diff options
Diffstat (limited to 'drivers/media/platform/mediatek')
114 files changed, 34118 insertions, 0 deletions
diff --git a/drivers/media/platform/mediatek/Kconfig b/drivers/media/platform/mediatek/Kconfig new file mode 100644 index 0000000000..84104e2cd0 --- /dev/null +++ b/drivers/media/platform/mediatek/Kconfig @@ -0,0 +1,9 @@ +# SPDX-License-Identifier: GPL-2.0-only + +comment "Mediatek media platform drivers" + +source "drivers/media/platform/mediatek/jpeg/Kconfig" +source "drivers/media/platform/mediatek/mdp/Kconfig" +source "drivers/media/platform/mediatek/vcodec/Kconfig" +source "drivers/media/platform/mediatek/vpu/Kconfig" +source "drivers/media/platform/mediatek/mdp3/Kconfig" diff --git a/drivers/media/platform/mediatek/Makefile b/drivers/media/platform/mediatek/Makefile new file mode 100644 index 0000000000..38e6ba917f --- /dev/null +++ b/drivers/media/platform/mediatek/Makefile @@ -0,0 +1,6 @@ +# SPDX-License-Identifier: GPL-2.0-only +obj-y += jpeg/ +obj-y += mdp/ +obj-y += vcodec/ +obj-y += vpu/ +obj-y += mdp3/ diff --git a/drivers/media/platform/mediatek/jpeg/Kconfig b/drivers/media/platform/mediatek/jpeg/Kconfig new file mode 100644 index 0000000000..39c4d1bc66 --- /dev/null +++ b/drivers/media/platform/mediatek/jpeg/Kconfig @@ -0,0 +1,16 @@ +# SPDX-License-Identifier: GPL-2.0-only +config VIDEO_MEDIATEK_JPEG + tristate "Mediatek JPEG Codec driver" + depends on V4L_MEM2MEM_DRIVERS + depends on MTK_IOMMU_V1 || MTK_IOMMU || COMPILE_TEST + depends on VIDEO_DEV + depends on ARCH_MEDIATEK || COMPILE_TEST + depends on MTK_SMI || (COMPILE_TEST && MTK_SMI=n) + select VIDEOBUF2_DMA_CONTIG + select V4L2_MEM2MEM_DEV + help + Mediatek jpeg codec driver provides HW capability to decode + JPEG format + + To compile this driver as a module, choose M here: the + module will be called mtk-jpeg diff --git a/drivers/media/platform/mediatek/jpeg/Makefile b/drivers/media/platform/mediatek/jpeg/Makefile new file mode 100644 index 0000000000..26e8485252 --- /dev/null +++ b/drivers/media/platform/mediatek/jpeg/Makefile @@ -0,0 +1,10 @@ +# SPDX-License-Identifier: GPL-2.0-only +obj-$(CONFIG_VIDEO_MEDIATEK_JPEG) += mtk_jpeg.o \ + mtk-jpeg-enc-hw.o \ + mtk-jpeg-dec-hw.o + +mtk_jpeg-y := mtk_jpeg_core.o \ + mtk_jpeg_dec_parse.o + +mtk-jpeg-enc-hw-y := mtk_jpeg_enc_hw.o +mtk-jpeg-dec-hw-y := mtk_jpeg_dec_hw.o diff --git a/drivers/media/platform/mediatek/jpeg/mtk_jpeg_core.c b/drivers/media/platform/mediatek/jpeg/mtk_jpeg_core.c new file mode 100644 index 0000000000..c3456c700c --- /dev/null +++ b/drivers/media/platform/mediatek/jpeg/mtk_jpeg_core.c @@ -0,0 +1,1965 @@ +// SPDX-License-Identifier: GPL-2.0-only +/* + * Copyright (c) 2016 MediaTek Inc. + * Author: Ming Hsiu Tsai <minghsiu.tsai@mediatek.com> + * Rick Chang <rick.chang@mediatek.com> + * Xia Jiang <xia.jiang@mediatek.com> + */ + +#include <linux/clk.h> +#include <linux/err.h> +#include <linux/interrupt.h> +#include <linux/io.h> +#include <linux/kernel.h> +#include <linux/module.h> +#include <linux/of_platform.h> +#include <linux/platform_device.h> +#include <linux/pm_runtime.h> +#include <linux/slab.h> +#include <linux/spinlock.h> +#include <media/v4l2-event.h> +#include <media/v4l2-mem2mem.h> +#include <media/v4l2-ioctl.h> +#include <media/videobuf2-core.h> +#include <media/videobuf2-dma-contig.h> + +#include "mtk_jpeg_enc_hw.h" +#include "mtk_jpeg_dec_hw.h" +#include "mtk_jpeg_core.h" +#include "mtk_jpeg_dec_parse.h" + +static struct mtk_jpeg_fmt mtk_jpeg_enc_formats[] = { + { + .fourcc = V4L2_PIX_FMT_JPEG, + .colplanes = 1, + .flags = MTK_JPEG_FMT_FLAG_CAPTURE, + }, + { + .fourcc = V4L2_PIX_FMT_NV12M, + .hw_format = JPEG_ENC_YUV_FORMAT_NV12, + .h_sample = {4, 4}, + .v_sample = {4, 2}, + .colplanes = 2, + .h_align = 4, + .v_align = 4, + .flags = MTK_JPEG_FMT_FLAG_OUTPUT, + }, + { + .fourcc = V4L2_PIX_FMT_NV21M, + .hw_format = JEPG_ENC_YUV_FORMAT_NV21, + .h_sample = {4, 4}, + .v_sample = {4, 2}, + .colplanes = 2, + .h_align = 4, + .v_align = 4, + .flags = MTK_JPEG_FMT_FLAG_OUTPUT, + }, + { + .fourcc = V4L2_PIX_FMT_YUYV, + .hw_format = JPEG_ENC_YUV_FORMAT_YUYV, + .h_sample = {8}, + .v_sample = {4}, + .colplanes = 1, + .h_align = 5, + .v_align = 3, + .flags = MTK_JPEG_FMT_FLAG_OUTPUT, + }, + { + .fourcc = V4L2_PIX_FMT_YVYU, + .hw_format = JPEG_ENC_YUV_FORMAT_YVYU, + .h_sample = {8}, + .v_sample = {4}, + .colplanes = 1, + .h_align = 5, + .v_align = 3, + .flags = MTK_JPEG_FMT_FLAG_OUTPUT, + }, +}; + +static struct mtk_jpeg_fmt mtk_jpeg_dec_formats[] = { + { + .fourcc = V4L2_PIX_FMT_JPEG, + .colplanes = 1, + .flags = MTK_JPEG_FMT_FLAG_OUTPUT, + }, + { + .fourcc = V4L2_PIX_FMT_YUV420M, + .h_sample = {4, 2, 2}, + .v_sample = {4, 2, 2}, + .colplanes = 3, + .h_align = 5, + .v_align = 4, + .flags = MTK_JPEG_FMT_FLAG_CAPTURE, + }, + { + .fourcc = V4L2_PIX_FMT_YUV422M, + .h_sample = {4, 2, 2}, + .v_sample = {4, 4, 4}, + .colplanes = 3, + .h_align = 5, + .v_align = 3, + .flags = MTK_JPEG_FMT_FLAG_CAPTURE, + }, +}; + +#define MTK_JPEG_ENC_NUM_FORMATS ARRAY_SIZE(mtk_jpeg_enc_formats) +#define MTK_JPEG_DEC_NUM_FORMATS ARRAY_SIZE(mtk_jpeg_dec_formats) +#define MTK_JPEG_MAX_RETRY_TIME 5000 + +enum { + MTK_JPEG_BUF_FLAGS_INIT = 0, + MTK_JPEG_BUF_FLAGS_LAST_FRAME = 1, +}; + +static int debug; +module_param(debug, int, 0644); + +static inline struct mtk_jpeg_ctx *ctrl_to_ctx(struct v4l2_ctrl *ctrl) +{ + return container_of(ctrl->handler, struct mtk_jpeg_ctx, ctrl_hdl); +} + +static inline struct mtk_jpeg_ctx *mtk_jpeg_fh_to_ctx(struct v4l2_fh *fh) +{ + return container_of(fh, struct mtk_jpeg_ctx, fh); +} + +static inline struct mtk_jpeg_src_buf *mtk_jpeg_vb2_to_srcbuf( + struct vb2_buffer *vb) +{ + return container_of(to_vb2_v4l2_buffer(vb), struct mtk_jpeg_src_buf, b); +} + +static int mtk_jpeg_querycap(struct file *file, void *priv, + struct v4l2_capability *cap) +{ + struct mtk_jpeg_dev *jpeg = video_drvdata(file); + + strscpy(cap->driver, jpeg->variant->dev_name, sizeof(cap->driver)); + strscpy(cap->card, jpeg->variant->dev_name, sizeof(cap->card)); + + return 0; +} + +static int vidioc_jpeg_enc_s_ctrl(struct v4l2_ctrl *ctrl) +{ + struct mtk_jpeg_ctx *ctx = ctrl_to_ctx(ctrl); + + switch (ctrl->id) { + case V4L2_CID_JPEG_RESTART_INTERVAL: + ctx->restart_interval = ctrl->val; + break; + case V4L2_CID_JPEG_COMPRESSION_QUALITY: + ctx->enc_quality = ctrl->val; + break; + case V4L2_CID_JPEG_ACTIVE_MARKER: + ctx->enable_exif = ctrl->val & V4L2_JPEG_ACTIVE_MARKER_APP1; + break; + } + + return 0; +} + +static const struct v4l2_ctrl_ops mtk_jpeg_enc_ctrl_ops = { + .s_ctrl = vidioc_jpeg_enc_s_ctrl, +}; + +static int mtk_jpeg_enc_ctrls_setup(struct mtk_jpeg_ctx *ctx) +{ + const struct v4l2_ctrl_ops *ops = &mtk_jpeg_enc_ctrl_ops; + struct v4l2_ctrl_handler *handler = &ctx->ctrl_hdl; + + v4l2_ctrl_handler_init(handler, 3); + + v4l2_ctrl_new_std(handler, ops, V4L2_CID_JPEG_RESTART_INTERVAL, 0, 100, + 1, 0); + v4l2_ctrl_new_std(handler, ops, V4L2_CID_JPEG_COMPRESSION_QUALITY, 48, + 100, 1, 90); + v4l2_ctrl_new_std(handler, ops, V4L2_CID_JPEG_ACTIVE_MARKER, 0, + V4L2_JPEG_ACTIVE_MARKER_APP1, 0, 0); + + if (handler->error) { + v4l2_ctrl_handler_free(&ctx->ctrl_hdl); + return handler->error; + } + + v4l2_ctrl_handler_setup(&ctx->ctrl_hdl); + + return 0; +} + +static int mtk_jpeg_enum_fmt(struct mtk_jpeg_fmt *mtk_jpeg_formats, int n, + struct v4l2_fmtdesc *f, u32 type) +{ + int i, num = 0; + + for (i = 0; i < n; ++i) { + if (mtk_jpeg_formats[i].flags & type) { + if (num == f->index) + break; + ++num; + } + } + + if (i >= n) + return -EINVAL; + + f->pixelformat = mtk_jpeg_formats[i].fourcc; + + return 0; +} + +static int mtk_jpeg_enum_fmt_vid_cap(struct file *file, void *priv, + struct v4l2_fmtdesc *f) +{ + struct mtk_jpeg_ctx *ctx = mtk_jpeg_fh_to_ctx(priv); + struct mtk_jpeg_dev *jpeg = ctx->jpeg; + + return mtk_jpeg_enum_fmt(jpeg->variant->formats, + jpeg->variant->num_formats, f, + MTK_JPEG_FMT_FLAG_CAPTURE); +} + +static int mtk_jpeg_enum_fmt_vid_out(struct file *file, void *priv, + struct v4l2_fmtdesc *f) +{ + struct mtk_jpeg_ctx *ctx = mtk_jpeg_fh_to_ctx(priv); + struct mtk_jpeg_dev *jpeg = ctx->jpeg; + + return mtk_jpeg_enum_fmt(jpeg->variant->formats, + jpeg->variant->num_formats, f, + MTK_JPEG_FMT_FLAG_OUTPUT); +} + +static struct mtk_jpeg_q_data *mtk_jpeg_get_q_data(struct mtk_jpeg_ctx *ctx, + enum v4l2_buf_type type) +{ + if (V4L2_TYPE_IS_OUTPUT(type)) + return &ctx->out_q; + return &ctx->cap_q; +} + +static struct mtk_jpeg_fmt * +mtk_jpeg_find_format(struct mtk_jpeg_fmt *mtk_jpeg_formats, int num_formats, + u32 pixelformat, unsigned int fmt_type) +{ + unsigned int k; + struct mtk_jpeg_fmt *fmt; + + for (k = 0; k < num_formats; k++) { + fmt = &mtk_jpeg_formats[k]; + + if (fmt->fourcc == pixelformat && fmt->flags & fmt_type) + return fmt; + } + + return NULL; +} + +static int mtk_jpeg_try_fmt_mplane(struct v4l2_pix_format_mplane *pix_mp, + struct mtk_jpeg_fmt *fmt) +{ + int i; + + pix_mp->field = V4L2_FIELD_NONE; + + pix_mp->num_planes = fmt->colplanes; + pix_mp->pixelformat = fmt->fourcc; + + if (fmt->fourcc == V4L2_PIX_FMT_JPEG) { + struct v4l2_plane_pix_format *pfmt = &pix_mp->plane_fmt[0]; + + pix_mp->height = clamp(pix_mp->height, MTK_JPEG_MIN_HEIGHT, + MTK_JPEG_MAX_HEIGHT); + pix_mp->width = clamp(pix_mp->width, MTK_JPEG_MIN_WIDTH, + MTK_JPEG_MAX_WIDTH); + + pfmt->bytesperline = 0; + /* Source size must be aligned to 128 */ + pfmt->sizeimage = round_up(pfmt->sizeimage, 128); + if (pfmt->sizeimage == 0) + pfmt->sizeimage = MTK_JPEG_DEFAULT_SIZEIMAGE; + return 0; + } + + /* other fourcc */ + pix_mp->height = clamp(round_up(pix_mp->height, fmt->v_align), + MTK_JPEG_MIN_HEIGHT, MTK_JPEG_MAX_HEIGHT); + pix_mp->width = clamp(round_up(pix_mp->width, fmt->h_align), + MTK_JPEG_MIN_WIDTH, MTK_JPEG_MAX_WIDTH); + + for (i = 0; i < fmt->colplanes; i++) { + struct v4l2_plane_pix_format *pfmt = &pix_mp->plane_fmt[i]; + u32 stride = pix_mp->width * fmt->h_sample[i] / 4; + u32 h = pix_mp->height * fmt->v_sample[i] / 4; + + pfmt->bytesperline = stride; + pfmt->sizeimage = stride * h; + } + return 0; +} + +static int mtk_jpeg_g_fmt_vid_mplane(struct file *file, void *priv, + struct v4l2_format *f) +{ + struct vb2_queue *vq; + struct mtk_jpeg_q_data *q_data = NULL; + struct v4l2_pix_format_mplane *pix_mp = &f->fmt.pix_mp; + struct mtk_jpeg_ctx *ctx = mtk_jpeg_fh_to_ctx(priv); + struct mtk_jpeg_dev *jpeg = ctx->jpeg; + int i; + + vq = v4l2_m2m_get_vq(ctx->fh.m2m_ctx, f->type); + if (!vq) + return -EINVAL; + + q_data = mtk_jpeg_get_q_data(ctx, f->type); + + pix_mp->width = q_data->pix_mp.width; + pix_mp->height = q_data->pix_mp.height; + pix_mp->field = V4L2_FIELD_NONE; + pix_mp->pixelformat = q_data->fmt->fourcc; + pix_mp->num_planes = q_data->fmt->colplanes; + pix_mp->colorspace = q_data->pix_mp.colorspace; + pix_mp->ycbcr_enc = q_data->pix_mp.ycbcr_enc; + pix_mp->xfer_func = q_data->pix_mp.xfer_func; + pix_mp->quantization = q_data->pix_mp.quantization; + + v4l2_dbg(1, debug, &jpeg->v4l2_dev, "(%d) g_fmt:%c%c%c%c wxh:%ux%u\n", + f->type, + (pix_mp->pixelformat & 0xff), + (pix_mp->pixelformat >> 8 & 0xff), + (pix_mp->pixelformat >> 16 & 0xff), + (pix_mp->pixelformat >> 24 & 0xff), + pix_mp->width, pix_mp->height); + + for (i = 0; i < pix_mp->num_planes; i++) { + struct v4l2_plane_pix_format *pfmt = &pix_mp->plane_fmt[i]; + + pfmt->bytesperline = q_data->pix_mp.plane_fmt[i].bytesperline; + pfmt->sizeimage = q_data->pix_mp.plane_fmt[i].sizeimage; + + v4l2_dbg(1, debug, &jpeg->v4l2_dev, + "plane[%d] bpl=%u, size=%u\n", + i, + pfmt->bytesperline, + pfmt->sizeimage); + } + return 0; +} + +static int mtk_jpeg_try_fmt_vid_cap_mplane(struct file *file, void *priv, + struct v4l2_format *f) +{ + struct mtk_jpeg_ctx *ctx = mtk_jpeg_fh_to_ctx(priv); + struct mtk_jpeg_dev *jpeg = ctx->jpeg; + struct mtk_jpeg_fmt *fmt; + + fmt = mtk_jpeg_find_format(jpeg->variant->formats, + jpeg->variant->num_formats, + f->fmt.pix_mp.pixelformat, + MTK_JPEG_FMT_FLAG_CAPTURE); + if (!fmt) + fmt = ctx->cap_q.fmt; + + v4l2_dbg(2, debug, &ctx->jpeg->v4l2_dev, "(%d) try_fmt:%c%c%c%c\n", + f->type, + (fmt->fourcc & 0xff), + (fmt->fourcc >> 8 & 0xff), + (fmt->fourcc >> 16 & 0xff), + (fmt->fourcc >> 24 & 0xff)); + + if (ctx->state != MTK_JPEG_INIT) { + mtk_jpeg_g_fmt_vid_mplane(file, priv, f); + return 0; + } + + return mtk_jpeg_try_fmt_mplane(&f->fmt.pix_mp, fmt); +} + +static int mtk_jpeg_try_fmt_vid_out_mplane(struct file *file, void *priv, + struct v4l2_format *f) +{ + struct mtk_jpeg_ctx *ctx = mtk_jpeg_fh_to_ctx(priv); + struct mtk_jpeg_dev *jpeg = ctx->jpeg; + struct mtk_jpeg_fmt *fmt; + + fmt = mtk_jpeg_find_format(jpeg->variant->formats, + jpeg->variant->num_formats, + f->fmt.pix_mp.pixelformat, + MTK_JPEG_FMT_FLAG_OUTPUT); + if (!fmt) + fmt = ctx->out_q.fmt; + + v4l2_dbg(2, debug, &ctx->jpeg->v4l2_dev, "(%d) try_fmt:%c%c%c%c\n", + f->type, + (fmt->fourcc & 0xff), + (fmt->fourcc >> 8 & 0xff), + (fmt->fourcc >> 16 & 0xff), + (fmt->fourcc >> 24 & 0xff)); + + if (ctx->state != MTK_JPEG_INIT) { + mtk_jpeg_g_fmt_vid_mplane(file, priv, f); + return 0; + } + + return mtk_jpeg_try_fmt_mplane(&f->fmt.pix_mp, fmt); +} + +static int mtk_jpeg_s_fmt_mplane(struct mtk_jpeg_ctx *ctx, + struct v4l2_format *f, unsigned int fmt_type) +{ + struct vb2_queue *vq; + struct mtk_jpeg_q_data *q_data = NULL; + struct v4l2_pix_format_mplane *pix_mp = &f->fmt.pix_mp; + struct mtk_jpeg_dev *jpeg = ctx->jpeg; + int i; + + vq = v4l2_m2m_get_vq(ctx->fh.m2m_ctx, f->type); + if (!vq) + return -EINVAL; + + q_data = mtk_jpeg_get_q_data(ctx, f->type); + + if (vb2_is_busy(vq)) { + v4l2_err(&jpeg->v4l2_dev, "queue busy\n"); + return -EBUSY; + } + + q_data->fmt = mtk_jpeg_find_format(jpeg->variant->formats, + jpeg->variant->num_formats, + pix_mp->pixelformat, fmt_type); + q_data->pix_mp.width = pix_mp->width; + q_data->pix_mp.height = pix_mp->height; + q_data->enc_crop_rect.width = pix_mp->width; + q_data->enc_crop_rect.height = pix_mp->height; + q_data->pix_mp.colorspace = V4L2_COLORSPACE_SRGB; + q_data->pix_mp.ycbcr_enc = V4L2_YCBCR_ENC_601; + q_data->pix_mp.xfer_func = V4L2_XFER_FUNC_SRGB; + q_data->pix_mp.quantization = V4L2_QUANTIZATION_FULL_RANGE; + + v4l2_dbg(1, debug, &jpeg->v4l2_dev, "(%d) s_fmt:%c%c%c%c wxh:%ux%u\n", + f->type, + (q_data->fmt->fourcc & 0xff), + (q_data->fmt->fourcc >> 8 & 0xff), + (q_data->fmt->fourcc >> 16 & 0xff), + (q_data->fmt->fourcc >> 24 & 0xff), + q_data->pix_mp.width, q_data->pix_mp.height); + + for (i = 0; i < q_data->fmt->colplanes; i++) { + q_data->pix_mp.plane_fmt[i].bytesperline = + pix_mp->plane_fmt[i].bytesperline; + q_data->pix_mp.plane_fmt[i].sizeimage = + pix_mp->plane_fmt[i].sizeimage; + + v4l2_dbg(1, debug, &jpeg->v4l2_dev, + "plane[%d] bpl=%u, size=%u\n", + i, q_data->pix_mp.plane_fmt[i].bytesperline, + q_data->pix_mp.plane_fmt[i].sizeimage); + } + + return 0; +} + +static int mtk_jpeg_s_fmt_vid_out_mplane(struct file *file, void *priv, + struct v4l2_format *f) +{ + int ret; + + ret = mtk_jpeg_try_fmt_vid_out_mplane(file, priv, f); + if (ret) + return ret; + + return mtk_jpeg_s_fmt_mplane(mtk_jpeg_fh_to_ctx(priv), f, + MTK_JPEG_FMT_FLAG_OUTPUT); +} + +static int mtk_jpeg_s_fmt_vid_cap_mplane(struct file *file, void *priv, + struct v4l2_format *f) +{ + int ret; + + ret = mtk_jpeg_try_fmt_vid_cap_mplane(file, priv, f); + if (ret) + return ret; + + return mtk_jpeg_s_fmt_mplane(mtk_jpeg_fh_to_ctx(priv), f, + MTK_JPEG_FMT_FLAG_CAPTURE); +} + +static void mtk_jpeg_queue_src_chg_event(struct mtk_jpeg_ctx *ctx) +{ + static const struct v4l2_event ev_src_ch = { + .type = V4L2_EVENT_SOURCE_CHANGE, + .u.src_change.changes = + V4L2_EVENT_SRC_CH_RESOLUTION, + }; + + v4l2_event_queue_fh(&ctx->fh, &ev_src_ch); +} + +static int mtk_jpeg_subscribe_event(struct v4l2_fh *fh, + const struct v4l2_event_subscription *sub) +{ + switch (sub->type) { + case V4L2_EVENT_SOURCE_CHANGE: + return v4l2_src_change_event_subscribe(fh, sub); + } + + return v4l2_ctrl_subscribe_event(fh, sub); +} + +static int mtk_jpeg_enc_g_selection(struct file *file, void *priv, + struct v4l2_selection *s) +{ + struct mtk_jpeg_ctx *ctx = mtk_jpeg_fh_to_ctx(priv); + + if (s->type != V4L2_BUF_TYPE_VIDEO_OUTPUT) + return -EINVAL; + + switch (s->target) { + case V4L2_SEL_TGT_CROP: + s->r = ctx->out_q.enc_crop_rect; + break; + case V4L2_SEL_TGT_CROP_BOUNDS: + case V4L2_SEL_TGT_CROP_DEFAULT: + s->r.width = ctx->out_q.pix_mp.width; + s->r.height = ctx->out_q.pix_mp.height; + s->r.left = 0; + s->r.top = 0; + break; + default: + return -EINVAL; + } + return 0; +} + +static int mtk_jpeg_dec_g_selection(struct file *file, void *priv, + struct v4l2_selection *s) +{ + struct mtk_jpeg_ctx *ctx = mtk_jpeg_fh_to_ctx(priv); + + if (s->type != V4L2_BUF_TYPE_VIDEO_CAPTURE) + return -EINVAL; + + switch (s->target) { + case V4L2_SEL_TGT_COMPOSE: + case V4L2_SEL_TGT_COMPOSE_DEFAULT: + s->r.width = ctx->out_q.pix_mp.width; + s->r.height = ctx->out_q.pix_mp.height; + s->r.left = 0; + s->r.top = 0; + break; + case V4L2_SEL_TGT_COMPOSE_BOUNDS: + case V4L2_SEL_TGT_COMPOSE_PADDED: + s->r.width = ctx->cap_q.pix_mp.width; + s->r.height = ctx->cap_q.pix_mp.height; + s->r.left = 0; + s->r.top = 0; + break; + default: + return -EINVAL; + } + return 0; +} + +static int mtk_jpeg_enc_s_selection(struct file *file, void *priv, + struct v4l2_selection *s) +{ + struct mtk_jpeg_ctx *ctx = mtk_jpeg_fh_to_ctx(priv); + + if (s->type != V4L2_BUF_TYPE_VIDEO_OUTPUT) + return -EINVAL; + + switch (s->target) { + case V4L2_SEL_TGT_CROP: + s->r.left = 0; + s->r.top = 0; + s->r.width = min(s->r.width, ctx->out_q.pix_mp.width); + s->r.height = min(s->r.height, ctx->out_q.pix_mp.height); + ctx->out_q.enc_crop_rect = s->r; + break; + default: + return -EINVAL; + } + + return 0; +} + +static int mtk_jpeg_qbuf(struct file *file, void *priv, struct v4l2_buffer *buf) +{ + struct v4l2_fh *fh = file->private_data; + struct mtk_jpeg_ctx *ctx = mtk_jpeg_fh_to_ctx(priv); + struct vb2_queue *vq; + struct vb2_buffer *vb; + struct mtk_jpeg_src_buf *jpeg_src_buf; + + if (buf->type != V4L2_BUF_TYPE_VIDEO_OUTPUT_MPLANE) + goto end; + + vq = v4l2_m2m_get_vq(fh->m2m_ctx, buf->type); + if (buf->index >= vq->num_buffers) { + dev_err(ctx->jpeg->dev, "buffer index out of range\n"); + return -EINVAL; + } + + vb = vq->bufs[buf->index]; + jpeg_src_buf = mtk_jpeg_vb2_to_srcbuf(vb); + jpeg_src_buf->bs_size = buf->m.planes[0].bytesused; + +end: + return v4l2_m2m_qbuf(file, fh->m2m_ctx, buf); +} + +static const struct v4l2_ioctl_ops mtk_jpeg_enc_ioctl_ops = { + .vidioc_querycap = mtk_jpeg_querycap, + .vidioc_enum_fmt_vid_cap = mtk_jpeg_enum_fmt_vid_cap, + .vidioc_enum_fmt_vid_out = mtk_jpeg_enum_fmt_vid_out, + .vidioc_try_fmt_vid_cap_mplane = mtk_jpeg_try_fmt_vid_cap_mplane, + .vidioc_try_fmt_vid_out_mplane = mtk_jpeg_try_fmt_vid_out_mplane, + .vidioc_g_fmt_vid_cap_mplane = mtk_jpeg_g_fmt_vid_mplane, + .vidioc_g_fmt_vid_out_mplane = mtk_jpeg_g_fmt_vid_mplane, + .vidioc_s_fmt_vid_cap_mplane = mtk_jpeg_s_fmt_vid_cap_mplane, + .vidioc_s_fmt_vid_out_mplane = mtk_jpeg_s_fmt_vid_out_mplane, + .vidioc_qbuf = v4l2_m2m_ioctl_qbuf, + .vidioc_subscribe_event = mtk_jpeg_subscribe_event, + .vidioc_g_selection = mtk_jpeg_enc_g_selection, + .vidioc_s_selection = mtk_jpeg_enc_s_selection, + + .vidioc_create_bufs = v4l2_m2m_ioctl_create_bufs, + .vidioc_prepare_buf = v4l2_m2m_ioctl_prepare_buf, + .vidioc_reqbufs = v4l2_m2m_ioctl_reqbufs, + .vidioc_querybuf = v4l2_m2m_ioctl_querybuf, + .vidioc_dqbuf = v4l2_m2m_ioctl_dqbuf, + .vidioc_expbuf = v4l2_m2m_ioctl_expbuf, + .vidioc_streamon = v4l2_m2m_ioctl_streamon, + .vidioc_streamoff = v4l2_m2m_ioctl_streamoff, + + .vidioc_unsubscribe_event = v4l2_event_unsubscribe, + + .vidioc_encoder_cmd = v4l2_m2m_ioctl_encoder_cmd, + .vidioc_try_encoder_cmd = v4l2_m2m_ioctl_try_encoder_cmd, +}; + +static const struct v4l2_ioctl_ops mtk_jpeg_dec_ioctl_ops = { + .vidioc_querycap = mtk_jpeg_querycap, + .vidioc_enum_fmt_vid_cap = mtk_jpeg_enum_fmt_vid_cap, + .vidioc_enum_fmt_vid_out = mtk_jpeg_enum_fmt_vid_out, + .vidioc_try_fmt_vid_cap_mplane = mtk_jpeg_try_fmt_vid_cap_mplane, + .vidioc_try_fmt_vid_out_mplane = mtk_jpeg_try_fmt_vid_out_mplane, + .vidioc_g_fmt_vid_cap_mplane = mtk_jpeg_g_fmt_vid_mplane, + .vidioc_g_fmt_vid_out_mplane = mtk_jpeg_g_fmt_vid_mplane, + .vidioc_s_fmt_vid_cap_mplane = mtk_jpeg_s_fmt_vid_cap_mplane, + .vidioc_s_fmt_vid_out_mplane = mtk_jpeg_s_fmt_vid_out_mplane, + .vidioc_qbuf = mtk_jpeg_qbuf, + .vidioc_subscribe_event = mtk_jpeg_subscribe_event, + .vidioc_g_selection = mtk_jpeg_dec_g_selection, + + .vidioc_create_bufs = v4l2_m2m_ioctl_create_bufs, + .vidioc_prepare_buf = v4l2_m2m_ioctl_prepare_buf, + .vidioc_reqbufs = v4l2_m2m_ioctl_reqbufs, + .vidioc_querybuf = v4l2_m2m_ioctl_querybuf, + .vidioc_dqbuf = v4l2_m2m_ioctl_dqbuf, + .vidioc_expbuf = v4l2_m2m_ioctl_expbuf, + .vidioc_streamon = v4l2_m2m_ioctl_streamon, + .vidioc_streamoff = v4l2_m2m_ioctl_streamoff, + + .vidioc_unsubscribe_event = v4l2_event_unsubscribe, + + .vidioc_decoder_cmd = v4l2_m2m_ioctl_decoder_cmd, + .vidioc_try_decoder_cmd = v4l2_m2m_ioctl_try_decoder_cmd, +}; + +static int mtk_jpeg_queue_setup(struct vb2_queue *q, + unsigned int *num_buffers, + unsigned int *num_planes, + unsigned int sizes[], + struct device *alloc_ctxs[]) +{ + struct mtk_jpeg_ctx *ctx = vb2_get_drv_priv(q); + struct mtk_jpeg_q_data *q_data = NULL; + struct mtk_jpeg_dev *jpeg = ctx->jpeg; + int i; + + v4l2_dbg(1, debug, &jpeg->v4l2_dev, "(%d) buf_req count=%u\n", + q->type, *num_buffers); + + q_data = mtk_jpeg_get_q_data(ctx, q->type); + if (!q_data) + return -EINVAL; + + if (*num_planes) { + for (i = 0; i < *num_planes; i++) + if (sizes[i] < q_data->pix_mp.plane_fmt[i].sizeimage) + return -EINVAL; + return 0; + } + + *num_planes = q_data->fmt->colplanes; + for (i = 0; i < q_data->fmt->colplanes; i++) { + sizes[i] = q_data->pix_mp.plane_fmt[i].sizeimage; + v4l2_dbg(1, debug, &jpeg->v4l2_dev, "sizeimage[%d]=%u\n", + i, sizes[i]); + } + + return 0; +} + +static int mtk_jpeg_buf_prepare(struct vb2_buffer *vb) +{ + struct mtk_jpeg_ctx *ctx = vb2_get_drv_priv(vb->vb2_queue); + struct mtk_jpeg_q_data *q_data = NULL; + struct v4l2_plane_pix_format plane_fmt = {}; + int i; + + q_data = mtk_jpeg_get_q_data(ctx, vb->vb2_queue->type); + if (!q_data) + return -EINVAL; + + for (i = 0; i < q_data->fmt->colplanes; i++) { + plane_fmt = q_data->pix_mp.plane_fmt[i]; + if (ctx->enable_exif && + q_data->fmt->fourcc == V4L2_PIX_FMT_JPEG) + vb2_set_plane_payload(vb, i, plane_fmt.sizeimage + + MTK_JPEG_MAX_EXIF_SIZE); + else + vb2_set_plane_payload(vb, i, plane_fmt.sizeimage); + } + + return 0; +} + +static bool mtk_jpeg_check_resolution_change(struct mtk_jpeg_ctx *ctx, + struct mtk_jpeg_dec_param *param) +{ + struct mtk_jpeg_dev *jpeg = ctx->jpeg; + struct mtk_jpeg_q_data *q_data; + + q_data = &ctx->out_q; + if (q_data->pix_mp.width != param->pic_w || + q_data->pix_mp.height != param->pic_h) { + v4l2_dbg(1, debug, &jpeg->v4l2_dev, "Picture size change\n"); + return true; + } + + q_data = &ctx->cap_q; + if (q_data->fmt != + mtk_jpeg_find_format(jpeg->variant->formats, + jpeg->variant->num_formats, param->dst_fourcc, + MTK_JPEG_FMT_FLAG_CAPTURE)) { + v4l2_dbg(1, debug, &jpeg->v4l2_dev, "format change\n"); + return true; + } + return false; +} + +static void mtk_jpeg_set_queue_data(struct mtk_jpeg_ctx *ctx, + struct mtk_jpeg_dec_param *param) +{ + struct mtk_jpeg_dev *jpeg = ctx->jpeg; + struct mtk_jpeg_q_data *q_data; + int i; + + q_data = &ctx->out_q; + q_data->pix_mp.width = param->pic_w; + q_data->pix_mp.height = param->pic_h; + + q_data = &ctx->cap_q; + q_data->pix_mp.width = param->dec_w; + q_data->pix_mp.height = param->dec_h; + q_data->fmt = mtk_jpeg_find_format(jpeg->variant->formats, + jpeg->variant->num_formats, + param->dst_fourcc, + MTK_JPEG_FMT_FLAG_CAPTURE); + + for (i = 0; i < q_data->fmt->colplanes; i++) { + q_data->pix_mp.plane_fmt[i].bytesperline = param->mem_stride[i]; + q_data->pix_mp.plane_fmt[i].sizeimage = param->comp_size[i]; + } + + v4l2_dbg(1, debug, &jpeg->v4l2_dev, + "set_parse cap:%c%c%c%c pic(%u, %u), buf(%u, %u)\n", + (param->dst_fourcc & 0xff), + (param->dst_fourcc >> 8 & 0xff), + (param->dst_fourcc >> 16 & 0xff), + (param->dst_fourcc >> 24 & 0xff), + param->pic_w, param->pic_h, + param->dec_w, param->dec_h); +} + +static void mtk_jpeg_enc_buf_queue(struct vb2_buffer *vb) +{ + struct mtk_jpeg_ctx *ctx = vb2_get_drv_priv(vb->vb2_queue); + struct mtk_jpeg_dev *jpeg = ctx->jpeg; + + v4l2_dbg(2, debug, &jpeg->v4l2_dev, "(%d) buf_q id=%d, vb=%p\n", + vb->vb2_queue->type, vb->index, vb); + + v4l2_m2m_buf_queue(ctx->fh.m2m_ctx, to_vb2_v4l2_buffer(vb)); +} + +static void mtk_jpeg_dec_buf_queue(struct vb2_buffer *vb) +{ + struct mtk_jpeg_ctx *ctx = vb2_get_drv_priv(vb->vb2_queue); + struct mtk_jpeg_dec_param *param; + struct mtk_jpeg_dev *jpeg = ctx->jpeg; + struct mtk_jpeg_src_buf *jpeg_src_buf; + bool header_valid; + + v4l2_dbg(2, debug, &jpeg->v4l2_dev, "(%d) buf_q id=%d, vb=%p\n", + vb->vb2_queue->type, vb->index, vb); + + if (vb->vb2_queue->type != V4L2_BUF_TYPE_VIDEO_OUTPUT_MPLANE) + goto end; + + jpeg_src_buf = mtk_jpeg_vb2_to_srcbuf(vb); + param = &jpeg_src_buf->dec_param; + memset(param, 0, sizeof(*param)); + + header_valid = mtk_jpeg_parse(param, (u8 *)vb2_plane_vaddr(vb, 0), + vb2_get_plane_payload(vb, 0)); + if (!header_valid) { + v4l2_err(&jpeg->v4l2_dev, "Header invalid.\n"); + vb2_buffer_done(vb, VB2_BUF_STATE_ERROR); + return; + } + + if (ctx->state == MTK_JPEG_INIT) { + struct vb2_queue *dst_vq = v4l2_m2m_get_vq( + ctx->fh.m2m_ctx, V4L2_BUF_TYPE_VIDEO_CAPTURE_MPLANE); + + mtk_jpeg_queue_src_chg_event(ctx); + mtk_jpeg_set_queue_data(ctx, param); + ctx->state = vb2_is_streaming(dst_vq) ? + MTK_JPEG_SOURCE_CHANGE : MTK_JPEG_RUNNING; + } +end: + v4l2_m2m_buf_queue(ctx->fh.m2m_ctx, to_vb2_v4l2_buffer(vb)); +} + +static struct vb2_v4l2_buffer *mtk_jpeg_buf_remove(struct mtk_jpeg_ctx *ctx, + enum v4l2_buf_type type) +{ + if (V4L2_TYPE_IS_OUTPUT(type)) + return v4l2_m2m_src_buf_remove(ctx->fh.m2m_ctx); + else + return v4l2_m2m_dst_buf_remove(ctx->fh.m2m_ctx); +} + +static void mtk_jpeg_enc_stop_streaming(struct vb2_queue *q) +{ + struct mtk_jpeg_ctx *ctx = vb2_get_drv_priv(q); + struct vb2_v4l2_buffer *vb; + + while ((vb = mtk_jpeg_buf_remove(ctx, q->type))) + v4l2_m2m_buf_done(vb, VB2_BUF_STATE_ERROR); +} + +static void mtk_jpeg_dec_stop_streaming(struct vb2_queue *q) +{ + struct mtk_jpeg_ctx *ctx = vb2_get_drv_priv(q); + struct vb2_v4l2_buffer *vb; + + /* + * STREAMOFF is an acknowledgment for source change event. + * Before STREAMOFF, we still have to return the old resolution and + * subsampling. Update capture queue when the stream is off. + */ + if (ctx->state == MTK_JPEG_SOURCE_CHANGE && + V4L2_TYPE_IS_CAPTURE(q->type)) { + struct mtk_jpeg_src_buf *src_buf; + + vb = v4l2_m2m_next_src_buf(ctx->fh.m2m_ctx); + src_buf = mtk_jpeg_vb2_to_srcbuf(&vb->vb2_buf); + mtk_jpeg_set_queue_data(ctx, &src_buf->dec_param); + ctx->state = MTK_JPEG_RUNNING; + } else if (V4L2_TYPE_IS_OUTPUT(q->type)) { + ctx->state = MTK_JPEG_INIT; + } + + while ((vb = mtk_jpeg_buf_remove(ctx, q->type))) + v4l2_m2m_buf_done(vb, VB2_BUF_STATE_ERROR); +} + +static const struct vb2_ops mtk_jpeg_dec_qops = { + .queue_setup = mtk_jpeg_queue_setup, + .buf_prepare = mtk_jpeg_buf_prepare, + .buf_queue = mtk_jpeg_dec_buf_queue, + .wait_prepare = vb2_ops_wait_prepare, + .wait_finish = vb2_ops_wait_finish, + .stop_streaming = mtk_jpeg_dec_stop_streaming, +}; + +static const struct vb2_ops mtk_jpeg_enc_qops = { + .queue_setup = mtk_jpeg_queue_setup, + .buf_prepare = mtk_jpeg_buf_prepare, + .buf_queue = mtk_jpeg_enc_buf_queue, + .wait_prepare = vb2_ops_wait_prepare, + .wait_finish = vb2_ops_wait_finish, + .stop_streaming = mtk_jpeg_enc_stop_streaming, +}; + +static void mtk_jpeg_set_dec_src(struct mtk_jpeg_ctx *ctx, + struct vb2_buffer *src_buf, + struct mtk_jpeg_bs *bs) +{ + bs->str_addr = vb2_dma_contig_plane_dma_addr(src_buf, 0); + bs->end_addr = bs->str_addr + + round_up(vb2_get_plane_payload(src_buf, 0), 16); + bs->size = round_up(vb2_plane_size(src_buf, 0), 128); +} + +static int mtk_jpeg_set_dec_dst(struct mtk_jpeg_ctx *ctx, + struct mtk_jpeg_dec_param *param, + struct vb2_buffer *dst_buf, + struct mtk_jpeg_fb *fb) +{ + int i; + + if (param->comp_num != dst_buf->num_planes) { + dev_err(ctx->jpeg->dev, "plane number mismatch (%u != %u)\n", + param->comp_num, dst_buf->num_planes); + return -EINVAL; + } + + for (i = 0; i < dst_buf->num_planes; i++) { + if (vb2_plane_size(dst_buf, i) < param->comp_size[i]) { + dev_err(ctx->jpeg->dev, + "buffer size is underflow (%lu < %u)\n", + vb2_plane_size(dst_buf, 0), + param->comp_size[i]); + return -EINVAL; + } + fb->plane_addr[i] = vb2_dma_contig_plane_dma_addr(dst_buf, i); + } + + return 0; +} + +static void mtk_jpeg_enc_device_run(void *priv) +{ + struct mtk_jpeg_ctx *ctx = priv; + struct mtk_jpeg_dev *jpeg = ctx->jpeg; + struct vb2_v4l2_buffer *src_buf, *dst_buf; + enum vb2_buffer_state buf_state = VB2_BUF_STATE_ERROR; + unsigned long flags; + int ret; + + src_buf = v4l2_m2m_next_src_buf(ctx->fh.m2m_ctx); + dst_buf = v4l2_m2m_next_dst_buf(ctx->fh.m2m_ctx); + + ret = pm_runtime_resume_and_get(jpeg->dev); + if (ret < 0) + goto enc_end; + + schedule_delayed_work(&jpeg->job_timeout_work, + msecs_to_jiffies(MTK_JPEG_HW_TIMEOUT_MSEC)); + + spin_lock_irqsave(&jpeg->hw_lock, flags); + + /* + * Resetting the hardware every frame is to ensure that all the + * registers are cleared. This is a hardware requirement. + */ + mtk_jpeg_enc_reset(jpeg->reg_base); + + mtk_jpeg_set_enc_src(ctx, jpeg->reg_base, &src_buf->vb2_buf); + mtk_jpeg_set_enc_dst(ctx, jpeg->reg_base, &dst_buf->vb2_buf); + mtk_jpeg_set_enc_params(ctx, jpeg->reg_base); + mtk_jpeg_enc_start(jpeg->reg_base); + spin_unlock_irqrestore(&jpeg->hw_lock, flags); + return; + +enc_end: + v4l2_m2m_src_buf_remove(ctx->fh.m2m_ctx); + v4l2_m2m_dst_buf_remove(ctx->fh.m2m_ctx); + v4l2_m2m_buf_done(src_buf, buf_state); + v4l2_m2m_buf_done(dst_buf, buf_state); + v4l2_m2m_job_finish(jpeg->m2m_dev, ctx->fh.m2m_ctx); +} + +static void mtk_jpeg_multicore_enc_device_run(void *priv) +{ + struct mtk_jpeg_ctx *ctx = priv; + struct mtk_jpeg_dev *jpeg = ctx->jpeg; + + queue_work(jpeg->workqueue, &ctx->jpeg_work); +} + +static void mtk_jpeg_multicore_dec_device_run(void *priv) +{ + struct mtk_jpeg_ctx *ctx = priv; + struct mtk_jpeg_dev *jpeg = ctx->jpeg; + + queue_work(jpeg->workqueue, &ctx->jpeg_work); +} + +static void mtk_jpeg_dec_device_run(void *priv) +{ + struct mtk_jpeg_ctx *ctx = priv; + struct mtk_jpeg_dev *jpeg = ctx->jpeg; + struct vb2_v4l2_buffer *src_buf, *dst_buf; + enum vb2_buffer_state buf_state = VB2_BUF_STATE_ERROR; + unsigned long flags; + struct mtk_jpeg_src_buf *jpeg_src_buf; + struct mtk_jpeg_bs bs; + struct mtk_jpeg_fb fb; + int ret; + + src_buf = v4l2_m2m_next_src_buf(ctx->fh.m2m_ctx); + dst_buf = v4l2_m2m_next_dst_buf(ctx->fh.m2m_ctx); + jpeg_src_buf = mtk_jpeg_vb2_to_srcbuf(&src_buf->vb2_buf); + + if (mtk_jpeg_check_resolution_change(ctx, &jpeg_src_buf->dec_param)) { + mtk_jpeg_queue_src_chg_event(ctx); + ctx->state = MTK_JPEG_SOURCE_CHANGE; + v4l2_m2m_job_finish(jpeg->m2m_dev, ctx->fh.m2m_ctx); + return; + } + + ret = pm_runtime_resume_and_get(jpeg->dev); + if (ret < 0) + goto dec_end; + + mtk_jpeg_set_dec_src(ctx, &src_buf->vb2_buf, &bs); + if (mtk_jpeg_set_dec_dst(ctx, &jpeg_src_buf->dec_param, &dst_buf->vb2_buf, &fb)) + goto dec_end; + + schedule_delayed_work(&jpeg->job_timeout_work, + msecs_to_jiffies(MTK_JPEG_HW_TIMEOUT_MSEC)); + + spin_lock_irqsave(&jpeg->hw_lock, flags); + mtk_jpeg_dec_reset(jpeg->reg_base); + mtk_jpeg_dec_set_config(jpeg->reg_base, + &jpeg_src_buf->dec_param, + jpeg_src_buf->bs_size, + &bs, + &fb); + mtk_jpeg_dec_start(jpeg->reg_base); + spin_unlock_irqrestore(&jpeg->hw_lock, flags); + return; + +dec_end: + v4l2_m2m_src_buf_remove(ctx->fh.m2m_ctx); + v4l2_m2m_dst_buf_remove(ctx->fh.m2m_ctx); + v4l2_m2m_buf_done(src_buf, buf_state); + v4l2_m2m_buf_done(dst_buf, buf_state); + v4l2_m2m_job_finish(jpeg->m2m_dev, ctx->fh.m2m_ctx); +} + +static int mtk_jpeg_dec_job_ready(void *priv) +{ + struct mtk_jpeg_ctx *ctx = priv; + + return (ctx->state == MTK_JPEG_RUNNING) ? 1 : 0; +} + +static const struct v4l2_m2m_ops mtk_jpeg_enc_m2m_ops = { + .device_run = mtk_jpeg_enc_device_run, +}; + +static const struct v4l2_m2m_ops mtk_jpeg_multicore_enc_m2m_ops = { + .device_run = mtk_jpeg_multicore_enc_device_run, +}; + +static const struct v4l2_m2m_ops mtk_jpeg_multicore_dec_m2m_ops = { + .device_run = mtk_jpeg_multicore_dec_device_run, +}; + +static const struct v4l2_m2m_ops mtk_jpeg_dec_m2m_ops = { + .device_run = mtk_jpeg_dec_device_run, + .job_ready = mtk_jpeg_dec_job_ready, +}; + +static int mtk_jpeg_queue_init(void *priv, struct vb2_queue *src_vq, + struct vb2_queue *dst_vq) +{ + struct mtk_jpeg_ctx *ctx = priv; + struct mtk_jpeg_dev *jpeg = ctx->jpeg; + int ret; + + src_vq->type = V4L2_BUF_TYPE_VIDEO_OUTPUT_MPLANE; + src_vq->io_modes = VB2_DMABUF | VB2_MMAP; + src_vq->drv_priv = ctx; + src_vq->buf_struct_size = sizeof(struct mtk_jpeg_src_buf); + src_vq->ops = jpeg->variant->qops; + src_vq->mem_ops = &vb2_dma_contig_memops; + src_vq->timestamp_flags = V4L2_BUF_FLAG_TIMESTAMP_COPY; + src_vq->lock = &ctx->jpeg->lock; + src_vq->dev = ctx->jpeg->dev; + ret = vb2_queue_init(src_vq); + if (ret) + return ret; + + dst_vq->type = V4L2_BUF_TYPE_VIDEO_CAPTURE_MPLANE; + dst_vq->io_modes = VB2_DMABUF | VB2_MMAP; + dst_vq->drv_priv = ctx; + dst_vq->buf_struct_size = sizeof(struct v4l2_m2m_buffer); + dst_vq->ops = jpeg->variant->qops; + dst_vq->mem_ops = &vb2_dma_contig_memops; + dst_vq->timestamp_flags = V4L2_BUF_FLAG_TIMESTAMP_COPY; + dst_vq->lock = &ctx->jpeg->lock; + dst_vq->dev = ctx->jpeg->dev; + ret = vb2_queue_init(dst_vq); + + return ret; +} + +static void mtk_jpeg_clk_on(struct mtk_jpeg_dev *jpeg) +{ + int ret; + + ret = clk_bulk_prepare_enable(jpeg->variant->num_clks, + jpeg->variant->clks); + if (ret) + dev_err(jpeg->dev, "Failed to open jpeg clk: %d\n", ret); +} + +static void mtk_jpeg_clk_off(struct mtk_jpeg_dev *jpeg) +{ + clk_bulk_disable_unprepare(jpeg->variant->num_clks, + jpeg->variant->clks); +} + +static void mtk_jpeg_set_default_params(struct mtk_jpeg_ctx *ctx) +{ + struct mtk_jpeg_q_data *q = &ctx->out_q; + struct mtk_jpeg_dev *jpeg = ctx->jpeg; + + ctx->fh.ctrl_handler = &ctx->ctrl_hdl; + q->pix_mp.colorspace = V4L2_COLORSPACE_SRGB; + q->pix_mp.ycbcr_enc = V4L2_YCBCR_ENC_601; + q->pix_mp.quantization = V4L2_QUANTIZATION_FULL_RANGE; + q->pix_mp.xfer_func = V4L2_XFER_FUNC_SRGB; + + q->fmt = mtk_jpeg_find_format(jpeg->variant->formats, + jpeg->variant->num_formats, + jpeg->variant->out_q_default_fourcc, + MTK_JPEG_FMT_FLAG_OUTPUT); + q->pix_mp.width = MTK_JPEG_MIN_WIDTH; + q->pix_mp.height = MTK_JPEG_MIN_HEIGHT; + mtk_jpeg_try_fmt_mplane(&q->pix_mp, q->fmt); + + q = &ctx->cap_q; + q->fmt = mtk_jpeg_find_format(jpeg->variant->formats, + jpeg->variant->num_formats, + jpeg->variant->cap_q_default_fourcc, + MTK_JPEG_FMT_FLAG_CAPTURE); + q->pix_mp.colorspace = V4L2_COLORSPACE_SRGB; + q->pix_mp.ycbcr_enc = V4L2_YCBCR_ENC_601; + q->pix_mp.quantization = V4L2_QUANTIZATION_FULL_RANGE; + q->pix_mp.xfer_func = V4L2_XFER_FUNC_SRGB; + q->pix_mp.width = MTK_JPEG_MIN_WIDTH; + q->pix_mp.height = MTK_JPEG_MIN_HEIGHT; + + mtk_jpeg_try_fmt_mplane(&q->pix_mp, q->fmt); +} + +static int mtk_jpeg_open(struct file *file) +{ + struct mtk_jpeg_dev *jpeg = video_drvdata(file); + struct video_device *vfd = video_devdata(file); + struct mtk_jpeg_ctx *ctx; + int ret = 0; + + ctx = kzalloc(sizeof(*ctx), GFP_KERNEL); + if (!ctx) + return -ENOMEM; + + if (mutex_lock_interruptible(&jpeg->lock)) { + ret = -ERESTARTSYS; + goto free; + } + + INIT_WORK(&ctx->jpeg_work, jpeg->variant->jpeg_worker); + INIT_LIST_HEAD(&ctx->dst_done_queue); + spin_lock_init(&ctx->done_queue_lock); + v4l2_fh_init(&ctx->fh, vfd); + file->private_data = &ctx->fh; + v4l2_fh_add(&ctx->fh); + + ctx->jpeg = jpeg; + ctx->fh.m2m_ctx = v4l2_m2m_ctx_init(jpeg->m2m_dev, ctx, + mtk_jpeg_queue_init); + if (IS_ERR(ctx->fh.m2m_ctx)) { + ret = PTR_ERR(ctx->fh.m2m_ctx); + goto error; + } + + if (jpeg->variant->cap_q_default_fourcc == V4L2_PIX_FMT_JPEG) { + ret = mtk_jpeg_enc_ctrls_setup(ctx); + if (ret) { + v4l2_err(&jpeg->v4l2_dev, "Failed to setup jpeg enc controls\n"); + goto error; + } + } else { + v4l2_ctrl_handler_init(&ctx->ctrl_hdl, 0); + } + + mtk_jpeg_set_default_params(ctx); + mutex_unlock(&jpeg->lock); + return 0; + +error: + v4l2_fh_del(&ctx->fh); + v4l2_fh_exit(&ctx->fh); + mutex_unlock(&jpeg->lock); +free: + kfree(ctx); + return ret; +} + +static int mtk_jpeg_release(struct file *file) +{ + struct mtk_jpeg_dev *jpeg = video_drvdata(file); + struct mtk_jpeg_ctx *ctx = mtk_jpeg_fh_to_ctx(file->private_data); + + mutex_lock(&jpeg->lock); + v4l2_m2m_ctx_release(ctx->fh.m2m_ctx); + v4l2_ctrl_handler_free(&ctx->ctrl_hdl); + v4l2_fh_del(&ctx->fh); + v4l2_fh_exit(&ctx->fh); + kfree(ctx); + mutex_unlock(&jpeg->lock); + return 0; +} + +static const struct v4l2_file_operations mtk_jpeg_fops = { + .owner = THIS_MODULE, + .open = mtk_jpeg_open, + .release = mtk_jpeg_release, + .poll = v4l2_m2m_fop_poll, + .unlocked_ioctl = video_ioctl2, + .mmap = v4l2_m2m_fop_mmap, +}; + +static void mtk_jpeg_job_timeout_work(struct work_struct *work) +{ + struct mtk_jpeg_dev *jpeg = container_of(work, struct mtk_jpeg_dev, + job_timeout_work.work); + struct mtk_jpeg_ctx *ctx; + struct vb2_v4l2_buffer *src_buf, *dst_buf; + + ctx = v4l2_m2m_get_curr_priv(jpeg->m2m_dev); + src_buf = v4l2_m2m_src_buf_remove(ctx->fh.m2m_ctx); + dst_buf = v4l2_m2m_dst_buf_remove(ctx->fh.m2m_ctx); + + jpeg->variant->hw_reset(jpeg->reg_base); + + pm_runtime_put(jpeg->dev); + + v4l2_m2m_buf_done(src_buf, VB2_BUF_STATE_ERROR); + v4l2_m2m_buf_done(dst_buf, VB2_BUF_STATE_ERROR); + v4l2_m2m_job_finish(jpeg->m2m_dev, ctx->fh.m2m_ctx); +} + +static int mtk_jpeg_single_core_init(struct platform_device *pdev, + struct mtk_jpeg_dev *jpeg_dev) +{ + struct mtk_jpeg_dev *jpeg = jpeg_dev; + int jpeg_irq, ret; + + INIT_DELAYED_WORK(&jpeg->job_timeout_work, + mtk_jpeg_job_timeout_work); + + jpeg->reg_base = devm_platform_ioremap_resource(pdev, 0); + if (IS_ERR(jpeg->reg_base)) { + ret = PTR_ERR(jpeg->reg_base); + return ret; + } + + jpeg_irq = platform_get_irq(pdev, 0); + if (jpeg_irq < 0) + return jpeg_irq; + + ret = devm_request_irq(&pdev->dev, + jpeg_irq, + jpeg->variant->irq_handler, + 0, + pdev->name, jpeg); + if (ret) { + dev_err(&pdev->dev, "Failed to request jpeg_irq %d (%d)\n", + jpeg_irq, ret); + return ret; + } + + ret = devm_clk_bulk_get(jpeg->dev, + jpeg->variant->num_clks, + jpeg->variant->clks); + if (ret) { + dev_err(&pdev->dev, "Failed to init clk\n"); + return ret; + } + + return 0; +} + +static int mtk_jpeg_probe(struct platform_device *pdev) +{ + struct mtk_jpeg_dev *jpeg; + struct device_node *child; + int num_child = 0; + int ret; + + jpeg = devm_kzalloc(&pdev->dev, sizeof(*jpeg), GFP_KERNEL); + if (!jpeg) + return -ENOMEM; + + mutex_init(&jpeg->lock); + spin_lock_init(&jpeg->hw_lock); + jpeg->dev = &pdev->dev; + jpeg->variant = of_device_get_match_data(jpeg->dev); + + platform_set_drvdata(pdev, jpeg); + + ret = devm_of_platform_populate(&pdev->dev); + if (ret) { + v4l2_err(&jpeg->v4l2_dev, "Master of platform populate failed."); + return -EINVAL; + } + + if (!jpeg->variant->multi_core) { + ret = mtk_jpeg_single_core_init(pdev, jpeg); + if (ret) { + v4l2_err(&jpeg->v4l2_dev, "mtk_jpeg_single_core_init failed."); + return -EINVAL; + } + } else { + init_waitqueue_head(&jpeg->hw_wq); + + for_each_child_of_node(pdev->dev.of_node, child) + num_child++; + + atomic_set(&jpeg->hw_rdy, num_child); + atomic_set(&jpeg->hw_index, 0); + + jpeg->workqueue = alloc_ordered_workqueue(MTK_JPEG_NAME, + WQ_MEM_RECLAIM + | WQ_FREEZABLE); + if (!jpeg->workqueue) + return -EINVAL; + } + + ret = v4l2_device_register(&pdev->dev, &jpeg->v4l2_dev); + if (ret) { + dev_err(&pdev->dev, "Failed to register v4l2 device\n"); + return -EINVAL; + } + + jpeg->m2m_dev = v4l2_m2m_init(jpeg->variant->m2m_ops); + + if (IS_ERR(jpeg->m2m_dev)) { + v4l2_err(&jpeg->v4l2_dev, "Failed to init mem2mem device\n"); + ret = PTR_ERR(jpeg->m2m_dev); + goto err_m2m_init; + } + + jpeg->vdev = video_device_alloc(); + if (!jpeg->vdev) { + ret = -ENOMEM; + goto err_vfd_jpeg_alloc; + } + snprintf(jpeg->vdev->name, sizeof(jpeg->vdev->name), + "%s", jpeg->variant->dev_name); + jpeg->vdev->fops = &mtk_jpeg_fops; + jpeg->vdev->ioctl_ops = jpeg->variant->ioctl_ops; + jpeg->vdev->minor = -1; + jpeg->vdev->release = video_device_release; + jpeg->vdev->lock = &jpeg->lock; + jpeg->vdev->v4l2_dev = &jpeg->v4l2_dev; + jpeg->vdev->vfl_dir = VFL_DIR_M2M; + jpeg->vdev->device_caps = V4L2_CAP_STREAMING | + V4L2_CAP_VIDEO_M2M_MPLANE; + + ret = video_register_device(jpeg->vdev, VFL_TYPE_VIDEO, -1); + if (ret) { + v4l2_err(&jpeg->v4l2_dev, "Failed to register video device\n"); + goto err_vfd_jpeg_register; + } + + video_set_drvdata(jpeg->vdev, jpeg); + v4l2_info(&jpeg->v4l2_dev, + "%s device registered as /dev/video%d (%d,%d)\n", + jpeg->variant->dev_name, jpeg->vdev->num, + VIDEO_MAJOR, jpeg->vdev->minor); + + pm_runtime_enable(&pdev->dev); + + return 0; + +err_vfd_jpeg_register: + video_device_release(jpeg->vdev); + +err_vfd_jpeg_alloc: + v4l2_m2m_release(jpeg->m2m_dev); + +err_m2m_init: + v4l2_device_unregister(&jpeg->v4l2_dev); + + return ret; +} + +static void mtk_jpeg_remove(struct platform_device *pdev) +{ + struct mtk_jpeg_dev *jpeg = platform_get_drvdata(pdev); + + pm_runtime_disable(&pdev->dev); + video_unregister_device(jpeg->vdev); + v4l2_m2m_release(jpeg->m2m_dev); + v4l2_device_unregister(&jpeg->v4l2_dev); +} + +static __maybe_unused int mtk_jpeg_pm_suspend(struct device *dev) +{ + struct mtk_jpeg_dev *jpeg = dev_get_drvdata(dev); + + mtk_jpeg_clk_off(jpeg); + + return 0; +} + +static __maybe_unused int mtk_jpeg_pm_resume(struct device *dev) +{ + struct mtk_jpeg_dev *jpeg = dev_get_drvdata(dev); + + mtk_jpeg_clk_on(jpeg); + + return 0; +} + +static __maybe_unused int mtk_jpeg_suspend(struct device *dev) +{ + struct mtk_jpeg_dev *jpeg = dev_get_drvdata(dev); + + v4l2_m2m_suspend(jpeg->m2m_dev); + return pm_runtime_force_suspend(dev); +} + +static __maybe_unused int mtk_jpeg_resume(struct device *dev) +{ + struct mtk_jpeg_dev *jpeg = dev_get_drvdata(dev); + int ret; + + ret = pm_runtime_force_resume(dev); + if (ret < 0) + return ret; + + v4l2_m2m_resume(jpeg->m2m_dev); + return ret; +} + +static const struct dev_pm_ops mtk_jpeg_pm_ops = { + SET_SYSTEM_SLEEP_PM_OPS(mtk_jpeg_suspend, mtk_jpeg_resume) + SET_RUNTIME_PM_OPS(mtk_jpeg_pm_suspend, mtk_jpeg_pm_resume, NULL) +}; + +static int mtk_jpegenc_get_hw(struct mtk_jpeg_ctx *ctx) +{ + struct mtk_jpegenc_comp_dev *comp_jpeg; + struct mtk_jpeg_dev *jpeg = ctx->jpeg; + unsigned long flags; + int hw_id = -1; + int i; + + spin_lock_irqsave(&jpeg->hw_lock, flags); + for (i = 0; i < MTK_JPEGENC_HW_MAX; i++) { + comp_jpeg = jpeg->enc_hw_dev[i]; + if (comp_jpeg->hw_state == MTK_JPEG_HW_IDLE) { + hw_id = i; + comp_jpeg->hw_state = MTK_JPEG_HW_BUSY; + break; + } + } + spin_unlock_irqrestore(&jpeg->hw_lock, flags); + + return hw_id; +} + +static int mtk_jpegenc_set_hw_param(struct mtk_jpeg_ctx *ctx, + int hw_id, + struct vb2_v4l2_buffer *src_buf, + struct vb2_v4l2_buffer *dst_buf) +{ + struct mtk_jpegenc_comp_dev *jpeg = ctx->jpeg->enc_hw_dev[hw_id]; + + jpeg->hw_param.curr_ctx = ctx; + jpeg->hw_param.src_buffer = src_buf; + jpeg->hw_param.dst_buffer = dst_buf; + + return 0; +} + +static int mtk_jpegenc_put_hw(struct mtk_jpeg_dev *jpeg, int hw_id) +{ + unsigned long flags; + + spin_lock_irqsave(&jpeg->hw_lock, flags); + jpeg->enc_hw_dev[hw_id]->hw_state = MTK_JPEG_HW_IDLE; + spin_unlock_irqrestore(&jpeg->hw_lock, flags); + + return 0; +} + +static int mtk_jpegdec_get_hw(struct mtk_jpeg_ctx *ctx) +{ + struct mtk_jpegdec_comp_dev *comp_jpeg; + struct mtk_jpeg_dev *jpeg = ctx->jpeg; + unsigned long flags; + int hw_id = -1; + int i; + + spin_lock_irqsave(&jpeg->hw_lock, flags); + for (i = 0; i < MTK_JPEGDEC_HW_MAX; i++) { + comp_jpeg = jpeg->dec_hw_dev[i]; + if (comp_jpeg->hw_state == MTK_JPEG_HW_IDLE) { + hw_id = i; + comp_jpeg->hw_state = MTK_JPEG_HW_BUSY; + break; + } + } + spin_unlock_irqrestore(&jpeg->hw_lock, flags); + + return hw_id; +} + +static int mtk_jpegdec_put_hw(struct mtk_jpeg_dev *jpeg, int hw_id) +{ + unsigned long flags; + + spin_lock_irqsave(&jpeg->hw_lock, flags); + jpeg->dec_hw_dev[hw_id]->hw_state = + MTK_JPEG_HW_IDLE; + spin_unlock_irqrestore(&jpeg->hw_lock, flags); + + return 0; +} + +static int mtk_jpegdec_set_hw_param(struct mtk_jpeg_ctx *ctx, + int hw_id, + struct vb2_v4l2_buffer *src_buf, + struct vb2_v4l2_buffer *dst_buf) +{ + struct mtk_jpegdec_comp_dev *jpeg = + ctx->jpeg->dec_hw_dev[hw_id]; + + jpeg->hw_param.curr_ctx = ctx; + jpeg->hw_param.src_buffer = src_buf; + jpeg->hw_param.dst_buffer = dst_buf; + + return 0; +} + +static irqreturn_t mtk_jpeg_enc_done(struct mtk_jpeg_dev *jpeg) +{ + struct mtk_jpeg_ctx *ctx; + struct vb2_v4l2_buffer *src_buf, *dst_buf; + enum vb2_buffer_state buf_state = VB2_BUF_STATE_ERROR; + u32 result_size; + + ctx = v4l2_m2m_get_curr_priv(jpeg->m2m_dev); + if (!ctx) { + v4l2_err(&jpeg->v4l2_dev, "Context is NULL\n"); + return IRQ_HANDLED; + } + + src_buf = v4l2_m2m_src_buf_remove(ctx->fh.m2m_ctx); + dst_buf = v4l2_m2m_dst_buf_remove(ctx->fh.m2m_ctx); + + result_size = mtk_jpeg_enc_get_file_size(jpeg->reg_base); + vb2_set_plane_payload(&dst_buf->vb2_buf, 0, result_size); + + buf_state = VB2_BUF_STATE_DONE; + + v4l2_m2m_buf_done(src_buf, buf_state); + v4l2_m2m_buf_done(dst_buf, buf_state); + v4l2_m2m_job_finish(jpeg->m2m_dev, ctx->fh.m2m_ctx); + pm_runtime_put(ctx->jpeg->dev); + return IRQ_HANDLED; +} + +static void mtk_jpegenc_worker(struct work_struct *work) +{ + struct mtk_jpegenc_comp_dev *comp_jpeg[MTK_JPEGENC_HW_MAX]; + enum vb2_buffer_state buf_state = VB2_BUF_STATE_ERROR; + struct mtk_jpeg_src_buf *jpeg_dst_buf; + struct vb2_v4l2_buffer *src_buf, *dst_buf; + int ret, i, hw_id = 0; + unsigned long flags; + + struct mtk_jpeg_ctx *ctx = container_of(work, + struct mtk_jpeg_ctx, + jpeg_work); + struct mtk_jpeg_dev *jpeg = ctx->jpeg; + + for (i = 0; i < MTK_JPEGENC_HW_MAX; i++) + comp_jpeg[i] = jpeg->enc_hw_dev[i]; + i = 0; + +retry_select: + hw_id = mtk_jpegenc_get_hw(ctx); + if (hw_id < 0) { + ret = wait_event_interruptible(jpeg->hw_wq, + atomic_read(&jpeg->hw_rdy) > 0); + if (ret != 0 || (i++ > MTK_JPEG_MAX_RETRY_TIME)) { + dev_err(jpeg->dev, "%s : %d, all HW are busy\n", + __func__, __LINE__); + v4l2_m2m_job_finish(jpeg->m2m_dev, ctx->fh.m2m_ctx); + return; + } + + goto retry_select; + } + + atomic_dec(&jpeg->hw_rdy); + src_buf = v4l2_m2m_next_src_buf(ctx->fh.m2m_ctx); + if (!src_buf) + goto getbuf_fail; + + dst_buf = v4l2_m2m_next_dst_buf(ctx->fh.m2m_ctx); + if (!dst_buf) + goto getbuf_fail; + + v4l2_m2m_buf_copy_metadata(src_buf, dst_buf, true); + + mtk_jpegenc_set_hw_param(ctx, hw_id, src_buf, dst_buf); + ret = pm_runtime_get_sync(comp_jpeg[hw_id]->dev); + if (ret < 0) { + dev_err(jpeg->dev, "%s : %d, pm_runtime_get_sync fail !!!\n", + __func__, __LINE__); + goto enc_end; + } + + ret = clk_prepare_enable(comp_jpeg[hw_id]->venc_clk.clks->clk); + if (ret) { + dev_err(jpeg->dev, "%s : %d, jpegenc clk_prepare_enable fail\n", + __func__, __LINE__); + goto enc_end; + } + + v4l2_m2m_src_buf_remove(ctx->fh.m2m_ctx); + v4l2_m2m_dst_buf_remove(ctx->fh.m2m_ctx); + + schedule_delayed_work(&comp_jpeg[hw_id]->job_timeout_work, + msecs_to_jiffies(MTK_JPEG_HW_TIMEOUT_MSEC)); + + spin_lock_irqsave(&comp_jpeg[hw_id]->hw_lock, flags); + jpeg_dst_buf = mtk_jpeg_vb2_to_srcbuf(&dst_buf->vb2_buf); + jpeg_dst_buf->curr_ctx = ctx; + jpeg_dst_buf->frame_num = ctx->total_frame_num; + ctx->total_frame_num++; + mtk_jpeg_enc_reset(comp_jpeg[hw_id]->reg_base); + mtk_jpeg_set_enc_dst(ctx, + comp_jpeg[hw_id]->reg_base, + &dst_buf->vb2_buf); + mtk_jpeg_set_enc_src(ctx, + comp_jpeg[hw_id]->reg_base, + &src_buf->vb2_buf); + mtk_jpeg_set_enc_params(ctx, comp_jpeg[hw_id]->reg_base); + mtk_jpeg_enc_start(comp_jpeg[hw_id]->reg_base); + v4l2_m2m_job_finish(jpeg->m2m_dev, ctx->fh.m2m_ctx); + spin_unlock_irqrestore(&comp_jpeg[hw_id]->hw_lock, flags); + + return; + +enc_end: + v4l2_m2m_src_buf_remove(ctx->fh.m2m_ctx); + v4l2_m2m_dst_buf_remove(ctx->fh.m2m_ctx); + v4l2_m2m_buf_done(src_buf, buf_state); + v4l2_m2m_buf_done(dst_buf, buf_state); +getbuf_fail: + atomic_inc(&jpeg->hw_rdy); + mtk_jpegenc_put_hw(jpeg, hw_id); + v4l2_m2m_job_finish(jpeg->m2m_dev, ctx->fh.m2m_ctx); +} + +static void mtk_jpegdec_worker(struct work_struct *work) +{ + struct mtk_jpeg_ctx *ctx = container_of(work, struct mtk_jpeg_ctx, + jpeg_work); + struct mtk_jpegdec_comp_dev *comp_jpeg[MTK_JPEGDEC_HW_MAX]; + enum vb2_buffer_state buf_state = VB2_BUF_STATE_ERROR; + struct mtk_jpeg_src_buf *jpeg_src_buf, *jpeg_dst_buf; + struct vb2_v4l2_buffer *src_buf, *dst_buf; + struct mtk_jpeg_dev *jpeg = ctx->jpeg; + int ret, i, hw_id = 0; + struct mtk_jpeg_bs bs; + struct mtk_jpeg_fb fb; + unsigned long flags; + + for (i = 0; i < MTK_JPEGDEC_HW_MAX; i++) + comp_jpeg[i] = jpeg->dec_hw_dev[i]; + i = 0; + +retry_select: + hw_id = mtk_jpegdec_get_hw(ctx); + if (hw_id < 0) { + ret = wait_event_interruptible_timeout(jpeg->hw_wq, + atomic_read(&jpeg->hw_rdy) > 0, + MTK_JPEG_HW_TIMEOUT_MSEC); + if (ret != 0 || (i++ > MTK_JPEG_MAX_RETRY_TIME)) { + dev_err(jpeg->dev, "%s : %d, all HW are busy\n", + __func__, __LINE__); + v4l2_m2m_job_finish(jpeg->m2m_dev, ctx->fh.m2m_ctx); + return; + } + + goto retry_select; + } + + atomic_dec(&jpeg->hw_rdy); + src_buf = v4l2_m2m_next_src_buf(ctx->fh.m2m_ctx); + if (!src_buf) + goto getbuf_fail; + + dst_buf = v4l2_m2m_next_dst_buf(ctx->fh.m2m_ctx); + if (!dst_buf) + goto getbuf_fail; + + v4l2_m2m_buf_copy_metadata(src_buf, dst_buf, true); + jpeg_src_buf = mtk_jpeg_vb2_to_srcbuf(&src_buf->vb2_buf); + jpeg_dst_buf = mtk_jpeg_vb2_to_srcbuf(&dst_buf->vb2_buf); + + if (mtk_jpeg_check_resolution_change(ctx, + &jpeg_src_buf->dec_param)) { + mtk_jpeg_queue_src_chg_event(ctx); + ctx->state = MTK_JPEG_SOURCE_CHANGE; + goto getbuf_fail; + } + + jpeg_src_buf->curr_ctx = ctx; + jpeg_src_buf->frame_num = ctx->total_frame_num; + jpeg_dst_buf->curr_ctx = ctx; + jpeg_dst_buf->frame_num = ctx->total_frame_num; + + mtk_jpegdec_set_hw_param(ctx, hw_id, src_buf, dst_buf); + ret = pm_runtime_get_sync(comp_jpeg[hw_id]->dev); + if (ret < 0) { + dev_err(jpeg->dev, "%s : %d, pm_runtime_get_sync fail !!!\n", + __func__, __LINE__); + goto dec_end; + } + + ret = clk_prepare_enable(comp_jpeg[hw_id]->jdec_clk.clks->clk); + if (ret) { + dev_err(jpeg->dev, "%s : %d, jpegdec clk_prepare_enable fail\n", + __func__, __LINE__); + goto clk_end; + } + + v4l2_m2m_src_buf_remove(ctx->fh.m2m_ctx); + v4l2_m2m_dst_buf_remove(ctx->fh.m2m_ctx); + + mtk_jpeg_set_dec_src(ctx, &src_buf->vb2_buf, &bs); + if (mtk_jpeg_set_dec_dst(ctx, + &jpeg_src_buf->dec_param, + &dst_buf->vb2_buf, &fb)) { + dev_err(jpeg->dev, "%s : %d, mtk_jpeg_set_dec_dst fail\n", + __func__, __LINE__); + goto setdst_end; + } + + schedule_delayed_work(&comp_jpeg[hw_id]->job_timeout_work, + msecs_to_jiffies(MTK_JPEG_HW_TIMEOUT_MSEC)); + + spin_lock_irqsave(&comp_jpeg[hw_id]->hw_lock, flags); + ctx->total_frame_num++; + mtk_jpeg_dec_reset(comp_jpeg[hw_id]->reg_base); + mtk_jpeg_dec_set_config(comp_jpeg[hw_id]->reg_base, + &jpeg_src_buf->dec_param, + jpeg_src_buf->bs_size, + &bs, + &fb); + mtk_jpeg_dec_start(comp_jpeg[hw_id]->reg_base); + v4l2_m2m_job_finish(jpeg->m2m_dev, ctx->fh.m2m_ctx); + spin_unlock_irqrestore(&comp_jpeg[hw_id]->hw_lock, flags); + + return; + +setdst_end: + clk_disable_unprepare(comp_jpeg[hw_id]->jdec_clk.clks->clk); +clk_end: + pm_runtime_put(comp_jpeg[hw_id]->dev); +dec_end: + v4l2_m2m_src_buf_remove(ctx->fh.m2m_ctx); + v4l2_m2m_dst_buf_remove(ctx->fh.m2m_ctx); + v4l2_m2m_buf_done(src_buf, buf_state); + v4l2_m2m_buf_done(dst_buf, buf_state); +getbuf_fail: + atomic_inc(&jpeg->hw_rdy); + mtk_jpegdec_put_hw(jpeg, hw_id); + v4l2_m2m_job_finish(jpeg->m2m_dev, ctx->fh.m2m_ctx); +} + +static irqreturn_t mtk_jpeg_enc_irq(int irq, void *priv) +{ + struct mtk_jpeg_dev *jpeg = priv; + u32 irq_status; + irqreturn_t ret = IRQ_NONE; + + cancel_delayed_work(&jpeg->job_timeout_work); + + irq_status = readl(jpeg->reg_base + JPEG_ENC_INT_STS) & + JPEG_ENC_INT_STATUS_MASK_ALLIRQ; + if (irq_status) + writel(0, jpeg->reg_base + JPEG_ENC_INT_STS); + + if (!(irq_status & JPEG_ENC_INT_STATUS_DONE)) + return ret; + + ret = mtk_jpeg_enc_done(jpeg); + return ret; +} + +static irqreturn_t mtk_jpeg_dec_irq(int irq, void *priv) +{ + struct mtk_jpeg_dev *jpeg = priv; + struct mtk_jpeg_ctx *ctx; + struct vb2_v4l2_buffer *src_buf, *dst_buf; + struct mtk_jpeg_src_buf *jpeg_src_buf; + enum vb2_buffer_state buf_state = VB2_BUF_STATE_ERROR; + u32 dec_irq_ret; + u32 dec_ret; + int i; + + cancel_delayed_work(&jpeg->job_timeout_work); + + dec_ret = mtk_jpeg_dec_get_int_status(jpeg->reg_base); + dec_irq_ret = mtk_jpeg_dec_enum_result(dec_ret); + ctx = v4l2_m2m_get_curr_priv(jpeg->m2m_dev); + if (!ctx) { + v4l2_err(&jpeg->v4l2_dev, "Context is NULL\n"); + return IRQ_HANDLED; + } + + src_buf = v4l2_m2m_src_buf_remove(ctx->fh.m2m_ctx); + dst_buf = v4l2_m2m_dst_buf_remove(ctx->fh.m2m_ctx); + jpeg_src_buf = mtk_jpeg_vb2_to_srcbuf(&src_buf->vb2_buf); + + if (dec_irq_ret >= MTK_JPEG_DEC_RESULT_UNDERFLOW) + mtk_jpeg_dec_reset(jpeg->reg_base); + + if (dec_irq_ret != MTK_JPEG_DEC_RESULT_EOF_DONE) { + dev_err(jpeg->dev, "decode failed\n"); + goto dec_end; + } + + for (i = 0; i < dst_buf->vb2_buf.num_planes; i++) + vb2_set_plane_payload(&dst_buf->vb2_buf, i, + jpeg_src_buf->dec_param.comp_size[i]); + + buf_state = VB2_BUF_STATE_DONE; + +dec_end: + v4l2_m2m_buf_done(src_buf, buf_state); + v4l2_m2m_buf_done(dst_buf, buf_state); + v4l2_m2m_job_finish(jpeg->m2m_dev, ctx->fh.m2m_ctx); + pm_runtime_put(ctx->jpeg->dev); + return IRQ_HANDLED; +} + +static struct clk_bulk_data mtk_jpeg_clocks[] = { + { .id = "jpgenc" }, +}; + +static struct clk_bulk_data mt8173_jpeg_dec_clocks[] = { + { .id = "jpgdec-smi" }, + { .id = "jpgdec" }, +}; + +static const struct mtk_jpeg_variant mt8173_jpeg_drvdata = { + .clks = mt8173_jpeg_dec_clocks, + .num_clks = ARRAY_SIZE(mt8173_jpeg_dec_clocks), + .formats = mtk_jpeg_dec_formats, + .num_formats = MTK_JPEG_DEC_NUM_FORMATS, + .qops = &mtk_jpeg_dec_qops, + .irq_handler = mtk_jpeg_dec_irq, + .hw_reset = mtk_jpeg_dec_reset, + .m2m_ops = &mtk_jpeg_dec_m2m_ops, + .dev_name = "mtk-jpeg-dec", + .ioctl_ops = &mtk_jpeg_dec_ioctl_ops, + .out_q_default_fourcc = V4L2_PIX_FMT_JPEG, + .cap_q_default_fourcc = V4L2_PIX_FMT_YUV420M, +}; + +static const struct mtk_jpeg_variant mtk_jpeg_drvdata = { + .clks = mtk_jpeg_clocks, + .num_clks = ARRAY_SIZE(mtk_jpeg_clocks), + .formats = mtk_jpeg_enc_formats, + .num_formats = MTK_JPEG_ENC_NUM_FORMATS, + .qops = &mtk_jpeg_enc_qops, + .irq_handler = mtk_jpeg_enc_irq, + .hw_reset = mtk_jpeg_enc_reset, + .m2m_ops = &mtk_jpeg_enc_m2m_ops, + .dev_name = "mtk-jpeg-enc", + .ioctl_ops = &mtk_jpeg_enc_ioctl_ops, + .out_q_default_fourcc = V4L2_PIX_FMT_YUYV, + .cap_q_default_fourcc = V4L2_PIX_FMT_JPEG, + .multi_core = false, +}; + +static struct mtk_jpeg_variant mtk8195_jpegenc_drvdata = { + .formats = mtk_jpeg_enc_formats, + .num_formats = MTK_JPEG_ENC_NUM_FORMATS, + .qops = &mtk_jpeg_enc_qops, + .m2m_ops = &mtk_jpeg_multicore_enc_m2m_ops, + .dev_name = "mtk-jpeg-enc", + .ioctl_ops = &mtk_jpeg_enc_ioctl_ops, + .out_q_default_fourcc = V4L2_PIX_FMT_YUYV, + .cap_q_default_fourcc = V4L2_PIX_FMT_JPEG, + .multi_core = true, + .jpeg_worker = mtk_jpegenc_worker, +}; + +static const struct mtk_jpeg_variant mtk8195_jpegdec_drvdata = { + .formats = mtk_jpeg_dec_formats, + .num_formats = MTK_JPEG_DEC_NUM_FORMATS, + .qops = &mtk_jpeg_dec_qops, + .m2m_ops = &mtk_jpeg_multicore_dec_m2m_ops, + .dev_name = "mtk-jpeg-dec", + .ioctl_ops = &mtk_jpeg_dec_ioctl_ops, + .out_q_default_fourcc = V4L2_PIX_FMT_JPEG, + .cap_q_default_fourcc = V4L2_PIX_FMT_YUV420M, + .multi_core = true, + .jpeg_worker = mtk_jpegdec_worker, +}; + +static const struct of_device_id mtk_jpeg_match[] = { + { + .compatible = "mediatek,mt8173-jpgdec", + .data = &mt8173_jpeg_drvdata, + }, + { + .compatible = "mediatek,mt2701-jpgdec", + .data = &mt8173_jpeg_drvdata, + }, + { + .compatible = "mediatek,mtk-jpgenc", + .data = &mtk_jpeg_drvdata, + }, + { + .compatible = "mediatek,mt8195-jpgenc", + .data = &mtk8195_jpegenc_drvdata, + }, + { + .compatible = "mediatek,mt8195-jpgdec", + .data = &mtk8195_jpegdec_drvdata, + }, + {}, +}; + +MODULE_DEVICE_TABLE(of, mtk_jpeg_match); + +static struct platform_driver mtk_jpeg_driver = { + .probe = mtk_jpeg_probe, + .remove_new = mtk_jpeg_remove, + .driver = { + .name = MTK_JPEG_NAME, + .of_match_table = mtk_jpeg_match, + .pm = &mtk_jpeg_pm_ops, + }, +}; + +module_platform_driver(mtk_jpeg_driver); + +MODULE_DESCRIPTION("MediaTek JPEG codec driver"); +MODULE_LICENSE("GPL v2"); diff --git a/drivers/media/platform/mediatek/jpeg/mtk_jpeg_core.h b/drivers/media/platform/mediatek/jpeg/mtk_jpeg_core.h new file mode 100644 index 0000000000..8ba6e757e1 --- /dev/null +++ b/drivers/media/platform/mediatek/jpeg/mtk_jpeg_core.h @@ -0,0 +1,306 @@ +/* SPDX-License-Identifier: GPL-2.0-only */ +/* + * Copyright (c) 2016 MediaTek Inc. + * Author: Ming Hsiu Tsai <minghsiu.tsai@mediatek.com> + * Rick Chang <rick.chang@mediatek.com> + * Xia Jiang <xia.jiang@mediatek.com> + */ + +#ifndef _MTK_JPEG_CORE_H +#define _MTK_JPEG_CORE_H + +#include <linux/clk.h> +#include <linux/interrupt.h> +#include <media/v4l2-ctrls.h> +#include <media/v4l2-device.h> +#include <media/v4l2-fh.h> +#include <media/videobuf2-v4l2.h> + +#include "mtk_jpeg_dec_hw.h" + +#define MTK_JPEG_NAME "mtk-jpeg" + +#define MTK_JPEG_FMT_FLAG_OUTPUT BIT(0) +#define MTK_JPEG_FMT_FLAG_CAPTURE BIT(1) + +#define MTK_JPEG_MIN_WIDTH 32U +#define MTK_JPEG_MIN_HEIGHT 32U +#define MTK_JPEG_MAX_WIDTH 65535U +#define MTK_JPEG_MAX_HEIGHT 65535U + +#define MTK_JPEG_DEFAULT_SIZEIMAGE (1 * 1024 * 1024) + +#define MTK_JPEG_HW_TIMEOUT_MSEC 1000 + +#define MTK_JPEG_MAX_EXIF_SIZE (64 * 1024) + +/** + * enum mtk_jpeg_ctx_state - states of the context state machine + * @MTK_JPEG_INIT: current state is initialized + * @MTK_JPEG_RUNNING: current state is running + * @MTK_JPEG_SOURCE_CHANGE: current state is source resolution change + */ +enum mtk_jpeg_ctx_state { + MTK_JPEG_INIT = 0, + MTK_JPEG_RUNNING, + MTK_JPEG_SOURCE_CHANGE, +}; + +/** + * struct mtk_jpeg_variant - mtk jpeg driver variant + * @clks: clock names + * @num_clks: numbers of clock + * @formats: jpeg driver's internal color format + * @num_formats: number of formats + * @qops: the callback of jpeg vb2_ops + * @irq_handler: jpeg irq handler callback + * @hw_reset: jpeg hardware reset callback + * @m2m_ops: the callback of jpeg v4l2_m2m_ops + * @dev_name: jpeg device name + * @ioctl_ops: the callback of jpeg v4l2_ioctl_ops + * @out_q_default_fourcc: output queue default fourcc + * @cap_q_default_fourcc: capture queue default fourcc + * @multi_core: mark jpeg hw is multi_core or not + * @jpeg_worker: jpeg dec or enc worker + */ +struct mtk_jpeg_variant { + struct clk_bulk_data *clks; + int num_clks; + struct mtk_jpeg_fmt *formats; + int num_formats; + const struct vb2_ops *qops; + irqreturn_t (*irq_handler)(int irq, void *priv); + void (*hw_reset)(void __iomem *base); + const struct v4l2_m2m_ops *m2m_ops; + const char *dev_name; + const struct v4l2_ioctl_ops *ioctl_ops; + u32 out_q_default_fourcc; + u32 cap_q_default_fourcc; + bool multi_core; + void (*jpeg_worker)(struct work_struct *work); +}; + +struct mtk_jpeg_src_buf { + u32 frame_num; + struct vb2_v4l2_buffer b; + struct list_head list; + u32 bs_size; + struct mtk_jpeg_dec_param dec_param; + + struct mtk_jpeg_ctx *curr_ctx; +}; + +enum mtk_jpeg_hw_state { + MTK_JPEG_HW_IDLE = 0, + MTK_JPEG_HW_BUSY = 1, +}; + +struct mtk_jpeg_hw_param { + struct vb2_v4l2_buffer *src_buffer; + struct vb2_v4l2_buffer *dst_buffer; + struct mtk_jpeg_ctx *curr_ctx; +}; + +enum mtk_jpegenc_hw_id { + MTK_JPEGENC_HW0, + MTK_JPEGENC_HW1, + MTK_JPEGENC_HW_MAX, +}; + +enum mtk_jpegdec_hw_id { + MTK_JPEGDEC_HW0, + MTK_JPEGDEC_HW1, + MTK_JPEGDEC_HW2, + MTK_JPEGDEC_HW_MAX, +}; + +/** + * struct mtk_jpegenc_clk - Structure used to store vcodec clock information + * @clks: JPEG encode clock + * @clk_num: JPEG encode clock numbers + */ +struct mtk_jpegenc_clk { + struct clk_bulk_data *clks; + int clk_num; +}; + +/** + * struct mtk_jpegdec_clk - Structure used to store vcodec clock information + * @clks: JPEG decode clock + * @clk_num: JPEG decode clock numbers + */ +struct mtk_jpegdec_clk { + struct clk_bulk_data *clks; + int clk_num; +}; + +/** + * struct mtk_jpegenc_comp_dev - JPEG COREX abstraction + * @dev: JPEG device + * @plat_dev: platform device data + * @reg_base: JPEG registers mapping + * @master_dev: mtk_jpeg_dev device + * @venc_clk: jpeg encode clock + * @jpegenc_irq: jpeg encode irq num + * @job_timeout_work: encode timeout workqueue + * @hw_param: jpeg encode hw parameters + * @hw_rdy: record hw ready + * @hw_state: record hw state + * @hw_lock: spinlock protecting the hw device resource + */ +struct mtk_jpegenc_comp_dev { + struct device *dev; + struct platform_device *plat_dev; + void __iomem *reg_base; + struct mtk_jpeg_dev *master_dev; + struct mtk_jpegenc_clk venc_clk; + int jpegenc_irq; + struct delayed_work job_timeout_work; + struct mtk_jpeg_hw_param hw_param; + enum mtk_jpeg_hw_state hw_state; + /* spinlock protecting the hw device resource */ + spinlock_t hw_lock; +}; + +/** + * struct mtk_jpegdec_comp_dev - JPEG COREX abstraction + * @dev: JPEG device + * @plat_dev: platform device data + * @reg_base: JPEG registers mapping + * @master_dev: mtk_jpeg_dev device + * @jdec_clk: mtk_jpegdec_clk + * @jpegdec_irq: jpeg decode irq num + * @job_timeout_work: decode timeout workqueue + * @hw_param: jpeg decode hw parameters + * @hw_state: record hw state + * @hw_lock: spinlock protecting hw + */ +struct mtk_jpegdec_comp_dev { + struct device *dev; + struct platform_device *plat_dev; + void __iomem *reg_base; + struct mtk_jpeg_dev *master_dev; + struct mtk_jpegdec_clk jdec_clk; + int jpegdec_irq; + struct delayed_work job_timeout_work; + struct mtk_jpeg_hw_param hw_param; + enum mtk_jpeg_hw_state hw_state; + /* spinlock protecting the hw device resource */ + spinlock_t hw_lock; +}; + +/** + * struct mtk_jpeg_dev - JPEG IP abstraction + * @lock: the mutex protecting this structure + * @hw_lock: spinlock protecting the hw device resource + * @workqueue: decode work queue + * @dev: JPEG device + * @v4l2_dev: v4l2 device for mem2mem mode + * @m2m_dev: v4l2 mem2mem device data + * @alloc_ctx: videobuf2 memory allocator's context + * @vdev: video device node for jpeg mem2mem mode + * @reg_base: JPEG registers mapping + * @job_timeout_work: IRQ timeout structure + * @variant: driver variant to be used + * @reg_encbase: jpg encode register base addr + * @enc_hw_dev: jpg encode hardware device + * @hw_wq: jpg wait queue + * @hw_rdy: jpg hw ready flag + * @reg_decbase: jpg decode register base addr + * @dec_hw_dev: jpg decode hardware device + * @hw_index: jpg hw index + */ +struct mtk_jpeg_dev { + struct mutex lock; + spinlock_t hw_lock; + struct workqueue_struct *workqueue; + struct device *dev; + struct v4l2_device v4l2_dev; + struct v4l2_m2m_dev *m2m_dev; + void *alloc_ctx; + struct video_device *vdev; + void __iomem *reg_base; + struct delayed_work job_timeout_work; + const struct mtk_jpeg_variant *variant; + + void __iomem *reg_encbase[MTK_JPEGENC_HW_MAX]; + struct mtk_jpegenc_comp_dev *enc_hw_dev[MTK_JPEGENC_HW_MAX]; + wait_queue_head_t hw_wq; + atomic_t hw_rdy; + + void __iomem *reg_decbase[MTK_JPEGDEC_HW_MAX]; + struct mtk_jpegdec_comp_dev *dec_hw_dev[MTK_JPEGDEC_HW_MAX]; + atomic_t hw_index; +}; + +/** + * struct mtk_jpeg_fmt - driver's internal color format data + * @fourcc: the fourcc code, 0 if not applicable + * @hw_format: hardware format value + * @h_sample: horizontal sample count of plane in 4 * 4 pixel image + * @v_sample: vertical sample count of plane in 4 * 4 pixel image + * @colplanes: number of color planes (1 for packed formats) + * @h_align: horizontal alignment order (align to 2^h_align) + * @v_align: vertical alignment order (align to 2^v_align) + * @flags: flags describing format applicability + */ +struct mtk_jpeg_fmt { + u32 fourcc; + u32 hw_format; + int h_sample[VIDEO_MAX_PLANES]; + int v_sample[VIDEO_MAX_PLANES]; + int colplanes; + int h_align; + int v_align; + u32 flags; +}; + +/** + * struct mtk_jpeg_q_data - parameters of one queue + * @fmt: driver-specific format of this queue + * @pix_mp: multiplanar format + * @enc_crop_rect: jpeg encoder crop information + */ +struct mtk_jpeg_q_data { + struct mtk_jpeg_fmt *fmt; + struct v4l2_pix_format_mplane pix_mp; + struct v4l2_rect enc_crop_rect; +}; + +/** + * struct mtk_jpeg_ctx - the device context data + * @jpeg: JPEG IP device for this context + * @out_q: source (output) queue information + * @cap_q: destination queue information + * @fh: V4L2 file handle + * @state: state of the context + * @enable_exif: enable exif mode of jpeg encoder + * @enc_quality: jpeg encoder quality + * @restart_interval: jpeg encoder restart interval + * @ctrl_hdl: controls handler + * @jpeg_work: jpeg encoder workqueue + * @total_frame_num: encoded frame number + * @dst_done_queue: encoded frame buffer queue + * @done_queue_lock: encoded frame operation spinlock + * @last_done_frame_num: the last encoded frame number + */ +struct mtk_jpeg_ctx { + struct mtk_jpeg_dev *jpeg; + struct mtk_jpeg_q_data out_q; + struct mtk_jpeg_q_data cap_q; + struct v4l2_fh fh; + enum mtk_jpeg_ctx_state state; + bool enable_exif; + u8 enc_quality; + u8 restart_interval; + struct v4l2_ctrl_handler ctrl_hdl; + + struct work_struct jpeg_work; + u32 total_frame_num; + struct list_head dst_done_queue; + /* spinlock protecting the encode done buffer */ + spinlock_t done_queue_lock; + u32 last_done_frame_num; +}; + +#endif /* _MTK_JPEG_CORE_H */ diff --git a/drivers/media/platform/mediatek/jpeg/mtk_jpeg_dec_hw.c b/drivers/media/platform/mediatek/jpeg/mtk_jpeg_dec_hw.c new file mode 100644 index 0000000000..4a6ee211e1 --- /dev/null +++ b/drivers/media/platform/mediatek/jpeg/mtk_jpeg_dec_hw.c @@ -0,0 +1,661 @@ +// SPDX-License-Identifier: GPL-2.0-only +/* + * Copyright (c) 2016 MediaTek Inc. + * Author: Ming Hsiu Tsai <minghsiu.tsai@mediatek.com> + * Rick Chang <rick.chang@mediatek.com> + */ + +#include <linux/clk.h> +#include <linux/interrupt.h> +#include <linux/irq.h> +#include <linux/io.h> +#include <linux/kernel.h> +#include <linux/module.h> +#include <linux/of.h> +#include <linux/platform_device.h> +#include <linux/pm_runtime.h> +#include <linux/slab.h> +#include <media/media-device.h> +#include <media/videobuf2-core.h> +#include <media/videobuf2-v4l2.h> +#include <media/v4l2-mem2mem.h> +#include <media/v4l2-dev.h> +#include <media/v4l2-device.h> +#include <media/v4l2-fh.h> +#include <media/v4l2-event.h> + +#include "mtk_jpeg_core.h" +#include "mtk_jpeg_dec_hw.h" + +#define MTK_JPEG_DUNUM_MASK(val) (((val) - 1) & 0x3) + +enum mtk_jpeg_color { + MTK_JPEG_COLOR_420 = 0x00221111, + MTK_JPEG_COLOR_422 = 0x00211111, + MTK_JPEG_COLOR_444 = 0x00111111, + MTK_JPEG_COLOR_422V = 0x00121111, + MTK_JPEG_COLOR_422X2 = 0x00412121, + MTK_JPEG_COLOR_422VX2 = 0x00222121, + MTK_JPEG_COLOR_400 = 0x00110000 +}; + +static const struct of_device_id mtk_jpegdec_hw_ids[] = { + { + .compatible = "mediatek,mt8195-jpgdec-hw", + }, + {}, +}; +MODULE_DEVICE_TABLE(of, mtk_jpegdec_hw_ids); + +static inline int mtk_jpeg_verify_align(u32 val, int align, u32 reg) +{ + if (val & (align - 1)) { + pr_err("mtk-jpeg: write reg %x without %d align\n", reg, align); + return -1; + } + + return 0; +} + +static int mtk_jpeg_decide_format(struct mtk_jpeg_dec_param *param) +{ + param->src_color = (param->sampling_w[0] << 20) | + (param->sampling_h[0] << 16) | + (param->sampling_w[1] << 12) | + (param->sampling_h[1] << 8) | + (param->sampling_w[2] << 4) | + (param->sampling_h[2]); + + param->uv_brz_w = 0; + switch (param->src_color) { + case MTK_JPEG_COLOR_444: + param->uv_brz_w = 1; + param->dst_fourcc = V4L2_PIX_FMT_YUV422M; + break; + case MTK_JPEG_COLOR_422X2: + case MTK_JPEG_COLOR_422: + param->dst_fourcc = V4L2_PIX_FMT_YUV422M; + break; + case MTK_JPEG_COLOR_422V: + case MTK_JPEG_COLOR_422VX2: + param->uv_brz_w = 1; + param->dst_fourcc = V4L2_PIX_FMT_YUV420M; + break; + case MTK_JPEG_COLOR_420: + param->dst_fourcc = V4L2_PIX_FMT_YUV420M; + break; + case MTK_JPEG_COLOR_400: + param->dst_fourcc = V4L2_PIX_FMT_GREY; + break; + default: + param->dst_fourcc = 0; + return -1; + } + + return 0; +} + +static void mtk_jpeg_calc_mcu(struct mtk_jpeg_dec_param *param) +{ + u32 factor_w, factor_h; + u32 i, comp, blk; + + factor_w = 2 + param->sampling_w[0]; + factor_h = 2 + param->sampling_h[0]; + param->mcu_w = (param->pic_w + (1 << factor_w) - 1) >> factor_w; + param->mcu_h = (param->pic_h + (1 << factor_h) - 1) >> factor_h; + param->total_mcu = param->mcu_w * param->mcu_h; + param->unit_num = ((param->pic_w + 7) >> 3) * ((param->pic_h + 7) >> 3); + param->blk_num = 0; + for (i = 0; i < MTK_JPEG_COMP_MAX; i++) { + param->blk_comp[i] = 0; + if (i >= param->comp_num) + continue; + param->blk_comp[i] = param->sampling_w[i] * + param->sampling_h[i]; + param->blk_num += param->blk_comp[i]; + } + + param->membership = 0; + for (i = 0, blk = 0, comp = 0; i < MTK_JPEG_BLOCK_MAX; i++) { + if (i < param->blk_num && comp < param->comp_num) { + u32 tmp; + + tmp = (0x04 + (comp & 0x3)); + param->membership |= tmp << (i * 3); + if (++blk == param->blk_comp[comp]) { + comp++; + blk = 0; + } + } else { + param->membership |= 7 << (i * 3); + } + } +} + +static void mtk_jpeg_calc_dma_group(struct mtk_jpeg_dec_param *param) +{ + u32 factor_mcu = 3; + + if (param->src_color == MTK_JPEG_COLOR_444 && + param->dst_fourcc == V4L2_PIX_FMT_YUV422M) + factor_mcu = 4; + else if (param->src_color == MTK_JPEG_COLOR_422V && + param->dst_fourcc == V4L2_PIX_FMT_YUV420M) + factor_mcu = 4; + else if (param->src_color == MTK_JPEG_COLOR_422X2 && + param->dst_fourcc == V4L2_PIX_FMT_YUV422M) + factor_mcu = 2; + else if (param->src_color == MTK_JPEG_COLOR_400 || + (param->src_color & 0x0FFFF) == 0) + factor_mcu = 4; + + param->dma_mcu = 1 << factor_mcu; + param->dma_group = param->mcu_w / param->dma_mcu; + param->dma_last_mcu = param->mcu_w % param->dma_mcu; + if (param->dma_last_mcu) + param->dma_group++; + else + param->dma_last_mcu = param->dma_mcu; +} + +static int mtk_jpeg_calc_dst_size(struct mtk_jpeg_dec_param *param) +{ + u32 i, padding_w; + u32 ds_row_h[3]; + u32 brz_w[3]; + + brz_w[0] = 0; + brz_w[1] = param->uv_brz_w; + brz_w[2] = brz_w[1]; + + for (i = 0; i < param->comp_num; i++) { + if (brz_w[i] > 3) + return -1; + + padding_w = param->mcu_w * MTK_JPEG_DCTSIZE * + param->sampling_w[i]; + /* output format is 420/422 */ + param->comp_w[i] = padding_w >> brz_w[i]; + param->comp_w[i] = round_up(param->comp_w[i], + MTK_JPEG_DCTSIZE); + param->img_stride[i] = i ? round_up(param->comp_w[i], 16) + : round_up(param->comp_w[i], 32); + ds_row_h[i] = (MTK_JPEG_DCTSIZE * param->sampling_h[i]); + } + param->dec_w = param->img_stride[0]; + param->dec_h = ds_row_h[0] * param->mcu_h; + + for (i = 0; i < MTK_JPEG_COMP_MAX; i++) { + /* They must be equal in frame mode. */ + param->mem_stride[i] = param->img_stride[i]; + param->comp_size[i] = param->mem_stride[i] * ds_row_h[i] * + param->mcu_h; + } + + param->y_size = param->comp_size[0]; + param->uv_size = param->comp_size[1]; + param->dec_size = param->y_size + (param->uv_size << 1); + + return 0; +} + +int mtk_jpeg_dec_fill_param(struct mtk_jpeg_dec_param *param) +{ + if (mtk_jpeg_decide_format(param)) + return -1; + + mtk_jpeg_calc_mcu(param); + mtk_jpeg_calc_dma_group(param); + if (mtk_jpeg_calc_dst_size(param)) + return -2; + + return 0; +} +EXPORT_SYMBOL_GPL(mtk_jpeg_dec_fill_param); + +u32 mtk_jpeg_dec_get_int_status(void __iomem *base) +{ + u32 ret; + + ret = readl(base + JPGDEC_REG_INTERRUPT_STATUS) & BIT_INQST_MASK_ALLIRQ; + if (ret) + writel(ret, base + JPGDEC_REG_INTERRUPT_STATUS); + + return ret; +} +EXPORT_SYMBOL_GPL(mtk_jpeg_dec_get_int_status); + +u32 mtk_jpeg_dec_enum_result(u32 irq_result) +{ + if (irq_result & BIT_INQST_MASK_EOF) + return MTK_JPEG_DEC_RESULT_EOF_DONE; + if (irq_result & BIT_INQST_MASK_PAUSE) + return MTK_JPEG_DEC_RESULT_PAUSE; + if (irq_result & BIT_INQST_MASK_UNDERFLOW) + return MTK_JPEG_DEC_RESULT_UNDERFLOW; + if (irq_result & BIT_INQST_MASK_OVERFLOW) + return MTK_JPEG_DEC_RESULT_OVERFLOW; + if (irq_result & BIT_INQST_MASK_ERROR_BS) + return MTK_JPEG_DEC_RESULT_ERROR_BS; + + return MTK_JPEG_DEC_RESULT_ERROR_UNKNOWN; +} +EXPORT_SYMBOL_GPL(mtk_jpeg_dec_enum_result); + +void mtk_jpeg_dec_start(void __iomem *base) +{ + writel(0, base + JPGDEC_REG_TRIG); +} +EXPORT_SYMBOL_GPL(mtk_jpeg_dec_start); + +static void mtk_jpeg_dec_soft_reset(void __iomem *base) +{ + writel(0x0000FFFF, base + JPGDEC_REG_INTERRUPT_STATUS); + writel(0x00, base + JPGDEC_REG_RESET); + writel(0x01, base + JPGDEC_REG_RESET); +} + +static void mtk_jpeg_dec_hard_reset(void __iomem *base) +{ + writel(0x00, base + JPGDEC_REG_RESET); + writel(0x10, base + JPGDEC_REG_RESET); +} + +void mtk_jpeg_dec_reset(void __iomem *base) +{ + mtk_jpeg_dec_soft_reset(base); + mtk_jpeg_dec_hard_reset(base); +} +EXPORT_SYMBOL_GPL(mtk_jpeg_dec_reset); + +static void mtk_jpeg_dec_set_brz_factor(void __iomem *base, u8 yscale_w, + u8 yscale_h, u8 uvscale_w, u8 uvscale_h) +{ + u32 val; + + val = (uvscale_h << 12) | (uvscale_w << 8) | + (yscale_h << 4) | yscale_w; + writel(val, base + JPGDEC_REG_BRZ_FACTOR); +} + +static void mtk_jpeg_dec_set_dst_bank0(void __iomem *base, u32 addr_y, + u32 addr_u, u32 addr_v) +{ + mtk_jpeg_verify_align(addr_y, 16, JPGDEC_REG_DEST_ADDR0_Y); + writel(addr_y, base + JPGDEC_REG_DEST_ADDR0_Y); + mtk_jpeg_verify_align(addr_u, 16, JPGDEC_REG_DEST_ADDR0_U); + writel(addr_u, base + JPGDEC_REG_DEST_ADDR0_U); + mtk_jpeg_verify_align(addr_v, 16, JPGDEC_REG_DEST_ADDR0_V); + writel(addr_v, base + JPGDEC_REG_DEST_ADDR0_V); +} + +static void mtk_jpeg_dec_set_dst_bank1(void __iomem *base, u32 addr_y, + u32 addr_u, u32 addr_v) +{ + writel(addr_y, base + JPGDEC_REG_DEST_ADDR1_Y); + writel(addr_u, base + JPGDEC_REG_DEST_ADDR1_U); + writel(addr_v, base + JPGDEC_REG_DEST_ADDR1_V); +} + +static void mtk_jpeg_dec_set_mem_stride(void __iomem *base, u32 stride_y, + u32 stride_uv) +{ + writel((stride_y & 0xFFFF), base + JPGDEC_REG_STRIDE_Y); + writel((stride_uv & 0xFFFF), base + JPGDEC_REG_STRIDE_UV); +} + +static void mtk_jpeg_dec_set_img_stride(void __iomem *base, u32 stride_y, + u32 stride_uv) +{ + writel((stride_y & 0xFFFF), base + JPGDEC_REG_IMG_STRIDE_Y); + writel((stride_uv & 0xFFFF), base + JPGDEC_REG_IMG_STRIDE_UV); +} + +static void mtk_jpeg_dec_set_pause_mcu_idx(void __iomem *base, u32 idx) +{ + writel(idx & 0x0003FFFFFF, base + JPGDEC_REG_PAUSE_MCU_NUM); +} + +static void mtk_jpeg_dec_set_dec_mode(void __iomem *base, u32 mode) +{ + writel(mode & 0x03, base + JPGDEC_REG_OPERATION_MODE); +} + +static void mtk_jpeg_dec_set_bs_write_ptr(void __iomem *base, u32 ptr) +{ + mtk_jpeg_verify_align(ptr, 16, JPGDEC_REG_FILE_BRP); + writel(ptr, base + JPGDEC_REG_FILE_BRP); +} + +static void mtk_jpeg_dec_set_bs_info(void __iomem *base, u32 addr, u32 size, + u32 bitstream_size) +{ + mtk_jpeg_verify_align(addr, 16, JPGDEC_REG_FILE_ADDR); + mtk_jpeg_verify_align(size, 128, JPGDEC_REG_FILE_TOTAL_SIZE); + writel(addr, base + JPGDEC_REG_FILE_ADDR); + writel(size, base + JPGDEC_REG_FILE_TOTAL_SIZE); + writel(bitstream_size, base + JPGDEC_REG_BIT_STREAM_SIZE); +} + +static void mtk_jpeg_dec_set_comp_id(void __iomem *base, u32 id_y, u32 id_u, + u32 id_v) +{ + u32 val; + + val = ((id_y & 0x00FF) << 24) | ((id_u & 0x00FF) << 16) | + ((id_v & 0x00FF) << 8); + writel(val, base + JPGDEC_REG_COMP_ID); +} + +static void mtk_jpeg_dec_set_total_mcu(void __iomem *base, u32 num) +{ + writel(num - 1, base + JPGDEC_REG_TOTAL_MCU_NUM); +} + +static void mtk_jpeg_dec_set_comp0_du(void __iomem *base, u32 num) +{ + writel(num - 1, base + JPGDEC_REG_COMP0_DATA_UNIT_NUM); +} + +static void mtk_jpeg_dec_set_du_membership(void __iomem *base, u32 member, + u32 gmc, u32 isgray) +{ + if (isgray) + member = 0x3FFFFFFC; + member |= (isgray << 31) | (gmc << 30); + writel(member, base + JPGDEC_REG_DU_CTRL); +} + +static void mtk_jpeg_dec_set_q_table(void __iomem *base, u32 id0, u32 id1, + u32 id2) +{ + u32 val; + + val = ((id0 & 0x0f) << 8) | ((id1 & 0x0f) << 4) | ((id2 & 0x0f) << 0); + writel(val, base + JPGDEC_REG_QT_ID); +} + +static void mtk_jpeg_dec_set_dma_group(void __iomem *base, u32 mcu_group, + u32 group_num, u32 last_mcu) +{ + u32 val; + + val = (((mcu_group - 1) & 0x00FF) << 16) | + (((group_num - 1) & 0x007F) << 8) | + ((last_mcu - 1) & 0x00FF); + writel(val, base + JPGDEC_REG_WDMA_CTRL); +} + +static void mtk_jpeg_dec_set_sampling_factor(void __iomem *base, u32 comp_num, + u32 y_w, u32 y_h, u32 u_w, + u32 u_h, u32 v_w, u32 v_h) +{ + u32 val; + u32 y_wh = (MTK_JPEG_DUNUM_MASK(y_w) << 2) | MTK_JPEG_DUNUM_MASK(y_h); + u32 u_wh = (MTK_JPEG_DUNUM_MASK(u_w) << 2) | MTK_JPEG_DUNUM_MASK(u_h); + u32 v_wh = (MTK_JPEG_DUNUM_MASK(v_w) << 2) | MTK_JPEG_DUNUM_MASK(v_h); + + if (comp_num == 1) + val = 0; + else + val = (y_wh << 8) | (u_wh << 4) | v_wh; + writel(val, base + JPGDEC_REG_DU_NUM); +} + +void mtk_jpeg_dec_set_config(void __iomem *base, + struct mtk_jpeg_dec_param *cfg, + u32 bitstream_size, + struct mtk_jpeg_bs *bs, + struct mtk_jpeg_fb *fb) +{ + mtk_jpeg_dec_set_brz_factor(base, 0, 0, cfg->uv_brz_w, 0); + mtk_jpeg_dec_set_dec_mode(base, 0); + mtk_jpeg_dec_set_comp0_du(base, cfg->unit_num); + mtk_jpeg_dec_set_total_mcu(base, cfg->total_mcu); + mtk_jpeg_dec_set_bs_info(base, bs->str_addr, bs->size, bitstream_size); + mtk_jpeg_dec_set_bs_write_ptr(base, bs->end_addr); + mtk_jpeg_dec_set_du_membership(base, cfg->membership, 1, + (cfg->comp_num == 1) ? 1 : 0); + mtk_jpeg_dec_set_comp_id(base, cfg->comp_id[0], cfg->comp_id[1], + cfg->comp_id[2]); + mtk_jpeg_dec_set_q_table(base, cfg->qtbl_num[0], + cfg->qtbl_num[1], cfg->qtbl_num[2]); + mtk_jpeg_dec_set_sampling_factor(base, cfg->comp_num, + cfg->sampling_w[0], + cfg->sampling_h[0], + cfg->sampling_w[1], + cfg->sampling_h[1], + cfg->sampling_w[2], + cfg->sampling_h[2]); + mtk_jpeg_dec_set_mem_stride(base, cfg->mem_stride[0], + cfg->mem_stride[1]); + mtk_jpeg_dec_set_img_stride(base, cfg->img_stride[0], + cfg->img_stride[1]); + mtk_jpeg_dec_set_dst_bank0(base, fb->plane_addr[0], + fb->plane_addr[1], fb->plane_addr[2]); + mtk_jpeg_dec_set_dst_bank1(base, 0, 0, 0); + mtk_jpeg_dec_set_dma_group(base, cfg->dma_mcu, cfg->dma_group, + cfg->dma_last_mcu); + mtk_jpeg_dec_set_pause_mcu_idx(base, cfg->total_mcu); +} +EXPORT_SYMBOL_GPL(mtk_jpeg_dec_set_config); + +static void mtk_jpegdec_put_buf(struct mtk_jpegdec_comp_dev *jpeg) +{ + struct mtk_jpeg_src_buf *dst_done_buf, *tmp_dst_done_buf; + struct vb2_v4l2_buffer *dst_buffer; + struct list_head *temp_entry; + struct list_head *pos = NULL; + struct mtk_jpeg_ctx *ctx; + unsigned long flags; + + ctx = jpeg->hw_param.curr_ctx; + if (unlikely(!ctx)) { + dev_err(jpeg->dev, "comp_jpeg ctx fail !!!\n"); + return; + } + + dst_buffer = jpeg->hw_param.dst_buffer; + if (!dst_buffer) { + dev_err(jpeg->dev, "comp_jpeg dst_buffer fail !!!\n"); + return; + } + + dst_done_buf = container_of(dst_buffer, struct mtk_jpeg_src_buf, b); + + spin_lock_irqsave(&ctx->done_queue_lock, flags); + list_add_tail(&dst_done_buf->list, &ctx->dst_done_queue); + while (!list_empty(&ctx->dst_done_queue) && + (pos != &ctx->dst_done_queue)) { + list_for_each_prev_safe(pos, temp_entry, &ctx->dst_done_queue) { + tmp_dst_done_buf = list_entry(pos, + struct mtk_jpeg_src_buf, + list); + if (tmp_dst_done_buf->frame_num == + ctx->last_done_frame_num) { + list_del(&tmp_dst_done_buf->list); + v4l2_m2m_buf_done(&tmp_dst_done_buf->b, + VB2_BUF_STATE_DONE); + ctx->last_done_frame_num++; + } + } + } + spin_unlock_irqrestore(&ctx->done_queue_lock, flags); +} + +static void mtk_jpegdec_timeout_work(struct work_struct *work) +{ + enum vb2_buffer_state buf_state = VB2_BUF_STATE_ERROR; + struct mtk_jpegdec_comp_dev *cjpeg = + container_of(work, struct mtk_jpegdec_comp_dev, + job_timeout_work.work); + struct mtk_jpeg_dev *master_jpeg = cjpeg->master_dev; + struct vb2_v4l2_buffer *src_buf, *dst_buf; + + src_buf = cjpeg->hw_param.src_buffer; + dst_buf = cjpeg->hw_param.dst_buffer; + v4l2_m2m_buf_copy_metadata(src_buf, dst_buf, true); + + mtk_jpeg_dec_reset(cjpeg->reg_base); + clk_disable_unprepare(cjpeg->jdec_clk.clks->clk); + pm_runtime_put(cjpeg->dev); + cjpeg->hw_state = MTK_JPEG_HW_IDLE; + atomic_inc(&master_jpeg->hw_rdy); + wake_up(&master_jpeg->hw_wq); + v4l2_m2m_buf_done(src_buf, buf_state); + mtk_jpegdec_put_buf(cjpeg); +} + +static irqreturn_t mtk_jpegdec_hw_irq_handler(int irq, void *priv) +{ + struct vb2_v4l2_buffer *src_buf, *dst_buf; + struct mtk_jpeg_src_buf *jpeg_src_buf; + enum vb2_buffer_state buf_state; + struct mtk_jpeg_ctx *ctx; + u32 dec_irq_ret; + u32 irq_status; + int i; + + struct mtk_jpegdec_comp_dev *jpeg = priv; + struct mtk_jpeg_dev *master_jpeg = jpeg->master_dev; + + cancel_delayed_work(&jpeg->job_timeout_work); + + ctx = jpeg->hw_param.curr_ctx; + src_buf = jpeg->hw_param.src_buffer; + dst_buf = jpeg->hw_param.dst_buffer; + v4l2_m2m_buf_copy_metadata(src_buf, dst_buf, true); + + irq_status = mtk_jpeg_dec_get_int_status(jpeg->reg_base); + dec_irq_ret = mtk_jpeg_dec_enum_result(irq_status); + if (dec_irq_ret >= MTK_JPEG_DEC_RESULT_UNDERFLOW) + mtk_jpeg_dec_reset(jpeg->reg_base); + + if (dec_irq_ret != MTK_JPEG_DEC_RESULT_EOF_DONE) + dev_warn(jpeg->dev, "Jpg Dec occurs unknown Err."); + + jpeg_src_buf = + container_of(src_buf, struct mtk_jpeg_src_buf, b); + + for (i = 0; i < dst_buf->vb2_buf.num_planes; i++) + vb2_set_plane_payload(&dst_buf->vb2_buf, i, + jpeg_src_buf->dec_param.comp_size[i]); + + buf_state = VB2_BUF_STATE_DONE; + v4l2_m2m_buf_done(src_buf, buf_state); + mtk_jpegdec_put_buf(jpeg); + pm_runtime_put(ctx->jpeg->dev); + clk_disable_unprepare(jpeg->jdec_clk.clks->clk); + + jpeg->hw_state = MTK_JPEG_HW_IDLE; + wake_up(&master_jpeg->hw_wq); + atomic_inc(&master_jpeg->hw_rdy); + + return IRQ_HANDLED; +} + +static int mtk_jpegdec_hw_init_irq(struct mtk_jpegdec_comp_dev *dev) +{ + struct platform_device *pdev = dev->plat_dev; + int ret; + + dev->jpegdec_irq = platform_get_irq(pdev, 0); + if (dev->jpegdec_irq < 0) + return dev->jpegdec_irq; + + ret = devm_request_irq(&pdev->dev, + dev->jpegdec_irq, + mtk_jpegdec_hw_irq_handler, + 0, + pdev->name, dev); + if (ret) { + dev_err(&pdev->dev, "Failed to devm_request_irq %d (%d)", + dev->jpegdec_irq, ret); + return ret; + } + + return 0; +} + +static void mtk_jpegdec_destroy_workqueue(void *data) +{ + destroy_workqueue(data); +} + +static int mtk_jpegdec_hw_probe(struct platform_device *pdev) +{ + struct mtk_jpegdec_clk *jpegdec_clk; + struct mtk_jpeg_dev *master_dev; + struct mtk_jpegdec_comp_dev *dev; + int ret, i; + + struct device *decs = &pdev->dev; + + if (!decs->parent) + return -EPROBE_DEFER; + + master_dev = dev_get_drvdata(decs->parent); + if (!master_dev) + return -EPROBE_DEFER; + + dev = devm_kzalloc(&pdev->dev, sizeof(*dev), GFP_KERNEL); + if (!dev) + return -ENOMEM; + + dev->plat_dev = pdev; + dev->dev = &pdev->dev; + + ret = devm_add_action_or_reset(&pdev->dev, + mtk_jpegdec_destroy_workqueue, + master_dev->workqueue); + if (ret) + return ret; + + spin_lock_init(&dev->hw_lock); + dev->hw_state = MTK_JPEG_HW_IDLE; + + INIT_DELAYED_WORK(&dev->job_timeout_work, + mtk_jpegdec_timeout_work); + + jpegdec_clk = &dev->jdec_clk; + + jpegdec_clk->clk_num = devm_clk_bulk_get_all(&pdev->dev, + &jpegdec_clk->clks); + if (jpegdec_clk->clk_num < 0) + return dev_err_probe(&pdev->dev, + jpegdec_clk->clk_num, + "Failed to get jpegdec clock count.\n"); + + dev->reg_base = devm_platform_ioremap_resource(pdev, 0); + if (IS_ERR(dev->reg_base)) + return PTR_ERR(dev->reg_base); + + ret = mtk_jpegdec_hw_init_irq(dev); + if (ret) + return dev_err_probe(&pdev->dev, ret, + "Failed to register JPEGDEC irq handler.\n"); + + i = atomic_add_return(1, &master_dev->hw_index) - 1; + master_dev->dec_hw_dev[i] = dev; + master_dev->reg_decbase[i] = dev->reg_base; + dev->master_dev = master_dev; + + platform_set_drvdata(pdev, dev); + pm_runtime_enable(&pdev->dev); + + return 0; +} + +static struct platform_driver mtk_jpegdec_hw_driver = { + .probe = mtk_jpegdec_hw_probe, + .driver = { + .name = "mtk-jpegdec-hw", + .of_match_table = mtk_jpegdec_hw_ids, + }, +}; + +module_platform_driver(mtk_jpegdec_hw_driver); + +MODULE_DESCRIPTION("MediaTek JPEG decode HW driver"); +MODULE_LICENSE("GPL"); diff --git a/drivers/media/platform/mediatek/jpeg/mtk_jpeg_dec_hw.h b/drivers/media/platform/mediatek/jpeg/mtk_jpeg_dec_hw.h new file mode 100644 index 0000000000..8c31c6b124 --- /dev/null +++ b/drivers/media/platform/mediatek/jpeg/mtk_jpeg_dec_hw.h @@ -0,0 +1,81 @@ +/* SPDX-License-Identifier: GPL-2.0-only */ +/* + * Copyright (c) 2016 MediaTek Inc. + * Author: Ming Hsiu Tsai <minghsiu.tsai@mediatek.com> + * Rick Chang <rick.chang@mediatek.com> + * Xia Jiang <xia.jiang@mediatek.com> + */ + +#ifndef _MTK_JPEG_DEC_HW_H +#define _MTK_JPEG_DEC_HW_H + +#include <media/videobuf2-core.h> + +#include "mtk_jpeg_dec_reg.h" + +#define MTK_JPEG_COMP_MAX 3 + +enum { + MTK_JPEG_DEC_RESULT_EOF_DONE = 0, + MTK_JPEG_DEC_RESULT_PAUSE = 1, + MTK_JPEG_DEC_RESULT_UNDERFLOW = 2, + MTK_JPEG_DEC_RESULT_OVERFLOW = 3, + MTK_JPEG_DEC_RESULT_ERROR_BS = 4, + MTK_JPEG_DEC_RESULT_ERROR_UNKNOWN = 6 +}; + +struct mtk_jpeg_dec_param { + u32 pic_w; + u32 pic_h; + u32 dec_w; + u32 dec_h; + u32 src_color; + u32 dst_fourcc; + u32 mcu_w; + u32 mcu_h; + u32 total_mcu; + u32 unit_num; + u32 comp_num; + u32 comp_id[MTK_JPEG_COMP_MAX]; + u32 sampling_w[MTK_JPEG_COMP_MAX]; + u32 sampling_h[MTK_JPEG_COMP_MAX]; + u32 qtbl_num[MTK_JPEG_COMP_MAX]; + u32 blk_num; + u32 blk_comp[MTK_JPEG_COMP_MAX]; + u32 membership; + u32 dma_mcu; + u32 dma_group; + u32 dma_last_mcu; + u32 img_stride[MTK_JPEG_COMP_MAX]; + u32 mem_stride[MTK_JPEG_COMP_MAX]; + u32 comp_w[MTK_JPEG_COMP_MAX]; + u32 comp_size[MTK_JPEG_COMP_MAX]; + u32 y_size; + u32 uv_size; + u32 dec_size; + u8 uv_brz_w; +}; + +struct mtk_jpeg_bs { + dma_addr_t str_addr; + dma_addr_t end_addr; + size_t size; +}; + +struct mtk_jpeg_fb { + dma_addr_t plane_addr[MTK_JPEG_COMP_MAX]; + size_t size; +}; + +int mtk_jpeg_dec_fill_param(struct mtk_jpeg_dec_param *param); +u32 mtk_jpeg_dec_get_int_status(void __iomem *dec_reg_base); +u32 mtk_jpeg_dec_enum_result(u32 irq_result); +void mtk_jpeg_dec_set_config(void __iomem *base, + struct mtk_jpeg_dec_param *cfg, + u32 bitstream_size, + struct mtk_jpeg_bs *bs, + struct mtk_jpeg_fb *fb); +void mtk_jpeg_dec_reset(void __iomem *dec_reg_base); +void mtk_jpeg_dec_start(void __iomem *dec_reg_base); + +#endif /* _MTK_JPEG_HW_H */ diff --git a/drivers/media/platform/mediatek/jpeg/mtk_jpeg_dec_parse.c b/drivers/media/platform/mediatek/jpeg/mtk_jpeg_dec_parse.c new file mode 100644 index 0000000000..bb9cdc9e0e --- /dev/null +++ b/drivers/media/platform/mediatek/jpeg/mtk_jpeg_dec_parse.c @@ -0,0 +1,147 @@ +// SPDX-License-Identifier: GPL-2.0-only +/* + * Copyright (c) 2016 MediaTek Inc. + * Author: Ming Hsiu Tsai <minghsiu.tsai@mediatek.com> + * Rick Chang <rick.chang@mediatek.com> + */ + +#include <linux/kernel.h> +#include <linux/videodev2.h> +#include <media/jpeg.h> + +#include "mtk_jpeg_dec_parse.h" + +struct mtk_jpeg_stream { + u8 *addr; + u32 size; + u32 curr; +}; + +static int read_byte(struct mtk_jpeg_stream *stream) +{ + if (stream->curr >= stream->size) + return -1; + return stream->addr[stream->curr++]; +} + +static int read_word_be(struct mtk_jpeg_stream *stream, u32 *word) +{ + u32 temp; + int byte; + + byte = read_byte(stream); + if (byte == -1) + return -1; + temp = byte << 8; + byte = read_byte(stream); + if (byte == -1) + return -1; + *word = (u32)byte | temp; + + return 0; +} + +static void read_skip(struct mtk_jpeg_stream *stream, long len) +{ + if (len <= 0) + return; + while (len--) + read_byte(stream); +} + +static bool mtk_jpeg_do_parse(struct mtk_jpeg_dec_param *param, u8 *src_addr_va, + u32 src_size) +{ + bool notfound = true; + struct mtk_jpeg_stream stream; + + stream.addr = src_addr_va; + stream.size = src_size; + stream.curr = 0; + + while (notfound) { + int i, length, byte; + u32 word; + + byte = read_byte(&stream); + if (byte == -1) + return false; + if (byte != 0xff) + continue; + do + byte = read_byte(&stream); + while (byte == 0xff); + if (byte == -1) + return false; + if (byte == 0) + continue; + + length = 0; + switch (byte) { + case JPEG_MARKER_SOF0: + /* length */ + if (read_word_be(&stream, &word)) + break; + + /* precision */ + if (read_byte(&stream) == -1) + break; + + if (read_word_be(&stream, &word)) + break; + param->pic_h = word; + + if (read_word_be(&stream, &word)) + break; + param->pic_w = word; + + param->comp_num = read_byte(&stream); + if (param->comp_num != 1 && param->comp_num != 3) + break; + + for (i = 0; i < param->comp_num; i++) { + param->comp_id[i] = read_byte(&stream); + if (param->comp_id[i] == -1) + break; + + /* sampling */ + byte = read_byte(&stream); + if (byte == -1) + break; + param->sampling_w[i] = (byte >> 4) & 0x0F; + param->sampling_h[i] = byte & 0x0F; + + param->qtbl_num[i] = read_byte(&stream); + if (param->qtbl_num[i] == -1) + break; + } + + notfound = !(i == param->comp_num); + break; + case JPEG_MARKER_RST ... JPEG_MARKER_RST + 7: + case JPEG_MARKER_SOI: + case JPEG_MARKER_EOI: + case JPEG_MARKER_TEM: + break; + default: + if (read_word_be(&stream, &word)) + break; + length = (long)word - 2; + read_skip(&stream, length); + break; + } + } + + return !notfound; +} + +bool mtk_jpeg_parse(struct mtk_jpeg_dec_param *param, u8 *src_addr_va, + u32 src_size) +{ + if (!mtk_jpeg_do_parse(param, src_addr_va, src_size)) + return false; + if (mtk_jpeg_dec_fill_param(param)) + return false; + + return true; +} diff --git a/drivers/media/platform/mediatek/jpeg/mtk_jpeg_dec_parse.h b/drivers/media/platform/mediatek/jpeg/mtk_jpeg_dec_parse.h new file mode 100644 index 0000000000..2918f15811 --- /dev/null +++ b/drivers/media/platform/mediatek/jpeg/mtk_jpeg_dec_parse.h @@ -0,0 +1,17 @@ +/* SPDX-License-Identifier: GPL-2.0-only */ +/* + * Copyright (c) 2016 MediaTek Inc. + * Author: Ming Hsiu Tsai <minghsiu.tsai@mediatek.com> + * Rick Chang <rick.chang@mediatek.com> + */ + +#ifndef _MTK_JPEG_PARSE_H +#define _MTK_JPEG_PARSE_H + +#include "mtk_jpeg_dec_hw.h" + +bool mtk_jpeg_parse(struct mtk_jpeg_dec_param *param, u8 *src_addr_va, + u32 src_size); + +#endif /* _MTK_JPEG_PARSE_H */ + diff --git a/drivers/media/platform/mediatek/jpeg/mtk_jpeg_dec_reg.h b/drivers/media/platform/mediatek/jpeg/mtk_jpeg_dec_reg.h new file mode 100644 index 0000000000..27b7711ca3 --- /dev/null +++ b/drivers/media/platform/mediatek/jpeg/mtk_jpeg_dec_reg.h @@ -0,0 +1,50 @@ +/* SPDX-License-Identifier: GPL-2.0-only */ +/* + * Copyright (c) 2016 MediaTek Inc. + * Author: Ming Hsiu Tsai <minghsiu.tsai@mediatek.com> + * Rick Chang <rick.chang@mediatek.com> + */ + +#ifndef _MTK_JPEG_REG_H +#define _MTK_JPEG_REG_H + +#define MTK_JPEG_BLOCK_MAX 10 +#define MTK_JPEG_DCTSIZE 8 + +#define BIT_INQST_MASK_ERROR_BS 0x20 +#define BIT_INQST_MASK_PAUSE 0x10 +#define BIT_INQST_MASK_OVERFLOW 0x04 +#define BIT_INQST_MASK_UNDERFLOW 0x02 +#define BIT_INQST_MASK_EOF 0x01 +#define BIT_INQST_MASK_ALLIRQ 0x37 + +#define JPGDEC_REG_RESET 0x0090 +#define JPGDEC_REG_BRZ_FACTOR 0x00f8 +#define JPGDEC_REG_DU_NUM 0x00fc +#define JPGDEC_REG_DEST_ADDR0_Y 0x0140 +#define JPGDEC_REG_DEST_ADDR0_U 0x0144 +#define JPGDEC_REG_DEST_ADDR0_V 0x0148 +#define JPGDEC_REG_DEST_ADDR1_Y 0x014c +#define JPGDEC_REG_DEST_ADDR1_U 0x0150 +#define JPGDEC_REG_DEST_ADDR1_V 0x0154 +#define JPGDEC_REG_STRIDE_Y 0x0158 +#define JPGDEC_REG_STRIDE_UV 0x015c +#define JPGDEC_REG_IMG_STRIDE_Y 0x0160 +#define JPGDEC_REG_IMG_STRIDE_UV 0x0164 +#define JPGDEC_REG_WDMA_CTRL 0x016c +#define JPGDEC_REG_PAUSE_MCU_NUM 0x0170 +#define JPGDEC_REG_OPERATION_MODE 0x017c +#define JPGDEC_REG_FILE_ADDR 0x0200 +#define JPGDEC_REG_COMP_ID 0x020c +#define JPGDEC_REG_TOTAL_MCU_NUM 0x0210 +#define JPGDEC_REG_COMP0_DATA_UNIT_NUM 0x0224 +#define JPGDEC_REG_DU_CTRL 0x023c +#define JPGDEC_REG_TRIG 0x0240 +#define JPGDEC_REG_FILE_BRP 0x0248 +#define JPGDEC_REG_FILE_TOTAL_SIZE 0x024c +#define JPGDEC_REG_QT_ID 0x0270 +#define JPGDEC_REG_INTERRUPT_STATUS 0x0274 +#define JPGDEC_REG_STATUS 0x0278 +#define JPGDEC_REG_BIT_STREAM_SIZE 0x0344 + +#endif /* _MTK_JPEG_REG_H */ diff --git a/drivers/media/platform/mediatek/jpeg/mtk_jpeg_enc_hw.c b/drivers/media/platform/mediatek/jpeg/mtk_jpeg_enc_hw.c new file mode 100644 index 0000000000..f8fa3b841c --- /dev/null +++ b/drivers/media/platform/mediatek/jpeg/mtk_jpeg_enc_hw.c @@ -0,0 +1,386 @@ +// SPDX-License-Identifier: GPL-2.0-only +/* + * Copyright (c) 2019 MediaTek Inc. + * Author: Xia Jiang <xia.jiang@mediatek.com> + * + */ + +#include <linux/clk.h> +#include <linux/interrupt.h> +#include <linux/irq.h> +#include <linux/io.h> +#include <linux/kernel.h> +#include <linux/mod_devicetable.h> +#include <linux/module.h> +#include <linux/platform_device.h> +#include <linux/pm_runtime.h> +#include <linux/slab.h> +#include <media/media-device.h> +#include <media/videobuf2-core.h> +#include <media/videobuf2-dma-contig.h> +#include <media/videobuf2-v4l2.h> +#include <media/v4l2-mem2mem.h> +#include <media/v4l2-dev.h> +#include <media/v4l2-device.h> +#include <media/v4l2-fh.h> +#include <media/v4l2-event.h> + +#include "mtk_jpeg_core.h" +#include "mtk_jpeg_enc_hw.h" + +static const struct mtk_jpeg_enc_qlt mtk_jpeg_enc_quality[] = { + {.quality_param = 34, .hardware_value = JPEG_ENC_QUALITY_Q34}, + {.quality_param = 39, .hardware_value = JPEG_ENC_QUALITY_Q39}, + {.quality_param = 48, .hardware_value = JPEG_ENC_QUALITY_Q48}, + {.quality_param = 60, .hardware_value = JPEG_ENC_QUALITY_Q60}, + {.quality_param = 64, .hardware_value = JPEG_ENC_QUALITY_Q64}, + {.quality_param = 68, .hardware_value = JPEG_ENC_QUALITY_Q68}, + {.quality_param = 74, .hardware_value = JPEG_ENC_QUALITY_Q74}, + {.quality_param = 80, .hardware_value = JPEG_ENC_QUALITY_Q80}, + {.quality_param = 82, .hardware_value = JPEG_ENC_QUALITY_Q82}, + {.quality_param = 84, .hardware_value = JPEG_ENC_QUALITY_Q84}, + {.quality_param = 87, .hardware_value = JPEG_ENC_QUALITY_Q87}, + {.quality_param = 90, .hardware_value = JPEG_ENC_QUALITY_Q90}, + {.quality_param = 92, .hardware_value = JPEG_ENC_QUALITY_Q92}, + {.quality_param = 95, .hardware_value = JPEG_ENC_QUALITY_Q95}, + {.quality_param = 97, .hardware_value = JPEG_ENC_QUALITY_Q97}, +}; + +static const struct of_device_id mtk_jpegenc_drv_ids[] = { + { + .compatible = "mediatek,mt8195-jpgenc-hw", + }, + {}, +}; +MODULE_DEVICE_TABLE(of, mtk_jpegenc_drv_ids); + +void mtk_jpeg_enc_reset(void __iomem *base) +{ + writel(0, base + JPEG_ENC_RSTB); + writel(JPEG_ENC_RESET_BIT, base + JPEG_ENC_RSTB); + writel(0, base + JPEG_ENC_CODEC_SEL); +} +EXPORT_SYMBOL_GPL(mtk_jpeg_enc_reset); + +u32 mtk_jpeg_enc_get_file_size(void __iomem *base) +{ + return readl(base + JPEG_ENC_DMA_ADDR0) - + readl(base + JPEG_ENC_DST_ADDR0); +} +EXPORT_SYMBOL_GPL(mtk_jpeg_enc_get_file_size); + +void mtk_jpeg_enc_start(void __iomem *base) +{ + u32 value; + + value = readl(base + JPEG_ENC_CTRL); + value |= JPEG_ENC_CTRL_INT_EN_BIT | JPEG_ENC_CTRL_ENABLE_BIT; + writel(value, base + JPEG_ENC_CTRL); +} +EXPORT_SYMBOL_GPL(mtk_jpeg_enc_start); + +void mtk_jpeg_set_enc_src(struct mtk_jpeg_ctx *ctx, void __iomem *base, + struct vb2_buffer *src_buf) +{ + int i; + dma_addr_t dma_addr; + + for (i = 0; i < src_buf->num_planes; i++) { + dma_addr = vb2_dma_contig_plane_dma_addr(src_buf, i) + + src_buf->planes[i].data_offset; + if (!i) + writel(dma_addr, base + JPEG_ENC_SRC_LUMA_ADDR); + else + writel(dma_addr, base + JPEG_ENC_SRC_CHROMA_ADDR); + } +} +EXPORT_SYMBOL_GPL(mtk_jpeg_set_enc_src); + +void mtk_jpeg_set_enc_dst(struct mtk_jpeg_ctx *ctx, void __iomem *base, + struct vb2_buffer *dst_buf) +{ + dma_addr_t dma_addr; + size_t size; + u32 dma_addr_offset; + u32 dma_addr_offsetmask; + + dma_addr = vb2_dma_contig_plane_dma_addr(dst_buf, 0); + dma_addr_offset = ctx->enable_exif ? MTK_JPEG_MAX_EXIF_SIZE : 0; + dma_addr_offsetmask = dma_addr & JPEG_ENC_DST_ADDR_OFFSET_MASK; + size = vb2_plane_size(dst_buf, 0); + + writel(dma_addr_offset & ~0xf, base + JPEG_ENC_OFFSET_ADDR); + writel(dma_addr_offsetmask & 0xf, base + JPEG_ENC_BYTE_OFFSET_MASK); + writel(dma_addr & ~0xf, base + JPEG_ENC_DST_ADDR0); + writel((dma_addr + size) & ~0xf, base + JPEG_ENC_STALL_ADDR0); +} +EXPORT_SYMBOL_GPL(mtk_jpeg_set_enc_dst); + +void mtk_jpeg_set_enc_params(struct mtk_jpeg_ctx *ctx, void __iomem *base) +{ + u32 value; + u32 width = ctx->out_q.enc_crop_rect.width; + u32 height = ctx->out_q.enc_crop_rect.height; + u32 enc_format = ctx->out_q.fmt->fourcc; + u32 bytesperline = ctx->out_q.pix_mp.plane_fmt[0].bytesperline; + u32 blk_num; + u32 img_stride; + u32 mem_stride; + u32 i, enc_quality; + u32 nr_enc_quality = ARRAY_SIZE(mtk_jpeg_enc_quality); + + value = width << 16 | height; + writel(value, base + JPEG_ENC_IMG_SIZE); + + if (enc_format == V4L2_PIX_FMT_NV12M || + enc_format == V4L2_PIX_FMT_NV21M) + /* + * Total 8 x 8 block number of luma and chroma. + * The number of blocks is counted from 0. + */ + blk_num = DIV_ROUND_UP(width, 16) * + DIV_ROUND_UP(height, 16) * 6 - 1; + else + blk_num = DIV_ROUND_UP(width, 16) * + DIV_ROUND_UP(height, 8) * 4 - 1; + writel(blk_num, base + JPEG_ENC_BLK_NUM); + + if (enc_format == V4L2_PIX_FMT_NV12M || + enc_format == V4L2_PIX_FMT_NV21M) { + /* 4:2:0 */ + img_stride = round_up(width, 16); + mem_stride = bytesperline; + } else { + /* 4:2:2 */ + img_stride = round_up(width * 2, 32); + mem_stride = img_stride; + } + writel(img_stride, base + JPEG_ENC_IMG_STRIDE); + writel(mem_stride, base + JPEG_ENC_STRIDE); + + enc_quality = mtk_jpeg_enc_quality[nr_enc_quality - 1].hardware_value; + for (i = 0; i < nr_enc_quality; i++) { + if (ctx->enc_quality <= mtk_jpeg_enc_quality[i].quality_param) { + enc_quality = mtk_jpeg_enc_quality[i].hardware_value; + break; + } + } + writel(enc_quality, base + JPEG_ENC_QUALITY); + + value = readl(base + JPEG_ENC_CTRL); + value &= ~JPEG_ENC_CTRL_YUV_FORMAT_MASK; + value |= (ctx->out_q.fmt->hw_format & 3) << 3; + if (ctx->enable_exif) + value |= JPEG_ENC_CTRL_FILE_FORMAT_BIT; + else + value &= ~JPEG_ENC_CTRL_FILE_FORMAT_BIT; + if (ctx->restart_interval) + value |= JPEG_ENC_CTRL_RESTART_EN_BIT; + else + value &= ~JPEG_ENC_CTRL_RESTART_EN_BIT; + writel(value, base + JPEG_ENC_CTRL); + + writel(ctx->restart_interval, base + JPEG_ENC_RST_MCU_NUM); +} +EXPORT_SYMBOL_GPL(mtk_jpeg_set_enc_params); + +static void mtk_jpegenc_put_buf(struct mtk_jpegenc_comp_dev *jpeg) +{ + struct mtk_jpeg_ctx *ctx; + struct vb2_v4l2_buffer *dst_buffer; + struct list_head *temp_entry; + struct list_head *pos = NULL; + struct mtk_jpeg_src_buf *dst_done_buf, *tmp_dst_done_buf; + unsigned long flags; + + ctx = jpeg->hw_param.curr_ctx; + if (!ctx) { + dev_err(jpeg->dev, "comp_jpeg ctx fail !!!\n"); + return; + } + + dst_buffer = jpeg->hw_param.dst_buffer; + if (!dst_buffer) { + dev_err(jpeg->dev, "comp_jpeg dst_buffer fail !!!\n"); + return; + } + + dst_done_buf = container_of(dst_buffer, + struct mtk_jpeg_src_buf, b); + + spin_lock_irqsave(&ctx->done_queue_lock, flags); + list_add_tail(&dst_done_buf->list, &ctx->dst_done_queue); + while (!list_empty(&ctx->dst_done_queue) && + (pos != &ctx->dst_done_queue)) { + list_for_each_prev_safe(pos, temp_entry, &ctx->dst_done_queue) { + tmp_dst_done_buf = list_entry(pos, + struct mtk_jpeg_src_buf, + list); + if (tmp_dst_done_buf->frame_num == + ctx->last_done_frame_num) { + list_del(&tmp_dst_done_buf->list); + v4l2_m2m_buf_done(&tmp_dst_done_buf->b, + VB2_BUF_STATE_DONE); + ctx->last_done_frame_num++; + } + } + } + spin_unlock_irqrestore(&ctx->done_queue_lock, flags); +} + +static void mtk_jpegenc_timeout_work(struct work_struct *work) +{ + struct delayed_work *dly_work = to_delayed_work(work); + struct mtk_jpegenc_comp_dev *cjpeg = + container_of(dly_work, + struct mtk_jpegenc_comp_dev, + job_timeout_work); + struct mtk_jpeg_dev *master_jpeg = cjpeg->master_dev; + enum vb2_buffer_state buf_state = VB2_BUF_STATE_ERROR; + struct vb2_v4l2_buffer *src_buf, *dst_buf; + + src_buf = cjpeg->hw_param.src_buffer; + dst_buf = cjpeg->hw_param.dst_buffer; + v4l2_m2m_buf_copy_metadata(src_buf, dst_buf, true); + + mtk_jpeg_enc_reset(cjpeg->reg_base); + clk_disable_unprepare(cjpeg->venc_clk.clks->clk); + pm_runtime_put(cjpeg->dev); + cjpeg->hw_state = MTK_JPEG_HW_IDLE; + atomic_inc(&master_jpeg->hw_rdy); + wake_up(&master_jpeg->hw_wq); + v4l2_m2m_buf_done(src_buf, buf_state); + mtk_jpegenc_put_buf(cjpeg); +} + +static irqreturn_t mtk_jpegenc_hw_irq_handler(int irq, void *priv) +{ + struct vb2_v4l2_buffer *src_buf, *dst_buf; + enum vb2_buffer_state buf_state; + struct mtk_jpeg_ctx *ctx; + u32 result_size; + u32 irq_status; + + struct mtk_jpegenc_comp_dev *jpeg = priv; + struct mtk_jpeg_dev *master_jpeg = jpeg->master_dev; + + cancel_delayed_work(&jpeg->job_timeout_work); + + ctx = jpeg->hw_param.curr_ctx; + src_buf = jpeg->hw_param.src_buffer; + dst_buf = jpeg->hw_param.dst_buffer; + v4l2_m2m_buf_copy_metadata(src_buf, dst_buf, true); + + irq_status = readl(jpeg->reg_base + JPEG_ENC_INT_STS) & + JPEG_ENC_INT_STATUS_MASK_ALLIRQ; + if (irq_status) + writel(0, jpeg->reg_base + JPEG_ENC_INT_STS); + if (!(irq_status & JPEG_ENC_INT_STATUS_DONE)) + dev_warn(jpeg->dev, "Jpg Enc occurs unknown Err."); + + result_size = mtk_jpeg_enc_get_file_size(jpeg->reg_base); + vb2_set_plane_payload(&dst_buf->vb2_buf, 0, result_size); + buf_state = VB2_BUF_STATE_DONE; + v4l2_m2m_buf_done(src_buf, buf_state); + mtk_jpegenc_put_buf(jpeg); + pm_runtime_put(ctx->jpeg->dev); + clk_disable_unprepare(jpeg->venc_clk.clks->clk); + + jpeg->hw_state = MTK_JPEG_HW_IDLE; + wake_up(&master_jpeg->hw_wq); + atomic_inc(&master_jpeg->hw_rdy); + + return IRQ_HANDLED; +} + +static int mtk_jpegenc_hw_init_irq(struct mtk_jpegenc_comp_dev *dev) +{ + struct platform_device *pdev = dev->plat_dev; + int ret; + + dev->jpegenc_irq = platform_get_irq(pdev, 0); + if (dev->jpegenc_irq < 0) + return dev->jpegenc_irq; + + ret = devm_request_irq(&pdev->dev, + dev->jpegenc_irq, + mtk_jpegenc_hw_irq_handler, + 0, + pdev->name, dev); + if (ret) { + dev_err(&pdev->dev, "Failed to devm_request_irq %d (%d)", + dev->jpegenc_irq, ret); + return ret; + } + + return 0; +} + +static int mtk_jpegenc_hw_probe(struct platform_device *pdev) +{ + struct mtk_jpegenc_clk *jpegenc_clk; + struct mtk_jpeg_dev *master_dev; + struct mtk_jpegenc_comp_dev *dev; + int ret, i; + + struct device *decs = &pdev->dev; + + if (!decs->parent) + return -EPROBE_DEFER; + + master_dev = dev_get_drvdata(decs->parent); + if (!master_dev) + return -EPROBE_DEFER; + + dev = devm_kzalloc(&pdev->dev, sizeof(*dev), GFP_KERNEL); + if (!dev) + return -ENOMEM; + + dev->plat_dev = pdev; + dev->dev = &pdev->dev; + + spin_lock_init(&dev->hw_lock); + dev->hw_state = MTK_JPEG_HW_IDLE; + + INIT_DELAYED_WORK(&dev->job_timeout_work, + mtk_jpegenc_timeout_work); + + jpegenc_clk = &dev->venc_clk; + + jpegenc_clk->clk_num = devm_clk_bulk_get_all(&pdev->dev, + &jpegenc_clk->clks); + if (jpegenc_clk->clk_num < 0) + return dev_err_probe(&pdev->dev, jpegenc_clk->clk_num, + "Failed to get jpegenc clock count\n"); + + dev->reg_base = devm_platform_ioremap_resource(pdev, 0); + if (IS_ERR(dev->reg_base)) + return PTR_ERR(dev->reg_base); + + ret = mtk_jpegenc_hw_init_irq(dev); + if (ret) + return ret; + + i = atomic_add_return(1, &master_dev->hw_index) - 1; + master_dev->enc_hw_dev[i] = dev; + master_dev->reg_encbase[i] = dev->reg_base; + dev->master_dev = master_dev; + + platform_set_drvdata(pdev, dev); + pm_runtime_enable(&pdev->dev); + + return 0; +} + +static struct platform_driver mtk_jpegenc_hw_driver = { + .probe = mtk_jpegenc_hw_probe, + .driver = { + .name = "mtk-jpegenc-hw", + .of_match_table = mtk_jpegenc_drv_ids, + }, +}; + +module_platform_driver(mtk_jpegenc_hw_driver); + +MODULE_DESCRIPTION("MediaTek JPEG encode HW driver"); +MODULE_LICENSE("GPL"); diff --git a/drivers/media/platform/mediatek/jpeg/mtk_jpeg_enc_hw.h b/drivers/media/platform/mediatek/jpeg/mtk_jpeg_enc_hw.h new file mode 100644 index 0000000000..61c60e4e58 --- /dev/null +++ b/drivers/media/platform/mediatek/jpeg/mtk_jpeg_enc_hw.h @@ -0,0 +1,91 @@ +/* SPDX-License-Identifier: GPL-2.0-only */ +/* + * Copyright (c) 2019 MediaTek Inc. + * Author: Xia Jiang <xia.jiang@mediatek.com> + * + */ + +#ifndef _MTK_JPEG_ENC_HW_H +#define _MTK_JPEG_ENC_HW_H + +#include <media/videobuf2-core.h> + +#include "mtk_jpeg_core.h" + +#define JPEG_ENC_INT_STATUS_DONE BIT(0) +#define JPEG_ENC_INT_STATUS_MASK_ALLIRQ 0x13 + +#define JPEG_ENC_DST_ADDR_OFFSET_MASK GENMASK(3, 0) + +#define JPEG_ENC_CTRL_YUV_FORMAT_MASK 0x18 +#define JPEG_ENC_CTRL_RESTART_EN_BIT BIT(10) +#define JPEG_ENC_CTRL_FILE_FORMAT_BIT BIT(5) +#define JPEG_ENC_CTRL_INT_EN_BIT BIT(2) +#define JPEG_ENC_CTRL_ENABLE_BIT BIT(0) +#define JPEG_ENC_RESET_BIT BIT(0) + +#define JPEG_ENC_YUV_FORMAT_YUYV 0 +#define JPEG_ENC_YUV_FORMAT_YVYU 1 +#define JPEG_ENC_YUV_FORMAT_NV12 2 +#define JEPG_ENC_YUV_FORMAT_NV21 3 + +#define JPEG_ENC_QUALITY_Q60 0x0 +#define JPEG_ENC_QUALITY_Q80 0x1 +#define JPEG_ENC_QUALITY_Q90 0x2 +#define JPEG_ENC_QUALITY_Q95 0x3 +#define JPEG_ENC_QUALITY_Q39 0x4 +#define JPEG_ENC_QUALITY_Q68 0x5 +#define JPEG_ENC_QUALITY_Q84 0x6 +#define JPEG_ENC_QUALITY_Q92 0x7 +#define JPEG_ENC_QUALITY_Q48 0x8 +#define JPEG_ENC_QUALITY_Q74 0xa +#define JPEG_ENC_QUALITY_Q87 0xb +#define JPEG_ENC_QUALITY_Q34 0xc +#define JPEG_ENC_QUALITY_Q64 0xe +#define JPEG_ENC_QUALITY_Q82 0xf +#define JPEG_ENC_QUALITY_Q97 0x10 + +#define JPEG_ENC_RSTB 0x100 +#define JPEG_ENC_CTRL 0x104 +#define JPEG_ENC_QUALITY 0x108 +#define JPEG_ENC_BLK_NUM 0x10C +#define JPEG_ENC_BLK_CNT 0x110 +#define JPEG_ENC_INT_STS 0x11c +#define JPEG_ENC_DST_ADDR0 0x120 +#define JPEG_ENC_DMA_ADDR0 0x124 +#define JPEG_ENC_STALL_ADDR0 0x128 +#define JPEG_ENC_OFFSET_ADDR 0x138 +#define JPEG_ENC_RST_MCU_NUM 0x150 +#define JPEG_ENC_IMG_SIZE 0x154 +#define JPEG_ENC_DEBUG_INFO0 0x160 +#define JPEG_ENC_DEBUG_INFO1 0x164 +#define JPEG_ENC_TOTAL_CYCLE 0x168 +#define JPEG_ENC_BYTE_OFFSET_MASK 0x16c +#define JPEG_ENC_SRC_LUMA_ADDR 0x170 +#define JPEG_ENC_SRC_CHROMA_ADDR 0x174 +#define JPEG_ENC_STRIDE 0x178 +#define JPEG_ENC_IMG_STRIDE 0x17c +#define JPEG_ENC_DCM_CTRL 0x300 +#define JPEG_ENC_CODEC_SEL 0x314 +#define JPEG_ENC_ULTRA_THRES 0x318 + +/** + * struct mtk_jpeg_enc_qlt - JPEG encoder quality data + * @quality_param: quality value + * @hardware_value: hardware value of quality + */ +struct mtk_jpeg_enc_qlt { + u8 quality_param; + u8 hardware_value; +}; + +void mtk_jpeg_enc_reset(void __iomem *base); +u32 mtk_jpeg_enc_get_file_size(void __iomem *base); +void mtk_jpeg_enc_start(void __iomem *enc_reg_base); +void mtk_jpeg_set_enc_src(struct mtk_jpeg_ctx *ctx, void __iomem *base, + struct vb2_buffer *src_buf); +void mtk_jpeg_set_enc_dst(struct mtk_jpeg_ctx *ctx, void __iomem *base, + struct vb2_buffer *dst_buf); +void mtk_jpeg_set_enc_params(struct mtk_jpeg_ctx *ctx, void __iomem *base); + +#endif /* _MTK_JPEG_ENC_HW_H */ diff --git a/drivers/media/platform/mediatek/mdp/Kconfig b/drivers/media/platform/mediatek/mdp/Kconfig new file mode 100644 index 0000000000..9f13a42899 --- /dev/null +++ b/drivers/media/platform/mediatek/mdp/Kconfig @@ -0,0 +1,17 @@ +# SPDX-License-Identifier: GPL-2.0-only +config VIDEO_MEDIATEK_MDP + tristate "Mediatek MDP driver" + depends on V4L_MEM2MEM_DRIVERS + depends on MTK_IOMMU || COMPILE_TEST + depends on VIDEO_DEV + depends on ARCH_MEDIATEK || COMPILE_TEST + depends on MTK_SMI || (COMPILE_TEST && MTK_SMI=n) + select VIDEOBUF2_DMA_CONTIG + select V4L2_MEM2MEM_DEV + select VIDEO_MEDIATEK_VPU + help + It is a v4l2 driver and present in Mediatek MT8173 SoCs. + The driver supports for scaling and color space conversion. + + To compile this driver as a module, choose M here: the + module will be called mtk-mdp. diff --git a/drivers/media/platform/mediatek/mdp/Makefile b/drivers/media/platform/mediatek/mdp/Makefile new file mode 100644 index 0000000000..b7c16ebecc --- /dev/null +++ b/drivers/media/platform/mediatek/mdp/Makefile @@ -0,0 +1,10 @@ +# SPDX-License-Identifier: GPL-2.0 +mtk-mdp-y += mtk_mdp_core.o +mtk-mdp-y += mtk_mdp_comp.o +mtk-mdp-y += mtk_mdp_m2m.o +mtk-mdp-y += mtk_mdp_regs.o +mtk-mdp-y += mtk_mdp_vpu.o + +obj-$(CONFIG_VIDEO_MEDIATEK_MDP) += mtk-mdp.o + +ccflags-y += -I$(srctree)/drivers/media/platform/mediatek/vpu diff --git a/drivers/media/platform/mediatek/mdp/mtk_mdp_comp.c b/drivers/media/platform/mediatek/mdp/mtk_mdp_comp.c new file mode 100644 index 0000000000..3501ac4112 --- /dev/null +++ b/drivers/media/platform/mediatek/mdp/mtk_mdp_comp.c @@ -0,0 +1,74 @@ +// SPDX-License-Identifier: GPL-2.0-only +/* + * Copyright (c) 2016 MediaTek Inc. + * Author: Ming Hsiu Tsai <minghsiu.tsai@mediatek.com> + */ + +#include <linux/clk.h> +#include <linux/device.h> +#include <linux/of.h> + +#include "mtk_mdp_comp.h" + + +void mtk_mdp_comp_clock_on(struct device *dev, struct mtk_mdp_comp *comp) +{ + int i, err; + + for (i = 0; i < ARRAY_SIZE(comp->clk); i++) { + if (IS_ERR(comp->clk[i])) + continue; + err = clk_prepare_enable(comp->clk[i]); + if (err) + dev_err(dev, + "failed to enable clock, err %d. type:%d i:%d\n", + err, comp->type, i); + } +} + +void mtk_mdp_comp_clock_off(struct device *dev, struct mtk_mdp_comp *comp) +{ + int i; + + for (i = 0; i < ARRAY_SIZE(comp->clk); i++) { + if (IS_ERR(comp->clk[i])) + continue; + clk_disable_unprepare(comp->clk[i]); + } +} + +int mtk_mdp_comp_init(struct device *dev, struct device_node *node, + struct mtk_mdp_comp *comp, + enum mtk_mdp_comp_type comp_type) +{ + int ret; + int i; + + comp->dev_node = of_node_get(node); + comp->type = comp_type; + + for (i = 0; i < ARRAY_SIZE(comp->clk); i++) { + comp->clk[i] = of_clk_get(node, i); + if (IS_ERR(comp->clk[i])) { + ret = dev_err_probe(dev, PTR_ERR(comp->clk[i]), + "Failed to get clock\n"); + goto put_dev; + } + + /* Only RDMA needs two clocks */ + if (comp->type != MTK_MDP_RDMA) + break; + } + + return 0; + +put_dev: + of_node_put(comp->dev_node); + + return ret; +} + +void mtk_mdp_comp_deinit(struct device *dev, struct mtk_mdp_comp *comp) +{ + of_node_put(comp->dev_node); +} diff --git a/drivers/media/platform/mediatek/mdp/mtk_mdp_comp.h b/drivers/media/platform/mediatek/mdp/mtk_mdp_comp.h new file mode 100644 index 0000000000..ae41dd3cd7 --- /dev/null +++ b/drivers/media/platform/mediatek/mdp/mtk_mdp_comp.h @@ -0,0 +1,46 @@ +/* SPDX-License-Identifier: GPL-2.0-only */ +/* + * Copyright (c) 2016 MediaTek Inc. + * Author: Ming Hsiu Tsai <minghsiu.tsai@mediatek.com> + */ + +#ifndef __MTK_MDP_COMP_H__ +#define __MTK_MDP_COMP_H__ + +/** + * enum mtk_mdp_comp_type - the MDP component + * @MTK_MDP_RDMA: Read DMA + * @MTK_MDP_RSZ: Riszer + * @MTK_MDP_WDMA: Write DMA + * @MTK_MDP_WROT: Write DMA with rotation + */ +enum mtk_mdp_comp_type { + MTK_MDP_RDMA, + MTK_MDP_RSZ, + MTK_MDP_WDMA, + MTK_MDP_WROT, +}; + +/** + * struct mtk_mdp_comp - the MDP's function component data + * @node: list node to track sibing MDP components + * @dev_node: component device node + * @clk: clocks required for component + * @type: component type + */ +struct mtk_mdp_comp { + struct list_head node; + struct device_node *dev_node; + struct clk *clk[2]; + enum mtk_mdp_comp_type type; +}; + +int mtk_mdp_comp_init(struct device *dev, struct device_node *node, + struct mtk_mdp_comp *comp, + enum mtk_mdp_comp_type comp_type); +void mtk_mdp_comp_deinit(struct device *dev, struct mtk_mdp_comp *comp); +void mtk_mdp_comp_clock_on(struct device *dev, struct mtk_mdp_comp *comp); +void mtk_mdp_comp_clock_off(struct device *dev, struct mtk_mdp_comp *comp); + + +#endif /* __MTK_MDP_COMP_H__ */ diff --git a/drivers/media/platform/mediatek/mdp/mtk_mdp_core.c b/drivers/media/platform/mediatek/mdp/mtk_mdp_core.c new file mode 100644 index 0000000000..917cdf38f2 --- /dev/null +++ b/drivers/media/platform/mediatek/mdp/mtk_mdp_core.c @@ -0,0 +1,313 @@ +// SPDX-License-Identifier: GPL-2.0-only +/* + * Copyright (c) 2015-2016 MediaTek Inc. + * Author: Houlong Wei <houlong.wei@mediatek.com> + * Ming Hsiu Tsai <minghsiu.tsai@mediatek.com> + */ + +#include <linux/clk.h> +#include <linux/device.h> +#include <linux/errno.h> +#include <linux/interrupt.h> +#include <linux/kernel.h> +#include <linux/module.h> +#include <linux/of.h> +#include <linux/of_address.h> +#include <linux/of_platform.h> +#include <linux/platform_device.h> +#include <linux/pm_runtime.h> +#include <linux/workqueue.h> + +#include "mtk_mdp_core.h" +#include "mtk_mdp_m2m.h" +#include "mtk_vpu.h" + +/* MDP debug log level (0-3). 3 shows all the logs. */ +int mtk_mdp_dbg_level; +EXPORT_SYMBOL(mtk_mdp_dbg_level); + +module_param(mtk_mdp_dbg_level, int, 0644); + +static const struct of_device_id mtk_mdp_comp_dt_ids[] __maybe_unused = { + { + .compatible = "mediatek,mt8173-mdp-rdma", + .data = (void *)MTK_MDP_RDMA + }, { + .compatible = "mediatek,mt8173-mdp-rsz", + .data = (void *)MTK_MDP_RSZ + }, { + .compatible = "mediatek,mt8173-mdp-wdma", + .data = (void *)MTK_MDP_WDMA + }, { + .compatible = "mediatek,mt8173-mdp-wrot", + .data = (void *)MTK_MDP_WROT + }, + { }, +}; + +static const struct of_device_id mtk_mdp_of_ids[] = { + { .compatible = "mediatek,mt8173-mdp", }, + { }, +}; +MODULE_DEVICE_TABLE(of, mtk_mdp_of_ids); + +static void mtk_mdp_clock_on(struct mtk_mdp_dev *mdp) +{ + struct device *dev = &mdp->pdev->dev; + struct mtk_mdp_comp *comp_node; + + list_for_each_entry(comp_node, &mdp->comp_list, node) + mtk_mdp_comp_clock_on(dev, comp_node); +} + +static void mtk_mdp_clock_off(struct mtk_mdp_dev *mdp) +{ + struct device *dev = &mdp->pdev->dev; + struct mtk_mdp_comp *comp_node; + + list_for_each_entry(comp_node, &mdp->comp_list, node) + mtk_mdp_comp_clock_off(dev, comp_node); +} + +static void mtk_mdp_wdt_worker(struct work_struct *work) +{ + struct mtk_mdp_dev *mdp = + container_of(work, struct mtk_mdp_dev, wdt_work); + struct mtk_mdp_ctx *ctx; + + mtk_mdp_err("Watchdog timeout"); + + list_for_each_entry(ctx, &mdp->ctx_list, list) { + mtk_mdp_dbg(0, "[%d] Change as state error", ctx->id); + mtk_mdp_ctx_state_lock_set(ctx, MTK_MDP_CTX_ERROR); + } +} + +static void mtk_mdp_reset_handler(void *priv) +{ + struct mtk_mdp_dev *mdp = priv; + + queue_work(mdp->wdt_wq, &mdp->wdt_work); +} + +void mtk_mdp_register_component(struct mtk_mdp_dev *mdp, + struct mtk_mdp_comp *comp) +{ + list_add(&comp->node, &mdp->comp_list); +} + +void mtk_mdp_unregister_component(struct mtk_mdp_dev *mdp, + struct mtk_mdp_comp *comp) +{ + list_del(&comp->node); +} + +static int mtk_mdp_probe(struct platform_device *pdev) +{ + struct mtk_mdp_dev *mdp; + struct device *dev = &pdev->dev; + struct device_node *node, *parent; + struct mtk_mdp_comp *comp, *comp_temp; + int ret = 0; + + mdp = devm_kzalloc(dev, sizeof(*mdp), GFP_KERNEL); + if (!mdp) + return -ENOMEM; + + mdp->id = pdev->id; + mdp->pdev = pdev; + INIT_LIST_HEAD(&mdp->comp_list); + INIT_LIST_HEAD(&mdp->ctx_list); + + mutex_init(&mdp->lock); + mutex_init(&mdp->vpulock); + + /* Old dts had the components as child nodes */ + node = of_get_next_child(dev->of_node, NULL); + if (node) { + of_node_put(node); + parent = dev->of_node; + dev_warn(dev, "device tree is out of date\n"); + } else { + parent = dev->of_node->parent; + } + + /* Iterate over sibling MDP function blocks */ + for_each_child_of_node(parent, node) { + const struct of_device_id *of_id; + enum mtk_mdp_comp_type comp_type; + + of_id = of_match_node(mtk_mdp_comp_dt_ids, node); + if (!of_id) + continue; + + if (!of_device_is_available(node)) { + dev_err(dev, "Skipping disabled component %pOF\n", + node); + continue; + } + + comp_type = (uintptr_t)of_id->data; + + comp = devm_kzalloc(dev, sizeof(*comp), GFP_KERNEL); + if (!comp) { + ret = -ENOMEM; + of_node_put(node); + goto err_comp; + } + + ret = mtk_mdp_comp_init(dev, node, comp, comp_type); + if (ret) { + of_node_put(node); + goto err_comp; + } + + mtk_mdp_register_component(mdp, comp); + } + + mdp->job_wq = create_singlethread_workqueue(MTK_MDP_MODULE_NAME); + if (!mdp->job_wq) { + dev_err(&pdev->dev, "unable to alloc job workqueue\n"); + ret = -ENOMEM; + goto err_alloc_job_wq; + } + + mdp->wdt_wq = create_singlethread_workqueue("mdp_wdt_wq"); + if (!mdp->wdt_wq) { + dev_err(&pdev->dev, "unable to alloc wdt workqueue\n"); + ret = -ENOMEM; + goto err_alloc_wdt_wq; + } + INIT_WORK(&mdp->wdt_work, mtk_mdp_wdt_worker); + + ret = v4l2_device_register(dev, &mdp->v4l2_dev); + if (ret) { + dev_err(&pdev->dev, "Failed to register v4l2 device\n"); + ret = -EINVAL; + goto err_dev_register; + } + + ret = mtk_mdp_register_m2m_device(mdp); + if (ret) { + v4l2_err(&mdp->v4l2_dev, "Failed to init mem2mem device\n"); + goto err_m2m_register; + } + + mdp->vpu_dev = vpu_get_plat_device(pdev); + ret = vpu_wdt_reg_handler(mdp->vpu_dev, mtk_mdp_reset_handler, mdp, + VPU_RST_MDP); + if (ret) { + dev_err(&pdev->dev, "Failed to register reset handler\n"); + goto err_m2m_register; + } + + platform_set_drvdata(pdev, mdp); + + ret = vb2_dma_contig_set_max_seg_size(&pdev->dev, DMA_BIT_MASK(32)); + if (ret) { + dev_err(&pdev->dev, "Failed to set vb2 dma mag seg size\n"); + goto err_m2m_register; + } + + pm_runtime_enable(dev); + dev_dbg(dev, "mdp-%d registered successfully\n", mdp->id); + + return 0; + +err_m2m_register: + v4l2_device_unregister(&mdp->v4l2_dev); + +err_dev_register: + destroy_workqueue(mdp->wdt_wq); + +err_alloc_wdt_wq: + destroy_workqueue(mdp->job_wq); + +err_alloc_job_wq: + +err_comp: + list_for_each_entry_safe(comp, comp_temp, &mdp->comp_list, node) { + mtk_mdp_unregister_component(mdp, comp); + mtk_mdp_comp_deinit(dev, comp); + } + + dev_dbg(dev, "err %d\n", ret); + return ret; +} + +static void mtk_mdp_remove(struct platform_device *pdev) +{ + struct mtk_mdp_dev *mdp = platform_get_drvdata(pdev); + struct mtk_mdp_comp *comp, *comp_temp; + + pm_runtime_disable(&pdev->dev); + vb2_dma_contig_clear_max_seg_size(&pdev->dev); + mtk_mdp_unregister_m2m_device(mdp); + v4l2_device_unregister(&mdp->v4l2_dev); + + destroy_workqueue(mdp->wdt_wq); + + destroy_workqueue(mdp->job_wq); + + list_for_each_entry_safe(comp, comp_temp, &mdp->comp_list, node) { + mtk_mdp_unregister_component(mdp, comp); + mtk_mdp_comp_deinit(&pdev->dev, comp); + } + + dev_dbg(&pdev->dev, "%s driver unloaded\n", pdev->name); +} + +static int __maybe_unused mtk_mdp_pm_suspend(struct device *dev) +{ + struct mtk_mdp_dev *mdp = dev_get_drvdata(dev); + + mtk_mdp_clock_off(mdp); + + return 0; +} + +static int __maybe_unused mtk_mdp_pm_resume(struct device *dev) +{ + struct mtk_mdp_dev *mdp = dev_get_drvdata(dev); + + mtk_mdp_clock_on(mdp); + + return 0; +} + +static int __maybe_unused mtk_mdp_suspend(struct device *dev) +{ + if (pm_runtime_suspended(dev)) + return 0; + + return mtk_mdp_pm_suspend(dev); +} + +static int __maybe_unused mtk_mdp_resume(struct device *dev) +{ + if (pm_runtime_suspended(dev)) + return 0; + + return mtk_mdp_pm_resume(dev); +} + +static const struct dev_pm_ops mtk_mdp_pm_ops = { + SET_SYSTEM_SLEEP_PM_OPS(mtk_mdp_suspend, mtk_mdp_resume) + SET_RUNTIME_PM_OPS(mtk_mdp_pm_suspend, mtk_mdp_pm_resume, NULL) +}; + +static struct platform_driver mtk_mdp_driver = { + .probe = mtk_mdp_probe, + .remove_new = mtk_mdp_remove, + .driver = { + .name = MTK_MDP_MODULE_NAME, + .pm = &mtk_mdp_pm_ops, + .of_match_table = mtk_mdp_of_ids, + } +}; + +module_platform_driver(mtk_mdp_driver); + +MODULE_AUTHOR("Houlong Wei <houlong.wei@mediatek.com>"); +MODULE_DESCRIPTION("Mediatek image processor driver"); +MODULE_LICENSE("GPL v2"); diff --git a/drivers/media/platform/mediatek/mdp/mtk_mdp_core.h b/drivers/media/platform/mediatek/mdp/mtk_mdp_core.h new file mode 100644 index 0000000000..a6e6dc3630 --- /dev/null +++ b/drivers/media/platform/mediatek/mdp/mtk_mdp_core.h @@ -0,0 +1,256 @@ +/* SPDX-License-Identifier: GPL-2.0-only */ +/* + * Copyright (c) 2015-2016 MediaTek Inc. + * Author: Houlong Wei <houlong.wei@mediatek.com> + * Ming Hsiu Tsai <minghsiu.tsai@mediatek.com> + */ + +#ifndef __MTK_MDP_CORE_H__ +#define __MTK_MDP_CORE_H__ + +#include <linux/videodev2.h> +#include <media/v4l2-ctrls.h> +#include <media/v4l2-device.h> +#include <media/v4l2-mem2mem.h> +#include <media/videobuf2-core.h> +#include <media/videobuf2-dma-contig.h> + +#include "mtk_mdp_vpu.h" +#include "mtk_mdp_comp.h" + + +#define MTK_MDP_MODULE_NAME "mtk-mdp" + +#define MTK_MDP_SHUTDOWN_TIMEOUT ((100*HZ)/1000) /* 100ms */ +#define MTK_MDP_MAX_CTRL_NUM 10 + +#define MTK_MDP_FMT_FLAG_OUTPUT BIT(0) +#define MTK_MDP_FMT_FLAG_CAPTURE BIT(1) + +#define MTK_MDP_VPU_INIT BIT(0) +#define MTK_MDP_CTX_ERROR BIT(5) + +/** + * struct mtk_mdp_pix_align - alignment of image + * @org_w: source alignment of width + * @org_h: source alignment of height + * @target_w: dst alignment of width + * @target_h: dst alignment of height + */ +struct mtk_mdp_pix_align { + u16 org_w; + u16 org_h; + u16 target_w; + u16 target_h; +}; + +/** + * struct mtk_mdp_fmt - the driver's internal color format data + * @pixelformat: the fourcc code for this format, 0 if not applicable + * @num_planes: number of physically non-contiguous data planes + * @num_comp: number of logical data planes + * @depth: per plane driver's private 'number of bits per pixel' + * @row_depth: per plane driver's private 'number of bits per pixel per row' + * @flags: flags indicating which operation mode format applies to + * MTK_MDP_FMT_FLAG_OUTPUT is used in OUTPUT stream + * MTK_MDP_FMT_FLAG_CAPTURE is used in CAPTURE stream + * @align: pointer to a pixel alignment struct, NULL if using default value + */ +struct mtk_mdp_fmt { + u32 pixelformat; + u16 num_planes; + u16 num_comp; + u8 depth[VIDEO_MAX_PLANES]; + u8 row_depth[VIDEO_MAX_PLANES]; + u32 flags; + struct mtk_mdp_pix_align *align; +}; + +/** + * struct mtk_mdp_addr - the image processor physical address set + * @addr: address of planes + */ +struct mtk_mdp_addr { + dma_addr_t addr[MTK_MDP_MAX_NUM_PLANE]; +}; + +/* struct mtk_mdp_ctrls - the image processor control set + * @rotate: rotation degree + * @hflip: horizontal flip + * @vflip: vertical flip + * @global_alpha: the alpha value of current frame + */ +struct mtk_mdp_ctrls { + struct v4l2_ctrl *rotate; + struct v4l2_ctrl *hflip; + struct v4l2_ctrl *vflip; + struct v4l2_ctrl *global_alpha; +}; + +/** + * struct mtk_mdp_frame - source/target frame properties + * @width: SRC : SRCIMG_WIDTH, DST : OUTPUTDMA_WHOLE_IMG_WIDTH + * @height: SRC : SRCIMG_HEIGHT, DST : OUTPUTDMA_WHOLE_IMG_HEIGHT + * @crop: cropped(source)/scaled(destination) size + * @payload: image size in bytes (w x h x bpp) + * @pitch: bytes per line of image in memory + * @addr: image frame buffer physical addresses + * @fmt: color format pointer + * @alpha: frame's alpha value + */ +struct mtk_mdp_frame { + u32 width; + u32 height; + struct v4l2_rect crop; + unsigned long payload[VIDEO_MAX_PLANES]; + unsigned int pitch[VIDEO_MAX_PLANES]; + struct mtk_mdp_addr addr; + const struct mtk_mdp_fmt *fmt; + u8 alpha; +}; + +/** + * struct mtk_mdp_variant - image processor variant information + * @pix_max: maximum limit of image size + * @pix_min: minimum limit of image size + * @pix_align: alignment of image + * @h_scale_up_max: maximum scale-up in horizontal + * @v_scale_up_max: maximum scale-up in vertical + * @h_scale_down_max: maximum scale-down in horizontal + * @v_scale_down_max: maximum scale-down in vertical + */ +struct mtk_mdp_variant { + struct mtk_mdp_pix_limit *pix_max; + struct mtk_mdp_pix_limit *pix_min; + struct mtk_mdp_pix_align *pix_align; + u16 h_scale_up_max; + u16 v_scale_up_max; + u16 h_scale_down_max; + u16 v_scale_down_max; +}; + +/** + * struct mtk_mdp_dev - abstraction for image processor entity + * @lock: the mutex protecting this data structure + * @vpulock: the mutex protecting the communication with VPU + * @pdev: pointer to the image processor platform device + * @variant: the IP variant information + * @id: image processor device index (0..MTK_MDP_MAX_DEVS) + * @comp_list: list of MDP function components + * @m2m_dev: v4l2 memory-to-memory device data + * @ctx_list: list of struct mtk_mdp_ctx + * @vdev: video device for image processor driver + * @v4l2_dev: V4L2 device to register video devices for. + * @job_wq: processor work queue + * @vpu_dev: VPU platform device + * @ctx_num: counter of active MTK MDP context + * @id_counter: An integer id given to the next opened context + * @wdt_wq: work queue for VPU watchdog + * @wdt_work: worker for VPU watchdog + */ +struct mtk_mdp_dev { + struct mutex lock; + struct mutex vpulock; + struct platform_device *pdev; + struct mtk_mdp_variant *variant; + u16 id; + struct list_head comp_list; + struct v4l2_m2m_dev *m2m_dev; + struct list_head ctx_list; + struct video_device *vdev; + struct v4l2_device v4l2_dev; + struct workqueue_struct *job_wq; + struct platform_device *vpu_dev; + int ctx_num; + unsigned long id_counter; + struct workqueue_struct *wdt_wq; + struct work_struct wdt_work; +}; + +/** + * struct mtk_mdp_ctx - the device context data + * @list: link to ctx_list of mtk_mdp_dev + * @s_frame: source frame properties + * @d_frame: destination frame properties + * @id: index of the context that this structure describes + * @flags: additional flags for image conversion + * @state: flags to keep track of user configuration + * Protected by slock + * @rotation: rotates the image by specified angle + * @hflip: mirror the picture horizontally + * @vflip: mirror the picture vertically + * @mdp_dev: the image processor device this context applies to + * @m2m_ctx: memory-to-memory device context + * @fh: v4l2 file handle + * @ctrl_handler: v4l2 controls handler + * @ctrls: image processor control set + * @ctrls_rdy: true if the control handler is initialized + * @colorspace: enum v4l2_colorspace; supplemental to pixelformat + * @ycbcr_enc: enum v4l2_ycbcr_encoding, Y'CbCr encoding + * @xfer_func: enum v4l2_xfer_func, colorspace transfer function + * @quant: enum v4l2_quantization, colorspace quantization + * @vpu: VPU instance + * @slock: the mutex protecting mtp_mdp_ctx.state + * @work: worker for image processing + */ +struct mtk_mdp_ctx { + struct list_head list; + struct mtk_mdp_frame s_frame; + struct mtk_mdp_frame d_frame; + u32 flags; + u32 state; + int id; + int rotation; + u32 hflip:1; + u32 vflip:1; + struct mtk_mdp_dev *mdp_dev; + struct v4l2_m2m_ctx *m2m_ctx; + struct v4l2_fh fh; + struct v4l2_ctrl_handler ctrl_handler; + struct mtk_mdp_ctrls ctrls; + bool ctrls_rdy; + enum v4l2_colorspace colorspace; + enum v4l2_ycbcr_encoding ycbcr_enc; + enum v4l2_xfer_func xfer_func; + enum v4l2_quantization quant; + + struct mtk_mdp_vpu vpu; + struct mutex slock; + struct work_struct work; +}; + +extern int mtk_mdp_dbg_level; + +void mtk_mdp_register_component(struct mtk_mdp_dev *mdp, + struct mtk_mdp_comp *comp); + +void mtk_mdp_unregister_component(struct mtk_mdp_dev *mdp, + struct mtk_mdp_comp *comp); + +#if defined(DEBUG) + +#define mtk_mdp_dbg(level, fmt, args...) \ + do { \ + if (mtk_mdp_dbg_level >= level) \ + pr_info("[MTK_MDP] level=%d %s(),%d: " fmt "\n", \ + level, __func__, __LINE__, ##args); \ + } while (0) + +#define mtk_mdp_err(fmt, args...) \ + pr_err("[MTK_MDP][ERROR] %s:%d: " fmt "\n", __func__, __LINE__, \ + ##args) + + +#define mtk_mdp_dbg_enter() mtk_mdp_dbg(3, "+") +#define mtk_mdp_dbg_leave() mtk_mdp_dbg(3, "-") + +#else + +#define mtk_mdp_dbg(level, fmt, args...) {} +#define mtk_mdp_err(fmt, args...) +#define mtk_mdp_dbg_enter() +#define mtk_mdp_dbg_leave() + +#endif + +#endif /* __MTK_MDP_CORE_H__ */ diff --git a/drivers/media/platform/mediatek/mdp/mtk_mdp_ipi.h b/drivers/media/platform/mediatek/mdp/mtk_mdp_ipi.h new file mode 100644 index 0000000000..b810c96695 --- /dev/null +++ b/drivers/media/platform/mediatek/mdp/mtk_mdp_ipi.h @@ -0,0 +1,120 @@ +/* SPDX-License-Identifier: GPL-2.0-only */ +/* + * Copyright (c) 2015-2016 MediaTek Inc. + * Author: Houlong Wei <houlong.wei@mediatek.com> + * Ming Hsiu Tsai <minghsiu.tsai@mediatek.com> + */ + +#ifndef __MTK_MDP_IPI_H__ +#define __MTK_MDP_IPI_H__ + +#define MTK_MDP_MAX_NUM_PLANE 3 + +enum mdp_ipi_msgid { + AP_MDP_INIT = 0xd000, + AP_MDP_DEINIT = 0xd001, + AP_MDP_PROCESS = 0xd002, + + VPU_MDP_INIT_ACK = 0xe000, + VPU_MDP_DEINIT_ACK = 0xe001, + VPU_MDP_PROCESS_ACK = 0xe002 +}; + +#pragma pack(push, 4) + +/** + * struct mdp_ipi_init - for AP_MDP_INIT + * @msg_id : AP_MDP_INIT + * @ipi_id : IPI_MDP + * @ap_inst : AP mtk_mdp_vpu address + */ +struct mdp_ipi_init { + uint32_t msg_id; + uint32_t ipi_id; + uint64_t ap_inst; +}; + +/** + * struct mdp_ipi_comm - for AP_MDP_PROCESS, AP_MDP_DEINIT + * @msg_id : AP_MDP_PROCESS, AP_MDP_DEINIT + * @ipi_id : IPI_MDP + * @ap_inst : AP mtk_mdp_vpu address + * @vpu_inst_addr : VPU MDP instance address + * @padding : Alignment padding + */ +struct mdp_ipi_comm { + uint32_t msg_id; + uint32_t ipi_id; + uint64_t ap_inst; + uint32_t vpu_inst_addr; + uint32_t padding; +}; + +/** + * struct mdp_ipi_comm_ack - for VPU_MDP_DEINIT_ACK, VPU_MDP_PROCESS_ACK + * @msg_id : VPU_MDP_DEINIT_ACK, VPU_MDP_PROCESS_ACK + * @ipi_id : IPI_MDP + * @ap_inst : AP mtk_mdp_vpu address + * @vpu_inst_addr : VPU MDP instance address + * @status : VPU exeuction result + */ +struct mdp_ipi_comm_ack { + uint32_t msg_id; + uint32_t ipi_id; + uint64_t ap_inst; + uint32_t vpu_inst_addr; + int32_t status; +}; + +/** + * struct mdp_config - configured for source/destination image + * @x : left + * @y : top + * @w : width + * @h : height + * @w_stride : bytes in horizontal + * @h_stride : bytes in vertical + * @crop_x : cropped left + * @crop_y : cropped top + * @crop_w : cropped width + * @crop_h : cropped height + * @format : color format + */ +struct mdp_config { + int32_t x; + int32_t y; + int32_t w; + int32_t h; + int32_t w_stride; + int32_t h_stride; + int32_t crop_x; + int32_t crop_y; + int32_t crop_w; + int32_t crop_h; + int32_t format; +}; + +struct mdp_buffer { + uint64_t addr_mva[MTK_MDP_MAX_NUM_PLANE]; + int32_t plane_size[MTK_MDP_MAX_NUM_PLANE]; + int32_t plane_num; +}; + +struct mdp_config_misc { + int32_t orientation; /* 0, 90, 180, 270 */ + int32_t hflip; /* 1 will enable the flip */ + int32_t vflip; /* 1 will enable the flip */ + int32_t alpha; /* global alpha */ +}; + +struct mdp_process_vsi { + struct mdp_config src_config; + struct mdp_buffer src_buffer; + struct mdp_config dst_config; + struct mdp_buffer dst_buffer; + struct mdp_config_misc misc; +}; + +#pragma pack(pop) + +#endif /* __MTK_MDP_IPI_H__ */ diff --git a/drivers/media/platform/mediatek/mdp/mtk_mdp_m2m.c b/drivers/media/platform/mediatek/mdp/mtk_mdp_m2m.c new file mode 100644 index 0000000000..f14779e759 --- /dev/null +++ b/drivers/media/platform/mediatek/mdp/mtk_mdp_m2m.c @@ -0,0 +1,1229 @@ +// SPDX-License-Identifier: GPL-2.0-only +/* + * Copyright (c) 2015-2016 MediaTek Inc. + * Author: Houlong Wei <houlong.wei@mediatek.com> + * Ming Hsiu Tsai <minghsiu.tsai@mediatek.com> + */ + +#include <linux/device.h> +#include <linux/errno.h> +#include <linux/kernel.h> +#include <linux/pm_runtime.h> +#include <linux/slab.h> +#include <linux/workqueue.h> +#include <media/v4l2-event.h> +#include <media/v4l2-ioctl.h> + +#include "mtk_mdp_core.h" +#include "mtk_mdp_m2m.h" +#include "mtk_mdp_regs.h" +#include "mtk_vpu.h" + + +/** + * struct mtk_mdp_pix_limit - image pixel size limits + * @org_w: source pixel width + * @org_h: source pixel height + * @target_rot_dis_w: pixel dst scaled width with the rotator is off + * @target_rot_dis_h: pixel dst scaled height with the rotator is off + * @target_rot_en_w: pixel dst scaled width with the rotator is on + * @target_rot_en_h: pixel dst scaled height with the rotator is on + */ +struct mtk_mdp_pix_limit { + u16 org_w; + u16 org_h; + u16 target_rot_dis_w; + u16 target_rot_dis_h; + u16 target_rot_en_w; + u16 target_rot_en_h; +}; + +static struct mtk_mdp_pix_align mtk_mdp_size_align = { + .org_w = 16, + .org_h = 16, + .target_w = 2, + .target_h = 2, +}; + +static const struct mtk_mdp_fmt mtk_mdp_formats[] = { + { + .pixelformat = V4L2_PIX_FMT_MT21C, + .depth = { 8, 4 }, + .row_depth = { 8, 8 }, + .num_planes = 2, + .num_comp = 2, + .align = &mtk_mdp_size_align, + .flags = MTK_MDP_FMT_FLAG_OUTPUT, + }, { + .pixelformat = V4L2_PIX_FMT_NV12M, + .depth = { 8, 4 }, + .row_depth = { 8, 8 }, + .num_planes = 2, + .num_comp = 2, + .flags = MTK_MDP_FMT_FLAG_OUTPUT | + MTK_MDP_FMT_FLAG_CAPTURE, + }, { + .pixelformat = V4L2_PIX_FMT_YUV420M, + .depth = { 8, 2, 2 }, + .row_depth = { 8, 4, 4 }, + .num_planes = 3, + .num_comp = 3, + .flags = MTK_MDP_FMT_FLAG_OUTPUT | + MTK_MDP_FMT_FLAG_CAPTURE, + }, { + .pixelformat = V4L2_PIX_FMT_YVU420, + .depth = { 12 }, + .row_depth = { 8 }, + .num_planes = 1, + .num_comp = 3, + .flags = MTK_MDP_FMT_FLAG_OUTPUT | + MTK_MDP_FMT_FLAG_CAPTURE, + } +}; + +static struct mtk_mdp_pix_limit mtk_mdp_size_max = { + .target_rot_dis_w = 4096, + .target_rot_dis_h = 4096, + .target_rot_en_w = 4096, + .target_rot_en_h = 4096, +}; + +static struct mtk_mdp_pix_limit mtk_mdp_size_min = { + .org_w = 16, + .org_h = 16, + .target_rot_dis_w = 16, + .target_rot_dis_h = 16, + .target_rot_en_w = 16, + .target_rot_en_h = 16, +}; + +/* align size for normal raster scan pixel format */ +static struct mtk_mdp_pix_align mtk_mdp_rs_align = { + .org_w = 2, + .org_h = 2, + .target_w = 2, + .target_h = 2, +}; + +static struct mtk_mdp_variant mtk_mdp_default_variant = { + .pix_max = &mtk_mdp_size_max, + .pix_min = &mtk_mdp_size_min, + .pix_align = &mtk_mdp_rs_align, + .h_scale_up_max = 32, + .v_scale_up_max = 32, + .h_scale_down_max = 32, + .v_scale_down_max = 128, +}; + +static const struct mtk_mdp_fmt *mtk_mdp_find_fmt(u32 pixelformat, u32 type) +{ + u32 i, flag; + + flag = V4L2_TYPE_IS_OUTPUT(type) ? MTK_MDP_FMT_FLAG_OUTPUT : + MTK_MDP_FMT_FLAG_CAPTURE; + + for (i = 0; i < ARRAY_SIZE(mtk_mdp_formats); ++i) { + if (!(mtk_mdp_formats[i].flags & flag)) + continue; + if (mtk_mdp_formats[i].pixelformat == pixelformat) + return &mtk_mdp_formats[i]; + } + return NULL; +} + +static const struct mtk_mdp_fmt *mtk_mdp_find_fmt_by_index(u32 index, u32 type) +{ + u32 i, flag, num = 0; + + flag = V4L2_TYPE_IS_OUTPUT(type) ? MTK_MDP_FMT_FLAG_OUTPUT : + MTK_MDP_FMT_FLAG_CAPTURE; + + for (i = 0; i < ARRAY_SIZE(mtk_mdp_formats); ++i) { + if (!(mtk_mdp_formats[i].flags & flag)) + continue; + if (index == num) + return &mtk_mdp_formats[i]; + num++; + } + return NULL; +} + +static void mtk_mdp_bound_align_image(u32 *w, unsigned int wmin, + unsigned int wmax, unsigned int align_w, + u32 *h, unsigned int hmin, + unsigned int hmax, unsigned int align_h) +{ + int org_w, org_h, step_w, step_h; + int walign, halign; + + org_w = *w; + org_h = *h; + walign = ffs(align_w) - 1; + halign = ffs(align_h) - 1; + v4l_bound_align_image(w, wmin, wmax, walign, h, hmin, hmax, halign, 0); + + step_w = 1 << walign; + step_h = 1 << halign; + if (*w < org_w && (*w + step_w) <= wmax) + *w += step_w; + if (*h < org_h && (*h + step_h) <= hmax) + *h += step_h; +} + +static const struct mtk_mdp_fmt *mtk_mdp_try_fmt_mplane(struct mtk_mdp_ctx *ctx, + struct v4l2_format *f) +{ + struct mtk_mdp_dev *mdp = ctx->mdp_dev; + struct mtk_mdp_variant *variant = mdp->variant; + struct v4l2_pix_format_mplane *pix_mp = &f->fmt.pix_mp; + const struct mtk_mdp_fmt *fmt; + u32 max_w, max_h, align_w, align_h; + u32 min_w, min_h, org_w, org_h; + int i; + + fmt = mtk_mdp_find_fmt(pix_mp->pixelformat, f->type); + if (!fmt) + fmt = mtk_mdp_find_fmt_by_index(0, f->type); + if (!fmt) { + dev_dbg(&ctx->mdp_dev->pdev->dev, + "pixelformat format 0x%X invalid\n", + pix_mp->pixelformat); + return NULL; + } + + pix_mp->field = V4L2_FIELD_NONE; + pix_mp->pixelformat = fmt->pixelformat; + if (V4L2_TYPE_IS_CAPTURE(f->type)) { + pix_mp->colorspace = ctx->colorspace; + pix_mp->xfer_func = ctx->xfer_func; + pix_mp->ycbcr_enc = ctx->ycbcr_enc; + pix_mp->quantization = ctx->quant; + } + + max_w = variant->pix_max->target_rot_dis_w; + max_h = variant->pix_max->target_rot_dis_h; + + if (fmt->align == NULL) { + /* use default alignment */ + align_w = variant->pix_align->org_w; + align_h = variant->pix_align->org_h; + } else { + align_w = fmt->align->org_w; + align_h = fmt->align->org_h; + } + + if (V4L2_TYPE_IS_OUTPUT(f->type)) { + min_w = variant->pix_min->org_w; + min_h = variant->pix_min->org_h; + } else { + min_w = variant->pix_min->target_rot_dis_w; + min_h = variant->pix_min->target_rot_dis_h; + } + + mtk_mdp_dbg(2, "[%d] type:%d, wxh:%ux%u, align:%ux%u, max:%ux%u", + ctx->id, f->type, pix_mp->width, pix_mp->height, + align_w, align_h, max_w, max_h); + /* + * To check if image size is modified to adjust parameter against + * hardware abilities + */ + org_w = pix_mp->width; + org_h = pix_mp->height; + + mtk_mdp_bound_align_image(&pix_mp->width, min_w, max_w, align_w, + &pix_mp->height, min_h, max_h, align_h); + + if (org_w != pix_mp->width || org_h != pix_mp->height) + mtk_mdp_dbg(1, "[%d] size change:%ux%u to %ux%u", ctx->id, + org_w, org_h, pix_mp->width, pix_mp->height); + pix_mp->num_planes = fmt->num_planes; + + for (i = 0; i < pix_mp->num_planes; ++i) { + int bpl = (pix_mp->width * fmt->row_depth[i]) / 8; + int sizeimage = (pix_mp->width * pix_mp->height * + fmt->depth[i]) / 8; + + pix_mp->plane_fmt[i].bytesperline = bpl; + if (pix_mp->plane_fmt[i].sizeimage < sizeimage) + pix_mp->plane_fmt[i].sizeimage = sizeimage; + mtk_mdp_dbg(2, "[%d] p%d, bpl:%d, sizeimage:%u (%u)", ctx->id, + i, bpl, pix_mp->plane_fmt[i].sizeimage, sizeimage); + } + + return fmt; +} + +static struct mtk_mdp_frame *mtk_mdp_ctx_get_frame(struct mtk_mdp_ctx *ctx, + enum v4l2_buf_type type) +{ + if (V4L2_TYPE_IS_OUTPUT(type)) + return &ctx->s_frame; + return &ctx->d_frame; +} + +static void mtk_mdp_check_crop_change(u32 new_w, u32 new_h, u32 *w, u32 *h) +{ + if (new_w != *w || new_h != *h) { + mtk_mdp_dbg(1, "size change:%dx%d to %dx%d", + *w, *h, new_w, new_h); + + *w = new_w; + *h = new_h; + } +} + +static int mtk_mdp_try_crop(struct mtk_mdp_ctx *ctx, u32 type, + struct v4l2_rect *r) +{ + struct mtk_mdp_frame *frame; + struct mtk_mdp_dev *mdp = ctx->mdp_dev; + struct mtk_mdp_variant *variant = mdp->variant; + u32 align_w, align_h, new_w, new_h; + u32 min_w, min_h, max_w, max_h; + + if (r->top < 0 || r->left < 0) { + dev_err(&ctx->mdp_dev->pdev->dev, + "doesn't support negative values for top & left\n"); + return -EINVAL; + } + + mtk_mdp_dbg(2, "[%d] type:%d, set wxh:%dx%d", ctx->id, type, + r->width, r->height); + + frame = mtk_mdp_ctx_get_frame(ctx, type); + max_w = frame->width; + max_h = frame->height; + new_w = r->width; + new_h = r->height; + + if (V4L2_TYPE_IS_OUTPUT(type)) { + align_w = 1; + align_h = 1; + min_w = 64; + min_h = 32; + } else { + align_w = variant->pix_align->target_w; + align_h = variant->pix_align->target_h; + if (ctx->ctrls.rotate->val == 90 || + ctx->ctrls.rotate->val == 270) { + max_w = frame->height; + max_h = frame->width; + min_w = variant->pix_min->target_rot_en_w; + min_h = variant->pix_min->target_rot_en_h; + new_w = r->height; + new_h = r->width; + } else { + min_w = variant->pix_min->target_rot_dis_w; + min_h = variant->pix_min->target_rot_dis_h; + } + } + + mtk_mdp_dbg(2, "[%d] align:%dx%d, min:%dx%d, new:%dx%d", ctx->id, + align_w, align_h, min_w, min_h, new_w, new_h); + + mtk_mdp_bound_align_image(&new_w, min_w, max_w, align_w, + &new_h, min_h, max_h, align_h); + + if (V4L2_TYPE_IS_CAPTURE(type) && + (ctx->ctrls.rotate->val == 90 || ctx->ctrls.rotate->val == 270)) + mtk_mdp_check_crop_change(new_h, new_w, + &r->width, &r->height); + else + mtk_mdp_check_crop_change(new_w, new_h, + &r->width, &r->height); + + /* adjust left/top if cropping rectangle is out of bounds */ + /* Need to add code to algin left value with 2's multiple */ + if (r->left + new_w > max_w) + r->left = max_w - new_w; + if (r->top + new_h > max_h) + r->top = max_h - new_h; + + if (r->left & 1) + r->left -= 1; + + mtk_mdp_dbg(2, "[%d] crop l,t,w,h:%d,%d,%d,%d, max:%dx%d", ctx->id, + r->left, r->top, r->width, + r->height, max_w, max_h); + return 0; +} + +static inline struct mtk_mdp_ctx *fh_to_ctx(struct v4l2_fh *fh) +{ + return container_of(fh, struct mtk_mdp_ctx, fh); +} + +static inline struct mtk_mdp_ctx *ctrl_to_ctx(struct v4l2_ctrl *ctrl) +{ + return container_of(ctrl->handler, struct mtk_mdp_ctx, ctrl_handler); +} + +void mtk_mdp_ctx_state_lock_set(struct mtk_mdp_ctx *ctx, u32 state) +{ + mutex_lock(&ctx->slock); + ctx->state |= state; + mutex_unlock(&ctx->slock); +} + +static bool mtk_mdp_ctx_state_is_set(struct mtk_mdp_ctx *ctx, u32 mask) +{ + bool ret; + + mutex_lock(&ctx->slock); + ret = (ctx->state & mask) == mask; + mutex_unlock(&ctx->slock); + return ret; +} + +static void mtk_mdp_set_frame_size(struct mtk_mdp_frame *frame, int width, + int height) +{ + frame->width = width; + frame->height = height; + frame->crop.width = width; + frame->crop.height = height; + frame->crop.left = 0; + frame->crop.top = 0; +} + +static int mtk_mdp_m2m_start_streaming(struct vb2_queue *q, unsigned int count) +{ + struct mtk_mdp_ctx *ctx = q->drv_priv; + int ret; + + ret = pm_runtime_resume_and_get(&ctx->mdp_dev->pdev->dev); + if (ret < 0) + mtk_mdp_dbg(1, "[%d] pm_runtime_resume_and_get failed:%d", + ctx->id, ret); + + return ret; +} + +static void *mtk_mdp_m2m_buf_remove(struct mtk_mdp_ctx *ctx, + enum v4l2_buf_type type) +{ + if (V4L2_TYPE_IS_OUTPUT(type)) + return v4l2_m2m_src_buf_remove(ctx->m2m_ctx); + else + return v4l2_m2m_dst_buf_remove(ctx->m2m_ctx); +} + +static void mtk_mdp_m2m_stop_streaming(struct vb2_queue *q) +{ + struct mtk_mdp_ctx *ctx = q->drv_priv; + struct vb2_buffer *vb; + + vb = mtk_mdp_m2m_buf_remove(ctx, q->type); + while (vb != NULL) { + v4l2_m2m_buf_done(to_vb2_v4l2_buffer(vb), VB2_BUF_STATE_ERROR); + vb = mtk_mdp_m2m_buf_remove(ctx, q->type); + } + + pm_runtime_put(&ctx->mdp_dev->pdev->dev); +} + +/* The color format (num_planes) must be already configured. */ +static void mtk_mdp_prepare_addr(struct mtk_mdp_ctx *ctx, + struct vb2_buffer *vb, + struct mtk_mdp_frame *frame, + struct mtk_mdp_addr *addr) +{ + u32 pix_size, planes, i; + + pix_size = frame->width * frame->height; + planes = min_t(u32, frame->fmt->num_planes, ARRAY_SIZE(addr->addr)); + for (i = 0; i < planes; i++) + addr->addr[i] = vb2_dma_contig_plane_dma_addr(vb, i); + + if (planes == 1) { + if (frame->fmt->pixelformat == V4L2_PIX_FMT_YVU420) { + addr->addr[1] = (dma_addr_t)(addr->addr[0] + pix_size); + addr->addr[2] = (dma_addr_t)(addr->addr[1] + + (pix_size >> 2)); + } else { + dev_err(&ctx->mdp_dev->pdev->dev, + "Invalid pixelformat:0x%x\n", + frame->fmt->pixelformat); + } + } + mtk_mdp_dbg(3, "[%d] planes:%d, size:%d, addr:%p,%p,%p", + ctx->id, planes, pix_size, (void *)addr->addr[0], + (void *)addr->addr[1], (void *)addr->addr[2]); +} + +static void mtk_mdp_m2m_get_bufs(struct mtk_mdp_ctx *ctx) +{ + struct mtk_mdp_frame *s_frame, *d_frame; + struct vb2_v4l2_buffer *src_vbuf, *dst_vbuf; + + s_frame = &ctx->s_frame; + d_frame = &ctx->d_frame; + + src_vbuf = v4l2_m2m_next_src_buf(ctx->m2m_ctx); + mtk_mdp_prepare_addr(ctx, &src_vbuf->vb2_buf, s_frame, &s_frame->addr); + + dst_vbuf = v4l2_m2m_next_dst_buf(ctx->m2m_ctx); + mtk_mdp_prepare_addr(ctx, &dst_vbuf->vb2_buf, d_frame, &d_frame->addr); + + dst_vbuf->vb2_buf.timestamp = src_vbuf->vb2_buf.timestamp; +} + +static void mtk_mdp_process_done(void *priv, int vb_state) +{ + struct mtk_mdp_dev *mdp = priv; + struct mtk_mdp_ctx *ctx; + struct vb2_v4l2_buffer *src_vbuf, *dst_vbuf; + + ctx = v4l2_m2m_get_curr_priv(mdp->m2m_dev); + if (!ctx) + return; + + src_vbuf = v4l2_m2m_src_buf_remove(ctx->m2m_ctx); + dst_vbuf = v4l2_m2m_dst_buf_remove(ctx->m2m_ctx); + + dst_vbuf->vb2_buf.timestamp = src_vbuf->vb2_buf.timestamp; + dst_vbuf->timecode = src_vbuf->timecode; + dst_vbuf->flags &= ~V4L2_BUF_FLAG_TSTAMP_SRC_MASK; + dst_vbuf->flags |= src_vbuf->flags & V4L2_BUF_FLAG_TSTAMP_SRC_MASK; + + v4l2_m2m_buf_done(src_vbuf, vb_state); + v4l2_m2m_buf_done(dst_vbuf, vb_state); + v4l2_m2m_job_finish(ctx->mdp_dev->m2m_dev, ctx->m2m_ctx); +} + +static void mtk_mdp_m2m_worker(struct work_struct *work) +{ + struct mtk_mdp_ctx *ctx = + container_of(work, struct mtk_mdp_ctx, work); + struct mtk_mdp_dev *mdp = ctx->mdp_dev; + enum vb2_buffer_state buf_state = VB2_BUF_STATE_ERROR; + int ret; + + if (mtk_mdp_ctx_state_is_set(ctx, MTK_MDP_CTX_ERROR)) { + dev_err(&mdp->pdev->dev, "ctx is in error state"); + goto worker_end; + } + + mtk_mdp_m2m_get_bufs(ctx); + + mtk_mdp_hw_set_input_addr(ctx, &ctx->s_frame.addr); + mtk_mdp_hw_set_output_addr(ctx, &ctx->d_frame.addr); + + mtk_mdp_hw_set_in_size(ctx); + mtk_mdp_hw_set_in_image_format(ctx); + + mtk_mdp_hw_set_out_size(ctx); + mtk_mdp_hw_set_out_image_format(ctx); + + mtk_mdp_hw_set_rotation(ctx); + mtk_mdp_hw_set_global_alpha(ctx); + + ret = mtk_mdp_vpu_process(&ctx->vpu); + if (ret) { + dev_err(&mdp->pdev->dev, "processing failed: %d", ret); + goto worker_end; + } + + buf_state = VB2_BUF_STATE_DONE; + +worker_end: + mtk_mdp_process_done(mdp, buf_state); +} + +static void mtk_mdp_m2m_device_run(void *priv) +{ + struct mtk_mdp_ctx *ctx = priv; + + queue_work(ctx->mdp_dev->job_wq, &ctx->work); +} + +static int mtk_mdp_m2m_queue_setup(struct vb2_queue *vq, + unsigned int *num_buffers, unsigned int *num_planes, + unsigned int sizes[], struct device *alloc_devs[]) +{ + struct mtk_mdp_ctx *ctx = vb2_get_drv_priv(vq); + struct mtk_mdp_frame *frame; + int i; + + frame = mtk_mdp_ctx_get_frame(ctx, vq->type); + *num_planes = frame->fmt->num_planes; + for (i = 0; i < frame->fmt->num_planes; i++) + sizes[i] = frame->payload[i]; + mtk_mdp_dbg(2, "[%d] type:%d, planes:%d, buffers:%d, size:%u,%u", + ctx->id, vq->type, *num_planes, *num_buffers, + sizes[0], sizes[1]); + return 0; +} + +static int mtk_mdp_m2m_buf_prepare(struct vb2_buffer *vb) +{ + struct mtk_mdp_ctx *ctx = vb2_get_drv_priv(vb->vb2_queue); + struct mtk_mdp_frame *frame; + int i; + + frame = mtk_mdp_ctx_get_frame(ctx, vb->vb2_queue->type); + + if (!V4L2_TYPE_IS_OUTPUT(vb->vb2_queue->type)) { + for (i = 0; i < frame->fmt->num_planes; i++) + vb2_set_plane_payload(vb, i, frame->payload[i]); + } + + return 0; +} + +static void mtk_mdp_m2m_buf_queue(struct vb2_buffer *vb) +{ + struct mtk_mdp_ctx *ctx = vb2_get_drv_priv(vb->vb2_queue); + + v4l2_m2m_buf_queue(ctx->m2m_ctx, to_vb2_v4l2_buffer(vb)); +} + +static const struct vb2_ops mtk_mdp_m2m_qops = { + .queue_setup = mtk_mdp_m2m_queue_setup, + .buf_prepare = mtk_mdp_m2m_buf_prepare, + .buf_queue = mtk_mdp_m2m_buf_queue, + .stop_streaming = mtk_mdp_m2m_stop_streaming, + .start_streaming = mtk_mdp_m2m_start_streaming, + .wait_prepare = vb2_ops_wait_prepare, + .wait_finish = vb2_ops_wait_finish, +}; + +static int mtk_mdp_m2m_querycap(struct file *file, void *fh, + struct v4l2_capability *cap) +{ + struct mtk_mdp_ctx *ctx = fh_to_ctx(fh); + struct mtk_mdp_dev *mdp = ctx->mdp_dev; + + strscpy(cap->driver, MTK_MDP_MODULE_NAME, sizeof(cap->driver)); + strscpy(cap->card, mdp->pdev->name, sizeof(cap->card)); + strscpy(cap->bus_info, "platform:mt8173", sizeof(cap->bus_info)); + + return 0; +} + +static int mtk_mdp_enum_fmt(struct v4l2_fmtdesc *f, u32 type) +{ + const struct mtk_mdp_fmt *fmt; + + fmt = mtk_mdp_find_fmt_by_index(f->index, type); + if (!fmt) + return -EINVAL; + + f->pixelformat = fmt->pixelformat; + + return 0; +} + +static int mtk_mdp_m2m_enum_fmt_vid_cap(struct file *file, void *priv, + struct v4l2_fmtdesc *f) +{ + return mtk_mdp_enum_fmt(f, V4L2_BUF_TYPE_VIDEO_CAPTURE_MPLANE); +} + +static int mtk_mdp_m2m_enum_fmt_vid_out(struct file *file, void *priv, + struct v4l2_fmtdesc *f) +{ + return mtk_mdp_enum_fmt(f, V4L2_BUF_TYPE_VIDEO_OUTPUT_MPLANE); +} + +static int mtk_mdp_m2m_g_fmt_mplane(struct file *file, void *fh, + struct v4l2_format *f) +{ + struct mtk_mdp_ctx *ctx = fh_to_ctx(fh); + struct mtk_mdp_frame *frame; + struct v4l2_pix_format_mplane *pix_mp; + int i; + + mtk_mdp_dbg(2, "[%d] type:%d", ctx->id, f->type); + + frame = mtk_mdp_ctx_get_frame(ctx, f->type); + pix_mp = &f->fmt.pix_mp; + + pix_mp->width = frame->width; + pix_mp->height = frame->height; + pix_mp->field = V4L2_FIELD_NONE; + pix_mp->pixelformat = frame->fmt->pixelformat; + pix_mp->num_planes = frame->fmt->num_planes; + pix_mp->colorspace = ctx->colorspace; + pix_mp->xfer_func = ctx->xfer_func; + pix_mp->ycbcr_enc = ctx->ycbcr_enc; + pix_mp->quantization = ctx->quant; + mtk_mdp_dbg(2, "[%d] wxh:%dx%d", ctx->id, + pix_mp->width, pix_mp->height); + + for (i = 0; i < pix_mp->num_planes; ++i) { + pix_mp->plane_fmt[i].bytesperline = (frame->width * + frame->fmt->row_depth[i]) / 8; + pix_mp->plane_fmt[i].sizeimage = (frame->width * + frame->height * frame->fmt->depth[i]) / 8; + + mtk_mdp_dbg(2, "[%d] p%d, bpl:%d, sizeimage:%d", ctx->id, i, + pix_mp->plane_fmt[i].bytesperline, + pix_mp->plane_fmt[i].sizeimage); + } + + return 0; +} + +static int mtk_mdp_m2m_try_fmt_mplane(struct file *file, void *fh, + struct v4l2_format *f) +{ + struct mtk_mdp_ctx *ctx = fh_to_ctx(fh); + + if (!mtk_mdp_try_fmt_mplane(ctx, f)) + return -EINVAL; + return 0; +} + +static int mtk_mdp_m2m_s_fmt_mplane(struct file *file, void *fh, + struct v4l2_format *f) +{ + struct mtk_mdp_ctx *ctx = fh_to_ctx(fh); + struct vb2_queue *vq; + struct mtk_mdp_frame *frame; + struct v4l2_pix_format_mplane *pix_mp; + const struct mtk_mdp_fmt *fmt; + int i; + + mtk_mdp_dbg(2, "[%d] type:%d", ctx->id, f->type); + + frame = mtk_mdp_ctx_get_frame(ctx, f->type); + fmt = mtk_mdp_try_fmt_mplane(ctx, f); + if (!fmt) { + mtk_mdp_err("[%d] try_fmt failed, type:%d", ctx->id, f->type); + return -EINVAL; + } + frame->fmt = fmt; + + vq = v4l2_m2m_get_vq(ctx->m2m_ctx, f->type); + if (vb2_is_streaming(vq)) { + dev_info(&ctx->mdp_dev->pdev->dev, "queue %d busy", f->type); + return -EBUSY; + } + + pix_mp = &f->fmt.pix_mp; + for (i = 0; i < frame->fmt->num_planes; i++) { + frame->payload[i] = pix_mp->plane_fmt[i].sizeimage; + frame->pitch[i] = pix_mp->plane_fmt[i].bytesperline; + } + + mtk_mdp_set_frame_size(frame, pix_mp->width, pix_mp->height); + if (V4L2_TYPE_IS_OUTPUT(f->type)) { + ctx->colorspace = pix_mp->colorspace; + ctx->xfer_func = pix_mp->xfer_func; + ctx->ycbcr_enc = pix_mp->ycbcr_enc; + ctx->quant = pix_mp->quantization; + } + + mtk_mdp_dbg(2, "[%d] type:%d, frame:%dx%d", ctx->id, f->type, + frame->width, frame->height); + + return 0; +} + +static int mtk_mdp_m2m_reqbufs(struct file *file, void *fh, + struct v4l2_requestbuffers *reqbufs) +{ + struct mtk_mdp_ctx *ctx = fh_to_ctx(fh); + + return v4l2_m2m_reqbufs(file, ctx->m2m_ctx, reqbufs); +} + +static int mtk_mdp_m2m_streamon(struct file *file, void *fh, + enum v4l2_buf_type type) +{ + struct mtk_mdp_ctx *ctx = fh_to_ctx(fh); + int ret; + + if (!mtk_mdp_ctx_state_is_set(ctx, MTK_MDP_VPU_INIT)) { + ret = mtk_mdp_vpu_init(&ctx->vpu); + if (ret < 0) { + dev_err(&ctx->mdp_dev->pdev->dev, + "vpu init failed %d\n", + ret); + return -EINVAL; + } + mtk_mdp_ctx_state_lock_set(ctx, MTK_MDP_VPU_INIT); + } + + return v4l2_m2m_streamon(file, ctx->m2m_ctx, type); +} + +static inline bool mtk_mdp_is_target_compose(u32 target) +{ + if (target == V4L2_SEL_TGT_COMPOSE_DEFAULT + || target == V4L2_SEL_TGT_COMPOSE_BOUNDS + || target == V4L2_SEL_TGT_COMPOSE) + return true; + return false; +} + +static inline bool mtk_mdp_is_target_crop(u32 target) +{ + if (target == V4L2_SEL_TGT_CROP_DEFAULT + || target == V4L2_SEL_TGT_CROP_BOUNDS + || target == V4L2_SEL_TGT_CROP) + return true; + return false; +} + +static int mtk_mdp_m2m_g_selection(struct file *file, void *fh, + struct v4l2_selection *s) +{ + struct mtk_mdp_frame *frame; + struct mtk_mdp_ctx *ctx = fh_to_ctx(fh); + bool valid = false; + + if (s->type == V4L2_BUF_TYPE_VIDEO_CAPTURE) { + if (mtk_mdp_is_target_compose(s->target)) + valid = true; + } else if (s->type == V4L2_BUF_TYPE_VIDEO_OUTPUT) { + if (mtk_mdp_is_target_crop(s->target)) + valid = true; + } + if (!valid) { + mtk_mdp_dbg(1, "[%d] invalid type:%d,%u", ctx->id, s->type, + s->target); + return -EINVAL; + } + + frame = mtk_mdp_ctx_get_frame(ctx, s->type); + + switch (s->target) { + case V4L2_SEL_TGT_COMPOSE_DEFAULT: + case V4L2_SEL_TGT_COMPOSE_BOUNDS: + case V4L2_SEL_TGT_CROP_BOUNDS: + case V4L2_SEL_TGT_CROP_DEFAULT: + s->r.left = 0; + s->r.top = 0; + s->r.width = frame->width; + s->r.height = frame->height; + return 0; + + case V4L2_SEL_TGT_COMPOSE: + case V4L2_SEL_TGT_CROP: + s->r.left = frame->crop.left; + s->r.top = frame->crop.top; + s->r.width = frame->crop.width; + s->r.height = frame->crop.height; + return 0; + } + + return -EINVAL; +} + +static int mtk_mdp_check_scaler_ratio(struct mtk_mdp_variant *var, int src_w, + int src_h, int dst_w, int dst_h, int rot) +{ + int tmp_w, tmp_h; + + if (rot == 90 || rot == 270) { + tmp_w = dst_h; + tmp_h = dst_w; + } else { + tmp_w = dst_w; + tmp_h = dst_h; + } + + if ((src_w / tmp_w) > var->h_scale_down_max || + (src_h / tmp_h) > var->v_scale_down_max || + (tmp_w / src_w) > var->h_scale_up_max || + (tmp_h / src_h) > var->v_scale_up_max) + return -EINVAL; + + return 0; +} + +static int mtk_mdp_m2m_s_selection(struct file *file, void *fh, + struct v4l2_selection *s) +{ + struct mtk_mdp_frame *frame; + struct mtk_mdp_ctx *ctx = fh_to_ctx(fh); + struct v4l2_rect new_r; + struct mtk_mdp_variant *variant = ctx->mdp_dev->variant; + int ret; + bool valid = false; + + if (s->type == V4L2_BUF_TYPE_VIDEO_CAPTURE) { + if (s->target == V4L2_SEL_TGT_COMPOSE) + valid = true; + } else if (s->type == V4L2_BUF_TYPE_VIDEO_OUTPUT) { + if (s->target == V4L2_SEL_TGT_CROP) + valid = true; + } + if (!valid) { + mtk_mdp_dbg(1, "[%d] invalid type:%d,%u", ctx->id, s->type, + s->target); + return -EINVAL; + } + + new_r = s->r; + ret = mtk_mdp_try_crop(ctx, s->type, &new_r); + if (ret) + return ret; + + if (mtk_mdp_is_target_crop(s->target)) + frame = &ctx->s_frame; + else + frame = &ctx->d_frame; + + /* Check to see if scaling ratio is within supported range */ + if (V4L2_TYPE_IS_OUTPUT(s->type)) + ret = mtk_mdp_check_scaler_ratio(variant, new_r.width, + new_r.height, ctx->d_frame.crop.width, + ctx->d_frame.crop.height, + ctx->ctrls.rotate->val); + else + ret = mtk_mdp_check_scaler_ratio(variant, + ctx->s_frame.crop.width, + ctx->s_frame.crop.height, new_r.width, + new_r.height, ctx->ctrls.rotate->val); + + if (ret) { + dev_info(&ctx->mdp_dev->pdev->dev, + "Out of scaler range"); + return -EINVAL; + } + + s->r = new_r; + frame->crop = new_r; + + return 0; +} + +static const struct v4l2_ioctl_ops mtk_mdp_m2m_ioctl_ops = { + .vidioc_querycap = mtk_mdp_m2m_querycap, + .vidioc_enum_fmt_vid_cap = mtk_mdp_m2m_enum_fmt_vid_cap, + .vidioc_enum_fmt_vid_out = mtk_mdp_m2m_enum_fmt_vid_out, + .vidioc_g_fmt_vid_cap_mplane = mtk_mdp_m2m_g_fmt_mplane, + .vidioc_g_fmt_vid_out_mplane = mtk_mdp_m2m_g_fmt_mplane, + .vidioc_try_fmt_vid_cap_mplane = mtk_mdp_m2m_try_fmt_mplane, + .vidioc_try_fmt_vid_out_mplane = mtk_mdp_m2m_try_fmt_mplane, + .vidioc_s_fmt_vid_cap_mplane = mtk_mdp_m2m_s_fmt_mplane, + .vidioc_s_fmt_vid_out_mplane = mtk_mdp_m2m_s_fmt_mplane, + .vidioc_reqbufs = mtk_mdp_m2m_reqbufs, + .vidioc_create_bufs = v4l2_m2m_ioctl_create_bufs, + .vidioc_expbuf = v4l2_m2m_ioctl_expbuf, + .vidioc_subscribe_event = v4l2_ctrl_subscribe_event, + .vidioc_unsubscribe_event = v4l2_event_unsubscribe, + .vidioc_querybuf = v4l2_m2m_ioctl_querybuf, + .vidioc_qbuf = v4l2_m2m_ioctl_qbuf, + .vidioc_dqbuf = v4l2_m2m_ioctl_dqbuf, + .vidioc_streamon = mtk_mdp_m2m_streamon, + .vidioc_streamoff = v4l2_m2m_ioctl_streamoff, + .vidioc_g_selection = mtk_mdp_m2m_g_selection, + .vidioc_s_selection = mtk_mdp_m2m_s_selection +}; + +static int mtk_mdp_m2m_queue_init(void *priv, struct vb2_queue *src_vq, + struct vb2_queue *dst_vq) +{ + struct mtk_mdp_ctx *ctx = priv; + int ret; + + memset(src_vq, 0, sizeof(*src_vq)); + src_vq->type = V4L2_BUF_TYPE_VIDEO_OUTPUT_MPLANE; + src_vq->io_modes = VB2_MMAP | VB2_DMABUF; + src_vq->drv_priv = ctx; + src_vq->ops = &mtk_mdp_m2m_qops; + src_vq->mem_ops = &vb2_dma_contig_memops; + src_vq->buf_struct_size = sizeof(struct v4l2_m2m_buffer); + src_vq->timestamp_flags = V4L2_BUF_FLAG_TIMESTAMP_COPY; + src_vq->dev = &ctx->mdp_dev->pdev->dev; + src_vq->lock = &ctx->mdp_dev->lock; + + ret = vb2_queue_init(src_vq); + if (ret) + return ret; + + memset(dst_vq, 0, sizeof(*dst_vq)); + dst_vq->type = V4L2_BUF_TYPE_VIDEO_CAPTURE_MPLANE; + dst_vq->io_modes = VB2_MMAP | VB2_DMABUF; + dst_vq->drv_priv = ctx; + dst_vq->ops = &mtk_mdp_m2m_qops; + dst_vq->mem_ops = &vb2_dma_contig_memops; + dst_vq->buf_struct_size = sizeof(struct v4l2_m2m_buffer); + dst_vq->timestamp_flags = V4L2_BUF_FLAG_TIMESTAMP_COPY; + dst_vq->dev = &ctx->mdp_dev->pdev->dev; + dst_vq->lock = &ctx->mdp_dev->lock; + + return vb2_queue_init(dst_vq); +} + +static int mtk_mdp_s_ctrl(struct v4l2_ctrl *ctrl) +{ + struct mtk_mdp_ctx *ctx = ctrl_to_ctx(ctrl); + struct mtk_mdp_dev *mdp = ctx->mdp_dev; + struct mtk_mdp_variant *variant = mdp->variant; + int ret = 0; + + if (ctrl->flags & V4L2_CTRL_FLAG_INACTIVE) + return 0; + + switch (ctrl->id) { + case V4L2_CID_HFLIP: + ctx->hflip = ctrl->val; + break; + case V4L2_CID_VFLIP: + ctx->vflip = ctrl->val; + break; + case V4L2_CID_ROTATE: + ret = mtk_mdp_check_scaler_ratio(variant, + ctx->s_frame.crop.width, + ctx->s_frame.crop.height, + ctx->d_frame.crop.width, + ctx->d_frame.crop.height, + ctx->ctrls.rotate->val); + + if (ret) + return -EINVAL; + + ctx->rotation = ctrl->val; + break; + case V4L2_CID_ALPHA_COMPONENT: + ctx->d_frame.alpha = ctrl->val; + break; + } + + return 0; +} + +static const struct v4l2_ctrl_ops mtk_mdp_ctrl_ops = { + .s_ctrl = mtk_mdp_s_ctrl, +}; + +static int mtk_mdp_ctrls_create(struct mtk_mdp_ctx *ctx) +{ + v4l2_ctrl_handler_init(&ctx->ctrl_handler, MTK_MDP_MAX_CTRL_NUM); + + ctx->ctrls.rotate = v4l2_ctrl_new_std(&ctx->ctrl_handler, + &mtk_mdp_ctrl_ops, V4L2_CID_ROTATE, 0, 270, 90, 0); + ctx->ctrls.hflip = v4l2_ctrl_new_std(&ctx->ctrl_handler, + &mtk_mdp_ctrl_ops, + V4L2_CID_HFLIP, + 0, 1, 1, 0); + ctx->ctrls.vflip = v4l2_ctrl_new_std(&ctx->ctrl_handler, + &mtk_mdp_ctrl_ops, + V4L2_CID_VFLIP, + 0, 1, 1, 0); + ctx->ctrls.global_alpha = v4l2_ctrl_new_std(&ctx->ctrl_handler, + &mtk_mdp_ctrl_ops, + V4L2_CID_ALPHA_COMPONENT, + 0, 255, 1, 0); + ctx->ctrls_rdy = ctx->ctrl_handler.error == 0; + + if (ctx->ctrl_handler.error) { + int err = ctx->ctrl_handler.error; + + v4l2_ctrl_handler_free(&ctx->ctrl_handler); + dev_err(&ctx->mdp_dev->pdev->dev, + "Failed to create control handlers\n"); + return err; + } + + return 0; +} + +static void mtk_mdp_set_default_params(struct mtk_mdp_ctx *ctx) +{ + struct mtk_mdp_dev *mdp = ctx->mdp_dev; + struct mtk_mdp_frame *frame; + + frame = mtk_mdp_ctx_get_frame(ctx, V4L2_BUF_TYPE_VIDEO_OUTPUT_MPLANE); + frame->fmt = mtk_mdp_find_fmt_by_index(0, + V4L2_BUF_TYPE_VIDEO_OUTPUT_MPLANE); + frame->width = mdp->variant->pix_min->org_w; + frame->height = mdp->variant->pix_min->org_h; + frame->payload[0] = frame->width * frame->height; + frame->payload[1] = frame->payload[0] / 2; + + frame = mtk_mdp_ctx_get_frame(ctx, V4L2_BUF_TYPE_VIDEO_CAPTURE_MPLANE); + frame->fmt = mtk_mdp_find_fmt_by_index(0, + V4L2_BUF_TYPE_VIDEO_CAPTURE_MPLANE); + frame->width = mdp->variant->pix_min->target_rot_dis_w; + frame->height = mdp->variant->pix_min->target_rot_dis_h; + frame->payload[0] = frame->width * frame->height; + frame->payload[1] = frame->payload[0] / 2; + +} + +static int mtk_mdp_m2m_open(struct file *file) +{ + struct mtk_mdp_dev *mdp = video_drvdata(file); + struct video_device *vfd = video_devdata(file); + struct mtk_mdp_ctx *ctx = NULL; + int ret; + struct v4l2_format default_format; + + ctx = kzalloc(sizeof(*ctx), GFP_KERNEL); + if (!ctx) + return -ENOMEM; + + if (mutex_lock_interruptible(&mdp->lock)) { + ret = -ERESTARTSYS; + goto err_lock; + } + + mutex_init(&ctx->slock); + ctx->id = mdp->id_counter++; + v4l2_fh_init(&ctx->fh, vfd); + file->private_data = &ctx->fh; + ret = mtk_mdp_ctrls_create(ctx); + if (ret) + goto error_ctrls; + + /* Use separate control handler per file handle */ + ctx->fh.ctrl_handler = &ctx->ctrl_handler; + v4l2_fh_add(&ctx->fh); + INIT_LIST_HEAD(&ctx->list); + + ctx->mdp_dev = mdp; + mtk_mdp_set_default_params(ctx); + + INIT_WORK(&ctx->work, mtk_mdp_m2m_worker); + ctx->m2m_ctx = v4l2_m2m_ctx_init(mdp->m2m_dev, ctx, + mtk_mdp_m2m_queue_init); + if (IS_ERR(ctx->m2m_ctx)) { + dev_err(&mdp->pdev->dev, "Failed to initialize m2m context"); + ret = PTR_ERR(ctx->m2m_ctx); + goto error_m2m_ctx; + } + ctx->fh.m2m_ctx = ctx->m2m_ctx; + if (mdp->ctx_num++ == 0) { + ret = vpu_load_firmware(mdp->vpu_dev); + if (ret < 0) { + dev_err(&mdp->pdev->dev, + "vpu_load_firmware failed %d\n", ret); + goto err_load_vpu; + } + + ret = mtk_mdp_vpu_register(mdp->pdev); + if (ret < 0) { + dev_err(&mdp->pdev->dev, + "mdp_vpu register failed %d\n", ret); + goto err_load_vpu; + } + } + + list_add(&ctx->list, &mdp->ctx_list); + mutex_unlock(&mdp->lock); + + /* Default format */ + memset(&default_format, 0, sizeof(default_format)); + default_format.type = V4L2_BUF_TYPE_VIDEO_OUTPUT_MPLANE; + default_format.fmt.pix_mp.width = 32; + default_format.fmt.pix_mp.height = 32; + default_format.fmt.pix_mp.pixelformat = V4L2_PIX_FMT_YUV420M; + mtk_mdp_m2m_s_fmt_mplane(file, &ctx->fh, &default_format); + default_format.type = V4L2_BUF_TYPE_VIDEO_CAPTURE_MPLANE; + mtk_mdp_m2m_s_fmt_mplane(file, &ctx->fh, &default_format); + + mtk_mdp_dbg(0, "%s [%d]", dev_name(&mdp->pdev->dev), ctx->id); + + return 0; + +err_load_vpu: + mdp->ctx_num--; + v4l2_m2m_ctx_release(ctx->m2m_ctx); +error_m2m_ctx: + v4l2_ctrl_handler_free(&ctx->ctrl_handler); +error_ctrls: + v4l2_fh_del(&ctx->fh); + v4l2_fh_exit(&ctx->fh); + mutex_unlock(&mdp->lock); +err_lock: + kfree(ctx); + + return ret; +} + +static int mtk_mdp_m2m_release(struct file *file) +{ + struct mtk_mdp_ctx *ctx = fh_to_ctx(file->private_data); + struct mtk_mdp_dev *mdp = ctx->mdp_dev; + + flush_workqueue(mdp->job_wq); + mutex_lock(&mdp->lock); + v4l2_m2m_ctx_release(ctx->m2m_ctx); + v4l2_ctrl_handler_free(&ctx->ctrl_handler); + v4l2_fh_del(&ctx->fh); + v4l2_fh_exit(&ctx->fh); + mtk_mdp_vpu_deinit(&ctx->vpu); + mdp->ctx_num--; + list_del_init(&ctx->list); + + mtk_mdp_dbg(0, "%s [%d]", dev_name(&mdp->pdev->dev), ctx->id); + + mutex_unlock(&mdp->lock); + kfree(ctx); + + return 0; +} + +static const struct v4l2_file_operations mtk_mdp_m2m_fops = { + .owner = THIS_MODULE, + .open = mtk_mdp_m2m_open, + .release = mtk_mdp_m2m_release, + .poll = v4l2_m2m_fop_poll, + .unlocked_ioctl = video_ioctl2, + .mmap = v4l2_m2m_fop_mmap, +}; + +static const struct v4l2_m2m_ops mtk_mdp_m2m_ops = { + .device_run = mtk_mdp_m2m_device_run, +}; + +int mtk_mdp_register_m2m_device(struct mtk_mdp_dev *mdp) +{ + struct device *dev = &mdp->pdev->dev; + int ret; + + mdp->variant = &mtk_mdp_default_variant; + mdp->vdev = video_device_alloc(); + if (!mdp->vdev) { + dev_err(dev, "failed to allocate video device\n"); + ret = -ENOMEM; + goto err_video_alloc; + } + mdp->vdev->device_caps = V4L2_CAP_VIDEO_M2M_MPLANE | V4L2_CAP_STREAMING; + mdp->vdev->fops = &mtk_mdp_m2m_fops; + mdp->vdev->ioctl_ops = &mtk_mdp_m2m_ioctl_ops; + mdp->vdev->release = video_device_release; + mdp->vdev->lock = &mdp->lock; + mdp->vdev->vfl_dir = VFL_DIR_M2M; + mdp->vdev->v4l2_dev = &mdp->v4l2_dev; + snprintf(mdp->vdev->name, sizeof(mdp->vdev->name), "%s:m2m", + MTK_MDP_MODULE_NAME); + video_set_drvdata(mdp->vdev, mdp); + + mdp->m2m_dev = v4l2_m2m_init(&mtk_mdp_m2m_ops); + if (IS_ERR(mdp->m2m_dev)) { + dev_err(dev, "failed to initialize v4l2-m2m device\n"); + ret = PTR_ERR(mdp->m2m_dev); + goto err_m2m_init; + } + + ret = video_register_device(mdp->vdev, VFL_TYPE_VIDEO, 2); + if (ret) { + dev_err(dev, "failed to register video device\n"); + goto err_vdev_register; + } + + v4l2_info(&mdp->v4l2_dev, "driver registered as /dev/video%d", + mdp->vdev->num); + return 0; + +err_vdev_register: + v4l2_m2m_release(mdp->m2m_dev); +err_m2m_init: + video_device_release(mdp->vdev); +err_video_alloc: + + return ret; +} + +void mtk_mdp_unregister_m2m_device(struct mtk_mdp_dev *mdp) +{ + video_unregister_device(mdp->vdev); + v4l2_m2m_release(mdp->m2m_dev); +} diff --git a/drivers/media/platform/mediatek/mdp/mtk_mdp_m2m.h b/drivers/media/platform/mediatek/mdp/mtk_mdp_m2m.h new file mode 100644 index 0000000000..485dbdbbf5 --- /dev/null +++ b/drivers/media/platform/mediatek/mdp/mtk_mdp_m2m.h @@ -0,0 +1,14 @@ +/* SPDX-License-Identifier: GPL-2.0-only */ +/* + * Copyright (c) 2016 MediaTek Inc. + * Author: Ming Hsiu Tsai <minghsiu.tsai@mediatek.com> + */ + +#ifndef __MTK_MDP_M2M_H__ +#define __MTK_MDP_M2M_H__ + +void mtk_mdp_ctx_state_lock_set(struct mtk_mdp_ctx *ctx, u32 state); +int mtk_mdp_register_m2m_device(struct mtk_mdp_dev *mdp); +void mtk_mdp_unregister_m2m_device(struct mtk_mdp_dev *mdp); + +#endif /* __MTK_MDP_M2M_H__ */ diff --git a/drivers/media/platform/mediatek/mdp/mtk_mdp_regs.c b/drivers/media/platform/mediatek/mdp/mtk_mdp_regs.c new file mode 100644 index 0000000000..ba476d50ae --- /dev/null +++ b/drivers/media/platform/mediatek/mdp/mtk_mdp_regs.c @@ -0,0 +1,148 @@ +// SPDX-License-Identifier: GPL-2.0-only +/* + * Copyright (c) 2015-2016 MediaTek Inc. + * Author: Houlong Wei <houlong.wei@mediatek.com> + * Ming Hsiu Tsai <minghsiu.tsai@mediatek.com> + */ + +#include <linux/platform_device.h> + +#include "mtk_mdp_core.h" +#include "mtk_mdp_regs.h" + + +#define MDP_COLORFMT_PACK(VIDEO, PLANE, COPLANE, HF, VF, BITS, GROUP, SWAP, ID)\ + (((VIDEO) << 27) | ((PLANE) << 24) | ((COPLANE) << 22) |\ + ((HF) << 20) | ((VF) << 18) | ((BITS) << 8) | ((GROUP) << 6) |\ + ((SWAP) << 5) | ((ID) << 0)) + +enum MDP_COLOR_ENUM { + MDP_COLOR_UNKNOWN = 0, + MDP_COLOR_NV12 = MDP_COLORFMT_PACK(0, 2, 1, 1, 1, 8, 1, 0, 12), + MDP_COLOR_I420 = MDP_COLORFMT_PACK(0, 3, 0, 1, 1, 8, 1, 0, 8), + MDP_COLOR_YV12 = MDP_COLORFMT_PACK(0, 3, 0, 1, 1, 8, 1, 1, 8), + /* Mediatek proprietary format */ + MDP_COLOR_420_MT21 = MDP_COLORFMT_PACK(5, 2, 1, 1, 1, 256, 1, 0, 12), +}; + +static int32_t mtk_mdp_map_color_format(int v4l2_format) +{ + switch (v4l2_format) { + case V4L2_PIX_FMT_NV12M: + case V4L2_PIX_FMT_NV12: + return MDP_COLOR_NV12; + case V4L2_PIX_FMT_MT21C: + return MDP_COLOR_420_MT21; + case V4L2_PIX_FMT_YUV420M: + case V4L2_PIX_FMT_YUV420: + return MDP_COLOR_I420; + case V4L2_PIX_FMT_YVU420: + return MDP_COLOR_YV12; + } + + mtk_mdp_err("Unknown format 0x%x", v4l2_format); + + return MDP_COLOR_UNKNOWN; +} + +void mtk_mdp_hw_set_input_addr(struct mtk_mdp_ctx *ctx, + struct mtk_mdp_addr *addr) +{ + struct mdp_buffer *src_buf = &ctx->vpu.vsi->src_buffer; + int i; + + for (i = 0; i < ARRAY_SIZE(addr->addr); i++) + src_buf->addr_mva[i] = (uint64_t)addr->addr[i]; +} + +void mtk_mdp_hw_set_output_addr(struct mtk_mdp_ctx *ctx, + struct mtk_mdp_addr *addr) +{ + struct mdp_buffer *dst_buf = &ctx->vpu.vsi->dst_buffer; + int i; + + for (i = 0; i < ARRAY_SIZE(addr->addr); i++) + dst_buf->addr_mva[i] = (uint64_t)addr->addr[i]; +} + +void mtk_mdp_hw_set_in_size(struct mtk_mdp_ctx *ctx) +{ + struct mtk_mdp_frame *frame = &ctx->s_frame; + struct mdp_config *config = &ctx->vpu.vsi->src_config; + + /* Set input pixel offset */ + config->crop_x = frame->crop.left; + config->crop_y = frame->crop.top; + + /* Set input cropped size */ + config->crop_w = frame->crop.width; + config->crop_h = frame->crop.height; + + /* Set input original size */ + config->x = 0; + config->y = 0; + config->w = frame->width; + config->h = frame->height; +} + +void mtk_mdp_hw_set_in_image_format(struct mtk_mdp_ctx *ctx) +{ + unsigned int i; + struct mtk_mdp_frame *frame = &ctx->s_frame; + struct mdp_config *config = &ctx->vpu.vsi->src_config; + struct mdp_buffer *src_buf = &ctx->vpu.vsi->src_buffer; + + src_buf->plane_num = frame->fmt->num_comp; + config->format = mtk_mdp_map_color_format(frame->fmt->pixelformat); + config->w_stride = 0; /* MDP will calculate it by color format. */ + config->h_stride = 0; /* MDP will calculate it by color format. */ + + for (i = 0; i < src_buf->plane_num; i++) + src_buf->plane_size[i] = frame->payload[i]; +} + +void mtk_mdp_hw_set_out_size(struct mtk_mdp_ctx *ctx) +{ + struct mtk_mdp_frame *frame = &ctx->d_frame; + struct mdp_config *config = &ctx->vpu.vsi->dst_config; + + config->crop_x = frame->crop.left; + config->crop_y = frame->crop.top; + config->crop_w = frame->crop.width; + config->crop_h = frame->crop.height; + config->x = 0; + config->y = 0; + config->w = frame->width; + config->h = frame->height; +} + +void mtk_mdp_hw_set_out_image_format(struct mtk_mdp_ctx *ctx) +{ + unsigned int i; + struct mtk_mdp_frame *frame = &ctx->d_frame; + struct mdp_config *config = &ctx->vpu.vsi->dst_config; + struct mdp_buffer *dst_buf = &ctx->vpu.vsi->dst_buffer; + + dst_buf->plane_num = frame->fmt->num_comp; + config->format = mtk_mdp_map_color_format(frame->fmt->pixelformat); + config->w_stride = 0; /* MDP will calculate it by color format. */ + config->h_stride = 0; /* MDP will calculate it by color format. */ + for (i = 0; i < dst_buf->plane_num; i++) + dst_buf->plane_size[i] = frame->payload[i]; +} + +void mtk_mdp_hw_set_rotation(struct mtk_mdp_ctx *ctx) +{ + struct mdp_config_misc *misc = &ctx->vpu.vsi->misc; + + misc->orientation = ctx->ctrls.rotate->val; + misc->hflip = ctx->ctrls.hflip->val; + misc->vflip = ctx->ctrls.vflip->val; +} + +void mtk_mdp_hw_set_global_alpha(struct mtk_mdp_ctx *ctx) +{ + struct mdp_config_misc *misc = &ctx->vpu.vsi->misc; + + misc->alpha = ctx->ctrls.global_alpha->val; +} diff --git a/drivers/media/platform/mediatek/mdp/mtk_mdp_regs.h b/drivers/media/platform/mediatek/mdp/mtk_mdp_regs.h new file mode 100644 index 0000000000..32cf202f23 --- /dev/null +++ b/drivers/media/platform/mediatek/mdp/mtk_mdp_regs.h @@ -0,0 +1,23 @@ +/* SPDX-License-Identifier: GPL-2.0-only */ +/* + * Copyright (c) 2016 MediaTek Inc. + * Author: Ming Hsiu Tsai <minghsiu.tsai@mediatek.com> + */ + +#ifndef __MTK_MDP_REGS_H__ +#define __MTK_MDP_REGS_H__ + + +void mtk_mdp_hw_set_input_addr(struct mtk_mdp_ctx *ctx, + struct mtk_mdp_addr *addr); +void mtk_mdp_hw_set_output_addr(struct mtk_mdp_ctx *ctx, + struct mtk_mdp_addr *addr); +void mtk_mdp_hw_set_in_size(struct mtk_mdp_ctx *ctx); +void mtk_mdp_hw_set_in_image_format(struct mtk_mdp_ctx *ctx); +void mtk_mdp_hw_set_out_size(struct mtk_mdp_ctx *ctx); +void mtk_mdp_hw_set_out_image_format(struct mtk_mdp_ctx *ctx); +void mtk_mdp_hw_set_rotation(struct mtk_mdp_ctx *ctx); +void mtk_mdp_hw_set_global_alpha(struct mtk_mdp_ctx *ctx); + + +#endif /* __MTK_MDP_REGS_H__ */ diff --git a/drivers/media/platform/mediatek/mdp/mtk_mdp_vpu.c b/drivers/media/platform/mediatek/mdp/mtk_mdp_vpu.c new file mode 100644 index 0000000000..b065ccd069 --- /dev/null +++ b/drivers/media/platform/mediatek/mdp/mtk_mdp_vpu.c @@ -0,0 +1,138 @@ +// SPDX-License-Identifier: GPL-2.0-only +/* + * Copyright (c) 2015-2016 MediaTek Inc. + * Author: Houlong Wei <houlong.wei@mediatek.com> + * Ming Hsiu Tsai <minghsiu.tsai@mediatek.com> + */ + +#include "mtk_mdp_core.h" +#include "mtk_mdp_vpu.h" +#include "mtk_vpu.h" + + +static inline struct mtk_mdp_ctx *vpu_to_ctx(struct mtk_mdp_vpu *vpu) +{ + return container_of(vpu, struct mtk_mdp_ctx, vpu); +} + +static void mtk_mdp_vpu_handle_init_ack(const struct mdp_ipi_comm_ack *msg) +{ + struct mtk_mdp_vpu *vpu = (struct mtk_mdp_vpu *) + (unsigned long)msg->ap_inst; + + /* mapping VPU address to kernel virtual address */ + vpu->vsi = (struct mdp_process_vsi *) + vpu_mapping_dm_addr(vpu->pdev, msg->vpu_inst_addr); + vpu->inst_addr = msg->vpu_inst_addr; +} + +static void mtk_mdp_vpu_ipi_handler(const void *data, unsigned int len, + void *priv) +{ + const struct mdp_ipi_comm_ack *msg = data; + unsigned int msg_id = msg->msg_id; + struct mtk_mdp_vpu *vpu = (struct mtk_mdp_vpu *) + (unsigned long)msg->ap_inst; + struct mtk_mdp_ctx *ctx; + + vpu->failure = msg->status; + if (!vpu->failure) { + switch (msg_id) { + case VPU_MDP_INIT_ACK: + mtk_mdp_vpu_handle_init_ack(data); + break; + case VPU_MDP_DEINIT_ACK: + case VPU_MDP_PROCESS_ACK: + break; + default: + ctx = vpu_to_ctx(vpu); + dev_err(&ctx->mdp_dev->pdev->dev, + "handle unknown ipi msg:0x%x\n", + msg_id); + break; + } + } else { + ctx = vpu_to_ctx(vpu); + mtk_mdp_dbg(0, "[%d]:msg 0x%x, failure:%d", ctx->id, + msg_id, vpu->failure); + } +} + +int mtk_mdp_vpu_register(struct platform_device *pdev) +{ + struct mtk_mdp_dev *mdp = platform_get_drvdata(pdev); + int err; + + err = vpu_ipi_register(mdp->vpu_dev, IPI_MDP, + mtk_mdp_vpu_ipi_handler, "mdp_vpu", NULL); + if (err) + dev_err(&mdp->pdev->dev, + "vpu_ipi_registration fail status=%d\n", err); + + return err; +} + +static int mtk_mdp_vpu_send_msg(void *msg, int len, struct mtk_mdp_vpu *vpu, + int id) +{ + struct mtk_mdp_ctx *ctx = vpu_to_ctx(vpu); + int err; + + if (!vpu->pdev) { + mtk_mdp_dbg(1, "[%d]:vpu pdev is NULL", ctx->id); + return -EINVAL; + } + + mutex_lock(&ctx->mdp_dev->vpulock); + err = vpu_ipi_send(vpu->pdev, (enum ipi_id)id, msg, len); + if (err) + dev_err(&ctx->mdp_dev->pdev->dev, + "vpu_ipi_send fail status %d\n", err); + mutex_unlock(&ctx->mdp_dev->vpulock); + + return err; +} + +static int mtk_mdp_vpu_send_ap_ipi(struct mtk_mdp_vpu *vpu, uint32_t msg_id) +{ + int err; + struct mdp_ipi_comm msg; + + msg.msg_id = msg_id; + msg.ipi_id = IPI_MDP; + msg.vpu_inst_addr = vpu->inst_addr; + msg.ap_inst = (unsigned long)vpu; + err = mtk_mdp_vpu_send_msg((void *)&msg, sizeof(msg), vpu, IPI_MDP); + if (!err && vpu->failure) + err = -EINVAL; + + return err; +} + +int mtk_mdp_vpu_init(struct mtk_mdp_vpu *vpu) +{ + int err; + struct mdp_ipi_init msg; + struct mtk_mdp_ctx *ctx = vpu_to_ctx(vpu); + + vpu->pdev = ctx->mdp_dev->vpu_dev; + + msg.msg_id = AP_MDP_INIT; + msg.ipi_id = IPI_MDP; + msg.ap_inst = (unsigned long)vpu; + err = mtk_mdp_vpu_send_msg((void *)&msg, sizeof(msg), vpu, IPI_MDP); + if (!err && vpu->failure) + err = -EINVAL; + + return err; +} + +int mtk_mdp_vpu_deinit(struct mtk_mdp_vpu *vpu) +{ + return mtk_mdp_vpu_send_ap_ipi(vpu, AP_MDP_DEINIT); +} + +int mtk_mdp_vpu_process(struct mtk_mdp_vpu *vpu) +{ + return mtk_mdp_vpu_send_ap_ipi(vpu, AP_MDP_PROCESS); +} diff --git a/drivers/media/platform/mediatek/mdp/mtk_mdp_vpu.h b/drivers/media/platform/mediatek/mdp/mtk_mdp_vpu.h new file mode 100644 index 0000000000..5a10205084 --- /dev/null +++ b/drivers/media/platform/mediatek/mdp/mtk_mdp_vpu.h @@ -0,0 +1,33 @@ +/* SPDX-License-Identifier: GPL-2.0-only */ +/* + * Copyright (c) 2015-2016 MediaTek Inc. + * Author: Houlong Wei <houlong.wei@mediatek.com> + * Ming Hsiu Tsai <minghsiu.tsai@mediatek.com> + */ + +#ifndef __MTK_MDP_VPU_H__ +#define __MTK_MDP_VPU_H__ + +#include "mtk_mdp_ipi.h" + + +/** + * struct mtk_mdp_vpu - VPU instance for MDP + * @pdev : pointer to the VPU platform device + * @inst_addr : VPU MDP instance address + * @failure : VPU execution result status + * @vsi : VPU shared information + */ +struct mtk_mdp_vpu { + struct platform_device *pdev; + uint32_t inst_addr; + int32_t failure; + struct mdp_process_vsi *vsi; +}; + +int mtk_mdp_vpu_register(struct platform_device *pdev); +int mtk_mdp_vpu_init(struct mtk_mdp_vpu *vpu); +int mtk_mdp_vpu_deinit(struct mtk_mdp_vpu *vpu); +int mtk_mdp_vpu_process(struct mtk_mdp_vpu *vpu); + +#endif /* __MTK_MDP_VPU_H__ */ diff --git a/drivers/media/platform/mediatek/mdp3/Kconfig b/drivers/media/platform/mediatek/mdp3/Kconfig new file mode 100644 index 0000000000..602329c447 --- /dev/null +++ b/drivers/media/platform/mediatek/mdp3/Kconfig @@ -0,0 +1,19 @@ +# SPDX-License-Identifier: GPL-2.0-only +config VIDEO_MEDIATEK_MDP3 + tristate "MediaTek MDP v3 driver" + depends on MTK_IOMMU || COMPILE_TEST + depends on VIDEO_DEV + depends on HAS_DMA + depends on REMOTEPROC + depends on MTK_MMSYS + depends on MTK_CMDQ + depends on MTK_SCP + select VIDEOBUF2_DMA_CONTIG + select V4L2_MEM2MEM_DEV + default n + help + It is a v4l2 driver and present in MediaTek MT8183 SoC. + The driver supports scaling and color space conversion. + + To compile this driver as a module, choose M here: the + module will be called mtk-mdp3. diff --git a/drivers/media/platform/mediatek/mdp3/Makefile b/drivers/media/platform/mediatek/mdp3/Makefile new file mode 100644 index 0000000000..2ee24195a2 --- /dev/null +++ b/drivers/media/platform/mediatek/mdp3/Makefile @@ -0,0 +1,6 @@ +# SPDX-License-Identifier: GPL-2.0-only +mtk-mdp3-y += mdp_cfg_data.o mtk-mdp3-core.o mtk-mdp3-vpu.o mtk-mdp3-regs.o +mtk-mdp3-y += mtk-mdp3-m2m.o +mtk-mdp3-y += mtk-mdp3-comp.o mtk-mdp3-cmdq.o + +obj-$(CONFIG_VIDEO_MEDIATEK_MDP3) += mtk-mdp3.o diff --git a/drivers/media/platform/mediatek/mdp3/mdp_cfg_data.c b/drivers/media/platform/mediatek/mdp3/mdp_cfg_data.c new file mode 100644 index 0000000000..502eeae0bf --- /dev/null +++ b/drivers/media/platform/mediatek/mdp3/mdp_cfg_data.c @@ -0,0 +1,453 @@ +// SPDX-License-Identifier: GPL-2.0-only +/* + * Copyright (c) 2023 MediaTek Inc. + * Author: Ping-Hsun Wu <ping-hsun.wu@mediatek.com> + */ + +#include "mtk-img-ipi.h" +#include "mtk-mdp3-cfg.h" +#include "mtk-mdp3-core.h" +#include "mtk-mdp3-comp.h" +#include "mtk-mdp3-regs.h" + +enum mt8183_mdp_comp_id { + /* ISP */ + MT8183_MDP_COMP_WPEI = 0, + MT8183_MDP_COMP_WPEO, /* 1 */ + MT8183_MDP_COMP_WPEI2, /* 2 */ + MT8183_MDP_COMP_WPEO2, /* 3 */ + MT8183_MDP_COMP_ISP_IMGI, /* 4 */ + MT8183_MDP_COMP_ISP_IMGO, /* 5 */ + MT8183_MDP_COMP_ISP_IMG2O, /* 6 */ + + /* IPU */ + MT8183_MDP_COMP_IPUI, /* 7 */ + MT8183_MDP_COMP_IPUO, /* 8 */ + + /* MDP */ + MT8183_MDP_COMP_CAMIN, /* 9 */ + MT8183_MDP_COMP_CAMIN2, /* 10 */ + MT8183_MDP_COMP_RDMA0, /* 11 */ + MT8183_MDP_COMP_AAL0, /* 12 */ + MT8183_MDP_COMP_CCORR0, /* 13 */ + MT8183_MDP_COMP_RSZ0, /* 14 */ + MT8183_MDP_COMP_RSZ1, /* 15 */ + MT8183_MDP_COMP_TDSHP0, /* 16 */ + MT8183_MDP_COMP_COLOR0, /* 17 */ + MT8183_MDP_COMP_PATH0_SOUT, /* 18 */ + MT8183_MDP_COMP_PATH1_SOUT, /* 19 */ + MT8183_MDP_COMP_WROT0, /* 20 */ + MT8183_MDP_COMP_WDMA, /* 21 */ + + /* Dummy Engine */ + MT8183_MDP_COMP_RDMA1, /* 22 */ + MT8183_MDP_COMP_RSZ2, /* 23 */ + MT8183_MDP_COMP_TDSHP1, /* 24 */ + MT8183_MDP_COMP_WROT1, /* 25 */ +}; + +static const struct of_device_id mt8183_mdp_probe_infra[MDP_INFRA_MAX] = { + [MDP_INFRA_MMSYS] = { .compatible = "mediatek,mt8183-mmsys" }, + [MDP_INFRA_MUTEX] = { .compatible = "mediatek,mt8183-disp-mutex" }, + [MDP_INFRA_SCP] = { .compatible = "mediatek,mt8183-scp" } +}; + +static const struct mdp_platform_config mt8183_plat_cfg = { + .rdma_support_10bit = true, + .rdma_rsz1_sram_sharing = true, + .rdma_upsample_repeat_only = true, + .rsz_disable_dcm_small_sample = false, + .wrot_filter_constraint = false, +}; + +static const u32 mt8183_mutex_idx[MDP_MAX_COMP_COUNT] = { + [MDP_COMP_RDMA0] = MUTEX_MOD_IDX_MDP_RDMA0, + [MDP_COMP_RSZ0] = MUTEX_MOD_IDX_MDP_RSZ0, + [MDP_COMP_RSZ1] = MUTEX_MOD_IDX_MDP_RSZ1, + [MDP_COMP_TDSHP0] = MUTEX_MOD_IDX_MDP_TDSHP0, + [MDP_COMP_WROT0] = MUTEX_MOD_IDX_MDP_WROT0, + [MDP_COMP_WDMA] = MUTEX_MOD_IDX_MDP_WDMA, + [MDP_COMP_AAL0] = MUTEX_MOD_IDX_MDP_AAL0, + [MDP_COMP_CCORR0] = MUTEX_MOD_IDX_MDP_CCORR0, +}; + +static const struct mdp_comp_data mt8183_mdp_comp_data[MDP_MAX_COMP_COUNT] = { + [MDP_COMP_WPEI] = { + {MDP_COMP_TYPE_WPEI, 0, MT8183_MDP_COMP_WPEI}, + {0, 0, 0} + }, + [MDP_COMP_WPEO] = { + {MDP_COMP_TYPE_EXTO, 2, MT8183_MDP_COMP_WPEO}, + {0, 0, 0} + }, + [MDP_COMP_WPEI2] = { + {MDP_COMP_TYPE_WPEI, 1, MT8183_MDP_COMP_WPEI2}, + {0, 0, 0} + }, + [MDP_COMP_WPEO2] = { + {MDP_COMP_TYPE_EXTO, 3, MT8183_MDP_COMP_WPEO2}, + {0, 0, 0} + }, + [MDP_COMP_ISP_IMGI] = { + {MDP_COMP_TYPE_IMGI, 0, MT8183_MDP_COMP_ISP_IMGI}, + {0, 0, 4} + }, + [MDP_COMP_ISP_IMGO] = { + {MDP_COMP_TYPE_EXTO, 0, MT8183_MDP_COMP_ISP_IMGO}, + {0, 0, 4} + }, + [MDP_COMP_ISP_IMG2O] = { + {MDP_COMP_TYPE_EXTO, 1, MT8183_MDP_COMP_ISP_IMG2O}, + {0, 0, 0} + }, + [MDP_COMP_CAMIN] = { + {MDP_COMP_TYPE_DL_PATH, 0, MT8183_MDP_COMP_CAMIN}, + {2, 2, 1} + }, + [MDP_COMP_CAMIN2] = { + {MDP_COMP_TYPE_DL_PATH, 1, MT8183_MDP_COMP_CAMIN2}, + {2, 4, 1} + }, + [MDP_COMP_RDMA0] = { + {MDP_COMP_TYPE_RDMA, 0, MT8183_MDP_COMP_RDMA0}, + {2, 0, 0} + }, + [MDP_COMP_CCORR0] = { + {MDP_COMP_TYPE_CCORR, 0, MT8183_MDP_COMP_CCORR0}, + {1, 0, 0} + }, + [MDP_COMP_RSZ0] = { + {MDP_COMP_TYPE_RSZ, 0, MT8183_MDP_COMP_RSZ0}, + {1, 0, 0} + }, + [MDP_COMP_RSZ1] = { + {MDP_COMP_TYPE_RSZ, 1, MT8183_MDP_COMP_RSZ1}, + {1, 0, 0} + }, + [MDP_COMP_TDSHP0] = { + {MDP_COMP_TYPE_TDSHP, 0, MT8183_MDP_COMP_TDSHP0}, + {0, 0, 0} + }, + [MDP_COMP_PATH0_SOUT] = { + {MDP_COMP_TYPE_PATH, 0, MT8183_MDP_COMP_PATH0_SOUT}, + {0, 0, 0} + }, + [MDP_COMP_PATH1_SOUT] = { + {MDP_COMP_TYPE_PATH, 1, MT8183_MDP_COMP_PATH1_SOUT}, + {0, 0, 0} + }, + [MDP_COMP_WROT0] = { + {MDP_COMP_TYPE_WROT, 0, MT8183_MDP_COMP_WROT0}, + {1, 0, 0} + }, + [MDP_COMP_WDMA] = { + {MDP_COMP_TYPE_WDMA, 0, MT8183_MDP_COMP_WDMA}, + {1, 0, 0} + }, +}; + +static const struct of_device_id mt8183_sub_comp_dt_ids[] = { + { + .compatible = "mediatek,mt8183-mdp3-wdma", + .data = (void *)MDP_COMP_TYPE_PATH, + }, { + .compatible = "mediatek,mt8183-mdp3-wrot", + .data = (void *)MDP_COMP_TYPE_PATH, + }, + {} +}; + +/* + * All 10-bit related formats are not added in the basic format list, + * please add the corresponding format settings before use. + */ +static const struct mdp_format mt8183_formats[] = { + { + .pixelformat = V4L2_PIX_FMT_GREY, + .mdp_color = MDP_COLOR_GREY, + .depth = { 8 }, + .row_depth = { 8 }, + .num_planes = 1, + .flags = MDP_FMT_FLAG_OUTPUT | MDP_FMT_FLAG_CAPTURE, + }, { + .pixelformat = V4L2_PIX_FMT_RGB565X, + .mdp_color = MDP_COLOR_BGR565, + .depth = { 16 }, + .row_depth = { 16 }, + .num_planes = 1, + .flags = MDP_FMT_FLAG_OUTPUT | MDP_FMT_FLAG_CAPTURE, + }, { + .pixelformat = V4L2_PIX_FMT_RGB565, + .mdp_color = MDP_COLOR_RGB565, + .depth = { 16 }, + .row_depth = { 16 }, + .num_planes = 1, + .flags = MDP_FMT_FLAG_OUTPUT | MDP_FMT_FLAG_CAPTURE, + }, { + .pixelformat = V4L2_PIX_FMT_RGB24, + .mdp_color = MDP_COLOR_RGB888, + .depth = { 24 }, + .row_depth = { 24 }, + .num_planes = 1, + .flags = MDP_FMT_FLAG_OUTPUT | MDP_FMT_FLAG_CAPTURE, + }, { + .pixelformat = V4L2_PIX_FMT_BGR24, + .mdp_color = MDP_COLOR_BGR888, + .depth = { 24 }, + .row_depth = { 24 }, + .num_planes = 1, + .flags = MDP_FMT_FLAG_OUTPUT | MDP_FMT_FLAG_CAPTURE, + }, { + .pixelformat = V4L2_PIX_FMT_ABGR32, + .mdp_color = MDP_COLOR_BGRA8888, + .depth = { 32 }, + .row_depth = { 32 }, + .num_planes = 1, + .flags = MDP_FMT_FLAG_OUTPUT | MDP_FMT_FLAG_CAPTURE, + }, { + .pixelformat = V4L2_PIX_FMT_ARGB32, + .mdp_color = MDP_COLOR_ARGB8888, + .depth = { 32 }, + .row_depth = { 32 }, + .num_planes = 1, + .flags = MDP_FMT_FLAG_OUTPUT | MDP_FMT_FLAG_CAPTURE, + }, { + .pixelformat = V4L2_PIX_FMT_UYVY, + .mdp_color = MDP_COLOR_UYVY, + .depth = { 16 }, + .row_depth = { 16 }, + .num_planes = 1, + .walign = 1, + .flags = MDP_FMT_FLAG_OUTPUT | MDP_FMT_FLAG_CAPTURE, + }, { + .pixelformat = V4L2_PIX_FMT_VYUY, + .mdp_color = MDP_COLOR_VYUY, + .depth = { 16 }, + .row_depth = { 16 }, + .num_planes = 1, + .walign = 1, + .flags = MDP_FMT_FLAG_OUTPUT | MDP_FMT_FLAG_CAPTURE, + }, { + .pixelformat = V4L2_PIX_FMT_YUYV, + .mdp_color = MDP_COLOR_YUYV, + .depth = { 16 }, + .row_depth = { 16 }, + .num_planes = 1, + .walign = 1, + .flags = MDP_FMT_FLAG_OUTPUT | MDP_FMT_FLAG_CAPTURE, + }, { + .pixelformat = V4L2_PIX_FMT_YVYU, + .mdp_color = MDP_COLOR_YVYU, + .depth = { 16 }, + .row_depth = { 16 }, + .num_planes = 1, + .walign = 1, + .flags = MDP_FMT_FLAG_OUTPUT | MDP_FMT_FLAG_CAPTURE, + }, { + .pixelformat = V4L2_PIX_FMT_YUV420, + .mdp_color = MDP_COLOR_I420, + .depth = { 12 }, + .row_depth = { 8 }, + .num_planes = 1, + .walign = 1, + .halign = 1, + .flags = MDP_FMT_FLAG_OUTPUT | MDP_FMT_FLAG_CAPTURE, + }, { + .pixelformat = V4L2_PIX_FMT_YVU420, + .mdp_color = MDP_COLOR_YV12, + .depth = { 12 }, + .row_depth = { 8 }, + .num_planes = 1, + .walign = 1, + .halign = 1, + .flags = MDP_FMT_FLAG_OUTPUT | MDP_FMT_FLAG_CAPTURE, + }, { + .pixelformat = V4L2_PIX_FMT_NV12, + .mdp_color = MDP_COLOR_NV12, + .depth = { 12 }, + .row_depth = { 8 }, + .num_planes = 1, + .walign = 1, + .halign = 1, + .flags = MDP_FMT_FLAG_OUTPUT | MDP_FMT_FLAG_CAPTURE, + }, { + .pixelformat = V4L2_PIX_FMT_NV21, + .mdp_color = MDP_COLOR_NV21, + .depth = { 12 }, + .row_depth = { 8 }, + .num_planes = 1, + .walign = 1, + .halign = 1, + .flags = MDP_FMT_FLAG_OUTPUT | MDP_FMT_FLAG_CAPTURE, + }, { + .pixelformat = V4L2_PIX_FMT_NV16, + .mdp_color = MDP_COLOR_NV16, + .depth = { 16 }, + .row_depth = { 8 }, + .num_planes = 1, + .walign = 1, + .flags = MDP_FMT_FLAG_OUTPUT, + }, { + .pixelformat = V4L2_PIX_FMT_NV61, + .mdp_color = MDP_COLOR_NV61, + .depth = { 16 }, + .row_depth = { 8 }, + .num_planes = 1, + .walign = 1, + .flags = MDP_FMT_FLAG_OUTPUT, + }, { + .pixelformat = V4L2_PIX_FMT_NV24, + .mdp_color = MDP_COLOR_NV24, + .depth = { 24 }, + .row_depth = { 8 }, + .num_planes = 1, + .flags = MDP_FMT_FLAG_OUTPUT, + }, { + .pixelformat = V4L2_PIX_FMT_NV42, + .mdp_color = MDP_COLOR_NV42, + .depth = { 24 }, + .row_depth = { 8 }, + .num_planes = 1, + .flags = MDP_FMT_FLAG_OUTPUT, + }, { + .pixelformat = V4L2_PIX_FMT_MT21C, + .mdp_color = MDP_COLOR_420_BLK_UFO, + .depth = { 8, 4 }, + .row_depth = { 8, 8 }, + .num_planes = 2, + .walign = 4, + .halign = 5, + .flags = MDP_FMT_FLAG_OUTPUT, + }, { + .pixelformat = V4L2_PIX_FMT_MM21, + .mdp_color = MDP_COLOR_420_BLK, + .depth = { 8, 4 }, + .row_depth = { 8, 8 }, + .num_planes = 2, + .walign = 4, + .halign = 5, + .flags = MDP_FMT_FLAG_OUTPUT, + }, { + .pixelformat = V4L2_PIX_FMT_NV12M, + .mdp_color = MDP_COLOR_NV12, + .depth = { 8, 4 }, + .row_depth = { 8, 8 }, + .num_planes = 2, + .walign = 1, + .halign = 1, + .flags = MDP_FMT_FLAG_OUTPUT | MDP_FMT_FLAG_CAPTURE, + }, { + .pixelformat = V4L2_PIX_FMT_NV21M, + .mdp_color = MDP_COLOR_NV21, + .depth = { 8, 4 }, + .row_depth = { 8, 8 }, + .num_planes = 2, + .walign = 1, + .halign = 1, + .flags = MDP_FMT_FLAG_OUTPUT | MDP_FMT_FLAG_CAPTURE, + }, { + .pixelformat = V4L2_PIX_FMT_NV16M, + .mdp_color = MDP_COLOR_NV16, + .depth = { 8, 8 }, + .row_depth = { 8, 8 }, + .num_planes = 2, + .walign = 1, + .flags = MDP_FMT_FLAG_OUTPUT, + }, { + .pixelformat = V4L2_PIX_FMT_NV61M, + .mdp_color = MDP_COLOR_NV61, + .depth = { 8, 8 }, + .row_depth = { 8, 8 }, + .num_planes = 2, + .walign = 1, + .flags = MDP_FMT_FLAG_OUTPUT, + }, { + .pixelformat = V4L2_PIX_FMT_YUV420M, + .mdp_color = MDP_COLOR_I420, + .depth = { 8, 2, 2 }, + .row_depth = { 8, 4, 4 }, + .num_planes = 3, + .walign = 1, + .halign = 1, + .flags = MDP_FMT_FLAG_OUTPUT | MDP_FMT_FLAG_CAPTURE, + }, { + .pixelformat = V4L2_PIX_FMT_YVU420M, + .mdp_color = MDP_COLOR_YV12, + .depth = { 8, 2, 2 }, + .row_depth = { 8, 4, 4 }, + .num_planes = 3, + .walign = 1, + .halign = 1, + .flags = MDP_FMT_FLAG_OUTPUT | MDP_FMT_FLAG_CAPTURE, + } +}; + +static const struct mdp_limit mt8183_mdp_def_limit = { + .out_limit = { + .wmin = 16, + .hmin = 16, + .wmax = 8176, + .hmax = 8176, + }, + .cap_limit = { + .wmin = 2, + .hmin = 2, + .wmax = 8176, + .hmax = 8176, + }, + .h_scale_up_max = 32, + .v_scale_up_max = 32, + .h_scale_down_max = 20, + .v_scale_down_max = 128, +}; + +static const struct mdp_pipe_info mt8183_pipe_info[] = { + [MDP_PIPE_WPEI] = {MDP_PIPE_WPEI, 0}, + [MDP_PIPE_WPEI2] = {MDP_PIPE_WPEI2, 1}, + [MDP_PIPE_IMGI] = {MDP_PIPE_IMGI, 2}, + [MDP_PIPE_RDMA0] = {MDP_PIPE_RDMA0, 3} +}; + +const struct mtk_mdp_driver_data mt8183_mdp_driver_data = { + .mdp_plat_id = MT8183, + .mdp_probe_infra = mt8183_mdp_probe_infra, + .mdp_cfg = &mt8183_plat_cfg, + .mdp_mutex_table_idx = mt8183_mutex_idx, + .comp_data = mt8183_mdp_comp_data, + .comp_data_len = ARRAY_SIZE(mt8183_mdp_comp_data), + .mdp_sub_comp_dt_ids = mt8183_sub_comp_dt_ids, + .format = mt8183_formats, + .format_len = ARRAY_SIZE(mt8183_formats), + .def_limit = &mt8183_mdp_def_limit, + .pipe_info = mt8183_pipe_info, + .pipe_info_len = ARRAY_SIZE(mt8183_pipe_info), +}; + +s32 mdp_cfg_get_id_inner(struct mdp_dev *mdp_dev, enum mtk_mdp_comp_id id) +{ + if (!mdp_dev) + return MDP_COMP_NONE; + if (id <= MDP_COMP_NONE || id >= MDP_MAX_COMP_COUNT) + return MDP_COMP_NONE; + + return mdp_dev->mdp_data->comp_data[id].match.inner_id; +} + +enum mtk_mdp_comp_id mdp_cfg_get_id_public(struct mdp_dev *mdp_dev, s32 inner_id) +{ + enum mtk_mdp_comp_id public_id = MDP_COMP_NONE; + u32 i; + + if (IS_ERR(mdp_dev) || !inner_id) + goto err_public_id; + + for (i = 0; i < MDP_MAX_COMP_COUNT; i++) { + if (mdp_dev->mdp_data->comp_data[i].match.inner_id == inner_id) { + public_id = i; + return public_id; + } + } + +err_public_id: + return public_id; +} diff --git a/drivers/media/platform/mediatek/mdp3/mdp_reg_ccorr.h b/drivers/media/platform/mediatek/mdp3/mdp_reg_ccorr.h new file mode 100644 index 0000000000..3b2c6531c1 --- /dev/null +++ b/drivers/media/platform/mediatek/mdp3/mdp_reg_ccorr.h @@ -0,0 +1,19 @@ +/* SPDX-License-Identifier: GPL-2.0-only */ +/* + * Copyright (c) 2022 MediaTek Inc. + * Author: Ping-Hsun Wu <ping-hsun.wu@mediatek.com> + */ + +#ifndef __MDP_REG_CCORR_H__ +#define __MDP_REG_CCORR_H__ + +#define MDP_CCORR_EN 0x000 +#define MDP_CCORR_CFG 0x020 +#define MDP_CCORR_SIZE 0x030 + +/* MASK */ +#define MDP_CCORR_EN_MASK 0x00000001 +#define MDP_CCORR_CFG_MASK 0x70001317 +#define MDP_CCORR_SIZE_MASK 0x1fff1fff + +#endif // __MDP_REG_CCORR_H__ diff --git a/drivers/media/platform/mediatek/mdp3/mdp_reg_rdma.h b/drivers/media/platform/mediatek/mdp3/mdp_reg_rdma.h new file mode 100644 index 0000000000..be4065e252 --- /dev/null +++ b/drivers/media/platform/mediatek/mdp3/mdp_reg_rdma.h @@ -0,0 +1,65 @@ +/* SPDX-License-Identifier: GPL-2.0-only */ +/* + * Copyright (c) 2022 MediaTek Inc. + * Author: Ping-Hsun Wu <ping-hsun.wu@mediatek.com> + */ + +#ifndef __MDP_REG_RDMA_H__ +#define __MDP_REG_RDMA_H__ + +#define MDP_RDMA_EN 0x000 +#define MDP_RDMA_RESET 0x008 +#define MDP_RDMA_CON 0x020 +#define MDP_RDMA_GMCIF_CON 0x028 +#define MDP_RDMA_SRC_CON 0x030 +#define MDP_RDMA_MF_BKGD_SIZE_IN_BYTE 0x060 +#define MDP_RDMA_MF_BKGD_SIZE_IN_PXL 0x068 +#define MDP_RDMA_MF_SRC_SIZE 0x070 +#define MDP_RDMA_MF_CLIP_SIZE 0x078 +#define MDP_RDMA_MF_OFFSET_1 0x080 +#define MDP_RDMA_SF_BKGD_SIZE_IN_BYTE 0x090 +#define MDP_RDMA_SRC_END_0 0x100 +#define MDP_RDMA_SRC_END_1 0x108 +#define MDP_RDMA_SRC_END_2 0x110 +#define MDP_RDMA_SRC_OFFSET_0 0x118 +#define MDP_RDMA_SRC_OFFSET_1 0x120 +#define MDP_RDMA_SRC_OFFSET_2 0x128 +#define MDP_RDMA_SRC_OFFSET_0_P 0x148 +#define MDP_RDMA_TRANSFORM_0 0x200 +#define MDP_RDMA_RESV_DUMMY_0 0x2a0 +#define MDP_RDMA_MON_STA_1 0x408 +#define MDP_RDMA_SRC_BASE_0 0xf00 +#define MDP_RDMA_SRC_BASE_1 0xf08 +#define MDP_RDMA_SRC_BASE_2 0xf10 +#define MDP_RDMA_UFO_DEC_LENGTH_BASE_Y 0xf20 +#define MDP_RDMA_UFO_DEC_LENGTH_BASE_C 0xf28 + +/* MASK */ +#define MDP_RDMA_EN_MASK 0x00000001 +#define MDP_RDMA_RESET_MASK 0x00000001 +#define MDP_RDMA_CON_MASK 0x00001110 +#define MDP_RDMA_GMCIF_CON_MASK 0xfffb3771 +#define MDP_RDMA_SRC_CON_MASK 0xf3ffffff +#define MDP_RDMA_MF_BKGD_SIZE_IN_BYTE_MASK 0x001fffff +#define MDP_RDMA_MF_BKGD_SIZE_IN_PXL_MASK 0x001fffff +#define MDP_RDMA_MF_SRC_SIZE_MASK 0x1fff1fff +#define MDP_RDMA_MF_CLIP_SIZE_MASK 0x1fff1fff +#define MDP_RDMA_MF_OFFSET_1_MASK 0x003f001f +#define MDP_RDMA_SF_BKGD_SIZE_IN_BYTE_MASK 0x001fffff +#define MDP_RDMA_SRC_END_0_MASK 0xffffffff +#define MDP_RDMA_SRC_END_1_MASK 0xffffffff +#define MDP_RDMA_SRC_END_2_MASK 0xffffffff +#define MDP_RDMA_SRC_OFFSET_0_MASK 0xffffffff +#define MDP_RDMA_SRC_OFFSET_1_MASK 0xffffffff +#define MDP_RDMA_SRC_OFFSET_2_MASK 0xffffffff +#define MDP_RDMA_SRC_OFFSET_0_P_MASK 0xffffffff +#define MDP_RDMA_TRANSFORM_0_MASK 0xff110777 +#define MDP_RDMA_RESV_DUMMY_0_MASK 0xffffffff +#define MDP_RDMA_MON_STA_1_MASK 0xffffffff +#define MDP_RDMA_SRC_BASE_0_MASK 0xffffffff +#define MDP_RDMA_SRC_BASE_1_MASK 0xffffffff +#define MDP_RDMA_SRC_BASE_2_MASK 0xffffffff +#define MDP_RDMA_UFO_DEC_LENGTH_BASE_Y_MASK 0xffffffff +#define MDP_RDMA_UFO_DEC_LENGTH_BASE_C_MASK 0xffffffff + +#endif // __MDP_REG_RDMA_H__ diff --git a/drivers/media/platform/mediatek/mdp3/mdp_reg_rsz.h b/drivers/media/platform/mediatek/mdp3/mdp_reg_rsz.h new file mode 100644 index 0000000000..484f6d6064 --- /dev/null +++ b/drivers/media/platform/mediatek/mdp3/mdp_reg_rsz.h @@ -0,0 +1,39 @@ +/* SPDX-License-Identifier: GPL-2.0-only */ +/* + * Copyright (c) 2022 MediaTek Inc. + * Author: Ping-Hsun Wu <ping-hsun.wu@mediatek.com> + */ + +#ifndef __MDP_REG_RSZ_H__ +#define __MDP_REG_RSZ_H__ + +#define PRZ_ENABLE 0x000 +#define PRZ_CONTROL_1 0x004 +#define PRZ_CONTROL_2 0x008 +#define PRZ_INPUT_IMAGE 0x010 +#define PRZ_OUTPUT_IMAGE 0x014 +#define PRZ_HORIZONTAL_COEFF_STEP 0x018 +#define PRZ_VERTICAL_COEFF_STEP 0x01c +#define PRZ_LUMA_HORIZONTAL_INTEGER_OFFSET 0x020 +#define PRZ_LUMA_HORIZONTAL_SUBPIXEL_OFFSET 0x024 +#define PRZ_LUMA_VERTICAL_INTEGER_OFFSET 0x028 +#define PRZ_LUMA_VERTICAL_SUBPIXEL_OFFSET 0x02c +#define PRZ_CHROMA_HORIZONTAL_INTEGER_OFFSET 0x030 +#define PRZ_CHROMA_HORIZONTAL_SUBPIXEL_OFFSET 0x034 + +/* MASK */ +#define PRZ_ENABLE_MASK 0x00010001 +#define PRZ_CONTROL_1_MASK 0xfffffff3 +#define PRZ_CONTROL_2_MASK 0x0ffffaff +#define PRZ_INPUT_IMAGE_MASK 0xffffffff +#define PRZ_OUTPUT_IMAGE_MASK 0xffffffff +#define PRZ_HORIZONTAL_COEFF_STEP_MASK 0x007fffff +#define PRZ_VERTICAL_COEFF_STEP_MASK 0x007fffff +#define PRZ_LUMA_HORIZONTAL_INTEGER_OFFSET_MASK 0x0000ffff +#define PRZ_LUMA_HORIZONTAL_SUBPIXEL_OFFSET_MASK 0x001fffff +#define PRZ_LUMA_VERTICAL_INTEGER_OFFSET_MASK 0x0000ffff +#define PRZ_LUMA_VERTICAL_SUBPIXEL_OFFSET_MASK 0x001fffff +#define PRZ_CHROMA_HORIZONTAL_INTEGER_OFFSET_MASK 0x0000ffff +#define PRZ_CHROMA_HORIZONTAL_SUBPIXEL_OFFSET_MASK 0x001fffff + +#endif // __MDP_REG_RSZ_H__ diff --git a/drivers/media/platform/mediatek/mdp3/mdp_reg_wdma.h b/drivers/media/platform/mediatek/mdp3/mdp_reg_wdma.h new file mode 100644 index 0000000000..0280e91c09 --- /dev/null +++ b/drivers/media/platform/mediatek/mdp3/mdp_reg_wdma.h @@ -0,0 +1,47 @@ +/* SPDX-License-Identifier: GPL-2.0-only */ +/* + * Copyright (c) 2022 MediaTek Inc. + * Author: Ping-Hsun Wu <ping-hsun.wu@mediatek.com> + */ + +#ifndef __MDP_REG_WDMA_H__ +#define __MDP_REG_WDMA_H__ + +#define WDMA_EN 0x008 +#define WDMA_RST 0x00c +#define WDMA_CFG 0x014 +#define WDMA_SRC_SIZE 0x018 +#define WDMA_CLIP_SIZE 0x01c +#define WDMA_CLIP_COORD 0x020 +#define WDMA_DST_W_IN_BYTE 0x028 +#define WDMA_ALPHA 0x02c +#define WDMA_BUF_CON2 0x03c +#define WDMA_DST_UV_PITCH 0x078 +#define WDMA_DST_ADDR_OFFSET 0x080 +#define WDMA_DST_U_ADDR_OFFSET 0x084 +#define WDMA_DST_V_ADDR_OFFSET 0x088 +#define WDMA_FLOW_CTRL_DBG 0x0a0 +#define WDMA_DST_ADDR 0xf00 +#define WDMA_DST_U_ADDR 0xf04 +#define WDMA_DST_V_ADDR 0xf08 + +/* MASK */ +#define WDMA_EN_MASK 0x00000001 +#define WDMA_RST_MASK 0x00000001 +#define WDMA_CFG_MASK 0xff03bff0 +#define WDMA_SRC_SIZE_MASK 0x3fff3fff +#define WDMA_CLIP_SIZE_MASK 0x3fff3fff +#define WDMA_CLIP_COORD_MASK 0x3fff3fff +#define WDMA_DST_W_IN_BYTE_MASK 0x0000ffff +#define WDMA_ALPHA_MASK 0x800000ff +#define WDMA_BUF_CON2_MASK 0xffffffff +#define WDMA_DST_UV_PITCH_MASK 0x0000ffff +#define WDMA_DST_ADDR_OFFSET_MASK 0x0fffffff +#define WDMA_DST_U_ADDR_OFFSET_MASK 0x0fffffff +#define WDMA_DST_V_ADDR_OFFSET_MASK 0x0fffffff +#define WDMA_FLOW_CTRL_DBG_MASK 0x0000f3ff +#define WDMA_DST_ADDR_MASK 0xffffffff +#define WDMA_DST_U_ADDR_MASK 0xffffffff +#define WDMA_DST_V_ADDR_MASK 0xffffffff + +#endif // __MDP_REG_WDMA_H__ diff --git a/drivers/media/platform/mediatek/mdp3/mdp_reg_wrot.h b/drivers/media/platform/mediatek/mdp3/mdp_reg_wrot.h new file mode 100644 index 0000000000..6d3ff0e2b6 --- /dev/null +++ b/drivers/media/platform/mediatek/mdp3/mdp_reg_wrot.h @@ -0,0 +1,55 @@ +/* SPDX-License-Identifier: GPL-2.0-only */ +/* + * Copyright (c) 2022 MediaTek Inc. + * Author: Ping-Hsun Wu <ping-hsun.wu@mediatek.com> + */ + +#ifndef __MDP_REG_WROT_H__ +#define __MDP_REG_WROT_H__ + +#define VIDO_CTRL 0x000 +#define VIDO_MAIN_BUF_SIZE 0x008 +#define VIDO_SOFT_RST 0x010 +#define VIDO_SOFT_RST_STAT 0x014 +#define VIDO_CROP_OFST 0x020 +#define VIDO_TAR_SIZE 0x024 +#define VIDO_OFST_ADDR 0x02c +#define VIDO_STRIDE 0x030 +#define VIDO_OFST_ADDR_C 0x038 +#define VIDO_STRIDE_C 0x03c +#define VIDO_DITHER 0x054 +#define VIDO_STRIDE_V 0x06c +#define VIDO_OFST_ADDR_V 0x068 +#define VIDO_RSV_1 0x070 +#define VIDO_IN_SIZE 0x078 +#define VIDO_ROT_EN 0x07c +#define VIDO_FIFO_TEST 0x080 +#define VIDO_MAT_CTRL 0x084 +#define VIDO_BASE_ADDR 0xf00 +#define VIDO_BASE_ADDR_C 0xf04 +#define VIDO_BASE_ADDR_V 0xf08 + +/* MASK */ +#define VIDO_CTRL_MASK 0xf530711f +#define VIDO_MAIN_BUF_SIZE_MASK 0x1fff7f77 +#define VIDO_SOFT_RST_MASK 0x00000001 +#define VIDO_SOFT_RST_STAT_MASK 0x00000001 +#define VIDO_TAR_SIZE_MASK 0x1fff1fff +#define VIDO_CROP_OFST_MASK 0x1fff1fff +#define VIDO_OFST_ADDR_MASK 0x0fffffff +#define VIDO_STRIDE_MASK 0x0000ffff +#define VIDO_OFST_ADDR_C_MASK 0x0fffffff +#define VIDO_STRIDE_C_MASK 0x0000ffff +#define VIDO_DITHER_MASK 0xff000001 +#define VIDO_STRIDE_V_MASK 0x0000ffff +#define VIDO_OFST_ADDR_V_MASK 0x0fffffff +#define VIDO_RSV_1_MASK 0xffffffff +#define VIDO_IN_SIZE_MASK 0x1fff1fff +#define VIDO_ROT_EN_MASK 0x00000001 +#define VIDO_FIFO_TEST_MASK 0x00000fff +#define VIDO_MAT_CTRL_MASK 0x000000f3 +#define VIDO_BASE_ADDR_MASK 0xffffffff +#define VIDO_BASE_ADDR_C_MASK 0xffffffff +#define VIDO_BASE_ADDR_V_MASK 0xffffffff + +#endif // __MDP_REG_WROT_H__ diff --git a/drivers/media/platform/mediatek/mdp3/mdp_sm_mt8183.h b/drivers/media/platform/mediatek/mdp3/mdp_sm_mt8183.h new file mode 100644 index 0000000000..8563708452 --- /dev/null +++ b/drivers/media/platform/mediatek/mdp3/mdp_sm_mt8183.h @@ -0,0 +1,144 @@ +/* SPDX-License-Identifier: GPL-2.0-only */ +/* + * Copyright (c) 2023 MediaTek Inc. + * Author: Ping-Hsun Wu <ping-hsun.wu@mediatek.com> + */ + +#ifndef __MDP_SM_MT8183_H__ +#define __MDP_SM_MT8183_H__ + +#include "mtk-mdp3-type.h" + +/* + * ISP-MDP generic output information + * MD5 of the target SCP prebuild: + * 2d995ddb5c3b0cf26e96d6a823481886 + */ + +#define IMG_MAX_SUBFRAMES_8183 14 + +struct img_comp_frame_8183 { + u32 output_disable:1; + u32 bypass:1; + u16 in_width; + u16 in_height; + u16 out_width; + u16 out_height; + struct img_crop crop; + u16 in_total_width; + u16 out_total_width; +} __packed; + +struct img_comp_subfrm_8183 { + u32 tile_disable:1; + struct img_region in; + struct img_region out; + struct img_offset luma; + struct img_offset chroma; + s16 out_vertical; /* Output vertical index */ + s16 out_horizontal; /* Output horizontal index */ +} __packed; + +struct mdp_rdma_subfrm_8183 { + u32 offset[IMG_MAX_PLANES]; + u32 offset_0_p; + u32 src; + u32 clip; + u32 clip_ofst; +} __packed; + +struct mdp_rdma_data_8183 { + u32 src_ctrl; + u32 control; + u32 iova[IMG_MAX_PLANES]; + u32 iova_end[IMG_MAX_PLANES]; + u32 mf_bkgd; + u32 mf_bkgd_in_pxl; + u32 sf_bkgd; + u32 ufo_dec_y; + u32 ufo_dec_c; + u32 transform; + struct mdp_rdma_subfrm_8183 subfrms[IMG_MAX_SUBFRAMES_8183]; +} __packed; + +struct mdp_rsz_subfrm_8183 { + u32 control2; + u32 src; + u32 clip; +} __packed; + +struct mdp_rsz_data_8183 { + u32 coeff_step_x; + u32 coeff_step_y; + u32 control1; + u32 control2; + struct mdp_rsz_subfrm_8183 subfrms[IMG_MAX_SUBFRAMES_8183]; +} __packed; + +struct mdp_wrot_subfrm_8183 { + u32 offset[IMG_MAX_PLANES]; + u32 src; + u32 clip; + u32 clip_ofst; + u32 main_buf; +} __packed; + +struct mdp_wrot_data_8183 { + u32 iova[IMG_MAX_PLANES]; + u32 control; + u32 stride[IMG_MAX_PLANES]; + u32 mat_ctrl; + u32 fifo_test; + u32 filter; + struct mdp_wrot_subfrm_8183 subfrms[IMG_MAX_SUBFRAMES_8183]; +} __packed; + +struct mdp_wdma_subfrm_8183 { + u32 offset[IMG_MAX_PLANES]; + u32 src; + u32 clip; + u32 clip_ofst; +} __packed; + +struct mdp_wdma_data_8183 { + u32 wdma_cfg; + u32 iova[IMG_MAX_PLANES]; + u32 w_in_byte; + u32 uv_stride; + struct mdp_wdma_subfrm_8183 subfrms[IMG_MAX_SUBFRAMES_8183]; +} __packed; + +struct isp_data_8183 { + u64 dl_flags; /* 1 << (enum mdp_comp_type) */ + u32 smxi_iova[4]; + u32 cq_idx; + u32 cq_iova; + u32 tpipe_iova[IMG_MAX_SUBFRAMES_8183]; +} __packed; + +struct img_compparam_8183 { + u16 type; /* enum mdp_comp_id */ + u16 id; /* engine alias_id */ + u32 input; + u32 outputs[IMG_MAX_HW_OUTPUTS]; + u32 num_outputs; + struct img_comp_frame_8183 frame; + struct img_comp_subfrm_8183 subfrms[IMG_MAX_SUBFRAMES_8183]; + u32 num_subfrms; + union { + struct mdp_rdma_data_8183 rdma; + struct mdp_rsz_data_8183 rsz; + struct mdp_wrot_data_8183 wrot; + struct mdp_wdma_data_8183 wdma; + struct isp_data_8183 isp; + }; +} __packed; + +struct img_config_8183 { + struct img_compparam_8183 components[IMG_MAX_COMPONENTS]; + u32 num_components; + struct img_mmsys_ctrl ctrls[IMG_MAX_SUBFRAMES_8183]; + u32 num_subfrms; +} __packed; + +#endif /* __MDP_SM_MT8183_H__ */ diff --git a/drivers/media/platform/mediatek/mdp3/mtk-img-ipi.h b/drivers/media/platform/mediatek/mdp3/mtk-img-ipi.h new file mode 100644 index 0000000000..22b8b9a10e --- /dev/null +++ b/drivers/media/platform/mediatek/mdp3/mtk-img-ipi.h @@ -0,0 +1,149 @@ +/* SPDX-License-Identifier: GPL-2.0-only */ +/* + * Copyright (c) 2022 MediaTek Inc. + * Author: Holmes Chiou <holmes.chiou@mediatek.com> + * Ping-Hsun Wu <ping-hsun.wu@mediatek.com> + */ + +#ifndef __MTK_IMG_IPI_H__ +#define __MTK_IMG_IPI_H__ + +#include <linux/err.h> +#include "mdp_sm_mt8183.h" +#include "mtk-mdp3-type.h" + +/* ISP-MDP generic input information */ + +#define IMG_IPI_INIT 1 +#define IMG_IPI_DEINIT 2 +#define IMG_IPI_FRAME 3 +#define IMG_IPI_DEBUG 4 + +struct img_timeval { + u32 tv_sec; + u32 tv_usec; +} __packed; + +struct img_addr { + u64 va; /* Used for Linux OS access */ + u32 pa; /* Used for CM4 access */ + u32 iova; /* Used for IOMMU HW access */ +} __packed; + +struct tuning_addr { + u64 present; + u32 pa; /* Used for CM4 access */ + u32 iova; /* Used for IOMMU HW access */ +} __packed; + +struct img_sw_addr { + u64 va; /* Used for APMCU access */ + u32 pa; /* Used for CM4 access */ +} __packed; + +struct img_plane_format { + u32 size; + u32 stride; +} __packed; + +struct img_pix_format { + u32 width; + u32 height; + u32 colorformat; /* enum mdp_color */ + u32 ycbcr_prof; /* enum mdp_ycbcr_profile */ + struct img_plane_format plane_fmt[IMG_MAX_PLANES]; +} __packed; + +struct img_image_buffer { + struct img_pix_format format; + u32 iova[IMG_MAX_PLANES]; + /* enum mdp_buffer_usage, FD or advanced ISP usages */ + u32 usage; +} __packed; + +#define IMG_SUBPIXEL_SHIFT 20 + +#define IMG_CTRL_FLAG_HFLIP BIT(0) +#define IMG_CTRL_FLAG_DITHER BIT(1) +#define IMG_CTRL_FLAG_SHARPNESS BIT(4) +#define IMG_CTRL_FLAG_HDR BIT(5) +#define IMG_CTRL_FLAG_DRE BIT(6) + +struct img_input { + struct img_image_buffer buffer; + u32 flags; /* HDR, DRE, dither */ +} __packed; + +struct img_output { + struct img_image_buffer buffer; + struct img_crop crop; + s32 rotation; + u32 flags; /* H-flip, sharpness, dither */ +} __packed; + +struct img_ipi_frameparam { + u32 index; + u32 frame_no; + struct img_timeval timestamp; + u32 type; /* enum mdp_stream_type */ + u32 state; + u32 num_inputs; + u32 num_outputs; + u64 drv_data; + struct img_input inputs[IMG_MAX_HW_INPUTS]; + struct img_output outputs[IMG_MAX_HW_OUTPUTS]; + struct tuning_addr tuning_data; + struct img_addr subfrm_data; + struct img_sw_addr config_data; + struct img_sw_addr self_data; +} __packed; + +struct img_sw_buffer { + u64 handle; /* Used for APMCU access */ + u32 scp_addr; /* Used for CM4 access */ +} __packed; + +struct img_ipi_param { + u32 usage; + struct img_sw_buffer frm_param; +} __packed; + +struct img_frameparam { + struct list_head list_entry; + struct img_ipi_frameparam frameparam; +} __packed; + +/* Platform config indicator */ +#define MT8183 8183 + +#define CFG_CHECK(plat, p_id) ((plat) == (p_id)) + +#define _CFG_OFST(plat, cfg, ofst) ((void *)(&((cfg)->config_##plat) + (ofst))) +#define CFG_OFST(plat, cfg, ofst) \ + (IS_ERR_OR_NULL(cfg) ? NULL : _CFG_OFST(plat, cfg, ofst)) + +#define _CFG_ADDR(plat, cfg, mem) (&((cfg)->config_##plat.mem)) +#define CFG_ADDR(plat, cfg, mem) \ + (IS_ERR_OR_NULL(cfg) ? NULL : _CFG_ADDR(plat, cfg, mem)) + +#define _CFG_GET(plat, cfg, mem) ((cfg)->config_##plat.mem) +#define CFG_GET(plat, cfg, mem) \ + (IS_ERR_OR_NULL(cfg) ? 0 : _CFG_GET(plat, cfg, mem)) + +#define _CFG_COMP(plat, comp, mem) ((comp)->comp_##plat.mem) +#define CFG_COMP(plat, comp, mem) \ + (IS_ERR_OR_NULL(comp) ? 0 : _CFG_COMP(plat, comp, mem)) + +struct img_config { + union { + struct img_config_8183 config_8183; + }; +} __packed; + +struct img_compparam { + union { + struct img_compparam_8183 comp_8183; + }; +} __packed; + +#endif /* __MTK_IMG_IPI_H__ */ diff --git a/drivers/media/platform/mediatek/mdp3/mtk-mdp3-cfg.h b/drivers/media/platform/mediatek/mdp3/mtk-mdp3-cfg.h new file mode 100644 index 0000000000..dee57cc4a9 --- /dev/null +++ b/drivers/media/platform/mediatek/mdp3/mtk-mdp3-cfg.h @@ -0,0 +1,20 @@ +/* SPDX-License-Identifier: GPL-2.0-only */ +/* + * Copyright (c) 2023 MediaTek Inc. + * Author: Ping-Hsun Wu <ping-hsun.wu@mediatek.com> + */ + +#ifndef __MTK_MDP3_CFG_H__ +#define __MTK_MDP3_CFG_H__ + +#include <linux/types.h> + +extern const struct mtk_mdp_driver_data mt8183_mdp_driver_data; + +struct mdp_dev; +enum mtk_mdp_comp_id; + +s32 mdp_cfg_get_id_inner(struct mdp_dev *mdp_dev, enum mtk_mdp_comp_id id); +enum mtk_mdp_comp_id mdp_cfg_get_id_public(struct mdp_dev *mdp_dev, s32 id); + +#endif /* __MTK_MDP3_CFG_H__ */ diff --git a/drivers/media/platform/mediatek/mdp3/mtk-mdp3-cmdq.c b/drivers/media/platform/mediatek/mdp3/mtk-mdp3-cmdq.c new file mode 100644 index 0000000000..6adac857a4 --- /dev/null +++ b/drivers/media/platform/mediatek/mdp3/mtk-mdp3-cmdq.c @@ -0,0 +1,523 @@ +// SPDX-License-Identifier: GPL-2.0-only +/* + * Copyright (c) 2022 MediaTek Inc. + * Author: Ping-Hsun Wu <ping-hsun.wu@mediatek.com> + */ + +#include <linux/mailbox_controller.h> +#include <linux/platform_device.h> +#include "mtk-mdp3-cmdq.h" +#include "mtk-mdp3-comp.h" +#include "mtk-mdp3-core.h" +#include "mtk-mdp3-m2m.h" +#include "mtk-img-ipi.h" + +#define MDP_PATH_MAX_COMPS IMG_MAX_COMPONENTS + +struct mdp_path { + struct mdp_dev *mdp_dev; + struct mdp_comp_ctx comps[MDP_PATH_MAX_COMPS]; + u32 num_comps; + const struct img_config *config; + const struct img_ipi_frameparam *param; + const struct v4l2_rect *composes[IMG_MAX_HW_OUTPUTS]; + struct v4l2_rect bounds[IMG_MAX_HW_OUTPUTS]; +}; + +#define has_op(ctx, op) \ + ((ctx)->comp->ops && (ctx)->comp->ops->op) + #define call_op(ctx, op, ...) \ + (has_op(ctx, op) ? (ctx)->comp->ops->op(ctx, ##__VA_ARGS__) : 0) + +static bool is_output_disabled(int p_id, const struct img_compparam *param, u32 count) +{ + u32 num = 0; + bool dis_output = false; + bool dis_tile = false; + + if (CFG_CHECK(MT8183, p_id)) { + num = CFG_COMP(MT8183, param, num_subfrms); + dis_output = CFG_COMP(MT8183, param, frame.output_disable); + dis_tile = CFG_COMP(MT8183, param, frame.output_disable); + } + + return (count < num) ? (dis_output || dis_tile) : true; +} + +static int mdp_path_subfrm_require(const struct mdp_path *path, + struct mdp_cmdq_cmd *cmd, + s32 *mutex_id, u32 count) +{ + const int p_id = path->mdp_dev->mdp_data->mdp_plat_id; + const struct mdp_comp_ctx *ctx; + const struct mtk_mdp_driver_data *data = path->mdp_dev->mdp_data; + struct device *dev = &path->mdp_dev->pdev->dev; + struct mtk_mutex **mutex = path->mdp_dev->mdp_mutex; + int id, index; + u32 num_comp = 0; + + if (CFG_CHECK(MT8183, p_id)) + num_comp = CFG_GET(MT8183, path->config, num_components); + + /* Decide which mutex to use based on the current pipeline */ + switch (path->comps[0].comp->public_id) { + case MDP_COMP_RDMA0: + index = MDP_PIPE_RDMA0; + break; + case MDP_COMP_ISP_IMGI: + index = MDP_PIPE_IMGI; + break; + case MDP_COMP_WPEI: + index = MDP_PIPE_WPEI; + break; + case MDP_COMP_WPEI2: + index = MDP_PIPE_WPEI2; + break; + default: + dev_err(dev, "Unknown pipeline and no mutex is assigned"); + return -EINVAL; + } + *mutex_id = data->pipe_info[index].mutex_id; + + /* Set mutex mod */ + for (index = 0; index < num_comp; index++) { + ctx = &path->comps[index]; + if (is_output_disabled(p_id, ctx->param, count)) + continue; + id = ctx->comp->public_id; + mtk_mutex_write_mod(mutex[*mutex_id], + data->mdp_mutex_table_idx[id], false); + } + + mtk_mutex_write_sof(mutex[*mutex_id], + MUTEX_SOF_IDX_SINGLE_MODE); + + return 0; +} + +static int mdp_path_subfrm_run(const struct mdp_path *path, + struct mdp_cmdq_cmd *cmd, + s32 *mutex_id, u32 count) +{ + const int p_id = path->mdp_dev->mdp_data->mdp_plat_id; + const struct mdp_comp_ctx *ctx; + struct device *dev = &path->mdp_dev->pdev->dev; + struct mtk_mutex **mutex = path->mdp_dev->mdp_mutex; + int index; + u32 num_comp = 0; + s32 event; + + if (-1 == *mutex_id) { + dev_err(dev, "Incorrect mutex id"); + return -EINVAL; + } + + if (CFG_CHECK(MT8183, p_id)) + num_comp = CFG_GET(MT8183, path->config, num_components); + + /* Wait WROT SRAM shared to DISP RDMA */ + /* Clear SOF event for each engine */ + for (index = 0; index < num_comp; index++) { + ctx = &path->comps[index]; + if (is_output_disabled(p_id, ctx->param, count)) + continue; + event = ctx->comp->gce_event[MDP_GCE_EVENT_SOF]; + if (event != MDP_GCE_NO_EVENT) + MM_REG_CLEAR(cmd, event); + } + + /* Enable the mutex */ + mtk_mutex_enable_by_cmdq(mutex[*mutex_id], (void *)&cmd->pkt); + + /* Wait SOF events and clear mutex modules (optional) */ + for (index = 0; index < num_comp; index++) { + ctx = &path->comps[index]; + if (is_output_disabled(p_id, ctx->param, count)) + continue; + event = ctx->comp->gce_event[MDP_GCE_EVENT_SOF]; + if (event != MDP_GCE_NO_EVENT) + MM_REG_WAIT(cmd, event); + } + + return 0; +} + +static int mdp_path_ctx_init(struct mdp_dev *mdp, struct mdp_path *path) +{ + const int p_id = mdp->mdp_data->mdp_plat_id; + void *param = NULL; + int index, ret; + u32 num_comp = 0; + + if (CFG_CHECK(MT8183, p_id)) + num_comp = CFG_GET(MT8183, path->config, num_components); + + if (num_comp < 1) + return -EINVAL; + + for (index = 0; index < num_comp; index++) { + if (CFG_CHECK(MT8183, p_id)) + param = (void *)CFG_ADDR(MT8183, path->config, components[index]); + ret = mdp_comp_ctx_config(mdp, &path->comps[index], + param, path->param); + if (ret) + return ret; + } + + return 0; +} + +static int mdp_path_config_subfrm(struct mdp_cmdq_cmd *cmd, + struct mdp_path *path, u32 count) +{ + const int p_id = path->mdp_dev->mdp_data->mdp_plat_id; + const struct img_mmsys_ctrl *ctrl = NULL; + const struct img_mux *set; + struct mdp_comp_ctx *ctx; + s32 mutex_id; + int index, ret; + u32 num_comp = 0; + + if (CFG_CHECK(MT8183, p_id)) + num_comp = CFG_GET(MT8183, path->config, num_components); + + if (CFG_CHECK(MT8183, p_id)) + ctrl = CFG_ADDR(MT8183, path->config, ctrls[count]); + + /* Acquire components */ + ret = mdp_path_subfrm_require(path, cmd, &mutex_id, count); + if (ret) + return ret; + /* Enable mux settings */ + for (index = 0; index < ctrl->num_sets; index++) { + set = &ctrl->sets[index]; + cmdq_pkt_write_mask(&cmd->pkt, set->subsys_id, set->reg, + set->value, 0xFFFFFFFF); + } + /* Config sub-frame information */ + for (index = (num_comp - 1); index >= 0; index--) { + ctx = &path->comps[index]; + if (is_output_disabled(p_id, ctx->param, count)) + continue; + ret = call_op(ctx, config_subfrm, cmd, count); + if (ret) + return ret; + } + /* Run components */ + ret = mdp_path_subfrm_run(path, cmd, &mutex_id, count); + if (ret) + return ret; + /* Wait components done */ + for (index = 0; index < num_comp; index++) { + ctx = &path->comps[index]; + if (is_output_disabled(p_id, ctx->param, count)) + continue; + ret = call_op(ctx, wait_comp_event, cmd); + if (ret) + return ret; + } + /* Advance to the next sub-frame */ + for (index = 0; index < num_comp; index++) { + ctx = &path->comps[index]; + ret = call_op(ctx, advance_subfrm, cmd, count); + if (ret) + return ret; + } + /* Disable mux settings */ + for (index = 0; index < ctrl->num_sets; index++) { + set = &ctrl->sets[index]; + cmdq_pkt_write_mask(&cmd->pkt, set->subsys_id, set->reg, + 0, 0xFFFFFFFF); + } + + return 0; +} + +static int mdp_path_config(struct mdp_dev *mdp, struct mdp_cmdq_cmd *cmd, + struct mdp_path *path) +{ + const int p_id = mdp->mdp_data->mdp_plat_id; + struct mdp_comp_ctx *ctx; + int index, count, ret; + u32 num_comp = 0; + u32 num_sub = 0; + + if (CFG_CHECK(MT8183, p_id)) + num_comp = CFG_GET(MT8183, path->config, num_components); + + if (CFG_CHECK(MT8183, p_id)) + num_sub = CFG_GET(MT8183, path->config, num_subfrms); + + /* Config path frame */ + /* Reset components */ + for (index = 0; index < num_comp; index++) { + ctx = &path->comps[index]; + ret = call_op(ctx, init_comp, cmd); + if (ret) + return ret; + } + /* Config frame mode */ + for (index = 0; index < num_comp; index++) { + const struct v4l2_rect *compose; + u32 out = 0; + + ctx = &path->comps[index]; + if (CFG_CHECK(MT8183, p_id)) + out = CFG_COMP(MT8183, ctx->param, outputs[0]); + + compose = path->composes[out]; + ret = call_op(ctx, config_frame, cmd, compose); + if (ret) + return ret; + } + + /* Config path sub-frames */ + for (count = 0; count < num_sub; count++) { + ret = mdp_path_config_subfrm(cmd, path, count); + if (ret) + return ret; + } + /* Post processing information */ + for (index = 0; index < num_comp; index++) { + ctx = &path->comps[index]; + ret = call_op(ctx, post_process, cmd); + if (ret) + return ret; + } + return 0; +} + +static int mdp_cmdq_pkt_create(struct cmdq_client *client, struct cmdq_pkt *pkt, + size_t size) +{ + struct device *dev; + dma_addr_t dma_addr; + + pkt->va_base = kzalloc(size, GFP_KERNEL); + if (!pkt->va_base) + return -ENOMEM; + + pkt->buf_size = size; + pkt->cl = (void *)client; + + dev = client->chan->mbox->dev; + dma_addr = dma_map_single(dev, pkt->va_base, pkt->buf_size, + DMA_TO_DEVICE); + if (dma_mapping_error(dev, dma_addr)) { + dev_err(dev, "dma map failed, size=%u\n", (u32)(u64)size); + kfree(pkt->va_base); + return -ENOMEM; + } + + pkt->pa_base = dma_addr; + + return 0; +} + +static void mdp_cmdq_pkt_destroy(struct cmdq_pkt *pkt) +{ + struct cmdq_client *client = (struct cmdq_client *)pkt->cl; + + dma_unmap_single(client->chan->mbox->dev, pkt->pa_base, pkt->buf_size, + DMA_TO_DEVICE); + kfree(pkt->va_base); + pkt->va_base = NULL; +} + +static void mdp_auto_release_work(struct work_struct *work) +{ + struct mdp_cmdq_cmd *cmd; + struct mdp_dev *mdp; + int id; + + cmd = container_of(work, struct mdp_cmdq_cmd, auto_release_work); + mdp = cmd->mdp; + + id = mdp->mdp_data->pipe_info[MDP_PIPE_RDMA0].mutex_id; + mtk_mutex_unprepare(mdp->mdp_mutex[id]); + mdp_comp_clocks_off(&mdp->pdev->dev, cmd->comps, + cmd->num_comps); + + atomic_dec(&mdp->job_count); + wake_up(&mdp->callback_wq); + + mdp_cmdq_pkt_destroy(&cmd->pkt); + kfree(cmd->comps); + cmd->comps = NULL; + kfree(cmd); + cmd = NULL; +} + +static void mdp_handle_cmdq_callback(struct mbox_client *cl, void *mssg) +{ + struct mdp_cmdq_cmd *cmd; + struct cmdq_cb_data *data; + struct mdp_dev *mdp; + struct device *dev; + int id; + + if (!mssg) { + pr_info("%s:no callback data\n", __func__); + return; + } + + data = (struct cmdq_cb_data *)mssg; + cmd = container_of(data->pkt, struct mdp_cmdq_cmd, pkt); + mdp = cmd->mdp; + dev = &mdp->pdev->dev; + + if (cmd->mdp_ctx) + mdp_m2m_job_finish(cmd->mdp_ctx); + + if (cmd->user_cmdq_cb) { + struct cmdq_cb_data user_cb_data; + + user_cb_data.sta = data->sta; + user_cb_data.pkt = data->pkt; + cmd->user_cmdq_cb(user_cb_data); + } + + INIT_WORK(&cmd->auto_release_work, mdp_auto_release_work); + if (!queue_work(mdp->clock_wq, &cmd->auto_release_work)) { + dev_err(dev, "%s:queue_work fail!\n", __func__); + id = mdp->mdp_data->pipe_info[MDP_PIPE_RDMA0].mutex_id; + mtk_mutex_unprepare(mdp->mdp_mutex[id]); + mdp_comp_clocks_off(&mdp->pdev->dev, cmd->comps, + cmd->num_comps); + + atomic_dec(&mdp->job_count); + wake_up(&mdp->callback_wq); + + mdp_cmdq_pkt_destroy(&cmd->pkt); + kfree(cmd->comps); + cmd->comps = NULL; + kfree(cmd); + cmd = NULL; + } +} + +int mdp_cmdq_send(struct mdp_dev *mdp, struct mdp_cmdq_param *param) +{ + struct mdp_path *path = NULL; + struct mdp_cmdq_cmd *cmd = NULL; + struct mdp_comp *comps = NULL; + struct device *dev = &mdp->pdev->dev; + const int p_id = mdp->mdp_data->mdp_plat_id; + int i, ret; + u32 num_comp = 0; + + atomic_inc(&mdp->job_count); + if (atomic_read(&mdp->suspended)) { + atomic_dec(&mdp->job_count); + return -ECANCELED; + } + + cmd = kzalloc(sizeof(*cmd), GFP_KERNEL); + if (!cmd) { + ret = -ENOMEM; + goto err_cancel_job; + } + + ret = mdp_cmdq_pkt_create(mdp->cmdq_clt, &cmd->pkt, SZ_16K); + if (ret) + goto err_free_cmd; + + if (CFG_CHECK(MT8183, p_id)) { + num_comp = CFG_GET(MT8183, param->config, num_components); + } else { + ret = -EINVAL; + goto err_destroy_pkt; + } + comps = kcalloc(num_comp, sizeof(*comps), GFP_KERNEL); + if (!comps) { + ret = -ENOMEM; + goto err_destroy_pkt; + } + + path = kzalloc(sizeof(*path), GFP_KERNEL); + if (!path) { + ret = -ENOMEM; + goto err_free_comps; + } + + i = mdp->mdp_data->pipe_info[MDP_PIPE_RDMA0].mutex_id; + ret = mtk_mutex_prepare(mdp->mdp_mutex[i]); + if (ret) { + dev_err(dev, "Fail to enable mutex clk\n"); + goto err_free_path; + } + + path->mdp_dev = mdp; + path->config = param->config; + path->param = param->param; + for (i = 0; i < param->param->num_outputs; i++) { + path->bounds[i].left = 0; + path->bounds[i].top = 0; + path->bounds[i].width = + param->param->outputs[i].buffer.format.width; + path->bounds[i].height = + param->param->outputs[i].buffer.format.height; + path->composes[i] = param->composes[i] ? + param->composes[i] : &path->bounds[i]; + } + ret = mdp_path_ctx_init(mdp, path); + if (ret) { + dev_err(dev, "mdp_path_ctx_init error\n"); + goto err_free_path; + } + + ret = mdp_path_config(mdp, cmd, path); + if (ret) { + dev_err(dev, "mdp_path_config error\n"); + goto err_free_path; + } + cmdq_pkt_finalize(&cmd->pkt); + + for (i = 0; i < num_comp; i++) + memcpy(&comps[i], path->comps[i].comp, + sizeof(struct mdp_comp)); + + mdp->cmdq_clt->client.rx_callback = mdp_handle_cmdq_callback; + cmd->mdp = mdp; + cmd->user_cmdq_cb = param->cmdq_cb; + cmd->user_cb_data = param->cb_data; + cmd->comps = comps; + cmd->num_comps = num_comp; + cmd->mdp_ctx = param->mdp_ctx; + + ret = mdp_comp_clocks_on(&mdp->pdev->dev, cmd->comps, cmd->num_comps); + if (ret) + goto err_free_path; + + dma_sync_single_for_device(mdp->cmdq_clt->chan->mbox->dev, + cmd->pkt.pa_base, cmd->pkt.cmd_buf_size, + DMA_TO_DEVICE); + ret = mbox_send_message(mdp->cmdq_clt->chan, &cmd->pkt); + if (ret < 0) { + dev_err(dev, "mbox send message fail %d!\n", ret); + goto err_clock_off; + } + mbox_client_txdone(mdp->cmdq_clt->chan, 0); + + kfree(path); + return 0; + +err_clock_off: + mdp_comp_clocks_off(&mdp->pdev->dev, cmd->comps, + cmd->num_comps); +err_free_path: + i = mdp->mdp_data->pipe_info[MDP_PIPE_RDMA0].mutex_id; + mtk_mutex_unprepare(mdp->mdp_mutex[i]); + kfree(path); +err_free_comps: + kfree(comps); +err_destroy_pkt: + mdp_cmdq_pkt_destroy(&cmd->pkt); +err_free_cmd: + kfree(cmd); +err_cancel_job: + atomic_dec(&mdp->job_count); + + return ret; +} +EXPORT_SYMBOL_GPL(mdp_cmdq_send); diff --git a/drivers/media/platform/mediatek/mdp3/mtk-mdp3-cmdq.h b/drivers/media/platform/mediatek/mdp3/mtk-mdp3-cmdq.h new file mode 100644 index 0000000000..43475b862d --- /dev/null +++ b/drivers/media/platform/mediatek/mdp3/mtk-mdp3-cmdq.h @@ -0,0 +1,43 @@ +/* SPDX-License-Identifier: GPL-2.0-only */ +/* + * Copyright (c) 2022 MediaTek Inc. + * Author: Ping-Hsun Wu <ping-hsun.wu@mediatek.com> + */ + +#ifndef __MTK_MDP3_CMDQ_H__ +#define __MTK_MDP3_CMDQ_H__ + +#include <linux/platform_device.h> +#include <linux/videodev2.h> +#include <linux/soc/mediatek/mtk-cmdq.h> +#include "mtk-img-ipi.h" + +struct platform_device *mdp_get_plat_device(struct platform_device *pdev); + +struct mdp_cmdq_param { + struct img_config *config; + struct img_ipi_frameparam *param; + const struct v4l2_rect *composes[IMG_MAX_HW_OUTPUTS]; + + void (*cmdq_cb)(struct cmdq_cb_data data); + void *cb_data; + void *mdp_ctx; +}; + +struct mdp_cmdq_cmd { + struct work_struct auto_release_work; + struct cmdq_pkt pkt; + s32 *event; + struct mdp_dev *mdp; + void (*user_cmdq_cb)(struct cmdq_cb_data data); + void *user_cb_data; + struct mdp_comp *comps; + void *mdp_ctx; + u8 num_comps; +}; + +struct mdp_dev; + +int mdp_cmdq_send(struct mdp_dev *mdp, struct mdp_cmdq_param *param); + +#endif /* __MTK_MDP3_CMDQ_H__ */ diff --git a/drivers/media/platform/mediatek/mdp3/mtk-mdp3-comp.c b/drivers/media/platform/mediatek/mdp3/mtk-mdp3-comp.c new file mode 100644 index 0000000000..667933ea15 --- /dev/null +++ b/drivers/media/platform/mediatek/mdp3/mtk-mdp3-comp.c @@ -0,0 +1,1208 @@ +// SPDX-License-Identifier: GPL-2.0-only +/* + * Copyright (c) 2022 MediaTek Inc. + * Author: Ping-Hsun Wu <ping-hsun.wu@mediatek.com> + */ + +#include <linux/clk.h> +#include <linux/of_platform.h> +#include <linux/of_address.h> +#include <linux/pm_runtime.h> +#include "mtk-mdp3-cfg.h" +#include "mtk-mdp3-comp.h" +#include "mtk-mdp3-core.h" +#include "mtk-mdp3-regs.h" + +#include "mdp_reg_rdma.h" +#include "mdp_reg_ccorr.h" +#include "mdp_reg_rsz.h" +#include "mdp_reg_wrot.h" +#include "mdp_reg_wdma.h" + +static u32 mdp_comp_alias_id[MDP_COMP_TYPE_COUNT]; +static int p_id; + +static inline const struct mdp_platform_config * +__get_plat_cfg(const struct mdp_comp_ctx *ctx) +{ + if (!ctx) + return NULL; + + return ctx->comp->mdp_dev->mdp_data->mdp_cfg; +} + +static s64 get_comp_flag(const struct mdp_comp_ctx *ctx) +{ + const struct mdp_platform_config *mdp_cfg = __get_plat_cfg(ctx); + u32 rdma0, rsz1; + + rdma0 = mdp_cfg_get_id_inner(ctx->comp->mdp_dev, MDP_COMP_RDMA0); + rsz1 = mdp_cfg_get_id_inner(ctx->comp->mdp_dev, MDP_COMP_RSZ1); + if (!rdma0 || !rsz1) + return MDP_COMP_NONE; + + if (mdp_cfg && mdp_cfg->rdma_rsz1_sram_sharing) + if (ctx->comp->inner_id == rdma0) + return BIT(rdma0) | BIT(rsz1); + + return BIT(ctx->comp->inner_id); +} + +static int init_rdma(struct mdp_comp_ctx *ctx, struct mdp_cmdq_cmd *cmd) +{ + const struct mdp_platform_config *mdp_cfg = __get_plat_cfg(ctx); + phys_addr_t base = ctx->comp->reg_base; + u8 subsys_id = ctx->comp->subsys_id; + s32 rdma0; + + rdma0 = mdp_cfg_get_id_inner(ctx->comp->mdp_dev, MDP_COMP_RDMA0); + if (!rdma0) + return -EINVAL; + + if (mdp_cfg && mdp_cfg->rdma_support_10bit) { + struct mdp_comp *prz1 = ctx->comp->mdp_dev->comp[MDP_COMP_RSZ1]; + + /* Disable RSZ1 */ + if (ctx->comp->inner_id == rdma0 && prz1) + MM_REG_WRITE(cmd, subsys_id, prz1->reg_base, PRZ_ENABLE, + 0x0, BIT(0)); + } + + /* Reset RDMA */ + MM_REG_WRITE(cmd, subsys_id, base, MDP_RDMA_RESET, BIT(0), BIT(0)); + MM_REG_POLL(cmd, subsys_id, base, MDP_RDMA_MON_STA_1, BIT(8), BIT(8)); + MM_REG_WRITE(cmd, subsys_id, base, MDP_RDMA_RESET, 0x0, BIT(0)); + return 0; +} + +static int config_rdma_frame(struct mdp_comp_ctx *ctx, + struct mdp_cmdq_cmd *cmd, + const struct v4l2_rect *compose) +{ + const struct mdp_platform_config *mdp_cfg = __get_plat_cfg(ctx); + u32 colorformat = ctx->input->buffer.format.colorformat; + bool block10bit = MDP_COLOR_IS_10BIT_PACKED(colorformat); + bool en_ufo = MDP_COLOR_IS_UFP(colorformat); + phys_addr_t base = ctx->comp->reg_base; + u8 subsys_id = ctx->comp->subsys_id; + u32 reg = 0; + + if (mdp_cfg && mdp_cfg->rdma_support_10bit) { + if (block10bit) + MM_REG_WRITE(cmd, subsys_id, base, + MDP_RDMA_RESV_DUMMY_0, 0x7, 0x7); + else + MM_REG_WRITE(cmd, subsys_id, base, + MDP_RDMA_RESV_DUMMY_0, 0x0, 0x7); + } + + /* Setup smi control */ + MM_REG_WRITE(cmd, subsys_id, base, MDP_RDMA_GMCIF_CON, + (7 << 4) + //burst type to 8 + (1 << 16), //enable pre-ultra + 0x00030071); + + /* Setup source frame info */ + if (CFG_CHECK(MT8183, p_id)) + reg = CFG_COMP(MT8183, ctx->param, rdma.src_ctrl); + MM_REG_WRITE(cmd, subsys_id, base, MDP_RDMA_SRC_CON, reg, + 0x03C8FE0F); + + if (mdp_cfg) + if (mdp_cfg->rdma_support_10bit && en_ufo) { + /* Setup source buffer base */ + if (CFG_CHECK(MT8183, p_id)) + reg = CFG_COMP(MT8183, ctx->param, rdma.ufo_dec_y); + MM_REG_WRITE(cmd, subsys_id, + base, MDP_RDMA_UFO_DEC_LENGTH_BASE_Y, + reg, 0xFFFFFFFF); + if (CFG_CHECK(MT8183, p_id)) + reg = CFG_COMP(MT8183, ctx->param, rdma.ufo_dec_c); + MM_REG_WRITE(cmd, subsys_id, + base, MDP_RDMA_UFO_DEC_LENGTH_BASE_C, + reg, 0xFFFFFFFF); + /* Set 10bit source frame pitch */ + if (block10bit) { + if (CFG_CHECK(MT8183, p_id)) + reg = CFG_COMP(MT8183, ctx->param, rdma.mf_bkgd_in_pxl); + MM_REG_WRITE(cmd, subsys_id, + base, MDP_RDMA_MF_BKGD_SIZE_IN_PXL, + reg, 0x001FFFFF); + } + } + + if (CFG_CHECK(MT8183, p_id)) + reg = CFG_COMP(MT8183, ctx->param, rdma.control); + MM_REG_WRITE(cmd, subsys_id, base, MDP_RDMA_CON, reg, + 0x1110); + /* Setup source buffer base */ + if (CFG_CHECK(MT8183, p_id)) + reg = CFG_COMP(MT8183, ctx->param, rdma.iova[0]); + MM_REG_WRITE(cmd, subsys_id, base, MDP_RDMA_SRC_BASE_0, reg, + 0xFFFFFFFF); + if (CFG_CHECK(MT8183, p_id)) + reg = CFG_COMP(MT8183, ctx->param, rdma.iova[1]); + MM_REG_WRITE(cmd, subsys_id, base, MDP_RDMA_SRC_BASE_1, reg, + 0xFFFFFFFF); + if (CFG_CHECK(MT8183, p_id)) + reg = CFG_COMP(MT8183, ctx->param, rdma.iova[2]); + MM_REG_WRITE(cmd, subsys_id, base, MDP_RDMA_SRC_BASE_2, reg, + 0xFFFFFFFF); + /* Setup source buffer end */ + if (CFG_CHECK(MT8183, p_id)) + reg = CFG_COMP(MT8183, ctx->param, rdma.iova_end[0]); + MM_REG_WRITE(cmd, subsys_id, base, MDP_RDMA_SRC_END_0, + reg, 0xFFFFFFFF); + if (CFG_CHECK(MT8183, p_id)) + reg = CFG_COMP(MT8183, ctx->param, rdma.iova_end[1]); + MM_REG_WRITE(cmd, subsys_id, base, MDP_RDMA_SRC_END_1, + reg, 0xFFFFFFFF); + if (CFG_CHECK(MT8183, p_id)) + reg = CFG_COMP(MT8183, ctx->param, rdma.iova_end[2]); + MM_REG_WRITE(cmd, subsys_id, base, MDP_RDMA_SRC_END_2, + reg, 0xFFFFFFFF); + /* Setup source frame pitch */ + if (CFG_CHECK(MT8183, p_id)) + reg = CFG_COMP(MT8183, ctx->param, rdma.mf_bkgd); + MM_REG_WRITE(cmd, subsys_id, base, MDP_RDMA_MF_BKGD_SIZE_IN_BYTE, + reg, 0x001FFFFF); + if (CFG_CHECK(MT8183, p_id)) + reg = CFG_COMP(MT8183, ctx->param, rdma.sf_bkgd); + MM_REG_WRITE(cmd, subsys_id, base, MDP_RDMA_SF_BKGD_SIZE_IN_BYTE, + reg, 0x001FFFFF); + /* Setup color transform */ + if (CFG_CHECK(MT8183, p_id)) + reg = CFG_COMP(MT8183, ctx->param, rdma.transform); + MM_REG_WRITE(cmd, subsys_id, base, MDP_RDMA_TRANSFORM_0, + reg, 0x0F110000); + + return 0; +} + +static int config_rdma_subfrm(struct mdp_comp_ctx *ctx, + struct mdp_cmdq_cmd *cmd, u32 index) +{ + const struct mdp_platform_config *mdp_cfg = __get_plat_cfg(ctx); + u32 colorformat = ctx->input->buffer.format.colorformat; + bool block10bit = MDP_COLOR_IS_10BIT_PACKED(colorformat); + bool en_ufo = MDP_COLOR_IS_UFP(colorformat); + phys_addr_t base = ctx->comp->reg_base; + u8 subsys_id = ctx->comp->subsys_id; + u32 csf_l = 0, csf_r = 0; + u32 reg = 0; + + /* Enable RDMA */ + MM_REG_WRITE(cmd, subsys_id, base, MDP_RDMA_EN, BIT(0), BIT(0)); + + /* Set Y pixel offset */ + if (CFG_CHECK(MT8183, p_id)) + reg = CFG_COMP(MT8183, ctx->param, rdma.subfrms[index].offset[0]); + MM_REG_WRITE(cmd, subsys_id, base, MDP_RDMA_SRC_OFFSET_0, + reg, 0xFFFFFFFF); + + /* Set 10bit UFO mode */ + if (mdp_cfg) { + if (mdp_cfg->rdma_support_10bit && block10bit && en_ufo) { + if (CFG_CHECK(MT8183, p_id)) + reg = CFG_COMP(MT8183, ctx->param, rdma.subfrms[index].offset_0_p); + MM_REG_WRITE(cmd, subsys_id, base, + MDP_RDMA_SRC_OFFSET_0_P, + reg, 0xFFFFFFFF); + } + } + + /* Set U pixel offset */ + if (CFG_CHECK(MT8183, p_id)) + reg = CFG_COMP(MT8183, ctx->param, rdma.subfrms[index].offset[1]); + MM_REG_WRITE(cmd, subsys_id, base, MDP_RDMA_SRC_OFFSET_1, + reg, 0xFFFFFFFF); + /* Set V pixel offset */ + if (CFG_CHECK(MT8183, p_id)) + reg = CFG_COMP(MT8183, ctx->param, rdma.subfrms[index].offset[2]); + MM_REG_WRITE(cmd, subsys_id, base, MDP_RDMA_SRC_OFFSET_2, + reg, 0xFFFFFFFF); + /* Set source size */ + if (CFG_CHECK(MT8183, p_id)) + reg = CFG_COMP(MT8183, ctx->param, rdma.subfrms[index].src); + MM_REG_WRITE(cmd, subsys_id, base, MDP_RDMA_MF_SRC_SIZE, reg, + 0x1FFF1FFF); + /* Set target size */ + if (CFG_CHECK(MT8183, p_id)) + reg = CFG_COMP(MT8183, ctx->param, rdma.subfrms[index].clip); + MM_REG_WRITE(cmd, subsys_id, base, MDP_RDMA_MF_CLIP_SIZE, + reg, 0x1FFF1FFF); + /* Set crop offset */ + if (CFG_CHECK(MT8183, p_id)) + reg = CFG_COMP(MT8183, ctx->param, rdma.subfrms[index].clip_ofst); + MM_REG_WRITE(cmd, subsys_id, base, MDP_RDMA_MF_OFFSET_1, + reg, 0x003F001F); + + if (CFG_CHECK(MT8183, p_id)) { + csf_l = CFG_COMP(MT8183, ctx->param, subfrms[index].in.left); + csf_r = CFG_COMP(MT8183, ctx->param, subfrms[index].in.right); + } + if (mdp_cfg && mdp_cfg->rdma_upsample_repeat_only) + if ((csf_r - csf_l + 1) > 320) + MM_REG_WRITE(cmd, subsys_id, base, + MDP_RDMA_RESV_DUMMY_0, BIT(2), BIT(2)); + + return 0; +} + +static int wait_rdma_event(struct mdp_comp_ctx *ctx, struct mdp_cmdq_cmd *cmd) +{ + struct device *dev = &ctx->comp->mdp_dev->pdev->dev; + phys_addr_t base = ctx->comp->reg_base; + u8 subsys_id = ctx->comp->subsys_id; + + if (ctx->comp->alias_id == 0) + MM_REG_WAIT(cmd, ctx->comp->gce_event[MDP_GCE_EVENT_EOF]); + else + dev_err(dev, "Do not support RDMA1_DONE event\n"); + + /* Disable RDMA */ + MM_REG_WRITE(cmd, subsys_id, base, MDP_RDMA_EN, 0x0, BIT(0)); + return 0; +} + +static const struct mdp_comp_ops rdma_ops = { + .get_comp_flag = get_comp_flag, + .init_comp = init_rdma, + .config_frame = config_rdma_frame, + .config_subfrm = config_rdma_subfrm, + .wait_comp_event = wait_rdma_event, +}; + +static int init_rsz(struct mdp_comp_ctx *ctx, struct mdp_cmdq_cmd *cmd) +{ + phys_addr_t base = ctx->comp->reg_base; + u8 subsys_id = ctx->comp->subsys_id; + + /* Reset RSZ */ + MM_REG_WRITE(cmd, subsys_id, base, PRZ_ENABLE, 0x10000, BIT(16)); + MM_REG_WRITE(cmd, subsys_id, base, PRZ_ENABLE, 0x0, BIT(16)); + /* Enable RSZ */ + MM_REG_WRITE(cmd, subsys_id, base, PRZ_ENABLE, BIT(0), BIT(0)); + return 0; +} + +static int config_rsz_frame(struct mdp_comp_ctx *ctx, + struct mdp_cmdq_cmd *cmd, + const struct v4l2_rect *compose) +{ + phys_addr_t base = ctx->comp->reg_base; + u8 subsys_id = ctx->comp->subsys_id; + bool bypass = FALSE; + u32 reg = 0; + + if (CFG_CHECK(MT8183, p_id)) + bypass = CFG_COMP(MT8183, ctx->param, frame.bypass); + + if (bypass) { + /* Disable RSZ */ + MM_REG_WRITE(cmd, subsys_id, base, PRZ_ENABLE, 0x0, BIT(0)); + return 0; + } + + if (CFG_CHECK(MT8183, p_id)) + reg = CFG_COMP(MT8183, ctx->param, rsz.control1); + MM_REG_WRITE(cmd, subsys_id, base, PRZ_CONTROL_1, reg, + 0x03FFFDF3); + if (CFG_CHECK(MT8183, p_id)) + reg = CFG_COMP(MT8183, ctx->param, rsz.control2); + MM_REG_WRITE(cmd, subsys_id, base, PRZ_CONTROL_2, reg, + 0x0FFFC290); + if (CFG_CHECK(MT8183, p_id)) + reg = CFG_COMP(MT8183, ctx->param, rsz.coeff_step_x); + MM_REG_WRITE(cmd, subsys_id, base, PRZ_HORIZONTAL_COEFF_STEP, + reg, 0x007FFFFF); + if (CFG_CHECK(MT8183, p_id)) + reg = CFG_COMP(MT8183, ctx->param, rsz.coeff_step_y); + MM_REG_WRITE(cmd, subsys_id, base, PRZ_VERTICAL_COEFF_STEP, + reg, 0x007FFFFF); + return 0; +} + +static int config_rsz_subfrm(struct mdp_comp_ctx *ctx, + struct mdp_cmdq_cmd *cmd, u32 index) +{ + const struct mdp_platform_config *mdp_cfg = __get_plat_cfg(ctx); + phys_addr_t base = ctx->comp->reg_base; + u8 subsys_id = ctx->comp->subsys_id; + u32 csf_l = 0, csf_r = 0; + u32 reg = 0; + + if (CFG_CHECK(MT8183, p_id)) + reg = CFG_COMP(MT8183, ctx->param, rsz.subfrms[index].control2); + MM_REG_WRITE(cmd, subsys_id, base, PRZ_CONTROL_2, reg, + 0x00003800); + if (CFG_CHECK(MT8183, p_id)) + reg = CFG_COMP(MT8183, ctx->param, rsz.subfrms[index].src); + MM_REG_WRITE(cmd, subsys_id, base, PRZ_INPUT_IMAGE, reg, + 0xFFFFFFFF); + + if (CFG_CHECK(MT8183, p_id)) { + csf_l = CFG_COMP(MT8183, ctx->param, subfrms[index].in.left); + csf_r = CFG_COMP(MT8183, ctx->param, subfrms[index].in.right); + } + if (mdp_cfg && mdp_cfg->rsz_disable_dcm_small_sample) + if ((csf_r - csf_l + 1) <= 16) + MM_REG_WRITE(cmd, subsys_id, base, PRZ_CONTROL_1, + BIT(27), BIT(27)); + + if (CFG_CHECK(MT8183, p_id)) + reg = CFG_COMP(MT8183, ctx->param, subfrms[index].luma.left); + MM_REG_WRITE(cmd, subsys_id, base, PRZ_LUMA_HORIZONTAL_INTEGER_OFFSET, + reg, 0xFFFF); + if (CFG_CHECK(MT8183, p_id)) + reg = CFG_COMP(MT8183, ctx->param, subfrms[index].luma.left_subpix); + MM_REG_WRITE(cmd, subsys_id, + base, PRZ_LUMA_HORIZONTAL_SUBPIXEL_OFFSET, + reg, 0x1FFFFF); + if (CFG_CHECK(MT8183, p_id)) + reg = CFG_COMP(MT8183, ctx->param, subfrms[index].luma.top); + MM_REG_WRITE(cmd, subsys_id, base, PRZ_LUMA_VERTICAL_INTEGER_OFFSET, + reg, 0xFFFF); + if (CFG_CHECK(MT8183, p_id)) + reg = CFG_COMP(MT8183, ctx->param, subfrms[index].luma.top_subpix); + MM_REG_WRITE(cmd, subsys_id, base, PRZ_LUMA_VERTICAL_SUBPIXEL_OFFSET, + reg, 0x1FFFFF); + if (CFG_CHECK(MT8183, p_id)) + reg = CFG_COMP(MT8183, ctx->param, subfrms[index].chroma.left); + MM_REG_WRITE(cmd, subsys_id, + base, PRZ_CHROMA_HORIZONTAL_INTEGER_OFFSET, + reg, 0xFFFF); + if (CFG_CHECK(MT8183, p_id)) + reg = CFG_COMP(MT8183, ctx->param, subfrms[index].chroma.left_subpix); + MM_REG_WRITE(cmd, subsys_id, + base, PRZ_CHROMA_HORIZONTAL_SUBPIXEL_OFFSET, + reg, 0x1FFFFF); + + if (CFG_CHECK(MT8183, p_id)) + reg = CFG_COMP(MT8183, ctx->param, rsz.subfrms[index].clip); + MM_REG_WRITE(cmd, subsys_id, base, PRZ_OUTPUT_IMAGE, reg, + 0xFFFFFFFF); + + return 0; +} + +static int advance_rsz_subfrm(struct mdp_comp_ctx *ctx, + struct mdp_cmdq_cmd *cmd, u32 index) +{ + const struct mdp_platform_config *mdp_cfg = __get_plat_cfg(ctx); + + if (mdp_cfg && mdp_cfg->rsz_disable_dcm_small_sample) { + phys_addr_t base = ctx->comp->reg_base; + u8 subsys_id = ctx->comp->subsys_id; + u32 csf_l = 0, csf_r = 0; + + if (CFG_CHECK(MT8183, p_id)) { + csf_l = CFG_COMP(MT8183, ctx->param, subfrms[index].in.left); + csf_r = CFG_COMP(MT8183, ctx->param, subfrms[index].in.right); + } + + if ((csf_r - csf_l + 1) <= 16) + MM_REG_WRITE(cmd, subsys_id, base, PRZ_CONTROL_1, 0x0, + BIT(27)); + } + + return 0; +} + +static const struct mdp_comp_ops rsz_ops = { + .get_comp_flag = get_comp_flag, + .init_comp = init_rsz, + .config_frame = config_rsz_frame, + .config_subfrm = config_rsz_subfrm, + .advance_subfrm = advance_rsz_subfrm, +}; + +static int init_wrot(struct mdp_comp_ctx *ctx, struct mdp_cmdq_cmd *cmd) +{ + phys_addr_t base = ctx->comp->reg_base; + u8 subsys_id = ctx->comp->subsys_id; + + /* Reset WROT */ + MM_REG_WRITE(cmd, subsys_id, base, VIDO_SOFT_RST, BIT(0), BIT(0)); + MM_REG_POLL(cmd, subsys_id, base, VIDO_SOFT_RST_STAT, BIT(0), BIT(0)); + MM_REG_WRITE(cmd, subsys_id, base, VIDO_SOFT_RST, 0x0, BIT(0)); + MM_REG_POLL(cmd, subsys_id, base, VIDO_SOFT_RST_STAT, 0x0, BIT(0)); + return 0; +} + +static int config_wrot_frame(struct mdp_comp_ctx *ctx, + struct mdp_cmdq_cmd *cmd, + const struct v4l2_rect *compose) +{ + const struct mdp_platform_config *mdp_cfg = __get_plat_cfg(ctx); + phys_addr_t base = ctx->comp->reg_base; + u8 subsys_id = ctx->comp->subsys_id; + u32 reg = 0; + + /* Write frame base address */ + if (CFG_CHECK(MT8183, p_id)) + reg = CFG_COMP(MT8183, ctx->param, wrot.iova[0]); + MM_REG_WRITE(cmd, subsys_id, base, VIDO_BASE_ADDR, reg, + 0xFFFFFFFF); + if (CFG_CHECK(MT8183, p_id)) + reg = CFG_COMP(MT8183, ctx->param, wrot.iova[1]); + MM_REG_WRITE(cmd, subsys_id, base, VIDO_BASE_ADDR_C, reg, + 0xFFFFFFFF); + if (CFG_CHECK(MT8183, p_id)) + reg = CFG_COMP(MT8183, ctx->param, wrot.iova[2]); + MM_REG_WRITE(cmd, subsys_id, base, VIDO_BASE_ADDR_V, reg, + 0xFFFFFFFF); + /* Write frame related registers */ + if (CFG_CHECK(MT8183, p_id)) + reg = CFG_COMP(MT8183, ctx->param, wrot.control); + MM_REG_WRITE(cmd, subsys_id, base, VIDO_CTRL, reg, + 0xF131510F); + /* Write frame Y pitch */ + if (CFG_CHECK(MT8183, p_id)) + reg = CFG_COMP(MT8183, ctx->param, wrot.stride[0]); + MM_REG_WRITE(cmd, subsys_id, base, VIDO_STRIDE, reg, + 0x0000FFFF); + /* Write frame UV pitch */ + if (CFG_CHECK(MT8183, p_id)) + reg = CFG_COMP(MT8183, ctx->param, wrot.stride[1]); + MM_REG_WRITE(cmd, subsys_id, base, VIDO_STRIDE_C, reg, + 0xFFFF); + if (CFG_CHECK(MT8183, p_id)) + reg = CFG_COMP(MT8183, ctx->param, wrot.stride[2]); + MM_REG_WRITE(cmd, subsys_id, base, VIDO_STRIDE_V, reg, + 0xFFFF); + /* Write matrix control */ + if (CFG_CHECK(MT8183, p_id)) + reg = CFG_COMP(MT8183, ctx->param, wrot.mat_ctrl); + MM_REG_WRITE(cmd, subsys_id, base, VIDO_MAT_CTRL, reg, 0xF3); + + /* Set the fixed ALPHA as 0xFF */ + MM_REG_WRITE(cmd, subsys_id, base, VIDO_DITHER, 0xFF000000, + 0xFF000000); + /* Set VIDO_EOL_SEL */ + MM_REG_WRITE(cmd, subsys_id, base, VIDO_RSV_1, BIT(31), BIT(31)); + /* Set VIDO_FIFO_TEST */ + if (CFG_CHECK(MT8183, p_id)) + reg = CFG_COMP(MT8183, ctx->param, wrot.fifo_test); + if (reg != 0) + MM_REG_WRITE(cmd, subsys_id, base, VIDO_FIFO_TEST, + reg, 0xFFF); + /* Filter enable */ + if (mdp_cfg && mdp_cfg->wrot_filter_constraint) { + if (CFG_CHECK(MT8183, p_id)) + reg = CFG_COMP(MT8183, ctx->param, wrot.filter); + MM_REG_WRITE(cmd, subsys_id, base, VIDO_MAIN_BUF_SIZE, + reg, 0x77); + } + + return 0; +} + +static int config_wrot_subfrm(struct mdp_comp_ctx *ctx, + struct mdp_cmdq_cmd *cmd, u32 index) +{ + phys_addr_t base = ctx->comp->reg_base; + u8 subsys_id = ctx->comp->subsys_id; + u32 reg = 0; + + /* Write Y pixel offset */ + if (CFG_CHECK(MT8183, p_id)) + reg = CFG_COMP(MT8183, ctx->param, wrot.subfrms[index].offset[0]); + MM_REG_WRITE(cmd, subsys_id, base, VIDO_OFST_ADDR, + reg, 0x0FFFFFFF); + /* Write U pixel offset */ + if (CFG_CHECK(MT8183, p_id)) + reg = CFG_COMP(MT8183, ctx->param, wrot.subfrms[index].offset[1]); + MM_REG_WRITE(cmd, subsys_id, base, VIDO_OFST_ADDR_C, + reg, 0x0FFFFFFF); + /* Write V pixel offset */ + if (CFG_CHECK(MT8183, p_id)) + reg = CFG_COMP(MT8183, ctx->param, wrot.subfrms[index].offset[2]); + MM_REG_WRITE(cmd, subsys_id, base, VIDO_OFST_ADDR_V, + reg, 0x0FFFFFFF); + /* Write source size */ + if (CFG_CHECK(MT8183, p_id)) + reg = CFG_COMP(MT8183, ctx->param, wrot.subfrms[index].src); + MM_REG_WRITE(cmd, subsys_id, base, VIDO_IN_SIZE, reg, + 0x1FFF1FFF); + /* Write target size */ + if (CFG_CHECK(MT8183, p_id)) + reg = CFG_COMP(MT8183, ctx->param, wrot.subfrms[index].clip); + MM_REG_WRITE(cmd, subsys_id, base, VIDO_TAR_SIZE, reg, + 0x1FFF1FFF); + if (CFG_CHECK(MT8183, p_id)) + reg = CFG_COMP(MT8183, ctx->param, wrot.subfrms[index].clip_ofst); + MM_REG_WRITE(cmd, subsys_id, base, VIDO_CROP_OFST, reg, + 0x1FFF1FFF); + + if (CFG_CHECK(MT8183, p_id)) + reg = CFG_COMP(MT8183, ctx->param, wrot.subfrms[index].main_buf); + MM_REG_WRITE(cmd, subsys_id, base, VIDO_MAIN_BUF_SIZE, + reg, 0x1FFF7F00); + + /* Enable WROT */ + MM_REG_WRITE(cmd, subsys_id, base, VIDO_ROT_EN, BIT(0), BIT(0)); + + return 0; +} + +static int wait_wrot_event(struct mdp_comp_ctx *ctx, struct mdp_cmdq_cmd *cmd) +{ + const struct mdp_platform_config *mdp_cfg = __get_plat_cfg(ctx); + struct device *dev = &ctx->comp->mdp_dev->pdev->dev; + phys_addr_t base = ctx->comp->reg_base; + u8 subsys_id = ctx->comp->subsys_id; + + if (ctx->comp->alias_id == 0) + MM_REG_WAIT(cmd, ctx->comp->gce_event[MDP_GCE_EVENT_EOF]); + else + dev_err(dev, "Do not support WROT1_DONE event\n"); + + if (mdp_cfg && mdp_cfg->wrot_filter_constraint) + MM_REG_WRITE(cmd, subsys_id, base, VIDO_MAIN_BUF_SIZE, 0x0, + 0x77); + + /* Disable WROT */ + MM_REG_WRITE(cmd, subsys_id, base, VIDO_ROT_EN, 0x0, BIT(0)); + + return 0; +} + +static const struct mdp_comp_ops wrot_ops = { + .get_comp_flag = get_comp_flag, + .init_comp = init_wrot, + .config_frame = config_wrot_frame, + .config_subfrm = config_wrot_subfrm, + .wait_comp_event = wait_wrot_event, +}; + +static int init_wdma(struct mdp_comp_ctx *ctx, struct mdp_cmdq_cmd *cmd) +{ + phys_addr_t base = ctx->comp->reg_base; + u8 subsys_id = ctx->comp->subsys_id; + + /* Reset WDMA */ + MM_REG_WRITE(cmd, subsys_id, base, WDMA_RST, BIT(0), BIT(0)); + MM_REG_POLL(cmd, subsys_id, base, WDMA_FLOW_CTRL_DBG, BIT(0), BIT(0)); + MM_REG_WRITE(cmd, subsys_id, base, WDMA_RST, 0x0, BIT(0)); + return 0; +} + +static int config_wdma_frame(struct mdp_comp_ctx *ctx, + struct mdp_cmdq_cmd *cmd, + const struct v4l2_rect *compose) +{ + phys_addr_t base = ctx->comp->reg_base; + u8 subsys_id = ctx->comp->subsys_id; + u32 reg = 0; + + MM_REG_WRITE(cmd, subsys_id, base, WDMA_BUF_CON2, 0x10101050, + 0xFFFFFFFF); + + /* Setup frame information */ + if (CFG_CHECK(MT8183, p_id)) + reg = CFG_COMP(MT8183, ctx->param, wdma.wdma_cfg); + MM_REG_WRITE(cmd, subsys_id, base, WDMA_CFG, reg, + 0x0F01B8F0); + /* Setup frame base address */ + if (CFG_CHECK(MT8183, p_id)) + reg = CFG_COMP(MT8183, ctx->param, wdma.iova[0]); + MM_REG_WRITE(cmd, subsys_id, base, WDMA_DST_ADDR, reg, + 0xFFFFFFFF); + if (CFG_CHECK(MT8183, p_id)) + reg = CFG_COMP(MT8183, ctx->param, wdma.iova[1]); + MM_REG_WRITE(cmd, subsys_id, base, WDMA_DST_U_ADDR, reg, + 0xFFFFFFFF); + if (CFG_CHECK(MT8183, p_id)) + reg = CFG_COMP(MT8183, ctx->param, wdma.iova[2]); + MM_REG_WRITE(cmd, subsys_id, base, WDMA_DST_V_ADDR, reg, + 0xFFFFFFFF); + /* Setup Y pitch */ + if (CFG_CHECK(MT8183, p_id)) + reg = CFG_COMP(MT8183, ctx->param, wdma.w_in_byte); + MM_REG_WRITE(cmd, subsys_id, base, WDMA_DST_W_IN_BYTE, + reg, 0x0000FFFF); + /* Setup UV pitch */ + if (CFG_CHECK(MT8183, p_id)) + reg = CFG_COMP(MT8183, ctx->param, wdma.uv_stride); + MM_REG_WRITE(cmd, subsys_id, base, WDMA_DST_UV_PITCH, + reg, 0x0000FFFF); + /* Set the fixed ALPHA as 0xFF */ + MM_REG_WRITE(cmd, subsys_id, base, WDMA_ALPHA, 0x800000FF, + 0x800000FF); + + return 0; +} + +static int config_wdma_subfrm(struct mdp_comp_ctx *ctx, + struct mdp_cmdq_cmd *cmd, u32 index) +{ + phys_addr_t base = ctx->comp->reg_base; + u8 subsys_id = ctx->comp->subsys_id; + u32 reg = 0; + + /* Write Y pixel offset */ + if (CFG_CHECK(MT8183, p_id)) + reg = CFG_COMP(MT8183, ctx->param, wdma.subfrms[index].offset[0]); + MM_REG_WRITE(cmd, subsys_id, base, WDMA_DST_ADDR_OFFSET, + reg, 0x0FFFFFFF); + /* Write U pixel offset */ + if (CFG_CHECK(MT8183, p_id)) + reg = CFG_COMP(MT8183, ctx->param, wdma.subfrms[index].offset[1]); + MM_REG_WRITE(cmd, subsys_id, base, WDMA_DST_U_ADDR_OFFSET, + reg, 0x0FFFFFFF); + /* Write V pixel offset */ + if (CFG_CHECK(MT8183, p_id)) + reg = CFG_COMP(MT8183, ctx->param, wdma.subfrms[index].offset[2]); + MM_REG_WRITE(cmd, subsys_id, base, WDMA_DST_V_ADDR_OFFSET, + reg, 0x0FFFFFFF); + /* Write source size */ + if (CFG_CHECK(MT8183, p_id)) + reg = CFG_COMP(MT8183, ctx->param, wdma.subfrms[index].src); + MM_REG_WRITE(cmd, subsys_id, base, WDMA_SRC_SIZE, reg, + 0x3FFF3FFF); + /* Write target size */ + if (CFG_CHECK(MT8183, p_id)) + reg = CFG_COMP(MT8183, ctx->param, wdma.subfrms[index].clip); + MM_REG_WRITE(cmd, subsys_id, base, WDMA_CLIP_SIZE, reg, + 0x3FFF3FFF); + /* Write clip offset */ + if (CFG_CHECK(MT8183, p_id)) + reg = CFG_COMP(MT8183, ctx->param, wdma.subfrms[index].clip_ofst); + MM_REG_WRITE(cmd, subsys_id, base, WDMA_CLIP_COORD, reg, + 0x3FFF3FFF); + + /* Enable WDMA */ + MM_REG_WRITE(cmd, subsys_id, base, WDMA_EN, BIT(0), BIT(0)); + + return 0; +} + +static int wait_wdma_event(struct mdp_comp_ctx *ctx, struct mdp_cmdq_cmd *cmd) +{ + phys_addr_t base = ctx->comp->reg_base; + u8 subsys_id = ctx->comp->subsys_id; + + MM_REG_WAIT(cmd, ctx->comp->gce_event[MDP_GCE_EVENT_EOF]); + /* Disable WDMA */ + MM_REG_WRITE(cmd, subsys_id, base, WDMA_EN, 0x0, BIT(0)); + return 0; +} + +static const struct mdp_comp_ops wdma_ops = { + .get_comp_flag = get_comp_flag, + .init_comp = init_wdma, + .config_frame = config_wdma_frame, + .config_subfrm = config_wdma_subfrm, + .wait_comp_event = wait_wdma_event, +}; + +static int init_ccorr(struct mdp_comp_ctx *ctx, struct mdp_cmdq_cmd *cmd) +{ + phys_addr_t base = ctx->comp->reg_base; + u8 subsys_id = ctx->comp->subsys_id; + + /* CCORR enable */ + MM_REG_WRITE(cmd, subsys_id, base, MDP_CCORR_EN, BIT(0), BIT(0)); + /* Relay mode */ + MM_REG_WRITE(cmd, subsys_id, base, MDP_CCORR_CFG, BIT(0), BIT(0)); + return 0; +} + +static int config_ccorr_subfrm(struct mdp_comp_ctx *ctx, + struct mdp_cmdq_cmd *cmd, u32 index) +{ + phys_addr_t base = ctx->comp->reg_base; + u8 subsys_id = ctx->comp->subsys_id; + u32 csf_l = 0, csf_r = 0; + u32 csf_t = 0, csf_b = 0; + u32 hsize, vsize; + + if (CFG_CHECK(MT8183, p_id)) { + csf_l = CFG_COMP(MT8183, ctx->param, subfrms[index].in.left); + csf_r = CFG_COMP(MT8183, ctx->param, subfrms[index].in.right); + csf_t = CFG_COMP(MT8183, ctx->param, subfrms[index].in.top); + csf_b = CFG_COMP(MT8183, ctx->param, subfrms[index].in.bottom); + } + + hsize = csf_r - csf_l + 1; + vsize = csf_b - csf_t + 1; + MM_REG_WRITE(cmd, subsys_id, base, MDP_CCORR_SIZE, + (hsize << 16) + (vsize << 0), 0x1FFF1FFF); + return 0; +} + +static const struct mdp_comp_ops ccorr_ops = { + .get_comp_flag = get_comp_flag, + .init_comp = init_ccorr, + .config_subfrm = config_ccorr_subfrm, +}; + +static const struct mdp_comp_ops *mdp_comp_ops[MDP_COMP_TYPE_COUNT] = { + [MDP_COMP_TYPE_RDMA] = &rdma_ops, + [MDP_COMP_TYPE_RSZ] = &rsz_ops, + [MDP_COMP_TYPE_WROT] = &wrot_ops, + [MDP_COMP_TYPE_WDMA] = &wdma_ops, + [MDP_COMP_TYPE_CCORR] = &ccorr_ops, +}; + +static const struct of_device_id mdp_comp_dt_ids[] __maybe_unused = { + { + .compatible = "mediatek,mt8183-mdp3-rdma", + .data = (void *)MDP_COMP_TYPE_RDMA, + }, { + .compatible = "mediatek,mt8183-mdp3-ccorr", + .data = (void *)MDP_COMP_TYPE_CCORR, + }, { + .compatible = "mediatek,mt8183-mdp3-rsz", + .data = (void *)MDP_COMP_TYPE_RSZ, + }, { + .compatible = "mediatek,mt8183-mdp3-wrot", + .data = (void *)MDP_COMP_TYPE_WROT, + }, { + .compatible = "mediatek,mt8183-mdp3-wdma", + .data = (void *)MDP_COMP_TYPE_WDMA, + }, + {} +}; + +static inline bool is_dma_capable(const enum mdp_comp_type type) +{ + return (type == MDP_COMP_TYPE_RDMA || + type == MDP_COMP_TYPE_WROT || + type == MDP_COMP_TYPE_WDMA); +} + +static inline bool is_bypass_gce_event(const enum mdp_comp_type type) +{ + /* + * Subcomponent PATH is only used for the direction of data flow and + * dose not need to wait for GCE event. + */ + return (type == MDP_COMP_TYPE_PATH); +} + +static int mdp_comp_get_id(struct mdp_dev *mdp, enum mdp_comp_type type, u32 alias_id) +{ + int i; + + for (i = 0; i < mdp->mdp_data->comp_data_len; i++) + if (mdp->mdp_data->comp_data[i].match.type == type && + mdp->mdp_data->comp_data[i].match.alias_id == alias_id) + return i; + return -ENODEV; +} + +int mdp_comp_clock_on(struct device *dev, struct mdp_comp *comp) +{ + int i, ret; + + /* Only DMA capable components need the pm control */ + if (comp->comp_dev && is_dma_capable(comp->type)) { + ret = pm_runtime_resume_and_get(comp->comp_dev); + if (ret < 0) { + dev_err(dev, + "Failed to get power, err %d. type:%d id:%d\n", + ret, comp->type, comp->inner_id); + return ret; + } + } + + for (i = 0; i < comp->clk_num; i++) { + if (IS_ERR_OR_NULL(comp->clks[i])) + continue; + ret = clk_prepare_enable(comp->clks[i]); + if (ret) { + dev_err(dev, + "Failed to enable clk %d. type:%d id:%d\n", + i, comp->type, comp->inner_id); + goto err_revert; + } + } + + return 0; + +err_revert: + while (--i >= 0) { + if (IS_ERR_OR_NULL(comp->clks[i])) + continue; + clk_disable_unprepare(comp->clks[i]); + } + if (comp->comp_dev && is_dma_capable(comp->type)) + pm_runtime_put_sync(comp->comp_dev); + + return ret; +} + +void mdp_comp_clock_off(struct device *dev, struct mdp_comp *comp) +{ + int i; + + for (i = 0; i < comp->clk_num; i++) { + if (IS_ERR_OR_NULL(comp->clks[i])) + continue; + clk_disable_unprepare(comp->clks[i]); + } + + if (comp->comp_dev && is_dma_capable(comp->type)) + pm_runtime_put(comp->comp_dev); +} + +int mdp_comp_clocks_on(struct device *dev, struct mdp_comp *comps, int num) +{ + int i, ret; + + for (i = 0; i < num; i++) { + ret = mdp_comp_clock_on(dev, &comps[i]); + if (ret) + return ret; + } + + return 0; +} + +void mdp_comp_clocks_off(struct device *dev, struct mdp_comp *comps, int num) +{ + int i; + + for (i = 0; i < num; i++) + mdp_comp_clock_off(dev, &comps[i]); +} + +static int mdp_get_subsys_id(struct mdp_dev *mdp, struct device *dev, + struct device_node *node, struct mdp_comp *comp) +{ + struct platform_device *comp_pdev; + struct cmdq_client_reg cmdq_reg; + int ret = 0; + int index = 0; + + if (!dev || !node || !comp) + return -EINVAL; + + comp_pdev = of_find_device_by_node(node); + + if (!comp_pdev) { + dev_err(dev, "get comp_pdev fail! comp public id=%d, inner id=%d, type=%d\n", + comp->public_id, comp->inner_id, comp->type); + return -ENODEV; + } + + index = mdp->mdp_data->comp_data[comp->public_id].info.dts_reg_ofst; + ret = cmdq_dev_get_client_reg(&comp_pdev->dev, &cmdq_reg, index); + if (ret != 0) { + dev_err(&comp_pdev->dev, "cmdq_dev_get_subsys fail!\n"); + put_device(&comp_pdev->dev); + return -EINVAL; + } + + comp->subsys_id = cmdq_reg.subsys; + dev_dbg(&comp_pdev->dev, "subsys id=%d\n", cmdq_reg.subsys); + put_device(&comp_pdev->dev); + + return 0; +} + +static void __mdp_comp_init(struct mdp_dev *mdp, struct device_node *node, + struct mdp_comp *comp) +{ + struct resource res; + phys_addr_t base; + int index; + + index = mdp->mdp_data->comp_data[comp->public_id].info.dts_reg_ofst; + if (of_address_to_resource(node, index, &res) < 0) + base = 0L; + else + base = res.start; + + comp->mdp_dev = mdp; + comp->regs = of_iomap(node, 0); + comp->reg_base = base; +} + +static int mdp_comp_init(struct mdp_dev *mdp, struct device_node *node, + struct mdp_comp *comp, enum mtk_mdp_comp_id id) +{ + struct device *dev = &mdp->pdev->dev; + struct platform_device *pdev_c; + int clk_ofst; + int i; + s32 event; + + if (id < 0 || id >= MDP_MAX_COMP_COUNT) { + dev_err(dev, "Invalid component id %d\n", id); + return -EINVAL; + } + + pdev_c = of_find_device_by_node(node); + if (!pdev_c) { + dev_warn(dev, "can't find platform device of node:%s\n", + node->name); + return -ENODEV; + } + + comp->comp_dev = &pdev_c->dev; + comp->public_id = id; + comp->type = mdp->mdp_data->comp_data[id].match.type; + comp->inner_id = mdp->mdp_data->comp_data[id].match.inner_id; + comp->alias_id = mdp->mdp_data->comp_data[id].match.alias_id; + comp->ops = mdp_comp_ops[comp->type]; + __mdp_comp_init(mdp, node, comp); + + comp->clk_num = mdp->mdp_data->comp_data[id].info.clk_num; + comp->clks = devm_kzalloc(dev, sizeof(struct clk *) * comp->clk_num, + GFP_KERNEL); + if (!comp->clks) + return -ENOMEM; + + clk_ofst = mdp->mdp_data->comp_data[id].info.clk_ofst; + + for (i = 0; i < comp->clk_num; i++) { + comp->clks[i] = of_clk_get(node, i + clk_ofst); + if (IS_ERR(comp->clks[i])) + break; + } + + mdp_get_subsys_id(mdp, dev, node, comp); + + /* Set GCE SOF event */ + if (is_bypass_gce_event(comp->type) || + of_property_read_u32_index(node, "mediatek,gce-events", + MDP_GCE_EVENT_SOF, &event)) + event = MDP_GCE_NO_EVENT; + + comp->gce_event[MDP_GCE_EVENT_SOF] = event; + + /* Set GCE EOF event */ + if (is_dma_capable(comp->type)) { + if (of_property_read_u32_index(node, "mediatek,gce-events", + MDP_GCE_EVENT_EOF, &event)) { + dev_err(dev, "Component id %d has no EOF\n", id); + return -EINVAL; + } + } else { + event = MDP_GCE_NO_EVENT; + } + + comp->gce_event[MDP_GCE_EVENT_EOF] = event; + + return 0; +} + +static void mdp_comp_deinit(struct mdp_comp *comp) +{ + if (!comp) + return; + + if (comp->comp_dev && comp->clks) { + devm_kfree(&comp->mdp_dev->pdev->dev, comp->clks); + comp->clks = NULL; + } + + if (comp->regs) + iounmap(comp->regs); +} + +static struct mdp_comp *mdp_comp_create(struct mdp_dev *mdp, + struct device_node *node, + enum mtk_mdp_comp_id id) +{ + struct device *dev = &mdp->pdev->dev; + struct mdp_comp *comp; + int ret; + + if (mdp->comp[id]) + return ERR_PTR(-EEXIST); + + comp = devm_kzalloc(dev, sizeof(*comp), GFP_KERNEL); + if (!comp) + return ERR_PTR(-ENOMEM); + + ret = mdp_comp_init(mdp, node, comp, id); + if (ret) { + devm_kfree(dev, comp); + return ERR_PTR(ret); + } + mdp->comp[id] = comp; + mdp->comp[id]->mdp_dev = mdp; + + dev_dbg(dev, "%s type:%d alias:%d public id:%d inner id:%d base:%#x regs:%p\n", + dev->of_node->name, comp->type, comp->alias_id, id, comp->inner_id, + (u32)comp->reg_base, comp->regs); + return comp; +} + +static int mdp_comp_sub_create(struct mdp_dev *mdp) +{ + struct device *dev = &mdp->pdev->dev; + struct device_node *node, *parent; + int ret = 0; + + parent = dev->of_node->parent; + + for_each_child_of_node(parent, node) { + const struct of_device_id *of_id; + enum mdp_comp_type type; + int id, alias_id; + struct mdp_comp *comp; + + of_id = of_match_node(mdp->mdp_data->mdp_sub_comp_dt_ids, node); + if (!of_id) + continue; + if (!of_device_is_available(node)) { + dev_dbg(dev, "Skipping disabled sub comp. %pOF\n", + node); + continue; + } + + type = (enum mdp_comp_type)(uintptr_t)of_id->data; + alias_id = mdp_comp_alias_id[type]; + id = mdp_comp_get_id(mdp, type, alias_id); + if (id < 0) { + dev_err(dev, + "Fail to get sub comp. id: type %d alias %d\n", + type, alias_id); + ret = -EINVAL; + goto err_free_node; + } + mdp_comp_alias_id[type]++; + + comp = mdp_comp_create(mdp, node, id); + if (IS_ERR(comp)) { + ret = PTR_ERR(comp); + goto err_free_node; + } + } + return ret; + +err_free_node: + of_node_put(node); + return ret; +} + +void mdp_comp_destroy(struct mdp_dev *mdp) +{ + int i; + + for (i = 0; i < ARRAY_SIZE(mdp->comp); i++) { + if (mdp->comp[i]) { + if (is_dma_capable(mdp->comp[i]->type)) + pm_runtime_disable(mdp->comp[i]->comp_dev); + mdp_comp_deinit(mdp->comp[i]); + devm_kfree(mdp->comp[i]->comp_dev, mdp->comp[i]); + mdp->comp[i] = NULL; + } + } +} + +int mdp_comp_config(struct mdp_dev *mdp) +{ + struct device *dev = &mdp->pdev->dev; + struct device_node *node, *parent; + int ret; + + memset(mdp_comp_alias_id, 0, sizeof(mdp_comp_alias_id)); + p_id = mdp->mdp_data->mdp_plat_id; + + parent = dev->of_node->parent; + /* Iterate over sibling MDP function blocks */ + for_each_child_of_node(parent, node) { + const struct of_device_id *of_id; + enum mdp_comp_type type; + int id, alias_id; + struct mdp_comp *comp; + + of_id = of_match_node(mdp_comp_dt_ids, node); + if (!of_id) + continue; + + if (!of_device_is_available(node)) { + dev_dbg(dev, "Skipping disabled component %pOF\n", + node); + continue; + } + + type = (enum mdp_comp_type)(uintptr_t)of_id->data; + alias_id = mdp_comp_alias_id[type]; + id = mdp_comp_get_id(mdp, type, alias_id); + if (id < 0) { + dev_err(dev, + "Fail to get component id: type %d alias %d\n", + type, alias_id); + continue; + } + mdp_comp_alias_id[type]++; + + comp = mdp_comp_create(mdp, node, id); + if (IS_ERR(comp)) { + ret = PTR_ERR(comp); + goto err_init_comps; + } + + /* Only DMA capable components need the pm control */ + if (!is_dma_capable(comp->type)) + continue; + pm_runtime_enable(comp->comp_dev); + } + + ret = mdp_comp_sub_create(mdp); + if (ret) + goto err_init_comps; + + return 0; + +err_init_comps: + mdp_comp_destroy(mdp); + return ret; +} + +int mdp_comp_ctx_config(struct mdp_dev *mdp, struct mdp_comp_ctx *ctx, + const struct img_compparam *param, + const struct img_ipi_frameparam *frame) +{ + struct device *dev = &mdp->pdev->dev; + enum mtk_mdp_comp_id public_id = MDP_COMP_NONE; + u32 arg; + int i, idx; + + if (!param) { + dev_err(dev, "Invalid component param"); + return -EINVAL; + } + + if (CFG_CHECK(MT8183, p_id)) + arg = CFG_COMP(MT8183, param, type); + else + return -EINVAL; + public_id = mdp_cfg_get_id_public(mdp, arg); + if (public_id < 0) { + dev_err(dev, "Invalid component id %d", public_id); + return -EINVAL; + } + + ctx->comp = mdp->comp[public_id]; + if (!ctx->comp) { + dev_err(dev, "Uninit component inner id %d", arg); + return -EINVAL; + } + + ctx->param = param; + if (CFG_CHECK(MT8183, p_id)) + arg = CFG_COMP(MT8183, param, input); + else + return -EINVAL; + ctx->input = &frame->inputs[arg]; + if (CFG_CHECK(MT8183, p_id)) + idx = CFG_COMP(MT8183, param, num_outputs); + else + return -EINVAL; + for (i = 0; i < idx; i++) { + if (CFG_CHECK(MT8183, p_id)) + arg = CFG_COMP(MT8183, param, outputs[i]); + else + return -EINVAL; + ctx->outputs[i] = &frame->outputs[arg]; + } + return 0; +} diff --git a/drivers/media/platform/mediatek/mdp3/mtk-mdp3-comp.h b/drivers/media/platform/mediatek/mdp3/mtk-mdp3-comp.h new file mode 100644 index 0000000000..20d2bcb77e --- /dev/null +++ b/drivers/media/platform/mediatek/mdp3/mtk-mdp3-comp.h @@ -0,0 +1,206 @@ +/* SPDX-License-Identifier: GPL-2.0-only */ +/* + * Copyright (c) 2022 MediaTek Inc. + * Author: Ping-Hsun Wu <ping-hsun.wu@mediatek.com> + */ + +#ifndef __MTK_MDP3_COMP_H__ +#define __MTK_MDP3_COMP_H__ + +#include "mtk-mdp3-cmdq.h" + +#define MM_REG_WRITE_MASK(cmd, id, base, ofst, val, mask, ...) \ + cmdq_pkt_write_mask(&((cmd)->pkt), id, \ + (base) + (ofst), (val), (mask), ##__VA_ARGS__) + +#define MM_REG_WRITE(cmd, id, base, ofst, val, mask, ...) \ +do { \ + typeof(mask) (m) = (mask); \ + MM_REG_WRITE_MASK(cmd, id, base, ofst, val, \ + (((m) & (ofst##_MASK)) == (ofst##_MASK)) ? \ + (0xffffffff) : (m), ##__VA_ARGS__); \ +} while (0) + +#define MM_REG_WAIT(cmd, evt) \ +do { \ + typeof(cmd) (c) = (cmd); \ + typeof(evt) (e) = (evt); \ + cmdq_pkt_wfe(&((c)->pkt), (e), true); \ +} while (0) + +#define MM_REG_WAIT_NO_CLEAR(cmd, evt) \ +do { \ + typeof(cmd) (c) = (cmd); \ + typeof(evt) (e) = (evt); \ + cmdq_pkt_wfe(&((c)->pkt), (e), false); \ +} while (0) + +#define MM_REG_CLEAR(cmd, evt) \ +do { \ + typeof(cmd) (c) = (cmd); \ + typeof(evt) (e) = (evt); \ + cmdq_pkt_clear_event(&((c)->pkt), (e)); \ +} while (0) + +#define MM_REG_SET_EVENT(cmd, evt) \ +do { \ + typeof(cmd) (c) = (cmd); \ + typeof(evt) (e) = (evt); \ + cmdq_pkt_set_event(&((c)->pkt), (e)); \ +} while (0) + +#define MM_REG_POLL_MASK(cmd, id, base, ofst, val, _mask, ...) \ +do { \ + typeof(_mask) (_m) = (_mask); \ + cmdq_pkt_poll_mask(&((cmd)->pkt), id, \ + (base) + (ofst), (val), (_m), ##__VA_ARGS__); \ +} while (0) + +#define MM_REG_POLL(cmd, id, base, ofst, val, mask, ...) \ +do { \ + typeof(mask) (m) = (mask); \ + MM_REG_POLL_MASK((cmd), id, base, ofst, val, \ + (((m) & (ofst##_MASK)) == (ofst##_MASK)) ? \ + (0xffffffff) : (m), ##__VA_ARGS__); \ +} while (0) + +enum mtk_mdp_comp_id { + MDP_COMP_NONE = -1, /* Invalid engine */ + + /* ISP */ + MDP_COMP_WPEI = 0, + MDP_COMP_WPEO, /* 1 */ + MDP_COMP_WPEI2, /* 2 */ + MDP_COMP_WPEO2, /* 3 */ + MDP_COMP_ISP_IMGI, /* 4 */ + MDP_COMP_ISP_IMGO, /* 5 */ + MDP_COMP_ISP_IMG2O, /* 6 */ + + /* IPU */ + MDP_COMP_IPUI, /* 7 */ + MDP_COMP_IPUO, /* 8 */ + + /* MDP */ + MDP_COMP_CAMIN, /* 9 */ + MDP_COMP_CAMIN2, /* 10 */ + MDP_COMP_RDMA0, /* 11 */ + MDP_COMP_AAL0, /* 12 */ + MDP_COMP_CCORR0, /* 13 */ + MDP_COMP_RSZ0, /* 14 */ + MDP_COMP_RSZ1, /* 15 */ + MDP_COMP_TDSHP0, /* 16 */ + MDP_COMP_COLOR0, /* 17 */ + MDP_COMP_PATH0_SOUT, /* 18 */ + MDP_COMP_PATH1_SOUT, /* 19 */ + MDP_COMP_WROT0, /* 20 */ + MDP_COMP_WDMA, /* 21 */ + + /* Dummy Engine */ + MDP_COMP_RDMA1, /* 22 */ + MDP_COMP_RSZ2, /* 23 */ + MDP_COMP_TDSHP1, /* 24 */ + MDP_COMP_WROT1, /* 25 */ + + MDP_MAX_COMP_COUNT /* ALWAYS keep at the end */ +}; + +enum mdp_comp_type { + MDP_COMP_TYPE_INVALID = 0, + + MDP_COMP_TYPE_RDMA, + MDP_COMP_TYPE_RSZ, + MDP_COMP_TYPE_WROT, + MDP_COMP_TYPE_WDMA, + MDP_COMP_TYPE_PATH, + + MDP_COMP_TYPE_TDSHP, + MDP_COMP_TYPE_COLOR, + MDP_COMP_TYPE_DRE, + MDP_COMP_TYPE_CCORR, + MDP_COMP_TYPE_HDR, + + MDP_COMP_TYPE_IMGI, + MDP_COMP_TYPE_WPEI, + MDP_COMP_TYPE_EXTO, /* External path */ + MDP_COMP_TYPE_DL_PATH, /* Direct-link path */ + + MDP_COMP_TYPE_COUNT /* ALWAYS keep at the end */ +}; + +#define MDP_GCE_NO_EVENT (-1) +enum { + MDP_GCE_EVENT_SOF = 0, + MDP_GCE_EVENT_EOF = 1, + MDP_GCE_EVENT_MAX, +}; + +struct mdp_comp_match { + enum mdp_comp_type type; + u32 alias_id; + s32 inner_id; +}; + +/* Used to describe the item order in MDP property */ +struct mdp_comp_info { + u32 clk_num; + u32 clk_ofst; + u32 dts_reg_ofst; +}; + +struct mdp_comp_data { + struct mdp_comp_match match; + struct mdp_comp_info info; +}; + +struct mdp_comp_ops; + +struct mdp_comp { + struct mdp_dev *mdp_dev; + void __iomem *regs; + phys_addr_t reg_base; + u8 subsys_id; + u8 clk_num; + struct clk **clks; + struct device *comp_dev; + enum mdp_comp_type type; + enum mtk_mdp_comp_id public_id; + s32 inner_id; + u32 alias_id; + s32 gce_event[MDP_GCE_EVENT_MAX]; + const struct mdp_comp_ops *ops; +}; + +struct mdp_comp_ctx { + struct mdp_comp *comp; + const struct img_compparam *param; + const struct img_input *input; + const struct img_output *outputs[IMG_MAX_HW_OUTPUTS]; +}; + +struct mdp_comp_ops { + s64 (*get_comp_flag)(const struct mdp_comp_ctx *ctx); + int (*init_comp)(struct mdp_comp_ctx *ctx, struct mdp_cmdq_cmd *cmd); + int (*config_frame)(struct mdp_comp_ctx *ctx, struct mdp_cmdq_cmd *cmd, + const struct v4l2_rect *compose); + int (*config_subfrm)(struct mdp_comp_ctx *ctx, + struct mdp_cmdq_cmd *cmd, u32 index); + int (*wait_comp_event)(struct mdp_comp_ctx *ctx, + struct mdp_cmdq_cmd *cmd); + int (*advance_subfrm)(struct mdp_comp_ctx *ctx, + struct mdp_cmdq_cmd *cmd, u32 index); + int (*post_process)(struct mdp_comp_ctx *ctx, struct mdp_cmdq_cmd *cmd); +}; + +struct mdp_dev; + +int mdp_comp_config(struct mdp_dev *mdp); +void mdp_comp_destroy(struct mdp_dev *mdp); +int mdp_comp_clock_on(struct device *dev, struct mdp_comp *comp); +void mdp_comp_clock_off(struct device *dev, struct mdp_comp *comp); +int mdp_comp_clocks_on(struct device *dev, struct mdp_comp *comps, int num); +void mdp_comp_clocks_off(struct device *dev, struct mdp_comp *comps, int num); +int mdp_comp_ctx_config(struct mdp_dev *mdp, struct mdp_comp_ctx *ctx, + const struct img_compparam *param, + const struct img_ipi_frameparam *frame); + +#endif /* __MTK_MDP3_COMP_H__ */ diff --git a/drivers/media/platform/mediatek/mdp3/mtk-mdp3-core.c b/drivers/media/platform/mediatek/mdp3/mtk-mdp3-core.c new file mode 100644 index 0000000000..cc44be10fd --- /dev/null +++ b/drivers/media/platform/mediatek/mdp3/mtk-mdp3-core.c @@ -0,0 +1,333 @@ +// SPDX-License-Identifier: GPL-2.0-only +/* + * Copyright (c) 2022 MediaTek Inc. + * Author: Ping-Hsun Wu <ping-hsun.wu@mediatek.com> + */ + +#include <linux/clk.h> +#include <linux/module.h> +#include <linux/of_platform.h> +#include <linux/platform_device.h> +#include <linux/pm_runtime.h> +#include <linux/remoteproc.h> +#include <linux/remoteproc/mtk_scp.h> +#include <media/videobuf2-dma-contig.h> + +#include "mtk-mdp3-core.h" +#include "mtk-mdp3-cfg.h" +#include "mtk-mdp3-m2m.h" + +static const struct of_device_id mdp_of_ids[] = { + { .compatible = "mediatek,mt8183-mdp3-rdma", + .data = &mt8183_mdp_driver_data, + }, + {}, +}; +MODULE_DEVICE_TABLE(of, mdp_of_ids); + +static struct platform_device *__get_pdev_by_id(struct platform_device *pdev, + enum mdp_infra_id id) +{ + struct device_node *node; + struct platform_device *mdp_pdev = NULL; + const struct mtk_mdp_driver_data *mdp_data; + const char *compat; + + if (!pdev) + return NULL; + + if (id < MDP_INFRA_MMSYS || id >= MDP_INFRA_MAX) { + dev_err(&pdev->dev, "Illegal infra id %d\n", id); + return NULL; + } + + mdp_data = of_device_get_match_data(&pdev->dev); + if (!mdp_data) { + dev_err(&pdev->dev, "have no driver data to find node\n"); + return NULL; + } + compat = mdp_data->mdp_probe_infra[id].compatible; + + node = of_find_compatible_node(NULL, NULL, compat); + if (WARN_ON(!node)) { + dev_err(&pdev->dev, "find node from id %d failed\n", id); + return NULL; + } + + mdp_pdev = of_find_device_by_node(node); + of_node_put(node); + if (WARN_ON(!mdp_pdev)) { + dev_err(&pdev->dev, "find pdev from id %d failed\n", id); + return NULL; + } + + return mdp_pdev; +} + +struct platform_device *mdp_get_plat_device(struct platform_device *pdev) +{ + struct device *dev = &pdev->dev; + struct device_node *mdp_node; + struct platform_device *mdp_pdev; + + mdp_node = of_parse_phandle(dev->of_node, MDP_PHANDLE_NAME, 0); + if (!mdp_node) { + dev_err(dev, "can't get node %s\n", MDP_PHANDLE_NAME); + return NULL; + } + + mdp_pdev = of_find_device_by_node(mdp_node); + of_node_put(mdp_node); + + return mdp_pdev; +} +EXPORT_SYMBOL_GPL(mdp_get_plat_device); + +int mdp_vpu_get_locked(struct mdp_dev *mdp) +{ + int ret = 0; + + if (mdp->vpu_count++ == 0) { + ret = rproc_boot(mdp->rproc_handle); + if (ret) { + dev_err(&mdp->pdev->dev, + "vpu_load_firmware failed %d\n", ret); + goto err_load_vpu; + } + ret = mdp_vpu_register(mdp); + if (ret) { + dev_err(&mdp->pdev->dev, + "mdp_vpu register failed %d\n", ret); + goto err_reg_vpu; + } + ret = mdp_vpu_dev_init(&mdp->vpu, mdp->scp, &mdp->vpu_lock); + if (ret) { + dev_err(&mdp->pdev->dev, + "mdp_vpu device init failed %d\n", ret); + goto err_init_vpu; + } + } + return 0; + +err_init_vpu: + mdp_vpu_unregister(mdp); +err_reg_vpu: +err_load_vpu: + mdp->vpu_count--; + return ret; +} + +void mdp_vpu_put_locked(struct mdp_dev *mdp) +{ + if (--mdp->vpu_count == 0) { + mdp_vpu_dev_deinit(&mdp->vpu); + mdp_vpu_unregister(mdp); + } +} + +void mdp_video_device_release(struct video_device *vdev) +{ + struct mdp_dev *mdp = (struct mdp_dev *)video_get_drvdata(vdev); + int i; + + scp_put(mdp->scp); + + destroy_workqueue(mdp->job_wq); + destroy_workqueue(mdp->clock_wq); + + pm_runtime_disable(&mdp->pdev->dev); + + vb2_dma_contig_clear_max_seg_size(&mdp->pdev->dev); + + mdp_comp_destroy(mdp); + for (i = 0; i < MDP_PIPE_MAX; i++) + mtk_mutex_put(mdp->mdp_mutex[i]); + + mdp_vpu_shared_mem_free(&mdp->vpu); + v4l2_m2m_release(mdp->m2m_dev); + kfree(mdp); +} + +static int mdp_probe(struct platform_device *pdev) +{ + struct device *dev = &pdev->dev; + struct mdp_dev *mdp; + struct platform_device *mm_pdev; + int ret, i, mutex_id; + + mdp = kzalloc(sizeof(*mdp), GFP_KERNEL); + if (!mdp) { + ret = -ENOMEM; + goto err_return; + } + + mdp->pdev = pdev; + mdp->mdp_data = of_device_get_match_data(&pdev->dev); + + mm_pdev = __get_pdev_by_id(pdev, MDP_INFRA_MMSYS); + if (!mm_pdev) { + ret = -ENODEV; + goto err_destroy_device; + } + mdp->mdp_mmsys = &mm_pdev->dev; + + mm_pdev = __get_pdev_by_id(pdev, MDP_INFRA_MUTEX); + if (WARN_ON(!mm_pdev)) { + ret = -ENODEV; + goto err_destroy_device; + } + for (i = 0; i < mdp->mdp_data->pipe_info_len; i++) { + mutex_id = mdp->mdp_data->pipe_info[i].mutex_id; + if (!IS_ERR_OR_NULL(mdp->mdp_mutex[mutex_id])) + continue; + mdp->mdp_mutex[mutex_id] = mtk_mutex_get(&mm_pdev->dev); + if (IS_ERR(mdp->mdp_mutex[mutex_id])) { + ret = PTR_ERR(mdp->mdp_mutex[mutex_id]); + goto err_free_mutex; + } + } + + ret = mdp_comp_config(mdp); + if (ret) { + dev_err(dev, "Failed to config mdp components\n"); + goto err_free_mutex; + } + + mdp->job_wq = alloc_workqueue(MDP_MODULE_NAME, WQ_FREEZABLE, 0); + if (!mdp->job_wq) { + dev_err(dev, "Unable to create job workqueue\n"); + ret = -ENOMEM; + goto err_deinit_comp; + } + + mdp->clock_wq = alloc_workqueue(MDP_MODULE_NAME "-clock", WQ_FREEZABLE, + 0); + if (!mdp->clock_wq) { + dev_err(dev, "Unable to create clock workqueue\n"); + ret = -ENOMEM; + goto err_destroy_job_wq; + } + + mm_pdev = __get_pdev_by_id(pdev, MDP_INFRA_SCP); + if (WARN_ON(!mm_pdev)) { + dev_err(&pdev->dev, "Could not get scp device\n"); + ret = -ENODEV; + goto err_destroy_clock_wq; + } + mdp->scp = platform_get_drvdata(mm_pdev); + mdp->rproc_handle = scp_get_rproc(mdp->scp); + dev_dbg(&pdev->dev, "MDP rproc_handle: %pK", mdp->rproc_handle); + + mutex_init(&mdp->vpu_lock); + mutex_init(&mdp->m2m_lock); + + mdp->cmdq_clt = cmdq_mbox_create(dev, 0); + if (IS_ERR(mdp->cmdq_clt)) { + ret = PTR_ERR(mdp->cmdq_clt); + goto err_put_scp; + } + + init_waitqueue_head(&mdp->callback_wq); + ida_init(&mdp->mdp_ida); + platform_set_drvdata(pdev, mdp); + + vb2_dma_contig_set_max_seg_size(&pdev->dev, DMA_BIT_MASK(32)); + + ret = v4l2_device_register(dev, &mdp->v4l2_dev); + if (ret) { + dev_err(dev, "Failed to register v4l2 device\n"); + ret = -EINVAL; + goto err_mbox_destroy; + } + + ret = mdp_m2m_device_register(mdp); + if (ret) { + v4l2_err(&mdp->v4l2_dev, "Failed to register m2m device\n"); + goto err_unregister_device; + } + + dev_dbg(dev, "mdp-%d registered successfully\n", pdev->id); + return 0; + +err_unregister_device: + v4l2_device_unregister(&mdp->v4l2_dev); +err_mbox_destroy: + cmdq_mbox_destroy(mdp->cmdq_clt); +err_put_scp: + scp_put(mdp->scp); +err_destroy_clock_wq: + destroy_workqueue(mdp->clock_wq); +err_destroy_job_wq: + destroy_workqueue(mdp->job_wq); +err_deinit_comp: + mdp_comp_destroy(mdp); +err_free_mutex: + for (i = 0; i < mdp->mdp_data->pipe_info_len; i++) + if (!IS_ERR_OR_NULL(mdp->mdp_mutex[i])) + mtk_mutex_put(mdp->mdp_mutex[i]); +err_destroy_device: + kfree(mdp); +err_return: + dev_dbg(dev, "Errno %d\n", ret); + return ret; +} + +static void mdp_remove(struct platform_device *pdev) +{ + struct mdp_dev *mdp = platform_get_drvdata(pdev); + + v4l2_device_unregister(&mdp->v4l2_dev); + + dev_dbg(&pdev->dev, "%s driver unloaded\n", pdev->name); +} + +static int __maybe_unused mdp_suspend(struct device *dev) +{ + struct mdp_dev *mdp = dev_get_drvdata(dev); + int ret; + + atomic_set(&mdp->suspended, 1); + + if (atomic_read(&mdp->job_count)) { + ret = wait_event_timeout(mdp->callback_wq, + !atomic_read(&mdp->job_count), + 2 * HZ); + if (ret == 0) { + dev_err(dev, + "%s:flushed cmdq task incomplete, count=%d\n", + __func__, atomic_read(&mdp->job_count)); + return -EBUSY; + } + } + + return 0; +} + +static int __maybe_unused mdp_resume(struct device *dev) +{ + struct mdp_dev *mdp = dev_get_drvdata(dev); + + atomic_set(&mdp->suspended, 0); + + return 0; +} + +static const struct dev_pm_ops mdp_pm_ops = { + SET_SYSTEM_SLEEP_PM_OPS(mdp_suspend, mdp_resume) +}; + +static struct platform_driver mdp_driver = { + .probe = mdp_probe, + .remove_new = mdp_remove, + .driver = { + .name = MDP_MODULE_NAME, + .pm = &mdp_pm_ops, + .of_match_table = mdp_of_ids, + }, +}; + +module_platform_driver(mdp_driver); + +MODULE_AUTHOR("Ping-Hsun Wu <ping-hsun.wu@mediatek.com>"); +MODULE_DESCRIPTION("MediaTek image processor 3 driver"); +MODULE_LICENSE("GPL"); diff --git a/drivers/media/platform/mediatek/mdp3/mtk-mdp3-core.h b/drivers/media/platform/mediatek/mdp3/mtk-mdp3-core.h new file mode 100644 index 0000000000..7e21d226ce --- /dev/null +++ b/drivers/media/platform/mediatek/mdp3/mtk-mdp3-core.h @@ -0,0 +1,108 @@ +/* SPDX-License-Identifier: GPL-2.0-only */ +/* + * Copyright (c) 2022 MediaTek Inc. + * Author: Ping-Hsun Wu <ping-hsun.wu@mediatek.com> + */ + +#ifndef __MTK_MDP3_CORE_H__ +#define __MTK_MDP3_CORE_H__ + +#include <media/v4l2-device.h> +#include <media/v4l2-mem2mem.h> +#include <linux/soc/mediatek/mtk-mmsys.h> +#include <linux/soc/mediatek/mtk-mutex.h> +#include "mtk-mdp3-comp.h" +#include "mtk-mdp3-vpu.h" + +#define MDP_MODULE_NAME "mtk-mdp3" +#define MDP_DEVICE_NAME "MediaTek MDP3" +#define MDP_PHANDLE_NAME "mediatek,mdp3" + +enum mdp_infra_id { + MDP_INFRA_MMSYS, + MDP_INFRA_MUTEX, + MDP_INFRA_SCP, + MDP_INFRA_MAX +}; + +enum mdp_buffer_usage { + MDP_BUFFER_USAGE_HW_READ, + MDP_BUFFER_USAGE_MDP, + MDP_BUFFER_USAGE_MDP2, + MDP_BUFFER_USAGE_ISP, + MDP_BUFFER_USAGE_WPE, +}; + +struct mdp_platform_config { + bool rdma_support_10bit; + bool rdma_rsz1_sram_sharing; + bool rdma_upsample_repeat_only; + bool rsz_disable_dcm_small_sample; + bool wrot_filter_constraint; +}; + +/* indicate which mutex is used by each pipepline */ +enum mdp_pipe_id { + MDP_PIPE_WPEI, + MDP_PIPE_WPEI2, + MDP_PIPE_IMGI, + MDP_PIPE_RDMA0, + MDP_PIPE_MAX +}; + +struct mtk_mdp_driver_data { + const int mdp_plat_id; + const struct of_device_id *mdp_probe_infra; + const struct mdp_platform_config *mdp_cfg; + const u32 *mdp_mutex_table_idx; + const struct mdp_comp_data *comp_data; + unsigned int comp_data_len; + const struct of_device_id *mdp_sub_comp_dt_ids; + const struct mdp_format *format; + unsigned int format_len; + const struct mdp_limit *def_limit; + const struct mdp_pipe_info *pipe_info; + unsigned int pipe_info_len; +}; + +struct mdp_dev { + struct platform_device *pdev; + struct device *mdp_mmsys; + struct mtk_mutex *mdp_mutex[MDP_PIPE_MAX]; + struct mdp_comp *comp[MDP_MAX_COMP_COUNT]; + const struct mtk_mdp_driver_data *mdp_data; + + struct workqueue_struct *job_wq; + struct workqueue_struct *clock_wq; + struct mdp_vpu_dev vpu; + struct mtk_scp *scp; + struct rproc *rproc_handle; + /* synchronization protect for accessing vpu working buffer info */ + struct mutex vpu_lock; + s32 vpu_count; + u32 id_count; + struct ida mdp_ida; + struct cmdq_client *cmdq_clt; + wait_queue_head_t callback_wq; + + struct v4l2_device v4l2_dev; + struct video_device *m2m_vdev; + struct v4l2_m2m_dev *m2m_dev; + /* synchronization protect for m2m device operation */ + struct mutex m2m_lock; + atomic_t suspended; + atomic_t job_count; +}; + +struct mdp_pipe_info { + enum mdp_pipe_id pipe_id; + u32 mutex_id; +}; + +int mdp_vpu_get_locked(struct mdp_dev *mdp); +void mdp_vpu_put_locked(struct mdp_dev *mdp); +int mdp_vpu_register(struct mdp_dev *mdp); +void mdp_vpu_unregister(struct mdp_dev *mdp); +void mdp_video_device_release(struct video_device *vdev); + +#endif /* __MTK_MDP3_CORE_H__ */ diff --git a/drivers/media/platform/mediatek/mdp3/mtk-mdp3-m2m.c b/drivers/media/platform/mediatek/mdp3/mtk-mdp3-m2m.c new file mode 100644 index 0000000000..a298c1b15b --- /dev/null +++ b/drivers/media/platform/mediatek/mdp3/mtk-mdp3-m2m.c @@ -0,0 +1,726 @@ +// SPDX-License-Identifier: GPL-2.0-only +/* + * Copyright (c) 2022 MediaTek Inc. + * Author: Ping-Hsun Wu <ping-hsun.wu@mediatek.com> + */ + +#include <linux/platform_device.h> +#include <media/v4l2-ioctl.h> +#include <media/v4l2-event.h> +#include <media/videobuf2-dma-contig.h> +#include "mtk-mdp3-m2m.h" + +static inline struct mdp_m2m_ctx *fh_to_ctx(struct v4l2_fh *fh) +{ + return container_of(fh, struct mdp_m2m_ctx, fh); +} + +static inline struct mdp_m2m_ctx *ctrl_to_ctx(struct v4l2_ctrl *ctrl) +{ + return container_of(ctrl->handler, struct mdp_m2m_ctx, ctrl_handler); +} + +static inline struct mdp_frame *ctx_get_frame(struct mdp_m2m_ctx *ctx, + enum v4l2_buf_type type) +{ + if (V4L2_TYPE_IS_OUTPUT(type)) + return &ctx->curr_param.output; + else + return &ctx->curr_param.captures[0]; +} + +static inline void mdp_m2m_ctx_set_state(struct mdp_m2m_ctx *ctx, u32 state) +{ + atomic_or(state, &ctx->curr_param.state); +} + +static inline bool mdp_m2m_ctx_is_state_set(struct mdp_m2m_ctx *ctx, u32 mask) +{ + return ((atomic_read(&ctx->curr_param.state) & mask) == mask); +} + +static void mdp_m2m_process_done(void *priv, int vb_state) +{ + struct mdp_m2m_ctx *ctx = priv; + struct vb2_v4l2_buffer *src_vbuf, *dst_vbuf; + + src_vbuf = (struct vb2_v4l2_buffer *) + v4l2_m2m_src_buf_remove(ctx->m2m_ctx); + dst_vbuf = (struct vb2_v4l2_buffer *) + v4l2_m2m_dst_buf_remove(ctx->m2m_ctx); + ctx->curr_param.frame_no = ctx->frame_count[MDP_M2M_SRC]; + src_vbuf->sequence = ctx->frame_count[MDP_M2M_SRC]++; + dst_vbuf->sequence = ctx->frame_count[MDP_M2M_DST]++; + v4l2_m2m_buf_copy_metadata(src_vbuf, dst_vbuf, true); + + v4l2_m2m_buf_done(src_vbuf, vb_state); + v4l2_m2m_buf_done(dst_vbuf, vb_state); + v4l2_m2m_job_finish(ctx->mdp_dev->m2m_dev, ctx->m2m_ctx); +} + +static void mdp_m2m_device_run(void *priv) +{ + struct mdp_m2m_ctx *ctx = priv; + struct mdp_frame *frame; + struct vb2_v4l2_buffer *src_vb, *dst_vb; + struct img_ipi_frameparam param = {}; + struct mdp_cmdq_param task = {}; + enum vb2_buffer_state vb_state = VB2_BUF_STATE_ERROR; + int ret; + + if (mdp_m2m_ctx_is_state_set(ctx, MDP_M2M_CTX_ERROR)) { + dev_err(&ctx->mdp_dev->pdev->dev, + "mdp_m2m_ctx is in error state\n"); + goto worker_end; + } + + param.frame_no = ctx->curr_param.frame_no; + param.type = ctx->curr_param.type; + param.num_inputs = 1; + param.num_outputs = 1; + + frame = ctx_get_frame(ctx, V4L2_BUF_TYPE_VIDEO_OUTPUT_MPLANE); + src_vb = v4l2_m2m_next_src_buf(ctx->m2m_ctx); + mdp_set_src_config(¶m.inputs[0], frame, &src_vb->vb2_buf); + + frame = ctx_get_frame(ctx, V4L2_BUF_TYPE_VIDEO_CAPTURE_MPLANE); + dst_vb = v4l2_m2m_next_dst_buf(ctx->m2m_ctx); + mdp_set_dst_config(¶m.outputs[0], frame, &dst_vb->vb2_buf); + + ret = mdp_vpu_process(&ctx->mdp_dev->vpu, ¶m); + if (ret) { + dev_err(&ctx->mdp_dev->pdev->dev, + "VPU MDP process failed: %d\n", ret); + goto worker_end; + } + + task.config = ctx->mdp_dev->vpu.config; + task.param = ¶m; + task.composes[0] = &frame->compose; + task.cmdq_cb = NULL; + task.cb_data = NULL; + task.mdp_ctx = ctx; + + ret = mdp_cmdq_send(ctx->mdp_dev, &task); + if (ret) { + dev_err(&ctx->mdp_dev->pdev->dev, + "CMDQ sendtask failed: %d\n", ret); + goto worker_end; + } + + return; + +worker_end: + mdp_m2m_process_done(ctx, vb_state); +} + +static int mdp_m2m_start_streaming(struct vb2_queue *q, unsigned int count) +{ + struct mdp_m2m_ctx *ctx = vb2_get_drv_priv(q); + struct mdp_frame *capture; + struct vb2_queue *vq; + int ret; + bool out_streaming, cap_streaming; + + if (V4L2_TYPE_IS_OUTPUT(q->type)) + ctx->frame_count[MDP_M2M_SRC] = 0; + + if (V4L2_TYPE_IS_CAPTURE(q->type)) + ctx->frame_count[MDP_M2M_DST] = 0; + + capture = ctx_get_frame(ctx, V4L2_BUF_TYPE_VIDEO_CAPTURE_MPLANE); + vq = v4l2_m2m_get_src_vq(ctx->m2m_ctx); + out_streaming = vb2_is_streaming(vq); + vq = v4l2_m2m_get_dst_vq(ctx->m2m_ctx); + cap_streaming = vb2_is_streaming(vq); + + /* Check to see if scaling ratio is within supported range */ + if ((V4L2_TYPE_IS_OUTPUT(q->type) && cap_streaming) || + (V4L2_TYPE_IS_CAPTURE(q->type) && out_streaming)) { + ret = mdp_check_scaling_ratio(&capture->crop.c, + &capture->compose, + capture->rotation, + ctx->curr_param.limit); + if (ret) { + dev_err(&ctx->mdp_dev->pdev->dev, + "Out of scaling range\n"); + return ret; + } + } + + if (!mdp_m2m_ctx_is_state_set(ctx, MDP_VPU_INIT)) { + ret = mdp_vpu_get_locked(ctx->mdp_dev); + if (ret) { + dev_err(&ctx->mdp_dev->pdev->dev, + "VPU init failed %d\n", ret); + return -EINVAL; + } + mdp_m2m_ctx_set_state(ctx, MDP_VPU_INIT); + } + + return 0; +} + +static struct vb2_v4l2_buffer *mdp_m2m_buf_remove(struct mdp_m2m_ctx *ctx, + unsigned int type) +{ + if (V4L2_TYPE_IS_OUTPUT(type)) + return (struct vb2_v4l2_buffer *) + v4l2_m2m_src_buf_remove(ctx->m2m_ctx); + else + return (struct vb2_v4l2_buffer *) + v4l2_m2m_dst_buf_remove(ctx->m2m_ctx); +} + +static void mdp_m2m_stop_streaming(struct vb2_queue *q) +{ + struct mdp_m2m_ctx *ctx = vb2_get_drv_priv(q); + struct vb2_v4l2_buffer *vb; + + vb = mdp_m2m_buf_remove(ctx, q->type); + while (vb) { + v4l2_m2m_buf_done(vb, VB2_BUF_STATE_ERROR); + vb = mdp_m2m_buf_remove(ctx, q->type); + } +} + +static int mdp_m2m_queue_setup(struct vb2_queue *q, + unsigned int *num_buffers, + unsigned int *num_planes, unsigned int sizes[], + struct device *alloc_devs[]) +{ + struct mdp_m2m_ctx *ctx = vb2_get_drv_priv(q); + struct v4l2_pix_format_mplane *pix_mp; + u32 i; + + pix_mp = &ctx_get_frame(ctx, q->type)->format.fmt.pix_mp; + + /* from VIDIOC_CREATE_BUFS */ + if (*num_planes) { + if (*num_planes != pix_mp->num_planes) + return -EINVAL; + for (i = 0; i < pix_mp->num_planes; ++i) + if (sizes[i] < pix_mp->plane_fmt[i].sizeimage) + return -EINVAL; + } else {/* from VIDIOC_REQBUFS */ + *num_planes = pix_mp->num_planes; + for (i = 0; i < pix_mp->num_planes; ++i) + sizes[i] = pix_mp->plane_fmt[i].sizeimage; + } + + return 0; +} + +static int mdp_m2m_buf_prepare(struct vb2_buffer *vb) +{ + struct mdp_m2m_ctx *ctx = vb2_get_drv_priv(vb->vb2_queue); + struct v4l2_pix_format_mplane *pix_mp; + struct vb2_v4l2_buffer *v4l2_buf = to_vb2_v4l2_buffer(vb); + u32 i; + + v4l2_buf->field = V4L2_FIELD_NONE; + + if (V4L2_TYPE_IS_CAPTURE(vb->type)) { + pix_mp = &ctx_get_frame(ctx, vb->type)->format.fmt.pix_mp; + for (i = 0; i < pix_mp->num_planes; ++i) { + vb2_set_plane_payload(vb, i, + pix_mp->plane_fmt[i].sizeimage); + } + } + return 0; +} + +static int mdp_m2m_buf_out_validate(struct vb2_buffer *vb) +{ + struct vb2_v4l2_buffer *v4l2_buf = to_vb2_v4l2_buffer(vb); + + v4l2_buf->field = V4L2_FIELD_NONE; + + return 0; +} + +static void mdp_m2m_buf_queue(struct vb2_buffer *vb) +{ + struct mdp_m2m_ctx *ctx = vb2_get_drv_priv(vb->vb2_queue); + struct vb2_v4l2_buffer *v4l2_buf = to_vb2_v4l2_buffer(vb); + + v4l2_buf->field = V4L2_FIELD_NONE; + + v4l2_m2m_buf_queue(ctx->m2m_ctx, to_vb2_v4l2_buffer(vb)); +} + +static const struct vb2_ops mdp_m2m_qops = { + .queue_setup = mdp_m2m_queue_setup, + .wait_prepare = vb2_ops_wait_prepare, + .wait_finish = vb2_ops_wait_finish, + .buf_prepare = mdp_m2m_buf_prepare, + .start_streaming = mdp_m2m_start_streaming, + .stop_streaming = mdp_m2m_stop_streaming, + .buf_queue = mdp_m2m_buf_queue, + .buf_out_validate = mdp_m2m_buf_out_validate, +}; + +static int mdp_m2m_querycap(struct file *file, void *fh, + struct v4l2_capability *cap) +{ + strscpy(cap->driver, MDP_MODULE_NAME, sizeof(cap->driver)); + strscpy(cap->card, MDP_DEVICE_NAME, sizeof(cap->card)); + + return 0; +} + +static int mdp_m2m_enum_fmt_mplane(struct file *file, void *fh, + struct v4l2_fmtdesc *f) +{ + struct mdp_m2m_ctx *ctx = fh_to_ctx(fh); + + return mdp_enum_fmt_mplane(ctx->mdp_dev, f); +} + +static int mdp_m2m_g_fmt_mplane(struct file *file, void *fh, + struct v4l2_format *f) +{ + struct mdp_m2m_ctx *ctx = fh_to_ctx(fh); + struct mdp_frame *frame; + struct v4l2_pix_format_mplane *pix_mp; + + frame = ctx_get_frame(ctx, f->type); + *f = frame->format; + pix_mp = &f->fmt.pix_mp; + pix_mp->colorspace = ctx->curr_param.colorspace; + pix_mp->xfer_func = ctx->curr_param.xfer_func; + pix_mp->ycbcr_enc = ctx->curr_param.ycbcr_enc; + pix_mp->quantization = ctx->curr_param.quant; + + return 0; +} + +static int mdp_m2m_s_fmt_mplane(struct file *file, void *fh, + struct v4l2_format *f) +{ + struct mdp_m2m_ctx *ctx = fh_to_ctx(fh); + struct mdp_frame *frame = ctx_get_frame(ctx, f->type); + struct mdp_frame *capture; + const struct mdp_format *fmt; + struct vb2_queue *vq; + + fmt = mdp_try_fmt_mplane(ctx->mdp_dev, f, &ctx->curr_param, ctx->id); + if (!fmt) + return -EINVAL; + + vq = v4l2_m2m_get_vq(ctx->m2m_ctx, f->type); + if (vb2_is_busy(vq)) + return -EBUSY; + + frame->format = *f; + frame->mdp_fmt = fmt; + frame->ycbcr_prof = mdp_map_ycbcr_prof_mplane(f, fmt->mdp_color); + frame->usage = V4L2_TYPE_IS_OUTPUT(f->type) ? + MDP_BUFFER_USAGE_HW_READ : MDP_BUFFER_USAGE_MDP; + + capture = ctx_get_frame(ctx, V4L2_BUF_TYPE_VIDEO_CAPTURE_MPLANE); + if (V4L2_TYPE_IS_OUTPUT(f->type)) { + capture->crop.c.left = 0; + capture->crop.c.top = 0; + capture->crop.c.width = f->fmt.pix_mp.width; + capture->crop.c.height = f->fmt.pix_mp.height; + ctx->curr_param.colorspace = f->fmt.pix_mp.colorspace; + ctx->curr_param.ycbcr_enc = f->fmt.pix_mp.ycbcr_enc; + ctx->curr_param.quant = f->fmt.pix_mp.quantization; + ctx->curr_param.xfer_func = f->fmt.pix_mp.xfer_func; + } else { + capture->compose.left = 0; + capture->compose.top = 0; + capture->compose.width = f->fmt.pix_mp.width; + capture->compose.height = f->fmt.pix_mp.height; + } + + return 0; +} + +static int mdp_m2m_try_fmt_mplane(struct file *file, void *fh, + struct v4l2_format *f) +{ + struct mdp_m2m_ctx *ctx = fh_to_ctx(fh); + + if (!mdp_try_fmt_mplane(ctx->mdp_dev, f, &ctx->curr_param, ctx->id)) + return -EINVAL; + + return 0; +} + +static int mdp_m2m_g_selection(struct file *file, void *fh, + struct v4l2_selection *s) +{ + struct mdp_m2m_ctx *ctx = fh_to_ctx(fh); + struct mdp_frame *frame; + bool valid = false; + + if (s->type == V4L2_BUF_TYPE_VIDEO_OUTPUT) + valid = mdp_target_is_crop(s->target); + else if (s->type == V4L2_BUF_TYPE_VIDEO_CAPTURE) + valid = mdp_target_is_compose(s->target); + + if (!valid) + return -EINVAL; + + switch (s->target) { + case V4L2_SEL_TGT_CROP: + if (s->type != V4L2_BUF_TYPE_VIDEO_OUTPUT) + return -EINVAL; + frame = ctx_get_frame(ctx, V4L2_BUF_TYPE_VIDEO_CAPTURE_MPLANE); + s->r = frame->crop.c; + return 0; + case V4L2_SEL_TGT_COMPOSE: + if (s->type != V4L2_BUF_TYPE_VIDEO_CAPTURE) + return -EINVAL; + frame = ctx_get_frame(ctx, V4L2_BUF_TYPE_VIDEO_CAPTURE_MPLANE); + s->r = frame->compose; + return 0; + case V4L2_SEL_TGT_CROP_DEFAULT: + case V4L2_SEL_TGT_CROP_BOUNDS: + if (s->type != V4L2_BUF_TYPE_VIDEO_OUTPUT) + return -EINVAL; + frame = ctx_get_frame(ctx, s->type); + s->r.left = 0; + s->r.top = 0; + s->r.width = frame->format.fmt.pix_mp.width; + s->r.height = frame->format.fmt.pix_mp.height; + return 0; + case V4L2_SEL_TGT_COMPOSE_DEFAULT: + case V4L2_SEL_TGT_COMPOSE_BOUNDS: + if (s->type != V4L2_BUF_TYPE_VIDEO_CAPTURE) + return -EINVAL; + frame = ctx_get_frame(ctx, s->type); + s->r.left = 0; + s->r.top = 0; + s->r.width = frame->format.fmt.pix_mp.width; + s->r.height = frame->format.fmt.pix_mp.height; + return 0; + } + return -EINVAL; +} + +static int mdp_m2m_s_selection(struct file *file, void *fh, + struct v4l2_selection *s) +{ + struct mdp_m2m_ctx *ctx = fh_to_ctx(fh); + struct mdp_frame *frame = ctx_get_frame(ctx, s->type); + struct mdp_frame *capture; + struct v4l2_rect r; + struct device *dev = &ctx->mdp_dev->pdev->dev; + bool valid = false; + int ret; + + if (s->type == V4L2_BUF_TYPE_VIDEO_OUTPUT) + valid = (s->target == V4L2_SEL_TGT_CROP); + else if (s->type == V4L2_BUF_TYPE_VIDEO_CAPTURE) + valid = (s->target == V4L2_SEL_TGT_COMPOSE); + + if (!valid) { + dev_dbg(dev, "[%s:%d] invalid type:%u target:%u", __func__, + ctx->id, s->type, s->target); + return -EINVAL; + } + + ret = mdp_try_crop(ctx, &r, s, frame); + if (ret) + return ret; + capture = ctx_get_frame(ctx, V4L2_BUF_TYPE_VIDEO_CAPTURE_MPLANE); + + if (mdp_target_is_crop(s->target)) + capture->crop.c = r; + else + capture->compose = r; + + s->r = r; + + return 0; +} + +static const struct v4l2_ioctl_ops mdp_m2m_ioctl_ops = { + .vidioc_querycap = mdp_m2m_querycap, + .vidioc_enum_fmt_vid_cap = mdp_m2m_enum_fmt_mplane, + .vidioc_enum_fmt_vid_out = mdp_m2m_enum_fmt_mplane, + .vidioc_g_fmt_vid_cap_mplane = mdp_m2m_g_fmt_mplane, + .vidioc_g_fmt_vid_out_mplane = mdp_m2m_g_fmt_mplane, + .vidioc_s_fmt_vid_cap_mplane = mdp_m2m_s_fmt_mplane, + .vidioc_s_fmt_vid_out_mplane = mdp_m2m_s_fmt_mplane, + .vidioc_try_fmt_vid_cap_mplane = mdp_m2m_try_fmt_mplane, + .vidioc_try_fmt_vid_out_mplane = mdp_m2m_try_fmt_mplane, + .vidioc_reqbufs = v4l2_m2m_ioctl_reqbufs, + .vidioc_querybuf = v4l2_m2m_ioctl_querybuf, + .vidioc_qbuf = v4l2_m2m_ioctl_qbuf, + .vidioc_expbuf = v4l2_m2m_ioctl_expbuf, + .vidioc_dqbuf = v4l2_m2m_ioctl_dqbuf, + .vidioc_create_bufs = v4l2_m2m_ioctl_create_bufs, + .vidioc_streamon = v4l2_m2m_ioctl_streamon, + .vidioc_streamoff = v4l2_m2m_ioctl_streamoff, + .vidioc_g_selection = mdp_m2m_g_selection, + .vidioc_s_selection = mdp_m2m_s_selection, + .vidioc_subscribe_event = v4l2_ctrl_subscribe_event, + .vidioc_unsubscribe_event = v4l2_event_unsubscribe, +}; + +static int mdp_m2m_queue_init(void *priv, + struct vb2_queue *src_vq, + struct vb2_queue *dst_vq) +{ + struct mdp_m2m_ctx *ctx = priv; + int ret; + + src_vq->type = V4L2_BUF_TYPE_VIDEO_OUTPUT_MPLANE; + src_vq->io_modes = VB2_MMAP | VB2_DMABUF; + src_vq->ops = &mdp_m2m_qops; + src_vq->mem_ops = &vb2_dma_contig_memops; + src_vq->drv_priv = ctx; + src_vq->buf_struct_size = sizeof(struct v4l2_m2m_buffer); + src_vq->timestamp_flags = V4L2_BUF_FLAG_TIMESTAMP_COPY; + src_vq->dev = &ctx->mdp_dev->pdev->dev; + src_vq->lock = &ctx->ctx_lock; + + ret = vb2_queue_init(src_vq); + if (ret) + return ret; + + dst_vq->type = V4L2_BUF_TYPE_VIDEO_CAPTURE_MPLANE; + dst_vq->io_modes = VB2_MMAP | VB2_DMABUF; + dst_vq->ops = &mdp_m2m_qops; + dst_vq->mem_ops = &vb2_dma_contig_memops; + dst_vq->drv_priv = ctx; + dst_vq->buf_struct_size = sizeof(struct v4l2_m2m_buffer); + dst_vq->timestamp_flags = V4L2_BUF_FLAG_TIMESTAMP_COPY; + dst_vq->dev = &ctx->mdp_dev->pdev->dev; + dst_vq->lock = &ctx->ctx_lock; + + return vb2_queue_init(dst_vq); +} + +static int mdp_m2m_s_ctrl(struct v4l2_ctrl *ctrl) +{ + struct mdp_m2m_ctx *ctx = ctrl_to_ctx(ctrl); + struct mdp_frame *capture; + + capture = ctx_get_frame(ctx, V4L2_BUF_TYPE_VIDEO_CAPTURE_MPLANE); + switch (ctrl->id) { + case V4L2_CID_HFLIP: + capture->hflip = ctrl->val; + break; + case V4L2_CID_VFLIP: + capture->vflip = ctrl->val; + break; + case V4L2_CID_ROTATE: + capture->rotation = ctrl->val; + break; + } + + return 0; +} + +static const struct v4l2_ctrl_ops mdp_m2m_ctrl_ops = { + .s_ctrl = mdp_m2m_s_ctrl, +}; + +static int mdp_m2m_ctrls_create(struct mdp_m2m_ctx *ctx) +{ + v4l2_ctrl_handler_init(&ctx->ctrl_handler, MDP_MAX_CTRLS); + ctx->ctrls.hflip = v4l2_ctrl_new_std(&ctx->ctrl_handler, + &mdp_m2m_ctrl_ops, V4L2_CID_HFLIP, + 0, 1, 1, 0); + ctx->ctrls.vflip = v4l2_ctrl_new_std(&ctx->ctrl_handler, + &mdp_m2m_ctrl_ops, V4L2_CID_VFLIP, + 0, 1, 1, 0); + ctx->ctrls.rotate = v4l2_ctrl_new_std(&ctx->ctrl_handler, + &mdp_m2m_ctrl_ops, + V4L2_CID_ROTATE, 0, 270, 90, 0); + + if (ctx->ctrl_handler.error) { + int err = ctx->ctrl_handler.error; + + v4l2_ctrl_handler_free(&ctx->ctrl_handler); + dev_err(&ctx->mdp_dev->pdev->dev, + "Failed to register controls\n"); + return err; + } + return 0; +} + +static int mdp_m2m_open(struct file *file) +{ + struct video_device *vdev = video_devdata(file); + struct mdp_dev *mdp = video_get_drvdata(vdev); + struct mdp_m2m_ctx *ctx; + struct device *dev = &mdp->pdev->dev; + int ret; + struct v4l2_format default_format = {}; + const struct mdp_limit *limit = mdp->mdp_data->def_limit; + + ctx = kzalloc(sizeof(*ctx), GFP_KERNEL); + if (!ctx) + return -ENOMEM; + + if (mutex_lock_interruptible(&mdp->m2m_lock)) { + ret = -ERESTARTSYS; + goto err_free_ctx; + } + + ret = ida_alloc(&mdp->mdp_ida, GFP_KERNEL); + if (ret < 0) + goto err_unlock_mutex; + ctx->id = ret; + + ctx->mdp_dev = mdp; + + v4l2_fh_init(&ctx->fh, vdev); + file->private_data = &ctx->fh; + ret = mdp_m2m_ctrls_create(ctx); + if (ret) + goto err_exit_fh; + + /* Use separate control handler per file handle */ + ctx->fh.ctrl_handler = &ctx->ctrl_handler; + v4l2_fh_add(&ctx->fh); + + mutex_init(&ctx->ctx_lock); + ctx->m2m_ctx = v4l2_m2m_ctx_init(mdp->m2m_dev, ctx, mdp_m2m_queue_init); + if (IS_ERR(ctx->m2m_ctx)) { + dev_err(dev, "Failed to initialize m2m context\n"); + ret = PTR_ERR(ctx->m2m_ctx); + goto err_release_handler; + } + ctx->fh.m2m_ctx = ctx->m2m_ctx; + + ctx->curr_param.ctx = ctx; + ret = mdp_frameparam_init(mdp, &ctx->curr_param); + if (ret) { + dev_err(dev, "Failed to initialize mdp parameter\n"); + goto err_release_m2m_ctx; + } + + mutex_unlock(&mdp->m2m_lock); + + /* Default format */ + default_format.type = V4L2_BUF_TYPE_VIDEO_OUTPUT_MPLANE; + default_format.fmt.pix_mp.width = limit->out_limit.wmin; + default_format.fmt.pix_mp.height = limit->out_limit.hmin; + default_format.fmt.pix_mp.pixelformat = V4L2_PIX_FMT_YUV420M; + mdp_m2m_s_fmt_mplane(file, &ctx->fh, &default_format); + default_format.type = V4L2_BUF_TYPE_VIDEO_CAPTURE_MPLANE; + mdp_m2m_s_fmt_mplane(file, &ctx->fh, &default_format); + + dev_dbg(dev, "%s:[%d]", __func__, ctx->id); + + return 0; + +err_release_m2m_ctx: + v4l2_m2m_ctx_release(ctx->m2m_ctx); +err_release_handler: + v4l2_ctrl_handler_free(&ctx->ctrl_handler); + v4l2_fh_del(&ctx->fh); +err_exit_fh: + v4l2_fh_exit(&ctx->fh); + ida_free(&mdp->mdp_ida, ctx->id); +err_unlock_mutex: + mutex_unlock(&mdp->m2m_lock); +err_free_ctx: + kfree(ctx); + + return ret; +} + +static int mdp_m2m_release(struct file *file) +{ + struct mdp_m2m_ctx *ctx = fh_to_ctx(file->private_data); + struct mdp_dev *mdp = video_drvdata(file); + struct device *dev = &mdp->pdev->dev; + + mutex_lock(&mdp->m2m_lock); + v4l2_m2m_ctx_release(ctx->m2m_ctx); + if (mdp_m2m_ctx_is_state_set(ctx, MDP_VPU_INIT)) + mdp_vpu_put_locked(mdp); + + v4l2_ctrl_handler_free(&ctx->ctrl_handler); + v4l2_fh_del(&ctx->fh); + v4l2_fh_exit(&ctx->fh); + ida_free(&mdp->mdp_ida, ctx->id); + mutex_unlock(&mdp->m2m_lock); + + dev_dbg(dev, "%s:[%d]", __func__, ctx->id); + kfree(ctx); + + return 0; +} + +static const struct v4l2_file_operations mdp_m2m_fops = { + .owner = THIS_MODULE, + .poll = v4l2_m2m_fop_poll, + .unlocked_ioctl = video_ioctl2, + .mmap = v4l2_m2m_fop_mmap, + .open = mdp_m2m_open, + .release = mdp_m2m_release, +}; + +static const struct v4l2_m2m_ops mdp_m2m_ops = { + .device_run = mdp_m2m_device_run, +}; + +int mdp_m2m_device_register(struct mdp_dev *mdp) +{ + struct device *dev = &mdp->pdev->dev; + int ret = 0; + + mdp->m2m_vdev = video_device_alloc(); + if (!mdp->m2m_vdev) { + dev_err(dev, "Failed to allocate video device\n"); + ret = -ENOMEM; + goto err_video_alloc; + } + mdp->m2m_vdev->device_caps = V4L2_CAP_VIDEO_M2M_MPLANE | + V4L2_CAP_STREAMING; + mdp->m2m_vdev->fops = &mdp_m2m_fops; + mdp->m2m_vdev->ioctl_ops = &mdp_m2m_ioctl_ops; + mdp->m2m_vdev->release = mdp_video_device_release; + mdp->m2m_vdev->lock = &mdp->m2m_lock; + mdp->m2m_vdev->vfl_dir = VFL_DIR_M2M; + mdp->m2m_vdev->v4l2_dev = &mdp->v4l2_dev; + snprintf(mdp->m2m_vdev->name, sizeof(mdp->m2m_vdev->name), "%s:m2m", + MDP_MODULE_NAME); + video_set_drvdata(mdp->m2m_vdev, mdp); + + mdp->m2m_dev = v4l2_m2m_init(&mdp_m2m_ops); + if (IS_ERR(mdp->m2m_dev)) { + dev_err(dev, "Failed to initialize v4l2-m2m device\n"); + ret = PTR_ERR(mdp->m2m_dev); + goto err_m2m_init; + } + + ret = video_register_device(mdp->m2m_vdev, VFL_TYPE_VIDEO, -1); + if (ret) { + dev_err(dev, "Failed to register video device\n"); + goto err_video_register; + } + + v4l2_info(&mdp->v4l2_dev, "Driver registered as /dev/video%d", + mdp->m2m_vdev->num); + return 0; + +err_video_register: + v4l2_m2m_release(mdp->m2m_dev); +err_m2m_init: + video_device_release(mdp->m2m_vdev); +err_video_alloc: + + return ret; +} + +void mdp_m2m_device_unregister(struct mdp_dev *mdp) +{ + video_unregister_device(mdp->m2m_vdev); +} + +void mdp_m2m_job_finish(struct mdp_m2m_ctx *ctx) +{ + enum vb2_buffer_state vb_state = VB2_BUF_STATE_DONE; + + mdp_m2m_process_done(ctx, vb_state); +} diff --git a/drivers/media/platform/mediatek/mdp3/mtk-mdp3-m2m.h b/drivers/media/platform/mediatek/mdp3/mtk-mdp3-m2m.h new file mode 100644 index 0000000000..dfc59e5b3b --- /dev/null +++ b/drivers/media/platform/mediatek/mdp3/mtk-mdp3-m2m.h @@ -0,0 +1,47 @@ +/* SPDX-License-Identifier: GPL-2.0-only */ +/* + * Copyright (c) 2022 MediaTek Inc. + * Author: Ping-Hsun Wu <ping-hsun.wu@mediatek.com> + */ + +#ifndef __MTK_MDP3_M2M_H__ +#define __MTK_MDP3_M2M_H__ + +#include <media/v4l2-ctrls.h> +#include "mtk-mdp3-core.h" +#include "mtk-mdp3-vpu.h" +#include "mtk-mdp3-regs.h" + +#define MDP_MAX_CTRLS 10 + +enum { + MDP_M2M_SRC = 0, + MDP_M2M_DST = 1, + MDP_M2M_MAX, +}; + +struct mdp_m2m_ctrls { + struct v4l2_ctrl *hflip; + struct v4l2_ctrl *vflip; + struct v4l2_ctrl *rotate; +}; + +struct mdp_m2m_ctx { + u32 id; + struct mdp_dev *mdp_dev; + struct v4l2_fh fh; + struct v4l2_ctrl_handler ctrl_handler; + struct mdp_m2m_ctrls ctrls; + struct v4l2_m2m_ctx *m2m_ctx; + u32 frame_count[MDP_M2M_MAX]; + + struct mdp_frameparam curr_param; + /* synchronization protect for mdp m2m context */ + struct mutex ctx_lock; +}; + +int mdp_m2m_device_register(struct mdp_dev *mdp); +void mdp_m2m_device_unregister(struct mdp_dev *mdp); +void mdp_m2m_job_finish(struct mdp_m2m_ctx *ctx); + +#endif /* __MTK_MDP3_M2M_H__ */ diff --git a/drivers/media/platform/mediatek/mdp3/mtk-mdp3-regs.c b/drivers/media/platform/mediatek/mdp3/mtk-mdp3-regs.c new file mode 100644 index 0000000000..9b436b911d --- /dev/null +++ b/drivers/media/platform/mediatek/mdp3/mtk-mdp3-regs.c @@ -0,0 +1,496 @@ +// SPDX-License-Identifier: GPL-2.0-only +/* + * Copyright (c) 2022 MediaTek Inc. + * Author: Ping-Hsun Wu <ping-hsun.wu@mediatek.com> + */ + +#include <linux/math64.h> +#include <media/v4l2-common.h> +#include <media/videobuf2-v4l2.h> +#include <media/videobuf2-dma-contig.h> +#include "mtk-mdp3-core.h" +#include "mtk-mdp3-regs.h" +#include "mtk-mdp3-m2m.h" + +static const struct mdp_format *mdp_find_fmt(const struct mtk_mdp_driver_data *mdp_data, + u32 pixelformat, u32 type) +{ + u32 i, flag; + + flag = V4L2_TYPE_IS_OUTPUT(type) ? MDP_FMT_FLAG_OUTPUT : + MDP_FMT_FLAG_CAPTURE; + for (i = 0; i < mdp_data->format_len; ++i) { + if (!(mdp_data->format[i].flags & flag)) + continue; + if (mdp_data->format[i].pixelformat == pixelformat) + return &mdp_data->format[i]; + } + return NULL; +} + +static const struct mdp_format *mdp_find_fmt_by_index(const struct mtk_mdp_driver_data *mdp_data, + u32 index, u32 type) +{ + u32 i, flag, num = 0; + + flag = V4L2_TYPE_IS_OUTPUT(type) ? MDP_FMT_FLAG_OUTPUT : + MDP_FMT_FLAG_CAPTURE; + for (i = 0; i < mdp_data->format_len; ++i) { + if (!(mdp_data->format[i].flags & flag)) + continue; + if (index == num) + return &mdp_data->format[i]; + num++; + } + return NULL; +} + +enum mdp_ycbcr_profile mdp_map_ycbcr_prof_mplane(struct v4l2_format *f, + u32 mdp_color) +{ + struct v4l2_pix_format_mplane *pix_mp = &f->fmt.pix_mp; + + if (MDP_COLOR_IS_RGB(mdp_color)) + return MDP_YCBCR_PROFILE_FULL_BT601; + + switch (pix_mp->colorspace) { + case V4L2_COLORSPACE_JPEG: + return MDP_YCBCR_PROFILE_JPEG; + case V4L2_COLORSPACE_REC709: + case V4L2_COLORSPACE_DCI_P3: + if (pix_mp->quantization == V4L2_QUANTIZATION_FULL_RANGE) + return MDP_YCBCR_PROFILE_FULL_BT709; + return MDP_YCBCR_PROFILE_BT709; + case V4L2_COLORSPACE_BT2020: + if (pix_mp->quantization == V4L2_QUANTIZATION_FULL_RANGE) + return MDP_YCBCR_PROFILE_FULL_BT2020; + return MDP_YCBCR_PROFILE_BT2020; + default: + if (pix_mp->quantization == V4L2_QUANTIZATION_FULL_RANGE) + return MDP_YCBCR_PROFILE_FULL_BT601; + return MDP_YCBCR_PROFILE_BT601; + } +} + +static void mdp_bound_align_image(u32 *w, u32 *h, + struct v4l2_frmsize_stepwise *s, + unsigned int salign) +{ + unsigned int org_w, org_h; + + org_w = *w; + org_h = *h; + v4l_bound_align_image(w, s->min_width, s->max_width, s->step_width, + h, s->min_height, s->max_height, s->step_height, + salign); + + s->min_width = org_w; + s->min_height = org_h; + v4l2_apply_frmsize_constraints(w, h, s); +} + +static int mdp_clamp_align(s32 *x, int min, int max, unsigned int align) +{ + unsigned int mask; + + if (min < 0 || max < 0) + return -ERANGE; + + /* Bits that must be zero to be aligned */ + mask = ~((1 << align) - 1); + + min = 0 ? 0 : ((min + ~mask) & mask); + max = max & mask; + if ((unsigned int)min > (unsigned int)max) + return -ERANGE; + + /* Clamp to aligned min and max */ + *x = clamp(*x, min, max); + + /* Round to nearest aligned value */ + if (align) + *x = (*x + (1 << (align - 1))) & mask; + return 0; +} + +int mdp_enum_fmt_mplane(struct mdp_dev *mdp, struct v4l2_fmtdesc *f) +{ + const struct mdp_format *fmt; + + fmt = mdp_find_fmt_by_index(mdp->mdp_data, f->index, f->type); + if (!fmt) + return -EINVAL; + + f->pixelformat = fmt->pixelformat; + return 0; +} + +const struct mdp_format *mdp_try_fmt_mplane(struct mdp_dev *mdp, + struct v4l2_format *f, + struct mdp_frameparam *param, + u32 ctx_id) +{ + struct device *dev = ¶m->ctx->mdp_dev->pdev->dev; + struct v4l2_pix_format_mplane *pix_mp = &f->fmt.pix_mp; + const struct mdp_format *fmt; + const struct mdp_pix_limit *pix_limit; + struct v4l2_frmsize_stepwise s; + u32 org_w, org_h; + unsigned int i; + + fmt = mdp_find_fmt(mdp->mdp_data, pix_mp->pixelformat, f->type); + if (!fmt) { + fmt = mdp_find_fmt_by_index(mdp->mdp_data, 0, f->type); + if (!fmt) { + dev_dbg(dev, "%d: pixelformat %c%c%c%c invalid", ctx_id, + (pix_mp->pixelformat & 0xff), + (pix_mp->pixelformat >> 8) & 0xff, + (pix_mp->pixelformat >> 16) & 0xff, + (pix_mp->pixelformat >> 24) & 0xff); + return NULL; + } + } + + pix_mp->field = V4L2_FIELD_NONE; + pix_mp->flags = 0; + pix_mp->pixelformat = fmt->pixelformat; + if (V4L2_TYPE_IS_CAPTURE(f->type)) { + pix_mp->colorspace = param->colorspace; + pix_mp->xfer_func = param->xfer_func; + pix_mp->ycbcr_enc = param->ycbcr_enc; + pix_mp->quantization = param->quant; + } + + pix_limit = V4L2_TYPE_IS_OUTPUT(f->type) ? ¶m->limit->out_limit : + ¶m->limit->cap_limit; + s.min_width = pix_limit->wmin; + s.max_width = pix_limit->wmax; + s.step_width = fmt->walign; + s.min_height = pix_limit->hmin; + s.max_height = pix_limit->hmax; + s.step_height = fmt->halign; + org_w = pix_mp->width; + org_h = pix_mp->height; + + mdp_bound_align_image(&pix_mp->width, &pix_mp->height, &s, fmt->salign); + if (org_w != pix_mp->width || org_h != pix_mp->height) + dev_dbg(dev, "%d: size change: %ux%u to %ux%u", ctx_id, + org_w, org_h, pix_mp->width, pix_mp->height); + + if (pix_mp->num_planes && pix_mp->num_planes != fmt->num_planes) + dev_dbg(dev, "%d num of planes change: %u to %u", ctx_id, + pix_mp->num_planes, fmt->num_planes); + pix_mp->num_planes = fmt->num_planes; + + for (i = 0; i < pix_mp->num_planes; ++i) { + u32 min_bpl = (pix_mp->width * fmt->row_depth[i]) >> 3; + u32 max_bpl = (pix_limit->wmax * fmt->row_depth[i]) >> 3; + u32 bpl = pix_mp->plane_fmt[i].bytesperline; + u32 min_si, max_si; + u32 si = pix_mp->plane_fmt[i].sizeimage; + u64 di; + + bpl = clamp(bpl, min_bpl, max_bpl); + pix_mp->plane_fmt[i].bytesperline = bpl; + + di = (u64)bpl * pix_mp->height * fmt->depth[i]; + min_si = (u32)div_u64(di, fmt->row_depth[i]); + di = (u64)bpl * s.max_height * fmt->depth[i]; + max_si = (u32)div_u64(di, fmt->row_depth[i]); + + si = clamp(si, min_si, max_si); + pix_mp->plane_fmt[i].sizeimage = si; + + dev_dbg(dev, "%d: p%u, bpl:%u [%u, %u], sizeimage:%u [%u, %u]", + ctx_id, i, bpl, min_bpl, max_bpl, si, min_si, max_si); + } + + return fmt; +} + +static int mdp_clamp_start(s32 *x, int min, int max, unsigned int align, + u32 flags) +{ + if (flags & V4L2_SEL_FLAG_GE) + max = *x; + if (flags & V4L2_SEL_FLAG_LE) + min = *x; + return mdp_clamp_align(x, min, max, align); +} + +static int mdp_clamp_end(s32 *x, int min, int max, unsigned int align, + u32 flags) +{ + if (flags & V4L2_SEL_FLAG_GE) + min = *x; + if (flags & V4L2_SEL_FLAG_LE) + max = *x; + return mdp_clamp_align(x, min, max, align); +} + +int mdp_try_crop(struct mdp_m2m_ctx *ctx, struct v4l2_rect *r, + const struct v4l2_selection *s, struct mdp_frame *frame) +{ + struct device *dev = &ctx->mdp_dev->pdev->dev; + s32 left, top, right, bottom; + u32 framew, frameh, walign, halign; + int ret; + + dev_dbg(dev, "%d target:%d, set:(%d,%d) %ux%u", ctx->id, + s->target, s->r.left, s->r.top, s->r.width, s->r.height); + + left = s->r.left; + top = s->r.top; + right = s->r.left + s->r.width; + bottom = s->r.top + s->r.height; + framew = frame->format.fmt.pix_mp.width; + frameh = frame->format.fmt.pix_mp.height; + + if (mdp_target_is_crop(s->target)) { + walign = 1; + halign = 1; + } else { + walign = frame->mdp_fmt->walign; + halign = frame->mdp_fmt->halign; + } + + dev_dbg(dev, "%d align:%u,%u, bound:%ux%u", ctx->id, + walign, halign, framew, frameh); + + ret = mdp_clamp_start(&left, 0, right, walign, s->flags); + if (ret) + return ret; + ret = mdp_clamp_start(&top, 0, bottom, halign, s->flags); + if (ret) + return ret; + ret = mdp_clamp_end(&right, left, framew, walign, s->flags); + if (ret) + return ret; + ret = mdp_clamp_end(&bottom, top, frameh, halign, s->flags); + if (ret) + return ret; + + r->left = left; + r->top = top; + r->width = right - left; + r->height = bottom - top; + + dev_dbg(dev, "%d crop:(%d,%d) %ux%u", ctx->id, + r->left, r->top, r->width, r->height); + return 0; +} + +int mdp_check_scaling_ratio(const struct v4l2_rect *crop, + const struct v4l2_rect *compose, s32 rotation, + const struct mdp_limit *limit) +{ + u32 crop_w, crop_h, comp_w, comp_h; + + crop_w = crop->width; + crop_h = crop->height; + if (90 == rotation || 270 == rotation) { + comp_w = compose->height; + comp_h = compose->width; + } else { + comp_w = compose->width; + comp_h = compose->height; + } + + if ((crop_w / comp_w) > limit->h_scale_down_max || + (crop_h / comp_h) > limit->v_scale_down_max || + (comp_w / crop_w) > limit->h_scale_up_max || + (comp_h / crop_h) > limit->v_scale_up_max) + return -ERANGE; + return 0; +} + +/* Stride that is accepted by MDP HW */ +static u32 mdp_fmt_get_stride(const struct mdp_format *fmt, + u32 bytesperline, unsigned int plane) +{ + enum mdp_color c = fmt->mdp_color; + u32 stride; + + stride = (bytesperline * MDP_COLOR_BITS_PER_PIXEL(c)) + / fmt->row_depth[0]; + if (plane == 0) + return stride; + if (plane < MDP_COLOR_GET_PLANE_COUNT(c)) { + if (MDP_COLOR_IS_BLOCK_MODE(c)) + stride = stride / 2; + return stride; + } + return 0; +} + +/* Stride that is accepted by MDP HW of format with contiguous planes */ +static u32 mdp_fmt_get_stride_contig(const struct mdp_format *fmt, + u32 pix_stride, unsigned int plane) +{ + enum mdp_color c = fmt->mdp_color; + u32 stride = pix_stride; + + if (plane == 0) + return stride; + if (plane < MDP_COLOR_GET_PLANE_COUNT(c)) { + stride = stride >> MDP_COLOR_GET_H_SUBSAMPLE(c); + if (MDP_COLOR_IS_UV_COPLANE(c) && !MDP_COLOR_IS_BLOCK_MODE(c)) + stride = stride * 2; + return stride; + } + return 0; +} + +/* Plane size that is accepted by MDP HW */ +static u32 mdp_fmt_get_plane_size(const struct mdp_format *fmt, + u32 stride, u32 height, unsigned int plane) +{ + enum mdp_color c = fmt->mdp_color; + u32 bytesperline; + + bytesperline = (stride * fmt->row_depth[0]) + / MDP_COLOR_BITS_PER_PIXEL(c); + if (plane == 0) + return bytesperline * height; + if (plane < MDP_COLOR_GET_PLANE_COUNT(c)) { + height = height >> MDP_COLOR_GET_V_SUBSAMPLE(c); + if (MDP_COLOR_IS_BLOCK_MODE(c)) + bytesperline = bytesperline * 2; + return bytesperline * height; + } + return 0; +} + +static void mdp_prepare_buffer(struct img_image_buffer *b, + struct mdp_frame *frame, struct vb2_buffer *vb) +{ + struct v4l2_pix_format_mplane *pix_mp = &frame->format.fmt.pix_mp; + unsigned int i; + + b->format.colorformat = frame->mdp_fmt->mdp_color; + b->format.ycbcr_prof = frame->ycbcr_prof; + for (i = 0; i < pix_mp->num_planes; ++i) { + u32 stride = mdp_fmt_get_stride(frame->mdp_fmt, + pix_mp->plane_fmt[i].bytesperline, i); + + b->format.plane_fmt[i].stride = stride; + b->format.plane_fmt[i].size = + mdp_fmt_get_plane_size(frame->mdp_fmt, stride, + pix_mp->height, i); + b->iova[i] = vb2_dma_contig_plane_dma_addr(vb, i); + } + for (; i < MDP_COLOR_GET_PLANE_COUNT(b->format.colorformat); ++i) { + u32 stride = mdp_fmt_get_stride_contig(frame->mdp_fmt, + b->format.plane_fmt[0].stride, i); + + b->format.plane_fmt[i].stride = stride; + b->format.plane_fmt[i].size = + mdp_fmt_get_plane_size(frame->mdp_fmt, stride, + pix_mp->height, i); + b->iova[i] = b->iova[i - 1] + b->format.plane_fmt[i - 1].size; + } + b->usage = frame->usage; +} + +void mdp_set_src_config(struct img_input *in, + struct mdp_frame *frame, struct vb2_buffer *vb) +{ + in->buffer.format.width = frame->format.fmt.pix_mp.width; + in->buffer.format.height = frame->format.fmt.pix_mp.height; + mdp_prepare_buffer(&in->buffer, frame, vb); +} + +static u32 mdp_to_fixed(u32 *r, struct v4l2_fract *f) +{ + u32 q; + + if (f->denominator == 0) { + *r = 0; + return 0; + } + + q = f->numerator / f->denominator; + *r = div_u64(((u64)f->numerator - q * f->denominator) << + IMG_SUBPIXEL_SHIFT, f->denominator); + return q; +} + +static void mdp_set_src_crop(struct img_crop *c, struct mdp_crop *crop) +{ + c->left = crop->c.left + + mdp_to_fixed(&c->left_subpix, &crop->left_subpix); + c->top = crop->c.top + + mdp_to_fixed(&c->top_subpix, &crop->top_subpix); + c->width = crop->c.width + + mdp_to_fixed(&c->width_subpix, &crop->width_subpix); + c->height = crop->c.height + + mdp_to_fixed(&c->height_subpix, &crop->height_subpix); +} + +static void mdp_set_orientation(struct img_output *out, + s32 rotation, bool hflip, bool vflip) +{ + u8 flip = 0; + + if (hflip) + flip ^= 1; + if (vflip) { + /* + * A vertical flip is equivalent to + * a 180-degree rotation with a horizontal flip + */ + rotation += 180; + flip ^= 1; + } + + out->rotation = rotation % 360; + if (flip != 0) + out->flags |= IMG_CTRL_FLAG_HFLIP; + else + out->flags &= ~IMG_CTRL_FLAG_HFLIP; +} + +void mdp_set_dst_config(struct img_output *out, + struct mdp_frame *frame, struct vb2_buffer *vb) +{ + out->buffer.format.width = frame->compose.width; + out->buffer.format.height = frame->compose.height; + mdp_prepare_buffer(&out->buffer, frame, vb); + mdp_set_src_crop(&out->crop, &frame->crop); + mdp_set_orientation(out, frame->rotation, frame->hflip, frame->vflip); +} + +int mdp_frameparam_init(struct mdp_dev *mdp, struct mdp_frameparam *param) +{ + struct mdp_frame *frame; + + if (!param) + return -EINVAL; + + INIT_LIST_HEAD(¶m->list); + param->limit = mdp->mdp_data->def_limit; + param->type = MDP_STREAM_TYPE_BITBLT; + + frame = ¶m->output; + frame->format.type = V4L2_BUF_TYPE_VIDEO_OUTPUT_MPLANE; + frame->mdp_fmt = mdp_try_fmt_mplane(mdp, &frame->format, param, 0); + frame->ycbcr_prof = + mdp_map_ycbcr_prof_mplane(&frame->format, + frame->mdp_fmt->mdp_color); + frame->usage = MDP_BUFFER_USAGE_HW_READ; + + param->num_captures = 1; + frame = ¶m->captures[0]; + frame->format.type = V4L2_BUF_TYPE_VIDEO_CAPTURE_MPLANE; + frame->mdp_fmt = mdp_try_fmt_mplane(mdp, &frame->format, param, 0); + frame->ycbcr_prof = + mdp_map_ycbcr_prof_mplane(&frame->format, + frame->mdp_fmt->mdp_color); + frame->usage = MDP_BUFFER_USAGE_MDP; + frame->crop.c.width = param->output.format.fmt.pix_mp.width; + frame->crop.c.height = param->output.format.fmt.pix_mp.height; + frame->compose.width = frame->format.fmt.pix_mp.width; + frame->compose.height = frame->format.fmt.pix_mp.height; + + return 0; +} diff --git a/drivers/media/platform/mediatek/mdp3/mtk-mdp3-regs.h b/drivers/media/platform/mediatek/mdp3/mtk-mdp3-regs.h new file mode 100644 index 0000000000..e9ab8ac2c0 --- /dev/null +++ b/drivers/media/platform/mediatek/mdp3/mtk-mdp3-regs.h @@ -0,0 +1,377 @@ +/* SPDX-License-Identifier: GPL-2.0-only */ +/* + * Copyright (c) 2022 MediaTek Inc. + * Author: Ping-Hsun Wu <ping-hsun.wu@mediatek.com> + */ + +#ifndef __MTK_MDP3_REGS_H__ +#define __MTK_MDP3_REGS_H__ + +#include <linux/videodev2.h> +#include <media/videobuf2-core.h> +#include "mtk-img-ipi.h" + +/* + * MDP native color code + * Plane count: 1, 2, 3 + * H-subsample: 0, 1, 2 + * V-subsample: 0, 1 + * Color group: 0-RGB, 1-YUV, 2-raw + */ +#define MDP_COLOR(COMPRESS, PACKED, LOOSE, VIDEO, PLANE, HF, VF, BITS, GROUP, SWAP, ID)\ + (((COMPRESS) << 29) | ((PACKED) << 27) | ((LOOSE) << 26) | ((VIDEO) << 23) |\ + ((PLANE) << 21) | ((HF) << 19) | ((VF) << 18) | ((BITS) << 8) |\ + ((GROUP) << 6) | ((SWAP) << 5) | ((ID) << 0)) + +#define MDP_COLOR_IS_COMPRESS(c) ((0x20000000 & (c)) >> 29) +#define MDP_COLOR_IS_10BIT_PACKED(c) ((0x08000000 & (c)) >> 27) +#define MDP_COLOR_IS_10BIT_LOOSE(c) (((0x0c000000 & (c)) >> 26) == 1) +#define MDP_COLOR_IS_10BIT_TILE(c) (((0x0c000000 & (c)) >> 26) == 3) +#define MDP_COLOR_IS_UFP(c) ((0x02000000 & (c)) >> 25) +#define MDP_COLOR_IS_INTERLACED(c) ((0x01000000 & (c)) >> 24) +#define MDP_COLOR_IS_BLOCK_MODE(c) ((0x00800000 & (c)) >> 23) +#define MDP_COLOR_GET_PLANE_COUNT(c) ((0x00600000 & (c)) >> 21) +#define MDP_COLOR_GET_H_SUBSAMPLE(c) ((0x00180000 & (c)) >> 19) +#define MDP_COLOR_GET_V_SUBSAMPLE(c) ((0x00040000 & (c)) >> 18) +#define MDP_COLOR_BITS_PER_PIXEL(c) ((0x0003ff00 & (c)) >> 8) +#define MDP_COLOR_GET_GROUP(c) ((0x000000c0 & (c)) >> 6) +#define MDP_COLOR_IS_SWAPPED(c) ((0x00000020 & (c)) >> 5) +#define MDP_COLOR_GET_UNIQUE_ID(c) ((0x0000001f & (c)) >> 0) +#define MDP_COLOR_GET_HW_FORMAT(c) ((0x0000001f & (c)) >> 0) + +#define MDP_COLOR_IS_RGB(c) (MDP_COLOR_GET_GROUP(c) == 0) +#define MDP_COLOR_IS_YUV(c) (MDP_COLOR_GET_GROUP(c) == 1) + +enum mdp_color { + MDP_COLOR_UNKNOWN = 0, + + /* MDP_COLOR_FULLG8 */ + MDP_COLOR_FULLG8_RGGB = MDP_COLOR(0, 0, 0, 0, 1, 0, 0, 8, 2, 0, 21), + MDP_COLOR_FULLG8_GRBG = MDP_COLOR(0, 0, 0, 0, 1, 0, 1, 8, 2, 0, 21), + MDP_COLOR_FULLG8_GBRG = MDP_COLOR(0, 0, 0, 0, 1, 1, 0, 8, 2, 0, 21), + MDP_COLOR_FULLG8_BGGR = MDP_COLOR(0, 0, 0, 0, 1, 1, 1, 8, 2, 0, 21), + MDP_COLOR_FULLG8 = MDP_COLOR_FULLG8_BGGR, + + /* MDP_COLOR_FULLG10 */ + MDP_COLOR_FULLG10_RGGB = MDP_COLOR(0, 0, 0, 0, 1, 0, 0, 10, 2, 0, 21), + MDP_COLOR_FULLG10_GRBG = MDP_COLOR(0, 0, 0, 0, 1, 0, 1, 10, 2, 0, 21), + MDP_COLOR_FULLG10_GBRG = MDP_COLOR(0, 0, 0, 0, 1, 1, 0, 10, 2, 0, 21), + MDP_COLOR_FULLG10_BGGR = MDP_COLOR(0, 0, 0, 0, 1, 1, 1, 10, 2, 0, 21), + MDP_COLOR_FULLG10 = MDP_COLOR_FULLG10_BGGR, + + /* MDP_COLOR_FULLG12 */ + MDP_COLOR_FULLG12_RGGB = MDP_COLOR(0, 0, 0, 0, 1, 0, 0, 12, 2, 0, 21), + MDP_COLOR_FULLG12_GRBG = MDP_COLOR(0, 0, 0, 0, 1, 0, 1, 12, 2, 0, 21), + MDP_COLOR_FULLG12_GBRG = MDP_COLOR(0, 0, 0, 0, 1, 1, 0, 12, 2, 0, 21), + MDP_COLOR_FULLG12_BGGR = MDP_COLOR(0, 0, 0, 0, 1, 1, 1, 12, 2, 0, 21), + MDP_COLOR_FULLG12 = MDP_COLOR_FULLG12_BGGR, + + /* MDP_COLOR_FULLG14 */ + MDP_COLOR_FULLG14_RGGB = MDP_COLOR(0, 0, 0, 0, 1, 0, 0, 14, 2, 0, 21), + MDP_COLOR_FULLG14_GRBG = MDP_COLOR(0, 0, 0, 0, 1, 0, 1, 14, 2, 0, 21), + MDP_COLOR_FULLG14_GBRG = MDP_COLOR(0, 0, 0, 0, 1, 1, 0, 14, 2, 0, 21), + MDP_COLOR_FULLG14_BGGR = MDP_COLOR(0, 0, 0, 0, 1, 1, 1, 14, 2, 0, 21), + MDP_COLOR_FULLG14 = MDP_COLOR_FULLG14_BGGR, + + MDP_COLOR_UFO10 = MDP_COLOR(0, 0, 0, 0, 1, 0, 0, 10, 2, 0, 24), + + /* MDP_COLOR_BAYER8 */ + MDP_COLOR_BAYER8_RGGB = MDP_COLOR(0, 0, 0, 0, 1, 0, 0, 8, 2, 0, 20), + MDP_COLOR_BAYER8_GRBG = MDP_COLOR(0, 0, 0, 0, 1, 0, 1, 8, 2, 0, 20), + MDP_COLOR_BAYER8_GBRG = MDP_COLOR(0, 0, 0, 0, 1, 1, 0, 8, 2, 0, 20), + MDP_COLOR_BAYER8_BGGR = MDP_COLOR(0, 0, 0, 0, 1, 1, 1, 8, 2, 0, 20), + MDP_COLOR_BAYER8 = MDP_COLOR_BAYER8_BGGR, + + /* MDP_COLOR_BAYER10 */ + MDP_COLOR_BAYER10_RGGB = MDP_COLOR(0, 0, 0, 0, 1, 0, 0, 10, 2, 0, 20), + MDP_COLOR_BAYER10_GRBG = MDP_COLOR(0, 0, 0, 0, 1, 0, 1, 10, 2, 0, 20), + MDP_COLOR_BAYER10_GBRG = MDP_COLOR(0, 0, 0, 0, 1, 1, 0, 10, 2, 0, 20), + MDP_COLOR_BAYER10_BGGR = MDP_COLOR(0, 0, 0, 0, 1, 1, 1, 10, 2, 0, 20), + MDP_COLOR_BAYER10 = MDP_COLOR_BAYER10_BGGR, + + /* MDP_COLOR_BAYER12 */ + MDP_COLOR_BAYER12_RGGB = MDP_COLOR(0, 0, 0, 0, 1, 0, 0, 12, 2, 0, 20), + MDP_COLOR_BAYER12_GRBG = MDP_COLOR(0, 0, 0, 0, 1, 0, 1, 12, 2, 0, 20), + MDP_COLOR_BAYER12_GBRG = MDP_COLOR(0, 0, 0, 0, 1, 1, 0, 12, 2, 0, 20), + MDP_COLOR_BAYER12_BGGR = MDP_COLOR(0, 0, 0, 0, 1, 1, 1, 12, 2, 0, 20), + MDP_COLOR_BAYER12 = MDP_COLOR_BAYER12_BGGR, + + /* MDP_COLOR_BAYER14 */ + MDP_COLOR_BAYER14_RGGB = MDP_COLOR(0, 0, 0, 0, 1, 0, 0, 14, 2, 0, 20), + MDP_COLOR_BAYER14_GRBG = MDP_COLOR(0, 0, 0, 0, 1, 0, 1, 14, 2, 0, 20), + MDP_COLOR_BAYER14_GBRG = MDP_COLOR(0, 0, 0, 0, 1, 1, 0, 14, 2, 0, 20), + MDP_COLOR_BAYER14_BGGR = MDP_COLOR(0, 0, 0, 0, 1, 1, 1, 14, 2, 0, 20), + MDP_COLOR_BAYER14 = MDP_COLOR_BAYER14_BGGR, + + MDP_COLOR_RGB48 = MDP_COLOR(0, 0, 0, 0, 1, 0, 0, 48, 0, 0, 23), + /* For bayer+mono raw-16 */ + MDP_COLOR_RGB565_RAW = MDP_COLOR(0, 0, 0, 0, 1, 0, 0, 16, 2, 0, 0), + + MDP_COLOR_BAYER8_UNPAK = MDP_COLOR(0, 0, 0, 0, 1, 0, 0, 8, 2, 0, 22), + MDP_COLOR_BAYER10_UNPAK = MDP_COLOR(0, 0, 0, 0, 1, 0, 0, 10, 2, 0, 22), + MDP_COLOR_BAYER12_UNPAK = MDP_COLOR(0, 0, 0, 0, 1, 0, 0, 12, 2, 0, 22), + MDP_COLOR_BAYER14_UNPAK = MDP_COLOR(0, 0, 0, 0, 1, 0, 0, 14, 2, 0, 22), + + /* Unified formats */ + MDP_COLOR_GREY = MDP_COLOR(0, 0, 0, 0, 1, 0, 0, 8, 1, 0, 7), + + MDP_COLOR_RGB565 = MDP_COLOR(0, 0, 0, 0, 1, 0, 0, 16, 0, 0, 0), + MDP_COLOR_BGR565 = MDP_COLOR(0, 0, 0, 0, 1, 0, 0, 16, 0, 1, 0), + MDP_COLOR_RGB888 = MDP_COLOR(0, 0, 0, 0, 1, 0, 0, 24, 0, 1, 1), + MDP_COLOR_BGR888 = MDP_COLOR(0, 0, 0, 0, 1, 0, 0, 24, 0, 0, 1), + MDP_COLOR_RGBA8888 = MDP_COLOR(0, 0, 0, 0, 1, 0, 0, 32, 0, 1, 2), + MDP_COLOR_BGRA8888 = MDP_COLOR(0, 0, 0, 0, 1, 0, 0, 32, 0, 0, 2), + MDP_COLOR_ARGB8888 = MDP_COLOR(0, 0, 0, 0, 1, 0, 0, 32, 0, 1, 3), + MDP_COLOR_ABGR8888 = MDP_COLOR(0, 0, 0, 0, 1, 0, 0, 32, 0, 0, 3), + + MDP_COLOR_UYVY = MDP_COLOR(0, 0, 0, 0, 1, 1, 0, 16, 1, 0, 4), + MDP_COLOR_VYUY = MDP_COLOR(0, 0, 0, 0, 1, 1, 0, 16, 1, 1, 4), + MDP_COLOR_YUYV = MDP_COLOR(0, 0, 0, 0, 1, 1, 0, 16, 1, 0, 5), + MDP_COLOR_YVYU = MDP_COLOR(0, 0, 0, 0, 1, 1, 0, 16, 1, 1, 5), + + MDP_COLOR_I420 = MDP_COLOR(0, 0, 0, 0, 3, 1, 1, 8, 1, 0, 8), + MDP_COLOR_YV12 = MDP_COLOR(0, 0, 0, 0, 3, 1, 1, 8, 1, 1, 8), + MDP_COLOR_I422 = MDP_COLOR(0, 0, 0, 0, 3, 1, 0, 8, 1, 0, 9), + MDP_COLOR_YV16 = MDP_COLOR(0, 0, 0, 0, 3, 1, 0, 8, 1, 1, 9), + MDP_COLOR_I444 = MDP_COLOR(0, 0, 0, 0, 3, 0, 0, 8, 1, 0, 10), + MDP_COLOR_YV24 = MDP_COLOR(0, 0, 0, 0, 3, 0, 0, 8, 1, 1, 10), + + MDP_COLOR_NV12 = MDP_COLOR(0, 0, 0, 0, 2, 1, 1, 8, 1, 0, 12), + MDP_COLOR_NV21 = MDP_COLOR(0, 0, 0, 0, 2, 1, 1, 8, 1, 1, 12), + MDP_COLOR_NV16 = MDP_COLOR(0, 0, 0, 0, 2, 1, 0, 8, 1, 0, 13), + MDP_COLOR_NV61 = MDP_COLOR(0, 0, 0, 0, 2, 1, 0, 8, 1, 1, 13), + MDP_COLOR_NV24 = MDP_COLOR(0, 0, 0, 0, 2, 0, 0, 8, 1, 0, 14), + MDP_COLOR_NV42 = MDP_COLOR(0, 0, 0, 0, 2, 0, 0, 8, 1, 1, 14), + + /* MediaTek proprietary formats */ + /* UFO encoded block mode */ + MDP_COLOR_420_BLK_UFO = MDP_COLOR(0, 0, 0, 5, 2, 1, 1, 256, 1, 0, 12), + /* Block mode */ + MDP_COLOR_420_BLK = MDP_COLOR(0, 0, 0, 1, 2, 1, 1, 256, 1, 0, 12), + /* Block mode + field mode */ + MDP_COLOR_420_BLKI = MDP_COLOR(0, 0, 0, 3, 2, 1, 1, 256, 1, 0, 12), + /* Block mode */ + MDP_COLOR_422_BLK = MDP_COLOR(0, 0, 0, 1, 1, 1, 0, 512, 1, 0, 4), + + MDP_COLOR_IYU2 = MDP_COLOR(0, 0, 0, 0, 1, 0, 0, 24, 1, 0, 25), + MDP_COLOR_YUV444 = MDP_COLOR(0, 0, 0, 0, 1, 0, 0, 24, 1, 0, 30), + + /* Packed 10-bit formats */ + MDP_COLOR_RGBA1010102 = MDP_COLOR(0, 1, 0, 0, 1, 0, 0, 32, 0, 1, 2), + MDP_COLOR_BGRA1010102 = MDP_COLOR(0, 1, 0, 0, 1, 0, 0, 32, 0, 0, 2), + /* Packed 10-bit UYVY */ + MDP_COLOR_UYVY_10P = MDP_COLOR(0, 1, 0, 0, 1, 1, 0, 20, 1, 0, 4), + /* Packed 10-bit NV21 */ + MDP_COLOR_NV21_10P = MDP_COLOR(0, 1, 0, 0, 2, 1, 1, 10, 1, 1, 12), + /* 10-bit block mode */ + MDP_COLOR_420_BLK_10_H = MDP_COLOR(0, 1, 0, 1, 2, 1, 1, 320, 1, 0, 12), + /* 10-bit HEVC tile mode */ + MDP_COLOR_420_BLK_10_V = MDP_COLOR(0, 1, 1, 1, 2, 1, 1, 320, 1, 0, 12), + /* UFO encoded 10-bit block mode */ + MDP_COLOR_420_BLK_U10_H = MDP_COLOR(0, 1, 0, 5, 2, 1, 1, 320, 1, 0, 12), + /* UFO encoded 10-bit HEVC tile mode */ + MDP_COLOR_420_BLK_U10_V = MDP_COLOR(0, 1, 1, 5, 2, 1, 1, 320, 1, 0, 12), + + /* Loose 10-bit formats */ + MDP_COLOR_UYVY_10L = MDP_COLOR(0, 0, 1, 0, 1, 1, 0, 20, 1, 0, 4), + MDP_COLOR_VYUY_10L = MDP_COLOR(0, 0, 1, 0, 1, 1, 0, 20, 1, 1, 4), + MDP_COLOR_YUYV_10L = MDP_COLOR(0, 0, 1, 0, 1, 1, 0, 20, 1, 0, 5), + MDP_COLOR_YVYU_10L = MDP_COLOR(0, 0, 1, 0, 1, 1, 0, 20, 1, 1, 5), + MDP_COLOR_NV12_10L = MDP_COLOR(0, 0, 1, 0, 2, 1, 1, 10, 1, 0, 12), + MDP_COLOR_NV21_10L = MDP_COLOR(0, 0, 1, 0, 2, 1, 1, 10, 1, 1, 12), + MDP_COLOR_NV16_10L = MDP_COLOR(0, 0, 1, 0, 2, 1, 0, 10, 1, 0, 13), + MDP_COLOR_NV61_10L = MDP_COLOR(0, 0, 1, 0, 2, 1, 0, 10, 1, 1, 13), + MDP_COLOR_YV12_10L = MDP_COLOR(0, 0, 1, 0, 3, 1, 1, 10, 1, 1, 8), + MDP_COLOR_I420_10L = MDP_COLOR(0, 0, 1, 0, 3, 1, 1, 10, 1, 0, 8), +}; + +static inline bool MDP_COLOR_IS_UV_COPLANE(enum mdp_color c) +{ + return (MDP_COLOR_GET_PLANE_COUNT(c) == 2 && MDP_COLOR_IS_YUV(c)); +} + +/* Minimum Y stride that is accepted by MDP HW */ +static inline u32 mdp_color_get_min_y_stride(enum mdp_color c, u32 width) +{ + return ((MDP_COLOR_BITS_PER_PIXEL(c) * width) + 4) >> 3; +} + +/* Minimum UV stride that is accepted by MDP HW */ +static inline u32 mdp_color_get_min_uv_stride(enum mdp_color c, u32 width) +{ + u32 min_stride; + + if (MDP_COLOR_GET_PLANE_COUNT(c) == 1) + return 0; + min_stride = mdp_color_get_min_y_stride(c, width) + >> MDP_COLOR_GET_H_SUBSAMPLE(c); + if (MDP_COLOR_IS_UV_COPLANE(c) && !MDP_COLOR_IS_BLOCK_MODE(c)) + min_stride = min_stride * 2; + return min_stride; +} + +/* Minimum Y plane size that is necessary in buffer */ +static inline u32 mdp_color_get_min_y_size(enum mdp_color c, + u32 width, u32 height) +{ + if (MDP_COLOR_IS_BLOCK_MODE(c)) + return ((MDP_COLOR_BITS_PER_PIXEL(c) * width) >> 8) * height; + return mdp_color_get_min_y_stride(c, width) * height; +} + +/* Minimum UV plane size that is necessary in buffer */ +static inline u32 mdp_color_get_min_uv_size(enum mdp_color c, + u32 width, u32 height) +{ + height = height >> MDP_COLOR_GET_V_SUBSAMPLE(c); + if (MDP_COLOR_IS_BLOCK_MODE(c) && (MDP_COLOR_GET_PLANE_COUNT(c) > 1)) + return ((MDP_COLOR_BITS_PER_PIXEL(c) * width) >> 8) * height; + return mdp_color_get_min_uv_stride(c, width) * height; +} + +/* Combine colorspace, xfer_func, ycbcr_encoding, and quantization */ +enum mdp_ycbcr_profile { + /* V4L2_YCBCR_ENC_601 and V4L2_QUANTIZATION_LIM_RANGE */ + MDP_YCBCR_PROFILE_BT601, + /* V4L2_YCBCR_ENC_709 and V4L2_QUANTIZATION_LIM_RANGE */ + MDP_YCBCR_PROFILE_BT709, + /* V4L2_YCBCR_ENC_601 and V4L2_QUANTIZATION_FULL_RANGE */ + MDP_YCBCR_PROFILE_JPEG, + MDP_YCBCR_PROFILE_FULL_BT601 = MDP_YCBCR_PROFILE_JPEG, + + /* Colorspaces not support for capture */ + /* V4L2_YCBCR_ENC_BT2020 and V4L2_QUANTIZATION_LIM_RANGE */ + MDP_YCBCR_PROFILE_BT2020, + /* V4L2_YCBCR_ENC_709 and V4L2_QUANTIZATION_FULL_RANGE */ + MDP_YCBCR_PROFILE_FULL_BT709, + /* V4L2_YCBCR_ENC_BT2020 and V4L2_QUANTIZATION_FULL_RANGE */ + MDP_YCBCR_PROFILE_FULL_BT2020, +}; + +#define MDP_FMT_FLAG_OUTPUT BIT(0) +#define MDP_FMT_FLAG_CAPTURE BIT(1) + +struct mdp_format { + u32 pixelformat; + u32 mdp_color; + u8 depth[VIDEO_MAX_PLANES]; + u8 row_depth[VIDEO_MAX_PLANES]; + u8 num_planes; + u8 walign; + u8 halign; + u8 salign; + u32 flags; +}; + +struct mdp_pix_limit { + u32 wmin; + u32 hmin; + u32 wmax; + u32 hmax; +}; + +struct mdp_limit { + struct mdp_pix_limit out_limit; + struct mdp_pix_limit cap_limit; + u32 h_scale_up_max; + u32 v_scale_up_max; + u32 h_scale_down_max; + u32 v_scale_down_max; +}; + +enum mdp_stream_type { + MDP_STREAM_TYPE_UNKNOWN, + MDP_STREAM_TYPE_BITBLT, + MDP_STREAM_TYPE_GPU_BITBLT, + MDP_STREAM_TYPE_DUAL_BITBLT, + MDP_STREAM_TYPE_2ND_BITBLT, + MDP_STREAM_TYPE_ISP_IC, + MDP_STREAM_TYPE_ISP_VR, + MDP_STREAM_TYPE_ISP_ZSD, + MDP_STREAM_TYPE_ISP_IP, + MDP_STREAM_TYPE_ISP_VSS, + MDP_STREAM_TYPE_ISP_ZSD_SLOW, + MDP_STREAM_TYPE_WPE, + MDP_STREAM_TYPE_WPE2, +}; + +struct mdp_crop { + struct v4l2_rect c; + struct v4l2_fract left_subpix; + struct v4l2_fract top_subpix; + struct v4l2_fract width_subpix; + struct v4l2_fract height_subpix; +}; + +struct mdp_frame { + struct v4l2_format format; + const struct mdp_format *mdp_fmt; + u32 ycbcr_prof; /* enum mdp_ycbcr_profile */ + u32 usage; /* enum mdp_buffer_usage */ + struct mdp_crop crop; + struct v4l2_rect compose; + s32 rotation; + u32 hflip:1; + u32 vflip:1; + u32 hdr:1; + u32 dre:1; + u32 sharpness:1; + u32 dither:1; +}; + +static inline bool mdp_target_is_crop(u32 target) +{ + return (target == V4L2_SEL_TGT_CROP) || + (target == V4L2_SEL_TGT_CROP_DEFAULT) || + (target == V4L2_SEL_TGT_CROP_BOUNDS); +} + +static inline bool mdp_target_is_compose(u32 target) +{ + return (target == V4L2_SEL_TGT_COMPOSE) || + (target == V4L2_SEL_TGT_COMPOSE_DEFAULT) || + (target == V4L2_SEL_TGT_COMPOSE_BOUNDS); +} + +#define MDP_MAX_CAPTURES IMG_MAX_HW_OUTPUTS + +#define MDP_VPU_INIT BIT(0) +#define MDP_M2M_CTX_ERROR BIT(1) + +struct mdp_frameparam { + struct list_head list; + struct mdp_m2m_ctx *ctx; + atomic_t state; + const struct mdp_limit *limit; + u32 type; /* enum mdp_stream_type */ + u32 frame_no; + struct mdp_frame output; + struct mdp_frame captures[MDP_MAX_CAPTURES]; + u32 num_captures; + enum v4l2_colorspace colorspace; + enum v4l2_ycbcr_encoding ycbcr_enc; + enum v4l2_xfer_func xfer_func; + enum v4l2_quantization quant; +}; + +struct mdp_dev; + +int mdp_enum_fmt_mplane(struct mdp_dev *mdp, struct v4l2_fmtdesc *f); +const struct mdp_format *mdp_try_fmt_mplane(struct mdp_dev *mdp, + struct v4l2_format *f, + struct mdp_frameparam *param, + u32 ctx_id); +enum mdp_ycbcr_profile mdp_map_ycbcr_prof_mplane(struct v4l2_format *f, + u32 mdp_color); +int mdp_try_crop(struct mdp_m2m_ctx *ctx, struct v4l2_rect *r, + const struct v4l2_selection *s, struct mdp_frame *frame); +int mdp_check_scaling_ratio(const struct v4l2_rect *crop, + const struct v4l2_rect *compose, s32 rotation, + const struct mdp_limit *limit); +void mdp_set_src_config(struct img_input *in, + struct mdp_frame *frame, struct vb2_buffer *vb); +void mdp_set_dst_config(struct img_output *out, + struct mdp_frame *frame, struct vb2_buffer *vb); +int mdp_frameparam_init(struct mdp_dev *mdp, struct mdp_frameparam *param); + +#endif /* __MTK_MDP3_REGS_H__ */ diff --git a/drivers/media/platform/mediatek/mdp3/mtk-mdp3-type.h b/drivers/media/platform/mediatek/mdp3/mtk-mdp3-type.h new file mode 100644 index 0000000000..ae03968061 --- /dev/null +++ b/drivers/media/platform/mediatek/mdp3/mtk-mdp3-type.h @@ -0,0 +1,53 @@ +/* SPDX-License-Identifier: GPL-2.0-only */ +/* + * Copyright (c) 2023 MediaTek Inc. + * Author: Ping-Hsun Wu <ping-hsun.wu@mediatek.com> + */ + +#ifndef __MTK_MDP3_TYPE_H__ +#define __MTK_MDP3_TYPE_H__ + +#include <linux/types.h> + +#define IMG_MAX_HW_INPUTS 3 +#define IMG_MAX_HW_OUTPUTS 4 +#define IMG_MAX_PLANES 3 +#define IMG_MAX_COMPONENTS 20 + +struct img_crop { + s32 left; + s32 top; + u32 width; + u32 height; + u32 left_subpix; + u32 top_subpix; + u32 width_subpix; + u32 height_subpix; +} __packed; + +struct img_region { + s32 left; + s32 right; + s32 top; + s32 bottom; +} __packed; + +struct img_offset { + s32 left; + s32 top; + u32 left_subpix; + u32 top_subpix; +} __packed; + +struct img_mux { + u32 reg; + u32 value; + u32 subsys_id; +} __packed; + +struct img_mmsys_ctrl { + struct img_mux sets[IMG_MAX_COMPONENTS * 2]; + u32 num_sets; +} __packed; + +#endif /* __MTK_MDP3_TYPE_H__ */ diff --git a/drivers/media/platform/mediatek/mdp3/mtk-mdp3-vpu.c b/drivers/media/platform/mediatek/mdp3/mtk-mdp3-vpu.c new file mode 100644 index 0000000000..49fc2e9d45 --- /dev/null +++ b/drivers/media/platform/mediatek/mdp3/mtk-mdp3-vpu.c @@ -0,0 +1,287 @@ +// SPDX-License-Identifier: GPL-2.0-only +/* + * Copyright (c) 2022 MediaTek Inc. + * Author: Ping-Hsun Wu <ping-hsun.wu@mediatek.com> + */ + +#include <linux/remoteproc.h> +#include <linux/remoteproc/mtk_scp.h> +#include "mtk-mdp3-vpu.h" +#include "mtk-mdp3-core.h" + +#define MDP_VPU_MESSAGE_TIMEOUT 500U + +static inline struct mdp_dev *vpu_to_mdp(struct mdp_vpu_dev *vpu) +{ + return container_of(vpu, struct mdp_dev, vpu); +} + +static int mdp_vpu_shared_mem_alloc(struct mdp_vpu_dev *vpu) +{ + struct device *dev; + + if (IS_ERR_OR_NULL(vpu)) + goto err_return; + + dev = scp_get_device(vpu->scp); + + if (!vpu->param) { + vpu->param = dma_alloc_wc(dev, vpu->param_size, + &vpu->param_addr, GFP_KERNEL); + if (!vpu->param) + goto err_return; + } + + if (!vpu->work) { + vpu->work = dma_alloc_wc(dev, vpu->work_size, + &vpu->work_addr, GFP_KERNEL); + if (!vpu->work) + goto err_free_param; + } + + if (!vpu->config) { + vpu->config = dma_alloc_wc(dev, vpu->config_size, + &vpu->config_addr, GFP_KERNEL); + if (!vpu->config) + goto err_free_work; + } + + return 0; + +err_free_work: + dma_free_wc(dev, vpu->work_size, vpu->work, vpu->work_addr); + vpu->work = NULL; +err_free_param: + dma_free_wc(dev, vpu->param_size, vpu->param, vpu->param_addr); + vpu->param = NULL; +err_return: + return -ENOMEM; +} + +void mdp_vpu_shared_mem_free(struct mdp_vpu_dev *vpu) +{ + struct device *dev; + + if (IS_ERR_OR_NULL(vpu)) + return; + + dev = scp_get_device(vpu->scp); + + if (vpu->param && vpu->param_addr) + dma_free_wc(dev, vpu->param_size, vpu->param, vpu->param_addr); + + if (vpu->work && vpu->work_addr) + dma_free_wc(dev, vpu->work_size, vpu->work, vpu->work_addr); + + if (vpu->config && vpu->config_addr) + dma_free_wc(dev, vpu->config_size, vpu->config, vpu->config_addr); +} + +static void mdp_vpu_ipi_handle_init_ack(void *data, unsigned int len, + void *priv) +{ + struct mdp_ipi_init_msg *msg = (struct mdp_ipi_init_msg *)data; + struct mdp_vpu_dev *vpu = + (struct mdp_vpu_dev *)(unsigned long)msg->drv_data; + + if (!vpu->work_size) + vpu->work_size = msg->work_size; + + vpu->status = msg->status; + complete(&vpu->ipi_acked); +} + +static void mdp_vpu_ipi_handle_deinit_ack(void *data, unsigned int len, + void *priv) +{ + struct mdp_ipi_deinit_msg *msg = (struct mdp_ipi_deinit_msg *)data; + struct mdp_vpu_dev *vpu = + (struct mdp_vpu_dev *)(unsigned long)msg->drv_data; + + vpu->status = msg->status; + complete(&vpu->ipi_acked); +} + +static void mdp_vpu_ipi_handle_frame_ack(void *data, unsigned int len, + void *priv) +{ + struct img_sw_addr *addr = (struct img_sw_addr *)data; + struct img_ipi_frameparam *param = + (struct img_ipi_frameparam *)(unsigned long)addr->va; + struct mdp_vpu_dev *vpu = + (struct mdp_vpu_dev *)(unsigned long)param->drv_data; + + if (param->state) { + struct mdp_dev *mdp = vpu_to_mdp(vpu); + + dev_err(&mdp->pdev->dev, "VPU MDP failure:%d\n", param->state); + } + vpu->status = param->state; + complete(&vpu->ipi_acked); +} + +int mdp_vpu_register(struct mdp_dev *mdp) +{ + int err; + struct mtk_scp *scp = mdp->scp; + struct device *dev = &mdp->pdev->dev; + + err = scp_ipi_register(scp, SCP_IPI_MDP_INIT, + mdp_vpu_ipi_handle_init_ack, NULL); + if (err) { + dev_err(dev, "scp_ipi_register failed %d\n", err); + goto err_ipi_init; + } + err = scp_ipi_register(scp, SCP_IPI_MDP_DEINIT, + mdp_vpu_ipi_handle_deinit_ack, NULL); + if (err) { + dev_err(dev, "scp_ipi_register failed %d\n", err); + goto err_ipi_deinit; + } + err = scp_ipi_register(scp, SCP_IPI_MDP_FRAME, + mdp_vpu_ipi_handle_frame_ack, NULL); + if (err) { + dev_err(dev, "scp_ipi_register failed %d\n", err); + goto err_ipi_frame; + } + return 0; + +err_ipi_frame: + scp_ipi_unregister(scp, SCP_IPI_MDP_DEINIT); +err_ipi_deinit: + scp_ipi_unregister(scp, SCP_IPI_MDP_INIT); +err_ipi_init: + + return err; +} + +void mdp_vpu_unregister(struct mdp_dev *mdp) +{ + scp_ipi_unregister(mdp->scp, SCP_IPI_MDP_INIT); + scp_ipi_unregister(mdp->scp, SCP_IPI_MDP_DEINIT); + scp_ipi_unregister(mdp->scp, SCP_IPI_MDP_FRAME); +} + +static int mdp_vpu_sendmsg(struct mdp_vpu_dev *vpu, enum scp_ipi_id id, + void *buf, unsigned int len) +{ + struct mdp_dev *mdp = vpu_to_mdp(vpu); + unsigned int t = MDP_VPU_MESSAGE_TIMEOUT; + int ret; + + if (!vpu->scp) { + dev_dbg(&mdp->pdev->dev, "vpu scp is NULL"); + return -EINVAL; + } + ret = scp_ipi_send(vpu->scp, id, buf, len, 2000); + + if (ret) { + dev_err(&mdp->pdev->dev, "scp_ipi_send failed %d\n", ret); + return -EPERM; + } + ret = wait_for_completion_timeout(&vpu->ipi_acked, + msecs_to_jiffies(t)); + if (!ret) + ret = -ETIME; + else if (vpu->status) + ret = -EINVAL; + else + ret = 0; + return ret; +} + +int mdp_vpu_dev_init(struct mdp_vpu_dev *vpu, struct mtk_scp *scp, + struct mutex *lock) +{ + struct mdp_ipi_init_msg msg = { + .drv_data = (unsigned long)vpu, + }; + struct mdp_dev *mdp = vpu_to_mdp(vpu); + int err; + + init_completion(&vpu->ipi_acked); + vpu->scp = scp; + vpu->lock = lock; + vpu->work_size = 0; + err = mdp_vpu_sendmsg(vpu, SCP_IPI_MDP_INIT, &msg, sizeof(msg)); + if (err) + goto err_work_size; + /* vpu work_size was set in mdp_vpu_ipi_handle_init_ack */ + + mutex_lock(vpu->lock); + vpu->work_size = ALIGN(vpu->work_size, 64); + vpu->param_size = ALIGN(sizeof(struct img_ipi_frameparam), 64); + vpu->config_size = ALIGN(sizeof(struct img_config), 64); + err = mdp_vpu_shared_mem_alloc(vpu); + mutex_unlock(vpu->lock); + if (err) { + dev_err(&mdp->pdev->dev, "VPU memory alloc fail!"); + goto err_mem_alloc; + } + + dev_dbg(&mdp->pdev->dev, + "VPU param:%pK pa:%pad sz:%zx, work:%pK pa:%pad sz:%zx, config:%pK pa:%pad sz:%zx", + vpu->param, &vpu->param_addr, vpu->param_size, + vpu->work, &vpu->work_addr, vpu->work_size, + vpu->config, &vpu->config_addr, vpu->config_size); + + msg.work_addr = vpu->work_addr; + msg.work_size = vpu->work_size; + err = mdp_vpu_sendmsg(vpu, SCP_IPI_MDP_INIT, &msg, sizeof(msg)); + if (err) + goto err_work_size; + + return 0; + +err_work_size: + switch (vpu->status) { + case -MDP_IPI_EBUSY: + err = -EBUSY; + break; + case -MDP_IPI_ENOMEM: + err = -ENOSPC; /* -ENOMEM */ + break; + } + return err; +err_mem_alloc: + return err; +} + +int mdp_vpu_dev_deinit(struct mdp_vpu_dev *vpu) +{ + struct mdp_ipi_deinit_msg msg = { + .drv_data = (unsigned long)vpu, + .work_addr = vpu->work_addr, + }; + + return mdp_vpu_sendmsg(vpu, SCP_IPI_MDP_DEINIT, &msg, sizeof(msg)); +} + +int mdp_vpu_process(struct mdp_vpu_dev *vpu, struct img_ipi_frameparam *param) +{ + struct mdp_dev *mdp = vpu_to_mdp(vpu); + struct img_sw_addr addr; + + mutex_lock(vpu->lock); + if (mdp_vpu_shared_mem_alloc(vpu)) { + dev_err(&mdp->pdev->dev, "VPU memory alloc fail!"); + mutex_unlock(vpu->lock); + return -ENOMEM; + } + + memset(vpu->param, 0, vpu->param_size); + memset(vpu->work, 0, vpu->work_size); + memset(vpu->config, 0, vpu->config_size); + + param->self_data.va = (unsigned long)vpu->work; + param->self_data.pa = vpu->work_addr; + param->config_data.va = (unsigned long)vpu->config; + param->config_data.pa = vpu->config_addr; + param->drv_data = (unsigned long)vpu; + memcpy(vpu->param, param, sizeof(*param)); + + addr.pa = vpu->param_addr; + addr.va = (unsigned long)vpu->param; + mutex_unlock(vpu->lock); + return mdp_vpu_sendmsg(vpu, SCP_IPI_MDP_FRAME, &addr, sizeof(addr)); +} diff --git a/drivers/media/platform/mediatek/mdp3/mtk-mdp3-vpu.h b/drivers/media/platform/mediatek/mdp3/mtk-mdp3-vpu.h new file mode 100644 index 0000000000..ad3551bc07 --- /dev/null +++ b/drivers/media/platform/mediatek/mdp3/mtk-mdp3-vpu.h @@ -0,0 +1,63 @@ +/* SPDX-License-Identifier: GPL-2.0-only */ +/* + * Copyright (c) 2022 MediaTek Inc. + * Author: Ping-Hsun Wu <ping-hsun.wu@mediatek.com> + */ + +#ifndef __MTK_MDP3_VPU_H__ +#define __MTK_MDP3_VPU_H__ + +#include <linux/platform_device.h> +#include "mtk-img-ipi.h" + +enum mdp_ipi_result { + MDP_IPI_SUCCESS = 0, + MDP_IPI_ENOMEM = 12, + MDP_IPI_EBUSY = 16, + MDP_IPI_EINVAL = 22, + MDP_IPI_EMINST = 24, + MDP_IPI_ERANGE = 34, + MDP_IPI_NR_ERRNO, + + MDP_IPI_EOTHER = MDP_IPI_NR_ERRNO, + MDP_IPI_PATH_CANT_MERGE, + MDP_IPI_OP_FAIL, +}; + +struct mdp_ipi_init_msg { + u32 status; + u64 drv_data; + u32 work_addr; /* [in] working buffer address */ + u32 work_size; /* [in] working buffer size */ +} __packed; + +struct mdp_ipi_deinit_msg { + u32 status; + u64 drv_data; + u32 work_addr; +} __packed; + +struct mdp_vpu_dev { + /* synchronization protect for accessing vpu working buffer info */ + struct mutex *lock; + struct mtk_scp *scp; + struct completion ipi_acked; + void *param; + dma_addr_t param_addr; + size_t param_size; + void *work; + dma_addr_t work_addr; + size_t work_size; + void *config; + dma_addr_t config_addr; + size_t config_size; + u32 status; +}; + +void mdp_vpu_shared_mem_free(struct mdp_vpu_dev *vpu); +int mdp_vpu_dev_init(struct mdp_vpu_dev *vpu, struct mtk_scp *scp, + struct mutex *lock /* for sync */); +int mdp_vpu_dev_deinit(struct mdp_vpu_dev *vpu); +int mdp_vpu_process(struct mdp_vpu_dev *vpu, struct img_ipi_frameparam *param); + +#endif /* __MTK_MDP3_VPU_H__ */ diff --git a/drivers/media/platform/mediatek/vcodec/Kconfig b/drivers/media/platform/mediatek/vcodec/Kconfig new file mode 100644 index 0000000000..74b00eb1bc --- /dev/null +++ b/drivers/media/platform/mediatek/vcodec/Kconfig @@ -0,0 +1,37 @@ +# SPDX-License-Identifier: GPL-2.0-only +config VIDEO_MEDIATEK_VCODEC_SCP + bool + +config VIDEO_MEDIATEK_VCODEC_VPU + bool + +config VIDEO_MEDIATEK_VCODEC + tristate "Mediatek Video Codec driver" + depends on V4L_MEM2MEM_DRIVERS + depends on MTK_IOMMU || COMPILE_TEST + depends on VIDEO_DEV + depends on ARCH_MEDIATEK || COMPILE_TEST + depends on VIDEO_MEDIATEK_VPU || MTK_SCP + # The two following lines ensure we have the same state ("m" or "y") as + # our dependencies, to avoid missing symbols during link. + depends on VIDEO_MEDIATEK_VPU || !VIDEO_MEDIATEK_VPU + depends on MTK_SCP || !MTK_SCP + depends on MTK_SMI || (COMPILE_TEST && MTK_SMI=n) + select VIDEOBUF2_DMA_CONTIG + select V4L2_MEM2MEM_DEV + select VIDEO_MEDIATEK_VCODEC_VPU if VIDEO_MEDIATEK_VPU + select VIDEO_MEDIATEK_VCODEC_SCP if MTK_SCP + select V4L2_H264 + select V4L2_VP9 + select MEDIA_CONTROLLER + select MEDIA_CONTROLLER_REQUEST_API + help + Mediatek video codec driver provides HW capability to + encode and decode in a range of video formats on MT8173 + and MT8183. + + Note that support for MT8173 requires VIDEO_MEDIATEK_VPU to + also be selected. Support for MT8183 depends on MTK_SCP. + + To compile this driver as modules, choose M here: the + modules will be called mtk-vcodec-dec and mtk-vcodec-enc. diff --git a/drivers/media/platform/mediatek/vcodec/Makefile b/drivers/media/platform/mediatek/vcodec/Makefile new file mode 100644 index 0000000000..014abbfbd9 --- /dev/null +++ b/drivers/media/platform/mediatek/vcodec/Makefile @@ -0,0 +1,5 @@ +# SPDX-License-Identifier: GPL-2.0 + +obj-y += common/ +obj-y += encoder/ +obj-y += decoder/ diff --git a/drivers/media/platform/mediatek/vcodec/common/Makefile b/drivers/media/platform/mediatek/vcodec/common/Makefile new file mode 100644 index 0000000000..d0479914df --- /dev/null +++ b/drivers/media/platform/mediatek/vcodec/common/Makefile @@ -0,0 +1,21 @@ +# SPDX-License-Identifier: GPL-2.0 + +obj-$(CONFIG_VIDEO_MEDIATEK_VCODEC) += mtk-vcodec-common.o + +mtk-vcodec-common-y := mtk_vcodec_intr.o \ + mtk_vcodec_util.o \ + mtk_vcodec_fw.o \ + +ifneq ($(CONFIG_VIDEO_MEDIATEK_VCODEC_VPU),) +mtk-vcodec-common-y += mtk_vcodec_fw_vpu.o +endif + +ifneq ($(CONFIG_VIDEO_MEDIATEK_VCODEC_SCP),) +mtk-vcodec-common-y += mtk_vcodec_fw_scp.o +endif + +ifneq ($(CONFIG_DEBUG_FS),) +obj-$(CONFIG_VIDEO_MEDIATEK_VCODEC) += mtk-vcodec-dbgfs.o + +mtk-vcodec-dbgfs-y := mtk_vcodec_dbgfs.o +endif
\ No newline at end of file diff --git a/drivers/media/platform/mediatek/vcodec/common/mtk_vcodec_cmn_drv.h b/drivers/media/platform/mediatek/vcodec/common/mtk_vcodec_cmn_drv.h new file mode 100644 index 0000000000..6087e27bd6 --- /dev/null +++ b/drivers/media/platform/mediatek/vcodec/common/mtk_vcodec_cmn_drv.h @@ -0,0 +1,147 @@ +/* SPDX-License-Identifier: GPL-2.0 */ +/* + * Copyright (c) 2023 MediaTek Inc. + * Author: Yunfei Dong <yunfei.dong@mediatek.com> + */ + +#ifndef _MTK_VCODEC_COM_DRV_H_ +#define _MTK_VCODEC_COM_DRV_H_ + +#include <linux/platform_device.h> +#include <linux/videodev2.h> +#include <media/v4l2-ctrls.h> +#include <media/v4l2-device.h> +#include <media/v4l2-ioctl.h> +#include <media/v4l2-mem2mem.h> +#include <media/videobuf2-core.h> + +#define MTK_VCODEC_MAX_PLANES 3 + +#define WAIT_INTR_TIMEOUT_MS 1000 + +/* + * enum mtk_q_type - Type of queue + */ +enum mtk_q_type { + MTK_Q_DATA_SRC = 0, + MTK_Q_DATA_DST = 1, +}; + +/* + * enum mtk_hw_reg_idx - MTK hw register base index + */ +enum mtk_hw_reg_idx { + VDEC_SYS, + VDEC_MISC, + VDEC_LD, + VDEC_TOP, + VDEC_CM, + VDEC_AD, + VDEC_AV, + VDEC_PP, + VDEC_HWD, + VDEC_HWQ, + VDEC_HWB, + VDEC_HWG, + NUM_MAX_VDEC_REG_BASE, + /* h264 encoder */ + VENC_SYS = NUM_MAX_VDEC_REG_BASE, + /* vp8 encoder */ + VENC_LT_SYS, + NUM_MAX_VCODEC_REG_BASE +}; + +/* + * struct mtk_vcodec_clk_info - Structure used to store clock name + */ +struct mtk_vcodec_clk_info { + const char *clk_name; + struct clk *vcodec_clk; +}; + +/* + * struct mtk_vcodec_clk - Structure used to store vcodec clock information + */ +struct mtk_vcodec_clk { + struct mtk_vcodec_clk_info *clk_info; + int clk_num; +}; + +/* + * struct mtk_vcodec_pm - Power management data structure + */ +struct mtk_vcodec_pm { + struct mtk_vcodec_clk vdec_clk; + struct mtk_vcodec_clk venc_clk; + struct device *dev; +}; + +/* + * enum mtk_vdec_hw_id - Hardware index used to separate + * different hardware + */ +enum mtk_vdec_hw_id { + MTK_VDEC_CORE, + MTK_VDEC_LAT0, + MTK_VDEC_LAT1, + MTK_VDEC_LAT_SOC, + MTK_VDEC_HW_MAX, +}; + +/** + * enum mtk_instance_state - The state of an MTK Vcodec instance. + * @MTK_STATE_FREE: default state when instance is created + * @MTK_STATE_INIT: vcodec instance is initialized + * @MTK_STATE_HEADER: vdec had sps/pps header parsed or venc + * had sps/pps header encoded + * @MTK_STATE_FLUSH: vdec is flushing. Only used by decoder + * @MTK_STATE_ABORT: vcodec should be aborted + */ +enum mtk_instance_state { + MTK_STATE_FREE = 0, + MTK_STATE_INIT = 1, + MTK_STATE_HEADER = 2, + MTK_STATE_FLUSH = 3, + MTK_STATE_ABORT = 4, +}; + +enum mtk_fmt_type { + MTK_FMT_DEC = 0, + MTK_FMT_ENC = 1, + MTK_FMT_FRAME = 2, +}; + +/* + * struct mtk_video_fmt - Structure used to store information about pixelformats + */ +struct mtk_video_fmt { + u32 fourcc; + enum mtk_fmt_type type; + u32 num_planes; + u32 flags; + struct v4l2_frmsize_stepwise frmsize; +}; + +/* + * struct mtk_q_data - Structure used to store information about queue + */ +struct mtk_q_data { + unsigned int visible_width; + unsigned int visible_height; + unsigned int coded_width; + unsigned int coded_height; + enum v4l2_field field; + unsigned int bytesperline[MTK_VCODEC_MAX_PLANES]; + unsigned int sizeimage[MTK_VCODEC_MAX_PLANES]; + const struct mtk_video_fmt *fmt; +}; + +/* + * enum mtk_instance_type - The type of an MTK Vcodec instance. + */ +enum mtk_instance_type { + MTK_INST_DECODER = 0, + MTK_INST_ENCODER = 1, +}; + +#endif /* _MTK_VCODEC_COM_DRV_H_ */ diff --git a/drivers/media/platform/mediatek/vcodec/common/mtk_vcodec_dbgfs.c b/drivers/media/platform/mediatek/vcodec/common/mtk_vcodec_dbgfs.c new file mode 100644 index 0000000000..5ad3797836 --- /dev/null +++ b/drivers/media/platform/mediatek/vcodec/common/mtk_vcodec_dbgfs.c @@ -0,0 +1,231 @@ +// SPDX-License-Identifier: GPL-2.0 +/* + * Copyright (c) 2023 MediaTek Inc. + * Author: Yunfei Dong <yunfei.dong@mediatek.com> + */ + +#include <linux/debugfs.h> + +#include "mtk_vcodec_dbgfs.h" +#include "../decoder/mtk_vcodec_dec_drv.h" +#include "../encoder/mtk_vcodec_enc_drv.h" +#include "mtk_vcodec_util.h" + +static void mtk_vdec_dbgfs_get_format_type(struct mtk_vcodec_dec_ctx *ctx, char *buf, + int *used, int total) +{ + int curr_len; + + switch (ctx->current_codec) { + case V4L2_PIX_FMT_H264_SLICE: + curr_len = snprintf(buf + *used, total - *used, + "\toutput format: h264 slice\n"); + break; + case V4L2_PIX_FMT_VP8_FRAME: + curr_len = snprintf(buf + *used, total - *used, + "\toutput format: vp8 slice\n"); + break; + case V4L2_PIX_FMT_VP9_FRAME: + curr_len = snprintf(buf + *used, total - *used, + "\toutput format: vp9 slice\n"); + break; + default: + curr_len = snprintf(buf + *used, total - *used, + "\tunsupported output format: 0x%x\n", + ctx->current_codec); + } + *used += curr_len; + + switch (ctx->capture_fourcc) { + case V4L2_PIX_FMT_MM21: + curr_len = snprintf(buf + *used, total - *used, + "\tcapture format: MM21\n"); + break; + case V4L2_PIX_FMT_MT21C: + curr_len = snprintf(buf + *used, total - *used, + "\tcapture format: MT21C\n"); + break; + default: + curr_len = snprintf(buf + *used, total - *used, + "\tunsupported capture format: 0x%x\n", + ctx->capture_fourcc); + } + *used += curr_len; +} + +static void mtk_vdec_dbgfs_get_help(char *buf, int *used, int total) +{ + int curr_len; + + curr_len = snprintf(buf + *used, total - *used, + "help: (1: echo -'info' > vdec 2: cat vdec)\n"); + *used += curr_len; + + curr_len = snprintf(buf + *used, total - *used, + "\t-picinfo: get resolution\n"); + *used += curr_len; + + curr_len = snprintf(buf + *used, total - *used, + "\t-format: get output & capture queue format\n"); + *used += curr_len; +} + +static ssize_t mtk_vdec_dbgfs_write(struct file *filp, const char __user *ubuf, + size_t count, loff_t *ppos) +{ + struct mtk_vcodec_dec_dev *vcodec_dev = filp->private_data; + struct mtk_vcodec_dbgfs *dbgfs = &vcodec_dev->dbgfs; + + mutex_lock(&dbgfs->dbgfs_lock); + dbgfs->buf_size = simple_write_to_buffer(dbgfs->dbgfs_buf, sizeof(dbgfs->dbgfs_buf), + ppos, ubuf, count); + mutex_unlock(&dbgfs->dbgfs_lock); + if (dbgfs->buf_size > 0) + return count; + + return dbgfs->buf_size; +} + +static ssize_t mtk_vdec_dbgfs_read(struct file *filp, char __user *ubuf, + size_t count, loff_t *ppos) +{ + struct mtk_vcodec_dec_dev *vcodec_dev = filp->private_data; + struct mtk_vcodec_dbgfs *dbgfs = &vcodec_dev->dbgfs; + struct mtk_vcodec_dbgfs_inst *dbgfs_inst; + struct mtk_vcodec_dec_ctx *ctx; + int total_len = 200 * (dbgfs->inst_count == 0 ? 1 : dbgfs->inst_count); + int used_len = 0, curr_len, ret; + bool dbgfs_index[MTK_VDEC_DBGFS_MAX] = {0}; + char *buf = kmalloc(total_len, GFP_KERNEL); + + if (!buf) + return -ENOMEM; + + if (strstr(dbgfs->dbgfs_buf, "-help") || dbgfs->buf_size == 1) { + mtk_vdec_dbgfs_get_help(buf, &used_len, total_len); + goto read_buffer; + } + + if (strstr(dbgfs->dbgfs_buf, "-picinfo")) + dbgfs_index[MTK_VDEC_DBGFS_PICINFO] = true; + + if (strstr(dbgfs->dbgfs_buf, "-format")) + dbgfs_index[MTK_VDEC_DBGFS_FORMAT] = true; + + mutex_lock(&dbgfs->dbgfs_lock); + list_for_each_entry(dbgfs_inst, &dbgfs->dbgfs_head, node) { + ctx = dbgfs_inst->vcodec_ctx; + + curr_len = snprintf(buf + used_len, total_len - used_len, + "inst[%d]:\n ", ctx->id); + used_len += curr_len; + + if (dbgfs_index[MTK_VDEC_DBGFS_PICINFO]) { + curr_len = snprintf(buf + used_len, total_len - used_len, + "\treal(%dx%d)=>align(%dx%d)\n", + ctx->picinfo.pic_w, ctx->picinfo.pic_h, + ctx->picinfo.buf_w, ctx->picinfo.buf_h); + used_len += curr_len; + } + + if (dbgfs_index[MTK_VDEC_DBGFS_FORMAT]) + mtk_vdec_dbgfs_get_format_type(ctx, buf, &used_len, total_len); + } + mutex_unlock(&dbgfs->dbgfs_lock); +read_buffer: + ret = simple_read_from_buffer(ubuf, count, ppos, buf, used_len); + kfree(buf); + return ret; +} + +static const struct file_operations vdec_fops = { + .open = simple_open, + .write = mtk_vdec_dbgfs_write, + .read = mtk_vdec_dbgfs_read, +}; + +void mtk_vcodec_dbgfs_create(struct mtk_vcodec_dec_ctx *ctx) +{ + struct mtk_vcodec_dbgfs_inst *dbgfs_inst; + struct mtk_vcodec_dec_dev *vcodec_dev = ctx->dev; + + dbgfs_inst = kzalloc(sizeof(*dbgfs_inst), GFP_KERNEL); + if (!dbgfs_inst) + return; + + list_add_tail(&dbgfs_inst->node, &vcodec_dev->dbgfs.dbgfs_head); + + vcodec_dev->dbgfs.inst_count++; + + dbgfs_inst->inst_id = ctx->id; + dbgfs_inst->vcodec_ctx = ctx; +} +EXPORT_SYMBOL_GPL(mtk_vcodec_dbgfs_create); + +void mtk_vcodec_dbgfs_remove(struct mtk_vcodec_dec_dev *vcodec_dev, int ctx_id) +{ + struct mtk_vcodec_dbgfs_inst *dbgfs_inst; + + list_for_each_entry(dbgfs_inst, &vcodec_dev->dbgfs.dbgfs_head, node) { + if (dbgfs_inst->inst_id == ctx_id) { + vcodec_dev->dbgfs.inst_count--; + list_del(&dbgfs_inst->node); + kfree(dbgfs_inst); + return; + } + } +} +EXPORT_SYMBOL_GPL(mtk_vcodec_dbgfs_remove); + +static void mtk_vcodec_dbgfs_vdec_init(struct mtk_vcodec_dec_dev *vcodec_dev) +{ + struct dentry *vcodec_root; + + vcodec_dev->dbgfs.vcodec_root = debugfs_create_dir("vcodec-dec", NULL); + if (IS_ERR(vcodec_dev->dbgfs.vcodec_root)) + dev_err(&vcodec_dev->plat_dev->dev, "create vcodec dir err:%ld\n", + PTR_ERR(vcodec_dev->dbgfs.vcodec_root)); + + vcodec_root = vcodec_dev->dbgfs.vcodec_root; + debugfs_create_x32("mtk_v4l2_dbg_level", 0644, vcodec_root, &mtk_v4l2_dbg_level); + debugfs_create_x32("mtk_vcodec_dbg", 0644, vcodec_root, &mtk_vcodec_dbg); + + vcodec_dev->dbgfs.inst_count = 0; + INIT_LIST_HEAD(&vcodec_dev->dbgfs.dbgfs_head); + debugfs_create_file("vdec", 0200, vcodec_root, vcodec_dev, &vdec_fops); + mutex_init(&vcodec_dev->dbgfs.dbgfs_lock); +} + +static void mtk_vcodec_dbgfs_venc_init(struct mtk_vcodec_enc_dev *vcodec_dev) +{ + struct dentry *vcodec_root; + + vcodec_dev->dbgfs.vcodec_root = debugfs_create_dir("vcodec-enc", NULL); + if (IS_ERR(vcodec_dev->dbgfs.vcodec_root)) + dev_err(&vcodec_dev->plat_dev->dev, "create venc dir err:%d\n", + IS_ERR(vcodec_dev->dbgfs.vcodec_root)); + + vcodec_root = vcodec_dev->dbgfs.vcodec_root; + debugfs_create_x32("mtk_v4l2_dbg_level", 0644, vcodec_root, &mtk_v4l2_dbg_level); + debugfs_create_x32("mtk_vcodec_dbg", 0644, vcodec_root, &mtk_vcodec_dbg); + + vcodec_dev->dbgfs.inst_count = 0; +} + +void mtk_vcodec_dbgfs_init(void *vcodec_dev, bool is_encode) +{ + if (is_encode) + mtk_vcodec_dbgfs_venc_init(vcodec_dev); + else + mtk_vcodec_dbgfs_vdec_init(vcodec_dev); +} +EXPORT_SYMBOL_GPL(mtk_vcodec_dbgfs_init); + +void mtk_vcodec_dbgfs_deinit(struct mtk_vcodec_dbgfs *dbgfs) +{ + debugfs_remove_recursive(dbgfs->vcodec_root); +} +EXPORT_SYMBOL_GPL(mtk_vcodec_dbgfs_deinit); + +MODULE_LICENSE("GPL v2"); +MODULE_DESCRIPTION("Mediatek video codec driver"); diff --git a/drivers/media/platform/mediatek/vcodec/common/mtk_vcodec_dbgfs.h b/drivers/media/platform/mediatek/vcodec/common/mtk_vcodec_dbgfs.h new file mode 100644 index 0000000000..073d2fedb5 --- /dev/null +++ b/drivers/media/platform/mediatek/vcodec/common/mtk_vcodec_dbgfs.h @@ -0,0 +1,74 @@ +/* SPDX-License-Identifier: GPL-2.0 */ +/* + * Copyright (c) 2023 MediaTek Inc. + * Author: Yunfei Dong <yunfei.dong@mediatek.com> + */ + +#ifndef __MTK_VCODEC_DBGFS_H__ +#define __MTK_VCODEC_DBGFS_H__ + +struct mtk_vcodec_dec_dev; +struct mtk_vcodec_dec_ctx; + +/* + * enum mtk_vdec_dbgfs_log_index - used to get different debug information + */ +enum mtk_vdec_dbgfs_log_index { + MTK_VDEC_DBGFS_PICINFO, + MTK_VDEC_DBGFS_FORMAT, + MTK_VDEC_DBGFS_MAX, +}; + +/** + * struct mtk_vcodec_dbgfs_inst - debugfs information for each inst + * @node: list node for each inst + * @vcodec_ctx: struct mtk_vcodec_dec_ctx + * @inst_id: index of the context that the same with ctx->id + */ +struct mtk_vcodec_dbgfs_inst { + struct list_head node; + struct mtk_vcodec_dec_ctx *vcodec_ctx; + int inst_id; +}; + +/** + * struct mtk_vcodec_dbgfs - dbgfs information + * @dbgfs_head: list head used to link each instance + * @vcodec_root: vcodec dbgfs entry + * @dbgfs_lock: dbgfs lock used to protect dbgfs_buf + * @dbgfs_buf: dbgfs buf used to store dbgfs cmd + * @buf_size: buffer size of dbgfs + * @inst_count: the count of total instance + */ +struct mtk_vcodec_dbgfs { + struct list_head dbgfs_head; + struct dentry *vcodec_root; + struct mutex dbgfs_lock; + char dbgfs_buf[1024]; + int buf_size; + int inst_count; +}; + +#if defined(CONFIG_DEBUG_FS) +void mtk_vcodec_dbgfs_create(struct mtk_vcodec_dec_ctx *ctx); +void mtk_vcodec_dbgfs_remove(struct mtk_vcodec_dec_dev *vcodec_dev, int ctx_id); +void mtk_vcodec_dbgfs_init(void *vcodec_dev, bool is_encode); +void mtk_vcodec_dbgfs_deinit(struct mtk_vcodec_dbgfs *dbgfs); +#else +static inline void mtk_vcodec_dbgfs_create(struct mtk_vcodec_dec_ctx *ctx) +{ +} + +static inline void mtk_vcodec_dbgfs_remove(struct mtk_vcodec_dec_dev *vcodec_dev, int ctx_id) +{ +} + +static inline void mtk_vcodec_dbgfs_init(void *vcodec_dev, bool is_encode) +{ +} + +static inline void mtk_vcodec_dbgfs_deinit(struct mtk_vcodec_dbgfs *dbgfs) +{ +} +#endif +#endif diff --git a/drivers/media/platform/mediatek/vcodec/common/mtk_vcodec_fw.c b/drivers/media/platform/mediatek/vcodec/common/mtk_vcodec_fw.c new file mode 100644 index 0000000000..08949b08fb --- /dev/null +++ b/drivers/media/platform/mediatek/vcodec/common/mtk_vcodec_fw.c @@ -0,0 +1,78 @@ +// SPDX-License-Identifier: GPL-2.0 + +#include "../decoder/mtk_vcodec_dec_drv.h" +#include "../encoder/mtk_vcodec_enc_drv.h" +#include "mtk_vcodec_fw_priv.h" + +struct mtk_vcodec_fw *mtk_vcodec_fw_select(void *priv, enum mtk_vcodec_fw_type type, + enum mtk_vcodec_fw_use fw_use) +{ + struct platform_device *plat_dev; + + if (fw_use == ENCODER) + plat_dev = ((struct mtk_vcodec_enc_dev *)priv)->plat_dev; + else + plat_dev = ((struct mtk_vcodec_dec_dev *)priv)->plat_dev; + + switch (type) { + case VPU: + return mtk_vcodec_fw_vpu_init(priv, fw_use); + case SCP: + return mtk_vcodec_fw_scp_init(priv, fw_use); + default: + dev_err(&plat_dev->dev, "Invalid vcodec fw type"); + return ERR_PTR(-EINVAL); + } +} +EXPORT_SYMBOL_GPL(mtk_vcodec_fw_select); + +void mtk_vcodec_fw_release(struct mtk_vcodec_fw *fw) +{ + fw->ops->release(fw); +} +EXPORT_SYMBOL_GPL(mtk_vcodec_fw_release); + +int mtk_vcodec_fw_load_firmware(struct mtk_vcodec_fw *fw) +{ + return fw->ops->load_firmware(fw); +} +EXPORT_SYMBOL_GPL(mtk_vcodec_fw_load_firmware); + +unsigned int mtk_vcodec_fw_get_vdec_capa(struct mtk_vcodec_fw *fw) +{ + return fw->ops->get_vdec_capa(fw); +} +EXPORT_SYMBOL_GPL(mtk_vcodec_fw_get_vdec_capa); + +unsigned int mtk_vcodec_fw_get_venc_capa(struct mtk_vcodec_fw *fw) +{ + return fw->ops->get_venc_capa(fw); +} +EXPORT_SYMBOL_GPL(mtk_vcodec_fw_get_venc_capa); + +void *mtk_vcodec_fw_map_dm_addr(struct mtk_vcodec_fw *fw, u32 mem_addr) +{ + return fw->ops->map_dm_addr(fw, mem_addr); +} +EXPORT_SYMBOL_GPL(mtk_vcodec_fw_map_dm_addr); + +int mtk_vcodec_fw_ipi_register(struct mtk_vcodec_fw *fw, int id, + mtk_vcodec_ipi_handler handler, + const char *name, void *priv) +{ + return fw->ops->ipi_register(fw, id, handler, name, priv); +} +EXPORT_SYMBOL_GPL(mtk_vcodec_fw_ipi_register); + +int mtk_vcodec_fw_ipi_send(struct mtk_vcodec_fw *fw, int id, void *buf, + unsigned int len, unsigned int wait) +{ + return fw->ops->ipi_send(fw, id, buf, len, wait); +} +EXPORT_SYMBOL_GPL(mtk_vcodec_fw_ipi_send); + +int mtk_vcodec_fw_get_type(struct mtk_vcodec_fw *fw) +{ + return fw->type; +} +EXPORT_SYMBOL_GPL(mtk_vcodec_fw_get_type); diff --git a/drivers/media/platform/mediatek/vcodec/common/mtk_vcodec_fw.h b/drivers/media/platform/mediatek/vcodec/common/mtk_vcodec_fw.h new file mode 100644 index 0000000000..300363a401 --- /dev/null +++ b/drivers/media/platform/mediatek/vcodec/common/mtk_vcodec_fw.h @@ -0,0 +1,44 @@ +/* SPDX-License-Identifier: GPL-2.0 */ + +#ifndef _MTK_VCODEC_FW_H_ +#define _MTK_VCODEC_FW_H_ + +#include <linux/remoteproc.h> +#include <linux/remoteproc/mtk_scp.h> + +#include "../../vpu/mtk_vpu.h" + +struct mtk_vcodec_dec_dev; +struct mtk_vcodec_enc_dev; + +enum mtk_vcodec_fw_type { + VPU, + SCP, +}; + +enum mtk_vcodec_fw_use { + DECODER, + ENCODER, +}; + +struct mtk_vcodec_fw; + +typedef void (*mtk_vcodec_ipi_handler) (void *data, + unsigned int len, void *priv); + +struct mtk_vcodec_fw *mtk_vcodec_fw_select(void *priv, enum mtk_vcodec_fw_type type, + enum mtk_vcodec_fw_use fw_use); +void mtk_vcodec_fw_release(struct mtk_vcodec_fw *fw); + +int mtk_vcodec_fw_load_firmware(struct mtk_vcodec_fw *fw); +unsigned int mtk_vcodec_fw_get_vdec_capa(struct mtk_vcodec_fw *fw); +unsigned int mtk_vcodec_fw_get_venc_capa(struct mtk_vcodec_fw *fw); +void *mtk_vcodec_fw_map_dm_addr(struct mtk_vcodec_fw *fw, u32 mem_addr); +int mtk_vcodec_fw_ipi_register(struct mtk_vcodec_fw *fw, int id, + mtk_vcodec_ipi_handler handler, + const char *name, void *priv); +int mtk_vcodec_fw_ipi_send(struct mtk_vcodec_fw *fw, int id, + void *buf, unsigned int len, unsigned int wait); +int mtk_vcodec_fw_get_type(struct mtk_vcodec_fw *fw); + +#endif /* _MTK_VCODEC_FW_H_ */ diff --git a/drivers/media/platform/mediatek/vcodec/common/mtk_vcodec_fw_priv.h b/drivers/media/platform/mediatek/vcodec/common/mtk_vcodec_fw_priv.h new file mode 100644 index 0000000000..99603accd8 --- /dev/null +++ b/drivers/media/platform/mediatek/vcodec/common/mtk_vcodec_fw_priv.h @@ -0,0 +1,52 @@ +/* SPDX-License-Identifier: GPL-2.0 */ + +#ifndef _MTK_VCODEC_FW_PRIV_H_ +#define _MTK_VCODEC_FW_PRIV_H_ + +#include "mtk_vcodec_fw.h" + +struct mtk_vcodec_dec_dev; +struct mtk_vcodec_enc_dev; + +struct mtk_vcodec_fw { + enum mtk_vcodec_fw_type type; + const struct mtk_vcodec_fw_ops *ops; + struct platform_device *pdev; + struct mtk_scp *scp; + enum mtk_vcodec_fw_use fw_use; +}; + +struct mtk_vcodec_fw_ops { + int (*load_firmware)(struct mtk_vcodec_fw *fw); + unsigned int (*get_vdec_capa)(struct mtk_vcodec_fw *fw); + unsigned int (*get_venc_capa)(struct mtk_vcodec_fw *fw); + void *(*map_dm_addr)(struct mtk_vcodec_fw *fw, u32 dtcm_dmem_addr); + int (*ipi_register)(struct mtk_vcodec_fw *fw, int id, + mtk_vcodec_ipi_handler handler, const char *name, + void *priv); + int (*ipi_send)(struct mtk_vcodec_fw *fw, int id, void *buf, + unsigned int len, unsigned int wait); + void (*release)(struct mtk_vcodec_fw *fw); +}; + +#if IS_ENABLED(CONFIG_VIDEO_MEDIATEK_VCODEC_VPU) +struct mtk_vcodec_fw *mtk_vcodec_fw_vpu_init(void *priv, enum mtk_vcodec_fw_use fw_use); +#else +static inline struct mtk_vcodec_fw * +mtk_vcodec_fw_vpu_init(void *priv, enum mtk_vcodec_fw_use fw_use) +{ + return ERR_PTR(-ENODEV); +} +#endif /* CONFIG_VIDEO_MEDIATEK_VCODEC_VPU */ + +#if IS_ENABLED(CONFIG_VIDEO_MEDIATEK_VCODEC_SCP) +struct mtk_vcodec_fw *mtk_vcodec_fw_scp_init(void *priv, enum mtk_vcodec_fw_use fw_use); +#else +static inline struct mtk_vcodec_fw * +mtk_vcodec_fw_scp_init(void *priv, enum mtk_vcodec_fw_use fw_use) +{ + return ERR_PTR(-ENODEV); +} +#endif /* CONFIG_VIDEO_MEDIATEK_VCODEC_SCP */ + +#endif /* _MTK_VCODEC_FW_PRIV_H_ */ diff --git a/drivers/media/platform/mediatek/vcodec/common/mtk_vcodec_fw_scp.c b/drivers/media/platform/mediatek/vcodec/common/mtk_vcodec_fw_scp.c new file mode 100644 index 0000000000..9e744d07a1 --- /dev/null +++ b/drivers/media/platform/mediatek/vcodec/common/mtk_vcodec_fw_scp.c @@ -0,0 +1,87 @@ +// SPDX-License-Identifier: GPL-2.0 + +#include "../decoder/mtk_vcodec_dec_drv.h" +#include "../encoder/mtk_vcodec_enc_drv.h" +#include "mtk_vcodec_fw_priv.h" + +static int mtk_vcodec_scp_load_firmware(struct mtk_vcodec_fw *fw) +{ + return rproc_boot(scp_get_rproc(fw->scp)); +} + +static unsigned int mtk_vcodec_scp_get_vdec_capa(struct mtk_vcodec_fw *fw) +{ + return scp_get_vdec_hw_capa(fw->scp); +} + +static unsigned int mtk_vcodec_scp_get_venc_capa(struct mtk_vcodec_fw *fw) +{ + return scp_get_venc_hw_capa(fw->scp); +} + +static void *mtk_vcodec_vpu_scp_dm_addr(struct mtk_vcodec_fw *fw, + u32 dtcm_dmem_addr) +{ + return scp_mapping_dm_addr(fw->scp, dtcm_dmem_addr); +} + +static int mtk_vcodec_scp_set_ipi_register(struct mtk_vcodec_fw *fw, int id, + mtk_vcodec_ipi_handler handler, + const char *name, void *priv) +{ + return scp_ipi_register(fw->scp, id, handler, priv); +} + +static int mtk_vcodec_scp_ipi_send(struct mtk_vcodec_fw *fw, int id, void *buf, + unsigned int len, unsigned int wait) +{ + return scp_ipi_send(fw->scp, id, buf, len, wait); +} + +static void mtk_vcodec_scp_release(struct mtk_vcodec_fw *fw) +{ + scp_put(fw->scp); +} + +static const struct mtk_vcodec_fw_ops mtk_vcodec_rproc_msg = { + .load_firmware = mtk_vcodec_scp_load_firmware, + .get_vdec_capa = mtk_vcodec_scp_get_vdec_capa, + .get_venc_capa = mtk_vcodec_scp_get_venc_capa, + .map_dm_addr = mtk_vcodec_vpu_scp_dm_addr, + .ipi_register = mtk_vcodec_scp_set_ipi_register, + .ipi_send = mtk_vcodec_scp_ipi_send, + .release = mtk_vcodec_scp_release, +}; + +struct mtk_vcodec_fw *mtk_vcodec_fw_scp_init(void *priv, enum mtk_vcodec_fw_use fw_use) +{ + struct mtk_vcodec_fw *fw; + struct platform_device *plat_dev; + struct mtk_scp *scp; + + if (fw_use == ENCODER) { + struct mtk_vcodec_enc_dev *enc_dev = priv; + + plat_dev = enc_dev->plat_dev; + } else if (fw_use == DECODER) { + struct mtk_vcodec_dec_dev *dec_dev = priv; + + plat_dev = dec_dev->plat_dev; + } else { + pr_err("Invalid fw_use %d (use a resonable fw id here)\n", fw_use); + return ERR_PTR(-EINVAL); + } + + scp = scp_get(plat_dev); + if (!scp) { + dev_err(&plat_dev->dev, "could not get vdec scp handle"); + return ERR_PTR(-EPROBE_DEFER); + } + + fw = devm_kzalloc(&plat_dev->dev, sizeof(*fw), GFP_KERNEL); + fw->type = SCP; + fw->ops = &mtk_vcodec_rproc_msg; + fw->scp = scp; + + return fw; +} diff --git a/drivers/media/platform/mediatek/vcodec/common/mtk_vcodec_fw_vpu.c b/drivers/media/platform/mediatek/vcodec/common/mtk_vcodec_fw_vpu.c new file mode 100644 index 0000000000..5e03b08865 --- /dev/null +++ b/drivers/media/platform/mediatek/vcodec/common/mtk_vcodec_fw_vpu.c @@ -0,0 +1,136 @@ +// SPDX-License-Identifier: GPL-2.0 + +#include "../decoder/mtk_vcodec_dec_drv.h" +#include "../encoder/mtk_vcodec_enc_drv.h" +#include "mtk_vcodec_fw_priv.h" + +static int mtk_vcodec_vpu_load_firmware(struct mtk_vcodec_fw *fw) +{ + return vpu_load_firmware(fw->pdev); +} + +static unsigned int mtk_vcodec_vpu_get_vdec_capa(struct mtk_vcodec_fw *fw) +{ + return vpu_get_vdec_hw_capa(fw->pdev); +} + +static unsigned int mtk_vcodec_vpu_get_venc_capa(struct mtk_vcodec_fw *fw) +{ + return vpu_get_venc_hw_capa(fw->pdev); +} + +static void *mtk_vcodec_vpu_map_dm_addr(struct mtk_vcodec_fw *fw, + u32 dtcm_dmem_addr) +{ + return vpu_mapping_dm_addr(fw->pdev, dtcm_dmem_addr); +} + +static int mtk_vcodec_vpu_set_ipi_register(struct mtk_vcodec_fw *fw, int id, + mtk_vcodec_ipi_handler handler, + const char *name, void *priv) +{ + /* + * The handler we receive takes a void * as its first argument. We + * cannot change this because it needs to be passed down to the rproc + * subsystem when SCP is used. VPU takes a const argument, which is + * more constrained, so the conversion below is safe. + */ + ipi_handler_t handler_const = (ipi_handler_t)handler; + + return vpu_ipi_register(fw->pdev, id, handler_const, name, priv); +} + +static int mtk_vcodec_vpu_ipi_send(struct mtk_vcodec_fw *fw, int id, void *buf, + unsigned int len, unsigned int wait) +{ + return vpu_ipi_send(fw->pdev, id, buf, len); +} + +static void mtk_vcodec_vpu_release(struct mtk_vcodec_fw *fw) +{ + put_device(&fw->pdev->dev); +} + +static void mtk_vcodec_vpu_reset_dec_handler(void *priv) +{ + struct mtk_vcodec_dec_dev *dev = priv; + struct mtk_vcodec_dec_ctx *ctx; + + dev_err(&dev->plat_dev->dev, "Watchdog timeout!!"); + + mutex_lock(&dev->dev_mutex); + list_for_each_entry(ctx, &dev->ctx_list, list) { + ctx->state = MTK_STATE_ABORT; + mtk_v4l2_vdec_dbg(0, ctx, "[%d] Change to state MTK_STATE_ABORT", ctx->id); + } + mutex_unlock(&dev->dev_mutex); +} + +static void mtk_vcodec_vpu_reset_enc_handler(void *priv) +{ + struct mtk_vcodec_enc_dev *dev = priv; + struct mtk_vcodec_enc_ctx *ctx; + + dev_err(&dev->plat_dev->dev, "Watchdog timeout!!"); + + mutex_lock(&dev->dev_mutex); + list_for_each_entry(ctx, &dev->ctx_list, list) { + ctx->state = MTK_STATE_ABORT; + mtk_v4l2_vdec_dbg(0, ctx, "[%d] Change to state MTK_STATE_ABORT", ctx->id); + } + mutex_unlock(&dev->dev_mutex); +} + +static const struct mtk_vcodec_fw_ops mtk_vcodec_vpu_msg = { + .load_firmware = mtk_vcodec_vpu_load_firmware, + .get_vdec_capa = mtk_vcodec_vpu_get_vdec_capa, + .get_venc_capa = mtk_vcodec_vpu_get_venc_capa, + .map_dm_addr = mtk_vcodec_vpu_map_dm_addr, + .ipi_register = mtk_vcodec_vpu_set_ipi_register, + .ipi_send = mtk_vcodec_vpu_ipi_send, + .release = mtk_vcodec_vpu_release, +}; + +struct mtk_vcodec_fw *mtk_vcodec_fw_vpu_init(void *priv, enum mtk_vcodec_fw_use fw_use) +{ + struct platform_device *fw_pdev; + struct platform_device *plat_dev; + struct mtk_vcodec_fw *fw; + enum rst_id rst_id; + + if (fw_use == ENCODER) { + struct mtk_vcodec_enc_dev *enc_dev = priv; + + plat_dev = enc_dev->plat_dev; + rst_id = VPU_RST_ENC; + } else if (fw_use == DECODER) { + struct mtk_vcodec_dec_dev *dec_dev = priv; + + plat_dev = dec_dev->plat_dev; + rst_id = VPU_RST_DEC; + } else { + pr_err("Invalid fw_use %d (use a resonable fw id here)\n", fw_use); + return ERR_PTR(-EINVAL); + } + + fw_pdev = vpu_get_plat_device(plat_dev); + if (!fw_pdev) { + dev_err(&plat_dev->dev, "firmware device is not ready"); + return ERR_PTR(-EINVAL); + } + + if (fw_use == DECODER) + vpu_wdt_reg_handler(fw_pdev, mtk_vcodec_vpu_reset_dec_handler, priv, rst_id); + else + vpu_wdt_reg_handler(fw_pdev, mtk_vcodec_vpu_reset_enc_handler, priv, rst_id); + + fw = devm_kzalloc(&plat_dev->dev, sizeof(*fw), GFP_KERNEL); + if (!fw) + return ERR_PTR(-ENOMEM); + fw->type = VPU; + fw->ops = &mtk_vcodec_vpu_msg; + fw->pdev = fw_pdev; + fw->fw_use = fw_use; + + return fw; +} diff --git a/drivers/media/platform/mediatek/vcodec/common/mtk_vcodec_intr.c b/drivers/media/platform/mediatek/vcodec/common/mtk_vcodec_intr.c new file mode 100644 index 0000000000..f203fc2563 --- /dev/null +++ b/drivers/media/platform/mediatek/vcodec/common/mtk_vcodec_intr.c @@ -0,0 +1,68 @@ +// SPDX-License-Identifier: GPL-2.0 +/* +* Copyright (c) 2016 MediaTek Inc. +* Author: Tiffany Lin <tiffany.lin@mediatek.com> +*/ + +#include <linux/errno.h> +#include <linux/wait.h> + +#include "../decoder/mtk_vcodec_dec_drv.h" +#include "../encoder/mtk_vcodec_enc_drv.h" +#include "mtk_vcodec_intr.h" + +int mtk_vcodec_wait_for_done_ctx(void *priv, int command, unsigned int timeout_ms, + unsigned int hw_id) +{ + int instance_type = *((int *)priv); + long timeout_jiff, ret; + int ctx_id, ctx_type, status = 0; + int *ctx_int_cond, *ctx_int_type; + wait_queue_head_t *ctx_queue; + struct platform_device *pdev; + + if (instance_type == DECODER) { + struct mtk_vcodec_dec_ctx *ctx; + + ctx = priv; + ctx_id = ctx->id; + ctx_type = ctx->type; + ctx_int_cond = ctx->int_cond; + ctx_int_type = ctx->int_type; + ctx_queue = ctx->queue; + pdev = ctx->dev->plat_dev; + } else { + struct mtk_vcodec_enc_ctx *ctx; + + ctx = priv; + ctx_id = ctx->id; + ctx_type = ctx->type; + ctx_int_cond = ctx->int_cond; + ctx_int_type = ctx->int_type; + ctx_queue = ctx->queue; + pdev = ctx->dev->plat_dev; + } + + timeout_jiff = msecs_to_jiffies(timeout_ms); + ret = wait_event_interruptible_timeout(ctx_queue[hw_id], + ctx_int_cond[hw_id], + timeout_jiff); + + if (!ret) { + status = -1; /* timeout */ + dev_err(&pdev->dev, "[%d] cmd=%d, type=%d, dec timeout=%ums (%d %d)", + ctx_id, command, ctx_type, timeout_ms, + ctx_int_cond[hw_id], ctx_int_type[hw_id]); + } else if (-ERESTARTSYS == ret) { + status = -1; + dev_err(&pdev->dev, "[%d] cmd=%d, type=%d, dec inter fail (%d %d)", + ctx_id, command, ctx_type, + ctx_int_cond[hw_id], ctx_int_type[hw_id]); + } + + ctx_int_cond[hw_id] = 0; + ctx_int_type[hw_id] = 0; + + return status; +} +EXPORT_SYMBOL(mtk_vcodec_wait_for_done_ctx); diff --git a/drivers/media/platform/mediatek/vcodec/common/mtk_vcodec_intr.h b/drivers/media/platform/mediatek/vcodec/common/mtk_vcodec_intr.h new file mode 100644 index 0000000000..3e3cc71ee5 --- /dev/null +++ b/drivers/media/platform/mediatek/vcodec/common/mtk_vcodec_intr.h @@ -0,0 +1,19 @@ +/* SPDX-License-Identifier: GPL-2.0 */ +/* +* Copyright (c) 2016 MediaTek Inc. +* Author: Tiffany Lin <tiffany.lin@mediatek.com> +*/ + +#ifndef _MTK_VCODEC_INTR_H_ +#define _MTK_VCODEC_INTR_H_ + +#define MTK_INST_IRQ_RECEIVED 0x1 + +struct mtk_vcodec_dec_ctx; +struct mtk_vcodec_enc_ctx; + +/* timeout is ms */ +int mtk_vcodec_wait_for_done_ctx(void *priv, int command, unsigned int timeout_ms, + unsigned int hw_id); + +#endif /* _MTK_VCODEC_INTR_H_ */ diff --git a/drivers/media/platform/mediatek/vcodec/common/mtk_vcodec_util.c b/drivers/media/platform/mediatek/vcodec/common/mtk_vcodec_util.c new file mode 100644 index 0000000000..9ce34a3b5e --- /dev/null +++ b/drivers/media/platform/mediatek/vcodec/common/mtk_vcodec_util.c @@ -0,0 +1,174 @@ +// SPDX-License-Identifier: GPL-2.0 +/* +* Copyright (c) 2016 MediaTek Inc. +* Author: PC Chen <pc.chen@mediatek.com> +* Tiffany Lin <tiffany.lin@mediatek.com> +*/ + +#include <linux/module.h> +#include <linux/of.h> +#include <linux/regmap.h> + +#include "../decoder/mtk_vcodec_dec_drv.h" +#include "../encoder/mtk_vcodec_enc_drv.h" +#include "../decoder/mtk_vcodec_dec_hw.h" + +#if defined(CONFIG_DEBUG_FS) +int mtk_vcodec_dbg; +EXPORT_SYMBOL(mtk_vcodec_dbg); + +int mtk_v4l2_dbg_level; +EXPORT_SYMBOL(mtk_v4l2_dbg_level); +#endif + +void __iomem *mtk_vcodec_get_reg_addr(void __iomem **reg_base, unsigned int reg_idx) +{ + if (reg_idx >= NUM_MAX_VCODEC_REG_BASE) { + pr_err(MTK_DBG_V4L2_STR "Invalid arguments, reg_idx=%d", reg_idx); + return NULL; + } + return reg_base[reg_idx]; +} +EXPORT_SYMBOL(mtk_vcodec_get_reg_addr); + +int mtk_vcodec_write_vdecsys(struct mtk_vcodec_dec_ctx *ctx, unsigned int reg, + unsigned int val) +{ + struct mtk_vcodec_dec_dev *dev = ctx->dev; + + if (dev->vdecsys_regmap) + return regmap_write(dev->vdecsys_regmap, reg, val); + + writel(val, dev->reg_base[VDEC_SYS] + reg); + + return 0; +} +EXPORT_SYMBOL(mtk_vcodec_write_vdecsys); + +int mtk_vcodec_mem_alloc(void *priv, struct mtk_vcodec_mem *mem) +{ + enum mtk_instance_type inst_type = *((unsigned int *)priv); + struct platform_device *plat_dev; + unsigned long size = mem->size; + int id; + + if (inst_type == MTK_INST_ENCODER) { + struct mtk_vcodec_enc_ctx *enc_ctx = priv; + + plat_dev = enc_ctx->dev->plat_dev; + id = enc_ctx->id; + } else { + struct mtk_vcodec_dec_ctx *dec_ctx = priv; + + plat_dev = dec_ctx->dev->plat_dev; + id = dec_ctx->id; + } + + mem->va = dma_alloc_coherent(&plat_dev->dev, size, &mem->dma_addr, GFP_KERNEL); + if (!mem->va) { + mtk_v4l2_err(plat_dev, "%s dma_alloc size=%ld failed!", + dev_name(&plat_dev->dev), size); + return -ENOMEM; + } + + mtk_v4l2_debug(plat_dev, 3, "[%d] - va = %p dma = 0x%lx size = 0x%lx", id, mem->va, + (unsigned long)mem->dma_addr, size); + + return 0; +} +EXPORT_SYMBOL(mtk_vcodec_mem_alloc); + +void mtk_vcodec_mem_free(void *priv, struct mtk_vcodec_mem *mem) +{ + enum mtk_instance_type inst_type = *((unsigned int *)priv); + struct platform_device *plat_dev; + unsigned long size = mem->size; + int id; + + if (inst_type == MTK_INST_ENCODER) { + struct mtk_vcodec_enc_ctx *enc_ctx = priv; + + plat_dev = enc_ctx->dev->plat_dev; + id = enc_ctx->id; + } else { + struct mtk_vcodec_dec_ctx *dec_ctx = priv; + + plat_dev = dec_ctx->dev->plat_dev; + id = dec_ctx->id; + } + + if (!mem->va) { + mtk_v4l2_err(plat_dev, "%s dma_free size=%ld failed!", + dev_name(&plat_dev->dev), size); + return; + } + + mtk_v4l2_debug(plat_dev, 3, "[%d] - va = %p dma = 0x%lx size = 0x%lx", id, mem->va, + (unsigned long)mem->dma_addr, size); + + dma_free_coherent(&plat_dev->dev, size, mem->va, mem->dma_addr); + mem->va = NULL; + mem->dma_addr = 0; + mem->size = 0; +} +EXPORT_SYMBOL(mtk_vcodec_mem_free); + +void *mtk_vcodec_get_hw_dev(struct mtk_vcodec_dec_dev *dev, int hw_idx) +{ + if (hw_idx >= MTK_VDEC_HW_MAX || hw_idx < 0 || !dev->subdev_dev[hw_idx]) { + dev_err(&dev->plat_dev->dev, "hw idx is out of range:%d", hw_idx); + return NULL; + } + + return dev->subdev_dev[hw_idx]; +} +EXPORT_SYMBOL(mtk_vcodec_get_hw_dev); + +void mtk_vcodec_set_curr_ctx(struct mtk_vcodec_dec_dev *vdec_dev, + struct mtk_vcodec_dec_ctx *ctx, int hw_idx) +{ + unsigned long flags; + struct mtk_vdec_hw_dev *subdev_dev; + + spin_lock_irqsave(&vdec_dev->irqlock, flags); + if (vdec_dev->vdec_pdata->is_subdev_supported) { + subdev_dev = mtk_vcodec_get_hw_dev(vdec_dev, hw_idx); + if (!subdev_dev) { + dev_err(&vdec_dev->plat_dev->dev, "Failed to get hw dev"); + spin_unlock_irqrestore(&vdec_dev->irqlock, flags); + return; + } + subdev_dev->curr_ctx = ctx; + } else { + vdec_dev->curr_ctx = ctx; + } + spin_unlock_irqrestore(&vdec_dev->irqlock, flags); +} +EXPORT_SYMBOL(mtk_vcodec_set_curr_ctx); + +struct mtk_vcodec_dec_ctx *mtk_vcodec_get_curr_ctx(struct mtk_vcodec_dec_dev *vdec_dev, + unsigned int hw_idx) +{ + unsigned long flags; + struct mtk_vcodec_dec_ctx *ctx; + struct mtk_vdec_hw_dev *subdev_dev; + + spin_lock_irqsave(&vdec_dev->irqlock, flags); + if (vdec_dev->vdec_pdata->is_subdev_supported) { + subdev_dev = mtk_vcodec_get_hw_dev(vdec_dev, hw_idx); + if (!subdev_dev) { + dev_err(&vdec_dev->plat_dev->dev, "Failed to get hw dev"); + spin_unlock_irqrestore(&vdec_dev->irqlock, flags); + return NULL; + } + ctx = subdev_dev->curr_ctx; + } else { + ctx = vdec_dev->curr_ctx; + } + spin_unlock_irqrestore(&vdec_dev->irqlock, flags); + return ctx; +} +EXPORT_SYMBOL(mtk_vcodec_get_curr_ctx); + +MODULE_LICENSE("GPL v2"); +MODULE_DESCRIPTION("Mediatek video codec driver"); diff --git a/drivers/media/platform/mediatek/vcodec/common/mtk_vcodec_util.h b/drivers/media/platform/mediatek/vcodec/common/mtk_vcodec_util.h new file mode 100644 index 0000000000..85f615cdd4 --- /dev/null +++ b/drivers/media/platform/mediatek/vcodec/common/mtk_vcodec_util.h @@ -0,0 +1,75 @@ +/* SPDX-License-Identifier: GPL-2.0 */ +/* +* Copyright (c) 2016 MediaTek Inc. +* Author: PC Chen <pc.chen@mediatek.com> +* Tiffany Lin <tiffany.lin@mediatek.com> +*/ + +#ifndef _MTK_VCODEC_UTIL_H_ +#define _MTK_VCODEC_UTIL_H_ + +#include <linux/types.h> +#include <linux/dma-direction.h> + +#define MTK_DBG_VCODEC_STR "[MTK_VCODEC]" +#define MTK_DBG_V4L2_STR "[MTK_V4L2]" + +struct mtk_vcodec_mem { + size_t size; + void *va; + dma_addr_t dma_addr; +}; + +struct mtk_vcodec_fb { + size_t size; + dma_addr_t dma_addr; +}; + +struct mtk_vcodec_dec_ctx; +struct mtk_vcodec_dec_dev; + +#undef pr_fmt +#define pr_fmt(fmt) "%s(),%d: " fmt, __func__, __LINE__ + +#define mtk_v4l2_err(plat_dev, fmt, args...) \ + dev_err(&(plat_dev)->dev, "[MTK_V4L2][ERROR] " fmt "\n", ##args) + +#define mtk_vcodec_err(inst_id, plat_dev, fmt, args...) \ + dev_err(&(plat_dev)->dev, "[MTK_VCODEC][ERROR][%d]: " fmt "\n", inst_id, ##args) + +#if defined(CONFIG_DEBUG_FS) +extern int mtk_v4l2_dbg_level; +extern int mtk_vcodec_dbg; + +#define mtk_v4l2_debug(plat_dev, level, fmt, args...) \ + do { \ + if (mtk_v4l2_dbg_level >= (level)) \ + dev_dbg(&(plat_dev)->dev, "[MTK_V4L2] %s, %d: " fmt "\n", \ + __func__, __LINE__, ##args); \ + } while (0) + +#define mtk_vcodec_debug(inst_id, plat_dev, fmt, args...) \ + do { \ + if (mtk_vcodec_dbg) \ + dev_dbg(&(plat_dev)->dev, "[MTK_VCODEC][%d]: %s, %d " fmt "\n", \ + inst_id, __func__, __LINE__, ##args); \ + } while (0) +#else +#define mtk_v4l2_debug(plat_dev, level, fmt, args...) \ + dev_dbg(&(plat_dev)->dev, "[MTK_V4L2]: " fmt "\n", ##args) + +#define mtk_vcodec_debug(inst_id, plat_dev, fmt, args...) \ + dev_dbg(&(plat_dev)->dev, "[MTK_VCODEC][%d]: " fmt "\n", inst_id, ##args) +#endif + +void __iomem *mtk_vcodec_get_reg_addr(void __iomem **reg_base, unsigned int reg_idx); +int mtk_vcodec_write_vdecsys(struct mtk_vcodec_dec_ctx *ctx, unsigned int reg, unsigned int val); +int mtk_vcodec_mem_alloc(void *priv, struct mtk_vcodec_mem *mem); +void mtk_vcodec_mem_free(void *priv, struct mtk_vcodec_mem *mem); +void mtk_vcodec_set_curr_ctx(struct mtk_vcodec_dec_dev *vdec_dev, + struct mtk_vcodec_dec_ctx *ctx, int hw_idx); +struct mtk_vcodec_dec_ctx *mtk_vcodec_get_curr_ctx(struct mtk_vcodec_dec_dev *vdec_dev, + unsigned int hw_idx); +void *mtk_vcodec_get_hw_dev(struct mtk_vcodec_dec_dev *dev, int hw_idx); + +#endif /* _MTK_VCODEC_UTIL_H_ */ diff --git a/drivers/media/platform/mediatek/vcodec/decoder/Makefile b/drivers/media/platform/mediatek/vcodec/decoder/Makefile new file mode 100644 index 0000000000..904cd22def --- /dev/null +++ b/drivers/media/platform/mediatek/vcodec/decoder/Makefile @@ -0,0 +1,25 @@ +# SPDX-License-Identifier: GPL-2.0 + +obj-$(CONFIG_VIDEO_MEDIATEK_VCODEC) += mtk-vcodec-dec.o \ + mtk-vcodec-dec-hw.o + +mtk-vcodec-dec-y := vdec/vdec_h264_if.o \ + vdec/vdec_vp8_if.o \ + vdec/vdec_vp8_req_if.o \ + vdec/vdec_vp9_if.o \ + vdec/vdec_vp9_req_lat_if.o \ + vdec/vdec_av1_req_lat_if.o \ + vdec/vdec_h264_req_if.o \ + vdec/vdec_h264_req_common.o \ + vdec/vdec_h264_req_multi_if.o \ + vdec/vdec_hevc_req_multi_if.o \ + mtk_vcodec_dec_drv.o \ + vdec_drv_if.o \ + vdec_vpu_if.o \ + vdec_msg_queue.o \ + mtk_vcodec_dec.o \ + mtk_vcodec_dec_stateful.o \ + mtk_vcodec_dec_stateless.o \ + mtk_vcodec_dec_pm.o \ + +mtk-vcodec-dec-hw-y := mtk_vcodec_dec_hw.o diff --git a/drivers/media/platform/mediatek/vcodec/decoder/mtk_vcodec_dec.c b/drivers/media/platform/mediatek/vcodec/decoder/mtk_vcodec_dec.c new file mode 100644 index 0000000000..91ed576d68 --- /dev/null +++ b/drivers/media/platform/mediatek/vcodec/decoder/mtk_vcodec_dec.c @@ -0,0 +1,1020 @@ +// SPDX-License-Identifier: GPL-2.0 +/* + * Copyright (c) 2016 MediaTek Inc. + * Author: PC Chen <pc.chen@mediatek.com> + * Tiffany Lin <tiffany.lin@mediatek.com> + */ + +#include <media/v4l2-event.h> +#include <media/v4l2-mem2mem.h> +#include <media/videobuf2-dma-contig.h> + +#include "mtk_vcodec_dec_drv.h" +#include "mtk_vcodec_dec.h" +#include "vdec_drv_if.h" +#include "mtk_vcodec_dec_pm.h" + +#define DFT_CFG_WIDTH MTK_VDEC_MIN_W +#define DFT_CFG_HEIGHT MTK_VDEC_MIN_H + +static const struct mtk_video_fmt * +mtk_vdec_find_format(struct v4l2_format *f, + const struct mtk_vcodec_dec_pdata *dec_pdata) +{ + const struct mtk_video_fmt *fmt; + unsigned int k; + + for (k = 0; k < *dec_pdata->num_formats; k++) { + fmt = &dec_pdata->vdec_formats[k]; + if (fmt->fourcc == f->fmt.pix_mp.pixelformat) + return fmt; + } + + return NULL; +} + +static bool mtk_vdec_get_cap_fmt(struct mtk_vcodec_dec_ctx *ctx, int format_index) +{ + const struct mtk_vcodec_dec_pdata *dec_pdata = ctx->dev->vdec_pdata; + const struct mtk_video_fmt *fmt; + struct mtk_q_data *q_data; + int num_frame_count = 0, i; + bool ret = false; + + fmt = &dec_pdata->vdec_formats[format_index]; + for (i = 0; i < *dec_pdata->num_formats; i++) { + if (dec_pdata->vdec_formats[i].type != MTK_FMT_FRAME) + continue; + + num_frame_count++; + } + + if (num_frame_count == 1 || (!ctx->is_10bit_bitstream && fmt->fourcc == V4L2_PIX_FMT_MM21)) + return true; + + q_data = &ctx->q_data[MTK_Q_DATA_SRC]; + switch (q_data->fmt->fourcc) { + case V4L2_PIX_FMT_H264_SLICE: + if (ctx->is_10bit_bitstream && fmt->fourcc == V4L2_PIX_FMT_MT2110R) + ret = true; + break; + case V4L2_PIX_FMT_VP9_FRAME: + case V4L2_PIX_FMT_AV1_FRAME: + case V4L2_PIX_FMT_HEVC_SLICE: + if (ctx->is_10bit_bitstream && fmt->fourcc == V4L2_PIX_FMT_MT2110T) + ret = true; + break; + default: + break; + } + + return ret; +} + +static struct mtk_q_data *mtk_vdec_get_q_data(struct mtk_vcodec_dec_ctx *ctx, + enum v4l2_buf_type type) +{ + if (V4L2_TYPE_IS_OUTPUT(type)) + return &ctx->q_data[MTK_Q_DATA_SRC]; + + return &ctx->q_data[MTK_Q_DATA_DST]; +} + +static int vidioc_try_decoder_cmd(struct file *file, void *priv, + struct v4l2_decoder_cmd *cmd) +{ + return v4l2_m2m_ioctl_try_decoder_cmd(file, priv, cmd); +} + + +static int vidioc_decoder_cmd(struct file *file, void *priv, + struct v4l2_decoder_cmd *cmd) +{ + struct mtk_vcodec_dec_ctx *ctx = fh_to_dec_ctx(priv); + struct vb2_queue *src_vq, *dst_vq; + int ret; + + ret = vidioc_try_decoder_cmd(file, priv, cmd); + if (ret) + return ret; + + mtk_v4l2_vdec_dbg(1, ctx, "decoder cmd=%u", cmd->cmd); + dst_vq = v4l2_m2m_get_vq(ctx->m2m_ctx, + V4L2_BUF_TYPE_VIDEO_CAPTURE_MPLANE); + switch (cmd->cmd) { + case V4L2_DEC_CMD_STOP: + src_vq = v4l2_m2m_get_vq(ctx->m2m_ctx, + V4L2_BUF_TYPE_VIDEO_OUTPUT_MPLANE); + if (!vb2_is_streaming(src_vq)) { + mtk_v4l2_vdec_dbg(1, ctx, "Output stream is off. No need to flush."); + return 0; + } + if (!vb2_is_streaming(dst_vq)) { + mtk_v4l2_vdec_dbg(1, ctx, "Capture stream is off. No need to flush."); + return 0; + } + v4l2_m2m_buf_queue(ctx->m2m_ctx, &ctx->empty_flush_buf.vb); + v4l2_m2m_try_schedule(ctx->m2m_ctx); + break; + + case V4L2_DEC_CMD_START: + vb2_clear_last_buffer_dequeued(dst_vq); + break; + + default: + return -EINVAL; + } + + return 0; +} + +void mtk_vdec_unlock(struct mtk_vcodec_dec_ctx *ctx) +{ + mutex_unlock(&ctx->dev->dec_mutex[ctx->hw_id]); +} + +void mtk_vdec_lock(struct mtk_vcodec_dec_ctx *ctx) +{ + mutex_lock(&ctx->dev->dec_mutex[ctx->hw_id]); +} + +void mtk_vcodec_dec_release(struct mtk_vcodec_dec_ctx *ctx) +{ + vdec_if_deinit(ctx); + ctx->state = MTK_STATE_FREE; +} + +void mtk_vcodec_dec_set_default_params(struct mtk_vcodec_dec_ctx *ctx) +{ + struct mtk_q_data *q_data; + + ctx->m2m_ctx->q_lock = &ctx->dev->dev_mutex; + ctx->fh.m2m_ctx = ctx->m2m_ctx; + ctx->fh.ctrl_handler = &ctx->ctrl_hdl; + INIT_WORK(&ctx->decode_work, ctx->dev->vdec_pdata->worker); + ctx->colorspace = V4L2_COLORSPACE_REC709; + ctx->ycbcr_enc = V4L2_YCBCR_ENC_DEFAULT; + ctx->quantization = V4L2_QUANTIZATION_DEFAULT; + ctx->xfer_func = V4L2_XFER_FUNC_DEFAULT; + + q_data = &ctx->q_data[MTK_Q_DATA_SRC]; + memset(q_data, 0, sizeof(struct mtk_q_data)); + q_data->visible_width = DFT_CFG_WIDTH; + q_data->visible_height = DFT_CFG_HEIGHT; + q_data->fmt = ctx->dev->vdec_pdata->default_out_fmt; + q_data->field = V4L2_FIELD_NONE; + + q_data->sizeimage[0] = DFT_CFG_WIDTH * DFT_CFG_HEIGHT; + q_data->bytesperline[0] = 0; + + q_data = &ctx->q_data[MTK_Q_DATA_DST]; + memset(q_data, 0, sizeof(struct mtk_q_data)); + q_data->visible_width = DFT_CFG_WIDTH; + q_data->visible_height = DFT_CFG_HEIGHT; + q_data->coded_width = DFT_CFG_WIDTH; + q_data->coded_height = DFT_CFG_HEIGHT; + q_data->fmt = ctx->dev->vdec_pdata->default_cap_fmt; + q_data->field = V4L2_FIELD_NONE; + + q_data->sizeimage[0] = q_data->coded_width * q_data->coded_height; + q_data->bytesperline[0] = q_data->coded_width; + q_data->sizeimage[1] = q_data->sizeimage[0] / 2; + q_data->bytesperline[1] = q_data->coded_width; +} + +static int vidioc_vdec_qbuf(struct file *file, void *priv, + struct v4l2_buffer *buf) +{ + struct mtk_vcodec_dec_ctx *ctx = fh_to_dec_ctx(priv); + + if (ctx->state == MTK_STATE_ABORT) { + mtk_v4l2_vdec_err(ctx, "[%d] Call on QBUF after unrecoverable error", ctx->id); + return -EIO; + } + + return v4l2_m2m_qbuf(file, ctx->m2m_ctx, buf); +} + +static int vidioc_vdec_dqbuf(struct file *file, void *priv, + struct v4l2_buffer *buf) +{ + struct mtk_vcodec_dec_ctx *ctx = fh_to_dec_ctx(priv); + + if (ctx->state == MTK_STATE_ABORT) { + mtk_v4l2_vdec_err(ctx, "[%d] Call on DQBUF after unrecoverable error", ctx->id); + return -EIO; + } + + return v4l2_m2m_dqbuf(file, ctx->m2m_ctx, buf); +} + +static int mtk_vcodec_dec_get_chip_name(void *priv) +{ + struct mtk_vcodec_dec_ctx *ctx = fh_to_dec_ctx(priv); + struct device *dev = &ctx->dev->plat_dev->dev; + + if (of_device_is_compatible(dev->of_node, "mediatek,mt8173-vcodec-dec")) + return 8173; + else if (of_device_is_compatible(dev->of_node, "mediatek,mt8183-vcodec-dec")) + return 8183; + else if (of_device_is_compatible(dev->of_node, "mediatek,mt8192-vcodec-dec")) + return 8192; + else if (of_device_is_compatible(dev->of_node, "mediatek,mt8195-vcodec-dec")) + return 8195; + else if (of_device_is_compatible(dev->of_node, "mediatek,mt8186-vcodec-dec")) + return 8186; + else if (of_device_is_compatible(dev->of_node, "mediatek,mt8188-vcodec-dec")) + return 8188; + else + return 8173; +} + +static int vidioc_vdec_querycap(struct file *file, void *priv, + struct v4l2_capability *cap) +{ + struct mtk_vcodec_dec_ctx *ctx = fh_to_dec_ctx(priv); + struct device *dev = &ctx->dev->plat_dev->dev; + int platform_name = mtk_vcodec_dec_get_chip_name(priv); + + strscpy(cap->driver, dev->driver->name, sizeof(cap->driver)); + snprintf(cap->card, sizeof(cap->card), "MT%d video decoder", platform_name); + + return 0; +} + +static int vidioc_vdec_subscribe_evt(struct v4l2_fh *fh, + const struct v4l2_event_subscription *sub) +{ + struct mtk_vcodec_dec_ctx *ctx = fh_to_dec_ctx(fh); + + if (ctx->dev->vdec_pdata->uses_stateless_api) + return v4l2_ctrl_subscribe_event(fh, sub); + + switch (sub->type) { + case V4L2_EVENT_EOS: + return v4l2_event_subscribe(fh, sub, 2, NULL); + case V4L2_EVENT_SOURCE_CHANGE: + return v4l2_src_change_event_subscribe(fh, sub); + default: + return v4l2_ctrl_subscribe_event(fh, sub); + } +} + +static int vidioc_try_fmt(struct mtk_vcodec_dec_ctx *ctx, struct v4l2_format *f, + const struct mtk_video_fmt *fmt) +{ + struct v4l2_pix_format_mplane *pix_fmt_mp = &f->fmt.pix_mp; + const struct v4l2_frmsize_stepwise *frmsize; + + pix_fmt_mp->field = V4L2_FIELD_NONE; + + /* Always apply frame size constraints from the coded side */ + if (V4L2_TYPE_IS_OUTPUT(f->type)) + frmsize = &fmt->frmsize; + else + frmsize = &ctx->q_data[MTK_Q_DATA_SRC].fmt->frmsize; + + pix_fmt_mp->width = clamp(pix_fmt_mp->width, MTK_VDEC_MIN_W, frmsize->max_width); + pix_fmt_mp->height = clamp(pix_fmt_mp->height, MTK_VDEC_MIN_H, frmsize->max_height); + + if (f->type == V4L2_BUF_TYPE_VIDEO_OUTPUT_MPLANE) { + pix_fmt_mp->num_planes = 1; + pix_fmt_mp->plane_fmt[0].bytesperline = 0; + } else if (f->type == V4L2_BUF_TYPE_VIDEO_CAPTURE_MPLANE) { + int tmp_w, tmp_h; + + /* + * Find next closer width align 64, heign align 64, size align + * 64 rectangle + * Note: This only get default value, the real HW needed value + * only available when ctx in MTK_STATE_HEADER state + */ + tmp_w = pix_fmt_mp->width; + tmp_h = pix_fmt_mp->height; + v4l_bound_align_image(&pix_fmt_mp->width, MTK_VDEC_MIN_W, frmsize->max_width, 6, + &pix_fmt_mp->height, MTK_VDEC_MIN_H, frmsize->max_height, 6, + 9); + + if (pix_fmt_mp->width < tmp_w && + (pix_fmt_mp->width + 64) <= frmsize->max_width) + pix_fmt_mp->width += 64; + if (pix_fmt_mp->height < tmp_h && + (pix_fmt_mp->height + 64) <= frmsize->max_height) + pix_fmt_mp->height += 64; + + mtk_v4l2_vdec_dbg(0, ctx, + "before resize wxh=%dx%d, after resize wxh=%dx%d, sizeimage=%d", + tmp_w, tmp_h, pix_fmt_mp->width, pix_fmt_mp->height, + pix_fmt_mp->width * pix_fmt_mp->height); + + pix_fmt_mp->num_planes = fmt->num_planes; + pix_fmt_mp->plane_fmt[0].sizeimage = + pix_fmt_mp->width * pix_fmt_mp->height; + pix_fmt_mp->plane_fmt[0].bytesperline = pix_fmt_mp->width; + + if (pix_fmt_mp->num_planes == 2) { + pix_fmt_mp->plane_fmt[1].sizeimage = + (pix_fmt_mp->width * pix_fmt_mp->height) / 2; + pix_fmt_mp->plane_fmt[1].bytesperline = + pix_fmt_mp->width; + } + } + + pix_fmt_mp->flags = 0; + return 0; +} + +static int vidioc_try_fmt_vid_cap_mplane(struct file *file, void *priv, + struct v4l2_format *f) +{ + const struct mtk_video_fmt *fmt; + struct mtk_vcodec_dec_ctx *ctx = fh_to_dec_ctx(priv); + const struct mtk_vcodec_dec_pdata *dec_pdata = ctx->dev->vdec_pdata; + + fmt = mtk_vdec_find_format(f, dec_pdata); + if (!fmt) { + f->fmt.pix.pixelformat = + ctx->q_data[MTK_Q_DATA_DST].fmt->fourcc; + fmt = mtk_vdec_find_format(f, dec_pdata); + } + + return vidioc_try_fmt(ctx, f, fmt); +} + +static int vidioc_try_fmt_vid_out_mplane(struct file *file, void *priv, + struct v4l2_format *f) +{ + struct v4l2_pix_format_mplane *pix_fmt_mp = &f->fmt.pix_mp; + const struct mtk_video_fmt *fmt; + struct mtk_vcodec_dec_ctx *ctx = fh_to_dec_ctx(priv); + const struct mtk_vcodec_dec_pdata *dec_pdata = ctx->dev->vdec_pdata; + + fmt = mtk_vdec_find_format(f, dec_pdata); + if (!fmt) { + f->fmt.pix.pixelformat = + ctx->q_data[MTK_Q_DATA_SRC].fmt->fourcc; + fmt = mtk_vdec_find_format(f, dec_pdata); + } + + if (pix_fmt_mp->plane_fmt[0].sizeimage == 0) { + mtk_v4l2_vdec_err(ctx, "sizeimage of output format must be given"); + return -EINVAL; + } + + return vidioc_try_fmt(ctx, f, fmt); +} + +static int vidioc_vdec_g_selection(struct file *file, void *priv, + struct v4l2_selection *s) +{ + struct mtk_vcodec_dec_ctx *ctx = fh_to_dec_ctx(priv); + struct mtk_q_data *q_data; + + if (s->type != V4L2_BUF_TYPE_VIDEO_CAPTURE) + return -EINVAL; + + q_data = &ctx->q_data[MTK_Q_DATA_DST]; + + switch (s->target) { + case V4L2_SEL_TGT_COMPOSE_DEFAULT: + s->r.left = 0; + s->r.top = 0; + s->r.width = ctx->picinfo.pic_w; + s->r.height = ctx->picinfo.pic_h; + break; + case V4L2_SEL_TGT_COMPOSE_BOUNDS: + s->r.left = 0; + s->r.top = 0; + s->r.width = ctx->picinfo.buf_w; + s->r.height = ctx->picinfo.buf_h; + break; + case V4L2_SEL_TGT_COMPOSE: + if (vdec_if_get_param(ctx, GET_PARAM_CROP_INFO, &(s->r))) { + /* set to default value if header info not ready yet*/ + s->r.left = 0; + s->r.top = 0; + s->r.width = q_data->visible_width; + s->r.height = q_data->visible_height; + } + break; + default: + return -EINVAL; + } + + if (ctx->state < MTK_STATE_HEADER) { + /* set to default value if header info not ready yet*/ + s->r.left = 0; + s->r.top = 0; + s->r.width = q_data->visible_width; + s->r.height = q_data->visible_height; + return 0; + } + + return 0; +} + +static int vidioc_vdec_s_selection(struct file *file, void *priv, + struct v4l2_selection *s) +{ + struct mtk_vcodec_dec_ctx *ctx = fh_to_dec_ctx(priv); + + if (s->type != V4L2_BUF_TYPE_VIDEO_CAPTURE) + return -EINVAL; + + switch (s->target) { + case V4L2_SEL_TGT_COMPOSE: + s->r.left = 0; + s->r.top = 0; + s->r.width = ctx->picinfo.pic_w; + s->r.height = ctx->picinfo.pic_h; + break; + default: + return -EINVAL; + } + + return 0; +} + +static int vidioc_vdec_s_fmt(struct file *file, void *priv, + struct v4l2_format *f) +{ + struct mtk_vcodec_dec_ctx *ctx = fh_to_dec_ctx(priv); + struct v4l2_pix_format_mplane *pix_mp; + struct mtk_q_data *q_data; + int ret = 0; + const struct mtk_video_fmt *fmt; + const struct mtk_vcodec_dec_pdata *dec_pdata = ctx->dev->vdec_pdata; + + mtk_v4l2_vdec_dbg(3, ctx, "[%d]", ctx->id); + + q_data = mtk_vdec_get_q_data(ctx, f->type); + if (!q_data) + return -EINVAL; + + pix_mp = &f->fmt.pix_mp; + /* + * Setting OUTPUT format after OUTPUT buffers are allocated is invalid + * if using the stateful API. + */ + if (!dec_pdata->uses_stateless_api && + f->type == V4L2_BUF_TYPE_VIDEO_OUTPUT_MPLANE && + vb2_is_busy(&ctx->m2m_ctx->out_q_ctx.q)) { + mtk_v4l2_vdec_err(ctx, "out_q_ctx buffers already requested"); + ret = -EBUSY; + } + + /* + * Setting CAPTURE format after CAPTURE buffers are allocated is + * invalid. + */ + if ((f->type == V4L2_BUF_TYPE_VIDEO_CAPTURE_MPLANE) && + vb2_is_busy(&ctx->m2m_ctx->cap_q_ctx.q)) { + mtk_v4l2_vdec_err(ctx, "cap_q_ctx buffers already requested"); + ret = -EBUSY; + } + + fmt = mtk_vdec_find_format(f, dec_pdata); + if (fmt == NULL) { + if (f->type == V4L2_BUF_TYPE_VIDEO_OUTPUT_MPLANE) { + f->fmt.pix.pixelformat = + dec_pdata->default_out_fmt->fourcc; + fmt = mtk_vdec_find_format(f, dec_pdata); + } else if (f->type == V4L2_BUF_TYPE_VIDEO_CAPTURE_MPLANE) { + f->fmt.pix.pixelformat = + dec_pdata->default_cap_fmt->fourcc; + fmt = mtk_vdec_find_format(f, dec_pdata); + } + } + if (fmt == NULL) + return -EINVAL; + + q_data->fmt = fmt; + vidioc_try_fmt(ctx, f, q_data->fmt); + if (f->type == V4L2_BUF_TYPE_VIDEO_OUTPUT_MPLANE) { + q_data->sizeimage[0] = pix_mp->plane_fmt[0].sizeimage; + q_data->coded_width = pix_mp->width; + q_data->coded_height = pix_mp->height; + + ctx->colorspace = pix_mp->colorspace; + ctx->ycbcr_enc = pix_mp->ycbcr_enc; + ctx->quantization = pix_mp->quantization; + ctx->xfer_func = pix_mp->xfer_func; + + ctx->current_codec = fmt->fourcc; + if (ctx->state == MTK_STATE_FREE) { + ret = vdec_if_init(ctx, q_data->fmt->fourcc); + if (ret) { + mtk_v4l2_vdec_err(ctx, "[%d]: vdec_if_init() fail ret=%d", + ctx->id, ret); + return -EINVAL; + } + ctx->state = MTK_STATE_INIT; + } + } else { + ctx->capture_fourcc = fmt->fourcc; + } + + /* + * If using the stateless API, S_FMT should have the effect of setting + * the CAPTURE queue resolution no matter which queue it was called on. + */ + if (dec_pdata->uses_stateless_api) { + ctx->picinfo.pic_w = pix_mp->width; + ctx->picinfo.pic_h = pix_mp->height; + + /* + * If get pic info fail, need to use the default pic info params, or + * v4l2-compliance will fail + */ + ret = vdec_if_get_param(ctx, GET_PARAM_PIC_INFO, &ctx->picinfo); + if (ret) { + mtk_v4l2_vdec_err(ctx, "[%d]Error!! Get GET_PARAM_PICTURE_INFO Fail", + ctx->id); + } + + ctx->last_decoded_picinfo = ctx->picinfo; + + if (ctx->q_data[MTK_Q_DATA_DST].fmt->num_planes == 1) { + ctx->q_data[MTK_Q_DATA_DST].sizeimage[0] = + ctx->picinfo.fb_sz[0] + + ctx->picinfo.fb_sz[1]; + ctx->q_data[MTK_Q_DATA_DST].bytesperline[0] = + ctx->picinfo.buf_w; + } else { + ctx->q_data[MTK_Q_DATA_DST].sizeimage[0] = + ctx->picinfo.fb_sz[0]; + ctx->q_data[MTK_Q_DATA_DST].bytesperline[0] = + ctx->picinfo.buf_w; + ctx->q_data[MTK_Q_DATA_DST].sizeimage[1] = + ctx->picinfo.fb_sz[1]; + ctx->q_data[MTK_Q_DATA_DST].bytesperline[1] = + ctx->picinfo.buf_w; + } + + ctx->q_data[MTK_Q_DATA_DST].coded_width = ctx->picinfo.buf_w; + ctx->q_data[MTK_Q_DATA_DST].coded_height = ctx->picinfo.buf_h; + mtk_v4l2_vdec_dbg(2, ctx, + "[%d] init() plane:%d wxh=%dx%d pic wxh=%dx%d sz=0x%x_0x%x", + ctx->id, pix_mp->num_planes, + ctx->picinfo.buf_w, ctx->picinfo.buf_h, + ctx->picinfo.pic_w, ctx->picinfo.pic_h, + ctx->q_data[MTK_Q_DATA_DST].sizeimage[0], + ctx->q_data[MTK_Q_DATA_DST].sizeimage[1]); + } + return 0; +} + +static int vidioc_enum_framesizes(struct file *file, void *priv, + struct v4l2_frmsizeenum *fsize) +{ + int i = 0; + struct mtk_vcodec_dec_ctx *ctx = fh_to_dec_ctx(priv); + const struct mtk_vcodec_dec_pdata *dec_pdata = ctx->dev->vdec_pdata; + + if (fsize->index != 0) + return -EINVAL; + + for (i = 0; i < *dec_pdata->num_formats; i++) { + if (fsize->pixel_format != dec_pdata->vdec_formats[i].fourcc) + continue; + + /* Only coded formats have frame sizes set */ + if (!dec_pdata->vdec_formats[i].frmsize.max_width) + return -ENOTTY; + + fsize->type = V4L2_FRMSIZE_TYPE_STEPWISE; + fsize->stepwise = dec_pdata->vdec_formats[i].frmsize; + + mtk_v4l2_vdec_dbg(1, ctx, "%x, %d %d %d %d %d %d", + ctx->dev->dec_capability, fsize->stepwise.min_width, + fsize->stepwise.max_width, fsize->stepwise.step_width, + fsize->stepwise.min_height, fsize->stepwise.max_height, + fsize->stepwise.step_height); + + return 0; + } + + return -EINVAL; +} + +static int vidioc_enum_fmt(struct v4l2_fmtdesc *f, void *priv, + bool output_queue) +{ + struct mtk_vcodec_dec_ctx *ctx = fh_to_dec_ctx(priv); + const struct mtk_vcodec_dec_pdata *dec_pdata = ctx->dev->vdec_pdata; + const struct mtk_video_fmt *fmt; + int i, j = 0; + + for (i = 0; i < *dec_pdata->num_formats; i++) { + if (output_queue && + dec_pdata->vdec_formats[i].type != MTK_FMT_DEC) + continue; + if (!output_queue && + dec_pdata->vdec_formats[i].type != MTK_FMT_FRAME) + continue; + + if (!output_queue && !mtk_vdec_get_cap_fmt(ctx, i)) + continue; + + if (j == f->index) + break; + ++j; + } + + if (i == *dec_pdata->num_formats) + return -EINVAL; + + fmt = &dec_pdata->vdec_formats[i]; + f->pixelformat = fmt->fourcc; + f->flags = fmt->flags; + + return 0; +} + +static int vidioc_vdec_enum_fmt_vid_cap(struct file *file, void *priv, + struct v4l2_fmtdesc *f) +{ + return vidioc_enum_fmt(f, priv, false); +} + +static int vidioc_vdec_enum_fmt_vid_out(struct file *file, void *priv, + struct v4l2_fmtdesc *f) +{ + return vidioc_enum_fmt(f, priv, true); +} + +static int vidioc_vdec_g_fmt(struct file *file, void *priv, + struct v4l2_format *f) +{ + struct mtk_vcodec_dec_ctx *ctx = fh_to_dec_ctx(priv); + struct v4l2_pix_format_mplane *pix_mp = &f->fmt.pix_mp; + struct vb2_queue *vq; + struct mtk_q_data *q_data; + + vq = v4l2_m2m_get_vq(ctx->m2m_ctx, f->type); + if (!vq) { + mtk_v4l2_vdec_err(ctx, "no vb2 queue for type=%d", f->type); + return -EINVAL; + } + + q_data = mtk_vdec_get_q_data(ctx, f->type); + + pix_mp->field = V4L2_FIELD_NONE; + pix_mp->colorspace = ctx->colorspace; + pix_mp->ycbcr_enc = ctx->ycbcr_enc; + pix_mp->quantization = ctx->quantization; + pix_mp->xfer_func = ctx->xfer_func; + + if ((f->type == V4L2_BUF_TYPE_VIDEO_CAPTURE_MPLANE) && + (ctx->state >= MTK_STATE_HEADER)) { + /* Until STREAMOFF is called on the CAPTURE queue + * (acknowledging the event), the driver operates as if + * the resolution hasn't changed yet. + * So we just return picinfo yet, and update picinfo in + * stop_streaming hook function + */ + q_data->sizeimage[0] = ctx->picinfo.fb_sz[0]; + q_data->sizeimage[1] = ctx->picinfo.fb_sz[1]; + q_data->bytesperline[0] = ctx->last_decoded_picinfo.buf_w; + q_data->bytesperline[1] = ctx->last_decoded_picinfo.buf_w; + q_data->coded_width = ctx->picinfo.buf_w; + q_data->coded_height = ctx->picinfo.buf_h; + ctx->last_decoded_picinfo.cap_fourcc = q_data->fmt->fourcc; + + /* + * Width and height are set to the dimensions + * of the movie, the buffer is bigger and + * further processing stages should crop to this + * rectangle. + */ + pix_mp->width = q_data->coded_width; + pix_mp->height = q_data->coded_height; + + /* + * Set pixelformat to the format in which mt vcodec + * outputs the decoded frame + */ + pix_mp->num_planes = q_data->fmt->num_planes; + pix_mp->pixelformat = q_data->fmt->fourcc; + pix_mp->plane_fmt[0].bytesperline = q_data->bytesperline[0]; + pix_mp->plane_fmt[0].sizeimage = q_data->sizeimage[0]; + pix_mp->plane_fmt[1].bytesperline = q_data->bytesperline[1]; + pix_mp->plane_fmt[1].sizeimage = q_data->sizeimage[1]; + + } else if (f->type == V4L2_BUF_TYPE_VIDEO_OUTPUT_MPLANE) { + /* + * This is run on OUTPUT + * The buffer contains compressed image + * so width and height have no meaning. + * Assign value here to pass v4l2-compliance test + */ + pix_mp->width = q_data->visible_width; + pix_mp->height = q_data->visible_height; + pix_mp->plane_fmt[0].bytesperline = q_data->bytesperline[0]; + pix_mp->plane_fmt[0].sizeimage = q_data->sizeimage[0]; + pix_mp->pixelformat = q_data->fmt->fourcc; + pix_mp->num_planes = q_data->fmt->num_planes; + } else { + pix_mp->width = q_data->coded_width; + pix_mp->height = q_data->coded_height; + pix_mp->num_planes = q_data->fmt->num_planes; + pix_mp->pixelformat = q_data->fmt->fourcc; + pix_mp->plane_fmt[0].bytesperline = q_data->bytesperline[0]; + pix_mp->plane_fmt[0].sizeimage = q_data->sizeimage[0]; + pix_mp->plane_fmt[1].bytesperline = q_data->bytesperline[1]; + pix_mp->plane_fmt[1].sizeimage = q_data->sizeimage[1]; + + mtk_v4l2_vdec_dbg(1, ctx, "[%d] type=%d state=%d Format information not ready!", + ctx->id, f->type, ctx->state); + } + + return 0; +} + +int vb2ops_vdec_queue_setup(struct vb2_queue *vq, unsigned int *nbuffers, + unsigned int *nplanes, unsigned int sizes[], + struct device *alloc_devs[]) +{ + struct mtk_vcodec_dec_ctx *ctx = vb2_get_drv_priv(vq); + struct mtk_q_data *q_data; + unsigned int i; + + q_data = mtk_vdec_get_q_data(ctx, vq->type); + + if (q_data == NULL) { + mtk_v4l2_vdec_err(ctx, "vq->type=%d err\n", vq->type); + return -EINVAL; + } + + if (*nplanes) { + if (vq->type == V4L2_BUF_TYPE_VIDEO_CAPTURE_MPLANE) { + if (*nplanes != q_data->fmt->num_planes) + return -EINVAL; + } else { + if (*nplanes != 1) + return -EINVAL; + } + for (i = 0; i < *nplanes; i++) { + if (sizes[i] < q_data->sizeimage[i]) + return -EINVAL; + } + } else { + if (vq->type == V4L2_BUF_TYPE_VIDEO_CAPTURE_MPLANE) + *nplanes = q_data->fmt->num_planes; + else + *nplanes = 1; + + for (i = 0; i < *nplanes; i++) + sizes[i] = q_data->sizeimage[i]; + } + + mtk_v4l2_vdec_dbg(1, ctx, + "[%d]\t type = %d, get %d plane(s), %d buffer(s) of size 0x%x 0x%x ", + ctx->id, vq->type, *nplanes, *nbuffers, sizes[0], sizes[1]); + + return 0; +} + +int vb2ops_vdec_buf_prepare(struct vb2_buffer *vb) +{ + struct mtk_vcodec_dec_ctx *ctx = vb2_get_drv_priv(vb->vb2_queue); + struct mtk_q_data *q_data; + int i; + + mtk_v4l2_vdec_dbg(3, ctx, "[%d] (%d) id=%d", + ctx->id, vb->vb2_queue->type, vb->index); + + q_data = mtk_vdec_get_q_data(ctx, vb->vb2_queue->type); + + for (i = 0; i < q_data->fmt->num_planes; i++) { + if (vb2_plane_size(vb, i) < q_data->sizeimage[i]) { + mtk_v4l2_vdec_err(ctx, "data will not fit into plane %d (%lu < %d)", + i, vb2_plane_size(vb, i), q_data->sizeimage[i]); + return -EINVAL; + } + if (!V4L2_TYPE_IS_OUTPUT(vb->type)) + vb2_set_plane_payload(vb, i, q_data->sizeimage[i]); + } + + return 0; +} + +void vb2ops_vdec_buf_finish(struct vb2_buffer *vb) +{ + struct mtk_vcodec_dec_ctx *ctx = vb2_get_drv_priv(vb->vb2_queue); + struct vb2_v4l2_buffer *vb2_v4l2; + struct mtk_video_dec_buf *buf; + bool buf_error; + + vb2_v4l2 = container_of(vb, struct vb2_v4l2_buffer, vb2_buf); + buf = container_of(vb2_v4l2, struct mtk_video_dec_buf, m2m_buf.vb); + mutex_lock(&ctx->lock); + if (vb->vb2_queue->type == V4L2_BUF_TYPE_VIDEO_CAPTURE_MPLANE) { + buf->queued_in_v4l2 = false; + buf->queued_in_vb2 = false; + } + buf_error = buf->error; + mutex_unlock(&ctx->lock); + + if (buf_error) { + mtk_v4l2_vdec_err(ctx, "Unrecoverable error on buffer."); + ctx->state = MTK_STATE_ABORT; + } +} + +int vb2ops_vdec_buf_init(struct vb2_buffer *vb) +{ + struct vb2_v4l2_buffer *vb2_v4l2 = container_of(vb, + struct vb2_v4l2_buffer, vb2_buf); + struct mtk_video_dec_buf *buf = container_of(vb2_v4l2, + struct mtk_video_dec_buf, m2m_buf.vb); + + if (vb->vb2_queue->type == V4L2_BUF_TYPE_VIDEO_CAPTURE_MPLANE) { + buf->used = false; + buf->queued_in_v4l2 = false; + } + + return 0; +} + +int vb2ops_vdec_start_streaming(struct vb2_queue *q, unsigned int count) +{ + struct mtk_vcodec_dec_ctx *ctx = vb2_get_drv_priv(q); + + if (ctx->state == MTK_STATE_FLUSH) + ctx->state = MTK_STATE_HEADER; + + return 0; +} + +void vb2ops_vdec_stop_streaming(struct vb2_queue *q) +{ + struct vb2_v4l2_buffer *src_buf = NULL, *dst_buf = NULL; + struct mtk_vcodec_dec_ctx *ctx = vb2_get_drv_priv(q); + int ret; + + mtk_v4l2_vdec_dbg(3, ctx, "[%d] (%d) state=(%x) ctx->decoded_frame_cnt=%d", + ctx->id, q->type, ctx->state, ctx->decoded_frame_cnt); + + if (q->type == V4L2_BUF_TYPE_VIDEO_OUTPUT_MPLANE) { + while ((src_buf = v4l2_m2m_src_buf_remove(ctx->m2m_ctx))) { + if (src_buf != &ctx->empty_flush_buf.vb) { + struct media_request *req = + src_buf->vb2_buf.req_obj.req; + v4l2_m2m_buf_done(src_buf, + VB2_BUF_STATE_ERROR); + if (req) + v4l2_ctrl_request_complete(req, &ctx->ctrl_hdl); + } + } + return; + } + + if (ctx->state >= MTK_STATE_HEADER) { + + /* Until STREAMOFF is called on the CAPTURE queue + * (acknowledging the event), the driver operates + * as if the resolution hasn't changed yet, i.e. + * VIDIOC_G_FMT< etc. return previous resolution. + * So we update picinfo here + */ + ctx->picinfo = ctx->last_decoded_picinfo; + + mtk_v4l2_vdec_dbg(2, ctx, + "[%d]-> new(%d,%d), old(%d,%d), real(%d,%d)", + ctx->id, ctx->last_decoded_picinfo.pic_w, + ctx->last_decoded_picinfo.pic_h, + ctx->picinfo.pic_w, ctx->picinfo.pic_h, + ctx->last_decoded_picinfo.buf_w, + ctx->last_decoded_picinfo.buf_h); + + ret = ctx->dev->vdec_pdata->flush_decoder(ctx); + if (ret) + mtk_v4l2_vdec_err(ctx, "DecodeFinal failed, ret=%d", ret); + } + ctx->state = MTK_STATE_FLUSH; + + while ((dst_buf = v4l2_m2m_dst_buf_remove(ctx->m2m_ctx))) { + vb2_set_plane_payload(&dst_buf->vb2_buf, 0, 0); + if (ctx->q_data[MTK_Q_DATA_DST].fmt->num_planes == 2) + vb2_set_plane_payload(&dst_buf->vb2_buf, 1, 0); + v4l2_m2m_buf_done(dst_buf, VB2_BUF_STATE_ERROR); + } + +} + +static void m2mops_vdec_device_run(void *priv) +{ + struct mtk_vcodec_dec_ctx *ctx = priv; + struct mtk_vcodec_dec_dev *dev = ctx->dev; + + queue_work(dev->decode_workqueue, &ctx->decode_work); +} + +static int m2mops_vdec_job_ready(void *m2m_priv) +{ + struct mtk_vcodec_dec_ctx *ctx = m2m_priv; + + mtk_v4l2_vdec_dbg(3, ctx, "[%d]", ctx->id); + + if (ctx->state == MTK_STATE_ABORT) + return 0; + + if ((ctx->last_decoded_picinfo.pic_w != ctx->picinfo.pic_w) || + (ctx->last_decoded_picinfo.pic_h != ctx->picinfo.pic_h)) + return 0; + + if (ctx->state != MTK_STATE_HEADER) + return 0; + + return 1; +} + +static void m2mops_vdec_job_abort(void *priv) +{ + struct mtk_vcodec_dec_ctx *ctx = priv; + + ctx->state = MTK_STATE_ABORT; +} + +const struct v4l2_m2m_ops mtk_vdec_m2m_ops = { + .device_run = m2mops_vdec_device_run, + .job_ready = m2mops_vdec_job_ready, + .job_abort = m2mops_vdec_job_abort, +}; + +const struct v4l2_ioctl_ops mtk_vdec_ioctl_ops = { + .vidioc_streamon = v4l2_m2m_ioctl_streamon, + .vidioc_streamoff = v4l2_m2m_ioctl_streamoff, + .vidioc_reqbufs = v4l2_m2m_ioctl_reqbufs, + .vidioc_querybuf = v4l2_m2m_ioctl_querybuf, + .vidioc_expbuf = v4l2_m2m_ioctl_expbuf, + + .vidioc_qbuf = vidioc_vdec_qbuf, + .vidioc_dqbuf = vidioc_vdec_dqbuf, + + .vidioc_try_fmt_vid_cap_mplane = vidioc_try_fmt_vid_cap_mplane, + .vidioc_try_fmt_vid_out_mplane = vidioc_try_fmt_vid_out_mplane, + + .vidioc_s_fmt_vid_cap_mplane = vidioc_vdec_s_fmt, + .vidioc_s_fmt_vid_out_mplane = vidioc_vdec_s_fmt, + .vidioc_g_fmt_vid_cap_mplane = vidioc_vdec_g_fmt, + .vidioc_g_fmt_vid_out_mplane = vidioc_vdec_g_fmt, + + .vidioc_create_bufs = v4l2_m2m_ioctl_create_bufs, + + .vidioc_enum_fmt_vid_cap = vidioc_vdec_enum_fmt_vid_cap, + .vidioc_enum_fmt_vid_out = vidioc_vdec_enum_fmt_vid_out, + .vidioc_enum_framesizes = vidioc_enum_framesizes, + + .vidioc_querycap = vidioc_vdec_querycap, + .vidioc_subscribe_event = vidioc_vdec_subscribe_evt, + .vidioc_unsubscribe_event = v4l2_event_unsubscribe, + .vidioc_g_selection = vidioc_vdec_g_selection, + .vidioc_s_selection = vidioc_vdec_s_selection, + + .vidioc_decoder_cmd = vidioc_decoder_cmd, + .vidioc_try_decoder_cmd = vidioc_try_decoder_cmd, +}; + +int mtk_vcodec_dec_queue_init(void *priv, struct vb2_queue *src_vq, + struct vb2_queue *dst_vq) +{ + struct mtk_vcodec_dec_ctx *ctx = priv; + int ret = 0; + + mtk_v4l2_vdec_dbg(3, ctx, "[%d]", ctx->id); + + src_vq->type = V4L2_BUF_TYPE_VIDEO_OUTPUT_MPLANE; + src_vq->io_modes = VB2_DMABUF | VB2_MMAP; + src_vq->drv_priv = ctx; + src_vq->buf_struct_size = sizeof(struct mtk_video_dec_buf); + src_vq->ops = ctx->dev->vdec_pdata->vdec_vb2_ops; + src_vq->mem_ops = &vb2_dma_contig_memops; + src_vq->timestamp_flags = V4L2_BUF_FLAG_TIMESTAMP_COPY; + src_vq->lock = &ctx->dev->dev_mutex; + src_vq->dev = &ctx->dev->plat_dev->dev; + src_vq->allow_cache_hints = 1; + + ret = vb2_queue_init(src_vq); + if (ret) { + mtk_v4l2_vdec_err(ctx, "Failed to initialize videobuf2 queue(output)"); + return ret; + } + dst_vq->type = V4L2_BUF_TYPE_VIDEO_CAPTURE_MPLANE; + dst_vq->io_modes = VB2_DMABUF | VB2_MMAP; + dst_vq->drv_priv = ctx; + dst_vq->buf_struct_size = sizeof(struct mtk_video_dec_buf); + dst_vq->ops = ctx->dev->vdec_pdata->vdec_vb2_ops; + dst_vq->mem_ops = &vb2_dma_contig_memops; + dst_vq->timestamp_flags = V4L2_BUF_FLAG_TIMESTAMP_COPY; + dst_vq->lock = &ctx->dev->dev_mutex; + dst_vq->dev = &ctx->dev->plat_dev->dev; + dst_vq->allow_cache_hints = 1; + + ret = vb2_queue_init(dst_vq); + if (ret) + mtk_v4l2_vdec_err(ctx, "Failed to initialize videobuf2 queue(capture)"); + + return ret; +} diff --git a/drivers/media/platform/mediatek/vcodec/decoder/mtk_vcodec_dec.h b/drivers/media/platform/mediatek/vcodec/decoder/mtk_vcodec_dec.h new file mode 100644 index 0000000000..ece27c880e --- /dev/null +++ b/drivers/media/platform/mediatek/vcodec/decoder/mtk_vcodec_dec.h @@ -0,0 +1,103 @@ +/* SPDX-License-Identifier: GPL-2.0 */ +/* + * Copyright (c) 2016 MediaTek Inc. + * Author: PC Chen <pc.chen@mediatek.com> + * Tiffany Lin <tiffany.lin@mediatek.com> + */ + +#ifndef _MTK_VCODEC_DEC_H_ +#define _MTK_VCODEC_DEC_H_ + +#include <media/videobuf2-core.h> +#include <media/v4l2-mem2mem.h> + +#include "mtk_vcodec_dec_drv.h" + +#define VCODEC_DEC_ALIGNED_64 64 +#define VCODEC_CAPABILITY_4K_DISABLED 0x10 +#define VCODEC_DEC_4K_CODED_WIDTH 4096U +#define VCODEC_DEC_4K_CODED_HEIGHT 2304U +#define MTK_VDEC_MAX_W 2048U +#define MTK_VDEC_MAX_H 1088U +#define MTK_VDEC_MIN_W 64U +#define MTK_VDEC_MIN_H 64U + +#define MTK_VDEC_IRQ_STATUS_DEC_SUCCESS 0x10000 + +/** + * struct vdec_fb - decoder frame buffer + * @base_y : Y plane memory info + * @base_c : C plane memory info + * @status : frame buffer status (vdec_fb_status) + */ +struct vdec_fb { + struct mtk_vcodec_mem base_y; + struct mtk_vcodec_mem base_c; + unsigned int status; +}; + +/** + * struct mtk_video_dec_buf - Private data related to each VB2 buffer. + * @m2m_buf: M2M buffer + * @list: link list + * @used: Capture buffer contain decoded frame data and keep in + * codec data structure + * @queued_in_vb2: Capture buffer is queue in vb2 + * @queued_in_v4l2: Capture buffer is in v4l2 driver, but not in vb2 + * queue yet + * @error: An unrecoverable error occurs on this buffer. + * @frame_buffer: Decode status, and buffer information of Capture buffer + * @bs_buffer: Output buffer info + * + * Note : These status information help us track and debug buffer state + */ +struct mtk_video_dec_buf { + struct v4l2_m2m_buffer m2m_buf; + + bool used; + bool queued_in_vb2; + bool queued_in_v4l2; + bool error; + + union { + struct vdec_fb frame_buffer; + struct mtk_vcodec_mem bs_buffer; + }; +}; + +extern const struct v4l2_ioctl_ops mtk_vdec_ioctl_ops; +extern const struct v4l2_m2m_ops mtk_vdec_m2m_ops; +extern const struct media_device_ops mtk_vcodec_media_ops; +extern const struct mtk_vcodec_dec_pdata mtk_vdec_8173_pdata; +extern const struct mtk_vcodec_dec_pdata mtk_vdec_8183_pdata; +extern const struct mtk_vcodec_dec_pdata mtk_lat_sig_core_pdata; +extern const struct mtk_vcodec_dec_pdata mtk_vdec_single_core_pdata; + + +/* + * mtk_vdec_lock/mtk_vdec_unlock are for ctx instance to + * get/release lock before/after access decoder hw. + * mtk_vdec_lock get decoder hw lock and set curr_ctx + * to ctx instance that get lock + */ +void mtk_vdec_unlock(struct mtk_vcodec_dec_ctx *ctx); +void mtk_vdec_lock(struct mtk_vcodec_dec_ctx *ctx); +int mtk_vcodec_dec_queue_init(void *priv, struct vb2_queue *src_vq, + struct vb2_queue *dst_vq); +void mtk_vcodec_dec_set_default_params(struct mtk_vcodec_dec_ctx *ctx); +void mtk_vcodec_dec_release(struct mtk_vcodec_dec_ctx *ctx); + +/* + * VB2 ops + */ +int vb2ops_vdec_queue_setup(struct vb2_queue *vq, unsigned int *nbuffers, + unsigned int *nplanes, unsigned int sizes[], + struct device *alloc_devs[]); +int vb2ops_vdec_buf_prepare(struct vb2_buffer *vb); +void vb2ops_vdec_buf_finish(struct vb2_buffer *vb); +int vb2ops_vdec_buf_init(struct vb2_buffer *vb); +int vb2ops_vdec_start_streaming(struct vb2_queue *q, unsigned int count); +void vb2ops_vdec_stop_streaming(struct vb2_queue *q); + + +#endif /* _MTK_VCODEC_DEC_H_ */ diff --git a/drivers/media/platform/mediatek/vcodec/decoder/mtk_vcodec_dec_drv.c b/drivers/media/platform/mediatek/vcodec/decoder/mtk_vcodec_dec_drv.c new file mode 100644 index 0000000000..0a89ce452a --- /dev/null +++ b/drivers/media/platform/mediatek/vcodec/decoder/mtk_vcodec_dec_drv.c @@ -0,0 +1,573 @@ +// SPDX-License-Identifier: GPL-2.0 +/* + * Copyright (c) 2016 MediaTek Inc. + * Author: PC Chen <pc.chen@mediatek.com> + * Tiffany Lin <tiffany.lin@mediatek.com> + */ + +#include <linux/bitfield.h> +#include <linux/slab.h> +#include <linux/interrupt.h> +#include <linux/irq.h> +#include <linux/mfd/syscon.h> +#include <linux/module.h> +#include <linux/of.h> +#include <linux/of_platform.h> +#include <linux/platform_device.h> +#include <linux/pm_runtime.h> +#include <linux/regmap.h> +#include <media/v4l2-event.h> +#include <media/v4l2-mem2mem.h> +#include <media/videobuf2-dma-contig.h> +#include <media/v4l2-device.h> + +#include "mtk_vcodec_dec.h" +#include "mtk_vcodec_dec_hw.h" +#include "mtk_vcodec_dec_pm.h" +#include "../common/mtk_vcodec_intr.h" + +static int mtk_vcodec_get_hw_count(struct mtk_vcodec_dec_ctx *ctx, struct mtk_vcodec_dec_dev *dev) +{ + switch (dev->vdec_pdata->hw_arch) { + case MTK_VDEC_PURE_SINGLE_CORE: + return MTK_VDEC_ONE_CORE; + case MTK_VDEC_LAT_SINGLE_CORE: + return MTK_VDEC_ONE_LAT_ONE_CORE; + default: + mtk_v4l2_vdec_err(ctx, "hw arch %d not supported", dev->vdec_pdata->hw_arch); + return MTK_VDEC_NO_HW; + } +} + +static bool mtk_vcodec_is_hw_active(struct mtk_vcodec_dec_dev *dev) +{ + u32 cg_status; + + if (dev->vdecsys_regmap) + return !regmap_test_bits(dev->vdecsys_regmap, VDEC_HW_ACTIVE_ADDR, + VDEC_HW_ACTIVE_MASK); + + cg_status = readl(dev->reg_base[VDEC_SYS] + VDEC_HW_ACTIVE_ADDR); + return !FIELD_GET(VDEC_HW_ACTIVE_MASK, cg_status); +} + +static irqreturn_t mtk_vcodec_dec_irq_handler(int irq, void *priv) +{ + struct mtk_vcodec_dec_dev *dev = priv; + struct mtk_vcodec_dec_ctx *ctx; + unsigned int dec_done_status = 0; + void __iomem *vdec_misc_addr = dev->reg_base[VDEC_MISC] + + VDEC_IRQ_CFG_REG; + + ctx = mtk_vcodec_get_curr_ctx(dev, MTK_VDEC_CORE); + + if (!mtk_vcodec_is_hw_active(dev)) { + mtk_v4l2_vdec_err(ctx, "DEC ISR, VDEC active is not 0x0"); + return IRQ_HANDLED; + } + + dec_done_status = readl(vdec_misc_addr); + ctx->irq_status = dec_done_status; + if ((dec_done_status & MTK_VDEC_IRQ_STATUS_DEC_SUCCESS) != + MTK_VDEC_IRQ_STATUS_DEC_SUCCESS) + return IRQ_HANDLED; + + /* clear interrupt */ + writel((readl(vdec_misc_addr) | VDEC_IRQ_CFG), + dev->reg_base[VDEC_MISC] + VDEC_IRQ_CFG_REG); + writel((readl(vdec_misc_addr) & ~VDEC_IRQ_CLR), + dev->reg_base[VDEC_MISC] + VDEC_IRQ_CFG_REG); + + wake_up_dec_ctx(ctx, MTK_INST_IRQ_RECEIVED, 0); + + mtk_v4l2_vdec_dbg(3, ctx, "wake up ctx %d, dec_done_status=%x", ctx->id, dec_done_status); + + return IRQ_HANDLED; +} + +static int mtk_vcodec_get_reg_bases(struct mtk_vcodec_dec_dev *dev) +{ + struct platform_device *pdev = dev->plat_dev; + int reg_num, i; + struct resource *res; + bool has_vdecsys_reg; + int num_max_vdec_regs; + static const char * const mtk_dec_reg_names[] = { + "misc", + "ld", + "top", + "cm", + "ad", + "av", + "pp", + "hwd", + "hwq", + "hwb", + "hwg" + }; + + /* + * If we have reg-names in devicetree, this means that we're on a new + * register organization, which implies that the VDEC_SYS iospace gets + * R/W through a syscon (regmap). + * Here we try to get the "misc" iostart only to check if we have reg-names + */ + res = platform_get_resource_byname(pdev, IORESOURCE_MEM, "misc"); + if (res) + has_vdecsys_reg = false; + else + has_vdecsys_reg = true; + + num_max_vdec_regs = has_vdecsys_reg ? NUM_MAX_VDEC_REG_BASE : + ARRAY_SIZE(mtk_dec_reg_names); + + /* Sizeof(u32) * 4 bytes for each register base. */ + reg_num = of_property_count_elems_of_size(pdev->dev.of_node, "reg", + sizeof(u32) * 4); + if (reg_num <= 0 || reg_num > num_max_vdec_regs) { + dev_err(&pdev->dev, "Invalid register property size: %d\n", reg_num); + return -EINVAL; + } + + if (has_vdecsys_reg) { + for (i = 0; i < reg_num; i++) { + dev->reg_base[i] = devm_platform_ioremap_resource(pdev, i); + if (IS_ERR(dev->reg_base[i])) + return PTR_ERR(dev->reg_base[i]); + + dev_dbg(&pdev->dev, "reg[%d] base=%p", i, dev->reg_base[i]); + } + } else { + for (i = 0; i < reg_num; i++) { + dev->reg_base[i+1] = devm_platform_ioremap_resource_byname(pdev, mtk_dec_reg_names[i]); + if (IS_ERR(dev->reg_base[i+1])) + return PTR_ERR(dev->reg_base[i+1]); + + dev_dbg(&pdev->dev, "reg[%d] base=%p", i + 1, dev->reg_base[i + 1]); + } + + dev->vdecsys_regmap = syscon_regmap_lookup_by_phandle(pdev->dev.of_node, + "mediatek,vdecsys"); + if (IS_ERR(dev->vdecsys_regmap)) { + dev_err(&pdev->dev, "Missing mediatek,vdecsys property"); + return PTR_ERR(dev->vdecsys_regmap); + } + } + + return 0; +} + +static int mtk_vcodec_init_dec_resources(struct mtk_vcodec_dec_dev *dev) +{ + struct platform_device *pdev = dev->plat_dev; + int ret; + + ret = mtk_vcodec_get_reg_bases(dev); + if (ret) + return ret; + + if (dev->vdec_pdata->is_subdev_supported) + return 0; + + dev->dec_irq = platform_get_irq(pdev, 0); + if (dev->dec_irq < 0) + return dev->dec_irq; + + irq_set_status_flags(dev->dec_irq, IRQ_NOAUTOEN); + ret = devm_request_irq(&pdev->dev, dev->dec_irq, + mtk_vcodec_dec_irq_handler, 0, pdev->name, dev); + if (ret) { + dev_err(&pdev->dev, "failed to install dev->dec_irq %d (%d)", + dev->dec_irq, ret); + return ret; + } + + ret = mtk_vcodec_init_dec_clk(pdev, &dev->pm); + if (ret < 0) { + dev_err(&pdev->dev, "failed to get mt vcodec clock source"); + return ret; + } + + pm_runtime_enable(&pdev->dev); + return 0; +} + +static int fops_vcodec_open(struct file *file) +{ + struct mtk_vcodec_dec_dev *dev = video_drvdata(file); + struct mtk_vcodec_dec_ctx *ctx = NULL; + int ret = 0, i, hw_count; + struct vb2_queue *src_vq; + + ctx = kzalloc(sizeof(*ctx), GFP_KERNEL); + if (!ctx) + return -ENOMEM; + + mutex_lock(&dev->dev_mutex); + ctx->id = dev->id_counter++; + v4l2_fh_init(&ctx->fh, video_devdata(file)); + file->private_data = &ctx->fh; + v4l2_fh_add(&ctx->fh); + INIT_LIST_HEAD(&ctx->list); + ctx->dev = dev; + if (ctx->dev->vdec_pdata->is_subdev_supported) { + hw_count = mtk_vcodec_get_hw_count(ctx, dev); + if (!hw_count || !dev->subdev_prob_done) { + ret = -EINVAL; + goto err_ctrls_setup; + } + + ret = dev->subdev_prob_done(dev); + if (ret) + goto err_ctrls_setup; + + for (i = 0; i < hw_count; i++) + init_waitqueue_head(&ctx->queue[i]); + } else { + init_waitqueue_head(&ctx->queue[0]); + } + mutex_init(&ctx->lock); + + ctx->type = MTK_INST_DECODER; + ret = dev->vdec_pdata->ctrls_setup(ctx); + if (ret) { + mtk_v4l2_vdec_err(ctx, "Failed to setup mt vcodec controls"); + goto err_ctrls_setup; + } + ctx->m2m_ctx = v4l2_m2m_ctx_init(dev->m2m_dev_dec, ctx, + &mtk_vcodec_dec_queue_init); + if (IS_ERR((__force void *)ctx->m2m_ctx)) { + ret = PTR_ERR((__force void *)ctx->m2m_ctx); + mtk_v4l2_vdec_err(ctx, "Failed to v4l2_m2m_ctx_init() (%d)", ret); + goto err_m2m_ctx_init; + } + src_vq = v4l2_m2m_get_vq(ctx->m2m_ctx, + V4L2_BUF_TYPE_VIDEO_OUTPUT_MPLANE); + ctx->empty_flush_buf.vb.vb2_buf.vb2_queue = src_vq; + mtk_vcodec_dec_set_default_params(ctx); + + if (v4l2_fh_is_singular(&ctx->fh)) { + /* + * Does nothing if firmware was already loaded. + */ + ret = mtk_vcodec_fw_load_firmware(dev->fw_handler); + if (ret < 0) { + /* + * Return 0 if downloading firmware successfully, + * otherwise it is failed + */ + mtk_v4l2_vdec_err(ctx, "failed to load firmware!"); + goto err_load_fw; + } + + dev->dec_capability = + mtk_vcodec_fw_get_vdec_capa(dev->fw_handler); + + mtk_v4l2_vdec_dbg(0, ctx, "decoder capability %x", dev->dec_capability); + } + + ctx->dev->vdec_pdata->init_vdec_params(ctx); + + list_add(&ctx->list, &dev->ctx_list); + mtk_vcodec_dbgfs_create(ctx); + + mutex_unlock(&dev->dev_mutex); + mtk_v4l2_vdec_dbg(0, ctx, "%s decoder [%d]", dev_name(&dev->plat_dev->dev), ctx->id); + return ret; + + /* Deinit when failure occurred */ +err_load_fw: + v4l2_m2m_ctx_release(ctx->m2m_ctx); +err_m2m_ctx_init: + v4l2_ctrl_handler_free(&ctx->ctrl_hdl); +err_ctrls_setup: + v4l2_fh_del(&ctx->fh); + v4l2_fh_exit(&ctx->fh); + kfree(ctx); + mutex_unlock(&dev->dev_mutex); + + return ret; +} + +static int fops_vcodec_release(struct file *file) +{ + struct mtk_vcodec_dec_dev *dev = video_drvdata(file); + struct mtk_vcodec_dec_ctx *ctx = fh_to_dec_ctx(file->private_data); + + mtk_v4l2_vdec_dbg(0, ctx, "[%d] decoder", ctx->id); + mutex_lock(&dev->dev_mutex); + + /* + * Call v4l2_m2m_ctx_release before mtk_vcodec_dec_release. First, it + * makes sure the worker thread is not running after vdec_if_deinit. + * Second, the decoder will be flushed and all the buffers will be + * returned in stop_streaming. + */ + v4l2_m2m_ctx_release(ctx->m2m_ctx); + mtk_vcodec_dec_release(ctx); + + v4l2_fh_del(&ctx->fh); + v4l2_fh_exit(&ctx->fh); + v4l2_ctrl_handler_free(&ctx->ctrl_hdl); + + mtk_vcodec_dbgfs_remove(dev, ctx->id); + list_del_init(&ctx->list); + kfree(ctx); + mutex_unlock(&dev->dev_mutex); + return 0; +} + +static const struct v4l2_file_operations mtk_vcodec_fops = { + .owner = THIS_MODULE, + .open = fops_vcodec_open, + .release = fops_vcodec_release, + .poll = v4l2_m2m_fop_poll, + .unlocked_ioctl = video_ioctl2, + .mmap = v4l2_m2m_fop_mmap, +}; + +static int mtk_vcodec_probe(struct platform_device *pdev) +{ + struct mtk_vcodec_dec_dev *dev; + struct video_device *vfd_dec; + phandle rproc_phandle; + enum mtk_vcodec_fw_type fw_type; + int i, ret; + + dev = devm_kzalloc(&pdev->dev, sizeof(*dev), GFP_KERNEL); + if (!dev) + return -ENOMEM; + + INIT_LIST_HEAD(&dev->ctx_list); + dev->plat_dev = pdev; + + dev->vdec_pdata = of_device_get_match_data(&pdev->dev); + if (!of_property_read_u32(pdev->dev.of_node, "mediatek,vpu", + &rproc_phandle)) { + fw_type = VPU; + } else if (!of_property_read_u32(pdev->dev.of_node, "mediatek,scp", + &rproc_phandle)) { + fw_type = SCP; + } else { + dev_dbg(&pdev->dev, "Could not get vdec IPI device"); + return -ENODEV; + } + dma_set_max_seg_size(&pdev->dev, UINT_MAX); + + dev->fw_handler = mtk_vcodec_fw_select(dev, fw_type, DECODER); + if (IS_ERR(dev->fw_handler)) + return PTR_ERR(dev->fw_handler); + + ret = mtk_vcodec_init_dec_resources(dev); + if (ret) { + dev_err(&pdev->dev, "Failed to init dec resources"); + goto err_dec_pm; + } + + if (IS_VDEC_LAT_ARCH(dev->vdec_pdata->hw_arch)) { + dev->core_workqueue = + alloc_ordered_workqueue("core-decoder", + WQ_MEM_RECLAIM | WQ_FREEZABLE); + if (!dev->core_workqueue) { + dev_dbg(&pdev->dev, "Failed to create core workqueue"); + ret = -EINVAL; + goto err_res; + } + } + + for (i = 0; i < MTK_VDEC_HW_MAX; i++) + mutex_init(&dev->dec_mutex[i]); + mutex_init(&dev->dev_mutex); + spin_lock_init(&dev->irqlock); + + snprintf(dev->v4l2_dev.name, sizeof(dev->v4l2_dev.name), "%s", + "[/MTK_V4L2_VDEC]"); + + ret = v4l2_device_register(&pdev->dev, &dev->v4l2_dev); + if (ret) { + dev_err(&pdev->dev, "v4l2_device_register err=%d", ret); + goto err_core_workq; + } + + vfd_dec = video_device_alloc(); + if (!vfd_dec) { + dev_err(&pdev->dev, "Failed to allocate video device"); + ret = -ENOMEM; + goto err_dec_alloc; + } + vfd_dec->fops = &mtk_vcodec_fops; + vfd_dec->ioctl_ops = &mtk_vdec_ioctl_ops; + vfd_dec->release = video_device_release; + vfd_dec->lock = &dev->dev_mutex; + vfd_dec->v4l2_dev = &dev->v4l2_dev; + vfd_dec->vfl_dir = VFL_DIR_M2M; + vfd_dec->device_caps = V4L2_CAP_VIDEO_M2M_MPLANE | + V4L2_CAP_STREAMING; + + snprintf(vfd_dec->name, sizeof(vfd_dec->name), "%s", + MTK_VCODEC_DEC_NAME); + video_set_drvdata(vfd_dec, dev); + dev->vfd_dec = vfd_dec; + platform_set_drvdata(pdev, dev); + + dev->m2m_dev_dec = v4l2_m2m_init(&mtk_vdec_m2m_ops); + if (IS_ERR((__force void *)dev->m2m_dev_dec)) { + dev_err(&pdev->dev, "Failed to init mem2mem dec device"); + ret = PTR_ERR((__force void *)dev->m2m_dev_dec); + goto err_dec_alloc; + } + + dev->decode_workqueue = + alloc_ordered_workqueue(MTK_VCODEC_DEC_NAME, + WQ_MEM_RECLAIM | WQ_FREEZABLE); + if (!dev->decode_workqueue) { + dev_err(&pdev->dev, "Failed to create decode workqueue"); + ret = -EINVAL; + goto err_event_workq; + } + + if (dev->vdec_pdata->is_subdev_supported) { + ret = of_platform_populate(pdev->dev.of_node, NULL, NULL, + &pdev->dev); + if (ret) { + dev_err(&pdev->dev, "Main device of_platform_populate failed."); + goto err_reg_cont; + } + } else { + set_bit(MTK_VDEC_CORE, dev->subdev_bitmap); + } + + atomic_set(&dev->dec_active_cnt, 0); + memset(dev->vdec_racing_info, 0, sizeof(dev->vdec_racing_info)); + mutex_init(&dev->dec_racing_info_mutex); + + ret = video_register_device(vfd_dec, VFL_TYPE_VIDEO, -1); + if (ret) { + dev_err(&pdev->dev, "Failed to register video device"); + goto err_reg_cont; + } + + if (dev->vdec_pdata->uses_stateless_api) { + v4l2_disable_ioctl(vfd_dec, VIDIOC_DECODER_CMD); + v4l2_disable_ioctl(vfd_dec, VIDIOC_TRY_DECODER_CMD); + + dev->mdev_dec.dev = &pdev->dev; + strscpy(dev->mdev_dec.model, MTK_VCODEC_DEC_NAME, + sizeof(dev->mdev_dec.model)); + + media_device_init(&dev->mdev_dec); + dev->mdev_dec.ops = &mtk_vcodec_media_ops; + dev->v4l2_dev.mdev = &dev->mdev_dec; + + ret = v4l2_m2m_register_media_controller(dev->m2m_dev_dec, dev->vfd_dec, + MEDIA_ENT_F_PROC_VIDEO_DECODER); + if (ret) { + dev_err(&pdev->dev, "Failed to register media controller"); + goto err_dec_mem_init; + } + + ret = media_device_register(&dev->mdev_dec); + if (ret) { + dev_err(&pdev->dev, "Failed to register media device"); + goto err_media_reg; + } + + dev_dbg(&pdev->dev, "media registered as /dev/media%d", vfd_dec->minor); + } + + mtk_vcodec_dbgfs_init(dev, false); + dev_dbg(&pdev->dev, "decoder registered as /dev/video%d", vfd_dec->minor); + + return 0; + +err_media_reg: + v4l2_m2m_unregister_media_controller(dev->m2m_dev_dec); +err_dec_mem_init: + video_unregister_device(vfd_dec); +err_reg_cont: + if (dev->vdec_pdata->uses_stateless_api) + media_device_cleanup(&dev->mdev_dec); + destroy_workqueue(dev->decode_workqueue); +err_event_workq: + v4l2_m2m_release(dev->m2m_dev_dec); +err_dec_alloc: + v4l2_device_unregister(&dev->v4l2_dev); +err_core_workq: + if (IS_VDEC_LAT_ARCH(dev->vdec_pdata->hw_arch)) + destroy_workqueue(dev->core_workqueue); +err_res: + if (!dev->vdec_pdata->is_subdev_supported) + pm_runtime_disable(dev->pm.dev); +err_dec_pm: + mtk_vcodec_fw_release(dev->fw_handler); + return ret; +} + +static const struct of_device_id mtk_vcodec_match[] = { + { + .compatible = "mediatek,mt8173-vcodec-dec", + .data = &mtk_vdec_8173_pdata, + }, + { + .compatible = "mediatek,mt8183-vcodec-dec", + .data = &mtk_vdec_8183_pdata, + }, + { + .compatible = "mediatek,mt8192-vcodec-dec", + .data = &mtk_lat_sig_core_pdata, + }, + { + .compatible = "mediatek,mt8186-vcodec-dec", + .data = &mtk_vdec_single_core_pdata, + }, + { + .compatible = "mediatek,mt8195-vcodec-dec", + .data = &mtk_lat_sig_core_pdata, + }, + { + .compatible = "mediatek,mt8188-vcodec-dec", + .data = &mtk_lat_sig_core_pdata, + }, + {}, +}; + +MODULE_DEVICE_TABLE(of, mtk_vcodec_match); + +static void mtk_vcodec_dec_remove(struct platform_device *pdev) +{ + struct mtk_vcodec_dec_dev *dev = platform_get_drvdata(pdev); + + destroy_workqueue(dev->decode_workqueue); + + if (media_devnode_is_registered(dev->mdev_dec.devnode)) { + media_device_unregister(&dev->mdev_dec); + v4l2_m2m_unregister_media_controller(dev->m2m_dev_dec); + media_device_cleanup(&dev->mdev_dec); + } + + if (dev->m2m_dev_dec) + v4l2_m2m_release(dev->m2m_dev_dec); + + if (dev->vfd_dec) + video_unregister_device(dev->vfd_dec); + + mtk_vcodec_dbgfs_deinit(&dev->dbgfs); + v4l2_device_unregister(&dev->v4l2_dev); + if (!dev->vdec_pdata->is_subdev_supported) + pm_runtime_disable(dev->pm.dev); + mtk_vcodec_fw_release(dev->fw_handler); +} + +static struct platform_driver mtk_vcodec_dec_driver = { + .probe = mtk_vcodec_probe, + .remove_new = mtk_vcodec_dec_remove, + .driver = { + .name = MTK_VCODEC_DEC_NAME, + .of_match_table = mtk_vcodec_match, + }, +}; + +module_platform_driver(mtk_vcodec_dec_driver); + +MODULE_LICENSE("GPL v2"); +MODULE_DESCRIPTION("Mediatek video codec V4L2 decoder driver"); diff --git a/drivers/media/platform/mediatek/vcodec/decoder/mtk_vcodec_dec_drv.h b/drivers/media/platform/mediatek/vcodec/decoder/mtk_vcodec_dec_drv.h new file mode 100644 index 0000000000..7e36b2c69b --- /dev/null +++ b/drivers/media/platform/mediatek/vcodec/decoder/mtk_vcodec_dec_drv.h @@ -0,0 +1,324 @@ +/* SPDX-License-Identifier: GPL-2.0 */ +/* + * Copyright (c) 2023 MediaTek Inc. + * Author: Yunfei Dong <yunfei.dong@mediatek.com> + */ + +#ifndef _MTK_VCODEC_DEC_DRV_H_ +#define _MTK_VCODEC_DEC_DRV_H_ + +#include "../common/mtk_vcodec_cmn_drv.h" +#include "../common/mtk_vcodec_dbgfs.h" +#include "../common/mtk_vcodec_fw_priv.h" +#include "../common/mtk_vcodec_util.h" +#include "vdec_msg_queue.h" + +#define MTK_VCODEC_DEC_NAME "mtk-vcodec-dec" + +#define IS_VDEC_LAT_ARCH(hw_arch) ((hw_arch) >= MTK_VDEC_LAT_SINGLE_CORE) +#define IS_VDEC_INNER_RACING(capability) ((capability) & MTK_VCODEC_INNER_RACING) + +/* + * enum mtk_vdec_format_types - Structure used to get supported + * format types according to decoder capability + */ +enum mtk_vdec_format_types { + MTK_VDEC_FORMAT_MM21 = 0x20, + MTK_VDEC_FORMAT_MT21C = 0x40, + MTK_VDEC_FORMAT_H264_SLICE = 0x100, + MTK_VDEC_FORMAT_VP8_FRAME = 0x200, + MTK_VDEC_FORMAT_VP9_FRAME = 0x400, + MTK_VDEC_FORMAT_AV1_FRAME = 0x800, + MTK_VDEC_FORMAT_HEVC_FRAME = 0x1000, + MTK_VCODEC_INNER_RACING = 0x20000, + MTK_VDEC_IS_SUPPORT_10BIT = 0x40000, +}; + +/* + * enum mtk_vdec_hw_count - Supported hardware count + */ +enum mtk_vdec_hw_count { + MTK_VDEC_NO_HW = 0, + MTK_VDEC_ONE_CORE, + MTK_VDEC_ONE_LAT_ONE_CORE, + MTK_VDEC_MAX_HW_COUNT, +}; + +/* + * enum mtk_vdec_hw_arch - Used to separate different hardware architecture + */ +enum mtk_vdec_hw_arch { + MTK_VDEC_PURE_SINGLE_CORE, + MTK_VDEC_LAT_SINGLE_CORE, +}; + +/** + * struct vdec_pic_info - picture size information + * @pic_w: picture width + * @pic_h: picture height + * @buf_w: picture buffer width (64 aligned up from pic_w) + * @buf_h: picture buffer heiht (64 aligned up from pic_h) + * @fb_sz: bitstream size of each plane + * E.g. suppose picture size is 176x144, + * buffer size will be aligned to 176x160. + * @cap_fourcc: fourcc number(may changed when resolution change) + * @reserved: align struct to 64-bit in order to adjust 32-bit and 64-bit os. + */ +struct vdec_pic_info { + unsigned int pic_w; + unsigned int pic_h; + unsigned int buf_w; + unsigned int buf_h; + unsigned int fb_sz[VIDEO_MAX_PLANES]; + unsigned int cap_fourcc; + unsigned int reserved; +}; + +/** + * struct mtk_vcodec_dec_pdata - compatible data for each IC + * @init_vdec_params: init vdec params + * @ctrls_setup: init vcodec dec ctrls + * @worker: worker to start a decode job + * @flush_decoder: function that flushes the decoder + * @get_cap_buffer: get capture buffer from capture queue + * @cap_to_disp: put capture buffer to disp list for lat and core arch + * @vdec_vb2_ops: struct vb2_ops + * + * @vdec_formats: supported video decoder formats + * @num_formats: count of video decoder formats + * @default_out_fmt: default output buffer format + * @default_cap_fmt: default capture buffer format + * + * @hw_arch: hardware arch is used to separate pure_sin_core and lat_sin_core + * + * @is_subdev_supported: whether support parent-node architecture(subdev) + * @uses_stateless_api: whether the decoder uses the stateless API with requests + */ +struct mtk_vcodec_dec_pdata { + void (*init_vdec_params)(struct mtk_vcodec_dec_ctx *ctx); + int (*ctrls_setup)(struct mtk_vcodec_dec_ctx *ctx); + void (*worker)(struct work_struct *work); + int (*flush_decoder)(struct mtk_vcodec_dec_ctx *ctx); + struct vdec_fb *(*get_cap_buffer)(struct mtk_vcodec_dec_ctx *ctx); + void (*cap_to_disp)(struct mtk_vcodec_dec_ctx *ctx, int error, + struct media_request *src_buf_req); + + const struct vb2_ops *vdec_vb2_ops; + + const struct mtk_video_fmt *vdec_formats; + const int *num_formats; + const struct mtk_video_fmt *default_out_fmt; + const struct mtk_video_fmt *default_cap_fmt; + + enum mtk_vdec_hw_arch hw_arch; + + bool is_subdev_supported; + bool uses_stateless_api; +}; + +/** + * struct mtk_vcodec_dec_ctx - Context (instance) private data. + * + * @type: type of decoder instance + * @dev: pointer to the mtk_vcodec_dec_dev of the device + * @list: link to ctx_list of mtk_vcodec_dec_dev + * + * @fh: struct v4l2_fh + * @m2m_ctx: pointer to the v4l2_m2m_ctx of the context + * @q_data: store information of input and output queue of the context + * @id: index of the context that this structure describes + * @state: state of the context + * + * @dec_if: hooked decoder driver interface + * @drv_handle: driver handle for specific decode/encode instance + * + * @picinfo: store picture info after header parsing + * @dpb_size: store dpb count after header parsing + * + * @int_cond: variable used by the waitqueue + * @int_type: type of the last interrupt + * @queue: waitqueue that can be used to wait for this context to finish + * @irq_status: irq status + * + * @ctrl_hdl: handler for v4l2 framework + * @decode_work: worker for the decoding + * @last_decoded_picinfo: pic information get from latest decode + * @empty_flush_buf: a fake size-0 capture buffer that indicates flush. Used + * for stateful decoder. + * @is_flushing: set to true if flushing is in progress. + * + * @current_codec: current set input codec, in V4L2 pixel format + * @capture_fourcc: capture queue type in V4L2 pixel format + * + * @colorspace: enum v4l2_colorspace; supplemental to pixelformat + * @ycbcr_enc: enum v4l2_ycbcr_encoding, Y'CbCr encoding + * @quantization: enum v4l2_quantization, colorspace quantization + * @xfer_func: enum v4l2_xfer_func, colorspace transfer function + * + * @decoded_frame_cnt: number of decoded frames + * @lock: protect variables accessed by V4L2 threads and worker thread such as + * mtk_video_dec_buf. + * @hw_id: hardware index used to identify different hardware. + * + * @msg_queue: msg queue used to store lat buffer information. + * @vpu_inst: vpu instance pointer. + * + * @is_10bit_bitstream: set to true if it's 10bit bitstream + */ +struct mtk_vcodec_dec_ctx { + enum mtk_instance_type type; + struct mtk_vcodec_dec_dev *dev; + struct list_head list; + + struct v4l2_fh fh; + struct v4l2_m2m_ctx *m2m_ctx; + struct mtk_q_data q_data[2]; + int id; + enum mtk_instance_state state; + + const struct vdec_common_if *dec_if; + void *drv_handle; + + struct vdec_pic_info picinfo; + int dpb_size; + + int int_cond[MTK_VDEC_HW_MAX]; + int int_type[MTK_VDEC_HW_MAX]; + wait_queue_head_t queue[MTK_VDEC_HW_MAX]; + unsigned int irq_status; + + struct v4l2_ctrl_handler ctrl_hdl; + struct work_struct decode_work; + struct vdec_pic_info last_decoded_picinfo; + struct v4l2_m2m_buffer empty_flush_buf; + bool is_flushing; + + u32 current_codec; + u32 capture_fourcc; + + enum v4l2_colorspace colorspace; + enum v4l2_ycbcr_encoding ycbcr_enc; + enum v4l2_quantization quantization; + enum v4l2_xfer_func xfer_func; + + int decoded_frame_cnt; + struct mutex lock; + int hw_id; + + struct vdec_msg_queue msg_queue; + void *vpu_inst; + + bool is_10bit_bitstream; +}; + +/** + * struct mtk_vcodec_dec_dev - driver data + * @v4l2_dev: V4L2 device to register video devices for. + * @vfd_dec: Video device for decoder + * @mdev_dec: Media device for decoder + * + * @m2m_dev_dec: m2m device for decoder + * @plat_dev: platform device + * @ctx_list: list of struct mtk_vcodec_ctx + * @curr_ctx: The context that is waiting for codec hardware + * + * @reg_base: Mapped address of MTK Vcodec registers. + * @vdec_pdata: decoder IC-specific data + * @vdecsys_regmap: VDEC_SYS register space passed through syscon + * + * @fw_handler: used to communicate with the firmware. + * @id_counter: used to identify current opened instance + * + * @dec_mutex: decoder hardware lock + * @dev_mutex: video_device lock + * @decode_workqueue: decode work queue + * + * @irqlock: protect data access by irq handler and work thread + * @dec_irq: decoder irq resource + * + * @pm: power management control + * @dec_capability: used to identify decode capability, ex: 4k + * + * @core_workqueue: queue used for core hardware decode + * + * @subdev_dev: subdev hardware device + * @subdev_prob_done: check whether all used hw device is prob done + * @subdev_bitmap: used to record hardware is ready or not + * + * @dec_active_cnt: used to mark whether need to record register value + * @vdec_racing_info: record register value + * @dec_racing_info_mutex: mutex lock used for inner racing mode + * @dbgfs: debug log related information + */ +struct mtk_vcodec_dec_dev { + struct v4l2_device v4l2_dev; + struct video_device *vfd_dec; + struct media_device mdev_dec; + + struct v4l2_m2m_dev *m2m_dev_dec; + struct platform_device *plat_dev; + struct list_head ctx_list; + struct mtk_vcodec_dec_ctx *curr_ctx; + + void __iomem *reg_base[NUM_MAX_VCODEC_REG_BASE]; + const struct mtk_vcodec_dec_pdata *vdec_pdata; + struct regmap *vdecsys_regmap; + + struct mtk_vcodec_fw *fw_handler; + u64 id_counter; + + /* decoder hardware mutex lock */ + struct mutex dec_mutex[MTK_VDEC_HW_MAX]; + struct mutex dev_mutex; + struct workqueue_struct *decode_workqueue; + + spinlock_t irqlock; + int dec_irq; + + struct mtk_vcodec_pm pm; + unsigned int dec_capability; + + struct workqueue_struct *core_workqueue; + + void *subdev_dev[MTK_VDEC_HW_MAX]; + int (*subdev_prob_done)(struct mtk_vcodec_dec_dev *vdec_dev); + DECLARE_BITMAP(subdev_bitmap, MTK_VDEC_HW_MAX); + + atomic_t dec_active_cnt; + u32 vdec_racing_info[132]; + /* Protects access to vdec_racing_info data */ + struct mutex dec_racing_info_mutex; + struct mtk_vcodec_dbgfs dbgfs; +}; + +static inline struct mtk_vcodec_dec_ctx *fh_to_dec_ctx(struct v4l2_fh *fh) +{ + return container_of(fh, struct mtk_vcodec_dec_ctx, fh); +} + +static inline struct mtk_vcodec_dec_ctx *ctrl_to_dec_ctx(struct v4l2_ctrl *ctrl) +{ + return container_of(ctrl->handler, struct mtk_vcodec_dec_ctx, ctrl_hdl); +} + +/* Wake up context wait_queue */ +static inline void +wake_up_dec_ctx(struct mtk_vcodec_dec_ctx *ctx, unsigned int reason, unsigned int hw_id) +{ + ctx->int_cond[hw_id] = 1; + ctx->int_type[hw_id] = reason; + wake_up_interruptible(&ctx->queue[hw_id]); +} + +#define mtk_vdec_err(ctx, fmt, args...) \ + mtk_vcodec_err((ctx)->id, (ctx)->dev->plat_dev, fmt, ##args) + +#define mtk_vdec_debug(ctx, fmt, args...) \ + mtk_vcodec_debug((ctx)->id, (ctx)->dev->plat_dev, fmt, ##args) + +#define mtk_v4l2_vdec_err(ctx, fmt, args...) mtk_v4l2_err((ctx)->dev->plat_dev, fmt, ##args) + +#define mtk_v4l2_vdec_dbg(level, ctx, fmt, args...) \ + mtk_v4l2_debug((ctx)->dev->plat_dev, level, fmt, ##args) + +#endif /* _MTK_VCODEC_DEC_DRV_H_ */ diff --git a/drivers/media/platform/mediatek/vcodec/decoder/mtk_vcodec_dec_hw.c b/drivers/media/platform/mediatek/vcodec/decoder/mtk_vcodec_dec_hw.c new file mode 100644 index 0000000000..881d5de41e --- /dev/null +++ b/drivers/media/platform/mediatek/vcodec/decoder/mtk_vcodec_dec_hw.c @@ -0,0 +1,201 @@ +// SPDX-License-Identifier: GPL-2.0 +/* + * Copyright (c) 2021 MediaTek Inc. + * Author: Yunfei Dong <yunfei.dong@mediatek.com> + */ + +#include <linux/interrupt.h> +#include <linux/irq.h> +#include <linux/module.h> +#include <linux/of.h> +#include <linux/of_device.h> +#include <linux/pm_runtime.h> +#include <linux/slab.h> + +#include "mtk_vcodec_dec.h" +#include "mtk_vcodec_dec_hw.h" +#include "mtk_vcodec_dec_pm.h" +#include "../common/mtk_vcodec_intr.h" + +static const struct of_device_id mtk_vdec_hw_match[] = { + { + .compatible = "mediatek,mtk-vcodec-lat", + .data = (void *)MTK_VDEC_LAT0, + }, + { + .compatible = "mediatek,mtk-vcodec-core", + .data = (void *)MTK_VDEC_CORE, + }, + { + .compatible = "mediatek,mtk-vcodec-lat-soc", + .data = (void *)MTK_VDEC_LAT_SOC, + }, + {}, +}; +MODULE_DEVICE_TABLE(of, mtk_vdec_hw_match); + +static int mtk_vdec_hw_prob_done(struct mtk_vcodec_dec_dev *vdec_dev) +{ + struct platform_device *pdev = vdec_dev->plat_dev; + struct device_node *subdev_node; + enum mtk_vdec_hw_id hw_idx; + const struct of_device_id *of_id; + int i; + + for (i = 0; i < ARRAY_SIZE(mtk_vdec_hw_match); i++) { + of_id = &mtk_vdec_hw_match[i]; + subdev_node = of_find_compatible_node(NULL, NULL, + of_id->compatible); + if (!subdev_node) + continue; + + of_node_put(subdev_node); + + hw_idx = (enum mtk_vdec_hw_id)(uintptr_t)of_id->data; + if (!test_bit(hw_idx, vdec_dev->subdev_bitmap)) { + dev_err(&pdev->dev, "vdec %d is not ready", hw_idx); + return -EAGAIN; + } + } + + return 0; +} + +static irqreturn_t mtk_vdec_hw_irq_handler(int irq, void *priv) +{ + struct mtk_vdec_hw_dev *dev = priv; + struct mtk_vcodec_dec_ctx *ctx; + u32 cg_status; + unsigned int dec_done_status; + void __iomem *vdec_misc_addr = dev->reg_base[VDEC_HW_MISC] + + VDEC_IRQ_CFG_REG; + + ctx = mtk_vcodec_get_curr_ctx(dev->main_dev, dev->hw_idx); + + /* check if HW active or not */ + cg_status = readl(dev->reg_base[VDEC_HW_SYS] + VDEC_HW_ACTIVE_ADDR); + if (cg_status & VDEC_HW_ACTIVE_MASK) { + mtk_v4l2_vdec_err(ctx, "vdec active is not 0x0 (0x%08x)", cg_status); + return IRQ_HANDLED; + } + + dec_done_status = readl(vdec_misc_addr); + if ((dec_done_status & MTK_VDEC_IRQ_STATUS_DEC_SUCCESS) != + MTK_VDEC_IRQ_STATUS_DEC_SUCCESS) + return IRQ_HANDLED; + + /* clear interrupt */ + writel(dec_done_status | VDEC_IRQ_CFG, vdec_misc_addr); + writel(dec_done_status & ~VDEC_IRQ_CLR, vdec_misc_addr); + + wake_up_dec_ctx(ctx, MTK_INST_IRQ_RECEIVED, dev->hw_idx); + + mtk_v4l2_vdec_dbg(3, ctx, "wake up ctx %d, dec_done_status=%x", + ctx->id, dec_done_status); + + return IRQ_HANDLED; +} + +static int mtk_vdec_hw_init_irq(struct mtk_vdec_hw_dev *dev) +{ + struct platform_device *pdev = dev->plat_dev; + int ret; + + dev->dec_irq = platform_get_irq(pdev, 0); + if (dev->dec_irq < 0) + return dev->dec_irq; + + irq_set_status_flags(dev->dec_irq, IRQ_NOAUTOEN); + ret = devm_request_irq(&pdev->dev, dev->dec_irq, + mtk_vdec_hw_irq_handler, 0, pdev->name, dev); + if (ret) { + dev_err(&pdev->dev, "Failed to install dev->dec_irq %d (%d)", + dev->dec_irq, ret); + return ret; + } + + return 0; +} + +static int mtk_vdec_hw_probe(struct platform_device *pdev) +{ + struct device *dev = &pdev->dev; + struct mtk_vdec_hw_dev *subdev_dev; + struct mtk_vcodec_dec_dev *main_dev; + const struct of_device_id *of_id; + int hw_idx; + int ret; + + if (!dev->parent) { + dev_err(dev, "no parent for hardware devices.\n"); + return -ENODEV; + } + + main_dev = dev_get_drvdata(dev->parent); + if (!main_dev) { + dev_err(dev, "failed to get parent driver data"); + return -EINVAL; + } + + subdev_dev = devm_kzalloc(dev, sizeof(*subdev_dev), GFP_KERNEL); + if (!subdev_dev) + return -ENOMEM; + + subdev_dev->plat_dev = pdev; + ret = mtk_vcodec_init_dec_clk(pdev, &subdev_dev->pm); + if (ret) + return ret; + + ret = devm_pm_runtime_enable(&pdev->dev); + if (ret) + return ret; + + of_id = of_match_device(mtk_vdec_hw_match, dev); + if (!of_id) { + dev_err(dev, "Can't get vdec subdev id.\n"); + return -EINVAL; + } + + hw_idx = (enum mtk_vdec_hw_id)(uintptr_t)of_id->data; + if (hw_idx >= MTK_VDEC_HW_MAX) { + dev_err(dev, "Hardware index %d not correct.\n", hw_idx); + return -EINVAL; + } + + main_dev->subdev_dev[hw_idx] = subdev_dev; + subdev_dev->hw_idx = hw_idx; + subdev_dev->main_dev = main_dev; + subdev_dev->reg_base[VDEC_HW_SYS] = main_dev->reg_base[VDEC_HW_SYS]; + set_bit(subdev_dev->hw_idx, main_dev->subdev_bitmap); + + if (IS_SUPPORT_VDEC_HW_IRQ(hw_idx)) { + ret = mtk_vdec_hw_init_irq(subdev_dev); + if (ret) + return ret; + } + + subdev_dev->reg_base[VDEC_HW_MISC] = + devm_platform_ioremap_resource(pdev, 0); + if (IS_ERR((__force void *)subdev_dev->reg_base[VDEC_HW_MISC])) { + ret = PTR_ERR((__force void *)subdev_dev->reg_base[VDEC_HW_MISC]); + return ret; + } + + if (!main_dev->subdev_prob_done) + main_dev->subdev_prob_done = mtk_vdec_hw_prob_done; + + platform_set_drvdata(pdev, subdev_dev); + return 0; +} + +static struct platform_driver mtk_vdec_driver = { + .probe = mtk_vdec_hw_probe, + .driver = { + .name = "mtk-vdec-comp", + .of_match_table = mtk_vdec_hw_match, + }, +}; +module_platform_driver(mtk_vdec_driver); + +MODULE_LICENSE("GPL v2"); +MODULE_DESCRIPTION("Mediatek video decoder hardware driver"); diff --git a/drivers/media/platform/mediatek/vcodec/decoder/mtk_vcodec_dec_hw.h b/drivers/media/platform/mediatek/vcodec/decoder/mtk_vcodec_dec_hw.h new file mode 100644 index 0000000000..83fe8b9428 --- /dev/null +++ b/drivers/media/platform/mediatek/vcodec/decoder/mtk_vcodec_dec_hw.h @@ -0,0 +1,59 @@ +/* SPDX-License-Identifier: GPL-2.0 */ +/* + * Copyright (c) 2021 MediaTek Inc. + * Author: Yunfei Dong <yunfei.dong@mediatek.com> + */ + +#ifndef _MTK_VCODEC_DEC_HW_H_ +#define _MTK_VCODEC_DEC_HW_H_ + +#include <linux/io.h> +#include <linux/platform_device.h> + +#include "mtk_vcodec_dec_drv.h" + +#define VDEC_HW_ACTIVE_ADDR 0x0 +#define VDEC_HW_ACTIVE_MASK BIT(4) +#define VDEC_IRQ_CFG 0x11 +#define VDEC_IRQ_CLR 0x10 +#define VDEC_IRQ_CFG_REG 0xa4 + +#define IS_SUPPORT_VDEC_HW_IRQ(hw_idx) ((hw_idx) != MTK_VDEC_LAT_SOC) + +/** + * enum mtk_vdec_hw_reg_idx - subdev hardware register base index + * @VDEC_HW_SYS : vdec soc register index + * @VDEC_HW_MISC: vdec misc register index + * @VDEC_HW_MAX : vdec supported max register index + */ +enum mtk_vdec_hw_reg_idx { + VDEC_HW_SYS, + VDEC_HW_MISC, + VDEC_HW_MAX +}; + +/** + * struct mtk_vdec_hw_dev - vdec hardware driver data + * @plat_dev: platform device + * @main_dev: main device + * @reg_base: mapped address of MTK Vcodec registers. + * + * @curr_ctx: the context that is waiting for codec hardware + * + * @dec_irq : decoder irq resource + * @pm : power management control + * @hw_idx : each hardware index + */ +struct mtk_vdec_hw_dev { + struct platform_device *plat_dev; + struct mtk_vcodec_dec_dev *main_dev; + void __iomem *reg_base[VDEC_HW_MAX]; + + struct mtk_vcodec_dec_ctx *curr_ctx; + + int dec_irq; + struct mtk_vcodec_pm pm; + int hw_idx; +}; + +#endif /* _MTK_VCODEC_DEC_HW_H_ */ diff --git a/drivers/media/platform/mediatek/vcodec/decoder/mtk_vcodec_dec_pm.c b/drivers/media/platform/mediatek/vcodec/decoder/mtk_vcodec_dec_pm.c new file mode 100644 index 0000000000..aefd3e9e30 --- /dev/null +++ b/drivers/media/platform/mediatek/vcodec/decoder/mtk_vcodec_dec_pm.c @@ -0,0 +1,262 @@ +// SPDX-License-Identifier: GPL-2.0 +/* + * Copyright (c) 2016 MediaTek Inc. + * Author: Tiffany Lin <tiffany.lin@mediatek.com> + */ + +#include <linux/clk.h> +#include <linux/interrupt.h> +#include <linux/of.h> +#include <linux/pm_runtime.h> + +#include "mtk_vcodec_dec_hw.h" +#include "mtk_vcodec_dec_pm.h" + +int mtk_vcodec_init_dec_clk(struct platform_device *pdev, struct mtk_vcodec_pm *pm) +{ + struct mtk_vcodec_clk *dec_clk; + struct mtk_vcodec_clk_info *clk_info; + int i = 0, ret; + + dec_clk = &pm->vdec_clk; + pm->dev = &pdev->dev; + + dec_clk->clk_num = + of_property_count_strings(pdev->dev.of_node, "clock-names"); + if (dec_clk->clk_num > 0) { + dec_clk->clk_info = devm_kcalloc(&pdev->dev, + dec_clk->clk_num, sizeof(*clk_info), + GFP_KERNEL); + if (!dec_clk->clk_info) + return -ENOMEM; + } else { + dev_err(&pdev->dev, "Failed to get vdec clock count"); + return -EINVAL; + } + + for (i = 0; i < dec_clk->clk_num; i++) { + clk_info = &dec_clk->clk_info[i]; + ret = of_property_read_string_index(pdev->dev.of_node, + "clock-names", i, &clk_info->clk_name); + if (ret) { + dev_err(&pdev->dev, "Failed to get clock name id = %d", i); + return ret; + } + clk_info->vcodec_clk = devm_clk_get(&pdev->dev, + clk_info->clk_name); + if (IS_ERR(clk_info->vcodec_clk)) { + dev_err(&pdev->dev, "devm_clk_get (%d)%s fail", i, clk_info->clk_name); + return PTR_ERR(clk_info->vcodec_clk); + } + } + + return 0; +} +EXPORT_SYMBOL_GPL(mtk_vcodec_init_dec_clk); + +static int mtk_vcodec_dec_pw_on(struct mtk_vcodec_pm *pm) +{ + int ret; + + ret = pm_runtime_resume_and_get(pm->dev); + if (ret) + dev_err(pm->dev, "pm_runtime_resume_and_get fail %d", ret); + + return ret; +} + +static void mtk_vcodec_dec_pw_off(struct mtk_vcodec_pm *pm) +{ + int ret; + + ret = pm_runtime_put(pm->dev); + if (ret && ret != -EAGAIN) + dev_err(pm->dev, "pm_runtime_put fail %d", ret); +} + +static void mtk_vcodec_dec_clock_on(struct mtk_vcodec_pm *pm) +{ + struct mtk_vcodec_clk *dec_clk; + int ret, i; + + dec_clk = &pm->vdec_clk; + for (i = 0; i < dec_clk->clk_num; i++) { + ret = clk_prepare_enable(dec_clk->clk_info[i].vcodec_clk); + if (ret) { + dev_err(pm->dev, "clk_prepare_enable %d %s fail %d", i, + dec_clk->clk_info[i].clk_name, ret); + goto error; + } + } + + return; +error: + for (i -= 1; i >= 0; i--) + clk_disable_unprepare(dec_clk->clk_info[i].vcodec_clk); +} + +static void mtk_vcodec_dec_clock_off(struct mtk_vcodec_pm *pm) +{ + struct mtk_vcodec_clk *dec_clk; + int i; + + dec_clk = &pm->vdec_clk; + for (i = dec_clk->clk_num - 1; i >= 0; i--) + clk_disable_unprepare(dec_clk->clk_info[i].vcodec_clk); +} + +static void mtk_vcodec_dec_enable_irq(struct mtk_vcodec_dec_dev *vdec_dev, int hw_idx) +{ + struct mtk_vdec_hw_dev *subdev_dev; + + if (!test_bit(hw_idx, vdec_dev->subdev_bitmap)) + return; + + if (vdec_dev->vdec_pdata->is_subdev_supported) { + subdev_dev = mtk_vcodec_get_hw_dev(vdec_dev, hw_idx); + if (subdev_dev) + enable_irq(subdev_dev->dec_irq); + else + dev_err(&vdec_dev->plat_dev->dev, "Failed to get hw dev\n"); + } else { + enable_irq(vdec_dev->dec_irq); + } +} + +static void mtk_vcodec_dec_disable_irq(struct mtk_vcodec_dec_dev *vdec_dev, int hw_idx) +{ + struct mtk_vdec_hw_dev *subdev_dev; + + if (!test_bit(hw_idx, vdec_dev->subdev_bitmap)) + return; + + if (vdec_dev->vdec_pdata->is_subdev_supported) { + subdev_dev = mtk_vcodec_get_hw_dev(vdec_dev, hw_idx); + if (subdev_dev) + disable_irq(subdev_dev->dec_irq); + else + dev_err(&vdec_dev->plat_dev->dev, "Failed to get hw dev\n"); + } else { + disable_irq(vdec_dev->dec_irq); + } +} + +static void mtk_vcodec_load_racing_info(struct mtk_vcodec_dec_ctx *ctx) +{ + void __iomem *vdec_racing_addr; + int j; + + mutex_lock(&ctx->dev->dec_racing_info_mutex); + if (atomic_inc_return(&ctx->dev->dec_active_cnt) == 1) { + vdec_racing_addr = ctx->dev->reg_base[VDEC_MISC] + 0x100; + for (j = 0; j < 132; j++) + writel(ctx->dev->vdec_racing_info[j], vdec_racing_addr + j * 4); + } + mutex_unlock(&ctx->dev->dec_racing_info_mutex); +} + +static void mtk_vcodec_record_racing_info(struct mtk_vcodec_dec_ctx *ctx) +{ + void __iomem *vdec_racing_addr; + int j; + + mutex_lock(&ctx->dev->dec_racing_info_mutex); + if (atomic_dec_and_test(&ctx->dev->dec_active_cnt)) { + vdec_racing_addr = ctx->dev->reg_base[VDEC_MISC] + 0x100; + for (j = 0; j < 132; j++) + ctx->dev->vdec_racing_info[j] = readl(vdec_racing_addr + j * 4); + } + mutex_unlock(&ctx->dev->dec_racing_info_mutex); +} + +static struct mtk_vcodec_pm *mtk_vcodec_dec_get_pm(struct mtk_vcodec_dec_dev *vdec_dev, + int hw_idx) +{ + struct mtk_vdec_hw_dev *subdev_dev; + + if (!test_bit(hw_idx, vdec_dev->subdev_bitmap)) + return NULL; + + if (vdec_dev->vdec_pdata->is_subdev_supported) { + subdev_dev = mtk_vcodec_get_hw_dev(vdec_dev, hw_idx); + if (subdev_dev) + return &subdev_dev->pm; + + dev_err(&vdec_dev->plat_dev->dev, "Failed to get hw dev\n"); + return NULL; + } + + return &vdec_dev->pm; +} + +static void mtk_vcodec_dec_child_dev_on(struct mtk_vcodec_dec_dev *vdec_dev, + int hw_idx) +{ + struct mtk_vcodec_pm *pm; + + pm = mtk_vcodec_dec_get_pm(vdec_dev, hw_idx); + if (pm) { + mtk_vcodec_dec_pw_on(pm); + mtk_vcodec_dec_clock_on(pm); + } + + if (hw_idx == MTK_VDEC_LAT0) { + pm = mtk_vcodec_dec_get_pm(vdec_dev, MTK_VDEC_LAT_SOC); + if (pm) { + mtk_vcodec_dec_pw_on(pm); + mtk_vcodec_dec_clock_on(pm); + } + } +} + +static void mtk_vcodec_dec_child_dev_off(struct mtk_vcodec_dec_dev *vdec_dev, + int hw_idx) +{ + struct mtk_vcodec_pm *pm; + + pm = mtk_vcodec_dec_get_pm(vdec_dev, hw_idx); + if (pm) { + mtk_vcodec_dec_clock_off(pm); + mtk_vcodec_dec_pw_off(pm); + } + + if (hw_idx == MTK_VDEC_LAT0) { + pm = mtk_vcodec_dec_get_pm(vdec_dev, MTK_VDEC_LAT_SOC); + if (pm) { + mtk_vcodec_dec_clock_off(pm); + mtk_vcodec_dec_pw_off(pm); + } + } +} + +void mtk_vcodec_dec_enable_hardware(struct mtk_vcodec_dec_ctx *ctx, int hw_idx) +{ + mutex_lock(&ctx->dev->dec_mutex[hw_idx]); + + if (IS_VDEC_LAT_ARCH(ctx->dev->vdec_pdata->hw_arch) && + hw_idx == MTK_VDEC_CORE) + mtk_vcodec_dec_child_dev_on(ctx->dev, MTK_VDEC_LAT0); + mtk_vcodec_dec_child_dev_on(ctx->dev, hw_idx); + + mtk_vcodec_dec_enable_irq(ctx->dev, hw_idx); + + if (IS_VDEC_INNER_RACING(ctx->dev->dec_capability)) + mtk_vcodec_load_racing_info(ctx); +} +EXPORT_SYMBOL_GPL(mtk_vcodec_dec_enable_hardware); + +void mtk_vcodec_dec_disable_hardware(struct mtk_vcodec_dec_ctx *ctx, int hw_idx) +{ + if (IS_VDEC_INNER_RACING(ctx->dev->dec_capability)) + mtk_vcodec_record_racing_info(ctx); + + mtk_vcodec_dec_disable_irq(ctx->dev, hw_idx); + + mtk_vcodec_dec_child_dev_off(ctx->dev, hw_idx); + if (IS_VDEC_LAT_ARCH(ctx->dev->vdec_pdata->hw_arch) && + hw_idx == MTK_VDEC_CORE) + mtk_vcodec_dec_child_dev_off(ctx->dev, MTK_VDEC_LAT0); + + mutex_unlock(&ctx->dev->dec_mutex[hw_idx]); +} +EXPORT_SYMBOL_GPL(mtk_vcodec_dec_disable_hardware); diff --git a/drivers/media/platform/mediatek/vcodec/decoder/mtk_vcodec_dec_pm.h b/drivers/media/platform/mediatek/vcodec/decoder/mtk_vcodec_dec_pm.h new file mode 100644 index 0000000000..87a50d589d --- /dev/null +++ b/drivers/media/platform/mediatek/vcodec/decoder/mtk_vcodec_dec_pm.h @@ -0,0 +1,17 @@ +/* SPDX-License-Identifier: GPL-2.0 */ +/* + * Copyright (c) 2016 MediaTek Inc. + * Author: Tiffany Lin <tiffany.lin@mediatek.com> + */ + +#ifndef _MTK_VCODEC_DEC_PM_H_ +#define _MTK_VCODEC_DEC_PM_H_ + +#include "mtk_vcodec_dec_drv.h" + +int mtk_vcodec_init_dec_clk(struct platform_device *pdev, struct mtk_vcodec_pm *pm); + +void mtk_vcodec_dec_enable_hardware(struct mtk_vcodec_dec_ctx *ctx, int hw_idx); +void mtk_vcodec_dec_disable_hardware(struct mtk_vcodec_dec_ctx *ctx, int hw_idx); + +#endif /* _MTK_VCODEC_DEC_PM_H_ */ diff --git a/drivers/media/platform/mediatek/vcodec/decoder/mtk_vcodec_dec_stateful.c b/drivers/media/platform/mediatek/vcodec/decoder/mtk_vcodec_dec_stateful.c new file mode 100644 index 0000000000..11ca2c2fba --- /dev/null +++ b/drivers/media/platform/mediatek/vcodec/decoder/mtk_vcodec_dec_stateful.c @@ -0,0 +1,623 @@ +// SPDX-License-Identifier: GPL-2.0 + +#include <media/v4l2-event.h> +#include <media/v4l2-mem2mem.h> +#include <media/videobuf2-dma-contig.h> + +#include "mtk_vcodec_dec.h" +#include "mtk_vcodec_dec_pm.h" +#include "vdec_drv_if.h" + +static struct mtk_video_fmt mtk_video_formats[] = { + { + .fourcc = V4L2_PIX_FMT_H264, + .type = MTK_FMT_DEC, + .num_planes = 1, + .flags = V4L2_FMT_FLAG_DYN_RESOLUTION, + .frmsize = { MTK_VDEC_MIN_W, MTK_VDEC_MAX_W, 16, + MTK_VDEC_MIN_H, MTK_VDEC_MAX_H, 16 }, + }, + { + .fourcc = V4L2_PIX_FMT_VP8, + .type = MTK_FMT_DEC, + .num_planes = 1, + .flags = V4L2_FMT_FLAG_DYN_RESOLUTION, + .frmsize = { MTK_VDEC_MIN_W, MTK_VDEC_MAX_W, 16, + MTK_VDEC_MIN_H, MTK_VDEC_MAX_H, 16 }, + }, + { + .fourcc = V4L2_PIX_FMT_VP9, + .type = MTK_FMT_DEC, + .num_planes = 1, + .flags = V4L2_FMT_FLAG_DYN_RESOLUTION, + .frmsize = { MTK_VDEC_MIN_W, MTK_VDEC_MAX_W, 16, + MTK_VDEC_MIN_H, MTK_VDEC_MAX_H, 16 }, + }, + { + .fourcc = V4L2_PIX_FMT_MT21C, + .type = MTK_FMT_FRAME, + .num_planes = 2, + }, +}; + +static const unsigned int num_supported_formats = + ARRAY_SIZE(mtk_video_formats); + +#define DEFAULT_OUT_FMT_IDX 0 +#define DEFAULT_CAP_FMT_IDX 3 + +/* + * This function tries to clean all display buffers, the buffers will return + * in display order. + * Note the buffers returned from codec driver may still be in driver's + * reference list. + */ +static struct vb2_buffer *get_display_buffer(struct mtk_vcodec_dec_ctx *ctx) +{ + struct vdec_fb *disp_frame_buffer = NULL; + struct mtk_video_dec_buf *dstbuf; + struct vb2_v4l2_buffer *vb; + + mtk_v4l2_vdec_dbg(3, ctx, "[%d]", ctx->id); + if (vdec_if_get_param(ctx, GET_PARAM_DISP_FRAME_BUFFER, + &disp_frame_buffer)) { + mtk_v4l2_vdec_err(ctx, "[%d]Cannot get param : GET_PARAM_DISP_FRAME_BUFFER", + ctx->id); + return NULL; + } + + if (!disp_frame_buffer) { + mtk_v4l2_vdec_dbg(3, ctx, "No display frame buffer"); + return NULL; + } + + dstbuf = container_of(disp_frame_buffer, struct mtk_video_dec_buf, + frame_buffer); + vb = &dstbuf->m2m_buf.vb; + mutex_lock(&ctx->lock); + if (dstbuf->used) { + mtk_v4l2_vdec_dbg(2, ctx, "[%d]status=%x queue id=%d to done_list %d", + ctx->id, disp_frame_buffer->status, + vb->vb2_buf.index, dstbuf->queued_in_vb2); + + v4l2_m2m_buf_done(vb, VB2_BUF_STATE_DONE); + ctx->decoded_frame_cnt++; + } + mutex_unlock(&ctx->lock); + return &vb->vb2_buf; +} + +/* + * This function tries to clean all capture buffers that are not used as + * reference buffers by codec driver any more + * In this case, we need re-queue buffer to vb2 buffer if user space + * already returns this buffer to v4l2 or this buffer is just the output of + * previous sps/pps/resolution change decode, or do nothing if user + * space still owns this buffer + */ +static struct vb2_buffer *get_free_buffer(struct mtk_vcodec_dec_ctx *ctx) +{ + struct mtk_video_dec_buf *dstbuf; + struct vdec_fb *free_frame_buffer = NULL; + struct vb2_v4l2_buffer *vb; + + if (vdec_if_get_param(ctx, GET_PARAM_FREE_FRAME_BUFFER, + &free_frame_buffer)) { + mtk_v4l2_vdec_err(ctx, "[%d] Error!! Cannot get param", ctx->id); + return NULL; + } + if (!free_frame_buffer) { + mtk_v4l2_vdec_dbg(3, ctx, " No free frame buffer"); + return NULL; + } + + mtk_v4l2_vdec_dbg(3, ctx, "[%d] tmp_frame_addr = 0x%p", ctx->id, + free_frame_buffer); + + dstbuf = container_of(free_frame_buffer, struct mtk_video_dec_buf, + frame_buffer); + vb = &dstbuf->m2m_buf.vb; + + mutex_lock(&ctx->lock); + if (dstbuf->used) { + if (dstbuf->queued_in_vb2 && dstbuf->queued_in_v4l2 && + free_frame_buffer->status == FB_ST_FREE) { + /* + * After decode sps/pps or non-display buffer, we don't + * need to return capture buffer to user space, but + * just re-queue this capture buffer to vb2 queue. + * This reduce overheads that dq/q unused capture + * buffer. In this case, queued_in_vb2 = true. + */ + mtk_v4l2_vdec_dbg(2, ctx, "[%d]status=%x queue id=%d to rdy_queue %d", + ctx->id, free_frame_buffer->status, + vb->vb2_buf.index, dstbuf->queued_in_vb2); + v4l2_m2m_buf_queue(ctx->m2m_ctx, vb); + } else if (!dstbuf->queued_in_vb2 && dstbuf->queued_in_v4l2) { + /* + * If buffer in v4l2 driver but not in vb2 queue yet, + * and we get this buffer from free_list, it means + * that codec driver do not use this buffer as + * reference buffer anymore. We should q buffer to vb2 + * queue, so later work thread could get this buffer + * for decode. In this case, queued_in_vb2 = false + * means this buffer is not from previous decode + * output. + */ + mtk_v4l2_vdec_dbg(2, ctx, + "[%d]status=%x queue id=%d to rdy_queue", + ctx->id, free_frame_buffer->status, + vb->vb2_buf.index); + v4l2_m2m_buf_queue(ctx->m2m_ctx, vb); + dstbuf->queued_in_vb2 = true; + } else { + /* + * Codec driver do not need to reference this capture + * buffer and this buffer is not in v4l2 driver. + * Then we don't need to do any thing, just add log when + * we need to debug buffer flow. + * When this buffer q from user space, it could + * directly q to vb2 buffer + */ + mtk_v4l2_vdec_dbg(3, ctx, "[%d]status=%x err queue id=%d %d %d", + ctx->id, free_frame_buffer->status, + vb->vb2_buf.index, dstbuf->queued_in_vb2, + dstbuf->queued_in_v4l2); + } + dstbuf->used = false; + } + mutex_unlock(&ctx->lock); + return &vb->vb2_buf; +} + +static void clean_display_buffer(struct mtk_vcodec_dec_ctx *ctx) +{ + while (get_display_buffer(ctx)) + ; +} + +static void clean_free_buffer(struct mtk_vcodec_dec_ctx *ctx) +{ + while (get_free_buffer(ctx)) + ; +} + +static void mtk_vdec_queue_res_chg_event(struct mtk_vcodec_dec_ctx *ctx) +{ + static const struct v4l2_event ev_src_ch = { + .type = V4L2_EVENT_SOURCE_CHANGE, + .u.src_change.changes = V4L2_EVENT_SRC_CH_RESOLUTION, + }; + + mtk_v4l2_vdec_dbg(1, ctx, "[%d]", ctx->id); + v4l2_event_queue_fh(&ctx->fh, &ev_src_ch); +} + +static int mtk_vdec_flush_decoder(struct mtk_vcodec_dec_ctx *ctx) +{ + bool res_chg; + int ret; + + ret = vdec_if_decode(ctx, NULL, NULL, &res_chg); + if (ret) + mtk_v4l2_vdec_err(ctx, "DecodeFinal failed, ret=%d", ret); + + clean_display_buffer(ctx); + clean_free_buffer(ctx); + + return 0; +} + +static void mtk_vdec_update_fmt(struct mtk_vcodec_dec_ctx *ctx, + unsigned int pixelformat) +{ + const struct mtk_video_fmt *fmt; + struct mtk_q_data *dst_q_data; + unsigned int k; + + dst_q_data = &ctx->q_data[MTK_Q_DATA_DST]; + for (k = 0; k < num_supported_formats; k++) { + fmt = &mtk_video_formats[k]; + if (fmt->fourcc == pixelformat) { + mtk_v4l2_vdec_dbg(1, ctx, "Update cap fourcc(%d -> %d)", + dst_q_data->fmt->fourcc, pixelformat); + dst_q_data->fmt = fmt; + return; + } + } + + mtk_v4l2_vdec_err(ctx, "Cannot get fourcc(%d), using init value", pixelformat); +} + +static int mtk_vdec_pic_info_update(struct mtk_vcodec_dec_ctx *ctx) +{ + unsigned int dpbsize = 0; + int ret; + + if (vdec_if_get_param(ctx, GET_PARAM_PIC_INFO, + &ctx->last_decoded_picinfo)) { + mtk_v4l2_vdec_err(ctx, "[%d]Error!! Cannot get param : GET_PARAM_PICTURE_INFO ERR", + ctx->id); + return -EINVAL; + } + + if (ctx->last_decoded_picinfo.pic_w == 0 || + ctx->last_decoded_picinfo.pic_h == 0 || + ctx->last_decoded_picinfo.buf_w == 0 || + ctx->last_decoded_picinfo.buf_h == 0) { + mtk_v4l2_vdec_err(ctx, "Cannot get correct pic info"); + return -EINVAL; + } + + if (ctx->last_decoded_picinfo.cap_fourcc != ctx->picinfo.cap_fourcc && + ctx->picinfo.cap_fourcc != 0) + mtk_vdec_update_fmt(ctx, ctx->picinfo.cap_fourcc); + + if (ctx->last_decoded_picinfo.pic_w == ctx->picinfo.pic_w || + ctx->last_decoded_picinfo.pic_h == ctx->picinfo.pic_h) + return 0; + + mtk_v4l2_vdec_dbg(1, ctx, "[%d]-> new(%d,%d), old(%d,%d), real(%d,%d)", ctx->id, + ctx->last_decoded_picinfo.pic_w, + ctx->last_decoded_picinfo.pic_h, ctx->picinfo.pic_w, + ctx->picinfo.pic_h, ctx->last_decoded_picinfo.buf_w, + ctx->last_decoded_picinfo.buf_h); + + ret = vdec_if_get_param(ctx, GET_PARAM_DPB_SIZE, &dpbsize); + if (dpbsize == 0) + mtk_v4l2_vdec_err(ctx, "Incorrect dpb size, ret=%d", ret); + + ctx->dpb_size = dpbsize; + + return ret; +} + +static void mtk_vdec_worker(struct work_struct *work) +{ + struct mtk_vcodec_dec_ctx *ctx = + container_of(work, struct mtk_vcodec_dec_ctx, decode_work); + struct mtk_vcodec_dec_dev *dev = ctx->dev; + struct vb2_v4l2_buffer *src_buf, *dst_buf; + struct mtk_vcodec_mem buf; + struct vdec_fb *pfb; + bool res_chg = false; + int ret; + struct mtk_video_dec_buf *dst_buf_info, *src_buf_info; + + src_buf = v4l2_m2m_next_src_buf(ctx->m2m_ctx); + if (!src_buf) { + v4l2_m2m_job_finish(dev->m2m_dev_dec, ctx->m2m_ctx); + mtk_v4l2_vdec_dbg(1, ctx, "[%d] src_buf empty!!", ctx->id); + return; + } + + dst_buf = v4l2_m2m_next_dst_buf(ctx->m2m_ctx); + if (!dst_buf) { + v4l2_m2m_job_finish(dev->m2m_dev_dec, ctx->m2m_ctx); + mtk_v4l2_vdec_dbg(1, ctx, "[%d] dst_buf empty!!", ctx->id); + return; + } + + dst_buf_info = + container_of(dst_buf, struct mtk_video_dec_buf, m2m_buf.vb); + + pfb = &dst_buf_info->frame_buffer; + pfb->base_y.va = vb2_plane_vaddr(&dst_buf->vb2_buf, 0); + pfb->base_y.dma_addr = + vb2_dma_contig_plane_dma_addr(&dst_buf->vb2_buf, 0); + pfb->base_y.size = ctx->picinfo.fb_sz[0]; + + pfb->base_c.va = vb2_plane_vaddr(&dst_buf->vb2_buf, 1); + pfb->base_c.dma_addr = + vb2_dma_contig_plane_dma_addr(&dst_buf->vb2_buf, 1); + pfb->base_c.size = ctx->picinfo.fb_sz[1]; + pfb->status = 0; + mtk_v4l2_vdec_dbg(3, ctx, "===>[%d] vdec_if_decode() ===>", ctx->id); + + mtk_v4l2_vdec_dbg(3, ctx, + "id=%d Framebuf pfb=%p VA=%p Y_DMA=%pad C_DMA=%pad Size=%zx", + dst_buf->vb2_buf.index, pfb, pfb->base_y.va, + &pfb->base_y.dma_addr, &pfb->base_c.dma_addr, pfb->base_y.size); + + if (src_buf == &ctx->empty_flush_buf.vb) { + mtk_v4l2_vdec_dbg(1, ctx, "Got empty flush input buffer."); + src_buf = v4l2_m2m_src_buf_remove(ctx->m2m_ctx); + + /* update dst buf status */ + dst_buf = v4l2_m2m_dst_buf_remove(ctx->m2m_ctx); + mutex_lock(&ctx->lock); + dst_buf_info->used = false; + mutex_unlock(&ctx->lock); + + vdec_if_decode(ctx, NULL, NULL, &res_chg); + clean_display_buffer(ctx); + vb2_set_plane_payload(&dst_buf->vb2_buf, 0, 0); + if (ctx->q_data[MTK_Q_DATA_DST].fmt->num_planes == 2) + vb2_set_plane_payload(&dst_buf->vb2_buf, 1, 0); + dst_buf->flags |= V4L2_BUF_FLAG_LAST; + v4l2_m2m_buf_done(dst_buf, VB2_BUF_STATE_DONE); + clean_free_buffer(ctx); + v4l2_m2m_job_finish(dev->m2m_dev_dec, ctx->m2m_ctx); + return; + } + + src_buf_info = + container_of(src_buf, struct mtk_video_dec_buf, m2m_buf.vb); + + buf.va = vb2_plane_vaddr(&src_buf->vb2_buf, 0); + buf.dma_addr = vb2_dma_contig_plane_dma_addr(&src_buf->vb2_buf, 0); + buf.size = (size_t)src_buf->vb2_buf.planes[0].bytesused; + if (!buf.va) { + v4l2_m2m_job_finish(dev->m2m_dev_dec, ctx->m2m_ctx); + mtk_v4l2_vdec_err(ctx, "[%d] id=%d src_addr is NULL!!", ctx->id, + src_buf->vb2_buf.index); + return; + } + mtk_v4l2_vdec_dbg(3, ctx, "[%d] Bitstream VA=%p DMA=%pad Size=%zx vb=%p", + ctx->id, buf.va, &buf.dma_addr, buf.size, src_buf); + dst_buf->vb2_buf.timestamp = src_buf->vb2_buf.timestamp; + dst_buf->timecode = src_buf->timecode; + mutex_lock(&ctx->lock); + dst_buf_info->used = true; + mutex_unlock(&ctx->lock); + src_buf_info->used = true; + + ret = vdec_if_decode(ctx, &buf, pfb, &res_chg); + + if (ret) { + mtk_v4l2_vdec_err(ctx, + "[%d] decode src[%d] sz=0x%zx pts=%llu dst[%d] ret=%d res_chg=%d", + ctx->id, src_buf->vb2_buf.index, buf.size, + src_buf->vb2_buf.timestamp, dst_buf->vb2_buf.index, ret, res_chg); + src_buf = v4l2_m2m_src_buf_remove(ctx->m2m_ctx); + if (ret == -EIO) { + mutex_lock(&ctx->lock); + src_buf_info->error = true; + mutex_unlock(&ctx->lock); + } + v4l2_m2m_buf_done(src_buf, VB2_BUF_STATE_ERROR); + } else if (!res_chg) { + /* + * we only return src buffer with VB2_BUF_STATE_DONE + * when decode success without resolution change + */ + src_buf = v4l2_m2m_src_buf_remove(ctx->m2m_ctx); + v4l2_m2m_buf_done(src_buf, VB2_BUF_STATE_DONE); + } + + dst_buf = v4l2_m2m_dst_buf_remove(ctx->m2m_ctx); + clean_display_buffer(ctx); + clean_free_buffer(ctx); + + if (!ret && res_chg) { + mtk_vdec_pic_info_update(ctx); + /* + * On encountering a resolution change in the stream. + * The driver must first process and decode all + * remaining buffers from before the resolution change + * point, so call flush decode here + */ + mtk_vdec_flush_decoder(ctx); + /* + * After all buffers containing decoded frames from + * before the resolution change point ready to be + * dequeued on the CAPTURE queue, the driver sends a + * V4L2_EVENT_SOURCE_CHANGE event for source change + * type V4L2_EVENT_SRC_CH_RESOLUTION + */ + mtk_vdec_queue_res_chg_event(ctx); + } + v4l2_m2m_job_finish(dev->m2m_dev_dec, ctx->m2m_ctx); +} + +static void vb2ops_vdec_stateful_buf_queue(struct vb2_buffer *vb) +{ + struct vb2_v4l2_buffer *src_buf; + struct mtk_vcodec_mem src_mem; + bool res_chg = false; + int ret; + unsigned int dpbsize = 1, i; + struct mtk_vcodec_dec_ctx *ctx = vb2_get_drv_priv(vb->vb2_queue); + struct vb2_v4l2_buffer *vb2_v4l2; + struct mtk_q_data *dst_q_data; + + mtk_v4l2_vdec_dbg(3, ctx, "[%d] (%d) id=%d, vb=%p", ctx->id, + vb->vb2_queue->type, vb->index, vb); + /* + * check if this buffer is ready to be used after decode + */ + if (vb->vb2_queue->type != V4L2_BUF_TYPE_VIDEO_OUTPUT_MPLANE) { + struct mtk_video_dec_buf *buf; + + vb2_v4l2 = to_vb2_v4l2_buffer(vb); + buf = container_of(vb2_v4l2, struct mtk_video_dec_buf, + m2m_buf.vb); + mutex_lock(&ctx->lock); + if (!buf->used) { + v4l2_m2m_buf_queue(ctx->m2m_ctx, vb2_v4l2); + buf->queued_in_vb2 = true; + buf->queued_in_v4l2 = true; + } else { + buf->queued_in_vb2 = false; + buf->queued_in_v4l2 = true; + } + mutex_unlock(&ctx->lock); + return; + } + + v4l2_m2m_buf_queue(ctx->m2m_ctx, to_vb2_v4l2_buffer(vb)); + + if (ctx->state != MTK_STATE_INIT) { + mtk_v4l2_vdec_dbg(3, ctx, "[%d] already init driver %d", ctx->id, ctx->state); + return; + } + + src_buf = v4l2_m2m_next_src_buf(ctx->m2m_ctx); + if (!src_buf) { + mtk_v4l2_vdec_err(ctx, "No src buffer"); + return; + } + + if (src_buf == &ctx->empty_flush_buf.vb) { + /* This shouldn't happen. Just in case. */ + mtk_v4l2_vdec_err(ctx, "Invalid flush buffer."); + v4l2_m2m_src_buf_remove(ctx->m2m_ctx); + return; + } + + src_mem.va = vb2_plane_vaddr(&src_buf->vb2_buf, 0); + src_mem.dma_addr = vb2_dma_contig_plane_dma_addr(&src_buf->vb2_buf, 0); + src_mem.size = (size_t)src_buf->vb2_buf.planes[0].bytesused; + mtk_v4l2_vdec_dbg(2, ctx, "[%d] buf id=%d va=%p dma=%pad size=%zx", ctx->id, + src_buf->vb2_buf.index, src_mem.va, &src_mem.dma_addr, src_mem.size); + + ret = vdec_if_decode(ctx, &src_mem, NULL, &res_chg); + if (ret || !res_chg) { + /* + * fb == NULL means to parse SPS/PPS header or + * resolution info in src_mem. Decode can fail + * if there is no SPS header or picture info + * in bs + */ + + src_buf = v4l2_m2m_src_buf_remove(ctx->m2m_ctx); + if (ret == -EIO) { + mtk_v4l2_vdec_err(ctx, "[%d] Unrecoverable error in vdec_if_decode.", + ctx->id); + ctx->state = MTK_STATE_ABORT; + v4l2_m2m_buf_done(src_buf, VB2_BUF_STATE_ERROR); + } else { + v4l2_m2m_buf_done(src_buf, VB2_BUF_STATE_DONE); + } + mtk_v4l2_vdec_dbg(ret ? 0 : 1, ctx, + "[%d] decode() src_buf=%d, size=%zu, fail=%d, res_chg=%d", + ctx->id, src_buf->vb2_buf.index, src_mem.size, ret, res_chg); + return; + } + + if (vdec_if_get_param(ctx, GET_PARAM_PIC_INFO, &ctx->picinfo)) { + mtk_v4l2_vdec_err(ctx, "[%d]Error!! Cannot get param : GET_PARAM_PICTURE_INFO ERR", + ctx->id); + return; + } + + ctx->last_decoded_picinfo = ctx->picinfo; + dst_q_data = &ctx->q_data[MTK_Q_DATA_DST]; + for (i = 0; i < dst_q_data->fmt->num_planes; i++) { + dst_q_data->sizeimage[i] = ctx->picinfo.fb_sz[i]; + dst_q_data->bytesperline[i] = ctx->picinfo.buf_w; + } + + mtk_v4l2_vdec_dbg(2, ctx, "[%d] init OK wxh=%dx%d pic wxh=%dx%d sz[0]=0x%x sz[1]=0x%x", + ctx->id, ctx->picinfo.buf_w, ctx->picinfo.buf_h, ctx->picinfo.pic_w, + ctx->picinfo.pic_h, dst_q_data->sizeimage[0], dst_q_data->sizeimage[1]); + + ret = vdec_if_get_param(ctx, GET_PARAM_DPB_SIZE, &dpbsize); + if (dpbsize == 0) + mtk_v4l2_vdec_err(ctx, "[%d] GET_PARAM_DPB_SIZE fail=%d", ctx->id, ret); + + ctx->dpb_size = dpbsize; + ctx->state = MTK_STATE_HEADER; + mtk_v4l2_vdec_dbg(1, ctx, "[%d] dpbsize=%d", ctx->id, ctx->dpb_size); + + mtk_vdec_queue_res_chg_event(ctx); +} + +static int mtk_vdec_g_v_ctrl(struct v4l2_ctrl *ctrl) +{ + struct mtk_vcodec_dec_ctx *ctx = ctrl_to_dec_ctx(ctrl); + int ret = 0; + + switch (ctrl->id) { + case V4L2_CID_MIN_BUFFERS_FOR_CAPTURE: + if (ctx->state >= MTK_STATE_HEADER) { + ctrl->val = ctx->dpb_size; + } else { + mtk_v4l2_vdec_dbg(0, ctx, "Seqinfo not ready"); + ctrl->val = 0; + } + break; + default: + ret = -EINVAL; + } + return ret; +} + +static const struct v4l2_ctrl_ops mtk_vcodec_dec_ctrl_ops = { + .g_volatile_ctrl = mtk_vdec_g_v_ctrl, +}; + +static int mtk_vcodec_dec_ctrls_setup(struct mtk_vcodec_dec_ctx *ctx) +{ + struct v4l2_ctrl *ctrl; + + v4l2_ctrl_handler_init(&ctx->ctrl_hdl, 1); + + ctrl = v4l2_ctrl_new_std(&ctx->ctrl_hdl, &mtk_vcodec_dec_ctrl_ops, + V4L2_CID_MIN_BUFFERS_FOR_CAPTURE, 0, 32, 1, 1); + ctrl->flags |= V4L2_CTRL_FLAG_VOLATILE; + v4l2_ctrl_new_std_menu(&ctx->ctrl_hdl, &mtk_vcodec_dec_ctrl_ops, + V4L2_CID_MPEG_VIDEO_VP9_PROFILE, + V4L2_MPEG_VIDEO_VP9_PROFILE_0, 0, + V4L2_MPEG_VIDEO_VP9_PROFILE_0); + /* + * H264. Baseline / Extended decoding is not supported. + */ + v4l2_ctrl_new_std_menu(&ctx->ctrl_hdl, &mtk_vcodec_dec_ctrl_ops, + V4L2_CID_MPEG_VIDEO_H264_PROFILE, V4L2_MPEG_VIDEO_H264_PROFILE_HIGH, + BIT(V4L2_MPEG_VIDEO_H264_PROFILE_BASELINE) | + BIT(V4L2_MPEG_VIDEO_H264_PROFILE_EXTENDED), + V4L2_MPEG_VIDEO_H264_PROFILE_MAIN); + + if (ctx->ctrl_hdl.error) { + mtk_v4l2_vdec_err(ctx, "Adding control failed %d", ctx->ctrl_hdl.error); + return ctx->ctrl_hdl.error; + } + + v4l2_ctrl_handler_setup(&ctx->ctrl_hdl); + return 0; +} + +static void mtk_init_vdec_params(struct mtk_vcodec_dec_ctx *ctx) +{ + unsigned int i; + + if (!(ctx->dev->dec_capability & VCODEC_CAPABILITY_4K_DISABLED)) { + for (i = 0; i < num_supported_formats; i++) { + if (mtk_video_formats[i].type != MTK_FMT_DEC) + continue; + + mtk_video_formats[i].frmsize.max_width = + VCODEC_DEC_4K_CODED_WIDTH; + mtk_video_formats[i].frmsize.max_height = + VCODEC_DEC_4K_CODED_HEIGHT; + } + } +} + +static struct vb2_ops mtk_vdec_frame_vb2_ops = { + .queue_setup = vb2ops_vdec_queue_setup, + .buf_prepare = vb2ops_vdec_buf_prepare, + .wait_prepare = vb2_ops_wait_prepare, + .wait_finish = vb2_ops_wait_finish, + .start_streaming = vb2ops_vdec_start_streaming, + + .buf_queue = vb2ops_vdec_stateful_buf_queue, + .buf_init = vb2ops_vdec_buf_init, + .buf_finish = vb2ops_vdec_buf_finish, + .stop_streaming = vb2ops_vdec_stop_streaming, +}; + +const struct mtk_vcodec_dec_pdata mtk_vdec_8173_pdata = { + .init_vdec_params = mtk_init_vdec_params, + .ctrls_setup = mtk_vcodec_dec_ctrls_setup, + .vdec_vb2_ops = &mtk_vdec_frame_vb2_ops, + .vdec_formats = mtk_video_formats, + .num_formats = &num_supported_formats, + .default_out_fmt = &mtk_video_formats[DEFAULT_OUT_FMT_IDX], + .default_cap_fmt = &mtk_video_formats[DEFAULT_CAP_FMT_IDX], + .worker = mtk_vdec_worker, + .flush_decoder = mtk_vdec_flush_decoder, + .is_subdev_supported = false, + .hw_arch = MTK_VDEC_PURE_SINGLE_CORE, +}; diff --git a/drivers/media/platform/mediatek/vcodec/decoder/mtk_vcodec_dec_stateless.c b/drivers/media/platform/mediatek/vcodec/decoder/mtk_vcodec_dec_stateless.c new file mode 100644 index 0000000000..e29c9c58f3 --- /dev/null +++ b/drivers/media/platform/mediatek/vcodec/decoder/mtk_vcodec_dec_stateless.c @@ -0,0 +1,756 @@ +// SPDX-License-Identifier: GPL-2.0 + +#include <media/videobuf2-v4l2.h> +#include <media/videobuf2-dma-contig.h> +#include <media/v4l2-event.h> +#include <media/v4l2-mem2mem.h> +#include <linux/module.h> + +#include "mtk_vcodec_dec.h" +#include "mtk_vcodec_dec_pm.h" +#include "vdec_drv_if.h" + +/** + * struct mtk_stateless_control - CID control type + * @cfg: control configuration + * @codec_type: codec type (V4L2 pixel format) for CID control type + */ +struct mtk_stateless_control { + struct v4l2_ctrl_config cfg; + int codec_type; +}; + +static const struct mtk_stateless_control mtk_stateless_controls[] = { + { + .cfg = { + .id = V4L2_CID_STATELESS_H264_SPS, + }, + .codec_type = V4L2_PIX_FMT_H264_SLICE, + }, + { + .cfg = { + .id = V4L2_CID_STATELESS_H264_PPS, + }, + .codec_type = V4L2_PIX_FMT_H264_SLICE, + }, + { + .cfg = { + .id = V4L2_CID_STATELESS_H264_SCALING_MATRIX, + }, + .codec_type = V4L2_PIX_FMT_H264_SLICE, + }, + { + .cfg = { + .id = V4L2_CID_STATELESS_H264_DECODE_PARAMS, + }, + .codec_type = V4L2_PIX_FMT_H264_SLICE, + }, + { + .cfg = { + .id = V4L2_CID_MPEG_VIDEO_H264_PROFILE, + .def = V4L2_MPEG_VIDEO_H264_PROFILE_MAIN, + .max = V4L2_MPEG_VIDEO_H264_PROFILE_HIGH, + .menu_skip_mask = + BIT(V4L2_MPEG_VIDEO_H264_PROFILE_BASELINE) | + BIT(V4L2_MPEG_VIDEO_H264_PROFILE_EXTENDED), + }, + .codec_type = V4L2_PIX_FMT_H264_SLICE, + }, + { + .cfg = { + .id = V4L2_CID_STATELESS_H264_DECODE_MODE, + .min = V4L2_STATELESS_H264_DECODE_MODE_FRAME_BASED, + .def = V4L2_STATELESS_H264_DECODE_MODE_FRAME_BASED, + .max = V4L2_STATELESS_H264_DECODE_MODE_FRAME_BASED, + }, + .codec_type = V4L2_PIX_FMT_H264_SLICE, + }, + { + .cfg = { + .id = V4L2_CID_STATELESS_H264_START_CODE, + .min = V4L2_STATELESS_H264_START_CODE_ANNEX_B, + .def = V4L2_STATELESS_H264_START_CODE_ANNEX_B, + .max = V4L2_STATELESS_H264_START_CODE_ANNEX_B, + }, + .codec_type = V4L2_PIX_FMT_H264_SLICE, + }, + { + .cfg = { + .id = V4L2_CID_STATELESS_VP8_FRAME, + }, + .codec_type = V4L2_PIX_FMT_VP8_FRAME, + }, + { + .cfg = { + .id = V4L2_CID_MPEG_VIDEO_VP8_PROFILE, + .min = V4L2_MPEG_VIDEO_VP8_PROFILE_0, + .def = V4L2_MPEG_VIDEO_VP8_PROFILE_0, + .max = V4L2_MPEG_VIDEO_VP8_PROFILE_3, + }, + .codec_type = V4L2_PIX_FMT_VP8_FRAME, + }, + { + .cfg = { + .id = V4L2_CID_STATELESS_VP9_FRAME, + }, + .codec_type = V4L2_PIX_FMT_VP9_FRAME, + }, + { + .cfg = { + .id = V4L2_CID_MPEG_VIDEO_VP9_PROFILE, + .min = V4L2_MPEG_VIDEO_VP9_PROFILE_0, + .def = V4L2_MPEG_VIDEO_VP9_PROFILE_0, + .max = V4L2_MPEG_VIDEO_VP9_PROFILE_3, + }, + .codec_type = V4L2_PIX_FMT_VP9_FRAME, + }, + { + .cfg = { + .id = V4L2_CID_STATELESS_HEVC_SPS, + }, + .codec_type = V4L2_PIX_FMT_HEVC_SLICE, + }, + { + .cfg = { + .id = V4L2_CID_STATELESS_HEVC_PPS, + }, + .codec_type = V4L2_PIX_FMT_HEVC_SLICE, + }, + { + .cfg = { + .id = V4L2_CID_STATELESS_HEVC_SCALING_MATRIX, + }, + .codec_type = V4L2_PIX_FMT_HEVC_SLICE, + }, + { + .cfg = { + .id = V4L2_CID_STATELESS_HEVC_DECODE_PARAMS, + }, + .codec_type = V4L2_PIX_FMT_HEVC_SLICE, + }, + { + .cfg = { + .id = V4L2_CID_MPEG_VIDEO_HEVC_PROFILE, + .def = V4L2_MPEG_VIDEO_HEVC_PROFILE_MAIN, + .max = V4L2_MPEG_VIDEO_HEVC_PROFILE_MAIN_10, + .menu_skip_mask = + BIT(V4L2_MPEG_VIDEO_HEVC_PROFILE_MAIN_STILL_PICTURE), + }, + .codec_type = V4L2_PIX_FMT_HEVC_SLICE, + }, + { + .cfg = { + .id = V4L2_CID_STATELESS_HEVC_DECODE_MODE, + .min = V4L2_STATELESS_HEVC_DECODE_MODE_FRAME_BASED, + .def = V4L2_STATELESS_HEVC_DECODE_MODE_FRAME_BASED, + .max = V4L2_STATELESS_HEVC_DECODE_MODE_FRAME_BASED, + }, + .codec_type = V4L2_PIX_FMT_HEVC_SLICE, + }, + { + .cfg = { + .id = V4L2_CID_STATELESS_HEVC_START_CODE, + .min = V4L2_STATELESS_HEVC_START_CODE_ANNEX_B, + .def = V4L2_STATELESS_HEVC_START_CODE_ANNEX_B, + .max = V4L2_STATELESS_HEVC_START_CODE_ANNEX_B, + }, + .codec_type = V4L2_PIX_FMT_HEVC_SLICE, + }, + { + .cfg = { + .id = V4L2_CID_STATELESS_AV1_SEQUENCE, + + }, + .codec_type = V4L2_PIX_FMT_AV1_FRAME, + }, + { + .cfg = { + .id = V4L2_CID_STATELESS_AV1_FRAME, + + }, + .codec_type = V4L2_PIX_FMT_AV1_FRAME, + }, + { + .cfg = { + .id = V4L2_CID_STATELESS_AV1_TILE_GROUP_ENTRY, + .dims = { V4L2_AV1_MAX_TILE_COUNT }, + + }, + .codec_type = V4L2_PIX_FMT_AV1_FRAME, + }, + { + .cfg = { + .id = V4L2_CID_MPEG_VIDEO_AV1_PROFILE, + .min = V4L2_MPEG_VIDEO_AV1_PROFILE_MAIN, + .def = V4L2_MPEG_VIDEO_AV1_PROFILE_MAIN, + .max = V4L2_MPEG_VIDEO_AV1_PROFILE_MAIN, + }, + .codec_type = V4L2_PIX_FMT_AV1_FRAME, + }, + { + .cfg = { + .id = V4L2_CID_MPEG_VIDEO_AV1_LEVEL, + .min = V4L2_MPEG_VIDEO_AV1_LEVEL_2_0, + .def = V4L2_MPEG_VIDEO_AV1_LEVEL_4_0, + .max = V4L2_MPEG_VIDEO_AV1_LEVEL_5_1, + }, + .codec_type = V4L2_PIX_FMT_AV1_FRAME, + }, +}; + +#define NUM_CTRLS ARRAY_SIZE(mtk_stateless_controls) + +static struct mtk_video_fmt mtk_video_formats[9]; + +static struct mtk_video_fmt default_out_format; +static struct mtk_video_fmt default_cap_format; +static unsigned int num_formats; + +static const struct v4l2_frmsize_stepwise stepwise_fhd = { + .min_width = MTK_VDEC_MIN_W, + .max_width = MTK_VDEC_MAX_W, + .step_width = 16, + .min_height = MTK_VDEC_MIN_H, + .max_height = MTK_VDEC_MAX_H, + .step_height = 16 +}; + +static void mtk_vdec_stateless_cap_to_disp(struct mtk_vcodec_dec_ctx *ctx, int error, + struct media_request *src_buf_req) +{ + struct vb2_v4l2_buffer *vb2_dst; + enum vb2_buffer_state state; + + if (error) + state = VB2_BUF_STATE_ERROR; + else + state = VB2_BUF_STATE_DONE; + + vb2_dst = v4l2_m2m_dst_buf_remove(ctx->m2m_ctx); + if (vb2_dst) { + v4l2_m2m_buf_done(vb2_dst, state); + mtk_v4l2_vdec_dbg(2, ctx, "free frame buffer id:%d to done list", + vb2_dst->vb2_buf.index); + } else { + mtk_v4l2_vdec_err(ctx, "dst buffer is NULL"); + } + + if (src_buf_req) + v4l2_ctrl_request_complete(src_buf_req, &ctx->ctrl_hdl); +} + +static struct vdec_fb *vdec_get_cap_buffer(struct mtk_vcodec_dec_ctx *ctx) +{ + struct mtk_video_dec_buf *framebuf; + struct vb2_v4l2_buffer *vb2_v4l2; + struct vb2_buffer *dst_buf; + struct vdec_fb *pfb; + + vb2_v4l2 = v4l2_m2m_next_dst_buf(ctx->m2m_ctx); + if (!vb2_v4l2) { + mtk_v4l2_vdec_dbg(1, ctx, "[%d] dst_buf empty!!", ctx->id); + return NULL; + } + + dst_buf = &vb2_v4l2->vb2_buf; + framebuf = container_of(vb2_v4l2, struct mtk_video_dec_buf, m2m_buf.vb); + + pfb = &framebuf->frame_buffer; + pfb->base_y.va = vb2_plane_vaddr(dst_buf, 0); + pfb->base_y.dma_addr = vb2_dma_contig_plane_dma_addr(dst_buf, 0); + pfb->base_y.size = ctx->q_data[MTK_Q_DATA_DST].sizeimage[0]; + + if (ctx->q_data[MTK_Q_DATA_DST].fmt->num_planes == 2) { + pfb->base_c.va = vb2_plane_vaddr(dst_buf, 1); + pfb->base_c.dma_addr = + vb2_dma_contig_plane_dma_addr(dst_buf, 1); + pfb->base_c.size = ctx->q_data[MTK_Q_DATA_DST].sizeimage[1]; + } + mtk_v4l2_vdec_dbg(1, ctx, + "id=%d Framebuf pfb=%p VA=%p Y/C_DMA=%pad_%pad Sz=%zx frame_count = %d", + dst_buf->index, pfb, pfb->base_y.va, &pfb->base_y.dma_addr, + &pfb->base_c.dma_addr, pfb->base_y.size, ctx->decoded_frame_cnt); + + return pfb; +} + +static void vb2ops_vdec_buf_request_complete(struct vb2_buffer *vb) +{ + struct mtk_vcodec_dec_ctx *ctx = vb2_get_drv_priv(vb->vb2_queue); + + v4l2_ctrl_request_complete(vb->req_obj.req, &ctx->ctrl_hdl); +} + +static void mtk_vdec_worker(struct work_struct *work) +{ + struct mtk_vcodec_dec_ctx *ctx = + container_of(work, struct mtk_vcodec_dec_ctx, decode_work); + struct mtk_vcodec_dec_dev *dev = ctx->dev; + struct vb2_v4l2_buffer *vb2_v4l2_src; + struct vb2_buffer *vb2_src; + struct mtk_vcodec_mem *bs_src; + struct mtk_video_dec_buf *dec_buf_src; + struct media_request *src_buf_req; + enum vb2_buffer_state state; + bool res_chg = false; + int ret; + + vb2_v4l2_src = v4l2_m2m_next_src_buf(ctx->m2m_ctx); + if (!vb2_v4l2_src) { + v4l2_m2m_job_finish(dev->m2m_dev_dec, ctx->m2m_ctx); + mtk_v4l2_vdec_dbg(1, ctx, "[%d] no available source buffer", ctx->id); + return; + } + + vb2_src = &vb2_v4l2_src->vb2_buf; + dec_buf_src = container_of(vb2_v4l2_src, struct mtk_video_dec_buf, + m2m_buf.vb); + bs_src = &dec_buf_src->bs_buffer; + + mtk_v4l2_vdec_dbg(3, ctx, "[%d] (%d) id=%d, vb=%p", ctx->id, + vb2_src->vb2_queue->type, vb2_src->index, vb2_src); + + bs_src->va = vb2_plane_vaddr(vb2_src, 0); + bs_src->dma_addr = vb2_dma_contig_plane_dma_addr(vb2_src, 0); + bs_src->size = (size_t)vb2_src->planes[0].bytesused; + if (!bs_src->va) { + v4l2_m2m_job_finish(dev->m2m_dev_dec, ctx->m2m_ctx); + mtk_v4l2_vdec_err(ctx, "[%d] id=%d source buffer is NULL", ctx->id, + vb2_src->index); + return; + } + + mtk_v4l2_vdec_dbg(3, ctx, "[%d] Bitstream VA=%p DMA=%pad Size=%zx vb=%p", + ctx->id, bs_src->va, &bs_src->dma_addr, bs_src->size, vb2_src); + /* Apply request controls. */ + src_buf_req = vb2_src->req_obj.req; + if (src_buf_req) + v4l2_ctrl_request_setup(src_buf_req, &ctx->ctrl_hdl); + else + mtk_v4l2_vdec_err(ctx, "vb2 buffer media request is NULL"); + + ret = vdec_if_decode(ctx, bs_src, NULL, &res_chg); + if (ret && ret != -EAGAIN) { + mtk_v4l2_vdec_err(ctx, + "[%d] decode src_buf[%d] sz=0x%zx pts=%llu ret=%d res_chg=%d", + ctx->id, vb2_src->index, bs_src->size, + vb2_src->timestamp, ret, res_chg); + if (ret == -EIO) { + mutex_lock(&ctx->lock); + dec_buf_src->error = true; + mutex_unlock(&ctx->lock); + } + } + + state = ret ? VB2_BUF_STATE_ERROR : VB2_BUF_STATE_DONE; + if (!IS_VDEC_LAT_ARCH(dev->vdec_pdata->hw_arch) || + ctx->current_codec == V4L2_PIX_FMT_VP8_FRAME) { + v4l2_m2m_buf_done_and_job_finish(dev->m2m_dev_dec, ctx->m2m_ctx, state); + if (src_buf_req) + v4l2_ctrl_request_complete(src_buf_req, &ctx->ctrl_hdl); + } else { + if (ret != -EAGAIN) { + v4l2_m2m_src_buf_remove(ctx->m2m_ctx); + v4l2_m2m_buf_done(vb2_v4l2_src, state); + } + v4l2_m2m_job_finish(dev->m2m_dev_dec, ctx->m2m_ctx); + } +} + +static void vb2ops_vdec_stateless_buf_queue(struct vb2_buffer *vb) +{ + struct mtk_vcodec_dec_ctx *ctx = vb2_get_drv_priv(vb->vb2_queue); + struct vb2_v4l2_buffer *vb2_v4l2 = to_vb2_v4l2_buffer(vb); + + mtk_v4l2_vdec_dbg(3, ctx, "[%d] (%d) id=%d, vb=%p", ctx->id, vb->vb2_queue->type, + vb->index, vb); + + mutex_lock(&ctx->lock); + v4l2_m2m_buf_queue(ctx->m2m_ctx, vb2_v4l2); + mutex_unlock(&ctx->lock); + if (vb->vb2_queue->type != V4L2_BUF_TYPE_VIDEO_OUTPUT_MPLANE) + return; + + /* If an OUTPUT buffer, we may need to update the state */ + if (ctx->state == MTK_STATE_INIT) { + ctx->state = MTK_STATE_HEADER; + mtk_v4l2_vdec_dbg(1, ctx, "Init driver from init to header."); + } else { + mtk_v4l2_vdec_dbg(3, ctx, "[%d] already init driver %d", ctx->id, ctx->state); + } +} + +static int mtk_vdec_flush_decoder(struct mtk_vcodec_dec_ctx *ctx) +{ + bool res_chg; + + return vdec_if_decode(ctx, NULL, NULL, &res_chg); +} + +static int mtk_vcodec_get_pic_info(struct mtk_vcodec_dec_ctx *ctx) +{ + struct mtk_q_data *q_data; + int ret = 0; + + q_data = &ctx->q_data[MTK_Q_DATA_DST]; + if (q_data->fmt->num_planes == 1) { + mtk_v4l2_vdec_err(ctx, "[%d]Error!! 10bit mode not support one plane", ctx->id); + return -EINVAL; + } + + ctx->capture_fourcc = q_data->fmt->fourcc; + ret = vdec_if_get_param(ctx, GET_PARAM_PIC_INFO, &ctx->picinfo); + if (ret) { + mtk_v4l2_vdec_err(ctx, "[%d]Error!! Get GET_PARAM_PICTURE_INFO Fail", ctx->id); + return ret; + } + + ctx->last_decoded_picinfo = ctx->picinfo; + + q_data->sizeimage[0] = ctx->picinfo.fb_sz[0]; + q_data->bytesperline[0] = ctx->picinfo.buf_w * 5 / 4; + + q_data->sizeimage[1] = ctx->picinfo.fb_sz[1]; + q_data->bytesperline[1] = ctx->picinfo.buf_w * 5 / 4; + + q_data->coded_width = ctx->picinfo.buf_w; + q_data->coded_height = ctx->picinfo.buf_h; + mtk_v4l2_vdec_dbg(1, ctx, "[%d] wxh=%dx%d pic wxh=%dx%d sz[0]=0x%x sz[1]=0x%x", + ctx->id, ctx->picinfo.buf_w, ctx->picinfo.buf_h, + ctx->picinfo.pic_w, ctx->picinfo.pic_h, + q_data->sizeimage[0], q_data->sizeimage[1]); + + return ret; +} + +static int mtk_vdec_s_ctrl(struct v4l2_ctrl *ctrl) +{ + struct mtk_vcodec_dec_ctx *ctx = ctrl_to_dec_ctx(ctrl); + struct v4l2_ctrl_h264_sps *h264; + struct v4l2_ctrl_hevc_sps *h265; + struct v4l2_ctrl_vp9_frame *frame; + struct v4l2_ctrl_av1_sequence *seq; + struct v4l2_ctrl *hdr_ctrl; + const struct mtk_vcodec_dec_pdata *dec_pdata = ctx->dev->vdec_pdata; + const struct mtk_video_fmt *fmt; + int i = 0, ret = 0; + + hdr_ctrl = ctrl; + if (!hdr_ctrl || !hdr_ctrl->p_new.p) + return -EINVAL; + + switch (hdr_ctrl->id) { + case V4L2_CID_STATELESS_H264_SPS: + h264 = (struct v4l2_ctrl_h264_sps *)hdr_ctrl->p_new.p; + + if (h264->bit_depth_chroma_minus8 == 2 && h264->bit_depth_luma_minus8 == 2) { + ctx->is_10bit_bitstream = true; + } else if (h264->bit_depth_chroma_minus8 != 0 && + h264->bit_depth_luma_minus8 != 0) { + mtk_v4l2_vdec_err(ctx, "H264: chroma_minus8:%d, luma_minus8:%d", + h264->bit_depth_chroma_minus8, + h264->bit_depth_luma_minus8); + return -EINVAL; + } + break; + case V4L2_CID_STATELESS_HEVC_SPS: + h265 = (struct v4l2_ctrl_hevc_sps *)hdr_ctrl->p_new.p; + + if (h265->bit_depth_chroma_minus8 == 2 && h265->bit_depth_luma_minus8 == 2) { + ctx->is_10bit_bitstream = true; + } else if (h265->bit_depth_chroma_minus8 != 0 && + h265->bit_depth_luma_minus8 != 0) { + mtk_v4l2_vdec_err(ctx, "HEVC: chroma_minus8:%d, luma_minus8:%d", + h265->bit_depth_chroma_minus8, + h265->bit_depth_luma_minus8); + return -EINVAL; + } + break; + case V4L2_CID_STATELESS_VP9_FRAME: + frame = (struct v4l2_ctrl_vp9_frame *)hdr_ctrl->p_new.p; + + if (frame->bit_depth == 10) { + ctx->is_10bit_bitstream = true; + } else if (frame->bit_depth != 8) { + mtk_v4l2_vdec_err(ctx, "VP9: bit_depth:%d", frame->bit_depth); + return -EINVAL; + } + break; + case V4L2_CID_STATELESS_AV1_SEQUENCE: + seq = (struct v4l2_ctrl_av1_sequence *)hdr_ctrl->p_new.p; + + if (seq->bit_depth == 10) { + ctx->is_10bit_bitstream = true; + } else if (seq->bit_depth != 8) { + mtk_v4l2_vdec_err(ctx, "AV1: bit_depth:%d", seq->bit_depth); + return -EINVAL; + } + break; + default: + mtk_v4l2_vdec_dbg(3, ctx, "Not supported to set ctrl id: 0x%x\n", hdr_ctrl->id); + return ret; + } + + if (!ctx->is_10bit_bitstream) + return ret; + + for (i = 0; i < *dec_pdata->num_formats; i++) { + fmt = &dec_pdata->vdec_formats[i]; + if (fmt->fourcc == V4L2_PIX_FMT_MT2110R && + hdr_ctrl->id == V4L2_CID_STATELESS_H264_SPS) { + ctx->q_data[MTK_Q_DATA_DST].fmt = fmt; + break; + } + + if (fmt->fourcc == V4L2_PIX_FMT_MT2110T && + (hdr_ctrl->id == V4L2_CID_STATELESS_HEVC_SPS || + hdr_ctrl->id == V4L2_CID_STATELESS_VP9_FRAME || + hdr_ctrl->id == V4L2_CID_STATELESS_AV1_SEQUENCE)) { + ctx->q_data[MTK_Q_DATA_DST].fmt = fmt; + break; + } + } + ret = mtk_vcodec_get_pic_info(ctx); + + return ret; +} + +static const struct v4l2_ctrl_ops mtk_vcodec_dec_ctrl_ops = { + .s_ctrl = mtk_vdec_s_ctrl, +}; + +static int mtk_vcodec_dec_ctrls_setup(struct mtk_vcodec_dec_ctx *ctx) +{ + unsigned int i; + + v4l2_ctrl_handler_init(&ctx->ctrl_hdl, NUM_CTRLS); + if (ctx->ctrl_hdl.error) { + mtk_v4l2_vdec_err(ctx, "v4l2_ctrl_handler_init failed\n"); + return ctx->ctrl_hdl.error; + } + + for (i = 0; i < NUM_CTRLS; i++) { + struct v4l2_ctrl_config cfg = mtk_stateless_controls[i].cfg; + cfg.ops = &mtk_vcodec_dec_ctrl_ops; + v4l2_ctrl_new_custom(&ctx->ctrl_hdl, &cfg, NULL); + if (ctx->ctrl_hdl.error) { + mtk_v4l2_vdec_err(ctx, "Adding control %d failed %d", i, + ctx->ctrl_hdl.error); + return ctx->ctrl_hdl.error; + } + } + + v4l2_ctrl_handler_setup(&ctx->ctrl_hdl); + + return 0; +} + +static int fops_media_request_validate(struct media_request *mreq) +{ + const unsigned int buffer_cnt = vb2_request_buffer_cnt(mreq); + + switch (buffer_cnt) { + case 1: + /* We expect exactly one buffer with the request */ + break; + case 0: + pr_debug(MTK_DBG_VCODEC_STR "No buffer provided with the request."); + return -ENOENT; + default: + pr_debug(MTK_DBG_VCODEC_STR "Too many buffers (%d) provided with the request.", + buffer_cnt); + return -EINVAL; + } + + return vb2_request_validate(mreq); +} + +const struct media_device_ops mtk_vcodec_media_ops = { + .req_validate = fops_media_request_validate, + .req_queue = v4l2_m2m_request_queue, +}; + +static void mtk_vcodec_add_formats(unsigned int fourcc, + struct mtk_vcodec_dec_ctx *ctx) +{ + struct mtk_vcodec_dec_dev *dev = ctx->dev; + const struct mtk_vcodec_dec_pdata *pdata = dev->vdec_pdata; + int count_formats = *pdata->num_formats; + + switch (fourcc) { + case V4L2_PIX_FMT_H264_SLICE: + case V4L2_PIX_FMT_VP8_FRAME: + case V4L2_PIX_FMT_VP9_FRAME: + case V4L2_PIX_FMT_HEVC_SLICE: + case V4L2_PIX_FMT_AV1_FRAME: + mtk_video_formats[count_formats].fourcc = fourcc; + mtk_video_formats[count_formats].type = MTK_FMT_DEC; + mtk_video_formats[count_formats].num_planes = 1; + mtk_video_formats[count_formats].frmsize = stepwise_fhd; + + if (!(ctx->dev->dec_capability & VCODEC_CAPABILITY_4K_DISABLED) && + fourcc != V4L2_PIX_FMT_VP8_FRAME) { + mtk_video_formats[count_formats].frmsize.max_width = + VCODEC_DEC_4K_CODED_WIDTH; + mtk_video_formats[count_formats].frmsize.max_height = + VCODEC_DEC_4K_CODED_HEIGHT; + } + break; + case V4L2_PIX_FMT_MM21: + case V4L2_PIX_FMT_MT21C: + case V4L2_PIX_FMT_MT2110T: + case V4L2_PIX_FMT_MT2110R: + mtk_video_formats[count_formats].fourcc = fourcc; + mtk_video_formats[count_formats].type = MTK_FMT_FRAME; + mtk_video_formats[count_formats].num_planes = 2; + break; + default: + mtk_v4l2_vdec_err(ctx, "Can not add unsupported format type"); + return; + } + + num_formats++; + mtk_v4l2_vdec_dbg(3, ctx, "num_formats: %d dec_capability: 0x%x", + count_formats, ctx->dev->dec_capability); +} + +static void mtk_vcodec_get_supported_formats(struct mtk_vcodec_dec_ctx *ctx) +{ + int cap_format_count = 0, out_format_count = 0; + + if (num_formats) + return; + + if (ctx->dev->dec_capability & MTK_VDEC_FORMAT_MT21C) { + mtk_vcodec_add_formats(V4L2_PIX_FMT_MT21C, ctx); + cap_format_count++; + } + if (ctx->dev->dec_capability & MTK_VDEC_IS_SUPPORT_10BIT) { + mtk_vcodec_add_formats(V4L2_PIX_FMT_MT2110T, ctx); + cap_format_count++; + mtk_vcodec_add_formats(V4L2_PIX_FMT_MT2110R, ctx); + cap_format_count++; + } + if (ctx->dev->dec_capability & MTK_VDEC_FORMAT_MM21) { + mtk_vcodec_add_formats(V4L2_PIX_FMT_MM21, ctx); + cap_format_count++; + } + if (ctx->dev->dec_capability & MTK_VDEC_FORMAT_H264_SLICE) { + mtk_vcodec_add_formats(V4L2_PIX_FMT_H264_SLICE, ctx); + out_format_count++; + } + if (ctx->dev->dec_capability & MTK_VDEC_FORMAT_VP8_FRAME) { + mtk_vcodec_add_formats(V4L2_PIX_FMT_VP8_FRAME, ctx); + out_format_count++; + } + if (ctx->dev->dec_capability & MTK_VDEC_FORMAT_VP9_FRAME) { + mtk_vcodec_add_formats(V4L2_PIX_FMT_VP9_FRAME, ctx); + out_format_count++; + } + if (ctx->dev->dec_capability & MTK_VDEC_FORMAT_HEVC_FRAME) { + mtk_vcodec_add_formats(V4L2_PIX_FMT_HEVC_SLICE, ctx); + out_format_count++; + } + if (ctx->dev->dec_capability & MTK_VDEC_FORMAT_AV1_FRAME) { + mtk_vcodec_add_formats(V4L2_PIX_FMT_AV1_FRAME, ctx); + out_format_count++; + } + + if (cap_format_count) + default_cap_format = mtk_video_formats[cap_format_count - 1]; + if (out_format_count) + default_out_format = + mtk_video_formats[cap_format_count + out_format_count - 1]; +} + +static void mtk_init_vdec_params(struct mtk_vcodec_dec_ctx *ctx) +{ + struct vb2_queue *src_vq; + + src_vq = v4l2_m2m_get_vq(ctx->m2m_ctx, + V4L2_BUF_TYPE_VIDEO_OUTPUT_MPLANE); + + if (!ctx->dev->vdec_pdata->is_subdev_supported) + ctx->dev->dec_capability |= + MTK_VDEC_FORMAT_H264_SLICE | MTK_VDEC_FORMAT_MM21; + mtk_vcodec_get_supported_formats(ctx); + + /* Support request api for output plane */ + src_vq->supports_requests = true; + src_vq->requires_requests = true; +} + +static int vb2ops_vdec_out_buf_validate(struct vb2_buffer *vb) +{ + struct vb2_v4l2_buffer *vbuf = to_vb2_v4l2_buffer(vb); + + vbuf->field = V4L2_FIELD_NONE; + return 0; +} + +static struct vb2_ops mtk_vdec_request_vb2_ops = { + .queue_setup = vb2ops_vdec_queue_setup, + .wait_prepare = vb2_ops_wait_prepare, + .wait_finish = vb2_ops_wait_finish, + .start_streaming = vb2ops_vdec_start_streaming, + .stop_streaming = vb2ops_vdec_stop_streaming, + + .buf_queue = vb2ops_vdec_stateless_buf_queue, + .buf_out_validate = vb2ops_vdec_out_buf_validate, + .buf_init = vb2ops_vdec_buf_init, + .buf_prepare = vb2ops_vdec_buf_prepare, + .buf_finish = vb2ops_vdec_buf_finish, + .buf_request_complete = vb2ops_vdec_buf_request_complete, +}; + +const struct mtk_vcodec_dec_pdata mtk_vdec_8183_pdata = { + .init_vdec_params = mtk_init_vdec_params, + .ctrls_setup = mtk_vcodec_dec_ctrls_setup, + .vdec_vb2_ops = &mtk_vdec_request_vb2_ops, + .vdec_formats = mtk_video_formats, + .num_formats = &num_formats, + .default_out_fmt = &default_out_format, + .default_cap_fmt = &default_cap_format, + .uses_stateless_api = true, + .worker = mtk_vdec_worker, + .flush_decoder = mtk_vdec_flush_decoder, + .cap_to_disp = mtk_vdec_stateless_cap_to_disp, + .get_cap_buffer = vdec_get_cap_buffer, + .is_subdev_supported = false, + .hw_arch = MTK_VDEC_PURE_SINGLE_CORE, +}; + +/* This platform data is used for one lat and one core architecture. */ +const struct mtk_vcodec_dec_pdata mtk_lat_sig_core_pdata = { + .init_vdec_params = mtk_init_vdec_params, + .ctrls_setup = mtk_vcodec_dec_ctrls_setup, + .vdec_vb2_ops = &mtk_vdec_request_vb2_ops, + .vdec_formats = mtk_video_formats, + .num_formats = &num_formats, + .default_out_fmt = &default_out_format, + .default_cap_fmt = &default_cap_format, + .uses_stateless_api = true, + .worker = mtk_vdec_worker, + .flush_decoder = mtk_vdec_flush_decoder, + .cap_to_disp = mtk_vdec_stateless_cap_to_disp, + .get_cap_buffer = vdec_get_cap_buffer, + .is_subdev_supported = true, + .hw_arch = MTK_VDEC_LAT_SINGLE_CORE, +}; + +const struct mtk_vcodec_dec_pdata mtk_vdec_single_core_pdata = { + .init_vdec_params = mtk_init_vdec_params, + .ctrls_setup = mtk_vcodec_dec_ctrls_setup, + .vdec_vb2_ops = &mtk_vdec_request_vb2_ops, + .vdec_formats = mtk_video_formats, + .num_formats = &num_formats, + .default_out_fmt = &default_out_format, + .default_cap_fmt = &default_cap_format, + .uses_stateless_api = true, + .worker = mtk_vdec_worker, + .flush_decoder = mtk_vdec_flush_decoder, + .cap_to_disp = mtk_vdec_stateless_cap_to_disp, + .get_cap_buffer = vdec_get_cap_buffer, + .is_subdev_supported = true, + .hw_arch = MTK_VDEC_PURE_SINGLE_CORE, +}; diff --git a/drivers/media/platform/mediatek/vcodec/decoder/vdec/vdec_av1_req_lat_if.c b/drivers/media/platform/mediatek/vcodec/decoder/vdec/vdec_av1_req_lat_if.c new file mode 100644 index 0000000000..2b6a5adbc4 --- /dev/null +++ b/drivers/media/platform/mediatek/vcodec/decoder/vdec/vdec_av1_req_lat_if.c @@ -0,0 +1,2209 @@ +// SPDX-License-Identifier: GPL-2.0 +/* + * Copyright (c) 2023 MediaTek Inc. + * Author: Xiaoyong Lu <xiaoyong.lu@mediatek.com> + */ + +#include <linux/module.h> +#include <linux/slab.h> +#include <media/videobuf2-dma-contig.h> + +#include "../mtk_vcodec_dec.h" +#include "../../common/mtk_vcodec_intr.h" +#include "../vdec_drv_base.h" +#include "../vdec_drv_if.h" +#include "../vdec_vpu_if.h" + +#define AV1_MAX_FRAME_BUF_COUNT (V4L2_AV1_TOTAL_REFS_PER_FRAME + 1) +#define AV1_TILE_BUF_SIZE 64 +#define AV1_SCALE_SUBPEL_BITS 10 +#define AV1_REF_SCALE_SHIFT 14 +#define AV1_REF_NO_SCALE BIT(AV1_REF_SCALE_SHIFT) +#define AV1_REF_INVALID_SCALE -1 +#define AV1_CDF_TABLE_BUFFER_SIZE 16384 +#define AV1_PRIMARY_REF_NONE 7 + +#define AV1_INVALID_IDX -1 + +#define AV1_DIV_ROUND_UP_POW2(value, n) \ +({ \ + typeof(n) _n = n; \ + typeof(value) _value = value; \ + (_value + (BIT(_n) >> 1)) >> _n; \ +}) + +#define AV1_DIV_ROUND_UP_POW2_SIGNED(value, n) \ +({ \ + typeof(n) _n_ = n; \ + typeof(value) _value_ = value; \ + (((_value_) < 0) ? -AV1_DIV_ROUND_UP_POW2(-(_value_), (_n_)) \ + : AV1_DIV_ROUND_UP_POW2((_value_), (_n_))); \ +}) + +#define BIT_FLAG(x, bit) (!!((x)->flags & (bit))) +#define SEGMENTATION_FLAG(x, name) (!!((x)->flags & V4L2_AV1_SEGMENTATION_FLAG_##name)) +#define QUANT_FLAG(x, name) (!!((x)->flags & V4L2_AV1_QUANTIZATION_FLAG_##name)) +#define SEQUENCE_FLAG(x, name) (!!((x)->flags & V4L2_AV1_SEQUENCE_FLAG_##name)) +#define FH_FLAG(x, name) (!!((x)->flags & V4L2_AV1_FRAME_FLAG_##name)) + +#define MINQ 0 +#define MAXQ 255 + +#define DIV_LUT_PREC_BITS 14 +#define DIV_LUT_BITS 8 +#define DIV_LUT_NUM BIT(DIV_LUT_BITS) +#define WARP_PARAM_REDUCE_BITS 6 +#define WARPEDMODEL_PREC_BITS 16 + +#define SEG_LVL_ALT_Q 0 +#define SECONDARY_FILTER_STRENGTH_NUM_BITS 2 + +static const short div_lut[DIV_LUT_NUM + 1] = { + 16384, 16320, 16257, 16194, 16132, 16070, 16009, 15948, 15888, 15828, 15768, + 15709, 15650, 15592, 15534, 15477, 15420, 15364, 15308, 15252, 15197, 15142, + 15087, 15033, 14980, 14926, 14873, 14821, 14769, 14717, 14665, 14614, 14564, + 14513, 14463, 14413, 14364, 14315, 14266, 14218, 14170, 14122, 14075, 14028, + 13981, 13935, 13888, 13843, 13797, 13752, 13707, 13662, 13618, 13574, 13530, + 13487, 13443, 13400, 13358, 13315, 13273, 13231, 13190, 13148, 13107, 13066, + 13026, 12985, 12945, 12906, 12866, 12827, 12788, 12749, 12710, 12672, 12633, + 12596, 12558, 12520, 12483, 12446, 12409, 12373, 12336, 12300, 12264, 12228, + 12193, 12157, 12122, 12087, 12053, 12018, 11984, 11950, 11916, 11882, 11848, + 11815, 11782, 11749, 11716, 11683, 11651, 11619, 11586, 11555, 11523, 11491, + 11460, 11429, 11398, 11367, 11336, 11305, 11275, 11245, 11215, 11185, 11155, + 11125, 11096, 11067, 11038, 11009, 10980, 10951, 10923, 10894, 10866, 10838, + 10810, 10782, 10755, 10727, 10700, 10673, 10645, 10618, 10592, 10565, 10538, + 10512, 10486, 10460, 10434, 10408, 10382, 10356, 10331, 10305, 10280, 10255, + 10230, 10205, 10180, 10156, 10131, 10107, 10082, 10058, 10034, 10010, 9986, + 9963, 9939, 9916, 9892, 9869, 9846, 9823, 9800, 9777, 9754, 9732, + 9709, 9687, 9664, 9642, 9620, 9598, 9576, 9554, 9533, 9511, 9489, + 9468, 9447, 9425, 9404, 9383, 9362, 9341, 9321, 9300, 9279, 9259, + 9239, 9218, 9198, 9178, 9158, 9138, 9118, 9098, 9079, 9059, 9039, + 9020, 9001, 8981, 8962, 8943, 8924, 8905, 8886, 8867, 8849, 8830, + 8812, 8793, 8775, 8756, 8738, 8720, 8702, 8684, 8666, 8648, 8630, + 8613, 8595, 8577, 8560, 8542, 8525, 8508, 8490, 8473, 8456, 8439, + 8422, 8405, 8389, 8372, 8355, 8339, 8322, 8306, 8289, 8273, 8257, + 8240, 8224, 8208, 8192, +}; + +/** + * struct vdec_av1_slice_init_vsi - VSI used to initialize instance + * @architecture: architecture type + * @reserved: reserved + * @core_vsi: for core vsi + * @cdf_table_addr: cdf table addr + * @cdf_table_size: cdf table size + * @iq_table_addr: iq table addr + * @iq_table_size: iq table size + * @vsi_size: share vsi structure size + */ +struct vdec_av1_slice_init_vsi { + u32 architecture; + u32 reserved; + u64 core_vsi; + u64 cdf_table_addr; + u32 cdf_table_size; + u64 iq_table_addr; + u32 iq_table_size; + u32 vsi_size; +}; + +/** + * struct vdec_av1_slice_mem - memory address and size + * @buf: dma_addr padding + * @dma_addr: buffer address + * @size: buffer size + * @dma_addr_end: buffer end address + * @padding: for padding + */ +struct vdec_av1_slice_mem { + union { + u64 buf; + dma_addr_t dma_addr; + }; + union { + size_t size; + dma_addr_t dma_addr_end; + u64 padding; + }; +}; + +/** + * struct vdec_av1_slice_state - decoding state + * @err : err type for decode + * @full : transcoded buffer is full or not + * @timeout : decode timeout or not + * @perf : performance enable + * @crc : hw checksum + * @out_size : hw output size + */ +struct vdec_av1_slice_state { + int err; + u32 full; + u32 timeout; + u32 perf; + u32 crc[16]; + u32 out_size; +}; + +/* + * enum vdec_av1_slice_resolution_level - resolution level + */ +enum vdec_av1_slice_resolution_level { + AV1_RES_NONE, + AV1_RES_FHD, + AV1_RES_4K, + AV1_RES_8K, +}; + +/* + * enum vdec_av1_slice_frame_type - av1 frame type + */ +enum vdec_av1_slice_frame_type { + AV1_KEY_FRAME = 0, + AV1_INTER_FRAME, + AV1_INTRA_ONLY_FRAME, + AV1_SWITCH_FRAME, + AV1_FRAME_TYPES, +}; + +/* + * enum vdec_av1_slice_reference_mode - reference mode type + */ +enum vdec_av1_slice_reference_mode { + AV1_SINGLE_REFERENCE = 0, + AV1_COMPOUND_REFERENCE, + AV1_REFERENCE_MODE_SELECT, + AV1_REFERENCE_MODES, +}; + +/** + * struct vdec_av1_slice_tile_group - info for each tile + * @num_tiles: tile number + * @tile_size: input size for each tile + * @tile_start_offset: tile offset to input buffer + */ +struct vdec_av1_slice_tile_group { + u32 num_tiles; + u32 tile_size[V4L2_AV1_MAX_TILE_COUNT]; + u32 tile_start_offset[V4L2_AV1_MAX_TILE_COUNT]; +}; + +/** + * struct vdec_av1_slice_scale_factors - scale info for each ref frame + * @is_scaled: frame is scaled or not + * @x_scale: frame width scale coefficient + * @y_scale: frame height scale coefficient + * @x_step: width step for x_scale + * @y_step: height step for y_scale + */ +struct vdec_av1_slice_scale_factors { + u8 is_scaled; + int x_scale; + int y_scale; + int x_step; + int y_step; +}; + +/** + * struct vdec_av1_slice_frame_refs - ref frame info + * @ref_fb_idx: ref slot index + * @ref_map_idx: ref frame index + * @scale_factors: scale factors for each ref frame + */ +struct vdec_av1_slice_frame_refs { + int ref_fb_idx; + int ref_map_idx; + struct vdec_av1_slice_scale_factors scale_factors; +}; + +/** + * struct vdec_av1_slice_gm - AV1 Global Motion parameters + * @wmtype: The type of global motion transform used + * @wmmat: gm_params + * @alpha: alpha info + * @beta: beta info + * @gamma: gamma info + * @delta: delta info + * @invalid: is invalid or not + */ +struct vdec_av1_slice_gm { + int wmtype; + int wmmat[8]; + short alpha; + short beta; + short gamma; + short delta; + char invalid; +}; + +/** + * struct vdec_av1_slice_sm - AV1 Skip Mode parameters + * @skip_mode_allowed: Skip Mode is allowed or not + * @skip_mode_present: specified that the skip_mode will be present or not + * @skip_mode_frame: specifies the frames to use for compound prediction + */ +struct vdec_av1_slice_sm { + u8 skip_mode_allowed; + u8 skip_mode_present; + int skip_mode_frame[2]; +}; + +/** + * struct vdec_av1_slice_seg - AV1 Segmentation params + * @segmentation_enabled: this frame makes use of the segmentation tool or not + * @segmentation_update_map: segmentation map are updated during the decoding frame + * @segmentation_temporal_update:segmentation map are coded relative the existing segmentaion map + * @segmentation_update_data: new parameters are about to be specified for each segment + * @feature_data: specifies the feature data for a segment feature + * @feature_enabled_mask: the corresponding feature value is coded or not. + * @segid_preskip: segment id will be read before the skip syntax element. + * @last_active_segid: the highest numbered segment id that has some enabled feature + */ +struct vdec_av1_slice_seg { + u8 segmentation_enabled; + u8 segmentation_update_map; + u8 segmentation_temporal_update; + u8 segmentation_update_data; + int feature_data[V4L2_AV1_MAX_SEGMENTS][V4L2_AV1_SEG_LVL_MAX]; + u16 feature_enabled_mask[V4L2_AV1_MAX_SEGMENTS]; + int segid_preskip; + int last_active_segid; +}; + +/** + * struct vdec_av1_slice_delta_q_lf - AV1 Loop Filter delta parameters + * @delta_q_present: specified whether quantizer index delta values are present + * @delta_q_res: specifies the left shift which should be applied to decoded quantizer index + * @delta_lf_present: specifies whether loop filter delta values are present + * @delta_lf_res: specifies the left shift which should be applied to decoded + * loop filter delta values + * @delta_lf_multi: specifies that separate loop filter deltas are sent for horizontal + * luma edges,vertical luma edges,the u edges, and the v edges. + */ +struct vdec_av1_slice_delta_q_lf { + u8 delta_q_present; + u8 delta_q_res; + u8 delta_lf_present; + u8 delta_lf_res; + u8 delta_lf_multi; +}; + +/** + * struct vdec_av1_slice_quantization - AV1 Quantization params + * @base_q_idx: indicates the base frame qindex. This is used for Y AC + * coefficients and as the base value for the other quantizers. + * @qindex: qindex + * @delta_qydc: indicates the Y DC quantizer relative to base_q_idx + * @delta_qudc: indicates the U DC quantizer relative to base_q_idx. + * @delta_quac: indicates the U AC quantizer relative to base_q_idx + * @delta_qvdc: indicates the V DC quantizer relative to base_q_idx + * @delta_qvac: indicates the V AC quantizer relative to base_q_idx + * @using_qmatrix: specifies that the quantizer matrix will be used to + * compute quantizers + * @qm_y: specifies the level in the quantizer matrix that should + * be used for luma plane decoding + * @qm_u: specifies the level in the quantizer matrix that should + * be used for chroma U plane decoding. + * @qm_v: specifies the level in the quantizer matrix that should be + * used for chroma V plane decoding + */ +struct vdec_av1_slice_quantization { + int base_q_idx; + int qindex[V4L2_AV1_MAX_SEGMENTS]; + int delta_qydc; + int delta_qudc; + int delta_quac; + int delta_qvdc; + int delta_qvac; + u8 using_qmatrix; + u8 qm_y; + u8 qm_u; + u8 qm_v; +}; + +/** + * struct vdec_av1_slice_lr - AV1 Loop Restauration parameters + * @use_lr: whether to use loop restoration + * @use_chroma_lr: whether to use chroma loop restoration + * @frame_restoration_type: specifies the type of restoration used for each plane + * @loop_restoration_size: pecifies the size of loop restoration units in units + * of samples in the current plane + */ +struct vdec_av1_slice_lr { + u8 use_lr; + u8 use_chroma_lr; + u8 frame_restoration_type[V4L2_AV1_NUM_PLANES_MAX]; + u32 loop_restoration_size[V4L2_AV1_NUM_PLANES_MAX]; +}; + +/** + * struct vdec_av1_slice_loop_filter - AV1 Loop filter parameters + * @loop_filter_level: an array containing loop filter strength values. + * @loop_filter_ref_deltas: contains the adjustment needed for the filter + * level based on the chosen reference frame + * @loop_filter_mode_deltas: contains the adjustment needed for the filter + * level based on the chosen mode + * @loop_filter_sharpness: indicates the sharpness level. The loop_filter_level + * and loop_filter_sharpness together determine when + * a block edge is filtered, and by how much the + * filtering can change the sample values + * @loop_filter_delta_enabled: filetr level depends on the mode and reference + * frame used to predict a block + */ +struct vdec_av1_slice_loop_filter { + u8 loop_filter_level[4]; + int loop_filter_ref_deltas[V4L2_AV1_TOTAL_REFS_PER_FRAME]; + int loop_filter_mode_deltas[2]; + u8 loop_filter_sharpness; + u8 loop_filter_delta_enabled; +}; + +/** + * struct vdec_av1_slice_cdef - AV1 CDEF parameters + * @cdef_damping: controls the amount of damping in the deringing filter + * @cdef_y_strength: specifies the strength of the primary filter and secondary filter + * @cdef_uv_strength: specifies the strength of the primary filter and secondary filter + * @cdef_bits: specifies the number of bits needed to specify which + * CDEF filter to apply + */ +struct vdec_av1_slice_cdef { + u8 cdef_damping; + u8 cdef_y_strength[8]; + u8 cdef_uv_strength[8]; + u8 cdef_bits; +}; + +/** + * struct vdec_av1_slice_mfmv - AV1 mfmv parameters + * @mfmv_valid_ref: mfmv_valid_ref + * @mfmv_dir: mfmv_dir + * @mfmv_ref_to_cur: mfmv_ref_to_cur + * @mfmv_ref_frame_idx: mfmv_ref_frame_idx + * @mfmv_count: mfmv_count + */ +struct vdec_av1_slice_mfmv { + u32 mfmv_valid_ref[3]; + u32 mfmv_dir[3]; + int mfmv_ref_to_cur[3]; + int mfmv_ref_frame_idx[3]; + int mfmv_count; +}; + +/** + * struct vdec_av1_slice_tile - AV1 Tile info + * @tile_cols: specifies the number of tiles across the frame + * @tile_rows: pecifies the number of tiles down the frame + * @mi_col_starts: an array specifying the start column + * @mi_row_starts: an array specifying the start row + * @context_update_tile_id: specifies which tile to use for the CDF update + * @uniform_tile_spacing_flag: tiles are uniformly spaced across the frame + * or the tile sizes are coded + */ +struct vdec_av1_slice_tile { + u8 tile_cols; + u8 tile_rows; + int mi_col_starts[V4L2_AV1_MAX_TILE_COLS + 1]; + int mi_row_starts[V4L2_AV1_MAX_TILE_ROWS + 1]; + u8 context_update_tile_id; + u8 uniform_tile_spacing_flag; +}; + +/** + * struct vdec_av1_slice_uncompressed_header - Represents an AV1 Frame Header OBU + * @use_ref_frame_mvs: use_ref_frame_mvs flag + * @order_hint: specifies OrderHintBits least significant bits of the expected + * @gm: global motion param + * @upscaled_width: the upscaled width + * @frame_width: frame's width + * @frame_height: frame's height + * @reduced_tx_set: frame is restricted to a reduced subset of the full + * set of transform types + * @tx_mode: specifies how the transform size is determined + * @uniform_tile_spacing_flag: tiles are uniformly spaced across the frame + * or the tile sizes are coded + * @interpolation_filter: specifies the filter selection used for performing inter prediction + * @allow_warped_motion: motion_mode may be present or not + * @is_motion_mode_switchable : euqlt to 0 specifies that only the SIMPLE motion mode will be used + * @reference_mode : frame reference mode selected + * @allow_high_precision_mv: specifies that motion vectors are specified to + * quarter pel precision or to eighth pel precision + * @allow_intra_bc: ubducates that intra block copy may be used in this frame + * @force_integer_mv: specifies motion vectors will always be integers or + * can contain fractional bits + * @allow_screen_content_tools: intra blocks may use palette encoding + * @error_resilient_mode: error resislent mode is enable/disable + * @frame_type: specifies the AV1 frame type + * @primary_ref_frame: specifies which reference frame contains the CDF values + * and other state that should be loaded at the start of the frame + * slots will be updated with the current frame after it is decoded + * @disable_frame_end_update_cdf:indicates the end of frame CDF update is disable or enable + * @disable_cdf_update: specified whether the CDF update in the symbol + * decoding process should be disables + * @skip_mode: av1 skip mode parameters + * @seg: av1 segmentaon parameters + * @delta_q_lf: av1 delta loop fileter + * @quant: av1 Quantization params + * @lr: av1 Loop Restauration parameters + * @superres_denom: the denominator for the upscaling ratio + * @loop_filter: av1 Loop filter parameters + * @cdef: av1 CDEF parameters + * @mfmv: av1 mfmv parameters + * @tile: av1 Tile info + * @frame_is_intra: intra frame + * @loss_less_array: loss less array + * @coded_loss_less: coded lsss less + * @mi_rows: size of mi unit in rows + * @mi_cols: size of mi unit in cols + */ +struct vdec_av1_slice_uncompressed_header { + u8 use_ref_frame_mvs; + int order_hint; + struct vdec_av1_slice_gm gm[V4L2_AV1_TOTAL_REFS_PER_FRAME]; + u32 upscaled_width; + u32 frame_width; + u32 frame_height; + u8 reduced_tx_set; + u8 tx_mode; + u8 uniform_tile_spacing_flag; + u8 interpolation_filter; + u8 allow_warped_motion; + u8 is_motion_mode_switchable; + u8 reference_mode; + u8 allow_high_precision_mv; + u8 allow_intra_bc; + u8 force_integer_mv; + u8 allow_screen_content_tools; + u8 error_resilient_mode; + u8 frame_type; + u8 primary_ref_frame; + u8 disable_frame_end_update_cdf; + u32 disable_cdf_update; + struct vdec_av1_slice_sm skip_mode; + struct vdec_av1_slice_seg seg; + struct vdec_av1_slice_delta_q_lf delta_q_lf; + struct vdec_av1_slice_quantization quant; + struct vdec_av1_slice_lr lr; + u32 superres_denom; + struct vdec_av1_slice_loop_filter loop_filter; + struct vdec_av1_slice_cdef cdef; + struct vdec_av1_slice_mfmv mfmv; + struct vdec_av1_slice_tile tile; + u8 frame_is_intra; + u8 loss_less_array[V4L2_AV1_MAX_SEGMENTS]; + u8 coded_loss_less; + u32 mi_rows; + u32 mi_cols; +}; + +/** + * struct vdec_av1_slice_seq_header - Represents an AV1 Sequence OBU + * @bitdepth: the bitdepth to use for the sequence + * @enable_superres: specifies whether the use_superres syntax element may be present + * @enable_filter_intra: specifies the use_filter_intra syntax element may be present + * @enable_intra_edge_filter: whether the intra edge filtering process should be enabled + * @enable_interintra_compound: specifies the mode info fo rinter blocks may + * contain the syntax element interintra + * @enable_masked_compound: specifies the mode info fo rinter blocks may + * contain the syntax element compound_type + * @enable_dual_filter: the inter prediction filter type may be specified independently + * @enable_jnt_comp: distance weights process may be used for inter prediction + * @mono_chrome: indicates the video does not contain U and V color planes + * @enable_order_hint: tools based on the values of order hints may be used + * @order_hint_bits: the number of bits used for the order_hint field at each frame + * @use_128x128_superblock: indicates superblocks contain 128*128 luma samples + * @subsampling_x: the chroma subsamling format + * @subsampling_y: the chroma subsamling format + * @max_frame_width: the maximum frame width for the frames represented by sequence + * @max_frame_height: the maximum frame height for the frames represented by sequence + */ +struct vdec_av1_slice_seq_header { + u8 bitdepth; + u8 enable_superres; + u8 enable_filter_intra; + u8 enable_intra_edge_filter; + u8 enable_interintra_compound; + u8 enable_masked_compound; + u8 enable_dual_filter; + u8 enable_jnt_comp; + u8 mono_chrome; + u8 enable_order_hint; + u8 order_hint_bits; + u8 use_128x128_superblock; + u8 subsampling_x; + u8 subsampling_y; + u32 max_frame_width; + u32 max_frame_height; +}; + +/** + * struct vdec_av1_slice_frame - Represents current Frame info + * @uh: uncompressed header info + * @seq: sequence header info + * @large_scale_tile: is large scale mode + * @cur_ts: current frame timestamp + * @prev_fb_idx: prev slot id + * @ref_frame_sign_bias: arrays for ref_frame sign bias + * @order_hints: arrays for ref_frame order hint + * @ref_frame_valid: arrays for valid ref_frame + * @ref_frame_map: map to slot frame info + * @frame_refs: ref_frame info + */ +struct vdec_av1_slice_frame { + struct vdec_av1_slice_uncompressed_header uh; + struct vdec_av1_slice_seq_header seq; + u8 large_scale_tile; + u64 cur_ts; + int prev_fb_idx; + u8 ref_frame_sign_bias[V4L2_AV1_TOTAL_REFS_PER_FRAME]; + u32 order_hints[V4L2_AV1_REFS_PER_FRAME]; + u32 ref_frame_valid[V4L2_AV1_REFS_PER_FRAME]; + int ref_frame_map[V4L2_AV1_TOTAL_REFS_PER_FRAME]; + struct vdec_av1_slice_frame_refs frame_refs[V4L2_AV1_REFS_PER_FRAME]; +}; + +/** + * struct vdec_av1_slice_work_buffer - work buffer for lat + * @mv_addr: mv buffer memory info + * @cdf_addr: cdf buffer memory info + * @segid_addr: segid buffer memory info + */ +struct vdec_av1_slice_work_buffer { + struct vdec_av1_slice_mem mv_addr; + struct vdec_av1_slice_mem cdf_addr; + struct vdec_av1_slice_mem segid_addr; +}; + +/** + * struct vdec_av1_slice_frame_info - frame info for each slot + * @frame_type: frame type + * @frame_is_intra: is intra frame + * @order_hint: order hint + * @order_hints: referece frame order hint + * @upscaled_width: upscale width + * @pic_pitch: buffer pitch + * @frame_width: frane width + * @frame_height: frame height + * @mi_rows: rows in mode info + * @mi_cols: cols in mode info + * @ref_count: mark to reference frame counts + */ +struct vdec_av1_slice_frame_info { + u8 frame_type; + u8 frame_is_intra; + int order_hint; + u32 order_hints[V4L2_AV1_REFS_PER_FRAME]; + u32 upscaled_width; + u32 pic_pitch; + u32 frame_width; + u32 frame_height; + u32 mi_rows; + u32 mi_cols; + int ref_count; +}; + +/** + * struct vdec_av1_slice_slot - slot info that needs to be saved in the global instance + * @frame_info: frame info for each slot + * @timestamp: time stamp info + */ +struct vdec_av1_slice_slot { + struct vdec_av1_slice_frame_info frame_info[AV1_MAX_FRAME_BUF_COUNT]; + u64 timestamp[AV1_MAX_FRAME_BUF_COUNT]; +}; + +/** + * struct vdec_av1_slice_fb - frame buffer for decoding + * @y: current y buffer address info + * @c: current c buffer address info + */ +struct vdec_av1_slice_fb { + struct vdec_av1_slice_mem y; + struct vdec_av1_slice_mem c; +}; + +/** + * struct vdec_av1_slice_vsi - exchange frame information between Main CPU and MicroP + * @bs: input buffer info + * @work_buffer: working buffe for hw + * @cdf_table: cdf_table buffer + * @cdf_tmp: cdf temp buffer + * @rd_mv: mv buffer for lat output , core input + * @ube: ube buffer + * @trans: transcoded buffer + * @err_map: err map buffer + * @row_info: row info buffer + * @fb: current y/c buffer + * @ref: ref y/c buffer + * @iq_table: iq table buffer + * @tile: tile buffer + * @slots: slots info for each frame + * @slot_id: current frame slot id + * @frame: current frame info + * @state: status after decode done + * @cur_lst_tile_id: tile id for large scale + */ +struct vdec_av1_slice_vsi { + /* lat */ + struct vdec_av1_slice_mem bs; + struct vdec_av1_slice_work_buffer work_buffer[AV1_MAX_FRAME_BUF_COUNT]; + struct vdec_av1_slice_mem cdf_table; + struct vdec_av1_slice_mem cdf_tmp; + /* LAT stage's output, Core stage's input */ + struct vdec_av1_slice_mem rd_mv; + struct vdec_av1_slice_mem ube; + struct vdec_av1_slice_mem trans; + struct vdec_av1_slice_mem err_map; + struct vdec_av1_slice_mem row_info; + /* core */ + struct vdec_av1_slice_fb fb; + struct vdec_av1_slice_fb ref[V4L2_AV1_REFS_PER_FRAME]; + struct vdec_av1_slice_mem iq_table; + /* lat and core share*/ + struct vdec_av1_slice_mem tile; + struct vdec_av1_slice_slot slots; + s8 slot_id; + struct vdec_av1_slice_frame frame; + struct vdec_av1_slice_state state; + u32 cur_lst_tile_id; +}; + +/** + * struct vdec_av1_slice_pfc - per-frame context that contains a local vsi. + * pass it from lat to core + * @vsi: local vsi. copy to/from remote vsi before/after decoding + * @ref_idx: reference buffer timestamp + * @seq: picture sequence + */ +struct vdec_av1_slice_pfc { + struct vdec_av1_slice_vsi vsi; + u64 ref_idx[V4L2_AV1_REFS_PER_FRAME]; + int seq; +}; + +/** + * struct vdec_av1_slice_instance - represent one av1 instance + * @ctx: pointer to codec's context + * @vpu: VPU instance + * @iq_table: iq table buffer + * @cdf_table: cdf table buffer + * @mv: mv working buffer + * @cdf: cdf working buffer + * @seg: segmentation working buffer + * @cdf_temp: cdf temp buffer + * @tile: tile buffer + * @slots: slots info + * @tile_group: tile_group entry + * @level: level of current resolution + * @width: width of last picture + * @height: height of last picture + * @frame_type: frame_type of last picture + * @irq_enabled: irq to Main CPU or MicroP + * @inneracing_mode: is inneracing mode + * @init_vsi: vsi used for initialized AV1 instance + * @vsi: vsi used for decoding/flush ... + * @core_vsi: vsi used for Core stage + * @seq: global picture sequence + */ +struct vdec_av1_slice_instance { + struct mtk_vcodec_dec_ctx *ctx; + struct vdec_vpu_inst vpu; + + struct mtk_vcodec_mem iq_table; + struct mtk_vcodec_mem cdf_table; + + struct mtk_vcodec_mem mv[AV1_MAX_FRAME_BUF_COUNT]; + struct mtk_vcodec_mem cdf[AV1_MAX_FRAME_BUF_COUNT]; + struct mtk_vcodec_mem seg[AV1_MAX_FRAME_BUF_COUNT]; + struct mtk_vcodec_mem cdf_temp; + struct mtk_vcodec_mem tile; + struct vdec_av1_slice_slot slots; + struct vdec_av1_slice_tile_group tile_group; + + /* for resolution change and get_pic_info */ + enum vdec_av1_slice_resolution_level level; + u32 width; + u32 height; + + u32 frame_type; + u32 irq_enabled; + u32 inneracing_mode; + + /* MicroP vsi */ + union { + struct vdec_av1_slice_init_vsi *init_vsi; + struct vdec_av1_slice_vsi *vsi; + }; + struct vdec_av1_slice_vsi *core_vsi; + int seq; +}; + +static int vdec_av1_slice_core_decode(struct vdec_lat_buf *lat_buf); + +static inline int vdec_av1_slice_get_msb(u32 n) +{ + if (n == 0) + return 0; + return 31 ^ __builtin_clz(n); +} + +static inline bool vdec_av1_slice_need_scale(u32 ref_width, u32 ref_height, + u32 this_width, u32 this_height) +{ + return ((this_width << 1) >= ref_width) && + ((this_height << 1) >= ref_height) && + (this_width <= (ref_width << 4)) && + (this_height <= (ref_height << 4)); +} + +static void *vdec_av1_get_ctrl_ptr(struct mtk_vcodec_dec_ctx *ctx, int id) +{ + struct v4l2_ctrl *ctrl = v4l2_ctrl_find(&ctx->ctrl_hdl, id); + + if (!ctrl) + return ERR_PTR(-EINVAL); + + return ctrl->p_cur.p; +} + +static int vdec_av1_slice_init_cdf_table(struct vdec_av1_slice_instance *instance) +{ + u8 *remote_cdf_table; + struct mtk_vcodec_dec_ctx *ctx; + struct vdec_av1_slice_init_vsi *vsi; + int ret; + + ctx = instance->ctx; + vsi = instance->vpu.vsi; + remote_cdf_table = mtk_vcodec_fw_map_dm_addr(ctx->dev->fw_handler, + (u32)vsi->cdf_table_addr); + if (IS_ERR(remote_cdf_table)) { + mtk_vdec_err(ctx, "failed to map cdf table\n"); + return PTR_ERR(remote_cdf_table); + } + + mtk_vdec_debug(ctx, "map cdf table to 0x%p\n", remote_cdf_table); + + if (instance->cdf_table.va) + mtk_vcodec_mem_free(ctx, &instance->cdf_table); + instance->cdf_table.size = vsi->cdf_table_size; + + ret = mtk_vcodec_mem_alloc(ctx, &instance->cdf_table); + if (ret) + return ret; + + memcpy(instance->cdf_table.va, remote_cdf_table, vsi->cdf_table_size); + + return 0; +} + +static int vdec_av1_slice_init_iq_table(struct vdec_av1_slice_instance *instance) +{ + u8 *remote_iq_table; + struct mtk_vcodec_dec_ctx *ctx; + struct vdec_av1_slice_init_vsi *vsi; + int ret; + + ctx = instance->ctx; + vsi = instance->vpu.vsi; + remote_iq_table = mtk_vcodec_fw_map_dm_addr(ctx->dev->fw_handler, + (u32)vsi->iq_table_addr); + if (IS_ERR(remote_iq_table)) { + mtk_vdec_err(ctx, "failed to map iq table\n"); + return PTR_ERR(remote_iq_table); + } + + mtk_vdec_debug(ctx, "map iq table to 0x%p\n", remote_iq_table); + + if (instance->iq_table.va) + mtk_vcodec_mem_free(ctx, &instance->iq_table); + instance->iq_table.size = vsi->iq_table_size; + + ret = mtk_vcodec_mem_alloc(ctx, &instance->iq_table); + if (ret) + return ret; + + memcpy(instance->iq_table.va, remote_iq_table, vsi->iq_table_size); + + return 0; +} + +static int vdec_av1_slice_get_new_slot(struct vdec_av1_slice_vsi *vsi) +{ + struct vdec_av1_slice_slot *slots = &vsi->slots; + int new_slot_idx = AV1_INVALID_IDX; + int i; + + for (i = 0; i < AV1_MAX_FRAME_BUF_COUNT; i++) { + if (slots->frame_info[i].ref_count == 0) { + new_slot_idx = i; + break; + } + } + + if (new_slot_idx != AV1_INVALID_IDX) { + slots->frame_info[new_slot_idx].ref_count++; + slots->timestamp[new_slot_idx] = vsi->frame.cur_ts; + } + + return new_slot_idx; +} + +static inline void vdec_av1_slice_clear_fb(struct vdec_av1_slice_frame_info *frame_info) +{ + memset((void *)frame_info, 0, sizeof(struct vdec_av1_slice_frame_info)); +} + +static void vdec_av1_slice_decrease_ref_count(struct vdec_av1_slice_slot *slots, int fb_idx) +{ + struct vdec_av1_slice_frame_info *frame_info = slots->frame_info; + + frame_info[fb_idx].ref_count--; + if (frame_info[fb_idx].ref_count < 0) { + frame_info[fb_idx].ref_count = 0; + pr_err(MTK_DBG_V4L2_STR "av1_error: %s() fb_idx %d decrease ref_count error\n", + __func__, fb_idx); + } + + vdec_av1_slice_clear_fb(&frame_info[fb_idx]); +} + +static void vdec_av1_slice_cleanup_slots(struct vdec_av1_slice_slot *slots, + struct vdec_av1_slice_frame *frame, + struct v4l2_ctrl_av1_frame *ctrl_fh) +{ + int slot_id, ref_id; + + for (ref_id = 0; ref_id < V4L2_AV1_TOTAL_REFS_PER_FRAME; ref_id++) + frame->ref_frame_map[ref_id] = AV1_INVALID_IDX; + + for (slot_id = 0; slot_id < AV1_MAX_FRAME_BUF_COUNT; slot_id++) { + u64 timestamp = slots->timestamp[slot_id]; + bool ref_used = false; + + /* ignored unused slots */ + if (slots->frame_info[slot_id].ref_count == 0) + continue; + + for (ref_id = 0; ref_id < V4L2_AV1_TOTAL_REFS_PER_FRAME; ref_id++) { + if (ctrl_fh->reference_frame_ts[ref_id] == timestamp) { + frame->ref_frame_map[ref_id] = slot_id; + ref_used = true; + } + } + + if (!ref_used) + vdec_av1_slice_decrease_ref_count(slots, slot_id); + } +} + +static void vdec_av1_slice_setup_slot(struct vdec_av1_slice_instance *instance, + struct vdec_av1_slice_vsi *vsi, + struct v4l2_ctrl_av1_frame *ctrl_fh) +{ + struct vdec_av1_slice_frame_info *cur_frame_info; + struct vdec_av1_slice_uncompressed_header *uh = &vsi->frame.uh; + int ref_id; + + memcpy(&vsi->slots, &instance->slots, sizeof(instance->slots)); + vdec_av1_slice_cleanup_slots(&vsi->slots, &vsi->frame, ctrl_fh); + vsi->slot_id = vdec_av1_slice_get_new_slot(vsi); + + if (vsi->slot_id == AV1_INVALID_IDX) { + mtk_v4l2_vdec_err(instance->ctx, "warning:av1 get invalid index slot\n"); + vsi->slot_id = 0; + } + cur_frame_info = &vsi->slots.frame_info[vsi->slot_id]; + cur_frame_info->frame_type = uh->frame_type; + cur_frame_info->frame_is_intra = ((uh->frame_type == AV1_INTRA_ONLY_FRAME) || + (uh->frame_type == AV1_KEY_FRAME)); + cur_frame_info->order_hint = uh->order_hint; + cur_frame_info->upscaled_width = uh->upscaled_width; + cur_frame_info->pic_pitch = 0; + cur_frame_info->frame_width = uh->frame_width; + cur_frame_info->frame_height = uh->frame_height; + cur_frame_info->mi_cols = ((uh->frame_width + 7) >> 3) << 1; + cur_frame_info->mi_rows = ((uh->frame_height + 7) >> 3) << 1; + + /* ensure current frame is properly mapped if referenced */ + for (ref_id = 0; ref_id < V4L2_AV1_TOTAL_REFS_PER_FRAME; ref_id++) { + u64 timestamp = vsi->slots.timestamp[vsi->slot_id]; + + if (ctrl_fh->reference_frame_ts[ref_id] == timestamp) + vsi->frame.ref_frame_map[ref_id] = vsi->slot_id; + } +} + +static int vdec_av1_slice_alloc_working_buffer(struct vdec_av1_slice_instance *instance, + struct vdec_av1_slice_vsi *vsi) +{ + struct mtk_vcodec_dec_ctx *ctx = instance->ctx; + enum vdec_av1_slice_resolution_level level; + u32 max_sb_w, max_sb_h, max_w, max_h, w, h; + int i, ret; + + w = vsi->frame.uh.frame_width; + h = vsi->frame.uh.frame_height; + + if (w > VCODEC_DEC_4K_CODED_WIDTH || h > VCODEC_DEC_4K_CODED_HEIGHT) + /* 8K */ + return -EINVAL; + + if (w > MTK_VDEC_MAX_W || h > MTK_VDEC_MAX_H) { + /* 4K */ + level = AV1_RES_4K; + max_w = VCODEC_DEC_4K_CODED_WIDTH; + max_h = VCODEC_DEC_4K_CODED_HEIGHT; + } else { + /* FHD */ + level = AV1_RES_FHD; + max_w = MTK_VDEC_MAX_W; + max_h = MTK_VDEC_MAX_H; + } + + if (level == instance->level) + return 0; + + mtk_vdec_debug(ctx, "resolution level changed from %u to %u, %ux%u", + instance->level, level, w, h); + + max_sb_w = DIV_ROUND_UP(max_w, 128); + max_sb_h = DIV_ROUND_UP(max_h, 128); + + for (i = 0; i < AV1_MAX_FRAME_BUF_COUNT; i++) { + if (instance->mv[i].va) + mtk_vcodec_mem_free(ctx, &instance->mv[i]); + instance->mv[i].size = max_sb_w * max_sb_h * SZ_1K; + ret = mtk_vcodec_mem_alloc(ctx, &instance->mv[i]); + if (ret) + goto err; + + if (instance->seg[i].va) + mtk_vcodec_mem_free(ctx, &instance->seg[i]); + instance->seg[i].size = max_sb_w * max_sb_h * 512; + ret = mtk_vcodec_mem_alloc(ctx, &instance->seg[i]); + if (ret) + goto err; + + if (instance->cdf[i].va) + mtk_vcodec_mem_free(ctx, &instance->cdf[i]); + instance->cdf[i].size = AV1_CDF_TABLE_BUFFER_SIZE; + ret = mtk_vcodec_mem_alloc(ctx, &instance->cdf[i]); + if (ret) + goto err; + } + + if (!instance->cdf_temp.va) { + instance->cdf_temp.size = (SZ_1K * 16 * 100); + ret = mtk_vcodec_mem_alloc(ctx, &instance->cdf_temp); + if (ret) + goto err; + vsi->cdf_tmp.buf = instance->cdf_temp.dma_addr; + vsi->cdf_tmp.size = instance->cdf_temp.size; + } + + if (instance->tile.va) + mtk_vcodec_mem_free(ctx, &instance->tile); + + instance->tile.size = AV1_TILE_BUF_SIZE * V4L2_AV1_MAX_TILE_COUNT; + ret = mtk_vcodec_mem_alloc(ctx, &instance->tile); + if (ret) + goto err; + + instance->level = level; + return 0; + +err: + instance->level = AV1_RES_NONE; + return ret; +} + +static void vdec_av1_slice_free_working_buffer(struct vdec_av1_slice_instance *instance) +{ + struct mtk_vcodec_dec_ctx *ctx = instance->ctx; + int i; + + for (i = 0; i < ARRAY_SIZE(instance->mv); i++) + mtk_vcodec_mem_free(ctx, &instance->mv[i]); + + for (i = 0; i < ARRAY_SIZE(instance->seg); i++) + mtk_vcodec_mem_free(ctx, &instance->seg[i]); + + for (i = 0; i < ARRAY_SIZE(instance->cdf); i++) + mtk_vcodec_mem_free(ctx, &instance->cdf[i]); + + mtk_vcodec_mem_free(ctx, &instance->tile); + mtk_vcodec_mem_free(ctx, &instance->cdf_temp); + mtk_vcodec_mem_free(ctx, &instance->cdf_table); + mtk_vcodec_mem_free(ctx, &instance->iq_table); + + instance->level = AV1_RES_NONE; +} + +static inline void vdec_av1_slice_vsi_from_remote(struct vdec_av1_slice_vsi *vsi, + struct vdec_av1_slice_vsi *remote_vsi) +{ + memcpy(&vsi->trans, &remote_vsi->trans, sizeof(vsi->trans)); + memcpy(&vsi->state, &remote_vsi->state, sizeof(vsi->state)); +} + +static inline void vdec_av1_slice_vsi_to_remote(struct vdec_av1_slice_vsi *vsi, + struct vdec_av1_slice_vsi *remote_vsi) +{ + memcpy(remote_vsi, vsi, sizeof(*vsi)); +} + +static int vdec_av1_slice_setup_lat_from_src_buf(struct vdec_av1_slice_instance *instance, + struct vdec_av1_slice_vsi *vsi, + struct vdec_lat_buf *lat_buf) +{ + struct vb2_v4l2_buffer *src; + struct vb2_v4l2_buffer *dst; + + src = v4l2_m2m_next_src_buf(instance->ctx->m2m_ctx); + if (!src) + return -EINVAL; + + lat_buf->src_buf_req = src->vb2_buf.req_obj.req; + dst = &lat_buf->ts_info; + v4l2_m2m_buf_copy_metadata(src, dst, true); + vsi->frame.cur_ts = dst->vb2_buf.timestamp; + + return 0; +} + +static short vdec_av1_slice_resolve_divisor_32(u32 D, short *shift) +{ + int f; + int e; + + *shift = vdec_av1_slice_get_msb(D); + /* e is obtained from D after resetting the most significant 1 bit. */ + e = D - ((u32)1 << *shift); + /* Get the most significant DIV_LUT_BITS (8) bits of e into f */ + if (*shift > DIV_LUT_BITS) + f = AV1_DIV_ROUND_UP_POW2(e, *shift - DIV_LUT_BITS); + else + f = e << (DIV_LUT_BITS - *shift); + if (f > DIV_LUT_NUM) + return -1; + *shift += DIV_LUT_PREC_BITS; + /* Use f as lookup into the precomputed table of multipliers */ + return div_lut[f]; +} + +static void vdec_av1_slice_get_shear_params(struct vdec_av1_slice_gm *gm_params) +{ + const int *mat = gm_params->wmmat; + short shift; + short y; + long long gv, dv; + + if (gm_params->wmmat[2] <= 0) + return; + + gm_params->alpha = clamp_val(mat[2] - (1 << WARPEDMODEL_PREC_BITS), S16_MIN, S16_MAX); + gm_params->beta = clamp_val(mat[3], S16_MIN, S16_MAX); + + y = vdec_av1_slice_resolve_divisor_32(abs(mat[2]), &shift) * (mat[2] < 0 ? -1 : 1); + + gv = ((long long)mat[4] * (1 << WARPEDMODEL_PREC_BITS)) * y; + gm_params->gamma = clamp_val((int)AV1_DIV_ROUND_UP_POW2_SIGNED(gv, shift), + S16_MIN, S16_MAX); + + dv = ((long long)mat[3] * mat[4]) * y; + gm_params->delta = clamp_val(mat[5] - (int)AV1_DIV_ROUND_UP_POW2_SIGNED(dv, shift) - + (1 << WARPEDMODEL_PREC_BITS), S16_MIN, S16_MAX); + + gm_params->alpha = AV1_DIV_ROUND_UP_POW2_SIGNED(gm_params->alpha, WARP_PARAM_REDUCE_BITS) * + (1 << WARP_PARAM_REDUCE_BITS); + gm_params->beta = AV1_DIV_ROUND_UP_POW2_SIGNED(gm_params->beta, WARP_PARAM_REDUCE_BITS) * + (1 << WARP_PARAM_REDUCE_BITS); + gm_params->gamma = AV1_DIV_ROUND_UP_POW2_SIGNED(gm_params->gamma, WARP_PARAM_REDUCE_BITS) * + (1 << WARP_PARAM_REDUCE_BITS); + gm_params->delta = AV1_DIV_ROUND_UP_POW2_SIGNED(gm_params->delta, WARP_PARAM_REDUCE_BITS) * + (1 << WARP_PARAM_REDUCE_BITS); +} + +static void vdec_av1_slice_setup_gm(struct vdec_av1_slice_gm *gm, + struct v4l2_av1_global_motion *ctrl_gm) +{ + u32 i, j; + + for (i = 0; i < V4L2_AV1_TOTAL_REFS_PER_FRAME; i++) { + gm[i].wmtype = ctrl_gm->type[i]; + for (j = 0; j < 6; j++) + gm[i].wmmat[j] = ctrl_gm->params[i][j]; + + gm[i].invalid = !!(ctrl_gm->invalid & BIT(i)); + gm[i].alpha = 0; + gm[i].beta = 0; + gm[i].gamma = 0; + gm[i].delta = 0; + if (gm[i].wmtype <= V4L2_AV1_WARP_MODEL_AFFINE) + vdec_av1_slice_get_shear_params(&gm[i]); + } +} + +static void vdec_av1_slice_setup_seg(struct vdec_av1_slice_seg *seg, + struct v4l2_av1_segmentation *ctrl_seg) +{ + u32 i, j; + + seg->segmentation_enabled = SEGMENTATION_FLAG(ctrl_seg, ENABLED); + seg->segmentation_update_map = SEGMENTATION_FLAG(ctrl_seg, UPDATE_MAP); + seg->segmentation_temporal_update = SEGMENTATION_FLAG(ctrl_seg, TEMPORAL_UPDATE); + seg->segmentation_update_data = SEGMENTATION_FLAG(ctrl_seg, UPDATE_DATA); + seg->segid_preskip = SEGMENTATION_FLAG(ctrl_seg, SEG_ID_PRE_SKIP); + seg->last_active_segid = ctrl_seg->last_active_seg_id; + + for (i = 0; i < V4L2_AV1_MAX_SEGMENTS; i++) { + seg->feature_enabled_mask[i] = ctrl_seg->feature_enabled[i]; + for (j = 0; j < V4L2_AV1_SEG_LVL_MAX; j++) + seg->feature_data[i][j] = ctrl_seg->feature_data[i][j]; + } +} + +static void vdec_av1_slice_setup_quant(struct vdec_av1_slice_quantization *quant, + struct v4l2_av1_quantization *ctrl_quant) +{ + quant->base_q_idx = ctrl_quant->base_q_idx; + quant->delta_qydc = ctrl_quant->delta_q_y_dc; + quant->delta_qudc = ctrl_quant->delta_q_u_dc; + quant->delta_quac = ctrl_quant->delta_q_u_ac; + quant->delta_qvdc = ctrl_quant->delta_q_v_dc; + quant->delta_qvac = ctrl_quant->delta_q_v_ac; + quant->qm_y = ctrl_quant->qm_y; + quant->qm_u = ctrl_quant->qm_u; + quant->qm_v = ctrl_quant->qm_v; + quant->using_qmatrix = QUANT_FLAG(ctrl_quant, USING_QMATRIX); +} + +static int vdec_av1_slice_get_qindex(struct vdec_av1_slice_uncompressed_header *uh, + int segmentation_id) +{ + struct vdec_av1_slice_seg *seg = &uh->seg; + struct vdec_av1_slice_quantization *quant = &uh->quant; + int data = 0, qindex = 0; + + if (seg->segmentation_enabled && + (seg->feature_enabled_mask[segmentation_id] & BIT(SEG_LVL_ALT_Q))) { + data = seg->feature_data[segmentation_id][SEG_LVL_ALT_Q]; + qindex = quant->base_q_idx + data; + return clamp_val(qindex, 0, MAXQ); + } + + return quant->base_q_idx; +} + +static void vdec_av1_slice_setup_lr(struct vdec_av1_slice_lr *lr, + struct v4l2_av1_loop_restoration *ctrl_lr) +{ + int i; + + lr->use_lr = 0; + lr->use_chroma_lr = 0; + for (i = 0; i < V4L2_AV1_NUM_PLANES_MAX; i++) { + lr->frame_restoration_type[i] = ctrl_lr->frame_restoration_type[i]; + lr->loop_restoration_size[i] = ctrl_lr->loop_restoration_size[i]; + if (lr->frame_restoration_type[i]) { + lr->use_lr = 1; + if (i > 0) + lr->use_chroma_lr = 1; + } + } +} + +static void vdec_av1_slice_setup_lf(struct vdec_av1_slice_loop_filter *lf, + struct v4l2_av1_loop_filter *ctrl_lf) +{ + int i; + + for (i = 0; i < ARRAY_SIZE(lf->loop_filter_level); i++) + lf->loop_filter_level[i] = ctrl_lf->level[i]; + + for (i = 0; i < V4L2_AV1_TOTAL_REFS_PER_FRAME; i++) + lf->loop_filter_ref_deltas[i] = ctrl_lf->ref_deltas[i]; + + for (i = 0; i < ARRAY_SIZE(lf->loop_filter_mode_deltas); i++) + lf->loop_filter_mode_deltas[i] = ctrl_lf->mode_deltas[i]; + + lf->loop_filter_sharpness = ctrl_lf->sharpness; + lf->loop_filter_delta_enabled = + BIT_FLAG(ctrl_lf, V4L2_AV1_LOOP_FILTER_FLAG_DELTA_ENABLED); +} + +static void vdec_av1_slice_setup_cdef(struct vdec_av1_slice_cdef *cdef, + struct v4l2_av1_cdef *ctrl_cdef) +{ + int i; + + cdef->cdef_damping = ctrl_cdef->damping_minus_3 + 3; + cdef->cdef_bits = ctrl_cdef->bits; + + for (i = 0; i < V4L2_AV1_CDEF_MAX; i++) { + if (ctrl_cdef->y_sec_strength[i] == 4) + ctrl_cdef->y_sec_strength[i] -= 1; + + if (ctrl_cdef->uv_sec_strength[i] == 4) + ctrl_cdef->uv_sec_strength[i] -= 1; + + cdef->cdef_y_strength[i] = + ctrl_cdef->y_pri_strength[i] << SECONDARY_FILTER_STRENGTH_NUM_BITS | + ctrl_cdef->y_sec_strength[i]; + cdef->cdef_uv_strength[i] = + ctrl_cdef->uv_pri_strength[i] << SECONDARY_FILTER_STRENGTH_NUM_BITS | + ctrl_cdef->uv_sec_strength[i]; + } +} + +static void vdec_av1_slice_setup_seq(struct vdec_av1_slice_seq_header *seq, + struct v4l2_ctrl_av1_sequence *ctrl_seq) +{ + seq->bitdepth = ctrl_seq->bit_depth; + seq->max_frame_width = ctrl_seq->max_frame_width_minus_1 + 1; + seq->max_frame_height = ctrl_seq->max_frame_height_minus_1 + 1; + seq->enable_superres = SEQUENCE_FLAG(ctrl_seq, ENABLE_SUPERRES); + seq->enable_filter_intra = SEQUENCE_FLAG(ctrl_seq, ENABLE_FILTER_INTRA); + seq->enable_intra_edge_filter = SEQUENCE_FLAG(ctrl_seq, ENABLE_INTRA_EDGE_FILTER); + seq->enable_interintra_compound = SEQUENCE_FLAG(ctrl_seq, ENABLE_INTERINTRA_COMPOUND); + seq->enable_masked_compound = SEQUENCE_FLAG(ctrl_seq, ENABLE_MASKED_COMPOUND); + seq->enable_dual_filter = SEQUENCE_FLAG(ctrl_seq, ENABLE_DUAL_FILTER); + seq->enable_jnt_comp = SEQUENCE_FLAG(ctrl_seq, ENABLE_JNT_COMP); + seq->mono_chrome = SEQUENCE_FLAG(ctrl_seq, MONO_CHROME); + seq->enable_order_hint = SEQUENCE_FLAG(ctrl_seq, ENABLE_ORDER_HINT); + seq->order_hint_bits = ctrl_seq->order_hint_bits; + seq->use_128x128_superblock = SEQUENCE_FLAG(ctrl_seq, USE_128X128_SUPERBLOCK); + seq->subsampling_x = SEQUENCE_FLAG(ctrl_seq, SUBSAMPLING_X); + seq->subsampling_y = SEQUENCE_FLAG(ctrl_seq, SUBSAMPLING_Y); +} + +static void vdec_av1_slice_setup_tile(struct vdec_av1_slice_frame *frame, + struct v4l2_av1_tile_info *ctrl_tile) +{ + struct vdec_av1_slice_seq_header *seq = &frame->seq; + struct vdec_av1_slice_tile *tile = &frame->uh.tile; + u32 mib_size_log2 = seq->use_128x128_superblock ? 5 : 4; + int i; + + tile->tile_cols = ctrl_tile->tile_cols; + tile->tile_rows = ctrl_tile->tile_rows; + tile->context_update_tile_id = ctrl_tile->context_update_tile_id; + tile->uniform_tile_spacing_flag = + BIT_FLAG(ctrl_tile, V4L2_AV1_TILE_INFO_FLAG_UNIFORM_TILE_SPACING); + + for (i = 0; i < tile->tile_cols + 1; i++) + tile->mi_col_starts[i] = + ALIGN(ctrl_tile->mi_col_starts[i], BIT(mib_size_log2)) >> mib_size_log2; + + for (i = 0; i < tile->tile_rows + 1; i++) + tile->mi_row_starts[i] = + ALIGN(ctrl_tile->mi_row_starts[i], BIT(mib_size_log2)) >> mib_size_log2; +} + +static void vdec_av1_slice_setup_uh(struct vdec_av1_slice_instance *instance, + struct vdec_av1_slice_frame *frame, + struct v4l2_ctrl_av1_frame *ctrl_fh) +{ + struct vdec_av1_slice_uncompressed_header *uh = &frame->uh; + int i; + + uh->use_ref_frame_mvs = FH_FLAG(ctrl_fh, USE_REF_FRAME_MVS); + uh->order_hint = ctrl_fh->order_hint; + vdec_av1_slice_setup_gm(uh->gm, &ctrl_fh->global_motion); + uh->upscaled_width = ctrl_fh->upscaled_width; + uh->frame_width = ctrl_fh->frame_width_minus_1 + 1; + uh->frame_height = ctrl_fh->frame_height_minus_1 + 1; + uh->mi_cols = ((uh->frame_width + 7) >> 3) << 1; + uh->mi_rows = ((uh->frame_height + 7) >> 3) << 1; + uh->reduced_tx_set = FH_FLAG(ctrl_fh, REDUCED_TX_SET); + uh->tx_mode = ctrl_fh->tx_mode; + uh->uniform_tile_spacing_flag = + BIT_FLAG(&ctrl_fh->tile_info, V4L2_AV1_TILE_INFO_FLAG_UNIFORM_TILE_SPACING); + uh->interpolation_filter = ctrl_fh->interpolation_filter; + uh->allow_warped_motion = FH_FLAG(ctrl_fh, ALLOW_WARPED_MOTION); + uh->is_motion_mode_switchable = FH_FLAG(ctrl_fh, IS_MOTION_MODE_SWITCHABLE); + uh->frame_type = ctrl_fh->frame_type; + uh->frame_is_intra = (uh->frame_type == V4L2_AV1_INTRA_ONLY_FRAME || + uh->frame_type == V4L2_AV1_KEY_FRAME); + + if (!uh->frame_is_intra && FH_FLAG(ctrl_fh, REFERENCE_SELECT)) + uh->reference_mode = AV1_REFERENCE_MODE_SELECT; + else + uh->reference_mode = AV1_SINGLE_REFERENCE; + + uh->allow_high_precision_mv = FH_FLAG(ctrl_fh, ALLOW_HIGH_PRECISION_MV); + uh->allow_intra_bc = FH_FLAG(ctrl_fh, ALLOW_INTRABC); + uh->force_integer_mv = FH_FLAG(ctrl_fh, FORCE_INTEGER_MV); + uh->allow_screen_content_tools = FH_FLAG(ctrl_fh, ALLOW_SCREEN_CONTENT_TOOLS); + uh->error_resilient_mode = FH_FLAG(ctrl_fh, ERROR_RESILIENT_MODE); + uh->primary_ref_frame = ctrl_fh->primary_ref_frame; + uh->disable_frame_end_update_cdf = + FH_FLAG(ctrl_fh, DISABLE_FRAME_END_UPDATE_CDF); + uh->disable_cdf_update = FH_FLAG(ctrl_fh, DISABLE_CDF_UPDATE); + uh->skip_mode.skip_mode_present = FH_FLAG(ctrl_fh, SKIP_MODE_PRESENT); + uh->skip_mode.skip_mode_frame[0] = + ctrl_fh->skip_mode_frame[0] - V4L2_AV1_REF_LAST_FRAME; + uh->skip_mode.skip_mode_frame[1] = + ctrl_fh->skip_mode_frame[1] - V4L2_AV1_REF_LAST_FRAME; + uh->skip_mode.skip_mode_allowed = ctrl_fh->skip_mode_frame[0] ? 1 : 0; + + vdec_av1_slice_setup_seg(&uh->seg, &ctrl_fh->segmentation); + uh->delta_q_lf.delta_q_present = QUANT_FLAG(&ctrl_fh->quantization, DELTA_Q_PRESENT); + uh->delta_q_lf.delta_q_res = 1 << ctrl_fh->quantization.delta_q_res; + uh->delta_q_lf.delta_lf_present = + BIT_FLAG(&ctrl_fh->loop_filter, V4L2_AV1_LOOP_FILTER_FLAG_DELTA_LF_PRESENT); + uh->delta_q_lf.delta_lf_res = ctrl_fh->loop_filter.delta_lf_res; + uh->delta_q_lf.delta_lf_multi = + BIT_FLAG(&ctrl_fh->loop_filter, V4L2_AV1_LOOP_FILTER_FLAG_DELTA_LF_MULTI); + vdec_av1_slice_setup_quant(&uh->quant, &ctrl_fh->quantization); + + uh->coded_loss_less = 1; + for (i = 0; i < V4L2_AV1_MAX_SEGMENTS; i++) { + uh->quant.qindex[i] = vdec_av1_slice_get_qindex(uh, i); + uh->loss_less_array[i] = + (uh->quant.qindex[i] == 0 && uh->quant.delta_qydc == 0 && + uh->quant.delta_quac == 0 && uh->quant.delta_qudc == 0 && + uh->quant.delta_qvac == 0 && uh->quant.delta_qvdc == 0); + + if (!uh->loss_less_array[i]) + uh->coded_loss_less = 0; + } + + vdec_av1_slice_setup_lr(&uh->lr, &ctrl_fh->loop_restoration); + uh->superres_denom = ctrl_fh->superres_denom; + vdec_av1_slice_setup_lf(&uh->loop_filter, &ctrl_fh->loop_filter); + vdec_av1_slice_setup_cdef(&uh->cdef, &ctrl_fh->cdef); + vdec_av1_slice_setup_tile(frame, &ctrl_fh->tile_info); +} + +static int vdec_av1_slice_setup_tile_group(struct vdec_av1_slice_instance *instance, + struct vdec_av1_slice_vsi *vsi) +{ + struct v4l2_ctrl_av1_tile_group_entry *ctrl_tge; + struct vdec_av1_slice_tile_group *tile_group = &instance->tile_group; + struct vdec_av1_slice_uncompressed_header *uh = &vsi->frame.uh; + struct vdec_av1_slice_tile *tile = &uh->tile; + struct v4l2_ctrl *ctrl; + u32 tge_size; + int i; + + ctrl = v4l2_ctrl_find(&instance->ctx->ctrl_hdl, V4L2_CID_STATELESS_AV1_TILE_GROUP_ENTRY); + if (!ctrl) + return -EINVAL; + + tge_size = ctrl->elems; + ctrl_tge = (struct v4l2_ctrl_av1_tile_group_entry *)ctrl->p_cur.p; + + tile_group->num_tiles = tile->tile_cols * tile->tile_rows; + + if (tile_group->num_tiles != tge_size || + tile_group->num_tiles > V4L2_AV1_MAX_TILE_COUNT) { + mtk_vdec_err(instance->ctx, "invalid tge_size %d, tile_num:%d\n", + tge_size, tile_group->num_tiles); + return -EINVAL; + } + + for (i = 0; i < tge_size; i++) { + if (i != ctrl_tge[i].tile_row * vsi->frame.uh.tile.tile_cols + + ctrl_tge[i].tile_col) { + mtk_vdec_err(instance->ctx, "invalid tge info %d, %d %d %d\n", + i, ctrl_tge[i].tile_row, ctrl_tge[i].tile_col, + vsi->frame.uh.tile.tile_rows); + return -EINVAL; + } + tile_group->tile_size[i] = ctrl_tge[i].tile_size; + tile_group->tile_start_offset[i] = ctrl_tge[i].tile_offset; + } + + return 0; +} + +static inline void vdec_av1_slice_setup_state(struct vdec_av1_slice_vsi *vsi) +{ + memset(&vsi->state, 0, sizeof(vsi->state)); +} + +static void vdec_av1_slice_setup_scale_factors(struct vdec_av1_slice_frame_refs *frame_ref, + struct vdec_av1_slice_frame_info *ref_frame_info, + struct vdec_av1_slice_uncompressed_header *uh) +{ + struct vdec_av1_slice_scale_factors *scale_factors = &frame_ref->scale_factors; + u32 ref_upscaled_width = ref_frame_info->upscaled_width; + u32 ref_frame_height = ref_frame_info->frame_height; + u32 frame_width = uh->frame_width; + u32 frame_height = uh->frame_height; + + if (!vdec_av1_slice_need_scale(ref_upscaled_width, ref_frame_height, + frame_width, frame_height)) { + scale_factors->x_scale = -1; + scale_factors->y_scale = -1; + scale_factors->is_scaled = 0; + return; + } + + scale_factors->x_scale = + ((ref_upscaled_width << AV1_REF_SCALE_SHIFT) + (frame_width >> 1)) / frame_width; + scale_factors->y_scale = + ((ref_frame_height << AV1_REF_SCALE_SHIFT) + (frame_height >> 1)) / frame_height; + scale_factors->is_scaled = + (scale_factors->x_scale != AV1_REF_INVALID_SCALE) && + (scale_factors->y_scale != AV1_REF_INVALID_SCALE) && + (scale_factors->x_scale != AV1_REF_NO_SCALE || + scale_factors->y_scale != AV1_REF_NO_SCALE); + scale_factors->x_step = + AV1_DIV_ROUND_UP_POW2(scale_factors->x_scale, + AV1_REF_SCALE_SHIFT - AV1_SCALE_SUBPEL_BITS); + scale_factors->y_step = + AV1_DIV_ROUND_UP_POW2(scale_factors->y_scale, + AV1_REF_SCALE_SHIFT - AV1_SCALE_SUBPEL_BITS); +} + +static unsigned char vdec_av1_slice_get_sign_bias(int a, + int b, + u8 enable_order_hint, + u8 order_hint_bits) +{ + int diff = 0; + int m = 0; + unsigned char result = 0; + + if (!enable_order_hint) + return 0; + + diff = a - b; + m = 1 << (order_hint_bits - 1); + diff = (diff & (m - 1)) - (diff & m); + + if (diff > 0) + result = 1; + + return result; +} + +static void vdec_av1_slice_setup_ref(struct vdec_av1_slice_pfc *pfc, + struct v4l2_ctrl_av1_frame *ctrl_fh) +{ + struct vdec_av1_slice_vsi *vsi = &pfc->vsi; + struct vdec_av1_slice_frame *frame = &vsi->frame; + struct vdec_av1_slice_slot *slots = &vsi->slots; + struct vdec_av1_slice_uncompressed_header *uh = &frame->uh; + struct vdec_av1_slice_seq_header *seq = &frame->seq; + struct vdec_av1_slice_frame_info *cur_frame_info = + &slots->frame_info[vsi->slot_id]; + struct vdec_av1_slice_frame_info *frame_info; + int i, slot_id; + + if (uh->frame_is_intra) + return; + + for (i = 0; i < V4L2_AV1_REFS_PER_FRAME; i++) { + int ref_idx = ctrl_fh->ref_frame_idx[i]; + + pfc->ref_idx[i] = ctrl_fh->reference_frame_ts[ref_idx]; + slot_id = frame->ref_frame_map[ref_idx]; + frame_info = &slots->frame_info[slot_id]; + if (slot_id == AV1_INVALID_IDX) { + pr_err(MTK_DBG_V4L2_STR "cannot match reference[%d] 0x%llx\n", i, + ctrl_fh->reference_frame_ts[ref_idx]); + frame->order_hints[i] = 0; + frame->ref_frame_valid[i] = 0; + continue; + } + + frame->frame_refs[i].ref_fb_idx = slot_id; + vdec_av1_slice_setup_scale_factors(&frame->frame_refs[i], + frame_info, uh); + if (!seq->enable_order_hint) + frame->ref_frame_sign_bias[i + 1] = 0; + else + frame->ref_frame_sign_bias[i + 1] = + vdec_av1_slice_get_sign_bias(frame_info->order_hint, + uh->order_hint, + seq->enable_order_hint, + seq->order_hint_bits); + + frame->order_hints[i] = ctrl_fh->order_hints[i + 1]; + cur_frame_info->order_hints[i] = frame->order_hints[i]; + frame->ref_frame_valid[i] = 1; + } +} + +static void vdec_av1_slice_get_previous(struct vdec_av1_slice_vsi *vsi) +{ + struct vdec_av1_slice_frame *frame = &vsi->frame; + + if (frame->uh.primary_ref_frame == AV1_PRIMARY_REF_NONE) + frame->prev_fb_idx = AV1_INVALID_IDX; + else + frame->prev_fb_idx = frame->frame_refs[frame->uh.primary_ref_frame].ref_fb_idx; +} + +static inline void vdec_av1_slice_setup_operating_mode(struct vdec_av1_slice_instance *instance, + struct vdec_av1_slice_frame *frame) +{ + frame->large_scale_tile = 0; +} + +static int vdec_av1_slice_setup_pfc(struct vdec_av1_slice_instance *instance, + struct vdec_av1_slice_pfc *pfc) +{ + struct v4l2_ctrl_av1_frame *ctrl_fh; + struct v4l2_ctrl_av1_sequence *ctrl_seq; + struct vdec_av1_slice_vsi *vsi = &pfc->vsi; + int ret = 0; + + /* frame header */ + ctrl_fh = (struct v4l2_ctrl_av1_frame *) + vdec_av1_get_ctrl_ptr(instance->ctx, + V4L2_CID_STATELESS_AV1_FRAME); + if (IS_ERR(ctrl_fh)) + return PTR_ERR(ctrl_fh); + + ctrl_seq = (struct v4l2_ctrl_av1_sequence *) + vdec_av1_get_ctrl_ptr(instance->ctx, + V4L2_CID_STATELESS_AV1_SEQUENCE); + if (IS_ERR(ctrl_seq)) + return PTR_ERR(ctrl_seq); + + /* setup vsi information */ + vdec_av1_slice_setup_seq(&vsi->frame.seq, ctrl_seq); + vdec_av1_slice_setup_uh(instance, &vsi->frame, ctrl_fh); + vdec_av1_slice_setup_operating_mode(instance, &vsi->frame); + + vdec_av1_slice_setup_state(vsi); + vdec_av1_slice_setup_slot(instance, vsi, ctrl_fh); + vdec_av1_slice_setup_ref(pfc, ctrl_fh); + vdec_av1_slice_get_previous(vsi); + + pfc->seq = instance->seq; + instance->seq++; + + return ret; +} + +static void vdec_av1_slice_setup_lat_buffer(struct vdec_av1_slice_instance *instance, + struct vdec_av1_slice_vsi *vsi, + struct mtk_vcodec_mem *bs, + struct vdec_lat_buf *lat_buf) +{ + struct vdec_av1_slice_work_buffer *work_buffer; + int i; + + vsi->bs.dma_addr = bs->dma_addr; + vsi->bs.size = bs->size; + + vsi->ube.dma_addr = lat_buf->ctx->msg_queue.wdma_addr.dma_addr; + vsi->ube.size = lat_buf->ctx->msg_queue.wdma_addr.size; + vsi->trans.dma_addr = lat_buf->ctx->msg_queue.wdma_wptr_addr; + /* used to store trans end */ + vsi->trans.dma_addr_end = lat_buf->ctx->msg_queue.wdma_rptr_addr; + vsi->err_map.dma_addr = lat_buf->wdma_err_addr.dma_addr; + vsi->err_map.size = lat_buf->wdma_err_addr.size; + vsi->rd_mv.dma_addr = lat_buf->rd_mv_addr.dma_addr; + vsi->rd_mv.size = lat_buf->rd_mv_addr.size; + + vsi->row_info.buf = 0; + vsi->row_info.size = 0; + + work_buffer = vsi->work_buffer; + + for (i = 0; i < AV1_MAX_FRAME_BUF_COUNT; i++) { + work_buffer[i].mv_addr.buf = instance->mv[i].dma_addr; + work_buffer[i].mv_addr.size = instance->mv[i].size; + work_buffer[i].segid_addr.buf = instance->seg[i].dma_addr; + work_buffer[i].segid_addr.size = instance->seg[i].size; + work_buffer[i].cdf_addr.buf = instance->cdf[i].dma_addr; + work_buffer[i].cdf_addr.size = instance->cdf[i].size; + } + + vsi->cdf_tmp.buf = instance->cdf_temp.dma_addr; + vsi->cdf_tmp.size = instance->cdf_temp.size; + + vsi->tile.buf = instance->tile.dma_addr; + vsi->tile.size = instance->tile.size; + memcpy(lat_buf->tile_addr.va, instance->tile.va, 64 * instance->tile_group.num_tiles); + + vsi->cdf_table.buf = instance->cdf_table.dma_addr; + vsi->cdf_table.size = instance->cdf_table.size; + vsi->iq_table.buf = instance->iq_table.dma_addr; + vsi->iq_table.size = instance->iq_table.size; +} + +static void vdec_av1_slice_setup_seg_buffer(struct vdec_av1_slice_instance *instance, + struct vdec_av1_slice_vsi *vsi) +{ + struct vdec_av1_slice_uncompressed_header *uh = &vsi->frame.uh; + struct mtk_vcodec_mem *buf; + + /* reset segment buffer */ + if (uh->primary_ref_frame == AV1_PRIMARY_REF_NONE || !uh->seg.segmentation_enabled) { + mtk_vdec_debug(instance->ctx, "reset seg %d\n", vsi->slot_id); + if (vsi->slot_id != AV1_INVALID_IDX) { + buf = &instance->seg[vsi->slot_id]; + memset(buf->va, 0, buf->size); + } + } +} + +static void vdec_av1_slice_setup_tile_buffer(struct vdec_av1_slice_instance *instance, + struct vdec_av1_slice_vsi *vsi, + struct mtk_vcodec_mem *bs) +{ + struct vdec_av1_slice_tile_group *tile_group = &instance->tile_group; + struct vdec_av1_slice_uncompressed_header *uh = &vsi->frame.uh; + struct vdec_av1_slice_tile *tile = &uh->tile; + u32 tile_num, tile_row, tile_col; + u32 allow_update_cdf = 0; + u32 sb_boundary_x_m1 = 0, sb_boundary_y_m1 = 0; + int tile_info_base; + u64 tile_buf_pa; + u32 *tile_info_buf = instance->tile.va; + u64 pa = (u64)bs->dma_addr; + + if (uh->disable_cdf_update == 0) + allow_update_cdf = 1; + + for (tile_num = 0; tile_num < tile_group->num_tiles; tile_num++) { + /* each uint32 takes place of 4 bytes */ + tile_info_base = (AV1_TILE_BUF_SIZE * tile_num) >> 2; + tile_row = tile_num / tile->tile_cols; + tile_col = tile_num % tile->tile_cols; + tile_info_buf[tile_info_base + 0] = (tile_group->tile_size[tile_num] << 3); + tile_buf_pa = pa + tile_group->tile_start_offset[tile_num]; + + /* save av1 tile high 4bits(bit 32-35) address in lower 4 bits position + * and clear original for hw requirement. + */ + tile_info_buf[tile_info_base + 1] = (tile_buf_pa & 0xFFFFFFF0ull) | + ((tile_buf_pa & 0xF00000000ull) >> 32); + tile_info_buf[tile_info_base + 2] = (tile_buf_pa & 0xFull) << 3; + + sb_boundary_x_m1 = + (tile->mi_col_starts[tile_col + 1] - tile->mi_col_starts[tile_col] - 1) & + 0x3f; + sb_boundary_y_m1 = + (tile->mi_row_starts[tile_row + 1] - tile->mi_row_starts[tile_row] - 1) & + 0x1ff; + + tile_info_buf[tile_info_base + 3] = (sb_boundary_y_m1 << 7) | sb_boundary_x_m1; + tile_info_buf[tile_info_base + 4] = ((allow_update_cdf << 18) | (1 << 16)); + + if (tile_num == tile->context_update_tile_id && + uh->disable_frame_end_update_cdf == 0) + tile_info_buf[tile_info_base + 4] |= (1 << 17); + + mtk_vdec_debug(instance->ctx, "// tile buf %d pos(%dx%d) offset 0x%x\n", + tile_num, tile_row, tile_col, tile_info_base); + mtk_vdec_debug(instance->ctx, "// %08x %08x %08x %08x\n", + tile_info_buf[tile_info_base + 0], + tile_info_buf[tile_info_base + 1], + tile_info_buf[tile_info_base + 2], + tile_info_buf[tile_info_base + 3]); + mtk_vdec_debug(instance->ctx, "// %08x %08x %08x %08x\n", + tile_info_buf[tile_info_base + 4], + tile_info_buf[tile_info_base + 5], + tile_info_buf[tile_info_base + 6], + tile_info_buf[tile_info_base + 7]); + } +} + +static int vdec_av1_slice_setup_lat(struct vdec_av1_slice_instance *instance, + struct mtk_vcodec_mem *bs, + struct vdec_lat_buf *lat_buf, + struct vdec_av1_slice_pfc *pfc) +{ + struct vdec_av1_slice_vsi *vsi = &pfc->vsi; + int ret; + + ret = vdec_av1_slice_setup_lat_from_src_buf(instance, vsi, lat_buf); + if (ret) + return ret; + + ret = vdec_av1_slice_setup_pfc(instance, pfc); + if (ret) + return ret; + + ret = vdec_av1_slice_setup_tile_group(instance, vsi); + if (ret) + return ret; + + ret = vdec_av1_slice_alloc_working_buffer(instance, vsi); + if (ret) + return ret; + + vdec_av1_slice_setup_seg_buffer(instance, vsi); + vdec_av1_slice_setup_tile_buffer(instance, vsi, bs); + vdec_av1_slice_setup_lat_buffer(instance, vsi, bs, lat_buf); + + return 0; +} + +static int vdec_av1_slice_update_lat(struct vdec_av1_slice_instance *instance, + struct vdec_lat_buf *lat_buf, + struct vdec_av1_slice_pfc *pfc) +{ + struct vdec_av1_slice_vsi *vsi; + + vsi = &pfc->vsi; + mtk_vdec_debug(instance->ctx, "frame %u LAT CRC 0x%08x, output size is %d\n", + pfc->seq, vsi->state.crc[0], vsi->state.out_size); + + /* buffer full, need to re-decode */ + if (vsi->state.full) { + /* buffer not enough */ + if (vsi->trans.dma_addr_end - vsi->trans.dma_addr == vsi->ube.size) + return -ENOMEM; + return -EAGAIN; + } + + instance->width = vsi->frame.uh.upscaled_width; + instance->height = vsi->frame.uh.frame_height; + instance->frame_type = vsi->frame.uh.frame_type; + + return 0; +} + +static int vdec_av1_slice_setup_core_to_dst_buf(struct vdec_av1_slice_instance *instance, + struct vdec_lat_buf *lat_buf) +{ + struct vb2_v4l2_buffer *dst; + + dst = v4l2_m2m_next_dst_buf(instance->ctx->m2m_ctx); + if (!dst) + return -EINVAL; + + v4l2_m2m_buf_copy_metadata(&lat_buf->ts_info, dst, true); + + return 0; +} + +static int vdec_av1_slice_setup_core_buffer(struct vdec_av1_slice_instance *instance, + struct vdec_av1_slice_pfc *pfc, + struct vdec_av1_slice_vsi *vsi, + struct vdec_fb *fb, + struct vdec_lat_buf *lat_buf) +{ + struct vb2_buffer *vb; + struct vb2_queue *vq; + int w, h, plane, size; + int i; + + plane = instance->ctx->q_data[MTK_Q_DATA_DST].fmt->num_planes; + w = vsi->frame.uh.upscaled_width; + h = vsi->frame.uh.frame_height; + size = ALIGN(w, VCODEC_DEC_ALIGNED_64) * ALIGN(h, VCODEC_DEC_ALIGNED_64); + + /* frame buffer */ + vsi->fb.y.dma_addr = fb->base_y.dma_addr; + if (plane == 1) + vsi->fb.c.dma_addr = fb->base_y.dma_addr + size; + else + vsi->fb.c.dma_addr = fb->base_c.dma_addr; + + /* reference buffers */ + vq = v4l2_m2m_get_vq(instance->ctx->m2m_ctx, V4L2_BUF_TYPE_VIDEO_CAPTURE_MPLANE); + if (!vq) + return -EINVAL; + + /* get current output buffer */ + vb = &v4l2_m2m_next_dst_buf(instance->ctx->m2m_ctx)->vb2_buf; + if (!vb) + return -EINVAL; + + /* get buffer address from vb2buf */ + for (i = 0; i < V4L2_AV1_REFS_PER_FRAME; i++) { + struct vdec_av1_slice_fb *vref = &vsi->ref[i]; + + vb = vb2_find_buffer(vq, pfc->ref_idx[i]); + if (!vb) { + memset(vref, 0, sizeof(*vref)); + continue; + } + + vref->y.dma_addr = vb2_dma_contig_plane_dma_addr(vb, 0); + if (plane == 1) + vref->c.dma_addr = vref->y.dma_addr + size; + else + vref->c.dma_addr = vb2_dma_contig_plane_dma_addr(vb, 1); + } + vsi->tile.dma_addr = lat_buf->tile_addr.dma_addr; + vsi->tile.size = lat_buf->tile_addr.size; + + return 0; +} + +static int vdec_av1_slice_setup_core(struct vdec_av1_slice_instance *instance, + struct vdec_fb *fb, + struct vdec_lat_buf *lat_buf, + struct vdec_av1_slice_pfc *pfc) +{ + struct vdec_av1_slice_vsi *vsi = &pfc->vsi; + int ret; + + ret = vdec_av1_slice_setup_core_to_dst_buf(instance, lat_buf); + if (ret) + return ret; + + ret = vdec_av1_slice_setup_core_buffer(instance, pfc, vsi, fb, lat_buf); + if (ret) + return ret; + + return 0; +} + +static int vdec_av1_slice_update_core(struct vdec_av1_slice_instance *instance, + struct vdec_lat_buf *lat_buf, + struct vdec_av1_slice_pfc *pfc) +{ + struct vdec_av1_slice_vsi *vsi = instance->core_vsi; + + mtk_vdec_debug(instance->ctx, "frame %u Y_CRC %08x %08x %08x %08x\n", + pfc->seq, vsi->state.crc[0], vsi->state.crc[1], + vsi->state.crc[2], vsi->state.crc[3]); + mtk_vdec_debug(instance->ctx, "frame %u C_CRC %08x %08x %08x %08x\n", + pfc->seq, vsi->state.crc[8], vsi->state.crc[9], + vsi->state.crc[10], vsi->state.crc[11]); + + return 0; +} + +static int vdec_av1_slice_init(struct mtk_vcodec_dec_ctx *ctx) +{ + struct vdec_av1_slice_instance *instance; + struct vdec_av1_slice_init_vsi *vsi; + int ret; + + instance = kzalloc(sizeof(*instance), GFP_KERNEL); + if (!instance) + return -ENOMEM; + + instance->ctx = ctx; + instance->vpu.id = SCP_IPI_VDEC_LAT; + instance->vpu.core_id = SCP_IPI_VDEC_CORE; + instance->vpu.ctx = ctx; + instance->vpu.codec_type = ctx->current_codec; + + ret = vpu_dec_init(&instance->vpu); + if (ret) { + mtk_vdec_err(ctx, "failed to init vpu dec, ret %d\n", ret); + goto error_vpu_init; + } + + /* init vsi and global flags */ + vsi = instance->vpu.vsi; + if (!vsi) { + mtk_vdec_err(ctx, "failed to get AV1 vsi\n"); + ret = -EINVAL; + goto error_vsi; + } + instance->init_vsi = vsi; + instance->core_vsi = mtk_vcodec_fw_map_dm_addr(ctx->dev->fw_handler, (u32)vsi->core_vsi); + + if (!instance->core_vsi) { + mtk_vdec_err(ctx, "failed to get AV1 core vsi\n"); + ret = -EINVAL; + goto error_vsi; + } + + if (vsi->vsi_size != sizeof(struct vdec_av1_slice_vsi)) + mtk_vdec_err(ctx, "remote vsi size 0x%x mismatch! expected: 0x%zx\n", + vsi->vsi_size, sizeof(struct vdec_av1_slice_vsi)); + + instance->irq_enabled = 1; + instance->inneracing_mode = IS_VDEC_INNER_RACING(instance->ctx->dev->dec_capability); + + mtk_vdec_debug(ctx, "vsi 0x%p core_vsi 0x%llx 0x%p, inneracing_mode %d\n", + vsi, vsi->core_vsi, instance->core_vsi, instance->inneracing_mode); + + ret = vdec_av1_slice_init_cdf_table(instance); + if (ret) + goto error_vsi; + + ret = vdec_av1_slice_init_iq_table(instance); + if (ret) + goto error_vsi; + + ctx->drv_handle = instance; + + return 0; +error_vsi: + vpu_dec_deinit(&instance->vpu); +error_vpu_init: + kfree(instance); + + return ret; +} + +static void vdec_av1_slice_deinit(void *h_vdec) +{ + struct vdec_av1_slice_instance *instance = h_vdec; + + if (!instance) + return; + mtk_vdec_debug(instance->ctx, "h_vdec 0x%p\n", h_vdec); + vpu_dec_deinit(&instance->vpu); + vdec_av1_slice_free_working_buffer(instance); + vdec_msg_queue_deinit(&instance->ctx->msg_queue, instance->ctx); + kfree(instance); +} + +static int vdec_av1_slice_flush(void *h_vdec, struct mtk_vcodec_mem *bs, + struct vdec_fb *fb, bool *res_chg) +{ + struct vdec_av1_slice_instance *instance = h_vdec; + int i; + + mtk_vdec_debug(instance->ctx, "flush ...\n"); + + vdec_msg_queue_wait_lat_buf_full(&instance->ctx->msg_queue); + + for (i = 0; i < AV1_MAX_FRAME_BUF_COUNT; i++) + vdec_av1_slice_clear_fb(&instance->slots.frame_info[i]); + + return vpu_dec_reset(&instance->vpu); +} + +static void vdec_av1_slice_get_pic_info(struct vdec_av1_slice_instance *instance) +{ + struct mtk_vcodec_dec_ctx *ctx = instance->ctx; + u32 data[3]; + + mtk_vdec_debug(ctx, "w %u h %u\n", ctx->picinfo.pic_w, ctx->picinfo.pic_h); + + data[0] = ctx->picinfo.pic_w; + data[1] = ctx->picinfo.pic_h; + data[2] = ctx->capture_fourcc; + vpu_dec_get_param(&instance->vpu, data, 3, GET_PARAM_PIC_INFO); + + ctx->picinfo.buf_w = ALIGN(ctx->picinfo.pic_w, VCODEC_DEC_ALIGNED_64); + ctx->picinfo.buf_h = ALIGN(ctx->picinfo.pic_h, VCODEC_DEC_ALIGNED_64); + ctx->picinfo.fb_sz[0] = instance->vpu.fb_sz[0]; + ctx->picinfo.fb_sz[1] = instance->vpu.fb_sz[1]; +} + +static inline void vdec_av1_slice_get_dpb_size(struct vdec_av1_slice_instance *instance, + u32 *dpb_sz) +{ + /* refer av1 specification */ + *dpb_sz = V4L2_AV1_TOTAL_REFS_PER_FRAME + 1; +} + +static void vdec_av1_slice_get_crop_info(struct vdec_av1_slice_instance *instance, + struct v4l2_rect *cr) +{ + struct mtk_vcodec_dec_ctx *ctx = instance->ctx; + + cr->left = 0; + cr->top = 0; + cr->width = ctx->picinfo.pic_w; + cr->height = ctx->picinfo.pic_h; + + mtk_vdec_debug(ctx, "l=%d, t=%d, w=%d, h=%d\n", + cr->left, cr->top, cr->width, cr->height); +} + +static int vdec_av1_slice_get_param(void *h_vdec, enum vdec_get_param_type type, void *out) +{ + struct vdec_av1_slice_instance *instance = h_vdec; + + switch (type) { + case GET_PARAM_PIC_INFO: + vdec_av1_slice_get_pic_info(instance); + break; + case GET_PARAM_DPB_SIZE: + vdec_av1_slice_get_dpb_size(instance, out); + break; + case GET_PARAM_CROP_INFO: + vdec_av1_slice_get_crop_info(instance, out); + break; + default: + mtk_vdec_err(instance->ctx, "invalid get parameter type=%d\n", type); + return -EINVAL; + } + + return 0; +} + +static int vdec_av1_slice_lat_decode(void *h_vdec, struct mtk_vcodec_mem *bs, + struct vdec_fb *fb, bool *res_chg) +{ + struct vdec_av1_slice_instance *instance = h_vdec; + struct vdec_lat_buf *lat_buf; + struct vdec_av1_slice_pfc *pfc; + struct vdec_av1_slice_vsi *vsi; + struct mtk_vcodec_dec_ctx *ctx; + int ret; + + if (!instance || !instance->ctx) + return -EINVAL; + + ctx = instance->ctx; + /* init msgQ for the first time */ + if (vdec_msg_queue_init(&ctx->msg_queue, ctx, + vdec_av1_slice_core_decode, sizeof(*pfc))) { + mtk_vdec_err(ctx, "failed to init AV1 msg queue\n"); + return -ENOMEM; + } + + /* bs NULL means flush decoder */ + if (!bs) + return vdec_av1_slice_flush(h_vdec, bs, fb, res_chg); + + lat_buf = vdec_msg_queue_dqbuf(&ctx->msg_queue.lat_ctx); + if (!lat_buf) { + mtk_vdec_err(ctx, "failed to get AV1 lat buf\n"); + return -EAGAIN; + } + pfc = (struct vdec_av1_slice_pfc *)lat_buf->private_data; + if (!pfc) { + ret = -EINVAL; + goto err_free_fb_out; + } + vsi = &pfc->vsi; + + ret = vdec_av1_slice_setup_lat(instance, bs, lat_buf, pfc); + if (ret) { + mtk_vdec_err(ctx, "failed to setup AV1 lat ret %d\n", ret); + goto err_free_fb_out; + } + + vdec_av1_slice_vsi_to_remote(vsi, instance->vsi); + ret = vpu_dec_start(&instance->vpu, NULL, 0); + if (ret) { + mtk_vdec_err(ctx, "failed to dec AV1 ret %d\n", ret); + goto err_free_fb_out; + } + if (instance->inneracing_mode) + vdec_msg_queue_qbuf(&ctx->msg_queue.core_ctx, lat_buf); + + if (instance->irq_enabled) { + ret = mtk_vcodec_wait_for_done_ctx(ctx, MTK_INST_IRQ_RECEIVED, + WAIT_INTR_TIMEOUT_MS, + MTK_VDEC_LAT0); + /* update remote vsi if decode timeout */ + if (ret) { + mtk_vdec_err(ctx, "AV1 Frame %d decode timeout %d\n", pfc->seq, ret); + WRITE_ONCE(instance->vsi->state.timeout, 1); + } + vpu_dec_end(&instance->vpu); + } + + vdec_av1_slice_vsi_from_remote(vsi, instance->vsi); + ret = vdec_av1_slice_update_lat(instance, lat_buf, pfc); + + /* LAT trans full, re-decode */ + if (ret == -EAGAIN) { + mtk_vdec_err(ctx, "AV1 Frame %d trans full\n", pfc->seq); + if (!instance->inneracing_mode) + vdec_msg_queue_qbuf(&ctx->msg_queue.lat_ctx, lat_buf); + return 0; + } + + /* LAT trans full, no more UBE or decode timeout */ + if (ret == -ENOMEM || vsi->state.timeout) { + mtk_vdec_err(ctx, "AV1 Frame %d insufficient buffer or timeout\n", pfc->seq); + if (!instance->inneracing_mode) + vdec_msg_queue_qbuf(&ctx->msg_queue.lat_ctx, lat_buf); + return -EBUSY; + } + vsi->trans.dma_addr_end += ctx->msg_queue.wdma_addr.dma_addr; + mtk_vdec_debug(ctx, "lat dma 1 0x%pad 0x%pad\n", + &pfc->vsi.trans.dma_addr, &pfc->vsi.trans.dma_addr_end); + + vdec_msg_queue_update_ube_wptr(&ctx->msg_queue, vsi->trans.dma_addr_end); + + if (!instance->inneracing_mode) + vdec_msg_queue_qbuf(&ctx->msg_queue.core_ctx, lat_buf); + memcpy(&instance->slots, &vsi->slots, sizeof(instance->slots)); + + return 0; + +err_free_fb_out: + vdec_msg_queue_qbuf(&ctx->msg_queue.lat_ctx, lat_buf); + + if (pfc) + mtk_vdec_err(ctx, "slice dec number: %d err: %d", pfc->seq, ret); + + return ret; +} + +static int vdec_av1_slice_core_decode(struct vdec_lat_buf *lat_buf) +{ + struct vdec_av1_slice_instance *instance; + struct vdec_av1_slice_pfc *pfc; + struct mtk_vcodec_dec_ctx *ctx = NULL; + struct vdec_fb *fb = NULL; + int ret = -EINVAL; + + if (!lat_buf) + return -EINVAL; + + pfc = lat_buf->private_data; + ctx = lat_buf->ctx; + if (!pfc || !ctx) + return -EINVAL; + + instance = ctx->drv_handle; + if (!instance) + goto err; + + fb = ctx->dev->vdec_pdata->get_cap_buffer(ctx); + if (!fb) { + ret = -EBUSY; + goto err; + } + + ret = vdec_av1_slice_setup_core(instance, fb, lat_buf, pfc); + if (ret) { + mtk_vdec_err(ctx, "vdec_av1_slice_setup_core\n"); + goto err; + } + vdec_av1_slice_vsi_to_remote(&pfc->vsi, instance->core_vsi); + ret = vpu_dec_core(&instance->vpu); + if (ret) { + mtk_vdec_err(ctx, "vpu_dec_core\n"); + goto err; + } + + if (instance->irq_enabled) { + ret = mtk_vcodec_wait_for_done_ctx(ctx, MTK_INST_IRQ_RECEIVED, + WAIT_INTR_TIMEOUT_MS, + MTK_VDEC_CORE); + /* update remote vsi if decode timeout */ + if (ret) { + mtk_vdec_err(ctx, "AV1 frame %d core timeout\n", pfc->seq); + WRITE_ONCE(instance->vsi->state.timeout, 1); + } + vpu_dec_core_end(&instance->vpu); + } + + ret = vdec_av1_slice_update_core(instance, lat_buf, pfc); + if (ret) { + mtk_vdec_err(ctx, "vdec_av1_slice_update_core\n"); + goto err; + } + + mtk_vdec_debug(ctx, "core dma_addr_end 0x%pad\n", + &instance->core_vsi->trans.dma_addr_end); + vdec_msg_queue_update_ube_rptr(&ctx->msg_queue, instance->core_vsi->trans.dma_addr_end); + + ctx->dev->vdec_pdata->cap_to_disp(ctx, 0, lat_buf->src_buf_req); + + return 0; + +err: + /* always update read pointer */ + vdec_msg_queue_update_ube_rptr(&ctx->msg_queue, pfc->vsi.trans.dma_addr_end); + + if (fb) + ctx->dev->vdec_pdata->cap_to_disp(ctx, 1, lat_buf->src_buf_req); + + return ret; +} + +const struct vdec_common_if vdec_av1_slice_lat_if = { + .init = vdec_av1_slice_init, + .decode = vdec_av1_slice_lat_decode, + .get_param = vdec_av1_slice_get_param, + .deinit = vdec_av1_slice_deinit, +}; diff --git a/drivers/media/platform/mediatek/vcodec/decoder/vdec/vdec_h264_if.c b/drivers/media/platform/mediatek/vcodec/decoder/vdec/vdec_h264_if.c new file mode 100644 index 0000000000..bf7dffe60d --- /dev/null +++ b/drivers/media/platform/mediatek/vcodec/decoder/vdec/vdec_h264_if.c @@ -0,0 +1,494 @@ +// SPDX-License-Identifier: GPL-2.0 +/* + * Copyright (c) 2016 MediaTek Inc. + * Author: PC Chen <pc.chen@mediatek.com> + */ + +#include <linux/module.h> +#include <linux/slab.h> + +#include "../vdec_drv_if.h" +#include "../mtk_vcodec_dec.h" +#include "../../common/mtk_vcodec_intr.h" +#include "../vdec_vpu_if.h" +#include "../vdec_drv_base.h" + +#define NAL_NON_IDR_SLICE 0x01 +#define NAL_IDR_SLICE 0x05 +#define NAL_H264_PPS 0x08 +#define NAL_TYPE(value) ((value) & 0x1F) + +#define BUF_PREDICTION_SZ (32 * 1024) + +#define MB_UNIT_LEN 16 + +/* motion vector size (bytes) for every macro block */ +#define HW_MB_STORE_SZ 64 + +#define H264_MAX_FB_NUM 17 +#define HDR_PARSING_BUF_SZ 1024 + +#define DEC_ERR_RET(ret) ((ret) >> 16) +#define H264_ERR_NOT_VALID 3 + +/** + * struct h264_fb - h264 decode frame buffer information + * @vdec_fb_va : virtual address of struct vdec_fb + * @y_fb_dma : dma address of Y frame buffer (luma) + * @c_fb_dma : dma address of C frame buffer (chroma) + * @poc : picture order count of frame buffer + * @reserved : for 8 bytes alignment + */ +struct h264_fb { + uint64_t vdec_fb_va; + uint64_t y_fb_dma; + uint64_t c_fb_dma; + int32_t poc; + uint32_t reserved; +}; + +/** + * struct h264_ring_fb_list - ring frame buffer list + * @fb_list : frame buffer array + * @read_idx : read index + * @write_idx : write index + * @count : buffer count in list + * @reserved : for 8 bytes alignment + */ +struct h264_ring_fb_list { + struct h264_fb fb_list[H264_MAX_FB_NUM]; + unsigned int read_idx; + unsigned int write_idx; + unsigned int count; + unsigned int reserved; +}; + +/** + * struct vdec_h264_dec_info - decode information + * @dpb_sz : decoding picture buffer size + * @resolution_changed : resolution change happen + * @realloc_mv_buf : flag to notify driver to re-allocate mv buffer + * @reserved : for 8 bytes alignment + * @bs_dma : Input bit-stream buffer dma address + * @y_fb_dma : Y frame buffer dma address + * @c_fb_dma : C frame buffer dma address + * @vdec_fb_va : VDEC frame buffer struct virtual address + */ +struct vdec_h264_dec_info { + uint32_t dpb_sz; + uint32_t resolution_changed; + uint32_t realloc_mv_buf; + uint32_t reserved; + uint64_t bs_dma; + uint64_t y_fb_dma; + uint64_t c_fb_dma; + uint64_t vdec_fb_va; +}; + +/** + * struct vdec_h264_vsi - shared memory for decode information exchange + * between VPU and Host. + * The memory is allocated by VPU then mapping to Host + * in vpu_dec_init() and freed in vpu_dec_deinit() + * by VPU. + * AP-W/R : AP is writer/reader on this item + * VPU-W/R: VPU is write/reader on this item + * @hdr_buf : Header parsing buffer (AP-W, VPU-R) + * @pred_buf_dma : HW working predication buffer dma address (AP-W, VPU-R) + * @mv_buf_dma : HW working motion vector buffer dma address (AP-W, VPU-R) + * @list_free : free frame buffer ring list (AP-W/R, VPU-W) + * @list_disp : display frame buffer ring list (AP-R, VPU-W) + * @dec : decode information (AP-R, VPU-W) + * @pic : picture information (AP-R, VPU-W) + * @crop : crop information (AP-R, VPU-W) + */ +struct vdec_h264_vsi { + unsigned char hdr_buf[HDR_PARSING_BUF_SZ]; + uint64_t pred_buf_dma; + uint64_t mv_buf_dma[H264_MAX_FB_NUM]; + struct h264_ring_fb_list list_free; + struct h264_ring_fb_list list_disp; + struct vdec_h264_dec_info dec; + struct vdec_pic_info pic; + struct v4l2_rect crop; +}; + +/** + * struct vdec_h264_inst - h264 decoder instance + * @num_nalu : how many nalus be decoded + * @ctx : point to mtk_vcodec_dec_ctx + * @pred_buf : HW working predication buffer + * @mv_buf : HW working motion vector buffer + * @vpu : VPU instance + * @vsi : VPU shared information + */ +struct vdec_h264_inst { + unsigned int num_nalu; + struct mtk_vcodec_dec_ctx *ctx; + struct mtk_vcodec_mem pred_buf; + struct mtk_vcodec_mem mv_buf[H264_MAX_FB_NUM]; + struct vdec_vpu_inst vpu; + struct vdec_h264_vsi *vsi; +}; + +static unsigned int get_mv_buf_size(unsigned int width, unsigned int height) +{ + return HW_MB_STORE_SZ * (width/MB_UNIT_LEN) * (height/MB_UNIT_LEN); +} + +static int allocate_predication_buf(struct vdec_h264_inst *inst) +{ + int err = 0; + + inst->pred_buf.size = BUF_PREDICTION_SZ; + err = mtk_vcodec_mem_alloc(inst->ctx, &inst->pred_buf); + if (err) { + mtk_vdec_err(inst->ctx, "failed to allocate ppl buf"); + return err; + } + + inst->vsi->pred_buf_dma = inst->pred_buf.dma_addr; + return 0; +} + +static void free_predication_buf(struct vdec_h264_inst *inst) +{ + struct mtk_vcodec_mem *mem = NULL; + + inst->vsi->pred_buf_dma = 0; + mem = &inst->pred_buf; + if (mem->va) + mtk_vcodec_mem_free(inst->ctx, mem); +} + +static int alloc_mv_buf(struct vdec_h264_inst *inst, struct vdec_pic_info *pic) +{ + int i; + int err; + struct mtk_vcodec_mem *mem = NULL; + unsigned int buf_sz = get_mv_buf_size(pic->buf_w, pic->buf_h); + + for (i = 0; i < H264_MAX_FB_NUM; i++) { + mem = &inst->mv_buf[i]; + if (mem->va) + mtk_vcodec_mem_free(inst->ctx, mem); + mem->size = buf_sz; + err = mtk_vcodec_mem_alloc(inst->ctx, mem); + if (err) { + mtk_vdec_err(inst->ctx, "failed to allocate mv buf"); + return err; + } + inst->vsi->mv_buf_dma[i] = mem->dma_addr; + } + + return 0; +} + +static void free_mv_buf(struct vdec_h264_inst *inst) +{ + int i; + struct mtk_vcodec_mem *mem = NULL; + + for (i = 0; i < H264_MAX_FB_NUM; i++) { + inst->vsi->mv_buf_dma[i] = 0; + mem = &inst->mv_buf[i]; + if (mem->va) + mtk_vcodec_mem_free(inst->ctx, mem); + } +} + +static int check_list_validity(struct vdec_h264_inst *inst, bool disp_list) +{ + struct h264_ring_fb_list *list; + + list = disp_list ? &inst->vsi->list_disp : &inst->vsi->list_free; + + if (list->count > H264_MAX_FB_NUM || + list->read_idx >= H264_MAX_FB_NUM || + list->write_idx >= H264_MAX_FB_NUM) { + mtk_vdec_err(inst->ctx, "%s list err: cnt=%d r_idx=%d w_idx=%d", + disp_list ? "disp" : "free", list->count, + list->read_idx, list->write_idx); + return -EINVAL; + } + + return 0; +} + +static void put_fb_to_free(struct vdec_h264_inst *inst, struct vdec_fb *fb) +{ + struct h264_ring_fb_list *list; + + if (fb) { + if (check_list_validity(inst, false)) + return; + + list = &inst->vsi->list_free; + if (list->count == H264_MAX_FB_NUM) { + mtk_vdec_err(inst->ctx, "[FB] put fb free_list full"); + return; + } + + mtk_vdec_debug(inst->ctx, "[FB] put fb into free_list @(%p, %llx)", + fb->base_y.va, (u64)fb->base_y.dma_addr); + + list->fb_list[list->write_idx].vdec_fb_va = (u64)(uintptr_t)fb; + list->write_idx = (list->write_idx == H264_MAX_FB_NUM - 1) ? + 0 : list->write_idx + 1; + list->count++; + } +} + +static void get_pic_info(struct vdec_h264_inst *inst, + struct vdec_pic_info *pic) +{ + *pic = inst->vsi->pic; + mtk_vdec_debug(inst->ctx, "pic(%d, %d), buf(%d, %d)", + pic->pic_w, pic->pic_h, pic->buf_w, pic->buf_h); + mtk_vdec_debug(inst->ctx, "fb size: Y(%d), C(%d)", pic->fb_sz[0], pic->fb_sz[1]); +} + +static void get_crop_info(struct vdec_h264_inst *inst, struct v4l2_rect *cr) +{ + cr->left = inst->vsi->crop.left; + cr->top = inst->vsi->crop.top; + cr->width = inst->vsi->crop.width; + cr->height = inst->vsi->crop.height; + + mtk_vdec_debug(inst->ctx, "l=%d, t=%d, w=%d, h=%d", cr->left, cr->top, + cr->width, cr->height); +} + +static void get_dpb_size(struct vdec_h264_inst *inst, unsigned int *dpb_sz) +{ + *dpb_sz = inst->vsi->dec.dpb_sz; + mtk_vdec_debug(inst->ctx, "sz=%d", *dpb_sz); +} + +static int vdec_h264_init(struct mtk_vcodec_dec_ctx *ctx) +{ + struct vdec_h264_inst *inst = NULL; + int err; + + inst = kzalloc(sizeof(*inst), GFP_KERNEL); + if (!inst) + return -ENOMEM; + + inst->ctx = ctx; + + inst->vpu.id = IPI_VDEC_H264; + inst->vpu.ctx = ctx; + + err = vpu_dec_init(&inst->vpu); + if (err) { + mtk_vdec_err(ctx, "vdec_h264 init err=%d", err); + goto error_free_inst; + } + + inst->vsi = (struct vdec_h264_vsi *)inst->vpu.vsi; + err = allocate_predication_buf(inst); + if (err) + goto error_deinit; + + mtk_vdec_debug(ctx, "H264 Instance >> %p", inst); + + ctx->drv_handle = inst; + return 0; + +error_deinit: + vpu_dec_deinit(&inst->vpu); + +error_free_inst: + kfree(inst); + return err; +} + +static void vdec_h264_deinit(void *h_vdec) +{ + struct vdec_h264_inst *inst = (struct vdec_h264_inst *)h_vdec; + + vpu_dec_deinit(&inst->vpu); + free_predication_buf(inst); + free_mv_buf(inst); + + kfree(inst); +} + +static int find_start_code(unsigned char *data, unsigned int data_sz) +{ + if (data_sz > 3 && data[0] == 0 && data[1] == 0 && data[2] == 1) + return 3; + + if (data_sz > 4 && data[0] == 0 && data[1] == 0 && data[2] == 0 && + data[3] == 1) + return 4; + + return -1; +} + +static int vdec_h264_decode(void *h_vdec, struct mtk_vcodec_mem *bs, + struct vdec_fb *fb, bool *res_chg) +{ + struct vdec_h264_inst *inst = (struct vdec_h264_inst *)h_vdec; + struct vdec_vpu_inst *vpu = &inst->vpu; + int nal_start_idx = 0; + int err = 0; + unsigned int nal_start; + unsigned int nal_type; + unsigned char *buf; + unsigned int buf_sz; + unsigned int data[2]; + uint64_t vdec_fb_va = (u64)(uintptr_t)fb; + uint64_t y_fb_dma = fb ? (u64)fb->base_y.dma_addr : 0; + uint64_t c_fb_dma = fb ? (u64)fb->base_c.dma_addr : 0; + + mtk_vdec_debug(inst->ctx, "+ [%d] FB y_dma=%llx c_dma=%llx va=%p", + ++inst->num_nalu, y_fb_dma, c_fb_dma, fb); + + /* bs NULL means flush decoder */ + if (bs == NULL) + return vpu_dec_reset(vpu); + + buf = (unsigned char *)bs->va; + buf_sz = bs->size; + nal_start_idx = find_start_code(buf, buf_sz); + if (nal_start_idx < 0) { + mtk_vdec_err(inst->ctx, "invalid nal start code"); + err = -EIO; + goto err_free_fb_out; + } + + nal_start = buf[nal_start_idx]; + nal_type = NAL_TYPE(buf[nal_start_idx]); + mtk_vdec_debug(inst->ctx, "\n + NALU[%d] type %d +\n", inst->num_nalu, + nal_type); + + if (nal_type == NAL_H264_PPS) { + buf_sz -= nal_start_idx; + if (buf_sz > HDR_PARSING_BUF_SZ) { + err = -EILSEQ; + goto err_free_fb_out; + } + memcpy(inst->vsi->hdr_buf, buf + nal_start_idx, buf_sz); + } + + inst->vsi->dec.bs_dma = (uint64_t)bs->dma_addr; + inst->vsi->dec.y_fb_dma = y_fb_dma; + inst->vsi->dec.c_fb_dma = c_fb_dma; + inst->vsi->dec.vdec_fb_va = vdec_fb_va; + + data[0] = buf_sz; + data[1] = nal_start; + err = vpu_dec_start(vpu, data, 2); + if (err) { + if (err > 0 && (DEC_ERR_RET(err) == H264_ERR_NOT_VALID)) { + mtk_vdec_err(inst->ctx, "- error bitstream - err = %d -", err); + err = -EIO; + } + goto err_free_fb_out; + } + + *res_chg = inst->vsi->dec.resolution_changed; + if (*res_chg) { + struct vdec_pic_info pic; + + mtk_vdec_debug(inst->ctx, "- resolution changed -"); + get_pic_info(inst, &pic); + + if (inst->vsi->dec.realloc_mv_buf) { + err = alloc_mv_buf(inst, &pic); + if (err) + goto err_free_fb_out; + } + } + + if (nal_type == NAL_NON_IDR_SLICE || nal_type == NAL_IDR_SLICE) { + /* wait decoder done interrupt */ + err = mtk_vcodec_wait_for_done_ctx(inst->ctx, + MTK_INST_IRQ_RECEIVED, + WAIT_INTR_TIMEOUT_MS, 0); + if (err) + goto err_free_fb_out; + + vpu_dec_end(vpu); + } + + mtk_vdec_debug(inst->ctx, "\n - NALU[%d] type=%d -\n", inst->num_nalu, nal_type); + return 0; + +err_free_fb_out: + put_fb_to_free(inst, fb); + mtk_vdec_err(inst->ctx, "\n - NALU[%d] err=%d -\n", inst->num_nalu, err); + return err; +} + +static void vdec_h264_get_fb(struct vdec_h264_inst *inst, + struct h264_ring_fb_list *list, + bool disp_list, struct vdec_fb **out_fb) +{ + struct vdec_fb *fb; + + if (check_list_validity(inst, disp_list)) + return; + + if (list->count == 0) { + mtk_vdec_debug(inst->ctx, "[FB] there is no %s fb", disp_list ? "disp" : "free"); + *out_fb = NULL; + return; + } + + fb = (struct vdec_fb *) + (uintptr_t)list->fb_list[list->read_idx].vdec_fb_va; + fb->status |= (disp_list ? FB_ST_DISPLAY : FB_ST_FREE); + + *out_fb = fb; + mtk_vdec_debug(inst->ctx, "[FB] get %s fb st=%d poc=%d %llx", + disp_list ? "disp" : "free", + fb->status, list->fb_list[list->read_idx].poc, + list->fb_list[list->read_idx].vdec_fb_va); + + list->read_idx = (list->read_idx == H264_MAX_FB_NUM - 1) ? + 0 : list->read_idx + 1; + list->count--; +} + +static int vdec_h264_get_param(void *h_vdec, enum vdec_get_param_type type, + void *out) +{ + struct vdec_h264_inst *inst = (struct vdec_h264_inst *)h_vdec; + + switch (type) { + case GET_PARAM_DISP_FRAME_BUFFER: + vdec_h264_get_fb(inst, &inst->vsi->list_disp, true, out); + break; + + case GET_PARAM_FREE_FRAME_BUFFER: + vdec_h264_get_fb(inst, &inst->vsi->list_free, false, out); + break; + + case GET_PARAM_PIC_INFO: + get_pic_info(inst, out); + break; + + case GET_PARAM_DPB_SIZE: + get_dpb_size(inst, out); + break; + + case GET_PARAM_CROP_INFO: + get_crop_info(inst, out); + break; + + default: + mtk_vdec_err(inst->ctx, "invalid get parameter type=%d", type); + return -EINVAL; + } + + return 0; +} + +const struct vdec_common_if vdec_h264_if = { + .init = vdec_h264_init, + .decode = vdec_h264_decode, + .get_param = vdec_h264_get_param, + .deinit = vdec_h264_deinit, +}; diff --git a/drivers/media/platform/mediatek/vcodec/decoder/vdec/vdec_h264_req_common.c b/drivers/media/platform/mediatek/vcodec/decoder/vdec/vdec_h264_req_common.c new file mode 100644 index 0000000000..5ca20d75dc --- /dev/null +++ b/drivers/media/platform/mediatek/vcodec/decoder/vdec/vdec_h264_req_common.c @@ -0,0 +1,322 @@ +// SPDX-License-Identifier: GPL-2.0 +/* + * Copyright (c) 2022 MediaTek Inc. + * Author: Yunfei Dong <yunfei.dong@mediatek.com> + */ + +#include "vdec_h264_req_common.h" + +/* get used parameters for sps/pps */ +#define GET_MTK_VDEC_FLAG(cond, flag) \ + { dst_param->cond = ((src_param->flags & flag) ? (1) : (0)); } +#define GET_MTK_VDEC_PARAM(param) \ + { dst_param->param = src_param->param; } + +void mtk_vdec_h264_get_ref_list(u8 *ref_list, + const struct v4l2_h264_reference *v4l2_ref_list, + int num_valid) +{ + u32 i; + + /* + * TODO The firmware does not support field decoding. Future + * implementation must use v4l2_ref_list[i].fields to obtain + * the reference field parity. + */ + + for (i = 0; i < num_valid; i++) + ref_list[i] = v4l2_ref_list[i].index; + + /* + * The firmware expects unused reflist entries to have the value 0x20. + */ + memset(&ref_list[num_valid], 0x20, 32 - num_valid); +} + +void *mtk_vdec_h264_get_ctrl_ptr(struct mtk_vcodec_dec_ctx *ctx, int id) +{ + struct v4l2_ctrl *ctrl = v4l2_ctrl_find(&ctx->ctrl_hdl, id); + + if (!ctrl) + return ERR_PTR(-EINVAL); + + return ctrl->p_cur.p; +} + +void mtk_vdec_h264_fill_dpb_info(struct mtk_vcodec_dec_ctx *ctx, + struct slice_api_h264_decode_param *decode_params, + struct mtk_h264_dpb_info *h264_dpb_info) +{ + const struct slice_h264_dpb_entry *dpb; + struct vb2_queue *vq; + struct vb2_buffer *vb; + struct vb2_v4l2_buffer *vb2_v4l2; + int index; + + vq = v4l2_m2m_get_vq(ctx->m2m_ctx, V4L2_BUF_TYPE_VIDEO_CAPTURE_MPLANE); + + for (index = 0; index < V4L2_H264_NUM_DPB_ENTRIES; index++) { + dpb = &decode_params->dpb[index]; + if (!(dpb->flags & V4L2_H264_DPB_ENTRY_FLAG_ACTIVE)) { + h264_dpb_info[index].reference_flag = 0; + continue; + } + + vb = vb2_find_buffer(vq, dpb->reference_ts); + if (!vb) { + dev_err(&ctx->dev->plat_dev->dev, + "Reference invalid: dpb_index(%d) reference_ts(%lld)", + index, dpb->reference_ts); + continue; + } + + /* 1 for short term reference, 2 for long term reference */ + if (!(dpb->flags & V4L2_H264_DPB_ENTRY_FLAG_LONG_TERM)) + h264_dpb_info[index].reference_flag = 1; + else + h264_dpb_info[index].reference_flag = 2; + + vb2_v4l2 = container_of(vb, struct vb2_v4l2_buffer, vb2_buf); + h264_dpb_info[index].field = vb2_v4l2->field; + + h264_dpb_info[index].y_dma_addr = + vb2_dma_contig_plane_dma_addr(vb, 0); + if (ctx->q_data[MTK_Q_DATA_DST].fmt->num_planes == 2) + h264_dpb_info[index].c_dma_addr = + vb2_dma_contig_plane_dma_addr(vb, 1); + else + h264_dpb_info[index].c_dma_addr = + h264_dpb_info[index].y_dma_addr + + ctx->picinfo.fb_sz[0]; + } +} + +void mtk_vdec_h264_copy_sps_params(struct mtk_h264_sps_param *dst_param, + const struct v4l2_ctrl_h264_sps *src_param) +{ + GET_MTK_VDEC_PARAM(chroma_format_idc); + GET_MTK_VDEC_PARAM(bit_depth_luma_minus8); + GET_MTK_VDEC_PARAM(bit_depth_chroma_minus8); + GET_MTK_VDEC_PARAM(log2_max_frame_num_minus4); + GET_MTK_VDEC_PARAM(pic_order_cnt_type); + GET_MTK_VDEC_PARAM(log2_max_pic_order_cnt_lsb_minus4); + GET_MTK_VDEC_PARAM(max_num_ref_frames); + GET_MTK_VDEC_PARAM(pic_width_in_mbs_minus1); + GET_MTK_VDEC_PARAM(pic_height_in_map_units_minus1); + + GET_MTK_VDEC_FLAG(separate_colour_plane_flag, + V4L2_H264_SPS_FLAG_SEPARATE_COLOUR_PLANE); + GET_MTK_VDEC_FLAG(qpprime_y_zero_transform_bypass_flag, + V4L2_H264_SPS_FLAG_QPPRIME_Y_ZERO_TRANSFORM_BYPASS); + GET_MTK_VDEC_FLAG(delta_pic_order_always_zero_flag, + V4L2_H264_SPS_FLAG_DELTA_PIC_ORDER_ALWAYS_ZERO); + GET_MTK_VDEC_FLAG(frame_mbs_only_flag, + V4L2_H264_SPS_FLAG_FRAME_MBS_ONLY); + GET_MTK_VDEC_FLAG(mb_adaptive_frame_field_flag, + V4L2_H264_SPS_FLAG_MB_ADAPTIVE_FRAME_FIELD); + GET_MTK_VDEC_FLAG(direct_8x8_inference_flag, + V4L2_H264_SPS_FLAG_DIRECT_8X8_INFERENCE); +} + +void mtk_vdec_h264_copy_pps_params(struct mtk_h264_pps_param *dst_param, + const struct v4l2_ctrl_h264_pps *src_param) +{ + GET_MTK_VDEC_PARAM(num_ref_idx_l0_default_active_minus1); + GET_MTK_VDEC_PARAM(num_ref_idx_l1_default_active_minus1); + GET_MTK_VDEC_PARAM(weighted_bipred_idc); + GET_MTK_VDEC_PARAM(pic_init_qp_minus26); + GET_MTK_VDEC_PARAM(chroma_qp_index_offset); + GET_MTK_VDEC_PARAM(second_chroma_qp_index_offset); + + GET_MTK_VDEC_FLAG(entropy_coding_mode_flag, + V4L2_H264_PPS_FLAG_ENTROPY_CODING_MODE); + GET_MTK_VDEC_FLAG(pic_order_present_flag, + V4L2_H264_PPS_FLAG_BOTTOM_FIELD_PIC_ORDER_IN_FRAME_PRESENT); + GET_MTK_VDEC_FLAG(weighted_pred_flag, + V4L2_H264_PPS_FLAG_WEIGHTED_PRED); + GET_MTK_VDEC_FLAG(deblocking_filter_control_present_flag, + V4L2_H264_PPS_FLAG_DEBLOCKING_FILTER_CONTROL_PRESENT); + GET_MTK_VDEC_FLAG(constrained_intra_pred_flag, + V4L2_H264_PPS_FLAG_CONSTRAINED_INTRA_PRED); + GET_MTK_VDEC_FLAG(redundant_pic_cnt_present_flag, + V4L2_H264_PPS_FLAG_REDUNDANT_PIC_CNT_PRESENT); + GET_MTK_VDEC_FLAG(transform_8x8_mode_flag, + V4L2_H264_PPS_FLAG_TRANSFORM_8X8_MODE); + GET_MTK_VDEC_FLAG(scaling_matrix_present_flag, + V4L2_H264_PPS_FLAG_SCALING_MATRIX_PRESENT); +} + +void mtk_vdec_h264_copy_slice_hd_params(struct mtk_h264_slice_hd_param *dst_param, + const struct v4l2_ctrl_h264_slice_params *src_param, + const struct v4l2_ctrl_h264_decode_params *dec_param) +{ + int temp; + + GET_MTK_VDEC_PARAM(first_mb_in_slice); + GET_MTK_VDEC_PARAM(slice_type); + GET_MTK_VDEC_PARAM(cabac_init_idc); + GET_MTK_VDEC_PARAM(slice_qp_delta); + GET_MTK_VDEC_PARAM(disable_deblocking_filter_idc); + GET_MTK_VDEC_PARAM(slice_alpha_c0_offset_div2); + GET_MTK_VDEC_PARAM(slice_beta_offset_div2); + GET_MTK_VDEC_PARAM(num_ref_idx_l0_active_minus1); + GET_MTK_VDEC_PARAM(num_ref_idx_l1_active_minus1); + + dst_param->frame_num = dec_param->frame_num; + dst_param->pic_order_cnt_lsb = dec_param->pic_order_cnt_lsb; + + dst_param->delta_pic_order_cnt_bottom = + dec_param->delta_pic_order_cnt_bottom; + dst_param->delta_pic_order_cnt0 = + dec_param->delta_pic_order_cnt0; + dst_param->delta_pic_order_cnt1 = + dec_param->delta_pic_order_cnt1; + + temp = dec_param->flags & V4L2_H264_DECODE_PARAM_FLAG_FIELD_PIC; + dst_param->field_pic_flag = temp ? 1 : 0; + + temp = dec_param->flags & V4L2_H264_DECODE_PARAM_FLAG_BOTTOM_FIELD; + dst_param->bottom_field_flag = temp ? 1 : 0; + + GET_MTK_VDEC_FLAG(direct_spatial_mv_pred_flag, + V4L2_H264_SLICE_FLAG_DIRECT_SPATIAL_MV_PRED); +} + +void mtk_vdec_h264_copy_scaling_matrix(struct slice_api_h264_scaling_matrix *dst_matrix, + const struct v4l2_ctrl_h264_scaling_matrix *src_matrix) +{ + memcpy(dst_matrix->scaling_list_4x4, src_matrix->scaling_list_4x4, + sizeof(dst_matrix->scaling_list_4x4)); + + memcpy(dst_matrix->scaling_list_8x8, src_matrix->scaling_list_8x8, + sizeof(dst_matrix->scaling_list_8x8)); +} + +void +mtk_vdec_h264_copy_decode_params(struct slice_api_h264_decode_param *dst_params, + const struct v4l2_ctrl_h264_decode_params *src_params, + const struct v4l2_h264_dpb_entry dpb[V4L2_H264_NUM_DPB_ENTRIES]) +{ + struct slice_h264_dpb_entry *dst_entry; + const struct v4l2_h264_dpb_entry *src_entry; + int i; + + for (i = 0; i < ARRAY_SIZE(dst_params->dpb); i++) { + dst_entry = &dst_params->dpb[i]; + src_entry = &dpb[i]; + + dst_entry->reference_ts = src_entry->reference_ts; + dst_entry->frame_num = src_entry->frame_num; + dst_entry->pic_num = src_entry->pic_num; + dst_entry->top_field_order_cnt = src_entry->top_field_order_cnt; + dst_entry->bottom_field_order_cnt = + src_entry->bottom_field_order_cnt; + dst_entry->flags = src_entry->flags; + } + + /* num_slices is a leftover from the old H.264 support and is ignored + * by the firmware. + */ + dst_params->num_slices = 0; + dst_params->nal_ref_idc = src_params->nal_ref_idc; + dst_params->top_field_order_cnt = src_params->top_field_order_cnt; + dst_params->bottom_field_order_cnt = src_params->bottom_field_order_cnt; + dst_params->flags = src_params->flags; +} + +static bool mtk_vdec_h264_dpb_entry_match(const struct v4l2_h264_dpb_entry *a, + const struct v4l2_h264_dpb_entry *b) +{ + return a->top_field_order_cnt == b->top_field_order_cnt && + a->bottom_field_order_cnt == b->bottom_field_order_cnt; +} + +/* + * Move DPB entries of dec_param that refer to a frame already existing in dpb + * into the already existing slot in dpb, and move other entries into new slots. + * + * This function is an adaptation of the similarly-named function in + * hantro_h264.c. + */ +void mtk_vdec_h264_update_dpb(const struct v4l2_ctrl_h264_decode_params *dec_param, + struct v4l2_h264_dpb_entry *dpb) +{ + DECLARE_BITMAP(new, ARRAY_SIZE(dec_param->dpb)) = { 0, }; + DECLARE_BITMAP(in_use, ARRAY_SIZE(dec_param->dpb)) = { 0, }; + DECLARE_BITMAP(used, ARRAY_SIZE(dec_param->dpb)) = { 0, }; + unsigned int i, j; + + /* Disable all entries by default, and mark the ones in use. */ + for (i = 0; i < ARRAY_SIZE(dec_param->dpb); i++) { + if (dpb[i].flags & V4L2_H264_DPB_ENTRY_FLAG_ACTIVE) + set_bit(i, in_use); + dpb[i].flags &= ~V4L2_H264_DPB_ENTRY_FLAG_ACTIVE; + } + + /* Try to match new DPB entries with existing ones by their POCs. */ + for (i = 0; i < ARRAY_SIZE(dec_param->dpb); i++) { + const struct v4l2_h264_dpb_entry *ndpb = &dec_param->dpb[i]; + + if (!(ndpb->flags & V4L2_H264_DPB_ENTRY_FLAG_ACTIVE)) + continue; + + /* + * To cut off some comparisons, iterate only on target DPB + * entries were already used. + */ + for_each_set_bit(j, in_use, ARRAY_SIZE(dec_param->dpb)) { + struct v4l2_h264_dpb_entry *cdpb; + + cdpb = &dpb[j]; + if (!mtk_vdec_h264_dpb_entry_match(cdpb, ndpb)) + continue; + + *cdpb = *ndpb; + set_bit(j, used); + /* Don't reiterate on this one. */ + clear_bit(j, in_use); + break; + } + + if (j == ARRAY_SIZE(dec_param->dpb)) + set_bit(i, new); + } + + /* For entries that could not be matched, use remaining free slots. */ + for_each_set_bit(i, new, ARRAY_SIZE(dec_param->dpb)) { + const struct v4l2_h264_dpb_entry *ndpb = &dec_param->dpb[i]; + struct v4l2_h264_dpb_entry *cdpb; + + /* + * Both arrays are of the same sizes, so there is no way + * we can end up with no space in target array, unless + * something is buggy. + */ + j = find_first_zero_bit(used, ARRAY_SIZE(dec_param->dpb)); + if (WARN_ON(j >= ARRAY_SIZE(dec_param->dpb))) + return; + + cdpb = &dpb[j]; + *cdpb = *ndpb; + set_bit(j, used); + } +} + +unsigned int mtk_vdec_h264_get_mv_buf_size(unsigned int width, unsigned int height) +{ + int unit_size = (width / MB_UNIT_LEN) * (height / MB_UNIT_LEN) + 8; + + return HW_MB_STORE_SZ * unit_size; +} + +int mtk_vdec_h264_find_start_code(unsigned char *data, unsigned int data_sz) +{ + if (data_sz > 3 && data[0] == 0 && data[1] == 0 && data[2] == 1) + return 3; + + if (data_sz > 4 && data[0] == 0 && data[1] == 0 && data[2] == 0 && + data[3] == 1) + return 4; + + return -1; +} diff --git a/drivers/media/platform/mediatek/vcodec/decoder/vdec/vdec_h264_req_common.h b/drivers/media/platform/mediatek/vcodec/decoder/vdec/vdec_h264_req_common.h new file mode 100644 index 0000000000..ac82be3360 --- /dev/null +++ b/drivers/media/platform/mediatek/vcodec/decoder/vdec/vdec_h264_req_common.h @@ -0,0 +1,277 @@ +/* SPDX-License-Identifier: GPL-2.0 */ +/* + * Copyright (c) 2022 MediaTek Inc. + * Author: Yunfei Dong <yunfei.dong@mediatek.com> + */ + +#ifndef _VDEC_H264_REQ_COMMON_H_ +#define _VDEC_H264_REQ_COMMON_H_ + +#include <linux/module.h> +#include <linux/slab.h> +#include <media/v4l2-h264.h> +#include <media/v4l2-mem2mem.h> +#include <media/videobuf2-dma-contig.h> + +#include "../mtk_vcodec_dec_drv.h" + +#define NAL_NON_IDR_SLICE 0x01 +#define NAL_IDR_SLICE 0x05 +#define NAL_TYPE(value) ((value) & 0x1F) + +#define BUF_PREDICTION_SZ (64 * 4096) +#define MB_UNIT_LEN 16 + +/* motion vector size (bytes) for every macro block */ +#define HW_MB_STORE_SZ 64 + +#define H264_MAX_MV_NUM 32 + +/** + * struct mtk_h264_dpb_info - h264 dpb information + * + * @y_dma_addr: Y bitstream physical address + * @c_dma_addr: CbCr bitstream physical address + * @reference_flag: reference picture flag (short/long term reference picture) + * @field: field picture flag + */ +struct mtk_h264_dpb_info { + dma_addr_t y_dma_addr; + dma_addr_t c_dma_addr; + int reference_flag; + int field; +}; + +/* + * struct mtk_h264_sps_param - parameters for sps + */ +struct mtk_h264_sps_param { + unsigned char chroma_format_idc; + unsigned char bit_depth_luma_minus8; + unsigned char bit_depth_chroma_minus8; + unsigned char log2_max_frame_num_minus4; + unsigned char pic_order_cnt_type; + unsigned char log2_max_pic_order_cnt_lsb_minus4; + unsigned char max_num_ref_frames; + unsigned char separate_colour_plane_flag; + unsigned short pic_width_in_mbs_minus1; + unsigned short pic_height_in_map_units_minus1; + unsigned int max_frame_nums; + unsigned char qpprime_y_zero_transform_bypass_flag; + unsigned char delta_pic_order_always_zero_flag; + unsigned char frame_mbs_only_flag; + unsigned char mb_adaptive_frame_field_flag; + unsigned char direct_8x8_inference_flag; + unsigned char reserved[3]; +}; + +/* + * struct mtk_h264_pps_param - parameters for pps + */ +struct mtk_h264_pps_param { + unsigned char num_ref_idx_l0_default_active_minus1; + unsigned char num_ref_idx_l1_default_active_minus1; + unsigned char weighted_bipred_idc; + char pic_init_qp_minus26; + char chroma_qp_index_offset; + char second_chroma_qp_index_offset; + unsigned char entropy_coding_mode_flag; + unsigned char pic_order_present_flag; + unsigned char deblocking_filter_control_present_flag; + unsigned char constrained_intra_pred_flag; + unsigned char weighted_pred_flag; + unsigned char redundant_pic_cnt_present_flag; + unsigned char transform_8x8_mode_flag; + unsigned char scaling_matrix_present_flag; + unsigned char reserved[2]; +}; + +/* + * struct mtk_h264_slice_hd_param - parameters for slice header + */ +struct mtk_h264_slice_hd_param { + unsigned int first_mb_in_slice; + unsigned int field_pic_flag; + unsigned int slice_type; + unsigned int frame_num; + int pic_order_cnt_lsb; + int delta_pic_order_cnt_bottom; + unsigned int bottom_field_flag; + unsigned int direct_spatial_mv_pred_flag; + int delta_pic_order_cnt0; + int delta_pic_order_cnt1; + unsigned int cabac_init_idc; + int slice_qp_delta; + unsigned int disable_deblocking_filter_idc; + int slice_alpha_c0_offset_div2; + int slice_beta_offset_div2; + unsigned int num_ref_idx_l0_active_minus1; + unsigned int num_ref_idx_l1_active_minus1; + unsigned int reserved; +}; + +/* + * struct slice_api_h264_scaling_matrix - parameters for scaling list + */ +struct slice_api_h264_scaling_matrix { + unsigned char scaling_list_4x4[6][16]; + unsigned char scaling_list_8x8[6][64]; +}; + +/* + * struct slice_h264_dpb_entry - each dpb information + */ +struct slice_h264_dpb_entry { + unsigned long long reference_ts; + unsigned short frame_num; + unsigned short pic_num; + /* Note that field is indicated by v4l2_buffer.field */ + int top_field_order_cnt; + int bottom_field_order_cnt; + unsigned int flags; +}; + +/* + * struct slice_api_h264_decode_param - parameters for decode. + */ +struct slice_api_h264_decode_param { + struct slice_h264_dpb_entry dpb[V4L2_H264_NUM_DPB_ENTRIES]; + unsigned short num_slices; + unsigned short nal_ref_idc; + unsigned char ref_pic_list_p0[32]; + unsigned char ref_pic_list_b0[32]; + unsigned char ref_pic_list_b1[32]; + int top_field_order_cnt; + int bottom_field_order_cnt; + unsigned int flags; +}; + +/** + * struct h264_fb - h264 decode frame buffer information + * + * @vdec_fb_va: virtual address of struct vdec_fb + * @y_fb_dma: dma address of Y frame buffer (luma) + * @c_fb_dma: dma address of C frame buffer (chroma) + * @poc: picture order count of frame buffer + * @reserved: for 8 bytes alignment + */ +struct h264_fb { + u64 vdec_fb_va; + u64 y_fb_dma; + u64 c_fb_dma; + s32 poc; + u32 reserved; +}; + +/** + * mtk_vdec_h264_get_ref_list - translate V4L2 reference list + * + * @ref_list: Mediatek reference picture list + * @v4l2_ref_list: V4L2 reference picture list + * @num_valid: used reference number + */ +void mtk_vdec_h264_get_ref_list(u8 *ref_list, + const struct v4l2_h264_reference *v4l2_ref_list, + int num_valid); + +/** + * mtk_vdec_h264_get_ctrl_ptr - get each CID contrl address. + * + * @ctx: v4l2 ctx + * @id: CID control ID + * + * Return: returns CID ctrl address. + */ +void *mtk_vdec_h264_get_ctrl_ptr(struct mtk_vcodec_dec_ctx *ctx, int id); + +/** + * mtk_vdec_h264_fill_dpb_info - get each CID contrl address. + * + * @ctx: v4l2 ctx + * @decode_params: slice decode params + * @h264_dpb_info: dpb buffer information + */ +void mtk_vdec_h264_fill_dpb_info(struct mtk_vcodec_dec_ctx *ctx, + struct slice_api_h264_decode_param *decode_params, + struct mtk_h264_dpb_info *h264_dpb_info); + +/** + * mtk_vdec_h264_copy_sps_params - get sps params. + * + * @dst_param: sps params for hw decoder + * @src_param: sps params from user driver + */ +void mtk_vdec_h264_copy_sps_params(struct mtk_h264_sps_param *dst_param, + const struct v4l2_ctrl_h264_sps *src_param); + +/** + * mtk_vdec_h264_copy_pps_params - get pps params. + * + * @dst_param: pps params for hw decoder + * @src_param: pps params from user driver + */ +void mtk_vdec_h264_copy_pps_params(struct mtk_h264_pps_param *dst_param, + const struct v4l2_ctrl_h264_pps *src_param); + +/** + * mtk_vdec_h264_copy_slice_hd_params - get slice header params. + * + * @dst_param: slice params for hw decoder + * @src_param: slice params from user driver + * @dec_param: decode params from user driver + */ +void mtk_vdec_h264_copy_slice_hd_params(struct mtk_h264_slice_hd_param *dst_param, + const struct v4l2_ctrl_h264_slice_params *src_param, + const struct v4l2_ctrl_h264_decode_params *dec_param); + +/** + * mtk_vdec_h264_copy_scaling_matrix - get each CID contrl address. + * + * @dst_matrix: scaling list params for hw decoder + * @src_matrix: scaling list params from user driver + */ +void mtk_vdec_h264_copy_scaling_matrix(struct slice_api_h264_scaling_matrix *dst_matrix, + const struct v4l2_ctrl_h264_scaling_matrix *src_matrix); + +/** + * mtk_vdec_h264_copy_decode_params - get decode params. + * + * @dst_params: dst params for hw decoder + * @src_params: decode params from user driver + * @dpb: dpb information + */ +void +mtk_vdec_h264_copy_decode_params(struct slice_api_h264_decode_param *dst_params, + const struct v4l2_ctrl_h264_decode_params *src_params, + const struct v4l2_h264_dpb_entry dpb[V4L2_H264_NUM_DPB_ENTRIES]); + +/** + * mtk_vdec_h264_update_dpb - updata dpb list. + * + * @dec_param: v4l2 control decode params + * @dpb: dpb entry informaton + */ +void mtk_vdec_h264_update_dpb(const struct v4l2_ctrl_h264_decode_params *dec_param, + struct v4l2_h264_dpb_entry *dpb); + +/** + * mtk_vdec_h264_find_start_code - find h264 start code using sofeware. + * + * @data: input buffer address + * @data_sz: input buffer size + * + * Return: returns start code position. + */ +int mtk_vdec_h264_find_start_code(unsigned char *data, unsigned int data_sz); + +/** + * mtk_vdec_h264_get_mv_buf_size - get mv buffer size. + * + * @width: picture width + * @height: picture height + * + * Return: returns mv buffer size. + */ +unsigned int mtk_vdec_h264_get_mv_buf_size(unsigned int width, unsigned int height); + +#endif diff --git a/drivers/media/platform/mediatek/vcodec/decoder/vdec/vdec_h264_req_if.c b/drivers/media/platform/mediatek/vcodec/decoder/vdec/vdec_h264_req_if.c new file mode 100644 index 0000000000..5600f1df65 --- /dev/null +++ b/drivers/media/platform/mediatek/vcodec/decoder/vdec/vdec_h264_req_if.c @@ -0,0 +1,444 @@ +// SPDX-License-Identifier: GPL-2.0 + +#include <linux/module.h> +#include <linux/slab.h> +#include <media/v4l2-mem2mem.h> +#include <media/v4l2-h264.h> +#include <media/videobuf2-dma-contig.h> + +#include "../mtk_vcodec_dec.h" +#include "../../common/mtk_vcodec_intr.h" +#include "../vdec_drv_base.h" +#include "../vdec_drv_if.h" +#include "../vdec_vpu_if.h" +#include "vdec_h264_req_common.h" + +/* + * struct mtk_h264_dec_slice_param - parameters for decode current frame + */ +struct mtk_h264_dec_slice_param { + struct mtk_h264_sps_param sps; + struct mtk_h264_pps_param pps; + struct slice_api_h264_scaling_matrix scaling_matrix; + struct slice_api_h264_decode_param decode_params; + struct mtk_h264_dpb_info h264_dpb_info[16]; +}; + +/** + * struct vdec_h264_dec_info - decode information + * @dpb_sz : decoding picture buffer size + * @resolution_changed : resoltion change happen + * @realloc_mv_buf : flag to notify driver to re-allocate mv buffer + * @cap_num_planes : number planes of capture buffer + * @bs_dma : Input bit-stream buffer dma address + * @y_fb_dma : Y frame buffer dma address + * @c_fb_dma : C frame buffer dma address + * @vdec_fb_va : VDEC frame buffer struct virtual address + */ +struct vdec_h264_dec_info { + u32 dpb_sz; + u32 resolution_changed; + u32 realloc_mv_buf; + u32 cap_num_planes; + u64 bs_dma; + u64 y_fb_dma; + u64 c_fb_dma; + u64 vdec_fb_va; +}; + +/** + * struct vdec_h264_vsi - shared memory for decode information exchange + * between VPU and Host. + * The memory is allocated by VPU then mapping to Host + * in vpu_dec_init() and freed in vpu_dec_deinit() + * by VPU. + * AP-W/R : AP is writer/reader on this item + * VPU-W/R: VPU is write/reader on this item + * @pred_buf_dma : HW working predication buffer dma address (AP-W, VPU-R) + * @mv_buf_dma : HW working motion vector buffer dma address (AP-W, VPU-R) + * @dec : decode information (AP-R, VPU-W) + * @pic : picture information (AP-R, VPU-W) + * @crop : crop information (AP-R, VPU-W) + * @h264_slice_params : the parameters that hardware use to decode + */ +struct vdec_h264_vsi { + u64 pred_buf_dma; + u64 mv_buf_dma[H264_MAX_MV_NUM]; + struct vdec_h264_dec_info dec; + struct vdec_pic_info pic; + struct v4l2_rect crop; + struct mtk_h264_dec_slice_param h264_slice_params; +}; + +/** + * struct vdec_h264_slice_inst - h264 decoder instance + * @num_nalu : how many nalus be decoded + * @ctx : point to mtk_vcodec_dec_ctx + * @pred_buf : HW working predication buffer + * @mv_buf : HW working motion vector buffer + * @vpu : VPU instance + * @vsi_ctx : Local VSI data for this decoding context + * @h264_slice_param : the parameters that hardware use to decode + * @dpb : decoded picture buffer used to store reference buffer information + */ +struct vdec_h264_slice_inst { + unsigned int num_nalu; + struct mtk_vcodec_dec_ctx *ctx; + struct mtk_vcodec_mem pred_buf; + struct mtk_vcodec_mem mv_buf[H264_MAX_MV_NUM]; + struct vdec_vpu_inst vpu; + struct vdec_h264_vsi vsi_ctx; + struct mtk_h264_dec_slice_param h264_slice_param; + + struct v4l2_h264_dpb_entry dpb[16]; +}; + +static int get_vdec_decode_parameters(struct vdec_h264_slice_inst *inst) +{ + const struct v4l2_ctrl_h264_decode_params *dec_params; + const struct v4l2_ctrl_h264_sps *sps; + const struct v4l2_ctrl_h264_pps *pps; + const struct v4l2_ctrl_h264_scaling_matrix *scaling_matrix; + struct mtk_h264_dec_slice_param *slice_param = &inst->h264_slice_param; + struct v4l2_h264_reflist_builder reflist_builder; + struct v4l2_h264_reference v4l2_p0_reflist[V4L2_H264_REF_LIST_LEN]; + struct v4l2_h264_reference v4l2_b0_reflist[V4L2_H264_REF_LIST_LEN]; + struct v4l2_h264_reference v4l2_b1_reflist[V4L2_H264_REF_LIST_LEN]; + u8 *p0_reflist = slice_param->decode_params.ref_pic_list_p0; + u8 *b0_reflist = slice_param->decode_params.ref_pic_list_b0; + u8 *b1_reflist = slice_param->decode_params.ref_pic_list_b1; + + dec_params = + mtk_vdec_h264_get_ctrl_ptr(inst->ctx, V4L2_CID_STATELESS_H264_DECODE_PARAMS); + if (IS_ERR(dec_params)) + return PTR_ERR(dec_params); + + sps = mtk_vdec_h264_get_ctrl_ptr(inst->ctx, V4L2_CID_STATELESS_H264_SPS); + if (IS_ERR(sps)) + return PTR_ERR(sps); + + pps = mtk_vdec_h264_get_ctrl_ptr(inst->ctx, V4L2_CID_STATELESS_H264_PPS); + if (IS_ERR(pps)) + return PTR_ERR(pps); + + scaling_matrix = + mtk_vdec_h264_get_ctrl_ptr(inst->ctx, V4L2_CID_STATELESS_H264_SCALING_MATRIX); + if (IS_ERR(scaling_matrix)) + return PTR_ERR(scaling_matrix); + + mtk_vdec_h264_update_dpb(dec_params, inst->dpb); + + mtk_vdec_h264_copy_sps_params(&slice_param->sps, sps); + mtk_vdec_h264_copy_pps_params(&slice_param->pps, pps); + mtk_vdec_h264_copy_scaling_matrix(&slice_param->scaling_matrix, scaling_matrix); + mtk_vdec_h264_copy_decode_params(&slice_param->decode_params, + dec_params, inst->dpb); + mtk_vdec_h264_fill_dpb_info(inst->ctx, &slice_param->decode_params, + slice_param->h264_dpb_info); + + /* Build the reference lists */ + v4l2_h264_init_reflist_builder(&reflist_builder, dec_params, sps, + inst->dpb); + v4l2_h264_build_p_ref_list(&reflist_builder, v4l2_p0_reflist); + v4l2_h264_build_b_ref_lists(&reflist_builder, v4l2_b0_reflist, + v4l2_b1_reflist); + + /* Adapt the built lists to the firmware's expectations */ + mtk_vdec_h264_get_ref_list(p0_reflist, v4l2_p0_reflist, reflist_builder.num_valid); + mtk_vdec_h264_get_ref_list(b0_reflist, v4l2_b0_reflist, reflist_builder.num_valid); + mtk_vdec_h264_get_ref_list(b1_reflist, v4l2_b1_reflist, reflist_builder.num_valid); + + memcpy(&inst->vsi_ctx.h264_slice_params, slice_param, + sizeof(inst->vsi_ctx.h264_slice_params)); + + return 0; +} + +static int allocate_predication_buf(struct vdec_h264_slice_inst *inst) +{ + int err; + + inst->pred_buf.size = BUF_PREDICTION_SZ; + err = mtk_vcodec_mem_alloc(inst->ctx, &inst->pred_buf); + if (err) { + mtk_vdec_err(inst->ctx, "failed to allocate ppl buf"); + return err; + } + + inst->vsi_ctx.pred_buf_dma = inst->pred_buf.dma_addr; + return 0; +} + +static void free_predication_buf(struct vdec_h264_slice_inst *inst) +{ + struct mtk_vcodec_mem *mem = &inst->pred_buf; + + inst->vsi_ctx.pred_buf_dma = 0; + if (mem->va) + mtk_vcodec_mem_free(inst->ctx, mem); +} + +static int alloc_mv_buf(struct vdec_h264_slice_inst *inst, + struct vdec_pic_info *pic) +{ + int i; + int err; + struct mtk_vcodec_mem *mem = NULL; + unsigned int buf_sz = mtk_vdec_h264_get_mv_buf_size(pic->buf_w, pic->buf_h); + + mtk_v4l2_vdec_dbg(3, inst->ctx, "size = 0x%x", buf_sz); + for (i = 0; i < H264_MAX_MV_NUM; i++) { + mem = &inst->mv_buf[i]; + if (mem->va) + mtk_vcodec_mem_free(inst->ctx, mem); + mem->size = buf_sz; + err = mtk_vcodec_mem_alloc(inst->ctx, mem); + if (err) { + mtk_vdec_err(inst->ctx, "failed to allocate mv buf"); + return err; + } + inst->vsi_ctx.mv_buf_dma[i] = mem->dma_addr; + } + + return 0; +} + +static void free_mv_buf(struct vdec_h264_slice_inst *inst) +{ + int i; + struct mtk_vcodec_mem *mem; + + for (i = 0; i < H264_MAX_MV_NUM; i++) { + inst->vsi_ctx.mv_buf_dma[i] = 0; + mem = &inst->mv_buf[i]; + if (mem->va) + mtk_vcodec_mem_free(inst->ctx, mem); + } +} + +static void get_pic_info(struct vdec_h264_slice_inst *inst, + struct vdec_pic_info *pic) +{ + struct mtk_vcodec_dec_ctx *ctx = inst->ctx; + + ctx->picinfo.buf_w = ALIGN(ctx->picinfo.pic_w, VCODEC_DEC_ALIGNED_64); + ctx->picinfo.buf_h = ALIGN(ctx->picinfo.pic_h, VCODEC_DEC_ALIGNED_64); + ctx->picinfo.fb_sz[0] = ctx->picinfo.buf_w * ctx->picinfo.buf_h; + ctx->picinfo.fb_sz[1] = ctx->picinfo.fb_sz[0] >> 1; + inst->vsi_ctx.dec.cap_num_planes = + ctx->q_data[MTK_Q_DATA_DST].fmt->num_planes; + + *pic = ctx->picinfo; + mtk_vdec_debug(inst->ctx, "pic(%d, %d), buf(%d, %d)", + ctx->picinfo.pic_w, ctx->picinfo.pic_h, + ctx->picinfo.buf_w, ctx->picinfo.buf_h); + mtk_vdec_debug(inst->ctx, "Y/C(%d, %d)", ctx->picinfo.fb_sz[0], + ctx->picinfo.fb_sz[1]); + + if (ctx->last_decoded_picinfo.pic_w != ctx->picinfo.pic_w || + ctx->last_decoded_picinfo.pic_h != ctx->picinfo.pic_h) { + inst->vsi_ctx.dec.resolution_changed = true; + if (ctx->last_decoded_picinfo.buf_w != ctx->picinfo.buf_w || + ctx->last_decoded_picinfo.buf_h != ctx->picinfo.buf_h) + inst->vsi_ctx.dec.realloc_mv_buf = true; + + mtk_v4l2_vdec_dbg(1, inst->ctx, "ResChg: (%d %d) : old(%d, %d) -> new(%d, %d)", + inst->vsi_ctx.dec.resolution_changed, + inst->vsi_ctx.dec.realloc_mv_buf, + ctx->last_decoded_picinfo.pic_w, + ctx->last_decoded_picinfo.pic_h, + ctx->picinfo.pic_w, ctx->picinfo.pic_h); + } +} + +static void get_crop_info(struct vdec_h264_slice_inst *inst, struct v4l2_rect *cr) +{ + cr->left = inst->vsi_ctx.crop.left; + cr->top = inst->vsi_ctx.crop.top; + cr->width = inst->vsi_ctx.crop.width; + cr->height = inst->vsi_ctx.crop.height; + + mtk_vdec_debug(inst->ctx, "l=%d, t=%d, w=%d, h=%d", + cr->left, cr->top, cr->width, cr->height); +} + +static void get_dpb_size(struct vdec_h264_slice_inst *inst, unsigned int *dpb_sz) +{ + *dpb_sz = inst->vsi_ctx.dec.dpb_sz; + mtk_vdec_debug(inst->ctx, "sz=%d", *dpb_sz); +} + +static int vdec_h264_slice_init(struct mtk_vcodec_dec_ctx *ctx) +{ + struct vdec_h264_slice_inst *inst; + int err; + + inst = kzalloc(sizeof(*inst), GFP_KERNEL); + if (!inst) + return -ENOMEM; + + inst->ctx = ctx; + + inst->vpu.id = SCP_IPI_VDEC_H264; + inst->vpu.ctx = ctx; + + err = vpu_dec_init(&inst->vpu); + if (err) { + mtk_vdec_err(ctx, "vdec_h264 init err=%d", err); + goto error_free_inst; + } + + memcpy(&inst->vsi_ctx, inst->vpu.vsi, sizeof(inst->vsi_ctx)); + inst->vsi_ctx.dec.resolution_changed = true; + inst->vsi_ctx.dec.realloc_mv_buf = true; + + err = allocate_predication_buf(inst); + if (err) + goto error_deinit; + + mtk_vdec_debug(ctx, "struct size = %zu,%zu,%zu,%zu\n", + sizeof(struct mtk_h264_sps_param), + sizeof(struct mtk_h264_pps_param), + sizeof(struct mtk_h264_dec_slice_param), + sizeof(struct mtk_h264_dpb_info)); + + mtk_vdec_debug(ctx, "H264 Instance >> %p", inst); + + ctx->drv_handle = inst; + return 0; + +error_deinit: + vpu_dec_deinit(&inst->vpu); + +error_free_inst: + kfree(inst); + return err; +} + +static void vdec_h264_slice_deinit(void *h_vdec) +{ + struct vdec_h264_slice_inst *inst = h_vdec; + + vpu_dec_deinit(&inst->vpu); + free_predication_buf(inst); + free_mv_buf(inst); + + kfree(inst); +} + +static int vdec_h264_slice_decode(void *h_vdec, struct mtk_vcodec_mem *bs, + struct vdec_fb *unused, bool *res_chg) +{ + struct vdec_h264_slice_inst *inst = h_vdec; + const struct v4l2_ctrl_h264_decode_params *dec_params = + mtk_vdec_h264_get_ctrl_ptr(inst->ctx, V4L2_CID_STATELESS_H264_DECODE_PARAMS); + struct vdec_vpu_inst *vpu = &inst->vpu; + struct mtk_video_dec_buf *src_buf_info; + struct mtk_video_dec_buf *dst_buf_info; + struct vdec_fb *fb; + u32 data[2]; + u64 y_fb_dma; + u64 c_fb_dma; + int err; + + inst->num_nalu++; + /* bs NULL means flush decoder */ + if (!bs) + return vpu_dec_reset(vpu); + + fb = inst->ctx->dev->vdec_pdata->get_cap_buffer(inst->ctx); + src_buf_info = container_of(bs, struct mtk_video_dec_buf, bs_buffer); + dst_buf_info = container_of(fb, struct mtk_video_dec_buf, frame_buffer); + + y_fb_dma = fb ? (u64)fb->base_y.dma_addr : 0; + c_fb_dma = fb ? (u64)fb->base_c.dma_addr : 0; + + mtk_vdec_debug(inst->ctx, "+ [%d] FB y_dma=%llx c_dma=%llx va=%p", + inst->num_nalu, y_fb_dma, c_fb_dma, fb); + + inst->vsi_ctx.dec.bs_dma = (uint64_t)bs->dma_addr; + inst->vsi_ctx.dec.y_fb_dma = y_fb_dma; + inst->vsi_ctx.dec.c_fb_dma = c_fb_dma; + inst->vsi_ctx.dec.vdec_fb_va = (u64)(uintptr_t)fb; + + v4l2_m2m_buf_copy_metadata(&src_buf_info->m2m_buf.vb, + &dst_buf_info->m2m_buf.vb, true); + err = get_vdec_decode_parameters(inst); + if (err) + goto err_free_fb_out; + + data[0] = bs->size; + /* + * Reconstruct the first byte of the NAL unit, as the firmware requests + * that information to be passed even though it is present in the stream + * itself... + */ + data[1] = (dec_params->nal_ref_idc << 5) | + ((dec_params->flags & V4L2_H264_DECODE_PARAM_FLAG_IDR_PIC) + ? 0x5 : 0x1); + + *res_chg = inst->vsi_ctx.dec.resolution_changed; + if (*res_chg) { + mtk_vdec_debug(inst->ctx, "- resolution changed -"); + if (inst->vsi_ctx.dec.realloc_mv_buf) { + err = alloc_mv_buf(inst, &inst->ctx->picinfo); + inst->vsi_ctx.dec.realloc_mv_buf = false; + if (err) + goto err_free_fb_out; + } + *res_chg = false; + } + + memcpy(inst->vpu.vsi, &inst->vsi_ctx, sizeof(inst->vsi_ctx)); + err = vpu_dec_start(vpu, data, 2); + if (err) + goto err_free_fb_out; + + /* wait decoder done interrupt */ + err = mtk_vcodec_wait_for_done_ctx(inst->ctx, + MTK_INST_IRQ_RECEIVED, + WAIT_INTR_TIMEOUT_MS, 0); + if (err) + goto err_free_fb_out; + vpu_dec_end(vpu); + + memcpy(&inst->vsi_ctx, inst->vpu.vsi, sizeof(inst->vsi_ctx)); + mtk_vdec_debug(inst->ctx, "\n - NALU[%d]", inst->num_nalu); + return 0; + +err_free_fb_out: + mtk_vdec_err(inst->ctx, "\n - NALU[%d] err=%d -\n", inst->num_nalu, err); + return err; +} + +static int vdec_h264_slice_get_param(void *h_vdec, enum vdec_get_param_type type, void *out) +{ + struct vdec_h264_slice_inst *inst = h_vdec; + + switch (type) { + case GET_PARAM_PIC_INFO: + get_pic_info(inst, out); + break; + + case GET_PARAM_DPB_SIZE: + get_dpb_size(inst, out); + break; + + case GET_PARAM_CROP_INFO: + get_crop_info(inst, out); + break; + + default: + mtk_vdec_err(inst->ctx, "invalid get parameter type=%d", type); + return -EINVAL; + } + + return 0; +} + +const struct vdec_common_if vdec_h264_slice_if = { + .init = vdec_h264_slice_init, + .decode = vdec_h264_slice_decode, + .get_param = vdec_h264_slice_get_param, + .deinit = vdec_h264_slice_deinit, +}; diff --git a/drivers/media/platform/mediatek/vcodec/decoder/vdec/vdec_h264_req_multi_if.c b/drivers/media/platform/mediatek/vcodec/decoder/vdec/vdec_h264_req_multi_if.c new file mode 100644 index 0000000000..0e741e0dc8 --- /dev/null +++ b/drivers/media/platform/mediatek/vcodec/decoder/vdec/vdec_h264_req_multi_if.c @@ -0,0 +1,848 @@ +// SPDX-License-Identifier: GPL-2.0 +/* + * Copyright (c) 2022 MediaTek Inc. + * Author: Yunfei Dong <yunfei.dong@mediatek.com> + */ + +#include <linux/module.h> +#include <linux/slab.h> +#include <media/v4l2-h264.h> +#include <media/v4l2-mem2mem.h> +#include <media/videobuf2-dma-contig.h> + +#include "../mtk_vcodec_dec.h" +#include "../../common/mtk_vcodec_intr.h" +#include "../vdec_drv_base.h" +#include "../vdec_drv_if.h" +#include "../vdec_vpu_if.h" +#include "vdec_h264_req_common.h" + +/** + * enum vdec_h264_core_dec_err_type - core decode error type + * + * @TRANS_BUFFER_FULL: trans buffer is full + * @SLICE_HEADER_FULL: slice header buffer is full + */ +enum vdec_h264_core_dec_err_type { + TRANS_BUFFER_FULL = 1, + SLICE_HEADER_FULL, +}; + +/** + * struct vdec_h264_slice_lat_dec_param - parameters for decode current frame + * + * @sps: h264 sps syntax parameters + * @pps: h264 pps syntax parameters + * @slice_header: h264 slice header syntax parameters + * @scaling_matrix: h264 scaling list parameters + * @decode_params: decoder parameters of each frame used for hardware decode + * @h264_dpb_info: dpb reference list + */ +struct vdec_h264_slice_lat_dec_param { + struct mtk_h264_sps_param sps; + struct mtk_h264_pps_param pps; + struct mtk_h264_slice_hd_param slice_header; + struct slice_api_h264_scaling_matrix scaling_matrix; + struct slice_api_h264_decode_param decode_params; + struct mtk_h264_dpb_info h264_dpb_info[V4L2_H264_NUM_DPB_ENTRIES]; +}; + +/** + * struct vdec_h264_slice_info - decode information + * + * @nal_info: nal info of current picture + * @timeout: Decode timeout: 1 timeout, 0 no timeount + * @bs_buf_size: bitstream size + * @bs_buf_addr: bitstream buffer dma address + * @y_fb_dma: Y frame buffer dma address + * @c_fb_dma: C frame buffer dma address + * @vdec_fb_va: VDEC frame buffer struct virtual address + * @crc: Used to check whether hardware's status is right + */ +struct vdec_h264_slice_info { + u16 nal_info; + u16 timeout; + u32 bs_buf_size; + u64 bs_buf_addr; + u64 y_fb_dma; + u64 c_fb_dma; + u64 vdec_fb_va; + u32 crc[8]; +}; + +/** + * struct vdec_h264_slice_vsi - shared memory for decode information exchange + * between SCP and Host. + * + * @wdma_err_addr: wdma error dma address + * @wdma_start_addr: wdma start dma address + * @wdma_end_addr: wdma end dma address + * @slice_bc_start_addr: slice bc start dma address + * @slice_bc_end_addr: slice bc end dma address + * @row_info_start_addr: row info start dma address + * @row_info_end_addr: row info end dma address + * @trans_start: trans start dma address + * @trans_end: trans end dma address + * @wdma_end_addr_offset: wdma end address offset + * + * @mv_buf_dma: HW working motion vector buffer + * dma address (AP-W, VPU-R) + * @dec: decode information (AP-R, VPU-W) + * @h264_slice_params: decode parameters for hw used + */ +struct vdec_h264_slice_vsi { + /* LAT dec addr */ + u64 wdma_err_addr; + u64 wdma_start_addr; + u64 wdma_end_addr; + u64 slice_bc_start_addr; + u64 slice_bc_end_addr; + u64 row_info_start_addr; + u64 row_info_end_addr; + u64 trans_start; + u64 trans_end; + u64 wdma_end_addr_offset; + + u64 mv_buf_dma[H264_MAX_MV_NUM]; + struct vdec_h264_slice_info dec; + struct vdec_h264_slice_lat_dec_param h264_slice_params; +}; + +/** + * struct vdec_h264_slice_share_info - shared information used to exchange + * message between lat and core + * + * @sps: sequence header information from user space + * @dec_params: decoder params from user space + * @h264_slice_params: decoder params used for hardware + * @trans_start: trans start dma address + * @trans_end: trans end dma address + * @nal_info: nal info of current picture + */ +struct vdec_h264_slice_share_info { + struct v4l2_ctrl_h264_sps sps; + struct v4l2_ctrl_h264_decode_params dec_params; + struct vdec_h264_slice_lat_dec_param h264_slice_params; + u64 trans_start; + u64 trans_end; + u16 nal_info; +}; + +/** + * struct vdec_h264_slice_inst - h264 decoder instance + * + * @slice_dec_num: how many picture be decoded + * @ctx: point to mtk_vcodec_dec_ctx + * @pred_buf: HW working predication buffer + * @mv_buf: HW working motion vector buffer + * @vpu: VPU instance + * @vsi: vsi used for lat + * @vsi_core: vsi used for core + * + * @vsi_ctx: Local VSI data for this decoding context + * @h264_slice_param: the parameters that hardware use to decode + * + * @resolution_changed:resolution changed + * @realloc_mv_buf: reallocate mv buffer + * @cap_num_planes: number of capture queue plane + * + * @dpb: decoded picture buffer used to store reference + * buffer information + *@is_field_bitstream: is field bitstream + */ +struct vdec_h264_slice_inst { + unsigned int slice_dec_num; + struct mtk_vcodec_dec_ctx *ctx; + struct mtk_vcodec_mem pred_buf; + struct mtk_vcodec_mem mv_buf[H264_MAX_MV_NUM]; + struct vdec_vpu_inst vpu; + struct vdec_h264_slice_vsi *vsi; + struct vdec_h264_slice_vsi *vsi_core; + + struct vdec_h264_slice_vsi vsi_ctx; + struct vdec_h264_slice_lat_dec_param h264_slice_param; + + unsigned int resolution_changed; + unsigned int realloc_mv_buf; + unsigned int cap_num_planes; + + struct v4l2_h264_dpb_entry dpb[16]; + bool is_field_bitstream; +}; + +static int vdec_h264_slice_fill_decode_parameters(struct vdec_h264_slice_inst *inst, + struct vdec_h264_slice_share_info *share_info) +{ + struct vdec_h264_slice_lat_dec_param *slice_param = &inst->vsi->h264_slice_params; + const struct v4l2_ctrl_h264_decode_params *dec_params; + const struct v4l2_ctrl_h264_scaling_matrix *src_matrix; + const struct v4l2_ctrl_h264_sps *sps; + const struct v4l2_ctrl_h264_pps *pps; + + dec_params = + mtk_vdec_h264_get_ctrl_ptr(inst->ctx, V4L2_CID_STATELESS_H264_DECODE_PARAMS); + if (IS_ERR(dec_params)) + return PTR_ERR(dec_params); + + src_matrix = + mtk_vdec_h264_get_ctrl_ptr(inst->ctx, V4L2_CID_STATELESS_H264_SCALING_MATRIX); + if (IS_ERR(src_matrix)) + return PTR_ERR(src_matrix); + + sps = mtk_vdec_h264_get_ctrl_ptr(inst->ctx, V4L2_CID_STATELESS_H264_SPS); + if (IS_ERR(sps)) + return PTR_ERR(sps); + + pps = mtk_vdec_h264_get_ctrl_ptr(inst->ctx, V4L2_CID_STATELESS_H264_PPS); + if (IS_ERR(pps)) + return PTR_ERR(pps); + + if (dec_params->flags & V4L2_H264_DECODE_PARAM_FLAG_FIELD_PIC) { + mtk_vdec_err(inst->ctx, "No support for H.264 field decoding."); + inst->is_field_bitstream = true; + return -EINVAL; + } + + mtk_vdec_h264_copy_sps_params(&slice_param->sps, sps); + mtk_vdec_h264_copy_pps_params(&slice_param->pps, pps); + mtk_vdec_h264_copy_scaling_matrix(&slice_param->scaling_matrix, src_matrix); + + memcpy(&share_info->sps, sps, sizeof(*sps)); + memcpy(&share_info->dec_params, dec_params, sizeof(*dec_params)); + + return 0; +} + +static int get_vdec_sig_decode_parameters(struct vdec_h264_slice_inst *inst) +{ + const struct v4l2_ctrl_h264_decode_params *dec_params; + const struct v4l2_ctrl_h264_sps *sps; + const struct v4l2_ctrl_h264_pps *pps; + const struct v4l2_ctrl_h264_scaling_matrix *scaling_matrix; + struct vdec_h264_slice_lat_dec_param *slice_param = &inst->h264_slice_param; + struct v4l2_h264_reflist_builder reflist_builder; + struct v4l2_h264_reference v4l2_p0_reflist[V4L2_H264_REF_LIST_LEN]; + struct v4l2_h264_reference v4l2_b0_reflist[V4L2_H264_REF_LIST_LEN]; + struct v4l2_h264_reference v4l2_b1_reflist[V4L2_H264_REF_LIST_LEN]; + u8 *p0_reflist = slice_param->decode_params.ref_pic_list_p0; + u8 *b0_reflist = slice_param->decode_params.ref_pic_list_b0; + u8 *b1_reflist = slice_param->decode_params.ref_pic_list_b1; + + dec_params = + mtk_vdec_h264_get_ctrl_ptr(inst->ctx, V4L2_CID_STATELESS_H264_DECODE_PARAMS); + if (IS_ERR(dec_params)) + return PTR_ERR(dec_params); + + sps = mtk_vdec_h264_get_ctrl_ptr(inst->ctx, V4L2_CID_STATELESS_H264_SPS); + if (IS_ERR(sps)) + return PTR_ERR(sps); + + pps = mtk_vdec_h264_get_ctrl_ptr(inst->ctx, V4L2_CID_STATELESS_H264_PPS); + if (IS_ERR(pps)) + return PTR_ERR(pps); + + scaling_matrix = + mtk_vdec_h264_get_ctrl_ptr(inst->ctx, V4L2_CID_STATELESS_H264_SCALING_MATRIX); + if (IS_ERR(scaling_matrix)) + return PTR_ERR(scaling_matrix); + + mtk_vdec_h264_update_dpb(dec_params, inst->dpb); + + mtk_vdec_h264_copy_sps_params(&slice_param->sps, sps); + mtk_vdec_h264_copy_pps_params(&slice_param->pps, pps); + mtk_vdec_h264_copy_scaling_matrix(&slice_param->scaling_matrix, scaling_matrix); + + mtk_vdec_h264_copy_decode_params(&slice_param->decode_params, dec_params, inst->dpb); + mtk_vdec_h264_fill_dpb_info(inst->ctx, &slice_param->decode_params, + slice_param->h264_dpb_info); + + /* Build the reference lists */ + v4l2_h264_init_reflist_builder(&reflist_builder, dec_params, sps, inst->dpb); + v4l2_h264_build_p_ref_list(&reflist_builder, v4l2_p0_reflist); + v4l2_h264_build_b_ref_lists(&reflist_builder, v4l2_b0_reflist, v4l2_b1_reflist); + + /* Adapt the built lists to the firmware's expectations */ + mtk_vdec_h264_get_ref_list(p0_reflist, v4l2_p0_reflist, reflist_builder.num_valid); + mtk_vdec_h264_get_ref_list(b0_reflist, v4l2_b0_reflist, reflist_builder.num_valid); + mtk_vdec_h264_get_ref_list(b1_reflist, v4l2_b1_reflist, reflist_builder.num_valid); + + memcpy(&inst->vsi_ctx.h264_slice_params, slice_param, + sizeof(inst->vsi_ctx.h264_slice_params)); + + return 0; +} + +static void vdec_h264_slice_fill_decode_reflist(struct vdec_h264_slice_inst *inst, + struct vdec_h264_slice_lat_dec_param *slice_param, + struct vdec_h264_slice_share_info *share_info) +{ + struct v4l2_ctrl_h264_decode_params *dec_params = &share_info->dec_params; + struct v4l2_ctrl_h264_sps *sps = &share_info->sps; + struct v4l2_h264_reflist_builder reflist_builder; + struct v4l2_h264_reference v4l2_p0_reflist[V4L2_H264_REF_LIST_LEN]; + struct v4l2_h264_reference v4l2_b0_reflist[V4L2_H264_REF_LIST_LEN]; + struct v4l2_h264_reference v4l2_b1_reflist[V4L2_H264_REF_LIST_LEN]; + u8 *p0_reflist = slice_param->decode_params.ref_pic_list_p0; + u8 *b0_reflist = slice_param->decode_params.ref_pic_list_b0; + u8 *b1_reflist = slice_param->decode_params.ref_pic_list_b1; + + mtk_vdec_h264_update_dpb(dec_params, inst->dpb); + + mtk_vdec_h264_copy_decode_params(&slice_param->decode_params, dec_params, + inst->dpb); + mtk_vdec_h264_fill_dpb_info(inst->ctx, &slice_param->decode_params, + slice_param->h264_dpb_info); + + mtk_v4l2_vdec_dbg(3, inst->ctx, "cur poc = %d\n", dec_params->bottom_field_order_cnt); + /* Build the reference lists */ + v4l2_h264_init_reflist_builder(&reflist_builder, dec_params, sps, + inst->dpb); + v4l2_h264_build_p_ref_list(&reflist_builder, v4l2_p0_reflist); + v4l2_h264_build_b_ref_lists(&reflist_builder, v4l2_b0_reflist, v4l2_b1_reflist); + + /* Adapt the built lists to the firmware's expectations */ + mtk_vdec_h264_get_ref_list(p0_reflist, v4l2_p0_reflist, reflist_builder.num_valid); + mtk_vdec_h264_get_ref_list(b0_reflist, v4l2_b0_reflist, reflist_builder.num_valid); + mtk_vdec_h264_get_ref_list(b1_reflist, v4l2_b1_reflist, reflist_builder.num_valid); +} + +static int vdec_h264_slice_alloc_mv_buf(struct vdec_h264_slice_inst *inst, + struct vdec_pic_info *pic) +{ + unsigned int buf_sz = mtk_vdec_h264_get_mv_buf_size(pic->buf_w, pic->buf_h); + struct mtk_vcodec_mem *mem; + int i, err; + + mtk_v4l2_vdec_dbg(3, inst->ctx, "size = 0x%x", buf_sz); + for (i = 0; i < H264_MAX_MV_NUM; i++) { + mem = &inst->mv_buf[i]; + if (mem->va) + mtk_vcodec_mem_free(inst->ctx, mem); + mem->size = buf_sz; + err = mtk_vcodec_mem_alloc(inst->ctx, mem); + if (err) { + mtk_vdec_err(inst->ctx, "failed to allocate mv buf"); + return err; + } + } + + return 0; +} + +static void vdec_h264_slice_free_mv_buf(struct vdec_h264_slice_inst *inst) +{ + int i; + struct mtk_vcodec_mem *mem; + + for (i = 0; i < H264_MAX_MV_NUM; i++) { + mem = &inst->mv_buf[i]; + if (mem->va) + mtk_vcodec_mem_free(inst->ctx, mem); + } +} + +static void vdec_h264_slice_get_pic_info(struct vdec_h264_slice_inst *inst) +{ + struct mtk_vcodec_dec_ctx *ctx = inst->ctx; + u32 data[3]; + + data[0] = ctx->picinfo.pic_w; + data[1] = ctx->picinfo.pic_h; + data[2] = ctx->capture_fourcc; + vpu_dec_get_param(&inst->vpu, data, 3, GET_PARAM_PIC_INFO); + + ctx->picinfo.buf_w = ALIGN(ctx->picinfo.pic_w, VCODEC_DEC_ALIGNED_64); + ctx->picinfo.buf_h = ALIGN(ctx->picinfo.pic_h, VCODEC_DEC_ALIGNED_64); + ctx->picinfo.fb_sz[0] = inst->vpu.fb_sz[0]; + ctx->picinfo.fb_sz[1] = inst->vpu.fb_sz[1]; + inst->cap_num_planes = + ctx->q_data[MTK_Q_DATA_DST].fmt->num_planes; + + mtk_vdec_debug(ctx, "pic(%d, %d), buf(%d, %d)", + ctx->picinfo.pic_w, ctx->picinfo.pic_h, + ctx->picinfo.buf_w, ctx->picinfo.buf_h); + mtk_vdec_debug(ctx, "Y/C(%d, %d)", ctx->picinfo.fb_sz[0], + ctx->picinfo.fb_sz[1]); + + if (ctx->last_decoded_picinfo.pic_w != ctx->picinfo.pic_w || + ctx->last_decoded_picinfo.pic_h != ctx->picinfo.pic_h) { + inst->resolution_changed = true; + if (ctx->last_decoded_picinfo.buf_w != ctx->picinfo.buf_w || + ctx->last_decoded_picinfo.buf_h != ctx->picinfo.buf_h) + inst->realloc_mv_buf = true; + + mtk_v4l2_vdec_dbg(1, inst->ctx, "resChg: (%d %d) : old(%d, %d) -> new(%d, %d)", + inst->resolution_changed, + inst->realloc_mv_buf, + ctx->last_decoded_picinfo.pic_w, + ctx->last_decoded_picinfo.pic_h, + ctx->picinfo.pic_w, ctx->picinfo.pic_h); + } +} + +static void vdec_h264_slice_get_crop_info(struct vdec_h264_slice_inst *inst, + struct v4l2_rect *cr) +{ + cr->left = 0; + cr->top = 0; + cr->width = inst->ctx->picinfo.pic_w; + cr->height = inst->ctx->picinfo.pic_h; + + mtk_vdec_debug(inst->ctx, "l=%d, t=%d, w=%d, h=%d", + cr->left, cr->top, cr->width, cr->height); +} + +static int vdec_h264_slice_init(struct mtk_vcodec_dec_ctx *ctx) +{ + struct vdec_h264_slice_inst *inst; + int err, vsi_size; + + inst = kzalloc(sizeof(*inst), GFP_KERNEL); + if (!inst) + return -ENOMEM; + + inst->ctx = ctx; + + inst->vpu.id = SCP_IPI_VDEC_LAT; + inst->vpu.core_id = SCP_IPI_VDEC_CORE; + inst->vpu.ctx = ctx; + inst->vpu.codec_type = ctx->current_codec; + inst->vpu.capture_type = ctx->capture_fourcc; + + err = vpu_dec_init(&inst->vpu); + if (err) { + mtk_vdec_err(ctx, "vdec_h264 init err=%d", err); + goto error_free_inst; + } + + vsi_size = round_up(sizeof(struct vdec_h264_slice_vsi), VCODEC_DEC_ALIGNED_64); + inst->vsi = inst->vpu.vsi; + inst->vsi_core = + (struct vdec_h264_slice_vsi *)(((char *)inst->vpu.vsi) + vsi_size); + inst->resolution_changed = true; + inst->realloc_mv_buf = true; + + mtk_vdec_debug(ctx, "lat struct size = %d,%d,%d,%d vsi: %d\n", + (int)sizeof(struct mtk_h264_sps_param), + (int)sizeof(struct mtk_h264_pps_param), + (int)sizeof(struct vdec_h264_slice_lat_dec_param), + (int)sizeof(struct mtk_h264_dpb_info), + vsi_size); + mtk_vdec_debug(ctx, "lat H264 instance >> %p, codec_type = 0x%x", + inst, inst->vpu.codec_type); + + ctx->drv_handle = inst; + return 0; + +error_free_inst: + kfree(inst); + return err; +} + +static void vdec_h264_slice_deinit(void *h_vdec) +{ + struct vdec_h264_slice_inst *inst = h_vdec; + + vpu_dec_deinit(&inst->vpu); + vdec_h264_slice_free_mv_buf(inst); + vdec_msg_queue_deinit(&inst->ctx->msg_queue, inst->ctx); + + kfree(inst); +} + +static int vdec_h264_slice_core_decode(struct vdec_lat_buf *lat_buf) +{ + struct vdec_fb *fb; + u64 vdec_fb_va; + u64 y_fb_dma, c_fb_dma; + int err, timeout, i; + struct mtk_vcodec_dec_ctx *ctx = lat_buf->ctx; + struct vdec_h264_slice_inst *inst = ctx->drv_handle; + struct vb2_v4l2_buffer *vb2_v4l2; + struct vdec_h264_slice_share_info *share_info = lat_buf->private_data; + struct mtk_vcodec_mem *mem; + struct vdec_vpu_inst *vpu = &inst->vpu; + + mtk_vdec_debug(ctx, "[h264-core] vdec_h264 core decode"); + memcpy(&inst->vsi_core->h264_slice_params, &share_info->h264_slice_params, + sizeof(share_info->h264_slice_params)); + + fb = ctx->dev->vdec_pdata->get_cap_buffer(ctx); + if (!fb) { + err = -EBUSY; + mtk_vdec_err(ctx, "fb buffer is NULL"); + goto vdec_dec_end; + } + + vdec_fb_va = (unsigned long)fb; + y_fb_dma = (u64)fb->base_y.dma_addr; + if (ctx->q_data[MTK_Q_DATA_DST].fmt->num_planes == 1) + c_fb_dma = + y_fb_dma + inst->ctx->picinfo.buf_w * inst->ctx->picinfo.buf_h; + else + c_fb_dma = (u64)fb->base_c.dma_addr; + + mtk_vdec_debug(ctx, "[h264-core] y/c addr = 0x%llx 0x%llx", y_fb_dma, c_fb_dma); + + inst->vsi_core->dec.y_fb_dma = y_fb_dma; + inst->vsi_core->dec.c_fb_dma = c_fb_dma; + inst->vsi_core->dec.vdec_fb_va = vdec_fb_va; + inst->vsi_core->dec.nal_info = share_info->nal_info; + inst->vsi_core->wdma_start_addr = + lat_buf->ctx->msg_queue.wdma_addr.dma_addr; + inst->vsi_core->wdma_end_addr = + lat_buf->ctx->msg_queue.wdma_addr.dma_addr + + lat_buf->ctx->msg_queue.wdma_addr.size; + inst->vsi_core->wdma_err_addr = lat_buf->wdma_err_addr.dma_addr; + inst->vsi_core->slice_bc_start_addr = lat_buf->slice_bc_addr.dma_addr; + inst->vsi_core->slice_bc_end_addr = lat_buf->slice_bc_addr.dma_addr + + lat_buf->slice_bc_addr.size; + inst->vsi_core->trans_start = share_info->trans_start; + inst->vsi_core->trans_end = share_info->trans_end; + for (i = 0; i < H264_MAX_MV_NUM; i++) { + mem = &inst->mv_buf[i]; + inst->vsi_core->mv_buf_dma[i] = mem->dma_addr; + } + + vb2_v4l2 = v4l2_m2m_next_dst_buf(ctx->m2m_ctx); + v4l2_m2m_buf_copy_metadata(&lat_buf->ts_info, vb2_v4l2, true); + + vdec_h264_slice_fill_decode_reflist(inst, &inst->vsi_core->h264_slice_params, + share_info); + + err = vpu_dec_core(vpu); + if (err) { + mtk_vdec_err(ctx, "core decode err=%d", err); + goto vdec_dec_end; + } + + /* wait decoder done interrupt */ + timeout = mtk_vcodec_wait_for_done_ctx(inst->ctx, MTK_INST_IRQ_RECEIVED, + WAIT_INTR_TIMEOUT_MS, MTK_VDEC_CORE); + if (timeout) + mtk_vdec_err(ctx, "core decode timeout: pic_%d", ctx->decoded_frame_cnt); + inst->vsi_core->dec.timeout = !!timeout; + + vpu_dec_core_end(vpu); + mtk_vdec_debug(ctx, "pic[%d] crc: 0x%x 0x%x 0x%x 0x%x 0x%x 0x%x 0x%x 0x%x", + ctx->decoded_frame_cnt, + inst->vsi_core->dec.crc[0], inst->vsi_core->dec.crc[1], + inst->vsi_core->dec.crc[2], inst->vsi_core->dec.crc[3], + inst->vsi_core->dec.crc[4], inst->vsi_core->dec.crc[5], + inst->vsi_core->dec.crc[6], inst->vsi_core->dec.crc[7]); + +vdec_dec_end: + vdec_msg_queue_update_ube_rptr(&lat_buf->ctx->msg_queue, share_info->trans_end); + ctx->dev->vdec_pdata->cap_to_disp(ctx, !!err, lat_buf->src_buf_req); + mtk_vdec_debug(ctx, "core decode done err=%d", err); + ctx->decoded_frame_cnt++; + return 0; +} + +static void vdec_h264_insert_startcode(struct mtk_vcodec_dec_dev *vcodec_dev, unsigned char *buf, + size_t *bs_size, struct mtk_h264_pps_param *pps) +{ + struct device *dev = &vcodec_dev->plat_dev->dev; + + /* Need to add pending data at the end of bitstream when bs_sz is small than + * 20 bytes for cavlc bitstream, or lat will decode fail. This pending data is + * useful for mt8192 and mt8195 platform. + * + * cavlc bitstream when entropy_coding_mode_flag is false. + */ + if (pps->entropy_coding_mode_flag || *bs_size > 20 || + !(of_device_is_compatible(dev->of_node, "mediatek,mt8192-vcodec-dec") || + of_device_is_compatible(dev->of_node, "mediatek,mt8195-vcodec-dec"))) + return; + + buf[*bs_size] = 0; + buf[*bs_size + 1] = 0; + buf[*bs_size + 2] = 1; + buf[*bs_size + 3] = 0xff; + (*bs_size) += 4; +} + +static int vdec_h264_slice_lat_decode(void *h_vdec, struct mtk_vcodec_mem *bs, + struct vdec_fb *fb, bool *res_chg) +{ + struct vdec_h264_slice_inst *inst = h_vdec; + struct vdec_vpu_inst *vpu = &inst->vpu; + struct mtk_video_dec_buf *src_buf_info; + int nal_start_idx, err, timeout = 0, i; + unsigned int data[2]; + struct vdec_lat_buf *lat_buf; + struct vdec_h264_slice_share_info *share_info; + unsigned char *buf; + struct mtk_vcodec_mem *mem; + + if (vdec_msg_queue_init(&inst->ctx->msg_queue, inst->ctx, + vdec_h264_slice_core_decode, + sizeof(*share_info))) + return -ENOMEM; + + /* bs NULL means flush decoder */ + if (!bs) { + vdec_msg_queue_wait_lat_buf_full(&inst->ctx->msg_queue); + return vpu_dec_reset(vpu); + } + + if (inst->is_field_bitstream) + return -EINVAL; + + lat_buf = vdec_msg_queue_dqbuf(&inst->ctx->msg_queue.lat_ctx); + if (!lat_buf) { + mtk_vdec_debug(inst->ctx, "failed to get lat buffer"); + return -EAGAIN; + } + share_info = lat_buf->private_data; + src_buf_info = container_of(bs, struct mtk_video_dec_buf, bs_buffer); + + buf = (unsigned char *)bs->va; + nal_start_idx = mtk_vdec_h264_find_start_code(buf, bs->size); + if (nal_start_idx < 0) { + err = -EINVAL; + goto err_free_fb_out; + } + + inst->vsi->dec.nal_info = buf[nal_start_idx]; + lat_buf->src_buf_req = src_buf_info->m2m_buf.vb.vb2_buf.req_obj.req; + v4l2_m2m_buf_copy_metadata(&src_buf_info->m2m_buf.vb, &lat_buf->ts_info, true); + + err = vdec_h264_slice_fill_decode_parameters(inst, share_info); + if (err) + goto err_free_fb_out; + + vdec_h264_insert_startcode(inst->ctx->dev, buf, &bs->size, + &share_info->h264_slice_params.pps); + + inst->vsi->dec.bs_buf_addr = (uint64_t)bs->dma_addr; + inst->vsi->dec.bs_buf_size = bs->size; + + *res_chg = inst->resolution_changed; + if (inst->resolution_changed) { + mtk_vdec_debug(inst->ctx, "- resolution changed -"); + if (inst->realloc_mv_buf) { + err = vdec_h264_slice_alloc_mv_buf(inst, &inst->ctx->picinfo); + inst->realloc_mv_buf = false; + if (err) + goto err_free_fb_out; + } + inst->resolution_changed = false; + } + for (i = 0; i < H264_MAX_MV_NUM; i++) { + mem = &inst->mv_buf[i]; + inst->vsi->mv_buf_dma[i] = mem->dma_addr; + } + inst->vsi->wdma_start_addr = lat_buf->ctx->msg_queue.wdma_addr.dma_addr; + inst->vsi->wdma_end_addr = lat_buf->ctx->msg_queue.wdma_addr.dma_addr + + lat_buf->ctx->msg_queue.wdma_addr.size; + inst->vsi->wdma_err_addr = lat_buf->wdma_err_addr.dma_addr; + inst->vsi->slice_bc_start_addr = lat_buf->slice_bc_addr.dma_addr; + inst->vsi->slice_bc_end_addr = lat_buf->slice_bc_addr.dma_addr + + lat_buf->slice_bc_addr.size; + + inst->vsi->trans_end = inst->ctx->msg_queue.wdma_rptr_addr; + inst->vsi->trans_start = inst->ctx->msg_queue.wdma_wptr_addr; + mtk_vdec_debug(inst->ctx, "lat:trans(0x%llx 0x%llx) err:0x%llx", + inst->vsi->wdma_start_addr, + inst->vsi->wdma_end_addr, + inst->vsi->wdma_err_addr); + + mtk_vdec_debug(inst->ctx, "slice(0x%llx 0x%llx) rprt((0x%llx 0x%llx))", + inst->vsi->slice_bc_start_addr, + inst->vsi->slice_bc_end_addr, + inst->vsi->trans_start, + inst->vsi->trans_end); + err = vpu_dec_start(vpu, data, 2); + if (err) { + mtk_vdec_debug(inst->ctx, "lat decode err: %d", err); + goto err_free_fb_out; + } + + share_info->trans_end = inst->ctx->msg_queue.wdma_addr.dma_addr + + inst->vsi->wdma_end_addr_offset; + share_info->trans_start = inst->ctx->msg_queue.wdma_wptr_addr; + share_info->nal_info = inst->vsi->dec.nal_info; + + if (IS_VDEC_INNER_RACING(inst->ctx->dev->dec_capability)) { + memcpy(&share_info->h264_slice_params, &inst->vsi->h264_slice_params, + sizeof(share_info->h264_slice_params)); + vdec_msg_queue_qbuf(&inst->ctx->msg_queue.core_ctx, lat_buf); + } + + /* wait decoder done interrupt */ + timeout = mtk_vcodec_wait_for_done_ctx(inst->ctx, MTK_INST_IRQ_RECEIVED, + WAIT_INTR_TIMEOUT_MS, MTK_VDEC_LAT0); + if (timeout) + mtk_vdec_err(inst->ctx, "lat decode timeout: pic_%d", inst->slice_dec_num); + inst->vsi->dec.timeout = !!timeout; + + err = vpu_dec_end(vpu); + if (err == SLICE_HEADER_FULL || err == TRANS_BUFFER_FULL) { + if (!IS_VDEC_INNER_RACING(inst->ctx->dev->dec_capability)) + vdec_msg_queue_qbuf(&inst->ctx->msg_queue.lat_ctx, lat_buf); + inst->slice_dec_num++; + mtk_vdec_err(inst->ctx, "lat dec fail: pic_%d err:%d", inst->slice_dec_num, err); + return -EINVAL; + } + + share_info->trans_end = inst->ctx->msg_queue.wdma_addr.dma_addr + + inst->vsi->wdma_end_addr_offset; + vdec_msg_queue_update_ube_wptr(&lat_buf->ctx->msg_queue, share_info->trans_end); + + if (!IS_VDEC_INNER_RACING(inst->ctx->dev->dec_capability)) { + memcpy(&share_info->h264_slice_params, &inst->vsi->h264_slice_params, + sizeof(share_info->h264_slice_params)); + vdec_msg_queue_qbuf(&inst->ctx->msg_queue.core_ctx, lat_buf); + } + mtk_vdec_debug(inst->ctx, "dec num: %d lat crc: 0x%x 0x%x 0x%x", inst->slice_dec_num, + inst->vsi->dec.crc[0], inst->vsi->dec.crc[1], inst->vsi->dec.crc[2]); + + inst->slice_dec_num++; + return 0; +err_free_fb_out: + vdec_msg_queue_qbuf(&inst->ctx->msg_queue.lat_ctx, lat_buf); + mtk_vdec_err(inst->ctx, "slice dec number: %d err: %d", inst->slice_dec_num, err); + return err; +} + +static int vdec_h264_slice_single_decode(void *h_vdec, struct mtk_vcodec_mem *bs, + struct vdec_fb *unused, bool *res_chg) +{ + struct vdec_h264_slice_inst *inst = h_vdec; + struct vdec_vpu_inst *vpu = &inst->vpu; + struct mtk_video_dec_buf *src_buf_info, *dst_buf_info; + struct vdec_fb *fb; + unsigned char *buf; + unsigned int data[2], i; + u64 y_fb_dma, c_fb_dma; + struct mtk_vcodec_mem *mem; + int err, nal_start_idx; + + /* bs NULL means flush decoder */ + if (!bs) + return vpu_dec_reset(vpu); + + fb = inst->ctx->dev->vdec_pdata->get_cap_buffer(inst->ctx); + src_buf_info = container_of(bs, struct mtk_video_dec_buf, bs_buffer); + dst_buf_info = container_of(fb, struct mtk_video_dec_buf, frame_buffer); + + y_fb_dma = fb ? (u64)fb->base_y.dma_addr : 0; + c_fb_dma = fb ? (u64)fb->base_c.dma_addr : 0; + mtk_vdec_debug(inst->ctx, "[h264-dec] [%d] y_dma=%llx c_dma=%llx", + inst->ctx->decoded_frame_cnt, y_fb_dma, c_fb_dma); + + inst->vsi_ctx.dec.bs_buf_addr = (u64)bs->dma_addr; + inst->vsi_ctx.dec.bs_buf_size = bs->size; + inst->vsi_ctx.dec.y_fb_dma = y_fb_dma; + inst->vsi_ctx.dec.c_fb_dma = c_fb_dma; + inst->vsi_ctx.dec.vdec_fb_va = (u64)(uintptr_t)fb; + + v4l2_m2m_buf_copy_metadata(&src_buf_info->m2m_buf.vb, + &dst_buf_info->m2m_buf.vb, true); + err = get_vdec_sig_decode_parameters(inst); + if (err) + goto err_free_fb_out; + + buf = (unsigned char *)bs->va; + nal_start_idx = mtk_vdec_h264_find_start_code(buf, bs->size); + if (nal_start_idx < 0) { + err = -EINVAL; + goto err_free_fb_out; + } + inst->vsi_ctx.dec.nal_info = buf[nal_start_idx]; + + *res_chg = inst->resolution_changed; + if (inst->resolution_changed) { + mtk_vdec_debug(inst->ctx, "- resolution changed -"); + if (inst->realloc_mv_buf) { + err = vdec_h264_slice_alloc_mv_buf(inst, &inst->ctx->picinfo); + inst->realloc_mv_buf = false; + if (err) + goto err_free_fb_out; + } + inst->resolution_changed = false; + + for (i = 0; i < H264_MAX_MV_NUM; i++) { + mem = &inst->mv_buf[i]; + inst->vsi_ctx.mv_buf_dma[i] = mem->dma_addr; + } + } + + memcpy(inst->vpu.vsi, &inst->vsi_ctx, sizeof(inst->vsi_ctx)); + err = vpu_dec_start(vpu, data, 2); + if (err) + goto err_free_fb_out; + + /* wait decoder done interrupt */ + err = mtk_vcodec_wait_for_done_ctx(inst->ctx, MTK_INST_IRQ_RECEIVED, + WAIT_INTR_TIMEOUT_MS, MTK_VDEC_CORE); + if (err) + mtk_vdec_err(inst->ctx, "decode timeout: pic_%d", inst->ctx->decoded_frame_cnt); + + inst->vsi->dec.timeout = !!err; + err = vpu_dec_end(vpu); + if (err) + goto err_free_fb_out; + + memcpy(&inst->vsi_ctx, inst->vpu.vsi, sizeof(inst->vsi_ctx)); + mtk_vdec_debug(inst->ctx, "pic[%d] crc: 0x%x 0x%x 0x%x 0x%x 0x%x 0x%x 0x%x 0x%x", + inst->ctx->decoded_frame_cnt, + inst->vsi_ctx.dec.crc[0], inst->vsi_ctx.dec.crc[1], + inst->vsi_ctx.dec.crc[2], inst->vsi_ctx.dec.crc[3], + inst->vsi_ctx.dec.crc[4], inst->vsi_ctx.dec.crc[5], + inst->vsi_ctx.dec.crc[6], inst->vsi_ctx.dec.crc[7]); + + inst->ctx->decoded_frame_cnt++; + return 0; + +err_free_fb_out: + mtk_vdec_err(inst->ctx, "dec frame number: %d err: %d", inst->ctx->decoded_frame_cnt, err); + return err; +} + +static int vdec_h264_slice_decode(void *h_vdec, struct mtk_vcodec_mem *bs, + struct vdec_fb *unused, bool *res_chg) +{ + struct vdec_h264_slice_inst *inst = h_vdec; + int ret; + + if (!h_vdec) + return -EINVAL; + + if (inst->ctx->dev->vdec_pdata->hw_arch == MTK_VDEC_PURE_SINGLE_CORE) + ret = vdec_h264_slice_single_decode(h_vdec, bs, unused, res_chg); + else + ret = vdec_h264_slice_lat_decode(h_vdec, bs, unused, res_chg); + + return ret; +} + +static int vdec_h264_slice_get_param(void *h_vdec, enum vdec_get_param_type type, + void *out) +{ + struct vdec_h264_slice_inst *inst = h_vdec; + + switch (type) { + case GET_PARAM_PIC_INFO: + vdec_h264_slice_get_pic_info(inst); + break; + case GET_PARAM_DPB_SIZE: + *(unsigned int *)out = 6; + break; + case GET_PARAM_CROP_INFO: + vdec_h264_slice_get_crop_info(inst, out); + break; + default: + mtk_vdec_err(inst->ctx, "invalid get parameter type=%d", type); + return -EINVAL; + } + return 0; +} + +const struct vdec_common_if vdec_h264_slice_multi_if = { + .init = vdec_h264_slice_init, + .decode = vdec_h264_slice_decode, + .get_param = vdec_h264_slice_get_param, + .deinit = vdec_h264_slice_deinit, +}; diff --git a/drivers/media/platform/mediatek/vcodec/decoder/vdec/vdec_hevc_req_multi_if.c b/drivers/media/platform/mediatek/vcodec/decoder/vdec/vdec_hevc_req_multi_if.c new file mode 100644 index 0000000000..06ed47df69 --- /dev/null +++ b/drivers/media/platform/mediatek/vcodec/decoder/vdec/vdec_hevc_req_multi_if.c @@ -0,0 +1,1092 @@ +// SPDX-License-Identifier: GPL-2.0 +/* + * Copyright (c) 2023 MediaTek Inc. + * Author: Yunfei Dong <yunfei.dong@mediatek.com> + */ + +#include <linux/module.h> +#include <linux/slab.h> +#include <media/videobuf2-dma-contig.h> + +#include "../mtk_vcodec_dec.h" +#include "../../common/mtk_vcodec_intr.h" +#include "../vdec_drv_base.h" +#include "../vdec_drv_if.h" +#include "../vdec_vpu_if.h" + +/* the size used to store hevc wrap information */ +#define VDEC_HEVC_WRAP_SZ (532 * SZ_1K) + +#define HEVC_MAX_MV_NUM 32 + +/* get used parameters for sps/pps */ +#define GET_HEVC_VDEC_FLAG(cond, flag) \ + { dst_param->cond = ((src_param->flags & (flag)) ? (1) : (0)); } +#define GET_HEVC_VDEC_PARAM(param) \ + { dst_param->param = src_param->param; } + +/** + * enum vdec_hevc_core_dec_err_type - core decode error type + * + * @TRANS_BUFFER_FULL: trans buffer is full + * @SLICE_HEADER_FULL: slice header buffer is full + */ +enum vdec_hevc_core_dec_err_type { + TRANS_BUFFER_FULL = 1, + SLICE_HEADER_FULL, +}; + +/** + * struct mtk_hevc_dpb_info - hevc dpb information + * + * @y_dma_addr: Y plane physical address + * @c_dma_addr: CbCr plane physical address + * @reference_flag: reference picture flag (short/long term reference picture) + * @field: field picture flag + */ +struct mtk_hevc_dpb_info { + dma_addr_t y_dma_addr; + dma_addr_t c_dma_addr; + int reference_flag; + int field; +}; + +/* + * struct mtk_hevc_sps_param - parameters for sps + */ +struct mtk_hevc_sps_param { + unsigned char video_parameter_set_id; + unsigned char seq_parameter_set_id; + unsigned short pic_width_in_luma_samples; + unsigned short pic_height_in_luma_samples; + unsigned char bit_depth_luma_minus8; + unsigned char bit_depth_chroma_minus8; + unsigned char log2_max_pic_order_cnt_lsb_minus4; + unsigned char sps_max_dec_pic_buffering_minus1; + unsigned char sps_max_num_reorder_pics; + unsigned char sps_max_latency_increase_plus1; + unsigned char log2_min_luma_coding_block_size_minus3; + unsigned char log2_diff_max_min_luma_coding_block_size; + unsigned char log2_min_luma_transform_block_size_minus2; + unsigned char log2_diff_max_min_luma_transform_block_size; + unsigned char max_transform_hierarchy_depth_inter; + unsigned char max_transform_hierarchy_depth_intra; + unsigned char pcm_sample_bit_depth_luma_minus1; + unsigned char pcm_sample_bit_depth_chroma_minus1; + unsigned char log2_min_pcm_luma_coding_block_size_minus3; + unsigned char log2_diff_max_min_pcm_luma_coding_block_size; + unsigned char num_short_term_ref_pic_sets; + unsigned char num_long_term_ref_pics_sps; + unsigned char chroma_format_idc; + unsigned char sps_max_sub_layers_minus1; + unsigned char separate_colour_plane; + unsigned char scaling_list_enabled; + unsigned char amp_enabled; + unsigned char sample_adaptive_offset; + unsigned char pcm_enabled; + unsigned char pcm_loop_filter_disabled; + unsigned char long_term_ref_pics_enabled; + unsigned char sps_temporal_mvp_enabled; + unsigned char strong_intra_smoothing_enabled; + unsigned char reserved[5]; +}; + +/* + * struct mtk_hevc_pps_param - parameters for pps + */ +struct mtk_hevc_pps_param { + unsigned char pic_parameter_set_id; + unsigned char num_extra_slice_header_bits; + unsigned char num_ref_idx_l0_default_active_minus1; + unsigned char num_ref_idx_l1_default_active_minus1; + char init_qp_minus26; + unsigned char diff_cu_qp_delta_depth; + char pps_cb_qp_offset; + char pps_cr_qp_offset; + unsigned char num_tile_columns_minus1; + unsigned char num_tile_rows_minus1; + unsigned char column_width_minus1[20]; + unsigned char row_height_minus1[22]; + char pps_beta_offset_div2; + char pps_tc_offset_div2; + unsigned char log2_parallel_merge_level_minus2; + char dependent_slice_segment_enabled; + char output_flag_present; + char sign_data_hiding_enabled; + char cabac_init_present; + char constrained_intra_pred; + char transform_skip_enabled; + char cu_qp_delta_enabled; + char pps_slice_chroma_qp_offsets_present; + char weighted_pred; + char weighted_bipred; + char transquant_bypass_enabled; + char pps_flag_tiles_enabled; + char entropy_coding_sync_enabled; + char loop_filter_across_tiles_enabled; + char pps_loop_filter_across_slices_enabled; + char deblocking_filter_override_enabled; + char pps_disable_deflocking_filter; + char lists_modification_present; + char slice_segment_header_extersion_present; + char deblocking_filter_control_present; + char uniform_spacing; + char reserved[6]; +}; + +/* + * struct mtk_hevc_slice_header_param - parameters for slice header + */ +struct mtk_hevc_slice_header_param { + unsigned int slice_type; + unsigned int num_active_ref_layer_pics; + int slice_qp; + int slice_qp_delta_cb; + int slice_qp_delta_cr; + int num_ref_idx[3]; + unsigned int col_ref_idx; + unsigned int five_minus_max_num_merge_cand; + int slice_deblocking_filter_beta_offset_div2; + int slice_deblocking_filter_tc_offset_div2; + unsigned char sao_enable_flag; + unsigned char sao_enable_flag_chroma; + unsigned char cabac_init_flag; + unsigned char slice_tmvp_flags_present; + unsigned char col_from_l0_flag; + unsigned char mvd_l1_zero_flag; + unsigned char slice_loop_filter_across_slices_enabled_flag; + unsigned char deblocking_filter_disable_flag; + unsigned int slice_reg0; + unsigned int slice_reg1; + unsigned int slice_reg2; + unsigned int num_rps_curr_temp_list; + unsigned int ref_list_mode; + int str_num_delta_pocs; + int str_num_negtive_pos_pics; + int num_long_term; + int num_long_term_sps; + unsigned int max_cu_width; + unsigned int max_cu_height; + unsigned int num_entry_point_offsets; + unsigned int last_lcu_x_in_tile[17]; + unsigned int last_lcu_y_in_tile[17]; + unsigned char nal_unit_type; +}; + +/* + * struct slice_api_hevc_scaling_matrix - parameters for scaling list + */ +struct slice_api_hevc_scaling_matrix { + unsigned char scaling_list_4x4[6][16]; + unsigned char scaling_list_8x8[6][64]; + unsigned char scaling_list_16x16[6][64]; + unsigned char scaling_list_32x32[2][64]; + unsigned char scaling_list_dc_coef_16x16[6]; + unsigned char scaling_list_dc_coef_32x32[2]; +}; + +/* + * struct slice_hevc_dpb_entry - each dpb information + */ +struct slice_hevc_dpb_entry { + u64 timestamp; + unsigned char flags; + unsigned char field_pic; + int pic_order_cnt_val; +}; + +/* + * struct slice_api_hevc_decode_param - parameters for decode. + */ +struct slice_api_hevc_decode_param { + struct slice_hevc_dpb_entry dpb[V4L2_HEVC_DPB_ENTRIES_NUM_MAX]; + int pic_order_cnt_val; + unsigned short short_term_ref_pic_set_size; + unsigned short long_term_ref_pic_set_size; + unsigned char num_active_dpb_entries; + unsigned char num_poc_st_curr_before; + unsigned char num_poc_st_curr_after; + unsigned char num_poc_lt_curr; + unsigned char poc_st_curr_before[V4L2_HEVC_DPB_ENTRIES_NUM_MAX]; + unsigned char poc_st_curr_after[V4L2_HEVC_DPB_ENTRIES_NUM_MAX]; + unsigned char poc_lt_curr[V4L2_HEVC_DPB_ENTRIES_NUM_MAX]; + unsigned char num_delta_pocs_of_ref_rps_idx; + int flags; +}; + +/** + * struct hevc_fb - hevc decode frame buffer information + * + * @vdec_fb_va: virtual address of struct vdec_fb + * @y_fb_dma: dma address of Y frame buffer (luma) + * @c_fb_dma: dma address of C frame buffer (chroma) + * @poc: picture order count of frame buffer + * @reserved: for 8 bytes alignment + */ +struct hevc_fb { + u64 vdec_fb_va; + u64 y_fb_dma; + u64 c_fb_dma; + s32 poc; + u32 reserved; +}; + +/** + * struct vdec_hevc_slice_lat_dec_param - parameters for decode current frame + * + * @sps: hevc sps syntax parameters + * @pps: hevc pps syntax parameters + * @slice_header: hevc slice header syntax parameters + * @scaling_matrix: hevc scaling list parameters + * @decode_params: decoder parameters of each frame used for hardware decode + * @hevc_dpb_info: dpb reference list + */ +struct vdec_hevc_slice_lat_dec_param { + struct mtk_hevc_sps_param sps; + struct mtk_hevc_pps_param pps; + struct mtk_hevc_slice_header_param slice_header; + struct slice_api_hevc_scaling_matrix scaling_matrix; + struct slice_api_hevc_decode_param decode_params; + struct mtk_hevc_dpb_info hevc_dpb_info[V4L2_HEVC_DPB_ENTRIES_NUM_MAX]; +}; + +/** + * struct vdec_hevc_slice_info - decode information + * + * @wdma_end_addr_offset: wdma end address offset + * @timeout: Decode timeout: 1 timeout, 0 no timeount + * @vdec_fb_va: VDEC frame buffer struct virtual address + * @crc: Used to check whether hardware's status is right + */ +struct vdec_hevc_slice_info { + u64 wdma_end_addr_offset; + u64 timeout; + u64 vdec_fb_va; + u32 crc[8]; +}; + +/* + * struct vdec_hevc_slice_mem - memory address and size + */ +struct vdec_hevc_slice_mem { + union { + u64 buf; + dma_addr_t dma_addr; + }; + union { + size_t size; + dma_addr_t dma_addr_end; + u64 padding; + }; +}; + +/** + * struct vdec_hevc_slice_fb - frame buffer for decoding + * @y: current y buffer address info + * @c: current c buffer address info + */ +struct vdec_hevc_slice_fb { + struct vdec_hevc_slice_mem y; + struct vdec_hevc_slice_mem c; +}; + +/** + * struct vdec_hevc_slice_vsi - shared memory for decode information exchange + * between SCP and Host. + * + * @bs: input buffer info + * + * @ube: ube buffer + * @trans: transcoded buffer + * @err_map: err map buffer + * @slice_bc: slice bc buffer + * @wrap: temp buffer + * + * @fb: current y/c buffer + * @mv_buf_dma: HW working motion vector buffer + * @dec: decode information (AP-R, VPU-W) + * @hevc_slice_params: decode parameters for hw used + */ +struct vdec_hevc_slice_vsi { + /* used in LAT stage */ + struct vdec_hevc_slice_mem bs; + + struct vdec_hevc_slice_mem ube; + struct vdec_hevc_slice_mem trans; + struct vdec_hevc_slice_mem err_map; + struct vdec_hevc_slice_mem slice_bc; + struct vdec_hevc_slice_mem wrap; + + struct vdec_hevc_slice_fb fb; + struct vdec_hevc_slice_mem mv_buf_dma[HEVC_MAX_MV_NUM]; + struct vdec_hevc_slice_info dec; + struct vdec_hevc_slice_lat_dec_param hevc_slice_params; +}; + +/** + * struct vdec_hevc_slice_share_info - shared information used to exchange + * message between lat and core + * + * @sps: sequence header information from user space + * @dec_params: decoder params from user space + * @hevc_slice_params: decoder params used for hardware + * @trans: trans buffer dma address + */ +struct vdec_hevc_slice_share_info { + struct v4l2_ctrl_hevc_sps sps; + struct v4l2_ctrl_hevc_decode_params dec_params; + struct vdec_hevc_slice_lat_dec_param hevc_slice_params; + struct vdec_hevc_slice_mem trans; +}; + +/** + * struct vdec_hevc_slice_inst - hevc decoder instance + * + * @slice_dec_num: how many picture be decoded + * @ctx: point to mtk_vcodec_dec_ctx + * @mv_buf: HW working motion vector buffer + * @vpu: VPU instance + * @vsi: vsi used for lat + * @vsi_core: vsi used for core + * @wrap_addr: wrap address used for hevc + * + * @hevc_slice_param: the parameters that hardware use to decode + * + * @resolution_changed: resolution changed + * @realloc_mv_buf: reallocate mv buffer + * @cap_num_planes: number of capture queue plane + */ +struct vdec_hevc_slice_inst { + unsigned int slice_dec_num; + struct mtk_vcodec_dec_ctx *ctx; + struct mtk_vcodec_mem mv_buf[HEVC_MAX_MV_NUM]; + struct vdec_vpu_inst vpu; + struct vdec_hevc_slice_vsi *vsi; + struct vdec_hevc_slice_vsi *vsi_core; + struct mtk_vcodec_mem wrap_addr; + + struct vdec_hevc_slice_lat_dec_param hevc_slice_param; + + unsigned int resolution_changed; + unsigned int realloc_mv_buf; + unsigned int cap_num_planes; +}; + +static unsigned int vdec_hevc_get_mv_buf_size(unsigned int width, unsigned int height) +{ + const unsigned int unit_size = (width / 16) * (height / 16) + 8; + + return 64 * unit_size; +} + +static void *vdec_hevc_get_ctrl_ptr(struct mtk_vcodec_dec_ctx *ctx, int id) +{ + struct v4l2_ctrl *ctrl = v4l2_ctrl_find(&ctx->ctrl_hdl, id); + + if (!ctrl) + return ERR_PTR(-EINVAL); + + return ctrl->p_cur.p; +} + +static void vdec_hevc_fill_dpb_info(struct mtk_vcodec_dec_ctx *ctx, + struct slice_api_hevc_decode_param *decode_params, + struct mtk_hevc_dpb_info *hevc_dpb_info) +{ + const struct slice_hevc_dpb_entry *dpb; + struct vb2_queue *vq; + struct vb2_buffer *vb; + int index; + + vq = v4l2_m2m_get_vq(ctx->m2m_ctx, V4L2_BUF_TYPE_VIDEO_CAPTURE_MPLANE); + for (index = 0; index < V4L2_HEVC_DPB_ENTRIES_NUM_MAX; index++) { + dpb = &decode_params->dpb[index]; + if (index >= decode_params->num_active_dpb_entries) + continue; + + vb = vb2_find_buffer(vq, dpb->timestamp); + if (!vb) { + dev_err(&ctx->dev->plat_dev->dev, + "Reference invalid: dpb_index(%d) timestamp(%lld)", + index, dpb->timestamp); + continue; + } + + hevc_dpb_info[index].field = dpb->field_pic; + + hevc_dpb_info[index].y_dma_addr = vb2_dma_contig_plane_dma_addr(vb, 0); + if (ctx->q_data[MTK_Q_DATA_DST].fmt->num_planes == 2) + hevc_dpb_info[index].c_dma_addr = vb2_dma_contig_plane_dma_addr(vb, 1); + else + hevc_dpb_info[index].c_dma_addr = + hevc_dpb_info[index].y_dma_addr + ctx->picinfo.fb_sz[0]; + } +} + +static void vdec_hevc_copy_sps_params(struct mtk_hevc_sps_param *dst_param, + const struct v4l2_ctrl_hevc_sps *src_param) +{ + GET_HEVC_VDEC_PARAM(video_parameter_set_id); + GET_HEVC_VDEC_PARAM(seq_parameter_set_id); + GET_HEVC_VDEC_PARAM(pic_width_in_luma_samples); + GET_HEVC_VDEC_PARAM(pic_height_in_luma_samples); + GET_HEVC_VDEC_PARAM(bit_depth_luma_minus8); + GET_HEVC_VDEC_PARAM(bit_depth_chroma_minus8); + GET_HEVC_VDEC_PARAM(log2_max_pic_order_cnt_lsb_minus4); + GET_HEVC_VDEC_PARAM(sps_max_dec_pic_buffering_minus1); + GET_HEVC_VDEC_PARAM(sps_max_num_reorder_pics); + GET_HEVC_VDEC_PARAM(sps_max_latency_increase_plus1); + GET_HEVC_VDEC_PARAM(log2_min_luma_coding_block_size_minus3); + GET_HEVC_VDEC_PARAM(log2_diff_max_min_luma_coding_block_size); + GET_HEVC_VDEC_PARAM(log2_min_luma_transform_block_size_minus2); + GET_HEVC_VDEC_PARAM(log2_diff_max_min_luma_transform_block_size); + GET_HEVC_VDEC_PARAM(max_transform_hierarchy_depth_inter); + GET_HEVC_VDEC_PARAM(max_transform_hierarchy_depth_intra); + GET_HEVC_VDEC_PARAM(pcm_sample_bit_depth_luma_minus1); + GET_HEVC_VDEC_PARAM(pcm_sample_bit_depth_chroma_minus1); + GET_HEVC_VDEC_PARAM(log2_min_pcm_luma_coding_block_size_minus3); + GET_HEVC_VDEC_PARAM(log2_diff_max_min_pcm_luma_coding_block_size); + GET_HEVC_VDEC_PARAM(num_short_term_ref_pic_sets); + GET_HEVC_VDEC_PARAM(num_long_term_ref_pics_sps); + GET_HEVC_VDEC_PARAM(chroma_format_idc); + GET_HEVC_VDEC_PARAM(sps_max_sub_layers_minus1); + + GET_HEVC_VDEC_FLAG(separate_colour_plane, + V4L2_HEVC_SPS_FLAG_SEPARATE_COLOUR_PLANE); + GET_HEVC_VDEC_FLAG(scaling_list_enabled, + V4L2_HEVC_SPS_FLAG_SCALING_LIST_ENABLED); + GET_HEVC_VDEC_FLAG(amp_enabled, + V4L2_HEVC_SPS_FLAG_AMP_ENABLED); + GET_HEVC_VDEC_FLAG(sample_adaptive_offset, + V4L2_HEVC_SPS_FLAG_SAMPLE_ADAPTIVE_OFFSET); + GET_HEVC_VDEC_FLAG(pcm_enabled, + V4L2_HEVC_SPS_FLAG_PCM_ENABLED); + GET_HEVC_VDEC_FLAG(pcm_loop_filter_disabled, + V4L2_HEVC_SPS_FLAG_PCM_LOOP_FILTER_DISABLED); + GET_HEVC_VDEC_FLAG(long_term_ref_pics_enabled, + V4L2_HEVC_SPS_FLAG_LONG_TERM_REF_PICS_PRESENT); + GET_HEVC_VDEC_FLAG(sps_temporal_mvp_enabled, + V4L2_HEVC_SPS_FLAG_SPS_TEMPORAL_MVP_ENABLED); + GET_HEVC_VDEC_FLAG(strong_intra_smoothing_enabled, + V4L2_HEVC_SPS_FLAG_STRONG_INTRA_SMOOTHING_ENABLED); +} + +static void vdec_hevc_copy_pps_params(struct mtk_hevc_pps_param *dst_param, + const struct v4l2_ctrl_hevc_pps *src_param) +{ + int i; + + GET_HEVC_VDEC_PARAM(pic_parameter_set_id); + GET_HEVC_VDEC_PARAM(num_extra_slice_header_bits); + GET_HEVC_VDEC_PARAM(num_ref_idx_l0_default_active_minus1); + GET_HEVC_VDEC_PARAM(num_ref_idx_l1_default_active_minus1); + GET_HEVC_VDEC_PARAM(init_qp_minus26); + GET_HEVC_VDEC_PARAM(diff_cu_qp_delta_depth); + GET_HEVC_VDEC_PARAM(pps_cb_qp_offset); + GET_HEVC_VDEC_PARAM(pps_cr_qp_offset); + GET_HEVC_VDEC_PARAM(num_tile_columns_minus1); + GET_HEVC_VDEC_PARAM(num_tile_rows_minus1); + GET_HEVC_VDEC_PARAM(init_qp_minus26); + GET_HEVC_VDEC_PARAM(diff_cu_qp_delta_depth); + GET_HEVC_VDEC_PARAM(pic_parameter_set_id); + GET_HEVC_VDEC_PARAM(num_extra_slice_header_bits); + GET_HEVC_VDEC_PARAM(num_ref_idx_l0_default_active_minus1); + GET_HEVC_VDEC_PARAM(num_ref_idx_l1_default_active_minus1); + GET_HEVC_VDEC_PARAM(pps_beta_offset_div2); + GET_HEVC_VDEC_PARAM(pps_tc_offset_div2); + GET_HEVC_VDEC_PARAM(log2_parallel_merge_level_minus2); + + for (i = 0; i < ARRAY_SIZE(src_param->column_width_minus1); i++) + GET_HEVC_VDEC_PARAM(column_width_minus1[i]); + for (i = 0; i < ARRAY_SIZE(src_param->row_height_minus1); i++) + GET_HEVC_VDEC_PARAM(row_height_minus1[i]); + + GET_HEVC_VDEC_FLAG(dependent_slice_segment_enabled, + V4L2_HEVC_PPS_FLAG_DEPENDENT_SLICE_SEGMENT_ENABLED); + GET_HEVC_VDEC_FLAG(output_flag_present, + V4L2_HEVC_PPS_FLAG_OUTPUT_FLAG_PRESENT); + GET_HEVC_VDEC_FLAG(sign_data_hiding_enabled, + V4L2_HEVC_PPS_FLAG_SIGN_DATA_HIDING_ENABLED); + GET_HEVC_VDEC_FLAG(cabac_init_present, + V4L2_HEVC_PPS_FLAG_CABAC_INIT_PRESENT); + GET_HEVC_VDEC_FLAG(constrained_intra_pred, + V4L2_HEVC_PPS_FLAG_CONSTRAINED_INTRA_PRED); + GET_HEVC_VDEC_FLAG(transform_skip_enabled, + V4L2_HEVC_PPS_FLAG_TRANSFORM_SKIP_ENABLED); + GET_HEVC_VDEC_FLAG(cu_qp_delta_enabled, + V4L2_HEVC_PPS_FLAG_CU_QP_DELTA_ENABLED); + GET_HEVC_VDEC_FLAG(pps_slice_chroma_qp_offsets_present, + V4L2_HEVC_PPS_FLAG_PPS_SLICE_CHROMA_QP_OFFSETS_PRESENT); + GET_HEVC_VDEC_FLAG(weighted_pred, + V4L2_HEVC_PPS_FLAG_WEIGHTED_PRED); + GET_HEVC_VDEC_FLAG(weighted_bipred, + V4L2_HEVC_PPS_FLAG_WEIGHTED_BIPRED); + GET_HEVC_VDEC_FLAG(transquant_bypass_enabled, + V4L2_HEVC_PPS_FLAG_TRANSQUANT_BYPASS_ENABLED); + GET_HEVC_VDEC_FLAG(pps_flag_tiles_enabled, + V4L2_HEVC_PPS_FLAG_TILES_ENABLED); + GET_HEVC_VDEC_FLAG(entropy_coding_sync_enabled, + V4L2_HEVC_PPS_FLAG_ENTROPY_CODING_SYNC_ENABLED); + GET_HEVC_VDEC_FLAG(loop_filter_across_tiles_enabled, + V4L2_HEVC_PPS_FLAG_LOOP_FILTER_ACROSS_TILES_ENABLED); + GET_HEVC_VDEC_FLAG(pps_loop_filter_across_slices_enabled, + V4L2_HEVC_PPS_FLAG_PPS_LOOP_FILTER_ACROSS_SLICES_ENABLED); + GET_HEVC_VDEC_FLAG(deblocking_filter_override_enabled, + V4L2_HEVC_PPS_FLAG_DEBLOCKING_FILTER_OVERRIDE_ENABLED); + GET_HEVC_VDEC_FLAG(pps_disable_deflocking_filter, + V4L2_HEVC_PPS_FLAG_PPS_DISABLE_DEBLOCKING_FILTER); + GET_HEVC_VDEC_FLAG(lists_modification_present, + V4L2_HEVC_PPS_FLAG_LISTS_MODIFICATION_PRESENT); + GET_HEVC_VDEC_FLAG(slice_segment_header_extersion_present, + V4L2_HEVC_PPS_FLAG_SLICE_SEGMENT_HEADER_EXTENSION_PRESENT); + GET_HEVC_VDEC_FLAG(deblocking_filter_control_present, + V4L2_HEVC_PPS_FLAG_DEBLOCKING_FILTER_CONTROL_PRESENT); + GET_HEVC_VDEC_FLAG(uniform_spacing, + V4L2_HEVC_PPS_FLAG_UNIFORM_SPACING); +} + +static void vdec_hevc_copy_scaling_matrix(struct slice_api_hevc_scaling_matrix *dst_matrix, + const struct v4l2_ctrl_hevc_scaling_matrix *src_matrix) +{ + memcpy(dst_matrix, src_matrix, sizeof(*src_matrix)); +} + +static void +vdec_hevc_copy_decode_params(struct slice_api_hevc_decode_param *dst_param, + const struct v4l2_ctrl_hevc_decode_params *src_param, + const struct v4l2_hevc_dpb_entry dpb[V4L2_HEVC_DPB_ENTRIES_NUM_MAX]) +{ + struct slice_hevc_dpb_entry *dst_entry; + const struct v4l2_hevc_dpb_entry *src_entry; + int i; + + for (i = 0; i < ARRAY_SIZE(dst_param->dpb); i++) { + dst_entry = &dst_param->dpb[i]; + src_entry = &dpb[i]; + + dst_entry->timestamp = src_entry->timestamp; + dst_entry->flags = src_entry->flags; + dst_entry->field_pic = src_entry->field_pic; + dst_entry->pic_order_cnt_val = src_entry->pic_order_cnt_val; + + GET_HEVC_VDEC_PARAM(poc_st_curr_before[i]); + GET_HEVC_VDEC_PARAM(poc_st_curr_after[i]); + GET_HEVC_VDEC_PARAM(poc_lt_curr[i]); + } + + GET_HEVC_VDEC_PARAM(pic_order_cnt_val); + GET_HEVC_VDEC_PARAM(short_term_ref_pic_set_size); + GET_HEVC_VDEC_PARAM(long_term_ref_pic_set_size); + GET_HEVC_VDEC_PARAM(num_active_dpb_entries); + GET_HEVC_VDEC_PARAM(num_poc_st_curr_before); + GET_HEVC_VDEC_PARAM(num_poc_st_curr_after); + GET_HEVC_VDEC_PARAM(num_delta_pocs_of_ref_rps_idx); + GET_HEVC_VDEC_PARAM(num_poc_lt_curr); + GET_HEVC_VDEC_PARAM(flags); +} + +static int vdec_hevc_slice_fill_decode_parameters(struct vdec_hevc_slice_inst *inst, + struct vdec_hevc_slice_share_info *share_info) +{ + struct vdec_hevc_slice_lat_dec_param *slice_param = &inst->vsi->hevc_slice_params; + const struct v4l2_ctrl_hevc_decode_params *dec_params; + const struct v4l2_ctrl_hevc_scaling_matrix *src_matrix; + const struct v4l2_ctrl_hevc_sps *sps; + const struct v4l2_ctrl_hevc_pps *pps; + + dec_params = + vdec_hevc_get_ctrl_ptr(inst->ctx, V4L2_CID_STATELESS_HEVC_DECODE_PARAMS); + if (IS_ERR(dec_params)) + return PTR_ERR(dec_params); + + src_matrix = + vdec_hevc_get_ctrl_ptr(inst->ctx, V4L2_CID_STATELESS_HEVC_SCALING_MATRIX); + if (IS_ERR(src_matrix)) + return PTR_ERR(src_matrix); + + sps = vdec_hevc_get_ctrl_ptr(inst->ctx, V4L2_CID_STATELESS_HEVC_SPS); + if (IS_ERR(sps)) + return PTR_ERR(sps); + + pps = vdec_hevc_get_ctrl_ptr(inst->ctx, V4L2_CID_STATELESS_HEVC_PPS); + if (IS_ERR(pps)) + return PTR_ERR(pps); + + vdec_hevc_copy_sps_params(&slice_param->sps, sps); + vdec_hevc_copy_pps_params(&slice_param->pps, pps); + vdec_hevc_copy_scaling_matrix(&slice_param->scaling_matrix, src_matrix); + + memcpy(&share_info->sps, sps, sizeof(*sps)); + memcpy(&share_info->dec_params, dec_params, sizeof(*dec_params)); + + slice_param->decode_params.num_poc_st_curr_before = dec_params->num_poc_st_curr_before; + slice_param->decode_params.num_poc_st_curr_after = dec_params->num_poc_st_curr_after; + slice_param->decode_params.num_poc_lt_curr = dec_params->num_poc_lt_curr; + slice_param->decode_params.num_delta_pocs_of_ref_rps_idx = + dec_params->num_delta_pocs_of_ref_rps_idx; + + return 0; +} + +static void vdec_hevc_slice_fill_decode_reflist(struct vdec_hevc_slice_inst *inst, + struct vdec_hevc_slice_lat_dec_param *slice_param, + struct vdec_hevc_slice_share_info *share_info) +{ + struct v4l2_ctrl_hevc_decode_params *dec_params = &share_info->dec_params; + + vdec_hevc_copy_decode_params(&slice_param->decode_params, dec_params, + share_info->dec_params.dpb); + + vdec_hevc_fill_dpb_info(inst->ctx, &slice_param->decode_params, + slice_param->hevc_dpb_info); +} + +static int vdec_hevc_slice_alloc_mv_buf(struct vdec_hevc_slice_inst *inst, + struct vdec_pic_info *pic) +{ + unsigned int buf_sz = vdec_hevc_get_mv_buf_size(pic->buf_w, pic->buf_h); + struct mtk_vcodec_mem *mem; + int i, err; + + mtk_v4l2_vdec_dbg(3, inst->ctx, "allocate mv buffer size = 0x%x", buf_sz); + for (i = 0; i < HEVC_MAX_MV_NUM; i++) { + mem = &inst->mv_buf[i]; + if (mem->va) + mtk_vcodec_mem_free(inst->ctx, mem); + mem->size = buf_sz; + err = mtk_vcodec_mem_alloc(inst->ctx, mem); + if (err) { + mtk_vdec_err(inst->ctx, "failed to allocate mv buf"); + return err; + } + } + + return 0; +} + +static void vdec_hevc_slice_free_mv_buf(struct vdec_hevc_slice_inst *inst) +{ + int i; + struct mtk_vcodec_mem *mem; + + for (i = 0; i < HEVC_MAX_MV_NUM; i++) { + mem = &inst->mv_buf[i]; + if (mem->va) + mtk_vcodec_mem_free(inst->ctx, mem); + } +} + +static void vdec_hevc_slice_get_pic_info(struct vdec_hevc_slice_inst *inst) +{ + struct mtk_vcodec_dec_ctx *ctx = inst->ctx; + u32 data[3]; + + data[0] = ctx->picinfo.pic_w; + data[1] = ctx->picinfo.pic_h; + data[2] = ctx->capture_fourcc; + vpu_dec_get_param(&inst->vpu, data, 3, GET_PARAM_PIC_INFO); + + ctx->picinfo.buf_w = ALIGN(ctx->picinfo.pic_w, VCODEC_DEC_ALIGNED_64); + ctx->picinfo.buf_h = ALIGN(ctx->picinfo.pic_h, VCODEC_DEC_ALIGNED_64); + ctx->picinfo.fb_sz[0] = inst->vpu.fb_sz[0]; + ctx->picinfo.fb_sz[1] = inst->vpu.fb_sz[1]; + inst->cap_num_planes = + ctx->q_data[MTK_Q_DATA_DST].fmt->num_planes; + + mtk_vdec_debug(ctx, "pic(%d, %d), buf(%d, %d)", + ctx->picinfo.pic_w, ctx->picinfo.pic_h, + ctx->picinfo.buf_w, ctx->picinfo.buf_h); + mtk_vdec_debug(ctx, "Y/C(%d, %d)", ctx->picinfo.fb_sz[0], + ctx->picinfo.fb_sz[1]); + + if (ctx->last_decoded_picinfo.pic_w != ctx->picinfo.pic_w || + ctx->last_decoded_picinfo.pic_h != ctx->picinfo.pic_h) { + inst->resolution_changed = true; + if (ctx->last_decoded_picinfo.buf_w != ctx->picinfo.buf_w || + ctx->last_decoded_picinfo.buf_h != ctx->picinfo.buf_h) + inst->realloc_mv_buf = true; + + mtk_v4l2_vdec_dbg(1, inst->ctx, "resChg: (%d %d) : old(%d, %d) -> new(%d, %d)", + inst->resolution_changed, + inst->realloc_mv_buf, + ctx->last_decoded_picinfo.pic_w, + ctx->last_decoded_picinfo.pic_h, + ctx->picinfo.pic_w, ctx->picinfo.pic_h); + } +} + +static void vdec_hevc_slice_get_crop_info(struct vdec_hevc_slice_inst *inst, + struct v4l2_rect *cr) +{ + cr->left = 0; + cr->top = 0; + cr->width = inst->ctx->picinfo.pic_w; + cr->height = inst->ctx->picinfo.pic_h; + + mtk_vdec_debug(inst->ctx, "l=%d, t=%d, w=%d, h=%d", + cr->left, cr->top, cr->width, cr->height); +} + +static int vdec_hevc_slice_setup_lat_buffer(struct vdec_hevc_slice_inst *inst, + struct mtk_vcodec_mem *bs, + struct vdec_lat_buf *lat_buf, + bool *res_chg) +{ + struct mtk_vcodec_mem *mem; + struct mtk_video_dec_buf *src_buf_info; + struct vdec_hevc_slice_share_info *share_info; + int i, err; + + inst->vsi->bs.dma_addr = (u64)bs->dma_addr; + inst->vsi->bs.size = bs->size; + + src_buf_info = container_of(bs, struct mtk_video_dec_buf, bs_buffer); + lat_buf->src_buf_req = src_buf_info->m2m_buf.vb.vb2_buf.req_obj.req; + v4l2_m2m_buf_copy_metadata(&src_buf_info->m2m_buf.vb, &lat_buf->ts_info, true); + + *res_chg = inst->resolution_changed; + if (inst->resolution_changed) { + mtk_vdec_debug(inst->ctx, "- resolution changed -"); + if (inst->realloc_mv_buf) { + err = vdec_hevc_slice_alloc_mv_buf(inst, &inst->ctx->picinfo); + inst->realloc_mv_buf = false; + if (err) + return err; + } + inst->resolution_changed = false; + } + + for (i = 0; i < HEVC_MAX_MV_NUM; i++) { + mem = &inst->mv_buf[i]; + inst->vsi->mv_buf_dma[i].dma_addr = mem->dma_addr; + inst->vsi->mv_buf_dma[i].size = mem->size; + } + + inst->vsi->ube.dma_addr = lat_buf->ctx->msg_queue.wdma_addr.dma_addr; + inst->vsi->ube.size = lat_buf->ctx->msg_queue.wdma_addr.size; + + inst->vsi->err_map.dma_addr = lat_buf->wdma_err_addr.dma_addr; + inst->vsi->err_map.size = lat_buf->wdma_err_addr.size; + + inst->vsi->slice_bc.dma_addr = lat_buf->slice_bc_addr.dma_addr; + inst->vsi->slice_bc.size = lat_buf->slice_bc_addr.size; + + inst->vsi->trans.dma_addr_end = inst->ctx->msg_queue.wdma_rptr_addr; + inst->vsi->trans.dma_addr = inst->ctx->msg_queue.wdma_wptr_addr; + + share_info = lat_buf->private_data; + share_info->trans.dma_addr = inst->vsi->trans.dma_addr; + share_info->trans.dma_addr_end = inst->vsi->trans.dma_addr_end; + + mtk_vdec_debug(inst->ctx, "lat: ube addr/size(0x%llx 0x%llx) err:0x%llx", + inst->vsi->ube.buf, + inst->vsi->ube.padding, + inst->vsi->err_map.buf); + + mtk_vdec_debug(inst->ctx, "slice addr/size(0x%llx 0x%llx) trans start/end((0x%llx 0x%llx))", + inst->vsi->slice_bc.buf, + inst->vsi->slice_bc.padding, + inst->vsi->trans.buf, + inst->vsi->trans.padding); + + return 0; +} + +static int vdec_hevc_slice_setup_core_buffer(struct vdec_hevc_slice_inst *inst, + struct vdec_hevc_slice_share_info *share_info, + struct vdec_lat_buf *lat_buf) +{ + struct mtk_vcodec_mem *mem; + struct mtk_vcodec_dec_ctx *ctx = inst->ctx; + struct vb2_v4l2_buffer *vb2_v4l2; + struct vdec_fb *fb; + u64 y_fb_dma, c_fb_dma; + int i; + + fb = ctx->dev->vdec_pdata->get_cap_buffer(ctx); + if (!fb) { + mtk_vdec_err(inst->ctx, "fb buffer is NULL"); + return -EBUSY; + } + + y_fb_dma = (u64)fb->base_y.dma_addr; + if (ctx->q_data[MTK_Q_DATA_DST].fmt->num_planes == 1) + c_fb_dma = + y_fb_dma + inst->ctx->picinfo.buf_w * inst->ctx->picinfo.buf_h; + else + c_fb_dma = (u64)fb->base_c.dma_addr; + + mtk_vdec_debug(inst->ctx, "[hevc-core] y/c addr = 0x%llx 0x%llx", y_fb_dma, c_fb_dma); + + inst->vsi_core->fb.y.dma_addr = y_fb_dma; + inst->vsi_core->fb.y.size = ctx->picinfo.fb_sz[0]; + inst->vsi_core->fb.c.dma_addr = c_fb_dma; + inst->vsi_core->fb.y.size = ctx->picinfo.fb_sz[1]; + + inst->vsi_core->dec.vdec_fb_va = (unsigned long)fb; + + inst->vsi_core->ube.dma_addr = lat_buf->ctx->msg_queue.wdma_addr.dma_addr; + inst->vsi_core->ube.size = lat_buf->ctx->msg_queue.wdma_addr.size; + + inst->vsi_core->err_map.dma_addr = lat_buf->wdma_err_addr.dma_addr; + inst->vsi_core->err_map.size = lat_buf->wdma_err_addr.size; + + inst->vsi_core->slice_bc.dma_addr = lat_buf->slice_bc_addr.dma_addr; + inst->vsi_core->slice_bc.size = lat_buf->slice_bc_addr.size; + + inst->vsi_core->trans.dma_addr = share_info->trans.dma_addr; + inst->vsi_core->trans.dma_addr_end = share_info->trans.dma_addr_end; + + inst->vsi_core->wrap.dma_addr = inst->wrap_addr.dma_addr; + inst->vsi_core->wrap.size = inst->wrap_addr.size; + + for (i = 0; i < HEVC_MAX_MV_NUM; i++) { + mem = &inst->mv_buf[i]; + inst->vsi_core->mv_buf_dma[i].dma_addr = mem->dma_addr; + inst->vsi_core->mv_buf_dma[i].size = mem->size; + } + + vb2_v4l2 = v4l2_m2m_next_dst_buf(ctx->m2m_ctx); + v4l2_m2m_buf_copy_metadata(&lat_buf->ts_info, vb2_v4l2, true); + + return 0; +} + +static int vdec_hevc_slice_init(struct mtk_vcodec_dec_ctx *ctx) +{ + struct vdec_hevc_slice_inst *inst; + int err, vsi_size; + + inst = kzalloc(sizeof(*inst), GFP_KERNEL); + if (!inst) + return -ENOMEM; + + inst->ctx = ctx; + + inst->vpu.id = SCP_IPI_VDEC_LAT; + inst->vpu.core_id = SCP_IPI_VDEC_CORE; + inst->vpu.ctx = ctx; + inst->vpu.codec_type = ctx->current_codec; + inst->vpu.capture_type = ctx->capture_fourcc; + + ctx->drv_handle = inst; + err = vpu_dec_init(&inst->vpu); + if (err) { + mtk_vdec_err(ctx, "vdec_hevc init err=%d", err); + goto error_free_inst; + } + + vsi_size = round_up(sizeof(struct vdec_hevc_slice_vsi), VCODEC_DEC_ALIGNED_64); + inst->vsi = inst->vpu.vsi; + inst->vsi_core = + (struct vdec_hevc_slice_vsi *)(((char *)inst->vpu.vsi) + vsi_size); + + inst->resolution_changed = true; + inst->realloc_mv_buf = true; + + inst->wrap_addr.size = VDEC_HEVC_WRAP_SZ; + err = mtk_vcodec_mem_alloc(ctx, &inst->wrap_addr); + if (err) + goto error_free_inst; + + mtk_vdec_debug(ctx, "lat struct size = %d,%d,%d,%d vsi: %d\n", + (int)sizeof(struct mtk_hevc_sps_param), + (int)sizeof(struct mtk_hevc_pps_param), + (int)sizeof(struct vdec_hevc_slice_lat_dec_param), + (int)sizeof(struct mtk_hevc_dpb_info), + vsi_size); + mtk_vdec_debug(ctx, "lat hevc instance >> %p, codec_type = 0x%x", + inst, inst->vpu.codec_type); + + return 0; +error_free_inst: + kfree(inst); + return err; +} + +static void vdec_hevc_slice_deinit(void *h_vdec) +{ + struct vdec_hevc_slice_inst *inst = h_vdec; + struct mtk_vcodec_mem *mem; + + vpu_dec_deinit(&inst->vpu); + vdec_hevc_slice_free_mv_buf(inst); + + mem = &inst->wrap_addr; + if (mem->va) + mtk_vcodec_mem_free(inst->ctx, mem); + + vdec_msg_queue_deinit(&inst->ctx->msg_queue, inst->ctx); + kfree(inst); +} + +static int vdec_hevc_slice_core_decode(struct vdec_lat_buf *lat_buf) +{ + int err, timeout; + struct mtk_vcodec_dec_ctx *ctx = lat_buf->ctx; + struct vdec_hevc_slice_inst *inst = ctx->drv_handle; + struct vdec_hevc_slice_share_info *share_info = lat_buf->private_data; + struct vdec_vpu_inst *vpu = &inst->vpu; + + mtk_vdec_debug(ctx, "[hevc-core] vdec_hevc core decode"); + memcpy(&inst->vsi_core->hevc_slice_params, &share_info->hevc_slice_params, + sizeof(share_info->hevc_slice_params)); + + err = vdec_hevc_slice_setup_core_buffer(inst, share_info, lat_buf); + if (err) + goto vdec_dec_end; + + vdec_hevc_slice_fill_decode_reflist(inst, &inst->vsi_core->hevc_slice_params, + share_info); + err = vpu_dec_core(vpu); + if (err) { + mtk_vdec_err(ctx, "core decode err=%d", err); + goto vdec_dec_end; + } + + /* wait decoder done interrupt */ + timeout = mtk_vcodec_wait_for_done_ctx(inst->ctx, MTK_INST_IRQ_RECEIVED, + WAIT_INTR_TIMEOUT_MS, MTK_VDEC_CORE); + if (timeout) + mtk_vdec_err(ctx, "core decode timeout: pic_%d", ctx->decoded_frame_cnt); + inst->vsi_core->dec.timeout = !!timeout; + + vpu_dec_core_end(vpu); + mtk_vdec_debug(ctx, "pic[%d] crc: 0x%x 0x%x 0x%x 0x%x 0x%x 0x%x 0x%x 0x%x", + ctx->decoded_frame_cnt, + inst->vsi_core->dec.crc[0], inst->vsi_core->dec.crc[1], + inst->vsi_core->dec.crc[2], inst->vsi_core->dec.crc[3], + inst->vsi_core->dec.crc[4], inst->vsi_core->dec.crc[5], + inst->vsi_core->dec.crc[6], inst->vsi_core->dec.crc[7]); + +vdec_dec_end: + vdec_msg_queue_update_ube_rptr(&lat_buf->ctx->msg_queue, share_info->trans.dma_addr_end); + ctx->dev->vdec_pdata->cap_to_disp(ctx, !!err, lat_buf->src_buf_req); + mtk_vdec_debug(ctx, "core decode done err=%d", err); + ctx->decoded_frame_cnt++; + return 0; +} + +static int vdec_hevc_slice_lat_decode(void *h_vdec, struct mtk_vcodec_mem *bs, + struct vdec_fb *fb, bool *res_chg) +{ + struct vdec_hevc_slice_inst *inst = h_vdec; + struct vdec_vpu_inst *vpu = &inst->vpu; + int err, timeout = 0; + unsigned int data[2]; + struct vdec_lat_buf *lat_buf; + struct vdec_hevc_slice_share_info *share_info; + + if (vdec_msg_queue_init(&inst->ctx->msg_queue, inst->ctx, + vdec_hevc_slice_core_decode, + sizeof(*share_info))) + return -ENOMEM; + + /* bs NULL means flush decoder */ + if (!bs) { + vdec_msg_queue_wait_lat_buf_full(&inst->ctx->msg_queue); + return vpu_dec_reset(vpu); + } + + lat_buf = vdec_msg_queue_dqbuf(&inst->ctx->msg_queue.lat_ctx); + if (!lat_buf) { + mtk_vdec_debug(inst->ctx, "failed to get lat buffer"); + return -EAGAIN; + } + + share_info = lat_buf->private_data; + err = vdec_hevc_slice_fill_decode_parameters(inst, share_info); + if (err) + goto err_free_fb_out; + + err = vdec_hevc_slice_setup_lat_buffer(inst, bs, lat_buf, res_chg); + if (err) + goto err_free_fb_out; + + err = vpu_dec_start(vpu, data, 2); + if (err) { + mtk_vdec_debug(inst->ctx, "lat decode err: %d", err); + goto err_free_fb_out; + } + + if (IS_VDEC_INNER_RACING(inst->ctx->dev->dec_capability)) { + memcpy(&share_info->hevc_slice_params, &inst->vsi->hevc_slice_params, + sizeof(share_info->hevc_slice_params)); + vdec_msg_queue_qbuf(&inst->ctx->msg_queue.core_ctx, lat_buf); + } + + /* wait decoder done interrupt */ + timeout = mtk_vcodec_wait_for_done_ctx(inst->ctx, MTK_INST_IRQ_RECEIVED, + WAIT_INTR_TIMEOUT_MS, MTK_VDEC_LAT0); + if (timeout) + mtk_vdec_err(inst->ctx, "lat decode timeout: pic_%d", inst->slice_dec_num); + inst->vsi->dec.timeout = !!timeout; + + err = vpu_dec_end(vpu); + if (err == SLICE_HEADER_FULL || err == TRANS_BUFFER_FULL) { + if (!IS_VDEC_INNER_RACING(inst->ctx->dev->dec_capability)) + vdec_msg_queue_qbuf(&inst->ctx->msg_queue.lat_ctx, lat_buf); + inst->slice_dec_num++; + mtk_vdec_err(inst->ctx, "lat dec fail: pic_%d err:%d", inst->slice_dec_num, err); + return -EINVAL; + } + + share_info->trans.dma_addr_end = inst->ctx->msg_queue.wdma_addr.dma_addr + + inst->vsi->dec.wdma_end_addr_offset; + vdec_msg_queue_update_ube_wptr(&lat_buf->ctx->msg_queue, share_info->trans.dma_addr_end); + + if (!IS_VDEC_INNER_RACING(inst->ctx->dev->dec_capability)) { + memcpy(&share_info->hevc_slice_params, &inst->vsi->hevc_slice_params, + sizeof(share_info->hevc_slice_params)); + vdec_msg_queue_qbuf(&inst->ctx->msg_queue.core_ctx, lat_buf); + } + mtk_vdec_debug(inst->ctx, "dec num: %d lat crc: 0x%x 0x%x 0x%x", inst->slice_dec_num, + inst->vsi->dec.crc[0], inst->vsi->dec.crc[1], inst->vsi->dec.crc[2]); + + inst->slice_dec_num++; + return 0; +err_free_fb_out: + vdec_msg_queue_qbuf(&inst->ctx->msg_queue.lat_ctx, lat_buf); + mtk_vdec_err(inst->ctx, "slice dec number: %d err: %d", inst->slice_dec_num, err); + return err; +} + +static int vdec_hevc_slice_decode(void *h_vdec, struct mtk_vcodec_mem *bs, + struct vdec_fb *unused, bool *res_chg) +{ + struct vdec_hevc_slice_inst *inst = h_vdec; + + if (!h_vdec || inst->ctx->dev->vdec_pdata->hw_arch == MTK_VDEC_PURE_SINGLE_CORE) + return -EINVAL; + + return vdec_hevc_slice_lat_decode(h_vdec, bs, unused, res_chg); +} + +static int vdec_hevc_slice_get_param(void *h_vdec, enum vdec_get_param_type type, + void *out) +{ + struct vdec_hevc_slice_inst *inst = h_vdec; + + switch (type) { + case GET_PARAM_PIC_INFO: + vdec_hevc_slice_get_pic_info(inst); + break; + case GET_PARAM_DPB_SIZE: + *(unsigned int *)out = 6; + break; + case GET_PARAM_CROP_INFO: + vdec_hevc_slice_get_crop_info(inst, out); + break; + default: + mtk_vdec_err(inst->ctx, "invalid get parameter type=%d", type); + return -EINVAL; + } + return 0; +} + +const struct vdec_common_if vdec_hevc_slice_multi_if = { + .init = vdec_hevc_slice_init, + .decode = vdec_hevc_slice_decode, + .get_param = vdec_hevc_slice_get_param, + .deinit = vdec_hevc_slice_deinit, +}; diff --git a/drivers/media/platform/mediatek/vcodec/decoder/vdec/vdec_vp8_if.c b/drivers/media/platform/mediatek/vcodec/decoder/vdec/vdec_vp8_if.c new file mode 100644 index 0000000000..19407f9bc7 --- /dev/null +++ b/drivers/media/platform/mediatek/vcodec/decoder/vdec/vdec_vp8_if.c @@ -0,0 +1,608 @@ +// SPDX-License-Identifier: GPL-2.0 +/* + * Copyright (c) 2016 MediaTek Inc. + * Author: Jungchang Tsao <jungchang.tsao@mediatek.com> + * PC Chen <pc.chen@mediatek.com> + */ + +#include <linux/slab.h> +#include "../vdec_drv_if.h" +#include "../mtk_vcodec_dec.h" +#include "../../common/mtk_vcodec_intr.h" +#include "../vdec_vpu_if.h" +#include "../vdec_drv_base.h" + +/* Decoding picture buffer size (3 reference frames plus current frame) */ +#define VP8_DPB_SIZE 4 + +/* HW working buffer size (bytes) */ +#define VP8_WORKING_BUF_SZ (45 * 4096) + +/* HW control register address */ +#define VP8_SEGID_DRAM_ADDR 0x3c +#define VP8_HW_VLD_ADDR 0x93C +#define VP8_HW_VLD_VALUE 0x940 +#define VP8_BSASET 0x100 +#define VP8_BSDSET 0x104 +#define VP8_RW_CKEN_SET 0x0 +#define VP8_RW_DCM_CON 0x18 +#define VP8_WO_VLD_SRST 0x108 +#define VP8_RW_MISC_SYS_SEL 0x84 +#define VP8_RW_MISC_SPEC_CON 0xC8 +#define VP8_WO_VLD_SRST 0x108 +#define VP8_RW_VP8_CTRL 0xA4 +#define VP8_RW_MISC_DCM_CON 0xEC +#define VP8_RW_MISC_SRST 0xF4 +#define VP8_RW_MISC_FUNC_CON 0xCC + +#define VP8_MAX_FRM_BUF_NUM 5 +#define VP8_MAX_FRM_BUF_NODE_NUM (VP8_MAX_FRM_BUF_NUM * 2) + +/* required buffer size (bytes) to store decode information */ +#define VP8_HW_SEGMENT_DATA_SZ 272 +#define VP8_HW_SEGMENT_UINT 4 + +#define VP8_DEC_TABLE_PROC_LOOP 96 +#define VP8_DEC_TABLE_UNIT 3 +#define VP8_DEC_TABLE_SZ 300 +#define VP8_DEC_TABLE_OFFSET 2 +#define VP8_DEC_TABLE_RW_UNIT 4 + +/** + * struct vdec_vp8_dec_info - decode misc information + * @working_buf_dma : working buffer dma address + * @prev_y_dma : previous decoded frame buffer Y plane address + * @cur_y_fb_dma : current plane Y frame buffer dma address + * @cur_c_fb_dma : current plane C frame buffer dma address + * @bs_dma : bitstream dma address + * @bs_sz : bitstream size + * @resolution_changed: resolution change flag 1 - changed, 0 - not change + * @show_frame : display this frame or not + * @wait_key_frame : wait key frame coming + */ +struct vdec_vp8_dec_info { + uint64_t working_buf_dma; + uint64_t prev_y_dma; + uint64_t cur_y_fb_dma; + uint64_t cur_c_fb_dma; + uint64_t bs_dma; + uint32_t bs_sz; + uint32_t resolution_changed; + uint32_t show_frame; + uint32_t wait_key_frame; +}; + +/** + * struct vdec_vp8_vsi - VPU shared information + * @dec : decoding information + * @pic : picture information + * @dec_table : decoder coefficient table + * @segment_buf : segmentation buffer + * @load_data : flag to indicate reload decode data + */ +struct vdec_vp8_vsi { + struct vdec_vp8_dec_info dec; + struct vdec_pic_info pic; + uint32_t dec_table[VP8_DEC_TABLE_SZ]; + uint32_t segment_buf[VP8_HW_SEGMENT_DATA_SZ][VP8_HW_SEGMENT_UINT]; + uint32_t load_data; +}; + +/** + * struct vdec_vp8_hw_reg_base - HW register base + * @misc : base address for misc + * @ld : base address for ld + * @top : base address for top + * @cm : base address for cm + * @hwd : base address for hwd + * @hwb : base address for hwb + */ +struct vdec_vp8_hw_reg_base { + void __iomem *misc; + void __iomem *ld; + void __iomem *top; + void __iomem *cm; + void __iomem *hwd; + void __iomem *hwb; +}; + +/** + * struct vdec_vp8_vpu_inst - VPU instance for VP8 decode + * @wq_hd : Wait queue to wait VPU message ack + * @signaled : 1 - Host has received ack message from VPU, 0 - not receive + * @failure : VPU execution result status 0 - success, others - fail + * @inst_addr : VPU decoder instance address + */ +struct vdec_vp8_vpu_inst { + wait_queue_head_t wq_hd; + int signaled; + int failure; + uint32_t inst_addr; +}; + +/* frame buffer (fb) list + * [available_fb_node_list] - decode fb are initialized to 0 and populated in + * [fb_use_list] - fb is set after decode and is moved to this list + * [fb_free_list] - fb is not needed for reference will be moved from + * [fb_use_list] to [fb_free_list] and + * once user remove fb from [fb_free_list], + * it is circulated back to [available_fb_node_list] + * [fb_disp_list] - fb is set after decode and is moved to this list + * once user remove fb from [fb_disp_list] it is + * circulated back to [available_fb_node_list] + */ + +/** + * struct vdec_vp8_inst - VP8 decoder instance + * @cur_fb : current frame buffer + * @dec_fb : decode frame buffer node + * @available_fb_node_list : list to store available frame buffer node + * @fb_use_list : list to store frame buffer in use + * @fb_free_list : list to store free frame buffer + * @fb_disp_list : list to store display ready frame buffer + * @working_buf : HW decoder working buffer + * @reg_base : HW register base address + * @frm_cnt : decode frame count + * @ctx : V4L2 context + * @vpu : VPU instance for decoder + * @vsi : VPU share information + */ +struct vdec_vp8_inst { + struct vdec_fb *cur_fb; + struct vdec_fb_node dec_fb[VP8_MAX_FRM_BUF_NODE_NUM]; + struct list_head available_fb_node_list; + struct list_head fb_use_list; + struct list_head fb_free_list; + struct list_head fb_disp_list; + struct mtk_vcodec_mem working_buf; + struct vdec_vp8_hw_reg_base reg_base; + unsigned int frm_cnt; + struct mtk_vcodec_dec_ctx *ctx; + struct vdec_vpu_inst vpu; + struct vdec_vp8_vsi *vsi; +}; + +static void get_hw_reg_base(struct vdec_vp8_inst *inst) +{ + void __iomem **reg_base = inst->ctx->dev->reg_base; + + inst->reg_base.top = mtk_vcodec_get_reg_addr(reg_base, VDEC_TOP); + inst->reg_base.cm = mtk_vcodec_get_reg_addr(reg_base, VDEC_CM); + inst->reg_base.hwd = mtk_vcodec_get_reg_addr(reg_base, VDEC_HWD); + inst->reg_base.misc = mtk_vcodec_get_reg_addr(reg_base, VDEC_MISC); + inst->reg_base.ld = mtk_vcodec_get_reg_addr(reg_base, VDEC_LD); + inst->reg_base.hwb = mtk_vcodec_get_reg_addr(reg_base, VDEC_HWB); +} + +static void write_hw_segmentation_data(struct vdec_vp8_inst *inst) +{ + int i, j; + u32 seg_id_addr; + u32 val; + void __iomem *cm = inst->reg_base.cm; + struct vdec_vp8_vsi *vsi = inst->vsi; + + seg_id_addr = readl(inst->reg_base.top + VP8_SEGID_DRAM_ADDR) >> 4; + + for (i = 0; i < ARRAY_SIZE(vsi->segment_buf); i++) { + for (j = ARRAY_SIZE(vsi->segment_buf[i]) - 1; j >= 0; j--) { + val = (1 << 16) + ((seg_id_addr + i) << 2) + j; + writel(val, cm + VP8_HW_VLD_ADDR); + + val = vsi->segment_buf[i][j]; + writel(val, cm + VP8_HW_VLD_VALUE); + } + } +} + +static void read_hw_segmentation_data(struct vdec_vp8_inst *inst) +{ + int i, j; + u32 seg_id_addr; + u32 val; + void __iomem *cm = inst->reg_base.cm; + struct vdec_vp8_vsi *vsi = inst->vsi; + + seg_id_addr = readl(inst->reg_base.top + VP8_SEGID_DRAM_ADDR) >> 4; + + for (i = 0; i < ARRAY_SIZE(vsi->segment_buf); i++) { + for (j = ARRAY_SIZE(vsi->segment_buf[i]) - 1; j >= 0; j--) { + val = ((seg_id_addr + i) << 2) + j; + writel(val, cm + VP8_HW_VLD_ADDR); + + val = readl(cm + VP8_HW_VLD_VALUE); + vsi->segment_buf[i][j] = val; + } + } +} + +/* reset HW and enable HW read/write data function */ +static void enable_hw_rw_function(struct vdec_vp8_inst *inst) +{ + u32 val = 0; + void __iomem *misc = inst->reg_base.misc; + void __iomem *ld = inst->reg_base.ld; + void __iomem *hwb = inst->reg_base.hwb; + void __iomem *hwd = inst->reg_base.hwd; + + mtk_vcodec_write_vdecsys(inst->ctx, VP8_RW_CKEN_SET, 0x1); + writel(0x101, ld + VP8_WO_VLD_SRST); + writel(0x101, hwb + VP8_WO_VLD_SRST); + + mtk_vcodec_write_vdecsys(inst->ctx, 0, 0x1); + val = readl(misc + VP8_RW_MISC_SRST); + writel((val & 0xFFFFFFFE), misc + VP8_RW_MISC_SRST); + + writel(0x1, misc + VP8_RW_MISC_SYS_SEL); + writel(0x17F, misc + VP8_RW_MISC_SPEC_CON); + writel(0x71201100, misc + VP8_RW_MISC_FUNC_CON); + writel(0x0, ld + VP8_WO_VLD_SRST); + writel(0x0, hwb + VP8_WO_VLD_SRST); + mtk_vcodec_write_vdecsys(inst->ctx, VP8_RW_DCM_CON, 0x1); + writel(0x1, misc + VP8_RW_MISC_DCM_CON); + writel(0x1, hwd + VP8_RW_VP8_CTRL); +} + +static void store_dec_table(struct vdec_vp8_inst *inst) +{ + int i, j; + u32 addr = 0, val = 0; + void __iomem *hwd = inst->reg_base.hwd; + u32 *p = &inst->vsi->dec_table[VP8_DEC_TABLE_OFFSET]; + + for (i = 0; i < VP8_DEC_TABLE_PROC_LOOP; i++) { + writel(addr, hwd + VP8_BSASET); + for (j = 0; j < VP8_DEC_TABLE_UNIT ; j++) { + val = *p++; + writel(val, hwd + VP8_BSDSET); + } + addr += VP8_DEC_TABLE_RW_UNIT; + } +} + +static void load_dec_table(struct vdec_vp8_inst *inst) +{ + int i; + u32 addr = 0; + u32 *p = &inst->vsi->dec_table[VP8_DEC_TABLE_OFFSET]; + void __iomem *hwd = inst->reg_base.hwd; + + for (i = 0; i < VP8_DEC_TABLE_PROC_LOOP; i++) { + writel(addr, hwd + VP8_BSASET); + /* read total 11 bytes */ + *p++ = readl(hwd + VP8_BSDSET); + *p++ = readl(hwd + VP8_BSDSET); + *p++ = readl(hwd + VP8_BSDSET) & 0xFFFFFF; + addr += VP8_DEC_TABLE_RW_UNIT; + } +} + +static void get_pic_info(struct vdec_vp8_inst *inst, struct vdec_pic_info *pic) +{ + *pic = inst->vsi->pic; + + mtk_vdec_debug(inst->ctx, "pic(%d, %d), buf(%d, %d)", + pic->pic_w, pic->pic_h, pic->buf_w, pic->buf_h); + mtk_vdec_debug(inst->ctx, "fb size: Y(%d), C(%d)", + pic->fb_sz[0], pic->fb_sz[1]); +} + +static void vp8_dec_finish(struct vdec_vp8_inst *inst) +{ + struct vdec_fb_node *node; + uint64_t prev_y_dma = inst->vsi->dec.prev_y_dma; + + mtk_vdec_debug(inst->ctx, "prev fb base dma=%llx", prev_y_dma); + + /* put last decode ok frame to fb_free_list */ + if (prev_y_dma != 0) { + list_for_each_entry(node, &inst->fb_use_list, list) { + struct vdec_fb *fb = (struct vdec_fb *)node->fb; + + if (prev_y_dma == (uint64_t)fb->base_y.dma_addr) { + list_move_tail(&node->list, + &inst->fb_free_list); + break; + } + } + } + + /* available_fb_node_list -> fb_use_list */ + node = list_first_entry(&inst->available_fb_node_list, + struct vdec_fb_node, list); + node->fb = inst->cur_fb; + list_move_tail(&node->list, &inst->fb_use_list); + + /* available_fb_node_list -> fb_disp_list */ + if (inst->vsi->dec.show_frame) { + node = list_first_entry(&inst->available_fb_node_list, + struct vdec_fb_node, list); + node->fb = inst->cur_fb; + list_move_tail(&node->list, &inst->fb_disp_list); + } +} + +static void move_fb_list_use_to_free(struct vdec_vp8_inst *inst) +{ + struct vdec_fb_node *node, *tmp; + + list_for_each_entry_safe(node, tmp, &inst->fb_use_list, list) + list_move_tail(&node->list, &inst->fb_free_list); +} + +static void init_list(struct vdec_vp8_inst *inst) +{ + int i; + + INIT_LIST_HEAD(&inst->available_fb_node_list); + INIT_LIST_HEAD(&inst->fb_use_list); + INIT_LIST_HEAD(&inst->fb_free_list); + INIT_LIST_HEAD(&inst->fb_disp_list); + + for (i = 0; i < ARRAY_SIZE(inst->dec_fb); i++) { + INIT_LIST_HEAD(&inst->dec_fb[i].list); + inst->dec_fb[i].fb = NULL; + list_add_tail(&inst->dec_fb[i].list, + &inst->available_fb_node_list); + } +} + +static void add_fb_to_free_list(struct vdec_vp8_inst *inst, void *fb) +{ + struct vdec_fb_node *node; + + if (fb) { + node = list_first_entry(&inst->available_fb_node_list, + struct vdec_fb_node, list); + node->fb = fb; + list_move_tail(&node->list, &inst->fb_free_list); + } +} + +static int alloc_working_buf(struct vdec_vp8_inst *inst) +{ + int err; + struct mtk_vcodec_mem *mem = &inst->working_buf; + + mem->size = VP8_WORKING_BUF_SZ; + err = mtk_vcodec_mem_alloc(inst->ctx, mem); + if (err) { + mtk_vdec_err(inst->ctx, "Cannot allocate working buffer"); + return err; + } + + inst->vsi->dec.working_buf_dma = (uint64_t)mem->dma_addr; + return 0; +} + +static void free_working_buf(struct vdec_vp8_inst *inst) +{ + struct mtk_vcodec_mem *mem = &inst->working_buf; + + if (mem->va) + mtk_vcodec_mem_free(inst->ctx, mem); + + inst->vsi->dec.working_buf_dma = 0; +} + +static int vdec_vp8_init(struct mtk_vcodec_dec_ctx *ctx) +{ + struct vdec_vp8_inst *inst; + int err; + + inst = kzalloc(sizeof(*inst), GFP_KERNEL); + if (!inst) + return -ENOMEM; + + inst->ctx = ctx; + + inst->vpu.id = IPI_VDEC_VP8; + inst->vpu.ctx = ctx; + + err = vpu_dec_init(&inst->vpu); + if (err) { + mtk_vdec_err(ctx, "vdec_vp8 init err=%d", err); + goto error_free_inst; + } + + inst->vsi = (struct vdec_vp8_vsi *)inst->vpu.vsi; + init_list(inst); + err = alloc_working_buf(inst); + if (err) + goto error_deinit; + + get_hw_reg_base(inst); + mtk_vdec_debug(ctx, "VP8 Instance >> %p", inst); + + ctx->drv_handle = inst; + return 0; + +error_deinit: + vpu_dec_deinit(&inst->vpu); +error_free_inst: + kfree(inst); + return err; +} + +static int vdec_vp8_decode(void *h_vdec, struct mtk_vcodec_mem *bs, + struct vdec_fb *fb, bool *res_chg) +{ + struct vdec_vp8_inst *inst = (struct vdec_vp8_inst *)h_vdec; + struct vdec_vp8_dec_info *dec = &inst->vsi->dec; + struct vdec_vpu_inst *vpu = &inst->vpu; + unsigned char *bs_va; + unsigned int data; + int err = 0; + uint64_t y_fb_dma; + uint64_t c_fb_dma; + + /* bs NULL means flush decoder */ + if (bs == NULL) { + move_fb_list_use_to_free(inst); + return vpu_dec_reset(vpu); + } + + y_fb_dma = fb ? (u64)fb->base_y.dma_addr : 0; + c_fb_dma = fb ? (u64)fb->base_c.dma_addr : 0; + + mtk_vdec_debug(inst->ctx, "+ [%d] FB y_dma=%llx c_dma=%llx fb=%p", + inst->frm_cnt, y_fb_dma, c_fb_dma, fb); + + inst->cur_fb = fb; + dec->bs_dma = (unsigned long)bs->dma_addr; + dec->bs_sz = bs->size; + dec->cur_y_fb_dma = y_fb_dma; + dec->cur_c_fb_dma = c_fb_dma; + + mtk_vdec_debug(inst->ctx, "\n + FRAME[%d] +\n", inst->frm_cnt); + + write_hw_segmentation_data(inst); + enable_hw_rw_function(inst); + store_dec_table(inst); + + bs_va = (unsigned char *)bs->va; + + /* retrieve width/hight and scale info from header */ + data = (*(bs_va + 9) << 24) | (*(bs_va + 8) << 16) | + (*(bs_va + 7) << 8) | *(bs_va + 6); + err = vpu_dec_start(vpu, &data, 1); + if (err) { + add_fb_to_free_list(inst, fb); + if (dec->wait_key_frame) { + mtk_vdec_debug(inst->ctx, "wait key frame !"); + return 0; + } + + goto error; + } + + if (dec->resolution_changed) { + mtk_vdec_debug(inst->ctx, "- resolution_changed -"); + *res_chg = true; + add_fb_to_free_list(inst, fb); + return 0; + } + + /* wait decoder done interrupt */ + mtk_vcodec_wait_for_done_ctx(inst->ctx, MTK_INST_IRQ_RECEIVED, + WAIT_INTR_TIMEOUT_MS, 0); + + if (inst->vsi->load_data) + load_dec_table(inst); + + vp8_dec_finish(inst); + read_hw_segmentation_data(inst); + + err = vpu_dec_end(vpu); + if (err) + goto error; + + mtk_vdec_debug(inst->ctx, "\n - FRAME[%d] - show=%d\n", inst->frm_cnt, dec->show_frame); + inst->frm_cnt++; + *res_chg = false; + return 0; + +error: + mtk_vdec_err(inst->ctx, "\n - FRAME[%d] - err=%d\n", inst->frm_cnt, err); + return err; +} + +static void get_disp_fb(struct vdec_vp8_inst *inst, struct vdec_fb **out_fb) +{ + struct vdec_fb_node *node; + struct vdec_fb *fb; + + node = list_first_entry_or_null(&inst->fb_disp_list, + struct vdec_fb_node, list); + if (node) { + list_move_tail(&node->list, &inst->available_fb_node_list); + fb = (struct vdec_fb *)node->fb; + fb->status |= FB_ST_DISPLAY; + mtk_vdec_debug(inst->ctx, "[FB] get disp fb %p st=%d", node->fb, fb->status); + } else { + fb = NULL; + mtk_vdec_debug(inst->ctx, "[FB] there is no disp fb"); + } + + *out_fb = fb; +} + +static void get_free_fb(struct vdec_vp8_inst *inst, struct vdec_fb **out_fb) +{ + struct vdec_fb_node *node; + struct vdec_fb *fb; + + node = list_first_entry_or_null(&inst->fb_free_list, + struct vdec_fb_node, list); + if (node) { + list_move_tail(&node->list, &inst->available_fb_node_list); + fb = (struct vdec_fb *)node->fb; + fb->status |= FB_ST_FREE; + mtk_vdec_debug(inst->ctx, "[FB] get free fb %p st=%d", node->fb, fb->status); + } else { + fb = NULL; + mtk_vdec_debug(inst->ctx, "[FB] there is no free fb"); + } + + *out_fb = fb; +} + +static void get_crop_info(struct vdec_vp8_inst *inst, struct v4l2_rect *cr) +{ + cr->left = 0; + cr->top = 0; + cr->width = inst->vsi->pic.pic_w; + cr->height = inst->vsi->pic.pic_h; + mtk_vdec_debug(inst->ctx, "get crop info l=%d, t=%d, w=%d, h=%d", + cr->left, cr->top, cr->width, cr->height); +} + +static int vdec_vp8_get_param(void *h_vdec, enum vdec_get_param_type type, + void *out) +{ + struct vdec_vp8_inst *inst = (struct vdec_vp8_inst *)h_vdec; + + switch (type) { + case GET_PARAM_DISP_FRAME_BUFFER: + get_disp_fb(inst, out); + break; + + case GET_PARAM_FREE_FRAME_BUFFER: + get_free_fb(inst, out); + break; + + case GET_PARAM_PIC_INFO: + get_pic_info(inst, out); + break; + + case GET_PARAM_CROP_INFO: + get_crop_info(inst, out); + break; + + case GET_PARAM_DPB_SIZE: + *((unsigned int *)out) = VP8_DPB_SIZE; + break; + + default: + mtk_vdec_err(inst->ctx, "invalid get parameter type=%d", type); + return -EINVAL; + } + + return 0; +} + +static void vdec_vp8_deinit(void *h_vdec) +{ + struct vdec_vp8_inst *inst = (struct vdec_vp8_inst *)h_vdec; + + vpu_dec_deinit(&inst->vpu); + free_working_buf(inst); + kfree(inst); +} + +const struct vdec_common_if vdec_vp8_if = { + .init = vdec_vp8_init, + .decode = vdec_vp8_decode, + .get_param = vdec_vp8_get_param, + .deinit = vdec_vp8_deinit, +}; diff --git a/drivers/media/platform/mediatek/vcodec/decoder/vdec/vdec_vp8_req_if.c b/drivers/media/platform/mediatek/vcodec/decoder/vdec/vdec_vp8_req_if.c new file mode 100644 index 0000000000..f64b21c071 --- /dev/null +++ b/drivers/media/platform/mediatek/vcodec/decoder/vdec/vdec_vp8_req_if.c @@ -0,0 +1,433 @@ +// SPDX-License-Identifier: GPL-2.0 +/* + * Copyright (c) 2021 MediaTek Inc. + * Author: Yunfei Dong <yunfei.dong@mediatek.com> + */ + +#include <linux/slab.h> +#include <media/v4l2-mem2mem.h> +#include <media/videobuf2-dma-contig.h> +#include <uapi/linux/v4l2-controls.h> + +#include "../mtk_vcodec_dec.h" +#include "../../common/mtk_vcodec_intr.h" +#include "../vdec_drv_base.h" +#include "../vdec_drv_if.h" +#include "../vdec_vpu_if.h" + +/* Decoding picture buffer size (3 reference frames plus current frame) */ +#define VP8_DPB_SIZE 4 + +/* HW working buffer size (bytes) */ +#define VP8_SEG_ID_SZ SZ_256K +#define VP8_PP_WRAPY_SZ SZ_64K +#define VP8_PP_WRAPC_SZ SZ_64K +#define VP8_VLD_PRED_SZ SZ_64K + +/** + * struct vdec_vp8_slice_info - decode misc information + * + * @vld_wrapper_dma: vld wrapper dma address + * @seg_id_buf_dma: seg id dma address + * @wrap_y_dma: wrap y dma address + * @wrap_c_dma: wrap y dma address + * @cur_y_fb_dma: current plane Y frame buffer dma address + * @cur_c_fb_dma: current plane C frame buffer dma address + * @bs_dma: bitstream dma address + * @bs_sz: bitstream size + * @resolution_changed:resolution change flag 1 - changed, 0 - not change + * @frame_header_type: current frame header type + * @wait_key_frame: wait key frame coming + * @crc: used to check whether hardware's status is right + * @reserved: reserved, currently unused + */ +struct vdec_vp8_slice_info { + u64 vld_wrapper_dma; + u64 seg_id_buf_dma; + u64 wrap_y_dma; + u64 wrap_c_dma; + u64 cur_y_fb_dma; + u64 cur_c_fb_dma; + u64 bs_dma; + u32 bs_sz; + u32 resolution_changed; + u32 frame_header_type; + u32 crc[8]; + u32 reserved; +}; + +/** + * struct vdec_vp8_slice_dpb_info - vp8 reference information + * + * @y_dma_addr: Y bitstream physical address + * @c_dma_addr: CbCr bitstream physical address + * @reference_flag: reference picture flag + * @reserved: 64bit align + */ +struct vdec_vp8_slice_dpb_info { + dma_addr_t y_dma_addr; + dma_addr_t c_dma_addr; + int reference_flag; + int reserved; +}; + +/** + * struct vdec_vp8_slice_vsi - VPU shared information + * + * @dec: decoding information + * @pic: picture information + * @vp8_dpb_info: reference buffer information + */ +struct vdec_vp8_slice_vsi { + struct vdec_vp8_slice_info dec; + struct vdec_pic_info pic; + struct vdec_vp8_slice_dpb_info vp8_dpb_info[3]; +}; + +/** + * struct vdec_vp8_slice_inst - VP8 decoder instance + * + * @seg_id_buf: seg buffer + * @wrap_y_buf: wrapper y buffer + * @wrap_c_buf: wrapper c buffer + * @vld_wrapper_buf: vld wrapper buffer + * @ctx: V4L2 context + * @vpu: VPU instance for decoder + * @vsi: VPU share information + */ +struct vdec_vp8_slice_inst { + struct mtk_vcodec_mem seg_id_buf; + struct mtk_vcodec_mem wrap_y_buf; + struct mtk_vcodec_mem wrap_c_buf; + struct mtk_vcodec_mem vld_wrapper_buf; + struct mtk_vcodec_dec_ctx *ctx; + struct vdec_vpu_inst vpu; + struct vdec_vp8_slice_vsi *vsi; +}; + +static void *vdec_vp8_slice_get_ctrl_ptr(struct mtk_vcodec_dec_ctx *ctx, int id) +{ + struct v4l2_ctrl *ctrl = v4l2_ctrl_find(&ctx->ctrl_hdl, id); + + if (!ctrl) + return ERR_PTR(-EINVAL); + + return ctrl->p_cur.p; +} + +static void vdec_vp8_slice_get_pic_info(struct vdec_vp8_slice_inst *inst) +{ + struct mtk_vcodec_dec_ctx *ctx = inst->ctx; + unsigned int data[3]; + + data[0] = ctx->picinfo.pic_w; + data[1] = ctx->picinfo.pic_h; + data[2] = ctx->capture_fourcc; + vpu_dec_get_param(&inst->vpu, data, 3, GET_PARAM_PIC_INFO); + + ctx->picinfo.buf_w = ALIGN(ctx->picinfo.pic_w, 64); + ctx->picinfo.buf_h = ALIGN(ctx->picinfo.pic_h, 64); + ctx->picinfo.fb_sz[0] = inst->vpu.fb_sz[0]; + ctx->picinfo.fb_sz[1] = inst->vpu.fb_sz[1]; + + inst->vsi->pic.pic_w = ctx->picinfo.pic_w; + inst->vsi->pic.pic_h = ctx->picinfo.pic_h; + inst->vsi->pic.buf_w = ctx->picinfo.buf_w; + inst->vsi->pic.buf_h = ctx->picinfo.buf_h; + inst->vsi->pic.fb_sz[0] = ctx->picinfo.fb_sz[0]; + inst->vsi->pic.fb_sz[1] = ctx->picinfo.fb_sz[1]; + mtk_vdec_debug(inst->ctx, "pic(%d, %d), buf(%d, %d)", + ctx->picinfo.pic_w, ctx->picinfo.pic_h, + ctx->picinfo.buf_w, ctx->picinfo.buf_h); + mtk_vdec_debug(inst->ctx, "fb size: Y(%d), C(%d)", + ctx->picinfo.fb_sz[0], ctx->picinfo.fb_sz[1]); +} + +static int vdec_vp8_slice_alloc_working_buf(struct vdec_vp8_slice_inst *inst) +{ + int err; + struct mtk_vcodec_mem *mem; + + mem = &inst->seg_id_buf; + mem->size = VP8_SEG_ID_SZ; + err = mtk_vcodec_mem_alloc(inst->ctx, mem); + if (err) { + mtk_vdec_err(inst->ctx, "Cannot allocate working buffer"); + return err; + } + inst->vsi->dec.seg_id_buf_dma = (u64)mem->dma_addr; + + mem = &inst->wrap_y_buf; + mem->size = VP8_PP_WRAPY_SZ; + err = mtk_vcodec_mem_alloc(inst->ctx, mem); + if (err) { + mtk_vdec_err(inst->ctx, "cannot allocate WRAP Y buffer"); + return err; + } + inst->vsi->dec.wrap_y_dma = (u64)mem->dma_addr; + + mem = &inst->wrap_c_buf; + mem->size = VP8_PP_WRAPC_SZ; + err = mtk_vcodec_mem_alloc(inst->ctx, mem); + if (err) { + mtk_vdec_err(inst->ctx, "cannot allocate WRAP C buffer"); + return err; + } + inst->vsi->dec.wrap_c_dma = (u64)mem->dma_addr; + + mem = &inst->vld_wrapper_buf; + mem->size = VP8_VLD_PRED_SZ; + err = mtk_vcodec_mem_alloc(inst->ctx, mem); + if (err) { + mtk_vdec_err(inst->ctx, "cannot allocate vld wrapper buffer"); + return err; + } + inst->vsi->dec.vld_wrapper_dma = (u64)mem->dma_addr; + + return 0; +} + +static void vdec_vp8_slice_free_working_buf(struct vdec_vp8_slice_inst *inst) +{ + struct mtk_vcodec_mem *mem; + + mem = &inst->seg_id_buf; + if (mem->va) + mtk_vcodec_mem_free(inst->ctx, mem); + inst->vsi->dec.seg_id_buf_dma = 0; + + mem = &inst->wrap_y_buf; + if (mem->va) + mtk_vcodec_mem_free(inst->ctx, mem); + inst->vsi->dec.wrap_y_dma = 0; + + mem = &inst->wrap_c_buf; + if (mem->va) + mtk_vcodec_mem_free(inst->ctx, mem); + inst->vsi->dec.wrap_c_dma = 0; + + mem = &inst->vld_wrapper_buf; + if (mem->va) + mtk_vcodec_mem_free(inst->ctx, mem); + inst->vsi->dec.vld_wrapper_dma = 0; +} + +static u64 vdec_vp8_slice_get_ref_by_ts(const struct v4l2_ctrl_vp8_frame *frame_header, + int index) +{ + switch (index) { + case 0: + return frame_header->last_frame_ts; + case 1: + return frame_header->golden_frame_ts; + case 2: + return frame_header->alt_frame_ts; + default: + break; + } + + return -1; +} + +static int vdec_vp8_slice_get_decode_parameters(struct vdec_vp8_slice_inst *inst) +{ + const struct v4l2_ctrl_vp8_frame *frame_header; + struct mtk_vcodec_dec_ctx *ctx = inst->ctx; + struct vb2_queue *vq; + struct vb2_buffer *vb; + u64 referenct_ts; + int index; + + frame_header = vdec_vp8_slice_get_ctrl_ptr(inst->ctx, V4L2_CID_STATELESS_VP8_FRAME); + if (IS_ERR(frame_header)) + return PTR_ERR(frame_header); + + vq = v4l2_m2m_get_vq(ctx->m2m_ctx, V4L2_BUF_TYPE_VIDEO_CAPTURE_MPLANE); + for (index = 0; index < 3; index++) { + referenct_ts = vdec_vp8_slice_get_ref_by_ts(frame_header, index); + vb = vb2_find_buffer(vq, referenct_ts); + if (!vb) { + if (!V4L2_VP8_FRAME_IS_KEY_FRAME(frame_header)) + mtk_vdec_err(inst->ctx, "reference invalid: index(%d) ts(%lld)", + index, referenct_ts); + inst->vsi->vp8_dpb_info[index].reference_flag = 0; + continue; + } + inst->vsi->vp8_dpb_info[index].reference_flag = 1; + + inst->vsi->vp8_dpb_info[index].y_dma_addr = + vb2_dma_contig_plane_dma_addr(vb, 0); + if (ctx->q_data[MTK_Q_DATA_DST].fmt->num_planes == 2) + inst->vsi->vp8_dpb_info[index].c_dma_addr = + vb2_dma_contig_plane_dma_addr(vb, 1); + else + inst->vsi->vp8_dpb_info[index].c_dma_addr = + inst->vsi->vp8_dpb_info[index].y_dma_addr + + ctx->picinfo.fb_sz[0]; + } + + inst->vsi->dec.frame_header_type = frame_header->flags >> 1; + + return 0; +} + +static int vdec_vp8_slice_init(struct mtk_vcodec_dec_ctx *ctx) +{ + struct vdec_vp8_slice_inst *inst; + int err; + + inst = kzalloc(sizeof(*inst), GFP_KERNEL); + if (!inst) + return -ENOMEM; + + inst->ctx = ctx; + + inst->vpu.id = SCP_IPI_VDEC_LAT; + inst->vpu.core_id = SCP_IPI_VDEC_CORE; + inst->vpu.ctx = ctx; + inst->vpu.codec_type = ctx->current_codec; + inst->vpu.capture_type = ctx->capture_fourcc; + + err = vpu_dec_init(&inst->vpu); + if (err) { + mtk_vdec_err(ctx, "vdec_vp8 init err=%d", err); + goto error_free_inst; + } + + inst->vsi = inst->vpu.vsi; + err = vdec_vp8_slice_alloc_working_buf(inst); + if (err) + goto error_deinit; + + mtk_vdec_debug(ctx, "vp8 struct size = %d vsi: %d\n", + (int)sizeof(struct v4l2_ctrl_vp8_frame), + (int)sizeof(struct vdec_vp8_slice_vsi)); + mtk_vdec_debug(ctx, "vp8:%p, codec_type = 0x%x vsi: 0x%p", + inst, inst->vpu.codec_type, inst->vpu.vsi); + + ctx->drv_handle = inst; + return 0; + +error_deinit: + vpu_dec_deinit(&inst->vpu); +error_free_inst: + kfree(inst); + return err; +} + +static int vdec_vp8_slice_decode(void *h_vdec, struct mtk_vcodec_mem *bs, + struct vdec_fb *fb, bool *res_chg) +{ + struct vdec_vp8_slice_inst *inst = h_vdec; + struct vdec_vpu_inst *vpu = &inst->vpu; + struct mtk_video_dec_buf *src_buf_info, *dst_buf_info; + unsigned int data; + u64 y_fb_dma, c_fb_dma; + int err, timeout; + + /* Resolution changes are never initiated by us */ + *res_chg = false; + + /* bs NULL means flush decoder */ + if (!bs) + return vpu_dec_reset(vpu); + + src_buf_info = container_of(bs, struct mtk_video_dec_buf, bs_buffer); + + fb = inst->ctx->dev->vdec_pdata->get_cap_buffer(inst->ctx); + dst_buf_info = container_of(fb, struct mtk_video_dec_buf, frame_buffer); + + y_fb_dma = fb ? (u64)fb->base_y.dma_addr : 0; + if (inst->ctx->q_data[MTK_Q_DATA_DST].fmt->num_planes == 1) + c_fb_dma = y_fb_dma + + inst->ctx->picinfo.buf_w * inst->ctx->picinfo.buf_h; + else + c_fb_dma = fb ? (u64)fb->base_c.dma_addr : 0; + + inst->vsi->dec.bs_dma = (u64)bs->dma_addr; + inst->vsi->dec.bs_sz = bs->size; + inst->vsi->dec.cur_y_fb_dma = y_fb_dma; + inst->vsi->dec.cur_c_fb_dma = c_fb_dma; + + mtk_vdec_debug(inst->ctx, "frame[%d] bs(%zu 0x%llx) y/c(0x%llx 0x%llx)", + inst->ctx->decoded_frame_cnt, + bs->size, (u64)bs->dma_addr, + y_fb_dma, c_fb_dma); + + v4l2_m2m_buf_copy_metadata(&src_buf_info->m2m_buf.vb, + &dst_buf_info->m2m_buf.vb, true); + + err = vdec_vp8_slice_get_decode_parameters(inst); + if (err) + goto error; + + err = vpu_dec_start(vpu, &data, 1); + if (err) { + mtk_vdec_debug(inst->ctx, "vp8 dec start err!"); + goto error; + } + + if (inst->vsi->dec.resolution_changed) { + mtk_vdec_debug(inst->ctx, "- resolution_changed -"); + *res_chg = true; + return 0; + } + + /* wait decode done interrupt */ + timeout = mtk_vcodec_wait_for_done_ctx(inst->ctx, MTK_INST_IRQ_RECEIVED, + 50, MTK_VDEC_CORE); + + err = vpu_dec_end(vpu); + if (err || timeout) + mtk_vdec_debug(inst->ctx, "vp8 dec error timeout:%d err: %d pic_%d", + timeout, err, inst->ctx->decoded_frame_cnt); + + mtk_vdec_debug(inst->ctx, "pic[%d] crc: 0x%x 0x%x 0x%x 0x%x 0x%x 0x%x 0x%x 0x%x", + inst->ctx->decoded_frame_cnt, + inst->vsi->dec.crc[0], inst->vsi->dec.crc[1], + inst->vsi->dec.crc[2], inst->vsi->dec.crc[3], + inst->vsi->dec.crc[4], inst->vsi->dec.crc[5], + inst->vsi->dec.crc[6], inst->vsi->dec.crc[7]); + + inst->ctx->decoded_frame_cnt++; +error: + return err; +} + +static int vdec_vp8_slice_get_param(void *h_vdec, enum vdec_get_param_type type, void *out) +{ + struct vdec_vp8_slice_inst *inst = h_vdec; + + switch (type) { + case GET_PARAM_PIC_INFO: + vdec_vp8_slice_get_pic_info(inst); + break; + case GET_PARAM_CROP_INFO: + mtk_vdec_debug(inst->ctx, "No need to get vp8 crop information."); + break; + case GET_PARAM_DPB_SIZE: + *((unsigned int *)out) = VP8_DPB_SIZE; + break; + default: + mtk_vdec_err(inst->ctx, "invalid get parameter type=%d", type); + return -EINVAL; + } + + return 0; +} + +static void vdec_vp8_slice_deinit(void *h_vdec) +{ + struct vdec_vp8_slice_inst *inst = h_vdec; + + vpu_dec_deinit(&inst->vpu); + vdec_vp8_slice_free_working_buf(inst); + kfree(inst); +} + +const struct vdec_common_if vdec_vp8_slice_if = { + .init = vdec_vp8_slice_init, + .decode = vdec_vp8_slice_decode, + .get_param = vdec_vp8_slice_get_param, + .deinit = vdec_vp8_slice_deinit, +}; diff --git a/drivers/media/platform/mediatek/vcodec/decoder/vdec/vdec_vp9_if.c b/drivers/media/platform/mediatek/vcodec/decoder/vdec/vdec_vp9_if.c new file mode 100644 index 0000000000..55355fa700 --- /dev/null +++ b/drivers/media/platform/mediatek/vcodec/decoder/vdec/vdec_vp9_if.c @@ -0,0 +1,1017 @@ +// SPDX-License-Identifier: GPL-2.0 +/* + * Copyright (c) 2016 MediaTek Inc. + * Author: Daniel Hsiao <daniel.hsiao@mediatek.com> + * Kai-Sean Yang <kai-sean.yang@mediatek.com> + * Tiffany Lin <tiffany.lin@mediatek.com> + */ + +#include <linux/fs.h> +#include <linux/slab.h> +#include <linux/syscalls.h> +#include <linux/delay.h> +#include <linux/time.h> + +#include "../../common/mtk_vcodec_intr.h" +#include "../vdec_drv_base.h" +#include "../vdec_vpu_if.h" + +#define VP9_SUPER_FRAME_BS_SZ 64 +#define MAX_VP9_DPB_SIZE 9 + +#define REFS_PER_FRAME 3 +#define MAX_NUM_REF_FRAMES 8 +#define VP9_MAX_FRM_BUF_NUM 9 +#define VP9_MAX_FRM_BUF_NODE_NUM (VP9_MAX_FRM_BUF_NUM * 2) +#define VP9_SEG_ID_SZ 0x12000 + +/** + * struct vp9_dram_buf - contains buffer info for vpu + * @va : cpu address + * @pa : iova address + * @sz : buffer size + * @padding : for 64 bytes alignment + */ +struct vp9_dram_buf { + unsigned long va; + unsigned long pa; + unsigned int sz; + unsigned int padding; +}; + +/** + * struct vp9_fb_info - contains frame buffer info + * @fb : frmae buffer + * @reserved : reserved field used by vpu + */ +struct vp9_fb_info { + struct vdec_fb *fb; + unsigned int reserved[32]; +}; + +/** + * struct vp9_ref_cnt_buf - contains reference buffer information + * @buf : referenced frame buffer + * @ref_cnt : referenced frame buffer's reference count. + * When reference count=0, remove it from reference list + */ +struct vp9_ref_cnt_buf { + struct vp9_fb_info buf; + unsigned int ref_cnt; +}; + +/** + * struct vp9_ref_buf - contains current frame's reference buffer information + * @buf : reference buffer + * @idx : reference buffer index to frm_bufs + * @reserved : reserved field used by vpu + */ +struct vp9_ref_buf { + struct vp9_fb_info *buf; + unsigned int idx; + unsigned int reserved[6]; +}; + +/** + * struct vp9_sf_ref_fb - contains frame buffer info + * @fb : super frame reference frame buffer + * @used : this reference frame info entry is used + * @padding : for 64 bytes size align + */ +struct vp9_sf_ref_fb { + struct vdec_fb fb; + int used; + int padding; +}; + +/* + * struct vdec_vp9_vsi - shared buffer between host and VPU firmware + * AP-W/R : AP is writer/reader on this item + * VPU-W/R: VPU is write/reader on this item + * @sf_bs_buf : super frame backup buffer (AP-W, VPU-R) + * @sf_ref_fb : record supoer frame reference buffer information + * (AP-R/W, VPU-R/W) + * @sf_next_ref_fb_idx : next available super frame (AP-W, VPU-R) + * @sf_frm_cnt : super frame count, filled by vpu (AP-R, VPU-W) + * @sf_frm_offset : super frame offset, filled by vpu (AP-R, VPU-W) + * @sf_frm_sz : super frame size, filled by vpu (AP-R, VPU-W) + * @sf_frm_idx : current super frame (AP-R, VPU-W) + * @sf_init : inform super frame info already parsed by vpu (AP-R, VPU-W) + * @fb : capture buffer (AP-W, VPU-R) + * @bs : bs buffer (AP-W, VPU-R) + * @cur_fb : current show capture buffer (AP-R/W, VPU-R/W) + * @pic_w : picture width (AP-R, VPU-W) + * @pic_h : picture height (AP-R, VPU-W) + * @buf_w : codec width (AP-R, VPU-W) + * @buf_h : coded height (AP-R, VPU-W) + * @buf_sz_y_bs : ufo compressed y plane size (AP-R, VPU-W) + * @buf_sz_c_bs : ufo compressed cbcr plane size (AP-R, VPU-W) + * @buf_len_sz_y : size used to store y plane ufo info (AP-R, VPU-W) + * @buf_len_sz_c : size used to store cbcr plane ufo info (AP-R, VPU-W) + + * @profile : profile sparsed from vpu (AP-R, VPU-W) + * @show_frame : [BIT(0)] display this frame or not (AP-R, VPU-W) + * [BIT(1)] reset segment data or not (AP-R, VPU-W) + * [BIT(2)] trig decoder hardware or not (AP-R, VPU-W) + * [BIT(3)] ask VPU to set bits(0~4) accordingly (AP-W, VPU-R) + * [BIT(4)] do not reset segment data before every frame (AP-R, VPU-W) + * @show_existing_frame : inform this frame is show existing frame + * (AP-R, VPU-W) + * @frm_to_show_idx : index to show frame (AP-R, VPU-W) + + * @refresh_frm_flags : indicate when frame need to refine reference count + * (AP-R, VPU-W) + * @resolution_changed : resolution change in this frame (AP-R, VPU-W) + + * @frm_bufs : maintain reference buffer info (AP-R/W, VPU-R/W) + * @ref_frm_map : maintain reference buffer map info (AP-R/W, VPU-R/W) + * @new_fb_idx : index to frm_bufs array (AP-R, VPU-W) + * @frm_num : decoded frame number, include sub-frame count (AP-R, VPU-W) + * @mv_buf : motion vector working buffer (AP-W, VPU-R) + * @frm_refs : maintain three reference buffer info (AP-R/W, VPU-R/W) + * @seg_id_buf : segmentation map working buffer (AP-W, VPU-R) + */ +struct vdec_vp9_vsi { + unsigned char sf_bs_buf[VP9_SUPER_FRAME_BS_SZ]; + struct vp9_sf_ref_fb sf_ref_fb[VP9_MAX_FRM_BUF_NUM-1]; + int sf_next_ref_fb_idx; + unsigned int sf_frm_cnt; + unsigned int sf_frm_offset[VP9_MAX_FRM_BUF_NUM-1]; + unsigned int sf_frm_sz[VP9_MAX_FRM_BUF_NUM-1]; + unsigned int sf_frm_idx; + unsigned int sf_init; + struct vdec_fb fb; + struct mtk_vcodec_mem bs; + struct vdec_fb cur_fb; + unsigned int pic_w; + unsigned int pic_h; + unsigned int buf_w; + unsigned int buf_h; + unsigned int buf_sz_y_bs; + unsigned int buf_sz_c_bs; + unsigned int buf_len_sz_y; + unsigned int buf_len_sz_c; + unsigned int profile; + unsigned int show_frame; + unsigned int show_existing_frame; + unsigned int frm_to_show_idx; + unsigned int refresh_frm_flags; + unsigned int resolution_changed; + + struct vp9_ref_cnt_buf frm_bufs[VP9_MAX_FRM_BUF_NUM]; + int ref_frm_map[MAX_NUM_REF_FRAMES]; + unsigned int new_fb_idx; + unsigned int frm_num; + struct vp9_dram_buf mv_buf; + + struct vp9_ref_buf frm_refs[REFS_PER_FRAME]; + struct vp9_dram_buf seg_id_buf; + +}; + +/* + * struct vdec_vp9_inst - vp9 decode instance + * @mv_buf : working buffer for mv + * @seg_id_buf : working buffer for segmentation map + * @dec_fb : vdec_fb node to link fb to different fb_xxx_list + * @available_fb_node_list : current available vdec_fb node + * @fb_use_list : current used or referenced vdec_fb + * @fb_free_list : current available to free vdec_fb + * @fb_disp_list : current available to display vdec_fb + * @cur_fb : current frame buffer + * @ctx : current decode context + * @vpu : vpu instance information + * @vsi : shared buffer between host and VPU firmware + * @total_frm_cnt : total frame count, it do not include sub-frames in super + * frame + * @mem : instance memory information + */ +struct vdec_vp9_inst { + struct mtk_vcodec_mem mv_buf; + struct mtk_vcodec_mem seg_id_buf; + + struct vdec_fb_node dec_fb[VP9_MAX_FRM_BUF_NODE_NUM]; + struct list_head available_fb_node_list; + struct list_head fb_use_list; + struct list_head fb_free_list; + struct list_head fb_disp_list; + struct vdec_fb *cur_fb; + struct mtk_vcodec_dec_ctx *ctx; + struct vdec_vpu_inst vpu; + struct vdec_vp9_vsi *vsi; + unsigned int total_frm_cnt; + struct mtk_vcodec_mem mem; +}; + +static bool vp9_is_sf_ref_fb(struct vdec_vp9_inst *inst, struct vdec_fb *fb) +{ + int i; + struct vdec_vp9_vsi *vsi = inst->vsi; + + for (i = 0; i < ARRAY_SIZE(vsi->sf_ref_fb); i++) { + if (fb == &vsi->sf_ref_fb[i].fb) + return true; + } + return false; +} + +static struct vdec_fb *vp9_rm_from_fb_use_list(struct vdec_vp9_inst + *inst, void *addr) +{ + struct vdec_fb *fb = NULL; + struct vdec_fb_node *node; + + list_for_each_entry(node, &inst->fb_use_list, list) { + fb = (struct vdec_fb *)node->fb; + if (fb->base_y.va == addr) { + list_move_tail(&node->list, + &inst->available_fb_node_list); + return fb; + } + } + + return NULL; +} + +static void vp9_add_to_fb_free_list(struct vdec_vp9_inst *inst, + struct vdec_fb *fb) +{ + struct vdec_fb_node *node; + + if (fb) { + node = list_first_entry_or_null(&inst->available_fb_node_list, + struct vdec_fb_node, list); + + if (node) { + node->fb = fb; + list_move_tail(&node->list, &inst->fb_free_list); + } + } else { + mtk_vdec_debug(inst->ctx, "No free fb node"); + } +} + +static void vp9_free_sf_ref_fb(struct vdec_fb *fb) +{ + struct vp9_sf_ref_fb *sf_ref_fb = + container_of(fb, struct vp9_sf_ref_fb, fb); + + sf_ref_fb->used = 0; +} + +static void vp9_ref_cnt_fb(struct vdec_vp9_inst *inst, int *idx, + int new_idx) +{ + struct vdec_vp9_vsi *vsi = inst->vsi; + int ref_idx = *idx; + + if (ref_idx >= 0 && vsi->frm_bufs[ref_idx].ref_cnt > 0) { + vsi->frm_bufs[ref_idx].ref_cnt--; + + if (vsi->frm_bufs[ref_idx].ref_cnt == 0) { + if (!vp9_is_sf_ref_fb(inst, + vsi->frm_bufs[ref_idx].buf.fb)) { + struct vdec_fb *fb; + + fb = vp9_rm_from_fb_use_list(inst, + vsi->frm_bufs[ref_idx].buf.fb->base_y.va); + vp9_add_to_fb_free_list(inst, fb); + } else + vp9_free_sf_ref_fb( + vsi->frm_bufs[ref_idx].buf.fb); + } + } + + *idx = new_idx; + vsi->frm_bufs[new_idx].ref_cnt++; +} + +static void vp9_free_all_sf_ref_fb(struct vdec_vp9_inst *inst) +{ + int i; + struct vdec_vp9_vsi *vsi = inst->vsi; + + for (i = 0; i < ARRAY_SIZE(vsi->sf_ref_fb); i++) { + if (vsi->sf_ref_fb[i].fb.base_y.va) { + mtk_vcodec_mem_free(inst->ctx, + &vsi->sf_ref_fb[i].fb.base_y); + mtk_vcodec_mem_free(inst->ctx, + &vsi->sf_ref_fb[i].fb.base_c); + vsi->sf_ref_fb[i].used = 0; + } + } +} + +/* For each sub-frame except the last one, the driver will dynamically + * allocate reference buffer by calling vp9_get_sf_ref_fb() + * The last sub-frame will use the original fb provided by the + * vp9_dec_decode() interface + */ +static int vp9_get_sf_ref_fb(struct vdec_vp9_inst *inst) +{ + int idx; + struct mtk_vcodec_mem *mem_basy_y; + struct mtk_vcodec_mem *mem_basy_c; + struct vdec_vp9_vsi *vsi = inst->vsi; + + for (idx = 0; + idx < ARRAY_SIZE(vsi->sf_ref_fb); + idx++) { + if (vsi->sf_ref_fb[idx].fb.base_y.va && + vsi->sf_ref_fb[idx].used == 0) { + return idx; + } + } + + for (idx = 0; + idx < ARRAY_SIZE(vsi->sf_ref_fb); + idx++) { + if (vsi->sf_ref_fb[idx].fb.base_y.va == NULL) + break; + } + + if (idx == ARRAY_SIZE(vsi->sf_ref_fb)) { + mtk_vdec_err(inst->ctx, "List Full"); + return -1; + } + + mem_basy_y = &vsi->sf_ref_fb[idx].fb.base_y; + mem_basy_y->size = vsi->buf_sz_y_bs + + vsi->buf_len_sz_y; + + if (mtk_vcodec_mem_alloc(inst->ctx, mem_basy_y)) { + mtk_vdec_err(inst->ctx, "Cannot allocate sf_ref_buf y_buf"); + return -1; + } + + mem_basy_c = &vsi->sf_ref_fb[idx].fb.base_c; + mem_basy_c->size = vsi->buf_sz_c_bs + + vsi->buf_len_sz_c; + + if (mtk_vcodec_mem_alloc(inst->ctx, mem_basy_c)) { + mtk_vdec_err(inst->ctx, "Cannot allocate sf_ref_fb c_buf"); + return -1; + } + vsi->sf_ref_fb[idx].used = 0; + + return idx; +} + +static bool vp9_alloc_work_buf(struct vdec_vp9_inst *inst) +{ + struct vdec_vp9_vsi *vsi = inst->vsi; + int result; + struct mtk_vcodec_mem *mem; + + unsigned int max_pic_w; + unsigned int max_pic_h; + + + if (!(inst->ctx->dev->dec_capability & + VCODEC_CAPABILITY_4K_DISABLED)) { + max_pic_w = VCODEC_DEC_4K_CODED_WIDTH; + max_pic_h = VCODEC_DEC_4K_CODED_HEIGHT; + } else { + max_pic_w = MTK_VDEC_MAX_W; + max_pic_h = MTK_VDEC_MAX_H; + } + + if ((vsi->pic_w > max_pic_w) || + (vsi->pic_h > max_pic_h)) { + mtk_vdec_err(inst->ctx, "Invalid w/h %d/%d", vsi->pic_w, vsi->pic_h); + return false; + } + + mtk_vdec_debug(inst->ctx, "BUF CHG(%d): w/h/sb_w/sb_h=%d/%d/%d/%d", + vsi->resolution_changed, vsi->pic_w, + vsi->pic_h, vsi->buf_w, vsi->buf_h); + + mem = &inst->mv_buf; + if (mem->va) + mtk_vcodec_mem_free(inst->ctx, mem); + + mem->size = ((vsi->buf_w / 64) * + (vsi->buf_h / 64) + 2) * 36 * 16; + result = mtk_vcodec_mem_alloc(inst->ctx, mem); + if (result) { + mem->size = 0; + mtk_vdec_err(inst->ctx, "Cannot allocate mv_buf"); + return false; + } + /* Set the va again */ + vsi->mv_buf.va = (unsigned long)mem->va; + vsi->mv_buf.pa = (unsigned long)mem->dma_addr; + vsi->mv_buf.sz = (unsigned int)mem->size; + + + mem = &inst->seg_id_buf; + if (mem->va) + mtk_vcodec_mem_free(inst->ctx, mem); + + mem->size = VP9_SEG_ID_SZ; + result = mtk_vcodec_mem_alloc(inst->ctx, mem); + if (result) { + mem->size = 0; + mtk_vdec_err(inst->ctx, "Cannot allocate seg_id_buf"); + return false; + } + /* Set the va again */ + vsi->seg_id_buf.va = (unsigned long)mem->va; + vsi->seg_id_buf.pa = (unsigned long)mem->dma_addr; + vsi->seg_id_buf.sz = (unsigned int)mem->size; + + + vp9_free_all_sf_ref_fb(inst); + vsi->sf_next_ref_fb_idx = vp9_get_sf_ref_fb(inst); + + return true; +} + +static bool vp9_add_to_fb_disp_list(struct vdec_vp9_inst *inst, + struct vdec_fb *fb) +{ + struct vdec_fb_node *node; + + if (!fb) { + mtk_vdec_err(inst->ctx, "fb == NULL"); + return false; + } + + node = list_first_entry_or_null(&inst->available_fb_node_list, + struct vdec_fb_node, list); + if (node) { + node->fb = fb; + list_move_tail(&node->list, &inst->fb_disp_list); + } else { + mtk_vdec_err(inst->ctx, "No available fb node"); + return false; + } + + return true; +} + +/* If any buffer updating is signaled it should be done here. */ +static void vp9_swap_frm_bufs(struct vdec_vp9_inst *inst) +{ + struct vdec_vp9_vsi *vsi = inst->vsi; + struct vp9_fb_info *frm_to_show; + int ref_index = 0, mask; + + for (mask = vsi->refresh_frm_flags; mask; mask >>= 1) { + if (mask & 1) + vp9_ref_cnt_fb(inst, &vsi->ref_frm_map[ref_index], + vsi->new_fb_idx); + ++ref_index; + } + + frm_to_show = &vsi->frm_bufs[vsi->new_fb_idx].buf; + vsi->frm_bufs[vsi->new_fb_idx].ref_cnt--; + + if (frm_to_show->fb != inst->cur_fb) { + /* This frame is show exist frame and no decode output + * copy frame data from frm_to_show to current CAPTURE + * buffer + */ + if ((frm_to_show->fb != NULL) && + (inst->cur_fb->base_y.size >= + frm_to_show->fb->base_y.size) && + (inst->cur_fb->base_c.size >= + frm_to_show->fb->base_c.size)) { + memcpy((void *)inst->cur_fb->base_y.va, + (void *)frm_to_show->fb->base_y.va, + frm_to_show->fb->base_y.size); + memcpy((void *)inst->cur_fb->base_c.va, + (void *)frm_to_show->fb->base_c.va, + frm_to_show->fb->base_c.size); + } else { + /* After resolution change case, current CAPTURE buffer + * may have less buffer size than frm_to_show buffer + * size + */ + if (frm_to_show->fb != NULL) + mtk_vdec_err(inst->ctx, + "base_y.size=%zu, frm_to_show: base_y.size=%zu", + inst->cur_fb->base_y.size, + frm_to_show->fb->base_y.size); + } + if (!vp9_is_sf_ref_fb(inst, inst->cur_fb)) { + if (vsi->show_frame & BIT(0)) + vp9_add_to_fb_disp_list(inst, inst->cur_fb); + } + } else { + if (!vp9_is_sf_ref_fb(inst, inst->cur_fb)) { + if (vsi->show_frame & BIT(0)) + vp9_add_to_fb_disp_list(inst, frm_to_show->fb); + } + } + + /* when ref_cnt ==0, move this fb to fb_free_list. v4l2 driver will + * clean fb_free_list + */ + if (vsi->frm_bufs[vsi->new_fb_idx].ref_cnt == 0) { + if (!vp9_is_sf_ref_fb( + inst, vsi->frm_bufs[vsi->new_fb_idx].buf.fb)) { + struct vdec_fb *fb; + + fb = vp9_rm_from_fb_use_list(inst, + vsi->frm_bufs[vsi->new_fb_idx].buf.fb->base_y.va); + + vp9_add_to_fb_free_list(inst, fb); + } else { + vp9_free_sf_ref_fb( + vsi->frm_bufs[vsi->new_fb_idx].buf.fb); + } + } + + /* if this super frame and it is not last sub-frame, get next fb for + * sub-frame decode + */ + if (vsi->sf_frm_cnt > 0 && vsi->sf_frm_idx != vsi->sf_frm_cnt - 1) + vsi->sf_next_ref_fb_idx = vp9_get_sf_ref_fb(inst); +} + +static bool vp9_wait_dec_end(struct vdec_vp9_inst *inst) +{ + struct mtk_vcodec_dec_ctx *ctx = inst->ctx; + + mtk_vcodec_wait_for_done_ctx(inst->ctx, + MTK_INST_IRQ_RECEIVED, + WAIT_INTR_TIMEOUT_MS, 0); + + if (ctx->irq_status & MTK_VDEC_IRQ_STATUS_DEC_SUCCESS) + return true; + else + return false; +} + +static struct vdec_vp9_inst *vp9_alloc_inst(struct mtk_vcodec_dec_ctx *ctx) +{ + int result; + struct mtk_vcodec_mem mem; + struct vdec_vp9_inst *inst; + + memset(&mem, 0, sizeof(mem)); + mem.size = sizeof(struct vdec_vp9_inst); + result = mtk_vcodec_mem_alloc(ctx, &mem); + if (result) + return NULL; + + inst = mem.va; + inst->mem = mem; + + return inst; +} + +static void vp9_free_inst(struct vdec_vp9_inst *inst) +{ + struct mtk_vcodec_mem mem; + + mem = inst->mem; + if (mem.va) + mtk_vcodec_mem_free(inst->ctx, &mem); +} + +static bool vp9_decode_end_proc(struct vdec_vp9_inst *inst) +{ + struct vdec_vp9_vsi *vsi = inst->vsi; + bool ret = false; + + if (!vsi->show_existing_frame) { + ret = vp9_wait_dec_end(inst); + if (!ret) { + mtk_vdec_err(inst->ctx, "Decode failed, Decode Timeout @[%d]", + vsi->frm_num); + return false; + } + + if (vpu_dec_end(&inst->vpu)) { + mtk_vdec_err(inst->ctx, "vp9_dec_vpu_end failed"); + return false; + } + mtk_vdec_debug(inst->ctx, "Decode Ok @%d (%d/%d)", vsi->frm_num, + vsi->pic_w, vsi->pic_h); + } else { + mtk_vdec_debug(inst->ctx, "Decode Ok @%d (show_existing_frame)", vsi->frm_num); + } + + vp9_swap_frm_bufs(inst); + vsi->frm_num++; + return true; +} + +static bool vp9_is_last_sub_frm(struct vdec_vp9_inst *inst) +{ + struct vdec_vp9_vsi *vsi = inst->vsi; + + if (vsi->sf_frm_cnt <= 0 || vsi->sf_frm_idx == vsi->sf_frm_cnt) + return true; + + return false; +} + +static struct vdec_fb *vp9_rm_from_fb_disp_list(struct vdec_vp9_inst *inst) +{ + struct vdec_fb_node *node; + struct vdec_fb *fb = NULL; + + node = list_first_entry_or_null(&inst->fb_disp_list, + struct vdec_fb_node, list); + if (node) { + fb = (struct vdec_fb *)node->fb; + fb->status |= FB_ST_DISPLAY; + list_move_tail(&node->list, &inst->available_fb_node_list); + mtk_vdec_debug(inst->ctx, "[FB] get disp fb %p st=%d", node->fb, fb->status); + } else + mtk_vdec_debug(inst->ctx, "[FB] there is no disp fb"); + + return fb; +} + +static bool vp9_add_to_fb_use_list(struct vdec_vp9_inst *inst, + struct vdec_fb *fb) +{ + struct vdec_fb_node *node; + + if (!fb) { + mtk_vdec_debug(inst->ctx, "fb == NULL"); + return false; + } + + node = list_first_entry_or_null(&inst->available_fb_node_list, + struct vdec_fb_node, list); + if (node) { + node->fb = fb; + list_move_tail(&node->list, &inst->fb_use_list); + } else { + mtk_vdec_err(inst->ctx, "No free fb node"); + return false; + } + return true; +} + +static void vp9_reset(struct vdec_vp9_inst *inst) +{ + struct vdec_fb_node *node, *tmp; + + list_for_each_entry_safe(node, tmp, &inst->fb_use_list, list) + list_move_tail(&node->list, &inst->fb_free_list); + + vp9_free_all_sf_ref_fb(inst); + inst->vsi->sf_next_ref_fb_idx = vp9_get_sf_ref_fb(inst); + + if (vpu_dec_reset(&inst->vpu)) + mtk_vdec_err(inst->ctx, "vp9_dec_vpu_reset failed"); + + /* Set the va again, since vpu_dec_reset will clear mv_buf in vpu */ + inst->vsi->mv_buf.va = (unsigned long)inst->mv_buf.va; + inst->vsi->mv_buf.pa = (unsigned long)inst->mv_buf.dma_addr; + inst->vsi->mv_buf.sz = (unsigned long)inst->mv_buf.size; + + /* Set the va again, since vpu_dec_reset will clear seg_id_buf in vpu */ + inst->vsi->seg_id_buf.va = (unsigned long)inst->seg_id_buf.va; + inst->vsi->seg_id_buf.pa = (unsigned long)inst->seg_id_buf.dma_addr; + inst->vsi->seg_id_buf.sz = (unsigned long)inst->seg_id_buf.size; + +} + +static void init_all_fb_lists(struct vdec_vp9_inst *inst) +{ + int i; + + INIT_LIST_HEAD(&inst->available_fb_node_list); + INIT_LIST_HEAD(&inst->fb_use_list); + INIT_LIST_HEAD(&inst->fb_free_list); + INIT_LIST_HEAD(&inst->fb_disp_list); + + for (i = 0; i < ARRAY_SIZE(inst->dec_fb); i++) { + INIT_LIST_HEAD(&inst->dec_fb[i].list); + inst->dec_fb[i].fb = NULL; + list_add_tail(&inst->dec_fb[i].list, + &inst->available_fb_node_list); + } +} + +static void get_pic_info(struct vdec_vp9_inst *inst, struct vdec_pic_info *pic) +{ + pic->fb_sz[0] = inst->vsi->buf_sz_y_bs + inst->vsi->buf_len_sz_y; + pic->fb_sz[1] = inst->vsi->buf_sz_c_bs + inst->vsi->buf_len_sz_c; + + pic->pic_w = inst->vsi->pic_w; + pic->pic_h = inst->vsi->pic_h; + pic->buf_w = inst->vsi->buf_w; + pic->buf_h = inst->vsi->buf_h; + + mtk_vdec_debug(inst->ctx, "pic(%d, %d), buf(%d, %d)", + pic->pic_w, pic->pic_h, pic->buf_w, pic->buf_h); + mtk_vdec_debug(inst->ctx, "fb size: Y(%d), C(%d)", pic->fb_sz[0], pic->fb_sz[1]); +} + +static void get_disp_fb(struct vdec_vp9_inst *inst, struct vdec_fb **out_fb) +{ + + *out_fb = vp9_rm_from_fb_disp_list(inst); + if (*out_fb) + (*out_fb)->status |= FB_ST_DISPLAY; +} + +static void get_free_fb(struct vdec_vp9_inst *inst, struct vdec_fb **out_fb) +{ + struct vdec_fb_node *node; + struct vdec_fb *fb = NULL; + + node = list_first_entry_or_null(&inst->fb_free_list, + struct vdec_fb_node, list); + if (node) { + list_move_tail(&node->list, &inst->available_fb_node_list); + fb = (struct vdec_fb *)node->fb; + fb->status |= FB_ST_FREE; + mtk_vdec_debug(inst->ctx, "[FB] get free fb %p st=%d", node->fb, fb->status); + } else { + mtk_vdec_debug(inst->ctx, "[FB] there is no free fb"); + } + + *out_fb = fb; +} + +static int validate_vsi_array_indexes(struct vdec_vp9_inst *inst, + struct vdec_vp9_vsi *vsi) { + if (vsi->sf_frm_idx >= VP9_MAX_FRM_BUF_NUM - 1) { + mtk_vdec_err(inst->ctx, "Invalid vsi->sf_frm_idx=%u.", vsi->sf_frm_idx); + return -EIO; + } + if (vsi->frm_to_show_idx >= VP9_MAX_FRM_BUF_NUM) { + mtk_vdec_err(inst->ctx, "Invalid vsi->frm_to_show_idx=%u.", vsi->frm_to_show_idx); + return -EIO; + } + if (vsi->new_fb_idx >= VP9_MAX_FRM_BUF_NUM) { + mtk_vdec_err(inst->ctx, "Invalid vsi->new_fb_idx=%u.", vsi->new_fb_idx); + return -EIO; + } + return 0; +} + +static void vdec_vp9_deinit(void *h_vdec) +{ + struct vdec_vp9_inst *inst = (struct vdec_vp9_inst *)h_vdec; + struct mtk_vcodec_mem *mem; + int ret = 0; + + ret = vpu_dec_deinit(&inst->vpu); + if (ret) + mtk_vdec_err(inst->ctx, "vpu_dec_deinit failed"); + + mem = &inst->mv_buf; + if (mem->va) + mtk_vcodec_mem_free(inst->ctx, mem); + + mem = &inst->seg_id_buf; + if (mem->va) + mtk_vcodec_mem_free(inst->ctx, mem); + + vp9_free_all_sf_ref_fb(inst); + vp9_free_inst(inst); +} + +static int vdec_vp9_init(struct mtk_vcodec_dec_ctx *ctx) +{ + struct vdec_vp9_inst *inst; + + inst = vp9_alloc_inst(ctx); + if (!inst) + return -ENOMEM; + + inst->total_frm_cnt = 0; + inst->ctx = ctx; + + inst->vpu.id = IPI_VDEC_VP9; + inst->vpu.ctx = ctx; + + if (vpu_dec_init(&inst->vpu)) { + mtk_vdec_err(inst->ctx, "vp9_dec_vpu_init failed"); + goto err_deinit_inst; + } + + inst->vsi = (struct vdec_vp9_vsi *)inst->vpu.vsi; + + inst->vsi->show_frame |= BIT(3); + + init_all_fb_lists(inst); + + ctx->drv_handle = inst; + return 0; + +err_deinit_inst: + vp9_free_inst(inst); + + return -EINVAL; +} + +static int vdec_vp9_decode(void *h_vdec, struct mtk_vcodec_mem *bs, + struct vdec_fb *fb, bool *res_chg) +{ + int ret = 0; + struct vdec_vp9_inst *inst = (struct vdec_vp9_inst *)h_vdec; + struct vdec_vp9_vsi *vsi = inst->vsi; + u32 data[3]; + int i; + + *res_chg = false; + + if ((bs == NULL) && (fb == NULL)) { + mtk_vdec_debug(inst->ctx, "[EOS]"); + vp9_reset(inst); + return ret; + } + + if (bs == NULL) { + mtk_vdec_err(inst->ctx, "bs == NULL"); + return -EINVAL; + } + + mtk_vdec_debug(inst->ctx, "Input BS Size = %zu", bs->size); + + while (1) { + struct vdec_fb *cur_fb = NULL; + + data[0] = *((unsigned int *)bs->va); + data[1] = *((unsigned int *)(bs->va + 4)); + data[2] = *((unsigned int *)(bs->va + 8)); + + vsi->bs = *bs; + + if (fb) + vsi->fb = *fb; + + if (!vsi->sf_init) { + unsigned int sf_bs_sz; + unsigned int sf_bs_off; + unsigned char *sf_bs_src; + unsigned char *sf_bs_dst; + + sf_bs_sz = bs->size > VP9_SUPER_FRAME_BS_SZ ? + VP9_SUPER_FRAME_BS_SZ : bs->size; + sf_bs_off = VP9_SUPER_FRAME_BS_SZ - sf_bs_sz; + sf_bs_src = bs->va + bs->size - sf_bs_sz; + sf_bs_dst = vsi->sf_bs_buf + sf_bs_off; + memcpy(sf_bs_dst, sf_bs_src, sf_bs_sz); + } else { + if ((vsi->sf_frm_cnt > 0) && + (vsi->sf_frm_idx < vsi->sf_frm_cnt)) { + unsigned int idx = vsi->sf_frm_idx; + + memcpy((void *)bs->va, + (void *)(bs->va + + vsi->sf_frm_offset[idx]), + vsi->sf_frm_sz[idx]); + } + } + + if (!(vsi->show_frame & BIT(4))) + memset(inst->seg_id_buf.va, 0, inst->seg_id_buf.size); + + ret = vpu_dec_start(&inst->vpu, data, 3); + if (ret) { + mtk_vdec_err(inst->ctx, "vpu_dec_start failed"); + goto DECODE_ERROR; + } + + if (vsi->show_frame & BIT(1)) { + memset(inst->seg_id_buf.va, 0, inst->seg_id_buf.size); + + if (vsi->show_frame & BIT(2)) { + ret = vpu_dec_start(&inst->vpu, NULL, 0); + if (ret) { + mtk_vdec_err(inst->ctx, "vpu trig decoder failed"); + goto DECODE_ERROR; + } + } + } + + ret = validate_vsi_array_indexes(inst, vsi); + if (ret) { + mtk_vdec_err(inst->ctx, "Invalid values from VPU."); + goto DECODE_ERROR; + } + + if (vsi->resolution_changed) { + if (!vp9_alloc_work_buf(inst)) { + ret = -EIO; + goto DECODE_ERROR; + } + } + + if (vsi->sf_frm_cnt > 0) { + cur_fb = &vsi->sf_ref_fb[vsi->sf_next_ref_fb_idx].fb; + + if (vsi->sf_frm_idx < vsi->sf_frm_cnt) + inst->cur_fb = cur_fb; + else + inst->cur_fb = fb; + } else { + inst->cur_fb = fb; + } + + vsi->frm_bufs[vsi->new_fb_idx].buf.fb = inst->cur_fb; + if (!vp9_is_sf_ref_fb(inst, inst->cur_fb)) + vp9_add_to_fb_use_list(inst, inst->cur_fb); + + mtk_vdec_debug(inst->ctx, "[#pic %d]", vsi->frm_num); + + if (vsi->show_existing_frame) + mtk_vdec_debug(inst->ctx, + "drv->new_fb_idx=%d, drv->frm_to_show_idx=%d", + vsi->new_fb_idx, vsi->frm_to_show_idx); + + if (vsi->show_existing_frame && (vsi->frm_to_show_idx < + VP9_MAX_FRM_BUF_NUM)) { + mtk_vdec_debug(inst->ctx, + "Skip Decode drv->new_fb_idx=%d, drv->frm_to_show_idx=%d", + vsi->new_fb_idx, vsi->frm_to_show_idx); + + vp9_ref_cnt_fb(inst, &vsi->new_fb_idx, + vsi->frm_to_show_idx); + } + + /* VPU assign the buffer pointer in its address space, + * reassign here + */ + for (i = 0; i < ARRAY_SIZE(vsi->frm_refs); i++) { + unsigned int idx = vsi->frm_refs[i].idx; + + vsi->frm_refs[i].buf = &vsi->frm_bufs[idx].buf; + } + + if (vsi->resolution_changed) { + *res_chg = true; + mtk_vdec_debug(inst->ctx, "VDEC_ST_RESOLUTION_CHANGED"); + + ret = 0; + goto DECODE_ERROR; + } + + if (!vp9_decode_end_proc(inst)) { + mtk_vdec_err(inst->ctx, "vp9_decode_end_proc"); + ret = -EINVAL; + goto DECODE_ERROR; + } + + if (vp9_is_last_sub_frm(inst)) + break; + + } + inst->total_frm_cnt++; + +DECODE_ERROR: + if (ret < 0) + vp9_add_to_fb_free_list(inst, fb); + + return ret; +} + +static void get_crop_info(struct vdec_vp9_inst *inst, struct v4l2_rect *cr) +{ + cr->left = 0; + cr->top = 0; + cr->width = inst->vsi->pic_w; + cr->height = inst->vsi->pic_h; + mtk_vdec_debug(inst->ctx, "get crop info l=%d, t=%d, w=%d, h=%d\n", + cr->left, cr->top, cr->width, cr->height); +} + +static int vdec_vp9_get_param(void *h_vdec, enum vdec_get_param_type type, + void *out) +{ + struct vdec_vp9_inst *inst = (struct vdec_vp9_inst *)h_vdec; + int ret = 0; + + switch (type) { + case GET_PARAM_DISP_FRAME_BUFFER: + get_disp_fb(inst, out); + break; + case GET_PARAM_FREE_FRAME_BUFFER: + get_free_fb(inst, out); + break; + case GET_PARAM_PIC_INFO: + get_pic_info(inst, out); + break; + case GET_PARAM_DPB_SIZE: + *((unsigned int *)out) = MAX_VP9_DPB_SIZE; + break; + case GET_PARAM_CROP_INFO: + get_crop_info(inst, out); + break; + default: + mtk_vdec_err(inst->ctx, "not supported param type %d", type); + ret = -EINVAL; + break; + } + + return ret; +} + +const struct vdec_common_if vdec_vp9_if = { + .init = vdec_vp9_init, + .decode = vdec_vp9_decode, + .get_param = vdec_vp9_get_param, + .deinit = vdec_vp9_deinit, +}; diff --git a/drivers/media/platform/mediatek/vcodec/decoder/vdec/vdec_vp9_req_lat_if.c b/drivers/media/platform/mediatek/vcodec/decoder/vdec/vdec_vp9_req_lat_if.c new file mode 100644 index 0000000000..e393e3e668 --- /dev/null +++ b/drivers/media/platform/mediatek/vcodec/decoder/vdec/vdec_vp9_req_lat_if.c @@ -0,0 +1,2216 @@ +// SPDX-License-Identifier: GPL-2.0 +/* + * Copyright (c) 2021 MediaTek Inc. + * Author: George Sun <george.sun@mediatek.com> + */ + +#include <linux/module.h> +#include <linux/slab.h> +#include <media/videobuf2-dma-contig.h> +#include <media/v4l2-vp9.h> + +#include "../mtk_vcodec_dec.h" +#include "../../common/mtk_vcodec_intr.h" +#include "../vdec_drv_base.h" +#include "../vdec_drv_if.h" +#include "../vdec_vpu_if.h" + +/* reset_frame_context defined in VP9 spec */ +#define VP9_RESET_FRAME_CONTEXT_NONE0 0 +#define VP9_RESET_FRAME_CONTEXT_NONE1 1 +#define VP9_RESET_FRAME_CONTEXT_SPEC 2 +#define VP9_RESET_FRAME_CONTEXT_ALL 3 + +#define VP9_TILE_BUF_SIZE 4096 +#define VP9_PROB_BUF_SIZE 2560 +#define VP9_COUNTS_BUF_SIZE 16384 + +#define HDR_FLAG(x) (!!((hdr)->flags & V4L2_VP9_FRAME_FLAG_##x)) +#define LF_FLAG(x) (!!((lf)->flags & V4L2_VP9_LOOP_FILTER_FLAG_##x)) +#define SEG_FLAG(x) (!!((seg)->flags & V4L2_VP9_SEGMENTATION_FLAG_##x)) +#define VP9_BAND_6(band) ((band) == 0 ? 3 : 6) + +/* + * struct vdec_vp9_slice_frame_ctx - vp9 prob tables footprint + */ +struct vdec_vp9_slice_frame_ctx { + struct { + u8 probs[6][3]; + u8 padding[2]; + } coef_probs[4][2][2][6]; + + u8 y_mode_prob[4][16]; + u8 switch_interp_prob[4][16]; + u8 seg[32]; /* ignore */ + u8 comp_inter_prob[16]; + u8 comp_ref_prob[16]; + u8 single_ref_prob[5][2]; + u8 single_ref_prob_padding[6]; + + u8 joint[3]; + u8 joint_padding[13]; + struct { + u8 sign; + u8 classes[10]; + u8 padding[5]; + } sign_classes[2]; + struct { + u8 class0[1]; + u8 bits[10]; + u8 padding[5]; + } class0_bits[2]; + struct { + u8 class0_fp[2][3]; + u8 fp[3]; + u8 class0_hp; + u8 hp; + u8 padding[5]; + } class0_fp_hp[2]; + + u8 uv_mode_prob[10][16]; + u8 uv_mode_prob_padding[2][16]; + + u8 partition_prob[16][4]; + + u8 inter_mode_probs[7][4]; + u8 skip_probs[4]; + + u8 tx_p8x8[2][4]; + u8 tx_p16x16[2][4]; + u8 tx_p32x32[2][4]; + u8 intra_inter_prob[8]; +}; + +/* + * struct vdec_vp9_slice_frame_counts - vp9 counts tables footprint + */ +struct vdec_vp9_slice_frame_counts { + union { + struct { + u32 band_0[3]; + u32 padding0[1]; + u32 band_1_5[5][6]; + u32 padding1[2]; + } eob_branch[4][2][2]; + u32 eob_branch_space[256 * 4]; + }; + + struct { + u32 band_0[3][4]; + u32 band_1_5[5][6][4]; + } coef_probs[4][2][2]; + + u32 intra_inter[4][2]; + u32 comp_inter[5][2]; + u32 comp_inter_padding[2]; + u32 comp_ref[5][2]; + u32 comp_ref_padding[2]; + u32 single_ref[5][2][2]; + u32 inter_mode[7][4]; + u32 y_mode[4][12]; + u32 uv_mode[10][10]; + u32 partition[16][4]; + u32 switchable_interp[4][4]; + + u32 tx_p8x8[2][2]; + u32 tx_p16x16[2][4]; + u32 tx_p32x32[2][4]; + + u32 skip[3][4]; + + u32 joint[4]; + + struct { + u32 sign[2]; + u32 class0[2]; + u32 classes[12]; + u32 bits[10][2]; + u32 padding[4]; + u32 class0_fp[2][4]; + u32 fp[4]; + u32 class0_hp[2]; + u32 hp[2]; + } mvcomp[2]; + + u32 reserved[126][4]; +}; + +/** + * struct vdec_vp9_slice_counts_map - vp9 counts tables to map + * v4l2_vp9_frame_symbol_counts + * @skip: skip counts. + * @y_mode: Y prediction mode counts. + * @filter: interpolation filter counts. + * @mv_joint: motion vector joint counts. + * @sign: motion vector sign counts. + * @classes: motion vector class counts. + * @class0: motion vector class0 bit counts. + * @bits: motion vector bits counts. + * @class0_fp: motion vector class0 fractional bit counts. + * @fp: motion vector fractional bit counts. + * @class0_hp: motion vector class0 high precision fractional bit counts. + * @hp: motion vector high precision fractional bit counts. + */ +struct vdec_vp9_slice_counts_map { + u32 skip[3][2]; + u32 y_mode[4][10]; + u32 filter[4][3]; + u32 sign[2][2]; + u32 classes[2][11]; + u32 class0[2][2]; + u32 bits[2][10][2]; + u32 class0_fp[2][2][4]; + u32 fp[2][4]; + u32 class0_hp[2][2]; + u32 hp[2][2]; +}; + +/* + * struct vdec_vp9_slice_uncompressed_header - vp9 uncompressed header syntax + * used for decoding + */ +struct vdec_vp9_slice_uncompressed_header { + u8 profile; + u8 last_frame_type; + u8 frame_type; + + u8 last_show_frame; + u8 show_frame; + u8 error_resilient_mode; + + u8 bit_depth; + u8 padding0[1]; + u16 last_frame_width; + u16 last_frame_height; + u16 frame_width; + u16 frame_height; + + u8 intra_only; + u8 reset_frame_context; + u8 ref_frame_sign_bias[4]; + u8 allow_high_precision_mv; + u8 interpolation_filter; + + u8 refresh_frame_context; + u8 frame_parallel_decoding_mode; + u8 frame_context_idx; + + /* loop_filter_params */ + u8 loop_filter_level; + u8 loop_filter_sharpness; + u8 loop_filter_delta_enabled; + s8 loop_filter_ref_deltas[4]; + s8 loop_filter_mode_deltas[2]; + + /* quantization_params */ + u8 base_q_idx; + s8 delta_q_y_dc; + s8 delta_q_uv_dc; + s8 delta_q_uv_ac; + + /* segmentation_params */ + u8 segmentation_enabled; + u8 segmentation_update_map; + u8 segmentation_tree_probs[7]; + u8 padding1[1]; + u8 segmentation_temporal_udpate; + u8 segmentation_pred_prob[3]; + u8 segmentation_update_data; + u8 segmentation_abs_or_delta_update; + u8 feature_enabled[8]; + s16 feature_value[8][4]; + + /* tile_info */ + u8 tile_cols_log2; + u8 tile_rows_log2; + u8 padding2[2]; + + u16 uncompressed_header_size; + u16 header_size_in_bytes; + + /* LAT OUT, CORE IN */ + u32 dequant[8][4]; +}; + +/* + * struct vdec_vp9_slice_compressed_header - vp9 compressed header syntax + * used for decoding. + */ +struct vdec_vp9_slice_compressed_header { + u8 tx_mode; + u8 ref_mode; + u8 comp_fixed_ref; + u8 comp_var_ref[2]; + u8 padding[3]; +}; + +/* + * struct vdec_vp9_slice_tiles - vp9 tile syntax + */ +struct vdec_vp9_slice_tiles { + u32 size[4][64]; + u32 mi_rows[4]; + u32 mi_cols[64]; + u8 actual_rows; + u8 padding[7]; +}; + +/* + * struct vdec_vp9_slice_reference - vp9 reference frame information + */ +struct vdec_vp9_slice_reference { + u16 frame_width; + u16 frame_height; + u8 bit_depth; + u8 subsampling_x; + u8 subsampling_y; + u8 padding; +}; + +/* + * struct vdec_vp9_slice_frame - vp9 syntax used for decoding + */ +struct vdec_vp9_slice_frame { + struct vdec_vp9_slice_uncompressed_header uh; + struct vdec_vp9_slice_compressed_header ch; + struct vdec_vp9_slice_tiles tiles; + struct vdec_vp9_slice_reference ref[3]; +}; + +/* + * struct vdec_vp9_slice_init_vsi - VSI used to initialize instance + */ +struct vdec_vp9_slice_init_vsi { + unsigned int architecture; + unsigned int reserved; + u64 core_vsi; + /* default frame context's position in MicroP */ + u64 default_frame_ctx; +}; + +/* + * struct vdec_vp9_slice_mem - memory address and size + */ +struct vdec_vp9_slice_mem { + union { + u64 buf; + dma_addr_t dma_addr; + }; + union { + size_t size; + dma_addr_t dma_addr_end; + u64 padding; + }; +}; + +/* + * struct vdec_vp9_slice_bs - input buffer for decoding + */ +struct vdec_vp9_slice_bs { + struct vdec_vp9_slice_mem buf; + struct vdec_vp9_slice_mem frame; +}; + +/* + * struct vdec_vp9_slice_fb - frame buffer for decoding + */ +struct vdec_vp9_slice_fb { + struct vdec_vp9_slice_mem y; + struct vdec_vp9_slice_mem c; +}; + +/* + * struct vdec_vp9_slice_state - decoding state + */ +struct vdec_vp9_slice_state { + int err; + unsigned int full; + unsigned int timeout; + unsigned int perf; + + unsigned int crc[12]; +}; + +/** + * struct vdec_vp9_slice_vsi - exchange decoding information + * between Main CPU and MicroP + * + * @bs: input buffer + * @fb: output buffer + * @ref: 3 reference buffers + * @mv: mv working buffer + * @seg: segmentation working buffer + * @tile: tile buffer + * @prob: prob table buffer, used to set/update prob table + * @counts: counts table buffer, used to update prob table + * @ube: general buffer + * @trans: trans buffer position in general buffer + * @err_map: error buffer + * @row_info: row info buffer + * @frame: decoding syntax + * @state: decoding state + */ +struct vdec_vp9_slice_vsi { + /* used in LAT stage */ + struct vdec_vp9_slice_bs bs; + /* used in Core stage */ + struct vdec_vp9_slice_fb fb; + struct vdec_vp9_slice_fb ref[3]; + + struct vdec_vp9_slice_mem mv[2]; + struct vdec_vp9_slice_mem seg[2]; + struct vdec_vp9_slice_mem tile; + struct vdec_vp9_slice_mem prob; + struct vdec_vp9_slice_mem counts; + + /* LAT stage's output, Core stage's input */ + struct vdec_vp9_slice_mem ube; + struct vdec_vp9_slice_mem trans; + struct vdec_vp9_slice_mem err_map; + struct vdec_vp9_slice_mem row_info; + + /* decoding parameters */ + struct vdec_vp9_slice_frame frame; + + struct vdec_vp9_slice_state state; +}; + +/** + * struct vdec_vp9_slice_pfc - per-frame context that contains a local vsi. + * pass it from lat to core + * + * @vsi: local vsi. copy to/from remote vsi before/after decoding + * @ref_idx: reference buffer index + * @seq: picture sequence + * @state: decoding state + */ +struct vdec_vp9_slice_pfc { + struct vdec_vp9_slice_vsi vsi; + + u64 ref_idx[3]; + + int seq; + + /* LAT/Core CRC */ + struct vdec_vp9_slice_state state[2]; +}; + +/* + * enum vdec_vp9_slice_resolution_level + */ +enum vdec_vp9_slice_resolution_level { + VP9_RES_NONE, + VP9_RES_FHD, + VP9_RES_4K, + VP9_RES_8K, +}; + +/* + * struct vdec_vp9_slice_ref - picture's width & height should kept + * for later decoding as reference picture + */ +struct vdec_vp9_slice_ref { + unsigned int width; + unsigned int height; +}; + +/** + * struct vdec_vp9_slice_instance - represent one vp9 instance + * + * @ctx: pointer to codec's context + * @vpu: VPU instance + * @seq: global picture sequence + * @level: level of current resolution + * @width: width of last picture + * @height: height of last picture + * @frame_type: frame_type of last picture + * @irq: irq to Main CPU or MicroP + * @show_frame: show_frame of last picture + * @dpb: picture information (width/height) for reference + * @mv: mv working buffer + * @seg: segmentation working buffer + * @tile: tile buffer + * @prob: prob table buffer, used to set/update prob table + * @counts: counts table buffer, used to update prob table + * @frame_ctx: 4 frame context according to VP9 Spec + * @frame_ctx_helper: 4 frame context according to newest kernel spec + * @dirty: state of each frame context + * @init_vsi: vsi used for initialized VP9 instance + * @vsi: vsi used for decoding/flush ... + * @core_vsi: vsi used for Core stage + * + * @sc_pfc: per frame context single core + * @counts_map: used map to counts_helper + * @counts_helper: counts table according to newest kernel spec + */ +struct vdec_vp9_slice_instance { + struct mtk_vcodec_dec_ctx *ctx; + struct vdec_vpu_inst vpu; + + int seq; + + enum vdec_vp9_slice_resolution_level level; + + /* for resolution change and get_pic_info */ + unsigned int width; + unsigned int height; + + /* for last_frame_type */ + unsigned int frame_type; + unsigned int irq; + + unsigned int show_frame; + + /* maintain vp9 reference frame state */ + struct vdec_vp9_slice_ref dpb[VB2_MAX_FRAME]; + + /* + * normal working buffers + * mv[0]/seg[0]/tile/prob/counts is used for LAT + * mv[1]/seg[1] is used for CORE + */ + struct mtk_vcodec_mem mv[2]; + struct mtk_vcodec_mem seg[2]; + struct mtk_vcodec_mem tile; + struct mtk_vcodec_mem prob; + struct mtk_vcodec_mem counts; + + /* 4 prob tables */ + struct vdec_vp9_slice_frame_ctx frame_ctx[4]; + /*4 helper tables */ + struct v4l2_vp9_frame_context frame_ctx_helper; + unsigned char dirty[4]; + + /* MicroP vsi */ + union { + struct vdec_vp9_slice_init_vsi *init_vsi; + struct vdec_vp9_slice_vsi *vsi; + }; + struct vdec_vp9_slice_vsi *core_vsi; + + struct vdec_vp9_slice_pfc sc_pfc; + struct vdec_vp9_slice_counts_map counts_map; + struct v4l2_vp9_frame_symbol_counts counts_helper; +}; + +/* + * all VP9 instances could share this default frame context. + */ +static struct vdec_vp9_slice_frame_ctx *vdec_vp9_slice_default_frame_ctx; +static DEFINE_MUTEX(vdec_vp9_slice_frame_ctx_lock); + +static int vdec_vp9_slice_core_decode(struct vdec_lat_buf *lat_buf); + +static int vdec_vp9_slice_init_default_frame_ctx(struct vdec_vp9_slice_instance *instance) +{ + struct vdec_vp9_slice_frame_ctx *remote_frame_ctx; + struct vdec_vp9_slice_frame_ctx *frame_ctx; + struct mtk_vcodec_dec_ctx *ctx; + struct vdec_vp9_slice_init_vsi *vsi; + int ret = 0; + + ctx = instance->ctx; + vsi = instance->vpu.vsi; + if (!ctx || !vsi) + return -EINVAL; + + remote_frame_ctx = mtk_vcodec_fw_map_dm_addr(ctx->dev->fw_handler, + (u32)vsi->default_frame_ctx); + if (!remote_frame_ctx) { + mtk_vdec_err(ctx, "failed to map default frame ctx\n"); + return -EINVAL; + } + + mutex_lock(&vdec_vp9_slice_frame_ctx_lock); + if (vdec_vp9_slice_default_frame_ctx) + goto out; + + frame_ctx = kmemdup(remote_frame_ctx, sizeof(*frame_ctx), GFP_KERNEL); + if (!frame_ctx) { + ret = -ENOMEM; + goto out; + } + + vdec_vp9_slice_default_frame_ctx = frame_ctx; + +out: + mutex_unlock(&vdec_vp9_slice_frame_ctx_lock); + + return ret; +} + +static int vdec_vp9_slice_alloc_working_buffer(struct vdec_vp9_slice_instance *instance, + struct vdec_vp9_slice_vsi *vsi) +{ + struct mtk_vcodec_dec_ctx *ctx = instance->ctx; + enum vdec_vp9_slice_resolution_level level; + /* super blocks */ + unsigned int max_sb_w; + unsigned int max_sb_h; + unsigned int max_w; + unsigned int max_h; + unsigned int w; + unsigned int h; + size_t size; + int ret; + int i; + + w = vsi->frame.uh.frame_width; + h = vsi->frame.uh.frame_height; + + if (w > VCODEC_DEC_4K_CODED_WIDTH || + h > VCODEC_DEC_4K_CODED_HEIGHT) { + return -EINVAL; + } else if (w > MTK_VDEC_MAX_W || h > MTK_VDEC_MAX_H) { + /* 4K */ + level = VP9_RES_4K; + max_w = VCODEC_DEC_4K_CODED_WIDTH; + max_h = VCODEC_DEC_4K_CODED_HEIGHT; + } else { + /* FHD */ + level = VP9_RES_FHD; + max_w = MTK_VDEC_MAX_W; + max_h = MTK_VDEC_MAX_H; + } + + if (level == instance->level) + return 0; + + mtk_vdec_debug(ctx, "resolution level changed, from %u to %u, %ux%u", + instance->level, level, w, h); + + max_sb_w = DIV_ROUND_UP(max_w, 64); + max_sb_h = DIV_ROUND_UP(max_h, 64); + ret = -ENOMEM; + + /* + * Lat-flush must wait core idle, otherwise core will + * use released buffers + */ + + size = (max_sb_w * max_sb_h + 2) * 576; + for (i = 0; i < 2; i++) { + if (instance->mv[i].va) + mtk_vcodec_mem_free(ctx, &instance->mv[i]); + instance->mv[i].size = size; + if (mtk_vcodec_mem_alloc(ctx, &instance->mv[i])) + goto err; + } + + size = (max_sb_w * max_sb_h * 32) + 256; + for (i = 0; i < 2; i++) { + if (instance->seg[i].va) + mtk_vcodec_mem_free(ctx, &instance->seg[i]); + instance->seg[i].size = size; + if (mtk_vcodec_mem_alloc(ctx, &instance->seg[i])) + goto err; + } + + if (!instance->tile.va) { + instance->tile.size = VP9_TILE_BUF_SIZE; + if (mtk_vcodec_mem_alloc(ctx, &instance->tile)) + goto err; + } + + if (!instance->prob.va) { + instance->prob.size = VP9_PROB_BUF_SIZE; + if (mtk_vcodec_mem_alloc(ctx, &instance->prob)) + goto err; + } + + if (!instance->counts.va) { + instance->counts.size = VP9_COUNTS_BUF_SIZE; + if (mtk_vcodec_mem_alloc(ctx, &instance->counts)) + goto err; + } + + instance->level = level; + return 0; + +err: + instance->level = VP9_RES_NONE; + return ret; +} + +static void vdec_vp9_slice_free_working_buffer(struct vdec_vp9_slice_instance *instance) +{ + struct mtk_vcodec_dec_ctx *ctx = instance->ctx; + int i; + + for (i = 0; i < ARRAY_SIZE(instance->mv); i++) { + if (instance->mv[i].va) + mtk_vcodec_mem_free(ctx, &instance->mv[i]); + } + for (i = 0; i < ARRAY_SIZE(instance->seg); i++) { + if (instance->seg[i].va) + mtk_vcodec_mem_free(ctx, &instance->seg[i]); + } + if (instance->tile.va) + mtk_vcodec_mem_free(ctx, &instance->tile); + if (instance->prob.va) + mtk_vcodec_mem_free(ctx, &instance->prob); + if (instance->counts.va) + mtk_vcodec_mem_free(ctx, &instance->counts); + + instance->level = VP9_RES_NONE; +} + +static void vdec_vp9_slice_vsi_from_remote(struct vdec_vp9_slice_vsi *vsi, + struct vdec_vp9_slice_vsi *remote_vsi, + int skip) +{ + struct vdec_vp9_slice_frame *rf; + struct vdec_vp9_slice_frame *f; + + /* + * compressed header + * dequant + * buffer position + * decode state + */ + if (!skip) { + rf = &remote_vsi->frame; + f = &vsi->frame; + memcpy(&f->ch, &rf->ch, sizeof(f->ch)); + memcpy(&f->uh.dequant, &rf->uh.dequant, sizeof(f->uh.dequant)); + memcpy(&vsi->trans, &remote_vsi->trans, sizeof(vsi->trans)); + } + + memcpy(&vsi->state, &remote_vsi->state, sizeof(vsi->state)); +} + +static void vdec_vp9_slice_vsi_to_remote(struct vdec_vp9_slice_vsi *vsi, + struct vdec_vp9_slice_vsi *remote_vsi) +{ + memcpy(remote_vsi, vsi, sizeof(*vsi)); +} + +static int vdec_vp9_slice_tile_offset(int idx, int mi_num, int tile_log2) +{ + int sbs = (mi_num + 7) >> 3; + int offset = ((idx * sbs) >> tile_log2) << 3; + + return min(offset, mi_num); +} + +static +int vdec_vp9_slice_setup_single_from_src_to_dst(struct vdec_vp9_slice_instance *instance) +{ + struct vb2_v4l2_buffer *src; + struct vb2_v4l2_buffer *dst; + + src = v4l2_m2m_next_src_buf(instance->ctx->m2m_ctx); + if (!src) + return -EINVAL; + + dst = v4l2_m2m_next_dst_buf(instance->ctx->m2m_ctx); + if (!dst) + return -EINVAL; + + v4l2_m2m_buf_copy_metadata(src, dst, true); + + return 0; +} + +static int vdec_vp9_slice_setup_lat_from_src_buf(struct vdec_vp9_slice_instance *instance, + struct vdec_lat_buf *lat_buf) +{ + struct vb2_v4l2_buffer *src; + struct vb2_v4l2_buffer *dst; + + src = v4l2_m2m_next_src_buf(instance->ctx->m2m_ctx); + if (!src) + return -EINVAL; + + lat_buf->src_buf_req = src->vb2_buf.req_obj.req; + + dst = &lat_buf->ts_info; + v4l2_m2m_buf_copy_metadata(src, dst, true); + return 0; +} + +static void vdec_vp9_slice_setup_hdr(struct vdec_vp9_slice_instance *instance, + struct vdec_vp9_slice_uncompressed_header *uh, + struct v4l2_ctrl_vp9_frame *hdr) +{ + int i; + + uh->profile = hdr->profile; + uh->last_frame_type = instance->frame_type; + uh->frame_type = !HDR_FLAG(KEY_FRAME); + uh->last_show_frame = instance->show_frame; + uh->show_frame = HDR_FLAG(SHOW_FRAME); + uh->error_resilient_mode = HDR_FLAG(ERROR_RESILIENT); + uh->bit_depth = hdr->bit_depth; + uh->last_frame_width = instance->width; + uh->last_frame_height = instance->height; + uh->frame_width = hdr->frame_width_minus_1 + 1; + uh->frame_height = hdr->frame_height_minus_1 + 1; + uh->intra_only = HDR_FLAG(INTRA_ONLY); + /* map v4l2 enum to values defined in VP9 spec for firmware */ + switch (hdr->reset_frame_context) { + case V4L2_VP9_RESET_FRAME_CTX_NONE: + uh->reset_frame_context = VP9_RESET_FRAME_CONTEXT_NONE0; + break; + case V4L2_VP9_RESET_FRAME_CTX_SPEC: + uh->reset_frame_context = VP9_RESET_FRAME_CONTEXT_SPEC; + break; + case V4L2_VP9_RESET_FRAME_CTX_ALL: + uh->reset_frame_context = VP9_RESET_FRAME_CONTEXT_ALL; + break; + default: + uh->reset_frame_context = VP9_RESET_FRAME_CONTEXT_NONE0; + break; + } + /* + * ref_frame_sign_bias specifies the intended direction + * of the motion vector in time for each reference frame. + * - INTRA_FRAME = 0, + * - LAST_FRAME = 1, + * - GOLDEN_FRAME = 2, + * - ALTREF_FRAME = 3, + * ref_frame_sign_bias[INTRA_FRAME] is always 0 + * and VDA only passes another 3 directions + */ + uh->ref_frame_sign_bias[0] = 0; + for (i = 0; i < 3; i++) + uh->ref_frame_sign_bias[i + 1] = + !!(hdr->ref_frame_sign_bias & (1 << i)); + uh->allow_high_precision_mv = HDR_FLAG(ALLOW_HIGH_PREC_MV); + uh->interpolation_filter = hdr->interpolation_filter; + uh->refresh_frame_context = HDR_FLAG(REFRESH_FRAME_CTX); + uh->frame_parallel_decoding_mode = HDR_FLAG(PARALLEL_DEC_MODE); + uh->frame_context_idx = hdr->frame_context_idx; + + /* tile info */ + uh->tile_cols_log2 = hdr->tile_cols_log2; + uh->tile_rows_log2 = hdr->tile_rows_log2; + + uh->uncompressed_header_size = hdr->uncompressed_header_size; + uh->header_size_in_bytes = hdr->compressed_header_size; +} + +static void vdec_vp9_slice_setup_frame_ctx(struct vdec_vp9_slice_instance *instance, + struct vdec_vp9_slice_uncompressed_header *uh, + struct v4l2_ctrl_vp9_frame *hdr) +{ + int error_resilient_mode; + int reset_frame_context; + int key_frame; + int intra_only; + int i; + + key_frame = HDR_FLAG(KEY_FRAME); + intra_only = HDR_FLAG(INTRA_ONLY); + error_resilient_mode = HDR_FLAG(ERROR_RESILIENT); + reset_frame_context = uh->reset_frame_context; + + /* + * according to "6.2 Uncompressed header syntax" in + * "VP9 Bitstream & Decoding Process Specification", + * reset @frame_context_idx when (FrameIsIntra || error_resilient_mode) + */ + if (key_frame || intra_only || error_resilient_mode) { + /* + * @reset_frame_context specifies + * whether the frame context should be + * reset to default values: + * 0 or 1 means do not reset any frame context + * 2 resets just the context specified in the frame header + * 3 resets all contexts + */ + if (key_frame || error_resilient_mode || + reset_frame_context == 3) { + /* use default table */ + for (i = 0; i < 4; i++) + instance->dirty[i] = 0; + } else if (reset_frame_context == 2) { + instance->dirty[uh->frame_context_idx] = 0; + } + uh->frame_context_idx = 0; + } +} + +static void vdec_vp9_slice_setup_loop_filter(struct vdec_vp9_slice_uncompressed_header *uh, + struct v4l2_vp9_loop_filter *lf) +{ + int i; + + uh->loop_filter_level = lf->level; + uh->loop_filter_sharpness = lf->sharpness; + uh->loop_filter_delta_enabled = LF_FLAG(DELTA_ENABLED); + for (i = 0; i < 4; i++) + uh->loop_filter_ref_deltas[i] = lf->ref_deltas[i]; + for (i = 0; i < 2; i++) + uh->loop_filter_mode_deltas[i] = lf->mode_deltas[i]; +} + +static void vdec_vp9_slice_setup_quantization(struct vdec_vp9_slice_uncompressed_header *uh, + struct v4l2_vp9_quantization *quant) +{ + uh->base_q_idx = quant->base_q_idx; + uh->delta_q_y_dc = quant->delta_q_y_dc; + uh->delta_q_uv_dc = quant->delta_q_uv_dc; + uh->delta_q_uv_ac = quant->delta_q_uv_ac; +} + +static void vdec_vp9_slice_setup_segmentation(struct vdec_vp9_slice_uncompressed_header *uh, + struct v4l2_vp9_segmentation *seg) +{ + int i; + int j; + + uh->segmentation_enabled = SEG_FLAG(ENABLED); + uh->segmentation_update_map = SEG_FLAG(UPDATE_MAP); + for (i = 0; i < 7; i++) + uh->segmentation_tree_probs[i] = seg->tree_probs[i]; + uh->segmentation_temporal_udpate = SEG_FLAG(TEMPORAL_UPDATE); + for (i = 0; i < 3; i++) + uh->segmentation_pred_prob[i] = seg->pred_probs[i]; + uh->segmentation_update_data = SEG_FLAG(UPDATE_DATA); + uh->segmentation_abs_or_delta_update = SEG_FLAG(ABS_OR_DELTA_UPDATE); + for (i = 0; i < 8; i++) { + uh->feature_enabled[i] = seg->feature_enabled[i]; + for (j = 0; j < 4; j++) + uh->feature_value[i][j] = seg->feature_data[i][j]; + } +} + +static int vdec_vp9_slice_setup_tile(struct vdec_vp9_slice_vsi *vsi, + struct v4l2_ctrl_vp9_frame *hdr) +{ + unsigned int rows_log2; + unsigned int cols_log2; + unsigned int rows; + unsigned int cols; + unsigned int mi_rows; + unsigned int mi_cols; + struct vdec_vp9_slice_tiles *tiles; + int offset; + int start; + int end; + int i; + + rows_log2 = hdr->tile_rows_log2; + cols_log2 = hdr->tile_cols_log2; + rows = 1 << rows_log2; + cols = 1 << cols_log2; + tiles = &vsi->frame.tiles; + tiles->actual_rows = 0; + + if (rows > 4 || cols > 64) + return -EINVAL; + + /* setup mi rows/cols information */ + mi_rows = (hdr->frame_height_minus_1 + 1 + 7) >> 3; + mi_cols = (hdr->frame_width_minus_1 + 1 + 7) >> 3; + + for (i = 0; i < rows; i++) { + start = vdec_vp9_slice_tile_offset(i, mi_rows, rows_log2); + end = vdec_vp9_slice_tile_offset(i + 1, mi_rows, rows_log2); + offset = end - start; + tiles->mi_rows[i] = (offset + 7) >> 3; + if (tiles->mi_rows[i]) + tiles->actual_rows++; + } + + for (i = 0; i < cols; i++) { + start = vdec_vp9_slice_tile_offset(i, mi_cols, cols_log2); + end = vdec_vp9_slice_tile_offset(i + 1, mi_cols, cols_log2); + offset = end - start; + tiles->mi_cols[i] = (offset + 7) >> 3; + } + + return 0; +} + +static void vdec_vp9_slice_setup_state(struct vdec_vp9_slice_vsi *vsi) +{ + memset(&vsi->state, 0, sizeof(vsi->state)); +} + +static void vdec_vp9_slice_setup_ref_idx(struct vdec_vp9_slice_pfc *pfc, + struct v4l2_ctrl_vp9_frame *hdr) +{ + pfc->ref_idx[0] = hdr->last_frame_ts; + pfc->ref_idx[1] = hdr->golden_frame_ts; + pfc->ref_idx[2] = hdr->alt_frame_ts; +} + +static int vdec_vp9_slice_setup_pfc(struct vdec_vp9_slice_instance *instance, + struct vdec_vp9_slice_pfc *pfc) +{ + struct v4l2_ctrl_vp9_frame *hdr; + struct vdec_vp9_slice_uncompressed_header *uh; + struct v4l2_ctrl *hdr_ctrl; + struct vdec_vp9_slice_vsi *vsi; + int ret; + + /* frame header */ + hdr_ctrl = v4l2_ctrl_find(&instance->ctx->ctrl_hdl, V4L2_CID_STATELESS_VP9_FRAME); + if (!hdr_ctrl || !hdr_ctrl->p_cur.p) + return -EINVAL; + + hdr = hdr_ctrl->p_cur.p; + vsi = &pfc->vsi; + uh = &vsi->frame.uh; + + /* setup vsi information */ + vdec_vp9_slice_setup_hdr(instance, uh, hdr); + vdec_vp9_slice_setup_frame_ctx(instance, uh, hdr); + vdec_vp9_slice_setup_loop_filter(uh, &hdr->lf); + vdec_vp9_slice_setup_quantization(uh, &hdr->quant); + vdec_vp9_slice_setup_segmentation(uh, &hdr->seg); + ret = vdec_vp9_slice_setup_tile(vsi, hdr); + if (ret) + return ret; + vdec_vp9_slice_setup_state(vsi); + + /* core stage needs buffer index to get ref y/c ... */ + vdec_vp9_slice_setup_ref_idx(pfc, hdr); + + pfc->seq = instance->seq; + instance->seq++; + + return 0; +} + +static int vdec_vp9_slice_setup_lat_buffer(struct vdec_vp9_slice_instance *instance, + struct vdec_vp9_slice_vsi *vsi, + struct mtk_vcodec_mem *bs, + struct vdec_lat_buf *lat_buf) +{ + int i; + + vsi->bs.buf.dma_addr = bs->dma_addr; + vsi->bs.buf.size = bs->size; + vsi->bs.frame.dma_addr = bs->dma_addr; + vsi->bs.frame.size = bs->size; + + for (i = 0; i < 2; i++) { + vsi->mv[i].dma_addr = instance->mv[i].dma_addr; + vsi->mv[i].size = instance->mv[i].size; + } + for (i = 0; i < 2; i++) { + vsi->seg[i].dma_addr = instance->seg[i].dma_addr; + vsi->seg[i].size = instance->seg[i].size; + } + vsi->tile.dma_addr = instance->tile.dma_addr; + vsi->tile.size = instance->tile.size; + vsi->prob.dma_addr = instance->prob.dma_addr; + vsi->prob.size = instance->prob.size; + vsi->counts.dma_addr = instance->counts.dma_addr; + vsi->counts.size = instance->counts.size; + + vsi->ube.dma_addr = lat_buf->ctx->msg_queue.wdma_addr.dma_addr; + vsi->ube.size = lat_buf->ctx->msg_queue.wdma_addr.size; + vsi->trans.dma_addr = lat_buf->ctx->msg_queue.wdma_wptr_addr; + /* used to store trans end */ + vsi->trans.dma_addr_end = lat_buf->ctx->msg_queue.wdma_rptr_addr; + vsi->err_map.dma_addr = lat_buf->wdma_err_addr.dma_addr; + vsi->err_map.size = lat_buf->wdma_err_addr.size; + + vsi->row_info.buf = 0; + vsi->row_info.size = 0; + + return 0; +} + +static int vdec_vp9_slice_setup_prob_buffer(struct vdec_vp9_slice_instance *instance, + struct vdec_vp9_slice_vsi *vsi) +{ + struct vdec_vp9_slice_frame_ctx *frame_ctx; + struct vdec_vp9_slice_uncompressed_header *uh; + + uh = &vsi->frame.uh; + + mtk_vdec_debug(instance->ctx, "ctx dirty %u idx %d\n", + instance->dirty[uh->frame_context_idx], + uh->frame_context_idx); + + if (instance->dirty[uh->frame_context_idx]) + frame_ctx = &instance->frame_ctx[uh->frame_context_idx]; + else + frame_ctx = vdec_vp9_slice_default_frame_ctx; + memcpy(instance->prob.va, frame_ctx, sizeof(*frame_ctx)); + + return 0; +} + +static void vdec_vp9_slice_setup_seg_buffer(struct vdec_vp9_slice_instance *instance, + struct vdec_vp9_slice_vsi *vsi, + struct mtk_vcodec_mem *buf) +{ + struct vdec_vp9_slice_uncompressed_header *uh; + + /* reset segment buffer */ + uh = &vsi->frame.uh; + if (uh->frame_type == 0 || + uh->intra_only || + uh->error_resilient_mode || + uh->frame_width != instance->width || + uh->frame_height != instance->height) { + mtk_vdec_debug(instance->ctx, "reset seg\n"); + memset(buf->va, 0, buf->size); + } +} + +/* + * parse tiles according to `6.4 Decode tiles syntax` + * in "vp9-bitstream-specification" + * + * frame contains uncompress header, compressed header and several tiles. + * this function parses tiles' position and size, stores them to tile buffer + * for decoding. + */ +static int vdec_vp9_slice_setup_tile_buffer(struct vdec_vp9_slice_instance *instance, + struct vdec_vp9_slice_vsi *vsi, + struct mtk_vcodec_mem *bs) +{ + struct vdec_vp9_slice_uncompressed_header *uh; + unsigned int rows_log2; + unsigned int cols_log2; + unsigned int rows; + unsigned int cols; + unsigned int mi_row; + unsigned int mi_col; + unsigned int offset; + unsigned int pa; + unsigned int size; + struct vdec_vp9_slice_tiles *tiles; + unsigned char *pos; + unsigned char *end; + unsigned char *va; + unsigned int *tb; + int i; + int j; + + uh = &vsi->frame.uh; + rows_log2 = uh->tile_rows_log2; + cols_log2 = uh->tile_cols_log2; + rows = 1 << rows_log2; + cols = 1 << cols_log2; + + if (rows > 4 || cols > 64) { + mtk_vdec_err(instance->ctx, "tile_rows %u tile_cols %u\n", rows, cols); + return -EINVAL; + } + + offset = uh->uncompressed_header_size + + uh->header_size_in_bytes; + if (bs->size <= offset) { + mtk_vdec_err(instance->ctx, "bs size %zu tile offset %u\n", bs->size, offset); + return -EINVAL; + } + + tiles = &vsi->frame.tiles; + /* setup tile buffer */ + + va = (unsigned char *)bs->va; + pos = va + offset; + end = va + bs->size; + /* truncated */ + pa = (unsigned int)bs->dma_addr + offset; + tb = instance->tile.va; + for (i = 0; i < rows; i++) { + for (j = 0; j < cols; j++) { + if (i == rows - 1 && + j == cols - 1) { + size = (unsigned int)(end - pos); + } else { + if (end - pos < 4) + return -EINVAL; + + size = (pos[0] << 24) | (pos[1] << 16) | + (pos[2] << 8) | pos[3]; + pos += 4; + pa += 4; + offset += 4; + if (end - pos < size) + return -EINVAL; + } + tiles->size[i][j] = size; + if (tiles->mi_rows[i]) { + *tb++ = (size << 3) + ((offset << 3) & 0x7f); + *tb++ = pa & ~0xf; + *tb++ = (pa << 3) & 0x7f; + mi_row = (tiles->mi_rows[i] - 1) & 0x1ff; + mi_col = (tiles->mi_cols[j] - 1) & 0x3f; + *tb++ = (mi_row << 6) + mi_col; + } + pos += size; + pa += size; + offset += size; + } + } + + return 0; +} + +static int vdec_vp9_slice_setup_lat(struct vdec_vp9_slice_instance *instance, + struct mtk_vcodec_mem *bs, + struct vdec_lat_buf *lat_buf, + struct vdec_vp9_slice_pfc *pfc) +{ + struct vdec_vp9_slice_vsi *vsi = &pfc->vsi; + int ret; + + ret = vdec_vp9_slice_setup_lat_from_src_buf(instance, lat_buf); + if (ret) + goto err; + + ret = vdec_vp9_slice_setup_pfc(instance, pfc); + if (ret) + goto err; + + ret = vdec_vp9_slice_alloc_working_buffer(instance, vsi); + if (ret) + goto err; + + ret = vdec_vp9_slice_setup_lat_buffer(instance, vsi, bs, lat_buf); + if (ret) + goto err; + + vdec_vp9_slice_setup_seg_buffer(instance, vsi, &instance->seg[0]); + + /* setup prob/tile buffers for LAT */ + + ret = vdec_vp9_slice_setup_prob_buffer(instance, vsi); + if (ret) + goto err; + + ret = vdec_vp9_slice_setup_tile_buffer(instance, vsi, bs); + if (ret) + goto err; + + return 0; + +err: + return ret; +} + +static +void vdec_vp9_slice_map_counts_eob_coef(unsigned int i, unsigned int j, unsigned int k, + struct vdec_vp9_slice_frame_counts *counts, + struct v4l2_vp9_frame_symbol_counts *counts_helper) +{ + u32 l = 0, m; + + /* + * helper eo -> mtk eo + * helpre e1 -> mtk c3 + * helper c0 -> c0 + * helper c1 -> c1 + * helper c2 -> c2 + */ + for (m = 0; m < 3; m++) { + counts_helper->coeff[i][j][k][l][m] = + (u32 (*)[3]) & counts->coef_probs[i][j][k].band_0[m]; + counts_helper->eob[i][j][k][l][m][0] = + &counts->eob_branch[i][j][k].band_0[m]; + counts_helper->eob[i][j][k][l][m][1] = + &counts->coef_probs[i][j][k].band_0[m][3]; + } + + for (l = 1; l < 6; l++) { + for (m = 0; m < 6; m++) { + counts_helper->coeff[i][j][k][l][m] = + (u32 (*)[3]) & counts->coef_probs[i][j][k].band_1_5[l - 1][m]; + counts_helper->eob[i][j][k][l][m][0] = + &counts->eob_branch[i][j][k].band_1_5[l - 1][m]; + counts_helper->eob[i][j][k][l][m][1] = + &counts->coef_probs[i][j][k].band_1_5[l - 1][m][3]; + } + } +} + +static void vdec_vp9_slice_counts_map_helper(struct vdec_vp9_slice_counts_map *counts_map, + struct vdec_vp9_slice_frame_counts *counts, + struct v4l2_vp9_frame_symbol_counts *counts_helper) +{ + int i, j, k; + + counts_helper->partition = &counts->partition; + counts_helper->intra_inter = &counts->intra_inter; + counts_helper->tx32p = &counts->tx_p32x32; + counts_helper->tx16p = &counts->tx_p16x16; + counts_helper->tx8p = &counts->tx_p8x8; + counts_helper->uv_mode = &counts->uv_mode; + + counts_helper->comp = &counts->comp_inter; + counts_helper->comp_ref = &counts->comp_ref; + counts_helper->single_ref = &counts->single_ref; + counts_helper->mv_mode = &counts->inter_mode; + counts_helper->mv_joint = &counts->joint; + + for (i = 0; i < ARRAY_SIZE(counts_map->skip); i++) + memcpy(counts_map->skip[i], counts->skip[i], + sizeof(counts_map->skip[0])); + counts_helper->skip = &counts_map->skip; + + for (i = 0; i < ARRAY_SIZE(counts_map->y_mode); i++) + memcpy(counts_map->y_mode[i], counts->y_mode[i], + sizeof(counts_map->y_mode[0])); + counts_helper->y_mode = &counts_map->y_mode; + + for (i = 0; i < ARRAY_SIZE(counts_map->filter); i++) + memcpy(counts_map->filter[i], counts->switchable_interp[i], + sizeof(counts_map->filter[0])); + counts_helper->filter = &counts_map->filter; + + for (i = 0; i < ARRAY_SIZE(counts_map->sign); i++) + memcpy(counts_map->sign[i], counts->mvcomp[i].sign, + sizeof(counts_map->sign[0])); + counts_helper->sign = &counts_map->sign; + + for (i = 0; i < ARRAY_SIZE(counts_map->classes); i++) + memcpy(counts_map->classes[i], counts->mvcomp[i].classes, + sizeof(counts_map->classes[0])); + counts_helper->classes = &counts_map->classes; + + for (i = 0; i < ARRAY_SIZE(counts_map->class0); i++) + memcpy(counts_map->class0[i], counts->mvcomp[i].class0, + sizeof(counts_map->class0[0])); + counts_helper->class0 = &counts_map->class0; + + for (i = 0; i < ARRAY_SIZE(counts_map->bits); i++) + for (j = 0; j < ARRAY_SIZE(counts_map->bits[0]); j++) + memcpy(counts_map->bits[i][j], counts->mvcomp[i].bits[j], + sizeof(counts_map->bits[0][0])); + counts_helper->bits = &counts_map->bits; + + for (i = 0; i < ARRAY_SIZE(counts_map->class0_fp); i++) + for (j = 0; j < ARRAY_SIZE(counts_map->class0_fp[0]); j++) + memcpy(counts_map->class0_fp[i][j], counts->mvcomp[i].class0_fp[j], + sizeof(counts_map->class0_fp[0][0])); + counts_helper->class0_fp = &counts_map->class0_fp; + + for (i = 0; i < ARRAY_SIZE(counts_map->fp); i++) + memcpy(counts_map->fp[i], counts->mvcomp[i].fp, + sizeof(counts_map->fp[0])); + counts_helper->fp = &counts_map->fp; + + for (i = 0; i < ARRAY_SIZE(counts_map->class0_hp); i++) + memcpy(counts_map->class0_hp[i], counts->mvcomp[i].class0_hp, + sizeof(counts_map->class0_hp[0])); + counts_helper->class0_hp = &counts_map->class0_hp; + + for (i = 0; i < ARRAY_SIZE(counts_map->hp); i++) + memcpy(counts_map->hp[i], counts->mvcomp[i].hp, sizeof(counts_map->hp[0])); + + counts_helper->hp = &counts_map->hp; + + for (i = 0; i < 4; i++) + for (j = 0; j < 2; j++) + for (k = 0; k < 2; k++) + vdec_vp9_slice_map_counts_eob_coef(i, j, k, counts, counts_helper); +} + +static void vdec_vp9_slice_map_to_coef(unsigned int i, unsigned int j, unsigned int k, + struct vdec_vp9_slice_frame_ctx *frame_ctx, + struct v4l2_vp9_frame_context *frame_ctx_helper) +{ + u32 l, m; + + for (l = 0; l < ARRAY_SIZE(frame_ctx_helper->coef[0][0][0]); l++) { + for (m = 0; m < VP9_BAND_6(l); m++) { + memcpy(frame_ctx_helper->coef[i][j][k][l][m], + frame_ctx->coef_probs[i][j][k][l].probs[m], + sizeof(frame_ctx_helper->coef[i][j][k][l][0])); + } + } +} + +static void vdec_vp9_slice_map_from_coef(unsigned int i, unsigned int j, unsigned int k, + struct vdec_vp9_slice_frame_ctx *frame_ctx, + struct v4l2_vp9_frame_context *frame_ctx_helper) +{ + u32 l, m; + + for (l = 0; l < ARRAY_SIZE(frame_ctx_helper->coef[0][0][0]); l++) { + for (m = 0; m < VP9_BAND_6(l); m++) { + memcpy(frame_ctx->coef_probs[i][j][k][l].probs[m], + frame_ctx_helper->coef[i][j][k][l][m], + sizeof(frame_ctx_helper->coef[i][j][k][l][0])); + } + } +} + +static +void vdec_vp9_slice_framectx_map_helper(bool frame_is_intra, + struct vdec_vp9_slice_frame_ctx *pre_frame_ctx, + struct vdec_vp9_slice_frame_ctx *frame_ctx, + struct v4l2_vp9_frame_context *frame_ctx_helper) +{ + struct v4l2_vp9_frame_mv_context *mv = &frame_ctx_helper->mv; + u32 i, j, k; + + for (i = 0; i < ARRAY_SIZE(frame_ctx_helper->coef); i++) + for (j = 0; j < ARRAY_SIZE(frame_ctx_helper->coef[0]); j++) + for (k = 0; k < ARRAY_SIZE(frame_ctx_helper->coef[0][0]); k++) + vdec_vp9_slice_map_to_coef(i, j, k, pre_frame_ctx, + frame_ctx_helper); + + /* + * use previous prob when frame is not intra or + * we should use the prob updated by the compressed header parse + */ + if (!frame_is_intra) + frame_ctx = pre_frame_ctx; + + for (i = 0; i < ARRAY_SIZE(frame_ctx_helper->tx8); i++) + memcpy(frame_ctx_helper->tx8[i], frame_ctx->tx_p8x8[i], + sizeof(frame_ctx_helper->tx8[0])); + + for (i = 0; i < ARRAY_SIZE(frame_ctx_helper->tx16); i++) + memcpy(frame_ctx_helper->tx16[i], frame_ctx->tx_p16x16[i], + sizeof(frame_ctx_helper->tx16[0])); + + for (i = 0; i < ARRAY_SIZE(frame_ctx_helper->tx32); i++) + memcpy(frame_ctx_helper->tx32[i], frame_ctx->tx_p32x32[i], + sizeof(frame_ctx_helper->tx32[0])); + + memcpy(frame_ctx_helper->skip, frame_ctx->skip_probs, sizeof(frame_ctx_helper->skip)); + + for (i = 0; i < ARRAY_SIZE(frame_ctx_helper->inter_mode); i++) + memcpy(frame_ctx_helper->inter_mode[i], frame_ctx->inter_mode_probs[i], + sizeof(frame_ctx_helper->inter_mode[0])); + + for (i = 0; i < ARRAY_SIZE(frame_ctx_helper->interp_filter); i++) + memcpy(frame_ctx_helper->interp_filter[i], frame_ctx->switch_interp_prob[i], + sizeof(frame_ctx_helper->interp_filter[0])); + + memcpy(frame_ctx_helper->is_inter, frame_ctx->intra_inter_prob, + sizeof(frame_ctx_helper->is_inter)); + + memcpy(frame_ctx_helper->comp_mode, frame_ctx->comp_inter_prob, + sizeof(frame_ctx_helper->comp_mode)); + + for (i = 0; i < ARRAY_SIZE(frame_ctx_helper->single_ref); i++) + memcpy(frame_ctx_helper->single_ref[i], frame_ctx->single_ref_prob[i], + sizeof(frame_ctx_helper->single_ref[0])); + + memcpy(frame_ctx_helper->comp_ref, frame_ctx->comp_ref_prob, + sizeof(frame_ctx_helper->comp_ref)); + + for (i = 0; i < ARRAY_SIZE(frame_ctx_helper->y_mode); i++) + memcpy(frame_ctx_helper->y_mode[i], frame_ctx->y_mode_prob[i], + sizeof(frame_ctx_helper->y_mode[0])); + + for (i = 0; i < ARRAY_SIZE(frame_ctx_helper->uv_mode); i++) + memcpy(frame_ctx_helper->uv_mode[i], frame_ctx->uv_mode_prob[i], + sizeof(frame_ctx_helper->uv_mode[0])); + + for (i = 0; i < ARRAY_SIZE(frame_ctx_helper->partition); i++) + memcpy(frame_ctx_helper->partition[i], frame_ctx->partition_prob[i], + sizeof(frame_ctx_helper->partition[0])); + + memcpy(mv->joint, frame_ctx->joint, sizeof(mv->joint)); + + for (i = 0; i < ARRAY_SIZE(mv->sign); i++) + mv->sign[i] = frame_ctx->sign_classes[i].sign; + + for (i = 0; i < ARRAY_SIZE(mv->classes); i++) + memcpy(mv->classes[i], frame_ctx->sign_classes[i].classes, + sizeof(mv->classes[i])); + + for (i = 0; i < ARRAY_SIZE(mv->class0_bit); i++) + mv->class0_bit[i] = frame_ctx->class0_bits[i].class0[0]; + + for (i = 0; i < ARRAY_SIZE(mv->bits); i++) + memcpy(mv->bits[i], frame_ctx->class0_bits[i].bits, sizeof(mv->bits[0])); + + for (i = 0; i < ARRAY_SIZE(mv->class0_fr); i++) + for (j = 0; j < ARRAY_SIZE(mv->class0_fr[0]); j++) + memcpy(mv->class0_fr[i][j], frame_ctx->class0_fp_hp[i].class0_fp[j], + sizeof(mv->class0_fr[0][0])); + + for (i = 0; i < ARRAY_SIZE(mv->fr); i++) + memcpy(mv->fr[i], frame_ctx->class0_fp_hp[i].fp, sizeof(mv->fr[0])); + + for (i = 0; i < ARRAY_SIZE(mv->class0_hp); i++) + mv->class0_hp[i] = frame_ctx->class0_fp_hp[i].class0_hp; + + for (i = 0; i < ARRAY_SIZE(mv->hp); i++) + mv->hp[i] = frame_ctx->class0_fp_hp[i].hp; +} + +static void vdec_vp9_slice_helper_map_framectx(struct v4l2_vp9_frame_context *frame_ctx_helper, + struct vdec_vp9_slice_frame_ctx *frame_ctx) +{ + struct v4l2_vp9_frame_mv_context *mv = &frame_ctx_helper->mv; + u32 i, j, k; + + for (i = 0; i < ARRAY_SIZE(frame_ctx_helper->tx8); i++) + memcpy(frame_ctx->tx_p8x8[i], frame_ctx_helper->tx8[i], + sizeof(frame_ctx_helper->tx8[0])); + + for (i = 0; i < ARRAY_SIZE(frame_ctx_helper->tx16); i++) + memcpy(frame_ctx->tx_p16x16[i], frame_ctx_helper->tx16[i], + sizeof(frame_ctx_helper->tx16[0])); + + for (i = 0; i < ARRAY_SIZE(frame_ctx_helper->tx32); i++) + memcpy(frame_ctx->tx_p32x32[i], frame_ctx_helper->tx32[i], + sizeof(frame_ctx_helper->tx32[0])); + + for (i = 0; i < ARRAY_SIZE(frame_ctx_helper->coef); i++) + for (j = 0; j < ARRAY_SIZE(frame_ctx_helper->coef[0]); j++) + for (k = 0; k < ARRAY_SIZE(frame_ctx_helper->coef[0][0]); k++) + vdec_vp9_slice_map_from_coef(i, j, k, frame_ctx, + frame_ctx_helper); + + memcpy(frame_ctx->skip_probs, frame_ctx_helper->skip, sizeof(frame_ctx_helper->skip)); + + for (i = 0; i < ARRAY_SIZE(frame_ctx_helper->inter_mode); i++) + memcpy(frame_ctx->inter_mode_probs[i], frame_ctx_helper->inter_mode[i], + sizeof(frame_ctx_helper->inter_mode[0])); + + for (i = 0; i < ARRAY_SIZE(frame_ctx_helper->interp_filter); i++) + memcpy(frame_ctx->switch_interp_prob[i], frame_ctx_helper->interp_filter[i], + sizeof(frame_ctx_helper->interp_filter[0])); + + memcpy(frame_ctx->intra_inter_prob, frame_ctx_helper->is_inter, + sizeof(frame_ctx_helper->is_inter)); + + memcpy(frame_ctx->comp_inter_prob, frame_ctx_helper->comp_mode, + sizeof(frame_ctx_helper->comp_mode)); + + for (i = 0; i < ARRAY_SIZE(frame_ctx_helper->single_ref); i++) + memcpy(frame_ctx->single_ref_prob[i], frame_ctx_helper->single_ref[i], + sizeof(frame_ctx_helper->single_ref[0])); + + memcpy(frame_ctx->comp_ref_prob, frame_ctx_helper->comp_ref, + sizeof(frame_ctx_helper->comp_ref)); + + for (i = 0; i < ARRAY_SIZE(frame_ctx_helper->y_mode); i++) + memcpy(frame_ctx->y_mode_prob[i], frame_ctx_helper->y_mode[i], + sizeof(frame_ctx_helper->y_mode[0])); + + for (i = 0; i < ARRAY_SIZE(frame_ctx_helper->uv_mode); i++) + memcpy(frame_ctx->uv_mode_prob[i], frame_ctx_helper->uv_mode[i], + sizeof(frame_ctx_helper->uv_mode[0])); + + for (i = 0; i < ARRAY_SIZE(frame_ctx_helper->partition); i++) + memcpy(frame_ctx->partition_prob[i], frame_ctx_helper->partition[i], + sizeof(frame_ctx_helper->partition[0])); + + memcpy(frame_ctx->joint, mv->joint, sizeof(mv->joint)); + + for (i = 0; i < ARRAY_SIZE(mv->sign); i++) + frame_ctx->sign_classes[i].sign = mv->sign[i]; + + for (i = 0; i < ARRAY_SIZE(mv->classes); i++) + memcpy(frame_ctx->sign_classes[i].classes, mv->classes[i], + sizeof(mv->classes[i])); + + for (i = 0; i < ARRAY_SIZE(mv->class0_bit); i++) + frame_ctx->class0_bits[i].class0[0] = mv->class0_bit[i]; + + for (i = 0; i < ARRAY_SIZE(mv->bits); i++) + memcpy(frame_ctx->class0_bits[i].bits, mv->bits[i], sizeof(mv->bits[0])); + + for (i = 0; i < ARRAY_SIZE(mv->class0_fr); i++) + for (j = 0; j < ARRAY_SIZE(mv->class0_fr[0]); j++) + memcpy(frame_ctx->class0_fp_hp[i].class0_fp[j], mv->class0_fr[i][j], + sizeof(mv->class0_fr[0][0])); + + for (i = 0; i < ARRAY_SIZE(mv->fr); i++) + memcpy(frame_ctx->class0_fp_hp[i].fp, mv->fr[i], sizeof(mv->fr[0])); + + for (i = 0; i < ARRAY_SIZE(mv->class0_hp); i++) + frame_ctx->class0_fp_hp[i].class0_hp = mv->class0_hp[i]; + + for (i = 0; i < ARRAY_SIZE(mv->hp); i++) + frame_ctx->class0_fp_hp[i].hp = mv->hp[i]; +} + +static int vdec_vp9_slice_update_prob(struct vdec_vp9_slice_instance *instance, + struct vdec_vp9_slice_vsi *vsi) +{ + struct vdec_vp9_slice_frame_ctx *pre_frame_ctx; + struct v4l2_vp9_frame_context *pre_frame_ctx_helper; + struct vdec_vp9_slice_frame_ctx *frame_ctx; + struct vdec_vp9_slice_frame_counts *counts; + struct v4l2_vp9_frame_symbol_counts *counts_helper; + struct vdec_vp9_slice_uncompressed_header *uh; + bool frame_is_intra; + bool use_128; + + uh = &vsi->frame.uh; + pre_frame_ctx = &instance->frame_ctx[uh->frame_context_idx]; + pre_frame_ctx_helper = &instance->frame_ctx_helper; + frame_ctx = (struct vdec_vp9_slice_frame_ctx *)instance->prob.va; + counts = (struct vdec_vp9_slice_frame_counts *)instance->counts.va; + counts_helper = &instance->counts_helper; + + if (!uh->refresh_frame_context) + return 0; + + if (!uh->frame_parallel_decoding_mode) { + vdec_vp9_slice_counts_map_helper(&instance->counts_map, counts, counts_helper); + + frame_is_intra = !vsi->frame.uh.frame_type || vsi->frame.uh.intra_only; + /* check default prob */ + if (!instance->dirty[uh->frame_context_idx]) + vdec_vp9_slice_framectx_map_helper(frame_is_intra, + vdec_vp9_slice_default_frame_ctx, + frame_ctx, + pre_frame_ctx_helper); + else + vdec_vp9_slice_framectx_map_helper(frame_is_intra, + pre_frame_ctx, + frame_ctx, + pre_frame_ctx_helper); + + use_128 = !frame_is_intra && !vsi->frame.uh.last_frame_type; + v4l2_vp9_adapt_coef_probs(pre_frame_ctx_helper, + counts_helper, + use_128, + frame_is_intra); + if (!frame_is_intra) + v4l2_vp9_adapt_noncoef_probs(pre_frame_ctx_helper, + counts_helper, + V4L2_VP9_REFERENCE_MODE_SINGLE_REFERENCE, + vsi->frame.uh.interpolation_filter, + vsi->frame.ch.tx_mode, + vsi->frame.uh.allow_high_precision_mv ? + V4L2_VP9_FRAME_FLAG_ALLOW_HIGH_PREC_MV : 0); + vdec_vp9_slice_helper_map_framectx(pre_frame_ctx_helper, pre_frame_ctx); + } else { + memcpy(pre_frame_ctx, frame_ctx, sizeof(*frame_ctx)); + } + + instance->dirty[uh->frame_context_idx] = 1; + + return 0; +} + +static int vdec_vp9_slice_update_single(struct vdec_vp9_slice_instance *instance, + struct vdec_vp9_slice_pfc *pfc) +{ + struct vdec_vp9_slice_vsi *vsi; + + vsi = &pfc->vsi; + memcpy(&pfc->state[0], &vsi->state, sizeof(vsi->state)); + + mtk_vdec_debug(instance->ctx, "Frame %u Y_CRC %08x %08x %08x %08x\n", + pfc->seq, vsi->state.crc[0], vsi->state.crc[1], + vsi->state.crc[2], vsi->state.crc[3]); + mtk_vdec_debug(instance->ctx, "Frame %u C_CRC %08x %08x %08x %08x\n", + pfc->seq, vsi->state.crc[4], vsi->state.crc[5], + vsi->state.crc[6], vsi->state.crc[7]); + + vdec_vp9_slice_update_prob(instance, vsi); + + instance->width = vsi->frame.uh.frame_width; + instance->height = vsi->frame.uh.frame_height; + instance->frame_type = vsi->frame.uh.frame_type; + instance->show_frame = vsi->frame.uh.show_frame; + + return 0; +} + +static int vdec_vp9_slice_update_lat(struct vdec_vp9_slice_instance *instance, + struct vdec_lat_buf *lat_buf, + struct vdec_vp9_slice_pfc *pfc) +{ + struct vdec_vp9_slice_vsi *vsi; + + vsi = &pfc->vsi; + memcpy(&pfc->state[0], &vsi->state, sizeof(vsi->state)); + + mtk_vdec_debug(instance->ctx, "Frame %u LAT CRC 0x%08x %lx %lx\n", + pfc->seq, vsi->state.crc[0], + (unsigned long)vsi->trans.dma_addr, + (unsigned long)vsi->trans.dma_addr_end); + + /* buffer full, need to re-decode */ + if (vsi->state.full) { + /* buffer not enough */ + if (vsi->trans.dma_addr_end - vsi->trans.dma_addr == + vsi->ube.size) + return -ENOMEM; + return -EAGAIN; + } + + vdec_vp9_slice_update_prob(instance, vsi); + + instance->width = vsi->frame.uh.frame_width; + instance->height = vsi->frame.uh.frame_height; + instance->frame_type = vsi->frame.uh.frame_type; + instance->show_frame = vsi->frame.uh.show_frame; + + return 0; +} + +static int vdec_vp9_slice_setup_core_to_dst_buf(struct vdec_vp9_slice_instance *instance, + struct vdec_lat_buf *lat_buf) +{ + struct vb2_v4l2_buffer *dst; + + dst = v4l2_m2m_next_dst_buf(instance->ctx->m2m_ctx); + if (!dst) + return -EINVAL; + + v4l2_m2m_buf_copy_metadata(&lat_buf->ts_info, dst, true); + return 0; +} + +static int vdec_vp9_slice_setup_core_buffer(struct vdec_vp9_slice_instance *instance, + struct vdec_vp9_slice_pfc *pfc, + struct vdec_vp9_slice_vsi *vsi, + struct vdec_fb *fb, + struct vdec_lat_buf *lat_buf) +{ + struct vb2_buffer *vb; + struct vb2_queue *vq; + struct vdec_vp9_slice_reference *ref; + int plane; + int size; + int w; + int h; + int i; + + plane = instance->ctx->q_data[MTK_Q_DATA_DST].fmt->num_planes; + w = vsi->frame.uh.frame_width; + h = vsi->frame.uh.frame_height; + size = ALIGN(w, 64) * ALIGN(h, 64); + + /* frame buffer */ + vsi->fb.y.dma_addr = fb->base_y.dma_addr; + if (plane == 1) + vsi->fb.c.dma_addr = fb->base_y.dma_addr + size; + else + vsi->fb.c.dma_addr = fb->base_c.dma_addr; + + /* reference buffers */ + vq = v4l2_m2m_get_vq(instance->ctx->m2m_ctx, + V4L2_BUF_TYPE_VIDEO_CAPTURE_MPLANE); + if (!vq) + return -EINVAL; + + /* get current output buffer */ + vb = &v4l2_m2m_next_dst_buf(instance->ctx->m2m_ctx)->vb2_buf; + if (!vb) + return -EINVAL; + + /* update internal buffer's width/height */ + for (i = 0; i < vq->num_buffers; i++) { + if (vb == vq->bufs[i]) { + instance->dpb[i].width = w; + instance->dpb[i].height = h; + break; + } + } + + /* + * get buffer's width/height from instance + * get buffer address from vb2buf + */ + for (i = 0; i < 3; i++) { + ref = &vsi->frame.ref[i]; + vb = vb2_find_buffer(vq, pfc->ref_idx[i]); + if (!vb) { + ref->frame_width = w; + ref->frame_height = h; + memset(&vsi->ref[i], 0, sizeof(vsi->ref[i])); + } else { + int idx = vb->index; + + ref->frame_width = instance->dpb[idx].width; + ref->frame_height = instance->dpb[idx].height; + vsi->ref[i].y.dma_addr = + vb2_dma_contig_plane_dma_addr(vb, 0); + if (plane == 1) + vsi->ref[i].c.dma_addr = + vsi->ref[i].y.dma_addr + size; + else + vsi->ref[i].c.dma_addr = + vb2_dma_contig_plane_dma_addr(vb, 1); + } + } + + return 0; +} + +static void vdec_vp9_slice_setup_single_buffer(struct vdec_vp9_slice_instance *instance, + struct vdec_vp9_slice_pfc *pfc, + struct vdec_vp9_slice_vsi *vsi, + struct mtk_vcodec_mem *bs, + struct vdec_fb *fb) +{ + int i; + + vsi->bs.buf.dma_addr = bs->dma_addr; + vsi->bs.buf.size = bs->size; + vsi->bs.frame.dma_addr = bs->dma_addr; + vsi->bs.frame.size = bs->size; + + for (i = 0; i < 2; i++) { + vsi->mv[i].dma_addr = instance->mv[i].dma_addr; + vsi->mv[i].size = instance->mv[i].size; + } + for (i = 0; i < 2; i++) { + vsi->seg[i].dma_addr = instance->seg[i].dma_addr; + vsi->seg[i].size = instance->seg[i].size; + } + vsi->tile.dma_addr = instance->tile.dma_addr; + vsi->tile.size = instance->tile.size; + vsi->prob.dma_addr = instance->prob.dma_addr; + vsi->prob.size = instance->prob.size; + vsi->counts.dma_addr = instance->counts.dma_addr; + vsi->counts.size = instance->counts.size; + + vsi->row_info.buf = 0; + vsi->row_info.size = 0; + + vdec_vp9_slice_setup_core_buffer(instance, pfc, vsi, fb, NULL); +} + +static int vdec_vp9_slice_setup_core(struct vdec_vp9_slice_instance *instance, + struct vdec_fb *fb, + struct vdec_lat_buf *lat_buf, + struct vdec_vp9_slice_pfc *pfc) +{ + struct vdec_vp9_slice_vsi *vsi = &pfc->vsi; + int ret; + + vdec_vp9_slice_setup_state(vsi); + + ret = vdec_vp9_slice_setup_core_to_dst_buf(instance, lat_buf); + if (ret) + goto err; + + ret = vdec_vp9_slice_setup_core_buffer(instance, pfc, vsi, fb, lat_buf); + if (ret) + goto err; + + vdec_vp9_slice_setup_seg_buffer(instance, vsi, &instance->seg[1]); + + return 0; + +err: + return ret; +} + +static int vdec_vp9_slice_setup_single(struct vdec_vp9_slice_instance *instance, + struct mtk_vcodec_mem *bs, + struct vdec_fb *fb, + struct vdec_vp9_slice_pfc *pfc) +{ + struct vdec_vp9_slice_vsi *vsi = &pfc->vsi; + int ret; + + ret = vdec_vp9_slice_setup_single_from_src_to_dst(instance); + if (ret) + goto err; + + ret = vdec_vp9_slice_setup_pfc(instance, pfc); + if (ret) + goto err; + + ret = vdec_vp9_slice_alloc_working_buffer(instance, vsi); + if (ret) + goto err; + + vdec_vp9_slice_setup_single_buffer(instance, pfc, vsi, bs, fb); + vdec_vp9_slice_setup_seg_buffer(instance, vsi, &instance->seg[0]); + + ret = vdec_vp9_slice_setup_prob_buffer(instance, vsi); + if (ret) + goto err; + + ret = vdec_vp9_slice_setup_tile_buffer(instance, vsi, bs); + if (ret) + goto err; + + return 0; + +err: + return ret; +} + +static int vdec_vp9_slice_update_core(struct vdec_vp9_slice_instance *instance, + struct vdec_lat_buf *lat_buf, + struct vdec_vp9_slice_pfc *pfc) +{ + struct vdec_vp9_slice_vsi *vsi; + + vsi = &pfc->vsi; + memcpy(&pfc->state[1], &vsi->state, sizeof(vsi->state)); + + mtk_vdec_debug(instance->ctx, "Frame %u Y_CRC %08x %08x %08x %08x\n", + pfc->seq, vsi->state.crc[0], vsi->state.crc[1], + vsi->state.crc[2], vsi->state.crc[3]); + mtk_vdec_debug(instance->ctx, "Frame %u C_CRC %08x %08x %08x %08x\n", + pfc->seq, vsi->state.crc[4], vsi->state.crc[5], + vsi->state.crc[6], vsi->state.crc[7]); + + return 0; +} + +static int vdec_vp9_slice_init(struct mtk_vcodec_dec_ctx *ctx) +{ + struct vdec_vp9_slice_instance *instance; + struct vdec_vp9_slice_init_vsi *vsi; + int ret; + + instance = kzalloc(sizeof(*instance), GFP_KERNEL); + if (!instance) + return -ENOMEM; + + instance->ctx = ctx; + instance->vpu.id = SCP_IPI_VDEC_LAT; + instance->vpu.core_id = SCP_IPI_VDEC_CORE; + instance->vpu.ctx = ctx; + instance->vpu.codec_type = ctx->current_codec; + + ret = vpu_dec_init(&instance->vpu); + if (ret) { + mtk_vdec_err(ctx, "failed to init vpu dec, ret %d\n", ret); + goto error_vpu_init; + } + + /* init vsi and global flags */ + + vsi = instance->vpu.vsi; + if (!vsi) { + mtk_vdec_err(ctx, "failed to get VP9 vsi\n"); + ret = -EINVAL; + goto error_vsi; + } + instance->init_vsi = vsi; + instance->core_vsi = mtk_vcodec_fw_map_dm_addr(ctx->dev->fw_handler, + (u32)vsi->core_vsi); + if (!instance->core_vsi) { + mtk_vdec_err(ctx, "failed to get VP9 core vsi\n"); + ret = -EINVAL; + goto error_vsi; + } + + instance->irq = 1; + + ret = vdec_vp9_slice_init_default_frame_ctx(instance); + if (ret) + goto error_default_frame_ctx; + + ctx->drv_handle = instance; + + return 0; + +error_default_frame_ctx: +error_vsi: + vpu_dec_deinit(&instance->vpu); +error_vpu_init: + kfree(instance); + return ret; +} + +static void vdec_vp9_slice_deinit(void *h_vdec) +{ + struct vdec_vp9_slice_instance *instance = h_vdec; + + if (!instance) + return; + + vpu_dec_deinit(&instance->vpu); + vdec_vp9_slice_free_working_buffer(instance); + vdec_msg_queue_deinit(&instance->ctx->msg_queue, instance->ctx); + kfree(instance); +} + +static int vdec_vp9_slice_flush(void *h_vdec, struct mtk_vcodec_mem *bs, + struct vdec_fb *fb, bool *res_chg) +{ + struct vdec_vp9_slice_instance *instance = h_vdec; + + mtk_vdec_debug(instance->ctx, "flush ...\n"); + if (instance->ctx->dev->vdec_pdata->hw_arch != MTK_VDEC_PURE_SINGLE_CORE) + vdec_msg_queue_wait_lat_buf_full(&instance->ctx->msg_queue); + return vpu_dec_reset(&instance->vpu); +} + +static void vdec_vp9_slice_get_pic_info(struct vdec_vp9_slice_instance *instance) +{ + struct mtk_vcodec_dec_ctx *ctx = instance->ctx; + unsigned int data[3]; + + mtk_vdec_debug(instance->ctx, "w %u h %u\n", ctx->picinfo.pic_w, ctx->picinfo.pic_h); + + data[0] = ctx->picinfo.pic_w; + data[1] = ctx->picinfo.pic_h; + data[2] = ctx->capture_fourcc; + vpu_dec_get_param(&instance->vpu, data, 3, GET_PARAM_PIC_INFO); + + ctx->picinfo.buf_w = ALIGN(ctx->picinfo.pic_w, 64); + ctx->picinfo.buf_h = ALIGN(ctx->picinfo.pic_h, 64); + ctx->picinfo.fb_sz[0] = instance->vpu.fb_sz[0]; + ctx->picinfo.fb_sz[1] = instance->vpu.fb_sz[1]; +} + +static void vdec_vp9_slice_get_dpb_size(struct vdec_vp9_slice_instance *instance, + unsigned int *dpb_sz) +{ + /* refer VP9 specification */ + *dpb_sz = 9; +} + +static int vdec_vp9_slice_get_param(void *h_vdec, enum vdec_get_param_type type, void *out) +{ + struct vdec_vp9_slice_instance *instance = h_vdec; + + switch (type) { + case GET_PARAM_PIC_INFO: + vdec_vp9_slice_get_pic_info(instance); + break; + case GET_PARAM_DPB_SIZE: + vdec_vp9_slice_get_dpb_size(instance, out); + break; + case GET_PARAM_CROP_INFO: + mtk_vdec_debug(instance->ctx, "No need to get vp9 crop information."); + break; + default: + mtk_vdec_err(instance->ctx, "invalid get parameter type=%d\n", type); + return -EINVAL; + } + + return 0; +} + +static int vdec_vp9_slice_single_decode(void *h_vdec, struct mtk_vcodec_mem *bs, + struct vdec_fb *fb, bool *res_chg) +{ + struct vdec_vp9_slice_instance *instance = h_vdec; + struct vdec_vp9_slice_pfc *pfc = &instance->sc_pfc; + struct vdec_vp9_slice_vsi *vsi; + struct mtk_vcodec_dec_ctx *ctx; + int ret; + + if (!instance || !instance->ctx) + return -EINVAL; + ctx = instance->ctx; + + /* bs NULL means flush decoder */ + if (!bs) + return vdec_vp9_slice_flush(h_vdec, bs, fb, res_chg); + + fb = ctx->dev->vdec_pdata->get_cap_buffer(ctx); + if (!fb) + return -EBUSY; + + vsi = &pfc->vsi; + + ret = vdec_vp9_slice_setup_single(instance, bs, fb, pfc); + if (ret) { + mtk_vdec_err(ctx, "Failed to setup VP9 single ret %d\n", ret); + return ret; + } + vdec_vp9_slice_vsi_to_remote(vsi, instance->vsi); + + ret = vpu_dec_start(&instance->vpu, NULL, 0); + if (ret) { + mtk_vdec_err(ctx, "Failed to dec VP9 ret %d\n", ret); + return ret; + } + + ret = mtk_vcodec_wait_for_done_ctx(ctx, MTK_INST_IRQ_RECEIVED, + WAIT_INTR_TIMEOUT_MS, MTK_VDEC_CORE); + /* update remote vsi if decode timeout */ + if (ret) { + mtk_vdec_err(ctx, "VP9 decode timeout %d\n", ret); + WRITE_ONCE(instance->vsi->state.timeout, 1); + } + + vpu_dec_end(&instance->vpu); + + vdec_vp9_slice_vsi_from_remote(vsi, instance->vsi, 0); + ret = vdec_vp9_slice_update_single(instance, pfc); + if (ret) { + mtk_vdec_err(ctx, "VP9 decode error: %d\n", ret); + return ret; + } + + instance->ctx->decoded_frame_cnt++; + return 0; +} + +static int vdec_vp9_slice_lat_decode(void *h_vdec, struct mtk_vcodec_mem *bs, + struct vdec_fb *fb, bool *res_chg) +{ + struct vdec_vp9_slice_instance *instance = h_vdec; + struct vdec_lat_buf *lat_buf; + struct vdec_vp9_slice_pfc *pfc; + struct vdec_vp9_slice_vsi *vsi; + struct mtk_vcodec_dec_ctx *ctx; + int ret; + + if (!instance || !instance->ctx) + return -EINVAL; + ctx = instance->ctx; + + /* init msgQ for the first time */ + if (vdec_msg_queue_init(&ctx->msg_queue, ctx, + vdec_vp9_slice_core_decode, + sizeof(*pfc))) + return -ENOMEM; + + /* bs NULL means flush decoder */ + if (!bs) + return vdec_vp9_slice_flush(h_vdec, bs, fb, res_chg); + + lat_buf = vdec_msg_queue_dqbuf(&instance->ctx->msg_queue.lat_ctx); + if (!lat_buf) { + mtk_vdec_debug(ctx, "Failed to get VP9 lat buf\n"); + return -EAGAIN; + } + pfc = (struct vdec_vp9_slice_pfc *)lat_buf->private_data; + if (!pfc) { + ret = -EINVAL; + goto err_free_fb_out; + } + vsi = &pfc->vsi; + + ret = vdec_vp9_slice_setup_lat(instance, bs, lat_buf, pfc); + if (ret) { + mtk_vdec_err(ctx, "Failed to setup VP9 lat ret %d\n", ret); + goto err_free_fb_out; + } + vdec_vp9_slice_vsi_to_remote(vsi, instance->vsi); + + ret = vpu_dec_start(&instance->vpu, NULL, 0); + if (ret) { + mtk_vdec_err(ctx, "Failed to dec VP9 ret %d\n", ret); + goto err_free_fb_out; + } + + if (instance->irq) { + ret = mtk_vcodec_wait_for_done_ctx(ctx, MTK_INST_IRQ_RECEIVED, + WAIT_INTR_TIMEOUT_MS, MTK_VDEC_LAT0); + /* update remote vsi if decode timeout */ + if (ret) { + mtk_vdec_err(ctx, "VP9 decode timeout %d pic %d\n", ret, pfc->seq); + WRITE_ONCE(instance->vsi->state.timeout, 1); + } + vpu_dec_end(&instance->vpu); + } + + vdec_vp9_slice_vsi_from_remote(vsi, instance->vsi, 0); + ret = vdec_vp9_slice_update_lat(instance, lat_buf, pfc); + + /* LAT trans full, no more UBE or decode timeout */ + if (ret) { + mtk_vdec_err(ctx, "VP9 decode error: %d\n", ret); + goto err_free_fb_out; + } + + mtk_vdec_debug(ctx, "lat dma addr: 0x%lx 0x%lx\n", + (unsigned long)pfc->vsi.trans.dma_addr, + (unsigned long)pfc->vsi.trans.dma_addr_end); + + vdec_msg_queue_update_ube_wptr(&ctx->msg_queue, + vsi->trans.dma_addr_end + + ctx->msg_queue.wdma_addr.dma_addr); + vdec_msg_queue_qbuf(&ctx->msg_queue.core_ctx, lat_buf); + + return 0; +err_free_fb_out: + vdec_msg_queue_qbuf(&ctx->msg_queue.lat_ctx, lat_buf); + return ret; +} + +static int vdec_vp9_slice_decode(void *h_vdec, struct mtk_vcodec_mem *bs, + struct vdec_fb *fb, bool *res_chg) +{ + struct vdec_vp9_slice_instance *instance = h_vdec; + int ret; + + if (instance->ctx->dev->vdec_pdata->hw_arch == MTK_VDEC_PURE_SINGLE_CORE) + ret = vdec_vp9_slice_single_decode(h_vdec, bs, fb, res_chg); + else + ret = vdec_vp9_slice_lat_decode(h_vdec, bs, fb, res_chg); + + return ret; +} + +static int vdec_vp9_slice_core_decode(struct vdec_lat_buf *lat_buf) +{ + struct vdec_vp9_slice_instance *instance; + struct vdec_vp9_slice_pfc *pfc; + struct mtk_vcodec_dec_ctx *ctx = NULL; + struct vdec_fb *fb = NULL; + int ret = -EINVAL; + + if (!lat_buf) + goto err; + + pfc = lat_buf->private_data; + ctx = lat_buf->ctx; + if (!pfc || !ctx) + goto err; + + instance = ctx->drv_handle; + if (!instance) + goto err; + + fb = ctx->dev->vdec_pdata->get_cap_buffer(ctx); + if (!fb) { + ret = -EBUSY; + goto err; + } + + ret = vdec_vp9_slice_setup_core(instance, fb, lat_buf, pfc); + if (ret) { + mtk_vdec_err(ctx, "vdec_vp9_slice_setup_core\n"); + goto err; + } + vdec_vp9_slice_vsi_to_remote(&pfc->vsi, instance->core_vsi); + + ret = vpu_dec_core(&instance->vpu); + if (ret) { + mtk_vdec_err(ctx, "vpu_dec_core\n"); + goto err; + } + + if (instance->irq) { + ret = mtk_vcodec_wait_for_done_ctx(ctx, MTK_INST_IRQ_RECEIVED, + WAIT_INTR_TIMEOUT_MS, MTK_VDEC_CORE); + /* update remote vsi if decode timeout */ + if (ret) { + mtk_vdec_err(ctx, "VP9 core timeout pic %d\n", pfc->seq); + WRITE_ONCE(instance->core_vsi->state.timeout, 1); + } + vpu_dec_core_end(&instance->vpu); + } + + vdec_vp9_slice_vsi_from_remote(&pfc->vsi, instance->core_vsi, 1); + ret = vdec_vp9_slice_update_core(instance, lat_buf, pfc); + if (ret) { + mtk_vdec_err(ctx, "vdec_vp9_slice_update_core\n"); + goto err; + } + + pfc->vsi.trans.dma_addr_end += ctx->msg_queue.wdma_addr.dma_addr; + mtk_vdec_debug(ctx, "core dma_addr_end 0x%lx\n", + (unsigned long)pfc->vsi.trans.dma_addr_end); + vdec_msg_queue_update_ube_rptr(&ctx->msg_queue, pfc->vsi.trans.dma_addr_end); + ctx->dev->vdec_pdata->cap_to_disp(ctx, 0, lat_buf->src_buf_req); + + return 0; + +err: + if (ctx && pfc) { + /* always update read pointer */ + vdec_msg_queue_update_ube_rptr(&ctx->msg_queue, pfc->vsi.trans.dma_addr_end); + + if (fb) + ctx->dev->vdec_pdata->cap_to_disp(ctx, 1, lat_buf->src_buf_req); + } + return ret; +} + +const struct vdec_common_if vdec_vp9_slice_lat_if = { + .init = vdec_vp9_slice_init, + .decode = vdec_vp9_slice_decode, + .get_param = vdec_vp9_slice_get_param, + .deinit = vdec_vp9_slice_deinit, +}; diff --git a/drivers/media/platform/mediatek/vcodec/decoder/vdec_drv_base.h b/drivers/media/platform/mediatek/vcodec/decoder/vdec_drv_base.h new file mode 100644 index 0000000000..f6abb93652 --- /dev/null +++ b/drivers/media/platform/mediatek/vcodec/decoder/vdec_drv_base.h @@ -0,0 +1,46 @@ +/* SPDX-License-Identifier: GPL-2.0 */ +/* + * Copyright (c) 2016 MediaTek Inc. + * Author: PC Chen <pc.chen@mediatek.com> + */ + +#ifndef _VDEC_DRV_BASE_ +#define _VDEC_DRV_BASE_ + +#include "vdec_drv_if.h" + +struct vdec_common_if { + /** + * (*init)() - initialize decode driver + * @ctx : [in] mtk v4l2 context + * @h_vdec : [out] driver handle + */ + int (*init)(struct mtk_vcodec_dec_ctx *ctx); + + /** + * (*decode)() - trigger decode + * @h_vdec : [in] driver handle + * @bs : [in] input bitstream + * @fb : [in] frame buffer to store decoded frame + * @res_chg : [out] resolution change happen + */ + int (*decode)(void *h_vdec, struct mtk_vcodec_mem *bs, + struct vdec_fb *fb, bool *res_chg); + + /** + * (*get_param)() - get driver's parameter + * @h_vdec : [in] driver handle + * @type : [in] input parameter type + * @out : [out] buffer to store query result + */ + int (*get_param)(void *h_vdec, enum vdec_get_param_type type, + void *out); + + /** + * (*deinit)() - deinitialize driver. + * @h_vdec : [in] driver handle to be deinit + */ + void (*deinit)(void *h_vdec); +}; + +#endif diff --git a/drivers/media/platform/mediatek/vcodec/decoder/vdec_drv_if.c b/drivers/media/platform/mediatek/vcodec/decoder/vdec_drv_if.c new file mode 100644 index 0000000000..d0b459b160 --- /dev/null +++ b/drivers/media/platform/mediatek/vcodec/decoder/vdec_drv_if.c @@ -0,0 +1,128 @@ +// SPDX-License-Identifier: GPL-2.0 +/* + * Copyright (c) 2016 MediaTek Inc. + * Author: PC Chen <pc.chen@mediatek.com> + * Tiffany Lin <tiffany.lin@mediatek.com> + */ + +#include <linux/interrupt.h> +#include <linux/kernel.h> +#include <linux/slab.h> + +#include "vdec_drv_if.h" +#include "mtk_vcodec_dec.h" +#include "vdec_drv_base.h" +#include "mtk_vcodec_dec_pm.h" + +int vdec_if_init(struct mtk_vcodec_dec_ctx *ctx, unsigned int fourcc) +{ + enum mtk_vdec_hw_arch hw_arch = ctx->dev->vdec_pdata->hw_arch; + int ret = 0; + + switch (fourcc) { + case V4L2_PIX_FMT_H264_SLICE: + if (!ctx->dev->vdec_pdata->is_subdev_supported) { + ctx->dec_if = &vdec_h264_slice_if; + ctx->hw_id = MTK_VDEC_CORE; + } else { + ctx->dec_if = &vdec_h264_slice_multi_if; + ctx->hw_id = IS_VDEC_LAT_ARCH(hw_arch) ? MTK_VDEC_LAT0 : MTK_VDEC_CORE; + } + break; + case V4L2_PIX_FMT_H264: + ctx->dec_if = &vdec_h264_if; + ctx->hw_id = MTK_VDEC_CORE; + break; + case V4L2_PIX_FMT_VP8_FRAME: + ctx->dec_if = &vdec_vp8_slice_if; + ctx->hw_id = MTK_VDEC_CORE; + break; + case V4L2_PIX_FMT_VP8: + ctx->dec_if = &vdec_vp8_if; + ctx->hw_id = MTK_VDEC_CORE; + break; + case V4L2_PIX_FMT_VP9: + ctx->dec_if = &vdec_vp9_if; + ctx->hw_id = MTK_VDEC_CORE; + break; + case V4L2_PIX_FMT_VP9_FRAME: + ctx->dec_if = &vdec_vp9_slice_lat_if; + ctx->hw_id = IS_VDEC_LAT_ARCH(hw_arch) ? MTK_VDEC_LAT0 : MTK_VDEC_CORE; + break; + case V4L2_PIX_FMT_HEVC_SLICE: + ctx->dec_if = &vdec_hevc_slice_multi_if; + ctx->hw_id = MTK_VDEC_LAT0; + break; + case V4L2_PIX_FMT_AV1_FRAME: + ctx->dec_if = &vdec_av1_slice_lat_if; + ctx->hw_id = MTK_VDEC_LAT0; + break; + default: + return -EINVAL; + } + + mtk_vcodec_dec_enable_hardware(ctx, ctx->hw_id); + ret = ctx->dec_if->init(ctx); + mtk_vcodec_dec_disable_hardware(ctx, ctx->hw_id); + + return ret; +} + +int vdec_if_decode(struct mtk_vcodec_dec_ctx *ctx, struct mtk_vcodec_mem *bs, + struct vdec_fb *fb, bool *res_chg) +{ + int ret = 0; + + if (bs) { + if ((bs->dma_addr & 63) != 0) { + mtk_v4l2_vdec_err(ctx, "bs dma_addr should 64 byte align"); + return -EINVAL; + } + } + + if (fb) { + if (((fb->base_y.dma_addr & 511) != 0) || + ((fb->base_c.dma_addr & 511) != 0)) { + mtk_v4l2_vdec_err(ctx, "frame buffer dma_addr should 512 byte align"); + return -EINVAL; + } + } + + if (!ctx->drv_handle) + return -EIO; + + mtk_vcodec_dec_enable_hardware(ctx, ctx->hw_id); + mtk_vcodec_set_curr_ctx(ctx->dev, ctx, ctx->hw_id); + ret = ctx->dec_if->decode(ctx->drv_handle, bs, fb, res_chg); + mtk_vcodec_set_curr_ctx(ctx->dev, NULL, ctx->hw_id); + mtk_vcodec_dec_disable_hardware(ctx, ctx->hw_id); + + return ret; +} + +int vdec_if_get_param(struct mtk_vcodec_dec_ctx *ctx, enum vdec_get_param_type type, + void *out) +{ + int ret = 0; + + if (!ctx->drv_handle) + return -EIO; + + mtk_vdec_lock(ctx); + ret = ctx->dec_if->get_param(ctx->drv_handle, type, out); + mtk_vdec_unlock(ctx); + + return ret; +} + +void vdec_if_deinit(struct mtk_vcodec_dec_ctx *ctx) +{ + if (!ctx->drv_handle) + return; + + mtk_vcodec_dec_enable_hardware(ctx, ctx->hw_id); + ctx->dec_if->deinit(ctx->drv_handle); + mtk_vcodec_dec_disable_hardware(ctx, ctx->hw_id); + + ctx->drv_handle = NULL; +} diff --git a/drivers/media/platform/mediatek/vcodec/decoder/vdec_drv_if.h b/drivers/media/platform/mediatek/vcodec/decoder/vdec_drv_if.h new file mode 100644 index 0000000000..bfd297c968 --- /dev/null +++ b/drivers/media/platform/mediatek/vcodec/decoder/vdec_drv_if.h @@ -0,0 +1,103 @@ +/* SPDX-License-Identifier: GPL-2.0 */ +/* + * Copyright (c) 2016 MediaTek Inc. + * Author: PC Chen <pc.chen@mediatek.com> + * Tiffany Lin <tiffany.lin@mediatek.com> + */ + +#ifndef _VDEC_DRV_IF_H_ +#define _VDEC_DRV_IF_H_ + +#include "mtk_vcodec_dec.h" + + +/** + * enum vdec_fb_status - decoder frame buffer status + * @FB_ST_NORMAL: initial state + * @FB_ST_DISPLAY: frame buffer is ready to be displayed + * @FB_ST_FREE: frame buffer is not used by decoder any more + */ +enum vdec_fb_status { + FB_ST_NORMAL = 0, + FB_ST_DISPLAY = (1 << 0), + FB_ST_FREE = (1 << 1) +}; + +/* For GET_PARAM_DISP_FRAME_BUFFER and GET_PARAM_FREE_FRAME_BUFFER, + * the caller does not own the returned buffer. The buffer will not be + * released before vdec_if_deinit. + * GET_PARAM_DISP_FRAME_BUFFER : get next displayable frame buffer, + * struct vdec_fb** + * GET_PARAM_FREE_FRAME_BUFFER : get non-referenced framebuffer, vdec_fb** + * GET_PARAM_PIC_INFO : get picture info, struct vdec_pic_info* + * GET_PARAM_CROP_INFO : get crop info, struct v4l2_crop* + * GET_PARAM_DPB_SIZE : get dpb size, unsigned int* + */ +enum vdec_get_param_type { + GET_PARAM_DISP_FRAME_BUFFER, + GET_PARAM_FREE_FRAME_BUFFER, + GET_PARAM_PIC_INFO, + GET_PARAM_CROP_INFO, + GET_PARAM_DPB_SIZE +}; + +/** + * struct vdec_fb_node - decoder frame buffer node + * @list : list to hold this node + * @fb : point to frame buffer (vdec_fb), fb could point to frame buffer and + * working buffer this is for maintain buffers in different state + */ +struct vdec_fb_node { + struct list_head list; + struct vdec_fb *fb; +}; + +extern const struct vdec_common_if vdec_h264_if; +extern const struct vdec_common_if vdec_h264_slice_if; +extern const struct vdec_common_if vdec_h264_slice_multi_if; +extern const struct vdec_common_if vdec_vp8_if; +extern const struct vdec_common_if vdec_vp8_slice_if; +extern const struct vdec_common_if vdec_vp9_if; +extern const struct vdec_common_if vdec_vp9_slice_lat_if; +extern const struct vdec_common_if vdec_hevc_slice_multi_if; +extern const struct vdec_common_if vdec_av1_slice_lat_if; + +/** + * vdec_if_init() - initialize decode driver + * @ctx : [in] v4l2 context + * @fourcc : [in] video format fourcc, V4L2_PIX_FMT_H264/VP8/VP9.. + */ +int vdec_if_init(struct mtk_vcodec_dec_ctx *ctx, unsigned int fourcc); + +/** + * vdec_if_deinit() - deinitialize decode driver + * @ctx : [in] v4l2 context + * + */ +void vdec_if_deinit(struct mtk_vcodec_dec_ctx *ctx); + +/** + * vdec_if_decode() - trigger decode + * @ctx : [in] v4l2 context + * @bs : [in] input bitstream + * @fb : [in] frame buffer to store decoded frame, when null means parse + * header only + * @res_chg : [out] resolution change happens if current bs have different + * picture width/height + * Note: To flush the decoder when reaching EOF, set input bitstream as NULL. + * + * Return: 0 on success. -EIO on unrecoverable error. + */ +int vdec_if_decode(struct mtk_vcodec_dec_ctx *ctx, struct mtk_vcodec_mem *bs, + struct vdec_fb *fb, bool *res_chg); + +/** + * vdec_if_get_param() - get driver's parameter + * @ctx : [in] v4l2 context + * @type : [in] input parameter type + * @out : [out] buffer to store query result + */ +int vdec_if_get_param(struct mtk_vcodec_dec_ctx *ctx, enum vdec_get_param_type type, + void *out); + +#endif diff --git a/drivers/media/platform/mediatek/vcodec/decoder/vdec_ipi_msg.h b/drivers/media/platform/mediatek/vcodec/decoder/vdec_ipi_msg.h new file mode 100644 index 0000000000..47070be2a9 --- /dev/null +++ b/drivers/media/platform/mediatek/vcodec/decoder/vdec_ipi_msg.h @@ -0,0 +1,153 @@ +/* SPDX-License-Identifier: GPL-2.0 */ +/* + * Copyright (c) 2016 MediaTek Inc. + * Author: PC Chen <pc.chen@mediatek.com> + */ + +#ifndef _VDEC_IPI_MSG_H_ +#define _VDEC_IPI_MSG_H_ + +/* + * enum vdec_ipi_msgid - message id between AP and VPU + * @AP_IPIMSG_XXX : AP to VPU cmd message id + * @VPU_IPIMSG_XXX_ACK : VPU ack AP cmd message id + */ +enum vdec_ipi_msgid { + AP_IPIMSG_DEC_INIT = 0xA000, + AP_IPIMSG_DEC_START = 0xA001, + AP_IPIMSG_DEC_END = 0xA002, + AP_IPIMSG_DEC_DEINIT = 0xA003, + AP_IPIMSG_DEC_RESET = 0xA004, + AP_IPIMSG_DEC_CORE = 0xA005, + AP_IPIMSG_DEC_CORE_END = 0xA006, + AP_IPIMSG_DEC_GET_PARAM = 0xA007, + + VPU_IPIMSG_DEC_INIT_ACK = 0xB000, + VPU_IPIMSG_DEC_START_ACK = 0xB001, + VPU_IPIMSG_DEC_END_ACK = 0xB002, + VPU_IPIMSG_DEC_DEINIT_ACK = 0xB003, + VPU_IPIMSG_DEC_RESET_ACK = 0xB004, + VPU_IPIMSG_DEC_CORE_ACK = 0xB005, + VPU_IPIMSG_DEC_CORE_END_ACK = 0xB006, + VPU_IPIMSG_DEC_GET_PARAM_ACK = 0xB007, +}; + +/** + * struct vdec_ap_ipi_cmd - generic AP to VPU ipi command format + * @msg_id : vdec_ipi_msgid + * @vpu_inst_addr : VPU decoder instance address. Used if ABI version < 2. + * @inst_id : instance ID. Used if the ABI version >= 2. + * @codec_type : codec fourcc + * @reserved : reserved param + */ +struct vdec_ap_ipi_cmd { + uint32_t msg_id; + union { + uint32_t vpu_inst_addr; + uint32_t inst_id; + }; + u32 codec_type; + u32 reserved; +}; + +/** + * struct vdec_vpu_ipi_ack - generic VPU to AP ipi command format + * @msg_id : vdec_ipi_msgid + * @status : VPU exeuction result + * @ap_inst_addr : AP video decoder instance address + */ +struct vdec_vpu_ipi_ack { + uint32_t msg_id; + int32_t status; + uint64_t ap_inst_addr; +}; + +/** + * struct vdec_ap_ipi_init - for AP_IPIMSG_DEC_INIT + * @msg_id : AP_IPIMSG_DEC_INIT + * @codec_type : codec fourcc + * @ap_inst_addr : AP video decoder instance address + */ +struct vdec_ap_ipi_init { + uint32_t msg_id; + u32 codec_type; + uint64_t ap_inst_addr; +}; + +/** + * struct vdec_ap_ipi_dec_start - for AP_IPIMSG_DEC_START + * @msg_id : AP_IPIMSG_DEC_START + * @vpu_inst_addr : VPU decoder instance address. Used if ABI version < 2. + * @inst_id : instance ID. Used if the ABI version >= 2. + * @data : Header info + * H264 decoder [0]:buf_sz [1]:nal_start + * VP8 decoder [0]:width/height + * VP9 decoder [0]:profile, [1][2] width/height + * @codec_type : codec fourcc + */ +struct vdec_ap_ipi_dec_start { + uint32_t msg_id; + union { + uint32_t vpu_inst_addr; + uint32_t inst_id; + }; + uint32_t data[3]; + u32 codec_type; +}; + +/** + * struct vdec_vpu_ipi_init_ack - for VPU_IPIMSG_DEC_INIT_ACK + * @msg_id : VPU_IPIMSG_DEC_INIT_ACK + * @status : VPU exeuction result + * @ap_inst_addr : AP vcodec_vpu_inst instance address + * @vpu_inst_addr : VPU decoder instance address + * @vdec_abi_version: ABI version of the firmware. Kernel can use it to + * ensure that it is compatible with the firmware. + * This field is not valid for MT8173 and must not be + * accessed for this chip. + * @inst_id : instance ID. Valid only if the ABI version >= 2. + */ +struct vdec_vpu_ipi_init_ack { + uint32_t msg_id; + int32_t status; + uint64_t ap_inst_addr; + uint32_t vpu_inst_addr; + uint32_t vdec_abi_version; + uint32_t inst_id; +}; + +/** + * struct vdec_ap_ipi_get_param - for AP_IPIMSG_DEC_GET_PARAM + * @msg_id : AP_IPIMSG_DEC_GET_PARAM + * @inst_id : instance ID. Used if the ABI version >= 2. + * @data : picture information + * @param_type : get param type + * @codec_type : Codec fourcc + */ +struct vdec_ap_ipi_get_param { + u32 msg_id; + u32 inst_id; + u32 data[4]; + u32 param_type; + u32 codec_type; +}; + +/** + * struct vdec_vpu_ipi_get_param_ack - for VPU_IPIMSG_DEC_GET_PARAM_ACK + * @msg_id : VPU_IPIMSG_DEC_GET_PARAM_ACK + * @status : VPU execution result + * @ap_inst_addr : AP vcodec_vpu_inst instance address + * @data : picture information from SCP. + * @param_type : get param type + * @reserved : reserved param + */ +struct vdec_vpu_ipi_get_param_ack { + u32 msg_id; + s32 status; + u64 ap_inst_addr; + u32 data[4]; + u32 param_type; + u32 reserved; +}; + +#endif diff --git a/drivers/media/platform/mediatek/vcodec/decoder/vdec_msg_queue.c b/drivers/media/platform/mediatek/vcodec/decoder/vdec_msg_queue.c new file mode 100644 index 0000000000..f283c4703d --- /dev/null +++ b/drivers/media/platform/mediatek/vcodec/decoder/vdec_msg_queue.c @@ -0,0 +1,372 @@ +// SPDX-License-Identifier: GPL-2.0 +/* + * Copyright (c) 2021 MediaTek Inc. + * Author: Yunfei Dong <yunfei.dong@mediatek.com> + */ + +#include <linux/freezer.h> +#include <linux/interrupt.h> +#include <linux/kthread.h> + +#include "mtk_vcodec_dec_drv.h" +#include "mtk_vcodec_dec_pm.h" +#include "vdec_msg_queue.h" + +#define VDEC_MSG_QUEUE_TIMEOUT_MS 1500 + +/* the size used to store lat slice header information */ +#define VDEC_LAT_SLICE_HEADER_SZ (640 * SZ_1K) + +/* the size used to store avc error information */ +#define VDEC_ERR_MAP_SZ_AVC (17 * SZ_1K) + +#define VDEC_RD_MV_BUFFER_SZ (((SZ_4K * 2304 >> 4) + SZ_1K) << 1) +#define VDEC_LAT_TILE_SZ (64 * V4L2_AV1_MAX_TILE_COUNT) + +/* core will read the trans buffer which decoded by lat to decode again. + * The trans buffer size of FHD and 4K bitstreams are different. + */ +static int vde_msg_queue_get_trans_size(int width, int height) +{ + if (width > 1920 || height > 1088) + return 30 * SZ_1M; + else + return 6 * SZ_1M; +} + +void vdec_msg_queue_init_ctx(struct vdec_msg_queue_ctx *ctx, int hardware_index) +{ + init_waitqueue_head(&ctx->ready_to_use); + INIT_LIST_HEAD(&ctx->ready_queue); + spin_lock_init(&ctx->ready_lock); + ctx->ready_num = 0; + ctx->hardware_index = hardware_index; +} + +static struct list_head *vdec_get_buf_list(int hardware_index, struct vdec_lat_buf *buf) +{ + switch (hardware_index) { + case MTK_VDEC_CORE: + return &buf->core_list; + case MTK_VDEC_LAT0: + return &buf->lat_list; + default: + return NULL; + } +} + +static void vdec_msg_queue_inc(struct vdec_msg_queue *msg_queue, int hardware_index) +{ + if (hardware_index == MTK_VDEC_CORE) + atomic_inc(&msg_queue->core_list_cnt); + else + atomic_inc(&msg_queue->lat_list_cnt); +} + +static void vdec_msg_queue_dec(struct vdec_msg_queue *msg_queue, int hardware_index) +{ + if (hardware_index == MTK_VDEC_CORE) + atomic_dec(&msg_queue->core_list_cnt); + else + atomic_dec(&msg_queue->lat_list_cnt); +} + +int vdec_msg_queue_qbuf(struct vdec_msg_queue_ctx *msg_ctx, struct vdec_lat_buf *buf) +{ + struct list_head *head; + + head = vdec_get_buf_list(msg_ctx->hardware_index, buf); + if (!head) { + mtk_v4l2_vdec_err(buf->ctx, "fail to qbuf: %d", msg_ctx->hardware_index); + return -EINVAL; + } + + spin_lock(&msg_ctx->ready_lock); + list_add_tail(head, &msg_ctx->ready_queue); + msg_ctx->ready_num++; + + vdec_msg_queue_inc(&buf->ctx->msg_queue, msg_ctx->hardware_index); + if (msg_ctx->hardware_index != MTK_VDEC_CORE) { + wake_up_all(&msg_ctx->ready_to_use); + } else { + if (!(buf->ctx->msg_queue.status & CONTEXT_LIST_QUEUED)) { + queue_work(buf->ctx->dev->core_workqueue, &buf->ctx->msg_queue.core_work); + buf->ctx->msg_queue.status |= CONTEXT_LIST_QUEUED; + } + } + + mtk_v4l2_vdec_dbg(3, buf->ctx, "enqueue buf type: %d addr: 0x%p num: %d", + msg_ctx->hardware_index, buf, msg_ctx->ready_num); + spin_unlock(&msg_ctx->ready_lock); + + return 0; +} + +static bool vdec_msg_queue_wait_event(struct vdec_msg_queue_ctx *msg_ctx) +{ + int ret; + + ret = wait_event_timeout(msg_ctx->ready_to_use, + !list_empty(&msg_ctx->ready_queue), + msecs_to_jiffies(VDEC_MSG_QUEUE_TIMEOUT_MS)); + if (!ret) + return false; + + return true; +} + +struct vdec_lat_buf *vdec_msg_queue_dqbuf(struct vdec_msg_queue_ctx *msg_ctx) +{ + struct vdec_lat_buf *buf; + struct list_head *head; + int ret; + + spin_lock(&msg_ctx->ready_lock); + if (list_empty(&msg_ctx->ready_queue)) { + spin_unlock(&msg_ctx->ready_lock); + + if (msg_ctx->hardware_index == MTK_VDEC_CORE) + return NULL; + + ret = vdec_msg_queue_wait_event(msg_ctx); + if (!ret) + return NULL; + spin_lock(&msg_ctx->ready_lock); + } + + if (msg_ctx->hardware_index == MTK_VDEC_CORE) + buf = list_first_entry(&msg_ctx->ready_queue, + struct vdec_lat_buf, core_list); + else + buf = list_first_entry(&msg_ctx->ready_queue, + struct vdec_lat_buf, lat_list); + + head = vdec_get_buf_list(msg_ctx->hardware_index, buf); + if (!head) { + spin_unlock(&msg_ctx->ready_lock); + mtk_v4l2_vdec_err(buf->ctx, "fail to dqbuf: %d", msg_ctx->hardware_index); + return NULL; + } + list_del(head); + vdec_msg_queue_dec(&buf->ctx->msg_queue, msg_ctx->hardware_index); + + msg_ctx->ready_num--; + mtk_v4l2_vdec_dbg(3, buf->ctx, "dqueue buf type:%d addr: 0x%p num: %d", + msg_ctx->hardware_index, buf, msg_ctx->ready_num); + spin_unlock(&msg_ctx->ready_lock); + + return buf; +} + +void vdec_msg_queue_update_ube_rptr(struct vdec_msg_queue *msg_queue, uint64_t ube_rptr) +{ + spin_lock(&msg_queue->lat_ctx.ready_lock); + msg_queue->wdma_rptr_addr = ube_rptr; + mtk_v4l2_vdec_dbg(3, msg_queue->ctx, "update ube rprt (0x%llx)", ube_rptr); + spin_unlock(&msg_queue->lat_ctx.ready_lock); +} + +void vdec_msg_queue_update_ube_wptr(struct vdec_msg_queue *msg_queue, uint64_t ube_wptr) +{ + spin_lock(&msg_queue->lat_ctx.ready_lock); + msg_queue->wdma_wptr_addr = ube_wptr; + mtk_v4l2_vdec_dbg(3, msg_queue->ctx, "update ube wprt: (0x%llx 0x%llx) offset: 0x%llx", + msg_queue->wdma_rptr_addr, msg_queue->wdma_wptr_addr, + ube_wptr); + spin_unlock(&msg_queue->lat_ctx.ready_lock); +} + +bool vdec_msg_queue_wait_lat_buf_full(struct vdec_msg_queue *msg_queue) +{ + if (atomic_read(&msg_queue->lat_list_cnt) == NUM_BUFFER_COUNT) { + mtk_v4l2_vdec_dbg(3, msg_queue->ctx, "wait buf full: (%d %d) ready:%d status:%d", + atomic_read(&msg_queue->lat_list_cnt), + atomic_read(&msg_queue->core_list_cnt), + msg_queue->lat_ctx.ready_num, msg_queue->status); + return true; + } + + msg_queue->flush_done = false; + vdec_msg_queue_qbuf(&msg_queue->core_ctx, &msg_queue->empty_lat_buf); + wait_event(msg_queue->core_dec_done, msg_queue->flush_done); + + mtk_v4l2_vdec_dbg(3, msg_queue->ctx, "flush done => ready_num:%d status:%d list(%d %d)", + msg_queue->lat_ctx.ready_num, msg_queue->status, + atomic_read(&msg_queue->lat_list_cnt), + atomic_read(&msg_queue->core_list_cnt)); + + return false; +} + +void vdec_msg_queue_deinit(struct vdec_msg_queue *msg_queue, + struct mtk_vcodec_dec_ctx *ctx) +{ + struct vdec_lat_buf *lat_buf; + struct mtk_vcodec_mem *mem; + int i; + + mem = &msg_queue->wdma_addr; + if (mem->va) + mtk_vcodec_mem_free(ctx, mem); + for (i = 0; i < NUM_BUFFER_COUNT; i++) { + lat_buf = &msg_queue->lat_buf[i]; + + mem = &lat_buf->wdma_err_addr; + if (mem->va) + mtk_vcodec_mem_free(ctx, mem); + + mem = &lat_buf->slice_bc_addr; + if (mem->va) + mtk_vcodec_mem_free(ctx, mem); + + mem = &lat_buf->rd_mv_addr; + if (mem->va) + mtk_vcodec_mem_free(ctx, mem); + + mem = &lat_buf->tile_addr; + if (mem->va) + mtk_vcodec_mem_free(ctx, mem); + + kfree(lat_buf->private_data); + lat_buf->private_data = NULL; + } + + if (msg_queue->wdma_addr.size) + cancel_work_sync(&msg_queue->core_work); +} + +static void vdec_msg_queue_core_work(struct work_struct *work) +{ + struct vdec_msg_queue *msg_queue = + container_of(work, struct vdec_msg_queue, core_work); + struct mtk_vcodec_dec_ctx *ctx = + container_of(msg_queue, struct mtk_vcodec_dec_ctx, msg_queue); + struct mtk_vcodec_dec_dev *dev = ctx->dev; + struct vdec_lat_buf *lat_buf; + + spin_lock(&msg_queue->core_ctx.ready_lock); + ctx->msg_queue.status &= ~CONTEXT_LIST_QUEUED; + spin_unlock(&msg_queue->core_ctx.ready_lock); + + lat_buf = vdec_msg_queue_dqbuf(&msg_queue->core_ctx); + if (!lat_buf) + return; + + if (lat_buf->is_last_frame) { + ctx->msg_queue.status = CONTEXT_LIST_DEC_DONE; + msg_queue->flush_done = true; + wake_up(&ctx->msg_queue.core_dec_done); + + return; + } + + ctx = lat_buf->ctx; + mtk_vcodec_dec_enable_hardware(ctx, MTK_VDEC_CORE); + mtk_vcodec_set_curr_ctx(dev, ctx, MTK_VDEC_CORE); + + lat_buf->core_decode(lat_buf); + + mtk_vcodec_set_curr_ctx(dev, NULL, MTK_VDEC_CORE); + mtk_vcodec_dec_disable_hardware(ctx, MTK_VDEC_CORE); + vdec_msg_queue_qbuf(&ctx->msg_queue.lat_ctx, lat_buf); + + if (!(ctx->msg_queue.status & CONTEXT_LIST_QUEUED) && + atomic_read(&msg_queue->core_list_cnt)) { + spin_lock(&msg_queue->core_ctx.ready_lock); + ctx->msg_queue.status |= CONTEXT_LIST_QUEUED; + spin_unlock(&msg_queue->core_ctx.ready_lock); + queue_work(ctx->dev->core_workqueue, &msg_queue->core_work); + } +} + +int vdec_msg_queue_init(struct vdec_msg_queue *msg_queue, + struct mtk_vcodec_dec_ctx *ctx, core_decode_cb_t core_decode, + int private_size) +{ + struct vdec_lat_buf *lat_buf; + int i, err; + + /* already init msg queue */ + if (msg_queue->wdma_addr.size) + return 0; + + vdec_msg_queue_init_ctx(&msg_queue->lat_ctx, MTK_VDEC_LAT0); + vdec_msg_queue_init_ctx(&msg_queue->core_ctx, MTK_VDEC_CORE); + INIT_WORK(&msg_queue->core_work, vdec_msg_queue_core_work); + + atomic_set(&msg_queue->lat_list_cnt, 0); + atomic_set(&msg_queue->core_list_cnt, 0); + init_waitqueue_head(&msg_queue->core_dec_done); + msg_queue->status = CONTEXT_LIST_EMPTY; + + msg_queue->wdma_addr.size = + vde_msg_queue_get_trans_size(ctx->picinfo.buf_w, + ctx->picinfo.buf_h); + err = mtk_vcodec_mem_alloc(ctx, &msg_queue->wdma_addr); + if (err) { + mtk_v4l2_vdec_err(ctx, "failed to allocate wdma_addr buf"); + msg_queue->wdma_addr.size = 0; + return -ENOMEM; + } + msg_queue->wdma_rptr_addr = msg_queue->wdma_addr.dma_addr; + msg_queue->wdma_wptr_addr = msg_queue->wdma_addr.dma_addr; + + msg_queue->empty_lat_buf.ctx = ctx; + msg_queue->empty_lat_buf.core_decode = NULL; + msg_queue->empty_lat_buf.is_last_frame = true; + + msg_queue->ctx = ctx; + for (i = 0; i < NUM_BUFFER_COUNT; i++) { + lat_buf = &msg_queue->lat_buf[i]; + + lat_buf->wdma_err_addr.size = VDEC_ERR_MAP_SZ_AVC; + err = mtk_vcodec_mem_alloc(ctx, &lat_buf->wdma_err_addr); + if (err) { + mtk_v4l2_vdec_err(ctx, "failed to allocate wdma_err_addr buf[%d]", i); + goto mem_alloc_err; + } + + lat_buf->slice_bc_addr.size = VDEC_LAT_SLICE_HEADER_SZ; + err = mtk_vcodec_mem_alloc(ctx, &lat_buf->slice_bc_addr); + if (err) { + mtk_v4l2_vdec_err(ctx, "failed to allocate wdma_addr buf[%d]", i); + goto mem_alloc_err; + } + + if (ctx->current_codec == V4L2_PIX_FMT_AV1_FRAME) { + lat_buf->rd_mv_addr.size = VDEC_RD_MV_BUFFER_SZ; + err = mtk_vcodec_mem_alloc(ctx, &lat_buf->rd_mv_addr); + if (err) { + mtk_v4l2_vdec_err(ctx, "failed to allocate rd_mv_addr buf[%d]", i); + goto mem_alloc_err; + } + + lat_buf->tile_addr.size = VDEC_LAT_TILE_SZ; + err = mtk_vcodec_mem_alloc(ctx, &lat_buf->tile_addr); + if (err) { + mtk_v4l2_vdec_err(ctx, "failed to allocate tile_addr buf[%d]", i); + goto mem_alloc_err; + } + } + + lat_buf->private_data = kzalloc(private_size, GFP_KERNEL); + if (!lat_buf->private_data) { + err = -ENOMEM; + goto mem_alloc_err; + } + + lat_buf->ctx = ctx; + lat_buf->core_decode = core_decode; + lat_buf->is_last_frame = false; + err = vdec_msg_queue_qbuf(&msg_queue->lat_ctx, lat_buf); + if (err) { + mtk_v4l2_vdec_err(ctx, "failed to qbuf buf[%d]", i); + goto mem_alloc_err; + } + } + return 0; + +mem_alloc_err: + vdec_msg_queue_deinit(msg_queue, ctx); + return err; +} diff --git a/drivers/media/platform/mediatek/vcodec/decoder/vdec_msg_queue.h b/drivers/media/platform/mediatek/vcodec/decoder/vdec_msg_queue.h new file mode 100644 index 0000000000..1d9beb9e4a --- /dev/null +++ b/drivers/media/platform/mediatek/vcodec/decoder/vdec_msg_queue.h @@ -0,0 +1,191 @@ +/* SPDX-License-Identifier: GPL-2.0 */ +/* + * Copyright (c) 2021 MediaTek Inc. + * Author: Yunfei Dong <yunfei.dong@mediatek.com> + */ + +#ifndef _VDEC_MSG_QUEUE_H_ +#define _VDEC_MSG_QUEUE_H_ + +#include <linux/sched.h> +#include <linux/semaphore.h> +#include <linux/slab.h> +#include <media/videobuf2-v4l2.h> + +#define NUM_BUFFER_COUNT 3 + +struct vdec_lat_buf; +struct mtk_vcodec_dec_ctx; +struct mtk_vcodec_dec_dev; +typedef int (*core_decode_cb_t)(struct vdec_lat_buf *lat_buf); + +/** + * enum core_ctx_status - Context decode status for core hardwre. + * @CONTEXT_LIST_EMPTY: No buffer queued on core hardware(must always be 0) + * @CONTEXT_LIST_QUEUED: Buffer queued to core work list + * @CONTEXT_LIST_DEC_DONE: context decode done + */ +enum core_ctx_status { + CONTEXT_LIST_EMPTY = 0, + CONTEXT_LIST_QUEUED, + CONTEXT_LIST_DEC_DONE, +}; + +/** + * struct vdec_msg_queue_ctx - represents a queue for buffers ready to be processed + * @ready_to_use: ready used queue used to signalize when get a job queue + * @ready_queue: list of ready lat buffer queues + * @ready_lock: spin lock to protect the lat buffer usage + * @ready_num: number of buffers ready to be processed + * @hardware_index: hardware id that this queue is used for + */ +struct vdec_msg_queue_ctx { + wait_queue_head_t ready_to_use; + struct list_head ready_queue; + /* protect lat buffer */ + spinlock_t ready_lock; + int ready_num; + int hardware_index; +}; + +/** + * struct vdec_lat_buf - lat buffer message used to store lat info for core decode + * @wdma_err_addr: wdma error address used for lat hardware + * @slice_bc_addr: slice bc address used for lat hardware + * @rd_mv_addr: mv addr for av1 lat hardware output, core hardware input + * @tile_addr: tile buffer for av1 core input + * @ts_info: need to set timestamp from output to capture + * @src_buf_req: output buffer media request object + * + * @private_data: shared information used to lat and core hardware + * @ctx: mtk vcodec context information + * @core_decode: different codec use different decode callback function + * @lat_list: add lat buffer to lat head list + * @core_list: add lat buffer to core head list + * + * @is_last_frame: meaning this buffer is the last frame + */ +struct vdec_lat_buf { + struct mtk_vcodec_mem wdma_err_addr; + struct mtk_vcodec_mem slice_bc_addr; + struct mtk_vcodec_mem rd_mv_addr; + struct mtk_vcodec_mem tile_addr; + struct vb2_v4l2_buffer ts_info; + struct media_request *src_buf_req; + + void *private_data; + struct mtk_vcodec_dec_ctx *ctx; + core_decode_cb_t core_decode; + struct list_head lat_list; + struct list_head core_list; + + bool is_last_frame; +}; + +/** + * struct vdec_msg_queue - used to store lat buffer message + * @lat_buf: lat buffer used to store lat buffer information + * @wdma_addr: wdma address used for ube + * @wdma_rptr_addr: ube read point + * @wdma_wptr_addr: ube write point + * @core_work: core hardware work + * @lat_ctx: used to store lat buffer list + * @core_ctx: used to store core buffer list + * + * @lat_list_cnt: used to record each instance lat list count + * @core_list_cnt: used to record each instance core list count + * @flush_done: core flush done status + * @empty_lat_buf: the last lat buf used to flush decode + * @core_dec_done: core work queue decode done event + * @status: current context decode status for core hardware + * @ctx: mtk vcodec context information + */ +struct vdec_msg_queue { + struct vdec_lat_buf lat_buf[NUM_BUFFER_COUNT]; + + struct mtk_vcodec_mem wdma_addr; + u64 wdma_rptr_addr; + u64 wdma_wptr_addr; + + struct work_struct core_work; + struct vdec_msg_queue_ctx lat_ctx; + struct vdec_msg_queue_ctx core_ctx; + + atomic_t lat_list_cnt; + atomic_t core_list_cnt; + bool flush_done; + struct vdec_lat_buf empty_lat_buf; + wait_queue_head_t core_dec_done; + int status; + struct mtk_vcodec_dec_ctx *ctx; +}; + +/** + * vdec_msg_queue_init - init lat buffer information. + * @msg_queue: used to store the lat buffer information + * @ctx: v4l2 ctx + * @core_decode: core decode callback for each codec + * @private_size: the private data size used to share with core + * + * Return: returns 0 if init successfully, or fail. + */ +int vdec_msg_queue_init(struct vdec_msg_queue *msg_queue, + struct mtk_vcodec_dec_ctx *ctx, core_decode_cb_t core_decode, + int private_size); + +/** + * vdec_msg_queue_init_ctx - used to init msg queue context information. + * @ctx: message queue context + * @hardware_index: hardware index + */ +void vdec_msg_queue_init_ctx(struct vdec_msg_queue_ctx *ctx, int hardware_index); + +/** + * vdec_msg_queue_qbuf - enqueue lat buffer to queue list. + * @ctx: message queue context + * @buf: current lat buffer + * + * Return: returns 0 if qbuf successfully, or fail. + */ +int vdec_msg_queue_qbuf(struct vdec_msg_queue_ctx *ctx, struct vdec_lat_buf *buf); + +/** + * vdec_msg_queue_dqbuf - dequeue lat buffer from queue list. + * @ctx: message queue context + * + * Return: returns not null if dq successfully, or fail. + */ +struct vdec_lat_buf *vdec_msg_queue_dqbuf(struct vdec_msg_queue_ctx *ctx); + +/** + * vdec_msg_queue_update_ube_rptr - used to updata the ube read point. + * @msg_queue: used to store the lat buffer information + * @ube_rptr: current ube read point + */ +void vdec_msg_queue_update_ube_rptr(struct vdec_msg_queue *msg_queue, uint64_t ube_rptr); + +/** + * vdec_msg_queue_update_ube_wptr - used to updata the ube write point. + * @msg_queue: used to store the lat buffer information + * @ube_wptr: current ube write point + */ +void vdec_msg_queue_update_ube_wptr(struct vdec_msg_queue *msg_queue, uint64_t ube_wptr); + +/** + * vdec_msg_queue_wait_lat_buf_full - used to check whether all lat buffer + * in lat list. + * @msg_queue: used to store the lat buffer information + * + * Return: returns true if successfully, or fail. + */ +bool vdec_msg_queue_wait_lat_buf_full(struct vdec_msg_queue *msg_queue); + +/** + * vdec_msg_queue_deinit - deinit lat buffer information. + * @msg_queue: used to store the lat buffer information + * @ctx: v4l2 ctx + */ +void vdec_msg_queue_deinit(struct vdec_msg_queue *msg_queue, + struct mtk_vcodec_dec_ctx *ctx); + +#endif diff --git a/drivers/media/platform/mediatek/vcodec/decoder/vdec_vpu_if.c b/drivers/media/platform/mediatek/vcodec/decoder/vdec_vpu_if.c new file mode 100644 index 0000000000..82e57ae983 --- /dev/null +++ b/drivers/media/platform/mediatek/vcodec/decoder/vdec_vpu_if.c @@ -0,0 +1,311 @@ +// SPDX-License-Identifier: GPL-2.0 +/* + * Copyright (c) 2016 MediaTek Inc. + * Author: PC Chen <pc.chen@mediatek.com> + */ + +#include "mtk_vcodec_dec_drv.h" +#include "vdec_drv_if.h" +#include "vdec_ipi_msg.h" +#include "vdec_vpu_if.h" + +static void handle_init_ack_msg(const struct vdec_vpu_ipi_init_ack *msg) +{ + struct vdec_vpu_inst *vpu = (struct vdec_vpu_inst *) + (unsigned long)msg->ap_inst_addr; + + mtk_vdec_debug(vpu->ctx, "+ ap_inst_addr = 0x%llx", msg->ap_inst_addr); + + /* mapping VPU address to kernel virtual address */ + /* the content in vsi is initialized to 0 in VPU */ + vpu->vsi = mtk_vcodec_fw_map_dm_addr(vpu->ctx->dev->fw_handler, + msg->vpu_inst_addr); + vpu->inst_addr = msg->vpu_inst_addr; + + mtk_vdec_debug(vpu->ctx, "- vpu_inst_addr = 0x%x", vpu->inst_addr); + + /* Set default ABI version if dealing with unversioned firmware. */ + vpu->fw_abi_version = 0; + /* + * Instance ID is only used if ABI version >= 2. Initialize it with + * garbage by default. + */ + vpu->inst_id = 0xdeadbeef; + + /* VPU firmware does not contain a version field. */ + if (mtk_vcodec_fw_get_type(vpu->ctx->dev->fw_handler) == VPU) + return; + + /* Check firmware version. */ + vpu->fw_abi_version = msg->vdec_abi_version; + mtk_vdec_debug(vpu->ctx, "firmware version 0x%x\n", vpu->fw_abi_version); + switch (vpu->fw_abi_version) { + case 1: + break; + case 2: + vpu->inst_id = msg->inst_id; + break; + default: + mtk_vdec_err(vpu->ctx, "unhandled firmware version 0x%x\n", vpu->fw_abi_version); + vpu->failure = 1; + break; + } +} + +static void handle_get_param_msg_ack(const struct vdec_vpu_ipi_get_param_ack *msg) +{ + struct vdec_vpu_inst *vpu = (struct vdec_vpu_inst *) + (unsigned long)msg->ap_inst_addr; + + mtk_vdec_debug(vpu->ctx, "+ ap_inst_addr = 0x%llx", msg->ap_inst_addr); + + /* param_type is enum vdec_get_param_type */ + switch (msg->param_type) { + case GET_PARAM_PIC_INFO: + vpu->fb_sz[0] = msg->data[0]; + vpu->fb_sz[1] = msg->data[1]; + break; + default: + mtk_vdec_err(vpu->ctx, "invalid get param type=%d", msg->param_type); + vpu->failure = 1; + break; + } +} + +static bool vpu_dec_check_ap_inst(struct mtk_vcodec_dec_dev *dec_dev, struct vdec_vpu_inst *vpu) +{ + struct mtk_vcodec_dec_ctx *ctx; + int ret = false; + + list_for_each_entry(ctx, &dec_dev->ctx_list, list) { + if (!IS_ERR_OR_NULL(ctx) && ctx->vpu_inst == vpu) { + ret = true; + break; + } + } + + return ret; +} + +/* + * vpu_dec_ipi_handler - Handler for VPU ipi message. + * + * @data: ipi message + * @len : length of ipi message + * @priv: callback private data which is passed by decoder when register. + * + * This function runs in interrupt context and it means there's an IPI MSG + * from VPU. + */ +static void vpu_dec_ipi_handler(void *data, unsigned int len, void *priv) +{ + struct mtk_vcodec_dec_dev *dec_dev; + const struct vdec_vpu_ipi_ack *msg = data; + struct vdec_vpu_inst *vpu; + + dec_dev = (struct mtk_vcodec_dec_dev *)priv; + vpu = (struct vdec_vpu_inst *)(unsigned long)msg->ap_inst_addr; + if (!priv || !vpu) { + pr_err(MTK_DBG_V4L2_STR "ap_inst_addr is NULL, did the SCP hang or crash?"); + return; + } + + if (!vpu_dec_check_ap_inst(dec_dev, vpu) || msg->msg_id < VPU_IPIMSG_DEC_INIT_ACK || + msg->msg_id > VPU_IPIMSG_DEC_GET_PARAM_ACK) { + mtk_v4l2_vdec_err(vpu->ctx, "vdec msg id not correctly => 0x%x", msg->msg_id); + vpu->failure = -EINVAL; + goto error; + } + + vpu->failure = msg->status; + if (msg->status != 0) + goto error; + + switch (msg->msg_id) { + case VPU_IPIMSG_DEC_INIT_ACK: + handle_init_ack_msg(data); + break; + + case VPU_IPIMSG_DEC_START_ACK: + case VPU_IPIMSG_DEC_END_ACK: + case VPU_IPIMSG_DEC_DEINIT_ACK: + case VPU_IPIMSG_DEC_RESET_ACK: + case VPU_IPIMSG_DEC_CORE_ACK: + case VPU_IPIMSG_DEC_CORE_END_ACK: + break; + + case VPU_IPIMSG_DEC_GET_PARAM_ACK: + handle_get_param_msg_ack(data); + break; + default: + mtk_vdec_err(vpu->ctx, "invalid msg=%X", msg->msg_id); + break; + } + +error: + vpu->signaled = 1; +} + +static int vcodec_vpu_send_msg(struct vdec_vpu_inst *vpu, void *msg, int len) +{ + int err, id, msgid; + + msgid = *(uint32_t *)msg; + mtk_vdec_debug(vpu->ctx, "id=%X", msgid); + + vpu->failure = 0; + vpu->signaled = 0; + + if (vpu->ctx->dev->vdec_pdata->hw_arch == MTK_VDEC_LAT_SINGLE_CORE) { + if (msgid == AP_IPIMSG_DEC_CORE || + msgid == AP_IPIMSG_DEC_CORE_END) + id = vpu->core_id; + else + id = vpu->id; + } else { + id = vpu->id; + } + + err = mtk_vcodec_fw_ipi_send(vpu->ctx->dev->fw_handler, id, msg, + len, 2000); + if (err) { + mtk_vdec_err(vpu->ctx, "send fail vpu_id=%d msg_id=%X status=%d", + id, msgid, err); + return err; + } + + return vpu->failure; +} + +static int vcodec_send_ap_ipi(struct vdec_vpu_inst *vpu, unsigned int msg_id) +{ + struct vdec_ap_ipi_cmd msg; + int err = 0; + + mtk_vdec_debug(vpu->ctx, "+ id=%X", msg_id); + + memset(&msg, 0, sizeof(msg)); + msg.msg_id = msg_id; + if (vpu->fw_abi_version < 2) + msg.vpu_inst_addr = vpu->inst_addr; + else + msg.inst_id = vpu->inst_id; + msg.codec_type = vpu->codec_type; + + err = vcodec_vpu_send_msg(vpu, &msg, sizeof(msg)); + mtk_vdec_debug(vpu->ctx, "- id=%X ret=%d", msg_id, err); + return err; +} + +int vpu_dec_init(struct vdec_vpu_inst *vpu) +{ + struct vdec_ap_ipi_init msg; + int err; + + init_waitqueue_head(&vpu->wq); + vpu->handler = vpu_dec_ipi_handler; + vpu->ctx->vpu_inst = vpu; + + err = mtk_vcodec_fw_ipi_register(vpu->ctx->dev->fw_handler, vpu->id, + vpu->handler, "vdec", vpu->ctx->dev); + if (err) { + mtk_vdec_err(vpu->ctx, "vpu_ipi_register fail status=%d", err); + return err; + } + + if (vpu->ctx->dev->vdec_pdata->hw_arch == MTK_VDEC_LAT_SINGLE_CORE) { + err = mtk_vcodec_fw_ipi_register(vpu->ctx->dev->fw_handler, + vpu->core_id, vpu->handler, + "vdec", vpu->ctx->dev); + if (err) { + mtk_vdec_err(vpu->ctx, "vpu_ipi_register core fail status=%d", err); + return err; + } + } + + memset(&msg, 0, sizeof(msg)); + msg.msg_id = AP_IPIMSG_DEC_INIT; + msg.ap_inst_addr = (unsigned long)vpu; + msg.codec_type = vpu->codec_type; + + mtk_vdec_debug(vpu->ctx, "vdec_inst=%p", vpu); + + err = vcodec_vpu_send_msg(vpu, (void *)&msg, sizeof(msg)); + mtk_vdec_debug(vpu->ctx, "- ret=%d", err); + return err; +} + +int vpu_dec_start(struct vdec_vpu_inst *vpu, uint32_t *data, unsigned int len) +{ + struct vdec_ap_ipi_dec_start msg; + int i; + int err = 0; + + if (len > ARRAY_SIZE(msg.data)) { + mtk_vdec_err(vpu->ctx, "invalid len = %d\n", len); + return -EINVAL; + } + + memset(&msg, 0, sizeof(msg)); + msg.msg_id = AP_IPIMSG_DEC_START; + if (vpu->fw_abi_version < 2) + msg.vpu_inst_addr = vpu->inst_addr; + else + msg.inst_id = vpu->inst_id; + + for (i = 0; i < len; i++) + msg.data[i] = data[i]; + msg.codec_type = vpu->codec_type; + + err = vcodec_vpu_send_msg(vpu, (void *)&msg, sizeof(msg)); + mtk_vdec_debug(vpu->ctx, "- ret=%d", err); + return err; +} + +int vpu_dec_get_param(struct vdec_vpu_inst *vpu, uint32_t *data, + unsigned int len, unsigned int param_type) +{ + struct vdec_ap_ipi_get_param msg; + int err; + + if (len > ARRAY_SIZE(msg.data)) { + mtk_vdec_err(vpu->ctx, "invalid len = %d\n", len); + return -EINVAL; + } + + memset(&msg, 0, sizeof(msg)); + msg.msg_id = AP_IPIMSG_DEC_GET_PARAM; + msg.inst_id = vpu->inst_id; + memcpy(msg.data, data, sizeof(unsigned int) * len); + msg.param_type = param_type; + msg.codec_type = vpu->codec_type; + + err = vcodec_vpu_send_msg(vpu, (void *)&msg, sizeof(msg)); + mtk_vdec_debug(vpu->ctx, "- ret=%d", err); + return err; +} + +int vpu_dec_core(struct vdec_vpu_inst *vpu) +{ + return vcodec_send_ap_ipi(vpu, AP_IPIMSG_DEC_CORE); +} + +int vpu_dec_end(struct vdec_vpu_inst *vpu) +{ + return vcodec_send_ap_ipi(vpu, AP_IPIMSG_DEC_END); +} + +int vpu_dec_core_end(struct vdec_vpu_inst *vpu) +{ + return vcodec_send_ap_ipi(vpu, AP_IPIMSG_DEC_CORE_END); +} + +int vpu_dec_deinit(struct vdec_vpu_inst *vpu) +{ + return vcodec_send_ap_ipi(vpu, AP_IPIMSG_DEC_DEINIT); +} + +int vpu_dec_reset(struct vdec_vpu_inst *vpu) +{ + return vcodec_send_ap_ipi(vpu, AP_IPIMSG_DEC_RESET); +} diff --git a/drivers/media/platform/mediatek/vcodec/decoder/vdec_vpu_if.h b/drivers/media/platform/mediatek/vcodec/decoder/vdec_vpu_if.h new file mode 100644 index 0000000000..fbb3f34a73 --- /dev/null +++ b/drivers/media/platform/mediatek/vcodec/decoder/vdec_vpu_if.h @@ -0,0 +1,120 @@ +/* SPDX-License-Identifier: GPL-2.0 */ +/* + * Copyright (c) 2016 MediaTek Inc. + * Author: PC Chen <pc.chen@mediatek.com> + */ + +#ifndef _VDEC_VPU_IF_H_ +#define _VDEC_VPU_IF_H_ + +struct mtk_vcodec_dec_ctx; + +/** + * struct vdec_vpu_inst - VPU instance for video codec + * @id : ipi msg id for each decoder + * @core_id : core id used to separate different hardware + * @vsi : driver structure allocated by VPU side and shared to AP side + * for control and info share + * @failure : VPU execution result status, 0: success, others: fail + * @inst_addr : VPU decoder instance address + * @fw_abi_version : ABI version of the firmware. + * @inst_id : if fw_abi_version >= 2, contains the instance ID to be given + * in place of inst_addr in messages. + * @signaled : 1 - Host has received ack message from VPU, 0 - not received + * @ctx : context for v4l2 layer integration + * @dev : platform device of VPU + * @wq : wait queue to wait VPU message ack + * @handler : ipi handler for each decoder + * @codec_type : use codec type to separate different codecs + * @capture_type: used capture type to separate different capture format + * @fb_sz : frame buffer size of each plane + */ +struct vdec_vpu_inst { + int id; + int core_id; + void *vsi; + int32_t failure; + uint32_t inst_addr; + uint32_t fw_abi_version; + uint32_t inst_id; + unsigned int signaled; + struct mtk_vcodec_dec_ctx *ctx; + wait_queue_head_t wq; + mtk_vcodec_ipi_handler handler; + unsigned int codec_type; + unsigned int capture_type; + unsigned int fb_sz[2]; +}; + +/** + * vpu_dec_init - init decoder instance and allocate required resource in VPU. + * + * @vpu: instance for vdec_vpu_inst + */ +int vpu_dec_init(struct vdec_vpu_inst *vpu); + +/** + * vpu_dec_start - start decoding, basically the function will be invoked once + * every frame. + * + * @vpu : instance for vdec_vpu_inst + * @data: meta data to pass bitstream info to VPU decoder + * @len : meta data length + */ +int vpu_dec_start(struct vdec_vpu_inst *vpu, uint32_t *data, unsigned int len); + +/** + * vpu_dec_end - end decoding, basically the function will be invoked once + * when HW decoding done interrupt received successfully. The + * decoder in VPU will continue to do reference frame management + * and check if there is a new decoded frame available to display. + * + * @vpu : instance for vdec_vpu_inst + */ +int vpu_dec_end(struct vdec_vpu_inst *vpu); + +/** + * vpu_dec_deinit - deinit decoder instance and resource freed in VPU. + * + * @vpu: instance for vdec_vpu_inst + */ +int vpu_dec_deinit(struct vdec_vpu_inst *vpu); + +/** + * vpu_dec_reset - reset decoder, use for flush decoder when end of stream or + * seek. Remainig non displayed frame will be pushed to display. + * + * @vpu: instance for vdec_vpu_inst + */ +int vpu_dec_reset(struct vdec_vpu_inst *vpu); + +/** + * vpu_dec_core - core start decoding, basically the function will be invoked once + * every frame. + * + * @vpu : instance for vdec_vpu_inst + */ +int vpu_dec_core(struct vdec_vpu_inst *vpu); + +/** + * vpu_dec_core_end - core end decoding, basically the function will be invoked once + * when core HW decoding done and receive interrupt successfully. The + * decoder in VPU will updata hardware information and deinit hardware + * and check if there is a new decoded frame available to display. + * + * @vpu : instance for vdec_vpu_inst + */ +int vpu_dec_core_end(struct vdec_vpu_inst *vpu); + +/** + * vpu_dec_get_param - get param from scp + * + * @vpu : instance for vdec_vpu_inst + * @data: meta data to pass bitstream info to VPU decoder + * @len : meta data length + * @param_type : get param type + */ +int vpu_dec_get_param(struct vdec_vpu_inst *vpu, uint32_t *data, + unsigned int len, unsigned int param_type); + +#endif diff --git a/drivers/media/platform/mediatek/vcodec/encoder/Makefile b/drivers/media/platform/mediatek/vcodec/encoder/Makefile new file mode 100644 index 0000000000..e621b5b7e5 --- /dev/null +++ b/drivers/media/platform/mediatek/vcodec/encoder/Makefile @@ -0,0 +1,11 @@ +# SPDX-License-Identifier: GPL-2.0 + +obj-$(CONFIG_VIDEO_MEDIATEK_VCODEC) += mtk-vcodec-enc.o + +mtk-vcodec-enc-y := venc/venc_vp8_if.o \ + venc/venc_h264_if.o \ + mtk_vcodec_enc.o \ + mtk_vcodec_enc_drv.o \ + mtk_vcodec_enc_pm.o \ + venc_drv_if.o \ + venc_vpu_if.o \ diff --git a/drivers/media/platform/mediatek/vcodec/encoder/mtk_vcodec_enc.c b/drivers/media/platform/mediatek/vcodec/encoder/mtk_vcodec_enc.c new file mode 100644 index 0000000000..04948d3eb0 --- /dev/null +++ b/drivers/media/platform/mediatek/vcodec/encoder/mtk_vcodec_enc.c @@ -0,0 +1,1460 @@ +// SPDX-License-Identifier: GPL-2.0 +/* +* Copyright (c) 2016 MediaTek Inc. +* Author: PC Chen <pc.chen@mediatek.com> +* Tiffany Lin <tiffany.lin@mediatek.com> +*/ + +#include <media/v4l2-event.h> +#include <media/v4l2-mem2mem.h> +#include <media/videobuf2-dma-contig.h> +#include <linux/pm_runtime.h> + +#include "mtk_vcodec_enc.h" +#include "venc_drv_if.h" + +#define MTK_VENC_MIN_W 160U +#define MTK_VENC_MIN_H 128U +#define MTK_VENC_HD_MAX_W 1920U +#define MTK_VENC_HD_MAX_H 1088U +#define MTK_VENC_4K_MAX_W 3840U +#define MTK_VENC_4K_MAX_H 2176U + +#define DFT_CFG_WIDTH MTK_VENC_MIN_W +#define DFT_CFG_HEIGHT MTK_VENC_MIN_H +#define MTK_MAX_CTRLS_HINT 20 + +#define MTK_DEFAULT_FRAMERATE_NUM 1001 +#define MTK_DEFAULT_FRAMERATE_DENOM 30000 +#define MTK_VENC_4K_CAPABILITY_ENABLE BIT(0) + +static void mtk_venc_worker(struct work_struct *work); + +static const struct v4l2_frmsize_stepwise mtk_venc_hd_framesizes = { + MTK_VENC_MIN_W, MTK_VENC_HD_MAX_W, 16, + MTK_VENC_MIN_H, MTK_VENC_HD_MAX_H, 16, +}; + +static const struct v4l2_frmsize_stepwise mtk_venc_4k_framesizes = { + MTK_VENC_MIN_W, MTK_VENC_4K_MAX_W, 16, + MTK_VENC_MIN_H, MTK_VENC_4K_MAX_H, 16, +}; + +static int vidioc_venc_s_ctrl(struct v4l2_ctrl *ctrl) +{ + struct mtk_vcodec_enc_ctx *ctx = ctrl_to_enc_ctx(ctrl); + struct mtk_enc_params *p = &ctx->enc_params; + int ret = 0; + + switch (ctrl->id) { + case V4L2_CID_MPEG_VIDEO_BITRATE_MODE: + mtk_v4l2_venc_dbg(2, ctx, "V4L2_CID_MPEG_VIDEO_BITRATE_MODE val= %d", ctrl->val); + if (ctrl->val != V4L2_MPEG_VIDEO_BITRATE_MODE_CBR) { + mtk_v4l2_venc_err(ctx, "Unsupported bitrate mode =%d", ctrl->val); + ret = -EINVAL; + } + break; + case V4L2_CID_MPEG_VIDEO_BITRATE: + mtk_v4l2_venc_dbg(2, ctx, "V4L2_CID_MPEG_VIDEO_BITRATE val = %d", ctrl->val); + p->bitrate = ctrl->val; + ctx->param_change |= MTK_ENCODE_PARAM_BITRATE; + break; + case V4L2_CID_MPEG_VIDEO_B_FRAMES: + mtk_v4l2_venc_dbg(2, ctx, "V4L2_CID_MPEG_VIDEO_B_FRAMES val = %d", ctrl->val); + p->num_b_frame = ctrl->val; + break; + case V4L2_CID_MPEG_VIDEO_FRAME_RC_ENABLE: + mtk_v4l2_venc_dbg(2, ctx, "V4L2_CID_MPEG_VIDEO_FRAME_RC_ENABLE val = %d", + ctrl->val); + p->rc_frame = ctrl->val; + break; + case V4L2_CID_MPEG_VIDEO_H264_MAX_QP: + mtk_v4l2_venc_dbg(2, ctx, "V4L2_CID_MPEG_VIDEO_H264_MAX_QP val = %d", ctrl->val); + p->h264_max_qp = ctrl->val; + break; + case V4L2_CID_MPEG_VIDEO_HEADER_MODE: + mtk_v4l2_venc_dbg(2, ctx, "V4L2_CID_MPEG_VIDEO_HEADER_MODE val = %d", ctrl->val); + p->seq_hdr_mode = ctrl->val; + break; + case V4L2_CID_MPEG_VIDEO_MB_RC_ENABLE: + mtk_v4l2_venc_dbg(2, ctx, "V4L2_CID_MPEG_VIDEO_MB_RC_ENABLE val = %d", ctrl->val); + p->rc_mb = ctrl->val; + break; + case V4L2_CID_MPEG_VIDEO_H264_PROFILE: + mtk_v4l2_venc_dbg(2, ctx, "V4L2_CID_MPEG_VIDEO_H264_PROFILE val = %d", ctrl->val); + p->h264_profile = ctrl->val; + break; + case V4L2_CID_MPEG_VIDEO_H264_LEVEL: + mtk_v4l2_venc_dbg(2, ctx, "V4L2_CID_MPEG_VIDEO_H264_LEVEL val = %d", ctrl->val); + p->h264_level = ctrl->val; + break; + case V4L2_CID_MPEG_VIDEO_H264_I_PERIOD: + mtk_v4l2_venc_dbg(2, ctx, "V4L2_CID_MPEG_VIDEO_H264_I_PERIOD val = %d", ctrl->val); + p->intra_period = ctrl->val; + ctx->param_change |= MTK_ENCODE_PARAM_INTRA_PERIOD; + break; + case V4L2_CID_MPEG_VIDEO_GOP_SIZE: + mtk_v4l2_venc_dbg(2, ctx, "V4L2_CID_MPEG_VIDEO_GOP_SIZE val = %d", ctrl->val); + p->gop_size = ctrl->val; + ctx->param_change |= MTK_ENCODE_PARAM_GOP_SIZE; + break; + case V4L2_CID_MPEG_VIDEO_VP8_PROFILE: + /* + * FIXME - what vp8 profiles are actually supported? + * The ctrl is added (with only profile 0 supported) for now. + */ + mtk_v4l2_venc_dbg(2, ctx, "V4L2_CID_MPEG_VIDEO_VP8_PROFILE val = %d", ctrl->val); + break; + case V4L2_CID_MPEG_VIDEO_FORCE_KEY_FRAME: + mtk_v4l2_venc_dbg(2, ctx, "V4L2_CID_MPEG_VIDEO_FORCE_KEY_FRAME"); + p->force_intra = 1; + ctx->param_change |= MTK_ENCODE_PARAM_FORCE_INTRA; + break; + default: + ret = -EINVAL; + break; + } + + return ret; +} + +static const struct v4l2_ctrl_ops mtk_vcodec_enc_ctrl_ops = { + .s_ctrl = vidioc_venc_s_ctrl, +}; + +static int vidioc_enum_fmt(struct v4l2_fmtdesc *f, + const struct mtk_video_fmt *formats, + size_t num_formats) +{ + if (f->index >= num_formats) + return -EINVAL; + + f->pixelformat = formats[f->index].fourcc; + + return 0; +} + +static const struct mtk_video_fmt * +mtk_venc_find_format(u32 fourcc, const struct mtk_vcodec_enc_pdata *pdata) +{ + const struct mtk_video_fmt *fmt; + unsigned int k; + + for (k = 0; k < pdata->num_capture_formats; k++) { + fmt = &pdata->capture_formats[k]; + if (fmt->fourcc == fourcc) + return fmt; + } + + for (k = 0; k < pdata->num_output_formats; k++) { + fmt = &pdata->output_formats[k]; + if (fmt->fourcc == fourcc) + return fmt; + } + + return NULL; +} + +static int vidioc_enum_framesizes(struct file *file, void *fh, + struct v4l2_frmsizeenum *fsize) +{ + const struct mtk_video_fmt *fmt; + struct mtk_vcodec_enc_ctx *ctx = fh_to_enc_ctx(fh); + + if (fsize->index != 0) + return -EINVAL; + + fmt = mtk_venc_find_format(fsize->pixel_format, + ctx->dev->venc_pdata); + if (!fmt) + return -EINVAL; + + fsize->type = V4L2_FRMSIZE_TYPE_STEPWISE; + + if (ctx->dev->enc_capability & MTK_VENC_4K_CAPABILITY_ENABLE) + fsize->stepwise = mtk_venc_4k_framesizes; + else + fsize->stepwise = mtk_venc_hd_framesizes; + + return 0; +} + +static int vidioc_enum_fmt_vid_cap(struct file *file, void *priv, + struct v4l2_fmtdesc *f) +{ + const struct mtk_vcodec_enc_pdata *pdata = + fh_to_enc_ctx(priv)->dev->venc_pdata; + + return vidioc_enum_fmt(f, pdata->capture_formats, + pdata->num_capture_formats); +} + +static int vidioc_enum_fmt_vid_out(struct file *file, void *priv, + struct v4l2_fmtdesc *f) +{ + const struct mtk_vcodec_enc_pdata *pdata = + fh_to_enc_ctx(priv)->dev->venc_pdata; + + return vidioc_enum_fmt(f, pdata->output_formats, + pdata->num_output_formats); +} + +static int mtk_vcodec_enc_get_chip_name(void *priv) +{ + struct mtk_vcodec_enc_ctx *ctx = fh_to_enc_ctx(priv); + struct device *dev = &ctx->dev->plat_dev->dev; + + if (of_device_is_compatible(dev->of_node, "mediatek,mt8173-vcodec-enc")) + return 8173; + else if (of_device_is_compatible(dev->of_node, "mediatek,mt8183-vcodec-enc")) + return 8183; + else if (of_device_is_compatible(dev->of_node, "mediatek,mt8192-vcodec-enc")) + return 8192; + else if (of_device_is_compatible(dev->of_node, "mediatek,mt8195-vcodec-enc")) + return 8195; + else if (of_device_is_compatible(dev->of_node, "mediatek,mt8188-vcodec-enc")) + return 8188; + else + return 8173; +} + +static int vidioc_venc_querycap(struct file *file, void *priv, + struct v4l2_capability *cap) +{ + struct mtk_vcodec_enc_ctx *ctx = fh_to_enc_ctx(priv); + struct device *dev = &ctx->dev->plat_dev->dev; + int platform_name = mtk_vcodec_enc_get_chip_name(priv); + + strscpy(cap->driver, dev->driver->name, sizeof(cap->driver)); + snprintf(cap->card, sizeof(cap->card), "MT%d video encoder", platform_name); + + return 0; +} + +static int vidioc_venc_s_parm(struct file *file, void *priv, + struct v4l2_streamparm *a) +{ + struct mtk_vcodec_enc_ctx *ctx = fh_to_enc_ctx(priv); + struct v4l2_fract *timeperframe = &a->parm.output.timeperframe; + + if (a->type != V4L2_BUF_TYPE_VIDEO_OUTPUT_MPLANE) + return -EINVAL; + + if (timeperframe->numerator == 0 || timeperframe->denominator == 0) { + timeperframe->numerator = MTK_DEFAULT_FRAMERATE_NUM; + timeperframe->denominator = MTK_DEFAULT_FRAMERATE_DENOM; + } + + ctx->enc_params.framerate_num = timeperframe->denominator; + ctx->enc_params.framerate_denom = timeperframe->numerator; + ctx->param_change |= MTK_ENCODE_PARAM_FRAMERATE; + + a->parm.output.capability = V4L2_CAP_TIMEPERFRAME; + + return 0; +} + +static int vidioc_venc_g_parm(struct file *file, void *priv, + struct v4l2_streamparm *a) +{ + struct mtk_vcodec_enc_ctx *ctx = fh_to_enc_ctx(priv); + + if (a->type != V4L2_BUF_TYPE_VIDEO_OUTPUT_MPLANE) + return -EINVAL; + + a->parm.output.capability = V4L2_CAP_TIMEPERFRAME; + a->parm.output.timeperframe.denominator = + ctx->enc_params.framerate_num; + a->parm.output.timeperframe.numerator = + ctx->enc_params.framerate_denom; + + return 0; +} + +static struct mtk_q_data *mtk_venc_get_q_data(struct mtk_vcodec_enc_ctx *ctx, + enum v4l2_buf_type type) +{ + if (V4L2_TYPE_IS_OUTPUT(type)) + return &ctx->q_data[MTK_Q_DATA_SRC]; + + return &ctx->q_data[MTK_Q_DATA_DST]; +} + +static void vidioc_try_fmt_cap(struct v4l2_format *f) +{ + f->fmt.pix_mp.field = V4L2_FIELD_NONE; + f->fmt.pix_mp.num_planes = 1; + f->fmt.pix_mp.plane_fmt[0].bytesperline = 0; + f->fmt.pix_mp.flags = 0; +} + +/* V4L2 specification suggests the driver corrects the format struct if any of + * the dimensions is unsupported + */ +static int vidioc_try_fmt_out(struct mtk_vcodec_enc_ctx *ctx, struct v4l2_format *f, + const struct mtk_video_fmt *fmt) +{ + struct v4l2_pix_format_mplane *pix_fmt_mp = &f->fmt.pix_mp; + int tmp_w, tmp_h; + unsigned int max_width, max_height; + + pix_fmt_mp->field = V4L2_FIELD_NONE; + + if (ctx->dev->enc_capability & MTK_VENC_4K_CAPABILITY_ENABLE) { + max_width = MTK_VENC_4K_MAX_W; + max_height = MTK_VENC_4K_MAX_H; + } else { + max_width = MTK_VENC_HD_MAX_W; + max_height = MTK_VENC_HD_MAX_H; + } + + pix_fmt_mp->height = clamp(pix_fmt_mp->height, MTK_VENC_MIN_H, max_height); + pix_fmt_mp->width = clamp(pix_fmt_mp->width, MTK_VENC_MIN_W, max_width); + + /* find next closer width align 16, heign align 32, size align + * 64 rectangle + */ + tmp_w = pix_fmt_mp->width; + tmp_h = pix_fmt_mp->height; + v4l_bound_align_image(&pix_fmt_mp->width, + MTK_VENC_MIN_W, + max_width, 4, + &pix_fmt_mp->height, + MTK_VENC_MIN_H, + max_height, 5, 6); + + if (pix_fmt_mp->width < tmp_w && (pix_fmt_mp->width + 16) <= max_width) + pix_fmt_mp->width += 16; + if (pix_fmt_mp->height < tmp_h && (pix_fmt_mp->height + 32) <= max_height) + pix_fmt_mp->height += 32; + + mtk_v4l2_venc_dbg(0, ctx, + "before resize wxh=%dx%d, after resize wxh=%dx%d, sizeimage=%d %d", + tmp_w, tmp_h, pix_fmt_mp->width, + pix_fmt_mp->height, + pix_fmt_mp->plane_fmt[0].sizeimage, + pix_fmt_mp->plane_fmt[1].sizeimage); + + pix_fmt_mp->num_planes = fmt->num_planes; + pix_fmt_mp->plane_fmt[0].sizeimage = + pix_fmt_mp->width * pix_fmt_mp->height + + ((ALIGN(pix_fmt_mp->width, 16) * 2) * 16); + pix_fmt_mp->plane_fmt[0].bytesperline = pix_fmt_mp->width; + + if (pix_fmt_mp->num_planes == 2) { + pix_fmt_mp->plane_fmt[1].sizeimage = + (pix_fmt_mp->width * pix_fmt_mp->height) / 2 + + (ALIGN(pix_fmt_mp->width, 16) * 16); + pix_fmt_mp->plane_fmt[2].sizeimage = 0; + pix_fmt_mp->plane_fmt[1].bytesperline = + pix_fmt_mp->width; + pix_fmt_mp->plane_fmt[2].bytesperline = 0; + } else if (pix_fmt_mp->num_planes == 3) { + pix_fmt_mp->plane_fmt[1].sizeimage = + pix_fmt_mp->plane_fmt[2].sizeimage = + (pix_fmt_mp->width * pix_fmt_mp->height) / 4 + + ((ALIGN(pix_fmt_mp->width, 16) / 2) * 16); + pix_fmt_mp->plane_fmt[1].bytesperline = + pix_fmt_mp->plane_fmt[2].bytesperline = + pix_fmt_mp->width / 2; + } + + pix_fmt_mp->flags = 0; + + return 0; +} + +static void mtk_venc_set_param(struct mtk_vcodec_enc_ctx *ctx, + struct venc_enc_param *param) +{ + struct mtk_q_data *q_data_src = &ctx->q_data[MTK_Q_DATA_SRC]; + struct mtk_enc_params *enc_params = &ctx->enc_params; + + switch (q_data_src->fmt->fourcc) { + case V4L2_PIX_FMT_YUV420M: + param->input_yuv_fmt = VENC_YUV_FORMAT_I420; + break; + case V4L2_PIX_FMT_YVU420M: + param->input_yuv_fmt = VENC_YUV_FORMAT_YV12; + break; + case V4L2_PIX_FMT_NV12M: + param->input_yuv_fmt = VENC_YUV_FORMAT_NV12; + break; + case V4L2_PIX_FMT_NV21M: + param->input_yuv_fmt = VENC_YUV_FORMAT_NV21; + break; + default: + mtk_v4l2_venc_err(ctx, "Unsupported fourcc =%d", q_data_src->fmt->fourcc); + break; + } + param->h264_profile = enc_params->h264_profile; + param->h264_level = enc_params->h264_level; + + /* Config visible resolution */ + param->width = q_data_src->visible_width; + param->height = q_data_src->visible_height; + /* Config coded resolution */ + param->buf_width = q_data_src->coded_width; + param->buf_height = q_data_src->coded_height; + param->frm_rate = enc_params->framerate_num / + enc_params->framerate_denom; + param->intra_period = enc_params->intra_period; + param->gop_size = enc_params->gop_size; + param->bitrate = enc_params->bitrate; + + mtk_v4l2_venc_dbg(0, ctx, + "fmt 0x%x, P/L %d/%d w/h %d/%d buf %d/%d fps/bps %d/%d gop %d i_per %d", + param->input_yuv_fmt, param->h264_profile, + param->h264_level, param->width, param->height, + param->buf_width, param->buf_height, + param->frm_rate, param->bitrate, + param->gop_size, param->intra_period); +} + +static int vidioc_venc_s_fmt_cap(struct file *file, void *priv, + struct v4l2_format *f) +{ + struct mtk_vcodec_enc_ctx *ctx = fh_to_enc_ctx(priv); + const struct mtk_vcodec_enc_pdata *pdata = ctx->dev->venc_pdata; + struct vb2_queue *vq; + struct mtk_q_data *q_data = mtk_venc_get_q_data(ctx, f->type); + int i, ret; + const struct mtk_video_fmt *fmt; + + vq = v4l2_m2m_get_vq(ctx->m2m_ctx, f->type); + if (!vq) { + mtk_v4l2_venc_err(ctx, "fail to get vq"); + return -EINVAL; + } + + if (vb2_is_busy(vq)) { + mtk_v4l2_venc_err(ctx, "queue busy"); + return -EBUSY; + } + + fmt = mtk_venc_find_format(f->fmt.pix.pixelformat, pdata); + if (!fmt) { + fmt = &ctx->dev->venc_pdata->capture_formats[0]; + f->fmt.pix.pixelformat = fmt->fourcc; + } + + q_data->fmt = fmt; + vidioc_try_fmt_cap(f); + + q_data->coded_width = f->fmt.pix_mp.width; + q_data->coded_height = f->fmt.pix_mp.height; + q_data->field = f->fmt.pix_mp.field; + + for (i = 0; i < f->fmt.pix_mp.num_planes; i++) { + struct v4l2_plane_pix_format *plane_fmt; + + plane_fmt = &f->fmt.pix_mp.plane_fmt[i]; + q_data->bytesperline[i] = plane_fmt->bytesperline; + q_data->sizeimage[i] = plane_fmt->sizeimage; + } + + if (ctx->state == MTK_STATE_FREE) { + ret = venc_if_init(ctx, q_data->fmt->fourcc); + if (ret) { + mtk_v4l2_venc_err(ctx, "venc_if_init failed=%d, codec type=%x", + ret, q_data->fmt->fourcc); + return -EBUSY; + } + ctx->state = MTK_STATE_INIT; + } + + return 0; +} + +static int vidioc_venc_s_fmt_out(struct file *file, void *priv, + struct v4l2_format *f) +{ + struct mtk_vcodec_enc_ctx *ctx = fh_to_enc_ctx(priv); + const struct mtk_vcodec_enc_pdata *pdata = ctx->dev->venc_pdata; + struct vb2_queue *vq; + struct mtk_q_data *q_data = mtk_venc_get_q_data(ctx, f->type); + int ret, i; + const struct mtk_video_fmt *fmt; + + vq = v4l2_m2m_get_vq(ctx->m2m_ctx, f->type); + if (!vq) { + mtk_v4l2_venc_err(ctx, "fail to get vq"); + return -EINVAL; + } + + if (vb2_is_busy(vq)) { + mtk_v4l2_venc_err(ctx, "queue busy"); + return -EBUSY; + } + + fmt = mtk_venc_find_format(f->fmt.pix.pixelformat, pdata); + if (!fmt) { + fmt = &ctx->dev->venc_pdata->output_formats[0]; + f->fmt.pix.pixelformat = fmt->fourcc; + } + + ret = vidioc_try_fmt_out(ctx, f, fmt); + if (ret) + return ret; + + q_data->fmt = fmt; + q_data->visible_width = f->fmt.pix_mp.width; + q_data->visible_height = f->fmt.pix_mp.height; + q_data->coded_width = f->fmt.pix_mp.width; + q_data->coded_height = f->fmt.pix_mp.height; + + q_data->field = f->fmt.pix_mp.field; + ctx->colorspace = f->fmt.pix_mp.colorspace; + ctx->ycbcr_enc = f->fmt.pix_mp.ycbcr_enc; + ctx->quantization = f->fmt.pix_mp.quantization; + ctx->xfer_func = f->fmt.pix_mp.xfer_func; + + for (i = 0; i < f->fmt.pix_mp.num_planes; i++) { + struct v4l2_plane_pix_format *plane_fmt; + + plane_fmt = &f->fmt.pix_mp.plane_fmt[i]; + q_data->bytesperline[i] = plane_fmt->bytesperline; + q_data->sizeimage[i] = plane_fmt->sizeimage; + } + + return 0; +} + +static int vidioc_venc_g_fmt(struct file *file, void *priv, + struct v4l2_format *f) +{ + struct v4l2_pix_format_mplane *pix = &f->fmt.pix_mp; + struct mtk_vcodec_enc_ctx *ctx = fh_to_enc_ctx(priv); + struct vb2_queue *vq; + struct mtk_q_data *q_data = mtk_venc_get_q_data(ctx, f->type); + int i; + + vq = v4l2_m2m_get_vq(ctx->m2m_ctx, f->type); + if (!vq) + return -EINVAL; + + + pix->width = q_data->coded_width; + pix->height = q_data->coded_height; + pix->pixelformat = q_data->fmt->fourcc; + pix->field = q_data->field; + pix->num_planes = q_data->fmt->num_planes; + for (i = 0; i < pix->num_planes; i++) { + pix->plane_fmt[i].bytesperline = q_data->bytesperline[i]; + pix->plane_fmt[i].sizeimage = q_data->sizeimage[i]; + } + + pix->flags = 0; + pix->colorspace = ctx->colorspace; + pix->ycbcr_enc = ctx->ycbcr_enc; + pix->quantization = ctx->quantization; + pix->xfer_func = ctx->xfer_func; + + return 0; +} + +static int vidioc_try_fmt_vid_cap_mplane(struct file *file, void *priv, + struct v4l2_format *f) +{ + const struct mtk_video_fmt *fmt; + struct mtk_vcodec_enc_ctx *ctx = fh_to_enc_ctx(priv); + const struct mtk_vcodec_enc_pdata *pdata = ctx->dev->venc_pdata; + + fmt = mtk_venc_find_format(f->fmt.pix.pixelformat, pdata); + if (!fmt) { + fmt = &ctx->dev->venc_pdata->capture_formats[0]; + f->fmt.pix.pixelformat = fmt->fourcc; + } + f->fmt.pix_mp.colorspace = ctx->colorspace; + f->fmt.pix_mp.ycbcr_enc = ctx->ycbcr_enc; + f->fmt.pix_mp.quantization = ctx->quantization; + f->fmt.pix_mp.xfer_func = ctx->xfer_func; + + vidioc_try_fmt_cap(f); + + return 0; +} + +static int vidioc_try_fmt_vid_out_mplane(struct file *file, void *priv, + struct v4l2_format *f) +{ + const struct mtk_video_fmt *fmt; + struct mtk_vcodec_enc_ctx *ctx = fh_to_enc_ctx(priv); + const struct mtk_vcodec_enc_pdata *pdata = ctx->dev->venc_pdata; + + fmt = mtk_venc_find_format(f->fmt.pix.pixelformat, pdata); + if (!fmt) { + fmt = &ctx->dev->venc_pdata->output_formats[0]; + f->fmt.pix.pixelformat = fmt->fourcc; + } + if (!f->fmt.pix_mp.colorspace) { + f->fmt.pix_mp.colorspace = V4L2_COLORSPACE_REC709; + f->fmt.pix_mp.ycbcr_enc = V4L2_YCBCR_ENC_DEFAULT; + f->fmt.pix_mp.quantization = V4L2_QUANTIZATION_DEFAULT; + f->fmt.pix_mp.xfer_func = V4L2_XFER_FUNC_DEFAULT; + } + + return vidioc_try_fmt_out(ctx, f, fmt); +} + +static int vidioc_venc_g_selection(struct file *file, void *priv, + struct v4l2_selection *s) +{ + struct mtk_vcodec_enc_ctx *ctx = fh_to_enc_ctx(priv); + struct mtk_q_data *q_data = mtk_venc_get_q_data(ctx, s->type); + + if (s->type != V4L2_BUF_TYPE_VIDEO_OUTPUT) + return -EINVAL; + + switch (s->target) { + case V4L2_SEL_TGT_CROP_DEFAULT: + case V4L2_SEL_TGT_CROP_BOUNDS: + s->r.top = 0; + s->r.left = 0; + s->r.width = q_data->coded_width; + s->r.height = q_data->coded_height; + break; + case V4L2_SEL_TGT_CROP: + s->r.top = 0; + s->r.left = 0; + s->r.width = q_data->visible_width; + s->r.height = q_data->visible_height; + break; + default: + return -EINVAL; + } + + return 0; +} + +static int vidioc_venc_s_selection(struct file *file, void *priv, + struct v4l2_selection *s) +{ + struct mtk_vcodec_enc_ctx *ctx = fh_to_enc_ctx(priv); + struct mtk_q_data *q_data = mtk_venc_get_q_data(ctx, s->type); + + if (s->type != V4L2_BUF_TYPE_VIDEO_OUTPUT) + return -EINVAL; + + switch (s->target) { + case V4L2_SEL_TGT_CROP: + /* Only support crop from (0,0) */ + s->r.top = 0; + s->r.left = 0; + s->r.width = min(s->r.width, q_data->coded_width); + s->r.height = min(s->r.height, q_data->coded_height); + q_data->visible_width = s->r.width; + q_data->visible_height = s->r.height; + break; + default: + return -EINVAL; + } + return 0; +} + +static int vidioc_venc_qbuf(struct file *file, void *priv, + struct v4l2_buffer *buf) +{ + struct mtk_vcodec_enc_ctx *ctx = fh_to_enc_ctx(priv); + + if (ctx->state == MTK_STATE_ABORT) { + mtk_v4l2_venc_err(ctx, "[%d] Call on QBUF after unrecoverable error", + ctx->id); + return -EIO; + } + + return v4l2_m2m_qbuf(file, ctx->m2m_ctx, buf); +} + +static int vidioc_venc_dqbuf(struct file *file, void *priv, + struct v4l2_buffer *buf) +{ + struct mtk_vcodec_enc_ctx *ctx = fh_to_enc_ctx(priv); + int ret; + + if (ctx->state == MTK_STATE_ABORT) { + mtk_v4l2_venc_err(ctx, "[%d] Call on QBUF after unrecoverable error", + ctx->id); + return -EIO; + } + + ret = v4l2_m2m_dqbuf(file, ctx->m2m_ctx, buf); + if (ret) + return ret; + + /* + * Complete flush if the user dequeued the 0-payload LAST buffer. + * We check the payload because a buffer with the LAST flag can also + * be seen during resolution changes. If we happen to be flushing at + * that time, the last buffer before the resolution changes could be + * misinterpreted for the buffer generated by the flush and terminate + * it earlier than we want. + */ + if (!V4L2_TYPE_IS_OUTPUT(buf->type) && + buf->flags & V4L2_BUF_FLAG_LAST && + buf->m.planes[0].bytesused == 0 && + ctx->is_flushing) { + /* + * Last CAPTURE buffer is dequeued, we can allow another flush + * to take place. + */ + ctx->is_flushing = false; + } + + return 0; +} + +static int vidioc_encoder_cmd(struct file *file, void *priv, + struct v4l2_encoder_cmd *cmd) +{ + struct mtk_vcodec_enc_ctx *ctx = fh_to_enc_ctx(priv); + struct vb2_queue *src_vq, *dst_vq; + int ret; + + if (ctx->state == MTK_STATE_ABORT) { + mtk_v4l2_venc_err(ctx, "[%d] Call to CMD after unrecoverable error", + ctx->id); + return -EIO; + } + + ret = v4l2_m2m_ioctl_try_encoder_cmd(file, priv, cmd); + if (ret) + return ret; + + /* Calling START or STOP is invalid if a flush is in progress */ + if (ctx->is_flushing) + return -EBUSY; + + mtk_v4l2_venc_dbg(1, ctx, "encoder cmd=%u", cmd->cmd); + + dst_vq = v4l2_m2m_get_vq(ctx->m2m_ctx, + V4L2_BUF_TYPE_VIDEO_CAPTURE_MPLANE); + switch (cmd->cmd) { + case V4L2_ENC_CMD_STOP: + src_vq = v4l2_m2m_get_vq(ctx->m2m_ctx, + V4L2_BUF_TYPE_VIDEO_OUTPUT_MPLANE); + if (!vb2_is_streaming(src_vq)) { + mtk_v4l2_venc_dbg(1, ctx, "Output stream is off. No need to flush."); + return 0; + } + if (!vb2_is_streaming(dst_vq)) { + mtk_v4l2_venc_dbg(1, ctx, "Capture stream is off. No need to flush."); + return 0; + } + ctx->is_flushing = true; + v4l2_m2m_buf_queue(ctx->m2m_ctx, &ctx->empty_flush_buf.vb); + v4l2_m2m_try_schedule(ctx->m2m_ctx); + break; + + case V4L2_ENC_CMD_START: + vb2_clear_last_buffer_dequeued(dst_vq); + break; + + default: + return -EINVAL; + } + + return 0; +} + +const struct v4l2_ioctl_ops mtk_venc_ioctl_ops = { + .vidioc_streamon = v4l2_m2m_ioctl_streamon, + .vidioc_streamoff = v4l2_m2m_ioctl_streamoff, + + .vidioc_reqbufs = v4l2_m2m_ioctl_reqbufs, + .vidioc_querybuf = v4l2_m2m_ioctl_querybuf, + .vidioc_qbuf = vidioc_venc_qbuf, + .vidioc_dqbuf = vidioc_venc_dqbuf, + + .vidioc_querycap = vidioc_venc_querycap, + .vidioc_enum_fmt_vid_cap = vidioc_enum_fmt_vid_cap, + .vidioc_enum_fmt_vid_out = vidioc_enum_fmt_vid_out, + .vidioc_enum_framesizes = vidioc_enum_framesizes, + + .vidioc_try_fmt_vid_cap_mplane = vidioc_try_fmt_vid_cap_mplane, + .vidioc_try_fmt_vid_out_mplane = vidioc_try_fmt_vid_out_mplane, + .vidioc_expbuf = v4l2_m2m_ioctl_expbuf, + .vidioc_subscribe_event = v4l2_ctrl_subscribe_event, + .vidioc_unsubscribe_event = v4l2_event_unsubscribe, + + .vidioc_s_parm = vidioc_venc_s_parm, + .vidioc_g_parm = vidioc_venc_g_parm, + .vidioc_s_fmt_vid_cap_mplane = vidioc_venc_s_fmt_cap, + .vidioc_s_fmt_vid_out_mplane = vidioc_venc_s_fmt_out, + + .vidioc_g_fmt_vid_cap_mplane = vidioc_venc_g_fmt, + .vidioc_g_fmt_vid_out_mplane = vidioc_venc_g_fmt, + + .vidioc_create_bufs = v4l2_m2m_ioctl_create_bufs, + .vidioc_prepare_buf = v4l2_m2m_ioctl_prepare_buf, + + .vidioc_g_selection = vidioc_venc_g_selection, + .vidioc_s_selection = vidioc_venc_s_selection, + + .vidioc_encoder_cmd = vidioc_encoder_cmd, + .vidioc_try_encoder_cmd = v4l2_m2m_ioctl_try_encoder_cmd, +}; + +static int vb2ops_venc_queue_setup(struct vb2_queue *vq, + unsigned int *nbuffers, + unsigned int *nplanes, + unsigned int sizes[], + struct device *alloc_devs[]) +{ + struct mtk_vcodec_enc_ctx *ctx = vb2_get_drv_priv(vq); + struct mtk_q_data *q_data = mtk_venc_get_q_data(ctx, vq->type); + unsigned int i; + + if (q_data == NULL) + return -EINVAL; + + if (*nplanes) { + if (*nplanes != q_data->fmt->num_planes) + return -EINVAL; + for (i = 0; i < *nplanes; i++) + if (sizes[i] < q_data->sizeimage[i]) + return -EINVAL; + } else { + *nplanes = q_data->fmt->num_planes; + for (i = 0; i < *nplanes; i++) + sizes[i] = q_data->sizeimage[i]; + } + + return 0; +} + +static int vb2ops_venc_buf_prepare(struct vb2_buffer *vb) +{ + struct mtk_vcodec_enc_ctx *ctx = vb2_get_drv_priv(vb->vb2_queue); + struct mtk_q_data *q_data = mtk_venc_get_q_data(ctx, vb->vb2_queue->type); + int i; + + for (i = 0; i < q_data->fmt->num_planes; i++) { + if (vb2_plane_size(vb, i) < q_data->sizeimage[i]) { + mtk_v4l2_venc_err(ctx, "data will not fit into plane %d (%lu < %d)", + i, vb2_plane_size(vb, i), q_data->sizeimage[i]); + return -EINVAL; + } + } + + return 0; +} + +static void vb2ops_venc_buf_queue(struct vb2_buffer *vb) +{ + struct mtk_vcodec_enc_ctx *ctx = vb2_get_drv_priv(vb->vb2_queue); + struct vb2_v4l2_buffer *vb2_v4l2 = + container_of(vb, struct vb2_v4l2_buffer, vb2_buf); + + struct mtk_video_enc_buf *mtk_buf = + container_of(vb2_v4l2, struct mtk_video_enc_buf, + m2m_buf.vb); + + if ((vb->vb2_queue->type == V4L2_BUF_TYPE_VIDEO_OUTPUT_MPLANE) && + (ctx->param_change != MTK_ENCODE_PARAM_NONE)) { + mtk_v4l2_venc_dbg(1, ctx, "[%d] Before id=%d encode parameter change %x", + ctx->id, vb2_v4l2->vb2_buf.index, ctx->param_change); + mtk_buf->param_change = ctx->param_change; + mtk_buf->enc_params = ctx->enc_params; + ctx->param_change = MTK_ENCODE_PARAM_NONE; + } + + v4l2_m2m_buf_queue(ctx->m2m_ctx, to_vb2_v4l2_buffer(vb)); +} + +static int vb2ops_venc_start_streaming(struct vb2_queue *q, unsigned int count) +{ + struct mtk_vcodec_enc_ctx *ctx = vb2_get_drv_priv(q); + struct venc_enc_param param; + int ret, pm_ret; + int i; + + /* Once state turn into MTK_STATE_ABORT, we need stop_streaming + * to clear it + */ + if ((ctx->state == MTK_STATE_ABORT) || (ctx->state == MTK_STATE_FREE)) { + ret = -EIO; + goto err_start_stream; + } + + /* Do the initialization when both start_streaming have been called */ + if (V4L2_TYPE_IS_OUTPUT(q->type)) { + if (!vb2_start_streaming_called(&ctx->m2m_ctx->cap_q_ctx.q)) + return 0; + } else { + if (!vb2_start_streaming_called(&ctx->m2m_ctx->out_q_ctx.q)) + return 0; + } + + ret = pm_runtime_resume_and_get(&ctx->dev->plat_dev->dev); + if (ret < 0) { + mtk_v4l2_venc_err(ctx, "pm_runtime_resume_and_get fail %d", ret); + goto err_start_stream; + } + + mtk_venc_set_param(ctx, ¶m); + ret = venc_if_set_param(ctx, VENC_SET_PARAM_ENC, ¶m); + if (ret) { + mtk_v4l2_venc_err(ctx, "venc_if_set_param failed=%d", ret); + ctx->state = MTK_STATE_ABORT; + goto err_set_param; + } + ctx->param_change = MTK_ENCODE_PARAM_NONE; + + if ((ctx->q_data[MTK_Q_DATA_DST].fmt->fourcc == V4L2_PIX_FMT_H264) && + (ctx->enc_params.seq_hdr_mode != + V4L2_MPEG_VIDEO_HEADER_MODE_SEPARATE)) { + ret = venc_if_set_param(ctx, + VENC_SET_PARAM_PREPEND_HEADER, + NULL); + if (ret) { + mtk_v4l2_venc_err(ctx, "venc_if_set_param failed=%d", ret); + ctx->state = MTK_STATE_ABORT; + goto err_set_param; + } + ctx->state = MTK_STATE_HEADER; + } + + return 0; + +err_set_param: + pm_ret = pm_runtime_put(&ctx->dev->plat_dev->dev); + if (pm_ret < 0) + mtk_v4l2_venc_err(ctx, "pm_runtime_put fail %d", pm_ret); + +err_start_stream: + for (i = 0; i < q->num_buffers; ++i) { + struct vb2_buffer *buf = vb2_get_buffer(q, i); + + /* + * FIXME: This check is not needed as only active buffers + * can be marked as done. + */ + if (buf && buf->state == VB2_BUF_STATE_ACTIVE) { + mtk_v4l2_venc_dbg(0, ctx, "[%d] id=%d, type=%d, %d->VB2_BUF_STATE_QUEUED", + ctx->id, i, q->type, (int)buf->state); + v4l2_m2m_buf_done(to_vb2_v4l2_buffer(buf), + VB2_BUF_STATE_QUEUED); + } + } + + return ret; +} + +static void vb2ops_venc_stop_streaming(struct vb2_queue *q) +{ + struct mtk_vcodec_enc_ctx *ctx = vb2_get_drv_priv(q); + struct vb2_v4l2_buffer *src_buf, *dst_buf; + int ret; + + mtk_v4l2_venc_dbg(2, ctx, "[%d]-> type=%d", ctx->id, q->type); + + if (q->type == V4L2_BUF_TYPE_VIDEO_CAPTURE_MPLANE) { + while ((dst_buf = v4l2_m2m_dst_buf_remove(ctx->m2m_ctx))) { + vb2_set_plane_payload(&dst_buf->vb2_buf, 0, 0); + v4l2_m2m_buf_done(dst_buf, VB2_BUF_STATE_ERROR); + } + /* STREAMOFF on the CAPTURE queue completes any ongoing flush */ + if (ctx->is_flushing) { + struct v4l2_m2m_buffer *b, *n; + + mtk_v4l2_venc_dbg(1, ctx, "STREAMOFF called while flushing"); + /* + * STREAMOFF could be called before the flush buffer is + * dequeued. Check whether empty flush buf is still in + * queue before removing it. + */ + v4l2_m2m_for_each_src_buf_safe(ctx->m2m_ctx, b, n) { + if (b == &ctx->empty_flush_buf) { + v4l2_m2m_src_buf_remove_by_buf(ctx->m2m_ctx, &b->vb); + break; + } + } + ctx->is_flushing = false; + } + } else { + while ((src_buf = v4l2_m2m_src_buf_remove(ctx->m2m_ctx))) { + if (src_buf != &ctx->empty_flush_buf.vb) + v4l2_m2m_buf_done(src_buf, VB2_BUF_STATE_ERROR); + } + if (ctx->is_flushing) { + /* + * If we are in the middle of a flush, put the flush + * buffer back into the queue so the next CAPTURE + * buffer gets returned with the LAST flag set. + */ + v4l2_m2m_buf_queue(ctx->m2m_ctx, + &ctx->empty_flush_buf.vb); + } + } + + if ((q->type == V4L2_BUF_TYPE_VIDEO_CAPTURE_MPLANE && + vb2_is_streaming(&ctx->m2m_ctx->out_q_ctx.q)) || + (q->type == V4L2_BUF_TYPE_VIDEO_OUTPUT_MPLANE && + vb2_is_streaming(&ctx->m2m_ctx->cap_q_ctx.q))) { + mtk_v4l2_venc_dbg(1, ctx, "[%d]-> q type %d out=%d cap=%d", + ctx->id, q->type, + vb2_is_streaming(&ctx->m2m_ctx->out_q_ctx.q), + vb2_is_streaming(&ctx->m2m_ctx->cap_q_ctx.q)); + return; + } + + /* Release the encoder if both streams are stopped. */ + ret = venc_if_deinit(ctx); + if (ret) + mtk_v4l2_venc_err(ctx, "venc_if_deinit failed=%d", ret); + + ret = pm_runtime_put(&ctx->dev->plat_dev->dev); + if (ret < 0) + mtk_v4l2_venc_err(ctx, "pm_runtime_put fail %d", ret); + + ctx->state = MTK_STATE_FREE; +} + +static int vb2ops_venc_buf_out_validate(struct vb2_buffer *vb) +{ + struct vb2_v4l2_buffer *vbuf = to_vb2_v4l2_buffer(vb); + + vbuf->field = V4L2_FIELD_NONE; + return 0; +} + +static const struct vb2_ops mtk_venc_vb2_ops = { + .queue_setup = vb2ops_venc_queue_setup, + .buf_out_validate = vb2ops_venc_buf_out_validate, + .buf_prepare = vb2ops_venc_buf_prepare, + .buf_queue = vb2ops_venc_buf_queue, + .wait_prepare = vb2_ops_wait_prepare, + .wait_finish = vb2_ops_wait_finish, + .start_streaming = vb2ops_venc_start_streaming, + .stop_streaming = vb2ops_venc_stop_streaming, +}; + +static int mtk_venc_encode_header(void *priv) +{ + struct mtk_vcodec_enc_ctx *ctx = priv; + int ret; + struct vb2_v4l2_buffer *src_buf, *dst_buf; + struct mtk_vcodec_mem bs_buf; + struct venc_done_result enc_result; + + dst_buf = v4l2_m2m_dst_buf_remove(ctx->m2m_ctx); + if (!dst_buf) { + mtk_v4l2_venc_dbg(1, ctx, "No dst buffer"); + return -EINVAL; + } + + bs_buf.va = vb2_plane_vaddr(&dst_buf->vb2_buf, 0); + bs_buf.dma_addr = vb2_dma_contig_plane_dma_addr(&dst_buf->vb2_buf, 0); + bs_buf.size = (size_t)dst_buf->vb2_buf.planes[0].length; + + mtk_v4l2_venc_dbg(1, ctx, + "[%d] buf id=%d va=0x%p dma_addr=0x%llx size=%zu", + ctx->id, dst_buf->vb2_buf.index, bs_buf.va, + (u64)bs_buf.dma_addr, bs_buf.size); + + ret = venc_if_encode(ctx, + VENC_START_OPT_ENCODE_SEQUENCE_HEADER, + NULL, &bs_buf, &enc_result); + + if (ret) { + vb2_set_plane_payload(&dst_buf->vb2_buf, 0, 0); + ctx->state = MTK_STATE_ABORT; + v4l2_m2m_buf_done(dst_buf, VB2_BUF_STATE_ERROR); + mtk_v4l2_venc_err(ctx, "venc_if_encode failed=%d", ret); + return -EINVAL; + } + src_buf = v4l2_m2m_next_src_buf(ctx->m2m_ctx); + if (src_buf) { + dst_buf->vb2_buf.timestamp = src_buf->vb2_buf.timestamp; + dst_buf->timecode = src_buf->timecode; + } else { + mtk_v4l2_venc_err(ctx, "No timestamp for the header buffer."); + } + + ctx->state = MTK_STATE_HEADER; + vb2_set_plane_payload(&dst_buf->vb2_buf, 0, enc_result.bs_size); + v4l2_m2m_buf_done(dst_buf, VB2_BUF_STATE_DONE); + + return 0; +} + +static int mtk_venc_param_change(struct mtk_vcodec_enc_ctx *ctx) +{ + struct venc_enc_param enc_prm; + struct vb2_v4l2_buffer *vb2_v4l2 = v4l2_m2m_next_src_buf(ctx->m2m_ctx); + struct mtk_video_enc_buf *mtk_buf; + int ret = 0; + + /* Don't upcast the empty flush buffer */ + if (vb2_v4l2 == &ctx->empty_flush_buf.vb) + return 0; + + mtk_buf = container_of(vb2_v4l2, struct mtk_video_enc_buf, m2m_buf.vb); + + memset(&enc_prm, 0, sizeof(enc_prm)); + if (mtk_buf->param_change == MTK_ENCODE_PARAM_NONE) + return 0; + + if (mtk_buf->param_change & MTK_ENCODE_PARAM_BITRATE) { + enc_prm.bitrate = mtk_buf->enc_params.bitrate; + mtk_v4l2_venc_dbg(1, ctx, "[%d] id=%d, change param br=%d", + ctx->id, vb2_v4l2->vb2_buf.index, enc_prm.bitrate); + ret |= venc_if_set_param(ctx, + VENC_SET_PARAM_ADJUST_BITRATE, + &enc_prm); + } + if (!ret && mtk_buf->param_change & MTK_ENCODE_PARAM_FRAMERATE) { + enc_prm.frm_rate = mtk_buf->enc_params.framerate_num / + mtk_buf->enc_params.framerate_denom; + mtk_v4l2_venc_dbg(1, ctx, "[%d] id=%d, change param fr=%d", + ctx->id, vb2_v4l2->vb2_buf.index, enc_prm.frm_rate); + ret |= venc_if_set_param(ctx, + VENC_SET_PARAM_ADJUST_FRAMERATE, + &enc_prm); + } + if (!ret && mtk_buf->param_change & MTK_ENCODE_PARAM_GOP_SIZE) { + enc_prm.gop_size = mtk_buf->enc_params.gop_size; + mtk_v4l2_venc_dbg(1, ctx, "change param intra period=%d", enc_prm.gop_size); + ret |= venc_if_set_param(ctx, + VENC_SET_PARAM_GOP_SIZE, + &enc_prm); + } + if (!ret && mtk_buf->param_change & MTK_ENCODE_PARAM_FORCE_INTRA) { + mtk_v4l2_venc_dbg(1, ctx, "[%d] id=%d, change param force I=%d", + ctx->id, vb2_v4l2->vb2_buf.index, + mtk_buf->enc_params.force_intra); + if (mtk_buf->enc_params.force_intra) + ret |= venc_if_set_param(ctx, + VENC_SET_PARAM_FORCE_INTRA, + NULL); + } + + mtk_buf->param_change = MTK_ENCODE_PARAM_NONE; + + if (ret) { + ctx->state = MTK_STATE_ABORT; + mtk_v4l2_venc_err(ctx, "venc_if_set_param %d failed=%d", + mtk_buf->param_change, ret); + return -1; + } + + return 0; +} + +/* + * v4l2_m2m_streamoff() holds dev_mutex and waits mtk_venc_worker() + * to call v4l2_m2m_job_finish(). + * If mtk_venc_worker() tries to acquire dev_mutex, it will deadlock. + * So this function must not try to acquire dev->dev_mutex. + * This means v4l2 ioctls and mtk_venc_worker() can run at the same time. + * mtk_venc_worker() should be carefully implemented to avoid bugs. + */ +static void mtk_venc_worker(struct work_struct *work) +{ + struct mtk_vcodec_enc_ctx *ctx = container_of(work, struct mtk_vcodec_enc_ctx, + encode_work); + struct vb2_v4l2_buffer *src_buf, *dst_buf; + struct venc_frm_buf frm_buf; + struct mtk_vcodec_mem bs_buf; + struct venc_done_result enc_result; + int ret, i; + + /* check dst_buf, dst_buf may be removed in device_run + * to stored encdoe header so we need check dst_buf and + * call job_finish here to prevent recursion + */ + dst_buf = v4l2_m2m_dst_buf_remove(ctx->m2m_ctx); + if (!dst_buf) { + v4l2_m2m_job_finish(ctx->dev->m2m_dev_enc, ctx->m2m_ctx); + return; + } + + src_buf = v4l2_m2m_src_buf_remove(ctx->m2m_ctx); + + /* + * If we see the flush buffer, send an empty buffer with the LAST flag + * to the client. is_flushing will be reset at the time the buffer + * is dequeued. + */ + if (src_buf == &ctx->empty_flush_buf.vb) { + vb2_set_plane_payload(&dst_buf->vb2_buf, 0, 0); + dst_buf->flags |= V4L2_BUF_FLAG_LAST; + v4l2_m2m_buf_done(dst_buf, VB2_BUF_STATE_DONE); + v4l2_m2m_job_finish(ctx->dev->m2m_dev_enc, ctx->m2m_ctx); + return; + } + + memset(&frm_buf, 0, sizeof(frm_buf)); + for (i = 0; i < src_buf->vb2_buf.num_planes ; i++) { + frm_buf.fb_addr[i].dma_addr = + vb2_dma_contig_plane_dma_addr(&src_buf->vb2_buf, i); + frm_buf.fb_addr[i].size = + (size_t)src_buf->vb2_buf.planes[i].length; + } + bs_buf.va = vb2_plane_vaddr(&dst_buf->vb2_buf, 0); + bs_buf.dma_addr = vb2_dma_contig_plane_dma_addr(&dst_buf->vb2_buf, 0); + bs_buf.size = (size_t)dst_buf->vb2_buf.planes[0].length; + + mtk_v4l2_venc_dbg(2, ctx, + "Framebuf PA=%llx Size=0x%zx;PA=0x%llx Size=0x%zx;PA=0x%llx Size=%zu", + (u64)frm_buf.fb_addr[0].dma_addr, frm_buf.fb_addr[0].size, + (u64)frm_buf.fb_addr[1].dma_addr, frm_buf.fb_addr[1].size, + (u64)frm_buf.fb_addr[2].dma_addr, frm_buf.fb_addr[2].size); + + ret = venc_if_encode(ctx, VENC_START_OPT_ENCODE_FRAME, + &frm_buf, &bs_buf, &enc_result); + + dst_buf->vb2_buf.timestamp = src_buf->vb2_buf.timestamp; + dst_buf->timecode = src_buf->timecode; + + if (enc_result.is_key_frm) + dst_buf->flags |= V4L2_BUF_FLAG_KEYFRAME; + + if (ret) { + v4l2_m2m_buf_done(src_buf, VB2_BUF_STATE_ERROR); + vb2_set_plane_payload(&dst_buf->vb2_buf, 0, 0); + v4l2_m2m_buf_done(dst_buf, VB2_BUF_STATE_ERROR); + mtk_v4l2_venc_err(ctx, "venc_if_encode failed=%d", ret); + } else { + v4l2_m2m_buf_done(src_buf, VB2_BUF_STATE_DONE); + vb2_set_plane_payload(&dst_buf->vb2_buf, 0, enc_result.bs_size); + v4l2_m2m_buf_done(dst_buf, VB2_BUF_STATE_DONE); + mtk_v4l2_venc_dbg(2, ctx, "venc_if_encode bs size=%d", + enc_result.bs_size); + } + + v4l2_m2m_job_finish(ctx->dev->m2m_dev_enc, ctx->m2m_ctx); + + mtk_v4l2_venc_dbg(1, ctx, "<=== src_buf[%d] dst_buf[%d] venc_if_encode ret=%d Size=%u===>", + src_buf->vb2_buf.index, dst_buf->vb2_buf.index, ret, enc_result.bs_size); +} + +static void m2mops_venc_device_run(void *priv) +{ + struct mtk_vcodec_enc_ctx *ctx = priv; + + if ((ctx->q_data[MTK_Q_DATA_DST].fmt->fourcc == V4L2_PIX_FMT_H264) && + (ctx->state != MTK_STATE_HEADER)) { + /* encode h264 sps/pps header */ + mtk_venc_encode_header(ctx); + queue_work(ctx->dev->encode_workqueue, &ctx->encode_work); + return; + } + + mtk_venc_param_change(ctx); + queue_work(ctx->dev->encode_workqueue, &ctx->encode_work); +} + +static int m2mops_venc_job_ready(void *m2m_priv) +{ + struct mtk_vcodec_enc_ctx *ctx = m2m_priv; + + if (ctx->state == MTK_STATE_ABORT || ctx->state == MTK_STATE_FREE) { + mtk_v4l2_venc_dbg(3, ctx, "[%d]Not ready: state=0x%x.", ctx->id, ctx->state); + return 0; + } + + return 1; +} + +static void m2mops_venc_job_abort(void *priv) +{ + struct mtk_vcodec_enc_ctx *ctx = priv; + + ctx->state = MTK_STATE_ABORT; +} + +const struct v4l2_m2m_ops mtk_venc_m2m_ops = { + .device_run = m2mops_venc_device_run, + .job_ready = m2mops_venc_job_ready, + .job_abort = m2mops_venc_job_abort, +}; + +void mtk_vcodec_enc_set_default_params(struct mtk_vcodec_enc_ctx *ctx) +{ + struct mtk_q_data *q_data; + + ctx->m2m_ctx->q_lock = &ctx->q_mutex; + ctx->fh.m2m_ctx = ctx->m2m_ctx; + ctx->fh.ctrl_handler = &ctx->ctrl_hdl; + INIT_WORK(&ctx->encode_work, mtk_venc_worker); + + ctx->colorspace = V4L2_COLORSPACE_REC709; + ctx->ycbcr_enc = V4L2_YCBCR_ENC_DEFAULT; + ctx->quantization = V4L2_QUANTIZATION_DEFAULT; + ctx->xfer_func = V4L2_XFER_FUNC_DEFAULT; + + q_data = &ctx->q_data[MTK_Q_DATA_SRC]; + memset(q_data, 0, sizeof(struct mtk_q_data)); + q_data->visible_width = DFT_CFG_WIDTH; + q_data->visible_height = DFT_CFG_HEIGHT; + q_data->coded_width = DFT_CFG_WIDTH; + q_data->coded_height = DFT_CFG_HEIGHT; + q_data->field = V4L2_FIELD_NONE; + + q_data->fmt = &ctx->dev->venc_pdata->output_formats[0]; + + v4l_bound_align_image(&q_data->coded_width, + MTK_VENC_MIN_W, + MTK_VENC_HD_MAX_W, 4, + &q_data->coded_height, + MTK_VENC_MIN_H, + MTK_VENC_HD_MAX_H, 5, 6); + + if (q_data->coded_width < DFT_CFG_WIDTH && + (q_data->coded_width + 16) <= MTK_VENC_HD_MAX_W) + q_data->coded_width += 16; + if (q_data->coded_height < DFT_CFG_HEIGHT && + (q_data->coded_height + 32) <= MTK_VENC_HD_MAX_H) + q_data->coded_height += 32; + + q_data->sizeimage[0] = + q_data->coded_width * q_data->coded_height+ + ((ALIGN(q_data->coded_width, 16) * 2) * 16); + q_data->bytesperline[0] = q_data->coded_width; + q_data->sizeimage[1] = + (q_data->coded_width * q_data->coded_height) / 2 + + (ALIGN(q_data->coded_width, 16) * 16); + q_data->bytesperline[1] = q_data->coded_width; + + q_data = &ctx->q_data[MTK_Q_DATA_DST]; + memset(q_data, 0, sizeof(struct mtk_q_data)); + q_data->coded_width = DFT_CFG_WIDTH; + q_data->coded_height = DFT_CFG_HEIGHT; + q_data->fmt = &ctx->dev->venc_pdata->capture_formats[0]; + q_data->field = V4L2_FIELD_NONE; + ctx->q_data[MTK_Q_DATA_DST].sizeimage[0] = + DFT_CFG_WIDTH * DFT_CFG_HEIGHT; + ctx->q_data[MTK_Q_DATA_DST].bytesperline[0] = 0; + + ctx->enc_params.framerate_num = MTK_DEFAULT_FRAMERATE_NUM; + ctx->enc_params.framerate_denom = MTK_DEFAULT_FRAMERATE_DENOM; +} + +int mtk_vcodec_enc_ctrls_setup(struct mtk_vcodec_enc_ctx *ctx) +{ + const struct v4l2_ctrl_ops *ops = &mtk_vcodec_enc_ctrl_ops; + struct v4l2_ctrl_handler *handler = &ctx->ctrl_hdl; + u8 h264_max_level; + + if (ctx->dev->enc_capability & MTK_VENC_4K_CAPABILITY_ENABLE) + h264_max_level = V4L2_MPEG_VIDEO_H264_LEVEL_5_1; + else + h264_max_level = V4L2_MPEG_VIDEO_H264_LEVEL_4_2; + + v4l2_ctrl_handler_init(handler, MTK_MAX_CTRLS_HINT); + + v4l2_ctrl_new_std(handler, ops, V4L2_CID_MIN_BUFFERS_FOR_OUTPUT, + 1, 1, 1, 1); + v4l2_ctrl_new_std(handler, ops, V4L2_CID_MPEG_VIDEO_BITRATE, + ctx->dev->venc_pdata->min_bitrate, + ctx->dev->venc_pdata->max_bitrate, 1, 4000000); + v4l2_ctrl_new_std(handler, ops, V4L2_CID_MPEG_VIDEO_B_FRAMES, + 0, 2, 1, 0); + v4l2_ctrl_new_std(handler, ops, V4L2_CID_MPEG_VIDEO_FRAME_RC_ENABLE, + 0, 1, 1, 1); + v4l2_ctrl_new_std(handler, ops, V4L2_CID_MPEG_VIDEO_H264_MAX_QP, + 0, 51, 1, 51); + v4l2_ctrl_new_std(handler, ops, V4L2_CID_MPEG_VIDEO_H264_I_PERIOD, + 0, 65535, 1, 0); + v4l2_ctrl_new_std(handler, ops, V4L2_CID_MPEG_VIDEO_GOP_SIZE, + 0, 65535, 1, 0); + v4l2_ctrl_new_std(handler, ops, V4L2_CID_MPEG_VIDEO_MB_RC_ENABLE, + 0, 1, 1, 0); + v4l2_ctrl_new_std(handler, ops, V4L2_CID_MPEG_VIDEO_FORCE_KEY_FRAME, + 0, 0, 0, 0); + v4l2_ctrl_new_std_menu(handler, ops, + V4L2_CID_MPEG_VIDEO_HEADER_MODE, + V4L2_MPEG_VIDEO_HEADER_MODE_JOINED_WITH_1ST_FRAME, + 0, V4L2_MPEG_VIDEO_HEADER_MODE_SEPARATE); + v4l2_ctrl_new_std_menu(handler, ops, V4L2_CID_MPEG_VIDEO_H264_PROFILE, + V4L2_MPEG_VIDEO_H264_PROFILE_HIGH, + ~((1 << V4L2_MPEG_VIDEO_H264_PROFILE_BASELINE) | + (1 << V4L2_MPEG_VIDEO_H264_PROFILE_MAIN) | + (1 << V4L2_MPEG_VIDEO_H264_PROFILE_HIGH)), + V4L2_MPEG_VIDEO_H264_PROFILE_HIGH); + v4l2_ctrl_new_std_menu(handler, ops, V4L2_CID_MPEG_VIDEO_H264_LEVEL, + h264_max_level, + 0, V4L2_MPEG_VIDEO_H264_LEVEL_4_0); + v4l2_ctrl_new_std_menu(handler, ops, V4L2_CID_MPEG_VIDEO_VP8_PROFILE, + V4L2_MPEG_VIDEO_VP8_PROFILE_0, 0, V4L2_MPEG_VIDEO_VP8_PROFILE_0); + v4l2_ctrl_new_std_menu(handler, ops, V4L2_CID_MPEG_VIDEO_BITRATE_MODE, + V4L2_MPEG_VIDEO_BITRATE_MODE_CBR, + ~(1 << V4L2_MPEG_VIDEO_BITRATE_MODE_CBR), + V4L2_MPEG_VIDEO_BITRATE_MODE_CBR); + + + if (handler->error) { + mtk_v4l2_venc_err(ctx, "Init control handler fail %d", handler->error); + return handler->error; + } + + v4l2_ctrl_handler_setup(&ctx->ctrl_hdl); + + return 0; +} + +int mtk_vcodec_enc_queue_init(void *priv, struct vb2_queue *src_vq, + struct vb2_queue *dst_vq) +{ + struct mtk_vcodec_enc_ctx *ctx = priv; + int ret; + + /* Note: VB2_USERPTR works with dma-contig because mt8173 + * support iommu + * https://patchwork.kernel.org/patch/8335461/ + * https://patchwork.kernel.org/patch/7596181/ + */ + src_vq->type = V4L2_BUF_TYPE_VIDEO_OUTPUT_MPLANE; + src_vq->io_modes = VB2_DMABUF | VB2_MMAP | VB2_USERPTR; + src_vq->drv_priv = ctx; + src_vq->buf_struct_size = sizeof(struct mtk_video_enc_buf); + src_vq->ops = &mtk_venc_vb2_ops; + src_vq->mem_ops = &vb2_dma_contig_memops; + src_vq->timestamp_flags = V4L2_BUF_FLAG_TIMESTAMP_COPY; + src_vq->lock = &ctx->q_mutex; + src_vq->dev = &ctx->dev->plat_dev->dev; + + ret = vb2_queue_init(src_vq); + if (ret) + return ret; + + dst_vq->type = V4L2_BUF_TYPE_VIDEO_CAPTURE_MPLANE; + dst_vq->io_modes = VB2_DMABUF | VB2_MMAP | VB2_USERPTR; + dst_vq->drv_priv = ctx; + dst_vq->buf_struct_size = sizeof(struct v4l2_m2m_buffer); + dst_vq->ops = &mtk_venc_vb2_ops; + dst_vq->mem_ops = &vb2_dma_contig_memops; + dst_vq->timestamp_flags = V4L2_BUF_FLAG_TIMESTAMP_COPY; + dst_vq->lock = &ctx->q_mutex; + dst_vq->dev = &ctx->dev->plat_dev->dev; + + return vb2_queue_init(dst_vq); +} + +int mtk_venc_unlock(struct mtk_vcodec_enc_ctx *ctx) +{ + struct mtk_vcodec_enc_dev *dev = ctx->dev; + + mutex_unlock(&dev->enc_mutex); + return 0; +} + +int mtk_venc_lock(struct mtk_vcodec_enc_ctx *ctx) +{ + struct mtk_vcodec_enc_dev *dev = ctx->dev; + + mutex_lock(&dev->enc_mutex); + return 0; +} + +void mtk_vcodec_enc_release(struct mtk_vcodec_enc_ctx *ctx) +{ + int ret = venc_if_deinit(ctx); + + if (ret) + mtk_v4l2_venc_err(ctx, "venc_if_deinit failed=%d", ret); + + ctx->state = MTK_STATE_FREE; +} diff --git a/drivers/media/platform/mediatek/vcodec/encoder/mtk_vcodec_enc.h b/drivers/media/platform/mediatek/vcodec/encoder/mtk_vcodec_enc.h new file mode 100644 index 0000000000..82246401ed --- /dev/null +++ b/drivers/media/platform/mediatek/vcodec/encoder/mtk_vcodec_enc.h @@ -0,0 +1,52 @@ +/* SPDX-License-Identifier: GPL-2.0 */ +/* +* Copyright (c) 2016 MediaTek Inc. +* Author: PC Chen <pc.chen@mediatek.com> +* Tiffany Lin <tiffany.lin@mediatek.com> +*/ + +#ifndef _MTK_VCODEC_ENC_H_ +#define _MTK_VCODEC_ENC_H_ + +#include <media/videobuf2-core.h> +#include <media/v4l2-mem2mem.h> + +#include "mtk_vcodec_enc_drv.h" + +#define MTK_VENC_IRQ_STATUS_SPS 0x1 +#define MTK_VENC_IRQ_STATUS_PPS 0x2 +#define MTK_VENC_IRQ_STATUS_FRM 0x4 +#define MTK_VENC_IRQ_STATUS_DRAM 0x8 +#define MTK_VENC_IRQ_STATUS_PAUSE 0x10 +#define MTK_VENC_IRQ_STATUS_SWITCH 0x20 + +#define MTK_VENC_IRQ_STATUS_OFFSET 0x05C +#define MTK_VENC_IRQ_ACK_OFFSET 0x060 + +/** + * struct mtk_video_enc_buf - Private data related to each VB2 buffer. + * @m2m_buf: M2M buffer + * @list: list that buffer link to + * @param_change: Types of encode parameter change before encoding this + * buffer + * @enc_params: Encode parameters changed before encode this buffer + */ +struct mtk_video_enc_buf { + struct v4l2_m2m_buffer m2m_buf; + + u32 param_change; + struct mtk_enc_params enc_params; +}; + +extern const struct v4l2_ioctl_ops mtk_venc_ioctl_ops; +extern const struct v4l2_m2m_ops mtk_venc_m2m_ops; + +int mtk_venc_unlock(struct mtk_vcodec_enc_ctx *ctx); +int mtk_venc_lock(struct mtk_vcodec_enc_ctx *ctx); +int mtk_vcodec_enc_queue_init(void *priv, struct vb2_queue *src_vq, + struct vb2_queue *dst_vq); +void mtk_vcodec_enc_release(struct mtk_vcodec_enc_ctx *ctx); +int mtk_vcodec_enc_ctrls_setup(struct mtk_vcodec_enc_ctx *ctx); +void mtk_vcodec_enc_set_default_params(struct mtk_vcodec_enc_ctx *ctx); + +#endif /* _MTK_VCODEC_ENC_H_ */ diff --git a/drivers/media/platform/mediatek/vcodec/encoder/mtk_vcodec_enc_drv.c b/drivers/media/platform/mediatek/vcodec/encoder/mtk_vcodec_enc_drv.c new file mode 100644 index 0000000000..6319f24bc7 --- /dev/null +++ b/drivers/media/platform/mediatek/vcodec/encoder/mtk_vcodec_enc_drv.c @@ -0,0 +1,482 @@ +// SPDX-License-Identifier: GPL-2.0 +/* +* Copyright (c) 2016 MediaTek Inc. +* Author: PC Chen <pc.chen@mediatek.com> +* Tiffany Lin <tiffany.lin@mediatek.com> +*/ + +#include <linux/slab.h> +#include <linux/interrupt.h> +#include <linux/irq.h> +#include <linux/module.h> +#include <linux/of.h> +#include <linux/platform_device.h> +#include <linux/pm_runtime.h> +#include <media/v4l2-event.h> +#include <media/v4l2-mem2mem.h> +#include <media/videobuf2-dma-contig.h> + +#include "mtk_vcodec_enc.h" +#include "mtk_vcodec_enc_pm.h" +#include "../common/mtk_vcodec_intr.h" + +static const struct mtk_video_fmt mtk_video_formats_output[] = { + { + .fourcc = V4L2_PIX_FMT_NV12M, + .type = MTK_FMT_FRAME, + .num_planes = 2, + }, + { + .fourcc = V4L2_PIX_FMT_NV21M, + .type = MTK_FMT_FRAME, + .num_planes = 2, + }, + { + .fourcc = V4L2_PIX_FMT_YUV420M, + .type = MTK_FMT_FRAME, + .num_planes = 3, + }, + { + .fourcc = V4L2_PIX_FMT_YVU420M, + .type = MTK_FMT_FRAME, + .num_planes = 3, + }, +}; + +static const struct mtk_video_fmt mtk_video_formats_capture_h264[] = { + { + .fourcc = V4L2_PIX_FMT_H264, + .type = MTK_FMT_ENC, + .num_planes = 1, + }, +}; + +static const struct mtk_video_fmt mtk_video_formats_capture_vp8[] = { + { + .fourcc = V4L2_PIX_FMT_VP8, + .type = MTK_FMT_ENC, + .num_planes = 1, + }, +}; + +static void clean_irq_status(unsigned int irq_status, void __iomem *addr) +{ + if (irq_status & MTK_VENC_IRQ_STATUS_PAUSE) + writel(MTK_VENC_IRQ_STATUS_PAUSE, addr); + + if (irq_status & MTK_VENC_IRQ_STATUS_SWITCH) + writel(MTK_VENC_IRQ_STATUS_SWITCH, addr); + + if (irq_status & MTK_VENC_IRQ_STATUS_DRAM) + writel(MTK_VENC_IRQ_STATUS_DRAM, addr); + + if (irq_status & MTK_VENC_IRQ_STATUS_SPS) + writel(MTK_VENC_IRQ_STATUS_SPS, addr); + + if (irq_status & MTK_VENC_IRQ_STATUS_PPS) + writel(MTK_VENC_IRQ_STATUS_PPS, addr); + + if (irq_status & MTK_VENC_IRQ_STATUS_FRM) + writel(MTK_VENC_IRQ_STATUS_FRM, addr); + +} +static irqreturn_t mtk_vcodec_enc_irq_handler(int irq, void *priv) +{ + struct mtk_vcodec_enc_dev *dev = priv; + struct mtk_vcodec_enc_ctx *ctx; + unsigned long flags; + void __iomem *addr; + int core_id; + + spin_lock_irqsave(&dev->irqlock, flags); + ctx = dev->curr_ctx; + spin_unlock_irqrestore(&dev->irqlock, flags); + + core_id = dev->venc_pdata->core_id; + if (core_id < 0 || core_id >= NUM_MAX_VCODEC_REG_BASE) { + mtk_v4l2_venc_err(ctx, "Invalid core id: %d, ctx id: %d", core_id, ctx->id); + return IRQ_HANDLED; + } + + mtk_v4l2_venc_dbg(1, ctx, "id: %d, core id: %d", ctx->id, core_id); + + addr = dev->reg_base[core_id] + MTK_VENC_IRQ_ACK_OFFSET; + + ctx->irq_status = readl(dev->reg_base[core_id] + + (MTK_VENC_IRQ_STATUS_OFFSET)); + + clean_irq_status(ctx->irq_status, addr); + + wake_up_enc_ctx(ctx, MTK_INST_IRQ_RECEIVED, 0); + return IRQ_HANDLED; +} + +static int fops_vcodec_open(struct file *file) +{ + struct mtk_vcodec_enc_dev *dev = video_drvdata(file); + struct mtk_vcodec_enc_ctx *ctx = NULL; + int ret = 0; + struct vb2_queue *src_vq; + + ctx = kzalloc(sizeof(*ctx), GFP_KERNEL); + if (!ctx) + return -ENOMEM; + + mutex_lock(&dev->dev_mutex); + /* + * Use simple counter to uniquely identify this context. Only + * used for logging. + */ + ctx->id = dev->id_counter++; + v4l2_fh_init(&ctx->fh, video_devdata(file)); + file->private_data = &ctx->fh; + v4l2_fh_add(&ctx->fh); + INIT_LIST_HEAD(&ctx->list); + ctx->dev = dev; + init_waitqueue_head(&ctx->queue[0]); + mutex_init(&ctx->q_mutex); + + ctx->type = MTK_INST_ENCODER; + ret = mtk_vcodec_enc_ctrls_setup(ctx); + if (ret) { + mtk_v4l2_venc_err(ctx, "Failed to setup controls() (%d)", ret); + goto err_ctrls_setup; + } + ctx->m2m_ctx = v4l2_m2m_ctx_init(dev->m2m_dev_enc, ctx, + &mtk_vcodec_enc_queue_init); + if (IS_ERR((__force void *)ctx->m2m_ctx)) { + ret = PTR_ERR((__force void *)ctx->m2m_ctx); + mtk_v4l2_venc_err(ctx, "Failed to v4l2_m2m_ctx_init() (%d)", ret); + goto err_m2m_ctx_init; + } + src_vq = v4l2_m2m_get_vq(ctx->m2m_ctx, + V4L2_BUF_TYPE_VIDEO_OUTPUT_MPLANE); + ctx->empty_flush_buf.vb.vb2_buf.vb2_queue = src_vq; + mtk_vcodec_enc_set_default_params(ctx); + + if (v4l2_fh_is_singular(&ctx->fh)) { + /* + * load fireware to checks if it was loaded already and + * does nothing in that case + */ + ret = mtk_vcodec_fw_load_firmware(dev->fw_handler); + if (ret < 0) { + /* + * Return 0 if downloading firmware successfully, + * otherwise it is failed + */ + mtk_v4l2_venc_err(ctx, "vpu_load_firmware failed!"); + goto err_load_fw; + } + + dev->enc_capability = + mtk_vcodec_fw_get_venc_capa(dev->fw_handler); + mtk_v4l2_venc_dbg(0, ctx, "encoder capability %x", dev->enc_capability); + } + + mtk_v4l2_venc_dbg(2, ctx, "Create instance [%d]@%p m2m_ctx=%p ", + ctx->id, ctx, ctx->m2m_ctx); + + list_add(&ctx->list, &dev->ctx_list); + + mutex_unlock(&dev->dev_mutex); + mtk_v4l2_venc_dbg(0, ctx, "%s encoder [%d]", dev_name(&dev->plat_dev->dev), + ctx->id); + return ret; + + /* Deinit when failure occurred */ +err_load_fw: + v4l2_m2m_ctx_release(ctx->m2m_ctx); +err_m2m_ctx_init: + v4l2_ctrl_handler_free(&ctx->ctrl_hdl); +err_ctrls_setup: + v4l2_fh_del(&ctx->fh); + v4l2_fh_exit(&ctx->fh); + kfree(ctx); + mutex_unlock(&dev->dev_mutex); + + return ret; +} + +static int fops_vcodec_release(struct file *file) +{ + struct mtk_vcodec_enc_dev *dev = video_drvdata(file); + struct mtk_vcodec_enc_ctx *ctx = fh_to_enc_ctx(file->private_data); + + mtk_v4l2_venc_dbg(1, ctx, "[%d] encoder", ctx->id); + mutex_lock(&dev->dev_mutex); + + v4l2_m2m_ctx_release(ctx->m2m_ctx); + mtk_vcodec_enc_release(ctx); + v4l2_fh_del(&ctx->fh); + v4l2_fh_exit(&ctx->fh); + v4l2_ctrl_handler_free(&ctx->ctrl_hdl); + + list_del_init(&ctx->list); + kfree(ctx); + mutex_unlock(&dev->dev_mutex); + return 0; +} + +static const struct v4l2_file_operations mtk_vcodec_fops = { + .owner = THIS_MODULE, + .open = fops_vcodec_open, + .release = fops_vcodec_release, + .poll = v4l2_m2m_fop_poll, + .unlocked_ioctl = video_ioctl2, + .mmap = v4l2_m2m_fop_mmap, +}; + +static int mtk_vcodec_probe(struct platform_device *pdev) +{ + struct mtk_vcodec_enc_dev *dev; + struct video_device *vfd_enc; + phandle rproc_phandle; + enum mtk_vcodec_fw_type fw_type; + int ret; + + dev = devm_kzalloc(&pdev->dev, sizeof(*dev), GFP_KERNEL); + if (!dev) + return -ENOMEM; + + INIT_LIST_HEAD(&dev->ctx_list); + dev->plat_dev = pdev; + + if (!of_property_read_u32(pdev->dev.of_node, "mediatek,vpu", + &rproc_phandle)) { + fw_type = VPU; + } else if (!of_property_read_u32(pdev->dev.of_node, "mediatek,scp", + &rproc_phandle)) { + fw_type = SCP; + } else { + dev_err(&pdev->dev, "[MTK VCODEC] Could not get venc IPI device"); + return -ENODEV; + } + dma_set_max_seg_size(&pdev->dev, UINT_MAX); + + dev->fw_handler = mtk_vcodec_fw_select(dev, fw_type, ENCODER); + if (IS_ERR(dev->fw_handler)) + return PTR_ERR(dev->fw_handler); + + dev->venc_pdata = of_device_get_match_data(&pdev->dev); + ret = mtk_vcodec_init_enc_clk(dev); + if (ret < 0) { + dev_err(&pdev->dev, "[MTK VCODEC] Failed to get mtk vcodec clock source!"); + goto err_enc_pm; + } + + pm_runtime_enable(&pdev->dev); + + dev->reg_base[dev->venc_pdata->core_id] = + devm_platform_ioremap_resource(pdev, 0); + if (IS_ERR(dev->reg_base[dev->venc_pdata->core_id])) { + ret = PTR_ERR(dev->reg_base[dev->venc_pdata->core_id]); + goto err_res; + } + + dev->enc_irq = platform_get_irq(pdev, 0); + if (dev->enc_irq < 0) { + ret = dev->enc_irq; + goto err_res; + } + + irq_set_status_flags(dev->enc_irq, IRQ_NOAUTOEN); + ret = devm_request_irq(&pdev->dev, dev->enc_irq, + mtk_vcodec_enc_irq_handler, + 0, pdev->name, dev); + if (ret) { + dev_err(&pdev->dev, + "[MTK VCODEC] Failed to install dev->enc_irq %d (%d) core_id (%d)", + dev->enc_irq, ret, dev->venc_pdata->core_id); + ret = -EINVAL; + goto err_res; + } + + mutex_init(&dev->enc_mutex); + mutex_init(&dev->dev_mutex); + spin_lock_init(&dev->irqlock); + + snprintf(dev->v4l2_dev.name, sizeof(dev->v4l2_dev.name), "%s", + "[MTK_V4L2_VENC]"); + + ret = v4l2_device_register(&pdev->dev, &dev->v4l2_dev); + if (ret) { + dev_err(&pdev->dev, "[MTK VCODEC] v4l2_device_register err=%d", ret); + goto err_res; + } + + /* allocate video device for encoder and register it */ + vfd_enc = video_device_alloc(); + if (!vfd_enc) { + dev_err(&pdev->dev, "[MTK VCODEC] Failed to allocate video device"); + ret = -ENOMEM; + goto err_enc_alloc; + } + vfd_enc->fops = &mtk_vcodec_fops; + vfd_enc->ioctl_ops = &mtk_venc_ioctl_ops; + vfd_enc->release = video_device_release; + vfd_enc->lock = &dev->dev_mutex; + vfd_enc->v4l2_dev = &dev->v4l2_dev; + vfd_enc->vfl_dir = VFL_DIR_M2M; + vfd_enc->device_caps = V4L2_CAP_VIDEO_M2M_MPLANE | + V4L2_CAP_STREAMING; + + snprintf(vfd_enc->name, sizeof(vfd_enc->name), "%s", + MTK_VCODEC_ENC_NAME); + video_set_drvdata(vfd_enc, dev); + dev->vfd_enc = vfd_enc; + platform_set_drvdata(pdev, dev); + + dev->m2m_dev_enc = v4l2_m2m_init(&mtk_venc_m2m_ops); + if (IS_ERR((__force void *)dev->m2m_dev_enc)) { + dev_err(&pdev->dev, "[MTK VCODEC] Failed to init mem2mem enc device"); + ret = PTR_ERR((__force void *)dev->m2m_dev_enc); + goto err_enc_mem_init; + } + + dev->encode_workqueue = + alloc_ordered_workqueue(MTK_VCODEC_ENC_NAME, + WQ_MEM_RECLAIM | + WQ_FREEZABLE); + if (!dev->encode_workqueue) { + dev_err(&pdev->dev, "[MTK VCODEC] Failed to create encode workqueue"); + ret = -EINVAL; + goto err_event_workq; + } + + ret = video_register_device(vfd_enc, VFL_TYPE_VIDEO, -1); + if (ret) { + dev_err(&pdev->dev, "[MTK VCODEC] Failed to register video device"); + goto err_enc_reg; + } + + mtk_vcodec_dbgfs_init(dev, true); + dev_dbg(&pdev->dev, "[MTK VCODEC] encoder %d registered as /dev/video%d", + dev->venc_pdata->core_id, vfd_enc->num); + + return 0; + +err_enc_reg: + destroy_workqueue(dev->encode_workqueue); +err_event_workq: + v4l2_m2m_release(dev->m2m_dev_enc); +err_enc_mem_init: + video_unregister_device(vfd_enc); +err_enc_alloc: + v4l2_device_unregister(&dev->v4l2_dev); +err_res: + pm_runtime_disable(dev->pm.dev); +err_enc_pm: + mtk_vcodec_fw_release(dev->fw_handler); + return ret; +} + +static const struct mtk_vcodec_enc_pdata mt8173_avc_pdata = { + .capture_formats = mtk_video_formats_capture_h264, + .num_capture_formats = ARRAY_SIZE(mtk_video_formats_capture_h264), + .output_formats = mtk_video_formats_output, + .num_output_formats = ARRAY_SIZE(mtk_video_formats_output), + .min_bitrate = 64, + .max_bitrate = 60000000, + .core_id = VENC_SYS, +}; + +static const struct mtk_vcodec_enc_pdata mt8173_vp8_pdata = { + .capture_formats = mtk_video_formats_capture_vp8, + .num_capture_formats = ARRAY_SIZE(mtk_video_formats_capture_vp8), + .output_formats = mtk_video_formats_output, + .num_output_formats = ARRAY_SIZE(mtk_video_formats_output), + .min_bitrate = 64, + .max_bitrate = 9000000, + .core_id = VENC_LT_SYS, +}; + +static const struct mtk_vcodec_enc_pdata mt8183_pdata = { + .uses_ext = true, + .capture_formats = mtk_video_formats_capture_h264, + .num_capture_formats = ARRAY_SIZE(mtk_video_formats_capture_h264), + .output_formats = mtk_video_formats_output, + .num_output_formats = ARRAY_SIZE(mtk_video_formats_output), + .min_bitrate = 64, + .max_bitrate = 40000000, + .core_id = VENC_SYS, +}; + +static const struct mtk_vcodec_enc_pdata mt8188_pdata = { + .uses_ext = true, + .capture_formats = mtk_video_formats_capture_h264, + .num_capture_formats = ARRAY_SIZE(mtk_video_formats_capture_h264), + .output_formats = mtk_video_formats_output, + .num_output_formats = ARRAY_SIZE(mtk_video_formats_output), + .min_bitrate = 64, + .max_bitrate = 50000000, + .core_id = VENC_SYS, + .uses_34bit = true, +}; + +static const struct mtk_vcodec_enc_pdata mt8192_pdata = { + .uses_ext = true, + .capture_formats = mtk_video_formats_capture_h264, + .num_capture_formats = ARRAY_SIZE(mtk_video_formats_capture_h264), + .output_formats = mtk_video_formats_output, + .num_output_formats = ARRAY_SIZE(mtk_video_formats_output), + .min_bitrate = 64, + .max_bitrate = 100000000, + .core_id = VENC_SYS, +}; + +static const struct mtk_vcodec_enc_pdata mt8195_pdata = { + .uses_ext = true, + .capture_formats = mtk_video_formats_capture_h264, + .num_capture_formats = ARRAY_SIZE(mtk_video_formats_capture_h264), + .output_formats = mtk_video_formats_output, + .num_output_formats = ARRAY_SIZE(mtk_video_formats_output), + .min_bitrate = 64, + .max_bitrate = 100000000, + .core_id = VENC_SYS, +}; + +static const struct of_device_id mtk_vcodec_enc_match[] = { + {.compatible = "mediatek,mt8173-vcodec-enc", + .data = &mt8173_avc_pdata}, + {.compatible = "mediatek,mt8173-vcodec-enc-vp8", + .data = &mt8173_vp8_pdata}, + {.compatible = "mediatek,mt8183-vcodec-enc", .data = &mt8183_pdata}, + {.compatible = "mediatek,mt8188-vcodec-enc", .data = &mt8188_pdata}, + {.compatible = "mediatek,mt8192-vcodec-enc", .data = &mt8192_pdata}, + {.compatible = "mediatek,mt8195-vcodec-enc", .data = &mt8195_pdata}, + {}, +}; +MODULE_DEVICE_TABLE(of, mtk_vcodec_enc_match); + +static void mtk_vcodec_enc_remove(struct platform_device *pdev) +{ + struct mtk_vcodec_enc_dev *dev = platform_get_drvdata(pdev); + + destroy_workqueue(dev->encode_workqueue); + if (dev->m2m_dev_enc) + v4l2_m2m_release(dev->m2m_dev_enc); + + if (dev->vfd_enc) + video_unregister_device(dev->vfd_enc); + + mtk_vcodec_dbgfs_deinit(&dev->dbgfs); + v4l2_device_unregister(&dev->v4l2_dev); + pm_runtime_disable(dev->pm.dev); + mtk_vcodec_fw_release(dev->fw_handler); +} + +static struct platform_driver mtk_vcodec_enc_driver = { + .probe = mtk_vcodec_probe, + .remove_new = mtk_vcodec_enc_remove, + .driver = { + .name = MTK_VCODEC_ENC_NAME, + .of_match_table = mtk_vcodec_enc_match, + }, +}; + +module_platform_driver(mtk_vcodec_enc_driver); + + +MODULE_LICENSE("GPL v2"); +MODULE_DESCRIPTION("Mediatek video codec V4L2 encoder driver"); diff --git a/drivers/media/platform/mediatek/vcodec/encoder/mtk_vcodec_enc_drv.h b/drivers/media/platform/mediatek/vcodec/encoder/mtk_vcodec_enc_drv.h new file mode 100644 index 0000000000..a042f607ed --- /dev/null +++ b/drivers/media/platform/mediatek/vcodec/encoder/mtk_vcodec_enc_drv.h @@ -0,0 +1,248 @@ +/* SPDX-License-Identifier: GPL-2.0 */ +/* + * Copyright (c) 2023 MediaTek Inc. + * Author: Yunfei Dong <yunfei.dong@mediatek.com> + */ + +#ifndef _MTK_VCODEC_ENC_DRV_H_ +#define _MTK_VCODEC_ENC_DRV_H_ + +#include "../common/mtk_vcodec_cmn_drv.h" +#include "../common/mtk_vcodec_dbgfs.h" +#include "../common/mtk_vcodec_fw_priv.h" +#include "../common/mtk_vcodec_util.h" + +#define MTK_VCODEC_ENC_NAME "mtk-vcodec-enc" + +#define MTK_ENC_CTX_IS_EXT(ctx) ((ctx)->dev->venc_pdata->uses_ext) +#define MTK_ENC_IOVA_IS_34BIT(ctx) ((ctx)->dev->venc_pdata->uses_34bit) + +/** + * struct mtk_vcodec_enc_pdata - compatible data for each IC + * + * @uses_ext: whether the encoder uses the extended firmware messaging format + * @min_bitrate: minimum supported encoding bitrate + * @max_bitrate: maximum supported encoding bitrate + * @capture_formats: array of supported capture formats + * @num_capture_formats: number of entries in capture_formats + * @output_formats: array of supported output formats + * @num_output_formats: number of entries in output_formats + * @core_id: stand for h264 or vp8 encode index + * @uses_34bit: whether the encoder uses 34-bit iova + */ +struct mtk_vcodec_enc_pdata { + bool uses_ext; + u64 min_bitrate; + u64 max_bitrate; + const struct mtk_video_fmt *capture_formats; + size_t num_capture_formats; + const struct mtk_video_fmt *output_formats; + size_t num_output_formats; + u8 core_id; + bool uses_34bit; +}; + +/* + * enum mtk_encode_param - General encoding parameters type + */ +enum mtk_encode_param { + MTK_ENCODE_PARAM_NONE = 0, + MTK_ENCODE_PARAM_BITRATE = (1 << 0), + MTK_ENCODE_PARAM_FRAMERATE = (1 << 1), + MTK_ENCODE_PARAM_INTRA_PERIOD = (1 << 2), + MTK_ENCODE_PARAM_FORCE_INTRA = (1 << 3), + MTK_ENCODE_PARAM_GOP_SIZE = (1 << 4), +}; + +/** + * struct mtk_enc_params - General encoding parameters + * @bitrate: target bitrate in bits per second + * @num_b_frame: number of b frames between p-frame + * @rc_frame: frame based rate control + * @rc_mb: macroblock based rate control + * @seq_hdr_mode: H.264 sequence header is encoded separately or joined + * with the first frame + * @intra_period: I frame period + * @gop_size: group of picture size, it's used as the intra frame period + * @framerate_num: frame rate numerator. ex: framerate_num=30 and + * framerate_denom=1 means FPS is 30 + * @framerate_denom: frame rate denominator. ex: framerate_num=30 and + * framerate_denom=1 means FPS is 30 + * @h264_max_qp: Max value for H.264 quantization parameter + * @h264_profile: V4L2 defined H.264 profile + * @h264_level: V4L2 defined H.264 level + * @force_intra: force/insert intra frame + */ +struct mtk_enc_params { + unsigned int bitrate; + unsigned int num_b_frame; + unsigned int rc_frame; + unsigned int rc_mb; + unsigned int seq_hdr_mode; + unsigned int intra_period; + unsigned int gop_size; + unsigned int framerate_num; + unsigned int framerate_denom; + unsigned int h264_max_qp; + unsigned int h264_profile; + unsigned int h264_level; + unsigned int force_intra; +}; + +/** + * struct mtk_vcodec_enc_ctx - Context (instance) private data. + * + * @type: type of encoder instance + * @dev: pointer to the mtk_vcodec_enc_dev of the device + * @list: link to ctx_list of mtk_vcodec_enc_dev + * + * @fh: struct v4l2_fh + * @m2m_ctx: pointer to the v4l2_m2m_ctx of the context + * @q_data: store information of input and output queue of the context + * @id: index of the context that this structure describes + * @state: state of the context + * @param_change: indicate encode parameter type + * @enc_params: encoding parameters + * + * @enc_if: hooked encoder driver interface + * @drv_handle: driver handle for specific decode/encode instance + * + * @int_cond: variable used by the waitqueue + * @int_type: type of the last interrupt + * @queue: waitqueue that can be used to wait for this context to finish + * @irq_status: irq status + * + * @ctrl_hdl: handler for v4l2 framework + * @encode_work: worker for the encoding + * @empty_flush_buf: a fake size-0 capture buffer that indicates flush. Used for encoder. + * @is_flushing: set to true if flushing is in progress. + * + * @colorspace: enum v4l2_colorspace; supplemental to pixelformat + * @ycbcr_enc: enum v4l2_ycbcr_encoding, Y'CbCr encoding + * @quantization: enum v4l2_quantization, colorspace quantization + * @xfer_func: enum v4l2_xfer_func, colorspace transfer function + * + * @q_mutex: vb2_queue mutex. + * @vpu_inst: vpu instance pointer. + */ +struct mtk_vcodec_enc_ctx { + enum mtk_instance_type type; + struct mtk_vcodec_enc_dev *dev; + struct list_head list; + + struct v4l2_fh fh; + struct v4l2_m2m_ctx *m2m_ctx; + struct mtk_q_data q_data[2]; + int id; + enum mtk_instance_state state; + enum mtk_encode_param param_change; + struct mtk_enc_params enc_params; + + const struct venc_common_if *enc_if; + void *drv_handle; + + int int_cond[MTK_VDEC_HW_MAX]; + int int_type[MTK_VDEC_HW_MAX]; + wait_queue_head_t queue[MTK_VDEC_HW_MAX]; + unsigned int irq_status; + + struct v4l2_ctrl_handler ctrl_hdl; + struct work_struct encode_work; + struct v4l2_m2m_buffer empty_flush_buf; + bool is_flushing; + + enum v4l2_colorspace colorspace; + enum v4l2_ycbcr_encoding ycbcr_enc; + enum v4l2_quantization quantization; + enum v4l2_xfer_func xfer_func; + + struct mutex q_mutex; + void *vpu_inst; +}; + +/** + * struct mtk_vcodec_enc_dev - driver data + * @v4l2_dev: V4L2 device to register video devices for. + * @vfd_enc: Video device for encoder. + * + * @m2m_dev_enc: m2m device for encoder. + * @plat_dev: platform device + * @ctx_list: list of struct mtk_vcodec_ctx + * @curr_ctx: The context that is waiting for codec hardware + * + * @reg_base: Mapped address of MTK Vcodec registers. + * @venc_pdata: encoder IC-specific data + * + * @fw_handler: used to communicate with the firmware. + * @id_counter: used to identify current opened instance + * + * @enc_mutex: encoder hardware lock. + * @dev_mutex: video_device lock + * @encode_workqueue: encode work queue + * + * @enc_irq: h264 encoder irq resource + * @irqlock: protect data access by irq handler and work thread + * + * @pm: power management control + * @enc_capability: used to identify encode capability + * @dbgfs: debug log related information + */ +struct mtk_vcodec_enc_dev { + struct v4l2_device v4l2_dev; + struct video_device *vfd_enc; + + struct v4l2_m2m_dev *m2m_dev_enc; + struct platform_device *plat_dev; + struct list_head ctx_list; + struct mtk_vcodec_enc_ctx *curr_ctx; + + void __iomem *reg_base[NUM_MAX_VCODEC_REG_BASE]; + const struct mtk_vcodec_enc_pdata *venc_pdata; + + struct mtk_vcodec_fw *fw_handler; + u64 id_counter; + + /* encoder hardware mutex lock */ + struct mutex enc_mutex; + struct mutex dev_mutex; + struct workqueue_struct *encode_workqueue; + + int enc_irq; + spinlock_t irqlock; + + struct mtk_vcodec_pm pm; + unsigned int enc_capability; + struct mtk_vcodec_dbgfs dbgfs; +}; + +static inline struct mtk_vcodec_enc_ctx *fh_to_enc_ctx(struct v4l2_fh *fh) +{ + return container_of(fh, struct mtk_vcodec_enc_ctx, fh); +} + +static inline struct mtk_vcodec_enc_ctx *ctrl_to_enc_ctx(struct v4l2_ctrl *ctrl) +{ + return container_of(ctrl->handler, struct mtk_vcodec_enc_ctx, ctrl_hdl); +} + +/* Wake up context wait_queue */ +static inline void +wake_up_enc_ctx(struct mtk_vcodec_enc_ctx *ctx, unsigned int reason, unsigned int hw_id) +{ + ctx->int_cond[hw_id] = 1; + ctx->int_type[hw_id] = reason; + wake_up_interruptible(&ctx->queue[hw_id]); +} + +#define mtk_venc_err(ctx, fmt, args...) \ + mtk_vcodec_err((ctx)->id, (ctx)->dev->plat_dev, fmt, ##args) + +#define mtk_venc_debug(ctx, fmt, args...) \ + mtk_vcodec_debug((ctx)->id, (ctx)->dev->plat_dev, fmt, ##args) + +#define mtk_v4l2_venc_err(ctx, fmt, args...) mtk_v4l2_err((ctx)->dev->plat_dev, fmt, ##args) + +#define mtk_v4l2_venc_dbg(level, ctx, fmt, args...) \ + mtk_v4l2_debug((ctx)->dev->plat_dev, level, fmt, ##args) + +#endif /* _MTK_VCODEC_ENC_DRV_H_ */ diff --git a/drivers/media/platform/mediatek/vcodec/encoder/mtk_vcodec_enc_pm.c b/drivers/media/platform/mediatek/vcodec/encoder/mtk_vcodec_enc_pm.c new file mode 100644 index 0000000000..3fce936e61 --- /dev/null +++ b/drivers/media/platform/mediatek/vcodec/encoder/mtk_vcodec_enc_pm.c @@ -0,0 +1,89 @@ +// SPDX-License-Identifier: GPL-2.0 +/* +* Copyright (c) 2016 MediaTek Inc. +* Author: Tiffany Lin <tiffany.lin@mediatek.com> +*/ + +#include <linux/clk.h> +#include <linux/of.h> +#include <linux/pm_runtime.h> + +#include "mtk_vcodec_enc_drv.h" +#include "mtk_vcodec_enc_pm.h" + +int mtk_vcodec_init_enc_clk(struct mtk_vcodec_enc_dev *mtkdev) +{ + struct platform_device *pdev; + struct mtk_vcodec_pm *pm; + struct mtk_vcodec_clk *enc_clk; + struct mtk_vcodec_clk_info *clk_info; + int ret, i; + + pdev = mtkdev->plat_dev; + pm = &mtkdev->pm; + memset(pm, 0, sizeof(struct mtk_vcodec_pm)); + pm->dev = &pdev->dev; + enc_clk = &pm->venc_clk; + + enc_clk->clk_num = of_property_count_strings(pdev->dev.of_node, + "clock-names"); + if (enc_clk->clk_num > 0) { + enc_clk->clk_info = devm_kcalloc(&pdev->dev, + enc_clk->clk_num, sizeof(*clk_info), + GFP_KERNEL); + if (!enc_clk->clk_info) + return -ENOMEM; + } else { + dev_err(pm->dev, "[MTK VCODEC] Failed to get venc clock count"); + return -EINVAL; + } + + for (i = 0; i < enc_clk->clk_num; i++) { + clk_info = &enc_clk->clk_info[i]; + ret = of_property_read_string_index(pdev->dev.of_node, + "clock-names", i, &clk_info->clk_name); + if (ret) { + dev_err(pm->dev, "[MTK VCODEC] venc failed to get clk name %d", i); + return ret; + } + clk_info->vcodec_clk = devm_clk_get(&pdev->dev, + clk_info->clk_name); + if (IS_ERR(clk_info->vcodec_clk)) { + dev_err(pm->dev, "[MTK VCODEC] venc devm_clk_get (%d)%s fail", i, + clk_info->clk_name); + return PTR_ERR(clk_info->vcodec_clk); + } + } + + return 0; +} + +void mtk_vcodec_enc_clock_on(struct mtk_vcodec_pm *pm) +{ + struct mtk_vcodec_clk *enc_clk = &pm->venc_clk; + int ret, i = 0; + + for (i = 0; i < enc_clk->clk_num; i++) { + ret = clk_prepare_enable(enc_clk->clk_info[i].vcodec_clk); + if (ret) { + dev_err(pm->dev, "[MTK VCODEC] venc clk_prepare_enable %d %s fail %d", i, + enc_clk->clk_info[i].clk_name, ret); + goto clkerr; + } + } + + return; + +clkerr: + for (i -= 1; i >= 0; i--) + clk_disable_unprepare(enc_clk->clk_info[i].vcodec_clk); +} + +void mtk_vcodec_enc_clock_off(struct mtk_vcodec_pm *pm) +{ + struct mtk_vcodec_clk *enc_clk = &pm->venc_clk; + int i = 0; + + for (i = enc_clk->clk_num - 1; i >= 0; i--) + clk_disable_unprepare(enc_clk->clk_info[i].vcodec_clk); +} diff --git a/drivers/media/platform/mediatek/vcodec/encoder/mtk_vcodec_enc_pm.h b/drivers/media/platform/mediatek/vcodec/encoder/mtk_vcodec_enc_pm.h new file mode 100644 index 0000000000..e50be05751 --- /dev/null +++ b/drivers/media/platform/mediatek/vcodec/encoder/mtk_vcodec_enc_pm.h @@ -0,0 +1,17 @@ +/* SPDX-License-Identifier: GPL-2.0 */ +/* +* Copyright (c) 2016 MediaTek Inc. +* Author: Tiffany Lin <tiffany.lin@mediatek.com> +*/ + +#ifndef _MTK_VCODEC_ENC_PM_H_ +#define _MTK_VCODEC_ENC_PM_H_ + +#include "mtk_vcodec_enc_drv.h" + +int mtk_vcodec_init_enc_clk(struct mtk_vcodec_enc_dev *dev); + +void mtk_vcodec_enc_clock_on(struct mtk_vcodec_pm *pm); +void mtk_vcodec_enc_clock_off(struct mtk_vcodec_pm *pm); + +#endif /* _MTK_VCODEC_ENC_PM_H_ */ diff --git a/drivers/media/platform/mediatek/vcodec/encoder/venc/venc_h264_if.c b/drivers/media/platform/mediatek/vcodec/encoder/venc/venc_h264_if.c new file mode 100644 index 0000000000..a68dac72c4 --- /dev/null +++ b/drivers/media/platform/mediatek/vcodec/encoder/venc/venc_h264_if.c @@ -0,0 +1,814 @@ +// SPDX-License-Identifier: GPL-2.0 +/* + * Copyright (c) 2016 MediaTek Inc. + * Author: Jungchang Tsao <jungchang.tsao@mediatek.com> + * Daniel Hsiao <daniel.hsiao@mediatek.com> + * PoChun Lin <pochun.lin@mediatek.com> + */ + +#include <linux/interrupt.h> +#include <linux/kernel.h> +#include <linux/slab.h> + +#include "../mtk_vcodec_enc_drv.h" +#include "../../common/mtk_vcodec_intr.h" +#include "../mtk_vcodec_enc.h" +#include "../mtk_vcodec_enc_pm.h" +#include "../venc_drv_base.h" +#include "../venc_ipi_msg.h" +#include "../venc_vpu_if.h" + +static const char h264_filler_marker[] = {0x0, 0x0, 0x0, 0x1, 0xc}; + +#define H264_FILLER_MARKER_SIZE ARRAY_SIZE(h264_filler_marker) +#define VENC_PIC_BITSTREAM_BYTE_CNT 0x0098 + +/* + * enum venc_h264_frame_type - h264 encoder output bitstream frame type + */ +enum venc_h264_frame_type { + VENC_H264_IDR_FRM, + VENC_H264_I_FRM, + VENC_H264_P_FRM, + VENC_H264_B_FRM, +}; + +/* + * enum venc_h264_vpu_work_buf - h264 encoder buffer index + */ +enum venc_h264_vpu_work_buf { + VENC_H264_VPU_WORK_BUF_RC_INFO, + VENC_H264_VPU_WORK_BUF_RC_CODE, + VENC_H264_VPU_WORK_BUF_REC_LUMA, + VENC_H264_VPU_WORK_BUF_REC_CHROMA, + VENC_H264_VPU_WORK_BUF_REF_LUMA, + VENC_H264_VPU_WORK_BUF_REF_CHROMA, + VENC_H264_VPU_WORK_BUF_MV_INFO_1, + VENC_H264_VPU_WORK_BUF_MV_INFO_2, + VENC_H264_VPU_WORK_BUF_SKIP_FRAME, + VENC_H264_VPU_WORK_BUF_MAX, +}; + +/* + * enum venc_h264_bs_mode - for bs_mode argument in h264_enc_vpu_encode + */ +enum venc_h264_bs_mode { + H264_BS_MODE_SPS, + H264_BS_MODE_PPS, + H264_BS_MODE_FRAME, +}; + +/* + * struct venc_h264_vpu_config - Structure for h264 encoder configuration + * AP-W/R : AP is writer/reader on this item + * VPU-W/R: VPU is write/reader on this item + * @input_fourcc: input fourcc + * @bitrate: target bitrate (in bps) + * @pic_w: picture width. Picture size is visible stream resolution, in pixels, + * to be used for display purposes; must be smaller or equal to buffer + * size. + * @pic_h: picture height + * @buf_w: buffer width. Buffer size is stream resolution in pixels aligned to + * hardware requirements. + * @buf_h: buffer height + * @gop_size: group of picture size (idr frame) + * @intra_period: intra frame period + * @framerate: frame rate in fps + * @profile: as specified in standard + * @level: as specified in standard + * @wfd: WFD mode 1:on, 0:off + */ +struct venc_h264_vpu_config { + u32 input_fourcc; + u32 bitrate; + u32 pic_w; + u32 pic_h; + u32 buf_w; + u32 buf_h; + u32 gop_size; + u32 intra_period; + u32 framerate; + u32 profile; + u32 level; + u32 wfd; +}; + +/* + * struct venc_h264_vpu_buf - Structure for buffer information + * AP-W/R : AP is writer/reader on this item + * VPU-W/R: VPU is write/reader on this item + * @iova: IO virtual address + * @vpua: VPU side memory addr which is used by RC_CODE + * @size: buffer size (in bytes) + */ +struct venc_h264_vpu_buf { + u32 iova; + u32 vpua; + u32 size; +}; + +/* + * struct venc_h264_vsi - Structure for VPU driver control and info share + * AP-W/R : AP is writer/reader on this item + * VPU-W/R: VPU is write/reader on this item + * This structure is allocated in VPU side and shared to AP side. + * @config: h264 encoder configuration + * @work_bufs: working buffer information in VPU side + * The work_bufs here is for storing the 'size' info shared to AP side. + * The similar item in struct venc_h264_inst is for memory allocation + * in AP side. The AP driver will copy the 'size' from here to the one in + * struct mtk_vcodec_mem, then invoke mtk_vcodec_mem_alloc to allocate + * the buffer. After that, bypass the 'dma_addr' to the 'iova' field here for + * register setting in VPU side. + */ +struct venc_h264_vsi { + struct venc_h264_vpu_config config; + struct venc_h264_vpu_buf work_bufs[VENC_H264_VPU_WORK_BUF_MAX]; +}; + +/** + * struct venc_h264_vpu_config_ext - Structure for h264 encoder configuration + * AP-W/R : AP is writer/reader on this item + * VPU-W/R: VPU is write/reader on this item + * @input_fourcc: input fourcc + * @bitrate: target bitrate (in bps) + * @pic_w: picture width. Picture size is visible stream resolution, in pixels, + * to be used for display purposes; must be smaller or equal to buffer + * size. + * @pic_h: picture height + * @buf_w: buffer width. Buffer size is stream resolution in pixels aligned to + * hardware requirements. + * @buf_h: buffer height + * @gop_size: group of picture size (idr frame) + * @intra_period: intra frame period + * @framerate: frame rate in fps + * @profile: as specified in standard + * @level: as specified in standard + * @wfd: WFD mode 1:on, 0:off + * @max_qp: max quant parameter + * @min_qp: min quant parameter + * @reserved: reserved configs + */ +struct venc_h264_vpu_config_ext { + u32 input_fourcc; + u32 bitrate; + u32 pic_w; + u32 pic_h; + u32 buf_w; + u32 buf_h; + u32 gop_size; + u32 intra_period; + u32 framerate; + u32 profile; + u32 level; + u32 wfd; + u32 max_qp; + u32 min_qp; + u32 reserved[8]; +}; + +/** + * struct venc_h264_vpu_buf_34 - Structure for 34-bit buffer information + * AP-W/R : AP is writer/reader on this item + * VPU-W/R: VPU is write/reader on this item + * @iova: 34-bit IO virtual address + * @vpua: VPU side memory addr which is used by RC_CODE + * @size: buffer size (in bytes) + */ +struct venc_h264_vpu_buf_34 { + u64 iova; + u32 vpua; + u32 size; +}; + +/** + * struct venc_h264_vsi_34 - Structure for VPU driver control and info share + * Used for 34-bit iova sharing + * @config: h264 encoder configuration + * @work_bufs: working buffer information in VPU side + */ +struct venc_h264_vsi_34 { + struct venc_h264_vpu_config_ext config; + struct venc_h264_vpu_buf_34 work_bufs[VENC_H264_VPU_WORK_BUF_MAX]; +}; + +/* + * struct venc_h264_inst - h264 encoder AP driver instance + * @hw_base: h264 encoder hardware register base + * @work_bufs: working buffer + * @pps_buf: buffer to store the pps bitstream + * @work_buf_allocated: working buffer allocated flag + * @frm_cnt: encoded frame count + * @prepend_hdr: when the v4l2 layer send VENC_SET_PARAM_PREPEND_HEADER cmd + * through h264_enc_set_param interface, it will set this flag and prepend the + * sps/pps in h264_enc_encode function. + * @vpu_inst: VPU instance to exchange information between AP and VPU + * @vsi: driver structure allocated by VPU side and shared to AP side for + * control and info share + * @vsi_34: driver structure allocated by VPU side and shared to AP side for + * control and info share, used for 34-bit iova sharing. + * @ctx: context for v4l2 layer integration + */ +struct venc_h264_inst { + void __iomem *hw_base; + struct mtk_vcodec_mem work_bufs[VENC_H264_VPU_WORK_BUF_MAX]; + struct mtk_vcodec_mem pps_buf; + bool work_buf_allocated; + unsigned int frm_cnt; + unsigned int skip_frm_cnt; + unsigned int prepend_hdr; + struct venc_vpu_inst vpu_inst; + struct venc_h264_vsi *vsi; + struct venc_h264_vsi_34 *vsi_34; + struct mtk_vcodec_enc_ctx *ctx; +}; + +static inline u32 h264_read_reg(struct venc_h264_inst *inst, u32 addr) +{ + return readl(inst->hw_base + addr); +} + +static unsigned int h264_get_profile(struct venc_h264_inst *inst, + unsigned int profile) +{ + switch (profile) { + case V4L2_MPEG_VIDEO_H264_PROFILE_BASELINE: + return 66; + case V4L2_MPEG_VIDEO_H264_PROFILE_MAIN: + return 77; + case V4L2_MPEG_VIDEO_H264_PROFILE_HIGH: + return 100; + case V4L2_MPEG_VIDEO_H264_PROFILE_CONSTRAINED_BASELINE: + mtk_venc_err(inst->ctx, "unsupported CONSTRAINED_BASELINE"); + return 0; + case V4L2_MPEG_VIDEO_H264_PROFILE_EXTENDED: + mtk_venc_err(inst->ctx, "unsupported EXTENDED"); + return 0; + default: + mtk_venc_debug(inst->ctx, "unsupported profile %d", profile); + return 100; + } +} + +static unsigned int h264_get_level(struct venc_h264_inst *inst, + unsigned int level) +{ + switch (level) { + case V4L2_MPEG_VIDEO_H264_LEVEL_1B: + mtk_venc_err(inst->ctx, "unsupported 1B"); + return 0; + case V4L2_MPEG_VIDEO_H264_LEVEL_1_0: + return 10; + case V4L2_MPEG_VIDEO_H264_LEVEL_1_1: + return 11; + case V4L2_MPEG_VIDEO_H264_LEVEL_1_2: + return 12; + case V4L2_MPEG_VIDEO_H264_LEVEL_1_3: + return 13; + case V4L2_MPEG_VIDEO_H264_LEVEL_2_0: + return 20; + case V4L2_MPEG_VIDEO_H264_LEVEL_2_1: + return 21; + case V4L2_MPEG_VIDEO_H264_LEVEL_2_2: + return 22; + case V4L2_MPEG_VIDEO_H264_LEVEL_3_0: + return 30; + case V4L2_MPEG_VIDEO_H264_LEVEL_3_1: + return 31; + case V4L2_MPEG_VIDEO_H264_LEVEL_3_2: + return 32; + case V4L2_MPEG_VIDEO_H264_LEVEL_4_0: + return 40; + case V4L2_MPEG_VIDEO_H264_LEVEL_4_1: + return 41; + case V4L2_MPEG_VIDEO_H264_LEVEL_4_2: + return 42; + case V4L2_MPEG_VIDEO_H264_LEVEL_5_0: + return 50; + case V4L2_MPEG_VIDEO_H264_LEVEL_5_1: + return 51; + default: + mtk_venc_debug(inst->ctx, "unsupported level %d", level); + return 31; + } +} + +static void h264_enc_free_work_buf(struct venc_h264_inst *inst) +{ + int i; + + /* Except the SKIP_FRAME buffers, + * other buffers need to be freed by AP. + */ + for (i = 0; i < VENC_H264_VPU_WORK_BUF_MAX; i++) { + if (i != VENC_H264_VPU_WORK_BUF_SKIP_FRAME) + mtk_vcodec_mem_free(inst->ctx, &inst->work_bufs[i]); + } + + mtk_vcodec_mem_free(inst->ctx, &inst->pps_buf); +} + +static int h264_enc_alloc_work_buf(struct venc_h264_inst *inst, bool is_34bit) +{ + struct venc_h264_vpu_buf *wb = NULL; + struct venc_h264_vpu_buf_34 *wb_34 = NULL; + int i; + u32 vpua, wb_size; + int ret = 0; + + if (is_34bit) + wb_34 = inst->vsi_34->work_bufs; + else + wb = inst->vsi->work_bufs; + + for (i = 0; i < VENC_H264_VPU_WORK_BUF_MAX; i++) { + /* + * This 'wb' structure is set by VPU side and shared to AP for + * buffer allocation and IO virtual addr mapping. For most of + * the buffers, AP will allocate the buffer according to 'size' + * field and store the IO virtual addr in 'iova' field. There + * are two exceptions: + * (1) RC_CODE buffer, it's pre-allocated in the VPU side, and + * save the VPU addr in the 'vpua' field. The AP will translate + * the VPU addr to the corresponding IO virtual addr and store + * in 'iova' field for reg setting in VPU side. + * (2) SKIP_FRAME buffer, it's pre-allocated in the VPU side, + * and save the VPU addr in the 'vpua' field. The AP will + * translate the VPU addr to the corresponding AP side virtual + * address and do some memcpy access to move to bitstream buffer + * assigned by v4l2 layer. + */ + if (is_34bit) { + inst->work_bufs[i].size = wb_34[i].size; + vpua = wb_34[i].vpua; + wb_size = wb_34[i].size; + } else { + inst->work_bufs[i].size = wb[i].size; + vpua = wb[i].vpua; + wb_size = wb[i].size; + } + + if (i == VENC_H264_VPU_WORK_BUF_SKIP_FRAME) { + struct mtk_vcodec_fw *handler; + + handler = inst->vpu_inst.ctx->dev->fw_handler; + inst->work_bufs[i].va = + mtk_vcodec_fw_map_dm_addr(handler, vpua); + inst->work_bufs[i].dma_addr = 0; + } else { + ret = mtk_vcodec_mem_alloc(inst->ctx, + &inst->work_bufs[i]); + if (ret) { + mtk_venc_err(inst->ctx, "cannot allocate buf %d", i); + goto err_alloc; + } + /* + * This RC_CODE is pre-allocated by VPU and saved in VPU + * addr. So we need use memcpy to copy RC_CODE from VPU + * addr into IO virtual addr in 'iova' field for reg + * setting in VPU side. + */ + if (i == VENC_H264_VPU_WORK_BUF_RC_CODE) { + struct mtk_vcodec_fw *handler; + void *tmp_va; + + handler = inst->vpu_inst.ctx->dev->fw_handler; + tmp_va = mtk_vcodec_fw_map_dm_addr(handler, + vpua); + memcpy(inst->work_bufs[i].va, tmp_va, wb_size); + } + } + if (is_34bit) + wb_34[i].iova = inst->work_bufs[i].dma_addr; + else + wb[i].iova = inst->work_bufs[i].dma_addr; + + mtk_venc_debug(inst->ctx, "work_buf[%d] va=0x%p iova=%pad size=%zu", + i, inst->work_bufs[i].va, + &inst->work_bufs[i].dma_addr, + inst->work_bufs[i].size); + } + + /* the pps_buf is used by AP side only */ + inst->pps_buf.size = 128; + ret = mtk_vcodec_mem_alloc(inst->ctx, &inst->pps_buf); + if (ret) { + mtk_venc_err(inst->ctx, "cannot allocate pps_buf"); + goto err_alloc; + } + + return ret; + +err_alloc: + h264_enc_free_work_buf(inst); + + return ret; +} + +static unsigned int h264_enc_wait_venc_done(struct venc_h264_inst *inst) +{ + unsigned int irq_status = 0; + struct mtk_vcodec_enc_ctx *ctx = (struct mtk_vcodec_enc_ctx *)inst->ctx; + + if (!mtk_vcodec_wait_for_done_ctx(ctx, MTK_INST_IRQ_RECEIVED, + WAIT_INTR_TIMEOUT_MS, 0)) { + irq_status = ctx->irq_status; + mtk_venc_debug(ctx, "irq_status %x <-", irq_status); + } + return irq_status; +} + +static int h264_frame_type(unsigned int frm_cnt, unsigned int gop_size, + unsigned int intra_period) +{ + if ((gop_size != 0 && (frm_cnt % gop_size) == 0) || + (frm_cnt == 0 && gop_size == 0)) { + /* IDR frame */ + return VENC_H264_IDR_FRM; + } else if ((intra_period != 0 && (frm_cnt % intra_period) == 0) || + (frm_cnt == 0 && intra_period == 0)) { + /* I frame */ + return VENC_H264_I_FRM; + } else { + return VENC_H264_P_FRM; /* Note: B frames are not supported */ + } +} + +static int h264_encode_sps(struct venc_h264_inst *inst, + struct mtk_vcodec_mem *bs_buf, + unsigned int *bs_size) +{ + int ret = 0; + unsigned int irq_status; + + ret = vpu_enc_encode(&inst->vpu_inst, H264_BS_MODE_SPS, NULL, bs_buf, NULL); + if (ret) + return ret; + + irq_status = h264_enc_wait_venc_done(inst); + if (irq_status != MTK_VENC_IRQ_STATUS_SPS) { + mtk_venc_err(inst->ctx, "expect irq status %d", MTK_VENC_IRQ_STATUS_SPS); + return -EINVAL; + } + + *bs_size = h264_read_reg(inst, VENC_PIC_BITSTREAM_BYTE_CNT); + mtk_venc_debug(inst->ctx, "bs size %d <-", *bs_size); + + return ret; +} + +static int h264_encode_pps(struct venc_h264_inst *inst, + struct mtk_vcodec_mem *bs_buf, + unsigned int *bs_size) +{ + int ret = 0; + unsigned int irq_status; + + ret = vpu_enc_encode(&inst->vpu_inst, H264_BS_MODE_PPS, NULL, bs_buf, NULL); + if (ret) + return ret; + + irq_status = h264_enc_wait_venc_done(inst); + if (irq_status != MTK_VENC_IRQ_STATUS_PPS) { + mtk_venc_err(inst->ctx, "expect irq status %d", MTK_VENC_IRQ_STATUS_PPS); + return -EINVAL; + } + + *bs_size = h264_read_reg(inst, VENC_PIC_BITSTREAM_BYTE_CNT); + mtk_venc_debug(inst->ctx, "bs size %d <-", *bs_size); + + return ret; +} + +static int h264_encode_header(struct venc_h264_inst *inst, + struct mtk_vcodec_mem *bs_buf, + unsigned int *bs_size) +{ + int ret = 0; + unsigned int bs_size_sps; + unsigned int bs_size_pps; + + ret = h264_encode_sps(inst, bs_buf, &bs_size_sps); + if (ret) + return ret; + + ret = h264_encode_pps(inst, &inst->pps_buf, &bs_size_pps); + if (ret) + return ret; + + memcpy(bs_buf->va + bs_size_sps, inst->pps_buf.va, bs_size_pps); + *bs_size = bs_size_sps + bs_size_pps; + + return ret; +} + +static int h264_encode_frame(struct venc_h264_inst *inst, + struct venc_frm_buf *frm_buf, + struct mtk_vcodec_mem *bs_buf, + unsigned int *bs_size) +{ + int ret = 0; + unsigned int gop_size; + unsigned int intra_period; + unsigned int irq_status; + struct venc_frame_info frame_info; + struct mtk_vcodec_enc_ctx *ctx = inst->ctx; + + mtk_venc_debug(ctx, "frm_cnt = %d\n ", inst->frm_cnt); + + if (MTK_ENC_IOVA_IS_34BIT(ctx)) { + gop_size = inst->vsi_34->config.gop_size; + intra_period = inst->vsi_34->config.intra_period; + } else { + gop_size = inst->vsi->config.gop_size; + intra_period = inst->vsi->config.intra_period; + } + frame_info.frm_count = inst->frm_cnt; + frame_info.skip_frm_count = inst->skip_frm_cnt; + frame_info.frm_type = h264_frame_type(inst->frm_cnt, gop_size, + intra_period); + mtk_venc_debug(ctx, "frm_count = %d,skip_frm_count =%d,frm_type=%d.\n", + frame_info.frm_count, frame_info.skip_frm_count, + frame_info.frm_type); + + ret = vpu_enc_encode(&inst->vpu_inst, H264_BS_MODE_FRAME, + frm_buf, bs_buf, &frame_info); + if (ret) + return ret; + + /* + * skip frame case: The skip frame buffer is composed by vpu side only, + * it does not trigger the hw, so skip the wait interrupt operation. + */ + if (inst->vpu_inst.state == VEN_IPI_MSG_ENC_STATE_SKIP) { + *bs_size = inst->vpu_inst.bs_size; + memcpy(bs_buf->va, + inst->work_bufs[VENC_H264_VPU_WORK_BUF_SKIP_FRAME].va, + *bs_size); + ++inst->frm_cnt; + ++inst->skip_frm_cnt; + return 0; + } + + irq_status = h264_enc_wait_venc_done(inst); + if (irq_status != MTK_VENC_IRQ_STATUS_FRM) { + mtk_venc_err(ctx, "irq_status=%d failed", irq_status); + return -EIO; + } + + *bs_size = h264_read_reg(inst, VENC_PIC_BITSTREAM_BYTE_CNT); + + ++inst->frm_cnt; + mtk_venc_debug(ctx, "frm %d bs_size %d key_frm %d <-", + inst->frm_cnt, *bs_size, inst->vpu_inst.is_key_frm); + + return 0; +} + +static void h264_encode_filler(struct venc_h264_inst *inst, void *buf, + int size) +{ + unsigned char *p = buf; + + if (size < H264_FILLER_MARKER_SIZE) { + mtk_venc_err(inst->ctx, "filler size too small %d", size); + return; + } + + memcpy(p, h264_filler_marker, ARRAY_SIZE(h264_filler_marker)); + size -= H264_FILLER_MARKER_SIZE; + p += H264_FILLER_MARKER_SIZE; + memset(p, 0xff, size); +} + +static int h264_enc_init(struct mtk_vcodec_enc_ctx *ctx) +{ + const bool is_ext = MTK_ENC_CTX_IS_EXT(ctx); + int ret = 0; + struct venc_h264_inst *inst; + + inst = kzalloc(sizeof(*inst), GFP_KERNEL); + if (!inst) + return -ENOMEM; + + inst->ctx = ctx; + inst->vpu_inst.ctx = ctx; + inst->vpu_inst.id = is_ext ? SCP_IPI_VENC_H264 : IPI_VENC_H264; + inst->hw_base = mtk_vcodec_get_reg_addr(inst->ctx->dev->reg_base, VENC_SYS); + + ret = vpu_enc_init(&inst->vpu_inst); + + if (MTK_ENC_IOVA_IS_34BIT(ctx)) + inst->vsi_34 = (struct venc_h264_vsi_34 *)inst->vpu_inst.vsi; + else + inst->vsi = (struct venc_h264_vsi *)inst->vpu_inst.vsi; + + if (ret) + kfree(inst); + else + ctx->drv_handle = inst; + + return ret; +} + +static int h264_enc_encode(void *handle, + enum venc_start_opt opt, + struct venc_frm_buf *frm_buf, + struct mtk_vcodec_mem *bs_buf, + struct venc_done_result *result) +{ + int ret = 0; + struct venc_h264_inst *inst = (struct venc_h264_inst *)handle; + struct mtk_vcodec_enc_ctx *ctx = inst->ctx; + + mtk_venc_debug(ctx, "opt %d ->", opt); + + enable_irq(ctx->dev->enc_irq); + + switch (opt) { + case VENC_START_OPT_ENCODE_SEQUENCE_HEADER: { + unsigned int bs_size_hdr; + + ret = h264_encode_header(inst, bs_buf, &bs_size_hdr); + if (ret) + goto encode_err; + + result->bs_size = bs_size_hdr; + result->is_key_frm = false; + break; + } + + case VENC_START_OPT_ENCODE_FRAME: { + int hdr_sz; + int hdr_sz_ext; + int filler_sz = 0; + const int bs_alignment = 128; + struct mtk_vcodec_mem tmp_bs_buf; + unsigned int bs_size_hdr; + unsigned int bs_size_frm; + + if (!inst->prepend_hdr) { + ret = h264_encode_frame(inst, frm_buf, bs_buf, + &result->bs_size); + if (ret) + goto encode_err; + result->is_key_frm = inst->vpu_inst.is_key_frm; + break; + } + + mtk_venc_debug(ctx, "h264_encode_frame prepend SPS/PPS"); + + ret = h264_encode_header(inst, bs_buf, &bs_size_hdr); + if (ret) + goto encode_err; + + hdr_sz = bs_size_hdr; + hdr_sz_ext = (hdr_sz & (bs_alignment - 1)); + if (hdr_sz_ext) { + filler_sz = bs_alignment - hdr_sz_ext; + if (hdr_sz_ext + H264_FILLER_MARKER_SIZE > bs_alignment) + filler_sz += bs_alignment; + h264_encode_filler(inst, bs_buf->va + hdr_sz, + filler_sz); + } + + tmp_bs_buf.va = bs_buf->va + hdr_sz + filler_sz; + tmp_bs_buf.dma_addr = bs_buf->dma_addr + hdr_sz + filler_sz; + tmp_bs_buf.size = bs_buf->size - (hdr_sz + filler_sz); + + ret = h264_encode_frame(inst, frm_buf, &tmp_bs_buf, + &bs_size_frm); + if (ret) + goto encode_err; + + result->bs_size = hdr_sz + filler_sz + bs_size_frm; + + mtk_venc_debug(ctx, "hdr %d filler %d frame %d bs %d", + hdr_sz, filler_sz, bs_size_frm, result->bs_size); + + inst->prepend_hdr = 0; + result->is_key_frm = inst->vpu_inst.is_key_frm; + break; + } + + default: + mtk_venc_err(ctx, "venc_start_opt %d not supported", opt); + ret = -EINVAL; + break; + } + +encode_err: + + disable_irq(ctx->dev->enc_irq); + mtk_venc_debug(ctx, "opt %d <-", opt); + + return ret; +} + +static void h264_enc_set_vsi_configs(struct venc_h264_inst *inst, + struct venc_enc_param *enc_prm) +{ + inst->vsi->config.input_fourcc = enc_prm->input_yuv_fmt; + inst->vsi->config.bitrate = enc_prm->bitrate; + inst->vsi->config.pic_w = enc_prm->width; + inst->vsi->config.pic_h = enc_prm->height; + inst->vsi->config.buf_w = enc_prm->buf_width; + inst->vsi->config.buf_h = enc_prm->buf_height; + inst->vsi->config.gop_size = enc_prm->gop_size; + inst->vsi->config.framerate = enc_prm->frm_rate; + inst->vsi->config.intra_period = enc_prm->intra_period; + inst->vsi->config.profile = + h264_get_profile(inst, enc_prm->h264_profile); + inst->vsi->config.level = + h264_get_level(inst, enc_prm->h264_level); + inst->vsi->config.wfd = 0; +} + +static void h264_enc_set_vsi_34_configs(struct venc_h264_inst *inst, + struct venc_enc_param *enc_prm) +{ + inst->vsi_34->config.input_fourcc = enc_prm->input_yuv_fmt; + inst->vsi_34->config.bitrate = enc_prm->bitrate; + inst->vsi_34->config.pic_w = enc_prm->width; + inst->vsi_34->config.pic_h = enc_prm->height; + inst->vsi_34->config.buf_w = enc_prm->buf_width; + inst->vsi_34->config.buf_h = enc_prm->buf_height; + inst->vsi_34->config.gop_size = enc_prm->gop_size; + inst->vsi_34->config.framerate = enc_prm->frm_rate; + inst->vsi_34->config.intra_period = enc_prm->intra_period; + inst->vsi_34->config.profile = + h264_get_profile(inst, enc_prm->h264_profile); + inst->vsi_34->config.level = + h264_get_level(inst, enc_prm->h264_level); + inst->vsi_34->config.wfd = 0; +} + +static int h264_enc_set_param(void *handle, + enum venc_set_param_type type, + struct venc_enc_param *enc_prm) +{ + int ret = 0; + struct venc_h264_inst *inst = (struct venc_h264_inst *)handle; + struct mtk_vcodec_enc_ctx *ctx = inst->ctx; + const bool is_34bit = MTK_ENC_IOVA_IS_34BIT(ctx); + + mtk_venc_debug(ctx, "->type=%d", type); + + switch (type) { + case VENC_SET_PARAM_ENC: + if (is_34bit) + h264_enc_set_vsi_34_configs(inst, enc_prm); + else + h264_enc_set_vsi_configs(inst, enc_prm); + ret = vpu_enc_set_param(&inst->vpu_inst, type, enc_prm); + if (ret) + break; + if (inst->work_buf_allocated) { + h264_enc_free_work_buf(inst); + inst->work_buf_allocated = false; + } + ret = h264_enc_alloc_work_buf(inst, is_34bit); + if (ret) + break; + inst->work_buf_allocated = true; + break; + + case VENC_SET_PARAM_PREPEND_HEADER: + inst->prepend_hdr = 1; + mtk_venc_debug(ctx, "set prepend header mode"); + break; + case VENC_SET_PARAM_FORCE_INTRA: + case VENC_SET_PARAM_GOP_SIZE: + case VENC_SET_PARAM_INTRA_PERIOD: + inst->frm_cnt = 0; + inst->skip_frm_cnt = 0; + fallthrough; + default: + ret = vpu_enc_set_param(&inst->vpu_inst, type, enc_prm); + break; + } + + return ret; +} + +static int h264_enc_deinit(void *handle) +{ + int ret = 0; + struct venc_h264_inst *inst = (struct venc_h264_inst *)handle; + + ret = vpu_enc_deinit(&inst->vpu_inst); + + if (inst->work_buf_allocated) + h264_enc_free_work_buf(inst); + + kfree(inst); + + return ret; +} + +const struct venc_common_if venc_h264_if = { + .init = h264_enc_init, + .encode = h264_enc_encode, + .set_param = h264_enc_set_param, + .deinit = h264_enc_deinit, +}; diff --git a/drivers/media/platform/mediatek/vcodec/encoder/venc/venc_vp8_if.c b/drivers/media/platform/mediatek/vcodec/encoder/venc/venc_vp8_if.c new file mode 100644 index 0000000000..05abca91e7 --- /dev/null +++ b/drivers/media/platform/mediatek/vcodec/encoder/venc/venc_vp8_if.c @@ -0,0 +1,441 @@ +// SPDX-License-Identifier: GPL-2.0 +/* + * Copyright (c) 2016 MediaTek Inc. + * Author: Daniel Hsiao <daniel.hsiao@mediatek.com> + * PoChun Lin <pochun.lin@mediatek.com> + */ + +#include <linux/interrupt.h> +#include <linux/kernel.h> +#include <linux/slab.h> + +#include "../mtk_vcodec_enc_drv.h" +#include "../../common/mtk_vcodec_intr.h" +#include "../mtk_vcodec_enc.h" +#include "../mtk_vcodec_enc_pm.h" +#include "../venc_drv_base.h" +#include "../venc_ipi_msg.h" +#include "../venc_vpu_if.h" + +#define VENC_BITSTREAM_FRAME_SIZE 0x0098 +#define VENC_BITSTREAM_HEADER_LEN 0x00e8 + +/* This ac_tag is vp8 frame tag. */ +#define MAX_AC_TAG_SIZE 10 + +/* + * enum venc_vp8_vpu_work_buf - vp8 encoder buffer index + */ +enum venc_vp8_vpu_work_buf { + VENC_VP8_VPU_WORK_BUF_LUMA, + VENC_VP8_VPU_WORK_BUF_LUMA2, + VENC_VP8_VPU_WORK_BUF_LUMA3, + VENC_VP8_VPU_WORK_BUF_CHROMA, + VENC_VP8_VPU_WORK_BUF_CHROMA2, + VENC_VP8_VPU_WORK_BUF_CHROMA3, + VENC_VP8_VPU_WORK_BUF_MV_INFO, + VENC_VP8_VPU_WORK_BUF_BS_HEADER, + VENC_VP8_VPU_WORK_BUF_PROB_BUF, + VENC_VP8_VPU_WORK_BUF_RC_INFO, + VENC_VP8_VPU_WORK_BUF_RC_CODE, + VENC_VP8_VPU_WORK_BUF_RC_CODE2, + VENC_VP8_VPU_WORK_BUF_RC_CODE3, + VENC_VP8_VPU_WORK_BUF_MAX, +}; + +/* + * struct venc_vp8_vpu_config - Structure for vp8 encoder configuration + * AP-W/R : AP is writer/reader on this item + * VPU-W/R: VPU is write/reader on this item + * @input_fourcc: input fourcc + * @bitrate: target bitrate (in bps) + * @pic_w: picture width. Picture size is visible stream resolution, in pixels, + * to be used for display purposes; must be smaller or equal to buffer + * size. + * @pic_h: picture height + * @buf_w: buffer width (with 16 alignment). Buffer size is stream resolution + * in pixels aligned to hardware requirements. + * @buf_h: buffer height (with 16 alignment) + * @gop_size: group of picture size (key frame) + * @framerate: frame rate in fps + * @ts_mode: temporal scalability mode (0: disable, 1: enable) + * support three temporal layers - 0: 7.5fps 1: 7.5fps 2: 15fps. + */ +struct venc_vp8_vpu_config { + u32 input_fourcc; + u32 bitrate; + u32 pic_w; + u32 pic_h; + u32 buf_w; + u32 buf_h; + u32 gop_size; + u32 framerate; + u32 ts_mode; +}; + +/* + * struct venc_vp8_vpu_buf - Structure for buffer information + * AP-W/R : AP is writer/reader on this item + * VPU-W/R: VPU is write/reader on this item + * @iova: IO virtual address + * @vpua: VPU side memory addr which is used by RC_CODE + * @size: buffer size (in bytes) + */ +struct venc_vp8_vpu_buf { + u32 iova; + u32 vpua; + u32 size; +}; + +/* + * struct venc_vp8_vsi - Structure for VPU driver control and info share + * AP-W/R : AP is writer/reader on this item + * VPU-W/R: VPU is write/reader on this item + * This structure is allocated in VPU side and shared to AP side. + * @config: vp8 encoder configuration + * @work_bufs: working buffer information in VPU side + * The work_bufs here is for storing the 'size' info shared to AP side. + * The similar item in struct venc_vp8_inst is for memory allocation + * in AP side. The AP driver will copy the 'size' from here to the one in + * struct mtk_vcodec_mem, then invoke mtk_vcodec_mem_alloc to allocate + * the buffer. After that, bypass the 'dma_addr' to the 'iova' field here for + * register setting in VPU side. + */ +struct venc_vp8_vsi { + struct venc_vp8_vpu_config config; + struct venc_vp8_vpu_buf work_bufs[VENC_VP8_VPU_WORK_BUF_MAX]; +}; + +/* + * struct venc_vp8_inst - vp8 encoder AP driver instance + * @hw_base: vp8 encoder hardware register base + * @work_bufs: working buffer + * @work_buf_allocated: working buffer allocated flag + * @frm_cnt: encoded frame count, it's used for I-frame judgement and + * reset when force intra cmd received. + * @ts_mode: temporal scalability mode (0: disable, 1: enable) + * support three temporal layers - 0: 7.5fps 1: 7.5fps 2: 15fps. + * @vpu_inst: VPU instance to exchange information between AP and VPU + * @vsi: driver structure allocated by VPU side and shared to AP side for + * control and info share + * @ctx: context for v4l2 layer integration + */ +struct venc_vp8_inst { + void __iomem *hw_base; + struct mtk_vcodec_mem work_bufs[VENC_VP8_VPU_WORK_BUF_MAX]; + bool work_buf_allocated; + unsigned int frm_cnt; + unsigned int ts_mode; + struct venc_vpu_inst vpu_inst; + struct venc_vp8_vsi *vsi; + struct mtk_vcodec_enc_ctx *ctx; +}; + +static inline u32 vp8_enc_read_reg(struct venc_vp8_inst *inst, u32 addr) +{ + return readl(inst->hw_base + addr); +} + +static void vp8_enc_free_work_buf(struct venc_vp8_inst *inst) +{ + int i; + + /* Buffers need to be freed by AP. */ + for (i = 0; i < VENC_VP8_VPU_WORK_BUF_MAX; i++) { + if (inst->work_bufs[i].size == 0) + continue; + mtk_vcodec_mem_free(inst->ctx, &inst->work_bufs[i]); + } +} + +static int vp8_enc_alloc_work_buf(struct venc_vp8_inst *inst) +{ + int i; + int ret = 0; + struct venc_vp8_vpu_buf *wb = inst->vsi->work_bufs; + + for (i = 0; i < VENC_VP8_VPU_WORK_BUF_MAX; i++) { + if (wb[i].size == 0) + continue; + /* + * This 'wb' structure is set by VPU side and shared to AP for + * buffer allocation and IO virtual addr mapping. For most of + * the buffers, AP will allocate the buffer according to 'size' + * field and store the IO virtual addr in 'iova' field. For the + * RC_CODEx buffers, they are pre-allocated in the VPU side + * because they are inside VPU SRAM, and save the VPU addr in + * the 'vpua' field. The AP will translate the VPU addr to the + * corresponding IO virtual addr and store in 'iova' field. + */ + inst->work_bufs[i].size = wb[i].size; + ret = mtk_vcodec_mem_alloc(inst->ctx, &inst->work_bufs[i]); + if (ret) { + mtk_venc_err(inst->ctx, "cannot alloc work_bufs[%d]", i); + goto err_alloc; + } + /* + * This RC_CODEx is pre-allocated by VPU and saved in VPU addr. + * So we need use memcpy to copy RC_CODEx from VPU addr into IO + * virtual addr in 'iova' field for reg setting in VPU side. + */ + if (i == VENC_VP8_VPU_WORK_BUF_RC_CODE || + i == VENC_VP8_VPU_WORK_BUF_RC_CODE2 || + i == VENC_VP8_VPU_WORK_BUF_RC_CODE3) { + struct mtk_vcodec_fw *handler; + void *tmp_va; + + handler = inst->vpu_inst.ctx->dev->fw_handler; + tmp_va = mtk_vcodec_fw_map_dm_addr(handler, + wb[i].vpua); + memcpy(inst->work_bufs[i].va, tmp_va, wb[i].size); + } + wb[i].iova = inst->work_bufs[i].dma_addr; + + mtk_venc_debug(inst->ctx, "work_bufs[%d] va=0x%p,iova=%pad,size=%zu", + i, inst->work_bufs[i].va, + &inst->work_bufs[i].dma_addr, + inst->work_bufs[i].size); + } + + return ret; + +err_alloc: + vp8_enc_free_work_buf(inst); + + return ret; +} + +static unsigned int vp8_enc_wait_venc_done(struct venc_vp8_inst *inst) +{ + unsigned int irq_status = 0; + struct mtk_vcodec_enc_ctx *ctx = (struct mtk_vcodec_enc_ctx *)inst->ctx; + + if (!mtk_vcodec_wait_for_done_ctx(ctx, MTK_INST_IRQ_RECEIVED, + WAIT_INTR_TIMEOUT_MS, 0)) { + irq_status = ctx->irq_status; + mtk_venc_debug(ctx, "isr return %x", irq_status); + } + return irq_status; +} + +/* + * Compose ac_tag, bitstream header and bitstream payload into + * one bitstream buffer. + */ +static int vp8_enc_compose_one_frame(struct venc_vp8_inst *inst, + struct mtk_vcodec_mem *bs_buf, + unsigned int *bs_size) +{ + unsigned int not_key; + u32 bs_frm_size; + u32 bs_hdr_len; + unsigned int ac_tag_size; + u8 ac_tag[MAX_AC_TAG_SIZE]; + u32 tag; + + bs_frm_size = vp8_enc_read_reg(inst, VENC_BITSTREAM_FRAME_SIZE); + bs_hdr_len = vp8_enc_read_reg(inst, VENC_BITSTREAM_HEADER_LEN); + + /* if a frame is key frame, not_key is 0 */ + not_key = !inst->vpu_inst.is_key_frm; + tag = (bs_hdr_len << 5) | 0x10 | not_key; + ac_tag[0] = tag & 0xff; + ac_tag[1] = (tag >> 8) & 0xff; + ac_tag[2] = (tag >> 16) & 0xff; + + /* key frame */ + if (not_key == 0) { + ac_tag_size = MAX_AC_TAG_SIZE; + ac_tag[3] = 0x9d; + ac_tag[4] = 0x01; + ac_tag[5] = 0x2a; + ac_tag[6] = inst->vsi->config.pic_w; + ac_tag[7] = inst->vsi->config.pic_w >> 8; + ac_tag[8] = inst->vsi->config.pic_h; + ac_tag[9] = inst->vsi->config.pic_h >> 8; + } else { + ac_tag_size = 3; + } + + if (bs_buf->size < bs_hdr_len + bs_frm_size + ac_tag_size) { + mtk_venc_err(inst->ctx, "bitstream buf size is too small(%zu)", bs_buf->size); + return -EINVAL; + } + + /* + * (1) The vp8 bitstream header and body are generated by the HW vp8 + * encoder separately at the same time. We cannot know the bitstream + * header length in advance. + * (2) From the vp8 spec, there is no stuffing byte allowed between the + * ac tag, bitstream header and bitstream body. + */ + memmove(bs_buf->va + bs_hdr_len + ac_tag_size, + bs_buf->va, bs_frm_size); + memcpy(bs_buf->va + ac_tag_size, + inst->work_bufs[VENC_VP8_VPU_WORK_BUF_BS_HEADER].va, + bs_hdr_len); + memcpy(bs_buf->va, ac_tag, ac_tag_size); + *bs_size = bs_frm_size + bs_hdr_len + ac_tag_size; + + return 0; +} + +static int vp8_enc_encode_frame(struct venc_vp8_inst *inst, + struct venc_frm_buf *frm_buf, + struct mtk_vcodec_mem *bs_buf, + unsigned int *bs_size) +{ + int ret = 0; + unsigned int irq_status; + + mtk_venc_debug(inst->ctx, "->frm_cnt=%d", inst->frm_cnt); + + ret = vpu_enc_encode(&inst->vpu_inst, 0, frm_buf, bs_buf, NULL); + if (ret) + return ret; + + irq_status = vp8_enc_wait_venc_done(inst); + if (irq_status != MTK_VENC_IRQ_STATUS_FRM) { + mtk_venc_err(inst->ctx, "irq_status=%d failed", irq_status); + return -EIO; + } + + if (vp8_enc_compose_one_frame(inst, bs_buf, bs_size)) { + mtk_venc_err(inst->ctx, "vp8_enc_compose_one_frame failed"); + return -EINVAL; + } + + inst->frm_cnt++; + mtk_venc_debug(inst->ctx, "<-size=%d key_frm=%d", *bs_size, inst->vpu_inst.is_key_frm); + + return ret; +} + +static int vp8_enc_init(struct mtk_vcodec_enc_ctx *ctx) +{ + int ret = 0; + struct venc_vp8_inst *inst; + + inst = kzalloc(sizeof(*inst), GFP_KERNEL); + if (!inst) + return -ENOMEM; + + inst->ctx = ctx; + inst->vpu_inst.ctx = ctx; + inst->vpu_inst.id = IPI_VENC_VP8; + inst->hw_base = mtk_vcodec_get_reg_addr(inst->ctx->dev->reg_base, VENC_LT_SYS); + + ret = vpu_enc_init(&inst->vpu_inst); + + inst->vsi = (struct venc_vp8_vsi *)inst->vpu_inst.vsi; + + if (ret) + kfree(inst); + else + ctx->drv_handle = inst; + + return ret; +} + +static int vp8_enc_encode(void *handle, + enum venc_start_opt opt, + struct venc_frm_buf *frm_buf, + struct mtk_vcodec_mem *bs_buf, + struct venc_done_result *result) +{ + int ret = 0; + struct venc_vp8_inst *inst = (struct venc_vp8_inst *)handle; + struct mtk_vcodec_enc_ctx *ctx = inst->ctx; + + enable_irq(ctx->dev->enc_irq); + + switch (opt) { + case VENC_START_OPT_ENCODE_FRAME: + ret = vp8_enc_encode_frame(inst, frm_buf, bs_buf, + &result->bs_size); + if (ret) + goto encode_err; + result->is_key_frm = inst->vpu_inst.is_key_frm; + break; + + default: + mtk_venc_err(ctx, "opt not support:%d", opt); + ret = -EINVAL; + break; + } + +encode_err: + + disable_irq(ctx->dev->enc_irq); + return ret; +} + +static int vp8_enc_set_param(void *handle, + enum venc_set_param_type type, + struct venc_enc_param *enc_prm) +{ + int ret = 0; + struct venc_vp8_inst *inst = (struct venc_vp8_inst *)handle; + + mtk_venc_debug(inst->ctx, "->type=%d", type); + + switch (type) { + case VENC_SET_PARAM_ENC: + inst->vsi->config.input_fourcc = enc_prm->input_yuv_fmt; + inst->vsi->config.bitrate = enc_prm->bitrate; + inst->vsi->config.pic_w = enc_prm->width; + inst->vsi->config.pic_h = enc_prm->height; + inst->vsi->config.buf_w = enc_prm->buf_width; + inst->vsi->config.buf_h = enc_prm->buf_height; + inst->vsi->config.gop_size = enc_prm->gop_size; + inst->vsi->config.framerate = enc_prm->frm_rate; + inst->vsi->config.ts_mode = inst->ts_mode; + ret = vpu_enc_set_param(&inst->vpu_inst, type, enc_prm); + if (ret) + break; + if (inst->work_buf_allocated) { + vp8_enc_free_work_buf(inst); + inst->work_buf_allocated = false; + } + ret = vp8_enc_alloc_work_buf(inst); + if (ret) + break; + inst->work_buf_allocated = true; + break; + + /* + * VENC_SET_PARAM_TS_MODE must be called before VENC_SET_PARAM_ENC + */ + case VENC_SET_PARAM_TS_MODE: + inst->ts_mode = 1; + mtk_venc_debug(inst->ctx, "set ts_mode"); + break; + + default: + ret = vpu_enc_set_param(&inst->vpu_inst, type, enc_prm); + break; + } + + return ret; +} + +static int vp8_enc_deinit(void *handle) +{ + int ret = 0; + struct venc_vp8_inst *inst = (struct venc_vp8_inst *)handle; + + ret = vpu_enc_deinit(&inst->vpu_inst); + + if (inst->work_buf_allocated) + vp8_enc_free_work_buf(inst); + + kfree(inst); + return ret; +} + +const struct venc_common_if venc_vp8_if = { + .init = vp8_enc_init, + .encode = vp8_enc_encode, + .set_param = vp8_enc_set_param, + .deinit = vp8_enc_deinit, +}; diff --git a/drivers/media/platform/mediatek/vcodec/encoder/venc_drv_base.h b/drivers/media/platform/mediatek/vcodec/encoder/venc_drv_base.h new file mode 100644 index 0000000000..856d50151b --- /dev/null +++ b/drivers/media/platform/mediatek/vcodec/encoder/venc_drv_base.h @@ -0,0 +1,53 @@ +/* SPDX-License-Identifier: GPL-2.0 */ +/* + * Copyright (c) 2016 MediaTek Inc. + * Author: Daniel Hsiao <daniel.hsiao@mediatek.com> + * Jungchang Tsao <jungchang.tsao@mediatek.com> + * Tiffany Lin <tiffany.lin@mediatek.com> + */ + +#ifndef _VENC_DRV_BASE_ +#define _VENC_DRV_BASE_ + +#include "mtk_vcodec_enc_drv.h" + +#include "venc_drv_if.h" + +struct venc_common_if { + /** + * (*init)() - initialize driver + * @ctx: [in] mtk v4l2 context + * @handle: [out] driver handle + */ + int (*init)(struct mtk_vcodec_enc_ctx *ctx); + + /** + * (*encode)() - trigger encode + * @handle: [in] driver handle + * @opt: [in] encode option + * @frm_buf: [in] frame buffer to store input frame + * @bs_buf: [in] bitstream buffer to store output bitstream + * @result: [out] encode result + */ + int (*encode)(void *handle, enum venc_start_opt opt, + struct venc_frm_buf *frm_buf, + struct mtk_vcodec_mem *bs_buf, + struct venc_done_result *result); + + /** + * (*set_param)() - set driver's parameter + * @handle: [in] driver handle + * @type: [in] parameter type + * @in: [in] buffer to store the parameter + */ + int (*set_param)(void *handle, enum venc_set_param_type type, + struct venc_enc_param *in); + + /** + * (*deinit)() - deinitialize driver. + * @handle: [in] driver handle + */ + int (*deinit)(void *handle); +}; + +#endif diff --git a/drivers/media/platform/mediatek/vcodec/encoder/venc_drv_if.c b/drivers/media/platform/mediatek/vcodec/encoder/venc_drv_if.c new file mode 100644 index 0000000000..1bdaecdd64 --- /dev/null +++ b/drivers/media/platform/mediatek/vcodec/encoder/venc_drv_if.c @@ -0,0 +1,100 @@ +// SPDX-License-Identifier: GPL-2.0 +/* + * Copyright (c) 2016 MediaTek Inc. + * Author: Daniel Hsiao <daniel.hsiao@mediatek.com> + * Jungchang Tsao <jungchang.tsao@mediatek.com> + * Tiffany Lin <tiffany.lin@mediatek.com> + */ + +#include <linux/interrupt.h> +#include <linux/kernel.h> +#include <linux/slab.h> + +#include "venc_drv_base.h" +#include "venc_drv_if.h" + +#include "mtk_vcodec_enc.h" +#include "mtk_vcodec_enc_pm.h" + +int venc_if_init(struct mtk_vcodec_enc_ctx *ctx, unsigned int fourcc) +{ + int ret = 0; + + switch (fourcc) { + case V4L2_PIX_FMT_VP8: + ctx->enc_if = &venc_vp8_if; + break; + case V4L2_PIX_FMT_H264: + ctx->enc_if = &venc_h264_if; + break; + default: + return -EINVAL; + } + + mtk_venc_lock(ctx); + mtk_vcodec_enc_clock_on(&ctx->dev->pm); + ret = ctx->enc_if->init(ctx); + mtk_vcodec_enc_clock_off(&ctx->dev->pm); + mtk_venc_unlock(ctx); + + return ret; +} + +int venc_if_set_param(struct mtk_vcodec_enc_ctx *ctx, + enum venc_set_param_type type, struct venc_enc_param *in) +{ + int ret = 0; + + mtk_venc_lock(ctx); + mtk_vcodec_enc_clock_on(&ctx->dev->pm); + ret = ctx->enc_if->set_param(ctx->drv_handle, type, in); + mtk_vcodec_enc_clock_off(&ctx->dev->pm); + mtk_venc_unlock(ctx); + + return ret; +} + +int venc_if_encode(struct mtk_vcodec_enc_ctx *ctx, + enum venc_start_opt opt, struct venc_frm_buf *frm_buf, + struct mtk_vcodec_mem *bs_buf, + struct venc_done_result *result) +{ + int ret = 0; + unsigned long flags; + + mtk_venc_lock(ctx); + + spin_lock_irqsave(&ctx->dev->irqlock, flags); + ctx->dev->curr_ctx = ctx; + spin_unlock_irqrestore(&ctx->dev->irqlock, flags); + + mtk_vcodec_enc_clock_on(&ctx->dev->pm); + ret = ctx->enc_if->encode(ctx->drv_handle, opt, frm_buf, + bs_buf, result); + mtk_vcodec_enc_clock_off(&ctx->dev->pm); + + spin_lock_irqsave(&ctx->dev->irqlock, flags); + ctx->dev->curr_ctx = NULL; + spin_unlock_irqrestore(&ctx->dev->irqlock, flags); + + mtk_venc_unlock(ctx); + return ret; +} + +int venc_if_deinit(struct mtk_vcodec_enc_ctx *ctx) +{ + int ret = 0; + + if (!ctx->drv_handle) + return 0; + + mtk_venc_lock(ctx); + mtk_vcodec_enc_clock_on(&ctx->dev->pm); + ret = ctx->enc_if->deinit(ctx->drv_handle); + mtk_vcodec_enc_clock_off(&ctx->dev->pm); + mtk_venc_unlock(ctx); + + ctx->drv_handle = NULL; + + return ret; +} diff --git a/drivers/media/platform/mediatek/vcodec/encoder/venc_drv_if.h b/drivers/media/platform/mediatek/vcodec/encoder/venc_drv_if.h new file mode 100644 index 0000000000..d00fb68b82 --- /dev/null +++ b/drivers/media/platform/mediatek/vcodec/encoder/venc_drv_if.h @@ -0,0 +1,169 @@ +/* SPDX-License-Identifier: GPL-2.0 */ +/* + * Copyright (c) 2016 MediaTek Inc. + * Author: Daniel Hsiao <daniel.hsiao@mediatek.com> + * Jungchang Tsao <jungchang.tsao@mediatek.com> + * Tiffany Lin <tiffany.lin@mediatek.com> + */ + +#ifndef _VENC_DRV_IF_H_ +#define _VENC_DRV_IF_H_ + +#include "mtk_vcodec_enc_drv.h" + +/* + * enum venc_yuv_fmt - The type of input yuv format + * (VPU related: If you change the order, you must also update the VPU codes.) + * @VENC_YUV_FORMAT_I420: I420 YUV format + * @VENC_YUV_FORMAT_YV12: YV12 YUV format + * @VENC_YUV_FORMAT_NV12: NV12 YUV format + * @VENC_YUV_FORMAT_NV21: NV21 YUV format + */ +enum venc_yuv_fmt { + VENC_YUV_FORMAT_I420 = 3, + VENC_YUV_FORMAT_YV12 = 5, + VENC_YUV_FORMAT_NV12 = 6, + VENC_YUV_FORMAT_NV21 = 7, +}; + +/* + * enum venc_start_opt - encode frame option used in venc_if_encode() + * @VENC_START_OPT_ENCODE_SEQUENCE_HEADER: encode SPS/PPS for H264 + * @VENC_START_OPT_ENCODE_FRAME: encode normal frame + */ +enum venc_start_opt { + VENC_START_OPT_ENCODE_SEQUENCE_HEADER, + VENC_START_OPT_ENCODE_FRAME, +}; + +/* + * enum venc_set_param_type - The type of set parameter used in + * venc_if_set_param() + * (VPU related: If you change the order, you must also update the VPU codes.) + * @VENC_SET_PARAM_ENC: set encoder parameters + * @VENC_SET_PARAM_FORCE_INTRA: force an intra frame + * @VENC_SET_PARAM_ADJUST_BITRATE: adjust bitrate (in bps) + * @VENC_SET_PARAM_ADJUST_FRAMERATE: set frame rate + * @VENC_SET_PARAM_GOP_SIZE: set IDR interval + * @VENC_SET_PARAM_INTRA_PERIOD: set I frame interval + * @VENC_SET_PARAM_SKIP_FRAME: set H264 skip one frame + * @VENC_SET_PARAM_PREPEND_HEADER: set H264 prepend SPS/PPS before IDR + * @VENC_SET_PARAM_TS_MODE: set VP8 temporal scalability mode + */ +enum venc_set_param_type { + VENC_SET_PARAM_ENC, + VENC_SET_PARAM_FORCE_INTRA, + VENC_SET_PARAM_ADJUST_BITRATE, + VENC_SET_PARAM_ADJUST_FRAMERATE, + VENC_SET_PARAM_GOP_SIZE, + VENC_SET_PARAM_INTRA_PERIOD, + VENC_SET_PARAM_SKIP_FRAME, + VENC_SET_PARAM_PREPEND_HEADER, + VENC_SET_PARAM_TS_MODE, +}; + +/* + * struct venc_enc_prm - encoder settings for VENC_SET_PARAM_ENC used in + * venc_if_set_param() + * @input_fourcc: input yuv format + * @h264_profile: V4L2 defined H.264 profile + * @h264_level: V4L2 defined H.264 level + * @width: image width + * @height: image height + * @buf_width: buffer width + * @buf_height: buffer height + * @frm_rate: frame rate in fps + * @intra_period: intra frame period + * @bitrate: target bitrate in bps + * @gop_size: group of picture size + */ +struct venc_enc_param { + enum venc_yuv_fmt input_yuv_fmt; + unsigned int h264_profile; + unsigned int h264_level; + unsigned int width; + unsigned int height; + unsigned int buf_width; + unsigned int buf_height; + unsigned int frm_rate; + unsigned int intra_period; + unsigned int bitrate; + unsigned int gop_size; +}; + +/** + * struct venc_frame_info - per-frame information to pass to the firmware. + * + * @frm_count: sequential number for this frame + * @skip_frm_count: number of frames skipped so far while decoding + * @frm_type: type of the frame, from enum venc_h264_frame_type + */ +struct venc_frame_info { + unsigned int frm_count; /* per frame update */ + unsigned int skip_frm_count; /* per frame update */ + unsigned int frm_type; /* per frame update */ +}; + +/* + * struct venc_frm_buf - frame buffer information used in venc_if_encode() + * @fb_addr: plane frame buffer addresses + */ +struct venc_frm_buf { + struct mtk_vcodec_fb fb_addr[MTK_VCODEC_MAX_PLANES]; +}; + +/* + * struct venc_done_result - This is return information used in venc_if_encode() + * @bs_size: output bitstream size + * @is_key_frm: output is key frame or not + */ +struct venc_done_result { + unsigned int bs_size; + bool is_key_frm; +}; + +extern const struct venc_common_if venc_h264_if; +extern const struct venc_common_if venc_vp8_if; + +/* + * venc_if_init - Create the driver handle + * @ctx: device context + * @fourcc: encoder input format + * Return: 0 if creating handle successfully, otherwise it is failed. + */ +int venc_if_init(struct mtk_vcodec_enc_ctx *ctx, unsigned int fourcc); + +/* + * venc_if_deinit - Release the driver handle + * @ctx: device context + * Return: 0 if releasing handle successfully, otherwise it is failed. + */ +int venc_if_deinit(struct mtk_vcodec_enc_ctx *ctx); + +/* + * venc_if_set_param - Set parameter to driver + * @ctx: device context + * @type: parameter type + * @in: input parameter + * Return: 0 if setting param successfully, otherwise it is failed. + */ +int venc_if_set_param(struct mtk_vcodec_enc_ctx *ctx, + enum venc_set_param_type type, + struct venc_enc_param *in); + +/* + * venc_if_encode - Encode one frame + * @ctx: device context + * @opt: encode frame option + * @frm_buf: input frame buffer information + * @bs_buf: output bitstream buffer infomraiton + * @result: encode result + * Return: 0 if encoding frame successfully, otherwise it is failed. + */ +int venc_if_encode(struct mtk_vcodec_enc_ctx *ctx, + enum venc_start_opt opt, + struct venc_frm_buf *frm_buf, + struct mtk_vcodec_mem *bs_buf, + struct venc_done_result *result); + +#endif /* _VENC_DRV_IF_H_ */ diff --git a/drivers/media/platform/mediatek/vcodec/encoder/venc_ipi_msg.h b/drivers/media/platform/mediatek/vcodec/encoder/venc_ipi_msg.h new file mode 100644 index 0000000000..bb16d96a7f --- /dev/null +++ b/drivers/media/platform/mediatek/vcodec/encoder/venc_ipi_msg.h @@ -0,0 +1,244 @@ +/* SPDX-License-Identifier: GPL-2.0 */ +/* + * Copyright (c) 2016 MediaTek Inc. + * Author: Jungchang Tsao <jungchang.tsao@mediatek.com> + * Daniel Hsiao <daniel.hsiao@mediatek.com> + * Tiffany Lin <tiffany.lin@mediatek.com> + */ + +#ifndef _VENC_IPI_MSG_H_ +#define _VENC_IPI_MSG_H_ + +#define AP_IPIMSG_VENC_BASE 0xC000 +#define VPU_IPIMSG_VENC_BASE 0xD000 + +/* + * enum venc_ipi_msg_id - message id between AP and VPU + * (ipi stands for inter-processor interrupt) + * @AP_IPIMSG_ENC_XXX: AP to VPU cmd message id + * @VPU_IPIMSG_ENC_XXX_DONE: VPU ack AP cmd message id + */ +enum venc_ipi_msg_id { + AP_IPIMSG_ENC_INIT = AP_IPIMSG_VENC_BASE, + AP_IPIMSG_ENC_SET_PARAM, + AP_IPIMSG_ENC_ENCODE, + AP_IPIMSG_ENC_DEINIT, + + VPU_IPIMSG_ENC_INIT_DONE = VPU_IPIMSG_VENC_BASE, + VPU_IPIMSG_ENC_SET_PARAM_DONE, + VPU_IPIMSG_ENC_ENCODE_DONE, + VPU_IPIMSG_ENC_DEINIT_DONE, +}; + +/** + * struct venc_ap_ipi_msg_init - AP to VPU init cmd structure + * @msg_id: message id (AP_IPIMSG_XXX_ENC_INIT) + * @reserved: reserved for future use. vpu is running in 32bit. Without + * this reserved field, if kernel run in 64bit. this struct size + * will be different between kernel and vpu + * @venc_inst: AP encoder instance + * (struct venc_vp8_inst/venc_h264_inst *) + */ +struct venc_ap_ipi_msg_init { + uint32_t msg_id; + uint32_t reserved; + uint64_t venc_inst; +}; + +/** + * struct venc_ap_ipi_msg_set_param - AP to VPU set_param cmd structure + * @msg_id: message id (AP_IPIMSG_XXX_ENC_SET_PARAM) + * @vpu_inst_addr: VPU encoder instance addr + * (struct venc_vp8_vsi/venc_h264_vsi *) + * @param_id: parameter id (venc_set_param_type) + * @data_item: number of items in the data array + * @data: data array to store the set parameters + */ +struct venc_ap_ipi_msg_set_param { + uint32_t msg_id; + uint32_t vpu_inst_addr; + uint32_t param_id; + uint32_t data_item; + uint32_t data[8]; +}; + +struct venc_ap_ipi_msg_set_param_ext { + struct venc_ap_ipi_msg_set_param base; + uint32_t data_ext[24]; +}; + +/** + * struct venc_ap_ipi_msg_enc - AP to VPU enc cmd structure + * @msg_id: message id (AP_IPIMSG_XXX_ENC_ENCODE) + * @vpu_inst_addr: VPU encoder instance addr + * (struct venc_vp8_vsi/venc_h264_vsi *) + * @bs_mode: bitstream mode for h264 + * (H264_BS_MODE_SPS/H264_BS_MODE_PPS/H264_BS_MODE_FRAME) + * @input_addr: pointer to input image buffer plane + * @bs_addr: pointer to output bit stream buffer + * @bs_size: bit stream buffer size + */ +struct venc_ap_ipi_msg_enc { + uint32_t msg_id; + uint32_t vpu_inst_addr; + uint32_t bs_mode; + uint32_t input_addr[3]; + uint32_t bs_addr; + uint32_t bs_size; +}; + +/** + * struct venc_ap_ipi_msg_enc_ext - AP to SCP extended enc cmd structure + * + * @base: base msg structure + * @data_item: number of items in the data array + * @data: data array to store the set parameters + */ +struct venc_ap_ipi_msg_enc_ext { + struct venc_ap_ipi_msg_enc base; + uint32_t data_item; + uint32_t data[32]; +}; + +/** + * struct venc_ap_ipi_msg_enc_ext_34 - AP to SCP extended enc cmd structure + * @msg_id: message id (AP_IPIMSG_XXX_ENC_ENCODE) + * @vpu_inst_addr: VPU encoder instance addr + * @bs_mode: bitstream mode for h264 + * @reserved: for struct padding + * @input_addr: input frame buffer 34 bit address + * @bs_addr: output bitstream buffer 34 bit address + * @bs_size: bitstream buffer size + * @data_item: number of items in the data array + * @data: data array to store the set parameters + */ +struct venc_ap_ipi_msg_enc_ext_34 { + u32 msg_id; + u32 vpu_inst_addr; + u32 bs_mode; + u32 reserved; + u64 input_addr[3]; + u64 bs_addr; + u32 bs_size; + u32 data_item; + u32 data[32]; +}; + +/** + * struct venc_ap_ipi_msg_deinit - AP to VPU deinit cmd structure + * @msg_id: message id (AP_IPIMSG_XXX_ENC_DEINIT) + * @vpu_inst_addr: VPU encoder instance addr + * (struct venc_vp8_vsi/venc_h264_vsi *) + */ +struct venc_ap_ipi_msg_deinit { + uint32_t msg_id; + uint32_t vpu_inst_addr; +}; + +/* + * enum venc_ipi_msg_status - VPU ack AP cmd status + */ +enum venc_ipi_msg_status { + VENC_IPI_MSG_STATUS_OK, + VENC_IPI_MSG_STATUS_FAIL, +}; + +/** + * struct venc_vpu_ipi_msg_common - VPU ack AP cmd common structure + * @msg_id: message id (VPU_IPIMSG_XXX_DONE) + * @status: cmd status (venc_ipi_msg_status) + * @venc_inst: AP encoder instance (struct venc_vp8_inst/venc_h264_inst *) + */ +struct venc_vpu_ipi_msg_common { + uint32_t msg_id; + uint32_t status; + uint64_t venc_inst; +}; + +/** + * struct venc_vpu_ipi_msg_init - VPU ack AP init cmd structure + * @msg_id: message id (VPU_IPIMSG_XXX_ENC_SET_PARAM_DONE) + * @status: cmd status (venc_ipi_msg_status) + * @venc_inst: AP encoder instance (struct venc_vp8_inst/venc_h264_inst *) + * @vpu_inst_addr: VPU encoder instance addr + * (struct venc_vp8_vsi/venc_h264_vsi *) + * @venc_abi_version: ABI version of the firmware. Kernel can use it to + * ensure that it is compatible with the firmware. + * For MT8173 the value of this field is undefined and + * should not be used. + */ +struct venc_vpu_ipi_msg_init { + uint32_t msg_id; + uint32_t status; + uint64_t venc_inst; + uint32_t vpu_inst_addr; + uint32_t venc_abi_version; +}; + +/** + * struct venc_vpu_ipi_msg_set_param - VPU ack AP set_param cmd structure + * @msg_id: message id (VPU_IPIMSG_XXX_ENC_SET_PARAM_DONE) + * @status: cmd status (venc_ipi_msg_status) + * @venc_inst: AP encoder instance (struct venc_vp8_inst/venc_h264_inst *) + * @param_id: parameter id (venc_set_param_type) + * @data_item: number of items in the data array + * @data: data array to store the return result + */ +struct venc_vpu_ipi_msg_set_param { + uint32_t msg_id; + uint32_t status; + uint64_t venc_inst; + uint32_t param_id; + uint32_t data_item; + uint32_t data[6]; +}; + +/** + * enum venc_ipi_msg_enc_state - Type of encode state + * @VEN_IPI_MSG_ENC_STATE_FRAME: one frame being encoded + * @VEN_IPI_MSG_ENC_STATE_PART: bit stream buffer full + * @VEN_IPI_MSG_ENC_STATE_SKIP: encoded skip frame + * @VEN_IPI_MSG_ENC_STATE_ERROR: encounter error + */ +enum venc_ipi_msg_enc_state { + VEN_IPI_MSG_ENC_STATE_FRAME, + VEN_IPI_MSG_ENC_STATE_PART, + VEN_IPI_MSG_ENC_STATE_SKIP, + VEN_IPI_MSG_ENC_STATE_ERROR, +}; + +/** + * struct venc_vpu_ipi_msg_enc - VPU ack AP enc cmd structure + * @msg_id: message id (VPU_IPIMSG_XXX_ENC_ENCODE_DONE) + * @status: cmd status (venc_ipi_msg_status) + * @venc_inst: AP encoder instance (struct venc_vp8_inst/venc_h264_inst *) + * @state: encode state (venc_ipi_msg_enc_state) + * @is_key_frm: whether the encoded frame is key frame + * @bs_size: encoded bitstream size + * @reserved: reserved for future use. vpu is running in 32bit. Without + * this reserved field, if kernel run in 64bit. this struct size + * will be different between kernel and vpu + */ +struct venc_vpu_ipi_msg_enc { + uint32_t msg_id; + uint32_t status; + uint64_t venc_inst; + uint32_t state; + uint32_t is_key_frm; + uint32_t bs_size; + uint32_t reserved; +}; + +/** + * struct venc_vpu_ipi_msg_deinit - VPU ack AP deinit cmd structure + * @msg_id: message id (VPU_IPIMSG_XXX_ENC_DEINIT_DONE) + * @status: cmd status (venc_ipi_msg_status) + * @venc_inst: AP encoder instance (struct venc_vp8_inst/venc_h264_inst *) + */ +struct venc_vpu_ipi_msg_deinit { + uint32_t msg_id; + uint32_t status; + uint64_t venc_inst; +}; + +#endif /* _VENC_IPI_MSG_H_ */ diff --git a/drivers/media/platform/mediatek/vcodec/encoder/venc_vpu_if.c b/drivers/media/platform/mediatek/vcodec/encoder/venc_vpu_if.c new file mode 100644 index 0000000000..84ad1cc6ad --- /dev/null +++ b/drivers/media/platform/mediatek/vcodec/encoder/venc_vpu_if.c @@ -0,0 +1,379 @@ +// SPDX-License-Identifier: GPL-2.0 +/* + * Copyright (c) 2016 MediaTek Inc. + * Author: PoChun Lin <pochun.lin@mediatek.com> + */ + +#include "mtk_vcodec_enc_drv.h" +#include "venc_ipi_msg.h" +#include "venc_vpu_if.h" + +static void handle_enc_init_msg(struct venc_vpu_inst *vpu, const void *data) +{ + const struct venc_vpu_ipi_msg_init *msg = data; + + vpu->inst_addr = msg->vpu_inst_addr; + vpu->vsi = mtk_vcodec_fw_map_dm_addr(vpu->ctx->dev->fw_handler, + msg->vpu_inst_addr); + + /* Firmware version field value is unspecified on MT8173. */ + if (mtk_vcodec_fw_get_type(vpu->ctx->dev->fw_handler) == VPU) + return; + + /* Check firmware version. */ + mtk_venc_debug(vpu->ctx, "firmware version: 0x%x\n", msg->venc_abi_version); + switch (msg->venc_abi_version) { + case 1: + break; + default: + mtk_venc_err(vpu->ctx, "unhandled firmware version 0x%x\n", + msg->venc_abi_version); + vpu->failure = 1; + break; + } +} + +static void handle_enc_encode_msg(struct venc_vpu_inst *vpu, const void *data) +{ + const struct venc_vpu_ipi_msg_enc *msg = data; + + vpu->state = msg->state; + vpu->bs_size = msg->bs_size; + vpu->is_key_frm = msg->is_key_frm; +} + +static bool vpu_enc_check_ap_inst(struct mtk_vcodec_enc_dev *enc_dev, struct venc_vpu_inst *vpu) +{ + struct mtk_vcodec_enc_ctx *ctx; + int ret = false; + + list_for_each_entry(ctx, &enc_dev->ctx_list, list) { + if (!IS_ERR_OR_NULL(ctx) && ctx->vpu_inst == vpu) { + ret = true; + break; + } + } + + return ret; +} + +static void vpu_enc_ipi_handler(void *data, unsigned int len, void *priv) +{ + struct mtk_vcodec_enc_dev *enc_dev; + const struct venc_vpu_ipi_msg_common *msg = data; + struct venc_vpu_inst *vpu; + + enc_dev = (struct mtk_vcodec_enc_dev *)priv; + vpu = (struct venc_vpu_inst *)(unsigned long)msg->venc_inst; + if (!priv || !vpu) { + pr_err(MTK_DBG_V4L2_STR "venc_inst is NULL, did the SCP hang or crash?"); + return; + } + + mtk_venc_debug(vpu->ctx, "msg_id %x inst %p status %d", msg->msg_id, vpu, msg->status); + if (!vpu_enc_check_ap_inst(enc_dev, vpu) || msg->msg_id < VPU_IPIMSG_ENC_INIT_DONE || + msg->msg_id > VPU_IPIMSG_ENC_DEINIT_DONE) { + mtk_v4l2_venc_err(vpu->ctx, "venc msg id not correctly => 0x%x", msg->msg_id); + vpu->failure = -EINVAL; + goto error; + } + + vpu->failure = (msg->status != VENC_IPI_MSG_STATUS_OK); + if (vpu->failure) { + mtk_venc_err(vpu->ctx, "vpu enc status failure %d", vpu->failure); + goto error; + } + + switch (msg->msg_id) { + case VPU_IPIMSG_ENC_INIT_DONE: + handle_enc_init_msg(vpu, data); + break; + case VPU_IPIMSG_ENC_SET_PARAM_DONE: + break; + case VPU_IPIMSG_ENC_ENCODE_DONE: + handle_enc_encode_msg(vpu, data); + break; + case VPU_IPIMSG_ENC_DEINIT_DONE: + break; + default: + mtk_venc_err(vpu->ctx, "unknown msg id %x", msg->msg_id); + break; + } + +error: + vpu->signaled = 1; +} + +static int vpu_enc_send_msg(struct venc_vpu_inst *vpu, void *msg, + int len) +{ + int status; + + if (!vpu->ctx->dev->fw_handler) { + mtk_venc_err(vpu->ctx, "inst dev is NULL"); + return -EINVAL; + } + + status = mtk_vcodec_fw_ipi_send(vpu->ctx->dev->fw_handler, vpu->id, msg, + len, 2000); + if (status) { + mtk_venc_err(vpu->ctx, "vpu_ipi_send msg_id %x len %d fail %d", + *(uint32_t *)msg, len, status); + return -EINVAL; + } + if (vpu->failure) + return -EINVAL; + + return 0; +} + +int vpu_enc_init(struct venc_vpu_inst *vpu) +{ + int status; + struct venc_ap_ipi_msg_init out; + + init_waitqueue_head(&vpu->wq_hd); + vpu->signaled = 0; + vpu->failure = 0; + vpu->ctx->vpu_inst = vpu; + + status = mtk_vcodec_fw_ipi_register(vpu->ctx->dev->fw_handler, vpu->id, + vpu_enc_ipi_handler, "venc", + vpu->ctx->dev); + + if (status) { + mtk_venc_err(vpu->ctx, "vpu_ipi_register fail %d", status); + return -EINVAL; + } + + memset(&out, 0, sizeof(out)); + out.msg_id = AP_IPIMSG_ENC_INIT; + out.venc_inst = (unsigned long)vpu; + if (vpu_enc_send_msg(vpu, &out, sizeof(out))) { + mtk_venc_err(vpu->ctx, "AP_IPIMSG_ENC_INIT fail"); + return -EINVAL; + } + + if (IS_ERR_OR_NULL(vpu->vsi)) { + mtk_venc_err(vpu->ctx, "invalid venc vsi"); + return -EINVAL; + } + + return 0; +} + +static unsigned int venc_enc_param_crop_right(struct venc_vpu_inst *vpu, + struct venc_enc_param *enc_prm) +{ + unsigned int img_crop_right = enc_prm->buf_width - enc_prm->width; + + return img_crop_right % 16; +} + +static unsigned int venc_enc_param_crop_bottom(struct venc_enc_param *enc_prm) +{ + return round_up(enc_prm->height, 16) - enc_prm->height; +} + +static unsigned int venc_enc_param_num_mb(struct venc_enc_param *enc_prm) +{ + return DIV_ROUND_UP(enc_prm->width, 16) * + DIV_ROUND_UP(enc_prm->height, 16); +} + +int vpu_enc_set_param(struct venc_vpu_inst *vpu, + enum venc_set_param_type id, + struct venc_enc_param *enc_param) +{ + const bool is_ext = MTK_ENC_CTX_IS_EXT(vpu->ctx); + size_t msg_size = is_ext ? + sizeof(struct venc_ap_ipi_msg_set_param_ext) : + sizeof(struct venc_ap_ipi_msg_set_param); + struct venc_ap_ipi_msg_set_param_ext out; + + mtk_venc_debug(vpu->ctx, "id %d ->", id); + + memset(&out, 0, sizeof(out)); + out.base.msg_id = AP_IPIMSG_ENC_SET_PARAM; + out.base.vpu_inst_addr = vpu->inst_addr; + out.base.param_id = id; + switch (id) { + case VENC_SET_PARAM_ENC: + if (is_ext) { + out.base.data_item = 3; + out.base.data[0] = + venc_enc_param_crop_right(vpu, enc_param); + out.base.data[1] = + venc_enc_param_crop_bottom(enc_param); + out.base.data[2] = venc_enc_param_num_mb(enc_param); + } else { + out.base.data_item = 0; + } + break; + case VENC_SET_PARAM_FORCE_INTRA: + out.base.data_item = 0; + break; + case VENC_SET_PARAM_ADJUST_BITRATE: + out.base.data_item = 1; + out.base.data[0] = enc_param->bitrate; + break; + case VENC_SET_PARAM_ADJUST_FRAMERATE: + out.base.data_item = 1; + out.base.data[0] = enc_param->frm_rate; + break; + case VENC_SET_PARAM_GOP_SIZE: + out.base.data_item = 1; + out.base.data[0] = enc_param->gop_size; + break; + case VENC_SET_PARAM_INTRA_PERIOD: + out.base.data_item = 1; + out.base.data[0] = enc_param->intra_period; + break; + case VENC_SET_PARAM_SKIP_FRAME: + out.base.data_item = 0; + break; + default: + mtk_venc_err(vpu->ctx, "id %d not supported", id); + return -EINVAL; + } + if (vpu_enc_send_msg(vpu, &out, msg_size)) { + mtk_venc_err(vpu->ctx, "AP_IPIMSG_ENC_SET_PARAM %d fail", id); + return -EINVAL; + } + + mtk_venc_debug(vpu->ctx, "id %d <-", id); + + return 0; +} + +static int vpu_enc_encode_32bits(struct venc_vpu_inst *vpu, + unsigned int bs_mode, + struct venc_frm_buf *frm_buf, + struct mtk_vcodec_mem *bs_buf, + struct venc_frame_info *frame_info) +{ + const bool is_ext = MTK_ENC_CTX_IS_EXT(vpu->ctx); + size_t msg_size = is_ext ? + sizeof(struct venc_ap_ipi_msg_enc_ext) : + sizeof(struct venc_ap_ipi_msg_enc); + struct venc_ap_ipi_msg_enc_ext out; + + mtk_venc_debug(vpu->ctx, "bs_mode %d ->", bs_mode); + + memset(&out, 0, sizeof(out)); + out.base.msg_id = AP_IPIMSG_ENC_ENCODE; + out.base.vpu_inst_addr = vpu->inst_addr; + out.base.bs_mode = bs_mode; + if (frm_buf) { + if ((frm_buf->fb_addr[0].dma_addr % 16 == 0) && + (frm_buf->fb_addr[1].dma_addr % 16 == 0) && + (frm_buf->fb_addr[2].dma_addr % 16 == 0)) { + out.base.input_addr[0] = frm_buf->fb_addr[0].dma_addr; + out.base.input_addr[1] = frm_buf->fb_addr[1].dma_addr; + out.base.input_addr[2] = frm_buf->fb_addr[2].dma_addr; + } else { + mtk_venc_err(vpu->ctx, "dma_addr not align to 16"); + return -EINVAL; + } + } + if (bs_buf) { + out.base.bs_addr = bs_buf->dma_addr; + out.base.bs_size = bs_buf->size; + } + if (is_ext && frame_info) { + out.data_item = 3; + out.data[0] = frame_info->frm_count; + out.data[1] = frame_info->skip_frm_count; + out.data[2] = frame_info->frm_type; + } + if (vpu_enc_send_msg(vpu, &out, msg_size)) { + mtk_venc_err(vpu->ctx, "AP_IPIMSG_ENC_ENCODE %d fail", bs_mode); + return -EINVAL; + } + + return 0; +} + +static int vpu_enc_encode_34bits(struct venc_vpu_inst *vpu, + unsigned int bs_mode, + struct venc_frm_buf *frm_buf, + struct mtk_vcodec_mem *bs_buf, + struct venc_frame_info *frame_info) +{ + struct venc_ap_ipi_msg_enc_ext_34 out; + size_t msg_size = sizeof(struct venc_ap_ipi_msg_enc_ext_34); + + mtk_venc_debug(vpu->ctx, "bs_mode %d ->", bs_mode); + + memset(&out, 0, sizeof(out)); + out.msg_id = AP_IPIMSG_ENC_ENCODE; + out.vpu_inst_addr = vpu->inst_addr; + out.bs_mode = bs_mode; + + if (frm_buf) { + if ((frm_buf->fb_addr[0].dma_addr % 16 == 0) && + (frm_buf->fb_addr[1].dma_addr % 16 == 0) && + (frm_buf->fb_addr[2].dma_addr % 16 == 0)) { + out.input_addr[0] = frm_buf->fb_addr[0].dma_addr; + out.input_addr[1] = frm_buf->fb_addr[1].dma_addr; + out.input_addr[2] = frm_buf->fb_addr[2].dma_addr; + } else { + mtk_venc_err(vpu->ctx, "dma_addr not align to 16"); + return -EINVAL; + } + } + if (bs_buf) { + out.bs_addr = bs_buf->dma_addr; + out.bs_size = bs_buf->size; + } + if (frame_info) { + out.data_item = 3; + out.data[0] = frame_info->frm_count; + out.data[1] = frame_info->skip_frm_count; + out.data[2] = frame_info->frm_type; + } + if (vpu_enc_send_msg(vpu, &out, msg_size)) { + mtk_venc_err(vpu->ctx, "AP_IPIMSG_ENC_ENCODE %d fail", bs_mode); + return -EINVAL; + } + + return 0; +} + +int vpu_enc_encode(struct venc_vpu_inst *vpu, unsigned int bs_mode, + struct venc_frm_buf *frm_buf, + struct mtk_vcodec_mem *bs_buf, + struct venc_frame_info *frame_info) +{ + int ret; + + if (MTK_ENC_IOVA_IS_34BIT(vpu->ctx)) + ret = vpu_enc_encode_34bits(vpu, bs_mode, + frm_buf, bs_buf, frame_info); + else + ret = vpu_enc_encode_32bits(vpu, bs_mode, + frm_buf, bs_buf, frame_info); + + if (ret) + return ret; + + mtk_venc_debug(vpu->ctx, "bs_mode %d state %d size %d key_frm %d <-", + bs_mode, vpu->state, vpu->bs_size, vpu->is_key_frm); + + return 0; +} + +int vpu_enc_deinit(struct venc_vpu_inst *vpu) +{ + struct venc_ap_ipi_msg_deinit out; + + memset(&out, 0, sizeof(out)); + out.msg_id = AP_IPIMSG_ENC_DEINIT; + out.vpu_inst_addr = vpu->inst_addr; + if (vpu_enc_send_msg(vpu, &out, sizeof(out))) { + mtk_venc_err(vpu->ctx, "AP_IPIMSG_ENC_DEINIT fail"); + return -EINVAL; + } + + return 0; +} diff --git a/drivers/media/platform/mediatek/vcodec/encoder/venc_vpu_if.h b/drivers/media/platform/mediatek/vcodec/encoder/venc_vpu_if.h new file mode 100644 index 0000000000..ede55fc3bd --- /dev/null +++ b/drivers/media/platform/mediatek/vcodec/encoder/venc_vpu_if.h @@ -0,0 +1,50 @@ +/* SPDX-License-Identifier: GPL-2.0 */ +/* + * Copyright (c) 2016 MediaTek Inc. + * Author: PoChun Lin <pochun.lin@mediatek.com> + */ + +#ifndef _VENC_VPU_IF_H_ +#define _VENC_VPU_IF_H_ + +#include "venc_drv_if.h" + +/* + * struct venc_vpu_inst - encoder VPU driver instance + * @wq_hd: wait queue used for vpu cmd trigger then wait vpu interrupt done + * @signaled: flag used for checking vpu interrupt done + * @failure: flag to show vpu cmd succeeds or not + * @state: enum venc_ipi_msg_enc_state + * @bs_size: bitstream size for skip frame case usage + * @is_key_frm: key frame flag + * @inst_addr: VPU instance addr + * @vsi: driver structure allocated by VPU side and shared to AP side for + * control and info share + * @id: the id of inter-processor interrupt + * @ctx: context for v4l2 layer integration + * @dev: device for v4l2 layer integration + */ +struct venc_vpu_inst { + wait_queue_head_t wq_hd; + int signaled; + int failure; + int state; + int bs_size; + int is_key_frm; + unsigned int inst_addr; + void *vsi; + int id; + struct mtk_vcodec_enc_ctx *ctx; +}; + +int vpu_enc_init(struct venc_vpu_inst *vpu); +int vpu_enc_set_param(struct venc_vpu_inst *vpu, + enum venc_set_param_type id, + struct venc_enc_param *param); +int vpu_enc_encode(struct venc_vpu_inst *vpu, unsigned int bs_mode, + struct venc_frm_buf *frm_buf, + struct mtk_vcodec_mem *bs_buf, + struct venc_frame_info *frame_info); +int vpu_enc_deinit(struct venc_vpu_inst *vpu); + +#endif diff --git a/drivers/media/platform/mediatek/vpu/Kconfig b/drivers/media/platform/mediatek/vpu/Kconfig new file mode 100644 index 0000000000..2a8443a93c --- /dev/null +++ b/drivers/media/platform/mediatek/vpu/Kconfig @@ -0,0 +1,15 @@ +# SPDX-License-Identifier: GPL-2.0-only + +config VIDEO_MEDIATEK_VPU + tristate "Mediatek Video Processor Unit" + depends on V4L_MEM2MEM_DRIVERS + depends on VIDEO_DEV + depends on ARCH_MEDIATEK || COMPILE_TEST + help + This driver provides downloading VPU firmware and + communicating with VPU. This driver for hw video + codec embedded in Mediatek's MT8173 SOCs. It is able + to handle video decoding/encoding in a range of formats. + + To compile this driver as a module, choose M here: the + module will be called mtk-vpu. diff --git a/drivers/media/platform/mediatek/vpu/Makefile b/drivers/media/platform/mediatek/vpu/Makefile new file mode 100644 index 0000000000..ecd2d392b8 --- /dev/null +++ b/drivers/media/platform/mediatek/vpu/Makefile @@ -0,0 +1,4 @@ +# SPDX-License-Identifier: GPL-2.0-only +mtk-vpu-y += mtk_vpu.o + +obj-$(CONFIG_VIDEO_MEDIATEK_VPU) += mtk-vpu.o diff --git a/drivers/media/platform/mediatek/vpu/mtk_vpu.c b/drivers/media/platform/mediatek/vpu/mtk_vpu.c new file mode 100644 index 0000000000..7243604a82 --- /dev/null +++ b/drivers/media/platform/mediatek/vpu/mtk_vpu.c @@ -0,0 +1,1055 @@ +// SPDX-License-Identifier: GPL-2.0-only +/* +* Copyright (c) 2016 MediaTek Inc. +* Author: Andrew-CT Chen <andrew-ct.chen@mediatek.com> +*/ +#include <linux/clk.h> +#include <linux/debugfs.h> +#include <linux/firmware.h> +#include <linux/interrupt.h> +#include <linux/iommu.h> +#include <linux/module.h> +#include <linux/of.h> +#include <linux/of_platform.h> +#include <linux/of_reserved_mem.h> +#include <linux/platform_device.h> +#include <linux/sched.h> +#include <linux/sizes.h> +#include <linux/dma-mapping.h> + +#include "mtk_vpu.h" + +/* + * VPU (video processor unit) is a tiny processor controlling video hardware + * related to video codec, scaling and color format converting. + * VPU interfaces with other blocks by share memory and interrupt. + */ + +#define INIT_TIMEOUT_MS 2000U +#define IPI_TIMEOUT_MS 2000U +#define VPU_IDLE_TIMEOUT_MS 1000U +#define VPU_FW_VER_LEN 16 + +/* maximum program/data TCM (Tightly-Coupled Memory) size */ +#define VPU_PTCM_SIZE (96 * SZ_1K) +#define VPU_DTCM_SIZE (32 * SZ_1K) +/* the offset to get data tcm address */ +#define VPU_DTCM_OFFSET 0x18000UL +/* daynamic allocated maximum extended memory size */ +#define VPU_EXT_P_SIZE SZ_1M +#define VPU_EXT_D_SIZE SZ_4M +/* maximum binary firmware size */ +#define VPU_P_FW_SIZE (VPU_PTCM_SIZE + VPU_EXT_P_SIZE) +#define VPU_D_FW_SIZE (VPU_DTCM_SIZE + VPU_EXT_D_SIZE) +/* the size of share buffer between Host and VPU */ +#define SHARE_BUF_SIZE 48 + +/* binary firmware name */ +#define VPU_P_FW "vpu_p.bin" +#define VPU_D_FW "vpu_d.bin" +#define VPU_P_FW_NEW "mediatek/mt8173/vpu_p.bin" +#define VPU_D_FW_NEW "mediatek/mt8173/vpu_d.bin" + +#define VPU_RESET 0x0 +#define VPU_TCM_CFG 0x0008 +#define VPU_PMEM_EXT0_ADDR 0x000C +#define VPU_PMEM_EXT1_ADDR 0x0010 +#define VPU_TO_HOST 0x001C +#define VPU_DMEM_EXT0_ADDR 0x0014 +#define VPU_DMEM_EXT1_ADDR 0x0018 +#define HOST_TO_VPU 0x0024 +#define VPU_IDLE_REG 0x002C +#define VPU_INT_STATUS 0x0034 +#define VPU_PC_REG 0x0060 +#define VPU_SP_REG 0x0064 +#define VPU_RA_REG 0x0068 +#define VPU_WDT_REG 0x0084 + +/* vpu inter-processor communication interrupt */ +#define VPU_IPC_INT BIT(8) +/* vpu idle state */ +#define VPU_IDLE_STATE BIT(23) + +/** + * enum vpu_fw_type - VPU firmware type + * + * @P_FW: program firmware + * @D_FW: data firmware + * + */ +enum vpu_fw_type { + P_FW, + D_FW, +}; + +/** + * struct vpu_mem - VPU extended program/data memory information + * + * @va: the kernel virtual memory address of VPU extended memory + * @pa: the physical memory address of VPU extended memory + * + */ +struct vpu_mem { + void *va; + dma_addr_t pa; +}; + +/** + * struct vpu_regs - VPU TCM and configuration registers + * + * @tcm: the register for VPU Tightly-Coupled Memory + * @cfg: the register for VPU configuration + * @irq: the irq number for VPU interrupt + */ +struct vpu_regs { + void __iomem *tcm; + void __iomem *cfg; + int irq; +}; + +/** + * struct vpu_wdt_handler - VPU watchdog reset handler + * + * @reset_func: reset handler + * @priv: private data + */ +struct vpu_wdt_handler { + void (*reset_func)(void *); + void *priv; +}; + +/** + * struct vpu_wdt - VPU watchdog workqueue + * + * @handler: VPU watchdog reset handler + * @ws: workstruct for VPU watchdog + * @wq: workqueue for VPU watchdog + */ +struct vpu_wdt { + struct vpu_wdt_handler handler[VPU_RST_MAX]; + struct work_struct ws; + struct workqueue_struct *wq; +}; + +/** + * struct vpu_run - VPU initialization status + * + * @signaled: the signal of vpu initialization completed + * @fw_ver: VPU firmware version + * @dec_capability: decoder capability which is not used for now and + * the value is reserved for future use + * @enc_capability: encoder capability which is not used for now and + * the value is reserved for future use + * @wq: wait queue for VPU initialization status + */ +struct vpu_run { + u32 signaled; + char fw_ver[VPU_FW_VER_LEN]; + unsigned int dec_capability; + unsigned int enc_capability; + wait_queue_head_t wq; +}; + +/** + * struct vpu_ipi_desc - VPU IPI descriptor + * + * @handler: IPI handler + * @name: the name of IPI handler + * @priv: the private data of IPI handler + */ +struct vpu_ipi_desc { + ipi_handler_t handler; + const char *name; + void *priv; +}; + +/** + * struct share_obj - DTCM (Data Tightly-Coupled Memory) buffer shared with + * AP and VPU + * + * @id: IPI id + * @len: share buffer length + * @share_buf: share buffer data + */ +struct share_obj { + s32 id; + u32 len; + unsigned char share_buf[SHARE_BUF_SIZE]; +}; + +/** + * struct mtk_vpu - vpu driver data + * @extmem: VPU extended memory information + * @reg: VPU TCM and configuration registers + * @run: VPU initialization status + * @wdt: VPU watchdog workqueue + * @ipi_desc: VPU IPI descriptor + * @recv_buf: VPU DTCM share buffer for receiving. The + * receive buffer is only accessed in interrupt context. + * @send_buf: VPU DTCM share buffer for sending + * @dev: VPU struct device + * @clk: VPU clock on/off + * @fw_loaded: indicate VPU firmware loaded + * @enable_4GB: VPU 4GB mode on/off + * @vpu_mutex: protect mtk_vpu (except recv_buf) and ensure only + * one client to use VPU service at a time. For example, + * suppose a client is using VPU to decode VP8. + * If the other client wants to encode VP8, + * it has to wait until VP8 decode completes. + * @wdt_refcnt: WDT reference count to make sure the watchdog can be + * disabled if no other client is using VPU service + * @ack_wq: The wait queue for each codec and mdp. When sleeping + * processes wake up, they will check the condition + * "ipi_id_ack" to run the corresponding action or + * go back to sleep. + * @ipi_id_ack: The ACKs for registered IPI function sending + * interrupt to VPU + * + */ +struct mtk_vpu { + struct vpu_mem extmem[2]; + struct vpu_regs reg; + struct vpu_run run; + struct vpu_wdt wdt; + struct vpu_ipi_desc ipi_desc[IPI_MAX]; + struct share_obj __iomem *recv_buf; + struct share_obj __iomem *send_buf; + struct device *dev; + struct clk *clk; + bool fw_loaded; + bool enable_4GB; + struct mutex vpu_mutex; /* for protecting vpu data data structure */ + u32 wdt_refcnt; + wait_queue_head_t ack_wq; + bool ipi_id_ack[IPI_MAX]; +}; + +static inline void vpu_cfg_writel(struct mtk_vpu *vpu, u32 val, u32 offset) +{ + writel(val, vpu->reg.cfg + offset); +} + +static inline u32 vpu_cfg_readl(struct mtk_vpu *vpu, u32 offset) +{ + return readl(vpu->reg.cfg + offset); +} + +static inline bool vpu_running(struct mtk_vpu *vpu) +{ + return vpu_cfg_readl(vpu, VPU_RESET) & BIT(0); +} + +static void vpu_clock_disable(struct mtk_vpu *vpu) +{ + /* Disable VPU watchdog */ + mutex_lock(&vpu->vpu_mutex); + if (!--vpu->wdt_refcnt) + vpu_cfg_writel(vpu, + vpu_cfg_readl(vpu, VPU_WDT_REG) & ~(1L << 31), + VPU_WDT_REG); + mutex_unlock(&vpu->vpu_mutex); + + clk_disable(vpu->clk); +} + +static int vpu_clock_enable(struct mtk_vpu *vpu) +{ + int ret; + + ret = clk_enable(vpu->clk); + if (ret) + return ret; + /* Enable VPU watchdog */ + mutex_lock(&vpu->vpu_mutex); + if (!vpu->wdt_refcnt++) + vpu_cfg_writel(vpu, + vpu_cfg_readl(vpu, VPU_WDT_REG) | (1L << 31), + VPU_WDT_REG); + mutex_unlock(&vpu->vpu_mutex); + + return ret; +} + +static void vpu_dump_status(struct mtk_vpu *vpu) +{ + dev_info(vpu->dev, + "vpu: run %x, pc = 0x%x, ra = 0x%x, sp = 0x%x, idle = 0x%x\n" + "vpu: int %x, hv = 0x%x, vh = 0x%x, wdt = 0x%x\n", + vpu_running(vpu), vpu_cfg_readl(vpu, VPU_PC_REG), + vpu_cfg_readl(vpu, VPU_RA_REG), vpu_cfg_readl(vpu, VPU_SP_REG), + vpu_cfg_readl(vpu, VPU_IDLE_REG), + vpu_cfg_readl(vpu, VPU_INT_STATUS), + vpu_cfg_readl(vpu, HOST_TO_VPU), + vpu_cfg_readl(vpu, VPU_TO_HOST), + vpu_cfg_readl(vpu, VPU_WDT_REG)); +} + +int vpu_ipi_register(struct platform_device *pdev, + enum ipi_id id, ipi_handler_t handler, + const char *name, void *priv) +{ + struct mtk_vpu *vpu = platform_get_drvdata(pdev); + struct vpu_ipi_desc *ipi_desc; + + if (!vpu) { + dev_err(&pdev->dev, "vpu device in not ready\n"); + return -EPROBE_DEFER; + } + + if (id < IPI_MAX && handler) { + ipi_desc = vpu->ipi_desc; + ipi_desc[id].name = name; + ipi_desc[id].handler = handler; + ipi_desc[id].priv = priv; + return 0; + } + + dev_err(&pdev->dev, "register vpu ipi id %d with invalid arguments\n", + id); + return -EINVAL; +} +EXPORT_SYMBOL_GPL(vpu_ipi_register); + +int vpu_ipi_send(struct platform_device *pdev, + enum ipi_id id, void *buf, + unsigned int len) +{ + struct mtk_vpu *vpu = platform_get_drvdata(pdev); + struct share_obj __iomem *send_obj = vpu->send_buf; + unsigned long timeout; + int ret = 0; + + if (id <= IPI_VPU_INIT || id >= IPI_MAX || + len > sizeof(send_obj->share_buf) || !buf) { + dev_err(vpu->dev, "failed to send ipi message\n"); + return -EINVAL; + } + + ret = vpu_clock_enable(vpu); + if (ret) { + dev_err(vpu->dev, "failed to enable vpu clock\n"); + return ret; + } + if (!vpu_running(vpu)) { + dev_err(vpu->dev, "vpu_ipi_send: VPU is not running\n"); + ret = -EINVAL; + goto clock_disable; + } + + mutex_lock(&vpu->vpu_mutex); + + /* Wait until VPU receives the last command */ + timeout = jiffies + msecs_to_jiffies(IPI_TIMEOUT_MS); + do { + if (time_after(jiffies, timeout)) { + dev_err(vpu->dev, "vpu_ipi_send: IPI timeout!\n"); + ret = -EIO; + vpu_dump_status(vpu); + goto mut_unlock; + } + } while (vpu_cfg_readl(vpu, HOST_TO_VPU)); + + memcpy_toio(send_obj->share_buf, buf, len); + writel(len, &send_obj->len); + writel(id, &send_obj->id); + + vpu->ipi_id_ack[id] = false; + /* send the command to VPU */ + vpu_cfg_writel(vpu, 0x1, HOST_TO_VPU); + + mutex_unlock(&vpu->vpu_mutex); + + /* wait for VPU's ACK */ + timeout = msecs_to_jiffies(IPI_TIMEOUT_MS); + ret = wait_event_timeout(vpu->ack_wq, vpu->ipi_id_ack[id], timeout); + vpu->ipi_id_ack[id] = false; + if (ret == 0) { + dev_err(vpu->dev, "vpu ipi %d ack time out !\n", id); + ret = -EIO; + vpu_dump_status(vpu); + goto clock_disable; + } + vpu_clock_disable(vpu); + + return 0; + +mut_unlock: + mutex_unlock(&vpu->vpu_mutex); +clock_disable: + vpu_clock_disable(vpu); + + return ret; +} +EXPORT_SYMBOL_GPL(vpu_ipi_send); + +static void vpu_wdt_reset_func(struct work_struct *ws) +{ + struct vpu_wdt *wdt = container_of(ws, struct vpu_wdt, ws); + struct mtk_vpu *vpu = container_of(wdt, struct mtk_vpu, wdt); + struct vpu_wdt_handler *handler = wdt->handler; + int index, ret; + + dev_info(vpu->dev, "vpu reset\n"); + ret = vpu_clock_enable(vpu); + if (ret) { + dev_err(vpu->dev, "[VPU] wdt enables clock failed %d\n", ret); + return; + } + mutex_lock(&vpu->vpu_mutex); + vpu_cfg_writel(vpu, 0x0, VPU_RESET); + vpu->fw_loaded = false; + mutex_unlock(&vpu->vpu_mutex); + vpu_clock_disable(vpu); + + for (index = 0; index < VPU_RST_MAX; index++) { + if (handler[index].reset_func) { + handler[index].reset_func(handler[index].priv); + dev_dbg(vpu->dev, "wdt handler func %d\n", index); + } + } +} + +int vpu_wdt_reg_handler(struct platform_device *pdev, + void wdt_reset(void *), + void *priv, enum rst_id id) +{ + struct mtk_vpu *vpu = platform_get_drvdata(pdev); + struct vpu_wdt_handler *handler; + + if (!vpu) { + dev_err(&pdev->dev, "vpu device in not ready\n"); + return -EPROBE_DEFER; + } + + handler = vpu->wdt.handler; + + if (id < VPU_RST_MAX && wdt_reset) { + dev_dbg(vpu->dev, "wdt register id %d\n", id); + mutex_lock(&vpu->vpu_mutex); + handler[id].reset_func = wdt_reset; + handler[id].priv = priv; + mutex_unlock(&vpu->vpu_mutex); + return 0; + } + + dev_err(vpu->dev, "register vpu wdt handler failed\n"); + return -EINVAL; +} +EXPORT_SYMBOL_GPL(vpu_wdt_reg_handler); + +unsigned int vpu_get_vdec_hw_capa(struct platform_device *pdev) +{ + struct mtk_vpu *vpu = platform_get_drvdata(pdev); + + return vpu->run.dec_capability; +} +EXPORT_SYMBOL_GPL(vpu_get_vdec_hw_capa); + +unsigned int vpu_get_venc_hw_capa(struct platform_device *pdev) +{ + struct mtk_vpu *vpu = platform_get_drvdata(pdev); + + return vpu->run.enc_capability; +} +EXPORT_SYMBOL_GPL(vpu_get_venc_hw_capa); + +void *vpu_mapping_dm_addr(struct platform_device *pdev, + u32 dtcm_dmem_addr) +{ + struct mtk_vpu *vpu = platform_get_drvdata(pdev); + + if (!dtcm_dmem_addr || + (dtcm_dmem_addr > (VPU_DTCM_SIZE + VPU_EXT_D_SIZE))) { + dev_err(vpu->dev, "invalid virtual data memory address\n"); + return ERR_PTR(-EINVAL); + } + + if (dtcm_dmem_addr < VPU_DTCM_SIZE) + return (__force void *)(dtcm_dmem_addr + vpu->reg.tcm + + VPU_DTCM_OFFSET); + + return vpu->extmem[D_FW].va + (dtcm_dmem_addr - VPU_DTCM_SIZE); +} +EXPORT_SYMBOL_GPL(vpu_mapping_dm_addr); + +struct platform_device *vpu_get_plat_device(struct platform_device *pdev) +{ + struct device *dev = &pdev->dev; + struct device_node *vpu_node; + struct platform_device *vpu_pdev; + + vpu_node = of_parse_phandle(dev->of_node, "mediatek,vpu", 0); + if (!vpu_node) { + dev_err(dev, "can't get vpu node\n"); + return NULL; + } + + vpu_pdev = of_find_device_by_node(vpu_node); + of_node_put(vpu_node); + if (WARN_ON(!vpu_pdev)) { + dev_err(dev, "vpu pdev failed\n"); + return NULL; + } + + return vpu_pdev; +} +EXPORT_SYMBOL_GPL(vpu_get_plat_device); + +/* load vpu program/data memory */ +static int load_requested_vpu(struct mtk_vpu *vpu, + u8 fw_type) +{ + size_t tcm_size = fw_type ? VPU_DTCM_SIZE : VPU_PTCM_SIZE; + size_t fw_size = fw_type ? VPU_D_FW_SIZE : VPU_P_FW_SIZE; + char *fw_name = fw_type ? VPU_D_FW : VPU_P_FW; + char *fw_new_name = fw_type ? VPU_D_FW_NEW : VPU_P_FW_NEW; + const struct firmware *vpu_fw; + size_t dl_size = 0; + size_t extra_fw_size = 0; + void *dest; + int ret; + + ret = request_firmware(&vpu_fw, fw_new_name, vpu->dev); + if (ret < 0) { + dev_info(vpu->dev, "Failed to load %s, %d, retry\n", + fw_new_name, ret); + + ret = request_firmware(&vpu_fw, fw_name, vpu->dev); + if (ret < 0) { + dev_err(vpu->dev, "Failed to load %s, %d\n", fw_name, + ret); + return ret; + } + } + dl_size = vpu_fw->size; + if (dl_size > fw_size) { + dev_err(vpu->dev, "fw %s size %zu is abnormal\n", fw_name, + dl_size); + release_firmware(vpu_fw); + return -EFBIG; + } + dev_dbg(vpu->dev, "Downloaded fw %s size: %zu.\n", + fw_name, + dl_size); + /* reset VPU */ + vpu_cfg_writel(vpu, 0x0, VPU_RESET); + + /* handle extended firmware size */ + if (dl_size > tcm_size) { + dev_dbg(vpu->dev, "fw size %zu > limited fw size %zu\n", + dl_size, tcm_size); + extra_fw_size = dl_size - tcm_size; + dev_dbg(vpu->dev, "extra_fw_size %zu\n", extra_fw_size); + dl_size = tcm_size; + } + dest = (__force void *)vpu->reg.tcm; + if (fw_type == D_FW) + dest += VPU_DTCM_OFFSET; + memcpy(dest, vpu_fw->data, dl_size); + /* download to extended memory if need */ + if (extra_fw_size > 0) { + dest = vpu->extmem[fw_type].va; + dev_dbg(vpu->dev, "download extended memory type %x\n", + fw_type); + memcpy(dest, vpu_fw->data + tcm_size, extra_fw_size); + } + + release_firmware(vpu_fw); + + return 0; +} + +int vpu_load_firmware(struct platform_device *pdev) +{ + struct mtk_vpu *vpu; + struct device *dev; + struct vpu_run *run; + int ret; + + if (!pdev) { + pr_err("VPU platform device is invalid\n"); + return -EINVAL; + } + + dev = &pdev->dev; + + vpu = platform_get_drvdata(pdev); + run = &vpu->run; + + mutex_lock(&vpu->vpu_mutex); + if (vpu->fw_loaded) { + mutex_unlock(&vpu->vpu_mutex); + return 0; + } + mutex_unlock(&vpu->vpu_mutex); + + ret = vpu_clock_enable(vpu); + if (ret) { + dev_err(dev, "enable clock failed %d\n", ret); + return ret; + } + + mutex_lock(&vpu->vpu_mutex); + + run->signaled = false; + dev_dbg(vpu->dev, "firmware request\n"); + /* Downloading program firmware to device*/ + ret = load_requested_vpu(vpu, P_FW); + if (ret < 0) { + dev_err(dev, "Failed to request %s, %d\n", VPU_P_FW, ret); + goto OUT_LOAD_FW; + } + + /* Downloading data firmware to device */ + ret = load_requested_vpu(vpu, D_FW); + if (ret < 0) { + dev_err(dev, "Failed to request %s, %d\n", VPU_D_FW, ret); + goto OUT_LOAD_FW; + } + + vpu->fw_loaded = true; + /* boot up vpu */ + vpu_cfg_writel(vpu, 0x1, VPU_RESET); + + ret = wait_event_interruptible_timeout(run->wq, + run->signaled, + msecs_to_jiffies(INIT_TIMEOUT_MS) + ); + if (ret == 0) { + ret = -ETIME; + dev_err(dev, "wait vpu initialization timeout!\n"); + goto OUT_LOAD_FW; + } else if (-ERESTARTSYS == ret) { + dev_err(dev, "wait vpu interrupted by a signal!\n"); + goto OUT_LOAD_FW; + } + + ret = 0; + dev_info(dev, "vpu is ready. Fw version %s\n", run->fw_ver); + +OUT_LOAD_FW: + mutex_unlock(&vpu->vpu_mutex); + vpu_clock_disable(vpu); + + return ret; +} +EXPORT_SYMBOL_GPL(vpu_load_firmware); + +static void vpu_init_ipi_handler(const void *data, unsigned int len, void *priv) +{ + struct mtk_vpu *vpu = priv; + const struct vpu_run *run = data; + + vpu->run.signaled = run->signaled; + strscpy(vpu->run.fw_ver, run->fw_ver, sizeof(vpu->run.fw_ver)); + vpu->run.dec_capability = run->dec_capability; + vpu->run.enc_capability = run->enc_capability; + wake_up_interruptible(&vpu->run.wq); +} + +#ifdef CONFIG_DEBUG_FS +static ssize_t vpu_debug_read(struct file *file, char __user *user_buf, + size_t count, loff_t *ppos) +{ + char buf[256]; + unsigned int len; + unsigned int running, pc, vpu_to_host, host_to_vpu, wdt, idle, ra, sp; + int ret; + struct device *dev = file->private_data; + struct mtk_vpu *vpu = dev_get_drvdata(dev); + + ret = vpu_clock_enable(vpu); + if (ret) { + dev_err(vpu->dev, "[VPU] enable clock failed %d\n", ret); + return 0; + } + + /* vpu register status */ + running = vpu_running(vpu); + pc = vpu_cfg_readl(vpu, VPU_PC_REG); + wdt = vpu_cfg_readl(vpu, VPU_WDT_REG); + host_to_vpu = vpu_cfg_readl(vpu, HOST_TO_VPU); + vpu_to_host = vpu_cfg_readl(vpu, VPU_TO_HOST); + ra = vpu_cfg_readl(vpu, VPU_RA_REG); + sp = vpu_cfg_readl(vpu, VPU_SP_REG); + idle = vpu_cfg_readl(vpu, VPU_IDLE_REG); + + vpu_clock_disable(vpu); + + if (running) { + len = snprintf(buf, sizeof(buf), "VPU is running\n\n" + "FW Version: %s\n" + "PC: 0x%x\n" + "WDT: 0x%x\n" + "Host to VPU: 0x%x\n" + "VPU to Host: 0x%x\n" + "SP: 0x%x\n" + "RA: 0x%x\n" + "idle: 0x%x\n", + vpu->run.fw_ver, pc, wdt, + host_to_vpu, vpu_to_host, sp, ra, idle); + } else { + len = snprintf(buf, sizeof(buf), "VPU not running\n"); + } + + return simple_read_from_buffer(user_buf, count, ppos, buf, len); +} + +static const struct file_operations vpu_debug_fops = { + .open = simple_open, + .read = vpu_debug_read, +}; +#endif /* CONFIG_DEBUG_FS */ + +static void vpu_free_ext_mem(struct mtk_vpu *vpu, u8 fw_type) +{ + struct device *dev = vpu->dev; + size_t fw_ext_size = fw_type ? VPU_EXT_D_SIZE : VPU_EXT_P_SIZE; + + dma_free_coherent(dev, fw_ext_size, vpu->extmem[fw_type].va, + vpu->extmem[fw_type].pa); +} + +static int vpu_alloc_ext_mem(struct mtk_vpu *vpu, u32 fw_type) +{ + struct device *dev = vpu->dev; + size_t fw_ext_size = fw_type ? VPU_EXT_D_SIZE : VPU_EXT_P_SIZE; + u32 vpu_ext_mem0 = fw_type ? VPU_DMEM_EXT0_ADDR : VPU_PMEM_EXT0_ADDR; + u32 vpu_ext_mem1 = fw_type ? VPU_DMEM_EXT1_ADDR : VPU_PMEM_EXT1_ADDR; + u32 offset_4gb = vpu->enable_4GB ? 0x40000000 : 0; + + vpu->extmem[fw_type].va = dma_alloc_coherent(dev, + fw_ext_size, + &vpu->extmem[fw_type].pa, + GFP_KERNEL); + if (!vpu->extmem[fw_type].va) { + dev_err(dev, "Failed to allocate the extended program memory\n"); + return -ENOMEM; + } + + /* Disable extend0. Enable extend1 */ + vpu_cfg_writel(vpu, 0x1, vpu_ext_mem0); + vpu_cfg_writel(vpu, (vpu->extmem[fw_type].pa & 0xFFFFF000) + offset_4gb, + vpu_ext_mem1); + + dev_info(dev, "%s extend memory phy=0x%llx virt=0x%p\n", + fw_type ? "Data" : "Program", + (unsigned long long)vpu->extmem[fw_type].pa, + vpu->extmem[fw_type].va); + + return 0; +} + +static void vpu_ipi_handler(struct mtk_vpu *vpu) +{ + struct share_obj __iomem *rcv_obj = vpu->recv_buf; + struct vpu_ipi_desc *ipi_desc = vpu->ipi_desc; + unsigned char data[SHARE_BUF_SIZE]; + s32 id = readl(&rcv_obj->id); + + memcpy_fromio(data, rcv_obj->share_buf, sizeof(data)); + if (id < IPI_MAX && ipi_desc[id].handler) { + ipi_desc[id].handler(data, readl(&rcv_obj->len), + ipi_desc[id].priv); + if (id > IPI_VPU_INIT) { + vpu->ipi_id_ack[id] = true; + wake_up(&vpu->ack_wq); + } + } else { + dev_err(vpu->dev, "No such ipi id = %d\n", id); + } +} + +static int vpu_ipi_init(struct mtk_vpu *vpu) +{ + /* Disable VPU to host interrupt */ + vpu_cfg_writel(vpu, 0x0, VPU_TO_HOST); + + /* shared buffer initialization */ + vpu->recv_buf = vpu->reg.tcm + VPU_DTCM_OFFSET; + vpu->send_buf = vpu->recv_buf + 1; + memset_io(vpu->recv_buf, 0, sizeof(struct share_obj)); + memset_io(vpu->send_buf, 0, sizeof(struct share_obj)); + + return 0; +} + +static irqreturn_t vpu_irq_handler(int irq, void *priv) +{ + struct mtk_vpu *vpu = priv; + u32 vpu_to_host; + int ret; + + /* + * Clock should have been enabled already. + * Enable again in case vpu_ipi_send times out + * and has disabled the clock. + */ + ret = clk_enable(vpu->clk); + if (ret) { + dev_err(vpu->dev, "[VPU] enable clock failed %d\n", ret); + return IRQ_NONE; + } + vpu_to_host = vpu_cfg_readl(vpu, VPU_TO_HOST); + if (vpu_to_host & VPU_IPC_INT) { + vpu_ipi_handler(vpu); + } else { + dev_err(vpu->dev, "vpu watchdog timeout! 0x%x", vpu_to_host); + queue_work(vpu->wdt.wq, &vpu->wdt.ws); + } + + /* VPU won't send another interrupt until we set VPU_TO_HOST to 0. */ + vpu_cfg_writel(vpu, 0x0, VPU_TO_HOST); + clk_disable(vpu->clk); + + return IRQ_HANDLED; +} + +#ifdef CONFIG_DEBUG_FS +static struct dentry *vpu_debugfs; +#endif +static int mtk_vpu_probe(struct platform_device *pdev) +{ + struct mtk_vpu *vpu; + struct device *dev; + int ret = 0; + + dev_dbg(&pdev->dev, "initialization\n"); + + dev = &pdev->dev; + vpu = devm_kzalloc(dev, sizeof(*vpu), GFP_KERNEL); + if (!vpu) + return -ENOMEM; + + vpu->dev = &pdev->dev; + vpu->reg.tcm = devm_platform_ioremap_resource_byname(pdev, "tcm"); + if (IS_ERR((__force void *)vpu->reg.tcm)) + return PTR_ERR((__force void *)vpu->reg.tcm); + + vpu->reg.cfg = devm_platform_ioremap_resource_byname(pdev, "cfg_reg"); + if (IS_ERR((__force void *)vpu->reg.cfg)) + return PTR_ERR((__force void *)vpu->reg.cfg); + + /* Get VPU clock */ + vpu->clk = devm_clk_get(dev, "main"); + if (IS_ERR(vpu->clk)) { + dev_err(dev, "get vpu clock failed\n"); + return PTR_ERR(vpu->clk); + } + + platform_set_drvdata(pdev, vpu); + + ret = clk_prepare(vpu->clk); + if (ret) { + dev_err(dev, "prepare vpu clock failed\n"); + return ret; + } + + /* VPU watchdog */ + vpu->wdt.wq = create_singlethread_workqueue("vpu_wdt"); + if (!vpu->wdt.wq) { + dev_err(dev, "initialize wdt workqueue failed\n"); + ret = -ENOMEM; + goto clk_unprepare; + } + INIT_WORK(&vpu->wdt.ws, vpu_wdt_reset_func); + mutex_init(&vpu->vpu_mutex); + + ret = vpu_clock_enable(vpu); + if (ret) { + dev_err(dev, "enable vpu clock failed\n"); + goto workqueue_destroy; + } + + dev_dbg(dev, "vpu ipi init\n"); + ret = vpu_ipi_init(vpu); + if (ret) { + dev_err(dev, "Failed to init ipi\n"); + goto disable_vpu_clk; + } + + /* register vpu initialization IPI */ + ret = vpu_ipi_register(pdev, IPI_VPU_INIT, vpu_init_ipi_handler, + "vpu_init", vpu); + if (ret) { + dev_err(dev, "Failed to register IPI_VPU_INIT\n"); + goto vpu_mutex_destroy; + } + +#ifdef CONFIG_DEBUG_FS + vpu_debugfs = debugfs_create_file("mtk_vpu", S_IRUGO, NULL, (void *)dev, + &vpu_debug_fops); +#endif + + /* Set PTCM to 96K and DTCM to 32K */ + vpu_cfg_writel(vpu, 0x2, VPU_TCM_CFG); + + vpu->enable_4GB = !!(totalram_pages() > (SZ_2G >> PAGE_SHIFT)); + dev_info(dev, "4GB mode %u\n", vpu->enable_4GB); + + if (vpu->enable_4GB) { + ret = of_reserved_mem_device_init(dev); + if (ret) + dev_info(dev, "init reserved memory failed\n"); + /* continue to use dynamic allocation if failed */ + } + + ret = vpu_alloc_ext_mem(vpu, D_FW); + if (ret) { + dev_err(dev, "Allocate DM failed\n"); + goto remove_debugfs; + } + + ret = vpu_alloc_ext_mem(vpu, P_FW); + if (ret) { + dev_err(dev, "Allocate PM failed\n"); + goto free_d_mem; + } + + init_waitqueue_head(&vpu->run.wq); + init_waitqueue_head(&vpu->ack_wq); + + ret = platform_get_irq(pdev, 0); + if (ret < 0) + goto free_p_mem; + vpu->reg.irq = ret; + ret = devm_request_irq(dev, vpu->reg.irq, vpu_irq_handler, 0, + pdev->name, vpu); + if (ret) { + dev_err(dev, "failed to request irq\n"); + goto free_p_mem; + } + + vpu_clock_disable(vpu); + dev_dbg(dev, "initialization completed\n"); + + return 0; + +free_p_mem: + vpu_free_ext_mem(vpu, P_FW); +free_d_mem: + vpu_free_ext_mem(vpu, D_FW); +remove_debugfs: + of_reserved_mem_device_release(dev); +#ifdef CONFIG_DEBUG_FS + debugfs_remove(vpu_debugfs); +#endif + memset(vpu->ipi_desc, 0, sizeof(struct vpu_ipi_desc) * IPI_MAX); +vpu_mutex_destroy: + mutex_destroy(&vpu->vpu_mutex); +disable_vpu_clk: + vpu_clock_disable(vpu); +workqueue_destroy: + destroy_workqueue(vpu->wdt.wq); +clk_unprepare: + clk_unprepare(vpu->clk); + + return ret; +} + +static const struct of_device_id mtk_vpu_match[] = { + { + .compatible = "mediatek,mt8173-vpu", + }, + {}, +}; +MODULE_DEVICE_TABLE(of, mtk_vpu_match); + +static void mtk_vpu_remove(struct platform_device *pdev) +{ + struct mtk_vpu *vpu = platform_get_drvdata(pdev); + +#ifdef CONFIG_DEBUG_FS + debugfs_remove(vpu_debugfs); +#endif + if (vpu->wdt.wq) + destroy_workqueue(vpu->wdt.wq); + vpu_free_ext_mem(vpu, P_FW); + vpu_free_ext_mem(vpu, D_FW); + mutex_destroy(&vpu->vpu_mutex); + clk_unprepare(vpu->clk); +} + +static int mtk_vpu_suspend(struct device *dev) +{ + struct mtk_vpu *vpu = dev_get_drvdata(dev); + unsigned long timeout; + int ret; + + ret = vpu_clock_enable(vpu); + if (ret) { + dev_err(dev, "failed to enable vpu clock\n"); + return ret; + } + + if (!vpu_running(vpu)) { + vpu_clock_disable(vpu); + clk_unprepare(vpu->clk); + return 0; + } + + mutex_lock(&vpu->vpu_mutex); + /* disable vpu timer interrupt */ + vpu_cfg_writel(vpu, vpu_cfg_readl(vpu, VPU_INT_STATUS) | VPU_IDLE_STATE, + VPU_INT_STATUS); + /* check if vpu is idle for system suspend */ + timeout = jiffies + msecs_to_jiffies(VPU_IDLE_TIMEOUT_MS); + do { + if (time_after(jiffies, timeout)) { + dev_err(dev, "vpu idle timeout\n"); + mutex_unlock(&vpu->vpu_mutex); + vpu_clock_disable(vpu); + return -EIO; + } + } while (!vpu_cfg_readl(vpu, VPU_IDLE_REG)); + + mutex_unlock(&vpu->vpu_mutex); + vpu_clock_disable(vpu); + clk_unprepare(vpu->clk); + + return 0; +} + +static int mtk_vpu_resume(struct device *dev) +{ + struct mtk_vpu *vpu = dev_get_drvdata(dev); + int ret; + + clk_prepare(vpu->clk); + ret = vpu_clock_enable(vpu); + if (ret) { + clk_unprepare(vpu->clk); + dev_err(dev, "failed to enable vpu clock\n"); + return ret; + } + + mutex_lock(&vpu->vpu_mutex); + /* enable vpu timer interrupt */ + vpu_cfg_writel(vpu, + vpu_cfg_readl(vpu, VPU_INT_STATUS) & ~(VPU_IDLE_STATE), + VPU_INT_STATUS); + mutex_unlock(&vpu->vpu_mutex); + vpu_clock_disable(vpu); + + return 0; +} + +static const struct dev_pm_ops mtk_vpu_pm = { + .suspend = mtk_vpu_suspend, + .resume = mtk_vpu_resume, +}; + +static struct platform_driver mtk_vpu_driver = { + .probe = mtk_vpu_probe, + .remove_new = mtk_vpu_remove, + .driver = { + .name = "mtk_vpu", + .pm = &mtk_vpu_pm, + .of_match_table = mtk_vpu_match, + }, +}; + +module_platform_driver(mtk_vpu_driver); + +MODULE_LICENSE("GPL v2"); +MODULE_DESCRIPTION("Mediatek Video Processor Unit driver"); diff --git a/drivers/media/platform/mediatek/vpu/mtk_vpu.h b/drivers/media/platform/mediatek/vpu/mtk_vpu.h new file mode 100644 index 0000000000..a56053ff13 --- /dev/null +++ b/drivers/media/platform/mediatek/vpu/mtk_vpu.h @@ -0,0 +1,188 @@ +/* SPDX-License-Identifier: GPL-2.0-only */ +/* +* Copyright (c) 2016 MediaTek Inc. +* Author: Andrew-CT Chen <andrew-ct.chen@mediatek.com> +*/ + +#ifndef _MTK_VPU_H +#define _MTK_VPU_H + +#include <linux/platform_device.h> + +/** + * DOC: VPU + * + * VPU (video processor unit) is a tiny processor controlling video hardware + * related to video codec, scaling and color format converting. + * VPU interfaces with other blocks by share memory and interrupt. + */ + +typedef void (*ipi_handler_t) (const void *data, + unsigned int len, + void *priv); + +/** + * enum ipi_id - the id of inter-processor interrupt + * + * @IPI_VPU_INIT: The interrupt from vpu is to notfiy kernel + * VPU initialization completed. + * IPI_VPU_INIT is sent from VPU when firmware is + * loaded. AP doesn't need to send IPI_VPU_INIT + * command to VPU. + * For other IPI below, AP should send the request + * to VPU to trigger the interrupt. + * @IPI_VDEC_H264: The interrupt from vpu is to notify kernel to + * handle H264 vidoe decoder job, and vice versa. + * Decode output format is always MT21 no matter what + * the input format is. + * @IPI_VDEC_VP8: The interrupt from is to notify kernel to + * handle VP8 video decoder job, and vice versa. + * Decode output format is always MT21 no matter what + * the input format is. + * @IPI_VDEC_VP9: The interrupt from vpu is to notify kernel to + * handle VP9 video decoder job, and vice versa. + * Decode output format is always MT21 no matter what + * the input format is. + * @IPI_VENC_H264: The interrupt from vpu is to notify kernel to + * handle H264 video encoder job, and vice versa. + * @IPI_VENC_VP8: The interrupt fro vpu is to notify kernel to + * handle VP8 video encoder job,, and vice versa. + * @IPI_MDP: The interrupt from vpu is to notify kernel to + * handle MDP (Media Data Path) job, and vice versa. + * @IPI_MAX: The maximum IPI number + */ + +enum ipi_id { + IPI_VPU_INIT = 0, + IPI_VDEC_H264, + IPI_VDEC_VP8, + IPI_VDEC_VP9, + IPI_VENC_H264, + IPI_VENC_VP8, + IPI_MDP, + IPI_MAX, +}; + +/** + * enum rst_id - reset id to register reset function for VPU watchdog timeout + * + * @VPU_RST_ENC: encoder reset id + * @VPU_RST_DEC: decoder reset id + * @VPU_RST_MDP: MDP (Media Data Path) reset id + * @VPU_RST_MAX: maximum reset id + */ +enum rst_id { + VPU_RST_ENC, + VPU_RST_DEC, + VPU_RST_MDP, + VPU_RST_MAX, +}; + +/** + * vpu_ipi_register - register an ipi function + * + * @pdev: VPU platform device + * @id: IPI ID + * @handler: IPI handler + * @name: IPI name + * @priv: private data for IPI handler + * + * Register an ipi function to receive ipi interrupt from VPU. + * + * Return: Return 0 if ipi registers successfully, otherwise it is failed. + */ +int vpu_ipi_register(struct platform_device *pdev, enum ipi_id id, + ipi_handler_t handler, const char *name, void *priv); + +/** + * vpu_ipi_send - send data from AP to vpu. + * + * @pdev: VPU platform device + * @id: IPI ID + * @buf: the data buffer + * @len: the data buffer length + * + * This function is thread-safe. When this function returns, + * VPU has received the data and starts the processing. + * When the processing completes, IPI handler registered + * by vpu_ipi_register will be called in interrupt context. + * + * Return: Return 0 if sending data successfully, otherwise it is failed. + **/ +int vpu_ipi_send(struct platform_device *pdev, + enum ipi_id id, void *buf, + unsigned int len); + +/** + * vpu_get_plat_device - get VPU's platform device + * + * @pdev: the platform device of the module requesting VPU platform + * device for using VPU API. + * + * Return: Return NULL if it is failed. + * otherwise it is VPU's platform device + **/ +struct platform_device *vpu_get_plat_device(struct platform_device *pdev); + +/** + * vpu_wdt_reg_handler - register a VPU watchdog handler + * + * @pdev: VPU platform device + * @vpu_wdt_reset_func(): the callback reset function + * @priv: the private data for reset function + * @priv: the private data for reset function + * @id: reset id + * + * Register a handler performing own tasks when vpu reset by watchdog + * + * Return: Return 0 if the handler is added successfully, + * otherwise it is failed. + **/ +int vpu_wdt_reg_handler(struct platform_device *pdev, + void vpu_wdt_reset_func(void *priv), + void *priv, enum rst_id id); + +/** + * vpu_get_vdec_hw_capa - get video decoder hardware capability + * + * @pdev: VPU platform device + * + * Return: video decoder hardware capability + **/ +unsigned int vpu_get_vdec_hw_capa(struct platform_device *pdev); + +/** + * vpu_get_venc_hw_capa - get video encoder hardware capability + * + * @pdev: VPU platform device + * + * Return: video encoder hardware capability + **/ +unsigned int vpu_get_venc_hw_capa(struct platform_device *pdev); + +/** + * vpu_load_firmware - download VPU firmware and boot it + * + * @pdev: VPU platform device + * + * Return: Return 0 if downloading firmware successfully, + * otherwise it is failed + **/ +int vpu_load_firmware(struct platform_device *pdev); + +/** + * vpu_mapping_dm_addr - Mapping DTCM/DMEM to kernel virtual address + * + * @pdev: VPU platform device + * @dtcm_dmem_addr: VPU's data memory address + * + * Mapping the VPU's DTCM (Data Tightly-Coupled Memory) / + * DMEM (Data Extended Memory) memory address to + * kernel virtual address. + * + * Return: Return ERR_PTR(-EINVAL) if mapping failed, + * otherwise the mapped kernel virtual address + **/ +void *vpu_mapping_dm_addr(struct platform_device *pdev, + u32 dtcm_dmem_addr); +#endif /* _MTK_VPU_H */ |