Adding upstream version 6.1.76.upstream/6.1.76

Signed-off-by: Daniel Baumann <daniel.baumann@progress-linux.org>

11 files changed, 3270 insertions, 0 deletions

diff --git a/drivers/media/platform/nvidia/Kconfig b/drivers/media/platform/nvidia/Kconfig
new file mode 100644
index 000000000..b211b4687
--- /dev/null
+++ b/drivers/media/platform/nvidia/Kconfig
@@ -0,0 +1,5 @@
+# SPDX-License-Identifier: GPL-2.0-only
+
+comment "NVidia media platform drivers"
+
+source "drivers/media/platform/nvidia/tegra-vde/Kconfig"
diff --git a/drivers/media/platform/nvidia/Makefile b/drivers/media/platform/nvidia/Makefile
new file mode 100644
index 000000000..428415ff8
--- /dev/null
+++ b/drivers/media/platform/nvidia/Makefile
@@ -0,0 +1,3 @@
+# SPDX-License-Identifier: GPL-2.0
+
+obj-y += tegra-vde/
diff --git a/drivers/media/platform/nvidia/tegra-vde/Kconfig b/drivers/media/platform/nvidia/tegra-vde/Kconfig
new file mode 100644
index 000000000..f7454823b
--- /dev/null
+++ b/drivers/media/platform/nvidia/tegra-vde/Kconfig
@@ -0,0 +1,17 @@
+config VIDEO_TEGRA_VDE
+	tristate "NVIDIA Tegra Video Decoder Engine driver"
+	depends on V4L_MEM2MEM_DRIVERS
+	depends on ARCH_TEGRA || COMPILE_TEST
+	depends on VIDEO_DEV
+	select DMA_SHARED_BUFFER
+	select IOMMU_IOVA
+	select MEDIA_CONTROLLER
+	select MEDIA_CONTROLLER_REQUEST_API
+	select SRAM
+	select VIDEOBUF2_DMA_CONTIG
+	select VIDEOBUF2_DMA_SG
+	select V4L2_H264
+	select V4L2_MEM2MEM_DEV
+	help
+	   Support for the NVIDIA Tegra video decoder unit.
+	   To compile this driver as a module choose m here.
diff --git a/drivers/media/platform/nvidia/tegra-vde/Makefile b/drivers/media/platform/nvidia/tegra-vde/Makefile
new file mode 100644
index 000000000..4e96f3305
--- /dev/null
+++ b/drivers/media/platform/nvidia/tegra-vde/Makefile
@@ -0,0 +1,3 @@
+# SPDX-License-Identifier: GPL-2.0
+tegra-vde-y := vde.o iommu.o dmabuf-cache.o h264.o v4l2.o
+obj-$(CONFIG_VIDEO_TEGRA_VDE)	+= tegra-vde.o
diff --git a/drivers/media/platform/nvidia/tegra-vde/dmabuf-cache.c b/drivers/media/platform/nvidia/tegra-vde/dmabuf-cache.c
new file mode 100644
index 000000000..69c346148
--- /dev/null
+++ b/drivers/media/platform/nvidia/tegra-vde/dmabuf-cache.c
@@ -0,0 +1,229 @@
+// SPDX-License-Identifier: GPL-2.0+
+/*
+ * NVIDIA Tegra Video decoder driver
+ *
+ * Copyright (C) 2016-2019 GRATE-DRIVER project
+ */
+
+#include <linux/dma-buf.h>
+#include <linux/iova.h>
+#include <linux/kernel.h>
+#include <linux/list.h>
+#include <linux/sched.h>
+#include <linux/slab.h>
+#include <linux/workqueue.h>
+#include <linux/module.h>
+
+#include "vde.h"
+
+MODULE_IMPORT_NS(DMA_BUF);
+
+struct tegra_vde_cache_entry {
+	enum dma_data_direction dma_dir;
+	struct dma_buf_attachment *a;
+	struct delayed_work dwork;
+	struct tegra_vde *vde;
+	struct list_head list;
+	struct sg_table *sgt;
+	struct iova *iova;
+	unsigned int refcnt;
+};
+
+static void tegra_vde_release_entry(struct tegra_vde_cache_entry *entry)
+{
+	struct dma_buf *dmabuf = entry->a->dmabuf;
+
+	WARN_ON_ONCE(entry->refcnt);
+
+	if (entry->vde->domain)
+		tegra_vde_iommu_unmap(entry->vde, entry->iova);
+
+	dma_buf_unmap_attachment(entry->a, entry->sgt, entry->dma_dir);
+	dma_buf_detach(dmabuf, entry->a);
+	dma_buf_put(dmabuf);
+
+	list_del(&entry->list);
+	kfree(entry);
+}
+
+static void tegra_vde_delayed_unmap(struct work_struct *work)
+{
+	struct tegra_vde_cache_entry *entry;
+	struct tegra_vde *vde;
+
+	entry = container_of(work, struct tegra_vde_cache_entry,
+			     dwork.work);
+	vde = entry->vde;
+
+	mutex_lock(&vde->map_lock);
+	tegra_vde_release_entry(entry);
+	mutex_unlock(&vde->map_lock);
+}
+
+int tegra_vde_dmabuf_cache_map(struct tegra_vde *vde,
+			       struct dma_buf *dmabuf,
+			       enum dma_data_direction dma_dir,
+			       struct dma_buf_attachment **ap,
+			       dma_addr_t *addrp)
+{
+	struct dma_buf_attachment *attachment;
+	struct tegra_vde_cache_entry *entry;
+	struct device *dev = vde->dev;
+	struct sg_table *sgt;
+	struct iova *iova;
+	int err;
+
+	mutex_lock(&vde->map_lock);
+
+	list_for_each_entry(entry, &vde->map_list, list) {
+		if (entry->a->dmabuf != dmabuf)
+			continue;
+
+		if (!cancel_delayed_work(&entry->dwork))
+			continue;
+
+		if (entry->dma_dir != dma_dir)
+			entry->dma_dir = DMA_BIDIRECTIONAL;
+
+		dma_buf_put(dmabuf);
+
+		if (vde->domain)
+			*addrp = iova_dma_addr(&vde->iova, entry->iova);
+		else
+			*addrp = sg_dma_address(entry->sgt->sgl);
+
+		goto ref;
+	}
+
+	attachment = dma_buf_attach(dmabuf, dev);
+	if (IS_ERR(attachment)) {
+		dev_err(dev, "Failed to attach dmabuf\n");
+		err = PTR_ERR(attachment);
+		goto err_unlock;
+	}
+
+	sgt = dma_buf_map_attachment(attachment, dma_dir);
+	if (IS_ERR(sgt)) {
+		dev_err(dev, "Failed to get dmabufs sg_table\n");
+		err = PTR_ERR(sgt);
+		goto err_detach;
+	}
+
+	if (!vde->domain && sgt->nents > 1) {
+		dev_err(dev, "Sparse DMA region is unsupported, please enable IOMMU\n");
+		err = -EINVAL;
+		goto err_unmap;
+	}
+
+	entry = kzalloc(sizeof(*entry), GFP_KERNEL);
+	if (!entry) {
+		err = -ENOMEM;
+		goto err_unmap;
+	}
+
+	if (vde->domain) {
+		err = tegra_vde_iommu_map(vde, sgt, &iova, dmabuf->size);
+		if (err)
+			goto err_free;
+
+		*addrp = iova_dma_addr(&vde->iova, iova);
+	} else {
+		*addrp = sg_dma_address(sgt->sgl);
+		iova = NULL;
+	}
+
+	INIT_DELAYED_WORK(&entry->dwork, tegra_vde_delayed_unmap);
+	list_add(&entry->list, &vde->map_list);
+
+	entry->dma_dir = dma_dir;
+	entry->iova = iova;
+	entry->vde = vde;
+	entry->sgt = sgt;
+	entry->a = attachment;
+ref:
+	entry->refcnt++;
+
+	*ap = entry->a;
+
+	mutex_unlock(&vde->map_lock);
+
+	return 0;
+
+err_free:
+	kfree(entry);
+err_unmap:
+	dma_buf_unmap_attachment(attachment, sgt, dma_dir);
+err_detach:
+	dma_buf_detach(dmabuf, attachment);
+err_unlock:
+	mutex_unlock(&vde->map_lock);
+
+	return err;
+}
+
+void tegra_vde_dmabuf_cache_unmap(struct tegra_vde *vde,
+				  struct dma_buf_attachment *a,
+				  bool release)
+{
+	struct tegra_vde_cache_entry *entry;
+
+	mutex_lock(&vde->map_lock);
+
+	list_for_each_entry(entry, &vde->map_list, list) {
+		if (entry->a != a)
+			continue;
+
+		WARN_ON_ONCE(!entry->refcnt);
+
+		if (--entry->refcnt == 0) {
+			if (release)
+				tegra_vde_release_entry(entry);
+			else
+				schedule_delayed_work(&entry->dwork, 5 * HZ);
+		}
+		break;
+	}
+
+	mutex_unlock(&vde->map_lock);
+}
+
+void tegra_vde_dmabuf_cache_unmap_sync(struct tegra_vde *vde)
+{
+	struct tegra_vde_cache_entry *entry, *tmp;
+
+	mutex_lock(&vde->map_lock);
+
+	list_for_each_entry_safe(entry, tmp, &vde->map_list, list) {
+		if (entry->refcnt)
+			continue;
+
+		if (!cancel_delayed_work(&entry->dwork))
+			continue;
+
+		tegra_vde_release_entry(entry);
+	}
+
+	mutex_unlock(&vde->map_lock);
+}
+
+void tegra_vde_dmabuf_cache_unmap_all(struct tegra_vde *vde)
+{
+	struct tegra_vde_cache_entry *entry, *tmp;
+
+	mutex_lock(&vde->map_lock);
+
+	while (!list_empty(&vde->map_list)) {
+		list_for_each_entry_safe(entry, tmp, &vde->map_list, list) {
+			if (!cancel_delayed_work(&entry->dwork))
+				continue;
+
+			tegra_vde_release_entry(entry);
+		}
+
+		mutex_unlock(&vde->map_lock);
+		schedule();
+		mutex_lock(&vde->map_lock);
+	}
+
+	mutex_unlock(&vde->map_lock);
+}
diff --git a/drivers/media/platform/nvidia/tegra-vde/h264.c b/drivers/media/platform/nvidia/tegra-vde/h264.c
new file mode 100644
index 000000000..204e474d5
--- /dev/null
+++ b/drivers/media/platform/nvidia/tegra-vde/h264.c
@@ -0,0 +1,950 @@
+// SPDX-License-Identifier: GPL-2.0+
+/*
+ * NVIDIA Tegra Video decoder driver
+ *
+ * Copyright (C) 2016-2022 Dmitry Osipenko <digetx@gmail.com>
+ *
+ */
+
+#include <linux/iopoll.h>
+#include <linux/pm_runtime.h>
+#include <linux/reset.h>
+#include <linux/slab.h>
+
+#include <media/v4l2-h264.h>
+
+#include "trace.h"
+#include "vde.h"
+
+#define FLAG_B_FRAME		0x1
+#define FLAG_REFERENCE		0x2
+
+struct tegra_vde_h264_frame {
+	unsigned int frame_num;
+	unsigned int flags;
+};
+
+struct tegra_vde_h264_decoder_ctx {
+	unsigned int dpb_frames_nb;
+	unsigned int dpb_ref_frames_with_earlier_poc_nb;
+	unsigned int baseline_profile;
+	unsigned int level_idc;
+	unsigned int log2_max_pic_order_cnt_lsb;
+	unsigned int log2_max_frame_num;
+	unsigned int pic_order_cnt_type;
+	unsigned int direct_8x8_inference_flag;
+	unsigned int pic_width_in_mbs;
+	unsigned int pic_height_in_mbs;
+	unsigned int pic_init_qp;
+	unsigned int deblocking_filter_control_present_flag;
+	unsigned int constrained_intra_pred_flag;
+	unsigned int chroma_qp_index_offset;
+	unsigned int pic_order_present_flag;
+	unsigned int num_ref_idx_l0_active_minus1;
+	unsigned int num_ref_idx_l1_active_minus1;
+};
+
+struct h264_reflists {
+	struct v4l2_h264_reference p[V4L2_H264_NUM_DPB_ENTRIES];
+	struct v4l2_h264_reference b0[V4L2_H264_NUM_DPB_ENTRIES];
+	struct v4l2_h264_reference b1[V4L2_H264_NUM_DPB_ENTRIES];
+};
+
+static int tegra_vde_wait_mbe(struct tegra_vde *vde)
+{
+	u32 tmp;
+
+	return readl_relaxed_poll_timeout(vde->mbe + 0x8C, tmp,
+					  tmp >= 0x10, 1, 100);
+}
+
+static int tegra_vde_setup_mbe_frame_idx(struct tegra_vde *vde,
+					 unsigned int refs_nb,
+					 bool setup_refs)
+{
+	u32 value, frame_idx_enb_mask = 0;
+	unsigned int frame_idx;
+	unsigned int idx;
+	int err;
+
+	tegra_vde_writel(vde, 0xD0000000 | (0 << 23), vde->mbe, 0x80);
+	tegra_vde_writel(vde, 0xD0200000 | (0 << 23), vde->mbe, 0x80);
+
+	err = tegra_vde_wait_mbe(vde);
+	if (err)
+		return err;
+
+	if (!setup_refs)
+		return 0;
+
+	for (idx = 0, frame_idx = 1; idx < refs_nb; idx++, frame_idx++) {
+		tegra_vde_writel(vde, 0xD0000000 | (frame_idx << 23),
+				 vde->mbe, 0x80);
+		tegra_vde_writel(vde, 0xD0200000 | (frame_idx << 23),
+				 vde->mbe, 0x80);
+
+		frame_idx_enb_mask |= frame_idx << (6 * (idx % 4));
+
+		if (idx % 4 == 3 || idx == refs_nb - 1) {
+			value = 0xC0000000;
+			value |= (idx >> 2) << 24;
+			value |= frame_idx_enb_mask;
+
+			tegra_vde_writel(vde, value, vde->mbe, 0x80);
+
+			err = tegra_vde_wait_mbe(vde);
+			if (err)
+				return err;
+
+			frame_idx_enb_mask = 0;
+		}
+	}
+
+	return 0;
+}
+
+static void tegra_vde_mbe_set_0xa_reg(struct tegra_vde *vde, int reg, u32 val)
+{
+	tegra_vde_writel(vde, 0xA0000000 | (reg << 24) | (val & 0xFFFF),
+			 vde->mbe, 0x80);
+	tegra_vde_writel(vde, 0xA0000000 | ((reg + 1) << 24) | (val >> 16),
+			 vde->mbe, 0x80);
+}
+
+static int tegra_vde_wait_bsev(struct tegra_vde *vde, bool wait_dma)
+{
+	struct device *dev = vde->dev;
+	u32 value;
+	int err;
+
+	err = readl_relaxed_poll_timeout(vde->bsev + INTR_STATUS, value,
+					 !(value & BIT(2)), 1, 100);
+	if (err) {
+		dev_err(dev, "BSEV unknown bit timeout\n");
+		return err;
+	}
+
+	err = readl_relaxed_poll_timeout(vde->bsev + INTR_STATUS, value,
+					 (value & BSE_ICMDQUE_EMPTY), 1, 100);
+	if (err) {
+		dev_err(dev, "BSEV ICMDQUE flush timeout\n");
+		return err;
+	}
+
+	if (!wait_dma)
+		return 0;
+
+	err = readl_relaxed_poll_timeout(vde->bsev + INTR_STATUS, value,
+					 !(value & BSE_DMA_BUSY), 1, 1000);
+	if (err) {
+		dev_err(dev, "BSEV DMA timeout\n");
+		return err;
+	}
+
+	return 0;
+}
+
+static int tegra_vde_push_to_bsev_icmdqueue(struct tegra_vde *vde,
+					    u32 value, bool wait_dma)
+{
+	tegra_vde_writel(vde, value, vde->bsev, ICMDQUE_WR);
+
+	return tegra_vde_wait_bsev(vde, wait_dma);
+}
+
+static void tegra_vde_setup_frameid(struct tegra_vde *vde,
+				    struct tegra_video_frame *frame,
+				    unsigned int frameid,
+				    u32 mbs_width, u32 mbs_height)
+{
+	u32 y_addr  = frame ? frame->y_addr  : 0x6CDEAD00;
+	u32 cb_addr = frame ? frame->cb_addr : 0x6CDEAD00;
+	u32 cr_addr = frame ? frame->cr_addr : 0x6CDEAD00;
+	u32 value1 = frame ? ((frame->luma_atoms_pitch << 16) | mbs_height) : 0;
+	u32 value2 = frame ? ((frame->chroma_atoms_pitch << 6) | 1) : 0;
+
+	tegra_vde_writel(vde, y_addr  >> 8, vde->frameid, 0x000 + frameid * 4);
+	tegra_vde_writel(vde, cb_addr >> 8, vde->frameid, 0x100 + frameid * 4);
+	tegra_vde_writel(vde, cr_addr >> 8, vde->frameid, 0x180 + frameid * 4);
+	tegra_vde_writel(vde, value1,       vde->frameid, 0x080 + frameid * 4);
+	tegra_vde_writel(vde, value2,       vde->frameid, 0x280 + frameid * 4);
+}
+
+static void tegra_setup_frameidx(struct tegra_vde *vde,
+				 struct tegra_video_frame *frames,
+				 unsigned int frames_nb,
+				 u32 mbs_width, u32 mbs_height)
+{
+	unsigned int idx;
+
+	for (idx = 0; idx < frames_nb; idx++)
+		tegra_vde_setup_frameid(vde, &frames[idx], idx,
+					mbs_width, mbs_height);
+
+	for (; idx < 17; idx++)
+		tegra_vde_setup_frameid(vde, NULL, idx, 0, 0);
+}
+
+static void tegra_vde_setup_iram_entry(struct tegra_vde *vde,
+				       unsigned int table,
+				       unsigned int row,
+				       u32 value1, u32 value2)
+{
+	u32 *iram_tables = vde->iram;
+
+	trace_vde_setup_iram_entry(table, row, value1, value2);
+
+	iram_tables[0x20 * table + row * 2 + 0] = value1;
+	iram_tables[0x20 * table + row * 2 + 1] = value2;
+}
+
+static void tegra_vde_setup_iram_tables(struct tegra_vde *vde,
+					struct tegra_video_frame *dpb_frames,
+					unsigned int ref_frames_nb,
+					unsigned int with_earlier_poc_nb)
+{
+	struct tegra_video_frame *frame;
+	int with_later_poc_nb;
+	u32 value, aux_addr;
+	unsigned int i, k;
+
+	trace_vde_ref_l0(dpb_frames[0].frame_num);
+
+	for (i = 0; i < 16; i++) {
+		if (i < ref_frames_nb) {
+			frame = &dpb_frames[i + 1];
+
+			aux_addr = frame->aux_addr;
+
+			value  = (i + 1) << 26;
+			value |= !(frame->flags & FLAG_B_FRAME) << 25;
+			value |= 1 << 24;
+			value |= frame->frame_num;
+		} else {
+			aux_addr = 0x6ADEAD00;
+			value = 0x3f;
+		}
+
+		tegra_vde_setup_iram_entry(vde, 0, i, value, aux_addr);
+		tegra_vde_setup_iram_entry(vde, 1, i, value, aux_addr);
+		tegra_vde_setup_iram_entry(vde, 2, i, value, aux_addr);
+		tegra_vde_setup_iram_entry(vde, 3, i, value, aux_addr);
+	}
+
+	if (!(dpb_frames[0].flags & FLAG_B_FRAME))
+		return;
+
+	if (with_earlier_poc_nb >= ref_frames_nb)
+		return;
+
+	with_later_poc_nb = ref_frames_nb - with_earlier_poc_nb;
+
+	trace_vde_ref_l1(with_later_poc_nb, with_earlier_poc_nb);
+
+	for (i = 0, k = with_earlier_poc_nb; i < with_later_poc_nb; i++, k++) {
+		frame = &dpb_frames[k + 1];
+
+		aux_addr = frame->aux_addr;
+
+		value  = (k + 1) << 26;
+		value |= !(frame->flags & FLAG_B_FRAME) << 25;
+		value |= 1 << 24;
+		value |= frame->frame_num;
+
+		tegra_vde_setup_iram_entry(vde, 2, i, value, aux_addr);
+	}
+
+	for (k = 0; i < ref_frames_nb; i++, k++) {
+		frame = &dpb_frames[k + 1];
+
+		aux_addr = frame->aux_addr;
+
+		value  = (k + 1) << 26;
+		value |= !(frame->flags & FLAG_B_FRAME) << 25;
+		value |= 1 << 24;
+		value |= frame->frame_num;
+
+		tegra_vde_setup_iram_entry(vde, 2, i, value, aux_addr);
+	}
+}
+
+static int tegra_vde_setup_hw_context(struct tegra_vde *vde,
+				      struct tegra_vde_h264_decoder_ctx *ctx,
+				      struct tegra_video_frame *dpb_frames,
+				      dma_addr_t bitstream_data_addr,
+				      size_t bitstream_data_size,
+				      unsigned int macroblocks_nb)
+{
+	struct device *dev = vde->dev;
+	u32 value;
+	int err;
+
+	tegra_vde_set_bits(vde, 0x000A, vde->sxe, 0xF0);
+	tegra_vde_set_bits(vde, 0x000B, vde->bsev, CMDQUE_CONTROL);
+	tegra_vde_set_bits(vde, 0x8002, vde->mbe, 0x50);
+	tegra_vde_set_bits(vde, 0x000A, vde->mbe, 0xA0);
+	tegra_vde_set_bits(vde, 0x000A, vde->ppe, 0x14);
+	tegra_vde_set_bits(vde, 0x000A, vde->ppe, 0x28);
+	tegra_vde_set_bits(vde, 0x0A00, vde->mce, 0x08);
+	tegra_vde_set_bits(vde, 0x000A, vde->tfe, 0x00);
+	tegra_vde_set_bits(vde, 0x0005, vde->vdma, 0x04);
+
+	tegra_vde_writel(vde, 0x00000000, vde->vdma, 0x1C);
+	tegra_vde_writel(vde, 0x00000000, vde->vdma, 0x00);
+	tegra_vde_writel(vde, 0x00000007, vde->vdma, 0x04);
+	tegra_vde_writel(vde, 0x00000007, vde->frameid, 0x200);
+	tegra_vde_writel(vde, 0x00000005, vde->tfe, 0x04);
+	tegra_vde_writel(vde, 0x00000000, vde->mbe, 0x84);
+	tegra_vde_writel(vde, 0x00000010, vde->sxe, 0x08);
+	tegra_vde_writel(vde, 0x00000150, vde->sxe, 0x54);
+	tegra_vde_writel(vde, 0x0000054C, vde->sxe, 0x58);
+	tegra_vde_writel(vde, 0x00000E34, vde->sxe, 0x5C);
+	tegra_vde_writel(vde, 0x063C063C, vde->mce, 0x10);
+	tegra_vde_writel(vde, 0x0003FC00, vde->bsev, INTR_STATUS);
+	tegra_vde_writel(vde, 0x0000150D, vde->bsev, BSE_CONFIG);
+	tegra_vde_writel(vde, 0x00000100, vde->bsev, BSE_INT_ENB);
+	tegra_vde_writel(vde, 0x00000000, vde->bsev, 0x98);
+	tegra_vde_writel(vde, 0x00000060, vde->bsev, 0x9C);
+
+	memset(vde->iram + 128, 0, macroblocks_nb / 2);
+
+	tegra_setup_frameidx(vde, dpb_frames, ctx->dpb_frames_nb,
+			     ctx->pic_width_in_mbs, ctx->pic_height_in_mbs);
+
+	tegra_vde_setup_iram_tables(vde, dpb_frames,
+				    ctx->dpb_frames_nb - 1,
+				    ctx->dpb_ref_frames_with_earlier_poc_nb);
+
+	/*
+	 * The IRAM mapping is write-combine, ensure that CPU buffers have
+	 * been flushed at this point.
+	 */
+	wmb();
+
+	tegra_vde_writel(vde, 0x00000000, vde->bsev, 0x8C);
+	tegra_vde_writel(vde, bitstream_data_addr + bitstream_data_size,
+			 vde->bsev, 0x54);
+
+	vde->bitstream_data_addr = bitstream_data_addr;
+
+	value = ctx->pic_width_in_mbs << 11 | ctx->pic_height_in_mbs << 3;
+
+	tegra_vde_writel(vde, value, vde->bsev, 0x88);
+
+	err = tegra_vde_wait_bsev(vde, false);
+	if (err)
+		return err;
+
+	err = tegra_vde_push_to_bsev_icmdqueue(vde, 0x800003FC, false);
+	if (err)
+		return err;
+
+	value = 0x01500000;
+	value |= ((vde->iram_lists_addr + 512) >> 2) & 0xFFFF;
+
+	err = tegra_vde_push_to_bsev_icmdqueue(vde, value, true);
+	if (err)
+		return err;
+
+	err = tegra_vde_push_to_bsev_icmdqueue(vde, 0x840F054C, false);
+	if (err)
+		return err;
+
+	err = tegra_vde_push_to_bsev_icmdqueue(vde, 0x80000080, false);
+	if (err)
+		return err;
+
+	value = 0x0E340000 | ((vde->iram_lists_addr >> 2) & 0xFFFF);
+
+	err = tegra_vde_push_to_bsev_icmdqueue(vde, value, true);
+	if (err)
+		return err;
+
+	value = 0x00800005;
+	value |= ctx->pic_width_in_mbs << 11;
+	value |= ctx->pic_height_in_mbs << 3;
+
+	tegra_vde_writel(vde, value, vde->sxe, 0x10);
+
+	value = !ctx->baseline_profile << 17;
+	value |= ctx->level_idc << 13;
+	value |= ctx->log2_max_pic_order_cnt_lsb << 7;
+	value |= ctx->pic_order_cnt_type << 5;
+	value |= ctx->log2_max_frame_num;
+
+	tegra_vde_writel(vde, value, vde->sxe, 0x40);
+
+	value = ctx->pic_init_qp << 25;
+	value |= !!(ctx->deblocking_filter_control_present_flag) << 2;
+	value |= !!ctx->pic_order_present_flag;
+
+	tegra_vde_writel(vde, value, vde->sxe, 0x44);
+
+	value = ctx->chroma_qp_index_offset;
+	value |= ctx->num_ref_idx_l0_active_minus1 << 5;
+	value |= ctx->num_ref_idx_l1_active_minus1 << 10;
+	value |= !!ctx->constrained_intra_pred_flag << 15;
+
+	tegra_vde_writel(vde, value, vde->sxe, 0x48);
+
+	value = 0x0C000000;
+	value |= !!(dpb_frames[0].flags & FLAG_B_FRAME) << 24;
+
+	tegra_vde_writel(vde, value, vde->sxe, 0x4C);
+
+	value = 0x03800000;
+	value |= bitstream_data_size & GENMASK(19, 15);
+
+	tegra_vde_writel(vde, value, vde->sxe, 0x68);
+
+	tegra_vde_writel(vde, bitstream_data_addr, vde->sxe, 0x6C);
+
+	if (vde->soc->supports_ref_pic_marking)
+		tegra_vde_writel(vde, vde->secure_bo->dma_addr, vde->sxe, 0x7c);
+
+	value = 0x10000005;
+	value |= ctx->pic_width_in_mbs << 11;
+	value |= ctx->pic_height_in_mbs << 3;
+
+	tegra_vde_writel(vde, value, vde->mbe, 0x80);
+
+	value = 0x26800000;
+	value |= ctx->level_idc << 4;
+	value |= !ctx->baseline_profile << 1;
+	value |= !!ctx->direct_8x8_inference_flag;
+
+	tegra_vde_writel(vde, value, vde->mbe, 0x80);
+
+	tegra_vde_writel(vde, 0xF4000001, vde->mbe, 0x80);
+	tegra_vde_writel(vde, 0x20000000, vde->mbe, 0x80);
+	tegra_vde_writel(vde, 0xF4000101, vde->mbe, 0x80);
+
+	value = 0x20000000;
+	value |= ctx->chroma_qp_index_offset << 8;
+
+	tegra_vde_writel(vde, value, vde->mbe, 0x80);
+
+	err = tegra_vde_setup_mbe_frame_idx(vde,
+					    ctx->dpb_frames_nb - 1,
+					    ctx->pic_order_cnt_type == 0);
+	if (err) {
+		dev_err(dev, "MBE frames setup failed %d\n", err);
+		return err;
+	}
+
+	tegra_vde_mbe_set_0xa_reg(vde, 0, 0x000009FC);
+	tegra_vde_mbe_set_0xa_reg(vde, 2, 0x61DEAD00);
+	tegra_vde_mbe_set_0xa_reg(vde, 4, 0x62DEAD00);
+	tegra_vde_mbe_set_0xa_reg(vde, 6, 0x63DEAD00);
+	tegra_vde_mbe_set_0xa_reg(vde, 8, dpb_frames[0].aux_addr);
+
+	value = 0xFC000000;
+	value |= !!(dpb_frames[0].flags & FLAG_B_FRAME) << 2;
+
+	if (!ctx->baseline_profile)
+		value |= !!(dpb_frames[0].flags & FLAG_REFERENCE) << 1;
+
+	tegra_vde_writel(vde, value, vde->mbe, 0x80);
+
+	err = tegra_vde_wait_mbe(vde);
+	if (err) {
+		dev_err(dev, "MBE programming failed %d\n", err);
+		return err;
+	}
+
+	return 0;
+}
+
+static void tegra_vde_decode_frame(struct tegra_vde *vde,
+				   unsigned int macroblocks_nb)
+{
+	reinit_completion(&vde->decode_completion);
+
+	tegra_vde_writel(vde, 0x00000001, vde->bsev, 0x8C);
+	tegra_vde_writel(vde, 0x20000000 | (macroblocks_nb - 1),
+			 vde->sxe, 0x00);
+}
+
+static int tegra_vde_validate_h264_ctx(struct device *dev,
+				       struct tegra_vde_h264_decoder_ctx *ctx)
+{
+	if (ctx->dpb_frames_nb == 0 || ctx->dpb_frames_nb > 17) {
+		dev_err(dev, "Bad DPB size %u\n", ctx->dpb_frames_nb);
+		return -EINVAL;
+	}
+
+	if (ctx->level_idc > 15) {
+		dev_err(dev, "Bad level value %u\n", ctx->level_idc);
+		return -EINVAL;
+	}
+
+	if (ctx->pic_init_qp > 52) {
+		dev_err(dev, "Bad pic_init_qp value %u\n", ctx->pic_init_qp);
+		return -EINVAL;
+	}
+
+	if (ctx->log2_max_pic_order_cnt_lsb > 16) {
+		dev_err(dev, "Bad log2_max_pic_order_cnt_lsb value %u\n",
+			ctx->log2_max_pic_order_cnt_lsb);
+		return -EINVAL;
+	}
+
+	if (ctx->log2_max_frame_num > 16) {
+		dev_err(dev, "Bad log2_max_frame_num value %u\n",
+			ctx->log2_max_frame_num);
+		return -EINVAL;
+	}
+
+	if (ctx->chroma_qp_index_offset > 31) {
+		dev_err(dev, "Bad chroma_qp_index_offset value %u\n",
+			ctx->chroma_qp_index_offset);
+		return -EINVAL;
+	}
+
+	if (ctx->pic_order_cnt_type > 2) {
+		dev_err(dev, "Bad pic_order_cnt_type value %u\n",
+			ctx->pic_order_cnt_type);
+		return -EINVAL;
+	}
+
+	if (ctx->num_ref_idx_l0_active_minus1 > 15) {
+		dev_err(dev, "Bad num_ref_idx_l0_active_minus1 value %u\n",
+			ctx->num_ref_idx_l0_active_minus1);
+		return -EINVAL;
+	}
+
+	if (ctx->num_ref_idx_l1_active_minus1 > 15) {
+		dev_err(dev, "Bad num_ref_idx_l1_active_minus1 value %u\n",
+			ctx->num_ref_idx_l1_active_minus1);
+		return -EINVAL;
+	}
+
+	if (!ctx->pic_width_in_mbs || ctx->pic_width_in_mbs > 127) {
+		dev_err(dev, "Bad pic_width_in_mbs value %u\n",
+			ctx->pic_width_in_mbs);
+		return -EINVAL;
+	}
+
+	if (!ctx->pic_height_in_mbs || ctx->pic_height_in_mbs > 127) {
+		dev_err(dev, "Bad pic_height_in_mbs value %u\n",
+			ctx->pic_height_in_mbs);
+		return -EINVAL;
+	}
+
+	return 0;
+}
+
+static int tegra_vde_decode_begin(struct tegra_vde *vde,
+				  struct tegra_vde_h264_decoder_ctx *ctx,
+				  struct tegra_video_frame *dpb_frames,
+				  dma_addr_t bitstream_data_addr,
+				  size_t bitstream_data_size)
+{
+	struct device *dev = vde->dev;
+	unsigned int macroblocks_nb;
+	int err;
+
+	err = mutex_lock_interruptible(&vde->lock);
+	if (err)
+		return err;
+
+	err = pm_runtime_resume_and_get(dev);
+	if (err < 0)
+		goto unlock;
+
+	/*
+	 * We rely on the VDE registers reset value, otherwise VDE
+	 * causes bus lockup.
+	 */
+	err = reset_control_assert(vde->rst_mc);
+	if (err) {
+		dev_err(dev, "DEC start: Failed to assert MC reset: %d\n",
+			err);
+		goto put_runtime_pm;
+	}
+
+	err = reset_control_reset(vde->rst);
+	if (err) {
+		dev_err(dev, "DEC start: Failed to reset HW: %d\n", err);
+		goto put_runtime_pm;
+	}
+
+	err = reset_control_deassert(vde->rst_mc);
+	if (err) {
+		dev_err(dev, "DEC start: Failed to deassert MC reset: %d\n",
+			err);
+		goto put_runtime_pm;
+	}
+
+	macroblocks_nb = ctx->pic_width_in_mbs * ctx->pic_height_in_mbs;
+
+	err = tegra_vde_setup_hw_context(vde, ctx, dpb_frames,
+					 bitstream_data_addr,
+					 bitstream_data_size,
+					 macroblocks_nb);
+	if (err)
+		goto put_runtime_pm;
+
+	tegra_vde_decode_frame(vde, macroblocks_nb);
+
+	return 0;
+
+put_runtime_pm:
+	pm_runtime_mark_last_busy(dev);
+	pm_runtime_put_autosuspend(dev);
+
+unlock:
+	mutex_unlock(&vde->lock);
+
+	return err;
+}
+
+static void tegra_vde_decode_abort(struct tegra_vde *vde)
+{
+	struct device *dev = vde->dev;
+	int err;
+
+	/*
+	 * At first reset memory client to avoid resetting VDE HW in the
+	 * middle of DMA which could result into memory corruption or hang
+	 * the whole system.
+	 */
+	err = reset_control_assert(vde->rst_mc);
+	if (err)
+		dev_err(dev, "DEC end: Failed to assert MC reset: %d\n", err);
+
+	err = reset_control_assert(vde->rst);
+	if (err)
+		dev_err(dev, "DEC end: Failed to assert HW reset: %d\n", err);
+
+	pm_runtime_mark_last_busy(dev);
+	pm_runtime_put_autosuspend(dev);
+
+	mutex_unlock(&vde->lock);
+}
+
+static int tegra_vde_decode_end(struct tegra_vde *vde)
+{
+	unsigned int read_bytes, macroblocks_nb;
+	struct device *dev = vde->dev;
+	dma_addr_t bsev_ptr;
+	long timeout;
+	int ret;
+
+	timeout = wait_for_completion_interruptible_timeout(
+			&vde->decode_completion, msecs_to_jiffies(1000));
+	if (timeout == 0) {
+		bsev_ptr = tegra_vde_readl(vde, vde->bsev, 0x10);
+		macroblocks_nb = tegra_vde_readl(vde, vde->sxe, 0xC8) & 0x1FFF;
+		read_bytes = bsev_ptr ? bsev_ptr - vde->bitstream_data_addr : 0;
+
+		dev_err(dev, "Decoding failed: read 0x%X bytes, %u macroblocks parsed\n",
+			read_bytes, macroblocks_nb);
+
+		ret = -EIO;
+	} else if (timeout < 0) {
+		ret = timeout;
+	} else {
+		ret = 0;
+	}
+
+	tegra_vde_decode_abort(vde);
+
+	return ret;
+}
+
+static struct vb2_buffer *get_ref_buf(struct tegra_ctx *ctx,
+				      struct vb2_v4l2_buffer *dst,
+				      unsigned int dpb_idx)
+{
+	const struct v4l2_h264_dpb_entry *dpb = ctx->h264.decode_params->dpb;
+	struct vb2_queue *cap_q = &ctx->fh.m2m_ctx->cap_q_ctx.q;
+	struct vb2_buffer *vb = NULL;
+
+	if (dpb[dpb_idx].flags & V4L2_H264_DPB_ENTRY_FLAG_ACTIVE)
+		vb = vb2_find_buffer(cap_q, dpb[dpb_idx].reference_ts);
+
+	/*
+	 * If a DPB entry is unused or invalid, address of current destination
+	 * buffer is returned.
+	 */
+	if (!vb)
+		return &dst->vb2_buf;
+
+	return vb;
+}
+
+static int tegra_vde_validate_vb_size(struct tegra_ctx *ctx,
+				      struct vb2_buffer *vb,
+				      unsigned int plane_id,
+				      size_t min_size)
+{
+	u64 offset = vb->planes[plane_id].data_offset;
+	struct device *dev = ctx->vde->dev;
+
+	if (offset + min_size > vb2_plane_size(vb, plane_id)) {
+		dev_err(dev, "Too small plane[%u] size %lu @0x%llX, should be at least %zu\n",
+			plane_id, vb2_plane_size(vb, plane_id), offset, min_size);
+		return -EINVAL;
+	}
+
+	return 0;
+}
+
+static int tegra_vde_h264_setup_frame(struct tegra_ctx *ctx,
+				      struct tegra_vde_h264_decoder_ctx *h264,
+				      struct v4l2_h264_reflist_builder *b,
+				      struct vb2_buffer *vb,
+				      unsigned int ref_id,
+				      unsigned int id)
+{
+	struct v4l2_pix_format_mplane *pixfmt = &ctx->decoded_fmt.fmt.pix_mp;
+	struct tegra_m2m_buffer *tb = vb_to_tegra_buf(vb);
+	struct tegra_ctx_h264 *h = &ctx->h264;
+	struct tegra_vde *vde = ctx->vde;
+	struct device *dev = vde->dev;
+	unsigned int cstride, lstride;
+	unsigned int flags = 0;
+	size_t lsize, csize;
+	int err, frame_num;
+
+	lsize = h264->pic_width_in_mbs * 16 * h264->pic_height_in_mbs * 16;
+	csize = h264->pic_width_in_mbs *  8 * h264->pic_height_in_mbs *  8;
+	lstride = pixfmt->plane_fmt[0].bytesperline;
+	cstride = pixfmt->plane_fmt[1].bytesperline;
+
+	err = tegra_vde_validate_vb_size(ctx, vb, 0, lsize);
+	if (err)
+		return err;
+
+	err = tegra_vde_validate_vb_size(ctx, vb, 1, csize);
+	if (err)
+		return err;
+
+	err = tegra_vde_validate_vb_size(ctx, vb, 2, csize);
+	if (err)
+		return err;
+
+	if (!tb->aux || tb->aux->size < csize) {
+		dev_err(dev, "Too small aux size %zd, should be at least %zu\n",
+			tb->aux ? tb->aux->size : -1, csize);
+		return -EINVAL;
+	}
+
+	if (id == 0) {
+		frame_num = h->decode_params->frame_num;
+
+		if (h->decode_params->nal_ref_idc)
+			flags |= FLAG_REFERENCE;
+	} else {
+		frame_num = b->refs[ref_id].frame_num;
+	}
+
+	if (tb->b_frame)
+		flags |= FLAG_B_FRAME;
+
+	vde->frames[id].flags = flags;
+	vde->frames[id].y_addr = tb->dma_addr[0];
+	vde->frames[id].cb_addr = tb->dma_addr[1];
+	vde->frames[id].cr_addr = tb->dma_addr[2];
+	vde->frames[id].aux_addr = tb->aux->dma_addr;
+	vde->frames[id].frame_num = frame_num & 0x7fffff;
+	vde->frames[id].luma_atoms_pitch = lstride / VDE_ATOM;
+	vde->frames[id].chroma_atoms_pitch = cstride / VDE_ATOM;
+
+	return 0;
+}
+
+static int tegra_vde_h264_setup_frames(struct tegra_ctx *ctx,
+				       struct tegra_vde_h264_decoder_ctx *h264)
+{
+	struct vb2_v4l2_buffer *src = v4l2_m2m_next_src_buf(ctx->fh.m2m_ctx);
+	struct vb2_v4l2_buffer *dst = v4l2_m2m_next_dst_buf(ctx->fh.m2m_ctx);
+	const struct v4l2_h264_dpb_entry *dpb = ctx->h264.decode_params->dpb;
+	struct tegra_m2m_buffer *tb = vb_to_tegra_buf(&dst->vb2_buf);
+	struct tegra_ctx_h264 *h = &ctx->h264;
+	struct v4l2_h264_reflist_builder b;
+	struct v4l2_h264_reference *dpb_id;
+	struct h264_reflists reflists;
+	struct vb2_buffer *ref;
+	unsigned int i;
+	int err;
+
+	/*
+	 * Tegra hardware requires information about frame's type, assuming
+	 * that frame consists of the same type slices. Userspace must tag
+	 * frame's type appropriately.
+	 *
+	 * Decoding of a non-uniform frames isn't supported by hardware and
+	 * require software preprocessing that we don't implement. Decoding
+	 * is expected to fail in this case. Such video streams are rare in
+	 * practice, so not a big deal.
+	 *
+	 * If userspace doesn't tell us frame's type, then we will try decode
+	 * as-is.
+	 */
+	v4l2_m2m_buf_copy_metadata(src, dst, true);
+
+	if (h->decode_params->flags & V4L2_H264_DECODE_PARAM_FLAG_BFRAME)
+		tb->b_frame = true;
+	else
+		tb->b_frame = false;
+
+	err = tegra_vde_h264_setup_frame(ctx, h264, NULL, &dst->vb2_buf, 0,
+					 h264->dpb_frames_nb++);
+	if (err)
+		return err;
+
+	if (!(h->decode_params->flags & (V4L2_H264_DECODE_PARAM_FLAG_PFRAME |
+					 V4L2_H264_DECODE_PARAM_FLAG_BFRAME)))
+		return 0;
+
+	v4l2_h264_init_reflist_builder(&b, h->decode_params, h->sps, dpb);
+
+	if (h->decode_params->flags & V4L2_H264_DECODE_PARAM_FLAG_BFRAME) {
+		v4l2_h264_build_b_ref_lists(&b, reflists.b0, reflists.b1);
+		dpb_id = reflists.b0;
+	} else {
+		v4l2_h264_build_p_ref_list(&b, reflists.p);
+		dpb_id = reflists.p;
+	}
+
+	for (i = 0; i < b.num_valid; i++) {
+		int dpb_idx = dpb_id[i].index;
+
+		ref = get_ref_buf(ctx, dst, dpb_idx);
+
+		err = tegra_vde_h264_setup_frame(ctx, h264, &b, ref, dpb_idx,
+						 h264->dpb_frames_nb++);
+		if (err)
+			return err;
+
+		if (b.refs[dpb_idx].top_field_order_cnt < b.cur_pic_order_count)
+			h264->dpb_ref_frames_with_earlier_poc_nb++;
+	}
+
+	return 0;
+}
+
+static unsigned int to_tegra_vde_h264_level_idc(unsigned int level_idc)
+{
+	switch (level_idc) {
+	case 11:
+		return 2;
+	case 12:
+		return 3;
+	case 13:
+		return 4;
+	case 20:
+		return 5;
+	case 21:
+		return 6;
+	case 22:
+		return 7;
+	case 30:
+		return 8;
+	case 31:
+		return 9;
+	case 32:
+		return 10;
+	case 40:
+		return 11;
+	case 41:
+		return 12;
+	case 42:
+		return 13;
+	case 50:
+		return 14;
+	default:
+		break;
+	}
+
+	return 15;
+}
+
+static int tegra_vde_h264_setup_context(struct tegra_ctx *ctx,
+					struct tegra_vde_h264_decoder_ctx *h264)
+{
+	struct tegra_ctx_h264 *h = &ctx->h264;
+	struct tegra_vde *vde = ctx->vde;
+	struct device *dev = vde->dev;
+	int err;
+
+	memset(h264, 0, sizeof(*h264));
+	memset(vde->frames, 0, sizeof(vde->frames));
+
+	tegra_vde_prepare_control_data(ctx, V4L2_CID_STATELESS_H264_DECODE_PARAMS);
+	tegra_vde_prepare_control_data(ctx, V4L2_CID_STATELESS_H264_SPS);
+	tegra_vde_prepare_control_data(ctx, V4L2_CID_STATELESS_H264_PPS);
+
+	/* CABAC unsupported by hardware, requires software preprocessing */
+	if (h->pps->flags & V4L2_H264_PPS_FLAG_ENTROPY_CODING_MODE)
+		return -EOPNOTSUPP;
+
+	if (h->decode_params->flags & V4L2_H264_DECODE_PARAM_FLAG_FIELD_PIC)
+		return -EOPNOTSUPP;
+
+	if (h->sps->profile_idc == 66)
+		h264->baseline_profile = 1;
+
+	if (h->sps->flags & V4L2_H264_SPS_FLAG_DIRECT_8X8_INFERENCE)
+		h264->direct_8x8_inference_flag = 1;
+
+	if (h->pps->flags & V4L2_H264_PPS_FLAG_CONSTRAINED_INTRA_PRED)
+		h264->constrained_intra_pred_flag = 1;
+
+	if (h->pps->flags & V4L2_H264_PPS_FLAG_DEBLOCKING_FILTER_CONTROL_PRESENT)
+		h264->deblocking_filter_control_present_flag = 1;
+
+	if (h->pps->flags & V4L2_H264_PPS_FLAG_BOTTOM_FIELD_PIC_ORDER_IN_FRAME_PRESENT)
+		h264->pic_order_present_flag = 1;
+
+	h264->level_idc				= to_tegra_vde_h264_level_idc(h->sps->level_idc);
+	h264->log2_max_pic_order_cnt_lsb	= h->sps->log2_max_pic_order_cnt_lsb_minus4 + 4;
+	h264->log2_max_frame_num		= h->sps->log2_max_frame_num_minus4 + 4;
+	h264->pic_order_cnt_type		= h->sps->pic_order_cnt_type;
+	h264->pic_width_in_mbs			= h->sps->pic_width_in_mbs_minus1 + 1;
+	h264->pic_height_in_mbs			= h->sps->pic_height_in_map_units_minus1 + 1;
+
+	h264->num_ref_idx_l0_active_minus1	= h->pps->num_ref_idx_l0_default_active_minus1;
+	h264->num_ref_idx_l1_active_minus1	= h->pps->num_ref_idx_l1_default_active_minus1;
+	h264->chroma_qp_index_offset		= h->pps->chroma_qp_index_offset & 0x1f;
+	h264->pic_init_qp			= h->pps->pic_init_qp_minus26 + 26;
+
+	err = tegra_vde_h264_setup_frames(ctx, h264);
+	if (err)
+		return err;
+
+	err = tegra_vde_validate_h264_ctx(dev, h264);
+	if (err)
+		return err;
+
+	return 0;
+}
+
+int tegra_vde_h264_decode_run(struct tegra_ctx *ctx)
+{
+	struct vb2_v4l2_buffer *src = v4l2_m2m_next_src_buf(ctx->fh.m2m_ctx);
+	struct tegra_m2m_buffer *bitstream = vb_to_tegra_buf(&src->vb2_buf);
+	size_t bitstream_size = vb2_get_plane_payload(&src->vb2_buf, 0);
+	struct tegra_vde_h264_decoder_ctx h264;
+	struct tegra_vde *vde = ctx->vde;
+	int err;
+
+	err = tegra_vde_h264_setup_context(ctx, &h264);
+	if (err)
+		return err;
+
+	err = tegra_vde_decode_begin(vde, &h264, vde->frames,
+				     bitstream->dma_addr[0],
+				     bitstream_size);
+	if (err)
+		return err;
+
+	return 0;
+}
+
+int tegra_vde_h264_decode_wait(struct tegra_ctx *ctx)
+{
+	return tegra_vde_decode_end(ctx->vde);
+}
diff --git a/drivers/media/platform/nvidia/tegra-vde/iommu.c b/drivers/media/platform/nvidia/tegra-vde/iommu.c
new file mode 100644
index 000000000..5521ed3e4
--- /dev/null
+++ b/drivers/media/platform/nvidia/tegra-vde/iommu.c
@@ -0,0 +1,157 @@
+// SPDX-License-Identifier: GPL-2.0+
+/*
+ * NVIDIA Tegra Video decoder driver
+ *
+ * Copyright (C) 2016-2019 GRATE-DRIVER project
+ */
+
+#include <linux/iommu.h>
+#include <linux/iova.h>
+#include <linux/kernel.h>
+#include <linux/platform_device.h>
+
+#if IS_ENABLED(CONFIG_ARM_DMA_USE_IOMMU)
+#include <asm/dma-iommu.h>
+#endif
+
+#include "vde.h"
+
+int tegra_vde_iommu_map(struct tegra_vde *vde,
+			struct sg_table *sgt,
+			struct iova **iovap,
+			size_t size)
+{
+	struct iova *iova;
+	unsigned long shift;
+	unsigned long end;
+	dma_addr_t addr;
+
+	end = vde->domain->geometry.aperture_end;
+	size = iova_align(&vde->iova, size);
+	shift = iova_shift(&vde->iova);
+
+	iova = alloc_iova(&vde->iova, size >> shift, end >> shift, true);
+	if (!iova)
+		return -ENOMEM;
+
+	addr = iova_dma_addr(&vde->iova, iova);
+
+	size = iommu_map_sgtable(vde->domain, addr, sgt,
+				 IOMMU_READ | IOMMU_WRITE);
+	if (!size) {
+		__free_iova(&vde->iova, iova);
+		return -ENXIO;
+	}
+
+	*iovap = iova;
+
+	return 0;
+}
+
+void tegra_vde_iommu_unmap(struct tegra_vde *vde, struct iova *iova)
+{
+	unsigned long shift = iova_shift(&vde->iova);
+	unsigned long size = iova_size(iova) << shift;
+	dma_addr_t addr = iova_dma_addr(&vde->iova, iova);
+
+	iommu_unmap(vde->domain, addr, size);
+	__free_iova(&vde->iova, iova);
+}
+
+int tegra_vde_iommu_init(struct tegra_vde *vde)
+{
+	struct device *dev = vde->dev;
+	struct iova *iova;
+	unsigned long order;
+	unsigned long shift;
+	int err;
+
+	vde->group = iommu_group_get(dev);
+	if (!vde->group)
+		return 0;
+
+#if IS_ENABLED(CONFIG_ARM_DMA_USE_IOMMU)
+	if (dev->archdata.mapping) {
+		struct dma_iommu_mapping *mapping = to_dma_iommu_mapping(dev);
+
+		arm_iommu_detach_device(dev);
+		arm_iommu_release_mapping(mapping);
+	}
+#endif
+	vde->domain = iommu_domain_alloc(&platform_bus_type);
+	if (!vde->domain) {
+		err = -ENOMEM;
+		goto put_group;
+	}
+
+	err = iova_cache_get();
+	if (err)
+		goto free_domain;
+
+	order = __ffs(vde->domain->pgsize_bitmap);
+	init_iova_domain(&vde->iova, 1UL << order, 0);
+
+	err = iommu_attach_group(vde->domain, vde->group);
+	if (err)
+		goto put_iova;
+
+	/*
+	 * We're using some static addresses that are not accessible by VDE
+	 * to trap invalid memory accesses.
+	 */
+	shift = iova_shift(&vde->iova);
+	iova = reserve_iova(&vde->iova, 0x60000000 >> shift,
+			    0x70000000 >> shift);
+	if (!iova) {
+		err = -ENOMEM;
+		goto detach_group;
+	}
+
+	vde->iova_resv_static_addresses = iova;
+
+	/*
+	 * BSEV's end-address wraps around due to integer overflow during
+	 * of hardware context preparation if IOVA is allocated at the end
+	 * of address space and VDE can't handle that. Hence simply reserve
+	 * the last page to avoid the problem.
+	 */
+	iova = reserve_iova(&vde->iova, 0xffffffff >> shift,
+			    (0xffffffff >> shift) + 1);
+	if (!iova) {
+		err = -ENOMEM;
+		goto unreserve_iova;
+	}
+
+	vde->iova_resv_last_page = iova;
+
+	return 0;
+
+unreserve_iova:
+	__free_iova(&vde->iova, vde->iova_resv_static_addresses);
+detach_group:
+	iommu_detach_group(vde->domain, vde->group);
+put_iova:
+	put_iova_domain(&vde->iova);
+	iova_cache_put();
+free_domain:
+	iommu_domain_free(vde->domain);
+put_group:
+	iommu_group_put(vde->group);
+
+	return err;
+}
+
+void tegra_vde_iommu_deinit(struct tegra_vde *vde)
+{
+	if (vde->domain) {
+		__free_iova(&vde->iova, vde->iova_resv_last_page);
+		__free_iova(&vde->iova, vde->iova_resv_static_addresses);
+		iommu_detach_group(vde->domain, vde->group);
+		put_iova_domain(&vde->iova);
+		iova_cache_put();
+		iommu_domain_free(vde->domain);
+		iommu_group_put(vde->group);
+
+		vde->domain = NULL;
+	}
+}
diff --git a/drivers/media/platform/nvidia/tegra-vde/trace.h b/drivers/media/platform/nvidia/tegra-vde/trace.h
new file mode 100644
index 000000000..7853ab095
--- /dev/null
+++ b/drivers/media/platform/nvidia/tegra-vde/trace.h
@@ -0,0 +1,95 @@
+/* SPDX-License-Identifier: GPL-2.0 */
+
+#undef TRACE_SYSTEM
+#define TRACE_SYSTEM tegra_vde
+
+#if !defined(TEGRA_VDE_TRACE_H) || defined(TRACE_HEADER_MULTI_READ)
+#define TEGRA_VDE_TRACE_H
+
+#include <linux/tracepoint.h>
+
+#include "vde.h"
+
+DECLARE_EVENT_CLASS(register_access,
+	TP_PROTO(struct tegra_vde *vde, void __iomem *base,
+		 u32 offset, u32 value),
+	TP_ARGS(vde, base, offset, value),
+	TP_STRUCT__entry(
+		__string(hw_name, tegra_vde_reg_base_name(vde, base))
+		__field(u32, offset)
+		__field(u32, value)
+	),
+	TP_fast_assign(
+		__assign_str(hw_name, tegra_vde_reg_base_name(vde, base));
+		__entry->offset = offset;
+		__entry->value = value;
+	),
+	TP_printk("%s:0x%03x 0x%08x", __get_str(hw_name), __entry->offset,
+		  __entry->value)
+);
+
+DEFINE_EVENT(register_access, vde_writel,
+	TP_PROTO(struct tegra_vde *vde, void __iomem *base,
+		 u32 offset, u32 value),
+	TP_ARGS(vde, base, offset, value));
+DEFINE_EVENT(register_access, vde_readl,
+	TP_PROTO(struct tegra_vde *vde, void __iomem *base,
+		 u32 offset, u32 value),
+	TP_ARGS(vde, base, offset, value));
+
+TRACE_EVENT(vde_setup_iram_entry,
+	TP_PROTO(unsigned int table, unsigned int row, u32 value, u32 aux_addr),
+	TP_ARGS(table, row, value, aux_addr),
+	TP_STRUCT__entry(
+		__field(unsigned int, table)
+		__field(unsigned int, row)
+		__field(u32, value)
+		__field(u32, aux_addr)
+	),
+	TP_fast_assign(
+		__entry->table = table;
+		__entry->row = row;
+		__entry->value = value;
+		__entry->aux_addr = aux_addr;
+	),
+	TP_printk("[%u][%u] = { 0x%08x (flags = \"%s\", frame_num = %u); 0x%08x }",
+		  __entry->table, __entry->row, __entry->value,
+		  __print_flags(__entry->value, " ", { (1 << 25), "B" }),
+		  __entry->value & 0x7FFFFF, __entry->aux_addr)
+);
+
+TRACE_EVENT(vde_ref_l0,
+	TP_PROTO(unsigned int frame_num),
+	TP_ARGS(frame_num),
+	TP_STRUCT__entry(
+		__field(unsigned int, frame_num)
+	),
+	TP_fast_assign(
+		__entry->frame_num = frame_num;
+	),
+	TP_printk("REF L0: DPB: Frame 0: frame_num = %u", __entry->frame_num)
+);
+
+TRACE_EVENT(vde_ref_l1,
+	TP_PROTO(unsigned int with_later_poc_nb,
+		 unsigned int with_earlier_poc_nb),
+	TP_ARGS(with_later_poc_nb, with_earlier_poc_nb),
+	TP_STRUCT__entry(
+		__field(unsigned int, with_later_poc_nb)
+		__field(unsigned int, with_earlier_poc_nb)
+	),
+	TP_fast_assign(
+		__entry->with_later_poc_nb = with_later_poc_nb;
+		__entry->with_earlier_poc_nb = with_earlier_poc_nb;
+	),
+	TP_printk("REF L1: with_later_poc_nb %u, with_earlier_poc_nb %u",
+		  __entry->with_later_poc_nb, __entry->with_earlier_poc_nb)
+);
+
+#endif /* TEGRA_VDE_TRACE_H */
+
+/* This part must be outside protection */
+#undef TRACE_INCLUDE_PATH
+#define TRACE_INCLUDE_PATH ../../drivers/media/platform/nvidia/tegra-vde
+#define TRACE_INCLUDE_FILE trace
+#include <trace/define_trace.h>
diff --git a/drivers/media/platform/nvidia/tegra-vde/v4l2.c b/drivers/media/platform/nvidia/tegra-vde/v4l2.c
new file mode 100644
index 000000000..bd8c207d5
--- /dev/null
+++ b/drivers/media/platform/nvidia/tegra-vde/v4l2.c
@@ -0,0 +1,1018 @@
+// SPDX-License-Identifier: GPL-2.0+
+/*
+ * NVIDIA Tegra Video decoder driver
+ *
+ * Copyright (C) 2019-2022 Dmitry Osipenko <digetx@gmail.com>
+ *
+ * Based on Cedrus driver by Bootlin.
+ * Copyright (C) 2016 Florent Revest <florent.revest@free-electrons.com>
+ * Copyright (C) 2018 Paul Kocialkowski <paul.kocialkowski@bootlin.com>
+ *
+ * Based on Rockchip driver by Collabora.
+ * Copyright (C) 2019 Boris Brezillon <boris.brezillon@collabora.com>
+ */
+
+#include <linux/err.h>
+#include <linux/slab.h>
+
+#include "vde.h"
+
+static const struct v4l2_ctrl_config ctrl_cfgs[] = {
+	{	.id = V4L2_CID_STATELESS_H264_DECODE_PARAMS,	},
+	{	.id = V4L2_CID_STATELESS_H264_SPS,		},
+	{	.id = V4L2_CID_STATELESS_H264_PPS,		},
+	{
+		.id = V4L2_CID_STATELESS_H264_DECODE_MODE,
+		.min = V4L2_STATELESS_H264_DECODE_MODE_FRAME_BASED,
+		.max = V4L2_STATELESS_H264_DECODE_MODE_FRAME_BASED,
+		.def = V4L2_STATELESS_H264_DECODE_MODE_FRAME_BASED,
+	},
+	{
+		.id = V4L2_CID_STATELESS_H264_START_CODE,
+		.min = V4L2_STATELESS_H264_START_CODE_ANNEX_B,
+		.max = V4L2_STATELESS_H264_START_CODE_ANNEX_B,
+		.def = V4L2_STATELESS_H264_START_CODE_ANNEX_B,
+	},
+	{
+		.id = V4L2_CID_MPEG_VIDEO_H264_PROFILE,
+		.min = V4L2_MPEG_VIDEO_H264_PROFILE_BASELINE,
+		.max = V4L2_MPEG_VIDEO_H264_PROFILE_MAIN,
+		.def = V4L2_MPEG_VIDEO_H264_PROFILE_MAIN,
+	},
+	{
+		.id = V4L2_CID_MPEG_VIDEO_H264_LEVEL,
+		.min = V4L2_MPEG_VIDEO_H264_LEVEL_1_0,
+		.max = V4L2_MPEG_VIDEO_H264_LEVEL_5_1,
+	},
+};
+
+static inline struct tegra_ctx *fh_to_tegra_ctx(struct v4l2_fh *fh)
+{
+	return container_of(fh, struct tegra_ctx, fh);
+}
+
+static void tegra_set_control_data(struct tegra_ctx *ctx, void *data, u32 id)
+{
+	switch (id) {
+	case V4L2_CID_STATELESS_H264_DECODE_PARAMS:
+		ctx->h264.decode_params = data;
+		break;
+	case V4L2_CID_STATELESS_H264_SPS:
+		ctx->h264.sps = data;
+		break;
+	case V4L2_CID_STATELESS_H264_PPS:
+		ctx->h264.pps = data;
+		break;
+	}
+}
+
+void tegra_vde_prepare_control_data(struct tegra_ctx *ctx, u32 id)
+{
+	unsigned int i;
+
+	for (i = 0; i < ARRAY_SIZE(ctrl_cfgs); i++) {
+		if (ctx->ctrls[i]->id == id) {
+			tegra_set_control_data(ctx, ctx->ctrls[i]->p_cur.p, id);
+			return;
+		}
+	}
+
+	tegra_set_control_data(ctx, NULL, id);
+}
+
+static int tegra_queue_setup(struct vb2_queue *vq,
+			     unsigned int *nbufs,
+			     unsigned int *num_planes,
+			     unsigned int sizes[],
+			     struct device *alloc_devs[])
+{
+	struct tegra_ctx *ctx = vb2_get_drv_priv(vq);
+	struct v4l2_format *f;
+	unsigned int i;
+
+	if (V4L2_TYPE_IS_OUTPUT(vq->type))
+		f = &ctx->coded_fmt;
+	else
+		f = &ctx->decoded_fmt;
+
+	if (*num_planes) {
+		if (*num_planes != f->fmt.pix_mp.num_planes)
+			return -EINVAL;
+
+		for (i = 0; i < f->fmt.pix_mp.num_planes; i++) {
+			if (sizes[i] < f->fmt.pix_mp.plane_fmt[i].sizeimage)
+				return -EINVAL;
+		}
+	} else {
+		*num_planes = f->fmt.pix_mp.num_planes;
+
+		for (i = 0; i < f->fmt.pix_mp.num_planes; i++)
+			sizes[i] = f->fmt.pix_mp.plane_fmt[i].sizeimage;
+	}
+
+	return 0;
+}
+
+static int tegra_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 void __tegra_buf_cleanup(struct vb2_buffer *vb, unsigned int i)
+{
+	struct vb2_queue *vq = vb->vb2_queue;
+	struct tegra_ctx *ctx = vb2_get_drv_priv(vq);
+	struct tegra_m2m_buffer *tb = vb_to_tegra_buf(vb);
+
+	while (i--) {
+		if (tb->a[i]) {
+			tegra_vde_dmabuf_cache_unmap(ctx->vde, tb->a[i], true);
+			tb->a[i] = NULL;
+		}
+
+		if (tb->iova[i]) {
+			tegra_vde_iommu_unmap(ctx->vde, tb->iova[i]);
+			tb->iova[i] = NULL;
+		}
+	}
+
+	if (tb->aux) {
+		tegra_vde_free_bo(tb->aux);
+		tb->aux = NULL;
+	}
+}
+
+static int tegra_buf_init(struct vb2_buffer *vb)
+{
+	struct vb2_queue *vq = vb->vb2_queue;
+	struct tegra_ctx *ctx = vb2_get_drv_priv(vq);
+	struct tegra_m2m_buffer *tb = vb_to_tegra_buf(vb);
+	struct tegra_vde *vde = ctx->vde;
+	enum dma_data_direction dma_dir;
+	struct sg_table *sgt;
+	unsigned int i;
+	int err;
+
+	if (V4L2_TYPE_IS_CAPTURE(vq->type) && vb->num_planes > 1) {
+		/*
+		 * Tegra decoder writes auxiliary data for I/P frames.
+		 * This data is needed for decoding of B frames.
+		 */
+		err = tegra_vde_alloc_bo(vde, &tb->aux, DMA_FROM_DEVICE,
+					 vb2_plane_size(vb, 1));
+		if (err)
+			return err;
+	}
+
+	if (V4L2_TYPE_IS_OUTPUT(vq->type))
+		dma_dir = DMA_TO_DEVICE;
+	else
+		dma_dir = DMA_FROM_DEVICE;
+
+	for (i = 0; i < vb->num_planes; i++) {
+		if (vq->memory == VB2_MEMORY_DMABUF) {
+			get_dma_buf(vb->planes[i].dbuf);
+
+			err = tegra_vde_dmabuf_cache_map(vde, vb->planes[i].dbuf,
+							 dma_dir, &tb->a[i],
+							 &tb->dma_base[i]);
+			if (err) {
+				dma_buf_put(vb->planes[i].dbuf);
+				goto cleanup;
+			}
+
+			continue;
+		}
+
+		if (vde->domain) {
+			sgt = vb2_dma_sg_plane_desc(vb, i);
+
+			err = tegra_vde_iommu_map(vde, sgt, &tb->iova[i],
+						  vb2_plane_size(vb, i));
+			if (err)
+				goto cleanup;
+
+			tb->dma_base[i] = iova_dma_addr(&vde->iova, tb->iova[i]);
+		} else {
+			tb->dma_base[i] = vb2_dma_contig_plane_dma_addr(vb, i);
+		}
+	}
+
+	return 0;
+
+cleanup:
+	__tegra_buf_cleanup(vb, i);
+
+	return err;
+}
+
+static void tegra_buf_cleanup(struct vb2_buffer *vb)
+{
+	__tegra_buf_cleanup(vb, vb->num_planes);
+}
+
+static int tegra_buf_prepare(struct vb2_buffer *vb)
+{
+	struct vb2_queue *vq = vb->vb2_queue;
+	struct tegra_ctx *ctx = vb2_get_drv_priv(vq);
+	struct tegra_m2m_buffer *tb = vb_to_tegra_buf(vb);
+	size_t hw_align, hw_size, hw_payload, size, offset;
+	struct v4l2_pix_format_mplane *pixfmt;
+	unsigned int i;
+	void *vb_data;
+
+	if (V4L2_TYPE_IS_OUTPUT(vq->type)) {
+		hw_align = BSEV_ALIGN;
+		pixfmt = &ctx->coded_fmt.fmt.pix_mp;
+	} else {
+		hw_align = FRAMEID_ALIGN;
+		pixfmt = &ctx->decoded_fmt.fmt.pix_mp;
+	}
+
+	for (i = 0; i < vb->num_planes; i++) {
+		offset = vb->planes[i].data_offset;
+
+		if (offset & (hw_align - 1))
+			return -EINVAL;
+
+		if (V4L2_TYPE_IS_CAPTURE(vq->type)) {
+			size = pixfmt->plane_fmt[i].sizeimage;
+			hw_payload = ALIGN(size, VDE_ATOM);
+		} else {
+			size = vb2_get_plane_payload(vb, i) - offset;
+			hw_payload = ALIGN(size + VDE_ATOM, SXE_BUFFER);
+		}
+
+		hw_size = offset + hw_payload;
+
+		if (vb2_plane_size(vb, i) < hw_size)
+			return -EINVAL;
+
+		vb2_set_plane_payload(vb, i, hw_payload);
+
+		if (V4L2_TYPE_IS_OUTPUT(vq->type)) {
+			vb_data = vb2_plane_vaddr(vb, i);
+
+			/*
+			 * Hardware requires zero-padding of coded data.
+			 * Otherwise it will fail to parse the trailing
+			 * data and abort the decoding.
+			 */
+			if (vb_data)
+				memset(vb_data + offset + size, 0,
+				       hw_size - offset - size);
+		}
+
+		tb->dma_addr[i] = tb->dma_base[i] + offset;
+	}
+
+	switch (pixfmt->pixelformat) {
+	case V4L2_PIX_FMT_YVU420M:
+		swap(tb->dma_addr[1], tb->dma_addr[2]);
+		break;
+	}
+
+	return 0;
+}
+
+static void tegra_buf_queue(struct vb2_buffer *vb)
+{
+	struct tegra_ctx *ctx = vb2_get_drv_priv(vb->vb2_queue);
+	struct vb2_v4l2_buffer *vbuf = to_vb2_v4l2_buffer(vb);
+
+	v4l2_m2m_buf_queue(ctx->fh.m2m_ctx, vbuf);
+}
+
+static void tegra_buf_request_complete(struct vb2_buffer *vb)
+{
+	struct tegra_ctx *ctx = vb2_get_drv_priv(vb->vb2_queue);
+
+	v4l2_ctrl_request_complete(vb->req_obj.req, &ctx->hdl);
+}
+
+static int tegra_start_streaming(struct vb2_queue *vq, unsigned int count)
+{
+	return 0;
+}
+
+static void tegra_stop_streaming(struct vb2_queue *vq)
+{
+	struct tegra_ctx *ctx = vb2_get_drv_priv(vq);
+
+	while (true) {
+		struct vb2_v4l2_buffer *vbuf;
+
+		if (V4L2_TYPE_IS_OUTPUT(vq->type))
+			vbuf = v4l2_m2m_src_buf_remove(ctx->fh.m2m_ctx);
+		else
+			vbuf = v4l2_m2m_dst_buf_remove(ctx->fh.m2m_ctx);
+
+		if (!vbuf)
+			break;
+
+		v4l2_ctrl_request_complete(vbuf->vb2_buf.req_obj.req, &ctx->hdl);
+		v4l2_m2m_buf_done(vbuf, VB2_BUF_STATE_ERROR);
+	}
+}
+
+static const struct vb2_ops tegra_qops = {
+	.queue_setup = tegra_queue_setup,
+	.buf_init = tegra_buf_init,
+	.buf_cleanup = tegra_buf_cleanup,
+	.buf_prepare = tegra_buf_prepare,
+	.buf_queue = tegra_buf_queue,
+	.buf_out_validate = tegra_buf_out_validate,
+	.buf_request_complete = tegra_buf_request_complete,
+	.start_streaming = tegra_start_streaming,
+	.stop_streaming = tegra_stop_streaming,
+	.wait_prepare = vb2_ops_wait_prepare,
+	.wait_finish = vb2_ops_wait_finish,
+};
+
+static int tegra_queue_init(void *priv,
+			    struct vb2_queue *src_vq,
+			    struct vb2_queue *dst_vq)
+{
+	struct tegra_ctx *ctx = priv;
+	struct tegra_vde *vde = ctx->vde;
+	const struct vb2_mem_ops *mem_ops;
+	unsigned long dma_attrs;
+	int err;
+
+	/*
+	 * TODO: Switch to use of vb2_dma_contig_memops uniformly once we
+	 * will add IOMMU_DOMAIN support for video decoder to tegra-smmu
+	 * driver. For now we need to stick with SG ops in order to be able
+	 * to get SGT table easily. This is suboptimal since SG mappings are
+	 * wasting CPU cache and we don't need that caching.
+	 */
+	if (vde->domain)
+		mem_ops = &vb2_dma_sg_memops;
+	else
+		mem_ops = &vb2_dma_contig_memops;
+
+	dma_attrs = DMA_ATTR_WRITE_COMBINE;
+
+	src_vq->buf_struct_size = sizeof(struct tegra_m2m_buffer);
+	src_vq->timestamp_flags = V4L2_BUF_FLAG_TIMESTAMP_COPY;
+	src_vq->type = V4L2_BUF_TYPE_VIDEO_OUTPUT_MPLANE;
+	src_vq->io_modes = VB2_DMABUF | VB2_MMAP;
+	src_vq->supports_requests = true;
+	src_vq->requires_requests = true;
+	src_vq->lock = &vde->v4l2_lock;
+	src_vq->dma_attrs = dma_attrs;
+	src_vq->mem_ops = mem_ops;
+	src_vq->ops = &tegra_qops;
+	src_vq->drv_priv = ctx;
+	src_vq->dev = vde->dev;
+
+	err = vb2_queue_init(src_vq);
+	if (err) {
+		v4l2_err(&vde->v4l2_dev,
+			 "failed to initialize src queue: %d\n", err);
+		return err;
+	}
+
+	/*
+	 * We may need to zero the end of bitstream in kernel if userspace
+	 * doesn't do that, hence kmap is needed for the coded data. It's not
+	 * needed for framebuffers.
+	 */
+	dma_attrs |= DMA_ATTR_NO_KERNEL_MAPPING;
+
+	dst_vq->buf_struct_size = sizeof(struct tegra_m2m_buffer);
+	dst_vq->timestamp_flags = V4L2_BUF_FLAG_TIMESTAMP_COPY;
+	dst_vq->type = V4L2_BUF_TYPE_VIDEO_CAPTURE_MPLANE;
+	dst_vq->io_modes = VB2_DMABUF | VB2_MMAP;
+	dst_vq->lock = &vde->v4l2_lock;
+	dst_vq->dma_attrs = dma_attrs;
+	dst_vq->mem_ops = mem_ops;
+	dst_vq->ops = &tegra_qops;
+	dst_vq->drv_priv = ctx;
+	dst_vq->dev = vde->dev;
+
+	err = vb2_queue_init(dst_vq);
+	if (err) {
+		v4l2_err(&vde->v4l2_dev,
+			 "failed to initialize dst queue: %d\n", err);
+		return err;
+	}
+
+	return 0;
+}
+
+static void tegra_reset_fmt(struct tegra_ctx *ctx, struct v4l2_format *f,
+			    u32 fourcc)
+{
+	memset(f, 0, sizeof(*f));
+	f->fmt.pix_mp.pixelformat = fourcc;
+	f->fmt.pix_mp.field = V4L2_FIELD_NONE;
+	f->fmt.pix_mp.xfer_func = V4L2_XFER_FUNC_DEFAULT;
+	f->fmt.pix_mp.ycbcr_enc = V4L2_YCBCR_ENC_DEFAULT;
+	f->fmt.pix_mp.colorspace = V4L2_COLORSPACE_REC709;
+	f->fmt.pix_mp.quantization = V4L2_QUANTIZATION_DEFAULT;
+}
+
+static void tegra_reset_coded_fmt(struct tegra_ctx *ctx)
+{
+	const struct tegra_vde_soc *soc = ctx->vde->soc;
+	struct v4l2_format *f = &ctx->coded_fmt;
+
+	ctx->coded_fmt_desc = &soc->coded_fmts[0];
+	tegra_reset_fmt(ctx, f, ctx->coded_fmt_desc->fourcc);
+
+	f->type = V4L2_BUF_TYPE_VIDEO_OUTPUT_MPLANE;
+	f->fmt.pix_mp.width = ctx->coded_fmt_desc->frmsize.min_width;
+	f->fmt.pix_mp.height = ctx->coded_fmt_desc->frmsize.min_height;
+}
+
+static void tegra_fill_pixfmt_mp(struct v4l2_pix_format_mplane *pixfmt,
+				 u32 pixelformat, u32 width, u32 height)
+{
+	const struct v4l2_format_info *info = v4l2_format_info(pixelformat);
+	struct v4l2_plane_pix_format *plane;
+	unsigned int i;
+
+	switch (pixelformat) {
+	case V4L2_PIX_FMT_YUV420M:
+	case V4L2_PIX_FMT_YVU420M:
+		pixfmt->width = width;
+		pixfmt->height = height;
+		pixfmt->pixelformat = pixelformat;
+		pixfmt->num_planes = info->mem_planes;
+
+		for (i = 0; i < pixfmt->num_planes; i++) {
+			unsigned int hdiv = (i == 0) ? 1 : 2;
+			unsigned int vdiv = (i == 0) ? 1 : 2;
+
+			/*
+			 * VDE is connected to Graphics Memory using 128bit port,
+			 * all memory accesses are made using 16B atoms.
+			 *
+			 * V4L requires Cb/Cr strides to be exactly half of the
+			 * Y stride, hence we're aligning Y to 16B x 2.
+			 */
+			plane = &pixfmt->plane_fmt[i];
+			plane->bytesperline = ALIGN(width, VDE_ATOM * 2) / hdiv;
+			plane->sizeimage = plane->bytesperline * height / vdiv;
+		}
+
+		break;
+	}
+}
+
+static void tegra_reset_decoded_fmt(struct tegra_ctx *ctx)
+{
+	struct v4l2_format *f = &ctx->decoded_fmt;
+
+	tegra_reset_fmt(ctx, f, ctx->coded_fmt_desc->decoded_fmts[0]);
+	f->type = V4L2_BUF_TYPE_VIDEO_CAPTURE_MPLANE;
+	tegra_fill_pixfmt_mp(&f->fmt.pix_mp,
+			     ctx->coded_fmt_desc->decoded_fmts[0],
+			     ctx->coded_fmt.fmt.pix_mp.width,
+			     ctx->coded_fmt.fmt.pix_mp.height);
+}
+
+static void tegra_job_finish(struct tegra_ctx *ctx,
+			     enum vb2_buffer_state result)
+{
+	v4l2_m2m_buf_done_and_job_finish(ctx->vde->m2m, ctx->fh.m2m_ctx,
+					 result);
+}
+
+static void tegra_decode_complete(struct work_struct *work)
+{
+	struct tegra_ctx *ctx = container_of(work, struct tegra_ctx, work);
+	int err;
+
+	err = ctx->coded_fmt_desc->decode_wait(ctx);
+	if (err)
+		tegra_job_finish(ctx, VB2_BUF_STATE_ERROR);
+	else
+		tegra_job_finish(ctx, VB2_BUF_STATE_DONE);
+}
+
+static int tegra_querycap(struct file *file, void *priv,
+			  struct v4l2_capability *cap)
+{
+	strscpy(cap->bus_info, "platform:tegra-vde", sizeof(cap->bus_info));
+	strscpy(cap->driver, "tegra-vde", sizeof(cap->driver));
+	strscpy(cap->card, "tegra-vde", sizeof(cap->card));
+
+	return 0;
+}
+
+static int tegra_enum_decoded_fmt(struct file *file, void *priv,
+				  struct v4l2_fmtdesc *f)
+{
+	struct tegra_ctx *ctx = fh_to_tegra_ctx(priv);
+
+	if (WARN_ON(!ctx->coded_fmt_desc))
+		return -EINVAL;
+
+	if (f->index >= ctx->coded_fmt_desc->num_decoded_fmts)
+		return -EINVAL;
+
+	f->pixelformat = ctx->coded_fmt_desc->decoded_fmts[f->index];
+
+	return 0;
+}
+
+static int tegra_g_decoded_fmt(struct file *file, void *priv,
+			       struct v4l2_format *f)
+{
+	struct tegra_ctx *ctx = fh_to_tegra_ctx(priv);
+
+	*f = ctx->decoded_fmt;
+	return 0;
+}
+
+static int tegra_try_decoded_fmt(struct file *file, void *priv,
+				 struct v4l2_format *f)
+{
+	struct v4l2_pix_format_mplane *pix_mp = &f->fmt.pix_mp;
+	struct tegra_ctx *ctx = fh_to_tegra_ctx(priv);
+	const struct tegra_coded_fmt_desc *coded_desc;
+	unsigned int i;
+
+	/*
+	 * The codec context should point to a coded format desc, if the format
+	 * on the coded end has not been set yet, it should point to the
+	 * default value.
+	 */
+	coded_desc = ctx->coded_fmt_desc;
+	if (WARN_ON(!coded_desc))
+		return -EINVAL;
+
+	if (!coded_desc->num_decoded_fmts)
+		return -EINVAL;
+
+	for (i = 0; i < coded_desc->num_decoded_fmts; i++) {
+		if (coded_desc->decoded_fmts[i] == pix_mp->pixelformat)
+			break;
+	}
+
+	if (i == coded_desc->num_decoded_fmts)
+		pix_mp->pixelformat = coded_desc->decoded_fmts[0];
+
+	/* always apply the frmsize constraint of the coded end */
+	v4l2_apply_frmsize_constraints(&pix_mp->width,
+				       &pix_mp->height,
+				       &coded_desc->frmsize);
+
+	tegra_fill_pixfmt_mp(pix_mp, pix_mp->pixelformat,
+			     pix_mp->width, pix_mp->height);
+	pix_mp->field = V4L2_FIELD_NONE;
+
+	return 0;
+}
+
+static int tegra_s_decoded_fmt(struct file *file, void *priv,
+			       struct v4l2_format *f)
+{
+	struct tegra_ctx *ctx = fh_to_tegra_ctx(priv);
+	struct vb2_queue *vq;
+	int err;
+
+	/* change not allowed if queue is busy */
+	vq = v4l2_m2m_get_vq(ctx->fh.m2m_ctx,
+			     V4L2_BUF_TYPE_VIDEO_CAPTURE_MPLANE);
+	if (vb2_is_busy(vq))
+		return -EBUSY;
+
+	err = tegra_try_decoded_fmt(file, priv, f);
+	if (err)
+		return err;
+
+	ctx->decoded_fmt = *f;
+
+	return 0;
+}
+
+static int tegra_enum_coded_fmt(struct file *file, void *priv,
+				struct v4l2_fmtdesc *f)
+{
+	struct tegra_ctx *ctx = fh_to_tegra_ctx(priv);
+	const struct tegra_vde_soc *soc = ctx->vde->soc;
+
+	if (f->index >= soc->num_coded_fmts)
+		return -EINVAL;
+
+	f->pixelformat = soc->coded_fmts[f->index].fourcc;
+
+	return 0;
+}
+
+static int tegra_g_coded_fmt(struct file *file, void *priv,
+			     struct v4l2_format *f)
+{
+	struct tegra_ctx *ctx = fh_to_tegra_ctx(priv);
+
+	*f = ctx->coded_fmt;
+	return 0;
+}
+
+static const struct tegra_coded_fmt_desc *
+tegra_find_coded_fmt_desc(struct tegra_ctx *ctx, u32 fourcc)
+{
+	const struct tegra_vde_soc *soc = ctx->vde->soc;
+	unsigned int i;
+
+	for (i = 0; i < soc->num_coded_fmts; i++) {
+		if (soc->coded_fmts[i].fourcc == fourcc)
+			return &soc->coded_fmts[i];
+	}
+
+	return NULL;
+}
+
+static int tegra_try_coded_fmt(struct file *file, void *priv,
+			       struct v4l2_format *f)
+{
+	struct v4l2_pix_format_mplane *pix_mp = &f->fmt.pix_mp;
+	struct tegra_ctx *ctx = fh_to_tegra_ctx(priv);
+	const struct tegra_vde_soc *soc = ctx->vde->soc;
+	int size = pix_mp->plane_fmt[0].sizeimage;
+	const struct tegra_coded_fmt_desc *desc;
+
+	desc = tegra_find_coded_fmt_desc(ctx, pix_mp->pixelformat);
+	if (!desc) {
+		pix_mp->pixelformat = soc->coded_fmts[0].fourcc;
+		desc = &soc->coded_fmts[0];
+	}
+
+	v4l2_apply_frmsize_constraints(&pix_mp->width,
+				       &pix_mp->height,
+				       &desc->frmsize);
+
+	pix_mp->plane_fmt[0].sizeimage = max(ALIGN(size, SXE_BUFFER), SZ_2M);
+	pix_mp->field = V4L2_FIELD_NONE;
+	pix_mp->num_planes = 1;
+
+	return 0;
+}
+
+static int tegra_s_coded_fmt(struct file *file, void *priv,
+			     struct v4l2_format *f)
+{
+	struct tegra_ctx *ctx = fh_to_tegra_ctx(priv);
+	struct v4l2_m2m_ctx *m2m_ctx = ctx->fh.m2m_ctx;
+	const struct tegra_coded_fmt_desc *desc;
+	struct vb2_queue *peer_vq, *vq;
+	struct v4l2_format *cap_fmt;
+	int err;
+
+	/*
+	 * In order to support dynamic resolution change, the decoder admits
+	 * a resolution change, as long as the pixelformat remains. Can't be
+	 * done if streaming.
+	 */
+	vq = v4l2_m2m_get_vq(m2m_ctx, V4L2_BUF_TYPE_VIDEO_OUTPUT_MPLANE);
+	if (vb2_is_streaming(vq) ||
+	    (vb2_is_busy(vq) &&
+	     f->fmt.pix_mp.pixelformat != ctx->coded_fmt.fmt.pix_mp.pixelformat))
+		return -EBUSY;
+
+	/*
+	 * Since format change on the OUTPUT queue will reset the CAPTURE
+	 * queue, we can't allow doing so when the CAPTURE queue has buffers
+	 * allocated.
+	 */
+	peer_vq = v4l2_m2m_get_vq(m2m_ctx, V4L2_BUF_TYPE_VIDEO_CAPTURE_MPLANE);
+	if (vb2_is_busy(peer_vq))
+		return -EBUSY;
+
+	err = tegra_try_coded_fmt(file, priv, f);
+	if (err)
+		return err;
+
+	desc = tegra_find_coded_fmt_desc(ctx, f->fmt.pix_mp.pixelformat);
+	if (!desc)
+		return -EINVAL;
+
+	ctx->coded_fmt_desc = desc;
+	ctx->coded_fmt = *f;
+
+	/*
+	 * Current decoded format might have become invalid with newly
+	 * selected codec, so reset it to default just to be safe and
+	 * keep internal driver state sane. User is mandated to set
+	 * the decoded format again after we return, so we don't need
+	 * anything smarter.
+	 *
+	 * Note that this will propagates any size changes to the decoded format.
+	 */
+	tegra_reset_decoded_fmt(ctx);
+
+	/* propagate colorspace information to capture */
+	cap_fmt = &ctx->decoded_fmt;
+	cap_fmt->fmt.pix_mp.xfer_func = f->fmt.pix_mp.xfer_func;
+	cap_fmt->fmt.pix_mp.ycbcr_enc = f->fmt.pix_mp.ycbcr_enc;
+	cap_fmt->fmt.pix_mp.colorspace = f->fmt.pix_mp.colorspace;
+	cap_fmt->fmt.pix_mp.quantization = f->fmt.pix_mp.quantization;
+
+	return 0;
+}
+
+static int tegra_enum_framesizes(struct file *file, void *priv,
+				 struct v4l2_frmsizeenum *fsize)
+{
+	struct tegra_ctx *ctx = fh_to_tegra_ctx(priv);
+	const struct tegra_coded_fmt_desc *fmt;
+
+	if (fsize->index)
+		return -EINVAL;
+
+	fmt = tegra_find_coded_fmt_desc(ctx, fsize->pixel_format);
+	if (!fmt)
+		return -EINVAL;
+
+	fsize->type = V4L2_FRMSIZE_TYPE_STEPWISE;
+	fsize->stepwise = fmt->frmsize;
+
+	return 0;
+}
+
+static const struct v4l2_ioctl_ops tegra_v4l2_ioctl_ops = {
+	.vidioc_querycap = tegra_querycap,
+	.vidioc_enum_framesizes = tegra_enum_framesizes,
+
+	.vidioc_try_fmt_vid_out_mplane = tegra_try_coded_fmt,
+	.vidioc_g_fmt_vid_out_mplane = tegra_g_coded_fmt,
+	.vidioc_s_fmt_vid_out_mplane = tegra_s_coded_fmt,
+	.vidioc_enum_fmt_vid_out = tegra_enum_coded_fmt,
+
+	.vidioc_try_fmt_vid_cap_mplane = tegra_try_decoded_fmt,
+	.vidioc_g_fmt_vid_cap_mplane = tegra_g_decoded_fmt,
+	.vidioc_s_fmt_vid_cap_mplane = tegra_s_decoded_fmt,
+	.vidioc_enum_fmt_vid_cap = tegra_enum_decoded_fmt,
+
+	.vidioc_reqbufs = v4l2_m2m_ioctl_reqbufs,
+	.vidioc_querybuf = v4l2_m2m_ioctl_querybuf,
+	.vidioc_qbuf = v4l2_m2m_ioctl_qbuf,
+	.vidioc_dqbuf = v4l2_m2m_ioctl_dqbuf,
+	.vidioc_prepare_buf = v4l2_m2m_ioctl_prepare_buf,
+	.vidioc_create_bufs = v4l2_m2m_ioctl_create_bufs,
+	.vidioc_expbuf = v4l2_m2m_ioctl_expbuf,
+
+	.vidioc_streamon = v4l2_m2m_ioctl_streamon,
+	.vidioc_streamoff = v4l2_m2m_ioctl_streamoff,
+
+	.vidioc_subscribe_event = v4l2_ctrl_subscribe_event,
+	.vidioc_unsubscribe_event = v4l2_event_unsubscribe,
+};
+
+static int tegra_init_ctrls(struct tegra_ctx *ctx)
+{
+	unsigned int i;
+	int err;
+
+	err = v4l2_ctrl_handler_init(&ctx->hdl, ARRAY_SIZE(ctrl_cfgs));
+	if (err)
+		return err;
+
+	for (i = 0; i < ARRAY_SIZE(ctrl_cfgs); i++) {
+		ctx->ctrls[i] = v4l2_ctrl_new_custom(&ctx->hdl, &ctrl_cfgs[i],
+						     NULL);
+		if (ctx->hdl.error) {
+			err = ctx->hdl.error;
+			goto free_ctrls;
+		}
+	}
+
+	err = v4l2_ctrl_handler_setup(&ctx->hdl);
+	if (err)
+		goto free_ctrls;
+
+	ctx->fh.ctrl_handler = &ctx->hdl;
+
+	return 0;
+
+free_ctrls:
+	v4l2_ctrl_handler_free(&ctx->hdl);
+
+	return err;
+}
+
+static int tegra_init_m2m(struct tegra_ctx *ctx)
+{
+	ctx->fh.m2m_ctx = v4l2_m2m_ctx_init(ctx->vde->m2m,
+					    ctx, tegra_queue_init);
+	if (IS_ERR(ctx->fh.m2m_ctx))
+		return PTR_ERR(ctx->fh.m2m_ctx);
+
+	return 0;
+}
+
+static int tegra_open(struct file *file)
+{
+	struct tegra_vde *vde = video_drvdata(file);
+	struct tegra_ctx *ctx;
+	int err;
+
+	ctx = kzalloc(offsetof(struct tegra_ctx, ctrls[ARRAY_SIZE(ctrl_cfgs)]),
+		      GFP_KERNEL);
+	if (!ctx)
+		return -ENOMEM;
+
+	ctx->vde = vde;
+	v4l2_fh_init(&ctx->fh, video_devdata(file));
+	INIT_WORK(&ctx->work, tegra_decode_complete);
+
+	err = tegra_init_ctrls(ctx);
+	if (err) {
+		v4l2_err(&vde->v4l2_dev, "failed to add controls: %d\n", err);
+		goto free_ctx;
+	}
+
+	err = tegra_init_m2m(ctx);
+	if (err) {
+		v4l2_err(&vde->v4l2_dev, "failed to initialize m2m: %d\n", err);
+		goto free_ctrls;
+	}
+
+	file->private_data = &ctx->fh;
+	v4l2_fh_add(&ctx->fh);
+
+	tegra_reset_coded_fmt(ctx);
+	tegra_try_coded_fmt(file, file->private_data, &ctx->coded_fmt);
+
+	tegra_reset_decoded_fmt(ctx);
+	tegra_try_decoded_fmt(file, file->private_data, &ctx->decoded_fmt);
+
+	return 0;
+
+free_ctrls:
+	v4l2_ctrl_handler_free(&ctx->hdl);
+free_ctx:
+	kfree(ctx);
+
+	return err;
+}
+
+static int tegra_release(struct file *file)
+{
+	struct v4l2_fh *fh = file->private_data;
+	struct tegra_ctx *ctx = fh_to_tegra_ctx(fh);
+	struct tegra_vde *vde = ctx->vde;
+
+	v4l2_fh_del(fh);
+	v4l2_m2m_ctx_release(fh->m2m_ctx);
+	v4l2_ctrl_handler_free(&ctx->hdl);
+	v4l2_fh_exit(fh);
+	kfree(ctx);
+
+	tegra_vde_dmabuf_cache_unmap_sync(vde);
+
+	return 0;
+}
+
+static const struct v4l2_file_operations tegra_v4l2_fops = {
+	.owner = THIS_MODULE,
+	.open = tegra_open,
+	.poll = v4l2_m2m_fop_poll,
+	.mmap = v4l2_m2m_fop_mmap,
+	.release = tegra_release,
+	.unlocked_ioctl = video_ioctl2,
+};
+
+static void tegra_device_run(void *priv)
+{
+	struct tegra_ctx *ctx = priv;
+	struct vb2_v4l2_buffer *src = v4l2_m2m_next_src_buf(ctx->fh.m2m_ctx);
+	struct media_request *src_req = src->vb2_buf.req_obj.req;
+	int err;
+
+	v4l2_ctrl_request_setup(src_req, &ctx->hdl);
+
+	err = ctx->coded_fmt_desc->decode_run(ctx);
+
+	v4l2_ctrl_request_complete(src_req, &ctx->hdl);
+
+	if (err)
+		tegra_job_finish(ctx, VB2_BUF_STATE_ERROR);
+	else
+		queue_work(ctx->vde->wq, &ctx->work);
+}
+
+static const struct v4l2_m2m_ops tegra_v4l2_m2m_ops = {
+	.device_run = tegra_device_run,
+};
+
+static int tegra_request_validate(struct media_request *req)
+{
+	unsigned int count;
+
+	count = vb2_request_buffer_cnt(req);
+	if (!count)
+		return -ENOENT;
+	else if (count > 1)
+		return -EINVAL;
+
+	return vb2_request_validate(req);
+}
+
+static const struct media_device_ops tegra_media_device_ops = {
+	.req_validate = tegra_request_validate,
+	.req_queue = v4l2_m2m_request_queue,
+};
+
+int tegra_vde_v4l2_init(struct tegra_vde *vde)
+{
+	struct device *dev = vde->dev;
+	int err;
+
+	mutex_init(&vde->v4l2_lock);
+	media_device_init(&vde->mdev);
+	video_set_drvdata(&vde->vdev, vde);
+
+	vde->vdev.lock = &vde->v4l2_lock,
+	vde->vdev.fops = &tegra_v4l2_fops,
+	vde->vdev.vfl_dir = VFL_DIR_M2M,
+	vde->vdev.release = video_device_release_empty,
+	vde->vdev.v4l2_dev = &vde->v4l2_dev;
+	vde->vdev.ioctl_ops = &tegra_v4l2_ioctl_ops,
+	vde->vdev.device_caps = V4L2_CAP_VIDEO_M2M_MPLANE | V4L2_CAP_STREAMING,
+
+	vde->v4l2_dev.mdev = &vde->mdev;
+	vde->mdev.ops = &tegra_media_device_ops;
+	vde->mdev.dev = dev;
+
+	strscpy(vde->mdev.model, "tegra-vde", sizeof(vde->mdev.model));
+	strscpy(vde->vdev.name,  "tegra-vde", sizeof(vde->vdev.name));
+	strscpy(vde->mdev.bus_info, "platform:tegra-vde",
+		sizeof(vde->mdev.bus_info));
+
+	vde->wq = create_workqueue("tegra-vde");
+	if (!vde->wq)
+		return -ENOMEM;
+
+	err = media_device_register(&vde->mdev);
+	if (err) {
+		dev_err(dev, "failed to register media device: %d\n", err);
+		goto clean_up_media_device;
+	}
+
+	err = v4l2_device_register(dev, &vde->v4l2_dev);
+	if (err) {
+		dev_err(dev, "failed to register v4l2 device: %d\n", err);
+		goto unreg_media_device;
+	}
+
+	err = video_register_device(&vde->vdev, VFL_TYPE_VIDEO, -1);
+	if (err) {
+		dev_err(dev, "failed to register video device: %d\n", err);
+		goto unreg_v4l2;
+	}
+
+	vde->m2m = v4l2_m2m_init(&tegra_v4l2_m2m_ops);
+	err = PTR_ERR_OR_ZERO(vde->m2m);
+	if (err) {
+		dev_err(dev, "failed to initialize m2m device: %d\n", err);
+		goto unreg_video_device;
+	}
+
+	err = v4l2_m2m_register_media_controller(vde->m2m, &vde->vdev,
+						 MEDIA_ENT_F_PROC_VIDEO_DECODER);
+	if (err) {
+		dev_err(dev, "failed to register media controller: %d\n", err);
+		goto release_m2m;
+	}
+
+	v4l2_info(&vde->v4l2_dev, "v4l2 device registered as /dev/video%d\n",
+		  vde->vdev.num);
+
+	return 0;
+
+release_m2m:
+	v4l2_m2m_release(vde->m2m);
+unreg_video_device:
+	video_unregister_device(&vde->vdev);
+unreg_v4l2:
+	v4l2_device_unregister(&vde->v4l2_dev);
+unreg_media_device:
+	media_device_unregister(&vde->mdev);
+clean_up_media_device:
+	media_device_cleanup(&vde->mdev);
+
+	destroy_workqueue(vde->wq);
+
+	return err;
+}
+
+void tegra_vde_v4l2_deinit(struct tegra_vde *vde)
+{
+	v4l2_m2m_unregister_media_controller(vde->m2m);
+	v4l2_m2m_release(vde->m2m);
+
+	video_unregister_device(&vde->vdev);
+	v4l2_device_unregister(&vde->v4l2_dev);
+
+	media_device_unregister(&vde->mdev);
+	media_device_cleanup(&vde->mdev);
+
+	destroy_workqueue(vde->wq);
+}
diff --git a/drivers/media/platform/nvidia/tegra-vde/vde.c b/drivers/media/platform/nvidia/tegra-vde/vde.c
new file mode 100644
index 000000000..f3e863a94
--- /dev/null
+++ b/drivers/media/platform/nvidia/tegra-vde/vde.c
@@ -0,0 +1,551 @@
+// SPDX-License-Identifier: GPL-2.0+
+/*
+ * NVIDIA Tegra Video decoder driver
+ *
+ * Copyright (C) 2016-2017 Dmitry Osipenko <digetx@gmail.com>
+ *
+ */
+
+#include <linux/clk.h>
+#include <linux/dma-buf.h>
+#include <linux/genalloc.h>
+#include <linux/interrupt.h>
+#include <linux/list.h>
+#include <linux/module.h>
+#include <linux/of_device.h>
+#include <linux/pm_runtime.h>
+#include <linux/reset.h>
+#include <linux/slab.h>
+#include <linux/uaccess.h>
+
+#include <soc/tegra/common.h>
+#include <soc/tegra/pmc.h>
+
+#include "vde.h"
+
+#define CREATE_TRACE_POINTS
+#include "trace.h"
+
+void tegra_vde_writel(struct tegra_vde *vde, u32 value,
+		      void __iomem *base, u32 offset)
+{
+	trace_vde_writel(vde, base, offset, value);
+
+	writel_relaxed(value, base + offset);
+}
+
+u32 tegra_vde_readl(struct tegra_vde *vde, void __iomem *base, u32 offset)
+{
+	u32 value = readl_relaxed(base + offset);
+
+	trace_vde_readl(vde, base, offset, value);
+
+	return value;
+}
+
+void tegra_vde_set_bits(struct tegra_vde *vde, u32 mask,
+			void __iomem *base, u32 offset)
+{
+	u32 value = tegra_vde_readl(vde, base, offset);
+
+	tegra_vde_writel(vde, value | mask, base, offset);
+}
+
+int tegra_vde_alloc_bo(struct tegra_vde *vde,
+		       struct tegra_vde_bo **ret_bo,
+		       enum dma_data_direction dma_dir,
+		       size_t size)
+{
+	struct device *dev = vde->dev;
+	struct tegra_vde_bo *bo;
+	int err;
+
+	bo = kzalloc(sizeof(*bo), GFP_KERNEL);
+	if (!bo)
+		return -ENOMEM;
+
+	bo->vde = vde;
+	bo->size = size;
+	bo->dma_dir = dma_dir;
+	bo->dma_attrs = DMA_ATTR_WRITE_COMBINE |
+			DMA_ATTR_NO_KERNEL_MAPPING;
+
+	if (!vde->domain)
+		bo->dma_attrs |= DMA_ATTR_FORCE_CONTIGUOUS;
+
+	bo->dma_cookie = dma_alloc_attrs(dev, bo->size, &bo->dma_handle,
+					 GFP_KERNEL, bo->dma_attrs);
+	if (!bo->dma_cookie) {
+		dev_err(dev, "Failed to allocate DMA buffer of size: %zu\n",
+			bo->size);
+		err = -ENOMEM;
+		goto free_bo;
+	}
+
+	err = dma_get_sgtable_attrs(dev, &bo->sgt, bo->dma_cookie,
+				    bo->dma_handle, bo->size, bo->dma_attrs);
+	if (err) {
+		dev_err(dev, "Failed to get DMA buffer SG table: %d\n", err);
+		goto free_attrs;
+	}
+
+	err = dma_map_sgtable(dev, &bo->sgt, bo->dma_dir, bo->dma_attrs);
+	if (err) {
+		dev_err(dev, "Failed to map DMA buffer SG table: %d\n", err);
+		goto free_table;
+	}
+
+	if (vde->domain) {
+		err = tegra_vde_iommu_map(vde, &bo->sgt, &bo->iova, bo->size);
+		if (err) {
+			dev_err(dev, "Failed to map DMA buffer IOVA: %d\n", err);
+			goto unmap_sgtable;
+		}
+
+		bo->dma_addr = iova_dma_addr(&vde->iova, bo->iova);
+	} else {
+		bo->dma_addr = sg_dma_address(bo->sgt.sgl);
+	}
+
+	*ret_bo = bo;
+
+	return 0;
+
+unmap_sgtable:
+	dma_unmap_sgtable(dev, &bo->sgt, bo->dma_dir, bo->dma_attrs);
+free_table:
+	sg_free_table(&bo->sgt);
+free_attrs:
+	dma_free_attrs(dev, bo->size, bo->dma_cookie, bo->dma_handle,
+		       bo->dma_attrs);
+free_bo:
+	kfree(bo);
+
+	return err;
+}
+
+void tegra_vde_free_bo(struct tegra_vde_bo *bo)
+{
+	struct tegra_vde *vde = bo->vde;
+	struct device *dev = vde->dev;
+
+	if (vde->domain)
+		tegra_vde_iommu_unmap(vde, bo->iova);
+
+	dma_unmap_sgtable(dev, &bo->sgt, bo->dma_dir, bo->dma_attrs);
+
+	sg_free_table(&bo->sgt);
+
+	dma_free_attrs(dev, bo->size, bo->dma_cookie, bo->dma_handle,
+		       bo->dma_attrs);
+	kfree(bo);
+}
+
+static irqreturn_t tegra_vde_isr(int irq, void *data)
+{
+	struct tegra_vde *vde = data;
+
+	if (completion_done(&vde->decode_completion))
+		return IRQ_NONE;
+
+	tegra_vde_set_bits(vde, 0, vde->frameid, 0x208);
+	complete(&vde->decode_completion);
+
+	return IRQ_HANDLED;
+}
+
+static __maybe_unused int tegra_vde_runtime_suspend(struct device *dev)
+{
+	struct tegra_vde *vde = dev_get_drvdata(dev);
+	int err;
+
+	if (!dev->pm_domain) {
+		err = tegra_powergate_power_off(TEGRA_POWERGATE_VDEC);
+		if (err) {
+			dev_err(dev, "Failed to power down HW: %d\n", err);
+			return err;
+		}
+	}
+
+	clk_disable_unprepare(vde->clk);
+	reset_control_release(vde->rst);
+	reset_control_release(vde->rst_mc);
+
+	return 0;
+}
+
+static __maybe_unused int tegra_vde_runtime_resume(struct device *dev)
+{
+	struct tegra_vde *vde = dev_get_drvdata(dev);
+	int err;
+
+	err = reset_control_acquire(vde->rst_mc);
+	if (err) {
+		dev_err(dev, "Failed to acquire mc reset: %d\n", err);
+		return err;
+	}
+
+	err = reset_control_acquire(vde->rst);
+	if (err) {
+		dev_err(dev, "Failed to acquire reset: %d\n", err);
+		goto release_mc_reset;
+	}
+
+	if (!dev->pm_domain) {
+		err = tegra_powergate_sequence_power_up(TEGRA_POWERGATE_VDEC,
+							vde->clk, vde->rst);
+		if (err) {
+			dev_err(dev, "Failed to power up HW : %d\n", err);
+			goto release_reset;
+		}
+	} else {
+		/*
+		 * tegra_powergate_sequence_power_up() leaves clocks enabled,
+		 * while GENPD not.
+		 */
+		err = clk_prepare_enable(vde->clk);
+		if (err) {
+			dev_err(dev, "Failed to enable clock: %d\n", err);
+			goto release_reset;
+		}
+	}
+
+	return 0;
+
+release_reset:
+	reset_control_release(vde->rst);
+release_mc_reset:
+	reset_control_release(vde->rst_mc);
+
+	return err;
+}
+
+static int tegra_vde_probe(struct platform_device *pdev)
+{
+	struct device *dev = &pdev->dev;
+	struct tegra_vde *vde;
+	int irq, err;
+
+	vde = devm_kzalloc(dev, sizeof(*vde), GFP_KERNEL);
+	if (!vde)
+		return -ENOMEM;
+
+	platform_set_drvdata(pdev, vde);
+
+	vde->soc = of_device_get_match_data(&pdev->dev);
+	vde->dev = dev;
+
+	vde->sxe = devm_platform_ioremap_resource_byname(pdev, "sxe");
+	if (IS_ERR(vde->sxe))
+		return PTR_ERR(vde->sxe);
+
+	vde->bsev = devm_platform_ioremap_resource_byname(pdev, "bsev");
+	if (IS_ERR(vde->bsev))
+		return PTR_ERR(vde->bsev);
+
+	vde->mbe = devm_platform_ioremap_resource_byname(pdev, "mbe");
+	if (IS_ERR(vde->mbe))
+		return PTR_ERR(vde->mbe);
+
+	vde->ppe = devm_platform_ioremap_resource_byname(pdev, "ppe");
+	if (IS_ERR(vde->ppe))
+		return PTR_ERR(vde->ppe);
+
+	vde->mce = devm_platform_ioremap_resource_byname(pdev, "mce");
+	if (IS_ERR(vde->mce))
+		return PTR_ERR(vde->mce);
+
+	vde->tfe = devm_platform_ioremap_resource_byname(pdev, "tfe");
+	if (IS_ERR(vde->tfe))
+		return PTR_ERR(vde->tfe);
+
+	vde->ppb = devm_platform_ioremap_resource_byname(pdev, "ppb");
+	if (IS_ERR(vde->ppb))
+		return PTR_ERR(vde->ppb);
+
+	vde->vdma = devm_platform_ioremap_resource_byname(pdev, "vdma");
+	if (IS_ERR(vde->vdma))
+		return PTR_ERR(vde->vdma);
+
+	vde->frameid = devm_platform_ioremap_resource_byname(pdev, "frameid");
+	if (IS_ERR(vde->frameid))
+		return PTR_ERR(vde->frameid);
+
+	vde->clk = devm_clk_get(dev, NULL);
+	if (IS_ERR(vde->clk)) {
+		err = PTR_ERR(vde->clk);
+		dev_err(dev, "Could not get VDE clk %d\n", err);
+		return err;
+	}
+
+	vde->rst = devm_reset_control_get_exclusive_released(dev, NULL);
+	if (IS_ERR(vde->rst)) {
+		err = PTR_ERR(vde->rst);
+		dev_err(dev, "Could not get VDE reset %d\n", err);
+		return err;
+	}
+
+	vde->rst_mc = devm_reset_control_get_optional_exclusive_released(dev, "mc");
+	if (IS_ERR(vde->rst_mc)) {
+		err = PTR_ERR(vde->rst_mc);
+		dev_err(dev, "Could not get MC reset %d\n", err);
+		return err;
+	}
+
+	irq = platform_get_irq_byname(pdev, "sync-token");
+	if (irq < 0)
+		return irq;
+
+	err = devm_request_irq(dev, irq, tegra_vde_isr, 0,
+			       dev_name(dev), vde);
+	if (err) {
+		dev_err(dev, "Could not request IRQ %d\n", err);
+		return err;
+	}
+
+	err = devm_tegra_core_dev_init_opp_table_common(dev);
+	if (err) {
+		dev_err(dev, "Could initialize OPP table %d\n", err);
+		return err;
+	}
+
+	vde->iram_pool = of_gen_pool_get(dev->of_node, "iram", 0);
+	if (!vde->iram_pool) {
+		dev_err(dev, "Could not get IRAM pool\n");
+		return -EPROBE_DEFER;
+	}
+
+	vde->iram = gen_pool_dma_alloc(vde->iram_pool,
+				       gen_pool_size(vde->iram_pool),
+				       &vde->iram_lists_addr);
+	if (!vde->iram) {
+		dev_err(dev, "Could not reserve IRAM\n");
+		return -ENOMEM;
+	}
+
+	INIT_LIST_HEAD(&vde->map_list);
+	mutex_init(&vde->map_lock);
+	mutex_init(&vde->lock);
+	init_completion(&vde->decode_completion);
+
+	err = tegra_vde_iommu_init(vde);
+	if (err) {
+		dev_err(dev, "Failed to initialize IOMMU: %d\n", err);
+		goto err_gen_free;
+	}
+
+	pm_runtime_enable(dev);
+	pm_runtime_use_autosuspend(dev);
+	pm_runtime_set_autosuspend_delay(dev, 300);
+
+	/*
+	 * VDE partition may be left ON after bootloader, hence let's
+	 * power-cycle it in order to put hardware into a predictable lower
+	 * power state.
+	 */
+	err = pm_runtime_resume_and_get(dev);
+	if (err)
+		goto err_pm_runtime;
+
+	pm_runtime_put(dev);
+
+	err = tegra_vde_alloc_bo(vde, &vde->secure_bo, DMA_FROM_DEVICE, 4096);
+	if (err) {
+		dev_err(dev, "Failed to allocate secure BO: %d\n", err);
+		goto err_pm_runtime;
+	}
+
+	err = tegra_vde_v4l2_init(vde);
+	if (err) {
+		dev_err(dev, "Failed to initialize V4L2: %d\n", err);
+		goto err_free_secure_bo;
+	}
+
+	return 0;
+
+err_free_secure_bo:
+	tegra_vde_free_bo(vde->secure_bo);
+err_pm_runtime:
+	pm_runtime_dont_use_autosuspend(dev);
+	pm_runtime_disable(dev);
+
+	tegra_vde_iommu_deinit(vde);
+
+err_gen_free:
+	gen_pool_free(vde->iram_pool, (unsigned long)vde->iram,
+		      gen_pool_size(vde->iram_pool));
+
+	return err;
+}
+
+static int tegra_vde_remove(struct platform_device *pdev)
+{
+	struct tegra_vde *vde = platform_get_drvdata(pdev);
+	struct device *dev = &pdev->dev;
+
+	tegra_vde_v4l2_deinit(vde);
+	tegra_vde_free_bo(vde->secure_bo);
+
+	/*
+	 * As it increments RPM usage_count even on errors, we don't need to
+	 * check the returned code here.
+	 */
+	pm_runtime_get_sync(dev);
+
+	pm_runtime_dont_use_autosuspend(dev);
+	pm_runtime_disable(dev);
+
+	/*
+	 * Balance RPM state, the VDE power domain is left ON and hardware
+	 * is clock-gated. It's safe to reboot machine now.
+	 */
+	pm_runtime_put_noidle(dev);
+	clk_disable_unprepare(vde->clk);
+
+	tegra_vde_dmabuf_cache_unmap_all(vde);
+	tegra_vde_iommu_deinit(vde);
+
+	gen_pool_free(vde->iram_pool, (unsigned long)vde->iram,
+		      gen_pool_size(vde->iram_pool));
+
+	return 0;
+}
+
+static void tegra_vde_shutdown(struct platform_device *pdev)
+{
+	/*
+	 * On some devices bootloader isn't ready to a power-gated VDE on
+	 * a warm-reboot, machine will hang in that case.
+	 */
+	pm_runtime_get_sync(&pdev->dev);
+}
+
+static __maybe_unused int tegra_vde_pm_suspend(struct device *dev)
+{
+	struct tegra_vde *vde = dev_get_drvdata(dev);
+	int err;
+
+	mutex_lock(&vde->lock);
+
+	err = pm_runtime_force_suspend(dev);
+	if (err < 0)
+		return err;
+
+	return 0;
+}
+
+static __maybe_unused int tegra_vde_pm_resume(struct device *dev)
+{
+	struct tegra_vde *vde = dev_get_drvdata(dev);
+	int err;
+
+	err = pm_runtime_force_resume(dev);
+	if (err < 0)
+		return err;
+
+	mutex_unlock(&vde->lock);
+
+	return 0;
+}
+
+static const struct dev_pm_ops tegra_vde_pm_ops = {
+	SET_RUNTIME_PM_OPS(tegra_vde_runtime_suspend,
+			   tegra_vde_runtime_resume,
+			   NULL)
+	SET_SYSTEM_SLEEP_PM_OPS(tegra_vde_pm_suspend,
+				tegra_vde_pm_resume)
+};
+
+static const u32 tegra124_decoded_fmts[] = {
+	/* TBD: T124 supports only a non-standard Tegra tiled format */
+};
+
+static const struct tegra_coded_fmt_desc tegra124_coded_fmts[] = {
+	{
+		.fourcc = V4L2_PIX_FMT_H264_SLICE,
+		.frmsize = {
+			.min_width = 16,
+			.max_width = 1920,
+			.step_width = 16,
+			.min_height = 16,
+			.max_height = 2032,
+			.step_height = 16,
+		},
+		.num_decoded_fmts = ARRAY_SIZE(tegra124_decoded_fmts),
+		.decoded_fmts = tegra124_decoded_fmts,
+		.decode_run = tegra_vde_h264_decode_run,
+		.decode_wait = tegra_vde_h264_decode_wait,
+	},
+};
+
+static const u32 tegra20_decoded_fmts[] = {
+	V4L2_PIX_FMT_YUV420M,
+	V4L2_PIX_FMT_YVU420M,
+};
+
+static const struct tegra_coded_fmt_desc tegra20_coded_fmts[] = {
+	{
+		.fourcc = V4L2_PIX_FMT_H264_SLICE,
+		.frmsize = {
+			.min_width = 16,
+			.max_width = 1920,
+			.step_width = 16,
+			.min_height = 16,
+			.max_height = 2032,
+			.step_height = 16,
+		},
+		.num_decoded_fmts = ARRAY_SIZE(tegra20_decoded_fmts),
+		.decoded_fmts = tegra20_decoded_fmts,
+		.decode_run = tegra_vde_h264_decode_run,
+		.decode_wait = tegra_vde_h264_decode_wait,
+	},
+};
+
+static const struct tegra_vde_soc tegra124_vde_soc = {
+	.supports_ref_pic_marking = true,
+	.coded_fmts = tegra124_coded_fmts,
+	.num_coded_fmts = ARRAY_SIZE(tegra124_coded_fmts),
+};
+
+static const struct tegra_vde_soc tegra114_vde_soc = {
+	.supports_ref_pic_marking = true,
+	.coded_fmts = tegra20_coded_fmts,
+	.num_coded_fmts = ARRAY_SIZE(tegra20_coded_fmts),
+};
+
+static const struct tegra_vde_soc tegra30_vde_soc = {
+	.supports_ref_pic_marking = false,
+	.coded_fmts = tegra20_coded_fmts,
+	.num_coded_fmts = ARRAY_SIZE(tegra20_coded_fmts),
+};
+
+static const struct tegra_vde_soc tegra20_vde_soc = {
+	.supports_ref_pic_marking = false,
+	.coded_fmts = tegra20_coded_fmts,
+	.num_coded_fmts = ARRAY_SIZE(tegra20_coded_fmts),
+};
+
+static const struct of_device_id tegra_vde_of_match[] = {
+	{ .compatible = "nvidia,tegra124-vde", .data = &tegra124_vde_soc },
+	{ .compatible = "nvidia,tegra114-vde", .data = &tegra114_vde_soc },
+	{ .compatible = "nvidia,tegra30-vde", .data = &tegra30_vde_soc },
+	{ .compatible = "nvidia,tegra20-vde", .data = &tegra20_vde_soc },
+	{ },
+};
+MODULE_DEVICE_TABLE(of, tegra_vde_of_match);
+
+static struct platform_driver tegra_vde_driver = {
+	.probe		= tegra_vde_probe,
+	.remove		= tegra_vde_remove,
+	.shutdown	= tegra_vde_shutdown,
+	.driver		= {
+		.name		= "tegra-vde",
+		.of_match_table = tegra_vde_of_match,
+		.pm		= &tegra_vde_pm_ops,
+	},
+};
+module_platform_driver(tegra_vde_driver);
+
+MODULE_DESCRIPTION("NVIDIA Tegra Video Decoder driver");
+MODULE_AUTHOR("Dmitry Osipenko <digetx@gmail.com>");
+MODULE_LICENSE("GPL");
diff --git a/drivers/media/platform/nvidia/tegra-vde/vde.h b/drivers/media/platform/nvidia/tegra-vde/vde.h
new file mode 100644
index 000000000..0fbb1f3d2
--- /dev/null
+++ b/drivers/media/platform/nvidia/tegra-vde/vde.h
@@ -0,0 +1,242 @@
+/* SPDX-License-Identifier: GPL-2.0+ */
+/*
+ * NVIDIA Tegra Video decoder driver
+ *
+ * Copyright (C) 2016-2019 GRATE-DRIVER project
+ */
+
+#ifndef TEGRA_VDE_H
+#define TEGRA_VDE_H
+
+#include <linux/completion.h>
+#include <linux/dma-direction.h>
+#include <linux/iova.h>
+#include <linux/list.h>
+#include <linux/mutex.h>
+#include <linux/types.h>
+#include <linux/workqueue.h>
+
+#include <media/media-device.h>
+#include <media/videobuf2-dma-contig.h>
+#include <media/videobuf2-dma-sg.h>
+#include <media/v4l2-ctrls.h>
+#include <media/v4l2-device.h>
+#include <media/v4l2-event.h>
+#include <media/v4l2-ioctl.h>
+#include <media/v4l2-mem2mem.h>
+
+#define ICMDQUE_WR		0x00
+#define CMDQUE_CONTROL		0x08
+#define INTR_STATUS		0x18
+#define BSE_INT_ENB		0x40
+#define BSE_CONFIG		0x44
+
+#define BSE_ICMDQUE_EMPTY	BIT(3)
+#define BSE_DMA_BUSY		BIT(23)
+
+#define BSEV_ALIGN		SZ_1
+#define FRAMEID_ALIGN		SZ_256
+#define SXE_BUFFER		SZ_32K
+#define VDE_ATOM		SZ_16
+
+struct clk;
+struct dma_buf;
+struct gen_pool;
+struct tegra_ctx;
+struct iommu_group;
+struct iommu_domain;
+struct reset_control;
+struct dma_buf_attachment;
+struct tegra_vde_h264_frame;
+struct tegra_vde_h264_decoder_ctx;
+
+struct tegra_video_frame {
+	struct dma_buf_attachment *y_dmabuf_attachment;
+	struct dma_buf_attachment *cb_dmabuf_attachment;
+	struct dma_buf_attachment *cr_dmabuf_attachment;
+	struct dma_buf_attachment *aux_dmabuf_attachment;
+	dma_addr_t y_addr;
+	dma_addr_t cb_addr;
+	dma_addr_t cr_addr;
+	dma_addr_t aux_addr;
+	u32 frame_num;
+	u32 flags;
+	u32 luma_atoms_pitch;
+	u32 chroma_atoms_pitch;
+};
+
+struct tegra_coded_fmt_desc {
+	u32 fourcc;
+	struct v4l2_frmsize_stepwise frmsize;
+	unsigned int num_decoded_fmts;
+	const u32 *decoded_fmts;
+	int (*decode_run)(struct tegra_ctx *ctx);
+	int (*decode_wait)(struct tegra_ctx *ctx);
+};
+
+struct tegra_vde_soc {
+	bool supports_ref_pic_marking;
+	const struct tegra_coded_fmt_desc *coded_fmts;
+	u32 num_coded_fmts;
+};
+
+struct tegra_vde_bo {
+	struct iova *iova;
+	struct sg_table sgt;
+	struct tegra_vde *vde;
+	enum dma_data_direction dma_dir;
+	unsigned long dma_attrs;
+	dma_addr_t dma_handle;
+	dma_addr_t dma_addr;
+	void *dma_cookie;
+	size_t size;
+};
+
+struct tegra_vde {
+	void __iomem *sxe;
+	void __iomem *bsev;
+	void __iomem *mbe;
+	void __iomem *ppe;
+	void __iomem *mce;
+	void __iomem *tfe;
+	void __iomem *ppb;
+	void __iomem *vdma;
+	void __iomem *frameid;
+	struct device *dev;
+	struct mutex lock;
+	struct mutex map_lock;
+	struct list_head map_list;
+	struct reset_control *rst;
+	struct reset_control *rst_mc;
+	struct gen_pool *iram_pool;
+	struct completion decode_completion;
+	struct clk *clk;
+	struct iommu_domain *domain;
+	struct iommu_group *group;
+	struct iova_domain iova;
+	struct iova *iova_resv_static_addresses;
+	struct iova *iova_resv_last_page;
+	const struct tegra_vde_soc *soc;
+	struct tegra_vde_bo *secure_bo;
+	dma_addr_t bitstream_data_addr;
+	dma_addr_t iram_lists_addr;
+	u32 *iram;
+	struct v4l2_device v4l2_dev;
+	struct v4l2_m2m_dev *m2m;
+	struct media_device mdev;
+	struct video_device vdev;
+	struct mutex v4l2_lock;
+	struct workqueue_struct *wq;
+	struct tegra_video_frame frames[V4L2_H264_NUM_DPB_ENTRIES + 1];
+};
+
+int tegra_vde_alloc_bo(struct tegra_vde *vde,
+		       struct tegra_vde_bo **ret_bo,
+		       enum dma_data_direction dma_dir,
+		       size_t size);
+void tegra_vde_free_bo(struct tegra_vde_bo *bo);
+
+struct tegra_ctx_h264 {
+	const struct v4l2_ctrl_h264_decode_params *decode_params;
+	const struct v4l2_ctrl_h264_sps *sps;
+	const struct v4l2_ctrl_h264_pps *pps;
+};
+
+struct tegra_ctx {
+	struct tegra_vde *vde;
+	struct tegra_ctx_h264 h264;
+	struct work_struct work;
+	struct v4l2_fh fh;
+	struct v4l2_ctrl_handler hdl;
+	struct v4l2_format coded_fmt;
+	struct v4l2_format decoded_fmt;
+	const struct tegra_coded_fmt_desc *coded_fmt_desc;
+	struct v4l2_ctrl *ctrls[];
+};
+
+struct tegra_m2m_buffer {
+	struct v4l2_m2m_buffer m2m;
+	struct dma_buf_attachment *a[VB2_MAX_PLANES];
+	dma_addr_t dma_base[VB2_MAX_PLANES];
+	dma_addr_t dma_addr[VB2_MAX_PLANES];
+	struct iova *iova[VB2_MAX_PLANES];
+	struct tegra_vde_bo *aux;
+	bool b_frame;
+};
+
+static inline struct tegra_m2m_buffer *
+vb_to_tegra_buf(struct vb2_buffer *vb)
+{
+	struct v4l2_m2m_buffer *m2m = container_of(vb, struct v4l2_m2m_buffer,
+						   vb.vb2_buf);
+
+	return container_of(m2m, struct tegra_m2m_buffer, m2m);
+}
+
+void tegra_vde_prepare_control_data(struct tegra_ctx *ctx, u32 id);
+
+void tegra_vde_writel(struct tegra_vde *vde, u32 value, void __iomem *base,
+		      u32 offset);
+u32 tegra_vde_readl(struct tegra_vde *vde, void __iomem *base, u32 offset);
+void tegra_vde_set_bits(struct tegra_vde *vde, u32 mask, void __iomem *base,
+			u32 offset);
+
+int tegra_vde_h264_decode_run(struct tegra_ctx *ctx);
+int tegra_vde_h264_decode_wait(struct tegra_ctx *ctx);
+
+int tegra_vde_iommu_init(struct tegra_vde *vde);
+void tegra_vde_iommu_deinit(struct tegra_vde *vde);
+int tegra_vde_iommu_map(struct tegra_vde *vde,
+			struct sg_table *sgt,
+			struct iova **iovap,
+			size_t size);
+void tegra_vde_iommu_unmap(struct tegra_vde *vde, struct iova *iova);
+
+int tegra_vde_dmabuf_cache_map(struct tegra_vde *vde,
+			       struct dma_buf *dmabuf,
+			       enum dma_data_direction dma_dir,
+			       struct dma_buf_attachment **ap,
+			       dma_addr_t *addrp);
+void tegra_vde_dmabuf_cache_unmap(struct tegra_vde *vde,
+				  struct dma_buf_attachment *a,
+				  bool release);
+void tegra_vde_dmabuf_cache_unmap_sync(struct tegra_vde *vde);
+void tegra_vde_dmabuf_cache_unmap_all(struct tegra_vde *vde);
+
+static __maybe_unused char const *
+tegra_vde_reg_base_name(struct tegra_vde *vde, void __iomem *base)
+{
+	if (vde->sxe == base)
+		return "SXE";
+
+	if (vde->bsev == base)
+		return "BSEV";
+
+	if (vde->mbe == base)
+		return "MBE";
+
+	if (vde->ppe == base)
+		return "PPE";
+
+	if (vde->mce == base)
+		return "MCE";
+
+	if (vde->tfe == base)
+		return "TFE";
+
+	if (vde->ppb == base)
+		return "PPB";
+
+	if (vde->vdma == base)
+		return "VDMA";
+
+	if (vde->frameid == base)
+		return "FRAMEID";
+
+	return "???";
+}
+
+int tegra_vde_v4l2_init(struct tegra_vde *vde);
+void tegra_vde_v4l2_deinit(struct tegra_vde *vde);
+
+#endif /* TEGRA_VDE_H */