1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
|
// SPDX-License-Identifier: GPL-2.0-only
/*
* Scaler library
*
* Copyright (c) 2013 Texas Instruments Inc.
*
* David Griego, <dagriego@biglakesoftware.com>
* Dale Farnsworth, <dale@farnsworth.org>
* Archit Taneja, <archit@ti.com>
*/
#include <linux/err.h>
#include <linux/io.h>
#include <linux/module.h>
#include <linux/platform_device.h>
#include <linux/slab.h>
#include "sc.h"
#include "sc_coeff.h"
void sc_dump_regs(struct sc_data *sc)
{
struct device *dev = &sc->pdev->dev;
#define DUMPREG(r) dev_dbg(dev, "%-35s %08x\n", #r, \
ioread32(sc->base + CFG_##r))
dev_dbg(dev, "SC Registers @ %pa:\n", &sc->res->start);
DUMPREG(SC0);
DUMPREG(SC1);
DUMPREG(SC2);
DUMPREG(SC3);
DUMPREG(SC4);
DUMPREG(SC5);
DUMPREG(SC6);
DUMPREG(SC8);
DUMPREG(SC9);
DUMPREG(SC10);
DUMPREG(SC11);
DUMPREG(SC12);
DUMPREG(SC13);
DUMPREG(SC17);
DUMPREG(SC18);
DUMPREG(SC19);
DUMPREG(SC20);
DUMPREG(SC21);
DUMPREG(SC22);
DUMPREG(SC23);
DUMPREG(SC24);
DUMPREG(SC25);
#undef DUMPREG
}
EXPORT_SYMBOL(sc_dump_regs);
/*
* set the horizontal scaler coefficients according to the ratio of output to
* input widths, after accounting for up to two levels of decimation
*/
void sc_set_hs_coeffs(struct sc_data *sc, void *addr, unsigned int src_w,
unsigned int dst_w)
{
int sixteenths;
int idx;
int i, j;
u16 *coeff_h = addr;
const u16 *cp;
if (dst_w > src_w) {
idx = HS_UP_SCALE;
} else {
if ((dst_w << 1) < src_w)
dst_w <<= 1; /* first level decimation */
if ((dst_w << 1) < src_w)
dst_w <<= 1; /* second level decimation */
if (dst_w == src_w) {
idx = HS_LE_16_16_SCALE;
} else {
sixteenths = (dst_w << 4) / src_w;
if (sixteenths < 8)
sixteenths = 8;
idx = HS_LT_9_16_SCALE + sixteenths - 8;
}
}
cp = scaler_hs_coeffs[idx];
for (i = 0; i < SC_NUM_PHASES * 2; i++) {
for (j = 0; j < SC_H_NUM_TAPS; j++)
*coeff_h++ = *cp++;
/*
* for each phase, the scaler expects space for 8 coefficients
* in it's memory. For the horizontal scaler, we copy the first
* 7 coefficients and skip the last slot to move to the next
* row to hold coefficients for the next phase
*/
coeff_h += SC_NUM_TAPS_MEM_ALIGN - SC_H_NUM_TAPS;
}
sc->load_coeff_h = true;
}
EXPORT_SYMBOL(sc_set_hs_coeffs);
/*
* set the vertical scaler coefficients according to the ratio of output to
* input heights
*/
void sc_set_vs_coeffs(struct sc_data *sc, void *addr, unsigned int src_h,
unsigned int dst_h)
{
int sixteenths;
int idx;
int i, j;
u16 *coeff_v = addr;
const u16 *cp;
if (dst_h > src_h) {
idx = VS_UP_SCALE;
} else if (dst_h == src_h) {
idx = VS_1_TO_1_SCALE;
} else {
sixteenths = (dst_h << 4) / src_h;
if (sixteenths < 8)
sixteenths = 8;
idx = VS_LT_9_16_SCALE + sixteenths - 8;
}
cp = scaler_vs_coeffs[idx];
for (i = 0; i < SC_NUM_PHASES * 2; i++) {
for (j = 0; j < SC_V_NUM_TAPS; j++)
*coeff_v++ = *cp++;
/*
* for the vertical scaler, we copy the first 5 coefficients and
* skip the last 3 slots to move to the next row to hold
* coefficients for the next phase
*/
coeff_v += SC_NUM_TAPS_MEM_ALIGN - SC_V_NUM_TAPS;
}
sc->load_coeff_v = true;
}
EXPORT_SYMBOL(sc_set_vs_coeffs);
void sc_config_scaler(struct sc_data *sc, u32 *sc_reg0, u32 *sc_reg8,
u32 *sc_reg17, unsigned int src_w, unsigned int src_h,
unsigned int dst_w, unsigned int dst_h)
{
struct device *dev = &sc->pdev->dev;
u32 val;
int dcm_x, dcm_shift;
bool use_rav;
unsigned long lltmp;
u32 lin_acc_inc, lin_acc_inc_u;
u32 col_acc_offset;
u16 factor = 0;
int row_acc_init_rav = 0, row_acc_init_rav_b = 0;
u32 row_acc_inc = 0, row_acc_offset = 0, row_acc_offset_b = 0;
/*
* location of SC register in payload memory with respect to the first
* register in the mmr address data block
*/
u32 *sc_reg9 = sc_reg8 + 1;
u32 *sc_reg12 = sc_reg8 + 4;
u32 *sc_reg13 = sc_reg8 + 5;
u32 *sc_reg24 = sc_reg17 + 7;
val = sc_reg0[0];
/* clear all the features(they may get enabled elsewhere later) */
val &= ~(CFG_SELFGEN_FID | CFG_TRIM | CFG_ENABLE_SIN2_VER_INTP |
CFG_INTERLACE_I | CFG_DCM_4X | CFG_DCM_2X | CFG_AUTO_HS |
CFG_ENABLE_EV | CFG_USE_RAV | CFG_INVT_FID | CFG_SC_BYPASS |
CFG_INTERLACE_O | CFG_Y_PK_EN | CFG_HP_BYPASS | CFG_LINEAR);
if (src_w == dst_w && src_h == dst_h) {
val |= CFG_SC_BYPASS;
sc_reg0[0] = val;
return;
}
/* we only support linear scaling for now */
val |= CFG_LINEAR;
/* configure horizontal scaler */
/* enable 2X or 4X decimation */
dcm_x = src_w / dst_w;
if (dcm_x > 4) {
val |= CFG_DCM_4X;
dcm_shift = 2;
} else if (dcm_x > 2) {
val |= CFG_DCM_2X;
dcm_shift = 1;
} else {
dcm_shift = 0;
}
lltmp = dst_w - 1;
lin_acc_inc = div64_u64(((u64)(src_w >> dcm_shift) - 1) << 24, lltmp);
lin_acc_inc_u = 0;
col_acc_offset = 0;
dev_dbg(dev, "hs config: src_w = %d, dst_w = %d, decimation = %s, lin_acc_inc = %08x\n",
src_w, dst_w, dcm_shift == 2 ? "4x" :
(dcm_shift == 1 ? "2x" : "none"), lin_acc_inc);
/* configure vertical scaler */
/* use RAV for vertical scaler if vertical downscaling is > 4x */
if (dst_h < (src_h >> 2)) {
use_rav = true;
val |= CFG_USE_RAV;
} else {
use_rav = false;
}
if (use_rav) {
/* use RAV */
factor = (u16) ((dst_h << 10) / src_h);
row_acc_init_rav = factor + ((1 + factor) >> 1);
if (row_acc_init_rav >= 1024)
row_acc_init_rav -= 1024;
row_acc_init_rav_b = row_acc_init_rav +
(1 + (row_acc_init_rav >> 1)) -
(1024 >> 1);
if (row_acc_init_rav_b < 0) {
row_acc_init_rav_b += row_acc_init_rav;
row_acc_init_rav *= 2;
}
dev_dbg(dev, "vs config(RAV): src_h = %d, dst_h = %d, factor = %d, acc_init = %08x, acc_init_b = %08x\n",
src_h, dst_h, factor, row_acc_init_rav,
row_acc_init_rav_b);
} else {
/* use polyphase */
row_acc_inc = ((src_h - 1) << 16) / (dst_h - 1);
row_acc_offset = 0;
row_acc_offset_b = 0;
dev_dbg(dev, "vs config(POLY): src_h = %d, dst_h = %d,row_acc_inc = %08x\n",
src_h, dst_h, row_acc_inc);
}
sc_reg0[0] = val;
sc_reg0[1] = row_acc_inc;
sc_reg0[2] = row_acc_offset;
sc_reg0[3] = row_acc_offset_b;
sc_reg0[4] = ((lin_acc_inc_u & CFG_LIN_ACC_INC_U_MASK) <<
CFG_LIN_ACC_INC_U_SHIFT) | (dst_w << CFG_TAR_W_SHIFT) |
(dst_h << CFG_TAR_H_SHIFT);
sc_reg0[5] = (src_w << CFG_SRC_W_SHIFT) | (src_h << CFG_SRC_H_SHIFT);
sc_reg0[6] = (row_acc_init_rav_b << CFG_ROW_ACC_INIT_RAV_B_SHIFT) |
(row_acc_init_rav << CFG_ROW_ACC_INIT_RAV_SHIFT);
*sc_reg9 = lin_acc_inc;
*sc_reg12 = col_acc_offset << CFG_COL_ACC_OFFSET_SHIFT;
*sc_reg13 = factor;
*sc_reg24 = (src_w << CFG_ORG_W_SHIFT) | (src_h << CFG_ORG_H_SHIFT);
}
EXPORT_SYMBOL(sc_config_scaler);
struct sc_data *sc_create(struct platform_device *pdev, const char *res_name)
{
struct sc_data *sc;
dev_dbg(&pdev->dev, "sc_create\n");
sc = devm_kzalloc(&pdev->dev, sizeof(*sc), GFP_KERNEL);
if (!sc) {
dev_err(&pdev->dev, "couldn't alloc sc_data\n");
return ERR_PTR(-ENOMEM);
}
sc->pdev = pdev;
sc->res = platform_get_resource_byname(pdev, IORESOURCE_MEM, res_name);
if (!sc->res) {
dev_err(&pdev->dev, "missing '%s' platform resources data\n",
res_name);
return ERR_PTR(-ENODEV);
}
sc->base = devm_ioremap_resource(&pdev->dev, sc->res);
if (IS_ERR(sc->base))
return ERR_CAST(sc->base);
return sc;
}
EXPORT_SYMBOL(sc_create);
MODULE_DESCRIPTION("TI VIP/VPE Scaler");
MODULE_AUTHOR("Texas Instruments Inc.");
MODULE_LICENSE("GPL v2");
|
