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
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
|
/*
* drivers/video/tegra/host/gk20a/ltc_common.c
*
* GK20A Graphics
*
* Copyright (c) 2011-2014, NVIDIA CORPORATION. All rights reserved.
*
* This program is free software; you can redistribute it and/or modify it
* under the terms and conditions of the GNU General Public License,
* version 2, as published by the Free Software Foundation.
*
* This program is distributed in the hope it will be useful, but WITHOUT
* ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
* FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for
* more details.
*
* You should have received a copy of the GNU General Public License
* along with this program. If not, see <http://www.gnu.org/licenses/>.
*/
#include <linux/dma-mapping.h>
#include <linux/delay.h>
#include "gk20a.h"
#include "gr_gk20a.h"
static int gk20a_determine_L2_size_bytes(struct gk20a *g)
{
const u32 gpuid = GK20A_GPUID(g->gpu_characteristics.arch,
g->gpu_characteristics.impl);
u32 lts_per_ltc;
u32 ways;
u32 sets;
u32 bytes_per_line;
u32 active_ltcs;
u32 cache_size;
u32 tmp;
u32 active_sets_value;
tmp = gk20a_readl(g, ltc_ltc0_lts0_tstg_cfg1_r());
ways = hweight32(ltc_ltc0_lts0_tstg_cfg1_active_ways_v(tmp));
active_sets_value = ltc_ltc0_lts0_tstg_cfg1_active_sets_v(tmp);
if (active_sets_value == ltc_ltc0_lts0_tstg_cfg1_active_sets_all_v()) {
sets = 64;
} else if (active_sets_value ==
ltc_ltc0_lts0_tstg_cfg1_active_sets_half_v()) {
sets = 32;
} else if (active_sets_value ==
ltc_ltc0_lts0_tstg_cfg1_active_sets_quarter_v()) {
sets = 16;
} else {
dev_err(dev_from_gk20a(g),
"Unknown constant %u for active sets",
(unsigned)active_sets_value);
sets = 0;
}
active_ltcs = g->gr.num_fbps;
/* chip-specific values */
switch (gpuid) {
case GK20A_GPUID_GK20A:
lts_per_ltc = 1;
bytes_per_line = 128;
break;
default:
dev_err(dev_from_gk20a(g), "Unknown GPU id 0x%02x\n",
(unsigned)gpuid);
lts_per_ltc = 0;
bytes_per_line = 0;
}
cache_size = active_ltcs * lts_per_ltc * ways * sets * bytes_per_line;
return cache_size;
}
/*
* Set the maximum number of ways that can have the "EVIST_LAST" class.
*/
static void gk20a_ltc_set_max_ways_evict_last(struct gk20a *g, u32 max_ways)
{
u32 mgmt_reg;
mgmt_reg = gk20a_readl(g, ltc_ltcs_ltss_tstg_set_mgmt_r()) &
~ltc_ltcs_ltss_tstg_set_mgmt_max_ways_evict_last_f(~0);
mgmt_reg |= ltc_ltcs_ltss_tstg_set_mgmt_max_ways_evict_last_f(max_ways);
gk20a_writel(g, ltc_ltcs_ltss_tstg_set_mgmt_r(), mgmt_reg);
}
/*
* Sets the ZBC color for the passed index.
*/
static void gk20a_ltc_set_zbc_color_entry(struct gk20a *g,
struct zbc_entry *color_val,
u32 index)
{
u32 i;
u32 real_index = index + GK20A_STARTOF_ZBC_TABLE;
gk20a_writel(g, ltc_ltcs_ltss_dstg_zbc_index_r(),
ltc_ltcs_ltss_dstg_zbc_index_address_f(real_index));
for (i = 0;
i < ltc_ltcs_ltss_dstg_zbc_color_clear_value__size_1_v(); i++)
gk20a_writel(g, ltc_ltcs_ltss_dstg_zbc_color_clear_value_r(i),
color_val->color_l2[i]);
}
/*
* Sets the ZBC depth for the passed index.
*/
static void gk20a_ltc_set_zbc_depth_entry(struct gk20a *g,
struct zbc_entry *depth_val,
u32 index)
{
u32 real_index = index + GK20A_STARTOF_ZBC_TABLE;
gk20a_writel(g, ltc_ltcs_ltss_dstg_zbc_index_r(),
ltc_ltcs_ltss_dstg_zbc_index_address_f(real_index));
gk20a_writel(g, ltc_ltcs_ltss_dstg_zbc_depth_clear_value_r(),
depth_val->depth);
}
/*
* Clear the L2 ZBC color table for the passed index.
*/
static void gk20a_ltc_clear_zbc_color_entry(struct gk20a *g, u32 index)
{
u32 i;
u32 real_index = index + GK20A_STARTOF_ZBC_TABLE;
gk20a_writel(g, ltc_ltcs_ltss_dstg_zbc_index_r(),
ltc_ltcs_ltss_dstg_zbc_index_address_f(real_index));
for (i = 0;
i < ltc_ltcs_ltss_dstg_zbc_color_clear_value__size_1_v(); i++)
gk20a_writel(g,
ltc_ltcs_ltss_dstg_zbc_color_clear_value_r(i), 0);
}
/*
* Clear the L2 ZBC depth entry for the passed index.
*/
static void gk20a_ltc_clear_zbc_depth_entry(struct gk20a *g, u32 index)
{
u32 real_index = index + GK20A_STARTOF_ZBC_TABLE;
gk20a_writel(g, ltc_ltcs_ltss_dstg_zbc_index_r(),
ltc_ltcs_ltss_dstg_zbc_index_address_f(real_index));
gk20a_writel(g, ltc_ltcs_ltss_dstg_zbc_depth_clear_value_r(), 0);
}
static int gk20a_ltc_init_zbc(struct gk20a *g, struct gr_gk20a *gr)
{
u32 i, j;
/* reset zbc clear */
for (i = 0; i < GK20A_SIZEOF_ZBC_TABLE -
GK20A_STARTOF_ZBC_TABLE; i++) {
gk20a_writel(g, ltc_ltcs_ltss_dstg_zbc_index_r(),
(gk20a_readl(g, ltc_ltcs_ltss_dstg_zbc_index_r()) &
~ltc_ltcs_ltss_dstg_zbc_index_address_f(~0)) |
ltc_ltcs_ltss_dstg_zbc_index_address_f(
i + GK20A_STARTOF_ZBC_TABLE));
for (j = 0; j < ltc_ltcs_ltss_dstg_zbc_color_clear_value__size_1_v(); j++)
gk20a_writel(g, ltc_ltcs_ltss_dstg_zbc_color_clear_value_r(j), 0);
gk20a_writel(g, ltc_ltcs_ltss_dstg_zbc_depth_clear_value_r(), 0);
}
gr_gk20a_clear_zbc_table(g, gr);
gr_gk20a_load_zbc_default_table(g, gr);
return 0;
}
static int gk20a_ltc_alloc_phys_cbc(struct gk20a *g,
size_t compbit_backing_size)
{
struct gr_gk20a *gr = &g->gr;
int order = ffs(compbit_backing_size >> PAGE_SHIFT);
struct page *pages;
struct sg_table *sgt;
int err = 0;
/* allocate few pages */
pages = alloc_pages(GFP_KERNEL, order);
if (!pages) {
gk20a_dbg(gpu_dbg_pte, "alloc_pages failed\n");
err = -ENOMEM;
goto err_alloc_pages;
}
/* clean up the pages */
memset(page_address(pages), 0, compbit_backing_size);
/* allocate room for placing the pages pointer.. */
gr->compbit_store.pages =
kzalloc(sizeof(*gr->compbit_store.pages), GFP_KERNEL);
if (!gr->compbit_store.pages) {
gk20a_dbg(gpu_dbg_pte, "failed to allocate pages struct");
err = -ENOMEM;
goto err_alloc_compbit_store;
}
err = gk20a_get_sgtable_from_pages(&g->dev->dev, &sgt, &pages, 0,
compbit_backing_size);
if (err) {
gk20a_dbg(gpu_dbg_pte, "could not get sg table for pages\n");
goto err_alloc_sg_table;
}
/* store the parameters to gr structure */
*gr->compbit_store.pages = pages;
gr->compbit_store.base_iova = sg_phys(sgt->sgl);
gr->compbit_store.size = compbit_backing_size;
gr->compbit_store.sgt = sgt;
return 0;
err_alloc_sg_table:
kfree(gr->compbit_store.pages);
gr->compbit_store.pages = NULL;
err_alloc_compbit_store:
__free_pages(pages, order);
err_alloc_pages:
return err;
}
static int gk20a_ltc_alloc_virt_cbc(struct gk20a *g,
size_t compbit_backing_size)
{
struct device *d = dev_from_gk20a(g);
struct gr_gk20a *gr = &g->gr;
DEFINE_DMA_ATTRS(attrs);
dma_addr_t iova;
int err;
dma_set_attr(DMA_ATTR_NO_KERNEL_MAPPING, &attrs);
gr->compbit_store.pages =
dma_alloc_attrs(d, compbit_backing_size, &iova,
GFP_KERNEL, &attrs);
if (!gr->compbit_store.pages) {
gk20a_err(dev_from_gk20a(g), "failed to allocate backing store for compbit : size %zu",
compbit_backing_size);
return -ENOMEM;
}
gr->compbit_store.base_iova = iova;
gr->compbit_store.size = compbit_backing_size;
err = gk20a_get_sgtable_from_pages(d,
&gr->compbit_store.sgt,
gr->compbit_store.pages, iova,
compbit_backing_size);
if (err) {
gk20a_err(dev_from_gk20a(g), "failed to allocate sgt for backing store");
return err;
}
return 0;
}
static void gk20a_ltc_init_cbc(struct gk20a *g, struct gr_gk20a *gr)
{
u32 max_size = gr->max_comptag_mem;
u32 max_comptag_lines = max_size << 3;
u32 compbit_base_post_divide;
u64 compbit_base_post_multiply64;
u64 compbit_store_base_iova;
u64 compbit_base_post_divide64;
if (IS_ENABLED(CONFIG_GK20A_PHYS_PAGE_TABLES))
compbit_store_base_iova = gr->compbit_store.base_iova;
else
compbit_store_base_iova = NV_MC_SMMU_VADDR_TRANSLATE(
gr->compbit_store.base_iova);
compbit_base_post_divide64 = compbit_store_base_iova >>
ltc_ltcs_ltss_cbc_base_alignment_shift_v();
do_div(compbit_base_post_divide64, gr->num_fbps);
compbit_base_post_divide = u64_lo32(compbit_base_post_divide64);
compbit_base_post_multiply64 = ((u64)compbit_base_post_divide *
gr->num_fbps) << ltc_ltcs_ltss_cbc_base_alignment_shift_v();
if (compbit_base_post_multiply64 < compbit_store_base_iova)
compbit_base_post_divide++;
gk20a_writel(g, ltc_ltcs_ltss_cbc_base_r(),
compbit_base_post_divide);
gk20a_dbg(gpu_dbg_info | gpu_dbg_map | gpu_dbg_pte,
"compbit base.pa: 0x%x,%08x cbc_base:0x%08x\n",
(u32)(compbit_store_base_iova >> 32),
(u32)(compbit_store_base_iova & 0xffffffff),
compbit_base_post_divide);
g->ops.ltc.cbc_ctrl(g, gk20a_cbc_op_invalidate,
0, max_comptag_lines - 1);
}
/* Flushes the compression bit cache as well as "data".
* Note: the name here is a bit of a misnomer. ELPG uses this
* internally... but ELPG doesn't have to be on to do it manually.
*/
static void gk20a_mm_g_elpg_flush_locked(struct gk20a *g)
{
u32 data;
s32 retry = 100;
gk20a_dbg_fn("");
/* Make sure all previous writes are committed to the L2. There's no
guarantee that writes are to DRAM. This will be a sysmembar internal
to the L2. */
gk20a_writel(g, ltc_ltcs_ltss_g_elpg_r(),
ltc_ltcs_ltss_g_elpg_flush_pending_f());
do {
data = gk20a_readl(g, ltc_ltc0_ltss_g_elpg_r());
if (ltc_ltc0_ltss_g_elpg_flush_v(data) ==
ltc_ltc0_ltss_g_elpg_flush_pending_v()) {
gk20a_dbg_info("g_elpg_flush 0x%x", data);
retry--;
usleep_range(20, 40);
} else
break;
} while (retry >= 0 || !tegra_platform_is_silicon());
if (retry < 0)
gk20a_warn(dev_from_gk20a(g),
"g_elpg_flush too many retries");
}
=======
>>>>>>> 7294a959b260... video: tegra: host: comptag init and clear
|