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
|
From 7281243966987a1febba2a13f836222577dbbf64 Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Christian=20K=C3=B6nig?= <christian.koenig@amd.com>
Date: Fri, 26 Jan 2018 15:00:43 +0100
Subject: [PATCH 3329/4131] drm/amdgpu: move waiting for VM flush into
gmc_v9_0_emit_flush_gpu_tlb
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit
Keep that at a common place instead of spread over all engines.
Signed-off-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Felix Kuehling <felix.kuehling@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Conflicts:
drivers/gpu/drm/amd/amdgpu/uvd_v7_0.c
drivers/gpu/drm/amd/amdgpu/vcn_v1_0.c
Change-Id: I464ca1ba93d0d24531771de6bed08a86bc69bc73
---
drivers/gpu/drm/amd/amdgpu/gfx_v9_0.c | 19 +++++++++----------
drivers/gpu/drm/amd/amdgpu/gmc_v9_0.c | 4 ++++
drivers/gpu/drm/amd/amdgpu/sdma_v4_0.c | 18 +++---------------
drivers/gpu/drm/amd/amdgpu/soc15.h | 3 ++-
drivers/gpu/drm/amd/amdgpu/uvd_v7_0.c | 20 ++++++--------------
drivers/gpu/drm/amd/amdgpu/vce_v4_0.c | 9 +++------
drivers/gpu/drm/amd/amdgpu/vcn_v1_0.c | 20 ++++++--------------
7 files changed, 33 insertions(+), 60 deletions(-)
diff --git a/drivers/gpu/drm/amd/amdgpu/gfx_v9_0.c b/drivers/gpu/drm/amd/amdgpu/gfx_v9_0.c
index 6bd55fb..05ccc94 100644
--- a/drivers/gpu/drm/amd/amdgpu/gfx_v9_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/gfx_v9_0.c
@@ -3693,15 +3693,8 @@ static void gfx_v9_0_ring_emit_vm_flush(struct amdgpu_ring *ring,
unsigned vmid, unsigned pasid,
uint64_t pd_addr)
{
- struct amdgpu_vmhub *hub = &ring->adev->vmhub[ring->funcs->vmhub];
- unsigned eng = ring->vm_inv_eng;
-
amdgpu_gmc_emit_flush_gpu_tlb(ring, vmid, pasid, pd_addr);
- /* wait for the invalidate to complete */
- gfx_v9_0_wait_reg_mem(ring, 0, 0, 0, hub->vm_inv_eng0_ack + eng,
- 0, 1 << vmid, 1 << vmid, 0x20);
-
/* compute doesn't have PFP */
if (ring->funcs->type == AMDGPU_RING_TYPE_GFX) {
/* sync PFP to ME, otherwise we might get invalid PFP reads */
@@ -4309,7 +4302,9 @@ static const struct amdgpu_ring_funcs gfx_v9_0_ring_funcs_gfx = {
.emit_frame_size = /* totally 242 maximum if 16 IBs */
5 + /* COND_EXEC */
7 + /* PIPELINE_SYNC */
- SOC15_FLUSH_GPU_TLB_NUM_WREG * 5 + 9 + /* VM_FLUSH */
+ SOC15_FLUSH_GPU_TLB_NUM_WREG * 5 +
+ SOC15_FLUSH_GPU_TLB_NUM_REG_WAIT * 7 +
+ 2 + /* VM_FLUSH */
8 + /* FENCE for VM_FLUSH */
20 + /* GDS switch */
4 + /* double SWITCH_BUFFER,
@@ -4359,7 +4354,9 @@ static const struct amdgpu_ring_funcs gfx_v9_0_ring_funcs_compute = {
7 + /* gfx_v9_0_ring_emit_hdp_flush */
5 + /* gfx_v9_0_ring_emit_hdp_invalidate */
7 + /* gfx_v9_0_ring_emit_pipeline_sync */
- SOC15_FLUSH_GPU_TLB_NUM_WREG * 5 + 9 + /* gfx_v9_0_ring_emit_vm_flush */
+ SOC15_FLUSH_GPU_TLB_NUM_WREG * 5 +
+ SOC15_FLUSH_GPU_TLB_NUM_REG_WAIT * 7 +
+ 2 + /* gfx_v9_0_ring_emit_vm_flush */
8 + 8 + 8, /* gfx_v9_0_ring_emit_fence x3 for user fence, vm fence */
.emit_ib_size = 4, /* gfx_v9_0_ring_emit_ib_compute */
.emit_ib = gfx_v9_0_ring_emit_ib_compute,
@@ -4392,7 +4389,9 @@ static const struct amdgpu_ring_funcs gfx_v9_0_ring_funcs_kiq = {
7 + /* gfx_v9_0_ring_emit_hdp_flush */
5 + /* gfx_v9_0_ring_emit_hdp_invalidate */
7 + /* gfx_v9_0_ring_emit_pipeline_sync */
- SOC15_FLUSH_GPU_TLB_NUM_WREG * 5 + 9 + /* gfx_v9_0_ring_emit_vm_flush */
+ SOC15_FLUSH_GPU_TLB_NUM_WREG * 5 +
+ SOC15_FLUSH_GPU_TLB_NUM_REG_WAIT * 7 +
+ 2 + /* gfx_v9_0_ring_emit_vm_flush */
8 + 8 + 8, /* gfx_v9_0_ring_emit_fence_kiq x3 for user fence, vm fence */
.emit_ib_size = 4, /* gfx_v9_0_ring_emit_ib_compute */
.emit_ib = gfx_v9_0_ring_emit_ib_compute,
diff --git a/drivers/gpu/drm/amd/amdgpu/gmc_v9_0.c b/drivers/gpu/drm/amd/amdgpu/gmc_v9_0.c
index 44eb22a..87e3883 100644
--- a/drivers/gpu/drm/amd/amdgpu/gmc_v9_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/gmc_v9_0.c
@@ -395,6 +395,10 @@ static uint64_t gmc_v9_0_emit_flush_gpu_tlb(struct amdgpu_ring *ring,
amdgpu_ring_emit_wreg(ring, hub->vm_inv_eng0_req + eng, req);
+ /* wait for the invalidate to complete */
+ amdgpu_ring_emit_reg_wait(ring, hub->vm_inv_eng0_ack + eng,
+ 1 << vmid, 1 << vmid);
+
return pd_addr;
}
diff --git a/drivers/gpu/drm/amd/amdgpu/sdma_v4_0.c b/drivers/gpu/drm/amd/amdgpu/sdma_v4_0.c
index 8fd08d6..4e4ca8f 100644
--- a/drivers/gpu/drm/amd/amdgpu/sdma_v4_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/sdma_v4_0.c
@@ -1136,21 +1136,7 @@ static void sdma_v4_0_ring_emit_vm_flush(struct amdgpu_ring *ring,
unsigned vmid, unsigned pasid,
uint64_t pd_addr)
{
- struct amdgpu_vmhub *hub = &ring->adev->vmhub[ring->funcs->vmhub];
- unsigned eng = ring->vm_inv_eng;
-
amdgpu_gmc_emit_flush_gpu_tlb(ring, vmid, pasid, pd_addr);
-
- /* wait for flush */
- amdgpu_ring_write(ring, SDMA_PKT_HEADER_OP(SDMA_OP_POLL_REGMEM) |
- SDMA_PKT_POLL_REGMEM_HEADER_HDP_FLUSH(0) |
- SDMA_PKT_POLL_REGMEM_HEADER_FUNC(3)); /* equal */
- amdgpu_ring_write(ring, (hub->vm_inv_eng0_ack + eng) << 2);
- amdgpu_ring_write(ring, 0);
- amdgpu_ring_write(ring, 1 << vmid); /* reference */
- amdgpu_ring_write(ring, 1 << vmid); /* mask */
- amdgpu_ring_write(ring, SDMA_PKT_POLL_REGMEM_DW5_RETRY_COUNT(0xfff) |
- SDMA_PKT_POLL_REGMEM_DW5_INTERVAL(10));
}
static void sdma_v4_0_ring_emit_wreg(struct amdgpu_ring *ring,
@@ -1599,7 +1585,9 @@ static const struct amdgpu_ring_funcs sdma_v4_0_ring_funcs = {
6 + /* sdma_v4_0_ring_emit_hdp_flush */
3 + /* sdma_v4_0_ring_emit_hdp_invalidate */
6 + /* sdma_v4_0_ring_emit_pipeline_sync */
- SOC15_FLUSH_GPU_TLB_NUM_WREG * 3 + 6 + /* sdma_v4_0_ring_emit_vm_flush */
+ /* sdma_v4_0_ring_emit_vm_flush */
+ SOC15_FLUSH_GPU_TLB_NUM_WREG * 3 +
+ SOC15_FLUSH_GPU_TLB_NUM_REG_WAIT * 6 +
10 + 10 + 10, /* sdma_v4_0_ring_emit_fence x3 for user fence, vm fence */
.emit_ib_size = 7 + 6, /* sdma_v4_0_ring_emit_ib */
.emit_ib = sdma_v4_0_ring_emit_ib,
diff --git a/drivers/gpu/drm/amd/amdgpu/soc15.h b/drivers/gpu/drm/amd/amdgpu/soc15.h
index 0f63ed6..f70da8a 100644
--- a/drivers/gpu/drm/amd/amdgpu/soc15.h
+++ b/drivers/gpu/drm/amd/amdgpu/soc15.h
@@ -27,7 +27,8 @@
#include "nbio_v6_1.h"
#include "nbio_v7_0.h"
-#define SOC15_FLUSH_GPU_TLB_NUM_WREG 4
+#define SOC15_FLUSH_GPU_TLB_NUM_WREG 4
+#define SOC15_FLUSH_GPU_TLB_NUM_REG_WAIT 1
extern const struct amd_ip_funcs soc15_common_ip_funcs;
diff --git a/drivers/gpu/drm/amd/amdgpu/uvd_v7_0.c b/drivers/gpu/drm/amd/amdgpu/uvd_v7_0.c
index 2c4ae6b..b031fb9 100644
--- a/drivers/gpu/drm/amd/amdgpu/uvd_v7_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/uvd_v7_0.c
@@ -1296,7 +1296,6 @@ static void uvd_v7_0_ring_emit_vm_flush(struct amdgpu_ring *ring,
uint64_t pd_addr)
{
struct amdgpu_vmhub *hub = &ring->adev->vmhub[ring->funcs->vmhub];
- unsigned eng = ring->vm_inv_eng;
uint32_t data0, data1, mask;
pd_addr = amdgpu_gmc_emit_flush_gpu_tlb(ring, vmid, pasid, pd_addr);
@@ -1306,12 +1305,6 @@ static void uvd_v7_0_ring_emit_vm_flush(struct amdgpu_ring *ring,
data1 = lower_32_bits(pd_addr);
mask = 0xffffffff;
uvd_v7_0_ring_emit_reg_wait(ring, data0, data1, mask);
-
- /* wait for flush */
- data0 = hub->vm_inv_eng0_ack + eng;
- data1 = 1 << vmid;
- mask = 1 << vmid;
- uvd_v7_0_ring_emit_reg_wait(ring, data0, data1, mask);
}
static void uvd_v7_0_ring_insert_nop(struct amdgpu_ring *ring, uint32_t count)
@@ -1344,17 +1337,12 @@ static void uvd_v7_0_enc_ring_emit_vm_flush(struct amdgpu_ring *ring,
uint64_t pd_addr)
{
struct amdgpu_vmhub *hub = &ring->adev->vmhub[ring->funcs->vmhub];
- unsigned eng = ring->vm_inv_eng;
pd_addr = amdgpu_gmc_emit_flush_gpu_tlb(ring, vmid, pasid, pd_addr);
/* wait for reg writes */
uvd_v7_0_enc_ring_emit_reg_wait(ring, hub->ctx0_ptb_addr_lo32 + vmid * 2,
lower_32_bits(pd_addr), 0xffffffff);
-
- /* wait for flush */
- uvd_v7_0_enc_ring_emit_reg_wait(ring, hub->vm_inv_eng0_ack + eng,
- 1 << vmid, 1 << vmid);
}
static void uvd_v7_0_enc_ring_emit_wreg(struct amdgpu_ring *ring,
@@ -1701,7 +1689,9 @@ static const struct amdgpu_ring_funcs uvd_v7_0_ring_vm_funcs = {
.emit_frame_size =
2 + /* uvd_v7_0_ring_emit_hdp_flush */
2 + /* uvd_v7_0_ring_emit_hdp_invalidate */
- SOC15_FLUSH_GPU_TLB_NUM_WREG * 6 + 16 + /* uvd_v7_0_ring_emit_vm_flush */
+ SOC15_FLUSH_GPU_TLB_NUM_WREG * 6 +
+ SOC15_FLUSH_GPU_TLB_NUM_REG_WAIT * 8 +
+ 8 + /* uvd_v7_0_ring_emit_vm_flush */
14 + 14, /* uvd_v7_0_ring_emit_fence x2 vm fence */
.emit_ib_size = 8, /* uvd_v7_0_ring_emit_ib */
.emit_ib = uvd_v7_0_ring_emit_ib,
@@ -1729,7 +1719,9 @@ static const struct amdgpu_ring_funcs uvd_v7_0_enc_ring_vm_funcs = {
.get_wptr = uvd_v7_0_enc_ring_get_wptr,
.set_wptr = uvd_v7_0_enc_ring_set_wptr,
.emit_frame_size =
- SOC15_FLUSH_GPU_TLB_NUM_WREG * 3 + 8 + /* uvd_v7_0_enc_ring_emit_vm_flush */
+ SOC15_FLUSH_GPU_TLB_NUM_WREG * 3 +
+ SOC15_FLUSH_GPU_TLB_NUM_REG_WAIT * 4 +
+ 4 + /* uvd_v7_0_enc_ring_emit_vm_flush */
5 + 5 + /* uvd_v7_0_enc_ring_emit_fence x2 vm fence */
1, /* uvd_v7_0_enc_ring_insert_end */
.emit_ib_size = 5, /* uvd_v7_0_enc_ring_emit_ib */
diff --git a/drivers/gpu/drm/amd/amdgpu/vce_v4_0.c b/drivers/gpu/drm/amd/amdgpu/vce_v4_0.c
index 2a4f73d..22c2067 100755
--- a/drivers/gpu/drm/amd/amdgpu/vce_v4_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/vce_v4_0.c
@@ -979,17 +979,12 @@ static void vce_v4_0_emit_vm_flush(struct amdgpu_ring *ring,
uint64_t pd_addr)
{
struct amdgpu_vmhub *hub = &ring->adev->vmhub[ring->funcs->vmhub];
- unsigned eng = ring->vm_inv_eng;
pd_addr = amdgpu_gmc_emit_flush_gpu_tlb(ring, vmid, pasid, pd_addr);
/* wait for reg writes */
vce_v4_0_emit_reg_wait(ring, hub->ctx0_ptb_addr_lo32 + vmid * 2,
lower_32_bits(pd_addr), 0xffffffff);
-
- /* wait for flush */
- vce_v4_0_emit_reg_wait(ring, hub->vm_inv_eng0_ack + eng,
- 1 << vmid, 1 << vmid);
}
static void vce_v4_0_emit_wreg(struct amdgpu_ring *ring,
@@ -1069,7 +1064,9 @@ static const struct amdgpu_ring_funcs vce_v4_0_ring_vm_funcs = {
.set_wptr = vce_v4_0_ring_set_wptr,
.parse_cs = amdgpu_vce_ring_parse_cs_vm,
.emit_frame_size =
- SOC15_FLUSH_GPU_TLB_NUM_WREG * 3 + 8 + /* vce_v4_0_emit_vm_flush */
+ SOC15_FLUSH_GPU_TLB_NUM_WREG * 3 +
+ SOC15_FLUSH_GPU_TLB_NUM_REG_WAIT * 4 +
+ 4 + /* vce_v4_0_emit_vm_flush */
5 + 5 + /* amdgpu_vce_ring_emit_fence x2 vm fence */
1, /* vce_v4_0_ring_insert_end */
.emit_ib_size = 5, /* vce_v4_0_ring_emit_ib */
diff --git a/drivers/gpu/drm/amd/amdgpu/vcn_v1_0.c b/drivers/gpu/drm/amd/amdgpu/vcn_v1_0.c
index 4fa86bd..3626321 100644
--- a/drivers/gpu/drm/amd/amdgpu/vcn_v1_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/vcn_v1_0.c
@@ -900,7 +900,6 @@ static void vcn_v1_0_dec_ring_emit_vm_flush(struct amdgpu_ring *ring,
uint64_t pd_addr)
{
struct amdgpu_vmhub *hub = &ring->adev->vmhub[ring->funcs->vmhub];
- unsigned eng = ring->vm_inv_eng;
uint32_t data0, data1, mask;
pd_addr = amdgpu_gmc_emit_flush_gpu_tlb(ring, vmid, pasid, pd_addr);
@@ -910,12 +909,6 @@ static void vcn_v1_0_dec_ring_emit_vm_flush(struct amdgpu_ring *ring,
data1 = lower_32_bits(pd_addr);
mask = 0xffffffff;
vcn_v1_0_dec_ring_emit_reg_wait(ring, data0, data1, mask);
-
- /* wait for flush */
- data0 = hub->vm_inv_eng0_ack + eng;
- data1 = 1 << vmid;
- mask = 1 << vmid;
- vcn_v1_0_dec_ring_emit_reg_wait(ring, data0, data1, mask);
}
static void vcn_v1_0_dec_ring_emit_wreg(struct amdgpu_ring *ring,
@@ -1045,17 +1038,12 @@ static void vcn_v1_0_enc_ring_emit_vm_flush(struct amdgpu_ring *ring,
uint64_t pd_addr)
{
struct amdgpu_vmhub *hub = &ring->adev->vmhub[ring->funcs->vmhub];
- unsigned eng = ring->vm_inv_eng;
pd_addr = amdgpu_gmc_emit_flush_gpu_tlb(ring, vmid, pasid, pd_addr);
/* wait for reg writes */
vcn_v1_0_enc_ring_emit_reg_wait(ring, hub->ctx0_ptb_addr_lo32 + vmid * 2,
lower_32_bits(pd_addr), 0xffffffff);
-
- /* wait for flush */
- vcn_v1_0_enc_ring_emit_reg_wait(ring, hub->vm_inv_eng0_ack + eng,
- 1 << vmid, 1 << vmid);
}
static void vcn_v1_0_enc_ring_emit_wreg(struct amdgpu_ring *ring,
@@ -1141,7 +1129,9 @@ static const struct amdgpu_ring_funcs vcn_v1_0_dec_ring_vm_funcs = {
.set_wptr = vcn_v1_0_dec_ring_set_wptr,
.emit_frame_size =
2 + /* vcn_v1_0_dec_ring_emit_hdp_invalidate */
- SOC15_FLUSH_GPU_TLB_NUM_WREG * 6 + 16 + /* vcn_v1_0_dec_ring_emit_vm_flush */
+ SOC15_FLUSH_GPU_TLB_NUM_WREG * 6 +
+ SOC15_FLUSH_GPU_TLB_NUM_REG_WAIT * 8 +
+ 8 + /* vcn_v1_0_dec_ring_emit_vm_flush */
14 + 14 + /* vcn_v1_0_dec_ring_emit_fence x2 vm fence */
6,
.emit_ib_size = 8, /* vcn_v1_0_dec_ring_emit_ib */
@@ -1171,7 +1161,9 @@ static const struct amdgpu_ring_funcs vcn_v1_0_enc_ring_vm_funcs = {
.get_wptr = vcn_v1_0_enc_ring_get_wptr,
.set_wptr = vcn_v1_0_enc_ring_set_wptr,
.emit_frame_size =
- SOC15_FLUSH_GPU_TLB_NUM_WREG * 3 + 8 + /* vcn_v1_0_enc_ring_emit_vm_flush */
+ SOC15_FLUSH_GPU_TLB_NUM_WREG * 3 +
+ SOC15_FLUSH_GPU_TLB_NUM_REG_WAIT * 4 +
+ 4 + /* vcn_v1_0_enc_ring_emit_vm_flush */
5 + 5 + /* vcn_v1_0_enc_ring_emit_fence x2 vm fence */
1, /* vcn_v1_0_enc_ring_insert_end */
.emit_ib_size = 5, /* vcn_v1_0_enc_ring_emit_ib */
--
2.7.4
|
