aboutsummaryrefslogtreecommitdiffstats
path: root/common/recipes-kernel/linux/files/0279-drm-amdgpu-cleanup-in-kernel-job-submission.patch
blob: 2bdc99f36e4990ac871ceb96126369409f229f7a (plain)
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
From 1eb0d8b2eedbd3da3fd627010743cab953537ef1 Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Christian=20K=C3=B6nig?= <christian.koenig@amd.com>
Date: Mon, 1 Feb 2016 12:20:25 +0100
Subject: [PATCH 0279/1110] drm/amdgpu: cleanup in kernel job submission
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

Add a job_alloc_with_ib helper and proper job submission.

Signed-off-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Alex Deucher <alexander.deucer@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Kalyan Alle <kalyan.alle@amd.com>
---
 drivers/gpu/drm/amd/amdgpu/amdgpu.h    | 16 ++----
 drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c | 90 ++++++++++------------------------
 2 files changed, 30 insertions(+), 76 deletions(-)

diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu.h b/drivers/gpu/drm/amd/amdgpu/amdgpu.h
index ea4e3aa..0e63bd3 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu.h
@@ -754,16 +754,12 @@ enum amdgpu_ring_type {
 };
 
 extern struct amd_sched_backend_ops amdgpu_sched_ops;
-int amdgpu_job_alloc(struct amdgpu_device *adev, unsigned num_ibs,
-                    struct amdgpu_job **job);
+int amdgpu_job_alloc_with_ib(struct amdgpu_device *adev, unsigned size,
+                             struct amdgpu_job **job);
+
 void amdgpu_job_free(struct amdgpu_job *job);
-int amdgpu_sched_ib_submit_kernel_helper(struct amdgpu_device *adev,
-					 struct amdgpu_ring *ring,
-					 struct amdgpu_ib *ibs,
-					 unsigned num_ibs,
-					 int (*free_job)(struct amdgpu_job *),
-					 void *owner,
-					 struct fence **fence);
+int amdgpu_job_submit(struct amdgpu_job *job, struct amdgpu_ring *ring,
+                      void *owner, struct fence **f);
 
 struct amdgpu_ring {
 	struct amdgpu_device		*adev;
@@ -954,7 +950,6 @@ int amdgpu_vm_bo_unmap(struct amdgpu_device *adev,
 		       uint64_t addr);
 void amdgpu_vm_bo_rmv(struct amdgpu_device *adev,
 		      struct amdgpu_bo_va *bo_va);
-int amdgpu_vm_free_job(struct amdgpu_job *job);
 
 /*
  * context related structures
@@ -1208,7 +1203,6 @@ struct amdgpu_job {
 	uint32_t		num_ibs;
 	void			*owner;
 	struct amdgpu_user_fence uf;
-	int (*free_job)(struct amdgpu_job *job);
 };
 #define to_amdgpu_job(sched_job)		\
 		container_of((sched_job), struct amdgpu_job, base)
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c
index ec9c967..e65f4a9 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c
@@ -320,15 +320,6 @@ static void amdgpu_vm_update_pages(struct amdgpu_device *adev,
 	}
 }
 
-int amdgpu_vm_free_job(struct amdgpu_job *job)
-{
-	int i;
-	for (i = 0; i < job->num_ibs; i++)
-		amdgpu_ib_free(job->adev, &job->ibs[i]);
-	kfree(job->ibs);
-	return 0;
-}
-
 /**
  * amdgpu_vm_clear_bo - initially clear the page dir/table
  *
@@ -342,7 +333,7 @@ static int amdgpu_vm_clear_bo(struct amdgpu_device *adev,
 {
 	struct amdgpu_ring *ring = adev->vm_manager.vm_pte_funcs_ring;
 	struct fence *fence = NULL;
-	struct amdgpu_ib *ib;
+        struct amdgpu_job *job;
 	unsigned entries;
 	uint64_t addr;
 	int r;
@@ -358,33 +349,24 @@ static int amdgpu_vm_clear_bo(struct amdgpu_device *adev,
 	addr = amdgpu_bo_gpu_offset(bo);
 	entries = amdgpu_bo_size(bo) / 8;
 
-	ib = kzalloc(sizeof(struct amdgpu_ib), GFP_KERNEL);
-	if (!ib)
+        r = amdgpu_job_alloc_with_ib(adev, 64, &job);
+        if (r)
 		goto error;
+        amdgpu_vm_update_pages(adev, NULL, 0, &job->ibs[0], addr, 0, entries,
+                                0, 0);
+        amdgpu_ring_pad_ib(ring, &job->ibs[0]);
 
-        r = amdgpu_ib_get(adev, NULL, 64, ib);
+        WARN_ON(job->ibs[0].length_dw > 64);
+        r = amdgpu_job_submit(job, ring, AMDGPU_FENCE_OWNER_VM, &fence);
 	if (r)
 		goto error_free;
 
-	ib->length_dw = 0;
-
-	amdgpu_vm_update_pages(adev, NULL, 0, ib, addr, 0, entries, 0, 0);
-        amdgpu_ring_pad_ib(ring, ib);
-
-	WARN_ON(ib->length_dw > 64);
-	r = amdgpu_sched_ib_submit_kernel_helper(adev, ring, ib, 1,
-						 &amdgpu_vm_free_job,
-						 AMDGPU_FENCE_OWNER_VM,
-						 &fence);
-	if (!r)
-		amdgpu_bo_fence(bo, fence, true);
+        amdgpu_bo_fence(bo, fence, true)
 	fence_put(fence);
 	return 0;
 
 error_free:
-	amdgpu_ib_free(adev, ib);
-	kfree(ib);
-
+        amdgpu_job_free(job);
 error:
 	return r;
 }
@@ -440,6 +422,7 @@ int amdgpu_vm_update_page_directory(struct amdgpu_device *adev,
 	uint32_t incr = AMDGPU_VM_PTE_COUNT * 8;
 	uint64_t last_pde = ~0, last_pt = ~0;
 	unsigned count = 0, pt_idx, ndw;
+        struct amdgpu_job *job;
 	struct amdgpu_ib *ib;
 	struct fence *fence = NULL;
 
@@ -450,18 +433,11 @@ int amdgpu_vm_update_page_directory(struct amdgpu_device *adev,
 
 	/* assume the worst case */
 	ndw += vm->max_pde_used * 6;
-
-	ib = kzalloc(sizeof(struct amdgpu_ib), GFP_KERNEL);
-	if (!ib)
-		return -ENOMEM;
-
-        r = amdgpu_ib_get(adev, NULL, ndw * 4, ib);
-	if (r) {
-		kfree(ib);
+        r = amdgpu_job_alloc_with_ib(adev, ndw * 4, &job);
+        if (r)
 		return r;
-	}
-	ib->length_dw = 0;
 
+        ib = &job->ibs[0];
 	/* walk over the address space and update the page directory */
 	for (pt_idx = 0; pt_idx <= vm->max_pde_used; ++pt_idx) {
                 struct amdgpu_bo *bo = vm->page_tables[pt_idx].entry.robj;
@@ -502,10 +478,7 @@ int amdgpu_vm_update_page_directory(struct amdgpu_device *adev,
                 amdgpu_ring_pad_ib(ring, ib);
 		amdgpu_sync_resv(adev, &ib->sync, pd->tbo.resv, AMDGPU_FENCE_OWNER_VM);
 		WARN_ON(ib->length_dw > ndw);
-		r = amdgpu_sched_ib_submit_kernel_helper(adev, ring, ib, 1,
-							 &amdgpu_vm_free_job,
-							 AMDGPU_FENCE_OWNER_VM,
-							 &fence);
+                r = amdgpu_job_submit(job, ring, AMDGPU_FENCE_OWNER_VM, &fence);
 		if (r)
 			goto error_free;
 
@@ -513,18 +486,13 @@ int amdgpu_vm_update_page_directory(struct amdgpu_device *adev,
 		fence_put(vm->page_directory_fence);
 		vm->page_directory_fence = fence_get(fence);
 		fence_put(fence);
-	}
-
-	if (ib->length_dw == 0) {
-		amdgpu_ib_free(adev, ib);
-		kfree(ib);
-	}
+        } else {
+                amdgpu_job_free(job);
 
 	return 0;
 
 error_free:
-	amdgpu_ib_free(adev, ib);
-	kfree(ib);
+        amdgpu_job_free(job);
 	return r;
 }
 
@@ -707,6 +675,7 @@ static int amdgpu_vm_bo_update_mapping(struct amdgpu_device *adev,
 	struct amdgpu_ring *ring = adev->vm_manager.vm_pte_funcs_ring;
         void *owner = AMDGPU_FENCE_OWNER_VM;
 	unsigned nptes, ncmds, ndw;
+        struct amdgpu_job *job;
 	struct amdgpu_ib *ib;
 	struct fence *f = NULL;
 	int r;
@@ -744,18 +713,13 @@ static int amdgpu_vm_bo_update_mapping(struct amdgpu_device *adev,
 		/* two extra commands for begin/end of fragment */
 		ndw += 2 * 10;
 	}
+        r = amdgpu_job_alloc_with_ib(adev, ndw * 4, &job);
+        if (r)
 
-	ib = kzalloc(sizeof(struct amdgpu_ib), GFP_KERNEL);
-	if (!ib)
-		return -ENOMEM;
-
-        r = amdgpu_ib_get(adev, NULL, ndw * 4, ib);
-	if (r) {
-		kfree(ib);
 		return r;
-	}
 
-	ib->length_dw = 0;
+        ib = &job->ibs[0];
+
         r = amdgpu_sync_resv(adev, &ib->sync, vm->page_directory->tbo.resv,
                              owner);
         if (r)
@@ -770,10 +734,7 @@ static int amdgpu_vm_bo_update_mapping(struct amdgpu_device *adev,
 
         amdgpu_ring_pad_ib(ring, ib);
 	WARN_ON(ib->length_dw > ndw);
-	r = amdgpu_sched_ib_submit_kernel_helper(adev, ring, ib, 1,
-						 &amdgpu_vm_free_job,
-						 AMDGPU_FENCE_OWNER_VM,
-						 &f);
+        r = amdgpu_job_submit(job, ring, AMDGPU_FENCE_OWNER_VM, &f);
 	if (r)
 		goto error_free;
 
@@ -786,8 +747,7 @@ static int amdgpu_vm_bo_update_mapping(struct amdgpu_device *adev,
 	return 0;
 
 error_free:
-	amdgpu_ib_free(adev, ib);
-	kfree(ib);
+        amdgpu_job_free(job);
 	return r;
 }
 
-- 
2.7.4