aboutsummaryrefslogtreecommitdiffstats
path: root/common/recipes-kernel/linux/files/0026-drm-amdgpu-unify-AMDGPU_CTX_MAX_CS_PENDING-and-amdgp.patch
blob: da13089685896c45053a306775b80f1ec8557038 (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
From 12db7286ff75575c9cac9afc5309c26e8ae21527 Mon Sep 17 00:00:00 2001
From: Chunming Zhou <David1.Zhou@amd.com>
Date: Thu, 10 Dec 2015 15:45:11 +0800
Subject: [PATCH 0026/1110] drm/amdgpu: unify AMDGPU_CTX_MAX_CS_PENDING and
 amdgpu_sched_jobs
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

Signed-off-by: Chunming Zhou <David1.Zhou@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
---
 drivers/gpu/drm/amd/amdgpu/amdgpu.h     |  5 ++---
 drivers/gpu/drm/amd/amdgpu/amdgpu_ctx.c | 25 ++++++++++++++++++-------
 2 files changed, 20 insertions(+), 10 deletions(-)

diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu.h b/drivers/gpu/drm/amd/amdgpu/amdgpu.h
index e85ed1b..f6563fa 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu.h
@@ -1022,11 +1022,9 @@ int amdgpu_vm_free_job(struct amdgpu_job *job);
  * context related structures
  */
 
-#define AMDGPU_CTX_MAX_CS_PENDING	16
-
 struct amdgpu_ctx_ring {
 	uint64_t		sequence;
-	struct fence		*fences[AMDGPU_CTX_MAX_CS_PENDING];
+	struct fence		**fences;
 	struct amd_sched_entity	entity;
 };
 
@@ -1035,6 +1033,7 @@ struct amdgpu_ctx {
 	struct amdgpu_device    *adev;
 	unsigned		reset_counter;
 	spinlock_t		ring_lock;
+	struct fence            **fences;
 	struct amdgpu_ctx_ring	rings[AMDGPU_MAX_RINGS];
 };
 
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ctx.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ctx.c
index 15e3416..ee121ec 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ctx.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ctx.c
@@ -35,15 +35,24 @@ int amdgpu_ctx_init(struct amdgpu_device *adev, enum amd_sched_priority pri,
 	ctx->adev = adev;
 	kref_init(&ctx->refcount);
 	spin_lock_init(&ctx->ring_lock);
-	for (i = 0; i < AMDGPU_MAX_RINGS; ++i)
-		ctx->rings[i].sequence = 1;
+	ctx->fences = kzalloc(sizeof(struct fence *) * amdgpu_sched_jobs *
+			 AMDGPU_MAX_RINGS, GFP_KERNEL);
+	if (!ctx->fences)
+		return -ENOMEM;
 
+	for (i = 0; i < AMDGPU_MAX_RINGS; ++i) {
+		ctx->rings[i].sequence = 1;
+		ctx->rings[i].fences = (void *)ctx->fences + sizeof(struct fence *) *
+			amdgpu_sched_jobs * i;
+	}
 	if (amdgpu_enable_scheduler) {
 		/* create context entity for each ring */
 		for (i = 0; i < adev->num_rings; i++) {
 			struct amd_sched_rq *rq;
-			if (pri >= AMD_SCHED_MAX_PRIORITY)
+			if (pri >= AMD_SCHED_MAX_PRIORITY) {
+				kfree(ctx->fences);
 				return -EINVAL;
+			}
 			rq = &adev->rings[i]->sched.sched_rq[pri];
 			r = amd_sched_entity_init(&adev->rings[i]->sched,
 						  &ctx->rings[i].entity,
@@ -56,6 +65,7 @@ int amdgpu_ctx_init(struct amdgpu_device *adev, enum amd_sched_priority pri,
 			for (j = 0; j < i; j++)
 				amd_sched_entity_fini(&adev->rings[j]->sched,
 						      &ctx->rings[j].entity);
+			kfree(ctx->fences);
 			return r;
 		}
 	}
@@ -71,8 +81,9 @@ void amdgpu_ctx_fini(struct amdgpu_ctx *ctx)
 		return;
 
 	for (i = 0; i < AMDGPU_MAX_RINGS; ++i)
-		for (j = 0; j < AMDGPU_CTX_MAX_CS_PENDING; ++j)
+		for (j = 0; j < amdgpu_sched_jobs; ++j)
 			fence_put(ctx->rings[i].fences[j]);
+	kfree(ctx->fences);
 
 	if (amdgpu_enable_scheduler) {
 		for (i = 0; i < adev->num_rings; i++)
@@ -241,7 +252,7 @@ uint64_t amdgpu_ctx_add_fence(struct amdgpu_ctx *ctx, struct amdgpu_ring *ring,
 	unsigned idx = 0;
 	struct fence *other = NULL;
 
-	idx = seq % AMDGPU_CTX_MAX_CS_PENDING;
+	idx = seq % amdgpu_sched_jobs;
 	other = cring->fences[idx];
 	if (other) {
 		signed long r;
@@ -276,12 +287,12 @@ struct fence *amdgpu_ctx_get_fence(struct amdgpu_ctx *ctx,
 	}
 
 
-	if (seq + AMDGPU_CTX_MAX_CS_PENDING < cring->sequence) {
+	if (seq + amdgpu_sched_jobs < cring->sequence) {
 		spin_unlock(&ctx->ring_lock);
 		return NULL;
 	}
 
-	fence = fence_get(cring->fences[seq % AMDGPU_CTX_MAX_CS_PENDING]);
+	fence = fence_get(cring->fences[seq % amdgpu_sched_jobs]);
 	spin_unlock(&ctx->ring_lock);
 
 	return fence;
-- 
2.7.4