dri-devel.lists.freedesktop.org archive mirror
 help / color / mirror / Atom feed
* [PATCH 1/9] drm/amdgpu: generally allow over-commit during BO allocation
@ 2022-11-25 10:21 Christian König
  2022-11-25 10:21 ` [PATCH 2/9] drm/ttm: remove ttm_bo_(un)lock_delayed_workqueue Christian König
                   ` (10 more replies)
  0 siblings, 11 replies; 34+ messages in thread
From: Christian König @ 2022-11-25 10:21 UTC (permalink / raw)
  To: Intel Graphics Development, dri-devel, amd-gfx

We already fallback to a dummy BO with no backing store when we
allocate GDS,GWS and OA resources and to GTT when we allocate VRAM.

Drop all those workarounds and generalize this for GTT as well. This
fixes ENOMEM issues with runaway applications which try to allocate/free
GTT in a loop and are otherwise only limited by the CPU speed.

The CS will wait for the cleanup of freed up BOs to satisfy the
various domain specific limits and so effectively throttle those
buggy applications down to a sane allocation behavior again.

Signed-off-by: Christian König <christian.koenig@amd.com>
---
 drivers/gpu/drm/amd/amdgpu/amdgpu_gem.c    | 16 +++-------------
 drivers/gpu/drm/amd/amdgpu/amdgpu_object.c |  6 +-----
 2 files changed, 4 insertions(+), 18 deletions(-)

diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_gem.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_gem.c
index a0780a4e3e61..62e98f1ad770 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_gem.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_gem.c
@@ -113,7 +113,7 @@ int amdgpu_gem_object_create(struct amdgpu_device *adev, unsigned long size,
 	bp.resv = resv;
 	bp.preferred_domain = initial_domain;
 	bp.flags = flags;
-	bp.domain = initial_domain;
+	bp.domain = initial_domain | AMDGPU_GEM_DOMAIN_CPU;
 	bp.bo_ptr_size = sizeof(struct amdgpu_bo);
 
 	r = amdgpu_bo_create_user(adev, &bp, &ubo);
@@ -332,20 +332,10 @@ int amdgpu_gem_create_ioctl(struct drm_device *dev, void *data,
 	}
 
 	initial_domain = (u32)(0xffffffff & args->in.domains);
-retry:
 	r = amdgpu_gem_object_create(adev, size, args->in.alignment,
-				     initial_domain,
-				     flags, ttm_bo_type_device, resv, &gobj);
+				     initial_domain, flags, ttm_bo_type_device,
+				     resv, &gobj);
 	if (r && r != -ERESTARTSYS) {
-		if (flags & AMDGPU_GEM_CREATE_CPU_ACCESS_REQUIRED) {
-			flags &= ~AMDGPU_GEM_CREATE_CPU_ACCESS_REQUIRED;
-			goto retry;
-		}
-
-		if (initial_domain == AMDGPU_GEM_DOMAIN_VRAM) {
-			initial_domain |= AMDGPU_GEM_DOMAIN_GTT;
-			goto retry;
-		}
 		DRM_DEBUG("Failed to allocate GEM object (%llu, %d, %llu, %d)\n",
 				size, initial_domain, args->in.alignment, r);
 	}
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_object.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_object.c
index 974e85d8b6cc..919bbea2e3ac 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_object.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_object.c
@@ -581,11 +581,7 @@ int amdgpu_bo_create(struct amdgpu_device *adev,
 		bo->flags |= AMDGPU_GEM_CREATE_VRAM_WIPE_ON_RELEASE;
 
 	bo->tbo.bdev = &adev->mman.bdev;
-	if (bp->domain & (AMDGPU_GEM_DOMAIN_GWS | AMDGPU_GEM_DOMAIN_OA |
-			  AMDGPU_GEM_DOMAIN_GDS))
-		amdgpu_bo_placement_from_domain(bo, AMDGPU_GEM_DOMAIN_CPU);
-	else
-		amdgpu_bo_placement_from_domain(bo, bp->domain);
+	amdgpu_bo_placement_from_domain(bo, bp->domain);
 	if (bp->type == ttm_bo_type_kernel)
 		bo->tbo.priority = 1;
 
-- 
2.34.1


^ permalink raw reply related	[flat|nested] 34+ messages in thread

end of thread, other threads:[~2023-06-15 12:04 UTC | newest]

Thread overview: 34+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2022-11-25 10:21 [PATCH 1/9] drm/amdgpu: generally allow over-commit during BO allocation Christian König
2022-11-25 10:21 ` [PATCH 2/9] drm/ttm: remove ttm_bo_(un)lock_delayed_workqueue Christian König
2022-11-25 10:21 ` [PATCH 3/9] drm/ttm: use per BO cleanup workers Christian König
2022-11-29 21:14   ` Felix Kuehling
2022-12-05 13:39     ` Christian König
2023-06-13 13:05       ` Karol Herbst
2023-06-13 13:59         ` Christian König
2023-06-13 14:18           ` Karol Herbst
2023-06-15 11:19             ` Christian König
2023-06-15 12:04               ` Karol Herbst
2022-11-25 10:21 ` [PATCH 4/9] drm/ttm: merge ttm_bo_api.h and ttm_bo_driver.h Christian König
2022-11-25 12:43   ` kernel test robot
2022-11-25 21:19   ` kernel test robot
2022-11-25 10:21 ` [PATCH 5/9] drm/nouveau: stop using ttm_bo_wait Christian König
2022-11-25 10:21 ` [PATCH 6/9] drm/qxl: " Christian König
2022-12-15 14:19   ` Christian König
2022-12-15 20:09     ` Dave Airlie
2022-11-25 10:21 ` [PATCH 7/9] drm/i915: " Christian König
2022-11-25 11:14   ` [Intel-gfx] " Tvrtko Ursulin
2022-11-25 12:46     ` Christian König
2022-11-29 18:05     ` Matthew Auld
2022-11-30 13:02       ` Tvrtko Ursulin
2022-11-30 14:06         ` Daniel Vetter
2022-12-05 19:58           ` Christian König
2022-12-06 18:03             ` Matthew Auld
2022-12-06 18:06               ` Christian König
2022-11-25 10:21 ` [PATCH 8/9] drm/ttm: use ttm_bo_wait_ctx instead of ttm_bo_wait Christian König
2022-11-25 10:21 ` [PATCH 9/9] drm/ttm: move ttm_bo_wait into VMWGFX Christian König
2022-11-25 18:18 ` [PATCH 1/9] drm/amdgpu: generally allow over-commit during BO allocation Alex Deucher
2022-12-05 13:41   ` Christian König
2022-11-28  6:00 ` Arunpravin Paneer Selvam
2022-12-10  6:15 ` Felix Kuehling
2022-12-10 14:12   ` Christian König
2022-12-11  1:13     ` Felix Kuehling

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).