* [PATCH] drm/amd/amdgpu: Bail out of BO node creation if not enough VRAM (v2)
@ 2019-06-11 16:54 StDenis, Tom
[not found] ` <20190611165444.17841-1-tom.stdenis-5C7GfCeVMHo@public.gmane.org>
0 siblings, 1 reply; 4+ messages in thread
From: StDenis, Tom @ 2019-06-11 16:54 UTC (permalink / raw)
To: amd-gfx-PD4FTy7X32lNgt0PjOBp9y5qC8QIuHrW@public.gmane.org; +Cc: StDenis, Tom
(v2): Return 0 and set mem->mm_node to NULL.
Signed-off-by: Tom St Denis <tom.stdenis@amd.com>
---
drivers/gpu/drm/amd/amdgpu/amdgpu_vram_mgr.c | 18 +++++++++++++-----
1 file changed, 13 insertions(+), 5 deletions(-)
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_vram_mgr.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_vram_mgr.c
index 8aea2f21b202..c2730e5c1081 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_vram_mgr.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_vram_mgr.c
@@ -276,7 +276,7 @@ static int amdgpu_vram_mgr_new(struct ttm_mem_type_manager *man,
struct drm_mm_node *nodes;
enum drm_mm_insert_mode mode;
unsigned long lpfn, num_nodes, pages_per_node, pages_left;
- uint64_t usage = 0, vis_usage = 0;
+ uint64_t vis_usage = 0;
unsigned i;
int r;
@@ -284,6 +284,14 @@ static int amdgpu_vram_mgr_new(struct ttm_mem_type_manager *man,
if (!lpfn)
lpfn = man->size;
+ /* bail out quickly if there's likely not enough VRAM for this BO */
+ atomic64_add(mem->num_pages << PAGE_SHIFT, &mgr->usage);
+ if (atomic64_read(&mgr->usage) > adev->gmc.mc_vram_size) {
+ atomic64_sub(mem->num_pages << PAGE_SHIFT, &mgr->usage);
+ mem->mm_node = NULL;
+ return 0;
+ }
+
if (place->flags & TTM_PL_FLAG_CONTIGUOUS) {
pages_per_node = ~0ul;
num_nodes = 1;
@@ -300,8 +308,10 @@ static int amdgpu_vram_mgr_new(struct ttm_mem_type_manager *man,
nodes = kvmalloc_array((uint32_t)num_nodes, sizeof(*nodes),
GFP_KERNEL | __GFP_ZERO);
- if (!nodes)
+ if (!nodes) {
+ atomic64_sub(mem->num_pages << PAGE_SHIFT, &mgr->usage);
return -ENOMEM;
+ }
mode = DRM_MM_INSERT_BEST;
if (place->flags & TTM_PL_FLAG_TOPDOWN)
@@ -321,7 +331,6 @@ static int amdgpu_vram_mgr_new(struct ttm_mem_type_manager *man,
if (unlikely(r))
break;
- usage += nodes[i].size << PAGE_SHIFT;
vis_usage += amdgpu_vram_mgr_vis_size(adev, &nodes[i]);
amdgpu_vram_mgr_virt_start(mem, &nodes[i]);
pages_left -= pages;
@@ -341,14 +350,12 @@ static int amdgpu_vram_mgr_new(struct ttm_mem_type_manager *man,
if (unlikely(r))
goto error;
- usage += nodes[i].size << PAGE_SHIFT;
vis_usage += amdgpu_vram_mgr_vis_size(adev, &nodes[i]);
amdgpu_vram_mgr_virt_start(mem, &nodes[i]);
pages_left -= pages;
}
spin_unlock(&mgr->lock);
- atomic64_add(usage, &mgr->usage);
atomic64_add(vis_usage, &mgr->vis_usage);
mem->mm_node = nodes;
@@ -359,6 +366,7 @@ static int amdgpu_vram_mgr_new(struct ttm_mem_type_manager *man,
while (i--)
drm_mm_remove_node(&nodes[i]);
spin_unlock(&mgr->lock);
+ atomic64_sub(mem->num_pages << PAGE_SHIFT, &mgr->usage);
kvfree(nodes);
return r == -ENOSPC ? 0 : r;
--
2.21.0
_______________________________________________
amd-gfx mailing list
amd-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/amd-gfx
^ permalink raw reply related [flat|nested] 4+ messages in thread[parent not found: <20190611165444.17841-1-tom.stdenis-5C7GfCeVMHo@public.gmane.org>]
* Re: [PATCH] drm/amd/amdgpu: Bail out of BO node creation if not enough VRAM (v2) [not found] ` <20190611165444.17841-1-tom.stdenis-5C7GfCeVMHo@public.gmane.org> @ 2019-06-12 7:26 ` Michel Dänzer [not found] ` <fd4c5152-2670-8ac5-e6cc-fff6f260f3ed-otUistvHUpPR7s880joybQ@public.gmane.org> 2019-06-12 8:00 ` Christian König 1 sibling, 1 reply; 4+ messages in thread From: Michel Dänzer @ 2019-06-12 7:26 UTC (permalink / raw) To: StDenis, Tom; +Cc: amd-gfx-PD4FTy7X32lNgt0PjOBp9y5qC8QIuHrW@public.gmane.org On 2019-06-11 6:54 p.m., StDenis, Tom wrote: > (v2): Return 0 and set mem->mm_node to NULL. > > Signed-off-by: Tom St Denis <tom.stdenis@amd.com> > --- > drivers/gpu/drm/amd/amdgpu/amdgpu_vram_mgr.c | 18 +++++++++++++----- > 1 file changed, 13 insertions(+), 5 deletions(-) > > [...] > > @@ -284,6 +284,14 @@ static int amdgpu_vram_mgr_new(struct ttm_mem_type_manager *man, > if (!lpfn) > lpfn = man->size; > > + /* bail out quickly if there's likely not enough VRAM for this BO */ > + atomic64_add(mem->num_pages << PAGE_SHIFT, &mgr->usage); > + if (atomic64_read(&mgr->usage) > adev->gmc.mc_vram_size) { Should probably use atomic64_add_return instead of atomic64_add + atomic64_read. Also, AFAICT this doesn't allow any VRAM overcommit, which seems a bit conservative? -- Earthling Michel Dänzer | https://www.amd.com Libre software enthusiast | Mesa and X developer _______________________________________________ amd-gfx mailing list amd-gfx@lists.freedesktop.org https://lists.freedesktop.org/mailman/listinfo/amd-gfx ^ permalink raw reply [flat|nested] 4+ messages in thread
[parent not found: <fd4c5152-2670-8ac5-e6cc-fff6f260f3ed-otUistvHUpPR7s880joybQ@public.gmane.org>]
* Re: [PATCH] drm/amd/amdgpu: Bail out of BO node creation if not enough VRAM (v2) [not found] ` <fd4c5152-2670-8ac5-e6cc-fff6f260f3ed-otUistvHUpPR7s880joybQ@public.gmane.org> @ 2019-06-12 8:03 ` Christian König 0 siblings, 0 replies; 4+ messages in thread From: Christian König @ 2019-06-12 8:03 UTC (permalink / raw) To: Michel Dänzer, StDenis, Tom Cc: amd-gfx-PD4FTy7X32lNgt0PjOBp9y5qC8QIuHrW@public.gmane.org Am 12.06.19 um 09:26 schrieb Michel Dänzer: > On 2019-06-11 6:54 p.m., StDenis, Tom wrote: >> (v2): Return 0 and set mem->mm_node to NULL. >> >> Signed-off-by: Tom St Denis <tom.stdenis@amd.com> >> --- >> drivers/gpu/drm/amd/amdgpu/amdgpu_vram_mgr.c | 18 +++++++++++++----- >> 1 file changed, 13 insertions(+), 5 deletions(-) >> >> [...] >> >> @@ -284,6 +284,14 @@ static int amdgpu_vram_mgr_new(struct ttm_mem_type_manager *man, >> if (!lpfn) >> lpfn = man->size; >> >> + /* bail out quickly if there's likely not enough VRAM for this BO */ >> + atomic64_add(mem->num_pages << PAGE_SHIFT, &mgr->usage); >> + if (atomic64_read(&mgr->usage) > adev->gmc.mc_vram_size) { > Should probably use atomic64_add_return instead of atomic64_add + > atomic64_read. > > Also, AFAICT this doesn't allow any VRAM overcommit, which seems a bit > conservative? This is the low level VRAM manager. Here we can't overcomit anyway and need to return early for TTM to evict things. Christian. _______________________________________________ amd-gfx mailing list amd-gfx@lists.freedesktop.org https://lists.freedesktop.org/mailman/listinfo/amd-gfx ^ permalink raw reply [flat|nested] 4+ messages in thread
* Re: [PATCH] drm/amd/amdgpu: Bail out of BO node creation if not enough VRAM (v2) [not found] ` <20190611165444.17841-1-tom.stdenis-5C7GfCeVMHo@public.gmane.org> 2019-06-12 7:26 ` Michel Dänzer @ 2019-06-12 8:00 ` Christian König 1 sibling, 0 replies; 4+ messages in thread From: Christian König @ 2019-06-12 8:00 UTC (permalink / raw) To: StDenis, Tom, amd-gfx-PD4FTy7X32lNgt0PjOBp9y5qC8QIuHrW@public.gmane.org Am 11.06.19 um 18:54 schrieb StDenis, Tom: > (v2): Return 0 and set mem->mm_node to NULL. > > Signed-off-by: Tom St Denis <tom.stdenis@amd.com> > --- > drivers/gpu/drm/amd/amdgpu/amdgpu_vram_mgr.c | 18 +++++++++++++----- > 1 file changed, 13 insertions(+), 5 deletions(-) > > diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_vram_mgr.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_vram_mgr.c > index 8aea2f21b202..c2730e5c1081 100644 > --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_vram_mgr.c > +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_vram_mgr.c > @@ -276,7 +276,7 @@ static int amdgpu_vram_mgr_new(struct ttm_mem_type_manager *man, > struct drm_mm_node *nodes; > enum drm_mm_insert_mode mode; > unsigned long lpfn, num_nodes, pages_per_node, pages_left; > - uint64_t usage = 0, vis_usage = 0; > + uint64_t vis_usage = 0; > unsigned i; > int r; > > @@ -284,6 +284,14 @@ static int amdgpu_vram_mgr_new(struct ttm_mem_type_manager *man, > if (!lpfn) > lpfn = man->size; > > + /* bail out quickly if there's likely not enough VRAM for this BO */ > + atomic64_add(mem->num_pages << PAGE_SHIFT, &mgr->usage); > + if (atomic64_read(&mgr->usage) > adev->gmc.mc_vram_size) { You should use atomic64_add_return here instead of two operations. Otherwise using an atomic doesn't make to much sense. Christian. > + atomic64_sub(mem->num_pages << PAGE_SHIFT, &mgr->usage); > + mem->mm_node = NULL; > + return 0; > + } > + > if (place->flags & TTM_PL_FLAG_CONTIGUOUS) { > pages_per_node = ~0ul; > num_nodes = 1; > @@ -300,8 +308,10 @@ static int amdgpu_vram_mgr_new(struct ttm_mem_type_manager *man, > > nodes = kvmalloc_array((uint32_t)num_nodes, sizeof(*nodes), > GFP_KERNEL | __GFP_ZERO); > - if (!nodes) > + if (!nodes) { > + atomic64_sub(mem->num_pages << PAGE_SHIFT, &mgr->usage); > return -ENOMEM; > + } > > mode = DRM_MM_INSERT_BEST; > if (place->flags & TTM_PL_FLAG_TOPDOWN) > @@ -321,7 +331,6 @@ static int amdgpu_vram_mgr_new(struct ttm_mem_type_manager *man, > if (unlikely(r)) > break; > > - usage += nodes[i].size << PAGE_SHIFT; > vis_usage += amdgpu_vram_mgr_vis_size(adev, &nodes[i]); > amdgpu_vram_mgr_virt_start(mem, &nodes[i]); > pages_left -= pages; > @@ -341,14 +350,12 @@ static int amdgpu_vram_mgr_new(struct ttm_mem_type_manager *man, > if (unlikely(r)) > goto error; > > - usage += nodes[i].size << PAGE_SHIFT; > vis_usage += amdgpu_vram_mgr_vis_size(adev, &nodes[i]); > amdgpu_vram_mgr_virt_start(mem, &nodes[i]); > pages_left -= pages; > } > spin_unlock(&mgr->lock); > > - atomic64_add(usage, &mgr->usage); > atomic64_add(vis_usage, &mgr->vis_usage); > > mem->mm_node = nodes; > @@ -359,6 +366,7 @@ static int amdgpu_vram_mgr_new(struct ttm_mem_type_manager *man, > while (i--) > drm_mm_remove_node(&nodes[i]); > spin_unlock(&mgr->lock); > + atomic64_sub(mem->num_pages << PAGE_SHIFT, &mgr->usage); > > kvfree(nodes); > return r == -ENOSPC ? 0 : r; _______________________________________________ amd-gfx mailing list amd-gfx@lists.freedesktop.org https://lists.freedesktop.org/mailman/listinfo/amd-gfx ^ permalink raw reply [flat|nested] 4+ messages in thread
end of thread, other threads:[~2019-06-12 8:03 UTC | newest]
Thread overview: 4+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2019-06-11 16:54 [PATCH] drm/amd/amdgpu: Bail out of BO node creation if not enough VRAM (v2) StDenis, Tom
[not found] ` <20190611165444.17841-1-tom.stdenis-5C7GfCeVMHo@public.gmane.org>
2019-06-12 7:26 ` Michel Dänzer
[not found] ` <fd4c5152-2670-8ac5-e6cc-fff6f260f3ed-otUistvHUpPR7s880joybQ@public.gmane.org>
2019-06-12 8:03 ` Christian König
2019-06-12 8:00 ` Christian König
This is a public inbox, see mirroring instructions for how to clone and mirror all data and code used for this inbox