RE: [PATCH 11/11] drm/amdgpu: stop removing BOs from the LRU during CS

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



Hi Christian

 

With the series patch set , amdgpu_vm_validate_pt_bos occasionally evicted amdgpu BOs failed and can’t

find the valid first busy bo . Another problem is that  during the first BOs get lock period will run into deadlock .

 

/* check if other user occupy memory too long time */

                if (!first_bo || !request_resv || !request_resv->lock.ctx) {

                        if (first_bo)

                                ttm_bo_put(first_bo);

                        return -EBUSY;

                }

                if (first_bo->resv == request_resv) {

                        ttm_bo_put(first_bo);

                        return -EBUSY;

                }

                if (ctx->interruptible)

                        ret = ww_mutex_lock_interruptible(&first_bo->resv->lock,

                                                          request_resv->lock.ctx);

                else

                        ret = ww_mutex_lock(&first_bo->resv->lock, request_resv->lock.ctx);

                if (ret) {

                        ttm_bo_put(first_bo);

                        if (ret == -EDEADLK) {

                                ret = -EAGAIN;

                        }

 

                        return ret;

                } 

 

Thanks

Prike

 

From: Christian König <ckoenig.leichtzumerken@xxxxxxxxx>
Sent: Wednesday, May 15, 2019 3:05 PM
To: Liang, Prike <Prike.Liang@xxxxxxx>; Marek Olšák <maraeo@xxxxxxxxx>
Cc: Zhou, David(ChunMing) <David1.Zhou@xxxxxxx>; dri-devel <dri-devel@xxxxxxxxxxxxxxxxxxxxx>; amd-gfx mailing list <amd-gfx@xxxxxxxxxxxxxxxxxxxxx>
Subject: Re: [PATCH 11/11] drm/amdgpu: stop removing BOs from the LRU during CS

 

[CAUTION: External Email]

Hi Prike,

no, that can lead to massive problems in a real OOM situation and is not something we can do here.

Christian.

Am 15.05.19 um 04:00 schrieb Liang, Prike:

Hi Christian ,

 

I just wonder when encounter ENOMEM error during pin amdgpu BOs can we retry validate again as below.

With the following simply patch the Abaqus pinned issue not observed.

 

diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_object.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_object.c

index 11cbf63..72a32f5 100644

--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_object.c

+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_object.c

@@ -902,11 +902,15 @@ int amdgpu_bo_pin_restricted(struct amdgpu_bo *bo, u32 domain,

                        bo->placements[i].lpfn = lpfn;

                bo->placements[i].flags |= TTM_PL_FLAG_NO_EVICT;

        }

-

+retry:

        r = ttm_bo_validate(&bo->tbo, &bo->placement, &ctx);

        if (unlikely(r)) {

-               dev_err(adev->dev, "%p pin failed\n", bo);

-               goto error;

+                if (r == -ENOMEM){

+                        goto retry;

+                } else {

+                       dev_err(adev->dev, "%p pin failed\n", bo);

+                       goto error;

+                }

        }

 

        bo->pin_count = 1;

 

 

Thanks,

Prike

 

From: Marek Olšák <maraeo@xxxxxxxxx>
Sent: Wednesday, May 15, 2019 3:33 AM
To: Christian König <ckoenig.leichtzumerken@xxxxxxxxx>
Cc: Zhou, David(ChunMing) <David1.Zhou@xxxxxxx>; Liang, Prike <Prike.Liang@xxxxxxx>; dri-devel <dri-devel@xxxxxxxxxxxxxxxxxxxxx>; amd-gfx mailing list <amd-gfx@xxxxxxxxxxxxxxxxxxxxx>
Subject: Re: [PATCH 11/11] drm/amdgpu: stop removing BOs from the LRU during CS

 

[CAUTION: External Email]

This series fixes the OOM errors. However, if I torture the kernel driver more, I can get it to deadlock and end up with unkillable processes. I can also get an OOM error. I just ran the test 5 times:

 

AMD_DEBUG=testgdsmm glxgears & AMD_DEBUG=testgdsmm glxgears & AMD_DEBUG=testgdsmm glxgears & AMD_DEBUG=testgdsmm glxgears & AMD_DEBUG=testgdsmm glxgears

 

Marek

 

On Tue, May 14, 2019 at 8:31 AM Christian König <ckoenig.leichtzumerken@xxxxxxxxx> wrote:

This avoids OOM situations when we have lots of threads
submitting at the same time.

Signed-off-by: Christian König <christian.koenig@xxxxxxx>
---
 drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c
index fff558cf385b..f9240a94217b 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c
@@ -648,7 +648,7 @@ static int amdgpu_cs_parser_bos(struct amdgpu_cs_parser *p,
        }

        r = ttm_eu_reserve_buffers(&p->ticket, &p->validated, true,
-                                  &duplicates, true);
+                                  &duplicates, false);
        if (unlikely(r != 0)) {
                if (r != -ERESTARTSYS)
                        DRM_ERROR("ttm_eu_reserve_buffers failed.\n");
--
2.17.1

_______________________________________________
amd-gfx mailing list
amd-gfx@xxxxxxxxxxxxxxxxxxxxx
https://lists.freedesktop.org/mailman/listinfo/amd-gfx

 

_______________________________________________
amd-gfx mailing list
amd-gfx@xxxxxxxxxxxxxxxxxxxxx
https://lists.freedesktop.org/mailman/listinfo/amd-gfx

[Index of Archives]     [Linux USB Devel]     [Linux Audio Users]     [Yosemite News]     [Linux Kernel]     [Linux SCSI]

  Powered by Linux