diff options
| author |   Dave Airlie <airlied@redhat.com> | 2017-07-21 14:00:40 +1000 |
|---|---|---|
| committer | Dave Airlie <airlied@redhat.com> | 2017-07-21 14:00:40 +1000 |
| commit | 22a548d04853297b3855dcac2881d0cb52e3b99c (patch) | |
| tree | 3e7d64542b720f1618b4cf2d125acfa3176b487e /drivers/gpu/drm/amd/amdgpu | |
| parent | Linux v4.13-rc1 (diff) | |
| parent | drm/amdgpu: Remove unused field kgd2kfd_shared_resources.num_mec (diff) | |
| download | linux-dev-22a548d04853297b3855dcac2881d0cb52e3b99c.tar.xz linux-dev-22a548d04853297b3855dcac2881d0cb52e3b99c.zip | |
Merge tag 'drm-amdkfd-fixes-2017-07-18' of git://people.freedesktop.org/~gabbayo/linux into drm-fixes
This is a pull request for 4.13-rc2/3. The first patch addresses a bug that
appears when doing over-subscription of queues, due to a miscalculation of
the number of available compute queues. The other three patches remove
dead code that was created from the bugfix.
* tag 'drm-amdkfd-fixes-2017-07-18' of git://people.freedesktop.org/~gabbayo/linux:
drm/amdgpu: Remove unused field kgd2kfd_shared_resources.num_mec
drm/radeon: Remove initialization of shared_resources.num_mec
drm/amdkfd: Remove unused references to shared_resources.num_mec
drm/amdgpu: Fix KFD oversubscription by tracking queues correctly
Diffstat (limited to 'drivers/gpu/drm/amd/amdgpu')
| -rw-r--r-- | drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.c | 3 |
1 files changed, 1 insertions, 2 deletions
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.c index 5f8ada1d872b..37971d9402e3 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.c @@ -101,7 +101,6 @@ void amdgpu_amdkfd_device_init(struct amdgpu_device *adev) if (adev->kfd) { struct kgd2kfd_shared_resources gpu_resources = { .compute_vmid_bitmap = 0xFF00, - .num_mec = adev->gfx.mec.num_mec, .num_pipe_per_mec = adev->gfx.mec.num_pipe_per_mec, .num_queue_per_pipe = adev->gfx.mec.num_queue_per_pipe }; @@ -122,7 +121,7 @@ void amdgpu_amdkfd_device_init(struct amdgpu_device *adev) /* According to linux/bitmap.h we shouldn't use bitmap_clear if * nbits is not compile time constant */ - last_valid_bit = adev->gfx.mec.num_mec + last_valid_bit = 1 /* only first MEC can have compute queues */ * adev->gfx.mec.num_pipe_per_mec * adev->gfx.mec.num_queue_per_pipe; for (i = last_valid_bit; i < KGD_MAX_QUEUES; ++i) |