linux/drivers/gpu/drm/amd/amdgpu/amdgpu_ib.c, branch v4.3

drm/amdgpu: fix fence wait in sync_fence, instead should be in sync_rings

2015-08-25T14:39:39Z

Signed-off-by: Christian König Signed-off-by: Chunming Zhou Reviewed-by: Christian K?nig

drm/amdgpu: remove v_seq handling from the scheduler v2

2015-08-25T14:39:16Z

Simply not used any more. Only keep 32bit atomic for fence sequence numbering. v2: trivial rebase Signed-off-by: Christian König Reviewed-by: Alex Deucher (v1) Reviewed-by: Jammy Zhou (v1) Reviewed-by: Chunming Zhou (v1)

drm/amdgpu: improve sa_bo->fence by kernel fence

2015-08-25T14:38:41Z

Signed-off-by: Chunming Zhou Reviewed-by: Christian K?nig

drm/amdgpu: fix zeroing all IB fields manually v2

2015-08-20T21:05:34Z

The problem now is that we don't necessarily call amdgpu_ib_get() in some error paths and so work with uninitialized data. Better require that the memory is already zeroed. v2: better commit message Signed-off-by: Christian König Reviewed-by: Chunming Zhou (v1) Reviewed-by: Alex Deucher

drm/amd: add scheduler fence implementation (v2)

2015-08-17T20:51:07Z

scheduler fence is based on kernel fence framework. v2: squash in Christian's build fix Signed-off-by: Chunming Zhou Reviewed-by: Christian K?nig

drm/amdgpu: fix seq in ctx_add_fence

2015-08-17T20:50:43Z

if enabling scheduler, then the queued seq is assigned when pushing job before emitting job. Signed-off-by: Chunming Zhou Reviewed-by: Christian K?nig

drm/amdgpu: fix UVD/VCE fence handling

2015-08-17T20:50:18Z

We need to return the sequence number to userspace even when we don't use user fences. Signed-off-by: Christian König Reviewed-by: Alex Deucher

drm/amdgpu: rework vm_grab_id interface

2015-08-17T20:50:16Z

This makes assigning VM IDs independent from the use of VM IDs. Signed-off-by: Christian König Reviewed-by: Jammy Zhou

drm/amdgpu: add user fence context map v2

2015-08-17T20:50:14Z

This is a prerequisite for the GPU scheduler to make the order of submission independent from the order of execution. v2: properly implement the locking Signed-off-by: Christian König Reviewed-by: Jammy Zhou Reviewed-by: Chunming Zhou

drm/amdgpu: deal with foreign fences in amdgpu_sync

2015-08-17T20:50:13Z

This also requires some error handling from the callers of that function. Signed-off-by: Christian König Reviewed-by: Jammy Zhou Reviewed-by: Chunming Zhou