Hi all,
Somewhat motivated by me looking at the v3d patch I went and dug around in the dma-fence code a bit. Result was a bit of doc polish and making a bunch of callbacks optional for cases where most implementations wanted the same default behaviour.
Comments, testing, feedback very much welcome.
Thanks, Daniel
Daniel Vetter (17): dma-fence: Some kerneldoc polish for dma-fence.h dma-fence: remove fill_driver_data callback dma-fence: Make ->enable_signaling optional dma-fence: Allow wait_any_timeout for all fences dma-fence: Make ->wait callback optional drm/amdgpu: Remove unecessary dma_fence_ops drm: Remove unecessary dma_fence_ops drm/etnaviv: Remove unecessary dma_fence_ops drm/i915: Remove unecessary dma_fence_ops drm/msm: Remove unecessary dma_fence_ops drm/nouveau: Remove unecessary dma_fence_ops drm/qxl: Remove unecessary dma_fence_ops drm/radeon: Remove custom dma_fence_ops->wait implementation drm/vc4: Remove unecessary dma_fence_ops drm/vgem: Remove unecessary dma_fence_ops drm/virtio: Remove unecessary dma_fence_ops dma-fence: Polish kernel-doc for dma-fence.c
Documentation/driver-api/dma-buf.rst | 6 + drivers/dma-buf/dma-fence-array.c | 1 - drivers/dma-buf/dma-fence.c | 163 ++++++++---- drivers/dma-buf/sw_sync.c | 1 - .../gpu/drm/amd/amdgpu/amdgpu_amdkfd_fence.c | 2 - drivers/gpu/drm/amd/amdgpu/amdgpu_fence.c | 1 - drivers/gpu/drm/drm_crtc.c | 7 - drivers/gpu/drm/drm_syncobj.c | 1 - drivers/gpu/drm/etnaviv/etnaviv_gpu.c | 7 - drivers/gpu/drm/i915/i915_gem_clflush.c | 7 - .../gpu/drm/i915/selftests/i915_sw_fence.c | 7 - drivers/gpu/drm/msm/msm_fence.c | 7 - drivers/gpu/drm/nouveau/nouveau_fence.c | 1 - drivers/gpu/drm/qxl/qxl_release.c | 7 - drivers/gpu/drm/radeon/radeon_fence.c | 63 ----- drivers/gpu/drm/scheduler/sched_fence.c | 11 - drivers/gpu/drm/vc4/vc4_fence.c | 7 - drivers/gpu/drm/vgem/vgem_fence.c | 14 -- drivers/gpu/drm/virtio/virtgpu_fence.c | 7 - include/linux/dma-fence.h | 231 +++++++++++------- 20 files changed, 267 insertions(+), 284 deletions(-)