completely rework the dma_resv semantic

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

Subject: completely rework the dma_resv semantic
From: "Christian König" <ckoenig.leichtzumerken@xxxxxxxxx>
Date: Mon, 29 Nov 2021 13:06:31 +0100
Cc: linaro-mm-sig@xxxxxxxxxxxxxxxx, dri-devel@xxxxxxxxxxxxxxxxxxxxx, linux-media@xxxxxxxxxxxxxxx

Hi everyone,

compared to the last version I've dropped the pruning as suggested by
Maarten, split the new DMA_RESV_USAGE_* patches from the general
introduction as suggeted by Daniel and renamed OTEHRS to BOOKKEEP as
suggested by Pekka.

Please take a look and review,
Christian.

Follow-Ups:
- [PATCH 23/28] dma-buf: specify usage while adding fences to dma_resv obj v2
  - From: Christian König
- [PATCH 22/28] dma-buf: add enum dma_resv_usage v3
  - From: Christian König
- [PATCH 27/28] amdgpu: remove DMA-buf fence workaround
  - From: Christian König
- [PATCH 25/28] dma-buf: add DMA_RESV_USAGE_BOOKKEEP
  - From: Christian König
- [PATCH 24/28] dma-buf: add DMA_RESV_USAGE_KERNEL
  - From: Christian König
- [PATCH 09/28] RDMA: use dma_resv_wait() instead of extracting the fence
  - From: Christian König
- [PATCH 19/28] drm: support more than one write fence in drm_gem_plane_helper_prepare_fb
  - From: Christian König
- [PATCH 16/28] dma-buf: finally make dma_resv_excl_fence private
  - From: Christian König
- [PATCH 26/28] dma-buf: wait for map to complete for static attachments
  - From: Christian König
- [PATCH 28/28] drm/ttm: remove bo->moving
  - From: Christian König
- [PATCH 21/28] drm/amdgpu: use dma_resv_get_singleton in amdgpu_pasid_free_cb
  - From: Christian König
- [PATCH 17/28] dma-buf: drop the DAG approach for the dma_resv object
  - From: Christian König
- [PATCH 12/28] drm/vmwgfx: stop using dma_resv_excl_fence
  - From: Christian König
- [PATCH 20/28] drm/nouveau: support more than one write fence in fenv50_wndw_prepare_fb
  - From: Christian König
- [PATCH 13/28] drm/radeon: stop using dma_resv_excl_fence
  - From: Christian König
- [PATCH 11/28] drm/nouveau: stop using dma_resv_excl_fence
  - From: Christian König
- [PATCH 10/28] drm/etnaviv: stop using dma_resv_excl_fence
  - From: Christian König
- [PATCH 07/28] dma-buf: drop excl_fence parameter from dma_resv_get_fences
  - From: Christian König
- [PATCH 14/28] drm/amdgpu: remove excl as shared workarounds
  - From: Christian König
- [PATCH 08/28] dma-buf: add dma_resv_get_singleton v2
  - From: Christian König
- [PATCH 06/28] dma-buf: finally make the dma_resv_list private
  - From: Christian König
- [PATCH 15/28] drm/amdgpu: use dma_resv_for_each_fence for CS workaround
  - From: Christian König
- [PATCH 18/28] dma-buf/drivers: make reserving a shared slot mandatory v2
  - From: Christian König
- [PATCH 02/28] drm/ttm: stop pruning fences after wait
  - From: Christian König
- [PATCH 05/28] dma-buf: add dma_resv_replace_fences
  - From: Christian König
- [PATCH 03/28] dma-buf: make fence mandatory for dma_resv_add_excl_fence v2
  - From: Christian König
- [PATCH 01/28] drm/i915: Remove dma_resv_prune
  - From: Christian König
- [PATCH 04/28] drm/qxl: use iterator instead of dma_resv_shared_list
  - From: Christian König

Prev by Date: Re: [PATCH 1/6] Documentation/gpu: Reorganize DC documentation
Next by Date: [PATCH 04/28] drm/qxl: use iterator instead of dma_resv_shared_list
Previous by thread: Re: completely rework the dma_resv semantic
Next by thread: [PATCH 04/28] drm/qxl: use iterator instead of dma_resv_shared_list
Index(es):
- Date
- Thread

[Index of Archives] [Linux DRI Users] [Linux Intel Graphics] [Linux USB Devel] [Video for Linux] [Linux Audio Users] [Yosemite News] [Linux Kernel] [Linux SCSI] [XFree86] [Linux USB Devel] [Video for Linux] [Linux Audio Users] [Linux Kernel] [Linux SCSI] [XFree86]