[0/3] RFC/T: dma_resv vs. mmap_sem

Message

Daniel Vetter Aug. 20, 2019, 2:53 p.m. UTC

Hi all,

As part of all the recent discussions around ttm and dma_resv I started
looking into this. The goal (at least somewhen in the near future) is to
have it all documented and the cross-driver semantics locked down as much
as possible.

One of the biggest issues there is how the dma_resv ww_mutex fits into the
overall locking hierarchy, especially wrt core mm locks. This patch series
should fix that.

Comments, review and especially testing (there's only so much you can do
with auditing) very much appreciated.

Cheers, Daniel

Daniel Vetter (3):
  dma_resv: prime lockdep annotations
  drm/nouveau: slowpath for pushbuf ioctl
  drm/ttm: remove ttm_bo_wait_unreserved

 drivers/dma-buf/dma-resv.c            | 12 ++++++
 drivers/gpu/drm/nouveau/nouveau_gem.c | 57 ++++++++++++++++++---------
 drivers/gpu/drm/ttm/ttm_bo.c          | 34 ----------------
 drivers/gpu/drm/ttm/ttm_bo_vm.c       | 26 +-----------
 include/drm/ttm/ttm_bo_api.h          |  1 -
 5 files changed, 51 insertions(+), 79 deletions(-)