On 11/01/16 09:17, Chris Wilson wrote:
If is simpler and leads to more readable code through the callstack if the allocation returns the allocated struct through the return value. The importance of this is that it no longer looks like we accidentally allocate requests as side-effect of calling certain functions. Signed-off-by: Chris Wilson <chris@xxxxxxxxxxxxxxxxxx> --- drivers/gpu/drm/i915/i915_drv.h | 3 +- drivers/gpu/drm/i915/i915_gem.c | 82 ++++++++++-------------------- drivers/gpu/drm/i915/i915_gem_execbuffer.c | 8 +-- drivers/gpu/drm/i915/i915_gem_request.c | 22 +++----- drivers/gpu/drm/i915/i915_gem_request.h | 6 +-- drivers/gpu/drm/i915/i915_trace.h | 15 +++--- drivers/gpu/drm/i915/intel_display.c | 25 +++++---- drivers/gpu/drm/i915/intel_lrc.c | 6 +-- drivers/gpu/drm/i915/intel_overlay.c | 24 ++++----- 9 files changed, 77 insertions(+), 114 deletions(-)
I would quite like to have reviewed this, since I think simplifying {i915_gem_}request_alloc() is a really good idea, and have submitted a patch to do just that. But I can't review it because it doesn't apply, even if I start from nightly and try to apply the whole series of 190 patches (it fails at patch 018/190, Separate out the seqno-barrier from engine->get_seqno).
Actually, it looks like the request_alloc() changes *would* apply cleanly without any of the rest of the pile of patches, except that this patch does more than it admits; in addition to the improvement to request_alloc() it also rewrites i915_gem_object_sync(), and *those* changes don't apply to nightly.
So this should be two separate patches, one to improve request_alloc() and then a separate one to update i915_gem_object_sync().
.Dave. _______________________________________________ Intel-gfx mailing list Intel-gfx@xxxxxxxxxxxxxxxxxxxxx http://lists.freedesktop.org/mailman/listinfo/intel-gfx
