summaryrefslogtreecommitdiff
path: root/drivers/gpu/drm/i915/i915_gem_render_state.c
diff options
context:
space:
mode:
authorChris Wilson <chris@chris-wilson.co.uk>2016-08-04 18:32:33 +0300
committerChris Wilson <chris@chris-wilson.co.uk>2016-08-04 22:20:00 +0300
commit305bc234a87fe8814149d36100b4b544caaddd00 (patch)
tree4eb58630d0ad2b3e79084ffc5f60da5d82e7ecaf /drivers/gpu/drm/i915/i915_gem_render_state.c
parent3272db53136f6be7555fb294db3a6e3f372b9380 (diff)
downloadlinux-305bc234a87fe8814149d36100b4b544caaddd00.tar.xz
drm/i915: Make i915_vma_pin() small and inline
Not only is i915_vma_pin() called for every single object on every single execbuf, it is usually a simple increment as the VMA is already bound for execution by the GPU. Rearrange the tests for unbound and pin_count overflow so that we can do the increment and test very cheaply and compact enough to inline the operation into execbuf. The trick used is to note that we can check for an overflow bit (keeping space available for it inside the flags) at the same time as checking the binding bits. Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk> Reviewed-by: Joonas Lahtinen <joonas.lahtinen@linux.intel.com> Link: http://patchwork.freedesktop.org/patch/msgid/1470324762-2545-17-git-send-email-chris@chris-wilson.co.uk
Diffstat (limited to 'drivers/gpu/drm/i915/i915_gem_render_state.c')
0 files changed, 0 insertions, 0 deletions