From: Brad Volkin <bradley.d.volkin@xxxxxxxxx> This is v2 of the series I sent here: http://lists.freedesktop.org/archives/intel-gfx/2014-June/047609.html I believe that I've addressed all of the feedback except * I didn't move the allocation of the shadow batch buffer into parse_cmds(). It didn't seem like it added much value and would maybe complicate the error handling in do_execbuffer(). * I kept the part about attaching the shadow batch to the request, though in perhaps a less invasive way. My concern here is with the scheduler possibly reordering requests, I don't know if we'd still be able to implement the busy tracking in the pool as suggested. The commit message for patch 4 still applies: we aren't ready for that change until the secure dispatch regression is resolved, but it's needed for testing. I've added patch 5 to use batch_len instead of object size, as an optimization. My testing didn't show any perf difference, but I don't have any libva benchmarks to run, and that's where it sounded like the issue would be. I just tacked the patch onto the end of the series rather than squashing it in so we can easily take it or leave it as desired. Brad Volkin (5): drm/i915: Implement a framework for batch buffer pools drm/i915: Use batch pools with the command parser drm/i915: Add a batch pool debugfs file drm/i915: Dispatch the shadow batch buffer drm/i915: Use batch length instead of object size in command parser Documentation/DocBook/drm.tmpl | 5 ++ drivers/gpu/drm/i915/Makefile | 1 + drivers/gpu/drm/i915/i915_cmd_parser.c | 88 +++++++++++++++---- drivers/gpu/drm/i915/i915_debugfs.c | 41 +++++++++ drivers/gpu/drm/i915/i915_dma.c | 1 + drivers/gpu/drm/i915/i915_drv.h | 26 ++++++ drivers/gpu/drm/i915/i915_gem.c | 10 +++ drivers/gpu/drm/i915/i915_gem_batch_pool.c | 133 +++++++++++++++++++++++++++++ drivers/gpu/drm/i915/i915_gem_execbuffer.c | 39 ++++++++- 9 files changed, 325 insertions(+), 19 deletions(-) create mode 100644 drivers/gpu/drm/i915/i915_gem_batch_pool.c -- 1.8.3.2 _______________________________________________ Intel-gfx mailing list Intel-gfx@xxxxxxxxxxxxxxxxxxxxx http://lists.freedesktop.org/mailman/listinfo/intel-gfx
