From: Thomas Hellström <thomas.hellstrom@xxxxxxxxx> The hwsp_gtt object is used for sub-allocation and could therefore be shared by many contexts causing unnecessary contention during concurrent context pinning. However since we're currently locking it only for pinning, it remains resident until we unpin it, and therefore it's safe to drop the lock early, allowing for concurrent thread access. Signed-off-by: Thomas Hellström <thomas.hellstrom@xxxxxxxxx> Reviewed-by: Chris Wilson <chris@xxxxxxxxxxxxxxxxxx> --- v2: Added additional clarifying code comments (Chris Wilson) --- drivers/gpu/drm/i915/gt/intel_context.c | 9 +++++++++ 1 file changed, 9 insertions(+) diff --git a/drivers/gpu/drm/i915/gt/intel_context.c b/drivers/gpu/drm/i915/gt/intel_context.c index 61b05cd4c47a..d301dda1b261 100644 --- a/drivers/gpu/drm/i915/gt/intel_context.c +++ b/drivers/gpu/drm/i915/gt/intel_context.c @@ -271,6 +271,15 @@ int __intel_context_do_pin_ww(struct intel_context *ce, i915_active_release(&ce->active); err_ctx_unpin: intel_context_post_unpin(ce); + + /* + * Unlock the hwsp_ggtt object since it's shared. + * In principle we can unlock all the global state locked above + * since it's pinned and doesn't need fencing, and will + * thus remain resident until it is explicitly unpinned. + */ + i915_gem_ww_unlock_single(ce->timeline->hwsp_ggtt->obj); + return err; } -- 2.25.1 _______________________________________________ Intel-gfx mailing list Intel-gfx@xxxxxxxxxxxxxxxxxxxxx https://lists.freedesktop.org/mailman/listinfo/intel-gfx
