From c75870d86f6a55a08ef0e67cf6cc25ae124ca9e7 Mon Sep 17 00:00:00 2001 From: Daniel Vetter Date: Mon, 22 Aug 2016 10:55:22 +0200 Subject: [PATCH] drm/i915: Ensure consistent control flow __i915_gem_active_get_rcu This issue here is (I think) purely theoretical, since a compiler would need to be especially foolish to recompute the value of i915_gem_request_completed right after it was already used. Hence the additional barrier() is also not really a restriction. But I believe this to be at least permissible, and since our rcu trickery is a beast it's worth to annotate all the corner cases. Chris proposed to instead just wrap a READ_ONCE around request->fence.seqno in i915_gem_request_completed. But that has a measurable impact on code size, and everywhere we hold a full reference to the underlying request it's also not needed. And personally I'd like to have just enough barriers and locking needed for correctness, but not more - it makes it much easier in the future to understand what's going on. Since the busy ioctl has now fully embraced it's races there's no point annotating it there too. We really only need it in active_get_rcu, since that function _must_ deliver a correct snapshot of the active fences (and not chase something else). v2: Polish the comment a bit more (Chris). Cc: Chris Wilson Cc: Joonas Lahtinen Reviewed-by: Chris Wilson Signed-off-by: Daniel Vetter Signed-off-by: Chris Wilson Link: http://patchwork.freedesktop.org/patch/msgid/1471856122-466-1-git-send-email-daniel.vetter@ffwll.ch --- drivers/gpu/drm/i915/i915_gem_request.h | 13 +++++++++++++ 1 file changed, 13 insertions(+) diff --git a/drivers/gpu/drm/i915/i915_gem_request.h b/drivers/gpu/drm/i915/i915_gem_request.h index 6c72bd8d9423..91014de8bfbc 100644 --- a/drivers/gpu/drm/i915/i915_gem_request.h +++ b/drivers/gpu/drm/i915/i915_gem_request.h @@ -472,6 +472,19 @@ __i915_gem_active_get_rcu(const struct i915_gem_active *active) if (!request || i915_gem_request_completed(request)) return NULL; + /* An especially silly compiler could decide to recompute the + * result of i915_gem_request_completed, more specifically + * re-emit the load for request->fence.seqno. A race would catch + * a later seqno value, which could flip the result from true to + * false. Which means part of the instructions below might not + * be executed, while later on instructions are executed. Due to + * barriers within the refcounting the inconsistency can't reach + * past the call to i915_gem_request_get_rcu, but not executing + * that while still executing i915_gem_request_put() creates + * havoc enough. Prevent this with a compiler barrier. + */ + barrier(); + request = i915_gem_request_get_rcu(request); /* What stops the following rcu_access_pointer() from occurring -- 2.20.1