[Intel-gfx] [PATCH] drm/i915: Propagate fence->error across semaphores

[Chris Wilson]

Replacing an inter-engine fence with a semaphore reduced the HW
execution latency, but that comes at a cost. For normal fences, we are
able to propagate the metadata such as errors along with the signaling.
For semaphores, we are missing this error propagation so add it in the
back channel we use to monitor the semaphore overload.

This raises a valid point on whether error propagation is sufficient in
the semaphore case if it is coupled to a fatal error, such as EFAULT. It
is not, and we should teach ourselves not to use a semaphore if we would
chain up to an external fence whose error we must not ignore.

Fixes: ef4688497512 ("drm/i915: Propagate fence errors")
Signed-off-by: Chris Wilson <chris at chris-wilson.co.uk>
Cc: Tvrtko Ursulin <tvrtko.ursulin at intel.com>
Cc: Matthew Auld <matthew.auld at intel.com>
---
 drivers/gpu/drm/i915/i915_request.c | 3 +++
 1 file changed, 3 insertions(+)

diff --git a/drivers/gpu/drm/i915/i915_request.c b/drivers/gpu/drm/i915/i915_request.c
index 9c5de07db47d..96a8c7a1be73 100644
--- a/drivers/gpu/drm/i915/i915_request.c
+++ b/drivers/gpu/drm/i915/i915_request.c
@@ -614,6 +614,9 @@ semaphore_notify(struct i915_sw_fence *fence, enum i915_sw_fence_notify state)
 
 	switch (state) {
 	case FENCE_COMPLETE:
+		if (unlikely(fence->error))
+			i915_request_set_error_once(rq, fence->error);
+
 		if (!(READ_ONCE(rq->sched.attr.priority) & I915_PRIORITY_NOSEMAPHORE)) {
 			i915_request_get(rq);
 			init_irq_work(&rq->semaphore_work, irq_semaphore_cb);
-- 
2.20.1