drm/i915: Be irqsafe inside reset
authorChris Wilson <chris@chris-wilson.co.uk>
Fri, 15 Jun 2018 09:31:36 +0000 (10:31 +0100)
committerChris Wilson <chris@chris-wilson.co.uk>
Fri, 15 Jun 2018 18:50:36 +0000 (19:50 +0100)
As we want to be able to call i915_reset_engine and co from a softirq or
timer context, we need to be irqsafe at all times. So we have to forgo
the simple spin_lock_irq for the full spin_lock_irqsave.

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20180615093137.14270-1-chris@chris-wilson.co.uk
drivers/gpu/drm/i915/i915_gem.c

index 977982a..5155e45 100644 (file)
@@ -3166,15 +3166,17 @@ i915_gem_reset_request(struct intel_engine_cs *engine,
                 */
                request = i915_gem_find_active_request(engine);
                if (request) {
+                       unsigned long flags;
+
                        i915_gem_context_mark_innocent(request->gem_context);
                        dma_fence_set_error(&request->fence, -EAGAIN);
 
                        /* Rewind the engine to replay the incomplete rq */
-                       spin_lock_irq(&engine->timeline.lock);
+                       spin_lock_irqsave(&engine->timeline.lock, flags);
                        request = list_prev_entry(request, link);
                        if (&request->link == &engine->timeline.requests)
                                request = NULL;
-                       spin_unlock_irq(&engine->timeline.lock);
+                       spin_unlock_irqrestore(&engine->timeline.lock, flags);
                }
        }