[Intel-gfx] [PATCH v2] drm/i915: Order assert forcewake test

[Chris Wilson]

Read the current value before computing the expected to ensure that if
the timer does complete early (against our will), it should not cause a
false positive.

v2: The local irq disable did not prevent the timer from running.

Signed-off-by: Chris Wilson <chris at chris-wilson.co.uk>
Cc: Tvrtko Ursulin <tvrtko.ursulin at intel.com>
---
 drivers/gpu/drm/i915/intel_uncore.c | 7 +++----
 1 file changed, 3 insertions(+), 4 deletions(-)

diff --git a/drivers/gpu/drm/i915/intel_uncore.c b/drivers/gpu/drm/i915/intel_uncore.c
index 2042c94c9cc9..bb9e0da30e94 100644
--- a/drivers/gpu/drm/i915/intel_uncore.c
+++ b/drivers/gpu/drm/i915/intel_uncore.c
@@ -758,19 +758,18 @@ void assert_forcewakes_active(struct intel_uncore *uncore,
 	 * Check that the caller has an explicit wakeref and we don't mistake
 	 * it for the auto wakeref.
 	 */
-	local_irq_disable();
 	for_each_fw_domain_masked(domain, fw_domains, uncore, tmp) {
+		unsigned int actual = READ_ONCE(domain->wake_count);
 		unsigned int expect = 1;
 
 		if (hrtimer_active(&domain->timer) && READ_ONCE(domain->active))
 			expect++; /* pending automatic release */
 
-		if (WARN(domain->wake_count < expect,
+		if (WARN(actual < expect,
 			 "Expected domain %d to be held awake by caller, count=%d\n",
-			 domain->id, domain->wake_count))
+			 domain->id, actual))
 			break;
 	}
-	local_irq_enable();
 }
 
 /* We give fast paths for the really cool registers */
-- 
2.20.1