drm/i915: Hold irq-off for the entire fake lock period

Sadly lockdep records when the irqs are re-enabled and then marks up the
fake lock as being irq-unsafe. Our hand is forced and so we must mark up
the entire fake lock critical section as irq-off.

Hopefully this is the last tweak required!

v2: Not quite, we need to mark the timeline spinlock as irqsafe. That
was a genuine bug being hidden by the earlier lockdep splat.

Fixes: d67739268c ("drm/i915/gt: Mark up the nested engine-pm timeline lock as irqsafe")
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Cc: Mika Kuoppala <mika.kuoppala@linux.intel.com>
Reviewed-by: Mika Kuoppala <mika.kuoppala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20190823132700.25286-2-chris@chris-wilson.co.uk
(cherry picked from commit 6dcb85a0ad)
Signed-off-by: Rodrigo Vivi <rodrigo.vivi@intel.com>

drivers/gpu/drm/i915/gt/intel_engine_pm.c

@@ -39,27 +39,32 @@ static int __engine_unpark(struct intel_wakeref *wf)
 
 #if IS_ENABLED(CONFIG_LOCKDEP)
 
-static inline void __timeline_mark_lock(struct intel_context *ce)
+static inline unsigned long __timeline_mark_lock(struct intel_context *ce)
 {
 	unsigned long flags;
 
 	local_irq_save(flags);
 	mutex_acquire(&ce->timeline->mutex.dep_map, 2, 0, _THIS_IP_);
-	local_irq_restore(flags);
+
+	return flags;
 }
 
-static inline void __timeline_mark_unlock(struct intel_context *ce)
+static inline void __timeline_mark_unlock(struct intel_context *ce,
+					  unsigned long flags)
 {
 	mutex_release(&ce->timeline->mutex.dep_map, 0, _THIS_IP_);
+	local_irq_restore(flags);
 }
 
 #else
 
-static inline void __timeline_mark_lock(struct intel_context *ce)
+static inline unsigned long __timeline_mark_lock(struct intel_context *ce)
 {
+	return 0;
 }
 
-static inline void __timeline_mark_unlock(struct intel_context *ce)
+static inline void __timeline_mark_unlock(struct intel_context *ce,
+					  unsigned long flags)
 {
 }
 
@@ -68,6 +73,8 @@ static inline void __timeline_mark_unlock(struct intel_context *ce)
 static bool switch_to_kernel_context(struct intel_engine_cs *engine)
 {
 	struct i915_request *rq;
+	unsigned long flags;
+	bool result = true;
 
 	/* Already inside the kernel context, safe to power down. */
 	if (engine->wakeref_serial == engine->serial)
@@ -89,12 +96,12 @@ static bool switch_to_kernel_context(struct intel_engine_cs *engine)
 	 * retiring the last request, thus all rings should be empty and
 	 * all timelines idle.
 	 */
-	__timeline_mark_lock(engine->kernel_context);
+	flags = __timeline_mark_lock(engine->kernel_context);
 
 	rq = __i915_request_create(engine->kernel_context, GFP_NOWAIT);
 	if (IS_ERR(rq))
 		/* Context switch failed, hope for the best! Maybe reset? */
-		return true;
+		goto out_unlock;
 
 	intel_timeline_enter(rq->timeline);
 
@@ -110,9 +117,10 @@ static bool switch_to_kernel_context(struct intel_engine_cs *engine)
 	__intel_wakeref_defer_park(&engine->wakeref);
 	__i915_request_queue(rq, NULL);
 
-	__timeline_mark_unlock(engine->kernel_context);
-
-	return false;
+	result = false;
+out_unlock:
+	__timeline_mark_unlock(engine->kernel_context, flags);
+	return result;
 }
 
 static int __engine_park(struct intel_wakeref *wf)

drivers/gpu/drm/i915/gt/intel_reset.c

@@ -792,6 +792,7 @@ static bool __intel_gt_unset_wedged(struct intel_gt *gt)
 {
 	struct intel_gt_timelines *timelines = &gt->timelines;
 	struct intel_timeline *tl;
+	unsigned long flags;
 
 	if (!test_bit(I915_WEDGED, &gt->reset.flags))
 		return true;
@@ -811,7 +812,7 @@ static bool __intel_gt_unset_wedged(struct intel_gt *gt)
 	 *
 	 * No more can be submitted until we reset the wedged bit.
 	 */
-	spin_lock(&timelines->lock);
+	spin_lock_irqsave(&timelines->lock, flags);
 	list_for_each_entry(tl, &timelines->active_list, link) {
 		struct i915_request *rq;
 
@@ -819,7 +820,7 @@ static bool __intel_gt_unset_wedged(struct intel_gt *gt)
 		if (!rq)
 			continue;
 
-		spin_unlock(&timelines->lock);
+		spin_unlock_irqrestore(&timelines->lock, flags);
 
 		/*
 		 * All internal dependencies (i915_requests) will have
@@ -832,10 +833,10 @@ static bool __intel_gt_unset_wedged(struct intel_gt *gt)
 		i915_request_put(rq);
 
 		/* Restart iteration after droping lock */
-		spin_lock(&timelines->lock);
+		spin_lock_irqsave(&timelines->lock, flags);
 		tl = list_entry(&timelines->active_list, typeof(*tl), link);
 	}
-	spin_unlock(&timelines->lock);
+	spin_unlock_irqrestore(&timelines->lock, flags);
 
 	intel_gt_sanitize(gt, false);
 

drivers/gpu/drm/i915/gt/intel_timeline.c

@@ -337,6 +337,7 @@ int intel_timeline_pin(struct intel_timeline *tl)
 void intel_timeline_enter(struct intel_timeline *tl)
 {
 	struct intel_gt_timelines *timelines = &tl->gt->timelines;
+	unsigned long flags;
 
 	lockdep_assert_held(&tl->mutex);
 
@@ -345,14 +346,15 @@ void intel_timeline_enter(struct intel_timeline *tl)
 		return;
 	GEM_BUG_ON(!tl->active_count); /* overflow? */
 
-	spin_lock(&timelines->lock);
+	spin_lock_irqsave(&timelines->lock, flags);
 	list_add(&tl->link, &timelines->active_list);
-	spin_unlock(&timelines->lock);
+	spin_unlock_irqrestore(&timelines->lock, flags);
 }
 
 void intel_timeline_exit(struct intel_timeline *tl)
 {
 	struct intel_gt_timelines *timelines = &tl->gt->timelines;
+	unsigned long flags;
 
 	lockdep_assert_held(&tl->mutex);
 
@@ -360,9 +362,9 @@ void intel_timeline_exit(struct intel_timeline *tl)
 	if (--tl->active_count)
 		return;
 
-	spin_lock(&timelines->lock);
+	spin_lock_irqsave(&timelines->lock, flags);
 	list_del(&tl->link);
-	spin_unlock(&timelines->lock);
+	spin_unlock_irqrestore(&timelines->lock, flags);
 
 	/*
 	 * Since this timeline is idle, all bariers upon which we were waiting

drivers/gpu/drm/i915/i915_gem.c

@@ -889,12 +889,13 @@ void i915_gem_runtime_suspend(struct drm_i915_private *i915)
 
 static long
 wait_for_timelines(struct drm_i915_private *i915,
-		   unsigned int flags, long timeout)
+		   unsigned int wait, long timeout)
 {
 	struct intel_gt_timelines *timelines = &i915->gt.timelines;
 	struct intel_timeline *tl;
+	unsigned long flags;
 
-	spin_lock(&timelines->lock);
+	spin_lock_irqsave(&timelines->lock, flags);
 	list_for_each_entry(tl, &timelines->active_list, link) {
 		struct i915_request *rq;
 
@@ -902,7 +903,7 @@ wait_for_timelines(struct drm_i915_private *i915,
 		if (!rq)
 			continue;
 
-		spin_unlock(&timelines->lock);
+		spin_unlock_irqrestore(&timelines->lock, flags);
 
 		/*
 		 * "Race-to-idle".
@@ -913,19 +914,19 @@ wait_for_timelines(struct drm_i915_private *i915,
 		 * want to complete as quickly as possible to avoid prolonged
 		 * stalls, so allow the gpu to boost to maximum clocks.
 		 */
-		if (flags & I915_WAIT_FOR_IDLE_BOOST)
+		if (wait & I915_WAIT_FOR_IDLE_BOOST)
 			gen6_rps_boost(rq);
 
-		timeout = i915_request_wait(rq, flags, timeout);
+		timeout = i915_request_wait(rq, wait, timeout);
 		i915_request_put(rq);
 		if (timeout < 0)
 			return timeout;
 
 		/* restart after reacquiring the lock */
-		spin_lock(&timelines->lock);
+		spin_lock_irqsave(&timelines->lock, flags);
 		tl = list_entry(&timelines->active_list, typeof(*tl), link);
 	}
-	spin_unlock(&timelines->lock);
+	spin_unlock_irqrestore(&timelines->lock, flags);
 
 	return timeout;
 }

drivers/gpu/drm/i915/i915_request.c

@@ -1465,9 +1465,10 @@ bool i915_retire_requests(struct drm_i915_private *i915)
 {
 	struct intel_gt_timelines *timelines = &i915->gt.timelines;
 	struct intel_timeline *tl, *tn;
+	unsigned long flags;
 	LIST_HEAD(free);
 
-	spin_lock(&timelines->lock);
+	spin_lock_irqsave(&timelines->lock, flags);
 	list_for_each_entry_safe(tl, tn, &timelines->active_list, link) {
 		if (!mutex_trylock(&tl->mutex))
 			continue;
@@ -1475,11 +1476,11 @@ bool i915_retire_requests(struct drm_i915_private *i915)
 		intel_timeline_get(tl);
 		GEM_BUG_ON(!tl->active_count);
 		tl->active_count++; /* pin the list element */
-		spin_unlock(&timelines->lock);
+		spin_unlock_irqrestore(&timelines->lock, flags);
 
 		retire_requests(tl);
 
-		spin_lock(&timelines->lock);
+		spin_lock_irqsave(&timelines->lock, flags);
 
 		/* Resume iteration after dropping lock */
 		list_safe_reset_next(tl, tn, link);
@@ -1494,7 +1495,7 @@ bool i915_retire_requests(struct drm_i915_private *i915)
 			list_add(&tl->link, &free);
 		}
 	}
-	spin_unlock(&timelines->lock);
+	spin_unlock_irqrestore(&timelines->lock, flags);
 
 	list_for_each_entry_safe(tl, tn, &free, link)
 		__intel_timeline_free(&tl->kref);