[Intel-gfx] [CI 3/3] drm/i915: Disable engine->irq_tasklet around resets

[Chris Wilson]

When we restart the engines, and we have active requests, a request on
the first engine may complete and queue a request to the second engine
before we try to restart the second engine. That queueing of the
request may race with the engine to restart, and so may corrupt the
current state. Disabling the engine->irq_tasklet prevents the two paths
from writing into ELSP simultaneously (and modifyin the execlists_port[]
at the same time).

Fixes: 821ed7df6e2a ("drm/i915: Update reset path to fix incomplete requests")
Testcase: igt/gem_exec_fence/await-hang
Signed-off-by: Chris Wilson <chris at chris-wilson.co.uk>
Cc: Tvrtko Ursulin <tvrtko.ursulin at intel.com>
Cc: Mika Kuoppala <mika.kuoppala at intel.com>
Reviewed-by: Mika Kuoppala <mika.kuoppala at intel.com>
---
 drivers/gpu/drm/i915/i915_gem.c | 15 +++++++++++++++
 1 file changed, 15 insertions(+)

diff --git a/drivers/gpu/drm/i915/i915_gem.c b/drivers/gpu/drm/i915/i915_gem.c
index 4b1832a925bf..bd311a30bed7 100644
--- a/drivers/gpu/drm/i915/i915_gem.c
+++ b/drivers/gpu/drm/i915/i915_gem.c
@@ -2641,6 +2641,15 @@ int i915_gem_reset_prepare(struct drm_i915_private *dev_priv)
 	for_each_engine(engine, dev_priv, id) {
 		struct drm_i915_gem_request *request;
 
+		/* Prevent request submission to the hardware until we have
+		 * completed the reset in i915_gem_reset_finish(). If a request
+		 * is completed by one engine, it may then queue a request
+		 * to a second via its engine->irq_tasklet *just* as we are
+		 * calling engine->init_hw() and also writing the ELSP.
+		 * Turning off the engine->irq_tasklet until the reset is over
+		 * prevents the race.
+		 */
+		tasklet_disable(&engine->irq_tasklet);
 		tasklet_kill(&engine->irq_tasklet);
 
 		if (engine_stalled(engine)) {
@@ -2780,7 +2789,13 @@ void i915_gem_reset(struct drm_i915_private *dev_priv)
 
 void i915_gem_reset_finish(struct drm_i915_private *dev_priv)
 {
+	struct intel_engine_cs *engine;
+	enum intel_engine_id id;
+
 	lockdep_assert_held(&dev_priv->drm.struct_mutex);
+
+	for_each_engine(engine, dev_priv, id)
+		tasklet_enable(&engine->irq_tasklet);
 }
 
 static void nop_submit_request(struct drm_i915_gem_request *request)
-- 
2.11.0