drm/i915: Hold forcewake for the duration of reset+restart

Resetting the engine requires us to hold the forcewake wakeref to
prevent RC6 trying to happen in the middle of the reset sequence. The
consequence of an unwanted RC6 event in the middle is that random state
is then saved to the powercontext and restored later, which may
overwrite the mmio state we need to preserve (e.g. PD_DIR_BASE in the
legacy ringbuffer reset_ring_common()).

This was noticed in the live_hangcheck selftests when Haswell would
sporadically fail to restart during igt_reset_queue().

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Mika Kuoppala <mika.kuoppala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20171009110301.21705-3-chris@chris-wilson.co.uk
Reviewed-by: Mika Kuoppala <mika.kuoppala@linux.intel.com>

diff --git a/drivers/gpu/drm/i915/i915_gem.c b/drivers/gpu/drm/i915/i915_gem.c
index 82a10036fb3880e10107d5c763721f862b4189af..eba23c239aae2d1cd97f91eea0a25303b62cdc71 100644 (file)
--- a/drivers/gpu/drm/i915/i915_gem.c
+++ b/drivers/gpu/drm/i915/i915_gem.c
@@ -2832,7 +2832,17 @@ i915_gem_reset_prepare_engine(struct intel_engine_cs *engine)
 {
        struct drm_i915_gem_request *request = NULL;
 
-       /* Prevent the signaler thread from updating the request
+       /*
+        * During the reset sequence, we must prevent the engine from
+        * entering RC6. As the context state is undefined until we restart
+        * the engine, if it does enter RC6 during the reset, the state
+        * written to the powercontext is undefined and so we may lose
+        * GPU state upon resume, i.e. fail to restart after a reset.
+        */
+       intel_uncore_forcewake_get(engine->i915, FORCEWAKE_ALL);
+
+       /*
+        * Prevent the signaler thread from updating the request
         * state (by calling dma_fence_signal) as we are processing
         * the reset. The write from the GPU of the seqno is
         * asynchronous and the signaler thread may see a different
@@ -2843,7 +2853,8 @@ i915_gem_reset_prepare_engine(struct intel_engine_cs *engine)
         */
        kthread_park(engine->breadcrumbs.signaler);
 
-       /* Prevent request submission to the hardware until we have
+       /*
+        * 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
@@ -3033,6 +3044,8 @@ void i915_gem_reset_finish_engine(struct intel_engine_cs *engine)
 {
        tasklet_enable(&engine->execlists.irq_tasklet);
        kthread_unpark(engine->breadcrumbs.signaler);
+
+       intel_uncore_forcewake_put(engine->i915, FORCEWAKE_ALL);
 }
 
 void i915_gem_reset_finish(struct drm_i915_private *dev_priv)