Donate to e Foundation | Murena handsets with /e/OS | Own a part of Murena! Learn more

Commit 8c185eca authored by Chris Wilson's avatar Chris Wilson
Browse files

drm/i915: Split I915_RESET_IN_PROGRESS into two flags 



I915_RESET_IN_PROGRESS is being used for both signaling the requirement
to i915_mutex_lock_interruptible() to avoid taking the struct_mutex and
to instruct a waiter (already holding the struct_mutex) to perform the
reset. To allow for a little more coordination, split these two meaning
into a couple of distinct flags. I915_RESET_BACKOFF tells
i915_mutex_lock_interruptible() not to acquire the mutex and
I915_RESET_HANDOFF tells the waiter to call i915_reset().

Signed-off-by: default avatarChris Wilson <chris@chris-wilson.co.uk>
Acked-by: default avatarMichel Thierry <michel.thierry@intel.com>
Reviewed-by: default avatarMika Kuoppala <mika.kuoppala@intel.com>
Link: http://patchwork.freedesktop.org/patch/msgid/20170316171305.12972-1-chris@chris-wilson.co.uk
parent 3fc03069

drivers/gpu/drm/i915/i915_debugfs.c

+9 −7
Original line number Original line Diff line number Diff line
@@ -1305,16 +1305,18 @@ static int i915_hangcheck_info(struct seq_file *m, void *unused) 
	enum intel_engine_id id;

	enum intel_engine_id id;





	if (test_bit(I915_WEDGED, &dev_priv->gpu_error.flags))

	if (test_bit(I915_WEDGED, &dev_priv->gpu_error.flags))

		seq_printf(m, "Wedged\n");

		seq_puts(m, "Wedged\n");

	if (test_bit(I915_RESET_IN_PROGRESS, &dev_priv->gpu_error.flags))

	if (test_bit(I915_RESET_BACKOFF, &dev_priv->gpu_error.flags))

		seq_printf(m, "Reset in progress\n");

		seq_puts(m, "Reset in progress: struct_mutex backoff\n");

	if (test_bit(I915_RESET_HANDOFF, &dev_priv->gpu_error.flags))

		seq_puts(m, "Reset in progress: reset handoff to waiter\n");

	if (waitqueue_active(&dev_priv->gpu_error.wait_queue))

	if (waitqueue_active(&dev_priv->gpu_error.wait_queue))

		seq_printf(m, "Waiter holding struct mutex\n");

		seq_puts(m, "Waiter holding struct mutex\n");

	if (waitqueue_active(&dev_priv->gpu_error.reset_queue))

	if (waitqueue_active(&dev_priv->gpu_error.reset_queue))

		seq_printf(m, "struct_mutex blocked for reset\n");

		seq_puts(m, "struct_mutex blocked for reset\n");





	if (!i915.enable_hangcheck) {

	if (!i915.enable_hangcheck) {

		seq_printf(m, "Hangcheck disabled\n");

		seq_puts(m, "Hangcheck disabled\n");

		return 0;

		return 0;

	}

	}
@@ -4127,7 +4129,7 @@ i915_wedged_set(void *data, u64 val) 
	 * while it is writing to 'i915_wedged'

	 * while it is writing to 'i915_wedged'

	 */

	 */





	if (i915_reset_in_progress(&dev_priv->gpu_error))

	if (i915_reset_backoff(&dev_priv->gpu_error))

		return -EAGAIN;

		return -EAGAIN;





	i915_handle_error(dev_priv, val,

	i915_handle_error(dev_priv, val,

drivers/gpu/drm/i915/i915_drv.c

+5 −2
Original line number Original line Diff line number Diff line
@@ -1815,8 +1815,9 @@ void i915_reset(struct drm_i915_private *dev_priv) 
	int ret;

	int ret;





	lockdep_assert_held(&dev_priv->drm.struct_mutex);

	lockdep_assert_held(&dev_priv->drm.struct_mutex);

	GEM_BUG_ON(!test_bit(I915_RESET_BACKOFF, &error->flags));





	if (!test_and_clear_bit(I915_RESET_IN_PROGRESS, &error->flags))

	if (!test_bit(I915_RESET_HANDOFF, &error->flags))

		return;

		return;





	/* Clear any previous failed attempts at recovery. Time to try again. */

	/* Clear any previous failed attempts at recovery. Time to try again. */
@@ -1869,7 +1870,9 @@ void i915_reset(struct drm_i915_private *dev_priv) 
wakeup:

wakeup:

	i915_gem_reset_finish(dev_priv);

	i915_gem_reset_finish(dev_priv);

	enable_irq(dev_priv->drm.irq);

	enable_irq(dev_priv->drm.irq);

	wake_up_bit(&error->flags, I915_RESET_IN_PROGRESS);



	clear_bit(I915_RESET_HANDOFF, &error->flags);

	wake_up_bit(&error->flags, I915_RESET_HANDOFF);

	return;

	return;





error:

error:

drivers/gpu/drm/i915/i915_drv.h

+35 −5
Original line number Original line Diff line number Diff line
@@ -1595,8 +1595,33 @@ struct i915_gpu_error { 
	 */

	 */

	unsigned long reset_count;

	unsigned long reset_count;





	/**

	 * flags: Control various stages of the GPU reset

	 *

	 * #I915_RESET_BACKOFF - When we start a reset, we want to stop any

	 * other users acquiring the struct_mutex. To do this we set the

	 * #I915_RESET_BACKOFF bit in the error flags when we detect a reset

	 * and then check for that bit before acquiring the struct_mutex (in

	 * i915_mutex_lock_interruptible()?). I915_RESET_BACKOFF serves a

	 * secondary role in preventing two concurrent global reset attempts.

	 *

	 * #I915_RESET_HANDOFF - To perform the actual GPU reset, we need the

	 * struct_mutex. We try to acquire the struct_mutex in the reset worker,

	 * but it may be held by some long running waiter (that we cannot

	 * interrupt without causing trouble). Once we are ready to do the GPU

	 * reset, we set the I915_RESET_HANDOFF bit and wakeup any waiters. If

	 * they already hold the struct_mutex and want to participate they can

	 * inspect the bit and do the reset directly, otherwise the worker

	 * waits for the struct_mutex.

	 *

	 * #I915_WEDGED - If reset fails and we can no longer use the GPU,

	 * we set the #I915_WEDGED bit. Prior to command submission, e.g.

	 * i915_gem_request_alloc(), this bit is checked and the sequence

	 * aborted (with -EIO reported to userspace) if set.

	 */

	unsigned long flags;

	unsigned long flags;

#define I915_RESET_IN_PROGRESS	0

#define I915_RESET_BACKOFF	0

#define I915_RESET_HANDOFF	1

#define I915_WEDGED		(BITS_PER_LONG - 1)

#define I915_WEDGED		(BITS_PER_LONG - 1)





	/**

	/**
@@ -3387,9 +3412,14 @@ i915_gem_find_active_request(struct intel_engine_cs *engine); 




void i915_gem_retire_requests(struct drm_i915_private *dev_priv);

void i915_gem_retire_requests(struct drm_i915_private *dev_priv);





static inline bool i915_reset_in_progress(struct i915_gpu_error *error)

static inline bool i915_reset_backoff(struct i915_gpu_error *error)

{

	return unlikely(test_bit(I915_RESET_BACKOFF, &error->flags));

}



static inline bool i915_reset_handoff(struct i915_gpu_error *error)

{

{

	return unlikely(test_bit(I915_RESET_IN_PROGRESS, &error->flags));

	return unlikely(test_bit(I915_RESET_HANDOFF, &error->flags));

}

}





static inline bool i915_terminally_wedged(struct i915_gpu_error *error)

static inline bool i915_terminally_wedged(struct i915_gpu_error *error)
@@ -3397,9 +3427,9 @@ static inline bool i915_terminally_wedged(struct i915_gpu_error *error) 
	return unlikely(test_bit(I915_WEDGED, &error->flags));

	return unlikely(test_bit(I915_WEDGED, &error->flags));

}

}





static inline bool i915_reset_in_progress_or_wedged(struct i915_gpu_error *error)

static inline bool i915_reset_backoff_or_wedged(struct i915_gpu_error *error)

{

{

	return i915_reset_in_progress(error) | i915_terminally_wedged(error);

	return i915_reset_backoff(error) | i915_terminally_wedged(error);

}

}





static inline u32 i915_reset_count(struct i915_gpu_error *error)

static inline u32 i915_reset_count(struct i915_gpu_error *error)

drivers/gpu/drm/i915/i915_gem.c

+1 −4
Original line number Original line Diff line number Diff line
@@ -103,16 +103,13 @@ i915_gem_wait_for_error(struct i915_gpu_error *error) 




	might_sleep();

	might_sleep();





	if (!i915_reset_in_progress(error))

		return 0;



	/*

	/*

	 * Only wait 10 seconds for the gpu reset to complete to avoid hanging

	 * Only wait 10 seconds for the gpu reset to complete to avoid hanging

	 * userspace. If it takes that long something really bad is going on and

	 * userspace. If it takes that long something really bad is going on and

	 * we should simply try to bail out and fail as gracefully as possible.

	 * we should simply try to bail out and fail as gracefully as possible.

	 */

	 */

	ret = wait_event_interruptible_timeout(error->reset_queue,

	ret = wait_event_interruptible_timeout(error->reset_queue,

					       !i915_reset_in_progress(error),

					       !i915_reset_backoff(error),

					       I915_RESET_TIMEOUT);

					       I915_RESET_TIMEOUT);

	if (ret == 0) {

	if (ret == 0) {

		DRM_ERROR("Timed out waiting for the gpu reset to complete\n");

		DRM_ERROR("Timed out waiting for the gpu reset to complete\n");

drivers/gpu/drm/i915/i915_gem_request.c

+1 −1
Original line number Original line Diff line number Diff line
@@ -1012,7 +1012,7 @@ bool __i915_spin_request(const struct drm_i915_gem_request *req, 




static bool __i915_wait_request_check_and_reset(struct drm_i915_gem_request *request)

static bool __i915_wait_request_check_and_reset(struct drm_i915_gem_request *request)

{

{

	if (likely(!i915_reset_in_progress(&request->i915->gpu_error)))

	if (likely(!i915_reset_handoff(&request->i915->gpu_error)))

		return false;

		return false;





	__set_current_state(TASK_RUNNING);

	__set_current_state(TASK_RUNNING);