Re: [Intel-gfx] [PATCH 1/2] drm/i915: Stop touching hangcheck.seqno from intel_engine_init_global_seqno()

[Chris Wilson]

On Fri, Apr 21, 2017 at 11:12:21AM +0300, Mika Kuoppala wrote:
> Chris Wilson  writes:
> 
> > The hangcheck runs independently to the main flow of seqno through the
> > driver. However, we have an odd coupling of the seqno reset that is
> > unwelcome, and if poked at just the right rate can cause spurious hangs
> > (e.g. gem_exec_whisper) on an apparently idle engine.
> >
> > Signed-off-by: Chris Wilson 
> > Cc: Mika Kuoppala 
> 
> Reviewed-by: Mika Kuoppala 

Ta, I'll send this by itself to CI to confirm that it is safe :)
-Chris

-- 
Chris Wilson, Intel Open Source Technology Centre
___
Intel-gfx mailing list
Intel-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/intel-gfx

Re: [Intel-gfx] [PATCH 1/2] drm/i915: Stop touching hangcheck.seqno from intel_engine_init_global_seqno()

[Mika Kuoppala]

Chris Wilson  writes:

> The hangcheck runs independently to the main flow of seqno through the
> driver. However, we have an odd coupling of the seqno reset that is
> unwelcome, and if poked at just the right rate can cause spurious hangs
> (e.g. gem_exec_whisper) on an apparently idle engine.
>
> Signed-off-by: Chris Wilson 
> Cc: Mika Kuoppala 

Reviewed-by: Mika Kuoppala 

> ---
>  drivers/gpu/drm/i915/intel_engine_cs.c | 4 +---
>  1 file changed, 1 insertion(+), 3 deletions(-)
>
> diff --git a/drivers/gpu/drm/i915/intel_engine_cs.c 
> b/drivers/gpu/drm/i915/intel_engine_cs.c
> index 7681d17ce454..f3cb7e931317 100644
> --- a/drivers/gpu/drm/i915/intel_engine_cs.c
> +++ b/drivers/gpu/drm/i915/intel_engine_cs.c
> @@ -265,6 +265,7 @@ void intel_engine_init_global_seqno(struct 
> intel_engine_cs *engine, u32 seqno)
>   struct drm_i915_private *dev_priv = engine->i915;
>  
>   GEM_BUG_ON(!intel_engine_is_idle(engine));
> + GEM_BUG_ON(i915_gem_active_isset(>timeline->last_request));
>  
>   /* Our semaphore implementation is strictly monotonic (i.e. we proceed
>* so long as the semaphore value in the register/page is greater
> @@ -284,9 +285,6 @@ void intel_engine_init_global_seqno(struct 
> intel_engine_cs *engine, u32 seqno)
>   intel_write_status_page(engine, I915_GEM_HWS_INDEX, seqno);
>   clear_bit(ENGINE_IRQ_BREADCRUMB, >irq_posted);
>  
> - GEM_BUG_ON(i915_gem_active_isset(>timeline->last_request));
> - engine->hangcheck.seqno = seqno;
> -
>   /* After manually advancing the seqno, fake the interrupt in case
>* there are any waiters for that seqno.
>*/
> -- 
> 2.11.0
___
Intel-gfx mailing list
Intel-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/intel-gfx

[Intel-gfx] [PATCH 1/2] drm/i915: Stop touching hangcheck.seqno from intel_engine_init_global_seqno()

[Chris Wilson]

The hangcheck runs independently to the main flow of seqno through the
driver. However, we have an odd coupling of the seqno reset that is
unwelcome, and if poked at just the right rate can cause spurious hangs
(e.g. gem_exec_whisper) on an apparently idle engine.

Signed-off-by: Chris Wilson 
Cc: Mika Kuoppala 
---
 drivers/gpu/drm/i915/intel_engine_cs.c | 4 +---
 1 file changed, 1 insertion(+), 3 deletions(-)

diff --git a/drivers/gpu/drm/i915/intel_engine_cs.c 
b/drivers/gpu/drm/i915/intel_engine_cs.c
index 7681d17ce454..f3cb7e931317 100644
--- a/drivers/gpu/drm/i915/intel_engine_cs.c
+++ b/drivers/gpu/drm/i915/intel_engine_cs.c
@@ -265,6 +265,7 @@ void intel_engine_init_global_seqno(struct intel_engine_cs 
*engine, u32 seqno)
struct drm_i915_private *dev_priv = engine->i915;
 
GEM_BUG_ON(!intel_engine_is_idle(engine));
+   GEM_BUG_ON(i915_gem_active_isset(>timeline->last_request));
 
/* Our semaphore implementation is strictly monotonic (i.e. we proceed
 * so long as the semaphore value in the register/page is greater
@@ -284,9 +285,6 @@ void intel_engine_init_global_seqno(struct intel_engine_cs 
*engine, u32 seqno)
intel_write_status_page(engine, I915_GEM_HWS_INDEX, seqno);
clear_bit(ENGINE_IRQ_BREADCRUMB, >irq_posted);
 
-   GEM_BUG_ON(i915_gem_active_isset(>timeline->last_request));
-   engine->hangcheck.seqno = seqno;
-
/* After manually advancing the seqno, fake the interrupt in case
 * there are any waiters for that seqno.
 */
-- 
2.11.0

___
Intel-gfx mailing list
Intel-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/intel-gfx