All of lore.kernel.org
 help / color / mirror / Atom feed
* [PATCH v2] drm/panfrost: Sync IRQ by job's timeout handler
@ 2023-07-23  0:01 ` Dmitry Osipenko
  0 siblings, 0 replies; 8+ messages in thread
From: Dmitry Osipenko @ 2023-07-23  0:01 UTC (permalink / raw)
  To: Rob Herring, Steven Price, Boris Brezillon
  Cc: kernel, linux-kernel, dri-devel

Panfrost IRQ handler may stuck for a long time, for example this happens
when there is a bad HDMI connection and HDMI handler takes a long time to
finish processing, holding Panfrost. Make Panfrost's job timeout handler
to sync IRQ before checking fence signal status in order to prevent
spurious job timeouts due to a slow IRQ processing.

Signed-off-by: Dmitry Osipenko <dmitry.osipenko@collabora.com>
---

Changelog:

v2: - Moved synchronize_irq() after first signal-check to avoid unnecessary
      blocking on syncing.

    - Added warn message about high interrupt latency.

 drivers/gpu/drm/panfrost/panfrost_job.c | 7 +++++++
 1 file changed, 7 insertions(+)

diff --git a/drivers/gpu/drm/panfrost/panfrost_job.c b/drivers/gpu/drm/panfrost/panfrost_job.c
index dbc597ab46fb..a7663d7847a2 100644
--- a/drivers/gpu/drm/panfrost/panfrost_job.c
+++ b/drivers/gpu/drm/panfrost/panfrost_job.c
@@ -720,6 +720,13 @@ static enum drm_gpu_sched_stat panfrost_job_timedout(struct drm_sched_job
 	if (dma_fence_is_signaled(job->done_fence))
 		return DRM_GPU_SCHED_STAT_NOMINAL;
 
+	synchronize_irq(pfdev->js->irq);
+
+	if (dma_fence_is_signaled(job->done_fence)) {
+		dev_warn(pfdev->dev, "unexpectedly high interrupt latency\n");
+		return DRM_GPU_SCHED_STAT_NOMINAL;
+	}
+
 	dev_err(pfdev->dev, "gpu sched timeout, js=%d, config=0x%x, status=0x%x, head=0x%x, tail=0x%x, sched_job=%p",
 		js,
 		job_read(pfdev, JS_CONFIG(js)),
-- 
2.41.0


^ permalink raw reply related	[flat|nested] 8+ messages in thread

* [PATCH v2] drm/panfrost: Sync IRQ by job's timeout handler
@ 2023-07-23  0:01 ` Dmitry Osipenko
  0 siblings, 0 replies; 8+ messages in thread
From: Dmitry Osipenko @ 2023-07-23  0:01 UTC (permalink / raw)
  To: Rob Herring, Steven Price, Boris Brezillon
  Cc: dri-devel, linux-kernel, kernel

Panfrost IRQ handler may stuck for a long time, for example this happens
when there is a bad HDMI connection and HDMI handler takes a long time to
finish processing, holding Panfrost. Make Panfrost's job timeout handler
to sync IRQ before checking fence signal status in order to prevent
spurious job timeouts due to a slow IRQ processing.

Signed-off-by: Dmitry Osipenko <dmitry.osipenko@collabora.com>
---

Changelog:

v2: - Moved synchronize_irq() after first signal-check to avoid unnecessary
      blocking on syncing.

    - Added warn message about high interrupt latency.

 drivers/gpu/drm/panfrost/panfrost_job.c | 7 +++++++
 1 file changed, 7 insertions(+)

diff --git a/drivers/gpu/drm/panfrost/panfrost_job.c b/drivers/gpu/drm/panfrost/panfrost_job.c
index dbc597ab46fb..a7663d7847a2 100644
--- a/drivers/gpu/drm/panfrost/panfrost_job.c
+++ b/drivers/gpu/drm/panfrost/panfrost_job.c
@@ -720,6 +720,13 @@ static enum drm_gpu_sched_stat panfrost_job_timedout(struct drm_sched_job
 	if (dma_fence_is_signaled(job->done_fence))
 		return DRM_GPU_SCHED_STAT_NOMINAL;
 
+	synchronize_irq(pfdev->js->irq);
+
+	if (dma_fence_is_signaled(job->done_fence)) {
+		dev_warn(pfdev->dev, "unexpectedly high interrupt latency\n");
+		return DRM_GPU_SCHED_STAT_NOMINAL;
+	}
+
 	dev_err(pfdev->dev, "gpu sched timeout, js=%d, config=0x%x, status=0x%x, head=0x%x, tail=0x%x, sched_job=%p",
 		js,
 		job_read(pfdev, JS_CONFIG(js)),
-- 
2.41.0


^ permalink raw reply related	[flat|nested] 8+ messages in thread

* Re: [PATCH v2] drm/panfrost: Sync IRQ by job's timeout handler
  2023-07-23  0:01 ` Dmitry Osipenko
@ 2023-07-24  6:57   ` Boris Brezillon
  -1 siblings, 0 replies; 8+ messages in thread
From: Boris Brezillon @ 2023-07-24  6:57 UTC (permalink / raw)
  To: Dmitry Osipenko; +Cc: kernel, linux-kernel, dri-devel, Steven Price

On Sun, 23 Jul 2023 03:01:42 +0300
Dmitry Osipenko <dmitry.osipenko@collabora.com> wrote:

> Panfrost IRQ handler may stuck for a long time, for example this happens
> when there is a bad HDMI connection and HDMI handler takes a long time to
> finish processing, holding Panfrost. Make Panfrost's job timeout handler
> to sync IRQ before checking fence signal status in order to prevent
> spurious job timeouts due to a slow IRQ processing.
> 
> Signed-off-by: Dmitry Osipenko <dmitry.osipenko@collabora.com>
> ---
> 
> Changelog:
> 
> v2: - Moved synchronize_irq() after first signal-check to avoid unnecessary
>       blocking on syncing.
> 
>     - Added warn message about high interrupt latency.
> 
>  drivers/gpu/drm/panfrost/panfrost_job.c | 7 +++++++
>  1 file changed, 7 insertions(+)
> 
> diff --git a/drivers/gpu/drm/panfrost/panfrost_job.c b/drivers/gpu/drm/panfrost/panfrost_job.c
> index dbc597ab46fb..a7663d7847a2 100644
> --- a/drivers/gpu/drm/panfrost/panfrost_job.c
> +++ b/drivers/gpu/drm/panfrost/panfrost_job.c
> @@ -720,6 +720,13 @@ static enum drm_gpu_sched_stat panfrost_job_timedout(struct drm_sched_job
>  	if (dma_fence_is_signaled(job->done_fence))
>  		return DRM_GPU_SCHED_STAT_NOMINAL;
>  
> +	synchronize_irq(pfdev->js->irq);

Can we add a comment here explaining why we're doing that?

> +
> +	if (dma_fence_is_signaled(job->done_fence)) {
> +		dev_warn(pfdev->dev, "unexpectedly high interrupt latency\n");
> +		return DRM_GPU_SCHED_STAT_NOMINAL;
> +	}
> +
>  	dev_err(pfdev->dev, "gpu sched timeout, js=%d, config=0x%x, status=0x%x, head=0x%x, tail=0x%x, sched_job=%p",
>  		js,
>  		job_read(pfdev, JS_CONFIG(js)),


^ permalink raw reply	[flat|nested] 8+ messages in thread

* Re: [PATCH v2] drm/panfrost: Sync IRQ by job's timeout handler
@ 2023-07-24  6:57   ` Boris Brezillon
  0 siblings, 0 replies; 8+ messages in thread
From: Boris Brezillon @ 2023-07-24  6:57 UTC (permalink / raw)
  To: Dmitry Osipenko
  Cc: Rob Herring, Steven Price, dri-devel, linux-kernel, kernel

On Sun, 23 Jul 2023 03:01:42 +0300
Dmitry Osipenko <dmitry.osipenko@collabora.com> wrote:

> Panfrost IRQ handler may stuck for a long time, for example this happens
> when there is a bad HDMI connection and HDMI handler takes a long time to
> finish processing, holding Panfrost. Make Panfrost's job timeout handler
> to sync IRQ before checking fence signal status in order to prevent
> spurious job timeouts due to a slow IRQ processing.
> 
> Signed-off-by: Dmitry Osipenko <dmitry.osipenko@collabora.com>
> ---
> 
> Changelog:
> 
> v2: - Moved synchronize_irq() after first signal-check to avoid unnecessary
>       blocking on syncing.
> 
>     - Added warn message about high interrupt latency.
> 
>  drivers/gpu/drm/panfrost/panfrost_job.c | 7 +++++++
>  1 file changed, 7 insertions(+)
> 
> diff --git a/drivers/gpu/drm/panfrost/panfrost_job.c b/drivers/gpu/drm/panfrost/panfrost_job.c
> index dbc597ab46fb..a7663d7847a2 100644
> --- a/drivers/gpu/drm/panfrost/panfrost_job.c
> +++ b/drivers/gpu/drm/panfrost/panfrost_job.c
> @@ -720,6 +720,13 @@ static enum drm_gpu_sched_stat panfrost_job_timedout(struct drm_sched_job
>  	if (dma_fence_is_signaled(job->done_fence))
>  		return DRM_GPU_SCHED_STAT_NOMINAL;
>  
> +	synchronize_irq(pfdev->js->irq);

Can we add a comment here explaining why we're doing that?

> +
> +	if (dma_fence_is_signaled(job->done_fence)) {
> +		dev_warn(pfdev->dev, "unexpectedly high interrupt latency\n");
> +		return DRM_GPU_SCHED_STAT_NOMINAL;
> +	}
> +
>  	dev_err(pfdev->dev, "gpu sched timeout, js=%d, config=0x%x, status=0x%x, head=0x%x, tail=0x%x, sched_job=%p",
>  		js,
>  		job_read(pfdev, JS_CONFIG(js)),


^ permalink raw reply	[flat|nested] 8+ messages in thread

* Re: [PATCH v2] drm/panfrost: Sync IRQ by job's timeout handler
  2023-07-23  0:01 ` Dmitry Osipenko
@ 2023-07-24  8:43   ` AngeloGioacchino Del Regno
  -1 siblings, 0 replies; 8+ messages in thread
From: AngeloGioacchino Del Regno @ 2023-07-24  8:43 UTC (permalink / raw)
  To: Dmitry Osipenko, Rob Herring, Steven Price, Boris Brezillon
  Cc: kernel, linux-kernel, dri-devel

Il 23/07/23 02:01, Dmitry Osipenko ha scritto:
> Panfrost IRQ handler may stuck for a long time, for example this happens
> when there is a bad HDMI connection and HDMI handler takes a long time to
> finish processing, holding Panfrost. Make Panfrost's job timeout handler
> to sync IRQ before checking fence signal status in order to prevent
> spurious job timeouts due to a slow IRQ processing.
> 
> Signed-off-by: Dmitry Osipenko <dmitry.osipenko@collabora.com>

Reviewed-by: AngeloGioacchino Del Regno <angelogioacchino.delregno@collabora.com>

Also, on MediaTek MT8192 and MT8195 Chromebooks:

Tested-by: AngeloGioacchino Del Regno <angelogioacchino.delregno@collabora.com>


^ permalink raw reply	[flat|nested] 8+ messages in thread

* Re: [PATCH v2] drm/panfrost: Sync IRQ by job's timeout handler
@ 2023-07-24  8:43   ` AngeloGioacchino Del Regno
  0 siblings, 0 replies; 8+ messages in thread
From: AngeloGioacchino Del Regno @ 2023-07-24  8:43 UTC (permalink / raw)
  To: Dmitry Osipenko, Rob Herring, Steven Price, Boris Brezillon
  Cc: dri-devel, linux-kernel, kernel

Il 23/07/23 02:01, Dmitry Osipenko ha scritto:
> Panfrost IRQ handler may stuck for a long time, for example this happens
> when there is a bad HDMI connection and HDMI handler takes a long time to
> finish processing, holding Panfrost. Make Panfrost's job timeout handler
> to sync IRQ before checking fence signal status in order to prevent
> spurious job timeouts due to a slow IRQ processing.
> 
> Signed-off-by: Dmitry Osipenko <dmitry.osipenko@collabora.com>

Reviewed-by: AngeloGioacchino Del Regno <angelogioacchino.delregno@collabora.com>

Also, on MediaTek MT8192 and MT8195 Chromebooks:

Tested-by: AngeloGioacchino Del Regno <angelogioacchino.delregno@collabora.com>


^ permalink raw reply	[flat|nested] 8+ messages in thread

* Re: [PATCH v2] drm/panfrost: Sync IRQ by job's timeout handler
  2023-07-24  6:57   ` Boris Brezillon
@ 2023-07-24  8:51     ` Steven Price
  -1 siblings, 0 replies; 8+ messages in thread
From: Steven Price @ 2023-07-24  8:51 UTC (permalink / raw)
  To: Boris Brezillon, Dmitry Osipenko; +Cc: kernel, linux-kernel, dri-devel

On 24/07/2023 07:57, Boris Brezillon wrote:
> On Sun, 23 Jul 2023 03:01:42 +0300
> Dmitry Osipenko <dmitry.osipenko@collabora.com> wrote:
> 
>> Panfrost IRQ handler may stuck for a long time, for example this happens
>> when there is a bad HDMI connection and HDMI handler takes a long time to
>> finish processing, holding Panfrost. Make Panfrost's job timeout handler
>> to sync IRQ before checking fence signal status in order to prevent
>> spurious job timeouts due to a slow IRQ processing.
>>
>> Signed-off-by: Dmitry Osipenko <dmitry.osipenko@collabora.com>
>> ---
>>
>> Changelog:
>>
>> v2: - Moved synchronize_irq() after first signal-check to avoid unnecessary
>>       blocking on syncing.
>>
>>     - Added warn message about high interrupt latency.
>>
>>  drivers/gpu/drm/panfrost/panfrost_job.c | 7 +++++++
>>  1 file changed, 7 insertions(+)
>>
>> diff --git a/drivers/gpu/drm/panfrost/panfrost_job.c b/drivers/gpu/drm/panfrost/panfrost_job.c
>> index dbc597ab46fb..a7663d7847a2 100644
>> --- a/drivers/gpu/drm/panfrost/panfrost_job.c
>> +++ b/drivers/gpu/drm/panfrost/panfrost_job.c
>> @@ -720,6 +720,13 @@ static enum drm_gpu_sched_stat panfrost_job_timedout(struct drm_sched_job
>>  	if (dma_fence_is_signaled(job->done_fence))
>>  		return DRM_GPU_SCHED_STAT_NOMINAL;
>>  
>> +	synchronize_irq(pfdev->js->irq);
> 
> Can we add a comment here explaining why we're doing that?

Agreed, with a comment added:

Reviewed-by: Steven Price <steven.price@arm.com>

>> +
>> +	if (dma_fence_is_signaled(job->done_fence)) {
>> +		dev_warn(pfdev->dev, "unexpectedly high interrupt latency\n");
>> +		return DRM_GPU_SCHED_STAT_NOMINAL;
>> +	}
>> +
>>  	dev_err(pfdev->dev, "gpu sched timeout, js=%d, config=0x%x, status=0x%x, head=0x%x, tail=0x%x, sched_job=%p",
>>  		js,
>>  		job_read(pfdev, JS_CONFIG(js)),
> 


^ permalink raw reply	[flat|nested] 8+ messages in thread

* Re: [PATCH v2] drm/panfrost: Sync IRQ by job's timeout handler
@ 2023-07-24  8:51     ` Steven Price
  0 siblings, 0 replies; 8+ messages in thread
From: Steven Price @ 2023-07-24  8:51 UTC (permalink / raw)
  To: Boris Brezillon, Dmitry Osipenko
  Cc: Rob Herring, dri-devel, linux-kernel, kernel

On 24/07/2023 07:57, Boris Brezillon wrote:
> On Sun, 23 Jul 2023 03:01:42 +0300
> Dmitry Osipenko <dmitry.osipenko@collabora.com> wrote:
> 
>> Panfrost IRQ handler may stuck for a long time, for example this happens
>> when there is a bad HDMI connection and HDMI handler takes a long time to
>> finish processing, holding Panfrost. Make Panfrost's job timeout handler
>> to sync IRQ before checking fence signal status in order to prevent
>> spurious job timeouts due to a slow IRQ processing.
>>
>> Signed-off-by: Dmitry Osipenko <dmitry.osipenko@collabora.com>
>> ---
>>
>> Changelog:
>>
>> v2: - Moved synchronize_irq() after first signal-check to avoid unnecessary
>>       blocking on syncing.
>>
>>     - Added warn message about high interrupt latency.
>>
>>  drivers/gpu/drm/panfrost/panfrost_job.c | 7 +++++++
>>  1 file changed, 7 insertions(+)
>>
>> diff --git a/drivers/gpu/drm/panfrost/panfrost_job.c b/drivers/gpu/drm/panfrost/panfrost_job.c
>> index dbc597ab46fb..a7663d7847a2 100644
>> --- a/drivers/gpu/drm/panfrost/panfrost_job.c
>> +++ b/drivers/gpu/drm/panfrost/panfrost_job.c
>> @@ -720,6 +720,13 @@ static enum drm_gpu_sched_stat panfrost_job_timedout(struct drm_sched_job
>>  	if (dma_fence_is_signaled(job->done_fence))
>>  		return DRM_GPU_SCHED_STAT_NOMINAL;
>>  
>> +	synchronize_irq(pfdev->js->irq);
> 
> Can we add a comment here explaining why we're doing that?

Agreed, with a comment added:

Reviewed-by: Steven Price <steven.price@arm.com>

>> +
>> +	if (dma_fence_is_signaled(job->done_fence)) {
>> +		dev_warn(pfdev->dev, "unexpectedly high interrupt latency\n");
>> +		return DRM_GPU_SCHED_STAT_NOMINAL;
>> +	}
>> +
>>  	dev_err(pfdev->dev, "gpu sched timeout, js=%d, config=0x%x, status=0x%x, head=0x%x, tail=0x%x, sched_job=%p",
>>  		js,
>>  		job_read(pfdev, JS_CONFIG(js)),
> 


^ permalink raw reply	[flat|nested] 8+ messages in thread

end of thread, other threads:[~2023-07-24  8:51 UTC | newest]

Thread overview: 8+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2023-07-23  0:01 [PATCH v2] drm/panfrost: Sync IRQ by job's timeout handler Dmitry Osipenko
2023-07-23  0:01 ` Dmitry Osipenko
2023-07-24  6:57 ` Boris Brezillon
2023-07-24  6:57   ` Boris Brezillon
2023-07-24  8:51   ` Steven Price
2023-07-24  8:51     ` Steven Price
2023-07-24  8:43 ` AngeloGioacchino Del Regno
2023-07-24  8:43   ` AngeloGioacchino Del Regno

This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.