Skip to contentSkip to conversation

drm/panfrost: Sync IRQ by job's timeout handler

PATCH Superseded Dmitry Osipenko 4 participants 4 messages Jul 23, 2023 – Jul 24, 2023

Subsystems (3)

v2 of 4 Superseded Latest is v4 →

lore
Share & tools
lore
#0 lore raw

PATCH v2 1/1 drm/panfrost: Sync IRQ by job's timeout handler

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.

---

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(+)
drivers/gpu/drm/panfrost/panfrost_job.c +7 -0
index dbc597ab46fb..a7663d7847a2 100644
@@ -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)),
Signature · 1 line
2.41.0
maintainer
#1 lore raw
Quoted 24 lines · On Sun, 23 Jul 2023 03:01:42 +0300 Dmitry Osipenko…

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.

---

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(+)
drivers/gpu/drm/panfrost/panfrost_job.c +0 -0
index dbc597ab46fb..a7663d7847a2 100644

drivers/gpu/drm/panfrost/panfrost_job.c static enum drm_gpu_sched_stat panfrost_job_timedout(struct drm_sched_job · line 723

Quoted 4 lines
drivers/gpu/drm/panfrost/panfrost_job.c +1 -0
@@ -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?

Quoted 9 lines
drivers/gpu/drm/panfrost/panfrost_job.c +6 -0
+
+	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)),
maintainer Reviewed-by
#2 lore raw
Quoted 30 lines · Boris Brezillon
Quoted 27 lines · On Sun, 23 Jul 2023 03:01:42 +0300 Dmitry Osipenko…

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.

---

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(+)
drivers/gpu/drm/panfrost/panfrost_job.c +0 -0
index dbc597ab46fb..a7663d7847a2 100644
@@ -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;
 
Quoted 3 lines

drivers/gpu/drm/panfrost/panfrost_job.c static enum drm_gpu_sched_stat panfrost_job_timedout(struct drm_sched_job · line 723

Quoted 1 line
drivers/gpu/drm/panfrost/panfrost_job.c +1 -0
+	synchronize_irq(pfdev->js->irq);

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

Agreed, with a comment added:

Quoted 10 lines
Quoted 9 lines
drivers/gpu/drm/panfrost/panfrost_job.c +6 -0
+
+	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)),
Reviewed-by Tested-by on v2 1/1
#3 lore raw
In reply to Dmitry Osipenko

Il 23/07/23 02:01, Dmitry Osipenko ha scritto:

Quoted 7 lines

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.

Also, on MediaTek MT8192 and MT8195 Chromebooks: