Hi Greg,
This patch introduced a race-condition that was fixed in 6e64d6b3a3c39655de56682ec83e894978d23412 ("drm/v3d: Assign job pointer to NULL before signaling the fence") - already in torvalds/master. Is it possible to push the two patches together? This way we wouldn't break any devices.
If possible, same thing for 5.4.
Best Regards, - Maíra
On 30/01/25 11:01, Greg Kroah-Hartman wrote:
5.10-stable review patch. If anyone has any objections, please let me know.
From: Maíra Canal mcanal@igalia.com
[ Upstream commit e4b5ccd392b92300a2b341705cc4805681094e49 ]
After a job completes, the corresponding pointer in the device must be set to NULL. Failing to do so triggers a warning when unloading the driver, as it appears the job is still active. To prevent this, assign the job pointer to NULL after completing the job, indicating the job has finished.
Fixes: 14d1d1908696 ("drm/v3d: Remove the bad signaled() implementation.") Signed-off-by: Maíra Canal mcanal@igalia.com Reviewed-by: Jose Maria Casanova Crespo jmcasanova@igalia.com Link: https://patchwork.freedesktop.org/patch/msgid/20250113154741.67520-1-mcanal@... Signed-off-by: Sasha Levin sashal@kernel.org
drivers/gpu/drm/v3d/v3d_irq.c | 4 ++++ 1 file changed, 4 insertions(+)
diff --git a/drivers/gpu/drm/v3d/v3d_irq.c b/drivers/gpu/drm/v3d/v3d_irq.c index c88686489b888..22aa02d75c5cc 100644 --- a/drivers/gpu/drm/v3d/v3d_irq.c +++ b/drivers/gpu/drm/v3d/v3d_irq.c @@ -103,6 +103,7 @@ v3d_irq(int irq, void *arg) trace_v3d_bcl_irq(&v3d->drm, fence->seqno); dma_fence_signal(&fence->base);
v3d->bin_job = NULL;@@ -112,6 +113,7 @@ v3d_irq(int irq, void *arg) trace_v3d_rcl_irq(&v3d->drm, fence->seqno); dma_fence_signal(&fence->base);
v3d->render_job = NULL;@@ -121,6 +123,7 @@ v3d_irq(int irq, void *arg) trace_v3d_csd_irq(&v3d->drm, fence->seqno); dma_fence_signal(&fence->base);
v3d->csd_job = NULL;@@ -157,6 +160,7 @@ v3d_hub_irq(int irq, void *arg) trace_v3d_tfu_irq(&v3d->drm, fence->seqno); dma_fence_signal(&fence->base);
v3d->tfu_job = NULL;