On 01.09.2026 15:18, Boris Brezillon wrote:
> On Fri, 28 Aug 2026 21:56:49 +0100
> Adrián Larumbe <[email protected]> wrote:
>
> > During device probe(), failure to do a PM get() will leave the usage_count
> > set to 0, which is the value assigned at device creation time. That means
> > when the autosuspend delay expires, runtime suspend callback won't be
> > invoked, so the device will remain powered on forever.
> >
> > On top of that, failure to call PM put() during device unplug means
> > Panfrost device's PM usage_count increases monotonically for every new
> > module reload.
> >
> > The combined outcome of both of the above was that devfreq OPP transition
> > notifications would be printed all the time, even when no jobs are being
> > submitted. This quickly fills the kernel ring buffer with junk.
> >
> > To fix the above:
> > - Manually adjust the PM refcnt at device probe and removal time.
> > - Ensure pm_runtime_dont_use_autosuspend is called in the wind-down path.
> > - Call pm_runtime_put_autosuspend() when device is ready to accept jobs
> >
> > Signed-off-by: Adrián Larumbe <[email protected]>
> > Fixes: 635430797d3f ("drm/panfrost: Rework runtime PM initialization")
> > Fixes: 876b15d2c88d ("drm/panfrost: Fix module unload")
>
> Because of how things are currently ordered, this won't be backportable
> without patch 2, which doesn't have a Fixes tag, so, either we put the
> fix first, and then we move code, or we make it a single commit.
I think I mentioned it in a previous reply. Maybe it's best to fix the
outstanding issues first
and then moving code into panfrost_{init|fini}().
> > ---
> > drivers/gpu/drm/panfrost/panfrost_device.c | 7 +++++++
> > 1 file changed, 7 insertions(+)
> >
> > diff --git a/drivers/gpu/drm/panfrost/panfrost_device.c
> > b/drivers/gpu/drm/panfrost/panfrost_device.c
> > index 99f7da2180f9..70c8109ea698 100644
> > --- a/drivers/gpu/drm/panfrost/panfrost_device.c
> > +++ b/drivers/gpu/drm/panfrost/panfrost_device.c
> > @@ -322,6 +322,7 @@ int panfrost_device_init(struct panfrost_device *pfdev)
> >
> > pm_runtime_set_active(pfdev->base.dev);
> > pm_runtime_mark_last_busy(pfdev->base.dev);
> > + pm_runtime_get_noresume(pfdev->base.dev);
> > pm_runtime_enable(pfdev->base.dev);
> > pm_runtime_set_autosuspend_delay(pfdev->base.dev, 50); /* ~3 frames */
> > pm_runtime_use_autosuspend(pfdev->base.dev);
> > @@ -334,9 +335,13 @@ int panfrost_device_init(struct panfrost_device *pfdev)
> > if (err < 0)
> > goto out_devreg;
> >
> > + pm_runtime_put_autosuspend(pfdev->base.dev);
> > +
> > return 0;
> >
> > out_devreg:
> > + pm_runtime_dont_use_autosuspend(pfdev->base.dev);
> > + pm_runtime_put_noidle(pfdev->base.dev);
> > pm_runtime_disable(pfdev->base.dev);
> > panfrost_device_disable_hw(pfdev);
> > panfrost_gem_fini(pfdev);
> > @@ -363,7 +368,9 @@ int panfrost_device_init(struct panfrost_device *pfdev)
> > void panfrost_device_fini(struct panfrost_device *pfdev)
> > {
> > pm_runtime_get_sync(pfdev->base.dev);
> > + pm_runtime_dont_use_autosuspend(pfdev->base.dev);
> > pm_runtime_disable(pfdev->base.dev);
> > + pm_runtime_put_noidle(pfdev->base.dev);
> >
> > panfrost_jm_stop_sched_jobs(pfdev);
> > panfrost_device_disable_hw(pfdev);
> >
Adrian Larumbe
