Recent changes to genpd prevents those PM domains being powered-on during
initialization from being powered-off during the boot sequence. Based upon
whether CONFIG_PM_CONFIG_PM_GENERIC_DOMAINS_OF is set of not, genpd relies
on the sync_state mechanism or the genpd_power_off_unused() (which is a
late_initcall_sync), to understand when it's okay to allow these PM domains
to be powered-off.
This new behaviour in genpd has lead to problems on different platforms.
Let's therefore restore the behavior of genpd_power_off_unused().
Moreover, let's introduce GENPD_FLAG_NO_STAY_ON, to allow genpd OF
providers to opt-out from the new behaviour.
Link: https://lore.kernel.org/all/20250701114733.636510-1-ulf.hansson@linaro.org/
Reported-by: Geert Uytterhoeven <geert@linux-m68k.org>
Link: https://lore.kernel.org/all/20250902-rk3576-lockup-regression-v1-1-c4a0c9daeb00@collabora.com/
Reported-by: Nicolas Frattaroli <nicolas.frattaroli@collabora.com>
Fixes: 0e789b491ba0 ("pmdomain: core: Leave powered-on genpds on until sync_state")
Fixes: 13a4b7fb6260 ("pmdomain: core: Leave powered-on genpds on until late_initcall_sync")
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
---
drivers/pmdomain/core.c | 20 ++++++++++++++------
include/linux/pm_domain.h | 7 +++++++
2 files changed, 21 insertions(+), 6 deletions(-)
diff --git a/drivers/pmdomain/core.c b/drivers/pmdomain/core.c
index 0006ab3d0789..61c2277c9ce3 100644
--- a/drivers/pmdomain/core.c
+++ b/drivers/pmdomain/core.c
@@ -187,6 +187,7 @@ static const struct genpd_lock_ops genpd_raw_spin_ops = {
#define genpd_is_opp_table_fw(genpd) (genpd->flags & GENPD_FLAG_OPP_TABLE_FW)
#define genpd_is_dev_name_fw(genpd) (genpd->flags & GENPD_FLAG_DEV_NAME_FW)
#define genpd_is_no_sync_state(genpd) (genpd->flags & GENPD_FLAG_NO_SYNC_STATE)
+#define genpd_is_no_stay_on(genpd) (genpd->flags & GENPD_FLAG_NO_STAY_ON)
static inline bool irq_safe_dev_in_sleep_domain(struct device *dev,
const struct generic_pm_domain *genpd)
@@ -1357,7 +1358,6 @@ static int genpd_runtime_resume(struct device *dev)
return ret;
}
-#ifndef CONFIG_PM_GENERIC_DOMAINS_OF
static bool pd_ignore_unused;
static int __init pd_ignore_unused_setup(char *__unused)
{
@@ -1382,9 +1382,6 @@ static int __init genpd_power_off_unused(void)
mutex_lock(&gpd_list_lock);
list_for_each_entry(genpd, &gpd_list, gpd_list_node) {
- genpd_lock(genpd);
- genpd->stay_on = false;
- genpd_unlock(genpd);
genpd_queue_power_off_work(genpd);
}
@@ -1393,7 +1390,6 @@ static int __init genpd_power_off_unused(void)
return 0;
}
late_initcall_sync(genpd_power_off_unused);
-#endif
#ifdef CONFIG_PM_SLEEP
@@ -2367,6 +2363,18 @@ static void genpd_lock_init(struct generic_pm_domain *genpd)
}
}
+#ifdef CONFIG_PM_GENERIC_DOMAINS_OF
+static void genpd_set_stay_on(struct generic_pm_domain *genpd, bool is_off)
+{
+ genpd->stay_on = !genpd_is_no_stay_on(genpd) && !is_off;
+}
+#else
+static void genpd_set_stay_on(struct generic_pm_domain *genpd, bool is_off)
+{
+ genpd->stay_on = false;
+}
+#endif
+
/**
* pm_genpd_init - Initialize a generic I/O PM domain object.
* @genpd: PM domain object to initialize.
@@ -2392,7 +2400,7 @@ int pm_genpd_init(struct generic_pm_domain *genpd,
INIT_WORK(&genpd->power_off_work, genpd_power_off_work_fn);
atomic_set(&genpd->sd_count, 0);
genpd->status = is_off ? GENPD_STATE_OFF : GENPD_STATE_ON;
- genpd->stay_on = !is_off;
+ genpd_set_stay_on(genpd, is_off);
genpd->sync_state = GENPD_SYNC_STATE_OFF;
genpd->device_count = 0;
genpd->provider = NULL;
diff --git a/include/linux/pm_domain.h b/include/linux/pm_domain.h
index c84edf217819..f67a2cb7d781 100644
--- a/include/linux/pm_domain.h
+++ b/include/linux/pm_domain.h
@@ -115,6 +115,12 @@ struct dev_pm_domain_list {
* genpd provider specific way, likely through a
* parent device node. This flag makes genpd to
* skip its internal support for this.
+ *
+ * GENPD_FLAG_NO_STAY_ON: For genpd OF providers a powered-on PM domain at
+ * initialization is prevented from being
+ * powered-off until the ->sync_state() callback is
+ * invoked. This flag informs genpd to allow a
+ * power-off without waiting for ->sync_state().
*/
#define GENPD_FLAG_PM_CLK (1U << 0)
#define GENPD_FLAG_IRQ_SAFE (1U << 1)
@@ -126,6 +132,7 @@ struct dev_pm_domain_list {
#define GENPD_FLAG_OPP_TABLE_FW (1U << 7)
#define GENPD_FLAG_DEV_NAME_FW (1U << 8)
#define GENPD_FLAG_NO_SYNC_STATE (1U << 9)
+#define GENPD_FLAG_NO_STAY_ON (1U << 10)
enum gpd_status {
GENPD_STATE_ON = 0, /* PM domain is on */
--
2.43.0Hi Ulf,
On Tue, 9 Sept 2025 at 13:11, Ulf Hansson <ulf.hansson@linaro.org> wrote:
> Recent changes to genpd prevents those PM domains being powered-on during
> initialization from being powered-off during the boot sequence. Based upon
> whether CONFIG_PM_CONFIG_PM_GENERIC_DOMAINS_OF is set of not, genpd relies
> on the sync_state mechanism or the genpd_power_off_unused() (which is a
> late_initcall_sync), to understand when it's okay to allow these PM domains
> to be powered-off.
>
> This new behaviour in genpd has lead to problems on different platforms.
> Let's therefore restore the behavior of genpd_power_off_unused().
> Moreover, let's introduce GENPD_FLAG_NO_STAY_ON, to allow genpd OF
> providers to opt-out from the new behaviour.
>
> Link: https://lore.kernel.org/all/20250701114733.636510-1-ulf.hansson@linaro.org/
> Reported-by: Geert Uytterhoeven <geert@linux-m68k.org>
> Link: https://lore.kernel.org/all/20250902-rk3576-lockup-regression-v1-1-c4a0c9daeb00@collabora.com/
> Reported-by: Nicolas Frattaroli <nicolas.frattaroli@collabora.com>
> Fixes: 0e789b491ba0 ("pmdomain: core: Leave powered-on genpds on until sync_state")
> Fixes: 13a4b7fb6260 ("pmdomain: core: Leave powered-on genpds on until late_initcall_sync")
> Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
Thanks for your patch!
Reviewed-by: Geert Uytterhoeven <geert+renesas@glider.be>
Tested-by: Geert Uytterhoeven <geert+renesas@glider.be>
> --- a/include/linux/pm_domain.h
> +++ b/include/linux/pm_domain.h
> @@ -115,6 +115,12 @@ struct dev_pm_domain_list {
> * genpd provider specific way, likely through a
> * parent device node. This flag makes genpd to
> * skip its internal support for this.
> + *
> + * GENPD_FLAG_NO_STAY_ON: For genpd OF providers a powered-on PM domain at
> + * initialization is prevented from being
> + * powered-off until the ->sync_state() callback is
> + * invoked. This flag informs genpd to allow a
> + * power-off without waiting for ->sync_state().
This also restores power-down of pmdomains after a failed device
probe (due to a real issue, or just -EPROBE_DEFER), possibly
interfering with other devices that are part of the same pmdomain(s)
but haven't been probed yet. E.g. what if your serial console is
part of the same pmdomain? Probably the pmdomain(s) should not
be powered down immediately, but only later, when either sync state
or genpd_power_off_unused() kicks in.
But this is a pre-existing issue, so not a blocked for this patch.
> */
> #define GENPD_FLAG_PM_CLK (1U << 0)
> #define GENPD_FLAG_IRQ_SAFE (1U << 1)
Gr{oetje,eeting}s,
Geert
--
Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- geert@linux-m68k.org
In personal conversations with technical people, I call myself a hacker. But
when I'm talking to journalists I just say "programmer" or something like that.
-- Linus Torvalds
On Thu, 11 Sept 2025 at 09:56, Geert Uytterhoeven <geert@linux-m68k.org> wrote:
>
> Hi Ulf,
>
> On Tue, 9 Sept 2025 at 13:11, Ulf Hansson <ulf.hansson@linaro.org> wrote:
> > Recent changes to genpd prevents those PM domains being powered-on during
> > initialization from being powered-off during the boot sequence. Based upon
> > whether CONFIG_PM_CONFIG_PM_GENERIC_DOMAINS_OF is set of not, genpd relies
> > on the sync_state mechanism or the genpd_power_off_unused() (which is a
> > late_initcall_sync), to understand when it's okay to allow these PM domains
> > to be powered-off.
> >
> > This new behaviour in genpd has lead to problems on different platforms.
> > Let's therefore restore the behavior of genpd_power_off_unused().
> > Moreover, let's introduce GENPD_FLAG_NO_STAY_ON, to allow genpd OF
> > providers to opt-out from the new behaviour.
> >
> > Link: https://lore.kernel.org/all/20250701114733.636510-1-ulf.hansson@linaro.org/
> > Reported-by: Geert Uytterhoeven <geert@linux-m68k.org>
> > Link: https://lore.kernel.org/all/20250902-rk3576-lockup-regression-v1-1-c4a0c9daeb00@collabora.com/
> > Reported-by: Nicolas Frattaroli <nicolas.frattaroli@collabora.com>
> > Fixes: 0e789b491ba0 ("pmdomain: core: Leave powered-on genpds on until sync_state")
> > Fixes: 13a4b7fb6260 ("pmdomain: core: Leave powered-on genpds on until late_initcall_sync")
> > Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
>
> Thanks for your patch!
>
> Reviewed-by: Geert Uytterhoeven <geert+renesas@glider.be>
> Tested-by: Geert Uytterhoeven <geert+renesas@glider.be>
>
> > --- a/include/linux/pm_domain.h
> > +++ b/include/linux/pm_domain.h
> > @@ -115,6 +115,12 @@ struct dev_pm_domain_list {
> > * genpd provider specific way, likely through a
> > * parent device node. This flag makes genpd to
> > * skip its internal support for this.
> > + *
> > + * GENPD_FLAG_NO_STAY_ON: For genpd OF providers a powered-on PM domain at
> > + * initialization is prevented from being
> > + * powered-off until the ->sync_state() callback is
> > + * invoked. This flag informs genpd to allow a
> > + * power-off without waiting for ->sync_state().
>
> This also restores power-down of pmdomains after a failed device
> probe (due to a real issue, or just -EPROBE_DEFER), possibly
> interfering with other devices that are part of the same pmdomain(s)
> but haven't been probed yet. E.g. what if your serial console is
> part of the same pmdomain? Probably the pmdomain(s) should not
> be powered down immediately, but only later, when either sync state
> or genpd_power_off_unused() kicks in.
That is exactly one of the problems we solve with the new sync_state
support in genpd.
If you have this problem for one of the PM domains, we should not set
GENPD_FLAG_NO_STAY_ON for it. In fact, that's exactly why I added this
flag, to allow it to be set on a per genpd basis. Otherwise we might
as well have used of_genpd_sync_state().
Down the road, if we can improve the sync_state support in genpd,
especially a more fine-grained control for those providers that use
of_genpd_add_provider_onecell(), I hope we should be able to remove
this flag.
[...]
Kind regards
Uffe
Am Dienstag, 9. September 2025, 13:11:20 Mitteleuropäische Sommerzeit schrieb Ulf Hansson:
> Recent changes to genpd prevents those PM domains being powered-on during
> initialization from being powered-off during the boot sequence. Based upon
> whether CONFIG_PM_CONFIG_PM_GENERIC_DOMAINS_OF is set of not, genpd relies
> on the sync_state mechanism or the genpd_power_off_unused() (which is a
> late_initcall_sync), to understand when it's okay to allow these PM domains
> to be powered-off.
>
> This new behaviour in genpd has lead to problems on different platforms.
> Let's therefore restore the behavior of genpd_power_off_unused().
> Moreover, let's introduce GENPD_FLAG_NO_STAY_ON, to allow genpd OF
> providers to opt-out from the new behaviour.
>
> Link: https://lore.kernel.org/all/20250701114733.636510-1-ulf.hansson@linaro.org/
> Reported-by: Geert Uytterhoeven <geert@linux-m68k.org>
> Link: https://lore.kernel.org/all/20250902-rk3576-lockup-regression-v1-1-c4a0c9daeb00@collabora.com/
> Reported-by: Nicolas Frattaroli <nicolas.frattaroli@collabora.com>
> Fixes: 0e789b491ba0 ("pmdomain: core: Leave powered-on genpds on until sync_state")
> Fixes: 13a4b7fb6260 ("pmdomain: core: Leave powered-on genpds on until late_initcall_sync")
> Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
NPU on rk3588 now again correctly gets its supplying regulator and
running said npu does not end up in a kernel panic :-)
Tested-by: Heiko Stuebner <heiko@sntech.de>
© 2016 - 2025 Red Hat, Inc.