Message ID | 1548846984-2044-1-git-send-email-vincent.guittot@linaro.org (mailing list archive) |
---|---|
State | Superseded |
Delegated to: | Geert Uytterhoeven |
Headers | show |
Series | [v2] PM-runtime: fix deadlock with ktime | expand |
On Wed, 30 Jan 2019 at 12:16, Vincent Guittot <vincent.guittot@linaro.org> wrote: > > A deadlock has been seen when swicthing clocksources which use PM runtime. > The call path is: > change_clocksource > ... > write_seqcount_begin > ... > timekeeping_update > ... > sh_cmt_clocksource_enable > ... > rpm_resume Hmm, isn't this already a problem in genpd then? In genpd's ->runtime_resume() callback (genpd_runtime_resume()) we call ktime_get() to measure the latency of running the callback. Or, perhaps irq_safe_dev_in_no_sleep_domain() returns true for the genpd + device in question, which means the ktime thingy is skipped. Geert? > pm_runtime_mark_last_busy > ktime_get > do > read_seqcount_begin > while read_seqcount_retry > .... > write_seqcount_end > > Although we should be safe because we haven't yet changed the clocksource > at that time, we can't because of seqcount protection. > > Use ktime_get_mono_fast_ns() instead which is lock safe for such case > > With ktime_get_mono_fast_ns, the timestamp is not guaranteed to be > monotonic across an update and as a result can goes backward. According to > update_fast_timekeeper() description: "In the worst case, this can > result is a slightly wrong timestamp (a few nanoseconds)". For > PM runtime autosuspend, this means only that the suspend decision can > be slightly sub optimal. > > Fixes: 8234f6734c5d ("PM-runtime: Switch autosuspend over to using hrtimers") > Reported-by: Biju Das <biju.das@bp.renesas.com> > Signed-off-by: Vincent Guittot <vincent.guittot@linaro.org> Reviewed-by: Ulf Hansson <ulf.hansson@linaro.org> Kind regards Uffe > --- > > - v2: Updated commit message to explain the impact of using > ktime_get_mono_fast_ns() > > drivers/base/power/runtime.c | 10 +++++----- > include/linux/pm_runtime.h | 2 +- > 2 files changed, 6 insertions(+), 6 deletions(-) > > diff --git a/drivers/base/power/runtime.c b/drivers/base/power/runtime.c > index 457be03..708a13f 100644 > --- a/drivers/base/power/runtime.c > +++ b/drivers/base/power/runtime.c > @@ -130,7 +130,7 @@ u64 pm_runtime_autosuspend_expiration(struct device *dev) > { > int autosuspend_delay; > u64 last_busy, expires = 0; > - u64 now = ktime_to_ns(ktime_get()); > + u64 now = ktime_get_mono_fast_ns(); > > if (!dev->power.use_autosuspend) > goto out; > @@ -909,7 +909,7 @@ static enum hrtimer_restart pm_suspend_timer_fn(struct hrtimer *timer) > * If 'expires' is after the current time, we've been called > * too early. > */ > - if (expires > 0 && expires < ktime_to_ns(ktime_get())) { > + if (expires > 0 && expires < ktime_get_mono_fast_ns()) { > dev->power.timer_expires = 0; > rpm_suspend(dev, dev->power.timer_autosuspends ? > (RPM_ASYNC | RPM_AUTO) : RPM_ASYNC); > @@ -928,7 +928,7 @@ static enum hrtimer_restart pm_suspend_timer_fn(struct hrtimer *timer) > int pm_schedule_suspend(struct device *dev, unsigned int delay) > { > unsigned long flags; > - ktime_t expires; > + u64 expires; > int retval; > > spin_lock_irqsave(&dev->power.lock, flags); > @@ -945,8 +945,8 @@ int pm_schedule_suspend(struct device *dev, unsigned int delay) > /* Other scheduled or pending requests need to be canceled. */ > pm_runtime_cancel_pending(dev); > > - expires = ktime_add(ktime_get(), ms_to_ktime(delay)); > - dev->power.timer_expires = ktime_to_ns(expires); > + expires = ktime_get_mono_fast_ns() + (u64)delay * NSEC_PER_MSEC); > + dev->power.timer_expires = expires; > dev->power.timer_autosuspends = 0; > hrtimer_start(&dev->power.suspend_timer, expires, HRTIMER_MODE_ABS); > > diff --git a/include/linux/pm_runtime.h b/include/linux/pm_runtime.h > index 54af4ee..fed5be7 100644 > --- a/include/linux/pm_runtime.h > +++ b/include/linux/pm_runtime.h > @@ -105,7 +105,7 @@ static inline bool pm_runtime_callbacks_present(struct device *dev) > > static inline void pm_runtime_mark_last_busy(struct device *dev) > { > - WRITE_ONCE(dev->power.last_busy, ktime_to_ns(ktime_get())); > + WRITE_ONCE(dev->power.last_busy, ktime_get_mono_fast_ns()); > } > > static inline bool pm_runtime_is_irq_safe(struct device *dev) > -- > 2.7.4 >
On Wed, Jan 30, 2019 at 12:16 PM Vincent Guittot <vincent.guittot@linaro.org> wrote: > > A deadlock has been seen when swicthing clocksources which use PM runtime. > The call path is: > change_clocksource > ... > write_seqcount_begin > ... > timekeeping_update > ... > sh_cmt_clocksource_enable > ... > rpm_resume > pm_runtime_mark_last_busy > ktime_get > do > read_seqcount_begin > while read_seqcount_retry > .... > write_seqcount_end > > Although we should be safe because we haven't yet changed the clocksource > at that time, we can't because of seqcount protection. > > Use ktime_get_mono_fast_ns() instead which is lock safe for such case > > With ktime_get_mono_fast_ns, the timestamp is not guaranteed to be > monotonic across an update and as a result can goes backward. According to > update_fast_timekeeper() description: "In the worst case, this can > result is a slightly wrong timestamp (a few nanoseconds)". For > PM runtime autosuspend, this means only that the suspend decision can > be slightly sub optimal. > > Fixes: 8234f6734c5d ("PM-runtime: Switch autosuspend over to using hrtimers") > Reported-by: Biju Das <biju.das@bp.renesas.com> > Signed-off-by: Vincent Guittot <vincent.guittot@linaro.org> I've queued this one up as a fix for 5.0, but unfortunately it clashes with the patch from Ladislav Michl at https://patchwork.kernel.org/patch/10755477/ which has been dropped now. Can you or Ladislav please rebase that patch on top of this one and repost? > --- > > - v2: Updated commit message to explain the impact of using > ktime_get_mono_fast_ns() > > drivers/base/power/runtime.c | 10 +++++----- > include/linux/pm_runtime.h | 2 +- > 2 files changed, 6 insertions(+), 6 deletions(-) > > diff --git a/drivers/base/power/runtime.c b/drivers/base/power/runtime.c > index 457be03..708a13f 100644 > --- a/drivers/base/power/runtime.c > +++ b/drivers/base/power/runtime.c > @@ -130,7 +130,7 @@ u64 pm_runtime_autosuspend_expiration(struct device *dev) > { > int autosuspend_delay; > u64 last_busy, expires = 0; > - u64 now = ktime_to_ns(ktime_get()); > + u64 now = ktime_get_mono_fast_ns(); > > if (!dev->power.use_autosuspend) > goto out; > @@ -909,7 +909,7 @@ static enum hrtimer_restart pm_suspend_timer_fn(struct hrtimer *timer) > * If 'expires' is after the current time, we've been called > * too early. > */ > - if (expires > 0 && expires < ktime_to_ns(ktime_get())) { > + if (expires > 0 && expires < ktime_get_mono_fast_ns()) { > dev->power.timer_expires = 0; > rpm_suspend(dev, dev->power.timer_autosuspends ? > (RPM_ASYNC | RPM_AUTO) : RPM_ASYNC); > @@ -928,7 +928,7 @@ static enum hrtimer_restart pm_suspend_timer_fn(struct hrtimer *timer) > int pm_schedule_suspend(struct device *dev, unsigned int delay) > { > unsigned long flags; > - ktime_t expires; > + u64 expires; > int retval; > > spin_lock_irqsave(&dev->power.lock, flags); > @@ -945,8 +945,8 @@ int pm_schedule_suspend(struct device *dev, unsigned int delay) > /* Other scheduled or pending requests need to be canceled. */ > pm_runtime_cancel_pending(dev); > > - expires = ktime_add(ktime_get(), ms_to_ktime(delay)); > - dev->power.timer_expires = ktime_to_ns(expires); > + expires = ktime_get_mono_fast_ns() + (u64)delay * NSEC_PER_MSEC); > + dev->power.timer_expires = expires; > dev->power.timer_autosuspends = 0; > hrtimer_start(&dev->power.suspend_timer, expires, HRTIMER_MODE_ABS); > > diff --git a/include/linux/pm_runtime.h b/include/linux/pm_runtime.h > index 54af4ee..fed5be7 100644 > --- a/include/linux/pm_runtime.h > +++ b/include/linux/pm_runtime.h > @@ -105,7 +105,7 @@ static inline bool pm_runtime_callbacks_present(struct device *dev) > > static inline void pm_runtime_mark_last_busy(struct device *dev) > { > - WRITE_ONCE(dev->power.last_busy, ktime_to_ns(ktime_get())); > + WRITE_ONCE(dev->power.last_busy, ktime_get_mono_fast_ns()); > } > > static inline bool pm_runtime_is_irq_safe(struct device *dev) > -- > 2.7.4 >
On Wed, 30 Jan 2019 at 14:06, Rafael J. Wysocki <rafael@kernel.org> wrote: > > On Wed, Jan 30, 2019 at 12:16 PM Vincent Guittot > <vincent.guittot@linaro.org> wrote: > > > > A deadlock has been seen when swicthing clocksources which use PM runtime. > > The call path is: > > change_clocksource > > ... > > write_seqcount_begin > > ... > > timekeeping_update > > ... > > sh_cmt_clocksource_enable > > ... > > rpm_resume > > pm_runtime_mark_last_busy > > ktime_get > > do > > read_seqcount_begin > > while read_seqcount_retry > > .... > > write_seqcount_end > > > > Although we should be safe because we haven't yet changed the clocksource > > at that time, we can't because of seqcount protection. > > > > Use ktime_get_mono_fast_ns() instead which is lock safe for such case > > > > With ktime_get_mono_fast_ns, the timestamp is not guaranteed to be > > monotonic across an update and as a result can goes backward. According to > > update_fast_timekeeper() description: "In the worst case, this can > > result is a slightly wrong timestamp (a few nanoseconds)". For > > PM runtime autosuspend, this means only that the suspend decision can > > be slightly sub optimal. > > > > Fixes: 8234f6734c5d ("PM-runtime: Switch autosuspend over to using hrtimers") > > Reported-by: Biju Das <biju.das@bp.renesas.com> > > Signed-off-by: Vincent Guittot <vincent.guittot@linaro.org> > > I've queued this one up as a fix for 5.0, but unfortunately it clashes > with the patch from Ladislav Michl at > https://patchwork.kernel.org/patch/10755477/ which has been dropped > now. Thanks for adding Ladislav in this thread. I'm sorry I forgot to add him in the loop. > > Can you or Ladislav please rebase that patch on top of this one and repost? Ladislav, Let me know if you prefer to rebase and repost your patch of if you want me to do. Regards, Vincent > > > --- > > > > - v2: Updated commit message to explain the impact of using > > ktime_get_mono_fast_ns() > > > > drivers/base/power/runtime.c | 10 +++++----- > > include/linux/pm_runtime.h | 2 +- > > 2 files changed, 6 insertions(+), 6 deletions(-) > > > > diff --git a/drivers/base/power/runtime.c b/drivers/base/power/runtime.c > > index 457be03..708a13f 100644 > > --- a/drivers/base/power/runtime.c > > +++ b/drivers/base/power/runtime.c > > @@ -130,7 +130,7 @@ u64 pm_runtime_autosuspend_expiration(struct device *dev) > > { > > int autosuspend_delay; > > u64 last_busy, expires = 0; > > - u64 now = ktime_to_ns(ktime_get()); > > + u64 now = ktime_get_mono_fast_ns(); > > > > if (!dev->power.use_autosuspend) > > goto out; > > @@ -909,7 +909,7 @@ static enum hrtimer_restart pm_suspend_timer_fn(struct hrtimer *timer) > > * If 'expires' is after the current time, we've been called > > * too early. > > */ > > - if (expires > 0 && expires < ktime_to_ns(ktime_get())) { > > + if (expires > 0 && expires < ktime_get_mono_fast_ns()) { > > dev->power.timer_expires = 0; > > rpm_suspend(dev, dev->power.timer_autosuspends ? > > (RPM_ASYNC | RPM_AUTO) : RPM_ASYNC); > > @@ -928,7 +928,7 @@ static enum hrtimer_restart pm_suspend_timer_fn(struct hrtimer *timer) > > int pm_schedule_suspend(struct device *dev, unsigned int delay) > > { > > unsigned long flags; > > - ktime_t expires; > > + u64 expires; > > int retval; > > > > spin_lock_irqsave(&dev->power.lock, flags); > > @@ -945,8 +945,8 @@ int pm_schedule_suspend(struct device *dev, unsigned int delay) > > /* Other scheduled or pending requests need to be canceled. */ > > pm_runtime_cancel_pending(dev); > > > > - expires = ktime_add(ktime_get(), ms_to_ktime(delay)); > > - dev->power.timer_expires = ktime_to_ns(expires); > > + expires = ktime_get_mono_fast_ns() + (u64)delay * NSEC_PER_MSEC); > > + dev->power.timer_expires = expires; > > dev->power.timer_autosuspends = 0; > > hrtimer_start(&dev->power.suspend_timer, expires, HRTIMER_MODE_ABS); > > > > diff --git a/include/linux/pm_runtime.h b/include/linux/pm_runtime.h > > index 54af4ee..fed5be7 100644 > > --- a/include/linux/pm_runtime.h > > +++ b/include/linux/pm_runtime.h > > @@ -105,7 +105,7 @@ static inline bool pm_runtime_callbacks_present(struct device *dev) > > > > static inline void pm_runtime_mark_last_busy(struct device *dev) > > { > > - WRITE_ONCE(dev->power.last_busy, ktime_to_ns(ktime_get())); > > + WRITE_ONCE(dev->power.last_busy, ktime_get_mono_fast_ns()); > > } > > > > static inline bool pm_runtime_is_irq_safe(struct device *dev) > > -- > > 2.7.4 > >
On Wed, Jan 30, 2019 at 02:18:49PM +0100, Vincent Guittot wrote: > On Wed, 30 Jan 2019 at 14:06, Rafael J. Wysocki <rafael@kernel.org> wrote: > > > > On Wed, Jan 30, 2019 at 12:16 PM Vincent Guittot > > <vincent.guittot@linaro.org> wrote: > > > > > > A deadlock has been seen when swicthing clocksources which use PM runtime. > > > The call path is: > > > change_clocksource > > > ... > > > write_seqcount_begin > > > ... > > > timekeeping_update > > > ... > > > sh_cmt_clocksource_enable > > > ... > > > rpm_resume > > > pm_runtime_mark_last_busy > > > ktime_get > > > do > > > read_seqcount_begin > > > while read_seqcount_retry > > > .... > > > write_seqcount_end > > > > > > Although we should be safe because we haven't yet changed the clocksource > > > at that time, we can't because of seqcount protection. > > > > > > Use ktime_get_mono_fast_ns() instead which is lock safe for such case > > > > > > With ktime_get_mono_fast_ns, the timestamp is not guaranteed to be > > > monotonic across an update and as a result can goes backward. According to > > > update_fast_timekeeper() description: "In the worst case, this can > > > result is a slightly wrong timestamp (a few nanoseconds)". For > > > PM runtime autosuspend, this means only that the suspend decision can > > > be slightly sub optimal. > > > > > > Fixes: 8234f6734c5d ("PM-runtime: Switch autosuspend over to using hrtimers") > > > Reported-by: Biju Das <biju.das@bp.renesas.com> > > > Signed-off-by: Vincent Guittot <vincent.guittot@linaro.org> > > > > I've queued this one up as a fix for 5.0, but unfortunately it clashes > > with the patch from Ladislav Michl at > > https://patchwork.kernel.org/patch/10755477/ which has been dropped > > now. > > Thanks for adding Ladislav in this thread. > I'm sorry I forgot to add him in the loop. > > > > > Can you or Ladislav please rebase that patch on top of this one and repost? > > Ladislav, > > Let me know if you prefer to rebase and repost your patch of if you > want me to do. I'll rebase it on top of Rafael's bleeding-edge branch. Best regards, ladis > Regards, > Vincent > > > > > > --- > > > > > > - v2: Updated commit message to explain the impact of using > > > ktime_get_mono_fast_ns() > > > > > > drivers/base/power/runtime.c | 10 +++++----- > > > include/linux/pm_runtime.h | 2 +- > > > 2 files changed, 6 insertions(+), 6 deletions(-) > > > > > > diff --git a/drivers/base/power/runtime.c b/drivers/base/power/runtime.c > > > index 457be03..708a13f 100644 > > > --- a/drivers/base/power/runtime.c > > > +++ b/drivers/base/power/runtime.c > > > @@ -130,7 +130,7 @@ u64 pm_runtime_autosuspend_expiration(struct device *dev) > > > { > > > int autosuspend_delay; > > > u64 last_busy, expires = 0; > > > - u64 now = ktime_to_ns(ktime_get()); > > > + u64 now = ktime_get_mono_fast_ns(); > > > > > > if (!dev->power.use_autosuspend) > > > goto out; > > > @@ -909,7 +909,7 @@ static enum hrtimer_restart pm_suspend_timer_fn(struct hrtimer *timer) > > > * If 'expires' is after the current time, we've been called > > > * too early. > > > */ > > > - if (expires > 0 && expires < ktime_to_ns(ktime_get())) { > > > + if (expires > 0 && expires < ktime_get_mono_fast_ns()) { > > > dev->power.timer_expires = 0; > > > rpm_suspend(dev, dev->power.timer_autosuspends ? > > > (RPM_ASYNC | RPM_AUTO) : RPM_ASYNC); > > > @@ -928,7 +928,7 @@ static enum hrtimer_restart pm_suspend_timer_fn(struct hrtimer *timer) > > > int pm_schedule_suspend(struct device *dev, unsigned int delay) > > > { > > > unsigned long flags; > > > - ktime_t expires; > > > + u64 expires; > > > int retval; > > > > > > spin_lock_irqsave(&dev->power.lock, flags); > > > @@ -945,8 +945,8 @@ int pm_schedule_suspend(struct device *dev, unsigned int delay) > > > /* Other scheduled or pending requests need to be canceled. */ > > > pm_runtime_cancel_pending(dev); > > > > > > - expires = ktime_add(ktime_get(), ms_to_ktime(delay)); > > > - dev->power.timer_expires = ktime_to_ns(expires); > > > + expires = ktime_get_mono_fast_ns() + (u64)delay * NSEC_PER_MSEC); > > > + dev->power.timer_expires = expires; > > > dev->power.timer_autosuspends = 0; > > > hrtimer_start(&dev->power.suspend_timer, expires, HRTIMER_MODE_ABS); > > > > > > diff --git a/include/linux/pm_runtime.h b/include/linux/pm_runtime.h > > > index 54af4ee..fed5be7 100644 > > > --- a/include/linux/pm_runtime.h > > > +++ b/include/linux/pm_runtime.h > > > @@ -105,7 +105,7 @@ static inline bool pm_runtime_callbacks_present(struct device *dev) > > > > > > static inline void pm_runtime_mark_last_busy(struct device *dev) > > > { > > > - WRITE_ONCE(dev->power.last_busy, ktime_to_ns(ktime_get())); > > > + WRITE_ONCE(dev->power.last_busy, ktime_get_mono_fast_ns()); > > > } > > > > > > static inline bool pm_runtime_is_irq_safe(struct device *dev) > > > -- > > > 2.7.4 > > >
On Wed, Jan 30, 2019 at 02:06:07PM +0100, Rafael J. Wysocki wrote: > On Wed, Jan 30, 2019 at 12:16 PM Vincent Guittot > <vincent.guittot@linaro.org> wrote: > > > > A deadlock has been seen when swicthing clocksources which use PM runtime. > > The call path is: > > change_clocksource > > ... > > write_seqcount_begin > > ... > > timekeeping_update > > ... > > sh_cmt_clocksource_enable > > ... > > rpm_resume > > pm_runtime_mark_last_busy > > ktime_get > > do > > read_seqcount_begin > > while read_seqcount_retry > > .... > > write_seqcount_end > > > > Although we should be safe because we haven't yet changed the clocksource > > at that time, we can't because of seqcount protection. > > > > Use ktime_get_mono_fast_ns() instead which is lock safe for such case > > > > With ktime_get_mono_fast_ns, the timestamp is not guaranteed to be > > monotonic across an update and as a result can goes backward. According to > > update_fast_timekeeper() description: "In the worst case, this can > > result is a slightly wrong timestamp (a few nanoseconds)". For > > PM runtime autosuspend, this means only that the suspend decision can > > be slightly sub optimal. > > > > Fixes: 8234f6734c5d ("PM-runtime: Switch autosuspend over to using hrtimers") > > Reported-by: Biju Das <biju.das@bp.renesas.com> > > Signed-off-by: Vincent Guittot <vincent.guittot@linaro.org> > > I've queued this one up as a fix for 5.0, but unfortunately it clashes > with the patch from Ladislav Michl at > https://patchwork.kernel.org/patch/10755477/ which has been dropped > now. > > Can you or Ladislav please rebase that patch on top of this one and repost? > > > --- > > > > - v2: Updated commit message to explain the impact of using > > ktime_get_mono_fast_ns() > > > > drivers/base/power/runtime.c | 10 +++++----- > > include/linux/pm_runtime.h | 2 +- > > 2 files changed, 6 insertions(+), 6 deletions(-) > > > > diff --git a/drivers/base/power/runtime.c b/drivers/base/power/runtime.c > > index 457be03..708a13f 100644 > > --- a/drivers/base/power/runtime.c > > +++ b/drivers/base/power/runtime.c > > @@ -130,7 +130,7 @@ u64 pm_runtime_autosuspend_expiration(struct device *dev) > > { > > int autosuspend_delay; > > u64 last_busy, expires = 0; > > - u64 now = ktime_to_ns(ktime_get()); > > + u64 now = ktime_get_mono_fast_ns(); > > > > if (!dev->power.use_autosuspend) > > goto out; > > @@ -909,7 +909,7 @@ static enum hrtimer_restart pm_suspend_timer_fn(struct hrtimer *timer) > > * If 'expires' is after the current time, we've been called > > * too early. > > */ > > - if (expires > 0 && expires < ktime_to_ns(ktime_get())) { > > + if (expires > 0 && expires < ktime_get_mono_fast_ns()) { > > dev->power.timer_expires = 0; > > rpm_suspend(dev, dev->power.timer_autosuspends ? > > (RPM_ASYNC | RPM_AUTO) : RPM_ASYNC); > > @@ -928,7 +928,7 @@ static enum hrtimer_restart pm_suspend_timer_fn(struct hrtimer *timer) > > int pm_schedule_suspend(struct device *dev, unsigned int delay) > > { > > unsigned long flags; > > - ktime_t expires; > > + u64 expires; > > int retval; > > > > spin_lock_irqsave(&dev->power.lock, flags); > > @@ -945,8 +945,8 @@ int pm_schedule_suspend(struct device *dev, unsigned int delay) > > /* Other scheduled or pending requests need to be canceled. */ > > pm_runtime_cancel_pending(dev); > > > > - expires = ktime_add(ktime_get(), ms_to_ktime(delay)); > > - dev->power.timer_expires = ktime_to_ns(expires); > > + expires = ktime_get_mono_fast_ns() + (u64)delay * NSEC_PER_MSEC); Just FYI, this one does not compile ^ I'm going to send updated optimization patch anyway as fixing this one will not break it :) > > + dev->power.timer_expires = expires; > > dev->power.timer_autosuspends = 0; > > hrtimer_start(&dev->power.suspend_timer, expires, HRTIMER_MODE_ABS); > > > > diff --git a/include/linux/pm_runtime.h b/include/linux/pm_runtime.h > > index 54af4ee..fed5be7 100644 > > --- a/include/linux/pm_runtime.h > > +++ b/include/linux/pm_runtime.h > > @@ -105,7 +105,7 @@ static inline bool pm_runtime_callbacks_present(struct device *dev) > > > > static inline void pm_runtime_mark_last_busy(struct device *dev) > > { > > - WRITE_ONCE(dev->power.last_busy, ktime_to_ns(ktime_get())); > > + WRITE_ONCE(dev->power.last_busy, ktime_get_mono_fast_ns()); > > } > > > > static inline bool pm_runtime_is_irq_safe(struct device *dev) > > -- > > 2.7.4 > >
Hi Ulf, On Wed, Jan 30, 2019 at 1:15 PM Ulf Hansson <ulf.hansson@linaro.org> wrote: > On Wed, 30 Jan 2019 at 12:16, Vincent Guittot > <vincent.guittot@linaro.org> wrote: > > > > A deadlock has been seen when swicthing clocksources which use PM runtime. > > The call path is: > > change_clocksource > > ... > > write_seqcount_begin > > ... > > timekeeping_update > > ... > > sh_cmt_clocksource_enable > > ... > > rpm_resume > > Hmm, isn't this already a problem in genpd then? > > In genpd's ->runtime_resume() callback (genpd_runtime_resume()) we > call ktime_get() to measure the latency of running the callback. > > Or, perhaps irq_safe_dev_in_no_sleep_domain() returns true for the > genpd + device in question, which means the ktime thingy is skipped. > > Geert? Correct. After adding dev_info(dev, "%s: pm_runtime_is_irq_safe = %d, genpd_is_irq_safe(%s) = %d => %s\n", __func__, pm_runtime_is_irq_safe(dev), genpd->name, genpd_is_irq_safe(genpd), ret ? "true" : "false"); I see on R-Car M2-W: sh_cmt ffca0000.timer: irq_safe_dev_in_no_sleep_domain: pm_runtime_is_irq_safe = 1, genpd_is_irq_safe(always-on) = 0 => true For all other devices, the result is false. Gr{oetje,eeting}s, Geert
On Tue, 5 Feb 2019 at 14:24, Geert Uytterhoeven <geert@linux-m68k.org> wrote: > > Hi Ulf, > > On Wed, Jan 30, 2019 at 1:15 PM Ulf Hansson <ulf.hansson@linaro.org> wrote: > > On Wed, 30 Jan 2019 at 12:16, Vincent Guittot > > <vincent.guittot@linaro.org> wrote: > > > > > > A deadlock has been seen when swicthing clocksources which use PM runtime. > > > The call path is: > > > change_clocksource > > > ... > > > write_seqcount_begin > > > ... > > > timekeeping_update > > > ... > > > sh_cmt_clocksource_enable > > > ... > > > rpm_resume > > > > Hmm, isn't this already a problem in genpd then? > > > > In genpd's ->runtime_resume() callback (genpd_runtime_resume()) we > > call ktime_get() to measure the latency of running the callback. > > > > Or, perhaps irq_safe_dev_in_no_sleep_domain() returns true for the > > genpd + device in question, which means the ktime thingy is skipped. > > > > Geert? > > Correct. After adding > > dev_info(dev, "%s: pm_runtime_is_irq_safe = %d, > genpd_is_irq_safe(%s) = %d => %s\n", __func__, > pm_runtime_is_irq_safe(dev), genpd->name, genpd_is_irq_safe(genpd), > ret ? "true" : "false"); > > I see on R-Car M2-W: > > sh_cmt ffca0000.timer: irq_safe_dev_in_no_sleep_domain: > pm_runtime_is_irq_safe = 1, genpd_is_irq_safe(always-on) = 0 => true > > For all other devices, the result is false. Thanks for confirming! I guess we have been lucky so far. Anyway, if runtime PM core converts to use ktime_get_mono_fast_ns(), we should probably consider to convert genpd to this as well. Don't you think? Kind regards Uffe
diff --git a/drivers/base/power/runtime.c b/drivers/base/power/runtime.c index 457be03..708a13f 100644 --- a/drivers/base/power/runtime.c +++ b/drivers/base/power/runtime.c @@ -130,7 +130,7 @@ u64 pm_runtime_autosuspend_expiration(struct device *dev) { int autosuspend_delay; u64 last_busy, expires = 0; - u64 now = ktime_to_ns(ktime_get()); + u64 now = ktime_get_mono_fast_ns(); if (!dev->power.use_autosuspend) goto out; @@ -909,7 +909,7 @@ static enum hrtimer_restart pm_suspend_timer_fn(struct hrtimer *timer) * If 'expires' is after the current time, we've been called * too early. */ - if (expires > 0 && expires < ktime_to_ns(ktime_get())) { + if (expires > 0 && expires < ktime_get_mono_fast_ns()) { dev->power.timer_expires = 0; rpm_suspend(dev, dev->power.timer_autosuspends ? (RPM_ASYNC | RPM_AUTO) : RPM_ASYNC); @@ -928,7 +928,7 @@ static enum hrtimer_restart pm_suspend_timer_fn(struct hrtimer *timer) int pm_schedule_suspend(struct device *dev, unsigned int delay) { unsigned long flags; - ktime_t expires; + u64 expires; int retval; spin_lock_irqsave(&dev->power.lock, flags); @@ -945,8 +945,8 @@ int pm_schedule_suspend(struct device *dev, unsigned int delay) /* Other scheduled or pending requests need to be canceled. */ pm_runtime_cancel_pending(dev); - expires = ktime_add(ktime_get(), ms_to_ktime(delay)); - dev->power.timer_expires = ktime_to_ns(expires); + expires = ktime_get_mono_fast_ns() + (u64)delay * NSEC_PER_MSEC); + dev->power.timer_expires = expires; dev->power.timer_autosuspends = 0; hrtimer_start(&dev->power.suspend_timer, expires, HRTIMER_MODE_ABS); diff --git a/include/linux/pm_runtime.h b/include/linux/pm_runtime.h index 54af4ee..fed5be7 100644 --- a/include/linux/pm_runtime.h +++ b/include/linux/pm_runtime.h @@ -105,7 +105,7 @@ static inline bool pm_runtime_callbacks_present(struct device *dev) static inline void pm_runtime_mark_last_busy(struct device *dev) { - WRITE_ONCE(dev->power.last_busy, ktime_to_ns(ktime_get())); + WRITE_ONCE(dev->power.last_busy, ktime_get_mono_fast_ns()); } static inline bool pm_runtime_is_irq_safe(struct device *dev)
A deadlock has been seen when swicthing clocksources which use PM runtime. The call path is: change_clocksource ... write_seqcount_begin ... timekeeping_update ... sh_cmt_clocksource_enable ... rpm_resume pm_runtime_mark_last_busy ktime_get do read_seqcount_begin while read_seqcount_retry .... write_seqcount_end Although we should be safe because we haven't yet changed the clocksource at that time, we can't because of seqcount protection. Use ktime_get_mono_fast_ns() instead which is lock safe for such case With ktime_get_mono_fast_ns, the timestamp is not guaranteed to be monotonic across an update and as a result can goes backward. According to update_fast_timekeeper() description: "In the worst case, this can result is a slightly wrong timestamp (a few nanoseconds)". For PM runtime autosuspend, this means only that the suspend decision can be slightly sub optimal. Fixes: 8234f6734c5d ("PM-runtime: Switch autosuspend over to using hrtimers") Reported-by: Biju Das <biju.das@bp.renesas.com> Signed-off-by: Vincent Guittot <vincent.guittot@linaro.org> --- - v2: Updated commit message to explain the impact of using ktime_get_mono_fast_ns() drivers/base/power/runtime.c | 10 +++++----- include/linux/pm_runtime.h | 2 +- 2 files changed, 6 insertions(+), 6 deletions(-)