lists.openwall.net   lists  /  announce  owl-users  owl-dev  john-users  john-dev  passwdqc-users  yescrypt  popa3d-users  /  oss-security  kernel-hardening  musl  sabotage  tlsify  passwords  /  crypt-dev  xvendor  /  Bugtraq  Full-Disclosure  linux-kernel  linux-netdev  linux-ext4  linux-hardening  linux-cve-announce  PHC 
Open Source and information security mailing list archives
 
Hash Suite: Windows password security audit tool. GUI, reports in PDF.
[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <CAJZ5v0iVLr2+PLXy2Y4pmGtZUOMGPVTpytRDgTTGbKt=Zq_LZA@mail.gmail.com>
Date: Thu, 4 Jul 2024 18:58:57 +0200
From: "Rafael J. Wysocki" <rafael@...nel.org>
To: Daniel Lezcano <daniel.lezcano@...aro.org>
Cc: "Rafael J. Wysocki" <rafael@...nel.org>, "Rafael J. Wysocki" <rjw@...ysocki.net>, 
	Linux PM <linux-pm@...r.kernel.org>, LKML <linux-kernel@...r.kernel.org>, 
	Lukasz Luba <lukasz.luba@....com>, 
	Srinivas Pandruvada <srinivas.pandruvada@...ux.intel.com>, Zhang Rui <rui.zhang@...el.com>, 
	Neil Armstrong <neil.armstrong@...aro.org>
Subject: Re: [PATCH v3] thermal: core: Call monitor_thermal_zone() if zone
 temperature is invalid

On Thu, Jul 4, 2024 at 6:53 PM Daniel Lezcano <daniel.lezcano@...aro.org> wrote:
>
> On 04/07/2024 16:21, Rafael J. Wysocki wrote:
> > On Thu, Jul 4, 2024 at 2:49 PM Daniel Lezcano <daniel.lezcano@...aro.org> wrote:
> >>
> >> On 04/07/2024 13:46, Rafael J. Wysocki wrote:
> >>> From: Rafael J. Wysocki <rafael.j.wysocki@...el.com>
> >>>
> >>> Commit 202aa0d4bb53 ("thermal: core: Do not call handle_thermal_trip()
> >>> if zone temperature is invalid") caused __thermal_zone_device_update()
> >>> to return early if the current thermal zone temperature was invalid.
> >>>
> >>> This was done to avoid running handle_thermal_trip() and governor
> >>> callbacks in that case which led to confusion.  However, it went too
> >>> far because monitor_thermal_zone() still needs to be called even when
> >>> the zone temperature is invalid to ensure that it will be updated
> >>> eventually in case thermal polling is enabled and the driver has no
> >>> other means to notify the core of zone temperature changes (for example,
> >>> it does not register an interrupt handler or ACPI notifier).
> >>>
> >>> Also if the .set_trips() zone callback is expected to set up monitoring
> >>> interrupts for a thermal zone, it needs to be provided with valid
> >>> boundaries and that can only be done if the zone temperature is known.
> >>>
> >>> Accordingly, to ensure that __thermal_zone_device_update() will
> >>> run again after a failing zone temperature check, make it call
> >>> monitor_thermal_zone() regardless of whether or not the zone
> >>> temperature is valid and make the latter schedule a thermal zone
> >>> temperature update if the zone temperature is invalid even if
> >>> polling is not enabled for the thermal zone (however, if this
> >>> continues to fail, give up after some time).
> >>
> >> Rafael,
> >>
> >> do we agree that we should fix somehow the current issue in this way
> >> because we are close to the merge window,
> >
> > Yes.
> >
> >> but the proper fix is not doing that ?
> >
> > We need to decide what to do in general when __thermal_zone_get_temp()
> > returns an error.  A proper fix would result from that, but it would
> > require more time than is available IMV.  We can properly fix this in
> > 6.11.
>
> Right, in general we should take care of returning values from the
> different functions, update_temperature(), etc... in order to have the
> thermal_zone_device_update() returning a value.
>
> So from there we can catch the result in the initialization function and
> do the proper actions.
>
>  From a higher perspective, IMO the code contains too many returning
> void functions. We should convert that into returning values and handle
> the error cases.
>
> > For 6.10 I see two options:
> >
> > 1. Apply the v2 of this patch:
> >
> > https://lore.kernel.org/linux-pm/2764814.mvXUDI8C0e@rjwysocki.net/
> >
> > I slightly prefer it because it is simpler and doesn't change the size
> > of struct thermal_zone_device.
>
> I agree
>
> >  However, the clear disadvantage of it
> > is that it will poke at dead thermal zones indefinitely.
>
> Yes, but the advantage of this disadvantage is it is so visible that
> buggy routine will be brought to the light, so they can be fixed. I
> don't think we should have so many, perhaps none.
>
> > The THERMAL_RECHECK_DELAY_MS value in it can be adjusted.  Maybe 250
> > ms would be a better choice?
>
> Yes
>
> > 2. Apply this patch (ie. v3)
> >
> > It is nicer to thermal zones that never become operational, but it may
> > miss thermal zones that become operational very late.
>
> I would keep this v3 as a backup in case there are too many complaints,
> but I doubt

OK, I'll go for the v2 with THERMAL_RECHECK_DELAY_MS equal to 250 ms.

Thanks!

> >>> Fixes: 202aa0d4bb53 ("thermal: core: Do not call handle_thermal_trip() if zone temperature is invalid")
> >>> Reported-by: Daniel Lezcano <daniel.lezcano@...aro.org>
> >>> Link: https://lore.kernel.org/linux-pm/dc1e6cba-352b-4c78-93b5-94dd033fca16@linaro.org
> >>> Link: https://lore.kernel.org/linux-pm/2764814.mvXUDI8C0e@rjwysocki.net
> >>> Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@...el.com>
> >>> ---
> >>>    drivers/thermal/thermal_core.c |   13 ++++++++++++-
> >>>    drivers/thermal/thermal_core.h |    9 +++++++++
> >>>    2 files changed, 21 insertions(+), 1 deletion(-)
> >>>
> >>> Index: linux-pm/drivers/thermal/thermal_core.c
> >>> ===================================================================
> >>> --- linux-pm.orig/drivers/thermal/thermal_core.c
> >>> +++ linux-pm/drivers/thermal/thermal_core.c
> >>> @@ -300,6 +300,14 @@ static void monitor_thermal_zone(struct
> >>>                thermal_zone_device_set_polling(tz, tz->passive_delay_jiffies);
> >>>        else if (tz->polling_delay_jiffies)
> >>>                thermal_zone_device_set_polling(tz, tz->polling_delay_jiffies);
> >>> +     else if (tz->temperature == THERMAL_TEMP_INVALID &&
> >>> +              tz->recheck_delay_jiffies <= THERMAL_MAX_RECHECK_DELAY) {
> >>> +             thermal_zone_device_set_polling(tz, tz->recheck_delay_jiffies);
> >>> +             /* Double the recheck delay for the next attempt. */
> >>> +             tz->recheck_delay_jiffies += tz->recheck_delay_jiffies;
> >>> +             if (tz->recheck_delay_jiffies > THERMAL_MAX_RECHECK_DELAY)
> >>> +                     dev_info(&tz->device, "Temperature unknown, giving up\n");
> >>> +     }
> >>>    }
> >>>
> >>>    static struct thermal_governor *thermal_get_tz_governor(struct thermal_zone_device *tz)
> >>> @@ -430,6 +438,7 @@ static void update_temperature(struct th
> >>>
> >>>        tz->last_temperature = tz->temperature;
> >>>        tz->temperature = temp;
> >>> +     tz->recheck_delay_jiffies = 1;
> >>>
> >>>        trace_thermal_temperature(tz);
> >>>
> >>> @@ -514,7 +523,7 @@ void __thermal_zone_device_update(struct
> >>>        update_temperature(tz);
> >>>
> >>>        if (tz->temperature == THERMAL_TEMP_INVALID)
> >>> -             return;
> >>> +             goto monitor;
> >>>
> >>>        tz->notify_event = event;
> >>>
> >>> @@ -536,6 +545,7 @@ void __thermal_zone_device_update(struct
> >>>
> >>>        thermal_debug_update_trip_stats(tz);
> >>>
> >>> +monitor:
> >>>        monitor_thermal_zone(tz);
> >>>    }
> >>>
> >>> @@ -1438,6 +1448,7 @@ thermal_zone_device_register_with_trips(
> >>>
> >>>        thermal_set_delay_jiffies(&tz->passive_delay_jiffies, passive_delay);
> >>>        thermal_set_delay_jiffies(&tz->polling_delay_jiffies, polling_delay);
> >>> +     tz->recheck_delay_jiffies = 1;
> >>>
> >>>        /* sys I/F */
> >>>        /* Add nodes that are always present via .groups */
> >>> Index: linux-pm/drivers/thermal/thermal_core.h
> >>> ===================================================================
> >>> --- linux-pm.orig/drivers/thermal/thermal_core.h
> >>> +++ linux-pm/drivers/thermal/thermal_core.h
> >>> @@ -67,6 +67,8 @@ struct thermal_governor {
> >>>     * @polling_delay_jiffies: number of jiffies to wait between polls when
> >>>     *                  checking whether trip points have been crossed (0 for
> >>>     *                  interrupt driven systems)
> >>> + * @recheck_delay_jiffies: delay after a failed thermal zone temperature check
> >>> + *                   before attempting to check it again
> >>>     * @temperature:    current temperature.  This is only for core code,
> >>>     *                  drivers should use thermal_zone_get_temp() to get the
> >>>     *                  current temperature
> >>> @@ -108,6 +110,7 @@ struct thermal_zone_device {
> >>>        int num_trips;
> >>>        unsigned long passive_delay_jiffies;
> >>>        unsigned long polling_delay_jiffies;
> >>> +     unsigned long recheck_delay_jiffies;
> >>>        int temperature;
> >>>        int last_temperature;
> >>>        int emul_temperature;
> >>> @@ -133,6 +136,12 @@ struct thermal_zone_device {
> >>>        struct thermal_trip_desc trips[] __counted_by(num_trips);
> >>>    };
> >>>
> >>> +/*
> >>> + * Maximum delay after a failing thermal zone temperature check before
> >>> + * attempting to check it again (in jiffies).
> >>> + */
> >>> +#define THERMAL_MAX_RECHECK_DELAY    (30 * HZ)
> >>> +
> >>>    /* Default Thermal Governor */
> >>>    #if defined(CONFIG_THERMAL_DEFAULT_GOV_STEP_WISE)
> >>>    #define DEFAULT_THERMAL_GOVERNOR       "step_wise"
> >>>
> >>>
> >>>
> >>
> >> --
> >> <http://www.linaro.org/> Linaro.org │ Open source software for ARM SoCs
> >>
> >> Follow Linaro:  <http://www.facebook.com/pages/Linaro> Facebook |
> >> <http://twitter.com/#!/linaroorg> Twitter |
> >> <http://www.linaro.org/linaro-blog/> Blog
> >>
> >>
>
> --
> <http://www.linaro.org/> Linaro.org │ Open source software for ARM SoCs
>
> Follow Linaro:  <http://www.facebook.com/pages/Linaro> Facebook |
> <http://twitter.com/#!/linaroorg> Twitter |
> <http://www.linaro.org/linaro-blog/> Blog
>

Powered by blists - more mailing lists

Powered by Openwall GNU/*/Linux Powered by OpenVZ