linux-kernel - Re: [PATCH v2] thermal/core: Clear all mitigation when thermal zone is disabled

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] [day] [month] [year] [list]

Date:   Thu, 30 Dec 2021 23:46:12 +0530
From:   Manaf Meethalavalappu Pallikunhi <quic_manafm@...cinc.com>
To:     "Rafael J. Wysocki" <rafael@...nel.org>
CC:     Daniel Lezcano <daniel.lezcano@...aro.org>,
        Amit Kucheria <amitk@...nel.org>,
        Zhang Rui <rui.zhang@...el.com>,
        Thara Gopinath <thara.gopinath@...aro.org>,
        Matthias Kaehlcke <mka@...omium.org>,
        Linux PM <linux-pm@...r.kernel.org>,
        Linux Kernel Mailing List <linux-kernel@...r.kernel.org>
Subject: Re: [PATCH v2] thermal/core: Clear all mitigation when thermal zone
 is disabled


On 12/30/2021 9:39 PM, Rafael J. Wysocki wrote:
> On Wed, Dec 29, 2021 at 8:03 AM Manaf Meethalavalappu Pallikunhi
> <quic_manafm@...cinc.com> wrote:
>> Whenever a thermal zone is in trip violated state, there is a chance
>> that the same thermal zone mode can be disabled either via thermal
>> core API or via thermal zone sysfs. Once it is disabled, the framework
>> bails out any re-evaluation of thermal zone. It leads to a case where
>> if it is already in mitigation state, it will stay the same state
>> until it is re-enabled.
> You seem to be arguing that disabling a thermal zone should prevent it
> from throttling anything, which is reasonable, but I'm not sure if the
> change below is sufficient for that.
>
>> To avoid above mentioned issue, on thermal zone disable request
>> reset thermal zone and clear mitigation for each trip explicitly.
>>
>> Signed-off-by: Manaf Meethalavalappu Pallikunhi <quic_manafm@...cinc.com>
>> ---
>>   drivers/thermal/thermal_core.c | 9 ++++++++-
>>   1 file changed, 8 insertions(+), 1 deletion(-)
>>
>> diff --git a/drivers/thermal/thermal_core.c b/drivers/thermal/thermal_core.c
>> index 51374f4..5f4e35b 100644
>> --- a/drivers/thermal/thermal_core.c
>> +++ b/drivers/thermal/thermal_core.c
>> @@ -427,6 +427,7 @@ static int thermal_zone_device_set_mode(struct thermal_zone_device *tz,
>>                                          enum thermal_device_mode mode)
>>   {
>>          int ret = 0;
>> +       int trip;
> This can be declared in the block in which it is used.
Sure, will update in next patch version
>
>>          mutex_lock(&tz->lock);
>>
>> @@ -449,8 +450,14 @@ static int thermal_zone_device_set_mode(struct thermal_zone_device *tz,
>>
>>          if (mode == THERMAL_DEVICE_ENABLED)
> The coding style asks for braces here if they are used after the else.
Sure will update in next version
>>                  thermal_notify_tz_enable(tz->id);
>> -       else
>> +       else {
>> +               /* make sure all previous throttlings are cleared */
>> +               thermal_zone_device_init(tz);
>> +               for (trip = 0; trip < tz->trips; trip++)
>> +                       handle_thermal_trip(tz, trip);
> So I'm not sure if this makes the throttling go away in all cases (eg.
> what if the current temperature is still above the given trip at this
> point?).

The thermal_zone_device_init() will reset current temperature with THERMAL_TEMP_INVALID. Then the following API
thandle_thermal_trip() doesn't call get_temp to sensor driver again instead it will use this reset temperature
value for each trip re-evaluation. Hence  for above mentioned case, I think it will clear the throttling
irrespective of what is the actual current temperature at that instant. Otherwise please correct me

May I know is there any other possible cases where throttling will not go away completely ?

>> +
>>                  thermal_notify_tz_disable(tz->id);
>> +       }
>>
>>          return ret;
>>   }
>>