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: <9821f60a-c562-12e8-bdd4-45dbfe88a117@linaro.org>
Date:   Sat, 27 Nov 2021 00:05:53 +0100
From:   Daniel Lezcano <daniel.lezcano@...aro.org>
To:     Doug Smythies <dsmythies@...us.net>
Cc:     "Rafael J. Wysocki" <rafael@...nel.org>,
        Linux PM list <linux-pm@...r.kernel.org>,
        Linux Kernel Mailing List <linux-kernel@...r.kernel.org>,
        Lukasz Luba <lukasz.luba@....com>,
        Greg Kroah-Hartman <gregkh@...uxfoundation.org>
Subject: Re: [PATCH v6 3/7] powercap/drivers/dtpm: Simplify the dtpm table

On 26/11/2021 22:56, Doug Smythies wrote:
> On Fri, Nov 26, 2021 at 9:43 AM Daniel Lezcano
> <daniel.lezcano@...aro.org> wrote:
>>
>> On 26/11/2021 18:21, Rafael J. Wysocki wrote:
>>> Hi Doug,
>>>
>>> On Fri, Nov 26, 2021 at 6:08 PM Doug Smythies <dsmythies@...us.net> wrote:
>>>>
>>>> Hi Daniel,
>>>>
>>>> This patch introduces a regression, at least on my test system.
>>>> I can no longer change CPU frequency scaling drivers, for example
>>>> from intel_cpufreq (A.K.A intel_pstate in passive mode) to intel_pstate
>>>> (A.K.A. active mode). The task just hangs forever.
>>>>
>>>> I bisected the kernel and got this commit as the result.
>>>> As a double check, I reverted this commit:
>>>> 7a89d7eacf8e84f2afb94db5ae9d9f9faa93f01c
>>>> on kernel 5.16-rc2 and the issue was resolved.
>>>>
>>>> While your email is fairly old, I observe that it was only included as of
>>>> kernel 5.16-rc1.
>>>>
>>>> Command Example that never completes:
>>>>
>>>> $ echo passive | sudo tee /sys/devices/system/cpu/intel_pstate/status
>>>>
>>>> syslog excerpt attached.
>>>
>>> This looks like it may be problematic:
>>>
>>> diff --git a/drivers/powercap/dtpm_cpu.c b/drivers/powercap/dtpm_cpu.c
>>> index f6076de39540..98841524a782 100644
>>> --- a/drivers/powercap/dtpm_cpu.c
>>> +++ b/drivers/powercap/dtpm_cpu.c
>>> @@ -204,7 +204,7 @@ static int cpuhp_dtpm_cpu_online(unsigned int cpu)
>>>        return ret;
>>> }
>>>
>>> -int dtpm_register_cpu(struct dtpm *parent)
>>> +static int __init dtpm_cpu_init(void)
>>> {
>>>        int ret;
>>>
>>> so please try to remove the __init annotation from dtpm_cpu_init() and
>>> see if that helps.
>>
>> Yes, actually that should be called only if it is configured properly.
>> The dtpm_cpu just initializes itself unconditionally, I did not figured
>> out there is the usually allyesconfig used by default by the distros.
>>
>> That should be fixed with a proper DT configuration [1]
> 
> I added your 5 patch set on top of 5.16-rc2 and confirm it fixes
> the issue. I tested both ways, with CONFIG_OF not set, forcing the
> CONFIG_DTPM stuff off, and with CONFIG_OF=y.
> 
> Oh, I used V2 of the patch set from earlier today.

Thanks Doug for testing.

For a -rc I should prevent the cpu to setup at boot time with a simple fix.


-- 
<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