linux-kernel - Re: [PATCH] PM / core: Clear the direct

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]

Date:   Thu, 4 Oct 2018 16:40:00 +0200
From:   Ulf Hansson <ulf.hansson@...aro.org>
To:     Alan Cooper <alcooperx@...il.com>
Cc:     "Rafael J. Wysocki" <rjw@...ysocki.net>,
        Linux PM <linux-pm@...r.kernel.org>,
        Linux Kernel Mailing List <linux-kernel@...r.kernel.org>,
        Alan Stern <stern@...land.harvard.edu>,
        Pavel Machek <pavel@....cz>
Subject: Re: [PATCH] PM / core: Clear the direct_complete flag on errors

On 4 October 2018 at 15:59, Alan Cooper <alcooperx@...il.com> wrote:
>> On 4 October 2018 at 11:08, Rafael J. Wysocki <rjw@...ysocki.net> wrote:
>> > From: Rafael J. Wysocki <rafael.j.wysocki@...el.com>
>> >
>> > If __device_suspend() returns early on an error or pending wakeup
>> > and the power.direct_complete flag has been set for the device
>> > already, the subsequent device_resume() will be confused by it
>> > and it will call pm_runtime_enable() incorrectly, as runtime PM
>> > has not been disabled for the device by __device_suspend().
>>
>> I think it would be fair to mention that is related to the async
>> suspend path, in dpm_suspend().
>
> This also fixed the issue and looks cleaner.
>
>>
>> >
>> > To avoid that, clear power.direct_complete if __device_suspend()
>> > is not going to disable runtime PM for the device before returning.
>>
>> Overall, by looking at the behavior in dpm_suspend() of async
>> suspended devices, it does look a bit fragile to me.
>>
>> My worries is that we put asynced suspended devices in the
>> dpm_suspended_list, no matter if the device was successfully suspended
>> or not. This differs from the no-async path.
>>
>> In the long run, maybe we should change that instead?
>
> I originally looked into this. Currently dmp_suspend moves async
> devices from the prepared list to the suspended list as they are
> queued and I looked at moving this to __device_suspend (after the
> checks for async_error and wake_pending) but realized that this would
> change normal resume ordering and was afraid that would be too
> disruptive.
>

I understand, however, I would be surprised if that would be a real
problem. But who knows. :-)

The hole async thing is opt-in, and it does also assume that
drivers/subsystems to cope with devices suspend/resume ordering to be
based upon device relationships, like parent/child,
device_link-supplier/consumers. If it's not working, drivers should
disable the async option, at least in my opinion.

Kind regards
Uffe