linux-kernel - Re: [PATCH] gpio: sprd: Change sprd_gpio lock to raw_spin

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]

Message-ID: <CAMRc=MfKLhD40X7Xc-U=tFjt=ipsDk43=cieb6AU5K7wLzVA=w@mail.gmail.com>
Date: Wed, 28 Jan 2026 01:10:24 -0800
From: Bartosz Golaszewski <brgl@...nel.org>
To: Andy Shevchenko <andriy.shevchenko@...el.com>
Cc: Sebastian Andrzej Siewior <bigeasy@...utronix.de>, Xuewen Yan <xuewen.yan@...soc.com>, linusw@...nel.org, 
	orsonzhai@...il.com, baolin.wang@...ux.alibaba.com, zhang.lyra@...il.com, 
	clrkwllms@...nel.org, rostedt@...dmis.org, linux-gpio@...r.kernel.org, 
	linux-kernel@...r.kernel.org, linux-rt-devel@...ts.linux.dev, 
	ke.wang@...soc.com, wenhua.lin@...soc.com, xuewen.yan94@...il.com, 
	Bartosz Golaszewski <brgl@...nel.org>
Subject: Re: [PATCH] gpio: sprd: Change sprd_gpio lock to raw_spin_lock

On Tue, 27 Jan 2026 20:36:16 +0100, Andy Shevchenko
<andriy.shevchenko@...el.com> said:
> On Tue, Jan 27, 2026 at 10:07:01AM +0100, Bartosz Golaszewski wrote:
>> On Mon, Jan 26, 2026 at 10:50 AM Sebastian Andrzej Siewior
>> <bigeasy@...utronix.de> wrote:
>> > On 2026-01-26 17:42:09 [+0800], Xuewen Yan wrote:
>
> ...
>
>> > > There was a lockdep warning in sprd_gpio:
>
>> > > [    6.258411][T329@C6]  el0t_64_sync+0x1c4/0x1c8
>> >
>> > This could be reduced to "BUG: Invalid wait context". The other bits
>> > provide to added value.
>> >
>> > > This is because the spin_lock would change to rt_mutex in PREEMPT_RT,
>> > > however the sprd_gpio->lock would use in hard-irq, this is unsafe.
>> > >
>> > > So change the spin_lock to raw_spin_lock to use the spinlock
>> >
>> > spinlock_t to raw_spinlock_t
>>
>> I fixed this in git.
>
> Not sure if you noticed the above comment by Sebastian and it's actually
> documented in Submitting Patches that backtraces should only carry the
> necessary information, so usually ~3-5 lines is enough and not two full pages
> like in the commit as I see in the repository.
>

Noted for the future, thanks.

Bartosz