[PATCH] drm/amdgpu: fix deadlock while reading mqd from debugfs

[Sharma, Shashank]

[AMD Official Use Only - General]

Hey Alex,
Sure, I will pick it up and push it to staging.

Regards
Shashank
________________________________
From: Alex Deucher <alexdeucher at gmail.com>
Sent: Monday, March 25, 2024 12:23 AM
To: Sharma, Shashank <Shashank.Sharma at amd.com>
Cc: Johannes Weiner <hannes at cmpxchg.org>; Christian König <ckoenig.leichtzumerken at gmail.com>; Deucher, Alexander <Alexander.Deucher at amd.com>; Koenig, Christian <Christian.Koenig at amd.com>; amd-gfx at lists.freedesktop.org <amd-gfx at lists.freedesktop.org>; dri-devel at lists.freedesktop.org <dri-devel at lists.freedesktop.org>; linux-kernel at vger.kernel.org <linux-kernel at vger.kernel.org>
Subject: Re: [PATCH] drm/amdgpu: fix deadlock while reading mqd from debugfs

On Sat, Mar 23, 2024 at 4:47 PM Sharma, Shashank
<shashank.sharma at amd.com> wrote:
>
>
> On 23/03/2024 15:52, Johannes Weiner wrote:
> > On Thu, Mar 14, 2024 at 01:09:57PM -0400, Johannes Weiner wrote:
> >> Hello,
> >>
> >> On Fri, Mar 08, 2024 at 12:32:33PM +0100, Christian König wrote:
> >>> Am 07.03.24 um 23:07 schrieb Johannes Weiner:
> >>>> Lastly I went with an open loop instead of a memcpy() as I wasn't
> >>>> sure if that memory is safe to address a byte at at time.
> >> Shashank pointed out to me in private that byte access would indeed be
> >> safe. However, after actually trying it it won't work because memcpy()
> >> doesn't play nice with mqd being volatile:
> >>
> >> /home/hannes/src/linux/linux/drivers/gpu/drm/amd/amdgpu/amdgpu_ring.c: In function 'amdgpu_debugfs_mqd_read':
> >> /home/hannes/src/linux/linux/drivers/gpu/drm/amd/amdgpu/amdgpu_ring.c:550:22: warning: passing argument 1 of '__builtin_dynamic_object_size' discards 'volatil' qualifier from pointer target type [-Wdiscarded-qualifiers]
> >>    550 |         memcpy(kbuf, mqd, ring->mqd_size);
> >>
> >> So I would propose leaving the patch as-is. Shashank, does that sound
> >> good to you?
> > Friendly ping :)
> >
> > Shashank, is your Reviewed-by still good for this patch, given the
> > above?
>
> Ah, sorry I missed this due to some parallel work, and just realized the
> memcpy/volatile limitation.
>
> I also feel the need of protecting MQD read under a lock to avoid
> parallel change in MQD while we do byte-by-byte copy, but I will add
> that in my to-do list.
>
> Please feel free to use my R-b.

Shashank, if the patch looks good, can you pick it up and apply it?

Alex


>
> - Shashank
>
> > Thanks
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.freedesktop.org/archives/amd-gfx/attachments/20240325/9fe76ab1/attachment.htm>