[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <CAHk-=wgiPxXzRNnfaXk7ozSWSu7fFU--kTmVjkDaTB05wwUk_g@mail.gmail.com>
Date: Sun, 21 Feb 2021 15:07:17 -0800
From: Linus Torvalds <torvalds@...ux-foundation.org>
To: Dave Airlie <airlied@...il.com>, Kevin Wang <kevin1.wang@....com>,
Alex Deucher <alexander.deucher@....com>
Cc: Daniel Vetter <daniel.vetter@...ll.ch>,
dri-devel <dri-devel@...ts.freedesktop.org>,
LKML <linux-kernel@...r.kernel.org>
Subject: Re: [git pull] drm for 5.12-rc1
On Thu, Feb 18, 2021 at 10:06 PM Dave Airlie <airlied@...il.com> wrote:
>
> Let me know if there are any issues,
gcc was happy, and I obviously already pushed out my merge, but then
when I did my clang build afterwards, it reports:
drivers/gpu/drm/amd/amdgpu/../pm/swsmu/smu_cmn.c:764:2: warning:
variable 'structure_size' is used uninitialized whenever switch
default is taken [-Wsometimes-uninitialized]
default:
^~~~~~~
drivers/gpu/drm/amd/amdgpu/../pm/swsmu/smu_cmn.c:770:23: note:
uninitialized use occurs here
memset(header, 0xFF, structure_size);
^~~~~~~~~~~~~~
and clang is very very right. That "default" case is completely
broken, and will generate a randomly sized memset. Not good.
Presumably that default case never happens, but if so it shouldn't exist.
Perhaps better yet, make the "default" case just do a "return" instead
of a break. Breaking out of the switch statement to code that cannot
possibly work is all kinds of mindless.
Kevin/Alex? This was introduced by commit de4b7cd8cb87
("drm/amd/pm/swsmu: unify the init soft gpu metrics function")
Linus
Powered by blists - more mailing lists