linux-kernel - Re: [PATCH RFC netfs] Fix uninitialized variable in netfs_retry_read

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 for Android: free password hash cracker in your pocket

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <4059210.1734654011@warthog.procyon.org.uk>
Date: Fri, 20 Dec 2024 00:20:11 +0000
From: David Howells <dhowells@...hat.com>
To: paulmck@...nel.org, Christian Brauner <brauner@...nel.org>
Cc: dhowells@...hat.com, jlayton@...nel.org, netfs@...ts.linux.dev,
    linux-fsdevel@...r.kernel.org, linux-kernel@...r.kernel.org,
    sfr@...b.auug.org.au, linux-next@...r.kernel.org
Subject: Re: [PATCH RFC netfs] Fix uninitialized variable in netfs_retry_read_subrequests()

Paul E. McKenney <paulmck@...nel.org> wrote:

> This should actually be considered more of a bug report than a patch.
> 
> Clang 18.1.8 (but not GCC 11.5.0) complains that the "subreq" local
> variable can be used uninitialized in netfs_retry_read_subrequests(),
> just after the abandon_after label.  This function is unusual in having
> three instances of this local variable.  The third and last one is clearly
> erroneous because there is a branch out of the enclosing do-while loop
> to the end of this function, and it looks like the intent is that the
> code at the end of this function be using the same value of the "subreq"
> local variable as is used within that do-while loop.
> 
> Therefore, take the obvious (if potentially quite misguided) approach
> of removing the third declaration of "subreq", instead simply setting
> it to NULL.

I think you're looking at the old version of my netfs-writeback branch that's
residing in Christian's vfs.netfs branch.  I've posted a new version of my
branch[1] without this problem and am hoping for Christian to update the
branch[2] so that Stephen can pull it into linux-next.

David

[1] https://lore.kernel.org/linux-fsdevel/20241216204124.3752367-1-dhowells@redhat.com/T/#t

[2] And hoping he'll remember to drop "[PATCH v5 26/32] Display waited-on page
index after 1min of waiting" for me.  I forgot to remove that debugging patch.