linux-kernel - Re: [PATCH] radix tree test suite: Fix uninitialized variable compilation warning

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, 10 Nov 2022 16:06:13 +0000
From:   Matthew Wilcox <willy@...radead.org>
To:     Andrew Morton <akpm@...ux-foundation.org>
Cc:     Rong Tao <rtoax@...mail.com>, Rong Tao <rongtao@...tc.cn>,
        wuchi <wuchi.zero@...il.com>,
        open list <linux-kernel@...r.kernel.org>, linux-mm@...ck.org
Subject: Re: [PATCH] radix tree test suite: Fix uninitialized variable
 compilation warning

On Wed, Nov 09, 2022 at 04:23:48PM -0800, Andrew Morton wrote:
> On Wed,  9 Nov 2022 22:34:25 +0800 Rong Tao <rtoax@...mail.com> wrote:
> 
> > [PATCH] radix tree test suite: Fix uninitialized variable compilation warning
> 
> This is not the test suite.
> 
> > We need to set an initial value for offset to eliminate compilation
> > warning.
> > 
> > How to reproduce warning:
> > 
> > $ make -C tools/testing/radix-tree
> > radix-tree.c: In function ‘radix_tree_tag_clear’:
> > radix-tree.c:1046:17: warning: ‘offset’ may be used uninitialized in this function [-Wmaybe-uninitialized]
> >  1046 |                 node_tag_clear(root, parent, tag, offset);
> >       |                 ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
> > 
> > ...
> >
> > --- a/lib/radix-tree.c
> > +++ b/lib/radix-tree.c
> > @@ -1029,7 +1029,7 @@ void *radix_tree_tag_clear(struct radix_tree_root *root,
> >  {
> >  	struct radix_tree_node *node, *parent;
> >  	unsigned long maxindex;
> > -	int offset;
> > +	int offset = 0;
> >  
> >  	radix_tree_load_root(root, &node, &maxindex);
> >  	if (index > maxindex)
> 
> Are we sure this isn't actually a bug?  What happens if the tree is empty?

If the tree is empty, then node is NULL and we never use offset.
The compiler is too stupid to know this.  This warning is only observed
when building the test suite and not when building the kernel itself.
I'm not sure the patch is worth it, tbh.