Re: [PATCH] mm/mempolicy: fix uninit-value in mpol_rebind_policy()

[Andrew Morton <akpm-de/tnXTf+JLsfHDXvbKv3WD2FQJk+8+b@public.gmane.org>]

On Mon, 16 May 2022 17:47:26 +0800 Wang Cheng <wanngchenng-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org> wrote:

> 
> ...
>
> This patch seems to fix below bug too.
> KMSAN: uninit-value in mpol_rebind_mm (2)
> https://syzkaller.appspot.com/bug?id=f2fecd0d7013f54ec4162f60743a2b28df40926b
> 
> The uninit-value is pol->w.cpuset_mems_allowed in mpol_rebind_policy().
> When syzkaller reproducer runs to the beginning of mpol_new(),
> 
> 	    mpol_new() mm/mempolicy.c
> 	  do_mbind() mm/mempolicy.c
> 	kernel_mbind() mm/mempolicy.c
> 
> `mode` is 1(MPOL_PREFERRED), nodes_empty(*nodes) is `true` and `flags`
> is 0. Then
> 
> 	mode = MPOL_LOCAL;
> 	...
> 	policy->mode = mode;
> 	policy->flags = flags;
> 
> will be executed. So in mpol_set_nodemask(),
> 
> 	    mpol_set_nodemask() mm/mempolicy.c
> 	  do_mbind()
> 	kernel_mbind()
> 
> pol->mode is 4(MPOL_LOCAL), that `nodemask` in `pol` is not initialized,
> which will be accessed in mpol_rebind_policy().

Thanks, I added the above to the changelog and I plan to import the
result into mm-stable later this week.

> IIUC, "#syz fix: mm/mempolicy: fix uninit-value in mpol_rebind_policy()"
> could be sent to syzbot+ad1b8c404f0959c4bfcc-Pl5Pbv+GP7P466ipTTIvnc23WoclnBCfAL8bYrjMMd8@public.gmane.org
> to attach the fixing commit to the bug. WDYT?

Could be.  The "syz fix" isn't a thing I've paid much attention to. 
I'll start doing so ;)