Re: vfs: lockdep splat with prepare_bprm_creds

From: Al Viro
Date: Sat Mar 16 2013 - 15:41:47 EST

On Fri, Mar 15, 2013 at 07:19:56PM +0100, Oleg Nesterov wrote:
> On 03/15, Al Viro wrote:
> >
> > On Fri, Mar 15, 2013 at 12:07:14AM -0400, Sasha Levin wrote:
> > > Hi all,
> > >
> > > While fuzzing with trinity inside a KVM tools guest running latest -next kernel
> > > I've stumbled on the following.
> > >
> > > Dave Jones reported something similar, but that seemed to involve cgroup's mutex
> > > and didn't seem like it was the same issue as this one.
> >
> > Lovely... It's an execve() attempt on a "binary" that is, in fact, a procfs
> > file (/proc/<pid>/stack),

> Cough... I am shy to disclose my ignorance, but could you explain how
> open_exec()->do_filp_open(MAY_EXEC) can succeed in this case? At least
> acl_permission_check() looks as if open_exec() should fail...

Umm... point. It might be a false positive, actually - some other
seq_file-based sucker (while chmod +x /proc/self/stack will fail,
chmod +x /proc/vmstat won't) that could be fed to execve(), leading to
1) kernel_read() from execve() can grab m.lock for *some* seq_file m,
while holding ->cred_guard_mutex
2) read() on /proc/self/stack tries to grab ->cred_guard_mutex,
while holding m.lock for a different seq_file m
... with lockdep having no idea that there's a reason why (1) and (2) can't
have the same seq_file involved, said reason being that all files with ->read()
trying to grab ->cred_guard_mutex don't have exec bit set *and* are impossible
to chmod.
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at
Please read the FAQ at