linux-kernel - Re: [PATCH 3/6] fs: clear FMODE

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]

Message-ID: <YrXwuf3lw/I1H64q@ZenIV>
Date:   Fri, 24 Jun 2022 18:13:29 +0100
From:   Al Viro <viro@...iv.linux.org.uk>
To:     "Jason A. Donenfeld" <Jason@...c4.com>
Cc:     linux-kernel@...r.kernel.org, Jens Axboe <axboe@...nel.dk>,
        linux-fsdevel@...r.kernel.org
Subject: Re: [PATCH 3/6] fs: clear FMODE_LSEEK if no llseek function

On Fri, Jun 24, 2022 at 07:09:17PM +0200, Jason A. Donenfeld wrote:
> Hi Al,
> 
> On Fri, Jun 24, 2022 at 06:04:19PM +0100, Al Viro wrote:
> > On Fri, Jun 24, 2022 at 06:56:28PM +0200, Jason A. Donenfeld wrote:
> > > This helps unify a longstanding wart where FMODE_LSEEK hasn't been
> > > uniformly unset when it should be.
> > > 
> > > Signed-off-by: Jason A. Donenfeld <Jason@...c4.com>
> > > ---
> > >  fs/file_table.c | 2 ++
> > >  fs/open.c       | 2 ++
> > >  2 files changed, 4 insertions(+)
> > > 
> > > diff --git a/fs/file_table.c b/fs/file_table.c
> > > index 5424e3a8df5f..15700b2e1b53 100644
> > > --- a/fs/file_table.c
> > > +++ b/fs/file_table.c
> > > @@ -241,6 +241,8 @@ static struct file *alloc_file(const struct path *path, int flags,
> > >  	if ((file->f_mode & FMODE_WRITE) &&
> > >  	     likely(fop->write || fop->write_iter))
> > >  		file->f_mode |= FMODE_CAN_WRITE;
> > > +	if ((file->f_mode & FMODE_LSEEK) && !file->f_op->llseek)
> > > +		file->f_mode &= ~FMODE_LSEEK;
> > 
> > 	Where would FMODE_LSEEK come from in this one?  ->f_mode is set
> > (in __alloc_file()) to OPEN_FMODE(flags); that does deal with FMODE_READ
> > and FMODE_WRITE, but FMODE_LSEEK will be clear...
> 
> >From the `int flags` parameter of the function. That's an O flag not an
> F flag, though, so I assume you mean that it's impossible to get LSEEK
> there in practice? If so, I'll drop this hunk.

	if (file->f_op->llseek)
		file->f_mode |= FMODE_LSEEK;

you want it to match what came in file_operations...