linux-kernel - Re: [Fwd: [PATCH 1/2] string: Add stracpy and stracpy

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:   Wed, 24 Jul 2019 03:43:22 -0700
From:   Joe Perches <joe@...ches.com>
To:     David Laight <David.Laight@...LAB.COM>,
        Julia Lawall <julia.lawall@...6.fr>
Cc:     cocci <cocci@...teme.lip6.fr>, LKML <linux-kernel@...r.kernel.org>
Subject: Re: [Fwd: [PATCH 1/2] string: Add stracpy and stracpy_pad
 mechanisms]

On Wed, 2019-07-24 at 10:28 +0000, David Laight wrote:
> From: Joe Perches
> > Sent: 24 July 2019 05:38
> > On Tue, 2019-07-23 at 23:27 -0500, Julia Lawall wrote:
> > > On Tue, 23 Jul 2019, Joe Perches wrote:
> > > > On Tue, 2019-07-23 at 22:54 -0500, Julia Lawall wrote:
> > > > > A seantic patch and the resulting output for the case where the third
> > > > > arugument is a constant is attached.  Likewise the resulting output on a
> > > > > recent linux-next.
[]
> > > > There is a problem with conversions with assignments
> > > > of strlcpy() so ideally the cocci script should make sure
> > > > any return value was not used before conversion.
> > > > 
> > > > This is not a provably good conversion:
> > > > 
> > > > drivers/s390/char/sclp_ftp.c
> > > > @@ -114,8 +114,7 @@ static int sclp_ftp_et7(const struct hmc
> > > >         sccb->evbuf.mdd.ftp.length = ftp->len;
> > > >         sccb->evbuf.mdd.ftp.bufaddr = virt_to_phys(ftp->buf);
> > > > 
> > > > -       len = strlcpy(sccb->evbuf.mdd.ftp.fident, ftp->fname,
> > > > -                     HMCDRV_FTP_FIDENT_MAX);
> > > > +       len = stracpy(sccb->evbuf.mdd.ftp.fident, ftp->fname);
[]
> > Any use of the strlcpy return value should not be converted
> > because the logic after an assignment or use of the return value
> > can not be assured to have the same behavior.
> 
> Most of the code probably expects the length to be that copied
> (so is broken if the string is truncated).

Fortunately there are relatively few uses of the return
value of strlcpy so it's not a large problem set.

Slightly unrepresentative counts:

$ git grep -P "^\s+strlcpy\b" | wc -l
1681
$ git grep -P "=\s*strlcpy\b" | wc -l
28