linux-kernel - Re: [PATCH] staging: rtl8192e: prefer strscpy over strncpy

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] [day] [month] [year] [list]

Date:   Tue, 5 Sep 2023 12:23:25 +0300
From:   Dan Carpenter <dan.carpenter@...aro.org>
To:     Michael Straube <straube.linux@...il.com>
Cc:     Greg KH <gregkh@...uxfoundation.org>, philipp.g.hortmann@...il.com,
        linux-staging@...ts.linux.dev, linux-kernel@...r.kernel.org
Subject: Re: [PATCH] staging: rtl8192e: prefer strscpy over strncpy

On Fri, Aug 25, 2023 at 07:13:42AM +0200, Michael Straube wrote:
> On 8/10/23 07:01, Dan Carpenter wrote:
> > On Wed, Aug 09, 2023 at 08:02:05PM +0200, Michael Straube wrote:
> > > On 8/9/23 14:21, Greg KH wrote:
> > > > On Sat, Aug 05, 2023 at 09:51:14AM +0200, Michael Straube wrote:
> > > > > Replace strncpy with strscpy in two places where the destination buffer
> > > > > should be NUL-terminated. Found by checkpatch.
> > > > > 
> > > > > WARNING: Prefer strscpy, strscpy_pad, or __nonstring over strncpy - see: https://github.com/KSPP/linux/issues/90
> > > > 
> > > > If a global search/replace could be done, it would have happend a long
> > > > time ago.
> > > > 
> > > > How was this tested?  The functions work differently, are you sure there
> > > > is no change in functionality here?
> > > > 
> > > 
> > > It was only compile tested. To me it looked as it does not change
> > > functionality, but looking a bit deeper at it I'm not sure anymore.
> > > So, we should leave it as is.
> > 
> > So there are three main differences between strncpy() and strcpy().
> > 
> > 1) The return.
> > 2) strncpy() will always write net->hidden_ssid_len bytes.  If the
> >     string to copy is smaller than net->hidden_ssid_len bytes it will
> >     fill the rest with zeroes.  This can be important for preventing
> >     information leaks.
> > 3) strscpy() will always add a NUL terminator where strncpy() just
> >     truncates a too long string without adding a terminator.
> > 
> > We want #3.  We don't care about #1.  The only thing to check is #2.
> > 
> > regards,
> > dan carpenter
> > 
> 
> Thank you Dan,
> 
> so in this case we should/could replace strncpy with strscpy_pad,
> correct?

I'm pretty sure that strscpy() was correct.  It requires some analysis
in how this is initialized and/or used.

Don't just automatically use strscpy_pad() to try avoid doing the
analysis.  ;)

regards,
dan carpenter