linux-kernel - Re: [PATCH v3] coccinelle: suggest replacing strncpy+truncation by strscpy

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:   Fri, 20 Jul 2018 07:49:59 +0200 (CEST)
From:   Julia Lawall <julia.lawall@...6.fr>
To:     Dominique Martinet <asmadeus@...ewreck.org>
cc:     Masahiro Yamada <yamada.masahiro@...ionext.com>,
        Ville Syrjälä <ville.syrjala@...ux.intel.com>,
        Gilles Muller <Gilles.Muller@...6.fr>,
        Nicolas Palix <nicolas.palix@...g.fr>,
        Michal Marek <michal.lkml@...kovi.net>, cocci@...teme.lip6.fr,
        linux-kernel@...r.kernel.org
Subject: Re: [PATCH v3] coccinelle: suggest replacing strncpy+truncation by
 strscpy



On Fri, 20 Jul 2018, Dominique Martinet wrote:

> Julia Lawall wrote on Fri, Jul 20, 2018:
> > > strscpy does however not clear the end of the destination buffer, so
> > > there is a risk of information leak if the full buffer is copied as is
> > > out of the kernel - this needs manual checking.
> >
> > As fasr as I can tell from lkml, only one of these patches has been
> > accepted?  There was also a concern about an information leak that there
> > was no response to.  Actually, I would prefer that more of the generated
> > patches are accepted before accepting the semantic patch, for something
> > that is not quite so obviously correct.
>
> As I'm pointing to the script which generated the patch in the generated
> patches, I got told that it would be better to get the coccinelle script
> accepted first, and asked others to hold on taking the patches at
> several places - I didn't resend any v2 of these with strscpy yet mostly
> for that reason.

I can't accept a semantic patch for which I can't judge the correctness.
It would be better to put a proper commit message in the individual
patches and get them accepted first.

The actual change is made by a script that is only a few lines long.  You
can put those lines in your commit message if you like.

> There were concerns for information leaks that I believe I adressed in
> the specific patch that was pointed out by the concern (I might have
> missed some?), but I'll take the time to check all the patches
> individually before resending as well as filling in better commit
> messages which also was one of the main concerns.
>
> I'm however a bit stuck if I'm waiting for the cocinelle script to be
> accepted to resend the patches, but you're waiting for the individual
> patches to be accepted to take the script... :)
>
>
> I guess there is no value in the script landing first by itself, I'll
> just remove the script path from the commit messages and resend the
> first few this weekend.

It's not that there is no value to the script.  The problem is that I
don't know if the script is correct - I'm not familiar with these string
functions.  Once the script is in the kernel, it stays there beyond your
patches, so I would prefer to know that it is correct up front, rather
than having to remove it afterwards.

julia