linux-kernel - Re: [PATCH 1/2] treewide: fix memory corruptions when TASK_COMM

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: <20120131190109.b53347fd.akpm@linux-foundation.org>
Date:	Tue, 31 Jan 2012 19:01:09 -0800
From:	Andrew Morton <akpm@...ux-foundation.org>
To:	Jan Engelhardt <jengelh@...ozas.de>
Cc:	linux-kernel@...r.kernel.org
Subject: Re: [PATCH 1/2] treewide: fix memory corruptions when TASK_COMM_LEN
 != 16

On Wed, 1 Feb 2012 03:15:50 +0100 (CET) Jan Engelhardt <jengelh@...ozas.de> wrote:

> 
> On Wednesday 2012-02-01 02:49, Andrew Morton wrote:
> >> 
> >> Did my patch not change the existing code sites using ->comm
> >> to always copy at most min(userbufsize aka 16, sizeof(t->comm)) bytes,
> >> thereby keeping the promise to userspace while at the same time
> >> making TASK_COMM_LEN's value freely choosable?
> >
> >That change is pretty pointless as long as we don't provide APIs to let
> >userspace access the expanded size.  And I've explained why we cannot
> >alter the existing APIs.
> 
> Ah yes, indeed. My reason for augmenting the size of t->comm was so
> that `ps afx` could show a more complete name of certain kernel
> threads' names. In this case, the kernel delivers the name via
> procfs via seq_printf("%s, t->comm),

Where does procfs do this?

> as do a few debug statements
> in the fashion of pr_debug("%s/%u ate my CPU", t->comm, t->pid).
> So maybe it was not /completely/ pointless.

I agree that the 16-char thing is irritatingly small.  But if we were
to increase it and to then utilise that increase, those userspace apps
which are still using the legacy prctl(PR_GET_NAME) would produce
pretty bad output.  Instead of "migration/0" and "migration/1" you'd
get "migration_threa" and "migration_threa".  And "flush-8:32" would
maddeningly become "flusher_thread-".

I suppose that would help motivate people to update their tools ;)
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/