linux-kernel - Re: [PATCH] proc: Don't allow empty /proc/PID/cmdline for user tasks

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]

Message-ID: <CA+55aFxdnFpNoufsad2cCic8iz2MYQ8h_KV+5sM+WF9KAuJk_Q@mail.gmail.com>
Date:   Thu, 17 May 2018 11:23:37 -0700
From:   Linus Torvalds <torvalds@...ux-foundation.org>
To:     Tejun Heo <tj@...nel.org>
Cc:     Linux Kernel Mailing List <linux-kernel@...r.kernel.org>,
        Linux API <linux-api@...r.kernel.org>,
        Andrew Morton <akpm@...ux-foundation.org>,
        kernel-team <kernel-team@...com>,
        Lennart Poettering <lennart@...ttering.net>
Subject: Re: [PATCH] proc: Don't allow empty /proc/PID/cmdline for user tasks

On Wed, May 16, 2018 at 6:21 PM Tejun Heo <tj@...nel.org> wrote:

> This patch fixes the issue by making proc_pid_cmdline_read() never
> return empty string for user tasks.

Ugh.

That function really is too damn ugly, and this just makes it worse.

Can we please split things up a bit before uglifying the code further?

IOW, the first step should be something like the attached patch, which
splits up all the tsk/mm error cases.

Then your patch could just be a trivial

     if (!*pos && !ret)
         ret = get_comm_cmdline(tsk, buf, count, pos);

in that new (and much simpler) proc_pid_cmdline_read() just before the
put_task_struct.

Hmm?

NOTE! This patch is *entirely* untested, but it builds and the conversion
was pretty much entirely mechanical.

And yes, the "get_mm_cmdline()" function is still too damn ugly, and should
still be cleaned up more, but it's at least _slightly_ simpler than it used
to be, and the new logic wouldn't go into that horrible thing.

                Linus

View attachment "patch.diff" of type "text/x-patch" (2382 bytes)