[<prev] [next>] [thread-next>] [day] [month] [year] [list]
Message-ID: <20190714095157.GA2276@avx2>
Date: Sun, 14 Jul 2019 12:51:57 +0300
From: Alexey Dobriyan <adobriyan@...il.com>
To: Linus Torvalds <torvalds@...ux-foundation.org>
Cc: Andrew Morton <akpm@...ux-foundation.org>, izbyshev@...ras.ru,
Oleg Nesterov <oleg@...hat.com>,
Michal Kubecek <mkubecek@...e.cz>, shasta@...corp.com,
linux-kernel@...r.kernel.org, security@...nel.org
Subject: Re: [PATCH] proc: revert /proc/*/cmdline rewrite
[re-add lists]
On Sat, Jul 13, 2019 at 10:50:20AM -0700, Linus Torvalds wrote:
> On Sat, Jul 13, 2019 at 12:29 AM Alexey Dobriyan <adobriyan@...il.com> wrote:
> >
> > /proc/*/cmdline continues to cause problems:
>
> If we're reverting this, then we should revert all the way back to the
> original fixed-length one that had the original semantics and was
> simple.
No, because all those Java applications have command lines measured in
dozens of kilobytes.
> What was the problem with the one-line fix instead?
The problem is that I can't even drag this trivia in out of _fear_ that
it is userspace observable:
https://marc.info/?t=155863429700002&r=1&w=4
[PATCH] elf: fix "start_code" evaluation
and yet the patch which did a regression and an infoleak continues
to be papered over and for which the only justification was
"simplify and clarify".
Powered by blists - more mailing lists