linux-kernel - Re: ipc,sem: sysv semaphore scalability

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: <CA+55aFyqhacFz9xc3igO2mpmwFHdMKGGMVbbrUiOOrkyjVxMXg@mail.gmail.com>
Date:	Sat, 30 Mar 2013 10:22:49 -0700
From:	Linus Torvalds <torvalds@...ux-foundation.org>
To:	Emmanuel Benisty <benisty.e@...il.com>
Cc:	Davidlohr Bueso <davidlohr.bueso@...com>,
	Dave Jones <davej@...hat.com>,
	Andrew Morton <akpm@...ux-foundation.org>,
	Rik van Riel <riel@...riel.com>,
	Linux Kernel Mailing List <linux-kernel@...r.kernel.org>,
	hhuang@...hat.com, "Low, Jason" <jason.low2@...com>,
	Michel Lespinasse <walken@...gle.com>,
	Larry Woodman <lwoodman@...hat.com>,
	"Vinod, Chegu" <chegu_vinod@...com>,
	Peter Hurley <peter@...leysoftware.com>
Subject: Re: ipc,sem: sysv semaphore scalability

On Fri, Mar 29, 2013 at 10:57 PM, Emmanuel Benisty <benisty.e@...il.com> wrote:
> On Sat, Mar 30, 2013 at 12:10 PM, Linus Torvalds
>>
>> This came from the gcc build?
>
> yes, very early in the build process, IIRC this line was repeated a
> few times and the build just stalled.

Ok, we're bringing out the crazy hacks now.

The attached patch is just insane, doesn't really even work in
general, and only even compiles on 64-bit. But it should work in
*practice* to find if somebody adds the same RCU head to the RCU lists
twice, and ignore the second time it happens (and give a warning that
hopefully pinpoints the backtrace).

It's ugly. It's broken. It may not work. In other words, I'm not proud
of it. But you seem to be the only one able to trigger the issue
easily, willing to try crazy crap, so "tag, you're it". Maybe this
gives us more information. And maybe it doesn't, and I'm totally wrong
about the whole "rcu head added twice" theory.

                        Linus

Download attachment "patch.diff" of type "application/octet-stream" (2253 bytes)