linux-kernel - Re: [slubllv7 06/17] slub: Add cmpxchg_double

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:	Tue, 12 Jul 2011 13:53:55 -0500 (CDT)
From:	Christoph Lameter <cl@...ux.com>
To:	"H. Peter Anvin" <hpa@...or.com>
cc:	Eric Dumazet <eric.dumazet@...il.com>,
	Pekka Enberg <penberg@...helsinki.fi>,
	David Rientjes <rientjes@...gle.com>,
	linux-kernel@...r.kernel.org, Thomas Gleixner <tglx@...utronix.de>
Subject: Re: [slubllv7 06/17] slub: Add cmpxchg_double_slab()

On Tue, 12 Jul 2011, H. Peter Anvin wrote:

> On 07/12/2011 09:47 AM, Christoph Lameter wrote:
> > On Tue, 12 Jul 2011, Eric Dumazet wrote:
> >
> >> Le mardi 12 juillet 2011 à 10:59 -0500, Christoph Lameter a écrit :
> >>
> >>> We do not need to have _count included. This is just there because the
> >>> field is in the way on 64 bit and we can only do 2x 64 bit cmpxchges.  On
> >>> 32 bi we can drop _count from "counters".
> >>
> >> OK, thanks for clarification.
> >
> > Still I'd like to get some ideas on how to make the whole thing much
> > cleaner. Isnt there some way to convert a struct to an unsigned long
> > without going through a union? And a way to convert a struct + 32 bit
> > atomic_t into a 64 bit unsigned long? Would simplify things significantly.
>
> If you know it is in memory you can cast ("pun") the pointer.

Well I use a page struct on the stack now to put the information in memory
and then use the union to cast it.

> If it's not in memory that can be inefficient, though.

Yeah. Isnt there some C trick to cast a word size struct to unsigned long?