From mboxrd@z Thu Jan 1 00:00:00 1970 Date: Wed, 26 Jul 2006 12:47:05 -0700 (PDT) From: Christoph Lameter Subject: Re: [patch 2/2] slab: always consider arch mandated alignment In-Reply-To: <44C7C46C.4090201@colorfullife.com> Message-ID: References: <20060723073500.GA10556@osiris.ibm.com> <20060723162427.GA10553@osiris.ibm.com> <20060726085113.GD9592@osiris.boeblingen.de.ibm.com> <20060726101340.GE9592@osiris.boeblingen.de.ibm.com> <20060726105204.GF9592@osiris.boeblingen.de.ibm.com> <44C7AF31.9000507@colorfullife.com> <44C7B842.5060606@colorfullife.com> <44C7C261.6050602@colorfullife.com> <44C7C46C.4090201@colorfullife.com> MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Sender: owner-linux-mm@kvack.org Return-Path: To: Manfred Spraul Cc: Pekka J Enberg , Heiko Carstens , Andrew Morton , linux-kernel@vger.kernel.org, linux-mm@kvack.org, Martin Schwidefsky List-ID: On Wed, 26 Jul 2006, Manfred Spraul wrote: > Christoph Lameter wrote: > > > A slab user is setting alignment in order to optimize performance not for > > correctness. Most users that I know of can live with misalignments. If that > > would not be the case then this code would never have worked. > > > > Which users do you know that set align and that can live with misalignments? > As I wrote, there are no such users in my (i386) kernel. The users of SLAB_HWCACHE_ALIGN can live with that. Systems running with slab debugging on must be very buggy at this point or we were very lucky: The list is a bit strange: >* the pmd structure (4096: hardware requirement) It is already exempted from debug since the size is 4096. >* the pgd structure (32 bytes: hardware requirement) We were lucky on that one in the past? This should break. >* the task structure (16 byte. fxsave) Would only break if floating point is used I think. >* sigqueue, pid: both request 4 byte alignment (based on __alignof__()). >Doesn't affect debugging. So also not relevant. We now want to say that SLAB_HWCACHE_ALIGN is only a suggestion to be disposed of if debug is on whereas an explicitly specified alignment must be enforced? -- To unsubscribe, send a message with 'unsubscribe linux-mm' in the body to majordomo@kvack.org. For more info on Linux MM, see: http://www.linux-mm.org/ . Don't email: email@kvack.org