From mboxrd@z Thu Jan 1 00:00:00 1970 From: Alexey Dobriyan Subject: Re: [PATCH] net: make sure struct dst_entry refcount is aligned on 64 bytes Date: Fri, 14 Nov 2008 14:35:58 +0300 Message-ID: <20081114113558.GA9692@x200.localdomain> References: <491D323B.9030802@cosmosbay.com> <20081114.005437.09284570.davem@davemloft.net> <491D3F18.5030505@cosmosbay.com> <20081114093613.GA2834@x200.localdomain> <491D5725.50006@cosmosbay.com> Mime-Version: 1.0 Content-Type: text/plain; charset=iso-8859-1 Content-Transfer-Encoding: QUOTED-PRINTABLE Cc: David Miller , netdev@vger.kernel.org, shemminger@vyatta.com, "Zhang, Yanmin" To: Eric Dumazet Return-path: Received: from ug-out-1314.google.com ([66.249.92.175]:21073 "EHLO ug-out-1314.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1750936AbYKNLc3 (ORCPT ); Fri, 14 Nov 2008 06:32:29 -0500 Received: by ug-out-1314.google.com with SMTP id 39so1613429ugf.37 for ; Fri, 14 Nov 2008 03:32:28 -0800 (PST) Content-Disposition: inline In-Reply-To: <491D5725.50006@cosmosbay.com> Sender: netdev-owner@vger.kernel.org List-ID: On Fri, Nov 14, 2008 at 11:47:01AM +0100, Eric Dumazet wrote: > Alexey Dobriyan a =E9crit : >> On Fri, Nov 14, 2008 at 10:04:24AM +0100, Eric Dumazet wrote: >>> David Miller a =E9crit : >>>> From: Eric Dumazet >>>> Date: Fri, 14 Nov 2008 09:09:31 +0100 >>>> >>>>> During tbench/oprofile sessions, I found that dst_release() was i= n third position. >>>> ... >>>>> Instead of first checking the refcount value, then decrement it, >>>>> we use atomic_dec_return() to help CPU to make the right memory t= ransaction >>>>> (ie getting the cache line in exclusive mode) >>>> ... >>>>> Signed-off-by: Eric Dumazet >>>> This looks great, applied, thanks Eric. >>>> >>> Thanks David >>> >>> >>> I think I understood some regressions here on 32bits=20 >>> >>> offsetof(struct dst_entry, __refcnt) is 0x7c again !!! >>> >>> This is really really bad for performance >>> >>> I believe this comes from a patch from Alexey Dobriyan >>> (commit def8b4faff5ca349beafbbfeb2c51f3602a6ef3a >>> net: reduce structures when XFRM=3Dn) >> >> Ick. > > Well, your patch is a good thing, we only need to make adjustments. > >> >>> This kills effort from Zhang Yanmin (and me...) >>> >>> (commit f1dd9c379cac7d5a76259e7dffcd5f8edc697d17 >>> [NET]: Fix tbench regression in 2.6.25-rc1) >>> >>> >>> Really we must find something so that this damned __refcnt is start= ing at 0x80 >> >> Make it last member? > > Yes, it will help tbench, but not machines that stress IP route cache > > (dst_use() must dirty the three fields "refcnt, __use , lastuse" ) > > Also, 'next' pointer should be in the same cache line, to speedup rou= te > cache lookups. Knowledge taken. > Next problem is that offsets depend on architecture being 32 or 64 bi= ts. > > On 64bit, offsetof(struct dst_entry, __refcnt) is 0xb0 : not very goo= d... I think all these constraints can be satisfied with clever rearranging = of dst_entry. Let me come up with alternative patch which still reduces dst slab size= =2E