linux-kernel - Re: [PATCH] slub: reduce overhead of slub

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, 28 Jun 2011 14:04:11 -0700
From:	Ben Greear <greearb@...delatech.com>
To:	David Rientjes <rientjes@...gle.com>
CC:	David Daney <ddaney.cavm@...il.com>,
	Christoph Lameter <cl@...ux.com>,
	Marcin Slusarz <marcin.slusarz@...il.com>,
	Pekka Enberg <penberg@...nel.org>,
	Matt Mackall <mpm@...enic.com>,
	LKML <linux-kernel@...r.kernel.org>, linux-mm@...ck.org
Subject: Re: [PATCH] slub: reduce overhead of slub_debug

On 06/28/2011 01:58 PM, David Rientjes wrote:
> On Tue, 28 Jun 2011, David Daney wrote:
>
>> On 06/28/2011 12:32 PM, Christoph Lameter wrote:
>>> On Sun, 26 Jun 2011, Marcin Slusarz wrote:
>>>
>>>> slub checks for poison one byte by one, which is highly inefficient
>>>> and shows up frequently as a highest cpu-eater in perf top.
>>>
>>> Ummm.. Performance improvements for debugging modes? If you need
>>> performance then switch off debuggin.
>>
>> There is no reason to make things gratuitously slow.  I don't know about the
>> merits of this particular patch, but I must disagree with the general
>> sentiment.
>>
>> We have high performance tracing, why not improve this as well.
>>
>> Just last week I was trying to find the cause of memory corruption that only
>> occurred at very high network packet rates.  Memory allocation speed was
>> definitely getting in the way of debugging.  For me, faster SLUB debugging
>> would be welcome.
>>
>
> SLUB debugging is useful only to diagnose issues or test new code, nobody
> is going to be enabling it in production environment.  We don't need 30
> new lines of code that make one thing slightly faster, in fact we'd prefer
> to have as simple and minimal code as possible for debugging features
> unless you're adding more debugging coverage.

If your problem happens under load, then the overhead of slub could significantly
change the behaviour of the system.  Anything that makes it more efficient without
unduly complicating code should be a win.  That posted patch wasn't all that
complicated, and even if it has bugs, it could be fixed easily enough.

Thanks,
Ben

> --
> To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
> the body of a message to majordomo@...r.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html
> Please read the FAQ at  http://www.tux.org/lkml/


-- 
Ben Greear <greearb@...delatech.com>
Candela Technologies Inc  http://www.candelatech.com

--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/