linux-kernel - Re: [PATCH -v2 5/5] x86: use dmi check to treat disabled cpus as hotplug cpus.

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: <4B4E5390.1070004@zytor.com>
Date:	Wed, 13 Jan 2010 15:13:20 -0800
From:	"H. Peter Anvin" <hpa@...or.com>
To:	Yinghai Lu <yinghai@...nel.org>
CC:	Suresh Siddha <suresh.b.siddha@...el.com>,
	Andi Kleen <andi@...stfloor.org>,
	Linus Torvalds <torvalds@...ux-foundation.org>,
	"ananth@...ibm.com" <ananth@...ibm.com>,
	Ingo Molnar <mingo@...e.hu>,
	Thomas Gleixner <tglx@...utronix.de>,
	Andrew Morton <akpm@...ux-foundation.org>,
	"linux-kernel@...r.kernel.org" <linux-kernel@...r.kernel.org>
Subject: Re: [PATCH -v2 5/5] x86: use dmi check to treat disabled cpus as
 hotplug cpus.

On 01/13/2010 03:02 PM, Yinghai Lu wrote:
> On 01/13/2010 02:49 PM, Suresh Siddha wrote:
>> On Wed, 2010-01-13 at 14:36 -0800, H. Peter Anvin wrote:
>>> On 01/13/2010 02:29 PM, Suresh Siddha wrote:
>>>>>
>>>>> Well, that *is* working around broken code, in this case the broken code
>>>>> is the percpu allocation strategy.
>>>>
>>>> Andi, Recently percpu folks changed the per-cpu static first chunk to
>>>> PMD SIZE right. I think that is what causing all this issue.
>>>>
>>>
>>> Please don't tell me we're allocating 2 MB per CPU and throwing away
>>> most of it...
>>
>> Looking at the percpu code, they do seem to free the unused memory in
>> that hole.
>>
>> Yinghai, what are the configurations that need MB's of per cpu area?
> 
> please check attached...
> 

I just built an allyesconfig kernel, and it has 1904K of percpu data.
1791K of that is in per_cpu__cpu_lock_stats.  So everything else is
~113K.  Still.  We should not be restricted from doing legitimately
large percpu allocations because of ghost cpus (and Tejun knows my
position on this), but it doesn't sound like the house is burning, either.

	-hpa
--
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/