linux-kernel - Re: [accounting regression since rc1] scheduler updates

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: <20070821112032.GA648@elte.hu>
Date:	Tue, 21 Aug 2007 13:20:32 +0200
From:	Ingo Molnar <mingo@...e.hu>
To:	Andi Kleen <andi@...stfloor.org>
Cc:	Martin Schwidefsky <schwidefsky@...ibm.com>,
	Christian Borntraeger <borntraeger@...ibm.com>,
	Linus Torvalds <torvalds@...ux-foundation.org>,
	Andrew Morton <akpm@...ux-foundation.org>,
	linux-kernel@...r.kernel.org,
	Jan Glauber <jang@...ux.vnet.ibm.com>,
	heiko.carstens@...ibm.com, Paul Mackerras <paulus@...ba.org>
Subject: Re: [accounting regression since rc1]  scheduler updates


* Andi Kleen <andi@...stfloor.org> wrote:

> > i know there are some incredibly broken (but rare) boxes where the bios 
> > will report it only knows C1 and do C2? Is that the case you are 
> > referring to, or is there something else too?
> 
> There are first a couple of older and not so old (Centaur) chips that 
> generally stop the TSC in C1.

there's not much we can do about them: the ACPI code doesnt measure 
their idle time, right? This is mostly for statistics purposes, so 
unless "broken" means tons of boxes, we dont have to have 100% coverage.

> And also some boxes who shouldn't have anything deeper than C2 have 
> trouble with the TSC. For example I got a Opteron machine (which 
> definitely shouldn't have any C2 since it's two socket) where the TSC 
> appears to stop or at least slow down a lot in C1.

how much is "a lot" in C1? There's an AMD TSC-slows-down-C1 erratum but 
that should be less than 1%. (which is fine enough for idle time 
measurements)

> And thirdly it's just unclean to add all kinds of custom hooks there. 
> It was already ugly in NOHZ, please don't continue that.

i'm not opposed to the idle notifiers but iirc the idle notifiers caused 
problems in themselves so part of them were reverted. We can do this 
more cleanly in .24 - it will make the benefits of the notifier cleanup 
even more apparent.

	Ingo
-
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/