netdev - Re: 2.6.24 BUG: soft lockup

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: <20080328141246.GD29218@elte.hu>
Date:	Fri, 28 Mar 2008 15:12:48 +0100
From:	Ingo Molnar <mingo@...e.hu>
To:	Herbert Xu <herbert@...dor.apana.org.au>
Cc:	jamal <hadi@...erus.ca>, "David S. Miller" <davem@...emloft.net>,
	Matheos.Worku@....COM, jesse.brandeburg@...el.com,
	jarkao2@...il.com, netdev@...r.kernel.org
Subject: Re: 2.6.24 BUG: soft lockup - CPU#X


* Herbert Xu <herbert@...dor.apana.org.au> wrote:

> I agree that using jiffies is a pretty coarse approximation of proper 
> scheduling.  However, in the absence of a better solution we have to 
> live with it.
> 
> Perhaps running these out of process context is the correct approach.

yes. Such anonymous work loops inside softirq context are a disaster to 
TCP determinism and a disaster to scheduling in general (the wrong guy 
gets credited with the overhead). Softirqs were a neat hack 10 years 
ago, now if we know the target task for some workload we should execute 
as much of the workload in that task's context as possible. (and even 
for stuff where we dont have a 'target task' - routing, filtering, etc. 
- it might be better to use kernel threads.)

	Ingo
--
To unsubscribe from this list: send the line "unsubscribe netdev" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html