linux-kernel - Re: frequent lockups in 3.18rc4

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 for Android: free password hash cracker in your pocket

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Date:	Sat, 6 Dec 2014 23:38:13 +0100 (CET)
From:	Thomas Gleixner <tglx@...utronix.de>
To:	Linus Torvalds <torvalds@...ux-foundation.org>
cc:	Dave Jones <davej@...hat.com>, Chris Mason <clm@...com>,
	Mike Galbraith <umgwanakikbuti@...il.com>,
	Ingo Molnar <mingo@...nel.org>,
	Peter Zijlstra <peterz@...radead.org>,
	Dâniel Fraga <fragabr@...il.com>,
	Sasha Levin <sasha.levin@...cle.com>,
	"Paul E. McKenney" <paulmck@...ux.vnet.ibm.com>,
	Linux Kernel Mailing List <linux-kernel@...r.kernel.org>
Subject: Re: frequent lockups in 3.18rc4

On Fri, 5 Dec 2014, Linus Torvalds wrote:
> On Fri, Dec 5, 2014 at 10:48 AM, Dave Jones <davej@...hat.com> wrote:
> >
> > In the meantime, I rebooted into the same kernel, and ran trinity
> > solely doing the lsetxattr syscalls.
> 
> Any particular reason for the lsetxattr guess? Just the last call
> chain? I don't recognize it from the other traces, but maybe I just
> didn't notice.
> 
> >   The load was a bit lower, so I
> > cranked up the number of child processes to 512, and then this
> > happened..
> 
> Ugh. "dump_trace()" being broken and looping forever? I don't actually

Looking at the callchain: up to the point where dump_stack() is called
everything is preemtible context. So dump_stack() would need to loop
for a few seconds to trigger the NMI watchdog.

> believe it, because this isn't even on the exception stack (well, the
> NMI dumper is, but that one worked fine - this is the "nested" dumping
> of just the allocation call chain)

I doubt that dump_trace() itself is broken, but the call site might
have handed in something which causes memory corruption. And looking
at set_track() and the completely undocumented way how it retrieves
the storage for the trace entries via get_track() makes my brain melt.

Thanks,

	tglx







--
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/