linux-kernel - Re: Question on lockdep and MAX_LOCK

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: <5111BD28.50407@candelatech.com>
Date:	Tue, 05 Feb 2013 18:17:12 -0800
From:	Ben Greear <greearb@...delatech.com>
To:	Steven Rostedt <rostedt@...dmis.org>
CC:	Linux Kernel Mailing List <linux-kernel@...r.kernel.org>,
	Peter Zijlstra <peterz@...radead.org>,
	Ingo Molnar <mingo@...nel.org>
Subject: Re: Question on lockdep and MAX_LOCK_DEPTH

On 02/05/2013 05:54 PM, Steven Rostedt wrote:

> I'm not sure the swapper task is part of the 'do_each_thread()' loop.
>
> Perhaps what you want to do is add a:
>
> 	lockdep_print_held_locks(current);

I'll add that and test...

> I'm curious. Does your code grab a read lock? If you grab the same read
> lock multiple times it adds to the list each time. Thus if you have
> anything like:
>
> 	for (i = 0; i < 100; i++ ) {
> 		read_lock(&lock);
> 	}
>
> 	for (i = 0; i < 100; i++) {
> 		read_unlock(&lock);
> 	}
>
> That will fill up the held locks quite a bit.
>
> The above code I showed is ridiculous and I doubt you have it, but if
> you have something that does lots of recursive reads for some reason,
> that could be an issue.

I have only one read/write lock in my module, and it looks like
I always lock it as a writer (will fix that soon for performance
reasons, but it should be valid locking I think).

I have no rcu locks at all in my module currently.

I've seen similar lockups on another machine that does not use
this module, but it uses a hacked up pktgen.  I haven't found
a test case that reproduces this on a clean upstream build,
but I am still fairly suspicious that it isn't my code.
Famous last words I'm sure :)

Thanks,
Ben

-- 
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/