Re: rb tree hrtimer lockup bug (found by perf_fuzzer)

From: Thomas Gleixner
Date: Thu Mar 20 2014 - 11:12:24 EST


On Thu, 20 Mar 2014, Vince Weaver wrote:

> On Thu, 20 Mar 2014, Thomas Gleixner wrote:
> > Find a more brute force method to get this decoded below. Full patch
> > this time, please zap the old one. That should spill out the recent
> > debug object calls and we can just hope, that we get enough history to
> > find this horror.
>
> The patch didn't compile, as "addr" was not known. I guessed and
> changed it to "oaddr", but if that's not what you meant then the
> trace results might be useless.

Yeah, it causes the explosion in the debug object code. This should
have been:

+ trace_printk("Tracking Object free: %s %p %pS\n",
+ descr->name, obj->object, obj->hint);


Sorry, -ENOTENOUGHCOFFEE to switch on DEBUG_OBJECTS_FREE=y before
compile testing.

Thanks,

tglx
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/