Re: perf: fuzzer triggers NULL pointer derefreence in x86_schedule_events

[Peter Zijlstra <peterz@infradead.org>]

On Mon, May 18, 2015 at 01:40:31PM -0400, Vince Weaver wrote:
> On Thu, 7 May 2015, Peter Zijlstra wrote:
> 
> > On Mon, May 04, 2015 at 12:32:56PM -0700, Stephane Eranian wrote:
> > > I think it is more likely related to the bitmask (idxmsk).  But then
> > > it is always allocated with the constraint even with the HT bug
> > > workaround.  So most, likely the index is bogus and you touch outside
> > > the idxmsk[] array.
> > 
> > [428232.701319] BUG: unable to handle kernel NULL pointer dereference at           (null)
> > 
> > But the thing really tried to touch NULL, not some random address that
> > faulted.
> > 
> > As always, Vince has found us a good puzzle ;-)
> 
> so the Haswell machine turned up the following oops that looks related.
> 
> Yet again we are ending up with a NULL pointer in the constraint table 
> somehow.
> 
> This maps to 
> 
> static bool __perf_sched_find_counter(struct perf_sched *sched)
> 
>         c = sched->events[sched->state.event]->hw.constraint;
> 
>         /* Prefer fixed purpose counters */
> --->	if (c->idxmsk64 & (~0ULL << INTEL_PMC_IDX_FIXED)) {
> 
> ffffffff81029ce4:       48 8b 55 88             mov    -0x78(%rbp),%rdx
> ffffffff81029ce8:       48 8b 04 c2             mov    (%rdx,%rax,8),%rax
> ffffffff81029cec:       ba 20 00 00 00          mov    $0x20,%edx
> ffffffff81029cf1:       48 8b 98 98 01 00 00    mov    0x198(%rax),%rbx
> ffffffff81029cf8:       4c 85 23                test   %r12,(%rbx)
> 
> 
> [306672.100641] BUG: unable to handle kernel NULL pointer dereference at           (null)
> [306672.109653] IP: [<ffffffff81029cf8>] perf_assign_events+0xa8/0x290

So new in this release is:

static void intel_put_event_constraints(struct cpu_hw_events *cpuc,
					struct perf_event *event)
{
	...

	/* cleanup dynamic constraint */
	if (c && (c->flags & PERF_X86_EVENT_DYNAMIC))
		event->hw.constraint = NULL;
}

Which is the only place that value is ever cleared...

Now, I've not quite figured out how that can intersect with scheduling,
typically we only call put_event_constraints() when we're done with the
event.