Re: [for-next][PATCH 14/14] tracing: Get trace_array ref counts when accessing trace files

[Steven Rostedt]

On Tue, 8 Apr 2014 13:32:31 -0400
Steven Rostedt <rostedt@xxxxxxxxxxx> wrote:

> On Tue, 08 Apr 2014 13:11:32 -0400
> Sasha Levin <sasha.levin@xxxxxxxxxx> wrote:
> 
> > >>
> > >> Although when I mount everything for the fuzzer I do it with '-onoexec,nosuid'
> > >> and the fuzzer is banned from testing mount(), so I'm not sure how it would
> > >> do that thing on it's own.
> > > 
> > > What does the -onoexec do? Not change the exec flag of files?
> > 
> > It lets you set it as executable, but it won't let you actually exec it.
> > 
> > > Also, I wonder if this has something to do with the syscall being a
> > > compat sys_exec and not a native one. Is userspace on your vm 32bit? Or
> > > does the fuzzer just try the different compat calls?
> > 
> > Right, everything is 64bit but the fuzzer tries 32bit calls as well.
> 
> When this bug is triggered, is there a way to let the fuzzer tell us
> what exactly the last syscall it sent was? That is, the syscall number
> plus all its arguments?

Just letting you know. I'm holding off looking too deep into this as it
seems to require very strange behavior to trigger (like perhaps execing
a tracing file). This only happens when running as root, right? If not,
then there's a bigger bug involved.

If you can get me more information on this bug, like what I suggested
above, then I'll take another look.

Thanks,

-- Steve
--
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/