[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-ID: <20140130085533.5436dc2a@gandalf.local.home>
Date: Thu, 30 Jan 2014 08:55:33 -0500
From: Steven Rostedt <rostedt@...dmis.org>
To: Dave Jones <davej@...hat.com>
Cc: Linux Kernel <linux-kernel@...r.kernel.org>,
Peter Zijlstra <peterz@...radead.org>,
Jiri Olsa <jolsa@...hat.com>
Subject: Re: FTRACE_WARN_ON((rec->flags & ~FTRACE_FL_MASK) == 0))
On Thu, 30 Jan 2014 00:32:32 -0500
Dave Jones <davej@...hat.com> wrote:
> On Wed, Jan 29, 2014 at 11:50:43PM -0500, Steven Rostedt wrote:
> > Are you running as root? If not, you found another way to get perf to start function tracing.
>
> Good point. In this case, I was trying some new experimental trinity code
> that starts as root, generates fd's, then drops privs before doing syscalls.
>
> So the "generate fds" part did some perf_event_open's as root, yeah.
>
> While that's less scary from a security pov than it was last time, it's still
> something that aparently needs fixing.
>
OK, then I'm not as worried (phew!). But you are correct, this needs to
be fixed. How do I run just perf commands from trinity? I do not think
this occurs from any other systemcall needing to be run.
Peter, Jiri,
Although we fixed the bug that let a normal user enable function
tracing from perf, we never figured out why perf screws up the function
tracing filter accounting. That will happen if you either use a
uninitialized ftrace hash, or modify the hash without using the ftrace
API. If it is the latter, it is probably done by some kind of use after
free.
-- Steve
--
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/
Powered by blists - more mailing lists