Re: [PATCH v2] fprobe: Ensure running fprobe_exit_handler() finished before calling rethook_free()

[Masami Hiramatsu (Google) <mhiramat@kernel.org>]

On Mon, 10 Jul 2023 18:04:22 -0400
Steven Rostedt <rostedt@goodmis.org> wrote:

> On Fri,  7 Jul 2023 23:03:19 +0900
> "Masami Hiramatsu (Google)" <mhiramat@kernel.org> wrote:
> 
> > From: Masami Hiramatsu (Google) <mhiramat@kernel.org>
> > 
> > Ensure running fprobe_exit_handler() has finished before
> > calling rethook_free() in the unregister_fprobe() so that caller can free
> > the fprobe right after unregister_fprobe().
> > 
> > unregister_fprobe() ensured that all running fprobe_entry/exit_handler()
> > have finished by calling unregister_ftrace_function() which synchronizes
> > RCU. But commit 5f81018753df ("fprobe: Release rethook after the ftrace_ops
> > is unregistered") changed to call rethook_free() after
> > unregister_ftrace_function(). So call rethook_stop() to make rethook
> > disabled before unregister_ftrace_function() and ensure it again.
> > 
> > Here is the possible code flow that can call the exit handler after
> > unregister_fprobe().
> > 
> > ------
> >  CPU1                              CPU2
> >  call unregister_fprobe(fp)
> >  ...
> >                                    __fprobe_handler()
> >                                    rethook_hook() on probed function
> >  unregister_ftrace_function()
> >                                    return from probed function
> >                                    rethook hooks
> >                                    find rh->handler == fprobe_exit_handler
> >                                    call fprobe_exit_handler()
> >  rethook_free():
> >    set rh->handler = NULL;
> >  return from unreigster_fprobe;
> >                                    call fp->exit_handler() <- (*)
> > ------
> > 
> > (*) At this point, the exit handler is called after returning from
> > unregister_fprobe().
> > 
> > This fixes it as following;
> > ------
> >  CPU1                              CPU2
> >  call unregister_fprobe()
> >  ...
> >  rethook_stop():
> >    set rh->handler = NULL;
> >                                    __fprobe_handler()
> >                                    rethook_hook() on probed function
> >  unregister_ftrace_function()
> >                                    return from probed function
> >                                    rethook hooks
> >                                    find rh->handler == NULL
> >                                    return from rethook
> >  rethook_free()
> >  return from unreigster_fprobe;
> > ------
> > 
> > 
> > Fixes: 5f81018753df ("fprobe: Release rethook after the ftrace_ops is unregistered")
> > Cc: stable@vger.kernel.org
> > Signed-off-by: Masami Hiramatsu (Google) <mhiramat@kernel.org>
> 
> Looks good.
> 
> Reviewed-by: Steven Rostedt (Google) <rostedt@goodmis.org>

Thank you :)

> 
> 
> > ---
> >  Changes in v2:
> >  - Update changelog to add a problematic code flow.
> 
> Nit, for making forensic analysis easier in the future, I now always add a
> link to the previous version. That is:
> 
> Changes since v1: https://lore.kernel.org/linux-trace-kernel/168796344232.46347.7947681068822514750.stgit@devnote2/
> - Update changelog to add a problematic code flow.

OK, I'll add it for an isolated patch too.

Thanks!

> 
> -- Steve
> 
> 
> > ---
> >  include/linux/rethook.h |    1 +
> >  kernel/trace/fprobe.c   |    3 +++
> >  kernel/trace/rethook.c  |   13 +++++++++++++
> >  3 files changed, 17 insertions(+)
> > 


-- 
Masami Hiramatsu (Google) <mhiramat@kernel.org>