Re: LOCKDEP: 3.9-rc1: mount.nfs/4272 still has locks held!

[Tejun Heo <tj@kernel.org>]

Hello, Jeff.

On Tue, Mar 05, 2013 at 06:39:41PM -0500, Jeff Layton wrote:
> Al was in the middle of his signal handling/execve rework though and I
> ran the idea past him. He pointedly told me that I was crazy for even
> considering it. This is rather non-trivial to handle since it means
> mucking around in a bunch of arch-specific code and dealing with all of
> the weirdo corner cases.

Hmmm... I tried it a couple years ago while I was working on ptrace
improvements to support checkpoint-restart in userland (PTRACE_SEIZE
and friends) and had a half-working code.  At the time, Oleg was
against the idea and something else came up so I didn't push it all
the way but IIRC it didn't need to touch any arch code.  It should be
able to just share the existing trap path with ptrace.

> Anyone up for working out how to handle a freeze event on a process
> that already has a pending signal, while it's being ptraced? It's
> probably best to chat with Al (cc'ed here) before you embark on this
> plan since he was just in that code recently.

Maybe Al sees something that I don't but AFAICS it should be doable in
generic code proper and I don't think it's gonna be that hard.

Oleg, are you still opposed to the idea of making freezer share trap
points with ptrace?

Thanks.

-- 
tejun