Re: [PATCH v2 net-next 00/14] af_unix: Rework GC.

[Kuniyuki Iwashima <kuniyu@amazon.com>]

From: Eric Dumazet <edumazet@google.com>
Date: Fri, 16 Feb 2024 22:28:11 +0100
> On Fri, Feb 16, 2024 at 10:06 PM Kuniyuki Iwashima <kuniyu@amazon.com> wrote:
> >
> > When we pass a file descriptor to an AF_UNIX socket via SCM_RIGTHS,
> > the underlying struct file of the inflight fd gets its refcount bumped.
> > If the fd is of an AF_UNIX socket, we need to track it in case it forms
> > cyclic references.
> >
> > Let's say we send a fd of AF_UNIX socket A to B and vice versa and
> > close() both sockets.
> >
> > When created, each socket's struct file initially has one reference.
> > After the fd exchange, both refcounts are bumped up to 2.  Then, close()
> > decreases both to 1.  From this point on, no one can touch the file/socket.
> 
> Note that I have pending syzbot reports about af_unix. (apparently
> syzbot found its way with SO_PEEK_OFF)

Interesting :)

> 
> I would prefer we wait a bit before reworking the whole thing.

Sure, I'll wait to avoid a sad situation that large refactoring
happens to fix the bug and we need to backport or cook another
fix for stable.