[systemd-bugs] [Bug 67273] logind writes out session state file before creating fifo so users end up in CLOSING state

[bugzilla-daemon at freedesktop.org]

https://bugs.freedesktop.org/show_bug.cgi?id=67273

--- Comment #4 from Ray Strode [halfline] <rstrode at redhat.com> ---
(In reply to comment #3)
> No it will briefly have "OPENING" in it. See session_get_state(): it will
> return OPENING if there's the scope job still pending. The problem was
> simply that we saved the state at a time where the scope job was finished
> but the fifo not set up yet...
Ah okay. When I first started looking into this issue with dwalsh, I was
originally looking at a stale git source checkout, and the behavior didn't
match what I was reading, so I updated the tree, and then pretty soon after we
saw the fifo_fd problem in the code.  The last time I read the
session_get_state() function I was looking at the stale source before updating,
and it didn't have an OPENING state.

> > 2) you send out the reply before saving the state file, so, in theory,
> > there's a small race where a client could get the reply and then proceed to
> > check the state file and still read stale information.  It's a pretty small
> > race though.
> 
> Hmm, so that's indeed something to fix! 
> 
> Done so now in 7c6e214aeb141e6d514ddd9237a956278c4afc8c.
cool.

-- 
You are receiving this mail because:
You are the QA Contact for the bug.
You are the assignee for the bug.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.freedesktop.org/archives/systemd-bugs/attachments/20130726/4777fbb2/attachment.html>