Control: tag -1 pending On Sun, Dec 27, 2015 at 04:00:50PM +0100, Michael Biebl wrote: > Am 26.12.2015 um 17:27 schrieb Colin Watson: > > Michael, this looks like a regression from your readiness notification > > changes that I applied recently. Please could you have a look? > > Attached is a patch on top of the existing one, which fixes the issue by > running the sd_notify() call only in the main process and not the > spawned off children. > It moves it right next to the existing SIGSTOP readiness-signal.
> From 60aab01587f4974261882b7c4066750f34522ea4 Mon Sep 17 00:00:00 2001 > From: Michael Biebl <[email protected]> > Date: Sun, 27 Dec 2015 15:58:03 +0100 > Subject: [PATCH] Call sd_notify() only for the main process > > Move the sd_notify() call next to the existing SIGSTOP readiness > notification so it is only run in the main process and not for the > spawned off children. It's not actually due to per-login child processes, which come rather later; the true reason is that when sshd is re-execed it sets inetd_flag. But your patch does the job either way and I'll just merge it into your previous patch. Thanks, -- Colin Watson [[email protected]]
