On Wed, Jun 10, 2009 at 9:17 AM, Thomas Pfaff<tpf...@tp76.info> wrote: > I'm getting a lot of "hotplugd[pid]: waitpid: Error 10" (No child processes) > when I attach and detach umass(4) devices. This happens almost every time. > > Seems hotplugd is getting an extra SIGCHLD from somewhere. I guess this > issue is pretty harmless (only pollutes the logs), but it got me curious. > > Adding a bit more verbosity to hotplugd(8) my /var/log/daemon shows this: ... > Jun 10 18:09:44 ws hotplugd[9309]: fork[5252]: child = 9309 > Jun 10 18:09:44 ws hotplugd[28984]: fork[5252]: child = 28984 > Jun 10 18:09:44 ws hotplugd[5943]: fork[5252]: child = 5943 > Jun 10 18:09:44 ws hotplugd[5252]: waitpid: child pid = 9309 > Jun 10 18:09:44 ws hotplugd[5252]: waitpid: child pid = 28984 > Jun 10 18:09:44 ws hotplugd[5252]: waitpid: child pid = 5943 > Jun 10 18:09:44 ws hotplugd[5252]: waitpid: No child processes
It's easy enough to change the code to suppress the error message in that case, but before doing so it would be interesting to see whether a ktrace of hotplugd across one of these incidents would show the extra SIGCHLD delivery. Can you give that a shot? Philip Guenther
