On 04/16/13 13:22, Gerd Hoffmann wrote: > On 04/16/13 13:10, Paolo Bonzini wrote: >> I'm happy to say that I'm not touching IOWatchPoll (well, almost: only >> for consistency). Instead, these patches try to make the code consistent >> (thus avoiding CRITICAL messages from glib) and to fix detection of pty >> connections; Gerd reported that it went berserk with polling. I think it >> is the same failure reported by Michael Hines, but I couldn't reproduce >> it with "-serial pty", only with the monitor. > > Things are back to normal with the series applied. > > Tested-by: Gerd Hoffmann <kra...@redhat.com> > > cheers, > Gerd
Trapped into the next issue. Added qmp monitor to the virtual machine, linked to unix socket. Start playing with the qom-* scrips in QMP. qemu hangs after the first qmp-list command (triggered by socket disconnect?). cheers, Gerd
