On Fri, Apr 19, 2013 at 8:03 PM, Stefan Hajnoczi <stefa...@redhat.com> wrote: > On Fri, Apr 19, 2013 at 01:58:40PM +0800, liu ping fan wrote: >> On Thu, Apr 18, 2013 at 10:34 PM, Stefan Hajnoczi <stefa...@gmail.com> wrote: >> > On Wed, Apr 17, 2013 at 04:39:15PM +0800, Liu Ping Fan wrote: >> > I'm a little worried that we're lacking G_IO_HUP | G_IO_ERR. Perhaps >> > disconnect and network errors will be ignored. >> > >> NetSocketState can do limited things about these situation, perhaps, >> implement net_socket_can_receive() to export such message to frontend >> ? > > net/socket.c uses G_IO_HUP | G_IO_ERR today, see > iohandler.c:qemu_iohandler_fill(). > > This patch *stops* using them. My question is: why stop and is it > correct? > No, it is not. It leaves dead file descriptors unhandled. Will fix up in next version
> Stefan