On Thu, Mar 08, 2018 at 05:50:29AM -0500, Frediano Ziglio wrote:
> >
> > Currently if we fail to set up the watch waiting for accept() to be
> > called on the socket, we still keep the network socket(s) open even if we
> > are not going to be able to use it. This commit makes sure it's closed a
> > set to -1 when such a failure occurs rather than having a half
> > initialized spice-server instance.
> > ---
> > server/reds.c | 47 +++++++++++++++++++++++++++++------------------
> > 1 file changed, 29 insertions(+), 18 deletions(-)
> >
> > diff --git a/server/reds.c b/server/reds.c
> > index a31ed4e96..829158ca3 100644
> > --- a/server/reds.c
> > +++ b/server/reds.c
> > @@ -2584,19 +2584,36 @@ void reds_set_client_mm_time_latency(RedsState *reds,
> > RedClient *client, uint32_
> > }
> > }
> >
> > +static void reds_cleanup_net(SpiceServer *reds)
> > +{
> > + if (reds->listen_socket != -1) {
> > + reds_core_watch_remove(reds, reds->listen_watch);
> > + if (reds->config->spice_listen_socket_fd != reds->listen_socket) {
> > + close(reds->listen_socket);
> > + }
> > + reds->listen_watch = NULL;
> > + reds->listen_socket = -1;
> > + }
> > + if (reds->secure_listen_socket != -1) {
> > + reds_core_watch_remove(reds, reds->secure_listen_watch);
> > + close(reds->secure_listen_socket);
> > + reds->secure_listen_watch = NULL;
> > + reds->secure_listen_socket = -1;
> > + }
> > +}
> > +
> > static int reds_init_net(RedsState *reds)
> > {
> > if (reds->config->spice_port != -1 || reds->config->spice_family ==
> > AF_UNIX) {
> > reds->listen_socket = reds_init_socket(reds->config->spice_addr,
> > reds->config->spice_port, reds->config->spice_family);
> > if (-1 == reds->listen_socket) {
> > - return -1;
> > + goto error;
> > }
> > reds->listen_watch = reds_core_watch_add(reds, reds->listen_socket,
> > SPICE_WATCH_EVENT_READ,
> > reds_accept, reds);
> > if (reds->listen_watch == NULL) {
> > - spice_warning("set fd handle failed");
> > - return -1;
> > + goto error;
> > }
> > }
> >
> > @@ -2604,14 +2621,13 @@ static int reds_init_net(RedsState *reds)
> > reds->secure_listen_socket =
> > reds_init_socket(reds->config->spice_addr,
> > reds->config->spice_secure_port,
> > reds->config->spice_family);
> > if (-1 == reds->secure_listen_socket) {
> > - return -1;
> > + goto error;
> > }
> > reds->secure_listen_watch = reds_core_watch_add(reds,
> > reds->secure_listen_socket,
> > SPICE_WATCH_EVENT_READ,
> > reds_accept_ssl_connection,
> > reds);
> > if (reds->secure_listen_watch == NULL) {
> > - spice_warning("set fd handle failed");
> > - return -1;
> > + goto error;
> > }
> > }
> >
> > @@ -2621,11 +2637,15 @@ static int reds_init_net(RedsState *reds)
> > SPICE_WATCH_EVENT_READ,
> > reds_accept, reds);
> > if (reds->listen_watch == NULL) {
> > - spice_warning("set fd handle failed");
> > - return -1;
> > + reds->listen_socket = -1;
>
> Is this necessary taking into account the check in reds_cleanup_net
>
> > + goto error;
> > }
> > }
> > return 0;
> > +
> > +error:
> > + reds_cleanup_net(reds);
>
> Maybe this call can be in the err label inside do_spice_init?
> This would reduce reds_init_net changes and improve do_spice_init too.
Good idea, I sent a patch doing that, though on second though, I would
have kept that error label doing the cleanup in reds_init_net, and added
a call to reds_cleanup_net(reds) in the err: label in do_spice_init().
Reason for that is that when do_foo() fails, I prefer if it cleans up
whatever it was trying to do. With v3, the caller has to call a method
to do the cleanup when this fails. Not a big deal, I don't think I'm
going to send a v4 just for that if people are fine with v3.
Christophe
Attachment:
signature.asc
Description: PGP signature
_______________________________________________ Spice-devel mailing list Spice-devel@xxxxxxxxxxxxxxxxxxxxx https://lists.freedesktop.org/mailman/listinfo/spice-devel