From: Simon McVittie <simon.mcvittie@collabora.co.uk>
Date: Wed, 12 Jan 2011 12:14:49 +0000 (+0000)
Subject: _dbus_loop_iterate: if the kernel says a fd is bad, stop watching it
X-Git-Tag: dbus-1.4.4~23
X-Git-Url: http://git.ipfire.org/cgi-bin/gitweb.cgi?a=commitdiff_plain;h=460e6282d68a928a78f16031ed1637bc7676e10e;p=thirdparty%2Fdbus.git

_dbus_loop_iterate: if the kernel says a fd is bad, stop watching it

Again, this shouldn't happen - modules are responsible for cleaning up
their watches - but the failure mode here is really bad: if we leave an
invalid fd in the set, every poll() call will instantly return, marking
it as POLLNVAL. The result is that dbus-daemon busy-loops on poll()
without responding to I/O, so the bad watch will probably never be
cleared up.

Bug: https://bugs.freedesktop.org/show_bug.cgi?id=32992
Bug-NB: NB#200248
Reviewed-by: Colin Walters <walters@verbum.org>
---

diff --git a/dbus/dbus-mainloop.c b/dbus/dbus-mainloop.c
index c37513829..43159a70b 100644
--- a/dbus/dbus-mainloop.c
+++ b/dbus/dbus-mainloop.c
@@ -829,6 +829,15 @@ _dbus_loop_iterate (DBusLoop     *loop,
                   
                   retval = TRUE;
                 }
+
+              if (_DBUS_UNLIKELY (fds[i].revents & _DBUS_POLLNVAL))
+                {
+                  _dbus_warn ("invalid request, socket fd %d not open\n",
+                      fds[i].fd);
+                  _dbus_watch_invalidate (wcb->watch);
+                  _dbus_loop_remove_watch (loop, wcb->watch, wcb->function,
+                      ((Callback *)wcb)->data);
+                }
             }
               
           ++i;