On Sat, 2007-06-02 at 17:15 +0000, Volker Braun wrote: > The original wpa_supplicant-0.5.7-use-syslog.patch is horrible, that is a > wifi-remote exploit waiting to happen. Great that you caught that! > Somebody needs to rewrite this atrocity without sprinkling sprintf > around... I think you should file an urgent RH bugzilla asap.
That said; if anyone wants to take a shot at adding real syslog support upstream, that would rock. The problem is that we don't want NM to capture the wpa_supplicant output like it used to, because that's icky. Since wpa_supplicant is a daemon, it should be writing to syslog and have configurable debug levels (as should NetworkManager). Upstream doesn't seem to have a lot of willpower to do this, but it greatly helps debugging problems with NetworkManager. Dan > Volker > > > On Sat, 02 Jun 2007 10:07:15 +0100, Jon Escombe wrote: > > This is caused by the redhat wpa_supplicant-0.5.7-use-syslog.patch > > overflowing a 2KBoutput buffer. The frame that's causing the error for > > me is 1400 bytes. When displayed in a "RX EAPOL - hexdump(len=1400): 00 > > 00 00" string, this requires a little over 4KB. This patch in this mail > > just increases the buffer size to 8KB, however I suspect there ought to > > be some boundary checking going on. > > Dan, is reporting here sufficient, or would a bugzilla entry help get > > this fixed? > > _______________________________________________ > NetworkManager-list mailing list > [email protected] > http://mail.gnome.org/mailman/listinfo/networkmanager-list _______________________________________________ NetworkManager-list mailing list [email protected] http://mail.gnome.org/mailman/listinfo/networkmanager-list
