site_archiver@lists.apple.com Delivered-To: darwin-dev@lists.apple.com Hi Ben, Thanks for the reply. Thanks, - Terry On Mar 28, 2007, at 2:01 PM, Ben Byer wrote: Ben On Mar 27, 2007, at 8:29 PM, Terry Simons wrote: Sweet. That was the problem. Can someone comment on whether this is expected behavior? ;) - Terry P.S. Thanks for the replies Matt. ;) On Mar 27, 2007, at 9:03 PM, Ben Byer wrote: On Mar 27, 2007, at 4:48 PM, Terry Simons wrote: Has anyone seen this before? Try applying this patch: http://cvs.tcpdump.org/cgi-bin/cvsweb/libpcap/inet.c?r1=1.72&r2=1.73 (explanation is provided in a comment inside the patch.) Ben _______________________________________________ Do not post admin requests to the list. They will be ignored. Darwin-dev mailing list (Darwin-dev@lists.apple.com) Help/Unsubscribe/Update your Subscription: http://lists.apple.com/mailman/options/darwin-dev/galimore%40mac.com _______________________________________________ Do not post admin requests to the list. They will be ignored. Darwin-dev mailing list (Darwin-dev@lists.apple.com) Help/Unsubscribe/Update your Subscription: http://lists.apple.com/mailman/options/darwin-dev/site_archiver%40lists.appl... The problem is going to affect any networking application that tries to enumerate the interfaces on the system and unknowingly opens the wlt interface. My application, for instance, doesn't use pcap at all. I am doing raw enumeration and socket handling. I suspect that Wireshark is doing some of its own interface handling as well... which might explain why it has the issue. I don't think it's a very big deal... The symptoms are obvious, and the fix is known so I'm a happy developer. ;) I haven't looked at the Atheros driver code, but my understanding is that this is an "unintended consequence" of the way the Atheros driver / BPF interface is implemented; I'm not aware of any immediate plans to change that code, so it's probably a safer bet to change the libpcap code, unless we find a lot of other bits of code that are broken by this ... quirk. Interesting that this only affects the Atheros systems. I had a friend try on a Broadcom based Intel Mac and it didn't exhibit the same issue, although he does have a "wlt" interface. Must be something specific to the Atheros driver's implementation of the monitor mode bits. My application, which appears to work fine on PowerPC Macs, causes the wireless card on Intel macs (MacBook Pro for sure) to lose association. I'm not sure what's causing this... but it definitely looks like a bug. I'm using BPF and calling network IOCTLs... so I'm pretty sure it's something in there, but I haven't been able to locate the exact issue yet. Any lurkers privy to the inner workings of the Atheros Airport driver? This email sent to galimore@mac.com This email sent to site_archiver@lists.apple.com