Hide Forgot
Description of problem: When an ipv6 link-local address is added (either automatically or manually) to a prism54pci (mac80211) device, the address will always be marked "tentative", and a "duplicate address detected!" message will be output in syslog. This is definitely not a duplicate address. It looks like the device is seeing its own broadcast packets, that the kernel sent out to perform ipv6 Neighbor Discovery. This may be similar to this bug: http://bughost.org/bugzilla/show_bug.cgi?id=265 Version-Release number of selected component (if applicable): kernel-2.6.21-1.3228.fc7 How reproducible: always Steps to Reproduce: 1. flush ipv6 addrs from the device (for consistency) # ip -6 addr flush wlan1 2. manually add a unique ipv6 addr to the device # ip -6 addr add fe80::209:5bff:feea:c06f/64 dev wlan1 scope link 3. look in syslog for the "duplicate address detected!" message 4. see that the address has been marked tentative # ip -6 addr show wlan1 7: wlan1: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qlen 1000 inet6 fe80::209:5bff:feea:c06f/64 scope link tentative valid_lft forever preferred_lft forever Actual results: "duplicate address detected!" in syslog, address is marked tentative and is unusable Expected results: no duplicate address detected, address not marked tentative Additional info: attached a wireshark capture of the ipv6 broadcast packets that the device sees when the address is added
Created attachment 158274 [details] ipv6 broadcast packets seen by the prism54pci device when adding a link-local ipv6 address
FWIW, I think this is a generic mac80211 problem.
*** Bug 250252 has been marked as a duplicate of this bug. ***
Please try the kernels here: http://koji.fedoraproject.org/koji/buildinfo?buildID=13002 Do these avoid the "duplicate address" problem?
Kernel 46 from koji seems to fix the problem on zd1211rw and IPv6 now works fine (it appears as if IPv4 DHCP setup is much faster as well, but I didn't test it thoroughly). Thanks!
2.6.22.1-50 no longer exhibits the "duplicate address" problem. Awesome, thanks John!