Bugzilla – Bug 624
duplicate address detected
Last modified: 2005-10-09 10:18:08
You need to log in before you can comment on or make changes to this bug.
This bug has been reported several times, most recently in Bug 508, which claims it is fixed in 1.0.2. However, I'm running 1.0.2 here and I'm getting the same problem. The patches that were submitted on those bugs differ from what is the existing code in is_network_packet(). They seem to have variants on addr1/addr3 and whether or not memcmp(...) is returned, or !memcmp(...) is returned. Which address should be correct? I might get some time to dabble on the weekend to see what works.
Created an attachment (id=330) [details] this patch fixes it for me(tm) Seems that the broadcast/multicast case was passing the relevant packets through. This patch reorders the code, and discards on src address != self.
Merged for 1.0.3 Thanks, James
The patch seems not applied correctly in ipw2200 1.0.3. Line 6326 in ipw2200 1.0.3 should have a "break" sentence.
The patch is changed a little bit to merge into 1.0.3, the return line "return !memcmp(header->addr1, priv->net_dev->dev_addr, ETH_ALEN); " is not deleted, so the "break" is not needed :) Mark the bug as verified.
I always have this problem with drivers : - 1.0.0 - 1.0.1 - 1.0.3 - 1.0.4 (with firmware 2.3, of course) my AP is an SMC2671W (only a 802.11b AP) that I have upgraded to the last firmware. Symptoms : 1) duplicate Packets, packet loss, fluctuating ping times 2) when traffic increase, the signal strengh periodicly decrease and reassociation with AP occurs. Changing ipw2200's rate (in my case to 1M) just reduce a little bit the problem (by reducing the number of "DUP !").
(In reply to comment #5) > Symptoms : > 1) duplicate Packets, packet loss, fluctuating ping times Yours is a different bug with this one. This bug says ipw2200 card receives its own packets. The bug you reported will be fix in the 1.0.5 (coming soon).
Setting back to verified, based on Crystal's comments.