Created attachment 327736 [details] config-3.3.8-r1 (working) With kernel =gentoo-sources-3.4.11, my usb wireless adapter Bus 002 Device 002: ID 0586:341a ZyXEL Communications Corp. NWD-270N Wireless N-lite USB Adapter does not work. I was running 3.2.21 before which was working fine. Wireless is also working with 3.3.8-r1 which I'm running now. I'm planning to test 3.5.7 soon.
Created attachment 327738 [details] config-3.4.11 (not working)
Created attachment 327740 [details] diff -uwdb /usr/src/linux-3.3.8-gentoo-r1/.config /usr/src/linux-3.4.11-gentoo/.config
In which way does it not work? Is the module not present? Do there appear relevant warnings or errors in `dmesg`? Does it load just fine but doesn't work regardless? Have you tried removing the module and loading it again to see if it works then?
Created attachment 327852 [details] config-3.5.7 (not working)
I have the driver rt2x00usb built-in in each case. In 3.5.7's case it seems to associate fine, but it doesn't seem to be able to transfer any data. In 3.3.8's case I get lots of phy0 -> rt2800usb_txdone_entry_check: Warning - TX status report missed for queue 2 entry 32 and phy0 -> rt2800usb_txdone_entry_check: Warning - Data pending for entry 46 in queue 2 but it works regardless.
Created attachment 327908 [details] 3.5.7-dmesg 3.5.7's log which doesn't seem to contain anything relevant
Have been looking further using phy0 -> rt2800usb_txdone_entry_check: Warning - TX status report missed for queue 2 entry 32 where I have done a diff between 3.3 and 3.5 and I see that for instance the if line has changed which is why you might not get the error anymore, as a point of reference I looked in the git blame of that file to see where this came from. The main commit that changed that line is commit 5f8f718ae1e1a6b4e16e67e9e67aff1864419f2e Author: Stanislaw Gruszka <sgruszka@redhat.com> Date: Wed Mar 14 11:16:20 2012 +0100 rt2x00: rt2800usb: do not check packedid for aggregated frames Tx statuses of aggregated subframes contain packetid of first subframe in the AMPDU. We can not identify AMPDU subframes based on packedid, so simply assume that status match first pending frame in the queue. Thats mostly the same what 2800pci do. Signed-off-by: Stanislaw Gruszka <sgruszka@redhat.com> Acked-by: Gertjan van Wingerde <gwingerde@gmail.com> Acked-by: Helmut Schaa <helmut.schaa@googlemail.com> Signed-off-by: John W. Linville <linville@tuxdriver.com> Looking up this commit message I found out that this commit was part of the wireless-next branch which was intended for v3.4, so it's most likely that this or another commit from these is to blame. You can see the other commit messages if you look for Stanislaw Gruszka in http://article.gmane.org/gmane.linux.kernel/1268671/match=Stanislaw+Gruszka and there are also two commits by Helmut Schaa but they might not be relevant. The interesting thing here is that they are about tx and txdone, which appears in your warning messages. Because we know which set of commits that most likely broke this for you and this set is very small, it might be feasible to do a git bisect to find the offending commit. As bad, use 0d9be8a4b7da71ef3b4ac8f6aa4fa225c1cb8e98 which is the last commit of his networking fixes done to rt2x00. As good, use d47a61aa228709fe1704e18a2f444661c10b81c0 which is the first commit before his networking fixes done to rt2x00. These aren't so many commits, so the bisect procedure should probably not take very long; about two or three reboots. Please read below link on how to perform this. http://wiki.gentoo.org/wiki/Kernel_git-bisect
Please provide bisect information as requested. You might want to test the latest kernel which is 3.7.1 as of this writing. Please comment here if you still have issues and we will reopen.
Hi, I'm running 3.7.9 now and it is the first version that is working well for me again. As this is my work machine (and I don't have another test machine) I have to be a little conservative, so I like to keep to kernel versions with high micro version and don't want to do any kernel bisects. Sorry for the trouble.