Created attachment 136301 [details] Trace that is flooding /var/log/messages and /var/log/journal I have noticed this bug after all my free space dissappeared from / . It is happening with all 3.14 kernels and probably all up to 3.15.0-0.rc4 that is the last one that I have checked. 3.13 is working fine. lsusb: Bus 004 Device 002: ID 0cf3:1006 Atheros Communications, Inc. TP-Link TL-WN322G v3 / TL-WN422G v2 802.11g [Atheros AR9271] uname -r: kernel 3.14.2-200.fc20.i686+PAE kernel 3.15.0-0.rc4.git0.1.fc21.i686+PAE When I insert wifi usb dongle into usb port it is not working and generate trace. But, that minor problem actually creates trouble - pattern that you can see in the attachment is flooding logs with such a rate that according to timestamps written in /var/log/messages ,in less that an hour log file grew from zero to 1.4GB! Probably it stoped when it have eaten all free space on /. Also it is flooding this location too: [root@localhost slobodan]# du -hs /var/log/journal/542c47fae5eb4d1d873a1b41ef83927e/ 2.4G /var/log/journal/542c47fae5eb4d1d873a1b41ef83927e/ I can reproduce this at any time with mentioned kernels.
Have you tried a git bisect? Are you able to do that?
TP-Link WN422G v2, same behaviour with freshly compiled kernel 3.15.1. The dongle works perfectly with Slackware's kernel 3.10.7.
I seem to have the same problem with my wifi dongle: ID 0846:9030 NetGear, Inc. WNA1100 Wireless-N 150 [Atheros AR9271] However the problem only happens if i connect the wifi dongle to my external USB-Hub. If i connect it directly to the computer the problem does not happen. I did a bisect but sadly only this commit came up: commit 3482528e9aced9234d4e2a4a9538c882a9aa5aa2 Author: Greg Kroah-Hartman <gregkh@linuxfoundation.org> Date: Thu Dec 19 15:42:20 2013 -0800 USB: core: remove CONFIG_USB_DEBUG usage
I did another bisect using CONFIG_USB_DEBUG=y and this is the commit which started the problem for me: Author: Oleksij Rempel <linux@rempel-privat.de> Date: Tue Aug 13 21:10:19 2013 +0200 ath9k_htc: do not use bulk on EP3 and EP4
commit 2b721118b7821107757eb1d37af4b60e877b27e7
I just reverted the commit on my 3.14-19, tested it with a Netgear WNA1100 and it works like a charm.
This bug is marked as NEEDINFO. The commit that introudces the bug has been found (Comment #5). What other info is needed?
Just lost in the shuffle... Any ath9k guys want to comment before we revert the commit in question? Given how old this bug is, we also should ask if the problem still happens?
I tested most recently on v3.18-rc6 and yes, the problem still happens.