Description of problem: kernel oops when wireless driver is passing traffic Version-Release number of selected component (if applicable): 2.6.20-1.3066.fc7 How reproducible: always Steps to Reproduce: 1. assosciate with ap and start a ping, seems to happen after 10 seconds 2. 3. Actual results: Expected results: Additional info: I'm not hundred percent sure about how, where and what here - since the logs seems to fill up with plenty debug messages. I'm happy to supply more details. BUG: warning at include/asm/dma-mapping.h:25/dma_map_single() (Not tainted) [<c04061e9>] show_trace_log_lvl+0x1a/0x2f [<c04067ad>] show_trace+0x12/0x14 [<c0406831>] dump_stack+0x16/0x18 [<f8b1e847>] d_tx+0x7f2/0xd69 [iwlwifi] [<f8a8f6c6>] __ieee80211_tx+0x20/0xec [mac80211] [<f8a91588>] ieee80211_master_start_xmit+0x410/0x50e [mac80211] [<c05b4352>] dev_hard_start_xmit+0x1ce/0x22b [<c05c1e2a>] __qdisc_run+0xdb/0x195 [<c05b5085>] net_tx_action+0xc2/0xeb [<c042bfa7>] __do_softirq+0x6f/0xe2 [<c0407420>] do_softirq+0x61/0xd0 [<c042bdd2>] local_bh_enable_ip+0xb9/0xdf [<c06161a5>] _spin_unlock_bh+0x25/0x28 [<f8a9dc86>] ieee80211_scan_completed+0xdf/0x109 [mac80211] [<f8a9f0d7>] ieee80211_sta_scan_work+0x63/0x199 [mac80211] [<c043510b>] run_workqueue+0x89/0x14e [<c0435b21>] worker_thread+0xf8/0x124 [<c043830f>] kthread+0xb3/0xdc [<c0405cd3>] kernel_thread_helper+0x7/0x10 =======================
Created attachment 152621 [details] Snippet from /var/log/messages with BUG Get this too, 25 seconds after association. Wireless (even unencrypted) not functioning.
Exact the same situation with 2.6.20-1.3069.fc7
*** Bug 237182 has been marked as a duplicate of this bug. ***
I don't get an oops, but the association does get dropped a few seconds after it initially succeeds, with plenty of log output like this (repeated many times): iwlwifi: Error sending RXON: time out after 500ms. iwlwifi: Error clearing ASSOC_MSK on current configuration (-110). iwlwifi: No space for Tx iwlwifi: Error sending RXON: ipw_queue_tx_hcmd failed: -28 iwlwifi: Error clearing ASSOC_MSK on current configuration (-28). I was running 3094 when this happened. 3056 is still the only kernel that works correctly for me with iwlwifi.
I have the same symptoms - no Oops but lots of those messages and the connection fails very soon after the initial success.
Bryan and Adam, I think you should be using bug 236828 instead.
Just confirming the bug still exists with 2.6.20-1.3104.fc7
I'm still seeing the same BUG warning in 2.6.21-1.3116.fc7.
I'm getting the same sort of problem in F7t4. I did a full install from the F7t4 Live CD on a ThinkPad T60p with the ipw3945 wireless card.
The simple trigger for this is to do a scan while you're associated.
Created attachment 153680 [details] 2.6.21-1.3117.fc7 kernel message log This is the result under today's 3117 kernel. I know just about zip about kernel debugging, but now there's a second BUG happening (see near end of log). I don't want to hijack this bug, so if that's unrelated, point me to the right component; however, the kernel's the only package changed on this machine between the last attempt and this one.
FWIW, the rtnl_setlink appears to be a different issue (but still in the kernel) and should be tracked separately.
some of this issues should be resolved in 0.0.15, see http://sourceforge.net/mailarchive/forum.php?thread_name=46362924.8040602%40linux.intel.com&forum_name=ipw3945-devel
John, I saw that you updated iwlwifi in cvs to 0.0.15 but 0.0.15 has another problem: no frames being sent when scanning while acciotated but this is fixed in the 0.0.16 release ;)
This problem looks to be fixed under the 3122 kernel with the new iwlwifi-0.0.16.
FYI, the newer kernels are at: http://people.redhat.com/davej/kernels/Fedora/fc7/kernel.repo