Because of problems with ath9k MiniPci card (see other thread) I tried to use a TP-Link TL-WN821N USB stick as a substitute radio for a high data rate mesh network. I am using compat-wireless-3.5-rc3-1-sn with only a small patch to enable ath9k_htc to switch to mesh point mode.
When transferring data (iperf) to another node these traces show up on console from time to time, most times in a bunch: [ 1844.137863] BUG: scheduling while atomic: kworker/u:2/840/0x00000200 [ 1844.141259] Modules linked in: ath9k_htc ipt_MASQUERADE iptable_nat nf_nat nf_conntrack_ipv4 nf_conntrack nf_defrag_ipv4 ip_tables x_tables ipv6 arc4 ath9k mac80211 cfg80211 ath9k_common ath9k_hw ftdi_sio rtc_cmos ath usbserial rtc_core evdev rtc_lib compat ext4 jbd2 crc16 ohci_hcd ehci_hcd r6040 usbcore libphy [last unloaded: scsi_wait_scan] [ 1844.173269] Pid: 840, comm: kworker/u:2 Not tainted 2.6.37-vortex86-sg #1 [ 1844.176849] Call Trace: [ 1844.179372] [<c1288982>] ? schedule+0x372/0x430 [ 1844.180804] [<d0c58985>] ? ath9k_hw_gettsf64+0x35/0x60 [ath9k_hw] [ 1844.183763] [<c1289428>] ? mutex_lock+0x8/0x20 [ 1844.188447] [<d08b82b8>] ? usb_hcd_submit_urb+0xa8/0x9c0 [usbcore] [ 1844.191491] [<c1288d89>] ? schedule_timeout+0xd9/0x1f0 [ 1844.196859] [<c102f5c0>] ? process_timeout+0x0/0x10 [ 1844.201924] [<c1288b9e>] ? wait_for_common+0x8e/0xf0 [ 1844.203748] [<c1020690>] ? default_wake_function+0x0/0x10 [ 1844.209429] [<d0feda4d>] ? ath9k_wmi_cmd+0x15d/0x200 [ath9k_htc] [ 1844.212321] [<d0ff419d>] ? ath9k_regread+0x2d/0x70 [ath9k_htc] [ 1844.215039] [<d0c58962>] ? ath9k_hw_gettsf64+0x12/0x60 [ath9k_hw] [ 1844.218005] [<d0ff0c68>] ? ath9k_htc_get_tsf+0x28/0x60 [ath9k_htc] [ 1844.221148] [<d0ddafe7>] ? ieee80211_iface_work+0xb7/0x330 [mac80211] [ 1844.224467] [<c10ad4ca>] ? kmem_cache_free+0x5a/0x90 [ 1844.229716] [<d0e094fd>] ? mesh_sync_adjust_tbtt+0xbd/0x2a0 [mac80211] [ 1844.233126] [<c10388b4>] ? process_one_work+0x104/0x390 [ 1844.235263] [<d0ddaf30>] ? ieee80211_iface_work+0x0/0x330 [mac80211] [ 1844.238489] [<c1038e07>] ? worker_thread+0xe7/0x3a0 [ 1844.240222] [<c1038d20>] ? worker_thread+0x0/0x3a0 [ 1844.245209] [<c103bdb4>] ? kthread+0x74/0x80 [ 1844.249667] [<c103bd40>] ? kthread+0x0/0x80 [ 1844.250701] [<c1002f56>] ? kernel_thread_helper+0x6/0x10 Any idea on what's causing this? Also, like in the other thread, I observed crashes instead of bug traces when using older compat-wireless builds. Although this seems to be an improvement, the overall stability of these Atheros chipsets or their linux drivers are not very reassuring... _______________________________________________ ath9k-devel mailing list ath9k-devel@lists.ath9k.org https://lists.ath9k.org/mailman/listinfo/ath9k-devel