[FS#13] Ath9k AP stays up for connected clients but doesn't show in scan on new ones

LEDE Bugs lede-bugs at lists.infradead.org
Thu Dec 29 05:22:53 PST 2016


The following task has a new comment added:

FS#13 - Ath9k AP stays up for connected clients but doesn't show in scan on new ones
User who did this - Johan van Zoomeren (amain)

----------
Hi Felix,

Thanks for the patch, but it does not seem to work. I compiled an image using your staging area, up to commit 2782d36a4053b9879ea84b9e29c347eb4471c639, and the problem is still there.

ath9k debug level:

# echo 0x00000409 > /sys/kernel/debug/ieee80211/phy0/ath9k/debug


iperf:

...
2016-12-29 11:34:49 [ 4] 286.5-287.0 sec 1.41 MBytes 23.7 Mbits/sec
2016-12-29 11:34:50 [ 4] 287.0-287.5 sec 1.21 MBytes 20.3 Mbits/sec
2016-12-29 11:34:50 [ 4] 287.5-288.0 sec 1.19 MBytes 19.9 Mbits/sec
2016-12-29 11:34:51 [ 4] 288.0-288.5 sec 1.34 MBytes 22.4 Mbits/sec
(connection lost)


router log:

..
Dec 29 11:33:22 debwrt kernel: [ 1383.295518] ath: phy0: NF calibrated [ctl] [chain 0] is -89
Dec 29 11:33:22 debwrt kernel: [ 1383.295555] ath: phy0: NF calibrated [ctl] [chain 1] is -88
Dec 29 11:33:22 debwrt kernel: [ 1383.295576] ath: phy0: NF calibrated [ctl] [chain 2] is -87
Dec 29 11:33:52 debwrt kernel: [ 1413.559811] ath: phy0: NF calibrated [ctl] [chain 0] is -88
Dec 29 11:33:52 debwrt kernel: [ 1413.559847] ath: phy0: NF calibrated [ctl] [chain 1] is -88
Dec 29 11:33:52 debwrt kernel: [ 1413.559869] ath: phy0: NF calibrated [ctl] [chain 2] is -87
Dec 29 11:34:22 debwrt kernel: [ 1443.626346] ath: phy0: NF calibrated [ctl] [chain 0] is -88
Dec 29 11:34:22 debwrt kernel: [ 1443.626382] ath: phy0: NF calibrated [ctl] [chain 1] is -89
Dec 29 11:34:22 debwrt kernel: [ 1443.626404] ath: phy0: NF calibrated [ctl] [chain 2] is -88
Dec 29 11:34:52 debwrt kernel: [ 1473.699523] ath: phy0: NF calibrated [ctl] [chain 0] is -89
Dec 29 11:34:52 debwrt kernel: [ 1473.699558] ath: phy0: NF calibrated [ctl] [chain 1] is -89
Dec 29 11:34:52 debwrt kernel: [ 1473.699580] ath: phy0: NF calibrated [ctl] [chain 2] is -89
Dec 29 11:34:52 debwrt kernel: [ 1473.799518] ath: phy0: tx hung, resetting the chip
Dec 29 11:34:52 debwrt kernel: [ 1473.830307] ath: phy0: NF calibrated [ctl] [chain 0] is -89
Dec 29 11:34:52 debwrt kernel: [ 1473.830340] ath: phy0: NF calibrated [ctl] [chain 1] is -89
Dec 29 11:34:52 debwrt kernel: [ 1473.830362] ath: phy0: NF calibrated [ctl] [chain 2] is -89
Dec 29 11:34:52 debwrt kernel: [ 1473.861072] ath: phy0: ah->misc_mode 0xc
Dec 29 11:34:52 debwrt kernel: [ 1473.863588] ath: phy0: enabling IQ Calibration
Dec 29 11:34:52 debwrt kernel: [ 1473.863608] ath: phy0: starting IQ Mismatch Calibration
Dec 29 11:34:52 debwrt kernel: [ 1473.863631] ath: phy0: Setting CFG 0x10a
...


After the connections gets lost, the log starts to show the "resetting the chip" message about every second for a while. See attached a larger syslog.

If there is anything a can do to help to debug or test, please let me know.

A way to get past the compile error already reported by Sven:

$ # revert commit: https://git.lede-project.org/?p=source.git;a=commit;h=2fd5ce9488d11c7e6eee7dc30f128bd12be889f5
$ make tools/libressl/clean
$ make tools/libressl/install
$ export LD_LIBRARY_PATH=$(readlink -f build_dir/host/libressl-2.5.0/ssl/.libs):$(readlink -f build_dir/host/libressl-2.5.0/crypto/.libs)
$ make -j4

----------

More information can be found at the following URL:
https://bugs.lede-project.org/index.php?do=details&task_id=13#comment1118



More information about the lede-bugs mailing list