Problem with Atheros adapter (TP-LINK TL-WDN4800, AR9380)

2015-02-04 Thread Willy Offermans
Dear FreeBSD friends,

Excuse me, I just became member of this mailling list, so I could not
respond to the correct mail thread.

About two months ago (December 14 and 15, there were some mails posted 
concerning trouble with the TP-LINK TL-WDN4800 PCI Express adapter. 
David and Adrian had a conversation about it. Adrian suggested David to 
switch to HEAD in order to fix some connection problems with some iMAC 
devices and the TP-LINK.

I wonder if this has fixed the problem with the TP-LINK TL-WDN4800 PCI
Express card. 

I'm just about to order the same card and I want to use it in
the same way David did. Of course I don't want to end up with unexpected
disconnections.

-- 
Met vriendelijke groeten,
With kind regards,
Mit freundlichen Gruessen,
De jrus wah,

Wiel

*
 W.K. Offermans

   Powered by 

(__)
 \\\'',)
   \/  \ ^
   .\._/_)

   www.FreeBSD.org
___
freebsd-wireless@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-wireless
To unsubscribe, send any mail to freebsd-wireless-unsubscr...@freebsd.org


Re: Problem with Atheros adapter (TP-LINK TL-WDN4800, AR9380)

2014-12-15 Thread David Blewett
I haven't hit a kernel panic, but several times I get to the point where
the wifi connection from Macbook no longer functions. The icon says it's
connected, but no network traffic passes. I see this in dmesg on the
FreeBSD box:

ath0: ath_tx_default_comp: bf 0xfe004e7a3b98: seqno 1873: bf_next not
NULL!
ath0: ath_tx_default_comp: bf 0xfe004e78aec0: seqno 1874: bf_next not
NULL!
ath0: ath_tx_default_comp: bf 0xfe004e7787e8: seqno 1875: bf_next not
NULL!
ath0: ath_tx_default_comp: bf 0xfe004e795fa8: seqno 1876: bf_next not
NULL!
ath0: ath_tx_default_comp: bf 0xfe004e786f00: seqno 1877: bf_next not
NULL!
ath0: ath_tx_default_comp: bf 0xfe004e791190: seqno 1878: bf_next not
NULL!
ath0: ath_tx_default_comp: bf 0xfe004e787560: seqno 1879: bf_next not
NULL!
ath0: ath_tx_default_comp: bf 0xfe004e7a9b38: seqno 1880: bf_next not
NULL!
ath0: ath_tx_default_comp: bf 0xfe004e77f910: seqno 1881: bf_next not
NULL!
ath0: ath_tx_default_comp: bf 0xfe004e79e720: seqno 1882: bf_next not
NULL!
ath0: ath_tx_default_comp: bf 0xfe004e777990: seqno 1883: bf_next not
NULL!
ath0: ath_tx_default_comp: bf 0xfe004e7a8e78: seqno 1884: bf_next not
NULL!
ath0: ath_tx_default_comp: bf 0xfe004e792180: seqno 1885: bf_next not
NULL!
ath0: ath_tx_default_comp: bf 0xfe004e7a2878: seqno 1886: bf_next not
NULL!
ath0: ath_tx_default_comp: bf 0xfe004e78f9a8: seqno 1887: bf_next not
NULL!
ath0: ath_tx_default_comp: bf 0xfe004e7a3208: seqno 1888: bf_next not
NULL!
ath0: ath_tx_default_comp: bf 0xfe004e7893a8: seqno 1889: bf_next not
NULL!
ath0: ath_tx_default_comp: bf 0xfe004e7a2218: seqno 1890: bf_next not
NULL!
ath0: ath_tx_default_comp: bf 0xfe004e782f40: seqno 1891: bf_next not
NULL!
ath0: ath_tx_default_comp: bf 0xfe004e77b620: seqno 1892: bf_next not
NULL!
ath0: ath_tx_default_comp: bf 0xfe004e78d368: seqno 1893: bf_next not
NULL!
ath0: ath_tx_default_comp: bf 0xfe004e783738: seqno 1894: bf_next not
NULL!
ath0: ath_tx_default_comp: bf 0xfe004e782748: seqno 1895: bf_next not
NULL!
ath0: ath_tx_default_comp: bf 0xfe004e7a5518: seqno 1896: bf_next not
NULL!
ath0: ath_tx_default_comp: bf 0xfe004e7a79c0: seqno 1897: bf_next not
NULL!
ath0: ath_tx_default_comp: bf 0xfe004e786d68: seqno 1898: bf_next not
NULL!
ath0: ath_tx_default_comp: bf 0xfe004e77d2d0: seqno 1899: bf_next not
NULL!
ath0: ath_tx_default_comp: bf 0xfe004e78b1f0: seqno 1900: bf_next not
NULL!
ath0: ath_tx_default_comp: bf 0xfe004e7a1090: seqno 1901: bf_next not
NULL!
ath0: ath_tx_default_comp: bf 0xfe004e77ac90: seqno 1902: bf_next not
NULL!
ath0: ath_tx_default_comp: bf 0xfe004e77c7a8: seqno 1903: bf_next not
NULL!
ath0: ath_tx_default_comp: bf 0xfe004e786240: seqno 1905: bf_next not
NULL!
ath0: ath_tx_default_comp: bf 0xfe004e79f578: seqno 1906: bf_next not
NULL!
wlan0: [28:cf:e9:12:ed:bb] pwr save q overflow, drops 179 (size 50)
wlan0: [28:cf:e9:12:ed:bb] pwr save q overflow, drops 180 (size 50)
wlan0: [28:cf:e9:12:ed:bb] pwr save q overflow, drops 181 (size 50)
wlan0: [28:cf:e9:12:ed:bb] pwr save q overflow, drops 182 (size 50)
wlan0: [28:cf:e9:12:ed:bb] pwr save q overflow, drops 183 (size 50)
wlan0: [28:cf:e9:12:ed:bb] pwr save q overflow, drops 184 (size 50)
wlan0: [28:cf:e9:12:ed:bb] pwr save q overflow, drops 185 (size 50)


Thanks,

David Blewett

On Sun, Dec 14, 2014 at 9:29 PM, David Blewett da...@dawninglight.net
wrote:

 ​Okay, I've built a custom kernel with if_ath_pci as a module, rebooted
 into the new kernel and got it running. I also enabled AH_DEBUG. Will let
 you know how things go this week.​

 Thanks,

 David Blewett

 On Sun, Dec 14, 2014 at 2:58 PM, David Blewett da...@dawninglight.net
 wrote:

 Okay, I can try to give that a go. Will take me a bit. I use ZFS on
 encrypted geli devices, and haven't tried building a custom kernel yet with
 this config.

 Thanks,

 David Blewett

 On Sun, Dec 14, 2014 at 2:55 PM, Adrian Chadd adr...@freebsd.org wrote:

 Aha! Listen time is 0. Hm!


 Just add this at line 1179 and recompile:


 if (ani_state-listen_time == 0)
 return;

 it's possible that it'll happen; it shoud be bailing out sooner rather
 than later.



 -adrian


 On 14 December 2014 at 11:53, David Blewett da...@dawninglight.net
 wrote:
  If this would be easier on IRC, I'm BlueAidan on Freenode.
 
  Output:
 
  (kgdb) list *0x8045161d
  0x8045161d is in ar9300_ani_ar_poll
  (/usr/src/sys/contrib/dev/ath/ath_hal/ar9300/ar9300_ani.c:1180).
  1175 */
  1176if (!DO_ANI(ah)) {
  1177return;
  1178}
  1179
  1180ofdm_phy_err_rate =
  1181ani_state-ofdm_phy_err_count * 1000 /
  ani_state-listen_time;
  1182cck_phy_err_rate =
  1183ani_state-cck_phy_err_count * 1000 /
  ani_state-listen_time;
  1184
 
  Thanks,
 
  David Blewett
 
  On Sun, Dec 14, 2014 at 2:39 PM, Adrian Chadd adr...@freebsd.org
 wrote:
 
  Hi!
 
  Hm, an integer divide by 

Re: Problem with Atheros adapter (TP-LINK TL-WDN4800, AR9380)

2014-12-15 Thread Adrian Chadd
Hi,

The ath_tx_default_comp stuff was fixed in -HEAD with a reasonably big
overhaul of the 11n aggregation code. It fixed a whole lot of bugs. I
don't think there's going to be a cheap solution to that besides
running -HEAD or backporting. I was kinda hoping another committer
would step up and backport the net80211 fixes from -HEAD to stable/10
once they're known to work out okay. I'm just too busy and
time-limited to look after both -HEAD and stable branches. :(

The second is a bit odd - mostly that indicates the STA has gone away
and we just haven't timed it out yet. (It'll tell the AP it's going to
sleep so it can scan, and it may just never decide to come back ..)



-adrian


On 15 December 2014 at 14:43, David Blewett da...@dawninglight.net wrote:
 I haven't hit a kernel panic, but several times I get to the point where the
 wifi connection from Macbook no longer functions. The icon says it's
 connected, but no network traffic passes. I see this in dmesg on the FreeBSD
 box:

 ath0: ath_tx_default_comp: bf 0xfe004e7a3b98: seqno 1873: bf_next not
 NULL!
 ath0: ath_tx_default_comp: bf 0xfe004e78aec0: seqno 1874: bf_next not
 NULL!
 ath0: ath_tx_default_comp: bf 0xfe004e7787e8: seqno 1875: bf_next not
 NULL!
 ath0: ath_tx_default_comp: bf 0xfe004e795fa8: seqno 1876: bf_next not
 NULL!
 ath0: ath_tx_default_comp: bf 0xfe004e786f00: seqno 1877: bf_next not
 NULL!
 ath0: ath_tx_default_comp: bf 0xfe004e791190: seqno 1878: bf_next not
 NULL!
 ath0: ath_tx_default_comp: bf 0xfe004e787560: seqno 1879: bf_next not
 NULL!
 ath0: ath_tx_default_comp: bf 0xfe004e7a9b38: seqno 1880: bf_next not
 NULL!
 ath0: ath_tx_default_comp: bf 0xfe004e77f910: seqno 1881: bf_next not
 NULL!
 ath0: ath_tx_default_comp: bf 0xfe004e79e720: seqno 1882: bf_next not
 NULL!
 ath0: ath_tx_default_comp: bf 0xfe004e777990: seqno 1883: bf_next not
 NULL!
 ath0: ath_tx_default_comp: bf 0xfe004e7a8e78: seqno 1884: bf_next not
 NULL!
 ath0: ath_tx_default_comp: bf 0xfe004e792180: seqno 1885: bf_next not
 NULL!
 ath0: ath_tx_default_comp: bf 0xfe004e7a2878: seqno 1886: bf_next not
 NULL!
 ath0: ath_tx_default_comp: bf 0xfe004e78f9a8: seqno 1887: bf_next not
 NULL!
 ath0: ath_tx_default_comp: bf 0xfe004e7a3208: seqno 1888: bf_next not
 NULL!
 ath0: ath_tx_default_comp: bf 0xfe004e7893a8: seqno 1889: bf_next not
 NULL!
 ath0: ath_tx_default_comp: bf 0xfe004e7a2218: seqno 1890: bf_next not
 NULL!
 ath0: ath_tx_default_comp: bf 0xfe004e782f40: seqno 1891: bf_next not
 NULL!
 ath0: ath_tx_default_comp: bf 0xfe004e77b620: seqno 1892: bf_next not
 NULL!
 ath0: ath_tx_default_comp: bf 0xfe004e78d368: seqno 1893: bf_next not
 NULL!
 ath0: ath_tx_default_comp: bf 0xfe004e783738: seqno 1894: bf_next not
 NULL!
 ath0: ath_tx_default_comp: bf 0xfe004e782748: seqno 1895: bf_next not
 NULL!
 ath0: ath_tx_default_comp: bf 0xfe004e7a5518: seqno 1896: bf_next not
 NULL!
 ath0: ath_tx_default_comp: bf 0xfe004e7a79c0: seqno 1897: bf_next not
 NULL!
 ath0: ath_tx_default_comp: bf 0xfe004e786d68: seqno 1898: bf_next not
 NULL!
 ath0: ath_tx_default_comp: bf 0xfe004e77d2d0: seqno 1899: bf_next not
 NULL!
 ath0: ath_tx_default_comp: bf 0xfe004e78b1f0: seqno 1900: bf_next not
 NULL!
 ath0: ath_tx_default_comp: bf 0xfe004e7a1090: seqno 1901: bf_next not
 NULL!
 ath0: ath_tx_default_comp: bf 0xfe004e77ac90: seqno 1902: bf_next not
 NULL!
 ath0: ath_tx_default_comp: bf 0xfe004e77c7a8: seqno 1903: bf_next not
 NULL!
 ath0: ath_tx_default_comp: bf 0xfe004e786240: seqno 1905: bf_next not
 NULL!
 ath0: ath_tx_default_comp: bf 0xfe004e79f578: seqno 1906: bf_next not
 NULL!
 wlan0: [28:cf:e9:12:ed:bb] pwr save q overflow, drops 179 (size 50)
 wlan0: [28:cf:e9:12:ed:bb] pwr save q overflow, drops 180 (size 50)
 wlan0: [28:cf:e9:12:ed:bb] pwr save q overflow, drops 181 (size 50)
 wlan0: [28:cf:e9:12:ed:bb] pwr save q overflow, drops 182 (size 50)
 wlan0: [28:cf:e9:12:ed:bb] pwr save q overflow, drops 183 (size 50)
 wlan0: [28:cf:e9:12:ed:bb] pwr save q overflow, drops 184 (size 50)
 wlan0: [28:cf:e9:12:ed:bb] pwr save q overflow, drops 185 (size 50)


 Thanks,

 David Blewett

 On Sun, Dec 14, 2014 at 9:29 PM, David Blewett da...@dawninglight.net
 wrote:

 Okay, I've built a custom kernel with if_ath_pci as a module, rebooted
 into the new kernel and got it running. I also enabled AH_DEBUG. Will let
 you know how things go this week.

 Thanks,

 David Blewett

 On Sun, Dec 14, 2014 at 2:58 PM, David Blewett da...@dawninglight.net
 wrote:

 Okay, I can try to give that a go. Will take me a bit. I use ZFS on
 encrypted geli devices, and haven't tried building a custom kernel yet with
 this config.

 Thanks,

 David Blewett

 On Sun, Dec 14, 2014 at 2:55 PM, Adrian Chadd adr...@freebsd.org wrote:

 Aha! Listen time is 0. Hm!


 Just add this at line 1179 and recompile:


 if (ani_state-listen_time == 0)
 return;

 it's possible that it'll 

Problem with Atheros adapter (TP-LINK TL-WDN4800, AR9380)

2014-12-14 Thread David Blewett
Hi All:

I've been having issues with a TP-LINK TL-WDN4800 PCI Express adapter [1]
that seem to have become worse since upgrading to 10.1-RELEASE. Since then,
I've had a couple of kernel panics. I use the adapter in hostapd mode as my
main WiFi access point. I would appreciate any help in making my setup more
reliable, and am open to trying to just about anything to help
troubleshoot.

The problems that I see are regular disconnections from a Macbook Pro
client. When this happens, sometimes the wifi icon says it is connected,
but I cannot actually transmit anything to the host (i.e., pings fail). The
only recourse is to turn the mbp's wifi off, then back on. This *usually*
succeeds, but sometimes does not. I also have an iPad that has similar
issues; turning wifi off and back on is usually sufficient to restore
connectivity.

This is my /etc/rc.conf:

wlans_ath0=wlan0
create_args_wlan0=wlanmode hostap bssid
ifconfig_wlan0=HOSTAP ssid MiddleEarth mode 11ng channel 2

and /etc/hostapd-wlan0.conf:

interface=wlan0
debug=2
ctrl_interface=/var/run/hostapd
ctrl_interface_group=wheel
ssid=MiddleEarth
wpa=2
wpa_passphrase=
wpa_key_mgmt=WPA-PSK
wpa_pairwise=CCMP

I see a lot of this in the logs:

ath0: stuck beacon; resetting (bmiss count 8)

However, from what I've read these are due to channel congestion and are to
be expected. The kernel panics look like this:

Fatal trap 18: integer divide fault while in kernel mode
cpuid = 1; apic id = 11
instruction pointer = 0x20:0x8045161d
stack pointer   = 0x28:0xfe00e43bb760
frame pointer   = 0x28:0xfe00e43bb7b0
code segment= base 0x0, limit 0xf, type 0x1b
= DPL 0, pres 1, long 1, def32 0, gran 1
processor eflags= interrupt enabled, ath0: stuck beacon; resetting
(bmiss count 8)
resume, IOPL = 0
current process = 12 (swi4: clock)
trap number = 18
panic: integer divide fault
cpuid = 0
KDB: stack backtrace:
#0 0x80963000 at kdb_backtrace+0x60
#1 0x80928125 at panic+0x155
#2 0x80d24f1f at trap_fatal+0x38f
#3 0x80d24b7c at trap+0x75c
#4 0x80d0a782 at calltrap+0x8
#5 0x8045b248 at ar9300_ani_poll_freebsd+0x48
#6 0x804093d6 at ath_calibrate+0xf6
#7 0x8093d557 at softclock_call_cc+0x177
#8 0x8093d994 at softclock+0x94
#9 0x808faf4b at intr_event_execute_handlers+0xab
#10 0x808fb396 at ithread_loop+0x96
#11 0x808f8b6a at fork_exit+0x9a

Here is the pciconf output:

ath0@pci0:2:0:0:class=0x028000 card=0x3112168c chip=0x0030168c rev=0x01
hdr=0x00
vendor = 'Atheros Communications Inc.'
device = 'AR9300 Wireless LAN adaptor'
class  = network

1. http://www.tp-link.com/lk/products/details/?model=TL-WDN4800

Thanks,

David Blewett
___
freebsd-wireless@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-wireless
To unsubscribe, send any mail to freebsd-wireless-unsubscr...@freebsd.org


Re: Problem with Atheros adapter (TP-LINK TL-WDN4800, AR9380)

2014-12-14 Thread Adrian Chadd
Hi!

Hm, an integer divide by zero fault. Tsk.

can you do this:

kgdb /boot/kernel/kernel
list *0x8045161d

Thanks!



-adrian


On 14 December 2014 at 11:24, David Blewett da...@dawninglight.net wrote:
 Hi All:

 I've been having issues with a TP-LINK TL-WDN4800 PCI Express adapter [1]
 that seem to have become worse since upgrading to 10.1-RELEASE. Since then,
 I've had a couple of kernel panics. I use the adapter in hostapd mode as my
 main WiFi access point. I would appreciate any help in making my setup more
 reliable, and am open to trying to just about anything to help
 troubleshoot.

 The problems that I see are regular disconnections from a Macbook Pro
 client. When this happens, sometimes the wifi icon says it is connected,
 but I cannot actually transmit anything to the host (i.e., pings fail). The
 only recourse is to turn the mbp's wifi off, then back on. This *usually*
 succeeds, but sometimes does not. I also have an iPad that has similar
 issues; turning wifi off and back on is usually sufficient to restore
 connectivity.

 This is my /etc/rc.conf:

 wlans_ath0=wlan0
 create_args_wlan0=wlanmode hostap bssid
 ifconfig_wlan0=HOSTAP ssid MiddleEarth mode 11ng channel 2

 and /etc/hostapd-wlan0.conf:

 interface=wlan0
 debug=2
 ctrl_interface=/var/run/hostapd
 ctrl_interface_group=wheel
 ssid=MiddleEarth
 wpa=2
 wpa_passphrase=
 wpa_key_mgmt=WPA-PSK
 wpa_pairwise=CCMP

 I see a lot of this in the logs:

 ath0: stuck beacon; resetting (bmiss count 8)

 However, from what I've read these are due to channel congestion and are to
 be expected. The kernel panics look like this:

 Fatal trap 18: integer divide fault while in kernel mode
 cpuid = 1; apic id = 11
 instruction pointer = 0x20:0x8045161d
 stack pointer   = 0x28:0xfe00e43bb760
 frame pointer   = 0x28:0xfe00e43bb7b0
 code segment= base 0x0, limit 0xf, type 0x1b
 = DPL 0, pres 1, long 1, def32 0, gran 1
 processor eflags= interrupt enabled, ath0: stuck beacon; resetting
 (bmiss count 8)
 resume, IOPL = 0
 current process = 12 (swi4: clock)
 trap number = 18
 panic: integer divide fault
 cpuid = 0
 KDB: stack backtrace:
 #0 0x80963000 at kdb_backtrace+0x60
 #1 0x80928125 at panic+0x155
 #2 0x80d24f1f at trap_fatal+0x38f
 #3 0x80d24b7c at trap+0x75c
 #4 0x80d0a782 at calltrap+0x8
 #5 0x8045b248 at ar9300_ani_poll_freebsd+0x48
 #6 0x804093d6 at ath_calibrate+0xf6
 #7 0x8093d557 at softclock_call_cc+0x177
 #8 0x8093d994 at softclock+0x94
 #9 0x808faf4b at intr_event_execute_handlers+0xab
 #10 0x808fb396 at ithread_loop+0x96
 #11 0x808f8b6a at fork_exit+0x9a

 Here is the pciconf output:

 ath0@pci0:2:0:0:class=0x028000 card=0x3112168c chip=0x0030168c rev=0x01
 hdr=0x00
 vendor = 'Atheros Communications Inc.'
 device = 'AR9300 Wireless LAN adaptor'
 class  = network

 1. http://www.tp-link.com/lk/products/details/?model=TL-WDN4800

 Thanks,

 David Blewett
 ___
 freebsd-wireless@freebsd.org mailing list
 http://lists.freebsd.org/mailman/listinfo/freebsd-wireless
 To unsubscribe, send any mail to freebsd-wireless-unsubscr...@freebsd.org
___
freebsd-wireless@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-wireless
To unsubscribe, send any mail to freebsd-wireless-unsubscr...@freebsd.org


Re: Problem with Atheros adapter (TP-LINK TL-WDN4800, AR9380)

2014-12-14 Thread David Blewett
If this would be easier on IRC, I'm BlueAidan on Freenode.

Output:

(kgdb) list *0x8045161d
0x8045161d is in ar9300_ani_ar_poll
(/usr/src/sys/contrib/dev/ath/ath_hal/ar9300/ar9300_ani.c:1180).
1175 */
1176if (!DO_ANI(ah)) {
1177return;
1178}
1179
1180ofdm_phy_err_rate =
1181ani_state-ofdm_phy_err_count * 1000 /
ani_state-listen_time;
1182cck_phy_err_rate =
1183ani_state-cck_phy_err_count * 1000 /
ani_state-listen_time;
1184

Thanks,

David Blewett

On Sun, Dec 14, 2014 at 2:39 PM, Adrian Chadd adr...@freebsd.org wrote:

 Hi!

 Hm, an integer divide by zero fault. Tsk.

 can you do this:

 kgdb /boot/kernel/kernel
 list *0x8045161d

 Thanks!



 -adrian


 On 14 December 2014 at 11:24, David Blewett da...@dawninglight.net
 wrote:
  Hi All:
 
  I've been having issues with a TP-LINK TL-WDN4800 PCI Express adapter [1]
  that seem to have become worse since upgrading to 10.1-RELEASE. Since
 then,
  I've had a couple of kernel panics. I use the adapter in hostapd mode as
 my
  main WiFi access point. I would appreciate any help in making my setup
 more
  reliable, and am open to trying to just about anything to help
  troubleshoot.
 
  The problems that I see are regular disconnections from a Macbook Pro
  client. When this happens, sometimes the wifi icon says it is connected,
  but I cannot actually transmit anything to the host (i.e., pings fail).
 The
  only recourse is to turn the mbp's wifi off, then back on. This *usually*
  succeeds, but sometimes does not. I also have an iPad that has similar
  issues; turning wifi off and back on is usually sufficient to restore
  connectivity.
 
  This is my /etc/rc.conf:
 
  wlans_ath0=wlan0
  create_args_wlan0=wlanmode hostap bssid
  ifconfig_wlan0=HOSTAP ssid MiddleEarth mode 11ng channel 2
 
  and /etc/hostapd-wlan0.conf:
 
  interface=wlan0
  debug=2
  ctrl_interface=/var/run/hostapd
  ctrl_interface_group=wheel
  ssid=MiddleEarth
  wpa=2
  wpa_passphrase=
  wpa_key_mgmt=WPA-PSK
  wpa_pairwise=CCMP
 
  I see a lot of this in the logs:
 
  ath0: stuck beacon; resetting (bmiss count 8)
 
  However, from what I've read these are due to channel congestion and are
 to
  be expected. The kernel panics look like this:
 
  Fatal trap 18: integer divide fault while in kernel mode
  cpuid = 1; apic id = 11
  instruction pointer = 0x20:0x8045161d
  stack pointer   = 0x28:0xfe00e43bb760
  frame pointer   = 0x28:0xfe00e43bb7b0
  code segment= base 0x0, limit 0xf, type 0x1b
  = DPL 0, pres 1, long 1, def32 0, gran 1
  processor eflags= interrupt enabled, ath0: stuck beacon; resetting
  (bmiss count 8)
  resume, IOPL = 0
  current process = 12 (swi4: clock)
  trap number = 18
  panic: integer divide fault
  cpuid = 0
  KDB: stack backtrace:
  #0 0x80963000 at kdb_backtrace+0x60
  #1 0x80928125 at panic+0x155
  #2 0x80d24f1f at trap_fatal+0x38f
  #3 0x80d24b7c at trap+0x75c
  #4 0x80d0a782 at calltrap+0x8
  #5 0x8045b248 at ar9300_ani_poll_freebsd+0x48
  #6 0x804093d6 at ath_calibrate+0xf6
  #7 0x8093d557 at softclock_call_cc+0x177
  #8 0x8093d994 at softclock+0x94
  #9 0x808faf4b at intr_event_execute_handlers+0xab
  #10 0x808fb396 at ithread_loop+0x96
  #11 0x808f8b6a at fork_exit+0x9a
 
  Here is the pciconf output:
 
  ath0@pci0:2:0:0:class=0x028000 card=0x3112168c chip=0x0030168c
 rev=0x01
  hdr=0x00
  vendor = 'Atheros Communications Inc.'
  device = 'AR9300 Wireless LAN adaptor'
  class  = network
 
  1. http://www.tp-link.com/lk/products/details/?model=TL-WDN4800
 
  Thanks,
 
  David Blewett
  ___
  freebsd-wireless@freebsd.org mailing list
  http://lists.freebsd.org/mailman/listinfo/freebsd-wireless
  To unsubscribe, send any mail to 
 freebsd-wireless-unsubscr...@freebsd.org

___
freebsd-wireless@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-wireless
To unsubscribe, send any mail to freebsd-wireless-unsubscr...@freebsd.org


Re: Problem with Atheros adapter (TP-LINK TL-WDN4800, AR9380)

2014-12-14 Thread Adrian Chadd
Aha! Listen time is 0. Hm!


Just add this at line 1179 and recompile:


if (ani_state-listen_time == 0)
return;

it's possible that it'll happen; it shoud be bailing out sooner rather
than later.



-adrian


On 14 December 2014 at 11:53, David Blewett da...@dawninglight.net wrote:
 If this would be easier on IRC, I'm BlueAidan on Freenode.

 Output:

 (kgdb) list *0x8045161d
 0x8045161d is in ar9300_ani_ar_poll
 (/usr/src/sys/contrib/dev/ath/ath_hal/ar9300/ar9300_ani.c:1180).
 1175 */
 1176if (!DO_ANI(ah)) {
 1177return;
 1178}
 1179
 1180ofdm_phy_err_rate =
 1181ani_state-ofdm_phy_err_count * 1000 /
 ani_state-listen_time;
 1182cck_phy_err_rate =
 1183ani_state-cck_phy_err_count * 1000 /
 ani_state-listen_time;
 1184

 Thanks,

 David Blewett

 On Sun, Dec 14, 2014 at 2:39 PM, Adrian Chadd adr...@freebsd.org wrote:

 Hi!

 Hm, an integer divide by zero fault. Tsk.

 can you do this:

 kgdb /boot/kernel/kernel
 list *0x8045161d

 Thanks!



 -adrian


 On 14 December 2014 at 11:24, David Blewett da...@dawninglight.net
 wrote:
  Hi All:
 
  I've been having issues with a TP-LINK TL-WDN4800 PCI Express adapter
  [1]
  that seem to have become worse since upgrading to 10.1-RELEASE. Since
  then,
  I've had a couple of kernel panics. I use the adapter in hostapd mode as
  my
  main WiFi access point. I would appreciate any help in making my setup
  more
  reliable, and am open to trying to just about anything to help
  troubleshoot.
 
  The problems that I see are regular disconnections from a Macbook Pro
  client. When this happens, sometimes the wifi icon says it is connected,
  but I cannot actually transmit anything to the host (i.e., pings fail).
  The
  only recourse is to turn the mbp's wifi off, then back on. This
  *usually*
  succeeds, but sometimes does not. I also have an iPad that has similar
  issues; turning wifi off and back on is usually sufficient to restore
  connectivity.
 
  This is my /etc/rc.conf:
 
  wlans_ath0=wlan0
  create_args_wlan0=wlanmode hostap bssid
  ifconfig_wlan0=HOSTAP ssid MiddleEarth mode 11ng channel 2
 
  and /etc/hostapd-wlan0.conf:
 
  interface=wlan0
  debug=2
  ctrl_interface=/var/run/hostapd
  ctrl_interface_group=wheel
  ssid=MiddleEarth
  wpa=2
  wpa_passphrase=
  wpa_key_mgmt=WPA-PSK
  wpa_pairwise=CCMP
 
  I see a lot of this in the logs:
 
  ath0: stuck beacon; resetting (bmiss count 8)
 
  However, from what I've read these are due to channel congestion and are
  to
  be expected. The kernel panics look like this:
 
  Fatal trap 18: integer divide fault while in kernel mode
  cpuid = 1; apic id = 11
  instruction pointer = 0x20:0x8045161d
  stack pointer   = 0x28:0xfe00e43bb760
  frame pointer   = 0x28:0xfe00e43bb7b0
  code segment= base 0x0, limit 0xf, type 0x1b
  = DPL 0, pres 1, long 1, def32 0, gran 1
  processor eflags= interrupt enabled, ath0: stuck beacon; resetting
  (bmiss count 8)
  resume, IOPL = 0
  current process = 12 (swi4: clock)
  trap number = 18
  panic: integer divide fault
  cpuid = 0
  KDB: stack backtrace:
  #0 0x80963000 at kdb_backtrace+0x60
  #1 0x80928125 at panic+0x155
  #2 0x80d24f1f at trap_fatal+0x38f
  #3 0x80d24b7c at trap+0x75c
  #4 0x80d0a782 at calltrap+0x8
  #5 0x8045b248 at ar9300_ani_poll_freebsd+0x48
  #6 0x804093d6 at ath_calibrate+0xf6
  #7 0x8093d557 at softclock_call_cc+0x177
  #8 0x8093d994 at softclock+0x94
  #9 0x808faf4b at intr_event_execute_handlers+0xab
  #10 0x808fb396 at ithread_loop+0x96
  #11 0x808f8b6a at fork_exit+0x9a
 
  Here is the pciconf output:
 
  ath0@pci0:2:0:0:class=0x028000 card=0x3112168c chip=0x0030168c
  rev=0x01
  hdr=0x00
  vendor = 'Atheros Communications Inc.'
  device = 'AR9300 Wireless LAN adaptor'
  class  = network
 
  1. http://www.tp-link.com/lk/products/details/?model=TL-WDN4800
 
  Thanks,
 
  David Blewett
  ___
  freebsd-wireless@freebsd.org mailing list
  http://lists.freebsd.org/mailman/listinfo/freebsd-wireless
  To unsubscribe, send any mail to
  freebsd-wireless-unsubscr...@freebsd.org
___
freebsd-wireless@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-wireless
To unsubscribe, send any mail to freebsd-wireless-unsubscr...@freebsd.org


Re: Problem with Atheros adapter (TP-LINK TL-WDN4800, AR9380)

2014-12-14 Thread David Blewett
Okay, I can try to give that a go. Will take me a bit. I use ZFS on
encrypted geli devices, and haven't tried building a custom kernel yet with
this config.

Thanks,

David Blewett

On Sun, Dec 14, 2014 at 2:55 PM, Adrian Chadd adr...@freebsd.org wrote:

 Aha! Listen time is 0. Hm!


 Just add this at line 1179 and recompile:


 if (ani_state-listen_time == 0)
 return;

 it's possible that it'll happen; it shoud be bailing out sooner rather
 than later.



 -adrian


 On 14 December 2014 at 11:53, David Blewett da...@dawninglight.net
 wrote:
  If this would be easier on IRC, I'm BlueAidan on Freenode.
 
  Output:
 
  (kgdb) list *0x8045161d
  0x8045161d is in ar9300_ani_ar_poll
  (/usr/src/sys/contrib/dev/ath/ath_hal/ar9300/ar9300_ani.c:1180).
  1175 */
  1176if (!DO_ANI(ah)) {
  1177return;
  1178}
  1179
  1180ofdm_phy_err_rate =
  1181ani_state-ofdm_phy_err_count * 1000 /
  ani_state-listen_time;
  1182cck_phy_err_rate =
  1183ani_state-cck_phy_err_count * 1000 /
  ani_state-listen_time;
  1184
 
  Thanks,
 
  David Blewett
 
  On Sun, Dec 14, 2014 at 2:39 PM, Adrian Chadd adr...@freebsd.org
 wrote:
 
  Hi!
 
  Hm, an integer divide by zero fault. Tsk.
 
  can you do this:
 
  kgdb /boot/kernel/kernel
  list *0x8045161d
 
  Thanks!
 
 
 
  -adrian
 
 
  On 14 December 2014 at 11:24, David Blewett da...@dawninglight.net
  wrote:
   Hi All:
  
   I've been having issues with a TP-LINK TL-WDN4800 PCI Express adapter
   [1]
   that seem to have become worse since upgrading to 10.1-RELEASE. Since
   then,
   I've had a couple of kernel panics. I use the adapter in hostapd mode
 as
   my
   main WiFi access point. I would appreciate any help in making my setup
   more
   reliable, and am open to trying to just about anything to help
   troubleshoot.
  
   The problems that I see are regular disconnections from a Macbook Pro
   client. When this happens, sometimes the wifi icon says it is
 connected,
   but I cannot actually transmit anything to the host (i.e., pings
 fail).
   The
   only recourse is to turn the mbp's wifi off, then back on. This
   *usually*
   succeeds, but sometimes does not. I also have an iPad that has similar
   issues; turning wifi off and back on is usually sufficient to restore
   connectivity.
  
   This is my /etc/rc.conf:
  
   wlans_ath0=wlan0
   create_args_wlan0=wlanmode hostap bssid
   ifconfig_wlan0=HOSTAP ssid MiddleEarth mode 11ng channel 2
  
   and /etc/hostapd-wlan0.conf:
  
   interface=wlan0
   debug=2
   ctrl_interface=/var/run/hostapd
   ctrl_interface_group=wheel
   ssid=MiddleEarth
   wpa=2
   wpa_passphrase=
   wpa_key_mgmt=WPA-PSK
   wpa_pairwise=CCMP
  
   I see a lot of this in the logs:
  
   ath0: stuck beacon; resetting (bmiss count 8)
  
   However, from what I've read these are due to channel congestion and
 are
   to
   be expected. The kernel panics look like this:
  
   Fatal trap 18: integer divide fault while in kernel mode
   cpuid = 1; apic id = 11
   instruction pointer = 0x20:0x8045161d
   stack pointer   = 0x28:0xfe00e43bb760
   frame pointer   = 0x28:0xfe00e43bb7b0
   code segment= base 0x0, limit 0xf, type 0x1b
   = DPL 0, pres 1, long 1, def32 0, gran 1
   processor eflags= interrupt enabled, ath0: stuck beacon; resetting
   (bmiss count 8)
   resume, IOPL = 0
   current process = 12 (swi4: clock)
   trap number = 18
   panic: integer divide fault
   cpuid = 0
   KDB: stack backtrace:
   #0 0x80963000 at kdb_backtrace+0x60
   #1 0x80928125 at panic+0x155
   #2 0x80d24f1f at trap_fatal+0x38f
   #3 0x80d24b7c at trap+0x75c
   #4 0x80d0a782 at calltrap+0x8
   #5 0x8045b248 at ar9300_ani_poll_freebsd+0x48
   #6 0x804093d6 at ath_calibrate+0xf6
   #7 0x8093d557 at softclock_call_cc+0x177
   #8 0x8093d994 at softclock+0x94
   #9 0x808faf4b at intr_event_execute_handlers+0xab
   #10 0x808fb396 at ithread_loop+0x96
   #11 0x808f8b6a at fork_exit+0x9a
  
   Here is the pciconf output:
  
   ath0@pci0:2:0:0:class=0x028000 card=0x3112168c chip=0x0030168c
   rev=0x01
   hdr=0x00
   vendor = 'Atheros Communications Inc.'
   device = 'AR9300 Wireless LAN adaptor'
   class  = network
  
   1. http://www.tp-link.com/lk/products/details/?model=TL-WDN4800
  
   Thanks,
  
   David Blewett
   ___
   freebsd-wireless@freebsd.org mailing list
   http://lists.freebsd.org/mailman/listinfo/freebsd-wireless
   To unsubscribe, send any mail to
   freebsd-wireless-unsubscr...@freebsd.org

___
freebsd-wireless@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-wireless
To unsubscribe, send any mail to freebsd-wireless-unsubscr...@freebsd.org


Re: Problem with Atheros adapter (TP-LINK TL-WDN4800, AR9380)

2014-12-14 Thread David Blewett
​Okay, I've built a custom kernel with if_ath_pci as a module, rebooted
into the new kernel and got it running. I also enabled AH_DEBUG. Will let
you know how things go this week.​

Thanks,

David Blewett

On Sun, Dec 14, 2014 at 2:58 PM, David Blewett da...@dawninglight.net
wrote:

 Okay, I can try to give that a go. Will take me a bit. I use ZFS on
 encrypted geli devices, and haven't tried building a custom kernel yet with
 this config.

 Thanks,

 David Blewett

 On Sun, Dec 14, 2014 at 2:55 PM, Adrian Chadd adr...@freebsd.org wrote:

 Aha! Listen time is 0. Hm!


 Just add this at line 1179 and recompile:


 if (ani_state-listen_time == 0)
 return;

 it's possible that it'll happen; it shoud be bailing out sooner rather
 than later.



 -adrian


 On 14 December 2014 at 11:53, David Blewett da...@dawninglight.net
 wrote:
  If this would be easier on IRC, I'm BlueAidan on Freenode.
 
  Output:
 
  (kgdb) list *0x8045161d
  0x8045161d is in ar9300_ani_ar_poll
  (/usr/src/sys/contrib/dev/ath/ath_hal/ar9300/ar9300_ani.c:1180).
  1175 */
  1176if (!DO_ANI(ah)) {
  1177return;
  1178}
  1179
  1180ofdm_phy_err_rate =
  1181ani_state-ofdm_phy_err_count * 1000 /
  ani_state-listen_time;
  1182cck_phy_err_rate =
  1183ani_state-cck_phy_err_count * 1000 /
  ani_state-listen_time;
  1184
 
  Thanks,
 
  David Blewett
 
  On Sun, Dec 14, 2014 at 2:39 PM, Adrian Chadd adr...@freebsd.org
 wrote:
 
  Hi!
 
  Hm, an integer divide by zero fault. Tsk.
 
  can you do this:
 
  kgdb /boot/kernel/kernel
  list *0x8045161d
 
  Thanks!
 
 
 
  -adrian
 
 
  On 14 December 2014 at 11:24, David Blewett da...@dawninglight.net
  wrote:
   Hi All:
  
   I've been having issues with a TP-LINK TL-WDN4800 PCI Express adapter
   [1]
   that seem to have become worse since upgrading to 10.1-RELEASE. Since
   then,
   I've had a couple of kernel panics. I use the adapter in hostapd
 mode as
   my
   main WiFi access point. I would appreciate any help in making my
 setup
   more
   reliable, and am open to trying to just about anything to help
   troubleshoot.
  
   The problems that I see are regular disconnections from a Macbook Pro
   client. When this happens, sometimes the wifi icon says it is
 connected,
   but I cannot actually transmit anything to the host (i.e., pings
 fail).
   The
   only recourse is to turn the mbp's wifi off, then back on. This
   *usually*
   succeeds, but sometimes does not. I also have an iPad that has
 similar
   issues; turning wifi off and back on is usually sufficient to restore
   connectivity.
  
   This is my /etc/rc.conf:
  
   wlans_ath0=wlan0
   create_args_wlan0=wlanmode hostap bssid
   ifconfig_wlan0=HOSTAP ssid MiddleEarth mode 11ng channel 2
  
   and /etc/hostapd-wlan0.conf:
  
   interface=wlan0
   debug=2
   ctrl_interface=/var/run/hostapd
   ctrl_interface_group=wheel
   ssid=MiddleEarth
   wpa=2
   wpa_passphrase=
   wpa_key_mgmt=WPA-PSK
   wpa_pairwise=CCMP
  
   I see a lot of this in the logs:
  
   ath0: stuck beacon; resetting (bmiss count 8)
  
   However, from what I've read these are due to channel congestion and
 are
   to
   be expected. The kernel panics look like this:
  
   Fatal trap 18: integer divide fault while in kernel mode
   cpuid = 1; apic id = 11
   instruction pointer = 0x20:0x8045161d
   stack pointer   = 0x28:0xfe00e43bb760
   frame pointer   = 0x28:0xfe00e43bb7b0
   code segment= base 0x0, limit 0xf, type 0x1b
   = DPL 0, pres 1, long 1, def32 0, gran 1
   processor eflags= interrupt enabled, ath0: stuck beacon;
 resetting
   (bmiss count 8)
   resume, IOPL = 0
   current process = 12 (swi4: clock)
   trap number = 18
   panic: integer divide fault
   cpuid = 0
   KDB: stack backtrace:
   #0 0x80963000 at kdb_backtrace+0x60
   #1 0x80928125 at panic+0x155
   #2 0x80d24f1f at trap_fatal+0x38f
   #3 0x80d24b7c at trap+0x75c
   #4 0x80d0a782 at calltrap+0x8
   #5 0x8045b248 at ar9300_ani_poll_freebsd+0x48
   #6 0x804093d6 at ath_calibrate+0xf6
   #7 0x8093d557 at softclock_call_cc+0x177
   #8 0x8093d994 at softclock+0x94
   #9 0x808faf4b at intr_event_execute_handlers+0xab
   #10 0x808fb396 at ithread_loop+0x96
   #11 0x808f8b6a at fork_exit+0x9a
  
   Here is the pciconf output:
  
   ath0@pci0:2:0:0:class=0x028000 card=0x3112168c chip=0x0030168c
   rev=0x01
   hdr=0x00
   vendor = 'Atheros Communications Inc.'
   device = 'AR9300 Wireless LAN adaptor'
   class  = network
  
   1. http://www.tp-link.com/lk/products/details/?model=TL-WDN4800
  
   Thanks,
  
   David Blewett
   ___
   freebsd-wireless@freebsd.org mailing list
   http://lists.freebsd.org/mailman/listinfo/freebsd-wireless
   To unsubscribe, send any mail to