Please note: This project is no longer active. The website is kept online for historic purposes only.
If you´re looking for a Linux driver for your Atheros WLAN device, you should continue here .

Ticket #282 (new defect)

Opened 14 years ago

Last modified 12 years ago

loading module ath_pci on a ultra sparc 5 freezes the system

Reported by: Assigned to:
Priority: critical Milestone:
Component: madwifi: driver Version: trunk
Keywords: Cc:
Patch is attached: 0 Pending:


Loading module ath_pci on a ultra sparc 5 freezes the system. ath_hal loads correctly.

Some info about the pci card:

D-Link DWL-G520

HW ver B2

FW ver 3.1.6

Info about Atheros and wlan driver:


wlan: (Atheros/multi-bss)

ath_rate_sample: 1.2

ath_pci: (Atheros/multi-bss)

Info about linux system:

slg@ultra:~# uname -a

Linux ultra #6 Sat Dec 24 15:55:00 CET 2005 sparc64 GNU/Linux

slg@ultra:~# gcc -v

Reading specs from /usr/lib/gcc-lib/sparc-linux-gnu/3.3.6/specs Configured with: ../src/configure -v --enable-languages=c,c++,ada --prefix=/usr --mandir=/usr/share/man --infodir=/usr/share/info --with-gxx-include-dir=/usr/include/c++/3.3 --enable-shared --enable-_cxa_atexit --with-system-zlib --enable-nls --without-included-gettext --enable-clocale=gnu --enable-debug --with-cpu=v7 sparc-linux-gnu Thread model: posix

gcc version 3.3.6 (Debian 1:3.3.6-10)

binutils 2.16.1cvs20051

Please let me know if I can help.

Regards, slg

Change History

01/04/06 07:30:20 changed by mrenzmann

  • version set to trunk.

I don't think this will cure your problem, nevertheless it would be good if you could retry with the latest revision of the madwifi-ng code (r1378 as of this writing).

Is there something similar like a kernel oops facility available on Ultra Sparcs? If so, please provide that.

01/30/06 14:49:11 changed by

I'm having the same problem on my u5, just hardlocks when loading ath_pci. Madwifi snapshot r1416, kernel, gcc-3.4.5, binutils-2.16.1. Card is a DWL-G510 rev. B1 (AR5215). Just ring my bell in anything you want to test.

01/30/06 15:01:27 changed by anonymous

Maybe it gives some additional hints if you try to insmod ath_pci with option ath_debug=0xffffffff. Would be really helpful if you could (hand)copy&paste the lines from the output here.

01/30/06 19:55:55 changed by

Managed to get some info in single-user mode:

(none) / # modprobe ath_pci ath_debug=0xffff ath_hal: (AR5210, AR5211, AR5212, RF5111, RF5112, RF2413, RF5413, REGOPS_FUNC, DFS) wlan: (Atheros/multi-bss) ath_rate_sample: 1.2 ath_pci: (Atheros/multi-bss) ath_attach: devid 0x1a

[A couple of seconds later...]

BUG: soft lockup detected on CPU#0! TSTATE: 0000000011009601 TPC: 00000000004050b0 TNPC: 00000000004050b4 Y: 00000000 Tainted: P TPC: <etrap_irq+0xec/0x118> g0: fffff8001f1c2f20 g1: 00000000ffffc000 g2: 0000000000000000 g3: 0000000000004004 g4: fffff8001f3cbc00 g5: 0000000000000000 g6: fffff8001f1c0000 g7: 0000000000000000 o0: 0000000000000001 o1: fffff8001f1c0400 o2: 00000000100001fc o3: 0000000000000000 o4: 0000000000000000 o5: 0000000000000004 sp: fffff8001f1c2661 ret_pc: 0000000000408cd0 RPC: <sparc64_realfault_common+0x10/0x20> l0: 0000000000684800 l1: 0000000911009600 l2: 0000000000408ccc l3: 0000000000000000 l4: 0000000000000002 l5: 000001ff00014000 l6: fffff8001f1c0000 l7: 0000000000000008 i0: 000001ff00010000 i1: 0000000100000004 i2: 00000000006a6400 i3: fffff8001f1c30f0 i4: 000000000069e03f i5: 0000000000000000 i6: fffff8001f1c27c1 i7: 0000000010014e24 I7: <zz016d9d41+0x44/0x1f4 [ath_hal]>

04/22/06 20:42:27 changed by dyqith

  • patch_attached changed.

any updates on this issue?

04/29/06 17:13:26 changed by

Cheers :)

I have the same problem. I offer my help with debug info, tester..

04/29/06 17:36:29 changed by dyqith

From an earlier post, please try these steps:

Run under console mode (to see the kernel panics/oops) Make sure none of the modules of madwifi is loaded ath_pci should load with "option ath_debug=0xffffffff" modprobe ath_pci (if no modprobe, load the modules in order of: ath_hal, ath_rate_sample, ath_pci, wlan, wlan_scan_sta, wlan_scan_ap)

If the kernel oops have the ksyms available, it'll be easier to debug.

05/10/06 06:02:18 changed by dyqith

Can someone try it with the latest svn revision and report back ?

05/10/06 17:02:47 changed by

Same hardlock as usual with r1545.

05/10/06 18:10:02 changed by dyqith

rastejante and gustavoz, are you both using sparc also or some other arch ?

There was a ticket that said sparc hal was messed up ?

05/10/06 18:49:17 changed by anonymous

An ultra 5 (sparc64) in my case. All ultra machines are basically sparc64. Gentoo is currently using gcc-3.4.x for the stable toolchain, there were some ABI changes for sparc that are documented here There's ticket #493 about the hal issues, problem is some distributions could use different gcc versions. 4.0.x won't break if built with 3.4.x, i have to check on 4.1.x but i don't think it'll break either, but 4.1.x is known to make bad code for sparc so it would be wise to stay far away for the time being. I could throw the card in a 280R but i doubt it'll help, and i don't know if madwifi-ng is pci domain aware.

05/10/06 18:50:03 changed by

Forgot to change my username above :-)

05/16/06 03:41:34 changed by dyqith

07/30/06 16:56:25 changed by xsteadfastx

i have the same problem under suse linux 10.1...x86 arch. i tried to get what does this problem. i dont know if its exactly the madwifi. sometimes it works and sometimes it freezes the system. maybe it has to do with acpi or something. no clue. i use the svn version from just now :)