Athereos wifi gives strange "no such device" errors

Gullible Jones · 2007-06-03 21:49:18

I've installed madwifi and udev loads the module (ath_pci) automatically on start... But the Network script complains the ath0 doesn't exist when I boot up:

Error for wireless request "Set ESSID" (8B1A) :
        Set failed on device ath0 ; no such device.
Error, ath0: ioctl SIOCGIFHWADDR: No such device

Likewise, ifconfig doesn't show ath0 and iwconfig shows no wireless extensions. But I've got the right stuff in rc.conf:

lo="lo 127.0.0.1"
eth0="dhcp"
ath0="dhcp"
INTERFACES=(lo eth0 ath0)

and in /etc/conf.d/wireless:

wlan_ath0="ath0 essid MyEssid"
WLAN_INTERFACES=(ath0)

Also, I don't know if this is part of the problem but lspci tells me this about my wifi:

03:00.0 Ethernet controller: Atheros Communications, Inc. Unknown device 001c (rev 01)

What gives? Is there a known software bug? Is this unsupported hardware, contrary to what I've read about this laptop?

(Hmm, I haven't tried loading it without the country code. I'll see if that works... No, still doesn't work. Bugger.)

Last edited by Gullible Jones (2007-06-03 21:56:08)

Penguin of Wonder · 2007-06-03 22:18:48

It is definitely a driver problem, here is my lspci for my Atheros card,

00:0b.0 Ethernet controller: Atheros Communications, Inc. AR5005G 802.11abg NIC (rev 01)

I'm still trying to think of why? Which chipset do you have? Maybe the driver doesn't support that specific one yet.

Last edited by Penguin of Wonder (2007-06-03 22:19:20)

Gullible Jones · 2007-06-03 23:39:37

Umm, some "Acer InviLink" thing. Just a sec... Uh, looks like it is AR5005G, i.e. supported. Not totally sure though, it might be a 5007eg (unsupported).

Here's another case. Sure enough dmesg | grep wifi gives:

wifi%d: unable to attach hardware: 'Hardware revision not supported' (HAL status 13)

Tried this "solution" quoted on the page:

The very next thing was to try to make the wireless chip work. This laptop came with a Atheros AR5005G wireless card inside. Using a Knoppix Live CD, I already knew it worked fine using the Madwifi drivers. Using my Ethernet port, I added the Livna repository to my system and installed the drivers from there, following these instructions. But they didn't work, at first. I read around, and eventually I saw somewhere that the modules have to be removed in a certain order (so as to free up the reference counts in the right order) and then probed in a specific order (no idea why). The order to remove: ath_pci, ath_hal, ath_rate_sample, ath_hal, wlan. Don't ask me why this order, or why one module is repeated, or anything. I just know this is what I did. Then I loaded ath_pci, and the rest just sort of snapped into place, and it worked, and I was able to configure the device. On subsequent boots, the device just worked without any extra effort required.

Not only did it not work, I discovered that ath_rate_sample never got loaded.

Another case here. And another solution in reply:

YOU DO NOT NEED NDISWRAPPER. M'kay? Don't even think about it.
Here's what's worked for me, tested with both WEP and WPA encryption in plain Ubuntu Feisty:
1) Enable the Atheros restricted driver at System>Administration>Restricted Drivers Manager.
2) Under System>Administration>Network it wants everything to be "Enable Roaming Mode" - WiFi and cabled Ethernet. Surprised me because I'd been using WiFi on the old Compaq via telling it directly which network to go to. Anyways. Use the "network" panel tool top right corner of Gnome to tell it which WiFi net to hit and specify parameters (DHCP, password, password type). It will also list which networks it finds and their password status (no shield means open port). There's also a "manual configuration" choice, works great.
Overall it's a damned decent machine, esp. for the money.

I'll see if there's anything like that in my BIOS.

(There's also some guy on the first page who says it's a BIOS bug, fixable by an update. I would certainly hope not.)

Edit: Looks like the above was about a GUI administration thing, not a BIOS thing. My stupid mistake. :\ At any rate, no clue what the BIOS version is, and flashing it might be hazardous... Shit, I hope I don't end up having to use ndiswrapper, I've heard mucho horror stories about it.

Last edited by Gullible Jones (2007-06-03 23:48:49)

bartsick · 2007-06-04 21:13:44

At least be absolutly sure what type of card / chipset you own before you load the wrong driver for it.

Check the NDIS wrapper list to find out what chipset is used in which card. (ndiswrapper has a new site and a non-working search feature atm, you'd better search for it manually)
http://ndiswrapper.sourceforge.net/joom … 3/id,list/

ndiswrapper + wpa_supplicant ain't real horror... I had to use it for my Asus WL-138G PCI card.
Check out this page @ the wiki:
http://wiki.archlinux.org/index.php/Ndi … supplicant

Bios upgrading, now that is horror. lol

Last edited by bartsick (2007-06-04 21:17:49)

fivre · 2007-06-05 19:38:42

I believe madwifi worked fine until a recent update (I have a 5507Z, AR5005G) and was puttering nicely along until last night. Anyway, here are some logs:

Some related looking stuff from dmesg:

ath_hal: module license 'Proprietary' taints kernel.
ath_hal: 0.9.18.0 (AR5210, AR5211, AR5212, RF5111, RF5112, RF2413, RF5413)
wlan: 0.8.4.2 (0.9.3.1)
ath_pci: 0.9.4.5 (0.9.3.1)

...Followed by that wifi%d error. I'm going to try downgrading the stuff that was upgraded and see if it works.

Okay, no luck there. Modprobing ath_pci manually doesn't help either. However, flicking the little switch that disables/enables wifi seems to work (the light, however, is always on.)

Also possible workarounds:
- Boot into Windows beforehand
- Go into BIOS setup beforehand
This seems to be a rather common problem, it has to do with some odd firmware switch apparently.

Edit 3: Okay, got it. The card apparently 'sleeps' if I remove my machine from any power source (remove the battery and disconnect the AC adapter.) However, booting back in, flicking the wifi switch, and rebooting fixes the problem.

Now if only I could disable the speakers...

Last edited by fivre (2007-06-05 20:53:05)

Gullible Jones · 2007-06-06 01:53:51

Fivre: Speakers are controlled by "Surround" rather than "PC Speaker", which does nothing at all AFAICT.

My laptop seemed to start with the wifi hardware on - the switch appeared to disable it (temporarily?), perhaps to prevent Vista from bombarding the user with unwanted popup messages, but cycling the power seemed to reenable it.

At any rate, it looks like I've got 5007eg, i.e. the bad stuff. Damn.

Gullible Jones · 2007-06-06 02:03:55

Okay, this is fucking insane. No support from HAL and no ETA for any semblence of support given, it looks like they're not even working on it. And apparently the drivers don't work with ndiswrapper unless you do some modification. From this Google-cached version of a deleted forum page:

Hate to respond to my own post, but I was able to get this card to work under linux. It is a b*tch, and I cannot post an exact guide, but I can give a few pointers:
1) madwifi does NOT work. The HAL does not yet support this particular chipset/revision.
2) you must use ndiswrapper. I used version 1.40, but I had to manually modify the source code. The windows driver uses a kernel function that is not supported in ndiswrapper (RtlCharToInteger). I manually added this function (along with some glue code) from the Wine source. I have notified the ndiswrapper developers that I have done so, so they will either request a patch from me, or they can simply implement their own version of this function. Be aware that without this modification to the source, this card WILL NOT WORK.
3) You will need the Windows (2000/XP) driver files. There is a catch-- you cannot simply unpack the driver files from the cd/download because they're in the new Installshield 12 format, which to my knowledge (correct me if I'm wrong) cannot be extracted with existing tools. If you have the driver installed in Windows, simply search your windows directory for the following 3 files: aw5006.sys, net2425.inf & net2425.cat. You will need these files on a partition that is accessible from your Linux distribution.
4) Once you modify / build / install ndiswrapper, follow the instructions for installing the driver. After this, you will need to follow your distribution-specific instructions for loading the ndiswrapper module at boot-time, in addition to setting up your wireless parameters, etc.
This was quite an oddessy, but it does work. Obviously, the best solution will (eventually) be to use madwifi's atheros chipset driver, but for the time being, this card CAN be used under Linux, but it's just a pain in the *ss.
Hope this helps.

Fuck.

But wait a minute, here's something from the madwifi site...

Atheros AR5007EG ¶
Chipset: AR2425 / AR5007EG
URL: http://atheros.com/pt/AR5007EG.htm
Supports: 802.11b 802.11g
Interface: PCI-Express x1
Device Information: 168c:001c
Notes: not supported by HAL as of 2007.03.12
Notes: 'lspci' on Intel Mac Mini results in 168c:001c. Works w/o HAL problem. 2007.05.03. However used Atheros chip unclear

Hmm. Wonder if it might finally work if I tried it with a CVS version of HAL.

(Of course, I'm starting to think that a PCMCIA wireless card might be in order...:rolleyes:)

Last edited by Gullible Jones (2007-06-06 02:05:42)

Gullible Jones · 2007-06-06 02:08:58

Hmmm wow, I didn't know before that madwifi uses proprietary blob-crap too.

(I wonder if this chipset is supported by OpenBSD. Might be time to borrow some code from them.)

Gullible Jones · 2007-06-06 02:33:44

And to top this gem off, Atheros' own website claims that this POS hardware is supported on Linux via madwifi. Fucking incredible.

I'll see if I can get ndiswrapper to work without any source modification. If not... Either I shell out some cash for a card or start taking a look at this laptop's return policy.

fivre · 2007-06-06 02:37:14

No, my alsamixer has two things that wor, 'PCI' and 'Front.' Both control headphones and speakers equally, there's no way to shut off the speakers.

I covered them with tape.

Gullible Jones · 2007-06-06 03:25:41

Fivre: try blacklisting the pcspkr module in rc.conf.

Re ndiswrapper: worked out better than I expected... Installed fine, module was able to load, etc. However, wlan0 still doesn't exist. dmesg gives me:

ndiswrapper (ntoskernel_exit:2650): object dbec3de0 type 2 was not freed, freeing it now
ndiswrapper version 1.43 loaded (smp=yes)
ndiswrapper: driver net5211 (,04/05/2007,5.3.0.35) loaded
ACPI: PCI Interrupt 0000:03:00.0[A] -> GSI 17 (level, low) -> IRQ 17
ndiswrapper (ZwClose:2194): closing handle 0xdeab25e8 not implemented
PCI: Setting latency timer of device 0000:03:00.0 to 64
ndiswrapper (NdisWriteErrorLogEntry:199): log: C0001389, count: 4, return_address: e051d054
ndiswrapper (NdisWriteErrorLogEntry:202): code: 0xde70b8cc
ndiswrapper (NdisWriteErrorLogEntry:202): code: 0x28
ndiswrapper (NdisWriteErrorLogEntry:202): code: 0xe03a5000
ndiswrapper (NdisWriteErrorLogEntry:202): code: 0xe03a5000
ndiswrapper (miniport_init:271): couldn't initialize device: C000009A
ndiswrapper (pnp_start_device:425): Windows driver couldn't initialize the device (C0000001)
unregister_netdevice: device eth%d/de70b000 never was registered
BUG: at net/core/dev.c:3259 unregister_netdevice()
 [<c02bfdcf>] unregister_netdev+0xf/0x20
 [<e04cd9bb>] wrap_ndis_remove_device+0x3b/0x200 [ndiswrapper]
 [<e04ce39d>] NdisDispatchPnp+0x12d/0xdb0 [ndiswrapper]
 [<c011f7d1>] __activate_task+0x21/0x40
 [<c0170bbd>] cache_alloc_refill+0x16d/0x510
 [<c011f569>] __wake_up_common+0x39/0x60
 [<c011fba8>] __wake_up+0x38/0x50
 [<e04c7842>] IoAllocateIrp+0x62/0xa0 [ndiswrapper]
 [<e04c83e5>] IoBuildAsynchronousFsdRequest+0x35/0x170 [ndiswrapper]
 [<e04c7b95>] IofCallDriver+0x55/0xc0 [ndiswrapper]
 [<e04ca10b>] IoSendIrpTopDev+0xbb/0x120 [ndiswrapper]
 [<e04ca1e3>] pnp_remove_device+0x73/0x1d0 [ndiswrapper]
 [<e04ca6b4>] wrap_pnp_start_device+0x1f4/0x290 [ndiswrapper]
 [<e04ca795>] wrap_pnp_start_pci_device+0x45/0x50 [ndiswrapper]
 [<c01b568f>] sysfs_make_dirent+0x2f/0x50
 [<c01b665e>] sysfs_create_link+0x6e/0x160
 [<c0235398>] pci_match_device+0xa8/0xc0
 [<c0235426>] pci_device_probe+0x56/0x80
 [<c028c206>] really_probe+0x66/0x180
 [<c028c369>] driver_probe_device+0x49/0xc0
 [<c033f49d>] klist_next+0x5d/0xa0
 [<c028c510>] __driver_attach+0x80/0x90
 [<c028b6ea>] bus_for_each_dev+0x3a/0x60
 [<c028c096>] driver_attach+0x16/0x20
 [<c028c490>] __driver_attach+0x0/0x90
 [<c028ba4b>] bus_add_driver+0x7b/0x1a0
 [<c0235604>] __pci_register_driver+0x84/0xd0
 [<e04bcbf7>] loader_init+0x107/0x230 [ndiswrapper]
 [<e04caaef>] wrap_procfs_init+0x3f/0xb0 [ndiswrapper]
 [<e014d0b5>] wrapper_init+0xb5/0xc2 [ndiswrapper]
 [<c0148d5d>] sys_init_module+0x15d/0x1970
 [<c0104232>] sysenter_past_esp+0x5f/0x85
 =======================
ndiswrapper (miniport_halt:335): device de70b500 is not initialized - not halting
ndiswrapper: device eth%d removed
ACPI: PCI interrupt for device 0000:03:00.0 disabled
ndiswrapper: probe of 0000:03:00.0 failed with error -22
usbcore: registered new interface driver ndiswrapper

Time to file a bug on the ndiswrapper website?

(ndiswrapper -l says it's there though:

net5211 : driver installed
        device (168C:001C) present

WTF?)

Last edited by Gullible Jones (2007-06-06 03:41:01)

Gullible Jones · 2007-06-06 22:29:48

Well I'll be damned, ndiswrapper works like a charm after rebooting. Maybe I needed to refresh my profile or something, though it doesn't look like that. Weird. Anyway...

[proteus@caleuche ~]$ ifconfig -a
eth0      Link encap:Ethernet  HWaddr 00:1B:24:06:BD:5F  
          inet addr:192.168.0.100  Bcast:192.168.0.255  Mask:255.255.255.0
          UP BROADCAST RUNNING MULTICAST  MTU:1500  Metric:1
          RX packets:946 errors:0 dropped:0 overruns:0 frame:0
          TX packets:379 errors:0 dropped:0 overruns:0 carrier:0
          collisions:0 txqueuelen:1000 
          RX bytes:367705 (359.0 Kb)  TX bytes:56878 (55.5 Kb)
          Interrupt:16 

lo        Link encap:Local Loopback  
          inet addr:127.0.0.1  Mask:255.0.0.0
          UP LOOPBACK RUNNING  MTU:16436  Metric:1
          RX packets:0 errors:0 dropped:0 overruns:0 frame:0
          TX packets:0 errors:0 dropped:0 overruns:0 carrier:0
          collisions:0 txqueuelen:0 
          RX bytes:0 (0.0 b)  TX bytes:0 (0.0 b)

wlan0     Link encap:Ethernet  HWaddr 00:19:7E:1E:5E:35  
          BROADCAST MULTICAST  MTU:1500  Metric:1
          RX packets:0 errors:0 dropped:0 overruns:0 frame:0
          TX packets:0 errors:0 dropped:0 overruns:0 carrier:0
          collisions:0 txqueuelen:1000 
          RX bytes:0 (0.0 b)  TX bytes:0 (0.0 b)
          Interrupt:17 Memory:34100000-34110000

Woo hoo!

Gullible Jones · 2007-06-08 03:19:34

Okay, it looks like it's fixed, but there's something weird going on with the network script... It's started giving me the output of ifconfig during boot, and on shutdown it gives the failure message with something about a device not existing. wlan0 obviously exists, so no idea what's up with that... Maybe I screwed up the script when I edited it last. I hope it's not anything totally wonky anyway...

Arch Linux

#1 2007-06-03 21:49:18

Athereos wifi gives strange "no such device" errors

#2 2007-06-03 22:18:48

Re: Athereos wifi gives strange "no such device" errors

#3 2007-06-03 23:39:37

Re: Athereos wifi gives strange "no such device" errors

#4 2007-06-04 21:13:44

Re: Athereos wifi gives strange "no such device" errors

#5 2007-06-05 19:38:42

Re: Athereos wifi gives strange "no such device" errors

#6 2007-06-06 01:53:51

Re: Athereos wifi gives strange "no such device" errors

#7 2007-06-06 02:03:55

Re: Athereos wifi gives strange "no such device" errors

#8 2007-06-06 02:08:58

Re: Athereos wifi gives strange "no such device" errors

#9 2007-06-06 02:33:44

Re: Athereos wifi gives strange "no such device" errors

#10 2007-06-06 02:37:14

Re: Athereos wifi gives strange "no such device" errors

#11 2007-06-06 03:25:41

Re: Athereos wifi gives strange "no such device" errors

#12 2007-06-06 22:29:48

Re: Athereos wifi gives strange "no such device" errors

#13 2007-06-08 03:19:34

Re: Athereos wifi gives strange "no such device" errors

Board footer