Please note: This project is no longer active.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 #198 (new enhancement)

Opened 16 years ago

Last modified 15 years ago

Hardware Support: AR5006XS

Reported by: Assigned to:
Priority: minor Milestone: version 1.0.0 - first stable release
Component: madwifi: HAL Version: trunk
Keywords: wmia-166ag Cc:
Patch is attached: 0 Pending:

Description (Last modified by mrenzmann)

the sparklan wmia-166ag is a 6th gen atheros mini-pci card. I added the pcid and modalias to the ath_pci driver to get it at least seen by the module, and it did, it passed it to the hal module and failed with this.

PCI: 0000:00:11.0 has unsupported PM cap regs version (7)
wifi%d: unable to attach hardware: 'No hardware present or device not yet supported' (HAL status 1)

the pcid for the ethernet and serial device is (both from the atheros board)

Ethernet: 0000:00:11.0 0200: 168c:ff16 (rev 01)
Serial:   0000:00:11.1 0700: 168c:ff96 (rev 01)

modalias is pci:v0000168Cd0000FF16sv00000000sd00000000bc02sc00i00

I used the latest pull from svn of the ng code. Here are all the versions.

ath_hal: (AR5210, AR5211, AR5212, RF5111, RF5112, RF2413, RF5413, DFS)
wlan: (Atheros/multi-bss)
ath_rate_sample: 1.2
ath_pci: (Atheros/multi-bss)

I dont know what else could be helpful info, I'll provide any required that i can though. Thanks in advance.

Change History

12/01/05 12:54:27 changed by mrenzmann

  • description changed.

I'm not sure that this is an issue that can be fixed via MadWifi. The message related to "unsupported pm regs" comes from the PCI layer of your kernel, as far as I can say. It might be worth to try to find out if you can do something to fix the cause of this message (google for the message, for example).

12/02/05 03:53:44 changed by

from what I googled, some people report the printk message as being nothing but a nuissance, the hardware still works, to the only message reporting why their hardware isn't working. I'm currently trying to get some help on lkml too.

Assuming it's nothing but a nuissance, would there be anything in the source part of the HAL module that would need to be updated to detect the card or anything in the other parts of madwifi ?

12/02/05 08:51:24 changed by mrenzmann

The "unable to attach hardware" message usually means that the driver aborted initializing your interface. Without initialization it's extremely unlikely that it will be available to your system. Do you have some examples (URLs) for these "nothing but a nuissance" comments?

12/02/05 22:43:44 changed by

the message i'm referring to is the "unsupported pm regs" since that's what we've been debating on _why_ the madwifi driver reports "unsupported hardware" etc. I know that if the driver aborts initialization it's not going to be available, that's not a question.

What was in question was if it was madwifi's code that doesn't support the card (which it definitely doesn't even detect as it is in trunk) or if it's the kernel's pci layer and hence, the interest in the printk the pci layer produces when loading the madwifi driver. I was simply reporting that there are multiple returns from google where people were recieving the same type of message but their hardware was still functional (which isn't the case for me obviously).

I think it's just a matter of the HAL not supporting my hardware since it seems to be a 6th gen chip that hasn't been added to the list of valid pcid's in the driver and who knows what goes on in the HAL for supporting cards.

If there is anything I need to do to add a card to madwifi assuming the HAL supports it, what do i need to do?

If it's the kernel's pci subsystem causing the problem, then i'll be directing all my interest to them, but so far I'm not convinced, since nobody has mentioned what, if anything, can be added to the madwifi code to properly recognize my card as a valid 6th gen atheros chipset. It didn't even detect the card before i put the pcid's and modalias info in myself for ath_pci, so i'm pretty sure there are other things i'm missing for the card to actually work with madwifi.

one of the first few results from googling the pci message. also gives some info on perhaps the printk not referring to the hardware not working.

I've yet to recieve any response on lkml about it. but i'll continue to look to see what's going on.

12/03/05 13:41:57 changed by

the chipset in the sparklan wmia-166ag is a 5414a. It seems likely i'll have to wait until the next HAL release from atheros to get any support. I've looked around the source everywhere obvious for card specific items but it seems that there isn't anything left i havn't done.

02/22/06 01:12:04 changed by anonymous

  • patch_attached changed.

I have also today compiled revision 1454 against kernel and there are still problems.

In first (cold) boot kernel loades ath_pci withnout any message about problems, but doesn't see the card. On second boot (warm) the driver is loaded without any problem. And is probably working.

I don't have sparklan card, it is some OEM, but PCI id is the same

  0000:00:11.0 0200: 168c:ff16 (rev 01)
    Flags: bus master, medium devsel, latency 64, IRQ 9
    Memory at 80100000 (32-bit, non-prefetchable) [size=64K]
    Capabilities: [44] Power Management version 7

  0000:00:11.1 0700: 168c:ff96 (rev 01) (prog-if 02)
    Subsystem: 168c:ee96
    Flags: bus master, medium devsel, latency 64, IRQ 9
    I/O ports at 1c00 [size=8]
    Memory at 80140000 (32-bit, non-prefetchable) [size=4K]
    Capabilities: [44] Power Management version 0

02/24/06 06:20:31 changed by mrenzmann

The previous comment sounds as if the problem could be related to what I've written in a comment to #370.

04/22/06 14:07:51 changed by

this problem also affects ALLNET ALL0281A cards. These cards have the same pcid for the ethernet and serial device - problems are as described in this ticket.

04/29/06 10:07:02 changed by anonymous

I found out that my ALLNET ALL0281A is able to be recognised with the corrent SVN-version (registred as 168c:0013 in the PCI chipset) on a more recent computer supporting PCI 2.2.

Only in a very old PentiumPro? board (Compaq Deskpro 4000) with a PCI version smaller than 2.2 is's reported as 168c:ff16 producing the problems as descibed in this ticket.

05/16/06 09:02:25 changed by dyqith

So this is a hardware problem ?

Can someone put it into the Compatibility page if its not there already ?


06/06/07 10:41:55 changed by

I have designed a seris of boards based on IXP425 processor. Board Ver1.0 can recognize my mini-pci wireless card(PA3416U-1MPC AR5BMB5 manufactured by ASKEY COMPUTER CORP.) with ID=168c:0013, and an AP is built with the board along with this miniPCI card; Board ver2.0 can not recognize the miniPCI card and reports two pci devices with unknown ID (168c:FF96 and 168c:FF16).NOTE: the two boards are running with exactly the same linux 2.6.10 and exactly the same miniPCI card. Board ver1.0 has only one miniPCI slot while the latter ver1.0 board has one miniPCI slot and another PCI slot, they use the same clock signal mux-ed by a clock driver chip. I don't know why ver2.0 board fails, but I can tell that is a hardware problem. If anybody has any idea, please contact me at And I'll post my solution here when the problem is solved.

07/26/07 02:15:39 changed by mentor

  • type changed from defect to enhancement.
  • summary changed from sparklan wmia-166ag has unsupported pm regs (not working) to Hardware Support: AR5006XS.