pci unsupported PM regs version (7), means hardware isn't working?

pci unsupported PM regs version (7), means hardware isn't working?

[Ed Sweetman]

I'm getting this warning when i try to load the madwifi drivers on my 
WRAP board for the WMIA-166AG mini-pci card using kernel 2.6.13.3 and 
the latest trunk of madwifi.  

This is the only error that's printed before the HAL driver for madwifi 
responds with "no hardware found or unsupported hardware" etc etc.   I 
had to add the pcid's to madwifi for it to even detect it enough to try 
and send it to the HAL module, but the madwifi dev team isn't looking at 
any bug reports because of this printk that's being made by the PCI 
subsystem in the kernel. 

So, does this printk mean anything (i've seen posts where the hardware 
producing it was working, the printks were just a nuissance) or does it 
indicate some issue the PCI subsystem is having in powering the card up 
and communicating with it.   In either case, I'd be more than happy with 
providing anyone able to patch the pci code with information i have on 
the card. 

If it's nothing but a harmless warning, i'll forward the response to the 
madwifi wiki and mailing list, so something can be done upstream to the 
hal module to work the card. 

Thanks in advance. 

Re: pci unsupported PM regs version (7), means hardware isn't working?

[Adam Belay]

On Thu, Dec 01, 2005 at 05:58:20PM -0500, Ed Sweetman wrote:
> I'm getting this warning when i try to load the madwifi drivers on my 
> WRAP board for the WMIA-166AG mini-pci card using kernel 2.6.13.3 and 
> the latest trunk of madwifi.  
> 
> This is the only error that's printed before the HAL driver for madwifi 
> responds with "no hardware found or unsupported hardware" etc etc.   I 
> had to add the pcid's to madwifi for it to even detect it enough to try 
> and send it to the HAL module, but the madwifi dev team isn't looking at 
> any bug reports because of this printk that's being made by the PCI 
> subsystem in the kernel. 
> 
> So, does this printk mean anything (i've seen posts where the hardware 
> producing it was working, the printks were just a nuissance) or does it 
> indicate some issue the PCI subsystem is having in powering the card up 
> and communicating with it.   In either case, I'd be more than happy with 
> providing anyone able to patch the pci code with information i have on 
> the card. 
> 
> If it's nothing but a harmless warning, i'll forward the response to the 
> madwifi wiki and mailing list, so something can be done upstream to the 
> hal module to work the card. 
> 
> Thanks in advance. 

The power management spec version reported by the driver doesn't exist as
far as I know.  Would it be possible to see the output of "lspci -xxx"?

Even if the kernel isn't able to understand the PM registers, we may want
to assume the device is in D0 and print a warning. The bar restore changes
possibly have changed this behavior.  As a result, pci_enable_device() would
return an error.

Thanks,
Adam