[ltp] T40p madwifi and acpi troubles

[Gilles Magnier]

Rob Browning wrote:
> Gilles Magnier <gilles@pulse-powah.net> writes:
> 
> 
>>i plan to make a bug repport to madwifi-users ML but archives of this
>>ML seems down at the moment and it may be a good idea to known
>>if i'm the only one having this problem.
>>
>>So, Have anyone successfully used madwifi drivers under 2.6 with acpi ?
>>if yes with whitch bios/EC version ?
> 
> 
> For what it's worth, I can only sucessfully swsuspend (echo 4 >
> /proc/acpi/sleep) and resume if I have my suspend script shutdown ath0
> and rmmod ath_pci before the suspend.  Otherwise it hangs.  If I
> recall correctly, the symptoms were similar to those you described.
> 

First thanks for the reply,

until now i did not try suspend, just booting and bringing up ath0 give me
this error and 99% packet loss.

testing a bit more makes me sure it's an acpi problem and not a bios 
selftunig
cpu speed (hopefully).

I've noticed that when processor is in C2 state madwifi drivers works 
well but
when in C3 it fails as said before, im currently looking at processor.c 
trying to force
it not using C3 state.

So removing processor.ko module (ie booting and not "modprobing" it) makes
everything working fine, including both echo 4 and echo 3 to 
/proc/acpi/sleep
and restore (dri disabled, usb module unloaded).

Gilles.