2008/6/11 Geoffrey McRae <[EMAIL PROTECTED]>:
> Hi everyone,
>

Hello and thanks for the feedback ;-)

> I am new to hardware hacking and I am looking to get the AR2454 device
> working in ath5k, however, due to my lack of understanding of the terms
> used (ie, what is a PHY?), I am finding it very hard to reverse engineer
> from trace dumps.
>

I guess you refer to 2425 or 5424, there is no 2454 device. Can you
please post the output of ath_info ?

> I have managed to add mmiotrace to a x86_64 2.6.25 kernel, and used the
> latest hal from Sam L. to get madwifi working with this device. I have
> also captured a complete trace and authentication with a WPA protected
> gateway, and altered the decode perl script to work with the new output
> format of mmiotrace.
>

Can you please share the output with the rest of us ? We NEED traces ;-)

> I have also produced a dump from the latest ath5k, and I have noticed a
> few things that seem wrong already...
>
> - The AR5K_PCICFG sleep register mask seems to be incorrect... ath5k is
> checking against 0x1000 where my trace shows checking against 0x10000. I
> have played around with a copy of ath5k and modified it to work this
> way, and it seems to be correct (for ar2454)
>

Yup i've also noticed that on other chips too but doesn't seem to be
the problem, card seems to wake up fine, i'll recheck it (for now it
works for me).

> - In my working trace, the value of AR5K_PHY_PLL should be set to 0xeb,
> which is not the case for ath5k.
>

You did an ath5k trace and it was different ?

It should be 0xeb, check out nic_wakeup, you should get for 2Ghz:
0x000000ab (AR5K_PHY_PLL_44MHZ_5212) | 0x00000040
(AR5K_PHY_PLL_RF5112) = 0x000000eb

What value do you get ?

BTW what snapshot of ath5k are you using ? The one that's currently on
compat-wireless and wireless-testing doesn't identify 2425 correctly
so you might get a different value (0x00000019) is that the case ?

> - When madwifi does a radio init (i think thats what it does), it is
> setting up PHY(17), but ath5k does not do this. I am not sure if this is
> a problem as I dont know what a PHY is.
>

It would be much easier if you shared the trace.
PHY(17) is settling time register, we set it up during reset (maybe
not in the same order).

> If anyone could direct me to some reading material, or send me some
> info, it would be great as I would love to learn how to do this so I can
> add the support myself.
>
> -Geoff
>

Start by checking out the various chips we are working on...
http://madwifi.org/wiki/Chipsets

For the rest check out the code, hw.c is mostly related to the MAC
chip, phy.c is related to the PHY chip and initvals.c contains the
initial settings for each chip combination (the static values we have
to write on each reset).

A PHY chip btw is the chip that deals with the PHYsical layer (in our
case the radio chip) ;-)

-- 
GPG ID: 0xD21DB2DB
As you read this post global entropy rises. Have Fun ;-)
Nick
_______________________________________________
ath5k-devel mailing list
ath5k-devel@lists.ath5k.org
https://lists.ath5k.org/mailman/listinfo/ath5k-devel

Reply via email to