[ath9k-devel] [RFC] ath9k: use correct init values for ar9100 devices

[Gabor Juhos <juhosg@openwrt.org>]

Sujith ?rta:
> Gabor Juhos wrote:
>> Well, I was curious, so I did it. :)
>>
>> After some cut&paste job, here is the diff between the original and the new
>> initvals from your patch:
>>
>> diff --git a/drivers/net/wireless/ath9k/initvals.h
>> b/drivers/net/wireless/ath9k/initvals.h
>> index d492363..d5a8aec 100644
>> --- a/drivers/net/wireless/ath9k/initvals.h
>> +++ b/drivers/net/wireless/ath9k/initvals.h
>> @@ -658,7 +657,7 @@ static const u32 ar5416Addac_9100[][2] = {
>>      {0x0000989c,  0x00000000 },
>>      {0x0000989c,  0x00000000 },
>>      {0x0000989c,  0x00000000 },
>> -    {0x000098cc,  0x00000000 },
>> +    {0x000098c4,  0x00000000 },
>>  };
>>
>>  static const u32 ar5416Modes[][6] = {
>> @@ -862,7 +861,7 @@ static const u32 ar5416Common[][2] = {
>>      { 0x00008134, 0x00000000 },
>>      { 0x00008138, 0x00000000 },
>>      { 0x0000813c, 0x00000000 },
>> -    { 0x00008144, 0x00000000 },
>> +    { 0x00008144, 0xffffffff },
>>      { 0x00008168, 0x00000000 },
>>      { 0x0000816c, 0x00000000 },
>>      { 0x00008170, 0x32143320 },
>>
> 
> Heh. :-)
> Dunno if those 2 values make any difference.

Indeed, these 2 values does not make too much difference.

The first one reverts this commit:
http://git.kernel.org/?p=linux/kernel/git/linville/wireless-testing.git;a=commitdiff;h=6753ddb4bebecab1039a2a2c8d51d6656d0d0513

With the second one we will use the same initial value (which is in the
ar5416Common_9100, ar5416Common_9160  ar5416Common_9280_2 and
ar5416Common_9285_1_2 arrays currently) for the AR_PHY_ERR_EIFS_MASK register.

Hm, the ar5416Common_9280 and ar5416Common_9285 arrays are using 0x00000000 for
the same register. By any chance, these values would be wrong in them also?

> 
>>>>> +#define AR_SREV_VERSION_9100		0x014
>>>> 0x14?
>>> Yep, this can be fixed.
>> I have to send a new patch, or you will take care of it?
> 
> Feel free to send a patch fixing this, and the macro cleanup.
> I'll just update the initvals, fix the check for AR9100 in HW attach 

Well, the _AR9100_OR_LATER would be the right one, simply the macro itself is
wrong. According to Luis' request, I would send a new patchset, then we can
discuss it further based on that.

> and the typo for AR9285 chipsets.

This one can go with a separated patch?

Regards,
Gabor