On Wed, Jan 6, 2010 at 11:55 PM, Rafał Miłecki <zaj...@gmail.com> wrote:
> 2010/1/6 Rafał Miłecki <zaj...@gmail.com>:
>> I've decided to made LP-PHY init code review for compatibility with specs
>> and found some issues.
>>
>> We lack some code for higher PHY revisions (up from 2/3) and for some
>> specific boards. Also I've found little mistake in specs which Larry just
>> corrected.
>>
>> Gábor: could you care for this parts of LP-PHY code?
>>
>> So far I've only checked
>> lpphy_read_band_sprom(dev);
>> and
>> lpphy_baseband_init(dev);
>>
>> I check for all conditions and operations (registers, masks, values). So far
>> didn't notice anything more serious than what's exposed by patch, so good
>> work Gábor :)
>
> There was actually one part I wasn't able to understand yesterday (it
> was late) and it still looks weird for me today... Could you verify my
> doubts, please?
>
> Specs:
> http://bcm-v4.sipsolutions.net/802.11/PHY/LP/ReadBandSrom
>
> Instead of "tmp2" we use "ofdmpo" (that's totally fine). However we do
> following read: "ofdmpo = bus->sprom.ofdm2gpo;" before "2 -> k"
> condition visible in specs.
>
> So we use sprom.ofdm2gpo for both conditions and we do not care for "opo".
>
> That is fine for rev. 8 as both: ofdm2gpo and opo are 0x142:
> http://bcm-v4.sipsolutions.net/SPROM
>
> However if we would meet LP-PHY on SSB with lower SPROM we wouldn't
> have ofdm2gpo and we should really use "opo" then.
>
> Am I right? Should we add "opo" reading for SPROM? Or maybe it's
> impossible to find LP-PHY on SSB with lower SPROM and we can keep
> using this work around?

It has been determined that the presence of "opo" in the LP-PHY part
of the code is Broadcrap resulting from the automated processes
Broadcom uses for generating their driver. There is no LP-PHY with
SPROM <v8.

>
> --
> Rafał
>



-- 
Vista: [V]iruses, [I]ntruders, [S]pyware, [T]rojans and [A]dware. :-)
_______________________________________________
Bcm43xx-dev mailing list
Bcm43xx-dev@lists.berlios.de
https://lists.berlios.de/mailman/listinfo/bcm43xx-dev

Reply via email to