[RFC] b43: N-PHY fixes - OK to cite wl MMIO traces in comments?

[Luca Casati]

Hi,

I have been getting b43 to work properly on a BCM43228 (14e4:4359,
core rev 30, N-PHY rev 16, radio 2057 rev 9) and have a small series
of N-PHY bug fixes I would like to submit:

  - TX tone sample table loaded as all zeros (operator precedence and
    a double fixed-point conversion before cordic_calc_iq())
  - TX IQ/LO calibration: inverted busy poll, result copied the wrong
    way, missing rev7+ cleanup
  - rev3+ RSSI calibrati is written instead
    of the selected VCM, plus a wrong distance term and rounding
  - 2057 RC calibration:re it starts
  - a missing field in the rev7 TX_PU override (compared to brcmsmac)

Where brcmsmac has the same logic I used it as the reference. Where it
doesn't cover this radior accesses against
an mmiotrace of the proprietary wl driver (6.30.223.271) running on my
own hardware. Only the M did not
disassemble the driver.

Some comments in the patches mention this, for example:

  /* Values from an mmiotrace of wl 6.30.223.271 on radio 2057 rev 9 */

b43 already has similar comments ("Extracted from MMIO dump of
6.30.223.141"), but I ware comments
naming wl and the trace acceptable, or would you prefer the values to
be described differentlyat all)?

Later I would also like ues (PAD tuning,
aux ADC midpoints) taken from the same traces, so the answer affects
those too.

The patches were developg assistant. I
will mark that in the patches according to the kernel's documentation.

Thanks,
Luca Casati