Re: [ath5k-devel] [PATCH 1/2] ath5k: fix wrong per rate target power eeprom reads for AR5K_EEPROM_MODE_11A

[Thomas Huehn <thomas@net.t-labs.tu-berlin.de>]

Hi Nick,

> There is nothing suspicious about it, it's what EEPROM docs say:
> 
> a) For 11a mode there are 10 calibration peers 


As I tested AR5413, AR5414 and AR5213 there are only 8 eeprom lines for
802.11a that provide valid per rate target power levels. I am curious if
this also holds for other chips.

> Also what do you mean wrong curve and how did you actually test your
> card and measured the tx power ?
> 


My first patch only changed the read iterations in function
ath5k_eeprom_read_target_rate_pwr_info() from 10 reads to 8. Having only
this, I measured received rssi values while changing the tx_power on a
sender. I did not observe any changes in rssi, so I went back to the
code to investigate.
And the function ath5k_eeprom_init_11a_pcal_freq() (my hardware eeprom
version is > 3.3) there is ath5k_eeprom_read_freq_list(..,10,..) called,
and those 10 iterations produce 2 wrong frequency peer values. If there
are only 8 iterations used, I am able to measure a 15 dB variation in
rssi values at the receiver side, while changing the tx_power on the
sender. For the chips I tested, ath5k_eeprom_read_freq_list() should
only read 8 pears for 802.11a.

With 'suspicious' I meant e.g. function
ath5k_eeprom_init_11a_pcal_freq(), as it checks the eeprom version and
if > 3.3 it calls
ath5k_eeprom_init_11a_pcal_freq(..,AR5K_EEPROM_N_5GHZ_CHAN,..), in the
else case, it reads 10 hardcoded freq.pears .. I am questioning if this
is correct, so maybe someone has the chance to test this.


Greetings Thomas