On Wed, May 20, 2009 at 05:29:38PM +0800, forr...@hifulltech wrote:
> Dear Bob,
>
> Another patch of ath5k FYI.
Hi Forrest,
Please include the patch in the email body instead of as an attachment,
it makes it easier to reply to. See Documentation/SubmittingPatches.
> diff -Naur linux-2.6.30-rc6/drivers/net/wireless/ath5k/eeprom.c
> linux-2.6.30-rc6/drivers/net/wireless/ath5k/eeprom.c
Please also add -p to diff...
> --- linux-2.6.30-rc6/drivers/net/wireless/ath5k/eeprom.c 2009-05-16
> 12:12:57.000000000 +0800
> +++ linux-2.6.30-rc6/drivers/net/wireless/ath5k/eeprom.c 2009-05-20
> 17:17:19.000000000 +0800
> @@ -415,6 +415,8 @@
> if (ee->ee_version < AR5K_EEPROM_VERSION_5_0)
> return 0;
>
> + --o;
> + AR5K_EEPROM_READ(o++, val);
> switch (mode){
> case AR5K_EEPROM_MODE_11A:
> ee->ee_switch_settling_turbo[mode] = (val >> 6) & 0x7f;
There wasn't a description of the patch so I couldn't spot the bug
immediately. The issue is we don't initialize val before using it (hmm,
why isn't that a compiler error?). Also the shift is wrong according to
legacy hal (here eeval is other part of ee_cck_ofdm_gain_delta):
if (ee->ee_version >= AR_EEPROM_VER5_0) {
ee->ee_switchSettlingTurbo[1] =
(eeval >> 8) & 0x7f;
ee->ee_txrxAttenTurbo[1] =
(eeval >> 15) & 0x1;
Luckily, this is all for turbo modes which we don't use, but it should
be fixed. Thanks for pointing it out!
--
Bob Copeland %% www.bobcopeland.com
_______________________________________________
ath5k-devel mailing list
ath5k-devel@lists.ath5k.org
https://lists.ath5k.org/mailman/listinfo/ath5k-devel
