2009/4/1 bitbucket <bitbuc...@yandex.ru>: > 01.04.09, 17:14, "Nick Kossifidis" <mickfl...@gmail.com>: > >> > /* No correction */ >> > - if (i_coffd == 0 || q_coffd == 0) >> > + if (i_coffd == 0 || q_coffd < 2) >> > goto done; >> > >> Are you sure that < 2 is ok ? > Yes. In Sam's hal (madwifi-free) they checked against 0 and >=2 . > >> Give me some time to look this through, i think i have one more fix >> for i/q calibration that adds some more accuracy... > I think, that it needs to more carefully check values of > MEAS_I,MEAS_Q,CORR_MEAS, > that read from card, because, I think, first nonzero value may be not perfect > value for calibration, IMHO. >
I think your patch can go in for now since it fixes i/q calibration, just please also rename 2191 #define AR5K_PHY_IQ_CORR_Q_Q_COFF 0x0000001f /* Mask for q correction info */ 2192 #define AR5K_PHY_IQ_CORR_Q_I_COFF 0x000007e0 /* Mask for i correction info */ 2193 #define AR5K_PHY_IQ_CORR_Q_I_COFF_S 5 to 2191 #define AR5K_PHY_IQ_CORR_Q_COFF 0x0000001f /* Mask for q correction info */ 2192 #define AR5K_PHY_IQ_CORR_I_COFF 0x000007e0 /* Mask for i correction info */ 2193 #define AR5K_PHY_IQ_CORR_I_COFF_S 5 and add some comments like Sam does so that we know what number types we put on hw. -- GPG ID: 0xD21DB2DB As you read this post global entropy rises. Have Fun ;-) Nick _______________________________________________ ath5k-devel mailing list ath5k-devel@lists.ath5k.org https://lists.ath5k.org/mailman/listinfo/ath5k-devel