today is "noise day" :) as i mentioned in an earlier mail, i did some noise tracing in madwifi and i compared the noise calculations we do in ath5k with what ath_hal_get_channel_noise() returns. the register value used for comparison was the last one read before ath_hal_get_channel_noise.
as a reminder, here are the calculations we do: AR5K_PHY_NF_M 0x000001ff AR5K_PHY_NF_RVAL(_n) (((_n) >> 19) & AR5K_PHY_NF_M) AR5K_PHY_NF_AVAL(_n) (-((_n) ^ AR5K_PHY_NF_M) + 1) i got only two different readings with madwifi so far, calculations step by step: 0x0d88fe00 1101100010001111111000000000 >>19 110110001 0x1B1 ^1ff 0x4E 78 +1 79 ath_hal_get_channel_noise: -91 0x0d19ce00 1101000110011100111000000000 >>19 110100011 0x1A3 ^1ff 0x5C 92 +1 93 ath_hal_get_channel_noise: -96 0x0d11ce00 1101000100011100111000000000 >>19 110100010 0x1A2 ^1ff 0x5D 93 +1 94 ath_hal_get_channel_noise: -95 0x0d09ce00 1101000010011100111000000000 >>19 110100001 0x1A1 ^1ff 0x5E 94 +1 95 ath_hal_get_channel_noise: -96 so in both cases what we calculate is different from what ath_hal_get_channel_noise() returns and i don't see a pattern. i know the HAL can do, and probably does >;> all sort of ugly things like moving averages to make life harder for us! anyhow maybe somebody else can see a pattern? i have a log up at http://br1.einfach.org/ath/noise-readings6.txt where the noise is between -96 and -95 if anyone cares to have a look. now, if i only knew how to create some background noise, to get more samples... microwave ovens? bruno _______________________________________________ ath5k-devel mailing list ath5k-devel@lists.ath5k.org https://lists.ath5k.org/mailman/listinfo/ath5k-devel
