Comment 14 for bug 1183065

So much for my theory in comment #1. Apparently the Wi-Fi power management on Android is much better than I expected.

Last night I activated Wi-Fi on my Nexus Galaxy (maguro) running JB last night, and with all other apps killed and on 2G, I barely saw a dent in the battery capacity. At the end of ~24 hours, it'd lost only ~30%.

Also, I missed the part in the initial description of disabling the AP ( or moving outside the AP's effective range ), so I did re-associated this morning, and then disabled from the AP. After two hours, this doesn't really seem to be having any more significant effect on battery usage either.

I've rebuilt a version of CM10.1 and will do some testing with wpa_supplicant's debug logging enabled ( and verbose ).

We've discussed that Android may be using device-specific driver libraries which may in turn use private ioctls. Hopefully I can verify use of these libraries via wpa_supplicant debug logging.