2012/5/1 Colin Guthrie <[email protected]>: > 'Twas brillig, and simple w8 at 01/05/12 14:38 did gyre and gimble: >>> > Anyway, checking commit logs between upstream 3.3.1 and 3.3.3 >>> > shows one commit that specifically touches the module you use... >>> > >>> > So can you try kernel-3.3.4-1.1.mga2 from: >>> > http://tmb.mine.nu/Mageia/Cauldron/bugs/3.3.4-rtlwifi-revert/ >>> > http://tmb2.mine.nu/Mageia/Cauldron/bugs/3.3.4-rtlwifi-revert/ >>> > >>> > If that works, you can ping upstream that: >>> > http://git.kernel.org/?p=linux/kernel/git/stable/linux-stable.git;a=commit;h=eb641120479de7ee703fabb5e6ee0ec7947474ba >>> > >>> > broke your wireless. >> In fact with this kernel the machine is able to detect the same wifi >> networks that sued to detect with kernel-3.3.1... > > Well that means that the commit listed above is problematic. > > You should tell the upstream folks about this as Thomas stated. Please > confirm when you've done this so the issue can be tracked accordingly.
I have been discussing this with Larry Finger (one of the realtek main developers) in "wireless <[email protected]>" ML and this Larry Finger what he said: "That patch keeps your device from running forever at a maximum of 1 Mbps by adjusting the initial gain of the RX amplifier. You may reverse it if you want so that you can detect more APs when you scan; however, most people prefer to get better performance from the device. That patch will not be reversed."
