On 9 April 2012 23:48, John Stultz john.stultz@linaro.org wrote:

On 04/09/2012 09:53 AM, Vishal Bhoj wrote:

...

Hi John,

I am having problems while merging the android 3.3 patches into a 3.4 rc1 kernel . Around 25 patches are failing . I wanted to know if someone has generated these androidization patches for 3.4 kernel :) .

I'm working on it. Hopefully will have something out today.

I could boot the upstream kernel to android GUI by enabling upstream

...

android drivers on 3.4 RC1 kernel.What I want to know here is what are the advantages of having all the android patches enabled.

So critically there isn't a whole lot. There's a few expected Android environment items, like ADB, apanic, the fiq_debugger, and paranoid networking. Then there are a number of smaller fixes for edge cases in areas like mmc, bluetooth, cgroup or ARM architecture items. Then its just further efficiency changes for the Android environment, like the interactive cpu gov and wakelocks. Finally there's the ION framework.

So yea, there's less criticality in using all of the android patchs if you're just trying to get a demo going, but its probably a good idea to use them w/ the fastmodel just so you can see how some of the more subtle power/performance tweaks interact, as well as getting real graphics integration going with ION.

Thanks for this info. Will give a try with your androidization branch for 3.4 rc1.

thanks -john