Hi, Qingzhan Thanks for your kind reply. And if you are also working on it or interestred in it, I'd like to exchange and communicate more with you ^_^
That's for certainty that there are great difference between 2.6.24's sys and 2.6.25's especially kobject/kset and related APIs, one example is kobject_create_and add() in 2.6.25 while it need several APIs to implemente it in 2.6.24. What I do is take the /sys/kernel in 2.6.24 as reference and port the modification of kobject.c of 2.6.25 (kobj_attribute, sysfs_ops, dynamic_kobj_ktype...) to driver/android/power.c in 2.6.24 then register the /sys/android_power subsystem. I'm still checking the code and behaviors but not clear the root cause yet ... BR/Mark On Thu, Jan 15, 2009 at 1:15 PM, Qingzhan Zhang <[email protected]> wrote: > I think you should your kernel ,there are some difference between 2.6.24's > sysfs and 2.6.25's. > > On Thu, Jan 15, 2009 at 5:36 PM, Android Groups <[email protected]>wrote: > >> >> Hi, all >> >> I'm porting Android 1.0 (2.6.25 based) to Marvell PXA9xx board with >> 2.6.24 kernel. Considering the workloads of drivers porting, I degrade >> the android to 2.6.24. At present, Many necessary features (ASHMEM, >> BINDER...) and drivers have been ported successfully and Android >> homescreen and applications can be displayed. But when I enabled the >> ANDROID_POWER( of course, some modification also have been performed >> accroding to the kobject/kset difference between 2.6.24 and 2.6.25 to >> register subsystem 'android_power' in /sys), the system stopped at >> kernel compression done like below: >> >> ------------------------------------------------------------------------------------ >> Uncompressing >> Linux............................................................. >> >> ................................................................................ >> ................ done, booting the kernel. >> >> ------------------------------------------------------------------------------------- >> >> I have checked the kernel source code carefully and thought that >> ANDROID_POWER shouldn't be the reason of system boot failure but it >> did. No other choices but turn to help in developer forum. Doesn't >> anybody account with the similar problems? and could anyone help me >> out? >> >> Thanks a lot! >> >> BR/Mark >> >> --~--~---------~--~----~------------~-------~--~----~ unsubscribe: [email protected] website: http://groups.google.com/group/android-porting -~----------~----~----~----~------~----~------~--~---
