Samsung Ics Kernel Sources are out!
- Thread starter iamnotkurtcobain
- Start date
This can also be used for i9000 or i9001 roms ? i mean it's Samsung after all, they didn't mention i9100 specifically.
I see a lot of people will be disapointed...Don't think all the CM9 problems with be ironed out with the Kernel source Code....
Sorry to be that guy...
Sorry to be that guy...
Goooood！！！
official open source？？？any Korean version drivers?? like NFC?? DMC?? and compass?
official open source？？？any Korean version drivers?? like NFC?? DMC?? and compass?
No, not everything will be fixed with kernel source... But having a hacked pre-compiled kernel patched to work with CM9 doesn't really do the build justice. So it can fix some of the issues and help stabilize the release.
No, it can't - and what makes you think they didn't mention I9100 specifically? I9100 is right there in the file name!
First we have to fix the kernel sources itself.
Samsungs 3.0.15 is not Linux 3.0.15.
Have fun pulling mainline patches. ;-)
They renamed mach-exynos4 to mach-exynos and added a bunch of exynos5 stuff.
Which is the future. Mainline did that from 3.1 to 3.2 (The rename, not the new exynos5 stuff - the rename was likely to accomodate putting exynos5 stuff in there just like OMAP2/3/4 are all lumped under mach-omap2 and even non-OMAP devices like AM33xx are slated to be put into the mach-omap2 tree)
We're seeing classic evidence of the delay for mainlining anything Android-specific (and mainlining stuff from arch/arm in general) into the kernel tree.
It would suck less if Samsung maintained public-facing git repos like TI does - but kgene's linux-samsung tree is kinda "meh" - mainline cpuidle support for Exynos is **** for example.
Last edited:
Similar threads
Top Liked Posts
-
There are no posts matching your filters.
-
85
First we have to fix the kernel sources itself.
Samsungs 3.0.15 is not Linux 3.0.15.
Have fun pulling mainline patches. ;-)
They renamed mach-exynos4 to mach-exynos and added a bunch of exynos5 stuff.3
Which is the future. Mainline did that from 3.1 to 3.2 (The rename, not the new exynos5 stuff - the rename was likely to accomodate putting exynos5 stuff in there just like OMAP2/3/4 are all lumped under mach-omap2 and even non-OMAP devices like AM33xx are slated to be put into the mach-omap2 tree)
We're seeing classic evidence of the delay for mainlining anything Android-specific (and mainlining stuff from arch/arm in general) into the kernel tree.
It would suck less if Samsung maintained public-facing git repos like TI does - but kgene's linux-samsung tree is kinda "meh" - mainline cpuidle support for Exynos is **** for example.2
Yup, removing that mali_r2p3 crap was one of the first fixes I had to do.
Second was removing a hardcoded forced CROSS_COMPILE override - due to a typo in the top-level Makefile (= instead of ?=), CROSS_COMPILE is forced to be Samsung's location regardless of whether you've set it as an environment variable.
It's pretty common for Sammy to leave in "cruft" that needs cleanup before getting a buildable kernel.
Also, building without USE_SEC_FIPS_WHATEVERITIS (see build_kernel.sh) and u1_defconfig will fail.
I managed to build a bootable kernel, then tried to integrate CWM, and since then (even after reverting the initramfs changes for CWM) I haven't been able to boot. Unfortunately you Euros had a 6-7 hour headstart on me.
I'm buying the resistors for a UART cable tonight (have everything else).2
We know for a fact that there are bugs that Samsung is too clueless to notice/fix... But they are also too clueless to pull the fixes from the community.
See the fuel_alerted wakelock bug for example... Fixed for around 6 months by the community, still broken in Samsung kernels.
