Related
This thread is NOT for general Q & A about using the kernel or configuring a specific version.
This thread IS for sharing source code or attempting to apply debugging to code changes.
In English: This is what a development thread really looks like. No downloads, No user manual.
Note to Mods: OT is welcome, so long as it is relevant to the kernel developer experience
Building the kernel 101:
A clean branch with only the build tools, ramdisks, and MacOS updates is at:
https://github.com/StarKissed/starkissed-kernel-trlte/tree/master
The stockish branch adds Linux patches, mainline updates, and SELinux mods
The deported branch is the customized version described in the OP feature list
You are going to need a few things, especially if building this kernel for MacOS.
GCC 4.7 (This is the version Samsung's documentation listed)
An updated mkbootimg (Must have DTS support)
A dtbtool (Included in the source, but only for Linux)
Insomnia (Not a requirement, but it will help you in the long run)
For MacOS, the tools are available from the following repositories:
https://github.com/StarKissed/android-toolchain-eabi-4.7
https://github.com/StarKissed/starkissed-build-tools
Let's jump ahead a moment for those of you who have built kernels and notice your boot.img is only around 11 MB while the versions you see listed here are all roughly 14 MB. This is where the new mkbootimg and dtbtool are necessary.
Attempting to retrieve the commandline arguments from the boot.img will not tell you one important piece:
Code:
./dtbtool -o dt.img -s 4096 -p ../scripts/dtc/ ../arch/arm/boot/dts/
chmod a+r dt.img
mkbootfs (ramdisk creation)
mkbootimg (mkbootimg params) --dt dt.img
This takes the DTSI that was built into a dtb as one of the first steps in the compile process and turns it into an image that is appended to the end of the boot.img. The dt.img just happens to be about 3 MB. (11 MB + 3 MB = 14 MB)
If you grabbed the original Samsung kernel source, you may notice that their default compile script directs the build to an "out" folder in the root of the kernel directory.
The obvious advantage to this is that their clean command runs into some errors, so the clean process is simply deleting the "out" folder and making a new one.
This doesn't work for a lot of the custom components you may want to add to your kernel, though. Some modifications, such as intelli-thermal, rely on a subdirectory that is referenced in the Makefile.
Building to this out folder will result in an error because it cannot find the output of that component, but building in the tree, or actual source, will also result in an error. The lesser of two evils is to fix in-tree builds.
Code:
Instructions for TRACE_INCLUDE_FILE in include/trace/define_trace.h say
"the path is relative to define_trace.h, not the file including it".
Fix in-tree compilation by making the path relative to define_trace.h.
What this means is that when you have a line that refers to the TRACE_INCLUDE_PATH or an include reference to a file that uses a relative location instead of a global, these need to be updated
The only two files that currently require updates in the stock source code are
Code:
drivers/scsi/ufs/ufs_test.c
#include -> #include "../sd.h"
and
Code:
drivers/video/msm/mdss/mdss_mdp_trace.h
#define TRACE_INCLUDE_PATH . -> #define TRACE_INCLUDE_PATH ../../drivers/video/msm/mdss/
If you want to be able to debug your kernel, you will need to enable last_kmsg.
Open up the arch/arm/configs/apq8084_sec_defconfig and add
Code:
CONFIG_SEC_DEBUG_NOCACHE_LOG_IN_LEVEL_LOW=y
CONFIG_PRINTK_PROCESS=y
CONFIG_SEC_DEBUG_LOW_LOG=y
More to come soon...
This is a great idea!
Looking fwd to seeing this thread take off...it will be very interesting for sure.
Sent from my SM-N910W8 using Tapatalk 2
@ktoonsez im back ready to play my man
@twistedumbrella my man i hope i dont become a ball buster on this thread but i have the 1st 2 questions on this thread:
1. How do you patch your kernels? (i use the p1 command.. is there another way?)
2. How did you manage to make the OC values show?
Disclaimer: I may have a lot of follow up questions
friedrich420 said:
@ktoonsez im back ready to play my man
@twistedumbrella my man i hope i dont become a ball buster on this thread but i have the 1st 2 questions on this thread:
1. How do you patch your kernels? (i use the p1 command.. is there another way?)
2. How did you manage to make the OC values show?
Disclaimer: I may have a lot of follow up questions
Click to expand...
Click to collapse
I use git am -3 so that any issue defaults to a merge then git mergetool to resolve it.
Switching the index in the dtsi that was modified was the only change to show it that I remember needing.
https://github.com/StarKissed/stark...mmit/695abe187620acf513b33237042925d60a99f0c1
friedrich420 said:
@ktoonsez im back ready to play my man
@twistedumbrella my man i hope i dont become a ball buster on this thread but i have the 1st 2 questions on this thread:
1. How do you patch your kernels? (i use the p1 command.. is there another way?)
2. How did you manage to make the OC values show?
Disclaimer: I may have a lot of follow up questions
Click to expand...
Click to collapse
Im back too.
ktoonsez said:
Im back too.
Click to expand...
Click to collapse
right on
any thoughts my man?
friedrich420 said:
right on
any thoughts my man?
Click to expand...
Click to collapse
Im in the process of just starting over from the "F" source and looking for testers on INTL thread. I wish I didnt have to start over but looks like Im going to have to
ktoonsez said:
Im in the process of just starting over from the "F" source and looking for testers on INTL thread. I wish I didnt have to start over but looks like Im going to have to
Click to expand...
Click to collapse
we know for sure that your source without bfq works fine... now its a matter of figuring out what of all the additions from yesterday's source to today's source caused the bootloop..
friedrich420 said:
we know for sure that your source without bfq works fine... now its a matter of figuring out what of all the additions from yesterday's source to today's source caused the bootloop..
Click to expand...
Click to collapse
I wonder if its any of the top 4 commits from here:
https://github.com/ktoonsez/KTNOTE4/commits/tw4.4?page=2
ktoonsez said:
I wonder if its any of the top 4 commits from here:
https://github.com/ktoonsez/KTNOTE4/commits/tw4.4?page=2
Click to expand...
Click to collapse
give me a little bit of time and ill revert it manually and compile again the one i compiled in the afternoon (your morning)
I got lost in Freedom Wars for a bit. I need to figure out an issue with thermal then I'll be back
friedrich420 said:
give me a little bit of time and ill revert it manually and compile again the one i compiled in the afternoon (your morning)
Click to expand...
Click to collapse
Here is a test one starting all over again, just has OC and Voltage control:
http://ktoonsez.jonathanjsimon.com/note4/TW/KT-NOTE4-KK4.4-TW-N910-11.13.2014D.zip
ktoonsez said:
Here is a test one starting all over again, just has OC and Voltage control:
http://ktoonsez.jonathanjsimon.com/note4/TW/KT-NOTE4-KK4.4-TW-N910-11.13.2014D.zip
Click to expand...
Click to collapse
Unfortunately same bahavior my man..
friedrich420 said:
Unfortunately same bahavior my man..
Click to expand...
Click to collapse
LOL, really? I just had 3 people boot up successfully, finally. You have the "F" or the "G"?
EDIT:
The guy I am testing with has the "G".
ktoonsez said:
LOL, really? I just had 3 people boot up successfully, finally. You have the "F" or the "G"?
Click to expand...
Click to collapse
the F...
friedrich420 said:
the F...
Click to expand...
Click to collapse
Just edited my post above, the guy that is answering me quick has the "G". Now Im really getting PISSED!!!!!!!!!! WTH!!
ktoonsez said:
Just edited my post above, the guy that is answering me quick has the "G". Now Im really getting PISSED!!!!!!!!!! WTH!!
Click to expand...
Click to collapse
Did anyone diff the ramdisk? If not, I can throw the two in git and see if anything lights up. I have the visual editor for Mac for exactly that.
ktoonsez said:
Just edited my post above, the guy that is answering me quick has the "G". Now Im really getting PISSED!!!!!!!!!! WTH!!
Click to expand...
Click to collapse
this may or may not confuse you more...
i made a test git repository : https://github.com/friedrich420/Note4-AEL-Kernel2
all i did was cherry pick your oc and voltage control commits..
It booted im running it @2.99ghz right now and its fine...
Could it be that you need the F source my man?
twistedumbrella said:
Did anyone diff the ramdisk? If not, I can throw the two in git and see if anything lights up. I have the visual editor for Mac for exactly that.
Click to expand...
Click to collapse
I started with stock N910F ramdisk in version "C", version "D" went back to my original universal ramdisk you can try the "C" and see if anything changes:
http://ktoonsez.jonathanjsimon.com/note4/TW/KT-NOTE4-KK4.4-TW-N910-11.13.2014C.zip
friedrich420 said:
this may or may not confuse you more...
i made a test git repository : https://github.com/friedrich420/Note4-AEL-Kernel2
all i did was cherry pick your oc and voltage control commits..
It booted im running it @2.99ghz right now and its fine...
Could it be that you need the F source my man?
Click to expand...
Click to collapse
That is exactly what I just merged in so the "D" that wont boot for you is pretty much the same thing, lol. See what happens on the "C" version which is stock N910F ramdisk.
Hello,
This is a stock SeLinux permissive kernel and init.d activated for Samsung Galaxy Note Edge (N915F) for Lollipop 5.0.1.
There are inputs that is working also on N915FY.
The kernel was compiled from Samsung source LL_CIS kernel using Google toolcahin v4.8.
If you want to test it, please flash in cwm the attached file and report back. Please make a backup of your actual kernel (boot.img) befor flashing this one.
Special thanks to @Micky387 who advised me to update the source kernel (MH driverI), @Wuby986 for his support and guidance and @daxgirl for his advices.
My kernel source for N915F (N915FOXE1BOC6) is here.
Happy flashing.
PS. Kernel updated. Same functionality but some different info in "About".
Well you got further than I did .. Kept getting some driver file referenced so it would not complied.. Thanks for effort I hope you work it out soon.. Unfortunately I don't have time at the moment ..
starbucks2010 said:
Hello,
Since few days I struggle to build a SeLinux permissive kernel for Samsung Galaxy Note Edge with Lollipop 5.0.1 but without success. Or with partial success.
I followed several guides from xda and other sites but it seems that everyone has his different way of doing it.
What I have done so far:
- I installed Ubuntu 14.04 with all updates and apt-get xxx
- I installed the Google toolchain arm-eabi-4.8
- I Installed Samsung kernel source for N915F_CIS from here
- I changed the path to the toolchain in Makefile and build_kernel.sh
- I run from kernel source folder: ./build_kernel.sh
- After 10-15 min of compilation I get a zImage file in //output/arch/arm/boot of ~ 9.8 Mb
- Using Android Image Kitchen v2.2 I integrated the zImage into a boot.img to be flashed with cwm
I flashed this new boot.img but the phone is rebooting every minute and in about I have "unknown"as baseband.
Anyone have an idea how to fix this?
Thank you in advance.
Click to expand...
Click to collapse
hi man. How are you ?
So many thing are change in lollipop . the modem and firmware modem dont call same as KK ROM.
So be sure you are this commit https://github.com/CyanogenMod/andr...mmit/784aa9a84b77090fa591bff77045341c3e3c6293
and add MHI module into tblte_defconfig
https://github.com/CyanogenMod/andr...mmit/1dd44659ee26c2d8b313c1346192e2bfa2ac7bb9
dont delete 18 and 19 line if you build for TW rom .
Try and report
micky387 said:
hi man. How are you ?
So many thing are change in lollipop . the modem and firmware modem dont call same as KK ROM.
So be sure you are this commit https://github.com/CyanogenMod/andr...mmit/784aa9a84b77090fa591bff77045341c3e3c6293
and add MHI module into tblte_defconfig
https://github.com/CyanogenMod/andr...mmit/1dd44659ee26c2d8b313c1346192e2bfa2ac7bb9
dont delete 18 and 19 line if you build for TW rom .
Try and report
Click to expand...
Click to collapse
Bonjour Mickey,
Thank you for your advice. I'll try it and let you know how it goes. Do you know if I have to flash also the modules together with the new kernel or only the boot.img alone ?
Merci.
starbucks2010 said:
Bonjour Mickey,
Thank you for your advice. I'll try it and let you know how it goes. Do you know if I have to flash also the modules together with the new kernel or only the boot.img alone ?
Merci.
Click to expand...
Click to collapse
No need to flash module. The MHI module will be in the kernel so you have boot.img with all inside.
starbucks2010 said:
Hello,
Since few days I struggle to build a SeLinux permissive kernel for Samsung Galaxy Note Edge with Lollipop 5.0.1 but without success. Or with partial success.
I followed several guides from xda and other sites but it seems that everyone has his different way of doing it.
What I have done so far:
- I installed Ubuntu 14.04 with all updates and apt-get xxx
- I installed the Google toolchain arm-eabi-4.8
- I Installed Samsung kernel source for N915F_CIS from here
- I changed the path to the toolchain in Makefile and build_kernel.sh
- I run from kernel source folder: ./build_kernel.sh
- After 10-15 min of compilation I get a zImage file in //output/arch/arm/boot of ~ 9.8 Mb
- Using Android Image Kitchen v2.2 I integrated the zImage into a boot.img to be flashed with cwm
I flashed this new boot.img but the phone is rebooting every minute and in about I have "unknown"as baseband.
Anyone have an idea how to fix this?
Thank you in advance.
Click to expand...
Click to collapse
Wait one more day..
Or maybe earlier
Finally done.
Make backup of your actual kernel, flash the attached zip from post 1in cwm and please report back.I didn't check the init.d support but it should work.
Good, I'll try.
Dev.
Just a small request please.
If you could please make sure a link to YOUR kernel sources in OP.
Any Linux kernel publicly released must have source to compile the exact binary (zImage) as repacked into your distributed boot.img. This must be current at time of every released version.
This is not just a XDA rule, but a legal GPL requirement, applying to all, even if modification is only minor. This is open-source
I would suggest the use of a github repo that helps keep track of your changes (commits), promotes open transparency, sharing and colaboration
If you have any questions on how to go about this, i would be happy to help via pm (but currently away with work), but would suggest to contact our friendly forum Mod.
Best of luck with your project. :thumbup:
Cheers.
UITA
Any chance that you would be willing to make this kernel compatible with the Sprint variety? Man.... I need my viper4android bro. Please. Lol
UpInTheAir said:
Dev.
Just a small request please.
If you could please make sure a link to YOUR kernel sources in OP.
Any Linux kernel publicly released must have source to compile the exact binary (zImage) as repacked into your distributed boot.img. This must be current at time of every released version.
This is not just a XDA rule, but a legal GPL requirement, applying to all, even if modification is only minor. This is open-source
I would suggest the use of a github repo that helps keep track of your changes (commits), promotes open transparency, sharing and colaboration
If you have any questions on how to go about this, i would be happy to help via pm (but currently away with work), but would suggest to contact our friendly forum Mod.
Best of luck with your project. :thumbup:
Cheers.
UITA
Click to expand...
Click to collapse
Thank you for your advices.The link to Samsung kernel source was already there . Anyway, I added it one more time just to be sure that is more visible.
For the github repo, I have to create an account and then I'll upload the files.
Best regards.
Well done mate ..... Yea
---------- Post added at 08:32 AM ---------- Previous post was at 07:36 AM ----------
Boot and works fine , nice one
Can this be used on N915FY as well?
I don't have any reports from N915FY users. You can try and tell us
Don't forget to make a backup first.
starbucks2010 said:
Thank you for your advices.The link to Samsung kernel source was already there . Anyway, I added it one more time just to be sure that is more visible.
For the github repo, I have to create an account and then I'll upload the files.
Best regards.
Click to expand...
Click to collapse
Hey, no problem.
Just to clarify....
If you've made any (even minor) change to "Samsung" source, then you have to upload "your" complete source and link "your" source, not "Samsung" So if you've made the kernel permissive or even use a different toolchain or path, technically you have modified from original and is now "your" source If you use a custom defconfig, don't forget it too
github is well worth the pain I Google for answers all the time (how to do this and that), but still screw things up (sometimes majorly) now and then.
dubart said:
Can this be used on N915FY as well?
Click to expand...
Click to collapse
Using this kernel with a N915FY. Great job btw I can finally use viper audio which I've missed so much. Major props dude!
I'm gonna flash this on sprint variant either that or throw this device out my window
dinaps86 said:
I'm gonna flash this on sprint variant either that or throw this device out my window
Click to expand...
Click to collapse
Lol it booted just no data. Yaaayyy!!!
I see that there are other people who are interested in Selinux permissive kernels.
So, if you want, I can add selinux permissive to any samsung kernel as long as you provide me 2 files: your original kernel (boot.img) from the ROM you want to flash and the Kernel.tar.gz - source kernel from Samsung site. In order to get Kernel.tar.gz file you'll need to download the kernel source file (~ 1Gb) from Samsung site, open the archive and upload ONLY the Kernel.tar.gz (~ 100-150 Mb) and the readme file from the archive to some sharing site. Then, I'll try to compile that kernel for you.
There are also downsides. I cannot stay up to date will all new commits done to the kernel source so I'll compile your kernels only with info that I have at that date. If you fork the github and find kernel source updates, I can add them.
PS. I only set up a working PC with Ubuntu and Google toolchain v4.8. If you find any information that this setup is not suitable for compiling your kernel, then there is no point of providing me the above files.
Best regards.
OP updated with some Antutu screen shots.
So, after destroying a tremendous amount of data and a whole lot of wait, I finally synced the whole OmniROM 6.0 source code via repo. Now I want to compile it for my device XOLO BLACK (Codename: BLACK) so, can somebody tell me how I can and what am I gonna need (I mean files, like vendor files, kernel etc.)?
Also, how can I extract its device tree? Because it's not available online.
You will need the full kernel source code to even begin and everything I have found says it is not released. Without it you are out of luck.
zelendel said:
You will need the full kernel source code to even begin and everything I have found says it is not released. Without it you are out of luck.
Click to expand...
Click to collapse
No no, kernel is here: https://github.com/hiveinside/xolo_kernel
What do I need now?
NutVolt said:
No no, kernel is here: https://github.com/hiveinside/xolo_kernel
What do I need now?
Click to expand...
Click to collapse
You will need the device tree as well as working on the errors in the kernel as they come up and they will. Also you may have to fix any issues with the source as it only contains the open source parts. Check out some of the building tuts on the site.
It will not be a simple matter of adding the kernel and stuff and then building. It normally takes weeks of bug hunting and fixing to get it to even boot.
zelendel said:
You will need the device tree as well as working on the errors in the kernel as they come up and they will. Also you may have to fix any issues with the source as it only contains the open source parts. Check out some of the building tuts on the site.
It will not be a simple matter of adding the kernel and stuff and then building. It normally takes weeks of bug hunting and fixing to get it to even boot.
Click to expand...
Click to collapse
I know it's gonna take long but if I can't even start then I'll feel very bad, all the hard work I've done to just fetch the source.
Anyway, how can I make a device tree? Can you teach me?
The good stuff: https://github.com/jcadduono/android_kernel_lge_msm8996/commits/android-7.0
Fork it!
Update: CAF branch has some problems at the moment, use this one instead for now.
So this is currently the v10d release. It's compatible with V20 Nougat v10d and G5 v20a.
Place toolchain (pick one up from here) in ~/build/toolchain & tar xf.
git clone https://[email protected]/yourname/android_kernel_lge_msm8996.git
If you want to change any config options:
./menuconfig.sh (optional target, ex. twrp or lge)
Ready?
./build.sh h850/h830/h918/us996/any model you want
Models are applied onto the target base defconfig, so your V20 kernels will build for all V20 variants and for all variants of the G5 as well.
Create your own by copying arch/arm64/configs/lge_defconfig to arch/arm64/configs/your-awesome_defconfig!
Grab:
build/arch/arm64/boot/Image.lz4-dtb [or] build/arch/arm64/boot/Image.gz-dtb
build/lib/modules
Place in:
https://github.com/jcadduono/lazyflasher/tree/kernel-flasher
make
You got yourself a magical all-powerful custom kernel installer, complete with your own custom kernel!
Notice that there's a few branches:
opensource is where I merge LG changes, when it's updated, android-7.0 is rebased on it
android-7.0 is a starting off point closest to absolute stock, base off here
android-7.0-security adds security and bug fixes from CAF on top of android-7.0 branch
twrp-7.0 is for Team Win Recovery Project (yes, it's ready, come with an unlocked device and I'll give it to you)
nethunter-7.0 is for the Kali NetHunter project (if you dare, some fun there)
Another issue was their weird build system, I dropped that. Replaced tuxera's exFAT prebuilt downloader with Samsung's GPL'd exFAT drivers instead. Unsure how well it works replacing it like that, we'll have to see.
I also added Qualcomm's Core Control which they just released the sources to somewhat recently. I believe it's already included on the G5 and probably V20 as a kernel module. Their released version doesn't have module support yet but it's fairly easy to add I guess with the leaked proprietary module version out there to pick pieces from.
Have fun!
Great Man !
thanks for this..
I've been following the root process since the beginning and I'm amazed at the relatively short amount of time it took. Huge kudos to @jcadduono! I will purchase this phone. Now, since the kernel needed to have it's encryption support modifed, are the patches added to a branch on your Github fork? I would like to compile a custom kernel, along with patching it with SultanXDA's SafetyNet boot bypass. Which branch will be appropiate to fork to flash on the TWRP/Rooted V20?
Thanks much @jcadduono! I'm going to start poking around in here and see if I can find anything particularly interesting.
zabracks said:
Thanks much @jcadduono! I'm going to start poking around in here and see if I can find anything particularly interesting.
Click to expand...
Click to collapse
Hey guys so I did get a kernel booting but not yet one built from CAF sources so it'll take a bit of work to figure out why.
There seems to be a couple occurrences of null pointer dereferences in early boot stage in both stock release and this CAF. Currently investigating (or trying to anyway, it's hard to find people that can flash my test builds and report back with logs - I do not own this phone)
jcadduono said:
Hey guys so I did get a kernel booting but not yet one built from CAF sources so it'll take a bit of work to figure out why.
There seems to be a couple occurrences of null pointer dereferences in early boot stage in both stock release and this CAF. Currently investigating (or trying to anyway, it's hard to find people that can flash my test builds and report back with logs - I do not own this phone)
Click to expand...
Click to collapse
I've got an H918, if you need somebody to test a kernel.
jcadduono said:
Hey guys so I did get a kernel booting but not yet one built from CAF sources so it'll take a bit of work to figure out why.
There seems to be a couple occurrences of null pointer dereferences in early boot stage in both stock release and this CAF. Currently investigating (or trying to anyway, it's hard to find people that can flash my test builds and report back with logs - I do not own this phone)
Click to expand...
Click to collapse
two things. Did you figure out how to compile boot.img or are you flashing the zimage? also hit me up on hangouts I've been wanting to dive into this device just looking for someone else to colaberate with.
EDIT: sent email in PM
albinoman887 said:
two things. Did you figure out how to compile boot.img or are you flashing the zimage? also hit me up on hangouts I've been wanting to dive into this device just looking for someone else to colaberate with.
EDIT: sent email in PM
Click to expand...
Click to collapse
I can compile it and flash it. There seems to be an issue with the sources LG has provided regarding display drivers. It seems any attempts at a stock kernel build results in a messed up screen once you hit the lock screen. I assume this has something to do with 2D graphics firmware and kernel framebuffer driver badly interacting. These sources are for v10b firmware, while the device appears to ship with v10d. I don't know if LG made any changes since then that could have caused this issue.
FYI the sources up right now are not my CAF merge, I had to move it to (same url but with -caf on the end) - I still have some issues to work out with the CAF merged sources (they aren't booting for whatever reason - not even TWRP).
It could take me a while to work on this as I still don't and likely never will own this device due to my lack of trust when it comes to LG.
jcadduono said:
I can compile it and flash it. There seems to be an issue with the sources LG has provided regarding display drivers. It seems any attempts at a stock kernel build results in a messed up screen once you hit the lock screen. I assume this has something to do with 2D graphics firmware and kernel framebuffer driver badly interacting. These sources are for v10b firmware, while the device appears to ship with v10d. I don't know if LG made any changes since then that could have caused this issue.
FYI the sources up right now are not my CAF merge, I had to move it to (same url but with -caf on the end) - I still have some issues to work out with the CAF merged sources (they aren't booting for whatever reason - not even TWRP).
It could take me a while to work on this as I still don't and likely never will own this device due to my lack of trust when it comes to LG.
Click to expand...
Click to collapse
sounds like the kgsl drivers. If you want we can work together on this. I want to get CM going too i just need some help. Im and experienced dev but havent jumped in the LG pool before. id look at the google pixel kernel source and grab the kgsl/adreno crap and update that.
i'll clone the source and check it out since i'm finally getting the stock rom downloaded (thanks for the wget tip)
jcadduono said:
I can compile it and flash it. There seems to be an issue with the sources LG has provided regarding display drivers. It seems any attempts at a stock kernel build results in a messed up screen once you hit the lock screen. I assume this has something to do with 2D graphics firmware and kernel framebuffer driver badly interacting. These sources are for v10b firmware, while the device appears to ship with v10d. I don't know if LG made any changes since then that could have caused this issue.
FYI the sources up right now are not my CAF merge, I had to move it to (same url but with -caf on the end) - I still have some issues to work out with the CAF merged sources (they aren't booting for whatever reason - not even TWRP).
It could take me a while to work on this as I still don't and likely never will own this device due to my lack of trust when it comes to LG.
Click to expand...
Click to collapse
which branch/repo is the one that boots but messed up display?
albinoman887 said:
sounds like the kgsl drivers. If you want we can work together on this. I want to get CM going too i just need some help. Im and experienced dev but havent jumped in the LG pool before. id look at the google pixel kernel source and grab the kgsl/adreno crap and update that.
i'll clone the source and check it out since i'm finally getting the stock rom downloaded (thanks for the wget tip)
Click to expand...
Click to collapse
Would absolutely LOVE to see CM on our phones! Not that it'll make things move faster, but I'll throw 100 bucks your way if it's done before Christmas
lightninbug said:
Would absolutely LOVE to see CM on our phones! Not that it'll make things move faster, but I'll throw 100 bucks your way if it's done before Christmas
Click to expand...
Click to collapse
I want Nethunter too
I'll be around to test any of the stuff, usually at night.
The v20 kernel is available on the lg open-source site, at least the h990 one is.
Sent from my LG-H990 using Tapatalk
---------- Post added at 03:36 PM ---------- Previous post was at 03:25 PM ----------
Artcrime said:
The v20 kernel is available on the lg open-source site, at least the h990 one is.
Sent from my LG-H990 using Tapatalk
Click to expand...
Click to collapse
There is v10d version for 918. opensource.lge.com forgive me if this is old news. a|c
albinoman887 said:
sounds like the kgsl drivers. If you want we can work together on this. I want to get CM going too i just need some help. Im and experienced dev but havent jumped in the LG pool before. id look at the google pixel kernel source and grab the kgsl/adreno crap and update that.
i'll clone the source and check it out since i'm finally getting the stock rom downloaded (thanks for the wget tip)
Click to expand...
Click to collapse
slayerh4x said:
I want Nethunter too
I'll be around to test any of the stuff, usually at night.
Click to expand...
Click to collapse
albinoman887 said:
two things. Did you figure out how to compile boot.img or are you flashing the zimage? also hit me up on hangouts I've been wanting to dive into this device just looking for someone else to colaberate with.
EDIT: sent email in PM
Click to expand...
Click to collapse
Sent from my LG-H990 using Tapatalk
I updated all the branches earlier today...should work fine now, so apparently was a source code issue that needed v10d update, not sure though no one's volunteered for testing (I don't do requests/testing over the forum, only on IRC)
Is a kernel the main thing stopping us from a CM rom?
what toolchain should we be using? got a link to a compiled one? i'm getting build errors I know are related to the toolchain but i've been out of the game for awhile. last i knew we needed gcc-4.9?
jcadduono said:
I updated all the branches earlier today...should work fine now, so apparently was a source code issue that needed v10d update, not sure though no one's volunteered for testing (I don't do requests/testing over the forum, only on IRC)
Click to expand...
Click to collapse
I'm there.
albinoman887 said:
what toolchain should we be using? got a link to a compiled one? i'm getting build errors I know are related to the toolchain but i've been out of the game for awhile. last i knew we needed gcc-4.9?
Click to expand...
Click to collapse
gcc-linaro-5.3.1-2016.05-x86_64_aarch64-linux-gnu.tar.xz .Be sure to change the PATH in menuconfig.sh
@jcadduono , thank you. You are truly an inspiration
Wondering if anyone can help me learn how to do this right...
So I was able to compile the kernel, but I didn't get an Image.gz (which im sure is fine). I did get an Image.lz4, Image.lz4-dtb, and Image (no extension). I figure only the .lz4* files are needed. I added all 3 to the lazyflasher, and ran make. The resulting zip flashed, but sent me straight to Bootloader. I then tried to fix by flashing the stock kernel, but this still did not boot. ReFlashed my ROM and it's booting now.
I used this toolchain gcc-linaro-5.3.1-2016.05-x86_64_aarch64-linux-gnu
Any guidance? My guess is it's flashing at the wrong point?
Here is the ZIP. DO NOT FLASH. This is only for someone who knows what they are looking at to tell me what is wrong with it. No modifications were done. Source was forked from OP, and compiled with default menuconfig, then build.sh h918
So I have successfully built OmniRom 8.1 Oreo for Shamu the zip file is a little over 500mb. The package flashes correctly but fails to boot. I have already tried to change the BoardConfig.mk file to be SeLinux=Permissive but it still fails to boot. My device is decrypted as well. What could be a source of the non boot issue other than what I have mentioned. This project is taking quite some time to complete.
HELP!!!!!
Does omnirom even have a 8.1 branch for shamu? I can't see it.
Elektroschmock said:
Does omnirom even have a 8.1 branch for shamu? I can't see it.
Click to expand...
Click to collapse
No it doesn't I am trying to bring it up, I know they don't officially support it for Oreo. They don't have a device maintainer for Shamu anymore. I have been using other device trees and filling in the required files for building Omni. Like I said I can get a successful build but in won't boot.
Rondeau79 said:
No it doesn't I am trying to bring it up, I know they don't officially support it for Oreo. They don't have a device maintainer for Shamu anymore. I have been using other device trees and filling in the required files for building Omni. Like I said I can get a successful build but in won't boot.
Click to expand...
Click to collapse
Well thats not how a bringup works. You need a ****load oft kernel and device commits for shamu to Boot Oreo.
The easiest way would be forking a working device and kernel repository...e.g. LineageOS. Next step would be adapting some makefiles and overlays.
Not much to do, as we have already done the Oreo bringup months ago.
Elektroschmock said:
Does omnirom even have a 8.1 branch for shamu? I can't see it.
Click to expand...
Click to collapse
Elektroschmock said:
Well thats not how a bringup works. You need a ****load oft kernel and device commits for shamu to Boot Oreo.
The easiest way would be forking a working device and kernel repository...e.g. LineageOS. Next step would be adapting some makefiles and overlays.
Not much to do, as we have already done the Oreo bringup months ago.
Click to expand...
Click to collapse
Yeah that's what I did. When the build first starts something in the vendor folder isn't being globed. It's only 99% of that folder that is accounted for. The build is successful but it just won't boot. I don't have much coding experience so this is really hard.
Rondeau79 said:
Yeah that's what I did. When the build first starts something in the vendor folder isn't being globed. It's only 99% of that folder that is accounted for. The build is successful but it just won't boot. I don't have much coding experience so this is really hard.
Click to expand...
Click to collapse
Do an -eng build and read logcat and dmesg via ADB. Usually you see whats wrong.
Rondeau79 said:
Yeah that's what I did. When the build first starts something in the vendor folder isn't being globed. It's only 99% of that folder that is accounted for. The build is successful but it just won't boot. I don't have much coding experience so this is really hard.
Click to expand...
Click to collapse
Hey how's it going? I know you've been trying hard to get an 8.1 Omni build running. FYI, @bthorne79 put out an Omni 8.1 build for Shamu back in December of last year. I tried it at the time and it was a nice build but if I remember correctly, the theming didn't work. I tried changing to a dark theme but it just stayed on default light theme. I always use a dark or black theme on Shamu and I'd rather not have to use substratum. Other than that, the rom ran very well.
Maybe give him a shout, after all you both have 79 in your name. Good luck, I definitely wouldn't mind seeing an Omni 8.1 build.
Mike T
webdroidmt said:
Hey how's it going? I know you've been trying hard to get an 8.1 Omni build running. FYI, @bthorne79 put out an Omni 8.1 build for Shamu back in December of last year. I tried it at the time and it was a nice build but if I remember correctly, the theming didn't work. I tried changing to a dark theme but it just stayed on default light theme. I always use a dark or black theme on Shamu and I'd rather not have to use substratum. Other than that, the rom ran very well.
Maybe give him a shout, after all you both have 79 in your name. Good luck, I definitely wouldn't mind seeing an Omni 8.1 build.
Mike T
Click to expand...
Click to collapse
Yeah this project has been a real headache! Nothing like before were most of the files and programming has been done. I am not done yet, I don't give up too easy. This has been quite an undertaking.
Follow this thread: https://forum.xda-developers.com/showpost.php?p=54118631&postcount=4
That'll give you an idea about what all things to consider while building for a device :fingers-crossed:
Nitin
Rondeau79 said:
Yeah this project has been a real headache! Nothing like before were most of the files and programming has been done. I am not done yet, I don't give up too easy. This has been quite an undertaking.
Click to expand...
Click to collapse
If you push your sources somewhere where we can see it, it should be easier for us to help you.
nitin.chobhe said:
Follow this thread: https://forum.xda-developers.com/showpost.php?p=54118631&postcount=4
That'll give you an idea about what all things to consider while building for a device :fingers-crossed:
Nitin
Click to expand...
Click to collapse
Yeah that guide is so outdated! I followed that kind of stuff years ago when I couldn't build from source. I have done many builds that boot for 3 devices. This is a different problem.
Elektroschmock said:
If you push your sources somewhere where we can see it, it should be easier for us to help you.
Click to expand...
Click to collapse
Yeah I am using Omni ROM source with device trees and kernel source from Lineage Oreo. I have modified files and make sources. I have all up to date vendor blobs cloned. There is no build errors just an issue with the system not booting. Pushing my source upstream would be to my personal GitHub account which I have never had to do. Usually I repo/clone source build and upload to Android File Host for working zips. I have never pushed anything up from my laptop!
Rondeau79 said:
Yeah that guide is so outdated! I followed that kind of stuff years ago when I couldn't build from source. I have done many builds that boot for 3 devices. This is a different problem.
Click to expand...
Click to collapse
Even though it is outdated it is still very much applicable. I linked that thread because you mentioned this.
Good that you know the problem. All the best!
Nitin
Rondeau79 said:
Yeah I am using Omni ROM source with device trees and kernel source from Lineage Oreo. I have modified files and make sources. I have all up to date vendor blobs cloned. There is no build errors just an issue with the system not booting. Pushing my source upstream would be to my personal GitHub account which I have never had to do. Usually I repo/clone source build and upload to Android File Host for working zips. I have never pushed anything up from my laptop!
Click to expand...
Click to collapse
It's not that you need to push it somewhere to get working zips, it's just for us to see your modifications.
Otherwise it is just guessing what you did.
Making it compile is one thing...making it work is a whole different story. When I did the Lineage Oreo bringup I had many things to fix untill it bootet properly.
Lineage's device tree for shamu depends on several other Lineage repositories. For example:
android_hardware_lineage_interfaces
android_hardware_qcom_gps
android_vendor_lineage
android_device_lineage_sepolicy
So just cloning the device repo without examining the reason for your 'bootloop' will get you nowhere.
I'd suggest you to do an engineering build which is easier to debug. Altought I would make this build permissive as selinux denials are a common reason for a bootloop.
How about trying a known working kernel such as popcorn? If the kernel works, you might be able to get some logs. And if you know whether your kernel is the problem, you are one step closer.
runekock said:
How about trying a known working kernel such as popcorn? If the kernel works, you might be able to get some logs. And if you know whether your kernel is the problem, you are one step closer.
Click to expand...
Click to collapse
If he is based of lineage's device tree, then Lineage's kernel is the best choice.
You should always be able to grab a log, no matter which kernel you are using.
Yeah willing to help out put your stuff up on git and give links
My webhost is down right meow but I can post my Oreo build I had of Omni here I didn't use Los trees i used aosp trees from camcory , joryb and mine cleanos
---------- Post added at 08:47 PM ---------- Previous post was at 08:28 PM ----------
i found my build i had it on afh if you want it to try this was early oreo so some things not fully functional i think gps was iffy https://www.androidfilehost.com/?fid=889964283620775520 i am syncing up omni oreo again for sure will help out if you want please pm
---------- Post added at 08:47 PM ---------- Previous post was at 08:47 PM ----------
nougat builds in that root folder also
bthorne79 said:
Yeah willing to help out put your stuff up on git and give links
Click to expand...
Click to collapse
My device and kernel tree GitHub link.
https://github.com/Rondeau79
Rondeau79 said:
My device and kernel tree GitHub link.
https://github.com/Rondeau79
Click to expand...
Click to collapse
Your probably going to want to fork the toolchains over that was used with cleanaosp
<!-- Prebuilts --> <project path="prebuilts/gcc/linux-x86/arm/arm-eabi-4.9" name="joryb/arm-eabi-7.x" clone-depth="1" remote="bb" revision="master"/> <project path="prebuilts/gcc/linux-x86/arm/arm-linux-androideabi-4.9" name="joryb/arm-linux-androideabi-7.x" clone-depth="1" remote="bb" revision="master"/>