• After 15+ years, we've made a big change: Android Forums is now Early Bird Club. Learn more here.

Root cm11 lucid2 l1v vs870 wip port log/thread

rbheromax said:
I'd recommend using an older compiler toolchain
Click to expand...

I switched to Google's gcc 4.4.3.
The fail still occurred so i disabled kernel tracers.
Issue gone. but with it is a new set of fals.
need to keep adding your flag to each folder i guess.
Not sure why its being so damn picky this time but its annoying.
I will reply with progress later.
 
Well u can add each missing flag to the top level Makefile. (In the root Dir of source)
Like if it says -W-no-unintialized you would add -W-unintialized to KBUILD_CFLAGS
 
rbheromax said:
Well u can add each missing flag to the top level Makefile. (In the root Dir of source)
Like if it says -W-no-unintialized you would add -W-unintialized to KBUILD_CFLAGS
Click to expand...

Thanks! I wasnt sure if kbuild flags worked with the cflags(im still a newbie dont hurt me pls)
Anyways, i fixed pretty much all of the uninitialized ones.
clock_8960.c Didnt like it if DEBUGFS config option was off. E.g. it declares some stuff with ifdef in one part, but another part that uses those vars doesnt have ifdef. So i have to add ifdef to the rest.
 
Try it without it. It may not be. I'll play around with the kernel to get it to compile with newer toolchains. Where is it
 
rbheromax said:
Try it without it. It may not be. I'll play around with the kernel to get it to compile with newer toolchains. Where is it
Click to expand...

When you disable debugfs in the config.
You have to comment each line out that references a debugfs function.
Also. i initially disabled debugfs because some of the fs stuff wouldnt compile.
And i had found that as an answer from a dev on xda
 
rbheromax said:
Try it without it. It may not be. I'll play around with the kernel to get it to compile with newer toolchains. Where is it
Click to expand...

UPDATE: A dev that i have on google hangouts told me debugfs is important for kmesg and logcat. So i used the meld tool to compare the original source tree with this one and uncommented all the debugfs lines.
Still working on the rest.
 
make[1]: Leaving directory `/media/android/cm11/kernel/lge/l1v'
make -C kernel/lge/l1v O=/media/android/cm11/out/target/product/l1v/obj/KERNEL_OBJ INSTALL_MOD_PATH=../../system ARCH=arm CROSS_COMPILE=" /media/android/cm11/prebuilts/gcc/linux-x86/arm/arm-eabi-4.7/bin/arm-eabi-" modules_install
make[1]: Entering directory `/media/android/cm11/kernel/lge/l1v'
INSTALL arch/arm/mach-msm/dma_test.ko
INSTALL arch/arm/mach-msm/msm-buspm-dev.ko
INSTALL arch/arm/mach-msm/reset_modem.ko
INSTALL block/test-iosched.ko
INSTALL crypto/ansi_cprng.ko
INSTALL drivers/bluetooth/bluetooth-power.ko
INSTALL drivers/crypto/msm/qce40.ko
INSTALL drivers/crypto/msm/qcedev.ko
INSTALL drivers/crypto/msm/qcrypto.ko
INSTALL drivers/gud/mcdrvmodule.ko
INSTALL drivers/gud/mckernelapi.ko
INSTALL drivers/media/radio/radio-iris-transport.ko
INSTALL drivers/media/video/gspca/gspca_main.ko
INSTALL drivers/misc/eeprom/eeprom_93cx6.ko
INSTALL drivers/mmc/card/mmc_block_test.ko
INSTALL drivers/mmc/card/mmc_test.ko
INSTALL drivers/scsi/scsi_wait_scan.ko
INSTALL drivers/spi/spidev.ko
INSTALL drivers/video/backlight/lcd.ko
DEPMOD 3.4.0-g44ec603-dirty
make[1]: Leaving directory `/media/android/cm11/kernel/lge/l1v'
mdpath=`find /media/android/cm11/out/target/product/l1v/system/lib/modules -type f -name modules.order`; if [ "$mdpath" != "" ];then mpath=`dirname $mdpath`; ko=`find $mpath/kernel -type f -name *.ko`; for i in $ko; do /media/android/cm11/prebuilts/gcc/linux-x86/arm/arm-eabi-4.7/bin/arm-eabi-strip --strip-unneeded $i; mv $i /media/android/cm11/out/target/product/l1v/system/lib/modules/; done; fi
mdpath=`find /media/android/cm11/out/target/product/l1v/system/lib/modules -type f -name modules.order`; if [ "$mdpath" != "" ];then mpath=`dirname $mdpath`; rm -rf $mpath; fi
mdpath=`find /media/android/cm11/out/target/product/l1v/system/lib/modules -type f -name modules.order`; if [ "$mdpath" != "" ];then mpath=`dirname $mdpath`; ko=`find $mpath/kernel -type f -name *.ko`; for i in $ko; do /media/android/cm11/prebuilts/gcc/linux-x86/arm/arm-eabi-4.7/bin/arm-eabi-strip --strip-unneeded $i; mv $i /media/android/cm11/out/target/product/l1v/system/lib/modules/; done; fi
mdpath=`find /media/android/cm11/out/target/product/l1v/system/lib/modules -type f -name modules.order`; if [ "$mdpath" != "" ];then mpath=`dirname $mdpath`; rm -rf $mpath; fi
target Prebuilt: (/media/android/cm11/out/target/product/l1v/kernel)
Install: /media/android/cm11/out/target/product/l1v/root/file_contexts
Click to expand...

Got it to compile with CM11 toolchains
Also you didn't have SELinux enabled in config so thats probably why the original port bootlooped instead of hanging.

https://bitbucket.org/rbheromax/kernel-l1v (goes in kernel/lge/l1v, stripped out FreedomKernel stuff)
https://bitbucket.org/rbheromax/android-l1v (goes in device/lge/l1v, bring-up for cm11)

that should get you a buildable kernel, bootable recovery, and possibly even a built rom that u can start hammering bugs out of
 
rbheromax said:
Got it to compile with CM11 toolchains
Also you didn't have SELinux enabled in config so thats probably why the original port bootlooped instead of hanging.

https://bitbucket.org/rbheromax/kernel-l1v (goes in kernel/lge/l1v, stripped out FreedomKernel stuff)
https://bitbucket.org/rbheromax/android-l1v (goes in device/lge/l1v, bring-up for cm11)

that should get you a buildable kernel, bootable recovery, and possibly even a built rom that u can start hammering bugs out of
Click to expand...

Uhhhhh, i got that working yesterday.
Thanks for the initiative I will add your stuff to the build env and keep working :D
But we have bigger problems.
DD isnt actually writing to the reocvery partition.
Im gonna post an img tonight and let people try it(with instructions of course).
But md5sum shows that neither loki or dd is actually changing the contents of recovery -> /dev/block/mmcblk0p19
 
rbheromax said:
grab a dmesg from adb shell
and dont write to that, write to /dev/block/platform/msm_sdcc.1/by-name/recovery
Click to expand...
Here is the shell shell - Pastebin.com
As you can see, i zeroed out the recovery partition, imaged it. And it matched md5sums.
But when i write the img to the partition. The partition's md5sum doesnt match the recovery.img's md5sum.
Either the recovery.img is too big and isnt getting completely written or something else is happening because dd doesnt indicate that the img is reaching the end of the partition at all.

output of cat /proc/partitions :
" 179 19 12288 mmcblk0p19"

output of ls -al /storage/sdcard1/recovery.img :
"-rwxrwxr-x system sdcard_rw 9467904 2014-09-16 19:36 recovery.img"

dmesg indicated nothing useful and loki doesnt do anything useful either.
dmesg has been repeating
<6>[69864.242903] max17043 check temp = 27, PreviousRcomp =0x5F
<6>[69864.244978] [Battery] max17043_work : soc(100), volt(4302)
<4>[69871.809147] >>>pet_watchdog_work [system rev: 10][raw ver: 7]
<6>[69874.661215] max17043 check temp = 27, PreviousRcomp =0x5F
<6>[69874.662009] [Battery] max17043_work : soc(100), volt(4301)
over and over.
------------
On another note, I looked at what you did with the config and and kernel. Thank you again, I was too lazy to pull the rest from the l1m tree :p
 
with recovery you;ll also want my fstab. its formatted differently to match newer cwm specs
it depends....your recovery is 9.2 mb
if your partition is only 8 mb then yea too big
but the writing that would only make recovery not bootable at, not boot the old version
i thnk your recovery partition should be ~16mb because thats what newer devices with bigger emmc's ship with. do you guys get fastboot when zero-ing out boot and recovery partitions? if so, thats where u can test boot the images and probably flash too. and from this state you'll still be able to get to download mode to restore a kdz from sammyz's method
 
lgrootnoob said:
Uhhhhh, i got that working yesterday.
Thanks for the initiative I will add your stuff to the build env and keep working :D
But we have bigger problems.
DD isnt actually writing to the reocvery partition.
Im gonna post an img tonight and let people try it(with instructions of course).
But md5sum shows that neither loki or dd is actually changing the contents of recovery -> /dev/block/mmcblk0p19
Click to expand...
Dude I feel bad I haven't been helping with the project much. Please tell me if you need anything.. Involve me more please
 
rbheromax said:
with recovery you;ll also want my fstab. its formatted differently to match newer cwm specs
it depends....your recovery is 9.2 mb
if your partition is only 8 mb then yea too big
but the writing that would only make recovery not bootable at, not boot the old version
i thnk your recovery partition should be ~16mb because thats what newer devices with bigger emmc's ship with. do you guys get fastboot when zero-ing out boot and recovery partitions? if so, thats where u can test boot the images and probably flash too. and from this state you'll still be able to get to download mode to restore a kdz from sammyz's method
Click to expand...
Actually i was _just_ thinking about that.
Cause I had wiped the boot partition of my optimus v a couple days ago and it gav me fastboot. xD
But anyways, I will try very soon and report back.
I do know that if you wipe the aboot, sammyz said that it mounts as a usb drive on your pc and gives you access to "modem" files and "adsp" files.
 
dont wipe aboot. it'll put ur phone in download mode
it'll show as: qhsusb_dload

from there you can directly write to and from the internal emmc while in linux because it mounts as a logical device
 
rbheromax said:
dont wipe aboot. it'll put ur phone in download mode
it'll show as: qhsusb_dload

from there you can directly write to and from the internal emmc while in linux because it mounts as a logical device
Click to expand...

12582912 Bytes to Megabytes Conversion Calculator
I dded from dev/zero to the recovery partition and this is what I got ^
it says its 12 megabytes.
So I think I need to adjust my "BOARD_RECOVERYIMAGE_PARTITION_SIZE := 50331648"
 
rbheromax said:
dont wipe aboot. it'll put ur phone in download mode
it'll show as: qhsusb_dload

from there you can directly write to and from the internal emmc while in linux because it mounts as a logical device
Click to expand...

Testing with:
BOARD_BOOTIMAGE_PARTITION_SIZE := 0x600000
# Maybe 0xBB8000 or 6291456 below, not sure. Still expirementing.
BOARD_RECOVERYIMAGE_PARTITION_SIZE := 0x600000
I _know_ the block size is 4096.
Some dude said to multiply by 512 bytes the output of cat /proc/partitions
And then sometimes its in hex.
dd gives us 12 mb in bytes. Soooo not sure.
 
You must log in or register to reply here.
Back
Top Bottom