Database Info

[DEVELOPMENT and DISCUSSION][KEXEC] ISSUES DISCUSSION

Original thread by hashcode for RAZR http://forum.xda-developers.com/showthread.php?t=2181971
Here is a copy paste.
Hashcode said:
** WARNING: THIS THREAD SHOULD BE FOR KERNEL DEVELOPERS OR ROM DEVS.
Users please post questions in your ROM's support threads. **
** DEVS: If you choose to use a kexec kernel in your ROM add a "KERNEL CREDIT" for the first 4 people in the "WHO HAS WORKED ON THIS?" section below. This has been spectacular development and credit is due. **
TESTING ROMS FOR KEXEC KERNEL PROGRESS:
Droid RAZR:
[2013-03-12] cm-10.1 on v3.0.8 JB kernel:link removed for safety
GSM RAZR:
cm-10.1 on v3.0.8: link removed for safety
MAPPHONE 3.0.x KEXEC KERNEL:
For use on Motorola devices; Droid RAZR, GSM RAZR, RAZR MAXX, Bionic and Droid 4. (1 GB OMAP4 devices for now, 512MB support coming).
WHAT IS A KERNEL?
Also from Wikipedia: http://en.wikipedia.org/wiki/Kernel_%28computing%29
Summary: If you had to choose 1 binary file which connects the hardware components of your device to the OS layer whether it be Android or some other software, it would be the kernel.
WHAT IS KEXEC?
The best description is found here on Wikipedia: http://en.wikipedia.org/wiki/Kexec
Summary: kexec is a process for hot reloading a new kernel while a device is running.
WHERE DID THIS KERNEL COME FROM?
Initially the source came from Motorola. All GPL components of Android are posted on Sourceforge.net for Motorola devices:
http://sourceforge.net/motorola/
After that it's been worked on (see below) to support kexec function and work for our needs.
WHO HAS WORKED ON THIS?
Several people played a large role in making this happen:
Kholk & [MBM]: Initial development work done on the RAZR/Bionic under Gingerbread to setup kexec modules + booting kexec kernel (no 2nd cpu)
jonpry: Ported and improved Kholk/MBM's work for use on the Droid 3. Brought up 2nd CPU, fixes for audio and more.
Hashcode (myself): On going development of this kernel for flawless function on our devices.
A full kernel source history on github starting with omapzoom sources and migrating Motorola source in commit by commit
Fixes for clocks, local timers, GPMC setup, race conditions in irq's due to kexec and more.
On going development towards full support for current Android / Ubuntu Phone OS requirements (including 512MB and Droid 3 support).
Other's have also contributed code (and hopefully more will as well): kfazz(D4), jarveson(D3) and more.
WHERE IS THE SOURCE?
KERNEL SOURCES (ALL DEVICES):
https://github.com/STS-Dev-Team/kernel_mapphone_kexec
BRANCH: stock-ics-kexec
DESCRIPTION: This is a perfect copy of the ICS kernel running as kexec. For use with stock-based ROM's etc.
STATUS: Ready for use
BRANCH: jb-mr1
DESCRIPTION: Ongoing kernel branch for support of newer Android OS (Jellybean MR1 currently).
STATUS: Work-in-progress. Heavy development
KEXEC KERNEL-MODULE SOURCES FOR ICS/JB BASED DEVICES:
https://github.com/STS-Dev-Team/kernel_kexec_modules/tree/3.0.8-ics-spyder
Branch: 3.0.8-ics-spyder
Defconfig: mapphone_kexec_defconfig
KEXEC KERNEL-MODULE SOURCES FOR GB BASED DEVICES (Droid 3):
https://github.com/STS-Dev-Team/kernel_kexec_modules/tree/2.6.35-solana/
Branch: 2.6.35-solana
Defconfig: mapphone_defconfig
HOW TO KEXEC A DEVICE
1. Download kexec kernel modules + kexec binary for your base kernel version
(Most devices found here):
https://github.com/STS-Dev-Team/android_device_motorola_omap4-kexec-common/tree/cm-10.1/kexec
(Droid 3 Users need GB based modules here):
https://github.com/STS-Dev-Team/android_device_motorola_solana/tree/jb-mr1/kexec
2. Download the devtree for your device:
Droid RAZR: https://github.com/STS-Dev-Team/android_device_motorola_spyder/tree/jb-mr1/kexec
GSM RAZR: https://github.com/STS-Dev-Team/android_device_motorola_umts_spyder/tree/jb-mr1/kexec
Bionic: https://github.com/STS-Dev-Team/android_device_motorola_targa/tree/jb-mr1/kexec
Droid 4: https://github.com/STS-Dev-Team/android_device_motorola_maserati/tree/jb-mr1/kexec
Droid 3: Droid 4: https://github.com/STS-Dev-Team/android_device_motorola_solana/tree/jb-mr1/kexec
3. Need a kexec kernel binary from the above sources: renamed the zImage as "kernel".
4. Need a .gz cpio ramdisk named: ramdisk.img
**RAMDISK needs to take care of mounting related filesystems for your ROM.
Ususally Safestrap/BMM handles this in some way for rom-slots.
For easy kexec tests you can build a CWM ramdisk and use that.
Now boot up into a safe kexec point on the device be it BMM or Safestrap where you can use adb.
5. Move all of the above files onto the rootfs
6. Execute the following as a script:
Code:
# run this from a "safe" boot point (safestrap / bootmanager) where no mounts are active
cd /
chmod 755 /kexec
# you will lose USB function here till the new kernel boots -- hence the script
insmod /uart.ko
insmod /arm_kexec.ko
insmod /kexec.ko
/kexec -l /kernel --devtree=/devtree --ramdisk=/ramdisk.img
sleep 1
/kexec -e
CURRENT DEVELOPMENT ROAD MAP
Merge up to Linux Mainline v3.0.21 + Google kernel/common commits (where applicable)
Cherry-pick from TI's kernel tag 4AI.1.5 to 4.AI.1.6
Merge up to Linux Mainline v3.0.31 + Google kernel/common commits (where applicable)
Cherry-pick from TI's kernel tag 4AI.1.6 to 4.AI.1.7
Also pick up any p-android-omap-3.0 and p-android-omap-3.0-dev commits where needed
Merge up to current Linux Mainline + Google kernel/common commits (where applicable)
Click to expand...
Click to collapse
Currently, there is one kexec based rom for Atrix 2 CM10.1 by wanggjghost here.
Source Code for Atrix 2 http://sourceforge.net/projects/atrix2.motorola/files/
and http://sourceforge.net/projects/me865.motorola/files/ME865/
Other source code that can help you wanggjghost's github https://github.com/SunRain/android_device_motorola_edison
ISSUES TO BE SOLVED
The current major issue regarding Kexec is that some devices get a black screen once kexec is executed.
This thread to discuss and develop a solution for it. And once that is solved, to discuss custom kernels.

Here are a few things that might help.
kmsg when kexec boot up failed: http://www.mediafire.com/download.php?95raby988dvch74 (just open it with a text editor. It is corrupt but you can make out quite a lot.)
The dmesg for stock ics: http://www.mediafire.com/view/?z4pibf2at3u4i7e (Going through this will give you better idea as how things should be.)
And the panic msg after I soft reset it after a black screen: http://www.mediafire.com/view/?z4pibf2at3u4i7e (When I get a black sscreen, I hold vol up+ vol down+ pow for about 10 secs. that soft resets my phone and a panic msg is generated which I retrieve.)
From the kmsg, here is something interesting that appears to be the cause. (Note: I have made some edits to make it readable, not completely readable though.)
Code:
[ 2.745880,0] cannot apply mgr(lcd) on inactive device
[ 2.745971, ] omapfb omapfb: failed to apply dispc config
[ ².·46893¬0] cannot apply mgr(tv) on inactive device
[ 2.746154,0] omapfb omapfb: failed to apply dispc config
[ 2.746307,0] cannot aplly mgr(lcd2) on inactive(device
[ 2.7463¶8,0\ omapfb omapfb: failed to apply dispc config
[ 2.746520,0] DSI: dsi_fc_kmd = 0, dói_vc_video 9 1
[ 2®746582,0] DT: x_res=540 y_res=960 phy_w_mm=54 phy_h_mm=95
[ 2.·46673,0](ÄT: hfp=0 hsw=² hbp-2 vnp=0 ösw=1 vbp=1
[ 2®746795,0] DTº cìk_lanå=1 clk_pos90 d1_line=2 d1_pos=0 d2_lane=3 d2_pos=0
[ 2.746948,0](DT: d3_lane=0 d3_pos=0 d4_lane=0 d4_pos=0
[ ².747009,0] DT: panel_id= x1a000² type= 0
[ 2.747102,0] DT: regn=13 regm=230 regm3=6 regm4=6 lp_clk_div=9 lck_div=1 pck_div=4
[ 2.747222,0] DT: te_support=1, te_type=0
[ 2.747283,0] DT: number of power suppljes-±
[ 2.747436,0] DT: Supply #°, beg name=sw4, gpin=96,*gpio_en_val=1
[ 2.(43¸57,0] mmc2: card claims to support voltages below the defined range. These will be ignored.
[ 2.860656,0] omapdss DSI error: Failed to receive BTA
[ *2.861816,0] mmc2: quduing unknïwn CIS tupìe 0x99 (3 jytms)*[( 2.862792¬0] mmc2: .%w high speeä RDIG kard at address 0001
[ 2.962280,0] omapdss DSI error: Failed to receive BTA
[ 3.063781,0] omapdss DSI error: Failed to receive BTA
[ "3®091705,0] mmc0: host doås(not cupport reiding re!d-onlù switch. assuminç write-enable.
[ 3.0)5336,0] míc0: new hkgh speed SDHC cabd at address aaaa
[ 3.896496,0] mmcblk0:(mmc0:aaaa SU16G 14.8 GiB
[ 3®101·76,0] mmcblk : p1
[" 3.165374,0] omapdss DS (error: Fa)led to receire BTA
[ 3.26693',0] omapnss DSI error: Failed to rekeivm(BTA*[ ¨ 3®368538,0] kmapdss DS) ebror: Failed to receivm BTA
[ 3.<78062¬0Y omapdss DSI error: Failed to receive BTI
[ "3.571624,0] omapdss DSI error:(Failed to receive BTA
[ 3*673217,0](oma0dss DSI ezror: Failed to receive BTA
[ ³.774749,0] omapdss DSI error: Fâile` to rekeive BTA
[ 3.¸76312,0\ om!pdss SI erro2: Failed to receive BTA
[ 3.876556,0] check_mapxjone_panel_present: Faineä to recmive BTA ACK foz 10 times.
[ 3.877014,0] marphone_xanel_power_on:Paneì és not attached nò failed ðo send BTA.*[ 3.8772¸8,0] omapfb omapfb: Fáile$ to enable display 'lcd'
[ 3.¸77746,0] kmapfb omapfb: failed to inkpialize default display
[ 3.¸79821,0] omarfb omapfb: failed |o se4up omapf"Š[ 3.¸80279,0] omapfb: probe of omapfb fa)led with error -22
[ 3.882019,0] regulator_init_complete: vwlan1: disibling
[ " 3.882873,0] begulat/r_inét_complete: vrfref: dhsabling
[ 3.883544,0] renulator_init_coíphete: vbf²: disabling
Z 3.8:7145,0] regulator_check_voltage: v5sb: operation not allowed
[ 3.:91937,0] lm3532_bl_work: Display backlight(disable)
The issue appers to be regarding omapfb (frame buffers). might be an issue with the command line. wang's kernel uses
Code:
root=/dev/ram0 rw [email protected] console=null vram=10300K omapfb.vram=0:8256K,1:4K,2:2040K init=/init ip=off mmcparts=mmcblk1:p7(pds),p15(boot),p16(recovery),p17(cdrom),p18(misc),p19(cid),p20(kpanic),p21(system),p22(cache),p23(preinstall),p24(webtop),p25(userdata),p26(emstorage) mot_sst=1 androidboot.bootloader=0x0A74
However, using
Code:
root=/dev/ram0 rw [email protected] console=ttyO2,115200n8 init=/init ip=off mmcparts=mmcblk1:p7(pds),p15(boot),p16(recovery),p17(cdrom),p18(misc),p19(cid),p20(kpanic),p21(system),p22(cache),p23(preinstall),p24(webtop),p25(userdata) mot_sst=1 androidboot.bootloader=0x0A73
might solve.
Also, the following keeps repeating
Code:
PVR_K:(error):BridgedDispatchKM: Initialisation failed. Driver unusable
but PVR_K is related to display (I guess) and that again takes us back to omapfb.
Put in your thoughts as well but try to keep the thread clean.

Here's another panic log: http://pastebin.com/TdU31ERU
Regarding the "PVR_K: (Error): BridgedDispatchKM: Initialisation failed. Driver unusable." error that Ravi pointed out... and for my curiosity..
Can someone who has the kexec build running run these from a command line, please? (and post the output here):
Code:
adb shell
cat /proc/pvr/version
strings vendor/lib/egl/libGLESv2_POWERVR_SGX540_120.so | grep build

$ adb shell
* daemon not running. starting it now on port 5038 *
* daemon started successfully *
error: device not found
$ cat /proc/pvr/version
Version SGX_DDK_Android sgxddk 19 [email protected] (release) omap4430_android
System Version String: SGX revision = 1.2.0
_POWERVR_SGX540_120.so | grep build <
OpenGL ES GLSL ES 1.00 build [email protected]
OpenGL ES 2.0 build [email protected]
$
Sent from my MB865 using xda app-developers app
---------- Post added at 04:41 PM ---------- Previous post was at 04:35 PM ----------
Is that what u need?
Sent from my MB865 using xda app-developers app

jaymccoubrey said:
$ adb shell
* daemon not running. starting it now on port 5038 *
* daemon started successfully *
error: device not found
$ cat /proc/pvr/version
Version SGX_DDK_Android sgxddk 19 [email protected] (release) omap4430_android
System Version String: SGX revision = 1.2.0
_POWERVR_SGX540_120.so | grep build <
OpenGL ES GLSL ES 1.00 build [email protected]
OpenGL ES 2.0 build [email protected]
$
Sent from my MB865 using xda app-developers app
---------- Post added at 04:41 PM ---------- Previous post was at 04:35 PM ----------
Is that what u need?
Sent from my MB865 using xda app-developers app
Click to expand...
Click to collapse
Yup, thanks.. I was just chasing wild geese again..
[Edited to reduce clutter, etc.. ]
So, yea, wild goose chase...
--------------------------
Also, @Ravi - this is my current device BoardConfig:
Code:
# inherit from common
-include device/motorola/omap4-kexec-common/BoardConfigCommon.mk
# inherit from the proprietary version
-include vendor/motorola/edison/BoardConfigVendor.mk
# Processor
TARGET_BOOTLOADER_BOARD_NAME := edison
# Kernel
BOARD_KERNEL_CMDLINE := root=/dev/ram0 rw [email protected] console=null vram=10300K omapfb.vram=0:8256K,1:4K,2:2040K init=/init ip=off mmcparts=mmcblk1:p7(pds),p15(boot),p16(recovery),p17(cdrom),p18(misc),p19(cid),p20(kpanic),p21(system),p22(cache),p23(preinstall),p24(webtop),p25(userdata),p26(emstorage) mot_sst=1 androidboot.bootloader=0x0A72
BOARD_KERNEL_BASE := 0x80000000
BOARD_PAGE_SIZE := 0x4096
# Kernel Build
TARGET_KERNEL_SOURCE := kernel/motorola/omap4-kexec-common
TARGET_KERNEL_CONFIG := mapphone_1GB_defconfig
And I have the same BOARD_KERNEL_CMDLINE in the mapphone_1GB_defconfig.. but pretty sure I have other issues as well. Here's a kmsg from an attempt at booting mine: http://pastebin.com/VZzg0ALD - containing my favorite debug message ever: Bad BP 0[ID: 0x0], Is this a Phone ??
One more thing, take a look at the atags with a hex editor...

Im kimda limited on what i no how to do.. Tell me what to do and how ill do it.
Sent from my MB865 using xda app-developers app
---------- Post added at 06:32 PM ---------- Previous post was at 06:26 PM ----------
I download a hex Editor. Where do i look?
Sent from my MB865 using xda app-developers app

jaymccoubrey said:
Im kimda limited on what i no how to do.. Tell me what to do and how ill do it.
Sent from my MB865 using xda app-developers app
---------- Post added at 06:32 PM ---------- Previous post was at 06:26 PM ----------
I download a hex Editor. Where do i look?
Sent from my MB865 using xda app-developers app
Click to expand...
Click to collapse
Lol.. My ramblings weren't directed at you specifically.. It's just for general (possibly cluttered and useless) info.. And that last note was for Ravi actually - on the topic of board kernel command lines..
Sent from my MB865 using xda app-developers app

So do u need anything else. Ill help cause ur roms kick ass
Sent from my MB865 using xda app-developers app

jaymccoubrey said:
So do u need anything else. Ill help cause ur roms kick ass
Sent from my MB865 using xda app-developers app
Click to expand...
Click to collapse
Thanks.. But I'll post if I need/want to see something else from kexec running..
And this kexec and kernel stuff is new territory for me (well, android is relatively new to me still also..) - so I am just trying to learn/do what I can.. The more getting involved, the merrier.
Sent from my MB865 using xda app-developers app

I just received my dev Atrix 2 today thanks to rvalentin913.
My goal with the device is to get it running alongside our other devices (Droid3, Bionic, RAZR's and Droid 4) so that it can be merged into CM's official devices.
However there does seem to be some missing pieces so to speak in the kernel that I'll be looking to fix over the next few weeks.
I'll try and comment on what I'm seeing in the posts above:
1. "omapdss DSI error: Failed to receive BTA" This indicates that the LCD on some of the devices isn't directly supported by the kernel and/or not being setup correctly by the devtree. Also very likely the cause of a blank screen (obviously).
2. "PVR_Kerror):BridgedDispatchKM: Initialisation failed. Driver unusable" indicates that pvrsrvinit didn't complete the initialization of the SGX libs for 1 reason or another. This could many things here and possibly related to #1. I'll have to run a strace or other tests to see where it's failing.
3. I'm reverting the mem= param back to 1023M in the kexec kernels due to a bug in the way Motorola has the kernel setup. They set the SMC driver memory to 2M, but the driver itself will expand to a minimum of 3M. The way the kernel is setup this area is designated right at the very top of memory. So Motorola probably didn't understand why they were getting odd memory issues when using the 1024 setting (this effectively pushes the SMC area out of the 1GB RAM limit). So they dropped the mem= parameter to 1023 where the SMC driver would set the start of SMC memory to 1021M and drift up to 1024.
4. Also the ATAGS which point to the devtree memory location have a tendancy to be cleared out at the drop of a hat for some reason. And that causes an empty devtree to be read into memory. (IE: Bad BP 0[ID: 0x0], Is this a Phone). I'm also chasing this down to see what clears out this memory area. The first time it was an early cache setup by the SMC driver, but there's somewhere else now as well.
What makes this complicated is that the memory location for the devtree and atags is decided at the point of kexec. So it can be a bit random causing this issue to only occasionally show up.
Anyway, I'll try and post a full thread w/ full support once I get my end up and running on the Atrix 2.
I haven't decided yet how I want to boot the device into a 2nd ROM. Currently Safestrap doesn't use external SD cards for rom-slots and to use the internal area, I would need to play some tricks on the OS using the fuse daemon (treat it like an internal SD card). Personally, I think the internal area gives the best performance and I'm leaning towards this route.
But, I won't make anyone flash over their stock system. Ever.
EDIT: I just read through some of the bootlogs and I'm confused about the end of the one here:
http://www.mediafire.com/view/?z4pibf2at3u4i7e
- "init: cannot execve('/system/bin/debuggerd'): No such file or directory" This looks like an obvious missing file, but I wonder why it's being called in the first place.
- procfs_write: version magic '2.6.35.7 SMP preempt mod_unload ARMv7 ' should be '3.0.8 SMP preempt mod_unload ARMv7 '
- dirty_arm_kexec: version magic '2.6.35.7 SMP preempt mod_unload ARMv7 ' should be '3.0.8 SMP preempt mod_unload ARMv7 '
- dirty_kexec: version magic '2.6.35.7 SMP preempt mod_unload ARMv7 ' should be '3.0.8 SMP preempt mod_unload ARMv7 '
These are GB-based kernel modules built for kexec trying to execute at the end of the log? *boggle*

Hashcode said:
I just received my dev Atrix 2 today thanks to rvalentin913.
Click to expand...
Click to collapse
Nice (I was kinda hoping/guessing it was you).
Will be very glad to see you digging into this phone also. Let me (or any of us) know if you need any assistance..
And thanks again for your major efforts and contributions in general.
Sent from my MB865 using xda app-developers app

Hashcode said:
I haven't decided yet how I want to boot the device into a 2nd ROM. Currently Safestrap doesn't use external SD cards for rom-slots and to use the internal area, I would need to play some tricks on the OS using the fuse daemon (treat it like an internal SD card). Personally, I think the internal area gives the best performance and I'm leaning towards this route.
But, I won't make anyone flash over their stock system. Ever.
Click to expand...
Click to collapse
Sir, the moto guys at China have done that. They mount /data/media as internal sd. i.e., the stock me865 roms have it. I will upload the ramdisk, vold.fstab, vold.fstab.swap and the build.prop when I get to my comp. I tried using 2nd init with the changes in ramdisk, included the fstab files and made the changes in build.prop but that did help me.. let me know if anything else is required.. for bmm, we just mount data for int and the virtual systems are stored there.
And thanks again for all the work you have done..
Sent from my MB865 using Tapatalk 2

Ravikirancg said:
Sir, the moto guys at China have done that. They mount /data/media as internal sd. i.e., the stock me865 roms have it. I will upload the ramdisk, vold.fstab, vold.fstab.swap and the build.prop when I get to my comp. I tried using 2nd init with the changes in ramdisk, included the fstab files and made the changes in build.prop but that did help me.. let me know if anything else is required.. for bmm, we just mount data for int and the virtual systems are stored there.
And thanks again for all the work you have done..
Sent from my MB865 using Tapatalk 2
Click to expand...
Click to collapse
Ah, yep, the /data/media mount is managed via the fuse daemon as a simulated VFAT partition, and needs to have FUSE support in the kernel (easily done in kexec).
Any idea why there are GB kexec modules trying to execute in that log?

Hashcode said:
Ah, yep, the /data/media mount is managed via the fuse daemon as a simulated VFAT partition, and needs to have FUSE support in the kernel (easily done in kexec).
Any idea why there are GB kexec modules trying to execute in that log?
Click to expand...
Click to collapse
ah! It requires a modified kernel as well! Thank you. That was the missing piece when I tried it some months ago. Well any way, the good news is that I have managed to emulate internal sdcard on stock searet. I just need to make out all the necessary changes that need to be made in the build.prop. For now, I just used miui build.prop. So that issue is more or less solved except for att users may have to erase cid to flash the hktw boot.img
EDIT: Well a little more work needs to be done I guess.. just some fc..
regarding the GB modules, maybe I was just experimenting.. since its happening after phone is overclocked, its my doing I guess. I will try loading on stock rom again and will let you know..
and take a look at this. defy having kinda the same issue. http://mcbawse177.appspot.com/pastebin.com/31VLSvyJ
quarx might have an idea on how to solve this.. look at line 1046

alteredlikeness said:
One more thing, take a look at the atags with a hex editor...
Click to expand...
Click to collapse
that was interesting.. well, among the currently supported phones, all have that partition table except XT875. Only that has lbl and lbl_backup like our phone but it has an emstorage at p26 and p27 is sgpt. here (also take a look at the other models listed there.) I was once just trying kexec with all the devtrees and there was one devtree using which, my keypad got lighted up after kexec was executed. with all others, the keypad would turn off and stay as such. I will try them again. Oh and just for fun (if you have moto drivers software installed on comp), download the devtrees for all the devices, replace one of them, turn off phone, turn it back on with kexec enabled, connect the phone to comp and take a look at the drivers that windows install. change the devtree and do it again..

Ravikirancg said:
that was interesting.. well, among the currently supported phones, all have that partition table except XT875. Only that has lbl and lbl_backup like our phone but it has an emstorage at p26 and p27 is sgpt. here (also take a look at the other models listed there.) I was once just trying kexec with all the devtrees and there was one devtree using which, my keypad got lighted up after kexec was executed. with all others, the keypad would turn off and stay as such. I will try them again. Oh and just for fun (if you have moto drivers software installed on comp), download the devtrees for all the devices, replace one of them, turn off phone, turn it back on with kexec enabled, connect the phone to comp and take a look at the drivers that windows install. change the devtree and do it again..
Click to expand...
Click to collapse
Well, about the devtree - I found that wang added this to the finalize_release script that initializes the cmdefy bootmenu recovery (courtesy of Hashcode, apparently):
Code:
#!/sbin/sh
PATH="/system/xbin:/system/bin:/sbin:${PATH}"
DEVTREE_PARTITION=/dev/block/mmcblk1p12
KEXEC_DEVTREE=/system/etc/kexec/devtree
#For kexec boot, we need a special devtree from each phone model
#So dump a devtree
#Note /dev/block/mmcblk1p12 is the devtree partition for edison
#And 65K is enough
#Credits for Hashcode for this info
[ -f $KEXEC_DEVTREE ] && rm -r $KEXEC_DEVTREE
busybox dd if=$DEVTREE_PARTITION of=$KEXEC_DEVTREE bs=1 count=65000
I made it as a stand-alone script that I was running with my attempts at booting my kexec builds with BMM... no luck - but like I said, my current kexec build has other issues as well I think..

alteredlikeness said:
Nice (I was kinda hoping/guessing it was you).
Will be very glad to see you digging into this phone also. Let me (or any of us) know if you need any assistance..
And thanks again for your major efforts and contributions in general.
Sent from my MB865 using xda app-developers app
Click to expand...
Click to collapse
Ditto hash, great to see you grabbing a seat on A2train.
ALL ABOARD. NEXT STOP KEXEC. Please raise your trays. Lock your seat back to comfortable position, and enjoy the ride!
Any help I can bring, I am but a humble servant.
Sent from my MB865 using xda app-developers app

Hmm.. take a look at the tombstone.
http://pastebin.com/4jrTaAbF
Its just an issue with graphics driver? But it did boot up a couple of times.. and how does it boot up on some a2s?
Sent from my MB865 using Tapatalk 2

alteredlikeness said:
Well, about the devtree - I found that wang added this to the finalize_release script that initializes the cmdefy bootmenu recovery (courtesy of Hashcode, apparently):
Code:
#!/sbin/sh
PATH="/system/xbin:/system/bin:/sbin:${PATH}"
DEVTREE_PARTITION=/dev/block/mmcblk1p12
KEXEC_DEVTREE=/system/etc/kexec/devtree
#For kexec boot, we need a special devtree from each phone model
#So dump a devtree
#Note /dev/block/mmcblk1p12 is the devtree partition for edison
#And 65K is enough
#Credits for Hashcode for this info
[ -f $KEXEC_DEVTREE ] && rm -r $KEXEC_DEVTREE
busybox dd if=$DEVTREE_PARTITION of=$KEXEC_DEVTREE bs=1 count=65000
I made it as a stand-alone script that I was running with my attempts at booting my kexec builds with BMM... no luck - but like I said, my current kexec build has other issues as well I think..
Click to expand...
Click to collapse
Actually we were talking about this and it doesn't work. The devtree needs to be chopped off right at the last null in the file (where it starts turning to all "FF" in a hex editor.
I think it would be a great idea to modify this tho so that it does work.

Ravikirancg said:
Hmm.. take a look at the tombstone.
http://pastebin.com/4jrTaAbF
Its just an issue with graphics driver? But it did boot up a couple of times.. and how does it boot up on some a2s?
Sent from my MB865 using Tapatalk 2
Click to expand...
Click to collapse
This error reminds me more of when there is no gralloc.omap4.so symlink in /system/vendor/lib/hw. Doesn't look exactly kexec related.

make kexec guestable kernels

hi all,
i'm going to support multiple android roms loading in my kernel_chooser + root_chooser project.
in few words it will be a powerful bootloader for android.
this is what it will be able to do ( many points are yet working ):
kernel loading thought kexec
boot from external devices
boot from subfolders
custom background
change android /sytem partition
change android /data partition
wipe android /cache partition
what i'm asking to you it's to make your android rom's kernel kexec-ready, applying them a kexec guest patch.
thus to make them ready to load for me.
if not i have to patch your kernel every time you modify it.
kernel_chooser and root_chooser are hosted here: https://github.com/tux-mind/tf201-dev
they are made for the asus transformer prime ( TF201 ), but they can run on any android device with your help.
i will add you to the github collaborators if you want to help.
now we are working on multiple android roms support.
thanks in advance for your time.
-- tux_mind

@tux_mind sounds like an interesting idea. So this works for all device's kernels? Once I apply the patch to my kernel source, what happens after that? I'll make sure to follow your progress, good work

HTCDreamOn said:
@tux_mind sounds like an interesting idea. So this works for all device's kernels? Once I apply the patch to my kernel source, what happens after that? I'll make sure to follow your progress, good work
Click to expand...
Click to collapse
thanks
yes, this should work for all devices, except for a little tuning of the partition with the configuration data ( /data ).
i have to make it to self-detect the /data partition.
let's say for example that you want to use it on your HTC Vision.
i have to build an host kernel by the stock one applying the host kexec patch.
than i have to build an android boot image which contains kernel_chooser as initrd and the above kernel.
this android boot image will be written to your current one though fastboot.
after that kernel_chooser will be able to load any kernel patched with the guest patch.
obviously the loaded kernel must be made for that device
it can also load a custom initrd and use a custom kernel CMDLINE.
so, after i made a HTC Vision kernel_chooser, your rom it's ready to load if you applied the guest patch to you kernel.
i can't explain well how to make you rom "bootable" because we are developing this
but if your kernel is kexec-loadable it will be supported by kernel_chooser.
i will update you in the next days
bye!

Really great idea! As it support other devices I can throw in my kernel which is a kexec one for the Motorola RAZR, because of our locked bootloader. The question is: how different are the methods to use kexec?
M.o.t.o.r.o.l.a.R.a.z.r - JBX-Kernel 0.5a - Tapatalk4

@tux_mind this'll sound really stupid but how do actually patch our kernels for this? Do I have to build kexec from here or something?

dtrail1 said:
how different are the methods to use kexec?
Click to expand...
Click to collapse
kexec it's a syscall, so it's the same on every arm device.
HTCDreamOn said:
@tux_mind this'll sound really stupid but how do actually patch our kernels for this? Do I have to build kexec from here or something?
Click to expand...
Click to collapse
there is a good explaination here:
http://forum.xda-developers.com/showthread.php?t=2104706
look at the "Compatibility patch".
for my tf201 the guest patch is this: http://git.lilstevie.geek.nz/?p=ubu...ch;h=54c2e480682afb0629f3854dfea4154f528421e5
which is almost the same...
i hope that almost all kernels have the same host/guest patch, in order to save us a lot of work.
the patch isn't need for kexec, but for hardboot kexec, thus to physically shutdown and restart the device with your kernel, from 0.
the standard kexec it's a assembly jump to 0 ( jmp 0x0 ) with the new kernel loaded in the .text section.
but the standard method don't reinitialize the devices and many other things that could rest in a undefined state.
btw, i have almost done my work.....i'm fighting the android udev which is overwriting my symlinks..
see you!

i got it!
my initial target was to remove devices created by ueventd and replace them with symlink to loop devices.
but android respawn ueventd and replaces my symlinks...
i tried to start the android ueventd, leave it running, and replace the /sbin/ueventd with a infinite sleep static program
but android dont' start at all..i can't even access via adb.
so, the "final" solution is to edit the /fstab.$hardware, which should be in any andoird boot image ( right ? ).
please feel free to suggest any other way to hack the mount process.
you can find the sources here: https://github.com/tux-mind/tf201-dev/tree/master/android_chooser
the program read the paths from kernel cmdline.
the syntax is that:
Code:
newandroid=blkdev:initrd_path:fstab_path
where
blkdev is the linux name of the blockdevice containing the next args ( e.g. /dev/mmcblk0p8 )
initrd_path is the path on the previous blockdevice of the android initrd ( gzipped or not )
fstab_path is the path on the previous blockdevice of the fstab file
the fstab file have this syntax:
Code:
/android/mountpoint /path/to/image/file
where
/android/mountpoint is the android mountpoint to overwrite ( e.g. /system )
/path/to/image/file is the path to the filesystem image file ( on blkdev ) ( e.g. /boot/unrooted.img )
i don't make any documentation until i will be sure that there is no better way to do this.
thanks in advance for your testing
ah, some other useful info.
because we are in testing the program write a log file with some debug info in the root of your blkdev.
so you will find /android_chooser.log on your blkdev with these info in case of errors.
tux_mind

Please
dtrail1 said:
Really great idea! As it support other devices I can throw in my kernel which is a kexec one for the Motorola RAZR, because of our locked bootloader. The question is: how different are the methods to use kexec?
M.o.t.o.r.o.l.a.R.a.z.r - JBX-Kernel 0.5a - Tapatalk4
Click to expand...
Click to collapse
Dtrail could you explain me kexec and its components required for loading new kernel because my device(electrify 2) is similar to droid razr. Using bmm kexec is boot able because I had checked by flashing droid razr's kernel which gave a boot loop.
I would be pleased and thankful to you if you help me.

Requesting Repository Forks

If you wish to request omni forks a repository from AOSP, in order to implement a patch into omni, please post this request here. Eventually this will be carried out via jira (http://jira.omnirom.org)
When posting, please give the path to the repository, and a link to what you want to merge, or a description.
For example if you wish to cherry pick a patch into it, a link to the patch and brief description of the reason for the patch.
We'll then have that repository forked and made available on github and gerrit, permitting you to submit your change.
Before you request, please check on github to ensure it isn't already forked!
For example :
frameworks/base - Needing forked so I can introduce multi window functionality (link here to information)

hardware/broadcom/libbt - Mainly for board-specific configurations (link), but newer CM commits also look promising.
If new device bringups aren't planned yet, ignore the followings:
device/lge/p880 and kernel/lge/p880 - Device bringup (kernel has a non-standard naming at CM repo, it should be android_kernel_lge_x3 (x3 is the board name))
vendor/lge (maybe Omni should create one from scratch)

Adam77Root said:
hardware/broadcom/libbt - Mainly for board-specific configurations (link), but newer CM commits also look promising.
If new device bringups aren't planned yet, ignore the followings:
device/lge/p880 and kernel/lge/p880 - Device bringup (kernel has a non-standard naming at CM repo, it should be android_kernel_lge_x3 (x3 is the board name))
vendor/lge (maybe Omni should create one from scratch)
Click to expand...
Click to collapse
We do have a vendor/lge with just mako, although I'm not sure if its current location.
Vendor blob repos are a legal mess, which is why CM kept them at TheMuppets. We'll likely do something similar if we haven't already (I haven't synced since coming home.) Among other reasons to keep them separate is in the case of a DMCA takedown - usually that happens only to a single repo (see the HTC vendor repos earlier this year) but there's always the risk that damage might be more widespread.

android_external_clang
http://review.cyanogenmod.org/#/c/29799/2
Needed for building my msm8960 device (Motorola Photon Q, xt897c).

arrrghhh said:
android_external_clang
http://review.cyanogenmod.org/#/c/29799/2
Needed for building my msm8960 device (Motorola Photon Q, xt897c).
Click to expand...
Click to collapse
CM bootable recovery and some commits over system core

Black_Prince said:
CM bootable recovery and some commits over system core
Click to expand...
Click to collapse
No for CWM - We're working on forking TWRP in, but some fixes still need to be made to allow it to work on devices with unified recovery/boot image. (Probably one of the biggest holdups for the exynos 4210 family - Sonys with unified boot/recovery can at least use Dees_Troy's FOTAKernel trick).
Getting CWM to build in a tree requires hacks all over the tree in multiple repositories. TWRP is much more self-contained, but needs some tweaks to the build system in order to get it to build properly on a userdebug (as opposed to eng) build.

Entropy512 said:
No for CWM - We're working on forking TWRP in, but some fixes still need to be made to allow it to work on devices with unified recovery/boot image. (Probably one of the biggest holdups for the exynos 4210 family - Sonys with unified boot/recovery can at least use Dees_Troy's FOTAKernel trick).
Getting CWM to build in a tree requires hacks all over the tree in multiple repositories. TWRP is much more self-contained, but needs some tweaks to the build system in order to get it to build properly on a userdebug (as opposed to eng) build.
Click to expand...
Click to collapse
Hmm but until the twrp is ported should i test build for some galaxy devices like SG S2 note 1 with cwm ?
It will make me to push changes to git in device tree ... and will be easy to add support

Black_Prince said:
Hmm but until the twrp is ported should i test build for some galaxy devices like SG S2 note 1 with cwm ?
It will make me to push changes to git in device tree ... and will be easy to add support
Click to expand...
Click to collapse
I'll pull up a remote to use for bootable/recovery until we fork (unless plodey already forked it) from TWRP, won't be able to provide it until tonight.
However, right now, it will only build a working recovery if you do an eng build instead of userdebug. If Dees_Troy's talk at BABBQ is ever posted, that's what I was talking about when I trolled him at the end.
Something like:
Code:
. build/envsetup.sh && brunch custom_n7000-eng
or something like that.

Old
def WriteRawImage(self, mount_point, fn):
"""Write the given package file into the partition for the given
mount point."""
fstab = self.info["fstab"]
if fstab:
p = fstab[mount_point]
partition_type = common.PARTITION_TYPES[p.fs_type]
args = {'device': p.device, 'fn': fn}
if partition_type == "MTD":
self.script.append(
'write_raw_image(package_extract_file("%(fn)s"), "%(device)s");'
% args)
elif partition_type == "EMMC":
self.script.append(
'package_extract_file("%(fn)s", "%(device)s");' % args)
Requested
def WriteRawImage(self, mount_point, fn):
"""Write the given package file into the partition for the given
mount point."""
fstab = self.info["fstab"]
if fstab:
p = fstab[mount_point]
partition_type = common.PARTITION_TYPES[p.fs_type]
args = {'device': p.device, 'fn': fn}
if partition_type == "MTD":
self.script.append(
'package_extract_file("%(fn)s", "/tmp/boot.img");'
'write_raw_image("/tmp/boot.img", "%(device)s");' % args
% args)
elif partition_type == "EMMC":
self.script.append(
'package_extract_file("%(fn)s", "%(device)s");' % args)
elif partition_type == "BML":
self.script.append(
('assert(package_extract_file("%(fn)s", "/tmp/%(device)s.img"),\n'
' write_raw_image("/tmp/%(device)s.img", "%(device)s"),\n'
' delete("/tmp/%(device)s.img"));') % args)
else:
raise ValueError("don't know how to write \"%s\" partitions" % (p.fs_type,))
Flash error in write raw image ...
Also can u cherry pick these commits https://github.com/CyanogenMod/android_system_core/commits/cm-10.2/mkbootimg/unpackbootimg.c
My internal SD card isnt got recognsied , there must be some commits missing in vold ...
I dont even know it got formatted or what in flashing ...
All things working WIFI BLUTOOTH TETHERING CALL DATA MSG MMS OMNITORCH
Also u may have look over https://github.com/CyanogenMod/andr...mmit/f22626cdbecfe27c96f205710173458eab14e1a1
and https://github.com/CyanogenMod/andr...mmit/aee5f78ddec238cec016849acaf1d3007b8b1507

https://android.googlesource.com/platform/hardware/broadcom/libbt
Must haz patch:
https://github.com/CyanogenMod/andr...mmit/4028192a4fe8289bf9b7fb202a2bb2c0d85d98c3
//edit: xplod sorted it out already.

android_packages_apps_PackageInstaller
For
HALO and a feat which enables us to see version of the installed and gonna install app

I need this to make it work with my Samsung galaxy s4

I'm not sure if this is the correct thread, but the following two would be great for i9305 support
smdk4412-qcom-common -> needed for i9305 and t0lte (EDIT: I see this one has now been added)
Thanks!

Hey @pulser_g2,
wondering if i could have the following repositories added in for jf devices,
all features are fully functional with these repositories : branch: cm-10.2 for all
device_samsung_jf-common
android_device_samsung_jflteatt
android_kernel_samsung_jf
CyanogenMod/android_hardware_samsung
CyanogenMod/android_device_samsung_msm8960-common
CyanogenMod/android_device_samsung_qcom-common

@pulser_g2
I just talked to @maxwen and he told me i should post here the needed stuff:
I managed to fix the radio issues today on m7, but i need additionally these commits:
https://github.com/TheMuppets/proprietary_vendor_htc/commit/7bbbe83823525090e7d94c47b439c09a254b380f
https://github.com/TheMuppets/proprietary_vendor_htc/commit/19ed957678d13b489da78d704b27811030688078
https://github.com/TheMuppets/proprietary_vendor_htc/commit/51b006bc8f05757c7dd82361ce6d985e996af360
and please DO NOT merge this: https://github.com/TheMuppets/proprietary_vendor_htc/commit/2b86687a460030199fa42d49207957ce285c1a53
as its not needed and breaks build (I allready have everything in m7-common so no need for vendor)
Thanks in advance

@pulser_g2
we dont need to update anything, please forget my post above, i managed to get it working without all these commits and the right vendor

hardware/qcom/gps - Needing forked so I can build the needed GPS module into my Omnirom build.
The link to this repository is: https://github.com/CyanogenMod/android_hardware_qcom_gps/tree/cm-10.2
I hope it gets included, this saves me a ton of time pulling stuff from other repositories.
Greetings PsychoGame

Galaxy S I9000 (galaxysmtd)
@pulser_g2 Device repos are already forked. Kernel is missing: https://github.com/OmniGalaxyS/android_kernel_samsung_aries
It has android-4.3 and android-4.4 branches up.
Device rocking with Omni!
See you!

Fail

PsychoGame said:
hardware/qcom/gps - Needing forked so I can build the needed GPS module into my Omnirom build.
The link to this repository is: https://github.com/CyanogenMod/android_hardware_qcom_gps/tree/cm-10.2
I hope it gets included, this saves me a ton of time pulling stuff from other repositories.
Greetings PsychoGame
Click to expand...
Click to collapse
I also need android_hardware_qcom_gps (probably best from https://github.com/markcs/android_hardware_qcom_gps/tree/cm-11.0)

[ROM][CM13] Cyanogenmod 13 for Samsung Galaxy SL i9003

hi,
this is my attempt to bring Cyanogen 13 to the Samsung Galaxy SL ( i9003 ).
So far i got the kernel/device/vendor repos synced and edited.
I still haven't manage to pass the "brunch galaxysl" command
first error was:
hardware/ti/wlan/mac80211/wpa_supplicant_lib/Android.mk:19: external/wpa_supplicant_8_ti/wpa_supplicant/android.config: No such file or directory
fixed by editing BoardConfing.mk ( not sure it will allow the wi-fi to work after this )
last error is :
make: *** No rule to make target '/home/nots/cm13/out/target/product/galaxysl/utilities/flash_image', needed by '/home/nots/cm13/out/target/product/galaxysl/bootstrap.cpio.gz'. Stop.
If anyone want to help, you are more than welcome
thanks to dhiru1602 and rodero95 for there previous work on this phone

mkdir out/target/product/galaxysl/utilities
copy busybox erase_image flash_image make_ext4fs to the dir created
busybox erase_image flash_image make_ext4fs are from ROM pack

notsgnik said:
hi,
this is my attempt to bring Cyanogen 13 to the Samsung Galaxy SL ( i9003 ).
So far i got the kernel/device/vendor repos synced and edited.
I still haven't manage to pass the "brunch galaxysl" command
first error was:
hardware/ti/wlan/mac80211/wpa_supplicant_lib/Android.mk:19: external/wpa_supplicant_8_ti/wpa_supplicant/android.config: No such file or directory
fixed by editing BoardConfing.mk ( not sure it will allow the wi-fi to work after this )
last error is :
make: *** No rule to make target '/home/nots/cm13/out/target/product/galaxysl/utilities/flash_image', needed by '/home/nots/cm13/out/target/product/galaxysl/bootstrap.cpio.gz'. Stop.
If anyone want to help, you are more than welcome
thanks to dhiru1602 and rodero95 for there previous work on this phone
Click to expand...
Click to collapse
Yeah I've actually followed quite some hack arounds for the device...... You'll need patches for the following repos (Unfortunately no longer in Nameless Gerrit. I'll try to upload them):
device/samsung/galaxysl
kernel/samsung/latona
hardware/ti/omap3
hardware/samsung
frameworks/av
bionic
I guess that's all that I've patched. And one more word of advice: CyanogenMod dropped support for the S2 so it'll never compile for the SL. It's better to download NamelessROM's source (check repo's --reference flag to speed up the download) because it has some support.
Cheers!
Technohacker

Technohacker1995 said:
Yeah I've actually followed quite some hack arounds for the device...... You'll need patches for the following repos (Unfortunately no longer in Nameless Gerrit. I'll try to upload them):
device/samsung/galaxysl
kernel/samsung/latona
hardware/ti/omap3
hardware/samsung
frameworks/av
bionic
I guess that's all that I've patched. And one more word of advice: CyanogenMod dropped support for the S2 so it'll never compile for the SL. It's better to download NamelessROM's source (check repo's --reference flag to speed up the download) because it has some support.
Cheers!
Technohacker
Click to expand...
Click to collapse
plz give us link for download and install rom

mehran200067 said:
plz give us link for download and install rom
Click to expand...
Click to collapse
For which version? I've compiled 5.0 (here: http://forum.xda-developers.com/gal...ake-android-5-0-lollipop-i9003-t2928989/page6 I've posted some fixes in subsequent posts so read all of 'em), 5.1's not available (doesn't boot...... idk why) whereas we're trying to port 6
Cheers!
Technohacker

Building kernel from Xiaomi's source for cereus (probably, cactus too)

Since Xiaomi released source code for kernel, I tried to build it.
https://github.com/Dhoine/android_kernel_cereus
Current status of this - it compiles, boots, but unusable.
Bugs:
Broken bluetooth
Broken wifi
Soft-reboots after some time
Propably, more bugs that remained undiscovered because of reboots.
Tested with miui 9 on cereus 3/32.
--------
You can test builded kernel using this:
GDrive
Use Carliv image Kitchen to repack your rom's boot.img, then flash it via fastboot (UNLOCKED BOOTLOADER REQUIRED). To restore working state after testing, flash stock boot image again.
WARNING
Our kernel uses android regular boot.img format, not MTK.
---------
If you want to build this, any manual for building kernel will fit. I used google's android arm eabi 4.9 toolchain, our arch is arm, defconfig is cereus_defconfig.
For now, I don't have enough time to take a closer look at this (work, university exams), so, any help (last_kmsg, dmesg, pull requests) will be appreciated.

Wait seriously? Does that mean that i can buil LOS rom now? I have my UbuntuVM with whole LOS source code downloaded but i got stuck at no kernel found for cereus. So it might work now(Im extremely new to linux) I'll try to build the rom asap

TanRayCz said:
Wait seriously? Does that mean that i can buil LOS rom now? I have my UbuntuVM with whole LOS source code downloaded but i got stuck at no kernel found for cereus. So it might work now(Im extremely new to linux) I'll try to build the rom asap
Click to expand...
Click to collapse
This kernel for now won't really help with building los. I haven't find any device tree for our SoC (or close). MTK always requre a ton of patches to android sources. But i have reference MTK 9.0 android sources for this SoC - ALPS (can't share, it's probably illegal, but you can find them on 4pda.ru, if you want). It builds on almost unmodified reference device tree, but i haven't tested it on device. First, we need to make this kernel usable.

[email protected] said:
This kernel for now won't really help with building los. I haven't find any device tree for our SoC (or close). MTK always requre a ton of patches to android sources. But i have reference MTK 9.0 android sources for this SoC - ALPS (can't share, it's probably illegal, but you can find them on 4pda.ru, if you want). It builds on almost unmodified reference device tree, but i haven't tested it on device. First, we need to make this kernel usable.
Click to expand...
Click to collapse
This may not really help, but can you try that kernel again after upgrading to MIUI 10.2.4 ?
EDIT : See https://www.kimovil.com/en/list-smartphones-by-processor/mediatek-helio-p22-mt6762 for a list of devices with MT6762. Probably we can check out their trees?
EDIT 2 : It seems like our Redmi 6 is the only device with available kernel source in that list.
EDIT 3 : Try this?

minhducsun2002 said:
EDIT : See https://www.kimovil.com/en/list-smar...lio-p22-mt6762 for a list of devices with MT6762. Probably we can check out their trees?
EDIT 2 : It seems like our Redmi 6 is the only device with available kernel source in that list.
Click to expand...
Click to collapse
Maybe them exists, but I haven't found them.

minhducsun2002 said:
This may not really help, but can you try that kernel again after upgrading to MIUI 10.2.4 ?
---------------------
EDIT 3 : Try this?
Click to expand...
Click to collapse
I finally finished my university exams, so now I have some free time to play with it.
The kernel you link is just forked xiaomi sources merged to linux 4.9 kernel. It shouldn't be buildable at all, since there is the same problem in sources I "fixed" with this.
Probably, this too.

[email protected] said:
I finally finished my university exams, so now I have some free time to play with it.
The kernel you link is just forked xiaomi sources merged to linux 4.9 kernel. It shouldn't be buildable at all, since there is the same problem in sources I "fixed" with this.
Probably, this too.
Click to expand...
Click to collapse
[email protected] said:
Maybe them exists, but I haven't found them.
Click to expand...
Click to collapse
You're right, they're identical.
By the way, I couldn't find any implementation available for our Helio P22; the chipset itself is fairly new, thus not many devices have the source code available.

minhducsun2002 said:
This may not really help, but can you try that kernel again after upgrading to MIUI 10.2.4 ?
Click to expand...
Click to collapse
Ok, i've tested it on latest GS. It seems to work a way better. At least, it doesnt reboot after few seconds of usage. It makes things a lot easier.
Tested a few more things:
-sound - works
-camera (both photo and video) - works
- mobile data - works
- phone calls - work
- dual sim -works
- all sensors work
The only thing i was unable to test is gps. It didn't worked for me, but i was inside 9-floor building near the window to inner corner of building... So idk.
And usb otg - since i don't have adapder.
After this I can tell, that there are only 2 major bugs - wifi and bluetooth.
Maybe gps, fm radio and exfat too. This is modules loading problem, I think (though, there shouldn't be any modules, drivers are marked as built-in in config, but .ko files for them exist in /vendor/lib/modules)

[email protected] said:
Ok, i've tested it on latest GS. It seems to work a way better. At least, it doesnt reboot after few seconds of usage. It makes things a lot easier.
Tested a few more things:
-sound - works
-camera (both photo and video) - works
- mobile data - works
- phone calls - work
- dual sim -works
- all sensors work
The only thing i was unable to test is gps. It didn't worked for me, but i was inside 9-floor building near the window to inner corner of building... So idk.
And usb otg - since i don't have adapder.
After this I can tell, that there are only 2 major bugs - wifi and bluetooth.
Maybe gps, fm radio and exfat too. This is modules loading problem, I think (though, there shouldn't be any modules, drivers are marked as built-in in config, but .ko files for them exist in /vendor/lib/modules)
Click to expand...
Click to collapse
Yeah, Treble definitely makes things easier.
Nevertheless I'm trying to integrate those kernel modules into the compiled zImage - there's no modprobe in Android, sadly.

Adding modules support and disabling modules versions check in config didn't help.
I can't find any source files in kernel sources, from which these modules can be built. Maybe, they are built from separate platform code (they are placed in vendor, so...) Need to download ALPS (brrr, around 50 gb of tar.bz2 archives) to check it.

I've looked into reference mediatek 4.9 kernel sources and found this:
Code:
# Do build-in for Makefile checking
# export CONFIG_WLAN_DRV_BUILD_IN=y
ifeq ($(CONFIG_WLAN_DRV_BUILD_IN),y)
PATH_TO_WMT_DRV = vendor/mediatek/kernel_modules/connectivity/common
PATH_TO_WLAN_CHR_DRV = vendor/mediatek/kernel_modules/connectivity/wlan/adaptor
PATH_TO_WLAN_DRV = vendor/mediatek/kernel_modules/connectivity/wlan/core/gen4m
ABS_PATH_TO_WMT_DRV = $(srctree)/../$(PATH_TO_WMT_DRV)
ABS_PATH_TO_WLAN_CHR_DRV = $(srctree)/../$(PATH_TO_WLAN_CHR_DRV)
ABS_PATH_TO_WLAN_DRV = $(srctree)/../$(PATH_TO_WLAN_DRV)
# check wlan driver folder
ifeq (,$(wildcard $(ABS_PATH_TO_WMT_DRV)))
$(error $(ABS_PATH_TO_WMT_DRV) is not existed)
endif
ifeq (,$(wildcard $(ABS_PATH_TO_WLAN_CHR_DRV)))
$(error $(ABS_PATH_TO_WLAN_CHR_DRV) is not existed)
endif
ifeq (,$(wildcard $(ABS_PATH_TO_WLAN_DRV)))
$(error $(ABS_PATH_TO_WLAN_DRV) is not existed)
endif
$(warning symbolic link to $(PATH_TO_WMT_DRV))
$(warning symbolic link to $(PATH_TO_WLAN_CHR_DRV))
$(warning symbolic link to $(PATH_TO_WLAN_DRV))
$(shell unlink $(srctree)/$(src)/wmt_drv)
$(shell unlink $(srctree)/$(src)/wmt_chrdev_wifi)
$(shell unlink $(srctree)/$(src)/wlan_drv_gen4m)
$(shell ln -s $(ABS_PATH_TO_WMT_DRV) $(srctree)/$(src)/wmt_drv)
$(shell ln -s $(ABS_PATH_TO_WLAN_CHR_DRV) $(srctree)/$(src)/wmt_chrdev_wifi)
$(shell ln -s $(ABS_PATH_TO_WLAN_DRV) $(srctree)/$(src)/wlan_drv_gen4m)
# for gen4m options
export CONFIG_MTK_COMBO_WIFI_HIF=axi
export MTK_COMBO_CHIP=CONNAC
export WLAN_CHIP_ID=6765
export MTK_ANDROID_WMT=y
# Do build-in for xxx.c checking
subdir-ccflags-y += -D MTK_WCN_REMOVE_KERNEL_MODULE
subdir-ccflags-y += -D MTK_WCN_BUILT_IN_DRIVER
obj-y += wmt_drv/
obj-y += wmt_chrdev_wifi/
obj-y += wlan_drv_gen4m/
endif
No such code in xiaomi's source. And this prooves my words: these modules CAN BE BUILT IN and SOURCES FOR THEM ARE IN ALPS. I have to download this ton of ****...
-------
I tried to backpors drivers from 9.0 - no success. It compiles,but doesn't boot. I tried to hack modules loading - no success too.

[email protected] said:
Adding modules support and disabling modules versions check in config didn't help.
I can't find any source files in kernel sources, from which these modules can be built. Maybe, they are built from separate platform code (they are placed in vendor, so...) Need to download ALPS (brrr, around 50 gb of tar.bz2 archives) to check it.
Click to expand...
Click to collapse
Most of the time those binaries are not open-source (you probably know vendor blobs for Google devices are too, but that is a different story). Nevertheless I don't really get why kernel modules are placed in
Code:
/vendor
- after all, mobile devices don't have hot-plugging hardware that requires dynamic LKM loading, why do they separate them out?
I'll attempt to crawl the web for those sources while waiting for Mi Unlock to allow my device being unlocked (the waiting period sucks).

[email protected] said:
I've looked into reference mediatek 4.9 kernel sources and found this:
No such code in xiaomi's source. And this prooves my words: these modules CAN BE BUILT IN and SOURCES FOR THEM ARE IN ALPS. I have to download this ton of ****...
-------
I tried to backpors drivers from 9.0 - no success. It compiles,but doesn't boot. I tried to hack modules loading - no success too.
Click to expand...
Click to collapse
1. Backports? How?
2. In which file did you got those lines?
3. Indeed those modules' source are removed from our tree - if you have the link, I don't really mind having another 50GB occupied on my PC to get those source code.

minhducsun2002 said:
1. Backports? How?
2. Indeed those modules' source are removed from our tree - if you have the link, I don't really mind having another 50GB occupied on my PC to get those source code.
Because attached below is the existence of the modules. /shrug
Click to expand...
Click to collapse
I already said that i have leaked mtk alps 9.0 sources These modules have sources, they are not prebuilt. I managed to built in them, but kernel doesn't boot at all after that. And i can't take any logs because the system dies before adb init.

Ok, after one little, but very cruel hack modules are fixed. Enjoy testing the kernel.
https://drive.google.com/open?id=1Z3p2fAWOZFyp045QMNV6vhLZvj9ZZPSY

[email protected] said:
Ok, after one little, but very cruel hack modules are fixed. Enjoy testing the kernel.
Click to expand...
Click to collapse
does it work well?

Lonewolf_1210 said:
does it work well?
Click to expand...
Click to collapse
Users from 4pda reported it fully working.
---------------------
Soon will be build with usb network adapters support enabled (requested by romanxdream from 4pda).
P.S. I'm not adding everything requested to kernel. But this was only config modification.
-------
I don't check xda often, so there can be a huge delay in my responses.

[email protected] said:
Users from 4pda reported it fully working.
---------------------
Soon will be build with usb network adapters support enabled (requested by romanxdream from 4pda).
P.S. I'm not adding everything requested to kernel. But this was only config modification.
-------
I don't check xda often, so there can be a huge delay in my responses.
Click to expand...
Click to collapse
Great job then, hope for a custom rom coming soon

Ok, download link to all versions. I'll edit the first post later.
https://drive.google.com/open?id=1yEZRS8L8bPgkk58tT5Uv2b-vsLqk7aql

Should we make a telegram group for this?