Database Develop

[Guide]USB Host Enabled Kernel for the G1/Dream

HOWTO: Compile USB Host Enabled Kernel for the G1/Dream
Hi everyone,
I've got my new HTC Dream and I love it. It's my first Android based device. Now I'm running CyanogenMod 6 Rom.
So this is what I have:
Model number: HTC Dream
Andreoid ver: 2.2
Kermel ver: 2.6.34.5-cynogenmod
Mod version: CyanodenMod-6.0.0-DS
Build number: FRF91
I followed these guides to compile usb host enabled kernel. Please thank Andrew de Quincey, Sven Killig , Michael Mitchell and Frank Sposaro for their awesome work.
Preparing Android Build Environment
mkdir ~/cm-kernel-usbhost/
cd ~/cm-kernel-usbhost/
git clone git://github.com/CyanogenMod/cm-kernel.git
cd cm-kernel
git branch -r
git checkout --track -b cm-2.6.34 origin/android-msm-2.6.34
Patch the kernel source tree
wget http://sven.killig.de/android/N1/2.2/usb_host/android-kernel_msm-bca5320_Nexus-One_usb-host.patch
patch -p1 < android-kernel_msm-bca5320_Nexus-One_usb-host.patch
open ~/cm-kernel-usbhost/ cm-kernel /drivers/usb/host/ehci-msm7201.c and delete line 313
Export a variable for the ARM cross compiler
export CCOMPILER=~/android-ndk-r4b/build/prebuilt/linux-x86/arm-eabi-4.4.0/bin/arm-eabi-
Download kernel .config here and move to .config
Edit!: Use attached config file!
mv g1-cm-kern-2.6.34-usbhost.config .config
Then start the build:
make ARCH=arm CROSS_COMPILE=$CCOMPILER
At this point you should have a kernel stored in ~/cm-kernel-usbhost/cm-kernel/arch/arm/boot/zImage
Now, get the kernel on the phone:
I use fastboot to boot from RAM (not persistent)
~/android-sdk-linux_86/tools/fastboot boot arch/arm/boot/zImage
Look above at the output from the kernel build. The last thing completed is the module building (total = modpost). Notice the hierarchal structure of the compiled modules within the filesystem. Ex: /system/drivers/usb/core/usbcore.ko
Copy modules to sdcard
Now we need to load the modules into the running kernel.
Open a terminal emulator on the phone (connectbot also works, connect to local)
get root:
su
then load the usb core and ehci modules:
insmod /sdcard/usbcore.ko
insmod /sdcard/ehci-hcd.ko
You can download compiled kernel and modules here.
Hope that this guide will be helpful to newbies like me.
Regards,
Andy
NB!
This kernel currently disables USB mount and adb over USB.
usb-storage.ko module is loaded fine but when a usb flash drive is pluged-in the phone reboots itself.
You need a powered hub or Y-cable and USB A Female / A Female Adaptor to connect your usb devices.
Now wifi is working!
UPDATE: 11.12.2010
How to make an OTA package and switch between usb host and otg.
wget ww2.cs.fsu.edu/~mitchell/android/ZipBuilder.tar
tar -xvf ZipBuilder.tar
cd ZipBuilder
./builder.sh
rename update_signed.zip usbhost_signed.zip and copy it to sd card.
make a Nandorid backup and split boot.img using split_bootimg.pl
rename boot.img-kernel to zImage and copy it to ~/cm-kernel-usbhost/cm-kernel/arch/arm/boot/
than run ./builder.sh again
rename update_signed.zip to otg_signed.zip and copy it to sd card.
Now you can flash these zip from recovery.

Any luck with Sven's img?

Yes, I tried to boot Sven's zImage but without success.
So now I'm trying to compile android-msm-2.6.34 kernel pathced with Sven's patch file.
The patch was successfully applied but when I started compilling this error occured:
In file included from drivers/usb/host/ehci-hcd.c:1153:
drivers/usb/host/ehci-msm7201.c: In function 'usb_hcd_msm7201_probe':
drivers/usb/host/ehci-msm7201.c:313: error: 'struct msm_hsusb_platform_data' has no member named 'phy_shutdown'
I don't know how and where to add this member.
I'd appreciate some help.
Regards Andy

kodovoimeji said:
In file included from drivers/usb/host/ehci-hcd.c:1153:
drivers/usb/host/ehci-msm7201.c: In function 'usb_hcd_msm7201_probe':
drivers/usb/host/ehci-msm7201.c:313: error: 'struct msm_hsusb_platform_data' has no member named 'phy_shutdown'
Click to expand...
Click to collapse
This seems to be defined only in the Hero Kernel. Delete line 313.

Thank you very much for quick reply!!!
Now everything is OK!
I used fastboot to boot from RAM and then load the usb core and ehci modules.
And...yeeeeeeee now my G1 is usb host device!
I'm using Y-cable from my external HDD.
I must compile some modules like usb-storage and start testing.
My usb mouse works great with usbhid.ko.
Thank you again!
P.S. I'll update the first post soon!

That's Great~! I've found this for several months! Could you post your detailed tutorial for us? I think most G1 owner should want it. Maybe, you could make a special patch for the roms of CM 6, and everyone could update kernel more effortlessly.
Cheers!
Fox

I updated the first post and now you can download compiled kernel-2.6.34 and modules.
I tried usb-storage.ko but when I pluged my usb flash drive the phone reboot itself.

kodovoimeji said:
I tried usb-storage.ko but when I pluged my usb flash drive the phone reboot itself.
Click to expand...
Click to collapse
Is there something in /data/dontpanic directly after the reboot?

sonic74 said:
Is there something in /data/dontpanic directly after the reboot?
Click to expand...
Click to collapse
No, it's empty.
I load these modules: usbcore.ko, ehci-hcd.ko, usb-libusual.ko, usb-storage.ko.
Usb-storage.ko depends on usb-libusual.ko
I'm curious if it's possible to run my ACR122 contactless smartcard reader.
I installed Debian Leny and ACR drivers on sdcard as I do on my PC.
This is the output when I run dmesg:
[ 196.719024] msm_hsusb msm_hsusb: GetStatus port 1 status 80001803 POWER sig=j CSC CONNECT
[ 196.719573] hub 1-0:1.0: port 1: status 0101 change 0001
[ 196.819427] hub 1-0:1.0: state 7 ports 1 chg 0002 evt 0000
[ 196.820007] hub 1-0:1.0: port 1, status 0101, change 0000, 12 Mb/s
[ 196.939056] usb 1-1: new full speed USB device using msm_hsusb and address 2
[ 197.089111] usb 1-1: ep0 maxpacket = 8
[ 197.093750] usb 1-1: skipped 1 descriptor after interface
[ 197.095733] usb 1-1: default language 0x0409
[ 197.098724] usb 1-1: udev 2, busnum 1, minor = 1
[ 197.099426] usb 1-1: New USB device found, idVendor=072f, idProduct=90cc
[ 197.100280] usb 1-1: New USB device strings: Mfr=1, Product=2, SerialNumber=0
[ 197.100769] usb 1-1: Product: CCID USB Reader
[ 197.101593] usb 1-1: Manufacturer:
[ 197.106353] usb 1-1: usb_probe_device
[ 197.106872] usb 1-1: configuration #1 chosen from 1 choice
[ 197.108795] usb 1-1: adding 1-1:1.0 (config #1, interface 0)
[ 197.116180] drivers/usb/core/inode.c: creating file '002'
[ 197.117492] hub 1-0:1.0: state 7 ports 1 chg 0000 evt 0002
Click to expand...
Click to collapse
But I can't see it when I run lsusb.
This is the output of pcscdaemon:
00000000 pcscdaemon.c:280:main() pcscd set to foreground with debug send to stderr
00000336 debuglog.c:239: DebugLogSetLevel() debug level=debug
00018596 pcscdaemon.c:498:main() pcsc-lite 1.4.102 deamon ready
01416524 hotplug_libhal.c:460:HPRegisterForHotplugEvents() error: dbus_bus_get: org.freedesktop.DBus.Error.NoServer: Failed to connect to socket /var/run/dbus/system_bus_socket: Connection refused
00000244 pcscdaemon.c:517:main() SVCServiceRunLoop returned
00000091 pcscdaemon.c523:at_exit() cleaning /var/run/pcscd
Click to expand...
Click to collapse

kodovoimeji said:
Usb-storage.ko depends on usb-libusual.ko
Click to expand...
Click to collapse
Strange, I haven't seen this module yet.
kodovoimeji said:
But I can't see it when I run lsusb.
Click to expand...
Click to collapse
You have to do
Code:
mount -t usbfs none /proc/bus/usb

Hi again,
thanks for your help. It seems that usb-storage.ko depends on usb-libusual.ko only when this modul is selected. I compiled kernel with build-in usb-storage but no matter i do my phone reboots after I pluged-in a usb flash storage. /data/dontpanic is empty.
Code:
mount -t usbfs none /proc/bus/usb
Now I can use lsusb.
I looked at .config but didn't find something about smartcard reader so if I manage to sort out drivers in Debian will it start working?
Kind regards
Andy

Please reupload link to kernel and modul.

http://ifile.it/ekountp/G1-usbhost-kern-and-modul-2.6.34.rar

Hi, newbie here. I was wondering whether once you have this setup, you can read read raw data directly from the usb port like you could in linux via /dev/hidraw2 or something similar?

Hello again,
why doesn't the wifi module work with this kernel?

is there any body going well of this?
my G1 detected the keyboard and my external HDD, I input the command 'lsusb' and 'dmesg', it seems working well; but a few moment later, it gives the message like this "can't enumerate usb devices in port 1"...what's more, I'm a noob with kernel and driver, anybody who can tell me how could I access the HDD and make the keyboard working, thans

Hi again,
at first I used Michael Mitchell's config file but than I decided to make my own config file:
1.Retrieve a working kernel config from your phone:
adb pull /proc/config.gz
gunzip config.gz
mv config .config
2.Customize the kernel:
make ARCH=arm CROSS_COMPILE=$CCOMPILER menuconfig
And now with my config file wifi is working.
There are CONFIG_USB_HIDDEV and CONFIG_HIDRAW options. I have basic skills in Linux so I can't help you.
I tried to compile USB Gadget Support as a module but receive this error:
Code:
CC [M] drivers/usb/gadget/msm72k_udc.o
drivers/usb/gadget/msm72k_udc.c: In function 'usb_phy_reset':
drivers/usb/gadget/msm72k_udc.c:1017: warning: unused variable 'retries'
drivers/usb/gadget/msm72k_udc.c:1016: warning: unused variable 'ret'
drivers/usb/gadget/msm72k_udc.c:1015: warning: unused variable 'val'
drivers/usb/gadget/msm72k_udc.c: At top level:
drivers/usb/gadget/msm72k_udc.c:1299: error: redefinition of 'msm_hsusb_set_vbus_state'
arch/arm/mach-msm/include/mach/board.h:121: note: previous definition of 'msm_hsusb_set_vbus_state' was here
make[2]: *** [drivers/usb/gadget/msm72k_udc.o] Error 1
make[1]: *** [drivers/usb/gadget] Error 2
make: *** [drivers] Error 2

hello ;D
Could i get a 2708 (15mb hack kernel mod) kernel with usb host support,please?;>
thanks for work,and congratulations for great idea!

Yeah I want too 2708 kernel with USB Host

Hi guys, I'm trying to get this working on the latest version of CyanogenMod 6 but am having troubles... has anyone else done this who fancies helping me out?
thanks!

[DEV TOOLS] - Pack & Repack Boot.img Files - for Windows USERS

Plagiarism will not be tolerated on XDA-Developers:
http://forum.xda-developers.com/showpost.php?p=35640046&postcount=36

Reserved - Possible Updates Using Cygwin

I'm gonna take this one just in case as well

Thx for the app but i have an empty ramdisk folder...and all of the ramdisk files can't created or newer or same age version exists...cydwin won't let me copy the log...im using huawei not SE phone..if it has something to do!! But i can see the contents at boot.img-ramdisk.gz

Thanks but please mirror to other filehoster because mf is buggy on my phone
Sent from my WT19i using xda premium

$ ./extractboot boot.img
./extractboot: ./split_bootimg.pl: /usr/bin/perl: bad interpreter: Permission denied
./extractboot: line 17: cpio: command not found
gzip: ../../boot.img-ramdisk.gz: No such file or directory
what i missed here?
edit: opps. missed the REQUIRED packages. testing again.

$ ./extractboot boot.img
Android Magice not found in ./boot.img/ Giving up.
gzip: ../../boot.img-ramdisk.gz: unexpected end of file
cpio: premature end of archive
Any help with this error? I am trying to split and edit a custom CM9/ICS boot.img

very useful tool, thanks

This isn't working for me on my htc evo 4g lte. I am unpacking the stock boot.img and then repacking it with NO changes and get bootloops.

your image png should not exceed 20KB and your logo.rle should not exceed 60kb Otherwise you will almost always bootloop

stef67000 said:
your image png should not exceed 20KB and your logo.rle should not exceed 60kb Otherwise you will almost always bootloop
Click to expand...
Click to collapse
I didn't change any pngs, logos, or anything. I simply unpacked, repacked, and flashed.

maxdamage2122 said:
I didn't change any pngs, logos, or anything. I simply unpacked, repacked, and flashed.
Click to expand...
Click to collapse
Check on the way your kernel should be build.. Ours use as base 0x20 but yours may change..
Sent from my LT18i using Tapatalk

jimbo77 said:
Check on the way your kernel should be build.. Ours use as base 0x20 but yours may change..
Sent from my LT18i using Tapatalk
Click to expand...
Click to collapse
Is there a way to check without the source? Source for our kernel isn't released yet.

I've tried to unpack the last boot.img from cm9 RC2 to get the new cwm 6.x but I have an issue
Code:
$ ./extractboot boot.img
Page size: 2048 (0x00000800)
Kernel size: 3229308 (0x0031467c)
Ramdisk size: 1789254 (0x001b4d46)
Second size: 0 (0x00000000)
Board name:
Command line:
Writing boot.img-kernel ... complete.
Writing boot.img-ramdisk.gz ... complete.
./extractboot: line 17: cpio: command not found

I use this way http://forum.xda-developers.com/showthread.php?t=1477845
how to change spash screen http://forum.xda-developers.com/showthread.php?t=1140406

You need to download the cpio package through cygic in order to have it fixed..
Sent from my SGS3... Yes I have it!!!!

jimbo77 said:
You need to download the cpio package through cygic in order to have it fixed..
Sent from my SGS3... Yes I have it!!!!
Click to expand...
Click to collapse
thanks

thomas.raines said:
$ ./extractboot boot.img
Android Magice not found in ./boot.img/ Giving up.
gzip: ../../boot.img-ramdisk.gz: unexpected end of file
cpio: premature end of archive
Any help with this error? I am trying to split and edit a custom CM9/ICS boot.img
Click to expand...
Click to collapse
Hello,I'm having the same error with CM9/10 ICS/JellyBean kernel boot.img,there are chances it comes from the kernel itself as it appears to be or should we better look into our configuration?

helo, i repacket doomlord kernel v22 and when im trying to flash it i got this error :
FAILED (remote: Wrong range:0x10008000-0x103202d7. Ok:0x00200000 - 0x03dfff)
It's somethink about size of kernel and "fake size maker" file, but i dont know how to fix it.

So I pulled a working one apart and replaced the kernel with the compiled one.
Code:
[email protected] ~
$ cd /boot
[email protected] /boot
$ ./extractboot boot.img
Page size: 2048 (0x00000800)
Kernel size: 4463500 (0x00441b8c)
Ramdisk size: 372027 (0x0005ad3b)
Second size: 0 (0x00000000)
Board name:
Command line: console=ttyHSL0,115200,n8 androidboot.hardware=qcom loglevel=0
Writing boot.img-kernel ... complete.
Writing boot.img-ramdisk.gz ... complete.
2881 blocks
[email protected] /boot
$ ./packboot
Using base address 0x40200000
Using command line 'console=ttyHSL0,115200,n8 androidboot.hardware=qcom loglevel=0'
No errors
I then put it back together and I'm getting the boot.img is like half the original size (even tho the kernel files were pretty close to almost deadon the same size) and when I try to send to boot I get this
Code:
>fastboot boot boot_new.img
downloading 'boot.img'...
OKAY [ 1.168s]
booting...
FAILED (remote: incomplete bootimage)
finished. total time: 1.168s
This is for a Pantech Burst (p9070) if that helps at all
Original boot.img
New repacked boot.img

[TOOL] New universal method of rooted for the Optimus 2x | Linux/Mac | (beta)

New universal method of rooted for the Optimus 2x
After much programming and experimenting with another great part, already I can deliver my latest work.
So far have only made for Linux and Mac. (there are methods to not be as done in CMD)
Apparently it seems very simple, but it cheats, it is quite complex, since I've programmed detection of errors and possible solutions to the problems of communication.
The program is smart, detects the operating system that we are using, Linux or iOS, and detects the version of ROM installed to determine installation done.
Also installs the CWM Recovery, and also knows which install depending on the GB ROM or ICS, all automatic, without having to think that option used.
Before you root it will ask you if the CWM Recovery is installed and if it is not you can do it at that time.
The method used is not new, I already used it in earlier versions of GB, the difference with the other methods is that you don't have to copy anything to the SD and things of these, simple and clean installation.
Well I hope that you have Linux or a Mac can root without having to boot the "tedious" Windows.
CHANGELOG
V-0.5
-Implemented detecting permissions
-Implemented detection system 32/64b
-Fixed bug, was not working in Ubuntu 11.04 and higher
V-0.4
-Detection of new possible connection failures.
-Detection of errors in routines Android
-Optimized routines
V-0.3
-Fixed bug of execution, did not work on some current versions of Ubuntu and derivatives
-New start menu.
-New header
V-0.2
-Re-raise with correction error in English language, it only pulled in Spanish
-Includes English language
-Changing the graphic appearance
V-0.1
-Only in Spanish
Click to expand...
Click to collapse
DOWNLOAD
Toolkit_ICS-Root_Stock-v-0.5_LINUX-MAC_by_Homero.rar - 13.40 MB

Thank's Man...and have a nice year 2013!!!...

Nice, it should became sticked.
Also... your program must be also noob-proof

Great work, homero! Thank you.
Sent from my LG-P990 using xda app-developers app

Homero again back with a bang...you're hulk of lg optimus forum..you rock bro

Finally some Linux love.... :thumbup:
Sent from my LG-P990 using xda app-developers app

Thanks homero
Enviado desde mi O2X

I have committed an error, and the English version, and does not run.
You can download the new corrected version or edit and modify.
Edit the root file and look for this line:
Code:
if [ "$ANSWER" = 1 ]; then
tools/es.sh
elif [ "$ANSWER" = 2 ]; then
tools/es.sh
elif [ "$ANSWER" = "x" ] || [ "$ANSWER" = "X" ]; then
Changing this:
Code:
elif [ "$ANSWER" = 2 ]; then
tools/es.sh
why another:
Code:
elif [ "$ANSWER" = 2 ]; then
tools/en.sh
Save and ready, already operates the English version

Homero2 said:
I have committed an error, and the English version, and does not run.
You can download the new corrected version or edit and modify.
Edit the root file and look for this line:
Code:
if [ "$ANSWER" = 1 ]; then
tools/es.sh
elif [ "$ANSWER" = 2 ]; then
tools/es.sh
elif [ "$ANSWER" = "x" ] || [ "$ANSWER" = "X" ]; then
Changing this:
Code:
elif [ "$ANSWER" = 2 ]; then
tools/es.sh
why another:
Code:
elif [ "$ANSWER" = 2 ]; then
tools/en.sh
Save and ready, already operates the English version
Click to expand...
Click to collapse
Beware of #!/bin/sh ... read -n command is bash command, and it does not work well in /bin/sh ...
Confirmed on NetRunner (kubuntu 12.10 ).
You should use #!/bin/bash. That applies to all *.sh scripts.

asmerito said:
Beware of #!/bin/sh ... read -n command is bash command, and it does not work well in /bin/sh ...
Confirmed on NetRunner (kubuntu 12.10 ).
You should use #!/bin/bash. That applies to all *.sh scripts.
Click to expand...
Click to collapse
OK, that is the answer that was waiting, since neither worked in Ubuntu 12.04
EDIT:
Re-raise corrected v-0.3

Though your tool says that flashing the recovery worked successful, I can't boot into recovery. When pushing vol- and power I'm just stuck at the pink LG logo (it's in a boot loop while holding vol- and power). Since the system is not rooted, yet, I also cannot boot into recovery with the CWM app.

thorstenhirsch said:
Though your tool says that flashing the recovery worked successful, I can't boot into recovery. When pushing vol- and power I'm just stuck at the pink LG logo (it's in a boot loop while holding vol- and power). Since the system is not rooted, yet, I also cannot boot into recovery with the CWM app.
Click to expand...
Click to collapse
Don't worry in excess, can be retrieved very easily, but before answer to these questions.
Android version?
did CWM previously installed?.
Terminal type? (P990, P990, etc.)

on slackware 14.0, expanded in /opt:
two patches applied;
files are executables with right permissions;
dir 'tools' is accessible;
lg2x is connected via USB (on, off, MTP, PC Suite, without battery and vol -, ...?)
but same result
Terminal: P990 (EU)
ROM: ICS stock
Previous ver: CM 7.2 stable
Previous CWM: 5.0.2

Homero2 said:
Android version?
did CWM previously installed?.
Terminal type? (P990, P990, etc.)
Click to expand...
Click to collapse
Android 4.0.4 (official V30A)
CWM - well, under GB layout I could install it (I was on CWM5 first and later switched to CWM6), but this was my first try on ICS layout
Terminal type - you mean the shell? I'm using Ubuntu 12.10, so /bin/sh did not work and I replaced all /bin/sh with /bin/bash, that means I was running bash when I "bricked" my phone.

kanzy said:
on slackware 14.0, expanded in /opt:
two patches applied;
files are executables with right permissions;
dir 'tools' is accessible;
lg2x is connected via USB (on, off, MTP, PC Suite, without battery and vol -, ...?)
but same result
Terminal: P990 (EU)
ROM: ICS stock
Previous ver: CM 7.2 stable
Previous CWM: 5.0.2
Click to expand...
Click to collapse
Is about understands that this method of single rooted is for Stock ROM versions.
Clarified the issue, I see errors jump to run ADB.
Try with the latest version that I uploaded, the V-0.4
thorstenhirsch said:
Android 4.0.4 (official V30A)
CWM - well, under GB layout I could install it (I was on CWM5 first and later switched to CWM6), but this was my first try on ICS layout
Terminal type - you mean the shell? I'm using Ubuntu 12.10, so /bin/sh did not work and I replaced all /bin/sh with /bin/bash, that means I was running bash when I "bricked" my phone.
Click to expand...
Click to collapse
Not be what could happen...
Doing tests I found that if you are in Recovery, and the SYSTEM partition is not mounted, the program cannot determine partition type should be used.
That error is already fixed in the V-0.4.
It can reflash the Recovery with one of the many versions that we have in the Forum, one in Flash only recovery or All in one.
I say this because it is the only thing that could have happened, since the rooted is not harmful, only copy two files in the system.

Using V-0.4, and it does not work...
Homero2 said:
Is about understands that this method of single rooted is for Stock ROM versions.
Clarified the issue, I see errors jump to run ADB.
Try with the latest version that I uploaded, the V-0.4
Not be what could happen...
Doing tests I found that if you are in Recovery, and the SYSTEM partition is not mounted, the program cannot determine partition type should be used.
That error is already fixed in the V-0.4.
It can reflash the Recovery with one of the many versions that we have in the Forum, one in Flash only recovery or All in one.
I say this because it is the only thing that could have happened, since the rooted is not harmful, only copy two files in the system.
Click to expand...
Click to collapse
Same problem as on slackware, got booloop after flashing recovery. Cannot enter recovery
After this step:
INSTRUCTIONS
CWM will be installed for version ICS V-4.0.4 of Android
Nvflash started
rcm version 0X20001_____________________________________________________________
System Information:
chip name: unknownh Recovery CWM
chip id: 0x20 major: 1 minor: 4
chip sku: 0xf
chip uid: 0x045070c04120b4d7
macrovision: disabled
hdcp: enabled
sbk burned: false
dk burned: false
boot device: emmc
operating mode: 3
device config strap: 0
device config fuse: 17
sdram config strap: 0
downloading bootloader -- load address: 0x108000 entry point: 0x108000
sending file: tools/fastboot.bin
/ 1027184/1027184 bytes sent
tools/fastboot.bin sent successfully
waiting for bootloader to initialize
bootloader downloaded successfully
sending file: tools/recovery_cwm-6_ics_opti.img
\ 7145472/7145472 bytes sent
tools/recovery_cwm-6_ics_opti.img sent successfully
everything went wrong, and my stock ICS got bootloop.
Using Kubuntu 12.10, stock ICS and V-0.4 of your tool.

Thank you mate! Gonna share it like a bee

asmerito said:
Same problem as on slackware, got booloop after flashing recovery. Cannot enter recovery
...
Click to expand...
Click to collapse
http://forum.xda-developers.com/showthread.php?t=921169

Fixed bug, now works in Ubuntu 11.04, 11.10, 12.04 and 12.10
An error has been detected by Jonki, Thank you!!

Homero2 said:
Fixed bug, now works in Ubuntu 11.04, 11.10, 12.04 and 12.10
An error has been detected by Jonki, Thank you!!
Click to expand...
Click to collapse
What was the problem?

Twrp (DEV-ONLY) compiling op6t (NEW-TEST) twrp-op6t-unified-test-0.6.img

I made this thread so all those trying to compile twrp for op6t can come together.
We can share ideas. What we have and have not got to work. What will work and what won't work.
Basically a thread where all information regarding twrp can be found in one location.
###########
#Please pm me #
##########
If using open source built roms and you have issues with twrp we may need to modify source in the rom or our twrp source.
Please report to rom maker as well.
Things to include.
What source 9.0 or 8.1
Compiling errors.
Post #2
Is where the download links are kept and updated.
Post #3 is where the other variations of twrp for source built roms or ports will be available to download if needed.

New test. twrp-op6t-test-0.3.img TEST ONLY!
@Dameon87 & myself have been working on this recovery for a while. We think it will get better. There are still bugs and issues with twrp. We would like your feedback.
Thanks
I am not responsible for anything that can happen to your device.
Rebuilt and fully working mke2fs
Decryption works
Storage works.
Please test and let us know what works and what doesn't.
twrp-op6t-test-0.2.img
https://www.androidfilehost.com/?fid=11410963190603862182
Updated twrp
Added more functionality ability to flash all zips and Ota's
Backup and restore all partitions including data.
Twrp-op6t-test-0.3.img
https://www.androidfilehost.com/?fid=11410963190603871199
New test twrp-unified-0.4.img
Everything working as usual
Cleaned up source and added @mauronofrio notch theme
Twrp-0.4.img
https://www.androidfilehost.com/?fid=11410963190603875438
Twrp installer thanks to @mauronofrio
https://www.androidfilehost.com/?fid=11410963190603876589
Updated twrp to decrypt 9.0.10
Twrp-0.5.img
https://www.androidfilehost.com/?fid=11410963190603879261
Twrp installer
https://www.androidfilehost.com/?fid=11410963190603879262
New test
Fixed screenshot color when taking screen shots in recovery.
Fixed touch vibration.
Please test and report back.
Twrp-0.6.img
https://www.androidfilehost.com/?fid...63190603880512
Twrp installer
https://www.androidfilehost.com/?fid...63190603880513
Device source
https://github.com/TWINNFAMOUS/android_device_oneplus_fajita
Bootable recovery source
https://github.com/OP6T-TWRP/android_bootable_recovery
Hope you all enjoy

This is where I will keep other twrp variations if needed for all source built roms and ports for our device. When available these will only be meant for testing. Thanks

twinnfamous said:
I made this thread so all those trying to compile twrp for op6t can come together.
We can share ideas. What we have and have not got to work. What will work and what won't work.
Basically a thread where all information regarding twrp can be found in one location.
Things to include.
What source 9.0 or 8.1
Compiling errors.
Click to expand...
Click to collapse
You're from the HTC u11 life! You made that phone awesome while I had it. Good to see you here!

I'm trying to build with 8.1 source and I get errors so build doesn't finish.
Is there anyone that's building successfully with 8.1?

Me, can you explain what are you doing exactly? Which source are you using?

mauronofrio said:
Me, can you explain what are you doing exactly? Which source are you using?
Click to expand...
Click to collapse
I'm trying with full omni source on 8.1. I've used the device source from both op6 & op6t without modifying just to see if it would be and it fails. I'll post output error soon as I can.
I don't have issues building with 9.0 source after I modify device source but it won't boot to twrp just stock recovery like it's not selecting the rite slot

twinnfamous said:
I'm trying to build with 8.1 source and I get errors so build doesn't finish.
Is there anyone that's building successfully with 8.1?
Click to expand...
Click to collapse
Make sure if you're using a newer version of Ubuntu with GCC 7+ you do: export LC_ALL=C
That should get you past the initial build errors.

Dameon87 said:
Make sure if you're using a newer version of Ubuntu with GCC 7+ you do: export LC_ALL=C
That should get you past the initial build errors.
Click to expand...
Click to collapse
Thanks for the info. I upgraded to linuxmint19 Recently. Guess I didn't pay attention to the changelog.

You also need to change two lines in the build/tools/buildinfo.sh :
echo "ro.build.version.release=$PLATFORM_VERSION"
echo "ro.build.version.security_patch=$PLATFORM_SECURITY_PATCH"
to:
echo "ro.build.version.release_orig=$PLATFORM_VERSION"
echo "ro.build.version.security_patch_orig=$PLATFORM_SECURITY_PATCH"

Dameon87 said:
You also need to change two lines in the build/tools/buildinfo.sh :
echo "ro.build.version.release=$PLATFORM_VERSION"
echo "ro.build.version.security_patch=$PLATFORM_SECURITY_PATCH"
to:
echo "ro.build.version.release_orig=$PLATFORM_VERSION"
echo "ro.build.version.security_patch_orig=$PLATFORM_SECURITY_PATCH"
Click to expand...
Click to collapse
And you need to increase the PLATFORM_VERSION to 16.1.0 ( in the new pixel 3 TWRPs is used 16 ) in build/core/version_defaults.mk to override Google's anti-rollback features

Dameon87 said:
You also need to change two lines in the build/tools/buildinfo.sh :
echo "ro.build.version.release=$PLATFORM_VERSION"
echo "ro.build.version.security_patch=$PLATFORM_SECURITY_PATCH"
to:
echo "ro.build.version.release_orig=$PLATFORM_VERSION"
echo "ro.build.version.security_patch_orig=$PLATFORM_SECURITY_PATCH"
Click to expand...
Click to collapse
mauronofrio said:
And you need to increase the PLATFORM_VERSION to 16.1.0 ( in the new pixel 3 TWRPs is used 16 ) in build/core/version_defaults.mk to override Google's anti-rollback features
Click to expand...
Click to collapse
The only issue I found was my lack of awareness. Ive been fighting with the 8.1 source all because the newer os I installed. I'm testing things out.
One thing I know is if the security patch level and date doesn't match what's in the phone then decryption won't work.
So I'm starting there.

I pulled this log from the phone it's only a part of it.
I've been trying everything and this is where it fails at in recovery.
e4crypt_initialize_global_de fail
So I started looking at any logs I can find in the phone and found this. Thought it might be helpful.
All the de folder locations are in the log
Also we use keymaster 4 that shows in The log as well.
I'm still using 8.1 source with 9.0 bootable recovery.
I'm thinking of focusing on trying to get get it to work with 9.0 source.
Log.
Mount system_root for mountting system
[ 1.120382] e4crypt_initialize_global_de
[ 1.120509] Keyring created with id 744456481 in process 530
[ 1.126746] I:List of Keymaster HALs found:
[ 1.126902] I:Keymaster HAL #1: Keymaster HAL: 4 from QTI SecurityLevel: TRUSTED_ENVIRONMENT HAL: [email protected]::IKeymasterDevice/default
[ 1.141729] Using Keymaster HAL: 4 from QTI for encryption. Security level:TRUSTED_ENVIRONMENT , HAL: [email protected]::IKeymasterDevice / defaulte4crypt_init_user0
[ 1.147584] Skipping non-de-key .
[ 1.147590] Skipping non-de-key ..
[ 1.152120] I:List of Keymaster HALs found:
[ 1.152266] I:Keymaster HAL #1: Keymaster HAL: 4 from QTI SecurityLevel: TRUSTED_ENVIRONMENT HAL: [email protected]::IKeymasterDevice/default
[ 1.160865] Using Keymaster HAL: 4 from QTI for encryption. Security level:TRUSTED_ENVIRONMENT , HAL: [email protected]::IKeymasterDevice / defaultInstalled de key for user 0
[ 1.160871] e4crypt_prepare_user_storage
[ 1.166977] ensure policy /data/system_de/0
[ 1.167599] I:Found policy 8372743e0b3d7924 at /data/system_de/0 which matches expected value
[ 1.167608] ensure policy /data/misc_de/0
[ 1.168197] I:Found policy 8372743e0b3d7924 at /data/misc_de/0 which matches expected value
[ 1.168206] ensure policy /data/user_de/0
[ 1.169338] I:Found policy 8372743e0b3d7924 at /data/user_de/0 which matches expected value
[ 1.358232] I:FBE enable...
[ 1.358334] Unable to locate gatekeeper password file '/data/system/gatekeeper.pattern.key'
[ 1.358412] (4294937945 ticks)(pid:530)[bootable/recovery/oem/src/main/oem.cpp]tree_init:90:efault password, decrept...
[ 1.358454] Unable to locate gatekeeper password file '/data/system/gatekeeper.pattern.key'
[ 1.358516] Skipping non-key .
[ 1.358549] Skipping non-key ..
[ 1.364930] I:List of Keymaster HALs found:
[ 1.365014] I:Keymaster HAL #1: Keymaster HAL: 4 from QTI SecurityLevel: TRUSTED_ENVIRONMENT HAL: [email protected]::IKeymasterDevice/default
[ 1.373672] Using Keymaster HAL: 4 from QTI for encryption. Security level:TRUSTED_ENVIRONMENT , HAL: [email protected]::IKeymasterDevice / defaultSuccessfully retrieved key
[ 1.373733] Installed ce key for user 0
[ 1.373765] e4crypt_prepare_user_storage
[ 1.377707] ensure policy /data/system_ce/0
[ 1.378426] I:Found policy d0acf3009500dfb4 at /data/system_ce/0 which matches expected value
[ 1.378462] ensure policy /data/misc_ce/0
[ 1.379037] I:Found policy d0acf3009500dfb4 at /data/misc_ce/0 which matches expected value
[ 1.379069] ensure policy /data/data
[ 1.380547] I:Found policy d0acf3009500dfb4 at /data/data which matches expected value
[ 1.380597] Decrypted Successfully!

twinnfamous said:
I pulled this log from the phone it's only a part of it.
I've been trying everything and this is where it fails at in recovery.
e4crypt_initialize_global_de fail
So I started looking at any logs I can find in the phone and found this. Thought it might be helpful.
All the de folder locations are in the log
Also we use keymaster 4 that shows in The log as well.
I'm still using 8.1 source with 9.0 bootable recovery.
I'm thinking of focusing on trying to get get it to work with 9.0 source.
Log.
Mount system_root for mountting system
[ 1.120382] e4crypt_initialize_global_de
[ 1.120509] Keyring created with id 744456481 in process 530
[ 1.126746] I:List of Keymaster HALs found:
[ 1.126902] I:Keymaster HAL #1: Keymaster HAL: 4 from QTI SecurityLevel: TRUSTED_ENVIRONMENT HAL: [email protected]::IKeymasterDevice/default
[ 1.141729] Using Keymaster HAL: 4 from QTI for encryption. Security level:TRUSTED_ENVIRONMENT , HAL: [email protected]::IKeymasterDevice / defaulte4crypt_init_user0
[ 1.147584] Skipping non-de-key .
[ 1.147590] Skipping non-de-key ..
[ 1.152120] I:List of Keymaster HALs found:
[ 1.152266] I:Keymaster HAL #1: Keymaster HAL: 4 from QTI SecurityLevel: TRUSTED_ENVIRONMENT HAL: [email protected]::IKeymasterDevice/default
[ 1.160865] Using Keymaster HAL: 4 from QTI for encryption. Security level:TRUSTED_ENVIRONMENT , HAL: [email protected]::IKeymasterDevice / defaultInstalled de key for user 0
[ 1.160871] e4crypt_prepare_user_storage
[ 1.166977] ensure policy /data/system_de/0
[ 1.167599] I:Found policy 8372743e0b3d7924 at /data/system_de/0 which matches expected value
[ 1.167608] ensure policy /data/misc_de/0
[ 1.168197] I:Found policy 8372743e0b3d7924 at /data/misc_de/0 which matches expected value
[ 1.168206] ensure policy /data/user_de/0
[ 1.169338] I:Found policy 8372743e0b3d7924 at /data/user_de/0 which matches expected value
[ 1.358232] I:FBE enable...
[ 1.358334] Unable to locate gatekeeper password file '/data/system/gatekeeper.pattern.key'
[ 1.358412] (4294937945 ticks)(pid:530)[bootable/recovery/oem/src/main/oem.cpp]tree_init:90:efault password, decrept...
[ 1.358454] Unable to locate gatekeeper password file '/data/system/gatekeeper.pattern.key'
[ 1.358516] Skipping non-key .
[ 1.358549] Skipping non-key ..
[ 1.364930] I:List of Keymaster HALs found:
[ 1.365014] I:Keymaster HAL #1: Keymaster HAL: 4 from QTI SecurityLevel: TRUSTED_ENVIRONMENT HAL: [email protected]::IKeymasterDevice/default
[ 1.373672] Using Keymaster HAL: 4 from QTI for encryption. Security level:TRUSTED_ENVIRONMENT , HAL: [email protected]::IKeymasterDevice / defaultSuccessfully retrieved key
[ 1.373733] Installed ce key for user 0
[ 1.373765] e4crypt_prepare_user_storage
[ 1.377707] ensure policy /data/system_ce/0
[ 1.378426] I:Found policy d0acf3009500dfb4 at /data/system_ce/0 which matches expected value
[ 1.378462] ensure policy /data/misc_ce/0
[ 1.379037] I:Found policy d0acf3009500dfb4 at /data/misc_ce/0 which matches expected value
[ 1.379069] ensure policy /data/data
[ 1.380547] I:Found policy d0acf3009500dfb4 at /data/data which matches expected value
[ 1.380597] Decrypted Successfully!
Click to expand...
Click to collapse
Was this taken from the default/stock recovery or just a standard boot? I've been using the twrp-8.1 branch (Which uses android_bootable_recovery 9.0 branch by default now). I've tried a multitude of things, but I simply just get the e4crypt_initialize_global_de fail error as well. I think the way things are done for the Pixel 3/XL are the way to proceed, but aside the device tree we don't have much else to go on.

Dameon87 said:
Was this taken from the default/stock recovery or just a standard boot? I've been using the twrp-8.1 branch (Which uses android_bootable_recovery 9.0 branch by default now). I've tried a multitude of things, but I simply just get the e4crypt_initialize_global_de fail error as well. I think the way things are done for the Pixel 3/XL are the way to proceed, but aside the device tree we don't have much else to go on.
Click to expand...
Click to collapse
Maybe Fstab files need modifying
The recovery.fstab I pulled out of the boot.Img has a data partition labeled f2fs
And a Data partition labeled ext4

twinnfamous said:
Maybe Fstab files need modifying
The recovery.fstab I pulled out of the boot.Img has a data partition labeled f2fs
And a Data partition labeled ext4
Click to expand...
Click to collapse
Already tried that but I didn't really get anything different. There's a few issues that are happening I think.
Dmesg repeats this over and over and over:
<3>[19700123_15:34:23.729788]@4 FG: fg_psy_get_property: unsupported property 27
<14>[19700123_15:34:26.049860]@4 init: starting service 'boot-1-0'...
<11>[19700123_15:34:26.050060]@4 init: property_set("ro.boottime.boot-1-0", "798060609590") failed: property already set
<11>[19700123_15:34:26.050448]@0 init: cannot execve('/sbin/[email protected]'): No such file or directory
<14>[19700123_15:34:26.051342]@4 init: Sending signal 9 to service 'boot-1-0' (pid 762) process group...
<14>[19700123_15:34:26.051372]@4 init: Successfully killed process cgroup uid 0 pid 762 in 0ms
Then just getting this according to the recovery.log:
I:File Based Encryption is present
e4crypt_initialize_global_de returned fail
e4crypt_initialize_global_de returned fail
e4crypt_initialize_global_de returned fail

Dameon87 said:
Already tried that but I didn't really get anything different. There's a few issues that are happening I think.
Dmesg repeats this over and over and over:
<3>[19700123_15:34:23.729788]@4 FG: fg_psy_get_property: unsupported property 27
<14>[19700123_15:34:26.049860]@4 init: starting service 'boot-1-0'...
<11>[19700123_15:34:26.050060]@4 init: property_set("ro.boottime.boot-1-0", "798060609590") failed: property already set
<11>[19700123_15:34:26.050448]@0 init: cannot execve('/sbin/[email protected]'): No such file or directory
<14>[19700123_15:34:26.051342]@4 init: Sending signal 9 to service 'boot-1-0' (pid 762) process group...
<14>[19700123_15:34:26.051372]@4 init: Successfully killed process cgroup uid 0 pid 762 in 0ms
Then just getting this according to the recovery.log:
I:File Based Encryption is present
e4crypt_initialize_global_de returned fail
e4crypt_initialize_global_de returned fail
e4crypt_initialize_global_de returned fail
Click to expand...
Click to collapse
This is probably why it won't even boot twrp built with 9.0

twinnfamous said:
This is probably why it won't even boot twrp built with 9.0
Click to expand...
Click to collapse
Are you doing a minimal build env, or pulling in the entire Omni tree? There is only a twrp-8.1 branch for the minimal manifest, but that does pull in android_bootable_recovery 9.0 branch. I can get both to compile and boot fine.

Dameon87 said:
Are you doing a minimal build env, or pulling in the entire Omni tree? There is only a twrp-8.1 branch for the minimal manifest, but that does pull in android_bootable_recovery 9.0 branch. I can get both to compile and boot fine.
Click to expand...
Click to collapse
Full sources on both 8.1 and 9.0

twinnfamous said:
Full sources on both 8.1 and 9.0
Click to expand...
Click to collapse
Alright I'm doing a full sync of the omni repo and will give that a try next. I had been just using the minimal manifest.

Problem flashing kernel

Hi all,
I'm doing some studies with the Pixel 4a 5g. I'd like to make some modifications to the kernel. For that reason, my first attempt was to build the original kernel (https://source.android.com/setup/build/building-kernels). I've compiled the source code with the following lines.
# repo init -u https://android.googlesource.com/kernel/manifest -b android-msm-bramble-4.19-android11-d1
# ./build/build.sh
The problem comes when flashing the image. Here's what I get:
./fastboot boot google-kernel/out/android-msm-pixel-4.19/dist/Image.lz4-dtb
creating boot image...
creating boot image - 9392128 bytes
Sending 'boot.img' (9172 KB) OKAY [ 0.372s]
Booting (bootloader) boot.img missing cmdline or OS version
At this points, the shell and the phone hang forever.
Any help will be much appreciated.
Thanks,
Adrian

Hi Adrian,
I've got the same error. Have you managed to address it?
Jon