Related
The original thread:http://forum.xda-developers.com/showthread.php?t=396782 needs an abridged version.
==================================================
Go to http://www.androidonhtc.com/ if you're just starting as it has the latest info. This thread is to highlight the info from the original thread only.
Latest Builds
Port Status
==================================================
Compiling Android Kernel for Kaiser
Modify initrd.gz files and CPIO handling
system.img mounting, editing and rebuilding with ext2/3
system.img mounting, editing and rebuilding with cramfs
Howto: Pull from git (new/update/resync)
[WIP] Configuring WiFi Interface
==================================================
system.img mounting, editing and rebuilding with ext2/3
seidler2547: Post:
Actually I've played with Android a bit for now, and I changed to ext3. It doesn't only work - it's much faster, too! Startup time during the blinking android is about half of what it was before.
How-To:
Code:
Code:
cd /tmp
# prepare dirs
mkdir a-sys
mkdir a-ext
# prepare image
dd if=/dev/zero of=/where/is/sdcard/system.img.new bs=1M count=64
mkfs.ext3 /where/is/sdcard/system.img.new
# mount old image and copy to new
mount -o loop /where/is/the/system.img a-sys
mount -o loop /where/is/sdcard/system.img.new a-ext
cp -a a-sys/* a-ext/
Now you can unmount the old image and happily edit in the new image. Don't forget to rename the system.img.new to system.img (after you have unmounted it).
In your initrd, in file init, where it says
Code:
losetup /dev/block/loop1 /sdcard/system.img
...
mount -t cramfs -o ro,noatime,nodiratime /dev/block/loop1 /system
change "-t cramfs" to "-t ext2" or "-t ext3". You can also change the path (/sdcard/system.img) there.
Click to expand...
Click to collapse
Modify initrd.gz files and CPIO handling
dcordes: Post:
There is no magick in the initrd files. They are .cpio.gz files, gzipped cpio balls. To extract a .cpio.gz file named initrd-android.cpio.gz simply do
Code:
gunzip initrd-android.cpio.gz && cpio -i < initrd.android.cpio
Then you have the extracted rootfs. The reverse way would be, assuming you are inside your rootfs folder:
Code:
find ./ | cpio -H newc -o | gzip > ../my-initr-android-with-custom-stuffs.cpio.gz
And yes, you can remove and add applications you find that way.
Click to expand...
Click to collapse
system.img mounting, editing and rebuilding with cramfs
dzo: Post:
Hi, you can't just use mkcramfs on the system folder because the permissions will be wrong. This is the script I use:
Code:
Code:
out/host/linux-x86/bin/genext2fs -d out/target/product/generic/system -b 80000 -a system.ext2
mount -o loop system.ext2 /mnt/system
cp /mnt/system/usr/keychars/qwerty2.kcm.bin /mnt/system/usr/keychars/vogue-ts.kcm.bin
cp com.google.android.maps.jar /mnt/system/framework
cp Maps.apk Street.apk /mnt/system/app
mkfs.cramfs /mnt/system system.img
umount /mnt/system
#pcp system.img :/Storage\ Card/system.img
This also puts the maps app in (just copy from one of my images) and the vogue keymap. Without the source for the ril you will also need to copy my RIL (libreference-ril.so).
Click to expand...
Click to collapse
[WIP] Configuring WiFi Interface
This has been able to initialize the interface, assign arbitrary IP addresses but can not go further at the moment.
Code:
# ifconfig tiwlan0 192.168.1.100
# ifconfig tiwlan0 up
error: SIOCSIFFLAGS (Cannot assign requested address)
# ifconfig tiwlan0
tiwlan0: ip 192.168.1.100 mask 255.255.255.0 flags (down broadcast multicast)
dmesg will show:
Code:
wlan: no version for "struct_module" found: kernel tainted.
TIWLAN: Driver loading
trout_wifi_power: 1
trout_wifi_reset: 0
trout_wifi_set_carddetect: 1
TIWLAN: Found SDIO control (vendor 0x104c, device 0x9066)
TIWLAN: Driver initialized (rc 0)
TIWLAN: Driver loaded
Android's built-in wireless settings seem to disable the interface beyond just interfering with it, therefore it's best to stay with terminal and using 'ash' will give you a shell with command history (up/down scroll).
markya23: Post:
Need to create a folder in you system image package called /etc/wifi and copy tiwlan.ini, wpa_supplicant.conf and fw1251r1c.bin.
Need to copy the wlan.ko to /lib/modules in the system image (create the dir if required).
Create the new system image and boot Android. Start the dev console and type:
Code:
cp /system/etc/wifi/wpa_supplicant.conf /data/misc/wifi/wpa_supplicant.conf
insmod /system/lib/modules/wlan.ko
wlan_loader -f /system/etc/wifi/Fw1251r1c.bin -e /proc/calibration -i /system/etc/wifi/tiwlan.ini
cd /data/local/tmp
wpa_supplicant -f -Dtiwlan0 -itiwlan0 -c/data/misc/wifi/wpa_supplicant.conf &
ifconfig tiwlan0 192.168.1.100 netmask 255.255.255.0
ifconfig tiwlan0 up
Click to expand...
Click to collapse
Compiling Android Kernel for Kaiser
dwaradzyn: Post:
Here are brief instructions on how to compile android kernel for Kaiser from git.linuxtogo.org repository. I assume that running OS is Linux and it has everything required to build x86 or ia64 kernel. Beside that latest git software should be installed. The shell is assumed to be bash.
1. Let's start with creating a directory for kernel in home directory:
Code:
mkdir ~/android-kernel
cd android-kernel
2. Next thing is to get the sources from repository. To make it happen (this could take a while, it downloads 280MB):
Code:
git clone git://git.linuxtogo.org/home/groups/mobile-linux/kernel.git
OUTPUT:
Code:
Initialized empty Git repository in /home/user/android-kernel/kernel/.git/
remote: Counting objects: 908251, done.
remote: Compressing objects: 100% (153970/153970), done.
remote: Total 908251 (delta 755115), reused 906063 (delta 753016)
Receiving objects: 100% (908251/908251), 281.86 MiB | 292 KiB/s, done.
Resolving deltas: 100% (755115/755115), done.
Checking out files: 100% (22584/22584), done.
3. The htc-msm branch is of our interest (again it could take a few seconds):
Click to expand...
Click to collapse
*** Update, poly_poly-man states we are working off of htc-vogue not htc-msm. I'm leaving the original code here but I would urge you to modify the next line as poly has suggested:
Code:
cd kernel
git checkout -b htc-msm origin/htc-msm
OUTPUT:
Code:
Branch htc-msm set up to track remote branch refs/remotes/origin/htc-msm.
Switched to a new branch "htc-msm"
4. Let's take care of arm toolchain. Download this file (64MB) into ~/android-kernel:
Code:
[url]http://www.codesourcery.com/gnu_toolchains/arm/portal/package2549/public/arm-none-linux-gnueabi/arm-2008q1-126-arm-none-linux-gnueabi-i686-pc-linux-gnu.tar.bz2[/url]
Unpack it:
Code:
cd ~/android-kernel
tar xjf arm-2008q1-126-arm-none-linux-gnueabi-i686-pc-linux-gnu.tar.bz2
5. Compile the kernel
Prepare default .config for Kaiser:
Code:
cd ~/android-kernel/kernel
make htckaiser_defconfig ARCH=arm
OUTPUT:
Code:
........
lots of output
........
# configuration written to .config
#
And finally compile the kernel to get zImage (takes a minute or two):
Code:
export PATH=~/android-kernel/arm-2008q1/bin:$PATH
make zImage ARCH=arm CROSS_COMPILE=arm-none-linux-gnueabi-
OUTPUT:
Code:
........
lots of output
........
Kernel: arch/arm/boot/zImage is ready
Now copy ~/android-kernel/kernel/arch/arm/boot/zImage to your phone and play with it.
Some ending tips:
A. You can compile earlier versions of sources in repository. To do that click on one of "commit" links on page:
Code:
http://git.linuxtogo.org/?p=groups/mobile-linux/kernel.git;a=summary
and read commit id (for example: f9d1bcea9342348623f5a57588044f76d8b649cd):
Code:
git reset --hard f9d1bcea9342348623f5a57588044f76d8b649cd
It will override any changes you made to files in ~/android-kernel/kernel.
B. Once you have downloaded git repository, you can swallow latest changes by issuing:
Code:
cd ~/android-kernel/kernel
git pull
C. If your machine has more than one cpus/cores you can speed up kernel compilation by adding -j <cores/cpus_number>, for example (dual core):
Code:
make -j 2 zImage ARCH=arm CROSS_COMPILE=arm-none-linux-gnueabi-
Click to expand...
Click to collapse
bad internet makes for double post. Please delete
wrong directions for kernel - we are working off of htc-vogue branch... not htc-msm...
can we make the internet work through the usb cable?
how does android know which device node is gps? it's not picking it up for kaiser...
if you enable gps in winmo (or enable it in smd0 - I believe the command is @startgps), smd7 is a nmea stream (acts as a serial GPS).... could a symlink possibly be the right solution to this?
Howto: Pull from git (new/update/resync)
This will download the latest from git:
dwaradzyn: Post:
Here are brief instructions on how to compile android kernel for Kaiser from git.linuxtogo.org repository. I assume that running OS is Linux and it has everything required to build x86 or ia64 kernel. Beside that latest git software should be installed. The shell is assumed to be bash.
1. Let's start with creating a directory for kernel in home directory:
Code:
mkdir ~/android-kernel
cd android-kernel
2. Next thing is to get the sources from repository. To make it happen (this could take a while, it downloads 280MB):
Code:
git clone git://git.linuxtogo.org/home/grou
ps/mobile-linux/kernel.git
OUTPUT:
Code:
Initialized empty Git repository in /home/user/android-kernel/kernel/.git/
remote: Counting objects: 908251, done.
remote: Compressing objects: 100% (153970/153970), done.
remote: Total 908251 (delta 755115), reused 906063 (delta 753016)
Receiving objects: 100% (908251/908251), 281.86 MiB | 292 KiB/s, done.
Resolving deltas: 100% (755115/755115), done.
Checking out files: 100% (22584/22584), done.
3. The htc-msm branch is of our interest (again it could take a few seconds):
Code:
cd kernel
git checkout -b htc-msm origin/htc-msm
OUTPUT:
Code:
Branch htc-msm set up to track remote branch refs/remotes/origin/htc-msm.
Switched to a new branch "htc-msm"
4. Let's take care of arm toolchain. Download this file (64MB) into ~/android-kernel:
Code:
http://www.codesourcery.com/gnu_toolchains/arm/portal/package2549/public/arm-none-linux-gnueabi/arm-2008q1-126-arm-none-linux-gnueabi-i686-pc-linux-gnu.tar.bz2
Unpack it:
Code:
cd ~/android-kernel
tar xjf arm-2008q1-126-arm-none-linux-gnueabi-i686-pc-linux-gnu.tar.bz2
5. Compile the kernel
Prepare default .config for Kaiser:
Code:
cd ~/android-kernel/kernel
make htckaiser_defconfig ARCH=arm
OUTPUT:
Code:
........
lots of output
........
# configuration written to .config
#
And finally compile the kernel to get zImage (takes a minute or two):
Code:
export PATH=~/android-kernel/arm-2008q1/bin:$PATH
make zImage ARCH=arm CROSS_COMPILE=arm-none-linux-gnueabi-
OUTPUT:
Code:
........
lots of output
........
Kernel: arch/arm/boot/zImage is ready
Now copy ~/android-kernel/kernel/arch/arm/boot/zImage to your phone and play with it.
Some ending tips:
A. You can compile earlier versions of sources in repository. To do that click on one of "commit" links on page:
http://git.linuxtogo.org/?p=groups/mobile-linux/kernel.git;a=summary
and read commit id (for example: f9d1bcea9342348623f5a57588044f76d8b649cd):
Code:
git reset --hard f9d1bcea9342348623f5a57588044f76d8b649cd
It will override any changes you made to files in ~/android-kernel/kernel.
B. Once you have downloaded git repository, you can swallow latest changes by issuing:
Code:
cd ~/android-kernel/kernel
git pull
C. If your machine has more than one cpus/cores you can speed up kernel compilation by adding -j <cores/cpus_number>, for example (dual core):
Code:
make -j 2 zImage ARCH=arm CROSS_COMPILE=arm-none-linux-gnueabi-
Click to expand...
Click to collapse
poly_poly-man: Post:
This will pull from git which will update/resync a git pull:
Code:
mkdir foo
cd foo
git init
git pull git://git.linuxtogo.org/home/groups/mobile-linux/kernel.git htc-vogue
Click to expand...
Click to collapse
I'll follow-up with some nice, full instructions...
1. prerequisites: arm-none-linux-gnueabi toolchain (gentoo users can use crossdev - otherwise.... uhh... idk?), git, a host toolchain (gentoo users have this by default, most other distros have this under "development" in their package managment... if you have gcc, you're probably set).
2. mkdir kernel
cd kernel
git init
git pull git://git.linuxtogo.org/home/groups/mobile-linux/kernel.git htc-vogue
3. make vogue_defconfig
4. make
5. cp arch/arm/boot/bzImage /path/to/sdcard/
6. to update, run the git pull command by itself again, run make (may have to do the config line again if it's changed) and cp.
Thanks for the post poly. I don't know how that's different from what the post I referenced as I'm not +4 at this stuff.. I did what you posted with android kernel from git and got a 1.2mb zImage that crashed HaRET.. I'm guessing this is my bad. What could I have overlooked? Thanks
enatefox said:
Thanks for the post poly. I don't know how that's different from what the post I referenced as I'm not +4 at this stuff.. I did what you posted with android kernel from git and got a 1.2mb zImage that crashed HaRET.. I'm guessing this is my bad. What could I have overlooked? Thanks
Click to expand...
Click to collapse
where'd you get your toolchain?
does building a regular (host arch) kernel work?
Also - what's the proper way to build a system.img by hand? I'm looking to modify that quite a bit, but can't find a persistent source tree besides the main one, which is seriously crippled.
I thought you were one of the experts, lol. I've been left with no support on how dzo, et all are customizing kernels so I've been in read only mode on the 'other thread' looking elsewhere for support.
Just wanted to ask you first, what's with the Dream radio? I know you posted the mods censored it but what's with the sig now? It piqued my interest... as I'm using (shudder) winmo on the regular while Android is being worked on I was hoping it would be worth looking into if it doesn't brick my phone.
Answers to your questions:
As I said, my own zImage is no go. Check this link (not for our phone but the links at the bottom are pretty useful): http://wiki.xda-developers.com/index.php?pagename=BlackstoneLinux#Runningx20.Linuxx20.onx20.blackstone
I got the toolchain from the steps I (re)posted on this thread:
http://forum.xda-developers.com/showpost.php?p=2269384&postcount=184 so that gave me a 1.2mb zImage where everyone's been posting 1.4mb-- I know there's something not right. As far as building a system.img by hand? I've taken existing ones either from posted bundles or from Android src directly. Maybe I suck (real possibility) but cupcake and 1.0 have been pretty flaky for me (there are system.img's included in the source). You should know how to mount and edit them though (look at the first post on this thread). My experience is the git source is useless unless you've got a G1-- I don't know how to make it run on Tilts. If it does work, then the answer to your question about host arch compiling is no-- it has to be ARMv5 for our phones. That's where this line comes in:
make zImage ARCH=arm CROSS_COMPILE=arm-none-linux-gnueabi-
Click to expand...
Click to collapse
Seriously, I need help learning how to modify the kernel-- I've been a "google will have the answer for me" Linux bystander for a while and while I'm good at what I've done I'm not so good at this uncharted territory. I'm really looking for help to figure out how to compile modules (saurik and dzo never got back to me about that) and all I've gotten is "wait while I do it myself" which is cool they're working on it but we're obviously here to work on it too.
I've even been trying to get Debian installed (familiar territory for me) so I can at least get stuff working. You do know that Android is nothing but a Java VM layer for Linux and it will (could be) put on any self-respecting distro. Personally, I want Debian on my phone with an Android chroot as that would trump all.
Sorry to rant but you seem more about figuring this out like me and I don't know who else is really working on this besides the dev-gods who have no tutorials.
enatefox said:
I thought you were one of the experts, lol. I've been left with no support on how dzo, et all are customizing kernels so I've been in read only mode on the 'other thread' looking elsewhere for support.
Just wanted to ask you first, what's with the Dream radio? I know you posted the mods censored it but what's with the sig now? It piqued my interest... as I'm using (shudder) winmo on the regular while Android is being worked on I was hoping it would be worth looking into if it doesn't brick my phone.
Click to expand...
Click to collapse
It never actually worked... maybe. My phone was reporting the wrong version on a *different* radio (1.65.21.18, was saying 19) before, and trying to flash this changed the version to be correct. 0x300 radios will never flash, and this as a 0x301 *will* brick your phone. Then again... like 2 people reported epic success... In other words, no, it never really existed.
Answers to your questions:
As I said, my own zImage is no go. Check this link (not for our phone but the links at the bottom are pretty useful): http://wiki.xda-developers.com/index.php?pagename=BlackstoneLinux#Runningx20.Linuxx20.onx20.blackstone
I got the toolchain from the steps I (re)posted on this thread:
http://forum.xda-developers.com/showpost.php?p=2269384&postcount=184 so that gave me a 1.2mb zImage where everyone's been posting 1.4mb-- I know there's something not right. As far as building a system.img by hand? I've taken existing ones either from posted bundles or from Android src directly. Maybe I suck (real possibility) but cupcake and 1.0 have been pretty flaky for me (there are system.img's included in the source). You should know how to mount and edit them though (look at the first post on this thread). My experience is the git source is useless unless you've got a G1-- I don't know how to make it run on Tilts. If it does work, then the answer to your question about host arch compiling is no-- it has to be ARMv5 for our phones. That's where this line comes in:
make zImage ARCH=arm CROSS_COMPILE=arm-none-linux-gnueabi-
Click to expand...
Click to collapse
unnecessary - zImage is implied, and the other two are in the Makefile already.
Seriously, I need help learning how to modify the kernel-- I've been a "google will have the answer for me" Linux bystander for a while and while I'm good at what I've done I'm not so good at this uncharted territory. I'm really looking for help to figure out how to compile modules (saurik and dzo never got back to me about that) and all I've gotten is "wait while I do it myself" which is cool they're working on it but we're obviously here to work on it too.
Click to expand...
Click to collapse
we have 0 modules at the moment (but perhaps support - I forget). Just compile stuff in - modules are very bad.
I've even been trying to get Debian installed (familiar territory for me) so I can at least get stuff working. You do know that Android is nothing but a Java VM layer for Linux and it will (could be) put on any self-respecting distro. Personally, I want Debian on my phone with an Android chroot as that would trump all.
Click to expand...
Click to collapse
It's a nice idea, but remember where a lot of the current coding is taking place - the ril, which is part of android (the system.img, at least), and doesn't go across to other distros quite as well. I think running Dalvik alongside whatever you're running in Debian might be too much for this phone's epically slow processor (compared to msm7201a).
Sorry to rant but you seem more about figuring this out like me and I don't know who else is really working on this besides the dev-gods who have no tutorials.
Click to expand...
Click to collapse
my instructions should work - they are what I use, at least.
Someone should change the instructions to (in the Wiki they are correct):
make ARCH=arm vogue_defconfig
I compiled my kernel, booted in Ubuntu, but had no touchscreen at all, I am not sure if the vogue_defconfig file takes care of everything necessary, so now I am checking everything via menuconfig to see whether eveything is all right or not.
DOMy
Do not Use Ext3 on SD cards
enatefox said:
seidler2547: Post:
Click to expand...
Click to collapse
You should NOT be using ext3 on any sort of flash ram device. You will wear out the medium. Journaling is just a bad idea in this sort of situation.
http://www.handhelds.org/hypermail/familiar/273/27320.html
3) ext3 is "very bad" because of the way it does journaling. It does journal, which seems like a good idea, but it also automatically periodically writes a lot of things out to the same secors on disk. I don't have first hand experience with this, but I remember somebody familiar with ext3 writing about this. It's method of journaling is not particularly intended for any kind of wear leveling at all.
Click to expand...
Click to collapse
http://www.mail-archive.com/[email protected]/msg38988.html
There are three disadvantages with the journaled file system:
- lower performance at write time, since there is the extra work of the
journal
- increased chance of damaging the SD card due to extra use of the
journal causing wearing
- increased space usage (for the journal)
Click to expand...
Click to collapse
And this is the general consensus within most of linux on ext3 and wearing mediums. I'm not sure if Andriod's kernel can do ext4, but ext4 can run without a journal.
Yet another reason not to use ext3 is that is near impossible to undelete something, something you can do with ext2.
-edit-
It looks like Android can use Ext4
http://thatsbadass.com/android/tag/ext4/
haha! is a good job, i like it
can't run android on my kaiser
hi folks,
i have read many threads here and on androidonhtc.com, i have read also the install instructions, but it seems im too stupid to run it.
it fails on "can't find system.img". i wil not flash it, i will run it from sd-card.
so, please can anyone attached an actually zip file which i just unpack und run haret.exe to work android on my kaiser ?
thanks and best regards
lenzen
Here is a simple way to compare two firmwares.
unpack the files and mount the file-systems as root on a Linux system:
Code:
mount -o loop /home/gregor/Desktop/CODE_I9000XWJM2/factoryfs.rfs /home/gregor/Desktop/dir1
mount -o loop /home/gregor/Desktop/I9000XWJM5/I9000XWJM5/CODE_I9000XWJM5.tar.md5_FILES/factoryfs.rfs /home/gregor/Desktop/dir2
Compare the two filesystems and generate a list:
Code:
rsync -cavn --delete /home/gregor/Desktop/dir1/ /home/gregor/Desktop/dir2/ > diff_JM2_JM5.txt
Examples attached
Wow, way more changes between JM2 and JM5 than you'd think!
this should be stickied or added to the big FAQ
i agree nice tip
Where did you get firmware version JH1? This is a new firmware right?
You can also use this command (change OLD and NEW with the dir1 and dir2 from rsync method):
Code:
LANG=C diff -ur OLD NEW | sed -e 's/^Binary files .* and \(.*\) differ$/Changed: \1/' | fgrep -v '.odex'
This shows in an IMHO somewhat nicer way which files were added or removed, and also shows exactly what changed in text files. It also removes odex files from the list, because they will change on every rebuild no matter how, even if the apps didn't change.
See this post for example output.
AllGamer said:
this should be stickied or added to the big FAQ
Click to expand...
Click to collapse
+1
This is awesome.
edit: oops, wrong thread
The idea of this exercise is (at least) to get a stable starting point for kernel development. The thing which is currently missing is a proper working .config. I have reconstructed it using differential analysis and in the process hoped to find which components have actually been activated and to uncover changes (or Easter eggs) in the sources which have not been advertised. Having a working and identical I9000XWJP6 kernel also means that open development can continue from the current official public release. From there the things possible are only limited by your imagination.
The following is a walk-through on how to build the kernel, description of pitfalls that will cause changes in .config to break, and some annotations on discoveries made in the process.
The things you need are:
Mandatory:
- I9000XWJP6 zImage : from your favorite location
- Source tree : opensource•samsung•com the GT-I9000 OpenSource Froyo Update JPM.zip
- Sourcery G++: www•codesourcery•com/sgpp/lite/arm/portal/release1039
- Tweak-Kit : <attached>
Optional:
- Arm enabled GCC and binutils, including a development libbfd.
- Lots of your favorite beverage
The md5sum of the zImage should be: 26e9d5d206baf1515144c6b8de6f10d2
It is critical that the Sourcery G++ version is 2009q3-67.
The Tweak-Kit contains the following components:
Readme.txt - You're reading it
mkvmlinux.cc - convert zImage to vmlinux and extract the init ramdisk image
I9000XWJP6_defconfig - default .config
stamp.patch - set date/time and such to original
style.patch - fix style related warnings
prototype.patch - fix prototype related warnings
error.patch - Recoverable errors
houston.patch - Unrecoverable errors
shadow.patch - Fixate data structures
I9000XWJP6.h - Fixated macros
I9000XWJP6.c - Entry point stubs
SOME THINGS I ENCOUNTERED IN THE PROCESS:
a) What I absolutely did not expect was that I found two different encodings of the build timestamp. I could deduce that the timezone was central Europe. I had the assumption it would be Asia or America.
b) What was to be expected is that the source tree is incomplete. The directories drivers/fsr and fs/rfs are missing. You can still compile the kernel as the missing files are used to build modules. Problems start when you change the config. Doing so will change entry points and data structures and your kernel might die a horrible death when it loads modules who are unaware on these changes. There is a workaround which I will explain later.
c) The weirdest thing I encountered were the functions enable_hlt() and disable_hlt(). The are located deep in the unwind tables, a section not intended for code. I spent many hours trying to figure out how they got there or why but I still have no clue. I found exactly only one way to reproduce this behaviour and it is certainly not due to a typo, accident or ignorance.
d) The kernel is not a production but a debug version. It has nearly all tracking/tracing/debug bells and whistles switched on. If the energy required to maintain the statistics where to emit light, you could use your Galaxy as a Christmas tree. Function profiling is enabled and has a considerable negative effect on performance, code is not optimized for size but speed, and unwind tables have been enabled which are not used. These have a really bad impact on footprint size. I really hope that the same compiler and settings are not used for the Android layer. Changing the config into a production version will not work (and crash) as the non-native modules expect the debugging hooks which will no longer exist. But the same workaround as above can be used.
e) The functionality of the power management domains have been optimized to oblivion due to the excessive placing of code disabling comments in large parts of the clock, power management and mach-aries.c. Maybe because the Galaxy hardware is too different than the evaluation boards, or the hardware is buggy and disabling the code makes it less unstable, or there was just not enough time to get the code working. Anyway, at this moment I have no oversight into what degree the absence of power domains influence battery usage.
f) When I started examining the binary code I was puzzled by snippets of code I could not reproduce. Even worse, I encountered snippets that were just questionable. Unusual instruction sequences, and resister usage. Thinking I bumped into a GCC bug, I started debugging the compiler and even tweaked instruction scheduling weights but with no satisfying outcome. I know that GCC is very stubborn with regard to saving and clobbering registers in/across function calls and the code I saw was just incorrect. I knew a different compiler was used and I suddenly realized that it may be more different than what first meets the eye. The culprit turned out to be Sourcery G++. It is a private maintained branch of GCC for reasons I have not investigated. Even the Sourcery assembler is tainted as it played a nasty trick on me with the enable_hlt/disable_hlt thing. I do not like the code I see and I am aiming into getting the sources stock GCC friendly with a working kernel. However, GCC and Sourcery generate code which seem difficult to mix, but I'm getting closer.
g) Compiler warnings. Many of the Samsung sources generate warnings, something I really dislike. In my opinion a warning is emitted for a piece of code which can be interpreted in several ways, leaving the compiler to choose which. Usually it will choose the wrong one. Most warnings were related to coding style shortcuts, a couple of incorrect function prototype resulting in functions that should return int to return random or falsely ignoring return values. There were also a couple of nasties like deference of uninitialized pointers, accessing out-of-bound data and mixing clock data-structures of different types. Included are a number of patches to fix them.
h) I looked deeper into why GCC and Sourcery won't mix and discovered that they have different implementations with regard to constant definition within enum declarations. Google points to the staring point "GCC bug 30260" where is written that the behaviour of enumeration constants has changed to becoming signed int. I have noticed that even explicit unsigned values will change to signed.
Here is an example of what is going wrong:
Take following declaration
Code:
enum rt_class_t { RT_TABLE_MAX=0xFFFFFFFF }
. GCC will consider RT_TABLE_MAX to be -1, and Sourcery will consider it 4294967295. Now, in net/ipv4/fib_rules.c there is this code snippet
Code:
for (u32 id = 1; id <= RT_TABLE_MAX; id++)
GCC will skip the loop, and Sourcery will have a hard time doing nothing.
There are more examples like calculating the location of physical memory or signed/unsigned comparisons. The compiler switches -fwrapv and -fstrict-overflow might influence things, but it general the behaviour is hardcoded and both compilers have a different flavour. I think it would be wiser to choose the GCC flavour as it is more widespread and thus better tested (and fixed).
i) GCC. I noticed that early versions of kernels compiled with GCC would not start. At first I thought it was because of Sourcery /GCC code generating differences. After a number of buxfixes (in error.patch) I suddenly noticed that the GCC kernel is working. My phone is running a GCC compiled production configured kernel for nearly a week.
j) "Houston, we've had a problem" with the light sensor. One of the compiler warnings brought me to the file drivers/sensor/optical/gp2a.c. There within are located two routines which read the light and proximity sensor. They seem copy-pasted identical, however the sensor value types are different as the proximity value is a char and the light intensity a double. What is more convenient than to simply change the data type of the supplied buffer in the function prototype. Now headache starts as the semantics of the read (and write) call say that the unit size is byte. So returning "1" indicates that only the first byte of the sensor value is copied. Also, there is no bounds/access checking so supplying an invalid pointer to the call will crash the kernel. So, assuming this is all one big mistake, I redesigned the function to do better (see houston.patch) and built a new kernel with it. To my utter surprise my battery charge extended from <24 hours to 2 days and 20 hours.
However... I also noticed that my backlight intensity level was constant at it's lowest although the setting was set to auto. I needed to know how the caller invokes the call, but after an extensive search of internet and android sources it is still something I have not found. Heuristics show that the reading the light sensor is called with a buffer length of 1, and the returned value is only accepted when returning a 1 and that the sensor value type is a double (8 bytes). This is wrong: read() semantics require that you supply a length of 8, and expect a return value of 8. This may be the base of many light sensor issues I found when Googling.
Anyway, I returned the code to it's original faulty behaviour, and being illuminated I disabled the auto backlight intensity and changed it to it's lowest setting to enjoy a longer life between battery charges.
TO CREATE YOUR KERNEL:
1) Prepare a working environment
1a) Unpack Sourcery G++. No installation needed, unpacking is sufficient
1b) Unpack the Samsung sources and cd to the location of the top-level Makefile.
1c) Unpack the zImage and the contents of the Tweak-Kit to the same location
1d) Make sure the zImage is called zImage.I9000XWJP6
2) The ramdisk image is required and can be extracted from zImage.I9000XWJP6
2a) Create an uncompressed image Image.I9000XWJP6
Code:
gcc scripts/binoffset.c -o scripts/binoffset
ofs=`scripts/binoffset zImage.I9000XWJP6 0x1f 0x8b 0x08 0x00 2>/dev/null`
dd ibs=$ofs skip=1 <zImage.I9000XWJP6 | gzip -c -d >Image.I9000XWJP6
2b) The Tweak-Kit includes mkvmlinux which converts the uncompressed binary image into a bfd object. You need an Arm enabled libbfd to get it working. This does not get installed by default so you need to deeplink into binutils. mkvmlinux locates and decodes the kallsym data and econstructs the symbol table. It then uses the values of __initramfs_start/end to extract the initramfs. If you are not bothered with the hassle, just use dd with hardcoded values.
Code:
g++ mkvmlinux.cc -o mkvmlinux -lbfd -liberty -lz [-I and -L that deeplink into binutils]
./mkvmlinux Image.I9000XWJP6 vmlinux.I9000XWJP6 -r initramfs.cpio
or
Code:
dd if=Image.I9000XWJP6 of=initramfs.cpio bs=1 count=2739712 skip=165568
3) Patch date/time and other environmental issues to the moment of original creation
Code:
patch -p1 <stamp.patch
4) Make your computer happy
Code:
# edit Makefile line 184 and update the macro CROSS_COMPILE=
cp I9000XWJP6_defconfig arch/arm/configs/I9000XWJP6_defconfig
make I9000XWJP6_defconfig
make
5) Verify that the kernel is identical
Code:
diff zImage.I9000XWJP6 arch/arm/boot/zImage
AND NOW FOR SOMETHING COMPLETELY DIFFERENT...
Tweaking the configuration will build you a new kernel but when your Galaxy powers on it will either die silently (hang) or experience a horrible death (reboot). The problem is that there are modules built from sources located in the removed directories drivers/fsr and fs/rfs. These modules were compiled with a specific data structure layout and entry points. These will surely change when re-configuring. The way to keep the non-native modules happy is to keep the structures and entry points intact.
The structure layout is influenced by the CONFIG_ macros. The structures can be fixed to reflect the state of the I9000XWJP6 kernel by replacing the CONFIG_ macro's by something that does not change after reconfiguration. For that I use a collection of 'shadow' macro's which have SHADOW_ as prefix. Because the data structures cannot expand, you cannot (easily) enable configure functionality which require extra fields in the data structures. Reducing functionality is highly seldom a problem.
If changing kernel functionality results in removal of entry points, then stubs are required for those entry points needed by the non-native modules
There are automated methods to verify that a new kernel abides to the above constraints. For the data structures the compiler must generate gstabs debug information. This is human readable and includes detailed structure descriptions. This information should be identical across re-configuration. However, the scripts get confused by anonymous structs which are by product of "typedef struct {" constructions. These need to be named, something shadow.patch also does.
The kernel modules have easily-readable symbol tables containing needed kernel entry points. These should all be present in all re-configured kernels. Validation tests that fail emit enough information to further fix data structures and entry points. The Tweak-Kit contains two files: I9000XWJP6.h containing the SHADOW_ macro's and I9000XWJP6.c for the stubs. Both were constructed in an on-demand basis for the reconfiguration I am currently using and both serve as good examples on what to do when validation fails.
Before reconfiguring, rebuild the kernel for usage as a validation checkpoint.
1) Undo the timestamp patches
Code:
patch -R -p1 <stamp.patch
2) Fix the warnings
Code:
patch -p1 <style.patch # style related issues
patch -p1 <prototype.patch # prototype related issues
patch -p1 <error.patch # bug fixing
3) Apply datastructure fixation, entrypoint stubbing and Makefile tweaking
Code:
patch -p1 <shadow.patch
cp I9000XWJP6.c arch/arm/plat-samsung
4) Before recompiling everything, you need to issue "make clean" first. However, the missing directories will now pose a problem as "make clean" will include their Makefiles and will fail if it can't. Just create empties to keep the build happy.
Code:
mkdir -p drivers/fsr fs/rfs
touch drivers/fsr/Makefile fs/rfs/Makefile
5) Optionally change the Makefile to point to your favorite compiler/toolchain. Please note that I am using GCC 4.4.5. GCC 4.5.1 is bumping into problems I haven't looked into yet.
Code:
# edit Makefile line 184 and update the macro CROSS_COMPILE
6) This build will generate gstab debug information. Unexpectingly, this might bite when combined with function profiling, so disable that. But do not CONFIG_FUNCTION_TRACER yet as that does more.
Code:
# edit Makefile line 553, disable the line containing KBUILD_CFLAGS += -pg
7) Unpack the initramfs image. The directory /lib/modules needs to be examined/updated
Code:
mkdir initramfs.dir
cd initramfs.dir
cpio -i --make-directories --preserve-modification-time --no-absolute-filenames <../initramfs.cpio
cd ..
8) Repack initfs as a tarball, as make clean will erase all the modules
Code:
tar cf initramfs.tar initramfs.dir
9) The initramfs image will contain new kernel modules. Make sure a new version will get generated.
Code:
# in .config line 80 point to the unpacked initram location
CONFIG_INITRAMFS_SOURCE="initramfs.dir"
# in .config lines 86-89, select your favourite compression
CONFIG_INITRAMFS_COMPRESSION_NONE=N
CONFIG_INITRAMFS_COMPRESSION_GZIP=Y
10) Build a new kernel
Code:
# not cleaning will confuse the verification
make clean
make CONFIG_DEBUG_INFO=y
# install the modules
tar xf initramfs.tar
cp `find drivers -name '*.ko'` initramfs.dir/lib/modules
# rebuild with a fresh new ramdisk image
rm usr/initramfs_data.cpio*
make CONFIG_DEBUG_INFO=y
11) Checkpoint structure/entrypoint information. This is architecture independent.
Code:
# extract structures. They are the entries with :T
objdump -G vmlinux | awk '{ print $7 }' | grep :T | sed 's/([^)]*)/()/g' | sed 's/=\*()//g' | sort -u > gstabs.ckp
# extract the entrypoints
nm vmlinux | grep 'r __ksymtab_' | awk '{ print $3 }' | sort > ksymtab.ckp
12) Do a test-run. Pack zImage and flash with Odin. If your Galaxy is up and running, I strongly suggest you make a backup of your environment. If you later change something and it breaks, then this is the best place to restart.
Code:
cp arch/arm/boot/zImage .
tar cf I9000XWJP6-2.6.32.9-test.tar zImage
13) Make your re-configuration. I really suggest you do not make too many changes in one go because it gives more work when the structure/entrypoint verification fails.
Code:
# re-configure. For this exercise, change the kernel to a more production version
CONFIG_CC_OPTIMIZE_FOR_SIZE=Y
CONFIG_DM_DEBUG=N
CONFIG_S3C_KEYPAD_DEBUG=N
CONFIG_DEBUG_FS=N
CONFIG_DEBUG_KERNEL=N
CONFIG_LATENCYTOP=N
CONFIG_FTRACE=N
CONFIG_ARM_UNWIND=N
CONFIG_DEBUG_USER=N
14) Build a new kernel
Code:
# not cleaning will confuse the verification
make clean
make CONFIG_DEBUG_INFO=y
# install the modules
tar xf initramfs.tar
cp `find drivers -name '*.ko'` initramfs.dir/lib/modules
# rebuild with a fresh new ramdisk image
rm usr/initramfs_data.cpio*
make CONFIG_DEBUG_INFO=y
You get "struct has no member named" errors if you have enabled subsystems that require data structures to change which are incompatible with the non-native modules.
15) Verify structure/entrypoint checkpoint
Code:
# extract/verify structures
objdump -G vmlinux | awk '{ print $7 }' | grep :T | sed 's/([^)]*)/()/g' | sed 's/=\*()//g' | sort -u > gstabs.t
# new/changed structures are tagged with '+'. Display only the changed ones
diff -U0 gstabs.ckp gstabs.t | grep '+' | grep ':T' | sed 's/+//' | sed 's/:T.*/:T/' | while read s; do
grep -q "$s" gstabs.ckp
if [ $? -eq 0 ]; then
echo $s;
fi
done
# extract/verify entrypoints
nm vmlinux | grep 'r __ksymtab_' | awk '{ print $3 }' | sort >ksymtab.t
# extract all entrypoints needed by the modules
nm `find initramfs.dir/ -name '*.ko'` >allkosym
# some symbols are referenced in other modules. Cross-reference and remove from list
grep ' U ' allkosym | sort -u | awk '{ print $2 }' | while read s; do
if ! egrep -q "^[0-9a-f]* . $s\$" allkosym; then
echo $s;
fi
done > allusym
# check that kernel has entrypoints for all final undefined symbols
cat allusym | while read s; do
if ! egrep -q "^__ksymtab_$s\$" ksymtab.t; then
echo $s;
fi
done
Both scripts will generate output if you have enabled subsystems that require datastructures to change. These do not necessarily have to be data structures needed by non-native modules. However, missing entrypoints are those used by the modules. If it's about datastructures, your best chance is to lookup the data type and see if there any #ifdef CONFIG_ macros that need to be changed into #ifdef SHADOW_. If it's a missing entry point, you need to add a stub in I9000XWJP6.c.
16) Do a test-run. Pack zImage and flash with Odin.
Code:
cp arch/arm/boot/zImage .
tar cf I9000XWJP6-2.6.32.9-test.tar zImage
17) If you want more, jump to step 13
18) When you are really done, rebuild a final and fresh kernel and initramfs image with debugging stuff removed. The -gstabs compiler switch slightly influences code generation.
Code:
# not cleaning will confuse the verification
make clean
make
# install the modules
tar xf initramfs.tar
cp `find drivers -name '*.ko'` initramfs.dir/lib/modules
# rebuild with a fresh new ramdisk image
rm usr/initramfs_data.cpio*
make
My uncompressed image has now shrunk from 14700623 to 11822559 bytes.
Happy Hacking...
[...and now to find a better workaround for those non-native modules.]
WoW, Nice work !! very good info for beginners like me
thx a lot for this tut and i've learnt a lot
btw, seems there r some typos or something is missing. i did it with (XXJPO):
Hexabit said:
make I9000XWJP6_defconfig
Click to expand...
Click to collapse
make defconfig I9000XWJP6_defconfig
- modified include/linux/a.out.h by removing the 2nd def for SEGMENT
- changed the boolean to lowercase for .config
PS i use the cpio extracted by myself coz i couldnt enable libbfd on my ubuntu x64
Good tips. Thanks.
Really insightful i hope the dev take all the tweaks into consideration and make a new optimized kernel
good job here!
I think it's the most amazing first post ever! It should be sticked or kept somewhere safe.
Awesome first post. Will have to work through this.
Great post with very interesting findings!
I'm no expert, so maybe my question is a bit silly:
Is Samsung's published code just a buggy and incomplete pre-release debug version? Then how can e.g. Voodoo get a good working kernel?
Or is the official firmware really built of this, so possibly full of strange bugs and missing optimizations?
For SHW-M110s development/discussion. (Korean Anycall SGS ONLY yes for any who don't know, we are officially part of the I9000 forum.
Koe1974 suggested this thread and will I think act as a co-OP on this discussion. Look for front page updates, links, whatever from him 3 posts down, (below the lost guy from China) in the future too.
Version 4 released
Version 4.0r1 released to fix Odin flashing problem
link to kernel
with tegrak_voodoo sound (v2) module, tegrak ext4 module, root (superuser.apk) and busybox 1.17.1 optional, safe mount option overides by default, auto detect lag fixed partitions (improves upgradeability and interchageability) [/B][/size] Compatible with previously z4modded ext2 setups. All ROM versions, SK05 through TA13 tested and released (link is below).
안녕하세요 to any of the Korean developers who find this. Please update us in English about what you are doing. Your English is probably MUCH better than my korean, and I live in Korea.. There are some people in China, Iran, Philippines, etc using this device who might be helped.
-------------------------------------------------------------------------------------------------------------------------
Ext4 (tegrak modules) z4build rooted voodoo sounds kernel link
A stock z4moded kernel with added ext4 support ( presently by "stealing" tegrak kernel modules.) and many tweaks to make it actually work.
This is now working with z4control to get an easy to apply ext4 lag fix!
------------------------------------------------------------------------------------------------------------------------------------
Bug discussion here please, usage discussion (how do get the file into odin?) .. maybe the general thread is better.
For now, our rooting guide is linked from my sig also, although the above linked kernel can also be made to provide root with no effort.
This type of initramfs modification can be done by unpacking a stock kernel making customizations and then applying z4build to it. But I used a z4modded kernel and then applied customizations and repacked. I did this because I set up to repack by hand anyway before knowing I wanted to use z4mod and even then, when I thought I needed to pack into a tegrak kernel (which z4build can't do). It turned out to cause complications, but also taught me a good bit about the process and about z4mod, and probably helped me find bugs.
Some other related useful links for Reference:
kernel extraction (commented by me specifically for tegrak lzma compressed initramfs and z4build split initramfs)
http://forum.xda-developers.com/wiki/index.php?title=Extract_initramfs_from_zImage
kernel repacking
http://forum.xda-developers.com/showthread.php?t=789712
It needs the initramfs to be cpio'd already something like this:
Code:
cd $initram
find ./ | cpio -H newc -o > $repackdir/newramfs.cpio
It also needs the editor.sh script modified to point to the cross-compiler.
By default it can only pack you initramfs into an image which previously had an uncompressed initramfs. This can be modified though easily.
kernel compiling
This old thread probably isn't too useful now..
http://forum.xda-developers.com/showthread.php?t=740740&page=2
We have much better info 3 posts down by koe1974.
Kernel sources here:
http://opensource.samsung.com/
search SHW-M110S.
The first froyo update has a nice readme with a link to the compiler (I don't have the link right this moment)
Not sure if we need to figure/find .configs, or if the ones included are ok to start. Just need to compile one once and see.
I got my compiler toolchain here:
http://www.codesourcery.com/sgpp/li...1-188-arm-none-eabi-i686-pc-linux-gnu.tar.bz2
I thought it's the one recommended is the original froyo source from samsung, but Koe says they point to 2009 version, so I don't know now.
............
I flashed the korean voodoo sk22 kernel just to boot it into recovery. It has CW mod recovery. Might be useful. (BTW I don't recommend flashing unfamiliar kernels haphazardly, ex: this one injected a file into my ROM that interefered with z4mod until I realized it ) Mine BTW injects only one empty directory (/etc/init.d) and nothing else. Remove the kernel and all other changes dissappear. (z4control adds a tad more, but very little, most of its additions self destruct after use.)
Post reserved.
Post Removed, as requested.
Sent from my GT-I9000 using XDA App
Links:
Korea's equivalent to XDA
Lilinser's GitHub - kernel repack, deodexer, etc.
Project-Voodoo - initramfs (SK05)
Just for hobbies - Voodoo for SL28
Tegrak Kernel
SHW-M110S intramfs Requires further research.
Older M110S from someone at MIT working on the M110S potential resource
bml7 & initramfs possible resource
more initramfs
initramfs SK05 Tested .. OK
How-Tos:
Basic How-To Build Environment with built kernel test
First this diff is from a z4modded stock SL28 image to the custom kernel image. It's not against stock. So it includes tegrak files and scripts I changed. Also note the -N option. It pretends like files that don't exist do.. Any binary files that "differ" are actually added from tegrak.
Code:
diff -rbpN z4mod_sl28/initramfs/init.rc z4grak-construction-sl28/initramfs/init.rc
*** z4mod_sl28/initramfs/init.rc 2011-01-13 02:20:12.000000000 -0500
--- z4grak-construction-sl28/initramfs/init.rc 2011-01-11 07:43:34.000000000 -0500
*************** loglevel 3
*** 58,63 ****
--- 58,72 ----
mount j4fs /dev/block/stl6 /mnt/.lfs
insmod /lib/modules/param.ko
+ #ext4 modules by woo
+ insmod /tegrak/lib/modules/mbcache.ko
+ insmod /tegrak/lib/modules/jbd2.ko
+ insmod /tegrak/lib/modules/ext4.ko
+
+ # tegrak system lagfix by woo
+ #
+ insmod /tegrak/lib/modules/tegrak_module.ko
+
# Backwards Compat - XXX: Going away in G*
symlink /mnt/sdcard /sdcard
*************** service vt /system/bin/vtserver
*** 728,734 ****
#user system
#group system
-
service dumpstate /system/bin/dumpstate -s
socket dumpstate stream 0660 shell log
disabled
--- 737,742 ----
*************** service dumpstate /system/bin/dumpstate
*** 739,747 ****
# oneshot
-
# Added by z4mod
service z4postinit /init
oneshot
--- 747,761 ----
# oneshot
# Added by z4mod
service z4postinit /init
oneshot
+ #install root ingore the mount type, it doesn't matter
+ # syntax looks a little strange to me.. we'll see if it works
+ mount rfs /dev/block/stl9 /system rw remount
+ cat /sbin/su > /system/bin/su
+ chown root /system/bin/su
+ chmod 4755 /system/bin/su
+ mount rfs /dev/block/stl9 /system ro remount
\ No newline at end of file
diff -rbpN z4mod_sl28/initramfs/lpm.rc z4grak-construction-sl28/initramfs/lpm.rc
*** z4mod_sl28/initramfs/lpm.rc 2011-01-13 02:20:12.000000000 -0500
--- z4grak-construction-sl28/initramfs/lpm.rc 2011-01-11 06:37:28.000000000 -0500
*************** on init
*** 16,21 ****
--- 16,26 ----
insmod /lib/modules/param.ko
insmod /lib/modules/vibrator.ko
+ #ext4 modules by woo
+ insmod /tegrak/lib/modules/mbcache.ko
+ insmod /tegrak/lib/modules/jbd2.ko
+ insmod /tegrak/lib/modules/ext4.ko
+
mount rfs /dev/block/stl9 /system check=no
mount rfs /dev/block/mmcblk0p2 /data nosuid nodev check=no
Binary files z4mod_sl28/initramfs/sbin/sslvpn and z4grak-construction-sl28/initramfs/sbin/sslvpn differ
Binary files z4mod_sl28/initramfs/tegrak/bin/mkfs.ext4 and z4grak-construction-sl28/initramfs/tegrak/bin/mkfs.ext4 differ
Binary files z4mod_sl28/initramfs/tegrak/bin/tune2fs and z4grak-construction-sl28/initramfs/tegrak/bin/tune2fs differ
Binary files z4mod_sl28/initramfs/tegrak/lib/modules/ext4.ko and z4grak-construction-sl28/initramfs/tegrak/lib/modules/ext4.ko differ
Binary files z4mod_sl28/initramfs/tegrak/lib/modules/jbd2.ko and z4grak-construction-sl28/initramfs/tegrak/lib/modules/jbd2.ko differ
Binary files z4mod_sl28/initramfs/tegrak/lib/modules/mbcache.ko and z4grak-construction-sl28/initramfs/tegrak/lib/modules/mbcache.ko differ
Binary files z4mod_sl28/initramfs/tegrak/lib/modules/tegrak_module.ko and z4grak-construction-sl28/initramfs/tegrak/lib/modules/tegrak_module.ko differ
I'm a bit confused about sslvpn It's in my SL28 , it's not in my z4moded SL28. everything else diff as expected. Maybe I just lost it, maybe z4mod removed it. It's a small unimportant mystery.
BTW it looks like what I believe are the recovery keys have changed from SK22 to SL28, so maybe using the wrong kernel breaks something.
and my slightly modified version of the extraction script with commented lines to deal with lzma.
it's much faster (well.. why not), and it handles direcories a little better. It need a "/" somewhere in the file name though so use "./zImage".
Code:
#!/bin/bash
#MUCH faster than dd bs=1 skip=blah
#
# syntas is fastdd file skip <length_in_bytes_optional>
# skip is NOT optional and should be set to 0 read from begining.
#
fastdd () {
#dd with a skip is crazy slower cause it forces bs=1
#credit goes to somebody on the internet.
local bs=1024
local file=$1
local skip=$2
local count=$3
(
dd bs=1 skip=$skip count=0 2>/dev/null
if [[ "$count" != "" ]]; then
dd bs=$bs count=$(($count / $bs))
dd bs=$(($count % $bs)) count=1
else
dd bs=1024 2> /dev/null
fi
) < "$file"
}
zImage=$1
basedir=${1%/*}
echo working directory $basedir
mkdir $basedir/initramfs
outdir=$basedir/initramfs/
#========================================================
# find start of gziped kernel object in the zImage file:
#========================================================
pos=`grep -P -a -b -m 1 --only-matching $'\x1F\x8B\x08' $zImage | cut -f 1 -d :`
echo "-I- Extracting kernel image from $zImage (start = $pos)"
echo
echo "*** Start of compressed kernel image:" $pos
#========================================================================
# the cpio archive might be gzipped too, so two gunzips could be needed:
#========================================================================
fastdd $zImage $pos | gunzip > /tmp/kernel.img
pos=`grep -P -a -b -m 1 --only-matching $'\x1F\x8B\x08' /tmp/kernel.img | cut -f 1 -d :`
#
# Use next one for tegrak secuere 11 SL28
# It's an lzma header
# It's found by looking for 5D 00 in the exact same place as cpio (070701) is found in stock.
# The long string of FF's is the real give away since 5D 00 is too vague.
# pos=`grep -P -a -b -m 1 --only-matching '\x{5D}\x{00}\x..\x{FF}\x{FF}\x{FF}\x{FF}\x{FF}\x{FF}' /tmp/kernel.img| cut -f 1 -d :`
echo
echo "*** gzip position in kernel.img :" $pos "(start of gzipped cpio)"
#===========================================================================
# find start and end of the "cpio" initramfs image inside the kernel object:
# ASCII cpio header starts with '070701'
# The end of the cpio archive is marked with an empty file named TRAILER!!!
#===========================================================================
if [ ! $pos = "" ]; then
echo "-I- Extracting compressed cpio image from kernel image (start = $pos)"
# use either one of the next two lines for gzip
# dd if=/tmp/kernel.img bs=1 skip=$pos | gunzip > /tmp/cpio.img
fastdd /tmp/kernel.img $pos |gunzip > /tmp/cpio.img
# comment above and uncomment one of next two lines for lzma, if decompressing tegrak image.
# dd if=/tmp/kernel.img bs=1 skip=$pos | unlzma > /tmp/cpio.img
# fastdd /tmp/kernel.img $pos | unlzma > /tmp/cpio.img
start=`grep -a -b -m 1 --only-matching '070701' /tmp/cpio.img | head -1 | cut -f 1 -d :`
end=`grep -a -b -m 1 --only-matching 'TRAILER!!!' /tmp/cpio.img | head -1 | cut -f 1 -d :`
inputfile=/tmp/cpio.img
else
echo "-I- Already uncompressed cpio.img, not decompressing"
start=`grep -a -b -m 1 --only-matching '070701' /tmp/kernel.img | head -1 | cut -f 1 -d :`
echo start $start
end=`grep -a -b -m 1 --only-matching 'TRAILER!!!' /tmp/kernel.img | head -1 | cut -f 1 -d :`
echo end $end
inputfile=/tmp/kernel.img
fi
end=$((end + 10))
count=$((end - start))
if (($count < 0)); then
echo "-E- Couldn't match start/end of the initramfs image."
exit
fi
echo "-I- Extracting initramfs image from $inputfile (start = $start, end = $end)"
echo inputfile: $inputfile
echo start $start
echo count $count
echo outdir $outdir
# dd if=$inputfile bs=1 skip=$start count=$count > $outdir/initramfs.cpio
fastdd $inputfile $start $count > $basedir/initramfs.cpio
cd $basedir
basedir=`pwd`
cd $outdir; cpio -v -i --no-absolute-filenames < $basedir/initramfs.cpio
cp /tmp/kernel.img $basedir/
More to come.
I like the way this is going, appagom, please put [MOD] in the title.
GL on the new thread, if you hope to have Koreans stumbled upon the thread it might be good to add more phrases like:
루팅
갤럭시s
안드로이드
프로요
업그레이드
업데이트
I never really fully utilized it myself. I hope you get more done here or I'll just take the move personally. Actually, now that you guys were moving into compiling and building I thought this would come. Also, you should look to get some Soju out of this with some donate links or something, even if you aren't looking to take in any cash perhaps you could use it as a seed pot for bounties.
Most importantly, I need to know what "the lost guy from China" said...ㅋㅋㅋ
@Koe, don't waste your time on Gingerbread, get us Honeycomb
Compile a Kernel in 13 Lucky Steps
If you are not familiar with Linux, you might have a hard time following this. Just as I am writing how to do get setup to compile a kernel and compile Android apps, others have written how to setup VirtualBox, Ubuntu, AndroidSDK, etc. Please see documentation provided by Oracle, Google, Ubuntu, etc. before you ask for help about VirtualBox, Ubuntu and the SDK. Thanks.
This is written for people who have used Linux but have not compiled much. Or for the brave at heart who are looking for a nice weekend project. This will tell/guide you through getting a system setup that will not destroy your existing OS.
My host OS is Ubuntu 10.10 64-bit. I wanted had to make an Ubuntu 10.04 32-bit 64-bit system for development because I didn't want to deal with 64-bit vs. 32-bit issues but since Gingerbread requires a 64-bit compiler, I had to. Might as well just use my host system, but since we're here let's keep going!
So you're about to begin. Let's just get one thing straight. No! This will not result in a kernel you can flash. “Then why do this?, you ask. Ask yourself that!
1. Install VirtualBox and the Oracle VM VirtualBox Extension Pack (Currently 4.0.0 r69151)
http://www.virtualbox.org/wiki/Downloads
2. Download an Ubuntu ISO (I suggest Ubuntu Desktop 10.04 64-bit)
http://www.ubuntu.com/desktop/get-ubuntu/download
3. Create a new VM in VirtualBox (You may/have to modify the settings)
Operating System: Linux
Version: Ubuntu64
Extended Features: IO APIC
Processors: 2
RAM: 2048MB
Video Mem: 128MB
HDD: 32GB
Enable PAE/NX
Shared Drive: (I use a shared folder, more detail later)
Click to expand...
Click to collapse
4. Install VirtualBox Guest Additions
5. Install Ubuntu and Update Ubuntu
TIP: Mount your Shared Folder with fstab (Optional but helpful)
If you chose to use a shared folder you can auto-mount it via /etc/fstab.
NOTE: I use a shared folder named andDEV and I mount it on my desktop (~/Desktop). Below is what I add to my /etc/fstab (You may/have to change it)
Code:
andDev /home/koe/Desktop/andDev vboxsf uid=1000,gid=1000 0 0
Click to expand...
Click to collapse
6. Enable multiverse and partner "Software Sources"
7. Install additional software: NOTE: I would also recommend installing preload, but it is not required.
Code:
sudo apt-get install qt3-dev-tools texinfo git-core gnupg flex bison gperf libsdl-dev libesd0-dev libwxgtk2.6-dev build-essential zip curl libncurses5-dev zlib1g-dev sun-java6-jdk eclipse ia32-libs
8. Get and Setup the AndroidSDK (Everything you need to know is there or just Google for help)
http://developer.android.com/sdk/index.html
NOTE: Ubuntu 10.04 does not have Java 5 in it's repositories. Follow this link to setup to Java 5
9. Get and Setup ADT Plugin for Eclipse (Everything you need to know is there or just Google for help)
http://developer.android.com/sdk/eclipse-adt.html#installing
TIP: At this point you might want to try Google's “Hello, Android” tutorial.
Click to expand...
Click to collapse
10. Download and Install the Sourcery G++ Lite for ARM EABI Toolchain (Currently arm-2010.09)
http://www.codesourcery.com/sgpp/lite/arm/portal/subscription3053
Look for and click the link for Recommended Release
Look for and click the link for IA32 GNU/Linux TAR
Extract the archive. You will have a folder named arm-2010.09
Make a directory in your home directory named CodeSourcery
Ex. mkdir ~/CodeSourcery
copy the entire arm-2010.09 folder into CodeSourcery
Click to expand...
Click to collapse
11. Update your $PATH
You should be familiar with this because you had to do it to setup the AndroidSDK
Append the following to your PATH in .bashrc
Code:
~/CodeSourcery/arm-2010.09/bin
12 Download and Prepare the Samsung Source Code (Currently SHW-M110S_Opensource_Froyo_update2.zip)
http://opensource.samsung.com/
Click Mobile - Mobile Phones
Look for and download SHW-M110S_Opensource_Froyo_update2.zip
Extract the archive. Inside the new folder are two more archives.
Extract SHW-M110S_Kernel.tar.gz Inside there is a new folder Kernel
You can copy this to a more convenient location. I copy it to my desktop.
In the Kernel folder is a file named Makefile. Open it with your editor of choice.
Go to line 184. You will see ...
CROSS_COMPILE ?= /opt/toolchains/arm-2009q3/bin/arm-none-linux-gnueabi-
You have to change it to something like below, but see the koe? That is my username so you have to change it to your username.
CROSS_COMPILE ?= /home/koe/CodeSourcery/arm-2010.09/bin/arm-none-eabi-
Save Makefile.
Click to expand...
Click to collapse
13. Compile a Kernel
NOTES:
1. Do not try to compile the code in your Shared Folder. It will fail.
2. When issuing these commands you will see lots of output during this part, most of which is not useful to you at this point.
3. The amount of time it takes for the final make command to run will depend on your computer.
Click to expand...
Click to collapse
Open a terminal window and move into the Kernel directory. Issue the following commands.
Code:
$ make shw-m110s_defconfig
$ make menuconfig
When the config editor opens do the following:
DOWN ARROW to Userspace binary formats and press ENTER
DOWN ARROW to Kernel support for a.out and ECOFF binaries and press SPACE
RIGHT ARROW to Exit and press ENTER
RIGHT ARROW to Exit and press ENTER
Press ENTER again and it will exit back to the command line
NOTE: If you have a powerful computer and you want to speed up the build time, make can be run as, make -j# The # represents how much it will try to do at once. $ make does 1 operation, make -j3 tries to do 3. I have a 2.66 Ghz dual-core CPU and I allow the VirtualBox guest OS access to both cores, so I use make -j3 The compile finishes in about 12 minutes and allows me to still use my host OS. For now, you might just want to run make without the -j option to get a full understanding of how long it takes. Later you can test with values.
Click to expand...
Click to collapse
WARNING: Time is relative. This will take some time ... go make a sandwich or maybe even watch a movie.
Code:
$ make
When you see $ again check the last couple of lines of output. You want to see …
OBJCOPY arch/arm/boot/zImage
Kernel: arch/arm/boot/zImage is ready
Click to expand...
Click to collapse
Congratulation! You just built a kernel for the m110s!
good write up.. Glad to see you hit the same a.out snag as me. Just makes me think it's not configured right and so I have low hopes. Will be fun to see what happens when you put an initramfs in it. I'd just unpack the stock one and try that first.
As for z4control.. I'm pretty interested in getting this working as something like this was the real reason I started messing with this. It seems there may be some issue with the z4mod's init wrapper not doing things it should but anyway.. just now seeing issues. flashed one kernel where I added some debug output added.. trying to understand it (not understanding it yet). I'm optimistic that I can track it down. edit:... definitely making progress, not quite there yet but getting closer.
appagom said:
good write up.. Glad to see you hit the same a.out snag as me. Just makes me think it's not configured right and so I have low hopes. Will be fun to see what happens when you put an initramfs in it. I'd just unpack the stock one and try that first.
Click to expand...
Click to collapse
Strictly to see if it would build completely and to get more info on how to do it, I did do an initramfs & kernel test build.
used the update2 kernel source
used the initramfs linked above (SHW-M110S intramfs Requires further research.)
ran find ./ | cpio -H newc -o > ~/Desktop/newramfs.cpio
added the cpio via menuconfig with no compression
It did build successfully and I ended up with a 7mb zImage vs. a 2.5mb.
There is no way in hell I am going to flash it because I do not know which initramfs (maybe from sk05) it is or what it contains but it did complete.
Now isn't this a kick in the nuts!
Since I got the toolchain all set I decided to focus on the Android source code. Following these directions .. http://source.android.com/source/download.html I got to "Building the code"
Code:
[email protected]:~/Desktop/myAnd$ source build/envsetup.sh
including device/htc/passion/vendorsetup.sh
including device/samsung/crespo/vendorsetup.sh
[email protected]:~/Desktop/myAnd$ lunch
You're building on Linux
Lunch menu... pick a combo:
1. full-eng
2. full_x86-eng
3. simulator
4. full_passion-userdebug
5. full_crespo-userdebug
Which would you like? [full-eng] 1
============================================
PLATFORM_VERSION_CODENAME=AOSP
PLATFORM_VERSION=AOSP
TARGET_PRODUCT=full
TARGET_BUILD_VARIANT=eng
TARGET_SIMULATOR=false
TARGET_BUILD_TYPE=release
TARGET_BUILD_APPS=
TARGET_ARCH=arm
TARGET_ARCH_VARIANT=armv5te
HOST_ARCH=x86
HOST_OS=linux
HOST_BUILD_TYPE=release
BUILD_ID=OPENMASTER
============================================
[email protected]:~/Desktop/myAnd$ make
============================================
PLATFORM_VERSION_CODENAME=AOSP
PLATFORM_VERSION=AOSP
TARGET_PRODUCT=full
TARGET_BUILD_VARIANT=eng
TARGET_SIMULATOR=false
TARGET_BUILD_TYPE=release
TARGET_BUILD_APPS=
TARGET_ARCH=arm
TARGET_ARCH_VARIANT=armv5te
HOST_ARCH=x86
HOST_OS=linux
HOST_BUILD_TYPE=release
BUILD_ID=OPENMASTER
============================================
Checking build tools versions...
build/core/main.mk:76: ************************************************************
build/core/main.mk:77: You are attempting to build on a 32-bit system.
build/core/main.mk:78: Only 64-bit build environments are supported beyond froyo/2.2.
build/core/main.mk:79: ************************************************************
build/core/main.mk:80: *** stop. Stop.
Well, 32-bit will do for now seeing how ::cough:: I'm not the one building gingerbread.
I guess my next how-to is going to be how to go out and buy a PC and setup a 64-bit development environment.
Well, 32-bit will do fro now see how ::cough:: I'm not the one building gingerbread.
I guess my next how-to is going to be how to go out and buy a PC and setup a 64-bit development environment.
Click to expand...
Click to collapse
Could try the gnu cross compiler I suppose maybe it's clever enough to work around it. Your computer really isn't 64 bit though? You could just install a 64 bit VM assuming it is.
koe1974 said:
added the cpio via menuconfig with no compression
Click to expand...
Click to collapse
Ok, that procedure is easier than it used to be, or than what I read it used to be. I'm not afraid to add one and flash it. Just busy with making one I like right now though.. then again.. sounds like a 10 minute interruption.. so.. we'll see... oh and I don't remember what was stopping xconfig from working, but something annoying.. it's quite a bit nicer than menuconfig. I need to learn how that works though.. cause really you should add modules from the compilation itself right.. so you cant have the initramfs already before you compile, unless make opens it and add the modules and re-cpio's it. Anyway.. I'll shoot first, ask later.
edit: just flashed.. it gave about 1 tenth of a blue bar and froze. Ok, that was fun, back to fixing and ext4 kernel up.
appagom said:
Could try the gnu cross compiler I suppose maybe it's clever enough to work around it. Your computer really isn't 64 bit though? You could just install a 64 bit VM assuming it is.
Click to expand...
Click to collapse
The VM I setup was 32bit, but it's 64bit now. I will just modify the How-To for a 64bit system.
4 bugs related z4control issues solved(ok 2 were probably the same as well as a fifth, so really 3)..
one, it's failing to create a directory it needs (/system/etc/init.d), reported, work around create by hand.
2) It's rfs formatting wrapper script which checked for voodoo, failed. Strange bug in sh actually.. workaround in script found and reported
bugs 3 and 4 details unimportant, were related to the mystery of where sslvpn went. z4build was splitting the initramfs by tacking it some of it brute force on to the end of the zImage. The init script found it again and unpacked the files before continuing. Oddly, it seems an uneeded because I fit the only big displaced file in just fine without splitting and I didn't even use heavy compression. Anyway, needed or not it caused two files to go missing, this one, and a small text file that reported the version of z4mod. this file prevented z4control from working.
I will not fix this tonight, but I mostly understand it (altough not in exact detail of precisely understanding some of the odd symptoms, but I think dealing with this will likely solve it all) My diffs could never find the version file, cuase I never unpacked it to know it even existed in the first place. Waaaah.. bug tracking is tiring. We should have this all running very shortly.
update: my first attempt at fixing the repacking bug.. resulted (I already feared this but hoped it would just go away) in a kernel that seems totaly complete, but where it seems something in the init wrapper isn't working... getting closer to finding it.
update: LAST BUG FIXED
The last bug(which I previously assumed was part of the repack bug) was actually aslo part of z4build, now documented here:
http://forum.xda-developers.com/showpost.php?p=10638535&postcount=1062
I'm pretty sure that one should really get it all going now. there are no lines of code left to break. I've tested almost every line now. I'll get a new kernel out, but will need to wait for an updated z4control. z4ziggy seems busy right now maybe (no complaint obviously).
I can fix this last bug in my roll out of the kernel. The other remaining ones can be worked around pretty easily, but not pretty for user instructions, better to wait probably.
SK05 Rebuild test (PASS)
We have doubts about if the shw-m110s_defconfig is what is used by samsung so I decided to try to make a working kernel from an initramfs and froyo_update source code.
General idea of what I did...
sk05 source code froyo_update (from samsung)
sk05 initramfs (link in post 4 "initramfs SK05 Tested .. OK")
cd initramfs_dir
find . -print0 | cpio -o0 -H newc | gzip -9 -f > some/path/initramfs.cpio.gz
cd kernel source code root directory
modify Makefile ln. 184
make aries_android_rfs_defconfig
make menuconfig
disable a.out
add path to initramfs archive
compression gzip
make
tar --format=ustar SK05rebuilt.tar zImage
Click to expand...
Click to collapse
Results in a zImage the same size (4.6mb) as samsung's, and it boots.
awesome..
I GOT IT
That was big fat B to solve. Debugging self destructing scripts that run in a startup environment with different mounts and different PATH variable is no fun and requires some imagination. I had to work around 2 more bugs in z4build but now I have user transparent workarounds to all of them.. plug and play. Unfortunately I have about 30 minutes more free time today so I probably wont post it today. I should clean up some comments and such in it first probably. We'll see.
appagom said:
awesome..
I GOT IT
Click to expand...
Click to collapse
Very nice!
Sent from my SHW-M110S using XDA App
the "awesome" comment was referring to you. Should be able to take the grak of z4grak with some compiling , but I'm happy with it in too.. doesn't matter.
Anyway.. I updated the kernel page (from the link on OP). The new kernel is there, but I pushed it out very fast. Didn't flash last repack myself yet after changing comments.. but I only changed comments. (I am running the posted copy now) It needs testing since my system has gotten a been used and non-pristine. Get z4control, you can use it to flash it if you want. the rest is all push button I hope.
Oh and default settings are less safe than Tegrak, not much if any better than ext2 actually. I think.. can't confirm right now. I recommend modifying etc/fstab to data=ordered,barrier=1 personally.. but that's what I like about z4.. you can decide for yourself.
Sent from my SHW-M110S using XDA App
Hello,
my Linux OS have a strange Python Version or setting, so the Build failed with:
Code:
FAILED: //art/build/apex:art-check-release-apex-gen generate art-check-release-apex-gen.dummy
Outputs: out/soong/.intermediates/art/build/apex/art-check-release-apex-gen/gen/art-check-release-apex-gen.dummy
Error: exited with code: 1
Command: rm -rf out/soong/.intermediates/art/build/apex/art-check-release-apex-gen/gen && out/soong/host/linux-x86/bin/sbox --sandbox-path out/soong/.temp --output-root out/soong/.intermediates/art/build/apex/art-check-release-apex-gen/gen -c 'out/soong/host/linux-x86/bin/art-apex-tester --debugfs out/soong/host/linux-x86/bin/debugfs --tmpdir __SBOX_OUT_DIR__ out/soong/.intermediates/art/build/apex/com.android.runtime.release/android_common_com.android.runtime.release/com.android.runtime.release.apex && touch __SBOX_OUT_FILES__' __SBOX_OUT_DIR__/art-check-release-apex-gen.dummy
Output:
.path_interposer: no python-exec wrapped executable found in /usr/lib/python-exec.
sbox command (out/soong/host/linux-x86/bin/art-apex-tester --debugfs out/soong/host/linux-x86/bin/debugfs --tmpdir out/soong/.temp/sbox333717900 out/soong/.intermediates/art/build/apex/com.android.runtime.release/android_common_com.android.runtime.release/com.android.runtime.release.apex && touch out/soong/.temp/sbox333717900/art-check-release-apex-gen.dummy) failed with err "exit status 127"
Its on the Finish-Line of the building Process. ( [ 99% 11533/11603] )
I try to use
Code:
virtualenv --version
virtualenv 20.0.18 from /usr/lib64/python3.6/site-packages/virtualenv/__init__.py
With the Python in:
prebuilts/python/linux-x86/2.7.5/bin/python2.7
Some hints for me?
Building Lineage 16, works fine, sure the build process is different.
Can i just use the Command from lineageos16, which create the zip file if all parts are finnished?
I've been having the same issue on Gentoo since LineageOS 17.1.
I have Python 2.7, 3.6, 3.7 and 3.8 installed.
If nobody has a solution and you find it, please comment with the solution, I'm sure we are not the only two people affected by this issue.
Thank you Yurienu,
easy would be to juse another live-Distro with a build Tree on a mounted file System.
However i think the error is in the tools of google, like apex or how the script use and call this function. Not sure if i like to debug this.
I think it because the path.interposer try to present a list of tools to art-check-release-apex-gen, from where the program to do its job. And a wrapper might failed and put its path there. When art-check do its job, try it and failed, because its not design to handle the error message (and so because of that bug).
For this speaks the error message:
But take a look on my error code: "Error: exited with code: 1" from the command: //art/build/apex:art-check-release-apex-gen generate art-check-release-apex-gen.dummy".
I think there is the bug. Whatever art-check-release-apex-gen do, it fail here. Then check, see the wrapper and say that there is no python.
I try to use python 2.7.18 (Gentoos Python 2.7), and it did not work.
I try to use virtuelenvwrapper*** and it failed but the error was like: "no Module with python 2.7.5, named virtuelenvwrapper". However i got this misconfigured. But it sound compelling. Because a Module for python 2.7.18, is maybe not byte compatible with python 2.7.5.
Think i try to use another distro for building this.
Bye the way, there is a Google Summer of code 2020 project, that's sound very interesting. So there is hope for a proper lineage and gentoo support in the future.
However, if i find a solution i share it here!
I stay a little bit longer with lineageos 16.0, because self build security updates are more important for me, as Android 10 is.
*** I just linked the prebuild binary an path as environment to /usr/local/bin/
Hello everyone,
I too met the same issue, the exact same content in the "out/error.log" than the first post; I am sharing the research I got around this issue, because I don't have the same conclusion.
First I tried to discard the error about ".path_interposer", and seek the error from the "sbox" software output.
I dig around "out/soong/host/linux-x86/bin/art-apex-tester", and got those information:
Code:
% file out/soong/host/linux-x86/bin/art-apex-tester
out/soong/host/linux-x86/bin/art-apex-tester: Zip archive data, made by v?[0x314], extract using at least v2.0, last modified Thu Nov
4 03:39:44 1999, uncompressed size 2619, method=store
% unzip -l out/soong/host/linux-x86/bin/art-apex-tester
Archive: out/soong/host/linux-x86/bin/art-apex-tester
Length Date Time Name
--------- ---------- ----- ----
2619 01-01-2008 00:00 __main__.py
28376 01-01-2008 00:00 art_apex_test.py
--------- -------
30995 2 files
It looks like it’s a Zip archive which can execute the Python script "__main__.py".
The "__main__.py" is a wrapper, which its goal is to call the other python file, "art_apex_test.py" with the correct parameters.
By reading its content, I noticed it searchs Python 3 executable, and calls the following command:
Code:
/usr/bin/python3 -S relative_path/to/art_apex_test.py […]
When I execute this command under a temporary directory, I have this output (still under the current directory as parent of the "out" directory):
Code:
python3 -S /tmp/art_apex_test.py --debugfs out/soong/host/linux-x86/bin/debugfs --tmpdir out/soong/.temp out/soong/.intermediates/art/build/apex/com.android.runtime.release/android_common_com.android.runtime.release/com.android.runtime.release.apex
--bitness=auto, trying to autodetect. This may be incorrect!
Detected multilib
I have the same ouput by either executing the Zip archive with the same arguments.
Unfortunately, the "tmpdir" "out/soong/.temp" remains empty.
What about "path_interposer" then? No idea, I found no match in either the two Python scripts.
And then I don't think it's related to the Python 2.7 embedded program.
Oddly enough, the full command with sbox has the same output, but not error. I start to think this error is in a build system file.
Hope to have help (a little).
Best regards,
thican said:
I too met the same issue, the exact same content in the "out/error.log" than the first post; I am sharing the research I got around this issue, because I don't have the same conclusion.
Click to expand...
Click to collapse
I've been having this issue a while also, turns out it's fixable (not that Gentoo won't override it every time you upgrade python) I linked /usr/bin/python3.7 to /usr/bin/python3 instead of /usr/bin/python-exec2c.
Now the apex tester thinger works fine, I'm guessing that inside the sandbox can't find /usr/lib/python-exec.
Hope this works for you guys as well.
SOLVED LineageOS 17.1: .path_interposer: no python-exec wrapped executable found
superfes said:
I've been having this issue a while also, turns out it's fixable (not that Gentoo won't override it every time you upgrade python) I linked /usr/bin/python3.7 to /usr/bin/python3 instead of /usr/bin/python-exec2c.
Now the apex tester thinger works fine, I'm guessing that inside the sandbox can't find /usr/lib/python-exec.
Hope this works for you guys as well.
Click to expand...
Click to collapse
OH! That's the trick, it finally works!
For an easy fix without needing to touch files hold by system, I just create a bin directory at the root of the home directory of the user, put its path inside the PATH environment variable, at its beginning (PATH="${HOME}/bin:${PATH}"), and then creating the symbolic links.
It was so easy it's so frustrating, having this issue for months, even during 16.0 branch IIRC.
Just to be sure, I create those links (I picked Python 3.8 for my example):
Code:
ln -s /usr/bin/python3.8 ~/bin/python3
ln -s python3 ~/bin/python
Now let's try this new build. Thank you very much for you help! :good:
PS: in your message, the links are backward, it's currently python3 which is linked to python-exec2c, and /usr/bin/python3.7{,m} are there own executable. I advice you to avoid editing your system files
@superfes and @thican you guys deserve three internets!
I was struggling with this error as well (not for 16.0, only for 17.1 -- Gentoo here as well). This fixed it!
So again: Thanks!
thican said:
OH! That's the trick, it finally works!
For an easy fix without needing to touch files hold by system, I just create a bin directory at the root of the home directory of the user, put its path inside the PATH environment variable, at its beginning (PATH="${HOME}/bin:${PATH}"), and then creating the symbolic links.
It was so easy it's so frustrating, having this issue for months, even during 16.0 branch IIRC.
Just to be sure, I create those links (I picked Python 3.8 for my example):
Code:
ln -s /usr/bin/python3.8 ~/bin/python3
ln -s python3 ~/bin/python
Click to expand...
Click to collapse
This is the way that worked fully for me. Thanks all! Running Gentoo latest hardened with python 3.7 as default.
Thank you superfes and thican!
I had not the time do investigate since covid19 blow up my personal workingdesk. Nice to go on with Lineage 17.1!
Tnx! Worked for me too!
With dev-lang/python-exec-2.4.6-r2
/usr/bin/python3 is a symlink to python-exec2c which cause this error.
Workaround is use vanilla python version of python-exec2c
before brunch command you can use like
Code:
[[ -d ~/bin/ ]] || mkdir ~/bin
[[ -e ~/bin/python3 ]] || ln -s /usr/lib/python-exec/python-exec2 ~/bin/python3
export PATH=~/bin:$PATH
Hence, in future shebang support of python or python3 will removed also.
762406 – [TRACKER] Packages that have failures with dev-lang/python-exec[-native-symlinks]
bugs.gentoo.org