Related
I've built it like I've always built my 2.3 kernel and I cant even get a stock kernel I compile to boot
you guys think it might be my CC? I've been using Sourcery G++ Lite 2011.03-42 for ARM EABI for a year almost and hasnt had any problems
I'm using supercurios kernel source also btw
any ideas?
ECOTOX said:
I've built it like I've always built my 2.3 kernel and I cant even get a stock kernel I compile to boot
you guys think it might be my CC? I've been using Sourcery G++ Lite 2011.03-42 for ARM EABI for a year almost and hasnt had any problems
I'm using supercurios kernel source also btw
any ideas?
Click to expand...
Click to collapse
"adb shell cat /proc/kmsg" output while it's trying to boot may be helpful.
I don't even get to that point, stops at Google logo and gives me nothing
PS I'm building for ICS (forgot to mention that)
Sent from my Nexus S 4G using Tapatalk
I'm not really sure. I have built from supercurio's source so I know it's not that.
Are you using this source (https://github.com/project-voodoo/linux_samsung)
Are you on the correct branch?
Have you made any changes?
Are you fastbooting the boot.img or using the anykernel script?
Hostile89 said:
I'm not really sure. I have built from supercurio's source so I know it's not that.
Are you using this source (https://github.com/project-voodoo/linux_samsung)
Are you on the correct branch?
Have you made any changes?
Are you fastbooting the boot.img or using the anykernel script?
Click to expand...
Click to collapse
No changes
Same source
Correct branch
Any kernel script
Only thing I haven't changed is CC =\
Sent from my Nexus S 4G using Tapatalk
ECOTOX said:
No changes
Same source
Correct branch
Any kernel script
Only thing I haven't changed is CC =\
Sent from my Nexus S 4G using Tapatalk
Click to expand...
Click to collapse
I have only ever used the one google supplies so I would say try that. It's the only thing I can think of
Hostile89 said:
I have only ever used the one google supplies so I would say try that. It's the only thing I can think of
Click to expand...
Click to collapse
Me too...I'll give a diff CC a try
Sent from my Nexus S 4G using Tapatalk
Solution
it was the CC, i used the default one and it worked.
apparently according to Ezekeel you need to add -mno-unalligned-access to the kbuild flags for the newer TC from CS
so if your using code sourcery add that to the build flags or just use the default toolchain
All,
Since we are working as a DEV team now I wanted to share a PM I had with another DEV on the board about our Kernel. I hope this can assist some others that have more experience with the Kernel like b_randon
g60madman said:
subpsyke,
I would love to upgrade our kernel for the Motorola Triumph. How did you go about figuring which Code Aurora vanilla kernel to start with? I took over development for CM7 from Whyzor and would be interested in your methodology.
Thanks in advance,
g60
Click to expand...
Click to collapse
subpsyke said:
1. I unpacked the original Huawei source to one directory, and downloaded the CodeAurora kernel to another.
2. I used the release tags as a reference: https://www.codeaurora.org/xwiki/bin/QAEP/froyo_almond
https://www.codeaurora.org/xwiki/bin/QAEP/froyo
https://www.codeaurora.org/gitweb/quic/la/?p=kernel/msm.git;a=tags
3. I reverted the codeaurora repository to tag releases (e.g.: git reset --hard M76XXTSNCJNLYA6010) and compared the result via "diff urN" and meld. I knew I was getting closer when you get a smaller diff in the patch size, and used meld to see if the differences were likely to be Huawei's additions vs. CodeAurora's changes. It was only after going forward and backwards between tags was I sure of the proper baseline.
I performed the same discovery process with the Samsung kernel for my GT-I5500, which used M76XXTSNCJNLYA6040 as a baseline.
Click to expand...
Click to collapse
g60madman said:
Sweet thanks! I will check it out and see what I can do
Click to expand...
Click to collapse
subpsyke said:
No problem
I forgot to mention the last steps:
4. Once you establish the baseline, create a new branch: git checkout -b newbranch
5. Overlay the changes from the vendor to a new commit*:
cp ~/blah/vendorkernel/ . ; git add . ; git commit -m "Initial import of vendor changes".
6. Once you've commited the vendor changes, you can use "git merge origin/froyo_almond" to move from the baseline to a newer revision. The froyo_almond branch is most suitable, as it's locked at 2.6.32, and the development focus seems to be on the msm7k chipset series. You could try updating to the android-msm-2.6.32 kernel, but it has more significant changes that will require adaptations of the board file, and may not be worthwhile, as development focus seems to be on newer chipsets.
7. Inevitably you'll get merge conflicts, as more than likely some vendor commits may interfere with upstream changes. For this, you'll need to use your own discretion in fixing up the code. I use the "git mergetool", with meld configured as my default editor, and manually checked all the conflicts.
Good luck!
*You may also want to fix up permissions etc., if your vendor source comes from a zip tarball. But it's purely cosmetic.
Click to expand...
Click to collapse
g60madman said:
So is almond the best flavor to start with. I have been using MSM/QSD for a while and in when TickerGuy originally created our device files for cm7 he listed in the readme
CAF information:
Branch: froyo_pumpkin
Tag: M7630AABBQMLZA2030
Didn't know if I should start with pumpkin or use almond. Let me know what you think?
Thanks again for the info!
Click to expand...
Click to collapse
subpsyke said:
Hmm... if your phone really has a msm7630 chipset, then yes, you should probably go for the froyo_pumpkin branch. The froyo_almond branch is only for the msm7627 and qsd8650 chipsets.
Click to expand...
Click to collapse
g60madman said:
Well thats the stupid thing, since the beginning we have used always used msm7x30 for our board config, But our stock ROM from Virgin Mobile the config was msm7k in the build.prop. However if you hit the Motorola Dev our pone clearly states Qualcomm MSM8655. I am leaning towards using the almond branch would that be correct?
Click to expand...
Click to collapse
subpsyke said:
According to wikipedia, it's MSM8655.
Look at the table here: https://www.codeaurora.org/xwiki/bin/QAEP/
The froyo_almond supports qsd8650, and froyo_pumpkin supports qsd8650a_st1x. I honestly have no idea what the difference is, but it's within the realm of possibility that your phone's chipset is that odd revision on the pumpkin branch.
If your kernel is using a pumpkin baseline, then you should continue along the same branch.
Click to expand...
Click to collapse
g60madman said:
I will download pumpkin and compare the kernel. I am not sure why TickerGuy started with pumpkin. That maybe what our build is based off of but I am not 100%.
Click to expand...
Click to collapse
subpsyke said:
Ok. When you clone the codeaurora git repository, you'll have all the branches included anyway.
Click to expand...
Click to collapse
....sympathy post...
Sent from my SPH-D700 using xda premium
(The froyo_almond supports qsd8650, and froyo_pumpkin supports qsd8650a_st1x.) Is this the difference between the Photon 4G which has a WiMax radio in it, and the Triumph which does not include a 4G radio? <--- nvm when they came out I had heard they were the same phone except 4G, apparently the Photon is a Tegra 2 device.
Okay so after looking into code Aurora more, as soon as I get my Ubuntu back up I am going to work on a vanilla froyo 2.6.32.9 kernel by using the froyo_pumpkin branch on the tag Karl gave us. Once I do that and have a commit that adds in Motorola changes, I am going to use the gingerbread_rel branch to try to get a 2.6.35 kernel booting on the phone, then use the ics_chocolate_rb7 branch to hopefully get the 3.x kernel.booting. those branches all seem to have the best support for msm7630 chipsets which I believe is the closest to the msm8655 chip only that it is clocked at 800mhz instead of 1ghz. If anyone else can lend any advice or help it would be swell!!
Sent from my Triumph using Tapatalk
b_randon14 said:
Okay so after looking into code Aurora more, as soon as I get my Ubuntu back up I am going to work on a vanilla froyo 2.6.32.9 kernel by using the froyo_pumpkin branch on the tag Karl gave us. Once I do that and have a commit that adds in Motorola changes, I am going to use the gingerbread_rel branch to try to get a 2.6.35 kernel booting on the phone, then use the ics_chocolate_rb7 branch to hopefully get the 3.x kernel.booting. those branches all seem to have the best support for msm7630 chipsets which I believe is the closest to the msm8655 chip only that it is clocked at 800mhz instead of 1ghz. If anyone else can lend any advice or help it would be swell!!
Sent from my Triumph using Tapatalk
Click to expand...
Click to collapse
Swell, I love that word. Here is an email I got from TickerGuy on the Kernel a fe months back
g60madman said:
TickerGuy,
I know you have moved on from the MT. Currently I have taken over development for CM7 from Whyzor and had a question for you.
When you designed the original device files you listed in the readme:
Branch: froyo_pumpkin
Tag: M7630AABBQMLZA2030
Is that really our branch from MSM/QSD? I know the phone has the MSM8655 chip. So I'm just trying to figure out why we use msm7x30 for the board configuration and not say msm7k or qsd8k?
Thanks in advance for any help you can offer
Click to expand...
Click to collapse
TickerGuy said:
I think the reason had to do with some of the peripheral chips -- it was a lot of fun getting this phone to work as it has a number of very odd things about it, especially in the GPS area.
Click to expand...
Click to collapse
So I think it's safe to say the route you are taking the right route. Also when building the Kernel do not forget to merge in the Wyzor fixes for the Video as I am using the new Andreno drivers. Just an FYI.
Yeah I will on the cm7 kernel. As of right now I'm gonna try to get a stock froyo kernel booting off code Aurora sources then go ffrom there.
Sent from my Triumph using Tapatalk
I've got my Linux mint 13 up and running so I'm going to try to get my build environment setup to build kernels either tonight or tomorrow and start pulling in source.
By the way, Linux mint 13 is pretty nice distro so far. I like it alot better than Ubuntu!
Sent from my Triumph using Tapatalk
b_randon14 said:
I've got my Linux mint 13 up and running so I'm going to try to get my build environment setup to build kernels either tonight or tomorrow and start pulling in source.
By the way, Linux mint 13 is pretty nice distro so far. I like it alot better than Ubuntu!
Sent from my Triumph using Tapatalk
Click to expand...
Click to collapse
Also you don't need to download the pumpkin branch simple download the kernel
git clone git://codeaurora.org/kernel/msm.git
git reset --hard M7630AABBQMLZA2030
The M7630AABBQMLZA2030 is the pumpkin branch and that should take us back to the vanilla kernel
Here is the pastebin link for the warning I was getting from the linker during the build of the code Aurora kernel.
http://pastebin.com/GLMBSz26
You can look at the kernel source on my github. Its the froyo pumpkin kernel repo.
The warnings cone from the gcc linker saying that it's trying to link a non executible section in built-in.o
I'm not sure where to start looking for the issue at. If anyone can lend any insight I would be grateful!
Sent from my Triumph using Tapatalk
I switched to the gcc 4.3.1 toolchain included with the cm7 source and it booted up. Worked just as good as the stock kernel. I'm gonna have to see why the newer toolchains are not compiling it right. I use linaro 4.6 on bKernel froyo which is based off motos source. So I don't see why it wont build this code right. But at least I got one to boot! !!
Sent from my Triumph using Tapatalk
b_randon14 said:
I switched to the gcc 4.3.1 toolchain included with the cm7 source and it booted up. Worked just as good as the stock kernel. I'm gonna have to see why the newer toolchains are not compiling it right. I use linaro 4.6 on bKernel froyo which is based off motos source. So I don't see why it wont build this code right. But at least I got one to boot! !!
Sent from my Triumph using Tapatalk
Click to expand...
Click to collapse
Good work brother!
Thanks. Now lets try to move on up to 2.6.35. My idea is to make a different patch between 2.6.32.9 and 2.6.35.7 and maybe that will simplify updating it. I'm not sure which gingerbread branch I'm gonna use for the 2.6.35 kernel!
Sent from my Triumph using Tapatalk
I would think the gingerbread branch, and use the M7630AABBQMLZA404025I.xml version. That is where I snagged the keyboard updates.
g60madman said:
I would think the gingerbread branch, and use the M7630AABBQMLZA404025I.xml version.
Click to expand...
Click to collapse
you can use source kernal .35 for device fih-fbo..we are the same drive only need to change touch driver..
Yeah I would use the fih kernels but we has issues with them rebooting on us.
Sent from my Triumph using Tapatalk
b_randon14 said:
Yeah I would use the fih kernels but we has issues with them rebooting on us.
Sent from my Triumph using Tapatalk
Click to expand...
Click to collapse
yes you need get a logcat and new baseband for this kernal...I have a file to solve the rendom reboot..but I can't help to get you for the baseband...
The kernel shouldn't have nothing to do with the baseband. Which file is it?
Sent from my Triumph using Tapatalk
b_randon14 said:
The kernel shouldn't have nothing to do with the baseband. Which file is it?
Sent from my Triumph using Tapatalk
Click to expand...
Click to collapse
yes I know,I mean in rom library need these file,I will give you file when I go back home
Anyone got any ideas to fix the issues with newer toolchains when building from the code Aurora source?
Sent from my Triumph using Tapatalk 2
Have you tried downloading one of the gingerbread repo's from code Aurora? I am not sure if they have a different version of the tool chain or not?
Hi all, so I've been wanting a really nice optimized Linaro recovery and have not been able to find one. So I decided to build my own and have found it to be very nice and stable, and of course to share with all of you xda peps
First off if you don't know what TWRP recovery is the original nexus 7 thread is HERE
Please read all of the original thread before flashing this recovery.
More info on how this recovery was built
Built using Linaro gcc 4.7 toolchain. I also built the toolchain from Linaro's gcc source. The toolchain source is HERE
This toolchain source gets update almost daily from linaro sources, but I don't normally have the time to build new toolchains daily. When I have time I will update it quite frequently.
Built off my own 4.1.2 kernel source. The kernel used to build this recovery was also compiled using the Linaro 4.7 toolchain.
I've added a few linaro recovery patches for interfaces to libpng. Those changes are HERE and HERE.
Built using Linaro bionic string routines optimizations.
Installation
Download the recovery image and flash in fastboot
Initial release 2.3.1.1 touch recovery
10.28.12 release 2.3.1.1 touch recovery
Kernel changes
Removed a lot of bloat from the kernel. I disabled GPU overclocking, user voltage control, cpu overclocking, a bunch of useless debugging stuff.
A short kernel changelog is HERE
Toolchain changes
Not much here. I included a static library to be used in the toolchain libiberty
A short changelog for the toolchain is HERE
Recovery changes
Built as engineering instead of userdebug.
11.5.12 release 2.3.1.1 touch recovery
Final android 4.1.2 linaro recovery version
Kernel changes
Mainline linux kernel upstream changes from linux-3.2.y
Changed kernel compression mode to GZIP and optimization level to -Os
Toolchain changes
Updated to latest linaro changes
Recovery changes
-O3 optimization level
Linaro strict-aliasing compiler flags optimization
Android-4.1.2 (This version is a final release, no more updates)
Size: 6.95 MB
MD5: e0f46f01556156b052b3779c9ed60e01
What? A Linaro recovery? I did not know there was such thing. I am downloading this now very excitedly.
Thank you. Very very helpful and nice.
OK... Now I need more info! I sorta understand the Linaro concept but my knowledge is limited. What's the reasons to base Recovery on it at this point? Any advantages, possible concerns? Will there be any noticeable differences? Just curious & wondering cause you said "you'd been wanting to make a recovery based on linaro".
Thank!
Sent from my Nexus 7 using Tapatalk 2
Hi men!
Thanks for your recovery.
But i experience some strange visual effects like distortion of the image or some lag effects.
---------- Post added at 04:21 PM ---------- Previous post was at 04:16 PM ----------
djd338 said:
OK... Now I need more info! I sorta understand the Linaro concept but my knowledge is limited. What's the reasons to base Recovery on it at this point? Any advantages, possible concerns? Will there be any noticeable differences? Just curious & wondering cause you said "you'd been wanting to make a recovery based on linaro".
Thank!
Sent from my Nexus 7 using Tapatalk 2
Click to expand...
Click to collapse
There is a discussion on this subject regarding some tests Ezekeel (XDA developper) made with different cross-compiler toolchains and those tests prooved that none of the compilers is better than another.
We heard a lot about linaro because when ICS was released, it was very laggy and linaro and is new compiler version 4.7 made ICS much smoother than before.
But for the pur performance linaro give no improvement if you compare with another one.
EDIT: I found the link of the test for you: http://forum.xda-developers.com/showpost.php?p=19872366&postcount=1
i remember this test,and also in my home test when i tried to build kernel,i don't see improvement using linaro or others toolchains..anyway it's great to have another thing to play on and see if it's best that the ufficial!
sert00 said:
i remember this test,and also in my home test when i tried to build kernel,i don't see improvement using linaro or others toolchains..anyway it's great to have another thing to play on and see if it's best that the ufficial!
Click to expand...
Click to collapse
Agree with you.
It wasn't for discredit the work of sparksco. I'm glad to test his work.
Just answer the question for the cross-compiler.
Thanks for the work sparksco
[email protected]_OC said:
Hi men!
There is a discussion on this subject regarding some tests Ezekeel (XDA developper) made with different cross-compiler toolchains and those tests prooved that none of the compilers is better than another.
We heard a lot about linaro because when ICS was released, it was very laggy and linaro and is new compiler version 4.7 made ICS much smoother than before.
But for the pur performance linaro give no improvement if you compare with another one.
EDIT: I found the link of the test for you: http://forum.xda-developers.com/showpost.php?p=19872366&postcount=1
Click to expand...
Click to collapse
That's one test with one toolchain by one developer. As far as I can tell he tested everything with one of linaro's really old toolchaons when they first released 4.6. So by looking at the dates I would guess linaro didn't add much to the toolchain at that point. There's also the GCC version to consider. This is using 4.7 and not 4.6. And lastly there's rom patches that linaro puts out that have nothing to do with the kernel but are used in the ROM building process when building recoveries. It's all debatable. I find this to be a bit smoother and backups seems to be a little faster but maybe it's just me.
[email protected]_OC said:
Hi men!
Thanks for your recovery.
But i experience some strange visual effects like distortion of the image or some lag effects.
Click to expand...
Click to collapse
Your going to have to provide more info than that. Your method of installing, what bootloader you have ect. Thanks.
Edit: flashing zip in recovery causes issues so I removed that method of installing.
sparksco said:
That's one test with one toolchain by one developer. As far as I can tell he tested everything with one of linaro's really old toolchaons when they first released 4.6. So by looking at the dates I would guess linaro didn't add much to the toolchain at that point. There's also the GCC version to consider. This is using 4.7 and not 4.6. And lastly there's rom patches that linaro puts out that have nothing to do with the kernel but are used in the ROM building process when building recoveries. It's all debatable. I find this to be a bit smoother and backups seems to be a little faster but maybe it's just me.
Click to expand...
Click to collapse
thanks.after give a try to this recovery,i agree with you with fact of possible quickest backup time.the general use is good,don't know if it's real an improvement,but i did a backup and at first look it seemed quicker..possible placebo effect,let's see what others say..
why you pulled cwm install version...bugged?
There's a lot of factors to consider. Just the fact that this is using a kernel I built with 4.7 from my own source code could improve things as well. FYI the kernel includes GPU overclocking.
sert00 said:
why you pulled cwm install version...bugged?
Click to expand...
Click to collapse
Read one post above yours...
Sent from my Nexus 7 using Tapatalk 2
sparksco said:
There's a lot of factors to consider. Just the fact that this is using a kernel I built with 4.7 from my own source code could improve things as well. FYI the kernel includes GPU overclocking.
Read one post above yours...
Sent from my Nexus 7 using Tapatalk 2
Click to expand...
Click to collapse
oh thanks,not saw the edit in the post!
New version is up.
Sent from my Nexus 7 using Tapatalk 2
I installed last night using goomanager, so I assume I have the previous version?
Anyway, it works great, so thanks.
stonebear said:
I installed last night using goomanager, so I assume I have the previous version?
Anyway, it works great, so thanks.
Click to expand...
Click to collapse
I am pretty sure you get the Official TWRP version from goomanager not this. Please correct me if I am wrong but I think that is what the Unofficial means.
zedorda said:
I am pretty sure you get the Official TWRP version from goomanager not this. Please correct me if I am wrong but I think that is what the Unofficial means.
Click to expand...
Click to collapse
Yea, I just realised that myself when I saw there were two threads.
What's the difference between this and ClockWorkMod? Is it more stable?
Neo3D said:
What's the difference between this and ClockWorkMod? Is it more stable?
Click to expand...
Click to collapse
Some more features on TWRP I think, especially the queue to flash multiple ZIP files. Flash it and see for yourself
modstorm said:
Some more features on TWRP I think, especially the queue to flash multiple ZIP files. Flash it and see for yourself
Click to expand...
Click to collapse
That, and I like the fact that you have an option to wipe cache/dalvik after flashing something
markj338 said:
That, and I like the fact that you have an option to wipe cache/dalvik after flashing something
Click to expand...
Click to collapse
That and TWRP can also be themed however I never found any themes :/
As many of you might have already heard about it: That using the Linaro toolchain to compile a ROM increases performance. But making stock kernel sources compatible to be compiled with Linaro is no easy task. But, the steps are similar.
When it is compatible to be compiled with GCC 4.3.3, why shouldn't it compile with GCC 4.7?
The problem lies in the code. The transition from GCC 4.3.3 to GCC 4.6, lot of things changed. GCC decided that people should start writing "cleaner code". So, new errors occur when compiling with GCC 4.6+. Example: "unused-but-set-variable". This was not an error in GCC 4.3.3, but GCC 4.7 made it a big issue :crying:
As I said before, GCC 4.3 can compile it, while GCC 4.7 can't. So, I decided to use certain __attribute__'s, and #pragmas, to ignore the errors that one might encounter when trying to compile with GCC 4.7.
Mainly the lines used were:
Code:
__attribute__((__unused__))
and,
Code:
#pragma GCC diagnostic ignored "-Wmaybe-uninitialized"
Steps:
1. Download the GCC 4.7 toolchain using the following command, and place it inside your android-toolchains folder:
Code:
git clone https://android.googlesource.com/platform/prebuilts/gcc/linux-x86/arm/arm-linux-androideabi-4.7
2. Download the kernel source compatible with GCC 4.7 from here
Code:
git clone git://github.com/vineethraj49/pico_kernel_stock.git
Edit the Makefile to correspond to the path of the GCC 4.7 toolchain :fingers-crossed:
Then, its simple,
Code:
make msm7627a_defconfig
and
Code:
make -jX
The code is compatible to be compiled with GCC 4.7 AS IS! You make modifications to code, you should & you WILL run into errors. Usually you can overcome errors by editing the required code, using #pragmas and __attribute__'s. If you ain't ready to work hard, please notify in the thread with the changes that you made to "msm7627a_defconfig", and the possibly a log, with the errors I will add the necessary __attribute__'s, and #pragmas :silly:
As for: Why didn't you make it compatible with Linaro toolchain?
The answer is that @Rishik999 is already working on making it compatible with Linaro, and I don't want to steal his idea
Credits:
Rishik999 for his idea about compiling with Linaro, and expect his kernel sources compatible with Linaro toolchain soon
http://gcc.gnu.org/gcc-4.6/porting_to.html
http://gcc.gnu.org/onlinedocs/gcc/Function-Attributes.html
http://gcc.gnu.org/onlinedocs/gcc-4.6.0/gcc/Diagnostic-Pragmas.html
Google
Some IBM docs, which helped.
http://web.mst.edu/~cpp/common/common_errors.html
Linus and the 2000+ people who contributed to the first Linux kernel.
Source: https://github.com/vineethraj49/pico_kernel_stock
Disclaimer: Everybody can use my work without asking permission, all my stuff comes with a "sharing is caring" xda philosophy licence... just credits would be nice
Reserved
Remember, kernel compilable without errors AS IS!
I will add some steps about how to ignore errors in this post, maybe. So that I don't have to edit a 100+ files again :silly:
Great work
Great work ! :good:
Why posted on general ?
All is well, but i doubt linaro toolchain makes a lot of difference while compiling kernels.
+ i dont think anyone uses stock kernel (2.3.5 android htc sense)
Yes
everyone is on kernel 3
Hpsgill said:
Yes
everyone is on kernel 3
Click to expand...
Click to collapse
im still on gingerbread..
ohh
a gb lover
only few users using gb
everyone is on ics,jb or sense 4
Hpsgill said:
ohh
a gb lover
only few users using gb
everyone is on ics,jb or sense 4
Click to expand...
Click to collapse
yes i love gingerbread IMO the most stable android version for our phone
how i wish cm7 will be complete and also im waiting for the sense 4.1 by cute prince to be stable enough like gb and waiting your sense ex build 6 too
I hope that _that will lead us in this topic because he seems to know away more than I do. I am here to learn and feel free to discuss anything that you like. No restrictions so we can get all the input from other users....
What is a toolchain?
After discussion with a few users, it is a mixed of toolchain types that they use.. According to my research, androideabi is targetting ROM build and optimize for the ROMs' binaries. It is fine when you use it to compile your kernel source but it is not optimized for the kernel compilation.
For kernel compiling, you should use the gnueabi toolchain because it uses the kernel's source for a specific kernel version during the toolcchain compiling for a better compatibility, I guess... However, some users reported that it was fine to use for ROM build also...
So the question is it matter what types of toolchains we are using? What are the benefits between the two? Does anyone see any difference between the two with users' experiences?
Here I will take this spot and fill it with useful info and links about what I have found on the web .... :good:
MythBusters XDA Edition: “Optimized” Compiler Toolchains
USING THE ANDROID TOOLCHAIN AS A STANDALONE COMPILER
ELinuxToolchains
The GNU Toolchain for ARM targets
ARM
lj50036 said:
Here I will take this spot and fill it with useful info and links about what I have found on the web .... :good:
MythBusters XDA Edition: “Optimized” Compiler Toolchains
Click to expand...
Click to collapse
It is great that you are joining the discussion because I have a lot of questions and some good optimizations while I tested with these toolchains. I will give what I know a long the way when the questions come up and hope we will have a better understanding what to use and not to use...
In the olden days, I used the 4.6.2 linaro toolchains and I have heard that a lot of people swear by DoomLord's prebuilts.
Just wanted to throw that out there. I personally have not tried anything above 4.7 yet but now I am tempted to
hardslog said:
In the olden days, I used the 4.6.2 linaro toolchains and I have heard that a lot of people swear by DoomLord's prebuilts.
Just wanted to throw that out there. I personally have not tried anything above 4.7 yet but now I am tempted to
Click to expand...
Click to collapse
Adding to your comment, I do see a performance improvement with different toolchains but some users said it is just a placebo...:crying: I am one of the trials and errors users with testing so nothing is going to stop me until proving by testing and users' experiences, haha...
BTW, I could not get the gcc-4.8/4.9 to work on our tf700 chipset yet because there are some graphical problems on linux kernel v3.1.10. I hope that someone can figure it out so we can test it...
There is a PAC rom in the TF300 forums that claims they are using SaberMod 4.8 without issues. http://forum.xda-developers.com/showthread.php?t=2501869
Furthermore there is a kernel (no longer in development it seems) in the TF300 forums that claims to use linaro 4.8 toolchains http://forum.xda-developers.com/showthread.php?t=2625580
hardslog said:
There is a PAC rom in the TF300 forums that claims they are using SaberMod 4.8 without issues. http://forum.xda-developers.com/showthread.php?t=2501869
Furthermore there is a kernel (no longer in development it seems) in the TF300 forums that claims to use linaro 4.8 toolchains http://forum.xda-developers.com/showthread.php?t=2625580
Click to expand...
Click to collapse
Thanks for the information...:good: I will look more into it when I have more time..
BTW, You should try the linaro toolchain for your kernel compilation but you should use the right kernel version that you intend to run. It is running very smooth... It takes less than 10 minutes to compile and test it out..
Cross Compiler Toolchains [Linaro GCC]
Hi,
Interesting thread but in my humble opinion should be in TF700's development section. So, I just used Christopher83's Toolchain for compiling _that's that10 kernel and flash it in CROMBi-kk RC3. As we have Tegra 3 Soc I used the toolchain with arm-cortex_a9-linux-gnueabi prefix which is optimized for Cortex-A9 cpu with Neon-VFPv3. I tested all the latest versions: 4.9 doesn't work at all (the TF700 was vibrating continuously!), the 4.8 had visual glitches but with 4.7 is working with no problems at all! Finally, from the same thread krislibaeer clarifies a bit the linaro prebuilt toolchains
here a little explanation:
arm-eabi toolchain: is for kernels
arm-linux-androideabi: is for rom building
so you use the arm-eabi toolchain for your kernels and the arm-linux-androideabi for roms
hope it helps a bit
so recommend is the arm-eabi toolchain for kernels
Click to expand...
Click to collapse
Hope that helps the discussion.
Cheers.
sziggins said:
Hi,
Interesting thread but in my humble opinion should be in TF700's development section. So, I just used Christopher83's Toolchain for compiling _that's that10 kernel and flash it in CROMBi-kk RC3. As we have Tegra 3 Soc I used the toolchain with arm-cortex_a9-linux-gnueabi prefix which is optimized for Cortex-A9 cpu with Neon-VFPv3. I tested all the latest versions: 4.9 doesn't work at all (the TF700 was vibrating continuously!), the 4.8 had visual glitches but with 4.7 is working with no problems at all! Finally, from the same thread krislibaeer clarifies a bit the linaro prebuilt toolchains
Hope that helps the discussion.
Cheers.
Click to expand...
Click to collapse
There are a few things that you need to pay attention to.
1. Neon-VFPv3 is for Cortex-a8 and not for a9. You may want to flag it as neon-fp16..
2. I believed that your toolchain is targetting linux kernel version 3.4.x or something but not for version 3.1.10.
3. I have the same issues with my owm builds gcc-4.8/4.9 without any solution.
4. Try some of -Ofast flag to see the improvement on v3.1.10
Good luck....:fingers-crossed:
LetMeKnow said:
There are a few things that you need to pay attention to.
1. Neon-VFPv3 is for Cortex-a8 and not for a9. You may want to flag it as neon-fp16..
2. I believed that your toolchain is targetting linux kernel version 3.4.x or something but not for version 3.1.10.
3. I have the same issues with my owm builds gcc-4.8/4.9 without any solution.
4. Try some of -Ofast flag to see the improvement on v3.1.10
Good luck....:fingers-crossed:
Click to expand...
Click to collapse
Just an FYI
I took the plunge and tried a new toolchain. Ended up trying a 4.9 linaro one for the Grimlock Kernel. Works like a champ on my TF300t. HOWEVER for some reason it will not even boot on a TF700. I'm told it vibrates and the screen goes all white or something. So here is the question:
Why would new toolchains work fine on a TF300 but not on a TF700? One of the transformers' great mysteries :laugh:
hardslog said:
Just an FYI
I took the plunge and tried a new toolchain. Ended up trying a 4.9 linaro one for the Grimlock Kernel. Works like a champ on my TF300t. HOWEVER for some reason it will not even boot on a TF700. I'm told it vibrates and the screen goes all white or something. So here is the question:
Why would new toolchains work fine on a TF300 but not on a TF700? One of the transformers' great mysteries :laugh:
Click to expand...
Click to collapse
Thanks for the information and very good quedtion....:good:
Here is my wild guess because the chipset is using in the tf700t, cortex-a9 t33... I checked the diffs on gcc4.7 and gcc4.9 and tried to match all libraries in hope that I could narrow down the bug but it was failed. There was one time that I succeeded boot into the tf700 with my compiled gcc4.9 and thought that I found the bug but if I rebooted it, it got back to the graphical issue, flicking screen... If I rebooted a few more times then the tf700 was working again. I did all my best to figure out the bug but it was a big failure at the end. That is how far it goes as of today... I don't know enough to solve the mysteries and hope that someone else will....:fingers-crossed:
LetMeKnow said:
Thanks for the information and very good quedtion....:good:
Here is my wild guess because the chipset is using in the tf700t, cortex-a9 t33... I checked the diffs on gcc4.7 and gcc4.9 and tried to match all libraries in hope that I could narrow down the bug but it was failed. There was one time that I succeeded boot into the tf700 with my compiled gcc4.9 and thought that I found the bug but if I rebooted it, it got back to the graphical issue, flicking screen... If I rebooted a few more times then the tf700 was working again. I did all my best to figure out the bug but it was a big failure at the end. That is how far it goes as of today... I don't know enough to solve the mysteries and hope that someone else will....:fingers-crossed:
Click to expand...
Click to collapse
Have you tried to compile a stock TF700 kernel with a 4.8 or 4.9 toolchain? I'm asking because _that kernel and Grimlock kernel actually change the cpu_speedo_id of the TF700 from 5 to 12
For reference check this commit: https://github.com/Hardslog/grimlock_kernel_asus_tegra3_unified/commit/50a19d0f6d6d03e6187a8fa7273be77755d72324#diff-c8f9ec2e1535a394abdd70e576a02ed7R160
I can only go so far with testing as I don't own a TF700........
hardslog said:
Have you tried to compile a stock TF700 kernel with a 4.8 or 4.9 toolchain? I'm asking because _that kernel and Grimlock kernel actually change the cpu_speedo_id of the TF700 from 5 to 12
For reference check this commit: https://github.com/Hardslog/grimlock_kernel_asus_tegra3_unified/commit/50a19d0f6d6d03e6187a8fa7273be77755d72324#diff-c8f9ec2e1535a394abdd70e576a02ed7R160
I can only go so far with testing as I don't own a TF700........
Click to expand...
Click to collapse
No, I have not but it is a good idea to try out. I have a few more days before leaving for two weeks... I will report back before the weekend, thanks again...:highfive:
BTW, have you try some -Ofast flags, not the -Ofast itself? Some of them are working very well with tf700 kernel..
Update: I don't have time to try your recommendation because I am preparing for my business trip. I will give it a test when I am back...