All - following up on a post in AndroidForums which referenced a comment on the root story here. The comment (from Gee):
Correct me if I’m wrong, but I remember hearing that the drivers/kernel are either encrypted, unusable, or simply unavailable. Meaning that while you can root the Galaxy S and put custom ROMs, you still can’t put unofficial ROMs (like Cyanogen) on it without expecting to break a lot of functionality.
That means that you WONT be putting Froyo on it until Samsung officially releases the ROM which can then later be modified. And if Samsung moves away from the Galaxy S after that update, you’ll never see a fully functioning Gingerbread ROM. Ever. That’s a hard pill for me to swallow.
Click to expand...
Click to collapse
Can anyone shed further light on this?
Sounds counter productive. I heard they will be releasing source code. Would be odd if they kept the rest under wraps or took extra measures to prevent access....
Methinks this is definitely worth further investigation. Need some experienced SGS modders to find this thread...
From what I have seen, the drivers are provided in binary form only (which is basically the same as how drivers are provided on Windows). So long as there isn't some code written into the drivers that would prohibit their running on non-official ROMs, there shouldn't be any problem using the drivers in custom ROMs.
I downloaded the entire sourcecode from the samsung website. Its a 150MB zip file and everything is in there in binary form. It must be compiled. I'm certainly not a linux or android dev but it seems like thats what you need. There's even readme.txt files in there to tell you how to build new versions of software.
I'm wondering how long it'll take for someone to put out a custom ROM for the captivate based on 2.2?
I read couples of threads and realize that if a CM mod is wanted, MS2 must have its bootloader unlock.
Since most work is done in /system partition, why do a bootloader modify is needed?
According to the update.zip, we can modify /system partition already, so why do we need bootloader unlock?
CyanogenMod needs its own kernel.
I guess that, since cyanogen is based on gingerbread now,
it runs with a more up-to-date kernel than the milestone's.
The locked bootloader and the efuse thing prevents updating the kernel.
It's possible that someone, using Cyanogen source, makes a backport running on our kernel, but I think it's a huge work...
However, the clever guys from freemymoto.com have created a hack that is capable of launching an other kernel after the legit one starts.
If salvation comes (Cyanogen alleluia ), it will certainly come from this project.
Actually, the hack works on DroidX, and Droid2. So it is 99.9% sure that with some work, it may run on milestone 2 as well.
So, to get Cyanogen work on MM2, we need :
- A very good developer, with android (of course !), linux kernel and kinit knowledge, able to port freemymoto's hack, and Cyanogen
- Hardware drivers for the Milestone (bluetooth, lan, gsm, 3g, touchscreen...)
If we find the golden dev, I personnally think that CyanogenMod could be out for milestone2 sooner than what most people think...
momus87 said:
CyanogenMod needs its own kernel.
I guess that, since cyanogen is based on gingerbread now,
it runs with a more up-to-date kernel than the milestone's.
The locked bootloader and the efuse thing prevents updating the kernel.
It's possible that someone, using Cyanogen source, makes a backport running on our kernel, but I think it's a huge work...
However, the clever guys from freemymoto.com have created a hack that is capable of launching an other kernel after the legit one starts.
If salvation comes (Cyanogen alleluia ), it will certainly come from this project.
Actually, the hack works on DroidX, and Droid2. So it is 99.9% sure that with some work, it may run on milestone 2 as well.
So, to get Cyanogen work on MM2, we need :
- A very good developer, with android (of course !), linux kernel and kinit knowledge, able to port freemymoto's hack, and Cyanogen
- Hardware drivers for the Milestone (bluetooth, lan, gsm, 3g, touchscreen...)
If we find the golden dev, I personnally think that CyanogenMod could be out for milestone2 sooner than what most people think...
Click to expand...
Click to collapse
I LOVE YOUR REPLY!
Well, this is the first reply to me in this forum.Ya...nobody reply me before...so bad..
Milestone has also a locked bootloader, but a CM7 port running on it. That is what makes me want to do something on my MS2.
I'm a C++ programmer, so I learn everything about this these days.
Thank you, really.
And I think I'd read all the freemymoto.com site RIGHT NOW~~
Well, if you think you can be the golden dev, you just made my day...
If you try this, I (and almost everybody here) will at least help providing the hardware drivers.
But the first step is to make the hack work on MM2...
Edit : For the milestone1, the bootloader security has been compromised using a leaked unofficial rom build,
which lead to custom kernels, without needing a hack like freemymoto's.
In fact, freemymoto's hack is based on a milestone1 project, which was abandoned after the rom leak.
It circumvents both bootloader lock and efuse by starting a second kernel after the first one.
(What I couldn't figure out is what happens of the first one... Still in memory? Still using some ressources?)
I read the site yesterday, however sadly, I don't have any experience about writing drivers of cell phone.
So I just sent a e-mail to them to ask if I can provide any help and what to do.
Hope I can help. I'll try my best.
For about the second kernel...I guess I have to read source to know what happened to them...
Anyway, thank you for your reply. At least I get a way to help^^
I didn't mean that we will have to write the drivers from scratch. But we will have to find them.
Hopefully, most of them are open-source. Wifi/bluetooth per example, is handled by a wl1271 chip. Drivers are available on texas instruments website...
You should get in touch with Pizzaroll, who is also trying to make the hack work on milestone2
It's really good news that there maybe a hack bootloader, i like my mm2 but it is really a pity that no much ROM.
liuyanghejerry...
Polish team try to broke bootloader with Boinc (many computers calculate 1 thing), maybe You and Yours friends can join to project?
In China lives maaaaany peoples, this can be speed up calculations
I don't think that distributed computing is the solution...
Even if the boinc project lets us find the signing key, we will have to do the same for the efuse thing.
The freemymoto's hack is a solution to circumvent both of these protections, and doesn't need any computing.
We only need someone to modify some parts of the code to make it work on Milestone2.
Tomszyn said:
liuyanghejerry...
Polish team try to broke bootloader with Boinc (many computers calculate 1 thing), maybe You and Yours friends can join to project?
In China lives maaaaany peoples, this can be speed up calculations
Click to expand...
Click to collapse
Currently I'm studying kexec method now, thanks to PizzaRoll and momus87's help^^
Well, in fact, I like to program and burn my brain. However, the Boinc method seems not really interesting...
Well Boinc is the best solution if you want to use bruteforce.
But even with a whole army of core i7, bruteforcing a 1024bits key is unthinkable actually...
Is this eFuse thing harder to break than the one used in iPhone? If it is then Mr Jobs should adopt a similar approach.
Hello everybody,
I'm an Android app developer for some time, but from some time ago until now I have been tweaking the framework, trying to enhance, customize and maybe correcting bugs, so I'm working on the emulator, but I would like to test it on my device (Atrix).
Since I'm a newbie on this matters and I found plenty of threads and sites/blogs, I got a little confused on what are the steps to be made, specially because I would like to build from source as to have my modifications flashed. I'm aware of the hardware implications for my specific device, so I decided to use a reported working rom source code for my device, but i'm a little lost. I thoutgh to be using Cyanogem but I don't know the specifics.
Since I already have the source code AOSP, I think It would be painful to make it work on my device (drivers, kernel, so on...).
I just want (for now) see my framework modifications on device. Anyone could point me to the right direction?
I thought of using this firmware here.
Thank you all.
I was wondering how hard it would be to alter the patches in the update.zip found here to work with the leon. I know the two phones are almost identical as seen here. as of right now there is only an update for the sprint/boost/virgin g-stylo's but there is supposed to be one coming for the t-mobile/mpcs one soon. i've taken the source and started tinkering with it. I've compiled the kernel changing TARGET_PRODUCT=g4stylusds_global_com to TARGET_PRODUCT=c50_mpcs_us but leaving the config as g4stylusds_global_com_defconfig and get far enough to get the zImage, but not the ko files. I have not yet compiled the rom, and it was about this point i started thinking about the update.zip DominicGresham was kind enough to provide us and maybe it would be easier than editing the source. any input would be appreciated. I like 6.x and would like to have it on my daily driver, and help everyone else that wants it for our device. i'm going to keep tinkering away until i get something besides a bootloop, and will attach the zimage i got after this test compile again on the kernel with the mpcs gstylo defconfig. i guess i should go ahead and try to compile the rom also and see how that goes. https://www.dropbox.com/sh/04pab4lbxncd7rr/AAA_GesYl74E35ikF4weaw1ca?dl=0
natediggler said:
I was wondering how hard it would be to alter the patches in the update.zip found here to work with the leon. I know the two phones are almost identical as seen here. as of right now there is only an update for the sprint/boost/virgin g-stylo's but there is supposed to be one coming for the t-mobile/mpcs one soon. i've taken the source and started tinkering with it. I've compiled the kernel changing TARGET_PRODUCT=g4stylusds_global_com to TARGET_PRODUCT=c50_mpcs_us but leaving the config as g4stylusds_global_com_defconfig and get far enough to get the zImage, but not the ko files. I have not yet compiled the rom, and it was about this point i started thinking about the update.zip DominicGresham was kind enough to provide us and maybe it would be easier than editing the source. any input would be appreciated. I like 6.x and would like to have it on my daily driver, and help everyone else that wants it for our device. i'm going to keep tinkering away until i get something besides a bootloop, and will attach the zimage i got after this test compile again on the kernel with the mpcs gstylo defconfig. i guess i should go ahead and try to compile the rom also and see how that goes. https://www.dropbox.com/sh/04pab4lbxncd7rr/AAA_GesYl74E35ikF4weaw1ca?dl=0
Click to expand...
Click to collapse
Hey are you still working on this rom? i would like to know because i can try to help in any way if needed
Hello XDA community,
After months of struggling im in desperate need of your help.
What is my goal?
To compile our own kernel so we can mess around with Qualcomm's trustzone.
Where did it all start?
It all started a few months ago back in November 2021 when me and a friend read a blog about a Trustzone exploit in Qualcomm's MSM8974 processor. (link to the blog)
Big noobs as we where, we where getting hyped an bought ourselves a Samsung Galaxy Note 3 with a MSM8974 processor. After messing around with the device we learned more and more about android and the workings of the software. We first started to root the phone and play around with ADB. After that we found that we can use Frida aswell. But all of it didn't lead to the result we wanted. So we started to dig deeper in to the write-up of the blog. We soon came to the conclusion that we where not even close to doing the right thing.
What happend next?
After we came to the conclusion that ADB and Frida is not enough we understood we need to build our own kernel for this exploit to work. So we do what we always do and that was start googling and watch YouTube. By doing that we found 2 different kernels:
Lineage 18.1
SM-N9005_EUR_LL_Opensource
We where over the moon because inside the folders we found the the files that need to be replaced for the exploit to work. So we replaced the files and try to compile the kernel by using this YouTube tutorial. Everything seemed to go well, we made the
Code:
make msm8974_sec_defconfig
and we run
Code:
make menuconfig
so far so good. the last and final step should be to run
Code:
make
to compile the kernel. And that is where it went wrong an ran in to a error.
Then what?
Then we started to compare the original files with the modified files and we found out that it wasn't just a few modifications in the code but the code was different. So we did what we always do, go back to the source. So back at the blog we thought maybe its because they modified a Nexus 5 at android 4.4.4.
Now what?
Now im trying to find the source of a android 4.4.4 kernel of msm8974 so we can compare those files to the modified files. So if anyone has useful info or knows where to find a kernel for android 4.4.4 for a Samsung Galaxy Note 3 (SM-N9005).
TLDR;
Im looking for a source of android 4.4.4. for a Samsung Galaxy Note 3 (SM-N9005) (hlte)
Well, unfortunately today almost no one uses android 4.4.4. even that have the exact same device, I am using Android 11 on my phone.
I recommend not wasting your time on a kernel for Android 4.4.4. And if you really want to make a kernel for this device I need one for Android 10 and above.
Also I think making a kernel for AOSPs like LineageOS or HavocOS or crDroid ETC is much easier.
A.M.2.K.M said:
Well, unfortunately today almost no one uses android 4.4.4. even that have the exact same device, I am using Android 11 on my phone.
I recommend not wasting your time on a kernel for Android 4.4.4. And if you really want to make a kernel for this device I need one for Android 10 and above.
Also I think making a kernel for AOSPs like LineageOS or HavocOS or crDroid ETC is much easier.
Click to expand...
Click to collapse
Thanks for your response I will look in to it. not sure if it will be easier because the files that needed to be changed are:
arch/arm/mach-msm/scm.c
arch/arm/mach-msm/include/mach/scm.h
drivers/misc/qseecom.c
SwAgLaNdeR said:
Thanks for your response I will look in to it. not sure if it will be easier because the files that needed to be changed are:
arch/arm/mach-msm/scm.c
arch/arm/mach-msm/include/mach/scm.h
drivers/misc/qseecom.c
Click to expand...
Click to collapse
Well, I don't know much about kernels.
I said it's easier because I thought it's a pure and raw operating system. But now according to your post it turns out it's not easier.
Plus almost no developers left for this now very old model .