Hi,
I'm a dev working on Mozilla's Boot2Gecko project. We've run into an issue with heimdall, and I'm trying to get ahold of the dev to discuss this further.
One of my coworkers created an issue on github, and I've added some comments, but neither one of us has received any replies, so I'm posting here to see if that helps.
The problem that we've run into, is that I can use heimdall fine when our system.img file is 100Mb or less, but get failures or corruption when system.img is 100Mb or larger.
My phone is a Samsung GT-I9100 (Samsung Galaxy S II). I upgraded the firmware to ICS using the XXLPQ firmware (I used this page: www dot theandroidsoul dot com slash xxlpq (the form won't let me post the url since I've posted less than 8 times), so its possible that the problem is related to newer bootloaders.
I have a system.img file that is 115Mb, and the command line I use is:
heimdall flash --factoryfs system.img
Under 1.3.1, there is a noticable pause at 87%, then it continues on to 100% and reports success, but the image that is burned into flash is corrupted.
Under 1.3.2, it stops at 87% and reports a failure.
Running with --verbose, the 87% corresponds to 800 packets of 128K each (which is 100 Mb)
If I reduce system.img to under 100 Mb, then both 1.3.1 and 1.3.2 work fine.
My dev machine is running Ubuntu 12.04, and I have a Windows 7 VirtualBox. I see the same behaviour from the Windows 7 VM as I see from ubuntu.
If I use odin (v1.83) from the Windows 7 VM, then it burns it properly.
I took a quick look at the source code and tried increasing the 800 limit (kMaxSequenceLength) and the 128K limit (kDefaultPacketSize), but those both fail as well.
I have packet wireshark captures of the USB streams under linux, and I installed a trial version of USBlyser under windows and have packet captures from both odin and heimdall, should those prove useful.
From the packet captures, odin seems to be transferring 1Mb at a time.
I had the same problems with my ET4G. I posted a pull request on the project's github. I'd link you there but I am not allowed since I am a new user, it is pull request #47. Works for me... maybe it will help you out.
bleffij said:
I had the same problems with my ET4G. I posted a pull request on the project's github. I'd link you there but I am not allowed since I am a new user, it is pull request #47. Works for me... maybe it will help you out.
Click to expand...
Click to collapse
this is certainly not fixed for me for an SHW-M250K (korean SGS II and compatible with I9100 ROMS with minor caveats).
I get extremely similar symptoms(identical as far as I can tell): Consistent corruption where files don't diff against the loopback mounted ROM files and are reprodicibly corrupt for any particular ROM. I saw this on.. I think it was 1.3.2 and then grabbed the latest 1.4RC1 from git and still see the same thing. Flashes with odin on the same hardware in fact in a windows vm within the same host OS.. work just fine. This is really bad especially since, sometimes the corruption isn't that obvious.
It seems to be scattered bits or bytes. Often the system boots but encounters various amounts of bugginess (depending on the EXACT rom, just adding or removing a file from the rom not surprisingly changes which bits and bytes are corrup)t. I guess it's many small chunks that are corrupt because I see typically maybe 80% of files are corrupt but still things kind of work. That tells me it's like a bit byte or small chunk here or there for every few MB probably. Enough to get most files but little enough it doesn't get every file or break them entirely.
I don't have a github account. Maybe I need to make one.
Related
As I have posted before, I have a script that uses wget to regularly pull a gif file off a server to check its status. This script worked when I was running Eugene's Vibrant9 ROM, but stopped when I installed Bionix Fusion 1.2. I am now running Nero v3, and it still does not work.
I initially suspected that the problem was with the version of wget, and that guided my investigations. But I have tried several wgets with the same results. To demonstrate, I've posted a test gif here:
http://216.97.237.135/tmp/test.gif
The gif is 544 bytes, but if I wget it from my phone over 3g I only get 308 bytes.
Further inspection has revealed that if I wget it from my phone over wifi I get all 544 bytes. In addition, if I tether my laptop to the phone using Mobile AP and wget from the laptop, I get 308 bytes. The problem is thus isolated to my 3g connection.
Given that it worked under a prior ROM, I now suspect that the problem is either the modem or the modem settings. Can someone suggest where I can find model settings that I might play with, or what different modems I might try installing (if that is even possible)?
I'm afraid that I find this issue fascinating yet frustrating, and I'm am running out of avenues to pursue. If I don't get any interest here, could someone suggest a different forum where the question might be of more interest? I'm posting here because I read this forum and there are a lot of smart people here, but this is a pretty narrow question....
There's been a few ROMs based on the stock kernel--one of them is mine--but nothing more in-depth. I'm setting out to change that, but I could use some help from more experienced folks.
What I've found so far:
- It uses the MSM7227_Surf Qualcomm System-on-chip (verified by disassembling a broken phone)
- which also has the Adreno200 GPU (built into MSM7227_Surf)
- It has 512MB flash partitioned very similarly to the Samsung Gio, with most partitions being at least the same if not sightly larger (with the exception of /system and /data being split differently such that /system is bigger than on the Gio and /data is smaller)
- Unlike the Ace/Gio, it uses a Broadcomm wifi chipset instead of the atheros at6000
What I'm stuck on at the moment is the partition map. How do I get the precise partition sizes when /proc/mtd doesn't exist?
I was able to figure it out. Since the partitions are nearly identical to the Samsung Galaxy Ace, I took a look at that phone's BoardConfig.mk and figured out the formula:
1) Get the partition list from 'cat /proc/partitions'
2) Multiply the reported # of blocks by 1024 to get the size in bytes
3) Convert the result from #2 to hexadecimal
Results:
BOARD_BOOTIMAGE_PARTITION_SIZE := 0x00A00000
BOARD_RECOVERYIMAGE_PARTITION_SIZE := 0x00A00000
BOARD_SYSTEMIMAGE_PARTITION_SIZE := 0x0DD00000
BOARD_USERDATAIMAGE_PARTITION_SIZE := 0x09800000
BOARD_FLASH_BLOCK_SIZE := 4096
EDIT: I just found where this phone puts the equivalent of /proc/mtd. It's in /proc/LinuStoreIII/bmlinfo:
Code:
FSR VERSION: FSR_1.2.1p1_b139_RTM
minor position size units id
1: 0x00000000-0x00180000 0x00180000 6 1
2: 0x00180000-0x00200000 0x00080000 2 2
3: 0x00200000-0x002c0000 0x000c0000 3 3
4: 0x002c0000-0x01bc0000 0x01900000 100 4
5: 0x01bc0000-0x024c0000 0x00900000 36 23
6: 0x024c0000-0x029c0000 0x00500000 20 25
7: 0x029c0000-0x02bc0000 0x00200000 8 5
8: 0x02bc0000-0x035c0000 0x00a00000 40 6
9: 0x035c0000-0x03fc0000 0x00a00000 40 7
10: 0x03fc0000-0x04080000 0x000c0000 3 8
11: 0x04080000-0x05980000 0x01900000 100 9
12: 0x05980000-0x13e00000 0x0e480000 914 21
13: 0x13e00000-0x1dc00000 0x09e00000 632 22
14: 0x1dc00000-0x1f500000 0x01900000 100 24
I just signed up a few minutes ago with this account to say thanks for doing this, from another guy who has just gotten this phone! (Well the Canadian version, same thing basically called the "Samsung Galaxy Q"/SGH-T589R) Not sure if I can do much, the closest thing I do compared to this is using Linux. I'll be rooting for you.
My initial goal was to get a Gingerbread kernel that a) supports the Galaxy Ace (the Gravity SMART's more popular cousin) and b) actually compiles, but I didn't have much luck. The kernel I wanted to use was built against some long-deprecated kernel source that I couldn't dig up--and applying the diff against the Samsung stock source tree failed to compile.
So, I've gotten a copy of the source to the ICS kernel developed by Maclaw & others [still too newb to post links] for the Samsung Gio and Samsung Ace.
Next step: finding the stuff unique to the Gravity SMART.
Fun fact: The Gravity SMART was originally code named the Gravity Touch 2, and this is reflected in the source code where the model's config flag is named CONFIG_MACH_GT2 (as opposed to CONFIG_MACH_COOPER for the Ace).
My initial suspicions are proving to be pretty spot-on. Quite a few of the changes involve simply adding "|| defined(CONFIG_MACH_GT2)" into an existing chain that typically includes the CONFIG_MACH_COOPER or CONFIG_MACH_GIO. However, there are a number of notable exceptions. Thanks to the power of a recursive grep, I've got a list of files and I'm slowly but surely getting the changes ported over.
I have a ways to go before I get a compiling kernel, but there's still one big hurdle to jump after I get all the changes brought in: Wifi.
The stock kernel uses a kernel module, dhd.ko, which appears to be a Broadcom driver. However, I suspect it may be customized/proprietary because the kernel config flag (CONFIG_BROADCOM_WIFI) isn't used anywhere in the source tree. Both the ICS and FroYo kernel sources have code for -other- broadcom drivers, but there may be compatibility issues. I also found a Broadcom driver on googlesource which I suspect was Samsung's starting point for the aforementioned kernel module, but it doesn't come with any instructions for compiling it so I'm not quite sure what to do with it. I'll probably start by just trying to use the kernel's built-in driver and poach the firmware file from the googlesource package, and see what happens.
OK, so a status update--
Using the list of files containing the CONFIG_MACH_GT2 tag as a guide, I went through and modified/ported the code as best seemed to make sense from the Samsung code into the ICS kernel.
Then began the fun of trying to compile the kernel.
The configuration routine wanted to eat my own CONFIG_MACH_GT2 flag, so I had to do a little more hacking to figure out how to modify the kernel configuration files to recognize it as a valid configuration option. For the benefit of others, I'll post the results of my research:
1) Edit arch/arm/tools/mach-types to add the machine type
2) Edit arch/arm/mach-msm/Kconfig to add the configuration option
3) run 'make distclean' to make sure that all the generated stuff gets regenerated properly.
Now it's recognizing the CONFIG_MACH_GT2 option, but one of the modules is failing to build. But, I'm too tired to tackle it right now so it will wait until another day.
It compiles!
... which of course doesn't mean it actually works.. I haven't had a chance to test it yet. Fortunately I have the ODIN files to restore ifwhen something goes horribly wrong.
gblues said:
It compiles!
... which of course doesn't mean it actually works.. I haven't had a chance to test it yet. Fortunately I have the ODIN files to restore ifwhen something goes horribly wrong.
Click to expand...
Click to collapse
Hoorah! Here's hoping it works!
Well, no dice. Stuck at "SAMSUNG" which, admittedly, is better than previous attempts at flashing a custom boot/recovery image (just got a blank screen on those attempts..)
In the process, though, I learned a few important things:
1) To flash any custom firmware, you need Odin 4.40 USA. Newer versions do not work.
2) How you make the .tar.md5 file matters! The files have to be added to the archive in the right order. I determined the order by extracting the stock firmware I acquired from samfirmware.com (using 'tar xvf') and then made sure that I added them in the same order when I created my version. If the files aren't in the correct order, Odin will crash when attempting to flash the phone.
Thank you!
I have been waiting for so long for someone to start something with this phone. I will help in any way I can, if you would like. I am a bit of a newb, but I have done a few things with a Dell streak 7 and an Optimus One, as well as interop unlocking my LG Quantum (a windows phone, but w/e )
I'm another step closer!
After the spectacular failure of my first attempts, I decided to try a smaller step--unpacking the stock boot image, making a useless change (putting "Hello, World!" at the top of /init.rc), packing it back up, and flashing it. When this also failed, I thought perhaps I was working against a locked bootloader.
After finding absolutely no information about the presence of a locked bootloader, and with similar phones not needing any special unlocking, I went a little more basic: what if I unpacked and repacked the boot.img without any changes? I did this, and I was surprised to find that the repacked file didn't match the original at all.
I did a little more research and playing around with mkbootimg, and successfully disassembled and re-assembled the original boot.img file (verified by comparing md5 checksums). I used the same settings to retry my "Hello World" test--and lo and behold, it WORKS!
Nicely done, I knew you could do it!
Awesome job! I can't wait to see how this works out!
Well, unfortunately, it hasn't. While I've been able to modify the contents of the ramdisk cpio, my kernels aren't booting. Or if they are, they're failing before anything useful happens.
Anyway, the trick to creating the boot.img and recovery.img:
The images are in the standard Android format, as documented here. In fact, the unpack-bootimg.pl from that page works brilliantly. However, to re-pack the image, it's best to use mkbootimg directly.
Let's say you've got your gzipped kernel named 'kernel.gz' and the gzipped cpio archive of your ramdisk named 'ramdisk.cpio.gz', and let's say you want your resulting image file to be named 'yourboot.img'. Then you'd use the following command:
Code:
mkbootimg --kernel kernel.gz --ramdisk ramdisk.cpio.gz --base 0x13600000 --pagesize 4096 -o yourboot.img
From further research, it appears that in order to boot a newer kernel than what shipped with it, the boot loaders need to be modified to accommodate it. The reason is that the access to the internal memory partitions is handled by a proprietary Samsung kernel module, which only exists for the kernel the phone shipped with. Even if I had the source for the Samsung module, the boot loader needs to support the newer version of the Samsung module.
So, what I'm going to try next is hacking the kernel to force it to report the same version as the phone expects, and see how spectacularly it fails well it works.
Well, fudging the kernel version in my 2.6.35.7 build didn't work, although just doing a straight compile of Samsung's kernel sources _did_ give me a booting kernel.
Well, the result of my extensive research is thus: there's no hope of getting Gingerbread on this phone without Samsung creating an official release. Here's why, from what I've gathered:
The bootloader does not appear to be locked; however, it does provide some key services that Samsung's proprietary kernel modules use to provide access to the flash partitions on the phone, kind of like the BIOS in your PC. The bootloader expects a specific kernel version (more specifically--and I apologize if the terminology is wrong--a specific kernel ABI). Simply faking the kernel version is not enough--it has to have exactly the same driver interface and kernel entrypoints as it expects. So even though I tried compiling the 2.6.35 kernel with a fake version number, the ABI is still wrong so the phone doesn't boot.
Hacking the boot loader is waaay outside of my skillset, although I did successfully dump it from my wife's busted phone. Judging from the results of my research, it's not something anyone's actually attempted. And given that screwing up the bootloader will hard-brick the phone, I'm not in a mood to experiment.
So basically, since Samsung is the only one with the code for the bootloader and the RFS drivers, it's up to them to create a GB release. Once they actually do so, then possibilities open up quite a bit.
Well you gave it your best shot so thank you! Maybe someone on here will come up with a way to get a non-Froyo version. Hell or even maybe if we're nice enough we'll get upcoming Jumping Jellybean available for this phone.
Thanks so much even for trying. I wish I could try and do something, but I would have no hope.
My sister's likely getting a Galaxy Q tomorrow, so I'll be following dev work on this device a bit.
Skip the RFS format. It's a pain in the neck to work with, and pretty much every benchmark out there shows it's not as fast as ext4 anyway.
Edit: Got another idea: try using the GB RFS kernel modules from the Gio/Ace/Fit/Mini...
If you want to try converting to ext4, make sure you have RFS format dumps of the /data and /cache partitions you can dd back into place in order to restore to stock.
Consider contacting Phiexz (who's sort of left XDA, but should be reachable at his own forum), Delanoister, Maclaw and the rest of the GalaxyICS team. Hyriand too. What they've done for the Gio is nothing short of outstanding.
Good luck,
Darkshado
gblues said:
my kernels aren't booting. Or if they are, they're failing before anything useful happens.
Click to expand...
Click to collapse
No ddms/logcat output, nothing?
For now simply follow the instructions and use the provided file download from the link posted below
http://forum.xda-developers.com/showpost.php?p=48392009&postcount=1
I have personaly tested this method in full on My AT&T Samsung Galaxy S4 ZOOM SM-C105AUCUAMJ2 (4.2.2 Jelly Bean)
although it should be safe to use on any AT&T Samsung Galaxy S4 ZOOM running an android version prior to Kit-Kat
All Credit for this, "exellent completely safe", Root method goes to its creator k1mu
And please Know that it is only being shared with you via his explicit permission.
EDIT:
Q&A for [ROOT] Saferoot: Root for AT&T Galaxy S4 ZOOM SM-C105AUCUAMJ2_On 4.2.2 JB
Please if you are not on an AT&T Galaxy S4 ZOOM C-150a, Feel Free to Read
but Do not seek help,instruction, or any other type of advice by posting in this thread as it is hard enough to find information about a specific device when the said device does not have a dedicated forum. Thank You for your Understanding in this.
Instructions & Advice
If ?'s arise
and time allows I will add device specific instruction as well as any advice, Please bear with me. Same as above Q&A Link
Will this same method also work on the SM-C105AUCUAMJ2?
kbracing6 said:
Will this same method also work on the SM-C105AUCUAMJ2?
Click to expand...
Click to collapse
Most Definitely as this was the reason for my post
Hey @Phatboyj420, as far as converting a raw OTA file to Odin files, here's what I can tell you. The OTA file should be filename.cfg, i.e. a cfg file. Believe it or not, this can be manipulated like an archive (like zip or rar or tar) so open that cfg file in 7zip, WinRar, or whatever. You're going to want to extract the files like (I can't guarantee this a full list, just the ones I can remember) modem.bin, NON-HLOS.bin, boot.img, recovery.img, aboot.mbn, rpm.mbn, sbl1.mbn, sbl2.mbn, sbl3.mbn (you might not have all 3, idk), and tz.mbn. Again, there may be others you need, and/or not all the ones I just listed might be necessary, they're just the ones I remember dealing with before. Oh, you're going to need to make sure the system, cache, and persdata partitions are all packed into .img.ext4 files, idk if they're like that in the OTA cfg archive.
Here's where it gets even hazier for me. So, I know from personal experience that you can take any of these files (I'll just use recovery.img for this example). In a linux terminal (I use Ubuntu for convenience's sake, my laptop has windows and ubuntu partitions) run this command:
Code:
tar -H ustar -c recovery.img > recovery.tar
So that's how you can take any one of those files I mentioned last paragraph and pack it into an Odin-flashable tar. I'm still a n00b when it comes to the linux terminal, so I can't really say the syntax for packing more than one at once.
So I guess I haven't really given you the complete process, but hopefully this is at least a decent starting point. Honestly I consider myself a hobbyist at best, not an expert. muniz_ri, who Devo7v mentioned earlier, did all the heavy work in this regard so he can probably help you much more, I just didn't want to volunteer him for the job
thisisapoorusernamechoice said:
Hey @Phatboyj420, as far as converting a raw OTA file to Odin files, here's what I can tell you. The OTA file should be filename.cfg, i.e. a cfg file. Believe it or not, this can be manipulated like an archive (like zip or rar or tar) so open that cfg file in 7zip, WinRar, or whatever. You're going to want to extract the files like (I can't guarantee this a full list, just the ones I can remember) modem.bin, NON-HLOS.bin, boot.img, recovery.img, aboot.mbn, rpm.mbn, sbl1.mbn, sbl2.mbn, sbl3.mbn (you might not have all 3, idk), and tz.mbn. Again, there may be others you need, and/or not all the ones I just listed might be necessary, they're just the ones I remember dealing with before. Oh, you're going to need to make sure the system, cache, and persdata partitions are all packed into .img.ext4 files, idk if they're like that in the OTA cfg archive.
Here's where it gets even hazier for me. So, I know from personal experience that you can take any of these files (I'll just use recovery.img for this example). In a linux terminal (I use Ubuntu for convenience's sake, my laptop has windows and ubuntu partitions) run this command:
Code:
tar -H ustar -c recovery.img > recovery.tar
So that's how you can take any one of those files I mentioned last paragraph and pack it into an Odin-flashable tar. I'm still a n00b when it comes to the linux terminal, so I can't really say the syntax for packing more than one at once.
So I guess I haven't really given you the complete process, but hopefully this is at least a decent starting point. Honestly I consider myself a hobbyist at best, not an expert. muniz_ri, who Devo7v mentioned earlier, did all the heavy work in this regard so he can probably help you much more, I just didn't want to volunteer him for the job
Click to expand...
Click to collapse
Thanks My guy,
This definately gives me a good jumping off point.
The phone shipped with JB-4.2.2 and I think there has been an AT&T OTA for KK so it would have to be a complete Firmware so thats good news.
I'm working on several projects at once so I don't know exactly when I'll get to this but when I do the first thing will be to verify whether the .img's for data/system/cache are img.ext4 or not and if not how to produce correctly.
Also I don't have a Linux Dev setup yet I'm running Windows On a 2011 Dell XPS-17_L702x with an intel 2720 quad core. So from your post I take it Ubuntu would be your suggestion for Linux Distro. and should I make a separate Linux Partition and run from it, or run from a VM-Box inside of windows? Which do you think would be most efficient?
Now that I think of it I'm going to start a dedicated thread for development discussion for this specific device. If I start the thread in the Development section for the general S$-ZOOM but title it specifically for the " AT&T-C105a_ZOOM " do you think the mods would want to move it to the general Q&A section?
" I would hope that, as it would pertain directly to the development of the specific device they would see fit to leave it in the development section.
Sorry for the randomness of my thoughts but I will link you to the thread when I get around to starting on it.
For now I'm going to get back to work on the S4_Active for My daughters B-day I intend to have it Rooted and rommed out for her. When I get that done I'll make my way back to this Project, and start the dedicated thread.
Thanks again,
" I look forward to future collaboration as it seem we share a similar Hobie at the least ",
Phatboyj
Unfortunately for the few lonely SM-C105a users still hanging on out there this does not appear to be a completely effective root. I get this in dmesg: <3>[ 2130.920856] c0 Restricted changing UID. PID = 11580(su) PPID = 11575(sh)
capt_planit said:
Unfortunately for the few lonely SM-C105a users still hanging on out there this does not appear to be a completely effective root. I get this in dmesg: <3>[ 2130.920856] c0 Restricted changing UID. PID = 11580(su) PPID = 11575(sh)
Click to expand...
Click to collapse
What is your build # because this is a fully effective root method for the ATT SM-C105a running build# SM-C105AUCUAMJ2
Please elaborate on where your dmesg is coming from.
Is it from running Saferoot to obtain Root?
Or is it from some other function you are attempting after obtaining root?
@Phatboyj420 Great to see you here--when I'm sure your S4 Zoom is long gone. Yeah, that's my build and I should point out that I get that in dmesg only for certain operations in the system folder (trying to copy modules I think). No biggie, for day to day this method works fine. I'm more interested in whatever happened with that unbrick image you were trying to make from dd. Did you ever test it? Was it effective? My current efforts are focused on developing a reliable unbrick method for our phones. Rather than the complete image I think we may need to extract and flash the original bootloader in many cases. Can you provide any feedback on this?
capt_planit said:
@Phatboyj420 Great to see you here--when I'm sure your S4 Zoom is long gone. Yeah, that's my build and I should point out that I get that in dmesg only for certain operations in the system folder (trying to copy modules I think). No biggie, for day to day this method works fine. I'm more interested in whatever happened with that unbrick image you were trying to make from dd. Did you ever test it? Was it effective? My current efforts are focused on developing a reliable unbrick method for our phones. Rather than the complete image I think we may need to extract and flash the original bootloader in many cases. Can you provide any feedback on this?
Click to expand...
Click to collapse
@ capt_planit
while I have moved on from using my [email protected] S4-Zoom, I do still have it and the dd dumps of it the problem with providing said dumps publicly is that some "idiot would inevitably flash the complete dd.image to there phone giving them an exact clone of my phone explicitly the IMEI via EFS parition are my concerns.
That being said if I know that the dd.images were to be used in a proper manner, by say a knowledgeable Dev. I would gladly supply them to further /Kickoff Development,
It sounds as if we are of like mind but I found myself at a stand still and did not recieve any response from the plea in my Sig.
But rest assured if there is something I can do to assist I will.
as far as an unbrick.img for the Zoom c105a
I did make one but have had no reason to use it to know if it works or if the SD unbrick method even works for the Zoom like it does on the Galaxy S3 that the method was originally discovered on.
Even if it does work the SD-Unbrick method only gets you to a state where Download
mode works SO unless I am mistaken and you can access adb through Download Mode witch I'm pretty sure you cant, we would still be at a stand still as there is no stock firmware publicly available for our device " Ludacris I know " but true none the less.
Edit:
1 thing we need is the OTA update from JB to KitKat available for our device we need some one to extract it before updating then and only then might we be able to create an install-able firmware for our device.
if I'm not mistaken after you download the OTA update you should be able to find it in /cache/fota just zip the entire fota folder move it to your sd-card and make it available to me and we'll make it happen from there.
If you don't want to except the update just delete it at this point and reboot.
...
...
I think this user @awwar describes an inability to access download mode>here. In anycase, I think that thread would be a great place to post your unbrick. image. I think your dd image seems too small. But if you still have it, that is what @moomoo was asking for when he started that thread. I can't provide mine, except as separate img files. My understanding is that flashing an efs image won't provide a real IMEI. I think flashing some combination of boot, system, cache (I'm surprised this would be necessary) and recovery should get the operator some kind of system. But so far it doesn't seem to work that way. Your help getting some working phone flash would be greatly appreciated. I believe, BTW that the OTA is dead...
If this is the wrong forum please let me know and I'll post it in the appropriate one
Lately me and a few others from Mainline4Lumia discord server have been trying to get little kernel to work properly on my Nokia Lumia 925 (windows phone) however it seems that after it copies to memory it just seem to fail with no log aside from what's displayed on the screen (via imbusho's bootshim), the Android4Lumia generic little kernel image (flashed as UEFI) seems to work up until it goes to read GPT on emmc where it entirely stops, tried disabling display however no signs of life/booting after flashing non A4L image.
the phone uses msms8960 (snapdragon s4 plus) with 1gb of ram and 16gb of storage . i was hoping to get some insight why it may stop after copying to memory, i do have UART on the phone if that helps. again please let me know if this is the right spot to post as i havent used xda much aside from reading and fixing my lumias
current resources being used
Little kernel for msms8960 target: https://github.com/feherneoh/bootloader_msm8227
Bootshim: https://github.com/imbushuo/boot-shim
don't entirely know source of A4L little kernel image
Lk2nd https://github.com/msm8916-mainline/lk2nd
Microsoft recovery tool for thor2: https://support.microsoft.com/en-us...tool-faq-2b186f06-7178-ed11-4cb6-5ed437f0855b
windows phone internals: https://github.com/ReneLergner/WPinternals
Hello,
So I need help with completely wiping a linux partition on my android SM-a137f (and computer) that has been there for privacy invasion reason/spying etc by someone at my work.
------------Skip to the bottom part for phone only part, the rest is about the computer issue , both related however as they infect each other and i see the same linux filesystem and patterns on both---------
I never used Linux before and I tried fighting this "trojan" in windows on my computerand fatigued myself after the 50th low level formatting, not understanding how it can just come back, plus self learning, if i use an antivirus that works decent, like install it within 2 minutes after formatting before it activates, it gave me a lifespan of +20 minutes perhaps, next time i did that, it had injected it and just spawned like 10 processors of it to lag me down. Plus installing these fake drivers, meaning it infected my kernel.
So as i dived in to the world of Linux i got my hands on rescuezilla, which is based on ubuntu, things started to clear up, very sophisiticated work, and insane. Looping stuff, a squash filesystems that is impossible to remove, and i mean impossible in my book, thats why i come here, because if you are going to suggest any permission changing command or unmount command, don't - i tried every single one in the book and it won't budge.
It's not even an I (i forgot what I stood for) in the permission, it's only like cr- r ---- r-- or something, two or three 3's
The overlay starts with a /cow and holds everything.
Since this will perhaps be a police matter unless he agrees to replace the devices this has killed (phone, im on my third now, computer and a very secure router) and I have more than enough evidence including a perfect clone of the whole harddrive of my work computer and laptop.
Anyway - i dont care about any data, wipe all, and more if possible. Its my 85th time formatting anyway so, i just can't get rid of it. It's "base" is the X: partition which would be the recovery one I believe, in windows. And it mounts itself as a ddrom and/or ram, virtual, but not virtual. Every program i've said has stated it's a virtual, and when i try anti-virtual removal stuff, it says its not a virtual disk. Read only that nothing can remove in windows at least, forget diskpart etc.
-----------------------------------------------------------Phone---------------------------------
I am willing to pay for proper help on this.
I just rooted it and I want to wipe that whole kernel on this phone. Not a factory reset - a clean wipe. Not a single file left. Is that possible? Or how else do I get rid of all of that stuff, I mean, there is a problem whe 280 system apps has on average 60 permissions each, up to 400.
Thanks a lot for any input in advance. I can't stress how much of a mental toll this has taken on me as i've been on this for 1 month straight not doing anything else. Since it's about my privacy, and It seems I've had one for the past 2 years with my every footstep, thing ive written or said, even a sleepshedule has been monitored.
Just a few screeshot examples
In Odin flash the firmware with a 4-file zip file.
ze7zez said:
In Odin flash the firmware with a 4-file zip file.
Click to expand...
Click to collapse
How do you mean? I done that several times, filled out the AP CP BL CSC etc, doesn't touch that partition
Mikez77 said:
How do you mean? I done that several times, filled out the AP CP BL CSC etc, doesn't touch that partition
Click to expand...
Click to collapse
Show a screenshot of the Odin window after the flashing is complete.
According to the OP, their computer is attacked within seconds of them logging in. Any antivirus only lasts a few minutes. Can't low level format the disk, because the attacker puts BIOS and virtual wifi spots into the UEFI partition, hacked their router and cell phones numerous times as well.
And they claim to know who this is, have a mountain of evidence but somehow won't go to the police with it. Search for the op's name and the keyword linux. They're carpet-bombing forums with paranoia.