Boot images for bootloop

[rchtk]

My phone is 10 months old and began bootlooping a few hours ago. I've kept it in a bootloop but it never booted neither system nor recovery.

It's still under warranty but I'll keep it a few days to try a few things. I'm going on vacation next week so I'll be without phone 10 months is just unacceptable from a QA point of view from Huawei and Google. Anyway...

Up to now I get mixed results: I can boot in recovery but can't fully boot. I'll try a few others way (either kernels or parameters) later during the weekend and update this post.

Boot image includes the kernel from O DP3 and recovery image the kernel from twrp 3.1.0.0. Current method is to limit the kernel to use less cores.

If your device is affected by the Boot Loop Of Death, you might try flashing this image to backup your data.

Twrp image:
twrp_3.1.1-bootloop (four cores)

Source code - Local manifest
https://github.com/architek/omnirom-twrp-angler_local-manifest

How to flash:
Reboot into bootloader mode (hold power + volume down)
unzip
fastboot flash recovery omni-7.1-twrp_3.1.1-bootloop.img

Feedback:
Tell me if that works out for you. Thanks.

Update
14th August : Binary link
21th August : Source code link

 

[DEVILOPS 007]

My phone is 10 months old and began bootlooping a few hours ago. I've kept it in a bootloop but it never booted neither system nor recovery.

It's still under waranty but I'll keep it a few days to try a few things. I'm going on vacation next week so I'll be without phone 10 months is just unacceptable from a QA point of view from Huawei and Google. Anyway...

Up to now I get mixed results: I can boot in recovery but can't fully boot. I'll try a few others way (either kernels or parameters) later during the weekend and update this post.

Boot image includes the kernel from O DP3 and recovery image the kernel from twrp 3.1.0.0. Current method is limiting the kernel to use 1 cpu.

Urls:
Xtwrp-3.1.0-0-angler.img
Xboot-opp3.170518.006-0bc919b4.img

How to use:
Reboot into bootloader mode (hold power + volume down)
fastboot boot Ximage.img

Will boot you either in recovery or system depending on the image you use.

Feedback:
Tell me if that works out for you. Thanks.

The best thing to do is to follow this method at either step 9 or step 10. https://forum.xda-developers.com/nexus-6p/general/guides-how-to-guides-beginners-t3206928

 

[rchtk]

The best thing to do is to follow this method at either step 9 or step 10. https://forum.xda-developers.com/nexus-6p/general/guides-how-to-guides-beginners-t3206928

Explain how it is better than what I posted.

 

[DEVILOPS 007]

Explain how it is better than what I posted.

Because you didn't share what link you were using information from and you didn't say that you tried following everything like flashing the system, boot, recovery, bootloader, vendor etc as that is what needed to fix bigger bricks. Also, you didn't take into account different factors like some things might not work making your method impossible. I'm not saying your method won't work for that certain situation but the link I posted is to one of the biggest people on xda as far as I know with an almost foolproof plan to fix bricks unless it is the blod.

 

[rchtk]

Because you didn't share what link you were using information from and you didn't say that you tried following everything like flashing the system, boot, recovery, bootloader, vendor etc as that is what needed to fix bigger bricks. Also, you didn't take into account different factors like some things might not work making your method impossible. I'm not saying your method won't work for that certain situation but the link I posted is to one of the biggest people on xda as far as I know with an almost foolproof plan to fix bricks unless it is the blod.

Yes, obviously, my post is for people experiencing blod. If you don't then you don't need these images.
The recovery image is not even flashed, its executed in RAM and allowed me to access my filesystem.
The system image goes up to moving dots but doesn't finish booting.

 

[DEVILOPS 007]

Yes, obviously, my post is for people experiencing blod. If you don't then you don't need these images.
The recovery image is not even flashed, its executed in RAM and allowed me to access my filesystem.
The system image goes up to moving dots but doesn't finish booting.

If it was a blod then your method wouldn't actually work. Your post helps with with something worse than a soft brick but not the infamous proper hard brick blod.

 

[rchtk]

If it was a blod then your method wouldn't actually work. Your post helps with with something worse than a soft brick but not the infamous proper hard brick blod.

The whole post is not about if it works or not. IT DOES WORK to get to recovery on my device which has a BLOD :silly:
Got it?

If you don't need these images, respects others and just don't post in this thread. I'm waiting for feedback on if it works on others too.

 

[DEVILOPS 007]

The whole post is not about if it works or not. IT DOES WORK to get to recovery on my device which has a BLOD :silly:
Got it?

If you don't need these images, respects others and just don't post in this thread. I'm waiting for feedback on if it works on others too.

I'm just saying that it doesn't work that often lying and the blod doesn't even let you get to recovery anyway but whatever. Believe whatever, believe what you want.

 

[outbackpaul]

My phone is 10 months old and began bootlooping a few hours ago. I've kept it in a bootloop but it never booted neither system nor recovery.

It's still under waranty but I'll keep it a few days to try a few things. I'm going on vacation next week so I'll be without phone 10 months is just unacceptable from a QA point of view from Huawei and Google. Anyway...

Up to now I get mixed results: I can boot in recovery but can't fully boot. I'll try a few others way (either kernels or parameters) later during the weekend and update this post.

Boot image includes the kernel from O DP3 and recovery image the kernel from twrp 3.1.0.0. Current method is to limit the kernel to use less cores.

Urls:
Xtwrp-3.1.0-0-2-angler.img
Xboot-opp3.170518.006-0bc919b4.img

How to use:
Reboot into bootloader mode (hold power + volume down)
fastboot boot Ximage.img

Will boot you either in recovery or system depending on the image you use.

Feedback:
Tell me if that works out for you. Thanks.

Ok, so I got it booted to the TWRP 3.1.0.0 from your post but I can't see any of my files. I had TWRP 3.0.3.0 installed.

What do I do next?

I haven't tried the other image yet. I'm just excited to see what my battery percentage is so I know if I'm in the 6-9% range as mentioned in the other thread about this problem.

---------- Post added at 06:45 PM ---------- Previous post was at 06:25 PM ----------

Ok, so I just tried the opp file and it's booted up to the Google logo and stops.

 

[rchtk]

Ok, so I got it booted to the TWRP 3.1.0.0 from your post but I can't see any of my files. I had TWRP 3.0.3.0 installed.

What do I do next?

I haven't tried the other image yet. I'm just excited to see what my battery percentage is so I know if I'm in the 6-9% range as mentioned in the other thread about this problem.

Glad it also works for you. Well that's weird that you don't see your files under /data .
Can you try to mount /data (and maybe open shell and issue dmesg) ? Were you encrypted?
That looks similar to me. Mine was encrypted and it actually failed to decrypt it. So I formated it in the hope that it would help to boot.

---------- Post added at 06:45 PM ---------- Previous post was at 06:25 PM ----------

Ok, so I just tried the opp file and it's booted up to the Google logo and stops.

With a modified ramdisk to remove ciphering and dm-verity, mine fails right away. With the boot img I posted, the moving dots show for like 5 minutes (creating cipher metadata?) so most of the partitions are mounted but it cannot fully boot.
I did a simple check of the emmc (dding mmc to /dev/null) which is ok but I actually need to compile a pattern write and read to confirm that.
Unfortunatly, I have no USB to serial device to see eventual kernel logs.

 

[rchtk]

Also, what are the modifications done to this recovery and boot image?

"Boot image includes the kernel from O DP3 and recovery image the kernel from twrp 3.1.0.0. Current method is to limit the kernel to use less cores."
I have several kernels with different modifications but the original ones are actually the ones that work better. These are original kernels but called with a maxcpus parameters (isolcpu would fit better but doesn't seem to do the job).
The recovery I posted is 5 cpus IIRC, its fast enough and works from RAM but not always when flashed. Flashing a recovery when cpu is lowered to 1 always works.
Again, on MY device, might not be the same for you.

Edit: links worked with proxy

Weird that going through a proxy helps. Whatever.

I'll try a few others things in the next days but I'm sending the device on Thursday at the latest... I'll try to get in touch with some kernel devs. Any idea is welcome!

 

[rchtk]

Tried booting the images. Unfortunately, both of them didn't work for me.

Sent from my XT1092 using Tapatalk

You fastboot boot Xrecovery and it didn't boot? fastboot flash followed by rebooting into recovery might not work.
This evening, I'll post one which limits to only one core.

For the system one, it has more requirements. Not sure if there is backward compatiblity (being able to start an old ROM/vendor with a very recent kernel).

I tried so many things. Only not the oven yet

 

[outbackpaul]

Glad it also works for you. Well that's weird that you don't see your files under /data .
Can you try to mount /data (and maybe open shell and issue dmesg) ? Were you encrypted?
That looks similar to me. Mine was encrypted and it actually failed to decrypt it. So I formated it in the hope that it would help to boot.

With a modified ramdisk to remove ciphering and dm-verity, mine fails right away. With the boot img I posted, the moving dots show for like 5 minutes (creating cipher metadata?) so most of the partitions are mounted but it cannot fully boot.
I did a simple check of the emmc (dding mmc to /dev/null) which is ok but I actually need to compile a pattern write and read to confirm that.
Unfortunatly, I have no USB to serial device to see eventual kernel logs.

I'm pretty sure I was encrypted. I don't have the phone with me at the moment.

So, if it's encrypted am I out of luck?

When it booted to TWRP it asked for a passcode. I just hit cancel. Is that what will let me see my data?

 

[rchtk]

I'm pretty sure I was encrypted. I don't have the phone with me at the moment.

So, if it's encrypted am I out of luck?

When it booted to TWRP it asked for a passcode. I just hit cancel. Is that what will let me see my data?

So you're in the same boat as me (hope you had backups like me).
Twrp 3.1 should be able to decrypt /data with the correct password. I guess it first tries with default_password and if a known pattern is not detected it asks for your passcode (pin to boot by default) until it succeeds.
But for me (encrypted with pin to boot), twrp wasn't able to decrypt. In my tests once I format /data, it gets encrypted again and twrp again fails to decrypt the partition.
Can you try with pass "default_password" (without quotes) or your pin to boot if you had set one?
Also do the same after flashing the one core version I just posted and rebooting into it.

 

[rchtk]

Eagerly waiting to try the one core only recovery.

http://dl.free.fr/oIkytaKDF

 

[rchtk]

I'm pretty sure I was encrypted. I don't have the phone with me at the moment.

So, if it's encrypted am I out of luck?

When it booted to TWRP it asked for a passcode. I just hit cancel. Is that what will let me see my data?

You should also try to *flash* the one core version and see if you're able to decrypt your data with it.

 

[outbackpaul]

So you're in the same boat as me (hope you had backups like me).
Twrp 3.1 should be able to decrypt /data with the correct password. I guess it first tries with default_password and if a known pattern is not detected it asks for your passcode (pin to boot by default) until it succeeds.
But for me (encrypted with pin to boot), twrp wasn't able to decrypt. In my tests once I format /data, it gets encrypted again and twrp again fails to decrypt the partition.
Can you try with pass "default_password" (without quotes) or your pin to boot if you had set one?
Also do the same after flashing the one core version I just posted and rebooting into it.

I don't even remember putting in a pin but I've tried every one that I remember ever using.

 

[rchtk]

One core version didn't work either. Guess I'm out of options.

Sent from my XT1092 using Tapatalk

It would be interesting to start from the same configuration by using the O DP3 factory image script flash-all.sh (removing -w if you want to keep your data).

Progress:
The fact that twrp can't decrypt data with the correct password is not understood. I'm getting a KM_ERROR_KEY_NOT_YET_VALID error message. Which is rather weird.
KM_TAG_ACTIVE_DATETIME requires comparison with a trusted UTC time source. If the current date and time is prior to the tag value, the method must return KM_ERROR_KEY_NOT_YET_VALID.
So I've formated /data, disabled encryption and verify and twrp sees my /data which is populated during system boot.
Regarding boot img, I've tried several things but I still fail to get past the jumping dots. It could be a consequence of the twrp problem about communication error with the keystore which has nothing to do with big cluster disabling (so there may be a general HW failure). I've tried playing a bit with big cluster settings but my attempts fail. Still a lot of things to try but I won't have much time in the next days.

 

[mnhs1010]

Nexus 6p bootlooping just

Tried both did not work
command used

Fastboot flash boot Xboot-opp3.170518.006-0bc919b4.img

Fastboot flash recovery Xtwrp-3.1.0-0-angler.img
and
Fastboot boot Xtwrp-3.1.0-0-angler.img


not working

 

[rchtk]

RIP

Game is over for me: I've let the phone charge at home and when trying to power it back, I just get a red blinking led. I always managed to power it back but this time not.
Which brings me to the conclusion that it looks power related. Battery, soldering, IC, motherboard (components are more prone to fail stop failures).
In the end, I've mostly changed kernel stuffs (and discovered a few kernel inconsistencies btw) but didn't have time to try a lot of things in userland nor investigate on the decryption failure.
My goal was just to get users to get back their data before throwing/sending back the brick.

So I'd suggest users to decrypt their /data before BLOD happens.
A bit frustrating but was fun anyway. Thanks to @Dees_Troy for the chat on IRC