Question Bootloop after C10 update - OnePlus Nord 2 5G

Hey,
I've recently updated my Nord 2 from A21 to C10. Phone was unlocked and rooted, so after having reflashed the original boot.img, I forced the installation of the official OTA through TWRP. I had to set ro.commonsoft.ota=OP515BL1 to make it work. After the installation, TWRP failed to mount /system, but that didn't surprised me. I checked that the boot partition has been well flashed.
Now every time I try to power on the phone, it directly tries to run into recovery mode. However it fails and start again and again...
Maybe the system tries to install the OTA using the original recovery, which of course fails, and because of an unknown reason, it doesn't reboot to system.
Because of the last update, fastboot is not accessible anymore using vol -, and BROM mode is not accessible using vol + / vol -.
I tried to crash the preloader using mtkclient but it didn't work.
I tried to use META mode to switch to fastboot, but preloader only answers "READY" (instead of "READYTOOBTSAF"), and nothing changes.
I try to reverse engineer preloader and lk but it's something new for me. META mode code is still present in the preloader, so I don't understand what's wrong with it. Maybe disabled by default on USB...
Does anyone has a solution to boot into BROM mode or make META mode work ?
Or maybe I could find DA authentication files somewhere ?

---

@Petitoto can you share a bit about how you got the meta command running?
I'm in a similar situation with a Nord 2T. While mtkclient can get some info out of the preloader, meta never seems to connect.
Code:
mtk gettargetconfig
Preloader - Status: Waiting for PreLoader VCOM, please connect mobile
Port - Device detected :)
Preloader - CPU: MT6893(Dimensity 1200)
Preloader - HW version: 0x0
Preloader - WDT: 0x10007000
Preloader - Uart: 0x11002000
Preloader - Brom payload addr: 0x100a00
Preloader - DA payload addr: 0x201000
Preloader - CQ_DMA addr: 0x10212000
Preloader - Var1: 0xa
Preloader - Disabling Watchdog...
Preloader - HW code: 0x950
Preloader - Target config: 0x5
Preloader - SBC enabled: True
Preloader - SLA enabled: False
Preloader - DAA enabled: True
Preloader - SWJTAG enabled: True
Preloader - EPP_PARAM at 0x600 after EMMC_BOOT/SDMMC_BOOT: False
Preloader - Root cert required: False
Preloader - Mem read auth: False
Preloader - Mem write auth: False
Preloader - Cmd 0xC8 blocked: False
Preloader - Get Target info
Preloader - HW subcode: 0x8a00
Preloader - HW Ver: 0xca00
Preloader - SW Ver: 0x0
Main - Getting target info...
Preloader - Target config: 0x5
Preloader - SBC enabled: True
Preloader - SLA enabled: False
Preloader - DAA enabled: True
Preloader - SWJTAG enabled: True
Preloader - EPP_PARAM at 0x600 after EMMC_BOOT/SDMMC_BOOT: False
Preloader - Root cert required: False
Preloader - Mem read auth: False
Preloader - Mem write auth: False
Preloader - Cmd 0xC8 blocked: False
Code:
mtk meta FASTBOOT
META - Status: Waiting for PreLoader VCOM, please connect mobile
META - Hint:
Power off the phone before connecting.
For preloader mode, don't press any hw button and connect usb.
...........
META - Hint:
Power off the phone before connecting.
For preloader mode, don't press any hw button and connect usb.
...........
META - Hint:
Power off the phone before connecting.
For preloader mode, don't press any hw button and connect usb.

---

Hey @Beanow,
I have the same gettargetconfig output, which indicates that the phone is not in BROM mode but stuck in preloader. Trying to interact with the preloader always lead to error because of the DAA (DAA_SIG_VERIFY_FAILED for example).
I have the same issue with mtkclient and meta mode. You can use the following modified mtk-bootseq.py:
py mtk-bootseq.py FASTBOOT COMXX (or python3 mtk-bootseq.py FASTBOOT /dev/ttyACMXX on linux).
Python:
import sys
import time
from serial import Serial
BOOTSEQ = bytes(sys.argv[1], "ascii")
DEVICE = sys.argv[2]
CONFIRM = b"READY" + BOOTSEQ[::-1]
while True:
try:
s = Serial(DEVICE, 115200, timeout=0.1)
print(".\n[+] Device detected")
break
except OSError as e:
sys.stdout.write("."); sys.stdout.flush()
time.sleep(0.1)
print("<-", s.read(256))
def send(bytes):
s.write(bytes)
print("->", str(bytes))
resp = s.read(256)
print("<-", str(resp))
return resp
resp = b''
while resp != CONFIRM:
resp = send(BOOTSEQ)
print("[+] Boot sequence sent")
On another device, it works and I get:
Code:
...............................
[+] Device detected
<- b'READYREADYREADYREADYREADY'
-> b'FASTBOOT'
<- b'READYTOOBTSAF'
[+] Boot sequence sent
However, on my Nord 2, I get:
Code:
...........................................
[+] Device detected
<- b'READYREADYREADYREADYREADY'
-> b'FASTBOOT'
<- b'READY'
-> b'FASTBOOT'
<- b''
-> b'FASTBOOT'
<- b''
Then the next s.write() is hanging.
I get the same result for any other boot mode. However, the code is still present in the preloader.
I unfolded my phone to try to find a test point. I tried all golden points but I only found:
- a point which loads preloader (and not BROM...) in the same way vol + / - do (in red in the picture)
- a point which boots the phone but without Android and OnePlus pictures (what's that ??) (in green)
{
"lightbox_close": "Close",
"lightbox_next": "Next",
"lightbox_previous": "Previous",
"lightbox_error": "The requested content cannot be loaded. Please try again later.",
"lightbox_start_slideshow": "Start slideshow",
"lightbox_stop_slideshow": "Stop slideshow",
"lightbox_full_screen": "Full screen",
"lightbox_thumbnails": "Thumbnails",
"lightbox_download": "Download",
"lightbox_share": "Share",
"lightbox_zoom": "Zoom",
"lightbox_new_window": "New window",
"lightbox_toggle_sidebar": "Toggle sidebar"
}
I don't know how test point is handled: if that's the role of preloader, it may have been disabled by the update (as the BROM and fastboot). We may need to find the DAT0 point of the eMMC to short it and prevent the BROM to find the preloader, making it to go in EDL mode. However, I think that this point isn't exposed, and I won't disassemble my phone further without beeing sure of success...

---

Thank you so much for the work so far!
Unfortunately I get no response at all on the Nord 2T.
Code:
.......................................
[+] Device detected
<- b''
-> b'FASTBOOT'
<- b''
-> b'FASTBOOT'
Traceback (most recent call last):
File "/media/droid-work/mtkclient/mtk-bootseq.py", line 31, in <module>
resp = send(BOOTSEQ)
File "/media/droid-work/mtkclient/mtk-bootseq.py", line 24, in send
resp = s.read(256)
File "/usr/lib/python3.10/site-packages/pyserial-3.5-py3.10.egg/serial/serialposix.py", line 595, in read
raise SerialException(
serial.serialutil.SerialException: device reports readiness to read but returned no data (device disconnected or multiple access on port?)
How did you connect to the device that you're getting these responses?
In my case, I need to use vol+, vol- and power, like mtkclient, or the ttyACM0 won't exist.
(I've got udevadm monitor up, watching for the usb/tty to be added)

---

Indeed, you need to run into preloader using vol +, vol -
Maybe a driver / python module issue. I've got similar issues on my linux. Try on windows or try to reinstall drivers.
It should work at least for the first answer. Else it means that your preloader doesn't send any data, which is not the case as mtkclient works.

---

I also tried a different baud, because a pl_lk log from oplusreserve2 partition suggested it may be used. No luck though. Note, this was a very old log I saved early on. Definitely not reflective of latest Nord 2T update.
Code:
[PLFM] boot_tag size = 0x0
BOOT_TAG_VERSION: 0
BOOT_REASON: 0
BOOT_MODE: 0
META_COM TYPE: 0
META_COM ID: 0
META_COM PORT: 285220864
META LOG DISABLE: 0
FAST META GPIO: 5906
LOG_COM PORT: 285220864
LOG_COM BAUD: 921600
LOG_COM EN: 1
LOG_COM SWITCH: 0
MEM_NUM: 2
MEM_SIZE: 0xAE7B
MEM_SIZE: 0xAE8D
I guess I'll try windows then

---

Code:
python mtk-bootseq.py FASTBOOT COM4
...................................................................................................................................
[+] Device detected
<- b''
-> b'FASTBOOT'
<- b''
-> b'FASTBOOT'
<- b''
-> b'FASTBOOT'
<- b''
Windows looks to behave similar. Though windows wouldn't take the MTK VCOM driver, so this is win10 default serial, in a VM over USB passthrough.

---

So, same result not in a VM. Though specifically with powershell I got the same output as you did.
Code:
...........................................
[+] Device detected
<- b'READYREADYREADYREADYREADY'
-> b'FASTBOOT'
<- b'READY'
-> b'FASTBOOT'
<- b''
-> b'FASTBOOT'
<- b''

---

This is really a helpfull post for us. I've already a oneplus nord 2 phn,from this post i know the more information about this phn.

---

Thank you so much.

---

@Beanow So same results...
It's weird that it doesn't work on Linux. Maybe an issue related to pyserial or connection settings.
What's preventing the device to be detected by mtkclient is line 54 in mtkclient/Library/meta.py: and cdc.pid == 0x2000 should be removed. So you can try to switch to fastboot using mtkclient on Linux, but with my Nord2 I get the same results as mtk-bootseq.py on Windows

---

Petitoto said:
@Beanow So same results...
It's weird that it doesn't work on Linux. Maybe an issue related to pyserial or connection settings.
What's preventing the device to be detected by mtkclient is line 54 in mtkclient/Library/meta.py: and cdc.pid == 0x2000 should be removed. So you can try to switch to fastboot using mtkclient on Linux, but with my Nord2 I get the same results as mtk-bootseq.py on Windows
Click to expand...
Click to collapse
Thanks for this. No need to switch to windows anymore, to use mtk client.

---

Petitoto said:
It's weird that it doesn't work on Linux. Maybe an issue related to pyserial or connection settings.
Click to expand...
Click to collapse
Is it 'not working' though? It's also weird to me that I had the same output as Linux using Windows' cmd, while there was READY spam in powershell. Same drivers, same python, same libraries, but different output?
I suspect that it might be a timing issue. Maybe the serial console doesn't care about or wait for input at all. And just spams READY a few times. It would be a matter of how fast the connection is established.
Perhaps as well there's a different subsystem sending commands to the 'meta' environment and the READY spam means it's processing those commands rather than whatever we're sending.
All theories, but I would find it really hard to believe there's a problem with Linux drivers / libraries for something as basic as a UART/serial console over USB.

---

Petitoto said:
@Beanow So same results...
It's weird that it doesn't work on Linux. Maybe an issue related to pyserial or connection settings.
What's preventing the device to be detected by mtkclient is line 54 in mtkclient/Library/meta.py: and cdc.pid == 0x2000 should be removed. So you can try to switch to fastboot using mtkclient on Linux, but with my Nord2 I get the same results as mtk-bootseq.py on Windows
Click to expand...
Click to collapse
I also suspected this PID check and tried to log the else cases, but never reaches those for me.
So removing the check didn't help for mtkclients' meta commands.

---

Is it 'not working' though? It's also weird to me that I had the same output as Linux using Windows' cmd, while there was READY spam in powershell. Same drivers, same python, same libraries, but different output?
Click to expand...
Click to collapse
Differents results when using cmd and powershell? There is really no reason for that. Unless it's not the same Python environment, with different pyserial for eg. I have issues to run mtk-bootseq on Linux, but always the same output on Windows' cmd.
I suspect that it might be a timing issue. Maybe the serial console doesn't care about or wait for input at all. And just spams READY a few times. It would be a matter of how fast the connection is established.
Click to expand...
Click to collapse
Maybe. On linux, I can get different results depending on baud rate, timeout (and luck?). If there is an issue related to the connection, it might explain why the preloader doesn't answer as expected. But as other commands (like mtk gettargetconfig, but also manually handshaking connections and gathering informations in pyserial) work well, I tend to think it's just disabled.
Perhaps as well there's a different subsystem sending commands to the 'meta' environment and the READY spam means it's processing those commands rather than whatever we're sending.
Click to expand...
Click to collapse
I don't really know how it works. The code is still present in the preloader. However this functionnality is not always enabled. Maybe reversing the preloader more or analysing the log you provided on Github might help to determine whether or not it is enabled. Moreover, even if we manage to switch to fastboot, if the bootloader has been fully disabled, we may face the issue of the preloader trying to run into a non existant fastboot. Maybe the FACTFACT mode may help to reset the device, but I don't really know a lot about this mode.
So removing the check didn't help for mtkclients' meta commands.
Click to expand...
Click to collapse
Once you removed this check, if you print the data sent by the preloader, you'll get the multiple "READY" like mtk-bootseq on Windows. Moreover, I can switch to fastboot using this command on another MTK device.

---

Dear Sir,
Do you have any method to recover my phone as the figure show?
Thank You