SDK manager flash failed

Autonomous Vehicles DRIVE AGX Orin DRIVE AGX Orin General
3

Hi sir
And the below is host’s dmesg log:
[四 12月 15 13:42:14 2022] usb 1-4: USB disconnect, device number 9
[四 12月 15 13:42:29 2022] usb 1-5: new full-speed USB device number 12 using xhci_hcd
[四 12月 15 13:42:29 2022] usb 1-5: New USB device found, idVendor=0955, idProduct=7045, bcdDevice= 0.01
[四 12月 15 13:42:29 2022] usb 1-5: New USB device strings: Mfr=1, Product=2, SerialNumber=3
[四 12月 15 13:42:29 2022] usb 1-5: Product: Tegra On-Platform Operator
[四 12月 15 13:42:29 2022] usb 1-5: Manufacturer: NVIDIA
[四 12月 15 13:42:29 2022] usb 1-5: SerialNumber: TOPODD55FB3B
[四 12月 15 13:42:29 2022] hid-generic 0003:0955:7045.0005: hiddev0,hidraw2: USB HID v1.10 Device [NVIDIA Tegra On-Platform Operator] on usb-0000:00:14.0-5/input0
[四 12月 15 13:42:29 2022] cdc_acm 1-5:1.1: ttyACM0: USB ACM device
[四 12月 15 13:42:29 2022] cdc_acm 1-5:1.3: ttyACM1: USB ACM device
[四 12月 15 13:42:29 2022] cdc_acm 1-5:1.5: ttyACM2: USB ACM device
[四 12月 15 13:42:29 2022] cdc_acm 1-5:1.7: ttyACM3: USB ACM device
[四 12月 15 13:46:04 2022] usb 1-11: new high-speed USB device number 13 using xhci_hcd
[四 12月 15 13:46:04 2022] usb 1-11: New USB device found, idVendor=0955, idProduct=7023, bcdDevice= 4.01
[四 12月 15 13:46:04 2022] usb 1-11: New USB device strings: Mfr=1, Product=2, SerialNumber=0
[四 12月 15 13:46:04 2022] usb 1-11: Product: APX
[四 12月 15 13:46:04 2022] usb 1-11: Manufacturer: NVIDIA Corp.
[四 12月 15 13:48:55 2022] wlp0s20f3: AP 88:40:3b:d6:12:97 changed bandwidth, new config is 2447 MHz, width 2 (2457/0 MHz)
[四 12月 15 13:49:22 2022] wlp0s20f3: AP 88:40:3b:d6:12:97 changed bandwidth, new config is 2447 MHz, width 1 (2447/0 MHz)
[四 12月 15 13:53:02 2022] usb 1-11: USB disconnect, device number 13
[四 12月 15 13:53:54 2022] usb 1-4: new high-speed USB device number 14 using xhci_hcd
[四 12月 15 13:53:55 2022] usb 1-4: New USB device found, idVendor=0955, idProduct=7023, bcdDevice= 4.01
[四 12月 15 13:53:55 2022] usb 1-4: New USB device strings: Mfr=1, Product=2, SerialNumber=0
[四 12月 15 13:53:55 2022] usb 1-4: Product: APX
[四 12月 15 13:53:55 2022] usb 1-4: Manufacturer: NVIDIA Corp.
[四 12月 15 13:54:51 2022] usb 1-4: USB disconnect, device number 14
[四 12月 15 13:54:52 2022] usb 1-4: new high-speed USB device number 15 using xhci_hcd
[四 12月 15 13:54:52 2022] usb 1-4: New USB device found, idVendor=0955, idProduct=7023, bcdDevice= 4.01
[四 12月 15 13:54:52 2022] usb 1-4: New USB device strings: Mfr=1, Product=2, SerialNumber=0
[四 12月 15 13:54:52 2022] usb 1-4: Product: APX
[四 12月 15 13:54:52 2022] usb 1-4: Manufacturer: NVIDIA Corp.
[四 12月 15 13:58:19 2022] usb 1-4: USB disconnect, device number 15
[四 12月 15 13:58:20 2022] usb 1-4: new high-speed USB device number 16 using xhci_hcd
[四 12月 15 13:58:20 2022] usb 1-4: New USB device found, idVendor=0955, idProduct=7023, bcdDevice= 4.01
[四 12月 15 13:58:20 2022] usb 1-4: New USB device strings: Mfr=1, Product=2, SerialNumber=0
[四 12月 15 13:58:20 2022] usb 1-4: Product: APX
[四 12月 15 13:58:20 2022] usb 1-4: Manufacturer: NVIDIA Corp.

4

Dear @sunjiandong,
Could you check AGX Orin Initial Setup - #17 by VickNV if it helps

5

Hi sir
I have check the usb. It is usb2.0 .
And whether the host will have adb node after the type-C connect to the host’s USB2.0 port?
Thanks!

6

Dear @sunjiandong,
I see “echo Failed to flash device: detected no board connected to host with USB cable” in above attached image. Make sure you have connected board to host.

Firstly please check below things.

  • Check if lsusb outputs Nvidia device after connecting host and target. You can verify lsusb output before and after the connection. Share the lsusb outputs.
  • Check if tegra can be set into recovery mode. You can connect to aurix console with minicom using sudo minicom -D /dev/ttyACM1 and run tegrarecovery x1 on and tegrareset x1. Now verify if you see an additional Nvidia device in lsusb output on host. After that put the target back in normal mode by running tegrarecovery x1 offandtegrareset x1` on aurix console.
  • If flashing fails, attach the complete sdkm logs for investigation.

We notice changing cables fixed flashing issues for couple of customers like in AGX Orin Initial Setup and DRIVE AGX Orin Flashing issue

7

Hi sir
Thanks for your reply. After set into the recovery mode, I can see the there have a APX port.Is it right? And I use the cable which is from the Nvidia.
nvsdkm.tar.gz (170.4 KB)

image
image824×384 85.5 KB

9

After checking your logs, we think it can be solved with the workaround mentioned at DRIVE AGX Orin Flashing issue - #5 by VickNV.
Please also see how the developer solved their problems in the topic.

10

Hi sir
From the case you list, I can see it can see it failed after the ADB port is enumerate but I failed after the Orin enter into recovery mode.
And I also change the type-c cable and use the usb2.0. It still failed.
Thanks!

11

Dear @sunjiandong,
Could you share the Tegra boot log and host’s dmesg log after using usb 2.0 port.

12

Hi sir
Do you need the Tegra log during flashing the device? How to capture?
Thanks!

13

Hi sir
The below is the Tegra boot log and the host’s dmesg log. Please help check.

host dmesg.log (98.7 KB)
Tegra boot.log (73.4 KB)

Thanks!

14

Dear @sunjiandong,
Thank you for sharing logs. I will share the logs for internal investigation.
Mean while, I would like to check manual flashing method instead of using sdkmanager.
Please try below and confirm

  • Put the target in recovery mode
sudo minicom -D /dev/ttyACM1 # on host
tegrarecovery x1 on #on Aurix console
tegrareset x1  #on Aurix console
  • bind and flash
cd drive-foundation
make -f make/Makefile.bind BOARD=p3710-10-s05 PCT=linux clean 
make -f make/Makefile.bind BOARD=p3710-10-s05 PCT=linux 
tools/flashtools/bootburn/bootburn.py -b p3710-10-s05 -B qspi
  • Put the target in normal mode
sudo minicom -D /dev/ttyACM1 # on host
tegrarecovery x1 off #on Aurix console
tegrareset x1  #on Aurix console

Please share the logs on host and aurix for these commands.

15

Hi sir
It refers as below when rung the command “make …”
aic@caic-HP-ZBook-Power-G7-Mobile-Workstation:~/nvidia/nvidia_sdk/DRIVE_OS_6.0.5_SDK_Linux_DRIVE_AGX_ORIN_DEVKITS/DRIVEOS/drive-foundation$ make -f make/Makefile.bind BOARD=p3710-10-s05 PCT=linux clean

Direct call of makefile depreciated
Do not use makefile directly. Please use bind_partitions script

make/Makefile.bind:26: *** stop。 停止。

Thanks@

16

Hi sir
The bleow is the log
The bind:
cd make/
./bind_partitions -b p3710-10-s05 linux
bind_partions.log (113.3 KB)

The bootburn:
cd …/
sudo tools/flashtools/bootburn/bootburn.py -b p3710-10-s05 -B qspi
bootburn.log (471.3 KB)

The aurix ttyACM1’s log:
aurix.log (4.3 KB)

17

Dear @sunjiandong,
Per bootburn.log, flashing is successful. Were you able to access Tegra A after putting the target in normal mode?

18

Hi sir
Yes, It can access successfully.
Does the bootburn have flash all images? Is there any other partitions need to flash? whether this command is the same with sdkmanager?
Thanks!

19

Dear @sunjiandong,
Yes. The above command flashes DRIVE OS onto target.

sdkmanager detects the board automatically using find_board_name and uses required configuration to flash board. I notice find_board_name seems to be failing as per sdkm logs

[2022-12-15 20:38:26,985 root DEBUG utilities.py 327 4622] Executing board_find_cmd python3 /home/caic/nvidia/nvidia_sdk/DRIVE_OS_6.0.5_SDK_Linux_DRIVE_AGX_ORIN_DEVKITS/DRIVEOS/drive-foundation/tools/flashtools/bootburn/bootburn.py --find_board_name -x /dev/ttyACM1 
[2022-12-15 20:38:30,419 root DEBUG pdk_bind.py 213 4622] Bind partitions failed !
Traceback (most recent call last):
  File "/dvs/git/dirty/git-master_foundation/out/embedded-foundation-generic-release-none/packaging/stage_driveinstaller/driveinstaller/module/pdk_bind.py", line 157, in pdk_bind
  File "/dvs/git/dirty/git-master_foundation/out/embedded-foundation-generic-release-none/packaging/stage_driveinstaller/driveinstaller/utils/utilities.py", line 333, in find_board_name
  File "/dvs/p4/build/sw/tools/linux/mods/python-3.6.3-glibc-2.5/lib/python3.6/subprocess.py", line 291, in check_call
subprocess.CalledProcessError: Command 'python3 /home/caic/nvidia/nvidia_sdk/DRIVE_OS_6.0.5_SDK_Linux_DRIVE_AGX_ORIN_DEVKITS/DRIVEOS/drive-foundation/tools/flashtools/bootburn/bootburn.py --find_board_name -x /dev/ttyACM1' returned non-zero exit status 43.
[2022-12-15 20:38:30,421 root ERROR pdk_bind.py 214 4622] Binding of partitions failed ! Check log for more details.

Could you check running the the same command and confirm if notice the same error.

20

Hi sir
Yes, it notice the same error when running this command manually. But it is ok when I add sudo before bootburn.py --find_board_name -x /dev/ttyACM1.
The ok log:
[bootburn]: [validateTargetBoardName(748)] : boardName p3710-10-s05 matches with the connected board
[bootburnTegra-A]: [StartNewSession_t23x(785)] : Target
[bootburnTegra-A]: [StartNewSession_t23x(786)] : Target ECID: 0x
[bootburnTegra-A]: [GetTargetECID(796)] : Target BR_CID: 0x80012344705DD2E30400000002030240

[bootburnTegra-A]: [GetTargetECID(797)] : Target ECID: 0x4705DD2E30400000002030240

[bootburnTegra-A]: [CheckFuseForAuthentication(814)] : Target Authentication State : SHA-2
[bootburnTegra-A]: [GenAdbSerialNum(575)] : ADB serial number is 8S2TQBHG8000001060I0

the failed:
[bootburn]: [setTargetConfigt23xInfoRomInfo(725)] : ****** s_InforomProdInfo 940-63710-0010
[bootburn]: [validateTargetBoardName(748)] : boardName p3710-10-s05 matches with the connected board
[bootburnTegra-A]: [executeShellCommand(164)] : shell command – /home/caic/nvidia/nvidia_sdk/DRIVE_OS_6.0.5_SDK_Linux_DRIVE_AGX_ORIN_DEVKITS/DRIVEOS/drive-foundation/tools/flashtools/flash/tegrarcm_v2 --instance /dev/bus/usb/001/017 --new_session --chip 0x23 failed
[bootburnTegra-A]: [executeShellCommand(166)] : Output: b’File rcm_state open failed\nERROR: failed to update RCM_STATE to 0\n’
[bootburnTegra-A]: [executeShellCommand(168)] : Return code: 16
[bootburnTegra-A]: [executeShellCommand(170)] : stdout: b’File rcm_state open failed\nERROR: failed to update RCM_STATE to 0\n’
[bootburnTegra-A]: [executeShellCommand(172)] : stderr: None
[bootburnTegra-A]: [sendRcmCommand(776)] : Unable to execute tegrarcm command – /home/caic/nvidia/nvidia_sdk/DRIVE_OS_6.0.5_SDK_Linux_DRIVE_AGX_ORIN_DEVKITS/DRIVEOS/drive-foundation/tools/flashtools/flash/tegrarcm_v2 --instance /dev/bus/usb/001/017 --new_session --chip 0x23
command line used was:
[‘tools/flashtools/bootburn/bootburn.py’, ‘–find_board_name’, ‘-x’, ‘/dev/ttyACM1’]

TEGRARCM – Unable to execute command

Exception 43 raised in bootburn_active
Traceback (most recent call last):

Thanks!

21

Dear @sunjiandong,
Thanks for update. It seems when you use sudo, you don’t see any issue. Do you see display on target after flashing with bootburn command?
Is there any permission issues for the user account? Could you check if launching sdkmanager with sudo helps to bypass this issue?

22

It looks something to do with your host environment.
Please try another host system or use the docker container way to install (follow Installation Methods).

23

@SivaRamaKrishnaNV @VickNV
I will out off office some days, I will have a try with your suggestions when I back.
Thanks for your reply.