Jetson orin NX 8G Flash jetson Linux fail

yes the host is Ubuntu 20.04,and not use the Virtual Box or VMware.
this PC flash the orin NX first time.

There is a known issue that some ubuntu 20.04 host has problem in flashing Orin.

Could you share the info of this host machine? Kernel version, which vendor is it.

it’s a Industry PC. if needed, i can change the OS to 18.04.
do you advise?

Do you have xavier nx devkit to test instead of custom board first?

no,we don’t have one.
last Friday, I took one pcs CPU+SSD to Nvidia agilent, the CPU+SSD can download the image and bring up ok.
update the debug uart log:
2304062021-debug_usb_log_fail.txt (50.0 KB)

Looks same as this post.

we have modify the DTSI in Linux_for_Tegra/bootloader/t186ref/BCT/tegra234-mb2-bct-misc-p3767-0000.dts, when bring-up, we check the log message, it became effect.
we don’t know how to make sure the CPU in RCM mode, find the USB device(0955:7423)?
无法确认这个芯片是否在RCM模式,通过USB设备?但是上位机已经有USB设备了,还是无法烧录,是什么条件或者缺少什么?

請問一下你現在的error是什麼?

如果lsusb可以看到0955:7423 而且flash tool可以開始跑, 那就是在rcm mode了

至於燒錄失敗就要看你的error log決定了. 到目前為止我從其他custom board那邊看下來大概有4~5種原因. 所以還是需要你們的log判斷

失败的原因似乎是cvm通讯异常,刷写日志如下:
[ 1.2807 ] MB2 Applet version 01.00.0000
[ 1.3180 ] 000000001f1e0506: E> NV3P_SERVER: Could not read eeprom for module cvm.

[ 1.4207 ] tegradevflash_v2 --oem platformdetails eeprom cvm /home/muyi/Desktop/sdk/rootfs/Linux_for_Tegra/bootloader/cvm.bin --chip 0x23 0
[ 1.4220 ] Error: Invalid option --chip
debug usb日志如下:
E> NV3P_SERVER: Failed to get eeprom handle for cvm from eeprom manager.
E> NV3P_SERVER: Could not read eeprom for module cvm.
debug_usb_log_1120.txt (6.0 KB)
flash_log_1120.txt (15.3 KB)

這邊麻煩注意一下… 我們的文件只請你把cvb eeprom關掉… 沒有說cvm eeprom也要關掉…
如果你不小心關掉了 麻煩改回來.

muyi@ubuntu:~/Desktop/sdk/rootfs/Linux_for_Tegra/bootloader$ cat t186ref/BCT/tegra234-mb2-bct-misc-p3767-0000.dts
/dts-v1/;

#include “tegra234-mb2-bct-common.dtsi”

/ {
mb2-misc {
eeprom {
cvm_eeprom_i2c_instance = <0>;
cvm_eeprom_i2c_slave_address = <0xa0>;
cvm_eeprom_read_size = <0x100>;
cvb_eeprom_i2c_instance = <0x0>;
cvb_eeprom_i2c_slave_address = <0xae>;
cvb_eeprom_read_size = <0>;
};
};
};

CVM部分,我们肯定没有修改。报错很多都和这个相关。
我们在烧录过程中,不论是命令,还是SDK,报错返回值有3,4,8,甚至会读不到RCM state.

如果cvm這部份沒有更動的話 可能是硬體部份的問題.

您估计是哪一块?烧录部分只涉及供电,usb,reset/recovery,nvme,上电时序。

这个是我拿了一片新的CPU+SSD,第一次刷机的log:
log如txt:
muyi_newCPU_PClog_230407.txt (16.5 KB)

[ 0.9964 ] Retrieving EEPROM data
[ 0.9965 ] tegrarcm_v2 --oem platformdetails eeprom cvm /home/nvidia/nvidia/nvidia_sdk/JetPack_5.1.1_Linux_JETSON_ORIN_NX_TARGETS/Linux_for_Tegra/bootloader/cvm.bin --chip 0x23 0
[ 1.0000 ] MB2 Applet version 01.00.0000
[ 1.0033 ] 000000001414032a: E> NV3P_SERVER: Could not read eeprom for module cvm.
[ 1.0464 ]
[ 1.0464 ]
[ 1.0517 ] tegradevflash_v2 --oem platformdetails eeprom cvm /home/nvidia/nvidia/nvidia_sdk/JetPack_5.1.1_Linux_JETSON_ORIN_NX_TARGETS/Linux_for_Tegra/bootloader/cvm.bin --chip 0x23 0
[ 1.0559 ] Error: Invalid option --chip

Hi,

建議你可以把log弄成txt然後附件. 直接複製貼上format都會跑掉…

另外flash error也需要UART那邊的log. 麻煩也附上一下 我猜應該還是cvm read failure.

把两个log都上传了。
PC_flash_log_1722.txt (15.4 KB)

debugUART_log_1722.txt (6.0 KB)

any comments?

麻煩參考design guide document.

debugUSB_log_ok_1816.txt (179.6 KB)
debugUSB_log_04101515.txt (46.6 KB)
附件有两份log:
debugUSB_log_ok-刷了很多次,有一次能够正常完成刷机,硬件不断电的情况下,刷完能够正常进入系统;
debugUSB_log_04101515-是刷完硬件断电重启,进不了系统;
麻烦您帮我看下,看看能不能看出是什么引起的。