*** Please note that this reply is generated by LLM automatically ***
The error message you’re seeing suggests that there’s an issue with the PCIe link on your Jetson AGX Orin module. Specifically, it’s complaining about the “Phy link never came up” on the PCIe interface.
Here are a few potential causes and solutions to this issue:
Hardware issue: It’s possible that there’s a hardware problem with the PCIe interface on your Jetson module or the device you’re trying to connect to it. Try swapping out the device or the cable to see if the issue persists.
Firmware issue: The error message mentions “UEFINVME” which suggests that the issue might be related to the UEFI firmware on your Jetson module. Try updating the firmware to the latest version to see if it resolves the issue.
Software issue: It’s possible that there’s a software issue that’s preventing the PCIe link from coming up. Try rebooting your Jetson module or reloading the PCIe driver to see if it resolves the issue.
Configuration issue: It’s possible that there’s a configuration issue that’s preventing the PCIe link from coming up. Try checking the PCIe configuration on your Jetson module to make sure it’s set up correctly.
In terms of comparing this issue to other potential issues, here are a few things to keep in mind:
Here are some specific commands you can try to troubleshoot the issue:
lspci -v to check the PCIe configuration on your Jetson moduledmesg | grep PCIe to check the kernel logs for any PCIe-related errorssudo jetson-io to check the device tree configuration on your Jetson moduleI hope this helps! Let me know if you have any other questions.
*** Please give the thumbs up if you get answers you like. Or provide feedback to help us improve the answer. ***
Nothing to configure. Default software already configure everything you need.
If custom board is not able to detect it, please check hardware signal and design.
No, it does not affect software configuration.
Please follow up the debug tips here
我们现在疑惑为什么在UEFI下能识别到NVME盘
我们在将C1的速度降到GEN 2后lspci能找到两个NVME的控制器，但是格式化NVME盘后，现在出现了几种不稳定的情况
1、lspci能找到两个NVME控制器，能找到两个盘
2、lspci能找到两个NVME控制器，但是只能找到一个盘（C1和C4都有可能）
3、lspci只能找到C4的控制器
4、重启时UEFI会报断言的错误 assert [nvmexpressdxe] /out/nvidia/bootloader/uefi/jetson_release/edk2/mdemodulepkg/bus/pci/nvmexpressdxe/nvmexpresshci.c(772): (private->cap.mpsmin + 12) <= 12
麻烦帮忙解答一下
2的狀況聽起來好像跟tegra pcie本身無關, 請問有dmesg能提供嗎？
dmesg_log51104_144918.txt (94.9 KB)
請問在這情況下的sudo lspci -vvv的結果跟偵測到時候的能分享一下嗎
看來出問題的時候link detection還是有點問題. 請問有pcie analyzer幫忙抓trace嗎？
有其他ssd可以測試嗎？
更换了我们之前在agx orin模组上测试稳定的盘，现在C4的盘能固定找到，但是C1的盘还是会有概率找不到，基本上跟我上面描述的情况一致
请问之前有碰到类似的案例吗？再次跟您确认下，C1上安装NVME是否需要修改pcie部分的驱动？
我们现在不方便测量信号质量，请问下还有什么排查思路吗？麻烦回复下，非常感谢
Suggest to check hardware signal and design.
