nvidia xavier平台SDK管理器安装DeepStream启动异常问题_专栏

1. 前言

硬件： Jetson AGX Xavier
DeepStream Version: DeepStream 5.0
SDK版本: Jetpack 4.4

安装Jetpack和DeepStream使用命令行SDK管理器和获取包依赖错误。

按照命令：

sdkmanager --cli install --logintype devzone --product Jetson
 --version 4.4 --targetos Linux --host --target P3448-0000
  --flash all --additionalsdk CUDA --additionalsdk "CUDA-X AI" 
  --additionalsdk "Computer Vision" --additionalsdk "NVDIA Container Runtime" 
  --additionalsdk Multimedia --additionalsdk DeepStream 
  --license accept

error：

info: You might want to run 'apt --fix-broken install' to correct these. 
info: The following packages have unmet dependencies: 
info:  nvidia-l4t-jetson-multimedia-api : Depends: nvidia-l4t-camera(= 32.4.3-20200625213809) but 32.4.3-20200625213407 is to be installed i
info: info: The following packages have unmet dependencies: 
info:  nvidia-l4t-jetson-multimedia-api : Depends: nvidia-l4t-camera (= 32.4.3-20200625213809) but 32.4.3-20200625213407 is to be installed 
info: Depends: nvidia-l4t-multimedia (= 32.4.3-20200625213809) but 32.4.3-20200625213407 is to be installed 
info: Depends: nvidia-l4t-multimedia-utils (= 32.4.3-20200625213809) but 32.4.3-20200625213407 is to be installed libgstrtspserver-1.0-0 libjansson4 librdkafka1=0.11.3-1
info: [ Package PreInstall Finished with Error ]
info: [ NV_DEEPSTREAM_TARGET_POST_INSTALL_COMP Install took 6s ] error: [error]: : [exec_command]: /bin/bash -c scp -q -o UserKnownHostsFile=/dev/null -o StrictHostKeyChecking=no /tmp/tmp_NV_DEEPSTREAM TARGET_POST_INSTALL_COMP.sh drishtic@192.168.55.1:~; 
ssh -t -q -o UserKnownHostsFile=/dev/null -o StrictHostKeyChecking=no drishtic@192.168.55.1 "~/tmp_NV_DEEPSTREAM_TARGET_POST_INSTALL_COMP.sh && rm -f ~/tmp_NV_DEEPSTREAM_TARGET_POST_INSTALL_COMP.sh"; 
[error]: info:error: command terminated with error 
info: Failed to execute commands with GenericInstaller

2. 调整板端型号选型

P3448-0000 更换成P2888-0004后，可解决上述error

继续安装报以下错误：

debug: running command < [ `lsusb | grep -c "0955:"` -ne 1 ] >
debug: command terminated with error
info: Only one NVIDIA device detected, as expected.
info: Start to check if ip and ssh up with default ip...
debug: running command < true >
info: command finished successfully
debug: running command < ip addr | grep 192.168.55.1 >
info:     inet 192.168.55.100/24 brd 192.168.55.255 scope global dynamic noprefixroute enp0s20f0u3
info: command finished successfully
debug: running command < true >
info: command finished successfully
debug: running command < nc -z -vv -w 5 192.168.55.1 22 >
info: nc: connect to 192.168.55.1 port 22 (tcp) failed: Connection refuseddebug: command terminated with error
error: Cannot connect to the device via SSH. Validate that SSH service is running on the device.

尝试使用sdkmanager GUI进行进行刷机，发现仍然启动异常：

sudo ./flash.sh jetson-xavier mmcblk0p1

如下表示刷写成功

但是仍然启动异常

3. 查看串口打印

使用micro usb进行调试

[0004.209] starting app kernel_boot_app
[0004.228] I> found decompressor handler: lz4-legacy
[0004.229] I> decompressing BMP blob ...
[0004.233] I> Kernel type = Normal
[0004.233] I> Loading kernel-bootctrl from partition
[0004.233] I> Loading partition kernel-bootctrl at 0xa42b0000 from device(0x1)
[0004.240] W> tegrabl_get_kernel_bootctrl: magic number(0x00000000) is invalid
[0004.240] W> tegrabl_get_kernel_bootctrl: use default dummy boot control data
[0004.241] I> ########## SD boot ##########
[0004.245] I> No sdcard
[0004.247] I> -0 params source = 
[0004.250] E> Blockdev open: exit error
[0004.253] E> SD boot failed, err: 724238353
[0004.257] I> ########## USB boot ##########
[0004.266] W> No valid slot number is found in scratch register
[0004.267] W> Return default slot: _a
[0004.283] I> USB Firmware Version: 60.06 release
[0004.338] I> regulator of usb2-0 already enabled
[0004.345] I> regulator of usb2-1 already enabled
[0004.352] I> regulator of usb2-2 already enabled
[0004.361] I> enabling 'vdd-5v-sata' regulator
[0005.365] E> failed to initialize xhci controller
[0005.366] E> Error in init of XUSB host driver, err: 79790026
[0005.366] E> Failed to initialize device 5-0
[0005.367] E> USB boot failed, err: 2037973030
[0005.367] I> ########## Fixed storage boot ##########
[0005.367] I> Already published: 00010003
[0005.368] I> Look for boot partition
[0005.371] I> Fallback: assuming 0th partition is boot partition
[0005.377] I> Detect filesystem
[0005.404] I> Loading extlinux.conf ...
[0005.404] I> rootfs path: /sdmmc_user/boot/extlinux/extlinux.conf
[0005.447] I> L4T boot options
[0005.447] I> [1]: "primary kernel"
[0005.448] I> Enter choice: 
[0008.449] I> Continuing with default option: 1
[0008.449] I> Loading kernel sig file from rootfs ...
[0008.449] I> rootfs path: /sdmmc_user/boot/Image.sig
[0008.468] I> Loading kernel binary from rootfs ...
[0008.468] I> rootfs path: /sdmmc_user/boot/Image
[0008.708] I> Validate kernel ...
[0008.708] I> T19x: Authenticate kernel (bin_type: 37), max size 0x5000000
[0009.030] I> No kernel-dtb binary path
[0009.031] W> No valid slot number is found in scratch register
[0009.031] W> Return default slot: _a
[0009.032] I> A/B: bin_type (38) slot 0
[0009.032] I> Loading kernel-dtb from partition
[0009.032] I> Loading partition kernel-dtb at 0x91000000 from device(0x1)
[0009.041] I> Validate kernel-dtb ...
[0009.041] I> T19x: Authenticate kernel-dtb (bin_type: 38), max size 0x400000
[0009.046] I> Loading ramdisk from rootfs ...
[0009.047] I> rootfs path: /sdmmc_user/boot/initrd
[0009.099] I> Kernel hdr @0xa42b0000
[0009.100] I> Kernel dtb @0x90000000
[0009.100] I> decompressor handler not found
[0009.100] I> Copying kernel image (34330640 bytes) from 0xa42b0000 to 0x80080000 ... [0009.106] I> Done
[0009.107] I> Updated bpmp info to DTB
[0009.108] I> Ramdisk: Base: 0x92000000; Size: 0x54ecaf
[0009.108] I> Updated initrd info to DTB
[0009.108] W> WARN: Fail to override "console=none" in commandline
[0009.112] E> tegrabl_linuxboot_add_disp_param, du 0 failed to get display params
[0009.120] E> tegrabl_linuxboot_add_disp_param, du 0 failed to get display params
[0009.127] E> tegrabl_linuxboot_add_disp_param, du 0 failed to get display params
[0009.134] W> No valid slot number is found in scratch register
[0009.140] W> Return default slot: _a
[0009.143] I> Active slot suffix: 
[0009.147] I> add_boot_slot_suffix: slot_suffix = 
[0009.151] I> Linux Cmdline: console=ttyTCU0,115200 video=tegrafb no_console_suspend=1 earlycon=tegra_comb_uart,mmio32,0x0c168000 gpt usbcore.old_scheme_first=1 tegraid=19.1.2.0.0 maxcpus=8 boot.slot_suffix= boot.ratchetvalues=0.4.2 vpr_resize sdhci_tegra.en_boot_part_access=1 
[0009.176] I> Updated bootarg info to DTB
[0009.179] W> MAC addr invalid!
[0009.182] E> Failed to get WIFI MAC address
[0009.186] W> MAC addr invalid!
[0009.189] E> Failed to get Bluetooth MAC address
[0009.194] I> eeprom_get_mac_addr: MAC (type: 2): 00:04:4b:cc:25:d2
[0009.200] W> "plugin-manager" doesn't exist, creating
[0009.205] I> Adding /chosen/plugin-manager/cvm
[0009.209] W> "chip-id" doesn't exist, creating
[0009.213] I> Adding /chosen/plugin-manager/chip-id
[0009.218] W> "configs" doesn't exist, creating
[0009.222] I> Adding /chosen/plugin-manager/configs
[0009.227] W> "ids" doesn't exist, creating
[0009.231] I> Adding /chosen/plugin-manager/ids
[0009.235] W> "odm-data" doesn't exist, creating
[0009.239] I> Adding /chosen/plugin-manager/odm-data
[0009.247] W> "memory" doesn't exist, creating
[0009.249] I> [0] START: 0x80000000, END: 0xac000000
[0009.253] I> [1] START: 0xac004000, END: 0xf09d0000
[0009.258] I> [2] START: 0xf09dc000, END: 0xf09e0000
[0009.262] I> dram_block larger than 80000000
[0009.266] I> [3] START: 0x100000000, END: 0x480000000
[0009.271] I> added [base:0x80000000, size:0x2c000000] to /memory
[0009.277] I> added [base:0xac200000, size:0x44600000] to /memory
[0009.283] I> added [base:0x100000000, size:0x380000000] to /memory
[0009.290] I> Updated memory info to DTB
[0009.293] E> add_disp_param: failed to get display params for du=0
[0009.299] W> "reset" doesn't exist, creating
[0009.304] I> NVG: Logical CPU: 0; MPIDR: 0x80000000
[0009.308] I> NVG: Logical CPU: 1; MPIDR: 0x80000001
[0009.312] I> NVG: Logical CPU: 2; MPIDR: 0x80000100
[0009.317] I> NVG: Logical CPU: 3; MPIDR: 0x80000101
[0009.322] I> NVG: Logical CPU: 4; MPIDR: 0x80000200
[0009.327] I> NVG: Logical CPU: 5; MPIDR: 0x80000201
[0009.331] I> NVG: Logical CPU: 6; MPIDR: 0x80000300
[0009.336] I> NVG: Logical CPU: 7; MPIDR: 0x80000301
[0009.342] W> "misc-data" doesn't exist, creating
[0009.345] I> Boot-device: eMMC
[0009.348] I> Add boot-sdmmc to plugin-manager/misc-data
[0009.353] I> Add storage-sdmmc to plugin-manager/misc-data
[0009.359] W> Unknown storage device
[0009.362] I> Add serial number:0421219065425 as DT property
[0009.369] I> Plugin-manager override starting
[0009.372] I> node /plugin-manager/fragement-tegra-wdt-en matches
[0009.379] I> node /plugin-manager/fragement-soft-wdt matches
[0009.387] I> node /plugin-manager/fragment-pcie-c5-rp matches
[0009.391] I> node /plugin-manager/fragment-tegra-ufs-lane10 matches
[0009.403] I> Disable plugin-manager status in FDT
[0009.403] I> Plugin-manager override finished successfully
[0009.405] I> tegrabl_load_kernel_and_dtb: Done
[0009.409] E> tegrabl_display_clear: display is not initialized
[0009.414] W> Boot logo display failed...
[0009.418] I> Kernel EP: 0x80080000, DTB: 0x90000000
[    0.000000] Booting Linux on physical CPU 0x0
[    0.000000] Linux version 4.9.140-tegra (buildbrain@mobile-u64-3357) (gcc version 7.3.1 20180425 [linaro-7.3-2018.05 revision d29120a424ecfbc167ef90065c0eeb7f91977701] (Linaro GCC 7.3-2018.05) ) #1 SMP PREEMPT Thu Jun 25 21:22:12 PDT 2020
[    0.000000] Boot CPU: AArch64 Processor [4e0f0040]
[    0.000000] OF: fdt:memory scan node memory, reg size 48,
[    0.000000] OF: fdt: - 80000000 ,  2c000000
[    0.000000] OF: fdt: - ac200000 ,  44600000
[    0.000000] OF: fdt: - 100000000 ,  380000000
[    0.000000] earlycon: tegra_comb_uart0 at MMIO32 0x000000000c168000 (options '')
[    0.000000] bootconsole [tegra_comb_uart0] enabled
[    6.179201] cgroup: cgroup2: unknown option "nsdelegate"
[    7.321425] using random self ethernet address
[    7.321548] using random host ethernet address

通过内核日志，像是Linux_for_Tegra对应的SDK源码不匹配

4. 查看正常系统的JetPack

到Nano(p3448)和Xavier (p2888)是完全不同的文件夹。
需要删除Xavier one，让sdkmanager重新下载

~/nvidia/nvidia_sdk$ ls
DRIVE                              JetPack_4.2.1_Linux_GA_P3448       JetPack_4.2.3_Linux_GA_P3310  JetPack_4.3_Linux_P2888              JetPack_4.4_DP_Linux_DP_JETSON_NANO_DEVKIT
JetPack_3.3.3_Linux_GA_JETSON_TX2  JetPack_4.2.1_Linux_GA_P3489       JetPack_4.2_Linux_P2888       JetPack_4.3_Linux_P2888-0060         JetPack_4.4_DP_Linux_DP_JETSON_XAVIER_NX
JetPack_4.2.1_EA_Linux_P3310       JetPack_4.2.2_Linux_GA_P2180       JetPack_4.2_Linux_P3310       JetPack_4.3_Linux_P3310              JetPack_4.4_Linux_JETSON_TX2
JetPack_4.2.1_EA_Linux_P3489       JetPack_4.2.2_Linux_GA_P2888-0060  JetPack_4.2_Linux_P3448       JetPack_4.3_Linux_P3448              JetPack_4.4_Linux_JETSON_XAVIER_NX
JetPack_4.2.1_Linux_GA_P2888       JetPack_4.2.2_Linux_GA_P3489       JetPack_4.2_Linux_P3489       JetPack_4.3_Linux_P3448-0020
JetPack_4.2.1_Linux_GA_P3310       JetPack_4.2.2_Linux_GA_P3489-0080  JetPack_4.3_Linux_P2180       JetPack_4.4_DP_Linux_DP_JETSON_NANO

通过SDKmanager也需要选择对应的板子

5. 完整刷机流程

（1）Xavier 进入 Recovery 模式
（2）lsusb

（3）Xavier用C-Pin microUSB线连接到X86-Ubuntu 18.04桌面
（4）启动sdkmanager
（5）登录Nvidia开发者帐号后，按照SDK Manager GUI中的步骤下载安装Jetson OS image和Flash Jetson OS
（6）调过IP设置步骤
（7）等待下载安装即可退出