Hi, nvidia team:
Our custom Thor board may stall occasionally when boot up.
here is the log:
开机卡死.log (113.0 KB)
BR//ChengKe
Hi, nvidia team:
Our custom Thor board may stall occasionally when boot up.
here is the log:
BR//ChengKe
Looking at your attached log. The hang appears to be in pwm_tegra_tachometer module registration/probe path, possibly associated with the tachometer@810c5c0000 DT node. I think the journald and systemd-modules-load failures are secondary symptoms to that.
# Is your pwm_tegra_tachometer loaded?
lsmod |grep pwm
pwm_tegra_tachometer 12288 0
pwm_fan 16384 0
pwm_tegra 12288 1
# If not, try loading
sudo modprobe -v pwm_tegra_tachometer
A Test you could do is temporarily blacklist pwm_tegra_tachometer to see if that skips the error. But, I think you will need pwm_tegra_tachometer for fans and cooling.
printf '%s\n' 'blacklist pwm_tegra_tachometer' | sudo tee /etc/modprobe.d/blacklist-pwm-tegra-tachometer.conf
sudo reboot
Jetson Module Adaptation and Bring-Up: Checklists
hi:nv team
我们想知道是什么原因导致的rt锁造成的堆栈,有没有patch修正,卡死之后,串口不能输入,系统也不重启,表现是开机不成功,目前复现了两次,我把另外一份log 也上传
log(1).txt (31.7 KB)
Check what compatible string your module expects:
modinfo obc_max96712 | grep alias
Then align your DT compatible property to match, or vice versa.
If max96712_c and max96712_d aren’t physically present on your custom carrier, disable those nodes:
&max96712_c {
status = "disabled";
};
&max96712_d {
status = "disabled";
};
This might be a bug in tegra_camera_rtcpu, it shouldn’t deadlock on device_link_drop_managed when camera devices fail to enumerate.
hi:nv team
Please analyze the log and determine which module caused the rt mutex issue.
Hi,
Please make sure you apply this to Jetpack 7.1:
Jetson AGX Thor FAQ
If the issue persists, please try to reproduce it on developer kit. We need to replicate it on developer kit first, and then investigate further.