[Jetson][Nano] Xorg task blocked when suspend/resume stress test

we hit Xorg task blocked issue when suspend/resume stress test,
it looks like tegra DC/HDMI driver’s problem in backtrace,
when hit this issue, HDMI output show blank screen with backlight,
dmesg as attachment,

dmesg (94.2 KB)

Hi NV,
I found a very similar issue as follow link,

it happens when enable DC/HDMI,

  1. Are you on devkit or your custom carrier board?

  2. Did you remember what state is the machine when below error gets printed?

[191727.521991] INFO: task Xorg:7432 blocked for more than 120 seconds.
[191727.522008] Not tainted 4.9.253-tegra #1
[191727.522016] “echo 0 > /proc/sys/kernel/hung_task_timeout_secs” disables this message.
[191727.522040] Xorg D 0 7432 7430 0x00000008
[191727.522048] Call trace:
[191727.522090] [] __switch_to+0x9c/0xc0
[191727.522112] [] __schedule+0x238/0x7d0
[191727.522125] [] schedule+0x40/0xa8
[191727.522139] [] schedule_preempt_disabled+0x28/0x40
[191727.522156] [] __mutex_lock_slowpath+0xcc/0x190
[191727.522168] [] mutex_lock+0x4c/0x60
[191727.522190] [] tegra_dc_sync_windows+0x1dc/0x220
[191727.522207] [] tegra_dc_update_winlut+0x104/0x2e8
[191727.522221] [] tegra_dc_update_lut+0x38/0xd8
[191727.522240] [] tegra_dc_ext_set_lut+0x10c/0x228
[191727.522253] [] tegra_dc_ioctl+0xbb4/0x1df0
[191727.522272] [] do_vfs_ioctl+0xb0/0x8d8
[191727.522284] [] SyS_ioctl+0x8c/0xa8
[191727.522298] [] el0_svc_naked+0x34/0x38
[191727.522565] INFO: task kworker/1:14:32080 blocked for more than 120 seconds.
[191727.522576] Not tainted 4.9.253-tegra #1
[191727.522584] “echo 0 > /proc/sys/kernel/hung_task_timeout_secs” disables this message.
[191727.522607] kworker/1:14 D 0 32080 2 0x00000000
[191727.522670] Workqueue: events tegra_hdmi_hpd_worker

Hi Wayne,
1.we tested on customer carrier board
2.I’m not sure “what state is the machine” do you mean? you mean HDMI output is normal or abnormal?

  1. Are you able to put your test over devkit? And is it rel-32.6?

  2. I mean how was the jetson nano trying to do when this error happened? It was trying to go to suspend mode? I see the last line was as below and it took 200 sec to print out error.

[191520.658309] PM: Preparing system for sleep (mem)

Hi Wayne,
1.we can try to reproduce, rel-32.6 is image version, right?
2.yes, system was going to suspend when hit the issue, no idea why stuck 200 sec then printed out error

Yes, using rel-32.6 image. Please also estimate how many times of suspend/resume are required to reproduce this issue.

Please also add some print inside kernel/nvidia/drivers/video/tegra/dc/window.c → int tegra_dc_sync_windows and see if this is due to ___wait_event goes timeout.

This topic was automatically closed 14 days after the last reply. New replies are no longer allowed.