Jtop 4.2.3 crashes when navigating to the Jtop GPU tab on an Nvidia Xavier AGX running Jetpack 5.0.2 GA.

dmbuck32 commented 1 year ago

Describe the bug

To Reproduce

Steps to reproduce the behaviour:

Download Jtop: sudo -H pip install -U jetson-stats
Restart Jtop service: sudo systemctl restart jtop.service
Run Jtop: jtop
Navigate to the GPU tab: Press 2
Jtop Crashes with error message.

Screenshots

Traceback (most recent call last):
  File "/usr/local/bin/jtop", line 8, in <module>
    sys.exit(main())
  File "/usr/local/lib/python3.8/dist-packages/jtop/__main__.py", line 159, in main
    curses.wrapper(JTOPGUI, jetson, pages, init_page=args.page,
  File "/usr/lib/python3.8/curses/__init__.py", line 105, in wrapper
    return func(stdscr, *args, **kwds)
  File "/usr/local/lib/python3.8/dist-packages/jtop/gui/jtopgui.py", line 100, in __init__
    self.run(loop, seconds)
  File "/usr/local/lib/python3.8/dist-packages/jtop/gui/jtopgui.py", line 129, in run
    self.draw()
  File "/usr/local/lib/python3.8/dist-packages/jtop/gui/jtopgui.py", line 143, in draw
    page.draw(self.key, self.mouse)
  File "/usr/local/lib/python3.8/dist-packages/jtop/gui/pgpu.py", line 209, in draw
    self.process_table.draw(first + 2 + gpu_height, 0, width, height_table, key, mouse)
  File "/usr/local/lib/python3.8/dist-packages/jtop/gui/lib/process_table.py", line 85, in draw
    return nprocess
UnboundLocalError: local variable 'nprocess' referenced before assignment

Expected behavior

The GPU tab to display a graph with the list of active GPU processes.

Board

Output from jetson_release -v:

Software part of jetson-stats 4.2.3 - (c) 2023, Raffaello Bonghi
Model: Jetson-AGX - Jetpack 5.0.2 GA [L4T 35.1.0]
Serial Number: [XXX Show with: jetson_release -s XXX]
Hardware:
 - 699-level Part Number: 699-82888-0004-400 P.0
 - P-Number: p2888-0004
 - Module: NVIDIA Jetson AGX Xavier (32 GB ram)
 - SoC: tegra194
 - CUDA Arch BIN: 7.2
Platform:
 - Machine: aarch64
 - System: Linux
 - Distribution: Ubuntu 20.04 focal
 - Release: 5.10.104-tegra
 - Python: 3.8.10
jtop:
 - Version: 4.2.3
 - Service: Active
Libraries:
 - CUDA: 11.4.239
 - cuDNN: 8.4.1.50
 - TensorRT: 5.0.2
 - VPI: 2.1.6
 - OpenCV: Not installed

Log from jtop.service

Attach here the output from: journalctl -u jtop.service -n 100 --no-pager

-- Logs begin at Mon 2023-03-27 17:54:06 UTC, end at Thu 2023-08-17 05:10:05 UTC. --
Aug 17 04:35:44 ghrunner2 jtop[1253304]: [INFO] jtop.core.power - Alarms VDDRQ - {'crit_alarm': 0, 'max_alarm': 0}
Aug 17 04:35:44 ghrunner2 jtop[1253304]: [INFO] jtop.core.power - Alarms SYS5V - {'crit_alarm': 0, 'max_alarm': 0}
Aug 17 04:35:44 ghrunner2 jtop[1253304]: [WARNING] jtop.core.power - Skipped "sum of shunt voltages" /sys/bus/i2c/devices/1-0041/hwmon/hwmon5/in7_label
Aug 17 04:35:44 ghrunner2 jtop[1253304]: [INFO] jtop.core.power - Alarms GPU - {'crit_alarm': 0, 'max_alarm': 0}
Aug 17 04:35:44 ghrunner2 jtop[1253304]: [INFO] jtop.core.power - Alarms CPU - {'crit_alarm': 0, 'max_alarm': 0}
Aug 17 04:35:44 ghrunner2 jtop[1253304]: [INFO] jtop.core.power - Alarms SOC - {'crit_alarm': 0, 'max_alarm': 0}
Aug 17 04:35:44 ghrunner2 jtop[1253304]: [WARNING] jtop.core.power - Skipped "sum of shunt voltages" /sys/bus/i2c/devices/1-0040/hwmon/hwmon4/in7_label
Aug 17 04:35:44 ghrunner2 jtop[1253304]: [INFO] jtop.core.power - Found I2C power monitor
Aug 17 04:35:44 ghrunner2 jtop[1253304]: [WARNING] jtop.core.power - Skipped usb-charger type=USB in=usb-charger
Aug 17 04:35:44 ghrunner2 jtop[1253304]: [INFO] jtop.core.fan - Fan pwmfan(1) found in /sys/class/hwmon/hwmon3
Aug 17 04:35:44 ghrunner2 jtop[1253304]: [INFO] jtop.core.fan - RPM pwm_tach found in /sys/class/hwmon/hwmon1
Aug 17 04:35:44 ghrunner2 jtop[1253304]: [INFO] jtop.core.jetson_clocks - jetson_clocks found in /usr/bin/jetson_clocks
Aug 17 04:35:44 ghrunner2 jtop[1253304]: [WARNING] jtop.core.nvpmodel - nvpmodel not available
Aug 17 04:35:44 ghrunner2 jtop[1253325]: [INFO] jtop.service - Initialization service
Aug 17 04:35:45 ghrunner2 jtop[1253325]: [INFO] jtop.core.jetson_clocks - Store jetson_clocks configuration in /usr/local/jtop/l4t_dfs.conf
Aug 17 04:35:45 ghrunner2 jtop[1253325]: [INFO] jtop.service - service ready
Aug 17 04:35:45 ghrunner2 jtop[1253325]: [INFO] jtop.service - jtop timer thread started 1000ms
Aug 17 04:35:55 ghrunner2 jtop[1253325]: [INFO] jtop.service - jtop timer thread close

Log from jetson-stats installation

Attach here the output from: sudo -H pip3 install --no-cache-dir -U jetson-stats

Requirement already up-to-date: jetson-stats in /usr/local/lib/python3.8/dist-packages (4.2.3)
Requirement already satisfied, skipping upgrade: distro in /usr/lib/python3/dist-packages (from jetson-stats) (1.4.0)
Requirement already satisfied, skipping upgrade: smbus2 in /usr/local/lib/python3.8/dist-packages (from jetson-stats) (0.4.2)

n4s commented 1 year ago

This occurs when no GPU process is created, the for at line 72 of process_table.py fails and the nprocess variable has no value. Adding nprocess = 0 just before the for fixes it.

_(workaround: edit your file at /usr/local/lib/python3.8/dist-packages/jtop/gui/lib/processtable.py and add the nprocess = 0 before the for at line 72)

rbonghi commented 10 months ago

This issue is fixed with the latest release of jetson-stats 4.2.4. Please update it!

sudo pip3 install -U jetson-stats

rbonghi / jetson_stats