Scheduled Maintenance: We are aware of an issue with Google, AOL, and Yahoo services as email providers which are blocking new registrations. We are trying to fix the issue and we have several internal and external support tickets in process to resolve the issue. Please see: viewtopic.php?t=158230

 

 

 

gpe 6E interrupts with 4.18 kernel on coffee lake laptop

Linux Kernel, Network, and Services configuration.
Post Reply
Message
Author
sofarsogood
Posts: 11
Joined: 2017-01-14 12:04

gpe 6E interrupts with 4.18 kernel on coffee lake laptop

#1 Post by sofarsogood »

Hi

I'm setting up a Coffee Lake laptop (Dell Precision 7530, i7-8750K 6 core, 16GB DDR4 RAM, Intel UHD 630, NVidia Quadro P1000, 512GB SSD) with Debian Buster. (4.18 kernel, minimal Xfce). TLP is installed and powertop reports all tunables as good except "VM writeback timeout". vgaswitcheroo shows the DIS graphics to be DynOff (nouveau driver, I don't need the NVidia card for the work I do).

With the system idle (just after boot, no network up, Xwindows not started, bluetooth rfkill'ed), I'm noticing 4-5 [9]acpi interrupts per second in powertop:

Code: Select all

*  *  *   Top 10 Power Consumers   *  *  *

Usage;Events/s;Category;Description;PW Estimate
100.0%;;Device;Radio device: dell-laptop;  4.58 W  
  0.0 pkts/s;;Device;Network interface: wlp111s0 (iwlwifi);  4.54 W  
 15.0%;;Device;Display backlight;  1.39 W  
  0.0%;  8.3;Timer;tick_sched_timer; 15.8 mW  
  0.0%;  4.9;kWork;fb_flashcursor; 9.35 mW  
  0.3%;  4.2;Interrupt;[9] acpi; 8.11 mW  
  0.0%;  1.5;Interrupt;[7] sched(softirq); 2.96 mW  
  0.0%;  0.8;Process;[rcu_sched]; 1.62 mW  
  0.0%;  0.7;kWork;acpi_ec_event_processor; 1.34 mW  
  0.0%;  0.7;kWork;acpi_ec_event_handler; 1.34 mW  
Leaving the system and looking in top shows that a kworker is the highest CPU user on the system:

Code: Select all

  PID USER      PR  NI    RES S  %CPU  %MEM     TIME+ COMMAND
  570 root      20            I               0:08.87 kworker/1:2-kec_query
    1 root      20       8292 S         0.1   0:01.56 systemd
  459 systemd+  20       5612 S         0.0   0:01.00 systemd-timesyn
   85 root      20            I               0:00.48 kworker/0:2-events
  107 root      20            I               0:00.41 kworker/6:1-mm_percpu_wq
  606 root      20       3700 S         0.0   0:00.38 irqbalance
  233 root      20            I               0:00.34 kworker/11:2-events
  333 root      20       4504 S         0.0   0:00.32 systemd-udevd
  763 root      20            I               0:00.26 kworker/5:2-rcu_gp
   84 root      20            I               0:00.25 kworker/0:1-kacpi_notify
  322 root      20       6848 S         0.0   0:00.22 systemd-journal
   73 root      20            I               0:00.18 kworker/10:0-events_power_efficient
  108 root      20            I               0:00.15 kworker/7:1-events
  604 root      20       5920 S         0.0   0:00.13 systemd-logind
  767 root      20            I               0:00.12 kworker/5:3-events_freezable
   10 root      20            I               0:00.11 rcu_sched
   98 root      20            I               0:00.11 kworker/4:1-mm_percpu_wq
   31 root      20            I               0:00.06 kworker/3:0-mm_percpu_wq
   92 root      39  19        S               0:00.06 khugepaged
  105 root      20            I               0:00.04 kworker/2:1-events_power_efficient
  615 root      20       3368 S         0.0   0:00.03 login
  864 adminu    20       3568 S         0.0   0:00.03 bash
   67 root      20            I               0:00.02 kworker/9:0-mm_percpu_wq
  191 root      20            I               0:00.02 kworker/u24:2-events_unbound
  607 message+  20       3604 S         0.0   0:00.01 dbus-daemon
 1013 adminu    20       3724 R         0.0   0:00.01 top
i've also tried disabling gpe6E interrupts. That stops the kworker using CPU and seems to have no detrimental effect on the laptop that I have been able to find so I could call it quits there but I'm curious to know what is going on.

I took a look using perf and only seemed to capture a bunch of generic routine names being called by the kworker that didn't really help identify what might be causing the interrupts.

Does anyone have any suggestions for how I might be able to troubleshoot this further ?

Thanks all

Post Reply