Unexpected realtime delay on task 0 on a HP8300

More
08 Mar 2024 03:22 #295436 by lrak
This started after updating to debian bookworm(12) I started getting these errors - can take a week or two of uptime.

First question - it task 0 the Linux task number or something from linuxCNC?

I've seen conflicting advise about kernel command line settings:

PCW wrote:
"Unless you have an actual latency issue I would not use any isolcpus
options at all"

vs linuxcnc.org/docs/html/install/latency-test.html

Not sure I really have a latency issue - could be some cron-job of systemd-timer job is getting fired off.

Right now I've turned off a couple of apt-timer jobs in systemd - need to wait for an error.
 

Please Log in or Create an account to join the conversation.

More
08 Mar 2024 04:09 #295437 by PCW
1. What is you servo thread period?
2. Do you have AMT disabled in the BIOS?
3. Do you have IRQ coalescing disabled? (man hm2_eth)
 
The following user(s) said Thank You: lrak

Please Log in or Create an account to join the conversation.

More
08 Mar 2024 06:12 - 08 Mar 2024 06:14 #295439 by lrak
@PCW

1-  in ~/linuxcnc/configs/machine.ini  I have
SERVO_PERIOD = 1000000

2 - grub command line:
GRUB_CMDLINE_LINUX_DEFAULT="quiet text isolcpus=2,3 apparmor=0 apm=off"

3 - I have in /etc/network/interfaces a line commented out
# hardware−irq−coalesce−rx−usecs 0

What ever is happening is rare - never happened while the machine was cutting. 
Last edit: 08 Mar 2024 06:14 by lrak.

Please Log in or Create an account to join the conversation.

More
08 Mar 2024 12:37 #295450 by PCW
# hardware−irq−coalesce−rx−usecs 0
 
This should not be commented out

AMT is something that must be disabled in the BIOS,
not something that can be changed in the kernel/grub
The following user(s) said Thank You: lrak

Please Log in or Create an account to join the conversation.

More
08 Mar 2024 14:33 #295459 by tommylight
Also while in BIOS, disable hyperthreading.
The following user(s) said Thank You: lrak

Please Log in or Create an account to join the conversation.

More
08 Mar 2024 21:57 - 08 Mar 2024 22:48 #295479 by lrak
I left things running overnight - got the same error - no time stamp<grump>.  (Is there a way to get a timestamp?).

Who or what is task-0 in the context of this error?

Trying enabling  hardware−irq−coalesce−rx−usecs 0

/etc/network/interfaces


# The loopback network interface
auto lo
iface lo inet loopback

# The primary network interface
auto eno1
allow-hotplug eno1
iface eno1 inet dhcp

#linuxCNC mesa interface
auto enp2s0
iface enp2s0 inet static
    address 192.168.10.1/24
    hardware−irq−coalesce−rx−usecs 0

,.,.
The mesa interface has a r8168 driver..

root@cnc:~# lspci |grep Ether
00:19.0 Ethernet controller: Intel Corporation 82579LM Gigabit Network Connection (Lewisville) (rev 04)
02:00.0 Ethernet controller: Realtek Semiconductor Co., Ltd. RTL8111/8168/8411 PCI Express Gigabit Ethernet Controller (rev 07)





,.,.
BIOS settings:

I do have AMT off
  I also have all the power saving stuff turned off.

I do have:
Turbo mode - enabled
multi-processor - enabled
,.,.

I found this in dmesg - I was thinking that the stock r8168 works now with bookworm? Do I still need to build a special one?

r8168: loading out-of-tree module taints kernel.
[    1.396296] r8168: module verification failed: signature and/or required key missing - tainting kernel
[    1.419277] ACPI Warning: SystemIO range 0x0000000000000428-0x000000000000042F conflicts with OpRegion 0x0000000000000400-0x000000000000047F (\PMIO) (20220331/utaddress-204)
[    1.419282] ACPI: OSL: Resource conflict; ACPI support missing from driver?
[    1.419285] ACPI Warning: SystemIO range 0x0000000000000540-0x000000000000054F conflicts with OpRegion 0x0000000000000500-0x0000000000000563 (\GPIO) (20220331/utaddress-204)
[    1.419287] ACPI: OSL: Resource conflict; ACPI support missing from driver?
[    1.419287] ACPI Warning: SystemIO range 0x0000000000000530-0x000000000000053F conflicts with OpRegion 0x0000000000000500-0x0000000000000563 (\GPIO) (20220331/utaddress-204)
[    1.419289] ACPI: OSL: Resource conflict; ACPI support missing from driver?
[    1.419290] ACPI Warning: SystemIO range 0x0000000000000500-0x000000000000052F conflicts with OpRegion 0x0000000000000500-0x0000000000000563 (\GPIO) (20220331/utaddress-204)
[    1.419292] ACPI: OSL: Resource conflict; ACPI support missing from driver?
[    1.419292] lpc_ich: Resource conflict(s) found affecting gpio_ich

,.,.
Latency test is showing after 10-min:
max servo jitter of 12104ns
max base thread of 7557ns

,.,.
A bit OT:
Just checked - coreboot does not support the HP8300..
There was a industrial intel board that we used for digital-video back in the '90s that I know didn't have any hiccups other than thermal throttling of the
MMU - which could be cured with the right heat-sink.  Wish I still had one..  
Such throttling still happens - some is independent of the management engines in the BIOSes  ..  Removing the thermal pads that go bad after a couple of years and replacing with good and proper heat-sink compound solves most of those problems.  Video apps put a LOT of load on the MMUs - had to add the right heat-sink and fan. 


 
Last edit: 08 Mar 2024 22:48 by lrak.

Please Log in or Create an account to join the conversation.

More
08 Mar 2024 22:03 #295481 by PCW
Do you have an add-in Ethernet card?
because the built in Ethernet MAC on
a HP8300 is an Intel device not a Realtek
(at least the ones I have all have Intel MACs)
The following user(s) said Thank You: lrak

Please Log in or Create an account to join the conversation.

More
08 Mar 2024 22:50 #295482 by lrak
@PCW
Was editing my comment while you replied.

Please Log in or Create an account to join the conversation.

More
08 Mar 2024 23:49 #295483 by PCW
Turbo mode should be off

Do you have an add-on network card?
The following user(s) said Thank You: lrak

Please Log in or Create an account to join the conversation.

More
10 Mar 2024 05:46 #295558 by lrak
@PCW who wrote "Do you have an add-on network card?"

Yes - The mesa interface is a realtek RTL8111/8168/8411 running via r8168 driver,

The onboard intel runs the LAN .

No error for 24hrs - restarting with turbo off.
 

Please Log in or Create an account to join the conversation.

Time to create page: 0.084 seconds
Powered by Kunena Forum