You are not logged in.

#26 2024-05-30 14:53:51

seth
Member
Registered: 2012-09-03
Posts: 54,433

Re: [Solved] MTU change on Killer E3100X causes hard reboot when its off

So it's the board or the NIC - in doubt the specific.
You could file a bug/quirk request at https://bugzilla.kernel.org/buglist.cgi … ct=Drivers and see whether there're others, justifying a general quirk.

Can you btw. prevent the reboot (and maybe get a backtrace intead) w/ https://wiki.archlinux.org/title/Kdump ?

Offline

#27 2024-05-30 16:31:17

Jeff_WuYo
Member
Registered: 2020-12-14
Posts: 58

Re: [Solved] MTU change on Killer E3100X causes hard reboot when its off

seth wrote:

So it's the board or the NIC - in doubt the specific.

I'm not sure. I can also trigger the bug on LMDE6 btw. Doesn't know how to get a live BSD for testing.

seth wrote:

maybe get a backtrace intead w/ Kdump?

I'll look into it, never used it before, might take me sometimes. Btw, the AUR package kdumpst mentioned in the wiki, it seems doesn't work with systemd-boot, which is what I use. How do I adapt it with systemd-boot, or just go with manual?
I know the info with Kdump is somewhat sensitive, but is there any info that could get me trouble, i.e. get attack on the Internet? The test OS I setup is close to bare minimum Archlinux install.

Last edited by Jeff_WuYo (2024-05-30 18:10:06)

Offline

#28 2024-05-30 20:50:10

seth
Member
Registered: 2012-09-03
Posts: 54,433

Re: [Solved] MTU change on Killer E3100X causes hard reboot when its off

https://nomadbsd.org/ - not sure whether the NIC is supported at all, though.

If you're manually setting up kdump nb. that you don't have to compile a kernel, the arch kernels have the necessaray configs enabled.
A backtrace of a kernel module is generally no privacy concern and there's no need nor reason in sharing the entire kernel memory dump.

There's also no guarantee that kdump will prevent a reboot (depends on what's actually triggering that, but an unprompted cold reboot is very much a hardware problem)

Offline

#29 2024-05-31 17:33:45

Jeff_WuYo
Member
Registered: 2020-12-14
Posts: 58

Re: [Solved] MTU change on Killer E3100X causes hard reboot when its off

seth wrote:

https://nomadbsd.org/ - not sure whether the NIC is supported at all, though.

Doesn't seem like it, but I finally got some info about device name. The photo I took during the BSD boot process, it says I226-K, and it failed to initialize. Under Linux, I didn't see "I226-K" show up in journal nor lspci. I probably should ask my motherboard vendor for support. I think I226-K is supported under linux igc?

As for kdump, I still struggle to get a dumpfile. If I trigger crash via sysrq, it seems loaded kernel and initramfs (Ctrl + Alt + Del works), but I couldn't get a output. It seems something about KMS driver which I have to unload or some sort. If I trigger the problem, it still behave the same, I couldn't control the system, neither sysrq nor Ctrl + Alt + Del. I've setup system service according to arch kdump wiki, but I didn't get any dumpfile.

Offline

#30 2024-06-13 21:23:09

seth
Member
Registered: 2012-09-03
Posts: 54,433

Re: [Solved] MTU change on Killer E3100X causes hard reboot when its off

It seems something about KMS driver which I have to unload or some sort

Hard to tell from "some sort", but either enable https://wiki.archlinux.org/title/NVIDIA … de_setting or try to use nouveau, though

If I trigger the problem, it still behave the same

that's moot if this doesn't trigger the kdump kernel sad

I think I226-K is supported under linux igc?

Your chip is

May 28 03:09:13 arch-z790-itx kernel: pci 0000:6e:00.0: [8086:3102] type 00 class 0x020000 PCIe Endpoint
May 28 03:09:13 arch-z790-itx kernel: pci 0000:6e:00.0: BAR 0 [mem 0x84a00000-0x84afffff]
May 28 03:09:13 arch-z790-itx kernel: pci 0000:6e:00.0: BAR 3 [mem 0x84b00000-0x84b03fff]
May 28 03:09:13 arch-z790-itx kernel: pci 0000:6e:00.0: ROM [mem 0x84900000-0x849fffff pref]
May 28 03:09:13 arch-z790-itx kernel: pci 0000:6e:00.0: PME# supported from D0 D3hot D3cold
May 28 03:09:13 arch-z790-itx kernel: pci 0000:6e:00.0: Adding to iommu group 24
May 28 03:09:13 arch-z790-itx kernel: igc 0000:6e:00.0: enabling device (0000 -> 0002)
May 28 03:09:13 arch-z790-itx kernel: igc 0000:6e:00.0: PTM enabled, 4ns granularity
May 28 03:09:13 arch-z790-itx kernel: igc 0000:6e:00.0 (unnamed net_device) (uninitialized): PHC added

Also https://cateee.net/lkddb/web-lkddb/IGC.html

https://forum.proxmox.com/threads/intel … ts.136908/ suggest this
1. might be a bug in the IGC module
2. can be worked around by disabling autoneg ?

Offline

#31 2024-06-18 11:07:36

Jeff_WuYo
Member
Registered: 2020-12-14
Posts: 58

Re: [Solved] MTU change on Killer E3100X causes hard reboot when its off

Haven't tested for a while. Never the less, I've test kernel 6.9.4-arch1-1 and 6.9.5-arch1-1, the problem was gone. I can set MTU properly even the NIC is down. I didn't test the lts kernel. The recent BIOS update doesn't seems to have any effect. I've test 6.9.2-arch1-1, the issue still there. Therefore, I'll mark this issue as solved. Thanks for yours help.

Offline

Board footer

Powered by FluxBB