Re: [BUG] Linux kernel crash in amd_smn_init during PVH Dom0 boot on AMD EPYC

[Arthur Borsboom <arthurborsboom@xxxxxxxxx>]

> Well, a pretty natural question: What amount of research have you done 
> yourself?

Lovely response. You really know how to motivate people and welcome
them into a community.

It has taken me about 1,5 day of work to find out why the screen
remained black, to determine where the problem was, how to retrieve
potential info by a serial-over-LAN console, subscribe to the mailing
list, and create this bug report to help.

But don't worry, I won't do any debugging anymore for the Xen project,
you have made sure of that. No need to respond. I will unsubscribe
myself from the mailing list after this mail.

Thanks for the warm welcome.
Enjoy your day

On Wed, 19 Aug 2026 at 08:50, Jan Beulich <jbeulich@xxxxxxxx> wrote:
>
> On 18.08.2026 23:47, Arthur Borsboom wrote:
> > Hi,
> >
> > I am encountering an early Linux kernel crash when booting a PVH
> > Dom0 on an AMD EPYC system (Family 19h Model 97). PV boot on the
> > same system works fine. The crash occurs during amd_smn_init due
> > to a divide-by-zero exception.
> >
> > There is a potentially related bug report discussing a Linux patch
> > addressing AMD SMN initialization under Xen PVH:
> >
> > https://patchew.org/linux/20260623211904.3674-1-jason.andryuk@xxxxxxx/
> >
> > It seems there was an ongoing discussion between Linux devs who
> > seem to have the need to align with Xen devs for a proper/future-proof fix.
> > I am submitting this report to share raw hardware trace data and
> > help move this forward.
> >
> > --- System Details ---
> > CPU: AMD EPYC 4344P 8-Core Processor (Family 25 / 0x19, Model 97
> > / 0x61, Stepping 2)
> > Motherboard: MSI MSIS366/S3661 (BIOS ES366AOC.10P 05/29/2025)
> > Xen Version: 4.20.0 (Command line: dom0_mem=8192M,max:8192M dom0=pvh
> > loglvl=all guest_loglvl=all sync_console noreboot console=com1,vga)
> > Linux Kernel: 7.1.5-arch1-2
> >
> > --- Kernel Crash Log ---
> > [...]
> > [    0.817348] Oops: divide error: 0000 [#1] SMP NOPTI
> > [    0.817348] fbcon: Taking over console
> > [    0.823766] CPU: 3 UID: 0 PID: 1 Comm: swapper/0 Not tainted
> > 7.1.5-arch1-2 #1 PREEMPT(full)
> > 33708094282b4038de0f4f0d201b68f6a0b5d9bf
> > [    0.839766] Hardware name: MSI MSIS366/S3661, BIOS ES366AOC.10P 
> > 05/29/2025
> > [    0.847766] RIP: 0010:amd_smn_init+0x1ed/0x280
> > [    0.852766] Code: d2 45 31 f6 45 31 ed 66 41 f7 f4 89 c5 48 89 df
> > e8 c8 c7 1b fd 48 89 c3 48 85 c0 0f 84 48 ff ff ff 44 89 e8 31 d2 45
> > 8d 7d 01 <66> f7 f5 66 85 d2 75 42 66 90 eb 2c 41 0f b7 ce 48 8d b3 d0
> > 00 00
> > [    0.870766] RSP: 0018:ffffc9000004bdc8 EFLAGS: 00010246
> > [    0.878766] RAX: 0000000000000000 RBX: ffff888102661000 RCX: 
> > 0000000000000007
> > [    0.887766] RDX: 0000000000000000 RSI: ffff888102661000 RDI: 
> > 0000000000000000
> > [    0.895766] RBP: 0000000000000000 R08: 0000000000000282 R09: 
> > ffff888100f57680
> > [    0.903766] R10: ffffea0004044580 R11: ffff888100042700 R12: 
> > 0000000000000002
> > [    0.911766] R13: 0000000000000000 R14: 0000000000000000 R15: 
> > 0000000000000001
> > [    0.919766] FS:  0000000000000000(0000) GS:ffff8882b2010000(0000)
> > knlGS:0000000000000000
> > [    0.927766] CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
> > [    0.935766] CR2: 0000000000000000 CR3: 0000000003822000 CR4: 
> > 0000000000750ef0
> > [    0.943766] PKRU: 55555554
> > [    0.943766] Call Trace:
> > [    0.948766]  <TASK>
> > [    0.951766]  ? __pfx_amd_smn_init+0x10/0x10
> > [    0.956766]  do_one_initcall+0x8e/0x3f0
> > [    0.959766]  kernel_init_freeable+0x24a/0x2d0
> > [    0.965766]  ? __pfx_kernel_init+0x10/0x10
> > [    0.967766]  kernel_init+0x1a/0x140
> > [    0.973766]  ret_from_fork+0x2a7/0x330
> > [    0.974766]  ? __pfx_kernel_init+0x10/0x10
> > [    0.982766]  ret_from_fork_asm+0x1a/0x30
> > [    0.988766]  </TASK>
> > [    0.989766] Modules linked in:
> > [    0.994772] ---[ end trace 0000000000000000 ]---
> > [    0.999769] RIP: 0010:amd_smn_init+0x1ed/0x280
> > [    1.004770] Code: d2 45 31 f6 45 31 ed 66 41 f7 f4 89 c5 48 89 df
> > e8 c8 c7 1b fd 48 89 c3 48 85 c0 0f 84 48 ff ff ff 44 89 e8 31 d2 45
> > 8d 7d 01 <66> f7 f5 66 85 d2 75 42 66 90 eb 2c 41 0f b7 ce 48 8d b3 d0
> > 00 00
> > [    1.004771] RSP: 0018:ffffc9000004bdc8 EFLAGS: 00010246
> > [    1.004773] RAX: 0000000000000000 RBX: ffff888102661000 RCX: 
> > 0000000000000007
> > [    1.004774] RDX: 0000000000000000 RSI: ffff888102661000 RDI: 
> > 0000000000000000
> > [    1.004775] RBP: 0000000000000000 R08: 0000000000000282 R09: 
> > ffff888100f57680
> > [    1.004776] R10: ffffea0004044580 R11: ffff888100042700 R12: 
> > 0000000000000002
> > [    1.004777] R13: 0000000000000000 R14: 0000000000000000 R15: 
> > 0000000000000001
> > [    1.004778] FS:  0000000000000000(0000) GS:ffff8882b2090000(0000)
> > knlGS:0000000000000000
> > [    1.004779] CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
> > [    1.004780] CR2: 0000000000000000 CR3: 0000000003822000 CR4: 
> > 0000000000750ef0
> > [    1.004782] PKRU: 55555554
> > [    1.004784] Kernel panic - not syncing: Attempted to kill init!
> > exitcode=0x0000000b
> > (XEN) Hardware Dom0 crashed: 'noreboot' set - not rebooting.
> >
> > Please let me know if any additional info would be helpful.
>
> Well, a pretty natural question: What amount of research have you done 
> yourself?
> You really should have found [1], or maybe its v1 counterpart, or the earlier
> report(s) of what I think is this same issue.
>
> Jan
>
> [1] 
> https://patchew.org/linux/20260814214255.83127-1-jason.andryuk@xxxxxxx/20260814214255.83127-2-jason.andryuk@xxxxxxx/
>