kern/55667: regression: XEN3_DOM0 fails to boot on

[kardel%netbsd.org@localhost]

>Number:         55667
>Category:       kern
>Synopsis:       regression: XEN3_DOM0 fails to boot on
>Confidential:   no
>Severity:       serious
>Priority:       high
>Responsible:    kern-bug-people
>State:          open
>Class:          sw-bug
>Submitter-Id:   net
>Arrival-Date:   Thu Sep 17 14:00:00 +0000 2020
>Originator:     kardel%netbsd.org@localhost
>Release:        NetBSD 9.99.55<->9.99.72
>Organization:
	
>Environment:
	
	
System: NetBSD Toblerone 9.99.55 NetBSD 9.99.55 (XEN3_DOM0) #1: Thu Apr 9 11:39:49 CEST 2020 kardel@Toblerone:/usr/src/sys/arch/amd64/compile/obj/XEN3_DOM0 amd64
Architecture: x86_64
Machine: amd64
>Description:
	XEN3_DOM0 used to boot flawlessly on this machine. Probably the
	MSI/MSIX interrupt rework introduced the regression.
	The boot stalls after attaching ums0 to wsmouse0. Basic
	timing seems to work as the IPMI driver logs the version
	16 seconds after boot and while the system seems hung.
	This is a regression as 9.99.55 used to work fine.

System info:
machdep.hypervisor = generic
machdep.idle-mechanism = xen
machdep.dmi.system-vendor = Supermicro
machdep.dmi.system-product = AS -2113S-WN24RT
machdep.dmi.system-version = 0123456789
machdep.dmi.bios-vendor = American Megatrends Inc.
machdep.dmi.bios-version = 2.0b
machdep.dmi.bios-date = 20191115
machdep.dmi.board-vendor = Supermicro
machdep.dmi.board-product = H11SSW-NT
machdep.dmi.board-version = 2.00
machdep.dmi.board-asset-tag = To be filled by O.E.M.
machdep.dmi.chassis-vendor = Supermicro
machdep.dmi.chassis-type = Supermicro
machdep.dmi.chassis-version = 0123456789
machdep.dmi.chassis-asset-tag = To be filled by O.E.M.
machdep.dmi.processor-vendor = Advanced Micro Devices, Inc.
machdep.dmi.processor-version = AMD EPYC 7302P 16-Core Processor               
machdep.dmi.processor-frequency = 3000 MHz
machdep.xen.version = 4.11.3nb1
machdep.xen.balloon.current = 4194304
machdep.xen.balloon.target = 4194304
machdep.xen.balloon.min = 2048
machdep.xen.balloon.max = 17179869180

	Working with jdolecek@ we have seen multiple scenarios (via patches
	to the nve driver) from the solid stall (right now) to boots with
	abysmally slow NVME access.

	Using xen 4.13 didn't change the boot stall.

	Other large Intel based systems run XEN3_DOM0 9.99.72 just fine.
	9.99.55 also runs on this system just fine.

>How-To-Repeat:
	Try to boot 9.99.72 XEN3_DOM0 on Supermicro AS -2113S-WN24RT. Watch it getting
	stuck.
>Fix:
	?

>Unformatted: