port-amd64/60313: interrupt storm on pciide on HP DL380p Gen8

["he%NetBSD.org@localhost via gnats" <gnats-admin%NetBSD.org@localhost>]

>Number:         60313
>Category:       port-amd64
>Synopsis:       interrupt storm on pciide on HP DL380p Gen8
>Confidential:   no
>Severity:       serious
>Priority:       medium
>Responsible:    port-amd64-maintainer
>State:          open
>Class:          support
>Submitter-Id:   net
>Arrival-Date:   Sat Jun 06 22:35:00 +0000 2026
>Originator:     Havard Eidnes
>Release:        NetBSD 11.0_RC4
>Organization:
	I try...
>Environment:
System: NetBSD stest.urc.uninett.no 11.0_RC4 NetBSD 11.0_RC4 (GENERIC) #0: Thu Jun  4 12:18:47 UTC 2026  mkrepro%mkrepro.NetBSD.org@localhost:/usr/src/sys/arch/amd64/compile/GENERIC amd64
Architecture: x86_64
Machine: amd64
>Description:
	I noticed "Intr *****" in "systat vm" view, and
	decided to see where that comes from, and it appears
	to come from here:

# vmstat -i | egrep 'interrupt|pin 17'
interrupt                    total   rate
ioapic0 pin 17          1331375677 338772
# dmesg | egrep 'pin 17'
[     1.089571] pciide0: using ioapic0 pin 17 for native-PCI interrupt
# dmesg | egrep 'pciide|atabus|atapi'
[     1.089571] pciide0 at pci0 dev 31 function 2: Intel C600/X79 SATA (rev. 0x05)
[     1.089571] pciide0: bus-master DMA support present, but unused (no driver support)
[     1.089571] pciide0: primary channel configured to native-PCI mode
[     1.089571] pciide0: using ioapic0 pin 17 for native-PCI interrupt
[     1.089571] atabus0 at pciide0 channel 0
[     1.089571] pciide0: secondary channel configured to native-PCI mode
[     1.089571] atabus1 at pciide0 channel 1
[    26.622404] atapibus0 at atabus1: 2 targets
[    26.682404] cd0 at atapibus0 drive 0: <hp      DVD A  DS8A8SH, 627225049827, KHE4> cdrom removable
# 

	Note, the CD drive is not in use at the moment.

	I cannot claim that this is a "new" problem.

	Attempt with "drvctl -d atapibus0" was unsuccessful in
	stemming the flow of interrupts, and "drvctl -d pciide0"
	was not met with success (possible kernel crash / DDB with
	no serial console, so now it's wedged...)

>How-To-Repeat:
	See above.
>Fix:
	Would be nice.