U2 panic: "!cpu_intr_p()" failed

Hi all,

I see this panic on my Ultra 2 on almost every boot:

  panic: kernel diagnostic assertion "!cpu_intr_p()" failed: file "/usr/cvs/2019-04-01/src/sys/kern/subr_xcall.c", line 436 high priority xcall for function 0x13afdc0

(see dmesg below).  When it doesn't panic here, it detects sd1 (which is the
same disk type as sd0) and continues.

If it doesn't panic at boot time, it will panic soon after when accessing
disks (most raidctl commands that read/write will cause it).  I've gone back
through older kernels and I haven't seen it before Feb 2019.  However, in
a 2019-02-01 kernel it only happens 1 boot in 10 or so.  Could it be timing
or locking related?  It looks the same as PR 54138 (although there is a
backtrace there).

Regards,

Julian

 - - - - - 8< - - - - - - - - - - - - - - - - - - - - - - - - - 8< - - - - - -

End of dmesg:

[   1.0118121] scsibus0: waiting 2 seconds for devices to settle...
[   1.0800171] scsibus1: waiting 2 seconds for devices to settle...
[   1.1543221] wskbd0 at kbd0 mux 1
[   1.2700191] IPsec: Initialized Security Association Processing.
[   3.1000191] sd0 at scsibus0 target 0 lun 0: <SEAGATE, ST318203LSUN18G, 034A> 
d
isk fixed
[   3.2000181] sd0: 17274 MB, 7508 cyl, 19 head, 248 sec, 512 bytes/sect x 35378
533 sectors
[   3.3000211] sd0: sync (100.00ns offset 15), 16-bit (20.000MB/s) transfers, ta
gged queueing
[   3.3976091] panic: kernel diagnostic assertion "!cpu_intr_p()" failed: file "
/usr/cvs/2019-04-01/src/sys/kern/subr_xcall.c", line 436 high priority xcall for
 function 0x13afdc0
[   3.5840211] cpu0: Begin traceback...
[   3.6288051] cpu0: End traceback...

 - - - - - 8< - - - - - - - - - - - - - - - - - - - - - - - - - 8< - - - - - -

From netbsd.gdb:

(gdb) l *0x13afdc0
0x13afdc0 is in nullop (/usr/cvs/2019-04-01/src/sys/kern/kern_stub.c:295).

References:
  https://nxr.netbsd.org/xref/src/sys/kern/subr_xcall.c#436
  https://nxr.netbsd.org/xref/src/sys/kern/kern_stub.c#291
  http://gnats.netbsd.org/54138