10.0 crash during sysupgrade in MP

[Riccardo Mottola <riccardo.mottola%libero.it@localhost>]

HI,

my SS20 is running 10.0 and has 2x SuperSPARCII CPUs (since I switched
with SS10 with HyperSPARC if you follow my odisseys).
It appeared reasonably usable (while with HS 10.0 was unusable), I
compiled some stuff, used it...

I thought to upgrade from 10.1 and am running the MP kernel.

Sysupgrade though crashed this way:

sysupgrade: I: Extracting base into /
 92%    150 MiB  455.96 KiB/s    00:27 ETA
Watchdog Reset
Type  help  for more information
<#2> ok [ 9573.4970405] nmi_hard: SMP botch.
[ 9573.4970405] cpu0: NMI: system interrupts:
0x400c0000<VME=0,SBUS=0,SC,T,ME>
[ 9573.4970405] SX STATUS: 00005400
[ 9573.4970405] SX ERROR : 00000000
[ 9573.4970405] SX DIAG  : 00004804
[ 9573.4970405] module0:
[ 9573.4970405]         mxcc error 0x0
[ 9573.4970405]         mxcc status 0xff1410002
[ 9573.4970405]         mxcc reset 0x0
[ 9573.4970405] module1:
[ 9573.4970405]         mxcc error 0x0
[ 9573.4970405]         mxcc status 0xff1100000
[ 9573.4970405]         mxcc reset 0x4 (WATCHDOG RESET)


Not nice! I restarted with single CPU kernel and it completed. So I
suppose there is (or was? it is 10.0 kernel of course) stability in MP
even with the SuperSPARC processors.

    sh /usr/sbin/postinstall -s /var/cache/sysupgrade/etc.tgz -s
/var/cache/sysupgrade/xetc.tgz -d / fix mtree opeh
Note that this may overwrite local changes.