At 21:00 Uhr +0100 9.11.2005, Hauke Fath wrote:
>What I noted, though, is
>
>Configuring network interfaces: wi0Before tsleep in 1131, file
>../../../../dev/s
>bus/stp4020.c: ticks = 40
>After tsleep in 1136, file ../../../../dev/sbus/stp4020.c: ticks = 40
>Before tsleep in 1131, file ../../../../dev/sbus/stp4020.c: ticks = 12
>After tsleep in 1136, file ../../../../dev/sbus/stp4020.c: ticks = 12
>Before tsleep in 1131, file ../../../../dev/sbus/stp4020.c: ticks = 2
>After tsleep in 1136, file ../../../../dev/sbus/stp4020.c: ticks = 2
>stray interrupt ipl 0x7 pc=0xf0007e84 npc=0xf0007e88 psr=404000c7<S,PS>
> hme0 bridge0stray interrupt ipl 0x7 pc=0xf0007e80 npc=0xf0007e84
>psr=404000c7<S
>,PS>
> pppoe0.
>
>in the bootlog which normally looks more like
>http://mail-index.netbsd.org/port-sparc/2005/06/09/0005.html. So there
>seems to be a timing-related issue.
>
>I shall try a single-cpu kernel next, just for laughs.

Bingo - it's an SMP issue:

NetBSD 3.0_BETA (SOLO) #0: Wed Nov  9 23:02:14 CET 2005
        hauke@pizza.causeuse.org:/usr/src/sys/arch/sparc/compile/SOLO
total memory = 319 MB
avail memory = 304 MB
bootpath: /iommu@f,e0000000/sbus@f,e0001000/espdma@f,400000/esp@f,800000/sd@0,0
mainbus0 (root): SUNW,SPARCstation-10: hostid 727021dd
cpu0 at mainbus0: mid 8: TMS390Z50 v0 or TMS390Z55 @ 75 MHz, on-chip FPU
cpu0: physical 20K instruction (64 b/l), 16K data (32 b/l), 1024K external
(32 b
/l): cache enabled
cpu at mainbus0 not configured

[...]

unmounting / (/dev/sd0a)... done
Before tsleep in 1131, file ../../../../dev/sbus/stp4020.c: ticks = 30
After tsleep in 1136, file ../../../../dev/sbus/stp4020.c: ticks = 30
halted

Program terminated
<#0> ok


And, while the DEBUG kernel on a single CPU is frighteningly slow, it has
none of that "stray interrupt" stuff from above. So that, also, seems SMP
related.

	hauke

--
"It's never straight up and down"     (DEVO)