On Wed, 28 Jan 2004 amoschos@in.gr wrote:

>Just purchased a SS10 and net-installed Netbsd 1.6.1.At the installation i used the
>GENERIC kernel (thus no SMP after the install).Once after the first boot i compiled a
>custom kernel enabling the MP and i got the common " ledma0 errors"!Started playing
>around with the conf but ended using the GENERIC.MP.
> Although it was reported that both CPUs (2 cpu modules exist) were initialized the
>output of the top command showed only once CPU0 (2 compiles in progress).

1.6.x only supports running on a single CPU, though other CPUs may be
found and initialised.

You need -current to run MP.

> The story ended like this:
>panic: nmi_hard: SMP botch.
>After this the machine was paniced and rebooted, panic and reboot , panic and reboot
>..

No idea what this is. Does it occur when you take the second CPU module
out? It may be a bug in the MP code.

As 1.6.x doesn't actually support MP beyond detection, you'd be just as
well using the GENERIC kernel.

>
>
>======================================================
>top
>load averages:  1.31,  1.30,  1.36
>23:05:28
>22 processes:  1 runnable, 20 sleeping, 1 on processor
>CPU states: 39.0% user,  0.0% nice,  8.9% system,  0.7% interrupt, 51.4% idle
>Memory: 45M Act, 4K Inact, 172K Wired, 1664K Exec, 39M File, 184M Free
>Swap:
>bash-2.05a# ME PRI NICE   SIZE   RES STATE      TIME   WCPU    CPU COMMAND
> 1801 root      61    0   532K  976K RUN/0      0:00 34.00%  1.66% cpp0
> 1800 root      10    0   136K  668K wait/0     0:00 11.00%  0.54% cc
>  483 root      10    0  1724K 2212K wait/0     0:06  0.05%  0.05% make
>  237 root       2    0   224K 1404K select/0   0:02  0.05%  0.05% telnetd
>  544 root      28    0   208K  920K CPU/0      0:27  0.00%  0.00% top
>    5 root      18    0     0K   29M syncer/0   0:21  0.00%  0.00% [ioflush]
>  193 root       2    0   224K 1380K select/0   0:20  0.00%  0.00% telnetd
>======================================================
>Info from the boot prom
>
>CPU_#0         TI, TMS390Z50(3.x) 0Mb External Cache
>CPU_#1         TI, TMS390Z50(3.x) 0Mb External Cache
>(I have noticed an article stating that modules with no external cache might have
>problems.Is it still an issue or it has been fixed?)
>======================================================
>At the continual reboots i was getting
>
>nmi_hard: SMP botch
>syncing disks... done
>Call traceback:
>pc = 0xf0258658 args = (0x0, 0x40400fe4, 0xf0002000, 0xf02ec400, 0xf02dacd8,
>0x0, 0xf02dac28) fp = 0xf02dac28
>pc = 0xf011d5e4 args = (0x100, 0x0, 0xfe012000, 0xf031c400, 0xf031c400,
>0xf02a2800, 0xf643c000, 0xf02dac90) fp = 0xf02dac90
>
>plus more.
>
>Any idea ?
>

"The problem with the French is that they don't have a word for entrepreneur"
  George W. Bush

----------------------------------------------------------------------
    /"\
    \ /    ASCII RIBBON CAMPAIGN - AGAINST HTML MAIL
     X                           - AGAINST MS ATTACHMENTS
    / \