Hi,

I just noticed something about the interrupt counts shown by
systat that seemed a bit strange, at least to me.  An excerpt
from the "vmstat" display shows:

    3 users    Load  2.05  2.85  3.08                  Thu Oct 11 11:53:15

          memory totals (in KB)             PAGING   SWAPPING      Interrupts
         real   virtual    free             in  out   in  out      1099 total
Active 135016    135016   86672     ops                             497 lev1
All    165488    165488  611328     pages                             7 lev5
                                                                    495 lev6
Proc:r  d  s  w    Csw   Trp   Sys  Int  Sof   Flt      3 forks     100 clock
     1     8        45   861  1512 1099       1691      1 fkppw
                                                        1 fksvm
  24.8% Sy   0.0% Us  74.7% Ni   0.0% In   0.6% Id        pwait
|    |    |    |    |    |    |    |    |    |    |       relck
============--------------------------------------        rlkok
                                                          noram
Namei         Sys-cache     Proc-cache                 32 ndcpy
    Calls     hits    %     hits     %                 30 fltcp
     3732     3716  100                               667 zfod
                                                       13 cow
Discs  cd0  wd0  md0                                   32 fmin
seeks                                                  42 ftarg
xfers         7                                           itarg
Kbyte       127                                       488 wired
%busy       1.4                                           pdfre

Notice the almost equal numbers for level1 and level6 interrupt.

I kept an eye on the display for some time during a kernel build,
and the maximum difference between these two numbers was less
than 10.  I also noticed that the "wired count" in the vmstat
display was very close to the interrupt level1 and level6
interrupt counters.

What is the reason for this?  Is this a double counting error?
Double interrupt posting error (or feature)?  Bug in systat?  Or
is this just the way it's supposed to be?

(This is on a Ultra 5 running -current from two days ago).

					-jarle