[Sunhelp] Look ma, thems be interrupts.

[Dale Ghent]

Perhaps someone from Sun who's lurking on this list can provide some
insight about a curious thing I observed.

I was setting up mirroring for my system disk tonight, and attached a
submirror to the mirror plex. While disksuite was doing it's resyncing, I
watched all the I/O activity commence. I ran a mpstat, and saw that one of
the CPUs in this dual-cpu Ultra2 was getting all the interrupts,
presumably all the ones generated by the fas driver:

CPU minf mjf xcal  intr ithr csw icsw migr smtx srw syscl usr sys wt idl
  0    0   0    0   675  575 123    0    1    0   0     8   0   3 97   0
  1    0   0    0     1    1 119    0    0    0   0     4   0   1  3  96
CPU minf mjf xcal  intr ithr csw icsw migr smtx srw syscl usr sys wt idl
  0    0   0    0   659  559 118   0     0    1   0     8   0   7 93   0
  1    0   0    0     1    1 116   0     1    0   0     8   0   0  1  99

I have always thought that it was more normal for the interrupts to be
distributed across all the CPUs on a mp system. Is this not the case for
some devices/drivers?

/dale