BSD 50% interrupt utilization in irq0/clock

Question

On OpenBSD on an Atom 450, with HPET configured in the BIOS and not, also with Hyperthreading/ACPI on-off, nothing seems to make a difference.

Here's my vmstat -iz

# vmstat 2
 procs    memory       page                    disk traps          cpu
 r b w    avm     fre  flt  re  pi  po  fr  sr wd0  int   sys   cs us sy id
 0 0 0   3488  981792  429   0   0   0   0   0  35 295771   906   89  1 51 48
 0 0 0   3492  981788   12   0   0   0   0   0   0 278490    37   10  0 50 50
^C
# vmstat -iz
interrupt                       total     rate
irq0/clock                      12115      102
irq96/inteldrm0                     0        0
irq133/ppb0                         0        0
irq96/em0                           0        0
irq84/ppb1                          0        0
irq81/em1                           0        0
irq131/ppb2                         0        0
irq82/em2                           0        0
irq132/ppb3                         0        0
irq83/em3                           0        0
irq84/uhci0                        28        0
irq83/uhci1                         0        0
irq84/ehci0                         2        0
irq80/pciide0                     877        7
irq82/pciide1                       0        0
irq81/ichiic0                       0        0
irq130/com0                       451        3
irq129/pckbc0                       0        0
Total                           13473      114

The irq0/clock continues to go out of control, this is about 1minute after booting....

# vmstat -s
       4096 bytes per page
     252939 pages managed
     245456 pages free
        865 pages active
       1015 pages inactive
          0 pages being paged out
          1 pages wired
          0 pages zeroed
          4 pages reserved for pagedaemon
          6 pages reserved for kernel
     132529 swap pages
          0 swap pages in use
          0 total anon's in system
          0 free anon's
      17558 page faults
      17368 traps
   47204661 interrupts
       4703 cpu context switches
         14 fpu context switches
      17854 software interrupts
      37197 syscalls
          0 pagein operations
          0 swap ins
          0 swap outs
        144 forks
         13 forks where vmspace is shared
         14 kernel map entries
          0 number of times the pagedaemon woke up
          0 revolutions of the clock hand
          0 pages freed by pagedaemon
          0 pages scanned by pagedaemon
          0 pages reactivated by pagedaemon
          0 busy pages found by pagedaemon
       9096 total name lookups
            cache hits (67% pos + 1% neg) system 0% per-directory
            deletions 0%, falsehits 0%, toolong 0%
      0 select collisions

Any ideas to reduce the load, thanks.

Answer 1

This perticular problem was likely asperbated by the fact that OpenBSD's sysctl is anemic. FBSD "poll" sysctl would of likely solved this much earlier.

The problem was due to a poorly written BIOS, specifically it's configuration and implmentation of a console/remote access serial port. It was spinning like crazy and generating these interrupts, even if not connected etc...