In Message <E18ZQy1-00063I-00@mail.jdwhite.org> ,
   Tracy Di Marco White <netbsd@gendalia.org> wrote:

=>The ATA controller is 133, and it isn't recognized, so things are a 
=>ittle slower than I expected.
=>pciide0 at pci0 dev 9 function 0: Nvidia Corporation product 0x0065 (rev. 0xa2)
=>pciide0: bus-master DMA support present, but unused (no driver support)
=>pciide0: primary channel configured to compatibility mode
=>pciide0: primary channel interrupting at irq 14

   I ran into the same thing with my Leadtek nForce-2 board.
Changing the config to include flags:

pciide* at pci? dev ? function ? flags 0x0001

made it much happier:

pciide0 at pci0 dev 9 function 0: Nvidia Corporation product 0x0065 (rev. 0xa2)
pciide0: bus-master DMA support present, used without full driver support
pciide0: primary channel configured to compatibility mode
pciide0: primary channel interrupting at irq 14

Before this change, doing a CVS update would take up to 50% of the
CPU (as reported by "systat vm") in interrupts, and this with an
Athlon XP 2100+.  Afterwards, it was closer to 1%, and I have
noticed no stability problems.

					Gary Duzan