Re: Possible IDE corruption bug?

[Jason Thorpe <thorpej%me.com@localhost>]

On Oct 28, 2026, at 3:00 PM, Rhialto <rhialto%falu.nl@localhost> wrote:

I used to use a PWS 433a with NetBSD 2.0 and earlier, where the Pyxis
chip set was buggy for the particular cmd ata controller. It seems I
have scsi in the box too, at least that is what my kernel config
suggests.

Some Pass 1 Pyxis parts are basically unusable for DMA to the on-board peripherals.  Pass 2 is supposed to have fixed all of the problems, but it has other quirks.

My suspicion is that disabling DMA for the on-board CMD IDE controller is going to be the solution here.  Back in the NetBSD 2.0 days, we didn’t have super good ways to do this in a generic way, but we have better facilities for it now.

-- thorpej