Calling all DS10 / DS10L owners!

Jason Thorpe <[email protected]> Mon, 16 Mar 2026 20:06:16 -0700
Newsgroups gmane.os.netbsd.ports.alpha
Message-ID <[email protected]>
Folks =E2=80=94

I recently nabbed a cheap slightly-beat-up-but-working DS10L because I =
want to get to the bottom of =E2=80=9Cwhy does IDE DMA not work properly =
on this machine?=E2=80=9D.

(See alpha/alpha/dec_6600.c rev 1.38 were I disabled IDE DMA for the =
DS10-type machines based on reports of =E2=80=9Cno va=E2=80=9D.)

I installed a 500G OWC SATA SSD in it with a PATA <-> SATA adaptron.

I spent some time digging only to determine that there is a large gap in =
the knowledge of WTF is going on with the rev 0xC1 ALi M5229 that=E2=80=99=
s installed in the machine.  I mean, I did learn that these IDE =
controllers are kind of terrible, to be sure, but not much enlightenment =
about the specific problem observed.

I did, however, discover two things:

- Yes, PIO mode really sucks, especially because the drive I installed =
can=E2=80=99t do multi-sector PIO.  SO MANY INTERRUPTS.

- Hey, even though it defaults to UDMA mode 2 and expels its lunch onto =
the floor, if I force it to UDMA mode 1, it seems to work ok.

I ran a bunch of file system data integrity tests on it overnight and =
all day, and everything seems .. fine?  So I=E2=80=99d like to check in =
a change to aceride.c that forces it down to UDMA 1 for rev 0xC1 chips, =
and re-enables IDE DMA at the DS10 platform level.

But I would love to have more people give it a spin.

If you have a DS10 and are willing to give this a whirl, hit me up =
off-list.

Merci d=E2=80=99avance!

-- thorpej