On Mon, 14 Feb 2000, Ken Nakata wrote:

> Frederick Bruckman wrote:
> > 
> > On Mon, 14 Feb 2000, Jonathan R. Hinds wrote:
> > 
> > > dmesg says the follow about the controller and hard drive:
> > >
> > > wdc0 at isa0 port 0x1f0-0x1f7 irq 14
> > > wd0 at wdc0 channel 0 drive 0: <QUANTUM FIREBALL1280A>
> > > wd0: drive supports 8-sector pio transfers, lba addressing
> > > wd0: 1222MB, 2484 cyl, 16 head, 63 sec, 512 bytes/sect x 2503872 sectors
> > > wd0: drive supports PIO mode 4, DMA mode 2
> > >
> > > Does this mean that it is using DMA? I guess that I will try another hard
> > > drive and see what happens.
> > 
> > Guess so. I would try "de-rating" to DMA mode 1, in the kernel config,
> > to try and work around the bug. PIO mode 3, DMA mode 1 is plenty fast,
> > if it works.
> 
> Hmm, this line here from dmesg:
> 
> > > wdc0 at isa0 port 0x1f0-0x1f7 irq 14
> 
> It's got no dma channel assigned.  Only an irq channel.  Even so, do you
> still get DMA transfer?

Hmm. This is what I get, vlb 486dx4, NetBSD-1.4.1+:

wdc0 at isa0 port 0x1f0-0x1f7 irq 14
wd0 at wdc0 channel 0 drive 0: <ST3491A>
wd0: drive supports 16-sector pio transfers, lba addressing
wd0: 408MB, 899 cyl, 15 head, 62 sec, 512 bytes/sect x 836070 sectors
wd0: drive supports PIO mode 3, DMA mode 1

It's not using DMA? I'm not swapping or building kernels on the
drive--it's my /usr--but it doesn't seem slow. On another machine, a
pentium 233, I had to force it to PIO mode 0 to be able to use a
similarly old drive (to transfer files), and that one's dog slow.