Re: Current kernel with http://gnats.netbsd.org/39965 and pmap patches

On Sat, Dec 20, 2008 at 6:49 AM, Izumi Tsutsui <[email protected]> wrote:
>> > The drives are probed properly (with interrupts)
>> > but interrupts are lost on read ops?
>>
>> Yes probing works fine with interrupts, but read ops are lost.
>
> By the way, which interrupt (level and vector) is used
> for the IDE interface on Falcon?

I don't know :/ Maybe we can find more information from some Atari documents.

> There is no intr_establish() function in wdc_mb.c
> and the wdcintr() handler seems invoked via dma function,
> but obviously wdc_mb doesn't use DMA xfers.
> (no necessary hooks in attachment, cf. macppc/dev/wdc_obio.c)
>
> If wdc_mb.c emulates interrupt polling via other dma device,
> I guess ATAC_CAP_NOIRQ is mandatory and claim_hw/free_hw hooks
> should be removed.
> (or we should use a real interrupt line to handle wdcintr()
>  if it's known)

I used NetBSD 3 based kernel that you posted on port-atari:
http://mail-index.netbsd.org/port-atari/2008/07/21/msg000038.html

IDE worked with that kernel (at least I was able to install base.tgz, etc.tgz
and comp.tgz sets from IDE drive), but NetBSD 3 kernel without patches panicked
with IDE based errors.

I can try wdc_mb.c with ATAC_CAP_NOIRQ enabled and claim_hw/free_hw hooks
removed.

-Tuomo