Re: Re: Atari machines running Debian?
> Michael Schmitz pí?e v So 02. 09. 2006 v 16:53 +0200:
> > The boot proceeds happily to complain about a nonexisting root device, so
> > porting the aranym natfeats to 2.6 would be next. Anything to watch out
> > for in that area, Petr?
>
> Michael,
>
> please remember that you don't need the ARAnyM patches for running Atari
> kernel on ARAnyM. They bring you the networking (which isn't going to
> fix the root device, I am afraid), proper identification
> in /proc/hardware (not useful but nice) and kernel halt that quits
> ARAnyM (just a nice bonus).
I shoehorned the natfeat patch into 2.6.17 but the ethernet device isn't
recognized. I didn't really adapt the driver probe function (the return
type changed, for one) so it's not exactly expected to work. I'll happily
test Roman's version of the natfeats patch
> So my guess is that for existing root device you need either Falcon IDE
> driver or something else entirely. Not sure what.
The falcon IDE driver is compiled in, and the hwif probe function is
properly executed. The IDE core driver should pick up from there, and
probe for devices. I positively _hate_ the mess that is IDE, so I'll need
a little help there - what's the proper course of events when probing for
devices? Do we need to set up more than just the hwif data? The Amiga IDE
drivers don't seem to do more, either. Falcon IDE shares the interrupt
with floppy, SCSI and DMA, that's where the stdma_lock hack comes in. Can
I plug in the locking functions somewhere sane, or will I need to patch
the core driver again?
An interrupt never comes in for IDE - do you generate interrupts on IDE
request complete, or just fake request completion immediately?
Michael
Reply to: