Re: Incoming to hq.fsmlabs.com:/home/bk/linuxppc_2_4

[Geert U. <ge...@li...>]

On Mon, 23 Apr 2001, Gabriel Paubert wrote:
> On Mon, 23 Apr 2001, Benjamin Herrenschmidt wrote:
> > >In the short term we could either do the pmac_ide_init thing like in
> > >2.2, or we could #define cli() to be ide_cli() in <asm/ide.h> and have
> > >that go through and do disable_irq for all the IDE irqs.
> > 
> > The pmac_ide_init() thing appear to be the "sanest" way to do it.
> > #define'ing cli() looks as hellish as the current workaround ;)
> 
> Indeed.
> 
> > I'm surprised as well that Linus ever accepted our IDE hack that
> > #define insw/outsw & friends in ide.h ... I'm wondering if we can
> > get rid of that or if it's still really useful ? Since we cleaned
> > up our generic insw/outsw functions to not do byteswap, I beleive
> > IDE should be able to use them "as is".
> 
> Yes. BTW how does it work on the APUS, the hardware swaps bytes it seems,
> but either:
> 
> - they don't have IDE (unlikely)

Of course APUS has IDE. It's just an Amiga with a PPC accel card.
Look at drivers/ide/gayle.c for the A1200/4000 builtin IDE.

BTW, there are also IDE expansion cards for Amiga.

> - the IDE controller takes this into account for PIO accesses, and byte
> swapping rules depend on who is the master (ugly)

Nothing has to be swapped. Note that this is not PCI.

> - we need to use the byte-swapping {in,out}s[wl] for IDE on APUS
> 
> - all the disks on APUS are byteswapped (so you can't read a CD) or
> you have to use fancy options.

No problem with disks and CDs. That problem exists on Atari :-)

> - nobody cares anymore about the APUS.

http://linux-apus.sourceforge.net/

> - all of the above :-)

Gr{oetje,eeting}s,

						Geert

--
Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- ge...@li...

In personal conversations with technical people, I call myself a hacker. But
when I'm talking to journalists I just say "programmer" or something like that.
							    -- Linus Torvalds