insw/outsw in ide.h

insw/outsw in ide.h

[David Monro]

> Is there any PPC platform that needs to use a byte-swapping insw/outsw in
> the IDE driver?  (I can't see why any platform would, but you never
> know...)

Yes - the IBM 850 does. Otherwise the firmware doesn't understand the
partition table and the CD drive doesn't work.

Cheers,

	David

** Sent via the linuxppc-dev mail list. See http://lists.linuxppc.org/

insw/outsw in ide.h

[Paul Mackerras]

Is there any PPC platform that needs to use a byte-swapping insw/outsw in
the IDE driver?  (I can't see why any platform would, but you never
know...)

I have changed the insw/outsw/insl/outsl in include/asm-ppc/io.h to be the
non-swapping form (in my tree at least, I'll push it to bk soon).  I would
like to take out the special definitions of insw/outsw in
include/asm-ppc/ide.h and just let the ide driver use the standard
(non-swapping) form.  Will this cause a problem for any PPC platforms?

--
Paul Mackerras, Senior Open Source Researcher, Linuxcare, Inc.
+61 2 6262 8990 tel, +61 2 6262 8991 fax
paulus@linuxcare.com.au, http://www.linuxcare.com.au/
Linuxcare.  Support for the revolution.

** Sent via the linuxppc-dev mail list. See http://lists.linuxppc.org/

RE: insw/outsw in ide.h

[Gary Thomas]

On 01-Mar-00 Paul Mackerras wrote:
>
> Is there any PPC platform that needs to use a byte-swapping insw/outsw in
> the IDE driver?  (I can't see why any platform would, but you never
> know...)
>
> I have changed the insw/outsw/insl/outsl in include/asm-ppc/io.h to be the
> non-swapping form (in my tree at least, I'll push it to bk soon).  I would
> like to take out the special definitions of insw/outsw in
> include/asm-ppc/ide.h and just let the ide driver use the standard
> (non-swapping) form.  Will this cause a problem for any PPC platforms?
>

I can't tell for sure, but won't this break existing [already installed]
systems?  I agree that the data path really doesn't matter, but once it's
on the media it probably does.

Have you tested this change on a system with an IDE drive that has a
live file system in place?

.. just my 2c

** Sent via the linuxppc-dev mail list. See http://lists.linuxppc.org/

RE: insw/outsw in ide.h

[Geert Uytterhoeven]

On Wed, 1 Mar 2000, Gary Thomas wrote:
> On 01-Mar-00 Paul Mackerras wrote:
> > Is there any PPC platform that needs to use a byte-swapping insw/outsw in
> > the IDE driver?  (I can't see why any platform would, but you never
> > know...)
> >
> > I have changed the insw/outsw/insl/outsl in include/asm-ppc/io.h to be the
> > non-swapping form (in my tree at least, I'll push it to bk soon).  I would
> > like to take out the special definitions of insw/outsw in
> > include/asm-ppc/ide.h and just let the ide driver use the standard
> > (non-swapping) form.  Will this cause a problem for any PPC platforms?
>
> I can't tell for sure, but won't this break existing [already installed]
> systems?  I agree that the data path really doesn't matter, but once it's
> on the media it probably does.
>
> Have you tested this change on a system with an IDE drive that has a
> live file system in place?

I know some people who formatted their disks on ia32 boxes and transferred them
to their CHRP boxes later, so the current code should work.

An easy way to find out is to try to mount a CD-ROM in an IDE CD-ROM drive. If
it works, endianness is OK.

Gr{oetje,eeting}s,

						Geert

--
Geert Uytterhoeven -- Linux/{m68k~Amiga,PPC~CHRP} -- geert@linux-m68k.org

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


** Sent via the linuxppc-dev mail list. See http://lists.linuxppc.org/

Re: insw/outsw in ide.h

[Cort Dougan]

I think the 850's need this.  I don't have one and haven't tested on one
but I recall some IDE silliness was necessary.

** Sent via the linuxppc-dev mail list. See http://lists.linuxppc.org/

Re: insw/outsw in ide.h

[Geert Uytterhoeven]

On Fri, 3 Mar 2000, Cort Dougan wrote:
> I think the 850's need this.  I don't have one and haven't tested on one
> but I recall some IDE silliness was necessary.

BTW, I'm in favor of having ide_*()-ops for all accesses, cfr. my comments in
the sources. Then we can distinguish between I/O port and I/O memory acces on a
per-machine/architecture base.

Gr{oetje,eeting}s,

						Geert

--
Geert Uytterhoeven ------------- Sony Software Development Center Europe (SDCE)
Geert.Uytterhoeven@sonycom.com ------------------- Sint-Stevens-Woluwestraat 55
Voice +32-2-7248638 Fax +32-2-7262686 ---------------- B-1130 Brussels, Belgium


** Sent via the linuxppc-dev mail list. See http://lists.linuxppc.org/

Re: insw/outsw in ide.h

[Dan Malek]

Geert Uytterhoeven wrote:


> BTW, I'm in favor of having ide_*()-ops for all accesses,....


Can I also have md.ide_*() for interrupts as well?  Please :-).
On embedded processors we have internal interrupt controllers
that require different initialization and ack'ing than the assumed
8259 type controllers.

Thanks.


	-- Dan

** Sent via the linuxppc-dev mail list. See http://lists.linuxppc.org/

Re: insw/outsw in ide.h

[Cort Dougan]

} Can I also have md.ide_*() for interrupts as well?  Please :-).
} On embedded processors we have internal interrupt controllers
} that require different initialization and ack'ing than the assumed
} 8259 type controllers.

You want ppc_md.ide_{insl,outsl} and so on?

** Sent via the linuxppc-dev mail list. See http://lists.linuxppc.org/