Re: More Amiga resource handling + misc

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

On Sat, 26 Aug 2000, Roman Zippel wrote:
> > This patch adds resource handling to the
> >   - Amiga mouse driver
> >   - Amiga keyboard driver
> >   - A3000 builtin SCSI driver
> >   - Amiga parallel port driver
> >   - Amiga timer driver
> >   - Amiga system RAM (I'm still interested in your /proc/iomem)
> 
> Here is the output with apus-2.4.0-test7:
> 
> 00000000-001fffff : Chip RAM
>   000be728-000be73f : Beep
>   000be750-001ffff7 : amifb [RAM]
> 00200000-009fffff : Zorro II mem
> 00bfeb00-00bfebff : amikeyb
> 00dd2000-00dd3fff : IDE
> 00dff00a-00dff00b : amigamouse [Denise]
> 00dff030-00dff033 : amiserial [Paula]
> 00dff0e0-00dff1ff : amifb [Denise/Lisa]
> 00e80000-00efffff : Zorro II exp
>   00e90000-00e9ffff : AriadNE2 Ethernet
>     00e90600-00e9067f : RTL8019AS
> 00f01060-00f2105f : Zorro device 21406400
                      ^^^^^^^^^^^^^^^^^^^^^
This would have been `Phase 5 CyberStorm Mk III [Accelerator and SCSI Host
Adapter]' if you would have applied my last Zorro patches as well :-)

> 07400000-07ffffff : 32-bit Fast RAM

That's A4000 motherboard RAM.

> 08000000-09ffffff : 32-bit Fast RAM

And that's CyberStorm RAM, I guess.

> 40000000-7fffffff : Zorro III cfg
> ff000000-ffffffff : Zorro III exp

No Zorro III cards?

> > It also makes the list of Amiga models __initdata, so it no longer consumes
> > precious kernel memory (and to compensate a bit for the extra memory used by
> > the resource management system :-)
> 
> BTW I have sometimes problems with __initdata, I get compiler errors like:
> 
> config.c:52: s_a500 causes a section type conflict
> 
> Anyone any idea? If I disable __initdata it works, I also find nothing
> suspicious in the precompiler output.

I saw that once with my cross-compiler (gcc 2.7.2.3/binutils 2.9.1.0.19) as
well. When I removed the object file and retried, it no longer happened.
Strange...

> > As usual these days, this patch isn't tested. But I have to store the patches
> > I cook _somewhere_ until I can test them myself, haven't I? ;-)
> 
> Most of the drivers want <linux/zorro.h>

You mean <linux/ioport.h>, for request_*region()?

> > +  /* request all RAM */
> > +  for (i = 0; i < m68k_num_memory; i++) {
> > +    ram_resource[i].name =
> > +      (m68k_memory[i].addr >= 0x01000000) ? "32-bit Fast RAM" :
> > +      (m68k_memory[i].addr < 0x00c00000) ? "16-bit Fast RAM" :
> > +      "16-bit Slow RAM";
> > +    ram_resource[i].start = m68k_memory[i].addr;
> > +    ram_resource[i].end = m68k_memory[i].addr+m68k_memory[i].size-1;
> > +    request_resource(&iomem_resource, &ram_resource[i]);
> > +  }
> 
> i is undefined.
> 
> > +static struct resource ram_resource[NUM_MEMINFO];
> > +
> 
> This too late defined.
> (this is only for apus, for m68k it looks correct, but untested here as
> well. :) ) Beside of this it works.

Ah, the APUS tree doesn't seem to have the other Amiga resource patches yet.

BTW, `diff -u arch/m68k/amiga/config.c arch/ppc/amiga/config.c' should show
almost no differences. We really have to find a better way to keep these in
sync...

> If X hadn't crashed, I could have created a patch (I was almost ready
> with the cvs import... :-( ).

Thanks _a_ _lot_ for testing!!

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