On Wed, 13 Jun 2001, Juergen Hannken-Illjes wrote:

> I tried my Duo280 some months ago. If it is docked (DuoDock II) the
> console detection didn't work. With the following change it came up:
>
> Index: arch/mac68k/dev/macfb.c
> ===================================================================
> RCS file: /cvsroot/syssrc/sys/arch/mac68k/dev/macfb.c,v
> retrieving revision 1.5
> diff -u -2 -r1.5 macfb.c
> --- arch/mac68k/dev/macfb.c	2000/10/22 05:05:39	1.5
> +++ arch/mac68k/dev/macfb.c	2001/01/14 20:44:35
> @@ -130,4 +130,9 @@
>  macfb_is_console(paddr_t addr)
>  {
> +	if (addr == 0xfe000000 && macfb_consaddr == 0xfe100000) {
> +		printf("HACK: assume console (%08x) for addr=%08x\n",
> +		       (unsigned)macfb_consaddr, (unsigned)addr);
> +		addr = macfb_consaddr;
> +	}
>  	if (addr != macfb_consaddr &&
>  	    (addr >= 0xf9000000 && addr <= 0xfeffffff)) {
>
	Woul that be likely to cause problems if the kernel were
	booted on any other mac models?


> There will be no dock scsi because the attachment of this special slot e
> doesn't work. I gave up because the LC40 has Bug E4 and I wasn't able
> to get around this bug (add a NOP before every F-Line instruction).

	If the only way to run on an LC40 with the E4 bug is to have
	a NOP before every F-Line instruction would there be any sense in
	adding a toolchain option to conditionally compile code that way,
	then at least affected LC40 owners could run (if they had another
	machine from which to compile).

-- 
		David/absolute		-- www.netbsd.org: No hype required --