Re: holes for ISA boards...

david parsons (o.r.c@p.e.l.l.p.o.r.t.l.a.n.d.o.r.u.s)
16 Jun 1999 10:18:28 -0700


In article <linux.kernel.E10uEH3-0004lP-00@kirtley.cl.cam.ac.uk>,
Stephen Early <Stephen.Early@cl.cam.ac.uk> wrote:
>In article <7k6k9q$n11@pell.pell.portland.or.us> you write:
>> In the simple case, setting mem=<foo> on the command-line should
>> circumvent any probing.
>
>No; probing still takes place, but the results are ignored when mem=
>is set. The problem with this motherboard was that calling int 0x15,
>eax=0xe820 made the BIOS print "Out of memory" and halt the machine,
>so specifying mem= doesn't help.

Well, the simple solution to that would be to #define
NO_FANCY_MEMORY and build a custom kernel. It's also dimly
possible that that one machine is running into a buglet in your
code (I certainly found some buglets in mine when I moved it over
to 2.2.9 from 2.0.28; though in my case it wasn't a subtle
interaction against one bios, but a infinite reboot machine on
every machine around :-() or it could be that the bios in question
is actually broken.

>If the e820 call code in setup.S is only enabled when we're compiling
>a 586+ kernel, we should be alright.

I can't say for your testsuite, but I own 486 machines that do
properly support e820, and which let me put in memory holes.

____
david parsons \bi/ and it's also possible it's some interaction with
\/ bzImages, which I have to use with 2.2.x :-(

-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu
Please read the FAQ at http://www.tux.org/lkml/