[Top] [All Lists]

Re: Problems booting a 4000SC.

Subject: Re: Problems booting a 4000SC.
From: Ralf Baechle <>
Date: Wed, 22 Oct 1997 17:54:20 -0700
In-reply-to: <>; from Charles A. Dempsey on Wed, Oct 22, 1997 at 11:10:29AM -0600
References: <>
On Wed, Oct 22, 1997 at 11:10:29AM -0600, Charles A. Dempsey wrote:
> I've got my Magnum 4000SC-50 set up, but I can't get a kernel to boot.
> Using either the precompiled milo-0.27.1 milo/pandora or a milo that I 
> compiled against linux-970710 (the corresponding pandora compiles but 
> doesn't work) the machine appears to load the kernel but then just sits 
> there with a blue screen. 
> At the "run a program" prompt;
> a:milo (floppy spins 1 min, drive light goes out, blue screen forever)
> a:pandora
> Pandora> boot 
>       Kernel text at 0x80000000, size 1211344 bytes
>       Kernel data at 0x80127bd0, size 0 bytes
>       Kernel bss at 0x80127bd0, size 127636
>       Kernel entry at 0x80000494
> (the floppy spins 8 min, drive light goes out, blue screen forever)

The awful low speed is "normal", the ARC firmware is Magnum is buggy.
Somebody should probably try to find a faster workaround ...

> Pandora> boot vmlinux (Same as above)
> I've tried this with three different 970710 kernels including the one 
> that works on my 4000PC. I believe I had this problem with my 4000PC-50, 
> but milo-0.27.1 with 970710 fixed it.
> The machine has the standard Magnum video card, 16 MB RAM, Matsushita CD 
> on ID=3, no hard drive.
> Does anyone have their 4000SC booting? What kernel/milo/pandora did you 
> use?

I suspect we've got still a bug in the handling of the SC/MC caches.
Hard to fix without having a machine with such a CPU.  Another cache
problem we have to fix are I/D caches with different line sizes like
in the Olivetti M700-10.  And we don't handle split I/D level 2
SC/MC caches correctly, but I so far I don't of any machine using such


<Prev in Thread] Current Thread [Next in Thread>