> Ulrich Windl wrote:
> > Yes! Yesterday I've made a boot disk with a real power kernel (>900k
> > uncompressed). Running it on a machine with 32MB I noticed that the
> > kernel always reads from the root disk (although it should be a
> > ramdisk, I'll have to re-check that). But I thought that there could
> > be a bug somewhere.
First let me add: The buffer cache works, but it seems to get filled
very slowly. The filesystem was minix. It seems 1.3.68 doesn't read
whole tracks, but only the sectors required (no read-ahead).
Up to now I wasn't able to find out why the traditional root disk
image can't be loaded as ramdisk (prompt asks for root disk). Maybe
that's a LILO problem.
> I've been following this thread, I have a Promise EIDE card w/WD4300
> disk, and all I am is impressed with the speed of everything. When I see
> an insect alert from the mail list, I try to duplicate [if I can] and
> not ONCE have I been able to duplicate/notice. [Thank you God!]
> What kind of HD/interface/hardware setup do you have?
Maybe you are right. It's hard to see a bug in here. The buffer only
gets filled slowly. I think 1.3.69 corrected some problems about the
buffer cache. I think that problem is independent of the hardware.
> I'm running a `no-name' Taiwan MB w/20meg RAM 486DX4 Soundblaster
> MultiCD, Sony CDU31A, Diamond Stealth 64 w/2meg DRAM, USR 14,400 Fax
> Modem. Kinda plain jane, but it works for me. Am I doing something
Your disk configuration is missing. (EIDE, which?)
A friend on mine has the same graphics card as you. He found out that
(running MS Windows) the card maps in some additional memory when
running SVGA modes. That caused an application to crash from time to
time. We had spent several hours searching for interrupt/DMA/ioport
conflicts, without finding anything. Things can be complicated.
Ulrich Windl Klinikum der Universitaet Regensburg
Rechenzentrum DV-med Franz-Josef-Strauss-Allee 11
Tel: +49 941 944-5879 D-93053 Regensburg
FAX: +49 941 944-5882
Just imagine my mail address were <Ulrich.Windl@rz.uni.r.de>...