Subject: Re: Boot netbsd 2.0 on RiscPC
To: David Brownlee <abs@NetBSD.org>
From: Peter Teichmann <firstname.lastname@example.org>
Date: 03/28/2005 23:33:05
Am Samstag, 26. M=E4rz 2005 02:13 schrieb David Brownlee:
> So to clarify - both the 2.0 !BtNetBSD using boot32, and
> direct use of the 2.0 boot32 fail for you, but work with
> an earlier BtNetBSD?
They do not work with boot32 as supplied with NetBSD 2.0 (well, they work=20
sometimes, but very seldom, maybe 1 out of 30 tries), and do work with=20
btnetbsd, both v1.4 2004/0/24 as supplied with 2.0, and 1.5 2002/03/24 as=20
supplied with the btnetbsd-0.99a archive (both did never fail).
> > Anyway, i tried it with several combinations of the following 6 SIMMs:
> > SIMM A,B: 32MB, with 2 banks 16MB each (the original ones)
> > SIMM C,D: 32MB, with 2 banks 16MB each (different type)
> > SIMM E,F: 64MB, with 1 bank of 64MB
> > Needless to say, all SIMMS were stable in the machines they were in
> > originally. BtNetBSD was able to boot with any combinations of them.
> > Boot32 was only able to boot with SIMM C or D in any of the slots, but
> > not with both of them. It failed with any other combinations I tried.
> > How can we explain this?
> That is very strange. Could you look at the messages boot32
> gives (load an invalid kernel - README works quite well :)
> and see if they are any different between one of C/D and
> one of A/B?
It could be that one difference beetween the SIMMs is that one could be EDO=
and the other FPM, but I do not know which is what. The 64MB ones are quite=
sure EDO. Would that make a difference?
Could you give me a hint where I can find "load an invalid kernel - README"=
guess I can find there information how to preserve the boot32 messages,=20
because they go away too fast if booting works.