On Wed, Nov 07, 2001 at 12:56:49PM +0000, Ben Harris wrote:
> On Wed, 7 Nov 2001, Ian Fry wrote:
> > On Wed, Nov 07, 2001 at 11:51:15AM +0000, Ben Harris wrote:
> > > On Wed, 7 Nov 2001, Ian Fry wrote:
> > > > I can now boot my A4000 to single user mode and enable swapping,
> > > > which I've never been able to do before. This stops all the 'VM
> > > > shortage, sleeping' messages :-) I still had to use a 'tuned' kernel
> > > > config, though - the FOURMEG config still doesn't quite work.
> > > Wow!  I'd almost got my A3020 to work this well, but not quite.
> > Well, I tried going to multi-user after I'd enabled swap, but sh dumped
> > core processing the rc scripts (this was using a userland from the latest
> > snapshot - I still want to try a more recent cross-built userland before
> > I start panicing about this).
> Ah.  That looks like you're in about the same state as me, then.  Have you
> got root on NFS?  For some reason, writing over NFS seems to be flakey at
> the moment, which makes getting valid coredumps to debug difficult.

Yes, currently, I have root on NFS. Is it just root on NFS, or swapping over
NFS? If the latter, I might try to get local swap setup.

One thing I forgot to mention is that taking out the 'ea' bug workaround
improves things *enormously*.

> > > > and vt220l8x8 could probably be made const (presumably, all the other
> > > > wsfonts will need similar fixes)
> > > No, it can't.  The wsfont machinery byte- and bit-swaps fonts in place
> > > if they turn out to be the wrong endianness for the framebuffer.
> > Ah, okay. Presumably it only does the byte/bit swapping once, when the
> > wsfont stuff is initialised?
> I think so.  It all looked a bit of a mess when I was last in that area,
> though.

Yes, I know - I've been looking around that code with the intention of
getting rasops to work in 4bpp modes, but I'm having trouble following some
of the code...

Ian.