5.0RC2 and negative runtime

[tponsford <tponsford%rnsmte.com@localhost>]

While doing some much needed spring cleaning, I uncovered a bunch o' 
vaxes hiding behind a very large alphaserver in the shed. They were once 
part of a alpha-vax-microvax VMS cluster I built up.I had almost forgot 
about them (well not forgot about them, just where I put them!) One 
very-large dead rattlesnake later ( it too, was hiding behind the 
alphaserver), I cleaned them off of all the dust and mouse-doody and did 
some inventory 1 vax4k-90 and 4k-60, two VLC's  and the odd 3100/38. 
Thankfully they all seemed no worse for wear.

Reading about the good progress Netbsd has made with current/5.0, I 
found a noisy scsi disk and installed 5.0RC2 on the Vax 4000-90. During 
the installation and for a bit later I would get this error message:

WARNING: negative runtime; monotonic clock has gone backwards

it then went away, but appeared again on the Vax4000-60. The TOY clocks 
checked out OK and it appears not to have affected anything..so far, but 
is it a bug, bad hardware?

Also it appears that there was no Xserver,ie., no Xorg,Xdec,or Xwhatever 
binary in the Xserver set. I was hoping to try the SPX driver on the 
graphic framebuffer, but it was MIA. Does the X/SPX driver need to be 
compiled out of the Xorg pkgs?

I also got the dreaded chicken and egg problem when trying to compile a 
package out of pkgsrc. I get the "Circular dependency detected" as it 
searches for gcc-2.95, but gcc-4.1.3 is installed. I thought 4.1.3 was 
generating good vax code, but do I need to install 2.95.? This is from 
the latest version of pkgsrc (03-2009)

I understand that 5.0 is self-hosting now. While I wasn't going to 
rebuild the whole system, It would be good if it could. I don't really 
want to start cross-compiling unless it's necessary, or the package is 
too big.
Except for these glitches 5.0 looks pretty good on a vax.
Any help or pointers on the monotonic clock, the spx framebuffer and the 
pkgsrc issue would be greatly appreciated!

Cheers
tom