> This question has been asked a few times, but never addressed.

Never?  Heck, I alone have posted messages about it every few months
(though admittedly I'm not certain how many of them have been on
port-sparc64; at least some have been on port-sparc).

> I messed with it a little while and it looks like the bootloader
> can't read past a certain block on the disk.

This is reasonably likely.  Some Sun ROM versions don't know how to use
10-byte CDBs, and 6-byte CDBs can't access beyond the first 1G of the
disk (a 6-byte CDB has only 21 bits of block number).  These versions
also aren't smart enough to return errors when attempting to access the
rest of the disk, instead just droppng the high bits of the block
number and thus getting the wrong block.

Of course, whether this is actually what's wrong in your case I can't
say for sure, but it matches your symptoms insofar as you've described
them.

> I've reinstalled with a small / and /usr on a separate partition, and
> my kernel that I built this morning works fine.

Note that a small boot partition is not alone enough; it also must be
early on the disk (within the first 1G).

> Is this a known issue?

It's a known issue with many machines.  Whether it's a known issue with
UltraAxes I couldn't say.

The ROM code version would be a useful value to report.

/~\ The ASCII				der Mouse
\ / Ribbon Campaign
 X  Against HTML	       mouse@rodents.montreal.qc.ca
/ \ Email!	     7D C8 61 52 5D E7 2D 39  4E F1 31 3E E8 B3 27 4B