[[ Time to reopen this again ;) ]]

Now that I have some working bootblocks, there's one or two more
questions.  To summarise - in the process of updating the pmax
bootblocks to something resembling the alpha bootblocks, I'm trying to
set up something that is multi-architecture friendly.  This is primarily
for CD-ROM distributions that can boot on multiple types of machines.
Here's some limitations we appear to be facing:

 a) Subdirectories are out because of the 7.5k code size limit on
    some architectures.
 b) The apparent ISO limitation of an 8.3 filename.  Todd Whitesel
    says that "pure" ISO filesystem allow more than this, but the
    code in sys/lib/libstand/cd9660.c couldn't find a "boot.pmax".

It would be "Really Nice(tm)" if we could get something that all (or
most) ports agree on.  I was talking to Jonathan about getting my stuff
in the tree, and he came up with the suggestion of calling the second
stage /${MACHINE}.btb - AFAIK .btb isn't used by anything else, and it
even sounds a little like "bootblock".

As far as backwards compatibility, on the pmax having the first stage
check for /pmax.btb and then /boot if that doesn't exist adds only 48
bytes...


There is also the question of what to call kernels on multi-arch boot
media.  One idea is /${MACHINE}, but hopefully a second stage wouldn't
have subdirectory constraints so something like /kernel/${MACHINE} may
make more sense.  Fortunately I don't need this one finalised before I
finish the rest of my pmax bootblock work.

Simon.