pl7 AHDI and ICD fun ...

[shenson@nyx.cs.du.edu (Stephen Henson)]

Hmmm pl7 may well not have the mount() stuff. However I've been using some
rather unusual methods which suggest I might have a solution to the ICD
bug (which is still ICD's fault not mine!) so the caches can be left on
when using huge partitions.
        The protection problem with AHDI seems rather serious. If anyone
has any suggestions (since none of mine seem to work) I'd love to hear
them. One possibility I'm toying with is to have a pseudo boot sector
which makes the sector size 512 bytes (where protection isn't a problem)
and to save the real boot sector data info elsewhere (behind the super block
or in the pseudo FAT's), minit would then restore this on request. The driver
could then try to read the last block at startup and if it failed (or logical 
lrecno bug was present) switch to physical mode.
        Of course I could always grasp the nettle and read the partition info
manually and interpret that and protect using a different partition id. This
only then leaves the problem of drive changes and drives with no TOS partitions
(e.g. how do I find out they exist?).
        Of course all of these problems are AHDI specific ... It seems like
AHDI may inherit ICD's crown as the number one 'pain in the arse driver
software', XHDI compliant software is currently in last place on this list
:-)
Steve.