I was wandering through the executive sources of RSX-11M-Plus V4.6
today, and found a bunch of code conditional on E$$LBN, to support
32-bit logical block numbers. I was surprised, to say the least.
The V4.6 Release Notes say the maximum disk size is 2**23-1. E$$LBN
is not defined. I don't see any sign of E$$LBN in the F11ACP sources,
where changes would definitely be required to support 32-bit LBNs.
In particular, the format-3 retrieval pointer (with 32-bit LBN and
16-bit count) from the ODS-1 spec would have to be implemented. (Or
maybe it has been, but isn't conditional on E$$LBN).
I wonder where else one would find 24-bit dependencies. Maybe only
where some programmer was trying to save a byte.
Anyway, it looks like back in 1996 Digital was starting to get ready
for L-A-R-G-E disks in RSX.
- Rob
--
Rob Brown b r o w n a t g m c l d o t c o m
G. Michaels Consulting Ltd. (780)438-9343 (voice)
Edmonton (780)437-3367 (FAX)
http://gmcl.com/
[cut]
> The V4.6 Release Notes say the maximum disk size is 2**23-1.
[cut]
Can't comment on RSX but I can remember fighting MicroVAX 3100
systems that used 6 byte SCSI commends and thus had trouble
booting >2GB disks. Later systems used 10 byte SCSI commands
thus large disks bootable. VMS used 10 byte SCSI commands
so it would happily use large disks as non-system disks.
Andy Stewart.
-Dave
> Section 1.4.8 of the release notes for M+ V4.6 indicate that the
> maximum size for DU devices is 2**24-1 (~8.6G)
You are right, my mistake.
> The FCP changes for 2/4 pointers was never included on the master
> pack, since it was somewhat pointless until BRU was sorted (which
> never happened).
Thanks for the info.