🚀 go-pugleaf

In article <1991Mar18.171629.29960@dg-rtp.dg.com> hardin@mavplus3.rtp.dg.com (Paul Hardin) writes:
>	The problem is that data can potentially be lost, with drives
>	of formatted capacity greater than 1.073742 GBytes!  

	Depends on a) how you calculate Gig (if K=1024, M=K*K, G=K*M), then
it's 1.0 Gig exactly (blocks 0 - $1fffff), b) logical block size.  The
1Gig limit is for 512 byte logical blocks (probably the most common).  For
1K logical blocks, it would be 2 Gig, etc.

	BTW, once I heard this I checked the commodore scsi drivers, and we
have the same problem (the biggest drive we had checked was ~9xx Mb).  Our
drivers (A590/A2091/A3000) do use 10-byte read/writes, but only if the
transfer size is > 256 blocks.  We're fixing this, of course.  At least we're
in good company (Dec, both Ultrix and VMS).  1/2 :-)

	Anyone playing with >1Gig drives should either use >512 byte blocks
(under 2.0 FS - don't forget to format the drive using the scsidirect
interface for 1K logical blocks) or just not use that area until they get
the fix (37.79 or better for A3000's).

	I wonder how many other people will be bitten by it.  Anyone know 
about Apple, nExt, Sun?

-- 
Randell Jesup, Keeper of AmigaDos, Commodore Engineering.
{uunet|rutgers}!cbmvax!jesup, jesup@cbmvax.commodore.com  BIX: rjesup  
The compiler runs
Like a swift-flowing river
I wait in silence.  (From "The Zen of Programming")  ;-)