From: makisara@vtinsx.ins.vtt.fi (Kai Makisara) Subject: Re: Linux .99pl4,5 can't read tape Date: 21 Apr 1993 18:28:10
In article <1r3h35INNs6i@senator-bedfellow.MIT.EDU> jfc@athena.mit.edu (John F Carr) writes:
...
I suspect the problem is that the READ command for SCSI tapes does not
specify a starting block, but implicitly requests data starting at the
current position. The UltraStor seems to be noticing that all tape requests
are for the same blocks, and returning data from the cache on the second and
subsequent reads. I'm not sure if this is a bug or a feature.
--
John Carr (jfc@athena.mit.edu)
The SCSI READ command for sequential-access devices does not have the
starting block field because it does not fit in the sequential-access
model. The transfer length field three bytes and thus fills most of the
space used for the starting block in a 6-byte READ for direct-access
devices. Since the firmware on the UltraStor board probably does not
treat differently sequential-access and direct-access devices, the
driver should take care of this.
I don't know about Adaptec 1542C but the 1542B does not have any cache.
Kai