[8714] in linux-scsi channel archive
Re: question on short transfers
daemon@ATHENA.MIT.EDU (sjbuller@us.ibm.com)
Tue Apr 25 19:32:32 2000
From: sjbuller@us.ibm.com
To: linux-scsi@vger.rutgers.edu
Message-ID: <872568CC.0080A97A.00@d53mta04h.boulder.ibm.com>
Date: Tue, 25 Apr 2000 16:00:08 -0600
Mime-Version: 1.0
Content-type: text/plain; charset=us-ascii
Content-Disposition: inline
Matthew Dharm <mdharm-scsi@one-eyed-alien.net> wrote:
>Personally, I think underflow is really an error condition, and should
>cause an auto-REQUEST_SENSE and sense data to be returned along with
>CHECK_CONDITION or some other message.
>
>Really, in the case of an underflow, we've most likely got some type of
>media problem. And it's very likely that even the part we _think_ got
>read/written correctly didn't actually.
>
>So I think we should rip this out.
>
>Matt Dharm
It is standard practice in variable block tape applications to always
issue a read for max block size. 99.999% of the reads result in a
check_condition for requested size != actual. Its not a media problem.
-steve
-
To unsubscribe from this list: send the line "unsubscribe linux-scsi" in
the body of a message to majordomo@vger.rutgers.edu
