Subject: Re: NCR Driver Problems
To: Scott Reynolds <firstname.lastname@example.org>
From: Stefan Esser <email@example.com.DE>
Date: 01/26/1996 02:18:28
On Jan 24, 23:07, Scott Reynolds wrote:
} Subject: Re: NCR Driver Problems
} On Wed, 24 Jan 1996, Stefan Esser wrote:
} > And there seem to be drives, that do some kind of
} > soft reset, making the drive unresponsive for as
} > much as a second or two.
} I realize this is a side topic, but could this possibly be thermal
} recalibration you're seeing?
No, the problem is that some programs do open the
raw device several times within fractions of a
second. The driver is using the drive, and for
this reason tagged commands have been issued and
are being processed.
The sd_open() code issues a START UNIT and the
driver sends several such commands (one for each
sd_open()) to the drive simultanously, each with
a unique simple tag.
There seem to be drives that don't like being
treated like this, and which return an error for
each of the START UNIT commands.
The correct fix is to only issue one such command
per drive (i.e. count the openings and issue the
command only if the drive has not been mounted or
opened before). The driver could of course first
stop using tagged commands, then send the START
UNIT command and finally revert to tagged command
operation, but I don't think the driver should have
to be made complex, if the problem can easily be
prevented in the generic SCSI code ...
Stefan Esser, Zentrum fuer Paralleles Rechnen Tel: +49 221 4706021
Universitaet zu Koeln, Weyertal 80, 50931 Koeln FAX: +49 221 4705160