A SENSE command is not dangerous. Its response comes from the Control
Unit and only reports status without altering it. If there is no device
at the address, a cc=3 is returned by the VM CP based on the lack of a
VDEV. 

About the only window for mischief, and it is just hypothetical, would
be if you could issue the SENSE between the time that a Unit Check is
presented by the hardware to CP and CP's doing a SENSE to get the
status. Like I say, hypothetical because I doubt that there really is
even that teeny-tiny window - if there is a unit check, CP will
undoubtedly do a SENSE before it presents the UC to the VM starts the
next virtual machine I/O. If the VIO is a SENSE, the saved status is
returned; if not, the status is cleared just as it would be in the real
hardware.


Regards, 
Richard Schuh 

 

> -----Original Message-----
> From: The IBM z/VM Operating System 
> [mailto:[EMAIL PROTECTED] On Behalf Of Imler, Steven J
> Sent: Wednesday, January 23, 2008 10:07 AM
> To: [email protected]
> Subject: Re: how to wait for dfsmsrm command
> 
> But isn't that a bit dangerous ... for the requestor to be 
> issuing tape commands and/or CCWs to the attached tape drive 
> while RMSMASTR is "doing its thing" mounting the tape for the 
> requestor?
> 
> JR (Steven) Imler
> CA
> Senior Software Engineer
> Tel:  +1 703 708 3479
> Fax:  +1 703 708 3267
> [EMAIL PROTECTED]
> 
> 
> > -----Original Message-----
> > From: The IBM z/VM Operating System
> > [mailto:[EMAIL PROTECTED] On Behalf Of Les Geer 
> (607-429-3580)
> > Sent: Wednesday, January 23, 2008 12:50 PM
> > To: [email protected]
> > Subject: Re: how to wait for dfsmsrm command
> > 
> > > } The only problem with using a tool like this is that while 
> > > DFSMS/VM is } mounting the tape for you, the tape drive 
> is attached 
> > > to the RMSMASTR } virtual machine, not the requestor's virtual 
> > > machine.  The tape drive is } not attached to the requestor until 
> > > the requested tape is loaded and the } drive is ready ...
> > 
> > Or first attach the drive to the requester's machine using the 
> > multiuser option before issuing the mount request.
> > 
> > 
> > Best Regards,
> > Les Geer
> > 
> > 
> 

Reply via email to