Locked in cache question -Reply -Reply

Jim Mcgrath jmcgrath at QNTM.COM
Fri May 16 11:57:16 PDT 1997


* From the SCSI Reflector (scsi at symbios.com), posted by:
* jmcgrath at qntm.com (Jim Mcgrath)
*
     I have an easy answer for all of this.  Since locking a cache never
     guarentees that you will get a cache hit (the device can always get
     an error) just treat it as advice - the host thinks this data will 
     be accessed often/soon.  If the device ignores the hint, then its
     performance will be lower.  If the device thinks obeying the hint
     would cost it too much (i.e. lock upo the device), it is free to
     ignore it.
     
     To my recollection, the mode page cache control bits are
     recommendations, not requirements, for this very reason (the
     only exception is write cache enable, since data integrity,
     not performance, issues could be raised).
     
     Jim
     


______________________________ Reply Separator _________________________________
Subject: Re: Locked in cache question -Reply -Reply
Author:  skinner at kodak.com (Brent Skinner) at SMTP
Date:    5/9/97 8:18 AM


* From the SCSI Reflector (scsi at symbios.com), posted by: 
* skinner at kodak.com (Brent Skinner)
*
Well, I sat all day looking at pictures on my whiteboard, trying to figure 
out how to support such a thing. It's going to take more code, but at least 
I won't have to throw away most of what I've done so far.
     
I'd recommend some wording around this kind of thing in the spec, but 
I'm probably the only one that could possibly see it that way.
     
So, if a host were to do a huge read starting at lba 100 and fill 
the cache, then lock the entire contents, followed by a read from 
lba 0, do I want to error or throw away some of the locked data in
order to make room for the new information? I'm sure our QA folks are 
going to try this.
     
From: PAT LaVarre 
> < I don't know enough about this subject to 
> < comment intelligently ... but the subject 
> < interests me enough that I'll comment
> < anyhow.
> ...
> > From ... (Brent Skinner)
> 
> Thanks for your polite attention.
> 
> I find your perspective as reasonable as
> mine ... but beyond that I find myself too 
> ignorant to add anything further ...
> 
> ... I'll be interested to see what in the 
> end you decide to do.
> 
> Cheers.   Pat LaVarre
> p.lavarre at ieee.org
> 
> 
     
     
"You learn all of your life...and die as a fool." -- Old Russian saying.
                                ____  ______
Brent Skinner                  |   / /      | Eastman Kodak Company 
<skinner at kodak.com>            |  / /       | Optical Storage Products
                               | / /        | 460 Buffalo Road
The Eastman Kodak Company      |( (  Kodak  | Rochester, NY 14652-3846 
doesn't know what I've written | \ \        | 
here, probably doesn't WANT to |  \ \       | Voice: 716-588-0389 
know and would, most likely,   |   \ \      | Fax:   716-588-5578 
disagree with me anyway.       -----  ------
*
* For SCSI Reflector information, send a message with
* 'info scsi' (no quotes) in the message body to majordomo at symbios.com
*
* For SCSI Reflector information, send a message with
* 'info scsi' (no quotes) in the message body to majordomo at symbios.com




More information about the T10 mailing list