Soft write protect is SPC

Gerry Houlder Gerry_Houlder at notes.seagate.com
Fri Jan 26 10:32:28 PST 1996


* From the SCSI Reflector, posted by:
* Gerry Houlder  <Gerry_Houlder at notes.seagate.com>
*
>If we add a Soft Write Protect field in the SPC we could have three separate 
>documents which use this field. The SPC, SSC, and SMC could/would all have 
>SWP.  Potentially, a medium changer device would have to keep track of the 
>SWP bit to be in compliance with all three documents. The obvious question is 
>which document would have precedence?  What happens in a multi-initiator 
>situation and we have SWP support in three different pages?  What is the 
>state of the SWP bit after a media change or other Unit Attention condition?
>What are the effects if the Mode Select cdb was sent with save pages set?

I am not aware that SMC has yet a third place for a SWP bit, but let's assume 
that
it does. The choice of which of these 3 SWP bits is required will be determined
by the customer. The customer will write their drivers to use one SWP function 
for
each device type. For example, there is already a customer base for tape devices
with a SWP function in tape unique page 9. These products won't support the
SPC page Ah or the SMC location in the same product. This is OK with the 
standard
because there isn't a required precedence for duplicative options.

Adding the function to a page common to all device types opens the possibility 
that
all devices can use the function without adding even more SWP bits into more
device unique pages. This will at least prevent further proliferation of other 
SWP
functions in more places.

>If we keep the Soft Write Protect field in any of the documents then I would 
>like to suggest a change in the additional sense code fields in SPC. If this 
>is not already done I would suggest adding a code for the Soft Write Protect, 
>such as, "27 01 - Write Protected, set by host".  The present code for write 
>protect could be changed to, "27 00 - Write Protected, set by physical switch".

I would not object to adding another ASC to differentiate these cases. If a 
system
company sees value in this, I'll be happy to add it.




More information about the T10 mailing list