[T10] Updates to "CAPPID corrections for incorporation of 22-057"

Bill Martin bill.martin at samsung.com
Wed May 24 06:13:24 PDT 2023


The cited text is modified in the latest proposal that I posted. There are two places that would be modified based on the issue that I stated. This is food for fodder.

Bill Martin

Chair INCITS T10
Co-Chair SNIA Technical Council

Chair SNIA CMSI

NVMe Board of Directors
SSD I/O Standards
Samsung Semiconductor, Inc.
Cell (408) 499-1839

From: t10-bounces at t10.org [mailto:t10-bounces at t10.org] On Behalf Of Ralph Weber
Sent: Tuesday, May 23, 2023 7:57 PM
To: T10 Reflector <t10 at t10.org>
Subject: Re: [T10] Updates to "CAPPID corrections for incorporation of 22-057"

This issues stated below appear to be similar to those raised by Gerry Houlder during a review of one of the seven revisions of 22-057. The result of Gerry's concern was the following list entry c) text in 6.5.1 "Mode pages overview"...

--- begin cited text ---
The device server shall terminate a MODE SELECT command with CHECK CONDITION status with the sense key set to ILLEGAL REQUEST and the additional sense code set to INVALID FIELD IN PARAMETER LIST, if the mode parameter header sent with that MODE SELECT command specifies a CAPPID bit that is set to one and:
a) the Capacity/Product Identification Mapping VPD page (see 6.6.6) is not supported;
b) the BLOCK DESCRIPTOR LENGTH field (see SPC-6) is set to zero; or
c) the mode parameter block descriptor contains a LOGICAL BLOCK LENGTH field (see 6.5.2.2 and 6.5.2.3) that specifies a value that is different from the current logical block length.
--- end cited text ---

Certainly, the proposal to add complexity to the VPD page definition will require modifications to the above cited text.

All the best, .Ralph

From: t10-bounces at t10.org<mailto:t10-bounces at t10.org> <t10-bounces at t10.org<mailto:t10-bounces at t10.org>> On Behalf Of Bill Martin
Sent: Tuesday, May 23, 2023 3:09 PM
To: T10 Reflector <t10 at t10.org<mailto:t10 at t10.org>>
Subject: Re: [T10] Updates to "CAPPID corrections for incorporation of 22-057"

I have received the following comments from Rich that should be considered as we look at the direction that we should be taking related to this:

Although it's possible to define the VPD page in such a way that it has different content for each logical block length, I believe there is extreme difficulty in defining when precisely that content changes. Does it change when a new logical block length is MODE SELECTed? Or when the device is formatted and the new logical block length actually takes effect?

This is why I've suggested that the VPD page be revised to list all possible combinations of logical block length and product ID / logical block count. That obviates the need to change the page when the block length is changed.

I think we have to solve the following conundrum:

1. MODE SELECT, CAPPID=1, no change in logical block length - this immediately changes the logical block count and the product ID (as revealed by the inquiry data). Very consistent with existing behavior - no format is needed.

2. MODE SELECT, CAPPID=1, change in logical block length - this requires a FORMAT UNIT to take effect from a functional standpoint (i.e. medium operations will behave differently, READ CAPACITY reports new data, etc.), but is the format required to produce the following:
a. a change in the CAPPID VPD page?
b. a change in the product ID in the standard inquiry data?

There are many tricky things to work out, and I certainly might have missed some of them.


Bill Martin
Chair INCITS T10
Co-Chair SNIA Technical Council
Chair SNIA CMSI
NVMe Board of Directors
SSD I/O Standards
Samsung Semiconductor, Inc.
Cell (408) 499-1839

From: t10-bounces at t10.org<mailto:t10-bounces at t10.org> [mailto:t10-bounces at t10.org] On Behalf Of Bill Martin
Sent: Tuesday, May 23, 2023 10:28 AM
To: T10 Reflector <t10 at t10.org<mailto:t10 at t10.org>>
Subject: [T10] Updates to "CAPPID corrections for incorporation of 22-057"

I have uploaded 23-044r2. There are two major changes:
1)      The definition of the logical unit's maximum capacity has been moved out of the "i.e.," clause into it's own paragraph as follows:

The logical unit's maximum capacity is a vendor specific value. If the Capacity/Product Identification Mapping VPD page (see 6.6.6) is supported, then that vendor specific value shall be less than (e.g., a reduced value due to depopulation) or equal to the value in the allowed number of logical blocks field in the Capacity/product identification descriptor with a value in the product identification field that equals the value of the product identification field in the standard Inquiry data.
2)      There is currently a restriction that CAPPID cannot be used if the logical block length is changed. The original proposal included wording that would have allowed that, but did not find all of the locations where this was prohibited. I have removed those additional places. This has the following discussion that needs to take place during the ad hoc meeting tomorrow and in the next CAP meeting.

Currently the value in the Capacity/product identification descriptor is a capacity that is in logical blocks without any reference to what the logical block length is. The issue is that two different scenarios could take place:
a)      The application client sends a MODE SELECT that changes the logical block length without setting the cappid bit to one and then sends a MODE SELECT that changes the capacity with the cappid bit set to 1. What does the capacity mean in this case?
b)      The application client sends a MODE SELECT that changes the logical block length and  the capacity with the cappid bit set to 1. What does the capacity mean in this case? Currently this is not allowed, but if the previous example is allowed, why should this example not be allowed?

Please download the new proposal, peruse it, and come with your favorite ammunition for the meeting tomorrow.  Fortunately it is a virtual meeting, so the flaming arrows will only be virtual. I know that I have potentially stirred up a hornets nest, but I feel that it is important to resolve this so that we can understand what customers really want.

Thank you,

Bill Martin
Chair INCITS T10
Co-Chair SNIA Technical Council
Chair SNIA CMSI
NVMe Board of Directors
SSD I/O Standards
Samsung Semiconductor, Inc.
Cell (408) 499-1839

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://www.t10.org/pipermail/t10/attachments/20230524/8139d77e/attachment.html>


More information about the T10 mailing list