[nvmewin] nvmewin Digest, Vol 47, Issue 6

Thomas Freeman thomas.freeman at hgst.com
Mon Nov 30 10:38:02 PST 2015 

Hi Suman,
Thank you for the helpful feedback.  Here are my responses to your list:

1.       I'll make these corrections.

2.       Good point. For the default case, I'll fail the command with a check condition.

3.       The shared buffer is in the srb extension - which should be a unique buffer per request. I don't see a case where the same request requires the driver to issue DSM and persistent reserve simultaneously. Please correct me if I'm wrong.

4.       Good suggestion - I'll make these corrections.

5.       Yes, a good idea. I'll make this correction.

I'm working on the corrections and re-validation now. I'll wait for your answer to number 3 before I re-submit my changes.

Thank you,
Tom Freeman
Software Engineer, Device Manager and Driver Development
HGST, a Western Digital company
thomas.freeman at hgst.com<mailto:thomas.freeman at hgst.com>
507-322-2311

[HGST_Logo_email]
3605 Hwy 52 N
Rochester, MN 55901
www.hgst.com<https://hgst.jiveon.com/external-link.jspa?url=http://www.hgst.com/>

From: SUMAN PRAKASH B [mailto:suman.p at samsung.com]
Sent: Thursday, November 26, 2015 7:50 AM
To: Thomas Freeman <thomas.freeman at hgst.com>; nvmewin-request at lists.openfabrics.org; nvmewin at lists.openfabrics.org
Cc: raymond.c.robles at intel.com; judy.brock at ssi.samsung.com; tru.nguyen at ssi.samsung.com
Subject: Re: nvmewin Digest, Vol 47, Issue 6


Hi Tom,

Sorry for the delay in reviewing your patch.



Please find below our review comments:

1) SCSIOP_PERSISTENT_RESERVE_IN and SCSIOP_PERSISTENT_RESERVE_OUT command opcodes needs to be added in SntiValidateNacaSetting().

2) For the default case in SntiValidateNacaSetting(), failure can be returned instead of success.  In future, if new command support is added, then that command opcode  needs to be added in SntiValidateNacaSetting(). If the support is not added in SntiValidateNacaSetting() and success is returned in default case, then this validation will be missed. Returning failure for the default case will catch these kind of issues.

3) For the DSM and Reservation commands, same memory buffer is used for both the commands.If both of these commands are issued in parallel, then synchronization issues may be observed.

4) NVMePreparePRPs() can be used in place of NVMeGetPhysAddr() in the function SntiTranslatePersistentReserveIn()[Line Number 3763]. NVMePreparePRPs() will take care of page alignment. So no need to check for page boundary crossing case of the memory buffer.[page boundary crossing verification is done in Line number 3777-3791]. The same can be applied in SntiTranslatePersistentReserveOut() [Line Number 4077]

5)  If the SntiTranslatePersReserveInResponse() returns the SNTI_SEQUENCE_IN_PROGRESS, then the SRB will not be completed. This value SNTI_SEQUENCE_IN_PROGRESS will be returned for default case.



Thanks,

Suman




------- Original Message -------
Sender : nvmewin-request at lists.openfabrics.org<nvmewin-request at lists.openfabrics.org>
Date   : Nov 26, 2015 04:21 (GMT+05:30)
Title  : nvmewin Digest, Vol 47, Issue 6

Send nvmewin mailing list submissions to
 nvmewin at lists.openfabrics.org

To subscribe or unsubscribe via the World Wide Web, visit
 http://lists.openfabrics.org/mailman/listinfo/nvmewin
or, via email, send a message with subject or body 'help' to
 nvmewin-request at lists.openfabrics.org

You can reach the person managing the list at
 nvmewin-owner at lists.openfabrics.org

When replying, please edit your Subject line so it is more specific
than "Re: Contents of nvmewin digest..."


Today's Topics:

   1. Re: Changes for SNTL 1.5 and SCSI Persistent Reserve In/Out
      (Robles, Raymond C)


----------------------------------------------------------------------

Message: 1
Date: Wed, 25 Nov 2015 22:51:52 +0000
From: "Robles, Raymond C" <raymond.c.robles at intel.com>
To: 'Thomas Freeman' <thomas.freeman at hgst.com>,
 "'nvmewin at lists.openfabrics.org'" <nvmewin at lists.openfabrics.org>
Subject: Re: [nvmewin] Changes for SNTL 1.5 and SCSI Persistent
 Reserve In/Out
Message-ID:
 <49158E750348AA499168FD41D88983607253C236 at fmsmsx117.amr.corp.intel.com>

Content-Type: text/plain; charset="us-ascii"


Hi Tom,

Sorry for the delay in retesting your updated patch. We have finished our review/tests and Intel approves this patch.

Now just need to wait approval from Samsung and PMC. Due to the holidays, we have a couple of options for this patch:


-          Wait for Samsung and PMC to come back with feedback/approval before pushing patch.

-          Push patch now (without approval from PMC and Samsung). Fix any remaining issues in future patches.

The second option allows us to move forward with the next patch from Intel for namespace management. The first option is still ideal as we would like to get as much feedback as possible.

Does anyone on this distribution have strong feelings for either option? Or an option I didn't describe?

Thanks,
Ray

From: Robles, Raymond C
Sent: Thursday, November 12, 2015 3:24 PM
To: Thomas Freeman; nvmewin at lists.openfabrics.org
Subject: RE: [nvmewin] Changes for SNTL 1.5 and SCSI Persistent Reserve In/Out

Thanks Tom! We'll rerun our tests.

Thanks,
Ray

From: Thomas Freeman [mailto:thomas.freeman at hgst.com]
Sent: Thursday, November 12, 2015 3:23 PM
To: Robles, Raymond C; nvmewin at lists.openfabrics.org<mailto:nvmewin at lists.openfabrics.org>
Subject: RE: [nvmewin] Changes for SNTL 1.5 and SCSI Persistent Reserve In/Out

Hi Ray,
There were issues with the IEKEY, RACQA, RRELA, RREGA and RTYPE.
I'm attaching a new file with those fixes - SNTL15ResInOut111215.zip (password "nvmehgst").

The code passes Scsi compliance testing (reservation is not enabled)
As before, I am doing a manual validation of PERSISTENT RESERVATION IN/OUT.

Tom Freeman
Software Engineer, Device Manager and Driver Development
HGST, a Western Digital company
thomas.freeman at hgst.com<mailto:thomas.freeman at hgst.com>
507-322-2311

[HGST_Logo_email]
3605 Hwy 52 N
Rochester, MN 55901
www.hgst.com<https://hgst.jiveon.com/external-link.jspa?url=http://www.hgst.com/>

From: Thomas Freeman
Sent: Tuesday, November 10, 2015 4:52 PM
To: 'Robles, Raymond C' <raymond.c.robles at intel.com<mailto:raymond.c.robles at intel.com>>; nvmewin at lists.openfabrics.org<mailto:nvmewin at lists.openfabrics.org>
Subject: RE: [nvmewin] Changes for SNTL 1.5 and SCSI Persistent Reserve In/Out

Hi Ray,
I don't have hardware to support NVM Reservation Register. So, my testing was limited to issuing individual Scsi Persistent Reserve commands and manually verifying the Scsi to NVMe translation.
As you observed, I missed the IEKEY field. My implementation was based on table 6-35 in the SNTL 1.5. That doesn't mention IEKEY. But, I overlooked the information about IEKEY from table 4-16.
I'll make that change and submit a new patch.

Tom Freeman
Software Engineer, Device Manager and Driver Development
HGST, a Western Digital company
thomas.freeman at hgst.com<mailto:thomas.freeman at hgst.com>
507-322-2311

[HGST_Logo_email]
3605 Hwy 52 N
Rochester, MN 55901
www.hgst.com<https://hgst.jiveon.com/external-link.jspa?url=http://www.hgst.com/>

From: Robles, Raymond C [mailto:raymond.c.robles at intel.com]
Sent: Tuesday, November 10, 2015 3:59 PM
To: Thomas Freeman <thomas.freeman at hgst.com<mailto:thomas.freeman at hgst.com>>; nvmewin at lists.openfabrics.org<mailto:nvmewin at lists.openfabrics.org>
Subject: RE: [nvmewin] Changes for SNTL 1.5 and SCSI Persistent Reserve In/Out

Hi Tom,

Intel has finished their internal testing of this patch. All tests passed with the exception of the MS SCSI Compliance Tests for Persistent Reserve In/Out (see attached file).

We're not sure if the failures are due to IEKEY field not being handled in Persistent Reserve Out command. Was this something that was intentionally left out? Were you able to try the SCSI compliance test?

Thanks,
Ray

From: nvmewin-bounces at lists.openfabrics.org<mailto:nvmewin-bounces at lists.openfabrics.org> [mailto:nvmewin-bounces at lists.openfabrics.org] On Behalf Of Robles, Raymond C
Sent: Monday, November 02, 2015 4:25 PM
To: Thomas Freeman; nvmewin at lists.openfabrics.org<mailto:nvmewin at lists.openfabrics.org>
Subject: Re: [nvmewin] Changes for SNTL 1.5 and SCSI Persistent Reserve In/Out

Hello all,

I extended the review period for this patch to allow for more time. We're coming up on the end of the year and I'd like to wrap up all planned patches.

Intel is wrapping up our review of this patch and will send out our feedback by the end of the week. Still awaiting feedback from Samsung and PMC. In order to keep to our original schedule, I'd like to wrap this patch up as soon as possible. Intel will be next with the namespace management patch.

Thanks,
Ray

From: nvmewin-bounces at lists.openfabrics.org<mailto:nvmewin-bounces at lists.openfabrics.org> [mailto:nvmewin-bounces at lists.openfabrics.org] On Behalf Of Robles, Raymond C
Sent: Wednesday, October 07, 2015 12:54 PM
To: Thomas Freeman; nvmewin at lists.openfabrics.org<mailto:nvmewin at lists.openfabrics.org>
Subject: Re: [nvmewin] Changes for SNTL 1.5 and SCSI Persistent Reserve In/Out

Tom,

Thanks for volunteering to submit this patch while Intel is wrapping up the namespace management patch.

All Reviewers,

Please provide feedback on HGST's patch by October 21st, 2015. Thanks!

Thanks,
Ray

From: Thomas Freeman [mailto:thomas.freeman at hgst.com]
Sent: Wednesday, October 07, 2015 12:45 PM
To: nvmewin at lists.openfabrics.org<mailto:nvmewin at lists.openfabrics.org>
Cc: Robles, Raymond C
Subject: Changes for SNTL 1.5 and SCSI Persistent Reserve In/Out

This patch includes changes to support SNTL version 1.5, SCSI Persistent Reserve In/Out and a variety of small fixes.

The following tests were successfully run on Windows 2008 R2, Windows 2012 and Windows 2012 R2
1 hour sdstress, 1 hour IOMETER, Quick and Slow format - MBR and GPT, Microsoft Scsi Compliance.

The above tests were run with an NVMe 1.1a compliant device that does not support NVMe Reservations.
Testing variations of Unit SN VPD page/Device ID Data Page (requiring V1.0 or NGUID support) and
SCSI Persistent Reserve In/Out, required setting breakpoints to manually alter the required Identify
data. At that point individual SCSI commands were issued and the driver's response (NVMe commands
issued and data returned) was manually verified.

The attached file, SNTL15ResInOut.zip, contains a Patch file, a copy of the source code and a Log file detailing the changes made.
The password for the file is "nvmehgst"

Tom Freeman
Software Engineer, Device Manager and Driver Development
HGST, a Western Digital company
thomas.freeman at hgst.com<mailto:thomas.freeman at hgst.com>
507-322-2311

[HGST_Logo_email]
3605 Hwy 52 N
Rochester, MN 55901
www.hgst.com<https://hgst.jiveon.com/external-link.jspa?url=http://www.hgst.com/>

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.openfabrics.org/pipermail/nvmewin/attachments/20151125/a06d1c5b/attachment.html>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: image003.gif
Type: image/gif
Size: 1117 bytes
Desc: image003.gif
URL: <http://lists.openfabrics.org/pipermail/nvmewin/attachments/20151125/a06d1c5b/attachment.gif>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: image004.png
Type: image/png
Size: 4274 bytes
Desc: image004.png
URL: <http://lists.openfabrics.org/pipermail/nvmewin/attachments/20151125/a06d1c5b/attachment.png>

------------------------------

_______________________________________________
nvmewin mailing list
nvmewin at lists.openfabrics.org
http://lists.openfabrics.org/mailman/listinfo/nvmewin


End of nvmewin Digest, Vol 47, Issue 6
**************************************

<p> </p><p> </p>
[Image removed by sender.]
HGST E-mail Confidentiality Notice & Disclaimer:
This e-mail and any files transmitted with it may contain confidential or legally privileged information of HGST and are intended solely for the use of the individual or entity to which they are addressed. If you are not the intended recipient, any disclosure, copying, distribution or any action taken or omitted to be taken in reliance on it, is prohibited.  If you have received this e-mail in error, please notify the sender immediately and delete the e-mail in its entirety from your system.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.openfabrics.org/pipermail/nvmewin/attachments/20151130/7622fca5/attachment.html>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: ~WRD000.jpg
Type: image/jpeg
Size: 823 bytes
Desc: ~WRD000.jpg
URL: <http://lists.openfabrics.org/pipermail/nvmewin/attachments/20151130/7622fca5/attachment.jpg>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: image001.png
Type: image/png
Size: 4274 bytes
Desc: image001.png
URL: <http://lists.openfabrics.org/pipermail/nvmewin/attachments/20151130/7622fca5/attachment.png>

More information about the nvmewin mailing list