From raymond.c.robles at intel.com  Tue Mar  3 14:38:00 2015
From: raymond.c.robles at intel.com (Robles, Raymond C)
Date: Tue, 3 Mar 2015 22:38:00 +0000
Subject: [nvmewin] OFA - 2015 Planning Meeting (2/27/2015) - Meeting Minutes
Message-ID: <49158E750348AA499168FD41D88983606B5F3740@fmsmsx117.amr.corp.intel.com>

OFA - 2015 Planning Meeting (2/27/2015)

Friday, February 27, 2015

1:54 PM

Agenda:
*         3rd and 4th reviewing companies
*         1.1/1.2 features
*         Releases
*         WHQL strategy



Attendees:
*         Ray, Carolyn (Intel)
*         Tom (HGST)
*         Austin (Dell)
*         Uma (SK Hynix)
*         Suman (Samsung)
*         <numerous others unidentified>



AR's from last meeting:
*         Dell (Austin): Identify initial WHQL strategy for 2015 (mid-February... next meeting to discuss).
o    Completed (2/27/2015): 2015 Releases will follow same WHQL strategy as 2014 releases.
*         Samsung (Judy): Follow up with management on becoming 3rd mandatory reviewing member company for OFA (2 weeks).
o    Completed (1/30/15): Samsung has agreed to become 3rd mandatory reviewing member company. Thanks Judy!
*         HGST (Ken): Follow up with management on becoming 3rd mandatory reviewing member company for OFA (2 weeks).
o    Completed (2/27/2015): HGST will be 4th reviewing company (code review only, no testing). Thanks Tom!
*         Intel (Ray): Provide preliminary list of 1.1/1.2 features, both mandatory and optional, to start discussion on what to include in 2015 releases.
o    Completed (2/27/2015): Ray sent out preliminary list for working group to review prior to meeting. Discussed list in last meeting (see minutes below).
*         Intel (Ray): Communicate with Ken Strandberg at OFA website for logistical changes (mid-February... next meeting).
o    Completed (2/12/2015): OFA executive team list updated.



Meeting Minutes:
*         3rd and 4th Reviewing Company - Samsung (Judy) and HGST (Tom)
o    Samsung: Code reviews and testing of patches
o    HGST: Code reviews only.
*         1.1/1.2 feature review and discussion
o    1.1 features
*         Tom (HGST): Would like to see all 1.1 features (except SGL) in OFA
*         General Acceptance of all features, except SGL...
*         Ray (Intel): SGL use case model of metadata buffer and host buffer...
o    1.2 features
*         Austin (Dell): AER: in the driver or not?
?  Having it in the driver, you don't limit this to just one host process/application
?  User WMI notification (mechanism) for listening to AER events
*         Tom (HGST): Atomicity: how does host know to set and use parameters correctly
?  Implication in the driver seems to be non-existent
?  Do we just punt on this feature?
*         Austin (Dell): Would like to see
?  Namespace management (optional)
?  Live firmware update (optional)
*         Can we make this seamless and not disruptive?
*         Adding this feature requires an arch decision on whether to re-enumerate or not
*         Re-enumeration is *almost* equivalent to a reset, which defeats purpose of live FW update
*         No re-enumeration is *dangerous* as some critical metadata may have changed (e.g. SN, NSZE, NCAP, NN, etc.)
o    Multipath features (requested by HGST)
o    Namespace Management (requested by Dell)
o    Live firmware Update (requested by Dell)
o    Atomicity Enhancements (requested by HGST)
*         2015 OFA releases will continue on same cadence as 2014 releases (Q2 and Q4 release)
*         WHQL: Each official release will undergo WHQL test suite. No submission to Microsoft necessary, but all WHQL tests must pass.



Candidate Features (2015)


*         Start I/O Concurrent Channels
*         Instead of learning cores, use perf opts. Shave off 2 seconds in init code... using Storport initialize perf opts. No need for source core interrupt steering code in driver.
*         Storport Notification usage. Hot plug and IOCTLs, this does not work, for Win8.1... StorportRequestTimer() function needed.
*         CFS bit monitoring and handling. Look into adding additional handling code just to monitor CSTS.CFS.
*         Read only support for devices at EOL. Detection at init, hot plug, or run-time.
*         Multipath feature set from 1.1 spec (Multi-path I/O, Namespace Sharing, Reservations)
*         Active/idle RTD3
*         Temperature thresholds (Get/Set Feature and AER trigger)
*         Namespace management (push to Q4 release)
*         Live firmware update
*         Atomicity Enhancements
*         WHQL test suite run (any bug fixes) - for each release of OFA
*         Win 10 support (push to Q4 release)



AR's:
*         Intel (Ray): Provide better customer/business case for SGL support in driver (3/20/2015)
*         Dell (Austin): Provide quick overview of how WMI can be used for AER async reporting to subscribed user apps (3/20/2015)
*         All: Requesting volunteers for patch submissions from candidate feature list for 2015.



Created with Microsoft OneNote 2013.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.openfabrics.org/pipermail/nvmewin/attachments/20150303/dd7a38ed/attachment.html>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: OFA - 2015 Planning Meeting (227.one
Type: application/msonenote
Size: 35552 bytes
Desc: OFA - 2015 Planning Meeting (227.one
URL: <http://lists.openfabrics.org/pipermail/nvmewin/attachments/20150303/dd7a38ed/attachment.bin>

From raymond.c.robles at intel.com  Tue Mar  3 15:03:05 2015
From: raymond.c.robles at intel.com (Robles, Raymond C)
Date: Tue, 3 Mar 2015 23:03:05 +0000
Subject: [nvmewin] OFA - 2015 Planning Meeting (2/27/2015) - Meeting
	Minutes
In-Reply-To: <49158E750348AA499168FD41D88983606B5F3740@fmsmsx117.amr.corp.intel.com>
References: <49158E750348AA499168FD41D88983606B5F3740@fmsmsx117.amr.corp.intel.com>
Message-ID: <49158E750348AA499168FD41D88983606B5F382F@fmsmsx117.amr.corp.intel.com>

Update below in BOLD on request for Namespace Management (request is for Q2 release, not Q4... error on my part in putting together meeting minutes).

Thanks to Tom for letting me know of the error.

Thanks,
Ray

From: Robles, Raymond C
Sent: Tuesday, March 03, 2015 3:38 PM
To: nvmewin at lists.openfabrics.org
Cc: Huffman, Amber; Robles, Raymond C
Subject: OFA - 2015 Planning Meeting (2/27/2015) - Meeting Minutes


OFA - 2015 Planning Meeting (2/27/2015)

Friday, February 27, 2015

1:54 PM

Agenda:
*         3rd and 4th reviewing companies
*         1.1/1.2 features
*         Releases
*         WHQL strategy



Attendees:
*         Ray, Carolyn (Intel)
*         Tom (HGST)
*         Austin (Dell)
*         Uma (SK Hynix)
*         Suman (Samsung)
*         <numerous others unidentified>



AR's from last meeting:
*         Dell (Austin): Identify initial WHQL strategy for 2015 (mid-February... next meeting to discuss).
o    Completed (2/27/2015): 2015 Releases will follow same WHQL strategy as 2014 releases.
*         Samsung (Judy): Follow up with management on becoming 3rd mandatory reviewing member company for OFA (2 weeks).
o    Completed (1/30/15): Samsung has agreed to become 3rd mandatory reviewing member company. Thanks Judy!
*         HGST (Ken): Follow up with management on becoming 3rd mandatory reviewing member company for OFA (2 weeks).
o    Completed (2/27/2015): HGST will be 4th reviewing company (code review only, no testing). Thanks Tom!
*         Intel (Ray): Provide preliminary list of 1.1/1.2 features, both mandatory and optional, to start discussion on what to include in 2015 releases.
o    Completed (2/27/2015): Ray sent out preliminary list for working group to review prior to meeting. Discussed list in last meeting (see minutes below).
*         Intel (Ray): Communicate with Ken Strandberg at OFA website for logistical changes (mid-February... next meeting).
o    Completed (2/12/2015): OFA executive team list updated.



Meeting Minutes:
*         3rd and 4th Reviewing Company - Samsung (Judy) and HGST (Tom)
o    Samsung: Code reviews and testing of patches
o    HGST: Code reviews only.
*         1.1/1.2 feature review and discussion
o    1.1 features
*         Tom (HGST): Would like to see all 1.1 features (except SGL) in OFA
*         General Acceptance of all features, except SGL...
*         Ray (Intel): SGL use case model of metadata buffer and host buffer...
o    1.2 features
*         Austin (Dell): AER: in the driver or not?
?  Having it in the driver, you don't limit this to just one host process/application
?  User WMI notification (mechanism) for listening to AER events
*         Tom (HGST): Atomicity: how does host know to set and use parameters correctly
?  Implication in the driver seems to be non-existent
?  Do we just punt on this feature?
*         Austin (Dell): Would like to see
?  Namespace management (optional)
?  Live firmware update (optional)
*         Can we make this seamless and not disruptive?
*         Adding this feature requires an arch decision on whether to re-enumerate or not
*         Re-enumeration is *almost* equivalent to a reset, which defeats purpose of live FW update
*         No re-enumeration is *dangerous* as some critical metadata may have changed (e.g. SN, NSZE, NCAP, NN, etc.)
o    Multipath features (requested by HGST)
o    Namespace Management (requested by Dell)
o    Live firmware Update (requested by Dell)
o    Atomicity Enhancements (requested by HGST)
*         2015 OFA releases will continue on same cadence as 2014 releases (Q2 and Q4 release)
*         WHQL: Each official release will undergo WHQL test suite. No submission to Microsoft necessary, but all WHQL tests must pass.



Candidate Features (2015)


*         Start I/O Concurrent Channels
*         Instead of learning cores, use perf opts. Shave off 2 seconds in init code... using Storport initialize perf opts. No need for source core interrupt steering code in driver.
*         Storport Notification usage. Hot plug and IOCTLs, this does not work, for Win8.1... StorportRequestTimer() function needed.
*         CFS bit monitoring and handling. Look into adding additional handling code just to monitor CSTS.CFS.
*         Read only support for devices at EOL. Detection at init, hot plug, or run-time.
*         Multipath feature set from 1.1 spec (Multi-path I/O, Namespace Sharing, Reservations)
*         Active/idle RTD3
*         Temperature thresholds (Get/Set Feature and AER trigger)
*         Namespace management (request for Q2 release)
*         Live firmware update
*         Atomicity Enhancements
*         WHQL test suite run (any bug fixes) - for each release of OFA
*         Win 10 support (push to Q4 release)



AR's:
*         Intel (Ray): Provide better customer/business case for SGL support in driver (3/20/2015)
*         Dell (Austin): Provide quick overview of how WMI can be used for AER async reporting to subscribed user apps (3/20/2015)
*         All: Requesting volunteers for patch submissions from candidate feature list for 2015.



Created with Microsoft OneNote 2013.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.openfabrics.org/pipermail/nvmewin/attachments/20150303/c0f3bad0/attachment.html>

From cheng.peng at memblaze.com  Wed Mar  4 22:50:19 2015
From: cheng.peng at memblaze.com (cheng.peng at memblaze.com)
Date: Thu, 5 Mar 2015 14:50:19 +0800
Subject: [nvmewin] don't create the disk device
Message-ID: <201503051450187766495@memblaze.com>

when driver is loaded, ScsiAdapter device is created, but disk device isn't created.

00000000 0.00000000 STORMINI: StorPortInitialize returns Status(0x0) 
00000001 0.02776694 STORMINI: Access Range, VirtualAddr=0xFFFFF8800654F000. 
00000002 0.02778550 STORMINI: NVMeFindAdapter: Stride Size set to 0x4 
00000003 0.02785815 STORMINI: NVMeEnumNumaCores: # of NUMA node(s) = 1. 
00000004 0.02786289 STORMINI: NVMeEnumNumaCores: Number of groups = 1. 
00000005 0.02786644 STORMINI: NVMeEnumNumaCores: Group(0) affinity mask(0x3). 
00000006 0.02786920 STORMINI: NVMeEnumNumaCores: Total 2 logical processors 
00000007 0.02787118 STORMINI: NVMeEnumNumaCores: NUMA Node#0 
00000008 0.02787394 STORMINI: Core mask is 0x3 in Group(0) 
00000009 0.02787749 STORMINI: There are 2 cores in Node#0. 
00000010 0.02787986 STORMINI: The total number of CPU cores 2. 
00000011 0.02829325 STORMINI: NVMeEnumMsiMessages: Msg granted=2 
00000012 0.02831062 STORMINI: NVMeAllocateMem: Succeeded! 
00000013 0.02832010 STORMINI: NVMeAllocateMem: Succeeded! 
00000014 0.02833590 STORMINI: NVMeInitSubQueue : SQ 0x0 pSubTDBL 0xFFFFF88006550000 at index 0x0 
00000015 0.02835406 STORMINI: NVMeInitCplQueue : CQ 0x0 pCplHDBL 0xFFFFF88006550004 at index 0x1 
00000016 0.02836788 STORMINI: NVMeEnableAdapter: Setting EN... 
00000017 0.02837617 STORMINI: NVMeAllocateMem: Succeeded! 
00000018 0.02838446 STORMINI: NVMeAllocateMem: Succeeded! 
00000019 0.02874849 STORMINI: NVMeRunningWaitOnRDY: Waiting... 
00000020 0.03745769 STORMINI: NVMeRunningWaitOnRDY: RDY has been set 
00000021 0.05304605 STORMINI: NVMeGetIdentifyStructures: IDEN_CONTROLLER 
00000022 0.05310448 STORMINI: NVMeInitCallback: Driver state: 33 
00000023 0.06864980 STORMINI: NVMeGetIdentifyStructures: IDEN_NAMESPACE 
00000024 0.06865493 STORMINI: NVMeGetIdentifyStructures: Get NS INFO for NSID 0x1 tgt lun 0x0 
00000025 0.06867468 STORMINI: NVMeInitCallback: Driver state: 34 
00000026 0.08686497 STORMINI: NVMeInitCallback: Driver state: 35 
00000027 0.09989244 STORMINI: NVMeInitCallback: Driver state: 35 
00000028 0.11546224 STORMINI: NVMeInitCallback: Driver state: 35 
00000029 0.11546777 STORMINI: pLbaRangeTypeEntry type : 0x0 lun id 0 nsid 0x1 
00000030 0.11547054 STORMINI: NVMeSetFeaturesCompletion: FYI LnuExt at 0 has been cleared (NSID not visible) 
00000031 0.13106994 STORMINI: NVMeAllocateMem: Succeeded! 
00000032 0.13109048 STORMINI: NVMeAllocateMem: Succeeded! 
00000033 0.13111298 STORMINI: NVMeAllocIoQueues: Core 0x0 ---> QueueID 0x1 
00000034 0.13112088 STORMINI: NVMeAllocateMem: Succeeded! 
00000035 0.13113983 STORMINI: NVMeAllocateMem: Succeeded! 
00000036 0.13116075 STORMINI: NVMeAllocIoQueues: Core 0x1 ---> QueueID 0x2 
00000037 0.13116352 STORMINI: NVMeMsiMapCores: Core(0x0)Msg#(0x0) 
00000038 0.13116549 STORMINI: NVMeMsiMapCores: Core(0x1)Msg#(0x1) 
00000039 0.13117695 STORMINI: NVMeInitSubQueue : SQ 0x1 pSubTDBL 0xFFFFF88006550008 at index 0x2 
00000040 0.13121486 STORMINI: NVMeInitSubQueue : SQ 0x2 pSubTDBL 0xFFFFF88006550010 at index 0x4 
00000041 0.13125157 STORMINI: NVMeInitCplQueue : CQ 0x1 pCplHDBL 0xFFFFF8800655000C at index 0x3 
00000042 0.13126460 STORMINI: NVMeInitCplQueue : CQ 0x2 pCplHDBL 0xFFFFF88006550014 at index 0x5 
00000043 0.14669226 STORMINI: NVMeInitCallback: Driver state: 38 
00000044 0.16231970 STORMINI: NVMeInitCallback: Driver state: 38 
00000045 0.17787489 STORMINI: NVMeInitCallback: Driver state: 39 
00000046 0.19351892 STORMINI: NVMeInitCallback: Driver state: 39 
00000047 0.22342743 STORMINI: FYI: SRB status 0x6 scsi 0x0 for CDB 0xa0 BTL 0 0 0 
00000048 0.22343650 STORMINI: FYI: SRB status 0x6 scsi 0x0 for CDB 0xa0 BTL 0 0 0 
00000049 0.22344086 STORMINI: FYI: SRB status 0x6 scsi 0x0 for CDB 0xa0 BTL 0 0 0 
00000050 0.22344598 STORMINI: FYI: SRB status 0x6 scsi 0x0 for CDB 0x12 BTL 0 0 0 
00000051 0.22344993 STORMINI: FYI: SRB status 0x6 scsi 0x0 for CDB 0x12 BTL 0 0 0 
00000052 0.22345348 STORMINI: FYI: SRB status 0x6 scsi 0x0 for CDB 0x12 BTL 0 0 0 
00000053 0.22369038 STORMINI: FYI: SRB status 0x6 scsi 0x8 for CDB 0x0 BTL 0 0 0 
00000054 0.22462179 STORMINI: NVMeRunning: StorPortReady... 

please tell me the reason, thank you 



cheng.peng at memblaze.com
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.openfabrics.org/pipermail/nvmewin/attachments/20150305/22bda209/attachment.html>

From cheng.peng at memblaze.com  Fri Mar 13 00:27:38 2015
From: cheng.peng at memblaze.com (cheng.peng at memblaze.com)
Date: Fri, 13 Mar 2015 15:27:38 +0800
Subject: [nvmewin] out of memory in the NVMePreparePRPs function
Message-ID: <2015031315273640372817@memblaze.com>

Hi,

I get a BSOD when the parameter TxLength of the NVMePreparePRPs function is assigned to a big number(e.g. 1572864)

In this case，pSrbExt->numberOfPrpEntries will be a wrong number due to out of memory

I have temporarily solved this problem by setting MAX_TX_SIZE to (2*1024*1024)

I think there should be a better solution

thank you



cheng.peng at memblaze.com
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.openfabrics.org/pipermail/nvmewin/attachments/20150313/6e2ff386/attachment.html>