[nvmewin] Is NVMe Generic Devices supported?

Fri Dec 18 19:22:56 PST 2015

Hi guys thanks for the feedbacks.

I am using windows 7 so the available hot fix on Microsoft's site to
add nvme driver natively is outdated. The driver version date is from
2006 (didn't knew NVMe Spec was that old). After the hot fix I
downloaded and installed Release 1.4 from SVN and had to manually
force the device driver to be Release 1.4 otherwise it was just using
the stornvme from the hot fix.

It works as a charm now!

So is windows 7 updates going to receive new driver releases?
Also 32bit SVN Release 1.4 reports driver version 1.3, is it ok?

Thanks!

Regards,
Angelo Silva Brito.
B.S. in Computer Engineering - UFPE Brazil
http://about.me/angelobrito
_________________________________________________

2015-12-18 19:55 GMT-03:00 Mike Berhan <mikeb at bustrace.com>:
> Since the NVMe driver source code is available, can you just build it and single step through the initialization sequence to see where it's failing?  If not, perhaps the WPP tracing capability built into the OFA driver can be engaged.
>
> Mike
>
> -----Original Message-----
> From: nvmewin-bounces at lists.openfabrics.org [mailto:nvmewin-bounces at lists.openfabrics.org] On Behalf Of Angelo Brito
> Sent: Friday, December 18, 2015 1:37 PM
> To: Iuliu Rus <iuliur at google.com>
> Cc: nvmewin <nvmewin at lists.openfabrics.org>
> Subject: Re: [nvmewin] Is NVMe Generic Devices supported?
>
> Hi Iuliu,
>
> I haven't seen any commands. No write were sent to my controller other than the MSIX configuration. Not even the Admin Queue's registers. I suspect that for some reason windows NVMe Driver is dropping off my controller after the PCIe configuration and windows events wasn't of much help.
> The events reported:
> - System
> - Provider
> [ Name] stornvme
> - EventID 11
> [ Qualifiers] 49156
> Level 2
> Task 0
> Keywords 0x80000000000000
> - TimeCreated
> [ SystemTime] 2015-12-18T18:28:31.145251600Z EventRecordID 2443 Channel System Computer schenberg-PC Security
> - EventData
> \Device\RaidPort3
> 0F00240001000000000000000B0004C002000000000000000000000000000000000000000000000001000000200004000B0004C0020000000000000000000000000000000000000004000000
> ________________________________
>
> Dados binários:
>
> Em Palavras
>
> 0000: 0024000F 00000001 00000000 C004000B
> 0008: 00000002 00000000 00000000 00000000
> 0010: 00000000 00000000 00000001 00040020
> 0018: C004000B 00000002 00000000 00000000
> 0020: 00000000 00000000 00000004
>
> Em Bytes
>
> 0000: 0F 00 24 00 01 00 00 00 ..$.....
> 0008: 00 00 00 00 0B 00 04 C0 .......À
> 0010: 02 00 00 00 00 00 00 00 ........
> 0018: 00 00 00 00 00 00 00 00 ........
> 0020: 00 00 00 00 00 00 00 00 ........
> 0028: 01 00 00 00 20 00 04 00 .... ...
> 0030: 0B 00 04 C0 02 00 00 00 ...À....
> 0038: 00 00 00 00 00 00 00 00 ........
> 0040: 00 00 00 00 00 00 00 00 ........
> 0048: 04 00 00 00 ....
>
> This pages suggest it is a hardware problem tough it works fine on Linux:
> https://support.microsoft.com/en-us/kb/154690
>
> Regards,
> Angelo Silva Brito.
> B.S. in Computer Engineering - UFPE Brazil http://about.me/angelobrito _________________________________________________
>
>
> 2015-12-18 14:29 GMT-05:00 Iuliu Rus <iuliur at google.com>:
>> Last time this happened to us it was because the device was responding
>> with an IdentifyController structure which had no firmware_revision
>> field and the fields were padded with zeroes instead of spaces. Does
>> your device get any commands on the admin queue from windows?
>>
>>
>> On Fri, Dec 18, 2015 at 11:10 AM, Angelo Brito <asb at cin.ufpe.br> wrote:
>>>
>>> Hello All,
>>>
>>> I have a generic NVMe device that works perfectly on Linux but does
>>> not work on Windows. The device is recognized as a Default NVMe
>>> Controller with Microsoft as vendor but issue an error code 10: This
>>> device cannot start. I am aware that this is a generic error code and
>>> could mean any thing.
>>>
>>> I have started to debug from the device perspective and the Host CPU
>>> just writes the MSI configuration registers during startup. I use
>>> Altera´s PCIe Hard IP to implement the PCIe and my controller is
>>> hosted on FPGA. The device was never started.
>>>
>>> Someone can share any more leads to debug? Why windows doesn´t start
>>> my device?
>>>
>>> Regards,
>>> Angelo Silva Brito.
>>> B.S. in Computer Engineering - UFPE Brazil
>>> http://about.me/angelobrito
>>> _________________________________________________
>>> _______________________________________________
>>> nvmewin mailing list
>>> nvmewin at lists.openfabrics.org
>>> http://lists.openfabrics.org/mailman/listinfo/nvmewin
>>
>>
>
>
> Regards,
> Angelo Silva Brito.
> B.S. in Computer Engineering - UFPE Brazil http://about.me/angelobrito _________________________________________________
>
>
> 2015-12-18 17:29 GMT-02:00 Iuliu Rus <iuliur at google.com>:
>> Last time this happened to us it was because the device was responding
>> with an IdentifyController structure which had no firmware_revision
>> field and the fields were padded with zeroes instead of spaces. Does
>> your device get any commands on the admin queue from windows?
>>
>>
>> On Fri, Dec 18, 2015 at 11:10 AM, Angelo Brito <asb at cin.ufpe.br> wrote:
>>>
>>> Hello All,
>>>
>>> I have a generic NVMe device that works perfectly on Linux but does
>>> not work on Windows. The device is recognized as a Default NVMe
>>> Controller with Microsoft as vendor but issue an error code 10: This
>>> device cannot start. I am aware that this is a generic error code and
>>> could mean any thing.
>>>
>>> I have started to debug from the device perspective and the Host CPU
>>> just writes the MSI configuration registers during startup. I use
>>> Altera´s PCIe Hard IP to implement the PCIe and my controller is
>>> hosted on FPGA. The device was never started.
>>>
>>> Someone can share any more leads to debug? Why windows doesn´t start
>>> my device?
>>>
>>> Regards,
>>> Angelo Silva Brito.
>>> B.S. in Computer Engineering - UFPE Brazil
>>> http://about.me/angelobrito
>>> _________________________________________________
>>> _______________________________________________
>>> nvmewin mailing list
>>> nvmewin at lists.openfabrics.org
>>> http://lists.openfabrics.org/mailman/listinfo/nvmewin
>>
>>
>
>
> Regards,
> Angelo Silva Brito.
> B.S. in Computer Engineering - UFPE Brazil http://about.me/angelobrito _________________________________________________
>
>
> 2015-12-18 14:29 GMT-05:00 Iuliu Rus <iuliur at google.com>:
>> Last time this happened to us it was because the device was responding
>> with an IdentifyController structure which had no firmware_revision
>> field and the fields were padded with zeroes instead of spaces. Does
>> your device get any commands on the admin queue from windows?
>>
>>
>> On Fri, Dec 18, 2015 at 11:10 AM, Angelo Brito <asb at cin.ufpe.br> wrote:
>>>
>>> Hello All,
>>>
>>> I have a generic NVMe device that works perfectly on Linux but does
>>> not work on Windows. The device is recognized as a Default NVMe
>>> Controller with Microsoft as vendor but issue an error code 10: This
>>> device cannot start. I am aware that this is a generic error code and
>>> could mean any thing.
>>>
>>> I have started to debug from the device perspective and the Host CPU
>>> just writes the MSI configuration registers during startup. I use
>>> Altera´s PCIe Hard IP to implement the PCIe and my controller is
>>> hosted on FPGA. The device was never started.
>>>
>>> Someone can share any more leads to debug? Why windows doesn´t start
>>> my device?
>>>
>>> Regards,
>>> Angelo Silva Brito.
>>> B.S. in Computer Engineering - UFPE Brazil
>>> http://about.me/angelobrito
>>> _________________________________________________
>>> _______________________________________________
>>> nvmewin mailing list
>>> nvmewin at lists.openfabrics.org
>>> http://lists.openfabrics.org/mailman/listinfo/nvmewin
>>
>>
> _______________________________________________
> nvmewin mailing list
> nvmewin at lists.openfabrics.org
> http://lists.openfabrics.org/mailman/listinfo/nvmewin
>