NVMe driver status

NVMe driver status

[Sunad Bhandary]

Hi Keith,

Currently there are two different variants of the NVMe driver, one with the
multi-queue implementation which is under active development and 
the one without it which is the current driver hosted in the NVMe git page. 

Going forward will the non-multi queue  driver also be maintained or will it
be done away with completely ?

Regards,
Sunad

NVMe driver status

[Keith Busch]

Hi Sunad,

The mainline going forward with kernel 3.19 will be blk-mq, and the
linux-nvme tree will be there as well on the next upstream merge.

I have my own repo for maintaining the bio-based version since a lot of
OS vendors are using older kernels in their minor release cycles. I can
push this out to a public repo on git.infradead.org if there's interest.

Thanks,
Keith

On Mon, 19 Jan 2015, Sunad Bhandary wrote:
> Hi Keith,
>
> Currently there are two different variants of the NVMe driver, one with the
> multi-queue implementation which is under active development and
> the one without it which is the current driver hosted in the NVMe git page.
>
> Going forward will the non-multi queue  driver also be maintained or will it
> be done away with completely ?
>
> Regards,
> Sunad

NVMe driver status

[Sunad Bhandary]

Hi Keith,

Thanks for clarifying the situation. 

If suppose the bio-based driver version is pushed out to a public repo, will
it be made  just a source repo or 
will it be actively maintained where new patches are acceptable and can be
merged into the repo ?

Thanks and regards,
Sunad

-----Original Message-----
From: Linux-nvme [mailto:linux-nvme-bounces@lists.infradead.org] On Behalf
Of Keith Busch
Sent: Monday, January 19, 2015 8:03 PM
To: Sunad Bhandary
Cc: 'Keith Busch'; linux-nvme at lists.infradead.org
Subject: Re: NVMe driver status

Hi Sunad,

The mainline going forward with kernel 3.19 will be blk-mq, and the
linux-nvme tree will be there as well on the next upstream merge.

I have my own repo for maintaining the bio-based version since a lot of OS
vendors are using older kernels in their minor release cycles. I can push
this out to a public repo on git.infradead.org if there's interest.

Thanks,
Keith

On Mon, 19 Jan 2015, Sunad Bhandary wrote:
> Hi Keith,
>
> Currently there are two different variants of the NVMe driver, one 
> with the multi-queue implementation which is under active development 
> and the one without it which is the current driver hosted in the NVMe git
page.
>
> Going forward will the non-multi queue  driver also be maintained or 
> will it be done away with completely ?
>
> Regards,
> Sunad

_______________________________________________
Linux-nvme mailing list
Linux-nvme at lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-nvme

NVMe driver status

[Keith Busch]

Okay, I added the repo to here:

http://git.infradead.org/users/kbusch/nvme-legacy.git

It's based off linux-stable v3.18.3, plus a merge from linux-nvme for
nvme parts that were not committed upstream by 3.18.

I don't know what kind of activity to expect this will get, but I'll be
happy to merge applicable upstream commits and apply patches for any new
features or bug fixes people want to see in the bio-based nvme driver.

On Tue, 20 Jan 2015, Sunad Bhandary wrote:
> Hi Keith,
>
> Thanks for clarifying the situation.
>
> If suppose the bio-based driver version is pushed out to a public repo, will
> it be made  just a source repo or
> will it be actively maintained where new patches are acceptable and can be
> merged into the repo ?
>
> Thanks and regards,
> Sunad
>
> -----Original Message-----
> From: Linux-nvme [mailto:linux-nvme-bounces at lists.infradead.org] On Behalf
> Of Keith Busch
> Sent: Monday, January 19, 2015 8:03 PM
> To: Sunad Bhandary
> Cc: 'Keith Busch'; linux-nvme at lists.infradead.org
> Subject: Re: NVMe driver status
>
> Hi Sunad,
>
> The mainline going forward with kernel 3.19 will be blk-mq, and the
> linux-nvme tree will be there as well on the next upstream merge.
>
> I have my own repo for maintaining the bio-based version since a lot of OS
> vendors are using older kernels in their minor release cycles. I can push
> this out to a public repo on git.infradead.org if there's interest.
>
> Thanks,
> Keith
>
> On Mon, 19 Jan 2015, Sunad Bhandary wrote:
>> Hi Keith,
>>
>> Currently there are two different variants of the NVMe driver, one
>> with the multi-queue implementation which is under active development
>> and the one without it which is the current driver hosted in the NVMe git
> page.
>>
>> Going forward will the non-multi queue  driver also be maintained or
>> will it be done away with completely ?
>>
>> Regards,
>> Sunad
>
> _______________________________________________
> Linux-nvme mailing list
> Linux-nvme at lists.infradead.org
> http://lists.infradead.org/mailman/listinfo/linux-nvme

NVMe driver status

[Sunad Bhandary]

Hi Keith,
Can we initiate some activity on the legacy-nvme maintenance?
 
Can you commit the async controller probe patch for the legacy driver also?
Or I can do this if you want.
Also, I'll be submitting the legacy-nvme version for a couple of patches
that we already submitted (write_long, split i/o in IOCtl path).

Going forward, how will the legacy-nvme be maintained?
The legacy-nvme was removed in 3.19. So are we going to have some
restrictions on the kernel versions supported by legacy nvme driver

Regards,
Sunad

-----Original Message-----
From: Linux-nvme [mailto:linux-nvme-bounces@lists.infradead.org] On Behalf
Of Keith Busch
Sent: Tuesday, January 20, 2015 9:17 PM
To: Sunad Bhandary
Cc: 'Keith Busch'; linux-nvme at lists.infradead.org
Subject: RE: NVMe driver status

Okay, I added the repo to here:

http://git.infradead.org/users/kbusch/nvme-legacy.git

It's based off linux-stable v3.18.3, plus a merge from linux-nvme for nvme
parts that were not committed upstream by 3.18.

I don't know what kind of activity to expect this will get, but I'll be
happy to merge applicable upstream commits and apply patches for any new
features or bug fixes people want to see in the bio-based nvme driver.

On Tue, 20 Jan 2015, Sunad Bhandary wrote:
> Hi Keith,
>
> Thanks for clarifying the situation.
>
> If suppose the bio-based driver version is pushed out to a public 
> repo, will it be made  just a source repo or will it be actively 
> maintained where new patches are acceptable and can be merged into the 
> repo ?
>
> Thanks and regards,
> Sunad
>
> -----Original Message-----
> From: Linux-nvme [mailto:linux-nvme-bounces at lists.infradead.org] On 
> Behalf Of Keith Busch
> Sent: Monday, January 19, 2015 8:03 PM
> To: Sunad Bhandary
> Cc: 'Keith Busch'; linux-nvme at lists.infradead.org
> Subject: Re: NVMe driver status
>
> Hi Sunad,
>
> The mainline going forward with kernel 3.19 will be blk-mq, and the 
> linux-nvme tree will be there as well on the next upstream merge.
>
> I have my own repo for maintaining the bio-based version since a lot 
> of OS vendors are using older kernels in their minor release cycles. I 
> can push this out to a public repo on git.infradead.org if there's
interest.
>
> Thanks,
> Keith
>
> On Mon, 19 Jan 2015, Sunad Bhandary wrote:
>> Hi Keith,
>>
>> Currently there are two different variants of the NVMe driver, one 
>> with the multi-queue implementation which is under active development 
>> and the one without it which is the current driver hosted in the NVMe 
>> git
> page.
>>
>> Going forward will the non-multi queue  driver also be maintained or 
>> will it be done away with completely ?
>>
>> Regards,
>> Sunad
>
> _______________________________________________
> Linux-nvme mailing list
> Linux-nvme at lists.infradead.org
> http://lists.infradead.org/mailman/listinfo/linux-nvme

_______________________________________________
Linux-nvme mailing list
Linux-nvme at lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-nvme

NVMe driver status

[David Darrington]

I'm interested in this also. I know that the there is a quarterly planning call for the ofed Windows NVMe driver. Would there be any interest in having occasional conference calls to discuss items such as these? I would be glad to setup a call, as long as Keith and/or Mathew agree that it is worthwhile and can attend.

-----Original Message-----
From: Linux-nvme [mailto:linux-nvme-bounces@lists.infradead.org] On Behalf Of Sunad Bhandary
Sent: Tuesday, March 10, 2015 10:21 AM
To: 'Keith Busch'
Cc: linux-nvme at lists.infradead.org
Subject: RE: NVMe driver status

Hi Keith,
Can we initiate some activity on the legacy-nvme maintenance?
 
Can you commit the async controller probe patch for the legacy driver also?
Or I can do this if you want.
Also, I'll be submitting the legacy-nvme version for a couple of patches that we already submitted (write_long, split i/o in IOCtl path).

Going forward, how will the legacy-nvme be maintained?
The legacy-nvme was removed in 3.19. So are we going to have some restrictions on the kernel versions supported by legacy nvme driver

Regards,
Sunad

-----Original Message-----
From: Linux-nvme [mailto:linux-nvme-bounces@lists.infradead.org] On Behalf Of Keith Busch
Sent: Tuesday, January 20, 2015 9:17 PM
To: Sunad Bhandary
Cc: 'Keith Busch'; linux-nvme at lists.infradead.org
Subject: RE: NVMe driver status

Okay, I added the repo to here:

http://git.infradead.org/users/kbusch/nvme-legacy.git

It's based off linux-stable v3.18.3, plus a merge from linux-nvme for nvme parts that were not committed upstream by 3.18.

I don't know what kind of activity to expect this will get, but I'll be happy to merge applicable upstream commits and apply patches for any new features or bug fixes people want to see in the bio-based nvme driver.

On Tue, 20 Jan 2015, Sunad Bhandary wrote:
> Hi Keith,
>
> Thanks for clarifying the situation.
>
> If suppose the bio-based driver version is pushed out to a public 
> repo, will it be made  just a source repo or will it be actively 
> maintained where new patches are acceptable and can be merged into the 
> repo ?
>
> Thanks and regards,
> Sunad
>
> -----Original Message-----
> From: Linux-nvme [mailto:linux-nvme-bounces at lists.infradead.org] On 
> Behalf Of Keith Busch
> Sent: Monday, January 19, 2015 8:03 PM
> To: Sunad Bhandary
> Cc: 'Keith Busch'; linux-nvme at lists.infradead.org
> Subject: Re: NVMe driver status
>
> Hi Sunad,
>
> The mainline going forward with kernel 3.19 will be blk-mq, and the 
> linux-nvme tree will be there as well on the next upstream merge.
>
> I have my own repo for maintaining the bio-based version since a lot 
> of OS vendors are using older kernels in their minor release cycles. I 
> can push this out to a public repo on git.infradead.org if there's
interest.
>
> Thanks,
> Keith
>
> On Mon, 19 Jan 2015, Sunad Bhandary wrote:
>> Hi Keith,
>>
>> Currently there are two different variants of the NVMe driver, one 
>> with the multi-queue implementation which is under active development 
>> and the one without it which is the current driver hosted in the NVMe 
>> git
> page.
>>
>> Going forward will the non-multi queue  driver also be maintained or 
>> will it be done away with completely ?
>>
>> Regards,
>> Sunad
>
> _______________________________________________
> Linux-nvme mailing list
> Linux-nvme at lists.infradead.org
> http://lists.infradead.org/mailman/listinfo/linux-nvme

_______________________________________________
Linux-nvme mailing list
Linux-nvme at lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-nvme


_______________________________________________
Linux-nvme mailing list
Linux-nvme at lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-nvme

NVMe driver status

[Keith Busch]

I'm happy to apply requested changes. I can't necessarily guarantee a
particular deadline, though I try to align with interested OS vendors
ahead of their release updates if they're using the older version.

Specifically on async probe and others, I'll get these merged in later
this week.

On Tue, 10 Mar 2015, David Darrington wrote:
> I'm interested in this also. I know that the there is a quarterly planning call for the ofed Windows NVMe driver. Would there be any interest in having occasional conference calls to discuss items such as these? I would be glad to setup a call, as long as Keith and/or Mathew agree that it is worthwhile and can attend.
>
> -----Original Message-----
> From: Linux-nvme [mailto:linux-nvme-bounces at lists.infradead.org] On Behalf Of Sunad Bhandary
> Sent: Tuesday, March 10, 2015 10:21 AM
> To: 'Keith Busch'
> Cc: linux-nvme at lists.infradead.org
> Subject: RE: NVMe driver status
>
> Hi Keith,
> Can we initiate some activity on the legacy-nvme maintenance?
>
> Can you commit the async controller probe patch for the legacy driver also?
> Or I can do this if you want.
> Also, I'll be submitting the legacy-nvme version for a couple of patches that we already submitted (write_long, split i/o in IOCtl path).
>
> Going forward, how will the legacy-nvme be maintained?
> The legacy-nvme was removed in 3.19. So are we going to have some restrictions on the kernel versions supported by legacy nvme driver

NVMe driver status

[Brandon Schulz]

Is there any kind of coordination among contributors on backports to legacy-nvme or is it just kind of a free-for-all until someone submits a patch on this list?

Thanks,
Brandon


Brandon Schulz
Systems & Software Architect 
SSD Device Development 
HGST, a Western Digital company
brandon.schulz at hgst.com
o: (507) 322-2257 
 
3605 Hwy 52 N 
Rochester, MN 55901
www.hgst.com


-----Original Message-----
From: Linux-nvme [mailto:linux-nvme-bounces@lists.infradead.org] On Behalf Of Keith Busch
Sent: Tuesday, March 10, 2015 3:02 PM
To: David Darrington
Cc: 'Keith Busch'; linux-nvme at lists.infradead.org; Sunad Bhandary
Subject: RE: NVMe driver status

I'm happy to apply requested changes. I can't necessarily guarantee a particular deadline, though I try to align with interested OS vendors ahead of their release updates if they're using the older version.

Specifically on async probe and others, I'll get these merged in later this week.

On Tue, 10 Mar 2015, David Darrington wrote:
> I'm interested in this also. I know that the there is a quarterly planning call for the ofed Windows NVMe driver. Would there be any interest in having occasional conference calls to discuss items such as these? I would be glad to setup a call, as long as Keith and/or Mathew agree that it is worthwhile and can attend.
>
> -----Original Message-----
> From: Linux-nvme [mailto:linux-nvme-bounces at lists.infradead.org] On 
> Behalf Of Sunad Bhandary
> Sent: Tuesday, March 10, 2015 10:21 AM
> To: 'Keith Busch'
> Cc: linux-nvme at lists.infradead.org
> Subject: RE: NVMe driver status
>
> Hi Keith,
> Can we initiate some activity on the legacy-nvme maintenance?
>
> Can you commit the async controller probe patch for the legacy driver also?
> Or I can do this if you want.
> Also, I'll be submitting the legacy-nvme version for a couple of patches that we already submitted (write_long, split i/o in IOCtl path).
>
> Going forward, how will the legacy-nvme be maintained?
> The legacy-nvme was removed in 3.19. So are we going to have some 
> restrictions on the kernel versions supported by legacy nvme driver

_______________________________________________
Linux-nvme mailing list
Linux-nvme at lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-nvme

NVMe driver status

[Keith Busch]

On Wed, 11 Mar 2015, Brandon Schulz wrote:
> Is there any kind of coordination among contributors on backports to
> legacy-nvme or is it just kind of a free-for-all until someone submits a
> patch on this list?

Coordination is done on the mailing lists, though that probably feels
a bit like a free-for-all.

I originally started the 3.18 fork so I can point distro vendors using
older kernels to a public tree to pull commits ahead of their updates. It
sounds like there is considerable outside interest, so if people are open
to co-maintaining, we can fork this to a github or something like that.

> Thanks,
> Brandon

NVMe driver status

[Brandon Schulz]

Yes, we are interested in co-maintaining...  something like Github would work for us.

Do you want to set something up, or how do you want to proceed?

Thanks,
Brandon


-----Original Message-----
From: Keith Busch [mailto:keith.busch@intel.com] 
Sent: Wednesday, March 11, 2015 1:40 PM
To: Brandon Schulz
Cc: Keith Busch; David Darrington; linux-nvme at lists.infradead.org
Subject: RE: NVMe driver status

On Wed, 11 Mar 2015, Brandon Schulz wrote:
> Is there any kind of coordination among contributors on backports to 
> legacy-nvme or is it just kind of a free-for-all until someone submits 
> a patch on this list?

Coordination is done on the mailing lists, though that probably feels a bit like a free-for-all.

I originally started the 3.18 fork so I can point distro vendors using older kernels to a public tree to pull commits ahead of their updates. It sounds like there is considerable outside interest, so if people are open to co-maintaining, we can fork this to a github or something like that.

> Thanks,
> Brandon

NVMe driver status

[Keith Busch]

On Wed, 11 Mar 2015, Brandon Schulz wrote:
> Yes, we are interested in co-maintaining...  something like Github would work for us.
>
> Do you want to set something up, or how do you want to proceed?

Are any of you folks at Linux Vault this week? We can put together an
impromptu BoF and sort this out real quick.

> Thanks,
> Brandon

NVMe driver status

[Brandon Schulz]

Sorry Keith - Dave and I are not at Linux Vault...

Brandon


-----Original Message-----
From: Keith Busch [mailto:keith.busch@intel.com] 
Sent: Wednesday, March 11, 2015 9:02 PM
To: Brandon Schulz
Cc: Keith Busch; David Darrington; linux-nvme at lists.infradead.org
Subject: RE: NVMe driver status

On Wed, 11 Mar 2015, Brandon Schulz wrote:
> Yes, we are interested in co-maintaining...  something like Github would work for us.
>
> Do you want to set something up, or how do you want to proceed?

Are any of you folks at Linux Vault this week? We can put together an impromptu BoF and sort this out real quick.

> Thanks,
> Brandon

NVMe driver status

[Brandon Schulz]

Keith -

Any more thoughts on how we proceed with setting up this co-maintained Github space?

Thanks,
Brandon



-----Original Message-----
From: Keith Busch [mailto:keith.busch@intel.com] 
Sent: Wednesday, March 11, 2015 9:02 PM
To: Brandon Schulz
Cc: Keith Busch; David Darrington; linux-nvme at lists.infradead.org
Subject: RE: NVMe driver status

On Wed, 11 Mar 2015, Brandon Schulz wrote:
> Yes, we are interested in co-maintaining...  something like Github would work for us.
>
> Do you want to set something up, or how do you want to proceed?

Are any of you folks at Linux Vault this week? We can put together an impromptu BoF and sort this out real quick.

> Thanks,
> Brandon

NVMe driver status

[Keith Busch]

If you setup a github project, I'll subscribe to it.

I merged all the upstream patches I intended to apply today and pushed to
'for-next' branch in the nvme-legacy git tree. I haven't fully tested yet,
otherwise I'd have just merged to master.

Also merged the lastest stable 3.18.y to master.

On Mon, 16 Mar 2015, Brandon Schulz wrote:
> Keith -
>
> Any more thoughts on how we proceed with setting up this co-maintained Github space?
>
> Thanks,
> Brandon