Re: 5.x branch mainline kernel push in two weeks?

2005-04-20 Thread Christoph Hellwig
On Wed, Apr 20, 2005 at 01:29:11AM -0700, Mike Christie wrote:
 Are you going to try and push someone like qla4010 which has a similar 
 interface to qla2xxx (it uses the msg box interface) to use the netlink 
 interface? With a lot of the iscsi stuff masked away it seems a little 
 difficult to go the extent open-iscsi did?

I'm not sure we should care too much about the qla4xxx driver.  It's
never been posted to linux-scsi, and the driver they pushed to SuSE
ignores everything we've said about scsi low-level drivers.  It looks
like the iSCSI part of Qlogic simply doesn't care about Linux (that is
upstream, not the major forks)
-
To unsubscribe from this list: send the line unsubscribe linux-scsi in
the body of a message to [EMAIL PROTECTED]
More majordomo info at  http://vger.kernel.org/majordomo-info.html


Re: 5.x branch mainline kernel push in two weeks?

2005-04-20 Thread Christoph Hellwig
On Wed, Apr 20, 2005 at 11:29:12AM -0700, Duane Grigsby wrote:
 No, that's not true; we are TOTALLY committed to getting all of our
 drivers in the upstream kernel. I have been working on the ISCSI driver
 cleanup to the level of our Fibre Channel driver as Andrew has done. I'm
 making it conform to the current linux-scsi API, ISCSI transport, and
 stripping out all of the failover baggage. My goal is the have something
 upstream in 2 - 3 weeks, if not before.   

Okay, sorry for implying the opposite.

So what API are you using for communicating with userspace right now?
How much userspace support do you need?  Is the current iscsi transport
class okay with you and what about the netlink version?

It would be very helpful to have you participating on the various iscsi
transport class and userspace interface discussions.
-
To unsubscribe from this list: send the line unsubscribe linux-scsi in
the body of a message to [EMAIL PROTECTED]
More majordomo info at  http://vger.kernel.org/majordomo-info.html


RE: 5.x branch mainline kernel push in two weeks?

2005-04-20 Thread Dmitry Yusupov
On Wed, 2005-04-20 at 11:29 -0700, Duane Grigsby wrote:
 No, that's not true; we are TOTALLY committed to getting all of our
 drivers in the upstream kernel. I have been working on the ISCSI driver
 cleanup to the level of our Fibre Channel driver as Andrew has done. I'm
 making it conform to the current linux-scsi API, ISCSI transport, and
 stripping out all of the failover baggage. My goal is the have something
 upstream in 2 - 3 weeks, if not before.   

But don't you think it is better to have integrated solution for all
iSCSI transports including QLogic's HBA and have unified user-space set
of tools? How about collaborate on this issue and come up with somewhat
generic which will fit with ongoing efforts we have today:
 
1) iSCSI/TCP
2) your QLogic's HBA
3) iSER/IB

I think those three projects combined could come up with generic iSCSI
solution(including user-space tools and API) which will really fit for
everyone else on the planet.

It would be really nice if you would participate in ongoing interface
discussion on next mailing lists:

1) [EMAIL PROTECTED]
2) [EMAIL PROTECTED]
3) [EMAIL PROTECTED]

Thanks!

 -Original Message-
 From: [EMAIL PROTECTED]
 [mailto:[EMAIL PROTECTED] On Behalf Of Christoph Hellwig
 Sent: Wednesday, April 20, 2005 1:49 AM
 To: [EMAIL PROTECTED]
 Cc: linux-iscsi development team; linux-scsi@vger.kernel.org
 Subject: Re: 5.x branch mainline kernel push in two weeks?
 
 On Wed, Apr 20, 2005 at 01:29:11AM -0700, Mike Christie wrote:
  Are you going to try and push someone like qla4010 which has a similar
 
  interface to qla2xxx (it uses the msg box interface) to use the
 netlink 
  interface? With a lot of the iscsi stuff masked away it seems a little
 
  difficult to go the extent open-iscsi did?
 
 I'm not sure we should care too much about the qla4xxx driver.  It's
 never been posted to linux-scsi, and the driver they pushed to SuSE
 ignores everything we've said about scsi low-level drivers.  It looks
 like the iSCSI part of Qlogic simply doesn't care about Linux (that is
 upstream, not the major forks)
 -
 To unsubscribe from this list: send the line unsubscribe linux-scsi in
 the body of a message to [EMAIL PROTECTED]
 More majordomo info at  http://vger.kernel.org/majordomo-info.html
 
 -
 To unsubscribe from this list: send the line unsubscribe linux-scsi in
 the body of a message to [EMAIL PROTECTED]
 More majordomo info at  http://vger.kernel.org/majordomo-info.html

-
To unsubscribe from this list: send the line unsubscribe linux-scsi in
the body of a message to [EMAIL PROTECTED]
More majordomo info at  http://vger.kernel.org/majordomo-info.html


RE: 5.x branch mainline kernel push in two weeks?

2005-04-20 Thread Duane Grigsby
So what API are you using for communicating with userspace right now?
How much userspace support do you need?  Is the current iscsi transport
class okay with you and what about the netlink version?
 the current driver uses ioctls, but I will be looking at what we need
to do to replace that with the ISCSI transport mechanism. 

How much userspace support do you need?  Is the current iscsi transport
class okay with you and what about the netlink version?
 Good question, I'm still looking into that. Not sure about netlink
version, I don't think the iscsi driver needs it. 

It would be very helpful to have you participating on the various iscsi
transport class and userspace interface discussions.
 I would be glad to participate in those discussions.  
 

-
To unsubscribe from this list: send the line unsubscribe linux-scsi in
the body of a message to [EMAIL PROTECTED]
More majordomo info at  http://vger.kernel.org/majordomo-info.html


RE: 5.x branch mainline kernel push in two weeks?

2005-04-20 Thread Duane Grigsby
Ok, I'll get on those discussions.

-Original Message-
From: Dmitry Yusupov [mailto:[EMAIL PROTECTED] 
Sent: Wednesday, April 20, 2005 11:50 AM
To: Duane Grigsby
Cc: Christoph Hellwig; [EMAIL PROTECTED]; linux-iscsi
development team; linux-scsi@vger.kernel.org
Subject: RE: 5.x branch mainline kernel push in two weeks?

On Wed, 2005-04-20 at 11:29 -0700, Duane Grigsby wrote:
 No, that's not true; we are TOTALLY committed to getting all of our
 drivers in the upstream kernel. I have been working on the ISCSI
driver
 cleanup to the level of our Fibre Channel driver as Andrew has done.
I'm
 making it conform to the current linux-scsi API, ISCSI transport, and
 stripping out all of the failover baggage. My goal is the have
something
 upstream in 2 - 3 weeks, if not before.   

But don't you think it is better to have integrated solution for all
iSCSI transports including QLogic's HBA and have unified user-space set
of tools? How about collaborate on this issue and come up with somewhat
generic which will fit with ongoing efforts we have today:
 
1) iSCSI/TCP
2) your QLogic's HBA
3) iSER/IB

I think those three projects combined could come up with generic iSCSI
solution(including user-space tools and API) which will really fit for
everyone else on the planet.

It would be really nice if you would participate in ongoing interface
discussion on next mailing lists:

1) [EMAIL PROTECTED]
2) [EMAIL PROTECTED]
3) [EMAIL PROTECTED]

Thanks!

 -Original Message-
 From: [EMAIL PROTECTED]
 [mailto:[EMAIL PROTECTED] On Behalf Of Christoph
Hellwig
 Sent: Wednesday, April 20, 2005 1:49 AM
 To: [EMAIL PROTECTED]
 Cc: linux-iscsi development team; linux-scsi@vger.kernel.org
 Subject: Re: 5.x branch mainline kernel push in two weeks?
 
 On Wed, Apr 20, 2005 at 01:29:11AM -0700, Mike Christie wrote:
  Are you going to try and push someone like qla4010 which has a
similar
 
  interface to qla2xxx (it uses the msg box interface) to use the
 netlink 
  interface? With a lot of the iscsi stuff masked away it seems a
little
 
  difficult to go the extent open-iscsi did?
 
 I'm not sure we should care too much about the qla4xxx driver.  It's
 never been posted to linux-scsi, and the driver they pushed to SuSE
 ignores everything we've said about scsi low-level drivers.  It looks
 like the iSCSI part of Qlogic simply doesn't care about Linux (that is
 upstream, not the major forks)
 -
 To unsubscribe from this list: send the line unsubscribe linux-scsi
in
 the body of a message to [EMAIL PROTECTED]
 More majordomo info at  http://vger.kernel.org/majordomo-info.html
 
 -
 To unsubscribe from this list: send the line unsubscribe linux-scsi
in
 the body of a message to [EMAIL PROTECTED]
 More majordomo info at  http://vger.kernel.org/majordomo-info.html


-
To unsubscribe from this list: send the line unsubscribe linux-scsi in
the body of a message to [EMAIL PROTECTED]
More majordomo info at  http://vger.kernel.org/majordomo-info.html