* RE: IP over FC and qlogic workarounds in fusion
@ 2004-08-30 22:29 Moore, Eric Dean
2004-09-06 10:13 ` Christoph Hellwig
0 siblings, 1 reply; 6+ messages in thread
From: Moore, Eric Dean @ 2004-08-30 22:29 UTC (permalink / raw)
To: Christoph Hellwig, andrew.vasquez; +Cc: linux-scsi, Shirron, Stephen
Here is feedback from Stephen Shirron, co-worker here
at LSI Logic:
To my knowledge QLogic has never even acknowledged that
they had a bug. In my opinion it is not
safe to remove this code. We do have versions of the FC
firmware that work around this QLogic bug, but it's hard
to know if the entire field is populated. It would not
be good to introduce a regression that decreases our
interoperability.
On Saturday, August 28, 2004 1:37 PM, Christoph Hellwig wrote:
>
> Does anyone know whether modern qlogic firmwares fixed the
> QLOGIC_NAA_WORKAROUND in mptlan.c? I'd love to get rid of it
> so fusion
> could use the little bit of common ip over fc infrastructure we have..
>
^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: IP over FC and qlogic workarounds in fusion
2004-08-30 22:29 IP over FC and qlogic workarounds in fusion Moore, Eric Dean
@ 2004-09-06 10:13 ` Christoph Hellwig
2004-09-07 4:16 ` Andrew Vasquez
0 siblings, 1 reply; 6+ messages in thread
From: Christoph Hellwig @ 2004-09-06 10:13 UTC (permalink / raw)
To: Moore, Eric Dean
Cc: Christoph Hellwig, andrew.vasquez, linux-scsi, Shirron, Stephen
On Mon, Aug 30, 2004 at 06:29:47PM -0400, Moore, Eric Dean wrote:
> Here is feedback from Stephen Shirron, co-worker here
> at LSI Logic:
>
> To my knowledge QLogic has never even acknowledged that
> they had a bug. In my opinion it is not
> safe to remove this code. We do have versions of the FC
> firmware that work around this QLogic bug, but it's hard
> to know if the entire field is populated. It would not
> be good to introduce a regression that decreases our
> interoperability.
Andrew, could you please try to to find the right people at Qlogic
to ACK or even better fix this?
Eric, are you and your LSI coworkers willing to integrate the workaround
with the net/802/fc.c code and make fusion use it?
^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: IP over FC and qlogic workarounds in fusion
2004-09-06 10:13 ` Christoph Hellwig
@ 2004-09-07 4:16 ` Andrew Vasquez
0 siblings, 0 replies; 6+ messages in thread
From: Andrew Vasquez @ 2004-09-07 4:16 UTC (permalink / raw)
To: Christoph Hellwig; +Cc: Moore, Eric Dean, linux-scsi, Shirron, Stephen
On Mon, 2004-09-06 at 03:13, Christoph Hellwig wrote:
> On Mon, Aug 30, 2004 at 06:29:47PM -0400, Moore, Eric Dean wrote:
> > Here is feedback from Stephen Shirron, co-worker here
> > at LSI Logic:
> >
> > To my knowledge QLogic has never even acknowledged that
> > they had a bug. In my opinion it is not
> > safe to remove this code. We do have versions of the FC
> > firmware that work around this QLogic bug, but it's hard
> > to know if the entire field is populated. It would not
> > be good to introduce a regression that decreases our
> > interoperability.
>
> Andrew, could you please try to to find the right people at Qlogic
> to ACK or even better fix this?
>
I'm not sure who Stephen originally worked with at QLogic, but yes, this
is a known problem (finally managed to track down several firmware
engineers) and unfortunately 'will not be addressed' in the future.
You'll also note that FC-IP support has effectively been extricated from
the 8.x series driver.
Regards,
Andrew Vasquez
^ permalink raw reply [flat|nested] 6+ messages in thread
* RE: IP over FC and qlogic workarounds in fusion
@ 2004-09-07 18:12 Shirron, Stephen
2004-09-07 18:14 ` Christoph Hellwig
0 siblings, 1 reply; 6+ messages in thread
From: Shirron, Stephen @ 2004-09-07 18:12 UTC (permalink / raw)
To: Christoph Hellwig, Moore, Eric Dean; +Cc: andrew.vasquez, linux-scsi
One problem I see with trying to remove this workaround from
our driver is that it would have to be coupled with a kernel
release, and then the newer driver would be bound to that release.
It would make it harder for customers to use a new driver kit with
an old kernel.
How is something like this typically addressed?
We are not opposed to making such a change.
stephen
-----Original Message-----
From: Christoph Hellwig [mailto:hch@lst.de]
Sent: Monday, September 06, 2004 6:14 AM
To: Moore, Eric Dean
Cc: Christoph Hellwig; andrew.vasquez@qlogic.com;
linux-scsi@vger.kernel.org; Shirron, Stephen
Subject: Re: IP over FC and qlogic workarounds in fusion
On Mon, Aug 30, 2004 at 06:29:47PM -0400, Moore, Eric Dean wrote:
> Here is feedback from Stephen Shirron, co-worker here
> at LSI Logic:
>
> To my knowledge QLogic has never even acknowledged that
> they had a bug. In my opinion it is not
> safe to remove this code. We do have versions of the FC
> firmware that work around this QLogic bug, but it's hard
> to know if the entire field is populated. It would not
> be good to introduce a regression that decreases our
> interoperability.
Andrew, could you please try to to find the right people at Qlogic
to ACK or even better fix this?
Eric, are you and your LSI coworkers willing to integrate the workaround
with the net/802/fc.c code and make fusion use it?
^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: IP over FC and qlogic workarounds in fusion
2004-09-07 18:12 Shirron, Stephen
@ 2004-09-07 18:14 ` Christoph Hellwig
0 siblings, 0 replies; 6+ messages in thread
From: Christoph Hellwig @ 2004-09-07 18:14 UTC (permalink / raw)
To: Shirron, Stephen; +Cc: Moore, Eric Dean, andrew.vasquez, linux-scsi
On Tue, Sep 07, 2004 at 12:12:10PM -0600, Shirron, Stephen wrote:
> One problem I see with trying to remove this workaround from
> our driver is that it would have to be coupled with a kernel
> release, and then the newer driver would be bound to that release.
> It would make it harder for customers to use a new driver kit with
> an old kernel.
>
> How is something like this typically addressed?
Backport the stack changes aswell when you want to run an older release.
There's a lot of recent additions additions that move code from drivers
to the stack and you need to make sure that the stack code is uptodate
to use all those features.
^ permalink raw reply [flat|nested] 6+ messages in thread
* IP over FC and qlogic workarounds in fusion
@ 2004-08-28 19:37 Christoph Hellwig
0 siblings, 0 replies; 6+ messages in thread
From: Christoph Hellwig @ 2004-08-28 19:37 UTC (permalink / raw)
To: Emoore, andrew.vasquez; +Cc: linux-scsi
Does anyone know whether modern qlogic firmwares fixed the
QLOGIC_NAA_WORKAROUND in mptlan.c? I'd love to get rid of it so fusion
could use the little bit of common ip over fc infrastructure we have..
^ permalink raw reply [flat|nested] 6+ messages in thread
end of thread, other threads:[~2004-09-07 18:14 UTC | newest]
Thread overview: 6+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2004-08-30 22:29 IP over FC and qlogic workarounds in fusion Moore, Eric Dean
2004-09-06 10:13 ` Christoph Hellwig
2004-09-07 4:16 ` Andrew Vasquez
-- strict thread matches above, loose matches on Subject: below --
2004-09-07 18:12 Shirron, Stephen
2004-09-07 18:14 ` Christoph Hellwig
2004-08-28 19:37 Christoph Hellwig
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox