Why is ide-scsi not provided in Arjan's 2.6 kernel?

[Willem Riede]

Unlike previous versions (last I tried that had it was
kernel-2.6.0-0.test8.1.66.i686.rpm)
kernel-smp-2.6.0-0.test9.1.76.i686.rpm does not have the ied-scsi module.

It's config has: 
configs/kernel-2.6.0-test9-i686.config:# CONFIG_BLK_DEV_IDESCSI is not set

This is really annoying, and actually leaves all owners of IDE OnStream tape drives
(where osst needs ide-scsi) in the cold! They will have to recompile every time :-(

Can this decision _please_ be revoked?

Thanks, Willem Riede.
(maintainer of the osst driver)

[Xose Vazquez Perez]

Willem Riede wrote:

>  Unlike previous versions (last I tried that had it was
> kernel-2.6.0-0.test8.1.66.i686.rpm)
>  kernel-smp-2.6.0-0.test9.1.76.i686.rpm does not have the ied-scsi module.
>  
>  It's config has: 
>  configs/kernel-2.6.0-test9-i686.config:# CONFIG_BLK_DEV_IDESCSI is not set
>  
>  This is really annoying, and actually leaves all owners of IDE OnStream tape drives
>  (where osst needs ide-scsi) in the cold! They will have to recompile every time :-(
>  
>  Can this decision _please_ be revoked? 
ide-scsi is broken

-- 
HTML mails are going to trash automagically

[Willem Riede]

On 2003.11.11 18:40, Xose Vazquez Perez wrote:
>  Willem Riede wrote:
>  
>  > Unlike previous versions (last I tried that had it was
> kernel-2.6.0-0.test8.1.66.i686.rpm)
>  > kernel-smp-2.6.0-0.test9.1.76.i686.rpm does not have the ied-scsi module.
>  > 
>  > It's config has: 
>  > configs/kernel-2.6.0-test9-i686.config:# CONFIG_BLK_DEV_IDESCSI is not set
>  > 
>  > This is really annoying, and actually leaves all owners of IDE OnStream tape drives
>  > (where osst needs ide-scsi) in the cold! They will have to recompile every time :-(
>  > 
>  > Can this decision _please_ be revoked?
>  
>  ide-scsi is broken 
Then why was it working just fine for me in test8?

And if it is not perfect, it needs fixing, not ignoring.

Regards, willem Riede.

[Xose Vazquez Perez]

Willem Riede wrote:

>  Then why was it working just fine for me in test8?
>  
>  And if it is not perfect, it needs fixing, not ignoring. 
from http://www.codemonkey.org.uk/post-halloween-2.5.txt

o  ide-scsi is completely broken in 2.6 currently. Known problem.
     If you need it either use 2.4 or fix it 8)

ask linux-kernel, they are the guilty ;-)

-- 
HTML mails are going to trash automagically

[nosp]

On Wed, 2003-11-12 at 00:02, Willem Riede wrote:
>  > ide-scsi is broken
>  
>  Then why was it working just fine for me in test8?
>  
>  And if it is not perfect, it needs fixing, not ignoring. 
You may want to check out the linux kernel mailing list as there is a
lot of traffic about ide-scsi (probably related).  Linus is involved and
basically it seems like it's broken and going to stay broken until
people contribute code.  So get out your vi/emacs or shut up :).

[Alan Cox]

>  You may want to check out the linux kernel mailing list as there is
> a
>  lot of traffic about ide-scsi (probably related).  Linus is involved and
>  basically it seems like it's broken and going to stay broken until
>  people contribute code.  So get out your vi/emacs or shut up :). 
Fixes were posted 6 months ago

[nosp]

On Wed, 2003-11-12 at 00:40, Alan Cox wrote:
>  > You may want to check out the linux kernel mailing list as there
> is a
>  > lot of traffic about ide-scsi (probably related).  Linus is involved and
>  > basically it seems like it's broken and going to stay broken until
>  > people contribute code.  So get out your vi/emacs or shut up :).
>  
>  Fixes were posted 6 months ago 
I should have also said: *Alan* and Linus and everyone is involved, so I
who am not worthy will bow out very quickly :).

For us mortals does this mean ide-scsi support is planned (targeted?
viable? hoped-for?) for FC2?

[Alan Cox]

>  I should have also said: *Alan* and Linus and everyone is involved,
> so I
>  who am not worthy will bow out very quickly :). 
Was involved. I got bored of wondering why nobody was fixing it in 2.6
when I fixed it all in 2.4. Ditto porting stuff like hotplug IDE, tho
I've not looked if that one was done.

[Willem Riede]

On 2003.11.11 19:52, Alan Cox wrote:
>  > I should have also said: *Alan* and Linus and everyone is
> involved, so I
>  > who am not worthy will bow out very quickly :).
>  
>  Was involved. I got bored of wondering why nobody was fixing it in 2.6
>  when I fixed it all in 2.4. Ditto porting stuff like hotplug IDE, tho
>  I've not looked if that one was done.
>   
And while I would never dare to compare my feeble efforts to Alan's, I have
posted on improvements to ide-scsi on linux-scsi list myself (1). 

When recently I decided to see where it stands, and tried Arjan's test 8 
kernel, I was pleasantly surprised at how well it worked. Hence also my 
disappointment to see it disabled.

And to "nosp" - I have every intention to get out my vi, but that doesn't
help if the result won't get used.

Thanks, Willem Riede.

(1) see a/o http://marc.theaimsgroup.com/?l=linux-scsi&m=104361480527780&w=2

[Alan Cox]

>  And to "nosp" - I have every intention to get out my vi,
> but that doesn't
>  help if the result won't get used.
>  
>  Thanks, Willem Riede.
>  
>  (1) see a/o http://marc.theaimsgroup.com/?l=linux-scsi&m=104361480527780&w=2
Nice. Thats way better than letting the IDE layer time it out as I did

[Willem Riede]

On 2003.11.12 18:40, Alan Cox wrote:
>  > 
>  > (1) see a/o
> http://marc.theaimsgroup.com/?l=linux-scsi&m=104361480527780&w=2
>  
>  Nice. Thats way better than letting the IDE layer time it out as I did 
Thank, Alan! Do you have any guidance as to what I can do to get ide-scsi resurrected?

Regards, Willem Riede.

[Alan Cox]

>  On 2003.11.12 18:40, Alan Cox wrote:
>  > > 
>  > > (1) see a/o
> http://marc.theaimsgroup.com/?l=linux-scsi&m=104361480527780&w=2
>  > 
>  > Nice. Thats way better than letting the IDE layer time it out as I did
>  
>  Thank, Alan! Do you have any guidance as to what I can do to get ide-scsi resurrected?
I guess submit the patch a few times. It'll get picked up in the end, even
if the distor folks do it. Certainly make sure arjan knows about it as 
the plan is for Fedora 2 to have 2.6.x, and we know that is likely to need
a lot of "escaped" patches sweeping up

[Xose Vazquez Perez]

Alan Cox wrote:

>  I guess submit the patch a few times. It'll get picked up in the
> end, even
>  if the distor folks do it. Certainly make sure arjan knows about it as 
>  the plan is for Fedora 2 to have 2.6.x, and we know that is likely to need
>  a lot of "escaped" patches sweeping up 
I hope to see a 2.4.x kernel in FC2 _too_. Otherwise it will be a lame
distribution.

--
HTML mails are going to trash automagically

[Chris Adams]

Once upon a time, Xose Vazquez Perez <xose(a)wanadoo.es&gt; said:
>  I hope to see a 2.4.x kernel in FC2 _too_. Otherwise it will be a
> lame
>  distribution. 
What would be lame about not including an old kernel?

Including multiple kernel versions (especially from different major
release trains) greatly increases complexity for very little gain.

-- 
Chris Adams <cmadams(a)hiwaay.net&gt;
Systems and Network Administrator - HiWAAY Internet Services
I don't speak for anybody but myself - that's enough trouble.

[Jeremy Katz]

On Thu, 2003-11-13 at 18:00, Xose Vazquez Perez wrote:
>  Alan Cox wrote:
>  > I guess submit the patch a few times. It'll get picked up in the end, even
>  > if the distor folks do it. Certainly make sure arjan knows about it as 
>  > the plan is for Fedora 2 to have 2.6.x, and we know that is likely to need
>  > a lot of "escaped" patches sweeping up
>  
>  I hope to see a 2.4.x kernel in FC2 _too_. Otherwise it will be a lame
>  distribution. 
As was already discussed elsewhere (and this is entirely the wrong list
at this point ;-), it is impossible to take advantage of all of the
benefits that you get with 2.6 and still leave a 2.4.x kernel.  Not to
mention the question of how do you present this to the user.

The right answer is to get 2.6.x to where it doesn't cause FC2 to be "a
lame distribution" by getting it tested and stabilized :-)

Cheers,

Jeremy