Re: smp dead lock of io_request_lock/queue_lock patch

[James Bottomley <James.Bottomley@steeleye.com>]

On Mon, 2004-01-12 at 10:52, Doug Ledford wrote:
> Well, the scsi-dledford-2.4 tree is intended to be someplace I can put
> all the stuff I'm having to carry forward in our kernels, so that's
> distinctly different than a driver update only tree.  I could do that
> separately and I have no problem doing that.

I'll take that as a "yes" then ;-)

Thanks for doing this, beacuse I really wasn't looking forward to trying
to sort it all out.

>  As for the other stuff,
> I'm not pushing to necessarily get any of my changes into mainline.  I
> would be happy if they make it there sometime as that would relieve load
> off of me, but at the same time I *am* making some changes to the core
> code (sorry Jens, but there are some ways in which the 2.4 core scsi
> code is just too broken to believe and leaving it broken just means
> dealing with everyone that points it out in bugzilla entries) and I know
> people are loath to change core code in mainline, so I kinda figured a
> lot of that stuff would either A) stay separate or B) only after myself
> and other interested parties agree on a patch and that patch is widely
> tested and known good then maybe it might get moved over, up to Marcelo.

I trust your judgement about this, so it sounds like we have the
beginnings of a good working model for 2.4

James