merging PCMCIA changes

[Greg KH]

On Sat, Jun 18, 2005 at 02:55:13PM -0700, Andrew Morton wrote:
> Dominik Brodowski <linux at dominikbrodowski.net> wrote:
> >
> > Hi,
> > 
> > On Sat, Jun 18, 2005 at 11:17:18AM -0700, Andrew Morton wrote:
> > > > Also, one other pci-related change very helpful for PCMCIA was in -mm for
> > > > quite some time and, to my knowledge, did not cause any problems. It was
> > > > titled 
> > > > 
> > > > gregkh-pci-pci-pci-transparent-bridge-handling-improvements-pci-core.patch
> > > > 
> > > > and was present until 2.6.12-rc5-mm2. While a different patch intending the
> > > > same functionality is in -mm, the related changes to the PCI infrastructure
> > > > cause severe headaches (see bug 
> > > > http://bugzilla.kernel.org/show_bug.cgi?id=4737 for example, other, similar
> > > > reports were posted to lkml.) Therefore I'd suggest to merge the patch
> > > > mentioned above at first, and fix up new PCI patches in -mm first.
> > > 
> > > The patches in -mm are staged after the git trees (mainly because I'm more
> > > flexible with rejects and things than people who are stuck using an SCM ;)
> > > 
> > > So I tend to wait until the bk/scm guys have merged.
> > 
> > However, with git-pci being broken at the moment, merging git-pci first does
> > not make sense. So isn't this a natural exception to your "normal" rule?
> > After all, gregkh-pci-pci-pci-transparent-bridge-handling-improvements-pci-core.patch
> > proved to do the right thing in -mm for quite some time. Or what am I
> > missing?
> > 
> 
> Yes, a number of people have reported the things in bug 4737.
> 
> I'd prefer that we get a patch in place which fixes up Greg's tree instead
> of doing something in parallel.
> 
> It's up to Greg, really.

Ok, I'll fix this all up before sending it off to Linus.

I need to get the driver core and devfs stuff off to him first, to make
it easier to merge the later stuff on in, so it might take me a day or
so (real world work is dragging me down...)

thanks,

greg k-h