Re: [PATCH] 2.6 SGI Altix I/O code reorganization

[Patrick Gefre <pfg@sgi.com>]

Luck, Tony wrote:
>>Yes, after looking at Grant's review/suggestion, we found that we can 
>>actually just use raw_pci_ops.  This will work well for us.  We have 
>>incoorporated this change.  No changes in pci/pci.c needed.
> 
> 
> Good.  Let's try to make some forward progress here.  I'd like
> to see the patches broken into a sequence something like this:
> 
> 1) Add new interfaces to header files to support any new API
>    needed by new files
> 2) Create all the new files (plain copies of old files where
>    a move is involved).
> 3) Functional changes to copied files.
> 4) Whitespace cleanup of copied files.
> 5) Point Makefiles to new files
> 6) Delete all the old/unused files.
> 7) Delete any API in headers that were only used by old files.
> 
> We'll need to coordinate with some other maintainrs for
> drivers/pci/hotplug/Kconfig and drivers/scsi/qla1280.c,
> but I'm ok with running all the other parts through the
> ia64 tree.
> 
> This follows the usual guidelines of a sequence of steps where
> the system is buildable+usable at each stage.
> 

Tony,

It had been suggested that we submit this as new code - since it can't be transitioned to. And I 
thought that was what we had decided on - a 'kill' patch and an 'add' patch. I can remove any 
Lindent'ing of older files if you don't want that. I will take out the Kconfig mod. I believe 
Christoph is the maintainer of the qla driver (he was one of the reviewers).

-- Pat