On Wed, 2010-10-13 at 15:28 -0700, Joe Perches wrote:
quoted text
> On Mon, 2010-10-11 at 17:00 -0700, Joe Perches wrote:
> > On Mon, 2010-10-11 at 16:52 -0700, Jeff Kirsher wrote:
> > > On Sun, Oct 10, 2010 at 13:42, Joe Perches <joe@perches.com> wrote:
> > > > Perhaps it's better to move drivers from the very populated
> > > > drivers/net directory into vendor specific directories similar
> > > > to the Atheros approach used for drivers/net/wireless/ath/
> > > NAK
> > > First, I think we need to keep the documentation in /Documentation/networking.
> > > Second, the changes are extensive and would create a lot of regression testing.
> > I don't see any actual changes here other than layout.
> > What kind of regression testing do you think necessary?
> 
> Jeff?
> 

Sorry I am not ignoring you, I was taking a closer look at your patch.

quoted text
> What regression testing would actually be done?
> 

The Makefile and Kconfig needs more work.  I applied your patch and none
of the Intel Wired drivers build.

The statement that there would be a lot of regression testing was in
reference to your response to Stephen that it would "allow consolidation
of common code".  Sorry for being vague about the regression testing.

In general, I do like the idea of moving all the Intel wired LAN drivers
into their own directory, like was Atheros has done in Wireless.

I am working on providing an updated RFC patch to resolve the
Makefile/Kconfig issues I found and few other minor issues I have
found. 

quoted text
> Any new objects are trivially validated against existing
> objects.
> 
> > > We have been looking at solutions like this for future
> > > drivers/hardware and is on the list of items we are currently working
> > > on, but feel it should not be made retroactively due to the regression
> > > testing and massive changes that would need to be made.
> > 
> > Might as well start somewhere.