Re: drivers/sh/maple/maple.c compile error

[Paul Mundt <lethal@linux-sh.org>]

On Mon, Jul 28, 2008 at 09:51:48PM +0100, Adrian McMenamin wrote:
> On Tue, 2008-07-29 at 05:44 +0900, Paul Mundt wrote:
> > On Mon, Jul 28, 2008 at 11:20:12PM +0300, Adrian Bunk wrote:
> > > Commit 306cfd630a4d121cf4e08b894d8b4c4cf106e57e
> > > (maple: tidy maple_driver code by removing redundant connect/disconnect)
> > > causes the following compile error:
> > > 
> > > <--  snip  -->
> > > 
> > > ...
> > >   CC      drivers/sh/maple/maple.o
> > > /home/bunk/linux/kernel-2.6/git/linux-2.6/drivers/sh/maple/maple.c: In function 'attach_matching_maple_driver':
> > > /home/bunk/linux/kernel-2.6/git/linux-2.6/drivers/sh/maple/maple.c:259: error: 'struct maple_driver' has no member named 'connect'
> > > /home/bunk/linux/kernel-2.6/git/linux-2.6/drivers/sh/maple/maple.c: In function 'maple_detach_driver':
> > > /home/bunk/linux/kernel-2.6/git/linux-2.6/drivers/sh/maple/maple.c:272: error: 'struct maple_driver' has no member named 'disconnect'
> > > /home/bunk/linux/kernel-2.6/git/linux-2.6/drivers/sh/maple/maple.c:273: error: 'struct maple_driver' has no member named 'disconnect'
> > > make[4]: *** [drivers/sh/maple/maple.o] Error 1
> > > 
> > > <--  snip  -->
> > > 
> > Adrian was supposed to send follow-up patches for this, which never
> > happened. I thought I had dropped this already, but in this case I'll
> > just revert it.
> 
> Actually I wasn't asked about this, but to submit a fix/replacement for
> a patch that didn't apply.
> 
> I will submit a patch for the whole thing then - as I cannot actually
> match the commit Adrian has quoted to anything.
> 
> I think the problem is that I am trying to get a change in the bus and
> in the drivers (which have different maintainers) to go in together and
> that has never quite worked. The code runs fine on my box and has done
> for months now.
> 
This seems to point to a serious problem in your working environment, as
you are neither able to view commits that are referenced or realize there
is a problem when patches fail to apply anywhere outside of your home
directory.

The problem is that the change you made to maple.c is incomplete, it
basically should have followed 2 other patches, one removing the logic
entirely from maple.c, and one removing it from the input driver, before
killing off the function pointers completely.  You were asked to fix up
and resend the input patch (which seems not to have happened due to a
problem in your work environment), but I'm unable to find anything
bordering on a maple patch that actually cleans up
attach_matching_maple_driver() and maple_detach_driver()
connect/disconnect references?

In any event, I'm going to revert this. Once you get your environment
fixed and are able to send a coherent series, we can revisit this. Dmitry
has already indicated that taking the input bits through my tree is fine,
so any process problems there are the least of the problem.