Re: [PATCH v4 05/41] counter: add HAS_IOPORT dependencies

[Niklas Schnelle]

On Fri, 2023-05-19 at 10:21 -0400, William Breathitt Gray wrote:
> On Fri, May 19, 2023 at 03:39:57PM +0200, Niklas Schnelle wrote:
> > On Fri, 2023-05-19 at 15:38 +0200, Niklas Schnelle wrote:
> > > On Fri, 2023-05-19 at 15:17 +0200, Niklas Schnelle wrote:
> > > > On Thu, 2023-05-18 at 21:26 -0400, William Breathitt Gray wrote:
> > > > > On Tue, May 16, 2023 at 01:00:01PM +0200, Niklas Schnelle wrote:
> > > > > > In a future patch HAS_IOPORT=n will result in inb()/outb() and friends
> > > > > > not being declared. We thus need to add HAS_IOPORT as dependency for
> > > > > > those drivers using them.
> > > > > > 
> > > > > > Co-developed-by: Arnd Bergmann <arnd@xxxxxxxxxx>
> > > > > > Signed-off-by: Arnd Bergmann <arnd@xxxxxxxxxx>
> > > > > > Signed-off-by: Niklas Schnelle <schnelle@xxxxxxxxxxxxx>
> > > > > 
> > > > > Hi Niklas,
> > > > > 
> > > > > The change itself is fine, but please update the description to reflect
> > > > > that this is adding a depends on HAS_IOPORT_MAP rather than HAS_IOPORT,
> > > > > along with the reason why it's needed (i.e. devm_ioport_map() is used).
> > > > > 
> > > > > Thanks,
> > > > > 
> > > > > William Breathitt Gray
> > > > > 
> > > > > 
> > > > 
> > > > Right, this clearly needs adjustment. I went with the following commit
> > > > message for v5:
> > > > 
> > > > "counter: add HAS_IOPORT_MAP dependency
> > > > 
> > > > The 104_QUAD_8 counter driver uses devm_ioport_map() without depending
> > > > on HAS_IOPORT_MAP. This causes compilation to fail on platforms such as
> > > > s390 which do not support I/O port mapping. Add the missing
> > > > HAS_IOPORT_MAP dependency to fix this."
> > > > 
> > > 
> > > Just noticed this isn't entirely correct. As devm_ioport_map() has an
> > > empty stub for HAS_IOPORT_MAP=n this doesn't lead to a compile error it
> > > just doesn't work. Will reword to "This causes the driver to not be
> > > useable on platforms ..."
> > 
> > s/useable/usable/
> 
> 104_QUAD_8 has an explicit dependency on PC104 and X86, so I don't think
> it would ever be used outside of x86 platforms. Does it still make sense
> to have the HAS_IOPORT_MAP dependency in this case?
> 
> William Breathitt Gray

Well, yes and no, you're right that it doesn't really cause compile
issues despite the "|| COMPILE_TEST" albeit the code could never work.
Still, I'd add the dependency. At the very least it serves as
documentation and maybe in the future someone will want to remove those
empty stubs for HAS_IOPORT_MAP=n.

Thanks
Niklas