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

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



On Mon, May 22, 2023 at 12:42:15PM +0200, Niklas Schnelle wrote:
> 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

Sure, that reasoning makes sense to me too, so let's go with the
explicit depends afterall.

By the way, I noticed two other modules that call devm_ioport_map() but
seem to be missing the HAS_IOPORT_MAP depends lines: the
drivers/iio/addac/stx104.c and drivers/iio/dac/cio-dac.c drivers. Do
these need respective patches as well?

As an aside, I haven't been following the previous patchsets closely so
forgive me if this has already been discussed in another thread: why
doesn't X86 automatically select HAS_IOPORT? Are there x86 platforms
that do not support ioport?

William Breathitt Gray

Attachment: signature.asc
Description: PGP signature


[Index of Archives]     [Linux USB Devel]     [Video for Linux]     [Linux Audio Users]     [Yosemite News]     [Linux Input]     [Linux Kernel]     [Linux SCSI]     [X.org]

  Powered by Linux