Re: [RFC v2 01/39] Kconfig: introduce HAS_IOPORT option and select it as necessary

[Bjorn Helgaas <helgaas@kernel.org>]

On Wed, May 04, 2022 at 11:31:28PM +0200, Arnd Bergmann wrote:
> On Wed, May 4, 2022 at 11:08 PM Bjorn Helgaas <helgaas@kernel.org> wrote:
> > On Fri, Apr 29, 2022 at 03:49:59PM +0200, Niklas Schnelle wrote:
> > > We introduce a new HAS_IOPORT Kconfig option to indicate support for
> > > I/O Port access. In a future patch HAS_IOPORT=n will disable compilation
> > > of the I/O accessor functions inb()/outb() and friends on architectures
> > > which can not meaningfully support legacy I/O spaces such as s390 or
> > > where such support is optional.
> >
> > So you plan to drop inb()/outb() on architectures where I/O port space
> > is optional?  So even platforms that have I/O port space may not be
> > able to use it?
> >
> > This feels like a lot of work where the main benefit is to keep
> > Kconfig from offering drivers that aren't of interest on s390.
> >
> > Granted, there may be issues where inb()/outb() does the wrong thing
> > such as dereferencing null pointers when I/O port space isn't
> > implemented.  I think that's a defect in inb()/outb() and could be
> > fixed there.
> 
> The current implementation in asm-generic/io.h implements inb()/outb()
> using readb()/writeb() with a fixed architecture specific offset.
> 
> There are three possible things that can happen here:
> 
> a) there is a host bridge driver that maps its I/O ports to this window,
>     and everything works
> b) the address range is reserved and accessible but no host bridge
>    driver has mapped its registers there, so an access causes a
>    page fault
> c) the architecture does not define an offset, and accessing low I/O
>     ports ends up as a NULL pointer dereference
> 
> The main goal is to avoid c), which is what happens on s390, but
> can also happen elsewhere. Catching b) would be nice as well,
> but is much harder to do from generic code as you'd need an
> architecture specific inline asm statement to insert a ex_table
> fixup, or a runtime conditional on each access.

Or s390 could implement its own inb().

I'm hearing that generic powerpc kernels have to run both on machines
that have I/O port space and those that don't.  That makes me think
s390 could do something similar.

Bjorn