Re: [RFC PATCH 2/2] asm-generic: Add new pci.h and use it

[Stafford Horne <shorne@gmail.com>]

On Fri, Jul 15, 2022 at 10:09:21AM +0200, Arnd Bergmann wrote:
> On Thu, Jul 14, 2022 at 11:46 PM Stafford Horne <shorne@gmail.com> wrote:
> >
> > The asm/pci.h used for many newer architectures share similar
> > definitions.  Move the common parts to asm-generic/pci.h to allow for
> > sharing code.
> 
> This looks very nice, thanks for doing it!
> 
> > Two things to note are:
> >
> >  - isa_dma_bridge_buggy, traditionally this is defined in asm/dma.h but
> >    these architectures avoid creating that file and add the definition
> >    to asm/pci.h.
> 
> I would prefer if we could just kill off this variable for non-x86, as it's
> only set to a nonzero value in two implementations that are both
> x86-specific and most of the references are gone. That does not have
> to be part of this series though, if you don't want to address it here, just
> add a comment to the new pci.h file.

I will look at it, maybe in a v3.

> >  - ARCH_GENERIC_PCI_MMAP_RESOURCE, csky does not define this so we
> >    undefine it after including asm-generic/pci.h.  Why doesn't csky
> >    define it?
> 
> Adding David Woodhouse to Cc, as he introduced this interface. As I
> understand it, this was meant as a replacement for the old
> architecture specific pci_mmap_page_range interface, and is ideally
> used everywhere.
> 
> It's probably something that slipped through the review of csky and
> should have been there.
> 
> As an aside, it seems the pci_mmap_page_range() cleanup was
> left almost complete, with sparc being the only one left after
> David Miller found a problem with the generic code. Not sure if
> this was ever resolved:
> https://lore.kernel.org/lkml/1519887203.622.3.camel@infradead.org/t/#u

I see, I will leave it foe now.

> > +#ifndef PCIBIOS_MIN_IO
> > +#define PCIBIOS_MIN_IO         0
> > +#endif
> > +
> > +#define PCIBIOS_MIN_MEM                0
> 
> We should probably #ifdef both of these for consistency.

OK.

> > +static inline int pci_get_legacy_ide_irq(struct pci_dev *dev, int channel)
> > +{
> > +       /* no legacy ide irq support */
> > +       return -ENODEV;
> > +}
> 
> And this can just go away now, according to what we found.

Yeah, that will be nice.

-Stafford