Re: [PATCH v5 RESEND 2/5] lib: move pci_iomap.c to drivers/pci/

[Philipp Stanner <pstanner@redhat.com>]

On Tue, 2024-01-23 at 14:20 -0600, Bjorn Helgaas wrote:
> On Thu, Jan 11, 2024 at 09:55:37AM +0100, Philipp Stanner wrote:
> > This file is guarded by an #ifdef CONFIG_PCI. It, consequently,
> > does not
> > belong to lib/ because it is not generic infrastructure.
> > 
> > Move the file to drivers/pci/ and implement the necessary changes
> > to
> > Makefiles and Kconfigs.
> > ...
> 
> > --- a/drivers/pci/Kconfig
> > +++ b/drivers/pci/Kconfig
> > @@ -13,6 +13,11 @@ config FORCE_PCI
> >         select HAVE_PCI
> >         select PCI
> >  
> > +# select this to provide a generic PCI iomap,
> > +# without PCI itself having to be defined
> > +config GENERIC_PCI_IOMAP
> > +       bool
> 
> > --- a/lib/pci_iomap.c
> > +++ b/drivers/pci/iomap.c
> > @@ -9,7 +9,6 @@
> >  
> >  #include <linux/export.h>
> >  
> > -#ifdef CONFIG_PCI
> 
> IIUC, in the case where CONFIG_GENERIC_PCI_IOMAP=y but CONFIG_PCI was
> not set, pci_iomap.c was compiled but produced no code because the
> entire file was wrapped with this #ifdef.
> 
> But after this patch, it looks like pci_iomap_range(),
> pci_iomap_wc_range(), etc., *will* be compiled?
> 
> Is that what you intend, or did I miss something?

They *will* be compiled when BOTH, CONFIG_PCI and
CONFIG_GENERIC_PCI_IOMAP have been set. It's a bit hard to see that in
the patch's diff. Here, look closely:

--- a/drivers/pci/Makefile
+++ b/drivers/pci/Makefile
@@ -14,6 +14,7 @@ ifdef CONFIG_PCI   <-----------
 obj-$(CONFIG_PROC_FS)		+= proc.o
 obj-$(CONFIG_SYSFS)		+= slot.o
 obj-$(CONFIG_ACPI)		+= pci-acpi.o
+obj-$(CONFIG_GENERIC_PCI_IOMAP) += iomap.o  <------------
 endif

So if I am not mistaken it behaves 100% as it did before

I prefered Makefile-logic over even more C Preprocessor to implement
that. The preprocessor has caused us so much trouble... :(

P.


> 
> Bjorn
>