[PATCH v2 00/15] PCI io.h clean-up

[nico@fluxnic.net (Nicolas Pitre)]

On Sat, 14 Jul 2012, Arnd Bergmann wrote:

> On Saturday 14 July 2012, Rob Herring wrote:
> > From: Rob Herring <rob.herring@calxeda.com>
> > 
> > This is the 2nd part of mach/io.h removals. This series removes io.h on
> > platforms with PCI by creating a fixed virtual I/O mapping and a common
> > __io() macro.
> > 
> > This version has changed a bit to accommodate Tegra converting its PCIe
> > host to a platform driver. Now the virtual space is only reserved during
> > early boot before .map_io() is called. The mapping is not created until
> > calling pci_ioremap_io which can be done at any point after vmalloc is
> > initialized.
> > 
> > I've gone back to fixed 64K windows for each PCI bus. This allows
> > removing all the i/o resource setup from the individually platforms and
> > placing it within the common ARM PCI code.
> > 
> > I've only tested versatilepb under qemu (with the model hacked up to
> > actually enable i/o space), so any testing is appreciated. iop3xx and
> > mv78xx0 have some risk of breaking as the PCI bus addresses are moved
> > to 0 from matching the cpu host bus addesss.
> > 
> > This is available from my git tree:
> > 
> > git://sources.calxeda.com/kernel/linux.git io-cleanup-pci
> 
> I would propose sticking it in the arm-soc tree as a staging branch for
> now to get some more testing on this. We can then decide in the merge
> window if we consider it good enough or whether we want to delay it
> to 3.7.
> 
> Any objections?

I suggested a cosmetic fix I'd like to see before this goes into 
mainline, but no objections otherwise.


Nicolas