On Mon, Aug 18, 2014 at 10:34:46PM +0100, Liviu Dudau wrote:

On Mon, Aug 18, 2014 at 03:26:04PM +0100, Catalin Marinas wrote:

quoted

On Tue, Aug 12, 2014 at 05:25:16PM +0100, Liviu Dudau wrote:

quoted

Some architectures do not have a simple view of the PCI I/O space
and instead use a range of CPU addresses that map to bus addresses.
For some architectures these ranges will be expressed by OF bindings
in a device tree file.

This patch introduces a pci_register_io_range() helper function with
a generic implementation that can be used by such architectures to
keep track of the I/O ranges described by the PCI bindings. If the
PCI_IOBASE macro is not defined that signals lack of support for PCI
and we return an error.

In order to retrieve the CPU address associated with an I/O port, a
new helper function pci_pio_to_address() is introduced. This will
search in the list of ranges registered with pci_register_io_range()
and return the CPU address that corresponds to the given port.

Cc: Grant Likely <redacted>
Cc: Rob Herring <robh+dt@kernel.org>
Cc: Arnd Bergmann <arnd@arndb.de>
Signed-off-by: Liviu Dudau <Liviu.Dudau@arm.com>
---
 drivers/of/address.c       | 95 ++++++++++++++++++++++++++++++++++++++++++++++
 include/linux/of_address.h |  2 +
 2 files changed, 97 insertions(+)

Actually, I think we moved too much code into drivers/of/. Is there
anything OF specific about this code to live here (apart from
OF_BAD_ADDR)? I think driver/pci/ more appropriate?

I have no strong preference for one place vs the other. While this does
not use any OF specific code, it should be called only from OF enabled
code. So the relationship is not one of depending on of_* but on being
dependent on.

In this case it's probably ok to keep them here for the time being.

I wonder whether it would also be useful for ACPI. But I have no idea
how ACPI would fit with the ACPI patches.

-- 
Catalin