On Wed, Feb 19, 2014 at 03:24:21PM +0000, Liam Girdwood wrote:
On Thu, 2014-02-20 at 00:00 +0900, Mark Brown wrote:
But lots of architectures do actually have these operations, I suspect looking at some of the existing users depending on PCI is enough if excessively strict (this will improve build coverage which tends to be useful even if the driver can't be run).
I know that the other architectures will implement their own ops, but no other architecture other than X86 will have Intel SST.
Yes, it should be something like
depends on (X86 || COMPILE_TEST) && WHATEVER
if there's a WHATEVER needed to guarantee the operation is there (I think PCI does the job). That way only people actively looking to build test will see the option on !X86 but people doing wider scale work can get build coverage more easily.
I can either send a new patch that additionally includes linux/io.h or send a V2 ?
The include of linux/io.h is needed no matter what - if it's being implicitly included on x86 then someone could change the headers and break it later on. If you can send a new patch right now that'd be great.