On Thu, Apr 28, 2016 at 11:24:38AM +0200, Alexander Gordeev wrote: > On Wed, Apr 27, 2016 at 06:23:27PM +0200, Alexander Gordeev wrote: > > On Wed, Apr 27, 2016 at 03:46:57PM +0200, Andrew Jones wrote: > > > On Wed, Apr 27, 2016 at 03:13:53PM +0200, Alexander Gordeev wrote: > > > > Make x86 consistent with other architectures and put ... > > > > diff --git a/lib/x86/asm/io.h b/lib/x86/asm/io.h > > > > new file mode 100644 > > > > index 0000000..68b10e5 > > > > --- /dev/null > > > > +++ b/lib/x86/asm/io.h > > > > @@ -0,0 +1,52 @@ > > > > +#ifndef _ASM_X86_IO_H_ > > > > +#define _ASM_X86_IO_H_ > > > > + > > > > +#include "asm/page.h" > > > > > > why is this include here? > > > > Just to conform to other archs which also include it. > > But probably it needs revising in the other archs instead.. > > I guess, I'd better to remove this one at this stage, since I do > not grasp the idea behind all archs header inclusion policy. > It seems it could be cleaned up further, and if so - it is better > to address in a separate patch. Agreed it would be a different patch. But in this case, none needed. The policy is to only include what's needed in source files. Also, header files should only include what's needed [most] of the time. I've made at least one exception, commit 957b6055, and other cases for convenience could probably be made, but case-by- case. Also, in this case, lib/arm[64]/asm/io.h includes asm/page.h because virt_to_phys/phys_to_virt builds on helpers that are in asm/page.h And, if you're wondering about asm/barrier.h, that's because asm-generic/io.h is included at the bottom, and we want to override the generic barriers before defining other generic functions that use them. Thanks, drew -- To unsubscribe from this list: send the line "unsubscribe kvm" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
