Re: [PATCH 2/2] xen/arm: support compressed kernels

[Ian Campbell <ian.campbell@citrix.com>]

On Wed, 2015-08-12 at 16:22 +0100, Stefano Stabellini wrote:
> -obj-bin-$(CONFIG_X86) += $(foreach n,decompress gunzip bunzip2 unxz
> > > unlzma unlzo unlz4 earlycpio,$(n).init.o)
> > > +obj-bin-y += $(foreach n,decompress gunzip bunzip2 unxz unlzma unlzo 
> > > unlz4 earlycpio,$(n).init.o)
> > 
> > I don't think we need/want earlycpio support on ARM (not yet anyway).
> 
> Unfortunately it is not possible to only compile some and not all
> because decompress.c makes use of them all.


earlycpio isn't a decompression algorithm though, it just happens to be
bundled into the same obj-bin. I don't see any sign of cpio support in
decompress.c (since that wouldn't make sense).

> 
> > >  obj-$(perfc)       += perfc.o
> > >  obj-$(crash_debug) += gdbstub.o
> > > diff --git a/xen/include/asm-arm/byteorder.h b/xen/include/asm
> > > -arm/byteorder.h
> > > index 9c712c4..3b7feda 100644
> > > --- a/xen/include/asm-arm/byteorder.h
> > > +++ b/xen/include/asm-arm/byteorder.h
> > > @@ -5,6 +5,8 @@
> > >  
> > >  #include <xen/byteorder/little_endian.h>
> > >  
> > > +#define CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS
> > 
> > While CONFIG_HAVE_UNALIGNED_ACCESS might be true on arm64 it may not be 
> > the
> > case that it is efficient. Also I'm not sure about arm32 at all.
> 
> It is true on arm archs >= v8,

I'm not sure about that.

>  so we should be fine. See:
> 
> http://lwn.net/Articles/540022/

Even this says it is not always the best option, but this is not fast path
code so never mind.

Also if it isn't true for armv7 you shouldn't set it unconditionally, it
needs either an ifdef or for arm{32,64}/byteorder.h to have it.

On x86 this #define is in config.h, please lets be consistent. Also that
makes more sense anyway since alignment != byteorder.

Ian.