[PATCHv22 14/16] X86: integrate CMA with DMA-mapping subsystem

[Arnd Bergmann]

On Wednesday 22 February 2012, Russell King - ARM Linux wrote:
> On Tue, Feb 21, 2012 at 04:18:02PM -0800, Andrew Morton wrote:
> > After a while I got it to compile for i386.  arm didn't go so well,
> > partly because arm allmodconfig is presently horked (something to do
> > with Kconfig not setting PHYS_OFFSET) and partly because arm defconfig
> > doesn't permit CMA to be set.  Got bored, gave up.
> 
> That's not going to get fixed, unfortunately.  It requires us to find
> some way to force various options to certain states on all*config
> builds, because not surprisingly a value of 'y', 'm' or 'n' doesn't
> work for integer or hex config options.
> 
> So the only way all*config can be used on ARM is with a seed config file
> to force various options to particular states to ensure that we end up
> with a sane configuration that avoids crap like that.
> 
> Alternatively, we need a way to tell kconfig that various options are to
> be set in certain ways in the Kconfig files for all*config to avoid it
> wanting values for hex or int options.

I usually set KCONFIG_ALLCONFIG to a file containing the extra options
I want, for another reason: As long as we are building platforms
separately, all{no,yes,mod}config and randconfig will always build
for the versatile platform instead of something more modern.

We could change that and make mach-vexpress the default, but we can
also wait until we have the initial multiplatform support done and
then use that one.

Building vexpress_defconfig (or most other defconfig, I would expect)
works fine with the CMA series applied. This should also work,
but there a few bugs in unrelated device drivers that would need to get
fixed first.

make ARCH=arm CROSS_COMPILE=arm-linux-gnueabi- -sj3 allmodconfig \ 	
	KCONFIG_ALLCONFIG=$PWD/arch/arm/configs/vexpress_defconfig

	Arnd