Re: [PATCH 12/21] dma-iommu: factor atomic pool allocations into helpers

[Robin Murphy]

On 18/04/2019 17:35, Christoph Hellwig wrote:
> On Thu, Apr 18, 2019 at 04:06:56PM +0100, Robin Murphy wrote:
> > OK, I'm still looking at mmap and get_sgtable, but for now I've pushed out
> > a partial branch that consolidates alloc and free in a way which makes
> > sense to me:
> >
> >    git://linux-arm.org/linux-rm  dma/rework
> >
> > Please let me know what you think.
>
>  From a very high level POV this looks ok, but sometimes a bit to
> convoluted to me.  The major issue why I went with the version I posted
> is that I can cleanly ifdef out the remap code in just a few sections.
> In this version it is spread out a lot more, and the use of IS_ENABLED
> means that we'd need a lot more stubs for functionality that won't
> ever be called but needs to be compilable.

OK, for some reason I'd convinced myself that mmap and get_sgtable would 
need changes to properly handle the new non-remapped CMA case - not sure 
how I reached that conclusion. On inspection they do appear to be broken 
for the non-coherent atomic case, but AFAICS that's been so forever...

Still, I've worked in the vm_map_pages() stuff pending in MM and given 
them the same treatment to finish the picture. Both x86_64_defconfig and 
i386_defconfig do indeed compile and link fine as I expected, so I 
really would like to understand the concern around #ifdefs better.

Robin.