clearly mark DMA_OPS support as an architecture feasture

[Christoph Hellwig]

Hi all,

we've had a long standing problems where drivers try to hook into the
DMA_OPS mechanisms to override them for something that is not DMA, or
to introduce additional dispatching.

Now that we are not using DMA_OPS support for dma-iommu and can build
kernels without DMA_OPS support on many common setups this becomes even
more problematic.

This series renames the option to ARCH_DMA_OPS and adds very explicit
comment to not use it in drivers.  The ipu6 and vdpa_sim/user drivers
that abuse the mechanism are made to depend on the option instead of
selecting it with a big comment, but I expect this to be fixed rather
sooner than later (I know the ipu6 maintainers are on it based on a
previous discussion).