[PATCH] arm64: dma: Drop cache invalidation from arch_dma_prep_coherent()

[Catalin Marinas]

On Tue, Aug 23, 2022 at 01:21:11PM +0100, Will Deacon wrote:
> arch_dma_prep_coherent() is called when preparing a non-cacheable region
> for a consistent DMA buffer allocation. Since the buffer pages may
> previously have been written via a cacheable mapping and consequently
> allocated as dirty cachelines, the purpose of this function is to remove
> these dirty lines from the cache, writing them back so that the
> non-coherent device is able to see them.
> 
> On arm64, this operation can be achieved with a clean to the point of
> coherency; a subsequent invalidation is not required and serves little
> purpose in the presence of a cacheable alias (e.g. the linear map),
> since clean lines can be speculatively fetched back into the cache after
> the invalidation operation has completed.
> 
> Relax the cache maintenance in arch_dma_prep_coherent() so that only a
> clean, and not a clean-and-invalidate operation is performed.
> 
> Cc: Catalin Marinas <catalin.marinas at arm.com>
> Cc: Mark Rutland <mark.rutland at arm.com>
> Cc: Robin Murphy <robin.murphy at arm.com>
> Cc: Christoph Hellwig <hch at lst.de>
> Cc: Ard Biesheuvel <ardb at kernel.org>
> Signed-off-by: Will Deacon <will at kernel.org>
> ---
> 
> I'm slightly wary about this change as other architectures seem to do
> clean+invalidate here, but I'd like to hear what others think in any
> case.

Given that we still have the cacheable alias around, I don't see much
point in the invalidation. So:

Reviewed-by: Catalin Marinas <catalin.marinas at arm.com>

(I was wondering why not just invalidate without clean but it could be
that the allocated memory was zeroed and we want that to make it to the
PoC)

-- 
Catalin