Re: [RFC PATCH 1/3] dma-mapping: make overriding GFP_* flags arch customizable

[Christoph Hellwig]

On Fri, Sep 27, 2019 at 02:33:14AM +0200, Halil Pasic wrote:
> Thank you for your feedback. Just to be sure we are on the same pager, I
> read commit a0be1db4304f like this:
> 1) virtio_pci_legacy needs to allocate the virtqueues so that the base
> address fits 44 bits
> 2) if 64 bit dma is possible they set coherent_dma_mask to
>   DMA_BIT_MASK(44) and dma_mask to DMA_BIT_MASK(64)
> 3) since the queues get allocated with coherent allocations 1) is
> satisfied
> 4) when the streaming mappings see a buffer that is beyond
>   DMA_BIT_MASK(44) then it has to treat it as not coherent memory
>   and do the syncing magic (which isn't actually required, just
>   a side effect of the workaround.

1-3 is correct, 4 is not.  The coherent mask is a little misnamed and
doesn't have to anything with coherency.  It is the mask for DMA
allocations, while the dma mask is for streaming mappings.

> I've already implemented a patch (see after the scissors line) that
> takes a similar route as commit a0be1db4304f, but I consider that a
> workaround at best. But if that is what the community wants... I have to
> get the job done one way or the other.

That patch (minus the comments about being a workaround) is what you
should have done from the beginning.