Re: problem with blk_queue_bounce_limit()

[Anton Blanchard]

>    But you're creating a new mapping for the old buffer.  What if you had
>    a DMA API implementation which consolidates multiple mapping attempts
>    of the same buffer into a single mapping entry (along with a reference
>    count)?  That would break the workaround.
>    
> I hope nobody is doing this, it would probably break other things
> we haven't considered yet.

Dave, we talked about this ages ago as a possible alternative to skb
recycling and persistent IOMMU mappings for those skbs.

Unfortunately you need a hash to map from all of memory to a pci bus
address for each host bridge (well IOMMU), and so far I cant think of
anything that wouldnt chew gobs of RAM.

Anton