Re: [Qemu-devel] Faster, generic IO/DMA model with vectored AIO?

[Blue Swirl]

I made a new patch sketching the system. It doesn't even compile, but
it should give a view how this would be put into work.

On the down side, new memory needs to be allocated for generation of
new vectors from previous ones, that may kill some of the performance.
Also, supporting DMA to MMIO registers can't be done with pure
translation only.

I'm not too happy about this model anymore, maybe the model I proposed
earlier is better. But it can't get to zero copy.

gdma_aiov.diff
Description: Text Data