Re: [Qemu-devel] [PATCH 1/9] Add stub functions for PCI device models to do PCI DMA

[Anthony Liguori <aliguori@us.ibm.com>]

On 10/02/2011 07:14 AM, Michael S. Tsirkin wrote:
> On Sun, Oct 02, 2011 at 02:01:10PM +0200, Avi Kivity wrote:
>>> Hmm, not entirely virtio specific, some devices use stX macros to do the
>>> conversion.  E.g. stw_be_phys and stl_le_phys are used in several
>>> places.
>>
>> These are fine - explicit endianness.
>
> Right. So changing these to e.g. stl_dma and assuming
> LE is default seems like a step backwards.

We're generalizing too much.

In general, the device model doesn't need atomic access functions.  That's 
because device model RAM access is not coherent with CPU RAM access.

Virtio is a very, very special case.  virtio requires coherent RAM access.

What we really ought to do is have a variant of the map API that allows for an 
invalidation callback.  That would allow us to map the ring for longer periods 
of time and then we could access the memory directly.

That would not only solve this problem but also improve overall performance.

Regards,

Anthony Liguori