On 10/29/07, Blue Swirl <[EMAIL PROTECTED]> wrote:
> We could cache the resolved addresses to overcome the additional setup
> overhead. Each stage should install cache invalidation callbacks or a
> method to call for recalculation of the addresses. For example IOMMU
> or ESPDMA mappings change very
On 10/28/07, Jamie Lokier <[EMAIL PROTECTED]> wrote:
> Blue Swirl wrote:
> > Currently scsi-disk provides a buffer. For true zero copy, this needs
> > to be changed so that instead the buffer is provided by the caller at
> > each stage until we reach the host memory. But I'll use the scsi-disk
> >
Blue Swirl wrote:
> Currently scsi-disk provides a buffer. For true zero copy, this needs
> to be changed so that instead the buffer is provided by the caller at
> each stage until we reach the host memory. But I'll use the scsi-disk
> buffer for now.
This might actually work in Qemu.
But in gene
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 regi
On 10/28/07, Paul Brook <[EMAIL PROTECTED]> wrote:
> > I changed Slirp output to use vectored IO to avoid the slowdown from
> > memcpy (see the patch for the work in progress, gives a small
> > performance improvement). But then I got the idea that using AIO would
> > be nice at the outgoing end of
> I changed Slirp output to use vectored IO to avoid the slowdown from
> memcpy (see the patch for the work in progress, gives a small
> performance improvement). But then I got the idea that using AIO would
> be nice at the outgoing end of the network IO processing. In fact,
> vectored AIO model c
Hi,
I changed Slirp output to use vectored IO to avoid the slowdown from
memcpy (see the patch for the work in progress, gives a small
performance improvement). But then I got the idea that using AIO would
be nice at the outgoing end of the network IO processing. In fact,
vectored AIO model could