Rusty Russell wrote:
Yeah, every time I get close to benchmarking I find another bug :( But I've
spent some time optimising the normal lguest net device, so we'll have a fair
comparison.
In theory vringfd will get us zero copy from guest sendfile out to external
machines. For anything else
On Tuesday 17 June 2008 00:02:55 Anthony Liguori wrote:
There's nothing that prevents zero-copy to be implemented for tun
without vringfd. In fact, I seem to recall that your earlier patches
implemented zero-copy :-)
They didn't actually work. You need to block until the data isn't being
Rusty Russell wrote:
On Sunday 15 June 2008 09:28:34 Anthony Liguori wrote:
Have you benchmarked the driver? I wasn't seeing great performance
myself although I think that was due to some bugs in the vringfd code.
Yeah, every time I get close to benchmarking I find another bug :(
Where a VLAN consists only of a single tap interface on the
host side and a single virtio interface on the guest side,
we can use the tun/tap driver's vringfd support to eliminate
copies.
We set up a vringfd for both rings and pass them to the
tun/tap driver. When the guest informs us of the