Re: V4V

[Jean Guyader <jean.guyader@gmail.com>]

On 25 May 2012 10:48, Stefano Stabellini
<stefano.stabellini@eu.citrix.com> wrote:
>
> On Thu, 24 May 2012, Jean Guyader wrote:
> > Some of the downsides to using the shared memory grant method:
> >     - This method imposes an implicit ordering on domain destruction.
> >       When this ordering is not honored the grantor domain cannot shutdown
> >       while the grantee still holds references. In the extreme case where
> >       the grantee domain hangs or crashes without releasing it granted
> >       pages, both domains can end up hung and unstoppable (the DEADBEEF
> >       issue).
>
> Is it still true? This looks like a serious issue.
>

I have tried to repro this issue the order day with libvchan but I
couldn't so it's probably fixed.
I suspect it has something to do with grant-ref and I don't think
libvchan uses grant refs.

>
> >     - You can't trust any ring structures because the entire set of pages
> >       that are granted are available to be written by the either guest.
> >     - The PV connect/disconnect state-machine is poorly implemented.
> >       There's no trivial mechanism to synchronize disconnecting/reconnecting
> >       and dom0 must also allow the two domains to see parts of xenstore
> >       belonging to the other domain in the process.
>
> We are starting to see this problem, trying to setup driver domains with
> libxl.
>
>
> >     - Using the grant-ref model and having to map grant pages on each
> >       transfer cause updates to V->P memory mappings and thus leads to
> >       TLB misses and flushes (TLB flushes being expensive operations).
>
> [snip]
>
> > I've done some benchmarks on V4V and libchan and the results were
> > pretty close between the the two if you use the same buffer size in both cases.
>
> It is strange that you cannot see any performance advantages using v4v. I
> was expecting quite a difference, especially on new numa machines.

The numbers for both system were in the 50MB/s 60MB/s ranges from domU to domU.
That was on a out of the box testing on a desktop type machine I
didn't look at making
any kind of tricks or adjustment to make it go faster.

Jean