Messages in this thread | | | Date | Wed, 15 Aug 2012 00:38:32 +0300 | From | "Michael S. Tsirkin" <> | Subject | Re: [PATCH v7 2/4] virtio_balloon: introduce migration primitives to balloon pages |
| |
On Tue, Aug 14, 2012 at 04:56:59PM -0400, Rik van Riel wrote: > On 08/14/2012 04:54 PM, Michael S. Tsirkin wrote: > > >To clarify, the global state that this patch adds, is ugly > >even if we didn't support multiple balloons yet. > >So I don't think I can accept such a patch. > >Rusty has a final word here, maybe he thinks differently. > > Before deciding that "does not support multiple balloon drivers > at once" is an issue, is there any use case at all for having > multiple balloon drivers active at a time? > > I do not see any.
For example, we had a proposal for a page-cache backed device. So it could be useful to have two, a regular balloon and a pagecache backed one. There could be other uses - it certainly looks like it works so how can you be sure it's useless?
And even ignoring that, global pointer to a device is an ugly hack and ugly hacks tend to explode.
And even ignoring estetics, and if we decide we are fine with a single balloon, it needs to fail gracefully not crash like it does now.
> -- > All rights reversed
| |