Am 17.07.2017 um 13:36 hat Eric Blake geschrieben: > On 07/14/2017 02:13 PM, Nir Soffer wrote: > > Eric, we are more concerned about using the current qemu version. > > > > We can use the fact that providing both size and backing format, > > qemu does not open the backing file, but this is not documented, and > > we don't want to base oVirt code on undocumented behavior. > > > > What we would like to have is: > > - qemu blessing for using this undocumented behaviour > > But how are you going to get that? By the time someone writes a > documentation patch, it won't land until qemu 2.10, but by then, they > might as well have written the -u patch instead.
In fact, soft freeze is tomorrow, so if we want to have -u in 2.10, it would be good to have a patch ready today. > Older versions will continue to have the older behavior, unless someone > backports a patch to give them the newer behavior - but if someone is > backporting -u, presumably they will also backport whatever > introspection mechanisms you would also use against upstream qemu to > learn if -u is present. If the introspection mechanism is not present > or gives the answer that -u is not present, then you can safely assume > the old behavior (because the new behavior is only going to be present > intentionally). But that is STILL not something that we are going to > explicitly document, because it makes more sense to implement the > working feature than to document the workaround to a missing feature. While I think adding -u today is reasonably realistic, I'm doubtful that we can get an introspection mechanism in place today. Perhaps we can declare it a bug fix, but I'd rather not rush something like that. How does libvirt detect qemu-img features today? Just try and then check the error message? Kevin
pgpi9aIveg1Wj.pgp
Description: PGP signature
