On Tue, Mar 8, 2011 at 1:31 PM, Kevin Wolf <kw...@redhat.com> wrote: > Am 08.03.2011 14:13, schrieb Stefan Hajnoczi: >> On Tue, Mar 8, 2011 at 11:47 AM, Kevin Wolf <kw...@redhat.com> wrote: >>> + 20 - 23: cluster_bits >>> + Number of bits that are used for addressing an offset >>> + within a cluster (1 << cluster_bits is the cluster >>> size) >>> + >>> + 24 - 31: size >>> + Virtual disk size in bytes >> >> Any constraints on these two fields that should be mentioned? > > For the size not that I'm aware of. > > For cluster_bits qemu restricts it to 512 <= cluster_size <= 2 MB. I > think we should add 512 as a lower limit, anything smaller doesn't make > sense and steals us bits that we want to use for flags. > > The 2 MB are more of an implementation limitation. Would you mention it > here? The format shouldn't have any problem with larger sizes.
It could be mentioned as an explicit implementation limit so that a third party implementing qcow2 support from scratch doesn't use the format in ways that won't work with QEMU's implementation. Stefan
