Hi,

I am thinking about compressed block device layer -- implemented in much
the same fashion as cgd/ccd/vnd.
This layer will use some underlying device and provide virtual block device of larger size.
It will store its content on underlying device in compressed form.
Compression will probably be per block basis.
In case compression ratio appears to be overestimated and real device got full,
layer would issue error notification to syslog and fallback to use swappable memory to
store its content.  This point of its behaviour is the most obscure for me currently.

How do you think can this layer be useful?  Is it worth implementing?