Other things we need to do with backend/blkfront Was:Re: Re: [PATCH 2 of 3] interface: Flesh out the BLKIF_OP_DISCARD description

[Konrad Rzeszutek Wilk <konrad.wilk@oracle.com>]

> Could we simplify and say that blkback won't expose discard support
> unless the underlying block device is correctly aligned for it? i.e.
> encourage people to align their underlying storage correctly? Presumably
> doing that has other benefits?

It got me thinking that we could do this - but I do not think that should
be spelled out in the interface. Rather it is up to the backend to either
expose it or not. The check for -1 in backend for that should do it. Keep
in mind that the discard operation is a hint, nothing else.

It also got me thinking about the aligment offset - which we do not
expose to the frontend. That is the one where the 63 sector DOS partition
ends up skewing up the whole disk layout. That is seperate from discard
operations.

It is more of a XenBus attribute. Then there is also the device serial
number which we don't expose either.