Re: [Qemu-devel] [Nbd] [PATCH v2] doc: Add NBD_CMD_BLOCK_STATUS extension

[Eric Blake <eblake@redhat.com>]

[-- Attachment #1: Type: text/plain, Size: 2675 bytes --]

On 04/04/2016 03:04 PM, Denis V. Lunev wrote:
> In v1 we have had 'status' field, which can have the
> following values for dirty request:
> 
> +      - `NBD_STATE_DIRTY` (0x0), LBA extent is dirty;
> +      - `NBD_STATE_CLEAN` (0x1), LBA extent is clean.
> 
> in the extent structure:
> 
> +        * 64 bits, offset (unsigned)
> +        * 32 bits, length (unsigned)
> +        * 16 bits, status (unsigned)

Between v1 and v2, we dropped 64-bit offset (offset is now implied, by
adding lengths of all earlier descriptors), and widened status from 16
bits to 32 bits (so that each descriptor is now naturally 8-byte aligned
and therefore easier to make a C array).

> 
> with an additional NBD_STATE_DIRTY_HOLE or (DIRTY_DEALLOCATED)
> we could report the entire state using one query. The user could be
> able to read entire state which is useful for backup at once.
> 
> Your current proposal is more tricky and it was misunderstood by Alex:
> 
> +        * 32 bits, status flags
> 
> and you describe flags as
> 
> +      - `NBD_STATE_HOLE` (bit 0); if set, the block represents a hole

> +      - `NBD_STATE_ZERO` (bit 1), if set, the block contents read as

> 
> +      - `NBD_STATE_CLEAN` (bit 2); if set, the block represents a

> and opinion of Alex was that all 3 bits could be set in reply to
> NBD_CMD_BLOCK_STATUS
> with NBD_CMD_FLAG_STATUS_DIRTY set.
> 
> This confused him. This confuses me too.

There's nothing that says that NBD_STATE_CLEAN can't be reassigned to
bit 0.  Conversely, we may want to add a future NBD_CMD_FLAG_FOO that
lets you read allocation and dirty information in the same call, in
which case having the bits be distinct will make that easier; but where
we would also make it obvious that the server is allowed to reject that
command flag as unsupported (we already state the server can reject
NBD_CMD_FLAG_STATUS_DIRTY with EINVAL as unsupported; and that it if
does not reject a dirtiness request but cannot otherwise report
anything, then the entire region is reported as dirty).

I don't have any strong opinions on whether NBD_STATE_CLEAN should
remain bit 2 or be renumbered to bit 0, although renumbering it to bit 0
would make it painfully obvious that you cannot query allocation and
dirtiness at the same time.

> 
> If allocated state is not replied to command with NBD_CMD_FLAG_STATUS_DIRTY
> then why to have different meaning of bits.

Because we still have room - no need to overlap the meaning of bits as
long as we have more bits to choose from.

-- 
Eric Blake   eblake redhat com    +1-919-301-3266
Libvirt virtualization library http://libvirt.org


[-- Attachment #2: OpenPGP digital signature --]
[-- Type: application/pgp-signature, Size: 604 bytes --]