Re: [Qemu-devel] [PATCH 2/2] Correct definition of NBD_CMD_FLAG_FUA

[Eric Blake <eblake@redhat.com>]

[-- Attachment #1: Type: text/plain, Size: 1243 bytes --]

On 03/31/2016 12:21 PM, Alex Bligh wrote:
> 
> On 31 Mar 2016, at 19:15, Alex Bligh <alex@alex.org.uk> wrote:
> 
>> It is doubtful whether anyone is using NBD_CMD_FLAG_FUA
>> at the moment in any case.
> 
> Drat. I spoke too soon. Qemu uses it, but presumably from its
> own .h file.

Yes, qemu has its own nbd.h, which still has nbd_request with a single
uint32_type that holds both flags and command type.  It wouldn't be too
hard to rework that to more closely match upstream NBD.

> 
> However, it's now nonsensical having it defined as 1<<16 in a
> 16 bit flags variable.

I don't see any problem with your patch on the NBD project side of
things; it's not like 'make install' is dumping a header into
/usr/include for client programs to reuse (which is _why_ qemu is using
its own nbd.h), because no one has really churned out an NBD-client
library for embedding in larger programs.

> 
> Should we produce a new name for it (and future command flags)
> that aren't shifted left 16 places, and just maintain the
> current value for compatibility?

I don't see the point.  Your fix looks correct.

-- 
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 --]