Re: [Nbd] Proposal to merge WRITE_ZEROES extension into master

[Wouter Verhelst <w@...112...>]

On Wed, Dec 14, 2016 at 01:02:37PM -0600, Eric Blake wrote:
> On 12/14/2016 12:52 PM, Wouter Verhelst wrote:
> > On Wed, Dec 14, 2016 at 06:20:57PM +0000, Alex Bligh wrote:
> >> Wouter,
> >>> Did the implementation of WRITE_ZEROES encounter any issues with the spec
> >>> that we might want to fix up? Or did things work out as expected?
> >>
> >> It was fine in nbdserver, and also fine in gonbdserver. Admittedly I didn't
> >> have a 'real' client to test it against.
> > 
> > Oh, I missed that part.
> > 
> > Is there a client anywhere that is more than a test thing? If not, then perhaps
> > hold off on merging for now until that exists too.
> 
> Qemu 2.8 (currently at -rc3, not officially released yet but should be
> soon, but none of the pending patches between -rc3 and the final release
> will affect nbd behavior) has both a client and a server implementation,
> and I'm also set up to turn on tracing in either the client or the
> server in order to trace what the other side sends over the wire.

That's good enough for me, thanks.

> An arbitrary client is a bit harder to test if you don't know how to provoke
> it into sending write zero commands, but qemu-io serves as a nice testbed
> that lets the qemu client send arbitrary commands to any server.

-- 
< ron> I mean, the main *practical* problem with C++, is there's like a dozen
       people in the world who think they really understand all of its rules,
       and pretty much all of them are just lying to themselves too.
 -- #debian-devel, OFTC, 2016-02-12