Re: [Nbd] Processing client's option list

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

On Sun, May 29, 2011 at 05:18:45PM +0200, Goswin von Brederlow wrote:
> Wouter Verhelst <w@...112...> writes:
> 
> > On Sun, May 29, 2011 at 12:48:12PM +0100, Alex Bligh wrote:
> >> --On 29 May 2011 13:33:23 +0200 Goswin von Brederlow
> >> <goswin-v-b@...186...> wrote:
> >> >Currently NBD_OPT_EXPORT_NAME is the only possible option and it is also
> >> >required. How about declaring that NBD_OPT_EXPORT_NAME must be the last
> >> >option being send and ends the negotiation?
> >> 
> >> With the current negotiation where all the client's options are sent at
> >> once, it would be fine. We could even use the current number for
> >> NBD_OPT_EXPORT_NAME for NBD_END_OPTIONS (or similar) and say that
> >> optionally it can take an export name for legacy compatibility.
> >> 
> >> We could then define NBD_OPT_EXPORT_NAME as some other value.
> >
> > I don't really like that. It muddles which is which, and could cause
> > confusion down the line. But there's not really a problem with stating
> > that NBD_OPT_EXPORT_NAME has to be the last option for now, and we can
> > deal with it if and when that ever needs to change.
> 
> Is the new protocol in use anywhere relevant?

It's been part of gentoo and Debian for a while, at least. I'm not going to
define what "relevant" is, that's not for me to decide.

> How long has it been out there in the wild?

It's been released in February.

> Maybe this could just be fixed without backward compatibility.

Not Going To Happen(tm)

-- 
The volume of a pizza of thickness a and radius z can be described by
the following formula:

pi zz a