On Mon, Mar 26, 2012 at 02:30:23PM +0900, Simon Horman wrote:
> Signed-off-by: Simon Horman <ho...@verge.net.au>
>
> ---
> v2
> As suggested by Ben Pfaff
> * Where Open Flow 1.2 breaks apart error codes defined
> in previous versions, provide all new definitions to
> previous versions and map the numeric error code to
> the first first definition supplied in ofp-errors.h.
> The case handled so far is:
> OFPERR_OFPBIC_BAD_EXP_TYPE -> { OFPERR_OFPBIC_BAD_EXPERIMENTER,
> OFPERR_OFPBIC_BAD_EXP_TYPE }
> * Correct name of OFPERR_OFPERR_BAD_ROLE, it should be
> OFPERR_OFPRRFC_BAD_ROLE.
> * Where Open Flow 1.2 adds error codes that were previously
> defined as Nicira extension errors define the later in terms
> of the new codes.
>
> Add some missing common openflow definitions
>
> Signed-off-by: Simon Horman <ho...@verge.net.au>
So, the more I looked at this the more a few things bothered me.
First, entirely getting rid of the checks for duplicates seemed risky;
it seemed like an invitation to get things wrong again later. Indeed,
when I added it back temporarily I found very quickly a typo here that
was obvious in retrospect:
/* NX1.0(1,258), NX1.1(1,258), OF1.1(4,7). Unsupported value in a match
field. */
OFPERR_OFPBMC_BAD_VALUE,
(s/OF1.1/OF1.2/ in case it's still not obvious.)
So I decided to improve that, by making explicit declarations of
expected duplicates mandatory.
And then I realized that the forms of the targets started to really
confuse me. OF, OF1.0, OF1.1, OF1.1only, OF1.2, and so on, with
meanings that aren't obvious. I decided that it's better to be more
consistent, so I changed these to, respectively, OF1.0+, OF1.0,
OF1.1+, OF1.1, and OF1.2, where the "+" consistently means "this
version and all later versions" and the absence means "this version
only".
Along the way I discovered and fixed some unit test failures and bugs
in extract-ofp-errors.
Anyway, I'll send this out a new series derived from this patch in a
minute. Simon, please review it.
Once we have that series in I'll look at the rest of this series.
Thanks,
Ben.
_______________________________________________
dev mailing list
dev@openvswitch.org
http://openvswitch.org/mailman/listinfo/dev
