Thanks, I'll push this series in a minute.

On Mon, Aug 11, 2014 at 02:11:02PM -0700, Jarno Rajahalme wrote:
> Acked-by: Jarno Rajahalme <jrajaha...@nicira.com>
> 
> On Aug 11, 2014, at 1:44 PM, Ben Pfaff <b...@nicira.com> wrote:
> 
> > Based on the OpenFlow 1.5 draft.
> > 
> > ONF-JIRA: EXT-425
> > Signed-off-by: Ben Pfaff <b...@nicira.com>
> > ---
> > lib/ofp-actions.c    |    2 +-
> > lib/ofp-errors.h     |    4 ++++
> > tests/ofp-actions.at |   12 +++++++++---
> > 3 files changed, 14 insertions(+), 4 deletions(-)
> > 
> > diff --git a/lib/ofp-actions.c b/lib/ofp-actions.c
> > index 8f3af19..cdcd72d 100644
> > --- a/lib/ofp-actions.c
> > +++ b/lib/ofp-actions.c
> > @@ -1965,7 +1965,7 @@ decode_OFPAT_RAW12_SET_FIELD(const struct 
> > ofp12_action_set_field *oasf,
> >     }
> > 
> >     if (NXM_HASMASK(oxm_header)) {
> > -        return OFPERR_OFPBAC_BAD_SET_TYPE;
> > +        return OFPERR_OFPBAC_BAD_SET_MASK;
> >     }
> >     mf = mf_from_nxm_header(oxm_header);
> >     if (!mf) {
> > diff --git a/lib/ofp-errors.h b/lib/ofp-errors.h
> > index d5c767b..e32b751 100644
> > --- a/lib/ofp-errors.h
> > +++ b/lib/ofp-errors.h
> > @@ -222,6 +222,10 @@ enum ofperr {
> >      * unsupported value, or modifies a read-only field. */
> >     OFPERR_OFPBAC_BAD_SET_ARGUMENT,
> > 
> > +    /* NX1.3-1.4(14), OF1.5+(2,16).  Field in Set-Field action has 
> > Has-Mask bit
> > +     * set to 1. */
> > +    OFPERR_OFPBAC_BAD_SET_MASK,
> > +
> >     /* NX1.0-1.1(2,256), NX1.2+(11).  Must-be-zero action argument had 
> > nonzero
> >      * value. */
> >     OFPERR_NXBAC_MUST_BE_ZERO,
> > diff --git a/tests/ofp-actions.at b/tests/ofp-actions.at
> > index 569f4ae..e9bd69a 100644
> > --- a/tests/ofp-actions.at
> > +++ b/tests/ofp-actions.at
> > @@ -502,9 +502,8 @@ AT_CHECK(
> >   [0], [expout], [experr])
> > AT_CLEANUP
> > 
> > -dnl Our primary goal here is to verify that the copy_field action gets
> > -dnl used instead of the Nicira extension reg_move in OpenFlow 1.5, so
> > -dnl the list of action tests is very short.
> > +dnl Our primary goal here is to verify that OpenFlow 1.5-specific changes,
> > +dnl so the list of tests is short.
> > AT_SETUP([OpenFlow 1.5 action translation])
> > AT_KEYWORDS([ofp-actions OF1.5])
> > AT_DATA([test-data], [dnl
> > @@ -514,6 +513,13 @@ AT_DATA([test-data], [dnl
> > # actions=move:NXM_OF_IN_PORT[]->NXM_OF_VLAN_TCI[]
> > 001c 0018 0010 0000 0000 0008 00000002 00000802 00000000
> > 
> > +dnl This action has a lot more wrong with it than the hasmask bit, but
> > +dnl the current OVS implementation checks for that first.
> > +# bad OpenFlow15 actions: OFPBAC_BAD_SET_MASK
> > +& ofp_actions|WARN|bad action at offset 0 (OFPBAC_BAD_SET_MASK):
> > +& 00000000  00 19 00 08 00 00 01 00-
> > +0019 0008 0000 0100
> > +
> > ])
> > sed '/^[[#&]]/d' < test-data > input.txt
> > sed -n 's/^# //p; /^$/p' < test-data > expout
> > -- 
> > 1.7.10.4
> > 
> > _______________________________________________
> > dev mailing list
> > dev@openvswitch.org
> > http://openvswitch.org/mailman/listinfo/dev
> 
_______________________________________________
dev mailing list
dev@openvswitch.org
http://openvswitch.org/mailman/listinfo/dev

Reply via email to