Seems reasonable.
--Justin
On Jul 19, 2012, at 12:24 AM, Ben Pfaff wrote:
> I don't think this corner case can come up in a real OpenFlow message,
> because the presence of the OpenFlow header guarantees that the ofpbuf's
> data is nonnull, but it did in a simple test that is coming up in a few
> commits.
>
> Signed-off-by: Ben Pfaff <b...@nicira.com>
> ---
> lib/nx-match.c | 12 ++++++++----
> 1 files changed, 8 insertions(+), 4 deletions(-)
>
> diff --git a/lib/nx-match.c b/lib/nx-match.c
> index 3e3b4c6..8b2a822 100644
> --- a/lib/nx-match.c
> +++ b/lib/nx-match.c
> @@ -99,6 +99,14 @@ nx_pull_match__(struct ofpbuf *b, unsigned int match_len,
> bool strict,
>
> assert((cookie != NULL) == (cookie_mask != NULL));
>
> + cls_rule_init_catchall(rule, priority);
> + if (cookie) {
> + *cookie = *cookie_mask = htonll(0);
> + }
> + if (!match_len) {
> + return 0;
> + }
> +
> p = ofpbuf_try_pull(b, ROUND_UP(match_len, 8));
> if (!p) {
> VLOG_DBG_RL(&rl, "nx_match length %u, rounded up to a "
> @@ -107,10 +115,6 @@ nx_pull_match__(struct ofpbuf *b, unsigned int
> match_len, bool strict,
> return OFPERR_OFPBMC_BAD_LEN;
> }
>
> - cls_rule_init_catchall(rule, priority);
> - if (cookie) {
> - *cookie = *cookie_mask = htonll(0);
> - }
> for (;
> (header = nx_entry_ok(p, match_len)) != 0;
> p += 4 + NXM_LENGTH(header), match_len -= 4 + NXM_LENGTH(header)) {
> --
> 1.7.2.5
>
> _______________________________________________
> dev mailing list
> dev@openvswitch.org
> http://openvswitch.org/mailman/listinfo/dev
_______________________________________________
dev mailing list
dev@openvswitch.org
http://openvswitch.org/mailman/listinfo/dev
