MLD group state string conversion fix

[sebastien.bini_stormshield.eu]

Summary

The ipv6 MLD group state string conversion function is broken (only used if KTR is turned on).

This merely adds the missing entry, to comply with the enum described on sys/netinet6/mld6_var.h:43:

/*
 * MLD per-group states.
 */
#define MLD_NOT_MEMBER			0 /* Can garbage collect group */
#define MLD_SILENT_MEMBER		1 /* Do not perform MLD for group */
#define MLD_REPORTING_MEMBER		2 /* MLDv1 we are reporter */
#define MLD_IDLE_MEMBER			3 /* MLDv1 we reported last */
#define MLD_LAZY_MEMBER			4 /* MLDv1 other member reporting */
#define MLD_SLEEPING_MEMBER		5 /* MLDv1 start query response */
#define MLD_AWAKENING_MEMBER		6 /* MLDv1 group timer will start */
#define MLD_G_QUERY_PENDING_MEMBER	7 /* MLDv2 group query pending */
#define MLD_SG_QUERY_PENDING_MEMBER	8 /* MLDv2 source query pending */
#define MLD_LEAVING_MEMBER		9 /* MLDv2 dying gasp (pending last */
					  /* retransmission of INCLUDE {}) */

I first noticed the issue on a FreeBSD 13.1.

[emaste]

sys/netinet6/in6_mcast.c

2866

Maybe == MLD_LEAVING_MEMBER - MLD_NOT_MEMBER + 1?

[sebastien.bini_stormshield.eu]

sys/netinet6/in6_mcast.c

2866

Could do. I am not so sure of the usefulness of this static assertion (won't break if a new entry is added, and won't run unless KTR is on).

At least in your version the boundaries will appear in a text search ¯\_(ツ)_/¯

[sebastien.bini_stormshield.eu]

• Code review comments

[emaste]

sys/netinet6/in6_mcast.c

2866

Indeed - I was thinking about suggesting it just be removed instead. Neither version would catch a new entry at the end of the list, but I think it's still fine to add

[emaste]

Can you confirm Sébastien BINI <[e-mail]> as the git author info?

[sebastien.bini_stormshield.eu]

@emaste Yes, it's good.