Skip site navigation (1)Skip section navigation (2)
Date:      Fri, 30 Nov 2012 22:22:49 -0800 (PST)
From:      Garrett Cooper <yanegomi@gmail.com>
To:        brde@optusnet.com.au
Cc:        freebsd-net@freebsd.org
Subject:   Re: [RFC] Better document net.inet6 sysctls and prune dead sysctls (fwd)
Message-ID:  <alpine.BSF.2.00.1211302222200.57210@toaster.local>
In-Reply-To: <alpine.BSF.2.00.1211302204460.57210@toaster.local>
References:  <alpine.BSF.2.00.1211302204460.57210@toaster.local>

next in thread | previous in thread | raw e-mail | index | archive | help
On Fri, 30 Nov 2012, Garrett Cooper wrote:

> On Tue, 27 Nov 2012, Bruce Evans wrote:
>
>> On Mon, 26 Nov 2012, Garrett Cooper wrote:
>>
>>> 	As noted in a previous thread, I set out to better document the 
>>> net.inet6 sysctls after having to tweak the knobs to get things to work 
>>> for TAHI, and this is the resulting draft (so far). I also took the 
>>> liberty of removing the ip6_rr_prune and icmp6_redirtimeout sysctls 
>>> because they weren't in use anywhere else in the sys/... portion of the 
>>> tree. I was wondering if there are any points that should be 
>>> corrected/clarified before I submit a PR with the resulting patch.
>> 
>> It would be good to fix the style bugs when changing lots.
>> 
>>> Index: sys/netinet6/in6_proto.c
>>> ===================================================================
>>> --- sys/netinet6/in6_proto.c	(revision 242903)
>>> +++ sys/netinet6/in6_proto.c	(working copy)
>>> @@ -443,7 +441,6 @@
>>> 
>>> /* ICMPV6 parameters */
>>> VNET_DEFINE(int, icmp6_rediraccept) = 1;/* accept and process redirects */
>>> -VNET_DEFINE(int, icmp6_redirtimeout) = 10 * 60;	/* 10 minutes */
>>> VNET_DEFINE(int, icmp6errppslim) = 100;		/* 100pps */
>>> /* control how to respond to NI queries */
>>> VNET_DEFINE(int, icmp6_nodeinfo) =
>>> @@ -515,15 +512,20 @@
>>> }
>>> 
>>> SYSCTL_VNET_INT(_net_inet6_ip6, IPV6CTL_FORWARDING, forwarding, 
>>> CTLFLAG_RW,
>>> -	&VNET_NAME(ip6_forwarding), 0, "");
>>> +	&VNET_NAME(ip6_forwarding), 0,
>>> +	"Forward IPv6 packets via node.");
>> 
>> All the sysctl names spell ipv6 without a v.  This is hard to fix.  This
>> bug is missing in the names of the sysctl numbers, but it is a bug to
>> used named numbers instead of OID_AUTO in code newer than OID_AUTO, and
>> ipv6 is much newer.
>> 
>> New style bug in almost every description.  Sysctl descriptions are not
>> normally terminated by a ".".
>> 
>> Old style bug in almost every ipv6 sysctl.  Large SYSCTL declarations are
>> normally normally-indented, with 4-space continuation indents.  This rule
>> is broken farily consistenly in ipv6 sysctl.  More ivp4 SYSCTL descriptions
>> are actually formatted normally.
>> 
>>> @@ -594,35 +613,45 @@
>>> #define	V_ip6_output_flowtable_size VNET(ip6_output_flowtable_size)
>>> 
>>> SYSCTL_VNET_INT(_net_inet6_ip6, OID_AUTO, output_flowtable_size, 
>>> CTLFLAG_RDTUN,
>>> -    &VNET_NAME(ip6_output_flowtable_size), 2048,
>>> -    "number of entries in the per-cpu output flow caches");
>>> +	&VNET_NAME(ip6_output_flowtable_size), 2048,
>>> +	"number of entries in the per-cpu output flow caches");
>>> ...
>> 
>> Here the formatting change is backwards (from normal continuation indent
>> to abnormal ipv6 continuation indent.
>> 
>> This and some other old sysctl descriptions have differnent style bugs:
>> they are normally terminated (not with a '.'), but are not normally
>> capitalized (with capaitals).
>> 
>>> SYSCTL_VNET_INT(_net_inet6_icmp6, ICMPV6CTL_ND6_PRUNE, nd6_prune, 
>>> CTLFLAG_RW,
>> 
>> Some mailer mangled the formatting (the line isn't too long and shouldn't
>> have been mangled).
>> 
>>> -	&VNET_NAME(nd6_prune), 0, "");
>>> +	&VNET_NAME(nd6_prune), 0,
>>> +	"Period (in seconds) for performing nd6 expiration checks of default 
>>> "
>> 
>> More mangling by some mailer.
>> 
>>> +	"routes and prefix lists");
>> 
>> However, the string is too long.  The message is obfuscated by splitting 
>> it.
>> This keeps the line length short in the source code, but it is still too
>> long in the output.  Sysctl descriptions should be no longer than 50 or
>> 60 characters, since the sysctl name will expand the output by 20 or 30
>> characters.  sysctl names longer than 20 or 30 are a larger bug, since
>> they are hard to write as well as hard to read.
>> 
>> This sysctl description is normally terminated (not with a ".").
>> 
>>> ...
>>> SYSCTL_VNET_INT(_net_inet6_icmp6, ICMPV6CTL_ND6_UMAXTRIES, nd6_umaxtries,
>>> -	CTLFLAG_RW, &VNET_NAME(nd6_umaxtries), 0, "");
>>> +	CTLFLAG_RW, &VNET_NAME(nd6_umaxtries), 0,
>>> +	"Maximum number of unicast queries to perform via nd6.");
>> 
>> Another type of sysctl misformatting is generated by names that are so
>> verbose that CTLFLAG* cannot be formatted normally (on the first line).
>> 
>> I missed earlier instances of these bugs.
>> 
>> Statistic on output style in sysctl descriptions: on freefall now:
>> sysctl -da:                    3546 lines
>> sysctl -da | grep -v ': *$':   3192 lines
>>    [this filters out sysctls with null descriptions.  The output is
>>    misformatted with a space before the null description]
>> sysctl -nda:                   3546 lines
>> sysctl -nda | grep -v '^: *$': 3192 lines
>>    [since this 3192 is the same as the number of sysctls with non-null
>>    descreiptions, there must be no multi-line descriptions with embedded
>>    newlines]
>> sysctl -nda | grep -v '^: *$' | grep -v '^[A-Z]':
>>                               1555 lines
>>    [most of non-capitalizations are style bugs.  The last 1395 of them
>>    are automatically generated for device sysctls.  This is part of 5
>>    lines of automatically generated spam for device sysctls (descriptions
>>    are duplicated ad spamium for %desc, %driver, %location, %pnpinfo and
>>    %parent).  A few are correct because they are for a proper name or
>>    keyword or a number]
>> sysctl -nda | grep -v '^: *$' | grep '\.$':
>>                                 74 lines
>>    [the style bug of terminating with a '.' is uncommon.  In a few cases
>>    it goes with the larger style bug of a multi-line description]
>> sysctl -nda | grep -v '^: *$' | grep '.... [81 dots]':
>>                                 11 lines
>>    [these lines are too long even without the sysctl names.  Abnormal
>>    termination is denser than usual in these lines (8 of 11)]
>> sysctl -nda | grep -v '^: *$' | grep '\..*\.':
>>                                  8 lines
>>    [2 of these are for multi-sentence descriptions].
>
> Bruce,
> 	Is this better? I tried to be consistent about using v6 properly when 
> dealing with protocols in order to match what the IETF did and I believe I 
> properly integrated in your comments.
> Thanks!
> -Garrett
>
> Index: sys/netinet6/in6_proto.c
> ===================================================================
> --- sys/netinet6/in6_proto.c	(revision 243557)
> +++ sys/netinet6/in6_proto.c	(working copy)
> @@ -131,7 +131,7 @@
> #endif
>
> /*
> - * TCP/IP protocol family: IP6, ICMP6, UDP, TCP.
> + * TCP/IP protocol family: IPv6, ICMPv6, UDPv6, TCPv6.
>  */
> FEATURE(inet6, "Internet Protocol version 6");
>
> @@ -382,9 +382,9 @@
>  */
> #ifndef	IPV6FORWARDING
> #ifdef GATEWAY6
> -#define	IPV6FORWARDING	1	/* forward IP6 packets not for us */
> +#define	IPV6FORWARDING	1	/* forward IPv6 packets */
> #else
> -#define	IPV6FORWARDING	0	/* don't forward IP6 packets not for 
> us */
> +#define	IPV6FORWARDING	0	/* don't forward IPv6 packets */
> #endif /* GATEWAY6 */
> #endif /* !IPV6FORWARDING */
>
> @@ -409,8 +409,6 @@
> VNET_DEFINE(int, ip6_auto_flowlabel) = 1;
> VNET_DEFINE(int, ip6_use_deprecated) = 1;/* allow deprecated addr
> 					 * (RFC2462 5.5.4) */
> -VNET_DEFINE(int, ip6_rr_prune) = 5;	/* router renumbering prefix
> -					 * walk list every 5 sec. */
> VNET_DEFINE(int, ip6_mcast_pmtu) = 0;	/* enable pMTU discovery for 
> multicast? */
> VNET_DEFINE(int, ip6_v6only) = 1;
>
> @@ -431,7 +429,7 @@
> VNET_DEFINE(int, pmtu_expire) = 60*10;
> VNET_DEFINE(int, pmtu_probe) = 60*2;
>
> -/* raw IP6 parameters */
> +/* raw IPV6 parameters */
> /*
>  * Nominal space allocated to a raw ip socket.
>  */
> @@ -443,13 +441,12 @@
>
> /* ICMPV6 parameters */
> VNET_DEFINE(int, icmp6_rediraccept) = 1;/* accept and process redirects */
> -VNET_DEFINE(int, icmp6_redirtimeout) = 10 * 60;	/* 10 minutes */
> VNET_DEFINE(int, icmp6errppslim) = 100;		/* 100pps */
> /* control how to respond to NI queries */
> VNET_DEFINE(int, icmp6_nodeinfo) =
>     (ICMP6_NODEINFO_FQDNOK|ICMP6_NODEINFO_NODEADDROK);
>
> -/* UDP on IP6 parameters */
> +/* UDP on IPv6 parameters */
> VNET_DEFINE(int, udp6_sendspace) = 9216;/* really max datagram size */
> VNET_DEFINE(int, udp6_recvspace) = 40 * (1024 + sizeof(struct 
> sockaddr_in6));
> 					/* 40 1K datagrams */
> @@ -461,12 +458,12 @@
> 	"Internet6 Family");
>
> /* net.inet6 */
> -SYSCTL_NODE(_net_inet6,	IPPROTO_IPV6,	ip6,	CTLFLAG_RW, 0, 
> "IP6");
> -SYSCTL_NODE(_net_inet6,	IPPROTO_ICMPV6,	icmp6,	CTLFLAG_RW, 0, 
> "ICMP6");
> -SYSCTL_NODE(_net_inet6,	IPPROTO_UDP,	udp6,	CTLFLAG_RW, 0, 
> "UDP6");
> -SYSCTL_NODE(_net_inet6,	IPPROTO_TCP,	tcp6,	CTLFLAG_RW, 0, 
> "TCP6");
> +SYSCTL_NODE(_net_inet6,	IPPROTO_IPV6,	ip6,	CTLFLAG_RW, 0, 
> "IPv6");
> +SYSCTL_NODE(_net_inet6,	IPPROTO_ICMPV6,	icmp6,	CTLFLAG_RW, 0, 
> "ICMPv6");
> +SYSCTL_NODE(_net_inet6,	IPPROTO_UDP,	udp6,	CTLFLAG_RW, 0, 
> "UDPv6");
> +SYSCTL_NODE(_net_inet6,	IPPROTO_TCP,	tcp6,	CTLFLAG_RW, 0, 
> "TCPv6");
> #ifdef SCTP
> -SYSCTL_NODE(_net_inet6,	IPPROTO_SCTP,	sctp6,	CTLFLAG_RW, 0, 
> "SCTP6");
> +SYSCTL_NODE(_net_inet6,	IPPROTO_SCTP,	sctp6,	CTLFLAG_RW, 0, 
> "SCTPv6");
> #endif
> #ifdef IPSEC
> SYSCTL_NODE(_net_inet6,	IPPROTO_ESP,	ipsec6,	CTLFLAG_RW, 0, 
> "IPSEC6");
> @@ -515,77 +512,99 @@
> }
>
> SYSCTL_VNET_INT(_net_inet6_ip6, IPV6CTL_FORWARDING, forwarding, CTLFLAG_RW,
> -	&VNET_NAME(ip6_forwarding), 0, "");
> +    &VNET_NAME(ip6_forwarding), 0,
> +    "Forward IPv6 packets via node");
> SYSCTL_VNET_INT(_net_inet6_ip6, IPV6CTL_SENDREDIRECTS, redirect, CTLFLAG_RW,
> -	&VNET_NAME(ip6_sendredirects), 0, "");
> +    &VNET_NAME(ip6_sendredirects), 0,
> +    "Redirect IPv6 packets via node");
> SYSCTL_VNET_INT(_net_inet6_ip6, IPV6CTL_DEFHLIM, hlim, CTLFLAG_RW,
> -	&VNET_NAME(ip6_defhlim), 0, "");
> +    &VNET_NAME(ip6_defhlim), 0,
> +    "Default hop limit for IPv6 unicast packets");
> SYSCTL_VNET_STRUCT(_net_inet6_ip6, IPV6CTL_STATS, stats, CTLFLAG_RW,
> -	&VNET_NAME(ip6stat), ip6stat, "");
> +    &VNET_NAME(ip6stat), ip6stat,
> +    "IPv6 statistics");
> SYSCTL_VNET_INT(_net_inet6_ip6, IPV6CTL_MAXFRAGPACKETS, maxfragpackets,
> -	CTLFLAG_RW, &VNET_NAME(ip6_maxfragpackets), 0, "");
> +    CTLFLAG_RW, &VNET_NAME(ip6_maxfragpackets), 0,
> +    "Maximum number of fragmented packets to process");
> SYSCTL_VNET_INT(_net_inet6_ip6, IPV6CTL_ACCEPT_RTADV, accept_rtadv,
> -	CTLFLAG_RW, &VNET_NAME(ip6_accept_rtadv), 0,
> -	"Default value of per-interface flag for accepting ICMPv6 Router"
> -	"Advertisement messages");
> +    CTLFLAG_RW, &VNET_NAME(ip6_accept_rtadv), 0,
> +    "Default value of per-interface flag for accepting ICMPv6 Router"
> +    "Advertisement messages");
> SYSCTL_VNET_INT(_net_inet6_ip6, IPV6CTL_NO_RADR, no_radr,
> -	CTLFLAG_RW, &VNET_NAME(ip6_no_radr), 0,
> -	"Default value of per-interface flag to control whether routers "
> -	"sending ICMPv6 RA messages on that interface are added into the "
> -	"default router list.");
> +    CTLFLAG_RW, &VNET_NAME(ip6_no_radr), 0,
> +    "Default value of per-interface flag to control whether routers "
> +    "sending ICMPv6 RA messages on that interface are added into the "
> +    "default router list");
> SYSCTL_VNET_INT(_net_inet6_ip6, IPV6CTL_NORBIT_RAIF, norbit_raif, 
> CTLFLAG_RW,
> -	&VNET_NAME(ip6_norbit_raif), 0,
> -	"Always set 0 to R flag in ICMPv6 NA messages when accepting RA"
> -	" on the interface.");
> +    &VNET_NAME(ip6_norbit_raif), 0,
> +    "Always set 0 to R flag in ICMPv6 NA messages when accepting RA "
> +    "on the interface");
> SYSCTL_VNET_INT(_net_inet6_ip6, IPV6CTL_RFC6204W3, rfc6204w3,
> -	CTLFLAG_RW, &VNET_NAME(ip6_rfc6204w3), 0,
> -	"Accept the default router list from ICMPv6 RA messages even "
> -	"when packet forwarding enabled.");
> +    CTLFLAG_RW, &VNET_NAME(ip6_rfc6204w3), 0,
> +    "Accept the default router list from ICMPv6 RA messages even when packet 
> "
> +    "forwarding enabled");
> SYSCTL_VNET_INT(_net_inet6_ip6, IPV6CTL_KEEPFAITH, keepfaith, CTLFLAG_RW,
> -	&VNET_NAME(ip6_keepfaith), 0, "");
> +    &VNET_NAME(ip6_keepfaith), 0,
> +    "");
> SYSCTL_VNET_INT(_net_inet6_ip6, IPV6CTL_LOG_INTERVAL, log_interval,
> -	CTLFLAG_RW, &VNET_NAME(ip6_log_interval), 0, "");
> +    CTLFLAG_RW, &VNET_NAME(ip6_log_interval), 0,
> +    "Period (in secs) for throttling logging for high-traffic operations");
> SYSCTL_VNET_INT(_net_inet6_ip6, IPV6CTL_HDRNESTLIMIT, hdrnestlimit,
> -	CTLFLAG_RW, &VNET_NAME(ip6_hdrnestlimit), 0, "");
> +    CTLFLAG_RW, &VNET_NAME(ip6_hdrnestlimit), 0,
> +    "Maximum number of nested header options to process");
> SYSCTL_VNET_INT(_net_inet6_ip6, IPV6CTL_DAD_COUNT, dad_count, CTLFLAG_RW,
> -	&VNET_NAME(ip6_dad_count), 0, "");
> +    &VNET_NAME(ip6_dad_count), 0,
> +    "Number of Duplicate Address Detection attempts to try before "
> +    "disabling interface.  Setting the value to 0 disables DAD");
> SYSCTL_VNET_INT(_net_inet6_ip6, IPV6CTL_AUTO_FLOWLABEL, auto_flowlabel,
> -	CTLFLAG_RW, &VNET_NAME(ip6_auto_flowlabel), 0, "");
> +    CTLFLAG_RW, &VNET_NAME(ip6_auto_flowlabel), 0,
> +    "Automatically attach a flowlabel to IPv6 packets");
> SYSCTL_VNET_INT(_net_inet6_ip6, IPV6CTL_DEFMCASTHLIM, defmcasthlim,
> -	CTLFLAG_RW, &VNET_NAME(ip6_defmcasthlim), 0, "");
> +    CTLFLAG_RW, &VNET_NAME(ip6_defmcasthlim), 0,
> +    "Default hop limit for ip6 multicast packets");
> SYSCTL_STRING(_net_inet6_ip6, IPV6CTL_KAME_VERSION, kame_version,
> -	CTLFLAG_RD, __KAME_VERSION, 0, "");
> +    CTLFLAG_RD, __KAME_VERSION, 0,
> +    "");
> SYSCTL_VNET_INT(_net_inet6_ip6, IPV6CTL_USE_DEPRECATED, use_deprecated,
> -	CTLFLAG_RW, &VNET_NAME(ip6_use_deprecated), 0, "");
> -SYSCTL_VNET_INT(_net_inet6_ip6, IPV6CTL_RR_PRUNE, rr_prune, CTLFLAG_RW,
> -	&VNET_NAME(ip6_rr_prune), 0, "");
> +    CTLFLAG_RW, &VNET_NAME(ip6_use_deprecated), 0,
> +    "Use deprecated IPv6 support");
> SYSCTL_VNET_INT(_net_inet6_ip6, IPV6CTL_USETEMPADDR, use_tempaddr,
> -	CTLFLAG_RW, &VNET_NAME(ip6_use_tempaddr), 0, "");
> +    CTLFLAG_RW, &VNET_NAME(ip6_use_tempaddr), 0,
> +    "Use a temporary address when establishing address via SLAAC");
> SYSCTL_VNET_PROC(_net_inet6_ip6, IPV6CTL_TEMPPLTIME, temppltime,
> -	CTLTYPE_INT|CTLFLAG_RW, &VNET_NAME(ip6_temp_preferred_lifetime), 0,
> -   	sysctl_ip6_temppltime, "I", "");
> +    CTLTYPE_INT|CTLFLAG_RW, &VNET_NAME(ip6_temp_preferred_lifetime), 0,
> +    sysctl_ip6_temppltime, "I",
> +    "Preferred lifetime (in secs) for a temporary address");
> SYSCTL_VNET_PROC(_net_inet6_ip6, IPV6CTL_TEMPVLTIME, tempvltime,
> -	CTLTYPE_INT|CTLFLAG_RW, &VNET_NAME(ip6_temp_valid_lifetime), 0,
> -   	sysctl_ip6_tempvltime, "I", "");
> +    CTLTYPE_INT|CTLFLAG_RW, &VNET_NAME(ip6_temp_valid_lifetime), 0,
> +    sysctl_ip6_tempvltime, "I",
> +    "Valid lifetime (in secs) for a temporary address");
> SYSCTL_VNET_INT(_net_inet6_ip6, IPV6CTL_V6ONLY, v6only,	CTLFLAG_RW,
> -	&VNET_NAME(ip6_v6only), 0, "");
> +    &VNET_NAME(ip6_v6only), 0,
> +    "Allow IPv4-mapped ip6 addresses per RFC 3493");
> SYSCTL_VNET_INT(_net_inet6_ip6, IPV6CTL_AUTO_LINKLOCAL, auto_linklocal,
> -	CTLFLAG_RW, &VNET_NAME(ip6_auto_linklocal), 0,
> -	"Default value of per-interface flag for automatically adding an 
> IPv6"
> -	" link-local address to interfaces when attached");
> -SYSCTL_VNET_STRUCT(_net_inet6_ip6, IPV6CTL_RIP6STATS, rip6stats, CTLFLAG_RW,
> -	&VNET_NAME(rip6stat), rip6stat, "");
> +    CTLFLAG_RW, &VNET_NAME(ip6_auto_linklocal), 0,
> +    "Default value of per-interface flag for automatically adding an ip6 "
> +    "link-local address to interfaces when attached");
> +SYSCTL_VNET_STRUCT(_net_inet6_ip6, IPV6CTL_RIPV6STATS, rip6stats, 
> CTLFLAG_RW,
> +    &VNET_NAME(rip6stat), rip6stat,
> +    "RIP6 statistics");
> SYSCTL_VNET_INT(_net_inet6_ip6, IPV6CTL_PREFER_TEMPADDR, prefer_tempaddr,
> -	CTLFLAG_RW, &VNET_NAME(ip6_prefer_tempaddr), 0, "");
> +    CTLFLAG_RW, &VNET_NAME(ip6_prefer_tempaddr), 0,
> +    "Prefer the temporary address assigned when performing NUD");
> SYSCTL_VNET_INT(_net_inet6_ip6, IPV6CTL_USE_DEFAULTZONE, use_defaultzone,
> -	CTLFLAG_RW, &VNET_NAME(ip6_use_defzone), 0,"");
> +    CTLFLAG_RW, &VNET_NAME(ip6_use_defzone), 0,
> +    "Use the default scope zone if not explicitly provided");
> SYSCTL_VNET_INT(_net_inet6_ip6, IPV6CTL_MAXFRAGS, maxfrags, CTLFLAG_RW,
> -	&VNET_NAME(ip6_maxfrags), 0, "");
> +    &VNET_NAME(ip6_maxfrags), 0,
> +    "Maximum number of fragments to hold in the reassembly queue");
> SYSCTL_VNET_INT(_net_inet6_ip6, IPV6CTL_MCAST_PMTU, mcast_pmtu, CTLFLAG_RW,
> -	&VNET_NAME(ip6_mcast_pmtu), 0, "");
> +    &VNET_NAME(ip6_mcast_pmtu), 0,
> +    "Enable multicast pMTU discovery");
> #ifdef IPSTEALTH
> SYSCTL_VNET_INT(_net_inet6_ip6, IPV6CTL_STEALTH, stealth, CTLFLAG_RW,
> -	&VNET_NAME(ip6stealth), 0, "");
> +    &VNET_NAME(ip6stealth), 0,
> +    "IPv6 stealth mode. No hop limit decrementation on forwarding");
> #endif
>
> #ifdef FLOWTABLE
> @@ -600,29 +619,39 @@
>
> /* net.inet6.icmp6 */
> SYSCTL_VNET_INT(_net_inet6_icmp6, ICMPV6CTL_REDIRACCEPT, rediraccept,
> -	CTLFLAG_RW, &VNET_NAME(icmp6_rediraccept), 0, "");
> -SYSCTL_VNET_INT(_net_inet6_icmp6, ICMPV6CTL_REDIRTIMEOUT, redirtimeout,
> -	CTLFLAG_RW, &VNET_NAME(icmp6_redirtimeout), 0, "");
> +    CTLFLAG_RW, &VNET_NAME(icmp6_rediraccept), 0,
> +    "Accept and process router redirects");
> SYSCTL_VNET_STRUCT(_net_inet6_icmp6, ICMPV6CTL_STATS, stats, CTLFLAG_RW,
> -	&VNET_NAME(icmp6stat), icmp6stat, "");
> +    &VNET_NAME(icmp6stat), icmp6stat,
> +    "ICMPv6 statistics");
> SYSCTL_VNET_INT(_net_inet6_icmp6, ICMPV6CTL_ND6_PRUNE, nd6_prune, 
> CTLFLAG_RW,
> -	&VNET_NAME(nd6_prune), 0, "");
> +    &VNET_NAME(nd6_prune), 0,
> +    "Period (in secs) for checking for stale router entries via nd6");
> SYSCTL_VNET_INT(_net_inet6_icmp6, ICMPV6CTL_ND6_DELAY, nd6_delay, 
> CTLFLAG_RW,
> -	&VNET_NAME(nd6_delay), 0, "");
> +    &VNET_NAME(nd6_delay), 0,
> +    "Delay (in secs) before performing NUD on stale nodes");
> SYSCTL_VNET_INT(_net_inet6_icmp6, ICMPV6CTL_ND6_UMAXTRIES, nd6_umaxtries,
> -	CTLFLAG_RW, &VNET_NAME(nd6_umaxtries), 0, "");
> +    CTLFLAG_RW, &VNET_NAME(nd6_umaxtries), 0,
> +    "Maximum number of unicast queries to perform via nd6");
> SYSCTL_VNET_INT(_net_inet6_icmp6, ICMPV6CTL_ND6_MMAXTRIES, nd6_mmaxtries,
> -	CTLFLAG_RW, &VNET_NAME(nd6_mmaxtries), 0, "");
> +    CTLFLAG_RW, &VNET_NAME(nd6_mmaxtries), 0,
> +    "Maximum number of multicast queries to perform via nd6");
> SYSCTL_VNET_INT(_net_inet6_icmp6, ICMPV6CTL_ND6_USELOOPBACK, 
> nd6_useloopback,
> -	CTLFLAG_RW, &VNET_NAME(nd6_useloopback), 0, "");
> +    CTLFLAG_RW, &VNET_NAME(nd6_useloopback), 0,
> +    "Use loopback interface for local traffic");
> SYSCTL_VNET_INT(_net_inet6_icmp6, ICMPV6CTL_NODEINFO, nodeinfo, CTLFLAG_RW,
> -	&VNET_NAME(icmp6_nodeinfo), 0, "");
> +    &VNET_NAME(icmp6_nodeinfo), 0,
> +    "Node information state");
> SYSCTL_VNET_INT(_net_inet6_icmp6, ICMPV6CTL_ERRPPSLIMIT, errppslimit,
> -	CTLFLAG_RW, &VNET_NAME(icmp6errppslim), 0, "");
> +    CTLFLAG_RW, &VNET_NAME(icmp6errppslim), 0,
> +    "Packets per second limit for ICMPv6 queries");
> SYSCTL_VNET_INT(_net_inet6_icmp6, ICMPV6CTL_ND6_MAXNUDHINT, nd6_maxnudhint,
> -	CTLFLAG_RW, &VNET_NAME(nd6_maxnudhint), 0, "");
> +    CTLFLAG_RW, &VNET_NAME(nd6_maxnudhint), 0,
> +    "Maximum number of upper layer hints");
> SYSCTL_VNET_INT(_net_inet6_icmp6, ICMPV6CTL_ND6_DEBUG, nd6_debug, 
> CTLFLAG_RW,
> -	&VNET_NAME(nd6_debug), 0, "");
> +    &VNET_NAME(nd6_debug), 0,
> +    "");
> SYSCTL_VNET_INT(_net_inet6_icmp6, ICMPV6CTL_ND6_ONLINKNSRFC4861,
> -	nd6_onlink_ns_rfc4861, CTLFLAG_RW, &VNET_NAME(nd6_onlink_ns_rfc4861),
> -	0, "Accept 'on-link' nd6 NS in compliance with RFC 4861.");
> +    nd6_onlink_ns_rfc4861,
> +    CTLFLAG_RW, &VNET_NAME(nd6_onlink_ns_rfc4861), 0,
> +    "Accept 'on-link' nd6 NS in compliance with RFC 4861");

Wrong To: address >_>.
-Garrett



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?alpine.BSF.2.00.1211302222200.57210>