From owner-freebsd-doc@FreeBSD.ORG Tue Feb 15 12:30:15 2011 Return-Path: Delivered-To: freebsd-doc@hub.freebsd.org Received: from mx1.freebsd.org (mx1.freebsd.org [IPv6:2001:4f8:fff6::34]) by hub.freebsd.org (Postfix) with ESMTP id 8869F106564A for ; Tue, 15 Feb 2011 12:30:15 +0000 (UTC) (envelope-from gnats@FreeBSD.org) Received: from freefall.freebsd.org (freefall.freebsd.org [IPv6:2001:4f8:fff6::28]) by mx1.freebsd.org (Postfix) with ESMTP id 5BDB68FC0C for ; Tue, 15 Feb 2011 12:30:15 +0000 (UTC) Received: from freefall.freebsd.org (localhost [127.0.0.1]) by freefall.freebsd.org (8.14.4/8.14.4) with ESMTP id p1FCUF7O045228 for ; Tue, 15 Feb 2011 12:30:15 GMT (envelope-from gnats@freefall.freebsd.org) Received: (from gnats@localhost) by freefall.freebsd.org (8.14.4/8.14.4/Submit) id p1FCUFNv045223; Tue, 15 Feb 2011 12:30:15 GMT (envelope-from gnats) Date: Tue, 15 Feb 2011 12:30:15 GMT Message-Id: <201102151230.p1FCUFNv045223@freefall.freebsd.org> To: freebsd-doc@FreeBSD.org From: Nikos Vassiliadis Cc: Subject: Re: docs/150917: icmp.4, wrong description of icmplim and icmplim_output sysctls X-BeenThere: freebsd-doc@freebsd.org X-Mailman-Version: 2.1.5 Precedence: list Reply-To: Nikos Vassiliadis List-Id: Documentation project List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Tue, 15 Feb 2011 12:30:15 -0000 The following reply was made to PR docs/150917; it has been noted by GNATS. From: Nikos Vassiliadis To: Giorgos Keramidas Cc: bug-followup@freebsd.org Subject: Re: docs/150917: icmp.4, wrong description of icmplim and icmplim_output sysctls Date: Tue, 15 Feb 2011 14:27:47 +0200 On 2/15/2011 9:47 AM, Giorgos Keramidas wrote: > On 2010-09-24 13:42, Nikos Vassiliadis wrote: >> The icmp manual page describes icmplim and icmplim_output wrongly. >> >> icmplim (integer) Bandwidth limit for ICMP replies in pack- >> ets/second. Used when icmplim_output is non-zero. >> Defaults to 200. >> >> icmplim_output (boolean) Enable/disable bandwidth limiting of ICMP >> replies. Defaults to true. >> >> The correct description [1] is this: >> >> If you do not want to see messages about this in your log files, but >> you still want the kernel to do response limiting, you can use the >> net.inet.icmp.icmplim_output sysctl variable to disable the output >> like this: >> >> # sysctl -w net.inet.icmp.icmplim_output=0 >> >> Finally, if you want to disable response limiting, you can set the >> net.inet.icmp.icmplim sysctl variable (see above for an example) to >> 0. Disabling response limiting is discouraged for the reasons listed >> above. > >> Index: src/share/man/man4/icmp.4 >> =================================================================== >> --- src/share/man/man4/icmp.4 (revision 213086) >> +++ src/share/man/man4/icmp.4 (working copy) >> @@ -179,15 +179,16 @@ >> the system replies to an ICMP Address Mask Request packet. >> Defaults to 0. >> .It Va icmplim >> -.Pq Vt integer >> -Bandwidth limit for ICMP replies in packets/second. >> -Used when >> -.Va icmplim_output >> -is non-zero. >> +.Pq Vt "unsigned integer" >> +Limit for ICMP or TCP RST responses in packets per second. >> +Response limiting is disabled by setting >> +.Va icmplim >> +to 0. > > I am not sure I like the idea of mentioning all the other protocols that > may send ICMP replies too, so this probably needs a bit of reworking. > For instance, UDP may send ICMP unreachable errors too. SCTP also. > > Are we going to list *all* the possible protocols by name? We don't list the protocols that initiated the response. As you mentioned, TCP|UDP|SCTP can initiate this rate-limiting function. Yet, all possible replies that this code handles are either ICMP or TCP. For example the current code does not handle SCTP ABORT, it *does* handle the case of an ICMP response generated by SCTP. The replies are defined here: http://fxr.watson.org/fxr/source/netinet/icmp_var.h#L99 > >> Defaults to 200. >> .It Va icmplim_output >> .Pq Vt boolean >> -Enable/disable bandwidth limiting of ICMP replies. >> +Enable/disable logging of >> +.Va icmplim . > > "logging of rate-limiting messages from icmplim" ? Yes, that's better. > >> Index: src/sys/netinet/ip_icmp.c >> =================================================================== >> --- src/sys/netinet/ip_icmp.c (revision 213086) >> +++ src/sys/netinet/ip_icmp.c (working copy) >> @@ -106,15 +106,15 @@ >> >> static VNET_DEFINE(int, icmplim) = 200; >> #define V_icmplim VNET(icmplim) >> -SYSCTL_VNET_INT(_net_inet_icmp, ICMPCTL_ICMPLIM, icmplim, CTLFLAG_RW, >> +SYSCTL_VNET_UINT(_net_inet_icmp, ICMPCTL_ICMPLIM, icmplim, CTLFLAG_RW, >> &VNET_NAME(icmplim), 0, >> - "Maximum number of ICMP responses per second"); >> + "Maximum number of ICMP or TCP RST responses per second"); >> >> static VNET_DEFINE(int, icmplim_output) = 1; >> #define V_icmplim_output VNET(icmplim_output) >> -SYSCTL_VNET_INT(_net_inet_icmp, OID_AUTO, icmplim_output, CTLFLAG_RW, >> +SYSCTL_VNET_UINT(_net_inet_icmp, OID_AUTO, icmplim_output, CTLFLAG_RW, >> &VNET_NAME(icmplim_output), 0, >> - "Enable rate limiting of ICMP responses"); >> + "Enable logging of enforced limit on ICMP or TCP RST responses"); > > This also has the problem that it does not mention UDP or SCTP replies > for port-unreachable, host-unreachable, etc. These are ICMP messages generated by UDP|SCTP. But... the sysctl's description doesn't have to be that accurate, if you ask me. I would like the manual page to be as accurate as possible. I find particularly unattractive the term bandwidth there! Thanks for handling this, Nikos