Skip to contentSkip to conversation

iproute2: drop equalize support.

PATCH Applied Andreas Henriksson 3 participants 3 messages Mar 14, 2009 – Mar 27, 2009

Subsystems (2)
lore
Share & tools
lore
#0 lore raw

PATCH 1/1 iproute2: drop equalize support.

Hello Stephen and netdev people!

Currently you can configure "equalize" and it looks all fine and dandy. The kernel has the interface defined, but apparently there's never actually been any implementation for it (only a never merged patch in the 2.4 era).

I'm suggesting to drop the code to give any potential users of this feature the benefit of receiving a proper error message. I see it unlikely that this will be implemented in the near future, but if it ever happens reviving the iproute2 side should be as easy as git revert this patch.

For more details see http://bugs.debian.org/149897

Regards,
Andreas Henriksson

debian/rules +0 -0
index 132455f..4d825ca 100755
doc/ip-cref.tex +0 -7
index bb4eb78..b333842 100644
@@ -1380,13 +1380,6 @@ database.
 even if it does not match any interface prefix. One application of this
 option may be found in~\cite{IP-TUNNELS}.
 
-\item \verb|equalize|
-
---- allow packet by packet randomization on multipath routes.
-Without this modifier, the route will be frozen to one selected
-nexthop, so that load splitting will only occur on per-flow base.
-\verb|equalize| only works if the kernel is patched.
-
 
 \end{itemize}
 
ip/iproute.c +3 -11
index 6a2ea05..529159b 100644
@@ -67,9 +67,9 @@ static void usage(void)
 	fprintf(stderr, "NODE_SPEC := [ TYPE ] PREFIX [ tos TOS ]\n");
 	fprintf(stderr, "             [ table TABLE_ID ] [ proto RTPROTO ]\n");
 	fprintf(stderr, "             [ scope SCOPE ] [ metric METRIC ]\n");
-	fprintf(stderr, "INFO_SPEC := NH OPTIONS FLAGS [ nexthop NH ]...\n");
+	fprintf(stderr, "INFO_SPEC := NH OPTIONS [ nexthop NH ]...\n");
 	fprintf(stderr, "NH := [ via ADDRESS ] [ dev STRING ] [ weight NUMBER ] NHFLAGS\n");
-	fprintf(stderr, "OPTIONS := FLAGS [ mtu NUMBER ] [ advmss NUMBER ]\n");
+	fprintf(stderr, "OPTIONS := [ mtu NUMBER ] [ advmss NUMBER ]\n");
 	fprintf(stderr, "           [ rtt TIME ] [ rttvar TIME ]\n");
 	fprintf(stderr, "           [ window NUMBER] [ cwnd NUMBER ] [ initcwnd NUMBER ]\n");
 	fprintf(stderr, "           [ ssthresh NUMBER ] [ realms REALM ] [ src ADDRESS ]\n");
@@ -78,7 +78,6 @@ static void usage(void)
 	fprintf(stderr, "          unreachable | prohibit | blackhole | nat ]\n");
 	fprintf(stderr, "TABLE_ID := [ local | main | default | all | NUMBER ]\n");
 	fprintf(stderr, "SCOPE := [ host | link | global | NUMBER ]\n");
-	fprintf(stderr, "FLAGS := [ equalize ]\n");
 	fprintf(stderr, "MP_ALGO := { rr | drr | random | wrandom }\n");
 	fprintf(stderr, "NHFLAGS := [ onlink | pervasive ]\n");
 	fprintf(stderr, "RTPROTO := [ kernel | boot | static | NUMBER ]\n");
@@ -387,8 +386,6 @@ int print_route(const struct sockaddr_nl *who, struct nlmsghdr *n, void *arg)
 		fprintf(fp, "onlink ");
 	if (r->rtm_flags & RTNH_F_PERVASIVE)
 		fprintf(fp, "pervasive ");
-	if (r->rtm_flags & RTM_F_EQUALIZE)
-		fprintf(fp, "equalize ");
 	if (r->rtm_flags & RTM_F_NOTIFY)
 		fprintf(fp, "notify ");
 
@@ -428,9 +425,7 @@ int print_route(const struct sockaddr_nl *who, struct nlmsghdr *n, void *arg)
 		PRTFL(FAST, "fastroute");
 		PRTFL(NOTIFY, "notify");
 		PRTFL(TPROXY, "proxy");
-#ifdef RTCF_EQUALIZE
-		PRTFL(EQUALIZE, "equalize");
-#endif
+
 		if (flags)
 			fprintf(fp, "%s%x> ", first ? "<" : "", flags);
 		if (tb[RTA_CACHEINFO]) {
@@ -859,9 +854,6 @@ int iproute_modify(int cmd, unsigned flags, int argc, char **argv)
 			addattr32(&req.n, sizeof(req), RTA_FLOW, realm);
 		} else if (strcmp(*argv, "onlink") == 0) {
 			req.r.rtm_flags |= RTNH_F_ONLINK;
-		} else if (matches(*argv, "equalize") == 0 ||
-			   strcmp(*argv, "eql") == 0) {
-			req.r.rtm_flags |= RTM_F_EQUALIZE;
 		} else if (strcmp(*argv, "nexthop") == 0) {
 			nhs_ok = 1;
 			break;
man/man8/ip.8 +2 -14
index e948d03..e4a57ff 100644
@@ -163,7 +163,7 @@ replace " | " monitor " } "
 .IR METRIC " ]"
 
 .ti -8
-.IR INFO_SPEC " := " "NH OPTIONS FLAGS" " ["
+.IR INFO_SPEC " := " "NH OPTIONS" " ["
 .B  nexthop
 .IR NH " ] ..."
 
@@ -177,7 +177,7 @@ replace " | " monitor " } "
 .IR NUMBER " ] " NHFLAGS
 
 .ti -8
-.IR OPTIONS " := " FLAGS " [ "
+.IR OPTIONS " := [ "
 .B  mtu
 .IR NUMBER " ] [ "
 .B  advmss
@@ -213,10 +213,6 @@ throw " | " unreachable " | " prohibit " | " blackhole " | " nat " ]"
 .IR NUMBER " ]"
 
 .ti -8
-.IR FLAGS " := [ "
-.BR equalize " ]"
-
-.ti -8
 .IR NHFLAGS " := [ "
 .BR onlink " | " pervasive " ]"
 
@@ -1582,14 +1578,6 @@ to assign (or not to assign) protocol tags.
 pretend that the nexthop is directly attached to this link,
 even if it does not match any interface prefix.
 
-.TP
-.B equalize
-allow packet by packet randomization on multipath routes.
-Without this modifier, the route will be frozen to one selected
-nexthop, so that load splitting will only occur on per-flow base.
-.B equalize
-only works if the kernel is patched.
-
 .SS ip route delete - delete route
 
 .B ip route del
#1 lore raw
Quoted 129 lines · On Sat, 14 Mar 2009 17:37:35 +0100 Andreas Henriksson <[email protected]> wrote:

Hello Stephen and netdev people!

Currently you can configure "equalize" and it looks all fine and dandy. The kernel has the interface defined, but apparently there's never actually been any implementation for it (only a never merged patch in the 2.4 era).

I'm suggesting to drop the code to give any potential users of this feature the benefit of receiving a proper error message. I see it unlikely that this will be implemented in the near future, but if it ever happens reviving the iproute2 side should be as easy as git revert this patch.

For more details see http://bugs.debian.org/149897

Regards,
Andreas Henriksson

debian/rules +0 -0
index 132455f..4d825ca 100755
doc/ip-cref.tex +0 -7
index bb4eb78..b333842 100644
@@ -1380,13 +1380,6 @@ database.
 even if it does not match any interface prefix. One application of this
 option may be found in~\cite{IP-TUNNELS}.
 
-\item \verb|equalize|
-
---- allow packet by packet randomization on multipath routes.
-Without this modifier, the route will be frozen to one selected
-nexthop, so that load splitting will only occur on per-flow base.
-\verb|equalize| only works if the kernel is patched.
-
 
 \end{itemize}
 
ip/iproute.c +3 -11
index 6a2ea05..529159b 100644
@@ -67,9 +67,9 @@ static void usage(void)
 	fprintf(stderr, "NODE_SPEC := [ TYPE ] PREFIX [ tos TOS ]\n");
 	fprintf(stderr, "             [ table TABLE_ID ] [ proto RTPROTO ]\n");
 	fprintf(stderr, "             [ scope SCOPE ] [ metric METRIC ]\n");
-	fprintf(stderr, "INFO_SPEC := NH OPTIONS FLAGS [ nexthop NH ]...\n");
+	fprintf(stderr, "INFO_SPEC := NH OPTIONS [ nexthop NH ]...\n");
 	fprintf(stderr, "NH := [ via ADDRESS ] [ dev STRING ] [ weight NUMBER ] NHFLAGS\n");
-	fprintf(stderr, "OPTIONS := FLAGS [ mtu NUMBER ] [ advmss NUMBER ]\n");
+	fprintf(stderr, "OPTIONS := [ mtu NUMBER ] [ advmss NUMBER ]\n");
 	fprintf(stderr, "           [ rtt TIME ] [ rttvar TIME ]\n");
 	fprintf(stderr, "           [ window NUMBER] [ cwnd NUMBER ] [ initcwnd NUMBER ]\n");
 	fprintf(stderr, "           [ ssthresh NUMBER ] [ realms REALM ] [ src ADDRESS ]\n");
@@ -78,7 +78,6 @@ static void usage(void)
 	fprintf(stderr, "          unreachable | prohibit | blackhole | nat ]\n");
 	fprintf(stderr, "TABLE_ID := [ local | main | default | all | NUMBER ]\n");
 	fprintf(stderr, "SCOPE := [ host | link | global | NUMBER ]\n");
-	fprintf(stderr, "FLAGS := [ equalize ]\n");
 	fprintf(stderr, "MP_ALGO := { rr | drr | random | wrandom }\n");
 	fprintf(stderr, "NHFLAGS := [ onlink | pervasive ]\n");
 	fprintf(stderr, "RTPROTO := [ kernel | boot | static | NUMBER ]\n");
@@ -387,8 +386,6 @@ int print_route(const struct sockaddr_nl *who, struct nlmsghdr *n, void *arg)
 		fprintf(fp, "onlink ");
 	if (r->rtm_flags & RTNH_F_PERVASIVE)
 		fprintf(fp, "pervasive ");
-	if (r->rtm_flags & RTM_F_EQUALIZE)
-		fprintf(fp, "equalize ");
 	if (r->rtm_flags & RTM_F_NOTIFY)
 		fprintf(fp, "notify ");
 
@@ -428,9 +425,7 @@ int print_route(const struct sockaddr_nl *who, struct nlmsghdr *n, void *arg)
 		PRTFL(FAST, "fastroute");
 		PRTFL(NOTIFY, "notify");
 		PRTFL(TPROXY, "proxy");
-#ifdef RTCF_EQUALIZE
-		PRTFL(EQUALIZE, "equalize");
-#endif
+
 		if (flags)
 			fprintf(fp, "%s%x> ", first ? "<" : "", flags);
 		if (tb[RTA_CACHEINFO]) {
@@ -859,9 +854,6 @@ int iproute_modify(int cmd, unsigned flags, int argc, char **argv)
 			addattr32(&req.n, sizeof(req), RTA_FLOW, realm);
 		} else if (strcmp(*argv, "onlink") == 0) {
 			req.r.rtm_flags |= RTNH_F_ONLINK;
-		} else if (matches(*argv, "equalize") == 0 ||
-			   strcmp(*argv, "eql") == 0) {
-			req.r.rtm_flags |= RTM_F_EQUALIZE;
 		} else if (strcmp(*argv, "nexthop") == 0) {
 			nhs_ok = 1;
 			break;
man/man8/ip.8 +2 -7
index e948d03..e4a57ff 100644
@@ -163,7 +163,7 @@ replace " | " monitor " } "
 .IR METRIC " ]"
 
 .ti -8
-.IR INFO_SPEC " := " "NH OPTIONS FLAGS" " ["
+.IR INFO_SPEC " := " "NH OPTIONS" " ["
 .B  nexthop
 .IR NH " ] ..."
 
@@ -177,7 +177,7 @@ replace " | " monitor " } "
 .IR NUMBER " ] " NHFLAGS
 
 .ti -8
-.IR OPTIONS " := " FLAGS " [ "
+.IR OPTIONS " := [ "
 .B  mtu
 .IR NUMBER " ] [ "
 .B  advmss
@@ -213,10 +213,6 @@ throw " | " unreachable " | " prohibit " | " blackhole " | " nat " ]"
 .IR NUMBER " ]"
 
 .ti -8
-.IR FLAGS " := [ "
-.BR equalize " ]"
-
-.ti -8
 .IR NHFLAGS " := [ "
 .BR onlink " | " pervasive " ]"
 
@@ -1582,14 +1578,6 @@ to assign (or not to assign) protocol tags.
 pretend that the nexthop is directly attached to this link,
 even if it does not match any interface prefix.
 
-.TP

man/man8/ip.8 to assign (or not to assign) protocol tags. · line 1583

Quoted 10 lines
man/man8/ip.8 +0 -7
-.B equalize
-allow packet by packet randomization on multipath routes.
-Without this modifier, the route will be frozen to one selected
-nexthop, so that load splitting will only occur on per-flow base.
-.B equalize
-only works if the kernel is patched.
-
 .SS ip route delete - delete route
 
 .B ip route del

applied, i think this was part of the aborted effort at more complex multipath support. If it ever returns, we can re-enable it.

#2 lore raw
Quoted 2 lines · Stephen Hemminger

applied, i think this was part of the aborted effort at more complex multipath support. If it ever returns, we can re-enable it.

Its actually a lot older than that, it used to be an external patch in the 2.4 era for per-packet multipath routing. I don't think anyone will miss it :)