| Message ID | 20200623164232.175846-2-brianvv@google.com |
|---|---|
| State | Accepted |
| Delegated to: | David Miller |
| Headers | show |
| Series | [v2,net-next,1/2] indirect_call_wrapper: extend indirect wrapper to support up to 4 calls | expand |
On Tue, 2020-06-23 at 09:42 -0700, Brian Vazquez wrote: > It was reported that a considerable amount of cycles were spent on the > expensive indirect calls on fib6_rule_lookup. This patch introduces an > inline helper called pol_route_func that uses the indirect_call_wrappers > to avoid the indirect calls. > > This patch saves around 50ns per call. > > Performance was measured on the receiver by checking the amount of > syncookies that server was able to generate under a synflood load. > > Traffic was generated using trafgen[1] which was pushing around 1Mpps on > a single queue. Receiver was using only one rx queue which help to > create a bottle neck and make the experiment rx-bounded. > > These are the syncookies generated over 10s from the different runs: > > Whithout the patch: > TcpExtSyncookiesSent 3553749 0.0 > TcpExtSyncookiesSent 3550895 0.0 > TcpExtSyncookiesSent 3553845 0.0 > TcpExtSyncookiesSent 3541050 0.0 > TcpExtSyncookiesSent 3539921 0.0 > TcpExtSyncookiesSent 3557659 0.0 > TcpExtSyncookiesSent 3526812 0.0 > TcpExtSyncookiesSent 3536121 0.0 > TcpExtSyncookiesSent 3529963 0.0 > TcpExtSyncookiesSent 3536319 0.0 > > With the patch: > TcpExtSyncookiesSent 3611786 0.0 > TcpExtSyncookiesSent 3596682 0.0 > TcpExtSyncookiesSent 3606878 0.0 > TcpExtSyncookiesSent 3599564 0.0 > TcpExtSyncookiesSent 3601304 0.0 > TcpExtSyncookiesSent 3609249 0.0 > TcpExtSyncookiesSent 3617437 0.0 > TcpExtSyncookiesSent 3608765 0.0 > TcpExtSyncookiesSent 3620205 0.0 > TcpExtSyncookiesSent 3601895 0.0 > > Without the patch the average is 354263 pkt/s or 2822 ns/pkt and with > the patch the average is 360738 pkt/s or 2772 ns/pkt which gives an > estimate of 50 ns per packet. > > [1] http://netsniff-ng.org/ > > Changelog since v1: > - Change ordering in the ICW (Paolo Abeni) > > Cc: Luigi Rizzo <lrizzo@google.com> > Cc: Paolo Abeni <pabeni@redhat.com> > Reported-by: Eric Dumazet <edumazet@google.com> > Signed-off-by: Brian Vazquez <brianvv@google.com> > --- > include/net/ip6_fib.h | 36 ++++++++++++++++++++++++++++++++++++ > net/ipv6/fib6_rules.c | 9 ++++++--- > net/ipv6/ip6_fib.c | 3 ++- > net/ipv6/route.c | 8 ++++---- > 4 files changed, 48 insertions(+), 8 deletions(-) > > diff --git a/include/net/ip6_fib.h b/include/net/ip6_fib.h > index 3f615a29766e..cc8356fd927f 100644 > --- a/include/net/ip6_fib.h > +++ b/include/net/ip6_fib.h > @@ -19,6 +19,7 @@ > #include <net/netlink.h> > #include <net/inetpeer.h> > #include <net/fib_notifier.h> > +#include <linux/indirect_call_wrapper.h> > > #ifdef CONFIG_IPV6_MULTIPLE_TABLES > #define FIB6_TABLE_HASHSZ 256 > @@ -552,6 +553,41 @@ struct bpf_iter__ipv6_route { > }; > #endif > > +INDIRECT_CALLABLE_DECLARE(struct rt6_info *ip6_pol_route_output(struct net *net, > + struct fib6_table *table, > + struct flowi6 *fl6, > + const struct sk_buff *skb, > + int flags)); > +INDIRECT_CALLABLE_DECLARE(struct rt6_info *ip6_pol_route_input(struct net *net, > + struct fib6_table *table, > + struct flowi6 *fl6, > + const struct sk_buff *skb, > + int flags)); > +INDIRECT_CALLABLE_DECLARE(struct rt6_info *__ip6_route_redirect(struct net *net, > + struct fib6_table *table, > + struct flowi6 *fl6, > + const struct sk_buff *skb, > + int flags)); > +INDIRECT_CALLABLE_DECLARE(struct rt6_info *ip6_pol_route_lookup(struct net *net, > + struct fib6_table *table, > + struct flowi6 *fl6, > + const struct sk_buff *skb, > + int flags)); > +static inline struct rt6_info *pol_lookup_func(pol_lookup_t lookup, > + struct net *net, > + struct fib6_table *table, > + struct flowi6 *fl6, > + const struct sk_buff *skb, > + int flags) > +{ > + return INDIRECT_CALL_4(lookup, > + ip6_pol_route_output, > + ip6_pol_route_input, > + ip6_pol_route_lookup, > + __ip6_route_redirect, > + net, table, fl6, skb, flags); > +} > + > #ifdef CONFIG_IPV6_MULTIPLE_TABLES > static inline bool fib6_has_custom_rules(const struct net *net) > { > diff --git a/net/ipv6/fib6_rules.c b/net/ipv6/fib6_rules.c > index fafe556d21e0..6053ef851555 100644 > --- a/net/ipv6/fib6_rules.c > +++ b/net/ipv6/fib6_rules.c > @@ -111,11 +111,13 @@ struct dst_entry *fib6_rule_lookup(struct net *net, struct flowi6 *fl6, > } else { > struct rt6_info *rt; > > - rt = lookup(net, net->ipv6.fib6_local_tbl, fl6, skb, flags); > + rt = pol_lookup_func(lookup, > + net, net->ipv6.fib6_local_tbl, fl6, skb, flags); > if (rt != net->ipv6.ip6_null_entry && rt->dst.error != -EAGAIN) > return &rt->dst; > ip6_rt_put_flags(rt, flags); > - rt = lookup(net, net->ipv6.fib6_main_tbl, fl6, skb, flags); > + rt = pol_lookup_func(lookup, > + net, net->ipv6.fib6_main_tbl, fl6, skb, flags); > if (rt->dst.error != -EAGAIN) > return &rt->dst; > ip6_rt_put_flags(rt, flags); > @@ -226,7 +228,8 @@ static int __fib6_rule_action(struct fib_rule *rule, struct flowi *flp, > goto out; > } > > - rt = lookup(net, table, flp6, arg->lookup_data, flags); > + rt = pol_lookup_func(lookup, > + net, table, flp6, arg->lookup_data, flags); > if (rt != net->ipv6.ip6_null_entry) { > err = fib6_rule_saddr(net, rule, flags, flp6, > ip6_dst_idev(&rt->dst)->dev); > diff --git a/net/ipv6/ip6_fib.c b/net/ipv6/ip6_fib.c > index 49ee89bbcba0..25a90f3f705c 100644 > --- a/net/ipv6/ip6_fib.c > +++ b/net/ipv6/ip6_fib.c > @@ -314,7 +314,8 @@ struct dst_entry *fib6_rule_lookup(struct net *net, struct flowi6 *fl6, > { > struct rt6_info *rt; > > - rt = lookup(net, net->ipv6.fib6_main_tbl, fl6, skb, flags); > + rt = pol_lookup_func(lookup, > + net, net->ipv6.fib6_main_tbl, fl6, skb, flags); > if (rt->dst.error == -EAGAIN) { > ip6_rt_put_flags(rt, flags); > rt = net->ipv6.ip6_null_entry; > diff --git a/net/ipv6/route.c b/net/ipv6/route.c > index 82cbb46a2a4f..5852039ca9cf 100644 > --- a/net/ipv6/route.c > +++ b/net/ipv6/route.c > @@ -1207,7 +1207,7 @@ static struct rt6_info *ip6_create_rt_rcu(const struct fib6_result *res) > return nrt; > } > > -static struct rt6_info *ip6_pol_route_lookup(struct net *net, > +INDIRECT_CALLABLE_SCOPE struct rt6_info *ip6_pol_route_lookup(struct net *net, > struct fib6_table *table, > struct flowi6 *fl6, > const struct sk_buff *skb, > @@ -2274,7 +2274,7 @@ struct rt6_info *ip6_pol_route(struct net *net, struct fib6_table *table, > } > EXPORT_SYMBOL_GPL(ip6_pol_route); > > -static struct rt6_info *ip6_pol_route_input(struct net *net, > +INDIRECT_CALLABLE_SCOPE struct rt6_info *ip6_pol_route_input(struct net *net, > struct fib6_table *table, > struct flowi6 *fl6, > const struct sk_buff *skb, > @@ -2465,7 +2465,7 @@ void ip6_route_input(struct sk_buff *skb) > &fl6, skb, flags)); > } > > -static struct rt6_info *ip6_pol_route_output(struct net *net, > +INDIRECT_CALLABLE_SCOPE struct rt6_info *ip6_pol_route_output(struct net *net, > struct fib6_table *table, > struct flowi6 *fl6, > const struct sk_buff *skb, > @@ -2912,7 +2912,7 @@ struct ip6rd_flowi { > struct in6_addr gateway; > }; > > -static struct rt6_info *__ip6_route_redirect(struct net *net, > +INDIRECT_CALLABLE_SCOPE struct rt6_info *__ip6_route_redirect(struct net *net, > struct fib6_table *table, > struct flowi6 *fl6, > const struct sk_buff *skb, Acked-by: Paolo Abeni <pabeni@redhat.com>
From: Brian Vazquez <brianvv@google.com> Date: Tue, 23 Jun 2020 09:42:32 -0700 > It was reported that a considerable amount of cycles were spent on the > expensive indirect calls on fib6_rule_lookup. This patch introduces an > inline helper called pol_route_func that uses the indirect_call_wrappers > to avoid the indirect calls. > > This patch saves around 50ns per call. > > Performance was measured on the receiver by checking the amount of > syncookies that server was able to generate under a synflood load. > > Traffic was generated using trafgen[1] which was pushing around 1Mpps on > a single queue. Receiver was using only one rx queue which help to > create a bottle neck and make the experiment rx-bounded. > > These are the syncookies generated over 10s from the different runs: > > Whithout the patch: ... > With the patch: ... > Without the patch the average is 354263 pkt/s or 2822 ns/pkt and with > the patch the average is 360738 pkt/s or 2772 ns/pkt which gives an > estimate of 50 ns per packet. > > [1] http://netsniff-ng.org/ > > Changelog since v1: > - Change ordering in the ICW (Paolo Abeni) > > Cc: Luigi Rizzo <lrizzo@google.com> > Cc: Paolo Abeni <pabeni@redhat.com> > Reported-by: Eric Dumazet <edumazet@google.com> > Signed-off-by: Brian Vazquez <brianvv@google.com> Applied.
diff --git a/include/net/ip6_fib.h b/include/net/ip6_fib.h index 3f615a29766e..cc8356fd927f 100644 --- a/include/net/ip6_fib.h +++ b/include/net/ip6_fib.h @@ -19,6 +19,7 @@ #include <net/netlink.h> #include <net/inetpeer.h> #include <net/fib_notifier.h> +#include <linux/indirect_call_wrapper.h> #ifdef CONFIG_IPV6_MULTIPLE_TABLES #define FIB6_TABLE_HASHSZ 256 @@ -552,6 +553,41 @@ struct bpf_iter__ipv6_route { }; #endif +INDIRECT_CALLABLE_DECLARE(struct rt6_info *ip6_pol_route_output(struct net *net, + struct fib6_table *table, + struct flowi6 *fl6, + const struct sk_buff *skb, + int flags)); +INDIRECT_CALLABLE_DECLARE(struct rt6_info *ip6_pol_route_input(struct net *net, + struct fib6_table *table, + struct flowi6 *fl6, + const struct sk_buff *skb, + int flags)); +INDIRECT_CALLABLE_DECLARE(struct rt6_info *__ip6_route_redirect(struct net *net, + struct fib6_table *table, + struct flowi6 *fl6, + const struct sk_buff *skb, + int flags)); +INDIRECT_CALLABLE_DECLARE(struct rt6_info *ip6_pol_route_lookup(struct net *net, + struct fib6_table *table, + struct flowi6 *fl6, + const struct sk_buff *skb, + int flags)); +static inline struct rt6_info *pol_lookup_func(pol_lookup_t lookup, + struct net *net, + struct fib6_table *table, + struct flowi6 *fl6, + const struct sk_buff *skb, + int flags) +{ + return INDIRECT_CALL_4(lookup, + ip6_pol_route_output, + ip6_pol_route_input, + ip6_pol_route_lookup, + __ip6_route_redirect, + net, table, fl6, skb, flags); +} + #ifdef CONFIG_IPV6_MULTIPLE_TABLES static inline bool fib6_has_custom_rules(const struct net *net) { diff --git a/net/ipv6/fib6_rules.c b/net/ipv6/fib6_rules.c index fafe556d21e0..6053ef851555 100644 --- a/net/ipv6/fib6_rules.c +++ b/net/ipv6/fib6_rules.c @@ -111,11 +111,13 @@ struct dst_entry *fib6_rule_lookup(struct net *net, struct flowi6 *fl6, } else { struct rt6_info *rt; - rt = lookup(net, net->ipv6.fib6_local_tbl, fl6, skb, flags); + rt = pol_lookup_func(lookup, + net, net->ipv6.fib6_local_tbl, fl6, skb, flags); if (rt != net->ipv6.ip6_null_entry && rt->dst.error != -EAGAIN) return &rt->dst; ip6_rt_put_flags(rt, flags); - rt = lookup(net, net->ipv6.fib6_main_tbl, fl6, skb, flags); + rt = pol_lookup_func(lookup, + net, net->ipv6.fib6_main_tbl, fl6, skb, flags); if (rt->dst.error != -EAGAIN) return &rt->dst; ip6_rt_put_flags(rt, flags); @@ -226,7 +228,8 @@ static int __fib6_rule_action(struct fib_rule *rule, struct flowi *flp, goto out; } - rt = lookup(net, table, flp6, arg->lookup_data, flags); + rt = pol_lookup_func(lookup, + net, table, flp6, arg->lookup_data, flags); if (rt != net->ipv6.ip6_null_entry) { err = fib6_rule_saddr(net, rule, flags, flp6, ip6_dst_idev(&rt->dst)->dev); diff --git a/net/ipv6/ip6_fib.c b/net/ipv6/ip6_fib.c index 49ee89bbcba0..25a90f3f705c 100644 --- a/net/ipv6/ip6_fib.c +++ b/net/ipv6/ip6_fib.c @@ -314,7 +314,8 @@ struct dst_entry *fib6_rule_lookup(struct net *net, struct flowi6 *fl6, { struct rt6_info *rt; - rt = lookup(net, net->ipv6.fib6_main_tbl, fl6, skb, flags); + rt = pol_lookup_func(lookup, + net, net->ipv6.fib6_main_tbl, fl6, skb, flags); if (rt->dst.error == -EAGAIN) { ip6_rt_put_flags(rt, flags); rt = net->ipv6.ip6_null_entry; diff --git a/net/ipv6/route.c b/net/ipv6/route.c index 82cbb46a2a4f..5852039ca9cf 100644 --- a/net/ipv6/route.c +++ b/net/ipv6/route.c @@ -1207,7 +1207,7 @@ static struct rt6_info *ip6_create_rt_rcu(const struct fib6_result *res) return nrt; } -static struct rt6_info *ip6_pol_route_lookup(struct net *net, +INDIRECT_CALLABLE_SCOPE struct rt6_info *ip6_pol_route_lookup(struct net *net, struct fib6_table *table, struct flowi6 *fl6, const struct sk_buff *skb, @@ -2274,7 +2274,7 @@ struct rt6_info *ip6_pol_route(struct net *net, struct fib6_table *table, } EXPORT_SYMBOL_GPL(ip6_pol_route); -static struct rt6_info *ip6_pol_route_input(struct net *net, +INDIRECT_CALLABLE_SCOPE struct rt6_info *ip6_pol_route_input(struct net *net, struct fib6_table *table, struct flowi6 *fl6, const struct sk_buff *skb, @@ -2465,7 +2465,7 @@ void ip6_route_input(struct sk_buff *skb) &fl6, skb, flags)); } -static struct rt6_info *ip6_pol_route_output(struct net *net, +INDIRECT_CALLABLE_SCOPE struct rt6_info *ip6_pol_route_output(struct net *net, struct fib6_table *table, struct flowi6 *fl6, const struct sk_buff *skb, @@ -2912,7 +2912,7 @@ struct ip6rd_flowi { struct in6_addr gateway; }; -static struct rt6_info *__ip6_route_redirect(struct net *net, +INDIRECT_CALLABLE_SCOPE struct rt6_info *__ip6_route_redirect(struct net *net, struct fib6_table *table, struct flowi6 *fl6, const struct sk_buff *skb,
It was reported that a considerable amount of cycles were spent on the expensive indirect calls on fib6_rule_lookup. This patch introduces an inline helper called pol_route_func that uses the indirect_call_wrappers to avoid the indirect calls. This patch saves around 50ns per call. Performance was measured on the receiver by checking the amount of syncookies that server was able to generate under a synflood load. Traffic was generated using trafgen[1] which was pushing around 1Mpps on a single queue. Receiver was using only one rx queue which help to create a bottle neck and make the experiment rx-bounded. These are the syncookies generated over 10s from the different runs: Whithout the patch: TcpExtSyncookiesSent 3553749 0.0 TcpExtSyncookiesSent 3550895 0.0 TcpExtSyncookiesSent 3553845 0.0 TcpExtSyncookiesSent 3541050 0.0 TcpExtSyncookiesSent 3539921 0.0 TcpExtSyncookiesSent 3557659 0.0 TcpExtSyncookiesSent 3526812 0.0 TcpExtSyncookiesSent 3536121 0.0 TcpExtSyncookiesSent 3529963 0.0 TcpExtSyncookiesSent 3536319 0.0 With the patch: TcpExtSyncookiesSent 3611786 0.0 TcpExtSyncookiesSent 3596682 0.0 TcpExtSyncookiesSent 3606878 0.0 TcpExtSyncookiesSent 3599564 0.0 TcpExtSyncookiesSent 3601304 0.0 TcpExtSyncookiesSent 3609249 0.0 TcpExtSyncookiesSent 3617437 0.0 TcpExtSyncookiesSent 3608765 0.0 TcpExtSyncookiesSent 3620205 0.0 TcpExtSyncookiesSent 3601895 0.0 Without the patch the average is 354263 pkt/s or 2822 ns/pkt and with the patch the average is 360738 pkt/s or 2772 ns/pkt which gives an estimate of 50 ns per packet. [1] http://netsniff-ng.org/ Changelog since v1: - Change ordering in the ICW (Paolo Abeni) Cc: Luigi Rizzo <lrizzo@google.com> Cc: Paolo Abeni <pabeni@redhat.com> Reported-by: Eric Dumazet <edumazet@google.com> Signed-off-by: Brian Vazquez <brianvv@google.com> --- include/net/ip6_fib.h | 36 ++++++++++++++++++++++++++++++++++++ net/ipv6/fib6_rules.c | 9 ++++++--- net/ipv6/ip6_fib.c | 3 ++- net/ipv6/route.c | 8 ++++---- 4 files changed, 48 insertions(+), 8 deletions(-)