[PATCH] lockfree rtcache using RCU

[PATCH] lockfree rtcache using RCU

[Dipankar Sarma]

This patch was discussed along with results months ago and Davem
asked me to send it to him for inclusion when RCU core is in.
Now that the RCU core is in, please include it. This however
depends on the RCU helper patches I had submitted earlier to
Linus which haven't yet been merged AFAICS. If however they have been
merged, then this can be applied.

It speeds up route cache lookup by 30-50% approximately as measured
with synthetic benchmark suggested by Dave.

http://marc.theaimsgroup.com/?l=linux-kernel&m=102258603404836&w=2

The entire discussion is in -
http://marc.theaimsgroup.com/?t=102258611100001&r=1&w=2

Thanks
-- 
Dipankar Sarma  <dipankar@in.ibm.com> http://lse.sourceforge.net
Linux Technology Center, IBM Software Lab, Bangalore, India.

rt_rcu-2.5.43-1.patch
---------------------

diff -urN linux-2.5.43-base/include/net/dst.h linux-2.5.43-rt_rcu/include/net/dst.h
--- linux-2.5.43-base/include/net/dst.h	Wed Oct 16 08:59:04 2002
+++ linux-2.5.43-rt_rcu/include/net/dst.h	Sat Oct 19 03:25:26 2002
@@ -9,6 +9,7 @@
 #define _NET_DST_H
 
 #include <linux/config.h>
+#include <linux/rcupdate.h>
 #include <net/neighbour.h>
 
 /*
@@ -62,6 +63,7 @@
 #endif
 
 	struct  dst_ops	        *ops;
+	struct rcu_head		rcu_head;
 		
 	char			info[0];
 };
diff -urN linux-2.5.43-base/net/ipv4/route.c linux-2.5.43-rt_rcu/net/ipv4/route.c
--- linux-2.5.43-base/net/ipv4/route.c	Wed Oct 16 08:59:06 2002
+++ linux-2.5.43-rt_rcu/net/ipv4/route.c	Sat Oct 19 04:05:16 2002
@@ -85,6 +85,7 @@
 #include <linux/mroute.h>
 #include <linux/netfilter_ipv4.h>
 #include <linux/random.h>
+#include <linux/rcupdate.h>
 #include <net/protocol.h>
 #include <net/ip.h>
 #include <net/route.h>
@@ -188,7 +189,7 @@
 
 struct rt_hash_bucket {
 	struct rtable	*chain;
-	rwlock_t	lock;
+	spinlock_t	lock;
 } __attribute__((__aligned__(8)));
 
 static struct rt_hash_bucket 	*rt_hash_table;
@@ -226,8 +227,8 @@
 		len = 128;
   	}
 	
+	rcu_read_lock();
 	for (i = rt_hash_mask; i >= 0; i--) {
-		read_lock_bh(&rt_hash_table[i].lock);
 		for (r = rt_hash_table[i].chain; r; r = r->u.rt_next) {
 			/*
 			 *	Spin through entries until we are ready
@@ -238,6 +239,7 @@
 				len = 0;
 				continue;
 			}
+			read_barrier_depends();
 			sprintf(temp, "%s\t%08lX\t%08lX\t%8X\t%d\t%u\t%d\t"
 				"%08lX\t%d\t%u\t%u\t%02X\t%d\t%1d\t%08X",
 				r->u.dst.dev ? r->u.dst.dev->name : "*",
@@ -262,14 +264,13 @@
 			sprintf(buffer + len, "%-127s\n", temp);
 			len += 128;
 			if (pos >= offset+length) {
-				read_unlock_bh(&rt_hash_table[i].lock);
 				goto done;
 			}
 		}
-		read_unlock_bh(&rt_hash_table[i].lock);
         }
 
 done:
+	rcu_read_unlock();
   	*start = buffer + len - (pos - offset);
   	len = pos - offset;
   	if (len > length)
@@ -318,13 +319,13 @@
   
 static __inline__ void rt_free(struct rtable *rt)
 {
-	dst_free(&rt->u.dst);
+	call_rcu(&rt->u.dst.rcu_head, (void (*)(void *))dst_free, &rt->u.dst);
 }
 
 static __inline__ void rt_drop(struct rtable *rt)
 {
 	ip_rt_put(rt);
-	dst_free(&rt->u.dst);
+	call_rcu(&rt->u.dst.rcu_head, (void (*)(void *))dst_free, &rt->u.dst);
 }
 
 static __inline__ int rt_fast_clean(struct rtable *rth)
@@ -377,7 +378,7 @@
 		i = (i + 1) & rt_hash_mask;
 		rthp = &rt_hash_table[i].chain;
 
-		write_lock(&rt_hash_table[i].lock);
+		spin_lock(&rt_hash_table[i].lock);
 		while ((rth = *rthp) != NULL) {
 			if (rth->u.dst.expires) {
 				/* Entry is expired even if it is in use */
@@ -396,7 +397,7 @@
 			*rthp = rth->u.rt_next;
 			rt_free(rth);
 		}
-		write_unlock(&rt_hash_table[i].lock);
+		spin_unlock(&rt_hash_table[i].lock);
 
 		/* Fallback loop breaker. */
 		if ((jiffies - now) > 0)
@@ -419,11 +420,11 @@
 	rt_deadline = 0;
 
 	for (i = rt_hash_mask; i >= 0; i--) {
-		write_lock_bh(&rt_hash_table[i].lock);
+		spin_lock_bh(&rt_hash_table[i].lock);
 		rth = rt_hash_table[i].chain;
 		if (rth)
 			rt_hash_table[i].chain = NULL;
-		write_unlock_bh(&rt_hash_table[i].lock);
+		spin_unlock_bh(&rt_hash_table[i].lock);
 
 		for (; rth; rth = next) {
 			next = rth->u.rt_next;
@@ -547,7 +548,7 @@
 
 			k = (k + 1) & rt_hash_mask;
 			rthp = &rt_hash_table[k].chain;
-			write_lock_bh(&rt_hash_table[k].lock);
+			spin_lock_bh(&rt_hash_table[k].lock);
 			while ((rth = *rthp) != NULL) {
 				if (!rt_may_expire(rth, tmo, expire)) {
 					tmo >>= 1;
@@ -558,7 +559,7 @@
 				rt_free(rth);
 				goal--;
 			}
-			write_unlock_bh(&rt_hash_table[k].lock);
+			spin_unlock_bh(&rt_hash_table[k].lock);
 			if (goal <= 0)
 				break;
 		}
@@ -619,7 +620,7 @@
 restart:
 	rthp = &rt_hash_table[hash].chain;
 
-	write_lock_bh(&rt_hash_table[hash].lock);
+	spin_lock_bh(&rt_hash_table[hash].lock);
 	while ((rth = *rthp) != NULL) {
 		if (memcmp(&rth->fl, &rt->fl, sizeof(rt->fl)) == 0) {
 			/* Put it first */
@@ -630,7 +631,7 @@
 			rth->u.dst.__use++;
 			dst_hold(&rth->u.dst);
 			rth->u.dst.lastuse = now;
-			write_unlock_bh(&rt_hash_table[hash].lock);
+			spin_unlock_bh(&rt_hash_table[hash].lock);
 
 			rt_drop(rt);
 			*rp = rth;
@@ -646,7 +647,7 @@
 	if (rt->rt_type == RTN_UNICAST || rt->fl.iif == 0) {
 		int err = arp_bind_neighbour(&rt->u.dst);
 		if (err) {
-			write_unlock_bh(&rt_hash_table[hash].lock);
+			spin_unlock_bh(&rt_hash_table[hash].lock);
 
 			if (err != -ENOBUFS) {
 				rt_drop(rt);
@@ -687,7 +688,7 @@
 	}
 #endif
 	rt_hash_table[hash].chain = rt;
-	write_unlock_bh(&rt_hash_table[hash].lock);
+	spin_unlock_bh(&rt_hash_table[hash].lock);
 	*rp = rt;
 	return 0;
 }
@@ -754,7 +755,7 @@
 {
 	struct rtable **rthp;
 
-	write_lock_bh(&rt_hash_table[hash].lock);
+	spin_lock_bh(&rt_hash_table[hash].lock);
 	ip_rt_put(rt);
 	for (rthp = &rt_hash_table[hash].chain; *rthp;
 	     rthp = &(*rthp)->u.rt_next)
@@ -763,7 +764,7 @@
 			rt_free(rt);
 			break;
 		}
-	write_unlock_bh(&rt_hash_table[hash].lock);
+	spin_unlock_bh(&rt_hash_table[hash].lock);
 }
 
 void ip_rt_redirect(u32 old_gw, u32 daddr, u32 new_gw,
@@ -802,10 +803,11 @@
 
 			rthp=&rt_hash_table[hash].chain;
 
-			read_lock(&rt_hash_table[hash].lock);
+			rcu_read_lock();
 			while ((rth = *rthp) != NULL) {
 				struct rtable *rt;
 
+				read_barrier_depends();
 				if (rth->fl.fl4_dst != daddr ||
 				    rth->fl.fl4_src != skeys[i] ||
 				    rth->fl.fl4_tos != tos ||
@@ -823,7 +825,7 @@
 					break;
 
 				dst_clone(&rth->u.dst);
-				read_unlock(&rt_hash_table[hash].lock);
+				rcu_read_unlock();
 
 				rt = dst_alloc(&ipv4_dst_ops);
 				if (rt == NULL) {
@@ -834,6 +836,7 @@
 
 				/* Copy all the information. */
 				*rt = *rth;
+ 				INIT_RCU_HEAD(&rt->u.dst.rcu_head);
 				rt->u.dst.__use		= 1;
 				atomic_set(&rt->u.dst.__refcnt, 1);
 				if (rt->u.dst.dev)
@@ -869,7 +872,7 @@
 					ip_rt_put(rt);
 				goto do_next;
 			}
-			read_unlock(&rt_hash_table[hash].lock);
+			rcu_read_unlock();
 		do_next:
 			;
 		}
@@ -1049,9 +1052,10 @@
 	for (i = 0; i < 2; i++) {
 		unsigned hash = rt_hash_code(daddr, skeys[i], tos);
 
-		read_lock(&rt_hash_table[hash].lock);
+		rcu_read_lock();
 		for (rth = rt_hash_table[hash].chain; rth;
 		     rth = rth->u.rt_next) {
+			read_barrier_depends();
 			if (rth->fl.fl4_dst == daddr &&
 			    rth->fl.fl4_src == skeys[i] &&
 			    rth->rt_dst  == daddr &&
@@ -1087,7 +1091,7 @@
 				}
 			}
 		}
-		read_unlock(&rt_hash_table[hash].lock);
+		rcu_read_unlock();
 	}
 	return est_mtu ? : new_mtu;
 }
@@ -1639,8 +1643,9 @@
 	tos &= IPTOS_RT_MASK;
 	hash = rt_hash_code(daddr, saddr ^ (iif << 5), tos);
 
-	read_lock(&rt_hash_table[hash].lock);
+	rcu_read_lock();
 	for (rth = rt_hash_table[hash].chain; rth; rth = rth->u.rt_next) {
+		read_barrier_depends();
 		if (rth->fl.fl4_dst == daddr &&
 		    rth->fl.fl4_src == saddr &&
 		    rth->fl.iif == iif &&
@@ -1653,12 +1658,12 @@
 			dst_hold(&rth->u.dst);
 			rth->u.dst.__use++;
 			rt_cache_stat[smp_processor_id()].in_hit++;
-			read_unlock(&rt_hash_table[hash].lock);
+			rcu_read_unlock();
 			skb->dst = (struct dst_entry*)rth;
 			return 0;
 		}
 	}
-	read_unlock(&rt_hash_table[hash].lock);
+	rcu_read_unlock();
 
 	/* Multicast recognition logic is moved from route cache to here.
 	   The problem was that too many Ethernet cards have broken/missing
@@ -1998,8 +2003,9 @@
 
 	hash = rt_hash_code(flp->fl4_dst, flp->fl4_src ^ (flp->oif << 5), flp->fl4_tos);
 
-	read_lock_bh(&rt_hash_table[hash].lock);
+	rcu_read_lock();
 	for (rth = rt_hash_table[hash].chain; rth; rth = rth->u.rt_next) {
+		read_barrier_depends();
 		if (rth->fl.fl4_dst == flp->fl4_dst &&
 		    rth->fl.fl4_src == flp->fl4_src &&
 		    rth->fl.iif == 0 &&
@@ -2013,12 +2019,12 @@
 			dst_hold(&rth->u.dst);
 			rth->u.dst.__use++;
 			rt_cache_stat[smp_processor_id()].out_hit++;
-			read_unlock_bh(&rt_hash_table[hash].lock);
+			rcu_read_unlock();
 			*rp = rth;
 			return 0;
 		}
 	}
-	read_unlock_bh(&rt_hash_table[hash].lock);
+	rcu_read_unlock();
 
 	return ip_route_output_slow(rp, flp);
 }	
@@ -2208,9 +2214,10 @@
 		if (h < s_h) continue;
 		if (h > s_h)
 			s_idx = 0;
-		read_lock_bh(&rt_hash_table[h].lock);
+		rcu_read_lock();
 		for (rt = rt_hash_table[h].chain, idx = 0; rt;
 		     rt = rt->u.rt_next, idx++) {
+			read_barrier_depends();
 			if (idx < s_idx)
 				continue;
 			skb->dst = dst_clone(&rt->u.dst);
@@ -2218,12 +2225,12 @@
 					 cb->nlh->nlmsg_seq,
 					 RTM_NEWROUTE, 1) <= 0) {
 				dst_release(xchg(&skb->dst, NULL));
-				read_unlock_bh(&rt_hash_table[h].lock);
+				rcu_read_unlock();
 				goto done;
 			}
 			dst_release(xchg(&skb->dst, NULL));
 		}
-		read_unlock_bh(&rt_hash_table[h].lock);
+		rcu_read_unlock();
 	}
 
 done:
@@ -2505,7 +2512,7 @@
 
 	rt_hash_mask--;
 	for (i = 0; i <= rt_hash_mask; i++) {
-		rt_hash_table[i].lock = RW_LOCK_UNLOCKED;
+		rt_hash_table[i].lock = SPIN_LOCK_UNLOCKED;
 		rt_hash_table[i].chain = NULL;
 	}
 

Re: [PATCH] lockfree rtcache using RCU

[David S. Miller]

   From: Dipankar Sarma <dipankar@in.ibm.com>
   Date: Sat, 19 Oct 2002 06:52:10 +0530

   This patch was discussed along with results months ago and Davem
   asked me to send it to him for inclusion when RCU core is in.
   Now that the RCU core is in, please include it.

Patch applied, thanks.

Re: [PATCH] lockfree rtcache using RCU

[David S. Miller]

Dipankar, I had to revert this patch.

Where is read_barrier_depends() defined?
I do not see it in current 2.5.x sources.

Re: [PATCH] lockfree rtcache using RCU

[Dipankar Sarma]

On Fri, Oct 18, 2002 at 06:38:11PM -0700, David S. Miller wrote:
> 
> Dipankar, I had to revert this patch.
> 
> Where is read_barrier_depends() defined?
> I do not see it in current 2.5.x sources.

Well, the read_barrier_depends helper patch is applied now to 
Linus' tree. Sorry, I should have been more clear.

Thanks
-- 
Dipankar Sarma  <dipankar@in.ibm.com> http://lse.sourceforge.net
Linux Technology Center, IBM Software Lab, Bangalore, India.