net: sched: Merge Qdisc::bstats and Qdisc::cpu_bstats data types

The only factor differentiating per-CPU bstats data type (struct
gnet_stats_basic_cpu) from the packed non-per-CPU one (struct
gnet_stats_basic_packed) was a u64_stats sync point inside the former.
The two data types are now equivalent: earlier commits added a u64_stats
sync point to the latter.

Combine both data types into "struct gnet_stats_basic_sync". This
eliminates redundancy and simplifies the bstats read/write APIs.

Use u64_stats_t for bstats "packets" and "bytes" data types. On 64-bit
architectures, u64_stats sync points do not use sequence counter
protection.

Signed-off-by: Ahmed S. Darwish <a.darwish@linutronix.de>
Signed-off-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de>
Signed-off-by: David S. Miller <davem@davemloft.net>

5 files changed, 40 insertions, 54 deletions

diff --git a/include/net/act_api.h b/include/net/act_api.h
index f19f7f4a463c..b5b624c7e488 100644
--- a/include/net/act_api.h
+++ b/include/net/act_api.h
@@ -30,13 +30,13 @@ struct tc_action {
 	atomic_t			tcfa_bindcnt;
 	int				tcfa_action;
 	struct tcf_t			tcfa_tm;
-	struct gnet_stats_basic_packed	tcfa_bstats;
-	struct gnet_stats_basic_packed	tcfa_bstats_hw;
+	struct gnet_stats_basic_sync	tcfa_bstats;
+	struct gnet_stats_basic_sync	tcfa_bstats_hw;
 	struct gnet_stats_queue		tcfa_qstats;
 	struct net_rate_estimator __rcu *tcfa_rate_est;
 	spinlock_t			tcfa_lock;
-	struct gnet_stats_basic_cpu __percpu *cpu_bstats;
-	struct gnet_stats_basic_cpu __percpu *cpu_bstats_hw;
+	struct gnet_stats_basic_sync __percpu *cpu_bstats;
+	struct gnet_stats_basic_sync __percpu *cpu_bstats_hw;
 	struct gnet_stats_queue __percpu *cpu_qstats;
 	struct tc_cookie	__rcu *act_cookie;
 	struct tcf_chain	__rcu *goto_chain;
@@ -206,7 +206,7 @@ static inline void tcf_action_update_bstats(struct tc_action *a,
 					    struct sk_buff *skb)
 {
 	if (likely(a->cpu_bstats)) {
-		bstats_cpu_update(this_cpu_ptr(a->cpu_bstats), skb);
+		bstats_update(this_cpu_ptr(a->cpu_bstats), skb);
 		return;
 	}
 	spin_lock(&a->tcfa_lock);
diff --git a/include/net/gen_stats.h b/include/net/gen_stats.h
index 304d792f7977..52b87588f467 100644
--- a/include/net/gen_stats.h
+++ b/include/net/gen_stats.h
@@ -7,15 +7,17 @@
 #include <linux/rtnetlink.h>
 #include <linux/pkt_sched.h>
 
-/* Note: this used to be in include/uapi/linux/gen_stats.h */
-struct gnet_stats_basic_packed {
-	__u64	bytes;
-	__u64	packets;
-	struct u64_stats_sync syncp;
-};
-
-struct gnet_stats_basic_cpu {
-	struct gnet_stats_basic_packed bstats;
+/* Throughput stats.
+ * Must be initialized beforehand with gnet_stats_basic_sync_init().
+ *
+ * If no reads can ever occur parallel to writes (e.g. stack-allocated
+ * bstats), then the internal stat values can be written to and read
+ * from directly. Otherwise, use _bstats_set/update() for writes and
+ * gnet_stats_add_basic() for reads.
+ */
+struct gnet_stats_basic_sync {
+	u64_stats_t bytes;
+	u64_stats_t packets;
 	struct u64_stats_sync syncp;
 } __aligned(2 * sizeof(u64));
 
@@ -35,7 +37,7 @@ struct gnet_dump {
 	struct tc_stats   tc_stats;
 };
 
-void gnet_stats_basic_packed_init(struct gnet_stats_basic_packed *b);
+void gnet_stats_basic_sync_init(struct gnet_stats_basic_sync *b);
 int gnet_stats_start_copy(struct sk_buff *skb, int type, spinlock_t *lock,
 			  struct gnet_dump *d, int padattr);
 
@@ -46,16 +48,16 @@ int gnet_stats_start_copy_compat(struct sk_buff *skb, int type,
 
 int gnet_stats_copy_basic(const seqcount_t *running,
 			  struct gnet_dump *d,
-			  struct gnet_stats_basic_cpu __percpu *cpu,
-			  struct gnet_stats_basic_packed *b);
+			  struct gnet_stats_basic_sync __percpu *cpu,
+			  struct gnet_stats_basic_sync *b);
 void gnet_stats_add_basic(const seqcount_t *running,
-			  struct gnet_stats_basic_packed *bstats,
-			  struct gnet_stats_basic_cpu __percpu *cpu,
-			  struct gnet_stats_basic_packed *b);
+			  struct gnet_stats_basic_sync *bstats,
+			  struct gnet_stats_basic_sync __percpu *cpu,
+			  struct gnet_stats_basic_sync *b);
 int gnet_stats_copy_basic_hw(const seqcount_t *running,
 			     struct gnet_dump *d,
-			     struct gnet_stats_basic_cpu __percpu *cpu,
-			     struct gnet_stats_basic_packed *b);
+			     struct gnet_stats_basic_sync __percpu *cpu,
+			     struct gnet_stats_basic_sync *b);
 int gnet_stats_copy_rate_est(struct gnet_dump *d,
 			     struct net_rate_estimator __rcu **ptr);
 int gnet_stats_copy_queue(struct gnet_dump *d,
@@ -68,14 +70,14 @@ int gnet_stats_copy_app(struct gnet_dump *d, void *st, int len);
 
 int gnet_stats_finish_copy(struct gnet_dump *d);
 
-int gen_new_estimator(struct gnet_stats_basic_packed *bstats,
-		      struct gnet_stats_basic_cpu __percpu *cpu_bstats,
+int gen_new_estimator(struct gnet_stats_basic_sync *bstats,
+		      struct gnet_stats_basic_sync __percpu *cpu_bstats,
 		      struct net_rate_estimator __rcu **rate_est,
 		      spinlock_t *lock,
 		      seqcount_t *running, struct nlattr *opt);
 void gen_kill_estimator(struct net_rate_estimator __rcu **ptr);
-int gen_replace_estimator(struct gnet_stats_basic_packed *bstats,
-			  struct gnet_stats_basic_cpu __percpu *cpu_bstats,
+int gen_replace_estimator(struct gnet_stats_basic_sync *bstats,
+			  struct gnet_stats_basic_sync __percpu *cpu_bstats,
 			  struct net_rate_estimator __rcu **ptr,
 			  spinlock_t *lock,
 			  seqcount_t *running, struct nlattr *opt);
diff --git a/include/net/netfilter/xt_rateest.h b/include/net/netfilter/xt_rateest.h
index 832ab69efda5..4c3809e141f4 100644
--- a/include/net/netfilter/xt_rateest.h
+++ b/include/net/netfilter/xt_rateest.h
@@ -6,7 +6,7 @@
 
 struct xt_rateest {
 	/* keep lock and bstats on same cache line to speedup xt_rateest_tg() */
-	struct gnet_stats_basic_packed	bstats;
+	struct gnet_stats_basic_sync	bstats;
 	spinlock_t			lock;
 
 
diff --git a/include/net/pkt_cls.h b/include/net/pkt_cls.h
index 83a6d0792180..4a5833108083 100644
--- a/include/net/pkt_cls.h
+++ b/include/net/pkt_cls.h
@@ -765,7 +765,7 @@ struct tc_cookie {
 };
 
 struct tc_qopt_offload_stats {
-	struct gnet_stats_basic_packed *bstats;
+	struct gnet_stats_basic_sync *bstats;
 	struct gnet_stats_queue *qstats;
 };
 
@@ -885,7 +885,7 @@ struct tc_gred_qopt_offload_params {
 };
 
 struct tc_gred_qopt_offload_stats {
-	struct gnet_stats_basic_packed bstats[MAX_DPs];
+	struct gnet_stats_basic_sync bstats[MAX_DPs];
 	struct gnet_stats_queue qstats[MAX_DPs];
 	struct red_stats *xstats[MAX_DPs];
 };
diff --git a/include/net/sch_generic.h b/include/net/sch_generic.h
index d7746aea3cec..7882e3aa6448 100644
--- a/include/net/sch_generic.h
+++ b/include/net/sch_generic.h
@@ -97,7 +97,7 @@ struct Qdisc {
 	struct netdev_queue	*dev_queue;
 
 	struct net_rate_estimator __rcu *rate_est;
-	struct gnet_stats_basic_cpu __percpu *cpu_bstats;
+	struct gnet_stats_basic_sync __percpu *cpu_bstats;
 	struct gnet_stats_queue	__percpu *cpu_qstats;
 	int			pad;
 	refcount_t		refcnt;
@@ -107,7 +107,7 @@ struct Qdisc {
 	 */
 	struct sk_buff_head	gso_skb ____cacheline_aligned_in_smp;
 	struct qdisc_skb_head	q;
-	struct gnet_stats_basic_packed bstats;
+	struct gnet_stats_basic_sync bstats;
 	seqcount_t		running;
 	struct gnet_stats_queue	qstats;
 	unsigned long		state;
@@ -849,16 +849,16 @@ static inline int qdisc_enqueue(struct sk_buff *skb, struct Qdisc *sch,
 	return sch->enqueue(skb, sch, to_free);
 }
 
-static inline void _bstats_update(struct gnet_stats_basic_packed *bstats,
+static inline void _bstats_update(struct gnet_stats_basic_sync *bstats,
 				  __u64 bytes, __u32 packets)
 {
 	u64_stats_update_begin(&bstats->syncp);
-	bstats->bytes += bytes;
-	bstats->packets += packets;
+	u64_stats_add(&bstats->bytes, bytes);
+	u64_stats_add(&bstats->packets, packets);
 	u64_stats_update_end(&bstats->syncp);
 }
 
-static inline void bstats_update(struct gnet_stats_basic_packed *bstats,
+static inline void bstats_update(struct gnet_stats_basic_sync *bstats,
 				 const struct sk_buff *skb)
 {
 	_bstats_update(bstats,
@@ -866,26 +866,10 @@ static inline void bstats_update(struct gnet_stats_basic_packed *bstats,
 		       skb_is_gso(skb) ? skb_shinfo(skb)->gso_segs : 1);
 }
 
-static inline void _bstats_cpu_update(struct gnet_stats_basic_cpu *bstats,
-				      __u64 bytes, __u32 packets)
-{
-	u64_stats_update_begin(&bstats->syncp);
-	_bstats_update(&bstats->bstats, bytes, packets);
-	u64_stats_update_end(&bstats->syncp);
-}
-
-static inline void bstats_cpu_update(struct gnet_stats_basic_cpu *bstats,
-				     const struct sk_buff *skb)
-{
-	u64_stats_update_begin(&bstats->syncp);
-	bstats_update(&bstats->bstats, skb);
-	u64_stats_update_end(&bstats->syncp);
-}
-
 static inline void qdisc_bstats_cpu_update(struct Qdisc *sch,
 					   const struct sk_buff *skb)
 {
-	bstats_cpu_update(this_cpu_ptr(sch->cpu_bstats), skb);
+	bstats_update(this_cpu_ptr(sch->cpu_bstats), skb);
 }
 
 static inline void qdisc_bstats_update(struct Qdisc *sch,
@@ -1317,7 +1301,7 @@ void psched_ppscfg_precompute(struct psched_pktrate *r, u64 pktrate64);
 struct mini_Qdisc {
 	struct tcf_proto *filter_list;
 	struct tcf_block *block;
-	struct gnet_stats_basic_cpu __percpu *cpu_bstats;
+	struct gnet_stats_basic_sync __percpu *cpu_bstats;
 	struct gnet_stats_queue	__percpu *cpu_qstats;
 	struct rcu_head rcu;
 };
@@ -1325,7 +1309,7 @@ struct mini_Qdisc {
 static inline void mini_qdisc_bstats_cpu_update(struct mini_Qdisc *miniq,
 						const struct sk_buff *skb)
 {
-	bstats_cpu_update(this_cpu_ptr(miniq->cpu_bstats), skb);
+	bstats_update(this_cpu_ptr(miniq->cpu_bstats), skb);
 }
 
 static inline void mini_qdisc_qstats_cpu_drop(struct mini_Qdisc *miniq)