net-memcg: Introduce net.core.memcg_exclusive sysctl.

If net.core.memcg_exclusive is 1 when sk->sk_memcg is allocated,
the socket is flagged with SK_MEMCG_EXCLUSIVE internally and skips
the global per-protocol memory accounting.

OTOH, for accept()ed child sockets, this flag is inherited from
the listening socket in sk_clone_lock() and set in __inet_accept().
This is to preserve the decision by BPF which will be supported later.

Given sk->sk_memcg can be accessed in the fast path, it would
be preferable to place the flag field in the same cache line as
sk->sk_memcg.

However, struct sock does not have such a 1-byte hole.

Let's store the flag in the lowest bit of sk->sk_memcg and check
it in mem_cgroup_sk_exclusive().

Tested with a script that creates local socket pairs and send()s a
bunch of data without recv()ing.

Setup:

  # mkdir /sys/fs/cgroup/test
  # echo $$ >> /sys/fs/cgroup/test/cgroup.procs
  # sysctl -q net.ipv4.tcp_mem="1000 1000 1000"

Without net.core.memcg_exclusive, charged to memcg & tcp_mem:

  # prlimit -n=524288:524288 bash -c "python3 pressure.py" &
  # cat /sys/fs/cgroup/test/memory.stat | grep sock
  sock 22642688 <-------------------------------------- charged to memcg
  # cat /proc/net/sockstat| grep TCP
  TCP: inuse 2006 orphan 0 tw 0 alloc 2008 mem 5376 <-- charged to tcp_mem
  # ss -tn | head -n 5
  State Recv-Q Send-Q Local Address:Port  Peer Address:Port
  ESTAB 2000   0          127.0.0.1:34479    127.0.0.1:53188
  ESTAB 2000   0          127.0.0.1:34479    127.0.0.1:49972
  ESTAB 2000   0          127.0.0.1:34479    127.0.0.1:53868
  ESTAB 2000   0          127.0.0.1:34479    127.0.0.1:53554
  # nstat | grep Pressure || echo no pressure
  TcpExtTCPMemoryPressures        1                  0.0

With net.core.memcg_exclusive=1, only charged to memcg:

  # sysctl -q net.core.memcg_exclusive=1
  # prlimit -n=524288:524288 bash -c "python3 pressure.py" &
  # cat /sys/fs/cgroup/test/memory.stat | grep sock
  sock 2757468160 <------------------------------------ charged to memcg
  # cat /proc/net/sockstat | grep TCP
  TCP: inuse 2006 orphan 0 tw 0 alloc 2008 mem 0 <- NOT charged to tcp_mem
  # ss -tn | head -n 5
  State Recv-Q Send-Q  Local Address:Port  Peer Address:Port
  ESTAB 111000 0           127.0.0.1:36019    127.0.0.1:49026
  ESTAB 110000 0           127.0.0.1:36019    127.0.0.1:45630
  ESTAB 110000 0           127.0.0.1:36019    127.0.0.1:44870
  ESTAB 111000 0           127.0.0.1:36019    127.0.0.1:45274
  # nstat | grep Pressure || echo no pressure
  no pressure

Signed-off-by: Kuniyuki Iwashima <[email protected]>

Author: q2ven

Documentation/admin-guide/sysctl/net.rst

@@ -212,6 +212,15 @@ mem_pcpu_rsv
 
 Per-cpu reserved forward alloc cache size in page units. Default 1MB per CPU.
 
+memcg_exclusive
+---------------
+
+Skip charging socket buffers to the per-protocol global memory accounting
+(controlled by net.ipv4.tcp_mem, etc) if they are already charged to the
+cgroup memory controller ("sock" in memory.stat file).
+
+Default: 0
+
 rmem_default
 ------------
 

include/net/netns/core.h

@@ -16,6 +16,9 @@ struct netns_core {
 	int	sysctl_optmem_max;
 	u8	sysctl_txrehash;
 	u8	sysctl_tstamp_allow_data;
+#ifdef CONFIG_MEMCG
+	u8	sysctl_memcg_exclusive;
+#endif
 
 #ifdef CONFIG_PROC_FS
 	struct prot_inuse __percpu *prot_inuse;

include/net/sock.h

@@ -2614,10 +2614,36 @@ static inline gfp_t gfp_memcg_charge(void)
 	return in_softirq() ? GFP_ATOMIC : GFP_KERNEL;
 }
 
+enum {
+	SK_MEMCG_EXCLUSIVE	= (1UL << 0),
+	SK_MEMCG_FLAG_MAX	= (1UL << 1),
+};
+
+#define SK_MEMCG_FLAG_MASK	(SK_MEMCG_FLAG_MAX - 1)
+#define SK_MEMCG_PTR_MASK	~SK_MEMCG_FLAG_MASK
+
 #ifdef CONFIG_MEMCG
 static inline struct mem_cgroup *mem_cgroup_from_sk(const struct sock *sk)
 {
-	return sk->sk_memcg;
+	unsigned long val = (unsigned long)sk->sk_memcg;
+
+	val &= SK_MEMCG_PTR_MASK;
+	return (struct mem_cgroup *)val;
+}
+
+static inline void mem_cgroup_sk_set_flags(struct sock *sk, unsigned short flags)
+{
+	unsigned long val = (unsigned long)mem_cgroup_from_sk(sk);
+
+	val |= flags;
+	sk->sk_memcg = (struct mem_cgroup *)val;
+}
+
+static inline unsigned short mem_cgroup_sk_get_flags(const struct sock *sk)
+{
+	unsigned long val = (unsigned long)sk->sk_memcg;
+
+	return val & SK_MEMCG_FLAG_MASK;
 }
 
 static inline bool mem_cgroup_sk_enabled(const struct sock *sk)
@@ -2627,7 +2653,7 @@ static inline bool mem_cgroup_sk_enabled(const struct sock *sk)
 
 static inline bool mem_cgroup_sk_exclusive(const struct sock *sk)
 {
-	return false;
+	return mem_cgroup_sk_get_flags(sk) & SK_MEMCG_EXCLUSIVE;
 }
 
 static inline bool mem_cgroup_sk_under_memory_pressure(const struct sock *sk)
@@ -2652,6 +2678,15 @@ static inline struct mem_cgroup *mem_cgroup_from_sk(const struct sock *sk)
 	return NULL;
 }
 
+static inline void mem_cgroup_sk_set_flags(struct sock *sk, unsigned short flags)
+{
+}
+
+static inline unsigned short mem_cgroup_sk_get_flags(const struct sock *sk)
+{
+	return 0;
+}
+
 static inline bool mem_cgroup_sk_enabled(const struct sock *sk)
 {
 	return false;

mm/memcontrol.c

@@ -4995,6 +4995,16 @@ void mem_cgroup_migrate(struct folio *old, struct folio *new)
 DEFINE_STATIC_KEY_FALSE(memcg_sockets_enabled_key);
 EXPORT_SYMBOL(memcg_sockets_enabled_key);
 
+static void mem_cgroup_sk_set(struct sock *sk, struct mem_cgroup *memcg)
+{
+	sk->sk_memcg = memcg;
+
+#ifdef CONFIG_NET
+	if (READ_ONCE(sock_net(sk)->core.sysctl_memcg_exclusive))
+		mem_cgroup_sk_set_flags(sk, SK_MEMCG_EXCLUSIVE);
+#endif
+}
+
 void mem_cgroup_sk_alloc(struct sock *sk)
 {
 	struct mem_cgroup *memcg;
@@ -5013,7 +5023,7 @@ void mem_cgroup_sk_alloc(struct sock *sk)
 	if (!cgroup_subsys_on_dfl(memory_cgrp_subsys) && !memcg1_tcpmem_active(memcg))
 		goto out;
 	if (css_tryget(&memcg->css))
-		sk->sk_memcg = memcg;
+		mem_cgroup_sk_set(sk, memcg);
 out:
 	rcu_read_unlock();
 }

net/core/sock.c

@@ -2520,6 +2520,7 @@ struct sock *sk_clone_lock(const struct sock *sk, const gfp_t priority)
 #ifdef CONFIG_MEMCG
 	/* sk->sk_memcg will be populated at accept() time */
 	newsk->sk_memcg = NULL;
+	mem_cgroup_sk_set_flags(newsk, mem_cgroup_sk_get_flags(sk));
 #endif
 
 	cgroup_sk_clone(&newsk->sk_cgrp_data);

net/core/sysctl_net_core.c

@@ -676,6 +676,17 @@ static struct ctl_table netns_core_table[] = {
 		.extra1		= SYSCTL_ZERO,
 		.extra2		= SYSCTL_ONE
 	},
+#ifdef CONFIG_MEMCG
+	{
+		.procname	= "memcg_exclusive",
+		.data		= &init_net.core.sysctl_memcg_exclusive,
+		.maxlen		= sizeof(u8),
+		.mode		= 0644,
+		.proc_handler	= proc_dou8vec_minmax,
+		.extra1		= SYSCTL_ZERO,
+		.extra2		= SYSCTL_ONE
+	},
+#endif
 	/* sysctl_core_net_init() will set the values after this
 	 * to readonly in network namespaces
 	 */

net/ipv4/af_inet.c

@@ -760,12 +760,16 @@ void __inet_accept(struct socket *sock, struct socket *newsock, struct sock *new
 	if (mem_cgroup_sockets_enabled &&
 	    (!IS_ENABLED(CONFIG_IP_SCTP) || sk_is_tcp(newsk))) {
 		gfp_t gfp = GFP_KERNEL | __GFP_NOFAIL;
+		unsigned short flags;
 
+		flags = mem_cgroup_sk_get_flags(newsk);
 		mem_cgroup_sk_alloc(newsk);
 
 		if (mem_cgroup_from_sk(newsk)) {
 			int amt;
 
+			mem_cgroup_sk_set_flags(newsk, flags);
+
 			/* The socket has not been accepted yet, no need
 			 * to look at newsk->sk_wmem_queued.
 			 */