Skip to contentSkip to conversation

tcp: replace hard coded GFP_KERNEL with sk_allocation

PATCH RESEND Applied Wu Fengguang 2 participants 4 messages Sep 3, 2009

Subsystems (5)
Networking Network drivers NFS client NFS server +1 more subsystems
lore
Share & tools
lore

Wu Fengguang

author
#0 lore raw

PATCH RESEND 1/1 tcp: replace hard coded GFP_KERNEL with sk_allocation

This fixed a lockdep warning which appeared when doing stress memory tests over NFS:

inconsistent {RECLAIM_FS-ON-W} -> {IN-RECLAIM_FS-W} usage.

page reclaim => nfs_writepage => tcp_sendmsg => lock sk_lock

mount_root => nfs_root_data => tcp_close => lock sk_lock =>
tcp_send_fin => alloc_skb_fclone => page reclaim

David raised a concern that if the allocation fails in tcp_send_fin(), and it's GFP_ATOMIC, we are going to yield() (which sleeps) and loop endlessly waiting for the allocation to succeed.

But fact is, the original GFP_KERNEL also sleeps. GFP_ATOMIC+yield() looks weird, but it is no worse the implicit sleep inside GFP_KERNEL. Both could loop endlessly under memory pressure.

CC Arnaldo Carvalho de Melo <[email protected]>
CC David S. Miller <[email protected]>
CC Herbert Xu <[email protected]>
Signed-off-by Wu Fengguang <[email protected]>
---
 net/ipv4/tcp.c        |    4 ++--
 net/ipv4/tcp_ipv4.c   |    5 +++--
 net/ipv4/tcp_output.c |    5 +++--
 3 files changed, 8 insertions(+), 6 deletions(-)
net/ipv4/tcp_ipv4.c +3 -2
@@ -970,8 +970,9 @@ static int tcp_v4_parse_md5_keys(struct 
 
 	if (!tcp_sk(sk)->md5sig_info) {
 		struct tcp_sock *tp = tcp_sk(sk);
-		struct tcp_md5sig_info *p = kzalloc(sizeof(*p), GFP_KERNEL);
+		struct tcp_md5sig_info *p;
 
+		p = kzalloc(sizeof(*p), sk->sk_allocation);
 		if (!p)
 			return -EINVAL;
 

@@ -979,7 +980,7 @@ static int tcp_v4_parse_md5_keys(struct 
 		sk->sk_route_caps &= ~NETIF_F_GSO_MASK;
 	}
 
-	newkey = kmemdup(cmd.tcpm_key, cmd.tcpm_keylen, GFP_KERNEL);
+	newkey = kmemdup(cmd.tcpm_key, cmd.tcpm_keylen, sk->sk_allocation);
 	if (!newkey)
 		return -ENOMEM;
 	return tcp_v4_md5_do_add(sk, sin->sin_addr.s_addr,
net/ipv4/tcp_output.c +3 -2
@@ -2100,7 +2100,8 @@ void tcp_send_fin(struct sock *sk)
 	} else {
 		/* Socket is locked, keep trying until memory is available. */
 		for (;;) {
-			skb = alloc_skb_fclone(MAX_TCP_HEADER, GFP_KERNEL);
+			skb = alloc_skb_fclone(MAX_TCP_HEADER,
+					       sk->sk_allocation);
 			if (skb)
 				break;
 			yield();

@@ -2358,7 +2359,7 @@ int tcp_connect(struct sock *sk)
 	sk->sk_wmem_queued += buff->truesize;
 	sk_mem_charge(sk, buff->truesize);
 	tp->packets_out += tcp_skb_pcount(buff);
-	tcp_transmit_skb(sk, buff, 1, GFP_KERNEL);
+	tcp_transmit_skb(sk, buff, 1, sk->sk_allocation);
 
 	/* We change tp->snd_nxt after the tcp_transmit_skb() call
 	 * in order to make this packet get counted in tcpOutSegs.
net/ipv4/tcp.c +2 -2
@@ -1834,7 +1834,7 @@ void tcp_close(struct sock *sk, long tim
 		/* Unread data was tossed, zap the connection. */
 		NET_INC_STATS_USER(sock_net(sk), LINUX_MIB_TCPABORTONCLOSE);
 		tcp_set_state(sk, TCP_CLOSE);
-		tcp_send_active_reset(sk, GFP_KERNEL);
+		tcp_send_active_reset(sk, sk->sk_allocation);
 	} else if (sock_flag(sk, SOCK_LINGER) && !sk->sk_lingertime) {
 		/* Check zero linger _after_ checking for unread data. */
 		sk->sk_prot->disconnect(sk, 0);

@@ -2666,7 +2666,7 @@ static struct tcp_md5sig_pool **__tcp_al
 		struct tcp_md5sig_pool *p;
 		struct crypto_hash *hash;
 
-		p = kzalloc(sizeof(*p), GFP_KERNEL);
+		p = kzalloc(sizeof(*p), sk->sk_allocation);
 		if (!p)
 			goto out_free;
 		*per_cpu_ptr(pool, cpu) = p;
maintainer Applied
#1 lore raw

From: Wu Fengguang <[email protected]>
Date: Thu, 3 Sep 2009 12:04:07 +0800

Quoted 14 lines

This fixed a lockdep warning which appeared when doing stress memory tests over NFS:

inconsistent {RECLAIM_FS-ON-W} -> {IN-RECLAIM_FS-W} usage.

page reclaim => nfs_writepage => tcp_sendmsg => lock sk_lock

mount_root => nfs_root_data => tcp_close => lock sk_lock =>
tcp_send_fin => alloc_skb_fclone => page reclaim

David raised a concern that if the allocation fails in tcp_send_fin(), and it's GFP_ATOMIC, we are going to yield() (which sleeps) and loop endlessly waiting for the allocation to succeed.

Quoted 8 lines

But fact is, the original GFP_KERNEL also sleeps. GFP_ATOMIC+yield() looks weird, but it is no worse the implicit sleep inside GFP_KERNEL. Both could loop endlessly under memory pressure.

CC Arnaldo Carvalho de Melo <[email protected]>
CC David S. Miller <[email protected]>
CC Herbert Xu <[email protected]>
Signed-off-by Wu Fengguang <[email protected]>

Applied to net-next-2.6, thanks!

maintainer
#2 lore raw

From: David Miller <[email protected]>
Date: Wed, 02 Sep 2009 23:10:17 -0700 (PDT)

Quoted 21 lines

From: Wu Fengguang <[email protected]>
Date: Thu, 3 Sep 2009 12:04:07 +0800

Quoted 18 lines

This fixed a lockdep warning which appeared when doing stress memory tests over NFS:

inconsistent {RECLAIM_FS-ON-W} -> {IN-RECLAIM_FS-W} usage.

page reclaim => nfs_writepage => tcp_sendmsg => lock sk_lock

mount_root => nfs_root_data => tcp_close => lock sk_lock =>
tcp_send_fin => alloc_skb_fclone => page reclaim

David raised a concern that if the allocation fails in tcp_send_fin(), and it's GFP_ATOMIC, we are going to yield() (which sleeps) and loop endlessly waiting for the allocation to succeed.

But fact is, the original GFP_KERNEL also sleeps. GFP_ATOMIC+yield() looks weird, but it is no worse the implicit sleep inside GFP_KERNEL. Both could loop endlessly under memory pressure.

Quoted 6 lines
Quoted 4 lines
CC Arnaldo Carvalho de Melo <[email protected]>
CC David S. Miller <[email protected]>
CC Herbert Xu <[email protected]>
Signed-off-by Wu Fengguang <[email protected]>

Applied to net-next-2.6, thanks!

You obviously didn't build test this with TCP MD5 support enabled, that fails.

I'm fixing it up, but if you're going to go through the motions of submitting a patch multiple times, at least do a thorough build test of the code you're changing.

author
#3 lore raw
Quoted 29 lines · David Miller

From: David Miller <[email protected]>
Date: Wed, 02 Sep 2009 23:10:17 -0700 (PDT)

Quoted 26 lines

From: Wu Fengguang <[email protected]>
Date: Thu, 3 Sep 2009 12:04:07 +0800

Quoted 22 lines

This fixed a lockdep warning which appeared when doing stress memory tests over NFS:

inconsistent {RECLAIM_FS-ON-W} -> {IN-RECLAIM_FS-W} usage.

page reclaim => nfs_writepage => tcp_sendmsg => lock sk_lock

mount_root => nfs_root_data => tcp_close => lock sk_lock =>
tcp_send_fin => alloc_skb_fclone => page reclaim

David raised a concern that if the allocation fails in tcp_send_fin(), and it's GFP_ATOMIC, we are going to yield() (which sleeps) and loop endlessly waiting for the allocation to succeed.

But fact is, the original GFP_KERNEL also sleeps. GFP_ATOMIC+yield() looks weird, but it is no worse the implicit sleep inside GFP_KERNEL. Both could loop endlessly under memory pressure.

CC Arnaldo Carvalho de Melo <[email protected]>
CC David S. Miller <[email protected]>
CC Herbert Xu <[email protected]>
Signed-off-by Wu Fengguang <[email protected]>
Quoted 4 lines
Quoted 1 line

Applied to net-next-2.6, thanks!

You obviously didn't build test this with TCP MD5 support enabled, that fails.

Ah sorry! I compile bare kernels on my laptop..

Quoted 3 lines

I'm fixing it up, but if you're going to go through the motions of submitting a patch multiple times, at least do a thorough build test of the code you're changing.

Good advice. I'll consider a build server.

Thanks,
Fengguang