[PATCH 1/2] tcp: bind() fix autoselection to share ports

The current code checks for conflicts when the application
requests a specific port.  If there is no conflict, then
the request is granted.

On the other hand, the port autoselection done by the kernel
fails when all ports are bound even when there is a port
with no conflict available.

The fix changes port autoselection to check if there is a
conflict and use it if not.

Signed-off-by: Flavio Leitner <redacted>
---
 net/ipv4/inet_connection_sock.c |    5 +++++
 1 files changed, 5 insertions(+), 0 deletions(-)

diff --git a/net/ipv4/inet_connection_sock.c b/net/ipv4/inet_connection_sock.c
index 2e4e244..ecd19b5 100644
--- a/net/ipv4/inet_connection_sock.c
+++ b/net/ipv4/inet_connection_sock.c

@@ -128,6 +128,11 @@ again:
 							goto have_snum;
 						}
 					}
+					if (!inet_csk(sk)->icsk_af_ops->bind_conflict(sk, tb)) {
+						spin_unlock(&head->lock);
+						snum = rover;
+						goto have_snum;
+					}
 					goto next;
 				}
 			break;

-- 
1.7.7.1

Port autoselection finds a port and then drop the lock,
then right after that, gets the hash bucket again and lock it.

Fix it to go direct.

Signed-off-by: Flavio Leitner <redacted>
---
 net/ipv4/inet_connection_sock.c |    6 ++----
 1 files changed, 2 insertions(+), 4 deletions(-)

diff --git a/net/ipv4/inet_connection_sock.c b/net/ipv4/inet_connection_sock.c
index ecd19b5..19d66ce 100644
--- a/net/ipv4/inet_connection_sock.c
+++ b/net/ipv4/inet_connection_sock.c

@@ -123,15 +123,13 @@ again:
 						smallest_size = tb->num_owners;
 						smallest_rover = rover;
 						if (atomic_read(&hashinfo->bsockets) > (high - low) + 1) {
-							spin_unlock(&head->lock);
 							snum = smallest_rover;
-							goto have_snum;
+							goto tb_found;
 						}
 					}
 					if (!inet_csk(sk)->icsk_af_ops->bind_conflict(sk, tb)) {
-						spin_unlock(&head->lock);
 						snum = rover;
-						goto have_snum;
+						goto tb_found;
 					}
 					goto next;
 				}

-- 
1.7.7.1

On 01/25/2012 04:34 PM, Flavio Leitner wrote:

The current code checks for conflicts when the application
requests a specific port.  If there is no conflict, then
the request is granted.

On the other hand, the port autoselection done by the kernel
fails when all ports are bound even when there is a port
with no conflict available.

The fix changes port autoselection to check if there is a
conflict and use it if not.

Signed-off-by: Flavio Leitner<redacted>

Signed-off-by: Marcelo Ricardo Leitner <redacted>

On 01/25/2012 04:34 PM, Flavio Leitner wrote:

Port autoselection finds a port and then drop the lock,
then right after that, gets the hash bucket again and lock it.

Fix it to go direct.

Signed-off-by: Flavio Leitner<redacted>

Signed-off-by: Marcelo Ricardo Leitner <redacted>

Le mercredi 25 janvier 2012 à 16:34 -0200, Flavio Leitner a écrit :

Acked-by: Eric Dumazet <redacted>

Le mercredi 25 janvier 2012 à 16:34 -0200, Flavio Leitner a écrit :

Port autoselection finds a port and then drop the lock,
then right after that, gets the hash bucket again and lock it.

Fix it to go direct.

Signed-off-by: Flavio Leitner <redacted>
---

Acked-by: Eric Dumazet <redacted>

From: Flavio Leitner <redacted>
Date: Wed, 25 Jan 2012 16:34:51 -0200

The current code checks for conflicts when the application
requests a specific port.  If there is no conflict, then
the request is granted.

On the other hand, the port autoselection done by the kernel
fails when all ports are bound even when there is a port
with no conflict available.

The fix changes port autoselection to check if there is a
conflict and use it if not.

Signed-off-by: Flavio Leitner <redacted>

Applied.

From: Flavio Leitner <redacted>
Date: Wed, 25 Jan 2012 16:34:52 -0200

Port autoselection finds a port and then drop the lock,
then right after that, gets the hash bucket again and lock it.

Fix it to go direct.

Signed-off-by: Flavio Leitner <redacted>

Applied, great work Flavio.