[net-next 1/2] tipc: update cluster capabilities if node deleted

[net-next 1/2] tipc: update cluster capabilities if node deleted

[Hoang Le]

There are two improvements when re-calculate cluster capabilities:

- When deleting a specific down node, need to re-calculate.
- In tipc_node_cleanup(), do not need to re-calculate if node
is still existing in cluster.

Acked-by: Jon Maloy <jon.maloy@ericsson.com>
Signed-off-by: Hoang Le <hoang.h.le@dektech.com.au>
---
 net/tipc/node.c | 12 +++++++++++-
 1 file changed, 11 insertions(+), 1 deletion(-)

diff --git a/net/tipc/node.c b/net/tipc/node.c
index 742c04756d72..a20fabd09e7e 100644
--- a/net/tipc/node.c
+++ b/net/tipc/node.c
@@ -665,6 +665,11 @@ static bool tipc_node_cleanup(struct tipc_node *peer)
 	}
 	tipc_node_write_unlock(peer);
 
+	if (!deleted) {
+		spin_unlock_bh(&tn->node_list_lock);
+		return deleted;
+	}
+
 	/* Calculate cluster capabilities */
 	tn->capabilities = TIPC_NODE_CAPABILITIES;
 	list_for_each_entry_rcu(temp_node, &tn->node_list, list) {
@@ -2041,7 +2046,7 @@ int tipc_nl_peer_rm(struct sk_buff *skb, struct genl_info *info)
 	struct net *net = sock_net(skb->sk);
 	struct tipc_net *tn = net_generic(net, tipc_net_id);
 	struct nlattr *attrs[TIPC_NLA_NET_MAX + 1];
-	struct tipc_node *peer;
+	struct tipc_node *peer, *temp_node;
 	u32 addr;
 	int err;
 
@@ -2082,6 +2087,11 @@ int tipc_nl_peer_rm(struct sk_buff *skb, struct genl_info *info)
 	tipc_node_write_unlock(peer);
 	tipc_node_delete(peer);
 
+	/* Calculate cluster capabilities */
+	tn->capabilities = TIPC_NODE_CAPABILITIES;
+	list_for_each_entry_rcu(temp_node, &tn->node_list, list) {
+		tn->capabilities &= temp_node->capabilities;
+	}
 	err = 0;
 err_out:
 	tipc_node_put(peer);
-- 
2.20.1

[net-next 2/2] tipc: reduce sensitive to retransmit failures

[Hoang Le]

With huge cluster (e.g >200nodes), the amount of that flow:
gap -> retransmit packet -> acked will take time in case of STATE_MSG
dropped/delayed because a lot of traffic. This lead to 1.5 sec tolerance
value criteria made link easy failure around 2nd, 3rd of failed
retransmission attempts.

Instead of re-introduced criteria of 99 faled retransmissions to fix the
issue, we increase failure detection timer to ten times tolerance value.

Fixes: 77cf8edbc0e7 ("tipc: simplify stale link failure criteria")
Acked-by: Jon Maloy <jon.maloy@ericsson.com>
Signed-off-by: Hoang Le <hoang.h.le@dektech.com.au>
---
 net/tipc/link.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/net/tipc/link.c b/net/tipc/link.c
index 038861bad72b..2aed7a958a8c 100644
--- a/net/tipc/link.c
+++ b/net/tipc/link.c
@@ -1087,7 +1087,7 @@ static bool link_retransmit_failure(struct tipc_link *l, struct tipc_link *r,
 		return false;
 
 	if (!time_after(jiffies, TIPC_SKB_CB(skb)->retr_stamp +
-			msecs_to_jiffies(r->tolerance)))
+			msecs_to_jiffies(r->tolerance * 10)))
 		return false;
 
 	hdr = buf_msg(skb);
-- 
2.20.1

RE: [net-next 1/2] tipc: update cluster capabilities if node deleted

[Jon Maloy]

Acked-by: Jon

> -----Original Message-----
> From: Hoang Le <hoang.h.le@dektech.com.au>
> Sent: 6-Nov-19 01:26
> To: Jon Maloy <jon.maloy@ericsson.com>; maloy@donjonn.com; netdev@vger.kernel.org; tipc-
> discussion@lists.sourceforge.net
> Subject: [net-next 1/2] tipc: update cluster capabilities if node deleted
> 
> There are two improvements when re-calculate cluster capabilities:
> 
> - When deleting a specific down node, need to re-calculate.
> - In tipc_node_cleanup(), do not need to re-calculate if node
> is still existing in cluster.
> 
> Acked-by: Jon Maloy <jon.maloy@ericsson.com>
> Signed-off-by: Hoang Le <hoang.h.le@dektech.com.au>
> ---
>  net/tipc/node.c | 12 +++++++++++-
>  1 file changed, 11 insertions(+), 1 deletion(-)
> 
> diff --git a/net/tipc/node.c b/net/tipc/node.c
> index 742c04756d72..a20fabd09e7e 100644
> --- a/net/tipc/node.c
> +++ b/net/tipc/node.c
> @@ -665,6 +665,11 @@ static bool tipc_node_cleanup(struct tipc_node *peer)
>  	}
>  	tipc_node_write_unlock(peer);
> 
> +	if (!deleted) {
> +		spin_unlock_bh(&tn->node_list_lock);
> +		return deleted;
> +	}
> +
>  	/* Calculate cluster capabilities */
>  	tn->capabilities = TIPC_NODE_CAPABILITIES;
>  	list_for_each_entry_rcu(temp_node, &tn->node_list, list) {
> @@ -2041,7 +2046,7 @@ int tipc_nl_peer_rm(struct sk_buff *skb, struct genl_info *info)
>  	struct net *net = sock_net(skb->sk);
>  	struct tipc_net *tn = net_generic(net, tipc_net_id);
>  	struct nlattr *attrs[TIPC_NLA_NET_MAX + 1];
> -	struct tipc_node *peer;
> +	struct tipc_node *peer, *temp_node;
>  	u32 addr;
>  	int err;
> 
> @@ -2082,6 +2087,11 @@ int tipc_nl_peer_rm(struct sk_buff *skb, struct genl_info *info)
>  	tipc_node_write_unlock(peer);
>  	tipc_node_delete(peer);
> 
> +	/* Calculate cluster capabilities */
> +	tn->capabilities = TIPC_NODE_CAPABILITIES;
> +	list_for_each_entry_rcu(temp_node, &tn->node_list, list) {
> +		tn->capabilities &= temp_node->capabilities;
> +	}
>  	err = 0;
>  err_out:
>  	tipc_node_put(peer);
> --
> 2.20.1

RE: [net-next 2/2] tipc: reduce sensitive to retransmit failures

[Jon Maloy]

Acked-by: Jon

> -----Original Message-----
> From: Hoang Le <hoang.h.le@dektech.com.au>
> Sent: 6-Nov-19 01:26
> To: Jon Maloy <jon.maloy@ericsson.com>; maloy@donjonn.com; netdev@vger.kernel.org; tipc-
> discussion@lists.sourceforge.net
> Subject: [net-next 2/2] tipc: reduce sensitive to retransmit failures
> 
> With huge cluster (e.g >200nodes), the amount of that flow:
> gap -> retransmit packet -> acked will take time in case of STATE_MSG
> dropped/delayed because a lot of traffic. This lead to 1.5 sec tolerance
> value criteria made link easy failure around 2nd, 3rd of failed
> retransmission attempts.
> 
> Instead of re-introduced criteria of 99 faled retransmissions to fix the
> issue, we increase failure detection timer to ten times tolerance value.
> 
> Fixes: 77cf8edbc0e7 ("tipc: simplify stale link failure criteria")
> Acked-by: Jon Maloy <jon.maloy@ericsson.com>
> Signed-off-by: Hoang Le <hoang.h.le@dektech.com.au>
> ---
>  net/tipc/link.c | 2 +-
>  1 file changed, 1 insertion(+), 1 deletion(-)
> 
> diff --git a/net/tipc/link.c b/net/tipc/link.c
> index 038861bad72b..2aed7a958a8c 100644
> --- a/net/tipc/link.c
> +++ b/net/tipc/link.c
> @@ -1087,7 +1087,7 @@ static bool link_retransmit_failure(struct tipc_link *l, struct tipc_link *r,
>  		return false;
> 
>  	if (!time_after(jiffies, TIPC_SKB_CB(skb)->retr_stamp +
> -			msecs_to_jiffies(r->tolerance)))
> +			msecs_to_jiffies(r->tolerance * 10)))
>  		return false;
> 
>  	hdr = buf_msg(skb);
> --
> 2.20.1

Re: [net-next 1/2] tipc: update cluster capabilities if node deleted

[David Miller]

From: Hoang Le <hoang.h.le@dektech.com.au>
Date: Wed,  6 Nov 2019 13:26:09 +0700

> There are two improvements when re-calculate cluster capabilities:
> 
> - When deleting a specific down node, need to re-calculate.
> - In tipc_node_cleanup(), do not need to re-calculate if node
> is still existing in cluster.
> 
> Acked-by: Jon Maloy <jon.maloy@ericsson.com>
> Signed-off-by: Hoang Le <hoang.h.le@dektech.com.au>

Applied.

Re: [net-next 2/2] tipc: reduce sensitive to retransmit failures

[David Miller]

From: Hoang Le <hoang.h.le@dektech.com.au>
Date: Wed,  6 Nov 2019 13:26:10 +0700

> With huge cluster (e.g >200nodes), the amount of that flow:
> gap -> retransmit packet -> acked will take time in case of STATE_MSG
> dropped/delayed because a lot of traffic. This lead to 1.5 sec tolerance
> value criteria made link easy failure around 2nd, 3rd of failed
> retransmission attempts.
> 
> Instead of re-introduced criteria of 99 faled retransmissions to fix the
> issue, we increase failure detection timer to ten times tolerance value.
> 
> Fixes: 77cf8edbc0e7 ("tipc: simplify stale link failure criteria")
> Acked-by: Jon Maloy <jon.maloy@ericsson.com>
> Signed-off-by: Hoang Le <hoang.h.le@dektech.com.au>

Applied.