From: Felix Fietkau <nbd@openwrt.org>
Date: Tue, 17 Sep 2013 10:07:52 +0000 (+0000)
Subject: kernel: backport a fix for multiqueue selection to prevent tcp reordering with fair... 
X-Git-Tag: reboot~9281
X-Git-Url: http://git.lede-project.org./?a=commitdiff_plain;h=9158a089836052f275342119b9e2a91de4335dbb;p=openwrt%2Fstaging%2Fnoltari.git

kernel: backport a fix for multiqueue selection to prevent tcp reordering with fair queueing

Signed-off-by: Felix Fietkau <nbd@openwrt.org>

SVN-Revision: 38016
---

diff --git a/target/linux/generic/patches-3.10/110-net_fix_multiqueue_selection.patch b/target/linux/generic/patches-3.10/110-net_fix_multiqueue_selection.patch
new file mode 100644
index 0000000000..6a07ee8c0d
--- /dev/null
+++ b/target/linux/generic/patches-3.10/110-net_fix_multiqueue_selection.patch
@@ -0,0 +1,41 @@
+From: Eric Dumazet <edumazet@google.com>
+
+commit 416186fbf8c5b4e4465 ("net: Split core bits of netdev_pick_tx
+into __netdev_pick_tx") added a bug that disables caching of queue
+index in the socket.
+
+This is the source of packet reorders for TCP flows, and
+again this is happening more often when using FQ pacing.
+
+Old code was doing
+
+if (queue_index != old_index)
+	sk_tx_queue_set(sk, queue_index);
+
+Alexander renamed the variables but forgot to change sk_tx_queue_set()
+2nd parameter.
+
+if (queue_index != new_index)
+	sk_tx_queue_set(sk, queue_index);
+
+This means we store -1 over and over in sk->sk_tx_queue_mapping
+
+Signed-off-by: Eric Dumazet <edumazet@google.com>
+Cc: Alexander Duyck <alexander.h.duyck@intel.com>
+Acked-by: Alexander Duyck <alexander.h.duyck@intel.com>
+
+---
+net/core/flow_dissector.c |    2 +-
+ 1 file changed, 1 insertion(+), 1 deletion(-)
+
+--- a/net/core/flow_dissector.c
++++ b/net/core/flow_dissector.c
+@@ -347,7 +347,7 @@ u16 __netdev_pick_tx(struct net_device *
+ 
+ 		if (queue_index != new_index && sk &&
+ 		    rcu_access_pointer(sk->sk_dst_cache))
+-			sk_tx_queue_set(sk, queue_index);
++			sk_tx_queue_set(sk, new_index);
+ 
+ 		queue_index = new_index;
+ 	}