Blame SOURCES/0181-tcp-ack-when-we-get-an-OOO-lost-packet.patch

f731ee
From 0000000000000000000000000000000000000000 Mon Sep 17 00:00:00 2001
f731ee
From: Josef Bacik <jbacik@fb.com>
f731ee
Date: Wed, 12 Aug 2015 08:16:22 -0700
f731ee
Subject: [PATCH] tcp: ack when we get an OOO/lost packet
f731ee
f731ee
While adding tcp window scaling support I was finding that I'd get some packet
f731ee
loss or reordering when transferring from large distances and grub would just
f731ee
timeout.  This is because we weren't ack'ing when we got our OOO packet, so the
f731ee
sender didn't know it needed to retransmit anything, so eventually it would fill
f731ee
the window and stop transmitting, and we'd time out.  Fix this by ACK'ing when
f731ee
we don't find our next sequence numbered packet.  With this fix I no longer time
f731ee
out.  Thanks,
f731ee
f731ee
Signed-off-by: Josef Bacik <jbacik@fb.com>
f731ee
---
f731ee
 grub-core/net/tcp.c | 5 ++++-
f731ee
 1 file changed, 4 insertions(+), 1 deletion(-)
f731ee
f731ee
diff --git a/grub-core/net/tcp.c b/grub-core/net/tcp.c
f731ee
index 2077f5519b2..fa29a2afc6c 100644
f731ee
--- a/grub-core/net/tcp.c
f731ee
+++ b/grub-core/net/tcp.c
f731ee
@@ -882,7 +882,10 @@ grub_net_recv_tcp_packet (struct grub_net_buff *nb,
f731ee
 	  grub_priority_queue_pop (sock->pq);
f731ee
 	}
f731ee
       if (grub_be_to_cpu32 (tcph->seqnr) != sock->their_cur_seq)
f731ee
-	return GRUB_ERR_NONE;
f731ee
+	{
f731ee
+	  ack (sock);
f731ee
+	  return GRUB_ERR_NONE;
f731ee
+	}
f731ee
       while (1)
f731ee
 	{
f731ee
 	  nb_top_p = grub_priority_queue_top (sock->pq);