diff options
| author | Thomas Huth <thuth@redhat.com> | 2015-11-26 23:06:21 +0100 |
|---|---|---|
| committer | Alexey Kardashevskiy <aik@ozlabs.ru> | 2015-12-01 17:02:04 +1100 |
| commit | a2e7a5c2be8cc894aa985ddb1179a06dae69a343 (patch) | |
| tree | 836951519b7b600611b40c63380143a8c3cdb667 /clients | |
| parent | b4c93802a5b2c72f096649c497ec9ff5708e4456 (diff) | |
| download | SLOF-a2e7a5c2be8cc894aa985ddb1179a06dae69a343.zip SLOF-a2e7a5c2be8cc894aa985ddb1179a06dae69a343.tar.gz SLOF-a2e7a5c2be8cc894aa985ddb1179a06dae69a343.tar.bz2 | |
Fix tracking of pending outgoing packets when handling ARP replies
pending_pkt.pkt_pending is never read, thus setting it to 0
in handle_arp() does not make much sense. On the other hand,
arp_table[i].pkt_pending is never set back to 0, but it is
checked for tracking which packets still have to be sent
in handle_arp(), thus we might send out the same queued packet
multiple times when receiving more than one ARP packet from the
other side.
So setting "pending_pkt.pkt_pending = 0" seems to be a typo, and
"arp_table[i].pkt_pending = 0" should be right instead.
Signed-off-by: Thomas Huth <thuth@redhat.com>
Reviewed-by: Nikunj A Dadhania <nikunj@linux.vnet.ibm.com>
Signed-off-by: Alexey Kardashevskiy <aik@ozlabs.ru>
Diffstat (limited to 'clients')
| -rw-r--r-- | clients/net-snk/app/netlib/ipv4.c | 2 |
1 files changed, 1 insertions, 1 deletions
diff --git a/clients/net-snk/app/netlib/ipv4.c b/clients/net-snk/app/netlib/ipv4.c index 8185de5..ee16b3d 100644 --- a/clients/net-snk/app/netlib/ipv4.c +++ b/clients/net-snk/app/netlib/ipv4.c @@ -758,7 +758,7 @@ handle_arp(int fd, uint8_t * packet, int32_t packetsize) memcpy(ethh -> dest_mac, arp_table[i].mac_addr, 6); send_ether(fd, pending_pkt.eth_frame, pending_pkt.eth_len); - pending_pkt.pkt_pending = 0; + arp_table[i].pkt_pending = 0; arp_table[i].eth_len = 0; } return 0; // no error |