diff options
| author |   Jeff Shaw <jeffrey.b.shaw@intel.com> | 2017-10-09 10:02:44 -0700 |
|---|---|---|
| committer | Jeff Shaw <jeffrey.b.shaw@intel.com> | 2017-10-09 10:10:53 -0700 |
| commit | 50a852c19dbf40a4564d33c2878285552b593427 (patch) | |
| tree | 8c83b64522381074a66966fc277688ae3ea6ea92 | |
| parent | bea87a678f755b9d205dc95802a38ee91c0d067e (diff) | |
[router] Set VLIB_RX metadata to valid sw_if_index
In normal cases, some multicast frames from the control plane are
dropped by VPP before being sent. We were not setting the VLIB_RX
interface to a valid sw_if_index, causing a segmentation fault
when updating counters in the process_drop node.
Change-Id: I950158b6c36c9056ee024d2524099db03bffbf35
Signed-off-by: Jeff Shaw <jeffrey.b.shaw@intel.com>
| -rw-r--r-- | router/router/tap_inject_node.c | 2 |
1 files changed, 1 insertions, 1 deletions
diff --git a/router/router/tap_inject_node.c b/router/router/tap_inject_node.c index 5429846..a7e46ce 100644 --- a/router/router/tap_inject_node.c +++ b/router/router/tap_inject_node.c @@ -237,7 +237,7 @@ tap_rx (vlib_main_t * vm, vlib_node_runtime_t * node, vlib_frame_t * f, int fd) b = vlib_get_buffer (vm, bi[0]); - vnet_buffer (b)->sw_if_index[VLIB_RX] = ~0; + vnet_buffer (b)->sw_if_index[VLIB_RX] = sw_if_index; vnet_buffer (b)->sw_if_index[VLIB_TX] = sw_if_index; n_bytes_left = n_bytes - VLIB_BUFFER_DATA_SIZE; |