On Mon, Nov 21, 2011 at 01:12:38PM -0800, Ethan Jackson wrote:
> > Another question comes to mind, though. ?Why don't the OpenFlow
> > enqueue actions generate datapath actions to set the queue id?
>
> I think it's because the relevant queue doesn't actually exist, so it
> falls back to an output
> Another question comes to mind, though. Why don't the OpenFlow
> enqueue actions generate datapath actions to set the queue id?
I think it's because the relevant queue doesn't actually exist, so it
falls back to an output action. To fix this, I would need to add
queues to the netdev-dummy modu
On Mon, Nov 21, 2011 at 12:49:51PM -0800, Ethan Jackson wrote:
> Here's an incremental.
The incremental looks OK.
Another question comes to mind, though. Why don't the OpenFlow
enqueue actions generate datapath actions to set the queue id?
___
dev mail
Here's an incremental.
---
tests/ofproto-dpif.at | 24 +---
1 files changed, 17 insertions(+), 7 deletions(-)
diff --git a/tests/ofproto-dpif.at b/tests/ofproto-dpif.at
index db2dccd..13c6fd7 100644
--- a/tests/ofproto-dpif.at
+++ b/tests/ofproto-dpif.at
@@ -64,7 +64,10 @@
On Mon, Nov 21, 2011 at 11:14:21AM -0800, Ethan Jackson wrote:
> +AT_SETUP([ofproto-dpif - output/flood flags])
> +dnl This test assumes that OpenFlow port numbers are allocated in order
> +dnl starting from one.
Do we actually reliably implement that, though? I'd expect the ports
to get added in
This patch adds basic tests to ofproto-dpif which verify that
output and flood actions respect the relevant OFPPC flags, and do
not loop back to the in_port.
---
tests/ofproto-dpif.at | 48
1 files changed, 48 insertions(+), 0 deletions(-)
diff -
