On Thu, Jul 01, 2021 at 08:46:46AM +0300, Ido Schimmel wrote: > On Wed, Jun 30, 2021 at 02:51:38PM +0200, Guillaume Nault wrote: > > diff --git a/tools/testing/selftests/net/forwarding/topo_nschain_lib.sh b/tools/testing/selftests/net/forwarding/topo_nschain_lib.sh > > new file mode 100644 > > index 000000000000..4c0bf2d7328a > > --- /dev/null > > +++ b/tools/testing/selftests/net/forwarding/topo_nschain_lib.sh > > @@ -0,0 +1,267 @@ > > +#!/bin/bash > > +# SPDX-License-Identifier: GPL-2.0 > > + > > +# A chain of 4 nodes connected with veth pairs. > > +# Each node lives in its own network namespace. > > Hi, > > The tests under tools/testing/selftests/net/forwarding/ are meant to use > VRFs as lightweight namespaces. This allows us to run the tests on both > physical switches with loopback cables and veth pairs, thereby > validating both the hardware and software datapaths. > > See tools/testing/selftests/net/forwarding/README It wasn't clear to me that net/forwarding was _only_ for VRF-based tests. > If the tests cannot be converted to VRFs, then I suggest moving them to > tools/testing/selftests/net/ I first tried to write this selftest using VRFs, but there were some problems that made me switch to namespaces (I don't remember precisely which ones, probably virtual tunnel devices in collect_md mode). I'll give VRF another try. And if I can't get to a satisfying solution, I'll move the selftest under net/.
