On 7/19/23 14:59, Ido Schimmel wrote:
On Tue, Jul 18, 2023 at 08:39:33PM +0200, Mirsad Todorovac wrote:
There is also a gotcha here: you do not delete all veths:
root@defiant:# ip link show
1: lo: <LOOPBACK,UP,LOWER_UP> mtu 65536 qdisc noqueue state UNKNOWN mode DEFAULT group default qlen 1000
link/loopback 00:00:00:00:00:00 brd 00:00:00:00:00:00
2: enp16s0: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc fq_codel state UP mode DEFAULT group default qlen 1000
link/ether 9c:6b:00:01:fb:80 brd ff:ff:ff:ff:ff:ff
root@defiant:# ./bridge_igmp.sh
[...]
root@defiant:# ip link show
1: lo: <LOOPBACK,UP,LOWER_UP> mtu 65536 qdisc noqueue state UNKNOWN mode DEFAULT group default qlen 1000
link/loopback 00:00:00:00:00:00 brd 00:00:00:00:00:00
2: enp16s0: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc fq_codel state UP mode DEFAULT group default qlen 1000
link/ether 9c:6b:00:01:fb:80 brd ff:ff:ff:ff:ff:ff
3: veth1@veth0: <BROADCAST,MULTICAST,M-DOWN> mtu 1500 qdisc noqueue state DOWN mode DEFAULT group default qlen 1000
link/ether b6:46:e6:4c:e4:00 brd ff:ff:ff:ff:ff:ff
4: veth0@veth1: <BROADCAST,MULTICAST,M-DOWN> mtu 1500 qdisc noqueue state DOWN mode DEFAULT group default qlen 1000
link/ether 2e:ff:7f:8a:6b:d4 brd ff:ff:ff:ff:ff:ff
5: veth3@veth2: <BROADCAST,MULTICAST,M-DOWN> mtu 1500 qdisc noqueue state DOWN mode DEFAULT group default qlen 1000
link/ether ba:33:37:81:dc:5b brd ff:ff:ff:ff:ff:ff
6: veth2@veth3: <BROADCAST,MULTICAST,M-DOWN> mtu 1500 qdisc noqueue state DOWN mode DEFAULT group default qlen 1000
link/ether f2:fd:0a:9b:94:17 brd ff:ff:ff:ff:ff:ff
root@defiant:#
These tests can be run with veth pairs or with loop backed physical
ports. We can't delete the latter and I don't see a clean way to delete
the veth pairs.
The following patch [1] changes the default to not create interfaces so
that by default these tests will be skipped [2]. Those who care about
running the tests can create a forwarding.config file (using
forwarding.config.sample as an example) and either create the veth pairs
themselves or opt-in for the interfaces to be created automatically
(setting NETIF_CREATE=yes), but not deleted.
[1]
diff --git a/tools/testing/selftests/net/forwarding/forwarding.config.sample b/tools/testing/selftests/net/forwarding/forwarding.config.sample
index 4a546509de90..b72f08dfd491 100644
--- a/tools/testing/selftests/net/forwarding/forwarding.config.sample
+++ b/tools/testing/selftests/net/forwarding/forwarding.config.sample
@@ -36,8 +36,9 @@ PAUSE_ON_FAIL=no
PAUSE_ON_CLEANUP=no
# Type of network interface to create
NETIF_TYPE=veth
-# Whether to create virtual interfaces (veth) or not
-NETIF_CREATE=yes
+# Whether to create virtual interfaces (veth) or not. Created interfaces are
+# not automatically deleted
+NETIF_CREATE=no
# Timeout (in seconds) before ping exits regardless of how many packets have
# been sent or received
PING_TIMEOUT=5
diff --git a/tools/testing/selftests/net/forwarding/lib.sh b/tools/testing/selftests/net/forwarding/lib.sh
index 9ddb68dd6a08..1b1bc634c63e 100755
--- a/tools/testing/selftests/net/forwarding/lib.sh
+++ b/tools/testing/selftests/net/forwarding/lib.sh
@@ -17,7 +17,7 @@ WAIT_TIME=${WAIT_TIME:=5}
PAUSE_ON_FAIL=${PAUSE_ON_FAIL:=no}
PAUSE_ON_CLEANUP=${PAUSE_ON_CLEANUP:=no}
NETIF_TYPE=${NETIF_TYPE:=veth}
-NETIF_CREATE=${NETIF_CREATE:=yes}
+NETIF_CREATE=${NETIF_CREATE:=no}
MCD=${MCD:=smcrouted}
MC_CLI=${MC_CLI:=smcroutectl}
PING_COUNT=${PING_COUNT:=10}
[2]
# ./bridge_igmp.sh
SKIP: could not find all required interfaces
Hi, Ido,
Please would you consider making a formal patch, so the vanilla 'make kselftest'
run would have some sensible defaults?
I am really interested in the network stack working right, though I really do
not have much opportunities testing like high speed links and bonded eth adapters :-(
Maybe that'll come :-/
Kind regards,
Mirsad