[PATCH net-next] Revert "net-loopback: set lo dev initial state to UP"

Subsystems: networking drivers, the rest

STALE2002d

2 messages, 2 authors, 2021-02-09 · open the first message on its own page

[PATCH net-next] Revert "net-loopback: set lo dev initial state to UP"

From: Petr Machata <petrm@nvidia.com>
Date: 2021-02-09 18:05:10

In commit c9dca822c729 ("net-loopback: set lo dev initial state to UP"),
linux started automatically bringing up the loopback device of a newly
created namespace. However, an existing user script might reasonably have
the following stanza when creating a new namespace -- and in fact at least
tools/testing/selftests/net/fib_nexthops.sh in Linux's very own testsuite
does:

 # set -e
 # ip netns add foo
 # ip -netns foo addr add 127.0.0.1/8 dev lo
 # ip -netns foo link set lo up
 # set +e

This will now fail, because the kernel reasonably rejects "ip addr add" of
a duplicate address. The described change of behavior therefore constitutes
a breakage. Revert it.

Fixes: c9dca822c729 ("net-loopback: set lo dev initial state to UP")
Signed-off-by: Petr Machata <petrm@nvidia.com>
---
 drivers/net/loopback.c | 6 ------
 1 file changed, 6 deletions(-)
diff --git a/drivers/net/loopback.c b/drivers/net/loopback.c
index 24487ec17f8b..a1c77cc00416 100644
--- a/drivers/net/loopback.c
+++ b/drivers/net/loopback.c
@@ -219,12 +219,6 @@ static __net_init int loopback_net_init(struct net *net)
 
 	BUG_ON(dev->ifindex != LOOPBACK_IFINDEX);
 	net->loopback_dev = dev;
-
-	/* bring loopback device UP */
-	rtnl_lock();
-	dev_open(dev, NULL);
-	rtnl_unlock();
-
 	return 0;
 
 out_free_netdev:
-- 
2.26.2

Re: [PATCH net-next] Revert "net-loopback: set lo dev initial state to UP"

From: Jakub Kicinski <kuba@kernel.org>
Date: 2021-02-09 18:49:37

On Tue, 9 Feb 2021 18:52:04 +0100 Petr Machata wrote:
In commit c9dca822c729 ("net-loopback: set lo dev initial state to UP"),
linux started automatically bringing up the loopback device of a newly
created namespace. However, an existing user script might reasonably have
the following stanza when creating a new namespace -- and in fact at least
tools/testing/selftests/net/fib_nexthops.sh in Linux's very own testsuite
does:

 # set -e
 # ip netns add foo
 # ip -netns foo addr add 127.0.0.1/8 dev lo
 # ip -netns foo link set lo up
 # set +e

This will now fail, because the kernel reasonably rejects "ip addr add" of
a duplicate address. The described change of behavior therefore constitutes
a breakage. Revert it.

Fixes: c9dca822c729 ("net-loopback: set lo dev initial state to UP")
Signed-off-by: Petr Machata <petrm@nvidia.com>
Reviewed-by: Jakub Kicinski <kuba@kernel.org>

Thanks!
Keyboard shortcuts
hback out one level
jnext message in thread
kprevious message in thread
ldrill in
Escclose help / fold thread tree
?toggle this help