[<prev] [next>] [thread-next>] [day] [month] [year] [list]
Message-Id: <20210228142613.1642938-1-idosch@idosch.org>
Date: Sun, 28 Feb 2021 16:26:11 +0200
From: Ido Schimmel <idosch@...sch.org>
To: netdev@...r.kernel.org
Cc: davem@...emloft.net, kuba@...nel.org, dsahern@...il.com,
roopa@...dia.com, sharpd@...dia.com, mlxsw@...dia.com,
Ido Schimmel <idosch@...dia.com>
Subject: [RFC PATCH net 0/2] nexthop: Do not flush blackhole nexthops when loopback goes down
From: Ido Schimmel <idosch@...dia.com>
Patch #1 prevents blackhole nexthops from being flushed when the
loopback device goes down given that as far as user space is concerned,
these nexthops do not have a nexthop device.
Patch #2 adds a test case.
This is a user visible change, so sending as RFC.
Personally, I think it is worth making the change. The flow is quite
obscure and therefore unlikely to result in any regressions, especially
when the nexthop API is quite new compared to the legacy API. In
addition, the current behavior is very puzzling to those not familiar
with the inner workings of the nexthop code.
Regardless, there are no regressions in fib_nexthops.sh with this
change:
# ./fib_nexthops.sh
...
Tests passed: 165
Tests failed: 0
Ido Schimmel (2):
nexthop: Do not flush blackhole nexthops when loopback goes down
selftests: fib_nexthops: Test blackhole nexthops when loopback goes
down
net/ipv4/nexthop.c | 10 +++++++---
tools/testing/selftests/net/fib_nexthops.sh | 8 ++++++++
2 files changed, 15 insertions(+), 3 deletions(-)
--
2.29.2
Powered by blists - more mailing lists