[<prev] [next>] [thread-next>] [day] [month] [year] [list]
Message-ID: <7e30b8e01fdfb76ff6f4e071348829e2f56767e8.1632477544.git.petrm@nvidia.com>
Date: Fri, 24 Sep 2021 12:04:27 +0200
From: Petr Machata <petrm@...dia.com>
To: <netdev@...r.kernel.org>
CC: "David S . Miller" <davem@...emloft.net>,
Jakub Kicinski <kuba@...nel.org>,
Petr Machata <petrm@...dia.com>
Subject: [PATCH net-next] selftests: net: fib_nexthops: Wait before checking reported idle time
The purpose of this test is to verify that after a short activity passes,
the reported time is reasonable: not zero (which could be reported by
mistake), and not something outrageous (which would be indicative of an
issue in used units).
However, the idle time is reported in units of clock_t, or hundredths of
second. If the initial sequence of commands is very quick, it is possible
that the idle time is reported as just flat-out zero. When this test was
recently enabled in our nightly regression, we started seeing spurious
failures for exactly this reason.
Therefore buffer the delay leading up to the test with a sleep, to make
sure there is no legitimate way of reporting 0.
Signed-off-by: Petr Machata <petrm@...dia.com>
---
tools/testing/selftests/net/fib_nexthops.sh | 1 +
1 file changed, 1 insertion(+)
diff --git a/tools/testing/selftests/net/fib_nexthops.sh b/tools/testing/selftests/net/fib_nexthops.sh
index 0d293391e9a4..b5a69ad191b0 100755
--- a/tools/testing/selftests/net/fib_nexthops.sh
+++ b/tools/testing/selftests/net/fib_nexthops.sh
@@ -2078,6 +2078,7 @@ basic_res()
"id 101 index 0 nhid 2 id 101 index 1 nhid 2 id 101 index 2 nhid 1 id 101 index 3 nhid 1"
log_test $? 0 "Dump all nexthop buckets in a group"
+ sleep 0.1
(( $($IP -j nexthop bucket list id 101 |
jq '[.[] | select(.bucket.idle_time > 0 and
.bucket.idle_time < 2)] | length') == 4 ))
--
2.31.1
Powered by blists - more mailing lists