selftests: forwarding: Two small changes
Subsystems (4)
cover selftests: forwarding: Two small changes
Two unrelated changes in this patchset:
- In patch #1, convert mirror tests from using ping directly to generating ICMP packets by mausezahn. Using ping in tests is error-prone, because ping is too smart. On a flaky system (notably in a simulator), when packets don't come quickly enough, more pings are sent, and that throws off counters. This was worked around in the past by just pinging more slowly, but using mausezahn avoids the issue as well without making the tests unnecessary slow.
- A missing stats_update callback was recently added to act_pedit. Now that iproute2 supports JSON dumping for pedit, extend in patch #2 the pedit_dsfield selftest with a check that would have caught the fact that the callback was missing.
Petr Machata (2):
selftests: forwarding: mirror_lib: Use mausezahn
selftests: forwarding: pedit_dsfield: Check counter value
tools/testing/selftests/net/forwarding/mirror_lib.sh | 6 ++---- tools/testing/selftests/net/forwarding/pedit_dsfield.sh | 7 ++++++- 2 files changed, 8 insertions(+), 5 deletions(-)
Signature · 1 line
2.20.1
PATCH 1/2 selftests: forwarding: mirror_lib: Use mausezahn
Using ping in tests is error-prone, because ping is too smart. On a flaky system (notably in a simulator), when packets don't come quickly enough, more pings are sent, and that throws off counters. Instead use mausezahn to generate ICMP echo request packets. That allows us to send them in quicker succession as well, because the reason the ping was made slow in the first place was to make the tests work on simulated systems.
--- tools/testing/selftests/net/forwarding/mirror_lib.sh | 6 ++---- 1 file changed, 2 insertions(+), 4 deletions(-)
tools/testing/selftests/net/forwarding/mirror_lib.sh +2 -4
@@ -29,11 +29,9 @@ mirror_test() local pref=$1; shift local expect=$1; shift - local ping_timeout=$((PING_TIMEOUT * 5)) local t0=$(tc_rule_stats_get $dev $pref) - ip vrf exec $vrf_name \ - ${PING} ${sip:+-I $sip} $dip -c 10 -i 0.5 -w $ping_timeout \ - &> /dev/null + $MZ $vrf_name ${sip:+-A $sip} -B $dip -a own -b bc -q \ + -c 10 -d 100ms -t icmp type=8 sleep 0.5 local t1=$(tc_rule_stats_get $dev $pref) local delta=$((t1 - t0))
Signature · 1 line
2.20.1
PATCH 2/2 selftests: forwarding: pedit_dsfield: Check counter value
In reply to Petr MachataA missing stats_update callback was recently added to act_pedit. Now that iproute2 supports JSON dumping for pedit, extend the pedit_dsfield selftest with a check that would have caught the fact that the callback was missing.
--- tools/testing/selftests/net/forwarding/pedit_dsfield.sh | 7 ++++++- 1 file changed, 6 insertions(+), 1 deletion(-)
tools/testing/selftests/net/forwarding/pedit_dsfield.sh +6 -1
@@ -132,7 +132,12 @@ do_test_pedit_dsfield_common() local pkts pkts=$(busywait "$TC_HIT_TIMEOUT" until_counter_is ">= 10" \ tc_rule_handle_stats_get "dev $h2 ingress" 101) - check_err $? "Expected to get 10 packets, but got $pkts." + check_err $? "Expected to get 10 packets on test probe, but got $pkts." + + pkts=$(tc_rule_handle_stats_get "$pedit_locus" 101) + ((pkts >= 10)) + check_err $? "Expected to get 10 packets on pedit rule, but got $pkts." + log_test "$pedit_locus pedit $pedit_action" }
Signature · 1 line
2.20.1
David Miller
maintainerFrom: Petr Machata <[email protected]>
Date: Fri, 29 May 2020 14:16:52 +0300
Quoted 10 lines
Two unrelated changes in this patchset:
- In patch #1, convert mirror tests from using ping directly to generating ICMP packets by mausezahn. Using ping in tests is error-prone, because ping is too smart. On a flaky system (notably in a simulator), when packets don't come quickly enough, more pings are sent, and that throws off counters. This was worked around in the past by just pinging more slowly, but using mausezahn avoids the issue as well without making the tests unnecessary slow.
Quoted 4 lines
- A missing
stats_updatecallback was recently added toact_pedit. Now that iproute2 supports JSON dumping for pedit, extend in patch #2 thepedit_dsfieldselftest with a check that would have caught the fact that the callback was missing.
Series applied, thanks.