[PATCH mptcp-next v11 0/8] add helpers and vars in mptcp_lib.sh, part 3

Geliang Tang posted 8 patches 2 months ago
Patches applied successfully (tree, apply log)
git fetch https://github.com/multipath-tcp/mptcp_net-next tags/patchew/cover.1709466624.git.tanggeliang@kylinos.cn
There is a newer version of this series
tools/testing/selftests/net/mptcp/diag.sh     | 19 ++---
.../selftests/net/mptcp/mptcp_connect.sh      | 76 ++++++++++---------
.../testing/selftests/net/mptcp/mptcp_join.sh | 28 +++----
.../testing/selftests/net/mptcp/mptcp_lib.sh  | 45 ++++++++---
.../selftests/net/mptcp/mptcp_sockopt.sh      | 33 ++++----
.../testing/selftests/net/mptcp/pm_netlink.sh |  5 +-
.../selftests/net/mptcp/simult_flows.sh       | 14 ++--
.../selftests/net/mptcp/userspace_pm.sh       | 22 ++----
8 files changed, 131 insertions(+), 111 deletions(-)
[PATCH mptcp-next v11 0/8] add helpers and vars in mptcp_lib.sh, part 3
Posted by Geliang Tang 2 months ago
From: Geliang Tang <tanggeliang@kylinos.cn>

v11:
 - fix the shellcheck error in v10

v10:
 - split two more patches (#4 and #5) from #6 and #7
 - update commit logs

v9:
 - move print_larger_title from patch 5 to patch 2
 - move MPTCP_LIB_TEST_FORMAT to a more suitable location in patch 4 and
   patch 5

v8:
 - address every comment from Matt in v7.

v7:
 - address every comment from Matt in v6.

v6:
 - only include the first 7 patches in v5
 - add mptcp_lib_pr_* helpers
 - add print_title helpers
 - align the output of simult_flows.sh

v5:
 - don't use mptcp_lib_print_test_counter in mptcp_join.sh, it breaks
   skip_test().

v4:
 - set test_cnt to 0, use ++counter in mptcp_lib_print_test_counter() to
   fix the following mismatched test counters:

# 012 userspace pm server fullmesh
#       syn                                 [ OK ]
#       synack                              [ OK ]
#       ack                                 [ OK ]
#       add                                 [ OK ]
#       echo                                [FAIL] got 1 ADD_ADDR echo[s] expected 2
# Server ns stats
# TcpPassiveOpens                 5                  0.0
# TcpInSegs                       25                 0.0
... ...
#
# 1 failure(s) has(ve) been detected:
#       - 13: userspace pm server fullmesh

v3:
 - fix shellcheck errors in v2

v2:
 - fix shellcheck errors in v1
 - print test results with counters

Geliang Tang (8):
  selftests: mptcp: print all error messages to stdout
  selftests: mptcp: connect: add dedicated port counter
  selftests: mptcp: connect: fix misaligned output
  selftests: mptcp: sockopt: print every test result
  selftests: mptcp: export TEST_COUNTER variable
  selftests: mptcp: add print_title in mptcp_lib
  selftests: mptcp: print test results with counters
  selftests: mptcp: print test results with colors

 tools/testing/selftests/net/mptcp/diag.sh     | 19 ++---
 .../selftests/net/mptcp/mptcp_connect.sh      | 76 ++++++++++---------
 .../testing/selftests/net/mptcp/mptcp_join.sh | 28 +++----
 .../testing/selftests/net/mptcp/mptcp_lib.sh  | 45 ++++++++---
 .../selftests/net/mptcp/mptcp_sockopt.sh      | 33 ++++----
 .../testing/selftests/net/mptcp/pm_netlink.sh |  5 +-
 .../selftests/net/mptcp/simult_flows.sh       | 14 ++--
 .../selftests/net/mptcp/userspace_pm.sh       | 22 ++----
 8 files changed, 131 insertions(+), 111 deletions(-)

-- 
2.40.1
Re: [PATCH mptcp-next v11 0/8] add helpers and vars in mptcp_lib.sh, part 3
Posted by Matthieu Baerts 2 months ago
Hi Geliang,

On 03/03/2024 12:52, Geliang Tang wrote:
> From: Geliang Tang <tanggeliang@kylinos.cn>
> 
> v11:
>  - fix the shellcheck error in v10
> 
> v10:
>  - split two more patches (#4 and #5) from #6 and #7
>  - update commit logs
> 
> v9:
>  - move print_larger_title from patch 5 to patch 2
>  - move MPTCP_LIB_TEST_FORMAT to a more suitable location in patch 4 and
>    patch 5
> 
> v8:
>  - address every comment from Matt in v7.

If you don't mind, please add a short changelog per individual patch.
Without that, it is quite hard to understand what you modified, if there
was something you didn't do on purpose or by accident, which patch has
moved where, etc. Without that, I end up re-reviewing the whole series
like it was a new one.

Apart from that, I think we are almost there. Two details on the two
first patches, then it looks like we might print 'OK', then 'FAIL' in
patch 4. And we might print 'OK', then some 'WARN' without '\n' at the
end in patch 8.


One important question: did you check the output when you force issues?
Just to make sure they are presented correctly, and mainly if we still
catch them!

Cheers,
Matt
-- 
Sponsored by the NGI0 Core fund.