-
Notifications
You must be signed in to change notification settings - Fork 6
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
tests: also run Join tests with Checksum support #1
base: latest
Are you sure you want to change the base?
Conversation
entrypoint.sh
Outdated
@@ -451,6 +452,10 @@ run_mptcp_connect_mmap() { | |||
_run_selftest_one_tap "${mptcp_connect_mmap_tap}" ./mptcp_connect.sh -m mmap | |||
} | |||
|
|||
run_mptcp_join_csum() { | |||
_run_selftest_one_tap "${mptcp_join_csum_tap}" ./mptcp_join.sh -C |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
_run_selftest_one_tap "${mptcp_join_csum_tap}" ./mptcp_join.sh -C | |
_run_selftest_one_tap "${mptcp_join_csum_tap}" ./mptcp_join.sh -fsabpkC |
This runs in about 1/4 the time on my system.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
@mjmartineau Thank you for the suggestion! (and sorry for the delay!)
That seems to add 3-4 minutes on a slow environment.
But what worry me is more to maintain a list of acceptable options. I mean: if we modify mptcp_join.sh
and maintain this list of tests there, that seems to be a more valuable solution (or better: some tests are ran twice in a normal situation: with an without checksum).
Or maybe enough to run mptcp_connect.sh a second time with checksum support? That seems to cover quite a few scenarios. For the ones with multiple paths, we have mptcp_join.sh that already cover some tests with checksum enabled.
In other words: drop this commit and only take yours?
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
I think it's important to get both the mptcp_connect.sh
coverage in #2 and (at least) some of the multiple-subflow cases in mptcp_join.sh
. Modifying mptcp_join.sh
would be fine with me, that way the checksum code would get exercised more by individual developers and other CI systems like kbuild.
Much like multipath-tcp#1, run the connect tests an extra time with the -C flag in order to exercise the checksum code in more scenarios. Signed-off-by: Mat Martineau <[email protected]>
d022f0b
to
ca3335b
Compare
To add more coverage. We already have some in mptcp_join.sh. Signed-off-by: Matthieu Baerts <[email protected]>
ca3335b
to
768b427
Compare
@mjmartineau with some delay, I'm reviving this PR: do you think we still need more tests to run with csum enabled? Now that tests are running on 4 instances, we have a bit more time. Sadly, not enough to run all But do you think it makes sense to run |
This needs to be covered by the CI.
Only downside: it takes ~25 minutes with the public CI and a debug
kernel.
cc: @mjmartineau