[Swan-dev] test run - issues cored dumps and output changes

[Andrew Cagney]

On 28 April 2018 at 15:14, Antony Antony <antony at phenome.org> wrote:
> I have a branch ike-rekey-aa that I would like to merge soon.
> It has many output changes. Mostly changes to changes to IKEv2 tests.
> I trying to fix as much as possible in a branch before merging it to the master.
> I notice noise in master so I am not sure about merging yet.
> It would be nice to clean up the master before merging.
>
> Here are some issues in master I noticed.
>
> A. assertion/coredump
>
> 1. ikev2-algo-esp-null-02
> http://testing.libreswan.org/results/testing/v3.22-1282-g57f735e-master/ikev2-algo-esp-null-02/OUTPUT/east.console.txt

The test is WIP

The core dump is a consequence of things being impaired, and the fix
can be pushed after the next release.

> 2. seccomp-02-tolerant
> http://testing.libreswan.org/results/testing/v3.22-1282-g57f735e-master/seccomp-02-tolerant/OUTPUT/west.console.diff

I posted about this to the list.  It's a consequence of building with
electric fence.

> 3. interop-ikev2-strongswan-22-cp-responder-psk on swantest. possibly at shutdown.
> https://swantest.libreswan.fi/results/blackswan/2018-04-24-swantest-3.22-1282-g57f735e0d-master/interop-ikev2-strongswan-22-cp-responder-psk/OUTPUT/road.core.txt

Looks like yet another shutdown race?

> 4. interop-ikev2-strongswan-37-initiator-digsig on swantest.
> https://swantest.libreswan.fi/results/blackswan/2018-04-24-swantest-3.22-1282-g57f735e0d-master/interop-ikev2-strongswan-37-initiator-digsig/OUTPUT/east.console.diff
>
> 5. interop-ikev2-strongswan-38-digsig-impair on swantest
> https://swantest.libreswan.fi/results/blackswan/2018-04-24-swantest-3.22-1282-g57f735e0d-master/interop-ikev2-strongswan-38-digsig-impair/OUTPUT/east.console.diff
>
> B: mismatch tests succeeding.
> 1. ikev2-asymmetric-16-auth-mismatch
> http://testing.libreswan.org/results/testing/v3.22-1282-g57f735e-master/ikev2-asymmetric-16-auth-mismatch/OUTPUT/west.console.diff

I think you mentioned this on IRC?  I pushed a 'fix' but then, after
discussion with Paul, yanked it.  Problem, it seems, is in the test.

> 2. ikev1-x509-09-san-email-match
> http://testing.libreswan.org/results/testing/v3.22-1282-g57f735e-master/ikev1-x509-09-san-email-match/OUTPUT/west.console.diff

> there are a bunch of such changes. Looks like refrense output is stale.
> -000 TIMESTAMP, 1024 RSA Key AwXXXXXXX (has private key), until TIMESTAMP ok
> +000 TIMESTAMP, 3072 RSA Key AwXXXXXXX (has private key), until TIMESTAMP ok

Right.  Someone forgot to update the test results.  Paul?

> There are bunch of :
> "003 "westnet-eastnet-subnets/2x2" #1: Authenticated using RSA"

I guess some are still lurking, I updated most cases.

> _______________________________________________
> Swan-dev mailing list
> Swan-dev at lists.libreswan.org
> https://lists.libreswan.org/mailman/listinfo/swan-dev