mirage-crypto is available on s390x since 0.10.3

[dinosaure]

According to #19657, this PR wants to lint mirage-crypto package with the s390x architecture (/cc @hannesm).

[hannesm]

thanks, looks good to me.

[dinosaure]

I would like to merge this PR but it seems that mirage-crypto (for older versions) has some issues - when the CI got some errors from tests in extra architectures. See specially:
https://opam.ci.ocaml.org/github/ocaml/opam-repository/commit/5a734e13b74379838c2011a6bae645cee9494552/variant/opam-2.1,extras,arm32,mirage-crypto.0.10.2,tests

It can be fixed by an other PR but we should look on it. Then, I want to have a double check from someone else to merge it.

[hannesm]

Thanks @dinosaure for looking into this. I agree this should be merged since it improves the current state. About the error, Worker stops running: Killed by signal -22, is there a way to get more output? Is there a way to figure out what "signal -22" is on arm32? Is there an easy way to reproduce? May it be some overcommit of resources on the arm32 runner?

[dinosaure]

About the error, Worker stops running: Killed by signal -22, is there a way to get more output? Is there a way to figure out what "signal -22" is on arm32? Is there an easy way to reproduce? May it be some overcommit of resources on the arm32 runner?

Currently we should be able to reproduce it with docker and qemu. Let's say that this PR does not want to solve this specific error. I will try to reproduce it as soon as I can on my side and come back with better outcomes directly on the mirage-crypto repository.

[mseri]

Let's merge in the meantime. @dinosaure can you mention the remaining "error -22" in an issue on the mirage-crypto repository so that we don't lose track of it?

[dinosaure]

@dinosaure can you mention the remaining "error -22" in an issue on the mirage-crypto repository so that we don't lose track of it?

I'm currently not able to reproduce it on armv7l but I don't have enough time to go deeper 😕