modstat: allow compat support to fail in tests

[brooks]

Summary

v1 and v2 versions of stuct module_stat are now gated by COMPAT_ ifdefs
so allow the otherwise valid query to fail with EINVAL.

Effort: CHERI upstreaming
Reported by: llm-postcr
Sponsored by: Innovate UK
Fixes: 1ebf46ecb627 ("modstat: don't copyout while holding a lock")

[brooks]

llm-postcr correctly reported that the v1 tests would now fail after 1ebf46ecb627 (thought it didn't realize they would pass on armv7 and powerpc). It also didn't report that v2 tests would fail on a kernel without COMPAT_FREEBSD13. I've update both to allow support to not be present.

I'm somewhat amused by llm-postcr's quibbling about the version gate for v1.

The gate also doesn't match where v1 came from. v1 was the only layout until           
commit a35261ef9153 (Jan 1999), so it is the FreeBSD 3.0 ELF ABI, not just             
a.out. In practice that probably doesn't matter, but the commit message's a.out        
justification isn't quite accurate.

It is true that ELF binaries that only supported v1 could exist, but 3.0.0 contains the compat code so no release binaries exist.

[markj]

tests/sys/kern/module_test.c

148

AFAICS there is no FEATURE macro to advertise COMPAT_43 support, unlike the FEATURE(compat_freebsd...) we have in kern_mib.c. If we had that, we could check for it here and get a more precise test. I'm not sure it's worth the effort.

[brooks]

Make it an error if modstat fails with a v2 struct when compat_freebsd13 is present