This was split out of bytecodealliance#10960 after some CI issues. The simplification of
s390x is split to bytecodealliance#10973 to show its failure but otherwise these changes
aren't expected to cause any issues (yet). Part of enabling this was in bytecodealliance#10974
where ASAN showed some issues as well.

prtest:full

This was split out of #10960 after some CI issues. The simplification of
s390x is split to #10973 to show its failure but otherwise these changes
aren't expected to cause any issues (yet). Part of enabling this was in #10974
where ASAN showed some issues as well.

prtest:full

* Fix ASAN in combination with throw intrinsic

Prior to this commit locally I would see a nondeterministic crash when
executing:

    RUSTFLAGS=-Zsanitizer=address \
      cargo +nightly run --release test filetests

The crash has a scary-looking internal assertion within the ASAN runtime
itself and local debugging seems to show that threads are required to
reproduce this and it's nondeterministic. I don't know why CI is
currently passing on all PRs when this is failing so frequently for
myself locally.

In any case it seems like there's a magical ASAN intrinsic to call
before `noreturn` things which I think is intended for thread-stopping
events which isn't a great fit for exceptions here but it seems to
resolve the crash.

Basically I have no idea why things crash locally, what this new
function does, nor why it fixes things. I am but a slave to the CI gods
in an attempt to make it so our CI is more reliable.

* Fix compile warning

This was split out of bytecodealliance#10960 after some CI issues. The simplification of
s390x is split to bytecodealliance#10973 to show its failure but otherwise these changes
aren't expected to cause any issues (yet). Part of enabling this was in bytecodealliance#10974
where ASAN showed some issues as well.

prtest:full