[NinoFloris]

Fixes #136936, a longstanding bug that was kept around as a quirk for zero-diffs at the time. Disasm tests here are a bit fragile, though arm has stabilized its instructions around this now, but I'd rather have them than not.

[azure-pipelines]

Azure Pipelines:
Successfully started running 5 pipeline(s).
10 pipeline(s) were filtered out due to trigger conditions.
There may be pipelines that require an authorized user to comment /azp run to run.

[dotnet-policy-service]

Tagging subscribers to this area: @JulieLeeMSFT, @jakobbotsch
See info in area-owners.md if you want to be subscribed.

[jakobbotsch]

Disasm tests here are a bit fragile, though arm has stabilized its instructions around this now, but I'd rather have them than not.

I don't agree. These tests are much too fragile, relying on exact promotion heuristics that are going to be changing in this release. I don't think the test is needed.

I would be ok with outerloop test that validates that a volatile load was not dropped by testing for a race.

[jakobbotsch]

a longstanding bug that was kept around as a quirk for zero-diffs at the time

FWIW, this used to have quite significant diffs around structs that were later proved to be not exposed. Looks like something changed since the diffs are very minimal now, however.