[ARM64] - IsValidCompareChain should return false if both operands are not integral types

[TIHan]

Description
Should resolve this assert: #74172

The assert was this in genCodeForConditionalCompare:

// No float support or swapping op1 and op2 to generate cmp reg, imm.assert(!varTypeIsFloating(op2Type));

Seems like float types are not supported on conditional compares, ccmp.

The PR, #71705, introduced compare chains.
Compare chains will generate conditional compares, but the checks to see if an op is a valid compare chain did not include any type checks. This PR adds the checks; if the operands are integral types, op is a valid compare chain.

Acceptance Criteria

• CI Passes

[ghost]

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

Issue Details

---

** Description **
Should resolve this assert: #74172

The assert was this in genCodeForConditionalCompare:

// No float support or swapping op1 and op2 to generate cmp reg, imm.assert(!varTypeIsFloating(op2Type));

Seems like float types are not supported on conditional compares, ccmp.

Compare chains will generate conditional compares, but the checks to see if an op is a valid compare chain did not include any type checks. This PR adds the checks; if the operands are integral types, op is a valid compare chain.

** Acceptance Criteria **

• CI Passes
• Run JITStress CI to see the test case succeed

Author:	TIHan
Assignees:	-
Labels:	area-CodeGen-coreclr
Milestone:	-

[TIHan]

@kunalspathak@BruceForstall This is ready.

[jakobbotsch]

ContainCheckCompareChain needs this check too. IsValidCompareChain is only used for the AND path because it wants to ensure both children can be converted before converting either of them.

[TIHan]

Yea, looks like we need to do that as well. Though, we do not have a test case that is failing as a result of not adding those checks there.

[jakobbotsch]

Indeed, that code won't be exercised until after #73472 is merged.

[a74nh]

Change looks good to me.

Thanks for fixing this! (and apologies for introducing it :) )

[TIHan]

@a74nh No need to apologize, not a problem at all! I mentioned that PR so we a history trail. I mean, @jakobbotsch has fixed bugs introduced by my PRs before, so it happens to everyone :)

[TIHan]

/backport to release/7.0

[github-actions]

Started backporting to release/7.0: https://github.com/dotnet/runtime/actions/runs/2980890529

[github-actions]

@TIHan an error occurred while backporting to release/7.0, please check the run log for details!

Error: @TIHan is not a repo collaborator, backporting is not allowed. If you're a collaborator please make sure your dotnet team membership visibility is set to Public on https://github.com/orgs/dotnet/people?query=TIHan

[TIHan]

@dotnet/jit-contrib Can someone run the backport release7.0? It says I'm not a collaborator at the moment.

[BruceForstall]

/backport to release/7.0

[github-actions]

Started backporting to release/7.0: https://github.com/dotnet/runtime/actions/runs/2981385893