Rename INS_FLGAS in consistent style for xarch emitter.

[Ruihan-Yin]

Fix #128266

This PR implements the short-term naming cleanup by standardizing the xarch instruction flags on the INS_FLAGS_* convention.

Converting insFlags to an enum class would require broader changes to shared helper signatures and the ARM code paths, where insFlags has a different purpose. Given the current release-cycle stage, that refactoring is deferred to a larger .NET 12 change.

[azure-pipelines]

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

[Copilot]

Pull request overview

Renames a set of xarch insFlags instruction-flag constants to a standardized INS_FLAGS_* naming convention and updates all in-repo call sites accordingly.

Changes:

• Renamed several xarch insFlags enum values in instr.h (x87 / AVX / APX / pseudo-name flags, plus related masks).
• Updated flag checks in instr.cpp and emitxarch.cpp to use the new identifiers.
• Updated xarch instruction metadata tables (instrsxarch.h) to reference the renamed flags.

Reviewed changes

Copilot reviewed 3 out of 4 changed files in this pull request and generated 1 comment.

File	Description
src/coreclr/jit/instr.h	Renames xarch insFlags constants and related masks in the enum definition.
src/coreclr/jit/instr.cpp	Updates instIsFP / instHasPseudoName to use renamed flags.
src/coreclr/jit/emitxarch.cpp	Updates EVEX/APX and AVX helper predicates to use renamed flags.
src/coreclr/jit/instrsxarch.h	Updates instruction metadata flag expressions to use renamed flags.

[Copilot]

Pull request overview

Copilot reviewed 3 out of 4 changed files in this pull request and generated no new comments.

Suppressed comments (1)

src/coreclr/jit/instr.h:247

• INS_FLAGS_ApxEvexMask introduces mixed-case acronym spelling (Apx/Evex) while nearby flags consistently use all-caps acronyms (e.g., Encoding_EVEX, Encoding_EVEX_APX_ONLY, Encoding_REX2, INS_FLAGS_IsDstDstSrcAVXInstruction). Since this PR’s goal is naming consistency, consider keeping acronyms consistently capitalized (e.g., INS_FLAGS_APX_EVEX_Mask) to match the established style in this enum.

    // Mask of all APX-EVEX related flags. An instruction matches this mask if it either only exists in the
    // APX EVEX space (Encoding_EVEX_APX_ONLY, e.g. crc32_apx/movbe_apx) or gains an APX encoding through the
    // NDD (INS_FLAGS_HasNDD) or NF (INS_FLAGS_HasNF) features. Use it to quickly test whether an instruction
    // has any APX-EVEX capability.
    INS_FLAGS_ApxEvexMask = (Encoding_EVEX_APX_ONLY | INS_FLAGS_HasNDD | INS_FLAGS_HasNF),

[azure-pipelines]

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

[Ruihan-Yin]

Hi @JulieLeeMSFT @tannergooding

This change would close #128266 with a less impactful approach.

For the long-term solution, I suggest tracking the refactoring to enum class insFlags in a separate issue, as it touches code across multiple platforms, and targeting it for a future cycle. This keeps the current fix scoped while preserving visibility of the broader cleanup effort.

Please let me know if you have a different idea or approach in mind.

[JulieLeeMSFT]

LGTM.
Just one question from LLM: The TARGET_AMD64-guarded regions (e.g. emitxarch.cpp:2340 CCMP/CTEST, IsApxExtendedEvexInstruction at 548-558) compile only under AMD64, so an x86-only CI leg would not catch a stale reference inside an AMD64 #ifdef. A clean AMD64 build + an unchanged SPMI asm diff is sufficient proof; just confirming if done.
@dhartglassMSFT, please follow up to merge this PR.

[Ruihan-Yin]

LGTM. Just one question from LLM: The TARGET_AMD64-guarded regions (e.g. emitxarch.cpp:2340 CCMP/CTEST, IsApxExtendedEvexInstruction at 548-558) compile only under AMD64, so an x86-only CI leg would not catch a stale reference inside an AMD64 #ifdef. A clean AMD64 build + an unchanged SPMI asm diff is sufficient proof; just confirming if done. @dhartglassMSFT, please follow up to merge this PR.

Yes, the x64 pipeline and my local AMD64 build & tests passed, confirming there are no stale references in AMD64-only code, primarily involving APX flags. The Windows x64 SuperPMI asm diff also reported no differences.