JIT: handle x86 tail call via helper when control expression is spilled

[AndyAyersMS]

Make sure we order the control expression after the spill.

Also add a range control to selectively enable instrumentation stress.

Fixes #120170.

[Copilot]

Pull Request Overview

This PR fixes a bug in the JIT compiler's handling of x86 tail calls via helper functions when the control expression gets spilled to a temporary variable. The fix ensures proper ordering of code by inserting the call target range before the call rather than after an argument, and then reordering all putargs to be adjacent to the call.

• Moves call target range insertion from after arg0 to before the call in LowerTailCallViaJitHelper
• Adds a call to MoveCFGCallArgs to properly reorder putargs after the range insertion
• Introduces a new configuration option JitInstrumentIfOptimizingRange to selectively control instrumentation stress testing

Reviewed Changes

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

File	Description
src/coreclr/jit/lower.cpp	Updates tail call lowering logic to handle spilled control expressions by changing insertion order and adding putarg reordering
src/coreclr/jit/jitconfigvalues.h	Adds new configuration string for instrumentation range control
src/coreclr/jit/compiler.cpp	Implements range-based filtering for the instrumentation optimization feature

[AndyAyersMS]

@jakobbotsch PTAL
cc @dotnet/jit-contrib

Small handful of diffs.

[dotnet-policy-service]

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

[AndyAyersMS]

/azp run runtime-coreclr pgostress, runtime-coreclr libraries-pgo

[azure-pipelines]

Azure Pipelines successfully started running 2 pipeline(s).

[AndyAyersMS]

Remaining failures look like timeouts / stackoverflow tester.

Tier1+instr is likely fairly slow; I wonder if we should invest in making it faster.

[AndyAyersMS]

/azp run runtime-coreclr pgostress, runtime-coreclr libraries-pgo

[azure-pipelines]

Azure Pipelines successfully started running 2 pipeline(s).

[AndyAyersMS]

pgostress failure is #110173

libraries-pgo is a bit messy, but failures are either timeouts (#120902), ooms (#120918), or one failure that seems timing related and unrelated to this PR:

x64 windows

DOTNET_JitDelegateProfiling=1
DOTNET_JitVTableProfiling=1
DOTNET_ReadyToRun=0
DOTNET_TC_QuickJitForLoops=1
DOTNET_TieredCompilation=1
DOTNET_TieredPGO=1

  Starting:    System.Transactions.Local.Tests (parallel test collections = on [4 threads], stop on fail = off
    System.Transactions.Tests.TransactionManagerTest.DefaultTimeout_MaxTimeout_Set_Get [FAIL]
      Assert.Equal() Failure: Values differ
      Expected: 00:02:00
      Actual:   00:01:00