Simplify condition to detect recursive inlining

[jkotas]

This change simplifies the conditon and also makes it more robust. There are situation where the IL code pointer can differ for the same method.

[jkotas]

Related to #88749

[ghost]

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

Issue Details

---

This change simplifies the conditon and also makes it more robust. There are situation where the IL code pointer can differ for the same method.

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

[AndyAyersMS]

Any simple explanation as to why this address might vary or differ during a single jit invocation?

I vaguely remember something like this a while back from the crossgen2 host where the IL buffer wasn't kept pinned for long enough.

[jkotas]

I vaguely remember something like this a while back from the crossgen2 host where the IL buffer wasn't kept pinned for long enough.

Right, that is one possible case. The AOT compilers cache the IL bodies (that can be computed or altered by trimming) and flush the cache every once in a while:

runtime/src/coreclr/tools/aot/ILCompiler.Compiler/Compiler/Compilation.cs

Lines 83 to 85 in 563a549

	// Flush the cache when it grows too big
	if (_methodILCache.Count > 1000)
	_methodILCache = new ILCache(_methodILCache.ILProvider);

Another one is profiler IL rewriting. The profilers can change the method IL body at will.