JIT: inconsitent logic around inlining pinvoke IL

[AndyAyersMS]

#113742 (comment)

impCheckForPInvokeCall contains profitability heuristics for inlining pinvoke IL stubs. Those can be overridden by inlining heuristics. Seems like neither is currently making the best decisions.

We should reconcile these and ideally centralize the logic in one place.

[dotnet-policy-service]

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