JIT: build pred lists before patchpoint expansion

[AndyAyersMS]

Move pred list building 3 phases earlier. It now happens just after instrumentation (or importation if we're not instrumenting) and just before expanding patchpoints.

Revise the patchpoint, indirect call, and post importer cleanup phases to do proper pred list maintenance.

Update the flow checker to handle cases we see when we haven't yet run the post importer cleanup.

Contributes to #80193.

[ghost]

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

Issue Details

---

Move pred list building 3 phases earlier. It now happens just after instrumentation (or importation if we're not instrumenting) and just before expanding patchpoints.

Revise the patchpoint, indirect call, and post importer cleanup phases to do proper pred list maintenance.

Update the flow checker to handle cases we see when we haven't yet run the post importer cleanup.

Contributes to #80193.

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

[AndyAyersMS]

@jakobbotsch PTAL
cc @dotnet/jit-contrib

Some minor diffs expected in Tier0 as we may leave blocks out of bbNum order and this perturbs LSRA. Did not seem worthwhile to quirk for this.

[AndyAyersMS]

/azp run runtime-coreclr jitstress, runtime-coreclr libraries-jitstress

[azure-pipelines]

Azure Pipelines successfully started running 2 pipeline(s).

[AndyAyersMS]

librares jitstress failure seems unrelated

  Starting:    System.Text.Json.Tests (parallel test collections = on, max threads = 2)

Assert failure(PID 28 [0x0000001c], Thread: 37 [0x0025]): obj->GetGCSafeMethodTable ()->HasFinalizer ()
    File: /__w/1/s/src/coreclr/vm/syncblk.cpp Line: 2039
    Image: /root/helix/work/correlation/dotnet