[NO-MERGE] Add access enforcement opts and access marker elimination to addHighLevelModulePipeline.

[zoecarver]

I want to see the performance gains of removing access markers in the middle of the pass pipeline. I think this could be the source of the performance improvements in #31423. After the benchmarks run, I'll close this PR.

[zoecarver]

@swift-ci please benchmark.

[swift-ci]

Performance: -O

Improvement	OLD	NEW	DELTA	RATIO
ObjectiveCBridgeStubToNSDateRef	3880	3440	-11.3%	1.13x (?)
DictionarySubscriptDefaultMutationOfObjects	1680	1500	-10.7%	1.12x (?)
StringToDataMedium	3900	3500	-10.3%	1.11x (?)

Code size: -O

Improvement	OLD	NEW	DELTA	RATIO
ArrayOfRef.o	8907	8795	-1.3%	1.01x

Performance: -Osize

Regression	OLD	NEW	DELTA	RATIO
MapReduceLazyCollectionShort	40	85	+112.5%	0.47x
FlattenListFlatMap	4185	5941	+42.0%	0.70x (?)
PointerArithmetics	31400	37100	+18.2%	0.85x
BinaryFloatingPointPropertiesBinade	31	34	+9.7%	0.91x (?)
ProtocolDispatch	314	342	+8.9%	0.92x
Improvement	OLD	NEW	DELTA	RATIO
ClassArrayGetter2	1830	130	-92.9%	14.08x
MapReduceClass2	198	25	-87.4%	7.92x
MapReduceClassShort2	362	200	-44.8%	1.81x
BinaryFloatingPointPropertiesNextUp	35	31	-11.4%	1.13x (?)
MapReduce	179	160	-10.6%	1.12x (?)
DictionarySubscriptDefaultMutationOfObjects	1720	1540	-10.5%	1.12x (?)

Code size: -Osize

Improvement	OLD	NEW	DELTA	RATIO
ArrayOfGenericRef.o	8304	8032	-3.3%	1.03x
ArrayOfRef.o	8581	8309	-3.2%	1.03x
DictionaryBridge.o	3309	3221	-2.7%	1.03x
LinkedList.o	2257	2200	-2.5%	1.03x
OpaqueConsumingUsers.o	2351	2303	-2.0%	1.02x
RGBHistogram.o	19552	19200	-1.8%	1.02x
ObjectAllocation.o	4298	4242	-1.3%	1.01x

Performance: -Onone

Regression	OLD	NEW	DELTA	RATIO
DataReplaceSmallBuffer	10400	11700	+12.5%	0.89x (?)
DictionaryBridgeToObjC_Access	857	958	+11.8%	0.89x (?)
ObjectiveCBridgeStubNSDateRefAccess	5345	5790	+8.3%	0.92x (?)
Improvement	OLD	NEW	DELTA	RATIO
ObjectiveCBridgeStubToNSDate2	680	620	-8.8%	1.10x (?)
Data.hash.Small	348	324	-6.9%	1.07x (?)

Code size: -swiftlibs

How to read the data

The tables contain differences in performance which are larger than 8% and differences in code size which are larger than 1%.

If you see any unexpected regressions, you should consider fixing the
regressions before you merge the PR.

Noise: Sometimes the performance results (not code size!) contain false
alarms. Unexpected regressions which are marked with '(?)' are probably noise.
If you see regressions which you cannot explain you can try to run the
benchmarks again. If regressions still show up, please consult with the
performance team (@eeckstein).

Hardware Overview

  Model Name: Mac Pro
  Model Identifier: MacPro6,1
  Processor Name: 12-Core Intel Xeon E5
  Processor Speed: 2.7 GHz
  Number of Processors: 1
  Total Number of Cores: 12
  L2 Cache (per Core): 256 KB
  L3 Cache: 30 MB
  Memory: 64 GB