* turn off composite scenario since it's flaky

* delete composite scenario

Avoid checking per object heap limit if the commit is not related to a specific object heap (#37725)

* Fix VeryLargeAmountOfEnumsToSerialize test

* User Parallel.For and correct indices

Fix Debug.Assert when using ReflectionMemberAccessor

* Added initial implementation of SpanSplitEnumerator

* Merged SpanSplitEnumerator and SpanSplitSequenceEnumerator

* Reordered .projitems entry

* Exposed System.Memory API additions

* Added SpanSplitEnumerator Tests

* Apply suggestions from code review
Co-authored-by: NLayomi Akinrinade <layomia@gmail.com>

* Apply suggestions from code review
Co-authored-by: NLayomi Akinrinade <layomia@gmail.com>

* Moved SpanSplitEnumerator to separate file

* Applied feedback to ReadOnlySpan.Split tests

* Renamed parameters/fields

* Removed mistaken compile include

* Fixed incorrect namespace on GetEnumeartor() return type

* Applied feedback for ReadOnlySpan.Split tests

* Added XML Documentation to public members

* Fixed cref reference to SpanSplitEnumerator in XML Documentation

* Fixed System.Memory.cs entries missing fully qualified identifier

* Apply suggestions from code review
Co-authored-by: NLayomi Akinrinade <layomia@gmail.com>

* Moved SpanSplit Tests to correct file

* Applied review feedback

* Removed trailing whitespace

* Fixed Unit Tests
Co-authored-by: NLayomi Akinrinade <layomia@gmail.com>

This change introduces initial provisions for dynamically loading
the native JIT based on the targeting OS and architecture; the change
expects the potentially multiple versions of clrjit to coexist with
Crossgen2 in the same folder, marked by a OS / arch suffix. Based
on these prerequisites the change adds ARM64 jobs to the Crossgen2
pipeline.

Thanks

Tomas

Co-authored-by: Nvargaz <vargaz@users.noreply.github.com>

* improve TLS frame parser

* retire SniHelper.cs

* feedback from review

* feedback from review

* feedback from review
Co-authored-by: NTomas Weinfurt <furt@Shining.local>

[master] Update dependencies from mono/linker Microsoft/vstest dotnet/llvm-project dotnet/xharness (#37643)

* Update dependencies from https://github.com/mono/linker build 20200609.1

Microsoft.NET.ILLink.Tasks
 From Version 5.0.0-preview.3.20306.1 -> To Version 5.0.0-preview.3.20309.1

* Update dependencies from https://github.com/microsoft/vstest build 20200609-09

Microsoft.NET.Test.Sdk
 From Version 16.7.0-preview-20200608-10 -> To Version 16.7.0-preview-20200609-09

* Update dependencies from https://github.com/dotnet/llvm-project build 20200608.1

runtime.linux-x64.Microsoft.NETCore.Runtime.Mono.LLVM.Tools , runtime.win-x64.Microsoft.NETCore.Runtime.Mono.LLVM.Tools , runtime.win-x64.Microsoft.NETCore.Runtime.Mono.LLVM.Sdk , runtime.osx.10.12-x64.Microsoft.NETCore.Runtime.Mono.LLVM.Tools , runtime.osx.10.12-x64.Microsoft.NETCore.Runtime.Mono.LLVM.Sdk , runtime.linux-x64.Microsoft.NETCore.Runtime.Mono.LLVM.Sdk
 From Version 9.0.1-alpha.1.20301.1 -> To Version 9.0.1-alpha.1.20308.1

* Update dependencies from https://github.com/dotnet/xharness build 20200608.1

Microsoft.DotNet.XHarness.TestRunners.Xunit
 From Version 1.0.0-prerelease.20305.1 -> To Version 1.0.0-prerelease.20308.1

* Update dependencies from https://github.com/mono/linker build 20200609.3

Microsoft.NET.ILLink.Tasks
 From Version 5.0.0-preview.3.20306.1 -> To Version 5.0.0-preview.3.20309.3

* Update dependencies from https://github.com/microsoft/vstest build 20200610-06

Microsoft.NET.Test.Sdk
 From Version 16.7.0-preview-20200608-10 -> To Version 16.8.0-preview-20200610-06

* Update dependencies from https://github.com/dotnet/xharness build 20200609.4

Microsoft.DotNet.XHarness.TestRunners.Xunit
 From Version 1.0.0-prerelease.20305.1 -> To Version 1.0.0-prerelease.20309.4
Co-authored-by: Ndotnet-maestro[bot] <dotnet-maestro[bot]@users.noreply.github.com>

* Profile driven layout
- Update to mpgo file format to hold weighted call graph data, and exclusive hit sample data
- Implementation of call graph data capture from both jitted and R2R code
- Add mechanism for parsing entrypoints from R2R files into arbitrary type systems insetad of just strings
- Update command line parser in dotnet-pgo to use named properties on class instead of arguments to main method
- Update dotnet-pgo to be able to use a etl.zip file
- Implement infrastructure for performing layout of PE file based on profile guided information
- Implement various profile guided method ordering routines
  - There are a variety of possible algorithms, these are simple to
  implement. More capable ones will generally use the call graph data.
  - Sort by exclusive weight
  - Sort by Hot (known to be used) and cold (not known to be used)
  - Sort by Hot (known to be used for more than 128 sample) Warm (known
  to be used) and Cold (not known to be used)
- Since the BulkType logging lock seems to be losing some types, skip the lock for MethodDetails data

* Add ILLink.Substitutions.xml files for System.Private.CoreLib.

Ported the settings from https://github.com/mono/mono/blob/eaa32d13659f0a6b6b5e62ddb49af68b1f9efb6c/sdks/wasm/src/linker-subs.xml and split them out as appropriate to reduce duplication across the different platform builds.

Contributes to #31785

* Remove all Platforms other than wasm to workaround linker bug.

See https://github.com/mono/linker/issues/1260

* Moving the System.Runtime.Intrinsics.Arm namespace in box

* Removing System.Runtime.Intrinsics.Experimental

Co-authored-by: NTomas Weinfurt <furt@Shining.local>

* Use [SkipLocalsInit] & remove code for ILLink to strip locals init

* Address review feedback

* Simply condition

* Clean up

* Fix condition

* Add System.Security.Cryptography.Algorithms to the pinvoke generator.

* Fix a stack overflow in WasmAppBuilder.

* [wasm] Unset MONO_LOG_MASK when running tests.

Its set to 'gc' by default which causes extra log messages to be added to the testResult.xml file.

* Turn off GC logging by default.

* adding the new version to avoid build failure

* update the assemblyin Pakcage version

This is a follow-up to #36918. It addresses one of the examples in #1007
where we remove a struct zero initialization but fail to clean up a dead
field assignment.

The change is not to mark a dependently promoted field as untracked
if we know that the struct local is no longer referenced.

I also addressed a couple of late cosmetic review comments from #36918.

No diffs in framework and benchmarks.

We had frame->ip used by EH and debugger and frame->state.ip used to save the ip of the next instruction to be executed when we return from the call. Since both represent more or less the same thing, unify them and avoid duplicate storing inside InterpFrame for each call. Because we use this ip with exception handling, we need to subtract 1 (similar to the jit) so the ip ends up being in the call instruction and not the next.
Co-authored-by: NBrzVlad <BrzVlad@users.noreply.github.com>

Extend the `PlatformNotSupportedException` for `OpenSslVersion` from `OSX` to `OSX, iOS, and tvOS`, as none of these Apple OS support OpenSsl.
Co-authored-by: NMitchell Hwang <mitchell.hwang@microsoft.com>

Looks like it was accidentally removed in https://github.com/dotnet/runtime/pull/37457

* Optimise MathF.CopySign and Math.CopySign

* Fix namespace error

* Clear up constant

* Remove x64 dependency for MathF method

* Remove unnecessary constant and unsafe from Math

* Remove unnecessary constant and unsafe from MathF

* Add ARM32/64 paths, and SSE41 path. Change integral constant to FP for perf and add integral constant for fallback

* Move conditional select to seperate helper methods, remove incorrect use of blendv

* Inline mask constants

* Correctly handle ISA not supported

* Remove old comment

* Fix syntax error

* Var -> explicit type

* Move helper method to sep class

As discussed w tanner

* move helper method

* Remove locals

* Fix namespace

* namespace

* Inline locals

* Change helper method to only accept float/double

* fix formatting

* fix sybntax

* Use a macro more widely used by unix platforms

* If time.h exists, clock_it must have been defined

* Check the existence of clockit_t

* Move type check to the location near other time checks

Need to load `System.Private.Corlib`

closes: https://github.com/dotnet/runtime/issues/37636

* SocketAsyncContext.Unix: remove Lock from IsReady

* Make QueueState an int

* Update comment

* cleanup comment

* Update for rebase

* Decrement sequenceNumber instead of using nullable int

* PR feedback

SocketsHttpHandlerTest_HttpClientHandlerTest_Http2.GetAsync_UnicodeHostName_SuccessStatusCodeInResponse is periodically failing with TimeoutException. It seems to be caused by the target server unavailability, this PR disables the test until a new unicode test host is found.

Fix race condition.

* Add ability for hostfxr_get_runtime_delegate to block on invalid delegate types.

This is required so that the runtime isn't loaded if the request has no chance of succeeding.

Fix spelling of initialization_options_t.

Add context_contract_version_set flag to initialization_options_t.

This allows hostpolicy to know whether it can rely on the version field of corehost_context_contract to be valid.
Also always initialize corehost_context_contract to {}.

* Allow app-based context to call hostfxr_get_runtime_delegate.

Currently, it may only request the load_assembly_and_get_function_pointer delegate.

Fix discrepancy in design doc.

Add error code HostApiUnsupportedScenario for blocks that might be removed. This gives the native host better information on why the request failed.

Add tests for running a component from an app context.