Runtime ingestion #13

* Update TargetFramework for Mono's SPC * Update TargetFramework for CoreCLR SPC as well

This PR has 2 objectives. First it makes the interp stack allocated in a single contiguous area of memory. This will enable us to do fast pointer bumping when switching between frames and not have to worry about scattered memory fragments. Second, it makes the stack space used by a frame (locals + execution stack) a compile time constant. This could prove useful for future optimizations, because we can tell at compile time at which stack offset every instruction outputs its result. In order to achieve this, localloc memory has been kicked out of the normal stack space into its own allocator, which is almost identical to the old stack fragment allocator. Co-authored-by: BrzVlad <BrzVlad@users.noreply.github.com>

I've also annotated CoreCLR's `Type.GetType` in the same spot as Mono. Mono's implementation was triggering warnings, which is what promted me to add it there, but we should annotate this nonetheless because absent intrinsic treatment in Linker, this would be unsafe.

* Use ProjectReference to build against CoreLib

* Devirtualized default value comparison * Fixed double comparison for ref types * Fixed serialization * Addressed issues

* fix helper multithreading issue -- don't re-read flags that can change * fix helper logging and abort on error * change strategy for protecting original method entry block when doing an OSR compile. Instead of marking it don't remove, use artificial ref count boost.

* move System.Formats.Cbor to standalone project * address feedback * address feedback * Remove Directory.Build.props altogether

* Fix PEFile::GetAssemblyLoadContext during early runtime stages The PEFile::GetAssemblyLoadContext was asserting if it was called before the m_pTPABinderContext was created in AppDomain::CreateBinderContext. That turned out to be a problem for a check performed in ReadyToRunInfo::Initialize. This change calls AppDomain::CreateBinderContext from the PEFile::SetupAssemblyLoadContext instead of calling AppDomain::GetTPABinderContext. That ensures that the PEFile::GetAssemblyLoadContext can always return valid AssemblyLoadContext.

…d Vector64<long>.CreateScalar (#38139)

* Enable Activity tracking on generic host * Add test * Remove trailing whitespace * Bad merge * React to W3C format change * Remove conditional fact

These nine tests fail with the JIT assertion failure '!comp->compDoOldStructRetyping()'. I'm not 100% sure whether this makes the job green again as I see a somewhat different error in the AdvSimd_Part3_r test so I'm reluctant to disable it with the same bug before more investigation. Thanks Tomas

[wasm]Don't assume SharedArrayBuffer exists.

…e in .NET 5.0 (#38177) * Bring back System.IO.Pipes.AccessControl package and make it available in .NET 5.0. * address suggestions * fix test error * Microsoft.Windows.Compatibility.pkgproj

* Update UnmanagedCallersOnlyAttribute API surface * Update comments and block unknown calling convention types for UnmanagedCallersOnlyAttribute. * UnmanagedCallersOnlyAttribute should not be inheritable.

* Fix Assertion failed 'rhs != UninitVal()' #35413 Issue was a preHeader blocks was inserted and the compacted with the previous block and we had a null value for bbReach during compaction. * Re-enable test b539509

Port the managed type system unit tests and wire them up to build.cmd -test so they run if the clr.tools subset is included in compilation. Port enough USG layout and `Array<T>` support so we can keep their respective tests running in this repo.

In the post-call part of a pinvoke inline call frame, it's not safe to start a stress mode GC in the window between checking `g_TrapReturningThreads` and the call to `CORINFO_HELP_STOP_FOR_GC`. The call instruction is already getting special treatement, but there may be other instructions between the check and call. Instead of trying to pattern match them all, suppress GC stress if `g_TrapReturningThreads` is true, the thread is in cooperative mode, and there's an active inline call frame. Closes #37236.

In runtime/src/coreclr/src/jit/unwindamd64.cpp line 516 frame register is printed as none if it is RAX, made the change in extensions.cs to reflect this.:

* Update perf proj from arcade * Add timeout for pr and internal jobs * Change to a variable that is set * Escape variable with single quotes * Change timeout from days to minutes * Fix typo * Change to default

Example recent run: https://dev.azure.com/dnceng/public/_build/results?buildId=698082&view=results

(1) Switch most legs to use composite mode; (2) Add ARM64 legs; (3) Add a single non-composite leg; (4) Reduce the frequency to 3 times a week in view of the increased level of testing. Thanks Tomas

…249) Disable more tests failing on Windows ARM64 with the bug dotnet/runtime#37883 (compDoOldStructRetyping assertion failure in JIT). I think this is currently the biggest known error bucket in CG2 composite testing. Thanks Tomas

* Re-enable Apphost.Bundle.Tests.BundleRename test * Remove StaticHostApp StaticHostTest used the StaticHostApp (which is effectively a copy of StandaloneApp) as a work-around for synchronization problems. Remove this work-around. * Synchronize GetNewTestArtifactPath() Add locking around GetNewTestArtifactPath() so that multiple threads attempting to create new copies of a TextFixture don't collide and work on the same path.

…ASM (#38266)

* update branding to review 8 * set reset flag to false

Corresponding issues have been fixed.

…failures (#38232) Co-authored-by: Mitchell Hwang <mitchell.hwang@microsoft.com>

…e.Xml (#38142) * remove unnecessary cross compilation * remove extra < * correct pinvoke path * use ridless for unix os * remove extra warn

* Enable Matrix3x2Tests.cs * Also enable Matrix3x2CreateScaleCenterTest1 and Matrix3x2CreateScaleCenterTest3

…s (#38252)

* Add trimming tests for System.Private.Xml.Linq changes * Address review feedback

…#38289) When the `configure` file is touched the autotools build will rerun it without the `emconfigure` wrapper like we do in mono.proj. This means instead of emscripten's `ar` or `ranlib` we're getting the system tools instead, causing compilation errors like: ``` libmono-icall-table.a: archive is missing an index; Use emar when creating libraries to ensure an index is created ``` Fix this by adding a dependency on configure as well so we rerun the whole autogen.sh process.

…38251)

If we have a very small root method that calls a large method that is marked with AggressiveInlining, we may fail to inline because of a budget check. Allow such inlines to bypass the check. Closes #38106.

* Annotate Microsoft.CSharp * Remove unnecessary NoWarn property

This is a fllow-up to #37335 and #37840. When an inline fails we replace `GT_RET_EXPR`with the original `GT_CALL` node. `GT_RET_EXPR`may end up in a basic block other than the original `GT_CALL` so we need to propagate basic block flags. Fixes #36588.

* Remove RCWAuxiliaryData * Remove unused functions

* add DefaultPropertyAttribute and DefaultEventAttribute to PrintDocument class and add references to System.ComponentModel.TypeConverter * add the DefaultPropertyAttribute and DefaultEventAttribute to PrintDocument in the ref Add DefaultEventAttribute and DefaultPropertyAttribute to PrintDocument in PrintDoument.Unix.cs

#38156 tracks work for solution

…8256) Fixes dotnet/runtime#38164.

Fixes dotnet/runtime#37358

Co-authored-by: Mitchell Hwang <mitchell.hwang@microsoft.com>

* Add telemetry to System.Net.Http This is the first of several PRs to add telemetry focused events and counters. A few things to note: * Added a new dedicated namespace for telemetry. Current feedback from partners is that our existing events are "confusing and a mess" due to their focus on low-level diagnostics and verbose debugging. See discussion in #37428 regarding reasons for a new namespace. * Please comment on the naming patterns for the events, counters, and display names, etc. I looked at ASP.NET events/counters and @anurse older guidance in https://gist.github.com/anurse/af1859663ac91c6cf69c820cebe92303. But there are inconsistencies between all of that. So, I'd like to get agreement on naming patterns for events/counters (i.e. noun-verb vs. verb-noun, present tense vs. past tense, etc.). * No automated tests were added at this time. The current discussions with others (reverse proxy team) is that automated tests are brittle in CI. But manual tests will be run. Contributes to #37428 * Address PR feedback * PR feedback * Added RequestAbort as separate event method with dimensions (host, port) * Address PR feedback 2 * RequestStart now includes host, port, scheme, path and version * RequestStart and RequestStop now bracket the lifetime of the request including reading the response body stream. If the response body is not complete read then disposing the stream will trigger RequestStop as well. * Address feedback 3 * Use exception filter to improve performance * Avoid negative 'current-requests' counter value * Address feedback 4

Allows importing a cert and key together easily from PEM files (one file with both parts, or two different files), and importing all of the certificates out of a single multi-PEM file into a collection. Co-authored-by: Jeremy Barton <jbarton@microsoft.com>

Fixes mono/mono#19612 Added missing signatures for OpenGLES binding in WebAssembly. Co-authored-by: davilovick <davilovick@users.noreply.github.com>

* Fix rethrow to retain stack When resolving types we were rethrowing after attempting to get implicit types. Switching to a parameterless rethrow to ensure that we keep the original stack. Add a test for roll-forward behavior. Add a few tests for settings events. Minor code style fixes. * Skip type resolution test on Mono The new type roll-forward test I added rolls backwards on Mono.

We use lowercase words, see https://github.com/dotnet/runtime/blob/d908270add914db0e9fb3ce72e93c410e8f0f95a/src/libraries/Microsoft.Extensions.Logging.Console/src/ConsoleLogger.cs#L234-L235

…net/xharness (#38268) * Update dependencies from https://github.com/mono/linker build 20200622.1 Microsoft.NET.ILLink.Tasks From Version 5.0.0-preview.3.20317.2 -> To Version 5.0.0-preview.3.20322.1 * Update dependencies from https://github.com/dotnet/llvm-project build 20200622.1 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.1.20315.1 -> To Version 9.0.1-alpha.1.20322.1 * Update dependencies from https://github.com/dotnet/xharness build 20200622.1 Microsoft.DotNet.XHarness.TestRunners.Xunit From Version 1.0.0-prerelease.20318.2 -> To Version 1.0.0-prerelease.20322.1 Co-authored-by: dotnet-maestro[bot] <dotnet-maestro[bot]@users.noreply.github.com>

Yet another batch of trimming annotations.

* Add Function tests. * Format code.

Tools like xharness or BenchmarkDotNet want to parse the output unmodified and the prefix makes that harder. It's enough if we only prefix stderr.

Fixes tests in dotnet/runtime that currently fail with `WASM-ERR: CANNOT HANDLE COOKIE VL`. Co-authored-by: akoeplinger <akoeplinger@users.noreply.github.com>

* Add an assert to avoid exception. * Delete two outdated comments about const casts.

And disable it for Mono, given our current bug. This is tested indirectly by the other test I've updated, but I think it ought to be more clear.

… or returning an unsupported T (#38241) * Ensure we throw NotSupportedException for any Vector<T> method taking or returning an unsupported T * Exit early for SIMDAsHWIntrinsics if the baseType is unsupported

…es" book (#38333)

…ive_trampoline. (#33454) Co-authored-by: jaykrell <jaykrell@users.noreply.github.com>

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

Introduces sync version of HttpClient.Send and all necessary changes to make it work synchronously (e.g.: HttpContent, HttpMessageHandler and their child classes). The change works properly only for HTTP1.1, for unsupported scenarios like HTTP2 throws. Testing the change uses existing tests calling HttpClient.SendAsync and introduces test argument calling the same test twice, once with HttpClient.SendAsync and then with HttpClient.Send. Resolves #32125

…equence APIs (#38075) * Add a note for SpanHelpers.SequenceCompareTo() and SequenceEqual() method * Use InsertSelectedScalar for WithUpper() and WithLower()

* Add netstandard2.1 target to System.Threading.Channels (#38046) * code review feedback

* Implement ConcurrentDictionary.GetEnumerator().Reset() Until we come up with a good pattern for adding a struct-based enumerator here, this enables the enumerator to be reused. If we choose to expose a struct-based implementation in the future, this will also make that trivial. * Update src/libraries/System.Collections.Concurrent/src/System/Collections/Concurrent/ConcurrentDictionary.cs

… (#38236) - Use existing Precode/Local import split - Remove special casing of ExternalMethodImport - Rename LocalMethodImport to DelayLoadMethodImport and fix comparison bug - Add new tests that covers these sorts of scenarios with delegates and function pointers

With this change, `mono_post_native_crash_handler` chains the crashing signal directly, instead of that having to happen after `mono_handle_native_crash` returns (which created a fragile repetitive code pattern). Finally, this way `mono_post_native_crash_handler` can also re-raise the signal right after chaining it. Since we remove our crashing signal handlers in `mono_handle_native_crash`, this should allow the OS to handle the signal as the final action. TODO: - [x] Investigate the possibly broken existing pattern where `mono_chain_signal` is called before `mono_handle_native_crash` , and if the crash is chained, `mono_handle_native_crash` doesn't run (example: `mini-runtime.c:mono_sigfpe_signal_handler`) - [x] Ensure that all of the crashing signal handlers are removed at the beginning of `mono_handle_native_crash` Fixes mono/mono#19860 Co-authored-by: alexischr <alexischr@users.noreply.github.com>

Restore and port Andrew's signature tests that had been reverted last year, originally from dotnet/corert@5575cc0

…GT_OBJ(lcl_addr)` and `GT_BLK(lcl_addr)` (#38314) Fixes #38070.

…ft.Extensions.Logging (#38338) Filed dotnet/runtime#38337 for the tests that cause a runtime assert. The rest are normal threading ignores.

This is needed in llvmonly mode, since that mode doesn't use the fastpath lazy rgctx fetch trampolines.  Co-authored-by: vargaz <vargaz@users.noreply.github.com>

Co-authored-by: Mitchell Hwang <mitchell.hwang@microsoft.com>

* sqdmlal * sqdmlal2 * sqmmlsl * sqmmlsl2 * sqdmulh * sqdmull * sqdmull2 * sqdrmlah * sqdrmlsh * sqdrmulh * sqxtn * sqxtn2 * sqxtun * sqxtun2 * uqxtn * uqxtn2 * sdot * udot

* Do not swallow exception on debug builds * fix contract

Do not permit OID properties to be changed after they have been set to a non-null value.

* Remove opcode validation when specifying relatedActivityId. * Remove test that is no longer needed.

… tests (#38368) Co-authored-by: Steve Pfister <steve.pfister@microsoft.com>

This allows the test suite to pass on WASM: `Tests run: 27708, Errors: 0, Failures: 0, Skipped: 107. Time: 159.380298s`

This allows the test suite to pass on WASM: `Tests run: 883, Errors: 0, Failures: 0, Skipped: 30. Time: 58.57082s`

Remove the defensive copies in shared framework assemblies, and remove the defensive copies in the Pkcs library except when targeting netstandard.

* Fix spill check for multireg local Check the individual spill flag for each field. Fix #38228

…#38363)

Fixes dotnet/runtime#936. A lot of this code is a port of what have in [corefxlab](https://github.com/dotnet/corefxlab/tree/master/src/System.Numerics.Experimental).

* Fix TPA map hash calculation. The point of issue is "the Turkish-I Problem". After locale changed, towupper() provide another result for "i" and different hash are calculated in case if file name have "i" letter. * Regression test for dotnet#37910

And update to the latest beta bits available.

…e phase (#38083) Avoid GT_COMMA insertion during Global Morph and CSE phases

Even though intuitively Reflection.Emit doesn't feel trimming friendly, it probably is if we enforce base types and interfaces of emitted things are preserved in their entirety.

Build Wasm for testing via Helix. A bunch of test suites are disabled on Wasm here, and will be re-enabled as we get the up and running.

* Don't set GCState to 1 in PInvoke Epilog Remove the incorrect setting of GCState and fix the liveness computation for compLvFrameListRoot for tailcalls. Fix #38192

* Update api-review-process.md * Update docs/project/api-review-process.md * more Co-authored-by: Dan Moseley <danmose@microsoft.com>

* termios: disable CR/LF mapping flags * fix mapping logic for '\n' characters * only disable cr/lf conversions on Console.ReadKey paths * add comments * address feedback * address feedback * move character conversions to the managed layer * fix issue * add key chord manual tests and doc manual testing * ensure terminal is initialized for Console.OpenStandardInput * Enable ICRNL when reading from UnixConsoleStream * add a manual test for OpenStandardInput * remove CR to LF conversion for Console.ReadLine * rename argument * add CR-to-LF transformation to Console.ReadOrPeek * use ConditionalFact attribute in test * remove dead code

FSW is not supported on wasm. Co-authored-by: Steve Pfister <steve.pfister@microsoft.com>

In #38163 I added a depth field to one of the inline policy base classes. A derived class already had a similar field. Unify in favor of the base class field. The inline policies implemented by these derived policy classes are sometimes used during jit stress (eg for random inlining). Fixes #38374.

* Fix nullable arguments in ObjectFactory * Address feedback

…38066) * Add trimming tests for System.ComponentModel.TypeConverter changes * Address review feedback * Address review feedback

… (#38317) We shouldn't need this anymore as the case it protects against should be covered by the new check added in #38246.

- Fix a crash in arm64 and NGen environment. - Iniitialize sectionIndex and pModule values in ExternalMethodFixupStub to call ExternalMethodFixupWorker

…Tests (#38355) Co-authored-by: Mitchell Hwang <mitchell.hwang@microsoft.com>

…38429)

This has three primary benefits: 1. In some cases, in particular for static fields where other static fields have non-default initialization, the C# compiler can't elide the default initialization, resulting in more IL. 2. It highlights places where the field is otherwise unused and can be deleted, but where the presence of the initialization was silencing the C# compiler's warnings about the unused field. 3. It increases consistency and decreases maintenance / code size / etc.

System.Linq.Expressions is not getting compiled without FEATURE_COMPILE so it started rotting. * Nullable annotations * Fix break introduced in #32768

* Add support for shared source in linker test code * Clean up DynamicDependencyAttribute usages * Clean up DynamicDependencyAttribute: review feedback + another test Co-authored-by: Jose Perez Rodriguez <joperezr@microsoft.com>

Co-authored-by: Thomas Levesque <thomaslevesque@users.noreply.github.com>

…option. (#38424)

- Require a signature that matches the actual target method definition - Except for using MethodImpl records that point at base type - In that case, apply Substitution to provide a signature matching environment that matches the parent type in the MemberRef - New test to cover the change in behavior with regards to interface MethodImpls This change allows specifying an unambiguous target when there was previously potential for ambiguity in the presence of generics and MethodImpl records.

…nt returns support (#37516) Covariant return implementation for Mono. --- There are also a couple of general Mono fixes in here, too: 1. Interpreter support for `MONO_TYPE_FNPTR` in `mint_type()` 2. a "signature assignment" mode for `mono_class_is_assignable_from` that doesn't treat `IntPtr[]` and `long[]` (or `int[]`) as interchangeable 3. support for unmanaged pointer comparisons in `mono_class_is_assignable_from` 4. Factor `mono_byref_type_is_assignable_from` out of `ves_icall_RuntimeTypeHandle_type_is_assignable_from` and make it generally usable. --- For covariant returns, there are a couple of pieces here: 1. Method overrides come in two flavors: implicit (matching by name and sig) and explicit (using a `.override` directive). * for explicit overrides, we have to check that the override has a signature that is subsumed by (its return type is a subtype of the return type of) the previous override, if any, and the declared method. * for implicit overrides, we have to check that the override has a signature that is subsumed by the previous most derived override (not just the one that we happened to find with a precisely matching signature). 2. If any override is marked by `[PreserveBaseOverridesAttribute]` then we need to find all the other slots that may have methods from the override chain and update them. (The attribute is expected to be applied to a `newslot virtual` method that is an explicit `.override` of some other method). There's an interesting test case where a class has _both_ an implicit and and explicit override for the same slot, arrived via different methods - in this case, the implicit one is supposed to be ignored. We implement this by doing the implicit slot filling first, then the explicit, and saving the covariant return checking until after both are done. --- Because the method (`MethodImpl`) is both an override of another virtual method, and is itself a `newslot`, it will occupy (at least) 2 slots in the vtable - its own newslot and the `MethodDecl` that it is overriding. If a third class subclasses the class with the `MethodImpl` or one of its descendents and itself overrides either the `MethodDecl` or (one of the) `MethodImpl` (either implicitly or with another explicit .override) the `PreserveBaseOverridesAttribute` is a signal to apply the new override that it should apply to all the slots. The way this works is when processing the overrides, we walk up the class hierarchy looking at the contents of the impl slot and if any of them have the attribute, we make sure to add a mapping to the "`override_map`" to replace any occurrences of a previous impl with the newest (most derived) impl. An (existing) later pass goes through the vtable and applies the override_map. Contributes to dotnet/runtime#37509 * [metadata] Implement support for PreserveBaseOverridesAttribute The attribute is expected to be applied to a `newslot virtual` method that is an explicit .override of some other method. Becuase the method (MethodImpl) is both an override of another virtual method, and is itself a `newslot`, it will occupy (at least) 2 slots in the vtable - its own newslot and the MethodDecl that it is overriding. If a third class subclasses the class with the MethodImpl or one of its descendents and itself overrides either the MethodDecl or (one of the) MethodImpl (either implicitly or with another explicit .override) the PreserveBaseOverridesAttribute is a signal to apply the new override that it should apply to all the slots. The way this works is when processing the overrides, we walk up the class hierarchy looking at the contents of the impl slot and if any of them have the attribute, we make sure to add a mapping to the "override_map" to replace any occurrences of a previous impl with the newest (most derived) impl. An (existing) later pass goes through the vtable and applies the override_map. * add FIXME for dynamic images * scan entire chain of inheritance, don't stop at the class with the attribute Makes OverrideMoreDerivedReturn.il pass * UnitTestDelegates is also passing * [metadata] Add mono_metadata_signature_equal_no_ret * inflate signature->ret when doing covariant checking * [class] Add mono_class_signature_is_assignable_from_checked In ECMA I.8.7.1 "assignment compatability for signature types" we need to distinguish IntPtr[] from int32[] and int64[]. The ordinary cast_class for an array type in Mono implements the "intermediate type" relation from I.8.7.3 "general assignment compatability" - IntPtr[] gets treated as inter-castable with int32[] or int64[], depending on platform pointer size. The new mono_class_signature_is_assignable_from_checked function does the finer relation for signatures, while the old mono_class_signature_is_assignable_from_checked preserves the existing behavior used elsewhere in the runtime. * Check every method previously in the slot against the new impl * UnitTest_GVM passes * [metadata] Don't infinite loop in mono_class_get_flags for a FNPTR When a MonoClass is a MONO_TYPE_FNPTR, the element_class is itself. We should probably not do that, but meanwhile, make `mono_class_get_flags` return something else for function pointer types. Makes mono_class_is_interface not overflow the stack * [metadata] Set MonoClass:cast_class for pointer types to the intermediate type Ignore signedness differences and set cast_class to the storage type, same as for arrays. This is used by mono_class_is_assignable_from when comparing IntPtr* vs ulong*, for example. * [metadata] Make mono_class_is_assignable_from work for unmanaged pointer types The issue is that we need to compare intermediate type or the reduced type of the element type, depending on whether we're doing compatability or general assignment compatability (upshot: ulong* and long* are always inter-assignable, but IntPtr* is not in signatures, but is for general storage) * [metadata] Pull mono_type_byref_is_assignable_from out of ves_icall_RuntimeTypeHandle_type_is_assignable_from Make it a separate internal function * [metadata] mono_type_byref_is_assignable_from for valuetypes and class types should check for identity X& and Y& where both are either a reference types or valuetypes are only assignable if X and Y are identical. The old code was checking for identity for valuetypes, but for reference types it would allow any two reference types at all, which is incorrect. * [class-init] Check for byref types in covariant return signatures Lets us pass Loader/classloader/MethodImpl/CovariantReturns/UnitTest/CompatibleWithTest.il from the CoreCLR testsuite * [interp] handle MONO_TYPE_FNPTR in mint_type It's just a pointer-sized integer, as far as the interpreter's type discipline is concerned * [class-init] Use a bit for covariant return method impls Allows us to check implicit overrides against previous impl's sig when the previous impl potentially had a covariant return type. Also use the bit to decide if we need to check signatures of the entire impl chain for explicit instantiations * ReturnTypeValidation unit tests work on Mono now * [class-init] check covariant vtable slots after overloads are applied The check has to be done after both implicit and explicit overrides have been applied because some implicit overrides must be ignored. If the check is done as soon as we detect that a covariant return method override is needed, we will incorrectly check implicit overrides that are to be ignored because a later explicit override applied to the same slot. * [metadata] Fix byref IsAssignableFrom The reflection API has an interesting behavior, whereas the ECMA definition is stricter. ``` public interface I {} public class B {} public class D : B, I {} public struct S : I { } class Program { static void Main(string[] args) { var tb = typeof (B).MakeByRefType (); var td = typeof (D).MakeByRefType (); var ti = typeof (I).MakeByRefType (); var ts = typeof (S).MakeByRefType (); Console.WriteLine ($"{tb.FullName} is assignable from {td.FullName} ? {tb.IsAssignableFrom (td) }"); Console.WriteLine ($"{td.FullName} is assignable from {tb.FullName} ? {td.IsAssignableFrom (tb) }"); Console.WriteLine ($"{ti.FullName} is assignable from {td.FullName} ? {ti.IsAssignableFrom (td) }"); Console.WriteLine ($"{td.FullName} is assignable from {ti.FullName} ? {td.IsAssignableFrom (ti) }"); Console.WriteLine ($"{ti.FullName} is assignable from {ts.FullName} ? {ti.IsAssignableFrom (ts) }"); Console.WriteLine ($"{ts.FullName} is assignable from {ti.FullName} ? {ts.IsAssignableFrom (ti) }"); } } // Expected Output: // B& is assignable from D& ? True // D& is assignable from B& ? False // I& is assignable from D& ? True // D& is assignable from I& ? False // I& is assignable from S& ? False // S& is assignable from I& ? False ``` * rename mono_class_signature_is_assignable No need for _checked suffix * rename mono_byref_type_is_assignable_from * [class-init] Move vtable setup to a separate file The code is getting quite long, move it to a separate file

* Remove redundant checks from PAL * Move check for AdvSimd support outside PAL * add check of SSE/SSE2 for x86/x64

Entry to crash reporting is now more properly handled by mono_dump_start() Co-authored-by: alexischr <alexischr@users.noreply.github.com>

* [wasm] Re-enable System.Threading.Tasks System.Threading.Channels Microsoft.Extensions.FileProviders.Physical * [wasm] Disable Serialization_Roundtrip in System.Threading.Channels.Tests Co-authored-by: Mitchell Hwang <mitchell.hwang@microsoft.com>

- Delete copy&paste code from CoreCLR that is only relevant for IL verification - Delete early out for CORINFO_ACCESS_ADDRESS. There is no check like that in old crossgen. Reduces number of methods that fail to compile in CoreLib significantly. - Ship internal calls without throwing and catching exceptions. Fixes #32663

Fix #38452

1. Don't mark fields of dependently promoted structs as untracked. 2. Remove some stores whose lhs local has a ref count of 1 when running late liveness. We can rely on ref counts since they are calculated right before the late liveness pass. 3. Remove dead GT_STORE_BLK in addition to GT_STOREIND in the late liveness pass. 4. Remove dead stores to untracked locals in the late liveness pass. 5. Allow optRemoveRedundantZeroInits to remove some redundant initializations of tracked locals. Move the phase to right after liveness so that SSA is correct after removing assignments to tracked locals.

This change allows inlining of generic dictionary lookups when the caller and callee are the same exact type and instantiation. Example: ``` class G<T> { static bool M1() => typeof(T); object M2() => new T[1]; bool M3() => Unsafe.SizeOf<T>(); static string M() { // Assume that T is string // All M1, M2, M3 are inlined after this change, but were not inlined before this change M1(); M2(); M3(); // Still not inlined - different instantiation, not the same exact type G<List<T>>.M1(); // Still not inlined - different type. OtherG<T>.M(); } } ```

…e.CoreLib and expose via System.Runtime. (#38473) * Move the calling convention types used in the new calling convention modopt-based encoding down to System.Private.CoreLib/System.Runtime. * Make S.R.CS.VisualC a partial facade so the tooling will generate the type forwards in the src assembly.

RyuJit would not inline methods that contained delegate invokes. Remove this limitation. Closes #10048. See also #37941.

…8311) * Adjust stack level * Reenable NarrowDouble() test * formatting * review comments * do not assert when isEBPbased=true * Add test coverage * updated the comment inside test

* [mini] Fix and testcase for mono/mono#19964 When we call a delegate that was bound to a specific function, don't do a virtual call. * [interp] Get interp method from delegate method_ptr if method is unset

…s (#36314) This helps the Windows 10 v1903 SDK installer run under wine-mono. It uses _AppDomain::CreateInstance to retrieve an ObjectHandle, and then tries to call _ObjectHandle::Unwrap() using the auto-dual interface. With these patches that call succeeds and it gets further, though it does not completely run (likely due to insufficiency in wine-mono's wpfgfx). This patch set does not implement accessor methods for public fields, mostly because the necessary bits are missing in MonoMethodBuilder, and plumbing them seems decidedly nontrivial. Co-authored-by: zfigura <zfigura@users.noreply.github.com>

* Introduce EventSource.IsSupported feature switch This allows for all of the EventSource logic to be trimmed by the ILLinker when the feature switch is set to false. Fix #37414 * Add test when EventSource IsSupported feature switch is disabled. * Update RemoteExecutor version and enable EventSource test

Adds a synchronous callback mode to EventPipe

* Fix regression in float/double sorting Several months back we moved the sorting logic for primtive types out of native code into managed. Doing so helped to make the logic reusable for spans and helped to reduce GC latency, and also actually helped with throughput in a variety of cases. But it ended up regressing throughput for sorting larger arrays of floating-point values, with float/double.CompareTo not getting inlined, and even if it were inlined, containing much more logic than was present in the native implementation. The native implementation did a pre-pass to move all NaNs to the front and then just used simple < and > comparison operations, so the managed implementation now does as well. * Special-case Half * Move Less/GreaterThan to each type that needs them

However, don't set the `COMPlus_GCStress` variable for non-Windows (bash) scripts, because we don't know if we'll be running on a platform that supports GCStress (such as macOS, which currently doesn't support it). All Windows platforms do always support GCStress in our test runs.

In a few MemoryCache hot paths, we're doing a syscall to get the time twice. It's far from ideal since the cost is a significant portion of the overall time. This is step 1, minor and non-breaking for ~30% gains in performance. Co-authored-by: Eric StJohn <ericstj@microsoft.com>

Tracked by dotnet/runtime#35884

* Arm64: support contained `GT_LCL_VAR/FLD_ADDR` in indir. Support contained `GT_LCL_VAR_ADDR, GT_LCL_FLD_ADDR` under `IND, STOREIND`. * XARCH: Support contained `GT_LCL_FLD_ADDR` in IND. Support contained `GT_LCL_FLD_ADDR` under `GT_STORE_IND, GT_IND`. * Clear `GTF_IND_ASG_LHS` after Rationalize for STORE_OBJ/BLK. We were doing this for `STOREIND`, but forgetting for `STORE_OBJ/BLK`. * Extract `LowerStoreIndirCommon`, `LowerIndir`. Extract without changes, I will need to add additional calls to them later. * Call the extracted funtions. Fix a few regressions with `NoRetyping`. * Create `LEA` of comples addr expr. Gives a few improvements when addr has an index. * Extract `LowerBlockStoreCommon`. * Tranform STORE_BLK/OBJ into STOREIND. When it is possible and profitable. * Don't tranform for GC and small types. For GC types we were not trying to contain addr(why?) when needed a barrier and it looked dangerous for me to do such a change(for 5.0). For small types we were generating `movzx rax` for `IND byte` instead of `mov ax` . * Update src/coreclr/src/jit/lower.cpp Co-authored-by: Carol Eidt <carol.eidt@microsoft.com> * Use `TODO-CQ`. * Extract `Rationalizer::RewriteIndir`. * Fix code review. Co-authored-by: Carol Eidt <carol.eidt@microsoft.com>

…312)

* [wasm] Enable the full gc by default * fix the condition * rename argument

Tracking: dotnet/runtime#38524

This increases the platform matrix tested by: gcstress-extra.yml gcstress0x3-gcstress0xc.yml r2r-extra.yml These pipelines each run twice weekly: once on Saturday, once on Sunday.

* make all-inclusive coreclr_static.lib and link into singlefilehost * Disable superhost linking on windows for now (issues with mixed Debug-on-Release builds) * shim coreclr policy between singlefile and standalone hosts * enable exports on singlefilehost executable * export __progname, perhaps it will make FreeBSD happy. * add environ too * mergeable wks and jit * ignore empty clr_path in single_file_bundle * hide R2R specific assert behind R2R ifdef * produce separate obj files from platform specific asm files for mergeable WKS - to make CMake happy * remove libcoreclr and libclrjit from netcoreapp/pkg * temporary disable setting up GS cookie on OSX superhost * renamed clrjit_static folder dll->static * extract JIT dllmain.cpp and eliminate clrjit_mergeable * make coreclr_static dependent on def/exports files instead of coreclr. * reducing double-build in cee_wks to just few files. Most of cee_wks is built once as an obj library and reused. * bring back cee_wks_mergeable and push shared object lib stuff lower to cee_wks_core * A comment about FreeBSD-specific workaround. * move finding unwind libs to a CMake function * since we have files compiled differently between embedded and standalone coreclr, we do not need EMBEDDED_RUNTIME marker. * renamed host_resolver_t.h and host_resolver_t.cpp to not have "_t" * path separating slash when forming CORECLR_STATIC_LIB_LOCATION

Two tests are failing due to delays being too small for background actions completion when test VM perf degrades. PR increases those delays.

This is an enhanced version of `mono_exception_get_managed_backtrace()` that takes an optional `const char *prefix` argument (to indent the stack-trace) and returns FALSE with a NULL backtrace on error. Co-authored-by: baulig <baulig@users.noreply.github.com>

* Remove VS 16.6 Preview 2 requirement * Clarify wording on Visual Studio community edition

…f not used in calling app (#38458) * JsonSerializer - Prevent Stack and Queue types from being pulled in if not used in calling app * Add comment about type fetch indirection * Load types based on target frameworks

…items (#38466) * remove unused usings and remove #nullable declarations * Wire half-precision float support to System.Half * address feedback * add test for ReadSingle handing double-precision encodings

…Send (#38502) * Reimplement CertificateAssetDownloader.DownloadAsset with HttpClient.Send * Address PR feedback

- Remove names pointing to non-existent members - Fix spaces in method signatures which break matching rules - Remove duplicate entries - Format file for consistency

…. (#38375)

`ResolveTokenInScope` now gets called with `typeOrMethodContext` that is a non-generic method. Make sure the method is on the same type.

Since the entire set of System.Security.Cryptography.* tests has been disabled on WASM recently in dotnet/runtime#37723, we need to let everyone know that those areas are not quite supported on WASM at the moment. In that case System.Security.Cryptography classes could throw PlatformNotSupportedException until some solution comes (dotnet/runtime#37669)

…38470) Co-authored-by: Mitchell Hwang <mitchell.hwang@microsoft.com>

The HostFileChangeMonitor relies on FSW, which is not supported on wasm. This change disables the tests that were failing because of it.

… (#38564) Resolves dotnet/runtime#37400 The regression test was not failing in Crossgen2 because somewhere along the way when we eyeball ported `getCallInfo` the ordering of these two if block was switched, so Crossgen2 was always treating such calls as non-virtual.

Addresses an issue that caused the testResult.xml to be empty on Helix runs. Adds options to only execute a specific class/method to the WASM runner.

…ash() for s390x (#38428) * Fix exception-s390x so mono_handle_native_crash() is called according to the changes made in mono/mono#19973 * Make roslyn the default compiler for s390x Co-authored-by: nealef <nealef@users.noreply.github.com>

On these tests the arm class layout happens to match the x64/arm64 layout and is different than the x86 layout. So, introduce arm specific test projects for these tests. Also in the recent test build conversion one project was overlooked, so update that project to work with this new scheme as well. Remove all exclusions for these and older variants of these tests. Closes #37470.

The exclude list entry path was incorrect

Disable the test due to non-deterministic failure in the CI. Enabling tracked by: #38013 Fixes: #35068

* Generate tables of signatures/transition functions. * Add mono_wasm_get_interp_to_native_trampoline () function to look up the transition function for a signature. * Cache the transition function in a imethod->method_data slot. Previously, we were calling a mono_wasm_interp_to_native_trampoline () function and a generated icall_trampoline_dispatch () function which added a lot of overhead.  Co-authored-by: vargaz <vargaz@users.noreply.github.com>

…#38454) mono/mono#19794 broke on iOS. Analyzing the result showed that, while the assembly was generated as expected, the linked result was different. Analyzing the LLVM code (which triggered the original fix) seems to imply that, on Apple platforms, they will not generate a label for it, only doing the 0-offset as we were doing before. Therefore, match the LLVM behavior by bringing back our previous logic when targetting Mach, otherwise use the new way of doing things. Fixes mono/mono#8806 (again) Co-authored-by: mathieubourgeois <mathieubourgeois@users.noreply.github.com>

It OOMs due to excessive allocations

…389) It does the following: 1. Upgrades the base image versions of HttpStress and SslStress tests containers to `nightly/sdk:5.0` 2. Wraps over `docker-compose` call to prevent it from accidentally reporting progress to `stderr `instead of `stdout` and thus breaking CI pipeline (see docker/compose#7346)

…Mono. (#37756) * Port EventPipeBuffer. * Port EventPipeBufferList. * Port EventPipeManager. * Merge internals source files. Changed GETTER_SETTER to only mandate use of inlined functions when calling betweeen source files for different types. Upgraded all use within each source file for direct struct access. * Changed function entry error checkin strategy. Only apply active checks on outer library API, ep.h, and use EP_ASSERT for all other functions. * Dropped requires_lock_held in function naming, switch to comment in header or in forward declare for functions that validates that lock is held. * Removed currently unused GETTER/SETTERS. * Moved more comments into sources from corresponding CoreCLR EventPipe sources. * Add EventPipeBuffer/EventPipeBufferManager unit tests. * Align to changes in native CoreCLR eventpipe* sources. * Add support starting default file session using env variables. * Disable eventpipe callback on WASM due to platform issues with delegates.

…(#38554) Co-authored-by: Mitchell Hwang <mitchell.hwang@microsoft.com>

* [interp] Add stfld opcode for storing valuetypes without refs * [interp] Remove MINT_VTRESULT use with byrefernce ctor intrinsic

…d parallelism (#38617)

* Added IsOppositeOfPrevMov() Added IsOppositeOfPrevMov() that will skip generating redundant mov. * review comments * Refactor code into IsRedundantMov() * dont optimize beyong IG boundary * add check to not cross IG boundary * Add more checks: - Change IG boundary check to exclude extended IGs. - Check the operand size before removing redundant movs. * fix the check * review comments * Add check to ensure if src/dst are both scalar or both vector * minor comment * Add check for insFormat * fix the typo * minor comments

* Update glossary.md Remove unused markdown definitions introduced in dotnet/corefx#38103 * Update source-generator-pinvokes.md Remove unused markdown definitions introduced in dotnet/runtime#33742

System.Decimal fields did not match Win32 DECIMAL type for historic reasons. It required type loader to have a quirk to artificially inflate System.Decimal alignment to match the alignment of Win32 DECIMAL type to make interop work well. This change is fixing the System.Decimal fields to match Win32 DECIMAL type and removing the quirk from the type loader since it does not belong there. The downsides are: - Slightly lower code quality on 32-bit platforms. 32-bit platforms are not our high performance targets anymore. - Explicit implementation of ISerializable is required on System.Decimal for binary serialization compatibility. Fixes #38390

…#38313) * Add AsyncCausalityTracer * cleanup * Handle multiple sessions * CR feedback * Remove AsyncCausalityTracer and replace it with direct calls to TplEventSource

- New command line parameter --verify-type-and-field-layout - Intended for use diagnosing bugs in crossgen2 only, not for general production use - New r2r precode fixups Verify_TypeLayout and Verify_FieldOffset - These allows the compiler to specify that a particular type/field layout must match with the runtime calculation. - If there is a mismatch, the runtime will failfast, even on release builds - Add use of new switch to crossgen2smoke test - Fix encoding bug in FieldFixupSignature. Check_FieldOffset and Verify_FieldOffset use unsigned offsets, not signed offsets

Fix issue link for wasm specific limitation.

…en feature switch is present (#38397) * Adding generation of substitution files for trimming out resources when feature switch is present * Adding support for CoreLib as well and fix issue with SR returning empty string * Cache result of AppContext Switch lookup * Fixing allConfigurations leg and addressing feedback * Fixing thread safety issue by removing caching * Addressing feedback * Apply suggestions from code review Co-authored-by: Jan Kotas <jkotas@microsoft.com> * Addressing feedback and adding tests * Fix typo Co-authored-by: Jan Kotas <jkotas@microsoft.com>

The default 10 is not enough stack frames to make sense of failures

- Disallow Encoding.GetEncoding("utf-7", ...) from returning an Encoding instance - Minor code cleanup by removing unused ifdefs - A compat switch is available to re-enable Encoding.GetEncoding("utf-7", ...)

S.S.C.Pkcs.ContentInfo..ctor(byte[]) didn't get spanified because it holds a reference to the array, and returns it via a property.

* Add exclusion category argument to interpreter runs * Add for Windows as well

…38602)

Fixes mono/mono#20049 Co-authored-by: thaystg <thaystg@users.noreply.github.com>

Fixes #38623

* Enable w35038 on VS. * fix for a contract.h * FIx other places. * another fix.

* Adding default string for stacktrace Word_At * Also adding defaults to other resources used on the stacktrace ToString method * Apply suggestions from code review Co-authored-by: Eric Erhardt <eric.erhardt@microsoft.com> Co-authored-by: Eric Erhardt <eric.erhardt@microsoft.com>

…signed (#38371) * ExtractNarrowingSaturateLower and ExtractNarrowingSaturateUpper * ExtractNarrowingSaturateUnsignedLower and ExtractNarrowingSaturateUnsignedUpper * ExtractNarrowingSaturateScalar and ExtractNarrowingSaturateUnsignedScalar

* Block support for a P/Invoke marked with UnmanagedCallersOnly. * Update crossgen2.

`fgComputeLifeUntrackedLocal` can be called multiple times for the same `lclVarNode` since `fgLocalVarLiveness` may call `fgInterBlockLocalVarLiveness` multiple times. Therefore, `m_promotedStructDeathVars` may be updated more than once for the same node. Fixes #38496.

R2R doesn't generate PInvoke stub for arm because of crossbitness issue. This patch enables PInvoke stub compilation for R2R, and updates native size, field offset, and so on based on target machine's bitness.

* [browser][wasm][tests] Add tests for TypedArrays * Address review comments

For Windows, add clrcompression.dll to the package. For Linux, add { libSystem.IO.Compression.Native.so, libSystem.Native.so, libSystem.Net.Security.Native.so, libSystem.Security.Cryptography.Native.OpenSsl.so } to the package and also remove the corresponding static libraries with the .a extension from crossgen2.deps.json.

* Require struct handler to be set after for CSE. * Reenable the test. * fix for empty cache * Update the comment.

…when -s WASM_OBJECT_FILES=0 is used. (#38370)

* Add ExcludeFromCodeCoverage.Justification property Delete some ExcludeFromCodeCoverages that aren't warranted, and add justifications almost everywhere else (only not doing so where a project builds for targets without the member). * Update src/libraries/Microsoft.CSharp/src/Microsoft/CSharp/RuntimeBinder/Semantics/Types/PointerType.cs Co-authored-by: Levi Broderick <GrabYourPitchforks@users.noreply.github.com> Co-authored-by: Levi Broderick <GrabYourPitchforks@users.noreply.github.com>

And use a testsuite that passes for the example

…o utf8 (#38641) * Change default output encoding to utf8 for wasm

…n that were fixed (#38658) Closes dotnet/runtime#32744

Fix #38455

There is a small breaking change in the SDK about how we run the XHarness tool - we are no longer installing the full .NET SDK to install the tool on every agent but rather installing it on the build machine and then shipping it as-is as Helix correlation payload. We then only need the .NET runtime to be installed on the Helix agents saving us bandwidth and cycles

…ers (#38646) * [mono] Implement getting the type modifiers for DynamicMethod parameters Fixes dotnet/runtime#36271 * Remove unused parameter from type_array_from_modifiers()

* Make Common.Tests pass * Remove Common test project from the exclusion list

* add TargetHostName to SslStream * fix unit tests * feedback from review

* move the interface check to be before versionbubble * updating fix to check for TypicalMethod of the target * Adding a comment to clarify the check * uncoditinoally invoke GetTypicalMethodDef * Add a smoke test. * moving the interface check to before version bubble Also adding a cross module test.

- Mimic the alignment and other algorithms present in coreclr - Add tests to crossgen2smoke to cover these scenarios - Update type system unit tests so that they can run in VS - Fix issue #38450

* Add platform-specific attributes Spec dotnet/runtime#33331 * Convert to xml doc * Update src/libraries/System.Private.CoreLib/src/System/Runtime/InteropServices/MinimumOSAttribute.cs Co-authored-by: Jeremy Barton <jbarton@microsoft.com> * Update src/libraries/System.Private.CoreLib/src/System/Runtime/InteropServices/ObsoletedInPlatformAttribute.cs Co-authored-by: Jeff Handley <jeffhandley@users.noreply.github.com> * Address code review * Add to ref assembly, test and fix build errors * Fix spacing, revert unwanted changes * Namespace was wrong, updated Co-authored-by: Jeremy Barton <jbarton@microsoft.com> Co-authored-by: Jeff Handley <jeffhandley@users.noreply.github.com> Co-authored-by: Buyaa Namnan <bunamnan@microsoft.com>

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Runtime ingestion #13

Runtime ingestion #13

Commits on Jun 21, 2020

Commits on Jun 22, 2020

Commits on Jun 23, 2020

Commits on Jun 24, 2020

Commits on Jun 25, 2020

Commits on Jun 26, 2020

Commits on Jun 27, 2020

Commits on Jun 28, 2020

Commits on Jun 29, 2020

Commits on Jun 30, 2020

Commits on Jul 1, 2020

Commits on Jul 14, 2020