| Age | Commit message (Collapse) | Author | Files | Lines |
|---|
|
* Disable running ILAsm round-trip build on COM (requires ilasm resource embedding support).
* ILAsmTestKind -> IlrtTestKind
|
|
- We can use our knowledge of object representation in the runtime to speed up type checks.
- We leave the ref T and the length deconstructed until the very end, optimizing register usage.
- The Length property getter is once again just a simple field accessor with no bitwise logic.
|
|
Fix SP check for x64/x86, remove for arm32/arm64
|
|
|
|
|
|
With this change, the JIT will recognize a call to BinaryPrimitives.ReverseEndianness and will emit a bswap instruction.
This logic is currently only hooked up for x86 and x64; ARM still uses fallback logic.
If the JIT can't emit a bswap instruction (for example, trying to emit a 64-bit bswap in a 32-bit process), it will fall back to a software implementation, so the APIs will work across all architectures.
|
|
Fix ValueNumber constant propagation with nested structs
|
|
Fixes #18672
Test case is here: JIT/Regression/JitBlue/GitHub_18672/GitHub_18672.cs
|
|
PMI asm diffs changes
|
|
preview1-03402-01, preview-27102-02, preview.18552.1, preview-27102-03, master-20181102-0044, master-20181102-0041, respectively (master) (#20745)
* Update BuildTools, CoreClr, CoreFx, CoreSetup, IbcData, PgoData to preview1-03402-01, preview-27102-02, preview.18552.1, preview-27102-03, master-20181102-0044, master-20181102-0041, respectively
* Disabled outdated test
|
|
|
|
As https://docs.microsoft.com/en-us/dotnet/api/system.runtime.interopservices.gchandle.free?view=netframework-4.7.2 says "The caller must ensure that for a given handle, Free is called only once.".
So delete the second call to `Free()`.
I believe it changes the desired behaviour for this test but it was added before 2010 (via source depo) and there are no comments about the regression that it was able to repro, so it is not worth to investigate that.
|
|
Fix expected Pri-1 test count
|
|
|
|
https://github.com/dotnet/coreclr/pull/20693 reduced the number of
Pri-1 tests significantly, which caused this error to appear on
some Pri-1 builds, e.g., Ubuntu x64 Checked builds.
|
|
* Respond to PMI asm diffs code review feedback
Add some try/finally code to improve robustness.
Remove direct invocation of run.cmd.
Also, simplify the generated asm directory hierarchy.
* Improve PMI script error code
Only return non-zero if there is a fatal error. For non-fatal error,
such as the existence of asm diffs, return zero. This will allow
subsequent "build" steps in netci.groovy to run, namely, the
archiving of the asm directory.
Later, determine a better way figure out how to deal with the existence
of asm diffs.
* Update dotnet SDK download links
|
|
Fix compile warnings
Fix compile warnings
Fix compile warnings
Remove duplicate tests. Move ReversePinvoke tests under StructMarshalling.
Clean up some compile errors.
More string cleanup.
More string fixes
Clean up consts and typedefs.
Fix typo.
Change TP_slen return type to match xplatform.h
Fix printf specifiers.
Don't define wcslen in xplatform if it's been macro redefined (to platformdefines.h's TP_slen)
Fix build breaks.
Remove extension from DllImport so it will correctly resolve xplat.
Fix printf specifiers again...
Fix Windows x86 calling convention mismatch.
Fix HRESULT size off-Windows
|
|
* Add PInvoke/NativeCallManagedComVisible tests
* Update to current infrastructure and disable off-Windows
* Fix incorrect S_OK value and made the native side actually return an HRESULT in hr.
|
|
Arm32: Handle live-out double reg in resolution
|
|
* Re-Enable windows with JitOptSensitive flag
The check in batch targets would keep us from running jitOptimiztionSensitiveTests
with COMPlus_TieredCompilation=0.
* Address pr feedback
|
|
In order to mitigate System.IO.FileLoadException we built our own xunit.console.dll with ConcurrentDictionary used for DependencyContextAssemblyCache.managedAssemblyCache instead of Dictionary and use this instead of the one pulled from NuGet.
|
|
When handling critical edges, live-in and live-out reg masks must contain both halves of double regs.
|
|
Move SSE/SSE2/AVX Load* intrinsic tests to the template framework
|
|
This is the last batch of deleting tests that are obviously redundant. The remaining framework API tests in CoreCLR can be dealt with as needed if it ever becomes a problem.
Fixes #12782
|
|
* Fix CCW with large numbers of arguments on arm64
- Correctly align the stack in COMToCLRDispatchHelper
- Add tests for naturally 16 byte aligned stack growth, and non-aligned growth
* New many arguments COM test
* Add support for the IL stub ETW diagnostic event
|
|
* Add a repro test.
* Fix the issue.
* Add comments
|
|
Closes #20651.
Also fix up some "near miss" cases for GenTreeField and GenTreeBoundsCheck,
where we get lucky and the importer currently splits trees with temps so the
currently ignored child nodes have no interesting side effects.
Revise GenTreeField a bit to pull more of the initialization work into the
constructor. Add a missing R2R field propagation for field nodes in GtClone
(evidently also never hit in practice).
|
|
(#20680)
Since moving to the helix queue, we have been failing because the execution policy on the vms is not set to unrestricted. To get around this, we need to explicitly pass -ExecutionPolicy Bypass.
|
|
* Interop/COM/NETClients/Primitives/NETClientPrimitives
* Interop/COM/NativeClients/Primitives
* Interop/COM/NETClients/IDispatch/NETClientIDispatch
* Interop/COM/NETClients/Aggregation/NETClientAggregation
|
|
If the jit decides it needs a return spill temp, and the return value
has already been spilled to a single-def temp, re-use the existing
for the return temp rather than creating a new one.
In conjunction with #20553 this allows late devirtualization for calls where
the object in the virtual call is the result of an inline that provides
a better type, and the objected formerly reached the call via one or more
intermediate temps.
Closes #15873.
|
|
|
|
Fixes failing test obsoleted by #20638 and contributes to #12782
Delete redundant CoreMangLib decimal tests
|
|
Add a run time check for Windows.Nano to all COM related tests.
|
|
|
|
* New Decimal.Remainder implementation
* Update CoreFX.issues.json
|
|
* Implement interface dispatch to COM for ARM64
- Implement missing stub (GenericComPlusCallStub)
* Re-enable IDispatch test for Arm64 (Fix issue #20580 )
|
|
Implement AVX2.BroadcastScalarToVector128/256
|
|
Disable b152292 for arm32 also
|
|
|
|
Stop attempting to move PSPSym during localloc for arm32/arm64
|
|
For arm64, this was creating a data corruption possibility, as the
calculations were done wrong.
In neither case was it necessary, as in the main function we always
access the PSPSym FP-relative, so when writing back the PSPSym, we
were always writing to exactly the same slot; nothing was actually
moving.
I can't recall or figure out why we were moving it in the first place,
or what might have changed.
Added a new variant of the localloc eh tests that has outgoing arguments,
to exercise the case where we need to re-establish the outgoing argument
space after the localloc.
|
|
|
|
|
|
Avx.Broadcast* tests to the template
|
|
|
|
|
|
|
|
|
|
* Avoid comparing capacity in StringBuilder.Equals
* Disabled corefx StringBuilder test
|
|
Disable for now while failures are investigated. See #20580.
|
