Brian Sullivan [Wed, 14 Feb 2018 22:03:27 +0000 (14:03 -0800)]
Merge pull request #16367 from briansull/vso-566984
Fix for CSE - Assertion failed '!"Badly linked tree"'
Michal Strehovský [Wed, 14 Feb 2018 21:04:47 +0000 (22:04 +0100)]
Mark sharedgenerics test as GCStressIncompatible (#16391)
Bruce Forstall [Wed, 14 Feb 2018 20:11:49 +0000 (12:11 -0800)]
Merge pull request #16380 from kbaladurin/fix-uaf
Histogram: doesn't allocate additional memory to store counts
Tarek Mahmoud Sayed [Wed, 14 Feb 2018 18:54:14 +0000 (10:54 -0800)]
Fix the fired asset in CompareInfo.IndexOf (#16373)
Koundinya Veluri [Wed, 14 Feb 2018 18:27:32 +0000 (10:27 -0800)]
Implement WaitHandle.SignalAndWait on Unix (#16383)
Part of fix for https://github.com/dotnet/coreclr/issues/10441
Sergey Andreenko [Wed, 14 Feb 2018 17:39:21 +0000 (09:39 -0800)]
Add VSD additional param for DIRECT calls. (#16267)
* add a virtualStubParamInfo as an argument
* try to delete duplicates
expect many failures for arm in lower because we delete: // Change the call type, so we can add the extra indirection here, rather than in codegen
* Revert legacy workaround from lower.
* Fix GetNonStandardAddedArgCount
Konstantin Baladurin [Wed, 14 Feb 2018 07:46:19 +0000 (10:46 +0300)]
Historam: doesn't allocate additional memory to store counts
It allows to remove deallocation in destuctor that can lead to uaf
for static Historgam's objects during shutdown.
David Mason [Wed, 14 Feb 2018 10:18:39 +0000 (02:18 -0800)]
Fix for issue #13153 - change SetGCConcurrent to TemporaryDisableConcurrentGC for profiler startup path. (#16314)
Michal Strehovský [Wed, 14 Feb 2018 10:16:09 +0000 (11:16 +0100)]
Handle ELEMENT_TYPE_PTR in ILCodeStream::EmitLDIND/EmitSTIND (#16381)
Fixes #16371.
Jan Vorlicek [Wed, 14 Feb 2018 09:15:19 +0000 (10:15 +0100)]
Fix ARM/ARM64 hijacking in tail calls (#16039)
* Fix ARM/ARM64 hijacking in tail calls
This change fixes an issue that can happen when a function that has tail
calls is hijacked. There are two potential issues:
1. When a function that tail calls another one is hijacked, the LR may be
stored at a different location in the stack frame of the tail call
target.
So just by performing tail call, the hijacked location becomes invalid and
unhijacking would corrupt stack by writing to that location.
2. There is a small window after the caller pops LR from the stack in its
epilog and before the tail called function pushes LR in its prolog when
the hijacked return address would not be not on the stack and so we would
not be able to unhijack.
The fix is to prevent hijacking of functions that contain tail calls.
* Enable the tailcall hijacking test for ARM64
The test JIT/Methodical/tailcall_v4/hijacking should be passing now on
ARM64.
Ahson Khan [Wed, 14 Feb 2018 02:18:46 +0000 (18:18 -0800)]
Add ReadOnlySpan string-like StartsWith and EndsWith API with globalization support (#16223)
* Add ReadOnlySpan string-like StartsWith API with globalization support
* Add ReadOnlySpan string-like EndsWith API with globalization support
* Fix nits from PR feedback.
* Address PR feedback
* Add unsafe keyword in CompareInfo.Unix methods.
* Address PR feedback and fix signature of Interop.Globalization.EndsWith
* Code formatting cleanup as suggested by VS.
* Revert "Code formatting cleanup as suggested by VS."
This reverts commit
20931fbf1ae113c6acaa3a4b2c0789b46d7ab06b.
* Creating Span specific methods and working around use of IsFastSort
* Code cleanup
* Improve performance by avoiding slicing and fix impl bug
* Remove commented out code and revert some calls to FindString
* Updating checks for Unix to mimic what IsFastSort does.
* Add length checks and include IsFastSort check within the loop
* Fix bounds checks to not include 0x80
* Call correct method when StringComparison is IgnoreCase.
* Update exception type in string compare to remain consistent
Brian Sullivan [Wed, 14 Feb 2018 00:17:43 +0000 (16:17 -0800)]
Fix clang format warnings
Brian Sullivan [Tue, 13 Feb 2018 23:48:38 +0000 (15:48 -0800)]
Check for the very uncommon case where when we are replacing a CSE use
and it contains a CSE def which itself has persistent side effects.
For such a case we simply abandon the replacement of the CSE use.
Jacek Blaszczynski [Tue, 6 Feb 2018 20:24:59 +0000 (21:24 +0100)]
Add Sse2 scalar hardware intrinsics tests
Jacek Blaszczynski [Sat, 27 Jan 2018 21:18:05 +0000 (22:18 +0100)]
Implement scalar Sse2 hardware intrinsics
Brian Robbins [Tue, 13 Feb 2018 22:34:31 +0000 (14:34 -0800)]
Fix registry path for ETW argument lookup. (#16347)
Wes Haggard [Tue, 13 Feb 2018 19:16:50 +0000 (11:16 -0800)]
Merge pull request #16344 from brunotag/patch-1
Update UsingYourBuild.md
Carol Eidt [Tue, 13 Feb 2018 15:42:38 +0000 (07:42 -0800)]
Merge pull request #16351 from fiigii/fixcont
Remove an incorrect assertion in containment analysis
Carol Eidt [Tue, 13 Feb 2018 15:34:34 +0000 (07:34 -0800)]
Merge pull request #15950 from dotnetrt/isafix
On all platforms first member of InstructionSet enum is InstructionSet_ILLEGAL
Fei Peng [Tue, 13 Feb 2018 07:42:34 +0000 (23:42 -0800)]
Fix an incorrect assertion in containment analysis
Michal Strehovský [Tue, 13 Feb 2018 07:19:30 +0000 (08:19 +0100)]
Fix ambiguous interface method error message (#16341)
Fixes #15864.
Michal Strehovský [Tue, 13 Feb 2018 04:08:29 +0000 (05:08 +0100)]
Define type initialization semantics for interface instance methods (#16340)
* Define type initialization semantics for interface instance methods
Accessing an instance method on an interface should trigger non-beforefieldinit class constructors (same as it does for valuetypes).
Fixes #15650.
Andy Ayers [Tue, 13 Feb 2018 03:31:51 +0000 (19:31 -0800)]
Improve perf of DelegateConstruct (#16345)
Remove a range lookup that's no longer needed.
See related issue #12438.
Fei Peng [Mon, 12 Feb 2018 21:51:32 +0000 (13:51 -0800)]
Fix SSE4.1 encoding
Fei Peng [Mon, 12 Feb 2018 20:51:31 +0000 (12:51 -0800)]
Implement SSE4.2 CompareGreaterThan
Fei Peng [Mon, 12 Feb 2018 20:50:51 +0000 (12:50 -0800)]
Implement SSE4.1 CompareEqual
Fei Peng [Mon, 12 Feb 2018 20:49:59 +0000 (12:49 -0800)]
reenable AVX2 compareEqual.Int64
Bruce Forstall [Tue, 13 Feb 2018 00:26:25 +0000 (16:26 -0800)]
Merge pull request #16181 from Buyduck/arm-memory-barrier
[RyuJIT/arm32] Support GTF_IND_VOLATILE and GTF_BLK_VOLATILE
Carol Eidt [Mon, 12 Feb 2018 23:00:30 +0000 (15:00 -0800)]
Merge pull request #16008 from sdmaclea/PR-ARM64-TEST-SIMD-HWINTRINSIC
[Arm64] Add Simd HWIntrinsic test
Matt Galbraith [Mon, 12 Feb 2018 20:56:46 +0000 (12:56 -0800)]
Merge pull request #16309 from MattGal/eol-some-linuxes
Remove EOL OSes and add new / missing ones
Bruno Tagliapietra [Mon, 12 Feb 2018 19:35:22 +0000 (08:35 +1300)]
Update UsingYourBuild.md
Apparently it needs a PlatformTarget to be specified as well when targeting x64 architecture
Andy Ayers [Mon, 12 Feb 2018 19:30:45 +0000 (11:30 -0800)]
JIT: Allow struct promotion in GS frames (#16220)
Enable struct promotion in methods where the jit is also emtting a stack guard.
Not allowing promotion seems like an unnecessary limitation and leads to poor
performance in methods with both Span and stackalloc.
Closes #16197.
Brian Robbins [Mon, 12 Feb 2018 19:15:40 +0000 (11:15 -0800)]
Use DateTime.ToFileTimeUtc to generate serializable DateTime for TraceLogging events. (#16317)
Steve MacLean [Thu, 1 Feb 2018 17:37:30 +0000 (12:37 -0500)]
Add TestInsert
Steve MacLean [Thu, 8 Feb 2018 17:26:47 +0000 (12:26 -0500)]
Improve diagnostics
Steve MacLean [Sat, 16 Dec 2017 01:04:06 +0000 (20:04 -0500)]
[Arm64] Add Simd HWIntrinsic test
Andy Ayers [Mon, 12 Feb 2018 17:16:53 +0000 (09:16 -0800)]
JIT: look for escaping byrefs (#16305)
If a byref is passed to a call or used in an assign, assume it escapes and
that subsequent dereferences may be made at arbitrary offsets.
This handles cases where a user takes the address of one field of a struct,
then uses that byref that to access other fields or non-field content within
the struct, and the byref formation and use are split across statements or
across caller-callee contexts.
Closes #16210.
dotnet-maestro-bot [Mon, 12 Feb 2018 12:26:26 +0000 (06:26 -0600)]
Update CoreClr, CoreFx, PgoData to preview2-26212-01, preview2-26212-01, master-
20180212-0035, respectively (#16335)
Jan Kotas [Mon, 12 Feb 2018 07:00:55 +0000 (23:00 -0800)]
Improve CMake detection on Windows when not in PATH (#16328)
Port dotnet/corert#5372
curl https://github.com/dotnet/corert/commit/
b723f90c611a2c79a0921f95c7299ebb7325eb59.patch | git am -p 3 --directory='src/pal/tools/' --reject
In CMake v10.2, the key `hklm:\SOFTWARE\Kitware` returns:
```powershell
Hive: HKEY_LOCAL_MACHINE\SOFTWARE\Kitware
Name Property
---- --------
CMake InstallDir : C:\Program Files\CMake\
```
with no space after `CMake` and property name `InstallDir`,
instead of `'(default)'`.
Koundinya Veluri [Mon, 12 Feb 2018 00:57:09 +0000 (16:57 -0800)]
Fix stack trace population to get proper source/line info for tier 1 methods (#16302)
Fixes https://github.com/dotnet/coreclr/issues/16224
dotnet-maestro-bot [Sun, 11 Feb 2018 16:34:54 +0000 (10:34 -0600)]
Update CoreClr, CoreFx to preview2-26211-03, preview2-26211-04, respectively (#16327)
Jan Kotas [Sun, 11 Feb 2018 16:34:43 +0000 (08:34 -0800)]
Delete PEFingerPrint.* (#16325)
Unnecessary dummy type on CoreCLR
dotnet-maestro-bot [Sun, 11 Feb 2018 13:07:04 +0000 (07:07 -0600)]
Update BuildTools, CoreClr to preview2-02511-02, preview2-26210-06, respectively (#16324)
dotnet-maestro-bot [Sat, 10 Feb 2018 15:27:20 +0000 (09:27 -0600)]
Update CoreClr to preview2-26210-04 (#16321)
Andy Ayers [Sat, 10 Feb 2018 11:57:44 +0000 (03:57 -0800)]
Allow jit internal errors to propagate to user visible exception (#16320)
May help users find workarounds for cases where the jits hit some
kind of implementation limit, as in #12480 or #14762.
dotnet-maestro-bot [Sat, 10 Feb 2018 11:27:03 +0000 (05:27 -0600)]
Update PgoData to master-
20180210-0042 (#16319)
Fei Peng [Thu, 8 Feb 2018 23:08:22 +0000 (15:08 -0800)]
Implement certain simple SIMD intrinsics for AVX/AVX2
Fei Peng [Fri, 9 Feb 2018 10:18:12 +0000 (02:18 -0800)]
Add AVX/AVX2 templete-drien tests and enable containment tests
Fei Peng [Thu, 8 Feb 2018 23:05:06 +0000 (15:05 -0800)]
Add new test template
Fei Peng [Thu, 8 Feb 2018 23:03:22 +0000 (15:03 -0800)]
Add test cases for some AVX/AVX2 intrinsics
dotnet-maestro-bot [Sat, 10 Feb 2018 04:58:29 +0000 (22:58 -0600)]
Update BuildTools, CoreClr, CoreFx, PgoData to preview2-02509-01, preview2-26209-08, preview2-26210-01, master-
20180209-0035, respectively (#16269)
Tanner Gooding [Fri, 9 Feb 2018 17:12:43 +0000 (09:12 -0800)]
Updating emitxarch to check TakesVexPrefix(ins) when adding REX bits.
Sung Yoon Whang [Fri, 9 Feb 2018 23:17:07 +0000 (15:17 -0800)]
Add back TraceFormat.Normal to avoid additional newlines from being added
Sung Yoon Whang [Fri, 9 Feb 2018 20:38:28 +0000 (12:38 -0800)]
Address PR comments
Signed-off-by: dotnet-bot <dotnet-bot@microsoft.com>
Sung Yoon Whang [Fri, 9 Feb 2018 03:28:17 +0000 (19:28 -0800)]
exclude errorSource from call to Environment.FailFast on Unix
Signed-off-by: dotnet-bot <dotnet-bot@microsoft.com>
John Doe [Mon, 5 Feb 2018 22:33:29 +0000 (14:33 -0800)]
dependancy -> dependency (dotnet/corert#5350)
Signed-off-by: dotnet-bot <dotnet-bot@microsoft.com>
Eugene Rozenfeld [Fri, 9 Feb 2018 23:45:37 +0000 (15:45 -0800)]
Fix for RETURNTRAP xarch codegen. (#16292)
This fixes a bug that caused an assert in GetSingleTempReg.
In this case we had an an interop call (resulting in RETURNTRAP in its epilog)
when there are live 256 bit values at the callsite.
The bug was that codegen for RETURNTRAP node requested a single temp register
and asserted that there was only one. In this case the set of temp registers
includes floating-point registers that may be needed for saving/restoring the upper
part of 256-bit registers. The fix was for codegen to request a single temp int register.
GetSingleTempReg will assert that there was only one int register in this case.
Tanner Gooding [Fri, 9 Feb 2018 02:05:02 +0000 (18:05 -0800)]
Fixing up fgValueNumberIntrinsic to use IsMathIntrinsic rather than checking individual cases.
David Mason [Fri, 9 Feb 2018 23:06:23 +0000 (15:06 -0800)]
Debugger api to set a breakpoint on offset 0 of all methods (#16303)
* Debugger api to set a breakpoint on offset 0 of all methods
* fix ifdef misplacement in pal cordebug.h
* add IID to cordebug_i.cpp
Carol Eidt [Fri, 9 Feb 2018 22:59:54 +0000 (14:59 -0800)]
Merge pull request #16308 from sdmaclea/PR-JIT.SIMD.VectorUtil.CheckValue
Fix JIT.SIMD.VectorUtil.CheckValue<T>
Matt Galbraith [Fri, 9 Feb 2018 21:13:49 +0000 (13:13 -0800)]
EOL Removals:
Fedora 23
RedHat 7.2
Ubuntu 16.10
Additions:
OSX 10.13
Ubuntu.17.10
Ubuntu 18.04
OpenSuse 42.3
SLES 12
Steve MacLean [Fri, 9 Feb 2018 21:01:05 +0000 (16:01 -0500)]
Fix JIT.SIMD.VectorUtil.CheckValue<T>
Brian Sullivan [Fri, 9 Feb 2018 21:07:51 +0000 (13:07 -0800)]
Remove an old JIT64 thread suspend work around (#16304)
JIT64 isn't ever used for CoreCLR
Stephen Toub [Fri, 9 Feb 2018 20:31:22 +0000 (15:31 -0500)]
Fix T[]->ArraySegment cast to return default for null (#16301)
Brian Sullivan [Fri, 9 Feb 2018 19:15:06 +0000 (11:15 -0800)]
Merge pull request #16248 from briansull/fix-16065
Fixes GitHub Issue 16065
Adam Sitnik [Fri, 9 Feb 2018 17:51:41 +0000 (18:51 +0100)]
Merge pull request #16217 from adamsitnik/eventPipeAssert
fix the Incorrect EventPipe Assert, fixes #16158
Jan Kotas [Fri, 9 Feb 2018 17:45:48 +0000 (09:45 -0800)]
Merge pull request #16300 from jkotas/revert
Revert "Debugger api to set a breakpoint on offset 0 of all methods"
Jan Kotas [Fri, 9 Feb 2018 17:40:42 +0000 (09:40 -0800)]
Revert "Debugger api to set a breakpoint on offset 0 of all methods (#15819)"
This reverts commit
3d689d00843618105e735c5647e1cb64e721a333.
Jan Kotas [Fri, 9 Feb 2018 17:40:36 +0000 (09:40 -0800)]
Revert "fix for issue #16289 (#16291)"
This reverts commit
cd3f6e77757729d3d02dfbec4af21cdfe83f5435.
Koundinya Veluri [Fri, 9 Feb 2018 16:28:59 +0000 (08:28 -0800)]
Fix GC reporting for slow tail calls with by-ref-like args (#16264)
Fix GC reporting for slow tail calls with by-ref-like args
Fixes https://github.com/dotnet/coreclr/issues/16229:
- Offsets recorded in the GC layout are intended to be offsets from the top of the stack, fixed by calculating the proper offset
Tanner Gooding [Thu, 8 Feb 2018 15:57:49 +0000 (07:57 -0800)]
Change compliement to complement
Tanner Gooding [Thu, 8 Feb 2018 00:47:04 +0000 (16:47 -0800)]
Fixing up the emitOutput handling for RR/RRR instructions
Tanner Gooding [Wed, 7 Feb 2018 04:38:51 +0000 (20:38 -0800)]
Enabling the named intrinsic support on SSE4.1 hardware for Math.Round, Math.Ceiling, and Math.Floor
Tanner Gooding [Wed, 7 Feb 2018 04:38:24 +0000 (20:38 -0800)]
Updating the emitter to more generally handle 4-Byte SSE4 instructions.
David Mason [Fri, 9 Feb 2018 08:30:02 +0000 (00:30 -0800)]
fix for issue #16289 (#16291)
Brian Sullivan [Fri, 9 Feb 2018 01:57:04 +0000 (17:57 -0800)]
Merge pull request #15595 from mikedn/rc-binop
Remove useless RangeCheck::Limit::keArray
Bruce Forstall [Fri, 9 Feb 2018 01:00:10 +0000 (17:00 -0800)]
Change corefx test process (#16263)
Instead of using the `/p:CoreCLROverridePath` switch when building
corefx, to bring in coreclr, build corefx normally. This builds
against the coreclr packages that corefx has already been validated
against in the CI, so the build should always work. Then, copy
over our built coreclr bits before testing. This method is expected
to have better behavior in most breaking change scenarios.
Wes Haggard [Fri, 9 Feb 2018 00:53:25 +0000 (16:53 -0800)]
Merge pull request #16281 from alpencolt/clr-extra-args-4-cmake
[Unix] Passing extra CMake args
Bruce Forstall [Thu, 8 Feb 2018 23:25:17 +0000 (15:25 -0800)]
Merge pull request #16278 from BredPet/fix_cfi_unwind_info
[RyuJIT/ARM32] Fix crash while generating CFI unwind info
Victor "Nate" Graf [Thu, 8 Feb 2018 23:07:20 +0000 (15:07 -0800)]
Adjust expected file size if ZapDisable is true (#16240)
Jarret Shook [Thu, 8 Feb 2018 22:43:24 +0000 (14:43 -0800)]
Merge pull request #15673 from tannergooding/generatelayout-switch
Updating build.cmd to support 'generatelayout' when 'skiptests' is specified.
David Mason [Thu, 8 Feb 2018 22:35:01 +0000 (14:35 -0800)]
Debugger api to set a breakpoint on offset 0 of all methods (#15819)
* Debugger api to set a breakpoint on offset 0 of all methods
* code review feedback
* more code review feedback
* respect IsIl in CorDbFunctionBreakoint()
* change SIMPLIFYING_ASSERT to SIMPLIFYING_ASSUMPTION_SUCCEEDED for hrs, it outputs a much better diagnostic message
Noah Falk [Thu, 8 Feb 2018 21:59:26 +0000 (13:59 -0800)]
Updating runtime version for tests (#16271)
Change
93ec62347a4e97b920513296eefb27bdacbecfe9 removed Span.NonPortableCast from System.Private.CoreLib.dll requiring tests
to use an updated version of System.Memory.dll that no longer refers to it.
Brian Robbins [Thu, 8 Feb 2018 21:24:52 +0000 (13:24 -0800)]
Explictly initialize EventData.Reserved for System.Private.CoreLib. (#16283)
Anirudh Agnihotry [Thu, 8 Feb 2018 21:03:50 +0000 (13:03 -0800)]
Extra semicolons left at the end of Combine function taking 4 paths. (#16284)
Jacek Blaszczynski [Wed, 7 Feb 2018 22:29:30 +0000 (23:29 +0100)]
Add tests for Sse2 memory fence instructions
Jacek Blaszczynski [Wed, 7 Feb 2018 22:28:38 +0000 (23:28 +0100)]
Implement Sse2 memory fence instructions
Carol Eidt [Thu, 8 Feb 2018 19:43:00 +0000 (11:43 -0800)]
Merge pull request #16268 from CarolEidt/Fix16254
Check mismatched types for SIMD copy
Carol Eidt [Thu, 8 Feb 2018 17:35:31 +0000 (09:35 -0800)]
Merge pull request #16260 from sdmaclea/PR-ARM64-Remove-SIMD-hacks
[Arm64] Remove SIMD hacks to get*TypeForStruct
Alexander Soldatov [Thu, 8 Feb 2018 17:26:59 +0000 (20:26 +0300)]
[Unix] Passing extra CMake args
Added support for compiler and linker additional arguments in CMake.
By using CLR_ADDITIONAL_LINKER_FLAGS and CLR_ADDITIONAL_COMPILER_OPTIONS
it's possible to pass toolchain options without changing Makefile.
Bruce Forstall [Thu, 8 Feb 2018 16:40:00 +0000 (08:40 -0800)]
Merge pull request #16279 from BruceForstall/FixArmCorefxAbsolutePaths
Make paths used by ARM corefx test script absolute
Bruce Forstall [Thu, 8 Feb 2018 16:31:09 +0000 (08:31 -0800)]
Make paths used by ARM corefx test script absolute
Jan Kotas [Thu, 8 Feb 2018 13:48:59 +0000 (05:48 -0800)]
Cleanup redundant stale COR header definition (#16261)
Petr Bred [Thu, 8 Feb 2018 12:04:44 +0000 (15:04 +0300)]
[RyuJIT/ARM32] Fix crash while generating CFI unwind info
Signed-off-by: Petr Bred <bredpetr@gmail.com>
Maryam Ariyan [Thu, 8 Feb 2018 06:55:11 +0000 (01:55 -0500)]
Throwing IO Related exception only on FileStream.Dispose(disposing:true) (#16250)
* Making Unix and Windows FileStream Dispose behavior more consistent and throwing IO related exception only on Dispose(true)
Fixes: #26734
* Commenting out SecurityException, also providing reason
Michal Strehovský [Thu, 8 Feb 2018 06:54:39 +0000 (07:54 +0100)]
Make it possible to Reflection.Emit default interface methods (#16257)
Fixes #15648.
John Doe [Thu, 8 Feb 2018 06:05:32 +0000 (22:05 -0800)]
Agruments -> Arguments (#16273)
Brian Sullivan [Wed, 7 Feb 2018 02:27:52 +0000 (18:27 -0800)]
Fixes GitHub Issue 16065
Bug in the CSE phase when a CSE use contains a CSE def
For the error cases we could unmark or discard a CSE def. This CSE def might be needed when
performing a subsequent CSE replacement. This could lead to an uninitialized value being
used for the CSE use.
Changed optUnmarkCSE method to return a bool instead of a void.
Changed optValuenumCSE_UnmarkCSEs to take a write-back mutable wbKeepList argument
Removed an unreachable code block in fgRemoveStmt and replaced it with a noway_assert
Added method headers for optUnmarkCSE and optUnmarkCSEs
Removed the code that allowed unmarking of a CSE def, instead we return false
Retyped the second arguments to optUnmarksCSEs and optValueNumCSE_UnmarkCSEs
to support appending new nodes to keep using wbKeepList
Code modifications to PerformCSE() to support collecting both persistent side effects and
any nested CSE defs found in CSE uses. Move the location of the call to
optValueNumCSE_UnmarkCSEs so that it comes before the decision to create a GT_COMMA node
for the cse use.
Carol Eidt [Thu, 8 Feb 2018 01:31:34 +0000 (17:31 -0800)]
Check mismatched types for SIMD copy
If we have a block copy from an enregisterable struct (today, that's just SIMD) to a different type target, it needs to be makred as address-taken, because the destination type is what's used for the copy, and all non-enregisterable destination types expect their source in memory.
Fix #16254
dotnet-maestro-bot [Thu, 8 Feb 2018 01:06:29 +0000 (19:06 -0600)]
Update BuildTools, CoreClr to preview2-02507-02, preview2-26207-11, respectively (#16256)