platform/upstream/coreclr.git
6 years agoPut default interface support behind a COMPlus switch (#15233)
Michal Strehovský [Tue, 28 Nov 2017 15:43:38 +0000 (16:43 +0100)]
Put default interface support behind a COMPlus switch (#15233)

7 years agoMerge pull request #15022 from MichalStrehovsky/master-mere
Michal Strehovský [Tue, 14 Nov 2017 21:03:14 +0000 (22:03 +0100)]
Merge pull request #15022 from MichalStrehovsky/master-mere

Merge master to dev/defaultintf

7 years agoMerge branch 'master' into master-mere
Michal Strehovský [Tue, 14 Nov 2017 18:13:35 +0000 (10:13 -0800)]
Merge branch 'master' into master-mere

7 years agoMerge pull request #15017 from BruceForstall/DisableArm64Tests
Bruce Forstall [Tue, 14 Nov 2017 17:46:09 +0000 (09:46 -0800)]
Merge pull request #15017 from BruceForstall/DisableArm64Tests

Disable arm64 tests that fail in Debug CI runs

7 years agoAdd object null checks in Memory<T> APIs to support default (#14816)
Ahson Khan [Tue, 14 Nov 2017 15:31:48 +0000 (07:31 -0800)]
Add object null checks in Memory<T> APIs to support default (#14816)

* Add object null checks in Memory<T> APIs to support default

* Changing Empty property to return default value.

* Update use of object null checks

* Addressing PR feedback.

* Removing typeof char check.

* Fix typo

7 years agoFix CppCodeGen break with latest XCode (dotnet/corert#4926)
Jan Kotas [Tue, 14 Nov 2017 12:13:22 +0000 (04:13 -0800)]
Fix CppCodeGen break with latest XCode (dotnet/corert#4926)

Latest XCode errors on ordered comparison of pointer with integer

Also, fixed some unnecessary always-true comparisons that the C++ compiler emitted warnings for.

Signed-off-by: dotnet-bot <dotnet-bot@microsoft.com>
7 years agoMerge pull request #14929 from sdmaclea/PR-ARM64-SIMD-genStructReturn
Carol Eidt [Tue, 14 Nov 2017 15:20:52 +0000 (07:20 -0800)]
Merge pull request #14929 from sdmaclea/PR-ARM64-SIMD-genStructReturn

[Arm64] SIMD genStructReturn

7 years agoMerge pull request #14934 from sdmaclea/PR-ARM64-SIMD-genMultiRegCallStoreToLocal
Carol Eidt [Tue, 14 Nov 2017 15:19:54 +0000 (07:19 -0800)]
Merge pull request #14934 from sdmaclea/PR-ARM64-SIMD-genMultiRegCallStoreToLocal

[Arm64] SIMD genMultiRegCallStoreToLocal

7 years agoMerge pull request #14975 from sdmaclea/PR-SIMD-genSIMDIntrinsicInit-0
Carol Eidt [Tue, 14 Nov 2017 15:18:06 +0000 (07:18 -0800)]
Merge pull request #14975 from sdmaclea/PR-SIMD-genSIMDIntrinsicInit-0

[ARM64] genSIMDIntrinsicInit contain 0

7 years agoMerge pull request #14976 from sdmaclea/PR-SIMD-genCodeForStoreLclVar-0
Carol Eidt [Tue, 14 Nov 2017 15:16:29 +0000 (07:16 -0800)]
Merge pull request #14976 from sdmaclea/PR-SIMD-genCodeForStoreLclVar-0

[ARM64] SIMD genCodeForStoreLclVar zero init

7 years agoUpdate BuildToolsVersion.txt
Michal Strehovský [Tue, 14 Nov 2017 11:58:36 +0000 (12:58 +0100)]
Update BuildToolsVersion.txt

7 years agoAdding Memory CopyTo APIs (#15010)
Ahson Khan [Tue, 14 Nov 2017 05:11:53 +0000 (21:11 -0800)]
Adding Memory CopyTo APIs (#15010)

7 years agoMerge pull request #15018 from dotnet-bot/from-tfs
Brian Sullivan [Tue, 14 Nov 2017 02:01:29 +0000 (18:01 -0800)]
Merge pull request #15018 from dotnet-bot/from-tfs

Merge changes from TFS

7 years agoMerge pull request #14944 from sdmaclea/PR-ARM64-SIMD-Partial-Callee-Save-SIMD12
Carol Eidt [Tue, 14 Nov 2017 00:31:57 +0000 (16:31 -0800)]
Merge pull request #14944 from sdmaclea/PR-ARM64-SIMD-Partial-Callee-Save-SIMD12

[ARM64] SIMD Partial Callee Save SIMD12

7 years agoDisable arm64 tests that fail in Debug CI runs
Bruce Forstall [Tue, 14 Nov 2017 00:25:30 +0000 (16:25 -0800)]
Disable arm64 tests that fail in Debug CI runs

Addresses #15016, until these can be investigated.

7 years agoMerge pull request #14953 from mikedn/ssa-mem-track
Bruce Forstall [Tue, 14 Nov 2017 00:05:17 +0000 (16:05 -0800)]
Merge pull request #14953 from mikedn/ssa-mem-track

Track actual SSA memory usage

7 years agoMerge pull request #15013 from BruceForstall/DisableTests
Bruce Forstall [Mon, 13 Nov 2017 23:26:15 +0000 (15:26 -0800)]
Merge pull request #15013 from BruceForstall/DisableTests

Disable ARM tests

7 years agoMerge pull request #14970 from jashook/better_smarty_archiving
Jarret Shook [Mon, 13 Nov 2017 23:11:23 +0000 (15:11 -0800)]
Merge pull request #14970 from jashook/better_smarty_archiving

Better smarty archiving

7 years agoDisable VectorConvert_r/VectorConvert_ro for ARM
Bruce Forstall [Mon, 13 Nov 2017 22:01:52 +0000 (14:01 -0800)]
Disable VectorConvert_r/VectorConvert_ro for ARM

The are now failing due to #14665.

7 years agoEnable long-running or resource-intensive GC tests
Bruce Forstall [Mon, 13 Nov 2017 21:56:28 +0000 (13:56 -0800)]
Enable long-running or resource-intensive GC tests

Specifically,
```
foregroundgc
lohfragmentation
536168
```

These are already disabled for arm64 and in issues.targets and the
various non-Windows *.txt files.

7 years agoMerge pull request #14988 from briansull/fix-523532
Brian Sullivan [Mon, 13 Nov 2017 21:38:56 +0000 (13:38 -0800)]
Merge pull request #14988 from briansull/fix-523532

Fix VSO 523532 - Possible mismatch in size on GT_EQ created for switch expansion

7 years agoMerge pull request #14921 from BruceForstall/EnableRyuJitArm32Crossgen
Bruce Forstall [Mon, 13 Nov 2017 21:19:22 +0000 (13:19 -0800)]
Merge pull request #14921 from BruceForstall/EnableRyuJitArm32Crossgen

Enable RyuJIT/arm32 crossgen compilation of S.P.C.dll

7 years agoMerge pull request dotnet/corert#4918 from dotnet/nmirror
Jan Kotas [Sun, 12 Nov 2017 04:45:53 +0000 (20:45 -0800)]
Merge pull request dotnet/corert#4918 from dotnet/nmirror

Merge nmirror to master

Signed-off-by: dotnet-bot <dotnet-bot@microsoft.com>
7 years agoMerge pull request #14984 from jashook/add_new_publish_vars
Jarret Shook [Mon, 13 Nov 2017 20:32:18 +0000 (12:32 -0800)]
Merge pull request #14984 from jashook/add_new_publish_vars

Use new publish vars

7 years agoUse new publish vars
jashook [Sat, 11 Nov 2017 00:06:31 +0000 (16:06 -0800)]
Use new publish vars

7 years agoBetter smarty archiving
jashook [Thu, 9 Nov 2017 23:35:37 +0000 (15:35 -0800)]
Better smarty archiving

7 years agofix DevDiv521437
dotnet-bot [Mon, 13 Nov 2017 18:32:12 +0000 (10:32 -0800)]
fix DevDiv521437

[tfs-changeset: 1680966]

7 years agocheck that VM set pAddr == nullptr (#15003)
Sergey Andreenko [Mon, 13 Nov 2017 18:27:26 +0000 (10:27 -0800)]
check that VM set pAddr == nullptr (#15003)

When VM returns addr != nullptr from getHelperFtn, it must set pAddr to nullptr.

7 years agoEnable RyuJIT/arm32 crossgen compilation of S.P.C.dll
Bruce Forstall [Wed, 8 Nov 2017 05:49:26 +0000 (21:49 -0800)]
Enable RyuJIT/arm32 crossgen compilation of S.P.C.dll

Currently, an ARM build and CI test run uses the legacy backend
arm32 JIT. This change enables a build of RyuJIT/arm32 as a
cross-targeting protojit.dll (altjit) in the cross tools (placed
next to the x86 hosted crossgen.exe in the build), and uses that
RyuJIT/arm32 cross-compiler to build System.Private.CoreLib.dll.
The CI system when testing RyuJIT/arm32 will use this to test
the crossgen'ed System.Private.CoreLib.dll built by RyuJIT/arm32
for all "arm" architecture jobs.

7 years agoMerge pull request #14920 from dotnet-maestro-bot/master-UpdateDependencies
Wes Haggard [Mon, 13 Nov 2017 17:43:41 +0000 (09:43 -0800)]
Merge pull request #14920 from dotnet-maestro-bot/master-UpdateDependencies

Update BuildTools, CoreClr, CoreFx, PgoData to prerelease-02210-05, preview1-25913-02, preview1-25913-02, master-20171110-0018, respectively (master)

7 years agoUpdate BuildTools, CoreClr, CoreFx, PgoData to prerelease-02210-05, preview1-25913...
dotnet-maestro-bot [Mon, 13 Nov 2017 14:41:28 +0000 (06:41 -0800)]
Update BuildTools, CoreClr, CoreFx, PgoData to prerelease-02210-05, preview1-25913-02, preview1-25913-02, master-20171110-0018, respectively

7 years agoFix sempahore spinning fix (#15002)
Stephen Toub [Mon, 13 Nov 2017 06:12:44 +0000 (01:12 -0500)]
Fix sempahore spinning fix (#15002)

A recent fix around SemaphoreSlim's spinning has led to it spinning even when spinning isn't necessary, resulting in measurable overheads when trying to acquire a semaphore that has count available.  This fixes that by putting back the check to avoid spinning if count is known to be available.

7 years agoPort https://github.com/dotnet/coreclr/pull/13148 to master (#14999)
Michal Strehovský [Sun, 12 Nov 2017 20:01:31 +0000 (21:01 +0100)]
Port https://github.com/dotnet/coreclr/pull/13148 to master (#14999)

7 years agoUpdating the xUnit Performance Api with bug fixes to the benchmark scenarios (#14986)
José Rivero [Sun, 12 Nov 2017 20:01:08 +0000 (12:01 -0800)]
Updating the xUnit Performance Api with bug fixes to the benchmark scenarios (#14986)

This newer version fixes collecting data on x86 where:
1. The Jit generate methods at runtime and that are not associated with any loaded module
2. Modules are loaded/unloaded multiple times
3. This newer version of the Api added some breaking changes to the Api interface

7 years agoAllow non-abstract non-public interface methods in ILAsm (#14998)
Michal Strehovský [Sun, 12 Nov 2017 19:58:20 +0000 (20:58 +0100)]
Allow non-abstract non-public interface methods in ILAsm (#14998)

This is to support the default interface method feature in C#.
See https://github.com/dotnet/csharplang/blob/5971e753a20a32e5f471a5439c88e667e32413f0/proposals/default-interface-methods.md

7 years agoRemove unnecessary Exception._exceptionMethodString field (#14992)
Stefan Nikolei [Sat, 11 Nov 2017 21:30:35 +0000 (22:30 +0100)]
Remove unnecessary Exception._exceptionMethodString field (#14992)

7 years agoFix implementation of GCToOSInterface::VirtualCommit to ditch the pages (#14979)
Jan Kotas [Sat, 11 Nov 2017 15:21:18 +0000 (07:21 -0800)]
Fix implementation of GCToOSInterface::VirtualCommit to ditch the pages (#14979)

The GC depends on the pages to be zeroed out when they are mapped in again

7 years agoMove '-all' from filename to args in Mac build (#14990)
Davis Goodin [Sat, 11 Nov 2017 01:41:08 +0000 (19:41 -0600)]
Move '-all' from filename to args in Mac build (#14990)

7 years agoFix VSO 523532 - Possible mismatch in size on GT_EQ created for switch expansion
Brian Sullivan [Sat, 11 Nov 2017 00:59:48 +0000 (16:59 -0800)]
Fix VSO 523532 - Possible mismatch in size on GT_EQ created for switch expansion

7 years agoDisable GC test requires much bigger amount of memory than arm32 machine may ever...
Roman Artemev [Fri, 10 Nov 2017 23:38:46 +0000 (15:38 -0800)]
Disable GC test requires much bigger amount of memory than arm32 machine may ever have (#14981)

7 years agoMerge pull request #14973 from sdmaclea/PR-ARM64-VECTOR-LD-REG-SIZE
Brian Sullivan [Fri, 10 Nov 2017 20:49:23 +0000 (12:49 -0800)]
Merge pull request #14973 from sdmaclea/PR-ARM64-VECTOR-LD-REG-SIZE

[Arm64] Do not change vector load register size

7 years agoRun clean -all instead of just -b (#14978)
Karthik Rajasekaran [Fri, 10 Nov 2017 20:03:48 +0000 (12:03 -0800)]
Run clean -all instead of just -b (#14978)

7 years ago[Arm64] Do not change vector load register size
Steve MacLean [Thu, 9 Nov 2017 17:13:51 +0000 (12:13 -0500)]
[Arm64] Do not change vector load register size

7 years ago[ARM64] SIMD genCodeForStoreLclVar zero init
Steve MacLean [Fri, 10 Nov 2017 16:09:33 +0000 (11:09 -0500)]
[ARM64] SIMD genCodeForStoreLclVar zero init

7 years ago[ARM64] genSIMDIntrinsicInit contain 0
Steve MacLean [Fri, 10 Nov 2017 17:57:21 +0000 (12:57 -0500)]
[ARM64] genSIMDIntrinsicInit contain 0

7 years agoMerge pull request #14974 from jcagme/master
JC Aguilera [Fri, 10 Nov 2017 18:45:29 +0000 (10:45 -0800)]
Merge pull request #14974 from jcagme/master

Removing retry parameters since they are not used anymore

7 years agoRemoving retry parameters since they are not used anymore
Juan Carlos Aguilera Mendez [Fri, 10 Nov 2017 17:49:09 +0000 (09:49 -0800)]
Removing retry parameters since they are not used anymore

7 years agoAdded UnixEpoch to DateTime and DateTimeOffset. (#14648)
Adrian Godong [Fri, 10 Nov 2017 16:25:50 +0000 (08:25 -0800)]
Added UnixEpoch to DateTime and DateTimeOffset. (#14648)

* Added UnixEpoch to DateTime and DateTimeOffset.

* Use ticks instead.

* Use DateTime.TicksTo1970

* Consolidated UnixEpochTicks

7 years agoSeparate Profile=On and Off for perf pipeline (#14963)
Michelle McDaniel [Fri, 10 Nov 2017 16:03:50 +0000 (08:03 -0800)]
Separate Profile=On and Off for perf pipeline (#14963)

* Separate Profile=On and Off for perf pipeline

In pipeline jobs, we can separate these two runs into separate runs.
Doing so decreases the time spent running windows jobs since we are not
running them in sequence. Only run Profile=On scenarios for PRs. This
should reduce the time spent running the PR job.

This change also adds the xunit max iteration parameters.

7 years ago[ARM64] SIMD Partial Callee Save SIMD12
Steve MacLean [Wed, 8 Nov 2017 18:47:38 +0000 (13:47 -0500)]
[ARM64] SIMD Partial Callee Save SIMD12

7 years agoMerge pull request #14708 from mikedn/rc-ssa
Joseph Tremoulet [Fri, 10 Nov 2017 14:42:22 +0000 (09:42 -0500)]
Merge pull request #14708 from mikedn/rc-ssa

Change RangeCheck to use the available SSA data

7 years agoFix build break when VS2015 is not installed. (#14948)
Vance Morrison [Fri, 10 Nov 2017 00:52:57 +0000 (16:52 -0800)]
Fix build break when VS2015 is not installed. (#14948)

* Fix build break when VS2015 is not installed.

If you ONLY install VS2017 (now the common scneario)
and build coreclr, it will fail  with COM with a Class_Not_Registered error

This is because a build-tools package is depending on an old COM object
that has gone way in VS2017.

While we sort getting a proper fix, this makes the work-around (which
is registering the DLL by hand) discoverable.

See https://github.com/dotnet/coreclr/issues/11305 for details.

* typo

7 years agoFix clean.cmd to work if no parameters are passed (#14969)
Matt Mitchell [Fri, 10 Nov 2017 00:02:30 +0000 (16:02 -0800)]
Fix clean.cmd to work if no parameters are passed (#14969)

7 years agoMerge pull request #14967 from jashook/small_netci_fixes
Jarret Shook [Thu, 9 Nov 2017 21:09:18 +0000 (13:09 -0800)]
Merge pull request #14967 from jashook/small_netci_fixes

Fix arm(64) jobs

7 years agoFix arm(64) jobs
jashook [Thu, 9 Nov 2017 20:53:33 +0000 (12:53 -0800)]
Fix arm(64) jobs

7 years ago[Arm64] SIMD genMultiRegCallStoreToLocal
Steve MacLean [Wed, 8 Nov 2017 17:32:45 +0000 (12:32 -0500)]
[Arm64] SIMD genMultiRegCallStoreToLocal

7 years ago[Arm64] SIMD genStructReturn
Steve MacLean [Tue, 7 Nov 2017 17:45:58 +0000 (12:45 -0500)]
[Arm64] SIMD genStructReturn

7 years agoMerge pull request #14961 from jashook/arm64_windows_nt_rework
Jarret Shook [Thu, 9 Nov 2017 18:36:14 +0000 (10:36 -0800)]
Merge pull request #14961 from jashook/arm64_windows_nt_rework

Rework Arm64 Windows jobs to flow jobs

7 years agoChange RangeCheck to use the available SSA data
Mike Danes [Thu, 26 Oct 2017 08:10:17 +0000 (11:10 +0300)]
Change RangeCheck to use the available SSA data

The necessary information (the assignment node and the block where it appears) can be obtained from the SSA data, there is no need to build a map of SSA definitions. Doing that consumes almost half of the time taken by RangeCheck and a sizeable chunk of memory.

7 years ago[Local GC] Unify background GC thread and server GC thread creation (#14821)
Sean Gillespie [Thu, 9 Nov 2017 17:58:47 +0000 (09:58 -0800)]
[Local GC] Unify background GC thread and server GC thread creation (#14821)

* Initial cut, ignoring thread affinity

* Integrate thread affinity

* Affinity for standalone Windows

* Add 'specialness' and the thread name as arguments to CreateThread

* First crack at unified implementation

* Set priority for server GC threads

* Remove unused parameter

* Address code review feedback and remove some dead code that broke the clang build

* Use char* on the interface instead of wchar_t (doesn't play well cross-platform)

* Rename IsGCSpecialThread -> CurrentThreadWasCreatedByGC

* Code review feedback and fix up the build

* rename CurrentThreadWasCreatedByGC -> WasCurrentThreadCreatedByGC

* Fix a contract violation when converting string encodings

* Thread::CreateUtilityThread returns a thread that is not suspended - restarting a non-suspended thread is incorrect

* CreateUnsuspendableThread -> CreateNonSuspendableThread

7 years agoMerge pull request #14882 from 4creators/cleanall
Wes Haggard [Thu, 9 Nov 2017 17:29:23 +0000 (09:29 -0800)]
Merge pull request #14882 from 4creators/cleanall

[Infrastructure] Improve clean.{cmd|sh} by adding git clean -xdf cleanup step, unifying implementation with corefx and adding VBCScompiler.exe kill logic and

7 years agoUpdate CoreCLR package to the freshly published version (#14962)
Michal Strehovský [Thu, 9 Nov 2017 16:49:49 +0000 (17:49 +0100)]
Update CoreCLR package to the freshly published version (#14962)

7 years agoRework Arm64 Windows jobs to flow jobs
jashook [Fri, 27 Oct 2017 18:22:19 +0000 (11:22 -0700)]
Rework Arm64 Windows jobs to flow jobs

This drops our dependency on the internal Arm64CI

7 years agoMerge pull request dotnet/corefx#25092 from brianrob/missing_define
Vance Morrison [Tue, 7 Nov 2017 19:07:39 +0000 (11:07 -0800)]
Merge pull request dotnet/corefx#25092 from brianrob/missing_define

Add PLATFORM_WINDOWS define for MDT EventSource package

Signed-off-by: dotnet-bot <dotnet-bot@microsoft.com>
7 years agoRemove Linux perf from pr pipeline job (#14943)
Michelle McDaniel [Thu, 9 Nov 2017 14:44:01 +0000 (06:44 -0800)]
Remove Linux perf from pr pipeline job (#14943)

* Remove Linux perf from pr pipeline job

We don't have a lot of linux perf machines and the sheer number of PR
jobs is causing them to be overloaded, so the perf leg is taking too
long. Disable them for prs for now until we can increase capacity.

* Disable baseline jobs

7 years agoRestore #13148 (#14958)
Michal Strehovský [Thu, 9 Nov 2017 10:39:08 +0000 (11:39 +0100)]
Restore #13148 (#14958)

Use ILAsm from this branch. This was lost in merges because the property got renamed.

We should really just port the ILAsm part of the feature to master...

7 years agoMerge pull request #14954 from dotnet/revert-14722-arm64_windows_nt_rework
Jarret Shook [Thu, 9 Nov 2017 06:57:59 +0000 (22:57 -0800)]
Merge pull request #14954 from dotnet/revert-14722-arm64_windows_nt_rework

Revert "Rework Arm64 Windows jobs to flow jobs"

7 years agoRevert "Rework Arm64 Windows jobs to flow jobs"
Jarret Shook [Thu, 9 Nov 2017 06:56:46 +0000 (22:56 -0800)]
Revert "Rework Arm64 Windows jobs to flow jobs"

7 years agoTrack actual SSA memory usage
Mike Danes [Thu, 9 Nov 2017 06:18:48 +0000 (08:18 +0200)]
Track actual SSA memory usage

Currently SsaBuilder use comp->getAllocator() in a lot of places so a lot of the memory it allocates lands in CMK_Generic instead of CMK_SSA.

Same for CopyProp phase.

7 years agoMerge pull request #14722 from jashook/arm64_windows_nt_rework
Jarret Shook [Thu, 9 Nov 2017 05:50:32 +0000 (21:50 -0800)]
Merge pull request #14722 from jashook/arm64_windows_nt_rework

Rework Arm64 Windows jobs to flow jobs

7 years agoRework Arm64 Windows jobs to flow jobs
jashook [Fri, 27 Oct 2017 18:22:19 +0000 (11:22 -0700)]
Rework Arm64 Windows jobs to flow jobs

This drops our dependency on the internal Arm64CI

7 years ago[RyuJIT/arm32] Fix GC hole in address mode generation requiring temps (#14949)
Bruce Forstall [Thu, 9 Nov 2017 05:27:01 +0000 (21:27 -0800)]
[RyuJIT/arm32] Fix GC hole in address mode generation requiring temps (#14949)

* [RyuJIT/arm32] Fix GC hole in address mode generation requiring temps

In cases requiring a temp register to compute a portion of an LEA,
where the partial computation involves a GCREF/BYREF base register,
the temp register must be marked as a BYREF, so it can get
updated during a GC.

This fixes a case like this:
```
add r2, r0, r1 LSL 2
ldr r0, [r2+24]
```

where `r2` needs to be marked BYREF.

Related to #14856.

* Formatting

7 years agoMerge pull request #14906 from stephentoub/add_memory_string
Stephen Toub [Thu, 9 Nov 2017 04:19:46 +0000 (23:19 -0500)]
Merge pull request #14906 from stephentoub/add_memory_string

Add string support to ReadOnlyMemory<char>

7 years agoMerge pull request #14918 from BruceForstall/NoLegacyUnsafeCpBlk
Bruce Forstall [Thu, 9 Nov 2017 02:14:17 +0000 (18:14 -0800)]
Merge pull request #14918 from BruceForstall/NoLegacyUnsafeCpBlk

Don't allow unsafe CpBlk for legacy backend

7 years agoMerge pull request #14935 from sdmaclea/PR-ARM64-SIMD-genCodeForLclFld
Carol Eidt [Thu, 9 Nov 2017 01:12:45 +0000 (17:12 -0800)]
Merge pull request #14935 from sdmaclea/PR-ARM64-SIMD-genCodeForLclFld

[ARM64] SIMD genCodeForLclFld

7 years agoAdd foreach support to Span<T> and ReadOnlySpan<T> (#14922)
Stephen Toub [Thu, 9 Nov 2017 01:02:50 +0000 (20:02 -0500)]
Add foreach support to Span<T> and ReadOnlySpan<T> (#14922)

Adds a ref struct Enumerator, and a corresponding GetEnumerator method to each.

7 years agoMerge pull request #14946 from MichalStrehovsky/master-merge
Michal Strehovský [Thu, 9 Nov 2017 00:09:17 +0000 (01:09 +0100)]
Merge pull request #14946 from MichalStrehovsky/master-merge

Merge master to dev/defaultintf

7 years agoMerge branch 'master' into master-merge
Michal Strehovský [Wed, 8 Nov 2017 23:51:45 +0000 (00:51 +0100)]
Merge branch 'master' into master-merge

7 years agoMerge pull request #14912 from briansull/add-b124443
Brian Sullivan [Wed, 8 Nov 2017 23:29:30 +0000 (15:29 -0800)]
Merge pull request #14912 from briansull/add-b124443

Added missing JIT test V1-M15-SP2/b124443/b124443.cs

7 years agoAdding placeholder Span debugger proxy (#14749)
Ahson Khan [Wed, 8 Nov 2017 22:29:51 +0000 (14:29 -0800)]
Adding placeholder Span debugger proxy (#14749)

* Adding placeholder Span debugger proxy.

* Remove unnecessary unsafe keyword.

7 years agoSmall number of bug fixes. (#14900)
José Rivero [Wed, 8 Nov 2017 22:09:14 +0000 (14:09 -0800)]
Small number of bug fixes. (#14900)

- Added stability prefix to the scenario benchmark (JitBench)
- Specify output directory to the `run-xunit-perf.cmd` script and avoid the extra step to xcopy files to the archive folder.
- Added a command line parser class to the illink scenario, and changed its behavior where it used to fail when a new command line option passed to xUnit was not recognized.
- Save the output log of the tests into the sandbox-logs folder.
- Updating the label of the machine pool used by the illink scenario

7 years agoMerge pull request #14942 from jcagme/master
Wes Haggard [Wed, 8 Nov 2017 21:48:12 +0000 (13:48 -0800)]
Merge pull request #14942 from jcagme/master

Adding IsPublic to CreateAzureContainer

7 years agoAdding IsPublic to CreateAzureContainer
Juan Carlos Aguilera Mendez [Wed, 8 Nov 2017 21:44:48 +0000 (13:44 -0800)]
Adding IsPublic to CreateAzureContainer

7 years agoChange Sym publish to use UnzipSymbolPackagesForPublish (#14941)
Karthik Rajasekaran [Wed, 8 Nov 2017 21:44:35 +0000 (13:44 -0800)]
Change Sym publish to use UnzipSymbolPackagesForPublish (#14941)

7 years agoAdd missing license header (#14936)
Jan Kotas [Wed, 8 Nov 2017 21:03:58 +0000 (13:03 -0800)]
Add missing license header (#14936)

7 years agoIncrease retry attempts in publish (#14940)
Karthik Rajasekaran [Wed, 8 Nov 2017 20:24:39 +0000 (12:24 -0800)]
Increase retry attempts in publish (#14940)

7 years agoPass in property SkipCreateWindowsPdbsFromPortablePdbs to fix (#14939)
Karthik Rajasekaran [Wed, 8 Nov 2017 20:06:29 +0000 (12:06 -0800)]
Pass in property SkipCreateWindowsPdbsFromPortablePdbs to fix (#14939)

7 years agoRemove RelativePath from Publish (#14938)
Karthik Rajasekaran [Wed, 8 Nov 2017 19:54:33 +0000 (11:54 -0800)]
Remove RelativePath from Publish (#14938)

7 years agoAdded missing test JIT/Regression/CLR-x86-JIT/V1-M15-SP2/b124443/b124443.cs
Brian Sullivan [Wed, 8 Nov 2017 00:30:21 +0000 (16:30 -0800)]
Added missing test JIT/Regression/CLR-x86-JIT/V1-M15-SP2/b124443/b124443.cs
Also added an IL version of this test for regression repro on CoreCLR

7 years agoMerge pull request #14928 from sdmaclea/PR-ARM64-SIMD-genPutArgStk
Carol Eidt [Wed, 8 Nov 2017 19:13:15 +0000 (11:13 -0800)]
Merge pull request #14928 from sdmaclea/PR-ARM64-SIMD-genPutArgStk

[Arm64] SIMD genPutArgStk

7 years agoMerge pull request #14937 from jcagme/master
JC Aguilera [Wed, 8 Nov 2017 18:59:12 +0000 (10:59 -0800)]
Merge pull request #14937 from jcagme/master

Updating BuildTools version

7 years agoUpdating BuildTools version
Juan Carlos Aguilera Mendez [Wed, 8 Nov 2017 18:39:38 +0000 (10:39 -0800)]
Updating BuildTools version

7 years ago[ARM64] SIMD genCodeForLclFld
Steve MacLean [Fri, 3 Nov 2017 17:28:38 +0000 (13:28 -0400)]
[ARM64] SIMD genCodeForLclFld

7 years agoMerge pull request #14861 from sdmaclea/PR-ARM64-SIMD-genFnPrologCalleeRegArgs
Carol Eidt [Wed, 8 Nov 2017 17:37:43 +0000 (09:37 -0800)]
Merge pull request #14861 from sdmaclea/PR-ARM64-SIMD-genFnPrologCalleeRegArgs

[Arm64] SIMD genFnPrologCalleeRegArgs

7 years agoRemove ConfigurationGroup from Publish Symbols (#14931)
Karthik Rajasekaran [Wed, 8 Nov 2017 17:28:47 +0000 (09:28 -0800)]
Remove ConfigurationGroup from Publish Symbols (#14931)

7 years agoUpdate clean.sh script to reflect changes introduced in clean.cmd, unify messages...
Jacek Blaszczynski [Wed, 8 Nov 2017 16:57:57 +0000 (17:57 +0100)]
Update clean.sh script to reflect changes introduced in clean.cmd, unify messages and functionality

7 years agoJIT: make suitably optimistic prejit inline assessments (#14850)
Andy Ayers [Wed, 8 Nov 2017 16:48:49 +0000 (08:48 -0800)]
JIT: make suitably optimistic prejit inline assessments (#14850)

When prejitting, the jit assesses whether each root method is a potential
inline candidate for any possible caller. Methods deemed un-inlinable in any
caller are marked in the runtime as "noinline" to save the jit some work
later on when it sees calls to these methods.

This assessment was too conservative and led to prejit-ordering dependences
for inlines. It also meant that prejitting was missing some inlines that
would have happened had we not done the prejit root assessment.

This change removes some of the prejit observation blockers. These mostly
will enable more prejit methods to become candidates. We also now track when
a method argument reaches a test.

When we are assessing profitability for a prejit root, assume the call site
best possible case.

Also, update the inline XML to capture the prejit assessments.

This increases the number of inlines considered and performed when prejitting
and will also lead to slightly more inlining when jitting. However we do not
expect a large througput impact when jitting -- the main impact of this change
is that inlining when prejitting is now just as aggressive as inlining when
jitting, and the decisions no longer depend on the order in which methods are
prejitted.

Closes #14441.
Closes #3482.

7 years ago[Arm64] SIMD genPutArgStk
Steve MacLean [Mon, 6 Nov 2017 21:59:42 +0000 (16:59 -0500)]
[Arm64] SIMD genPutArgStk

7 years agoAdd MemoryMarshal.AsMemory
Stephen Toub [Wed, 8 Nov 2017 15:01:50 +0000 (10:01 -0500)]
Add MemoryMarshal.AsMemory

7 years agoAdd string support to ReadOnlyMemory<char>
Stephen Toub [Tue, 7 Nov 2017 22:00:08 +0000 (17:00 -0500)]
Add string support to ReadOnlyMemory<char>