Egor Chesakov [Wed, 30 Jan 2019 22:59:29 +0000 (14:59 -0800)]
Disable tests consistently failing in AzDO CI (#22304)
* JIT\jit64\opt\cse\hugeexpr1 on Linux/arm
Related issue: https://github.com/dotnet/coreclr/issues/22260
* baseservices\threading\interlocked\compareexchange\CompareExchangeTClass
* baseservices\threading\interlocked\compareexchange\CompareExchangeTClass_1
Related issue: https://github.com/dotnet/coreclr/issues/22303
Eugene Rozenfeld [Tue, 29 Jan 2019 19:26:49 +0000 (11:26 -0800)]
Remove GTF_ADDR_ONSTACK and IsVarAddr.
IsVarAddr was checking GTF_ADDR_ONSTACK to determine if
the GT_ADDR node is an address of a local. This change removes both
GTF_ADDR_ONSTACK and IsVarAddr and uses IsLocalAdrExpr instead.
IsLocalAddrExpr uses opcodes to determine if GT_ADDR node is
a local address.
GTF_ADDR_ONSTACK flag is ancient, added before 2002 so I couldn't find
the checkin that introduced it.
I changed the assert to a check and an assignment since simplifications
inside fgMorphArgs between
https://github.com/dotnet/coreclr/blob/
1a1e4c4d5a8030cb8d82a2e5b06c2ab357b92534/src/jit/morph.cpp#L3709
(which causes https://github.com/dotnet/coreclr/blob/
1a1e4c4d5a8030cb8d82a2e5b06c2ab357b92534/src/jit/morph.cpp#L3057)
and
https://github.com/dotnet/coreclr/blob/
1a1e4c4d5a8030cb8d82a2e5b06c2ab357b92534/src/jit/morph.cpp#L3790
may result in more GT_ADDR nodes recognized by IsLocalAdrExpr.
x86 and x64 pmi frameworks had no code diffs and some gcinfo reductions
(15 methods with gcinfo diffs in x86).
Fixes #22190.
Stephen Toub [Wed, 30 Jan 2019 21:22:27 +0000 (16:22 -0500)]
Switch CoreLib over to using shared Environment from corefx (#22106)
This requires merging/adapting the implementation with EnvironmentAugments (which goes away completely), the shared files, what corert has, etc.
Michal Strehovský [Wed, 30 Jan 2019 20:16:13 +0000 (21:16 +0100)]
Fix CorError header file generation (#22294)
The checked in batch script was only usable if you have a Razzle environment and even then it didn't update the right files.
Meir Blachman [Wed, 30 Jan 2019 15:44:57 +0000 (17:44 +0200)]
fixed pdf link in dotnet-standards.md (#22292)
Michal Strehovský [Wed, 30 Jan 2019 12:41:44 +0000 (13:41 +0100)]
Fix namespace of the AmbiguousImplementationException (#22291)
Moving to the namespace that was approved in dotnet/corefx#34124.
Michal Strehovský [Tue, 29 Jan 2019 21:38:31 +0000 (22:38 +0100)]
Recover some of the size lost with AsyncMethodBuilder/Task unification (#6913)
When we unified the implementations of these across all of our runtimes, we lost all size optimizations people have been doing on the Project N side over the past six years.
This restores a bit of the loss. For one sample app with lots of async usage, this removes 2.1 MB of generic instantiations.
There is more we can do, but I can't spend time on that right now. These two things jumped out on me when I was looking at it back in December and were an easy fix I wanted to do for a while.
Signed-off-by: dotnet-bot <dotnet-bot@microsoft.com>
Markus Kitsinger [Wed, 30 Jan 2019 04:57:23 +0000 (22:57 -0600)]
Add AmbiguousImplementationException (#22280)
Markus Kitsinger [Wed, 30 Jan 2019 04:56:32 +0000 (22:56 -0600)]
Do not cache error string in SerializationException (#22281)
Egor Chesakov [Wed, 30 Jan 2019 01:17:18 +0000 (17:17 -0800)]
RunCrossGen in ReadyToRun tests, set TimeoutPerTest, set FailOnWorkItemFailure (#22261)
* For ReadyToRun tests in AzDO set RunCrossGen=true during Helix test running
* Set __TestTimeout and introduce TimeoutPerTestInMinutes for test jobs
* Set FailOnWorkItemFailure - fail the AzDO step on Helix work item "catastrophic failure"
Marek Safar [Wed, 30 Jan 2019 00:35:09 +0000 (01:35 +0100)]
Move Marshal::StringTo* methods to shared partition (#22266)
Tanner Gooding [Tue, 29 Jan 2019 21:34:01 +0000 (13:34 -0800)]
Fixing several of the HWIntrinsic SoftwareFallback functions to be captureless (#22263)
Maoni Stephens [Tue, 29 Jan 2019 21:09:32 +0000 (13:09 -0800)]
To support container scenario, 2 HardLimit configs are added - (#22180)
GCHeapHardLimit - specifies a hard limit for the GC heap
GCHeapHardLimitPercent - specifies a percentage of the physical memory this process is allowed to use
If both are specified, GCHeapHardLimit is checked first and only when it's not specified
would we check GCHeapHardLimitPercent.
If neither is specified but the process is running inside a container with a memory
limit specified, we will take this as the hard limit:
max (20mb, 75% of the memory limit on the container)
If one of the HardLimit configs is specified, and the process is running inside a container
with a memory limit, the GC heap usage will not exceed the HardLimit but the total memory
is still the memory limit on the container so when we calculate the memory load it's based
off the container memory limit.
An example,
process is running inside a container with 200mb limit
user also specified GCHeapHardLimit as 100mb.
if 50mb out of the 100mb is used for GC, and 100mb is used for other things, the memory load
is (50 + 100)/200 = 75%.
Some notes on these configs -
+ The limit is the commit size.
+ This is only supported on 64-bit.
+ For Server GC the minimum *reserved* segment size is 16mb per heap, this is to avoid the
scenario where the hard limit is small but the process can use many procs and we end up
with tiny segments which doesn't make sense. We then keep track of the committed on the segments
so the total does not exceed the hard limit.
Marek Safar [Tue, 29 Jan 2019 01:16:16 +0000 (02:16 +0100)]
Add Marshal ansi conversion to shared partition (#22243)
* Add Marshal ansi conversion to shared partition
* Ignore bestFit/throwOnUnmappableChar on Unix
* Use same assert in StringToAnsiString
* wideCharLen parameter rename
mikedn [Mon, 28 Jan 2019 19:38:19 +0000 (21:38 +0200)]
Fix various fgMorphInitBlock issues (#21820)
* Fix various fgMorphInitBlock issues
* Remove unnecessary destAddr variable
* Extend/fix comment headers
Andy Ayers [Mon, 28 Jan 2019 18:45:45 +0000 (10:45 -0800)]
Fix finding entry point unwind record for a method with fragmented un… (#22202)
On Arm and Arm64 unwind records can only cover a limited range of code
(512K and 1MB respectively). So for methods larger than this the jit
will emit multiple "fragment" unwind records to cover the full method code
range. Only the first of these describes the behavior of the method prolog.
When mapping an offset back to a method's entry point unwind, make sure to
find this "root" unwind record instead of one of the internal fragments.
Fixes #19209.
dotnet-maestro-bot [Mon, 28 Jan 2019 14:16:58 +0000 (06:16 -0800)]
Update PgoData to master-
20190128.1 (#22237)
Ben Adams [Mon, 28 Jan 2019 02:29:19 +0000 (03:29 +0100)]
string.EndsWith use SequenceEqual not SequenceCompareTo (#22207)
TW [Mon, 28 Jan 2019 02:26:53 +0000 (04:26 +0200)]
Improve BinaryReader.ReadUInt32() perf by 30% when using MemoryStream (#22102)
BinaryReader.ReadInt32() has nice optimization which was missing from the ReadUInt32() version. Now both implementations are aligned.
ReadUInt32 and all the 64 bits types are now 30% to 50% faster than original implementation.
Jan Kotas [Sun, 27 Jan 2019 19:48:53 +0000 (11:48 -0800)]
Keep CoreCLR-specific temporary hack under ifdef
Signed-off-by: dotnet-bot <dotnet-bot@microsoft.com>
Ben Adams [Mon, 28 Jan 2019 02:01:22 +0000 (03:01 +0100)]
Shrink Task.Delay when used without cancellation (#22233)
Stephen Toub [Sun, 27 Jan 2019 19:15:08 +0000 (14:15 -0500)]
Remove several uses of VoidTaskResult (#22238)
* Remove defunct netstandard code from ValueTask
* Remove several uses of VoidTaskResult
Currently TrySetResult/Canceled/Exception live on `Task<T>`. There's no reason `TrySetCanceled` and `TrySetException` need to live there, as they only access state from the base `Task`, and so can be moved down. `TrySetResult` needs the `TResult`, however in a variety of cases `Task<T>` is used with a `VoidTaskResult`, and for such cases we can just have a parameterless `TrySetResult()` on the base class as well, which can be used any time there is no `TResult` or when `default(TResult)` is the desired result. This lets us switch several cases where we were using `Task<VoidTaskResult>` to just be `Task`, which saves 8 bytes on the task instance on 64-bit. It also avoids an Interlocked.Exchange as part of the TrySetResult call.
This primarily affects Task.Delay and the non-generic variants of Task.WhenAll, ValueTask.AsTask(), Task.FromCanceled, and Task.FromException.
Filip Navara [Sun, 27 Jan 2019 19:12:30 +0000 (20:12 +0100)]
Move Timer to shared CoreLib partition. (#22231)
* Move Timer to shared CoreLib partition.
* Move SignalNoCallbacksRunning back to shared code, add static EventWaitHandle.Set.
* Remove AppDomain references from shared Timer code, reshuffle some code out of SetTimer in non-shared code.
* Change m_ prefix to match code style.
dotnet-maestro-bot [Sun, 27 Jan 2019 04:26:13 +0000 (20:26 -0800)]
Update PgoData to master-
20190126.1 (#22198)
Filip Navara [Sun, 27 Jan 2019 04:25:06 +0000 (05:25 +0100)]
Make ResourceManager use Assembly instead of RuntimeAssembly internally. (#21979)
Marek Safar [Sun, 27 Jan 2019 01:08:53 +0000 (02:08 +0100)]
Move Marshal::ZeroFree* to shared partition (#22230)
* Move Marshal::ZeroFree* to shared partition
* Add managed version of SysStringByteLen
Jan Kotas [Sat, 26 Jan 2019 17:26:16 +0000 (09:26 -0800)]
Cleanup stackoverflow handling leftovers (#22228)
Stephen Toub [Sat, 26 Jan 2019 11:26:35 +0000 (06:26 -0500)]
Factor WinRT GetFolderPath into static accessible via reflection (#34873)
Signed-off-by: dotnet-bot <dotnet-bot@microsoft.com>
Marek Safar [Sat, 26 Jan 2019 02:32:24 +0000 (03:32 +0100)]
Move more Marshal methods to shared partition (#22205)
* Move more Marshal methods to shared partition
* Implement ThrowExceptionForHR in managed
Egor Chesakov [Sat, 26 Jan 2019 01:05:33 +0000 (17:05 -0800)]
Submit to one Helix queue on pull requests and CI builds. Add CI trigger for master branch (#22211)
This limits pull request and CI (push) triggered jobs to running only on one *default* Helix queue:
Linux arm32 - Ubuntu.1404.Arm32.Open
Linux arm64 - Ubuntu.1804.Arm64.Open
Linux x64 - Ubuntu.1804.Amd64.Open
Windows_NT x86/x64 - Windows.10.Amd64.Open
Windows_NT arm32/arm64 - we don't run these in AzDO
OSX - OSX.1013.Amd64.Open
Scheduled and manually triggered jobs will use more than one Helix queues for more comprehensive checking.
This enables CI (push) trigger for master branch. During CI build the following scenarios for Pri1 tests are going to run:
- normal
- no_tiered_compilation
Egor Chesakov [Sat, 26 Jan 2019 01:02:12 +0000 (17:02 -0800)]
Switch to dnceng-linux-external-temp pool (#22220)
Sung Yoon Whang [Fri, 25 Jan 2019 21:35:37 +0000 (13:35 -0800)]
Fix eventsource metadata exception (#21993)
* Do not throw an exception when we fail to generate metadata for eventsource
* some build failures
* return null directly here
* remove FAILURE_UNSUPPORTED_TYPE
* address PR comments
* address more pr comments
* Fix potential incorrect generation of metadata blob when property metadata generation fails
Sergey Andreenko [Fri, 25 Jan 2019 20:21:09 +0000 (12:21 -0800)]
this test should not be run under gcstress (#22214)
Ben Adams [Fri, 25 Jan 2019 19:03:33 +0000 (20:03 +0100)]
Skip rangecheck in string.EndsWith(char) (#22206)
Jeremy Koritzinsky [Fri, 25 Jan 2019 16:37:43 +0000 (08:37 -0800)]
Remove reflection-based wraper for AssemblyDependencyResolver. (#22195)
Now that AssemblyDependencyResolver is exposed in CoreFX and has propogated back to CoreCLR, we can remove the reflection-based wrapper for the unit tests for it.
Ben Adams [Fri, 25 Jan 2019 16:17:58 +0000 (17:17 +0100)]
Use AggressiveOptimization for intrinsics-based SpanHelpers (#22191)
* Use AggressiveOptimization for intrinsics-based SpanHelpers
* Remove from LastIndexOfAny
Marek Safar [Fri, 25 Jan 2019 15:58:14 +0000 (16:58 +0100)]
Reduce Task's static ctor dependencies (#22172)
* Reduce Task's static ctor dependencies
* Remove defensive check
* Revert "Remove defensive check"
This reverts commit
df15ab0350e8f44fc5cbb1a18c5211cd1de11989.
Elliot Prior [Fri, 25 Jan 2019 13:59:30 +0000 (13:59 +0000)]
Cancel CTS immediately if delay is zero. (#18098)
The intended behaviour on specifying a delay of zero is that the CancellationTokenSource is cancelled immediately - before this change, it would depend on the scheduling of the timer callback which could lead to seeing a non-cancelled token. Now, it marks itself as cancelled immediately, without invoking any callbacks (as this method is called only from within the constructor of the object, before any callbacks could have been registered.
Brian Sullivan [Fri, 25 Jan 2019 11:20:23 +0000 (03:20 -0800)]
Fixes the desktop build (#22197)
* Fixes the desktop build
warning C4311: 'reinterpret_cast': pointer truncation from 'BYTE *' to 'uint32_t' f:\j6\codegenmirror\src\ndp\clr\src\jit\emit.cpp(5636): warning C4302: 'reinterpret_cast': truncation from 'BYTE *' to 'uint32_t'
* Fix formatting
Ben Adams [Fri, 25 Jan 2019 02:48:43 +0000 (03:48 +0100)]
Speedup .SequenceCompareTo(byte, ...) (#22127)
* Speedup .SequenceCompareTo(byte, ...)
* Rename jump location
* Better annotations for clarity
Jan Vorlicek [Fri, 25 Jan 2019 01:34:39 +0000 (02:34 +0100)]
Fix collectible AssemblyLoadContext weak handle (#22193)
Native runtime keeps short weak handle for collectible AssemblyLoadContext
until the unload time when it gets converted to a strong one. A test
running with gc stress has recently hit a problem caused by the fact
that the weak handle is "short weak" and thus doesn't track
resurrection. Here is what happened:
* The test code has lost reference to the AssemblyLoadContext
* Finalizer was just invoked on that context due to that and so the weak
short handle was destroyed.
* Native runtime called AssemblyLoadContext.Resolve with the now dead
handle, the function tried to get the managed object and failed.
The fix is to change the handle type to long weak so that it doesn't get
destroyed before the finalization.
José Rivero [Fri, 25 Jan 2019 01:03:47 +0000 (17:03 -0800)]
Updating MicrosoftDiagnosticsTracingTraceEventPackageVersion (#22071)
Dotnet-GitSync-Bot [Fri, 25 Jan 2019 01:02:36 +0000 (17:02 -0800)]
Fix argument checks in ThreadPool.RegisterWaitForSingleObject. (#6887) (#22196)
Signed-off-by: dotnet-bot <dotnet-bot@microsoft.com>
Jan Kotas [Fri, 25 Jan 2019 01:01:09 +0000 (17:01 -0800)]
Rename TplEtwProvider to TplEventSource (#22194)
* Rename TplEtwProvider to TplEventSource
* Rename etwLog to log
Marek Safar [Fri, 25 Jan 2019 01:00:33 +0000 (02:00 +0100)]
Move Marshal to shared partition (#22146)
* Move Marshal to shared partition
* Moved code optimizations
* Cosmetic changes
* Avoid duplicate arguments checks bys shipping Encoding.UTF8.GetString
Andy Ayers [Fri, 25 Jan 2019 00:18:23 +0000 (16:18 -0800)]
Fix SOS GCInfo for Arm/Arm64 (#22188)
Add missing DECODE_HAS_TAILCALLS flag.
Closes #22163
Tanner Gooding [Thu, 24 Jan 2019 21:58:28 +0000 (13:58 -0800)]
Fixing some issues with the decimal parser/formatter (#22070)
* Fixing the decimal formatting so that -0 prints as 0
* Fixing the decimal parser to consider digits still in the buffer as part of the zero/non-zero tail check
Mike McLaughlin [Thu, 24 Jan 2019 21:37:47 +0000 (13:37 -0800)]
Disable exception handling on MacOS when DLL like DAC, etc. (#22171)
Marek Safar [Thu, 24 Jan 2019 20:56:41 +0000 (21:56 +0100)]
Move TaskContinuation to shared partition (#22167)
* Move TaskContinuation to shared partition
* Remove TASK_STATE_THREAD_WAS_ABORTED from Task
* Use already existing ThrowAsync
Carol Eidt [Thu, 24 Jan 2019 20:23:51 +0000 (12:23 -0800)]
Merge pull request #14393 from mikedn/float-neg-abs
Improve `new Vector2/3/4(c1, c2, c3, c4)` codegen
Mike Danes [Thu, 24 Jan 2019 09:28:28 +0000 (11:28 +0200)]
Update SSE2 test
Mike Danes [Thu, 13 Sep 2018 19:24:04 +0000 (22:24 +0300)]
Fix importer spilling in the presence of assignment side effects
Atomic ops like GT_CMPXCHG and some HW intrinsic nodes act like assignements so impAppendStmt has to account for them. They can be top level nodes or they can appear in the RHS of a GT_ASG node that perhaps is not considered to have an assignment side effect itself.
Egor Chesakov [Thu, 24 Jan 2019 18:25:46 +0000 (10:25 -0800)]
Use .asArray/.asString syntax for scenarios and helixQueues (#22173)
Eugene Rozenfeld [Thu, 24 Jan 2019 17:26:53 +0000 (09:26 -0800)]
Fix test build warning. (#22168)
I decided to suppress the warning for the test instead of
removing the test for test matrix completeness.
Fixes #21519.
dotnet-maestro-bot [Thu, 24 Jan 2019 17:25:59 +0000 (09:25 -0800)]
Update PgoData to master-
20190124.1 (#22181)
Michal Strehovský [Thu, 24 Jan 2019 15:25:00 +0000 (16:25 +0100)]
Fix UWP VirtualAlloc import (dotnet/corert#6882)
Signed-off-by: dotnet-bot <dotnet-bot@microsoft.com>
Filip Navara [Thu, 24 Jan 2019 17:25:33 +0000 (18:25 +0100)]
Remove obsolete thread abortion flags. (#22185)
Andrey Kvochko [Thu, 24 Jan 2019 16:09:34 +0000 (19:09 +0300)]
Disable PSPSym in filter prolog for CoreRT (#22184)
Koundinya Veluri [Thu, 24 Jan 2019 15:30:31 +0000 (07:30 -0800)]
Work around build failure (#22138)
Fixes https://github.com/dotnet/coreclr/issues/22103
- There were reports of build failure from https://github.com/dotnet/coreclr/pull/21292, worked around it for now with a todo
Jan Kotas [Thu, 24 Jan 2019 06:29:26 +0000 (22:29 -0800)]
IOCompletionCallback moved to shared partition
Marek Safar [Thu, 24 Jan 2019 06:18:48 +0000 (07:18 +0100)]
Move IOCompletionCallback to shared partition (dotnet/corert#6875)
Signed-off-by: dotnet-bot <dotnet-bot@microsoft.com>
Filip Navara [Thu, 24 Jan 2019 12:34:09 +0000 (13:34 +0100)]
Remove no-op holder stack validation. (#22182)
Filip Navara [Thu, 24 Jan 2019 07:39:36 +0000 (08:39 +0100)]
Fix incorrect usage of wcscpy_s/wcscat_s. (#22153)
Calum Grant [Thu, 24 Jan 2019 07:38:46 +0000 (07:38 +0000)]
Fix accidental assignment inside an assert. (#22148)
dotnet-maestro-bot [Thu, 24 Jan 2019 06:54:32 +0000 (22:54 -0800)]
Update CoreFx, PgoData to preview.19073.1, master-
20190123.4, respectively (#22112)
Stephen Toub [Thu, 24 Jan 2019 03:47:03 +0000 (22:47 -0500)]
Enable warnings-as-errors for System.Private.CoreLib (#22164)
And disable several NoWarns.
Ben Adams [Thu, 24 Jan 2019 03:33:49 +0000 (04:33 +0100)]
Speedup SpanHelpers.IndexOf{Any}(byte, ...) (#22118)
* Speedup SpanHelpers.IndexOf(byte)
* 128 * 2 alignment
* Move TrailingZeroCountFallback to common SpanHelpers
So it can be used by other types than byte
* Speedup SpanHelpers.IndexOfAny(byte, ...)
* Indent for support flags
* More helpers, constency in local names/formatting, feedback
* Skip bounds check in software fallback
Maoni Stephens [Thu, 24 Jan 2019 03:06:35 +0000 (19:06 -0800)]
this test should not be run under gcstress (#22175)
Stephen Toub [Thu, 24 Jan 2019 01:51:34 +0000 (20:51 -0500)]
Implement IAsyncDisposable.ConfigureAwait (#22160)
Jan Kotas [Wed, 23 Jan 2019 17:52:31 +0000 (09:52 -0800)]
Fix CoreRT build after moving Task.cs to shared partition
Signed-off-by: dotnet-bot <dotnet-bot@microsoft.com>
Aditya Mandaleeka [Thu, 24 Jan 2019 01:07:21 +0000 (17:07 -0800)]
Merge pull request #22003 from adityamandaleeka/writebarrier_updates_arm64
Write barrier optimizations for ARM64 Windows
Jeremy Koritzinsky [Thu, 24 Jan 2019 01:02:09 +0000 (17:02 -0800)]
Add GC.KeepAlive to keep collectible ALC alive across usage. (#22133)
Egor Chesakov [Wed, 23 Jan 2019 22:04:12 +0000 (14:04 -0800)]
Disable superpmicollection in r2r jobs until https://github.com/dotnet/coreclr/issues/22027 is fixed (#22085)
Jan Vorlicek [Wed, 23 Jan 2019 21:59:27 +0000 (22:59 +0100)]
Add unloadability testing tool (#22064)
* Add unloadability testing tool
This change adds the unloadability testing tool that can be used to run
coreclr tests or any other .NET core app inside of unloadable
AssemblyLoadContext, unload it after its execution completes and verify
that the unload succeeded. It has also various additional testing
options:
* memory leak testing
* running multiple iterations of the load/run/unload sequence
* optional breaking into debugger at various interesting stages (before
executing the test assembly, after executing it, on unload failure)
* delegated load when the AssemblyLoadContext that loads the test
assembly delegates loading of all the dependencies to another
AssemblyLoadContext
Filip Navara [Wed, 23 Jan 2019 21:35:11 +0000 (22:35 +0100)]
Move common ThreadPool code to shared CoreLib partition. (#22115)
* Move common ThreadPool code to shared CoreLib partition.
* Move numWorkingThreads counting out of shared code.
* Rename ThreadPoolGlobals.vmTpInitialized to threadPoolInitialized.
Remove exception handling from ThreadPoolWorkQueue.Dispatch.
Remove obsolete code from ThreadPoolWorkQueue.Dispatch.
Rename ThreadPool.EnsureVMInitialized to ThreadPool.EnsureInitialized.
Filip Navara [Wed, 23 Jan 2019 21:13:01 +0000 (22:13 +0100)]
Remove all traces of FEATURE_STACK_PROBE. (#22149)
Filip Navara [Tue, 22 Jan 2019 17:14:53 +0000 (18:14 +0100)]
Update parameter names in PInvoke signatures to match MSDN documentation.
Signed-off-by: dotnet-bot <dotnet-bot@microsoft.com>
Filip Navara [Tue, 22 Jan 2019 13:25:08 +0000 (14:25 +0100)]
Build fixes.
Signed-off-by: dotnet-bot <dotnet-bot@microsoft.com>
Filip Navara [Wed, 23 Jan 2019 20:47:47 +0000 (21:47 +0100)]
Fix possibly unintentional use of bit-wise or operator. (#22152)
Mike Danes [Wed, 23 Jan 2019 20:44:29 +0000 (22:44 +0200)]
Remove obsolete LEGACY_BACKEND ifdef
Egor Chesakov [Wed, 23 Jan 2019 20:29:35 +0000 (12:29 -0800)]
Disable OSX.1012.Amd64.Open in Azure DevOps (#22158)
Eugene Rozenfeld [Sat, 19 Jan 2019 00:15:32 +0000 (16:15 -0800)]
Fix R2R failure in one of ObjectStackAllocation tests.
AllocateSimpleClassAndCheckTypeNoHelper fails because
the object is allocated on the heap in R2R mode. The reason for that is
that we always use a helper for isinst in R2R mode and helper calls
cause us to assume the object escapes.
We don't have to use a helper in this case (even for R2R), #22086 tracks
fixing that.
For now the fix is not to check that the object was allocated on the stack
in this test case.
Fixes #22026.
Eugene Rozenfeld [Wed, 23 Jan 2019 00:47:33 +0000 (16:47 -0800)]
Fix for JIT\Directed\arglist\vararg test case.
TestEchoShortNoVararg was failing and disabled. The reason it was
failing was a mismatch between DllImport declaration for echo_short and
the native function: the former specified 2-byte arg and return while
the latter specified 1-byte arg and return. I fixed that and also added
a couple of test cases with 1-byte arg and 2-byte return and
vice versa.
Fixes #19704.
Jeremy Koritzinsky [Wed, 23 Jan 2019 18:34:54 +0000 (10:34 -0800)]
Don't redirect stdlib wchar functions to our test platform shims. (#22072)
* Don't redirect stdlib wchar functions to our test platform shims.
* Fix bad find-replaces.
* Remove unused overload and last wchar shadowing.
Bruce Forstall [Wed, 23 Jan 2019 18:12:30 +0000 (10:12 -0800)]
Merge pull request #22053 from BruceForstall/FixSimd12StoreLclFld
Fix SIMD12 GT_STORE_LCL_FLD
alexk000 [Wed, 23 Jan 2019 17:36:20 +0000 (12:36 -0500)]
Remove variable name duplication (#22155)
Filip Navara [Wed, 23 Jan 2019 17:35:31 +0000 (18:35 +0100)]
Remove unused thread abortion methods. (#22147)
Marek Safar [Wed, 23 Jan 2019 16:44:56 +0000 (17:44 +0100)]
Move Task to shared partition (#22105)
Egor Chesakov [Wed, 23 Jan 2019 16:11:49 +0000 (08:11 -0800)]
Use Windows.10.Arm64.Open with Jenkins ONLY (#22144)
Bruce Forstall [Fri, 18 Jan 2019 01:10:24 +0000 (17:10 -0800)]
Fix SIMD12 GT_STORE_LCL_FLD
It was ignoring the local field offset, by checking for the wrong
operator.
This appeared as a recent regression, but I couldn't find any code
(including test) change that would have caused this.
The failure only happens for JitStress with remorphing stress. This
creates the STORE_LCL_FLD. Otherwise, it looks like we normally
end up with IND(LCL_FLD_ADDR). Maybe we should be morphing these
somewhere?
No x64 PMI asm diffs in corelib or the JIT\SIMD tree. With JitStress=1
there are exactly 6 asm diffs in the Vector3Interop test, which was the
one failing.
The same fix is made for both xarch and arm64 code generators.
Fixes #21935
Ben Adams [Wed, 23 Jan 2019 03:05:25 +0000 (04:05 +0100)]
Remove duplicate requiresSynchronization params attribute (#22143)
Egor Chesakov [Wed, 23 Jan 2019 00:46:47 +0000 (16:46 -0800)]
Don't use Ubuntu.1604.Arm64.Open in Azure DevOps (#22135)
Michal Strehovský [Tue, 22 Jan 2019 20:40:56 +0000 (21:40 +0100)]
Decouple JIT_FLAG_PREJIT from HW intrinsics (#22117)
Per dotnet/corert#6836 (comment) moving the decision whether to enable HW intrinsics to the EE side.
Michal Strehovský [Tue, 22 Jan 2019 17:09:31 +0000 (18:09 +0100)]
Restrict HW intrinsic name check to intrinsic types (#22116)
Avoids having to compare names of all types in CoreLib to see if they're HW intrinsics.
Michal Strehovský [Tue, 22 Jan 2019 17:08:31 +0000 (18:08 +0100)]
Do not throw from TraceResolver (#22126)
Fixes #22059.
Filip Navara [Tue, 22 Jan 2019 12:16:45 +0000 (13:16 +0100)]
Move MemoryFailPoint to shared CoreLib partition (#22104)
* Move MemoryFailPoint to shared CoreLib partition.
* Split MemoryFailPoint into Unix and Windows versions.
* Replace MemoryFailPoint.GetMemorySettings FCall with GC.GetSegmentSize to make sharing with CoreRT easier.
Jan Kotas [Tue, 22 Jan 2019 05:02:55 +0000 (21:02 -0800)]
Fix CoreRT build breaks
Signed-off-by: dotnet-bot <dotnet-bot@microsoft.com>
Marek Safar [Tue, 22 Jan 2019 03:28:13 +0000 (04:28 +0100)]
Move TplEtwProvider to shared partition (#22120)
Stephen Toub [Tue, 22 Jan 2019 02:56:55 +0000 (21:56 -0500)]
Remove two unnecessary ToArray calls (#22119)
We can just copy directly from the List, rather than first converting the list to an array and then copying that.
Stephen Toub [Mon, 21 Jan 2019 21:20:18 +0000 (16:20 -0500)]
Avoid intermediate strings in some StringBuilder formatting (#22111)
* Avoid intermediate strings in some StringBuilder formatting
Several appends to string builders are first creating strings and then appending those strings, and they're not using the existing Append(primitive) overloads because they want to customize the format string or provider as part of appending the value. This fixes a few of those cases, using another internal AppendSpanFormattable overload on StringBuilder.
* Address PR feedback