platform/upstream/coreclr.git
6 years agoAdd Decimal.TryFormat span-based method (#15145)
Stephen Toub [Wed, 22 Nov 2017 05:13:37 +0000 (00:13 -0500)]
Add Decimal.TryFormat span-based method (#15145)

* Move decimal formatting to managed code

- Move decimal formatting to shared (decimal itself is still not shared)
- Delete VM's decimal formatting code

* Add Decimal.TryFormat span-based method

* Workaround ProjectN compiler bug

Apply the same attribution that's applied to the code in corert.

* Address PR feedback

* Temporarily undo explicit layout change

Seeing whether it's the cause of these failures on Unix (Ubuntu, CentOS, and OSX):
```
JIT.Directed.coverage.oldtests.lclfldadd_cs_r.lclfldadd_cs_r
JIT.Directed.perffix.primitivevt.mixed1_cs_ro.mixed1_cs_ro
JIT.Methodical.fp.exgen.1000w1d_cs_r.1000w1d_cs_r
JIT.Methodical.fp.exgen.1000w1d_cs_ro.1000w1d_cs_ro
JIT.Methodical.fp.exgen.1000w1d_cs_do.1000w1d_cs_do
JIT.Methodical.fp.exgen.1000w1d_cs_d.1000w1d_cs_d
JIT.Methodical.fp.exgen.10w5d_cs_d.10w5d_cs_d
JIT.Methodical.fp.exgen.10w5d_cs_do.10w5d_cs_do
JIT.Methodical.fp.exgen.10w5d_cs_ro.10w5d_cs_ro
JIT.Methodical.fp.exgen.200w1d-02_cs_do.200w1d-02_cs_do
JIT.Methodical.fp.exgen.10w5d_cs_r.10w5d_cs_r
JIT.Methodical.fp.exgen.200w1d-02_cs_ro.200w1d-02_cs_ro
JIT.Directed.perffix.primitivevt.mixed1_cs_do.mixed1_cs_do
```

Example failure:
```
FAILED   - JIT/Methodical/fp/exgen/200w1d-02_cs_ro/200w1d-02_cs_ro.sh
               BEGIN EXECUTION
               /mnt/j/workspace/dotnet_coreclr/master/checked_ubuntu_tst_prtest/bin/tests/Windows_NT.x64.Checked/Tests/coreoverlay/corerun 200w1d-02_cs_ro.exe

               Unhandled Exception: System.OverflowException: Value was either too large or too small for an Int32.
                  at System.Convert.ThrowInt32OverflowException()
                  at System.Convert.ToInt32(Int64 value)
                  at testout1.Func_0()
                  at testout1.Main()
               ./200w1d-02_cs_ro.sh: line 243: 101339 Aborted                 (core dumped) $_DebuggerFullPath "$CORE_ROOT/corerun" $ExePath $CLRTestExecutionArguments
               Expected: 100
               Actual: 134
               END EXECUTION - FAILED
```

6 years agoformat spmi source (#15144)
Sergey Andreenko [Wed, 22 Nov 2017 02:20:16 +0000 (18:20 -0800)]
format spmi source (#15144)

6 years agoDelete unused leftovers from binary serialization (#15140)
Jan Kotas [Tue, 21 Nov 2017 22:44:58 +0000 (14:44 -0800)]
Delete unused leftovers from binary serialization (#15140)

6 years agoMerge pull request dotnet/corert#4971 from dotnet/nmirror
Michal Strehovský [Tue, 21 Nov 2017 18:26:57 +0000 (19:26 +0100)]
Merge pull request dotnet/corert#4971 from dotnet/nmirror

Merge nmirror to master

Signed-off-by: dotnet-bot <dotnet-bot@microsoft.com>
6 years agoMove SafeBuffer and a few other files to shared CoreLib partition (#15141)
Jan Kotas [Tue, 21 Nov 2017 19:20:00 +0000 (11:20 -0800)]
Move SafeBuffer and a few other files to shared CoreLib partition (#15141)

6 years agoMerge pull request #15138 from BruceForstall/FixArmPri1
Bruce Forstall [Tue, 21 Nov 2017 18:42:07 +0000 (10:42 -0800)]
Merge pull request #15138 from BruceForstall/FixArmPri1

Temporary fix for arm/armlb/arm64 builds: always build Pri 1 tests

6 years agoMove UnmanagedMemoryAccessor to shared CoreLib partition (#15137)
Jan Kotas [Tue, 21 Nov 2017 14:01:12 +0000 (06:01 -0800)]
Move UnmanagedMemoryAccessor to shared CoreLib partition (#15137)

6 years agoTemporary fix for arm/armlb/arm64 builds: always build Pri 1 tests
Bruce Forstall [Tue, 21 Nov 2017 05:57:21 +0000 (21:57 -0800)]
Temporary fix for arm/armlb/arm64 builds: always build Pri 1 tests

A fix like #15075 will replace this soon.

6 years agoMerge pull request #15133 from BruceForstall/FixMissingLabel
Bruce Forstall [Tue, 21 Nov 2017 05:47:20 +0000 (21:47 -0800)]
Merge pull request #15133 from BruceForstall/FixMissingLabel

Fix genPrepForEHCodegen for arm64

6 years agoMerge pull request #15131 from BruceForstall/RemoveX86LegacyBackendTesting
Bruce Forstall [Tue, 21 Nov 2017 05:39:24 +0000 (21:39 -0800)]
Merge pull request #15131 from BruceForstall/RemoveX86LegacyBackendTesting

Remove x86 legacy_backend build and CI testing

6 years agoVersion: Use int.TryFormat & StringBuilder.Append(int) (#15132)
Justin Van Patten [Tue, 21 Nov 2017 03:49:23 +0000 (19:49 -0800)]
Version: Use int.TryFormat & StringBuilder.Append(int) (#15132)

Now that `int.TryFormat` is available, along with the more efficient
`StringBuilder.Append(int)`, use them in `Version`.

6 years agoMerge pull request #15034 from CarolEidt/FixDivRegs
Carol Eidt [Tue, 21 Nov 2017 02:13:14 +0000 (18:13 -0800)]
Merge pull request #15034 from CarolEidt/FixDivRegs

Set reg conflicts on contained IND

6 years agoReduce allocations in StringBuilder.AppendFormat for primitive types (#15110)
Justin Van Patten [Tue, 21 Nov 2017 02:11:58 +0000 (18:11 -0800)]
Reduce allocations in StringBuilder.AppendFormat for primitive types (#15110)

Use the new `TryFormat` APIs to avoid string allocations for primitive types
inside `StringBuilder.AppendFormat`, used by `string.Format`/interpolated
strings.

6 years agoChange quotes.
Santiago Fernandez Madero [Tue, 21 Nov 2017 00:29:58 +0000 (16:29 -0800)]
Change quotes.

6 years agoRemove x86 legacy_backend build and CI testing
Bruce Forstall [Tue, 21 Nov 2017 00:23:22 +0000 (16:23 -0800)]
Remove x86 legacy_backend build and CI testing

6 years agofix the adding (#15121)
Sergey Andreenko [Tue, 21 Nov 2017 00:20:56 +0000 (16:20 -0800)]
fix the adding (#15121)

6 years agoAdd a core dump example.
Santiago Fernandez Madero [Tue, 21 Nov 2017 00:10:57 +0000 (16:10 -0800)]
Add a core dump example.

6 years agoFix typo (#15129)
dotnet bot [Mon, 20 Nov 2017 23:54:50 +0000 (15:54 -0800)]
Fix typo (#15129)

Signed-off-by: dotnet-bot <dotnet-bot@microsoft.com>
6 years agoSet reg conflicts on contained IND
Carol Eidt [Wed, 15 Nov 2017 05:37:12 +0000 (21:37 -0800)]
Set reg conflicts on contained IND

For shift, rotate, mod and div, we were only excluding conflicting registers on the non-fixed operand if it is not contained. However, we still want to do that for the base and index (if any) on a contained IND.
Also, fix the LocationInfoListNodePool to attribute its memory usage to CMK_LSRA.

6 years agoFix a few of createdump bugs. (#15079)
Mike McLaughlin [Mon, 20 Nov 2017 23:02:22 +0000 (15:02 -0800)]
Fix a few of createdump bugs. (#15079)

1) Fix createdump p_vaddr bias/base problem.
2) createdump was loading the DAC from the directory where createdump was executed from and not from the libcoreclr.so path. Fixed this.
3) Don't use the dac if the libcoreclr.so module is not found.
4) Remove the QI of ICoreDebugDataTarget4 in datatarget.cpp. Forgot to do this when the VirtualUnwind function was removed. Caused crash.

6 years agoMerge pull request #15120 from sdmaclea/PR-ARM64-SIMD-SIMD12-genCodeForStoreLcl
Carol Eidt [Mon, 20 Nov 2017 22:58:37 +0000 (14:58 -0800)]
Merge pull request #15120 from sdmaclea/PR-ARM64-SIMD-SIMD12-genCodeForStoreLcl

[Arm64] SIMD12 genCodeForStoreLcl*

6 years agoMerge pull request #15087 from BruceForstall/FixJmpGsCheck
Bruce Forstall [Mon, 20 Nov 2017 22:29:55 +0000 (14:29 -0800)]
Merge pull request #15087 from BruceForstall/FixJmpGsCheck

Fix RyuJIT/arm32 GS cookie check before JMP call

6 years agoMerge pull request #15084 from BruceForstall/FixMulLongBashing3
Bruce Forstall [Mon, 20 Nov 2017 22:27:25 +0000 (14:27 -0800)]
Merge pull request #15084 from BruceForstall/FixMulLongBashing3

Fix RyuJIT/arm32 GT_MUL_LONG node bashing

6 years agoMerge pull request #15086 from BruceForstall/FixBitCastGenTreeDefn2
Bruce Forstall [Mon, 20 Nov 2017 22:23:57 +0000 (14:23 -0800)]
Merge pull request #15086 from BruceForstall/FixBitCastGenTreeDefn2

Fix GenTree definition for GT_BITCAST

6 years agoFix debuggable GenTree (#15081)
Bruce Forstall [Mon, 20 Nov 2017 22:23:25 +0000 (14:23 -0800)]
Fix debuggable GenTree (#15081)

* Fix debuggable GenTree

Since the `DEBUGGABLE_GENTREE` hack was introduced, many new
GenTree types have been introduced, and gtstructs.h has shifted
around. This change fixes it so we should see accurate GenTree
vtable pointers for all cases where SetOper gets called.

* Formatting

6 years agoUpdate jitbench harness to capture steady state data
Andy Ayers [Fri, 17 Nov 2017 00:14:06 +0000 (16:14 -0800)]
Update jitbench harness to capture steady state data

Update jitbench SHA to pick up new version that reports steady state
median response time, with fractional digits.

6 years agoFix genPrepForEHCodegen for arm64
Bruce Forstall [Mon, 20 Nov 2017 22:05:03 +0000 (14:05 -0800)]
Fix genPrepForEHCodegen for arm64

Enable amd64 code for arm64 as well, to match the expectation of
where labels exist in genReportEH. Fixes #14831, a unique case
during MinOpts (and probably only minopts) where dead code follows
a call/finally block sequence.

6 years agoMerge pull request #15115 from CarolEidt/Enable14665Tests
Carol Eidt [Mon, 20 Nov 2017 21:07:56 +0000 (13:07 -0800)]
Merge pull request #15115 from CarolEidt/Enable14665Tests

Re-enable tests disabled for #14665

6 years agoMerge pull request #14933 from sdmaclea/PR-ARM64-SIMD-Normalize-Struct-Return
Bruce Forstall [Mon, 20 Nov 2017 21:06:26 +0000 (13:06 -0800)]
Merge pull request #14933 from sdmaclea/PR-ARM64-SIMD-Normalize-Struct-Return

[Arm64] SIMD Normalize struct return type

6 years agoMerge pull request #12552 from mikedn/switch-bt
Bruce Forstall [Mon, 20 Nov 2017 21:01:09 +0000 (13:01 -0800)]
Merge pull request #12552 from mikedn/switch-bt

Use BT in switch lowering

6 years agoMerge pull request #15117 from sdmaclea/PR-ARM64-SIMD-SIMDIntrinsicSet-life
Bruce Forstall [Mon, 20 Nov 2017 20:57:33 +0000 (12:57 -0800)]
Merge pull request #15117 from sdmaclea/PR-ARM64-SIMD-SIMDIntrinsicSet-life

[Arm64] SIMD fix SIMDIntrinsicSet* op2 life

6 years ago[Arm64] SIMD12 genCodeForStoreLcl*
Steve MacLean [Mon, 20 Nov 2017 20:34:03 +0000 (15:34 -0500)]
[Arm64] SIMD12 genCodeForStoreLcl*

6 years agoMove MemoryStream to shared CoreLib partition (#15116)
Jan Kotas [Mon, 20 Nov 2017 20:19:01 +0000 (12:19 -0800)]
Move MemoryStream to shared CoreLib partition (#15116)

- Apply formating and other minor changes from CoreRT
    - Prefer async implementations from CoreCLR since they are more recent and better optimized
- Apply similar changes to UnmanagedMemoryStream as well to keep the two close

6 years ago[Arm64] SIMD fix SIMDIntrinsicSet* op2 life
Steve MacLean [Mon, 20 Nov 2017 17:40:09 +0000 (12:40 -0500)]
[Arm64] SIMD fix SIMDIntrinsicSet* op2 life

6 years agoRe-enable tests disabled for #14665
Carol Eidt [Mon, 20 Nov 2017 18:15:58 +0000 (10:15 -0800)]
Re-enable tests disabled for #14665

6 years agoMove TaskToApm to shared CoreLib partition (#15113)
Jan Kotas [Mon, 20 Nov 2017 16:42:47 +0000 (08:42 -0800)]
Move TaskToApm to shared CoreLib partition (#15113)

- Get TaskToApm in sync with CoreFX copy and move it to shared CoreLib partition
- Delete redundant __Error file
- Delete remaining uses of InternalBlockCopy and replace it with BlockCopy

6 years agoMerge pull request #15088 from BruceForstall/FixArmStressLclFlds
Bruce Forstall [Mon, 20 Nov 2017 05:26:24 +0000 (21:26 -0800)]
Merge pull request #15088 from BruceForstall/FixArmStressLclFlds

Fix STRESS_LCL_FLDS for ARM

6 years agoUpdate BuildTools, CoreClr, CoreFx to prerelease-02219-01, preview1-25920-01, preview...
dotnet-maestro-bot [Mon, 20 Nov 2017 04:01:28 +0000 (20:01 -0800)]
Update BuildTools, CoreClr, CoreFx to prerelease-02219-01, preview1-25920-01, preview1-25920-01, respectively (#15099)

6 years agoAllow ByRef return types in Dynamic Method (#15106)
Alex [Mon, 20 Nov 2017 02:32:25 +0000 (04:32 +0200)]
Allow ByRef return types in Dynamic Method (#15106)

* Allow ByRef return types in Dynamic Method

* Additional cleanup

6 years agoFix two issues in managed number formatting (#15105)
Stephen Toub [Sun, 19 Nov 2017 21:10:30 +0000 (16:10 -0500)]
Fix two issues in managed number formatting (#15105)

* Fix ValueStringBuilder to return correct array to pool

When growing the builder, it's erroneously returning the new array rather than the old one.

* Fix number formatting to not share static buffer

Was missing [ThreadStatic] on a static, such that parallel formatting of numbers might end up sharing the same buffer.  Oops.  But instead of just adding [ThreadStatic], I've switched to starting with a relatively small buffer on the stack.  It should suffice for the majority of performed formatting, and the ValueStringBuilder will grow for longer needs.

* Work around C# compiler warning about stack space being exposed

6 years agoDelete FriendAccessAllowedAttribute and associated dead code (#15101)
Jan Kotas [Sun, 19 Nov 2017 15:41:14 +0000 (07:41 -0800)]
Delete FriendAccessAllowedAttribute and associated dead code (#15101)

- Preserved commented out FriendAccessAllowed on types and methods that are still used as friends in System.Runtime.WindowsRuntime

6 years agoMove debugger attributes to shared partition (#15100)
Jan Kotas [Sun, 19 Nov 2017 15:39:48 +0000 (07:39 -0800)]
Move debugger attributes to shared partition (#15100)

- Break down DebuggerAttributes.cs into more files, apply cleanup done in CoreRT
- Move two outliers (DebuggerStepperBoundaryAttribute, DebuggerVisualizerAttribute) from CoreFX to CoreLib so that all debugger attributes are together

6 years agoDelete SuppressUnmanagedCodeSecurity on QCalls
Jan Kotas [Sun, 19 Nov 2017 13:49:08 +0000 (05:49 -0800)]
Delete SuppressUnmanagedCodeSecurity on QCalls

6 years agoUse BT in switch lowering
Mike Danes [Sat, 29 Apr 2017 10:50:26 +0000 (13:50 +0300)]
Use BT in switch lowering

6 years agoDelete unnecesary SuppressUnmanagedCodeSecurity annotations (#15102)
Jan Kotas [Sun, 19 Nov 2017 06:41:39 +0000 (22:41 -0800)]
Delete unnecesary SuppressUnmanagedCodeSecurity annotations (#15102)

6 years agoMerge pull request #15069 from stephentoub/coreclrnumbershared
Jan Kotas [Sun, 19 Nov 2017 03:24:14 +0000 (19:24 -0800)]
Merge pull request #15069 from stephentoub/coreclrnumbershared

Add TryFormat span-based methods to most primitive types

6 years agoUse TryFormat in StringBuilder
Stephen Toub [Fri, 17 Nov 2017 00:05:50 +0000 (19:05 -0500)]
Use TryFormat in StringBuilder

Avoid string allocations when appending primitives to StringBuilder.  We try to format into the existing array space, and only fall back to using ToString when there isn't enough room and we're going to grow the builder anyway.

6 years agoAdd TryFormat methods to {S}Byte, {U}Int16/32/64, and Boolean
Stephen Toub [Thu, 16 Nov 2017 18:59:51 +0000 (13:59 -0500)]
Add TryFormat methods to {S}Byte, {U}Int16/32/64, and Boolean

6 years agoChange FormatFixed to avoid bounds checks
Stephen Toub [Thu, 16 Nov 2017 17:40:04 +0000 (12:40 -0500)]
Change FormatFixed to avoid bounds checks

Match the original native code's use of pointers.

6 years agoRemove now dead code from managed parsing/formatting.
Stephen Toub [Wed, 15 Nov 2017 23:18:20 +0000 (18:18 -0500)]
Remove now dead code from managed parsing/formatting.

- Remove dead "bigNumber" code.
- Remove custom wcslen function. Use String's.
- Delete dead fcalls from runtime. FormatInt32, FormatUInt32, FormatInt64, and FormatUInt64 are no longer needed.  Delete them and many of the helpers used only by them.

6 years agoImprove performance of managed formatting code
Stephen Toub [Wed, 15 Nov 2017 17:54:06 +0000 (12:54 -0500)]
Improve performance of managed formatting code

- Optimize NumberBuffer passing by reference instead of value. It's a large struct of ~50 bytes; copying it around has non-trivial cost.
- Replace formatting StringBuilder with ref struct and stack allocation. Avoids lots of allocation and associated throughput costs.
- Improve perf of 'D' formatting of 32-bit and 64-bit integers.
- Remove array allocations accessing NumberFormatInfo props.
- Accessing array properties like PercentGroupSizes clones the corresponding field.  That's unnecessary here, as we don't mutate the array.
- Remove int[] allocation from NumberToStringFormat. Span makes it easy to start with stack space and grow to an allocated array as needed.
- Improve perf of hex formatting of integers. Including removing some sizable allocations.
- Manually inline several hot functions called in only one place.
- Tweak some range comparisons in ParseFormatSpecifier.
- Avoid large stackallocs in NumberToString{Fixed}. It's incurring non-trivial overheads.
- Tweak perf of ValueStringBuilder. In particular, make Append(string) faster for the single-char case, which is extremely common in integer formatting due to its prevalence in strings in NumberFormatInfo.

6 years agoMove integer parsing/formatting logic to shared
Stephen Toub [Tue, 14 Nov 2017 21:49:43 +0000 (16:49 -0500)]
Move integer parsing/formatting logic to shared

Moves existing managed parsing code to shared, to be shared with corert.
Takes the managed formatting port from corert and moves that to shared
as well.

6 years agoDelete unused Thread::YieldTask (#15091)
Jan Kotas [Sat, 18 Nov 2017 20:02:58 +0000 (12:02 -0800)]
Delete unused Thread::YieldTask (#15091)

6 years agoUpdate BuildTools, CoreClr, CoreFx, PgoData to prerelease-02217-03, preview1-25918...
dotnet-maestro-bot [Sat, 18 Nov 2017 16:25:54 +0000 (08:25 -0800)]
Update BuildTools, CoreClr, CoreFx, PgoData to prerelease-02217-03, preview1-25918-04, preview1-25917-04, master-20171116-0044, respectively (#15094)

6 years agoFix #15057 MemoryStream.CopyToAsync calls Write / WriteAsync with count=0 (#15070)
David Nissimoff [Sat, 18 Nov 2017 15:11:06 +0000 (07:11 -0800)]
Fix #15057 MemoryStream.CopyToAsync calls Write / WriteAsync with count=0 (#15070)

6 years agoRename mscorlib to System.Private.CoreLib in documentation (#15090)
Jan Kotas [Sat, 18 Nov 2017 15:10:35 +0000 (07:10 -0800)]
Rename mscorlib to System.Private.CoreLib in documentation (#15090)

Fixes #15078

6 years agoFix up ConfiguredValueTaskAwaitable signature for new compiler checks (#15092)
Jan Kotas [Sat, 18 Nov 2017 09:03:29 +0000 (01:03 -0800)]
Fix up ConfiguredValueTaskAwaitable signature for new compiler checks (#15092)

6 years agoFix MacOSX build break (#15093)
Jan Kotas [Sat, 18 Nov 2017 09:02:40 +0000 (01:02 -0800)]
Fix MacOSX build break (#15093)

Fixes #15005

6 years agoFix STRESS_LCL_FLDS for ARM
Bruce Forstall [Sat, 18 Nov 2017 02:01:10 +0000 (18:01 -0800)]
Fix STRESS_LCL_FLDS for ARM

The code in lvaStressLclFldCB to force alignment for ARM was
using the ADDR node, not the LCL_VAR node. For LONG vars, this meant
computing an incorrect offset into the morphed BLK type local
(4 instead of 8).

Partially fixes #14862

6 years agoFix RyuJIT/arm32 GS cookie check before JMP call
Bruce Forstall [Sat, 18 Nov 2017 00:48:56 +0000 (16:48 -0800)]
Fix RyuJIT/arm32 GS cookie check before JMP call

The GS cookie check was using r2/r3 registers, after they had been
reloaded as outgoing argument registers for the JMP call, thus
trashing them. Change the temp regs used to r12/lr, the only
non-argument, non-callee-saved registers available on arm32.

Partially fixes #14862

6 years agoRemove deprecated Copy Task (#15050)
Karthik Rajasekaran [Fri, 17 Nov 2017 23:26:52 +0000 (15:26 -0800)]
Remove deprecated Copy Task (#15050)

6 years agoFix GenTree definition for GT_BITCAST
Bruce Forstall [Fri, 17 Nov 2017 23:15:29 +0000 (15:15 -0800)]
Fix GenTree definition for GT_BITCAST

Nothing should be defined as GenTreeUnOp; use GenTreeOp instead.

6 years agoFix RyuJIT/arm32 GT_MUL_LONG node bashing
Bruce Forstall [Fri, 17 Nov 2017 23:19:15 +0000 (15:19 -0800)]
Fix RyuJIT/arm32 GT_MUL_LONG node bashing

For RyuJIT/arm32, during decomposition we can bash GT_MUL to
GT_MUL_LONG, which changes its type from GenTreeOp to
GenTreeMultiRegOp. When we do this, we need to initialize
the additional fields in GenTreeMultiRegOp. If we don't,
we get asserts in JitDumps, and presumably other bad things
happen.

6 years agoEnable FEATURE_DATATARGET4 for Unix Arm builds (#15060)
Chris Sienkiewicz [Fri, 17 Nov 2017 21:55:43 +0000 (13:55 -0800)]
Enable FEATURE_DATATARGET4 for Unix Arm builds (#15060)

6 years agofix for il to native mapping etw events (#15038)
David Mason [Fri, 17 Nov 2017 19:07:31 +0000 (11:07 -0800)]
fix for il to native mapping etw events (#15038)

After investigation the failures look unrelated to my changes. Thanks Jan and Noah.

6 years agoGeneric hardware intrinsics throw exception for non-numeric types (#15068)
Fei Peng [Fri, 17 Nov 2017 17:42:29 +0000 (09:42 -0800)]
Generic hardware intrinsics throw exception for non-numeric types (#15068)

6 years agoDo not check stmts after the tail call candidate if the optimization has been aborted...
Sergey Andreenko [Fri, 17 Nov 2017 02:48:18 +0000 (18:48 -0800)]
Do not check stmts after the tail call candidate if the optimization has been aborted already. (#15065)

Fix #15056

6 years agoMerge pull request #15066 from BruceForstall/CleanUpGroovy2
Bruce Forstall [Fri, 17 Nov 2017 01:09:18 +0000 (17:09 -0800)]
Merge pull request #15066 from BruceForstall/CleanUpGroovy2

Fix two typos in trigger phrases

6 years agoFix two typos in trigger phrases
Bruce Forstall [Fri, 17 Nov 2017 00:58:33 +0000 (16:58 -0800)]
Fix two typos in trigger phrases

6 years agoClean up netci.groovy (#15030)
Bruce Forstall [Thu, 16 Nov 2017 23:20:57 +0000 (15:20 -0800)]
Clean up netci.groovy (#15030)

* Clean up netci.groovy

1. Don't repeat entire list of JIT stress and R2R scenarios over and over;
use functions to check for scenario groups.
2. Merge handling of R2R baseline and stress scenarios.
3. Add function to determine if a scenario is a JIT stress scenario.
4. Fix JIT stress mode display name generator to handle CoreFx tests
as well, preserving the way they are currently displayed.
5. Fix R2R display name generator to handle baseline R2R scenario.
6. Fix bug where x86 R2R baseline job wasn't created.
7. Fix bugs where triggering some jobs, such as R2R jobs, would trigger lots
of other jobs with insufficient trigger strings. Thus, more trigger strings
now have architecture + scenario in them. E.g., "Innerloop" jobs now much
be triggered using a phrase containing "Innerloop".
8. Simplify code that creates builds, and test run commands.
9. Fix mis-casing of COMPlus_JITMinOpts in R2R stress list.
10. Expand list of valid ARM Windows scenarios to match total list of
scenarios, but with some scenarios commented out. Add the
jitstress2 + jitstressregs to the acceptable list. Add note about the
many scenarios in this list that are not in the full scenario list. Add
function to determine if a scenario is in this list.
11. Support crossgen of frameworks using altjit
12. Don't support 'r2r' scenario on arm/armlb/arm64 Windows.

6 years agofix format (#15064)
Sergey Andreenko [Thu, 16 Nov 2017 23:09:19 +0000 (15:09 -0800)]
fix format (#15064)

6 years agoMerge pull request #15063 from dotnet-bot/from-tfs
Brian Sullivan [Thu, 16 Nov 2017 22:52:41 +0000 (14:52 -0800)]
Merge pull request #15063 from dotnet-bot/from-tfs

Merge changes from TFS

6 years agoFix desktop build breaks.
dotnet-bot [Thu, 16 Nov 2017 21:17:49 +0000 (13:17 -0800)]
Fix desktop build breaks.

[tfs-changeset: 1681307]

6 years agoFix 14455: _opt_relrotarg_valref test failure on amd64. (#15032)
Sergey Andreenko [Thu, 16 Nov 2017 18:38:46 +0000 (10:38 -0800)]
Fix 14455: _opt_relrotarg_valref test failure on amd64. (#15032)

* fix the issue

Do not zero-initialize temps, that were created by jit and do not have GC refs.
The confusion was that `varDsc->lvIsTemp` means `short live` variable, when we wanted to ask is it IL temp or not.

* add a non-stress repro.

6 years agodon't use Compiler:: if we are currently in this namescape. (#15045)
Sergey Andreenko [Thu, 16 Nov 2017 04:36:41 +0000 (20:36 -0800)]
don't use Compiler:: if we are currently in this namescape. (#15045)

I was confused that we were outside of Compiler::

6 years agoHashCode based on xxHash32 (#14863)
Jonathan Dickinson [Thu, 16 Nov 2017 02:47:00 +0000 (18:47 -0800)]
HashCode based on xxHash32 (#14863)

* HashCode based on xxHash32

Works by maintaining the xxHash32 state variables (v1 -> v4, length) as well as a queue of values that fall outside of the block size (16 bytes/4 ints). The seed is initialized to random bytes.

Further details, unit tests and history: https://github.com/dotnet/corefx/pull/25013

6 years agoAdded EmitCalli overload (#15040)
Ivan Cronyn [Thu, 16 Nov 2017 02:37:22 +0000 (02:37 +0000)]
Added EmitCalli overload (#15040)

6 years agoImprove SSA dominator computation memory usage (#14965)
mikedn [Wed, 15 Nov 2017 23:14:22 +0000 (01:14 +0200)]
Improve SSA dominator computation memory usage (#14965)

* Improve SSA DF computation memory usage

DF(b) can be stored in a vector instead of a hashtable (set). Nothing needs a O(1) membership test, the duplicates that may be generated during DF construction can be easily avoided by observing that:
* each block in the graph is processed exactly once
* during processing the block is added to the relevant DF vectors, at the end of each vector. If the same DF vector is encountered multiple times then checking if the block is already a member is trivial.

* Use the same block BitVec for topo sort and IDom

* Add BitVecOps::TryAddElemD

Makes the code more readable and avoids the duplicated IsShort test that a separate IsMember/AddElemD may generate.

* Improve SSA IDF computation memory usage

Like DF(b), IDF(b) can be stored in a vector if duplicates are avoided. This can be done by using a BitVector like TopologicalSort and ComputeImmediateDom already do. It's cheaper than using a hashtable (even though it requires a "clear" for every block that has a frontier).

Also, storing IDF(b) into a vector makes it easier to track newly added blocks - they're at the end of the vector. Using a separate "delta" set is no longer needed.

6 years agochange labels to use new Ubuntu machine pool (#15043)
Victor "Nate" Graf [Wed, 15 Nov 2017 22:28:57 +0000 (14:28 -0800)]
change labels to use new Ubuntu machine pool (#15043)

6 years agoMerge pull request #15031 from dotnet-maestro-bot/master-UpdateDependencies
JC Aguilera [Wed, 15 Nov 2017 22:04:21 +0000 (14:04 -0800)]
Merge pull request #15031 from dotnet-maestro-bot/master-UpdateDependencies

Update BuildTools, CoreClr, CoreFx, PgoData to prerelease-02215-01, preview1-25915-02, preview1-25915-02, master-20171115-1118, respectively (master)

6 years agoMerge pull request #15039 from fiigii/fix
Bruce Forstall [Wed, 15 Nov 2017 21:32:04 +0000 (13:32 -0800)]
Merge pull request #15039 from fiigii/fix

Fix #15035

6 years agoFire tracepoints for methods with no template (#14987)
Victor "Nate" Graf [Wed, 15 Nov 2017 21:29:47 +0000 (13:29 -0800)]
Fire tracepoints for methods with no template (#14987)

6 years agoAdding null check for implicit cast from array/arraysegment to Span (#15044)
Ahson Khan [Wed, 15 Nov 2017 21:25:01 +0000 (13:25 -0800)]
Adding null check for implicit cast from array/arraysegment to Span (#15044)

6 years agoRe-enable Standalone GC job, running daily (#15015)
Sean Gillespie [Wed, 15 Nov 2017 21:18:42 +0000 (13:18 -0800)]
Re-enable Standalone GC job, running daily (#15015)

* Re-enable Standalone GC job, running daily

* osGroup is already defined

* Fix failed assertion

6 years agoUpdate BuildTools, CoreClr, CoreFx, PgoData to prerelease-02215-01, preview1-25915...
dotnet-maestro-bot [Wed, 15 Nov 2017 19:37:53 +0000 (11:37 -0800)]
Update BuildTools, CoreClr, CoreFx, PgoData to prerelease-02215-01, preview1-25915-02, preview1-25915-02, master-20171115-1118, respectively

6 years agoDelete move to return spill in the implicit tail call in the inlinee with several...
Sergey Andreenko [Wed, 15 Nov 2017 18:52:27 +0000 (10:52 -0800)]
Delete move to return spill in the implicit tail call in the inlinee with several returns. (#14945)

* add fgNeedReturnSpillTemp
* fix the issue
* add the repro without stress mode

6 years agoFix #15035
Fei Peng [Wed, 15 Nov 2017 09:22:59 +0000 (01:22 -0800)]
Fix #15035

6 years agoRemove out-of-use method and properties from appdomain (#15029)
Zhukov Maxim [Wed, 15 Nov 2017 02:30:43 +0000 (05:30 +0300)]
Remove out-of-use method and properties from appdomain (#15029)

Not supported on CoreCLR appdomain's method and properties were removed

#15001

6 years agoEnable Globalization Invariant mode using Environment Variable (#15024)
Tarek Mahmoud Sayed [Wed, 15 Nov 2017 01:45:08 +0000 (17:45 -0800)]
Enable Globalization Invariant mode using Environment Variable (#15024)

* Enable Globalization Invariant mode using Environment Variable

* Prefix the environment variable name with DOTNET_

* Address the feedback

6 years agoUpdate BuildTools, CoreClr, CoreFx, PgoData to prerelease-02214-03, preview1-25914...
dotnet-maestro-bot [Wed, 15 Nov 2017 00:46:49 +0000 (16:46 -0800)]
Update BuildTools, CoreClr, CoreFx, PgoData to prerelease-02214-03, preview1-25914-01, preview1-25913-03, master-20171113-1530, respectively (#15007)

6 years agomove fgFixupStructReturn (#14966)
Sergey Andreenko [Wed, 15 Nov 2017 00:30:52 +0000 (16:30 -0800)]
move fgFixupStructReturn (#14966)

It was done twice during the tail call optimization and in the args morphing. The problem is if tail call optimization was aborted it was not clear was `fgFixupStructReturn` called or not.

6 years agoJIT: look through GT_RET_EXPR when forming inline candidates
Andy Ayers [Fri, 10 Nov 2017 01:07:12 +0000 (17:07 -0800)]
JIT: look through GT_RET_EXPR when forming inline candidates

If an inline candidate has args that come from calls that are inline
candidates, the arg trees for the candidate may be GT_RET_EXPRs. Since
the jit inlines in preorder, any arg inlines will have been resolved
by the time we get to the candidate. So, the jit can look back through
the GT_RET_EXPR to get a better handle on the actual arg tree.

Doing this has several small benefits:
* It short-circuits parts of the subsequent GT_RET_EXPR update and so
saves a little bit of throughput.
* It potentially allows for more streamlined arg passing since the actual
node side effects may be less constraining than the placeholder effects on
the GT_RET_EXPR (which must be "worst case").
* It may unblock inlines as actual arg values can be seen when evaluating
the candidate for profitability.

During testing I found cases where looking at the actual arg seemed to
degrade information. This turned out to be a misuse of `argHasLdargaOp`
to indicate that a caller-supplied argument was potentially locally
aliased.

This misuse was blocking type propagation when the actual arg tree was seen to
be an aliasable local reference instead of the formerly opaque GT_RET_EXPR.
To fix this, I split out the aliased arg case with a different flag. All
that the jit needs to ensure in the aliased case is that it does not directly
substitute the arg; the arg must be evaluated at the point of the call. The
type of the actual can safely propagate to the arg as long as the arg itself
is not modifiable.

Added a utility for walking back through GT_RET_EXPRs and updated the
other case where the jit was doing this to use the utility as well.

Closes #14174.

6 years agoMerge pull request #15014 from fiigii/vexencoding
Carol Eidt [Tue, 14 Nov 2017 22:39:05 +0000 (14:39 -0800)]
Merge pull request #15014 from fiigii/vexencoding

Change VEX-encoding selection to avoid AVX-SSE transition penalties

6 years agoMerge pull request #15026 from jashook/re_add_arm64_stress
Jarret Shook [Tue, 14 Nov 2017 22:01:34 +0000 (14:01 -0800)]
Merge pull request #15026 from jashook/re_add_arm64_stress

Add jitstress arm64 jobs again

6 years agoMerge pull request #14914 from CarolEidt/Fix14665
Carol Eidt [Tue, 14 Nov 2017 21:59:31 +0000 (13:59 -0800)]
Merge pull request #14914 from CarolEidt/Fix14665

ARMARCH: no cascaded adds in addr mode

6 years agosmall changes to get coreclr building on FreeBSD again (#14964)
Tomas Weinfurt [Tue, 14 Nov 2017 21:44:25 +0000 (13:44 -0800)]
small changes to get coreclr building on FreeBSD again (#14964)

6 years agoAdd jitstress arm64 jobs again
jashook [Tue, 14 Nov 2017 21:22:48 +0000 (13:22 -0800)]
Add jitstress arm64 jobs again

6 years agoMerge pull request #15023 from BruceForstall/FixJitOpts
Bruce Forstall [Tue, 14 Nov 2017 20:59:01 +0000 (12:59 -0800)]
Merge pull request #15023 from BruceForstall/FixJitOpts

Fix RunTest.cmd argument parsing

6 years agoChange VEX-encoding selection to avoid AVX-SSE transition penalties
Fei Peng [Tue, 14 Nov 2017 05:42:57 +0000 (21:42 -0800)]
Change VEX-encoding selection to avoid AVX-SSE transition penalties

6 years agoFix RunTest.cmd argument parsing
Bruce Forstall [Tue, 14 Nov 2017 18:23:00 +0000 (10:23 -0800)]
Fix RunTest.cmd argument parsing

The "jitminopts" and "jitforcerelocs" options were eating whatever
option followed them.