platform/upstream/coreclr.git
7 years agoMerge pull request #13364 from hqueue/arm/ryujit/issue_13363_regression_of_13190
Bruce Forstall [Mon, 14 Aug 2017 22:16:44 +0000 (15:16 -0700)]
Merge pull request #13364 from hqueue/arm/ryujit/issue_13363_regression_of_13190

[RyuJIT/ARM32] Fix regression

7 years ago[Linux/ARM] Fix managed breakpoints (#13316)
Igor Kulaychuk [Mon, 14 Aug 2017 21:30:56 +0000 (00:30 +0300)]
[Linux/ARM] Fix managed breakpoints (#13316)

* [Linux/ARM] Fix managed breakpoints

This commit introduces the following changes in order to enable
ICorDebug-based debuggers to use breakpoints on ARM Linux:

* Use 0xde01 as breakpoint instruction on ARM Linux.
  ARM reference recommends to use 0xdefe as a breakpoint instruction,
  but Linux kernel generates SIGILL for this instruction.
  The 0xde01 instruction causes the kernel to generate SIGTRAP.

* Fix SIGTRAP handling on ARM Linux.
  Unlike x86, when SIGTRAP happens on ARM Linux, the PC points
  at the break instruction. But the rest of the code expects
  that it points to an instruction after the break,
  so we adjust the PC at the start of HandleHardwareException().

* Enable ARM single stepping for PAL.
  Handle single stepping for PAL path the same way as for non-PAL path.
  Also enable ArmSingleStepper executable buffer by allocating it
  from system global loader executable heap.

* Hande ARM single step only when debugger is attached, fix comments and code style

* Pass existing Thread object to HandleArmSingleStep

7 years agoMerge pull request #13350 from BruceForstall/FixVSO478352
Bruce Forstall [Mon, 14 Aug 2017 17:30:38 +0000 (10:30 -0700)]
Merge pull request #13350 from BruceForstall/FixVSO478352

Fix double-indirect p/invoke call generation

7 years agoAdded SetThreadDescription to set the unmanaged thread name (#12593)
Alois-xx [Mon, 14 Aug 2017 17:26:28 +0000 (19:26 +0200)]
Added SetThreadDescription to set the unmanaged thread name (#12593)

* Added SetThreadDescription to set the unmanaged thread name as well when a managed thread name was set.
This will show up in future debuggers which know how to read that information or in ETW traces in the Thread Name column.

* use printf  instead of wprintf which exists on all platforms.

* Removed printf
Ensure that GetProceAddress is only called once to when the method is not present.
Potential perf hit should be negligible since setting a thread name can only happen once per managed thread.

* - Moved SetThreadName code to winfix.cpp as proposed
- Finalizer and threadpool threads get their name
- GCToEEInterface::CreateBackgroundThread is also named
- but regular GC threads have no name because when I included utilcode.h things did break apart.

* Fix for data race in g_pfnSetThreadDescription

* Fix string literals on unix builds.

* Fixed nits
Settled thread name on ".NET Core ThreadPool"

7 years agoIncrease timeout and log the result for a couple of WaitAny tests that expect Abandon...
Koundinya Veluri [Mon, 14 Aug 2017 11:23:16 +0000 (04:23 -0700)]
Increase timeout and log the result for a couple of WaitAny tests that expect AbandonedMutexException (#13298)

7 years ago[RyuJIT/ARM32] Fix regression
Hyung-Kyu Choi [Mon, 14 Aug 2017 10:18:13 +0000 (19:18 +0900)]
[RyuJIT/ARM32] Fix regression

- This fixes regression cuased by PR #13190
- Restore code for resolving registers for multi register operator

Signed-off-by: Hyung-Kyu Choi <hk0110.choi@samsung.com>
7 years agoTypo (#13360)
John Doe [Mon, 14 Aug 2017 06:47:41 +0000 (23:47 -0700)]
Typo (#13360)

7 years agoUpdate CoreClr, CoreFx to preview2-25614-01, preview2-25614-01, respectively (#13359)
dotnet-maestro-bot [Mon, 14 Aug 2017 06:47:19 +0000 (23:47 -0700)]
Update CoreClr, CoreFx to preview2-25614-01, preview2-25614-01, respectively (#13359)

7 years agoFinish moving Guid to shared partition
Jan Kotas [Sun, 13 Aug 2017 15:35:37 +0000 (08:35 -0700)]
Finish moving Guid to shared partition

Signed-off-by: dotnet-bot <dotnet-bot@microsoft.com>
7 years agoUpdate CoreClr, CoreFx to preview2-25613-02, preview2-25613-02, respectively (#13300)
dotnet-maestro-bot [Sun, 13 Aug 2017 19:14:04 +0000 (12:14 -0700)]
Update CoreClr, CoreFx to preview2-25613-02, preview2-25613-02, respectively (#13300)

7 years agoFollow-up: Add enforcepgo for release Windows_NT default release build (#13319)
Victor "Nate" Graf [Fri, 11 Aug 2017 23:43:51 +0000 (16:43 -0700)]
Follow-up: Add enforcepgo for release Windows_NT default release build (#13319)

* Add enforcepgo for release Windows_NT default release build

* Restrict enforment of PGO to PGO supported builds

* Expand the usage of enforcpgo

* Fix break in the if-else control flow

7 years agoMerge pull request #13332 from wateret/fix-ryuarm-hfa
Bruce Forstall [Fri, 11 Aug 2017 23:43:21 +0000 (16:43 -0700)]
Merge pull request #13332 from wateret/fix-ryuarm-hfa

[RyuJIT/armel] Fix PutArgReg BuildRefPositions

7 years agoMove files to shared CoreLib partition (#13342)
Jan Kotas [Fri, 11 Aug 2017 23:17:00 +0000 (16:17 -0700)]
Move files to shared CoreLib partition (#13342)

7 years agoFix double-indirect p/invoke call generation
Bruce Forstall [Fri, 11 Aug 2017 22:15:29 +0000 (15:15 -0700)]
Fix double-indirect p/invoke call generation

When the signature to getAddressOfPInvokeTarget() was changed,
the double-indirect CT_USER_FUNC case was accidentally changed
to pass a non-null `addr` to emitIns_Call(). This leads to asserts
when ngen'ing (in my test, ngen of System.dll on desktop using `LEGACY_BACKEND`
cross-compiling arm32 altjit). For some reason, x86 is more permissive
in the emitIns_Call() assert for this case, but doesn't use the
non-null addr.

Fixes VSO 478352

7 years agoFix bug where whitespace breaks arg proccessing (#13320)
Victor "Nate" Graf [Fri, 11 Aug 2017 22:10:29 +0000 (15:10 -0700)]
Fix bug where whitespace breaks arg proccessing (#13320)

* Fix bug where whitespace breaks arg proccessing

* Fix whitespace proccessing bug in buil-text.cmd

7 years agoMerge pull request #13338 from jashook/log_crossgen_command
Jarret Shook [Fri, 11 Aug 2017 21:08:55 +0000 (14:08 -0700)]
Merge pull request #13338 from jashook/log_crossgen_command

Log the crossgen command in build.sh.

7 years agoChanges that didn't make it to PR#12982 (#13055)
Cyd Haselton [Fri, 11 Aug 2017 20:14:36 +0000 (15:14 -0500)]
Changes that didn't make it to PR#12982 (#13055)

* Changes that didn't make it to PR#12982

* Recommended changes to volatile.h

Added two templates to cast away volatility due to clang 3.8 changes

* Update runtime.Linux.Microsoft.NETCore.Runtime.CoreCLR.props

Fixed erroneous extra Platform conditional line

7 years agoLog the crossgen command in build.sh.
jashook [Fri, 11 Aug 2017 15:56:20 +0000 (08:56 -0700)]
Log the crossgen command in build.sh.

This makes it easier to copy paste the commnand if it fails.

7 years agoFixed CoreRT issue https://github.com/dotnet/coreclr/issues/13195 (#13335)
sergey ignatov [Fri, 11 Aug 2017 14:39:04 +0000 (17:39 +0300)]
Fixed CoreRT issue https://github.com/dotnet/coreclr/issues/13195 (#13335)

7 years ago[RyuJIT/armel] Fix PutArgReg BuildRefPositions
Hanjoung Lee [Fri, 11 Aug 2017 07:19:02 +0000 (16:19 +0900)]
[RyuJIT/armel] Fix PutArgReg BuildRefPositions

We should fix registers for GT_PUTARG_REG, not GT_COPY or GT_MUL_LONG.
And merge the `if` statement with GT_PUTARG_SPLIT case.

7 years agoMerge pull request #13306 from hseok-oh/ryujit/fix_13156
Carol Eidt [Fri, 11 Aug 2017 06:31:17 +0000 (23:31 -0700)]
Merge pull request #13306 from hseok-oh/ryujit/fix_13156

Fix lclVar move node type inserted on lsra resolve phase

7 years agoFix genSpillVar
Hyeongseok Oh [Fri, 11 Aug 2017 05:00:02 +0000 (14:00 +0900)]
Fix genSpillVar

Fix genSpillVar() instead of insertMove()

7 years agoFix the race condition in mutex test using another manual reset event (#13308)
Ahson Ahmed Khan [Fri, 11 Aug 2017 02:56:03 +0000 (19:56 -0700)]
Fix the race condition in mutex test using another manual reset event (#13308)

* Fix the race condition by using a manual reset event instead of gc.keepalive

* Addressing PR comment

7 years agoMerge pull request #13285 from shimingsg/v-shige/add-perftc-GCLarge
Shiming Ge [Fri, 11 Aug 2017 01:37:26 +0000 (09:37 +0800)]
Merge pull request #13285 from shimingsg/v-shige/add-perftc-GCLarge

add perf test case GCLarge

7 years agoMerge pull request #13299 from pgavlin/GH13161
Pat Gavlin [Thu, 10 Aug 2017 23:02:03 +0000 (16:02 -0700)]
Merge pull request #13299 from pgavlin/GH13161

Fix issue #13161.

7 years agoImplement BitConverter Span-based APIs (#13278)
Nina Chikanov [Thu, 10 Aug 2017 23:01:27 +0000 (16:01 -0700)]
Implement BitConverter Span-based APIs (#13278)

Finished first checkbox of issue 22355: Implement in
System.Private.CoreLib in coreclr

7 years agoMerge pull request #13304 from BruceForstall/FixVSO471672
Bruce Forstall [Thu, 10 Aug 2017 23:00:37 +0000 (16:00 -0700)]
Merge pull request #13304 from BruceForstall/FixVSO471672

Copy additional GenTreeBoundsChk fields in gtCloneExpr

7 years agoJIT: modify box/unbox/isinst/castclass expansions for fast jitting (#13188)
Andy Ayers [Thu, 10 Aug 2017 22:17:31 +0000 (15:17 -0700)]
JIT: modify box/unbox/isinst/castclass expansions for fast jitting (#13188)

When the jit is generating code in debug/minopts/rare-block modes,
we'd prefer it to generate code more quickly and worry less about overall
generated code performance.

Generally speaking smaller intermediate and final code should correlate
well with faster jitting.

This change alters the expansions of box, unbox, isinst, and castclass when
generating code for minopts, debug, or in rarely run blocks. In such modes
the jit estimates whether an inline sequence or general helper call would
result in more compact code, and then chooses the smaller sequence.

This reduces generated code size around 2.5% in a variety of scenarios,
and roughly translates to a 1.5% improvement in time spent jitting.

Similar strategies can be applied to other complex operations during
importation. That work is forthcoming.

7 years agoRevert "Add enforcepgo for release Windows_NT default release build (#13275)" (#13318)
Jan Kotas [Thu, 10 Aug 2017 21:24:33 +0000 (14:24 -0700)]
Revert "Add enforcepgo for release Windows_NT default release build (#13275)" (#13318)

This reverts commit b07a921524ec9e8c8f5e482707f9fc1b9b86d95e.

7 years agoMerge pull request #13307 from wateret/cleanup-armel-lsracond
Bruce Forstall [Thu, 10 Aug 2017 16:20:59 +0000 (09:20 -0700)]
Merge pull request #13307 from wateret/cleanup-armel-lsracond

[RyuJIT/armel] Cleanup with OperIsMultiRegOp()

7 years agoMerge pull request #13309 from wateret/arm-getregcount
Bruce Forstall [Thu, 10 Aug 2017 16:09:11 +0000 (09:09 -0700)]
Merge pull request #13309 from wateret/arm-getregcount

[RyuJIT/armel] Introduce GetRegCount() for MultiRegOp

7 years agoAdd enforcepgo for release Windows_NT default release build (#13275)
Victor "Nate" Graf [Thu, 10 Aug 2017 15:47:44 +0000 (08:47 -0700)]
Add enforcepgo for release Windows_NT default release build (#13275)

* Add enforcepgo for release Windows_NT default release build

* Restrict enforment of PGO to PGO supported builds

* Expand the usage of enforcpgo

* Fix break in the if-else control flow

7 years agoMerge pull request #13286 from JosephTremoulet/TailLoopScratch
Joseph Tremoulet [Thu, 10 Aug 2017 14:19:38 +0000 (10:19 -0400)]
Merge pull request #13286 from JosephTremoulet/TailLoopScratch

Ensure scratch BB in tail recursion -> loop xform

7 years ago[RyuJIT/armel] Cleanup with OperIsMultiRegOp()
Hanjoung Lee [Thu, 10 Aug 2017 03:24:45 +0000 (12:24 +0900)]
[RyuJIT/armel] Cleanup with OperIsMultiRegOp()

Use OperIsMultiRegOp() instead of separated conditions

7 years agoremove empty line
Shiming Ge (Pactera Technologies) [Thu, 10 Aug 2017 07:49:59 +0000 (00:49 -0700)]
remove empty line

7 years ago[RyuJIT/armel] Introduce GetRegCount() for MultiRegOp
Hanjoung Lee [Thu, 10 Aug 2017 06:35:38 +0000 (15:35 +0900)]
[RyuJIT/armel] Introduce GetRegCount() for MultiRegOp

Abstract getting reg count with method GetRegCount()
Added the case when the reg count is 0.

7 years agoMerge pull request #13284 from wateret/fix-13281
Carol Eidt [Thu, 10 Aug 2017 04:27:20 +0000 (21:27 -0700)]
Merge pull request #13284 from wateret/fix-13281

[RyuJIT/arm32] Fix Regression

7 years agoFix lclVar move node type inserted in lsra phase
Hyeongseok Oh [Thu, 10 Aug 2017 01:55:51 +0000 (10:55 +0900)]
Fix lclVar move node type inserted in lsra phase

When we try to insert lclVal move node (load, store) on lsra phase,
fix type as TYP_INT when size is smaller than TYP_INT

7 years agoLinkBench: Fix Roslyn job
Swaroop Sridhar [Wed, 9 Aug 2017 21:32:26 +0000 (14:32 -0700)]
LinkBench: Fix Roslyn job

This change fixes a build-break in the LinkBench job, because
of a failure in the Roslyn Build.
With this change, Roslyn Benchmark is built using MsBuild (similar to
Roslyn's Build.cmd) instead of directly calling dotnet publish.

7 years agoCopy additional GenTreeBoundsChk fields in gtCloneExpr
Bruce Forstall [Thu, 10 Aug 2017 01:18:31 +0000 (18:18 -0700)]
Copy additional GenTreeBoundsChk fields in gtCloneExpr

We were previously losing the gtStkDepth field, which led to
an assert with the STRESS_CLONE_EXPR stress mode with array
bounds checks in function argument lists that had already
pushed stack arguments.

7 years agoMerge pull request #13190 from jyoungyun/ryujit/use_the_regtype_of_refposition_instea...
Carol Eidt [Thu, 10 Aug 2017 01:17:12 +0000 (18:17 -0700)]
Merge pull request #13190 from jyoungyun/ryujit/use_the_regtype_of_refposition_instead_of_its_type_in_the_gt_putarg_split_case

[ARM32/RyuJIT] Use regtype instead of node's type in PUTARG_SPLIT case

7 years agoadd the license header to the C# file
Shiming Ge (Pactera Technologies) [Thu, 10 Aug 2017 01:07:54 +0000 (18:07 -0700)]
add the license header to the C# file

7 years agoFix issue #13161.
Pat Gavlin [Wed, 9 Aug 2017 23:54:43 +0000 (16:54 -0700)]
Fix issue #13161.

This issue was a crash when calling R2R-compiled code from JITted code
on Windows/ARM32. The caller had a live value in `r4` that was scribbled
over by the callee. This change adds the necessary call to
`regSet.rsTrackRegTrash` which ensures that `r4` is saved/restored in
the prolog/epilog when it is used for R2R indirect calls.

7 years agoFixed issue #13282. (#13291)
Mike McLaughlin [Wed, 9 Aug 2017 21:41:13 +0000 (14:41 -0700)]
Fixed issue #13282. (#13291)

Changed the open process token from TOKEN_READ to TOKEN_QUERY.

7 years agoMerge pull request #13234 from janvorli/add-rhel6-official-rid
Wes Haggard [Wed, 9 Aug 2017 21:35:22 +0000 (14:35 -0700)]
Merge pull request #13234 from janvorli/add-rhel6-official-rid

Add RHEL 6 official RID

7 years agoMerge pull request #13268 from dotnet-maestro-bot/master-UpdateDependencies
Wes Haggard [Wed, 9 Aug 2017 21:18:44 +0000 (14:18 -0700)]
Merge pull request #13268 from dotnet-maestro-bot/master-UpdateDependencies

Update CoreClr, CoreFx to preview2-25609-02, preview2-25609-01, respectively (master)

7 years agoAdd fail-fast check after lvaGrabTemp in impImportBlockCode (#13271)
Sergey Andreenko [Wed, 9 Aug 2017 20:55:02 +0000 (13:55 -0700)]
Add fail-fast check after lvaGrabTemp in impImportBlockCode (#13271)

Fix #13231.

* responce Brian's comment.

7 years agoMerge pull request #12802 from pgavlin/VSO462269
Pat Gavlin [Wed, 9 Aug 2017 18:25:29 +0000 (11:25 -0700)]
Merge pull request #12802 from pgavlin/VSO462269

Fix VSO 462269.

7 years agotypo (#13287)
John Doe [Wed, 9 Aug 2017 18:09:28 +0000 (11:09 -0700)]
typo (#13287)

7 years agoStart at running ILLink to find dead code in Corelib (#13266)
Dan Moseley [Wed, 9 Aug 2017 18:08:32 +0000 (11:08 -0700)]
Start at running ILLink to find dead code in Corelib (#13266)

* Remove some dead code

* Remove from native side

* More

* Remove nls consts

* Fix break

* Fix break

7 years agoUpdate ryujit-tutorial.md
Carol Eidt [Wed, 9 Aug 2017 17:52:23 +0000 (10:52 -0700)]
Update ryujit-tutorial.md

Format the dump info for getting the flow graph

7 years agoFix VSO 462269.
Pat Gavlin [Thu, 13 Jul 2017 19:59:29 +0000 (12:59 -0700)]
Fix VSO 462269.

When decomposing a long compare on 32-bit platforms, the operands to the
decomposed compare must be sign- or zero-extended appropriately.

7 years agoUpdate CoreClr, CoreFx to preview2-25609-02, preview2-25609-01, respectively
dotnet-maestro-bot [Wed, 9 Aug 2017 12:55:33 +0000 (05:55 -0700)]
Update CoreClr, CoreFx to preview2-25609-02, preview2-25609-01, respectively

7 years agoMerge pull request #13270 from stephentoub/refactor_taskwait
Stephen Toub [Wed, 9 Aug 2017 12:45:24 +0000 (08:45 -0400)]
Merge pull request #13270 from stephentoub/refactor_taskwait

Refactor large Task NoOptimization bodies into own methods

7 years agoadd perf test case GCLarge
Shiming Ge (Pactera Technologies) [Wed, 9 Aug 2017 09:36:11 +0000 (02:36 -0700)]
add perf test case GCLarge

7 years ago[RyuJIT/arm32] Fix Regression
Hanjoung Lee [Wed, 9 Aug 2017 03:11:29 +0000 (12:11 +0900)]
[RyuJIT/arm32] Fix Regression

Fix regression caused by #13023

Fix #13281

7 years agoEnsure scratch BB in tail recursion -> loop xform
Joseph Tremoulet [Wed, 9 Aug 2017 02:11:23 +0000 (22:11 -0400)]
Ensure scratch BB in tail recursion -> loop xform

Our loop detection requires a predecessor outside the loop, so without
this the loops often don't get optimized as such.

7 years agoMerge pull request #13276 from dotnet-bot/from-tfs
Brian Sullivan [Wed, 9 Aug 2017 01:22:34 +0000 (18:22 -0700)]
Merge pull request #13276 from dotnet-bot/from-tfs

Merge changes from TFS

7 years agoCache both string and its length during GDBJIT setup (#13221)
Jonghyun Park [Wed, 9 Aug 2017 00:38:50 +0000 (09:38 +0900)]
Cache both string and its length during GDBJIT setup (#13221)

7 years agoAdd Encoding Span-based APIs (#13269)
Stephen Toub [Tue, 8 Aug 2017 21:45:16 +0000 (17:45 -0400)]
Add Encoding Span-based APIs (#13269)

* Add Encoding Span-based APIs

* Address PR feedback

7 years agoFix fo CoreCLR cofde formating error
Brian Sullivan [Tue, 8 Aug 2017 21:11:07 +0000 (14:11 -0700)]
Fix fo CoreCLR cofde formating error
Add extra spacing

[tfs-changeset: 1669566]

7 years agoFix possibly uninitialized warning
Brian Sullivan [Tue, 8 Aug 2017 20:45:48 +0000 (13:45 -0700)]
Fix possibly uninitialized warning

[tfs-changeset: 1669561]

7 years agoChange Task.WaitAll/Any overloads to call WaitAll/AnyCore directly
Stephen Toub [Tue, 8 Aug 2017 18:55:06 +0000 (14:55 -0400)]
Change Task.WaitAll/Any overloads to call WaitAll/AnyCore directly

Avoid multiple unnecessary levels of NoOptimization frames.

7 years agoAdd StringBuilder Span-based APIs (#13163)
Kevin Gosse [Tue, 8 Aug 2017 18:53:26 +0000 (20:53 +0200)]
Add StringBuilder Span-based APIs (#13163)

* Add StringBuilder Span-based APIs

* Factorized ThreadSafeCopy methods

The char[] based one was called only by CopyTo so it wasn't needed anymore. Replaced by a Span based version

* Fix possible overflow

7 years agoMerge pull request #12168 from gbalykov/remove-relocations-readonly
Bruce Forstall [Tue, 8 Aug 2017 18:28:27 +0000 (11:28 -0700)]
Merge pull request #12168 from gbalykov/remove-relocations-readonly

Partially remove relocations from SECTION_Readonly

7 years agoAdd build-time check to enforce PGO compilation for applicable DLL files (#13258)
Victor "Nate" Graf [Tue, 8 Aug 2017 18:14:42 +0000 (11:14 -0700)]
Add build-time check to enforce PGO compilation for applicable DLL files (#13258)

* Added PGO execution verification script

Added a script which can help verify whether or not PGO was used
to compile PE files passed to it

* Added target to enforce PGO compilation

* Remove OS checking for enforcepgo

7 years agoUse logf() instead of JitDump() (#13265)
Bruce Forstall [Tue, 8 Aug 2017 17:52:57 +0000 (10:52 -0700)]
Use logf() instead of JitDump() (#13265)

* Use logf() instead of JitDump()

This allows more JitDump output to get routed through the CLR
logging facility.

* Formatting

7 years agoImprove performance of string.IndexOfAny for 2 & 3 char searches (#13219)
Bruce Bowyer-Smyth [Tue, 8 Aug 2017 17:25:31 +0000 (03:25 +1000)]
Improve performance of string.IndexOfAny for 2 & 3 char searches (#13219)

7 years agoMerge pull request #13261 from pgavlin/VSO475427
Pat Gavlin [Tue, 8 Aug 2017 17:04:20 +0000 (10:04 -0700)]
Merge pull request #13261 from pgavlin/VSO475427

Fix VSO 475427.

7 years agoMerge pull request #13263 from BruceForstall/FixDesktopX86SPMIBuild
Bruce Forstall [Tue, 8 Aug 2017 16:40:34 +0000 (09:40 -0700)]
Merge pull request #13263 from BruceForstall/FixDesktopX86SPMIBuild

Fix SuperPMI x86 desktop build

7 years agoMerge pull request #13259 from pgavlin/VSO469210
Pat Gavlin [Tue, 8 Aug 2017 16:36:05 +0000 (09:36 -0700)]
Merge pull request #13259 from pgavlin/VSO469210

Fix VSO 469210.

7 years agoFormat code.
Pat Gavlin [Tue, 8 Aug 2017 16:35:47 +0000 (09:35 -0700)]
Format code.

7 years agoAdd smoke test for event pipe on Linux (#13240)
Victor "Nate" Graf [Tue, 8 Aug 2017 15:46:56 +0000 (08:46 -0700)]
Add smoke test for event pipe on Linux (#13240)

* Add smoke test for eventpipe

* Clean up some code from a different version

* Use more descriptive project and namespace names

7 years agoRefactor large Task NoOptimization bodies into own methods
Stephen Toub [Tue, 8 Aug 2017 15:45:49 +0000 (11:45 -0400)]
Refactor large Task NoOptimization bodies into own methods

The methods apparently need to be marked as [MethodImpl(MethodImplOptions.NoOptimization)] to aid the Visual Studio debugger in picking out these methods and their arguments, but it's causing a few chunky functions to have no optimizations on the whole body.  Separating these bodies out.

7 years agoMerge pull request #13260 from pgavlin/VSO471670
Pat Gavlin [Tue, 8 Aug 2017 15:21:13 +0000 (08:21 -0700)]
Merge pull request #13260 from pgavlin/VSO471670

Fix VSO 471670.

7 years ago[ARM32/RyuJIT] Set the source candidates for PutArgSplit nodes
Jiyoung Yun [Tue, 8 Aug 2017 09:44:43 +0000 (18:44 +0900)]
[ARM32/RyuJIT] Set the source candidates for PutArgSplit nodes

The Use RefPosition of PutArgSplit should use the same register with the Def RefPosition.

7 years ago[ARM32/RyuJIT] Use regtype instead of node's type in PUTARG_SPLIT case
Jiyoung Yun [Thu, 3 Aug 2017 05:31:12 +0000 (14:31 +0900)]
[ARM32/RyuJIT] Use regtype instead of node's type in PUTARG_SPLIT case

The size of TYP_STRUCT in definition is 1 and it is lower than sizeof(int).
It means TYP_STRUCT should use the node's type instead of its type.
In PUTARG_SPLIT case, it wil use its regtype of child nodes.

7 years agoUpdate CoreClr, CoreFx to preview2-25608-01, preview2-25608-01, respectively (#13262)
dotnet-maestro-bot [Tue, 8 Aug 2017 08:16:52 +0000 (01:16 -0700)]
Update CoreClr, CoreFx to preview2-25608-01, preview2-25608-01, respectively (#13262)

7 years agoMerge pull request #13023 from wateret/fix-12994
Carol Eidt [Tue, 8 Aug 2017 04:37:49 +0000 (21:37 -0700)]
Merge pull request #13023 from wateret/fix-12994

[RyuJIT/armel] Make RefPosition arg regs fixed

7 years agoRemove some dead code (#13257)
Dan Moseley [Tue, 8 Aug 2017 03:37:50 +0000 (20:37 -0700)]
Remove some dead code (#13257)

7 years agoFix SuperPMI x86 desktop build
Bruce Forstall [Tue, 8 Aug 2017 00:28:22 +0000 (17:28 -0700)]
Fix SuperPMI x86 desktop build

7 years agoFix VSO 475427.
Pat Gavlin [Tue, 8 Aug 2017 00:05:46 +0000 (17:05 -0700)]
Fix VSO 475427.

In addition to repredicting if the frame size is large enough to require
that we reserve `r10`, we also need to throw out any previous
prediction. If we do not, we may decide to reuse it if uses less stack
space than the reprediction.

7 years agoFix VSO 471670.
Pat Gavlin [Mon, 7 Aug 2017 23:41:56 +0000 (16:41 -0700)]
Fix VSO 471670.

If we remove a NOP during rationalize that is unused, we need to ensure
that its operand is also marked as an unused value.

7 years agoFix VSO 469210.
Pat Gavlin [Mon, 7 Aug 2017 23:26:56 +0000 (16:26 -0700)]
Fix VSO 469210.

This bug arises when ARM32/Legacy backend generates an RMW operator
whose source operand is an addressing mode. In that case, `sched_AM`
needs to choose a register to use for the value that will be loaded from
memory. Originally it was doing so via `rsPickReg`, but that function
may spill registers that are in use, including the register that is
supposed to be modified by the RMW operator. All other points in
`sched_AM` that need to allocate a register use `rsPickFreeReg` in order
to avoid issues like this. This change updates the use of `rsPickReg` to
instead use `rsPickFreeReg`.

7 years agoMerge pull request #13250 from JosephTremoulet/CaptureHead
Joseph Tremoulet [Mon, 7 Aug 2017 23:07:21 +0000 (19:07 -0400)]
Merge pull request #13250 from JosephTremoulet/CaptureHead

Stop rejecting loops with backward exits

7 years agoSwaping the order of how scenarios are run. (#13255)
José Rivero [Mon, 7 Aug 2017 21:57:10 +0000 (14:57 -0700)]
Swaping the order of how scenarios are run. (#13255)

- Currently ILLink scenario is broken and the x64 scenario leg does not run JitBench
  By swaping the scenario it will let JitBench run while ILLink is being fixed.

7 years agoMerge pull request #13236 from briansull/arm-issue
Brian Sullivan [Mon, 7 Aug 2017 21:09:12 +0000 (14:09 -0700)]
Merge pull request #13236 from briansull/arm-issue

Updated fix for issue #13093

7 years agoFix Region DisplayName (#13249)
Tarek Mahmoud Sayed [Mon, 7 Aug 2017 19:48:39 +0000 (12:48 -0700)]
Fix Region DisplayName (#13249)

This change fix the logic when getting the display name from the OS and when we use the native name instead

7 years agoCleanup code access security from the unmanaged runtime (#13241)
Jan Kotas [Mon, 7 Aug 2017 19:16:45 +0000 (12:16 -0700)]
Cleanup code access security from the unmanaged runtime (#13241)

7 years agoStop rejecting loops with backward exits
Joseph Tremoulet [Mon, 7 Aug 2017 18:11:39 +0000 (11:11 -0700)]
Stop rejecting loops with backward exits

Remove some convoluted code from `optFindNaturalLoops` that has the effect
of rejecting loops with exits that are (lexically) backwards branches.
The comments in the removed code indicate that the intent was to identify
loops which include blocks lexically before `top`, but since `top` is
defined as `head->bbNext`, any such loop was doomed to be rejected by the
code (also being removed) that rejects a loop that "captures head".

When a loop does have a branch back to a block prior to `top`, the
subsequent code that identifies loop exits will identify it as such, and
the subsequent code that rejects predecessors other than `head` of loop
blocks ensures that the backwards branch truly was an exit.

This change still leaves `top` and `first` as separate concepts in the
loop table, to allow for subsequently improvements to loop detection to
identify such loops, though today they're always the same block.

7 years agoRemoved the legacy JIT32 assert regarding 4-byte alignment inArenaAllocator::allocate...
Brian Sullivan [Fri, 4 Aug 2017 21:39:05 +0000 (14:39 -0700)]
Removed the legacy JIT32 assert regarding 4-byte alignment inArenaAllocator::allocateMemory
Immediately after this assert we roundUp to an pointer size allocation amount.

Stopped using two implementation of ArenaAllocator::allocateMemory
Instead we add ifdef DEBUG parts to the common version
Move the implementation of ArenaAllocator::allocateMemory to the header file

Fixed a couple of allocations in assertionprop to use CMF_AssertionProp so that we correctly
attribute which phase uses the memory being allocated.
Added range check for the array writes in Compiler::optMapComplementary

Removed the AlignUp to 4-byte in operator new and new[] in compiler.hpp

7 years agofix callsignalhandlerwrapper(arm) (#13114)
chunseoklee [Mon, 7 Aug 2017 17:51:02 +0000 (02:51 +0900)]
fix callsignalhandlerwrapper(arm) (#13114)

issue: #13099
1. add unwind information
2. add free_stack

7 years agoUpdate CoreClr, CoreFx to preview2-25607-01, preview2-25607-02, respectively (#13133)
dotnet-maestro-bot [Mon, 7 Aug 2017 17:48:01 +0000 (10:48 -0700)]
Update CoreClr, CoreFx to preview2-25607-01, preview2-25607-02, respectively (#13133)

7 years agoUse attribute names without namespaces in a few places (#13248)
dotnet bot [Mon, 7 Aug 2017 17:47:05 +0000 (10:47 -0700)]
Use attribute names without namespaces in a few places (#13248)

Signed-off-by: dotnet-bot <dotnet-bot@microsoft.com>
7 years agoReduce size of array enumerators for x86 (#13157)
James Ko [Mon, 7 Aug 2017 16:55:51 +0000 (12:55 -0400)]
Reduce size of array enumerators for x86 (#13157)

7 years agoUse placement new to initialize ELF Header (instead of memcpy) (#13218)
Jonghyun Park [Mon, 7 Aug 2017 16:27:16 +0000 (01:27 +0900)]
Use placement new to initialize ELF Header (instead of memcpy) (#13218)

7 years agoEnable string format tests on Linux. (#13191)
Jim Ma [Mon, 7 Aug 2017 16:25:25 +0000 (04:25 +1200)]
Enable string format tests on Linux. (#13191)

Modified string format tests so that it can be supported on Linux.

Fix #11655

7 years agoUpdate buildtools to 2.0.0-prerelease-01903-01. (#13228)
Davis Goodin [Mon, 7 Aug 2017 15:00:42 +0000 (10:00 -0500)]
Update buildtools to 2.0.0-prerelease-01903-01. (#13228)

7 years agoReview fixes
Gleb Balykov [Mon, 7 Aug 2017 10:33:58 +0000 (13:33 +0300)]
Review fixes

7 years agoMove files to shared partition (#13244)
Jan Kotas [Mon, 7 Aug 2017 12:56:05 +0000 (05:56 -0700)]
Move files to shared partition (#13244)