Vance Morrison [Thu, 27 Apr 2017 19:43:38 +0000 (12:43 -0700)]
Turn on unwind information at informational level rather than Info level. (#11257)
This is bascically a bug fix. Quite a while ago we wanted to be able to collect stacks (which needed JIT names, and unwind information)
without collecting all verbose CLR events (in particular the GC allocation Tick events). We did this in most places but missed this one.
Bruce Forstall [Thu, 27 Apr 2017 19:41:43 +0000 (12:41 -0700)]
Remove unused TreeNodeInfo::isHelperCallWithKills field (#11259)
Remove unused TreeNodeInfo::isHelperCallWithKills field
Bruce Forstall [Thu, 27 Apr 2017 18:21:27 +0000 (11:21 -0700)]
Merge pull request #11224 from parjong/fix/arm_unaligned_ldobj
[ARM/Linux] Honor unaligned prefix for ldobj instruction
Gaurav Khanna [Thu, 27 Apr 2017 17:46:47 +0000 (10:46 -0700)]
[WIP] Enable binding to CoreLib as NI image (#11040)
Enable binding to CoreLib as NI image
Bruce Forstall [Thu, 27 Apr 2017 17:45:36 +0000 (10:45 -0700)]
Merge pull request #11255 from mskvortsov/FixPutArgReg
[RyuJIT/ARM32] Update putarg_reg codegen
Bruce Forstall [Thu, 27 Apr 2017 17:19:10 +0000 (10:19 -0700)]
Merge pull request #11258 from mskvortsov/FixConsumeReporting
Do not report out-of-order consuming if a node wasn't consumed at all
Sean Gillespie [Thu, 27 Apr 2017 17:09:32 +0000 (10:09 -0700)]
Fix a double-free encountered when failing to initialize a handle table bucket (#11244)
* Fix a double-free encountered when failing to initialize a handle table bucket
* Don't free m_bucket in the destructor of HandleTableBucketHolder
Hanjoung Lee [Thu, 27 Apr 2017 16:39:47 +0000 (01:39 +0900)]
Fix unittest `LPSTRTest` (#11096)
* Fix unittest `LPSTRTest`
The test case Interop.StringMarshalling.LPSTR.LPSTRTest
always returned exit code 100(pass) even if it actually failed.
This implements `ExitTest()` in the same manner with `LPTSTRTest`.
* Fix unittest `LPSTRTest` native
Brian Sullivan [Thu, 27 Apr 2017 15:53:42 +0000 (08:53 -0700)]
Merge pull request #11219 from sdmaclea/PR-ARM64-CpObj-Use-LDP-STP
[Arm64] Use ldp/stp in genCodeForCpObj()
Brian Sullivan [Thu, 27 Apr 2017 15:51:39 +0000 (08:51 -0700)]
Merge pull request #11243 from sdmaclea/PR-ARM64-LDP-ALLOW-GCREFS
[Arm64] Update emitOutputInstr to support GC refs in ldp forms
Mikhail Skvortcov [Thu, 27 Apr 2017 14:53:53 +0000 (17:53 +0300)]
Do not report out-of-order consuming if node wasn't consumed at all
Jonghyun Park [Thu, 27 Apr 2017 14:50:39 +0000 (23:50 +0900)]
[x86/Linux] Disable USE_CALLER_SP_IN_FUNCLET in EH (#11247)
Mikhail Skvortcov [Thu, 27 Apr 2017 10:54:35 +0000 (13:54 +0300)]
RyuJIT/ARM32: Update GT_PUTARG_REG codegen.
Steve MacLean [Thu, 27 Apr 2017 04:50:37 +0000 (00:50 -0400)]
[Arm64/Unix] Revise JIT_MemCpy (#11143)
* [Arm64/Unix] Revise JIT_MemCpy
Use ldp/stp
Correctly handle short copy lengths
Simplify code & pseudo code
Use uint*_t to make pseudo code more readable
Felipe Pessoto [Thu, 27 Apr 2017 04:50:04 +0000 (01:50 -0300)]
Fix building with only VS2017 installed. (#11182)
The fix was originally made in https://github.com/dotnet/corefx/pull/18700
William Godbe [Thu, 27 Apr 2017 01:55:16 +0000 (18:55 -0700)]
Merge pull request #11240 from wtgodbe/SourceBranch
Use SourceBranch from pipebuild to define Helix Branch
Koundinya Veluri [Thu, 27 Apr 2017 01:42:48 +0000 (18:42 -0700)]
Fix to respect readonly prefix for multidimensional array Address method (#11245)
Fixes #9414
- Added missing 'else' back from regression in https://github.com/dotnet/coreclr/commit/
97b4ff0b438261ba11b357008630076054a6f25d#diff-edc46b80f57431489a82311948a8234dL6424, which has the effect of passing null to the Address method's instParam to indicate the readonly prefix in order to bypass the type check
Gaurav Khanna [Wed, 26 Apr 2017 23:53:58 +0000 (16:53 -0700)]
Fix filename display in exception messages (#11238)
dotnet bot [Wed, 26 Apr 2017 23:21:04 +0000 (16:21 -0700)]
Update CoreClr, CoreFx to preview2-25226-02, preview2-25226-02, respectively (#11127)
Jan Kotas [Wed, 26 Apr 2017 21:01:22 +0000 (14:01 -0700)]
Delete moved file
Faizur Rahman [Wed, 26 Apr 2017 03:39:46 +0000 (20:39 -0700)]
Move DllNotFoundException, EntryPointNotFoundException to CoreLib (#3415)
Signed-off-by: dotnet-bot <dotnet-bot@microsoft.com>
Koundinya Veluri [Wed, 26 Apr 2017 23:08:53 +0000 (16:08 -0700)]
Fix timer callback time drift (#11220)
Fixes #6408:
- The change to S.T.Timer is the actual fix, but fixed the timer firing on the thread pool timer thread similarly as well
Steve MacLean [Wed, 26 Apr 2017 22:53:08 +0000 (18:53 -0400)]
Fix comment
[Arm64] Update emitOutputInstr to support GC refs in ldp forms
Pat Gavlin [Wed, 26 Apr 2017 22:37:39 +0000 (15:37 -0700)]
Merge pull request #11235 from pgavlin/VSO424019
Require a frame pointer in methods with throw blocks for sources that…
wtgodbe [Wed, 26 Apr 2017 21:54:04 +0000 (14:54 -0700)]
Use SourceBranch from pipebuild to define Helix Branch
Sean Gillespie [Wed, 26 Apr 2017 21:50:19 +0000 (14:50 -0700)]
Fix and re-enable some GC finalizer tests (#11216)
* Fix and re-enable some GC finalizer tests
* Make the Usage test a little more robust
[Arm64] CpObj ldp/stp review feedback
Rework asserts
Add comments
Format
Sven Boemer [Wed, 26 Apr 2017 21:02:36 +0000 (14:02 -0700)]
Merge pull request #11178 from sbomer/ilrt_factoring
Refactor ildasm round trip logic
Pat Gavlin [Wed, 26 Apr 2017 20:23:59 +0000 (13:23 -0700)]
Require a frame pointer in methods with throw blocks for sources that have mismatched stack depths.
This requirement effectively disables double-aligned frames (which are
effectively a special case of ESP frames). Without this change it is
possible for the JIT to generate a throw block with sources that have
mismatched offsets, which can lead to GC holes.
Fixes VSO 424019.
Andy Ayers [Wed, 26 Apr 2017 20:04:14 +0000 (13:04 -0700)]
Merge pull request #11218 from AndyAyersMS/FixReturnTypeMismatch
JIT: fix return type mismatch inlining block
Michael Letterle [Fri, 31 Mar 2017 15:31:53 +0000 (11:31 -0400)]
Lookup System.Runtime and mscorlib references by Assembly name
Rather than by alias.
Fixes #10595
William Godbe [Wed, 26 Apr 2017 18:29:52 +0000 (11:29 -0700)]
Merge pull request #11229 from wtgodbe/CrossgenFix
Exit on failure of HelixPrep.proj
wtgodbe [Wed, 26 Apr 2017 18:29:10 +0000 (11:29 -0700)]
Exit on failure of HelixPrep.proj
Sven Boemer [Mon, 24 Apr 2017 17:12:38 +0000 (10:12 -0700)]
Refactor ildasm round trip logic
This change factors the Ildasm round trip logic so that it should not
require a separate test build. The ildasmroundtrip flag is no longer
passed to build-test.cmd; instead the wrapper scripts are always
generated with a check for an environment variable that can be set from
runtests.
Also remove separate test build for ilrt in the ci scripts
Jan Vorlicek [Wed, 26 Apr 2017 14:10:13 +0000 (07:10 -0700)]
Enable build with clang 4.0 (#11226)
This change enables build with clang 4.0 and fixes a bunch of new errors that
the stricter compiler was reporting.
Jan Vorlicek [Wed, 26 Apr 2017 13:48:20 +0000 (06:48 -0700)]
Fix thread wait to be resilient to real time clock changes on Unix (#11213)
This change fixes a problem that happens when a thread uses a timed wait
and the real time clock change in the system (e.g. due to the user action,
time server sync, leap second, etc). In such case, the wait may end up waiting
for shorter or longer period of time.
It fixes the issue for Thread.Sleep and timed wait on a process local synchronization
primitives by using CLOCK_MONOTONIC instead of CLOCK_REALTIME for waiting on
cond vars.
Unfortunately, the waits on cross process synchronication primitives like
named mutexes always use CLOCK_REALTIME, so this fix works on process local primitives
only.
Jonghyun Park [Wed, 26 Apr 2017 12:49:07 +0000 (21:49 +0900)]
[x86/Linux] Stack alignment check in Hijack Helpers (#11221)
Jonghyun Park [Wed, 26 Apr 2017 12:18:33 +0000 (21:18 +0900)]
[x86/Linux] Use appropriate FCALL macro for DateMarshaler__ConvertToNative (#11194)
Michal Strehovský [Tue, 25 Apr 2017 20:13:36 +0000 (13:13 -0700)]
Merge pull request dotnet/corert#3429 from dotnet/nmirror
Merge nmirror to master
Signed-off-by: dotnet-bot <dotnet-bot@microsoft.com>
Alan Lawrence [Wed, 26 Apr 2017 11:14:11 +0000 (12:14 +0100)]
Comment explaining use of FlushProcessWriteBuffers in StompWriteBarrier. (#10924)
Jan Vorlicek [Wed, 26 Apr 2017 08:52:35 +0000 (01:52 -0700)]
Fix few clang analyzer warnings (#11185)
This change fixes two issues reported by the clang analyzer. In both cases, uninitialized
variable value was used in an error code path.
Jonghyun Park [Wed, 26 Apr 2017 06:24:44 +0000 (15:24 +0900)]
[ARM/Linux] Honor unaligned prefix for ldobj instruction
Bruce Forstall [Wed, 26 Apr 2017 04:17:48 +0000 (21:17 -0700)]
Merge pull request #11107 from BruceForstall/NumSuccImprovement
Num succ improvement
Bruce Forstall [Wed, 26 Apr 2017 04:16:29 +0000 (21:16 -0700)]
Merge pull request #11204 from BruceForstall/FixArmAltjitGcinfo
Fix GCInfo library used by ARM/ARM64 protononjit builds
William Godbe [Wed, 26 Apr 2017 04:16:09 +0000 (21:16 -0700)]
Merge pull request #11215 from wtgodbe/ActuallyDoCrossgen
Set RunCrossGen=true for non-Windows R2R runs
[Arm64] Use ldp/stp in genCodeForCpObj()
Jonghyun Park [Tue, 25 Apr 2017 23:31:29 +0000 (08:31 +0900)]
Move DynamicHelperFrame::UpdateRegDisplay into cgenx86.cpp (#11193)
* Move DynamicHelperFrame::UpdateRegDisplay into cgenx86.cpp
* Fix x86/Windows link error
Andy Ayers [Tue, 25 Apr 2017 23:06:44 +0000 (16:06 -0700)]
JIT: fix return type mismatch inlining block
If caller passes a directly substitutable expression whose
type is compatible with but not the same as the callee argument type,
and that argument is the callee's return value, the jit may
reject inlining the method because of a return type mismatch.
Fix is to retype the expressions we use in the callee's body
as the callee's type.
Closes #11211.
wtgodbe [Tue, 25 Apr 2017 22:40:22 +0000 (15:40 -0700)]
Set RunCrossGen=true for non-Windows R2R runs
Pat Gavlin [Tue, 25 Apr 2017 22:36:37 +0000 (15:36 -0700)]
Merge pull request #11134 from pgavlin/VSO406163
Fix fgNewBBinRegion when inserting into filters.
Roman Artemev [Tue, 25 Apr 2017 22:24:09 +0000 (15:24 -0700)]
Merge pull request #11110 from rartemev/issue_9880
Made behavior to conform to ECMA for 'tail.call, pop, ret' case
mskvortsov [Tue, 25 Apr 2017 22:01:43 +0000 (01:01 +0300)]
[RyuJIT/ARM32] Move genJumpKindsForTree into common part (#11103)
* Move genJumpKindsForTree() into codegenarmarch.cpp.
* Edit genJumpKindsForTree().
* Introduce genCodeForJumpTrue() for both ArmArch and XArch.
* Fix GT_JTRUE codegen for ARM32 by using genCodeForJumpTrue().
Bruce Forstall [Tue, 25 Apr 2017 16:43:38 +0000 (09:43 -0700)]
Fix GCInfo library used by ARM/ARM64 protononjit builds
Otherwise, using the normal GCInfo library, the altjits will
assert creating GC info.
Create new "cross-target" builds of the GCInfo library.
Roman Artemev [Tue, 25 Apr 2017 21:31:38 +0000 (14:31 -0700)]
Merge pull request #11118 from rartemev/issue_10094
Fixed netci.groove scripts to make CentOS and Ubuntu R2R jitstressregs builds worked.
Roman Artemev [Thu, 20 Apr 2017 00:49:25 +0000 (17:49 -0700)]
Made behavior to conform to ECMA for 'tail.call, pop, ret' case
William Godbe [Tue, 25 Apr 2017 19:56:08 +0000 (12:56 -0700)]
Merge pull request #11206 from wtgodbe/ContainerName
Update R2R Container Names
wtgodbe [Tue, 25 Apr 2017 19:55:37 +0000 (12:55 -0700)]
Update R2R Container Names
Roman Artemev [Tue, 25 Apr 2017 00:01:06 +0000 (17:01 -0700)]
Testing. DO NOT MERGE ME
Jonghyun Park [Tue, 25 Apr 2017 18:40:40 +0000 (03:40 +0900)]
Allow FEATURE_INTERPRETER to be enabled by command-line option (#11188)
* Allow FEATURE_INTERPRETER to be enabled by command-line option
* Set FEATURE_INTERPRETER as 0
William Godbe [Tue, 25 Apr 2017 18:29:37 +0000 (11:29 -0700)]
Merge pull request #11187 from wtgodbe/NonWindowsR2R
Enable running R2R tests in Helix on non-Windows
wtgodbe [Mon, 24 Apr 2017 20:41:42 +0000 (13:41 -0700)]
Enable running R2R tests in Helix on non-Windows
Bruce Forstall [Tue, 25 Apr 2017 18:14:27 +0000 (11:14 -0700)]
Merge pull request #11203 from mskvortsov/FixLclFld
[RyuJIT/ARM32] Update lclFld codegen
Hyung-Kyu Choi [Tue, 25 Apr 2017 17:58:22 +0000 (02:58 +0900)]
[ARM32] clang 3.9 as a default for ARM cross build (#11064)
* [ARM32] clang 3.9 as a default for ARM cross build
Use clang 3.9 as a default compiler for ARM cross build
* Both arm and armel will be built using clang 3.9 as a default
* ARM CI will use clang 3.9 as a default
* ARM pipieline build will use clang 3.9 as a default
Signed-off-by: Hyung-Kyu Choi <hk0110.choi@samsung.com>
* [ARM32] Update docs describing clang-3.9 and ARM cross build
Signed-off-by: Hyung-Kyu Choi <hk0110.choi@samsung.com>
* [ARM] Check minimum version requirement of clang for ARM cross build
Signed-off-by: Hyung-Kyu Choi <hk0110.choi@samsung.com>
Pat Gavlin [Fri, 21 Apr 2017 17:26:05 +0000 (10:26 -0700)]
Fix fgNewBBinRegion when inserting into filters.
In order to ensure that the GC operates properly when handling
exceptions, the CLR ABI requires that control exits a filter region
throguh its terminal instruction. `fgNewBBinRegion` was violating this
invariant, however, which lead to GC stress failures when the GC was
invoked immediately after a filter finished executing but before control
entered any handlers. This change fixes the behavior of
`fgNewBBinRegion` when inserting new blocks into filter regions.
There are two cases to consider when inserting a new BB into a filter
region:
1. The filter region consists of multiple blocks
2. The filter region consists of a single block
The first case is straightforward: instead of inserting a new BB after
the last block of the filter, insert it before the last block of the
filter. Because the filter contains multiple blocks, the predecessor of
the filter must also be in the filter region.
The latter case requires splitting the single block, as inserting a new
block before the single block would otherwise change control flow (the
filter would begin with the new block rather than the single block). In
order to acheive this with minimal complexity, this change first inserts
a `BBJ_ALWAYS` block that transfers control to the existing single
block, then inserts a second new block immediately before the existing
single block. The second new block is then returned. To put it
visually, the transformation is from this:
```
filter start: BBN (BBJ_EHFILTERRET)
```
to this:
```
filter start: BBN + 1 (BBJ_ALWAYS -> BBN)
BBN + 2 (jumpKind)
BBN (BBJ_EHFILTERRET)
```
and the function returns `BBN + 2`.
Fixes VSO 406163.
dotnet bot [Tue, 25 Apr 2017 17:05:21 +0000 (10:05 -0700)]
Merge pull request dotnet/corert#3418 from dotnet/nmirror (#11186)
Merge nmirror to master
Signed-off-by: dotnet-bot <dotnet-bot@microsoft.com>
Bruce Forstall [Tue, 25 Apr 2017 16:58:16 +0000 (09:58 -0700)]
Merge pull request #11191 from BruceForstall/FixFuncletAssert
Fix funclet begin assert
Bruce Forstall [Tue, 25 Apr 2017 15:40:26 +0000 (08:40 -0700)]
Merge pull request #11080 from BruceForstall/MeasureNoway
Add code to measure dynamic noway_assert usage
Mikhail Skvortcov [Tue, 25 Apr 2017 12:04:03 +0000 (15:04 +0300)]
RyuJIT/ARM32: Update GT_LCL_FLD codegen.
Hanjoung Lee [Tue, 25 Apr 2017 09:07:38 +0000 (18:07 +0900)]
[x86/Linux] Do not report untracked from filters (#11195)
Related Issue : #11101
Bruce Forstall [Wed, 19 Apr 2017 20:37:34 +0000 (13:37 -0700)]
Add code to measure dynamic noway_assert usage
Set COMPlus_JitMeasureNowayAssert=1 to enable measurement.
Set COMPlus_JitMeasureNowayAssertFile to a path/file to send
output to a file instead of stdout.
Code is under `#if MEASURE_NOWAY` which is enabled for DEBUG
only, though it can be manually enabled for non-DEBUG builds.
Bruce Forstall [Mon, 24 Apr 2017 18:12:14 +0000 (11:12 -0700)]
Fix funclet begin assert
Bruce Forstall [Tue, 25 Apr 2017 06:49:01 +0000 (23:49 -0700)]
Merge pull request #11196 from wateret/fix-formatting
Fix formatting fail
Bruce Forstall [Tue, 25 Apr 2017 06:31:17 +0000 (23:31 -0700)]
Merge pull request #11183 from sdmaclea/PR-FIX-11181
Fix 11181
Hyeongseok Oh [Tue, 25 Apr 2017 06:08:38 +0000 (15:08 +0900)]
[x86/Linux] Change CI docker image (#11098)
* [x86/Linux] Change CI docker image
Change CI docker image to support clang-3.9
* Update x86_ci_script.sh
change docker tag name
* Update x86_ci_script.sh
environment set correctly: ROOTFS_DIR
* Update x86_ci_script.sh
change docker image name
Hanjoung Lee [Tue, 25 Apr 2017 05:51:14 +0000 (14:51 +0900)]
Fix formatting fail
Jim Ma [Tue, 25 Apr 2017 05:40:36 +0000 (17:40 +1200)]
Ensure using the given executable path. (#10525)
* Ensure using the given executable path.
Fix the bug that corerun trying to load the executable from current directory even if the user has specified a full path of the executable.
Fix #5631
* Revert "Ensure using the given executable path."
This reverts commit
d237e6329f85132429176a0644cf6d93c9437ff4.
* Partially implemented.
* Ensure app context using correct IL file path.
Extracting simple name from the given file path and look up its value in tpa list. If the value does not equal to the given file path, we update the value with correct given file path.
Fix #5631
* Re-arrange the slashIndex to eliminate unneccessary code.
* A few performance improvement with a memory leak bug fixed.
* Added the absolute path of target assembly to tpa list.
Fix #5631
* Added the path of target assembly to tpa on Windows.
Fix #5631
Mike McLaughlin [Tue, 25 Apr 2017 00:12:36 +0000 (17:12 -0700)]
Add "createdump" command to SOS. (#11161)
Uses the createdump utility code as a library.
Changed the "-m, --micro" option to "-t, --triage".
Lots of cleanup.
Andy Ayers [Mon, 24 Apr 2017 23:42:38 +0000 (16:42 -0700)]
Merge pull request #11152 from AndyAyersMS/FixTailCallStressAssert
JIT: simple fix for the tail call stress assert
Bruce Forstall [Mon, 24 Apr 2017 23:28:08 +0000 (16:28 -0700)]
Merge pull request #11157 from mikedn/reg-encoding
Simplify emitter's RegEncoding
Bruce Forstall [Mon, 24 Apr 2017 23:02:18 +0000 (16:02 -0700)]
Merge pull request #11108 from BruceForstall/RangedForAllSuccessorIter
Use C++ ranged 'for' for AllSuccessorIter use
Bruce Forstall [Mon, 24 Apr 2017 21:28:51 +0000 (14:28 -0700)]
Merge pull request #11172 from mskvortsov/ArithOverflowChecks
[RyuJIT/ARM32] Enable arithmetic overflow checks
Russ Keldorph [Sun, 23 Apr 2017 13:42:14 +0000 (06:42 -0700)]
Support other toolsets in Win/ARM64 build
Fix 11181
Bruce Forstall [Mon, 24 Apr 2017 18:15:35 +0000 (11:15 -0700)]
Fix cmake warning about mismatched if/endif comment (#11176)
Steve MacLean [Mon, 24 Apr 2017 17:51:29 +0000 (13:51 -0400)]
Remove PAL FileAlignment restriction (#10959)
* Remove PAL FileAlignment restriction
* Address PR #10959 feedback
* Fix amd64 crossgen error
* Respond to review feedback
* Fix arm32 regression
* Prepare to remove VIRTUAL_PAGE_* from map.cpp
Also simplify previous section code
* Rename function to GetVirtualPageSize()
Sean Gillespie [Mon, 24 Apr 2017 17:44:25 +0000 (10:44 -0700)]
[Local GC] Ensure that handle creation returns null on failure instead of throwing (#11092)
* [Local GC] Ensure that handle creation returns null on failure instead of throwing
* Fix some clang pedantry about jumping past variable initialization
* Throw OOM if initialization of handle store fails
* Perform the handle OOM check in each handle helper
Sergey Andreenko [Mon, 24 Apr 2017 16:39:25 +0000 (09:39 -0700)]
do not use = {nullptr} initialization. (#11153)
It worked incorrectly on desktop.
Bruce Forstall [Thu, 20 Apr 2017 16:49:24 +0000 (09:49 -0700)]
Split NumSucc/GetSucc into two versions
One set takes Compiler*, one doesn't.
It's simpler, and hopefully faster.
Bruce Forstall [Mon, 24 Apr 2017 15:43:37 +0000 (08:43 -0700)]
Merge pull request #11169 from hseok-oh/ryujit/fix_11128
[ARM32/RyuJIT] Fix failure related with struct argument passing
Bruce Forstall [Mon, 24 Apr 2017 15:29:15 +0000 (08:29 -0700)]
Merge pull request #11091 from BruceForstall/BlockRefactor
Refactoring: move some BasicBlock functions to block.cpp
Mikhail Skvortcov [Mon, 24 Apr 2017 11:10:20 +0000 (14:10 +0300)]
RyuJIT/ARM32: Enable arithmetic overflow checks.
Jonghyun Park [Mon, 24 Apr 2017 10:20:38 +0000 (19:20 +0900)]
Support FEATURE_DBGIPC override (#11164)
Jonghyun Park [Mon, 24 Apr 2017 08:39:52 +0000 (17:39 +0900)]
[x86/Linux] Remove unnecessary ResumeEsp field (#11167)
Hyeongseok Oh [Mon, 24 Apr 2017 08:34:47 +0000 (17:34 +0900)]
[ARM32/RyuJIT] Fix failure related with struct argument passing
Block using NYI_ARM when struct value split between registers and stack
Fix type assign error for struct slot
Jonghyun Park [Mon, 24 Apr 2017 08:24:46 +0000 (17:24 +0900)]
[x86/Linux] Fix unmatched NESTED_END in CallEHFilterFunclet (#11163)
OmariO [Sun, 23 Apr 2017 20:35:24 +0000 (23:35 +0300)]
11106: Use RuntimeHelpers.IsReferenceOrContainsReferences in Dictionary.Remove methods (#11158)
Mike Danes [Sat, 22 Apr 2017 13:41:56 +0000 (16:41 +0300)]
Simplify emitter's RegEncoding
VC++ does't recognize that `(c ? x - 16 : x) & 7` is just `x & 7` and we end up with 5 instructions instead of 1.
Sean Gillespie [Sat, 22 Apr 2017 11:41:32 +0000 (04:41 -0700)]
Mark GetGenerationWR2 as GCStressIncompatible across all platforms (#11151)
Sean Gillespie [Sat, 22 Apr 2017 11:39:09 +0000 (04:39 -0700)]
Remove destructor from GCEvent and instead rely on the OS to clean up (#11132)
* Remove destructor from GCEvent and instead rely on the OS to clean up
* Add a comment justifying the lack of destructor
* wording: many -> all