Rama krishnan Raghupathy [Tue, 15 Mar 2016 01:53:57 +0000 (18:53 -0700)]
Merge pull request #3634 from ramarag/testbuild
Enabling standalone builds for Test Project of any Priority
William Godbe [Tue, 15 Mar 2016 00:31:44 +0000 (17:31 -0700)]
Merge pull request #3702 from wtgodbe/disableEH
Disable failing Hardware EH test on non-windows
Rama krishnan Raghupathy [Mon, 14 Mar 2016 23:49:25 +0000 (16:49 -0700)]
Merge pull request #3654 from ramarag/lttngperf
Using StackString to improve memory consumption in Lttng EventProvider
wtgodbe [Mon, 14 Mar 2016 22:28:03 +0000 (15:28 -0700)]
Disable failing Hardware EH test on non-windows
Tarek Mahmoud Sayed [Mon, 14 Mar 2016 22:02:01 +0000 (15:02 -0700)]
Merge pull request #3695 from tarekgh/FixUmAlQuraCalendarStaticDataInitialization
Fix UmAlQuraCalendar static data initilaization
Kyungwoo Lee [Mon, 14 Mar 2016 21:05:40 +0000 (14:05 -0700)]
Merge pull request #3679 from kyulee1/fixgchole
ARM64: Fix GC hole in indirect call site
Brian Sullivan [Mon, 14 Mar 2016 20:00:30 +0000 (13:00 -0700)]
Merge pull request #3696 from briansull/update-tests-lst
Update tests lst for ARM64
Brian Sullivan [Mon, 14 Mar 2016 19:57:39 +0000 (12:57 -0700)]
Merge branch 'master' into update-tests-lst
Brian Sullivan [Mon, 14 Mar 2016 19:56:18 +0000 (12:56 -0700)]
Updated with NEW_PASS tests
Kyungwoo Lee [Mon, 14 Mar 2016 18:06:11 +0000 (11:06 -0700)]
ARM64: Fix GC hole in indirect call site
This fixes dotnet#3663.
Indirect call (```br``` or ```blr```) target is encoded with a register
which the first operand internally represents.
Unfortunately, call sites use the first two operands to hold GC
callee-save registers.
So, this GC register information was overridden by the call target operand
in the indirect(virtual) call sites.
The fix is to use 3rd/4th operands instead of 1st/2nd operands to hold GC info.
Ideally we should use different field name and also ensure constness when
we set up the operand so that it's never written more than once.
https://github.com/dotnet/coreclr/issues/3693 is filed.
Tarek Mahmoud Sayed [Mon, 14 Mar 2016 19:28:17 +0000 (12:28 -0700)]
Fix UmAlQuraCalendar static data initilaization
The current way of declaring the static data is cauing null reference exception
when the static constructor get called. the reason is the static initialization
method InitDateMapping is using the static array rawData which is not really
get initilaized yet when InitDateMapping execute.
The fix is moving rawData inside InitDateMapping.
https://github.com/dotnet/corefx/issues/6774#issuecomment-
196044580
Kyungwoo Lee [Mon, 14 Mar 2016 18:21:27 +0000 (11:21 -0700)]
Merge pull request #3687 from kyulee1/fixzeroinit
ARM64: Fix ZeroInit of Locals
Russ Keldorph [Mon, 14 Mar 2016 16:41:36 +0000 (09:41 -0700)]
Merge pull request #3659 from RussKeldorph/stresstags
Add test mode exclusions
Kyungwoo Lee [Sun, 13 Mar 2016 20:11:18 +0000 (13:11 -0700)]
ARM64: Fix ZeroInit of Locals
When zero-initializing locals, JIT emits wrong instruction sequence --
e.g, 28 byte zero-intialization as shown below.
The issue was JIT passed wrong arguments to emitIns_R_R_I.
Before (Fail)
```
stp xzr, xzr, [x2],#16
str xzr, [x2,#2] --> just two byte offset (no x2 post-increment)
str wzr, [x2]
```
After (Pass)
```
stp xzr, xzr, [x2],#16
str xzr, [x2],#8
str wzr, [x2]
```
Jan Kotas [Sun, 13 Mar 2016 21:55:50 +0000 (14:55 -0700)]
Merge pull request #3685 from jkotas/ifdefs
Fix ifdefs
Tarek Mahmoud Sayed [Sun, 13 Mar 2016 20:47:07 +0000 (13:47 -0700)]
Merge pull request #3662 from tarekgh/FixGetYearForJapaneseCalendar
Fix calculating the calendar year in JapaneseLunisolarCalendar
Kyungwoo Lee [Sun, 13 Mar 2016 20:10:06 +0000 (13:10 -0700)]
Merge pull request #3673 from briansull/update-tests-lst
ARM64: Major Update for new tests to run using Smarty.exe
Jan Kotas [Sun, 13 Mar 2016 19:09:22 +0000 (12:09 -0700)]
Remove remaining strongname validation on Windows
CodeCLR always skips strongname validation. Some code for strong name validation was still compiled for Windows, but not for Unix. This change removes the differences between Windows and Unix by never compiling the strong name validation code in.
Jan Kotas [Sun, 13 Mar 2016 13:03:17 +0000 (06:03 -0700)]
Remove incorrect use of FEATURE_LEGACYSURFACE ifdef
Sergiy Kuryata [Sun, 13 Mar 2016 06:17:12 +0000 (22:17 -0800)]
Merge pull request #3683 from sergiy-k/runtestsfix1
Exclude netcore50aot assemblies in runtests.sh
Russ Keldorph [Fri, 11 Mar 2016 00:34:47 +0000 (16:34 -0800)]
Add test mode exclusions
Add the ability to exclude individual tests that are incompatible with
certain test modes. If a test is incompatible with or too slow under
GCStress, add
<GCStressIncompatible>true</GCStressIncompatible>
as a property to the test's project file. If a test requires optimization
or is sensitive to precise optimization patterns, use
<JitOptimizationSensitive>true</JitOptimizationSensitive>
Sergiy Kuryata [Sun, 13 Mar 2016 03:47:48 +0000 (19:47 -0800)]
Exclude netcore50aot assemblies in runtests.sh
Jan Kotas [Sat, 12 Mar 2016 07:26:24 +0000 (23:26 -0800)]
Merge pull request #3671 from brianrob/multicorejitthread
Use Default Stack Size for Multi-Core JIT Playback Thread
Sergiy Kuryata [Sat, 12 Mar 2016 04:18:52 +0000 (20:18 -0800)]
Merge pull request #3678 from sergiy-k/runtestsfix
Patch the runtests.sh script to work around an issue with corefx assemblies
Jan Kotas [Sat, 12 Mar 2016 03:49:27 +0000 (19:49 -0800)]
Merge pull request #3674 from jkotas/tail-call
Fallback to JIT for explicit tailcalls in R2R
Sergiy Kuryata [Sat, 12 Mar 2016 02:17:15 +0000 (18:17 -0800)]
Patch the runtests.sh script to work around an issue with corefx assemblies
Change https://github.com/dotnet/corefx/pull/6779 in CoreFx added corert
assemblies to the mix of assemblies that are built for CoreCLR on Unix
which broke the logic in runtest.sh that creates an overlay for running
tests. This change creates a work around for the problem.
Jan Kotas [Sat, 12 Mar 2016 00:39:41 +0000 (16:39 -0800)]
Merge pull request #3590 from jkotas/deadcode
Delete dead code
Chris McKinsey [Fri, 11 Mar 2016 23:46:56 +0000 (15:46 -0800)]
Merge pull request #3631 from cmckinsey/DevDiv_199169
Fix bad codegen for initblk with large fill value
Matt Mitchell [Fri, 11 Mar 2016 23:17:45 +0000 (15:17 -0800)]
Merge pull request #3675 from mmitche/fix-names
Fix name on trigger
Matt Mitchell [Fri, 11 Mar 2016 23:16:30 +0000 (15:16 -0800)]
Fix name on trigger
Jan Kotas [Fri, 11 Mar 2016 22:52:58 +0000 (14:52 -0800)]
Fallback to JIT for explicit tailcalls in R2R
Brian Sullivan [Fri, 11 Mar 2016 22:27:25 +0000 (14:27 -0800)]
Major Update for new tests
Updated Tests.lst and preserved Categories tags
We can run all of the Pro, Pri1 tests
We currently run only the ones that were passing
from the previous Tests.lst file
DDCloud [Fri, 11 Mar 2016 03:39:22 +0000 (19:39 -0800)]
Using StackString to improve memory consumption in Lttng EventProvider
Matt Mitchell [Fri, 11 Mar 2016 22:09:13 +0000 (14:09 -0800)]
Merge pull request #3672 from mmitche/update-badges
Update badges to point to new master branch jobs
Matt Mitchell [Fri, 11 Mar 2016 22:02:50 +0000 (14:02 -0800)]
Update badges to point to new master branch jobs
Matt Mitchell [Fri, 11 Mar 2016 21:14:34 +0000 (13:14 -0800)]
Merge pull request #3197 from mmitche/branchify
[Do Not Merge] Move coreclr to branch model
Jan Kotas [Fri, 11 Mar 2016 19:14:49 +0000 (11:14 -0800)]
Merge pull request #3658 from leemgs/upstream-doc-crossbuild
Documentation: Remove the "patching Urcu" section
Tarek Mahmoud Sayed [Fri, 11 Mar 2016 18:44:48 +0000 (10:44 -0800)]
Fix calculating the calendar year in JapaneseLunisolarCalendar
The months in this calendar is lunar months. We can have the case when a new era
starts in the middle of a month which means part of the month will belong to one
era and the rest will belong to the new era. When calculating the calendar year
number for dates which in the rest of the month and exist in the new started era,
we should still use the old era info instead of the new era info because
the rest of the month still belong to the year of last era.
Example of such date is Feb 2nd 1989. Using such date while the era starts
on Jan 8th 1989 make this date belong to that new era which it is part of
the month started with previous era. Using the new era info will have the year
we are calculating equal to the era offset (i.e. year = m_EraInfo[i].yearOffset)
and we’ll end returning zero for the year. The fix is to use the previous era
info to calculate the year
issue https://github.com/dotnet/corefx/issues/6775
Sejong Oh [Fri, 11 Mar 2016 19:04:24 +0000 (11:04 -0800)]
Merge pull request #3645 from sejongoh/increase_gctest_time_out
Increase timeout and crond schedule for GC stress testing
Matt Mitchell [Fri, 11 Mar 2016 18:45:32 +0000 (10:45 -0800)]
Move coreclr to branched CI model
This moves coreclr to the branch-based CI model. In this model, instead of having the CI generate jobs out of dotnet/coreclr's master branch for all branches (using all sorts of odd naming suffixes of course), the master CI lists tell the CI which branches it should look at for CI config info. It passes the branch name to the netci file, which then can use it to tell the SCM to pull code from specific branches or set branch specific PR triggers.
Brian Sullivan [Fri, 11 Mar 2016 18:03:03 +0000 (10:03 -0800)]
Merge pull request #3653 from briansull/migrate-tags
Add migrate-tags.py script for migrating Smarty tags
Sean Gillespie [Fri, 11 Mar 2016 17:57:38 +0000 (09:57 -0800)]
Merge pull request #3252 from mjp41/master
Fixed parent pointers becoming corrupt in GC benchmark for Red/Black trees
Brian Sullivan [Fri, 11 Mar 2016 17:52:27 +0000 (09:52 -0800)]
Added Header and other Codereview feedback
Geunsik Lim [Fri, 11 Mar 2016 10:30:12 +0000 (19:30 +0900)]
Documentation: Remove Patching Urcu Section
Atomic function parameter types Urcu is already fixed.
We don't need any manual patching for the cross build.
Let's remove this section from now on.
Signed-off-by: Geunsik Lim geunsik.lim@samsung.com
Signed-off-by: Prajwal A N an.prajwal@samsung.com
Signed-off-by: MyungJoo Ham myungjoo.ham@samsung.com
Sejong Oh [Thu, 10 Mar 2016 23:57:09 +0000 (15:57 -0800)]
a This is a combination of 2 commits.
Increase timeout and crond schedule for GC stress testing
Sejong Oh [Fri, 11 Mar 2016 07:31:18 +0000 (23:31 -0800)]
Merge pull request #3630 from sejongoh/pull_down_coredistool_windows
Pull down coredistool package on Windows
Brian Robbins [Fri, 11 Mar 2016 06:35:56 +0000 (22:35 -0800)]
Update Multi-Core JIT to use the default stack size for the JIT playback thread.
Brian Sullivan [Fri, 11 Mar 2016 03:31:35 +0000 (19:31 -0800)]
Add migrate-tags.py script for migrating Smarty tags
Also added a test case for the script
Chris McKinsey [Wed, 9 Mar 2016 23:03:48 +0000 (15:03 -0800)]
Fix bad codegen for initblk with large fill value
When the initialization value of an init block is constant we attempt to
generate an inline expansion sequences to execute the stores. In order to
generate wider stores the constant value is expanded so that each byte
position holds the initial byte value. In some cases the constant on the
initblk may be a value not in the range of 0..255 and this constant
expansion will generate the wrong value. This change fixes the x64, x86,
and arm64 (comment only) expanses to use only the lower byte value of this
constant. The ARM32 backend already does this, and the other JIT32/JIT64
JITs shipping in Desktop also do this correctly.
Rama Krishnan Raghupathy [Thu, 10 Mar 2016 23:45:22 +0000 (15:45 -0800)]
Enabling standalone builds for Test Project of any Priority
Sergiy Kuryata [Fri, 11 Mar 2016 02:02:04 +0000 (18:02 -0800)]
Merge pull request #3647 from dagood/fix-pri1-test-dependencies
Fix priority 1 tests: add missing dependencies
Andy Ayers [Fri, 11 Mar 2016 02:01:36 +0000 (18:01 -0800)]
Merge pull request #3632 from AndyAyersMS/InlineRefactorMoveProfitability
Inline refactoring: move profitability assessment to LegacyPolicy
Rahul Kumar [Fri, 11 Mar 2016 01:55:27 +0000 (17:55 -0800)]
Merge pull request #3652 from rahku/EHTestPort
exclude test baseservices\exceptions\regressions\Dev11\147911
Rahul Kumar [Fri, 11 Mar 2016 01:03:11 +0000 (17:03 -0800)]
exclude test baseservices\exceptions\regressions\Dev11\147911
John Chen [Fri, 11 Mar 2016 00:42:28 +0000 (16:42 -0800)]
Merge pull request #3644 from JohnChen0/master
Fix a GC hole in a Ready-to-Run helper
Davis Goodin [Fri, 11 Mar 2016 00:04:53 +0000 (18:04 -0600)]
Add new dependencies to tests where package upgrade moved types.
Eugene Rozenfeld [Thu, 10 Mar 2016 23:19:24 +0000 (15:19 -0800)]
Merge pull request #3641 from erozenfeld/GetCORSystemDirectoryInternaLFix
Fix for a bug in GetCORSystemDirectoryInternaL
Sejong Oh [Wed, 9 Mar 2016 02:11:57 +0000 (18:11 -0800)]
Pull down coredistool package on Windows
Rahul Kumar [Thu, 10 Mar 2016 23:05:34 +0000 (15:05 -0800)]
Merge pull request #3640 from rahku/EHTestPort
Fix arm64 build break
John Chen (CLR) [Thu, 10 Mar 2016 23:04:44 +0000 (15:04 -0800)]
Fix a GC hole in a Ready-to-Run helper
Assembly routines DelayLoad_Helper* were putting an argument at a
wrong location on the stack. This resulted in a GC hole that caused
random crashes, such as dotnet/cli#1785 and dotnet/roslyn#9632.
Aditya Mandaleeka [Thu, 10 Mar 2016 22:29:43 +0000 (14:29 -0800)]
Merge pull request #3603 from bendono/doc
Fix spelling and grammar.
Eugene Rozenfeld [Thu, 10 Mar 2016 21:05:23 +0000 (13:05 -0800)]
Fix for a bug in GetCORSystemDirectoryInternaL
Change
7045ca7a6de381b382a709dc0a61f49ca713d160 introduced a bug in
GetCORSystemDirectoryInternaL. The second parameter of GetRuntimeDirectory
is in/out so passing a pointer to an uninitialized cchBuffer was wrong.
The fix is to initialize cchBuffer to MAX_PATH - 1.
This was causing failures in some non CoreClr ngen scenarios.
Matthew Parkinson [Thu, 10 Mar 2016 20:37:41 +0000 (20:37 +0000)]
Addressing code review feedback.
Jan Vorlicek [Thu, 10 Mar 2016 20:32:38 +0000 (21:32 +0100)]
Merge pull request #3639 from svick/patch-1
Remove incorrect escaping in example in BOTR
Rahul Kumar [Thu, 10 Mar 2016 19:55:30 +0000 (11:55 -0800)]
Fix arm64 build break
Sejong Oh [Thu, 10 Mar 2016 19:08:15 +0000 (11:08 -0800)]
Merge pull request #3638 from sejongoh/quick_path_fix_for_init_tool_sh
Fix shell command to execute init-tool.sh on CI
Petr Onderka [Thu, 10 Mar 2016 18:32:25 +0000 (19:32 +0100)]
Remove incorrect escaping in example in BOTR
The example was casting to `ArrayMethodDesc\*` (note the backslash, possibly there to avoid italics formatting in some previous version?), which doesn't make sense. Now it casts to `ArrayMethodDesc*`.
Also, since I'm already changing this snippet, enable syntax highlighting as C++ for it.
Davis Goodin [Thu, 10 Mar 2016 18:13:45 +0000 (12:13 -0600)]
Merge pull request #3589 from dagood/test-dependency-validation
Add dependency validation/tooled-upgrade and make dependencies valid under new rules
Sejong Oh [Thu, 10 Mar 2016 17:39:10 +0000 (09:39 -0800)]
Fix shell command to execute init-tool.sh on CI
Davis Goodin [Thu, 10 Mar 2016 15:28:41 +0000 (09:28 -0600)]
Fix the centralized GC test project.jsons and other new project.jsons from merge.
Davis Goodin [Wed, 9 Mar 2016 22:45:07 +0000 (16:45 -0600)]
Upgrade new test project.json files from merge.
Davis Goodin [Tue, 8 Mar 2016 17:52:24 +0000 (11:52 -0600)]
Add documentation for dependency validation.
Davis Goodin [Thu, 3 Mar 2016 23:18:51 +0000 (17:18 -0600)]
Add System.Threading.Thread to tests to restore missing types.
Upgrade System.Runtime dependency for JIT tests to get GCHandle.
Change system.runtime 4.0.20 to 4.1.0 rc2 dependency to try to copy the runtimes to CORE_ROOT.
Fix new project.json files.
Fix hardcoded Microsoft.DotNet.xunit.performance version.
Upgrade System.Runtime.InteropServices to prerelease level to avoid type conflicts with System.Runtime.
Davis Goodin [Thu, 3 Mar 2016 22:30:38 +0000 (16:30 -0600)]
Update dependencies on packages that don't exist.
First checks to see if a stable version of that dependency was restored, and if not, found the stable.minor.patch version with the prerelease spec of the dependency.
Davis Goodin [Thu, 3 Mar 2016 21:41:17 +0000 (15:41 -0600)]
Add working validation rules for xunit packages and apply them.
Davis Goodin [Thu, 3 Mar 2016 21:28:52 +0000 (15:28 -0600)]
Remove potentially unused System.Numerics.Vectors.WindowsRuntime reference and spare WindowsRuntime.
Davis Goodin [Thu, 3 Mar 2016 21:05:15 +0000 (15:05 -0600)]
Update all versions to validate.
Davis Goodin [Thu, 3 Mar 2016 01:31:29 +0000 (19:31 -0600)]
Add validation to test build.
Don't validate generated files, and make the xunit.runner.msbuild workaround only apply to that specific package.
Jan Kotas [Thu, 10 Mar 2016 15:01:30 +0000 (07:01 -0800)]
Merge pull request #3629 from leemgs/upstream-crossbuild2
Add comment about disabling crossgen command for cross build
Sean Gillespie [Thu, 10 Mar 2016 02:57:22 +0000 (18:57 -0800)]
Merge pull request #3394 from swgillespie/gctests
Initial bringup of new set of ported GC tests.
Andy Ayers [Thu, 10 Mar 2016 00:04:11 +0000 (16:04 -0800)]
Inline refactoring: move profitability assessment to LegacyPolicy
LegacyPolicy now encapsulates all the computations needed to evaluate
whether an inline is profitable or not.
This completes the main objective of the refactoring effort, which
was to preserve and encapsulate the current inliner's behavior.
Rahul Kumar [Thu, 10 Mar 2016 01:52:13 +0000 (17:52 -0800)]
Merge pull request #3489 from rahku/EHTestPort
Port EH tests.
Rahul Kumar [Wed, 2 Mar 2016 22:36:15 +0000 (14:36 -0800)]
Port EH tests.
Geunsik Lim [Thu, 10 Mar 2016 01:46:23 +0000 (10:46 +0900)]
Remove unnecessary expressions.
Signed-off-by: Geunsik Lim <geunsik.lim@samsung.com>
Sean Gillespie [Fri, 26 Feb 2016 22:56:07 +0000 (14:56 -0800)]
Initial bringup of new set of ported GC tests.
Jose Perez Rodriguez [Thu, 10 Mar 2016 01:23:12 +0000 (17:23 -0800)]
Merge pull request #3598 from joperezr/DisableOpenSourceSign
Disable OpenSourceSign and update BuildTools
Jose Perez Rodriguez [Wed, 9 Mar 2016 00:51:35 +0000 (16:51 -0800)]
Disable OpenSourceSign and update Build Tools
Geunsik Lim [Thu, 10 Mar 2016 00:56:27 +0000 (09:56 +0900)]
Add comment about disabling crossgen command for cross build
As we all know, the current cross build generates exectuable
'crossgen' with arm architecture (without x86 architecture).
So, it can not be run to make mscorlib.ni.dll from mscorlib.dll
on X86-based build machine.
Let's add a comment why we have to skip crossgen command in
build_mscorlib() of build.sh in case of cross build for Linux/ARM.
Signed-off-by: Geunsik Lim <geunsik.lim@samsung.com>
Rama krishnan Raghupathy [Thu, 10 Mar 2016 00:38:45 +0000 (16:38 -0800)]
Merge pull request #3599 from ramarag/arm64nativetests
Enable native test binaries to be built for Arm64
Andy Ayers [Wed, 9 Mar 2016 23:57:20 +0000 (15:57 -0800)]
Merge pull request #3626 from AndyAyersMS/InlineRefactorFixNames
Fix various naming convention issues in inliner code
Aditya Mandaleeka [Wed, 9 Mar 2016 22:57:00 +0000 (14:57 -0800)]
Merge pull request #3624 from adityamandaleeka/remove_threadstatic_duplicates
Remove duplicate ThreadStatic tests
Pat Gavlin [Wed, 9 Mar 2016 22:32:39 +0000 (14:32 -0800)]
Merge pull request #3625 from pgavlin/SupportPackageVersion
Accept a build number arg in BuildRyuJitPackage.
Rama Krishnan Raghupathy [Wed, 9 Mar 2016 02:26:13 +0000 (18:26 -0800)]
Enable native test binaries to be built for Arm64
Andy Ayers [Wed, 9 Mar 2016 19:47:42 +0000 (11:47 -0800)]
Fix various naming convention issues in inliner code
Per the jit coding conventions, rename all class member functions to be
CamelCase. Rename all class data members to begin with `m_`.
tijoytom [Wed, 9 Mar 2016 21:45:14 +0000 (13:45 -0800)]
Merge pull request #3562 from tijoytom/master
Improve interop test coverage.
Pat Gavlin [Wed, 9 Mar 2016 21:13:02 +0000 (13:13 -0800)]
Accept a build number arg in BuildRyuJitPackage.
Andy Ayers [Wed, 9 Mar 2016 20:52:52 +0000 (12:52 -0800)]
Merge pull request #3602 from AndyAyersMS/InlineRefactorNativeSize
Inline refactoring: move native callsite estimator
Aditya Mandaleeka [Wed, 9 Mar 2016 20:17:19 +0000 (12:17 -0800)]
Remove duplicate ThreadStatic tests.
tijoytom [Wed, 9 Mar 2016 19:14:09 +0000 (11:14 -0800)]
Merge branch 'master' of https://github.com/dotnet/coreclr
Andy Ayers [Tue, 8 Mar 2016 23:13:56 +0000 (15:13 -0800)]
Inline refactoring: move native callsite estimator
Move the native call site estimation logic into `LegacyPolicy`. Rework
the remaining per-callsite bonus multipliers so that they fit into the
observation framework. Remove `hasNativeSizeEstimate` and instead
check that the inline is a candidate and its observation is a
discretionary inline. Fix header comment in `fgFindJumpTargets`.