Bruce Forstall [Fri, 2 Apr 2021 20:43:43 +0000 (13:43 -0700)]
Disable Test49826 for ilasm round-trip testing (#50649)
Tracking issue: https://github.com/dotnet/runtime/issues/49985
monojenkins [Fri, 2 Apr 2021 19:55:42 +0000 (15:55 -0400)]
Correcting proclib function that did not adhere to filemap wrappers (#50643)
During our upgrade of the mono runtime we ran into a situation where `mono_pe_file_map` was being called as part of opening an image now on android. This function did not currently adhere to the filemap wrapper functions so the difference in data types being used by Unity caused crashes.
Co-authored-by: UnityAlex <UnityAlex@users.noreply.github.com>
Vladimir Sadov [Fri, 2 Apr 2021 18:08:03 +0000 (11:08 -0700)]
should be GetThreadNULLOk (otherwise debugger crashes in checked) (#50614)
Aleksey Kliger (λgeek) [Fri, 2 Apr 2021 18:00:28 +0000 (14:00 -0400)]
Fix setting or getting pointer fields with reflection (#50460)
Setting a pointer via reflection incorrectly added an extra layer of
indirection, and getting a pointer via reflection incorrectly
dereferenced the value once.
As a result, setting and then getting only via reflection worked
correctly, but doing either step by directly accessing the field and the
other via reflection was broken.
Also adjusts `FieldInfoTest` to test this, for both static and instance
fields (the set path especially is different for those.)
Fixes https://github.com/mono/mono/issues/20872
This change is released under the MIT license.
Co-authored-by: Sebastian Paarmann <sebastian@s-paarmann.de>
Jose Perez Rodriguez [Fri, 2 Apr 2021 16:23:28 +0000 (09:23 -0700)]
Moving Suppress message to assembly level and adding Scope in order to get right suppressing behavior (#50626)
* Moving Suppress message to assembly level and adding Scope in order to get right suppressing behavior
* Fixing indentation
Jeremy Barton [Fri, 2 Apr 2021 16:10:08 +0000 (09:10 -0700)]
Use EVP_PKEY for RSA signing operations
With this change all RSA private key operations (excluding import/export) use the EVP_PKEY APIs.
* RSAPaddingProcessor is no longer used in conjunction with the private keys, on Linux.
* The pal_rsa.c copy of HasPrivateKey has been removed.
Marek Safar [Fri, 2 Apr 2021 16:08:36 +0000 (18:08 +0200)]
More precise descriptor for Mono's Enum/String types (#50640)
Maxim Lipnin [Fri, 2 Apr 2021 15:36:44 +0000 (18:36 +0300)]
[iOS/tvOS] Fix TargetOS case-sensitivity in tests (#50582)
* Fix TargetOS case-sensitivity in tests
* Fix TargetOS in AppleAppBuilder
* Remove the switch
Vitek Karas [Fri, 2 Apr 2021 14:02:56 +0000 (07:02 -0700)]
Allow exact duplicate input entries when bundling single file (#50476)
SDK can (and will) produce inputs for bundling which contain exact duplicate entries (same source path and same target relative path). Currently publishing such app as non-single-file works just fine (files are overwritten), but publishing it as single-file fails.
Fixing this in the SDK seems like a rather complex problem - see dotnet/sdk#16576 for more details.
There's no harm in allowing exact duplicates and ignoring them when bundling (only one copy of the file is bundled) as that will be the same behavior as non-single-file publish.
If the duplicates are not exact (different source path) then still reject those. SDK currently allows that, but it's very problematic (effectively random output).
Zoltan Varga [Fri, 2 Apr 2021 10:49:29 +0000 (06:49 -0400)]
[wasm] Fix the handling of float arguments in the icall table generator. (#50627)
The runtime icall tables use "single" instead of "float".
Marek Safar [Fri, 2 Apr 2021 08:46:08 +0000 (10:46 +0200)]
Remove AssemblyName from SPC descriptor (#50605)
It's not needed as the runtime bridge uses MonoAssemblyName
Marek Safar [Fri, 2 Apr 2021 08:22:24 +0000 (10:22 +0200)]
Remove unused MonoIOError file (#50610)
Marek Safar [Fri, 2 Apr 2021 08:11:05 +0000 (10:11 +0200)]
Use existing HexConvertor method for ascii to hex digit conversion (#50558)
monojenkins [Fri, 2 Apr 2021 07:08:11 +0000 (03:08 -0400)]
mono_breakpoint_clean_code always returns TRUE. (#50599)
Co-authored-by: jaykrell <jaykrell@users.noreply.github.com>
Egor Bogatov [Fri, 2 Apr 2021 05:10:39 +0000 (08:10 +0300)]
Simplify Comparer_get_Default (#50602)
Bruce Forstall [Fri, 2 Apr 2021 04:54:00 +0000 (21:54 -0700)]
Disable test b426654 for GCStress (#50628)
It has been failing with timeouts on GCStress=3.
Tracking issue: https://github.com/dotnet/runtime/issues/50615
Aleksey Kliger (λgeek) [Fri, 2 Apr 2021 04:33:28 +0000 (00:33 -0400)]
[mono] Add back mono_register_config_for_assembly (#50584)
Can't drop API symbols.
Aaron Robinson [Fri, 2 Apr 2021 02:29:46 +0000 (19:29 -0700)]
Split up complicated assert into parts. (#50620)
In an effort to narrow down a hard to repro GCStress issue, the assert
is being split into parts.
Bruce Forstall [Fri, 2 Apr 2021 02:17:21 +0000 (19:17 -0700)]
Disable stackoverflowtester under gcstress (#50617)
This test currently fails under many gcstress legs.
Tracking issue: https://github.com/dotnet/runtime/issues/46279
Bruce Forstall [Fri, 2 Apr 2021 01:49:55 +0000 (18:49 -0700)]
Disable TestDynamicAssembly tests (#50592)
These fail under GCStress.
Issue: https://github.com/dotnet/runtime/issues/48579
Tomáš Rylek [Fri, 2 Apr 2021 01:24:14 +0000 (18:24 -0700)]
Fix formatting string for assembly MVID mismatch (#50474)
During development of the MVID check change I at some point ended
up with mismatched printf specifiers and arguments. This change fixes
this inconsistency.
Thanks
Tomas
David Fowler [Fri, 2 Apr 2021 00:04:24 +0000 (17:04 -0700)]
Pool the underlying list and dictionary in scopes. (#50463)
- This change pools a set of scopes assuming they are short lived.
- One breaking change is that after disposal, pooled scopes will throw if services are accessed afterwards on the scope.
- Modified test to throw after dispose
Bill Wert [Thu, 1 Apr 2021 23:35:08 +0000 (16:35 -0700)]
Fix variable reference (#50623)
Jan Kotas [Thu, 1 Apr 2021 23:21:35 +0000 (16:21 -0700)]
Fix x86 warning on latest VS dogfood (#50607)
* Fix x86 warning on latest VS dogfood
* Delete unnecessary warning disable
Steve Pfister [Thu, 1 Apr 2021 21:23:01 +0000 (17:23 -0400)]
Exclude tvOS aot functional test from ci run (#50591)
Addresses https://github.com/dotnet/runtime/issues/50589
Bill Wert [Thu, 1 Apr 2021 19:00:27 +0000 (12:00 -0700)]
Android scenario fixes (#50596)
* fix path to workitems
* update queue
dotnet-maestro[bot] [Thu, 1 Apr 2021 18:30:40 +0000 (20:30 +0200)]
Update dependencies from https://github.com/mono/linker build
20210331.1 (#50555)
Microsoft.NET.ILLink.Tasks
From Version 6.0.100-preview.2.21180.4 -> To Version 6.0.100-preview.2.21181.1
Co-authored-by: dotnet-maestro[bot] <dotnet-maestro[bot]@users.noreply.github.com>
David Wrighton [Thu, 1 Apr 2021 17:47:20 +0000 (10:47 -0700)]
Fix dumping R2R embedded PGO data (#50533)
- Handle case where assembly has no embedded pgo data
- Handle case where method in file has no embedded pgo data but some other methods do
Aaron Robinson [Thu, 1 Apr 2021 17:20:22 +0000 (10:20 -0700)]
Support DOTNET_ prefix (#50507)
* Support and prefer DOTNET_ prefix.
Add support for EnableDOTNETPrefix to disable respecting the DOTNET_ prefix.
Enable environment variable caching on non-Windows platforms.
Bruce Forstall [Thu, 1 Apr 2021 16:36:19 +0000 (09:36 -0700)]
Fix GC overreporting with generics context + GS cookie (#50544)
This is a case of a generics context in `this` where we need to keep
`this` alive (but aren't required to by the VM) with a GS cookie check
(due to "JitStress: STRESS_UNSAFE_BUFFER_CHECKS"), and maybe other
JitStress modes, where `genEmitGSCookieCheck` decides to use `rcx`
to load the cookie, even though it is alive. `this` is also mostly
in a register, but it reported live as a stack slot. When the GS
cookie label gets created, the "codegen" gc info hasn\92t been updated
with the killed register, so the emitter label gets the wrong GC info
and brings rcx back alive just for the `nop`. GC cookie checks kind
of do their own register allocation since they\92re part of the epilogs.
The GS cookie code checks for keep alive using "lvRegister", but that
isn't set, since the cookie is mostly alive in a stack slock, not
always in a register. Use `lvIsInReg()` instead, so GS cookie check
will use RDX to load the cookie instead.
Fixes #50404
Jan Kotas [Thu, 1 Apr 2021 15:31:07 +0000 (08:31 -0700)]
Fix default nothrow behavior for MethodIL.GetObject (#50539)
Maxim Lipnin [Thu, 1 Apr 2021 14:31:37 +0000 (17:31 +0300)]
Fix closing tag (#50552)
Fixing the oversight in #50475
Tom Deseyn [Thu, 1 Apr 2021 13:56:14 +0000 (15:56 +0200)]
Free temporary enums on failure (#50471)
* Free temporary enums on failure
* PR feedback
SingleAccretion [Thu, 1 Apr 2021 13:27:57 +0000 (16:27 +0300)]
Delete missed mono/netcore from the documentation (#50557)
Maxim Lipnin [Thu, 1 Apr 2021 06:54:57 +0000 (09:54 +0300)]
Set iOSSimulator as TargetOS in case of empty value for the iOS functional tests (#50475)
Stephen Toub [Thu, 1 Apr 2021 05:32:03 +0000 (01:32 -0400)]
Add Random.Shared (#50297)
* Add Random.Shared
* Disable on wasm test requiring parallelism
Steve Pfister [Thu, 1 Apr 2021 04:24:08 +0000 (00:24 -0400)]
Bring up tvOS x64 staging pipeline leg (#50326)
Contributes to #33381
Elinor Fung [Thu, 1 Apr 2021 04:14:29 +0000 (21:14 -0700)]
[AndroidCrypto] Fix some leaks from missed local ref releases (#50534)
Mike McLaughlin [Thu, 1 Apr 2021 03:39:07 +0000 (20:39 -0700)]
Fix A/V in ApplyUpdate API (#50452)
* Fix A/V in ApplyUpdate API
Add thread suspend around the update API.
Issue: https://github.com/dotnet/runtime/issues/50445
* Remove thread suspension; switch to ResetCodeEntryPoint()
Tlakaelel Axayakatl Ceja [Thu, 1 Apr 2021 03:19:34 +0000 (20:19 -0700)]
Add class as a target for RequiresUnreferencedCodeAttribute (#50064)
Stephen Toub [Thu, 1 Apr 2021 00:35:42 +0000 (20:35 -0400)]
Remove a bunch of X509 PAL closures from FindCore
Alexander Köplinger [Thu, 1 Apr 2021 00:10:14 +0000 (02:10 +0200)]
Cleanup name of test exclusion itemgroup (#50513)
Follow up from feedback in https://github.com/dotnet/runtime/pull/50332
Levi Broderick [Thu, 1 Apr 2021 00:08:42 +0000 (17:08 -0700)]
Add ExceptionDispatchInfo.SetRemoteStackTrace (#50392)
Elinor Fung [Thu, 1 Apr 2021 00:06:50 +0000 (17:06 -0700)]
Stop skipping entire System.Net.Http functional test suite on Android (#50509)
It is skipped based on flakiness/slowness on a Checked/Debug runtime, which we are not currently
using for Android tests. Put the test suite under Android test ProjectExclusions instead.
dotnet-maestro[bot] [Thu, 1 Apr 2021 00:06:35 +0000 (20:06 -0400)]
Update dependencies from https://github.com/mono/linker build
20210330.4 (#50481)
Microsoft.NET.ILLink.Tasks
From Version 6.0.100-preview.2.21180.1 -> To Version 6.0.100-preview.2.21180.4
Co-authored-by: dotnet-maestro[bot] <dotnet-maestro[bot]@users.noreply.github.com>
Ulrich Weigand [Thu, 1 Apr 2021 00:06:05 +0000 (02:06 +0200)]
Big-endian fix: Accessing OBJECTHEADER in ImageConverter (#50499)
* Big-endian fix: Accessing OBJECTHEADER in ImageConverter
The GetBitmapStream routine in ImageConverter is supposed to
skip an embedded OBJECTHEADER structure. However, this
structure needs to be read in little-endian mode, not in
native byte order.
Fixes the ImageConverterTest.ImageWithOleHeader test on
big-endian systems.
* Update src/libraries/System.Drawing.Common/src/System/Drawing/ImageConverter.cs
Co-authored-by: Stephen Toub <stoub@microsoft.com>
Stephen Toub [Thu, 1 Apr 2021 00:05:51 +0000 (20:05 -0400)]
Avoid closure in WebSocketHandle.ConnectAsync (#50502)
Stephen Toub [Thu, 1 Apr 2021 00:05:10 +0000 (20:05 -0400)]
Fix accidental closure in Regex.Replace (#50512)
And sprinkle some "static" around to help avoid this in the future.
Bill Wert [Wed, 31 Mar 2021 22:54:45 +0000 (15:54 -0700)]
fix XML tag (#50523)
Bill Wert [Wed, 31 Mar 2021 22:43:50 +0000 (15:43 -0700)]
Fix linux test legs (#50518)
dotnet-maestro[bot] [Wed, 31 Mar 2021 21:46:37 +0000 (21:46 +0000)]
[main] Update dependencies from dotnet/arcade dotnet/icu dotnet/xharness dotnet/llvm-project (#50416)
[main] Update dependencies from dotnet/arcade dotnet/icu dotnet/xharness dotnet/llvm-project
- Revert publishing branch changes
https://github.com/dotnet/arcade/issues/6987
- reintroduce publishing branch workaround
https://github.com/dotnet/arcade/issues/6987
Natalia Kondratyeva [Wed, 31 Mar 2021 21:40:43 +0000 (23:40 +0200)]
Fix msquic.dll loading (#50506)
Alexander Köplinger [Wed, 31 Mar 2021 21:39:24 +0000 (23:39 +0200)]
Reenable Android.Device_Emulator.Aot.Test.csproj (#50332)
* Reenable Android.Device_Emulator.Aot.Test.csproj
Closes https://github.com/dotnet/runtime/issues/49757
* Update tests.proj
Stephen Toub [Wed, 31 Mar 2021 21:25:52 +0000 (17:25 -0400)]
Avoid closure in ValidationContext if serviceProvider is null (#50483)
Bill Wert [Wed, 31 Mar 2021 21:06:48 +0000 (14:06 -0700)]
add android scenarios (#50407)
Juan Hoyos [Wed, 31 Mar 2021 20:34:25 +0000 (13:34 -0700)]
Fix usage of process_vm_readv in createdump (#50477)
* Fix usage of process_vm_readv in createdump; fallback to pread64
* No-op logging
Prashanth Govindarajan [Wed, 31 Mar 2021 20:08:23 +0000 (13:08 -0700)]
Prohibit BF unless the app opts in (#48527)
* Prohibit BF unless the app opts in
* Address feedback
* First unit test
* 2nd unit test. Checkpoint
* sq
* sq
* sq
* Address Levi's feedback
* sq
* Address nits
* Linker changes and tests
* sq
* sq
* Linker warnings ids
* Address comments
* Change trimming test so linker can detect a pattern
* sq
* Address comments
* sq
Co-authored-by: Eric Erhardt <eric.erhardt@microsoft.com>
Elinor Fung [Wed, 31 Mar 2021 19:09:35 +0000 (12:09 -0700)]
[AndroidCrypto] Make SslStreamCertificateContext treat an empty chain as not finding intermediates (#50195)
Andy Gocke [Wed, 31 Mar 2021 19:08:18 +0000 (12:08 -0700)]
Also build bundleproj for PGO (#50434)
I previously disabled publishing all but Microsoft.NETCore.App.sfxproj,
which publishes the dotnet-runtime-internal package. This is almost, but
not quite, the right package. It doesn't contain the host dependencies,
although it does contain the right runtime files. The right package is
Microsoft.NETCore.App.bundleproj, which builds dotnet-runtime. Unfortunately,
I didn't catch this in manual testing because I think copying over the temp
bits reused the host dependencies from the previous install location. I've
wiped this from scratch and it looks fine now. (crossed fingers)
Andy Gocke [Wed, 31 Mar 2021 18:34:05 +0000 (11:34 -0700)]
Add option for building a test exe as single file (#42972)
* Add option for building a test exe as single file
* Remove left over test
* Add target to exclude references from single-file
* First attempt at adding a CI job
* Opt-in specific libraries for single-file testing support
Start with System.Collections as all tests pass.
* Config testing using single-file in build.cmd
* Change yml suffix name to SingleFile
* Windows_NT_x64 -> windows_x64
* Fix for helix queueing
* Respond to host rename
* Change TargetOS to check for windows
* chmod test exe on linux
* Direct singlefilehost to the locally built copy
* Adjust singlefilehost copy+call
* Add .exe suffix on Windows
* Move libraries after hosts build to allow for libs.test to depend on hosts build
* Split up host and libs packaging and tests
* Move host packaging
* Move pretest up
* Move packages up as well
* Reorder libs pretest and libs.packages
* Add isSingleFile build parameter to limit Linux Helix jobs
* Typo
* Change conditional check
* Fix yml
* Fix yml
* Fix neq
* Fix subsets
* Typo
* Fix
* Adjust assert
* Include code from Michal to skip failing test
* Remove empty ItemGroup
* Update src/libraries/Common/tests/SingleFileTestRunner/SingleFileTestRunner.cs
Co-authored-by: Michal Strehovský <MichalStrehovsky@users.noreply.github.com>
* Update eng/testing/tests.singlefile.targets
Co-authored-by: Michal Strehovský <MichalStrehovsky@users.noreply.github.com>
* Apply suggestions from code review
Co-authored-by: Michal Strehovský <MichalStrehovsky@users.noreply.github.com>
* Use ProjectExclusions
* Update eng/testing/tests.singlefile.targets
Co-authored-by: Michal Strehovský <MichalStrehovsky@users.noreply.github.com>
* Revert changes
* Remove host build from tests
Co-authored-by: Michal Strehovský <MichalStrehovsky@users.noreply.github.com>
David Wrighton [Wed, 31 Mar 2021 17:50:25 +0000 (10:50 -0700)]
Refactor missing reference errors to allow not throwing for all cases (#50437)
- Build the concept of a cacheable resolution failure
- Plumb it through the Ecma type loader and the required public api surfaces
- Use it within the Mibc parser to avoid throwing
Bruce Forstall [Wed, 31 Mar 2021 17:41:10 +0000 (10:41 -0700)]
Support private SuperPMI MCH file stores (#50268)
* Support private SuperPMI MCH file stores
In addition to the default Azure Storage SuperPMI location,
support private data stores via the new `-private_store` argument
to `replay` and `asmdiffs`. Private stores are typically file
system locations, or, on Windows, UNC paths.
There already exists the `-mch_files` argument, which allows you
to use any specified set of MCH files. However, when specifying
`-mch_files`, only those files are used and the normal stores are
ignored.
The private store mechanism is more convenient in some scenarios,
as it treats the specified private stores equivalently to the default
Azure Storage store.
In addition, both private stores and `-mch_files` handle downloading
ZIP files, un-ZIPping, and caching their contents.
Finally, a new top-level command `upload-private` is introduced to
help create private stores by specifying a set of MCH files to
use. It ZIP compresses the files to upload.
* Fix `-base_jit_path`
If `-base_jit_path` is specified on the command line, ensure it is a
full path before using it.
Levi Broderick [Wed, 31 Mar 2021 17:30:24 +0000 (10:30 -0700)]
Add IsWellKnownStringComparer methods (#50312)
Zoltan Varga [Wed, 31 Mar 2021 17:21:16 +0000 (13:21 -0400)]
[mono] Fix the handling of jit icall wrappers in mono_aot_method_hash (), the name might be different at runtime if DISABLE_JIT is set. (#50482)
Fan Yang [Wed, 31 Mar 2021 17:03:27 +0000 (13:03 -0400)]
Update testing-android.md (#50484)
Eric Erhardt [Wed, 31 Mar 2021 16:13:49 +0000 (11:13 -0500)]
Rename CreateInstanceT_StructWithoutDefaultConstructor_ThrowsMissingMethodException test (#50191)
This doesn't throw an exception, so the test shouldn't be named "ThrowsMissingMethodException"
Dong-Heon Jung [Wed, 31 Mar 2021 15:42:59 +0000 (00:42 +0900)]
Fix a crash in WriteMethodProfileDataLogFile (#50473)
pEmitter is not assigned in https://github.com/dotnet/runtime/pull/50242
Stephen Toub [Wed, 31 Mar 2021 13:30:24 +0000 (09:30 -0400)]
Add AsyncMethodBuilderOverride and PoolingAsyncValueTaskMethodBuilders (#50116)
* Add AsyncMethodBuilderOverride and PoolingAsyncValueTaskMethodBuilders
* Revise based on C# LDM changes to model
* Fix API compat errors
Ivan Beňovic [Wed, 31 Mar 2021 11:13:59 +0000 (13:13 +0200)]
Allow bounded channel to be created with drop delegate (#50331)
* Bounded channel can be created with drop delegate.
- Add additional CreateBounded overload with delegate parameter that will be called when item is being dropped from channel
- Added unit tests
* Fix typo in comment.
* Apply suggestions from code review
Co-authored-by: Stephen Toub <stoub@microsoft.com>
* Call drop delegate outside of lock statement.
* Use overload of CreateBounded method instead of calling ctor directly.
* Code review suggestions refactor.
* Move Monitor.Enter before try and use local scoped parent variable everywhere.
* Drop delegate should not be called while sync lock is held.
Enqueuing of new item should be done while sync lock is being held.
Added additional tests.
* Rerun gitlab CI.
* Do not run deadlock test for bounded channels if platform do not support threading.
* Apply suggestions from code review
Co-authored-by: Ivan Benovic <ivan.benovic@innovatrics.com>
Co-authored-by: Stephen Toub <stoub@microsoft.com>
xtqqczze [Wed, 31 Mar 2021 11:02:11 +0000 (12:02 +0100)]
Fix typo in `string.Join` (#50467)
Follow-up #44032.
Adam Sitnik [Wed, 31 Mar 2021 11:01:20 +0000 (13:01 +0200)]
[FileStream] Rename "Legacy" to "Net5Compat" (#50470)
* rename Legacy => Net5Compat (System.Private.CoreLib)
* rename Legacy => Net5Compat (test projects)
Vlad Brezae [Wed, 31 Mar 2021 08:32:05 +0000 (11:32 +0300)]
[interp] Fix tracing (#50415)
Ilya [Wed, 31 Mar 2021 08:02:48 +0000 (13:02 +0500)]
Add tracking a recursion depth in System.IO.FileSystem (#48148)
* Add tracking a depth recursion in System.IO.FileSystem
* Update src/libraries/System.IO.FileSystem/src/System/IO/EnumerationOptions.cs
Co-authored-by: Carlos Sanchez <1175054+carlossanlop@users.noreply.github.com>
* Fix docs
* Address feedbacks
* Fix typo
* Add MaxRecursionDepth value check
* Fix net48 build: Add string to MS.IO.Redist resx file
* Fix test and initialization.
* Address feedback
Co-authored-by: Carlos Sanchez <1175054+carlossanlop@users.noreply.github.com>
Co-authored-by: David Cantu <dacantu@microsoft.com>
Stephen Toub [Wed, 31 Mar 2021 07:23:25 +0000 (03:23 -0400)]
Rename Random.LegacyImpl (#50456)
To be more specific about what it is.
Aleksey Kliger (λgeek) [Wed, 31 Mar 2021 07:06:12 +0000 (03:06 -0400)]
[tests] Add Mono issue to CallConvMemberFunction tests (#50461)
Adam Sitnik [Wed, 31 Mar 2021 06:53:52 +0000 (08:53 +0200)]
ensure MetadataReader fast path works with every FileStream on every OS (#50367)
Adam Sitnik [Wed, 31 Mar 2021 06:42:18 +0000 (08:42 +0200)]
Dont cache file length when handle has been exposed (#50424)
Zoltan Varga [Wed, 31 Mar 2021 05:12:47 +0000 (01:12 -0400)]
[mono] Address comments from PR #50328. (#50462)
* Use g_slist_concat ().
* Add MemoryManagerLock to the lock tracer.
* Remove some dead code.
Buyaa Namnan [Wed, 31 Mar 2021 04:41:14 +0000 (21:41 -0700)]
Resolve unknown platform name warnings (#50193)
* Add targets as supported platforms
* Update analyzer version
* Update comment, remove '' != 'false' condition
Brian Robbins [Wed, 31 Mar 2021 04:28:50 +0000 (21:28 -0700)]
Specify the Size of List<T> Instances in InvokeTypeInfo (#50438)
* Initialize List<T> instances in InvokeTypeInfo with known sizes.
* Switch from List<T> to arrays.
Aaron Robinson [Wed, 31 Mar 2021 03:51:24 +0000 (20:51 -0700)]
Merge REGUTIL logic for environment variables into CLRConfig (#50393)
* Merge REGUTIL logic for environment variables into CLRConfig
Remove REGUTIL
Remove unused CLR configuration options
David Pine [Wed, 31 Mar 2021 00:20:57 +0000 (19:20 -0500)]
Added the `ConfigureDefaults` API, as an extension method of `IHostBuilder` (#50447)
* Added the ConfigureDefaults API, as an extension method on the IHostBuilder interface. Fix #36003
* Apply suggestions from code review
Co-authored-by: Eric Erhardt <eric.erhardt@microsoft.com>
* Update src/libraries/Microsoft.Extensions.Hosting/src/Host.cs
Co-authored-by: Eric Erhardt <eric.erhardt@microsoft.com>
Tomáš Rylek [Tue, 30 Mar 2021 23:44:40 +0000 (16:44 -0700)]
Fix handling of array constructors in Crossgen2 (#48557)
Fixes: https://github.com/dotnet/runtime/issues/48204
Alexander Köplinger [Tue, 30 Mar 2021 23:42:47 +0000 (01:42 +0200)]
Make CLOCK_MONOTONIC handling more precise in pal_threading.c (#50441)
The root cause of the problem in https://github.com/dotnet/runtime/pull/50388 turned out to be because we do not have `pthread_condattr_setclock(..., CLOCK_MONOTONIC);` on iOS.
This caused the timeout argument for `pthread_cond_timedwait` to be interpreted as an _absolute system time_ (in seconds since the Unix epoch), however the value we got back from `clock_gettime(CLOCK_MONOTONIC, ...)` was actually some value based on the uptime of the system.
Since the uptime is much smaller than the system time the wait immediately returned.
Harden the handling by adding a check for `HAVE_PTHREAD_CONDATTR_SETCLOCK` like we already do in `SystemNative_LowLevelMonitor_Create()`
Zoltan Varga [Tue, 30 Mar 2021 23:18:03 +0000 (19:18 -0400)]
[mono] Remove domain lock (#50328)
* [mono] Reduce domain lock usage.
* Avoid double locking the loader+domain lock, its not needed.
* Use the mem manager lock and the loader lock in a few places instead of
the domain lock.
* Use lock free code in mono_jit_runtime_invoke ().
* Avoid locking around mono_jit_info_table_add (), it does locking itself.
* [mono] Internalize access to MonoMemoryManager.mp to memory-manager.c.
Also add a separate non-coop lock to protect it, since its accessed
very frequently.
* Use a separate lock for each memory manager.
* Remove the domain lock.
Zoltan Varga [Tue, 30 Mar 2021 23:17:23 +0000 (19:17 -0400)]
[mono] Move the jit info table definitions into a separate jit-info.h file, it no longer depends on domains. (#50423)
Bruce Forstall [Tue, 30 Mar 2021 22:45:56 +0000 (15:45 -0700)]
Rename the superpmi.py asmdiffs output directory (#50403)
* Rename the superpmi.py asmdiffs output directory
Previously, the directory name was `asm.os.arch.build_flavor`.
Now, use the MCH file name instead. Since the MCH filename of our
collections has the target os/arch/flavor, those components will still
be there (but will be based on the MCH file, not the host; this is
important if you are doing cross-compile asm diffs). It will
also have the MCH base name, such as "benchmarks.run", to help
distinguish different runs. Of course, if there is a conflict with
a previous run, `.1`, `.2`, etc., is appended to the directory name.
Also, add a new, optional `-tag` option to `superpmi.py asmdiffs` to
allow specifying an additional word to prefix the artifacts folder
base name. This might help more easily distinguish different experiments.
So, `superpmi.py asmdiffs -filter benchmarks -tag test2` might give you
dasm files in these directories:
```
c:\runtime\artifacts\spmi\asm.test2.benchmarks.run.windows.x64.checked.3\base
c:\runtime\artifacts\spmi\asm.test2.benchmarks.run.windows.x64.checked.3\diff
```
* Review feedback
Created `create_artifacts_base_name()` function.
Fixed `run_and_log` to display command line of command being
invoked at the specified log level. Currently only affects invoking
`jit-analyze`, which makes the `jit-analyze` command line visible
on the console, not just in the log file.
Maoni Stephens [Tue, 30 Mar 2021 22:29:29 +0000 (15:29 -0700)]
expose alloc data on dbi (#50341)
we are already exposing this data for sos, should make this available for debuggers to use as well. it's very lightweight and would be very useful for things like showing exactly how many bytes were allocated on the thread between 2 breakpoints.
Alexander Köplinger [Tue, 30 Mar 2021 16:52:30 +0000 (18:52 +0200)]
Use OptimizeForSize condition in SPC only on Browser (#50413)
The size impact doesn't matter for tvOS.
David Cantú [Tue, 30 Mar 2021 16:45:10 +0000 (09:45 -0700)]
FileSystemGlobbing: Allow rootDir paths ending with separator to match files correctly (#45139)
* Allow rootDir paths ending with separator to match files correctly
* Avoid running tests with windows-like absolute paths on non-windows platforms
* Address test suggestion
Bruce Forstall [Tue, 30 Mar 2021 16:42:04 +0000 (09:42 -0700)]
Fix issues with JitDump (#50395)
1. On 64 bit platforms, with "long" bitset types (> 64 bits), when
converting to string format for display, we were only displaying
the low 32 bits of every 64 bit chunk.
2. Add display of the `IGF_ALIGN` flag.
dotnet-maestro[bot] [Tue, 30 Mar 2021 16:25:00 +0000 (11:25 -0500)]
Update dependencies from https://github.com/mono/linker build
20210330.1 (#50417)
Microsoft.NET.ILLink.Tasks
From Version 6.0.100-preview.2.21177.1 -> To Version 6.0.100-preview.2.21180.1
Co-authored-by: dotnet-maestro[bot] <dotnet-maestro[bot]@users.noreply.github.com>
Omair Majid [Tue, 30 Mar 2021 16:08:02 +0000 (12:08 -0400)]
Clean up allocated memory on error in CordbEval::NewParameterizedObject (#50371)
pArgData is allocated in this piece of code; if we are failing and
returning early, we should deallocate this memory to avoid a memory
leak.
Tomas Weinfurt [Tue, 30 Mar 2021 16:04:42 +0000 (09:04 -0700)]
increase timout on ConnectTimeout_TimesOutSSLAuth_Throws test (#50351)
Ryan Lucia [Tue, 30 Mar 2021 14:31:36 +0000 (10:31 -0400)]
Fix waits on iOS (#50388)
I don't understand why CLOCK_MONOTONIC doesn't work even in newer simulators here, but this will at least fix the problem for now.
Jan Kotas [Tue, 30 Mar 2021 14:10:16 +0000 (07:10 -0700)]
Revert "Fix release builds with latest VS dogfood (#49948)" (#50398)
This reverts commit
3961d91812c997af18ff1ebbdac53bcc4d84fb02.
This workaround is no longer needed.
Vlad Brezae [Tue, 30 Mar 2021 11:23:03 +0000 (14:23 +0300)]
[interp] Add a few super instructions (#50361)
* [interp] Fix PROFILE_INTERP build
* [interp] Add ldc.i8.0 opcode, replacing ldnull on 64bit
We still need to use this opcode in more places, instead of the generic ldc.i8
* [interp] Add a few super instructions
We depend on the cprop pass to init local_ref_count. We only add a super instruction if the definition of the an intermediary result is done in the same basic block (the dreg is local var) and it is not used anywhere else, otherwise we can't clear the instruction.
* [interp] Improve the marvin block intrinsic
By removing also 2 additional ldloca instructions, and enabling cprop for these vars that no longer have their address taken.
dotnet-maestro[bot] [Tue, 30 Mar 2021 11:21:28 +0000 (07:21 -0400)]
[main] Update dependencies from dotnet/runtime dotnet/xharness (#50358)
* Update dependencies from https://github.com/dotnet/runtime build
20210328.6
Microsoft.NETCore.DotNetHost , Microsoft.NETCore.DotNetHostPolicy , Microsoft.NETCore.ILAsm , runtime.native.System.IO.Ports , Microsoft.NET.Sdk.IL , System.Runtime.CompilerServices.Unsafe , System.Text.Json
From Version 6.0.0-preview.3.21169.6 -> To Version 6.0.0-preview.4.21178.6
* Update dependencies from https://github.com/dotnet/xharness build
20210329.2
Microsoft.DotNet.XHarness.CLI , Microsoft.DotNet.XHarness.TestRunners.Xunit
From Version 1.0.0-prerelease.21176.1 -> To Version 1.0.0-prerelease.21179.2
Co-authored-by: dotnet-maestro[bot] <dotnet-maestro[bot]@users.noreply.github.com>
Ben Adams [Tue, 30 Mar 2021 11:19:31 +0000 (12:19 +0100)]
Add "Help Wanted" badge (#50352)
* Add "Help Wanted" badge
* Move help wanted second
Neale Ferguson [Tue, 30 Mar 2021 10:22:26 +0000 (20:22 +1000)]
s390x: Fixes to tailcall and array parameter fix (#49466)
### Small s390x fixes
- Add another tailcall disqualifying condition
- Minor formatting change
### New Array Parameter fix
We have been seeing msbuild tasks fail due to an OverflowException being raised in `(wrapper_alloc)_object:AllocVector` because the number of elements appears to be negative (0xffffffff00000088).
However, this is just the result of a missing type conversion somehow. The original C# code is simply:
```
_metadataStringTable = new string[_reader.ReadInt32()];
```
The IL corresponding to that line reads:
```
converting (in B2: stack: 2) IL_0021: callvirt 0x0a00035e
cmethod = int System.IO.BinaryReader:ReadInt32 ()
Call requires: 1 parameters
converting (in B2: stack: 2) IL_0026: newarr 0x010000b6
```
Note that the return from the `ReadInt32` call is of type `int` (i.e. 32 bits) and is passed unmodified as argument to `newarr`. This remains the same in the initial stages of Mono JIT compilation:
```
call_membase R43 <- [R45 + 0xe8] [int System.IO.BinaryReader:ReadInt32 ()] [s390_r2 <- R44] clobbers: c
il_seq_point il: 0x26, nonempty-stack
newarr R46 <- R43
```
But later the `newarr` is implemented in terms of a function call:
```
call_membase R43 <- [R45 + 0xe8] [int System.IO.BinaryReader:ReadInt32 ()] [s390_r2 <- R44] clobbers: c
il_seq_point il: 0x26, nonempty-stack
i8const R81 <- [
2929315620864]
move R83 <- R81
move R84 <- R43
call R46 <- [(wrapper alloc) object object:AllocVector (intptr,intptr)] [s390_r2 <- R83] [s390_r3 <- R84] clobbers: c
```
And here the argument is of type `intptr` (i.e. 64 bits). But the return value of ReadInt32 is still passed through directly to `AllocVector` without any conversion, and this remains true in the final assembler code:
```
b2: 0d e1 basr %r14,%r1 // ReadInt32
b4: b9 04 00 32 lgr %r3,%r2 // <<- simple move of the return value
b8: c0 28 00 00 02 aa iihf %r2,682
be: c0 29 08 d1 28 00 iilf %r2,
147924992
c4: c0 e8 00 00 03 ff iihf %r14,1023
ca: c0 e9 9d ec 97 08 iilf %r14,
2649528072
d0: 0d ee basr %r14,%r14 // AllocVector
```
This becomes a problem because the implementation of `AllocVector` does indeed expect a 64-bit input, and throws an exception if that is negative.
In addition, the implementation of `ReadInt32` only computes a 32-bit result in the low 32 bits of the register, and leaves the upper half undefined. Due to the particular code sequence we get a return with the upper half nonzero even though the lower half is a positive integer:
```
e6: e3 20 f0 e0 00 14 lgf %r2,224(%r15) <<- if this is a 32-bit value with the top bit set, then the upper half of %r2 is now 0xffffffff
ec: e3 20 f0 b8 00 50 sty %r2,184(%r15)
f2: b9 04 00 32 lgr %r3,%r2 <<- and so is %r3
f6: c0 01 00 ff 00 ff lgfi %r0,
16711935
fc: b9 e4 00 43 ngrk %r4,%r3,%r0
100: b9 14 00 24 lgfr %r2,%r4
104: 88 20 00 08 srl %r2,8
108: 89 40 00 18 sll %r4,24
10c: b9 e6 40 22 ogrk %r2,%r2,%r4
110: c0 01 ff 00 ff 00 lgfi %r0,-
16711936 <<-- the upper half of %r0 is also 0xffffffff
116: b9 e4 00 43 ngrk %r4,%r3,%r0 <<-- and so is %r4
11a: b9 14 00 34 lgfr %r3,%r4
11e: 89 30 00 08 sll %r3,8
122: 88 40 00 18 srl %r4,24
126: b9 e6 40 33 ogrk %r3,%r3,%r4 <<- and now %r3
12a: b9 08 00 23 agr %r2,%r3 <<- and therefore %r2, where the upper half for 0 before
12e: e3 20 f0 b8 00 50 sty %r2,184(%r15)
134: a7 fb 00 f8 aghi %r15,248
138: eb 6e f0 30 00 04 lmg %r6,%r14,48(%r15)
13e: 07 fe br %r14 <<- and here it is returned unchanged
```