There are roughly three categories:
1. Already implemented or completely unused defines
2. Defines we can remove because they're no longer useful:
- `MAJOR_IN_MKDEV` and `MAJOR_IN_SYSMACROS`
Not used in the runtime anymore.
- `HAVE_EPOLL` and `HAVE_KQUEUE`
Not used in the runtime anymore.
- `ANDROID_UNIFIED_HEADERS` and `HAVE_ANDROID_NDK_VERSION_H`
We're requiring a recent Android NDK now that only supports unified headers.
- `MONO_DL_NEED_USCORE`
It doesn't seem needed anymore.
dlsym() on macOS automatically prepends an underscore and from what I can find it was mostly for old *BSDs.
- `GLIBC_BEFORE_2_3_4_SCHED_SETAFFINITY`
We don't support such an old glibc anymore.
- `HAVE_SIOCGIFCONF`
We don't use this and related networking code in the runtime anymore.
- `DISABLE_HW_TRAPS`
We have `MONO_ARCH_DISABLE_HW_TRAPS` instead.
- `MONO_BIG_ARRAYS`
No longer needed and probably broken due to missing BCL changes.
- `ENABLE_MONODROID` and replace `ENABLE_MONOTOUCH` with `TARGET_APPLE_MOBILE`
ENABLE_MONODROID was unused and MonoTouch was an earlier name for Xamarin.iOS, replace it with something more descriptive
- `MONO_JEMALLOC`
We haven't used it and don't plan to.
- `MONO_XEN_OPT`
Xen is not a primary scenario for Mono anymore so we probably don't need this optimization.
3. Defines which are still useful and are implemented by this PR
- Fix `MONO_KEYWORD_THREAD` build and implement `HAVE_TLS_MODEL_ATTR`
- Boehm/Null GC support
- `MONO_SMALL_CONFIG`
- `USE_MALLOC_FOR_MEMPOOLS`
- `ENABLE_DTRACE`
- `ENABLE_EXPERIMENT_TIERED` and remove dynamic experiments
- `ENABLE_OVERRIDABLE_ALLOCATORS`
- `MONO_ARCH_ILP32`
- `RISCV_FPABI`
- `__mono_ppc64__`
- `HAVE_ARMV5/6`. There is an open question about what we'll do with `HAVE_ARMV7`, added a TODO comment.
* Add Dot intrinsic
* WIP saddvq/uaddvq
* Add Dot and Sum impl for Arm64
* Add faddv
* Fix the Dot and Sum implementations
* Cleanup
* Cleanup
* Fix type checking
* Fix type checking
* Fix indentation
* Fix edge case with single element vectors
* Improve implementation
* Use ApiCompatTask in ApiCompat.proj
As ApiCompat now has a task that allows to run in-proc, use that instead
of shelving out. Also use items instead of properties to define
the attribute exclusion list.
* Update ApiCompat version
* [mono] Fix parsing options added by utils/options.h when using --response.
* [mono][jit] Add support for WASM EH.
WASM EH uses the newly added llvm catchpad instructions instead of
the old landingpad instructions. Add a --wasm-exceptions option
to control whenever the old JS style or the new WASM style code
is generated.
* Add support for the new WASM Exception Handling feature.
It is controlled by the $(WasmExceptionHandling) msbuild property.
Its only used in AOT mode, the interpreter mode wouldn't benefit from it.
The only file in the runtime which contains a throw or a catch is
llvm-runtime.cpp, so create 2 small static libraries from this file:
* libmono-wasm-eh-js - compiled to use JS style EH
* libmono-wasm-eh-wasm - compiled to use WASM style EH
This avoids the need to compile 2 versions of the whole runtime.
Requires emscripten 3.x.
* Add missing quote
* Update src/mono/mono/mini/CMakeLists.txt
Co-authored-by: Aleksey Kliger (λgeek) <[email protected]>
* Update src/mono/CMakeLists.txt
Co-authored-by: Aleksey Kliger (λgeek) <[email protected]>
Co-authored-by: Radek Doulik <[email protected]>
Co-authored-by: Aleksey Kliger (λgeek) <[email protected]>
Microsoft.NET.ILLink.Tasks
From Version 7.0.100-1.22158.1 -> To Version 7.0.100-1.22160.1
Co-authored-by: dotnet-maestro[bot] <dotnet-maestro[bot]@users.noreply.github.com>
* Remove duplicated code - I know that tests should be descriptive, but repeating 100 times that we want to capture output doesn't feel necessary.
* For some of them move more stuff into the shared test state - this improves perf as we avoid repeating the setup (copying files around) for each test case, we do it once for the entire class
I also changed some of the tests to "Theory" as it's easier to read that way.
* `SDKLookup.cs` - moved most of the state in to the shared state to speed up the tests.
* Adding more cases into the "theory" tests
* Most notably for the framework resolution I added variations on the TFM (which will be needed when we implement disable of MLL)
* Adding new tests mostly around "list runtimes" (various forms), "list sdks" (various forms) and errors (which also list runtimes or sdks)
* Ported all of the `MultiLevelLookupFramework` tests over to the `FrameworkResolution` and `DependencyResolutions` suites which have a more robust test infra and can run the same tests much faster. Along the way I added lot more variations on top of the existing tests:
* `PerAssemblyVersionResolutionMultipleFrameworks.cs` - this is actually not an MLL test, but I moved it to the new infra to make it much faster
* `MultipleHives.cs` - MLL framework resolution tests
For SDK resolution I kept the `MultiLevelSDKLookup.cs` just removed code duplication and added new variants.
For the core reviewers: I promise I didn't remove any single test case in spirit with these exceptions:
* We had tests which validated that framework resolution is not affected by frameworks in current directory and also by frameworks in the user's directory. I left some basic test for the current directory check, but I completely removed the user's directory variant as the product simply doesn't have any code around that anymore.
* Remove "user" directory from SDK resolution tests
The product simply doesn't have any knowledge about user directory SDK lookup, so there's little value in testing for it (even though the test expects it to not have any impact).
Co-authored-by: Elinor Fung <[email protected]>
This changes the one-shot decryption to allow decrypting directly in to user-supplied buffers, even in the presence of padding, so that a copy and CryptoPool rental can be avoided.
If the buffer is large enough to accommodate the padding, then decrypt directly in to the supplied buffer, zero the padding, and indicate the de-padded amount in bytesWritten.
If the buffer is possibly large enough, then decrypt all but the last block in to the buffer, and do the last block on the stack. If the de-padded amount in the stack is small enough, then copy the remaining. If it is too large, then zero everything in the user buffer.
* Overhaul how regex generated code is structured
* Address PR feedback
* Change parser to return RegexMethod
Clean up how the data is structured.
* Add a test for same method name in multiple types
* Add arm64 intrinsics for Narrow
* Fix integer types
* Fix more issues with integer types
* Add some comments
* Nit
* Change var names
* Fix format
* Fix format again
We never read that field in the runtime and it was causing unnecessary disk IO during startup (5ms on my iOS device).
It was also never set on Windows already.
System.Net.MsQuic.Transport
From Version 7.0.0-alpha.1.22159.1 -> To Version 7.0.0-alpha.1.22160.2
Co-authored-by: dotnet-maestro[bot] <dotnet-maestro[bot]@users.noreply.github.com>
* Use StartsWith(..., OrdinalIgnoreCase) in RegexCompiler / source generator
When we encounter a sequence of sets representing case-insensitive ASCII, we can simplify the code generated to just call StartsWith, which both makes it more readable but also takes advantage of the new JIT optimization to lower that into efficient vectorized comparisons based on the supplied literal.
This also cleans up some formatting in the source generator emitted code to make things much more concise and less noisy.
* Address PR feedback
* disable timing out test on mono
* Update src/libraries/System.Text.Json/tests/System.Text.Json.Tests/Serialization/CustomConverterTests/CustomConverterTests.cs
* Update dependencies from https://github.com/dotnet/arcade build 20220307.6
Microsoft.DotNet.XUnitConsoleRunner , Microsoft.DotNet.CodeAnalysis , Microsoft.DotNet.Build.Tasks.Workloads , Microsoft.DotNet.Build.Tasks.Templating , Microsoft.DotNet.Build.Tasks.TargetFramework , Microsoft.DotNet.Build.Tasks.Packaging , Microsoft.DotNet.Build.Tasks.Installers , Microsoft.DotNet.Build.Tasks.Feed , Microsoft.DotNet.Build.Tasks.Archives , Microsoft.DotNet.Arcade.Sdk , Microsoft.DotNet.ApiCompat , Microsoft.DotNet.XUnitExtensions , Microsoft.DotNet.GenAPI , Microsoft.DotNet.VersionTools.Tasks , Microsoft.DotNet.GenFacades , Microsoft.DotNet.SharedFramework.Sdk , Microsoft.DotNet.RemoteExecutor , Microsoft.DotNet.PackageTesting , Microsoft.DotNet.Helix.Sdk
From Version 2.5.1-beta.22154.3 -> To Version 2.5.1-beta.22157.6
* Update dependencies from https://github.com/dotnet/icu build 20220307.1
Microsoft.NETCore.Runtime.ICU.Transport
From Version 7.0.0-preview.3.22128.1 -> To Version 7.0.0-preview.3.22157.1
* Update dependencies from https://github.com/dotnet/xharness build 20220307.2
Microsoft.DotNet.XHarness.CLI , Microsoft.DotNet.XHarness.TestRunners.Common , Microsoft.DotNet.XHarness.TestRunners.Xunit
From Version 1.0.0-prerelease.22153.2 -> To Version 1.0.0-prerelease.22157.2
* Update dependencies from https://github.com/dotnet/roslyn-analyzers build 20220307.1
Microsoft.CodeAnalysis.NetAnalyzers
From Version 7.0.0-preview1.22153.1 -> To Version 7.0.0-preview1.22157.1
* Update dependencies from https://github.com/dotnet/hotreload-utils build 20220307.2
Microsoft.DotNet.HotReload.Utils.Generator.BuildTool
From Version 1.0.2-alpha.0.22128.2 -> To Version 1.0.2-alpha.0.22157.2
* Update dependencies from https://github.com/dotnet/llvm-project build 20220307.1
runtime.win-x64.Microsoft.NETCore.Runtime.ObjWriter , runtime.win-arm64.Microsoft.NETCore.Runtime.ObjWriter , runtime.osx.10.12-x64.Microsoft.NETCore.Runtime.ObjWriter , runtime.osx.11.0-arm64.Microsoft.NETCore.Runtime.ObjWriter , runtime.linux-x64.Microsoft.NETCore.Runtime.ObjWriter , runtime.linux-musl-x64.Microsoft.NETCore.Runtime.ObjWriter , runtime.linux-musl-arm64.Microsoft.NETCore.Runtime.ObjWriter , runtime.linux-arm64.Microsoft.NETCore.Runtime.ObjWriter
From Version 1.0.0-alpha.1.22128.1 -> To Version 1.0.0-alpha.1.22157.1
* Use full names for GenericHandler passed to xharness
* Update dependencies from https://github.com/dotnet/msquic build 20220309.1
System.Net.MsQuic.Transport
From Version 7.0.0-alpha.1.21529.3 -> To Version 7.0.0-alpha.1.22159.1
* Update dependencies from https://github.com/dotnet/xharness build 20220310.1
Microsoft.DotNet.XHarness.CLI , Microsoft.DotNet.XHarness.TestRunners.Common , Microsoft.DotNet.XHarness.TestRunners.Xunit
From Version 1.0.0-prerelease.22153.2 -> To Version 1.0.0-prerelease.22160.1
Co-authored-by: dotnet-maestro[bot] <dotnet-maestro[bot]@users.noreply.github.com>
Co-authored-by: Ankit Jain <[email protected]>
* reproduce issue: collecting disposables even if composite token has alredy been changed
* stop collecting disposables if composite token has already been changed
* Update src/libraries/Microsoft.Extensions.Primitives/tests/CompositeChangeTokenTest.cs
Co-authored-by: Eric Erhardt <[email protected]>
* Update src/libraries/Microsoft.Extensions.Primitives/tests/CompositeChangeTokenTest.cs
Co-authored-by: Eric Erhardt <[email protected]>
Co-authored-by: Eric Erhardt <[email protected]>
Fixes#44689
As of #65148, libraries use different approach to handling OpenSSL errors. The original assert which would be hit if a previous operation left errors in the queue is no longer necessary as all OpenSSL calls are prepended by `ERR_clear_error()` to make sure the latest (and most relevant) error is reported.
Details:
In gen 1 GCs, we recompute the budget at the end of the GC. However, the amount consumed in the current GC is nowhere taken into account. That causes us to do at most alternate gen 0 and gen 1 GCs, we cannot do back-to-back gen 1 GCs.
If we take the gen 1 budget as the amount of memory we are aiming to be used for gen 1, then it seems illogical to just drop the amount used in this GC on the floor.
The fix is to actually subtract the amount of gen 1 budget used in this GC from dd_new_allocation for gen 1. This is a bit awkward for server GC, because we even out the budget and the amount to subtract from dd_new_allocation across heaps after having computed it for each heap individually.