- 11 9月, 2022 3 次提交
-
-
由 Jan Kotas 提交于
* Flush instruction cache after thunk pool allocation Fixes #74710 * More precise ifdef
-
由 Andy Ayers 提交于
If a struct has an exposed field, we can't safely reason about is value in local assertion prop. Closes #75249.
-
由 Stephen Toub 提交于
Also consolidated some duplicated code.
-
- 10 9月, 2022 16 次提交
-
-
由 Larry Ewing 提交于
Backport the workload changes from net7 but and a versionless manifest that represents the netappcurrent so that version bumps will require fewer immediate changes in net9.0 - `Wasm.Build.Tests` installs `wasm-tools`, and `wasm-tools-net7` - and uses a sdk with both `wasm-tools`, and `wasm-tools-net7` for running, with projects built for `net7.0` - `net6.0` workload install is attempted, but fails because the `6.0.10` package isn't on public feeds yet Also, - Provision chromedriver also, alongwith chrome - [host] Track change in forwardConsole parameter to forwardConsoleToWS - Drop UnixFilePermissions.xml, as it is not needed anymore Co-authored-by: NAnkit Jain <radical@gmail.com>
-
由 Stephen Toub 提交于
-
由 Lakshan Fernando 提交于
* Disable MonitoringIsEnabled with issue 75302 * Narrow the test exclusion
-
由 Aaron Robinson 提交于
-
由 Jeremy Koritzinsky 提交于
-
由 SingleAccretion 提交于
In all cases where it was used, "lvPromoted" is a better check.
-
由 Eric Erhardt 提交于
* Make System.Transactions.Local trimmable on Windows Remove `IsTrimmable=false` from the project, so this assembly is still trimmed with `partial` trimming on Windows. Add a "LibraryBuild" ILLink warning, so when the distributed transaction code is not trimmed, the app developer gets a warning that it is not guaranteed to work. Fix #75031 * Fix x86 build. Move the ILLink suppression to a method that is completely trimmed on x86.
-
由 David Cantú 提交于
* Fix buffer too small error on prefix * Add test * Remove var usage * Remove assert * Update src/libraries/System.Formats.Tar/tests/TarWriter/TarWriter.WriteEntry.Tests.cs
-
由 Radek Zikmund 提交于
-
由 Andy Ayers 提交于
In #74384 I modified morph to generate assertions before a block op was morphed, but after the source and dest were morphed. This misses generating some assertions that arise once the expanded form is known. We now re-run assertion gen on the expanded tree, when we do an `OneAsgBlock` expansion. Other cases my also prove profitable. Closes #75229.
-
由 Filip Navara 提交于
Allow lldb to skip over null reference exceptions when debugging NativeAOT code (ported over from Mono) (#75332)
-
由 Ulrich Weigand 提交于
* Fix XmlBinaryReader/XmlBinaryWriter on big-endian systems Properly byte-swap float/double/decimal types. Handle array types correctly on big-endian systems. Added more test cases based on PR 71478. Fixes #74494
-
由 Miha Zupan 提交于
-
由 Stefan 提交于
-
由 Tomas Weinfurt 提交于
-
由 Kevin Jones 提交于
It's the same thing as the managed implementation now
-
- 09 9月, 2022 21 次提交
-
-
由 Stephen Toub 提交于
-
由 Alexander Köplinger 提交于
TypeRefererenceHandle -> TypeReferenceHandle
-
由 Kevin Jones 提交于
-
由 Stephen Toub 提交于
-
由 Stephen Toub 提交于
There's only a few of these constructed, so it's not a big deal, but might as well.
-
由 Stephen Toub 提交于
-
由 Stephen Toub 提交于
-
由 Jan Dupej 提交于
* - [mono] JIT now optimizes box+callvirt opcode sequence into a call * - [mono] fixed resolution of method in box+callvirt optimization * - [mono] more constraints on box+callvirt optimization to prevent invalid references * - [mono] JIT now optimizes box+callvirt opcode sequence into a call * - [mono] fixed resolution of method in box+callvirt optimization * - [mono] more constraints on box+callvirt optimization to prevent invalid references * - [mono] box+callvirt optimization in JIT, #32166 * - [mono] JIT fixed certain edge cases in box+callvirt optimization #32166 * - [mono] cleaned up JIT changes for box+callvirt optimization #32166 * - [mono] removed temporary objects for issue 32166 * - [mono] fixed error object usage in box+callvirt optimized handler * [mono] JIT no longer attempts to optimize box+callvirt if the callee has type parameters #32166. Fixed related code style issues. * [mono] Addressed final code style issues #32166. * [mono] box+callvirt optimization is now disabled when the resolved callee has type params #32166 * [mono][jit] Refined conditions under which box+callvirt optimization can occur. #32166 * [mono][jit] box+callvirt optimization is now disabled for LLVM AOT. #32166 * [mono][jit] box+callvirt no longer optimized in AOT when generics are involved. * Update src/mono/mono/mini/method-to-ir.c [mono] Referencing issue with a full link, instead of number. Co-authored-by: NAlexander Köplinger <alex.koeplinger@outlook.com> Co-authored-by: NAlexander Köplinger <alex.koeplinger@outlook.com>
-
由 Alexander Köplinger 提交于
The VerifyArchitecture test wasn't handling armv6 process running on arm64, only armv7. Also disabled tests that are failing on s390x/ppc64le with ActiveIssue.
-
由 Johan Lorensson 提交于
[Mono] Race in init_method when using LLVM AOT. When using LLVM AOT codegen, init_method updates two GOT slots. These slots are initialized as part of init_method, but there is a race between initialization of the two slots. Current implementation can have two threads running init_method for the same method, but as soon as: [got_slots [pindex]] = addr store is visible, it will trigger other threads to return back from init_method, but since that could happen before the corresponding LLVM AOT const slot is set, second thread will return to method calling init_method, load the LLVM aot const, and crash when trying to use it (since its still NULL). This crash is very rare but have been identified on x86/x64 CPU's, when one thread is either preempted between updating regular GOT slot and LLVM GOT slot or store into LLVM GOT slot gets delayed in store buffer. I have also been able to emulate the scenario in debugger, triggering the issue and crashing in the method loading from LLVM aot const slot. Fix change order of updates and make sure the update of LLVM aot const slot happens before memory_barrier, since: got [got_slots [pindex]] = addr; have release semantics in relation to addr and update of LLVM aot const slot. Fix also add acquire/release semantics for ji->type in init_method since it is used to guard if a thread ignores a patch or not and it should not be re-ordered with previous stores, since it can cause similar race conditions with updated slots.
-
由 Adeel Mujahid 提交于
-
由 Matous Kozak 提交于
* sqrt call * Vector128 Sqrt HW intrinsics support * Sqrt move checks before ifdef
-
由 Adam Sitnik 提交于
-
由 Tomas Weinfurt 提交于
-
由 Manish Godse 提交于
-
由 Vladimir Sadov 提交于
-
由 Zoltan Varga 提交于
-
由 Olli Saarikivi 提交于
Pass integer bound into innermost matching loop rather than slicing the input span, which was causing end anchors to improperly match before input end. Add test that fails when timeout check boundaries get treated as end-of-input.
-
由 Akhil Indurti 提交于
The DirectoryEntryConverter's implementation is in the System.DirectoryServices.Design namespace. However, the reference was mistakenly put in the System.DirectoryServices namespace. This was flagged by the new ApiCompat tool, where the type argument to the TypeConverter attribute was different (since they're in different namespaces).
-
由 Ahmet İbrahim AKSOY 提交于
-
由 Andy Ayers 提交于
In local assertion prop, optimize out stores of zeros to locations that are already known to be zero. Don't do this for integer locals as it can lead to excessive loop cloning. Update VN to handle the `struct LCL_FLD = 0` case better.
-