Changes

Summary

  1. [LangRef] Document that DataLayout defaults to little-endian (details)
  2. [SystemZ] Provide size values for PATCHPOINT, STACKMAP and FENTRY_CALL. (details)
  3. [lldb] [Communication] Add a WriteAll() method that resumes writing (details)
  4. [MLIR] Fix AffineExpr getLargestKnownDivisor for ceildiv and floordiv (details)
  5. [clangd] Add a way to enable IncludeCleaner through config (details)
  6. [lldb] Fix PlatformAppleSimulator for a458ef4f (details)
  7. [lldb] [lldb-gdbserver] Unify listen/connect code to use ConnectionFileDescriptor (details)
  8. [AMDGPU] Use standard MachineBasicBlock::getFallThrough method. NFCI. (details)
  9. [AMDGPU] Fix setcc combine for i128 (details)
  10. [lldb] [Host] Move port predicate-related logic to gdb-remote (details)
  11. [clangd] Remove tricky integration test that flakes/fails on some platforms. (details)
  12. [x86] add test for vector increment; NFC (details)
  13. [x86] use cast instead of dyn_cast for unchecked usage; NFC (details)
  14. [fir][NFC] Fix indentation for fir.alloca builders (details)
Commit c3dce37a55e1aac60ac371a596dee8f7591d77ea by fraser
[LangRef] Document that DataLayout defaults to little-endian

Little-endian has apparently been the default since 2014.

Reviewed By: nikic

Differential Revision: https://reviews.llvm.org/D112316
The file was modifiedllvm/docs/LangRef.rst
Commit 9f8872779a8556d408e5fc88d4edd3eb50b2a24e by paulsson
[SystemZ] Provide size values for PATCHPOINT, STACKMAP and FENTRY_CALL.

All instructions must have a correct size value close to emission when
SystemZLongBranch runs, or a necessary branch relaxation may be missed.

This patch also adds an assert for instruction sizes in SystemZLongBranch.

Review: Ulrich Weigand
The file was modifiedllvm/lib/Target/SystemZ/SystemZInstrInfo.cpp
The file was modifiedllvm/test/CodeGen/SystemZ/patchpoint.ll
The file was modifiedllvm/lib/Target/SystemZ/SystemZLongBranch.cpp
Commit f279e50fd0f0035e0205f4d36e7e5a8e0112fc24 by mgorny
[lldb] [Communication] Add a WriteAll() method that resumes writing

Add a Communication::WriteAll() that resumes Write() if the initial call
did not write all data.  Use it in GDBRemoteCommunication when sending
packets in order to fix handling partial writes (i.e. just resume/retry
them rather than erring out).  This fixes LLDB failures when writing
large packets to a pty.

Differential Revision: https://reviews.llvm.org/D112169
The file was modifiedlldb/unittests/Core/CommunicationTest.cpp
The file was modifiedlldb/source/Plugins/Process/gdb-remote/GDBRemoteCommunication.cpp
The file was modifiedlldb/include/lldb/Core/Communication.h
The file was modifiedlldb/source/Core/Communication.cpp
Commit 41a8b4600731b0ce702adb6087c73133089c78d5 by uday
[MLIR] Fix AffineExpr getLargestKnownDivisor for ceildiv and floordiv

Fix AffineExpr `getLargestKnownDivisor` for ceil/floor div cases.
In these cases, nothing can be inferred on the divisor of the
result.

Add test case for `mod` as well.

Differential Revision: https://reviews.llvm.org/D112523
The file was modifiedmlir/test/Dialect/Affine/unroll.mlir
The file was modifiedmlir/lib/IR/AffineExpr.cpp
Commit 6d314ee570975610ef85e8c56659f00857b49744 by kbobyrev
[clangd] Add a way to enable IncludeCleaner through config

This is useful for dogfooding the feature and spotting bugs.

Reviewed By: sammccall

Differential Revision: https://reviews.llvm.org/D111870
The file was modifiedclang-tools-extra/clangd/unittests/ConfigCompileTests.cpp
The file was modifiedclang-tools-extra/clangd/IncludeCleaner.h
The file was modifiedclang-tools-extra/clangd/Config.h
The file was modifiedclang-tools-extra/clangd/ConfigYAML.cpp
The file was modifiedclang-tools-extra/clangd/unittests/ConfigYAMLTests.cpp
The file was modifiedclang-tools-extra/clangd/ConfigFragment.h
The file was modifiedclang-tools-extra/clangd/ConfigCompile.cpp
The file was modifiedclang-tools-extra/clangd/Diagnostics.cpp
The file was modifiedclang-tools-extra/clangd/Diagnostics.h
The file was modifiedclang-tools-extra/clangd/unittests/DiagnosticsTests.cpp
The file was modifiedclang-tools-extra/clangd/IncludeCleaner.cpp
The file was modifiedclang-tools-extra/clangd/ParsedAST.cpp
Commit 93c7ed8c3f8e2b2966e565758d0992ec1b07a11f by pavel
[lldb] Fix PlatformAppleSimulator for a458ef4f
The file was modifiedlldb/source/Plugins/Platform/MacOSX/PlatformAppleSimulator.cpp
Commit 58d28b931f92f5ea2a6a01e088b794ee6ebd05e7 by mgorny
[lldb] [lldb-gdbserver] Unify listen/connect code to use ConnectionFileDescriptor

Unify the listen and connect code inside lldb-server to use
ConnectionFileDescriptor uniformly rather than a mix of it and Acceptor.
This involves:

- adding a function to map legacy values of host:port parameter
  (including legacy server URLs) into CFD-style URLs

- adding a callback to return "local socket id" (i.e. UNIX socket path
  or TCP port number) between listen() and accept() calls in CFD

- adding a "unix-abstract-accept" scheme to CFD

As an additional advantage, this permits lldb-server to accept any URL
known to CFD including the new serial:// scheme.  Effectively,
lldb-server can now listen on the serial port.  Tests for connecting
over a pty are added to test that.

Differential Revision: https://reviews.llvm.org/D111964
The file was modifiedlldb/source/Plugins/Process/gdb-remote/GDBRemoteCommunicationServerLLGS.h
The file was addedlldb/test/API/tools/lldb-server/TestPtyServer.py
The file was addedlldb/unittests/Process/gdb-remote/GDBRemoteCommunicationServerLLGSTest.cpp
The file was modifiedlldb/include/lldb/Host/posix/ConnectionFileDescriptorPosix.h
The file was modifiedlldb/source/Host/posix/ConnectionFileDescriptorPosix.cpp
The file was modifiedlldb/unittests/Process/gdb-remote/CMakeLists.txt
The file was modifiedlldb/source/Plugins/Process/gdb-remote/GDBRemoteCommunicationServerLLGS.cpp
The file was modifiedlldb/tools/lldb-server/lldb-gdbserver.cpp
Commit c8e5aef1a09d4973c43fa8dd704e3fa59418c72d by jay.foad
[AMDGPU] Use standard MachineBasicBlock::getFallThrough method. NFCI.

Differential Revision: https://reviews.llvm.org/D101825
The file was modifiedllvm/lib/Target/AMDGPU/SILowerControlFlow.cpp
Commit 487f15603e7394658423ab0a0c9afd3d51cd068d by sebastian.neubauer
[AMDGPU] Fix setcc combine for i128

The combine asserted if constants could not be represented as uint64_t.
Use APInts to fix this.

Differential Revision: https://reviews.llvm.org/D112416
The file was modifiedllvm/test/CodeGen/AMDGPU/setcc64.ll
The file was modifiedllvm/lib/Target/AMDGPU/SIISelLowering.cpp
Commit 4373f3595f8e37f6183d9880ee5b4eb59cba3852 by mgorny
[lldb] [Host] Move port predicate-related logic to gdb-remote

Remove the port predicate from Socket and ConnectionFileDescriptor,
and move it to gdb-remote.  It is specifically relevant to the threading
used inside gdb-remote and with the new port callback API, we can
reliably move it there.  While at it, switch from the custom Predicate
to std::promise/std::future.

Differential Revision: https://reviews.llvm.org/D112357
The file was modifiedlldb/unittests/Host/SocketTest.cpp
The file was modifiedlldb/source/Host/common/Socket.cpp
The file was modifiedlldb/source/Host/posix/ConnectionFileDescriptorPosix.cpp
The file was modifiedlldb/source/Plugins/Process/gdb-remote/GDBRemoteCommunication.cpp
The file was modifiedlldb/source/Plugins/Process/gdb-remote/GDBRemoteCommunication.h
The file was modifiedlldb/include/lldb/Host/Socket.h
The file was modifiedlldb/include/lldb/Host/posix/ConnectionFileDescriptorPosix.h
The file was modifiedlldb/unittests/Host/SocketTestUtilities.cpp
Commit 51be7061d025139ba66869d5d99c7157a3ae9edd by sam.mccall
[clangd] Remove tricky integration test that flakes/fails on some platforms.

This functionality is covered by unittests, will add a better
integration test soon.
The file was removedclang-tools-extra/clangd/test/crash.test
Commit bffc6574971e22a470c2114c865b29ccb9cfe3d7 by spatel
[x86] add test for vector increment; NFC
The file was modifiedllvm/test/CodeGen/X86/combine-sub.ll
Commit 2ab0148c140dc54efea111b50b1650b7e3bdbfeb by spatel
[x86] use cast instead of dyn_cast for unchecked usage; NFC

This was noted as an independent clean-up in D112464.
The file was modifiedllvm/lib/Target/X86/X86ISelLowering.cpp
Commit 1189d97d12a354376832a11eff1b3940b95be3f4 by clementval
[fir][NFC] Fix indentation for fir.alloca builders
The file was modifiedflang/include/flang/Optimizer/Dialect/FIROps.td