aboutsummaryrefslogtreecommitdiff
path: root/llvm/utils
AgeCommit message (Collapse)AuthorFilesLines
5 hours[gn build] Manually partially port #108276 (#149595)Arthur Eubanks1-0/+1
Fixes windows gn clang build. Don't worry about exporting clang symbols on windows in gn build for now.
9 hours[gn build] Port 58be6226eb89Nico Weber1-0/+2
9 hours[gn build] Port 287b9447cc12LLVM GN Syncbot1-0/+1
10 hours[gn build] Port f78c4ce55bc4LLVM GN Syncbot1-0/+1
29 hours[utils] Add a script to update regclass numbers in MIR tests (#142761)Jay Foad1-0/+27
32 hours[gn build] Port aa7ada1dfbe2LLVM GN Syncbot2-0/+2
32 hours[gn build] Port 7355ea3f6b21LLVM GN Syncbot2-0/+2
32 hours[gn build] Port 401b5ccf6b50LLVM GN Syncbot1-0/+1
3 days[gn build] Port 6c257754d955LLVM GN Syncbot1-1/+0
3 days[gn build] Port d344c383e20aLLVM GN Syncbot1-0/+1
4 days[TableGen] Add some -time-phases support in CodeGenRegisters (#149309)Jay Foad2-1/+12
4 days[TableGen] Add `getName()` to error messages for better debugging (#149531)Shaoce SUN1-2/+4
Including the name helps quickly locate the corresponding Instruction that caused the issue.
4 days[libc++] Remove unused _LIBCPP_HAS_NO_STD_MODULES macro from __config_site ↵Louis Dionne1-1/+0
(#148902) Since 1d6b6132f, that macro isn't used anywhere anymore.
4 days[gn build] Port 8f3e78f9715cLLVM GN Syncbot1-0/+1
4 days[gn] port d994487db78 (llvm-ir2vec)Nico Weber2-0/+10
5 daysPartially Reapply "RuntimeLibcalls: Add methods to recognize libcall names ↵Matt Arsenault1-1/+12
(#149001)" This partially reverts commit a96121089b9c94e08c6632f91f2dffc73c0ffa28. Drop the IRSymtab changes for now
5 daysRevert "RuntimeLibcalls: Add methods to recognize libcall names (#149001)"Nikita Popov1-12/+1
This reverts commit 45477add8dfe9851605697bd908b49f0ec244625. This causes a significant LTO compile-time regression.
5 daysRuntimeLibcalls: Add methods to recognize libcall names (#149001)Matt Arsenault1-1/+12
Also replace the current static DenseMap of preserved symbol names in the Symtab hack with this. That was broken statefulness across compiles, so this at least fixes that. However this is still broken, llvm-as shouldn't really depend on the triple.
5 daysUse Parallel xz for test-suite sources. (#149389)Tobias Hieta1-1/+1
5 days[utils][TableGen] Make some non-bitmask enums iterable (#148647)Krzysztof Parzyszek1-2/+20
Additionally, add sentinel values <Enum>::First_ and <Enum>::Last_ to each one of those enums. This will allow using `enum_seq_inclusive` to generate the list of enum-typed values of any generated scoped (non-bitmask) enum.
6 days[Debugify] Add 'acceptance-test' mode for the debugify report script (#147574)Stephen Tozer1-33/+121
For the purposes of setting up CI that makes use of debugify, this patch adds an alternative mode for the llvm-original-di-preservation.py script, which produces terminal-friendly(-ish) YAML output instead of an HTML report, and sets the return code to 1 if the input file contains errors, or 0 if the input file contains no errors or does not exist, making it simple to use it in CI. This introduces a small change in existing usage, in that the path for the HTML report file is now passed with `--report-file <path>` rather than as a positional argument; I could make the argparse logic work without this change, but I believe that is simpler to understand this way, and to my knowledge debugify isn't currently being used in automated environments where changing this might cause issues. As a small change while passing by, I also changed `-compress` to `--compress`, for consistency. As a note for reviewers, the reason that we treat a non-existent input file as a pass is that this is actually the expected state: we use clang to compile numerous files, passing a filepath for debugify errors. Any errors found by debugify will be written to this file; if none are found, the file is untouched. This is also mentioned in a code comment, but I think it useful to state upfront. Finally, the justification for adding a new mode to this script instead of adding a separate script for the separate functionality is that this script understands debugify's output, and performs some deduplication that is useful for clarifying the resulting output. Writing a new script would require duplicating logic unnecessarily, and risks the scripts falling out-of-sync if changes are made to debugify's output.
6 days[gn build] Port 2194bca2b78cLLVM GN Syncbot1-0/+1
6 days[TableGen] Add a bitvector of members of CodeGenRegisterClass (#149122)Jay Foad3-18/+15
This makes CodeGenRegisterClass::contains fast. Use this to simplify inferMatchingSuperRegClass.
6 days[TableGen] Remove unnecessary sortAndUniqueRegisters (#149125)Jay Foad1-3/+0
Each of the SRSets is already sorted and unique because it is a filtered version of RC->getMembers() which is already sorted and unique.
6 days[lld-macho] Move Linker Optimization Hints pass to a separate fileDaniel Bertalan1-0/+1
Moving it away from the arm64 `TargetInfo` class will let us enable it more easily for arm64_32 and the soon-to-be-added arm64e target as well. This is the NFC part of #148964
6 days[Clang] Reintroduce obsolete libclang symbols to avoid an ABI break (#149079)Sirraide1-0/+1
For more context, see https://github.com/llvm/llvm-project/pull/119269#issuecomment-3075444493, but briefly, when removing ARCMigrate, I also removed some symbols in libclang, which constitutes an ABI break that we don’t want, so this pr reintroduces the removed symbols; the declarations are marked as deprecated for future removal, and the implementations print an error and do nothing, which is what we used to do when ARCMigrate was disabled.
6 days[gn build] Port 88a498c3b110LLVM GN Syncbot1-1/+1
6 daysRemove workarounds for NDK versions before 2017's r16. (#148879)enh-google1-1/+0
7 days[gn build] Port 5ae49393957cLLVM GN Syncbot1-0/+1
7 days[gn build] Port 3e4153c97b54LLVM GN Syncbot1-0/+1
7 days[gn build] Port 7f2bcd97cf4aLLVM GN Syncbot1-0/+1
7 days[gn build] Port eec98bef84b8LLVM GN Syncbot1-0/+1
7 days[gn build] Port ee9b84f5bed7LLVM GN Syncbot1-0/+1
7 days[gn build] Port 0d2e11f3e834LLVM GN Syncbot4-5/+1
7 days[gn] port 5951c44573Nico Weber1-0/+2
7 daysTableGen: Use StringOffsetTable for RuntimeLibcall names (#148839)Matt Arsenault1-10/+17
7 daysBump version to 22.0.0-gitllvmorg-22-initTobias Hieta3-3/+3
8 daysremote-exec: Detect and propagate signal death in the remote process.Peter Collingbourne1-1/+10
If the remote process died with a signal, this will be exposed by ssh as an exit code in the range 128 < rc < 160. We may be running under `not --crash` which will expect us to also die with a signal, so send the signal to ourselves so that wait4() in `not` will detect the signal. Speculative fix for failing buildbot: https://lab.llvm.org/buildbot/#/builders/193/builds/9070
8 days[RISCV] Relax destination instruction dag operand matching in ↵Sudharsan Veeravalli1-7/+3
CompresInstEmitter (#148660) We have some 48-bit instructions in the `Xqci` spec that currently cannot be compressed to their 32-bit variants due to the constraint in `CompressInstEmitter` on destination instruction operands not being allowed to mismatch with the DAG operands. For eg. the` QC_E_ADDI` instruction can be compressed to the `ADDI` instruction when the immediate is signed-12 bit but this is currently not possible since the `QC_E_ADDI` instruction has `GPRNoX0` register operands while the `ADDI` instruction has `GPR` register operands leading to an operand type validation error. I think we can remove the check that only source instruction operands can mismatch with the corresponding DAG operands and rely on the fact that we check if the DAG register operand type is a subclass of the instruction register operand type.
8 days[NFC][TableGen][DecoderEmitter] Eliminate `indent` for a few functions (#148718)Rahul Joshi1-71/+55
Eliminate the `indent` argument for functions which are always called with `indent(0)`.
8 daysReapply "compiler-rt: Introduce runtime functions for emulated PAC."Peter Collingbourne3-0/+4
This reverts commit 0c0aa56cdcf1fe3970a5f3875db412530512fc07. This time with the following fixes for buildbot failures: - Add underscore prefixes to symbol names on Apple platforms. - Modify the test so that it skips the crash tests on platforms where they are not expected to pass: - Platforms that implement FEAT_PAuth but not FEAT_FPAC (e.g. Apple M1, Cortex-A78C) - Platforms where DA key is disabled (e.g. older Linux kernels, Linux kernels with PAC disabled, likely Windows) Original commit message follows: The emulated PAC runtime functions emulate the ARMv8.3a pointer authentication instructions and are intended for use in heterogeneous testing environments. For more information, see the associated RFC: https://discourse.llvm.org/t/rfc-emulated-pac/85557 Reviewers: mstorsjo, pawosm-arm, atrosinenko Reviewed By: atrosinenko Pull Request: https://github.com/llvm/llvm-project/pull/148094
8 days[gn build] Port b9ccc0ca38dcLLVM GN Syncbot1-1/+4
9 days[llvm][release] Mark gh commands in package verify exampleDavid Spickett1-2/+2
9 days[gn build] Port 59b39c0031edLLVM GN Syncbot1-0/+1
11 days[gn build] Port 2f4a804e6eb5LLVM GN Syncbot1-0/+1
11 days[gn build] Port bf94c8ddb321LLVM GN Syncbot1-0/+1
12 days[gn build] Port def7bbbe17c6LLVM GN Syncbot1-0/+1
12 days[gn build] Port 3076794e924fLLVM GN Syncbot1-0/+1
12 days[NFC][TableGen] Minor code cleanup in SearchableTableEmitter (#147856)Rahul Joshi1-71/+87
- Add braces around if/else bodies per LLVM coding standards. - Use range for loops and structured bindings. - use auto for variables initialized with `dyn_cast`. - Refactor `compareBy` to also use early return in the comparison loop by extracting the comparison into lambdas.
12 daysRevert "compiler-rt: Introduce runtime functions for emulated PAC."Peter Collingbourne3-4/+0
As well as followup "builtins: Speculative MSVC fix." This reverts commits 5b1db59fb87b4146f827d17396f54ef30ae0dc40 and f1c4df5b7bb79efb3e9be7fa5f8176506499d0a6. Needs fixes for failing tests which will take time to implement.
generated by cgit v1.1 at 2025-07-22 22:43:46 +0000