LLVM and SPIRV-LLVM-Translator pulldown (WW29) #6420
Merged
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This can now be used with -static-libgcc. Differential Revision: https://reviews.llvm.org/D129218
…rocessing. This mostly affects PGO/LTO builds which use module flags describing the call graph. Fixes Issue #51893. Reviewed By: dexonsmith Differential Revision: https://reviews.llvm.org/D125999
Create a TargetCharacteristics class to centralize the few items of target specific information that are relevant to semantics. Use the new class for all target queries, including derived type component layout modeling. Future work will initialize this class with target information provided or forwarded by the drivers, and use it to fold layout-dependent intrinsic functions like TRANSFER(). Differential Revision: https://reviews.llvm.org/D129018 Updates: Attempts to work around build issues on Windows.
Reverting while I investigate build failures This reverts commit e3dc568.
Compiler can generate calls to some functions implicitly, even under constraints of freestanding environment. Make sure these functions are available in our runtime objects. Fixes test failures on some systems after https://reviews.llvm.org/D128960. Reviewed By: yota9 Differential Revision: https://reviews.llvm.org/D129168
`dxil` is an architecture supported by the DirectX backend. These intrinsics will likely be shared with other DirectX architectures like `dxbc`. Using a common prefix `dx` will make it more intuitive. Also the `dx` prefix is already set in the Triple, which causes intrinsics described here to be unmatchable via the ClangBuiltin mechanism.
Adds optional dependency on cpp-httplib, a lightweight header-only HTTP server. Reviewed By: phosek Differential Revision: https://reviews.llvm.org/D113218
Proposing to move the check for scalar MASS conversion from constructor of PPCTargetLowering to the lowerLibCallBase function which decides about the lowering. The Target machine option Options.PPCGenScalarMASSEntries is set in PPCTargetMachine.cpp. But an object of the class PPCTargetLowering is created in one of the included header files. So, the constructor will run before setting PPCGenScalarMASSEntries to correct value. So, we cannot check this option in the constructor. Differential: https://reviews.llvm.org/D128653 Reviewer: @bmahjour
The number of spaces between `#` and `pragma` can differ due to different indention levels in the preprocessor directives. Therefore allow any number of spaces. The test used to put an exclamation mark in its diagnostic. This adds little benefit and only makes it harder to copy the offending filename. As drive-by this exclamation mark has been removed. Reviewed By: ldionne, #libc Differential Revision: https://reviews.llvm.org/D129047
This provides a minimal HTTP server interface and an implementation wrapping [[ https://github.com/yhirose/cpp-httplib | cpp-httplib ]] in the Debuginfod library. If the Curl HTTP client is available (D112753) the server is tested by pinging it with the client. Reviewed By: dblaikie Differential Revision: https://reviews.llvm.org/D114415
After checking the libc++abi.dylib shipped in macOS 10.13, I can confirm that it contains the align_val_t variants of operator new and operator delete. However, the libc++abi.dylib shipped on macOS 10.12 does not. Differential Revision: https://reviews.llvm.org/D129198
Introduce an off-by default `-Winvalid-utf8` warning that detects invalid UTF-8 code units sequences in comments. Invalid UTF-8 in other places is already diagnosed, as that cannot appear in identifiers and other grammar constructs. The warning is off by default as its likely to be somewhat disruptive otherwise. This warning allows clang to conform to the yet-to be approved WG21 "P2295R5 Support for UTF-8 as a portable source file encoding" paper. Reviewed By: aaron.ballman, #clang-language-wg Differential Revision: https://reviews.llvm.org/D128059
…nsulation Instead of marking private symbols with internal_linkage (which leads to one copy per translation unit -- rather wasteful), use an ABI tag that gets rev'd with each libc++ version. That way, we know that we can't have name collisions between implementation-detail functions across libc++ versions, so we'll never violate the ODR. However, within a single program, each symbol still has a proper name with external linkage, which means that the linker is free to deduplicate symbols even across TUs. This actually means that we can guarantee that versions of libc++ can be mixed within the same program without ever having to take a code size hit, and without having to manually opt-in -- it should just work out of the box. Differential Revision: https://reviews.llvm.org/D127444
Vectors are defined to be tightly packed, regardless of the element type. The AsmPrinter didn't realize this, and was allocating extra padding. Fixes llvm/llvm-project#49286 Fixes llvm/llvm-project#53246 Fixes llvm/llvm-project#55522 Differential Revision: https://reviews.llvm.org/D129164
…cking local debuginfo. This library implements the class `DebuginfodCollection`, which scans a set of directories for binaries, classifying them according to whether they contain debuginfo. This also provides the `DebuginfodServer`, an `HTTPServer` which serves debuginfod's `/debuginfo` and `/executable` endpoints. This is intended as the final new supporting library required for `llvm-debuginfod`. As implemented here, `DebuginfodCollection` only finds ELF binaries and DWARF debuginfo. All other files are ignored. However, the class interface is format-agnostic. Generalizing to support other platforms will require refactoring of LLVM's object parsing libraries to eliminate use of `report_fatal_error` ([[ https://github.com/llvm/llvm-project/blob/main/llvm/lib/Object/WasmObjectFile.cpp#L74 | e.g. when reading WASM files ]]), so that the debuginfod daemon does not crash when it encounters a malformed file on the disk. The `DebuginfodCollection` is tested by end-to-end tests of the debuginfod server (D114846). Reviewed By: mysterymath Differential Revision: https://reviews.llvm.org/D114845
The length of valid codepoints was incorrectly calculated which was not caught before because the absence of tests for the valid codepoints scenario. Differential Revision: https://reviews.llvm.org/D129223
This adds special-case treatment for parameter packs in make_unique-like functions to forward parameter names to inlay hints. The parameter packs are being resolved recursively by traversing the function body of forwarding functions looking for expressions matching the (std::forwarded) parameters expanded from a pack. The implementation checks whether parameters are being passed by (rvalue) reference or value and adds reference inlay hints accordingly. The traversal has a limited recursion stack depth, and recursive calls like std::make_tuple are cut off to avoid hinting duplicate parameter names. Reviewed By: sammccall Differential Revision: https://reviews.llvm.org/D124690
This reverts commit 4174f0c. Also revert follow-up "[Clang] Fix invalid utf-8 detection" This reverts commit bf45e27. The second commit broke tests, see comments on https://reviews.llvm.org/D129223, and it sounds like the first commit isn't valid without the second one. So reverting both for now.
Silence warning from MSVC when handling ##__VA_ARGS Reviewed By: aartbik Differential Revision: https://reviews.llvm.org/D129227
Reviewed By: ktras Differential Revision: https://reviews.llvm.org/D128957
…el#1512) This change is incomplete, and will not properly convert every single LLVM IR file to SPIR-V. However, it will convert many files. Areas that are known to require more investigation: * ptr-valued extractvalue, especially from cmpxchg instructions * return values of functions * llvm.*.annotation support * function pointer extensions * pointer types in structs Original commit: KhronosGroup/SPIRV-LLVM-Translator@21fee92
Substitution rules in Itanium mangled names are unfortunately more complex and harder to parse than they should be, and there were two bugs that cropped in the current implementation (as found by running c++filt on the resulting names and discovering that they didn't demangle correctly). The first bug was that unqualified names double-registered the underlying pointer type. A mangled pointer type like PU3AS419__spirv_DeviceEvent should generate three substitutable names: * 19__spirv_DeviceEvent [a user-defined type gets an entry] * U3AS419__spirv_DeviceEvent [qualifiers of a pointer get an entry] * PU3AS419__spirv_DeviceEvent [the pointer as a whole gets an entry] However, if there's no qualifiers--we get a P19__spirv_DeviceEvent instead-- the original code was still adding multiple entries as follows: * 19__spirv_DeviceEvent [a user-defined type gets an entry] * 19__spirv_DeviceEvent [qualifiers of a pointer, but there are none!] * P19__spirv_DeviceEvent [the pointer as a whole gets an entry] The second bug is in the representation of the block pointer type for kernels: this is mangled as U13block_pointerFvvE. This type actually gets two entries: * FvvE [void(void) function type] * U13block_pointerFvvE [block_pointer-qualified function type] We weren't adding any entries for either of these types, causing later type entries to be offset by two. With this patch, these type entries *still* aren't added, so the mangling is incorrect if multiple block types are present in a function declaration. Since such functions don't exist for OpenCL/SPIR-V, it's okay that these cases are not correctly handled. Original commit: KhronosGroup/SPIRV-LLVM-Translator@9b280f4
* Implement support for @llvm.sadd.sat intrinsic. * Rename LowerSaddWithOverflow to LowerSaddIntrinsics. Original commit: KhronosGroup/SPIRV-LLVM-Translator@0af7e64
Emit (in OCLToSPIRV) and decode (in SPIRVToOCL) the SignExtend and ZeroExtend Image Operands introduced by SPIR-V version 1.4 to preserve signedness of image read and write operations. This allows distinguishing between e.g. `read_imagei` and `read_imageui` and enables a lossless LLVM -> SPIR-V -> LLVM conversion for those OpenCL builtins. Before SPIR-V 1.4, there was no way to represent the signedness of the image operations and llvm-spirv defaulted to signed. This commit leaves that behaviour unchanged if llvm-spirv's maximum version is set to SPIR-V 1.3 or below. Add the `--spirv-max-version=1.3` flag to some tests that rely on the old behaviour. Original commit: KhronosGroup/SPIRV-LLVM-Translator@bb4ead0
dwoodwor-intel
added
the
disable-lint
This should resolve the issues with: https://lab.llvm.org/buildbot/#/builders/92/builds/29439
|
It looks like the sphinx build failure has been fixed upstream; I've cherry-picked that commit. |
This reverts commit 4256e4e. This change enabled a test that is failing on Windows, so I'm leaving it out of this week's pulldown.
|
A couple more test fixes:
|
|
I'm seeing some test failures that look like they're caused by intel/llvm-test-suite#1076 — they should be fixed by #6400 so I've merged that in |
|
@intel/llvm-gatekeepers, it looks like all of the (non-lint) tests are passing now, except that ProtexIP step crashed/was terminated after running for 20 hours. Is that expected for pulldowns? Should I go ahead with the merge? |
No, it's not expected. I expect this job to work for any PR. If you are 100% sure that the changes you did on top of pulled ones are fine, you can merge it. @dwoodwor-intel, do you track the llvm-spirv changes made here? be9bfe6 reverts on of the commits. Will you merge it with next pulldown or cherry-pick separately? |
Those should be fine; I'll go ahead with the merge.
Yes, I've updated the commit hash in the message above to not include the reverted commit and my plan is to merge it with the next pulldown instead. |
|
/merge |
bb-sycl
approved these changes
Jul 12, 2022
|
Tue 12 Jul 2022 05:29:40 PM UTC --- Start to merge the commit into sycl branch. It will take several minutes. |
|
Tue 12 Jul 2022 06:00:24 PM UTC --- Merge the branch in this PR to base automatically. Will close the PR later. |
|
I hunted down the Windows bug in |
|
Thanks a lot! |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
LLVM: llvm/llvm-project@132d711
SPIRV-LLVM-Translator: KhronosGroup/SPIRV-LLVM-Translator@18d1f41