Fix rdar://64538073 #3099

ahatanaka · 2021-07-15T06:20:37Z

No description provided.

ahatanaka · 2021-07-15T06:29:24Z

The new patch requires users to create a source file and pass the file name to the command line. I made this change since I wasn't able to find a clean way to teach the driver to process the arguments and create the correct CC1 invocations without passing a source file.

jansvoboda11

The new patch requires users to create a source file and pass the file name to the command line. I made this change since I wasn't able to find a clean way to teach the driver to process the arguments and create the correct CC1 invocations without passing a source file.

This sounds like we're exposing an implementation detail to the clients and forcing them to work around it. Would it be possible to create the file internally in the dependency scanner? (Probably just in an in-memory VFS.)

The contents of the file don't play any role when scanning dependencies by module name, is that correct?

jansvoboda11 · 2021-07-15T08:10:40Z

clang/include/clang-c/Dependencies.h

+ * but get the dependencies by module name alone.
+ */
+CINDEX_LINKAGE CXFileDependencies *
+clang_experimental_DependencyScannerWorkerByModName_getFileDependencies_v0(


Since this now doesn't require a special worker, I think clang_experimental_DependencyScannerWorker_getModuleDependencies_v0 would be a clearer name. WDYT?

It's not clear to me whether the function name should end with getFileDependencies_v0 or getModuleDependencies_v0. The return type CXFileDependencies contains both FileDeps and ModuleDeps, so it seems to me that a name ending with getFileAndModuleDependencies_v0 would be more accurate, but the two functions declared above this one both use getFileDependencies.

jansvoboda11 · 2021-07-15T08:11:53Z

clang/include/clang-c/Dependencies.h

+clang_experimental_DependencyScannerWorkerByModName_getFileDependencies_v0(
+    CXDependencyScannerWorker Worker, int argc, const char *const *argv,
+    const char *WorkingDirectory, CXModuleDiscoveredCallback *MDC,
+    void *Context, CXString *error, const char *LookedUpModuleName);


Just ModuleName might be clearer. Maybe this should be moved after the argv parameter, since those are the main inputs to the function.

jansvoboda11 · 2021-07-15T08:14:34Z

clang/include/clang/Frontend/FrontendActions.h

@@ -45,6 +45,14 @@ class ReadPCHAndPreprocessAction : public FrontendAction {
  bool usesPreprocessorOnly() const override { return false; }
 };

+class ReadPCHAndPreprocessByModNameAction : public ReadPCHAndPreprocessAction {


Does this need to inherit from ReadPCHAndPreprocessAction? I think PreprocessorFrontendAction should be just fine, since precompiled headers don't affect dependencies of a module.

Would it be incorrect to inherit from PreprocessOnlyAction instead?

jansvoboda11 · 2021-07-15T08:22:53Z

clang/test/ClangScanDeps/modules-full-by-mod-name.cpp

+//
+// RUN: echo %t.dir > %t.result
+// RUN: clang-scan-deps -compilation-database %t.cdb -j 4 -format experimental-full \
+// RUN:   -generate-modules-path-args -module-files-dir %t.dir/custom \


The -generate-modules-path-args and -module-files-dir options are orthogonal to whether we're scanning dependencies of a module by its name or a TU by its source file. You can simplify this test by removing the code testing these.

hyp · 2021-07-24T18:50:22Z

You need to land it on next now instead of apple/main :)

collection Fixes PR51210.

This patch adds the -fminimize-whitespace with the following effects: * If combined with -E, remove as much non-line-breaking whitespace as possible. * If combined with -E -P, removes as much whitespace as possible, including line-breaks. The motivation is to reduce the amount of insignificant changes in the preprocessed output with source files where only whitespace has been changed (add/remove comments, clang-format, etc.) which is in particular useful with ccache. A patch for ccache for using this flag has been proposed to ccache as well: ccache/ccache#815, which will use -fnormalize-whitespace when clang-13 has been detected, and additionally uses -P in "unify_mode". ccache already had a unify_mode in an older version which was removed because of problems that using the preprocessor itself does not have (such that the custom tokenizer did not recognize C++11 raw strings). This patch slightly reorganizes which part is responsible for adding newlines that are required for semantics. It is now either startNewLineIfNeeded() or MoveToLine() but never both; this avoids the ShouldUpdateCurrentLine workaround and avoids redundant lines being inserted in some cases. It also fixes a mandatory newline not inserted after a _Pragma("...") that is expanded into a #pragma. Reviewed By: aaron.ballman Differential Revision: https://reviews.llvm.org/D104601

…f hasTrait<ReturnLike>. This CL adds a new RegionBranchTerminatorOpInterface to query information about operands that can be passed to successor regions. Similar to the BranchOpInterface, it allows to freely define the involved operands. However, in contrast to the BranchOpInterface, it expects an additional region number to distinguish between various use cases which might require different operands passed to different regions. Moreover, we added new utility functions (namely getMutableRegionBranchSuccessorOperands and getRegionBranchSuccessorOperands) to query (mutable) operand ranges for operations equiped with the ReturnLike trait and/or implementing the newly added interface. This simplifies reasoning about terminators in the scope of the nested regions. We also adjusted the SCF.ConditionOp to benefit from the newly added capabilities. Differential Revision: https://reviews.llvm.org/D105018

…orm. This allows ORC to execute code containing Objective-C and Swift classes and methods (provided that the language runtime is loaded into the executor).

Add 'enconding' attribute visitor. Without it ASM printer doesn't use attribute aliases for 'enconding'. Reviewed By: rriddle Differential Revision: https://reviews.llvm.org/D105554

Suppress only current warning on openmp-clang-x86_64-linux-debian Reviewed By: jdoerfert Differential Revision: https://reviews.llvm.org/D106777

Default to building the amdgpu plugin to use dlopen when hsa is not found instead of disabling it. Reviewed By: jdoerfert Differential Revision: https://reviews.llvm.org/D106600

…angler"" This reverts commit 2aa0cf1. Get rid of reference to the temporary.

This patch extends support for (scalable-vector) splats in the DAGCombiner via the `ISD::matchBinaryPredicate` function, which enable a variety of simple combines of constants. Users of this function may now have to distinguish between `BUILD_VECTOR` and `SPLAT_VECTOR` vector operands. The way of dealing with this in-tree follows the approach added for `ISD::matchUnaryPredicate` implemented in D94501. Reviewed By: craig.topper Differential Revision: https://reviews.llvm.org/D106575

I have added a new FastMathFlags parameter to getArithmeticReductionCost to indicate what type of reduction we are performing: 1. Tree-wise. This is the typical fast-math reduction that involves continually splitting a vector up into halves and adding each half together until we get a scalar result. This is the default behaviour for integers, whereas for floating point we only do this if reassociation is allowed. 2. Ordered. This now allows us to estimate the cost of performing a strict vector reduction by treating it as a series of scalar operations in lane order. This is the case when FP reassociation is not permitted. For scalable vectors this is more difficult because at compile time we do not know how many lanes there are, and so we use the worst case maximum vscale value. I have also fixed getTypeBasedIntrinsicInstrCost to pass in the FastMathFlags, which meant fixing up some X86 tests where we always assumed the vector.reduce.fadd/mul intrinsics were 'fast'. New tests have been added here: Analysis/CostModel/AArch64/reduce-fadd.ll Analysis/CostModel/AArch64/sve-intrinsics.ll Transforms/LoopVectorize/AArch64/strict-fadd-cost.ll Transforms/LoopVectorize/AArch64/sve-strict-fadd-cost.ll Differential Revision: https://reviews.llvm.org/D105432

Reviewed By: tmatheson Differential Revision: https://reviews.llvm.org/D106635

Thanks to gAlfonso-bit for the patch! Differential Revision: https://reviews.llvm.org/D106691

- `scf.yield` in the "after" region supplies new arguments to the "before" region. Differential Revision: https://reviews.llvm.org/D106806

These are required by MemIsShadow for checking if an address actually is shadow memory. Differential Revision: https://reviews.llvm.org/D105745

Change SetCurrentThread*() logic not to include the zero padding in PID/TID that was a side effect of 02ef0f5. This should fix problems caused by sending 64-bit integers to 32-bit servers. Reported by Ted Woodward. Differential Revision: https://reviews.llvm.org/D106832

See llvm.org/pr51221

This is empty for now, but we will add a check that TBI is enabled once the tagged pointer ABI for zircon is finalized. This depends on D105667. Differential Revision: https://reviews.llvm.org/D105668

Ensure that libSupport does not carry any static global initializer. libSupport can be embedded in use cases where we don't want to load all cl::opt unless we want to parse the command line. ManagedStatic can be used to enable lazy-initialization of globals. The -Werror=global-constructors is only added on platform that have support for the flag and for which std::mutex does not have a global destructor. This is ensured by having CMake trying to compile a file with a global mutex before adding the flag to libSupport.

Allegedly the DWARF backend ignores this field of DIEnumerator, but we set it nonetheless in case we decide to use it in the future. Alternatively, we could remove it, but it is simpler to pass down the signed bit as it is in the AST for now. Implemented to address comments on D106585

Before MASSV only supported P8 and P9 on AIX ans Linux . This patch proposes MASSV to add support of P7 and P10 only on AIX too. Differential: https://reviews.llvm.org/D106678

This will just be a call into __sanitizer_fill_shadow defined in the fuchsia source tree. This depends on D105663. Differential Revision: https://reviews.llvm.org/D105664

Differential revision: https://reviews.llvm.org/D106494

…list of module and file dependencies of a module for a particular compiler invocation rdar://64538073

- Define a subclass of FrontendAction and call loadModule there. - Remove the API for creating a worker just for clang_experimental_DependencyScannerWorkerByModName_getFileDependencies_v1. - Add the capability to get the dependencies based on the module name to clang-scan-deps.

- Create a fake virtual source file instead of asking users to create one and pass it to the command line. - Remove the "translation-units" output from test case modules-full-by-mod-name.cpp. The information isn't useful as the source file name isn't passed on the command line in this case. - Inherit from `PreprocessOnlyAction`.

ahatanaka · 2021-07-27T00:57:59Z

Hmm, I think rebasing onto next didn't go well. I'll open a new pull request.

ahatanaka requested review from Bigcheese, hyp and jansvoboda11 July 15, 2021 06:20

jansvoboda11 requested changes Jul 15, 2021

View reviewed changes

git apple-llvm automerger and others added 24 commits July 26, 2021 03:47

Merge commit '0d3e4d9d4d2c' from llvm.org/main into next

e87fdd9

[Object] make SourceMgr available to MCContext during inline asm symbols

1558bb8

collection Fixes PR51210.

Merge commit '1558bb80c01b' from llvm.org/main into next

39da0cb

Merge commit 'ae6b40000238' from llvm.org/main into next

e4ff089

Merge commit '0425332015f4' from llvm.org/main into next

64bc830

[ORC][ORC-RT] Add initial Objective-C and Swift support to MachOPlatf…

cdcc354

…orm. This allows ORC to execute code containing Objective-C and Swift classes and methods (provided that the language runtime is loaded into the executor).

Merge commit 'cdcc35476833' from llvm.org/main into next

4b5a48e

[libc] fix LibcUnitTestMain when building with shared libraries

47afd43

Merge commit '47afd43eaa9b' from llvm.org/main into next

dcd8d50

[mlir] Fix RankedTensorType::walkImmediateSubElements method

eb6c63c

Add 'enconding' attribute visitor. Without it ASM printer doesn't use attribute aliases for 'enconding'. Reviewed By: rriddle Differential Revision: https://reviews.llvm.org/D105554

Merge commit 'eb6c63cb0b6e' from llvm.org/main into next

ddd397a

[libomptarget][nfc] Squash unused variable warning

93fe84d

Suppress only current warning on openmp-clang-x86_64-linux-debian Reviewed By: jdoerfert Differential Revision: https://reviews.llvm.org/D106777

Merge commit '93fe84d32fea' from llvm.org/main into next

1d56ed9

[libomptarget] Build amdgpu plugin without hsa

2a613a7

Default to building the amdgpu plugin to use dlopen when hsa is not found instead of disabling it. Reviewed By: jdoerfert Differential Revision: https://reviews.llvm.org/D106600

Merge commit '2a613a779044' from llvm.org/main into next

e69f71a

Revert "Revert D106562 "[clangd] Get rid of arg adjusters in CommandM…

0a3c796

…angler"" This reverts commit 2aa0cf1. Get rid of reference to the temporary.

Merge commit '0a3c7960cba1' from llvm.org/main into next

b376c11

Merge commit 'f924a3d47492' from llvm.org/main into next

4810df4

Merge commit '0aff1798b572' from llvm.org/main into next

6297390

[AArch64][AsmParser] NFC: Parser.getTok().getLoc() -> getLoc()

e6ff917

Reviewed By: tmatheson Differential Revision: https://reviews.llvm.org/D106635

git apple-llvm automerger and others added 26 commits July 26, 2021 15:17

Merge commit '172051a1f4b1' from llvm.org/main into next

3443373

[libc++] Fix spacing in <vector>. NFCI.

1e1b570

Thanks to gAlfonso-bit for the patch! Differential Revision: https://reviews.llvm.org/D106691

Merge commit '1e1b5706c3f5' from llvm.org/main into next

f646654

[MLIR][SCF][NFC] Fix typo in documentation of scf.while

58aa388

- `scf.yield` in the "after" region supplies new arguments to the "before" region. Differential Revision: https://reviews.llvm.org/D106806

Merge commit '58aa3881bac6' from llvm.org/main into next

58854b6

[compiler-rt][hwasan][fuchsia] Define shadow bound globals

b50fb58

These are required by MemIsShadow for checking if an address actually is shadow memory. Differential Revision: https://reviews.llvm.org/D105745

Merge commit 'b50fb58695b4' from llvm.org/main into next

a4c7df2

Merge commit '3c3269559ba9' from llvm.org/main into next

f1ca31e

Disable the new enum i128 test under ASan, it uncovers an existing leak

a9b114c

See llvm.org/pr51221

Merge commit 'a9b114c5dd68' from llvm.org/main into next

579cf81

[compiler-rt][hwasan][fuchsia] Implement InitializeOsSupport

71af002

This is empty for now, but we will add a check that TBI is enabled once the tagged pointer ABI for zircon is finalized. This depends on D105667. Differential Revision: https://reviews.llvm.org/D105668

Merge commit '71af002d151e' from llvm.org/main into next

5bb9033

Merge commit 'beff86e8ff42' from llvm.org/main into next

53146d4

Merge commit 'f9f56488e02d' from llvm.org/main into next

e6a2915

[PowerPC] Add pwr7 and pwr10 support to IBM MASSV pass on AIX

45951ad

Before MASSV only supported P8 and P9 on AIX ans Linux . This patch proposes MASSV to add support of P7 and P10 only on AIX too. Differential: https://reviews.llvm.org/D106678

Merge commit '45951ad3231c' from llvm.org/main into next

9895f2e

[compiler-rt][hwasan][fuchsia] Implement TagMemoryAligned for fuchsia

b49a3bf

This will just be a call into __sanitizer_fill_shadow defined in the fuchsia source tree. This depends on D105663. Differential Revision: https://reviews.llvm.org/D105664

Merge commit 'b49a3bf7c094' from llvm.org/main into next

c530c06

[AArch64][GlobalISel] Legalize ctpop s128

97e95fe

Differential revision: https://reviews.llvm.org/D106494

Merge commit '97e95fea53fc' from llvm.org/main into next

a26b671

[clang][clang-scan-deps] Add an experimental C API for returning the …

d14aa82

…list of module and file dependencies of a module for a particular compiler invocation rdar://64538073

ahatanaka force-pushed the PR-64538073 branch from 6a4bbe6 to 6d94900 Compare July 27, 2021 00:54

ahatanaka closed this Jul 27, 2021

ahatanaka mentioned this pull request Jul 27, 2021

Add an API for clang dependency scanner to perform module lookup by name alone #3127

Merged

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Uh oh!

Fix rdar://64538073 #3099

Fix rdar://64538073 #3099

Uh oh!

ahatanaka commented Jul 15, 2021

Uh oh!

ahatanaka commented Jul 15, 2021

Uh oh!

jansvoboda11 left a comment

Uh oh!

jansvoboda11 Jul 15, 2021

Uh oh!

ahatanaka Jul 26, 2021

Uh oh!

jansvoboda11 Jul 15, 2021

Uh oh!

jansvoboda11 Jul 15, 2021

Uh oh!

ahatanaka Jul 26, 2021

Uh oh!

jansvoboda11 Jul 15, 2021

Uh oh!

hyp commented Jul 24, 2021

Uh oh!

ahatanaka commented Jul 27, 2021

Uh oh!

Uh oh!

Fix rdar://64538073 #3099

Fix rdar://64538073 #3099

Uh oh!

Conversation

ahatanaka commented Jul 15, 2021

Uh oh!

ahatanaka commented Jul 15, 2021

Uh oh!

jansvoboda11 left a comment

Choose a reason for hiding this comment

Uh oh!

jansvoboda11 Jul 15, 2021

Choose a reason for hiding this comment

Uh oh!

ahatanaka Jul 26, 2021

Choose a reason for hiding this comment

Uh oh!

jansvoboda11 Jul 15, 2021

Choose a reason for hiding this comment

Uh oh!

jansvoboda11 Jul 15, 2021

Choose a reason for hiding this comment

Uh oh!

ahatanaka Jul 26, 2021

Choose a reason for hiding this comment

Uh oh!

jansvoboda11 Jul 15, 2021

Choose a reason for hiding this comment

Uh oh!

hyp commented Jul 24, 2021

Uh oh!

ahatanaka commented Jul 27, 2021

Uh oh!

Uh oh!