[MLIR]-tosa-optional-decompositions triggers Assertion Failure `isa<To>(Val) && "cast<Ty>() argument of incompatible type!"'

[xisang0]

Test on commir: 6548b63
steps to reproduce:

mlir-opt test.mlir -tosa-optional-decompositions

test case:

module {
  func.func @test_build_mult_and_shift(%arg0: tensor<1x32x32x8x!quant.uniform<i8:f32, 0.015684768557548523>>, %arg1: tensor<16x1x1x8x!quant.uniform<i8<-127:127>:f32, 0.015680249780416489>>, %arg2: tensor<16xi32>) -> tensor<1x32x32x16x!quant.uniform<i8:f32, 0.078431375324726104>> {
    %0 = "tosa.conv2d"(%arg0, %arg1, %arg2) {pad = array<i64: 1, 1, 2, 2>, dilation = array<i64: 2, 1>, stride = array<i64: 1, 1>, quantization_info = #tosa.conv_quant<input_zp = -1, weight_zp = 0>} : (tensor<1x32x32x8x!quant.uniform<i8:f32, 0.015684768557548523>>, tensor<16x1x1x8x!quant.uniform<i8<-127:127>:f32, 0.015680249780416489>>, tensor<16xi32>) -> tensor<1x32x32x16x!quant.uniform<i8:f32, 0.078431375324726104>>
    return %0 : tensor<1x32x32x16x!quant.uniform<i8:f32, 0.078431375324726104>>
  }
}

crash trace:

mlir-opt: /home/workdir/llvm-project-19/llvm/include/llvm/Support/Casting.h:572: decltype(auto) llvm::cast(From &) [To = mlir::IntegerType, From = mlir::Type]: Assertion `isa<To>(Val) && "cast<Ty>() argument of incompatible type!"' failed.
PLEASE submit a bug report to https://github.com/llvm/llvm-project/issues/ and include the crash backtrace.
Stack dump:
0.      Program arguments: ./mlir-opt /home/workdir/test.mlir -tosa-optional-decompositions
 #0 0x00005a1ed4fa23b8 llvm::sys::PrintStackTrace(llvm::raw_ostream&, int) (./mlir-opt+0x10723b8)
 #1 0x00005a1ed4f9fede llvm::sys::RunSignalHandlers() (./mlir-opt+0x106fede)
 #2 0x00005a1ed4fa2dc8 SignalHandler(int) Signals.cpp:0:0
 #3 0x00007441f106d520 (/lib/x86_64-linux-gnu/libc.so.6+0x42520)
 #4 0x00007441f10c19fc pthread_kill (/lib/x86_64-linux-gnu/libc.so.6+0x969fc)
 #5 0x00007441f106d476 gsignal (/lib/x86_64-linux-gnu/libc.so.6+0x42476)
 #6 0x00007441f10537f3 abort (/lib/x86_64-linux-gnu/libc.so.6+0x287f3)
 #7 0x00007441f105371b (/lib/x86_64-linux-gnu/libc.so.6+0x2871b)
 #8 0x00007441f1064e96 (/lib/x86_64-linux-gnu/libc.so.6+0x39e96)
 #9 0x00005a1ed753916c (./mlir-opt+0x360916c)
#10 0x00005a1ed758a914 (anonymous namespace)::Conv2DIsFullyConnected::matchAndRewrite(mlir::tosa::Conv2DOp, mlir::PatternRewriter&) const TosaDecomposeConv2D.cpp:0:0
#11 0x00005a1eda28718a void llvm::function_ref<void ()>::callback_fn<mlir::PatternApplicator::matchAndRewrite(mlir::Operation*, mlir::PatternRewriter&, llvm::function_ref<bool (mlir::Pattern const&)>, llvm::function_ref<void (mlir::Pattern const&)>, llvm::function_ref<llvm::LogicalResult (mlir::Pattern const&)>)::$_0>(long) PatternApplicator.cpp:0:0
#12 0x00005a1eda283dcf mlir::PatternApplicator::matchAndRewrite(mlir::Operation*, mlir::PatternRewriter&, llvm::function_ref<bool (mlir::Pattern const&)>, llvm::function_ref<void (mlir::Pattern const&)>, llvm::function_ref<llvm::LogicalResult (mlir::Pattern const&)>) (./mlir-opt+0x6353dcf)
#13 0x00005a1ed82976de (anonymous namespace)::GreedyPatternRewriteDriver::processWorklist() GreedyPatternRewriteDriver.cpp:0:0
#14 0x00005a1ed829417b mlir::applyPatternsAndFoldGreedily(mlir::Region&, mlir::FrozenRewritePatternSet const&, mlir::GreedyRewriteConfig, bool*) (./mlir-opt+0x436417b)
#15 0x00005a1ed756c1cb (anonymous namespace)::TosaOptionalDecompositions::runOnOperation() TosaOptionalDecompositions.cpp:0:0
#16 0x00005a1ed821e6df mlir::detail::OpToOpPassAdaptor::run(mlir::Pass*, mlir::Operation*, mlir::AnalysisManager, bool, unsigned int) (./mlir-opt+0x42ee6df)
#17 0x00005a1ed821eec2 mlir::detail::OpToOpPassAdaptor::runPipeline(mlir::OpPassManager&, mlir::Operation*, mlir::AnalysisManager, bool, unsigned int, mlir::PassInstrumentor*, mlir::PassInstrumentation::PipelineParentInfo const*) (./mlir-opt+0x42eeec2)
#18 0x00005a1ed82245de auto void mlir::parallelForEach<__gnu_cxx::__normal_iterator<mlir::detail::OpToOpPassAdaptor::runOnOperationAsyncImpl(bool)::OpPMInfo*, std::vector<mlir::detail::OpToOpPassAdaptor::runOnOperationAsyncImpl(bool)::OpPMInfo, std::allocator<mlir::detail::OpToOpPassAdaptor::runOnOperationAsyncImpl(bool)::OpPMInfo>>>, mlir::detail::OpToOpPassAdaptor::runOnOperationAsyncImpl(bool)::$_0>(mlir::MLIRContext*, __gnu_cxx::__normal_iterator<mlir::detail::OpToOpPassAdaptor::runOnOperationAsyncImpl(bool)::OpPMInfo*, std::vector<mlir::detail::OpToOpPassAdaptor::runOnOperationAsyncImpl(bool)::OpPMInfo, std::allocator<mlir::detail::OpToOpPassAdaptor::runOnOperationAsyncImpl(bool)::OpPMInfo>>>, __gnu_cxx::__normal_iterator<mlir::detail::OpToOpPassAdaptor::runOnOperationAsyncImpl(bool)::OpPMInfo*, std::vector<mlir::detail::OpToOpPassAdaptor::runOnOperationAsyncImpl(bool)::OpPMInfo, std::allocator<mlir::detail::OpToOpPassAdaptor::runOnOperationAsyncImpl(bool)::OpPMInfo>>>, mlir::detail::OpToOpPassAdaptor::runOnOperationAsyncImpl(bool)::$_0&&)::'lambda'(__gnu_cxx::__normal_iterator<mlir::detail::OpToOpPassAdaptor::runOnOperationAsyncImpl(bool)::OpPMInfo*, std::vector<mlir::detail::OpToOpPassAdaptor::runOnOperationAsyncImpl(bool)::OpPMInfo, std::allocator<mlir::detail::OpToOpPassAdaptor::runOnOperationAsyncImpl(bool)::OpPMInfo>>>&&)::operator()<mlir::detail::OpToOpPassAdaptor::runOnOperationAsyncImpl(bool)::OpPMInfo&>(__gnu_cxx::__normal_iterator<mlir::detail::OpToOpPassAdaptor::runOnOperationAsyncImpl(bool)::OpPMInfo*, std::vector<mlir::detail::OpToOpPassAdaptor::runOnOperationAsyncImpl(bool)::OpPMInfo, std::allocator<mlir::detail::OpToOpPassAdaptor::runOnOperationAsyncImpl(bool)::OpPMInfo>>>&&) const Pass.cpp:0:0
#19 0x00005a1ed82206cb mlir::detail::OpToOpPassAdaptor::runOnOperationAsyncImpl(bool) (./mlir-opt+0x42f06cb)
#20 0x00005a1ed821e83c mlir::detail::OpToOpPassAdaptor::run(mlir::Pass*, mlir::Operation*, mlir::AnalysisManager, bool, unsigned int) (./mlir-opt+0x42ee83c)
#21 0x00005a1ed821eec2 mlir::detail::OpToOpPassAdaptor::runPipeline(mlir::OpPassManager&, mlir::Operation*, mlir::AnalysisManager, bool, unsigned int, mlir::PassInstrumentor*, mlir::PassInstrumentation::PipelineParentInfo const*) (./mlir-opt+0x42eeec2)
#22 0x00005a1ed822166e mlir::PassManager::run(mlir::Operation*) (./mlir-opt+0x42f166e)
#23 0x00005a1ed821a192 performActions(llvm::raw_ostream&, std::shared_ptr<llvm::SourceMgr> const&, mlir::MLIRContext*, mlir::MlirOptMainConfig const&) MlirOptMain.cpp:0:0
#24 0x00005a1ed8219dfb llvm::LogicalResult llvm::function_ref<llvm::LogicalResult (std::unique_ptr<llvm::MemoryBuffer, std::default_delete<llvm::MemoryBuffer>>, llvm::raw_ostream&)>::callback_fn<mlir::MlirOptMain(llvm::raw_ostream&, std::unique_ptr<llvm::MemoryBuffer, std::default_delete<llvm::MemoryBuffer>>, mlir::DialectRegistry&, mlir::MlirOptMainConfig const&)::$_0>(long, std::unique_ptr<llvm::MemoryBuffer, std::default_delete<llvm::MemoryBuffer>>, llvm::raw_ostream&) MlirOptMain.cpp:0:0
#25 0x00005a1ed82c3ae5 mlir::splitAndProcessBuffer(std::unique_ptr<llvm::MemoryBuffer, std::default_delete<llvm::MemoryBuffer>>, llvm::function_ref<llvm::LogicalResult (std::unique_ptr<llvm::MemoryBuffer, std::default_delete<llvm::MemoryBuffer>>, llvm::raw_ostream&)>, llvm::raw_ostream&, llvm::StringRef, llvm::StringRef) (./mlir-opt+0x4393ae5)
#26 0x00005a1ed8214ff2 mlir::MlirOptMain(llvm::raw_ostream&, std::unique_ptr<llvm::MemoryBuffer, std::default_delete<llvm::MemoryBuffer>>, mlir::DialectRegistry&, mlir::MlirOptMainConfig const&) (./mlir-opt+0x42e4ff2)
#27 0x00005a1ed82152a3 mlir::MlirOptMain(int, char**, llvm::StringRef, llvm::StringRef, mlir::DialectRegistry&) (./mlir-opt+0x42e52a3)
#28 0x00005a1ed82154b2 mlir::MlirOptMain(int, char**, llvm::StringRef, mlir::DialectRegistry&) (./mlir-opt+0x42e54b2)
#29 0x00005a1ed4f815d7 main (./mlir-opt+0x10515d7)
#30 0x00007441f1054d90 (/lib/x86_64-linux-gnu/libc.so.6+0x29d90)
#31 0x00007441f1054e40 __libc_start_main (/lib/x86_64-linux-gnu/libc.so.6+0x29e40)
#32 0x00005a1ed4f81145 _start (./mlir-opt+0x1051145)
Aborted (core dumped)

[xisang0]

This issue has been fixed already.

[CoTinker]

Fixed by #126152.