LLVM ERROR: pattern '(anonymous namespace)::AdjustCallingConventionForReturn' does not support 1:N conversion

[AnaT246]

There's an error with this code
torch-mlir-opt -pass-pipeline='builtin.module(torchscript-module-to-torch-backend-pipeline)' ~/tmp/torch_raw.mlir | tee ~/tmp/torch.mlir
This is the torch_raw.mlir file:

`module` attributes {torch.debug_module_name = "SimpleModule"} {
  func.func private @__torch__.SimpleModule.forward(%arg0: !torch.nn.Module<"__torch__.SimpleModule">, %arg1: !torch.tensor {torch.type_bound = !torch.vtensor<[3,4,4],f32>}) -> !torch.tuple<tensor, tensor> {
    %sign, %logabsdet = torch.aten.linalg_slogdet %arg1 : !torch.tensor -> !torch.tensor, !torch.tensor
    %1 = torch.prim.TupleConstruct %sign, %logabsdet : !torch.tensor, !torch.tensor -> !torch.tuple<tensor, tensor>
    return %1 : !torch.tuple<tensor, tensor>
  }
  torch.class_type @__torch__.SimpleModule {
    torch.attr private "training" : !torch.bool
    torch.attr private "_is_full_backward_hook" : !torch.optional<bool>
    torch.method "forward", @__torch__.SimpleModule.forward
  }
  %false = torch.constant.bool false
  %none = torch.constant.none
  %0 = torch.nn_module {
    torch.slot "training", %false : !torch.bool
    torch.slot "_is_full_backward_hook", %none : !torch.none
  } : !torch.nn.Module<"__torch__.SimpleModule">
}

And this is the error I get:

LLVM ERROR: pattern '(anonymous namespace)::AdjustCallingConventionForReturn' does not support 1:N conversion
PLEASE submit a bug report to https://github.com/llvm/llvm-project/issues/ and include the crash backtrace.
Stack dump:
0.      Program arguments: torch-mlir-opt -pass-pipeline=builtin.module(torchscript-module-to-torch-backend-pipeline) /home/ana_todorovic/tmp/torch_raw.mlir
 #0 0x000055fc16d93928 llvm::sys::PrintStackTrace(llvm::raw_ostream&, int) (/home/ana_todorovic/torch-mlir/build/bin/torch-mlir-opt+0x28eb928)
 #1 0x000055fc16d9156e llvm::sys::RunSignalHandlers() (/home/ana_todorovic/torch-mlir/build/bin/torch-mlir-opt+0x28e956e)
 #2 0x000055fc16d93fb8 SignalHandler(int) Signals.cpp:0:0
 #3 0x00007f3d8a386330 (/lib/x86_64-linux-gnu/libc.so.6+0x45330)
 #4 0x00007f3d8a3dfb2c pthread_kill (/lib/x86_64-linux-gnu/libc.so.6+0x9eb2c)
 #5 0x00007f3d8a38627e raise (/lib/x86_64-linux-gnu/libc.so.6+0x4527e)
 #6 0x00007f3d8a3698ff abort (/lib/x86_64-linux-gnu/libc.so.6+0x288ff)
 #7 0x000055fc16d5d82c llvm::report_fatal_error(llvm::Twine const&, bool) (/home/ana_todorovic/torch-mlir/build/bin/torch-mlir-opt+0x28b582c)
 #8 0x000055fc167384b7 (/home/ana_todorovic/torch-mlir/build/bin/torch-mlir-opt+0x22904b7)
 #9 0x000055fc154f839e mlir::OpConversionPattern<mlir::func::ReturnOp>::matchAndRewrite(mlir::func::ReturnOp, mlir::func::ReturnOpGenericAdaptor<llvm::ArrayRef<mlir::ValueRange>>, mlir::ConversionPatternRewriter&) const AdjustCallingConventions.cpp:0:0
#10 0x000055fc154f7b25 mlir::OpConversionPattern<mlir::func::ReturnOp>::matchAndRewrite(mlir::Operation*, llvm::ArrayRef<mlir::ValueRange>, mlir::ConversionPatternRewriter&) const AdjustCallingConventions.cpp:0:0
#11 0x000055fc16738656 mlir::ConversionPattern::matchAndRewrite(mlir::Operation*, mlir::PatternRewriter&) const (/home/ana_todorovic/torch-mlir/build/bin/torch-mlir-opt+0x2290656)
#12 0x000055fc16775c8e 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 
#13 0x000055fc1677250b 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&)>) (/home/ana_todorovic/torch-mlir/build/bin/torch-mlir-opt+0x22ca50b)
#14 0x000055fc167398bb (anonymous namespace)::OperationLegalizer::legalize(mlir::Operation*, mlir::ConversionPatternRewriter&) DialectConversion.cpp:0:0
#15 0x000055fc16738777 mlir::OperationConverter::convert(mlir::ConversionPatternRewriter&, mlir::Operation*) (/home/ana_todorovic/torch-mlir/build/bin/torch-mlir-opt+0x2290777)
#16 0x000055fc16739aff mlir::OperationConverter::convertOperations(llvm::ArrayRef<mlir::Operation*>) (/home/ana_todorovic/torch-mlir/build/bin/torch-mlir-opt+0x2291aff)        
#17 0x000055fc1673ed9b mlir::applyPartialConversion(mlir::Operation*, mlir::ConversionTarget const&, mlir::FrozenRewritePatternSet const&, mlir::ConversionConfig) (/home/ana_todorovic/torch-mlir/build/bin/torch-mlir-opt+0x2296d9b)
#18 0x000055fc154f1803 (anonymous namespace)::AdjustCallingConventionsPass::runOnOperation() AdjustCallingConventions.cpp:0:0
#19 0x000055fc167c8d72 mlir::detail::OpToOpPassAdaptor::run(mlir::Pass*, mlir::Operation*, mlir::AnalysisManager, bool, unsigned int) (/home/ana_todorovic/torch-mlir/build/bin/torch-mlir-opt+0x2320d72)
#20 0x000055fc167c9760 mlir::detail::OpToOpPassAdaptor::runPipeline(mlir::OpPassManager&, mlir::Operation*, mlir::AnalysisManager, bool, unsigned int, mlir::PassInstrumentor*, mlir::PassInstrumentation::PipelineParentInfo const*) (/home/ana_todorovic/torch-mlir/build/bin/torch-mlir-opt+0x2321760)
#21 0x000055fc167cbe1c mlir::PassManager::run(mlir::Operation*) (/home/ana_todorovic/torch-mlir/build/bin/torch-mlir-opt+0x2323e1c)
#22 0x000055fc14fc9e1b performActions(llvm::raw_ostream&, std::shared_ptr<llvm::SourceMgr> const&, mlir::MLIRContext*, mlir::MlirOptMainConfig const&) MlirOptMain.cpp:0:0      
#23 0x000055fc14fc9a7f 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
#24 0x000055fc16d2bc65 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) (/home/ana_todorovic/torch-mlir/build/bin/torch-mlir-opt+0x2883c65)
#25 0x000055fc14fc10e5 mlir::MlirOptMain(llvm::raw_ostream&, std::unique_ptr<llvm::MemoryBuffer, std::default_delete<llvm::MemoryBuffer>>, mlir::DialectRegistry&, mlir::MlirOptMainConfig const&) (/home/ana_todorovic/torch-mlir/build/bin/torch-mlir-opt+0xb190e5)
#26 0x000055fc14fc1393 mlir::MlirOptMain(int, char**, llvm::StringRef, llvm::StringRef, mlir::DialectRegistry&) (/home/ana_todorovic/torch-mlir/build/bin/torch-mlir-opt+0xb19393)
#27 0x000055fc14fc15a2 mlir::MlirOptMain(int, char**, llvm::StringRef, mlir::DialectRegistry&) (/home/ana_todorovic/torch-mlir/build/bin/torch-mlir-opt+0xb195a2)
#28 0x000055fc14fbf862 main (/home/ana_todorovic/torch-mlir/build/bin/torch-mlir-opt+0xb17862)
#29 0x00007f3d8a36b1ca (/lib/x86_64-linux-gnu/libc.so.6+0x2a1ca)
#30 0x00007f3d8a36b28b __libc_start_main (/lib/x86_64-linux-gnu/libc.so.6+0x2a28b)
#31 0x000055fc14fbf495 _start (/home/ana_todorovic/torch-mlir/build/bin/torch-mlir-opt+0xb17495)

[vivekkhandelwal1]

@praveen-g-ctt, if you have some cycles, can you please take a look at this?

[praveen-g-ctt]

I see a PR related to this #4034 . This should fix the issue