| Issue |
179847
|
| Summary |
mlir-opt crash: Assertion isIntOrFloat() failed in getIntOrFloatBitWidth during --math-expand-ops with math.ctlz on index type
|
| Labels |
new issue
|
| Assignees |
|
| Reporter |
xvv4402
|
Mlir-opt version: 22.1.0-rc1
Bug Op: **math.ctlz**
Bug Pass: **--math-expand-ops**
Commenting out does not trigger the error: **memref.realloc**
Detailed Description: The crash occurs during the --math-expand-ops pass when it encounters a math.ctlz operation using the index type. While removing memref.realloc appears to "fix" the issue, this is actually due to Dead Code Elimination (DCE); without the realloc operation consuming the result, the pass simply skips the unused math.ctlz and avoids the buggy code path.
```mlir
module {
func.func @main() {
%0 = "tosa.const"() <{values = dense<0> : tensor<1xi32>}> : () -> tensor<1xi32>
%1 = gpu.block_id x
%2 = math.ctlz %1 : index
%3 = index.maxu %2, %1
%4 = bufferization.to_buffer %0 : tensor<1xi32> to memref<1xi32>
%5 = memref.realloc %4(%3) : memref<1xi32> to memref<?xi32>
return
}
}
```
```
mlir-opt --math-expand-ops temp.mlir
```
```
mlir-opt: llvm-project-llvmorg-22.1.0-rc1/mlir/lib/IR/Types.cpp:123: unsigned int mlir::Type::getIntOrFloatBitWidth() const: Assertion `isIntOrFloat() && "only integers and floats have a bitwidth"' failed.
PLEASE submit a bug report to https://github.com/llvm/llvm-project/issues/ and include the crash backtrace and instructions to reproduce the bug.
Stack dump:
0. Program arguments: llvmorg-22.1.0-install/bin/mlir-opt --math-expand-ops temp.mlir
#0 0x00005757bd0a9e32 llvm::sys::PrintStackTrace(llvm::raw_ostream&, int) (llvmorg-22.1.0-install/bin/mlir-opt+0x252fe32)
#1 0x00005757bd0a65af llvm::sys::RunSignalHandlers() (llvmorg-22.1.0-install/bin/mlir-opt+0x252c5af)
#2 0x00005757bd0a66fc SignalHandler(int, siginfo_t*, void*) Signals.cpp:0:0
#3 0x00007eb4a5245330 (/lib/x86_64-linux-gnu/libc.so.6+0x45330)
#4 0x00007eb4a529eb2c __pthread_kill_implementation ./nptl/pthread_kill.c:44:76
#5 0x00007eb4a529eb2c __pthread_kill_internal ./nptl/pthread_kill.c:78:10
#6 0x00007eb4a529eb2c pthread_kill ./nptl/pthread_kill.c:89:10
#7 0x00007eb4a524527e raise ./signal/../sysdeps/posix/raise.c:27:6
#8 0x00007eb4a52288ff abort ./stdlib/abort.c:81:7
#9 0x00007eb4a522881b _nl_load_domain ./intl/loadmsgcat.c:1177:9
#10 0x00007eb4a523b517 (/lib/x86_64-linux-gnu/libc.so.6+0x3b517)
#11 0x00005757c4cf56b6 mlir::Type::getIntOrFloatBitWidth() const (llvmorg-22.1.0-install/bin/mlir-opt+0xa17b6b6)
#12 0x00005757be625927 convertCtlzOp(mlir::math::CountLeadingZerosOp, mlir::PatternRewriter&) ExpandOps.cpp:0:0
#13 0x00005757c48d4e03 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&)>)::'lambda'()::operator()() const PatternApplicator.cpp:0:0
#14 0x00005757c48d63af 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&)>) (llvmorg-22.1.0-install/bin/mlir-opt+0x9d5c3af)
#15 0x00005757c489b3c0 (anonymous namespace)::GreedyPatternRewriteDriver::processWorklist() GreedyPatternRewriteDriver.cpp:0:0
#16 0x00005757c48a0060 mlir::applyPatternsGreedily(mlir::Region&, mlir::FrozenRewritePatternSet const&, mlir::GreedyRewriteConfig, bool*) (llvmorg-22.1.0-install/bin/mlir-opt+0x9d26060)
#17 0x00005757be624e71 (anonymous namespace)::MathExpandOpsPass::runOnOperation() ExpandOps.cpp:0:0
#18 0x00005757c4949756 mlir::detail::OpToOpPassAdaptor::run(mlir::Pass*, mlir::Operation*, mlir::AnalysisManager, bool, unsigned int) (llvmorg-22.1.0-install/bin/mlir-opt+0x9dcf756)
```
![]()
_______________________________________________
llvm-bugs mailing list
[email protected]
https://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-bugs
