| Issue |
177826
|
| Summary |
mlir-opt --convert-vector-to-xegpu crashes on vector.maskedstore with constant mask
|
| Labels |
new issue
|
| Assignees |
|
| Reporter |
Subway2023
|
# Description
mlir-opt version: 22.1.0-rc1
Running mlir-opt with the --convert-vector-to-xegpu pass on a module containing a vector.maskedstore with a constant mask triggers an assertion failure
# Reproduce
location: If the **vector.store** is commented out, the crash no longer occurs.
```
module {
func.func @main() {
%0 = vector.step : vector<16xindex>
%1 = vector.constant_mask [16] : vector<16xi1>
%alloc = memref.alloc() : memref<16x16xindex>
%c0 = arith.constant 0 : index
vector.store %0, %alloc[%c0, %c0] : memref<16x16xindex>, vector<16xindex>
return
}
}
```
```
mlir-opt --convert-vector-to-xegpu test.mlir
```
```
mlir-opt: /mnt/sdd1/mlir/tool/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: /mnt/sdd1/sbw/mlir/tool/llvmorg-22.1.0-install/bin/mlir-opt --convert-vector-to-xegpu /mnt/sdd1/sbw/mlir/data/genProgram_1224/250410_96d2/mlir_steps/250410_96d2_5.mlir
#0 0x0000641bd912be32 llvm::sys::PrintStackTrace(llvm::raw_ostream&, int) (/mnt/sdd1/sbw/mlir/tool/llvmorg-22.1.0-install/bin/mlir-opt+0x252fe32)
#1 0x0000641bd91285af llvm::sys::RunSignalHandlers() (/mnt/sdd1/sbw/mlir/tool/llvmorg-22.1.0-install/bin/mlir-opt+0x252c5af)
#2 0x0000641bd91286fc SignalHandler(int, siginfo_t*, void*) Signals.cpp:0:0
#3 0x000074def4445330 (/lib/x86_64-linux-gnu/libc.so.6+0x45330)
#4 0x000074def449eb2c __pthread_kill_implementation ./nptl/pthread_kill.c:44:76
#5 0x000074def449eb2c __pthread_kill_internal ./nptl/pthread_kill.c:78:10
#6 0x000074def449eb2c pthread_kill ./nptl/pthread_kill.c:89:10
#7 0x000074def444527e raise ./signal/../sysdeps/posix/raise.c:27:6
#8 0x000074def44288ff abort ./stdlib/abort.c:81:7
#9 0x000074def442881b _nl_load_domain ./intl/loadmsgcat.c:1177:9
#10 0x000074def443b517 (/lib/x86_64-linux-gnu/libc.so.6+0x3b517)
#11 0x0000641be0d776b6 mlir::Type::getIntOrFloatBitWidth() const (/mnt/sdd1/sbw/mlir/tool/llvmorg-22.1.0-install/bin/mlir-opt+0xa17b6b6)
```
![]()
_______________________________________________
llvm-bugs mailing list
[email protected]
https://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-bugs
