Segmentation fault when running tritonbench flash attention with `--causal`

[yjk21]

Describe the bug

I'm running the benchmarking command from the ws branch, but added the --causal flag, i.e.:

TORCH_CUDA_ARCH_LIST=9.0a cuda-gdb --args python run.py --op flash_attention --only triton_tutorial_flash_v2_ws,triton_tutorial_flash_v2_tma_ws,triton_tutorial_flash_v2 --num-inputs 1 --seq-len 4096 --metrics tflops --batch 8 --n-heads 16 --d-head 128 --causal

I'm seeing a segfault here:

Thread 1 "python" received signal SIGSEGV, Segmentation fault.
0x00007d54b9f6b32e in mlir::detail::IROperandBase::insertInto<mlir::IRObjectWithUseList<mlir::OpOperand> > (useList=0x5b34c5d18750, this=0x5b34c5cef190) at /root/.triton/llvm/llvm-b5cc222d-ubuntu-x64/include/mlir/IR/UseDefLists.h:101
101           nextUse->back = &nextUse;

Without the flag it looks WAI.

Environment details

Tritonbench at 3a5dccb159834968567a2e45e561dc1aeaa8f8a8
Meta triton at 67f51cc1420cabeb6bf4d28c1813e38ea9a92e20

[htyu]

Thanks for reporting the issue. We will take a look.

[pchen7e2]

#284 has nothing to do with this issue. The tool just mistakenly thought the phrase "#18" in my PR summary referred to this issue