Error: Cannot lower continuous masked loads

[manbearian]

created from #7.

I'm unsure of the nature of this failure; further investigation is needed.

repros.zip

triton-shared-opt -triton-to-linalg 26.mlir
triton-shared-opt -triton-to-linalg 37.mlir
triton-shared-opt -triton-to-linalg 55.mlir
triton-shared-opt -triton-to-linalg 76.mlir

Error output:

+++/home/ianb/test/ttirs_linalg_failed/26.mlir
/home/ianb/test/ttirs_linalg_failed/26.mlir:24:11: error: Cannot lower continuous masked loads
    %13 = tt.load %10, %12, %cst_1 {cache = 1 : i32, evict = 1 : i32, isVolatile = false} : tensor<1x128xi64>
          ^
/home/ianb/test/ttirs_linalg_failed/26.mlir:24:11: note: see current operation: %47 = "tt.load"(%36, %40, %8) <{cache = 1 : i32, evict = 1 : i32, isVolatile = false, operand_segment_sizes = array<i32: 1, 1, 1>}> : (tensor<1x128x!tt.ptr<i64, 1>>, tensor<1x128xi1>, tensor<1x128xi64>) -> tensor<1x128xi64>
/home/ianb/test/ttirs_linalg_failed/26.mlir:24:11: error: failed to legalize operation 'tt.load'
    %13 = tt.load %10, %12, %cst_1 {cache = 1 : i32, evict = 1 : i32, isVolatile = false} : tensor<1x128xi64>
          ^
/home/ianb/test/ttirs_linalg_failed/26.mlir:24:11: note: see current operation: %44 = "tt.load"(%36, %40, %8) <{cache = 1 : i32, evict = 1 : i32, isVolatile = false, operand_segment_sizes = array<i32: 1, 1, 1>}> : (tensor<1x128x!tt.ptr<i64, 1>>, tensor<1x128xi1>, tensor<1x128xi64>) -> tensor<1x128xi64>