[torch.mlir] option -allow-unregistered-dialect doesn't address torch dialect

[vfdff]

• For test, it is descriptived with torch mlir, and I add option -allow-unregistered-dialect for this custom dialect, but it still report error.

<source>:15:13: error: Dialect `torch' not found for custom op 'torch.constant.bool' 
    %true = torch.constant.bool true

• Related torch mlir test

module {
  func.func @main(
    %input: !torch.vtensor<[2,3,10],f32>,
    %kernel_size: !torch.list<int>,
    %stride: !torch.list<int>,
    %padding: !torch.list<int>,
    %ceil_mode: !torch.bool,
    %count_include_pad: !torch.bool
  ) -> !torch.vtensor<[2,3,4],f32> {
    %true = torch.constant.bool true
    %false = torch.constant.bool false
    %int0 = torch.constant.int 0
    %int2 = torch.constant.int 2
    %int3 = torch.constant.int 3
    %0 = torch.prim.ListConstruct %int3 : (!torch.int) -> !torch.list<int>
    %1 = torch.prim.ListConstruct %int2 : (!torch.int) -> !torch.list<int>
    %2 = torch.prim.ListConstruct %int0 : (!torch.int) -> !torch.list<int>
    %output = torch.aten.avg_pool1d %input, %kernel_size, %stride, %padding, %ceil_mode, %count_include_pad 
        : !torch.vtensor<[2,3,10],f32>, !torch.list<int>, !torch.list<int>, !torch.list<int>, !torch.bool, !torch.bool -> !torch.vtensor<[2,3,4],f32>
    return %output : !torch.vtensor<[2,3,4],f32>
  }
}

[llvmbot]

@llvm/issue-subscribers-mlir

Author: Allen (vfdff)

* For [test](https://godbolt.org/z/9cr7MhqsK), it is descriptived with torch mlir, and I add option **-allow-unregistered-dialect** for this custom dialect, but it still report error. ``` <source>:15:13: error: Dialect `torch' not found for custom op 'torch.constant.bool' %true = torch.constant.bool true ```

• Related torch mlir test

module {
  func.func @<!-- -->main(
    %input: !torch.vtensor&lt;[2,3,10],f32&gt;,
    %kernel_size: !torch.list&lt;int&gt;,
    %stride: !torch.list&lt;int&gt;,
    %padding: !torch.list&lt;int&gt;,
    %ceil_mode: !torch.bool,
    %count_include_pad: !torch.bool
  ) -&gt; !torch.vtensor&lt;[2,3,4],f32&gt; {
    %true = torch.constant.bool true
    %false = torch.constant.bool false
    %int0 = torch.constant.int 0
    %int2 = torch.constant.int 2
    %int3 = torch.constant.int 3
    %0 = torch.prim.ListConstruct %int3 : (!torch.int) -&gt; !torch.list&lt;int&gt;
    %1 = torch.prim.ListConstruct %int2 : (!torch.int) -&gt; !torch.list&lt;int&gt;
    %2 = torch.prim.ListConstruct %int0 : (!torch.int) -&gt; !torch.list&lt;int&gt;
    %output = torch.aten.avg_pool1d %input, %kernel_size, %stride, %padding, %ceil_mode, %count_include_pad 
        : !torch.vtensor&lt;[2,3,10],f32&gt;, !torch.list&lt;int&gt;, !torch.list&lt;int&gt;, !torch.list&lt;int&gt;, !torch.bool, !torch.bool -&gt; !torch.vtensor&lt;[2,3,4],f32&gt;
    return %output : !torch.vtensor&lt;[2,3,4],f32&gt;
  }
}

[joker-eph]

-allow-unregistered-dialect can't help with the custom ASM syntax, only the generic one.