<table border="1" cellspacing="0" cellpadding="8">
    <tr>
        <th>Issue</th>
        <td>
            <a href=https://github.com/llvm/llvm-project/issues/62128>62128</a>
        </td>
    </tr>

    <tr>
        <th>Summary</th>
        <td>
            [MLIR] `--view-op-graph` doesn't handle graph regions (backedges)
        </td>
    </tr>

    <tr>
      <th>Labels</th>
      <td>
            good first issue,
            mlir
      </td>
    </tr>

    <tr>
      <th>Assignees</th>
      <td>
      </td>
    </tr>

    <tr>
      <th>Reporter</th>
      <td>
          mortbopet
      </td>
    </tr>
</table>

<pre>
    As a trivial example, consider the following:

```mlir
module {
  %add = arith.addi %c0, %c1 : index
  %c0 = arith.constant 0 : index
  %c1 = arith.constant 1 : index
}
```

renders the following .dot graph: `mlir-opt --view-op-graph`:
```dot
digraph G {
  compound = true;
 subgraph cluster_1 {
    v2 [label = " ", shape = plain];
    label = "builtin.module : ())\n";
    subgraph cluster_3 {
      v4 [label = " ", shape = plain];
      label = "";
      v5 [label = "arith.addi : (index))\n", shape = ellipse];
      v6 [label = "arith.constant : (index))\n\nvalue: 0 : index", shape = ellipse];
 v7 [label = "arith.constant : (index))\n\nvalue: 1 : index", shape = ellipse];
    }
  }
  v0 -> v5 [label = "0", style = solid];
 v0 -> v5 [label = "1", style = solid];
}
```



I suspect the only required change to the `view-op-graph` pass will be to do a pre-pass on all ops in a region, assign node names to them, and _then_ iterate and emit the operations.
</pre>
<img width="1px" height="1px" alt="" src="http://email.email.llvm.org/o/eJysVctu6zYTfhp6M5BAUveFFkn8-8cB2k1fIKDEscSWIlWSUk7evqDknNhx0hsKyDI1l--bGQ6Hwns1GMSWFI-kOB7EEkbr2sm60NkZw6Gz8rV98CAgOLUqoQG_i2nWSPgT9NZ4JdFBGBHOVmv7osxAsgdCj4S-vUu6P5NWbhdNVi4agVSP-zcA4YWQEkh2BOFUGFMhpYrSnkaiuGBAsgdQRuL3K6-eXjnFeIIwAejntuwz2ztcUh0_RH6djkMj0fnblCGVNsDgxDxGsEuyiZ0DJMmq8CWxc7KrS_penzd8acMukWozgv9f16a302wXs1cnuAVJ9qbzS7c79HrxAd0zu3YEWDmQ4lGLDvXmTTiPv1hSP4oZN-GshTKkOL7DAsCNT7coHZRJ3_Ytpshrwpv4FE8mQl4734WV3YYFsOb_NrAPoX2kBliLO-jrltpj37f7JoEbatRazR7vydfyC_Qf_fQFQ_FkVqHj5j3ctOffYV6r_4CU_UNSAPhxEm6WK4WEZP_7rND0DTm86h3ZW63kbTJfu7O_dv_zw0k4I8WjmsSw5cLrMYTZxwPHT4SfFo8u2bQ-HVQYly5KemsCmpD2diL8xMosq5s8I_zEM05pxViRdPJcdo3giSiLMsmbmiU1y4pEdFVT5nWGXdWlsxli3a_D2d7fwC9-xj5sM8Ma_QoOf1-UQwn9KMyAEOymIyX9OCxgFt7Di9Iaus1OWhAwO0w2hTUgtAY7e1AGBDgclDWxhvtkB2MlghET-gvJtCmNhOcwonkGFdCJgJsIJ3UJco5CZY1PD7LNZJM14oAtK2uWlwXj_DC2RYNZjRVij42skJddw_OqYKyT5_zMmoNqOeUZzVlOm6zOeVrQuqkl7WrJBD13JckpTkLpVOt1Sq0bDsr7BduSM14fts7w28XE-WCthLNyPsBmszcK4Xy7VDiPd5drI07SLYMnOdXKB_-OHFTQ2y3380_ffiHFMdb6fjSDtOgN4VWAURipcZ_pl7L6eMQ60f-GckBPeHNYnG5vW2xvq0svRfbLXzI7-yv2gfDTloAn_LTl-UcAAAD__8uEKZw">