[Bug tree-optimization/128484] loop-carried signed bit-field load and its sign extract stay outside the cutoff test

[rguenth at gcc dot gnu.org via Gcc-bugs]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=128484

Richard Biener <rguenth at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
   Last reconfirmed|                            |2026-11-06
             Status|UNCONFIRMED                 |NEW
     Ever confirmed|0                           |1

--- Comment #4 from Richard Biener <rguenth at gcc dot gnu.org> ---
The sink pass does not sink across the backedge (which would require PHI
translation).  The issue with the testcase that it requires "reverse PRE" to

    int j = 0;
    int id;
    for (int i = 0; i < n; ++i)
      {
        id = ext[i].id; 
        if (r2[i] < cutoff)
          {
            if (id >= lo && id <= hi)
              out[j++] = i;
          }
      }

before sinking can then perform the sinking.  Thus we need to eliminate
the loop carried dependency here (or not introduce it in the first place
if that was not in the original testcase).