[llvm] [RISCV] Fix incorrect folding of select on ctlz/cttz (PR #155231)

[via llvm-commits]

================
@@ -18754,7 +18758,6 @@ static SDValue foldSelectOfCTTZOrCTLZ(SDNode *N, SelectionDAG &DAG) {
                               CountZeroes.getValueType(), CountZeroesArgument);
   }
 
-  unsigned BitWidth = CountZeroes.getValueSizeInBits();
   SDValue BitWidthMinusOne =
       DAG.getConstant(BitWidth - 1, SDLoc(N), CountZeroes.getValueType());
----------------
MitchBriles wrote:

This wouldn't work because it's handling the case where the MSB is set (which only happens with an input of 0). For example, in the case of `i7`, that mask would be `0b111` which is the max output of `ctlz`/`cttz` anyway.

https://github.com/llvm/llvm-project/pull/155231