Remove dead (fneg (select)) pattern. XorCombine modifications are sufficient for the fneg-modifier-casting.ll regression except gfx1100

Author: chrisjbris

llvm/lib/Target/AMDGPU/SIInstructions.td

@@ -2379,10 +2379,6 @@ def : AMDGPUPat <
 let True16Predicate = NotHasTrue16BitInsts in {
 def : ROTRPattern <V_ALIGNBIT_B32_e64>;
 
-// Prevents regression in fneg-modifier-casting.ll along with modifications to XorCombine() when v2i32 or is legal.
-def : AMDGPUPat <
-  (fneg (select i1:$src0, (f32 (bitconvert i32:$src1)), (f32 (bitconvert i32:$src2)))),
-    (V_CNDMASK_B32_e64 (i32 1), $src2, (i32 1), $src1, $src0)>;
 
 def : GCNPat<(i32 (trunc (srl i64:$src0, (and i32:$src1, (i32 31))))),
           (V_ALIGNBIT_B32_e64 (i32 (EXTRACT_SUBREG (i64 $src0), sub1)),