[AMDGPU] Fix using wrong register in frame index shift (llvm#101649)

In case of v_mad we have materialized the offset in vgpr and mad is
performed in wave space, later vgpr have to be shifted back in lane
space. [llvm#99556](llvm#99556)
introduces a bug.

Co-authored-by: Pankajdwivedi-25 <[email protected]>
Change-Id: Ic02338cf087cb3c95964b7273609a73a3e7b2d03

Author: vikramRH

llvm/lib/Target/AMDGPU/SIRegisterInfo.cpp

@@ -2757,7 +2757,7 @@ bool SIRegisterInfo::eliminateFrameIndex(MachineBasicBlock::iterator MI,
                 BuildMI(*MBB, MI, DL, TII->get(AMDGPU::V_LSHRREV_B32_e64),
                         TmpResultReg)
                     .addImm(ST.getWavefrontSizeLog2())
-                    .addReg(FrameReg);
+                    .addReg(TmpResultReg);
               }
 
               Register NewDest = IsCopy ? ResultReg

llvm/test/CodeGen/AMDGPU/eliminate-frame-index-s-mov-b32.mir

@@ -2264,7 +2264,7 @@ body:   |
     ; GFX8-NEXT: S_CMP_EQ_I32 $sgpr4, $sgpr5, implicit-def $scc
     ; GFX8-NEXT: $vgpr0 = V_MOV_B32_e32 64, implicit $exec
     ; GFX8-NEXT: $vgpr0 = V_MAD_U32_U24_e64 killed $vgpr0, 64, $sgpr32, 0, implicit $exec
-    ; GFX8-NEXT: $vgpr0 = V_LSHRREV_B32_e64 6, $sgpr32, implicit $exec
+    ; GFX8-NEXT: $vgpr0 = V_LSHRREV_B32_e64 6, $vgpr0, implicit $exec
     ; GFX8-NEXT: $sgpr4 = V_READFIRSTLANE_B32 $vgpr0, implicit $exec
     ; GFX8-NEXT: S_NOP 0, implicit $sgpr0_sgpr1_sgpr2_sgpr3_sgpr4_sgpr5_sgpr6_sgpr7
     ; GFX8-NEXT: S_NOP 0, implicit $sgpr8_sgpr9_sgpr10_sgpr11_sgpr12_sgpr13_sgpr14_sgpr15
@@ -2657,7 +2657,7 @@ body:             |
     ; GFX8-NEXT: S_CMP_EQ_I32 $sgpr4, $sgpr5, implicit-def $scc
     ; GFX8-NEXT: $vgpr0 = V_MOV_B32_e32 68, implicit $exec
     ; GFX8-NEXT: $vgpr0 = V_MAD_U32_U24_e64 killed $vgpr0, 64, $sgpr32, 0, implicit $exec
-    ; GFX8-NEXT: $vgpr0 = V_LSHRREV_B32_e64 6, $sgpr32, implicit $exec
+    ; GFX8-NEXT: $vgpr0 = V_LSHRREV_B32_e64 6, $vgpr0, implicit $exec
     ; GFX8-NEXT: $sgpr4 = V_READFIRSTLANE_B32 $vgpr0, implicit $exec
     ; GFX8-NEXT: S_NOP 0, implicit $sgpr0_sgpr1_sgpr2_sgpr3_sgpr4_sgpr5_sgpr6_sgpr7
     ; GFX8-NEXT: S_NOP 0, implicit $sgpr8_sgpr9_sgpr10_sgpr11_sgpr12_sgpr13_sgpr14_sgpr15