[AMDGPU][CodeGen] enable D16Writes32BitVgpr for gfx12 (#165587)

The issue in #157795 also
reproducible for gfx12.

This should only be used in true16 mode. Although gfx12 not yet has
true16 mode enable, there could be downstream branch to test true16 mode
manually thus enable this workaround now

Author: broxigarchen

llvm/lib/Target/AMDGPU/AMDGPU.td

@@ -2069,6 +2069,7 @@ def FeatureISAVersion12 : FeatureSet<
    FeatureMemoryAtomicFAddF32DenormalSupport,
    FeatureBVHDualAndBVH8Insts,
    FeatureWaitsBeforeSystemScopeStores,
+   FeatureD16Writes32BitVgpr
    ]>;
 
 def FeatureISAVersion12_50 : FeatureSet<
@@ -2143,6 +2144,7 @@ def FeatureISAVersion12_50 : FeatureSet<
    FeatureSupportsXNACK,
    FeatureXNACK,
    FeatureClusters,
+   FeatureD16Writes32BitVgpr,
 ]>;
 
 def FeatureISAVersion12_51 : FeatureSet<

llvm/test/CodeGen/AMDGPU/spillv16.ll

@@ -1,6 +1,8 @@
 ; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py UTC_ARGS: --version 2
 ; RUN: llc -mtriple=amdgcn-amd-amdhsa -mcpu=gfx1100 -mattr=+real-true16 -enable-misched=0 -post-RA-scheduler=0 -stress-regalloc=8 < %s | FileCheck %s -check-prefixes=GCN,GCN-TRUE16
 ; RUN: llc -mtriple=amdgcn-amd-amdhsa -mcpu=gfx1100 -mattr=-real-true16 -enable-misched=0 -post-RA-scheduler=0 -stress-regalloc=8 < %s | FileCheck %s -check-prefixes=GCN,GCN-FAKE16
+; RUN: llc -mtriple=amdgcn-amd-amdhsa -mcpu=gfx1200 -mattr=+real-true16,+d16-write-vgpr32 -enable-misched=0 -post-RA-scheduler=0 -stress-regalloc=8 < %s | FileCheck %s -check-prefixes=GFX12-TRUE16,GFX12-TRUE16-D16W32
+; RUN: llc -mtriple=amdgcn-amd-amdhsa -mcpu=gfx1200 -mattr=+real-true16,-d16-write-vgpr32 -enable-misched=0 -post-RA-scheduler=0 -stress-regalloc=8 < %s | FileCheck %s -check-prefixes=GFX12-TRUE16,GFX12-TRUE16-D16W16
 ; RUN: llc -mtriple=amdgcn-amd-amdhsa -mcpu=gfx1250 -mattr=+real-true16 -enable-misched=0 -post-RA-scheduler=0 -stress-regalloc=8 < %s | FileCheck %s -check-prefixes=GFX1250,GFX1250-TRUE16
 ; RUN: llc -mtriple=amdgcn-amd-amdhsa -mcpu=gfx1250 -mattr=-real-true16 -enable-misched=0 -post-RA-scheduler=0 -stress-regalloc=8 < %s | FileCheck %s -check-prefixes=GFX1250,GFX1250-FAKE16
 
@@ -35,6 +37,26 @@ define void @spill_i16_alu() {
 ; GCN-FAKE16-NEXT:    s_waitcnt_vscnt null, 0x0
 ; GCN-FAKE16-NEXT:    s_setpc_b64 s[30:31]
 ;
+; GFX12-TRUE16-LABEL: spill_i16_alu:
+; GFX12-TRUE16:       ; %bb.0: ; %entry
+; GFX12-TRUE16-NEXT:    s_wait_loadcnt_dscnt 0x0
+; GFX12-TRUE16-NEXT:    s_wait_expcnt 0x0
+; GFX12-TRUE16-NEXT:    s_wait_samplecnt 0x0
+; GFX12-TRUE16-NEXT:    s_wait_bvhcnt 0x0
+; GFX12-TRUE16-NEXT:    s_wait_kmcnt 0x0
+; GFX12-TRUE16-NEXT:    scratch_load_d16_b16 v0, off, s32 scope:SCOPE_SYS
+; GFX12-TRUE16-NEXT:    s_wait_loadcnt 0x0
+; GFX12-TRUE16-NEXT:    v_add_nc_u16 v0.l, 0x7b, v0.l
+; GFX12-TRUE16-NEXT:    scratch_store_b16 off, v0, s32 offset:2 ; 2-byte Folded Spill
+; GFX12-TRUE16-NEXT:    ;;#ASMSTART
+; GFX12-TRUE16-NEXT:    ;;#ASMEND
+; GFX12-TRUE16-NEXT:    scratch_load_d16_b16 v0, off, s32 offset:2 th:TH_LOAD_LU ; 2-byte Folded Reload
+; GFX12-TRUE16-NEXT:    s_wait_loadcnt 0x0
+; GFX12-TRUE16-NEXT:    s_wait_storecnt 0x0
+; GFX12-TRUE16-NEXT:    scratch_store_b16 off, v0, s32 scope:SCOPE_SYS
+; GFX12-TRUE16-NEXT:    s_wait_storecnt 0x0
+; GFX12-TRUE16-NEXT:    s_setpc_b64 s[30:31]
+;
 ; GFX1250-TRUE16-LABEL: spill_i16_alu:
 ; GFX1250-TRUE16:       ; %bb.0: ; %entry
 ; GFX1250-TRUE16-NEXT:    s_wait_loadcnt_dscnt 0x0
@@ -126,6 +148,56 @@ define void @spill_i16_alu_two_vals() {
 ; GCN-FAKE16-NEXT:    s_waitcnt_vscnt null, 0x0
 ; GCN-FAKE16-NEXT:    s_setpc_b64 s[30:31]
 ;
+; GFX12-TRUE16-D16W32-LABEL: spill_i16_alu_two_vals:
+; GFX12-TRUE16-D16W32:       ; %bb.0: ; %entry
+; GFX12-TRUE16-D16W32-NEXT:    s_wait_loadcnt_dscnt 0x0
+; GFX12-TRUE16-D16W32-NEXT:    s_wait_expcnt 0x0
+; GFX12-TRUE16-D16W32-NEXT:    s_wait_samplecnt 0x0
+; GFX12-TRUE16-D16W32-NEXT:    s_wait_bvhcnt 0x0
+; GFX12-TRUE16-D16W32-NEXT:    s_wait_kmcnt 0x0
+; GFX12-TRUE16-D16W32-NEXT:    scratch_load_d16_b16 v0, off, s32 scope:SCOPE_SYS
+; GFX12-TRUE16-D16W32-NEXT:    s_wait_loadcnt 0x0
+; GFX12-TRUE16-D16W32-NEXT:    v_add_nc_u16 v0.l, 0x7b, v0.l
+; GFX12-TRUE16-D16W32-NEXT:    scratch_store_b16 off, v0, s32 offset:6 ; 2-byte Folded Spill
+; GFX12-TRUE16-D16W32-NEXT:    ;;#ASMSTART
+; GFX12-TRUE16-D16W32-NEXT:    ;;#ASMEND
+; GFX12-TRUE16-D16W32-NEXT:    scratch_load_d16_b16 v0, off, s32 offset:4 scope:SCOPE_SYS
+; GFX12-TRUE16-D16W32-NEXT:    s_wait_loadcnt 0x0
+; GFX12-TRUE16-D16W32-NEXT:    scratch_load_d16_hi_b16 v0, off, s32 offset:6 th:TH_LOAD_LU ; 2-byte Folded Reload
+; GFX12-TRUE16-D16W32-NEXT:    s_wait_loadcnt 0x0
+; GFX12-TRUE16-D16W32-NEXT:    v_add_nc_u16 v0.l, 0x7b, v0.l
+; GFX12-TRUE16-D16W32-NEXT:    s_wait_storecnt 0x0
+; GFX12-TRUE16-D16W32-NEXT:    scratch_store_d16_hi_b16 off, v0, s32 scope:SCOPE_SYS
+; GFX12-TRUE16-D16W32-NEXT:    s_wait_storecnt 0x0
+; GFX12-TRUE16-D16W32-NEXT:    scratch_store_b16 off, v0, s32 offset:4 scope:SCOPE_SYS
+; GFX12-TRUE16-D16W32-NEXT:    s_wait_storecnt 0x0
+; GFX12-TRUE16-D16W32-NEXT:    s_setpc_b64 s[30:31]
+;
+; GFX12-TRUE16-D16W16-LABEL: spill_i16_alu_two_vals:
+; GFX12-TRUE16-D16W16:       ; %bb.0: ; %entry
+; GFX12-TRUE16-D16W16-NEXT:    s_wait_loadcnt_dscnt 0x0
+; GFX12-TRUE16-D16W16-NEXT:    s_wait_expcnt 0x0
+; GFX12-TRUE16-D16W16-NEXT:    s_wait_samplecnt 0x0
+; GFX12-TRUE16-D16W16-NEXT:    s_wait_bvhcnt 0x0
+; GFX12-TRUE16-D16W16-NEXT:    s_wait_kmcnt 0x0
+; GFX12-TRUE16-D16W16-NEXT:    scratch_load_d16_b16 v0, off, s32 scope:SCOPE_SYS
+; GFX12-TRUE16-D16W16-NEXT:    s_wait_loadcnt 0x0
+; GFX12-TRUE16-D16W16-NEXT:    v_add_nc_u16 v0.l, 0x7b, v0.l
+; GFX12-TRUE16-D16W16-NEXT:    scratch_store_b16 off, v0, s32 offset:6 ; 2-byte Folded Spill
+; GFX12-TRUE16-D16W16-NEXT:    ;;#ASMSTART
+; GFX12-TRUE16-D16W16-NEXT:    ;;#ASMEND
+; GFX12-TRUE16-D16W16-NEXT:    scratch_load_d16_b16 v0, off, s32 offset:4 scope:SCOPE_SYS
+; GFX12-TRUE16-D16W16-NEXT:    s_wait_loadcnt 0x0
+; GFX12-TRUE16-D16W16-NEXT:    scratch_load_d16_hi_b16 v0, off, s32 offset:6 th:TH_LOAD_LU ; 2-byte Folded Reload
+; GFX12-TRUE16-D16W16-NEXT:    v_add_nc_u16 v0.l, 0x7b, v0.l
+; GFX12-TRUE16-D16W16-NEXT:    s_wait_loadcnt 0x0
+; GFX12-TRUE16-D16W16-NEXT:    s_wait_storecnt 0x0
+; GFX12-TRUE16-D16W16-NEXT:    scratch_store_d16_hi_b16 off, v0, s32 scope:SCOPE_SYS
+; GFX12-TRUE16-D16W16-NEXT:    s_wait_storecnt 0x0
+; GFX12-TRUE16-D16W16-NEXT:    scratch_store_b16 off, v0, s32 offset:4 scope:SCOPE_SYS
+; GFX12-TRUE16-D16W16-NEXT:    s_wait_storecnt 0x0
+; GFX12-TRUE16-D16W16-NEXT:    s_setpc_b64 s[30:31]
+;
 ; GFX1250-TRUE16-LABEL: spill_i16_alu_two_vals:
 ; GFX1250-TRUE16:       ; %bb.0: ; %entry
 ; GFX1250-TRUE16-NEXT:    s_wait_loadcnt_dscnt 0x0
@@ -223,6 +295,25 @@ define void @spill_i16() {
 ; GCN-FAKE16-NEXT:    s_waitcnt_vscnt null, 0x0
 ; GCN-FAKE16-NEXT:    s_setpc_b64 s[30:31]
 ;
+; GFX12-TRUE16-LABEL: spill_i16:
+; GFX12-TRUE16:       ; %bb.0: ; %entry
+; GFX12-TRUE16-NEXT:    s_wait_loadcnt_dscnt 0x0
+; GFX12-TRUE16-NEXT:    s_wait_expcnt 0x0
+; GFX12-TRUE16-NEXT:    s_wait_samplecnt 0x0
+; GFX12-TRUE16-NEXT:    s_wait_bvhcnt 0x0
+; GFX12-TRUE16-NEXT:    s_wait_kmcnt 0x0
+; GFX12-TRUE16-NEXT:    scratch_load_d16_b16 v0, off, s32 scope:SCOPE_SYS
+; GFX12-TRUE16-NEXT:    s_wait_loadcnt 0x0
+; GFX12-TRUE16-NEXT:    scratch_store_b16 off, v0, s32 offset:2 ; 2-byte Folded Spill
+; GFX12-TRUE16-NEXT:    ;;#ASMSTART
+; GFX12-TRUE16-NEXT:    ;;#ASMEND
+; GFX12-TRUE16-NEXT:    scratch_load_d16_b16 v0, off, s32 offset:2 th:TH_LOAD_LU ; 2-byte Folded Reload
+; GFX12-TRUE16-NEXT:    s_wait_loadcnt 0x0
+; GFX12-TRUE16-NEXT:    s_wait_storecnt 0x0
+; GFX12-TRUE16-NEXT:    scratch_store_b16 off, v0, s32 scope:SCOPE_SYS
+; GFX12-TRUE16-NEXT:    s_wait_storecnt 0x0
+; GFX12-TRUE16-NEXT:    s_setpc_b64 s[30:31]
+;
 ; GFX1250-LABEL: spill_i16:
 ; GFX1250:       ; %bb.0: ; %entry
 ; GFX1250-NEXT:    s_wait_loadcnt_dscnt 0x0
@@ -282,6 +373,25 @@ define void @spill_half() {
 ; GCN-FAKE16-NEXT:    s_waitcnt_vscnt null, 0x0
 ; GCN-FAKE16-NEXT:    s_setpc_b64 s[30:31]
 ;
+; GFX12-TRUE16-LABEL: spill_half:
+; GFX12-TRUE16:       ; %bb.0: ; %entry
+; GFX12-TRUE16-NEXT:    s_wait_loadcnt_dscnt 0x0
+; GFX12-TRUE16-NEXT:    s_wait_expcnt 0x0
+; GFX12-TRUE16-NEXT:    s_wait_samplecnt 0x0
+; GFX12-TRUE16-NEXT:    s_wait_bvhcnt 0x0
+; GFX12-TRUE16-NEXT:    s_wait_kmcnt 0x0
+; GFX12-TRUE16-NEXT:    scratch_load_d16_b16 v0, off, s32 scope:SCOPE_SYS
+; GFX12-TRUE16-NEXT:    s_wait_loadcnt 0x0
+; GFX12-TRUE16-NEXT:    scratch_store_b16 off, v0, s32 offset:2 ; 2-byte Folded Spill
+; GFX12-TRUE16-NEXT:    ;;#ASMSTART
+; GFX12-TRUE16-NEXT:    ;;#ASMEND
+; GFX12-TRUE16-NEXT:    scratch_load_d16_b16 v0, off, s32 offset:2 th:TH_LOAD_LU ; 2-byte Folded Reload
+; GFX12-TRUE16-NEXT:    s_wait_loadcnt 0x0
+; GFX12-TRUE16-NEXT:    s_wait_storecnt 0x0
+; GFX12-TRUE16-NEXT:    scratch_store_b16 off, v0, s32 scope:SCOPE_SYS
+; GFX12-TRUE16-NEXT:    s_wait_storecnt 0x0
+; GFX12-TRUE16-NEXT:    s_setpc_b64 s[30:31]
+;
 ; GFX1250-LABEL: spill_half:
 ; GFX1250:       ; %bb.0: ; %entry
 ; GFX1250-NEXT:    s_wait_loadcnt_dscnt 0x0
@@ -341,6 +451,25 @@ define void @spill_i16_from_v2i16() {
 ; GCN-FAKE16-NEXT:    s_waitcnt_vscnt null, 0x0
 ; GCN-FAKE16-NEXT:    s_setpc_b64 s[30:31]
 ;
+; GFX12-TRUE16-LABEL: spill_i16_from_v2i16:
+; GFX12-TRUE16:       ; %bb.0: ; %entry
+; GFX12-TRUE16-NEXT:    s_wait_loadcnt_dscnt 0x0
+; GFX12-TRUE16-NEXT:    s_wait_expcnt 0x0
+; GFX12-TRUE16-NEXT:    s_wait_samplecnt 0x0
+; GFX12-TRUE16-NEXT:    s_wait_bvhcnt 0x0
+; GFX12-TRUE16-NEXT:    s_wait_kmcnt 0x0
+; GFX12-TRUE16-NEXT:    scratch_load_d16_b16 v0, off, s32 offset:2 scope:SCOPE_SYS
+; GFX12-TRUE16-NEXT:    s_wait_loadcnt 0x0
+; GFX12-TRUE16-NEXT:    scratch_store_b16 off, v0, s32 offset:8 ; 2-byte Folded Spill
+; GFX12-TRUE16-NEXT:    ;;#ASMSTART
+; GFX12-TRUE16-NEXT:    ;;#ASMEND
+; GFX12-TRUE16-NEXT:    scratch_load_d16_b16 v0, off, s32 offset:8 th:TH_LOAD_LU ; 2-byte Folded Reload
+; GFX12-TRUE16-NEXT:    s_wait_loadcnt 0x0
+; GFX12-TRUE16-NEXT:    s_wait_storecnt 0x0
+; GFX12-TRUE16-NEXT:    scratch_store_b16 off, v0, s32 offset:2 scope:SCOPE_SYS
+; GFX12-TRUE16-NEXT:    s_wait_storecnt 0x0
+; GFX12-TRUE16-NEXT:    s_setpc_b64 s[30:31]
+;
 ; GFX1250-LABEL: spill_i16_from_v2i16:
 ; GFX1250:       ; %bb.0: ; %entry
 ; GFX1250-NEXT:    s_wait_loadcnt_dscnt 0x0
@@ -414,13 +543,39 @@ define void @spill_2xi16_from_v2i16() {
 ; GCN-FAKE16-NEXT:    s_waitcnt_vscnt null, 0x0
 ; GCN-FAKE16-NEXT:    s_setpc_b64 s[30:31]
 ;
+; GFX12-TRUE16-LABEL: spill_2xi16_from_v2i16:
+; GFX12-TRUE16:       ; %bb.0: ; %entry
+; GFX12-TRUE16-NEXT:    s_wait_loadcnt_dscnt 0x0
+; GFX12-TRUE16-NEXT:    s_wait_expcnt 0x0
+; GFX12-TRUE16-NEXT:    s_wait_samplecnt 0x0
+; GFX12-TRUE16-NEXT:    s_wait_bvhcnt 0x0
+; GFX12-TRUE16-NEXT:    s_wait_kmcnt 0x0
+; GFX12-TRUE16-NEXT:    scratch_load_d16_b16 v0, off, s32 offset:2 scope:SCOPE_SYS
+; GFX12-TRUE16-NEXT:    s_wait_loadcnt 0x0
+; GFX12-TRUE16-NEXT:    scratch_store_b16 off, v0, s32 offset:8 ; 2-byte Folded Spill
+; GFX12-TRUE16-NEXT:    scratch_load_d16_b16 v0, off, s32 scope:SCOPE_SYS
+; GFX12-TRUE16-NEXT:    s_wait_loadcnt 0x0
+; GFX12-TRUE16-NEXT:    scratch_store_b16 off, v0, s32 offset:10 ; 2-byte Folded Spill
+; GFX12-TRUE16-NEXT:    ;;#ASMSTART
+; GFX12-TRUE16-NEXT:    ;;#ASMEND
+; GFX12-TRUE16-NEXT:    scratch_load_d16_b16 v0, off, s32 offset:8 th:TH_LOAD_LU ; 2-byte Folded Reload
+; GFX12-TRUE16-NEXT:    s_wait_loadcnt 0x0
+; GFX12-TRUE16-NEXT:    s_wait_storecnt 0x0
+; GFX12-TRUE16-NEXT:    scratch_store_b16 off, v0, s32 offset:2 scope:SCOPE_SYS
+; GFX12-TRUE16-NEXT:    s_wait_storecnt 0x0
+; GFX12-TRUE16-NEXT:    scratch_load_d16_b16 v0, off, s32 offset:10 th:TH_LOAD_LU ; 2-byte Folded Reload
+; GFX12-TRUE16-NEXT:    s_wait_loadcnt 0x0
+; GFX12-TRUE16-NEXT:    scratch_store_b16 off, v0, s32 scope:SCOPE_SYS
+; GFX12-TRUE16-NEXT:    s_wait_storecnt 0x0
+; GFX12-TRUE16-NEXT:    s_setpc_b64 s[30:31]
+;
 ; GFX1250-TRUE16-LABEL: spill_2xi16_from_v2i16:
 ; GFX1250-TRUE16:       ; %bb.0: ; %entry
 ; GFX1250-TRUE16-NEXT:    s_wait_loadcnt_dscnt 0x0
 ; GFX1250-TRUE16-NEXT:    s_wait_kmcnt 0x0
 ; GFX1250-TRUE16-NEXT:    scratch_load_u16 v0, off, s32 offset:2 scope:SCOPE_SYS
 ; GFX1250-TRUE16-NEXT:    s_wait_loadcnt 0x0
-; GFX1250-TRUE16-NEXT:    s_clause 0x1
+; GFX1250-TRUE16-NEXT:    s_clause 0x1 ; 4-byte Folded Spill
 ; GFX1250-TRUE16-NEXT:    scratch_store_b32 off, v0, s32 offset:12
 ; GFX1250-TRUE16-NEXT:    scratch_load_u16 v0, off, s32 scope:SCOPE_SYS
 ; GFX1250-TRUE16-NEXT:    s_wait_loadcnt 0x0
@@ -444,7 +599,7 @@ define void @spill_2xi16_from_v2i16() {
 ; GFX1250-FAKE16-NEXT:    s_wait_kmcnt 0x0
 ; GFX1250-FAKE16-NEXT:    scratch_load_u16 v0, off, s32 offset:2 scope:SCOPE_SYS
 ; GFX1250-FAKE16-NEXT:    s_wait_loadcnt 0x0
-; GFX1250-FAKE16-NEXT:    s_clause 0x1
+; GFX1250-FAKE16-NEXT:    s_clause 0x1 ; 4-byte Folded Spill
 ; GFX1250-FAKE16-NEXT:    scratch_store_b32 off, v0, s32 offset:8
 ; GFX1250-FAKE16-NEXT:    scratch_load_u16 v0, off, s32 scope:SCOPE_SYS
 ; GFX1250-FAKE16-NEXT:    s_wait_loadcnt 0x0
@@ -520,6 +675,32 @@ define void @spill_2xi16_from_v2i16_one_free_reg() {
 ; GCN-FAKE16-NEXT:    s_waitcnt_vscnt null, 0x0
 ; GCN-FAKE16-NEXT:    s_setpc_b64 s[30:31]
 ;
+; GFX12-TRUE16-LABEL: spill_2xi16_from_v2i16_one_free_reg:
+; GFX12-TRUE16:       ; %bb.0: ; %entry
+; GFX12-TRUE16-NEXT:    s_wait_loadcnt_dscnt 0x0
+; GFX12-TRUE16-NEXT:    s_wait_expcnt 0x0
+; GFX12-TRUE16-NEXT:    s_wait_samplecnt 0x0
+; GFX12-TRUE16-NEXT:    s_wait_bvhcnt 0x0
+; GFX12-TRUE16-NEXT:    s_wait_kmcnt 0x0
+; GFX12-TRUE16-NEXT:    scratch_load_d16_b16 v0, off, s32 offset:2 scope:SCOPE_SYS
+; GFX12-TRUE16-NEXT:    s_wait_loadcnt 0x0
+; GFX12-TRUE16-NEXT:    scratch_store_b16 off, v0, s32 offset:8 ; 2-byte Folded Spill
+; GFX12-TRUE16-NEXT:    scratch_load_d16_b16 v0, off, s32 scope:SCOPE_SYS
+; GFX12-TRUE16-NEXT:    s_wait_loadcnt 0x0
+; GFX12-TRUE16-NEXT:    scratch_store_b16 off, v0, s32 offset:10 ; 2-byte Folded Spill
+; GFX12-TRUE16-NEXT:    ;;#ASMSTART
+; GFX12-TRUE16-NEXT:    ;;#ASMEND
+; GFX12-TRUE16-NEXT:    scratch_load_d16_b16 v0, off, s32 offset:8 th:TH_LOAD_LU ; 2-byte Folded Reload
+; GFX12-TRUE16-NEXT:    s_wait_loadcnt 0x0
+; GFX12-TRUE16-NEXT:    s_wait_storecnt 0x0
+; GFX12-TRUE16-NEXT:    scratch_store_b16 off, v0, s32 offset:2 scope:SCOPE_SYS
+; GFX12-TRUE16-NEXT:    s_wait_storecnt 0x0
+; GFX12-TRUE16-NEXT:    scratch_load_d16_b16 v0, off, s32 offset:10 th:TH_LOAD_LU ; 2-byte Folded Reload
+; GFX12-TRUE16-NEXT:    s_wait_loadcnt 0x0
+; GFX12-TRUE16-NEXT:    scratch_store_b16 off, v0, s32 scope:SCOPE_SYS
+; GFX12-TRUE16-NEXT:    s_wait_storecnt 0x0
+; GFX12-TRUE16-NEXT:    s_setpc_b64 s[30:31]
+;
 ; GFX1250-TRUE16-LABEL: spill_2xi16_from_v2i16_one_free_reg:
 ; GFX1250-TRUE16:       ; %bb.0: ; %entry
 ; GFX1250-TRUE16-NEXT:    s_wait_loadcnt_dscnt 0x0
@@ -595,6 +776,25 @@ define void @spill_v2i16() {
 ; GCN-NEXT:    s_waitcnt_vscnt null, 0x0
 ; GCN-NEXT:    s_setpc_b64 s[30:31]
 ;
+; GFX12-TRUE16-LABEL: spill_v2i16:
+; GFX12-TRUE16:       ; %bb.0: ; %entry
+; GFX12-TRUE16-NEXT:    s_wait_loadcnt_dscnt 0x0
+; GFX12-TRUE16-NEXT:    s_wait_expcnt 0x0
+; GFX12-TRUE16-NEXT:    s_wait_samplecnt 0x0
+; GFX12-TRUE16-NEXT:    s_wait_bvhcnt 0x0
+; GFX12-TRUE16-NEXT:    s_wait_kmcnt 0x0
+; GFX12-TRUE16-NEXT:    scratch_load_b32 v0, off, s32 offset:4 scope:SCOPE_SYS
+; GFX12-TRUE16-NEXT:    s_wait_loadcnt 0x0
+; GFX12-TRUE16-NEXT:    scratch_store_b32 off, v0, s32 offset:8 ; 4-byte Folded Spill
+; GFX12-TRUE16-NEXT:    ;;#ASMSTART
+; GFX12-TRUE16-NEXT:    ;;#ASMEND
+; GFX12-TRUE16-NEXT:    scratch_load_b32 v0, off, s32 offset:8 th:TH_LOAD_LU ; 4-byte Folded Reload
+; GFX12-TRUE16-NEXT:    s_wait_loadcnt 0x0
+; GFX12-TRUE16-NEXT:    s_wait_storecnt 0x0
+; GFX12-TRUE16-NEXT:    scratch_store_b32 off, v0, s32 offset:4 scope:SCOPE_SYS
+; GFX12-TRUE16-NEXT:    s_wait_storecnt 0x0
+; GFX12-TRUE16-NEXT:    s_setpc_b64 s[30:31]
+;
 ; GFX1250-LABEL: spill_v2i16:
 ; GFX1250:       ; %bb.0: ; %entry
 ; GFX1250-NEXT:    s_wait_loadcnt_dscnt 0x0