diff --git a/llvm/lib/Target/AArch64/AArch64ISelLowering.cpp b/llvm/lib/Target/AArch64/AArch64ISelLowering.cpp
index 4fef93cc5aec5..836a5819beba0 100644
--- a/llvm/lib/Target/AArch64/AArch64ISelLowering.cpp
+++ b/llvm/lib/Target/AArch64/AArch64ISelLowering.cpp
@@ -24135,9 +24135,6 @@ static SDValue combineStoreValueFPToInt(StoreSDNode *ST,
                                         TargetLowering::DAGCombinerInfo &DCI,
                                         SelectionDAG &DAG,
                                         const AArch64Subtarget *Subtarget) {
-  // Limit to post-legalization in order to avoid peeling truncating stores.
-  if (DCI.isBeforeLegalize())
-    return SDValue();
   if (!Subtarget->isNeonAvailable())
     return SDValue();
   // Source operand is already a vector.
@@ -24174,6 +24171,13 @@ static SDValue combineStoreValueFPToInt(StoreSDNode *ST,
   SDValue VecFP = DAG.getNode(ISD::SCALAR_TO_VECTOR, DL, VecSrcVT, FPSrc);
   SDValue VecConv = DAG.getNode(Value.getOpcode(), DL, VecDstVT, VecFP);
 
+  if (ST->isTruncatingStore()) {
+    EVT NewVecDstVT = EVT::getVectorVT(
+        *DAG.getContext(), ST->getMemoryVT(),
+        VecDstVT.getFixedSizeInBits() / ST->getMemoryVT().getFixedSizeInBits());
+    VecConv = DAG.getNode(AArch64ISD::NVCAST, DL, NewVecDstVT, VecConv);
+  }
+
   SDValue Zero = DAG.getVectorIdxConstant(0, DL);
   SDValue Extracted =
       DAG.getNode(ISD::EXTRACT_VECTOR_ELT, DL, VT, VecConv, Zero);
diff --git a/llvm/lib/Target/AArch64/AArch64InstrInfo.td b/llvm/lib/Target/AArch64/AArch64InstrInfo.td
index 251fd44b6ea31..a62de87b072e9 100644
--- a/llvm/lib/Target/AArch64/AArch64InstrInfo.td
+++ b/llvm/lib/Target/AArch64/AArch64InstrInfo.td
@@ -9273,8 +9273,12 @@ multiclass St1LanePost128Pat<SDPatternOperator scalar_store, Operand VecIndex,
 
 defm : St1LanePost128Pat<post_truncsti8, VectorIndexB, v16i8, i32, ST1i8_POST,
                          1>;
+defm : St1LanePost128Pat<post_truncsti8, VectorIndexB, v16i8, i64, ST1i8_POST,
+                         1>;
 defm : St1LanePost128Pat<post_truncsti16, VectorIndexH, v8i16, i32, ST1i16_POST,
                          2>;
+defm : St1LanePost128Pat<post_truncsti16, VectorIndexH, v8i16, i64, ST1i16_POST,
+                         2>;
 defm : St1LanePost128Pat<post_store, VectorIndexS, v4i32, i32, ST1i32_POST, 4>;
 defm : St1LanePost128Pat<post_store, VectorIndexS, v4f32, f32, ST1i32_POST, 4>;
 defm : St1LanePost128Pat<post_store, VectorIndexD, v2i64, i64, ST1i64_POST, 8>;
diff --git a/llvm/test/CodeGen/AArch64/store-float-conversion.ll b/llvm/test/CodeGen/AArch64/store-float-conversion.ll
index c46801fc16714..1d4073f673edb 100644
--- a/llvm/test/CodeGen/AArch64/store-float-conversion.ll
+++ b/llvm/test/CodeGen/AArch64/store-float-conversion.ll
@@ -27,6 +27,20 @@ entry:
   ret void
 }
 
+define ptr @f32_to_s8_inc(float %f, ptr %dst) {
+; CHECK-LABEL: f32_to_s8_inc:
+; CHECK:       // %bb.0: // %entry
+; CHECK-NEXT:    fcvtzs s0, s0
+; CHECK-NEXT:    st1 { v0.b }[0], [x0], #1
+; CHECK-NEXT:    ret
+entry:
+  %conv = fptosi float %f to i32
+  %trunc = trunc i32 %conv to i8
+  %next = getelementptr i8, ptr %dst, i64 1
+  store i8 %trunc, ptr %dst
+  ret ptr %next
+}
+
 define void @f32_to_u16(float %f, ptr %dst) {
 ; CHECK-LABEL: f32_to_u16:
 ; CHECK:       // %bb.0: // %entry
@@ -53,6 +67,20 @@ entry:
   ret void
 }
 
+define ptr @f32_to_s16_inc(float %f, ptr %dst) {
+; CHECK-LABEL: f32_to_s16_inc:
+; CHECK:       // %bb.0: // %entry
+; CHECK-NEXT:    fcvtzs s0, s0
+; CHECK-NEXT:    st1 { v0.h }[0], [x0], #2
+; CHECK-NEXT:    ret
+entry:
+  %conv = fptosi float %f to i32
+  %trunc = trunc i32 %conv to i16
+  %next = getelementptr i16, ptr %dst, i64 1
+  store i16 %trunc, ptr %dst
+  ret ptr %next
+}
+
 define void @f32_to_u32(float %f, ptr %dst) {
 ; CHECK-LABEL: f32_to_u32:
 ; CHECK:       // %bb.0: // %entry
@@ -77,6 +105,19 @@ entry:
   ret void
 }
 
+define ptr @f32_to_s32_inc(float %f, ptr %dst) {
+; CHECK-LABEL: f32_to_s32_inc:
+; CHECK:       // %bb.0: // %entry
+; CHECK-NEXT:    fcvtzs s0, s0
+; CHECK-NEXT:    st1 { v0.s }[0], [x0], #4
+; CHECK-NEXT:    ret
+entry:
+  %conv = fptosi float %f to i32
+  %next = getelementptr i32, ptr %dst, i64 1
+  store i32 %conv, ptr %dst
+  ret ptr %next
+}
+
 define void @f32_to_s64(float %f, ptr %dst) {
 ; CHECK-LABEL: f32_to_s64:
 ; CHECK:       // %bb.0: // %entry
@@ -115,6 +156,93 @@ entry:
   ret void
 }
 
+define ptr @f64_to_s64_inc(double %d, ptr %dst) {
+; CHECK-LABEL: f64_to_s64_inc:
+; CHECK:       // %bb.0: // %entry
+; CHECK-NEXT:    fcvtzs d0, d0
+; CHECK-NEXT:    st1 { v0.d }[0], [x0], #8
+; CHECK-NEXT:    ret
+entry:
+  %conv = fptosi double %d to i64
+  %next = getelementptr i64, ptr %dst, i64 1
+  store i64 %conv, ptr %dst
+  ret ptr %next
+}
+
+define void @f64_to_u8(double %d, ptr %dst) {
+; CHECK-LABEL: f64_to_u8:
+; CHECK:       // %bb.0:
+; CHECK-NEXT:    fcvtzu d0, d0
+; CHECK-NEXT:    str b0, [x0]
+; CHECK-NEXT:    ret
+  %conv = fptoui double %d to i64
+  %trunc = trunc i64 %conv to i8
+  store i8 %trunc, ptr %dst
+  ret void
+}
+
+define void @f64_to_s8(double %d, ptr %dst) {
+; CHECK-LABEL: f64_to_s8:
+; CHECK:       // %bb.0:
+; CHECK-NEXT:    fcvtzs d0, d0
+; CHECK-NEXT:    str b0, [x0]
+; CHECK-NEXT:    ret
+  %conv = fptosi double %d to i64
+  %trunc = trunc i64 %conv to i8
+  store i8 %trunc, ptr %dst
+  ret void
+}
+
+define ptr @f64_to_s8_inc(double %d, ptr %dst) {
+; CHECK-LABEL: f64_to_s8_inc:
+; CHECK:       // %bb.0:
+; CHECK-NEXT:    fcvtzs d0, d0
+; CHECK-NEXT:    st1 { v0.b }[0], [x0], #1
+; CHECK-NEXT:    ret
+  %conv = fptosi double %d to i64
+  %trunc = trunc i64 %conv to i8
+  store i8 %trunc, ptr %dst
+  %next = getelementptr i8, ptr %dst, i64 1
+  ret ptr %next
+}
+
+define void @f64_to_u16(double %d, ptr %dst) {
+; CHECK-LABEL: f64_to_u16:
+; CHECK:       // %bb.0:
+; CHECK-NEXT:    fcvtzu d0, d0
+; CHECK-NEXT:    str h0, [x0]
+; CHECK-NEXT:    ret
+  %conv = fptoui double %d to i64
+  %trunc = trunc i64 %conv to i16
+  store i16 %trunc, ptr %dst
+  ret void
+}
+
+define void @f64_to_s16(double %d, ptr %dst) {
+; CHECK-LABEL: f64_to_s16:
+; CHECK:       // %bb.0:
+; CHECK-NEXT:    fcvtzs d0, d0
+; CHECK-NEXT:    str h0, [x0]
+; CHECK-NEXT:    ret
+  %conv = fptosi double %d to i64
+  %trunc = trunc i64 %conv to i16
+  store i16 %trunc, ptr %dst
+  ret void
+}
+
+define ptr @f64_to_s16_inc(double %d, ptr %dst) {
+; CHECK-LABEL: f64_to_s16_inc:
+; CHECK:       // %bb.0:
+; CHECK-NEXT:    fcvtzs d0, d0
+; CHECK-NEXT:    st1 { v0.h }[0], [x0], #2
+; CHECK-NEXT:    ret
+  %conv = fptosi double %d to i64
+  %trunc = trunc i64 %conv to i16
+  %next = getelementptr i16, ptr %dst, i64 1
+  store i16 %trunc, ptr %dst
+  ret ptr %next
+}
+
 define i32 @f32_to_i32_multiple_uses(float %f, ptr %dst) {
 ; CHECK-LABEL: f32_to_i32_multiple_uses:
 ; CHECK:       // %bb.0: // %entry
diff --git a/llvm/test/CodeGen/AArch64/tbl-loops.ll b/llvm/test/CodeGen/AArch64/tbl-loops.ll
index 5fc996ad921ff..223698ba225a8 100644
--- a/llvm/test/CodeGen/AArch64/tbl-loops.ll
+++ b/llvm/test/CodeGen/AArch64/tbl-loops.ll
@@ -64,8 +64,7 @@ define void @loop1(ptr noalias nocapture noundef writeonly %dst, ptr nocapture n
 ; CHECK-NEXT:    fcsel s2, s0, s3, mi
 ; CHECK-NEXT:    subs w10, w10, #1
 ; CHECK-NEXT:    fcvtzs s2, s2
-; CHECK-NEXT:    fmov w11, s2
-; CHECK-NEXT:    strb w11, [x9], #1
+; CHECK-NEXT:    st1 { v2.b }[0], [x9], #1
 ; CHECK-NEXT:    b.ne .LBB0_7
 ; CHECK-NEXT:  .LBB0_8: // %for.cond.cleanup
 ; CHECK-NEXT:    ret