diff --git a/src/cmd/compile/internal/ppc64/ssa.go b/src/cmd/compile/internal/ppc64/ssa.go
index 3888aa6527..3e20c44a4c 100644
--- a/src/cmd/compile/internal/ppc64/ssa.go
+++ b/src/cmd/compile/internal/ppc64/ssa.go
@@ -1781,6 +1781,9 @@ func ssaGenValue(s *gc.SSAGenState, v *ssa.Value) {
 		pp := s.Call(v)
 		pp.To.Reg = ppc64.REG_LR
 
+		// Insert a hint this is not a subroutine return.
+		pp.SetFrom3(obj.Addr{Type: obj.TYPE_CONST, Offset: 1})
+
 		if gc.Ctxt.Flag_shared {
 			// When compiling Go into PIC, the function we just
 			// called via pointer might have been implemented in
diff --git a/src/cmd/internal/obj/ppc64/asm9.go b/src/cmd/internal/obj/ppc64/asm9.go
index 775d27d8e8..41e263b2c0 100644
--- a/src/cmd/internal/obj/ppc64/asm9.go
+++ b/src/cmd/internal/obj/ppc64/asm9.go
@@ -334,6 +334,7 @@ var optab = []Optab{
 	{ABC, C_SCON, C_REG, C_NONE, C_SBRA, 16, 4, 0},
 	{ABC, C_SCON, C_REG, C_NONE, C_LBRA, 17, 4, 0},
 	{ABR, C_NONE, C_NONE, C_NONE, C_LR, 18, 4, 0},
+	{ABR, C_NONE, C_NONE, C_SCON, C_LR, 18, 4, 0},
 	{ABR, C_NONE, C_NONE, C_NONE, C_CTR, 18, 4, 0},
 	{ABR, C_REG, C_NONE, C_NONE, C_CTR, 18, 4, 0},
 	{ABR, C_NONE, C_NONE, C_NONE, C_ZOREG, 15, 8, 0},
@@ -2844,6 +2845,7 @@ func (c *ctxt9) asmout(p *obj.Prog, o *Optab, out []uint32) {
 
 	case 18: /* br/bl (lr/ctr); bc/bcl bo,bi,(lr/ctr) */
 		var v int32
+		var bh uint32 = 0
 		if p.As == ABC || p.As == ABCL {
 			v = c.regoff(&p.From) & 31
 		} else {
@@ -2865,6 +2867,15 @@ func (c *ctxt9) asmout(p *obj.Prog, o *Optab, out []uint32) {
 			v = 0
 		}
 
+		// Insert optional branch hint for bclr[l]/bcctr[l]
+		if p.From3Type() != obj.TYPE_NONE {
+			bh = uint32(p.GetFrom3().Offset)
+			if bh == 2 || bh > 3 {
+				log.Fatalf("BH must be 0,1,3 for %v", p)
+			}
+			o1 |= bh << 11
+		}
+
 		if p.As == ABL || p.As == ABCL {
 			o1 |= 1
 		}