cmd/compile: make link register allocatable in non-leaf functions

We save and restore the link register in non-leaf functions because
it is clobbered by CALLs. It is therefore available for general
purpose use.

Only enabled on s390x currently. The RC4 benchmarks in particular
benefit from the extra register:

name     old speed     new speed     delta
RC4_128  243MB/s ± 2%  341MB/s ± 2%  +40.46%  (p=0.008 n=5+5)
RC4_1K   267MB/s ± 0%  359MB/s ± 1%  +34.32%  (p=0.008 n=5+5)
RC4_8K   271MB/s ± 0%  362MB/s ± 0%  +33.61%  (p=0.008 n=5+5)

Change-Id: Id23bff95e771da9425353da2f32668b8e34ba09f
Reviewed-on: https://go-review.googlesource.com/30597
Reviewed-by: Cherry Zhang <cherryyz@google.com>
Run-TryBot: Michael Munday <munday@ca.ibm.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>

src/cmd/compile/internal/ssa/regalloc.go

@@ -460,6 +460,18 @@ func (s *regAllocState) allocValToReg(v *Value, mask regMask, nospill bool, line
 	return c
 }
 
+// isLeaf reports whether f performs any calls.
+func isLeaf(f *Func) bool {
+	for _, b := range f.Blocks {
+		for _, v := range b.Values {
+			if opcodeTable[v.Op].call {
+				return false
+			}
+		}
+	}
+	return true
+}
+
 func (s *regAllocState) init(f *Func) {
 	s.f = f
 	s.registers = f.Config.registers
@@ -510,6 +522,12 @@ func (s *regAllocState) init(f *Func) {
 			s.allocatable &^= 1 << 12 // R12
 		}
 	}
+	if s.f.Config.LinkReg != -1 {
+		if isLeaf(f) {
+			// Leaf functions don't save/restore the link register.
+			s.allocatable &^= 1 << uint(s.f.Config.LinkReg)
+		}
+	}
 	if s.f.Config.ctxt.Flag_dynlink {
 		switch s.f.Config.arch {
 		case "amd64":