cmd/internal/obj/s390x: add atomic operation instructions

Adds the following s390x instructions from the interlocked access
facility:

 * LAA(G)  - load and add
 * LAAL(G) - load and add logical
 * LAN(G)  - load and and
 * LAX(G)  - load and exclusive or
 * LAO(G)  - load and or

These instructions can be used for atomic arithmetic/logical
operations. The atomic packages will be updated in future CLs.

Change-Id: Idc850ac6749b3e778fda3da66bcd864f6b1df375
Reviewed-on: https://go-review.googlesource.com/27871Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>

src/cmd/asm/internal/asm/testdata/s390x.s

@@ -78,6 +78,17 @@ TEXT main·foo(SB),7,$16-0 // TEXT main.foo(SB), 7, $16-0
 	DIVWU	R1, R2                // b90400a0b90400b2b99700a1b904002b
 	DIVWU	R1, R2, R3            // b90400a0b90400b2b99700a1b904003b
 
+	LAA	R1, R2, 524287(R3)    // eb213fff7ff8
+	LAAG	R4, R5, -524288(R6)   // eb54600080e8
+	LAAL	R7, R8, 8192(R9)      // eb87900002fa
+	LAALG	R10, R11, -8192(R12)  // ebbac000feea
+	LAN	R1, R2, (R3)          // eb21300000f4
+	LANG	R4, R5, (R6)          // eb54600000e4
+	LAX	R7, R8, (R9)          // eb87900000f7
+	LAXG	R10, R11, (R12)       // ebbac00000e7
+	LAO	R1, R2, (R3)          // eb21300000f6
+	LAOG	R4, R5, (R6)          // eb54600000e6
+
 	XC	$8, (R15), n-8(SP)       // XC  (R15), $8, n-8(SP)       // d707f010f000
 	NC	$8, (R15), n-8(SP)       // NC  (R15), $8, n-8(SP)       // d407f010f000
 	OC	$8, (R15), n-8(SP)       // OC  (R15), $8, n-8(SP)       // d607f010f000

src/cmd/internal/obj/s390x/a.out.go

@@ -364,6 +364,18 @@ const (
 	ALA
 	ALAY
 
+	// interlocked load and op
+	ALAA
+	ALAAG
+	ALAAL
+	ALAALG
+	ALAN
+	ALANG
+	ALAX
+	ALAXG
+	ALAO
+	ALAOG
+
 	// load/store multiple
 	ALMY
 	ALMG

src/cmd/internal/obj/s390x/anames.go

@@ -134,6 +134,16 @@ var Anames = []string{
 	"LARL",
 	"LA",
 	"LAY",
+	"LAA",
+	"LAAG",
+	"LAAL",
+	"LAALG",
+	"LAN",
+	"LANG",
+	"LAX",
+	"LAXG",
+	"LAO",
+	"LAOG",
 	"LMY",
 	"LMG",
 	"STMY",

src/cmd/internal/obj/s390x/asmz.go

@@ -137,6 +137,9 @@ var optab = []Optab{
 	Optab{AMOVBZ, C_ADDR, C_NONE, C_NONE, C_REG, 75, 0},
 	Optab{AMOVB, C_ADDR, C_NONE, C_NONE, C_REG, 75, 0},
 
+	// interlocked load and op
+	Optab{ALAAG, C_REG, C_REG, C_NONE, C_LOREG, 99, 0},
+
 	// integer arithmetic
 	Optab{AADD, C_REG, C_REG, C_NONE, C_REG, 2, 0},
 	Optab{AADD, C_REG, C_NONE, C_NONE, C_REG, 2, 0},
@@ -809,6 +812,16 @@ func buildop(ctxt *obj.Link) {
 			opset(ASTCKC, r)
 			opset(ASTCKE, r)
 			opset(ASTCKF, r)
+		case ALAAG:
+			opset(ALAA, r)
+			opset(ALAAL, r)
+			opset(ALAALG, r)
+			opset(ALAN, r)
+			opset(ALANG, r)
+			opset(ALAX, r)
+			opset(ALAXG, r)
+			opset(ALAO, r)
+			opset(ALAOG, r)
 		case ASTMG:
 			opset(ASTMY, r)
 		case ALMG:
@@ -3800,6 +3813,39 @@ func asmout(ctxt *obj.Link, asm *[]byte) {
 			zRSY(op_LMG, uint32(rstart), uint32(rend), uint32(reg), uint32(offset), asm)
 		}
 
+	case 99: // interlocked load and op
+		if p.To.Index != 0 {
+			ctxt.Diag("cannot use indexed address")
+		}
+		offset := regoff(ctxt, &p.To)
+		if offset < -DISP20/2 || offset >= DISP20/2 {
+			ctxt.Diag("%v does not fit into 20-bit signed integer", offset)
+		}
+		var opcode uint32
+		switch p.As {
+		case ALAA:
+			opcode = op_LAA
+		case ALAAG:
+			opcode = op_LAAG
+		case ALAAL:
+			opcode = op_LAAL
+		case ALAALG:
+			opcode = op_LAALG
+		case ALAN:
+			opcode = op_LAN
+		case ALANG:
+			opcode = op_LANG
+		case ALAX:
+			opcode = op_LAX
+		case ALAXG:
+			opcode = op_LAXG
+		case ALAO:
+			opcode = op_LAO
+		case ALAOG:
+			opcode = op_LAOG
+		}
+		zRSY(opcode, uint32(p.Reg), uint32(p.From.Reg), uint32(p.To.Reg), uint32(offset), asm)
+
 	case 100: // VRX STORE
 		op, m3, _ := vop(p.As)
 		if p.From3 != nil {