From: bill_ofarrell <billo@ca.ibm.com>
Date: Fri, 26 Oct 2018 21:33:51 +0000 (-0400)
Subject: cmd/asm: add s390x VMSLG instruction variants
X-Git-Tag: go1.12beta1~606
X-Git-Url: http://www.git.cypherpunks.su/?a=commitdiff_plain;h=3f3142ad99155424007b8e8182265b5b716b44ee;p=gostls13.git

cmd/asm: add s390x VMSLG instruction variants

VMSLG has three variants on z14 and later machines. These variants are used in "limbified" squaring:
VMSLEG: Even Shift Indication -- the even-indexed intermediate result is doubled
VMSLOG: Odd Shift Indication -- the odd-indexed intermediate result is doubled
VMSLEOG: Even and Odd Shift Indication -- both intermediate results are doubled
Limbified squaring is very useful for high performance cryptographic algorithms, such as
elliptic curve. This change allows these instructions to be used in Go assembly.

Change-Id: Iaad577b07320205539f99b3cb37a2a984882721b
Reviewed-on: https://go-review.googlesource.com/c/145180
Reviewed-by: Michael Munday <mike.munday@ibm.com>
---

diff --git a/src/cmd/asm/internal/asm/testdata/s390x.s b/src/cmd/asm/internal/asm/testdata/s390x.s
index ad70d2af44..0e50303d70 100644
--- a/src/cmd/asm/internal/asm/testdata/s390x.s
+++ b/src/cmd/asm/internal/asm/testdata/s390x.s
@@ -369,6 +369,9 @@ TEXT main·foo(SB),DUPOK|NOSPLIT,$16-0 // TEXT main.foo(SB), DUPOK|NOSPLIT, $16-
 	VSTEH	$7, V31, (R2)           // e7f020007809
 	VSTEB	$15, V29, 4094(R12)     // e7d0cffef808
 	VMSLG	V21, V22, V23, V24      // e78563007fb8
+	VMSLEG  V21, V22, V23, V24      // e78563807fb8
+	VMSLOG  V21, V22, V23, V24      // e78563407fb8
+	VMSLEOG V21, V22, V23, V24      // e78563c07fb8
 
 	RET
 	RET	foo(SB)
diff --git a/src/cmd/internal/obj/s390x/a.out.go b/src/cmd/internal/obj/s390x/a.out.go
index 9ee02a2d0d..af321f6131 100644
--- a/src/cmd/internal/obj/s390x/a.out.go
+++ b/src/cmd/internal/obj/s390x/a.out.go
@@ -945,6 +945,9 @@ const (
 	AVUPLHW
 	AVUPLF
 	AVMSLG
+	AVMSLEG
+	AVMSLOG
+	AVMSLEOG
 
 	// binary
 	ABYTE
diff --git a/src/cmd/internal/obj/s390x/anames.go b/src/cmd/internal/obj/s390x/anames.go
index 2d6ea5abb4..9cea9f962d 100644
--- a/src/cmd/internal/obj/s390x/anames.go
+++ b/src/cmd/internal/obj/s390x/anames.go
@@ -678,6 +678,9 @@ var Anames = []string{
 	"VUPLHW",
 	"VUPLF",
 	"VMSLG",
+	"VMSLEG",
+	"VMSLOG",
+	"VMSLEOG",
 	"BYTE",
 	"WORD",
 	"DWORD",
diff --git a/src/cmd/internal/obj/s390x/asmz.go b/src/cmd/internal/obj/s390x/asmz.go
index 359610c41d..4e43d27790 100644
--- a/src/cmd/internal/obj/s390x/asmz.go
+++ b/src/cmd/internal/obj/s390x/asmz.go
@@ -1379,6 +1379,9 @@ func buildop(ctxt *obj.Link) {
 			opset(AVSBCBIQ, r)
 			opset(AVSBIQ, r)
 			opset(AVMSLG, r)
+			opset(AVMSLEG, r)
+			opset(AVMSLOG, r)
+			opset(AVMSLEOG, r)
 		case AVSEL:
 			opset(AVFMADB, r)
 			opset(AWFMADB, r)
diff --git a/src/cmd/internal/obj/s390x/vector.go b/src/cmd/internal/obj/s390x/vector.go
index 3f1d900367..62adcf6d4a 100644
--- a/src/cmd/internal/obj/s390x/vector.go
+++ b/src/cmd/internal/obj/s390x/vector.go
@@ -45,7 +45,7 @@ func vop(as obj.As) (opcode, es, cs uint32) {
 		return op_VAC, 0, 0
 	case AVACQ:
 		return op_VAC, 4, 0
-	case AVMSLG:
+	case AVMSLG, AVMSLEG, AVMSLOG, AVMSLEOG:
 		return op_VMSL, 3, 0
 	case AVACCC:
 		return op_VACCC, 0, 0
@@ -1058,6 +1058,12 @@ func singleElementMask(as obj.As) uint32 {
 		AWFTCIDB,
 		AWFIDB:
 		return 8
+	case AVMSLEG:
+		return 8
+	case AVMSLOG:
+		return 4
+	case AVMSLEOG:
+		return 12
 	}
 	return 0
 }