[release-branch.go1.8] cmd/internal/obj/x86: use LEAx rather than ADDx when calling DUFFxxxx via GOT
DUFFZERO on 386 is not marked as clobbering flags, but rewriteToUseGot rewrote
"ADUFFZERO $offset" to "MOVL runtime.duffxxx@GOT, CX; ADDL $offset, CX; CALL CX"
which does. Luckily the fix is easier than figuring out what the problem was:
replace the ADDL $offset, CX with LEAL $offset(CX), CX.
On amd64 DUFFZERO clobbers flags, on arm, arm64 and ppc64 ADD does not clobber
flags and s390x does not use the duff functions, so I'm fairly confident this
is the only fix required.
I don't know how to write a test though.
Change-Id: I69b0958f5f45771d61db5f5ecb4ded94e8960d4d
Reviewed-on: https://go-review.googlesource.com/41821
Run-TryBot: Michael Hudson-Doyle <michael.hudson@canonical.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Keith Randall <khr@golang.org>
diff --git a/src/cmd/internal/obj/x86/obj6.go b/src/cmd/internal/obj/x86/obj6.go
index eb6f867..f3dfd32 100644
--- a/src/cmd/internal/obj/x86/obj6.go
+++ b/src/cmd/internal/obj/x86/obj6.go
@@ -322,15 +322,13 @@
// Rewrite p, if necessary, to access global data via the global offset table.
func rewriteToUseGot(ctxt *obj.Link, p *obj.Prog) {
- var add, lea, mov obj.As
+ var lea, mov obj.As
var reg int16
if p.Mode == 64 {
- add = AADDQ
lea = ALEAQ
mov = AMOVQ
reg = REG_R15
} else {
- add = AADDL
lea = ALEAL
mov = AMOVL
reg = REG_CX
@@ -347,8 +345,10 @@
// ADUFFxxx $offset
// becomes
// $MOV runtime.duffxxx@GOT, $reg
- // $ADD $offset, $reg
+ // $LEA $offset($reg), $reg
// CALL $reg
+ // (we use LEAx rather than ADDx because ADDx clobbers
+ // flags and duffzero on 386 does not otherwise do so)
var sym *obj.LSym
if p.As == obj.ADUFFZERO {
sym = obj.Linklookup(ctxt, "runtime.duffzero", 0)
@@ -365,9 +365,10 @@
p.To.Offset = 0
p.To.Sym = nil
p1 := obj.Appendp(ctxt, p)
- p1.As = add
- p1.From.Type = obj.TYPE_CONST
+ p1.As = lea
+ p1.From.Type = obj.TYPE_MEM
p1.From.Offset = offset
+ p1.From.Reg = reg
p1.To.Type = obj.TYPE_REG
p1.To.Reg = reg
p2 := obj.Appendp(ctxt, p1)
