ec/asm/ecp_nistz256-{!x86_64}.pl: fix scatter_w7 function.

The ecp_nistz256_scatter_w7 function is called when application
attempts to use custom generator, i.e. rarely. Even though non-x86_64
versions were wrong, it didn't affect point operations, they were just
not as fast as expected.
Reviewed-by: NRich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/6738)

crypto/ec/asm/ecp_nistz256-armv4.pl

@@ -894,13 +894,13 @@ ecp_nistz256_scatter_w7:
 .Loop_scatter_w7:
 	ldr	$mask,[$inp],#4
 	subs	$index,$index,#1
-	strb	$mask,[$out,#64*0-1]
+	strb	$mask,[$out,#64*0]
 	mov	$mask,$mask,lsr#8
-	strb	$mask,[$out,#64*1-1]
+	strb	$mask,[$out,#64*1]
 	mov	$mask,$mask,lsr#8
-	strb	$mask,[$out,#64*2-1]
+	strb	$mask,[$out,#64*2]
 	mov	$mask,$mask,lsr#8
-	strb	$mask,[$out,#64*3-1]
+	strb	$mask,[$out,#64*3]
 	add	$out,$out,#64*4
 	bne	.Loop_scatter_w7
 

crypto/ec/asm/ecp_nistz256-armv8.pl

@@ -1776,21 +1776,21 @@ ecp_nistz256_scatter_w7:
 	prfm	pstl1strm,[$out,#4096+64*5]
 	prfm	pstl1strm,[$out,#4096+64*6]
 	prfm	pstl1strm,[$out,#4096+64*7]
-	strb	w3,[$out,#64*0-1]
+	strb	w3,[$out,#64*0]
 	lsr	x3,x3,#8
-	strb	w3,[$out,#64*1-1]
+	strb	w3,[$out,#64*1]
 	lsr	x3,x3,#8
-	strb	w3,[$out,#64*2-1]
+	strb	w3,[$out,#64*2]
 	lsr	x3,x3,#8
-	strb	w3,[$out,#64*3-1]
+	strb	w3,[$out,#64*3]
 	lsr	x3,x3,#8
-	strb	w3,[$out,#64*4-1]
+	strb	w3,[$out,#64*4]
 	lsr	x3,x3,#8
-	strb	w3,[$out,#64*5-1]
+	strb	w3,[$out,#64*5]
 	lsr	x3,x3,#8
-	strb	w3,[$out,#64*6-1]
+	strb	w3,[$out,#64*6]
 	lsr	x3,x3,#8
-	strb	w3,[$out,#64*7-1]
+	strb	w3,[$out,#64*7]
 	add	$out,$out,#64*8
 	b.ne	.Loop_scatter_w7
 

crypto/ec/asm/ecp_nistz256-ppc64.pl

@@ -2297,21 +2297,21 @@ ecp_nistz256_scatter_w7:
 
 .Loop_scatter_w7:
 	ldu	r0,8($inp)
-	stb	r0,64*0-1($out)
+	stb	r0,64*0($out)
 	srdi	r0,r0,8
-	stb	r0,64*1-1($out)
+	stb	r0,64*1($out)
 	srdi	r0,r0,8
-	stb	r0,64*2-1($out)
+	stb	r0,64*2($out)
 	srdi	r0,r0,8
-	stb	r0,64*3-1($out)
+	stb	r0,64*3($out)
 	srdi	r0,r0,8
-	stb	r0,64*4-1($out)
+	stb	r0,64*4($out)
 	srdi	r0,r0,8
-	stb	r0,64*5-1($out)
+	stb	r0,64*5($out)
 	srdi	r0,r0,8
-	stb	r0,64*6-1($out)
+	stb	r0,64*6($out)
 	srdi	r0,r0,8
-	stb	r0,64*7-1($out)
+	stb	r0,64*7($out)
 	addi	$out,$out,64*8
 	bdnz	.Loop_scatter_w7
 

crypto/ec/asm/ecp_nistz256-sparcv9.pl

@@ -1531,13 +1531,13 @@ ecp_nistz256_scatter_w7:
 	ld	[$inp],%l0
 	add	$inp,4,$inp
 	subcc	$index,1,$index
-	stb	%l0,[$out+64*0-1]
+	stb	%l0,[$out+64*0]
 	srl	%l0,8,%l1
-	stb	%l1,[$out+64*1-1]
+	stb	%l1,[$out+64*1]
 	srl	%l0,16,%l2
-	stb	%l2,[$out+64*2-1]
+	stb	%l2,[$out+64*2]
 	srl	%l0,24,%l3
-	stb	%l3,[$out+64*3-1]
+	stb	%l3,[$out+64*3]
 	bne	.Loop_scatter_w7
 	add	$out,64*4,$out
 

crypto/ec/asm/ecp_nistz256-x86.pl

@@ -1179,7 +1179,7 @@ for ($i=0;$i<7;$i++) {
 	&mov	("esi",&wparam(1));
 	&mov	("ebp",&wparam(2));
 
-	&lea	("edi",&DWP(-1,"edi","ebp"));
+	&lea	("edi",&DWP(0,"edi","ebp"));
 	&mov	("ebp",64/4);
 &set_label("scatter_w7_loop");
 	&mov	("eax",&DWP(0,"esi"));