deps: Update openssl asm files.

thaliproject · Aug 20, 2016 · 66639d3 · 66639d3
1 parent 4f289ae
commit 66639d3
Show file tree

Hide file tree

Showing 11 changed files with 1,920 additions and 571 deletions.
diff --git a/deps/openssl/asm/arm-elf-gas/bn/armv4-gf2m.S b/deps/openssl/asm/arm-elf-gas/bn/armv4-gf2m.S
@@ -11,13 +11,13 @@
 mul_1x1_neon:
 	vshl.u64	d2,d16,#8	@ q1-q3 are slided
 
-	vmull.p8	q0,d16,d17	@ a·bb
+	vmull.p8	q0,d16,d17	@ a·bb
 	vshl.u64	d4,d16,#16
-	vmull.p8	q1,d2,d17	@ a<<8·bb
+	vmull.p8	q1,d2,d17	@ a<<8·bb
 	vshl.u64	d6,d16,#24
-	vmull.p8	q2,d4,d17	@ a<<16·bb
+	vmull.p8	q2,d4,d17	@ a<<16·bb
 	vshr.u64	d2,#8
-	vmull.p8	q3,d6,d17	@ a<<24·bb
+	vmull.p8	q3,d6,d17	@ a<<24·bb
 	vshl.u64	d3,#24
 	veor		d0,d2
 	vshr.u64	d4,#16
@@ -132,20 +132,20 @@ bn_GF2m_mul_2x2:
 
 	vmov	d16,d18
 	vmov	d17,d19
-	bl	mul_1x1_neon		@ a1·b1
+	bl	mul_1x1_neon		@ a1·b1
 	vmov	d22,d0
 
 	vmov	d16,d20
 	vmov	d17,d21
-	bl	mul_1x1_neon		@ a0·b0
+	bl	mul_1x1_neon		@ a0·b0
 	vmov	d23,d0
 
 	veor	d16,d20,d18
 	veor	d17,d21,d19
 	veor	d20,d23,d22
-	bl	mul_1x1_neon		@ (a0+a1)·(b0+b1)
+	bl	mul_1x1_neon		@ (a0+a1)·(b0+b1)
 
-	veor	d0,d20			@ (a0+a1)·(b0+b1)-a0·b0-a1·b1
+	veor	d0,d20			@ (a0+a1)·(b0+b1)-a0·b0-a1·b1
 	vshl.u64 d1,d0,#32
 	vshr.u64 d0,d0,#32
 	veor	d23,d1
@@ -165,7 +165,7 @@ bn_GF2m_mul_2x2:
 	mov	r12,#7<<2
 	sub	sp,sp,#32		@ allocate tab[8]
 
-	bl	mul_1x1_ialu		@ a1·b1
+	bl	mul_1x1_ialu		@ a1·b1
 	str	r5,[r10,#8]
 	str	r4,[r10,#12]
 
@@ -175,13 +175,13 @@ bn_GF2m_mul_2x2:
 	 eor	r2,r2,r1
 	eor	r0,r0,r3
 	 eor	r1,r1,r2
-	bl	mul_1x1_ialu		@ a0·b0
+	bl	mul_1x1_ialu		@ a0·b0
 	str	r5,[r10]
 	str	r4,[r10,#4]
 
 	eor	r1,r1,r2
 	eor	r0,r0,r3
-	bl	mul_1x1_ialu		@ (a1+a0)·(b1+b0)
+	bl	mul_1x1_ialu		@ (a1+a0)·(b1+b0)
 	ldmia	r10,{r6-r9}
 	eor	r5,r5,r4
 	eor	r4,r4,r7

diff --git a/deps/openssl/asm/arm-elf-gas/modes/ghash-armv4.S b/deps/openssl/asm/arm-elf-gas/modes/ghash-armv4.S
@@ -368,8 +368,8 @@ gcm_ghash_neon:
 	vdup.8		d4,d28[0]	@ broadcast lowest byte
 .Linner_neon:
 	subs		r1,r1,#1
-	vmull.p8	q9,d1,d4		@ H.lo·Xi[i]
-	vmull.p8	q8,d0,d4		@ H.hi·Xi[i]
+	vmull.p8	q9,d1,d4		@ H.lo·Xi[i]
+	vmull.p8	q8,d0,d4		@ H.hi·Xi[i]
 	vext.8		q14,q12,#1		@ IN>>=8
 
 	veor		q10,q13		@ modulo-scheduled part
@@ -382,7 +382,7 @@ gcm_ghash_neon:
 	vsli.8		d2,d3,#1		@ compose the "carry" byte
 	vext.8		q10,q12,#1		@ Z>>=8
 
-	vmull.p8	q11,d2,d5		@ "carry"·0xe1
+	vmull.p8	q11,d2,d5		@ "carry"·0xe1
 	vshr.u8		d2,d3,#7		@ save Z's bottom bit
 	vext.8		q13,q9,q12,#1	@ Qlo>>=8
 	veor		q10,q8

diff --git a/deps/openssl/asm/x64-elf-gas/bn/x86_64-mont.s b/deps/openssl/asm/x64-elf-gas/bn/x86_64-mont.s
@@ -31,6 +31,21 @@ bn_mul_mont:
 
 	movq	%r11,8(%rsp,%r9,8)
 .Lmul_body:
+
+
+
+
+
+
+	subq	%rsp,%r11
+	andq	$-4096,%r11
+.Lmul_page_walk:
+	movq	(%rsp,%r11,1),%r10
+	subq	$4096,%r11
+.byte	0x66,0x2e
+
+	jnc	.Lmul_page_walk
+
 	movq	%rdx,%r12
 	movq	(%r8),%r8
 	movq	(%r12),%rbx
@@ -228,6 +243,15 @@ bn_mul4x_mont:
 
 	movq	%r11,8(%rsp,%r9,8)
 .Lmul4x_body:
+	subq	%rsp,%r11
+	andq	$-4096,%r11
+.Lmul4x_page_walk:
+	movq	(%rsp,%r11,1),%r10
+	subq	$4096,%r11
+.byte	0x2e
+
+	jnc	.Lmul4x_page_walk
+
 	movq	%rdi,16(%rsp,%r9,8)
 	movq	%rdx,%r12
 	movq	(%r8),%r8
@@ -610,6 +634,7 @@ bn_mul4x_mont:
 .align	16
 bn_sqr4x_mont:
 .Lsqr4x_enter:
+	movq	%rsp,%rax
 	pushq	%rbx
 	pushq	%rbp
 	pushq	%r12
@@ -618,13 +643,25 @@ bn_sqr4x_mont:
 	pushq	%r15
 
 	shll	$3,%r9d
-	xorq	%r10,%r10
 	movq	%rsp,%r11
-	subq	%r9,%r10
+	negq	%r9
 	movq	(%r8),%r8
-	leaq	-72(%rsp,%r10,2),%rsp
+	leaq	-72(%rsp,%r9,2),%rsp
 	andq	$-1024,%rsp
 
+	subq	%rsp,%r11
+	andq	$-4096,%r11
+.Lsqr4x_page_walk:
+	movq	(%rsp,%r11,1),%r10
+	subq	$4096,%r11
+.byte	0x2e
+
+	jnc	.Lsqr4x_page_walk
+
+	movq	%r9,%r10
+	negq	%r9
+	leaq	-48(%rax),%r11
+