openssl asm: reverted to AESNI-x86 code to gnutls 3.4.x code

The newer code was creating position dependent code.

3 files changed, 1283 insertions, 2193 deletions

diff --git a/lib/accelerated/x86/coff/aesni-x86.s b/lib/accelerated/x86/coff/aesni-x86.s
index 2c535e0917..502be77883 100644
--- a/lib/accelerated/x86/coff/aesni-x86.s
+++ b/lib/accelerated/x86/coff/aesni-x86.s
@@ -60,10 +60,7 @@ _aesni_encrypt:
 	leal	16(%edx),%edx
 	jnz	.L000enc1_loop_1
 .byte	102,15,56,221,209
-	pxor	%xmm0,%xmm0
-	pxor	%xmm1,%xmm1
 	movups	%xmm2,(%eax)
-	pxor	%xmm2,%xmm2
 	ret
 .globl	_aesni_decrypt
 .def	_aesni_decrypt;	.scl	2;	.type	32;	.endef
@@ -86,87 +83,31 @@ _aesni_decrypt:
 	leal	16(%edx),%edx
 	jnz	.L001dec1_loop_2
 .byte	102,15,56,223,209
-	pxor	%xmm0,%xmm0
-	pxor	%xmm1,%xmm1
 	movups	%xmm2,(%eax)
-	pxor	%xmm2,%xmm2
-	ret
-.def	__aesni_encrypt2;	.scl	3;	.type	32;	.endef
-.align	16
-__aesni_encrypt2:
-	movups	(%edx),%xmm0
-	shll	$4,%ecx
-	movups	16(%edx),%xmm1
-	xorps	%xmm0,%xmm2
-	pxor	%xmm0,%xmm3
-	movups	32(%edx),%xmm0
-	leal	32(%edx,%ecx,1),%edx
-	negl	%ecx
-	addl	$16,%ecx
-.L002enc2_loop:
-.byte	102,15,56,220,209
-.byte	102,15,56,220,217
-	movups	(%edx,%ecx,1),%xmm1
-	addl	$32,%ecx
-.byte	102,15,56,220,208
-.byte	102,15,56,220,216
-	movups	-16(%edx,%ecx,1),%xmm0
-	jnz	.L002enc2_loop
-.byte	102,15,56,220,209
-.byte	102,15,56,220,217
-.byte	102,15,56,221,208
-.byte	102,15,56,221,216
-	ret
-.def	__aesni_decrypt2;	.scl	3;	.type	32;	.endef
-.align	16
-__aesni_decrypt2:
-	movups	(%edx),%xmm0
-	shll	$4,%ecx
-	movups	16(%edx),%xmm1
-	xorps	%xmm0,%xmm2
-	pxor	%xmm0,%xmm3
-	movups	32(%edx),%xmm0
-	leal	32(%edx,%ecx,1),%edx
-	negl	%ecx
-	addl	$16,%ecx
-.L003dec2_loop:
-.byte	102,15,56,222,209
-.byte	102,15,56,222,217
-	movups	(%edx,%ecx,1),%xmm1
-	addl	$32,%ecx
-.byte	102,15,56,222,208
-.byte	102,15,56,222,216
-	movups	-16(%edx,%ecx,1),%xmm0
-	jnz	.L003dec2_loop
-.byte	102,15,56,222,209
-.byte	102,15,56,222,217
-.byte	102,15,56,223,208
-.byte	102,15,56,223,216
 	ret
 .def	__aesni_encrypt3;	.scl	3;	.type	32;	.endef
 .align	16
 __aesni_encrypt3:
 	movups	(%edx),%xmm0
-	shll	$4,%ecx
+	shrl	$1,%ecx
 	movups	16(%edx),%xmm1
+	leal	32(%edx),%edx
 	xorps	%xmm0,%xmm2
 	pxor	%xmm0,%xmm3
 	pxor	%xmm0,%xmm4
-	movups	32(%edx),%xmm0
-	leal	32(%edx,%ecx,1),%edx
-	negl	%ecx
-	addl	$16,%ecx
-.L004enc3_loop:
+	movups	(%edx),%xmm0
+.L002enc3_loop:
 .byte	102,15,56,220,209
 .byte	102,15,56,220,217
+	decl	%ecx
 .byte	102,15,56,220,225
-	movups	(%edx,%ecx,1),%xmm1
-	addl	$32,%ecx
+	movups	16(%edx),%xmm1
 .byte	102,15,56,220,208
 .byte	102,15,56,220,216
+	leal	32(%edx),%edx
 .byte	102,15,56,220,224
-	movups	-16(%edx,%ecx,1),%xmm0
-	jnz	.L004enc3_loop
+	movups	(%edx),%xmm0
+	jnz	.L002enc3_loop
 .byte	102,15,56,220,209
 .byte	102,15,56,220,217
 .byte	102,15,56,220,225
@@ -178,26 +119,25 @@ __aesni_encrypt3:
 .align	16
 __aesni_decrypt3:
 	movups	(%edx),%xmm0
-	shll	$4,%ecx
+	shrl	$1,%ecx
 	movups	16(%edx),%xmm1
+	leal	32(%edx),%edx
 	xorps	%xmm0,%xmm2
 	pxor	%xmm0,%xmm3
 	pxor	%xmm0,%xmm4
-	movups	32(%edx),%xmm0
-	leal	32(%edx,%ecx,1),%edx
-	negl	%ecx
-	addl	$16,%ecx
-.L005dec3_loop:
+	movups	(%edx),%xmm0
+.L003dec3_loop:
 .byte	102,15,56,222,209
 .byte	102,15,56,222,217
+	decl	%ecx
 .byte	102,15,56,222,225
-	movups	(%edx,%ecx,1),%xmm1
-	addl	$32,%ecx
+	movups	16(%edx),%xmm1
 .byte	102,15,56,222,208
 .byte	102,15,56,222,216
+	leal	32(%edx),%edx
 .byte	102,15,56,222,224
-	movups	-16(%edx,%ecx,1),%xmm0
-	jnz	.L005dec3_loop
+	movups	(%edx),%xmm0
+	jnz	.L003dec3_loop
 .byte	102,15,56,222,209
 .byte	102,15,56,222,217
 .byte	102,15,56,222,225
@@ -210,29 +150,27 @@ __aesni_decrypt3:
 __aesni_encrypt4:
 	movups	(%edx),%xmm0
 	movups	16(%edx),%xmm1
-	shll	$4,%ecx
+	shrl	$1,%ecx
+	leal	32(%edx),%edx
 	xorps	%xmm0,%xmm2
 	pxor	%xmm0,%xmm3
 	pxor	%xmm0,%xmm4
 	pxor	%xmm0,%xmm5
-	movups	32(%edx),%xmm0
-	leal	32(%edx,%ecx,1),%edx
-	negl	%ecx
-.byte	15,31,64,0
-	addl	$16,%ecx
-.L006enc4_loop:
+	movups	(%edx),%xmm0
+.L004enc4_loop:
 .byte	102,15,56,220,209
 .byte	102,15,56,220,217
+	decl	%ecx
 .byte	102,15,56,220,225
 .byte	102,15,56,220,233
-	movups	(%edx,%ecx,1),%xmm1
-	addl	$32,%ecx
+	movups	16(%edx),%xmm1
 .byte	102,15,56,220,208
 .byte	102,15,56,220,216
+	leal	32(%edx),%edx
 .byte	102,15,56,220,224
 .byte	102,15,56,220,232
-	movups	-16(%edx,%ecx,1),%xmm0
-	jnz	.L006enc4_loop
+	movups	(%edx),%xmm0
+	jnz	.L004enc4_loop
 .byte	102,15,56,220,209
 .byte	102,15,56,220,217
 .byte	102,15,56,220,225
@@ -247,29 +185,27 @@ __aesni_encrypt4:
 __aesni_decrypt4:
 	movups	(%edx),%xmm0
 	movups	16(%edx),%xmm1
-	shll	$4,%ecx
+	shrl	$1,%ecx
+	leal	32(%edx),%edx
 	xorps	%xmm0,%xmm2
 	pxor	%xmm0,%xmm3
 	pxor	%xmm0,%xmm4
 	pxor	%xmm0,%xmm5
-	movups	32(%edx),%xmm0
-	leal	32(%edx,%ecx,1),%edx
-	negl	%ecx
-.byte	15,31,64,0
-	addl	$16,%ecx
-.L007dec4_loop:
+	movups	(%edx),%xmm0
+.L005dec4_loop:
 .byte	102,15,56,222,209
 .byte	102,15,56,222,217
+	decl	%ecx
 .byte	102,15,56,222,225
 .byte	102,15,56,222,233
-	movups	(%edx,%ecx,1),%xmm1
-	addl	$32,%ecx
+	movups	16(%edx),%xmm1
 .byte	102,15,56,222,208
 .byte	102,15,56,222,216
+	leal	32(%edx),%edx
 .byte	102,15,56,222,224
 .byte	102,15,56,222,232
-	movups	-16(%edx,%ecx,1),%xmm0
-	jnz	.L007dec4_loop
+	movups	(%edx),%xmm0
+	jnz	.L005dec4_loop
 .byte	102,15,56,222,209
 .byte	102,15,56,222,217
 .byte	102,15,56,222,225
@@ -283,42 +219,45 @@ __aesni_decrypt4:
 .align	16
 __aesni_encrypt6:
 	movups	(%edx),%xmm0
-	shll	$4,%ecx
+	shrl	$1,%ecx
 	movups	16(%edx),%xmm1
+	leal	32(%edx),%edx
 	xorps	%xmm0,%xmm2
 	pxor	%xmm0,%xmm3
-	pxor	%xmm0,%xmm4
 .byte	102,15,56,220,209
-	pxor	%xmm0,%xmm5
-	pxor	%xmm0,%xmm6
+	pxor	%xmm0,%xmm4
 .byte	102,15,56,220,217
-	leal	32(%edx,%ecx,1),%edx
-	negl	%ecx
+	pxor	%xmm0,%xmm5
+	decl	%ecx
 .byte	102,15,56,220,225
+	pxor	%xmm0,%xmm6
+.byte	102,15,56,220,233
 	pxor	%xmm0,%xmm7
-	movups	(%edx,%ecx,1),%xmm0
-	addl	$16,%ecx
-	jmp	.L008_aesni_encrypt6_inner
+.byte	102,15,56,220,241
+	movups	(%edx),%xmm0
+.byte	102,15,56,220,249
+	jmp	.L_aesni_encrypt6_enter
 .align	16
-.L009enc6_loop:
+.L006enc6_loop:
 .byte	102,15,56,220,209
 .byte	102,15,56,220,217
+	decl	%ecx
 .byte	102,15,56,220,225
-.L008_aesni_encrypt6_inner:
 .byte	102,15,56,220,233
 .byte	102,15,56,220,241
 .byte	102,15,56,220,249
+.align	16
 .L_aesni_encrypt6_enter:
-	movups	(%edx,%ecx,1),%xmm1
-	addl	$32,%ecx
+	movups	16(%edx),%xmm1
 .byte	102,15,56,220,208
 .byte	102,15,56,220,216
+	leal	32(%edx),%edx
 .byte	102,15,56,220,224
 .byte	102,15,56,220,232
 .byte	102,15,56,220,240
 .byte	102,15,56,220,248
-	movups	-16(%edx,%ecx,1),%xmm0
-	jnz	.L009enc6_loop
+	movups	(%edx),%xmm0
+	jnz	.L006enc6_loop
 .byte	102,15,56,220,209
 .byte	102,15,56,220,217
 .byte	102,15,56,220,225
@@ -336,42 +275,45 @@ __aesni_encrypt6:
 .align	16
 __aesni_decrypt6:
 	movups	(%edx),%xmm0
-	shll	$4,%ecx
+	shrl	$1,%ecx
 	movups	16(%edx),%xmm1
+	leal	32(%edx),%edx
 	xorps	%xmm0,%xmm2
 	pxor	%xmm0,%xmm3
-	pxor	%xmm0,%xmm4
 .byte	102,15,56,222,209
-	pxor	%xmm0,%xmm5
-	pxor	%xmm0,%xmm6
+	pxor	%xmm0,%xmm4
 .byte	102,15,56,222,217
-	leal	32(%edx,%ecx,1),%edx
-	negl	%ecx
+	pxor	%xmm0,%xmm5
+	decl	%ecx
 .byte	102,15,56,222,225
+	pxor	%xmm0,%xmm6
+.byte	102,15,56,222,233
 	pxor	%xmm0,%xmm7
-	movups	(%edx,%ecx,1),%xmm0
-	addl	$16,%ecx
-	jmp	.L010_aesni_decrypt6_inner
+.byte	102,15,56,222,241
+	movups	(%edx),%xmm0
+.byte	102,15,56,222,249
+	jmp	.L_aesni_decrypt6_enter
 .align	16
-.L011dec6_loop:
+.L007dec6_loop:
 .byte	102,15,56,222,209
 .byte	102,15,56,222,217
+	decl	%ecx
 .byte	102,15,56,222,225
-.L010_aesni_decrypt6_inner:
 .byte	102,15,56,222,233
 .byte	102,15,56,222,241
 .byte	102,15,56,222,249
+.align	16
 .L_aesni_decrypt6_enter:
-	movups	(%edx,%ecx,1),%xmm1
-	addl	$32,%ecx
+	movups	16(%edx),%xmm1
 .byte	102,15,56,222,208
 .byte	102,15,56,222,216
+	leal	32(%edx),%edx
 .byte	102,15,56,222,224
 .byte	102,15,56,222,232
 .byte	102,15,56,222,240
 .byte	102,15,56,222,248
-	movups	-16(%edx,%ecx,1),%xmm0
-	jnz	.L011dec6_loop
+	movups	(%edx),%xmm0
+	jnz	.L007dec6_loop
 .byte	102,15,56,222,209
 .byte	102,15,56,222,217
 .byte	102,15,56,222,225
@@ -400,14 +342,14 @@ _aesni_ecb_encrypt:
 	movl	32(%esp),%edx
 	movl	36(%esp),%ebx
 	andl	$-16,%eax
-	jz	.L012ecb_ret
+	jz	.L008ecb_ret
 	movl	240(%edx),%ecx
 	testl	%ebx,%ebx
-	jz	.L013ecb_decrypt
+	jz	.L009ecb_decrypt
 	movl	%edx,%ebp
 	movl	%ecx,%ebx
 	cmpl	$96,%eax
-	jb	.L014ecb_enc_tail
+	jb	.L010ecb_enc_tail
 	movdqu	(%esi),%xmm2
 	movdqu	16(%esi),%xmm3
 	movdqu	32(%esi),%xmm4
@@ -416,9 +358,9 @@ _aesni_ecb_encrypt:
 	movdqu	80(%esi),%xmm7
 	leal	96(%esi),%esi
 	subl	$96,%eax
-	jmp	.L015ecb_enc_loop6_enter
+	jmp	.L011ecb_enc_loop6_enter
 .align	16
-.L016ecb_enc_loop6:
+.L012ecb_enc_loop6:
 	movups	%xmm2,(%edi)
 	movdqu	(%esi),%xmm2
 	movups	%xmm3,16(%edi)
@@ -433,12 +375,12 @@ _aesni_ecb_encrypt:
 	leal	96(%edi),%edi
 	movdqu	80(%esi),%xmm7
 	leal	96(%esi),%esi
-.L015ecb_enc_loop6_enter:
+.L011ecb_enc_loop6_enter:
 	call	__aesni_encrypt6
 	movl	%ebp,%edx
 	movl	%ebx,%ecx
 	subl	$96,%eax
-	jnc	.L016ecb_enc_loop6
+	jnc	.L012ecb_enc_loop6
 	movups	%xmm2,(%edi)
 	movups	%xmm3,16(%edi)
 	movups	%xmm4,32(%edi)
@@ -447,18 +389,18 @@ _aesni_ecb_encrypt:
 	movups	%xmm7,80(%edi)
 	leal	96(%edi),%edi
 	addl	$96,%eax
-	jz	.L012ecb_ret
-.L014ecb_enc_tail:
+	jz	.L008ecb_ret
+.L010ecb_enc_tail:
 	movups	(%esi),%xmm2
 	cmpl	$32,%eax
-	jb	.L017ecb_enc_one
+	jb	.L013ecb_enc_one
 	movups	16(%esi),%xmm3
-	je	.L018ecb_enc_two
+	je	.L014ecb_enc_two
 	movups	32(%esi),%xmm4
 	cmpl	$64,%eax
-	jb	.L019ecb_enc_three
+	jb	.L015ecb_enc_three
 	movups	48(%esi),%xmm5
-	je	.L020ecb_enc_four
+	je	.L016ecb_enc_four
 	movups	64(%esi),%xmm6
 	xorps	%xmm7,%xmm7
 	call	__aesni_encrypt6
@@ -467,49 +409,50 @@ _aesni_ecb_encrypt:
 	movups	%xmm4,32(%edi)
 	movups	%xmm5,48(%edi)
 	movups	%xmm6,64(%edi)
-	jmp	.L012ecb_ret
+	jmp	.L008ecb_ret
 .align	16
-.L017ecb_enc_one:
+.L013ecb_enc_one:
 	movups	(%edx),%xmm0
 	movups	16(%edx),%xmm1
 	leal	32(%edx),%edx
 	xorps	%xmm0,%xmm2
-.L021enc1_loop_3:
+.L017enc1_loop_3:
 .byte	102,15,56,220,209
 	decl	%ecx
 	movups	(%edx),%xmm1
 	leal	16(%edx),%edx
-	jnz	.L021enc1_loop_3
+	jnz	.L017enc1_loop_3
 .byte	102,15,56,221,209
 	movups	%xmm2,(%edi)
-	jmp	.L012ecb_ret
+	jmp	.L008ecb_ret
 .align	16
-.L018ecb_enc_two:
-	call	__aesni_encrypt2
+.L014ecb_enc_two:
+	xorps	%xmm4,%xmm4
+	call	__aesni_encrypt3
 	movups	%xmm2,(%edi)
 	movups	%xmm3,16(%edi)
-	jmp	.L012ecb_ret
+	jmp	.L008ecb_ret
 .align	16
-.L019ecb_enc_three:
+.L015ecb_enc_three:
 	call	__aesni_encrypt3
 	movups	%xmm2,(%edi)
 	movups	%xmm3,16(%edi)
 	movups	%xmm4,32(%edi)
-	jmp	.L012ecb_ret
+	jmp	.L008ecb_ret
 .align	16
-.L020ecb_enc_four:
+.L016ecb_enc_four:
 	call	__aesni_encrypt4
 	movups	%xmm2,(%edi)
 	movups	%xmm3,16(%edi)
 	movups	%xmm4,32(%edi)
 	movups	%xmm5,48(%edi)
-	jmp	.L012ecb_ret
+	jmp	.L008ecb_ret
 .align	16
-.L013ecb_decrypt:
+.L009ecb_decrypt:
 	movl	%edx,%ebp
 	movl	%ecx,%ebx
 	cmpl	$96,%eax
-	jb	.L022ecb_dec_tail
+	jb	.L018ecb_dec_tail
 	movdqu	(%esi),%xmm2
 	movdqu	16(%esi),%xmm3
 	movdqu	32(%esi),%xmm4
@@ -518,9 +461,9 @@ _aesni_ecb_encrypt:
 	movdqu	80(%esi),%xmm7
 	leal	96(%esi),%esi
 	subl	$96,%eax
-	jmp	.L023ecb_dec_loop6_enter
+	jmp	.L019ecb_dec_loop6_enter
 .align	16
-.L024ecb_dec_loop6:
+.L020ecb_dec_loop6:
 	movups	%xmm2,(%edi)
 	movdqu	(%esi),%xmm2
 	movups	%xmm3,16(%edi)
@@ -535,12 +478,12 @@ _aesni_ecb_encrypt:
 	leal	96(%edi),%edi
 	movdqu	80(%esi),%xmm7
 	leal	96(%esi),%esi
-.L023ecb_dec_loop6_enter:
+.L019ecb_dec_loop6_enter:
 	call	__aesni_decrypt6
 	movl	%ebp,%edx
 	movl	%ebx,%ecx
 	subl	$96,%eax
-	jnc	.L024ecb_dec_loop6
+	jnc	.L020ecb_dec_loop6
 	movups	%xmm2,(%edi)
 	movups	%xmm3,16(%edi)
 	movups	%xmm4,32(%edi)
@@ -549,18 +492,18 @@ _aesni_ecb_encrypt:
 	movups	%xmm7,80(%edi)
 	leal	96(%edi),%edi
 	addl	$96,%eax
-	jz	.L012ecb_ret
-.L022ecb_dec_tail:
+	jz	.L008ecb_ret
+.L018ecb_dec_tail:
 	movups	(%esi),%xmm2
 	cmpl	$32,%eax
-	jb	.L025ecb_dec_one
+	jb	.L021ecb_dec_one
 	movups	16(%esi),%xmm3
-	je	.L026ecb_dec_two
+	je	.L022ecb_dec_two
 	movups	32(%esi),%xmm4
 	cmpl	$64,%eax
-	jb	.L027ecb_dec_three
+	jb	.L023ecb_dec_three
 	movups	48(%esi),%xmm5
-	je	.L028ecb_dec_four
+	je	.L024ecb_dec_four
 	movups	64(%esi),%xmm6
 	xorps	%xmm7,%xmm7
 	call	__aesni_decrypt6
@@ -569,51 +512,44 @@ _aesni_ecb_encrypt:
 	movups	%xmm4,32(%edi)
 	movups	%xmm5,48(%edi)
 	movups	%xmm6,64(%edi)
-	jmp	.L012ecb_ret
+	jmp	.L008ecb_ret
 .align	16
-.L025ecb_dec_one:
+.L021ecb_dec_one:
 	movups	(%edx),%xmm0
 	movups	16(%edx),%xmm1
 	leal	32(%edx),%edx
 	xorps	%xmm0,%xmm2
-.L029dec1_loop_4:
+.L025dec1_loop_4:
 .byte	102,15,56,222,209
 	decl	%ecx
 	movups	(%edx),%xmm1
 	leal	16(%edx),%edx
-	jnz	.L029dec1_loop_4
+	jnz	.L025dec1_loop_4
 .byte	102,15,56,223,209
 	movups	%xmm2,(%edi)
-	jmp	.L012ecb_ret
+	jmp	.L008ecb_ret
 .align	16
-.L026ecb_dec_two:
-	call	__aesni_decrypt2
+.L022ecb_dec_two:
+	xorps	%xmm4,%xmm4
+	call	__aesni_decrypt3
 	movups	%xmm2,(%edi)
 	movups	%xmm3,16(%edi)
-	jmp	.L012ecb_ret
+	jmp	.L008ecb_ret
 .align	16
-.L027ecb_dec_three:
+.L023ecb_dec_three:
 	call	__aesni_decrypt3
 	movups	%xmm2,(%edi)
 	movups	%xmm3,16(%edi)
 	movups	%xmm4,32(%edi)
-	jmp	.L012ecb_ret
+	jmp	.L008ecb_ret
 .align	16
-.L028ecb_dec_four:
+.L024ecb_dec_four:
 	call	__aesni_decrypt4
 	movups	%xmm2,(%edi)
 	movups	%xmm3,16(%edi)
 	movups	%xmm4,32(%edi)
 	movups	%xmm5,48(%edi)
-.L012ecb_ret:
-	pxor	%xmm0,%xmm0
-	pxor	%xmm1,%xmm1
-	pxor	%xmm2,%xmm2
-	pxor	%xmm3,%xmm3
-	pxor	%xmm4,%xmm4
-	pxor	%xmm5,%xmm5
-	pxor	%xmm6,%xmm6
-	pxor	%xmm7,%xmm7
+.L008ecb_ret:
 	popl	%edi
 	popl	%esi
 	popl	%ebx
@@ -651,56 +587,48 @@ _aesni_ccm64_encrypt_blocks:
 	movl	%ebp,20(%esp)
 	movl	%ebp,24(%esp)
 	movl	%ebp,28(%esp)
-	shll	$4,%ecx
-	movl	$16,%ebx
+	shrl	$1,%ecx
 	leal	(%edx),%ebp
 	movdqa	(%esp),%xmm5
 	movdqa	%xmm7,%xmm2
-	leal	32(%edx,%ecx,1),%edx
-	subl	%ecx,%ebx
+	movl	%ecx,%ebx
 .byte	102,15,56,0,253
-.L030ccm64_enc_outer:
+.L026ccm64_enc_outer:
 	movups	(%ebp),%xmm0
 	movl	%ebx,%ecx
 	movups	(%esi),%xmm6
 	xorps	%xmm0,%xmm2
 	movups	16(%ebp),%xmm1
 	xorps	%xmm6,%xmm0
+	leal	32(%ebp),%edx
 	xorps	%xmm0,%xmm3
-	movups	32(%ebp),%xmm0
-.L031ccm64_enc2_loop:
+	movups	(%edx),%xmm0
+.L027ccm64_enc2_loop:
 .byte	102,15,56,220,209
+	decl	%ecx
 .byte	102,15,56,220,217
-	movups	(%edx,%ecx,1),%xmm1
-	addl	$32,%ecx
+	movups	16(%edx),%xmm1
 .byte	102,15,56,220,208
+	leal	32(%edx),%edx
 .byte	102,15,56,220,216
-	movups	-16(%edx,%ecx,1),%xmm0
-	jnz	.L031ccm64_enc2_loop
+	movups	(%edx),%xmm0
+	jnz	.L027ccm64_enc2_loop
 .byte	102,15,56,220,209
 .byte	102,15,56,220,217
 	paddq	16(%esp),%xmm7
-	decl	%eax
 .byte	102,15,56,221,208
 .byte	102,15,56,221,216
+	decl	%eax
 	leal	16(%esi),%esi
 	xorps	%xmm2,%xmm6
 	movdqa	%xmm7,%xmm2
 	movups	%xmm6,(%edi)
-.byte	102,15,56,0,213
 	leal	16(%edi),%edi
-	jnz	.L030ccm64_enc_outer
+.byte	102,15,56,0,213
+	jnz	.L026ccm64_enc_outer
 	movl	48(%esp),%esp
 	movl	40(%esp),%edi
 	movups	%xmm3,(%edi)
-	pxor	%xmm0,%xmm0
-	pxor	%xmm1,%xmm1
-	pxor	%xmm2,%xmm2
-	pxor	%xmm3,%xmm3
-	pxor	%xmm4,%xmm4
-	pxor	%xmm5,%xmm5
-	pxor	%xmm6,%xmm6
-	pxor	%xmm7,%xmm7
 	popl	%edi
 	popl	%esi
 	popl	%ebx
@@ -747,82 +675,71 @@ _aesni_ccm64_decrypt_blocks:
 	movups	16(%edx),%xmm1
 	leal	32(%edx),%edx
 	xorps	%xmm0,%xmm2
-.L032enc1_loop_5:
+.L028enc1_loop_5:
 .byte	102,15,56,220,209
 	decl	%ecx
 	movups	(%edx),%xmm1
 	leal	16(%edx),%edx
-	jnz	.L032enc1_loop_5
+	jnz	.L028enc1_loop_5
 .byte	102,15,56,221,209
-	shll	$4,%ebx
-	movl	$16,%ecx
 	movups	(%esi),%xmm6
 	paddq	16(%esp),%xmm7
 	leal	16(%esi),%esi
-	subl	%ebx,%ecx
-	leal	32(%ebp,%ebx,1),%edx
-	movl	%ecx,%ebx
-	jmp	.L033ccm64_dec_outer
+	jmp	.L029ccm64_dec_outer
 .align	16
-.L033ccm64_dec_outer:
+.L029ccm64_dec_outer:
 	xorps	%xmm2,%xmm6
 	movdqa	%xmm7,%xmm2
+	movl	%ebx,%ecx
 	movups	%xmm6,(%edi)
 	leal	16(%edi),%edi
 .byte	102,15,56,0,213
 	subl	$1,%eax
-	jz	.L034ccm64_dec_break
+	jz	.L030ccm64_dec_break
 	movups	(%ebp),%xmm0
-	movl	%ebx,%ecx
+	shrl	$1,%ecx
 	movups	16(%ebp),%xmm1
 	xorps	%xmm0,%xmm6
+	leal	32(%ebp),%edx
 	xorps	%xmm0,%xmm2
 	xorps	%xmm6,%xmm3
-	movups	32(%ebp),%xmm0
-.L035ccm64_dec2_loop:
+	movups	(%edx),%xmm0
+.L031ccm64_dec2_loop:
 .byte	102,15,56,220,209
+	decl	%ecx
 .byte	102,15,56,220,217
-	movups	(%edx,%ecx,1),%xmm1
-	addl	$32,%ecx
+	movups	16(%edx),%xmm1
 .byte	102,15,56,220,208
+	leal	32(%edx),%edx
 .byte	102,15,56,220,216
-	movups	-16(%edx,%ecx,1),%xmm0
-	jnz	.L035ccm64_dec2_loop
+	movups	(%edx),%xmm0
+	jnz	.L031ccm64_dec2_loop
 	movups	(%esi),%xmm6
 	paddq	16(%esp),%xmm7
 .byte	102,15,56,220,209
 .byte	102,15,56,220,217
+	leal	16(%esi),%esi
 .byte	102,15,56,221,208
 .byte	102,15,56,221,216
-	leal	16(%esi),%esi
-	jmp	.L033ccm64_dec_outer
+	jmp	.L029ccm64_dec_outer
 .align	16
-.L034ccm64_dec_break:
-	movl	240(%ebp),%ecx
+.L030ccm64_dec_break:
 	movl	%ebp,%edx
 	movups	(%edx),%xmm0
 	movups	16(%edx),%xmm1
 	xorps	%xmm0,%xmm6
 	leal	32(%edx),%edx
 	xorps	%xmm6,%xmm3
-.L036enc1_loop_6:
+.L032enc1_loop_6:
 .byte	102,15,56,220,217
 	decl	%ecx
 	movups	(%edx),%xmm1
 	leal	16(%edx),%edx
-	jnz	.L036enc1_loop_6
+	jnz	.L032enc1_loop_6
 .byte	102,15,56,221,217
 	movl	48(%esp),%esp
 	movl	40(%esp),%edi
 	movups	%xmm3,(%edi)
-	pxor	%xmm0,%xmm0
-	pxor	%xmm1,%xmm1
-	pxor	%xmm2,%xmm2
-	pxor	%xmm3,%xmm3
-	pxor	%xmm4,%xmm4
-	pxor	%xmm5,%xmm5
-	pxor	%xmm6,%xmm6
-	pxor	%xmm7,%xmm7
 	popl	%edi
 	popl	%esi
 	popl	%ebx
@@ -847,7 +764,7 @@ _aesni_ctr32_encrypt_blocks:
 	andl	$-16,%esp
 	movl	%ebp,80(%esp)
 	cmpl	$1,%eax
-	je	.L037ctr32_one_shortcut
+	je	.L033ctr32_one_shortcut
 	movdqu	(%ebx),%xmm7
 	movl	$202182159,(%esp)
 	movl	$134810123,4(%esp)
@@ -863,59 +780,63 @@ _aesni_ctr32_encrypt_blocks:
 .byte	102,15,58,34,253,3
 	movl	240(%edx),%ecx
 	bswap	%ebx
-	pxor	%xmm0,%xmm0
 	pxor	%xmm1,%xmm1
+	pxor	%xmm0,%xmm0
 	movdqa	(%esp),%xmm2
-.byte	102,15,58,34,195,0
+.byte	102,15,58,34,203,0
 	leal	3(%ebx),%ebp
-.byte	102,15,58,34,205,0
+.byte	102,15,58,34,197,0
 	incl	%ebx
-.byte	102,15,58,34,195,1
+.byte	102,15,58,34,203,1
 	incl	%ebp
-.byte	102,15,58,34,205,1
+.byte	102,15,58,34,197,1
 	incl	%ebx
-.byte	102,15,58,34,195,2
+.byte	102,15,58,34,203,2
 	incl	%ebp
-.byte	102,15,58,34,205,2
-	movdqa	%xmm0,48(%esp)
-.byte	102,15,56,0,194
-	movdqu	(%edx),%xmm6
-	movdqa	%xmm1,64(%esp)
+.byte	102,15,58,34,197,2
+	movdqa	%xmm1,48(%esp)
 .byte	102,15,56,0,202
-	pshufd	$192,%xmm0,%xmm2
-	pshufd	$128,%xmm0,%xmm3
+	movdqa	%xmm0,64(%esp)
+.byte	102,15,56,0,194
+	pshufd	$192,%xmm1,%xmm2
+	pshufd	$128,%xmm1,%xmm3
 	cmpl	$6,%eax
-	jb	.L038ctr32_tail
-	pxor	%xmm6,%xmm7
-	shll	$4,%ecx
-	movl	$16,%ebx
+	jb	.L034ctr32_tail
 	movdqa	%xmm7,32(%esp)
+	shrl	$1,%ecx
 	movl	%edx,%ebp
-	subl	%ecx,%ebx
-	leal	32(%edx,%ecx,1),%edx
+	movl	%ecx,%ebx
 	subl	$6,%eax
-	jmp	.L039ctr32_loop6
-.align	16
-.L039ctr32_loop6:
-	pshufd	$64,%xmm0,%xmm4
-	movdqa	32(%esp),%xmm0
-	pshufd	$192,%xmm1,%xmm5
+	jmp	.L035ctr32_loop6
+.align	16
+.L035ctr32_loop6:
+	pshufd	$64,%xmm1,%xmm4
+	movdqa	32(%esp),%xmm1
+	pshufd	$192,%xmm0,%xmm5
+	por	%xmm1,%xmm2
+	pshufd	$128,%xmm0,%xmm6
+	por	%xmm1,%xmm3
+	pshufd	$64,%xmm0,%xmm7
+	por	%xmm1,%xmm4
+	por	%xmm1,%xmm5
+	por	%xmm1,%xmm6
+	por	%xmm1,%xmm7
+	movups	(%ebp),%xmm0
+	movups	16(%ebp),%xmm1
+	leal	32(%ebp),%edx
+	decl	%ecx
 	pxor	%xmm0,%xmm2
-	pshufd	$128,%xmm1,%xmm6
 	pxor	%xmm0,%xmm3
-	pshufd	$64,%xmm1,%xmm7
-	movups	16(%ebp),%xmm1
-	pxor	%xmm0,%xmm4
-	pxor	%xmm0,%xmm5
 .byte	102,15,56,220,209
-	pxor	%xmm0,%xmm6
-	pxor	%xmm0,%xmm7
+	pxor	%xmm0,%xmm4
 .byte	102,15,56,220,217
-	movups	32(%ebp),%xmm0
-	movl	%ebx,%ecx
+	pxor	%xmm0,%xmm5
 .byte	102,15,56,220,225
+	pxor	%xmm0,%xmm6
 .byte	102,15,56,220,233
+	pxor	%xmm0,%xmm7
 .byte	102,15,56,220,241
+	movups	(%edx),%xmm0
 .byte	102,15,56,220,249
 	call	.L_aesni_encrypt6_enter
 	movups	(%esi),%xmm1
@@ -926,51 +847,51 @@ _aesni_ctr32_encrypt_blocks:
 	movups	%xmm2,(%edi)
 	movdqa	16(%esp),%xmm0
 	xorps	%xmm1,%xmm4
-	movdqa	64(%esp),%xmm1
+	movdqa	48(%esp),%xmm1
 	movups	%xmm3,16(%edi)
 	movups	%xmm4,32(%edi)
 	paddd	%xmm0,%xmm1
-	paddd	48(%esp),%xmm0
+	paddd	64(%esp),%xmm0
 	movdqa	(%esp),%xmm2
 	movups	48(%esi),%xmm3
 	movups	64(%esi),%xmm4
 	xorps	%xmm3,%xmm5
 	movups	80(%esi),%xmm3
 	leal	96(%esi),%esi
-	movdqa	%xmm0,48(%esp)
-.byte	102,15,56,0,194
+	movdqa	%xmm1,48(%esp)
+.byte	102,15,56,0,202
 	xorps	%xmm4,%xmm6
 	movups	%xmm5,48(%edi)
 	xorps	%xmm3,%xmm7
-	movdqa	%xmm1,64(%esp)
-.byte	102,15,56,0,202
+	movdqa	%xmm0,64(%esp)
+.byte	102,15,56,0,194
 	movups	%xmm6,64(%edi)
-	pshufd	$192,%xmm0,%xmm2
+	pshufd	$192,%xmm1,%xmm2
 	movups	%xmm7,80(%edi)
 	leal	96(%edi),%edi
-	pshufd	$128,%xmm0,%xmm3
+	movl	%ebx,%ecx
+	pshufd	$128,%xmm1,%xmm3
 	subl	$6,%eax
-	jnc	.L039ctr32_loop6
+	jnc	.L035ctr32_loop6
 	addl	$6,%eax
-	jz	.L040ctr32_ret
-	movdqu	(%ebp),%xmm7
+	jz	.L036ctr32_ret
 	movl	%ebp,%edx
-	pxor	32(%esp),%xmm7
-	movl	240(%ebp),%ecx
-.L038ctr32_tail:
+	leal	1(,%ecx,2),%ecx
+	movdqa	32(%esp),%xmm7
+.L034ctr32_tail:
 	por	%xmm7,%xmm2
 	cmpl	$2,%eax
-	jb	.L041ctr32_one
-	pshufd	$64,%xmm0,%xmm4
+	jb	.L037ctr32_one
+	pshufd	$64,%xmm1,%xmm4
 	por	%xmm7,%xmm3
-	je	.L042ctr32_two
-	pshufd	$192,%xmm1,%xmm5
+	je	.L038ctr32_two
+	pshufd	$192,%xmm0,%xmm5
 	por	%xmm7,%xmm4
 	cmpl	$4,%eax
-	jb	.L043ctr32_three
-	pshufd	$128,%xmm1,%xmm6
+	jb	.L039ctr32_three
+	pshufd	$128,%xmm0,%xmm6
 	por	%xmm7,%xmm5
-	je	.L044ctr32_four
+	je	.L040ctr32_four
 	por	%xmm7,%xmm6
 	call	__aesni_encrypt6
 	movups	(%esi),%xmm1
@@ -988,39 +909,39 @@ _aesni_ctr32_encrypt_blocks:
 	movups	%xmm4,32(%edi)
 	movups	%xmm5,48(%edi)
 	movups	%xmm6,64(%edi)
-	jmp	.L040ctr32_ret
+	jmp	.L036ctr32_ret
 .align	16
-.L037ctr32_one_shortcut:
+.L033ctr32_one_shortcut:
 	movups	(%ebx),%xmm2
 	movl	240(%edx),%ecx
-.L041ctr32_one:
+.L037ctr32_one:
 	movups	(%edx),%xmm0
 	movups	16(%edx),%xmm1
 	leal	32(%edx),%edx
 	xorps	%xmm0,%xmm2
-.L045enc1_loop_7:
+.L041enc1_loop_7:
 .byte	102,15,56,220,209
 	decl	%ecx
 	movups	(%edx),%xmm1
 	leal	16(%edx),%edx
-	jnz	.L045enc1_loop_7
+	jnz	.L041enc1_loop_7
 .byte	102,15,56,221,209
 	movups	(%esi),%xmm6
 	xorps	%xmm2,%xmm6
 	movups	%xmm6,(%edi)
-	jmp	.L040ctr32_ret
+	jmp	.L036ctr32_ret
 .align	16
-.L042ctr32_two:
-	call	__aesni_encrypt2
+.L038ctr32_two:
+	call	__aesni_encrypt3
 	movups	(%esi),%xmm5
 	movups	16(%esi),%xmm6
 	xorps	%xmm5,%xmm2
 	xorps	%xmm6,%xmm3
 	movups	%xmm2,(%edi)
 	movups	%xmm3,16(%edi)
-	jmp	.L040ctr32_ret
+	jmp	.L036ctr32_ret
 .align	16
-.L043ctr32_three:
+.L039ctr32_three:
 	call	__aesni_encrypt3
 	movups	(%esi),%xmm5
 	movups	16(%esi),%xmm6
@@ -1031,9 +952,9 @@ _aesni_ctr32_encrypt_blocks:
 	xorps	%xmm7,%xmm4
 	movups	%xmm3,16(%edi)
 	movups	%xmm4,32(%edi)
-	jmp	.L040ctr32_ret
+	jmp	.L036ctr32_ret
 .align	16
-.L044ctr32_four:
+.L040ctr32_four:
 	call	__aesni_encrypt4
 	movups	(%esi),%xmm6
 	movups	16(%esi),%xmm7
@@ -1047,18 +968,7 @@ _aesni_ctr32_encrypt_blocks:
 	xorps	%xmm0,%xmm5
 	movups	%xmm4,32(%edi)
 	movups	%xmm5,48(%edi)
-.L040ctr32_ret:
-	pxor	%xmm0,%xmm0
-	pxor	%xmm1,%xmm1
-	pxor	%xmm2,%xmm2
-	pxor	%xmm3,%xmm3
-	pxor	%xmm4,%xmm4
-	movdqa	%xmm0,32(%esp)
-	pxor	%xmm5,%xmm5
-	movdqa	%xmm0,48(%esp)
-	pxor	%xmm6,%xmm6
-	movdqa	%xmm0,64(%esp)
-	pxor	%xmm7,%xmm7
+.L036ctr32_ret:
 	movl	80(%esp),%esp
 	popl	%edi
 	popl	%esi
@@ -1082,12 +992,12 @@ _aesni_xts_encrypt:
 	movups	16(%edx),%xmm1
 	leal	32(%edx),%edx
 	xorps	%xmm0,%xmm2
-.L046enc1_loop_8:
+.L042enc1_loop_8:
 .byte	102,15,56,220,209
 	decl	%ecx
 	movups	(%edx),%xmm1
 	leal	16(%edx),%edx
-	jnz	.L046enc1_loop_8
+	jnz	.L042enc1_loop_8
 .byte	102,15,56,221,209
 	movl	20(%esp),%esi
 	movl	24(%esp),%edi
@@ -1111,14 +1021,12 @@ _aesni_xts_encrypt:
 	movl	%edx,%ebp
 	movl	%ecx,%ebx
 	subl	$96,%eax
-	jc	.L047xts_enc_short
-	shll	$4,%ecx
-	movl	$16,%ebx
-	subl	%ecx,%ebx
-	leal	32(%edx,%ecx,1),%edx
-	jmp	.L048xts_enc_loop6
+	jc	.L043xts_enc_short
+	shrl	$1,%ecx
+	movl	%ecx,%ebx
+	jmp	.L044xts_enc_loop6
 .align	16
-.L048xts_enc_loop6:
+.L044xts_enc_loop6:
 	pshufd	$19,%xmm0,%xmm2
 	pxor	%xmm0,%xmm0
 	movdqa	%xmm1,(%esp)
@@ -1154,7 +1062,6 @@ _aesni_xts_encrypt:
 	pand	%xmm3,%xmm7
 	movups	(%esi),%xmm2
 	pxor	%xmm1,%xmm7
-	movl	%ebx,%ecx
 	movdqu	16(%esi),%xmm3
 	xorps	%xmm0,%xmm2
 	movdqu	32(%esi),%xmm4
@@ -1170,17 +1077,19 @@ _aesni_xts_encrypt:
 	movdqa	%xmm7,80(%esp)
 	pxor	%xmm1,%xmm7
 	movups	16(%ebp),%xmm1
+	leal	32(%ebp),%edx
 	pxor	16(%esp),%xmm3
-	pxor	32(%esp),%xmm4
 .byte	102,15,56,220,209
-	pxor	48(%esp),%xmm5
-	pxor	64(%esp),%xmm6
+	pxor	32(%esp),%xmm4
 .byte	102,15,56,220,217
-	pxor	%xmm0,%xmm7
-	movups	32(%ebp),%xmm0
+	pxor	48(%esp),%xmm5
+	decl	%ecx
 .byte	102,15,56,220,225
+	pxor	64(%esp),%xmm6
 .byte	102,15,56,220,233
+	pxor	%xmm0,%xmm7
 .byte	102,15,56,220,241
+	movups	(%edx),%xmm0
 .byte	102,15,56,220,249
 	call	.L_aesni_encrypt6_enter
 	movdqa	80(%esp),%xmm1
@@ -1205,25 +1114,26 @@ _aesni_xts_encrypt:
 	paddq	%xmm1,%xmm1
 	pand	%xmm3,%xmm2
 	pcmpgtd	%xmm1,%xmm0
+	movl	%ebx,%ecx
 	pxor	%xmm2,%xmm1
 	subl	$96,%eax
-	jnc	.L048xts_enc_loop6
-	movl	240(%ebp),%ecx
+	jnc	.L044xts_enc_loop6
+	leal	1(,%ecx,2),%ecx
 	movl	%ebp,%edx
 	movl	%ecx,%ebx
-.L047xts_enc_short:
+.L043xts_enc_short:
 	addl	$96,%eax
-	jz	.L049xts_enc_done6x
+	jz	.L045xts_enc_done6x
 	movdqa	%xmm1,%xmm5
 	cmpl	$32,%eax
-	jb	.L050xts_enc_one
+	jb	.L046xts_enc_one
 	pshufd	$19,%xmm0,%xmm2
 	pxor	%xmm0,%xmm0
 	paddq	%xmm1,%xmm1
 	pand	%xmm3,%xmm2
 	pcmpgtd	%xmm1,%xmm0
 	pxor	%xmm2,%xmm1
-	je	.L051xts_enc_two
+	je	.L047xts_enc_two
 	pshufd	$19,%xmm0,%xmm2
 	pxor	%xmm0,%xmm0
 	movdqa	%xmm1,%xmm6
@@ -1232,7 +1142,7 @@ _aesni_xts_encrypt:
 	pcmpgtd	%xmm1,%xmm0
 	pxor	%xmm2,%xmm1
 	cmpl	$64,%eax
-	jb	.L052xts_enc_three
+	jb	.L048xts_enc_three
 	pshufd	$19,%xmm0,%xmm2
 	pxor	%xmm0,%xmm0
 	movdqa	%xmm1,%xmm7
@@ -1242,7 +1152,7 @@ _aesni_xts_encrypt:
 	pxor	%xmm2,%xmm1
 	movdqa	%xmm5,(%esp)
 	movdqa	%xmm6,16(%esp)
-	je	.L053xts_enc_four
+	je	.L049xts_enc_four
 	movdqa	%xmm7,32(%esp)
 	pshufd	$19,%xmm0,%xmm7
 	movdqa	%xmm1,48(%esp)
@@ -1274,9 +1184,9 @@ _aesni_xts_encrypt:
 	movups	%xmm5,48(%edi)
 	movups	%xmm6,64(%edi)
 	leal	80(%edi),%edi
-	jmp	.L054xts_enc_done
+	jmp	.L050xts_enc_done
 .align	16
-.L050xts_enc_one:
+.L046xts_enc_one:
 	movups	(%esi),%xmm2
 	leal	16(%esi),%esi
 	xorps	%xmm5,%xmm2
@@ -1284,36 +1194,37 @@ _aesni_xts_encrypt:
 	movups	16(%edx),%xmm1
 	leal	32(%edx),%edx
 	xorps	%xmm0,%xmm2
-.L055enc1_loop_9:
+.L051enc1_loop_9:
 .byte	102,15,56,220,209
 	decl	%ecx
 	movups	(%edx),%xmm1
 	leal	16(%edx),%edx
-	jnz	.L055enc1_loop_9
+	jnz	.L051enc1_loop_9
 .byte	102,15,56,221,209
 	xorps	%xmm5,%xmm2
 	movups	%xmm2,(%edi)
 	leal	16(%edi),%edi
 	movdqa	%xmm5,%xmm1
-	jmp	.L054xts_enc_done
+	jmp	.L050xts_enc_done
 .align	16
-.L051xts_enc_two:
+.L047xts_enc_two:
 	movaps	%xmm1,%xmm6
 	movups	(%esi),%xmm2
 	movups	16(%esi),%xmm3
 	leal	32(%esi),%esi
 	xorps	%xmm5,%xmm2
 	xorps	%xmm6,%xmm3
-	call	__aesni_encrypt2
+	xorps	%xmm4,%xmm4
+	call	__aesni_encrypt3
 	xorps	%xmm5,%xmm2
 	xorps	%xmm6,%xmm3
 	movups	%xmm2,(%edi)
 	movups	%xmm3,16(%edi)
 	leal	32(%edi),%edi
 	movdqa	%xmm6,%xmm1
-	jmp	.L054xts_enc_done
+	jmp	.L050xts_enc_done
 .align	16
-.L052xts_enc_three:
+.L048xts_enc_three:
 	movaps	%xmm1,%xmm7
 	movups	(%esi),%xmm2
 	movups	16(%esi),%xmm3
@@ -1331,9 +1242,9 @@ _aesni_xts_encrypt:
 	movups	%xmm4,32(%edi)
 	leal	48(%edi),%edi
 	movdqa	%xmm7,%xmm1
-	jmp	.L054xts_enc_done
+	jmp	.L050xts_enc_done
 .align	16
-.L053xts_enc_four:
+.L049xts_enc_four:
 	movaps	%xmm1,%xmm6
 	movups	(%esi),%xmm2
 	movups	16(%esi),%xmm3
@@ -1355,28 +1266,28 @@ _aesni_xts_encrypt:
 	movups	%xmm5,48(%edi)
 	leal	64(%edi),%edi
 	movdqa	%xmm6,%xmm1
-	jmp	.L054xts_enc_done
+	jmp	.L050xts_enc_done
 .align	16
-.L049xts_enc_done6x:
+.L045xts_enc_done6x:
 	movl	112(%esp),%eax
 	andl	$15,%eax
-	jz	.L056xts_enc_ret
+	jz	.L052xts_enc_ret
 	movdqa	%xmm1,%xmm5
 	movl	%eax,112(%esp)
-	jmp	.L057xts_enc_steal
+	jmp	.L053xts_enc_steal
 .align	16
-.L054xts_enc_done:
+.L050xts_enc_done:
 	movl	112(%esp),%eax
 	pxor	%xmm0,%xmm0
 	andl	$15,%eax
-	jz	.L056xts_enc_ret
+	jz	.L052xts_enc_ret
 	pcmpgtd	%xmm1,%xmm0
 	movl	%eax,112(%esp)
 	pshufd	$19,%xmm0,%xmm5
 	paddq	%xmm1,%xmm1
 	pand	96(%esp),%xmm5
 	pxor	%xmm1,%xmm5
-.L057xts_enc_steal:
+.L053xts_enc_steal:
 	movzbl	(%esi),%ecx
 	movzbl	-16(%edi),%edx
 	leal	1(%esi),%esi
@@ -1384,7 +1295,7 @@ _aesni_xts_encrypt:
 	movb	%dl,(%edi)
 	leal	1(%edi),%edi
 	subl	$1,%eax
-	jnz	.L057xts_enc_steal
+	jnz	.L053xts_enc_steal
 	subl	112(%esp),%edi
 	movl	%ebp,%edx
 	movl	%ebx,%ecx
@@ -1394,30 +1305,16 @@ _aesni_xts_encrypt:
 	movups	16(%edx),%xmm1
 	leal	32(%edx),%edx
 	xorps	%xmm0,%xmm2
-.L058enc1_loop_10:
+.L054enc1_loop_10:
 .byte	102,15,56,220,209
 	decl	%ecx
 	movups	(%edx),%xmm1
 	leal	16(%edx),%edx
-	jnz	.L058enc1_loop_10
+	jnz	.L054enc1_loop_10
 .byte	102,15,56,221,209
 	xorps	%xmm5,%xmm2
 	movups	%xmm2,-16(%edi)
-.L056xts_enc_ret:
-	pxor	%xmm0,%xmm0
-	pxor	%xmm1,%xmm1
-	pxor	%xmm2,%xmm2
-	movdqa	%xmm0,(%esp)
-	pxor	%xmm3,%xmm3
-	movdqa	%xmm0,16(%esp)
-	pxor	%xmm4,%xmm4
-	movdqa	%xmm0,32(%esp)
-	pxor	%xmm5,%xmm5
-	movdqa	%xmm0,48(%esp)
-	pxor	%xmm6,%xmm6
-	movdqa	%xmm0,64(%esp)
-	pxor	%xmm7,%xmm7
-	movdqa	%xmm0,80(%esp)
+.L052xts_enc_ret:
 	movl	116(%esp),%esp
 	popl	%edi
 	popl	%esi
@@ -1441,12 +1338,12 @@ _aesni_xts_decrypt:
 	movups	16(%edx),%xmm1
 	leal	32(%edx),%edx
 	xorps	%xmm0,%xmm2
-.L059enc1_loop_11:
+.L055enc1_loop_11:
 .byte	102,15,56,220,209
 	decl	%ecx
 	movups	(%edx),%xmm1
 	leal	16(%edx),%edx
-	jnz	.L059enc1_loop_11
+	jnz	.L055enc1_loop_11
 .byte	102,15,56,221,209
 	movl	20(%esp),%esi
 	movl	24(%esp),%edi
@@ -1475,14 +1372,12 @@ _aesni_xts_decrypt:
 	pcmpgtd	%xmm1,%xmm0
 	andl	$-16,%eax
 	subl	$96,%eax
-	jc	.L060xts_dec_short
-	shll	$4,%ecx
-	movl	$16,%ebx
-	subl	%ecx,%ebx
-	leal	32(%edx,%ecx,1),%edx
-	jmp	.L061xts_dec_loop6
+	jc	.L056xts_dec_short
+	shrl	$1,%ecx
+	movl	%ecx,%ebx
+	jmp	.L057xts_dec_loop6
 .align	16
-.L061xts_dec_loop6:
+.L057xts_dec_loop6:
 	pshufd	$19,%xmm0,%xmm2
 	pxor	%xmm0,%xmm0
 	movdqa	%xmm1,(%esp)
@@ -1518,7 +1413,6 @@ _aesni_xts_decrypt:
 	pand	%xmm3,%xmm7
 	movups	(%esi),%xmm2
 	pxor	%xmm1,%xmm7
-	movl	%ebx,%ecx
 	movdqu	16(%esi),%xmm3
 	xorps	%xmm0,%xmm2
 	movdqu	32(%esi),%xmm4
@@ -1534,17 +1428,19 @@ _aesni_xts_decrypt:
 	movdqa	%xmm7,80(%esp)
 	pxor	%xmm1,%xmm7
 	movups	16(%ebp),%xmm1
+	leal	32(%ebp),%edx
 	pxor	16(%esp),%xmm3
-	pxor	32(%esp),%xmm4
 .byte	102,15,56,222,209
-	pxor	48(%esp),%xmm5
-	pxor	64(%esp),%xmm6
+	pxor	32(%esp),%xmm4
 .byte	102,15,56,222,217
-	pxor	%xmm0,%xmm7
-	movups	32(%ebp),%xmm0
+	pxor	48(%esp),%xmm5
+	decl	%ecx
 .byte	102,15,56,222,225
+	pxor	64(%esp),%xmm6
 .byte	102,15,56,222,233
+	pxor	%xmm0,%xmm7
 .byte	102,15,56,222,241
+	movups	(%edx),%xmm0
 .byte	102,15,56,222,249
 	call	.L_aesni_decrypt6_enter
 	movdqa	80(%esp),%xmm1
@@ -1569,25 +1465,26 @@ _aesni_xts_decrypt:
 	paddq	%xmm1,%xmm1
 	pand	%xmm3,%xmm2
 	pcmpgtd	%xmm1,%xmm0
+	movl	%ebx,%ecx
 	pxor	%xmm2,%xmm1
 	subl	$96,%eax
-	jnc	.L061xts_dec_loop6
-	movl	240(%ebp),%ecx
+	jnc	.L057xts_dec_loop6
+	leal	1(,%ecx,2),%ecx
 	movl	%ebp,%edx
 	movl	%ecx,%ebx
-.L060xts_dec_short:
+.L056xts_dec_short:
 	addl	$96,%eax
-	jz	.L062xts_dec_done6x
+	jz	.L058xts_dec_done6x
 	movdqa	%xmm1,%xmm5
 	cmpl	$32,%eax
-	jb	.L063xts_dec_one
+	jb	.L059xts_dec_one
 	pshufd	$19,%xmm0,%xmm2
 	pxor	%xmm0,%xmm0
 	paddq	%xmm1,%xmm1
 	pand	%xmm3,%xmm2
 	pcmpgtd	%xmm1,%xmm0
 	pxor	%xmm2,%xmm1
-	je	.L064xts_dec_two
+	je	.L060xts_dec_two
 	pshufd	$19,%xmm0,%xmm2
 	pxor	%xmm0,%xmm0
 	movdqa	%xmm1,%xmm6
@@ -1596,7 +1493,7 @@ _aesni_xts_decrypt:
 	pcmpgtd	%xmm1,%xmm0
 	pxor	%xmm2,%xmm1
 	cmpl	$64,%eax
-	jb	.L065xts_dec_three
+	jb	.L061xts_dec_three
 	pshufd	$19,%xmm0,%xmm2
 	pxor	%xmm0,%xmm0
 	movdqa	%xmm1,%xmm7
@@ -1606,7 +1503,7 @@ _aesni_xts_decrypt:
 	pxor	%xmm2,%xmm1
 	movdqa	%xmm5,(%esp)
 	movdqa	%xmm6,16(%esp)
-	je	.L066xts_dec_four
+	je	.L062xts_dec_four
 	movdqa	%xmm7,32(%esp)
 	pshufd	$19,%xmm0,%xmm7
 	movdqa	%xmm1,48(%esp)
@@ -1638,9 +1535,9 @@ _aesni_xts_decrypt:
 	movups	%xmm5,48(%edi)
 	movups	%xmm6,64(%edi)
 	leal	80(%edi),%edi
-	jmp	.L067xts_dec_done
+	jmp	.L063xts_dec_done
 .align	16
-.L063xts_dec_one:
+.L059xts_dec_one:
 	movups	(%esi),%xmm2
 	leal	16(%esi),%esi
 	xorps	%xmm5,%xmm2
@@ -1648,36 +1545,36 @@ _aesni_xts_decrypt:
 	movups	16(%edx),%xmm1
 	leal	32(%edx),%edx
 	xorps	%xmm0,%xmm2
-.L068dec1_loop_12:
+.L064dec1_loop_12:
 .byte	102,15,56,222,209
 	decl	%ecx
 	movups	(%edx),%xmm1
 	leal	16(%edx),%edx
-	jnz	.L068dec1_loop_12
+	jnz	.L064dec1_loop_12
 .byte	102,15,56,223,209
 	xorps	%xmm5,%xmm2
 	movups	%xmm2,(%edi)
 	leal	16(%edi),%edi
 	movdqa	%xmm5,%xmm1
-	jmp	.L067xts_dec_done
+	jmp	.L063xts_dec_done
 .align	16
-.L064xts_dec_two:
+.L060xts_dec_two:
 	movaps	%xmm1,%xmm6
 	movups	(%esi),%xmm2
 	movups	16(%esi),%xmm3
 	leal	32(%esi),%esi
 	xorps	%xmm5,%xmm2
 	xorps	%xmm6,%xmm3
-	call	__aesni_decrypt2
+	call	__aesni_decrypt3
 	xorps	%xmm5,%xmm2
 	xorps	%xmm6,%xmm3
 	movups	%xmm2,(%edi)
 	movups	%xmm3,16(%edi)
 	leal	32(%edi),%edi
 	movdqa	%xmm6,%xmm1
-	jmp	.L067xts_dec_done
+	jmp	.L063xts_dec_done
 .align	16
-.L065xts_dec_three:
+.L061xts_dec_three:
 	movaps	%xmm1,%xmm7
 	movups	(%esi),%xmm2
 	movups	16(%esi),%xmm3
@@ -1695,9 +1592,9 @@ _aesni_xts_decrypt:
 	movups	%xmm4,32(%edi)
 	leal	48(%edi),%edi
 	movdqa	%xmm7,%xmm1
-	jmp	.L067xts_dec_done
+	jmp	.L063xts_dec_done
 .align	16
-.L066xts_dec_four:
+.L062xts_dec_four:
 	movaps	%xmm1,%xmm6
 	movups	(%esi),%xmm2
 	movups	16(%esi),%xmm3
@@ -1719,20 +1616,20 @@ _aesni_xts_decrypt:
 	movups	%xmm5,48(%edi)
 	leal	64(%edi),%edi
 	movdqa	%xmm6,%xmm1
-	jmp	.L067xts_dec_done
+	jmp	.L063xts_dec_done
 .align	16
-.L062xts_dec_done6x:
+.L058xts_dec_done6x:
 	movl	112(%esp),%eax
 	andl	$15,%eax
-	jz	.L069xts_dec_ret
+	jz	.L065xts_dec_ret
 	movl	%eax,112(%esp)
-	jmp	.L070xts_dec_only_one_more
+	jmp	.L066xts_dec_only_one_more
 .align	16
-.L067xts_dec_done:
+.L063xts_dec_done:
 	movl	112(%esp),%eax
 	pxor	%xmm0,%xmm0
 	andl	$15,%eax
-	jz	.L069xts_dec_ret
+	jz	.L065xts_dec_ret
 	pcmpgtd	%xmm1,%xmm0
 	movl	%eax,112(%esp)
 	pshufd	$19,%xmm0,%xmm2
@@ -1742,7 +1639,7 @@ _aesni_xts_decrypt:
 	pand	%xmm3,%xmm2
 	pcmpgtd	%xmm1,%xmm0
 	pxor	%xmm2,%xmm1
-.L070xts_dec_only_one_more:
+.L066xts_dec_only_one_more:
 	pshufd	$19,%xmm0,%xmm5
 	movdqa	%xmm1,%xmm6
 	paddq	%xmm1,%xmm1
@@ -1756,16 +1653,16 @@ _aesni_xts_decrypt:
 	movups	16(%edx),%xmm1
 	leal	32(%edx),%edx
 	xorps	%xmm0,%xmm2
-.L071dec1_loop_13:
+.L067dec1_loop_13:
 .byte	102,15,56,222,209
 	decl	%ecx
 	movups	(%edx),%xmm1
 	leal	16(%edx),%edx
-	jnz	.L071dec1_loop_13
+	jnz	.L067dec1_loop_13
 .byte	102,15,56,223,209
 	xorps	%xmm5,%xmm2
 	movups	%xmm2,(%edi)
-.L072xts_dec_steal:
+.L068xts_dec_steal:
 	movzbl	16(%esi),%ecx
 	movzbl	(%edi),%edx
 	leal	1(%esi),%esi
@@ -1773,7 +1670,7 @@ _aesni_xts_decrypt:
 	movb	%dl,16(%edi)
 	leal	1(%edi),%edi
 	subl	$1,%eax
-	jnz	.L072xts_dec_steal
+	jnz	.L068xts_dec_steal
 	subl	112(%esp),%edi
 	movl	%ebp,%edx
 	movl	%ebx,%ecx
@@ -1783,30 +1680,16 @@ _aesni_xts_decrypt:
 	movups	16(%edx),%xmm1
 	leal	32(%edx),%edx
 	xorps	%xmm0,%xmm2
-.L073dec1_loop_14:
+.L069dec1_loop_14:
 .byte	102,15,56,222,209
 	decl	%ecx
 	movups	(%edx),%xmm1
 	leal	16(%edx),%edx
-	jnz	.L073dec1_loop_14
+	jnz	.L069dec1_loop_14
 .byte	102,15,56,223,209
 	xorps	%xmm6,%xmm2
 	movups	%xmm2,(%edi)
-.L069xts_dec_ret:
-	pxor	%xmm0,%xmm0
-	pxor	%xmm1,%xmm1
-	pxor	%xmm2,%xmm2
-	movdqa	%xmm0,(%esp)
-	pxor	%xmm3,%xmm3
-	movdqa	%xmm0,16(%esp)
-	pxor	%xmm4,%xmm4
-	movdqa	%xmm0,32(%esp)
-	pxor	%xmm5,%xmm5
-	movdqa	%xmm0,48(%esp)
-	pxor	%xmm6,%xmm6
-	movdqa	%xmm0,64(%esp)
-	pxor	%xmm7,%xmm7
-	movdqa	%xmm0,80(%esp)
+.L065xts_dec_ret:
 	movl	116(%esp),%esp
 	popl	%edi
 	popl	%esi
@@ -1831,7 +1714,7 @@ _aesni_cbc_encrypt:
 	movl	32(%esp),%edx
 	movl	36(%esp),%ebp
 	testl	%eax,%eax
-	jz	.L074cbc_abort
+	jz	.L070cbc_abort
 	cmpl	$0,40(%esp)
 	xchgl	%esp,%ebx
 	movups	(%ebp),%xmm7
@@ -1839,14 +1722,14 @@ _aesni_cbc_encrypt:
 	movl	%edx,%ebp
 	movl	%ebx,16(%esp)
 	movl	%ecx,%ebx
-	je	.L075cbc_decrypt
+	je	.L071cbc_decrypt
 	movaps	%xmm7,%xmm2
 	cmpl	$16,%eax
-	jb	.L076cbc_enc_tail
+	jb	.L072cbc_enc_tail
 	subl	$16,%eax
-	jmp	.L077cbc_enc_loop
+	jmp	.L073cbc_enc_loop
 .align	16
-.L077cbc_enc_loop:
+.L073cbc_enc_loop:
 	movups	(%esi),%xmm7
 	leal	16(%esi),%esi
 	movups	(%edx),%xmm0
@@ -1854,25 +1737,24 @@ _aesni_cbc_encrypt:
 	xorps	%xmm0,%xmm7
 	leal	32(%edx),%edx
 	xorps	%xmm7,%xmm2
-.L078enc1_loop_15:
+.L074enc1_loop_15:
 .byte	102,15,56,220,209
 	decl	%ecx
 	movups	(%edx),%xmm1
 	leal	16(%edx),%edx
-	jnz	.L078enc1_loop_15
+	jnz	.L074enc1_loop_15
 .byte	102,15,56,221,209
 	movl	%ebx,%ecx
 	movl	%ebp,%edx
 	movups	%xmm2,(%edi)
 	leal	16(%edi),%edi
 	subl	$16,%eax
-	jnc	.L077cbc_enc_loop
+	jnc	.L073cbc_enc_loop
 	addl	$16,%eax
-	jnz	.L076cbc_enc_tail
+	jnz	.L072cbc_enc_tail
 	movaps	%xmm2,%xmm7
-	pxor	%xmm2,%xmm2
-	jmp	.L079cbc_ret
-.L076cbc_enc_tail:
+	jmp	.L075cbc_ret
+.L072cbc_enc_tail:
 	movl	%eax,%ecx
 .long	2767451785
 	movl	$16,%ecx
@@ -1883,20 +1765,20 @@ _aesni_cbc_encrypt:
 	movl	%ebx,%ecx
 	movl	%edi,%esi
 	movl	%ebp,%edx
-	jmp	.L077cbc_enc_loop
+	jmp	.L073cbc_enc_loop
 .align	16
-.L075cbc_decrypt:
+.L071cbc_decrypt:
 	cmpl	$80,%eax
-	jbe	.L080cbc_dec_tail
+	jbe	.L076cbc_dec_tail
 	movaps	%xmm7,(%esp)
 	subl	$80,%eax
-	jmp	.L081cbc_dec_loop6_enter
+	jmp	.L077cbc_dec_loop6_enter
 .align	16
-.L082cbc_dec_loop6:
+.L078cbc_dec_loop6:
 	movaps	%xmm0,(%esp)
 	movups	%xmm7,(%edi)
 	leal	16(%edi),%edi
-.L081cbc_dec_loop6_enter:
+.L077cbc_dec_loop6_enter:
 	movdqu	(%esi),%xmm2
 	movdqu	16(%esi),%xmm3
 	movdqu	32(%esi),%xmm4
@@ -1926,28 +1808,28 @@ _aesni_cbc_encrypt:
 	movups	%xmm6,64(%edi)
 	leal	80(%edi),%edi
 	subl	$96,%eax
-	ja	.L082cbc_dec_loop6
+	ja	.L078cbc_dec_loop6
 	movaps	%xmm7,%xmm2
 	movaps	%xmm0,%xmm7
 	addl	$80,%eax
-	jle	.L083cbc_dec_clear_tail_collected
+	jle	.L079cbc_dec_tail_collected
 	movups	%xmm2,(%edi)
 	leal	16(%edi),%edi
-.L080cbc_dec_tail:
+.L076cbc_dec_tail:
 	movups	(%esi),%xmm2
 	movaps	%xmm2,%xmm6
 	cmpl	$16,%eax
-	jbe	.L084cbc_dec_one
+	jbe	.L080cbc_dec_one
 	movups	16(%esi),%xmm3
 	movaps	%xmm3,%xmm5
 	cmpl	$32,%eax
-	jbe	.L085cbc_dec_two
+	jbe	.L081cbc_dec_two
 	movups	32(%esi),%xmm4
 	cmpl	$48,%eax
-	jbe	.L086cbc_dec_three
+	jbe	.L082cbc_dec_three
 	movups	48(%esi),%xmm5
 	cmpl	$64,%eax
-	jbe	.L087cbc_dec_four
+	jbe	.L083cbc_dec_four
 	movups	64(%esi),%xmm6
 	movaps	%xmm7,(%esp)
 	movups	(%esi),%xmm2
@@ -1965,62 +1847,56 @@ _aesni_cbc_encrypt:
 	xorps	%xmm0,%xmm6
 	movups	%xmm2,(%edi)
 	movups	%xmm3,16(%edi)
-	pxor	%xmm3,%xmm3
 	movups	%xmm4,32(%edi)
-	pxor	%xmm4,%xmm4
 	movups	%xmm5,48(%edi)
-	pxor	%xmm5,%xmm5
 	leal	64(%edi),%edi
 	movaps	%xmm6,%xmm2
-	pxor	%xmm6,%xmm6
 	subl	$80,%eax
-	jmp	.L088cbc_dec_tail_collected
+	jmp	.L079cbc_dec_tail_collected
 .align	16
-.L084cbc_dec_one:
+.L080cbc_dec_one:
 	movups	(%edx),%xmm0
 	movups	16(%edx),%xmm1
 	leal	32(%edx),%edx
 	xorps	%xmm0,%xmm2
-.L089dec1_loop_16:
+.L084dec1_loop_16:
 .byte	102,15,56,222,209
 	decl	%ecx
 	movups	(%edx),%xmm1
 	leal	16(%edx),%edx
-	jnz	.L089dec1_loop_16
+	jnz	.L084dec1_loop_16
 .byte	102,15,56,223,209
 	xorps	%xmm7,%xmm2
 	movaps	%xmm6,%xmm7
 	subl	$16,%eax
-	jmp	.L088cbc_dec_tail_collected
+	jmp	.L079cbc_dec_tail_collected
 .align	16
-.L085cbc_dec_two:
-	call	__aesni_decrypt2
+.L081cbc_dec_two:
+	xorps	%xmm4,%xmm4
+	call	__aesni_decrypt3
 	xorps	%xmm7,%xmm2
 	xorps	%xmm6,%xmm3
 	movups	%xmm2,(%edi)
 	movaps	%xmm3,%xmm2
-	pxor	%xmm3,%xmm3
 	leal	16(%edi),%edi
 	movaps	%xmm5,%xmm7
 	subl	$32,%eax
-	jmp	.L088cbc_dec_tail_collected
+	jmp	.L079cbc_dec_tail_collected
 .align	16
-.L086cbc_dec_three:
+.L082cbc_dec_three:
 	call	__aesni_decrypt3
 	xorps	%xmm7,%xmm2
 	xorps	%xmm6,%xmm3
 	xorps	%xmm5,%xmm4
 	movups	%xmm2,(%edi)
 	movaps	%xmm4,%xmm2
-	pxor	%xmm4,%xmm4
 	movups	%xmm3,16(%edi)
-	pxor	%xmm3,%xmm3
 	leal	32(%edi),%edi
 	movups	32(%esi),%xmm7
 	subl	$48,%eax
-	jmp	.L088cbc_dec_tail_collected
+	jmp	.L079cbc_dec_tail_collected
 .align	16
-.L087cbc_dec_four:
+.L083cbc_dec_four:
 	call	__aesni_decrypt4
 	movups	16(%esi),%xmm1
 	movups	32(%esi),%xmm0
@@ -2030,44 +1906,28 @@ _aesni_cbc_encrypt:
 	movups	%xmm2,(%edi)
 	xorps	%xmm1,%xmm4
 	movups	%xmm3,16(%edi)
-	pxor	%xmm3,%xmm3
 	xorps	%xmm0,%xmm5
 	movups	%xmm4,32(%edi)
-	pxor	%xmm4,%xmm4
 	leal	48(%edi),%edi
 	movaps	%xmm5,%xmm2
-	pxor	%xmm5,%xmm5
 	subl	$64,%eax
-	jmp	.L088cbc_dec_tail_collected
-.align	16
-.L083cbc_dec_clear_tail_collected:
-	pxor	%xmm3,%xmm3
-	pxor	%xmm4,%xmm4
-	pxor	%xmm5,%xmm5
-	pxor	%xmm6,%xmm6
-.L088cbc_dec_tail_collected:
+.L079cbc_dec_tail_collected:
 	andl	$15,%eax
-	jnz	.L090cbc_dec_tail_partial
+	jnz	.L085cbc_dec_tail_partial
 	movups	%xmm2,(%edi)
-	pxor	%xmm0,%xmm0
-	jmp	.L079cbc_ret
+	jmp	.L075cbc_ret
 .align	16
-.L090cbc_dec_tail_partial:
+.L085cbc_dec_tail_partial:
 	movaps	%xmm2,(%esp)
-	pxor	%xmm0,%xmm0
 	movl	$16,%ecx
 	movl	%esp,%esi
 	subl	%eax,%ecx
 .long	2767451785
-	movdqa	%xmm2,(%esp)
-.L079cbc_ret:
+.L075cbc_ret:
 	movl	16(%esp),%esp
 	movl	36(%esp),%ebp
-	pxor	%xmm2,%xmm2
-	pxor	%xmm1,%xmm1
 	movups	%xmm7,(%ebp)
-	pxor	%xmm7,%xmm7
-.L074cbc_abort:
+.L070cbc_abort:
 	popl	%edi
 	popl	%esi
 	popl	%ebx
@@ -2076,62 +1936,52 @@ _aesni_cbc_encrypt:
 .def	__aesni_set_encrypt_key;	.scl	3;	.type	32;	.endef
 .align	16
 __aesni_set_encrypt_key:
-	pushl	%ebp
-	pushl	%ebx
 	testl	%eax,%eax
-	jz	.L091bad_pointer
+	jz	.L086bad_pointer
 	testl	%edx,%edx
-	jz	.L091bad_pointer
-	call	.L092pic
-.L092pic:
-	popl	%ebx
-	leal	.Lkey_const-.L092pic(%ebx),%ebx
-	leal	__gnutls_x86_cpuid_s,%ebp
+	jz	.L086bad_pointer
 	movups	(%eax),%xmm0
 	xorps	%xmm4,%xmm4
-	movl	4(%ebp),%ebp
 	leal	16(%edx),%edx
-	andl	$268437504,%ebp
 	cmpl	$256,%ecx
-	je	.L09314rounds
+	je	.L08714rounds
 	cmpl	$192,%ecx
-	je	.L09412rounds
+	je	.L08812rounds
 	cmpl	$128,%ecx
-	jne	.L095bad_keybits
+	jne	.L089bad_keybits
 .align	16
-.L09610rounds:
-	cmpl	$268435456,%ebp
-	je	.L09710rounds_alt
+.L09010rounds:
 	movl	$9,%ecx
 	movups	%xmm0,-16(%edx)
 .byte	102,15,58,223,200,1
-	call	.L098key_128_cold
+	call	.L091key_128_cold
 .byte	102,15,58,223,200,2
-	call	.L099key_128
+	call	.L092key_128
 .byte	102,15,58,223,200,4
-	call	.L099key_128
+	call	.L092key_128
 .byte	102,15,58,223,200,8
-	call	.L099key_128
+	call	.L092key_128
 .byte	102,15,58,223,200,16
-	call	.L099key_128
+	call	.L092key_128
 .byte	102,15,58,223,200,32
-	call	.L099key_128
+	call	.L092key_128
 .byte	102,15,58,223,200,64
-	call	.L099key_128
+	call	.L092key_128
 .byte	102,15,58,223,200,128
-	call	.L099key_128
+	call	.L092key_128
 .byte	102,15,58,223,200,27
-	call	.L099key_128
+	call	.L092key_128
 .byte	102,15,58,223,200,54
-	call	.L099key_128
+	call	.L092key_128
 	movups	%xmm0,(%edx)
 	movl	%ecx,80(%edx)
-	jmp	.L100good_key
+	xorl	%eax,%eax
+	ret
 .align	16
-.L099key_128:
+.L092key_128:
 	movups	%xmm0,(%edx)
 	leal	16(%edx),%edx
-.L098key_128_cold:
+.L091key_128_cold:
 	shufps	$16,%xmm0,%xmm4
 	xorps	%xmm4,%xmm0
 	shufps	$140,%xmm0,%xmm4
@@ -2140,91 +1990,38 @@ __aesni_set_encrypt_key:
 	xorps	%xmm1,%xmm0
 	ret
 .align	16
-.L09710rounds_alt:
-	movdqa	(%ebx),%xmm5
-	movl	$8,%ecx
-	movdqa	32(%ebx),%xmm4
-	movdqa	%xmm0,%xmm2
-	movdqu	%xmm0,-16(%edx)
-.L101loop_key128:
-.byte	102,15,56,0,197
-.byte	102,15,56,221,196
-	pslld	$1,%xmm4
-	leal	16(%edx),%edx
-	movdqa	%xmm2,%xmm3
-	pslldq	$4,%xmm2
-	pxor	%xmm2,%xmm3
-	pslldq	$4,%xmm2
-	pxor	%xmm2,%xmm3
-	pslldq	$4,%xmm2
-	pxor	%xmm3,%xmm2
-	pxor	%xmm2,%xmm0
-	movdqu	%xmm0,-16(%edx)
-	movdqa	%xmm0,%xmm2
-	decl	%ecx
-	jnz	.L101loop_key128
-	movdqa	48(%ebx),%xmm4
-.byte	102,15,56,0,197
-.byte	102,15,56,221,196
-	pslld	$1,%xmm4
-	movdqa	%xmm2,%xmm3
-	pslldq	$4,%xmm2
-	pxor	%xmm2,%xmm3
-	pslldq	$4,%xmm2
-	pxor	%xmm2,%xmm3
-	pslldq	$4,%xmm2
-	pxor	%xmm3,%xmm2
-	pxor	%xmm2,%xmm0
-	movdqu	%xmm0,(%edx)
-	movdqa	%xmm0,%xmm2
-.byte	102,15,56,0,197
-.byte	102,15,56,221,196
-	movdqa	%xmm2,%xmm3
-	pslldq	$4,%xmm2
-	pxor	%xmm2,%xmm3
-	pslldq	$4,%xmm2
-	pxor	%xmm2,%xmm3
-	pslldq	$4,%xmm2
-	pxor	%xmm3,%xmm2
-	pxor	%xmm2,%xmm0
-	movdqu	%xmm0,16(%edx)
-	movl	$9,%ecx
-	movl	%ecx,96(%edx)
-	jmp	.L100good_key
-.align	16
-.L09412rounds:
+.L08812rounds:
 	movq	16(%eax),%xmm2
-	cmpl	$268435456,%ebp
-	je	.L10212rounds_alt
 	movl	$11,%ecx
 	movups	%xmm0,-16(%edx)
 .byte	102,15,58,223,202,1
-	call	.L103key_192a_cold
+	call	.L093key_192a_cold
 .byte	102,15,58,223,202,2
-	call	.L104key_192b
+	call	.L094key_192b
 .byte	102,15,58,223,202,4
-	call	.L105key_192a
+	call	.L095key_192a
 .byte	102,15,58,223,202,8
-	call	.L104key_192b
+	call	.L094key_192b
 .byte	102,15,58,223,202,16
-	call	.L105key_192a
+	call	.L095key_192a
 .byte	102,15,58,223,202,32
-	call	.L104key_192b
+	call	.L094key_192b
 .byte	102,15,58,223,202,64
-	call	.L105key_192a
+	call	.L095key_192a
 .byte	102,15,58,223,202,128
-	call	.L104key_192b
+	call	.L094key_192b
 	movups	%xmm0,(%edx)
 	movl	%ecx,48(%edx)
-	jmp	.L100good_key
+	xorl	%eax,%eax
+	ret
 .align	16
-.L105key_192a:
+.L095key_192a:
 	movups	%xmm0,(%edx)
 	leal	16(%edx),%edx
 .align	16
-.L103key_192a_cold:
+.L093key_192a_cold:
 	movaps	%xmm2,%xmm5
-.L106key_192b_warm:
+.L096key_192b_warm:
 	shufps	$16,%xmm0,%xmm4
 	movdqa	%xmm2,%xmm3
 	xorps	%xmm4,%xmm0
@@ -2238,90 +2035,56 @@ __aesni_set_encrypt_key:
 	pxor	%xmm3,%xmm2
 	ret
 .align	16
-.L104key_192b:
+.L094key_192b:
 	movaps	%xmm0,%xmm3
 	shufps	$68,%xmm0,%xmm5
 	movups	%xmm5,(%edx)
 	shufps	$78,%xmm2,%xmm3
 	movups	%xmm3,16(%edx)
 	leal	32(%edx),%edx
-	jmp	.L106key_192b_warm
-.align	16
-.L10212rounds_alt:
-	movdqa	16(%ebx),%xmm5
-	movdqa	32(%ebx),%xmm4
-	movl	$8,%ecx
-	movdqu	%xmm0,-16(%edx)
-.L107loop_key192:
-	movq	%xmm2,(%edx)
-	movdqa	%xmm2,%xmm1
-.byte	102,15,56,0,213
-.byte	102,15,56,221,212
-	pslld	$1,%xmm4
-	leal	24(%edx),%edx
-	movdqa	%xmm0,%xmm3
-	pslldq	$4,%xmm0
-	pxor	%xmm0,%xmm3
-	pslldq	$4,%xmm0
-	pxor	%xmm0,%xmm3
-	pslldq	$4,%xmm0
-	pxor	%xmm3,%xmm0
-	pshufd	$255,%xmm0,%xmm3
-	pxor	%xmm1,%xmm3
-	pslldq	$4,%xmm1
-	pxor	%xmm1,%xmm3
-	pxor	%xmm2,%xmm0
-	pxor	%xmm3,%xmm2
-	movdqu	%xmm0,-16(%edx)
-	decl	%ecx
-	jnz	.L107loop_key192
-	movl	$11,%ecx
-	movl	%ecx,32(%edx)
-	jmp	.L100good_key
+	jmp	.L096key_192b_warm
 .align	16
-.L09314rounds:
+.L08714rounds:
 	movups	16(%eax),%xmm2
-	leal	16(%edx),%edx
-	cmpl	$268435456,%ebp
-	je	.L10814rounds_alt
 	movl	$13,%ecx
+	leal	16(%edx),%edx
 	movups	%xmm0,-32(%edx)
 	movups	%xmm2,-16(%edx)
 .byte	102,15,58,223,202,1
-	call	.L109key_256a_cold
+	call	.L097key_256a_cold
 .byte	102,15,58,223,200,1
-	call	.L110key_256b
+	call	.L098key_256b
 .byte	102,15,58,223,202,2
-	call	.L111key_256a
+	call	.L099key_256a
 .byte	102,15,58,223,200,2
-	call	.L110key_256b
+	call	.L098key_256b
 .byte	102,15,58,223,202,4
-	call	.L111key_256a
+	call	.L099key_256a
 .byte	102,15,58,223,200,4
-	call	.L110key_256b
+	call	.L098key_256b
 .byte	102,15,58,223,202,8
-	call	.L111key_256a
+	call	.L099key_256a
 .byte	102,15,58,223,200,8
-	call	.L110key_256b
+	call	.L098key_256b
 .byte	102,15,58,223,202,16
-	call	.L111key_256a
+	call	.L099key_256a
 .byte	102,15,58,223,200,16
-	call	.L110key_256b
+	call	.L098key_256b
 .byte	102,15,58,223,202,32
-	call	.L111key_256a
+	call	.L099key_256a
 .byte	102,15,58,223,200,32
-	call	.L110key_256b
+	call	.L098key_256b
 .byte	102,15,58,223,202,64
-	call	.L111key_256a
+	call	.L099key_256a
 	movups	%xmm0,(%edx)
 	movl	%ecx,16(%edx)
 	xorl	%eax,%eax
-	jmp	.L100good_key
+	ret
 .align	16
-.L111key_256a:
+.L099key_256a:
 	movups	%xmm2,(%edx)
 	leal	16(%edx),%edx
-.L109key_256a_cold:
+.L097key_256a_cold:
 	shufps	$16,%xmm0,%xmm4
 	xorps	%xmm4,%xmm0
 	shufps	$140,%xmm0,%xmm4
@@ -2330,7 +2093,7 @@ __aesni_set_encrypt_key:
 	xorps	%xmm1,%xmm0
 	ret
 .align	16
-.L110key_256b:
+.L098key_256b:
 	movups	%xmm0,(%edx)
 	leal	16(%edx),%edx
 	shufps	$16,%xmm2,%xmm4
@@ -2340,70 +2103,13 @@ __aesni_set_encrypt_key:
 	shufps	$170,%xmm1,%xmm1
 	xorps	%xmm1,%xmm2
 	ret
-.align	16
-.L10814rounds_alt:
-	movdqa	(%ebx),%xmm5
-	movdqa	32(%ebx),%xmm4
-	movl	$7,%ecx
-	movdqu	%xmm0,-32(%edx)
-	movdqa	%xmm2,%xmm1
-	movdqu	%xmm2,-16(%edx)
-.L112loop_key256:
-.byte	102,15,56,0,213
-.byte	102,15,56,221,212
-	movdqa	%xmm0,%xmm3
-	pslldq	$4,%xmm0
-	pxor	%xmm0,%xmm3
-	pslldq	$4,%xmm0
-	pxor	%xmm0,%xmm3
-	pslldq	$4,%xmm0
-	pxor	%xmm3,%xmm0
-	pslld	$1,%xmm4
-	pxor	%xmm2,%xmm0
-	movdqu	%xmm0,(%edx)
-	decl	%ecx
-	jz	.L113done_key256
-	pshufd	$255,%xmm0,%xmm2
-	pxor	%xmm3,%xmm3
-.byte	102,15,56,221,211
-	movdqa	%xmm1,%xmm3
-	pslldq	$4,%xmm1
-	pxor	%xmm1,%xmm3
-	pslldq	$4,%xmm1
-	pxor	%xmm1,%xmm3
-	pslldq	$4,%xmm1
-	pxor	%xmm3,%xmm1
-	pxor	%xmm1,%xmm2
-	movdqu	%xmm2,16(%edx)
-	leal	32(%edx),%edx
-	movdqa	%xmm2,%xmm1
-	jmp	.L112loop_key256
-.L113done_key256:
-	movl	$13,%ecx
-	movl	%ecx,16(%edx)
-.L100good_key:
-	pxor	%xmm0,%xmm0
-	pxor	%xmm1,%xmm1
-	pxor	%xmm2,%xmm2
-	pxor	%xmm3,%xmm3
-	pxor	%xmm4,%xmm4
-	pxor	%xmm5,%xmm5
-	xorl	%eax,%eax
-	popl	%ebx
-	popl	%ebp
-	ret
 .align	4
-.L091bad_pointer:
+.L086bad_pointer:
 	movl	$-1,%eax
-	popl	%ebx
-	popl	%ebp
 	ret
 .align	4
-.L095bad_keybits:
-	pxor	%xmm0,%xmm0
+.L089bad_keybits:
 	movl	$-2,%eax
-	popl	%ebx
-	popl	%ebp
 	ret
 .globl	_aesni_set_encrypt_key
 .def	_aesni_set_encrypt_key;	.scl	2;	.type	32;	.endef
@@ -2427,7 +2133,7 @@ _aesni_set_decrypt_key:
 	movl	12(%esp),%edx
 	shll	$4,%ecx
 	testl	%eax,%eax
-	jnz	.L114dec_key_ret
+	jnz	.L100dec_key_ret
 	leal	16(%edx,%ecx,1),%eax
 	movups	(%edx),%xmm0
 	movups	(%eax),%xmm1
@@ -2435,7 +2141,7 @@ _aesni_set_decrypt_key:
 	movups	%xmm1,(%edx)
 	leal	16(%edx),%edx
 	leal	-16(%eax),%eax
-.L115dec_key_inverse:
+.L101dec_key_inverse:
 	movups	(%edx),%xmm0
 	movups	(%eax),%xmm1
 .byte	102,15,56,219,192
@@ -2445,24 +2151,15 @@ _aesni_set_decrypt_key:
 	movups	%xmm0,16(%eax)
 	movups	%xmm1,-16(%edx)
 	cmpl	%edx,%eax
-	ja	.L115dec_key_inverse
+	ja	.L101dec_key_inverse
 	movups	(%edx),%xmm0
 .byte	102,15,56,219,192
 	movups	%xmm0,(%edx)
-	pxor	%xmm0,%xmm0
-	pxor	%xmm1,%xmm1
 	xorl	%eax,%eax
-.L114dec_key_ret:
+.L100dec_key_ret:
 	ret
-.align	64
-.Lkey_const:
-.long	202313229,202313229,202313229,202313229
-.long	67569157,67569157,67569157,67569157
-.long	1,1,1,1
-.long	27,27,27,27
 .byte	65,69,83,32,102,111,114,32,73,110,116,101,108,32,65,69
 .byte	83,45,78,73,44,32,67,82,89,80,84,79,71,65,77,83
 .byte	32,98,121,32,60,97,112,112,114,111,64,111,112,101,110,115
 .byte	115,108,46,111,114,103,62,0
-.comm	__gnutls_x86_cpuid_s,16
 
diff --git a/lib/accelerated/x86/elf/aesni-x86.s b/lib/accelerated/x86/elf/aesni-x86.s
index 73d623cbda..5d70f2568f 100644
--- a/lib/accelerated/x86/elf/aesni-x86.s
+++ b/lib/accelerated/x86/elf/aesni-x86.s
@@ -60,10 +60,7 @@ aesni_encrypt:
 	leal	16(%edx),%edx
 	jnz	.L000enc1_loop_1
 .byte	102,15,56,221,209
-	pxor	%xmm0,%xmm0
-	pxor	%xmm1,%xmm1
 	movups	%xmm2,(%eax)
-	pxor	%xmm2,%xmm2
 	ret
 .size	aesni_encrypt,.-.L_aesni_encrypt_begin
 .globl	aesni_decrypt
@@ -87,90 +84,32 @@ aesni_decrypt:
 	leal	16(%edx),%edx
 	jnz	.L001dec1_loop_2
 .byte	102,15,56,223,209
-	pxor	%xmm0,%xmm0
-	pxor	%xmm1,%xmm1
 	movups	%xmm2,(%eax)
-	pxor	%xmm2,%xmm2
 	ret
 .size	aesni_decrypt,.-.L_aesni_decrypt_begin
-.type	_aesni_encrypt2,@function
-.align	16
-_aesni_encrypt2:
-	movups	(%edx),%xmm0
-	shll	$4,%ecx
-	movups	16(%edx),%xmm1
-	xorps	%xmm0,%xmm2
-	pxor	%xmm0,%xmm3
-	movups	32(%edx),%xmm0
-	leal	32(%edx,%ecx,1),%edx
-	negl	%ecx
-	addl	$16,%ecx
-.L002enc2_loop:
-.byte	102,15,56,220,209
-.byte	102,15,56,220,217
-	movups	(%edx,%ecx,1),%xmm1
-	addl	$32,%ecx
-.byte	102,15,56,220,208
-.byte	102,15,56,220,216
-	movups	-16(%edx,%ecx,1),%xmm0
-	jnz	.L002enc2_loop
-.byte	102,15,56,220,209
-.byte	102,15,56,220,217
-.byte	102,15,56,221,208
-.byte	102,15,56,221,216
-	ret
-.size	_aesni_encrypt2,.-_aesni_encrypt2
-.type	_aesni_decrypt2,@function
-.align	16
-_aesni_decrypt2:
-	movups	(%edx),%xmm0
-	shll	$4,%ecx
-	movups	16(%edx),%xmm1
-	xorps	%xmm0,%xmm2
-	pxor	%xmm0,%xmm3
-	movups	32(%edx),%xmm0
-	leal	32(%edx,%ecx,1),%edx
-	negl	%ecx
-	addl	$16,%ecx
-.L003dec2_loop:
-.byte	102,15,56,222,209
-.byte	102,15,56,222,217
-	movups	(%edx,%ecx,1),%xmm1
-	addl	$32,%ecx
-.byte	102,15,56,222,208
-.byte	102,15,56,222,216
-	movups	-16(%edx,%ecx,1),%xmm0
-	jnz	.L003dec2_loop
-.byte	102,15,56,222,209
-.byte	102,15,56,222,217
-.byte	102,15,56,223,208
-.byte	102,15,56,223,216
-	ret
-.size	_aesni_decrypt2,.-_aesni_decrypt2
 .type	_aesni_encrypt3,@function
 .align	16
 _aesni_encrypt3:
 	movups	(%edx),%xmm0
-	shll	$4,%ecx
+	shrl	$1,%ecx
 	movups	16(%edx),%xmm1
+	leal	32(%edx),%edx
 	xorps	%xmm0,%xmm2
 	pxor	%xmm0,%xmm3
 	pxor	%xmm0,%xmm4
-	movups	32(%edx),%xmm0
-	leal	32(%edx,%ecx,1),%edx
-	negl	%ecx
-	addl	$16,%ecx
-.L004enc3_loop:
+	movups	(%edx),%xmm0
+.L002enc3_loop:
 .byte	102,15,56,220,209
 .byte	102,15,56,220,217
+	decl	%ecx
 .byte	102,15,56,220,225
-	movups	(%edx,%ecx,1),%xmm1
-	addl	$32,%ecx
+	movups	16(%edx),%xmm1
 .byte	102,15,56,220,208
 .byte	102,15,56,220,216
+	leal	32(%edx),%edx
 .byte	102,15,56,220,224
-	movups	-16(%edx,%ecx,1),%xmm0
-	jnz	.L004enc3_loop
+	movups	(%edx),%xmm0
+	jnz	.L002enc3_loop
 .byte	102,15,56,220,209
 .byte	102,15,56,220,217
 .byte	102,15,56,220,225
@@ -183,26 +122,25 @@ _aesni_encrypt3:
 .align	16
 _aesni_decrypt3:
 	movups	(%edx),%xmm0
-	shll	$4,%ecx
+	shrl	$1,%ecx
 	movups	16(%edx),%xmm1
+	leal	32(%edx),%edx
 	xorps	%xmm0,%xmm2
 	pxor	%xmm0,%xmm3
 	pxor	%xmm0,%xmm4
-	movups	32(%edx),%xmm0
-	leal	32(%edx,%ecx,1),%edx
-	negl	%ecx
-	addl	$16,%ecx
-.L005dec3_loop:
+	movups	(%edx),%xmm0
+.L003dec3_loop:
 .byte	102,15,56,222,209
 .byte	102,15,56,222,217
+	decl	%ecx
 .byte	102,15,56,222,225
-	movups	(%edx,%ecx,1),%xmm1
-	addl	$32,%ecx
+	movups	16(%edx),%xmm1
 .byte	102,15,56,222,208
 .byte	102,15,56,222,216
+	leal	32(%edx),%edx
 .byte	102,15,56,222,224
-	movups	-16(%edx,%ecx,1),%xmm0
-	jnz	.L005dec3_loop
+	movups	(%edx),%xmm0
+	jnz	.L003dec3_loop
 .byte	102,15,56,222,209
 .byte	102,15,56,222,217
 .byte	102,15,56,222,225
@@ -216,29 +154,27 @@ _aesni_decrypt3:
 _aesni_encrypt4:
 	movups	(%edx),%xmm0
 	movups	16(%edx),%xmm1
-	shll	$4,%ecx
+	shrl	$1,%ecx
+	leal	32(%edx),%edx
 	xorps	%xmm0,%xmm2
 	pxor	%xmm0,%xmm3
 	pxor	%xmm0,%xmm4
 	pxor	%xmm0,%xmm5
-	movups	32(%edx),%xmm0
-	leal	32(%edx,%ecx,1),%edx
-	negl	%ecx
-.byte	15,31,64,0
-	addl	$16,%ecx
-.L006enc4_loop:
+	movups	(%edx),%xmm0
+.L004enc4_loop:
 .byte	102,15,56,220,209
 .byte	102,15,56,220,217
+	decl	%ecx
 .byte	102,15,56,220,225
 .byte	102,15,56,220,233
-	movups	(%edx,%ecx,1),%xmm1
-	addl	$32,%ecx
+	movups	16(%edx),%xmm1
 .byte	102,15,56,220,208
 .byte	102,15,56,220,216
+	leal	32(%edx),%edx
 .byte	102,15,56,220,224
 .byte	102,15,56,220,232
-	movups	-16(%edx,%ecx,1),%xmm0
-	jnz	.L006enc4_loop
+	movups	(%edx),%xmm0
+	jnz	.L004enc4_loop
 .byte	102,15,56,220,209
 .byte	102,15,56,220,217
 .byte	102,15,56,220,225
@@ -254,29 +190,27 @@ _aesni_encrypt4:
 _aesni_decrypt4:
 	movups	(%edx),%xmm0
 	movups	16(%edx),%xmm1
-	shll	$4,%ecx
+	shrl	$1,%ecx
+	leal	32(%edx),%edx
 	xorps	%xmm0,%xmm2
 	pxor	%xmm0,%xmm3
 	pxor	%xmm0,%xmm4
 	pxor	%xmm0,%xmm5
-	movups	32(%edx),%xmm0
-	leal	32(%edx,%ecx,1),%edx
-	negl	%ecx
-.byte	15,31,64,0
-	addl	$16,%ecx
-.L007dec4_loop:
+	movups	(%edx),%xmm0
+.L005dec4_loop:
 .byte	102,15,56,222,209
 .byte	102,15,56,222,217
+	decl	%ecx
 .byte	102,15,56,222,225
 .byte	102,15,56,222,233
-	movups	(%edx,%ecx,1),%xmm1
-	addl	$32,%ecx
+	movups	16(%edx),%xmm1
 .byte	102,15,56,222,208
 .byte	102,15,56,222,216
+	leal	32(%edx),%edx
 .byte	102,15,56,222,224
 .byte	102,15,56,222,232
-	movups	-16(%edx,%ecx,1),%xmm0
-	jnz	.L007dec4_loop
+	movups	(%edx),%xmm0
+	jnz	.L005dec4_loop
 .byte	102,15,56,222,209
 .byte	102,15,56,222,217
 .byte	102,15,56,222,225
@@ -291,42 +225,45 @@ _aesni_decrypt4:
 .align	16
 _aesni_encrypt6:
 	movups	(%edx),%xmm0
-	shll	$4,%ecx
+	shrl	$1,%ecx
 	movups	16(%edx),%xmm1
+	leal	32(%edx),%edx
 	xorps	%xmm0,%xmm2
 	pxor	%xmm0,%xmm3
-	pxor	%xmm0,%xmm4
 .byte	102,15,56,220,209
-	pxor	%xmm0,%xmm5
-	pxor	%xmm0,%xmm6
+	pxor	%xmm0,%xmm4
 .byte	102,15,56,220,217
-	leal	32(%edx,%ecx,1),%edx
-	negl	%ecx
+	pxor	%xmm0,%xmm5
+	decl	%ecx
 .byte	102,15,56,220,225
+	pxor	%xmm0,%xmm6
+.byte	102,15,56,220,233
 	pxor	%xmm0,%xmm7
-	movups	(%edx,%ecx,1),%xmm0
-	addl	$16,%ecx
-	jmp	.L008_aesni_encrypt6_inner
+.byte	102,15,56,220,241
+	movups	(%edx),%xmm0
+.byte	102,15,56,220,249
+	jmp	.L_aesni_encrypt6_enter
 .align	16
-.L009enc6_loop:
+.L006enc6_loop:
 .byte	102,15,56,220,209
 .byte	102,15,56,220,217
+	decl	%ecx
 .byte	102,15,56,220,225
-.L008_aesni_encrypt6_inner:
 .byte	102,15,56,220,233
 .byte	102,15,56,220,241
 .byte	102,15,56,220,249
+.align	16
 .L_aesni_encrypt6_enter:
-	movups	(%edx,%ecx,1),%xmm1
-	addl	$32,%ecx
+	movups	16(%edx),%xmm1
 .byte	102,15,56,220,208
 .byte	102,15,56,220,216
+	leal	32(%edx),%edx
 .byte	102,15,56,220,224
 .byte	102,15,56,220,232
 .byte	102,15,56,220,240
 .byte	102,15,56,220,248
-	movups	-16(%edx,%ecx,1),%xmm0
-	jnz	.L009enc6_loop
+	movups	(%edx),%xmm0
+	jnz	.L006enc6_loop
 .byte	102,15,56,220,209
 .byte	102,15,56,220,217
 .byte	102,15,56,220,225
@@ -345,42 +282,45 @@ _aesni_encrypt6:
 .align	16
 _aesni_decrypt6:
 	movups	(%edx),%xmm0
-	shll	$4,%ecx
+	shrl	$1,%ecx
 	movups	16(%edx),%xmm1
+	leal	32(%edx),%edx
 	xorps	%xmm0,%xmm2
 	pxor	%xmm0,%xmm3
-	pxor	%xmm0,%xmm4
 .byte	102,15,56,222,209
-	pxor	%xmm0,%xmm5
-	pxor	%xmm0,%xmm6
+	pxor	%xmm0,%xmm4
 .byte	102,15,56,222,217
-	leal	32(%edx,%ecx,1),%edx
-	negl	%ecx
+	pxor	%xmm0,%xmm5
+	decl	%ecx
 .byte	102,15,56,222,225
+	pxor	%xmm0,%xmm6
+.byte	102,15,56,222,233
 	pxor	%xmm0,%xmm7
-	movups	(%edx,%ecx,1),%xmm0
-	addl	$16,%ecx
-	jmp	.L010_aesni_decrypt6_inner
+.byte	102,15,56,222,241
+	movups	(%edx),%xmm0
+.byte	102,15,56,222,249
+	jmp	.L_aesni_decrypt6_enter
 .align	16
-.L011dec6_loop:
+.L007dec6_loop:
 .byte	102,15,56,222,209
 .byte	102,15,56,222,217
+	decl	%ecx
 .byte	102,15,56,222,225
-.L010_aesni_decrypt6_inner:
 .byte	102,15,56,222,233
 .byte	102,15,56,222,241
 .byte	102,15,56,222,249
+.align	16
 .L_aesni_decrypt6_enter:
-	movups	(%edx,%ecx,1),%xmm1
-	addl	$32,%ecx
+	movups	16(%edx),%xmm1
 .byte	102,15,56,222,208
 .byte	102,15,56,222,216
+	leal	32(%edx),%edx
 .byte	102,15,56,222,224
 .byte	102,15,56,222,232
 .byte	102,15,56,222,240
 .byte	102,15,56,222,248
-	movups	-16(%edx,%ecx,1),%xmm0
-	jnz	.L011dec6_loop
+	movups	(%edx),%xmm0
+	jnz	.L007dec6_loop
 .byte	102,15,56,222,209
 .byte	102,15,56,222,217
 .byte	102,15,56,222,225
@@ -410,14 +350,14 @@ aesni_ecb_encrypt:
 	movl	32(%esp),%edx
 	movl	36(%esp),%ebx
 	andl	$-16,%eax
-	jz	.L012ecb_ret
+	jz	.L008ecb_ret
 	movl	240(%edx),%ecx
 	testl	%ebx,%ebx
-	jz	.L013ecb_decrypt
+	jz	.L009ecb_decrypt
 	movl	%edx,%ebp
 	movl	%ecx,%ebx
 	cmpl	$96,%eax
-	jb	.L014ecb_enc_tail
+	jb	.L010ecb_enc_tail
 	movdqu	(%esi),%xmm2
 	movdqu	16(%esi),%xmm3
 	movdqu	32(%esi),%xmm4
@@ -426,9 +366,9 @@ aesni_ecb_encrypt:
 	movdqu	80(%esi),%xmm7
 	leal	96(%esi),%esi
 	subl	$96,%eax
-	jmp	.L015ecb_enc_loop6_enter
+	jmp	.L011ecb_enc_loop6_enter
 .align	16
-.L016ecb_enc_loop6:
+.L012ecb_enc_loop6:
 	movups	%xmm2,(%edi)
 	movdqu	(%esi),%xmm2
 	movups	%xmm3,16(%edi)
@@ -443,12 +383,12 @@ aesni_ecb_encrypt:
 	leal	96(%edi),%edi
 	movdqu	80(%esi),%xmm7
 	leal	96(%esi),%esi
-.L015ecb_enc_loop6_enter:
+.L011ecb_enc_loop6_enter:
 	call	_aesni_encrypt6
 	movl	%ebp,%edx
 	movl	%ebx,%ecx
 	subl	$96,%eax
-	jnc	.L016ecb_enc_loop6
+	jnc	.L012ecb_enc_loop6
 	movups	%xmm2,(%edi)
 	movups	%xmm3,16(%edi)
 	movups	%xmm4,32(%edi)
@@ -457,18 +397,18 @@ aesni_ecb_encrypt:
 	movups	%xmm7,80(%edi)
 	leal	96(%edi),%edi
 	addl	$96,%eax
-	jz	.L012ecb_ret
-.L014ecb_enc_tail:
+	jz	.L008ecb_ret
+.L010ecb_enc_tail:
 	movups	(%esi),%xmm2
 	cmpl	$32,%eax
-	jb	.L017ecb_enc_one
+	jb	.L013ecb_enc_one
 	movups	16(%esi),%xmm3
-	je	.L018ecb_enc_two
+	je	.L014ecb_enc_two
 	movups	32(%esi),%xmm4
 	cmpl	$64,%eax
-	jb	.L019ecb_enc_three
+	jb	.L015ecb_enc_three
 	movups	48(%esi),%xmm5
-	je	.L020ecb_enc_four
+	je	.L016ecb_enc_four
 	movups	64(%esi),%xmm6
 	xorps	%xmm7,%xmm7
 	call	_aesni_encrypt6
@@ -477,49 +417,50 @@ aesni_ecb_encrypt:
 	movups	%xmm4,32(%edi)
 	movups	%xmm5,48(%edi)
 	movups	%xmm6,64(%edi)
-	jmp	.L012ecb_ret
+	jmp	.L008ecb_ret
 .align	16
-.L017ecb_enc_one:
+.L013ecb_enc_one:
 	movups	(%edx),%xmm0
 	movups	16(%edx),%xmm1
 	leal	32(%edx),%edx
 	xorps	%xmm0,%xmm2
-.L021enc1_loop_3:
+.L017enc1_loop_3:
 .byte	102,15,56,220,209
 	decl	%ecx
 	movups	(%edx),%xmm1
 	leal	16(%edx),%edx
-	jnz	.L021enc1_loop_3
+	jnz	.L017enc1_loop_3
 .byte	102,15,56,221,209
 	movups	%xmm2,(%edi)
-	jmp	.L012ecb_ret
+	jmp	.L008ecb_ret
 .align	16
-.L018ecb_enc_two:
-	call	_aesni_encrypt2
+.L014ecb_enc_two:
+	xorps	%xmm4,%xmm4
+	call	_aesni_encrypt3
 	movups	%xmm2,(%edi)
 	movups	%xmm3,16(%edi)
-	jmp	.L012ecb_ret
+	jmp	.L008ecb_ret
 .align	16
-.L019ecb_enc_three:
+.L015ecb_enc_three:
 	call	_aesni_encrypt3
 	movups	%xmm2,(%edi)
 	movups	%xmm3,16(%edi)
 	movups	%xmm4,32(%edi)
-	jmp	.L012ecb_ret
+	jmp	.L008ecb_ret
 .align	16
-.L020ecb_enc_four:
+.L016ecb_enc_four:
 	call	_aesni_encrypt4
 	movups	%xmm2,(%edi)
 	movups	%xmm3,16(%edi)
 	movups	%xmm4,32(%edi)
 	movups	%xmm5,48(%edi)
-	jmp	.L012ecb_ret
+	jmp	.L008ecb_ret
 .align	16
-.L013ecb_decrypt:
+.L009ecb_decrypt:
 	movl	%edx,%ebp
 	movl	%ecx,%ebx
 	cmpl	$96,%eax
-	jb	.L022ecb_dec_tail
+	jb	.L018ecb_dec_tail
 	movdqu	(%esi),%xmm2
 	movdqu	16(%esi),%xmm3
 	movdqu	32(%esi),%xmm4
@@ -528,9 +469,9 @@ aesni_ecb_encrypt:
 	movdqu	80(%esi),%xmm7
 	leal	96(%esi),%esi
 	subl	$96,%eax
-	jmp	.L023ecb_dec_loop6_enter
+	jmp	.L019ecb_dec_loop6_enter
 .align	16
-.L024ecb_dec_loop6:
+.L020ecb_dec_loop6:
 	movups	%xmm2,(%edi)
 	movdqu	(%esi),%xmm2
 	movups	%xmm3,16(%edi)
@@ -545,12 +486,12 @@ aesni_ecb_encrypt:
 	leal	96(%edi),%edi
 	movdqu	80(%esi),%xmm7
 	leal	96(%esi),%esi
-.L023ecb_dec_loop6_enter:
+.L019ecb_dec_loop6_enter:
 	call	_aesni_decrypt6
 	movl	%ebp,%edx
 	movl	%ebx,%ecx
 	subl	$96,%eax
-	jnc	.L024ecb_dec_loop6
+	jnc	.L020ecb_dec_loop6
 	movups	%xmm2,(%edi)
 	movups	%xmm3,16(%edi)
 	movups	%xmm4,32(%edi)
@@ -559,18 +500,18 @@ aesni_ecb_encrypt:
 	movups	%xmm7,80(%edi)
 	leal	96(%edi),%edi
 	addl	$96,%eax
-	jz	.L012ecb_ret
-.L022ecb_dec_tail:
+	jz	.L008ecb_ret
+.L018ecb_dec_tail:
 	movups	(%esi),%xmm2
 	cmpl	$32,%eax
-	jb	.L025ecb_dec_one
+	jb	.L021ecb_dec_one
 	movups	16(%esi),%xmm3
-	je	.L026ecb_dec_two
+	je	.L022ecb_dec_two
 	movups	32(%esi),%xmm4
 	cmpl	$64,%eax
-	jb	.L027ecb_dec_three
+	jb	.L023ecb_dec_three
 	movups	48(%esi),%xmm5
-	je	.L028ecb_dec_four
+	je	.L024ecb_dec_four
 	movups	64(%esi),%xmm6
 	xorps	%xmm7,%xmm7
 	call	_aesni_decrypt6
@@ -579,51 +520,44 @@ aesni_ecb_encrypt:
 	movups	%xmm4,32(%edi)
 	movups	%xmm5,48(%edi)
 	movups	%xmm6,64(%edi)
-	jmp	.L012ecb_ret
+	jmp	.L008ecb_ret
 .align	16
-.L025ecb_dec_one:
+.L021ecb_dec_one:
 	movups	(%edx),%xmm0
 	movups	16(%edx),%xmm1
 	leal	32(%edx),%edx
 	xorps	%xmm0,%xmm2
-.L029dec1_loop_4:
+.L025dec1_loop_4:
 .byte	102,15,56,222,209
 	decl	%ecx
 	movups	(%edx),%xmm1
 	leal	16(%edx),%edx
-	jnz	.L029dec1_loop_4
+	jnz	.L025dec1_loop_4
 .byte	102,15,56,223,209
 	movups	%xmm2,(%edi)
-	jmp	.L012ecb_ret
+	jmp	.L008ecb_ret
 .align	16
-.L026ecb_dec_two:
-	call	_aesni_decrypt2
+.L022ecb_dec_two:
+	xorps	%xmm4,%xmm4
+	call	_aesni_decrypt3
 	movups	%xmm2,(%edi)
 	movups	%xmm3,16(%edi)
-	jmp	.L012ecb_ret
+	jmp	.L008ecb_ret
 .align	16
-.L027ecb_dec_three:
+.L023ecb_dec_three:
 	call	_aesni_decrypt3
 	movups	%xmm2,(%edi)
 	movups	%xmm3,16(%edi)
 	movups	%xmm4,32(%edi)
-	jmp	.L012ecb_ret
+	jmp	.L008ecb_ret
 .align	16
-.L028ecb_dec_four:
+.L024ecb_dec_four:
 	call	_aesni_decrypt4
 	movups	%xmm2,(%edi)
 	movups	%xmm3,16(%edi)
 	movups	%xmm4,32(%edi)
 	movups	%xmm5,48(%edi)
-.L012ecb_ret:
-	pxor	%xmm0,%xmm0
-	pxor	%xmm1,%xmm1
-	pxor	%xmm2,%xmm2
-	pxor	%xmm3,%xmm3
-	pxor	%xmm4,%xmm4
-	pxor	%xmm5,%xmm5
-	pxor	%xmm6,%xmm6
-	pxor	%xmm7,%xmm7
+.L008ecb_ret:
 	popl	%edi
 	popl	%esi
 	popl	%ebx
@@ -662,56 +596,48 @@ aesni_ccm64_encrypt_blocks:
 	movl	%ebp,20(%esp)
 	movl	%ebp,24(%esp)
 	movl	%ebp,28(%esp)
-	shll	$4,%ecx
-	movl	$16,%ebx
+	shrl	$1,%ecx
 	leal	(%edx),%ebp
 	movdqa	(%esp),%xmm5
 	movdqa	%xmm7,%xmm2
-	leal	32(%edx,%ecx,1),%edx
-	subl	%ecx,%ebx
+	movl	%ecx,%ebx
 .byte	102,15,56,0,253
-.L030ccm64_enc_outer:
+.L026ccm64_enc_outer:
 	movups	(%ebp),%xmm0
 	movl	%ebx,%ecx
 	movups	(%esi),%xmm6
 	xorps	%xmm0,%xmm2
 	movups	16(%ebp),%xmm1
 	xorps	%xmm6,%xmm0
+	leal	32(%ebp),%edx
 	xorps	%xmm0,%xmm3
-	movups	32(%ebp),%xmm0
-.L031ccm64_enc2_loop:
+	movups	(%edx),%xmm0
+.L027ccm64_enc2_loop:
 .byte	102,15,56,220,209
+	decl	%ecx
 .byte	102,15,56,220,217
-	movups	(%edx,%ecx,1),%xmm1
-	addl	$32,%ecx
+	movups	16(%edx),%xmm1
 .byte	102,15,56,220,208
+	leal	32(%edx),%edx
 .byte	102,15,56,220,216
-	movups	-16(%edx,%ecx,1),%xmm0
-	jnz	.L031ccm64_enc2_loop
+	movups	(%edx),%xmm0
+	jnz	.L027ccm64_enc2_loop
 .byte	102,15,56,220,209
 .byte	102,15,56,220,217
 	paddq	16(%esp),%xmm7
-	decl	%eax
 .byte	102,15,56,221,208
 .byte	102,15,56,221,216
+	decl	%eax
 	leal	16(%esi),%esi
 	xorps	%xmm2,%xmm6
 	movdqa	%xmm7,%xmm2
 	movups	%xmm6,(%edi)
-.byte	102,15,56,0,213
 	leal	16(%edi),%edi
-	jnz	.L030ccm64_enc_outer
+.byte	102,15,56,0,213
+	jnz	.L026ccm64_enc_outer
 	movl	48(%esp),%esp
 	movl	40(%esp),%edi
 	movups	%xmm3,(%edi)
-	pxor	%xmm0,%xmm0
-	pxor	%xmm1,%xmm1
-	pxor	%xmm2,%xmm2
-	pxor	%xmm3,%xmm3
-	pxor	%xmm4,%xmm4
-	pxor	%xmm5,%xmm5
-	pxor	%xmm6,%xmm6
-	pxor	%xmm7,%xmm7
 	popl	%edi
 	popl	%esi
 	popl	%ebx
@@ -759,82 +685,71 @@ aesni_ccm64_decrypt_blocks:
 	movups	16(%edx),%xmm1
 	leal	32(%edx),%edx
 	xorps	%xmm0,%xmm2
-.L032enc1_loop_5:
+.L028enc1_loop_5:
 .byte	102,15,56,220,209
 	decl	%ecx
 	movups	(%edx),%xmm1
 	leal	16(%edx),%edx
-	jnz	.L032enc1_loop_5
+	jnz	.L028enc1_loop_5
 .byte	102,15,56,221,209
-	shll	$4,%ebx
-	movl	$16,%ecx
 	movups	(%esi),%xmm6
 	paddq	16(%esp),%xmm7
 	leal	16(%esi),%esi
-	subl	%ebx,%ecx
-	leal	32(%ebp,%ebx,1),%edx
-	movl	%ecx,%ebx
-	jmp	.L033ccm64_dec_outer
+	jmp	.L029ccm64_dec_outer
 .align	16
-.L033ccm64_dec_outer:
+.L029ccm64_dec_outer:
 	xorps	%xmm2,%xmm6
 	movdqa	%xmm7,%xmm2
+	movl	%ebx,%ecx
 	movups	%xmm6,(%edi)
 	leal	16(%edi),%edi
 .byte	102,15,56,0,213
 	subl	$1,%eax
-	jz	.L034ccm64_dec_break
+	jz	.L030ccm64_dec_break
 	movups	(%ebp),%xmm0
-	movl	%ebx,%ecx
+	shrl	$1,%ecx
 	movups	16(%ebp),%xmm1
 	xorps	%xmm0,%xmm6
+	leal	32(%ebp),%edx
 	xorps	%xmm0,%xmm2
 	xorps	%xmm6,%xmm3
-	movups	32(%ebp),%xmm0
-.L035ccm64_dec2_loop:
+	movups	(%edx),%xmm0
+.L031ccm64_dec2_loop:
 .byte	102,15,56,220,209
+	decl	%ecx
 .byte	102,15,56,220,217
-	movups	(%edx,%ecx,1),%xmm1
-	addl	$32,%ecx
+	movups	16(%edx),%xmm1
 .byte	102,15,56,220,208
+	leal	32(%edx),%edx
 .byte	102,15,56,220,216
-	movups	-16(%edx,%ecx,1),%xmm0
-	jnz	.L035ccm64_dec2_loop
+	movups	(%edx),%xmm0
+	jnz	.L031ccm64_dec2_loop
 	movups	(%esi),%xmm6
 	paddq	16(%esp),%xmm7
 .byte	102,15,56,220,209
 .byte	102,15,56,220,217
+	leal	16(%esi),%esi
 .byte	102,15,56,221,208
 .byte	102,15,56,221,216
-	leal	16(%esi),%esi
-	jmp	.L033ccm64_dec_outer
+	jmp	.L029ccm64_dec_outer
 .align	16
-.L034ccm64_dec_break:
-	movl	240(%ebp),%ecx
+.L030ccm64_dec_break:
 	movl	%ebp,%edx
 	movups	(%edx),%xmm0
 	movups	16(%edx),%xmm1
 	xorps	%xmm0,%xmm6
 	leal	32(%edx),%edx
 	xorps	%xmm6,%xmm3
-.L036enc1_loop_6:
+.L032enc1_loop_6:
 .byte	102,15,56,220,217
 	decl	%ecx
 	movups	(%edx),%xmm1
 	leal	16(%edx),%edx
-	jnz	.L036enc1_loop_6
+	jnz	.L032enc1_loop_6
 .byte	102,15,56,221,217
 	movl	48(%esp),%esp
 	movl	40(%esp),%edi
 	movups	%xmm3,(%edi)
-	pxor	%xmm0,%xmm0
-	pxor	%xmm1,%xmm1
-	pxor	%xmm2,%xmm2
-	pxor	%xmm3,%xmm3
-	pxor	%xmm4,%xmm4
-	pxor	%xmm5,%xmm5
-	pxor	%xmm6,%xmm6
-	pxor	%xmm7,%xmm7
 	popl	%edi
 	popl	%esi
 	popl	%ebx
@@ -860,7 +775,7 @@ aesni_ctr32_encrypt_blocks:
 	andl	$-16,%esp
 	movl	%ebp,80(%esp)
 	cmpl	$1,%eax
-	je	.L037ctr32_one_shortcut
+	je	.L033ctr32_one_shortcut
 	movdqu	(%ebx),%xmm7
 	movl	$202182159,(%esp)
 	movl	$134810123,4(%esp)
@@ -876,59 +791,63 @@ aesni_ctr32_encrypt_blocks:
 .byte	102,15,58,34,253,3
 	movl	240(%edx),%ecx
 	bswap	%ebx
-	pxor	%xmm0,%xmm0
 	pxor	%xmm1,%xmm1
+	pxor	%xmm0,%xmm0
 	movdqa	(%esp),%xmm2
-.byte	102,15,58,34,195,0
+.byte	102,15,58,34,203,0
 	leal	3(%ebx),%ebp
-.byte	102,15,58,34,205,0
+.byte	102,15,58,34,197,0
 	incl	%ebx
-.byte	102,15,58,34,195,1
+.byte	102,15,58,34,203,1
 	incl	%ebp
-.byte	102,15,58,34,205,1
+.byte	102,15,58,34,197,1
 	incl	%ebx
-.byte	102,15,58,34,195,2
+.byte	102,15,58,34,203,2
 	incl	%ebp
-.byte	102,15,58,34,205,2
-	movdqa	%xmm0,48(%esp)
-.byte	102,15,56,0,194
-	movdqu	(%edx),%xmm6
-	movdqa	%xmm1,64(%esp)
+.byte	102,15,58,34,197,2
+	movdqa	%xmm1,48(%esp)
 .byte	102,15,56,0,202
-	pshufd	$192,%xmm0,%xmm2
-	pshufd	$128,%xmm0,%xmm3
+	movdqa	%xmm0,64(%esp)
+.byte	102,15,56,0,194
+	pshufd	$192,%xmm1,%xmm2
+	pshufd	$128,%xmm1,%xmm3
 	cmpl	$6,%eax
-	jb	.L038ctr32_tail
-	pxor	%xmm6,%xmm7
-	shll	$4,%ecx
-	movl	$16,%ebx
+	jb	.L034ctr32_tail
 	movdqa	%xmm7,32(%esp)
+	shrl	$1,%ecx
 	movl	%edx,%ebp
-	subl	%ecx,%ebx
-	leal	32(%edx,%ecx,1),%edx
+	movl	%ecx,%ebx
 	subl	$6,%eax
-	jmp	.L039ctr32_loop6
-.align	16
-.L039ctr32_loop6:
-	pshufd	$64,%xmm0,%xmm4
-	movdqa	32(%esp),%xmm0
-	pshufd	$192,%xmm1,%xmm5
+	jmp	.L035ctr32_loop6
+.align	16
+.L035ctr32_loop6:
+	pshufd	$64,%xmm1,%xmm4
+	movdqa	32(%esp),%xmm1
+	pshufd	$192,%xmm0,%xmm5
+	por	%xmm1,%xmm2
+	pshufd	$128,%xmm0,%xmm6
+	por	%xmm1,%xmm3
+	pshufd	$64,%xmm0,%xmm7
+	por	%xmm1,%xmm4
+	por	%xmm1,%xmm5
+	por	%xmm1,%xmm6
+	por	%xmm1,%xmm7
+	movups	(%ebp),%xmm0
+	movups	16(%ebp),%xmm1
+	leal	32(%ebp),%edx
+	decl	%ecx
 	pxor	%xmm0,%xmm2
-	pshufd	$128,%xmm1,%xmm6
 	pxor	%xmm0,%xmm3
-	pshufd	$64,%xmm1,%xmm7
-	movups	16(%ebp),%xmm1
-	pxor	%xmm0,%xmm4
-	pxor	%xmm0,%xmm5
 .byte	102,15,56,220,209
-	pxor	%xmm0,%xmm6
-	pxor	%xmm0,%xmm7
+	pxor	%xmm0,%xmm4
 .byte	102,15,56,220,217
-	movups	32(%ebp),%xmm0
-	movl	%ebx,%ecx
+	pxor	%xmm0,%xmm5
 .byte	102,15,56,220,225
+	pxor	%xmm0,%xmm6
 .byte	102,15,56,220,233
+	pxor	%xmm0,%xmm7
 .byte	102,15,56,220,241
+	movups	(%edx),%xmm0
 .byte	102,15,56,220,249
 	call	.L_aesni_encrypt6_enter
 	movups	(%esi),%xmm1
@@ -939,51 +858,51 @@ aesni_ctr32_encrypt_blocks:
 	movups	%xmm2,(%edi)
 	movdqa	16(%esp),%xmm0
 	xorps	%xmm1,%xmm4
-	movdqa	64(%esp),%xmm1
+	movdqa	48(%esp),%xmm1
 	movups	%xmm3,16(%edi)
 	movups	%xmm4,32(%edi)
 	paddd	%xmm0,%xmm1
-	paddd	48(%esp),%xmm0
+	paddd	64(%esp),%xmm0
 	movdqa	(%esp),%xmm2
 	movups	48(%esi),%xmm3
 	movups	64(%esi),%xmm4
 	xorps	%xmm3,%xmm5
 	movups	80(%esi),%xmm3
 	leal	96(%esi),%esi
-	movdqa	%xmm0,48(%esp)
-.byte	102,15,56,0,194
+	movdqa	%xmm1,48(%esp)
+.byte	102,15,56,0,202
 	xorps	%xmm4,%xmm6
 	movups	%xmm5,48(%edi)
 	xorps	%xmm3,%xmm7
-	movdqa	%xmm1,64(%esp)
-.byte	102,15,56,0,202
+	movdqa	%xmm0,64(%esp)
+.byte	102,15,56,0,194
 	movups	%xmm6,64(%edi)
-	pshufd	$192,%xmm0,%xmm2
+	pshufd	$192,%xmm1,%xmm2
 	movups	%xmm7,80(%edi)
 	leal	96(%edi),%edi
-	pshufd	$128,%xmm0,%xmm3
+	movl	%ebx,%ecx
+	pshufd	$128,%xmm1,%xmm3
 	subl	$6,%eax
-	jnc	.L039ctr32_loop6
+	jnc	.L035ctr32_loop6
 	addl	$6,%eax
-	jz	.L040ctr32_ret
-	movdqu	(%ebp),%xmm7
+	jz	.L036ctr32_ret
 	movl	%ebp,%edx
-	pxor	32(%esp),%xmm7
-	movl	240(%ebp),%ecx
-.L038ctr32_tail:
+	leal	1(,%ecx,2),%ecx
+	movdqa	32(%esp),%xmm7
+.L034ctr32_tail:
 	por	%xmm7,%xmm2
 	cmpl	$2,%eax
-	jb	.L041ctr32_one
-	pshufd	$64,%xmm0,%xmm4
+	jb	.L037ctr32_one
+	pshufd	$64,%xmm1,%xmm4
 	por	%xmm7,%xmm3
-	je	.L042ctr32_two
-	pshufd	$192,%xmm1,%xmm5
+	je	.L038ctr32_two
+	pshufd	$192,%xmm0,%xmm5
 	por	%xmm7,%xmm4
 	cmpl	$4,%eax
-	jb	.L043ctr32_three
-	pshufd	$128,%xmm1,%xmm6
+	jb	.L039ctr32_three
+	pshufd	$128,%xmm0,%xmm6
 	por	%xmm7,%xmm5
-	je	.L044ctr32_four
+	je	.L040ctr32_four
 	por	%xmm7,%xmm6
 	call	_aesni_encrypt6
 	movups	(%esi),%xmm1
@@ -1001,39 +920,39 @@ aesni_ctr32_encrypt_blocks:
 	movups	%xmm4,32(%edi)
 	movups	%xmm5,48(%edi)
 	movups	%xmm6,64(%edi)
-	jmp	.L040ctr32_ret
+	jmp	.L036ctr32_ret
 .align	16
-.L037ctr32_one_shortcut:
+.L033ctr32_one_shortcut:
 	movups	(%ebx),%xmm2
 	movl	240(%edx),%ecx
-.L041ctr32_one:
+.L037ctr32_one:
 	movups	(%edx),%xmm0
 	movups	16(%edx),%xmm1
 	leal	32(%edx),%edx
 	xorps	%xmm0,%xmm2
-.L045enc1_loop_7:
+.L041enc1_loop_7:
 .byte	102,15,56,220,209
 	decl	%ecx
 	movups	(%edx),%xmm1
 	leal	16(%edx),%edx
-	jnz	.L045enc1_loop_7
+	jnz	.L041enc1_loop_7
 .byte	102,15,56,221,209
 	movups	(%esi),%xmm6
 	xorps	%xmm2,%xmm6
 	movups	%xmm6,(%edi)
-	jmp	.L040ctr32_ret
+	jmp	.L036ctr32_ret
 .align	16
-.L042ctr32_two:
-	call	_aesni_encrypt2
+.L038ctr32_two:
+	call	_aesni_encrypt3
 	movups	(%esi),%xmm5
 	movups	16(%esi),%xmm6
 	xorps	%xmm5,%xmm2
 	xorps	%xmm6,%xmm3
 	movups	%xmm2,(%edi)
 	movups	%xmm3,16(%edi)
-	jmp	.L040ctr32_ret
+	jmp	.L036ctr32_ret
 .align	16
-.L043ctr32_three:
+.L039ctr32_three:
 	call	_aesni_encrypt3
 	movups	(%esi),%xmm5
 	movups	16(%esi),%xmm6
@@ -1044,9 +963,9 @@ aesni_ctr32_encrypt_blocks:
 	xorps	%xmm7,%xmm4
 	movups	%xmm3,16(%edi)
 	movups	%xmm4,32(%edi)
-	jmp	.L040ctr32_ret
+	jmp	.L036ctr32_ret
 .align	16
-.L044ctr32_four:
+.L040ctr32_four:
 	call	_aesni_encrypt4
 	movups	(%esi),%xmm6
 	movups	16(%esi),%xmm7
@@ -1060,18 +979,7 @@ aesni_ctr32_encrypt_blocks:
 	xorps	%xmm0,%xmm5
 	movups	%xmm4,32(%edi)
 	movups	%xmm5,48(%edi)
-.L040ctr32_ret:
-	pxor	%xmm0,%xmm0
-	pxor	%xmm1,%xmm1
-	pxor	%xmm2,%xmm2
-	pxor	%xmm3,%xmm3
-	pxor	%xmm4,%xmm4
-	movdqa	%xmm0,32(%esp)
-	pxor	%xmm5,%xmm5
-	movdqa	%xmm0,48(%esp)
-	pxor	%xmm6,%xmm6
-	movdqa	%xmm0,64(%esp)
-	pxor	%xmm7,%xmm7
+.L036ctr32_ret:
 	movl	80(%esp),%esp
 	popl	%edi
 	popl	%esi
@@ -1096,12 +1004,12 @@ aesni_xts_encrypt:
 	movups	16(%edx),%xmm1
 	leal	32(%edx),%edx
 	xorps	%xmm0,%xmm2
-.L046enc1_loop_8:
+.L042enc1_loop_8:
 .byte	102,15,56,220,209
 	decl	%ecx
 	movups	(%edx),%xmm1
 	leal	16(%edx),%edx
-	jnz	.L046enc1_loop_8
+	jnz	.L042enc1_loop_8
 .byte	102,15,56,221,209
 	movl	20(%esp),%esi
 	movl	24(%esp),%edi
@@ -1125,14 +1033,12 @@ aesni_xts_encrypt:
 	movl	%edx,%ebp
 	movl	%ecx,%ebx
 	subl	$96,%eax
-	jc	.L047xts_enc_short
-	shll	$4,%ecx
-	movl	$16,%ebx
-	subl	%ecx,%ebx
-	leal	32(%edx,%ecx,1),%edx
-	jmp	.L048xts_enc_loop6
+	jc	.L043xts_enc_short
+	shrl	$1,%ecx
+	movl	%ecx,%ebx
+	jmp	.L044xts_enc_loop6
 .align	16
-.L048xts_enc_loop6:
+.L044xts_enc_loop6:
 	pshufd	$19,%xmm0,%xmm2
 	pxor	%xmm0,%xmm0
 	movdqa	%xmm1,(%esp)
@@ -1168,7 +1074,6 @@ aesni_xts_encrypt:
 	pand	%xmm3,%xmm7
 	movups	(%esi),%xmm2
 	pxor	%xmm1,%xmm7
-	movl	%ebx,%ecx
 	movdqu	16(%esi),%xmm3
 	xorps	%xmm0,%xmm2
 	movdqu	32(%esi),%xmm4
@@ -1184,17 +1089,19 @@ aesni_xts_encrypt:
 	movdqa	%xmm7,80(%esp)
 	pxor	%xmm1,%xmm7
 	movups	16(%ebp),%xmm1
+	leal	32(%ebp),%edx
 	pxor	16(%esp),%xmm3
-	pxor	32(%esp),%xmm4
 .byte	102,15,56,220,209
-	pxor	48(%esp),%xmm5
-	pxor	64(%esp),%xmm6
+	pxor	32(%esp),%xmm4
 .byte	102,15,56,220,217
-	pxor	%xmm0,%xmm7
-	movups	32(%ebp),%xmm0
+	pxor	48(%esp),%xmm5
+	decl	%ecx
 .byte	102,15,56,220,225
+	pxor	64(%esp),%xmm6
 .byte	102,15,56,220,233
+	pxor	%xmm0,%xmm7
 .byte	102,15,56,220,241
+	movups	(%edx),%xmm0
 .byte	102,15,56,220,249
 	call	.L_aesni_encrypt6_enter
 	movdqa	80(%esp),%xmm1
@@ -1219,25 +1126,26 @@ aesni_xts_encrypt:
 	paddq	%xmm1,%xmm1
 	pand	%xmm3,%xmm2
 	pcmpgtd	%xmm1,%xmm0
+	movl	%ebx,%ecx
 	pxor	%xmm2,%xmm1
 	subl	$96,%eax
-	jnc	.L048xts_enc_loop6
-	movl	240(%ebp),%ecx
+	jnc	.L044xts_enc_loop6
+	leal	1(,%ecx,2),%ecx
 	movl	%ebp,%edx
 	movl	%ecx,%ebx
-.L047xts_enc_short:
+.L043xts_enc_short:
 	addl	$96,%eax
-	jz	.L049xts_enc_done6x
+	jz	.L045xts_enc_done6x
 	movdqa	%xmm1,%xmm5
 	cmpl	$32,%eax
-	jb	.L050xts_enc_one
+	jb	.L046xts_enc_one
 	pshufd	$19,%xmm0,%xmm2
 	pxor	%xmm0,%xmm0
 	paddq	%xmm1,%xmm1
 	pand	%xmm3,%xmm2
 	pcmpgtd	%xmm1,%xmm0
 	pxor	%xmm2,%xmm1
-	je	.L051xts_enc_two
+	je	.L047xts_enc_two
 	pshufd	$19,%xmm0,%xmm2
 	pxor	%xmm0,%xmm0
 	movdqa	%xmm1,%xmm6
@@ -1246,7 +1154,7 @@ aesni_xts_encrypt:
 	pcmpgtd	%xmm1,%xmm0
 	pxor	%xmm2,%xmm1
 	cmpl	$64,%eax
-	jb	.L052xts_enc_three
+	jb	.L048xts_enc_three
 	pshufd	$19,%xmm0,%xmm2
 	pxor	%xmm0,%xmm0
 	movdqa	%xmm1,%xmm7
@@ -1256,7 +1164,7 @@ aesni_xts_encrypt:
 	pxor	%xmm2,%xmm1
 	movdqa	%xmm5,(%esp)
 	movdqa	%xmm6,16(%esp)
-	je	.L053xts_enc_four
+	je	.L049xts_enc_four
 	movdqa	%xmm7,32(%esp)
 	pshufd	$19,%xmm0,%xmm7
 	movdqa	%xmm1,48(%esp)
@@ -1288,9 +1196,9 @@ aesni_xts_encrypt:
 	movups	%xmm5,48(%edi)
 	movups	%xmm6,64(%edi)
 	leal	80(%edi),%edi
-	jmp	.L054xts_enc_done
+	jmp	.L050xts_enc_done
 .align	16
-.L050xts_enc_one:
+.L046xts_enc_one:
 	movups	(%esi),%xmm2
 	leal	16(%esi),%esi
 	xorps	%xmm5,%xmm2
@@ -1298,36 +1206,37 @@ aesni_xts_encrypt:
 	movups	16(%edx),%xmm1
 	leal	32(%edx),%edx
 	xorps	%xmm0,%xmm2
-.L055enc1_loop_9:
+.L051enc1_loop_9:
 .byte	102,15,56,220,209
 	decl	%ecx
 	movups	(%edx),%xmm1
 	leal	16(%edx),%edx
-	jnz	.L055enc1_loop_9
+	jnz	.L051enc1_loop_9
 .byte	102,15,56,221,209
 	xorps	%xmm5,%xmm2
 	movups	%xmm2,(%edi)
 	leal	16(%edi),%edi
 	movdqa	%xmm5,%xmm1
-	jmp	.L054xts_enc_done
+	jmp	.L050xts_enc_done
 .align	16
-.L051xts_enc_two:
+.L047xts_enc_two:
 	movaps	%xmm1,%xmm6
 	movups	(%esi),%xmm2
 	movups	16(%esi),%xmm3
 	leal	32(%esi),%esi
 	xorps	%xmm5,%xmm2
 	xorps	%xmm6,%xmm3
-	call	_aesni_encrypt2
+	xorps	%xmm4,%xmm4
+	call	_aesni_encrypt3
 	xorps	%xmm5,%xmm2
 	xorps	%xmm6,%xmm3
 	movups	%xmm2,(%edi)
 	movups	%xmm3,16(%edi)
 	leal	32(%edi),%edi
 	movdqa	%xmm6,%xmm1
-	jmp	.L054xts_enc_done
+	jmp	.L050xts_enc_done
 .align	16
-.L052xts_enc_three:
+.L048xts_enc_three:
 	movaps	%xmm1,%xmm7
 	movups	(%esi),%xmm2
 	movups	16(%esi),%xmm3
@@ -1345,9 +1254,9 @@ aesni_xts_encrypt:
 	movups	%xmm4,32(%edi)
 	leal	48(%edi),%edi
 	movdqa	%xmm7,%xmm1
-	jmp	.L054xts_enc_done
+	jmp	.L050xts_enc_done
 .align	16
-.L053xts_enc_four:
+.L049xts_enc_four:
 	movaps	%xmm1,%xmm6
 	movups	(%esi),%xmm2
 	movups	16(%esi),%xmm3
@@ -1369,28 +1278,28 @@ aesni_xts_encrypt:
 	movups	%xmm5,48(%edi)
 	leal	64(%edi),%edi
 	movdqa	%xmm6,%xmm1
-	jmp	.L054xts_enc_done
+	jmp	.L050xts_enc_done
 .align	16
-.L049xts_enc_done6x:
+.L045xts_enc_done6x:
 	movl	112(%esp),%eax
 	andl	$15,%eax
-	jz	.L056xts_enc_ret
+	jz	.L052xts_enc_ret
 	movdqa	%xmm1,%xmm5
 	movl	%eax,112(%esp)
-	jmp	.L057xts_enc_steal
+	jmp	.L053xts_enc_steal
 .align	16
-.L054xts_enc_done:
+.L050xts_enc_done:
 	movl	112(%esp),%eax
 	pxor	%xmm0,%xmm0
 	andl	$15,%eax
-	jz	.L056xts_enc_ret
+	jz	.L052xts_enc_ret
 	pcmpgtd	%xmm1,%xmm0
 	movl	%eax,112(%esp)
 	pshufd	$19,%xmm0,%xmm5
 	paddq	%xmm1,%xmm1
 	pand	96(%esp),%xmm5
 	pxor	%xmm1,%xmm5
-.L057xts_enc_steal:
+.L053xts_enc_steal:
 	movzbl	(%esi),%ecx
 	movzbl	-16(%edi),%edx
 	leal	1(%esi),%esi
@@ -1398,7 +1307,7 @@ aesni_xts_encrypt:
 	movb	%dl,(%edi)
 	leal	1(%edi),%edi
 	subl	$1,%eax
-	jnz	.L057xts_enc_steal
+	jnz	.L053xts_enc_steal
 	subl	112(%esp),%edi
 	movl	%ebp,%edx
 	movl	%ebx,%ecx
@@ -1408,30 +1317,16 @@ aesni_xts_encrypt:
 	movups	16(%edx),%xmm1
 	leal	32(%edx),%edx
 	xorps	%xmm0,%xmm2
-.L058enc1_loop_10:
+.L054enc1_loop_10:
 .byte	102,15,56,220,209
 	decl	%ecx
 	movups	(%edx),%xmm1
 	leal	16(%edx),%edx
-	jnz	.L058enc1_loop_10
+	jnz	.L054enc1_loop_10
 .byte	102,15,56,221,209
 	xorps	%xmm5,%xmm2
 	movups	%xmm2,-16(%edi)
-.L056xts_enc_ret:
-	pxor	%xmm0,%xmm0
-	pxor	%xmm1,%xmm1
-	pxor	%xmm2,%xmm2
-	movdqa	%xmm0,(%esp)
-	pxor	%xmm3,%xmm3
-	movdqa	%xmm0,16(%esp)
-	pxor	%xmm4,%xmm4
-	movdqa	%xmm0,32(%esp)
-	pxor	%xmm5,%xmm5
-	movdqa	%xmm0,48(%esp)
-	pxor	%xmm6,%xmm6
-	movdqa	%xmm0,64(%esp)
-	pxor	%xmm7,%xmm7
-	movdqa	%xmm0,80(%esp)
+.L052xts_enc_ret:
 	movl	116(%esp),%esp
 	popl	%edi
 	popl	%esi
@@ -1456,12 +1351,12 @@ aesni_xts_decrypt:
 	movups	16(%edx),%xmm1
 	leal	32(%edx),%edx
 	xorps	%xmm0,%xmm2
-.L059enc1_loop_11:
+.L055enc1_loop_11:
 .byte	102,15,56,220,209
 	decl	%ecx
 	movups	(%edx),%xmm1
 	leal	16(%edx),%edx
-	jnz	.L059enc1_loop_11
+	jnz	.L055enc1_loop_11
 .byte	102,15,56,221,209
 	movl	20(%esp),%esi
 	movl	24(%esp),%edi
@@ -1490,14 +1385,12 @@ aesni_xts_decrypt:
 	pcmpgtd	%xmm1,%xmm0
 	andl	$-16,%eax
 	subl	$96,%eax
-	jc	.L060xts_dec_short
-	shll	$4,%ecx
-	movl	$16,%ebx
-	subl	%ecx,%ebx
-	leal	32(%edx,%ecx,1),%edx
-	jmp	.L061xts_dec_loop6
+	jc	.L056xts_dec_short
+	shrl	$1,%ecx
+	movl	%ecx,%ebx
+	jmp	.L057xts_dec_loop6
 .align	16
-.L061xts_dec_loop6:
+.L057xts_dec_loop6:
 	pshufd	$19,%xmm0,%xmm2
 	pxor	%xmm0,%xmm0
 	movdqa	%xmm1,(%esp)
@@ -1533,7 +1426,6 @@ aesni_xts_decrypt:
 	pand	%xmm3,%xmm7
 	movups	(%esi),%xmm2
 	pxor	%xmm1,%xmm7
-	movl	%ebx,%ecx
 	movdqu	16(%esi),%xmm3
 	xorps	%xmm0,%xmm2
 	movdqu	32(%esi),%xmm4
@@ -1549,17 +1441,19 @@ aesni_xts_decrypt:
 	movdqa	%xmm7,80(%esp)
 	pxor	%xmm1,%xmm7
 	movups	16(%ebp),%xmm1
+	leal	32(%ebp),%edx
 	pxor	16(%esp),%xmm3
-	pxor	32(%esp),%xmm4
 .byte	102,15,56,222,209
-	pxor	48(%esp),%xmm5
-	pxor	64(%esp),%xmm6
+	pxor	32(%esp),%xmm4
 .byte	102,15,56,222,217
-	pxor	%xmm0,%xmm7
-	movups	32(%ebp),%xmm0
+	pxor	48(%esp),%xmm5
+	decl	%ecx
 .byte	102,15,56,222,225
+	pxor	64(%esp),%xmm6
 .byte	102,15,56,222,233
+	pxor	%xmm0,%xmm7
 .byte	102,15,56,222,241
+	movups	(%edx),%xmm0
 .byte	102,15,56,222,249
 	call	.L_aesni_decrypt6_enter
 	movdqa	80(%esp),%xmm1
@@ -1584,25 +1478,26 @@ aesni_xts_decrypt:
 	paddq	%xmm1,%xmm1
 	pand	%xmm3,%xmm2
 	pcmpgtd	%xmm1,%xmm0
+	movl	%ebx,%ecx
 	pxor	%xmm2,%xmm1
 	subl	$96,%eax
-	jnc	.L061xts_dec_loop6
-	movl	240(%ebp),%ecx
+	jnc	.L057xts_dec_loop6
+	leal	1(,%ecx,2),%ecx
 	movl	%ebp,%edx
 	movl	%ecx,%ebx
-.L060xts_dec_short:
+.L056xts_dec_short:
 	addl	$96,%eax
-	jz	.L062xts_dec_done6x
+	jz	.L058xts_dec_done6x
 	movdqa	%xmm1,%xmm5
 	cmpl	$32,%eax
-	jb	.L063xts_dec_one
+	jb	.L059xts_dec_one
 	pshufd	$19,%xmm0,%xmm2
 	pxor	%xmm0,%xmm0
 	paddq	%xmm1,%xmm1
 	pand	%xmm3,%xmm2
 	pcmpgtd	%xmm1,%xmm0
 	pxor	%xmm2,%xmm1
-	je	.L064xts_dec_two
+	je	.L060xts_dec_two
 	pshufd	$19,%xmm0,%xmm2
 	pxor	%xmm0,%xmm0
 	movdqa	%xmm1,%xmm6
@@ -1611,7 +1506,7 @@ aesni_xts_decrypt:
 	pcmpgtd	%xmm1,%xmm0
 	pxor	%xmm2,%xmm1
 	cmpl	$64,%eax
-	jb	.L065xts_dec_three
+	jb	.L061xts_dec_three
 	pshufd	$19,%xmm0,%xmm2
 	pxor	%xmm0,%xmm0
 	movdqa	%xmm1,%xmm7
@@ -1621,7 +1516,7 @@ aesni_xts_decrypt:
 	pxor	%xmm2,%xmm1
 	movdqa	%xmm5,(%esp)
 	movdqa	%xmm6,16(%esp)
-	je	.L066xts_dec_four
+	je	.L062xts_dec_four
 	movdqa	%xmm7,32(%esp)
 	pshufd	$19,%xmm0,%xmm7
 	movdqa	%xmm1,48(%esp)
@@ -1653,9 +1548,9 @@ aesni_xts_decrypt:
 	movups	%xmm5,48(%edi)
 	movups	%xmm6,64(%edi)
 	leal	80(%edi),%edi
-	jmp	.L067xts_dec_done
+	jmp	.L063xts_dec_done
 .align	16
-.L063xts_dec_one:
+.L059xts_dec_one:
 	movups	(%esi),%xmm2
 	leal	16(%esi),%esi
 	xorps	%xmm5,%xmm2
@@ -1663,36 +1558,36 @@ aesni_xts_decrypt:
 	movups	16(%edx),%xmm1
 	leal	32(%edx),%edx
 	xorps	%xmm0,%xmm2
-.L068dec1_loop_12:
+.L064dec1_loop_12:
 .byte	102,15,56,222,209
 	decl	%ecx
 	movups	(%edx),%xmm1
 	leal	16(%edx),%edx
-	jnz	.L068dec1_loop_12
+	jnz	.L064dec1_loop_12
 .byte	102,15,56,223,209
 	xorps	%xmm5,%xmm2
 	movups	%xmm2,(%edi)
 	leal	16(%edi),%edi
 	movdqa	%xmm5,%xmm1
-	jmp	.L067xts_dec_done
+	jmp	.L063xts_dec_done
 .align	16
-.L064xts_dec_two:
+.L060xts_dec_two:
 	movaps	%xmm1,%xmm6
 	movups	(%esi),%xmm2
 	movups	16(%esi),%xmm3
 	leal	32(%esi),%esi
 	xorps	%xmm5,%xmm2
 	xorps	%xmm6,%xmm3
-	call	_aesni_decrypt2
+	call	_aesni_decrypt3
 	xorps	%xmm5,%xmm2
 	xorps	%xmm6,%xmm3
 	movups	%xmm2,(%edi)
 	movups	%xmm3,16(%edi)
 	leal	32(%edi),%edi
 	movdqa	%xmm6,%xmm1
-	jmp	.L067xts_dec_done
+	jmp	.L063xts_dec_done
 .align	16
-.L065xts_dec_three:
+.L061xts_dec_three:
 	movaps	%xmm1,%xmm7
 	movups	(%esi),%xmm2
 	movups	16(%esi),%xmm3
@@ -1710,9 +1605,9 @@ aesni_xts_decrypt:
 	movups	%xmm4,32(%edi)
 	leal	48(%edi),%edi
 	movdqa	%xmm7,%xmm1
-	jmp	.L067xts_dec_done
+	jmp	.L063xts_dec_done
 .align	16
-.L066xts_dec_four:
+.L062xts_dec_four:
 	movaps	%xmm1,%xmm6
 	movups	(%esi),%xmm2
 	movups	16(%esi),%xmm3
@@ -1734,20 +1629,20 @@ aesni_xts_decrypt:
 	movups	%xmm5,48(%edi)
 	leal	64(%edi),%edi
 	movdqa	%xmm6,%xmm1
-	jmp	.L067xts_dec_done
+	jmp	.L063xts_dec_done
 .align	16
-.L062xts_dec_done6x:
+.L058xts_dec_done6x:
 	movl	112(%esp),%eax
 	andl	$15,%eax
-	jz	.L069xts_dec_ret
+	jz	.L065xts_dec_ret
 	movl	%eax,112(%esp)
-	jmp	.L070xts_dec_only_one_more
+	jmp	.L066xts_dec_only_one_more
 .align	16
-.L067xts_dec_done:
+.L063xts_dec_done:
 	movl	112(%esp),%eax
 	pxor	%xmm0,%xmm0
 	andl	$15,%eax
-	jz	.L069xts_dec_ret
+	jz	.L065xts_dec_ret
 	pcmpgtd	%xmm1,%xmm0
 	movl	%eax,112(%esp)
 	pshufd	$19,%xmm0,%xmm2
@@ -1757,7 +1652,7 @@ aesni_xts_decrypt:
 	pand	%xmm3,%xmm2
 	pcmpgtd	%xmm1,%xmm0
 	pxor	%xmm2,%xmm1
-.L070xts_dec_only_one_more:
+.L066xts_dec_only_one_more:
 	pshufd	$19,%xmm0,%xmm5
 	movdqa	%xmm1,%xmm6
 	paddq	%xmm1,%xmm1
@@ -1771,16 +1666,16 @@ aesni_xts_decrypt:
 	movups	16(%edx),%xmm1
 	leal	32(%edx),%edx
 	xorps	%xmm0,%xmm2
-.L071dec1_loop_13:
+.L067dec1_loop_13:
 .byte	102,15,56,222,209
 	decl	%ecx
 	movups	(%edx),%xmm1
 	leal	16(%edx),%edx
-	jnz	.L071dec1_loop_13
+	jnz	.L067dec1_loop_13
 .byte	102,15,56,223,209
 	xorps	%xmm5,%xmm2
 	movups	%xmm2,(%edi)
-.L072xts_dec_steal:
+.L068xts_dec_steal:
 	movzbl	16(%esi),%ecx
 	movzbl	(%edi),%edx
 	leal	1(%esi),%esi
@@ -1788,7 +1683,7 @@ aesni_xts_decrypt:
 	movb	%dl,16(%edi)
 	leal	1(%edi),%edi
 	subl	$1,%eax
-	jnz	.L072xts_dec_steal
+	jnz	.L068xts_dec_steal
 	subl	112(%esp),%edi
 	movl	%ebp,%edx
 	movl	%ebx,%ecx
@@ -1798,30 +1693,16 @@ aesni_xts_decrypt:
 	movups	16(%edx),%xmm1
 	leal	32(%edx),%edx
 	xorps	%xmm0,%xmm2
-.L073dec1_loop_14:
+.L069dec1_loop_14:
 .byte	102,15,56,222,209
 	decl	%ecx
 	movups	(%edx),%xmm1
 	leal	16(%edx),%edx
-	jnz	.L073dec1_loop_14
+	jnz	.L069dec1_loop_14
 .byte	102,15,56,223,209
 	xorps	%xmm6,%xmm2
 	movups	%xmm2,(%edi)
-.L069xts_dec_ret:
-	pxor	%xmm0,%xmm0
-	pxor	%xmm1,%xmm1
-	pxor	%xmm2,%xmm2
-	movdqa	%xmm0,(%esp)
-	pxor	%xmm3,%xmm3
-	movdqa	%xmm0,16(%esp)
-	pxor	%xmm4,%xmm4
-	movdqa	%xmm0,32(%esp)
-	pxor	%xmm5,%xmm5
-	movdqa	%xmm0,48(%esp)
-	pxor	%xmm6,%xmm6
-	movdqa	%xmm0,64(%esp)
-	pxor	%xmm7,%xmm7
-	movdqa	%xmm0,80(%esp)
+.L065xts_dec_ret:
 	movl	116(%esp),%esp
 	popl	%edi
 	popl	%esi
@@ -1847,7 +1728,7 @@ aesni_cbc_encrypt:
 	movl	32(%esp),%edx
 	movl	36(%esp),%ebp
 	testl	%eax,%eax
-	jz	.L074cbc_abort
+	jz	.L070cbc_abort
 	cmpl	$0,40(%esp)
 	xchgl	%esp,%ebx
 	movups	(%ebp),%xmm7
@@ -1855,14 +1736,14 @@ aesni_cbc_encrypt:
 	movl	%edx,%ebp
 	movl	%ebx,16(%esp)
 	movl	%ecx,%ebx
-	je	.L075cbc_decrypt
+	je	.L071cbc_decrypt
 	movaps	%xmm7,%xmm2
 	cmpl	$16,%eax
-	jb	.L076cbc_enc_tail
+	jb	.L072cbc_enc_tail
 	subl	$16,%eax
-	jmp	.L077cbc_enc_loop
+	jmp	.L073cbc_enc_loop
 .align	16
-.L077cbc_enc_loop:
+.L073cbc_enc_loop:
 	movups	(%esi),%xmm7
 	leal	16(%esi),%esi
 	movups	(%edx),%xmm0
@@ -1870,25 +1751,24 @@ aesni_cbc_encrypt:
 	xorps	%xmm0,%xmm7
 	leal	32(%edx),%edx
 	xorps	%xmm7,%xmm2
-.L078enc1_loop_15:
+.L074enc1_loop_15:
 .byte	102,15,56,220,209
 	decl	%ecx
 	movups	(%edx),%xmm1
 	leal	16(%edx),%edx
-	jnz	.L078enc1_loop_15
+	jnz	.L074enc1_loop_15
 .byte	102,15,56,221,209
 	movl	%ebx,%ecx
 	movl	%ebp,%edx
 	movups	%xmm2,(%edi)
 	leal	16(%edi),%edi
 	subl	$16,%eax
-	jnc	.L077cbc_enc_loop
+	jnc	.L073cbc_enc_loop
 	addl	$16,%eax
-	jnz	.L076cbc_enc_tail
+	jnz	.L072cbc_enc_tail
 	movaps	%xmm2,%xmm7
-	pxor	%xmm2,%xmm2
-	jmp	.L079cbc_ret
-.L076cbc_enc_tail:
+	jmp	.L075cbc_ret
+.L072cbc_enc_tail:
 	movl	%eax,%ecx
 .long	2767451785
 	movl	$16,%ecx
@@ -1899,20 +1779,20 @@ aesni_cbc_encrypt:
 	movl	%ebx,%ecx
 	movl	%edi,%esi
 	movl	%ebp,%edx
-	jmp	.L077cbc_enc_loop
+	jmp	.L073cbc_enc_loop
 .align	16
-.L075cbc_decrypt:
+.L071cbc_decrypt:
 	cmpl	$80,%eax
-	jbe	.L080cbc_dec_tail
+	jbe	.L076cbc_dec_tail
 	movaps	%xmm7,(%esp)
 	subl	$80,%eax
-	jmp	.L081cbc_dec_loop6_enter
+	jmp	.L077cbc_dec_loop6_enter
 .align	16
-.L082cbc_dec_loop6:
+.L078cbc_dec_loop6:
 	movaps	%xmm0,(%esp)
 	movups	%xmm7,(%edi)
 	leal	16(%edi),%edi
-.L081cbc_dec_loop6_enter:
+.L077cbc_dec_loop6_enter:
 	movdqu	(%esi),%xmm2
 	movdqu	16(%esi),%xmm3
 	movdqu	32(%esi),%xmm4
@@ -1942,28 +1822,28 @@ aesni_cbc_encrypt:
 	movups	%xmm6,64(%edi)
 	leal	80(%edi),%edi
 	subl	$96,%eax
-	ja	.L082cbc_dec_loop6
+	ja	.L078cbc_dec_loop6
 	movaps	%xmm7,%xmm2
 	movaps	%xmm0,%xmm7
 	addl	$80,%eax
-	jle	.L083cbc_dec_clear_tail_collected
+	jle	.L079cbc_dec_tail_collected
 	movups	%xmm2,(%edi)
 	leal	16(%edi),%edi
-.L080cbc_dec_tail:
+.L076cbc_dec_tail:
 	movups	(%esi),%xmm2
 	movaps	%xmm2,%xmm6
 	cmpl	$16,%eax
-	jbe	.L084cbc_dec_one
+	jbe	.L080cbc_dec_one
 	movups	16(%esi),%xmm3
 	movaps	%xmm3,%xmm5
 	cmpl	$32,%eax
-	jbe	.L085cbc_dec_two
+	jbe	.L081cbc_dec_two
 	movups	32(%esi),%xmm4
 	cmpl	$48,%eax
-	jbe	.L086cbc_dec_three
+	jbe	.L082cbc_dec_three
 	movups	48(%esi),%xmm5
 	cmpl	$64,%eax
-	jbe	.L087cbc_dec_four
+	jbe	.L083cbc_dec_four
 	movups	64(%esi),%xmm6
 	movaps	%xmm7,(%esp)
 	movups	(%esi),%xmm2
@@ -1981,62 +1861,56 @@ aesni_cbc_encrypt:
 	xorps	%xmm0,%xmm6
 	movups	%xmm2,(%edi)
 	movups	%xmm3,16(%edi)
-	pxor	%xmm3,%xmm3
 	movups	%xmm4,32(%edi)
-	pxor	%xmm4,%xmm4
 	movups	%xmm5,48(%edi)
-	pxor	%xmm5,%xmm5
 	leal	64(%edi),%edi
 	movaps	%xmm6,%xmm2
-	pxor	%xmm6,%xmm6
 	subl	$80,%eax
-	jmp	.L088cbc_dec_tail_collected
+	jmp	.L079cbc_dec_tail_collected
 .align	16
-.L084cbc_dec_one:
+.L080cbc_dec_one:
 	movups	(%edx),%xmm0
 	movups	16(%edx),%xmm1
 	leal	32(%edx),%edx
 	xorps	%xmm0,%xmm2
-.L089dec1_loop_16:
+.L084dec1_loop_16:
 .byte	102,15,56,222,209
 	decl	%ecx
 	movups	(%edx),%xmm1
 	leal	16(%edx),%edx
-	jnz	.L089dec1_loop_16
+	jnz	.L084dec1_loop_16
 .byte	102,15,56,223,209
 	xorps	%xmm7,%xmm2
 	movaps	%xmm6,%xmm7
 	subl	$16,%eax
-	jmp	.L088cbc_dec_tail_collected
+	jmp	.L079cbc_dec_tail_collected
 .align	16
-.L085cbc_dec_two:
-	call	_aesni_decrypt2
+.L081cbc_dec_two:
+	xorps	%xmm4,%xmm4
+	call	_aesni_decrypt3
 	xorps	%xmm7,%xmm2
 	xorps	%xmm6,%xmm3
 	movups	%xmm2,(%edi)
 	movaps	%xmm3,%xmm2
-	pxor	%xmm3,%xmm3
 	leal	16(%edi),%edi
 	movaps	%xmm5,%xmm7
 	subl	$32,%eax
-	jmp	.L088cbc_dec_tail_collected
+	jmp	.L079cbc_dec_tail_collected
 .align	16
-.L086cbc_dec_three:
+.L082cbc_dec_three:
 	call	_aesni_decrypt3
 	xorps	%xmm7,%xmm2
 	xorps	%xmm6,%xmm3
 	xorps	%xmm5,%xmm4
 	movups	%xmm2,(%edi)
 	movaps	%xmm4,%xmm2
-	pxor	%xmm4,%xmm4
 	movups	%xmm3,16(%edi)
-	pxor	%xmm3,%xmm3
 	leal	32(%edi),%edi
 	movups	32(%esi),%xmm7
 	subl	$48,%eax
-	jmp	.L088cbc_dec_tail_collected
+	jmp	.L079cbc_dec_tail_collected
 .align	16
-.L087cbc_dec_four:
+.L083cbc_dec_four:
 	call	_aesni_decrypt4
 	movups	16(%esi),%xmm1
 	movups	32(%esi),%xmm0
@@ -2046,44 +1920,28 @@ aesni_cbc_encrypt:
 	movups	%xmm2,(%edi)
 	xorps	%xmm1,%xmm4
 	movups	%xmm3,16(%edi)
-	pxor	%xmm3,%xmm3
 	xorps	%xmm0,%xmm5
 	movups	%xmm4,32(%edi)
-	pxor	%xmm4,%xmm4
 	leal	48(%edi),%edi
 	movaps	%xmm5,%xmm2
-	pxor	%xmm5,%xmm5
 	subl	$64,%eax
-	jmp	.L088cbc_dec_tail_collected
-.align	16
-.L083cbc_dec_clear_tail_collected:
-	pxor	%xmm3,%xmm3
-	pxor	%xmm4,%xmm4
-	pxor	%xmm5,%xmm5
-	pxor	%xmm6,%xmm6
-.L088cbc_dec_tail_collected:
+.L079cbc_dec_tail_collected:
 	andl	$15,%eax
-	jnz	.L090cbc_dec_tail_partial
+	jnz	.L085cbc_dec_tail_partial
 	movups	%xmm2,(%edi)
-	pxor	%xmm0,%xmm0
-	jmp	.L079cbc_ret
+	jmp	.L075cbc_ret
 .align	16
-.L090cbc_dec_tail_partial:
+.L085cbc_dec_tail_partial:
 	movaps	%xmm2,(%esp)
-	pxor	%xmm0,%xmm0
 	movl	$16,%ecx
 	movl	%esp,%esi
 	subl	%eax,%ecx
 .long	2767451785
-	movdqa	%xmm2,(%esp)
-.L079cbc_ret:
+.L075cbc_ret:
 	movl	16(%esp),%esp
 	movl	36(%esp),%ebp
-	pxor	%xmm2,%xmm2
-	pxor	%xmm1,%xmm1
 	movups	%xmm7,(%ebp)
-	pxor	%xmm7,%xmm7
-.L074cbc_abort:
+.L070cbc_abort:
 	popl	%edi
 	popl	%esi
 	popl	%ebx
@@ -2093,62 +1951,52 @@ aesni_cbc_encrypt:
 .type	_aesni_set_encrypt_key,@function
 .align	16
 _aesni_set_encrypt_key:
-	pushl	%ebp
-	pushl	%ebx
 	testl	%eax,%eax
-	jz	.L091bad_pointer
+	jz	.L086bad_pointer
 	testl	%edx,%edx
-	jz	.L091bad_pointer
-	call	.L092pic
-.L092pic:
-	popl	%ebx
-	leal	.Lkey_const-.L092pic(%ebx),%ebx
-	leal	_gnutls_x86_cpuid_s,%ebp
+	jz	.L086bad_pointer
 	movups	(%eax),%xmm0
 	xorps	%xmm4,%xmm4
-	movl	4(%ebp),%ebp
 	leal	16(%edx),%edx
-	andl	$268437504,%ebp
 	cmpl	$256,%ecx
-	je	.L09314rounds
+	je	.L08714rounds
 	cmpl	$192,%ecx
-	je	.L09412rounds
+	je	.L08812rounds
 	cmpl	$128,%ecx
-	jne	.L095bad_keybits
+	jne	.L089bad_keybits
 .align	16
-.L09610rounds:
-	cmpl	$268435456,%ebp
-	je	.L09710rounds_alt
+.L09010rounds:
 	movl	$9,%ecx
 	movups	%xmm0,-16(%edx)
 .byte	102,15,58,223,200,1
-	call	.L098key_128_cold
+	call	.L091key_128_cold
 .byte	102,15,58,223,200,2
-	call	.L099key_128
+	call	.L092key_128
 .byte	102,15,58,223,200,4
-	call	.L099key_128
+	call	.L092key_128
 .byte	102,15,58,223,200,8
-	call	.L099key_128
+	call	.L092key_128
 .byte	102,15,58,223,200,16
-	call	.L099key_128
+	call	.L092key_128
 .byte	102,15,58,223,200,32
-	call	.L099key_128
+	call	.L092key_128
 .byte	102,15,58,223,200,64
-	call	.L099key_128
+	call	.L092key_128
 .byte	102,15,58,223,200,128
-	call	.L099key_128
+	call	.L092key_128
 .byte	102,15,58,223,200,27
-	call	.L099key_128
+	call	.L092key_128
 .byte	102,15,58,223,200,54
-	call	.L099key_128
+	call	.L092key_128
 	movups	%xmm0,(%edx)
 	movl	%ecx,80(%edx)
-	jmp	.L100good_key
+	xorl	%eax,%eax
+	ret
 .align	16
-.L099key_128:
+.L092key_128:
 	movups	%xmm0,(%edx)
 	leal	16(%edx),%edx
-.L098key_128_cold:
+.L091key_128_cold:
 	shufps	$16,%xmm0,%xmm4
 	xorps	%xmm4,%xmm0
 	shufps	$140,%xmm0,%xmm4
@@ -2157,91 +2005,38 @@ _aesni_set_encrypt_key:
 	xorps	%xmm1,%xmm0
 	ret
 .align	16
-.L09710rounds_alt:
-	movdqa	(%ebx),%xmm5
-	movl	$8,%ecx
-	movdqa	32(%ebx),%xmm4
-	movdqa	%xmm0,%xmm2
-	movdqu	%xmm0,-16(%edx)
-.L101loop_key128:
-.byte	102,15,56,0,197
-.byte	102,15,56,221,196
-	pslld	$1,%xmm4
-	leal	16(%edx),%edx
-	movdqa	%xmm2,%xmm3
-	pslldq	$4,%xmm2
-	pxor	%xmm2,%xmm3
-	pslldq	$4,%xmm2
-	pxor	%xmm2,%xmm3
-	pslldq	$4,%xmm2
-	pxor	%xmm3,%xmm2
-	pxor	%xmm2,%xmm0
-	movdqu	%xmm0,-16(%edx)
-	movdqa	%xmm0,%xmm2
-	decl	%ecx
-	jnz	.L101loop_key128
-	movdqa	48(%ebx),%xmm4
-.byte	102,15,56,0,197
-.byte	102,15,56,221,196
-	pslld	$1,%xmm4
-	movdqa	%xmm2,%xmm3
-	pslldq	$4,%xmm2
-	pxor	%xmm2,%xmm3
-	pslldq	$4,%xmm2
-	pxor	%xmm2,%xmm3
-	pslldq	$4,%xmm2
-	pxor	%xmm3,%xmm2
-	pxor	%xmm2,%xmm0
-	movdqu	%xmm0,(%edx)
-	movdqa	%xmm0,%xmm2
-.byte	102,15,56,0,197
-.byte	102,15,56,221,196
-	movdqa	%xmm2,%xmm3
-	pslldq	$4,%xmm2
-	pxor	%xmm2,%xmm3
-	pslldq	$4,%xmm2
-	pxor	%xmm2,%xmm3
-	pslldq	$4,%xmm2
-	pxor	%xmm3,%xmm2
-	pxor	%xmm2,%xmm0
-	movdqu	%xmm0,16(%edx)
-	movl	$9,%ecx
-	movl	%ecx,96(%edx)
-	jmp	.L100good_key
-.align	16
-.L09412rounds:
+.L08812rounds:
 	movq	16(%eax),%xmm2
-	cmpl	$268435456,%ebp
-	je	.L10212rounds_alt
 	movl	$11,%ecx
 	movups	%xmm0,-16(%edx)
 .byte	102,15,58,223,202,1
-	call	.L103key_192a_cold
+	call	.L093key_192a_cold
 .byte	102,15,58,223,202,2
-	call	.L104key_192b
+	call	.L094key_192b
 .byte	102,15,58,223,202,4
-	call	.L105key_192a
+	call	.L095key_192a
 .byte	102,15,58,223,202,8
-	call	.L104key_192b
+	call	.L094key_192b
 .byte	102,15,58,223,202,16
-	call	.L105key_192a
+	call	.L095key_192a
 .byte	102,15,58,223,202,32
-	call	.L104key_192b
+	call	.L094key_192b
 .byte	102,15,58,223,202,64
-	call	.L105key_192a
+	call	.L095key_192a
 .byte	102,15,58,223,202,128
-	call	.L104key_192b
+	call	.L094key_192b
 	movups	%xmm0,(%edx)
 	movl	%ecx,48(%edx)
-	jmp	.L100good_key
+	xorl	%eax,%eax
+	ret
 .align	16
-.L105key_192a:
+.L095key_192a:
 	movups	%xmm0,(%edx)
 	leal	16(%edx),%edx
 .align	16
-.L103key_192a_cold:
+.L093key_192a_cold:
 	movaps	%xmm2,%xmm5
-.L106key_192b_warm:
+.L096key_192b_warm:
 	shufps	$16,%xmm0,%xmm4
 	movdqa	%xmm2,%xmm3
 	xorps	%xmm4,%xmm0
@@ -2255,90 +2050,56 @@ _aesni_set_encrypt_key:
 	pxor	%xmm3,%xmm2
 	ret
 .align	16
-.L104key_192b:
+.L094key_192b:
 	movaps	%xmm0,%xmm3
 	shufps	$68,%xmm0,%xmm5
 	movups	%xmm5,(%edx)
 	shufps	$78,%xmm2,%xmm3
 	movups	%xmm3,16(%edx)
 	leal	32(%edx),%edx
-	jmp	.L106key_192b_warm
-.align	16
-.L10212rounds_alt:
-	movdqa	16(%ebx),%xmm5
-	movdqa	32(%ebx),%xmm4
-	movl	$8,%ecx
-	movdqu	%xmm0,-16(%edx)
-.L107loop_key192:
-	movq	%xmm2,(%edx)
-	movdqa	%xmm2,%xmm1
-.byte	102,15,56,0,213
-.byte	102,15,56,221,212
-	pslld	$1,%xmm4
-	leal	24(%edx),%edx
-	movdqa	%xmm0,%xmm3
-	pslldq	$4,%xmm0
-	pxor	%xmm0,%xmm3
-	pslldq	$4,%xmm0
-	pxor	%xmm0,%xmm3
-	pslldq	$4,%xmm0
-	pxor	%xmm3,%xmm0
-	pshufd	$255,%xmm0,%xmm3
-	pxor	%xmm1,%xmm3
-	pslldq	$4,%xmm1
-	pxor	%xmm1,%xmm3
-	pxor	%xmm2,%xmm0
-	pxor	%xmm3,%xmm2
-	movdqu	%xmm0,-16(%edx)
-	decl	%ecx
-	jnz	.L107loop_key192
-	movl	$11,%ecx
-	movl	%ecx,32(%edx)
-	jmp	.L100good_key
+	jmp	.L096key_192b_warm
 .align	16
-.L09314rounds:
+.L08714rounds:
 	movups	16(%eax),%xmm2
-	leal	16(%edx),%edx
-	cmpl	$268435456,%ebp
-	je	.L10814rounds_alt
 	movl	$13,%ecx
+	leal	16(%edx),%edx
 	movups	%xmm0,-32(%edx)
 	movups	%xmm2,-16(%edx)
 .byte	102,15,58,223,202,1
-	call	.L109key_256a_cold
+	call	.L097key_256a_cold
 .byte	102,15,58,223,200,1
-	call	.L110key_256b
+	call	.L098key_256b
 .byte	102,15,58,223,202,2
-	call	.L111key_256a
+	call	.L099key_256a
 .byte	102,15,58,223,200,2
-	call	.L110key_256b
+	call	.L098key_256b
 .byte	102,15,58,223,202,4
-	call	.L111key_256a
+	call	.L099key_256a
 .byte	102,15,58,223,200,4
-	call	.L110key_256b
+	call	.L098key_256b
 .byte	102,15,58,223,202,8
-	call	.L111key_256a
+	call	.L099key_256a
 .byte	102,15,58,223,200,8
-	call	.L110key_256b
+	call	.L098key_256b
 .byte	102,15,58,223,202,16
-	call	.L111key_256a
+	call	.L099key_256a
 .byte	102,15,58,223,200,16
-	call	.L110key_256b
+	call	.L098key_256b
 .byte	102,15,58,223,202,32
-	call	.L111key_256a
+	call	.L099key_256a
 .byte	102,15,58,223,200,32
-	call	.L110key_256b
+	call	.L098key_256b
 .byte	102,15,58,223,202,64
-	call	.L111key_256a
+	call	.L099key_256a
 	movups	%xmm0,(%edx)
 	movl	%ecx,16(%edx)
 	xorl	%eax,%eax
-	jmp	.L100good_key
+	ret
 .align	16
-.L111key_256a:
+.L099key_256a:
 	movups	%xmm2,(%edx)
 	leal	16(%edx),%edx
-.L109key_256a_cold:
+.L097key_256a_cold:
 	shufps	$16,%xmm0,%xmm4
 	xorps	%xmm4,%xmm0
 	shufps	$140,%xmm0,%xmm4
@@ -2347,7 +2108,7 @@ _aesni_set_encrypt_key:
 	xorps	%xmm1,%xmm0
 	ret
 .align	16
-.L110key_256b:
+.L098key_256b:
 	movups	%xmm0,(%edx)
 	leal	16(%edx),%edx
 	shufps	$16,%xmm2,%xmm4
@@ -2357,70 +2118,13 @@ _aesni_set_encrypt_key:
 	shufps	$170,%xmm1,%xmm1
 	xorps	%xmm1,%xmm2
 	ret
-.align	16
-.L10814rounds_alt:
-	movdqa	(%ebx),%xmm5
-	movdqa	32(%ebx),%xmm4
-	movl	$7,%ecx
-	movdqu	%xmm0,-32(%edx)
-	movdqa	%xmm2,%xmm1
-	movdqu	%xmm2,-16(%edx)
-.L112loop_key256:
-.byte	102,15,56,0,213
-.byte	102,15,56,221,212
-	movdqa	%xmm0,%xmm3
-	pslldq	$4,%xmm0
-	pxor	%xmm0,%xmm3
-	pslldq	$4,%xmm0
-	pxor	%xmm0,%xmm3
-	pslldq	$4,%xmm0
-	pxor	%xmm3,%xmm0
-	pslld	$1,%xmm4
-	pxor	%xmm2,%xmm0
-	movdqu	%xmm0,(%edx)
-	decl	%ecx
-	jz	.L113done_key256
-	pshufd	$255,%xmm0,%xmm2
-	pxor	%xmm3,%xmm3
-.byte	102,15,56,221,211
-	movdqa	%xmm1,%xmm3
-	pslldq	$4,%xmm1
-	pxor	%xmm1,%xmm3
-	pslldq	$4,%xmm1
-	pxor	%xmm1,%xmm3
-	pslldq	$4,%xmm1
-	pxor	%xmm3,%xmm1
-	pxor	%xmm1,%xmm2
-	movdqu	%xmm2,16(%edx)
-	leal	32(%edx),%edx
-	movdqa	%xmm2,%xmm1
-	jmp	.L112loop_key256
-.L113done_key256:
-	movl	$13,%ecx
-	movl	%ecx,16(%edx)
-.L100good_key:
-	pxor	%xmm0,%xmm0
-	pxor	%xmm1,%xmm1
-	pxor	%xmm2,%xmm2
-	pxor	%xmm3,%xmm3
-	pxor	%xmm4,%xmm4
-	pxor	%xmm5,%xmm5
-	xorl	%eax,%eax
-	popl	%ebx
-	popl	%ebp
-	ret
 .align	4
-.L091bad_pointer:
+.L086bad_pointer:
 	movl	$-1,%eax
-	popl	%ebx
-	popl	%ebp
 	ret
 .align	4
-.L095bad_keybits:
-	pxor	%xmm0,%xmm0
+.L089bad_keybits:
 	movl	$-2,%eax
-	popl	%ebx
-	popl	%ebp
 	ret
 .size	_aesni_set_encrypt_key,.-_aesni_set_encrypt_key
 .globl	aesni_set_encrypt_key
@@ -2446,7 +2150,7 @@ aesni_set_decrypt_key:
 	movl	12(%esp),%edx
 	shll	$4,%ecx
 	testl	%eax,%eax
-	jnz	.L114dec_key_ret
+	jnz	.L100dec_key_ret
 	leal	16(%edx,%ecx,1),%eax
 	movups	(%edx),%xmm0
 	movups	(%eax),%xmm1
@@ -2454,7 +2158,7 @@ aesni_set_decrypt_key:
 	movups	%xmm1,(%edx)
 	leal	16(%edx),%edx
 	leal	-16(%eax),%eax
-.L115dec_key_inverse:
+.L101dec_key_inverse:
 	movups	(%edx),%xmm0
 	movups	(%eax),%xmm1
 .byte	102,15,56,219,192
@@ -2464,26 +2168,20 @@ aesni_set_decrypt_key:
 	movups	%xmm0,16(%eax)
 	movups	%xmm1,-16(%edx)
 	cmpl	%edx,%eax
-	ja	.L115dec_key_inverse
+	ja	.L101dec_key_inverse
 	movups	(%edx),%xmm0
 .byte	102,15,56,219,192
 	movups	%xmm0,(%edx)
-	pxor	%xmm0,%xmm0
-	pxor	%xmm1,%xmm1
 	xorl	%eax,%eax
-.L114dec_key_ret:
+.L100dec_key_ret:
 	ret
 .size	aesni_set_decrypt_key,.-.L_aesni_set_decrypt_key_begin
-.align	64
-.Lkey_const:
-.long	202313229,202313229,202313229,202313229
-.long	67569157,67569157,67569157,67569157
-.long	1,1,1,1
-.long	27,27,27,27
 .byte	65,69,83,32,102,111,114,32,73,110,116,101,108,32,65,69
 .byte	83,45,78,73,44,32,67,82,89,80,84,79,71,65,77,83
 .byte	32,98,121,32,60,97,112,112,114,111,64,111,112,101,110,115
 .byte	115,108,46,111,114,103,62,0
-.comm	_gnutls_x86_cpuid_s,16,4
+
 
 .section .note.GNU-stack,"",%progbits
+
+
diff --git a/lib/accelerated/x86/macosx/aesni-x86.s b/lib/accelerated/x86/macosx/aesni-x86.s
index 275ab58ec5..09ca1cbc5c 100644
--- a/lib/accelerated/x86/macosx/aesni-x86.s
+++ b/lib/accelerated/x86/macosx/aesni-x86.s
@@ -59,10 +59,7 @@ L000enc1_loop_1:
 	leal	16(%edx),%edx
 	jnz	L000enc1_loop_1
 .byte	102,15,56,221,209
-	pxor	%xmm0,%xmm0
-	pxor	%xmm1,%xmm1
 	movups	%xmm2,(%eax)
-	pxor	%xmm2,%xmm2
 	ret
 .globl	_aesni_decrypt
 .align	4
@@ -84,84 +81,30 @@ L001dec1_loop_2:
 	leal	16(%edx),%edx
 	jnz	L001dec1_loop_2
 .byte	102,15,56,223,209
-	pxor	%xmm0,%xmm0
-	pxor	%xmm1,%xmm1
 	movups	%xmm2,(%eax)
-	pxor	%xmm2,%xmm2
-	ret
-.align	4
-__aesni_encrypt2:
-	movups	(%edx),%xmm0
-	shll	$4,%ecx
-	movups	16(%edx),%xmm1
-	xorps	%xmm0,%xmm2
-	pxor	%xmm0,%xmm3
-	movups	32(%edx),%xmm0
-	leal	32(%edx,%ecx,1),%edx
-	negl	%ecx
-	addl	$16,%ecx
-L002enc2_loop:
-.byte	102,15,56,220,209
-.byte	102,15,56,220,217
-	movups	(%edx,%ecx,1),%xmm1
-	addl	$32,%ecx
-.byte	102,15,56,220,208
-.byte	102,15,56,220,216
-	movups	-16(%edx,%ecx,1),%xmm0
-	jnz	L002enc2_loop
-.byte	102,15,56,220,209
-.byte	102,15,56,220,217
-.byte	102,15,56,221,208
-.byte	102,15,56,221,216
-	ret
-.align	4
-__aesni_decrypt2:
-	movups	(%edx),%xmm0
-	shll	$4,%ecx
-	movups	16(%edx),%xmm1
-	xorps	%xmm0,%xmm2
-	pxor	%xmm0,%xmm3
-	movups	32(%edx),%xmm0
-	leal	32(%edx,%ecx,1),%edx
-	negl	%ecx
-	addl	$16,%ecx
-L003dec2_loop:
-.byte	102,15,56,222,209
-.byte	102,15,56,222,217
-	movups	(%edx,%ecx,1),%xmm1
-	addl	$32,%ecx
-.byte	102,15,56,222,208
-.byte	102,15,56,222,216
-	movups	-16(%edx,%ecx,1),%xmm0
-	jnz	L003dec2_loop
-.byte	102,15,56,222,209
-.byte	102,15,56,222,217
-.byte	102,15,56,223,208
-.byte	102,15,56,223,216
 	ret
 .align	4
 __aesni_encrypt3:
 	movups	(%edx),%xmm0
-	shll	$4,%ecx
+	shrl	$1,%ecx
 	movups	16(%edx),%xmm1
+	leal	32(%edx),%edx
 	xorps	%xmm0,%xmm2
 	pxor	%xmm0,%xmm3
 	pxor	%xmm0,%xmm4
-	movups	32(%edx),%xmm0
-	leal	32(%edx,%ecx,1),%edx
-	negl	%ecx
-	addl	$16,%ecx
-L004enc3_loop:
+	movups	(%edx),%xmm0
+L002enc3_loop:
 .byte	102,15,56,220,209
 .byte	102,15,56,220,217
+	decl	%ecx
 .byte	102,15,56,220,225
-	movups	(%edx,%ecx,1),%xmm1
-	addl	$32,%ecx
+	movups	16(%edx),%xmm1
 .byte	102,15,56,220,208
 .byte	102,15,56,220,216
+	leal	32(%edx),%edx
 .byte	102,15,56,220,224
-	movups	-16(%edx,%ecx,1),%xmm0
-	jnz	L004enc3_loop
+	movups	(%edx),%xmm0
+	jnz	L002enc3_loop
 .byte	102,15,56,220,209
 .byte	102,15,56,220,217
 .byte	102,15,56,220,225
@@ -172,26 +115,25 @@ L004enc3_loop:
 .align	4
 __aesni_decrypt3:
 	movups	(%edx),%xmm0
-	shll	$4,%ecx
+	shrl	$1,%ecx
 	movups	16(%edx),%xmm1
+	leal	32(%edx),%edx
 	xorps	%xmm0,%xmm2
 	pxor	%xmm0,%xmm3
 	pxor	%xmm0,%xmm4
-	movups	32(%edx),%xmm0
-	leal	32(%edx,%ecx,1),%edx
-	negl	%ecx
-	addl	$16,%ecx
-L005dec3_loop:
+	movups	(%edx),%xmm0
+L003dec3_loop:
 .byte	102,15,56,222,209
 .byte	102,15,56,222,217
+	decl	%ecx
 .byte	102,15,56,222,225
-	movups	(%edx,%ecx,1),%xmm1
-	addl	$32,%ecx
+	movups	16(%edx),%xmm1
 .byte	102,15,56,222,208
 .byte	102,15,56,222,216
+	leal	32(%edx),%edx
 .byte	102,15,56,222,224
-	movups	-16(%edx,%ecx,1),%xmm0
-	jnz	L005dec3_loop
+	movups	(%edx),%xmm0
+	jnz	L003dec3_loop
 .byte	102,15,56,222,209
 .byte	102,15,56,222,217
 .byte	102,15,56,222,225
@@ -203,29 +145,27 @@ L005dec3_loop:
 __aesni_encrypt4:
 	movups	(%edx),%xmm0
 	movups	16(%edx),%xmm1
-	shll	$4,%ecx
+	shrl	$1,%ecx
+	leal	32(%edx),%edx
 	xorps	%xmm0,%xmm2
 	pxor	%xmm0,%xmm3
 	pxor	%xmm0,%xmm4
 	pxor	%xmm0,%xmm5
-	movups	32(%edx),%xmm0
-	leal	32(%edx,%ecx,1),%edx
-	negl	%ecx
-.byte	15,31,64,0
-	addl	$16,%ecx
-L006enc4_loop:
+	movups	(%edx),%xmm0
+L004enc4_loop:
 .byte	102,15,56,220,209
 .byte	102,15,56,220,217
+	decl	%ecx
 .byte	102,15,56,220,225
 .byte	102,15,56,220,233
-	movups	(%edx,%ecx,1),%xmm1
-	addl	$32,%ecx
+	movups	16(%edx),%xmm1
 .byte	102,15,56,220,208
 .byte	102,15,56,220,216
+	leal	32(%edx),%edx
 .byte	102,15,56,220,224
 .byte	102,15,56,220,232
-	movups	-16(%edx,%ecx,1),%xmm0
-	jnz	L006enc4_loop
+	movups	(%edx),%xmm0
+	jnz	L004enc4_loop
 .byte	102,15,56,220,209
 .byte	102,15,56,220,217
 .byte	102,15,56,220,225
@@ -239,29 +179,27 @@ L006enc4_loop:
 __aesni_decrypt4:
 	movups	(%edx),%xmm0
 	movups	16(%edx),%xmm1
-	shll	$4,%ecx
+	shrl	$1,%ecx
+	leal	32(%edx),%edx
 	xorps	%xmm0,%xmm2
 	pxor	%xmm0,%xmm3
 	pxor	%xmm0,%xmm4
 	pxor	%xmm0,%xmm5
-	movups	32(%edx),%xmm0
-	leal	32(%edx,%ecx,1),%edx
-	negl	%ecx
-.byte	15,31,64,0
-	addl	$16,%ecx
-L007dec4_loop:
+	movups	(%edx),%xmm0
+L005dec4_loop:
 .byte	102,15,56,222,209
 .byte	102,15,56,222,217
+	decl	%ecx
 .byte	102,15,56,222,225
 .byte	102,15,56,222,233
-	movups	(%edx,%ecx,1),%xmm1
-	addl	$32,%ecx
+	movups	16(%edx),%xmm1
 .byte	102,15,56,222,208
 .byte	102,15,56,222,216
+	leal	32(%edx),%edx
 .byte	102,15,56,222,224
 .byte	102,15,56,222,232
-	movups	-16(%edx,%ecx,1),%xmm0
-	jnz	L007dec4_loop
+	movups	(%edx),%xmm0
+	jnz	L005dec4_loop
 .byte	102,15,56,222,209
 .byte	102,15,56,222,217
 .byte	102,15,56,222,225
@@ -274,42 +212,45 @@ L007dec4_loop:
 .align	4
 __aesni_encrypt6:
 	movups	(%edx),%xmm0
-	shll	$4,%ecx
+	shrl	$1,%ecx
 	movups	16(%edx),%xmm1
+	leal	32(%edx),%edx
 	xorps	%xmm0,%xmm2
 	pxor	%xmm0,%xmm3
-	pxor	%xmm0,%xmm4
 .byte	102,15,56,220,209
-	pxor	%xmm0,%xmm5
-	pxor	%xmm0,%xmm6
+	pxor	%xmm0,%xmm4
 .byte	102,15,56,220,217
-	leal	32(%edx,%ecx,1),%edx
-	negl	%ecx
+	pxor	%xmm0,%xmm5
+	decl	%ecx
 .byte	102,15,56,220,225
+	pxor	%xmm0,%xmm6
+.byte	102,15,56,220,233
 	pxor	%xmm0,%xmm7
-	movups	(%edx,%ecx,1),%xmm0
-	addl	$16,%ecx
-	jmp	L008_aesni_encrypt6_inner
+.byte	102,15,56,220,241
+	movups	(%edx),%xmm0
+.byte	102,15,56,220,249
+	jmp	L_aesni_encrypt6_enter
 .align	4,0x90
-L009enc6_loop:
+L006enc6_loop:
 .byte	102,15,56,220,209
 .byte	102,15,56,220,217
+	decl	%ecx
 .byte	102,15,56,220,225
-L008_aesni_encrypt6_inner:
 .byte	102,15,56,220,233
 .byte	102,15,56,220,241
 .byte	102,15,56,220,249
+.align	4,0x90
 L_aesni_encrypt6_enter:
-	movups	(%edx,%ecx,1),%xmm1
-	addl	$32,%ecx
+	movups	16(%edx),%xmm1
 .byte	102,15,56,220,208
 .byte	102,15,56,220,216
+	leal	32(%edx),%edx
 .byte	102,15,56,220,224
 .byte	102,15,56,220,232
 .byte	102,15,56,220,240
 .byte	102,15,56,220,248
-	movups	-16(%edx,%ecx,1),%xmm0
-	jnz	L009enc6_loop
+	movups	(%edx),%xmm0
+	jnz	L006enc6_loop
 .byte	102,15,56,220,209
 .byte	102,15,56,220,217
 .byte	102,15,56,220,225
@@ -326,42 +267,45 @@ L_aesni_encrypt6_enter:
 .align	4
 __aesni_decrypt6:
 	movups	(%edx),%xmm0
-	shll	$4,%ecx
+	shrl	$1,%ecx
 	movups	16(%edx),%xmm1
+	leal	32(%edx),%edx
 	xorps	%xmm0,%xmm2
 	pxor	%xmm0,%xmm3
-	pxor	%xmm0,%xmm4
 .byte	102,15,56,222,209
-	pxor	%xmm0,%xmm5
-	pxor	%xmm0,%xmm6
+	pxor	%xmm0,%xmm4
 .byte	102,15,56,222,217
-	leal	32(%edx,%ecx,1),%edx
-	negl	%ecx
+	pxor	%xmm0,%xmm5
+	decl	%ecx
 .byte	102,15,56,222,225
+	pxor	%xmm0,%xmm6
+.byte	102,15,56,222,233
 	pxor	%xmm0,%xmm7
-	movups	(%edx,%ecx,1),%xmm0
-	addl	$16,%ecx
-	jmp	L010_aesni_decrypt6_inner
+.byte	102,15,56,222,241
+	movups	(%edx),%xmm0
+.byte	102,15,56,222,249
+	jmp	L_aesni_decrypt6_enter
 .align	4,0x90
-L011dec6_loop:
+L007dec6_loop:
 .byte	102,15,56,222,209
 .byte	102,15,56,222,217
+	decl	%ecx
 .byte	102,15,56,222,225
-L010_aesni_decrypt6_inner:
 .byte	102,15,56,222,233
 .byte	102,15,56,222,241
 .byte	102,15,56,222,249
+.align	4,0x90
 L_aesni_decrypt6_enter:
-	movups	(%edx,%ecx,1),%xmm1
-	addl	$32,%ecx
+	movups	16(%edx),%xmm1
 .byte	102,15,56,222,208
 .byte	102,15,56,222,216
+	leal	32(%edx),%edx
 .byte	102,15,56,222,224
 .byte	102,15,56,222,232
 .byte	102,15,56,222,240
 .byte	102,15,56,222,248
-	movups	-16(%edx,%ecx,1),%xmm0
-	jnz	L011dec6_loop
+	movups	(%edx),%xmm0
+	jnz	L007dec6_loop
 .byte	102,15,56,222,209
 .byte	102,15,56,222,217
 .byte	102,15,56,222,225
@@ -389,14 +333,14 @@ L_aesni_ecb_encrypt_begin:
 	movl	32(%esp),%edx
 	movl	36(%esp),%ebx
 	andl	$-16,%eax
-	jz	L012ecb_ret
+	jz	L008ecb_ret
 	movl	240(%edx),%ecx
 	testl	%ebx,%ebx
-	jz	L013ecb_decrypt
+	jz	L009ecb_decrypt
 	movl	%edx,%ebp
 	movl	%ecx,%ebx
 	cmpl	$96,%eax
-	jb	L014ecb_enc_tail
+	jb	L010ecb_enc_tail
 	movdqu	(%esi),%xmm2
 	movdqu	16(%esi),%xmm3
 	movdqu	32(%esi),%xmm4
@@ -405,9 +349,9 @@ L_aesni_ecb_encrypt_begin:
 	movdqu	80(%esi),%xmm7
 	leal	96(%esi),%esi
 	subl	$96,%eax
-	jmp	L015ecb_enc_loop6_enter
+	jmp	L011ecb_enc_loop6_enter
 .align	4,0x90
-L016ecb_enc_loop6:
+L012ecb_enc_loop6:
 	movups	%xmm2,(%edi)
 	movdqu	(%esi),%xmm2
 	movups	%xmm3,16(%edi)
@@ -422,12 +366,12 @@ L016ecb_enc_loop6:
 	leal	96(%edi),%edi
 	movdqu	80(%esi),%xmm7
 	leal	96(%esi),%esi
-L015ecb_enc_loop6_enter:
+L011ecb_enc_loop6_enter:
 	call	__aesni_encrypt6
 	movl	%ebp,%edx
 	movl	%ebx,%ecx
 	subl	$96,%eax
-	jnc	L016ecb_enc_loop6
+	jnc	L012ecb_enc_loop6
 	movups	%xmm2,(%edi)
 	movups	%xmm3,16(%edi)
 	movups	%xmm4,32(%edi)
@@ -436,18 +380,18 @@ L015ecb_enc_loop6_enter:
 	movups	%xmm7,80(%edi)
 	leal	96(%edi),%edi
 	addl	$96,%eax
-	jz	L012ecb_ret
-L014ecb_enc_tail:
+	jz	L008ecb_ret
+L010ecb_enc_tail:
 	movups	(%esi),%xmm2
 	cmpl	$32,%eax
-	jb	L017ecb_enc_one
+	jb	L013ecb_enc_one
 	movups	16(%esi),%xmm3
-	je	L018ecb_enc_two
+	je	L014ecb_enc_two
 	movups	32(%esi),%xmm4
 	cmpl	$64,%eax
-	jb	L019ecb_enc_three
+	jb	L015ecb_enc_three
 	movups	48(%esi),%xmm5
-	je	L020ecb_enc_four
+	je	L016ecb_enc_four
 	movups	64(%esi),%xmm6
 	xorps	%xmm7,%xmm7
 	call	__aesni_encrypt6
@@ -456,49 +400,50 @@ L014ecb_enc_tail:
 	movups	%xmm4,32(%edi)
 	movups	%xmm5,48(%edi)
 	movups	%xmm6,64(%edi)
-	jmp	L012ecb_ret
+	jmp	L008ecb_ret
 .align	4,0x90
-L017ecb_enc_one:
+L013ecb_enc_one:
 	movups	(%edx),%xmm0
 	movups	16(%edx),%xmm1
 	leal	32(%edx),%edx
 	xorps	%xmm0,%xmm2
-L021enc1_loop_3:
+L017enc1_loop_3:
 .byte	102,15,56,220,209
 	decl	%ecx
 	movups	(%edx),%xmm1
 	leal	16(%edx),%edx
-	jnz	L021enc1_loop_3
+	jnz	L017enc1_loop_3
 .byte	102,15,56,221,209
 	movups	%xmm2,(%edi)
-	jmp	L012ecb_ret
+	jmp	L008ecb_ret
 .align	4,0x90
-L018ecb_enc_two:
-	call	__aesni_encrypt2
+L014ecb_enc_two:
+	xorps	%xmm4,%xmm4
+	call	__aesni_encrypt3
 	movups	%xmm2,(%edi)
 	movups	%xmm3,16(%edi)
-	jmp	L012ecb_ret
+	jmp	L008ecb_ret
 .align	4,0x90
-L019ecb_enc_three:
+L015ecb_enc_three:
 	call	__aesni_encrypt3
 	movups	%xmm2,(%edi)
 	movups	%xmm3,16(%edi)
 	movups	%xmm4,32(%edi)
-	jmp	L012ecb_ret
+	jmp	L008ecb_ret
 .align	4,0x90
-L020ecb_enc_four:
+L016ecb_enc_four:
 	call	__aesni_encrypt4
 	movups	%xmm2,(%edi)
 	movups	%xmm3,16(%edi)
 	movups	%xmm4,32(%edi)
 	movups	%xmm5,48(%edi)
-	jmp	L012ecb_ret
+	jmp	L008ecb_ret
 .align	4,0x90
-L013ecb_decrypt:
+L009ecb_decrypt:
 	movl	%edx,%ebp
 	movl	%ecx,%ebx
 	cmpl	$96,%eax
-	jb	L022ecb_dec_tail
+	jb	L018ecb_dec_tail
 	movdqu	(%esi),%xmm2
 	movdqu	16(%esi),%xmm3
 	movdqu	32(%esi),%xmm4
@@ -507,9 +452,9 @@ L013ecb_decrypt:
 	movdqu	80(%esi),%xmm7
 	leal	96(%esi),%esi
 	subl	$96,%eax
-	jmp	L023ecb_dec_loop6_enter
+	jmp	L019ecb_dec_loop6_enter
 .align	4,0x90
-L024ecb_dec_loop6:
+L020ecb_dec_loop6:
 	movups	%xmm2,(%edi)
 	movdqu	(%esi),%xmm2
 	movups	%xmm3,16(%edi)
@@ -524,12 +469,12 @@ L024ecb_dec_loop6:
 	leal	96(%edi),%edi
 	movdqu	80(%esi),%xmm7
 	leal	96(%esi),%esi
-L023ecb_dec_loop6_enter:
+L019ecb_dec_loop6_enter:
 	call	__aesni_decrypt6
 	movl	%ebp,%edx
 	movl	%ebx,%ecx
 	subl	$96,%eax
-	jnc	L024ecb_dec_loop6
+	jnc	L020ecb_dec_loop6
 	movups	%xmm2,(%edi)
 	movups	%xmm3,16(%edi)
 	movups	%xmm4,32(%edi)
@@ -538,18 +483,18 @@ L023ecb_dec_loop6_enter:
 	movups	%xmm7,80(%edi)
 	leal	96(%edi),%edi
 	addl	$96,%eax
-	jz	L012ecb_ret
-L022ecb_dec_tail:
+	jz	L008ecb_ret
+L018ecb_dec_tail:
 	movups	(%esi),%xmm2
 	cmpl	$32,%eax
-	jb	L025ecb_dec_one
+	jb	L021ecb_dec_one
 	movups	16(%esi),%xmm3
-	je	L026ecb_dec_two
+	je	L022ecb_dec_two
 	movups	32(%esi),%xmm4
 	cmpl	$64,%eax
-	jb	L027ecb_dec_three
+	jb	L023ecb_dec_three
 	movups	48(%esi),%xmm5
-	je	L028ecb_dec_four
+	je	L024ecb_dec_four
 	movups	64(%esi),%xmm6
 	xorps	%xmm7,%xmm7
 	call	__aesni_decrypt6
@@ -558,51 +503,44 @@ L022ecb_dec_tail:
 	movups	%xmm4,32(%edi)
 	movups	%xmm5,48(%edi)
 	movups	%xmm6,64(%edi)
-	jmp	L012ecb_ret
+	jmp	L008ecb_ret
 .align	4,0x90
-L025ecb_dec_one:
+L021ecb_dec_one:
 	movups	(%edx),%xmm0
 	movups	16(%edx),%xmm1
 	leal	32(%edx),%edx
 	xorps	%xmm0,%xmm2
-L029dec1_loop_4:
+L025dec1_loop_4:
 .byte	102,15,56,222,209
 	decl	%ecx
 	movups	(%edx),%xmm1
 	leal	16(%edx),%edx
-	jnz	L029dec1_loop_4
+	jnz	L025dec1_loop_4
 .byte	102,15,56,223,209
 	movups	%xmm2,(%edi)
-	jmp	L012ecb_ret
+	jmp	L008ecb_ret
 .align	4,0x90
-L026ecb_dec_two:
-	call	__aesni_decrypt2
+L022ecb_dec_two:
+	xorps	%xmm4,%xmm4
+	call	__aesni_decrypt3
 	movups	%xmm2,(%edi)
 	movups	%xmm3,16(%edi)
-	jmp	L012ecb_ret
+	jmp	L008ecb_ret
 .align	4,0x90
-L027ecb_dec_three:
+L023ecb_dec_three:
 	call	__aesni_decrypt3
 	movups	%xmm2,(%edi)
 	movups	%xmm3,16(%edi)
 	movups	%xmm4,32(%edi)
-	jmp	L012ecb_ret
+	jmp	L008ecb_ret
 .align	4,0x90
-L028ecb_dec_four:
+L024ecb_dec_four:
 	call	__aesni_decrypt4
 	movups	%xmm2,(%edi)
 	movups	%xmm3,16(%edi)
 	movups	%xmm4,32(%edi)
 	movups	%xmm5,48(%edi)
-L012ecb_ret:
-	pxor	%xmm0,%xmm0
-	pxor	%xmm1,%xmm1
-	pxor	%xmm2,%xmm2
-	pxor	%xmm3,%xmm3
-	pxor	%xmm4,%xmm4
-	pxor	%xmm5,%xmm5
-	pxor	%xmm6,%xmm6
-	pxor	%xmm7,%xmm7
+L008ecb_ret:
 	popl	%edi
 	popl	%esi
 	popl	%ebx
@@ -639,56 +577,48 @@ L_aesni_ccm64_encrypt_blocks_begin:
 	movl	%ebp,20(%esp)
 	movl	%ebp,24(%esp)
 	movl	%ebp,28(%esp)
-	shll	$4,%ecx
-	movl	$16,%ebx
+	shrl	$1,%ecx
 	leal	(%edx),%ebp
 	movdqa	(%esp),%xmm5
 	movdqa	%xmm7,%xmm2
-	leal	32(%edx,%ecx,1),%edx
-	subl	%ecx,%ebx
+	movl	%ecx,%ebx
 .byte	102,15,56,0,253
-L030ccm64_enc_outer:
+L026ccm64_enc_outer:
 	movups	(%ebp),%xmm0
 	movl	%ebx,%ecx
 	movups	(%esi),%xmm6
 	xorps	%xmm0,%xmm2
 	movups	16(%ebp),%xmm1
 	xorps	%xmm6,%xmm0
+	leal	32(%ebp),%edx
 	xorps	%xmm0,%xmm3
-	movups	32(%ebp),%xmm0
-L031ccm64_enc2_loop:
+	movups	(%edx),%xmm0
+L027ccm64_enc2_loop:
 .byte	102,15,56,220,209
+	decl	%ecx
 .byte	102,15,56,220,217
-	movups	(%edx,%ecx,1),%xmm1
-	addl	$32,%ecx
+	movups	16(%edx),%xmm1
 .byte	102,15,56,220,208
+	leal	32(%edx),%edx
 .byte	102,15,56,220,216
-	movups	-16(%edx,%ecx,1),%xmm0
-	jnz	L031ccm64_enc2_loop
+	movups	(%edx),%xmm0
+	jnz	L027ccm64_enc2_loop
 .byte	102,15,56,220,209
 .byte	102,15,56,220,217
 	paddq	16(%esp),%xmm7
-	decl	%eax
 .byte	102,15,56,221,208
 .byte	102,15,56,221,216
+	decl	%eax
 	leal	16(%esi),%esi
 	xorps	%xmm2,%xmm6
 	movdqa	%xmm7,%xmm2
 	movups	%xmm6,(%edi)
-.byte	102,15,56,0,213
 	leal	16(%edi),%edi
-	jnz	L030ccm64_enc_outer
+.byte	102,15,56,0,213
+	jnz	L026ccm64_enc_outer
 	movl	48(%esp),%esp
 	movl	40(%esp),%edi
 	movups	%xmm3,(%edi)
-	pxor	%xmm0,%xmm0
-	pxor	%xmm1,%xmm1
-	pxor	%xmm2,%xmm2
-	pxor	%xmm3,%xmm3
-	pxor	%xmm4,%xmm4
-	pxor	%xmm5,%xmm5
-	pxor	%xmm6,%xmm6
-	pxor	%xmm7,%xmm7
 	popl	%edi
 	popl	%esi
 	popl	%ebx
@@ -734,82 +664,71 @@ L_aesni_ccm64_decrypt_blocks_begin:
 	movups	16(%edx),%xmm1
 	leal	32(%edx),%edx
 	xorps	%xmm0,%xmm2
-L032enc1_loop_5:
+L028enc1_loop_5:
 .byte	102,15,56,220,209
 	decl	%ecx
 	movups	(%edx),%xmm1
 	leal	16(%edx),%edx
-	jnz	L032enc1_loop_5
+	jnz	L028enc1_loop_5
 .byte	102,15,56,221,209
-	shll	$4,%ebx
-	movl	$16,%ecx
 	movups	(%esi),%xmm6
 	paddq	16(%esp),%xmm7
 	leal	16(%esi),%esi
-	subl	%ebx,%ecx
-	leal	32(%ebp,%ebx,1),%edx
-	movl	%ecx,%ebx
-	jmp	L033ccm64_dec_outer
+	jmp	L029ccm64_dec_outer
 .align	4,0x90
-L033ccm64_dec_outer:
+L029ccm64_dec_outer:
 	xorps	%xmm2,%xmm6
 	movdqa	%xmm7,%xmm2
+	movl	%ebx,%ecx
 	movups	%xmm6,(%edi)
 	leal	16(%edi),%edi
 .byte	102,15,56,0,213
 	subl	$1,%eax
-	jz	L034ccm64_dec_break
+	jz	L030ccm64_dec_break
 	movups	(%ebp),%xmm0
-	movl	%ebx,%ecx
+	shrl	$1,%ecx
 	movups	16(%ebp),%xmm1
 	xorps	%xmm0,%xmm6
+	leal	32(%ebp),%edx
 	xorps	%xmm0,%xmm2
 	xorps	%xmm6,%xmm3
-	movups	32(%ebp),%xmm0
-L035ccm64_dec2_loop:
+	movups	(%edx),%xmm0
+L031ccm64_dec2_loop:
 .byte	102,15,56,220,209
+	decl	%ecx
 .byte	102,15,56,220,217
-	movups	(%edx,%ecx,1),%xmm1
-	addl	$32,%ecx
+	movups	16(%edx),%xmm1
 .byte	102,15,56,220,208
+	leal	32(%edx),%edx
 .byte	102,15,56,220,216
-	movups	-16(%edx,%ecx,1),%xmm0
-	jnz	L035ccm64_dec2_loop
+	movups	(%edx),%xmm0
+	jnz	L031ccm64_dec2_loop
 	movups	(%esi),%xmm6
 	paddq	16(%esp),%xmm7
 .byte	102,15,56,220,209
 .byte	102,15,56,220,217
+	leal	16(%esi),%esi
 .byte	102,15,56,221,208
 .byte	102,15,56,221,216
-	leal	16(%esi),%esi
-	jmp	L033ccm64_dec_outer
+	jmp	L029ccm64_dec_outer
 .align	4,0x90
-L034ccm64_dec_break:
-	movl	240(%ebp),%ecx
+L030ccm64_dec_break:
 	movl	%ebp,%edx
 	movups	(%edx),%xmm0
 	movups	16(%edx),%xmm1
 	xorps	%xmm0,%xmm6
 	leal	32(%edx),%edx
 	xorps	%xmm6,%xmm3
-L036enc1_loop_6:
+L032enc1_loop_6:
 .byte	102,15,56,220,217
 	decl	%ecx
 	movups	(%edx),%xmm1
 	leal	16(%edx),%edx
-	jnz	L036enc1_loop_6
+	jnz	L032enc1_loop_6
 .byte	102,15,56,221,217
 	movl	48(%esp),%esp
 	movl	40(%esp),%edi
 	movups	%xmm3,(%edi)
-	pxor	%xmm0,%xmm0
-	pxor	%xmm1,%xmm1
-	pxor	%xmm2,%xmm2
-	pxor	%xmm3,%xmm3
-	pxor	%xmm4,%xmm4
-	pxor	%xmm5,%xmm5
-	pxor	%xmm6,%xmm6
-	pxor	%xmm7,%xmm7
 	popl	%edi
 	popl	%esi
 	popl	%ebx
@@ -833,7 +752,7 @@ L_aesni_ctr32_encrypt_blocks_begin:
 	andl	$-16,%esp
 	movl	%ebp,80(%esp)
 	cmpl	$1,%eax
-	je	L037ctr32_one_shortcut
+	je	L033ctr32_one_shortcut
 	movdqu	(%ebx),%xmm7
 	movl	$202182159,(%esp)
 	movl	$134810123,4(%esp)
@@ -849,59 +768,63 @@ L_aesni_ctr32_encrypt_blocks_begin:
 .byte	102,15,58,34,253,3
 	movl	240(%edx),%ecx
 	bswap	%ebx
-	pxor	%xmm0,%xmm0
 	pxor	%xmm1,%xmm1
+	pxor	%xmm0,%xmm0
 	movdqa	(%esp),%xmm2
-.byte	102,15,58,34,195,0
+.byte	102,15,58,34,203,0
 	leal	3(%ebx),%ebp
-.byte	102,15,58,34,205,0
+.byte	102,15,58,34,197,0
 	incl	%ebx
-.byte	102,15,58,34,195,1
+.byte	102,15,58,34,203,1
 	incl	%ebp
-.byte	102,15,58,34,205,1
+.byte	102,15,58,34,197,1
 	incl	%ebx
-.byte	102,15,58,34,195,2
+.byte	102,15,58,34,203,2
 	incl	%ebp
-.byte	102,15,58,34,205,2
-	movdqa	%xmm0,48(%esp)
-.byte	102,15,56,0,194
-	movdqu	(%edx),%xmm6
-	movdqa	%xmm1,64(%esp)
+.byte	102,15,58,34,197,2
+	movdqa	%xmm1,48(%esp)
 .byte	102,15,56,0,202
-	pshufd	$192,%xmm0,%xmm2
-	pshufd	$128,%xmm0,%xmm3
+	movdqa	%xmm0,64(%esp)
+.byte	102,15,56,0,194
+	pshufd	$192,%xmm1,%xmm2
+	pshufd	$128,%xmm1,%xmm3
 	cmpl	$6,%eax
-	jb	L038ctr32_tail
-	pxor	%xmm6,%xmm7
-	shll	$4,%ecx
-	movl	$16,%ebx
+	jb	L034ctr32_tail
 	movdqa	%xmm7,32(%esp)
+	shrl	$1,%ecx
 	movl	%edx,%ebp
-	subl	%ecx,%ebx
-	leal	32(%edx,%ecx,1),%edx
+	movl	%ecx,%ebx
 	subl	$6,%eax
-	jmp	L039ctr32_loop6
+	jmp	L035ctr32_loop6
 .align	4,0x90
-L039ctr32_loop6:
-	pshufd	$64,%xmm0,%xmm4
-	movdqa	32(%esp),%xmm0
-	pshufd	$192,%xmm1,%xmm5
+L035ctr32_loop6:
+	pshufd	$64,%xmm1,%xmm4
+	movdqa	32(%esp),%xmm1
+	pshufd	$192,%xmm0,%xmm5
+	por	%xmm1,%xmm2
+	pshufd	$128,%xmm0,%xmm6
+	por	%xmm1,%xmm3
+	pshufd	$64,%xmm0,%xmm7
+	por	%xmm1,%xmm4
+	por	%xmm1,%xmm5
+	por	%xmm1,%xmm6
+	por	%xmm1,%xmm7
+	movups	(%ebp),%xmm0
+	movups	16(%ebp),%xmm1
+	leal	32(%ebp),%edx
+	decl	%ecx
 	pxor	%xmm0,%xmm2
-	pshufd	$128,%xmm1,%xmm6
 	pxor	%xmm0,%xmm3
-	pshufd	$64,%xmm1,%xmm7
-	movups	16(%ebp),%xmm1
-	pxor	%xmm0,%xmm4
-	pxor	%xmm0,%xmm5
 .byte	102,15,56,220,209
-	pxor	%xmm0,%xmm6
-	pxor	%xmm0,%xmm7
+	pxor	%xmm0,%xmm4
 .byte	102,15,56,220,217
-	movups	32(%ebp),%xmm0
-	movl	%ebx,%ecx
+	pxor	%xmm0,%xmm5
 .byte	102,15,56,220,225
+	pxor	%xmm0,%xmm6
 .byte	102,15,56,220,233
+	pxor	%xmm0,%xmm7
 .byte	102,15,56,220,241
+	movups	(%edx),%xmm0
 .byte	102,15,56,220,249
 	call	L_aesni_encrypt6_enter
 	movups	(%esi),%xmm1
@@ -912,51 +835,51 @@ L039ctr32_loop6:
 	movups	%xmm2,(%edi)
 	movdqa	16(%esp),%xmm0
 	xorps	%xmm1,%xmm4
-	movdqa	64(%esp),%xmm1
+	movdqa	48(%esp),%xmm1
 	movups	%xmm3,16(%edi)
 	movups	%xmm4,32(%edi)
 	paddd	%xmm0,%xmm1
-	paddd	48(%esp),%xmm0
+	paddd	64(%esp),%xmm0
 	movdqa	(%esp),%xmm2
 	movups	48(%esi),%xmm3
 	movups	64(%esi),%xmm4
 	xorps	%xmm3,%xmm5
 	movups	80(%esi),%xmm3
 	leal	96(%esi),%esi
-	movdqa	%xmm0,48(%esp)
-.byte	102,15,56,0,194
+	movdqa	%xmm1,48(%esp)
+.byte	102,15,56,0,202
 	xorps	%xmm4,%xmm6
 	movups	%xmm5,48(%edi)
 	xorps	%xmm3,%xmm7
-	movdqa	%xmm1,64(%esp)
-.byte	102,15,56,0,202
+	movdqa	%xmm0,64(%esp)
+.byte	102,15,56,0,194
 	movups	%xmm6,64(%edi)
-	pshufd	$192,%xmm0,%xmm2
+	pshufd	$192,%xmm1,%xmm2
 	movups	%xmm7,80(%edi)
 	leal	96(%edi),%edi
-	pshufd	$128,%xmm0,%xmm3
+	movl	%ebx,%ecx
+	pshufd	$128,%xmm1,%xmm3
 	subl	$6,%eax
-	jnc	L039ctr32_loop6
+	jnc	L035ctr32_loop6
 	addl	$6,%eax
-	jz	L040ctr32_ret
-	movdqu	(%ebp),%xmm7
+	jz	L036ctr32_ret
 	movl	%ebp,%edx
-	pxor	32(%esp),%xmm7
-	movl	240(%ebp),%ecx
-L038ctr32_tail:
+	leal	1(,%ecx,2),%ecx
+	movdqa	32(%esp),%xmm7
+L034ctr32_tail:
 	por	%xmm7,%xmm2
 	cmpl	$2,%eax
-	jb	L041ctr32_one
-	pshufd	$64,%xmm0,%xmm4
+	jb	L037ctr32_one
+	pshufd	$64,%xmm1,%xmm4
 	por	%xmm7,%xmm3
-	je	L042ctr32_two
-	pshufd	$192,%xmm1,%xmm5
+	je	L038ctr32_two
+	pshufd	$192,%xmm0,%xmm5
 	por	%xmm7,%xmm4
 	cmpl	$4,%eax
-	jb	L043ctr32_three
-	pshufd	$128,%xmm1,%xmm6
+	jb	L039ctr32_three
+	pshufd	$128,%xmm0,%xmm6
 	por	%xmm7,%xmm5
-	je	L044ctr32_four
+	je	L040ctr32_four
 	por	%xmm7,%xmm6
 	call	__aesni_encrypt6
 	movups	(%esi),%xmm1
@@ -974,39 +897,39 @@ L038ctr32_tail:
 	movups	%xmm4,32(%edi)
 	movups	%xmm5,48(%edi)
 	movups	%xmm6,64(%edi)
-	jmp	L040ctr32_ret
+	jmp	L036ctr32_ret
 .align	4,0x90
-L037ctr32_one_shortcut:
+L033ctr32_one_shortcut:
 	movups	(%ebx),%xmm2
 	movl	240(%edx),%ecx
-L041ctr32_one:
+L037ctr32_one:
 	movups	(%edx),%xmm0
 	movups	16(%edx),%xmm1
 	leal	32(%edx),%edx
 	xorps	%xmm0,%xmm2
-L045enc1_loop_7:
+L041enc1_loop_7:
 .byte	102,15,56,220,209
 	decl	%ecx
 	movups	(%edx),%xmm1
 	leal	16(%edx),%edx
-	jnz	L045enc1_loop_7
+	jnz	L041enc1_loop_7
 .byte	102,15,56,221,209
 	movups	(%esi),%xmm6
 	xorps	%xmm2,%xmm6
 	movups	%xmm6,(%edi)
-	jmp	L040ctr32_ret
+	jmp	L036ctr32_ret
 .align	4,0x90
-L042ctr32_two:
-	call	__aesni_encrypt2
+L038ctr32_two:
+	call	__aesni_encrypt3
 	movups	(%esi),%xmm5
 	movups	16(%esi),%xmm6
 	xorps	%xmm5,%xmm2
 	xorps	%xmm6,%xmm3
 	movups	%xmm2,(%edi)
 	movups	%xmm3,16(%edi)
-	jmp	L040ctr32_ret
+	jmp	L036ctr32_ret
 .align	4,0x90
-L043ctr32_three:
+L039ctr32_three:
 	call	__aesni_encrypt3
 	movups	(%esi),%xmm5
 	movups	16(%esi),%xmm6
@@ -1017,9 +940,9 @@ L043ctr32_three:
 	xorps	%xmm7,%xmm4
 	movups	%xmm3,16(%edi)
 	movups	%xmm4,32(%edi)
-	jmp	L040ctr32_ret
+	jmp	L036ctr32_ret
 .align	4,0x90
-L044ctr32_four:
+L040ctr32_four:
 	call	__aesni_encrypt4
 	movups	(%esi),%xmm6
 	movups	16(%esi),%xmm7
@@ -1033,18 +956,7 @@ L044ctr32_four:
 	xorps	%xmm0,%xmm5
 	movups	%xmm4,32(%edi)
 	movups	%xmm5,48(%edi)
-L040ctr32_ret:
-	pxor	%xmm0,%xmm0
-	pxor	%xmm1,%xmm1
-	pxor	%xmm2,%xmm2
-	pxor	%xmm3,%xmm3
-	pxor	%xmm4,%xmm4
-	movdqa	%xmm0,32(%esp)
-	pxor	%xmm5,%xmm5
-	movdqa	%xmm0,48(%esp)
-	pxor	%xmm6,%xmm6
-	movdqa	%xmm0,64(%esp)
-	pxor	%xmm7,%xmm7
+L036ctr32_ret:
 	movl	80(%esp),%esp
 	popl	%edi
 	popl	%esi
@@ -1067,12 +979,12 @@ L_aesni_xts_encrypt_begin:
 	movups	16(%edx),%xmm1
 	leal	32(%edx),%edx
 	xorps	%xmm0,%xmm2
-L046enc1_loop_8:
+L042enc1_loop_8:
 .byte	102,15,56,220,209
 	decl	%ecx
 	movups	(%edx),%xmm1
 	leal	16(%edx),%edx
-	jnz	L046enc1_loop_8
+	jnz	L042enc1_loop_8
 .byte	102,15,56,221,209
 	movl	20(%esp),%esi
 	movl	24(%esp),%edi
@@ -1096,14 +1008,12 @@ L046enc1_loop_8:
 	movl	%edx,%ebp
 	movl	%ecx,%ebx
 	subl	$96,%eax
-	jc	L047xts_enc_short
-	shll	$4,%ecx
-	movl	$16,%ebx
-	subl	%ecx,%ebx
-	leal	32(%edx,%ecx,1),%edx
-	jmp	L048xts_enc_loop6
+	jc	L043xts_enc_short
+	shrl	$1,%ecx
+	movl	%ecx,%ebx
+	jmp	L044xts_enc_loop6
 .align	4,0x90
-L048xts_enc_loop6:
+L044xts_enc_loop6:
 	pshufd	$19,%xmm0,%xmm2
 	pxor	%xmm0,%xmm0
 	movdqa	%xmm1,(%esp)
@@ -1139,7 +1049,6 @@ L048xts_enc_loop6:
 	pand	%xmm3,%xmm7
 	movups	(%esi),%xmm2
 	pxor	%xmm1,%xmm7
-	movl	%ebx,%ecx
 	movdqu	16(%esi),%xmm3
 	xorps	%xmm0,%xmm2
 	movdqu	32(%esi),%xmm4
@@ -1155,17 +1064,19 @@ L048xts_enc_loop6:
 	movdqa	%xmm7,80(%esp)
 	pxor	%xmm1,%xmm7
 	movups	16(%ebp),%xmm1
+	leal	32(%ebp),%edx
 	pxor	16(%esp),%xmm3
-	pxor	32(%esp),%xmm4
 .byte	102,15,56,220,209
-	pxor	48(%esp),%xmm5
-	pxor	64(%esp),%xmm6
+	pxor	32(%esp),%xmm4
 .byte	102,15,56,220,217
-	pxor	%xmm0,%xmm7
-	movups	32(%ebp),%xmm0
+	pxor	48(%esp),%xmm5
+	decl	%ecx
 .byte	102,15,56,220,225
+	pxor	64(%esp),%xmm6
 .byte	102,15,56,220,233
+	pxor	%xmm0,%xmm7
 .byte	102,15,56,220,241
+	movups	(%edx),%xmm0
 .byte	102,15,56,220,249
 	call	L_aesni_encrypt6_enter
 	movdqa	80(%esp),%xmm1
@@ -1190,25 +1101,26 @@ L048xts_enc_loop6:
 	paddq	%xmm1,%xmm1
 	pand	%xmm3,%xmm2
 	pcmpgtd	%xmm1,%xmm0
+	movl	%ebx,%ecx
 	pxor	%xmm2,%xmm1
 	subl	$96,%eax
-	jnc	L048xts_enc_loop6
-	movl	240(%ebp),%ecx
+	jnc	L044xts_enc_loop6
+	leal	1(,%ecx,2),%ecx
 	movl	%ebp,%edx
 	movl	%ecx,%ebx
-L047xts_enc_short:
+L043xts_enc_short:
 	addl	$96,%eax
-	jz	L049xts_enc_done6x
+	jz	L045xts_enc_done6x
 	movdqa	%xmm1,%xmm5
 	cmpl	$32,%eax
-	jb	L050xts_enc_one
+	jb	L046xts_enc_one
 	pshufd	$19,%xmm0,%xmm2
 	pxor	%xmm0,%xmm0
 	paddq	%xmm1,%xmm1
 	pand	%xmm3,%xmm2
 	pcmpgtd	%xmm1,%xmm0
 	pxor	%xmm2,%xmm1
-	je	L051xts_enc_two
+	je	L047xts_enc_two
 	pshufd	$19,%xmm0,%xmm2
 	pxor	%xmm0,%xmm0
 	movdqa	%xmm1,%xmm6
@@ -1217,7 +1129,7 @@ L047xts_enc_short:
 	pcmpgtd	%xmm1,%xmm0
 	pxor	%xmm2,%xmm1
 	cmpl	$64,%eax
-	jb	L052xts_enc_three
+	jb	L048xts_enc_three
 	pshufd	$19,%xmm0,%xmm2
 	pxor	%xmm0,%xmm0
 	movdqa	%xmm1,%xmm7
@@ -1227,7 +1139,7 @@ L047xts_enc_short:
 	pxor	%xmm2,%xmm1
 	movdqa	%xmm5,(%esp)
 	movdqa	%xmm6,16(%esp)
-	je	L053xts_enc_four
+	je	L049xts_enc_four
 	movdqa	%xmm7,32(%esp)
 	pshufd	$19,%xmm0,%xmm7
 	movdqa	%xmm1,48(%esp)
@@ -1259,9 +1171,9 @@ L047xts_enc_short:
 	movups	%xmm5,48(%edi)
 	movups	%xmm6,64(%edi)
 	leal	80(%edi),%edi
-	jmp	L054xts_enc_done
+	jmp	L050xts_enc_done
 .align	4,0x90
-L050xts_enc_one:
+L046xts_enc_one:
 	movups	(%esi),%xmm2
 	leal	16(%esi),%esi
 	xorps	%xmm5,%xmm2
@@ -1269,36 +1181,37 @@ L050xts_enc_one:
 	movups	16(%edx),%xmm1
 	leal	32(%edx),%edx
 	xorps	%xmm0,%xmm2
-L055enc1_loop_9:
+L051enc1_loop_9:
 .byte	102,15,56,220,209
 	decl	%ecx
 	movups	(%edx),%xmm1
 	leal	16(%edx),%edx
-	jnz	L055enc1_loop_9
+	jnz	L051enc1_loop_9
 .byte	102,15,56,221,209
 	xorps	%xmm5,%xmm2
 	movups	%xmm2,(%edi)
 	leal	16(%edi),%edi
 	movdqa	%xmm5,%xmm1
-	jmp	L054xts_enc_done
+	jmp	L050xts_enc_done
 .align	4,0x90
-L051xts_enc_two:
+L047xts_enc_two:
 	movaps	%xmm1,%xmm6
 	movups	(%esi),%xmm2
 	movups	16(%esi),%xmm3
 	leal	32(%esi),%esi
 	xorps	%xmm5,%xmm2
 	xorps	%xmm6,%xmm3
-	call	__aesni_encrypt2
+	xorps	%xmm4,%xmm4
+	call	__aesni_encrypt3
 	xorps	%xmm5,%xmm2
 	xorps	%xmm6,%xmm3
 	movups	%xmm2,(%edi)
 	movups	%xmm3,16(%edi)
 	leal	32(%edi),%edi
 	movdqa	%xmm6,%xmm1
-	jmp	L054xts_enc_done
+	jmp	L050xts_enc_done
 .align	4,0x90
-L052xts_enc_three:
+L048xts_enc_three:
 	movaps	%xmm1,%xmm7
 	movups	(%esi),%xmm2
 	movups	16(%esi),%xmm3
@@ -1316,9 +1229,9 @@ L052xts_enc_three:
 	movups	%xmm4,32(%edi)
 	leal	48(%edi),%edi
 	movdqa	%xmm7,%xmm1
-	jmp	L054xts_enc_done
+	jmp	L050xts_enc_done
 .align	4,0x90
-L053xts_enc_four:
+L049xts_enc_four:
 	movaps	%xmm1,%xmm6
 	movups	(%esi),%xmm2
 	movups	16(%esi),%xmm3
@@ -1340,28 +1253,28 @@ L053xts_enc_four:
 	movups	%xmm5,48(%edi)
 	leal	64(%edi),%edi
 	movdqa	%xmm6,%xmm1
-	jmp	L054xts_enc_done
+	jmp	L050xts_enc_done
 .align	4,0x90
-L049xts_enc_done6x:
+L045xts_enc_done6x:
 	movl	112(%esp),%eax
 	andl	$15,%eax
-	jz	L056xts_enc_ret
+	jz	L052xts_enc_ret
 	movdqa	%xmm1,%xmm5
 	movl	%eax,112(%esp)
-	jmp	L057xts_enc_steal
+	jmp	L053xts_enc_steal
 .align	4,0x90
-L054xts_enc_done:
+L050xts_enc_done:
 	movl	112(%esp),%eax
 	pxor	%xmm0,%xmm0
 	andl	$15,%eax
-	jz	L056xts_enc_ret
+	jz	L052xts_enc_ret
 	pcmpgtd	%xmm1,%xmm0
 	movl	%eax,112(%esp)
 	pshufd	$19,%xmm0,%xmm5
 	paddq	%xmm1,%xmm1
 	pand	96(%esp),%xmm5
 	pxor	%xmm1,%xmm5
-L057xts_enc_steal:
+L053xts_enc_steal:
 	movzbl	(%esi),%ecx
 	movzbl	-16(%edi),%edx
 	leal	1(%esi),%esi
@@ -1369,7 +1282,7 @@ L057xts_enc_steal:
 	movb	%dl,(%edi)
 	leal	1(%edi),%edi
 	subl	$1,%eax
-	jnz	L057xts_enc_steal
+	jnz	L053xts_enc_steal
 	subl	112(%esp),%edi
 	movl	%ebp,%edx
 	movl	%ebx,%ecx
@@ -1379,30 +1292,16 @@ L057xts_enc_steal:
 	movups	16(%edx),%xmm1
 	leal	32(%edx),%edx
 	xorps	%xmm0,%xmm2
-L058enc1_loop_10:
+L054enc1_loop_10:
 .byte	102,15,56,220,209
 	decl	%ecx
 	movups	(%edx),%xmm1
 	leal	16(%edx),%edx
-	jnz	L058enc1_loop_10
+	jnz	L054enc1_loop_10
 .byte	102,15,56,221,209
 	xorps	%xmm5,%xmm2
 	movups	%xmm2,-16(%edi)
-L056xts_enc_ret:
-	pxor	%xmm0,%xmm0
-	pxor	%xmm1,%xmm1
-	pxor	%xmm2,%xmm2
-	movdqa	%xmm0,(%esp)
-	pxor	%xmm3,%xmm3
-	movdqa	%xmm0,16(%esp)
-	pxor	%xmm4,%xmm4
-	movdqa	%xmm0,32(%esp)
-	pxor	%xmm5,%xmm5
-	movdqa	%xmm0,48(%esp)
-	pxor	%xmm6,%xmm6
-	movdqa	%xmm0,64(%esp)
-	pxor	%xmm7,%xmm7
-	movdqa	%xmm0,80(%esp)
+L052xts_enc_ret:
 	movl	116(%esp),%esp
 	popl	%edi
 	popl	%esi
@@ -1425,12 +1324,12 @@ L_aesni_xts_decrypt_begin:
 	movups	16(%edx),%xmm1
 	leal	32(%edx),%edx
 	xorps	%xmm0,%xmm2
-L059enc1_loop_11:
+L055enc1_loop_11:
 .byte	102,15,56,220,209
 	decl	%ecx
 	movups	(%edx),%xmm1
 	leal	16(%edx),%edx
-	jnz	L059enc1_loop_11
+	jnz	L055enc1_loop_11
 .byte	102,15,56,221,209
 	movl	20(%esp),%esi
 	movl	24(%esp),%edi
@@ -1459,14 +1358,12 @@ L059enc1_loop_11:
 	pcmpgtd	%xmm1,%xmm0
 	andl	$-16,%eax
 	subl	$96,%eax
-	jc	L060xts_dec_short
-	shll	$4,%ecx
-	movl	$16,%ebx
-	subl	%ecx,%ebx
-	leal	32(%edx,%ecx,1),%edx
-	jmp	L061xts_dec_loop6
+	jc	L056xts_dec_short
+	shrl	$1,%ecx
+	movl	%ecx,%ebx
+	jmp	L057xts_dec_loop6
 .align	4,0x90
-L061xts_dec_loop6:
+L057xts_dec_loop6:
 	pshufd	$19,%xmm0,%xmm2
 	pxor	%xmm0,%xmm0
 	movdqa	%xmm1,(%esp)
@@ -1502,7 +1399,6 @@ L061xts_dec_loop6:
 	pand	%xmm3,%xmm7
 	movups	(%esi),%xmm2
 	pxor	%xmm1,%xmm7
-	movl	%ebx,%ecx
 	movdqu	16(%esi),%xmm3
 	xorps	%xmm0,%xmm2
 	movdqu	32(%esi),%xmm4
@@ -1518,17 +1414,19 @@ L061xts_dec_loop6:
 	movdqa	%xmm7,80(%esp)
 	pxor	%xmm1,%xmm7
 	movups	16(%ebp),%xmm1
+	leal	32(%ebp),%edx
 	pxor	16(%esp),%xmm3
-	pxor	32(%esp),%xmm4
 .byte	102,15,56,222,209
-	pxor	48(%esp),%xmm5
-	pxor	64(%esp),%xmm6
+	pxor	32(%esp),%xmm4
 .byte	102,15,56,222,217
-	pxor	%xmm0,%xmm7
-	movups	32(%ebp),%xmm0
+	pxor	48(%esp),%xmm5
+	decl	%ecx
 .byte	102,15,56,222,225
+	pxor	64(%esp),%xmm6
 .byte	102,15,56,222,233
+	pxor	%xmm0,%xmm7
 .byte	102,15,56,222,241
+	movups	(%edx),%xmm0
 .byte	102,15,56,222,249
 	call	L_aesni_decrypt6_enter
 	movdqa	80(%esp),%xmm1
@@ -1553,25 +1451,26 @@ L061xts_dec_loop6:
 	paddq	%xmm1,%xmm1
 	pand	%xmm3,%xmm2
 	pcmpgtd	%xmm1,%xmm0
+	movl	%ebx,%ecx
 	pxor	%xmm2,%xmm1
 	subl	$96,%eax
-	jnc	L061xts_dec_loop6
-	movl	240(%ebp),%ecx
+	jnc	L057xts_dec_loop6
+	leal	1(,%ecx,2),%ecx
 	movl	%ebp,%edx
 	movl	%ecx,%ebx
-L060xts_dec_short:
+L056xts_dec_short:
 	addl	$96,%eax
-	jz	L062xts_dec_done6x
+	jz	L058xts_dec_done6x
 	movdqa	%xmm1,%xmm5
 	cmpl	$32,%eax
-	jb	L063xts_dec_one
+	jb	L059xts_dec_one
 	pshufd	$19,%xmm0,%xmm2
 	pxor	%xmm0,%xmm0
 	paddq	%xmm1,%xmm1
 	pand	%xmm3,%xmm2
 	pcmpgtd	%xmm1,%xmm0
 	pxor	%xmm2,%xmm1
-	je	L064xts_dec_two
+	je	L060xts_dec_two
 	pshufd	$19,%xmm0,%xmm2
 	pxor	%xmm0,%xmm0
 	movdqa	%xmm1,%xmm6
@@ -1580,7 +1479,7 @@ L060xts_dec_short:
 	pcmpgtd	%xmm1,%xmm0
 	pxor	%xmm2,%xmm1
 	cmpl	$64,%eax
-	jb	L065xts_dec_three
+	jb	L061xts_dec_three
 	pshufd	$19,%xmm0,%xmm2
 	pxor	%xmm0,%xmm0
 	movdqa	%xmm1,%xmm7
@@ -1590,7 +1489,7 @@ L060xts_dec_short:
 	pxor	%xmm2,%xmm1
 	movdqa	%xmm5,(%esp)
 	movdqa	%xmm6,16(%esp)
-	je	L066xts_dec_four
+	je	L062xts_dec_four
 	movdqa	%xmm7,32(%esp)
 	pshufd	$19,%xmm0,%xmm7
 	movdqa	%xmm1,48(%esp)
@@ -1622,9 +1521,9 @@ L060xts_dec_short:
 	movups	%xmm5,48(%edi)
 	movups	%xmm6,64(%edi)
 	leal	80(%edi),%edi
-	jmp	L067xts_dec_done
+	jmp	L063xts_dec_done
 .align	4,0x90
-L063xts_dec_one:
+L059xts_dec_one:
 	movups	(%esi),%xmm2
 	leal	16(%esi),%esi
 	xorps	%xmm5,%xmm2
@@ -1632,36 +1531,36 @@ L063xts_dec_one:
 	movups	16(%edx),%xmm1
 	leal	32(%edx),%edx
 	xorps	%xmm0,%xmm2
-L068dec1_loop_12:
+L064dec1_loop_12:
 .byte	102,15,56,222,209
 	decl	%ecx
 	movups	(%edx),%xmm1
 	leal	16(%edx),%edx
-	jnz	L068dec1_loop_12
+	jnz	L064dec1_loop_12
 .byte	102,15,56,223,209
 	xorps	%xmm5,%xmm2
 	movups	%xmm2,(%edi)
 	leal	16(%edi),%edi
 	movdqa	%xmm5,%xmm1
-	jmp	L067xts_dec_done
+	jmp	L063xts_dec_done
 .align	4,0x90
-L064xts_dec_two:
+L060xts_dec_two:
 	movaps	%xmm1,%xmm6
 	movups	(%esi),%xmm2
 	movups	16(%esi),%xmm3
 	leal	32(%esi),%esi
 	xorps	%xmm5,%xmm2
 	xorps	%xmm6,%xmm3
-	call	__aesni_decrypt2
+	call	__aesni_decrypt3
 	xorps	%xmm5,%xmm2
 	xorps	%xmm6,%xmm3
 	movups	%xmm2,(%edi)
 	movups	%xmm3,16(%edi)
 	leal	32(%edi),%edi
 	movdqa	%xmm6,%xmm1
-	jmp	L067xts_dec_done
+	jmp	L063xts_dec_done
 .align	4,0x90
-L065xts_dec_three:
+L061xts_dec_three:
 	movaps	%xmm1,%xmm7
 	movups	(%esi),%xmm2
 	movups	16(%esi),%xmm3
@@ -1679,9 +1578,9 @@ L065xts_dec_three:
 	movups	%xmm4,32(%edi)
 	leal	48(%edi),%edi
 	movdqa	%xmm7,%xmm1
-	jmp	L067xts_dec_done
+	jmp	L063xts_dec_done
 .align	4,0x90
-L066xts_dec_four:
+L062xts_dec_four:
 	movaps	%xmm1,%xmm6
 	movups	(%esi),%xmm2
 	movups	16(%esi),%xmm3
@@ -1703,20 +1602,20 @@ L066xts_dec_four:
 	movups	%xmm5,48(%edi)
 	leal	64(%edi),%edi
 	movdqa	%xmm6,%xmm1
-	jmp	L067xts_dec_done
+	jmp	L063xts_dec_done
 .align	4,0x90
-L062xts_dec_done6x:
+L058xts_dec_done6x:
 	movl	112(%esp),%eax
 	andl	$15,%eax
-	jz	L069xts_dec_ret
+	jz	L065xts_dec_ret
 	movl	%eax,112(%esp)
-	jmp	L070xts_dec_only_one_more
+	jmp	L066xts_dec_only_one_more
 .align	4,0x90
-L067xts_dec_done:
+L063xts_dec_done:
 	movl	112(%esp),%eax
 	pxor	%xmm0,%xmm0
 	andl	$15,%eax
-	jz	L069xts_dec_ret
+	jz	L065xts_dec_ret
 	pcmpgtd	%xmm1,%xmm0
 	movl	%eax,112(%esp)
 	pshufd	$19,%xmm0,%xmm2
@@ -1726,7 +1625,7 @@ L067xts_dec_done:
 	pand	%xmm3,%xmm2
 	pcmpgtd	%xmm1,%xmm0
 	pxor	%xmm2,%xmm1
-L070xts_dec_only_one_more:
+L066xts_dec_only_one_more:
 	pshufd	$19,%xmm0,%xmm5
 	movdqa	%xmm1,%xmm6
 	paddq	%xmm1,%xmm1
@@ -1740,16 +1639,16 @@ L070xts_dec_only_one_more:
 	movups	16(%edx),%xmm1
 	leal	32(%edx),%edx
 	xorps	%xmm0,%xmm2
-L071dec1_loop_13:
+L067dec1_loop_13:
 .byte	102,15,56,222,209
 	decl	%ecx
 	movups	(%edx),%xmm1
 	leal	16(%edx),%edx
-	jnz	L071dec1_loop_13
+	jnz	L067dec1_loop_13
 .byte	102,15,56,223,209
 	xorps	%xmm5,%xmm2
 	movups	%xmm2,(%edi)
-L072xts_dec_steal:
+L068xts_dec_steal:
 	movzbl	16(%esi),%ecx
 	movzbl	(%edi),%edx
 	leal	1(%esi),%esi
@@ -1757,7 +1656,7 @@ L072xts_dec_steal:
 	movb	%dl,16(%edi)
 	leal	1(%edi),%edi
 	subl	$1,%eax
-	jnz	L072xts_dec_steal
+	jnz	L068xts_dec_steal
 	subl	112(%esp),%edi
 	movl	%ebp,%edx
 	movl	%ebx,%ecx
@@ -1767,30 +1666,16 @@ L072xts_dec_steal:
 	movups	16(%edx),%xmm1
 	leal	32(%edx),%edx
 	xorps	%xmm0,%xmm2
-L073dec1_loop_14:
+L069dec1_loop_14:
 .byte	102,15,56,222,209
 	decl	%ecx
 	movups	(%edx),%xmm1
 	leal	16(%edx),%edx
-	jnz	L073dec1_loop_14
+	jnz	L069dec1_loop_14
 .byte	102,15,56,223,209
 	xorps	%xmm6,%xmm2
 	movups	%xmm2,(%edi)
-L069xts_dec_ret:
-	pxor	%xmm0,%xmm0
-	pxor	%xmm1,%xmm1
-	pxor	%xmm2,%xmm2
-	movdqa	%xmm0,(%esp)
-	pxor	%xmm3,%xmm3
-	movdqa	%xmm0,16(%esp)
-	pxor	%xmm4,%xmm4
-	movdqa	%xmm0,32(%esp)
-	pxor	%xmm5,%xmm5
-	movdqa	%xmm0,48(%esp)
-	pxor	%xmm6,%xmm6
-	movdqa	%xmm0,64(%esp)
-	pxor	%xmm7,%xmm7
-	movdqa	%xmm0,80(%esp)
+L065xts_dec_ret:
 	movl	116(%esp),%esp
 	popl	%edi
 	popl	%esi
@@ -1814,7 +1699,7 @@ L_aesni_cbc_encrypt_begin:
 	movl	32(%esp),%edx
 	movl	36(%esp),%ebp
 	testl	%eax,%eax
-	jz	L074cbc_abort
+	jz	L070cbc_abort
 	cmpl	$0,40(%esp)
 	xchgl	%esp,%ebx
 	movups	(%ebp),%xmm7
@@ -1822,14 +1707,14 @@ L_aesni_cbc_encrypt_begin:
 	movl	%edx,%ebp
 	movl	%ebx,16(%esp)
 	movl	%ecx,%ebx
-	je	L075cbc_decrypt
+	je	L071cbc_decrypt
 	movaps	%xmm7,%xmm2
 	cmpl	$16,%eax
-	jb	L076cbc_enc_tail
+	jb	L072cbc_enc_tail
 	subl	$16,%eax
-	jmp	L077cbc_enc_loop
+	jmp	L073cbc_enc_loop
 .align	4,0x90
-L077cbc_enc_loop:
+L073cbc_enc_loop:
 	movups	(%esi),%xmm7
 	leal	16(%esi),%esi
 	movups	(%edx),%xmm0
@@ -1837,25 +1722,24 @@ L077cbc_enc_loop:
 	xorps	%xmm0,%xmm7
 	leal	32(%edx),%edx
 	xorps	%xmm7,%xmm2
-L078enc1_loop_15:
+L074enc1_loop_15:
 .byte	102,15,56,220,209
 	decl	%ecx
 	movups	(%edx),%xmm1
 	leal	16(%edx),%edx
-	jnz	L078enc1_loop_15
+	jnz	L074enc1_loop_15
 .byte	102,15,56,221,209
 	movl	%ebx,%ecx
 	movl	%ebp,%edx
 	movups	%xmm2,(%edi)
 	leal	16(%edi),%edi
 	subl	$16,%eax
-	jnc	L077cbc_enc_loop
+	jnc	L073cbc_enc_loop
 	addl	$16,%eax
-	jnz	L076cbc_enc_tail
+	jnz	L072cbc_enc_tail
 	movaps	%xmm2,%xmm7
-	pxor	%xmm2,%xmm2
-	jmp	L079cbc_ret
-L076cbc_enc_tail:
+	jmp	L075cbc_ret
+L072cbc_enc_tail:
 	movl	%eax,%ecx
 .long	2767451785
 	movl	$16,%ecx
@@ -1866,20 +1750,20 @@ L076cbc_enc_tail:
 	movl	%ebx,%ecx
 	movl	%edi,%esi
 	movl	%ebp,%edx
-	jmp	L077cbc_enc_loop
+	jmp	L073cbc_enc_loop
 .align	4,0x90
-L075cbc_decrypt:
+L071cbc_decrypt:
 	cmpl	$80,%eax
-	jbe	L080cbc_dec_tail
+	jbe	L076cbc_dec_tail
 	movaps	%xmm7,(%esp)
 	subl	$80,%eax
-	jmp	L081cbc_dec_loop6_enter
+	jmp	L077cbc_dec_loop6_enter
 .align	4,0x90
-L082cbc_dec_loop6:
+L078cbc_dec_loop6:
 	movaps	%xmm0,(%esp)
 	movups	%xmm7,(%edi)
 	leal	16(%edi),%edi
-L081cbc_dec_loop6_enter:
+L077cbc_dec_loop6_enter:
 	movdqu	(%esi),%xmm2
 	movdqu	16(%esi),%xmm3
 	movdqu	32(%esi),%xmm4
@@ -1909,28 +1793,28 @@ L081cbc_dec_loop6_enter:
 	movups	%xmm6,64(%edi)
 	leal	80(%edi),%edi
 	subl	$96,%eax
-	ja	L082cbc_dec_loop6
+	ja	L078cbc_dec_loop6
 	movaps	%xmm7,%xmm2
 	movaps	%xmm0,%xmm7
 	addl	$80,%eax
-	jle	L083cbc_dec_clear_tail_collected
+	jle	L079cbc_dec_tail_collected
 	movups	%xmm2,(%edi)
 	leal	16(%edi),%edi
-L080cbc_dec_tail:
+L076cbc_dec_tail:
 	movups	(%esi),%xmm2
 	movaps	%xmm2,%xmm6
 	cmpl	$16,%eax
-	jbe	L084cbc_dec_one
+	jbe	L080cbc_dec_one
 	movups	16(%esi),%xmm3
 	movaps	%xmm3,%xmm5
 	cmpl	$32,%eax
-	jbe	L085cbc_dec_two
+	jbe	L081cbc_dec_two
 	movups	32(%esi),%xmm4
 	cmpl	$48,%eax
-	jbe	L086cbc_dec_three
+	jbe	L082cbc_dec_three
 	movups	48(%esi),%xmm5
 	cmpl	$64,%eax
-	jbe	L087cbc_dec_four
+	jbe	L083cbc_dec_four
 	movups	64(%esi),%xmm6
 	movaps	%xmm7,(%esp)
 	movups	(%esi),%xmm2
@@ -1948,62 +1832,56 @@ L080cbc_dec_tail:
 	xorps	%xmm0,%xmm6
 	movups	%xmm2,(%edi)
 	movups	%xmm3,16(%edi)
-	pxor	%xmm3,%xmm3
 	movups	%xmm4,32(%edi)
-	pxor	%xmm4,%xmm4
 	movups	%xmm5,48(%edi)
-	pxor	%xmm5,%xmm5
 	leal	64(%edi),%edi
 	movaps	%xmm6,%xmm2
-	pxor	%xmm6,%xmm6
 	subl	$80,%eax
-	jmp	L088cbc_dec_tail_collected
+	jmp	L079cbc_dec_tail_collected
 .align	4,0x90
-L084cbc_dec_one:
+L080cbc_dec_one:
 	movups	(%edx),%xmm0
 	movups	16(%edx),%xmm1
 	leal	32(%edx),%edx
 	xorps	%xmm0,%xmm2
-L089dec1_loop_16:
+L084dec1_loop_16:
 .byte	102,15,56,222,209
 	decl	%ecx
 	movups	(%edx),%xmm1
 	leal	16(%edx),%edx
-	jnz	L089dec1_loop_16
+	jnz	L084dec1_loop_16
 .byte	102,15,56,223,209
 	xorps	%xmm7,%xmm2
 	movaps	%xmm6,%xmm7
 	subl	$16,%eax
-	jmp	L088cbc_dec_tail_collected
+	jmp	L079cbc_dec_tail_collected
 .align	4,0x90
-L085cbc_dec_two:
-	call	__aesni_decrypt2
+L081cbc_dec_two:
+	xorps	%xmm4,%xmm4
+	call	__aesni_decrypt3
 	xorps	%xmm7,%xmm2
 	xorps	%xmm6,%xmm3
 	movups	%xmm2,(%edi)
 	movaps	%xmm3,%xmm2
-	pxor	%xmm3,%xmm3
 	leal	16(%edi),%edi
 	movaps	%xmm5,%xmm7
 	subl	$32,%eax
-	jmp	L088cbc_dec_tail_collected
+	jmp	L079cbc_dec_tail_collected
 .align	4,0x90
-L086cbc_dec_three:
+L082cbc_dec_three:
 	call	__aesni_decrypt3
 	xorps	%xmm7,%xmm2
 	xorps	%xmm6,%xmm3
 	xorps	%xmm5,%xmm4
 	movups	%xmm2,(%edi)
 	movaps	%xmm4,%xmm2
-	pxor	%xmm4,%xmm4
 	movups	%xmm3,16(%edi)
-	pxor	%xmm3,%xmm3
 	leal	32(%edi),%edi
 	movups	32(%esi),%xmm7
 	subl	$48,%eax
-	jmp	L088cbc_dec_tail_collected
+	jmp	L079cbc_dec_tail_collected
 .align	4,0x90
-L087cbc_dec_four:
+L083cbc_dec_four:
 	call	__aesni_decrypt4
 	movups	16(%esi),%xmm1
 	movups	32(%esi),%xmm0
@@ -2013,44 +1891,28 @@ L087cbc_dec_four:
 	movups	%xmm2,(%edi)
 	xorps	%xmm1,%xmm4
 	movups	%xmm3,16(%edi)
-	pxor	%xmm3,%xmm3
 	xorps	%xmm0,%xmm5
 	movups	%xmm4,32(%edi)
-	pxor	%xmm4,%xmm4
 	leal	48(%edi),%edi
 	movaps	%xmm5,%xmm2
-	pxor	%xmm5,%xmm5
 	subl	$64,%eax
-	jmp	L088cbc_dec_tail_collected
-.align	4,0x90
-L083cbc_dec_clear_tail_collected:
-	pxor	%xmm3,%xmm3
-	pxor	%xmm4,%xmm4
-	pxor	%xmm5,%xmm5
-	pxor	%xmm6,%xmm6
-L088cbc_dec_tail_collected:
+L079cbc_dec_tail_collected:
 	andl	$15,%eax
-	jnz	L090cbc_dec_tail_partial
+	jnz	L085cbc_dec_tail_partial
 	movups	%xmm2,(%edi)
-	pxor	%xmm0,%xmm0
-	jmp	L079cbc_ret
+	jmp	L075cbc_ret
 .align	4,0x90
-L090cbc_dec_tail_partial:
+L085cbc_dec_tail_partial:
 	movaps	%xmm2,(%esp)
-	pxor	%xmm0,%xmm0
 	movl	$16,%ecx
 	movl	%esp,%esi
 	subl	%eax,%ecx
 .long	2767451785
-	movdqa	%xmm2,(%esp)
-L079cbc_ret:
+L075cbc_ret:
 	movl	16(%esp),%esp
 	movl	36(%esp),%ebp
-	pxor	%xmm2,%xmm2
-	pxor	%xmm1,%xmm1
 	movups	%xmm7,(%ebp)
-	pxor	%xmm7,%xmm7
-L074cbc_abort:
+L070cbc_abort:
 	popl	%edi
 	popl	%esi
 	popl	%ebx
@@ -2058,62 +1920,52 @@ L074cbc_abort:
 	ret
 .align	4
 __aesni_set_encrypt_key:
-	pushl	%ebp
-	pushl	%ebx
 	testl	%eax,%eax
-	jz	L091bad_pointer
+	jz	L086bad_pointer
 	testl	%edx,%edx
-	jz	L091bad_pointer
-	call	L092pic
-L092pic:
-	popl	%ebx
-	leal	Lkey_const-L092pic(%ebx),%ebx
-	movl	L__gnutls_x86_cpuid_s$non_lazy_ptr-Lkey_const(%ebx),%ebp
+	jz	L086bad_pointer
 	movups	(%eax),%xmm0
 	xorps	%xmm4,%xmm4
-	movl	4(%ebp),%ebp
 	leal	16(%edx),%edx
-	andl	$268437504,%ebp
 	cmpl	$256,%ecx
-	je	L09314rounds
+	je	L08714rounds
 	cmpl	$192,%ecx
-	je	L09412rounds
+	je	L08812rounds
 	cmpl	$128,%ecx
-	jne	L095bad_keybits
+	jne	L089bad_keybits
 .align	4,0x90
-L09610rounds:
-	cmpl	$268435456,%ebp
-	je	L09710rounds_alt
+L09010rounds:
 	movl	$9,%ecx
 	movups	%xmm0,-16(%edx)
 .byte	102,15,58,223,200,1
-	call	L098key_128_cold
+	call	L091key_128_cold
 .byte	102,15,58,223,200,2
-	call	L099key_128
+	call	L092key_128
 .byte	102,15,58,223,200,4
-	call	L099key_128
+	call	L092key_128
 .byte	102,15,58,223,200,8
-	call	L099key_128
+	call	L092key_128
 .byte	102,15,58,223,200,16
-	call	L099key_128
+	call	L092key_128
 .byte	102,15,58,223,200,32
-	call	L099key_128
+	call	L092key_128
 .byte	102,15,58,223,200,64
-	call	L099key_128
+	call	L092key_128
 .byte	102,15,58,223,200,128
-	call	L099key_128
+	call	L092key_128
 .byte	102,15,58,223,200,27
-	call	L099key_128
+	call	L092key_128
 .byte	102,15,58,223,200,54
-	call	L099key_128
+	call	L092key_128
 	movups	%xmm0,(%edx)
 	movl	%ecx,80(%edx)
-	jmp	L100good_key
+	xorl	%eax,%eax
+	ret
 .align	4,0x90
-L099key_128:
+L092key_128:
 	movups	%xmm0,(%edx)
 	leal	16(%edx),%edx
-L098key_128_cold:
+L091key_128_cold:
 	shufps	$16,%xmm0,%xmm4
 	xorps	%xmm4,%xmm0
 	shufps	$140,%xmm0,%xmm4
@@ -2122,91 +1974,38 @@ L098key_128_cold:
 	xorps	%xmm1,%xmm0
 	ret
 .align	4,0x90
-L09710rounds_alt:
-	movdqa	(%ebx),%xmm5
-	movl	$8,%ecx
-	movdqa	32(%ebx),%xmm4
-	movdqa	%xmm0,%xmm2
-	movdqu	%xmm0,-16(%edx)
-L101loop_key128:
-.byte	102,15,56,0,197
-.byte	102,15,56,221,196
-	pslld	$1,%xmm4
-	leal	16(%edx),%edx
-	movdqa	%xmm2,%xmm3
-	pslldq	$4,%xmm2
-	pxor	%xmm2,%xmm3
-	pslldq	$4,%xmm2
-	pxor	%xmm2,%xmm3
-	pslldq	$4,%xmm2
-	pxor	%xmm3,%xmm2
-	pxor	%xmm2,%xmm0
-	movdqu	%xmm0,-16(%edx)
-	movdqa	%xmm0,%xmm2
-	decl	%ecx
-	jnz	L101loop_key128
-	movdqa	48(%ebx),%xmm4
-.byte	102,15,56,0,197
-.byte	102,15,56,221,196
-	pslld	$1,%xmm4
-	movdqa	%xmm2,%xmm3
-	pslldq	$4,%xmm2
-	pxor	%xmm2,%xmm3
-	pslldq	$4,%xmm2
-	pxor	%xmm2,%xmm3
-	pslldq	$4,%xmm2
-	pxor	%xmm3,%xmm2
-	pxor	%xmm2,%xmm0
-	movdqu	%xmm0,(%edx)
-	movdqa	%xmm0,%xmm2
-.byte	102,15,56,0,197
-.byte	102,15,56,221,196
-	movdqa	%xmm2,%xmm3
-	pslldq	$4,%xmm2
-	pxor	%xmm2,%xmm3
-	pslldq	$4,%xmm2
-	pxor	%xmm2,%xmm3
-	pslldq	$4,%xmm2
-	pxor	%xmm3,%xmm2
-	pxor	%xmm2,%xmm0
-	movdqu	%xmm0,16(%edx)
-	movl	$9,%ecx
-	movl	%ecx,96(%edx)
-	jmp	L100good_key
-.align	4,0x90
-L09412rounds:
+L08812rounds:
 	movq	16(%eax),%xmm2
-	cmpl	$268435456,%ebp
-	je	L10212rounds_alt
 	movl	$11,%ecx
 	movups	%xmm0,-16(%edx)
 .byte	102,15,58,223,202,1
-	call	L103key_192a_cold
+	call	L093key_192a_cold
 .byte	102,15,58,223,202,2
-	call	L104key_192b
+	call	L094key_192b
 .byte	102,15,58,223,202,4
-	call	L105key_192a
+	call	L095key_192a
 .byte	102,15,58,223,202,8
-	call	L104key_192b
+	call	L094key_192b
 .byte	102,15,58,223,202,16
-	call	L105key_192a
+	call	L095key_192a
 .byte	102,15,58,223,202,32
-	call	L104key_192b
+	call	L094key_192b
 .byte	102,15,58,223,202,64
-	call	L105key_192a
+	call	L095key_192a
 .byte	102,15,58,223,202,128
-	call	L104key_192b
+	call	L094key_192b
 	movups	%xmm0,(%edx)
 	movl	%ecx,48(%edx)
-	jmp	L100good_key
+	xorl	%eax,%eax
+	ret
 .align	4,0x90
-L105key_192a:
+L095key_192a:
 	movups	%xmm0,(%edx)
 	leal	16(%edx),%edx
 .align	4,0x90
-L103key_192a_cold:
+L093key_192a_cold:
 	movaps	%xmm2,%xmm5
-L106key_192b_warm:
+L096key_192b_warm:
 	shufps	$16,%xmm0,%xmm4
 	movdqa	%xmm2,%xmm3
 	xorps	%xmm4,%xmm0
@@ -2220,90 +2019,56 @@ L106key_192b_warm:
 	pxor	%xmm3,%xmm2
 	ret
 .align	4,0x90
-L104key_192b:
+L094key_192b:
 	movaps	%xmm0,%xmm3
 	shufps	$68,%xmm0,%xmm5
 	movups	%xmm5,(%edx)
 	shufps	$78,%xmm2,%xmm3
 	movups	%xmm3,16(%edx)
 	leal	32(%edx),%edx
-	jmp	L106key_192b_warm
+	jmp	L096key_192b_warm
 .align	4,0x90
-L10212rounds_alt:
-	movdqa	16(%ebx),%xmm5
-	movdqa	32(%ebx),%xmm4
-	movl	$8,%ecx
-	movdqu	%xmm0,-16(%edx)
-L107loop_key192:
-	movq	%xmm2,(%edx)
-	movdqa	%xmm2,%xmm1
-.byte	102,15,56,0,213
-.byte	102,15,56,221,212
-	pslld	$1,%xmm4
-	leal	24(%edx),%edx
-	movdqa	%xmm0,%xmm3
-	pslldq	$4,%xmm0
-	pxor	%xmm0,%xmm3
-	pslldq	$4,%xmm0
-	pxor	%xmm0,%xmm3
-	pslldq	$4,%xmm0
-	pxor	%xmm3,%xmm0
-	pshufd	$255,%xmm0,%xmm3
-	pxor	%xmm1,%xmm3
-	pslldq	$4,%xmm1
-	pxor	%xmm1,%xmm3
-	pxor	%xmm2,%xmm0
-	pxor	%xmm3,%xmm2
-	movdqu	%xmm0,-16(%edx)
-	decl	%ecx
-	jnz	L107loop_key192
-	movl	$11,%ecx
-	movl	%ecx,32(%edx)
-	jmp	L100good_key
-.align	4,0x90
-L09314rounds:
+L08714rounds:
 	movups	16(%eax),%xmm2
-	leal	16(%edx),%edx
-	cmpl	$268435456,%ebp
-	je	L10814rounds_alt
 	movl	$13,%ecx
+	leal	16(%edx),%edx
 	movups	%xmm0,-32(%edx)
 	movups	%xmm2,-16(%edx)
 .byte	102,15,58,223,202,1
-	call	L109key_256a_cold
+	call	L097key_256a_cold
 .byte	102,15,58,223,200,1
-	call	L110key_256b
+	call	L098key_256b
 .byte	102,15,58,223,202,2
-	call	L111key_256a
+	call	L099key_256a
 .byte	102,15,58,223,200,2
-	call	L110key_256b
+	call	L098key_256b
 .byte	102,15,58,223,202,4
-	call	L111key_256a
+	call	L099key_256a
 .byte	102,15,58,223,200,4
-	call	L110key_256b
+	call	L098key_256b
 .byte	102,15,58,223,202,8
-	call	L111key_256a
+	call	L099key_256a
 .byte	102,15,58,223,200,8
-	call	L110key_256b
+	call	L098key_256b
 .byte	102,15,58,223,202,16
-	call	L111key_256a
+	call	L099key_256a
 .byte	102,15,58,223,200,16
-	call	L110key_256b
+	call	L098key_256b
 .byte	102,15,58,223,202,32
-	call	L111key_256a
+	call	L099key_256a
 .byte	102,15,58,223,200,32
-	call	L110key_256b
+	call	L098key_256b
 .byte	102,15,58,223,202,64
-	call	L111key_256a
+	call	L099key_256a
 	movups	%xmm0,(%edx)
 	movl	%ecx,16(%edx)
 	xorl	%eax,%eax
-	jmp	L100good_key
+	ret
 .align	4,0x90
-L111key_256a:
+L099key_256a:
 	movups	%xmm2,(%edx)
 	leal	16(%edx),%edx
-L109key_256a_cold:
+L097key_256a_cold:
 	shufps	$16,%xmm0,%xmm4
 	xorps	%xmm4,%xmm0
 	shufps	$140,%xmm0,%xmm4
@@ -2312,7 +2077,7 @@ L109key_256a_cold:
 	xorps	%xmm1,%xmm0
 	ret
 .align	4,0x90
-L110key_256b:
+L098key_256b:
 	movups	%xmm0,(%edx)
 	leal	16(%edx),%edx
 	shufps	$16,%xmm2,%xmm4
@@ -2322,70 +2087,13 @@ L110key_256b:
 	shufps	$170,%xmm1,%xmm1
 	xorps	%xmm1,%xmm2
 	ret
-.align	4,0x90
-L10814rounds_alt:
-	movdqa	(%ebx),%xmm5
-	movdqa	32(%ebx),%xmm4
-	movl	$7,%ecx
-	movdqu	%xmm0,-32(%edx)
-	movdqa	%xmm2,%xmm1
-	movdqu	%xmm2,-16(%edx)
-L112loop_key256:
-.byte	102,15,56,0,213
-.byte	102,15,56,221,212
-	movdqa	%xmm0,%xmm3
-	pslldq	$4,%xmm0
-	pxor	%xmm0,%xmm3
-	pslldq	$4,%xmm0
-	pxor	%xmm0,%xmm3
-	pslldq	$4,%xmm0
-	pxor	%xmm3,%xmm0
-	pslld	$1,%xmm4
-	pxor	%xmm2,%xmm0
-	movdqu	%xmm0,(%edx)
-	decl	%ecx
-	jz	L113done_key256
-	pshufd	$255,%xmm0,%xmm2
-	pxor	%xmm3,%xmm3
-.byte	102,15,56,221,211
-	movdqa	%xmm1,%xmm3
-	pslldq	$4,%xmm1
-	pxor	%xmm1,%xmm3
-	pslldq	$4,%xmm1
-	pxor	%xmm1,%xmm3
-	pslldq	$4,%xmm1
-	pxor	%xmm3,%xmm1
-	pxor	%xmm1,%xmm2
-	movdqu	%xmm2,16(%edx)
-	leal	32(%edx),%edx
-	movdqa	%xmm2,%xmm1
-	jmp	L112loop_key256
-L113done_key256:
-	movl	$13,%ecx
-	movl	%ecx,16(%edx)
-L100good_key:
-	pxor	%xmm0,%xmm0
-	pxor	%xmm1,%xmm1
-	pxor	%xmm2,%xmm2
-	pxor	%xmm3,%xmm3
-	pxor	%xmm4,%xmm4
-	pxor	%xmm5,%xmm5
-	xorl	%eax,%eax
-	popl	%ebx
-	popl	%ebp
-	ret
 .align	2,0x90
-L091bad_pointer:
+L086bad_pointer:
 	movl	$-1,%eax
-	popl	%ebx
-	popl	%ebp
 	ret
 .align	2,0x90
-L095bad_keybits:
-	pxor	%xmm0,%xmm0
+L089bad_keybits:
 	movl	$-2,%eax
-	popl	%ebx
-	popl	%ebp
 	ret
 .globl	_aesni_set_encrypt_key
 .align	4
@@ -2407,7 +2115,7 @@ L_aesni_set_decrypt_key_begin:
 	movl	12(%esp),%edx
 	shll	$4,%ecx
 	testl	%eax,%eax
-	jnz	L114dec_key_ret
+	jnz	L100dec_key_ret
 	leal	16(%edx,%ecx,1),%eax
 	movups	(%edx),%xmm0
 	movups	(%eax),%xmm1
@@ -2415,7 +2123,7 @@ L_aesni_set_decrypt_key_begin:
 	movups	%xmm1,(%edx)
 	leal	16(%edx),%edx
 	leal	-16(%eax),%eax
-L115dec_key_inverse:
+L101dec_key_inverse:
 	movups	(%edx),%xmm0
 	movups	(%eax),%xmm1
 .byte	102,15,56,219,192
@@ -2425,28 +2133,15 @@ L115dec_key_inverse:
 	movups	%xmm0,16(%eax)
 	movups	%xmm1,-16(%edx)
 	cmpl	%edx,%eax
-	ja	L115dec_key_inverse
+	ja	L101dec_key_inverse
 	movups	(%edx),%xmm0
 .byte	102,15,56,219,192
 	movups	%xmm0,(%edx)
-	pxor	%xmm0,%xmm0
-	pxor	%xmm1,%xmm1
 	xorl	%eax,%eax
-L114dec_key_ret:
+L100dec_key_ret:
 	ret
-.align	6,0x90
-Lkey_const:
-.long	202313229,202313229,202313229,202313229
-.long	67569157,67569157,67569157,67569157
-.long	1,1,1,1
-.long	27,27,27,27
 .byte	65,69,83,32,102,111,114,32,73,110,116,101,108,32,65,69
 .byte	83,45,78,73,44,32,67,82,89,80,84,79,71,65,77,83
 .byte	32,98,121,32,60,97,112,112,114,111,64,111,112,101,110,115
 .byte	115,108,46,111,114,103,62,0
-.section __IMPORT,__pointers,non_lazy_symbol_pointers
-L__gnutls_x86_cpuid_s$non_lazy_ptr:
-.indirect_symbol	__gnutls_x86_cpuid_s
-.long	0
-.comm	__gnutls_x86_cpuid_s,16,2