Added Mike Hamburg's SSSE3 AES implementation.

14 files changed, 1813 insertions, 12 deletions

diff --git a/lib/accelerated/x86/coff/aes-ssse3-x86.s b/lib/accelerated/x86/coff/aes-ssse3-x86.s
new file mode 100644
index 0000000000..6894b14b7c
--- /dev/null
+++ b/lib/accelerated/x86/coff/aes-ssse3-x86.s
@@ -0,0 +1,662 @@
+######################################################################
+## Constant-time SSSE3 AES core implementation.
+## version 0.1
+##
+## By Mike Hamburg (Stanford University), 2009
+## Public domain.
+##
+## For details see http://shiftleft.org/papers/vector_aes/ and
+## http://crypto.stanford.edu/vpaes/.
+#
+# *** This file is auto-generated ***
+#
+.file	"vpaes-x86.s"
+.text
+.align	64
+.L_vpaes_consts:
+.long	218628480,235210255,168496130,67568393
+.long	252381056,17041926,33884169,51187212
+.long	252645135,252645135,252645135,252645135
+.long	1512730624,3266504856,1377990664,3401244816
+.long	830229760,1275146365,2969422977,3447763452
+.long	3411033600,2979783055,338359620,2782886510
+.long	4209124096,907596821,221174255,1006095553
+.long	191964160,3799684038,3164090317,1589111125
+.long	182528256,1777043520,2877432650,3265356744
+.long	1874708224,3503451415,3305285752,363511674
+.long	1606117888,3487855781,1093350906,2384367825
+.long	197121,67569157,134941193,202313229
+.long	67569157,134941193,202313229,197121
+.long	134941193,202313229,197121,67569157
+.long	202313229,197121,67569157,134941193
+.long	33619971,100992007,168364043,235736079
+.long	235736079,33619971,100992007,168364043
+.long	168364043,235736079,33619971,100992007
+.long	100992007,168364043,235736079,33619971
+.long	50462976,117835012,185207048,252579084
+.long	252314880,51251460,117574920,184942860
+.long	184682752,252054788,50987272,118359308
+.long	118099200,185467140,251790600,50727180
+.long	2946363062,528716217,1300004225,1881839624
+.long	1532713819,1532713819,1532713819,1532713819
+.long	3602276352,4288629033,3737020424,4153884961
+.long	1354558464,32357713,2958822624,3775749553
+.long	1201988352,132424512,1572796698,503232858
+.long	2213177600,1597421020,4103937655,675398315
+.long	2749646592,4273543773,1511898873,121693092
+.long	3040248576,1103263732,2871565598,1608280554
+.long	2236667136,2588920351,482954393,64377734
+.long	3069987328,291237287,2117370568,3650299247
+.long	533321216,3573750986,2572112006,1401264716
+.long	1339849704,2721158661,548607111,3445553514
+.long	2128193280,3054596040,2183486460,1257083700
+.long	655635200,1165381986,3923443150,2344132524
+.long	190078720,256924420,290342170,357187870
+.long	1610966272,2263057382,4103205268,309794674
+.long	2592527872,2233205587,1335446729,3402964816
+.long	3973531904,3225098121,3002836325,1918774430
+.long	3870401024,2102906079,2284471353,4117666579
+.long	617007872,1021508343,366931923,691083277
+.long	2528395776,3491914898,2968704004,1613121270
+.long	3445188352,3247741094,844474987,4093578302
+.long	651481088,1190302358,1689581232,574775300
+.long	4289380608,206939853,2555985458,2489840491
+.long	2130264064,327674451,3566485037,3349835193
+.long	2470714624,316102159,3636825756,3393945945
+.byte	86,101,99,116,111,114,32,80,101,114,109,117,116,97,116,105
+.byte	111,110,32,65,69,83,32,102,111,114,32,120,56,54,47,83
+.byte	83,83,69,51,44,32,77,105,107,101,32,72,97,109,98,117
+.byte	114,103,32,40,83,116,97,110,102,111,114,100,32,85,110,105
+.byte	118,101,114,115,105,116,121,41,0
+.align	64
+.def	__vpaes_preheat;	.scl	3;	.type	32;	.endef
+.align	16
+__vpaes_preheat:
+	addl	(%esp),%ebp
+	movdqa	-48(%ebp),%xmm7
+	movdqa	-16(%ebp),%xmm6
+	ret
+.def	__vpaes_encrypt_core;	.scl	3;	.type	32;	.endef
+.align	16
+__vpaes_encrypt_core:
+	movl	$16,%ecx
+	movl	240(%edx),%eax
+	movdqa	%xmm6,%xmm1
+	movdqa	(%ebp),%xmm2
+	pandn	%xmm0,%xmm1
+	pand	%xmm6,%xmm0
+	movdqu	(%edx),%xmm5
+.byte	102,15,56,0,208
+	movdqa	16(%ebp),%xmm0
+	pxor	%xmm5,%xmm2
+	psrld	$4,%xmm1
+	addl	$16,%edx
+.byte	102,15,56,0,193
+	leal	192(%ebp),%ebx
+	pxor	%xmm2,%xmm0
+	jmp	.L000enc_entry
+.align	16
+.L001enc_loop:
+	movdqa	32(%ebp),%xmm4
+	movdqa	48(%ebp),%xmm0
+.byte	102,15,56,0,226
+.byte	102,15,56,0,195
+	pxor	%xmm5,%xmm4
+	movdqa	64(%ebp),%xmm5
+	pxor	%xmm4,%xmm0
+	movdqa	-64(%ebx,%ecx,1),%xmm1
+.byte	102,15,56,0,234
+	movdqa	80(%ebp),%xmm2
+	movdqa	(%ebx,%ecx,1),%xmm4
+.byte	102,15,56,0,211
+	movdqa	%xmm0,%xmm3
+	pxor	%xmm5,%xmm2
+.byte	102,15,56,0,193
+	addl	$16,%edx
+	pxor	%xmm2,%xmm0
+.byte	102,15,56,0,220
+	addl	$16,%ecx
+	pxor	%xmm0,%xmm3
+.byte	102,15,56,0,193
+	andl	$48,%ecx
+	subl	$1,%eax
+	pxor	%xmm3,%xmm0
+.L000enc_entry:
+	movdqa	%xmm6,%xmm1
+	movdqa	-32(%ebp),%xmm5
+	pandn	%xmm0,%xmm1
+	psrld	$4,%xmm1
+	pand	%xmm6,%xmm0
+.byte	102,15,56,0,232
+	movdqa	%xmm7,%xmm3
+	pxor	%xmm1,%xmm0
+.byte	102,15,56,0,217
+	movdqa	%xmm7,%xmm4
+	pxor	%xmm5,%xmm3
+.byte	102,15,56,0,224
+	movdqa	%xmm7,%xmm2
+	pxor	%xmm5,%xmm4
+.byte	102,15,56,0,211
+	movdqa	%xmm7,%xmm3
+	pxor	%xmm0,%xmm2
+.byte	102,15,56,0,220
+	movdqu	(%edx),%xmm5
+	pxor	%xmm1,%xmm3
+	jnz	.L001enc_loop
+	movdqa	96(%ebp),%xmm4
+	movdqa	112(%ebp),%xmm0
+.byte	102,15,56,0,226
+	pxor	%xmm5,%xmm4
+.byte	102,15,56,0,195
+	movdqa	64(%ebx,%ecx,1),%xmm1
+	pxor	%xmm4,%xmm0
+.byte	102,15,56,0,193
+	ret
+.def	__vpaes_decrypt_core;	.scl	3;	.type	32;	.endef
+.align	16
+__vpaes_decrypt_core:
+	leal	608(%ebp),%ebx
+	movl	240(%edx),%eax
+	movdqa	%xmm6,%xmm1
+	movdqa	-64(%ebx),%xmm2
+	pandn	%xmm0,%xmm1
+	movl	%eax,%ecx
+	psrld	$4,%xmm1
+	movdqu	(%edx),%xmm5
+	shll	$4,%ecx
+	pand	%xmm6,%xmm0
+.byte	102,15,56,0,208
+	movdqa	-48(%ebx),%xmm0
+	xorl	$48,%ecx
+.byte	102,15,56,0,193
+	andl	$48,%ecx
+	pxor	%xmm5,%xmm2
+	movdqa	176(%ebp),%xmm5
+	pxor	%xmm2,%xmm0
+	addl	$16,%edx
+	leal	-352(%ebx,%ecx,1),%ecx
+	jmp	.L002dec_entry
+.align	16
+.L003dec_loop:
+	movdqa	-32(%ebx),%xmm4
+	movdqa	-16(%ebx),%xmm1
+.byte	102,15,56,0,226
+.byte	102,15,56,0,203
+	pxor	%xmm4,%xmm0
+	movdqa	(%ebx),%xmm4
+	pxor	%xmm1,%xmm0
+	movdqa	16(%ebx),%xmm1
+.byte	102,15,56,0,226
+.byte	102,15,56,0,197
+.byte	102,15,56,0,203
+	pxor	%xmm4,%xmm0
+	movdqa	32(%ebx),%xmm4
+	pxor	%xmm1,%xmm0
+	movdqa	48(%ebx),%xmm1
+.byte	102,15,56,0,226
+.byte	102,15,56,0,197
+.byte	102,15,56,0,203
+	pxor	%xmm4,%xmm0
+	movdqa	64(%ebx),%xmm4
+	pxor	%xmm1,%xmm0
+	movdqa	80(%ebx),%xmm1
+.byte	102,15,56,0,226
+.byte	102,15,56,0,197
+.byte	102,15,56,0,203
+	pxor	%xmm4,%xmm0
+	addl	$16,%edx
+.byte	102,15,58,15,237,12
+	pxor	%xmm1,%xmm0
+	subl	$1,%eax
+.L002dec_entry:
+	movdqa	%xmm6,%xmm1
+	movdqa	-32(%ebp),%xmm2
+	pandn	%xmm0,%xmm1
+	pand	%xmm6,%xmm0
+	psrld	$4,%xmm1
+.byte	102,15,56,0,208
+	movdqa	%xmm7,%xmm3
+	pxor	%xmm1,%xmm0
+.byte	102,15,56,0,217
+	movdqa	%xmm7,%xmm4
+	pxor	%xmm2,%xmm3
+.byte	102,15,56,0,224
+	pxor	%xmm2,%xmm4
+	movdqa	%xmm7,%xmm2
+.byte	102,15,56,0,211
+	movdqa	%xmm7,%xmm3
+	pxor	%xmm0,%xmm2
+.byte	102,15,56,0,220
+	movdqu	(%edx),%xmm0
+	pxor	%xmm1,%xmm3
+	jnz	.L003dec_loop
+	movdqa	96(%ebx),%xmm4
+.byte	102,15,56,0,226
+	pxor	%xmm0,%xmm4
+	movdqa	112(%ebx),%xmm0
+	movdqa	(%ecx),%xmm2
+.byte	102,15,56,0,195
+	pxor	%xmm4,%xmm0
+.byte	102,15,56,0,194
+	ret
+.def	__vpaes_schedule_core;	.scl	3;	.type	32;	.endef
+.align	16
+__vpaes_schedule_core:
+	addl	(%esp),%ebp
+	movdqu	(%esi),%xmm0
+	movdqa	320(%ebp),%xmm2
+	movdqa	%xmm0,%xmm3
+	leal	(%ebp),%ebx
+	movdqa	%xmm2,4(%esp)
+	call	__vpaes_schedule_transform
+	movdqa	%xmm0,%xmm7
+	testl	%edi,%edi
+	jnz	.L004schedule_am_decrypting
+	movdqu	%xmm0,(%edx)
+	jmp	.L005schedule_go
+.L004schedule_am_decrypting:
+	movdqa	256(%ebp,%ecx,1),%xmm1
+.byte	102,15,56,0,217
+	movdqu	%xmm3,(%edx)
+	xorl	$48,%ecx
+.L005schedule_go:
+	cmpl	$192,%eax
+	ja	.L006schedule_256
+	je	.L007schedule_192
+.L008schedule_128:
+	movl	$10,%eax
+.L009loop_schedule_128:
+	call	__vpaes_schedule_round
+	decl	%eax
+	jz	.L010schedule_mangle_last
+	call	__vpaes_schedule_mangle
+	jmp	.L009loop_schedule_128
+.align	16
+.L007schedule_192:
+	movdqu	8(%esi),%xmm0
+	call	__vpaes_schedule_transform
+	movdqa	%xmm0,%xmm6
+	pxor	%xmm4,%xmm4
+	movhlps	%xmm4,%xmm6
+	movl	$4,%eax
+.L011loop_schedule_192:
+	call	__vpaes_schedule_round
+.byte	102,15,58,15,198,8
+	call	__vpaes_schedule_mangle
+	call	__vpaes_schedule_192_smear
+	call	__vpaes_schedule_mangle
+	call	__vpaes_schedule_round
+	decl	%eax
+	jz	.L010schedule_mangle_last
+	call	__vpaes_schedule_mangle
+	call	__vpaes_schedule_192_smear
+	jmp	.L011loop_schedule_192
+.align	16
+.L006schedule_256:
+	movdqu	16(%esi),%xmm0
+	call	__vpaes_schedule_transform
+	movl	$7,%eax
+.L012loop_schedule_256:
+	call	__vpaes_schedule_mangle
+	movdqa	%xmm0,%xmm6
+	call	__vpaes_schedule_round
+	decl	%eax
+	jz	.L010schedule_mangle_last
+	call	__vpaes_schedule_mangle
+	pshufd	$255,%xmm0,%xmm0
+	movdqa	%xmm7,20(%esp)
+	movdqa	%xmm6,%xmm7
+	call	.L_vpaes_schedule_low_round
+	movdqa	20(%esp),%xmm7
+	jmp	.L012loop_schedule_256
+.align	16
+.L010schedule_mangle_last:
+	leal	384(%ebp),%ebx
+	testl	%edi,%edi
+	jnz	.L013schedule_mangle_last_dec
+	movdqa	256(%ebp,%ecx,1),%xmm1
+.byte	102,15,56,0,193
+	leal	352(%ebp),%ebx
+	addl	$32,%edx
+.L013schedule_mangle_last_dec:
+	addl	$-16,%edx
+	pxor	336(%ebp),%xmm0
+	call	__vpaes_schedule_transform
+	movdqu	%xmm0,(%edx)
+	pxor	%xmm0,%xmm0
+	pxor	%xmm1,%xmm1
+	pxor	%xmm2,%xmm2
+	pxor	%xmm3,%xmm3
+	pxor	%xmm4,%xmm4
+	pxor	%xmm5,%xmm5
+	pxor	%xmm6,%xmm6
+	pxor	%xmm7,%xmm7
+	ret
+.def	__vpaes_schedule_192_smear;	.scl	3;	.type	32;	.endef
+.align	16
+__vpaes_schedule_192_smear:
+	pshufd	$128,%xmm6,%xmm1
+	pshufd	$254,%xmm7,%xmm0
+	pxor	%xmm1,%xmm6
+	pxor	%xmm1,%xmm1
+	pxor	%xmm0,%xmm6
+	movdqa	%xmm6,%xmm0
+	movhlps	%xmm1,%xmm6
+	ret
+.def	__vpaes_schedule_round;	.scl	3;	.type	32;	.endef
+.align	16
+__vpaes_schedule_round:
+	movdqa	8(%esp),%xmm2
+	pxor	%xmm1,%xmm1
+.byte	102,15,58,15,202,15
+.byte	102,15,58,15,210,15
+	pxor	%xmm1,%xmm7
+	pshufd	$255,%xmm0,%xmm0
+.byte	102,15,58,15,192,1
+	movdqa	%xmm2,8(%esp)
+.L_vpaes_schedule_low_round:
+	movdqa	%xmm7,%xmm1
+	pslldq	$4,%xmm7
+	pxor	%xmm1,%xmm7
+	movdqa	%xmm7,%xmm1
+	pslldq	$8,%xmm7
+	pxor	%xmm1,%xmm7
+	pxor	336(%ebp),%xmm7
+	movdqa	-16(%ebp),%xmm4
+	movdqa	-48(%ebp),%xmm5
+	movdqa	%xmm4,%xmm1
+	pandn	%xmm0,%xmm1
+	psrld	$4,%xmm1
+	pand	%xmm4,%xmm0
+	movdqa	-32(%ebp),%xmm2
+.byte	102,15,56,0,208
+	pxor	%xmm1,%xmm0
+	movdqa	%xmm5,%xmm3
+.byte	102,15,56,0,217
+	pxor	%xmm2,%xmm3
+	movdqa	%xmm5,%xmm4
+.byte	102,15,56,0,224
+	pxor	%xmm2,%xmm4
+	movdqa	%xmm5,%xmm2
+.byte	102,15,56,0,211
+	pxor	%xmm0,%xmm2
+	movdqa	%xmm5,%xmm3
+.byte	102,15,56,0,220
+	pxor	%xmm1,%xmm3
+	movdqa	32(%ebp),%xmm4
+.byte	102,15,56,0,226
+	movdqa	48(%ebp),%xmm0
+.byte	102,15,56,0,195
+	pxor	%xmm4,%xmm0
+	pxor	%xmm7,%xmm0
+	movdqa	%xmm0,%xmm7
+	ret
+.def	__vpaes_schedule_transform;	.scl	3;	.type	32;	.endef
+.align	16
+__vpaes_schedule_transform:
+	movdqa	-16(%ebp),%xmm2
+	movdqa	%xmm2,%xmm1
+	pandn	%xmm0,%xmm1
+	psrld	$4,%xmm1
+	pand	%xmm2,%xmm0
+	movdqa	(%ebx),%xmm2
+.byte	102,15,56,0,208
+	movdqa	16(%ebx),%xmm0
+.byte	102,15,56,0,193
+	pxor	%xmm2,%xmm0
+	ret
+.def	__vpaes_schedule_mangle;	.scl	3;	.type	32;	.endef
+.align	16
+__vpaes_schedule_mangle:
+	movdqa	%xmm0,%xmm4
+	movdqa	128(%ebp),%xmm5
+	testl	%edi,%edi
+	jnz	.L014schedule_mangle_dec
+	addl	$16,%edx
+	pxor	336(%ebp),%xmm4
+.byte	102,15,56,0,229
+	movdqa	%xmm4,%xmm3
+.byte	102,15,56,0,229
+	pxor	%xmm4,%xmm3
+.byte	102,15,56,0,229
+	pxor	%xmm4,%xmm3
+	jmp	.L015schedule_mangle_both
+.align	16
+.L014schedule_mangle_dec:
+	movdqa	-16(%ebp),%xmm2
+	leal	416(%ebp),%esi
+	movdqa	%xmm2,%xmm1
+	pandn	%xmm4,%xmm1
+	psrld	$4,%xmm1
+	pand	%xmm2,%xmm4
+	movdqa	(%esi),%xmm2
+.byte	102,15,56,0,212
+	movdqa	16(%esi),%xmm3
+.byte	102,15,56,0,217
+	pxor	%xmm2,%xmm3
+.byte	102,15,56,0,221
+	movdqa	32(%esi),%xmm2
+.byte	102,15,56,0,212
+	pxor	%xmm3,%xmm2
+	movdqa	48(%esi),%xmm3
+.byte	102,15,56,0,217
+	pxor	%xmm2,%xmm3
+.byte	102,15,56,0,221
+	movdqa	64(%esi),%xmm2
+.byte	102,15,56,0,212
+	pxor	%xmm3,%xmm2
+	movdqa	80(%esi),%xmm3
+.byte	102,15,56,0,217
+	pxor	%xmm2,%xmm3
+.byte	102,15,56,0,221
+	movdqa	96(%esi),%xmm2
+.byte	102,15,56,0,212
+	pxor	%xmm3,%xmm2
+	movdqa	112(%esi),%xmm3
+.byte	102,15,56,0,217
+	pxor	%xmm2,%xmm3
+	addl	$-16,%edx
+.L015schedule_mangle_both:
+	movdqa	256(%ebp,%ecx,1),%xmm1
+.byte	102,15,56,0,217
+	addl	$-16,%ecx
+	andl	$48,%ecx
+	movdqu	%xmm3,(%edx)
+	ret
+.globl	_vpaes_set_encrypt_key
+.def	_vpaes_set_encrypt_key;	.scl	2;	.type	32;	.endef
+.align	16
+_vpaes_set_encrypt_key:
+.L_vpaes_set_encrypt_key_begin:
+	pushl	%ebp
+	pushl	%ebx
+	pushl	%esi
+	pushl	%edi
+	movl	20(%esp),%esi
+	leal	-56(%esp),%ebx
+	movl	24(%esp),%eax
+	andl	$-16,%ebx
+	movl	28(%esp),%edx
+	xchgl	%esp,%ebx
+	movl	%ebx,48(%esp)
+	movl	%eax,%ebx
+	shrl	$5,%ebx
+	addl	$5,%ebx
+	movl	%ebx,240(%edx)
+	movl	$48,%ecx
+	movl	$0,%edi
+	leal	.L_vpaes_consts+0x30-.L016pic_point,%ebp
+	call	__vpaes_schedule_core
+.L016pic_point:
+	movl	48(%esp),%esp
+	xorl	%eax,%eax
+	popl	%edi
+	popl	%esi
+	popl	%ebx
+	popl	%ebp
+	ret
+.globl	_vpaes_set_decrypt_key
+.def	_vpaes_set_decrypt_key;	.scl	2;	.type	32;	.endef
+.align	16
+_vpaes_set_decrypt_key:
+.L_vpaes_set_decrypt_key_begin:
+	pushl	%ebp
+	pushl	%ebx
+	pushl	%esi
+	pushl	%edi
+	movl	20(%esp),%esi
+	leal	-56(%esp),%ebx
+	movl	24(%esp),%eax
+	andl	$-16,%ebx
+	movl	28(%esp),%edx
+	xchgl	%esp,%ebx
+	movl	%ebx,48(%esp)
+	movl	%eax,%ebx
+	shrl	$5,%ebx
+	addl	$5,%ebx
+	movl	%ebx,240(%edx)
+	shll	$4,%ebx
+	leal	16(%edx,%ebx,1),%edx
+	movl	$1,%edi
+	movl	%eax,%ecx
+	shrl	$1,%ecx
+	andl	$32,%ecx
+	xorl	$32,%ecx
+	leal	.L_vpaes_consts+0x30-.L017pic_point,%ebp
+	call	__vpaes_schedule_core
+.L017pic_point:
+	movl	48(%esp),%esp
+	xorl	%eax,%eax
+	popl	%edi
+	popl	%esi
+	popl	%ebx
+	popl	%ebp
+	ret
+.globl	_vpaes_encrypt
+.def	_vpaes_encrypt;	.scl	2;	.type	32;	.endef
+.align	16
+_vpaes_encrypt:
+.L_vpaes_encrypt_begin:
+	pushl	%ebp
+	pushl	%ebx
+	pushl	%esi
+	pushl	%edi
+	leal	.L_vpaes_consts+0x30-.L018pic_point,%ebp
+	call	__vpaes_preheat
+.L018pic_point:
+	movl	20(%esp),%esi
+	leal	-56(%esp),%ebx
+	movl	24(%esp),%edi
+	andl	$-16,%ebx
+	movl	28(%esp),%edx
+	xchgl	%esp,%ebx
+	movl	%ebx,48(%esp)
+	movdqu	(%esi),%xmm0
+	call	__vpaes_encrypt_core
+	movdqu	%xmm0,(%edi)
+	movl	48(%esp),%esp
+	popl	%edi
+	popl	%esi
+	popl	%ebx
+	popl	%ebp
+	ret
+.globl	_vpaes_decrypt
+.def	_vpaes_decrypt;	.scl	2;	.type	32;	.endef
+.align	16
+_vpaes_decrypt:
+.L_vpaes_decrypt_begin:
+	pushl	%ebp
+	pushl	%ebx
+	pushl	%esi
+	pushl	%edi
+	leal	.L_vpaes_consts+0x30-.L019pic_point,%ebp
+	call	__vpaes_preheat
+.L019pic_point:
+	movl	20(%esp),%esi
+	leal	-56(%esp),%ebx
+	movl	24(%esp),%edi
+	andl	$-16,%ebx
+	movl	28(%esp),%edx
+	xchgl	%esp,%ebx
+	movl	%ebx,48(%esp)
+	movdqu	(%esi),%xmm0
+	call	__vpaes_decrypt_core
+	movdqu	%xmm0,(%edi)
+	movl	48(%esp),%esp
+	popl	%edi
+	popl	%esi
+	popl	%ebx
+	popl	%ebp
+	ret
+.globl	_vpaes_cbc_encrypt
+.def	_vpaes_cbc_encrypt;	.scl	2;	.type	32;	.endef
+.align	16
+_vpaes_cbc_encrypt:
+.L_vpaes_cbc_encrypt_begin:
+	pushl	%ebp
+	pushl	%ebx
+	pushl	%esi
+	pushl	%edi
+	movl	20(%esp),%esi
+	movl	24(%esp),%edi
+	movl	28(%esp),%eax
+	movl	32(%esp),%edx
+	subl	$16,%eax
+	jc	.L020cbc_abort
+	leal	-56(%esp),%ebx
+	movl	36(%esp),%ebp
+	andl	$-16,%ebx
+	movl	40(%esp),%ecx
+	xchgl	%esp,%ebx
+	movdqu	(%ebp),%xmm1
+	subl	%esi,%edi
+	movl	%ebx,48(%esp)
+	movl	%edi,(%esp)
+	movl	%edx,4(%esp)
+	movl	%ebp,8(%esp)
+	movl	%eax,%edi
+	leal	.L_vpaes_consts+0x30-.L021pic_point,%ebp
+	call	__vpaes_preheat
+.L021pic_point:
+	cmpl	$0,%ecx
+	je	.L022cbc_dec_loop
+	jmp	.L023cbc_enc_loop
+.align	16
+.L023cbc_enc_loop:
+	movdqu	(%esi),%xmm0
+	pxor	%xmm1,%xmm0
+	call	__vpaes_encrypt_core
+	movl	(%esp),%ebx
+	movl	4(%esp),%edx
+	movdqa	%xmm0,%xmm1
+	movdqu	%xmm0,(%ebx,%esi,1)
+	leal	16(%esi),%esi
+	subl	$16,%edi
+	jnc	.L023cbc_enc_loop
+	jmp	.L024cbc_done
+.align	16
+.L022cbc_dec_loop:
+	movdqu	(%esi),%xmm0
+	movdqa	%xmm1,16(%esp)
+	movdqa	%xmm0,32(%esp)
+	call	__vpaes_decrypt_core
+	movl	(%esp),%ebx
+	movl	4(%esp),%edx
+	pxor	16(%esp),%xmm0
+	movdqa	32(%esp),%xmm1
+	movdqu	%xmm0,(%ebx,%esi,1)
+	leal	16(%esi),%esi
+	subl	$16,%edi
+	jnc	.L022cbc_dec_loop
+.L024cbc_done:
+	movl	8(%esp),%ebx
+	movl	48(%esp),%esp
+	movdqu	%xmm1,(%ebx)
+.L020cbc_abort:
+	popl	%edi
+	popl	%esi
+	popl	%ebx
+	popl	%ebp
+	ret
+
+.section .note.GNU-stack,"",%progbits
diff --git a/lib/accelerated/x86/coff/aes-ssse3-x86_64.s b/lib/accelerated/x86/coff/aes-ssse3-x86_64.s
new file mode 100644
index 0000000000..f8dbd266be
--- /dev/null
+++ b/lib/accelerated/x86/coff/aes-ssse3-x86_64.s
@@ -0,0 +1,1137 @@
+######################################################################
+## Constant-time SSSE3 AES core implementation.
+## version 0.1
+##
+## By Mike Hamburg (Stanford University), 2009
+## Public domain.
+##
+## For details see http://shiftleft.org/papers/vector_aes/ and
+## http://crypto.stanford.edu/vpaes/.
+#
+# *** This file is auto-generated ***
+#
+.text	
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+.def	_vpaes_encrypt_core;	.scl 3;	.type 32;	.endef
+.p2align	4
+_vpaes_encrypt_core:
+	movq	%rdx,%r9
+	movq	$16,%r11
+	movl	240(%rdx),%eax
+	movdqa	%xmm9,%xmm1
+	movdqa	.Lk_ipt(%rip),%xmm2
+	pandn	%xmm0,%xmm1
+	movdqu	(%r9),%xmm5
+	psrld	$4,%xmm1
+	pand	%xmm9,%xmm0
+.byte	102,15,56,0,208
+	movdqa	.Lk_ipt+16(%rip),%xmm0
+.byte	102,15,56,0,193
+	pxor	%xmm5,%xmm2
+	addq	$16,%r9
+	pxor	%xmm2,%xmm0
+	leaq	.Lk_mc_backward(%rip),%r10
+	jmp	.Lenc_entry
+
+.p2align	4
+.Lenc_loop:
+
+	movdqa	%xmm13,%xmm4
+	movdqa	%xmm12,%xmm0
+.byte	102,15,56,0,226
+.byte	102,15,56,0,195
+	pxor	%xmm5,%xmm4
+	movdqa	%xmm15,%xmm5
+	pxor	%xmm4,%xmm0
+	movdqa	-64(%r11,%r10,1),%xmm1
+.byte	102,15,56,0,234
+	movdqa	(%r11,%r10,1),%xmm4
+	movdqa	%xmm14,%xmm2
+.byte	102,15,56,0,211
+	movdqa	%xmm0,%xmm3
+	pxor	%xmm5,%xmm2
+.byte	102,15,56,0,193
+	addq	$16,%r9
+	pxor	%xmm2,%xmm0
+.byte	102,15,56,0,220
+	addq	$16,%r11
+	pxor	%xmm0,%xmm3
+.byte	102,15,56,0,193
+	andq	$48,%r11
+	subq	$1,%rax
+	pxor	%xmm3,%xmm0
+
+.Lenc_entry:
+
+	movdqa	%xmm9,%xmm1
+	movdqa	%xmm11,%xmm5
+	pandn	%xmm0,%xmm1
+	psrld	$4,%xmm1
+	pand	%xmm9,%xmm0
+.byte	102,15,56,0,232
+	movdqa	%xmm10,%xmm3
+	pxor	%xmm1,%xmm0
+.byte	102,15,56,0,217
+	movdqa	%xmm10,%xmm4
+	pxor	%xmm5,%xmm3
+.byte	102,15,56,0,224
+	movdqa	%xmm10,%xmm2
+	pxor	%xmm5,%xmm4
+.byte	102,15,56,0,211
+	movdqa	%xmm10,%xmm3
+	pxor	%xmm0,%xmm2
+.byte	102,15,56,0,220
+	movdqu	(%r9),%xmm5
+	pxor	%xmm1,%xmm3
+	jnz	.Lenc_loop
+
+
+	movdqa	-96(%r10),%xmm4
+	movdqa	-80(%r10),%xmm0
+.byte	102,15,56,0,226
+	pxor	%xmm5,%xmm4
+.byte	102,15,56,0,195
+	movdqa	64(%r11,%r10,1),%xmm1
+	pxor	%xmm4,%xmm0
+.byte	102,15,56,0,193
+	.byte	0xf3,0xc3
+
+
+
+
+
+
+
+.def	_vpaes_decrypt_core;	.scl 3;	.type 32;	.endef
+.p2align	4
+_vpaes_decrypt_core:
+	movq	%rdx,%r9
+	movl	240(%rdx),%eax
+	movdqa	%xmm9,%xmm1
+	movdqa	.Lk_dipt(%rip),%xmm2
+	pandn	%xmm0,%xmm1
+	movq	%rax,%r11
+	psrld	$4,%xmm1
+	movdqu	(%r9),%xmm5
+	shlq	$4,%r11
+	pand	%xmm9,%xmm0
+.byte	102,15,56,0,208
+	movdqa	.Lk_dipt+16(%rip),%xmm0
+	xorq	$48,%r11
+	leaq	.Lk_dsbd(%rip),%r10
+.byte	102,15,56,0,193
+	andq	$48,%r11
+	pxor	%xmm5,%xmm2
+	movdqa	.Lk_mc_forward+48(%rip),%xmm5
+	pxor	%xmm2,%xmm0
+	addq	$16,%r9
+	addq	%r10,%r11
+	jmp	.Ldec_entry
+
+.p2align	4
+.Ldec_loop:
+
+
+
+	movdqa	-32(%r10),%xmm4
+	movdqa	-16(%r10),%xmm1
+.byte	102,15,56,0,226
+.byte	102,15,56,0,203
+	pxor	%xmm4,%xmm0
+	movdqa	0(%r10),%xmm4
+	pxor	%xmm1,%xmm0
+	movdqa	16(%r10),%xmm1
+
+.byte	102,15,56,0,226
+.byte	102,15,56,0,197
+.byte	102,15,56,0,203
+	pxor	%xmm4,%xmm0
+	movdqa	32(%r10),%xmm4
+	pxor	%xmm1,%xmm0
+	movdqa	48(%r10),%xmm1
+
+.byte	102,15,56,0,226
+.byte	102,15,56,0,197
+.byte	102,15,56,0,203
+	pxor	%xmm4,%xmm0
+	movdqa	64(%r10),%xmm4
+	pxor	%xmm1,%xmm0
+	movdqa	80(%r10),%xmm1
+
+.byte	102,15,56,0,226
+.byte	102,15,56,0,197
+.byte	102,15,56,0,203
+	pxor	%xmm4,%xmm0
+	addq	$16,%r9
+.byte	102,15,58,15,237,12
+	pxor	%xmm1,%xmm0
+	subq	$1,%rax
+
+.Ldec_entry:
+
+	movdqa	%xmm9,%xmm1
+	pandn	%xmm0,%xmm1
+	movdqa	%xmm11,%xmm2
+	psrld	$4,%xmm1
+	pand	%xmm9,%xmm0
+.byte	102,15,56,0,208
+	movdqa	%xmm10,%xmm3
+	pxor	%xmm1,%xmm0
+.byte	102,15,56,0,217
+	movdqa	%xmm10,%xmm4
+	pxor	%xmm2,%xmm3
+.byte	102,15,56,0,224
+	pxor	%xmm2,%xmm4
+	movdqa	%xmm10,%xmm2
+.byte	102,15,56,0,211
+	movdqa	%xmm10,%xmm3
+	pxor	%xmm0,%xmm2
+.byte	102,15,56,0,220
+	movdqu	(%r9),%xmm0
+	pxor	%xmm1,%xmm3
+	jnz	.Ldec_loop
+
+
+	movdqa	96(%r10),%xmm4
+.byte	102,15,56,0,226
+	pxor	%xmm0,%xmm4
+	movdqa	112(%r10),%xmm0
+	movdqa	-352(%r11),%xmm2
+.byte	102,15,56,0,195
+	pxor	%xmm4,%xmm0
+.byte	102,15,56,0,194
+	.byte	0xf3,0xc3
+
+
+
+
+
+
+
+.def	_vpaes_schedule_core;	.scl 3;	.type 32;	.endef
+.p2align	4
+_vpaes_schedule_core:
+
+
+
+
+
+	call	_vpaes_preheat		
+	movdqa	.Lk_rcon(%rip),%xmm8
+	movdqu	(%rdi),%xmm0
+
+
+	movdqa	%xmm0,%xmm3
+	leaq	.Lk_ipt(%rip),%r11
+	call	_vpaes_schedule_transform
+	movdqa	%xmm0,%xmm7
+
+	leaq	.Lk_sr(%rip),%r10
+	testq	%rcx,%rcx
+	jnz	.Lschedule_am_decrypting
+
+
+	movdqu	%xmm0,(%rdx)
+	jmp	.Lschedule_go
+
+.Lschedule_am_decrypting:
+
+	movdqa	(%r8,%r10,1),%xmm1
+.byte	102,15,56,0,217
+	movdqu	%xmm3,(%rdx)
+	xorq	$48,%r8
+
+.Lschedule_go:
+	cmpl	$192,%esi
+	ja	.Lschedule_256
+	je	.Lschedule_192
+
+
+
+
+
+
+
+
+
+
+.Lschedule_128:
+	movl	$10,%esi
+
+.Loop_schedule_128:
+	call	_vpaes_schedule_round
+	decq	%rsi
+	jz	.Lschedule_mangle_last
+	call	_vpaes_schedule_mangle	
+	jmp	.Loop_schedule_128
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+.p2align	4
+.Lschedule_192:
+	movdqu	8(%rdi),%xmm0
+	call	_vpaes_schedule_transform	
+	movdqa	%xmm0,%xmm6
+	pxor	%xmm4,%xmm4
+	movhlps	%xmm4,%xmm6
+	movl	$4,%esi
+
+.Loop_schedule_192:
+	call	_vpaes_schedule_round
+.byte	102,15,58,15,198,8
+	call	_vpaes_schedule_mangle	
+	call	_vpaes_schedule_192_smear
+	call	_vpaes_schedule_mangle	
+	call	_vpaes_schedule_round
+	decq	%rsi
+	jz	.Lschedule_mangle_last
+	call	_vpaes_schedule_mangle	
+	call	_vpaes_schedule_192_smear
+	jmp	.Loop_schedule_192
+
+
+
+
+
+
+
+
+
+
+
+.p2align	4
+.Lschedule_256:
+	movdqu	16(%rdi),%xmm0
+	call	_vpaes_schedule_transform	
+	movl	$7,%esi
+
+.Loop_schedule_256:
+	call	_vpaes_schedule_mangle	
+	movdqa	%xmm0,%xmm6
+
+
+	call	_vpaes_schedule_round
+	decq	%rsi
+	jz	.Lschedule_mangle_last
+	call	_vpaes_schedule_mangle	
+
+
+	pshufd	$255,%xmm0,%xmm0
+	movdqa	%xmm7,%xmm5
+	movdqa	%xmm6,%xmm7
+	call	_vpaes_schedule_low_round
+	movdqa	%xmm5,%xmm7
+
+	jmp	.Loop_schedule_256
+
+
+
+
+
+
+
+
+
+
+
+
+.p2align	4
+.Lschedule_mangle_last:
+
+	leaq	.Lk_deskew(%rip),%r11
+	testq	%rcx,%rcx
+	jnz	.Lschedule_mangle_last_dec
+
+
+	movdqa	(%r8,%r10,1),%xmm1
+.byte	102,15,56,0,193
+	leaq	.Lk_opt(%rip),%r11
+	addq	$32,%rdx
+
+.Lschedule_mangle_last_dec:
+	addq	$-16,%rdx
+	pxor	.Lk_s63(%rip),%xmm0
+	call	_vpaes_schedule_transform 
+	movdqu	%xmm0,(%rdx)
+
+
+	pxor	%xmm0,%xmm0
+	pxor	%xmm1,%xmm1
+	pxor	%xmm2,%xmm2
+	pxor	%xmm3,%xmm3
+	pxor	%xmm4,%xmm4
+	pxor	%xmm5,%xmm5
+	pxor	%xmm6,%xmm6
+	pxor	%xmm7,%xmm7
+	.byte	0xf3,0xc3
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+.def	_vpaes_schedule_192_smear;	.scl 3;	.type 32;	.endef
+.p2align	4
+_vpaes_schedule_192_smear:
+	pshufd	$128,%xmm6,%xmm1
+	pshufd	$254,%xmm7,%xmm0
+	pxor	%xmm1,%xmm6
+	pxor	%xmm1,%xmm1
+	pxor	%xmm0,%xmm6
+	movdqa	%xmm6,%xmm0
+	movhlps	%xmm1,%xmm6
+	.byte	0xf3,0xc3
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+.def	_vpaes_schedule_round;	.scl 3;	.type 32;	.endef
+.p2align	4
+_vpaes_schedule_round:
+
+	pxor	%xmm1,%xmm1
+.byte	102,65,15,58,15,200,15
+.byte	102,69,15,58,15,192,15
+	pxor	%xmm1,%xmm7
+
+
+	pshufd	$255,%xmm0,%xmm0
+.byte	102,15,58,15,192,1
+
+
+
+
+_vpaes_schedule_low_round:
+
+	movdqa	%xmm7,%xmm1
+	pslldq	$4,%xmm7
+	pxor	%xmm1,%xmm7
+	movdqa	%xmm7,%xmm1
+	pslldq	$8,%xmm7
+	pxor	%xmm1,%xmm7
+	pxor	.Lk_s63(%rip),%xmm7
+
+
+	movdqa	%xmm9,%xmm1
+	pandn	%xmm0,%xmm1
+	psrld	$4,%xmm1
+	pand	%xmm9,%xmm0
+	movdqa	%xmm11,%xmm2
+.byte	102,15,56,0,208
+	pxor	%xmm1,%xmm0
+	movdqa	%xmm10,%xmm3
+.byte	102,15,56,0,217
+	pxor	%xmm2,%xmm3
+	movdqa	%xmm10,%xmm4
+.byte	102,15,56,0,224
+	pxor	%xmm2,%xmm4
+	movdqa	%xmm10,%xmm2
+.byte	102,15,56,0,211
+	pxor	%xmm0,%xmm2
+	movdqa	%xmm10,%xmm3
+.byte	102,15,56,0,220
+	pxor	%xmm1,%xmm3
+	movdqa	%xmm13,%xmm4
+.byte	102,15,56,0,226
+	movdqa	%xmm12,%xmm0
+.byte	102,15,56,0,195
+	pxor	%xmm4,%xmm0
+
+
+	pxor	%xmm7,%xmm0
+	movdqa	%xmm0,%xmm7
+	.byte	0xf3,0xc3
+
+
+
+
+
+
+
+
+
+
+
+.def	_vpaes_schedule_transform;	.scl 3;	.type 32;	.endef
+.p2align	4
+_vpaes_schedule_transform:
+	movdqa	%xmm9,%xmm1
+	pandn	%xmm0,%xmm1
+	psrld	$4,%xmm1
+	pand	%xmm9,%xmm0
+	movdqa	(%r11),%xmm2
+.byte	102,15,56,0,208
+	movdqa	16(%r11),%xmm0
+.byte	102,15,56,0,193
+	pxor	%xmm2,%xmm0
+	.byte	0xf3,0xc3
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+.def	_vpaes_schedule_mangle;	.scl 3;	.type 32;	.endef
+.p2align	4
+_vpaes_schedule_mangle:
+	movdqa	%xmm0,%xmm4
+	movdqa	.Lk_mc_forward(%rip),%xmm5
+	testq	%rcx,%rcx
+	jnz	.Lschedule_mangle_dec
+
+
+	addq	$16,%rdx
+	pxor	.Lk_s63(%rip),%xmm4
+.byte	102,15,56,0,229
+	movdqa	%xmm4,%xmm3
+.byte	102,15,56,0,229
+	pxor	%xmm4,%xmm3
+.byte	102,15,56,0,229
+	pxor	%xmm4,%xmm3
+
+	jmp	.Lschedule_mangle_both
+.p2align	4
+.Lschedule_mangle_dec:
+
+	leaq	.Lk_dksd(%rip),%r11
+	movdqa	%xmm9,%xmm1
+	pandn	%xmm4,%xmm1
+	psrld	$4,%xmm1
+	pand	%xmm9,%xmm4
+
+	movdqa	0(%r11),%xmm2
+.byte	102,15,56,0,212
+	movdqa	16(%r11),%xmm3
+.byte	102,15,56,0,217
+	pxor	%xmm2,%xmm3
+.byte	102,15,56,0,221
+
+	movdqa	32(%r11),%xmm2
+.byte	102,15,56,0,212
+	pxor	%xmm3,%xmm2
+	movdqa	48(%r11),%xmm3
+.byte	102,15,56,0,217
+	pxor	%xmm2,%xmm3
+.byte	102,15,56,0,221
+
+	movdqa	64(%r11),%xmm2
+.byte	102,15,56,0,212
+	pxor	%xmm3,%xmm2
+	movdqa	80(%r11),%xmm3
+.byte	102,15,56,0,217
+	pxor	%xmm2,%xmm3
+.byte	102,15,56,0,221
+
+	movdqa	96(%r11),%xmm2
+.byte	102,15,56,0,212
+	pxor	%xmm3,%xmm2
+	movdqa	112(%r11),%xmm3
+.byte	102,15,56,0,217
+	pxor	%xmm2,%xmm3
+
+	addq	$-16,%rdx
+
+.Lschedule_mangle_both:
+	movdqa	(%r8,%r10,1),%xmm1
+.byte	102,15,56,0,217
+	addq	$-16,%r8
+	andq	$48,%r8
+	movdqu	%xmm3,(%rdx)
+	.byte	0xf3,0xc3
+
+
+
+
+
+.globl	vpaes_set_encrypt_key
+.def	vpaes_set_encrypt_key;	.scl 2;	.type 32;	.endef
+.p2align	4
+vpaes_set_encrypt_key:
+	movq	%rdi,8(%rsp)
+	movq	%rsi,16(%rsp)
+	movq	%rsp,%rax
+.LSEH_begin_vpaes_set_encrypt_key:
+	movq	%rcx,%rdi
+	movq	%rdx,%rsi
+	movq	%r8,%rdx
+
+	leaq	-184(%rsp),%rsp
+	movaps	%xmm6,16(%rsp)
+	movaps	%xmm7,32(%rsp)
+	movaps	%xmm8,48(%rsp)
+	movaps	%xmm9,64(%rsp)
+	movaps	%xmm10,80(%rsp)
+	movaps	%xmm11,96(%rsp)
+	movaps	%xmm12,112(%rsp)
+	movaps	%xmm13,128(%rsp)
+	movaps	%xmm14,144(%rsp)
+	movaps	%xmm15,160(%rsp)
+.Lenc_key_body:
+	movl	%esi,%eax
+	shrl	$5,%eax
+	addl	$5,%eax
+	movl	%eax,240(%rdx)
+
+	movl	$0,%ecx
+	movl	$48,%r8d
+	call	_vpaes_schedule_core
+	movaps	16(%rsp),%xmm6
+	movaps	32(%rsp),%xmm7
+	movaps	48(%rsp),%xmm8
+	movaps	64(%rsp),%xmm9
+	movaps	80(%rsp),%xmm10
+	movaps	96(%rsp),%xmm11
+	movaps	112(%rsp),%xmm12
+	movaps	128(%rsp),%xmm13
+	movaps	144(%rsp),%xmm14
+	movaps	160(%rsp),%xmm15
+	leaq	184(%rsp),%rsp
+.Lenc_key_epilogue:
+	xorl	%eax,%eax
+	movq	8(%rsp),%rdi
+	movq	16(%rsp),%rsi
+	.byte	0xf3,0xc3
+.LSEH_end_vpaes_set_encrypt_key:
+
+.globl	vpaes_set_decrypt_key
+.def	vpaes_set_decrypt_key;	.scl 2;	.type 32;	.endef
+.p2align	4
+vpaes_set_decrypt_key:
+	movq	%rdi,8(%rsp)
+	movq	%rsi,16(%rsp)
+	movq	%rsp,%rax
+.LSEH_begin_vpaes_set_decrypt_key:
+	movq	%rcx,%rdi
+	movq	%rdx,%rsi
+	movq	%r8,%rdx
+
+	leaq	-184(%rsp),%rsp
+	movaps	%xmm6,16(%rsp)
+	movaps	%xmm7,32(%rsp)
+	movaps	%xmm8,48(%rsp)
+	movaps	%xmm9,64(%rsp)
+	movaps	%xmm10,80(%rsp)
+	movaps	%xmm11,96(%rsp)
+	movaps	%xmm12,112(%rsp)
+	movaps	%xmm13,128(%rsp)
+	movaps	%xmm14,144(%rsp)
+	movaps	%xmm15,160(%rsp)
+.Ldec_key_body:
+	movl	%esi,%eax
+	shrl	$5,%eax
+	addl	$5,%eax
+	movl	%eax,240(%rdx)
+	shll	$4,%eax
+	leaq	16(%rdx,%rax,1),%rdx
+
+	movl	$1,%ecx
+	movl	%esi,%r8d
+	shrl	$1,%r8d
+	andl	$32,%r8d
+	xorl	$32,%r8d
+	call	_vpaes_schedule_core
+	movaps	16(%rsp),%xmm6
+	movaps	32(%rsp),%xmm7
+	movaps	48(%rsp),%xmm8
+	movaps	64(%rsp),%xmm9
+	movaps	80(%rsp),%xmm10
+	movaps	96(%rsp),%xmm11
+	movaps	112(%rsp),%xmm12
+	movaps	128(%rsp),%xmm13
+	movaps	144(%rsp),%xmm14
+	movaps	160(%rsp),%xmm15
+	leaq	184(%rsp),%rsp
+.Ldec_key_epilogue:
+	xorl	%eax,%eax
+	movq	8(%rsp),%rdi
+	movq	16(%rsp),%rsi
+	.byte	0xf3,0xc3
+.LSEH_end_vpaes_set_decrypt_key:
+
+.globl	vpaes_encrypt
+.def	vpaes_encrypt;	.scl 2;	.type 32;	.endef
+.p2align	4
+vpaes_encrypt:
+	movq	%rdi,8(%rsp)
+	movq	%rsi,16(%rsp)
+	movq	%rsp,%rax
+.LSEH_begin_vpaes_encrypt:
+	movq	%rcx,%rdi
+	movq	%rdx,%rsi
+	movq	%r8,%rdx
+
+	leaq	-184(%rsp),%rsp
+	movaps	%xmm6,16(%rsp)
+	movaps	%xmm7,32(%rsp)
+	movaps	%xmm8,48(%rsp)
+	movaps	%xmm9,64(%rsp)
+	movaps	%xmm10,80(%rsp)
+	movaps	%xmm11,96(%rsp)
+	movaps	%xmm12,112(%rsp)
+	movaps	%xmm13,128(%rsp)
+	movaps	%xmm14,144(%rsp)
+	movaps	%xmm15,160(%rsp)
+.Lenc_body:
+	movdqu	(%rdi),%xmm0
+	call	_vpaes_preheat
+	call	_vpaes_encrypt_core
+	movdqu	%xmm0,(%rsi)
+	movaps	16(%rsp),%xmm6
+	movaps	32(%rsp),%xmm7
+	movaps	48(%rsp),%xmm8
+	movaps	64(%rsp),%xmm9
+	movaps	80(%rsp),%xmm10
+	movaps	96(%rsp),%xmm11
+	movaps	112(%rsp),%xmm12
+	movaps	128(%rsp),%xmm13
+	movaps	144(%rsp),%xmm14
+	movaps	160(%rsp),%xmm15
+	leaq	184(%rsp),%rsp
+.Lenc_epilogue:
+	movq	8(%rsp),%rdi
+	movq	16(%rsp),%rsi
+	.byte	0xf3,0xc3
+.LSEH_end_vpaes_encrypt:
+
+.globl	vpaes_decrypt
+.def	vpaes_decrypt;	.scl 2;	.type 32;	.endef
+.p2align	4
+vpaes_decrypt:
+	movq	%rdi,8(%rsp)
+	movq	%rsi,16(%rsp)
+	movq	%rsp,%rax
+.LSEH_begin_vpaes_decrypt:
+	movq	%rcx,%rdi
+	movq	%rdx,%rsi
+	movq	%r8,%rdx
+
+	leaq	-184(%rsp),%rsp
+	movaps	%xmm6,16(%rsp)
+	movaps	%xmm7,32(%rsp)
+	movaps	%xmm8,48(%rsp)
+	movaps	%xmm9,64(%rsp)
+	movaps	%xmm10,80(%rsp)
+	movaps	%xmm11,96(%rsp)
+	movaps	%xmm12,112(%rsp)
+	movaps	%xmm13,128(%rsp)
+	movaps	%xmm14,144(%rsp)
+	movaps	%xmm15,160(%rsp)
+.Ldec_body:
+	movdqu	(%rdi),%xmm0
+	call	_vpaes_preheat
+	call	_vpaes_decrypt_core
+	movdqu	%xmm0,(%rsi)
+	movaps	16(%rsp),%xmm6
+	movaps	32(%rsp),%xmm7
+	movaps	48(%rsp),%xmm8
+	movaps	64(%rsp),%xmm9
+	movaps	80(%rsp),%xmm10
+	movaps	96(%rsp),%xmm11
+	movaps	112(%rsp),%xmm12
+	movaps	128(%rsp),%xmm13
+	movaps	144(%rsp),%xmm14
+	movaps	160(%rsp),%xmm15
+	leaq	184(%rsp),%rsp
+.Ldec_epilogue:
+	movq	8(%rsp),%rdi
+	movq	16(%rsp),%rsi
+	.byte	0xf3,0xc3
+.LSEH_end_vpaes_decrypt:
+.globl	vpaes_cbc_encrypt
+.def	vpaes_cbc_encrypt;	.scl 2;	.type 32;	.endef
+.p2align	4
+vpaes_cbc_encrypt:
+	movq	%rdi,8(%rsp)
+	movq	%rsi,16(%rsp)
+	movq	%rsp,%rax
+.LSEH_begin_vpaes_cbc_encrypt:
+	movq	%rcx,%rdi
+	movq	%rdx,%rsi
+	movq	%r8,%rdx
+	movq	%r9,%rcx
+	movq	40(%rsp),%r8
+	movq	48(%rsp),%r9
+
+	xchgq	%rcx,%rdx
+	subq	$16,%rcx
+	jc	.Lcbc_abort
+	leaq	-184(%rsp),%rsp
+	movaps	%xmm6,16(%rsp)
+	movaps	%xmm7,32(%rsp)
+	movaps	%xmm8,48(%rsp)
+	movaps	%xmm9,64(%rsp)
+	movaps	%xmm10,80(%rsp)
+	movaps	%xmm11,96(%rsp)
+	movaps	%xmm12,112(%rsp)
+	movaps	%xmm13,128(%rsp)
+	movaps	%xmm14,144(%rsp)
+	movaps	%xmm15,160(%rsp)
+.Lcbc_body:
+	movdqu	(%r8),%xmm6
+	subq	%rdi,%rsi
+	call	_vpaes_preheat
+	cmpl	$0,%r9d
+	je	.Lcbc_dec_loop
+	jmp	.Lcbc_enc_loop
+.p2align	4
+.Lcbc_enc_loop:
+	movdqu	(%rdi),%xmm0
+	pxor	%xmm6,%xmm0
+	call	_vpaes_encrypt_core
+	movdqa	%xmm0,%xmm6
+	movdqu	%xmm0,(%rsi,%rdi,1)
+	leaq	16(%rdi),%rdi
+	subq	$16,%rcx
+	jnc	.Lcbc_enc_loop
+	jmp	.Lcbc_done
+.p2align	4
+.Lcbc_dec_loop:
+	movdqu	(%rdi),%xmm0
+	movdqa	%xmm0,%xmm7
+	call	_vpaes_decrypt_core
+	pxor	%xmm6,%xmm0
+	movdqa	%xmm7,%xmm6
+	movdqu	%xmm0,(%rsi,%rdi,1)
+	leaq	16(%rdi),%rdi
+	subq	$16,%rcx
+	jnc	.Lcbc_dec_loop
+.Lcbc_done:
+	movdqu	%xmm6,(%r8)
+	movaps	16(%rsp),%xmm6
+	movaps	32(%rsp),%xmm7
+	movaps	48(%rsp),%xmm8
+	movaps	64(%rsp),%xmm9
+	movaps	80(%rsp),%xmm10
+	movaps	96(%rsp),%xmm11
+	movaps	112(%rsp),%xmm12
+	movaps	128(%rsp),%xmm13
+	movaps	144(%rsp),%xmm14
+	movaps	160(%rsp),%xmm15
+	leaq	184(%rsp),%rsp
+.Lcbc_epilogue:
+.Lcbc_abort:
+	movq	8(%rsp),%rdi
+	movq	16(%rsp),%rsi
+	.byte	0xf3,0xc3
+.LSEH_end_vpaes_cbc_encrypt:
+
+
+
+
+
+
+.def	_vpaes_preheat;	.scl 3;	.type 32;	.endef
+.p2align	4
+_vpaes_preheat:
+	leaq	.Lk_s0F(%rip),%r10
+	movdqa	-32(%r10),%xmm10
+	movdqa	-16(%r10),%xmm11
+	movdqa	0(%r10),%xmm9
+	movdqa	48(%r10),%xmm13
+	movdqa	64(%r10),%xmm12
+	movdqa	80(%r10),%xmm15
+	movdqa	96(%r10),%xmm14
+	.byte	0xf3,0xc3
+
+
+
+
+
+
+
+.p2align	6
+_vpaes_consts:
+.Lk_inv:
+.quad	0x0E05060F0D080180, 0x040703090A0B0C02
+.quad	0x01040A060F0B0780, 0x030D0E0C02050809
+
+.Lk_s0F:
+.quad	0x0F0F0F0F0F0F0F0F, 0x0F0F0F0F0F0F0F0F
+
+.Lk_ipt:
+.quad	0xC2B2E8985A2A7000, 0xCABAE09052227808
+.quad	0x4C01307D317C4D00, 0xCD80B1FCB0FDCC81
+
+.Lk_sb1:
+.quad	0xB19BE18FCB503E00, 0xA5DF7A6E142AF544
+.quad	0x3618D415FAE22300, 0x3BF7CCC10D2ED9EF
+.Lk_sb2:
+.quad	0xE27A93C60B712400, 0x5EB7E955BC982FCD
+.quad	0x69EB88400AE12900, 0xC2A163C8AB82234A
+.Lk_sbo:
+.quad	0xD0D26D176FBDC700, 0x15AABF7AC502A878
+.quad	0xCFE474A55FBB6A00, 0x8E1E90D1412B35FA
+
+.Lk_mc_forward:
+.quad	0x0407060500030201, 0x0C0F0E0D080B0A09
+.quad	0x080B0A0904070605, 0x000302010C0F0E0D
+.quad	0x0C0F0E0D080B0A09, 0x0407060500030201
+.quad	0x000302010C0F0E0D, 0x080B0A0904070605
+
+.Lk_mc_backward:
+.quad	0x0605040702010003, 0x0E0D0C0F0A09080B
+.quad	0x020100030E0D0C0F, 0x0A09080B06050407
+.quad	0x0E0D0C0F0A09080B, 0x0605040702010003
+.quad	0x0A09080B06050407, 0x020100030E0D0C0F
+
+.Lk_sr:
+.quad	0x0706050403020100, 0x0F0E0D0C0B0A0908
+.quad	0x030E09040F0A0500, 0x0B06010C07020D08
+.quad	0x0F060D040B020900, 0x070E050C030A0108
+.quad	0x0B0E0104070A0D00, 0x0306090C0F020508
+
+.Lk_rcon:
+.quad	0x1F8391B9AF9DEEB6, 0x702A98084D7C7D81
+
+.Lk_s63:
+.quad	0x5B5B5B5B5B5B5B5B, 0x5B5B5B5B5B5B5B5B
+
+.Lk_opt:
+.quad	0xFF9F4929D6B66000, 0xF7974121DEBE6808
+.quad	0x01EDBD5150BCEC00, 0xE10D5DB1B05C0CE0
+
+.Lk_deskew:
+.quad	0x07E4A34047A4E300, 0x1DFEB95A5DBEF91A
+.quad	0x5F36B5DC83EA6900, 0x2841C2ABF49D1E77
+
+
+
+
+
+.Lk_dksd:
+.quad	0xFEB91A5DA3E44700, 0x0740E3A45A1DBEF9
+.quad	0x41C277F4B5368300, 0x5FDC69EAAB289D1E
+.Lk_dksb:
+.quad	0x9A4FCA1F8550D500, 0x03D653861CC94C99
+.quad	0x115BEDA7B6FC4A00, 0xD993256F7E3482C8
+.Lk_dkse:
+.quad	0xD5031CCA1FC9D600, 0x53859A4C994F5086
+.quad	0xA23196054FDC7BE8, 0xCD5EF96A20B31487
+.Lk_dks9:
+.quad	0xB6116FC87ED9A700, 0x4AED933482255BFC
+.quad	0x4576516227143300, 0x8BB89FACE9DAFDCE
+
+
+
+
+
+.Lk_dipt:
+.quad	0x0F505B040B545F00, 0x154A411E114E451A
+.quad	0x86E383E660056500, 0x12771772F491F194
+
+.Lk_dsb9:
+.quad	0x851C03539A86D600, 0xCAD51F504F994CC9
+.quad	0xC03B1789ECD74900, 0x725E2C9EB2FBA565
+.Lk_dsbd:
+.quad	0x7D57CCDFE6B1A200, 0xF56E9B13882A4439
+.quad	0x3CE2FAF724C6CB00, 0x2931180D15DEEFD3
+.Lk_dsbb:
+.quad	0xD022649296B44200, 0x602646F6B0F2D404
+.quad	0xC19498A6CD596700, 0xF3FF0C3E3255AA6B
+.Lk_dsbe:
+.quad	0x46F2929626D4D000, 0x2242600464B4F6B0
+.quad	0x0C55A6CDFFAAC100, 0x9467F36B98593E32
+.Lk_dsbo:
+.quad	0x1387EA537EF94000, 0xC7AA6DB9D4943E2D
+.quad	0x12D7560F93441D00, 0xCA4B8159D8C58E9C
+.byte	86,101,99,116,111,114,32,80,101,114,109,117,116,97,105,111,110,32,65,69,83,32,102,111,114,32,120,56,54,95,54,52,47,83,83,83,69,51,44,32,77,105,107,101,32,72,97,109,98,117,114,103,32,40,83,116,97,110,102,111,114,100,32,85,110,105,118,101,114,115,105,116,121,41,0
+.p2align	6
+
+
+.def	se_handler;	.scl 3;	.type 32;	.endef
+.p2align	4
+se_handler:
+	pushq	%rsi
+	pushq	%rdi
+	pushq	%rbx
+	pushq	%rbp
+	pushq	%r12
+	pushq	%r13
+	pushq	%r14
+	pushq	%r15
+	pushfq
+	subq	$64,%rsp
+
+	movq	120(%r8),%rax
+	movq	248(%r8),%rbx
+
+	movq	8(%r9),%rsi
+	movq	56(%r9),%r11
+
+	movl	0(%r11),%r10d
+	leaq	(%rsi,%r10,1),%r10
+	cmpq	%r10,%rbx
+	jb	.Lin_prologue
+
+	movq	152(%r8),%rax
+
+	movl	4(%r11),%r10d
+	leaq	(%rsi,%r10,1),%r10
+	cmpq	%r10,%rbx
+	jae	.Lin_prologue
+
+	leaq	16(%rax),%rsi
+	leaq	512(%r8),%rdi
+	movl	$20,%ecx
+.long	0xa548f3fc		
+	leaq	184(%rax),%rax
+
+.Lin_prologue:
+	movq	8(%rax),%rdi
+	movq	16(%rax),%rsi
+	movq	%rax,152(%r8)
+	movq	%rsi,168(%r8)
+	movq	%rdi,176(%r8)
+
+	movq	40(%r9),%rdi
+	movq	%r8,%rsi
+	movl	$154,%ecx
+.long	0xa548f3fc		
+
+	movq	%r9,%rsi
+	xorq	%rcx,%rcx
+	movq	8(%rsi),%rdx
+	movq	0(%rsi),%r8
+	movq	16(%rsi),%r9
+	movq	40(%rsi),%r10
+	leaq	56(%rsi),%r11
+	leaq	24(%rsi),%r12
+	movq	%r10,32(%rsp)
+	movq	%r11,40(%rsp)
+	movq	%r12,48(%rsp)
+	movq	%rcx,56(%rsp)
+	call	*__imp_RtlVirtualUnwind(%rip)
+
+	movl	$1,%eax
+	addq	$64,%rsp
+	popfq
+	popq	%r15
+	popq	%r14
+	popq	%r13
+	popq	%r12
+	popq	%rbp
+	popq	%rbx
+	popq	%rdi
+	popq	%rsi
+	.byte	0xf3,0xc3
+
+
+.section	.pdata
+.p2align	2
+.rva	.LSEH_begin_vpaes_set_encrypt_key
+.rva	.LSEH_end_vpaes_set_encrypt_key
+.rva	.LSEH_info_vpaes_set_encrypt_key
+
+.rva	.LSEH_begin_vpaes_set_decrypt_key
+.rva	.LSEH_end_vpaes_set_decrypt_key
+.rva	.LSEH_info_vpaes_set_decrypt_key
+
+.rva	.LSEH_begin_vpaes_encrypt
+.rva	.LSEH_end_vpaes_encrypt
+.rva	.LSEH_info_vpaes_encrypt
+
+.rva	.LSEH_begin_vpaes_decrypt
+.rva	.LSEH_end_vpaes_decrypt
+.rva	.LSEH_info_vpaes_decrypt
+
+.rva	.LSEH_begin_vpaes_cbc_encrypt
+.rva	.LSEH_end_vpaes_cbc_encrypt
+.rva	.LSEH_info_vpaes_cbc_encrypt
+
+.section	.xdata
+.p2align	3
+.LSEH_info_vpaes_set_encrypt_key:
+.byte	9,0,0,0
+.rva	se_handler
+.rva	.Lenc_key_body,.Lenc_key_epilogue	
+.LSEH_info_vpaes_set_decrypt_key:
+.byte	9,0,0,0
+.rva	se_handler
+.rva	.Ldec_key_body,.Ldec_key_epilogue	
+.LSEH_info_vpaes_encrypt:
+.byte	9,0,0,0
+.rva	se_handler
+.rva	.Lenc_body,.Lenc_epilogue		
+.LSEH_info_vpaes_decrypt:
+.byte	9,0,0,0
+.rva	se_handler
+.rva	.Ldec_body,.Ldec_epilogue		
+.LSEH_info_vpaes_cbc_encrypt:
+.byte	9,0,0,0
+.rva	se_handler
+.rva	.Lcbc_body,.Lcbc_epilogue		
+
+.section .note.GNU-stack,"",%progbits
diff --git a/lib/accelerated/x86/coff/aesni-x86.s b/lib/accelerated/x86/coff/aesni-x86.s
index 1970712d30..9c982a20b5 100644
--- a/lib/accelerated/x86/coff/aesni-x86.s
+++ b/lib/accelerated/x86/coff/aesni-x86.s
@@ -1,4 +1,4 @@
-# Copyright (c) 2011-2012, Andy Polyakov <appro@openssl.org>
+# Copyright (c) 2011-2013, Andy Polyakov <appro@openssl.org>
 # All rights reserved.
 #
 # Redistribution and use in source and binary forms, with or without
diff --git a/lib/accelerated/x86/coff/aesni-x86_64.s b/lib/accelerated/x86/coff/aesni-x86_64.s
index 85b51085a5..30b822394b 100644
--- a/lib/accelerated/x86/coff/aesni-x86_64.s
+++ b/lib/accelerated/x86/coff/aesni-x86_64.s
@@ -1,4 +1,4 @@
-# Copyright (c) 2011-2012, Andy Polyakov <appro@openssl.org>
+# Copyright (c) 2011-2013, Andy Polyakov <appro@openssl.org>
 # All rights reserved.
 #
 # Redistribution and use in source and binary forms, with or without
diff --git a/lib/accelerated/x86/coff/cpuid-x86.s b/lib/accelerated/x86/coff/cpuid-x86.s
index f35cfba63a..9931ff05ba 100644
--- a/lib/accelerated/x86/coff/cpuid-x86.s
+++ b/lib/accelerated/x86/coff/cpuid-x86.s
@@ -1,5 +1,6 @@
 #
-# Copyright (C) 2011-2012 Free Software Foundation, Inc.
+# Copyright (C) 2011-2013 Free Software Foundation, Inc.
+# Copyright (C) 2013 Nikos Mavrogiannopoulos
 #
 # Author: Nikos Mavrogiannopoulos
 #
diff --git a/lib/accelerated/x86/coff/cpuid-x86_64.s b/lib/accelerated/x86/coff/cpuid-x86_64.s
index 033df92ebf..3add1900cf 100644
--- a/lib/accelerated/x86/coff/cpuid-x86_64.s
+++ b/lib/accelerated/x86/coff/cpuid-x86_64.s
@@ -1,5 +1,6 @@
 #
-# Copyright (C) 2011-2012 Free Software Foundation, Inc.
+# Copyright (C) 2011-2013 Free Software Foundation, Inc.
+# Copyright (C) 2013 Nikos Mavrogiannopoulos
 #
 # Author: Nikos Mavrogiannopoulos
 #
diff --git a/lib/accelerated/x86/coff/e_padlock-x86.s b/lib/accelerated/x86/coff/e_padlock-x86.s
index d51d62ff73..328e6462f6 100644
--- a/lib/accelerated/x86/coff/e_padlock-x86.s
+++ b/lib/accelerated/x86/coff/e_padlock-x86.s
@@ -1,4 +1,4 @@
-# Copyright (c) 2011-2012, Andy Polyakov <appro@openssl.org>
+# Copyright (c) 2011-2013, Andy Polyakov <appro@openssl.org>
 # All rights reserved.
 #
 # Redistribution and use in source and binary forms, with or without
diff --git a/lib/accelerated/x86/coff/e_padlock-x86_64.s b/lib/accelerated/x86/coff/e_padlock-x86_64.s
index 14c62fd176..6b73825042 100644
--- a/lib/accelerated/x86/coff/e_padlock-x86_64.s
+++ b/lib/accelerated/x86/coff/e_padlock-x86_64.s
@@ -1,4 +1,4 @@
-# Copyright (c) 2011-2012, Andy Polyakov <appro@openssl.org>
+# Copyright (c) 2011-2013, Andy Polyakov <appro@openssl.org>
 # All rights reserved.
 #
 # Redistribution and use in source and binary forms, with or without
diff --git a/lib/accelerated/x86/coff/ghash-x86_64.s b/lib/accelerated/x86/coff/ghash-x86_64.s
index 951ee891b9..d61c82b9cc 100644
--- a/lib/accelerated/x86/coff/ghash-x86_64.s
+++ b/lib/accelerated/x86/coff/ghash-x86_64.s
@@ -1,4 +1,4 @@
-# Copyright (c) 2011-2012, Andy Polyakov <appro@openssl.org>
+# Copyright (c) 2011-2013, Andy Polyakov <appro@openssl.org>
 # All rights reserved.
 #
 # Redistribution and use in source and binary forms, with or without
diff --git a/lib/accelerated/x86/coff/sha1-ssse3-x86.s b/lib/accelerated/x86/coff/sha1-ssse3-x86.s
index 9bd41a0de4..450f574e16 100644
--- a/lib/accelerated/x86/coff/sha1-ssse3-x86.s
+++ b/lib/accelerated/x86/coff/sha1-ssse3-x86.s
@@ -1,4 +1,4 @@
-# Copyright (c) 2011-2012, Andy Polyakov <appro@openssl.org>
+# Copyright (c) 2011-2013, Andy Polyakov <appro@openssl.org>
 # All rights reserved.
 #
 # Redistribution and use in source and binary forms, with or without
diff --git a/lib/accelerated/x86/coff/sha1-ssse3-x86_64.s b/lib/accelerated/x86/coff/sha1-ssse3-x86_64.s
index 75868a42c6..98fd50dcc1 100644
--- a/lib/accelerated/x86/coff/sha1-ssse3-x86_64.s
+++ b/lib/accelerated/x86/coff/sha1-ssse3-x86_64.s
@@ -1,4 +1,4 @@
-# Copyright (c) 2011-2012, Andy Polyakov <appro@openssl.org>
+# Copyright (c) 2011-2013, Andy Polyakov <appro@openssl.org>
 # All rights reserved.
 #
 # Redistribution and use in source and binary forms, with or without
diff --git a/lib/accelerated/x86/coff/sha256-ssse3-x86.s b/lib/accelerated/x86/coff/sha256-ssse3-x86.s
index 6fe27746ce..117b2bd413 100644
--- a/lib/accelerated/x86/coff/sha256-ssse3-x86.s
+++ b/lib/accelerated/x86/coff/sha256-ssse3-x86.s
@@ -1,4 +1,4 @@
-# Copyright (c) 2011-2012, Andy Polyakov <appro@openssl.org>
+# Copyright (c) 2011-2013, Andy Polyakov <appro@openssl.org>
 # All rights reserved.
 #
 # Redistribution and use in source and binary forms, with or without
diff --git a/lib/accelerated/x86/coff/sha512-ssse3-x86.s b/lib/accelerated/x86/coff/sha512-ssse3-x86.s
index 79098da5c2..d68eeffb4a 100644
--- a/lib/accelerated/x86/coff/sha512-ssse3-x86.s
+++ b/lib/accelerated/x86/coff/sha512-ssse3-x86.s
@@ -1,4 +1,4 @@
-# Copyright (c) 2011-2012, Andy Polyakov <appro@openssl.org>
+# Copyright (c) 2011-2013, Andy Polyakov <appro@openssl.org>
 # All rights reserved.
 #
 # Redistribution and use in source and binary forms, with or without
diff --git a/lib/accelerated/x86/coff/sha512-ssse3-x86_64.s b/lib/accelerated/x86/coff/sha512-ssse3-x86_64.s
index bbb2661f26..dd80574c89 100644
--- a/lib/accelerated/x86/coff/sha512-ssse3-x86_64.s
+++ b/lib/accelerated/x86/coff/sha512-ssse3-x86_64.s
@@ -1,4 +1,4 @@
-# Copyright (c) 2011-2012, Andy Polyakov <appro@openssl.org>
+# Copyright (c) 2011-2013, Andy Polyakov <appro@openssl.org>
 # All rights reserved.
 #
 # Redistribution and use in source and binary forms, with or without