Update to openssl-1.0.1

1 files changed, 310 insertions, 134 deletions

diff --git a/openssl/crypto/aes/asm/aes-ppc.pl b/openssl/crypto/aes/asm/aes-ppc.pl
index f82c5e181..7c52cbe5f 100644
--- a/openssl/crypto/aes/asm/aes-ppc.pl
+++ b/openssl/crypto/aes/asm/aes-ppc.pl
@@ -7,7 +7,7 @@
 # details see http://www.openssl.org/~appro/cryptogams/.
 # ====================================================================
 
-# Needs more work: key setup, page boundaries, CBC routine...
+# Needs more work: key setup, CBC routine...
 #
 # ppc_AES_[en|de]crypt perform at 18 cycles per byte processed with
 # 128-bit key, which is ~40% better than 64-bit code generated by gcc
@@ -18,7 +18,7 @@
 
 # February 2010
 #
-# Rescheduling instructions to favour Power6 pipeline gives 10%
+# Rescheduling instructions to favour Power6 pipeline gave 10%
 # performance improvement on the platfrom in question (and marginal
 # improvement even on others). It should be noted that Power6 fails
 # to process byte in 18 cycles, only in 23, because it fails to issue
@@ -33,11 +33,13 @@ $flavour = shift;
 
 if ($flavour =~ /64/) {
 	$SIZE_T	=8;
+	$LRSAVE	=2*$SIZE_T;
 	$STU	="stdu";
 	$POP	="ld";
 	$PUSH	="std";
 } elsif ($flavour =~ /32/) {
 	$SIZE_T	=4;
+	$LRSAVE	=$SIZE_T;
 	$STU	="stwu";
 	$POP	="lwz";
 	$PUSH	="stw";
@@ -116,15 +118,19 @@ LAES_Te:
 	addi	$Tbl0,$Tbl0,`128-8`
 	mtlr	r0
 	blr
-	.space	`32-24`
+	.long	0
+	.byte	0,12,0x14,0,0,0,0,0
+	.space	`64-9*4`
 LAES_Td:
 	mflr	r0
 	bcl	20,31,\$+4
 	mflr	$Tbl0	;    vvvvvvvv "distance" between . and 1st data entry
-	addi	$Tbl0,$Tbl0,`128-8-32+2048+256`
+	addi	$Tbl0,$Tbl0,`128-64-8+2048+256`
 	mtlr	r0
 	blr
-	.space	`128-32-24`
+	.long	0
+	.byte	0,12,0x14,0,0,0,0,0
+	.space	`128-64-9*4`
 ___
 &_data_word(
 	0xc66363a5, 0xf87c7c84, 0xee777799, 0xf67b7b8d,
@@ -328,10 +334,9 @@ $code.=<<___;
 .globl	.AES_encrypt
 .align	7
 .AES_encrypt:
-	mflr	r0
 	$STU	$sp,-$FRAME($sp)
+	mflr	r0
 
-	$PUSH	r0,`$FRAME-$SIZE_T*21`($sp)
 	$PUSH	$toc,`$FRAME-$SIZE_T*20`($sp)
 	$PUSH	r13,`$FRAME-$SIZE_T*19`($sp)
 	$PUSH	r14,`$FRAME-$SIZE_T*18`($sp)
@@ -352,7 +357,14 @@ $code.=<<___;
 	$PUSH	r29,`$FRAME-$SIZE_T*3`($sp)
 	$PUSH	r30,`$FRAME-$SIZE_T*2`($sp)
 	$PUSH	r31,`$FRAME-$SIZE_T*1`($sp)
+	$PUSH	r0,`$FRAME+$LRSAVE`($sp)
+
+	andi.	$t0,$inp,3
+	andi.	$t1,$out,3
+	or.	$t0,$t0,$t1
+	bne	Lenc_unaligned
 
+Lenc_unaligned_ok:
 	lwz	$s0,0($inp)
 	lwz	$s1,4($inp)
 	lwz	$s2,8($inp)
@@ -363,8 +375,80 @@ $code.=<<___;
 	stw	$s1,4($out)
 	stw	$s2,8($out)
 	stw	$s3,12($out)
+	b	Lenc_done
+
+Lenc_unaligned:
+	subfic	$t0,$inp,4096
+	subfic	$t1,$out,4096
+	andi.	$t0,$t0,4096-16
+	beq	Lenc_xpage
+	andi.	$t1,$t1,4096-16
+	bne	Lenc_unaligned_ok
+
+Lenc_xpage:
+	lbz	$acc00,0($inp)
+	lbz	$acc01,1($inp)
+	lbz	$acc02,2($inp)
+	lbz	$s0,3($inp)
+	lbz	$acc04,4($inp)
+	lbz	$acc05,5($inp)
+	lbz	$acc06,6($inp)
+	lbz	$s1,7($inp)
+	lbz	$acc08,8($inp)
+	lbz	$acc09,9($inp)
+	lbz	$acc10,10($inp)
+	insrwi	$s0,$acc00,8,0
+	lbz	$s2,11($inp)
+	insrwi	$s1,$acc04,8,0
+	lbz	$acc12,12($inp)
+	insrwi	$s0,$acc01,8,8
+	lbz	$acc13,13($inp)
+	insrwi	$s1,$acc05,8,8
+	lbz	$acc14,14($inp)
+	insrwi	$s0,$acc02,8,16
+	lbz	$s3,15($inp)
+	insrwi	$s1,$acc06,8,16
+	insrwi	$s2,$acc08,8,0
+	insrwi	$s3,$acc12,8,0
+	insrwi	$s2,$acc09,8,8
+	insrwi	$s3,$acc13,8,8
+	insrwi	$s2,$acc10,8,16
+	insrwi	$s3,$acc14,8,16
+
+	bl	LAES_Te
+	bl	Lppc_AES_encrypt_compact
+
+	extrwi	$acc00,$s0,8,0
+	extrwi	$acc01,$s0,8,8
+	stb	$acc00,0($out)
+	extrwi	$acc02,$s0,8,16
+	stb	$acc01,1($out)
+	stb	$acc02,2($out)
+	extrwi	$acc04,$s1,8,0
+	stb	$s0,3($out)
+	extrwi	$acc05,$s1,8,8
+	stb	$acc04,4($out)
+	extrwi	$acc06,$s1,8,16
+	stb	$acc05,5($out)
+	stb	$acc06,6($out)
+	extrwi	$acc08,$s2,8,0
+	stb	$s1,7($out)
+	extrwi	$acc09,$s2,8,8
+	stb	$acc08,8($out)
+	extrwi	$acc10,$s2,8,16
+	stb	$acc09,9($out)
+	stb	$acc10,10($out)
+	extrwi	$acc12,$s3,8,0
+	stb	$s2,11($out)
+	extrwi	$acc13,$s3,8,8
+	stb	$acc12,12($out)
+	extrwi	$acc14,$s3,8,16
+	stb	$acc13,13($out)
+	stb	$acc14,14($out)
+	stb	$s3,15($out)
 
-	$POP	r0,`$FRAME-$SIZE_T*21`($sp)
+Lenc_done:
+	$POP	r0,`$FRAME+$LRSAVE`($sp)
 	$POP	$toc,`$FRAME-$SIZE_T*20`($sp)
 	$POP	r13,`$FRAME-$SIZE_T*19`($sp)
 	$POP	r14,`$FRAME-$SIZE_T*18`($sp)
@@ -388,18 +472,21 @@ $code.=<<___;
 	mtlr	r0
 	addi	$sp,$sp,$FRAME
 	blr
+	.long	0
+	.byte	0,12,4,1,0x80,18,3,0
+	.long	0
 
 .align	5
 Lppc_AES_encrypt:
 	lwz	$acc00,240($key)
-	lwz	$t0,0($key)
-	lwz	$t1,4($key)
-	lwz	$t2,8($key)
-	lwz	$t3,12($key)
 	addi	$Tbl1,$Tbl0,3
+	lwz	$t0,0($key)
 	addi	$Tbl2,$Tbl0,2
+	lwz	$t1,4($key)
 	addi	$Tbl3,$Tbl0,1
+	lwz	$t2,8($key)
 	addi	$acc00,$acc00,-1
+	lwz	$t3,12($key)
 	addi	$key,$key,16
 	xor	$s0,$s0,$t0
 	xor	$s1,$s1,$t1
@@ -413,44 +500,44 @@ Lenc_loop:
 	rlwinm	$acc02,$s2,`32-24+3`,21,28
 	rlwinm	$acc03,$s3,`32-24+3`,21,28
 	lwz	$t0,0($key)
-	lwz	$t1,4($key)
 	rlwinm	$acc04,$s1,`32-16+3`,21,28
+	lwz	$t1,4($key)
 	rlwinm	$acc05,$s2,`32-16+3`,21,28
 	lwz	$t2,8($key)
-	lwz	$t3,12($key)
 	rlwinm	$acc06,$s3,`32-16+3`,21,28
+	lwz	$t3,12($key)
 	rlwinm	$acc07,$s0,`32-16+3`,21,28
 	lwzx	$acc00,$Tbl0,$acc00
-	lwzx	$acc01,$Tbl0,$acc01
 	rlwinm	$acc08,$s2,`32-8+3`,21,28
+	lwzx	$acc01,$Tbl0,$acc01
 	rlwinm	$acc09,$s3,`32-8+3`,21,28
 	lwzx	$acc02,$Tbl0,$acc02
-	lwzx	$acc03,$Tbl0,$acc03
 	rlwinm	$acc10,$s0,`32-8+3`,21,28
+	lwzx	$acc03,$Tbl0,$acc03
 	rlwinm	$acc11,$s1,`32-8+3`,21,28
 	lwzx	$acc04,$Tbl1,$acc04
-	lwzx	$acc05,$Tbl1,$acc05
 	rlwinm	$acc12,$s3,`0+3`,21,28
+	lwzx	$acc05,$Tbl1,$acc05
 	rlwinm	$acc13,$s0,`0+3`,21,28
 	lwzx	$acc06,$Tbl1,$acc06
-	lwzx	$acc07,$Tbl1,$acc07
 	rlwinm	$acc14,$s1,`0+3`,21,28
+	lwzx	$acc07,$Tbl1,$acc07
 	rlwinm	$acc15,$s2,`0+3`,21,28
 	lwzx	$acc08,$Tbl2,$acc08
-	lwzx	$acc09,$Tbl2,$acc09
 	xor	$t0,$t0,$acc00
+	lwzx	$acc09,$Tbl2,$acc09
 	xor	$t1,$t1,$acc01
 	lwzx	$acc10,$Tbl2,$acc10
-	lwzx	$acc11,$Tbl2,$acc11
 	xor	$t2,$t2,$acc02
+	lwzx	$acc11,$Tbl2,$acc11
 	xor	$t3,$t3,$acc03
 	lwzx	$acc12,$Tbl3,$acc12
-	lwzx	$acc13,$Tbl3,$acc13
 	xor	$t0,$t0,$acc04
+	lwzx	$acc13,$Tbl3,$acc13
 	xor	$t1,$t1,$acc05
 	lwzx	$acc14,$Tbl3,$acc14
-	lwzx	$acc15,$Tbl3,$acc15
 	xor	$t2,$t2,$acc06
+	lwzx	$acc15,$Tbl3,$acc15
 	xor	$t3,$t3,$acc07
 	xor	$t0,$t0,$acc08
 	xor	$t1,$t1,$acc09
@@ -466,60 +553,60 @@ Lenc_loop:
 	addi	$Tbl2,$Tbl0,2048
 	nop
 	lwz	$t0,0($key)
-	lwz	$t1,4($key)
 	rlwinm	$acc00,$s0,`32-24`,24,31
+	lwz	$t1,4($key)
 	rlwinm	$acc01,$s1,`32-24`,24,31
 	lwz	$t2,8($key)
-	lwz	$t3,12($key)
 	rlwinm	$acc02,$s2,`32-24`,24,31
+	lwz	$t3,12($key)
 	rlwinm	$acc03,$s3,`32-24`,24,31
 	lwz	$acc08,`2048+0`($Tbl0)	! prefetch Te4
-	lwz	$acc09,`2048+32`($Tbl0)
 	rlwinm	$acc04,$s1,`32-16`,24,31
+	lwz	$acc09,`2048+32`($Tbl0)
 	rlwinm	$acc05,$s2,`32-16`,24,31
 	lwz	$acc10,`2048+64`($Tbl0)
-	lwz	$acc11,`2048+96`($Tbl0)
 	rlwinm	$acc06,$s3,`32-16`,24,31
+	lwz	$acc11,`2048+96`($Tbl0)
 	rlwinm	$acc07,$s0,`32-16`,24,31
 	lwz	$acc12,`2048+128`($Tbl0)
-	lwz	$acc13,`2048+160`($Tbl0)
 	rlwinm	$acc08,$s2,`32-8`,24,31
+	lwz	$acc13,`2048+160`($Tbl0)
 	rlwinm	$acc09,$s3,`32-8`,24,31
 	lwz	$acc14,`2048+192`($Tbl0)
-	lwz	$acc15,`2048+224`($Tbl0)
 	rlwinm	$acc10,$s0,`32-8`,24,31
+	lwz	$acc15,`2048+224`($Tbl0)
 	rlwinm	$acc11,$s1,`32-8`,24,31
 	lbzx	$acc00,$Tbl2,$acc00
-	lbzx	$acc01,$Tbl2,$acc01
 	rlwinm	$acc12,$s3,`0`,24,31
+	lbzx	$acc01,$Tbl2,$acc01
 	rlwinm	$acc13,$s0,`0`,24,31
 	lbzx	$acc02,$Tbl2,$acc02
-	lbzx	$acc03,$Tbl2,$acc03
 	rlwinm	$acc14,$s1,`0`,24,31
+	lbzx	$acc03,$Tbl2,$acc03
 	rlwinm	$acc15,$s2,`0`,24,31
 	lbzx	$acc04,$Tbl2,$acc04
-	lbzx	$acc05,$Tbl2,$acc05
 	rlwinm	$s0,$acc00,24,0,7
+	lbzx	$acc05,$Tbl2,$acc05
 	rlwinm	$s1,$acc01,24,0,7
 	lbzx	$acc06,$Tbl2,$acc06
-	lbzx	$acc07,$Tbl2,$acc07
 	rlwinm	$s2,$acc02,24,0,7
+	lbzx	$acc07,$Tbl2,$acc07
 	rlwinm	$s3,$acc03,24,0,7
 	lbzx	$acc08,$Tbl2,$acc08
-	lbzx	$acc09,$Tbl2,$acc09
 	rlwimi	$s0,$acc04,16,8,15
+	lbzx	$acc09,$Tbl2,$acc09
 	rlwimi	$s1,$acc05,16,8,15
 	lbzx	$acc10,$Tbl2,$acc10
-	lbzx	$acc11,$Tbl2,$acc11
 	rlwimi	$s2,$acc06,16,8,15
+	lbzx	$acc11,$Tbl2,$acc11
 	rlwimi	$s3,$acc07,16,8,15
 	lbzx	$acc12,$Tbl2,$acc12
-	lbzx	$acc13,$Tbl2,$acc13
 	rlwimi	$s0,$acc08,8,16,23
+	lbzx	$acc13,$Tbl2,$acc13
 	rlwimi	$s1,$acc09,8,16,23
 	lbzx	$acc14,$Tbl2,$acc14
-	lbzx	$acc15,$Tbl2,$acc15
 	rlwimi	$s2,$acc10,8,16,23
+	lbzx	$acc15,$Tbl2,$acc15
 	rlwimi	$s3,$acc11,8,16,23
 	or	$s0,$s0,$acc12
 	or	$s1,$s1,$acc13
@@ -530,29 +617,31 @@ Lenc_loop:
 	xor	$s2,$s2,$t2
 	xor	$s3,$s3,$t3
 	blr
+	.long	0
+	.byte	0,12,0x14,0,0,0,0,0
 
 .align	4
 Lppc_AES_encrypt_compact:
 	lwz	$acc00,240($key)
-	lwz	$t0,0($key)
-	lwz	$t1,4($key)
-	lwz	$t2,8($key)
-	lwz	$t3,12($key)
 	addi	$Tbl1,$Tbl0,2048
+	lwz	$t0,0($key)
 	lis	$mask80,0x8080
+	lwz	$t1,4($key)
 	lis	$mask1b,0x1b1b
-	addi	$key,$key,16
+	lwz	$t2,8($key)
 	ori	$mask80,$mask80,0x8080
+	lwz	$t3,12($key)
 	ori	$mask1b,$mask1b,0x1b1b
+	addi	$key,$key,16
 	mtctr	$acc00
 .align	4
 Lenc_compact_loop:
 	xor	$s0,$s0,$t0
 	xor	$s1,$s1,$t1
-	xor	$s2,$s2,$t2
-	xor	$s3,$s3,$t3
 	rlwinm	$acc00,$s0,`32-24`,24,31
+	xor	$s2,$s2,$t2
 	rlwinm	$acc01,$s1,`32-24`,24,31
+	xor	$s3,$s3,$t3
 	rlwinm	$acc02,$s2,`32-24`,24,31
 	rlwinm	$acc03,$s3,`32-24`,24,31
 	rlwinm	$acc04,$s1,`32-16`,24,31
@@ -560,48 +649,48 @@ Lenc_compact_loop:
 	rlwinm	$acc06,$s3,`32-16`,24,31
 	rlwinm	$acc07,$s0,`32-16`,24,31
 	lbzx	$acc00,$Tbl1,$acc00
-	lbzx	$acc01,$Tbl1,$acc01
 	rlwinm	$acc08,$s2,`32-8`,24,31
+	lbzx	$acc01,$Tbl1,$acc01
 	rlwinm	$acc09,$s3,`32-8`,24,31
 	lbzx	$acc02,$Tbl1,$acc02
-	lbzx	$acc03,$Tbl1,$acc03
 	rlwinm	$acc10,$s0,`32-8`,24,31
+	lbzx	$acc03,$Tbl1,$acc03
 	rlwinm	$acc11,$s1,`32-8`,24,31
 	lbzx	$acc04,$Tbl1,$acc04
-	lbzx	$acc05,$Tbl1,$acc05
 	rlwinm	$acc12,$s3,`0`,24,31
+	lbzx	$acc05,$Tbl1,$acc05
 	rlwinm	$acc13,$s0,`0`,24,31
 	lbzx	$acc06,$Tbl1,$acc06
-	lbzx	$acc07,$Tbl1,$acc07
 	rlwinm	$acc14,$s1,`0`,24,31
+	lbzx	$acc07,$Tbl1,$acc07
 	rlwinm	$acc15,$s2,`0`,24,31
 	lbzx	$acc08,$Tbl1,$acc08
-	lbzx	$acc09,$Tbl1,$acc09
 	rlwinm	$s0,$acc00,24,0,7
+	lbzx	$acc09,$Tbl1,$acc09
 	rlwinm	$s1,$acc01,24,0,7
 	lbzx	$acc10,$Tbl1,$acc10
-	lbzx	$acc11,$Tbl1,$acc11
 	rlwinm	$s2,$acc02,24,0,7
+	lbzx	$acc11,$Tbl1,$acc11
 	rlwinm	$s3,$acc03,24,0,7
 	lbzx	$acc12,$Tbl1,$acc12
-	lbzx	$acc13,$Tbl1,$acc13
 	rlwimi	$s0,$acc04,16,8,15
+	lbzx	$acc13,$Tbl1,$acc13
 	rlwimi	$s1,$acc05,16,8,15
 	lbzx	$acc14,$Tbl1,$acc14
-	lbzx	$acc15,$Tbl1,$acc15
 	rlwimi	$s2,$acc06,16,8,15
+	lbzx	$acc15,$Tbl1,$acc15
 	rlwimi	$s3,$acc07,16,8,15
 	rlwimi	$s0,$acc08,8,16,23
 	rlwimi	$s1,$acc09,8,16,23
 	rlwimi	$s2,$acc10,8,16,23
 	rlwimi	$s3,$acc11,8,16,23
 	lwz	$t0,0($key)
-	lwz	$t1,4($key)
 	or	$s0,$s0,$acc12
+	lwz	$t1,4($key)
 	or	$s1,$s1,$acc13
 	lwz	$t2,8($key)
-	lwz	$t3,12($key)
 	or	$s2,$s2,$acc14
+	lwz	$t3,12($key)
 	or	$s3,$s3,$acc15
 
 	addi	$key,$key,16
@@ -612,12 +701,12 @@ Lenc_compact_loop:
 	and	$acc02,$s2,$mask80
 	and	$acc03,$s3,$mask80
 	srwi	$acc04,$acc00,7		# r1>>7
-	srwi	$acc05,$acc01,7
-	srwi	$acc06,$acc02,7
-	srwi	$acc07,$acc03,7
 	andc	$acc08,$s0,$mask80	# r0&0x7f7f7f7f
+	srwi	$acc05,$acc01,7
 	andc	$acc09,$s1,$mask80
+	srwi	$acc06,$acc02,7
 	andc	$acc10,$s2,$mask80
+	srwi	$acc07,$acc03,7
 	andc	$acc11,$s3,$mask80
 	sub	$acc00,$acc00,$acc04	# r1-(r1>>7)
 	sub	$acc01,$acc01,$acc05
@@ -633,32 +722,32 @@ Lenc_compact_loop:
 	and	$acc03,$acc03,$mask1b
 	xor	$acc00,$acc00,$acc08	# r2
 	xor	$acc01,$acc01,$acc09
+	 rotlwi	$acc12,$s0,16		# ROTATE(r0,16)
 	xor	$acc02,$acc02,$acc10
+	 rotlwi	$acc13,$s1,16
 	xor	$acc03,$acc03,$acc11
+	 rotlwi	$acc14,$s2,16
 
-	rotlwi	$acc12,$s0,16		# ROTATE(r0,16)
-	rotlwi	$acc13,$s1,16
-	rotlwi	$acc14,$s2,16
-	rotlwi	$acc15,$s3,16
 	xor	$s0,$s0,$acc00		# r0^r2
+	rotlwi	$acc15,$s3,16
 	xor	$s1,$s1,$acc01
-	xor	$s2,$s2,$acc02
-	xor	$s3,$s3,$acc03
 	rotrwi	$s0,$s0,24		# ROTATE(r2^r0,24)
+	xor	$s2,$s2,$acc02
 	rotrwi	$s1,$s1,24
+	xor	$s3,$s3,$acc03
 	rotrwi	$s2,$s2,24
-	rotrwi	$s3,$s3,24
 	xor	$s0,$s0,$acc00		# ROTATE(r2^r0,24)^r2
+	rotrwi	$s3,$s3,24
 	xor	$s1,$s1,$acc01
 	xor	$s2,$s2,$acc02
 	xor	$s3,$s3,$acc03
 	rotlwi	$acc08,$acc12,8		# ROTATE(r0,24)
-	rotlwi	$acc09,$acc13,8
-	rotlwi	$acc10,$acc14,8
-	rotlwi	$acc11,$acc15,8
 	xor	$s0,$s0,$acc12		#
+	rotlwi	$acc09,$acc13,8
 	xor	$s1,$s1,$acc13
+	rotlwi	$acc10,$acc14,8
 	xor	$s2,$s2,$acc14
+	rotlwi	$acc11,$acc15,8
 	xor	$s3,$s3,$acc15
 	xor	$s0,$s0,$acc08		#
 	xor	$s1,$s1,$acc09
@@ -673,14 +762,15 @@ Lenc_compact_done:
 	xor	$s2,$s2,$t2
 	xor	$s3,$s3,$t3
 	blr
+	.long	0
+	.byte	0,12,0x14,0,0,0,0,0
 
 .globl	.AES_decrypt
 .align	7
 .AES_decrypt:
-	mflr	r0
 	$STU	$sp,-$FRAME($sp)
+	mflr	r0
 
-	$PUSH	r0,`$FRAME-$SIZE_T*21`($sp)
 	$PUSH	$toc,`$FRAME-$SIZE_T*20`($sp)
 	$PUSH	r13,`$FRAME-$SIZE_T*19`($sp)
 	$PUSH	r14,`$FRAME-$SIZE_T*18`($sp)
@@ -701,7 +791,14 @@ Lenc_compact_done:
 	$PUSH	r29,`$FRAME-$SIZE_T*3`($sp)
 	$PUSH	r30,`$FRAME-$SIZE_T*2`($sp)
 	$PUSH	r31,`$FRAME-$SIZE_T*1`($sp)
+	$PUSH	r0,`$FRAME+$LRSAVE`($sp)
 
+	andi.	$t0,$inp,3
+	andi.	$t1,$out,3
+	or.	$t0,$t0,$t1
+	bne	Ldec_unaligned
+
+Ldec_unaligned_ok:
 	lwz	$s0,0($inp)
 	lwz	$s1,4($inp)
 	lwz	$s2,8($inp)
@@ -712,8 +809,80 @@ Lenc_compact_done:
 	stw	$s1,4($out)
 	stw	$s2,8($out)
 	stw	$s3,12($out)
+	b	Ldec_done
+
+Ldec_unaligned:
+	subfic	$t0,$inp,4096
+	subfic	$t1,$out,4096
+	andi.	$t0,$t0,4096-16
+	beq	Ldec_xpage
+	andi.	$t1,$t1,4096-16
+	bne	Ldec_unaligned_ok
+
+Ldec_xpage:
+	lbz	$acc00,0($inp)
+	lbz	$acc01,1($inp)
+	lbz	$acc02,2($inp)
+	lbz	$s0,3($inp)
+	lbz	$acc04,4($inp)
+	lbz	$acc05,5($inp)
+	lbz	$acc06,6($inp)
+	lbz	$s1,7($inp)
+	lbz	$acc08,8($inp)
+	lbz	$acc09,9($inp)
+	lbz	$acc10,10($inp)
+	insrwi	$s0,$acc00,8,0
+	lbz	$s2,11($inp)
+	insrwi	$s1,$acc04,8,0
+	lbz	$acc12,12($inp)
+	insrwi	$s0,$acc01,8,8
+	lbz	$acc13,13($inp)
+	insrwi	$s1,$acc05,8,8
+	lbz	$acc14,14($inp)
+	insrwi	$s0,$acc02,8,16
+	lbz	$s3,15($inp)
+	insrwi	$s1,$acc06,8,16
+	insrwi	$s2,$acc08,8,0
+	insrwi	$s3,$acc12,8,0
+	insrwi	$s2,$acc09,8,8
+	insrwi	$s3,$acc13,8,8
+	insrwi	$s2,$acc10,8,16
+	insrwi	$s3,$acc14,8,16
+
+	bl	LAES_Td
+	bl	Lppc_AES_decrypt_compact
 
-	$POP	r0,`$FRAME-$SIZE_T*21`($sp)
+	extrwi	$acc00,$s0,8,0
+	extrwi	$acc01,$s0,8,8
+	stb	$acc00,0($out)
+	extrwi	$acc02,$s0,8,16
+	stb	$acc01,1($out)
+	stb	$acc02,2($out)
+	extrwi	$acc04,$s1,8,0
+	stb	$s0,3($out)
+	extrwi	$acc05,$s1,8,8
+	stb	$acc04,4($out)
+	extrwi	$acc06,$s1,8,16
+	stb	$acc05,5($out)
+	stb	$acc06,6($out)
+	extrwi	$acc08,$s2,8,0
+	stb	$s1,7($out)
+	extrwi	$acc09,$s2,8,8
+	stb	$acc08,8($out)
+	extrwi	$acc10,$s2,8,16
+	stb	$acc09,9($out)
+	stb	$acc10,10($out)
+	extrwi	$acc12,$s3,8,0
+	stb	$s2,11($out)
+	extrwi	$acc13,$s3,8,8
+	stb	$acc12,12($out)
+	extrwi	$acc14,$s3,8,16
+	stb	$acc13,13($out)
+	stb	$acc14,14($out)
+	stb	$s3,15($out)
+
+Ldec_done:
+	$POP	r0,`$FRAME+$LRSAVE`($sp)
 	$POP	$toc,`$FRAME-$SIZE_T*20`($sp)
 	$POP	r13,`$FRAME-$SIZE_T*19`($sp)
 	$POP	r14,`$FRAME-$SIZE_T*18`($sp)
@@ -737,18 +906,21 @@ Lenc_compact_done:
 	mtlr	r0
 	addi	$sp,$sp,$FRAME
 	blr
+	.long	0
+	.byte	0,12,4,1,0x80,18,3,0
+	.long	0
 
 .align	5
 Lppc_AES_decrypt:
 	lwz	$acc00,240($key)
-	lwz	$t0,0($key)
-	lwz	$t1,4($key)
-	lwz	$t2,8($key)
-	lwz	$t3,12($key)
 	addi	$Tbl1,$Tbl0,3
+	lwz	$t0,0($key)
 	addi	$Tbl2,$Tbl0,2
+	lwz	$t1,4($key)
 	addi	$Tbl3,$Tbl0,1
+	lwz	$t2,8($key)
 	addi	$acc00,$acc00,-1
+	lwz	$t3,12($key)
 	addi	$key,$key,16
 	xor	$s0,$s0,$t0
 	xor	$s1,$s1,$t1
@@ -762,44 +934,44 @@ Ldec_loop:
 	rlwinm	$acc02,$s2,`32-24+3`,21,28
 	rlwinm	$acc03,$s3,`32-24+3`,21,28
 	lwz	$t0,0($key)
-	lwz	$t1,4($key)
 	rlwinm	$acc04,$s3,`32-16+3`,21,28
+	lwz	$t1,4($key)
 	rlwinm	$acc05,$s0,`32-16+3`,21,28
 	lwz	$t2,8($key)
-	lwz	$t3,12($key)
 	rlwinm	$acc06,$s1,`32-16+3`,21,28
+	lwz	$t3,12($key)
 	rlwinm	$acc07,$s2,`32-16+3`,21,28
 	lwzx	$acc00,$Tbl0,$acc00
-	lwzx	$acc01,$Tbl0,$acc01
 	rlwinm	$acc08,$s2,`32-8+3`,21,28
+	lwzx	$acc01,$Tbl0,$acc01
 	rlwinm	$acc09,$s3,`32-8+3`,21,28
 	lwzx	$acc02,$Tbl0,$acc02
-	lwzx	$acc03,$Tbl0,$acc03
 	rlwinm	$acc10,$s0,`32-8+3`,21,28
+	lwzx	$acc03,$Tbl0,$acc03
 	rlwinm	$acc11,$s1,`32-8+3`,21,28
 	lwzx	$acc04,$Tbl1,$acc04
-	lwzx	$acc05,$Tbl1,$acc05
 	rlwinm	$acc12,$s1,`0+3`,21,28
+	lwzx	$acc05,$Tbl1,$acc05
 	rlwinm	$acc13,$s2,`0+3`,21,28
 	lwzx	$acc06,$Tbl1,$acc06
-	lwzx	$acc07,$Tbl1,$acc07
 	rlwinm	$acc14,$s3,`0+3`,21,28
+	lwzx	$acc07,$Tbl1,$acc07
 	rlwinm	$acc15,$s0,`0+3`,21,28
 	lwzx	$acc08,$Tbl2,$acc08
-	lwzx	$acc09,$Tbl2,$acc09
 	xor	$t0,$t0,$acc00
+	lwzx	$acc09,$Tbl2,$acc09
 	xor	$t1,$t1,$acc01
 	lwzx	$acc10,$Tbl2,$acc10
-	lwzx	$acc11,$Tbl2,$acc11
 	xor	$t2,$t2,$acc02
+	lwzx	$acc11,$Tbl2,$acc11
 	xor	$t3,$t3,$acc03
 	lwzx	$acc12,$Tbl3,$acc12
-	lwzx	$acc13,$Tbl3,$acc13
 	xor	$t0,$t0,$acc04
+	lwzx	$acc13,$Tbl3,$acc13
 	xor	$t1,$t1,$acc05
 	lwzx	$acc14,$Tbl3,$acc14
-	lwzx	$acc15,$Tbl3,$acc15
 	xor	$t2,$t2,$acc06
+	lwzx	$acc15,$Tbl3,$acc15
 	xor	$t3,$t3,$acc07
 	xor	$t0,$t0,$acc08
 	xor	$t1,$t1,$acc09
@@ -815,56 +987,56 @@ Ldec_loop:
 	addi	$Tbl2,$Tbl0,2048
 	nop
 	lwz	$t0,0($key)
-	lwz	$t1,4($key)
 	rlwinm	$acc00,$s0,`32-24`,24,31
+	lwz	$t1,4($key)
 	rlwinm	$acc01,$s1,`32-24`,24,31
 	lwz	$t2,8($key)
-	lwz	$t3,12($key)
 	rlwinm	$acc02,$s2,`32-24`,24,31
+	lwz	$t3,12($key)
 	rlwinm	$acc03,$s3,`32-24`,24,31
 	lwz	$acc08,`2048+0`($Tbl0)	! prefetch Td4
-	lwz	$acc09,`2048+32`($Tbl0)
 	rlwinm	$acc04,$s3,`32-16`,24,31
+	lwz	$acc09,`2048+32`($Tbl0)
 	rlwinm	$acc05,$s0,`32-16`,24,31
 	lwz	$acc10,`2048+64`($Tbl0)
-	lwz	$acc11,`2048+96`($Tbl0)
 	lbzx	$acc00,$Tbl2,$acc00
+	lwz	$acc11,`2048+96`($Tbl0)
 	lbzx	$acc01,$Tbl2,$acc01
 	lwz	$acc12,`2048+128`($Tbl0)
-	lwz	$acc13,`2048+160`($Tbl0)
 	rlwinm	$acc06,$s1,`32-16`,24,31
+	lwz	$acc13,`2048+160`($Tbl0)
 	rlwinm	$acc07,$s2,`32-16`,24,31
 	lwz	$acc14,`2048+192`($Tbl0)
-	lwz	$acc15,`2048+224`($Tbl0)
 	rlwinm	$acc08,$s2,`32-8`,24,31
+	lwz	$acc15,`2048+224`($Tbl0)
 	rlwinm	$acc09,$s3,`32-8`,24,31
 	lbzx	$acc02,$Tbl2,$acc02
-	lbzx	$acc03,$Tbl2,$acc03
 	rlwinm	$acc10,$s0,`32-8`,24,31
+	lbzx	$acc03,$Tbl2,$acc03
 	rlwinm	$acc11,$s1,`32-8`,24,31
 	lbzx	$acc04,$Tbl2,$acc04
-	lbzx	$acc05,$Tbl2,$acc05
 	rlwinm	$acc12,$s1,`0`,24,31
+	lbzx	$acc05,$Tbl2,$acc05
 	rlwinm	$acc13,$s2,`0`,24,31
 	lbzx	$acc06,$Tbl2,$acc06
-	lbzx	$acc07,$Tbl2,$acc07
 	rlwinm	$acc14,$s3,`0`,24,31
+	lbzx	$acc07,$Tbl2,$acc07
 	rlwinm	$acc15,$s0,`0`,24,31
 	lbzx	$acc08,$Tbl2,$acc08
-	lbzx	$acc09,$Tbl2,$acc09
 	rlwinm	$s0,$acc00,24,0,7
+	lbzx	$acc09,$Tbl2,$acc09
 	rlwinm	$s1,$acc01,24,0,7
 	lbzx	$acc10,$Tbl2,$acc10
-	lbzx	$acc11,$Tbl2,$acc11
 	rlwinm	$s2,$acc02,24,0,7
+	lbzx	$acc11,$Tbl2,$acc11
 	rlwinm	$s3,$acc03,24,0,7
 	lbzx	$acc12,$Tbl2,$acc12
-	lbzx	$acc13,$Tbl2,$acc13
 	rlwimi	$s0,$acc04,16,8,15
+	lbzx	$acc13,$Tbl2,$acc13
 	rlwimi	$s1,$acc05,16,8,15
 	lbzx	$acc14,$Tbl2,$acc14
-	lbzx	$acc15,$Tbl2,$acc15
 	rlwimi	$s2,$acc06,16,8,15
+	lbzx	$acc15,$Tbl2,$acc15
 	rlwimi	$s3,$acc07,16,8,15
 	rlwimi	$s0,$acc08,8,16,23
 	rlwimi	$s1,$acc09,8,16,23
@@ -879,20 +1051,22 @@ Ldec_loop:
 	xor	$s2,$s2,$t2
 	xor	$s3,$s3,$t3
 	blr
+	.long	0
+	.byte	0,12,0x14,0,0,0,0,0
 
 .align	4
 Lppc_AES_decrypt_compact:
 	lwz	$acc00,240($key)
-	lwz	$t0,0($key)
-	lwz	$t1,4($key)
-	lwz	$t2,8($key)
-	lwz	$t3,12($key)
 	addi	$Tbl1,$Tbl0,2048
+	lwz	$t0,0($key)
 	lis	$mask80,0x8080
+	lwz	$t1,4($key)
 	lis	$mask1b,0x1b1b
-	addi	$key,$key,16
+	lwz	$t2,8($key)
 	ori	$mask80,$mask80,0x8080
+	lwz	$t3,12($key)
 	ori	$mask1b,$mask1b,0x1b1b
+	addi	$key,$key,16
 ___
 $code.=<<___ if ($SIZE_T==8);
 	insrdi	$mask80,$mask80,32,0
@@ -904,10 +1078,10 @@ $code.=<<___;
 Ldec_compact_loop:
 	xor	$s0,$s0,$t0
 	xor	$s1,$s1,$t1
-	xor	$s2,$s2,$t2
-	xor	$s3,$s3,$t3
 	rlwinm	$acc00,$s0,`32-24`,24,31
+	xor	$s2,$s2,$t2
 	rlwinm	$acc01,$s1,`32-24`,24,31
+	xor	$s3,$s3,$t3
 	rlwinm	$acc02,$s2,`32-24`,24,31
 	rlwinm	$acc03,$s3,`32-24`,24,31
 	rlwinm	$acc04,$s3,`32-16`,24,31
@@ -915,48 +1089,48 @@ Ldec_compact_loop:
 	rlwinm	$acc06,$s1,`32-16`,24,31
 	rlwinm	$acc07,$s2,`32-16`,24,31
 	lbzx	$acc00,$Tbl1,$acc00
-	lbzx	$acc01,$Tbl1,$acc01
 	rlwinm	$acc08,$s2,`32-8`,24,31
+	lbzx	$acc01,$Tbl1,$acc01
 	rlwinm	$acc09,$s3,`32-8`,24,31
 	lbzx	$acc02,$Tbl1,$acc02
-	lbzx	$acc03,$Tbl1,$acc03
 	rlwinm	$acc10,$s0,`32-8`,24,31
+	lbzx	$acc03,$Tbl1,$acc03
 	rlwinm	$acc11,$s1,`32-8`,24,31
 	lbzx	$acc04,$Tbl1,$acc04
-	lbzx	$acc05,$Tbl1,$acc05
 	rlwinm	$acc12,$s1,`0`,24,31
+	lbzx	$acc05,$Tbl1,$acc05
 	rlwinm	$acc13,$s2,`0`,24,31
 	lbzx	$acc06,$Tbl1,$acc06
-	lbzx	$acc07,$Tbl1,$acc07
 	rlwinm	$acc14,$s3,`0`,24,31
+	lbzx	$acc07,$Tbl1,$acc07
 	rlwinm	$acc15,$s0,`0`,24,31
 	lbzx	$acc08,$Tbl1,$acc08
-	lbzx	$acc09,$Tbl1,$acc09
 	rlwinm	$s0,$acc00,24,0,7
+	lbzx	$acc09,$Tbl1,$acc09
 	rlwinm	$s1,$acc01,24,0,7
 	lbzx	$acc10,$Tbl1,$acc10
-	lbzx	$acc11,$Tbl1,$acc11
 	rlwinm	$s2,$acc02,24,0,7
+	lbzx	$acc11,$Tbl1,$acc11
 	rlwinm	$s3,$acc03,24,0,7
 	lbzx	$acc12,$Tbl1,$acc12
-	lbzx	$acc13,$Tbl1,$acc13
 	rlwimi	$s0,$acc04,16,8,15
+	lbzx	$acc13,$Tbl1,$acc13
 	rlwimi	$s1,$acc05,16,8,15
 	lbzx	$acc14,$Tbl1,$acc14
-	lbzx	$acc15,$Tbl1,$acc15
 	rlwimi	$s2,$acc06,16,8,15
+	lbzx	$acc15,$Tbl1,$acc15
 	rlwimi	$s3,$acc07,16,8,15
 	rlwimi	$s0,$acc08,8,16,23
 	rlwimi	$s1,$acc09,8,16,23
 	rlwimi	$s2,$acc10,8,16,23
 	rlwimi	$s3,$acc11,8,16,23
 	lwz	$t0,0($key)
-	lwz	$t1,4($key)
 	or	$s0,$s0,$acc12
+	lwz	$t1,4($key)
 	or	$s1,$s1,$acc13
 	lwz	$t2,8($key)
-	lwz	$t3,12($key)
 	or	$s2,$s2,$acc14
+	lwz	$t3,12($key)
 	or	$s3,$s3,$acc15
 
 	addi	$key,$key,16
@@ -1030,12 +1204,12 @@ $code.=<<___ if ($SIZE_T==4);
 	and	$acc02,$s2,$mask80
 	and	$acc03,$s3,$mask80
 	srwi	$acc04,$acc00,7		# r1>>7
-	srwi	$acc05,$acc01,7
-	srwi	$acc06,$acc02,7
-	srwi	$acc07,$acc03,7
 	andc	$acc08,$s0,$mask80	# r0&0x7f7f7f7f
+	srwi	$acc05,$acc01,7
 	andc	$acc09,$s1,$mask80
+	srwi	$acc06,$acc02,7
 	andc	$acc10,$s2,$mask80
+	srwi	$acc07,$acc03,7
 	andc	$acc11,$s3,$mask80
 	sub	$acc00,$acc00,$acc04	# r1-(r1>>7)
 	sub	$acc01,$acc01,$acc05
@@ -1059,12 +1233,12 @@ $code.=<<___ if ($SIZE_T==4);
 	and	$acc06,$acc02,$mask80
 	and	$acc07,$acc03,$mask80
 	srwi	$acc08,$acc04,7		# r1>>7
-	srwi	$acc09,$acc05,7
-	srwi	$acc10,$acc06,7
-	srwi	$acc11,$acc07,7
 	andc	$acc12,$acc00,$mask80	# r2&0x7f7f7f7f
+	srwi	$acc09,$acc05,7
 	andc	$acc13,$acc01,$mask80
+	srwi	$acc10,$acc06,7
 	andc	$acc14,$acc02,$mask80
+	srwi	$acc11,$acc07,7
 	andc	$acc15,$acc03,$mask80
 	sub	$acc04,$acc04,$acc08	# r1-(r1>>7)
 	sub	$acc05,$acc05,$acc09
@@ -1085,13 +1259,13 @@ $code.=<<___ if ($SIZE_T==4);
 
 	and	$acc08,$acc04,$mask80	# r1=r4&0x80808080
 	and	$acc09,$acc05,$mask80
-	and	$acc10,$acc06,$mask80
-	and	$acc11,$acc07,$mask80
 	srwi	$acc12,$acc08,7		# r1>>7
+	and	$acc10,$acc06,$mask80
 	srwi	$acc13,$acc09,7
+	and	$acc11,$acc07,$mask80
 	srwi	$acc14,$acc10,7
-	srwi	$acc15,$acc11,7
 	sub	$acc08,$acc08,$acc12	# r1-(r1>>7)
+	srwi	$acc15,$acc11,7
 	sub	$acc09,$acc09,$acc13
 	sub	$acc10,$acc10,$acc14
 	sub	$acc11,$acc11,$acc15
@@ -1124,10 +1298,10 @@ ___
 $code.=<<___;
 	rotrwi	$s0,$s0,8		# = ROTATE(r0,8)
 	rotrwi	$s1,$s1,8
-	rotrwi	$s2,$s2,8
-	rotrwi	$s3,$s3,8
 	xor	$s0,$s0,$acc00		# ^= r2^r0
+	rotrwi	$s2,$s2,8
 	xor	$s1,$s1,$acc01
+	rotrwi	$s3,$s3,8
 	xor	$s2,$s2,$acc02
 	xor	$s3,$s3,$acc03
 	xor	$acc00,$acc00,$acc08
@@ -1135,32 +1309,32 @@ $code.=<<___;
 	xor	$acc02,$acc02,$acc10
 	xor	$acc03,$acc03,$acc11
 	xor	$s0,$s0,$acc04		# ^= r4^r0
-	xor	$s1,$s1,$acc05
-	xor	$s2,$s2,$acc06
-	xor	$s3,$s3,$acc07
 	rotrwi	$acc00,$acc00,24
+	xor	$s1,$s1,$acc05
 	rotrwi	$acc01,$acc01,24
+	xor	$s2,$s2,$acc06
 	rotrwi	$acc02,$acc02,24
+	xor	$s3,$s3,$acc07
 	rotrwi	$acc03,$acc03,24
 	xor	$acc04,$acc04,$acc08
 	xor	$acc05,$acc05,$acc09
 	xor	$acc06,$acc06,$acc10
 	xor	$acc07,$acc07,$acc11
 	xor	$s0,$s0,$acc08		# ^= r8 [^((r4^r0)^(r2^r0)=r4^r2)]
-	xor	$s1,$s1,$acc09
-	xor	$s2,$s2,$acc10
-	xor	$s3,$s3,$acc11
 	rotrwi	$acc04,$acc04,16
+	xor	$s1,$s1,$acc09
 	rotrwi	$acc05,$acc05,16
+	xor	$s2,$s2,$acc10
 	rotrwi	$acc06,$acc06,16
+	xor	$s3,$s3,$acc11
 	rotrwi	$acc07,$acc07,16
 	xor	$s0,$s0,$acc00		# ^= ROTATE(r8^r2^r0,24)
-	xor	$s1,$s1,$acc01
-	xor	$s2,$s2,$acc02
-	xor	$s3,$s3,$acc03
 	rotrwi	$acc08,$acc08,8
+	xor	$s1,$s1,$acc01
 	rotrwi	$acc09,$acc09,8
+	xor	$s2,$s2,$acc02
 	rotrwi	$acc10,$acc10,8
+	xor	$s3,$s3,$acc03
 	rotrwi	$acc11,$acc11,8
 	xor	$s0,$s0,$acc04		# ^= ROTATE(r8^r4^r0,16)
 	xor	$s1,$s1,$acc05
@@ -1179,7 +1353,9 @@ Ldec_compact_done:
 	xor	$s2,$s2,$t2
 	xor	$s3,$s3,$t3
 	blr
-.long	0
+	.long	0
+	.byte	0,12,0x14,0,0,0,0,0
+
 .asciz	"AES for PPC, CRYPTOGAMS by <appro\@openssl.org>"
 .align	7
 ___