blob: d97df627d813b9eb582692275f15c7ccb232918b [file] [log] [blame]
diff --git a/Configure b/Configure
index de78469..26743bb 100755
--- a/Configure
+++ b/Configure
@@ -136,7 +136,8 @@ my $alpha_asm="alphacpuid.o:bn_asm.o alpha-mont.o:::::sha1-alpha.o:::::::ghash-a
my $mips32_asm=":bn-mips.o::aes_cbc.o aes-mips.o:::sha1-mips.o sha256-mips.o::::::::";
my $mips64_asm=":bn-mips.o mips-mont.o::aes_cbc.o aes-mips.o:::sha1-mips.o sha256-mips.o sha512-mips.o::::::::";
my $s390x_asm="s390xcap.o s390xcpuid.o:bn-s390x.o s390x-mont.o s390x-gf2m.o::aes-s390x.o aes-ctr.o aes-xts.o:::sha1-s390x.o sha256-s390x.o sha512-s390x.o::rc4-s390x.o:::::ghash-s390x.o:";
-my $armv4_asm="armcap.o armv4cpuid.o:bn_asm.o armv4-mont.o armv4-gf2m.o::aes_cbc.o aes-armv4.o:::sha1-armv4-large.o sha256-armv4.o sha512-armv4.o:::::::ghash-armv4.o::void";
+my $armv4_asm="armcap.o armv4cpuid.o:bn_asm.o armv4-mont.o armv4-gf2m.o::aes_cbc.o aes-armv4.o bsaes-armv7.o aesv8-armx.o:::sha1-armv4-large.o sha256-armv4.o sha512-armv4.o:::::::ghash-armv4.o ghashv8-armx.o::void";
+my $aarch64_asm="armcap.o arm64cpuid.o mem_clr.o:::aes_core.o aes_cbc.o aesv8-armx.o:::sha1-armv8.o sha256-armv8.o sha512-armv8.o:::::::ghashv8-armx.o:";
my $parisc11_asm="pariscid.o:bn_asm.o parisc-mont.o::aes_core.o aes_cbc.o aes-parisc.o:::sha1-parisc.o sha256-parisc.o sha512-parisc.o::rc4-parisc.o:::::ghash-parisc.o::32";
my $parisc20_asm="pariscid.o:pa-risc2W.o parisc-mont.o::aes_core.o aes_cbc.o aes-parisc.o:::sha1-parisc.o sha256-parisc.o sha512-parisc.o::rc4-parisc.o:::::ghash-parisc.o::64";
my $ppc32_asm="ppccpuid.o ppccap.o:bn-ppc.o ppc-mont.o ppc64-mont.o::aes_core.o aes_cbc.o aes-ppc.o:::sha1-ppc.o sha256-ppc.o::::::::";
@@ -350,6 +351,7 @@ my %table=(
# It's believed that majority of ARM toolchains predefine appropriate -march.
# If you compiler does not, do complement config command line with one!
"linux-armv4", "gcc:-DTERMIO -O3 -Wall::-D_REENTRANT::-ldl:BN_LLONG RC4_CHAR RC4_CHUNK DES_INT DES_UNROLL BF_PTR:${armv4_asm}:dlfcn:linux-shared:-fPIC::.so.\$(SHLIB_MAJOR).\$(SHLIB_MINOR)",
+"linux-aarch64","gcc:-DTERMIO -O3 -Wall::-D_REENTRANT::-ldl:SIXTY_FOUR_BIT_LONG RC4_CHAR RC4_CHUNK DES_INT DES_UNROLL BF_PTR:${aarch64_asm}:linux64:dlfcn:linux-shared:-fPIC::.so.\$(SHLIB_MAJOR).\$(SHLIB_MINOR)",
#### IA-32 targets...
"linux-ia32-icc", "icc:-DL_ENDIAN -DTERMIO -O2 -no_cpprt::-D_REENTRANT::-ldl:BN_LLONG ${x86_gcc_des} ${x86_gcc_opts}:${x86_elf_asm}:dlfcn:linux-shared:-KPIC::.so.\$(SHLIB_MAJOR).\$(SHLIB_MINOR)",
"linux-elf", "gcc:-DL_ENDIAN -DTERMIO -O3 -fomit-frame-pointer -Wall::-D_REENTRANT::-ldl:BN_LLONG ${x86_gcc_des} ${x86_gcc_opts}:${x86_elf_asm}:dlfcn:linux-shared:-fPIC::.so.\$(SHLIB_MAJOR).\$(SHLIB_MINOR)",
@@ -1503,7 +1505,7 @@ if ($rmd160_obj =~ /\.o$/)
}
if ($aes_obj =~ /\.o$/)
{
- $cflags.=" -DAES_ASM";
+ $cflags.=" -DAES_ASM" if ($aes_obj =~ m/\baes\-/);;
# aes-ctr.o is not a real file, only indication that assembler
# module implements AES_ctr32_encrypt...
$cflags.=" -DAES_CTR_ASM" if ($aes_obj =~ s/\s*aes\-ctr\.o//);
@@ -1525,7 +1527,7 @@ else {
$wp_obj="wp_block.o";
}
$cmll_obj=$cmll_enc unless ($cmll_obj =~ /.o$/);
-if ($modes_obj =~ /ghash/)
+if ($modes_obj =~ /ghash\-/)
{
$cflags.=" -DGHASH_ASM";
}
diff --git a/config b/config
index 41fa2a6..dff7df7 100755
--- a/config
+++ b/config
@@ -644,6 +644,7 @@ case "$GUESSOS" in
armv[1-3]*-*-linux2) OUT="linux-generic32" ;;
armv[7-9]*-*-linux2) OUT="linux-armv4"; options="$options -march=armv7-a" ;;
arm*-*-linux2) OUT="linux-armv4" ;;
+ aarch64-*-linux2) OUT="linux-aarch64" ;;
sh*b-*-linux2) OUT="linux-generic32"; options="$options -DB_ENDIAN" ;;
sh*-*-linux2) OUT="linux-generic32"; options="$options -DL_ENDIAN" ;;
m68k*-*-linux2) OUT="linux-generic32"; options="$options -DB_ENDIAN" ;;
diff --git a/crypto/aes/Makefile b/crypto/aes/Makefile
index 45ede0a..9181a1a 100644
--- a/crypto/aes/Makefile
+++ b/crypto/aes/Makefile
@@ -78,9 +78,15 @@ aes-parisc.s: asm/aes-parisc.pl
aes-mips.S: asm/aes-mips.pl
$(PERL) asm/aes-mips.pl $(PERLASM_SCHEME) $@
+aesv8-armx.S: asm/aesv8-armx.pl
+ $(PERL) asm/aesv8-armx.pl $(PERLASM_SCHEME) $@
+aesv8-armx.o: aesv8-armx.S
+
# GNU make "catch all"
aes-%.S: asm/aes-%.pl; $(PERL) $< $(PERLASM_SCHEME) > $@
aes-armv4.o: aes-armv4.S
+bsaes-%.S: asm/bsaes-%.pl; $(PERL) $< $(PERLASM_SCHEME) $@
+bsaes-armv7.o: bsaes-armv7.S
files:
$(PERL) $(TOP)/util/files.pl Makefile >> $(TOP)/MINFO
diff --git a/crypto/aes/asm/aes-armv4.pl b/crypto/aes/asm/aes-armv4.pl
index 86b86c4..4f89170 100644
--- a/crypto/aes/asm/aes-armv4.pl
+++ b/crypto/aes/asm/aes-armv4.pl
@@ -1,7 +1,7 @@
#!/usr/bin/env perl
# ====================================================================
-# Written by Andy Polyakov <appro@fy.chalmers.se> for the OpenSSL
+# Written by Andy Polyakov <appro@openssl.org> for the OpenSSL
# project. The module is, however, dual licensed under OpenSSL and
# CRYPTOGAMS licenses depending on where you obtain it. For further
# details see http://www.openssl.org/~appro/cryptogams/.
@@ -51,9 +51,23 @@ $key="r11";
$rounds="r12";
$code=<<___;
-#include "arm_arch.h"
+#ifndef __KERNEL__
+# include "arm_arch.h"
+#else
+# define __ARM_ARCH__ __LINUX_ARM_ARCH__
+#endif
+
.text
+#if __ARM_ARCH__<7
+.code 32
+#else
+.syntax unified
+# ifdef __thumb2__
+.thumb
+# else
.code 32
+# endif
+#endif
.type AES_Te,%object
.align 5
@@ -167,7 +181,11 @@ AES_Te:
.type AES_encrypt,%function
.align 5
AES_encrypt:
+#if __ARM_ARCH__<7
sub r3,pc,#8 @ AES_encrypt
+#else
+ adr r3,AES_encrypt
+#endif
stmdb sp!,{r1,r4-r12,lr}
mov $rounds,r0 @ inp
mov $key,r2
@@ -409,11 +427,21 @@ _armv4_AES_encrypt:
.align 5
private_AES_set_encrypt_key:
_armv4_AES_set_encrypt_key:
+#if __ARM_ARCH__<7
sub r3,pc,#8 @ AES_set_encrypt_key
+#else
+ adr r3,private_AES_set_encrypt_key
+#endif
teq r0,#0
+#if __ARM_ARCH__>=7
+ itt eq @ Thumb2 thing, sanity check in ARM
+#endif
moveq r0,#-1
beq .Labrt
teq r2,#0
+#if __ARM_ARCH__>=7
+ itt eq @ Thumb2 thing, sanity check in ARM
+#endif
moveq r0,#-1
beq .Labrt
@@ -422,6 +450,9 @@ _armv4_AES_set_encrypt_key:
teq r1,#192
beq .Lok
teq r1,#256
+#if __ARM_ARCH__>=7
+ itt ne @ Thumb2 thing, sanity check in ARM
+#endif
movne r0,#-1
bne .Labrt
@@ -576,6 +607,9 @@ _armv4_AES_set_encrypt_key:
str $s2,[$key,#-16]
subs $rounds,$rounds,#1
str $s3,[$key,#-12]
+#if __ARM_ARCH__>=7
+ itt eq @ Thumb2 thing, sanity check in ARM
+#endif
subeq r2,$key,#216
beq .Ldone
@@ -645,6 +679,9 @@ _armv4_AES_set_encrypt_key:
str $s2,[$key,#-24]
subs $rounds,$rounds,#1
str $s3,[$key,#-20]
+#if __ARM_ARCH__>=7
+ itt eq @ Thumb2 thing, sanity check in ARM
+#endif
subeq r2,$key,#256
beq .Ldone
@@ -674,11 +711,17 @@ _armv4_AES_set_encrypt_key:
str $i3,[$key,#-4]
b .L256_loop
+.align 2
.Ldone: mov r0,#0
ldmia sp!,{r4-r12,lr}
-.Labrt: tst lr,#1
+.Labrt:
+#if __ARM_ARCH__>=5
+ ret @ bx lr
+#else
+ tst lr,#1
moveq pc,lr @ be binary compatible with V4, yet
bx lr @ interoperable with Thumb ISA:-)
+#endif
.size private_AES_set_encrypt_key,.-private_AES_set_encrypt_key
.global private_AES_set_decrypt_key
@@ -688,34 +731,57 @@ private_AES_set_decrypt_key:
str lr,[sp,#-4]! @ push lr
bl _armv4_AES_set_encrypt_key
teq r0,#0
- ldrne lr,[sp],#4 @ pop lr
+ ldr lr,[sp],#4 @ pop lr
bne .Labrt
- stmdb sp!,{r4-r12}
+ mov r0,r2 @ AES_set_encrypt_key preserves r2,
+ mov r1,r2 @ which is AES_KEY *key
+ b _armv4_AES_set_enc2dec_key
+.size private_AES_set_decrypt_key,.-private_AES_set_decrypt_key
- ldr $rounds,[r2,#240] @ AES_set_encrypt_key preserves r2,
- mov $key,r2 @ which is AES_KEY *key
- mov $i1,r2
- add $i2,r2,$rounds,lsl#4
+@ void AES_set_enc2dec_key(const AES_KEY *inp,AES_KEY *out)
+.global AES_set_enc2dec_key
+.type AES_set_enc2dec_key,%function
+.align 5
+AES_set_enc2dec_key:
+_armv4_AES_set_enc2dec_key:
+ stmdb sp!,{r4-r12,lr}
+
+ ldr $rounds,[r0,#240]
+ mov $i1,r0 @ input
+ add $i2,r0,$rounds,lsl#4
+ mov $key,r1 @ ouput
+ add $tbl,r1,$rounds,lsl#4
+ str $rounds,[r1,#240]
+
+.Linv: ldr $s0,[$i1],#16
+ ldr $s1,[$i1,#-12]
+ ldr $s2,[$i1,#-8]
+ ldr $s3,[$i1,#-4]
+ ldr $t1,[$i2],#-16
+ ldr $t2,[$i2,#16+4]
+ ldr $t3,[$i2,#16+8]
+ ldr $i3,[$i2,#16+12]
+ str $s0,[$tbl],#-16
+ str $s1,[$tbl,#16+4]
+ str $s2,[$tbl,#16+8]
+ str $s3,[$tbl,#16+12]
+ str $t1,[$key],#16
+ str $t2,[$key,#-12]
+ str $t3,[$key,#-8]
+ str $i3,[$key,#-4]
+ teq $i1,$i2
+ bne .Linv
-.Linv: ldr $s0,[$i1]
+ ldr $s0,[$i1]
ldr $s1,[$i1,#4]
ldr $s2,[$i1,#8]
ldr $s3,[$i1,#12]
- ldr $t1,[$i2]
- ldr $t2,[$i2,#4]
- ldr $t3,[$i2,#8]
- ldr $i3,[$i2,#12]
- str $s0,[$i2],#-16
- str $s1,[$i2,#16+4]
- str $s2,[$i2,#16+8]
- str $s3,[$i2,#16+12]
- str $t1,[$i1],#16
- str $t2,[$i1,#-12]
- str $t3,[$i1,#-8]
- str $i3,[$i1,#-4]
- teq $i1,$i2
- bne .Linv
+ str $s0,[$key]
+ str $s1,[$key,#4]
+ str $s2,[$key,#8]
+ str $s3,[$key,#12]
+ sub $key,$key,$rounds,lsl#3
___
$mask80=$i1;
$mask1b=$i2;
@@ -773,7 +839,7 @@ $code.=<<___;
moveq pc,lr @ be binary compatible with V4, yet
bx lr @ interoperable with Thumb ISA:-)
#endif
-.size private_AES_set_decrypt_key,.-private_AES_set_decrypt_key
+.size AES_set_enc2dec_key,.-AES_set_enc2dec_key
.type AES_Td,%object
.align 5
@@ -883,7 +949,11 @@ AES_Td:
.type AES_decrypt,%function
.align 5
AES_decrypt:
+#if __ARM_ARCH__<7
sub r3,pc,#8 @ AES_decrypt
+#else
+ adr r3,AES_decrypt
+#endif
stmdb sp!,{r1,r4-r12,lr}
mov $rounds,r0 @ inp
mov $key,r2
@@ -1080,8 +1150,9 @@ _armv4_AES_decrypt:
ldrb $t3,[$tbl,$i3] @ Td4[s0>>0]
and $i3,lr,$s1,lsr#8
+ add $s1,$tbl,$s1,lsr#24
ldrb $i1,[$tbl,$i1] @ Td4[s1>>0]
- ldrb $s1,[$tbl,$s1,lsr#24] @ Td4[s1>>24]
+ ldrb $s1,[$s1] @ Td4[s1>>24]
ldrb $i2,[$tbl,$i2] @ Td4[s1>>16]
eor $s0,$i1,$s0,lsl#24
ldrb $i3,[$tbl,$i3] @ Td4[s1>>8]
@@ -1094,7 +1165,8 @@ _armv4_AES_decrypt:
ldrb $i2,[$tbl,$i2] @ Td4[s2>>0]
and $i3,lr,$s2,lsr#16
- ldrb $s2,[$tbl,$s2,lsr#24] @ Td4[s2>>24]
+ add $s2,$tbl,$s2,lsr#24
+ ldrb $s2,[$s2] @ Td4[s2>>24]
eor $s0,$s0,$i1,lsl#8
ldrb $i3,[$tbl,$i3] @ Td4[s2>>16]
eor $s1,$i2,$s1,lsl#16
@@ -1106,8 +1178,9 @@ _armv4_AES_decrypt:
ldrb $i2,[$tbl,$i2] @ Td4[s3>>8]
and $i3,lr,$s3 @ i2
+ add $s3,$tbl,$s3,lsr#24
ldrb $i3,[$tbl,$i3] @ Td4[s3>>0]
- ldrb $s3,[$tbl,$s3,lsr#24] @ Td4[s3>>24]
+ ldrb $s3,[$s3] @ Td4[s3>>24]
eor $s0,$s0,$i1,lsl#16
ldr $i1,[$key,#0]
eor $s1,$s1,$i2,lsl#8
@@ -1130,5 +1203,15 @@ _armv4_AES_decrypt:
___
$code =~ s/\bbx\s+lr\b/.word\t0xe12fff1e/gm; # make it possible to compile with -march=armv4
+$code =~ s/\bret\b/bx\tlr/gm;
+
+open SELF,$0;
+while(<SELF>) {
+ next if (/^#!/);
+ last if (!s/^#/@/ and !/^$/);
+ print;
+}
+close SELF;
+
print $code;
close STDOUT; # enforce flush
diff --git a/crypto/aes/asm/aesv8-armx.pl b/crypto/aes/asm/aesv8-armx.pl
new file mode 100755
index 0000000..415dc04
--- /dev/null
+++ b/crypto/aes/asm/aesv8-armx.pl
@@ -0,0 +1,980 @@
+#!/usr/bin/env perl
+#
+# ====================================================================
+# Written by Andy Polyakov <appro@openssl.org> for the OpenSSL
+# project. The module is, however, dual licensed under OpenSSL and
+# CRYPTOGAMS licenses depending on where you obtain it. For further
+# details see http://www.openssl.org/~appro/cryptogams/.
+# ====================================================================
+#
+# This module implements support for ARMv8 AES instructions. The
+# module is endian-agnostic in sense that it supports both big- and
+# little-endian cases. As does it support both 32- and 64-bit modes
+# of operation. Latter is achieved by limiting amount of utilized
+# registers to 16, which implies additional instructions. This has
+# no effect on mighty Apple A7, as results are literally equal to
+# the theoretical estimates based on instruction latencies and issue
+# rate. It remains to be seen how does it affect other platforms...
+#
+# Performance in cycles per byte processed with 128-bit key:
+#
+# CBC enc CBC dec CTR
+# Apple A7 2.39 1.20 1.20
+# Cortex-A5x n/a n/a n/a
+
+$flavour = shift;
+open STDOUT,">".shift;
+
+$prefix="aes_v8";
+
+$code=<<___;
+#include "arm_arch.h"
+
+#if __ARM_ARCH__>=7
+.text
+___
+$code.=".arch armv8-a+crypto\n" if ($flavour =~ /64/);
+$code.=".fpu neon\n.code 32\n" if ($flavour !~ /64/);
+
+# Assembler mnemonics are an eclectic mix of 32- and 64-bit syntax,
+# NEON is mostly 32-bit mnemonics, integer - mostly 64. Goal is to
+# maintain both 32- and 64-bit codes within single module and
+# transliterate common code to either flavour with regex vodoo.
+#
+{{{
+my ($inp,$bits,$out,$ptr,$rounds)=("x0","w1","x2","x3","w12");
+my ($zero,$rcon,$mask,$in0,$in1,$tmp,$key)=
+ $flavour=~/64/? map("q$_",(0..6)) : map("q$_",(0..3,8..10));
+
+
+$code.=<<___;
+.align 5
+rcon:
+.long 0x01,0x01,0x01,0x01
+.long 0x0c0f0e0d,0x0c0f0e0d,0x0c0f0e0d,0x0c0f0e0d // rotate-n-splat
+.long 0x1b,0x1b,0x1b,0x1b
+
+.globl ${prefix}_set_encrypt_key
+.type ${prefix}_set_encrypt_key,%function
+.align 5
+${prefix}_set_encrypt_key:
+.Lenc_key:
+___
+$code.=<<___ if ($flavour =~ /64/);
+ stp x29,x30,[sp,#-16]!
+ add x29,sp,#0
+___
+$code.=<<___;
+ adr $ptr,rcon
+ cmp $bits,#192
+
+ veor $zero,$zero,$zero
+ vld1.8 {$in0},[$inp],#16
+ mov $bits,#8 // reuse $bits
+ vld1.32 {$rcon,$mask},[$ptr],#32
+
+ b.lt .Loop128
+ b.eq .L192
+ b .L256
+
+.align 4
+.Loop128:
+ vtbl.8 $key,{$in0},$mask
+ vext.8 $tmp,$zero,$in0,#12
+ vst1.32 {$in0},[$out],#16
+ aese $key,$zero
+ subs $bits,$bits,#1
+
+ veor $in0,$in0,$tmp
+ vext.8 $tmp,$zero,$tmp,#12
+ veor $in0,$in0,$tmp
+ vext.8 $tmp,$zero,$tmp,#12
+ veor $key,$key,$rcon
+ veor $in0,$in0,$tmp
+ vshl.u8 $rcon,$rcon,#1
+ veor $in0,$in0,$key
+ b.ne .Loop128
+
+ vld1.32 {$rcon},[$ptr]
+
+ vtbl.8 $key,{$in0},$mask
+ vext.8 $tmp,$zero,$in0,#12
+ vst1.32 {$in0},[$out],#16
+ aese $key,$zero
+
+ veor $in0,$in0,$tmp
+ vext.8 $tmp,$zero,$tmp,#12
+ veor $in0,$in0,$tmp
+ vext.8 $tmp,$zero,$tmp,#12
+ veor $key,$key,$rcon
+ veor $in0,$in0,$tmp
+ vshl.u8 $rcon,$rcon,#1
+ veor $in0,$in0,$key
+
+ vtbl.8 $key,{$in0},$mask
+ vext.8 $tmp,$zero,$in0,#12
+ vst1.32 {$in0},[$out],#16
+ aese $key,$zero
+
+ veor $in0,$in0,$tmp
+ vext.8 $tmp,$zero,$tmp,#12
+ veor $in0,$in0,$tmp
+ vext.8 $tmp,$zero,$tmp,#12
+ veor $key,$key,$rcon
+ veor $in0,$in0,$tmp
+ veor $in0,$in0,$key
+ vst1.32 {$in0},[$out]
+ add $out,$out,#0x50
+
+ mov $rounds,#10
+ b .Ldone
+
+.align 4
+.L192:
+ vld1.8 {$in1},[$inp],#8
+ vmov.i8 $key,#8 // borrow $key
+ vst1.32 {$in0},[$out],#16
+ vsub.i8 $mask,$mask,$key // adjust the mask
+
+.Loop192:
+ vtbl.8 $key,{$in1},$mask
+ vext.8 $tmp,$zero,$in0,#12
+ vst1.32 {$in1},[$out],#8
+ aese $key,$zero
+ subs $bits,$bits,#1
+
+ veor $in0,$in0,$tmp
+ vext.8 $tmp,$zero,$tmp,#12
+ veor $in0,$in0,$tmp
+ vext.8 $tmp,$zero,$tmp,#12
+ veor $in0,$in0,$tmp
+
+ vdup.32 $tmp,${in0}[3]
+ veor $tmp,$tmp,$in1
+ veor $key,$key,$rcon
+ vext.8 $in1,$zero,$in1,#12
+ vshl.u8 $rcon,$rcon,#1
+ veor $in1,$in1,$tmp
+ veor $in0,$in0,$key
+ veor $in1,$in1,$key
+ vst1.32 {$in0},[$out],#16
+ b.ne .Loop192
+
+ mov $rounds,#12
+ add $out,$out,#0x20
+ b .Ldone
+
+.align 4
+.L256:
+ vld1.8 {$in1},[$inp]
+ mov $bits,#7
+ mov $rounds,#14
+ vst1.32 {$in0},[$out],#16
+
+.Loop256:
+ vtbl.8 $key,{$in1},$mask
+ vext.8 $tmp,$zero,$in0,#12
+ vst1.32 {$in1},[$out],#16
+ aese $key,$zero
+ subs $bits,$bits,#1
+
+ veor $in0,$in0,$tmp
+ vext.8 $tmp,$zero,$tmp,#12
+ veor $in0,$in0,$tmp
+ vext.8 $tmp,$zero,$tmp,#12
+ veor $key,$key,$rcon
+ veor $in0,$in0,$tmp
+ vshl.u8 $rcon,$rcon,#1
+ veor $in0,$in0,$key
+ vst1.32 {$in0},[$out],#16
+ b.eq .Ldone
+
+ vdup.32 $key,${in0}[3] // just splat
+ vext.8 $tmp,$zero,$in1,#12
+ aese $key,$zero
+
+ veor $in1,$in1,$tmp
+ vext.8 $tmp,$zero,$tmp,#12
+ veor $in1,$in1,$tmp
+ vext.8 $tmp,$zero,$tmp,#12
+ veor $in1,$in1,$tmp
+
+ veor $in1,$in1,$key
+ b .Loop256
+
+.Ldone:
+ str $rounds,[$out]
+
+ eor x0,x0,x0 // return value
+ `"ldr x29,[sp],#16" if ($flavour =~ /64/)`
+ ret
+.size ${prefix}_set_encrypt_key,.-${prefix}_set_encrypt_key
+
+.globl ${prefix}_set_decrypt_key
+.type ${prefix}_set_decrypt_key,%function
+.align 5
+${prefix}_set_decrypt_key:
+___
+$code.=<<___ if ($flavour =~ /64/);
+ stp x29,x30,[sp,#-16]!
+ add x29,sp,#0
+___
+$code.=<<___ if ($flavour !~ /64/);
+ stmdb sp!,{r4,lr}
+___
+$code.=<<___;
+ bl .Lenc_key
+
+ sub $out,$out,#240 // restore original $out
+ mov x4,#-16
+ add $inp,$out,x12,lsl#4 // end of key schedule
+
+ vld1.32 {v0.16b},[$out]
+ vld1.32 {v1.16b},[$inp]
+ vst1.32 {v0.16b},[$inp],x4
+ vst1.32 {v1.16b},[$out],#16
+
+.Loop_imc:
+ vld1.32 {v0.16b},[$out]
+ vld1.32 {v1.16b},[$inp]
+ aesimc v0.16b,v0.16b
+ aesimc v1.16b,v1.16b
+ vst1.32 {v0.16b},[$inp],x4
+ vst1.32 {v1.16b},[$out],#16
+ cmp $inp,$out
+ b.hi .Loop_imc
+
+ vld1.32 {v0.16b},[$out]
+ aesimc v0.16b,v0.16b
+ vst1.32 {v0.16b},[$inp]
+
+ eor x0,x0,x0 // return value
+___
+$code.=<<___ if ($flavour !~ /64/);
+ ldmia sp!,{r4,pc}
+___
+$code.=<<___ if ($flavour =~ /64/);
+ ldp x29,x30,[sp],#16
+ ret
+___
+$code.=<<___;
+.size ${prefix}_set_decrypt_key,.-${prefix}_set_decrypt_key
+___
+}}}
+{{{
+sub gen_block () {
+my $dir = shift;
+my ($e,$mc) = $dir eq "en" ? ("e","mc") : ("d","imc");
+my ($inp,$out,$key)=map("x$_",(0..2));
+my $rounds="w3";
+my ($rndkey0,$rndkey1,$inout)=map("q$_",(0..3));
+
+$code.=<<___;
+.globl ${prefix}_${dir}crypt
+.type ${prefix}_${dir}crypt,%function
+.align 5
+${prefix}_${dir}crypt:
+ ldr $rounds,[$key,#240]
+ vld1.32 {$rndkey0},[$key],#16
+ vld1.8 {$inout},[$inp]
+ sub $rounds,$rounds,#2
+ vld1.32 {$rndkey1},[$key],#16
+
+.Loop_${dir}c:
+ aes$e $inout,$rndkey0
+ vld1.32 {$rndkey0},[$key],#16
+ aes$mc $inout,$inout
+ subs $rounds,$rounds,#2
+ aes$e $inout,$rndkey1
+ vld1.32 {$rndkey1},[$key],#16
+ aes$mc $inout,$inout
+ b.gt .Loop_${dir}c
+
+ aes$e $inout,$rndkey0
+ vld1.32 {$rndkey0},[$key]
+ aes$mc $inout,$inout
+ aes$e $inout,$rndkey1
+ veor $inout,$inout,$rndkey0
+
+ vst1.8 {$inout},[$out]
+ ret
+.size ${prefix}_${dir}crypt,.-${prefix}_${dir}crypt
+___
+}
+&gen_block("en");
+&gen_block("de");
+}}}
+{{{
+my ($inp,$out,$len,$key,$ivp)=map("x$_",(0..4)); my $enc="w5";
+my ($rounds,$cnt,$key_,$step,$step1)=($enc,"w6","x7","x8","x12");
+my ($dat0,$dat1,$in0,$in1,$tmp0,$tmp1,$ivec,$rndlast)=map("q$_",(0..7));
+
+my ($dat,$tmp,$rndzero_n_last)=($dat0,$tmp0,$tmp1);
+
+### q8-q15 preloaded key schedule
+
+$code.=<<___;
+.globl ${prefix}_cbc_encrypt
+.type ${prefix}_cbc_encrypt,%function
+.align 5
+${prefix}_cbc_encrypt:
+___
+$code.=<<___ if ($flavour =~ /64/);
+ stp x29,x30,[sp,#-16]!
+ add x29,sp,#0
+___
+$code.=<<___ if ($flavour !~ /64/);
+ mov ip,sp
+ stmdb sp!,{r4-r8,lr}
+ vstmdb sp!,{d8-d15} @ ABI specification says so
+ ldmia ip,{r4-r5} @ load remaining args
+___
+$code.=<<___;
+ subs $len,$len,#16
+ mov $step,#16
+ b.lo .Lcbc_abort
+ cclr $step,eq
+
+ cmp $enc,#0 // en- or decrypting?
+ ldr $rounds,[$key,#240]
+ and $len,$len,#-16
+ vld1.8 {$ivec},[$ivp]
+ vld1.8 {$dat},[$inp],$step
+
+ vld1.32 {q8-q9},[$key] // load key schedule...
+ sub $rounds,$rounds,#6
+ add $key_,$key,x5,lsl#4 // pointer to last 7 round keys
+ sub $rounds,$rounds,#2
+ vld1.32 {q10-q11},[$key_],#32
+ vld1.32 {q12-q13},[$key_],#32
+ vld1.32 {q14-q15},[$key_],#32
+ vld1.32 {$rndlast},[$key_]
+
+ add $key_,$key,#32
+ mov $cnt,$rounds
+ b.eq .Lcbc_dec
+
+ cmp $rounds,#2
+ veor $dat,$dat,$ivec
+ veor $rndzero_n_last,q8,$rndlast
+ b.eq .Lcbc_enc128
+
+.Loop_cbc_enc:
+ aese $dat,q8
+ vld1.32 {q8},[$key_],#16
+ aesmc $dat,$dat
+ subs $cnt,$cnt,#2
+ aese $dat,q9
+ vld1.32 {q9},[$key_],#16
+ aesmc $dat,$dat
+ b.gt .Loop_cbc_enc
+
+ aese $dat,q8
+ aesmc $dat,$dat
+ subs $len,$len,#16
+ aese $dat,q9
+ aesmc $dat,$dat
+ cclr $step,eq
+ aese $dat,q10
+ aesmc $dat,$dat
+ add $key_,$key,#16
+ aese $dat,q11
+ aesmc $dat,$dat
+ vld1.8 {q8},[$inp],$step
+ aese $dat,q12
+ aesmc $dat,$dat
+ veor q8,q8,$rndzero_n_last
+ aese $dat,q13
+ aesmc $dat,$dat
+ vld1.32 {q9},[$key_],#16 // re-pre-load rndkey[1]
+ aese $dat,q14
+ aesmc $dat,$dat
+ aese $dat,q15
+
+ mov $cnt,$rounds
+ veor $ivec,$dat,$rndlast
+ vst1.8 {$ivec},[$out],#16
+ b.hs .Loop_cbc_enc
+
+ b .Lcbc_done
+
+.align 5
+.Lcbc_enc128:
+ vld1.32 {$in0-$in1},[$key_]
+ aese $dat,q8
+ aesmc $dat,$dat
+ b .Lenter_cbc_enc128
+.Loop_cbc_enc128:
+ aese $dat,q8
+ aesmc $dat,$dat
+ vst1.8 {$ivec},[$out],#16
+.Lenter_cbc_enc128:
+ aese $dat,q9
+ aesmc $dat,$dat
+ subs $len,$len,#16
+ aese $dat,$in0
+ aesmc $dat,$dat
+ cclr $step,eq
+ aese $dat,$in1
+ aesmc $dat,$dat
+ aese $dat,q10
+ aesmc $dat,$dat
+ aese $dat,q11
+ aesmc $dat,$dat
+ vld1.8 {q8},[$inp],$step
+ aese $dat,q12
+ aesmc $dat,$dat
+ aese $dat,q13
+ aesmc $dat,$dat
+ aese $dat,q14
+ aesmc $dat,$dat
+ veor q8,q8,$rndzero_n_last
+ aese $dat,q15
+ veor $ivec,$dat,$rndlast
+ b.hs .Loop_cbc_enc128
+
+ vst1.8 {$ivec},[$out],#16
+ b .Lcbc_done
+
+.align 5
+.Lcbc_dec128:
+ vld1.32 {$tmp0-$tmp1},[$key_]
+ veor $ivec,$ivec,$rndlast
+ veor $in0,$dat0,$rndlast
+ mov $step1,$step
+
+.Loop2x_cbc_dec128:
+ aesd $dat0,q8
+ aesd $dat1,q8
+ aesimc $dat0,$dat0
+ aesimc $dat1,$dat1
+ subs $len,$len,#32
+ aesd $dat0,q9
+ aesd $dat1,q9
+ aesimc $dat0,$dat0
+ aesimc $dat1,$dat1
+ cclr $step,lo
+ aesd $dat0,$tmp0
+ aesd $dat1,$tmp0
+ aesimc $dat0,$dat0
+ aesimc $dat1,$dat1
+ cclr $step1,ls
+ aesd $dat0,$tmp1
+ aesd $dat1,$tmp1
+ aesimc $dat0,$dat0
+ aesimc $dat1,$dat1
+ aesd $dat0,q10
+ aesd $dat1,q10
+ aesimc $dat0,$dat0
+ aesimc $dat1,$dat1
+ aesd $dat0,q11
+ aesd $dat1,q11
+ aesimc $dat0,$dat0
+ aesimc $dat1,$dat1
+ aesd $dat0,q12
+ aesd $dat1,q12
+ aesimc $dat0,$dat0
+ aesimc $dat1,$dat1
+ aesd $dat0,q13
+ aesd $dat1,q13
+ aesimc $dat0,$dat0
+ aesimc $dat1,$dat1
+ aesd $dat0,q14
+ aesd $dat1,q14
+ aesimc $dat0,$dat0
+ aesimc $dat1,$dat1
+ aesd $dat0,q15
+ aesd $dat1,q15
+
+ veor $ivec,$ivec,$dat0
+ vld1.8 {$dat0},[$inp],$step
+ veor $in0,$in0,$dat1
+ vld1.8 {$dat1},[$inp],$step1
+ vst1.8 {$ivec},[$out],#16
+ veor $ivec,$in1,$rndlast
+ vst1.8 {$in0},[$out],#16
+ veor $in0,$dat0,$rndlast
+ vorr $in1,$dat1,$dat1
+ b.hs .Loop2x_cbc_dec128
+
+ adds $len,$len,#32
+ veor $ivec,$ivec,$rndlast
+ b.eq .Lcbc_done
+ veor $in0,$in0,$rndlast
+ b .Lcbc_dec_tail
+
+.align 5
+.Lcbc_dec:
+ subs $len,$len,#16
+ vorr $in0,$dat,$dat
+ b.lo .Lcbc_dec_tail
+
+ cclr $step,eq
+ cmp $rounds,#2
+ vld1.8 {$dat1},[$inp],$step
+ vorr $in1,$dat1,$dat1
+ b.eq .Lcbc_dec128
+
+.Loop2x_cbc_dec:
+ aesd $dat0,q8
+ aesd $dat1,q8
+ vld1.32 {q8},[$key_],#16
+ aesimc $dat0,$dat0
+ aesimc $dat1,$dat1
+ subs $cnt,$cnt,#2
+ aesd $dat0,q9
+ aesd $dat1,q9
+ vld1.32 {q9},[$key_],#16
+ aesimc $dat0,$dat0
+ aesimc $dat1,$dat1
+ b.gt .Loop2x_cbc_dec
+
+ aesd $dat0,q8
+ aesd $dat1,q8
+ aesimc $dat0,$dat0
+ aesimc $dat1,$dat1
+ veor $tmp0,$ivec,$rndlast
+ veor $tmp1,$in0,$rndlast
+ aesd $dat0,q9
+ aesd $dat1,q9
+ aesimc $dat0,$dat0
+ aesimc $dat1,$dat1
+ vorr $ivec,$in1,$in1
+ subs $len,$len,#32
+ aesd $dat0,q10
+ aesd $dat1,q10
+ aesimc $dat0,$dat0
+ cclr $step,lo
+ aesimc $dat1,$dat1
+ mov $key_,$key
+ aesd $dat0,q11
+ aesd $dat1,q11
+ aesimc $dat0,$dat0
+ vld1.8 {$in0},[$inp],$step
+ aesimc $dat1,$dat1
+ cclr $step,ls
+ aesd $dat0,q12
+ aesd $dat1,q12
+ aesimc $dat0,$dat0
+ aesimc $dat1,$dat1
+ vld1.8 {$in1},[$inp],$step
+ aesd $dat0,q13
+ aesd $dat1,q13
+ aesimc $dat0,$dat0
+ aesimc $dat1,$dat1
+ vld1.32 {q8},[$key_],#16 // re-pre-load rndkey[0]
+ aesd $dat0,q14
+ aesd $dat1,q14
+ aesimc $dat0,$dat0
+ aesimc $dat1,$dat1
+ vld1.32 {q9},[$key_],#16 // re-pre-load rndkey[1]
+ aesd $dat0,q15
+ aesd $dat1,q15
+
+ mov $cnt,$rounds
+ veor $tmp0,$tmp0,$dat0
+ veor $tmp1,$tmp1,$dat1
+ vorr $dat0,$in0,$in0
+ vst1.8 {$tmp0},[$out],#16
+ vorr $dat1,$in1,$in1
+ vst1.8 {$tmp1},[$out],#16
+ b.hs .Loop2x_cbc_dec
+
+ adds $len,$len,#32
+ b.eq .Lcbc_done
+
+.Lcbc_dec_tail:
+ aesd $dat,q8
+ vld1.32 {q8},[$key_],#16
+ aesimc $dat,$dat
+ subs $cnt,$cnt,#2
+ aesd $dat,q9
+ vld1.32 {q9},[$key_],#16
+ aesimc $dat,$dat
+ b.gt .Lcbc_dec_tail
+
+ aesd $dat,q8
+ aesimc $dat,$dat
+ aesd $dat,q9
+ aesimc $dat,$dat
+ veor $tmp,$ivec,$rndlast
+ aesd $dat,q10
+ aesimc $dat,$dat
+ vorr $ivec,$in0,$in0
+ aesd $dat,q11
+ aesimc $dat,$dat
+ aesd $dat,q12
+ aesimc $dat,$dat
+ aesd $dat,q13
+ aesimc $dat,$dat
+ aesd $dat,q14
+ aesimc $dat,$dat
+ aesd $dat,q15
+
+ veor $tmp,$tmp,$dat
+ vst1.8 {$tmp},[$out],#16
+
+.Lcbc_done:
+ vst1.8 {$ivec},[$ivp]
+.Lcbc_abort:
+___
+$code.=<<___ if ($flavour !~ /64/);
+ vldmia sp!,{d8-d15}
+ ldmia sp!,{r4-r8,pc}
+___
+$code.=<<___ if ($flavour =~ /64/);
+ ldr x29,[sp],#16
+ ret
+___
+$code.=<<___;
+.size ${prefix}_cbc_encrypt,.-${prefix}_cbc_encrypt
+___
+}}}
+{{{
+my ($inp,$out,$len,$key,$ivp)=map("x$_",(0..4));
+my ($rounds,$cnt,$key_,$ctr,$tctr,$tctr1)=("w5","w6","x7","w8","w9","w10");
+my ($dat0,$dat1,$in0,$in1,$tmp0,$tmp1,$ivec,$rndlast)=map("q$_",(0..7));
+
+my ($dat,$tmp)=($dat0,$tmp0);
+
+### q8-q15 preloaded key schedule
+
+$code.=<<___;
+.globl ${prefix}_ctr32_encrypt_blocks
+.type ${prefix}_ctr32_encrypt_blocks,%function
+.align 5
+${prefix}_ctr32_encrypt_blocks:
+___
+$code.=<<___ if ($flavour =~ /64/);
+ stp x29,x30,[sp,#-16]!
+ add x29,sp,#0
+___
+$code.=<<___ if ($flavour !~ /64/);
+ mov ip,sp
+ stmdb sp!,{r4-r10,lr}
+ vstmdb sp!,{d8-d15} @ ABI specification says so
+ ldr r4, [ip] @ load remaining arg
+___
+$code.=<<___;
+ ldr $rounds,[$key,#240]
+
+ ldr $ctr, [$ivp, #12]
+ vld1.32 {$dat0},[$ivp]
+
+ vld1.32 {q8-q9},[$key] // load key schedule...
+ sub $rounds,$rounds,#6
+ add $key_,$key,x5,lsl#4 // pointer to last 7 round keys
+ sub $rounds,$rounds,#2
+ vld1.32 {q10-q11},[$key_],#32
+ vld1.32 {q12-q13},[$key_],#32
+ vld1.32 {q14-q15},[$key_],#32
+ vld1.32 {$rndlast},[$key_]
+
+ add $key_,$key,#32
+ mov $cnt,$rounds
+
+ subs $len,$len,#2
+ b.lo .Lctr32_tail
+
+#ifndef __ARMEB__
+ rev $ctr, $ctr
+#endif
+ vorr $dat1,$dat0,$dat0
+ add $ctr, $ctr, #1
+ vorr $ivec,$dat0,$dat0
+ rev $tctr1, $ctr
+ cmp $rounds,#2
+ vmov.32 ${dat1}[3],$tctr1
+ b.eq .Lctr32_128
+
+.Loop2x_ctr32:
+ aese $dat0,q8
+ aese $dat1,q8
+ vld1.32 {q8},[$key_],#16
+ aesmc $dat0,$dat0
+ aesmc $dat1,$dat1
+ subs $cnt,$cnt,#2
+ aese $dat0,q9
+ aese $dat1,q9
+ vld1.32 {q9},[$key_],#16
+ aesmc $dat0,$dat0
+ aesmc $dat1,$dat1
+ b.gt .Loop2x_ctr32
+
+ aese $dat0,q8
+ aese $dat1,q8
+ aesmc $tmp0,$dat0
+ vorr $dat0,$ivec,$ivec
+ aesmc $tmp1,$dat1
+ vorr $dat1,$ivec,$ivec
+ aese $tmp0,q9
+ aese $tmp1,q9
+ vld1.8 {$in0},[$inp],#16
+ aesmc $tmp0,$tmp0
+ vld1.8 {$in1},[$inp],#16
+ aesmc $tmp1,$tmp1
+ add $ctr,$ctr,#1
+ aese $tmp0,q10
+ aese $tmp1,q10
+ rev $tctr,$ctr
+ aesmc $tmp0,$tmp0
+ aesmc $tmp1,$tmp1
+ add $ctr,$ctr,#1
+ aese $tmp0,q11
+ aese $tmp1,q11
+ veor $in0,$in0,$rndlast
+ rev $tctr1,$ctr
+ aesmc $tmp0,$tmp0
+ aesmc $tmp1,$tmp1
+ veor $in1,$in1,$rndlast
+ mov $key_,$key
+ aese $tmp0,q12
+ aese $tmp1,q12
+ subs $len,$len,#2
+ aesmc $tmp0,$tmp0
+ aesmc $tmp1,$tmp1
+ vld1.32 {q8-q9},[$key_],#32 // re-pre-load rndkey[0-1]
+ aese $tmp0,q13
+ aese $tmp1,q13
+ aesmc $tmp0,$tmp0
+ aesmc $tmp1,$tmp1
+ aese $tmp0,q14
+ aese $tmp1,q14
+ vmov.32 ${dat0}[3], $tctr
+ aesmc $tmp0,$tmp0
+ vmov.32 ${dat1}[3], $tctr1
+ aesmc $tmp1,$tmp1
+ aese $tmp0,q15
+ aese $tmp1,q15
+
+ mov $cnt,$rounds
+ veor $in0,$in0,$tmp0
+ veor $in1,$in1,$tmp1
+ vst1.8 {$in0},[$out],#16
+ vst1.8 {$in1},[$out],#16
+ b.hs .Loop2x_ctr32
+
+ adds $len,$len,#2
+ b.eq .Lctr32_done
+ b .Lctr32_tail
+
+.Lctr32_128:
+ vld1.32 {$tmp0-$tmp1},[$key_]
+
+.Loop2x_ctr32_128:
+ aese $dat0,q8
+ aese $dat1,q8
+ aesmc $dat0,$dat0
+ vld1.8 {$in0},[$inp],#16
+ aesmc $dat1,$dat1
+ vld1.8 {$in1},[$inp],#16
+ aese $dat0,q9
+ aese $dat1,q9
+ add $ctr,$ctr,#1
+ aesmc $dat0,$dat0
+ aesmc $dat1,$dat1
+ rev $tctr,$ctr
+ aese $dat0,$tmp0
+ aese $dat1,$tmp0
+ add $ctr,$ctr,#1
+ aesmc $dat0,$dat0
+ aesmc $dat1,$dat1
+ rev $tctr1,$ctr
+ aese $dat0,$tmp1
+ aese $dat1,$tmp1
+ subs $len,$len,#2
+ aesmc $dat0,$dat0
+ aesmc $dat1,$dat1
+ aese $dat0,q10
+ aese $dat1,q10
+ aesmc $dat0,$dat0
+ aesmc $dat1,$dat1
+ aese $dat0,q11
+ aese $dat1,q11
+ aesmc $dat0,$dat0
+ aesmc $dat1,$dat1
+ aese $dat0,q12
+ aese $dat1,q12
+ aesmc $dat0,$dat0
+ aesmc $dat1,$dat1
+ aese $dat0,q13
+ aese $dat1,q13
+ aesmc $dat0,$dat0
+ aesmc $dat1,$dat1
+ aese $dat0,q14
+ aese $dat1,q14
+ aesmc $dat0,$dat0
+ aesmc $dat1,$dat1
+ veor $in0,$in0,$rndlast
+ aese $dat0,q15
+ veor $in1,$in1,$rndlast
+ aese $dat1,q15
+
+ veor $in0,$in0,$dat0
+ vorr $dat0,$ivec,$ivec
+ veor $in1,$in1,$dat1
+ vorr $dat1,$ivec,$ivec
+ vst1.8 {$in0},[$out],#16
+ vmov.32 ${dat0}[3], $tctr
+ vst1.8 {$in1},[$out],#16
+ vmov.32 ${dat1}[3], $tctr1
+ b.hs .Loop2x_ctr32_128
+
+ adds $len,$len,#2
+ b.eq .Lctr32_done
+
+.Lctr32_tail:
+ aese $dat,q8
+ vld1.32 {q8},[$key_],#16
+ aesmc $dat,$dat
+ subs $cnt,$cnt,#2
+ aese $dat,q9
+ vld1.32 {q9},[$key_],#16
+ aesmc $dat,$dat
+ b.gt .Lctr32_tail
+
+ aese $dat,q8
+ aesmc $dat,$dat
+ aese $dat,q9
+ aesmc $dat,$dat
+ vld1.8 {$in0},[$inp]
+ aese $dat,q10
+ aesmc $dat,$dat
+ aese $dat,q11
+ aesmc $dat,$dat
+ aese $dat,q12
+ aesmc $dat,$dat
+ aese $dat,q13
+ aesmc $dat,$dat
+ aese $dat,q14
+ aesmc $dat,$dat
+ veor $in0,$in0,$rndlast
+ aese $dat,q15
+
+ veor $in0,$in0,$dat
+ vst1.8 {$in0},[$out]
+
+.Lctr32_done:
+___
+$code.=<<___ if ($flavour !~ /64/);
+ vldmia sp!,{d8-d15}
+ ldmia sp!,{r4-r10,pc}
+___
+$code.=<<___ if ($flavour =~ /64/);
+ ldr x29,[sp],#16
+ ret
+___
+$code.=<<___;
+.size ${prefix}_ctr32_encrypt_blocks,.-${prefix}_ctr32_encrypt_blocks
+___
+}}}
+$code.=<<___;
+#endif
+___
+########################################
+if ($flavour =~ /64/) { ######## 64-bit code
+ my %opcode = (
+ "aesd" => 0x4e285800, "aese" => 0x4e284800,
+ "aesimc"=> 0x4e287800, "aesmc" => 0x4e286800 );
+
+ local *unaes = sub {
+ my ($mnemonic,$arg)=@_;
+
+ $arg =~ m/[qv]([0-9]+)[^,]*,\s*[qv]([0-9]+)/o &&
+ sprintf ".inst\t0x%08x\t//%s %s",
+ $opcode{$mnemonic}|$1|($2<<5),
+ $mnemonic,$arg;
+ };
+
+ foreach(split("\n",$code)) {
+ s/\`([^\`]*)\`/eval($1)/geo;
+
+ s/\bq([0-9]+)\b/"v".($1<8?$1:$1+8).".16b"/geo; # old->new registers
+ s/@\s/\/\//o; # old->new style commentary
+
+ #s/[v]?(aes\w+)\s+([qv].*)/unaes($1,$2)/geo or
+ s/cclr\s+([wx])([^,]+),\s*([a-z]+)/csel $1$2,$1zr,$1$2,$3/o or
+ s/vmov\.i8/movi/o or # fix up legacy mnemonics
+ s/vext\.8/ext/o or
+ s/vrev32\.8/rev32/o or
+ s/vtst\.8/cmtst/o or
+ s/vshr/ushr/o or
+ s/^(\s+)v/$1/o or # strip off v prefix
+ s/\bbx\s+lr\b/ret/o;
+
+ # fix up remainig legacy suffixes
+ s/\.[ui]?8//o;
+ m/\],#8/o and s/\.16b/\.8b/go;
+ s/\.[ui]?32//o and s/\.16b/\.4s/go;
+ s/\.[ui]?64//o and s/\.16b/\.2d/go;
+ s/\.[42]([sd])\[([0-3])\]/\.$1\[$2\]/o;
+
+ print $_,"\n";
+ }
+} else { ######## 32-bit code
+ my %opcode = (
+ "aesd" => 0xf3b00340, "aese" => 0xf3b00300,
+ "aesimc"=> 0xf3b003c0, "aesmc" => 0xf3b00380 );
+
+ local *unaes = sub {
+ my ($mnemonic,$arg)=@_;
+
+ if ($arg =~ m/[qv]([0-9]+)[^,]*,\s*[qv]([0-9]+)/o) {
+ my $word = $opcode{$mnemonic}|(($1&7)<<13)|(($1&8)<<19)
+ |(($2&7)<<1) |(($2&8)<<2);
+ # since ARMv7 instructions are always encoded little-endian.
+ # correct solution is to use .inst directive, but older
+ # assemblers don't implement it:-(
+ sprintf ".byte\t0x%02x,0x%02x,0x%02x,0x%02x\t@ %s %s",
+ $word&0xff,($word>>8)&0xff,
+ ($word>>16)&0xff,($word>>24)&0xff,
+ $mnemonic,$arg;
+ }
+ };
+
+ sub unvtbl {
+ my $arg=shift;
+
+ $arg =~ m/q([0-9]+),\s*\{q([0-9]+)\},\s*q([0-9]+)/o &&
+ sprintf "vtbl.8 d%d,{q%d},d%d\n\t".
+ "vtbl.8 d%d,{q%d},d%d", 2*$1,$2,2*$3, 2*$1+1,$2,2*$3+1;
+ }
+
+ sub unvdup32 {
+ my $arg=shift;
+
+ $arg =~ m/q([0-9]+),\s*q([0-9]+)\[([0-3])\]/o &&
+ sprintf "vdup.32 q%d,d%d[%d]",$1,2*$2+($3>>1),$3&1;
+ }
+
+ sub unvmov32 {
+ my $arg=shift;
+
+ $arg =~ m/q([0-9]+)\[([0-3])\],(.*)/o &&
+ sprintf "vmov.32 d%d[%d],%s",2*$1+($2>>1),$2&1,$3;
+ }
+
+ foreach(split("\n",$code)) {
+ s/\`([^\`]*)\`/eval($1)/geo;
+
+ s/\b[wx]([0-9]+)\b/r$1/go; # new->old registers
+ s/\bv([0-9])\.[12468]+[bsd]\b/q$1/go; # new->old registers
+ s/\/\/\s?/@ /o; # new->old style commentary
+
+ # fix up remainig new-style suffixes
+ s/\{q([0-9]+)\},\s*\[(.+)\],#8/sprintf "{d%d},[$2]!",2*$1/eo or
+ s/\],#[0-9]+/]!/o;
+
+ s/[v]?(aes\w+)\s+([qv].*)/unaes($1,$2)/geo or
+ s/cclr\s+([^,]+),\s*([a-z]+)/mov$2 $1,#0/o or
+ s/vtbl\.8\s+(.*)/unvtbl($1)/geo or
+ s/vdup\.32\s+(.*)/unvdup32($1)/geo or
+ s/vmov\.32\s+(.*)/unvmov32($1)/geo or
+ s/^(\s+)b\./$1b/o or
+ s/^(\s+)ret/$1bx\tlr/o;
+
+ print $_,"\n";
+ }
+}
+
+close STDOUT;
diff --git a/crypto/aes/asm/bsaes-armv7.pl b/crypto/aes/asm/bsaes-armv7.pl
new file mode 100644
index 0000000..f3d96d9
--- /dev/null
+++ b/crypto/aes/asm/bsaes-armv7.pl
@@ -0,0 +1,2467 @@
+#!/usr/bin/env perl
+
+# ====================================================================
+# Written by Andy Polyakov <appro@openssl.org> for the OpenSSL
+# project. The module is, however, dual licensed under OpenSSL and
+# CRYPTOGAMS licenses depending on where you obtain it. For further
+# details see http://www.openssl.org/~appro/cryptogams/.
+#
+# Specific modes and adaptation for Linux kernel by Ard Biesheuvel
+# <ard.biesheuvel@linaro.org>. Permission to use under GPL terms is
+# granted.
+# ====================================================================
+
+# Bit-sliced AES for ARM NEON
+#
+# February 2012.
+#
+# This implementation is direct adaptation of bsaes-x86_64 module for
+# ARM NEON. Except that this module is endian-neutral [in sense that
+# it can be compiled for either endianness] by courtesy of vld1.8's
+# neutrality. Initial version doesn't implement interface to OpenSSL,
+# only low-level primitives and unsupported entry points, just enough
+# to collect performance results, which for Cortex-A8 core are:
+#
+# encrypt 19.5 cycles per byte processed with 128-bit key
+# decrypt 22.1 cycles per byte processed with 128-bit key
+# key conv. 440 cycles per 128-bit key/0.18 of 8x block
+#
+# Snapdragon S4 encrypts byte in 17.6 cycles and decrypts in 19.7,
+# which is [much] worse than anticipated (for further details see
+# http://www.openssl.org/~appro/Snapdragon-S4.html).
+#
+# Cortex-A15 manages in 14.2/16.1 cycles [when integer-only code
+# manages in 20.0 cycles].
+#
+# When comparing to x86_64 results keep in mind that NEON unit is
+# [mostly] single-issue and thus can't [fully] benefit from
+# instruction-level parallelism. And when comparing to aes-armv4
+# results keep in mind key schedule conversion overhead (see
+# bsaes-x86_64.pl for further details)...
+#
+# <appro@openssl.org>
+
+# April-August 2013
+#
+# Add CBC, CTR and XTS subroutines, adapt for kernel use.
+#
+# <ard.biesheuvel@linaro.org>
+
+while (($output=shift) && ($output!~/^\w[\w\-]*\.\w+$/)) {}
+open STDOUT,">$output";
+
+my ($inp,$out,$len,$key)=("r0","r1","r2","r3");
+my @XMM=map("q$_",(0..15));
+
+{
+my ($key,$rounds,$const)=("r4","r5","r6");
+
+sub Dlo() { shift=~m|q([1]?[0-9])|?"d".($1*2):""; }
+sub Dhi() { shift=~m|q([1]?[0-9])|?"d".($1*2+1):""; }
+
+sub Sbox {
+# input in lsb > [b0, b1, b2, b3, b4, b5, b6, b7] < msb
+# output in lsb > [b0, b1, b4, b6, b3, b7, b2, b5] < msb
+my @b=@_[0..7];
+my @t=@_[8..11];
+my @s=@_[12..15];
+ &InBasisChange (@b);
+ &Inv_GF256 (@b[6,5,0,3,7,1,4,2],@t,@s);
+ &OutBasisChange (@b[7,1,4,2,6,5,0,3]);
+}
+
+sub InBasisChange {
+# input in lsb > [b0, b1, b2, b3, b4, b5, b6, b7] < msb
+# output in lsb > [b6, b5, b0, b3, b7, b1, b4, b2] < msb
+my @b=@_[0..7];
+$code.=<<___;
+ veor @b[2], @b[2], @b[1]
+ veor @b[5], @b[5], @b[6]
+ veor @b[3], @b[3], @b[0]
+ veor @b[6], @b[6], @b[2]
+ veor @b[5], @b[5], @b[0]
+
+ veor @b[6], @b[6], @b[3]
+ veor @b[3], @b[3], @b[7]
+ veor @b[7], @b[7], @b[5]
+ veor @b[3], @b[3], @b[4]
+ veor @b[4], @b[4], @b[5]
+
+ veor @b[2], @b[2], @b[7]
+ veor @b[3], @b[3], @b[1]
+ veor @b[1], @b[1], @b[5]
+___
+}
+
+sub OutBasisChange {
+# input in lsb > [b0, b1, b2, b3, b4, b5, b6, b7] < msb
+# output in lsb > [b6, b1, b2, b4, b7, b0, b3, b5] < msb
+my @b=@_[0..7];
+$code.=<<___;
+ veor @b[0], @b[0], @b[6]
+ veor @b[1], @b[1], @b[4]
+ veor @b[4], @b[4], @b[6]
+ veor @b[2], @b[2], @b[0]
+ veor @b[6], @b[6], @b[1]
+
+ veor @b[1], @b[1], @b[5]
+ veor @b[5], @b[5], @b[3]
+ veor @b[3], @b[3], @b[7]
+ veor @b[7], @b[7], @b[5]
+ veor @b[2], @b[2], @b[5]
+
+ veor @b[4], @b[4], @b[7]
+___
+}
+
+sub InvSbox {
+# input in lsb > [b0, b1, b2, b3, b4, b5, b6, b7] < msb
+# output in lsb > [b0, b1, b6, b4, b2, b7, b3, b5] < msb
+my @b=@_[0..7];
+my @t=@_[8..11];
+my @s=@_[12..15];
+ &InvInBasisChange (@b);
+ &Inv_GF256 (@b[5,1,2,6,3,7,0,4],@t,@s);
+ &InvOutBasisChange (@b[3,7,0,4,5,1,2,6]);
+}
+
+sub InvInBasisChange { # OutBasisChange in reverse (with twist)
+my @b=@_[5,1,2,6,3,7,0,4];
+$code.=<<___
+ veor @b[1], @b[1], @b[7]
+ veor @b[4], @b[4], @b[7]
+
+ veor @b[7], @b[7], @b[5]
+ veor @b[1], @b[1], @b[3]
+ veor @b[2], @b[2], @b[5]
+ veor @b[3], @b[3], @b[7]
+
+ veor @b[6], @b[6], @b[1]
+ veor @b[2], @b[2], @b[0]
+ veor @b[5], @b[5], @b[3]
+ veor @b[4], @b[4], @b[6]
+ veor @b[0], @b[0], @b[6]
+ veor @b[1], @b[1], @b[4]
+___
+}
+
+sub InvOutBasisChange { # InBasisChange in reverse
+my @b=@_[2,5,7,3,6,1,0,4];
+$code.=<<___;
+ veor @b[1], @b[1], @b[5]
+ veor @b[2], @b[2], @b[7]
+
+ veor @b[3], @b[3], @b[1]
+ veor @b[4], @b[4], @b[5]
+ veor @b[7], @b[7], @b[5]
+ veor @b[3], @b[3], @b[4]
+ veor @b[5], @b[5], @b[0]
+ veor @b[3], @b[3], @b[7]
+ veor @b[6], @b[6], @b[2]
+ veor @b[2], @b[2], @b[1]
+ veor @b[6], @b[6], @b[3]
+
+ veor @b[3], @b[3], @b[0]
+ veor @b[5], @b[5], @b[6]
+___
+}
+
+sub Mul_GF4 {
+#;*************************************************************
+#;* Mul_GF4: Input x0-x1,y0-y1 Output x0-x1 Temp t0 (8) *
+#;*************************************************************
+my ($x0,$x1,$y0,$y1,$t0,$t1)=@_;
+$code.=<<___;
+ veor $t0, $y0, $y1
+ vand $t0, $t0, $x0
+ veor $x0, $x0, $x1
+ vand $t1, $x1, $y0
+ vand $x0, $x0, $y1
+ veor $x1, $t1, $t0
+ veor $x0, $x0, $t1
+___
+}
+
+sub Mul_GF4_N { # not used, see next subroutine
+# multiply and scale by N
+my ($x0,$x1,$y0,$y1,$t0)=@_;
+$code.=<<___;
+ veor $t0, $y0, $y1
+ vand $t0, $t0, $x0
+ veor $x0, $x0, $x1
+ vand $x1, $x1, $y0
+ vand $x0, $x0, $y1
+ veor $x1, $x1, $x0
+ veor $x0, $x0, $t0
+___
+}
+
+sub Mul_GF4_N_GF4 {
+# interleaved Mul_GF4_N and Mul_GF4
+my ($x0,$x1,$y0,$y1,$t0,
+ $x2,$x3,$y2,$y3,$t1)=@_;
+$code.=<<___;
+ veor $t0, $y0, $y1
+ veor $t1, $y2, $y3
+ vand $t0, $t0, $x0
+ vand $t1, $t1, $x2
+ veor $x0, $x0, $x1
+ veor $x2, $x2, $x3
+ vand $x1, $x1, $y0
+ vand $x3, $x3, $y2
+ vand $x0, $x0, $y1
+ vand $x2, $x2, $y3
+ veor $x1, $x1, $x0
+ veor $x2, $x2, $x3
+ veor $x0, $x0, $t0
+ veor $x3, $x3, $t1
+___
+}
+sub Mul_GF16_2 {
+my @x=@_[0..7];
+my @y=@_[8..11];
+my @t=@_[12..15];
+$code.=<<___;
+ veor @t[0], @x[0], @x[2]
+ veor @t[1], @x[1], @x[3]
+___
+ &Mul_GF4 (@x[0], @x[1], @y[0], @y[1], @t[2..3]);
+$code.=<<___;
+ veor @y[0], @y[0], @y[2]
+ veor @y[1], @y[1], @y[3]
+___
+ Mul_GF4_N_GF4 (@t[0], @t[1], @y[0], @y[1], @t[3],
+ @x[2], @x[3], @y[2], @y[3], @t[2]);
+$code.=<<___;
+ veor @x[0], @x[0], @t[0]
+ veor @x[2], @x[2], @t[0]
+ veor @x[1], @x[1], @t[1]
+ veor @x[3], @x[3], @t[1]
+
+ veor @t[0], @x[4], @x[6]
+ veor @t[1], @x[5], @x[7]
+___
+ &Mul_GF4_N_GF4 (@t[0], @t[1], @y[0], @y[1], @t[3],
+ @x[6], @x[7], @y[2], @y[3], @t[2]);
+$code.=<<___;
+ veor @y[0], @y[0], @y[2]
+ veor @y[1], @y[1], @y[3]
+___
+ &Mul_GF4 (@x[4], @x[5], @y[0], @y[1], @t[2..3]);
+$code.=<<___;
+ veor @x[4], @x[4], @t[0]
+ veor @x[6], @x[6], @t[0]
+ veor @x[5], @x[5], @t[1]
+ veor @x[7], @x[7], @t[1]
+___
+}
+sub Inv_GF256 {
+#;********************************************************************
+#;* Inv_GF256: Input x0-x7 Output x0-x7 Temp t0-t3,s0-s3 (144) *
+#;********************************************************************
+my @x=@_[0..7];
+my @t=@_[8..11];
+my @s=@_[12..15];
+# direct optimizations from hardware
+$code.=<<___;
+ veor @t[3], @x[4], @x[6]
+ veor @t[2], @x[5], @x[7]
+ veor @t[1], @x[1], @x[3]
+ veor @s[1], @x[7], @x[6]
+ vmov @t[0], @t[2]
+ veor @s[0], @x[0], @x[2]
+
+ vorr @t[2], @t[2], @t[1]
+ veor @s[3], @t[3], @t[0]
+ vand @s[2], @t[3], @s[0]
+ vorr @t[3], @t[3], @s[0]
+ veor @s[0], @s[0], @t[1]
+ vand @t[0], @t[0], @t[1]
+ veor @t[1], @x[3], @x[2]
+ vand @s[3], @s[3], @s[0]
+ vand @s[1], @s[1], @t[1]
+ veor @t[1], @x[4], @x[5]
+ veor @s[0], @x[1], @x[0]
+ veor @t[3], @t[3], @s[1]
+ veor @t[2], @t[2], @s[1]
+ vand @s[1], @t[1], @s[0]
+ vorr @t[1], @t[1], @s[0]
+ veor @t[3], @t[3], @s[3]
+ veor @t[0], @t[0], @s[1]
+ veor @t[2], @t[2], @s[2]
+ veor @t[1], @t[1], @s[3]
+ veor @t[0], @t[0], @s[2]
+ vand @s[0], @x[7], @x[3]
+ veor @t[1], @t[1], @s[2]
+ vand @s[1], @x[6], @x[2]
+ vand @s[2], @x[5], @x[1]
+ vorr @s[3], @x[4], @x[0]
+ veor @t[3], @t[3], @s[0]
+ veor @t[1], @t[1], @s[2]
+ veor @t[0], @t[0], @s[3]
+ veor @t[2], @t[2], @s[1]
+
+ @ Inv_GF16 \t0, \t1, \t2, \t3, \s0, \s1, \s2, \s3
+
+ @ new smaller inversion
+
+ vand @s[2], @t[3], @t[1]
+ vmov @s[0], @t[0]
+
+ veor @s[1], @t[2], @s[2]
+ veor @s[3], @t[0], @s[2]
+ veor @s[2], @t[0], @s[2] @ @s[2]=@s[3]
+
+ vbsl @s[1], @t[1], @t[0]
+ vbsl @s[3], @t[3], @t[2]
+ veor @t[3], @t[3], @t[2]
+
+ vbsl @s[0], @s[1], @s[2]
+ vbsl @t[0], @s[2], @s[1]
+
+ vand @s[2], @s[0], @s[3]
+ veor @t[1], @t[1], @t[0]
+
+ veor @s[2], @s[2], @t[3]
+___
+# output in s3, s2, s1, t1
+
+# Mul_GF16_2 \x0, \x1, \x2, \x3, \x4, \x5, \x6, \x7, \t2, \t3, \t0, \t1, \s0, \s1, \s2, \s3
+
+# Mul_GF16_2 \x0, \x1, \x2, \x3, \x4, \x5, \x6, \x7, \s3, \s2, \s1, \t1, \s0, \t0, \t2, \t3
+ &Mul_GF16_2(@x,@s[3,2,1],@t[1],@s[0],@t[0,2,3]);
+
+### output msb > [x3,x2,x1,x0,x7,x6,x5,x4] < lsb
+}
+
+# AES linear components
+
+sub ShiftRows {
+my @x=@_[0..7];
+my @t=@_[8..11];
+my $mask=pop;
+$code.=<<___;
+ vldmia $key!, {@t[0]-@t[3]}
+ veor @t[0], @t[0], @x[0]
+ veor @t[1], @t[1], @x[1]
+ vtbl.8 `&Dlo(@x[0])`, {@t[0]}, `&Dlo($mask)`
+ vtbl.8 `&Dhi(@x[0])`, {@t[0]}, `&Dhi($mask)`
+ vldmia $key!, {@t[0]}
+ veor @t[2], @t[2], @x[2]
+ vtbl.8 `&Dlo(@x[1])`, {@t[1]}, `&Dlo($mask)`
+ vtbl.8 `&Dhi(@x[1])`, {@t[1]}, `&Dhi($mask)`
+ vldmia $key!, {@t[1]}
+ veor @t[3], @t[3], @x[3]
+ vtbl.8 `&Dlo(@x[2])`, {@t[2]}, `&Dlo($mask)`
+ vtbl.8 `&Dhi(@x[2])`, {@t[2]}, `&Dhi($mask)`
+ vldmia $key!, {@t[2]}
+ vtbl.8 `&Dlo(@x[3])`, {@t[3]}, `&Dlo($mask)`
+ vtbl.8 `&Dhi(@x[3])`, {@t[3]}, `&Dhi($mask)`
+ vldmia $key!, {@t[3]}
+ veor @t[0], @t[0], @x[4]
+ veor @t[1], @t[1], @x[5]
+ vtbl.8 `&Dlo(@x[4])`, {@t[0]}, `&Dlo($mask)`
+ vtbl.8 `&Dhi(@x[4])`, {@t[0]}, `&Dhi($mask)`
+ veor @t[2], @t[2], @x[6]
+ vtbl.8 `&Dlo(@x[5])`, {@t[1]}, `&Dlo($mask)`
+ vtbl.8 `&Dhi(@x[5])`, {@t[1]}, `&Dhi($mask)`
+ veor @t[3], @t[3], @x[7]
+ vtbl.8 `&Dlo(@x[6])`, {@t[2]}, `&Dlo($mask)`
+ vtbl.8 `&Dhi(@x[6])`, {@t[2]}, `&Dhi($mask)`
+ vtbl.8 `&Dlo(@x[7])`, {@t[3]}, `&Dlo($mask)`
+ vtbl.8 `&Dhi(@x[7])`, {@t[3]}, `&Dhi($mask)`
+___
+}
+
+sub MixColumns {
+# modified to emit output in order suitable for feeding back to aesenc[last]
+my @x=@_[0..7];
+my @t=@_[8..15];
+my $inv=@_[16]; # optional
+$code.=<<___;
+ vext.8 @t[0], @x[0], @x[0], #12 @ x0 <<< 32
+ vext.8 @t[1], @x[1], @x[1], #12
+ veor @x[0], @x[0], @t[0] @ x0 ^ (x0 <<< 32)
+ vext.8 @t[2], @x[2], @x[2], #12
+ veor @x[1], @x[1], @t[1]
+ vext.8 @t[3], @x[3], @x[3], #12
+ veor @x[2], @x[2], @t[2]
+ vext.8 @t[4], @x[4], @x[4], #12
+ veor @x[3], @x[3], @t[3]
+ vext.8 @t[5], @x[5], @x[5], #12
+ veor @x[4], @x[4], @t[4]
+ vext.8 @t[6], @x[6], @x[6], #12
+ veor @x[5], @x[5], @t[5]
+ vext.8 @t[7], @x[7], @x[7], #12
+ veor @x[6], @x[6], @t[6]
+
+ veor @t[1], @t[1], @x[0]
+ veor @x[7], @x[7], @t[7]
+ vext.8 @x[0], @x[0], @x[0], #8 @ (x0 ^ (x0 <<< 32)) <<< 64)
+ veor @t[2], @t[2], @x[1]
+ veor @t[0], @t[0], @x[7]
+ veor @t[1], @t[1], @x[7]
+ vext.8 @x[1], @x[1], @x[1], #8
+ veor @t[5], @t[5], @x[4]
+ veor @x[0], @x[0], @t[0]
+ veor @t[6], @t[6], @x[5]
+ veor @x[1], @x[1], @t[1]
+ vext.8 @t[0], @x[4], @x[4], #8
+ veor @t[4], @t[4], @x[3]
+ vext.8 @t[1], @x[5], @x[5], #8
+ veor @t[7], @t[7], @x[6]
+ vext.8 @x[4], @x[3], @x[3], #8
+ veor @t[3], @t[3], @x[2]
+ vext.8 @x[5], @x[7], @x[7], #8
+ veor @t[4], @t[4], @x[7]
+ vext.8 @x[3], @x[6], @x[6], #8
+ veor @t[3], @t[3], @x[7]
+ vext.8 @x[6], @x[2], @x[2], #8
+ veor @x[7], @t[1], @t[5]
+___
+$code.=<<___ if (!$inv);
+ veor @x[2], @t[0], @t[4]
+ veor @x[4], @x[4], @t[3]
+ veor @x[5], @x[5], @t[7]
+ veor @x[3], @x[3], @t[6]
+ @ vmov @x[2], @t[0]
+ veor @x[6], @x[6], @t[2]
+ @ vmov @x[7], @t[1]
+___
+$code.=<<___ if ($inv);
+ veor @t[3], @t[3], @x[4]
+ veor @x[5], @x[5], @t[7]
+ veor @x[2], @x[3], @t[6]
+ veor @x[3], @t[0], @t[4]
+ veor @x[4], @x[6], @t[2]
+ vmov @x[6], @t[3]
+ @ vmov @x[7], @t[1]
+___
+}
+
+sub InvMixColumns_orig {
+my @x=@_[0..7];
+my @t=@_[8..15];
+
+$code.=<<___;
+ @ multiplication by 0x0e
+ vext.8 @t[7], @x[7], @x[7], #12
+ vmov @t[2], @x[2]
+ veor @x[2], @x[2], @x[5] @ 2 5
+ veor @x[7], @x[7], @x[5] @ 7 5
+ vext.8 @t[0], @x[0], @x[0], #12
+ vmov @t[5], @x[5]
+ veor @x[5], @x[5], @x[0] @ 5 0 [1]
+ veor @x[0], @x[0], @x[1] @ 0 1
+ vext.8 @t[1], @x[1], @x[1], #12
+ veor @x[1], @x[1], @x[2] @ 1 25
+ veor @x[0], @x[0], @x[6] @ 01 6 [2]
+ vext.8 @t[3], @x[3], @x[3], #12
+ veor @x[1], @x[1], @x[3] @ 125 3 [4]
+ veor @x[2], @x[2], @x[0] @ 25 016 [3]
+ veor @x[3], @x[3], @x[7] @ 3 75
+ veor @x[7], @x[7], @x[6] @ 75 6 [0]
+ vext.8 @t[6], @x[6], @x[6], #12
+ vmov @t[4], @x[4]
+ veor @x[6], @x[6], @x[4] @ 6 4
+ veor @x[4], @x[4], @x[3] @ 4 375 [6]
+ veor @x[3], @x[3], @x[7] @ 375 756=36
+ veor @x[6], @x[6], @t[5] @ 64 5 [7]
+ veor @x[3], @x[3], @t[2] @ 36 2
+ vext.8 @t[5], @t[5], @t[5], #12
+ veor @x[3], @x[3], @t[4] @ 362 4 [5]
+___
+ my @y = @x[7,5,0,2,1,3,4,6];
+$code.=<<___;
+ @ multiplication by 0x0b
+ veor @y[1], @y[1], @y[0]
+ veor @y[0], @y[0], @t[0]
+ vext.8 @t[2], @t[2], @t[2], #12
+ veor @y[1], @y[1], @t[1]
+ veor @y[0], @y[0], @t[5]
+ vext.8 @t[4], @t[4], @t[4], #12
+ veor @y[1], @y[1], @t[6]
+ veor @y[0], @y[0], @t[7]
+ veor @t[7], @t[7], @t[6] @ clobber t[7]
+
+ veor @y[3], @y[3], @t[0]
+ veor @y[1], @y[1], @y[0]
+ vext.8 @t[0], @t[0], @t[0], #12
+ veor @y[2], @y[2], @t[1]
+ veor @y[4], @y[4], @t[1]
+ vext.8 @t[1], @t[1], @t[1], #12
+ veor @y[2], @y[2], @t[2]
+ veor @y[3], @y[3], @t[2]
+ veor @y[5], @y[5], @t[2]
+ veor @y[2], @y[2], @t[7]
+ vext.8 @t[2], @t[2], @t[2], #12
+ veor @y[3], @y[3], @t[3]
+ veor @y[6], @y[6], @t[3]
+ veor @y[4], @y[4], @t[3]
+ veor @y[7], @y[7], @t[4]
+ vext.8 @t[3], @t[3], @t[3], #12
+ veor @y[5], @y[5], @t[4]
+ veor @y[7], @y[7], @t[7]
+ veor @t[7], @t[7], @t[5] @ clobber t[7] even more
+ veor @y[3], @y[3], @t[5]
+ veor @y[4], @y[4], @t[4]
+
+ veor @y[5], @y[5], @t[7]
+ vext.8 @t[4], @t[4], @t[4], #12
+ veor @y[6], @y[6], @t[7]
+ veor @y[4], @y[4], @t[7]
+
+ veor @t[7], @t[7], @t[5]
+ vext.8 @t[5], @t[5], @t[5], #12
+
+ @ multiplication by 0x0d
+ veor @y[4], @y[4], @y[7]
+ veor @t[7], @t[7], @t[6] @ restore t[7]
+ veor @y[7], @y[7], @t[4]
+ vext.8 @t[6], @t[6], @t[6], #12
+ veor @y[2], @y[2], @t[0]
+ veor @y[7], @y[7], @t[5]
+ vext.8 @t[7], @t[7], @t[7], #12
+ veor @y[2], @y[2], @t[2]
+
+ veor @y[3], @y[3], @y[1]
+ veor @y[1], @y[1], @t[1]
+ veor @y[0], @y[0], @t[0]
+ veor @y[3], @y[3], @t[0]
+ veor @y[1], @y[1], @t[5]
+ veor @y[0], @y[0], @t[5]
+ vext.8 @t[0], @t[0], @t[0], #12
+ veor @y[1], @y[1], @t[7]
+ veor @y[0], @y[0], @t[6]
+ veor @y[3], @y[3], @y[1]
+ veor @y[4], @y[4], @t[1]
+ vext.8 @t[1], @t[1], @t[1], #12
+
+ veor @y[7], @y[7], @t[7]
+ veor @y[4], @y[4], @t[2]
+ veor @y[5], @y[5], @t[2]
+ veor @y[2], @y[2], @t[6]
+ veor @t[6], @t[6], @t[3] @ clobber t[6]
+ vext.8 @t[2], @t[2], @t[2], #12
+ veor @y[4], @y[4], @y[7]
+ veor @y[3], @y[3], @t[6]
+
+ veor @y[6], @y[6], @t[6]
+ veor @y[5], @y[5], @t[5]
+ vext.8 @t[5], @t[5], @t[5], #12
+ veor @y[6], @y[6], @t[4]
+ vext.8 @t[4], @t[4], @t[4], #12
+ veor @y[5], @y[5], @t[6]
+ veor @y[6], @y[6], @t[7]
+ vext.8 @t[7], @t[7], @t[7], #12
+ veor @t[6], @t[6], @t[3] @ restore t[6]
+ vext.8 @t[3], @t[3], @t[3], #12
+
+ @ multiplication by 0x09
+ veor @y[4], @y[4], @y[1]
+ veor @t[1], @t[1], @y[1] @ t[1]=y[1]
+ veor @t[0], @t[0], @t[5] @ clobber t[0]
+ vext.8 @t[6], @t[6], @t[6], #12
+ veor @t[1], @t[1], @t[5]
+ veor @y[3], @y[3], @t[0]
+ veor @t[0], @t[0], @y[0] @ t[0]=y[0]
+ veor @t[1], @t[1], @t[6]
+ veor @t[6], @t[6], @t[7] @ clobber t[6]
+ veor @y[4], @y[4], @t[1]
+ veor @y[7], @y[7], @t[4]
+ veor @y[6], @y[6], @t[3]
+ veor @y[5], @y[5], @t[2]
+ veor @t[4], @t[4], @y[4] @ t[4]=y[4]
+ veor @t[3], @t[3], @y[3] @ t[3]=y[3]
+ veor @t[5], @t[5], @y[5] @ t[5]=y[5]
+ veor @t[2], @t[2], @y[2] @ t[2]=y[2]
+ veor @t[3], @t[3], @t[7]
+ veor @XMM[5], @t[5], @t[6]
+ veor @XMM[6], @t[6], @y[6] @ t[6]=y[6]
+ veor @XMM[2], @t[2], @t[6]
+ veor @XMM[7], @t[7], @y[7] @ t[7]=y[7]
+
+ vmov @XMM[0], @t[0]
+ vmov @XMM[1], @t[1]
+ @ vmov @XMM[2], @t[2]
+ vmov @XMM[3], @t[3]
+ vmov @XMM[4], @t[4]
+ @ vmov @XMM[5], @t[5]
+ @ vmov @XMM[6], @t[6]
+ @ vmov @XMM[7], @t[7]
+___
+}
+
+sub InvMixColumns {
+my @x=@_[0..7];
+my @t=@_[8..15];
+
+# Thanks to Jussi Kivilinna for providing pointer to
+#
+# | 0e 0b 0d 09 | | 02 03 01 01 | | 05 00 04 00 |
+# | 09 0e 0b 0d | = | 01 02 03 01 | x | 00 05 00 04 |
+# | 0d 09 0e 0b | | 01 01 02 03 | | 04 00 05 00 |
+# | 0b 0d 09 0e | | 03 01 01 02 | | 00 04 00 05 |
+
+$code.=<<___;
+ @ multiplication by 0x05-0x00-0x04-0x00
+ vext.8 @t[0], @x[0], @x[0], #8
+ vext.8 @t[6], @x[6], @x[6], #8
+ vext.8 @t[7], @x[7], @x[7], #8
+ veor @t[0], @t[0], @x[0]
+ vext.8 @t[1], @x[1], @x[1], #8
+ veor @t[6], @t[6], @x[6]
+ vext.8 @t[2], @x[2], @x[2], #8
+ veor @t[7], @t[7], @x[7]
+ vext.8 @t[3], @x[3], @x[3], #8
+ veor @t[1], @t[1], @x[1]
+ vext.8 @t[4], @x[4], @x[4], #8
+ veor @t[2], @t[2], @x[2]
+ vext.8 @t[5], @x[5], @x[5], #8
+ veor @t[3], @t[3], @x[3]
+ veor @t[4], @t[4], @x[4]
+ veor @t[5], @t[5], @x[5]
+
+ veor @x[0], @x[0], @t[6]
+ veor @x[1], @x[1], @t[6]
+ veor @x[2], @x[2], @t[0]
+ veor @x[4], @x[4], @t[2]
+ veor @x[3], @x[3], @t[1]
+ veor @x[1], @x[1], @t[7]
+ veor @x[2], @x[2], @t[7]
+ veor @x[4], @x[4], @t[6]
+ veor @x[5], @x[5], @t[3]
+ veor @x[3], @x[3], @t[6]
+ veor @x[6], @x[6], @t[4]
+ veor @x[4], @x[4], @t[7]
+ veor @x[5], @x[5], @t[7]
+ veor @x[7], @x[7], @t[5]
+___
+ &MixColumns (@x,@t,1); # flipped 2<->3 and 4<->6
+}
+
+sub swapmove {
+my ($a,$b,$n,$mask,$t)=@_;
+$code.=<<___;
+ vshr.u64 $t, $b, #$n
+ veor $t, $t, $a
+ vand $t, $t, $mask
+ veor $a, $a, $t
+ vshl.u64 $t, $t, #$n
+ veor $b, $b, $t
+___
+}
+sub swapmove2x {
+my ($a0,$b0,$a1,$b1,$n,$mask,$t0,$t1)=@_;
+$code.=<<___;
+ vshr.u64 $t0, $b0, #$n
+ vshr.u64 $t1, $b1, #$n
+ veor $t0, $t0, $a0
+ veor $t1, $t1, $a1
+ vand $t0, $t0, $mask
+ vand $t1, $t1, $mask
+ veor $a0, $a0, $t0
+ vshl.u64 $t0, $t0, #$n
+ veor $a1, $a1, $t1
+ vshl.u64 $t1, $t1, #$n
+ veor $b0, $b0, $t0
+ veor $b1, $b1, $t1
+___
+}
+
+sub bitslice {
+my @x=reverse(@_[0..7]);
+my ($t0,$t1,$t2,$t3)=@_[8..11];
+$code.=<<___;
+ vmov.i8 $t0,#0x55 @ compose .LBS0
+ vmov.i8 $t1,#0x33 @ compose .LBS1
+___
+ &swapmove2x(@x[0,1,2,3],1,$t0,$t2,$t3);
+ &swapmove2x(@x[4,5,6,7],1,$t0,$t2,$t3);
+$code.=<<___;
+ vmov.i8 $t0,#0x0f @ compose .LBS2
+___
+ &swapmove2x(@x[0,2,1,3],2,$t1,$t2,$t3);
+ &swapmove2x(@x[4,6,5,7],2,$t1,$t2,$t3);
+
+ &swapmove2x(@x[0,4,1,5],4,$t0,$t2,$t3);
+ &swapmove2x(@x[2,6,3,7],4,$t0,$t2,$t3);
+}
+
+$code.=<<___;
+#ifndef __KERNEL__
+# include "arm_arch.h"
+
+# define VFP_ABI_PUSH vstmdb sp!,{d8-d15}
+# define VFP_ABI_POP vldmia sp!,{d8-d15}
+# define VFP_ABI_FRAME 0x40
+#else
+# define VFP_ABI_PUSH
+# define VFP_ABI_POP
+# define VFP_ABI_FRAME 0
+# define BSAES_ASM_EXTENDED_KEY
+# define XTS_CHAIN_TWEAK
+# define __ARM_ARCH__ __LINUX_ARM_ARCH__
+#endif
+
+#ifdef __thumb__
+# define adrl adr
+#endif
+
+#if __ARM_ARCH__>=7
+.text
+.syntax unified @ ARMv7-capable assembler is expected to handle this
+#ifdef __thumb2__
+.thumb
+#else
+.code 32
+#endif
+
+.fpu neon
+
+.type _bsaes_decrypt8,%function
+.align 4
+_bsaes_decrypt8:
+ adr $const,_bsaes_decrypt8
+ vldmia $key!, {@XMM[9]} @ round 0 key
+ add $const,$const,#.LM0ISR-_bsaes_decrypt8
+
+ vldmia $const!, {@XMM[8]} @ .LM0ISR
+ veor @XMM[10], @XMM[0], @XMM[9] @ xor with round0 key
+ veor @XMM[11], @XMM[1], @XMM[9]
+ vtbl.8 `&Dlo(@XMM[0])`, {@XMM[10]}, `&Dlo(@XMM[8])`
+ vtbl.8 `&Dhi(@XMM[0])`, {@XMM[10]}, `&Dhi(@XMM[8])`
+ veor @XMM[12], @XMM[2], @XMM[9]
+ vtbl.8 `&Dlo(@XMM[1])`, {@XMM[11]}, `&Dlo(@XMM[8])`
+ vtbl.8 `&Dhi(@XMM[1])`, {@XMM[11]}, `&Dhi(@XMM[8])`
+ veor @XMM[13], @XMM[3], @XMM[9]
+ vtbl.8 `&Dlo(@XMM[2])`, {@XMM[12]}, `&Dlo(@XMM[8])`
+ vtbl.8 `&Dhi(@XMM[2])`, {@XMM[12]}, `&Dhi(@XMM[8])`
+ veor @XMM[14], @XMM[4], @XMM[9]
+ vtbl.8 `&Dlo(@XMM[3])`, {@XMM[13]}, `&Dlo(@XMM[8])`
+ vtbl.8 `&Dhi(@XMM[3])`, {@XMM[13]}, `&Dhi(@XMM[8])`
+ veor @XMM[15], @XMM[5], @XMM[9]
+ vtbl.8 `&Dlo(@XMM[4])`, {@XMM[14]}, `&Dlo(@XMM[8])`
+ vtbl.8 `&Dhi(@XMM[4])`, {@XMM[14]}, `&Dhi(@XMM[8])`
+ veor @XMM[10], @XMM[6], @XMM[9]
+ vtbl.8 `&Dlo(@XMM[5])`, {@XMM[15]}, `&Dlo(@XMM[8])`
+ vtbl.8 `&Dhi(@XMM[5])`, {@XMM[15]}, `&Dhi(@XMM[8])`
+ veor @XMM[11], @XMM[7], @XMM[9]
+ vtbl.8 `&Dlo(@XMM[6])`, {@XMM[10]}, `&Dlo(@XMM[8])`
+ vtbl.8 `&Dhi(@XMM[6])`, {@XMM[10]}, `&Dhi(@XMM[8])`
+ vtbl.8 `&Dlo(@XMM[7])`, {@XMM[11]}, `&Dlo(@XMM[8])`
+ vtbl.8 `&Dhi(@XMM[7])`, {@XMM[11]}, `&Dhi(@XMM[8])`
+___
+ &bitslice (@XMM[0..7, 8..11]);
+$code.=<<___;
+ sub $rounds,$rounds,#1
+ b .Ldec_sbox
+.align 4
+.Ldec_loop:
+___
+ &ShiftRows (@XMM[0..7, 8..12]);
+$code.=".Ldec_sbox:\n";
+ &InvSbox (@XMM[0..7, 8..15]);
+$code.=<<___;
+ subs $rounds,$rounds,#1
+ bcc .Ldec_done
+___
+ &InvMixColumns (@XMM[0,1,6,4,2,7,3,5, 8..15]);
+$code.=<<___;
+ vldmia $const, {@XMM[12]} @ .LISR
+ ite eq @ Thumb2 thing, sanity check in ARM
+ addeq $const,$const,#0x10
+ bne .Ldec_loop
+ vldmia $const, {@XMM[12]} @ .LISRM0
+ b .Ldec_loop
+.align 4
+.Ldec_done:
+___
+ &bitslice (@XMM[0,1,6,4,2,7,3,5, 8..11]);
+$code.=<<___;
+ vldmia $key, {@XMM[8]} @ last round key
+ veor @XMM[6], @XMM[6], @XMM[8]
+ veor @XMM[4], @XMM[4], @XMM[8]
+ veor @XMM[2], @XMM[2], @XMM[8]
+ veor @XMM[7], @XMM[7], @XMM[8]
+ veor @XMM[3], @XMM[3], @XMM[8]
+ veor @XMM[5], @XMM[5], @XMM[8]
+ veor @XMM[0], @XMM[0], @XMM[8]
+ veor @XMM[1], @XMM[1], @XMM[8]
+ bx lr
+.size _bsaes_decrypt8,.-_bsaes_decrypt8
+
+.type _bsaes_const,%object
+.align 6
+_bsaes_const:
+.LM0ISR: @ InvShiftRows constants
+ .quad 0x0a0e0206070b0f03, 0x0004080c0d010509
+.LISR:
+ .quad 0x0504070602010003, 0x0f0e0d0c080b0a09
+.LISRM0:
+ .quad 0x01040b0e0205080f, 0x0306090c00070a0d
+.LM0SR: @ ShiftRows constants
+ .quad 0x0a0e02060f03070b, 0x0004080c05090d01
+.LSR:
+ .quad 0x0504070600030201, 0x0f0e0d0c0a09080b
+.LSRM0:
+ .quad 0x0304090e00050a0f, 0x01060b0c0207080d
+.LM0:
+ .quad 0x02060a0e03070b0f, 0x0004080c0105090d
+.LREVM0SR:
+ .quad 0x090d01050c000408, 0x03070b0f060a0e02
+.asciz "Bit-sliced AES for NEON, CRYPTOGAMS by <appro\@openssl.org>"
+.align 6
+.size _bsaes_const,.-_bsaes_const
+
+.type _bsaes_encrypt8,%function
+.align 4
+_bsaes_encrypt8:
+ adr $const,_bsaes_encrypt8
+ vldmia $key!, {@XMM[9]} @ round 0 key
+ sub $const,$const,#_bsaes_encrypt8-.LM0SR
+
+ vldmia $const!, {@XMM[8]} @ .LM0SR
+_bsaes_encrypt8_alt:
+ veor @XMM[10], @XMM[0], @XMM[9] @ xor with round0 key
+ veor @XMM[11], @XMM[1], @XMM[9]
+ vtbl.8 `&Dlo(@XMM[0])`, {@XMM[10]}, `&Dlo(@XMM[8])`
+ vtbl.8 `&Dhi(@XMM[0])`, {@XMM[10]}, `&Dhi(@XMM[8])`
+ veor @XMM[12], @XMM[2], @XMM[9]
+ vtbl.8 `&Dlo(@XMM[1])`, {@XMM[11]}, `&Dlo(@XMM[8])`
+ vtbl.8 `&Dhi(@XMM[1])`, {@XMM[11]}, `&Dhi(@XMM[8])`
+ veor @XMM[13], @XMM[3], @XMM[9]
+ vtbl.8 `&Dlo(@XMM[2])`, {@XMM[12]}, `&Dlo(@XMM[8])`
+ vtbl.8 `&Dhi(@XMM[2])`, {@XMM[12]}, `&Dhi(@XMM[8])`
+ veor @XMM[14], @XMM[4], @XMM[9]
+ vtbl.8 `&Dlo(@XMM[3])`, {@XMM[13]}, `&Dlo(@XMM[8])`
+ vtbl.8 `&Dhi(@XMM[3])`, {@XMM[13]}, `&Dhi(@XMM[8])`
+ veor @XMM[15], @XMM[5], @XMM[9]
+ vtbl.8 `&Dlo(@XMM[4])`, {@XMM[14]}, `&Dlo(@XMM[8])`
+ vtbl.8 `&Dhi(@XMM[4])`, {@XMM[14]}, `&Dhi(@XMM[8])`
+ veor @XMM[10], @XMM[6], @XMM[9]
+ vtbl.8 `&Dlo(@XMM[5])`, {@XMM[15]}, `&Dlo(@XMM[8])`
+ vtbl.8 `&Dhi(@XMM[5])`, {@XMM[15]}, `&Dhi(@XMM[8])`
+ veor @XMM[11], @XMM[7], @XMM[9]
+ vtbl.8 `&Dlo(@XMM[6])`, {@XMM[10]}, `&Dlo(@XMM[8])`
+ vtbl.8 `&Dhi(@XMM[6])`, {@XMM[10]}, `&Dhi(@XMM[8])`
+ vtbl.8 `&Dlo(@XMM[7])`, {@XMM[11]}, `&Dlo(@XMM[8])`
+ vtbl.8 `&Dhi(@XMM[7])`, {@XMM[11]}, `&Dhi(@XMM[8])`
+_bsaes_encrypt8_bitslice:
+___
+ &bitslice (@XMM[0..7, 8..11]);
+$code.=<<___;
+ sub $rounds,$rounds,#1
+ b .Lenc_sbox
+.align 4
+.Lenc_loop:
+___
+ &ShiftRows (@XMM[0..7, 8..12]);
+$code.=".Lenc_sbox:\n";
+ &Sbox (@XMM[0..7, 8..15]);
+$code.=<<___;
+ subs $rounds,$rounds,#1
+ bcc .Lenc_done
+___
+ &MixColumns (@XMM[0,1,4,6,3,7,2,5, 8..15]);
+$code.=<<___;
+ vldmia $const, {@XMM[12]} @ .LSR
+ ite eq @ Thumb2 thing, samity check in ARM
+ addeq $const,$const,#0x10
+ bne .Lenc_loop
+ vldmia $const, {@XMM[12]} @ .LSRM0
+ b .Lenc_loop
+.align 4
+.Lenc_done:
+___
+ # output in lsb > [t0, t1, t4, t6, t3, t7, t2, t5] < msb
+ &bitslice (@XMM[0,1,4,6,3,7,2,5, 8..11]);
+$code.=<<___;
+ vldmia $key, {@XMM[8]} @ last round key
+ veor @XMM[4], @XMM[4], @XMM[8]
+ veor @XMM[6], @XMM[6], @XMM[8]
+ veor @XMM[3], @XMM[3], @XMM[8]
+ veor @XMM[7], @XMM[7], @XMM[8]
+ veor @XMM[2], @XMM[2], @XMM[8]
+ veor @XMM[5], @XMM[5], @XMM[8]
+ veor @XMM[0], @XMM[0], @XMM[8]
+ veor @XMM[1], @XMM[1], @XMM[8]
+ bx lr
+.size _bsaes_encrypt8,.-_bsaes_encrypt8
+___
+}
+{
+my ($out,$inp,$rounds,$const)=("r12","r4","r5","r6");
+
+sub bitslice_key {
+my @x=reverse(@_[0..7]);
+my ($bs0,$bs1,$bs2,$t2,$t3)=@_[8..12];
+
+ &swapmove (@x[0,1],1,$bs0,$t2,$t3);
+$code.=<<___;
+ @ &swapmove(@x[2,3],1,$t0,$t2,$t3);
+ vmov @x[2], @x[0]
+ vmov @x[3], @x[1]
+___
+ #&swapmove2x(@x[4,5,6,7],1,$t0,$t2,$t3);
+
+ &swapmove2x (@x[0,2,1,3],2,$bs1,$t2,$t3);
+$code.=<<___;
+ @ &swapmove2x(@x[4,6,5,7],2,$t1,$t2,$t3);
+ vmov @x[4], @x[0]
+ vmov @x[6], @x[2]
+ vmov @x[5], @x[1]
+ vmov @x[7], @x[3]
+___
+ &swapmove2x (@x[0,4,1,5],4,$bs2,$t2,$t3);
+ &swapmove2x (@x[2,6,3,7],4,$bs2,$t2,$t3);
+}
+
+$code.=<<___;
+.type _bsaes_key_convert,%function
+.align 4
+_bsaes_key_convert:
+ adr $const,_bsaes_key_convert
+ vld1.8 {@XMM[7]}, [$inp]! @ load round 0 key
+ sub $const,$const,#_bsaes_key_convert-.LM0
+ vld1.8 {@XMM[15]}, [$inp]! @ load round 1 key
+
+ vmov.i8 @XMM[8], #0x01 @ bit masks
+ vmov.i8 @XMM[9], #0x02
+ vmov.i8 @XMM[10], #0x04
+ vmov.i8 @XMM[11], #0x08
+ vmov.i8 @XMM[12], #0x10
+ vmov.i8 @XMM[13], #0x20
+ vldmia $const, {@XMM[14]} @ .LM0
+
+#ifdef __ARMEL__
+ vrev32.8 @XMM[7], @XMM[7]
+ vrev32.8 @XMM[15], @XMM[15]
+#endif
+ sub $rounds,$rounds,#1
+ vstmia $out!, {@XMM[7]} @ save round 0 key
+ b .Lkey_loop
+
+.align 4
+.Lkey_loop:
+ vtbl.8 `&Dlo(@XMM[7])`,{@XMM[15]},`&Dlo(@XMM[14])`
+ vtbl.8 `&Dhi(@XMM[7])`,{@XMM[15]},`&Dhi(@XMM[14])`
+ vmov.i8 @XMM[6], #0x40
+ vmov.i8 @XMM[15], #0x80
+
+ vtst.8 @XMM[0], @XMM[7], @XMM[8]
+ vtst.8 @XMM[1], @XMM[7], @XMM[9]
+ vtst.8 @XMM[2], @XMM[7], @XMM[10]
+ vtst.8 @XMM[3], @XMM[7], @XMM[11]
+ vtst.8 @XMM[4], @XMM[7], @XMM[12]
+ vtst.8 @XMM[5], @XMM[7], @XMM[13]
+ vtst.8 @XMM[6], @XMM[7], @XMM[6]
+ vtst.8 @XMM[7], @XMM[7], @XMM[15]
+ vld1.8 {@XMM[15]}, [$inp]! @ load next round key
+ vmvn @XMM[0], @XMM[0] @ "pnot"
+ vmvn @XMM[1], @XMM[1]
+ vmvn @XMM[5], @XMM[5]
+ vmvn @XMM[6], @XMM[6]
+#ifdef __ARMEL__
+ vrev32.8 @XMM[15], @XMM[15]
+#endif
+ subs $rounds,$rounds,#1
+ vstmia $out!,{@XMM[0]-@XMM[7]} @ write bit-sliced round key
+ bne .Lkey_loop
+
+ vmov.i8 @XMM[7],#0x63 @ compose .L63
+ @ don't save last round key
+ bx lr
+.size _bsaes_key_convert,.-_bsaes_key_convert
+___
+}
+
+if (0) { # following four functions are unsupported interface
+ # used for benchmarking...
+$code.=<<___;
+.globl bsaes_enc_key_convert
+.type bsaes_enc_key_convert,%function
+.align 4
+bsaes_enc_key_convert:
+ stmdb sp!,{r4-r6,lr}
+ vstmdb sp!,{d8-d15} @ ABI specification says so
+
+ ldr r5,[$inp,#240] @ pass rounds
+ mov r4,$inp @ pass key
+ mov r12,$out @ pass key schedule
+ bl _bsaes_key_convert
+ veor @XMM[7],@XMM[7],@XMM[15] @ fix up last round key
+ vstmia r12, {@XMM[7]} @ save last round key
+
+ vldmia sp!,{d8-d15}
+ ldmia sp!,{r4-r6,pc}
+.size bsaes_enc_key_convert,.-bsaes_enc_key_convert
+
+.globl bsaes_encrypt_128
+.type bsaes_encrypt_128,%function
+.align 4
+bsaes_encrypt_128:
+ stmdb sp!,{r4-r6,lr}
+ vstmdb sp!,{d8-d15} @ ABI specification says so
+.Lenc128_loop:
+ vld1.8 {@XMM[0]-@XMM[1]}, [$inp]! @ load input
+ vld1.8 {@XMM[2]-@XMM[3]}, [$inp]!
+ mov r4,$key @ pass the key
+ vld1.8 {@XMM[4]-@XMM[5]}, [$inp]!
+ mov r5,#10 @ pass rounds
+ vld1.8 {@XMM[6]-@XMM[7]}, [$inp]!
+
+ bl _bsaes_encrypt8
+
+ vst1.8 {@XMM[0]-@XMM[1]}, [$out]! @ write output
+ vst1.8 {@XMM[4]}, [$out]!
+ vst1.8 {@XMM[6]}, [$out]!
+ vst1.8 {@XMM[3]}, [$out]!
+ vst1.8 {@XMM[7]}, [$out]!
+ vst1.8 {@XMM[2]}, [$out]!
+ subs $len,$len,#0x80
+ vst1.8 {@XMM[5]}, [$out]!
+ bhi .Lenc128_loop
+
+ vldmia sp!,{d8-d15}
+ ldmia sp!,{r4-r6,pc}
+.size bsaes_encrypt_128,.-bsaes_encrypt_128
+
+.globl bsaes_dec_key_convert
+.type bsaes_dec_key_convert,%function
+.align 4
+bsaes_dec_key_convert:
+ stmdb sp!,{r4-r6,lr}
+ vstmdb sp!,{d8-d15} @ ABI specification says so
+
+ ldr r5,[$inp,#240] @ pass rounds
+ mov r4,$inp @ pass key
+ mov r12,$out @ pass key schedule
+ bl _bsaes_key_convert
+ vldmia $out, {@XMM[6]}
+ vstmia r12, {@XMM[15]} @ save last round key
+ veor @XMM[7], @XMM[7], @XMM[6] @ fix up round 0 key
+ vstmia $out, {@XMM[7]}
+
+ vldmia sp!,{d8-d15}
+ ldmia sp!,{r4-r6,pc}
+.size bsaes_dec_key_convert,.-bsaes_dec_key_convert
+
+.globl bsaes_decrypt_128
+.type bsaes_decrypt_128,%function
+.align 4
+bsaes_decrypt_128:
+ stmdb sp!,{r4-r6,lr}
+ vstmdb sp!,{d8-d15} @ ABI specification says so
+.Ldec128_loop:
+ vld1.8 {@XMM[0]-@XMM[1]}, [$inp]! @ load input
+ vld1.8 {@XMM[2]-@XMM[3]}, [$inp]!
+ mov r4,$key @ pass the key
+ vld1.8 {@XMM[4]-@XMM[5]}, [$inp]!
+ mov r5,#10 @ pass rounds
+ vld1.8 {@XMM[6]-@XMM[7]}, [$inp]!
+
+ bl _bsaes_decrypt8
+
+ vst1.8 {@XMM[0]-@XMM[1]}, [$out]! @ write output
+ vst1.8 {@XMM[6]}, [$out]!
+ vst1.8 {@XMM[4]}, [$out]!
+ vst1.8 {@XMM[2]}, [$out]!
+ vst1.8 {@XMM[7]}, [$out]!
+ vst1.8 {@XMM[3]}, [$out]!
+ subs $len,$len,#0x80
+ vst1.8 {@XMM[5]}, [$out]!
+ bhi .Ldec128_loop
+
+ vldmia sp!,{d8-d15}
+ ldmia sp!,{r4-r6,pc}
+.size bsaes_decrypt_128,.-bsaes_decrypt_128
+___
+}
+{
+my ($inp,$out,$len,$key, $ivp,$fp,$rounds)=map("r$_",(0..3,8..10));
+my ($keysched)=("sp");
+
+$code.=<<___;
+.extern AES_cbc_encrypt
+.extern AES_decrypt
+
+.global bsaes_cbc_encrypt
+.type bsaes_cbc_encrypt,%function
+.align 5
+bsaes_cbc_encrypt:
+#ifndef __KERNEL__
+ cmp $len, #128
+#ifndef __thumb__
+ blo AES_cbc_encrypt
+#else
+ bhs 1f
+ b AES_cbc_encrypt
+1:
+#endif
+#endif
+
+ @ it is up to the caller to make sure we are called with enc == 0
+
+ mov ip, sp
+ stmdb sp!, {r4-r10, lr}
+ VFP_ABI_PUSH
+ ldr $ivp, [ip] @ IV is 1st arg on the stack
+ mov $len, $len, lsr#4 @ len in 16 byte blocks
+ sub sp, #0x10 @ scratch space to carry over the IV
+ mov $fp, sp @ save sp
+
+ ldr $rounds, [$key, #240] @ get # of rounds
+#ifndef BSAES_ASM_EXTENDED_KEY
+ @ allocate the key schedule on the stack
+ sub r12, sp, $rounds, lsl#7 @ 128 bytes per inner round key
+ add r12, #`128-32` @ sifze of bit-slices key schedule
+
+ @ populate the key schedule
+ mov r4, $key @ pass key
+ mov r5, $rounds @ pass # of rounds
+ mov sp, r12 @ sp is $keysched
+ bl _bsaes_key_convert
+ vldmia $keysched, {@XMM[6]}
+ vstmia r12, {@XMM[15]} @ save last round key
+ veor @XMM[7], @XMM[7], @XMM[6] @ fix up round 0 key
+ vstmia $keysched, {@XMM[7]}
+#else
+ ldr r12, [$key, #244]
+ eors r12, #1
+ beq 0f
+
+ @ populate the key schedule
+ str r12, [$key, #244]
+ mov r4, $key @ pass key
+ mov r5, $rounds @ pass # of rounds
+ add r12, $key, #248 @ pass key schedule
+ bl _bsaes_key_convert
+ add r4, $key, #248
+ vldmia r4, {@XMM[6]}
+ vstmia r12, {@XMM[15]} @ save last round key
+ veor @XMM[7], @XMM[7], @XMM[6] @ fix up round 0 key
+ vstmia r4, {@XMM[7]}
+
+.align 2
+0:
+#endif
+
+ vld1.8 {@XMM[15]}, [$ivp] @ load IV
+ b .Lcbc_dec_loop
+
+.align 4
+.Lcbc_dec_loop:
+ subs $len, $len, #0x8
+ bmi .Lcbc_dec_loop_finish
+
+ vld1.8 {@XMM[0]-@XMM[1]}, [$inp]! @ load input
+ vld1.8 {@XMM[2]-@XMM[3]}, [$inp]!
+#ifndef BSAES_ASM_EXTENDED_KEY
+ mov r4, $keysched @ pass the key
+#else
+ add r4, $key, #248
+#endif
+ vld1.8 {@XMM[4]-@XMM[5]}, [$inp]!
+ mov r5, $rounds
+ vld1.8 {@XMM[6]-@XMM[7]}, [$inp]
+ sub $inp, $inp, #0x60
+ vstmia $fp, {@XMM[15]} @ put aside IV
+
+ bl _bsaes_decrypt8
+
+ vldmia $fp, {@XMM[14]} @ reload IV
+ vld1.8 {@XMM[8]-@XMM[9]}, [$inp]! @ reload input
+ veor @XMM[0], @XMM[0], @XMM[14] @ ^= IV
+ vld1.8 {@XMM[10]-@XMM[11]}, [$inp]!
+ veor @XMM[1], @XMM[1], @XMM[8]
+ veor @XMM[6], @XMM[6], @XMM[9]
+ vld1.8 {@XMM[12]-@XMM[13]}, [$inp]!
+ veor @XMM[4], @XMM[4], @XMM[10]
+ veor @XMM[2], @XMM[2], @XMM[11]
+ vld1.8 {@XMM[14]-@XMM[15]}, [$inp]!
+ veor @XMM[7], @XMM[7], @XMM[12]
+ vst1.8 {@XMM[0]-@XMM[1]}, [$out]! @ write output
+ veor @XMM[3], @XMM[3], @XMM[13]
+ vst1.8 {@XMM[6]}, [$out]!
+ veor @XMM[5], @XMM[5], @XMM[14]
+ vst1.8 {@XMM[4]}, [$out]!
+ vst1.8 {@XMM[2]}, [$out]!
+ vst1.8 {@XMM[7]}, [$out]!
+ vst1.8 {@XMM[3]}, [$out]!
+ vst1.8 {@XMM[5]}, [$out]!
+
+ b .Lcbc_dec_loop
+
+.Lcbc_dec_loop_finish:
+ adds $len, $len, #8
+ beq .Lcbc_dec_done
+
+ vld1.8 {@XMM[0]}, [$inp]! @ load input
+ cmp $len, #2
+ blo .Lcbc_dec_one
+ vld1.8 {@XMM[1]}, [$inp]!
+#ifndef BSAES_ASM_EXTENDED_KEY
+ mov r4, $keysched @ pass the key
+#else
+ add r4, $key, #248
+#endif
+ mov r5, $rounds
+ vstmia $fp, {@XMM[15]} @ put aside IV
+ beq .Lcbc_dec_two
+ vld1.8 {@XMM[2]}, [$inp]!
+ cmp $len, #4
+ blo .Lcbc_dec_three
+ vld1.8 {@XMM[3]}, [$inp]!
+ beq .Lcbc_dec_four
+ vld1.8 {@XMM[4]}, [$inp]!
+ cmp $len, #6
+ blo .Lcbc_dec_five
+ vld1.8 {@XMM[5]}, [$inp]!
+ beq .Lcbc_dec_six
+ vld1.8 {@XMM[6]}, [$inp]!
+ sub $inp, $inp, #0x70
+
+ bl _bsaes_decrypt8
+
+ vldmia $fp, {@XMM[14]} @ reload IV
+ vld1.8 {@XMM[8]-@XMM[9]}, [$inp]! @ reload input
+ veor @XMM[0], @XMM[0], @XMM[14] @ ^= IV
+ vld1.8 {@XMM[10]-@XMM[11]}, [$inp]!
+ veor @XMM[1], @XMM[1], @XMM[8]
+ veor @XMM[6], @XMM[6], @XMM[9]
+ vld1.8 {@XMM[12]-@XMM[13]}, [$inp]!
+ veor @XMM[4], @XMM[4], @XMM[10]
+ veor @XMM[2], @XMM[2], @XMM[11]
+ vld1.8 {@XMM[15]}, [$inp]!
+ veor @XMM[7], @XMM[7], @XMM[12]
+ vst1.8 {@XMM[0]-@XMM[1]}, [$out]! @ write output
+ veor @XMM[3], @XMM[3], @XMM[13]
+ vst1.8 {@XMM[6]}, [$out]!
+ vst1.8 {@XMM[4]}, [$out]!
+ vst1.8 {@XMM[2]}, [$out]!
+ vst1.8 {@XMM[7]}, [$out]!
+ vst1.8 {@XMM[3]}, [$out]!
+ b .Lcbc_dec_done
+.align 4
+.Lcbc_dec_six:
+ sub $inp, $inp, #0x60
+ bl _bsaes_decrypt8
+ vldmia $fp,{@XMM[14]} @ reload IV
+ vld1.8 {@XMM[8]-@XMM[9]}, [$inp]! @ reload input
+ veor @XMM[0], @XMM[0], @XMM[14] @ ^= IV
+ vld1.8 {@XMM[10]-@XMM[11]}, [$inp]!
+ veor @XMM[1], @XMM[1], @XMM[8]
+ veor @XMM[6], @XMM[6], @XMM[9]
+ vld1.8 {@XMM[12]}, [$inp]!
+ veor @XMM[4], @XMM[4], @XMM[10]
+ veor @XMM[2], @XMM[2], @XMM[11]
+ vld1.8 {@XMM[15]}, [$inp]!
+ veor @XMM[7], @XMM[7], @XMM[12]
+ vst1.8 {@XMM[0]-@XMM[1]}, [$out]! @ write output
+ vst1.8 {@XMM[6]}, [$out]!
+ vst1.8 {@XMM[4]}, [$out]!
+ vst1.8 {@XMM[2]}, [$out]!
+ vst1.8 {@XMM[7]}, [$out]!
+ b .Lcbc_dec_done
+.align 4
+.Lcbc_dec_five:
+ sub $inp, $inp, #0x50
+ bl _bsaes_decrypt8
+ vldmia $fp, {@XMM[14]} @ reload IV
+ vld1.8 {@XMM[8]-@XMM[9]}, [$inp]! @ reload input
+ veor @XMM[0], @XMM[0], @XMM[14] @ ^= IV
+ vld1.8 {@XMM[10]-@XMM[11]}, [$inp]!
+ veor @XMM[1], @XMM[1], @XMM[8]
+ veor @XMM[6], @XMM[6], @XMM[9]
+ vld1.8 {@XMM[15]}, [$inp]!
+ veor @XMM[4], @XMM[4], @XMM[10]
+ vst1.8 {@XMM[0]-@XMM[1]}, [$out]! @ write output
+ veor @XMM[2], @XMM[2], @XMM[11]
+ vst1.8 {@XMM[6]}, [$out]!
+ vst1.8 {@XMM[4]}, [$out]!
+ vst1.8 {@XMM[2]}, [$out]!
+ b .Lcbc_dec_done
+.align 4
+.Lcbc_dec_four:
+ sub $inp, $inp, #0x40
+ bl _bsaes_decrypt8
+ vldmia $fp, {@XMM[14]} @ reload IV
+ vld1.8 {@XMM[8]-@XMM[9]}, [$inp]! @ reload input
+ veor @XMM[0], @XMM[0], @XMM[14] @ ^= IV
+ vld1.8 {@XMM[10]}, [$inp]!
+ veor @XMM[1], @XMM[1], @XMM[8]
+ veor @XMM[6], @XMM[6], @XMM[9]
+ vld1.8 {@XMM[15]}, [$inp]!
+ veor @XMM[4], @XMM[4], @XMM[10]
+ vst1.8 {@XMM[0]-@XMM[1]}, [$out]! @ write output
+ vst1.8 {@XMM[6]}, [$out]!
+ vst1.8 {@XMM[4]}, [$out]!
+ b .Lcbc_dec_done
+.align 4
+.Lcbc_dec_three:
+ sub $inp, $inp, #0x30
+ bl _bsaes_decrypt8
+ vldmia $fp, {@XMM[14]} @ reload IV
+ vld1.8 {@XMM[8]-@XMM[9]}, [$inp]! @ reload input
+ veor @XMM[0], @XMM[0], @XMM[14] @ ^= IV
+ vld1.8 {@XMM[15]}, [$inp]!
+ veor @XMM[1], @XMM[1], @XMM[8]
+ veor @XMM[6], @XMM[6], @XMM[9]
+ vst1.8 {@XMM[0]-@XMM[1]}, [$out]! @ write output
+ vst1.8 {@XMM[6]}, [$out]!
+ b .Lcbc_dec_done
+.align 4
+.Lcbc_dec_two:
+ sub $inp, $inp, #0x20
+ bl _bsaes_decrypt8
+ vldmia $fp, {@XMM[14]} @ reload IV
+ vld1.8 {@XMM[8]}, [$inp]! @ reload input
+ veor @XMM[0], @XMM[0], @XMM[14] @ ^= IV
+ vld1.8 {@XMM[15]}, [$inp]! @ reload input
+ veor @XMM[1], @XMM[1], @XMM[8]
+ vst1.8 {@XMM[0]-@XMM[1]}, [$out]! @ write output
+ b .Lcbc_dec_done
+.align 4
+.Lcbc_dec_one:
+ sub $inp, $inp, #0x10
+ mov $rounds, $out @ save original out pointer
+ mov $out, $fp @ use the iv scratch space as out buffer
+ mov r2, $key
+ vmov @XMM[4],@XMM[15] @ just in case ensure that IV
+ vmov @XMM[5],@XMM[0] @ and input are preserved
+ bl AES_decrypt
+ vld1.8 {@XMM[0]}, [$fp,:64] @ load result
+ veor @XMM[0], @XMM[0], @XMM[4] @ ^= IV
+ vmov @XMM[15], @XMM[5] @ @XMM[5] holds input
+ vst1.8 {@XMM[0]}, [$rounds] @ write output
+
+.Lcbc_dec_done:
+#ifndef BSAES_ASM_EXTENDED_KEY
+ vmov.i32 q0, #0
+ vmov.i32 q1, #0
+.Lcbc_dec_bzero: @ wipe key schedule [if any]
+ vstmia $keysched!, {q0-q1}
+ cmp $keysched, $fp
+ bne .Lcbc_dec_bzero
+#endif
+
+ mov sp, $fp
+ add sp, #0x10 @ add sp,$fp,#0x10 is no good for thumb
+ vst1.8 {@XMM[15]}, [$ivp] @ return IV
+ VFP_ABI_POP
+ ldmia sp!, {r4-r10, pc}
+.size bsaes_cbc_encrypt,.-bsaes_cbc_encrypt
+___
+}
+{
+my ($inp,$out,$len,$key, $ctr,$fp,$rounds)=(map("r$_",(0..3,8..10)));
+my $const = "r6"; # shared with _bsaes_encrypt8_alt
+my $keysched = "sp";
+
+$code.=<<___;
+.extern AES_encrypt
+.global bsaes_ctr32_encrypt_blocks
+.type bsaes_ctr32_encrypt_blocks,%function
+.align 5
+bsaes_ctr32_encrypt_blocks:
+ cmp $len, #8 @ use plain AES for
+ blo .Lctr_enc_short @ small sizes
+
+ mov ip, sp
+ stmdb sp!, {r4-r10, lr}
+ VFP_ABI_PUSH
+ ldr $ctr, [ip] @ ctr is 1st arg on the stack
+ sub sp, sp, #0x10 @ scratch space to carry over the ctr
+ mov $fp, sp @ save sp
+
+ ldr $rounds, [$key, #240] @ get # of rounds
+#ifndef BSAES_ASM_EXTENDED_KEY
+ @ allocate the key schedule on the stack
+ sub r12, sp, $rounds, lsl#7 @ 128 bytes per inner round key
+ add r12, #`128-32` @ size of bit-sliced key schedule
+
+ @ populate the key schedule
+ mov r4, $key @ pass key
+ mov r5, $rounds @ pass # of rounds
+ mov sp, r12 @ sp is $keysched
+ bl _bsaes_key_convert
+ veor @XMM[7],@XMM[7],@XMM[15] @ fix up last round key
+ vstmia r12, {@XMM[7]} @ save last round key
+
+ vld1.8 {@XMM[0]}, [$ctr] @ load counter
+ add $ctr, $const, #.LREVM0SR-.LM0 @ borrow $ctr
+ vldmia $keysched, {@XMM[4]} @ load round0 key
+#else
+ ldr r12, [$key, #244]
+ eors r12, #1
+ beq 0f
+
+ @ populate the key schedule
+ str r12, [$key, #244]
+ mov r4, $key @ pass key
+ mov r5, $rounds @ pass # of rounds
+ add r12, $key, #248 @ pass key schedule
+ bl _bsaes_key_convert
+ veor @XMM[7],@XMM[7],@XMM[15] @ fix up last round key
+ vstmia r12, {@XMM[7]} @ save last round key
+
+.align 2
+0: add r12, $key, #248
+ vld1.8 {@XMM[0]}, [$ctr] @ load counter
+ adrl $ctr, .LREVM0SR @ borrow $ctr
+ vldmia r12, {@XMM[4]} @ load round0 key
+ sub sp, #0x10 @ place for adjusted round0 key
+#endif
+
+ vmov.i32 @XMM[8],#1 @ compose 1<<96
+ veor @XMM[9],@XMM[9],@XMM[9]
+ vrev32.8 @XMM[0],@XMM[0]
+ vext.8 @XMM[8],@XMM[9],@XMM[8],#4
+ vrev32.8 @XMM[4],@XMM[4]
+ vadd.u32 @XMM[9],@XMM[8],@XMM[8] @ compose 2<<96
+ vstmia $keysched, {@XMM[4]} @ save adjusted round0 key
+ b .Lctr_enc_loop
+
+.align 4
+.Lctr_enc_loop:
+ vadd.u32 @XMM[10], @XMM[8], @XMM[9] @ compose 3<<96
+ vadd.u32 @XMM[1], @XMM[0], @XMM[8] @ +1
+ vadd.u32 @XMM[2], @XMM[0], @XMM[9] @ +2
+ vadd.u32 @XMM[3], @XMM[0], @XMM[10] @ +3
+ vadd.u32 @XMM[4], @XMM[1], @XMM[10]
+ vadd.u32 @XMM[5], @XMM[2], @XMM[10]
+ vadd.u32 @XMM[6], @XMM[3], @XMM[10]
+ vadd.u32 @XMM[7], @XMM[4], @XMM[10]
+ vadd.u32 @XMM[10], @XMM[5], @XMM[10] @ next counter
+
+ @ Borrow prologue from _bsaes_encrypt8 to use the opportunity
+ @ to flip byte order in 32-bit counter
+
+ vldmia $keysched, {@XMM[9]} @ load round0 key
+#ifndef BSAES_ASM_EXTENDED_KEY
+ add r4, $keysched, #0x10 @ pass next round key
+#else
+ add r4, $key, #`248+16`
+#endif
+ vldmia $ctr, {@XMM[8]} @ .LREVM0SR
+ mov r5, $rounds @ pass rounds
+ vstmia $fp, {@XMM[10]} @ save next counter
+ sub $const, $ctr, #.LREVM0SR-.LSR @ pass constants
+
+ bl _bsaes_encrypt8_alt
+
+ subs $len, $len, #8
+ blo .Lctr_enc_loop_done
+
+ vld1.8 {@XMM[8]-@XMM[9]}, [$inp]! @ load input
+ vld1.8 {@XMM[10]-@XMM[11]}, [$inp]!
+ veor @XMM[0], @XMM[8]
+ veor @XMM[1], @XMM[9]
+ vld1.8 {@XMM[12]-@XMM[13]}, [$inp]!
+ veor @XMM[4], @XMM[10]
+ veor @XMM[6], @XMM[11]
+ vld1.8 {@XMM[14]-@XMM[15]}, [$inp]!
+ veor @XMM[3], @XMM[12]
+ vst1.8 {@XMM[0]-@XMM[1]}, [$out]! @ write output
+ veor @XMM[7], @XMM[13]
+ veor @XMM[2], @XMM[14]
+ vst1.8 {@XMM[4]}, [$out]!
+ veor @XMM[5], @XMM[15]
+ vst1.8 {@XMM[6]}, [$out]!
+ vmov.i32 @XMM[8], #1 @ compose 1<<96
+ vst1.8 {@XMM[3]}, [$out]!
+ veor @XMM[9], @XMM[9], @XMM[9]
+ vst1.8 {@XMM[7]}, [$out]!
+ vext.8 @XMM[8], @XMM[9], @XMM[8], #4
+ vst1.8 {@XMM[2]}, [$out]!
+ vadd.u32 @XMM[9],@XMM[8],@XMM[8] @ compose 2<<96
+ vst1.8 {@XMM[5]}, [$out]!
+ vldmia $fp, {@XMM[0]} @ load counter
+
+ bne .Lctr_enc_loop
+ b .Lctr_enc_done
+
+.align 4
+.Lctr_enc_loop_done:
+ add $len, $len, #8
+ vld1.8 {@XMM[8]}, [$inp]! @ load input
+ veor @XMM[0], @XMM[8]
+ vst1.8 {@XMM[0]}, [$out]! @ write output
+ cmp $len, #2
+ blo .Lctr_enc_done
+ vld1.8 {@XMM[9]}, [$inp]!
+ veor @XMM[1], @XMM[9]
+ vst1.8 {@XMM[1]}, [$out]!
+ beq .Lctr_enc_done
+ vld1.8 {@XMM[10]}, [$inp]!
+ veor @XMM[4], @XMM[10]
+ vst1.8 {@XMM[4]}, [$out]!
+ cmp $len, #4
+ blo .Lctr_enc_done
+ vld1.8 {@XMM[11]}, [$inp]!
+ veor @XMM[6], @XMM[11]
+ vst1.8 {@XMM[6]}, [$out]!
+ beq .Lctr_enc_done
+ vld1.8 {@XMM[12]}, [$inp]!
+ veor @XMM[3], @XMM[12]
+ vst1.8 {@XMM[3]}, [$out]!
+ cmp $len, #6
+ blo .Lctr_enc_done
+ vld1.8 {@XMM[13]}, [$inp]!
+ veor @XMM[7], @XMM[13]
+ vst1.8 {@XMM[7]}, [$out]!
+ beq .Lctr_enc_done
+ vld1.8 {@XMM[14]}, [$inp]
+ veor @XMM[2], @XMM[14]
+ vst1.8 {@XMM[2]}, [$out]!
+
+.Lctr_enc_done:
+ vmov.i32 q0, #0
+ vmov.i32 q1, #0
+#ifndef BSAES_ASM_EXTENDED_KEY
+.Lctr_enc_bzero: @ wipe key schedule [if any]
+ vstmia $keysched!, {q0-q1}
+ cmp $keysched, $fp
+ bne .Lctr_enc_bzero
+#else
+ vstmia $keysched, {q0-q1}
+#endif
+
+ mov sp, $fp
+ add sp, #0x10 @ add sp,$fp,#0x10 is no good for thumb
+ VFP_ABI_POP
+ ldmia sp!, {r4-r10, pc} @ return
+
+.align 4
+.Lctr_enc_short:
+ ldr ip, [sp] @ ctr pointer is passed on stack
+ stmdb sp!, {r4-r8, lr}
+
+ mov r4, $inp @ copy arguments
+ mov r5, $out
+ mov r6, $len
+ mov r7, $key
+ ldr r8, [ip, #12] @ load counter LSW
+ vld1.8 {@XMM[1]}, [ip] @ load whole counter value
+#ifdef __ARMEL__
+ rev r8, r8
+#endif
+ sub sp, sp, #0x10
+ vst1.8 {@XMM[1]}, [sp,:64] @ copy counter value
+ sub sp, sp, #0x10
+
+.Lctr_enc_short_loop:
+ add r0, sp, #0x10 @ input counter value
+ mov r1, sp @ output on the stack
+ mov r2, r7 @ key
+
+ bl AES_encrypt
+
+ vld1.8 {@XMM[0]}, [r4]! @ load input
+ vld1.8 {@XMM[1]}, [sp,:64] @ load encrypted counter
+ add r8, r8, #1
+#ifdef __ARMEL__
+ rev r0, r8
+ str r0, [sp, #0x1c] @ next counter value
+#else
+ str r8, [sp, #0x1c] @ next counter value
+#endif
+ veor @XMM[0],@XMM[0],@XMM[1]
+ vst1.8 {@XMM[0]}, [r5]! @ store output
+ subs r6, r6, #1
+ bne .Lctr_enc_short_loop
+
+ vmov.i32 q0, #0
+ vmov.i32 q1, #0
+ vstmia sp!, {q0-q1}
+
+ ldmia sp!, {r4-r8, pc}
+.size bsaes_ctr32_encrypt_blocks,.-bsaes_ctr32_encrypt_blocks
+___
+}
+{
+######################################################################
+# void bsaes_xts_[en|de]crypt(const char *inp,char *out,size_t len,
+# const AES_KEY *key1, const AES_KEY *key2,
+# const unsigned char iv[16]);
+#
+my ($inp,$out,$len,$key,$rounds,$magic,$fp)=(map("r$_",(7..10,1..3)));
+my $const="r6"; # returned by _bsaes_key_convert
+my $twmask=@XMM[5];
+my @T=@XMM[6..7];
+
+$code.=<<___;
+.globl bsaes_xts_encrypt
+.type bsaes_xts_encrypt,%function
+.align 4
+bsaes_xts_encrypt:
+ mov ip, sp
+ stmdb sp!, {r4-r10, lr} @ 0x20
+ VFP_ABI_PUSH
+ mov r6, sp @ future $fp
+
+ mov $inp, r0
+ mov $out, r1
+ mov $len, r2
+ mov $key, r3
+
+ sub r0, sp, #0x10 @ 0x10
+ bic r0, #0xf @ align at 16 bytes
+ mov sp, r0
+
+#ifdef XTS_CHAIN_TWEAK
+ ldr r0, [ip] @ pointer to input tweak
+#else
+ @ generate initial tweak
+ ldr r0, [ip, #4] @ iv[]
+ mov r1, sp
+ ldr r2, [ip, #0] @ key2
+ bl AES_encrypt
+ mov r0,sp @ pointer to initial tweak
+#endif
+
+ ldr $rounds, [$key, #240] @ get # of rounds
+ mov $fp, r6
+#ifndef BSAES_ASM_EXTENDED_KEY
+ @ allocate the key schedule on the stack
+ sub r12, sp, $rounds, lsl#7 @ 128 bytes per inner round key
+ @ add r12, #`128-32` @ size of bit-sliced key schedule
+ sub r12, #`32+16` @ place for tweak[9]
+
+ @ populate the key schedule
+ mov r4, $key @ pass key
+ mov r5, $rounds @ pass # of rounds
+ mov sp, r12
+ add r12, #0x90 @ pass key schedule
+ bl _bsaes_key_convert
+ veor @XMM[7], @XMM[7], @XMM[15] @ fix up last round key
+ vstmia r12, {@XMM[7]} @ save last round key
+#else
+ ldr r12, [$key, #244]
+ eors r12, #1
+ beq 0f
+
+ str r12, [$key, #244]
+ mov r4, $key @ pass key
+ mov r5, $rounds @ pass # of rounds
+ add r12, $key, #248 @ pass key schedule
+ bl _bsaes_key_convert
+ veor @XMM[7], @XMM[7], @XMM[15] @ fix up last round key
+ vstmia r12, {@XMM[7]}
+
+.align 2
+0: sub sp, #0x90 @ place for tweak[9]
+#endif
+
+ vld1.8 {@XMM[8]}, [r0] @ initial tweak
+ adr $magic, .Lxts_magic
+
+ subs $len, #0x80
+ blo .Lxts_enc_short
+ b .Lxts_enc_loop
+
+.align 4
+.Lxts_enc_loop:
+ vldmia $magic, {$twmask} @ load XTS magic
+ vshr.s64 @T[0], @XMM[8], #63
+ mov r0, sp
+ vand @T[0], @T[0], $twmask
+___
+for($i=9;$i<16;$i++) {
+$code.=<<___;
+ vadd.u64 @XMM[$i], @XMM[$i-1], @XMM[$i-1]
+ vst1.64 {@XMM[$i-1]}, [r0,:128]!
+ vswp `&Dhi("@T[0]")`,`&Dlo("@T[0]")`
+ vshr.s64 @T[1], @XMM[$i], #63
+ veor @XMM[$i], @XMM[$i], @T[0]
+ vand @T[1], @T[1], $twmask
+___
+ @T=reverse(@T);
+
+$code.=<<___ if ($i>=10);
+ vld1.8 {@XMM[$i-10]}, [$inp]!
+___
+$code.=<<___ if ($i>=11);
+ veor @XMM[$i-11], @XMM[$i-11], @XMM[$i-3]
+___
+}
+$code.=<<___;
+ vadd.u64 @XMM[8], @XMM[15], @XMM[15]
+ vst1.64 {@XMM[15]}, [r0,:128]!
+ vswp `&Dhi("@T[0]")`,`&Dlo("@T[0]")`
+ veor @XMM[8], @XMM[8], @T[0]
+ vst1.64 {@XMM[8]}, [r0,:128] @ next round tweak
+
+ vld1.8 {@XMM[6]-@XMM[7]}, [$inp]!
+ veor @XMM[5], @XMM[5], @XMM[13]
+#ifndef BSAES_ASM_EXTENDED_KEY
+ add r4, sp, #0x90 @ pass key schedule
+#else
+ add r4, $key, #248 @ pass key schedule
+#endif
+ veor @XMM[6], @XMM[6], @XMM[14]
+ mov r5, $rounds @ pass rounds
+ veor @XMM[7], @XMM[7], @XMM[15]
+ mov r0, sp
+
+ bl _bsaes_encrypt8
+
+ vld1.64 {@XMM[ 8]-@XMM[ 9]}, [r0,:128]!
+ vld1.64 {@XMM[10]-@XMM[11]}, [r0,:128]!
+ veor @XMM[0], @XMM[0], @XMM[ 8]
+ vld1.64 {@XMM[12]-@XMM[13]}, [r0,:128]!
+ veor @XMM[1], @XMM[1], @XMM[ 9]
+ veor @XMM[8], @XMM[4], @XMM[10]
+ vst1.8 {@XMM[0]-@XMM[1]}, [$out]!
+ veor @XMM[9], @XMM[6], @XMM[11]
+ vld1.64 {@XMM[14]-@XMM[15]}, [r0,:128]!
+ veor @XMM[10], @XMM[3], @XMM[12]
+ vst1.8 {@XMM[8]-@XMM[9]}, [$out]!
+ veor @XMM[11], @XMM[7], @XMM[13]
+ veor @XMM[12], @XMM[2], @XMM[14]
+ vst1.8 {@XMM[10]-@XMM[11]}, [$out]!
+ veor @XMM[13], @XMM[5], @XMM[15]
+ vst1.8 {@XMM[12]-@XMM[13]}, [$out]!
+
+ vld1.64 {@XMM[8]}, [r0,:128] @ next round tweak
+
+ subs $len, #0x80
+ bpl .Lxts_enc_loop
+
+.Lxts_enc_short:
+ adds $len, #0x70
+ bmi .Lxts_enc_done
+
+ vldmia $magic, {$twmask} @ load XTS magic
+ vshr.s64 @T[0], @XMM[8], #63
+ mov r0, sp
+ vand @T[0], @T[0], $twmask
+___
+for($i=9;$i<16;$i++) {
+$code.=<<___;
+ vadd.u64 @XMM[$i], @XMM[$i-1], @XMM[$i-1]
+ vst1.64 {@XMM[$i-1]}, [r0,:128]!
+ vswp `&Dhi("@T[0]")`,`&Dlo("@T[0]")`
+ vshr.s64 @T[1], @XMM[$i], #63
+ veor @XMM[$i], @XMM[$i], @T[0]
+ vand @T[1], @T[1], $twmask
+___
+ @T=reverse(@T);
+
+$code.=<<___ if ($i>=10);
+ vld1.8 {@XMM[$i-10]}, [$inp]!
+ subs $len, #0x10
+ bmi .Lxts_enc_`$i-9`
+___
+$code.=<<___ if ($i>=11);
+ veor @XMM[$i-11], @XMM[$i-11], @XMM[$i-3]
+___
+}
+$code.=<<___;
+ sub $len, #0x10
+ vst1.64 {@XMM[15]}, [r0,:128] @ next round tweak
+
+ vld1.8 {@XMM[6]}, [$inp]!
+ veor @XMM[5], @XMM[5], @XMM[13]
+#ifndef BSAES_ASM_EXTENDED_KEY
+ add r4, sp, #0x90 @ pass key schedule
+#else
+ add r4, $key, #248 @ pass key schedule
+#endif
+ veor @XMM[6], @XMM[6], @XMM[14]
+ mov r5, $rounds @ pass rounds
+ mov r0, sp
+
+ bl _bsaes_encrypt8
+
+ vld1.64 {@XMM[ 8]-@XMM[ 9]}, [r0,:128]!
+ vld1.64 {@XMM[10]-@XMM[11]}, [r0,:128]!
+ veor @XMM[0], @XMM[0], @XMM[ 8]
+ vld1.64 {@XMM[12]-@XMM[13]}, [r0,:128]!
+ veor @XMM[1], @XMM[1], @XMM[ 9]
+ veor @XMM[8], @XMM[4], @XMM[10]
+ vst1.8 {@XMM[0]-@XMM[1]}, [$out]!
+ veor @XMM[9], @XMM[6], @XMM[11]
+ vld1.64 {@XMM[14]}, [r0,:128]!
+ veor @XMM[10], @XMM[3], @XMM[12]
+ vst1.8 {@XMM[8]-@XMM[9]}, [$out]!
+ veor @XMM[11], @XMM[7], @XMM[13]
+ veor @XMM[12], @XMM[2], @XMM[14]
+ vst1.8 {@XMM[10]-@XMM[11]}, [$out]!
+ vst1.8 {@XMM[12]}, [$out]!
+
+ vld1.64 {@XMM[8]}, [r0,:128] @ next round tweak
+ b .Lxts_enc_done
+.align 4
+.Lxts_enc_6:
+ vst1.64 {@XMM[14]}, [r0,:128] @ next round tweak
+
+ veor @XMM[4], @XMM[4], @XMM[12]
+#ifndef BSAES_ASM_EXTENDED_KEY
+ add r4, sp, #0x90 @ pass key schedule
+#else
+ add r4, $key, #248 @ pass key schedule
+#endif
+ veor @XMM[5], @XMM[5], @XMM[13]
+ mov r5, $rounds @ pass rounds
+ mov r0, sp
+
+ bl _bsaes_encrypt8
+
+ vld1.64 {@XMM[ 8]-@XMM[ 9]}, [r0,:128]!
+ vld1.64 {@XMM[10]-@XMM[11]}, [r0,:128]!
+ veor @XMM[0], @XMM[0], @XMM[ 8]
+ vld1.64 {@XMM[12]-@XMM[13]}, [r0,:128]!
+ veor @XMM[1], @XMM[1], @XMM[ 9]
+ veor @XMM[8], @XMM[4], @XMM[10]
+ vst1.8 {@XMM[0]-@XMM[1]}, [$out]!
+ veor @XMM[9], @XMM[6], @XMM[11]
+ veor @XMM[10], @XMM[3], @XMM[12]
+ vst1.8 {@XMM[8]-@XMM[9]}, [$out]!
+ veor @XMM[11], @XMM[7], @XMM[13]
+ vst1.8 {@XMM[10]-@XMM[11]}, [$out]!
+
+ vld1.64 {@XMM[8]}, [r0,:128] @ next round tweak
+ b .Lxts_enc_done
+
+@ put this in range for both ARM and Thumb mode adr instructions
+.align 5
+.Lxts_magic:
+ .quad 1, 0x87
+
+.align 5
+.Lxts_enc_5:
+ vst1.64 {@XMM[13]}, [r0,:128] @ next round tweak
+
+ veor @XMM[3], @XMM[3], @XMM[11]
+#ifndef BSAES_ASM_EXTENDED_KEY
+ add r4, sp, #0x90 @ pass key schedule
+#else
+ add r4, $key, #248 @ pass key schedule
+#endif
+ veor @XMM[4], @XMM[4], @XMM[12]
+ mov r5, $rounds @ pass rounds
+ mov r0, sp
+
+ bl _bsaes_encrypt8
+
+ vld1.64 {@XMM[ 8]-@XMM[ 9]}, [r0,:128]!
+ vld1.64 {@XMM[10]-@XMM[11]}, [r0,:128]!
+ veor @XMM[0], @XMM[0], @XMM[ 8]
+ vld1.64 {@XMM[12]}, [r0,:128]!
+ veor @XMM[1], @XMM[1], @XMM[ 9]
+ veor @XMM[8], @XMM[4], @XMM[10]
+ vst1.8 {@XMM[0]-@XMM[1]}, [$out]!
+ veor @XMM[9], @XMM[6], @XMM[11]
+ veor @XMM[10], @XMM[3], @XMM[12]
+ vst1.8 {@XMM[8]-@XMM[9]}, [$out]!
+ vst1.8 {@XMM[10]}, [$out]!
+
+ vld1.64 {@XMM[8]}, [r0,:128] @ next round tweak
+ b .Lxts_enc_done
+.align 4
+.Lxts_enc_4:
+ vst1.64 {@XMM[12]}, [r0,:128] @ next round tweak
+
+ veor @XMM[2], @XMM[2], @XMM[10]
+#ifndef BSAES_ASM_EXTENDED_KEY
+ add r4, sp, #0x90 @ pass key schedule
+#else
+ add r4, $key, #248 @ pass key schedule
+#endif
+ veor @XMM[3], @XMM[3], @XMM[11]
+ mov r5, $rounds @ pass rounds
+ mov r0, sp
+
+ bl _bsaes_encrypt8
+
+ vld1.64 {@XMM[ 8]-@XMM[ 9]}, [r0,:128]!
+ vld1.64 {@XMM[10]-@XMM[11]}, [r0,:128]!
+ veor @XMM[0], @XMM[0], @XMM[ 8]
+ veor @XMM[1], @XMM[1], @XMM[ 9]
+ veor @XMM[8], @XMM[4], @XMM[10]
+ vst1.8 {@XMM[0]-@XMM[1]}, [$out]!
+ veor @XMM[9], @XMM[6], @XMM[11]
+ vst1.8 {@XMM[8]-@XMM[9]}, [$out]!
+
+ vld1.64 {@XMM[8]}, [r0,:128] @ next round tweak
+ b .Lxts_enc_done
+.align 4
+.Lxts_enc_3:
+ vst1.64 {@XMM[11]}, [r0,:128] @ next round tweak
+
+ veor @XMM[1], @XMM[1], @XMM[9]
+#ifndef BSAES_ASM_EXTENDED_KEY
+ add r4, sp, #0x90 @ pass key schedule
+#else
+ add r4, $key, #248 @ pass key schedule
+#endif
+ veor @XMM[2], @XMM[2], @XMM[10]
+ mov r5, $rounds @ pass rounds
+ mov r0, sp
+
+ bl _bsaes_encrypt8
+
+ vld1.64 {@XMM[8]-@XMM[9]}, [r0,:128]!
+ vld1.64 {@XMM[10]}, [r0,:128]!
+ veor @XMM[0], @XMM[0], @XMM[ 8]
+ veor @XMM[1], @XMM[1], @XMM[ 9]
+ veor @XMM[8], @XMM[4], @XMM[10]
+ vst1.8 {@XMM[0]-@XMM[1]}, [$out]!
+ vst1.8 {@XMM[8]}, [$out]!
+
+ vld1.64 {@XMM[8]}, [r0,:128] @ next round tweak
+ b .Lxts_enc_done
+.align 4
+.Lxts_enc_2:
+ vst1.64 {@XMM[10]}, [r0,:128] @ next round tweak
+
+ veor @XMM[0], @XMM[0], @XMM[8]
+#ifndef BSAES_ASM_EXTENDED_KEY
+ add r4, sp, #0x90 @ pass key schedule
+#else
+ add r4, $key, #248 @ pass key schedule
+#endif
+ veor @XMM[1], @XMM[1], @XMM[9]
+ mov r5, $rounds @ pass rounds
+ mov r0, sp
+
+ bl _bsaes_encrypt8
+
+ vld1.64 {@XMM[8]-@XMM[9]}, [r0,:128]!
+ veor @XMM[0], @XMM[0], @XMM[ 8]
+ veor @XMM[1], @XMM[1], @XMM[ 9]
+ vst1.8 {@XMM[0]-@XMM[1]}, [$out]!
+
+ vld1.64 {@XMM[8]}, [r0,:128] @ next round tweak
+ b .Lxts_enc_done
+.align 4
+.Lxts_enc_1:
+ mov r0, sp
+ veor @XMM[0], @XMM[8]
+ mov r1, sp
+ vst1.8 {@XMM[0]}, [sp,:128]
+ mov r2, $key
+ mov r4, $fp @ preserve fp
+
+ bl AES_encrypt
+
+ vld1.8 {@XMM[0]}, [sp,:128]
+ veor @XMM[0], @XMM[0], @XMM[8]
+ vst1.8 {@XMM[0]}, [$out]!
+ mov $fp, r4
+
+ vmov @XMM[8], @XMM[9] @ next round tweak
+
+.Lxts_enc_done:
+#ifndef XTS_CHAIN_TWEAK
+ adds $len, #0x10
+ beq .Lxts_enc_ret
+ sub r6, $out, #0x10
+
+.Lxts_enc_steal:
+ ldrb r0, [$inp], #1
+ ldrb r1, [$out, #-0x10]
+ strb r0, [$out, #-0x10]
+ strb r1, [$out], #1
+
+ subs $len, #1
+ bhi .Lxts_enc_steal
+
+ vld1.8 {@XMM[0]}, [r6]
+ mov r0, sp
+ veor @XMM[0], @XMM[0], @XMM[8]
+ mov r1, sp
+ vst1.8 {@XMM[0]}, [sp,:128]
+ mov r2, $key
+ mov r4, $fp @ preserve fp
+
+ bl AES_encrypt
+
+ vld1.8 {@XMM[0]}, [sp,:128]
+ veor @XMM[0], @XMM[0], @XMM[8]
+ vst1.8 {@XMM[0]}, [r6]
+ mov $fp, r4
+#endif
+
+.Lxts_enc_ret:
+ bic r0, $fp, #0xf
+ vmov.i32 q0, #0
+ vmov.i32 q1, #0
+#ifdef XTS_CHAIN_TWEAK
+ ldr r1, [$fp, #0x20+VFP_ABI_FRAME] @ chain tweak
+#endif
+.Lxts_enc_bzero: @ wipe key schedule [if any]
+ vstmia sp!, {q0-q1}
+ cmp sp, r0
+ bne .Lxts_enc_bzero
+
+ mov sp, $fp
+#ifdef XTS_CHAIN_TWEAK
+ vst1.8 {@XMM[8]}, [r1]
+#endif
+ VFP_ABI_POP
+ ldmia sp!, {r4-r10, pc} @ return
+
+.size bsaes_xts_encrypt,.-bsaes_xts_encrypt
+
+.globl bsaes_xts_decrypt
+.type bsaes_xts_decrypt,%function
+.align 4
+bsaes_xts_decrypt:
+ mov ip, sp
+ stmdb sp!, {r4-r10, lr} @ 0x20
+ VFP_ABI_PUSH
+ mov r6, sp @ future $fp
+
+ mov $inp, r0
+ mov $out, r1
+ mov $len, r2
+ mov $key, r3
+
+ sub r0, sp, #0x10 @ 0x10
+ bic r0, #0xf @ align at 16 bytes
+ mov sp, r0
+
+#ifdef XTS_CHAIN_TWEAK
+ ldr r0, [ip] @ pointer to input tweak
+#else
+ @ generate initial tweak
+ ldr r0, [ip, #4] @ iv[]
+ mov r1, sp
+ ldr r2, [ip, #0] @ key2
+ bl AES_encrypt
+ mov r0, sp @ pointer to initial tweak
+#endif
+
+ ldr $rounds, [$key, #240] @ get # of rounds
+ mov $fp, r6
+#ifndef BSAES_ASM_EXTENDED_KEY
+ @ allocate the key schedule on the stack
+ sub r12, sp, $rounds, lsl#7 @ 128 bytes per inner round key
+ @ add r12, #`128-32` @ size of bit-sliced key schedule
+ sub r12, #`32+16` @ place for tweak[9]
+
+ @ populate the key schedule
+ mov r4, $key @ pass key
+ mov r5, $rounds @ pass # of rounds
+ mov sp, r12
+ add r12, #0x90 @ pass key schedule
+ bl _bsaes_key_convert
+ add r4, sp, #0x90
+ vldmia r4, {@XMM[6]}
+ vstmia r12, {@XMM[15]} @ save last round key
+ veor @XMM[7], @XMM[7], @XMM[6] @ fix up round 0 key
+ vstmia r4, {@XMM[7]}
+#else
+ ldr r12, [$key, #244]
+ eors r12, #1
+ beq 0f
+
+ str r12, [$key, #244]
+ mov r4, $key @ pass key
+ mov r5, $rounds @ pass # of rounds
+ add r12, $key, #248 @ pass key schedule
+ bl _bsaes_key_convert
+ add r4, $key, #248
+ vldmia r4, {@XMM[6]}
+ vstmia r12, {@XMM[15]} @ save last round key
+ veor @XMM[7], @XMM[7], @XMM[6] @ fix up round 0 key
+ vstmia r4, {@XMM[7]}
+
+.align 2
+0: sub sp, #0x90 @ place for tweak[9]
+#endif
+ vld1.8 {@XMM[8]}, [r0] @ initial tweak
+ adr $magic, .Lxts_magic
+
+ tst $len, #0xf @ if not multiple of 16
+ it ne @ Thumb2 thing, sanity check in ARM
+ subne $len, #0x10 @ subtract another 16 bytes
+ subs $len, #0x80
+
+ blo .Lxts_dec_short
+ b .Lxts_dec_loop
+
+.align 4
+.Lxts_dec_loop:
+ vldmia $magic, {$twmask} @ load XTS magic
+ vshr.s64 @T[0], @XMM[8], #63
+ mov r0, sp
+ vand @T[0], @T[0], $twmask
+___
+for($i=9;$i<16;$i++) {
+$code.=<<___;
+ vadd.u64 @XMM[$i], @XMM[$i-1], @XMM[$i-1]
+ vst1.64 {@XMM[$i-1]}, [r0,:128]!
+ vswp `&Dhi("@T[0]")`,`&Dlo("@T[0]")`
+ vshr.s64 @T[1], @XMM[$i], #63
+ veor @XMM[$i], @XMM[$i], @T[0]
+ vand @T[1], @T[1], $twmask
+___
+ @T=reverse(@T);
+
+$code.=<<___ if ($i>=10);
+ vld1.8 {@XMM[$i-10]}, [$inp]!
+___
+$code.=<<___ if ($i>=11);
+ veor @XMM[$i-11], @XMM[$i-11], @XMM[$i-3]
+___
+}
+$code.=<<___;
+ vadd.u64 @XMM[8], @XMM[15], @XMM[15]
+ vst1.64 {@XMM[15]}, [r0,:128]!
+ vswp `&Dhi("@T[0]")`,`&Dlo("@T[0]")`
+ veor @XMM[8], @XMM[8], @T[0]
+ vst1.64 {@XMM[8]}, [r0,:128] @ next round tweak
+
+ vld1.8 {@XMM[6]-@XMM[7]}, [$inp]!
+ veor @XMM[5], @XMM[5], @XMM[13]
+#ifndef BSAES_ASM_EXTENDED_KEY
+ add r4, sp, #0x90 @ pass key schedule
+#else
+ add r4, $key, #248 @ pass key schedule
+#endif
+ veor @XMM[6], @XMM[6], @XMM[14]
+ mov r5, $rounds @ pass rounds
+ veor @XMM[7], @XMM[7], @XMM[15]
+ mov r0, sp
+
+ bl _bsaes_decrypt8
+
+ vld1.64 {@XMM[ 8]-@XMM[ 9]}, [r0,:128]!
+ vld1.64 {@XMM[10]-@XMM[11]}, [r0,:128]!
+ veor @XMM[0], @XMM[0], @XMM[ 8]
+ vld1.64 {@XMM[12]-@XMM[13]}, [r0,:128]!
+ veor @XMM[1], @XMM[1], @XMM[ 9]
+ veor @XMM[8], @XMM[6], @XMM[10]
+ vst1.8 {@XMM[0]-@XMM[1]}, [$out]!
+ veor @XMM[9], @XMM[4], @XMM[11]
+ vld1.64 {@XMM[14]-@XMM[15]}, [r0,:128]!
+ veor @XMM[10], @XMM[2], @XMM[12]
+ vst1.8 {@XMM[8]-@XMM[9]}, [$out]!
+ veor @XMM[11], @XMM[7], @XMM[13]
+ veor @XMM[12], @XMM[3], @XMM[14]
+ vst1.8 {@XMM[10]-@XMM[11]}, [$out]!
+ veor @XMM[13], @XMM[5], @XMM[15]
+ vst1.8 {@XMM[12]-@XMM[13]}, [$out]!
+
+ vld1.64 {@XMM[8]}, [r0,:128] @ next round tweak
+
+ subs $len, #0x80
+ bpl .Lxts_dec_loop
+
+.Lxts_dec_short:
+ adds $len, #0x70
+ bmi .Lxts_dec_done
+
+ vldmia $magic, {$twmask} @ load XTS magic
+ vshr.s64 @T[0], @XMM[8], #63
+ mov r0, sp
+ vand @T[0], @T[0], $twmask
+___
+for($i=9;$i<16;$i++) {
+$code.=<<___;
+ vadd.u64 @XMM[$i], @XMM[$i-1], @XMM[$i-1]
+ vst1.64 {@XMM[$i-1]}, [r0,:128]!
+ vswp `&Dhi("@T[0]")`,`&Dlo("@T[0]")`
+ vshr.s64 @T[1], @XMM[$i], #63
+ veor @XMM[$i], @XMM[$i], @T[0]
+ vand @T[1], @T[1], $twmask
+___
+ @T=reverse(@T);
+
+$code.=<<___ if ($i>=10);
+ vld1.8 {@XMM[$i-10]}, [$inp]!
+ subs $len, #0x10
+ bmi .Lxts_dec_`$i-9`
+___
+$code.=<<___ if ($i>=11);
+ veor @XMM[$i-11], @XMM[$i-11], @XMM[$i-3]
+___
+}
+$code.=<<___;
+ sub $len, #0x10
+ vst1.64 {@XMM[15]}, [r0,:128] @ next round tweak
+
+ vld1.8 {@XMM[6]}, [$inp]!
+ veor @XMM[5], @XMM[5], @XMM[13]
+#ifndef BSAES_ASM_EXTENDED_KEY
+ add r4, sp, #0x90 @ pass key schedule
+#else
+ add r4, $key, #248 @ pass key schedule
+#endif
+ veor @XMM[6], @XMM[6], @XMM[14]
+ mov r5, $rounds @ pass rounds
+ mov r0, sp
+
+ bl _bsaes_decrypt8
+
+ vld1.64 {@XMM[ 8]-@XMM[ 9]}, [r0,:128]!
+ vld1.64 {@XMM[10]-@XMM[11]}, [r0,:128]!
+ veor @XMM[0], @XMM[0], @XMM[ 8]
+ vld1.64 {@XMM[12]-@XMM[13]}, [r0,:128]!
+ veor @XMM[1], @XMM[1], @XMM[ 9]
+ veor @XMM[8], @XMM[6], @XMM[10]
+ vst1.8 {@XMM[0]-@XMM[1]}, [$out]!
+ veor @XMM[9], @XMM[4], @XMM[11]
+ vld1.64 {@XMM[14]}, [r0,:128]!
+ veor @XMM[10], @XMM[2], @XMM[12]
+ vst1.8 {@XMM[8]-@XMM[9]}, [$out]!
+ veor @XMM[11], @XMM[7], @XMM[13]
+ veor @XMM[12], @XMM[3], @XMM[14]
+ vst1.8 {@XMM[10]-@XMM[11]}, [$out]!
+ vst1.8 {@XMM[12]}, [$out]!
+
+ vld1.64 {@XMM[8]}, [r0,:128] @ next round tweak
+ b .Lxts_dec_done
+.align 4
+.Lxts_dec_6:
+ vst1.64 {@XMM[14]}, [r0,:128] @ next round tweak
+
+ veor @XMM[4], @XMM[4], @XMM[12]
+#ifndef BSAES_ASM_EXTENDED_KEY
+ add r4, sp, #0x90 @ pass key schedule
+#else
+ add r4, $key, #248 @ pass key schedule
+#endif
+ veor @XMM[5], @XMM[5], @XMM[13]
+ mov r5, $rounds @ pass rounds
+ mov r0, sp
+
+ bl _bsaes_decrypt8
+
+ vld1.64 {@XMM[ 8]-@XMM[ 9]}, [r0,:128]!
+ vld1.64 {@XMM[10]-@XMM[11]}, [r0,:128]!
+ veor @XMM[0], @XMM[0], @XMM[ 8]
+ vld1.64 {@XMM[12]-@XMM[13]}, [r0,:128]!
+ veor @XMM[1], @XMM[1], @XMM[ 9]
+ veor @XMM[8], @XMM[6], @XMM[10]
+ vst1.8 {@XMM[0]-@XMM[1]}, [$out]!
+ veor @XMM[9], @XMM[4], @XMM[11]
+ veor @XMM[10], @XMM[2], @XMM[12]
+ vst1.8 {@XMM[8]-@XMM[9]}, [$out]!
+ veor @XMM[11], @XMM[7], @XMM[13]
+ vst1.8 {@XMM[10]-@XMM[11]}, [$out]!
+
+ vld1.64 {@XMM[8]}, [r0,:128] @ next round tweak
+ b .Lxts_dec_done
+.align 4
+.Lxts_dec_5:
+ vst1.64 {@XMM[13]}, [r0,:128] @ next round tweak
+
+ veor @XMM[3], @XMM[3], @XMM[11]
+#ifndef BSAES_ASM_EXTENDED_KEY
+ add r4, sp, #0x90 @ pass key schedule
+#else
+ add r4, $key, #248 @ pass key schedule
+#endif
+ veor @XMM[4], @XMM[4], @XMM[12]
+ mov r5, $rounds @ pass rounds
+ mov r0, sp
+
+ bl _bsaes_decrypt8
+
+ vld1.64 {@XMM[ 8]-@XMM[ 9]}, [r0,:128]!
+ vld1.64 {@XMM[10]-@XMM[11]}, [r0,:128]!
+ veor @XMM[0], @XMM[0], @XMM[ 8]
+ vld1.64 {@XMM[12]}, [r0,:128]!
+ veor @XMM[1], @XMM[1], @XMM[ 9]
+ veor @XMM[8], @XMM[6], @XMM[10]
+ vst1.8 {@XMM[0]-@XMM[1]}, [$out]!
+ veor @XMM[9], @XMM[4], @XMM[11]
+ veor @XMM[10], @XMM[2], @XMM[12]
+ vst1.8 {@XMM[8]-@XMM[9]}, [$out]!
+ vst1.8 {@XMM[10]}, [$out]!
+
+ vld1.64 {@XMM[8]}, [r0,:128] @ next round tweak
+ b .Lxts_dec_done
+.align 4
+.Lxts_dec_4:
+ vst1.64 {@XMM[12]}, [r0,:128] @ next round tweak
+
+ veor @XMM[2], @XMM[2], @XMM[10]
+#ifndef BSAES_ASM_EXTENDED_KEY
+ add r4, sp, #0x90 @ pass key schedule
+#else
+ add r4, $key, #248 @ pass key schedule
+#endif
+ veor @XMM[3], @XMM[3], @XMM[11]
+ mov r5, $rounds @ pass rounds
+ mov r0, sp
+
+ bl _bsaes_decrypt8
+
+ vld1.64 {@XMM[ 8]-@XMM[ 9]}, [r0,:128]!
+ vld1.64 {@XMM[10]-@XMM[11]}, [r0,:128]!
+ veor @XMM[0], @XMM[0], @XMM[ 8]
+ veor @XMM[1], @XMM[1], @XMM[ 9]
+ veor @XMM[8], @XMM[6], @XMM[10]
+ vst1.8 {@XMM[0]-@XMM[1]}, [$out]!
+ veor @XMM[9], @XMM[4], @XMM[11]
+ vst1.8 {@XMM[8]-@XMM[9]}, [$out]!
+
+ vld1.64 {@XMM[8]}, [r0,:128] @ next round tweak
+ b .Lxts_dec_done
+.align 4
+.Lxts_dec_3:
+ vst1.64 {@XMM[11]}, [r0,:128] @ next round tweak
+
+ veor @XMM[1], @XMM[1], @XMM[9]
+#ifndef BSAES_ASM_EXTENDED_KEY
+ add r4, sp, #0x90 @ pass key schedule
+#else
+ add r4, $key, #248 @ pass key schedule
+#endif
+ veor @XMM[2], @XMM[2], @XMM[10]
+ mov r5, $rounds @ pass rounds
+ mov r0, sp
+
+ bl _bsaes_decrypt8
+
+ vld1.64 {@XMM[8]-@XMM[9]}, [r0,:128]!
+ vld1.64 {@XMM[10]}, [r0,:128]!
+ veor @XMM[0], @XMM[0], @XMM[ 8]
+ veor @XMM[1], @XMM[1], @XMM[ 9]
+ veor @XMM[8], @XMM[6], @XMM[10]
+ vst1.8 {@XMM[0]-@XMM[1]}, [$out]!
+ vst1.8 {@XMM[8]}, [$out]!
+
+ vld1.64 {@XMM[8]}, [r0,:128] @ next round tweak
+ b .Lxts_dec_done
+.align 4
+.Lxts_dec_2:
+ vst1.64 {@XMM[10]}, [r0,:128] @ next round tweak
+
+ veor @XMM[0], @XMM[0], @XMM[8]
+#ifndef BSAES_ASM_EXTENDED_KEY
+ add r4, sp, #0x90 @ pass key schedule
+#else
+ add r4, $key, #248 @ pass key schedule
+#endif
+ veor @XMM[1], @XMM[1], @XMM[9]
+ mov r5, $rounds @ pass rounds
+ mov r0, sp
+
+ bl _bsaes_decrypt8
+
+ vld1.64 {@XMM[8]-@XMM[9]}, [r0,:128]!
+ veor @XMM[0], @XMM[0], @XMM[ 8]
+ veor @XMM[1], @XMM[1], @XMM[ 9]
+ vst1.8 {@XMM[0]-@XMM[1]}, [$out]!
+
+ vld1.64 {@XMM[8]}, [r0,:128] @ next round tweak
+ b .Lxts_dec_done
+.align 4
+.Lxts_dec_1:
+ mov r0, sp
+ veor @XMM[0], @XMM[8]
+ mov r1, sp
+ vst1.8 {@XMM[0]}, [sp,:128]
+ mov r2, $key
+ mov r4, $fp @ preserve fp
+ mov r5, $magic @ preserve magic
+
+ bl AES_decrypt
+
+ vld1.8 {@XMM[0]}, [sp,:128]
+ veor @XMM[0], @XMM[0], @XMM[8]
+ vst1.8 {@XMM[0]}, [$out]!
+ mov $fp, r4
+ mov $magic, r5
+
+ vmov @XMM[8], @XMM[9] @ next round tweak
+
+.Lxts_dec_done:
+#ifndef XTS_CHAIN_TWEAK
+ adds $len, #0x10
+ beq .Lxts_dec_ret
+
+ @ calculate one round of extra tweak for the stolen ciphertext
+ vldmia $magic, {$twmask}
+ vshr.s64 @XMM[6], @XMM[8], #63
+ vand @XMM[6], @XMM[6], $twmask
+ vadd.u64 @XMM[9], @XMM[8], @XMM[8]
+ vswp `&Dhi("@XMM[6]")`,`&Dlo("@XMM[6]")`
+ veor @XMM[9], @XMM[9], @XMM[6]
+
+ @ perform the final decryption with the last tweak value
+ vld1.8 {@XMM[0]}, [$inp]!
+ mov r0, sp
+ veor @XMM[0], @XMM[0], @XMM[9]
+ mov r1, sp
+ vst1.8 {@XMM[0]}, [sp,:128]
+ mov r2, $key
+ mov r4, $fp @ preserve fp
+
+ bl AES_decrypt
+
+ vld1.8 {@XMM[0]}, [sp,:128]
+ veor @XMM[0], @XMM[0], @XMM[9]
+ vst1.8 {@XMM[0]}, [$out]
+
+ mov r6, $out
+.Lxts_dec_steal:
+ ldrb r1, [$out]
+ ldrb r0, [$inp], #1
+ strb r1, [$out, #0x10]
+ strb r0, [$out], #1
+
+ subs $len, #1
+ bhi .Lxts_dec_steal
+
+ vld1.8 {@XMM[0]}, [r6]
+ mov r0, sp
+ veor @XMM[0], @XMM[8]
+ mov r1, sp
+ vst1.8 {@XMM[0]}, [sp,:128]
+ mov r2, $key
+
+ bl AES_decrypt
+
+ vld1.8 {@XMM[0]}, [sp,:128]
+ veor @XMM[0], @XMM[0], @XMM[8]
+ vst1.8 {@XMM[0]}, [r6]
+ mov $fp, r4
+#endif
+
+.Lxts_dec_ret:
+ bic r0, $fp, #0xf
+ vmov.i32 q0, #0
+ vmov.i32 q1, #0