Initial import

29 files changed, 14638 insertions, 0 deletions

diff --git a/arch/arm/crypto/.gitignore b/arch/arm/crypto/.gitignore
new file mode 100644
index 000000000..6231d36b3
--- /dev/null
+++ b/arch/arm/crypto/.gitignore
@@ -0,0 +1 @@
+aesbs-core.S
diff --git a/arch/arm/crypto/Kconfig b/arch/arm/crypto/Kconfig
new file mode 100644
index 000000000..8da2207b0
--- /dev/null
+++ b/arch/arm/crypto/Kconfig
@@ -0,0 +1,130 @@
+
+menuconfig ARM_CRYPTO
+	bool "ARM Accelerated Cryptographic Algorithms"
+	depends on ARM
+	help
+	  Say Y here to choose from a selection of cryptographic algorithms
+	  implemented using ARM specific CPU features or instructions.
+
+if ARM_CRYPTO
+
+config CRYPTO_SHA1_ARM
+	tristate "SHA1 digest algorithm (ARM-asm)"
+	select CRYPTO_SHA1
+	select CRYPTO_HASH
+	help
+	  SHA-1 secure hash standard (FIPS 180-1/DFIPS 180-2) implemented
+	  using optimized ARM assembler.
+
+config CRYPTO_SHA1_ARM_NEON
+	tristate "SHA1 digest algorithm (ARM NEON)"
+	depends on KERNEL_MODE_NEON
+	select CRYPTO_SHA1_ARM
+	select CRYPTO_SHA1
+	select CRYPTO_HASH
+	help
+	  SHA-1 secure hash standard (FIPS 180-1/DFIPS 180-2) implemented
+	  using optimized ARM NEON assembly, when NEON instructions are
+	  available.
+
+config CRYPTO_SHA1_ARM_CE
+	tristate "SHA1 digest algorithm (ARM v8 Crypto Extensions)"
+	depends on KERNEL_MODE_NEON
+	select CRYPTO_SHA1_ARM
+	select CRYPTO_HASH
+	help
+	  SHA-1 secure hash standard (FIPS 180-1/DFIPS 180-2) implemented
+	  using special ARMv8 Crypto Extensions.
+
+config CRYPTO_SHA2_ARM_CE
+	tristate "SHA-224/256 digest algorithm (ARM v8 Crypto Extensions)"
+	depends on KERNEL_MODE_NEON
+	select CRYPTO_SHA256_ARM
+	select CRYPTO_HASH
+	help
+	  SHA-256 secure hash standard (DFIPS 180-2) implemented
+	  using special ARMv8 Crypto Extensions.
+
+config CRYPTO_SHA256_ARM
+	tristate "SHA-224/256 digest algorithm (ARM-asm and NEON)"
+	select CRYPTO_HASH
+	depends on !CPU_V7M
+	help
+	  SHA-256 secure hash standard (DFIPS 180-2) implemented
+	  using optimized ARM assembler and NEON, when available.
+
+config CRYPTO_SHA512_ARM_NEON
+	tristate "SHA384 and SHA512 digest algorithm (ARM NEON)"
+	depends on KERNEL_MODE_NEON
+	select CRYPTO_SHA512
+	select CRYPTO_HASH
+	help
+	  SHA-512 secure hash standard (DFIPS 180-2) implemented
+	  using ARM NEON instructions, when available.
+
+	  This version of SHA implements a 512 bit hash with 256 bits of
+	  security against collision attacks.
+
+	  This code also includes SHA-384, a 384 bit hash with 192 bits
+	  of security against collision attacks.
+
+config CRYPTO_AES_ARM
+	tristate "AES cipher algorithms (ARM-asm)"
+	depends on ARM
+	select CRYPTO_ALGAPI
+	select CRYPTO_AES
+	help
+	  Use optimized AES assembler routines for ARM platforms.
+
+	  AES cipher algorithms (FIPS-197). AES uses the Rijndael
+	  algorithm.
+
+	  Rijndael appears to be consistently a very good performer in
+	  both hardware and software across a wide range of computing
+	  environments regardless of its use in feedback or non-feedback
+	  modes. Its key setup time is excellent, and its key agility is
+	  good. Rijndael's very low memory requirements make it very well
+	  suited for restricted-space environments, in which it also
+	  demonstrates excellent performance. Rijndael's operations are
+	  among the easiest to defend against power and timing attacks.
+
+	  The AES specifies three key sizes: 128, 192 and 256 bits
+
+	  See <http://csrc.nist.gov/encryption/aes/> for more information.
+
+config CRYPTO_AES_ARM_BS
+	tristate "Bit sliced AES using NEON instructions"
+	depends on KERNEL_MODE_NEON
+	select CRYPTO_ALGAPI
+	select CRYPTO_AES_ARM
+	select CRYPTO_ABLK_HELPER
+	help
+	  Use a faster and more secure NEON based implementation of AES in CBC,
+	  CTR and XTS modes
+
+	  Bit sliced AES gives around 45% speedup on Cortex-A15 for CTR mode
+	  and for XTS mode encryption, CBC and XTS mode decryption speedup is
+	  around 25%. (CBC encryption speed is not affected by this driver.)
+	  This implementation does not rely on any lookup tables so it is
+	  believed to be invulnerable to cache timing attacks.
+
+config CRYPTO_AES_ARM_CE
+	tristate "Accelerated AES using ARMv8 Crypto Extensions"
+	depends on KERNEL_MODE_NEON
+	select CRYPTO_ALGAPI
+	select CRYPTO_ABLK_HELPER
+	help
+	  Use an implementation of AES in CBC, CTR and XTS modes that uses
+	  ARMv8 Crypto Extensions
+
+config CRYPTO_GHASH_ARM_CE
+	tristate "PMULL-accelerated GHASH using ARMv8 Crypto Extensions"
+	depends on KERNEL_MODE_NEON
+	select CRYPTO_HASH
+	select CRYPTO_CRYPTD
+	help
+	  Use an implementation of GHASH (used by the GCM AEAD chaining mode)
+	  that uses the 64x64 to 128 bit polynomial multiplication (vmull.p64)
+	  that is part of the ARMv8 Crypto Extensions
+
+endif
diff --git a/arch/arm/crypto/Makefile b/arch/arm/crypto/Makefile
new file mode 100644
index 000000000..6ea828241
--- /dev/null
+++ b/arch/arm/crypto/Makefile
@@ -0,0 +1,48 @@
+#
+# Arch-specific CryptoAPI modules.
+#
+
+obj-$(CONFIG_CRYPTO_AES_ARM) += aes-arm.o
+obj-$(CONFIG_CRYPTO_AES_ARM_BS) += aes-arm-bs.o
+obj-$(CONFIG_CRYPTO_SHA1_ARM) += sha1-arm.o
+obj-$(CONFIG_CRYPTO_SHA1_ARM_NEON) += sha1-arm-neon.o
+obj-$(CONFIG_CRYPTO_SHA256_ARM) += sha256-arm.o
+obj-$(CONFIG_CRYPTO_SHA512_ARM_NEON) += sha512-arm-neon.o
+
+ce-obj-$(CONFIG_CRYPTO_AES_ARM_CE) += aes-arm-ce.o
+ce-obj-$(CONFIG_CRYPTO_SHA1_ARM_CE) += sha1-arm-ce.o
+ce-obj-$(CONFIG_CRYPTO_SHA2_ARM_CE) += sha2-arm-ce.o
+ce-obj-$(CONFIG_CRYPTO_GHASH_ARM_CE) += ghash-arm-ce.o
+
+ifneq ($(ce-obj-y)$(ce-obj-m),)
+ifeq ($(call as-instr,.fpu crypto-neon-fp-armv8,y,n),y)
+obj-y += $(ce-obj-y)
+obj-m += $(ce-obj-m)
+else
+$(warning These ARMv8 Crypto Extensions modules need binutils 2.23 or higher)
+$(warning $(ce-obj-y) $(ce-obj-m))
+endif
+endif
+
+aes-arm-y	:= aes-armv4.o aes_glue.o
+aes-arm-bs-y	:= aesbs-core.o aesbs-glue.o
+sha1-arm-y	:= sha1-armv4-large.o sha1_glue.o
+sha1-arm-neon-y	:= sha1-armv7-neon.o sha1_neon_glue.o
+sha256-arm-neon-$(CONFIG_KERNEL_MODE_NEON) := sha256_neon_glue.o
+sha256-arm-y	:= sha256-core.o sha256_glue.o $(sha256-arm-neon-y)
+sha512-arm-neon-y := sha512-armv7-neon.o sha512_neon_glue.o
+sha1-arm-ce-y	:= sha1-ce-core.o sha1-ce-glue.o
+sha2-arm-ce-y	:= sha2-ce-core.o sha2-ce-glue.o
+aes-arm-ce-y	:= aes-ce-core.o aes-ce-glue.o
+ghash-arm-ce-y	:= ghash-ce-core.o ghash-ce-glue.o
+
+quiet_cmd_perl = PERL    $@
+      cmd_perl = $(PERL) $(<) > $(@)
+
+$(src)/aesbs-core.S_shipped: $(src)/bsaes-armv7.pl
+	$(call cmd,perl)
+
+$(src)/sha256-core.S_shipped: $(src)/sha256-armv4.pl
+	$(call cmd,perl)
+
+.PRECIOUS: $(obj)/aesbs-core.S $(obj)/sha256-core.S
diff --git a/arch/arm/crypto/aes-armv4.S b/arch/arm/crypto/aes-armv4.S
new file mode 100644
index 000000000..ebb9761fb
--- /dev/null
+++ b/arch/arm/crypto/aes-armv4.S
@@ -0,0 +1,1089 @@
+#define __ARM_ARCH__ __LINUX_ARM_ARCH__
+@ ====================================================================
+@ Written by Andy Polyakov <appro@fy.chalmers.se> 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/.
+@ ====================================================================
+
+@ AES for ARMv4
+
+@ January 2007.
+@
+@ Code uses single 1K S-box and is >2 times faster than code generated
+@ by gcc-3.4.1. This is thanks to unique feature of ARMv4 ISA, which
+@ allows to merge logical or arithmetic operation with shift or rotate
+@ in one instruction and emit combined result every cycle. The module
+@ is endian-neutral. The performance is ~42 cycles/byte for 128-bit
+@ key [on single-issue Xscale PXA250 core].
+
+@ May 2007.
+@
+@ AES_set_[en|de]crypt_key is added.
+
+@ July 2010.
+@
+@ Rescheduling for dual-issue pipeline resulted in 12% improvement on
+@ Cortex A8 core and ~25 cycles per byte processed with 128-bit key.
+
+@ February 2011.
+@
+@ Profiler-assisted and platform-specific optimization resulted in 16%
+@ improvement on Cortex A8 core and ~21.5 cycles per byte.
+
+@ A little glue here to select the correct code below for the ARM CPU
+@ that is being targetted.
+
+#include <linux/linkage.h>
+#include <asm/assembler.h>
+
+.text
+
+.type	AES_Te,%object
+.align	5
+AES_Te:
+.word	0xc66363a5, 0xf87c7c84, 0xee777799, 0xf67b7b8d
+.word	0xfff2f20d, 0xd66b6bbd, 0xde6f6fb1, 0x91c5c554
+.word	0x60303050, 0x02010103, 0xce6767a9, 0x562b2b7d
+.word	0xe7fefe19, 0xb5d7d762, 0x4dababe6, 0xec76769a
+.word	0x8fcaca45, 0x1f82829d, 0x89c9c940, 0xfa7d7d87
+.word	0xeffafa15, 0xb25959eb, 0x8e4747c9, 0xfbf0f00b
+.word	0x41adadec, 0xb3d4d467, 0x5fa2a2fd, 0x45afafea
+.word	0x239c9cbf, 0x53a4a4f7, 0xe4727296, 0x9bc0c05b
+.word	0x75b7b7c2, 0xe1fdfd1c, 0x3d9393ae, 0x4c26266a
+.word	0x6c36365a, 0x7e3f3f41, 0xf5f7f702, 0x83cccc4f
+.word	0x6834345c, 0x51a5a5f4, 0xd1e5e534, 0xf9f1f108
+.word	0xe2717193, 0xabd8d873, 0x62313153, 0x2a15153f
+.word	0x0804040c, 0x95c7c752, 0x46232365, 0x9dc3c35e
+.word	0x30181828, 0x379696a1, 0x0a05050f, 0x2f9a9ab5
+.word	0x0e070709, 0x24121236, 0x1b80809b, 0xdfe2e23d
+.word	0xcdebeb26, 0x4e272769, 0x7fb2b2cd, 0xea75759f
+.word	0x1209091b, 0x1d83839e, 0x582c2c74, 0x341a1a2e
+.word	0x361b1b2d, 0xdc6e6eb2, 0xb45a5aee, 0x5ba0a0fb
+.word	0xa45252f6, 0x763b3b4d, 0xb7d6d661, 0x7db3b3ce
+.word	0x5229297b, 0xdde3e33e, 0x5e2f2f71, 0x13848497
+.word	0xa65353f5, 0xb9d1d168, 0x00000000, 0xc1eded2c
+.word	0x40202060, 0xe3fcfc1f, 0x79b1b1c8, 0xb65b5bed
+.word	0xd46a6abe, 0x8dcbcb46, 0x67bebed9, 0x7239394b
+.word	0x944a4ade, 0x984c4cd4, 0xb05858e8, 0x85cfcf4a
+.word	0xbbd0d06b, 0xc5efef2a, 0x4faaaae5, 0xedfbfb16
+.word	0x864343c5, 0x9a4d4dd7, 0x66333355, 0x11858594
+.word	0x8a4545cf, 0xe9f9f910, 0x04020206, 0xfe7f7f81
+.word	0xa05050f0, 0x783c3c44, 0x259f9fba, 0x4ba8a8e3
+.word	0xa25151f3, 0x5da3a3fe, 0x804040c0, 0x058f8f8a
+.word	0x3f9292ad, 0x219d9dbc, 0x70383848, 0xf1f5f504
+.word	0x63bcbcdf, 0x77b6b6c1, 0xafdada75, 0x42212163
+.word	0x20101030, 0xe5ffff1a, 0xfdf3f30e, 0xbfd2d26d
+.word	0x81cdcd4c, 0x180c0c14, 0x26131335, 0xc3ecec2f
+.word	0xbe5f5fe1, 0x359797a2, 0x884444cc, 0x2e171739
+.word	0x93c4c457, 0x55a7a7f2, 0xfc7e7e82, 0x7a3d3d47
+.word	0xc86464ac, 0xba5d5de7, 0x3219192b, 0xe6737395
+.word	0xc06060a0, 0x19818198, 0x9e4f4fd1, 0xa3dcdc7f
+.word	0x44222266, 0x542a2a7e, 0x3b9090ab, 0x0b888883
+.word	0x8c4646ca, 0xc7eeee29, 0x6bb8b8d3, 0x2814143c
+.word	0xa7dede79, 0xbc5e5ee2, 0x160b0b1d, 0xaddbdb76
+.word	0xdbe0e03b, 0x64323256, 0x743a3a4e, 0x140a0a1e
+.word	0x924949db, 0x0c06060a, 0x4824246c, 0xb85c5ce4
+.word	0x9fc2c25d, 0xbdd3d36e, 0x43acacef, 0xc46262a6
+.word	0x399191a8, 0x319595a4, 0xd3e4e437, 0xf279798b
+.word	0xd5e7e732, 0x8bc8c843, 0x6e373759, 0xda6d6db7
+.word	0x018d8d8c, 0xb1d5d564, 0x9c4e4ed2, 0x49a9a9e0
+.word	0xd86c6cb4, 0xac5656fa, 0xf3f4f407, 0xcfeaea25
+.word	0xca6565af, 0xf47a7a8e, 0x47aeaee9, 0x10080818
+.word	0x6fbabad5, 0xf0787888, 0x4a25256f, 0x5c2e2e72
+.word	0x381c1c24, 0x57a6a6f1, 0x73b4b4c7, 0x97c6c651
+.word	0xcbe8e823, 0xa1dddd7c, 0xe874749c, 0x3e1f1f21
+.word	0x964b4bdd, 0x61bdbddc, 0x0d8b8b86, 0x0f8a8a85
+.word	0xe0707090, 0x7c3e3e42, 0x71b5b5c4, 0xcc6666aa
+.word	0x904848d8, 0x06030305, 0xf7f6f601, 0x1c0e0e12
+.word	0xc26161a3, 0x6a35355f, 0xae5757f9, 0x69b9b9d0
+.word	0x17868691, 0x99c1c158, 0x3a1d1d27, 0x279e9eb9
+.word	0xd9e1e138, 0xebf8f813, 0x2b9898b3, 0x22111133
+.word	0xd26969bb, 0xa9d9d970, 0x078e8e89, 0x339494a7
+.word	0x2d9b9bb6, 0x3c1e1e22, 0x15878792, 0xc9e9e920
+.word	0x87cece49, 0xaa5555ff, 0x50282878, 0xa5dfdf7a
+.word	0x038c8c8f, 0x59a1a1f8, 0x09898980, 0x1a0d0d17
+.word	0x65bfbfda, 0xd7e6e631, 0x844242c6, 0xd06868b8
+.word	0x824141c3, 0x299999b0, 0x5a2d2d77, 0x1e0f0f11
+.word	0x7bb0b0cb, 0xa85454fc, 0x6dbbbbd6, 0x2c16163a
+@ Te4[256]
+.byte	0x63, 0x7c, 0x77, 0x7b, 0xf2, 0x6b, 0x6f, 0xc5
+.byte	0x30, 0x01, 0x67, 0x2b, 0xfe, 0xd7, 0xab, 0x76
+.byte	0xca, 0x82, 0xc9, 0x7d, 0xfa, 0x59, 0x47, 0xf0
+.byte	0xad, 0xd4, 0xa2, 0xaf, 0x9c, 0xa4, 0x72, 0xc0
+.byte	0xb7, 0xfd, 0x93, 0x26, 0x36, 0x3f, 0xf7, 0xcc
+.byte	0x34, 0xa5, 0xe5, 0xf1, 0x71, 0xd8, 0x31, 0x15
+.byte	0x04, 0xc7, 0x23, 0xc3, 0x18, 0x96, 0x05, 0x9a
+.byte	0x07, 0x12, 0x80, 0xe2, 0xeb, 0x27, 0xb2, 0x75
+.byte	0x09, 0x83, 0x2c, 0x1a, 0x1b, 0x6e, 0x5a, 0xa0
+.byte	0x52, 0x3b, 0xd6, 0xb3, 0x29, 0xe3, 0x2f, 0x84
+.byte	0x53, 0xd1, 0x00, 0xed, 0x20, 0xfc, 0xb1, 0x5b
+.byte	0x6a, 0xcb, 0xbe, 0x39, 0x4a, 0x4c, 0x58, 0xcf
+.byte	0xd0, 0xef, 0xaa, 0xfb, 0x43, 0x4d, 0x33, 0x85
+.byte	0x45, 0xf9, 0x02, 0x7f, 0x50, 0x3c, 0x9f, 0xa8
+.byte	0x51, 0xa3, 0x40, 0x8f, 0x92, 0x9d, 0x38, 0xf5
+.byte	0xbc, 0xb6, 0xda, 0x21, 0x10, 0xff, 0xf3, 0xd2
+.byte	0xcd, 0x0c, 0x13, 0xec, 0x5f, 0x97, 0x44, 0x17
+.byte	0xc4, 0xa7, 0x7e, 0x3d, 0x64, 0x5d, 0x19, 0x73
+.byte	0x60, 0x81, 0x4f, 0xdc, 0x22, 0x2a, 0x90, 0x88
+.byte	0x46, 0xee, 0xb8, 0x14, 0xde, 0x5e, 0x0b, 0xdb
+.byte	0xe0, 0x32, 0x3a, 0x0a, 0x49, 0x06, 0x24, 0x5c
+.byte	0xc2, 0xd3, 0xac, 0x62, 0x91, 0x95, 0xe4, 0x79
+.byte	0xe7, 0xc8, 0x37, 0x6d, 0x8d, 0xd5, 0x4e, 0xa9
+.byte	0x6c, 0x56, 0xf4, 0xea, 0x65, 0x7a, 0xae, 0x08
+.byte	0xba, 0x78, 0x25, 0x2e, 0x1c, 0xa6, 0xb4, 0xc6
+.byte	0xe8, 0xdd, 0x74, 0x1f, 0x4b, 0xbd, 0x8b, 0x8a
+.byte	0x70, 0x3e, 0xb5, 0x66, 0x48, 0x03, 0xf6, 0x0e
+.byte	0x61, 0x35, 0x57, 0xb9, 0x86, 0xc1, 0x1d, 0x9e
+.byte	0xe1, 0xf8, 0x98, 0x11, 0x69, 0xd9, 0x8e, 0x94
+.byte	0x9b, 0x1e, 0x87, 0xe9, 0xce, 0x55, 0x28, 0xdf
+.byte	0x8c, 0xa1, 0x89, 0x0d, 0xbf, 0xe6, 0x42, 0x68
+.byte	0x41, 0x99, 0x2d, 0x0f, 0xb0, 0x54, 0xbb, 0x16
+@ rcon[]
+.word	0x01000000, 0x02000000, 0x04000000, 0x08000000
+.word	0x10000000, 0x20000000, 0x40000000, 0x80000000
+.word	0x1B000000, 0x36000000, 0, 0, 0, 0, 0, 0
+.size	AES_Te,.-AES_Te
+
+@ void AES_encrypt(const unsigned char *in, unsigned char *out,
+@ 		 const AES_KEY *key) {
+.align	5
+ENTRY(AES_encrypt)
+	adr	r3,AES_encrypt
+	stmdb   sp!,{r1,r4-r12,lr}
+	mov	r12,r0		@ inp
+	mov	r11,r2
+	sub	r10,r3,#AES_encrypt-AES_Te	@ Te
+#if __ARM_ARCH__<7
+	ldrb	r0,[r12,#3]	@ load input data in endian-neutral
+	ldrb	r4,[r12,#2]	@ manner...
+	ldrb	r5,[r12,#1]
+	ldrb	r6,[r12,#0]
+	orr	r0,r0,r4,lsl#8
+	ldrb	r1,[r12,#7]
+	orr	r0,r0,r5,lsl#16
+	ldrb	r4,[r12,#6]
+	orr	r0,r0,r6,lsl#24
+	ldrb	r5,[r12,#5]
+	ldrb	r6,[r12,#4]
+	orr	r1,r1,r4,lsl#8
+	ldrb	r2,[r12,#11]
+	orr	r1,r1,r5,lsl#16
+	ldrb	r4,[r12,#10]
+	orr	r1,r1,r6,lsl#24
+	ldrb	r5,[r12,#9]
+	ldrb	r6,[r12,#8]
+	orr	r2,r2,r4,lsl#8
+	ldrb	r3,[r12,#15]
+	orr	r2,r2,r5,lsl#16
+	ldrb	r4,[r12,#14]
+	orr	r2,r2,r6,lsl#24
+	ldrb	r5,[r12,#13]
+	ldrb	r6,[r12,#12]
+	orr	r3,r3,r4,lsl#8
+	orr	r3,r3,r5,lsl#16
+	orr	r3,r3,r6,lsl#24
+#else
+	ldr	r0,[r12,#0]
+	ldr	r1,[r12,#4]
+	ldr	r2,[r12,#8]
+	ldr	r3,[r12,#12]
+#ifdef __ARMEL__
+	rev	r0,r0
+	rev	r1,r1
+	rev	r2,r2
+	rev	r3,r3
+#endif
+#endif
+	bl	_armv4_AES_encrypt
+
+	ldr	r12,[sp],#4		@ pop out
+#if __ARM_ARCH__>=7
+#ifdef __ARMEL__
+	rev	r0,r0
+	rev	r1,r1
+	rev	r2,r2
+	rev	r3,r3
+#endif
+	str	r0,[r12,#0]
+	str	r1,[r12,#4]
+	str	r2,[r12,#8]
+	str	r3,[r12,#12]
+#else
+	mov	r4,r0,lsr#24		@ write output in endian-neutral
+	mov	r5,r0,lsr#16		@ manner...
+	mov	r6,r0,lsr#8
+	strb	r4,[r12,#0]
+	strb	r5,[r12,#1]
+	mov	r4,r1,lsr#24
+	strb	r6,[r12,#2]
+	mov	r5,r1,lsr#16
+	strb	r0,[r12,#3]
+	mov	r6,r1,lsr#8
+	strb	r4,[r12,#4]
+	strb	r5,[r12,#5]
+	mov	r4,r2,lsr#24
+	strb	r6,[r12,#6]
+	mov	r5,r2,lsr#16
+	strb	r1,[r12,#7]
+	mov	r6,r2,lsr#8
+	strb	r4,[r12,#8]
+	strb	r5,[r12,#9]
+	mov	r4,r3,lsr#24
+	strb	r6,[r12,#10]
+	mov	r5,r3,lsr#16
+	strb	r2,[r12,#11]
+	mov	r6,r3,lsr#8
+	strb	r4,[r12,#12]
+	strb	r5,[r12,#13]
+	strb	r6,[r12,#14]
+	strb	r3,[r12,#15]
+#endif
+	ldmia	sp!,{r4-r12,pc}
+ENDPROC(AES_encrypt)
+
+.type   _armv4_AES_encrypt,%function
+.align	2
+_armv4_AES_encrypt:
+	str	lr,[sp,#-4]!		@ push lr
+	ldmia	r11!,{r4-r7}
+	eor	r0,r0,r4
+	ldr	r12,[r11,#240-16]
+	eor	r1,r1,r5
+	eor	r2,r2,r6
+	eor	r3,r3,r7
+	sub	r12,r12,#1
+	mov	lr,#255
+
+	and	r7,lr,r0
+	and	r8,lr,r0,lsr#8
+	and	r9,lr,r0,lsr#16
+	mov	r0,r0,lsr#24
+.Lenc_loop:
+	ldr	r4,[r10,r7,lsl#2]	@ Te3[s0>>0]
+	and	r7,lr,r1,lsr#16	@ i0
+	ldr	r5,[r10,r8,lsl#2]	@ Te2[s0>>8]
+	and	r8,lr,r1
+	ldr	r6,[r10,r9,lsl#2]	@ Te1[s0>>16]
+	and	r9,lr,r1,lsr#8
+	ldr	r0,[r10,r0,lsl#2]	@ Te0[s0>>24]
+	mov	r1,r1,lsr#24
+
+	ldr	r7,[r10,r7,lsl#2]	@ Te1[s1>>16]
+	ldr	r8,[r10,r8,lsl#2]	@ Te3[s1>>0]
+	ldr	r9,[r10,r9,lsl#2]	@ Te2[s1>>8]
+	eor	r0,r0,r7,ror#8
+	ldr	r1,[r10,r1,lsl#2]	@ Te0[s1>>24]
+	and	r7,lr,r2,lsr#8	@ i0
+	eor	r5,r5,r8,ror#8
+	and	r8,lr,r2,lsr#16	@ i1
+	eor	r6,r6,r9,ror#8
+	and	r9,lr,r2
+	ldr	r7,[r10,r7,lsl#2]	@ Te2[s2>>8]
+	eor	r1,r1,r4,ror#24
+	ldr	r8,[r10,r8,lsl#2]	@ Te1[s2>>16]
+	mov	r2,r2,lsr#24
+
+	ldr	r9,[r10,r9,lsl#2]	@ Te3[s2>>0]
+	eor	r0,r0,r7,ror#16
+	ldr	r2,[r10,r2,lsl#2]	@ Te0[s2>>24]
+	and	r7,lr,r3		@ i0
+	eor	r1,r1,r8,ror#8
+	and	r8,lr,r3,lsr#8	@ i1
+	eor	r6,r6,r9,ror#16
+	and	r9,lr,r3,lsr#16	@ i2
+	ldr	r7,[r10,r7,lsl#2]	@ Te3[s3>>0]
+	eor	r2,r2,r5,ror#16
+	ldr	r8,[r10,r8,lsl#2]	@ Te2[s3>>8]
+	mov	r3,r3,lsr#24
+
+	ldr	r9,[r10,r9,lsl#2]	@ Te1[s3>>16]
+	eor	r0,r0,r7,ror#24
+	ldr	r7,[r11],#16
+	eor	r1,r1,r8,ror#16
+	ldr	r3,[r10,r3,lsl#2]	@ Te0[s3>>24]
+	eor	r2,r2,r9,ror#8
+	ldr	r4,[r11,#-12]
+	eor	r3,r3,r6,ror#8
+
+	ldr	r5,[r11,#-8]
+	eor	r0,r0,r7
+	ldr	r6,[r11,#-4]
+	and	r7,lr,r0
+	eor	r1,r1,r4
+	and	r8,lr,r0,lsr#8
+	eor	r2,r2,r5
+	and	r9,lr,r0,lsr#16
+	eor	r3,r3,r6
+	mov	r0,r0,lsr#24
+
+	subs	r12,r12,#1
+	bne	.Lenc_loop
+
+	add	r10,r10,#2
+
+	ldrb	r4,[r10,r7,lsl#2]	@ Te4[s0>>0]
+	and	r7,lr,r1,lsr#16	@ i0
+	ldrb	r5,[r10,r8,lsl#2]	@ Te4[s0>>8]
+	and	r8,lr,r1
+	ldrb	r6,[r10,r9,lsl#2]	@ Te4[s0>>16]
+	and	r9,lr,r1,lsr#8
+	ldrb	r0,[r10,r0,lsl#2]	@ Te4[s0>>24]
+	mov	r1,r1,lsr#24
+
+	ldrb	r7,[r10,r7,lsl#2]	@ Te4[s1>>16]
+	ldrb	r8,[r10,r8,lsl#2]	@ Te4[s1>>0]
+	ldrb	r9,[r10,r9,lsl#2]	@ Te4[s1>>8]
+	eor	r0,r7,r0,lsl#8
+	ldrb	r1,[r10,r1,lsl#2]	@ Te4[s1>>24]
+	and	r7,lr,r2,lsr#8	@ i0
+	eor	r5,r8,r5,lsl#8
+	and	r8,lr,r2,lsr#16	@ i1
+	eor	r6,r9,r6,lsl#8
+	and	r9,lr,r2
+	ldrb	r7,[r10,r7,lsl#2]	@ Te4[s2>>8]
+	eor	r1,r4,r1,lsl#24
+	ldrb	r8,[r10,r8,lsl#2]	@ Te4[s2>>16]
+	mov	r2,r2,lsr#24
+
+	ldrb	r9,[r10,r9,lsl#2]	@ Te4[s2>>0]
+	eor	r0,r7,r0,lsl#8
+	ldrb	r2,[r10,r2,lsl#2]	@ Te4[s2>>24]
+	and	r7,lr,r3		@ i0
+	eor	r1,r1,r8,lsl#16
+	and	r8,lr,r3,lsr#8	@ i1
+	eor	r6,r9,r6,lsl#8
+	and	r9,lr,r3,lsr#16	@ i2
+	ldrb	r7,[r10,r7,lsl#2]	@ Te4[s3>>0]
+	eor	r2,r5,r2,lsl#24
+	ldrb	r8,[r10,r8,lsl#2]	@ Te4[s3>>8]
+	mov	r3,r3,lsr#24
+
+	ldrb	r9,[r10,r9,lsl#2]	@ Te4[s3>>16]
+	eor	r0,r7,r0,lsl#8
+	ldr	r7,[r11,#0]
+	ldrb	r3,[r10,r3,lsl#2]	@ Te4[s3>>24]
+	eor	r1,r1,r8,lsl#8
+	ldr	r4,[r11,#4]
+	eor	r2,r2,r9,lsl#16
+	ldr	r5,[r11,#8]
+	eor	r3,r6,r3,lsl#24
+	ldr	r6,[r11,#12]
+
+	eor	r0,r0,r7
+	eor	r1,r1,r4
+	eor	r2,r2,r5
+	eor	r3,r3,r6
+
+	sub	r10,r10,#2
+	ldr	pc,[sp],#4		@ pop and return
+.size	_armv4_AES_encrypt,.-_armv4_AES_encrypt
+
+.align	5
+ENTRY(private_AES_set_encrypt_key)
+_armv4_AES_set_encrypt_key:
+	adr	r3,_armv4_AES_set_encrypt_key
+	teq	r0,#0
+	moveq	r0,#-1
+	beq	.Labrt
+	teq	r2,#0
+	moveq	r0,#-1
+	beq	.Labrt
+
+	teq	r1,#128
+	beq	.Lok
+	teq	r1,#192
+	beq	.Lok
+	teq	r1,#256
+	movne	r0,#-1
+	bne	.Labrt
+
+.Lok:	stmdb   sp!,{r4-r12,lr}
+	sub	r10,r3,#_armv4_AES_set_encrypt_key-AES_Te-1024	@ Te4
+
+	mov	r12,r0		@ inp
+	mov	lr,r1			@ bits
+	mov	r11,r2			@ key
+
+#if __ARM_ARCH__<7
+	ldrb	r0,[r12,#3]	@ load input data in endian-neutral
+	ldrb	r4,[r12,#2]	@ manner...
+	ldrb	r5,[r12,#1]
+	ldrb	r6,[r12,#0]
+	orr	r0,r0,r4,lsl#8
+	ldrb	r1,[r12,#7]
+	orr	r0,r0,r5,lsl#16
+	ldrb	r4,[r12,#6]
+	orr	r0,r0,r6,lsl#24
+	ldrb	r5,[r12,#5]
+	ldrb	r6,[r12,#4]
+	orr	r1,r1,r4,lsl#8
+	ldrb	r2,[r12,#11]
+	orr	r1,r1,r5,lsl#16
+	ldrb	r4,[r12,#10]
+	orr	r1,r1,r6,lsl#24
+	ldrb	r5,[r12,#9]
+	ldrb	r6,[r12,#8]
+	orr	r2,r2,r4,lsl#8
+	ldrb	r3,[r12,#15]
+	orr	r2,r2,r5,lsl#16
+	ldrb	r4,[r12,#14]
+	orr	r2,r2,r6,lsl#24
+	ldrb	r5,[r12,#13]
+	ldrb	r6,[r12,#12]
+	orr	r3,r3,r4,lsl#8
+	str	r0,[r11],#16
+	orr	r3,r3,r5,lsl#16
+	str	r1,[r11,#-12]
+	orr	r3,r3,r6,lsl#24
+	str	r2,[r11,#-8]
+	str	r3,[r11,#-4]
+#else
+	ldr	r0,[r12,#0]
+	ldr	r1,[r12,#4]
+	ldr	r2,[r12,#8]
+	ldr	r3,[r12,#12]
+#ifdef __ARMEL__
+	rev	r0,r0
+	rev	r1,r1
+	rev	r2,r2
+	rev	r3,r3
+#endif
+	str	r0,[r11],#16
+	str	r1,[r11,#-12]
+	str	r2,[r11,#-8]
+	str	r3,[r11,#-4]
+#endif
+
+	teq	lr,#128
+	bne	.Lnot128
+	mov	r12,#10
+	str	r12,[r11,#240-16]
+	add	r6,r10,#256			@ rcon
+	mov	lr,#255
+
+.L128_loop:
+	and	r5,lr,r3,lsr#24
+	and	r7,lr,r3,lsr#16
+	ldrb	r5,[r10,r5]
+	and	r8,lr,r3,lsr#8
+	ldrb	r7,[r10,r7]
+	and	r9,lr,r3
+	ldrb	r8,[r10,r8]
+	orr	r5,r5,r7,lsl#24
+	ldrb	r9,[r10,r9]
+	orr	r5,r5,r8,lsl#16
+	ldr	r4,[r6],#4			@ rcon[i++]
+	orr	r5,r5,r9,lsl#8
+	eor	r5,r5,r4
+	eor	r0,r0,r5			@ rk[4]=rk[0]^...
+	eor	r1,r1,r0			@ rk[5]=rk[1]^rk[4]
+	str	r0,[r11],#16
+	eor	r2,r2,r1			@ rk[6]=rk[2]^rk[5]
+	str	r1,[r11,#-12]
+	eor	r3,r3,r2			@ rk[7]=rk[3]^rk[6]
+	str	r2,[r11,#-8]
+	subs	r12,r12,#1
+	str	r3,[r11,#-4]
+	bne	.L128_loop
+	sub	r2,r11,#176
+	b	.Ldone
+
+.Lnot128:
+#if __ARM_ARCH__<7
+	ldrb	r8,[r12,#19]
+	ldrb	r4,[r12,#18]
+	ldrb	r5,[r12,#17]
+	ldrb	r6,[r12,#16]
+	orr	r8,r8,r4,lsl#8
+	ldrb	r9,[r12,#23]
+	orr	r8,r8,r5,lsl#16
+	ldrb	r4,[r12,#22]
+	orr	r8,r8,r6,lsl#24
+	ldrb	r5,[r12,#21]
+	ldrb	r6,[r12,#20]
+	orr	r9,r9,r4,lsl#8
+	orr	r9,r9,r5,lsl#16
+	str	r8,[r11],#8
+	orr	r9,r9,r6,lsl#24
+	str	r9,[r11,#-4]
+#else
+	ldr	r8,[r12,#16]
+	ldr	r9,[r12,#20]
+#ifdef __ARMEL__
+	rev	r8,r8
+	rev	r9,r9
+#endif
+	str	r8,[r11],#8
+	str	r9,[r11,#-4]
+#endif
+
+	teq	lr,#192
+	bne	.Lnot192
+	mov	r12,#12
+	str	r12,[r11,#240-24]
+	add	r6,r10,#256			@ rcon
+	mov	lr,#255
+	mov	r12,#8
+
+.L192_loop:
+	and	r5,lr,r9,lsr#24
+	and	r7,lr,r9,lsr#16
+	ldrb	r5,[r10,r5]
+	and	r8,lr,r9,lsr#8
+	ldrb	r7,[r10,r7]
+	and	r9,lr,r9
+	ldrb	r8,[r10,r8]
+	orr	r5,r5,r7,lsl#24
+	ldrb	r9,[r10,r9]
+	orr	r5,r5,r8,lsl#16
+	ldr	r4,[r6],#4			@ rcon[i++]
+	orr	r5,r5,r9,lsl#8
+	eor	r9,r5,r4
+	eor	r0,r0,r9			@ rk[6]=rk[0]^...
+	eor	r1,r1,r0			@ rk[7]=rk[1]^rk[6]
+	str	r0,[r11],#24
+	eor	r2,r2,r1			@ rk[8]=rk[2]^rk[7]
+	str	r1,[r11,#-20]
+	eor	r3,r3,r2			@ rk[9]=rk[3]^rk[8]
+	str	r2,[r11,#-16]
+	subs	r12,r12,#1
+	str	r3,[r11,#-12]
+	subeq	r2,r11,#216
+	beq	.Ldone
+
+	ldr	r7,[r11,#-32]
+	ldr	r8,[r11,#-28]
+	eor	r7,r7,r3			@ rk[10]=rk[4]^rk[9]
+	eor	r9,r8,r7			@ rk[11]=rk[5]^rk[10]
+	str	r7,[r11,#-8]
+	str	r9,[r11,#-4]
+	b	.L192_loop
+
+.Lnot192:
+#if __ARM_ARCH__<7
+	ldrb	r8,[r12,#27]
+	ldrb	r4,[r12,#26]
+	ldrb	r5,[r12,#25]
+	ldrb	r6,[r12,#24]
+	orr	r8,r8,r4,lsl#8
+	ldrb	r9,[r12,#31]
+	orr	r8,r8,r5,lsl#16
+	ldrb	r4,[r12,#30]
+	orr	r8,r8,r6,lsl#24
+	ldrb	r5,[r12,#29]
+	ldrb	r6,[r12,#28]
+	orr	r9,r9,r4,lsl#8
+	orr	r9,r9,r5,lsl#16
+	str	r8,[r11],#8
+	orr	r9,r9,r6,lsl#24
+	str	r9,[r11,#-4]
+#else
+	ldr	r8,[r12,#24]
+	ldr	r9,[r12,#28]
+#ifdef __ARMEL__
+	rev	r8,r8
+	rev	r9,r9
+#endif
+	str	r8,[r11],#8
+	str	r9,[r11,#-4]
+#endif
+
+	mov	r12,#14
+	str	r12,[r11,#240-32]
+	add	r6,r10,#256			@ rcon
+	mov	lr,#255
+	mov	r12,#7
+
+.L256_loop:
+	and	r5,lr,r9,lsr#24
+	and	r7,lr,r9,lsr#16
+	ldrb	r5,[r10,r5]
+	and	r8,lr,r9,lsr#8
+	ldrb	r7,[r10,r7]
+	and	r9,lr,r9
+	ldrb	r8,[r10,r8]
+	orr	r5,r5,r7,lsl#24
+	ldrb	r9,[r10,r9]
+	orr	r5,r5,r8,lsl#16
+	ldr	r4,[r6],#4			@ rcon[i++]
+	orr	r5,r5,r9,lsl#8
+	eor	r9,r5,r4
+	eor	r0,r0,r9			@ rk[8]=rk[0]^...
+	eor	r1,r1,r0			@ rk[9]=rk[1]^rk[8]
+	str	r0,[r11],#32
+	eor	r2,r2,r1			@ rk[10]=rk[2]^rk[9]
+	str	r1,[r11,#-28]
+	eor	r3,r3,r2			@ rk[11]=rk[3]^rk[10]
+	str	r2,[r11,#-24]
+	subs	r12,r12,#1
+	str	r3,[r11,#-20]
+	subeq	r2,r11,#256
+	beq	.Ldone
+
+	and	r5,lr,r3
+	and	r7,lr,r3,lsr#8
+	ldrb	r5,[r10,r5]
+	and	r8,lr,r3,lsr#16
+	ldrb	r7,[r10,r7]
+	and	r9,lr,r3,lsr#24
+	ldrb	r8,[r10,r8]
+	orr	r5,r5,r7,lsl#8
+	ldrb	r9,[r10,r9]
+	orr	r5,r5,r8,lsl#16
+	ldr	r4,[r11,#-48]
+	orr	r5,r5,r9,lsl#24
+
+	ldr	r7,[r11,#-44]
+	ldr	r8,[r11,#-40]
+	eor	r4,r4,r5			@ rk[12]=rk[4]^...
+	ldr	r9,[r11,#-36]
+	eor	r7,r7,r4			@ rk[13]=rk[5]^rk[12]
+	str	r4,[r11,#-16]
+	eor	r8,r8,r7			@ rk[14]=rk[6]^rk[13]
+	str	r7,[r11,#-12]
+	eor	r9,r9,r8			@ rk[15]=rk[7]^rk[14]
+	str	r8,[r11,#-8]
+	str	r9,[r11,#-4]
+	b	.L256_loop
+
+.Ldone:	mov	r0,#0
+	ldmia   sp!,{r4-r12,lr}
+.Labrt:	ret	lr
+ENDPROC(private_AES_set_encrypt_key)
+
+.align	5
+ENTRY(private_AES_set_decrypt_key)
+	str	lr,[sp,#-4]!            @ push lr
+#if 0
+	@ kernel does both of these in setkey so optimise this bit out by
+	@ expecting the key to already have the enc_key work done (see aes_glue.c)
+	bl	_armv4_AES_set_encrypt_key
+#else
+	mov	r0,#0
+#endif
+	teq	r0,#0
+	ldrne	lr,[sp],#4              @ pop lr
+	bne	.Labrt
+
+	stmdb   sp!,{r4-r12}
+
+	ldr	r12,[r2,#240]	@ AES_set_encrypt_key preserves r2,
+	mov	r11,r2			@ which is AES_KEY *key
+	mov	r7,r2
+	add	r8,r2,r12,lsl#4
+
+.Linv:	ldr	r0,[r7]
+	ldr	r1,[r7,#4]
+	ldr	r2,[r7,#8]
+	ldr	r3,[r7,#12]
+	ldr	r4,[r8]
+	ldr	r5,[r8,#4]
+	ldr	r6,[r8,#8]
+	ldr	r9,[r8,#12]
+	str	r0,[r8],#-16
+	str	r1,[r8,#16+4]
+	str	r2,[r8,#16+8]
+	str	r3,[r8,#16+12]
+	str	r4,[r7],#16
+	str	r5,[r7,#-12]
+	str	r6,[r7,#-8]
+	str	r9,[r7,#-4]
+	teq	r7,r8
+	bne	.Linv
+	ldr	r0,[r11,#16]!		@ prefetch tp1
+	mov	r7,#0x80
+	mov	r8,#0x1b
+	orr	r7,r7,#0x8000
+	orr	r8,r8,#0x1b00
+	orr	r7,r7,r7,lsl#16
+	orr	r8,r8,r8,lsl#16
+	sub	r12,r12,#1
+	mvn	r9,r7
+	mov	r12,r12,lsl#2	@ (rounds-1)*4
+
+.Lmix:	and	r4,r0,r7
+	and	r1,r0,r9
+	sub	r4,r4,r4,lsr#7
+	and	r4,r4,r8
+	eor	r1,r4,r1,lsl#1	@ tp2
+
+	and	r4,r1,r7
+	and	r2,r1,r9
+	sub	r4,r4,r4,lsr#7
+	and	r4,r4,r8
+	eor	r2,r4,r2,lsl#1	@ tp4
+
+	and	r4,r2,r7
+	and	r3,r2,r9
+	sub	r4,r4,r4,lsr#7
+	and	r4,r4,r8
+	eor	r3,r4,r3,lsl#1	@ tp8
+
+	eor	r4,r1,r2
+	eor	r5,r0,r3		@ tp9
+	eor	r4,r4,r3		@ tpe
+	eor	r4,r4,r1,ror#24
+	eor	r4,r4,r5,ror#24	@ ^= ROTATE(tpb=tp9^tp2,8)
+	eor	r4,r4,r2,ror#16
+	eor	r4,r4,r5,ror#16	@ ^= ROTATE(tpd=tp9^tp4,16)
+	eor	r4,r4,r5,ror#8	@ ^= ROTATE(tp9,24)
+
+	ldr	r0,[r11,#4]		@ prefetch tp1
+	str	r4,[r11],#4
+	subs	r12,r12,#1
+	bne	.Lmix
+
+	mov	r0,#0
+	ldmia	sp!,{r4-r12,pc}
+ENDPROC(private_AES_set_decrypt_key)
+
+.type	AES_Td,%object
+.align	5
+AES_Td:
+.word	0x51f4a750, 0x7e416553, 0x1a17a4c3, 0x3a275e96
+.word	0x3bab6bcb, 0x1f9d45f1, 0xacfa58ab, 0x4be30393
+.word	0x2030fa55, 0xad766df6, 0x88cc7691, 0xf5024c25
+.word	0x4fe5d7fc, 0xc52acbd7, 0x26354480, 0xb562a38f
+.word	0xdeb15a49, 0x25ba1b67, 0x45ea0e98, 0x5dfec0e1
+.word	0xc32f7502, 0x814cf012, 0x8d4697a3, 0x6bd3f9c6
+.word	0x038f5fe7, 0x15929c95, 0xbf6d7aeb, 0x955259da
+.word	0xd4be832d, 0x587421d3, 0x49e06929, 0x8ec9c844
+.word	0x75c2896a, 0xf48e7978, 0x99583e6b, 0x27b971dd
+.word	0xbee14fb6, 0xf088ad17, 0xc920ac66, 0x7dce3ab4
+.word	0x63df4a18, 0xe51a3182, 0x97513360, 0x62537f45
+.word	0xb16477e0, 0xbb6bae84, 0xfe81a01c, 0xf9082b94
+.word	0x70486858, 0x8f45fd19, 0x94de6c87, 0x527bf8b7
+.word	0xab73d323, 0x724b02e2, 0xe31f8f57, 0x6655ab2a
+.word	0xb2eb2807, 0x2fb5c203, 0x86c57b9a, 0xd33708a5
+.word	0x302887f2, 0x23bfa5b2, 0x02036aba, 0xed16825c
+.word	0x8acf1c2b, 0xa779b492, 0xf307f2f0, 0x4e69e2a1
+.word	0x65daf4cd, 0x0605bed5, 0xd134621f, 0xc4a6fe8a
+.word	0x342e539d, 0xa2f355a0, 0x058ae132, 0xa4f6eb75
+.word	0x0b83ec39, 0x4060efaa, 0x5e719f06, 0xbd6e1051
+.word	0x3e218af9, 0x96dd063d, 0xdd3e05ae, 0x4de6bd46
+.word	0x91548db5, 0x71c45d05, 0x0406d46f, 0x605015ff
+.word	0x1998fb24, 0xd6bde997, 0x894043cc, 0x67d99e77
+.word	0xb0e842bd, 0x07898b88, 0xe7195b38, 0x79c8eedb
+.word	0xa17c0a47, 0x7c420fe9, 0xf8841ec9, 0x00000000
+.word	0x09808683, 0x322bed48, 0x1e1170ac, 0x6c5a724e
+.word	0xfd0efffb, 0x0f853856, 0x3daed51e, 0x362d3927
+.word	0x0a0fd964, 0x685ca621, 0x9b5b54d1, 0x24362e3a
+.word	0x0c0a67b1, 0x9357e70f, 0xb4ee96d2, 0x1b9b919e
+.word	0x80c0c54f, 0x61dc20a2, 0x5a774b69, 0x1c121a16
+.word	0xe293ba0a, 0xc0a02ae5, 0x3c22e043, 0x121b171d
+.word	0x0e090d0b, 0xf28bc7ad, 0x2db6a8b9, 0x141ea9c8
+.word	0x57f11985, 0xaf75074c, 0xee99ddbb, 0xa37f60fd
+.word	0xf701269f, 0x5c72f5bc, 0x44663bc5, 0x5bfb7e34
+.word	0x8b432976, 0xcb23c6dc, 0xb6edfc68, 0xb8e4f163
+.word	0xd731dcca, 0x42638510, 0x13972240, 0x84c61120
+.word	0x854a247d, 0xd2bb3df8, 0xaef93211, 0xc729a16d
+.word	0x1d9e2f4b, 0xdcb230f3, 0x0d8652ec, 0x77c1e3d0
+.word	0x2bb3166c, 0xa970b999, 0x119448fa, 0x47e96422
+.word	0xa8fc8cc4, 0xa0f03f1a, 0x567d2cd8, 0x223390ef
+.word	0x87494ec7, 0xd938d1c1, 0x8ccaa2fe, 0x98d40b36
+.word	0xa6f581cf, 0xa57ade28, 0xdab78e26, 0x3fadbfa4
+.word	0x2c3a9de4, 0x5078920d, 0x6a5fcc9b, 0x547e4662
+.word	0xf68d13c2, 0x90d8b8e8, 0x2e39f75e, 0x82c3aff5
+.word	0x9f5d80be, 0x69d0937c, 0x6fd52da9, 0xcf2512b3
+.word	0xc8ac993b, 0x10187da7, 0xe89c636e, 0xdb3bbb7b
+.word	0xcd267809, 0x6e5918f4, 0xec9ab701, 0x834f9aa8
+.word	0xe6956e65, 0xaaffe67e, 0x21bccf08, 0xef15e8e6
+.word	0xbae79bd9, 0x4a6f36ce, 0xea9f09d4, 0x29b07cd6
+.word	0x31a4b2af, 0x2a3f2331, 0xc6a59430, 0x35a266c0
+.word	0x744ebc37, 0xfc82caa6, 0xe090d0b0, 0x33a7d815
+.word	0xf104984a, 0x41ecdaf7, 0x7fcd500e, 0x1791f62f
+.word	0x764dd68d, 0x43efb04d, 0xccaa4d54, 0xe49604df
+.word	0x9ed1b5e3, 0x4c6a881b, 0xc12c1fb8, 0x4665517f
+.word	0x9d5eea04, 0x018c355d, 0xfa877473, 0xfb0b412e
+.word	0xb3671d5a, 0x92dbd252, 0xe9105633, 0x6dd64713
+.word	0x9ad7618c, 0x37a10c7a, 0x59f8148e, 0xeb133c89
+.word	0xcea927ee, 0xb761c935, 0xe11ce5ed, 0x7a47b13c
+.word	0x9cd2df59, 0x55f2733f, 0x1814ce79, 0x73c737bf
+.word	0x53f7cdea, 0x5ffdaa5b, 0xdf3d6f14, 0x7844db86
+.word	0xcaaff381, 0xb968c43e, 0x3824342c, 0xc2a3405f
+.word	0x161dc372, 0xbce2250c, 0x283c498b, 0xff0d9541
+.word	0x39a80171, 0x080cb3de, 0xd8b4e49c, 0x6456c190
+.word	0x7bcb8461, 0xd532b670, 0x486c5c74, 0xd0b85742
+@ Td4[256]
+.byte	0x52, 0x09, 0x6a, 0xd5, 0x30, 0x36, 0xa5, 0x38
+.byte	0xbf, 0x40, 0xa3, 0x9e, 0x81, 0xf3, 0xd7, 0xfb
+.byte	0x7c, 0xe3, 0x39, 0x82, 0x9b, 0x2f, 0xff, 0x87
+.byte	0x34, 0x8e, 0x43, 0x44, 0xc4, 0xde, 0xe9, 0xcb
+.byte	0x54, 0x7b, 0x94, 0x32, 0xa6, 0xc2, 0x23, 0x3d
+.byte	0xee, 0x4c, 0x95, 0x0b, 0x42, 0xfa, 0xc3, 0x4e
+.byte	0x08, 0x2e, 0xa1, 0x66, 0x28, 0xd9, 0x24, 0xb2
+.byte	0x76, 0x5b, 0xa2, 0x49, 0x6d, 0x8b, 0xd1, 0x25
+.byte	0x72, 0xf8, 0xf6, 0x64, 0x86, 0x68, 0x98, 0x16
+.byte	0xd4, 0xa4, 0x5c, 0xcc, 0x5d, 0x65, 0xb6, 0x92
+.byte	0x6c, 0x70, 0x48, 0x50, 0xfd, 0xed, 0xb9, 0xda
+.byte	0x5e, 0x15, 0x46, 0x57, 0xa7, 0x8d, 0x9d, 0x84
+.byte	0x90, 0xd8, 0xab, 0x00, 0x8c, 0xbc, 0xd3, 0x0a
+.byte	0xf7, 0xe4, 0x58, 0x05, 0xb8, 0xb3, 0x45, 0x06
+.byte	0xd0, 0x2c, 0x1e, 0x8f, 0xca, 0x3f, 0x0f, 0x02
+.byte	0xc1, 0xaf, 0xbd, 0x03, 0x01, 0x13, 0x8a, 0x6b
+.byte	0x3a, 0x91, 0x11, 0x41, 0x4f, 0x67, 0xdc, 0xea
+.byte	0x97, 0xf2, 0xcf, 0xce, 0xf0, 0xb4, 0xe6, 0x73
+.byte	0x96, 0xac, 0x74, 0x22, 0xe7, 0xad, 0x35, 0x85
+.byte	0xe2, 0xf9, 0x37, 0xe8, 0x1c, 0x75, 0xdf, 0x6e
+.byte	0x47, 0xf1, 0x1a, 0x71, 0x1d, 0x29, 0xc5, 0x89
+.byte	0x6f, 0xb7, 0x62, 0x0e, 0xaa, 0x18, 0xbe, 0x1b
+.byte	0xfc, 0x56, 0x3e, 0x4b, 0xc6, 0xd2, 0x79, 0x20
+.byte	0x9a, 0xdb, 0xc0, 0xfe, 0x78, 0xcd, 0x5a, 0xf4
+.byte	0x1f, 0xdd, 0xa8, 0x33, 0x88, 0x07, 0xc7, 0x31
+.byte	0xb1, 0x12, 0x10, 0x59, 0x27, 0x80, 0xec, 0x5f
+.byte	0x60, 0x51, 0x7f, 0xa9, 0x19, 0xb5, 0x4a, 0x0d
+.byte	0x2d, 0xe5, 0x7a, 0x9f, 0x93, 0xc9, 0x9c, 0xef
+.byte	0xa0, 0xe0, 0x3b, 0x4d, 0xae, 0x2a, 0xf5, 0xb0
+.byte	0xc8, 0xeb, 0xbb, 0x3c, 0x83, 0x53, 0x99, 0x61
+.byte	0x17, 0x2b, 0x04, 0x7e, 0xba, 0x77, 0xd6, 0x26
+.byte	0xe1, 0x69, 0x14, 0x63, 0x55, 0x21, 0x0c, 0x7d
+.size	AES_Td,.-AES_Td
+
+@ void AES_decrypt(const unsigned char *in, unsigned char *out,
+@ 		 const AES_KEY *key) {
+.align	5
+ENTRY(AES_decrypt)
+	adr	r3,AES_decrypt
+	stmdb   sp!,{r1,r4-r12,lr}
+	mov	r12,r0		@ inp
+	mov	r11,r2
+	sub	r10,r3,#AES_decrypt-AES_Td		@ Td
+#if __ARM_ARCH__<7
+	ldrb	r0,[r12,#3]	@ load input data in endian-neutral
+	ldrb	r4,[r12,#2]	@ manner...
+	ldrb	r5,[r12,#1]
+	ldrb	r6,[r12,#0]
+	orr	r0,r0,r4,lsl#8
+	ldrb	r1,[r12,#7]
+	orr	r0,r0,r5,lsl#16
+	ldrb	r4,[r12,#6]
+	orr	r0,r0,r6,lsl#24
+	ldrb	r5,[r12,#5]
+	ldrb	r6,[r12,#4]
+	orr	r1,r1,r4,lsl#8
+	ldrb	r2,[r12,#11]
+	orr	r1,r1,r5,lsl#16
+	ldrb	r4,[r12,#10]
+	orr	r1,r1,r6,lsl#24
+	ldrb	r5,[r12,#9]
+	ldrb	r6,[r12,#8]
+	orr	r2,r2,r4,lsl#8
+	ldrb	r3,[r12,#15]
+	orr	r2,r2,r5,lsl#16
+	ldrb	r4,[r12,#14]
+	orr	r2,r2,r6,lsl#24
+	ldrb	r5,[r12,#13]
+	ldrb	r6,[r12,#12]
+	orr	r3,r3,r4,lsl#8
+	orr	r3,r3,r5,lsl#16
+	orr	r3,r3,r6,lsl#24
+#else
+	ldr	r0,[r12,#0]
+	ldr	r1,[r12,#4]
+	ldr	r2,[r12,#8]
+	ldr	r3,[r12,#12]
+#ifdef __ARMEL__
+	rev	r0,r0
+	rev	r1,r1
+	rev	r2,r2
+	rev	r3,r3
+#endif
+#endif
+	bl	_armv4_AES_decrypt
+
+	ldr	r12,[sp],#4		@ pop out
+#if __ARM_ARCH__>=7
+#ifdef __ARMEL__
+	rev	r0,r0
+	rev	r1,r1
+	rev	r2,r2
+	rev	r3,r3
+#endif
+	str	r0,[r12,#0]
+	str	r1,[r12,#4]
+	str	r2,[r12,#8]
+	str	r3,[r12,#12]
+#else
+	mov	r4,r0,lsr#24		@ write output in endian-neutral
+	mov	r5,r0,lsr#16		@ manner...
+	mov	r6,r0,lsr#8
+	strb	r4,[r12,#0]
+	strb	r5,[r12,#1]
+	mov	r4,r1,lsr#24
+	strb	r6,[r12,#2]
+	mov	r5,r1,lsr#16
+	strb	r0,[r12,#3]
+	mov	r6,r1,lsr#8
+	strb	r4,[r12,#4]
+	strb	r5,[r12,#5]
+	mov	r4,r2,lsr#24
+	strb	r6,[r12,#6]
+	mov	r5,r2,lsr#16
+	strb	r1,[r12,#7]
+	mov	r6,r2,lsr#8
+	strb	r4,[r12,#8]
+	strb	r5,[r12,#9]
+	mov	r4,r3,lsr#24
+	strb	r6,[r12,#10]
+	mov	r5,r3,lsr#16
+	strb	r2,[r12,#11]
+	mov	r6,r3,lsr#8
+	strb	r4,[r12,#12]
+	strb	r5,[r12,#13]
+	strb	r6,[r12,#14]
+	strb	r3,[r12,#15]
+#endif
+	ldmia	sp!,{r4-r12,pc}
+ENDPROC(AES_decrypt)
+
+.type   _armv4_AES_decrypt,%function
+.align	2
+_armv4_AES_decrypt:
+	str	lr,[sp,#-4]!		@ push lr
+	ldmia	r11!,{r4-r7}
+	eor	r0,r0,r4
+	ldr	r12,[r11,#240-16]
+	eor	r1,r1,r5
+	eor	r2,r2,r6
+	eor	r3,r3,r7
+	sub	r12,r12,#1
+	mov	lr,#255
+
+	and	r7,lr,r0,lsr#16
+	and	r8,lr,r0,lsr#8
+	and	r9,lr,r0
+	mov	r0,r0,lsr#24
+.Ldec_loop:
+	ldr	r4,[r10,r7,lsl#2]	@ Td1[s0>>16]
+	and	r7,lr,r1		@ i0
+	ldr	r5,[r10,r8,lsl#2]	@ Td2[s0>>8]
+	and	r8,lr,r1,lsr#16
+	ldr	r6,[r10,r9,lsl#2]	@ Td3[s0>>0]
+	and	r9,lr,r1,lsr#8
+	ldr	r0,[r10,r0,lsl#2]	@ Td0[s0>>24]
+	mov	r1,r1,lsr#24
+
+	ldr	r7,[r10,r7,lsl#2]	@ Td3[s1>>0]
+	ldr	r8,[r10,r8,lsl#2]	@ Td1[s1>>16]
+	ldr	r9,[r10,r9,lsl#2]	@ Td2[s1>>8]
+	eor	r0,r0,r7,ror#24
+	ldr	r1,[r10,r1,lsl#2]	@ Td0[s1>>24]
+	and	r7,lr,r2,lsr#8	@ i0
+	eor	r5,r8,r5,ror#8
+	and	r8,lr,r2		@ i1
+	eor	r6,r9,r6,ror#8
+	and	r9,lr,r2,lsr#16
+	ldr	r7,[r10,r7,lsl#2]	@ Td2[s2>>8]
+	eor	r1,r1,r4,ror#8
+	ldr	r8,[r10,r8,lsl#2]	@ Td3[s2>>0]
+	mov	r2,r2,lsr#24
+
+	ldr	r9,[r10,r9,lsl#2]	@ Td1[s2>>16]
+	eor	r0,r0,r7,ror#16
+	ldr	r2,[r10,r2,lsl#2]	@ Td0[s2>>24]
+	and	r7,lr,r3,lsr#16	@ i0
+	eor	r1,r1,r8,ror#24
+	and	r8,lr,r3,lsr#8	@ i1
+	eor	r6,r9,r6,ror#8
+	and	r9,lr,r3		@ i2
+	ldr	r7,[r10,r7,lsl#2]	@ Td1[s3>>16]
+	eor	r2,r2,r5,ror#8
+	ldr	r8,[r10,r8,lsl#2]	@ Td2[s3>>8]
+	mov	r3,r3,lsr#24
+
+	ldr	r9,[r10,r9,lsl#2]	@ Td3[s3>>0]
+	eor	r0,r0,r7,ror#8
+	ldr	r7,[r11],#16
+	eor	r1,r1,r8,ror#16
+	ldr	r3,[r10,r3,lsl#2]	@ Td0[s3>>24]
+	eor	r2,r2,r9,ror#24
+
+	ldr	r4,[r11,#-12]
+	eor	r0,r0,r7
+	ldr	r5,[r11,#-8]
+	eor	r3,r3,r6,ror#8
+	ldr	r6,[r11,#-4]
+	and	r7,lr,r0,lsr#16
+	eor	r1,r1,r4
+	and	r8,lr,r0,lsr#8
+	eor	r2,r2,r5
+	and	r9,lr,r0
+	eor	r3,r3,r6
+	mov	r0,r0,lsr#24
+
+	subs	r12,r12,#1
+	bne	.Ldec_loop
+
+	add	r10,r10,#1024
+
+	ldr	r5,[r10,#0]		@ prefetch Td4
+	ldr	r6,[r10,#32]
+	ldr	r4,[r10,#64]
+	ldr	r5,[r10,#96]
+	ldr	r6,[r10,#128]
+	ldr	r4,[r10,#160]
+	ldr	r5,[r10,#192]
+	ldr	r6,[r10,#224]
+
+	ldrb	r0,[r10,r0]		@ Td4[s0>>24]
+	ldrb	r4,[r10,r7]		@ Td4[s0>>16]
+	and	r7,lr,r1		@ i0
+	ldrb	r5,[r10,r8]		@ Td4[s0>>8]
+	and	r8,lr,r1,lsr#16
+	ldrb	r6,[r10,r9]		@ Td4[s0>>0]
+	and	r9,lr,r1,lsr#8
+
+	ldrb	r7,[r10,r7]		@ Td4[s1>>0]
+ ARM(	ldrb	r1,[r10,r1,lsr#24]  )	@ Td4[s1>>24]
+ THUMB(	add	r1,r10,r1,lsr#24    ) 	@ Td4[s1>>24]
+ THUMB(	ldrb	r1,[r1]		    )
+	ldrb	r8,[r10,r8]		@ Td4[s1>>16]
+	eor	r0,r7,r0,lsl#24
+	ldrb	r9,[r10,r9]		@ Td4[s1>>8]
+	eor	r1,r4,r1,lsl#8
+	and	r7,lr,r2,lsr#8	@ i0
+	eor	r5,r5,r8,lsl#8
+	and	r8,lr,r2		@ i1
+	ldrb	r7,[r10,r7]		@ Td4[s2>>8]
+	eor	r6,r6,r9,lsl#8
+	ldrb	r8,[r10,r8]		@ Td4[s2>>0]
+	and	r9,lr,r2,lsr#16
+
+ ARM(	ldrb	r2,[r10,r2,lsr#24]  )	@ Td4[s2>>24]
+ THUMB(	add	r2,r10,r2,lsr#24    )	@ Td4[s2>>24]
+ THUMB(	ldrb	r2,[r2]		    )
+	eor	r0,r0,r7,lsl#8
+	ldrb	r9,[r10,r9]		@ Td4[s2>>16]
+	eor	r1,r8,r1,lsl#16
+	and	r7,lr,r3,lsr#16	@ i0
+	eor	r2,r5,r2,lsl#16
+	and	r8,lr,r3,lsr#8	@ i1
+	ldrb	r7,[r10,r7]		@ Td4[s3>>16]
+	eor	r6,r6,r9,lsl#16
+	ldrb	r8,[r10,r8]		@ Td4[s3>>8]
+	and	r9,lr,r3		@ i2
+
+	ldrb	r9,[r10,r9]		@ Td4[s3>>0]
+ ARM(	ldrb	r3,[r10,r3,lsr#24]  )	@ Td4[s3>>24]
+ THUMB(	add	r3,r10,r3,lsr#24    )	@ Td4[s3>>24]
+ THUMB(	ldrb	r3,[r3]		    )
+	eor	r0,r0,r7,lsl#16
+	ldr	r7,[r11,#0]
+	eor	r1,r1,r8,lsl#8
+	ldr	r4,[r11,#4]
+	eor	r2,r9,r2,lsl#8
+	ldr	r5,[r11,#8]
+	eor	r3,r6,r3,lsl#24
+	ldr	r6,[r11,#12]
+
+	eor	r0,r0,r7
+	eor	r1,r1,r4
+	eor	r2,r2,r5
+	eor	r3,r3,r6
+
+	sub	r10,r10,#1024
+	ldr	pc,[sp],#4		@ pop and return
+.size	_armv4_AES_decrypt,.-_armv4_AES_decrypt
+.asciz	"AES for ARMv4, CRYPTOGAMS by <appro@openssl.org>"
+.align	2
diff --git a/arch/arm/crypto/aes-ce-core.S b/arch/arm/crypto/aes-ce-core.S
new file mode 100644
index 000000000..8cfa468ee
--- /dev/null
+++ b/arch/arm/crypto/aes-ce-core.S
@@ -0,0 +1,518 @@
+/*
+ * aes-ce-core.S - AES in CBC/CTR/XTS mode using ARMv8 Crypto Extensions
+ *
+ * Copyright (C) 2015 Linaro Ltd <ard.biesheuvel@linaro.org>
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2 as
+ * published by the Free Software Foundation.
+ */
+
+#include <linux/linkage.h>
+#include <asm/assembler.h>
+
+	.text
+	.fpu		crypto-neon-fp-armv8
+	.align		3
+
+	.macro		enc_round, state, key
+	aese.8		\state, \key
+	aesmc.8		\state, \state
+	.endm
+
+	.macro		dec_round, state, key
+	aesd.8		\state, \key
+	aesimc.8	\state, \state
+	.endm
+
+	.macro		enc_dround, key1, key2
+	enc_round	q0, \key1
+	enc_round	q0, \key2
+	.endm
+
+	.macro		dec_dround, key1, key2
+	dec_round	q0, \key1
+	dec_round	q0, \key2
+	.endm
+
+	.macro		enc_fround, key1, key2, key3
+	enc_round	q0, \key1
+	aese.8		q0, \key2
+	veor		q0, q0, \key3
+	.endm
+
+	.macro		dec_fround, key1, key2, key3
+	dec_round	q0, \key1
+	aesd.8		q0, \key2
+	veor		q0, q0, \key3
+	.endm
+
+	.macro		enc_dround_3x, key1, key2
+	enc_round	q0, \key1
+	enc_round	q1, \key1
+	enc_round	q2, \key1
+	enc_round	q0, \key2
+	enc_round	q1, \key2
+	enc_round	q2, \key2
+	.endm
+
+	.macro		dec_dround_3x, key1, key2
+	dec_round	q0, \key1
+	dec_round	q1, \key1
+	dec_round	q2, \key1
+	dec_round	q0, \key2
+	dec_round	q1, \key2
+	dec_round	q2, \key2
+	.endm
+
+	.macro		enc_fround_3x, key1, key2, key3
+	enc_round	q0, \key1
+	enc_round	q1, \key1
+	enc_round	q2, \key1
+	aese.8		q0, \key2
+	aese.8		q1, \key2
+	aese.8		q2, \key2
+	veor		q0, q0, \key3
+	veor		q1, q1, \key3
+	veor		q2, q2, \key3
+	.endm
+
+	.macro		dec_fround_3x, key1, key2, key3
+	dec_round	q0, \key1
+	dec_round	q1, \key1
+	dec_round	q2, \key1
+	aesd.8		q0, \key2
+	aesd.8		q1, \key2
+	aesd.8		q2, \key2
+	veor		q0, q0, \key3
+	veor		q1, q1, \key3
+	veor		q2, q2, \key3
+	.endm
+
+	.macro		do_block, dround, fround
+	cmp		r3, #12			@ which key size?
+	vld1.8		{q10-q11}, [ip]!
+	\dround		q8, q9
+	vld1.8		{q12-q13}, [ip]!
+	\dround		q10, q11
+	vld1.8		{q10-q11}, [ip]!
+	\dround		q12, q13
+	vld1.8		{q12-q13}, [ip]!
+	\dround		q10, q11
+	blo		0f			@ AES-128: 10 rounds
+	vld1.8		{q10-q11}, [ip]!
+	beq		1f			@ AES-192: 12 rounds
+	\dround		q12, q13
+	vld1.8		{q12-q13}, [ip]
+	\dround		q10, q11
+0:	\fround		q12, q13, q14
+	bx		lr
+
+1:	\dround		q12, q13
+	\fround		q10, q11, q14
+	bx		lr
+	.endm
+
+	/*
+	 * Internal, non-AAPCS compliant functions that implement the core AES
+	 * transforms. These should preserve all registers except q0 - q2 and ip
+	 * Arguments:
+	 *   q0        : first in/output block
+	 *   q1        : second in/output block (_3x version only)
+	 *   q2        : third in/output block (_3x version only)
+	 *   q8        : first round key
+	 *   q9        : secound round key
+	 *   ip        : address of 3rd round key
+	 *   q14       : final round key
+	 *   r3        : number of rounds
+	 */
+	.align		6
+aes_encrypt:
+	add		ip, r2, #32		@ 3rd round key
+.Laes_encrypt_tweak:
+	do_block	enc_dround, enc_fround
+ENDPROC(aes_encrypt)
+
+	.align		6
+aes_decrypt:
+	add		ip, r2, #32		@ 3rd round key
+	do_block	dec_dround, dec_fround
+ENDPROC(aes_decrypt)
+
+	.align		6
+aes_encrypt_3x:
+	add		ip, r2, #32		@ 3rd round key
+	do_block	enc_dround_3x, enc_fround_3x
+ENDPROC(aes_encrypt_3x)
+
+	.align		6
+aes_decrypt_3x:
+	add		ip, r2, #32		@ 3rd round key
+	do_block	dec_dround_3x, dec_fround_3x
+ENDPROC(aes_decrypt_3x)
+
+	.macro		prepare_key, rk, rounds
+	add		ip, \rk, \rounds, lsl #4
+	vld1.8		{q8-q9}, [\rk]		@ load first 2 round keys
+	vld1.8		{q14}, [ip]		@ load last round key
+	.endm
+
+	/*
+	 * aes_ecb_encrypt(u8 out[], u8 const in[], u8 const rk[], int rounds,
+	 *		   int blocks)
+	 * aes_ecb_decrypt(u8 out[], u8 const in[], u8 const rk[], int rounds,
+	 *		   int blocks)
+	 */
+ENTRY(ce_aes_ecb_encrypt)
+	push		{r4, lr}
+	ldr		r4, [sp, #8]
+	prepare_key	r2, r3
+.Lecbencloop3x:
+	subs		r4, r4, #3
+	bmi		.Lecbenc1x
+	vld1.8		{q0-q1}, [r1, :64]!
+	vld1.8		{q2}, [r1, :64]!
+	bl		aes_encrypt_3x
+	vst1.8		{q0-q1}, [r0, :64]!
+	vst1.8		{q2}, [r0, :64]!
+	b		.Lecbencloop3x
+.Lecbenc1x:
+	adds		r4, r4, #3
+	beq		.Lecbencout
+.Lecbencloop:
+	vld1.8		{q0}, [r1, :64]!
+	bl		aes_encrypt
+	vst1.8		{q0}, [r0, :64]!
+	subs		r4, r4, #1
+	bne		.Lecbencloop
+.Lecbencout:
+	pop		{r4, pc}
+ENDPROC(ce_aes_ecb_encrypt)
+
+ENTRY(ce_aes_ecb_decrypt)
+	push		{r4, lr}
+	ldr		r4, [sp, #8]
+	prepare_key	r2, r3
+.Lecbdecloop3x:
+	subs		r4, r4, #3
+	bmi		.Lecbdec1x
+	vld1.8		{q0-q1}, [r1, :64]!
+	vld1.8		{q2}, [r1, :64]!
+	bl		aes_decrypt_3x
+	vst1.8		{q0-q1}, [r0, :64]!
+	vst1.8		{q2}, [r0, :64]!
+	b		.Lecbdecloop3x
+.Lecbdec1x:
+	adds		r4, r4, #3
+	beq		.Lecbdecout
+.Lecbdecloop:
+	vld1.8		{q0}, [r1, :64]!
+	bl		aes_decrypt
+	vst1.8		{q0}, [r0, :64]!
+	subs		r4, r4, #1
+	bne		.Lecbdecloop
+.Lecbdecout:
+	pop		{r4, pc}
+ENDPROC(ce_aes_ecb_decrypt)
+
+	/*
+	 * aes_cbc_encrypt(u8 out[], u8 const in[], u8 const rk[], int rounds,
+	 *		   int blocks, u8 iv[])
+	 * aes_cbc_decrypt(u8 out[], u8 const in[], u8 const rk[], int rounds,
+	 *		   int blocks, u8 iv[])
+	 */
+ENTRY(ce_aes_cbc_encrypt)
+	push		{r4-r6, lr}
+	ldrd		r4, r5, [sp, #16]
+	vld1.8		{q0}, [r5]
+	prepare_key	r2, r3
+.Lcbcencloop:
+	vld1.8		{q1}, [r1, :64]!	@ get next pt block
+	veor		q0, q0, q1		@ ..and xor with iv
+	bl		aes_encrypt
+	vst1.8		{q0}, [r0, :64]!
+	subs		r4, r4, #1
+	bne		.Lcbcencloop
+	vst1.8		{q0}, [r5]
+	pop		{r4-r6, pc}
+ENDPROC(ce_aes_cbc_encrypt)
+
+ENTRY(ce_aes_cbc_decrypt)
+	push		{r4-r6, lr}
+	ldrd		r4, r5, [sp, #16]
+	vld1.8		{q6}, [r5]		@ keep iv in q6
+	prepare_key	r2, r3
+.Lcbcdecloop3x:
+	subs		r4, r4, #3
+	bmi		.Lcbcdec1x
+	vld1.8		{q0-q1}, [r1, :64]!
+	vld1.8		{q2}, [r1, :64]!
+	vmov		q3, q0
+	vmov		q4, q1
+	vmov		q5, q2
+	bl		aes_decrypt_3x
+	veor		q0, q0, q6
+	veor		q1, q1, q3
+	veor		q2, q2, q4
+	vmov		q6, q5
+	vst1.8		{q0-q1}, [r0, :64]!
+	vst1.8		{q2}, [r0, :64]!
+	b		.Lcbcdecloop3x
+.Lcbcdec1x:
+	adds		r4, r4, #3
+	beq		.Lcbcdecout
+	vmov		q15, q14		@ preserve last round key
+.Lcbcdecloop:
+	vld1.8		{q0}, [r1, :64]!	@ get next ct block
+	veor		q14, q15, q6		@ combine prev ct with last key
+	vmov		q6, q0
+	bl		aes_decrypt
+	vst1.8		{q0}, [r0, :64]!
+	subs		r4, r4, #1
+	bne		.Lcbcdecloop
+.Lcbcdecout:
+	vst1.8		{q6}, [r5]		@ keep iv in q6
+	pop		{r4-r6, pc}
+ENDPROC(ce_aes_cbc_decrypt)
+
+	/*
+	 * aes_ctr_encrypt(u8 out[], u8 const in[], u8 const rk[], int rounds,
+	 *		   int blocks, u8 ctr[])
+	 */
+ENTRY(ce_aes_ctr_encrypt)
+	push		{r4-r6, lr}
+	ldrd		r4, r5, [sp, #16]
+	vld1.8		{q6}, [r5]		@ load ctr
+	prepare_key	r2, r3
+	vmov		r6, s27			@ keep swabbed ctr in r6
+	rev		r6, r6
+	cmn		r6, r4			@ 32 bit overflow?
+	bcs		.Lctrloop
+.Lctrloop3x:
+	subs		r4, r4, #3
+	bmi		.Lctr1x
+	add		r6, r6, #1
+	vmov		q0, q6
+	vmov		q1, q6
+	rev		ip, r6
+	add		r6, r6, #1
+	vmov		q2, q6
+	vmov		s7, ip
+	rev		ip, r6
+	add		r6, r6, #1
+	vmov		s11, ip
+	vld1.8		{q3-q4}, [r1, :64]!
+	vld1.8		{q5}, [r1, :64]!
+	bl		aes_encrypt_3x
+	veor		q0, q0, q3
+	veor		q1, q1, q4
+	veor		q2, q2, q5
+	rev		ip, r6
+	vst1.8		{q0-q1}, [r0, :64]!
+	vst1.8		{q2}, [r0, :64]!
+	vmov		s27, ip
+	b		.Lctrloop3x
+.Lctr1x:
+	adds		r4, r4, #3
+	beq		.Lctrout
+.Lctrloop:
+	vmov		q0, q6
+	bl		aes_encrypt
+	subs		r4, r4, #1
+	bmi		.Lctrhalfblock		@ blocks < 0 means 1/2 block
+	vld1.8		{q3}, [r1, :64]!
+	veor		q3, q0, q3
+	vst1.8		{q3}, [r0, :64]!
+
+	adds		r6, r6, #1		@ increment BE ctr
+	rev		ip, r6
+	vmov		s27, ip
+	bcs		.Lctrcarry
+	teq		r4, #0
+	bne		.Lctrloop
+.Lctrout:
+	vst1.8		{q6}, [r5]
+	pop		{r4-r6, pc}
+
+.Lctrhalfblock:
+	vld1.8		{d1}, [r1, :64]
+	veor		d0, d0, d1
+	vst1.8		{d0}, [r0, :64]
+	pop		{r4-r6, pc}
+
+.Lctrcarry:
+	.irp		sreg, s26, s25, s24
+	vmov		ip, \sreg		@ load next word of ctr
+	rev		ip, ip			@ ... to handle the carry
+	adds		ip, ip, #1
+	rev		ip, ip
+	vmov		\sreg, ip
+	bcc		0f
+	.endr
+0:	teq		r4, #0
+	beq		.Lctrout
+	b		.Lctrloop
+ENDPROC(ce_aes_ctr_encrypt)
+
+	/*
+	 * aes_xts_encrypt(u8 out[], u8 const in[], u8 const rk1[], int rounds,
+	 *		   int blocks, u8 iv[], u8 const rk2[], int first)
+	 * aes_xts_decrypt(u8 out[], u8 const in[], u8 const rk1[], int rounds,
+	 *		   int blocks, u8 iv[], u8 const rk2[], int first)
+	 */
+
+	.macro		next_tweak, out, in, const, tmp
+	vshr.s64	\tmp, \in, #63
+	vand		\tmp, \tmp, \const
+	vadd.u64	\out, \in, \in
+	vext.8		\tmp, \tmp, \tmp, #8
+	veor		\out, \out, \tmp
+	.endm
+
+	.align		3
+.Lxts_mul_x:
+	.quad		1, 0x87
+
+ce_aes_xts_init:
+	vldr		d14, .Lxts_mul_x
+	vldr		d15, .Lxts_mul_x + 8
+
+	ldrd		r4, r5, [sp, #16]	@ load args
+	ldr		r6, [sp, #28]
+	vld1.8		{q0}, [r5]		@ load iv
+	teq		r6, #1			@ start of a block?
+	bxne		lr
+
+	@ Encrypt the IV in q0 with the second AES key. This should only
+	@ be done at the start of a block.
+	ldr		r6, [sp, #24]		@ load AES key 2
+	prepare_key	r6, r3
+	add		ip, r6, #32		@ 3rd round key of key 2
+	b		.Laes_encrypt_tweak	@ tail call
+ENDPROC(ce_aes_xts_init)
+
+ENTRY(ce_aes_xts_encrypt)
+	push		{r4-r6, lr}
+
+	bl		ce_aes_xts_init		@ run shared prologue
+	prepare_key	r2, r3
+	vmov		q3, q0
+
+	teq		r6, #0			@ start of a block?
+	bne		.Lxtsenc3x
+
+.Lxtsencloop3x:
+	next_tweak	q3, q3, q7, q6
+.Lxtsenc3x:
+	subs		r4, r4, #3
+	bmi		.Lxtsenc1x
+	vld1.8		{q0-q1}, [r1, :64]!	@ get 3 pt blocks
+	vld1.8		{q2}, [r1, :64]!
+	next_tweak	q4, q3, q7, q6
+	veor		q0, q0, q3
+	next_tweak	q5, q4, q7, q6
+	veor		q1, q1, q4
+	veor		q2, q2, q5
+	bl		aes_encrypt_3x
+	veor		q0, q0, q3
+	veor		q1, q1, q4
+	veor		q2, q2, q5
+	vst1.8		{q0-q1}, [r0, :64]!	@ write 3 ct blocks
+	vst1.8		{q2}, [r0, :64]!
+	vmov		q3, q5
+	teq		r4, #0
+	beq		.Lxtsencout
+	b		.Lxtsencloop3x
+.Lxtsenc1x:
+	adds		r4, r4, #3
+	beq		.Lxtsencout
+.Lxtsencloop:
+	vld1.8		{q0}, [r1, :64]!
+	veor		q0, q0, q3
+	bl		aes_encrypt
+	veor		q0, q0, q3
+	vst1.8		{q0}, [r0, :64]!
+	subs		r4, r4, #1
+	beq		.Lxtsencout
+	next_tweak	q3, q3, q7, q6
+	b		.Lxtsencloop
+.Lxtsencout:
+	vst1.8		{q3}, [r5]
+	pop		{r4-r6, pc}
+ENDPROC(ce_aes_xts_encrypt)
+
+
+ENTRY(ce_aes_xts_decrypt)
+	push		{r4-r6, lr}
+
+	bl		ce_aes_xts_init		@ run shared prologue
+	prepare_key	r2, r3
+	vmov		q3, q0
+
+	teq		r6, #0			@ start of a block?
+	bne		.Lxtsdec3x
+
+.Lxtsdecloop3x:
+	next_tweak	q3, q3, q7, q6
+.Lxtsdec3x:
+	subs		r4, r4, #3
+	bmi		.Lxtsdec1x
+	vld1.8		{q0-q1}, [r1, :64]!	@ get 3 ct blocks
+	vld1.8		{q2}, [r1, :64]!
+	next_tweak	q4, q3, q7, q6
+	veor		q0, q0, q3
+	next_tweak	q5, q4, q7, q6
+	veor		q1, q1, q4
+	veor		q2, q2, q5
+	bl		aes_decrypt_3x
+	veor		q0, q0, q3
+	veor		q1, q1, q4
+	veor		q2, q2, q5
+	vst1.8		{q0-q1}, [r0, :64]!	@ write 3 pt blocks
+	vst1.8		{q2}, [r0, :64]!
+	vmov		q3, q5
+	teq		r4, #0
+	beq		.Lxtsdecout
+	b		.Lxtsdecloop3x
+.Lxtsdec1x:
+	adds		r4, r4, #3
+	beq		.Lxtsdecout
+.Lxtsdecloop:
+	vld1.8		{q0}, [r1, :64]!
+	veor		q0, q0, q3
+	add		ip, r2, #32		@ 3rd round key
+	bl		aes_decrypt
+	veor		q0, q0, q3
+	vst1.8		{q0}, [r0, :64]!
+	subs		r4, r4, #1
+	beq		.Lxtsdecout
+	next_tweak	q3, q3, q7, q6
+	b		.Lxtsdecloop
+.Lxtsdecout:
+	vst1.8		{q3}, [r5]
+	pop		{r4-r6, pc}
+ENDPROC(ce_aes_xts_decrypt)
+
+	/*
+	 * u32 ce_aes_sub(u32 input) - use the aese instruction to perform the
+	 *                             AES sbox substitution on each byte in
+	 *                             'input'
+	 */
+ENTRY(ce_aes_sub)
+	vdup.32		q1, r0
+	veor		q0, q0, q0
+	aese.8		q0, q1
+	vmov		r0, s0
+	bx		lr
+ENDPROC(ce_aes_sub)
+
+	/*
+	 * void ce_aes_invert(u8 *dst, u8 *src) - perform the Inverse MixColumns
+	 *                                        operation on round key *src
+	 */
+ENTRY(ce_aes_invert)
+	vld1.8		{q0}, [r1]
+	aesimc.8	q0, q0
+	vst1.8		{q0}, [r0]
+	bx		lr
+ENDPROC(ce_aes_invert)
diff --git a/arch/arm/crypto/aes-ce-glue.c b/arch/arm/crypto/aes-ce-glue.c
new file mode 100644
index 000000000..b445a5d56
--- /dev/null
+++ b/arch/arm/crypto/aes-ce-glue.c
@@ -0,0 +1,524 @@
+/*
+ * aes-ce-glue.c - wrapper code for ARMv8 AES
+ *
+ * Copyright (C) 2015 Linaro Ltd <ard.biesheuvel@linaro.org>
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2 as
+ * published by the Free Software Foundation.
+ */
+
+#include <asm/hwcap.h>
+#include <asm/neon.h>
+#include <asm/hwcap.h>
+#include <crypto/aes.h>
+#include <crypto/ablk_helper.h>
+#include <crypto/algapi.h>
+#include <linux/module.h>
+
+MODULE_DESCRIPTION("AES-ECB/CBC/CTR/XTS using ARMv8 Crypto Extensions");
+MODULE_AUTHOR("Ard Biesheuvel <ard.biesheuvel@linaro.org>");
+MODULE_LICENSE("GPL v2");
+
+/* defined in aes-ce-core.S */
+asmlinkage u32 ce_aes_sub(u32 input);
+asmlinkage void ce_aes_invert(void *dst, void *src);
+
+asmlinkage void ce_aes_ecb_encrypt(u8 out[], u8 const in[], u8 const rk[],
+				   int rounds, int blocks);
+asmlinkage void ce_aes_ecb_decrypt(u8 out[], u8 const in[], u8 const rk[],
+				   int rounds, int blocks);
+
+asmlinkage void ce_aes_cbc_encrypt(u8 out[], u8 const in[], u8 const rk[],
+				   int rounds, int blocks, u8 iv[]);
+asmlinkage void ce_aes_cbc_decrypt(u8 out[], u8 const in[], u8 const rk[],
+				   int rounds, int blocks, u8 iv[]);
+
+asmlinkage void ce_aes_ctr_encrypt(u8 out[], u8 const in[], u8 const rk[],
+				   int rounds, int blocks, u8 ctr[]);
+
+asmlinkage void ce_aes_xts_encrypt(u8 out[], u8 const in[], u8 const rk1[],
+				   int rounds, int blocks, u8 iv[],
+				   u8 const rk2[], int first);
+asmlinkage void ce_aes_xts_decrypt(u8 out[], u8 const in[], u8 const rk1[],
+				   int rounds, int blocks, u8 iv[],
+				   u8 const rk2[], int first);
+
+struct aes_block {
+	u8 b[AES_BLOCK_SIZE];
+};
+
+static int num_rounds(struct crypto_aes_ctx *ctx)
+{
+	/*
+	 * # of rounds specified by AES:
+	 * 128 bit key		10 rounds
+	 * 192 bit key		12 rounds
+	 * 256 bit key		14 rounds
+	 * => n byte key	=> 6 + (n/4) rounds
+	 */
+	return 6 + ctx->key_length / 4;
+}
+
+static int ce_aes_expandkey(struct crypto_aes_ctx *ctx, const u8 *in_key,
+			    unsigned int key_len)
+{
+	/*
+	 * The AES key schedule round constants
+	 */
+	static u8 const rcon[] = {
+		0x01, 0x02, 0x04, 0x08, 0x10, 0x20, 0x40, 0x80, 0x1b, 0x36,
+	};
+
+	u32 kwords = key_len / sizeof(u32);
+	struct aes_block *key_enc, *key_dec;
+	int i, j;
+
+	if (key_len != AES_KEYSIZE_128 &&
+	    key_len != AES_KEYSIZE_192 &&
+	    key_len != AES_KEYSIZE_256)
+		return -EINVAL;
+
+	memcpy(ctx->key_enc, in_key, key_len);
+	ctx->key_length = key_len;
+
+	kernel_neon_begin();
+	for (i = 0; i < sizeof(rcon); i++) {
+		u32 *rki = ctx->key_enc + (i * kwords);
+		u32 *rko = rki + kwords;
+
+		rko[0] = ror32(ce_aes_sub(rki[kwords - 1]), 8);
+		rko[0] = rko[0] ^ rki[0] ^ rcon[i];
+		rko[1] = rko[0] ^ rki[1];
+		rko[2] = rko[1] ^ rki[2];
+		rko[3] = rko[2] ^ rki[3];
+
+		if (key_len == AES_KEYSIZE_192) {
+			if (i >= 7)
+				break;
+			rko[4] = rko[3] ^ rki[4];
+			rko[5] = rko[4] ^ rki[5];
+		} else if (key_len == AES_KEYSIZE_256) {
+			if (i >= 6)
+				break;
+			rko[4] = ce_aes_sub(rko[3]) ^ rki[4];
+			rko[5] = rko[4] ^ rki[5];
+			rko[6] = rko[5] ^ rki[6];
+			rko[7] = rko[6] ^ rki[7];
+		}
+	}
+
+	/*
+	 * Generate the decryption keys for the Equivalent Inverse Cipher.
+	 * This involves reversing the order of the round keys, and applying
+	 * the Inverse Mix Columns transformation on all but the first and
+	 * the last one.
+	 */
+	key_enc = (struct aes_block *)ctx->key_enc;
+	key_dec = (struct aes_block *)ctx->key_dec;
+	j = num_rounds(ctx);
+
+	key_dec[0] = key_enc[j];
+	for (i = 1, j--; j > 0; i++, j--)
+		ce_aes_invert(key_dec + i, key_enc + j);
+	key_dec[i] = key_enc[0];
+
+	kernel_neon_end();
+	return 0;
+}
+
+static int ce_aes_setkey(struct crypto_tfm *tfm, const u8 *in_key,
+			 unsigned int key_len)
+{
+	struct crypto_aes_ctx *ctx = crypto_tfm_ctx(tfm);
+	int ret;
+
+	ret = ce_aes_expandkey(ctx, in_key, key_len);
+	if (!ret)
+		return 0;
+
+	tfm->crt_flags |= CRYPTO_TFM_RES_BAD_KEY_LEN;
+	return -EINVAL;
+}
+
+struct crypto_aes_xts_ctx {
+	struct crypto_aes_ctx key1;
+	struct crypto_aes_ctx __aligned(8) key2;
+};
+
+static int xts_set_key(struct crypto_tfm *tfm, const u8 *in_key,
+		       unsigned int key_len)
+{
+	struct crypto_aes_xts_ctx *ctx = crypto_tfm_ctx(tfm);
+	int ret;
+
+	ret = ce_aes_expandkey(&ctx->key1, in_key, key_len / 2);
+	if (!ret)
+		ret = ce_aes_expandkey(&ctx->key2, &in_key[key_len / 2],
+				       key_len / 2);
+	if (!ret)
+		return 0;
+
+	tfm->crt_flags |= CRYPTO_TFM_RES_BAD_KEY_LEN;
+	return -EINVAL;
+}
+
+static int ecb_encrypt(struct blkcipher_desc *desc, struct scatterlist *dst,
+		       struct scatterlist *src, unsigned int nbytes)
+{
+	struct crypto_aes_ctx *ctx = crypto_blkcipher_ctx(desc->tfm);
+	struct blkcipher_walk walk;
+	unsigned int blocks;
+	int err;
+
+	desc->flags &= ~CRYPTO_TFM_REQ_MAY_SLEEP;
+	blkcipher_walk_init(&walk, dst, src, nbytes);
+	err = blkcipher_walk_virt(desc, &walk);
+
+	kernel_neon_begin();
+	while ((blocks = (walk.nbytes / AES_BLOCK_SIZE))) {
+		ce_aes_ecb_encrypt(walk.dst.virt.addr, walk.src.virt.addr,
+				   (u8 *)ctx->key_enc, num_rounds(ctx), blocks);
+		err = blkcipher_walk_done(desc, &walk,
+					  walk.nbytes % AES_BLOCK_SIZE);
+	}
+	kernel_neon_end();
+	return err;
+}
+
+static int ecb_decrypt(struct blkcipher_desc *desc, struct scatterlist *dst,
+		       struct scatterlist *src, unsigned int nbytes)
+{
+	struct crypto_aes_ctx *ctx = crypto_blkcipher_ctx(desc->tfm);
+	struct blkcipher_walk walk;
+	unsigned int blocks;
+	int err;
+
+	desc->flags &= ~CRYPTO_TFM_REQ_MAY_SLEEP;
+	blkcipher_walk_init(&walk, dst, src, nbytes);
+	err = blkcipher_walk_virt(desc, &walk);
+
+	kernel_neon_begin();
+	while ((blocks = (walk.nbytes / AES_BLOCK_SIZE))) {
+		ce_aes_ecb_decrypt(walk.dst.virt.addr, walk.src.virt.addr,
+				   (u8 *)ctx->key_dec, num_rounds(ctx), blocks);
+		err = blkcipher_walk_done(desc, &walk,
+					  walk.nbytes % AES_BLOCK_SIZE);
+	}
+	kernel_neon_end();
+	return err;
+}
+
+static int cbc_encrypt(struct blkcipher_desc *desc, struct scatterlist *dst,
+		       struct scatterlist *src, unsigned int nbytes)
+{
+	struct crypto_aes_ctx *ctx = crypto_blkcipher_ctx(desc->tfm);
+	struct blkcipher_walk walk;
+	unsigned int blocks;
+	int err;
+
+	desc->flags &= ~CRYPTO_TFM_REQ_MAY_SLEEP;
+	blkcipher_walk_init(&walk, dst, src, nbytes);
+	err = blkcipher_walk_virt(desc, &walk);
+
+	kernel_neon_begin();
+	while ((blocks = (walk.nbytes / AES_BLOCK_SIZE))) {
+		ce_aes_cbc_encrypt(walk.dst.virt.addr, walk.src.virt.addr,
+				   (u8 *)ctx->key_enc, num_rounds(ctx), blocks,
+				   walk.iv);
+		err = blkcipher_walk_done(desc, &walk,
+					  walk.nbytes % AES_BLOCK_SIZE);
+	}
+	kernel_neon_end();
+	return err;
+}
+
+static int cbc_decrypt(struct blkcipher_desc *desc, struct scatterlist *dst,
+		       struct scatterlist *src, unsigned int nbytes)
+{
+	struct crypto_aes_ctx *ctx = crypto_blkcipher_ctx(desc->tfm);
+	struct blkcipher_walk walk;
+	unsigned int blocks;
+	int err;
+
+	desc->flags &= ~CRYPTO_TFM_REQ_MAY_SLEEP;
+	blkcipher_walk_init(&walk, dst, src, nbytes);
+	err = blkcipher_walk_virt(desc, &walk);
+
+	kernel_neon_begin();
+	while ((blocks = (walk.nbytes / AES_BLOCK_SIZE))) {
+		ce_aes_cbc_decrypt(walk.dst.virt.addr, walk.src.virt.addr,
+				   (u8 *)ctx->key_dec, num_rounds(ctx), blocks,
+				   walk.iv);
+		err = blkcipher_walk_done(desc, &walk,
+					  walk.nbytes % AES_BLOCK_SIZE);
+	}
+	kernel_neon_end();
+	return err;
+}
+
+static int ctr_encrypt(struct blkcipher_desc *desc, struct scatterlist *dst,
+		       struct scatterlist *src, unsigned int nbytes)
+{
+	struct crypto_aes_ctx *ctx = crypto_blkcipher_ctx(desc->tfm);
+	struct blkcipher_walk walk;
+	int err, blocks;
+
+	desc->flags &= ~CRYPTO_TFM_REQ_MAY_SLEEP;
+	blkcipher_walk_init(&walk, dst, src, nbytes);
+	err = blkcipher_walk_virt_block(desc, &walk, AES_BLOCK_SIZE);
+
+	kernel_neon_begin();
+	while ((blocks = (walk.nbytes / AES_BLOCK_SIZE))) {
+		ce_aes_ctr_encrypt(walk.dst.virt.addr, walk.src.virt.addr,
+				   (u8 *)ctx->key_enc, num_rounds(ctx), blocks,
+				   walk.iv);
+		nbytes -= blocks * AES_BLOCK_SIZE;
+		if (nbytes && nbytes == walk.nbytes % AES_BLOCK_SIZE)
+			break;
+		err = blkcipher_walk_done(desc, &walk,
+					  walk.nbytes % AES_BLOCK_SIZE);
+	}
+	if (nbytes) {
+		u8 *tdst = walk.dst.virt.addr + blocks * AES_BLOCK_SIZE;
+		u8 *tsrc = walk.src.virt.addr + blocks * AES_BLOCK_SIZE;
+		u8 __aligned(8) tail[AES_BLOCK_SIZE];
+
+		/*
+		 * Minimum alignment is 8 bytes, so if nbytes is <= 8, we need
+		 * to tell aes_ctr_encrypt() to only read half a block.
+		 */
+		blocks = (nbytes <= 8) ? -1 : 1;
+
+		ce_aes_ctr_encrypt(tail, tsrc, (u8 *)ctx->key_enc,
+				   num_rounds(ctx), blocks, walk.iv);
+		memcpy(tdst, tail, nbytes);
+		err = blkcipher_walk_done(desc, &walk, 0);
+	}
+	kernel_neon_end();
+
+	return err;
+}
+
+static int xts_encrypt(struct blkcipher_desc *desc, struct scatterlist *dst,
+		       struct scatterlist *src, unsigned int nbytes)
+{
+	struct crypto_aes_xts_ctx *ctx = crypto_blkcipher_ctx(desc->tfm);
+	int err, first, rounds = num_rounds(&ctx->key1);
+	struct blkcipher_walk walk;
+	unsigned int blocks;
+
+	desc->flags &= ~CRYPTO_TFM_REQ_MAY_SLEEP;
+	blkcipher_walk_init(&walk, dst, src, nbytes);
+	err = blkcipher_walk_virt(desc, &walk);
+
+	kernel_neon_begin();
+	for (first = 1; (blocks = (walk.nbytes / AES_BLOCK_SIZE)); first = 0) {
+		ce_aes_xts_encrypt(walk.dst.virt.addr, walk.src.virt.addr,
+				   (u8 *)ctx->key1.key_enc, rounds, blocks,
+				   walk.iv, (u8 *)ctx->key2.key_enc, first);
+		err = blkcipher_walk_done(desc, &walk,
+					  walk.nbytes % AES_BLOCK_SIZE);
+	}
+	kernel_neon_end();
+
+	return err;
+}
+
+static int xts_decrypt(struct blkcipher_desc *desc, struct scatterlist *dst,
+		       struct scatterlist *src, unsigned int nbytes)
+{
+	struct crypto_aes_xts_ctx *ctx = crypto_blkcipher_ctx(desc->tfm);
+	int err, first, rounds = num_rounds(&ctx->key1);
+	struct blkcipher_walk walk;
+	unsigned int blocks;
+
+	desc->flags &= ~CRYPTO_TFM_REQ_MAY_SLEEP;
+	blkcipher_walk_init(&walk, dst, src, nbytes);
+	err = blkcipher_walk_virt(desc, &walk);
+
+	kernel_neon_begin();
+	for (first = 1; (blocks = (walk.nbytes / AES_BLOCK_SIZE)); first = 0) {
+		ce_aes_xts_decrypt(walk.dst.virt.addr, walk.src.virt.addr,
+				   (u8 *)ctx->key1.key_dec, rounds, blocks,
+				   walk.iv, (u8 *)ctx->key2.key_enc, first);
+		err = blkcipher_walk_done(desc, &walk,
+					  walk.nbytes % AES_BLOCK_SIZE);
+	}
+	kernel_neon_end();
+
+	return err;
+}
+
+static struct crypto_alg aes_algs[] = { {
+	.cra_name		= "__ecb-aes-ce",
+	.cra_driver_name	= "__driver-ecb-aes-ce",
+	.cra_priority		= 0,
+	.cra_flags		= CRYPTO_ALG_TYPE_BLKCIPHER |
+				  CRYPTO_ALG_INTERNAL,
+	.cra_blocksize		= AES_BLOCK_SIZE,
+	.cra_ctxsize		= sizeof(struct crypto_aes_ctx),
+	.cra_alignmask		= 7,
+	.cra_type		= &crypto_blkcipher_type,
+	.cra_module		= THIS_MODULE,
+	.cra_blkcipher = {
+		.min_keysize	= AES_MIN_KEY_SIZE,
+		.max_keysize	= AES_MAX_KEY_SIZE,
+		.ivsize		= AES_BLOCK_SIZE,
+		.setkey		= ce_aes_setkey,
+		.encrypt	= ecb_encrypt,
+		.decrypt	= ecb_decrypt,
+	},
+}, {
+	.cra_name		= "__cbc-aes-ce",
+	.cra_driver_name	= "__driver-cbc-aes-ce",
+	.cra_priority		= 0,
+	.cra_flags		= CRYPTO_ALG_TYPE_BLKCIPHER |
+				  CRYPTO_ALG_INTERNAL,
+	.cra_blocksize		= AES_BLOCK_SIZE,
+	.cra_ctxsize		= sizeof(struct crypto_aes_ctx),
+	.cra_alignmask		= 7,
+	.cra_type		= &crypto_blkcipher_type,
+	.cra_module		= THIS_MODULE,
+	.cra_blkcipher = {
+		.min_keysize	= AES_MIN_KEY_SIZE,
+		.max_keysize	= AES_MAX_KEY_SIZE,
+		.ivsize		= AES_BLOCK_SIZE,
+		.setkey		= ce_aes_setkey,
+		.encrypt	= cbc_encrypt,
+		.decrypt	= cbc_decrypt,
+	},
+}, {
+	.cra_name		= "__ctr-aes-ce",
+	.cra_driver_name	= "__driver-ctr-aes-ce",
+	.cra_priority		= 0,
+	.cra_flags		= CRYPTO_ALG_TYPE_BLKCIPHER |
+				  CRYPTO_ALG_INTERNAL,
+	.cra_blocksize		= 1,
+	.cra_ctxsize		= sizeof(struct crypto_aes_ctx),
+	.cra_alignmask		= 7,
+	.cra_type		= &crypto_blkcipher_type,
+	.cra_module		= THIS_MODULE,
+	.cra_blkcipher = {
+		.min_keysize	= AES_MIN_KEY_SIZE,
+		.max_keysize	= AES_MAX_KEY_SIZE,
+		.ivsize		= AES_BLOCK_SIZE,
+		.setkey		= ce_aes_setkey,
+		.encrypt	= ctr_encrypt,
+		.decrypt	= ctr_encrypt,
+	},
+}, {
+	.cra_name		= "__xts-aes-ce",
+	.cra_driver_name	= "__driver-xts-aes-ce",
+	.cra_priority		= 0,
+	.cra_flags		= CRYPTO_ALG_TYPE_BLKCIPHER |
+				  CRYPTO_ALG_INTERNAL,
+	.cra_blocksize		= AES_BLOCK_SIZE,
+	.cra_ctxsize		= sizeof(struct crypto_aes_xts_ctx),
+	.cra_alignmask		= 7,
+	.cra_type		= &crypto_blkcipher_type,
+	.cra_module		= THIS_MODULE,
+	.cra_blkcipher = {
+		.min_keysize	= 2 * AES_MIN_KEY_SIZE,
+		.max_keysize	= 2 * AES_MAX_KEY_SIZE,
+		.ivsize		= AES_BLOCK_SIZE,
+		.setkey		= xts_set_key,
+		.encrypt	= xts_encrypt,
+		.decrypt	= xts_decrypt,
+	},
+}, {
+	.cra_name		= "ecb(aes)",
+	.cra_driver_name	= "ecb-aes-ce",
+	.cra_priority		= 300,
+	.cra_flags		= CRYPTO_ALG_TYPE_ABLKCIPHER|CRYPTO_ALG_ASYNC,
+	.cra_blocksize		= AES_BLOCK_SIZE,
+	.cra_ctxsize		= sizeof(struct async_helper_ctx),
+	.cra_alignmask		= 7,
+	.cra_type		= &crypto_ablkcipher_type,
+	.cra_module		= THIS_MODULE,
+	.cra_init		= ablk_init,
+	.cra_exit		= ablk_exit,
+	.cra_ablkcipher = {
+		.min_keysize	= AES_MIN_KEY_SIZE,
+		.max_keysize	= AES_MAX_KEY_SIZE,
+		.ivsize		= AES_BLOCK_SIZE,
+		.setkey		= ablk_set_key,
+		.encrypt	= ablk_encrypt,
+		.decrypt	= ablk_decrypt,
+	}
+}, {
+	.cra_name		= "cbc(aes)",
+	.cra_driver_name	= "cbc-aes-ce",
+	.cra_priority		= 300,
+	.cra_flags		= CRYPTO_ALG_TYPE_ABLKCIPHER|CRYPTO_ALG_ASYNC,
+	.cra_blocksize		= AES_BLOCK_SIZE,
+	.cra_ctxsize		= sizeof(struct async_helper_ctx),
+	.cra_alignmask		= 7,
+	.cra_type		= &crypto_ablkcipher_type,
+	.cra_module		= THIS_MODULE,
+	.cra_init		= ablk_init,
+	.cra_exit		= ablk_exit,
+	.cra_ablkcipher = {
+		.min_keysize	= AES_MIN_KEY_SIZE,
+		.max_keysize	= AES_MAX_KEY_SIZE,
+		.ivsize		= AES_BLOCK_SIZE,
+		.setkey		= ablk_set_key,
+		.encrypt	= ablk_encrypt,
+		.decrypt	= ablk_decrypt,
+	}
+}, {
+	.cra_name		= "ctr(aes)",
+	.cra_driver_name	= "ctr-aes-ce",
+	.cra_priority		= 300,
+	.cra_flags		= CRYPTO_ALG_TYPE_ABLKCIPHER|CRYPTO_ALG_ASYNC,
+	.cra_blocksize		= 1,
+	.cra_ctxsize		= sizeof(struct async_helper_ctx),
+	.cra_alignmask		= 7,
+	.cra_type		= &crypto_ablkcipher_type,
+	.cra_module		= THIS_MODULE,
+	.cra_init		= ablk_init,
+	.cra_exit		= ablk_exit,
+	.cra_ablkcipher = {
+		.min_keysize	= AES_MIN_KEY_SIZE,
+		.max_keysize	= AES_MAX_KEY_SIZE,
+		.ivsize		= AES_BLOCK_SIZE,
+		.setkey		= ablk_set_key,
+		.encrypt	= ablk_encrypt,
+		.decrypt	= ablk_decrypt,
+	}
+}, {
+	.cra_name		= "xts(aes)",
+	.cra_driver_name	= "xts-aes-ce",
+	.cra_priority		= 300,
+	.cra_flags		= CRYPTO_ALG_TYPE_ABLKCIPHER|CRYPTO_ALG_ASYNC,
+	.cra_blocksize		= AES_BLOCK_SIZE,
+	.cra_ctxsize		= sizeof(struct async_helper_ctx),
+	.cra_alignmask		= 7,
+	.cra_type		= &crypto_ablkcipher_type,
+	.cra_module		= THIS_MODULE,
+	.cra_init		= ablk_init,
+	.cra_exit		= ablk_exit,
+	.cra_ablkcipher = {
+		.min_keysize	= 2 * AES_MIN_KEY_SIZE,
+		.max_keysize	= 2 * AES_MAX_KEY_SIZE,
+		.ivsize		= AES_BLOCK_SIZE,
+		.setkey		= ablk_set_key,
+		.encrypt	= ablk_encrypt,
+		.decrypt	= ablk_decrypt,
+	}
+} };
+
+static int __init aes_init(void)
+{
+	if (!(elf_hwcap2 & HWCAP2_AES))
+		return -ENODEV;
+	return crypto_register_algs(aes_algs, ARRAY_SIZE(aes_algs));
+}
+
+static void __exit aes_exit(void)
+{
+	crypto_unregister_algs(aes_algs, ARRAY_SIZE(aes_algs));
+}
+
+module_init(aes_init);
+module_exit(aes_exit);
diff --git a/arch/arm/crypto/aes_glue.c b/arch/arm/crypto/aes_glue.c
new file mode 100644
index 000000000..0409b8f89
--- /dev/null
+++ b/arch/arm/crypto/aes_glue.c
@@ -0,0 +1,98 @@
+/*
+ * Glue Code for the asm optimized version of the AES Cipher Algorithm
+ */
+
+#include <linux/module.h>
+#include <linux/crypto.h>
+#include <crypto/aes.h>
+
+#include "aes_glue.h"
+
+EXPORT_SYMBOL(AES_encrypt);
+EXPORT_SYMBOL(AES_decrypt);
+EXPORT_SYMBOL(private_AES_set_encrypt_key);
+EXPORT_SYMBOL(private_AES_set_decrypt_key);
+
+static void aes_encrypt(struct crypto_tfm *tfm, u8 *dst, const u8 *src)
+{
+	struct AES_CTX *ctx = crypto_tfm_ctx(tfm);
+	AES_encrypt(src, dst, &ctx->enc_key);
+}
+
+static void aes_decrypt(struct crypto_tfm *tfm, u8 *dst, const u8 *src)
+{
+	struct AES_CTX *ctx = crypto_tfm_ctx(tfm);
+	AES_decrypt(src, dst, &ctx->dec_key);
+}
+
+static int aes_set_key(struct crypto_tfm *tfm, const u8 *in_key,
+		unsigned int key_len)
+{
+	struct AES_CTX *ctx = crypto_tfm_ctx(tfm);
+
+	switch (key_len) {
+	case AES_KEYSIZE_128:
+		key_len = 128;
+		break;
+	case AES_KEYSIZE_192:
+		key_len = 192;
+		break;
+	case AES_KEYSIZE_256:
+		key_len = 256;
+		break;
+	default:
+		tfm->crt_flags |= CRYPTO_TFM_RES_BAD_KEY_LEN;
+		return -EINVAL;
+	}
+
+	if (private_AES_set_encrypt_key(in_key, key_len, &ctx->enc_key) == -1) {
+		tfm->crt_flags |= CRYPTO_TFM_RES_BAD_KEY_LEN;
+		return -EINVAL;
+	}
+	/* private_AES_set_decrypt_key expects an encryption key as input */
+	ctx->dec_key = ctx->enc_key;
+	if (private_AES_set_decrypt_key(in_key, key_len, &ctx->dec_key) == -1) {
+		tfm->crt_flags |= CRYPTO_TFM_RES_BAD_KEY_LEN;
+		return -EINVAL;
+	}
+	return 0;
+}
+
+static struct crypto_alg aes_alg = {
+	.cra_name		= "aes",
+	.cra_driver_name	= "aes-asm",
+	.cra_priority		= 200,
+	.cra_flags		= CRYPTO_ALG_TYPE_CIPHER,
+	.cra_blocksize		= AES_BLOCK_SIZE,
+	.cra_ctxsize		= sizeof(struct AES_CTX),
+	.cra_module		= THIS_MODULE,
+	.cra_list		= LIST_HEAD_INIT(aes_alg.cra_list),
+	.cra_u	= {
+		.cipher	= {
+			.cia_min_keysize	= AES_MIN_KEY_SIZE,
+			.cia_max_keysize	= AES_MAX_KEY_SIZE,
+			.cia_setkey		= aes_set_key,
+			.cia_encrypt		= aes_encrypt,
+			.cia_decrypt		= aes_decrypt
+		}
+	}
+};
+
+static int __init aes_init(void)
+{
+	return crypto_register_alg(&aes_alg);
+}
+
+static void __exit aes_fini(void)
+{
+	crypto_unregister_alg(&aes_alg);
+}
+
+module_init(aes_init);
+module_exit(aes_fini);
+
+MODULE_DESCRIPTION("Rijndael (AES) Cipher Algorithm (ASM)");
+MODULE_LICENSE("GPL");
+MODULE_ALIAS_CRYPTO("aes");
+MODULE_ALIAS_CRYPTO("aes-asm");
+MODULE_AUTHOR("David McCullough <ucdevel@gmail.com>");
diff --git a/arch/arm/crypto/aes_glue.h b/arch/arm/crypto/aes_glue.h
new file mode 100644
index 000000000..cca3e51eb
--- /dev/null
+++ b/arch/arm/crypto/aes_glue.h
@@ -0,0 +1,19 @@
+
+#define AES_MAXNR 14
+
+struct AES_KEY {
+	unsigned int rd_key[4 * (AES_MAXNR + 1)];
+	int rounds;
+};
+
+struct AES_CTX {
+	struct AES_KEY enc_key;
+	struct AES_KEY dec_key;
+};
+
+asmlinkage void AES_encrypt(const u8 *in, u8 *out, struct AES_KEY *ctx);
+asmlinkage void AES_decrypt(const u8 *in, u8 *out, struct AES_KEY *ctx);
+asmlinkage int private_AES_set_decrypt_key(const unsigned char *userKey,
+					   const int bits, struct AES_KEY *key);
+asmlinkage int private_AES_set_encrypt_key(const unsigned char *userKey,
+					   const int bits, struct AES_KEY *key);
diff --git a/arch/arm/crypto/aesbs-core.S_shipped b/arch/arm/crypto/aesbs-core.S_shipped
new file mode 100644
index 000000000..1d1800f71
--- /dev/null
+++ b/arch/arm/crypto/aesbs-core.S_shipped
@@ -0,0 +1,2548 @@
+
+@ ====================================================================
+@ 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>
+
+#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__
+# define __ARM_MAX_ARCH__ 7
+#endif
+
+#ifdef __thumb__
+# define adrl adr
+#endif
+
+#if __ARM_MAX_ARCH__>=7
+.arch	armv7-a
+.fpu	neon
+
+.text
+.syntax	unified 	@ ARMv7-capable assembler is expected to handle this
+#ifdef __thumb2__
+.thumb
+#else
+.code   32
+#endif
+
+.type	_bsaes_decrypt8,%function
+.align	4
+_bsaes_decrypt8:
+	adr	r6,_bsaes_decrypt8
+	vldmia	r4!, {q9}		@ round 0 key
+	add	r6,r6,#.LM0ISR-_bsaes_decrypt8
+
+	vldmia	r6!, {q8}		@ .LM0ISR
+	veor	q10, q0, q9	@ xor with round0 key
+	veor	q11, q1, q9
+	 vtbl.8	d0, {q10}, d16
+	 vtbl.8	d1, {q10}, d17
+	veor	q12, q2, q9
+	 vtbl.8	d2, {q11}, d16
+	 vtbl.8	d3, {q11}, d17
+	veor	q13, q3, q9
+	 vtbl.8	d4, {q12}, d16
+	 vtbl.8	d5, {q12}, d17
+	veor	q14, q4, q9
+	 vtbl.8	d6, {q13}, d16
+	 vtbl.8	d7, {q13}, d17
+	veor	q15, q5, q9
+	 vtbl.8	d8, {q14}, d16
+	 vtbl.8	d9, {q14}, d17
+	veor	q10, q6, q9
+	 vtbl.8	d10, {q15}, d16
+	 vtbl.8	d11, {q15}, d17
+	veor	q11, q7, q9
+	 vtbl.8	d12, {q10}, d16
+	 vtbl.8	d13, {q10}, d17
+	 vtbl.8	d14, {q11}, d16
+	 vtbl.8	d15, {q11}, d17
+	vmov.i8	q8,#0x55			@ compose .LBS0
+	vmov.i8	q9,#0x33			@ compose .LBS1
+	vshr.u64	q10, q6, #1
+	 vshr.u64	q11, q4, #1
+	veor		q10, q10, q7
+	 veor		q11, q11, q5
+	vand		q10, q10, q8
+	 vand		q11, q11, q8
+	veor		q7, q7, q10
+	vshl.u64	q10, q10, #1
+	 veor		q5, q5, q11
+	 vshl.u64	q11, q11, #1
+	veor		q6, q6, q10
+	 veor		q4, q4, q11
+	vshr.u64	q10, q2, #1
+	 vshr.u64	q11, q0, #1
+	veor		q10, q10, q3
+	 veor		q11, q11, q1
+	vand		q10, q10, q8
+	 vand		q11, q11, q8
+	veor		q3, q3, q10
+	vshl.u64	q10, q10, #1
+	 veor		q1, q1, q11
+	 vshl.u64	q11, q11, #1
+	veor		q2, q2, q10
+	 veor		q0, q0, q11
+	vmov.i8	q8,#0x0f			@ compose .LBS2
+	vshr.u64	q10, q5, #2
+	 vshr.u64	q11, q4, #2
+	veor		q10, q10, q7
+	 veor		q11, q11, q6
+	vand		q10, q10, q9
+	 vand		q11, q11, q9
+	veor		q7, q7, q10
+	vshl.u64	q10, q10, #2
+	 veor		q6, q6, q11
+	 vshl.u64	q11, q11, #2
+	veor		q5, q5, q10
+	 veor		q4, q4, q11
+	vshr.u64	q10, q1, #2
+	 vshr.u64	q11, q0, #2
+	veor		q10, q10, q3
+	 veor		q11, q11, q2
+	vand		q10, q10, q9
+	 vand		q11, q11, q9
+	veor		q3, q3, q10
+	vshl.u64	q10, q10, #2
+	 veor		q2, q2, q11
+	 vshl.u64	q11, q11, #2
+	veor		q1, q1, q10
+	 veor		q0, q0, q11
+	vshr.u64	q10, q3, #4
+	 vshr.u64	q11, q2, #4
+	veor		q10, q10, q7
+	 veor		q11, q11, q6
+	vand		q10, q10, q8
+	 vand		q11, q11, q8
+	veor		q7, q7, q10
+	vshl.u64	q10, q10, #4
+	 veor		q6, q6, q11
+	 vshl.u64	q11, q11, #4
+	veor		q3, q3, q10
+	 veor		q2, q2, q11
+	vshr.u64	q10, q1, #4
+	 vshr.u64	q11, q0, #4
+	veor		q10, q10, q5
+	 veor		q11, q11, q4
+	vand		q10, q10, q8
+	 vand		q11, q11, q8
+	veor		q5, q5, q10
+	vshl.u64	q10, q10, #4
+	 veor		q4, q4, q11
+	 vshl.u64	q11, q11, #4
+	veor		q1, q1, q10
+	 veor		q0, q0, q11
+	sub	r5,r5,#1
+	b	.Ldec_sbox
+.align	4
+.Ldec_loop:
+	vldmia	r4!, {q8-q11}
+	veor	q8, q8, q0
+	veor	q9, q9, q1
+	vtbl.8	d0, {q8}, d24
+	vtbl.8	d1, {q8}, d25
+	vldmia	r4!, {q8}
+	veor	q10, q10, q2
+	vtbl.8	d2, {q9}, d24
+	vtbl.8	d3, {q9}, d25
+	vldmia	r4!, {q9}
+	veor	q11, q11, q3
+	vtbl.8	d4, {q10}, d24
+	vtbl.8	d5, {q10}, d25
+	vldmia	r4!, {q10}
+	vtbl.8	d6, {q11}, d24
+	vtbl.8	d7, {q11}, d25
+	vldmia	r4!, {q11}
+	veor	q8, q8, q4
+	veor	q9, q9, q5
+	vtbl.8	d8, {q8}, d24
+	vtbl.8	d9, {q8}, d25
+	veor	q10, q10, q6
+	vtbl.8	d10, {q9}, d24
+	vtbl.8	d11, {q9}, d25
+	veor	q11, q11, q7
+	vtbl.8	d12, {q10}, d24
+	vtbl.8	d13, {q10}, d25
+	vtbl.8	d14, {q11}, d24
+	vtbl.8	d15, {q11}, d25
+.Ldec_sbox:
+	 veor	q1, q1, q4
+	veor	q3, q3, q4
+
+	veor	q4, q4, q7
+	 veor	q1, q1, q6
+	veor	q2, q2, q7
+	veor	q6, q6, q4
+
+	veor	q0, q0, q1
+	veor	q2, q2, q5
+	 veor	q7, q7, q6
+	veor	q3, q3, q0
+	veor	q5, q5, q0
+	veor	q1, q1, q3
+	veor	q11, q3, q0
+	veor	q10, q7, q4
+	veor	q9, q1, q6
+	veor	q13, q4, q0
+	 vmov	q8, q10
+	veor	q12, q5, q2
+
+	vorr	q10, q10, q9
+	veor	q15, q11, q8
+	vand	q14, q11, q12
+	vorr	q11, q11, q12
+	veor	q12, q12, q9
+	vand	q8, q8, q9
+	veor	q9, q6, q2
+	vand	q15, q15, q12
+	vand	q13, q13, q9
+	veor	q9, q3, q7
+	veor	q12, q1, q5
+	veor	q11, q11, q13
+	veor	q10, q10, q13
+	vand	q13, q9, q12
+	vorr	q9, q9, q12
+	veor	q11, q11, q15
+	veor	q8, q8, q13
+	veor	q10, q10, q14
+	veor	q9, q9, q15
+	veor	q8, q8, q14
+	vand	q12, q4, q6
+	veor	q9, q9, q14
+	vand	q13, q0, q2
+	vand	q14, q7, q1
+	vorr	q15, q3, q5
+	veor	q11, q11, q12
+	veor	q9, q9, q14
+	veor	q8, q8, q15
+	veor	q10, q10, q13
+
+	@ Inv_GF16 	0, 	1, 	2, 	3, s0, s1, s2, s3
+
+	@ new smaller inversion
+
+	vand	q14, q11, q9
+	vmov	q12, q8
+
+	veor	q13, q10, q14
+	veor	q15, q8, q14
+	veor	q14, q8, q14	@ q14=q15
+
+	vbsl	q13, q9, q8
+	vbsl	q15, q11, q10
+	veor	q11, q11, q10
+
+	vbsl	q12, q13, q14
+	vbsl	q8, q14, q13
+
+	vand	q14, q12, q15
+	veor	q9, q9, q8
+
+	veor	q14, q14, q11
+	veor	q12, q5, q2
+	veor	q8, q1, q6
+	veor 	q10, q15, q14
+	vand	q10, q10, q5
+	veor	q5, q5, q1
+	vand	q11, q1, q15
+	vand	q5, q5, q14
+	veor	q1, q11, q10
+	veor	q5, q5, q11
+	veor	q15, q15, q13
+	veor	q14, q14, q9
+	veor	q11, q15, q14
+	 veor 	q10, q13, q9
+	vand	q11, q11, q12
+	 vand	q10, q10, q2
+	veor	q12, q12, q8
+	 veor	q2, q2, q6
+	vand	q8, q8, q15
+	 vand	q6, q6, q13
+	vand	q12, q12, q14
+	 vand	q2, q2, q9
+	veor	q8, q8, q12
+	 veor	q2, q2, q6
+	veor	q12, q12, q11
+	 veor	q6, q6, q10
+	veor	q5, q5, q12
+	veor	q2, q2, q12
+	veor	q1, q1, q8
+	veor	q6, q6, q8
+
+	veor	q12, q3, q0
+	veor	q8, q7, q4
+	veor	q11, q15, q14
+	 veor 	q10, q13, q9
+	vand	q11, q11, q12
+	 vand	q10, q10, q0
+	veor	q12, q12, q8
+	 veor	q0, q0, q4
+	vand	q8, q8, q15
+	 vand	q4, q4, q13
+	vand	q12, q12, q14
+	 vand	q0, q0, q9
+	veor	q8, q8, q12
+	 veor	q0, q0, q4
+	veor	q12, q12, q11
+	 veor	q4, q4, q10
+	veor	q15, q15, q13
+	veor	q14, q14, q9
+	veor 	q10, q15, q14
+	vand	q10, q10, q3
+	veor	q3, q3, q7
+	vand	q11, q7, q15
+	vand	q3, q3, q14
+	veor	q7, q11, q10
+	veor	q3, q3, q11
+	veor	q3, q3, q12
+	veor	q0, q0, q12
+	veor	q7, q7, q8
+	veor	q4, q4, q8
+	veor	q1, q1, q7
+	veor	q6, q6, q5
+
+	veor	q4, q4, q1
+	veor	q2, q2, q7
+	veor	q5, q5, q7
+	veor	q4, q4, q2
+	 veor 	q7, q7, q0
+	veor	q4, q4, q5
+	 veor	q3, q3, q6
+	 veor	q6, q6, q1
+	veor	q3, q3, q4
+
+	veor	q4, q4, q0
+	veor	q7, q7, q3
+	subs	r5,r5,#1
+	bcc	.Ldec_done
+	@ multiplication by 0x05-0x00-0x04-0x00
+	vext.8	q8, q0, q0, #8
+	vext.8	q14, q3, q3, #8
+	vext.8	q15, q5, q5, #8
+	veor	q8, q8, q0
+	vext.8	q9, q1, q1, #8
+	veor	q14, q14, q3
+	vext.8	q10, q6, q6, #8
+	veor	q15, q15, q5
+	vext.8	q11, q4, q4, #8
+	veor	q9, q9, q1
+	vext.8	q12, q2, q2, #8
+	veor	q10, q10, q6
+	vext.8	q13, q7, q7, #8
+	veor	q11, q11, q4
+	veor	q12, q12, q2
+	veor	q13, q13, q7
+
+	 veor	q0, q0, q14
+	 veor	q1, q1, q14
+	 veor	q6, q6, q8
+	 veor	q2, q2, q10
+	 veor	q4, q4, q9
+	 veor	q1, q1, q15
+	 veor	q6, q6, q15
+	 veor	q2, q2, q14
+	 veor	q7, q7, q11
+	 veor	q4, q4, q14
+	 veor	q3, q3, q12
+	 veor	q2, q2, q15
+	 veor	q7, q7, q15
+	 veor	q5, q5, q13
+	vext.8	q8, q0, q0, #12	@ x0 <<< 32
+	vext.8	q9, q1, q1, #12
+	 veor	q0, q0, q8		@ x0 ^ (x0 <<< 32)
+	vext.8	q10, q6, q6, #12
+	 veor	q1, q1, q9
+	vext.8	q11, q4, q4, #12
+	 veor	q6, q6, q10
+	vext.8	q12, q2, q2, #12
+	 veor	q4, q4, q11
+	vext.8	q13, q7, q7, #12
+	 veor	q2, q2, q12
+	vext.8	q14, q3, q3, #12
+	 veor	q7, q7, q13
+	vext.8	q15, q5, q5, #12
+	 veor	q3, q3, q14
+
+	veor	q9, q9, q0
+	 veor	q5, q5, q15
+	 vext.8	q0, q0, q0, #8		@ (x0 ^ (x0 <<< 32)) <<< 64)
+	veor	q10, q10, q1
+	veor	q8, q8, q5
+	veor	q9, q9, q5
+	 vext.8	q1, q1, q1, #8
+	veor	q13, q13, q2
+	 veor	q0, q0, q8
+	veor	q14, q14, q7
+	 veor	q1, q1, q9
+	 vext.8	q8, q2, q2, #8
+	veor	q12, q12, q4
+	 vext.8	q9, q7, q7, #8
+	veor	q15, q15, q3
+	 vext.8	q2, q4, q4, #8
+	veor	q11, q11, q6
+	 vext.8	q7, q5, q5, #8
+	veor	q12, q12, q5
+	 vext.8	q4, q3, q3, #8
+	veor	q11, q11, q5
+	 vext.8	q3, q6, q6, #8
+	veor	q5, q9, q13
+	veor	q11, q11, q2
+	veor	q7, q7, q15
+	veor	q6, q4, q14
+	veor	q4, q8, q12
+	veor	q2, q3, q10
+	vmov	q3, q11
+	 @ vmov	q5, q9
+	vldmia	r6, {q12}		@ .LISR
+	ite	eq				@ Thumb2 thing, sanity check in ARM
+	addeq	r6,r6,#0x10
+	bne	.Ldec_loop
+	vldmia	r6, {q12}		@ .LISRM0
+	b	.Ldec_loop
+.align	4
+.Ldec_done:
+	vmov.i8	q8,#0x55			@ compose .LBS0
+	vmov.i8	q9,#0x33			@ compose .LBS1
+	vshr.u64	q10, q3, #1
+	 vshr.u64	q11, q2, #1
+	veor		q10, q10, q5
+	 veor		q11, q11, q7
+	vand		q10, q10, q8
+	 vand		q11, q11, q8
+	veor		q5, q5, q10
+	vshl.u64	q10, q10, #1
+	 veor		q7, q7, q11
+	 vshl.u64	q11, q11, #1
+	veor		q3, q3, q10
+	 veor		q2, q2, q11
+	vshr.u64	q10, q6, #1
+	 vshr.u64	q11, q0, #1
+	veor		q10, q10, q4
+	 veor		q11, q11, q1
+	vand		q10, q10, q8
+	 vand		q11, q11, q8
+	veor		q4, q4, q10
+	vshl.u64	q10, q10, #1
+	 veor		q1, q1, q11
+	 vshl.u64	q11, q11, #1
+	veor		q6, q6, q10
+	 veor		q0, q0, q11
+	vmov.i8	q8,#0x0f			@ compose .LBS2
+	vshr.u64	q10, q7, #2
+	 vshr.u64	q11, q2, #2
+	veor		q10, q10, q5
+	 veor		q11, q11, q3
+	vand		q10, q10, q9
+	 vand		q11, q11, q9
+	veor		q5, q5, q10
+	vshl.u64	q10, q10, #2
+	 veor		q3, q3, q11
+	 vshl.u64	q11, q11, #2
+	veor		q7, q7, q10
+	 veor		q2, q2, q11
+	vshr.u64	q10, q1, #2
+	 vshr.u64	q11, q0, #2
+	veor		q10, q10, q4
+	 veor		q11, q11, q6
+	vand		q10, q10, q9
+	 vand		q11, q11, q9
+	veor		q4, q4, q10
+	vshl.u64	q10, q10, #2
+	 veor		q6, q6, q11
+	 vshl.u64	q11, q11, #2
+	veor		q1, q1, q10
+	 veor		q0, q0, q11
+	vshr.u64	q10, q4, #4
+	 vshr.u64	q11, q6, #4
+	veor		q10, q10, q5
+	 veor		q11, q11, q3
+	vand		q10, q10, q8
+	 vand		q11, q11, q8
+	veor		q5, q5, q10
+	vshl.u64	q10, q10, #4
+	 veor		q3, q3, q11
+	 vshl.u64	q11, q11, #4
+	veor		q4, q4, q10
+	 veor		q6, q6, q11
+	vshr.u64	q10, q1, #4
+	 vshr.u64	q11, q0, #4
+	veor		q10, q10, q7
+	 veor		q11, q11, q2
+	vand		q10, q10, q8
+	 vand		q11, q11, q8
+	veor		q7, q7, q10
+	vshl.u64	q10, q10, #4
+	 veor		q2, q2, q11
+	 vshl.u64	q11, q11, #4
+	veor		q1, q1, q10
+	 veor		q0, q0, q11
+	vldmia	r4, {q8}			@ last round key
+	veor	q6, q6, q8
+	veor	q4, q4, q8
+	veor	q2, q2, q8
+	veor	q7, q7, q8
+	veor	q3, q3, q8
+	veor	q5, q5, q8
+	veor	q0, q0, q8
+	veor	q1, q1, q8
+	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	r6,_bsaes_encrypt8
+	vldmia	r4!, {q9}		@ round 0 key
+	sub	r6,r6,#_bsaes_encrypt8-.LM0SR
+
+	vldmia	r6!, {q8}		@ .LM0SR
+_bsaes_encrypt8_alt:
+	veor	q10, q0, q9	@ xor with round0 key
+	veor	q11, q1, q9
+	 vtbl.8	d0, {q10}, d16
+	 vtbl.8	d1, {q10}, d17
+	veor	q12, q2, q9
+	 vtbl.8	d2, {q11}, d16
+	 vtbl.8	d3, {q11}, d17
+	veor	q13, q3, q9
+	 vtbl.8	d4, {q12}, d16
+	 vtbl.8	d5, {q12}, d17
+	veor	q14, q4, q9
+	 vtbl.8	d6, {q13}, d16
+	 vtbl.8	d7, {q13}, d17
+	veor	q15, q5, q9
+	 vtbl.8	d8, {q14}, d16
+	 vtbl.8	d9, {q14}, d17
+	veor	q10, q6, q9
+	 vtbl.8	d10, {q15}, d16
+	 vtbl.8	d11, {q15}, d17
+	veor	q11, q7, q9
+	 vtbl.8	d12, {q10}, d16
+	 vtbl.8	d13, {q10}, d17
+	 vtbl.8	d14, {q11}, d16
+	 vtbl.8	d15, {q11}, d17
+_bsaes_encrypt8_bitslice:
+	vmov.i8	q8,#0x55			@ compose .LBS0
+	vmov.i8	q9,#0x33			@ compose .LBS1
+	vshr.u64	q10, q6, #1
+	 vshr.u64	q11, q4, #1
+	veor		q10, q10, q7
+	 veor		q11, q11, q5
+	vand		q10, q10, q8
+	 vand		q11, q11, q8
+	veor		q7, q7, q10
+	vshl.u64	q10, q10, #1
+	 veor		q5, q5, q11
+	 vshl.u64	q11, q11, #1
+	veor		q6, q6, q10
+	 veor		q4, q4, q11
+	vshr.u64	q10, q2, #1
+	 vshr.u64	q11, q0, #1
+	veor		q10, q10, q3
+	 veor		q11, q11, q1
+	vand		q10, q10, q8
+	 vand		q11, q11, q8
+	veor		q3, q3, q10
+	vshl.u64	q10, q10, #1
+	 veor		q1, q1, q11
+	 vshl.u64	q11, q11, #1
+	veor		q2, q2, q10
+	 veor		q0, q0, q11
+	vmov.i8	q8,#0x0f			@ compose .LBS2
+	vshr.u64	q10, q5, #2
+	 vshr.u64	q11, q4, #2
+	veor		q10, q10, q7
+	 veor		q11, q11, q6
+	vand		q10, q10, q9
+	 vand		q11, q11, q9
+	veor		q7, q7, q10
+	vshl.u64	q10, q10, #2
+	 veor		q6, q6, q11
+	 vshl.u64	q11, q11, #2
+	veor		q5, q5, q10
+	 veor		q4, q4, q11
+	vshr.u64	q10, q1, #2
+	 vshr.u64	q11, q0, #2
+	veor		q10, q10, q3
+	 veor		q11, q11, q2
+	vand		q10, q10, q9
+	 vand		q11, q11, q9
+	veor		q3, q3, q10
+	vshl.u64	q10, q10, #2
+	 veor		q2, q2, q11
+	 vshl.u64	q11, q11, #2
+	veor		q1, q1, q10
+	 veor		q0, q0, q11
+	vshr.u64	q10, q3, #4
+	 vshr.u64	q11, q2, #4
+	veor		q10, q10, q7
+	 veor		q11, q11, q6
+	vand		q10, q10, q8
+	 vand		q11, q11, q8
+	veor		q7, q7, q10
+	vshl.u64	q10, q10, #4
+	 veor		q6, q6, q11
+	 vshl.u64	q11, q11, #4
+	veor		q3, q3, q10
+	 veor		q2, q2, q11
+	vshr.u64	q10, q1, #4
+	 vshr.u64	q11, q0, #4
+	veor		q10, q10, q5
+	 veor		q11, q11, q4
+	vand		q10, q10, q8
+	 vand		q11, q11, q8
+	veor		q5, q5, q10
+	vshl.u64	q10, q10, #4
+	 veor		q4, q4, q11
+	 vshl.u64	q11, q11, #4
+	veor		q1, q1, q10
+	 veor		q0, q0, q11
+	sub	r5,r5,#1
+	b	.Lenc_sbox
+.align	4
+.Lenc_loop:
+	vldmia	r4!, {q8-q11}
+	veor	q8, q8, q0
+	veor	q9, q9, q1
+	vtbl.8	d0, {q8}, d24
+	vtbl.8	d1, {q8}, d25
+	vldmia	r4!, {q8}
+	veor	q10, q10, q2
+	vtbl.8	d2, {q9}, d24
+	vtbl.8	d3, {q9}, d25
+	vldmia	r4!, {q9}
+	veor	q11, q11, q3
+	vtbl.8	d4, {q10}, d24
+	vtbl.8	d5, {q10}, d25
+	vldmia	r4!, {q10}
+	vtbl.8	d6, {q11}, d24
+	vtbl.8	d7, {q11}, d25
+	vldmia	r4!, {q11}
+	veor	q8, q8, q4
+	veor	q9, q9, q5
+	vtbl.8	d8, {q8}, d24
+	vtbl.8	d9, {q8}, d25
+	veor	q10, q10, q6
+	vtbl.8	d10, {q9}, d24
+	vtbl.8	d11, {q9}, d25
+	veor	q11, q11, q7
+	vtbl.8	d12, {q10}, d24
+	vtbl.8	d13, {q10}, d25
+	vtbl.8	d14, {q11}, d24
+	vtbl.8	d15, {q11}, d25
+.Lenc_sbox:
+	veor	q2, q2, q1
+	veor	q5, q5, q6
+	veor	q3, q3, q0
+	veor	q6, q6, q2
+	veor	q5, q5, q0
+
+	veor	q6, q6, q3
+	veor	q3, q3, q7
+	veor	q7, q7, q5
+	veor	q3, q3, q4
+	veor	q4, q4, q5
+
+	veor	q2, q2, q7
+	veor	q3, q3, q1
+	veor	q1, q1, q5
+	veor	q11, q7, q4
+	veor	q10, q1, q2
+	veor	q9, q5, q3
+	veor	q13, q2, q4
+	 vmov	q8, q10
+	veor	q12, q6, q0
+
+	vorr	q10, q10, q9
+	veor	q15, q11, q8
+	vand	q14, q11, q12
+	vorr	q11, q11, q12
+	veor	q12, q12, q9
+	vand	q8, q8, q9
+	veor	q9, q3, q0
+	vand	q15, q15, q12
+	vand	q13, q13, q9
+	veor	q9, q7, q1
+	veor	q12, q5, q6
+	veor	q11, q11, q13
+	veor	q10, q10, q13
+	vand	q13, q9, q12
+	vorr	q9, q9, q12
+	veor	q11, q11, q15
+	veor	q8, q8, q13
+	veor	q10, q10, q14
+	veor	q9, q9, q15
+	veor	q8, q8, q14
+	vand	q12, q2, q3
+	veor	q9, q9, q14
+	vand	q13, q4, q0
+	vand	q14, q1, q5
+	vorr	q15, q7, q6
+	veor	q11, q11, q12
+	veor	q9, q9, q14
+	veor	q8, q8, q15
+	veor	q10, q10, q13
+
+	@ Inv_GF16 	0, 	1, 	2, 	3, s0, s1, s2, s3
+
+	@ new smaller inversion
+
+	vand	q14, q11, q9
+	vmov	q12, q8
+
+	veor	q13, q10, q14
+	veor	q15, q8, q14
+	veor	q14, q8, q14	@ q14=q15
+
+	vbsl	q13, q9, q8
+	vbsl	q15, q11, q10
+	veor	q11, q11, q10
+
+	vbsl	q12, q13, q14
+	vbsl	q8, q14, q13
+
+	vand	q14, q12, q15
+	veor	q9, q9, q8
+
+	veor	q14, q14, q11
+	veor	q12, q6, q0
+	veor	q8, q5, q3
+	veor 	q10, q15, q14
+	vand	q10, q10, q6
+	veor	q6, q6, q5
+	vand	q11, q5, q15
+	vand	q6, q6, q14
+	veor	q5, q11, q10
+	veor	q6, q6, q11
+	veor	q15, q15, q13
+	veor	q14, q14, q9
+	veor	q11, q15, q14
+	 veor 	q10, q13, q9
+	vand	q11, q11, q12
+	 vand	q10, q10, q0
+	veor	q12, q12, q8
+	 veor	q0, q0, q3
+	vand	q8, q8, q15
+	 vand	q3, q3, q13
+	vand	q12, q12, q14
+	 vand	q0, q0, q9
+	veor	q8, q8, q12
+	 veor	q0, q0, q3
+	veor	q12, q12, q11
+	 veor	q3, q3, q10
+	veor	q6, q6, q12
+	veor	q0, q0, q12
+	veor	q5, q5, q8
+	veor	q3, q3, q8
+
+	veor	q12, q7, q4
+	veor	q8, q1, q2
+	veor	q11, q15, q14
+	 veor 	q10, q13, q9
+	vand	q11, q11, q12
+	 vand	q10, q10, q4
+	veor	q12, q12, q8
+	 veor	q4, q4, q2
+	vand	q8, q8, q15
+	 vand	q2, q2, q13
+	vand	q12, q12, q14
+	 vand	q4, q4, q9
+	veor	q8, q8, q12
+	 veor	q4, q4, q2
+	veor	q12, q12, q11
+	 veor	q2, q2, q10
+	veor	q15, q15, q13
+	veor	q14, q14, q9
+	veor 	q10, q15, q14
+	vand	q10, q10, q7
+	veor	q7, q7, q1
+	vand	q11, q1, q15
+	vand	q7, q7, q14
+	veor	q1, q11, q10
+	veor	q7, q7, q11
+	veor	q7, q7, q12
+	veor	q4, q4, q12
+	veor	q1, q1, q8
+	veor	q2, q2, q8
+	veor	q7, q7, q0
+	veor	q1, q1, q6
+	veor	q6, q6, q0
+	veor	q4, q4, q7
+	veor	q0, q0, q1
+
+	veor	q1, q1, q5
+	veor	q5, q5, q2
+	veor	q2, q2, q3
+	veor	q3, q3, q5
+	veor	q4, q4, q5
+
+	veor	q6, q6, q3
+	subs	r5,r5,#1
+	bcc	.Lenc_done
+	vext.8	q8, q0, q0, #12	@ x0 <<< 32
+	vext.8	q9, q1, q1, #12
+	 veor	q0, q0, q8		@ x0 ^ (x0 <<< 32)
+	vext.8	q10, q4, q4, #12
+	 veor	q1, q1, q9
+	vext.8	q11, q6, q6, #12
+	 veor	q4, q4, q10
+	vext.8	q12, q3, q3, #12
+	 veor	q6, q6, q11
+	vext.8	q13, q7, q7, #12
+	 veor	q3, q3, q12
+	vext.8	q14, q2, q2, #12
+	 veor	q7, q7, q13
+	vext.8	q15, q5, q5, #12
+	 veor	q2, q2, q14
+
+	veor	q9, q9, q0
+	 veor	q5, q5, q15
+	 vext.8	q0, q0, q0, #8		@ (x0 ^ (x0 <<< 32)) <<< 64)
+	veor	q10, q10, q1
+	veor	q8, q8, q5
+	veor	q9, q9, q5
+	 vext.8	q1, q1, q1, #8
+	veor	q13, q13, q3
+	 veor	q0, q0, q8
+	veor	q14, q14, q7
+	 veor	q1, q1, q9
+	 vext.8	q8, q3, q3, #8
+	veor	q12, q12, q6
+	 vext.8	q9, q7, q7, #8
+	veor	q15, q15, q2
+	 vext.8	q3, q6, q6, #8
+	veor	q11, q11, q4
+	 vext.8	q7, q5, q5, #8
+	veor	q12, q12, q5
+	 vext.8	q6, q2, q2, #8
+	veor	q11, q11, q5
+	 vext.8	q2, q4, q4, #8
+	veor	q5, q9, q13
+	veor	q4, q8, q12
+	veor	q3, q3, q11
+	veor	q7, q7, q15
+	veor	q6, q6, q14
+	 @ vmov	q4, q8
+	veor	q2, q2, q10
+	 @ vmov	q5, q9
+	vldmia	r6, {q12}		@ .LSR
+	ite	eq				@ Thumb2 thing, samity check in ARM
+	addeq	r6,r6,#0x10
+	bne	.Lenc_loop
+	vldmia	r6, {q12}		@ .LSRM0
+	b	.Lenc_loop
+.align	4
+.Lenc_done:
+	vmov.i8	q8,#0x55			@ compose .LBS0
+	vmov.i8	q9,#0x33			@ compose .LBS1
+	vshr.u64	q10, q2, #1
+	 vshr.u64	q11, q3, #1
+	veor		q10, q10, q5
+	 veor		q11, q11, q7
+	vand		q10, q10, q8
+	 vand		q11, q11, q8
+	veor		q5, q5, q10
+	vshl.u64	q10, q10, #1
+	 veor		q7, q7, q11
+	 vshl.u64	q11, q11, #1
+	veor		q2, q2, q10
+	 veor		q3, q3, q11
+	vshr.u64	q10, q4, #1
+	 vshr.u64	q11, q0, #1
+	veor		q10, q10, q6
+	 veor		q11, q11, q1
+	vand		q10, q10, q8
+	 vand		q11, q11, q8
+	veor		q6, q6, q10
+	vshl.u64	q10, q10, #1
+	 veor		q1, q1, q11
+	 vshl.u64	q11, q11, #1
+	veor		q4, q4, q10
+	 veor		q0, q0, q11
+	vmov.i8	q8,#0x0f			@ compose .LBS2
+	vshr.u64	q10, q7, #2
+	 vshr.u64	q11, q3, #2
+	veor		q10, q10, q5
+	 veor		q11, q11, q2
+	vand		q10, q10, q9
+	 vand		q11, q11, q9
+	veor		q5, q5, q10
+	vshl.u64	q10, q10, #2
+	 veor		q2, q2, q11
+	 vshl.u64	q11, q11, #2
+	veor		q7, q7, q10
+	 veor		q3, q3, q11
+	vshr.u64	q10, q1, #2
+	 vshr.u64	q11, q0, #2
+	veor		q10, q10, q6
+	 veor		q11, q11, q4
+	vand		q10, q10, q9
+	 vand		q11, q11, q9
+	veor		q6, q6, q10
+	vshl.u64	q10, q10, #2
+	 veor		q4, q4, q11
+	 vshl.u64	q11, q11, #2
+	veor		q1, q1, q10
+	 veor		q0, q0, q11
+	vshr.u64	q10, q6, #4
+	 vshr.u64	q11, q4, #4
+	veor		q10, q10, q5
+	 veor		q11, q11, q2
+	vand		q10, q10, q8
+	 vand		q11, q11, q8
+	veor		q5, q5, q10
+	vshl.u64	q10, q10, #4
+	 veor		q2, q2, q11
+	 vshl.u64	q11, q11, #4
+	veor		q6, q6, q10
+	 veor		q4, q4, q11
+	vshr.u64	q10, q1, #4
+	 vshr.u64	q11, q0, #4
+	veor		q10, q10, q7
+	 veor		q11, q11, q3
+	vand		q10, q10, q8
+	 vand		q11, q11, q8
+	veor		q7, q7, q10
+	vshl.u64	q10, q10, #4
+	 veor		q3, q3, q11
+	 vshl.u64	q11, q11, #4
+	veor		q1, q1, q10
+	 veor		q0, q0, q11
+	vldmia	r4, {q8}			@ last round key
+	veor	q4, q4, q8
+	veor	q6, q6, q8
+	veor	q3, q3, q8
+	veor	q7, q7, q8
+	veor	q2, q2, q8
+	veor	q5, q5, q8
+	veor	q0, q0, q8
+	veor	q1, q1, q8
+	bx	lr
+.size	_bsaes_encrypt8,.-_bsaes_encrypt8
+.type	_bsaes_key_convert,%function
+.align	4
+_bsaes_key_convert:
+	adr	r6,_bsaes_key_convert
+	vld1.8	{q7},  [r4]!		@ load round 0 key
+	sub	r6,r6,#_bsaes_key_convert-.LM0
+	vld1.8	{q15}, [r4]!		@ load round 1 key
+
+	vmov.i8	q8,  #0x01			@ bit masks
+	vmov.i8	q9,  #0x02
+	vmov.i8	q10, #0x04
+	vmov.i8	q11, #0x08
+	vmov.i8	q12, #0x10
+	vmov.i8	q13, #0x20
+	vldmia	r6, {q14}		@ .LM0
+
+#ifdef __ARMEL__
+	vrev32.8	q7,  q7
+	vrev32.8	q15, q15
+#endif
+	sub	r5,r5,#1
+	vstmia	r12!, {q7}		@ save round 0 key
+	b	.Lkey_loop
+
+.align	4
+.Lkey_loop:
+	vtbl.8	d14,{q15},d28
+	vtbl.8	d15,{q15},d29
+	vmov.i8	q6,  #0x40
+	vmov.i8	q15, #0x80
+
+	vtst.8	q0, q7, q8
+	vtst.8	q1, q7, q9
+	vtst.8	q2, q7, q10
+	vtst.8	q3, q7, q11
+	vtst.8	q4, q7, q12
+	vtst.8	q5, q7, q13
+	vtst.8	q6, q7, q6
+	vtst.8	q7, q7, q15
+	vld1.8	{q15}, [r4]!		@ load next round key
+	vmvn	q0, q0		@ "pnot"
+	vmvn	q1, q1
+	vmvn	q5, q5
+	vmvn	q6, q6
+#ifdef __ARMEL__
+	vrev32.8	q15, q15
+#endif
+	subs	r5,r5,#1
+	vstmia	r12!,{q0-q7}		@ write bit-sliced round key
+	bne	.Lkey_loop
+
+	vmov.i8	q7,#0x63			@ compose .L63
+	@ don't save last round key
+	bx	lr
+.size	_bsaes_key_convert,.-_bsaes_key_convert
+.extern AES_cbc_encrypt
+.extern AES_decrypt
+
+.global	bsaes_cbc_encrypt
+.type	bsaes_cbc_encrypt,%function
+.align	5
+bsaes_cbc_encrypt:
+#ifndef	__KERNEL__
+	cmp	r2, #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	r8, [ip]			@ IV is 1st arg on the stack
+	mov	r2, r2, lsr#4		@ len in 16 byte blocks
+	sub	sp, #0x10			@ scratch space to carry over the IV
+	mov	r9, sp				@ save sp
+
+	ldr	r10, [r3, #240]		@ get # of rounds
+#ifndef	BSAES_ASM_EXTENDED_KEY
+	@ allocate the key schedule on the stack
+	sub	r12, sp, r10, lsl#7		@ 128 bytes per inner round key
+	add	r12, #96			@ sifze of bit-slices key schedule
+
+	@ populate the key schedule
+	mov	r4, r3			@ pass key
+	mov	r5, r10			@ pass # of rounds
+	mov	sp, r12				@ sp is sp
+	bl	_bsaes_key_convert
+	vldmia	sp, {q6}
+	vstmia	r12,  {q15}		@ save last round key
+	veor	q7, q7, q6	@ fix up round 0 key
+	vstmia	sp, {q7}
+#else
+	ldr	r12, [r3, #244]
+	eors	r12, #1
+	beq	0f
+
+	@ populate the key schedule
+	str	r12, [r3, #244]
+	mov	r4, r3			@ pass key
+	mov	r5, r10			@ pass # of rounds
+	add	r12, r3, #248			@ pass key schedule
+	bl	_bsaes_key_convert
+	add	r4, r3, #248
+	vldmia	r4, {q6}
+	vstmia	r12, {q15}			@ save last round key
+	veor	q7, q7, q6	@ fix up round 0 key
+	vstmia	r4, {q7}
+
+.align	2
+0:
+#endif
+
+	vld1.8	{q15}, [r8]		@ load IV
+	b	.Lcbc_dec_loop
+
+.align	4
+.Lcbc_dec_loop:
+	subs	r2, r2, #0x8
+	bmi	.Lcbc_dec_loop_finish
+
+	vld1.8	{q0-q1}, [r0]!	@ load input
+	vld1.8	{q2-q3}, [r0]!
+#ifndef	BSAES_ASM_EXTENDED_KEY
+	mov	r4, sp			@ pass the key
+#else
+	add	r4, r3, #248
+#endif
+	vld1.8	{q4-q5}, [r0]!
+	mov	r5, r10
+	vld1.8	{q6-q7}, [r0]
+	sub	r0, r0, #0x60
+	vstmia	r9, {q15}			@ put aside IV
+
+	bl	_bsaes_decrypt8
+
+	vldmia	r9, {q14}			@ reload IV
+	vld1.8	{q8-q9}, [r0]!	@ reload input
+	veor	q0, q0, q14	@ ^= IV
+	vld1.8	{q10-q11}, [r0]!
+	veor	q1, q1, q8
+	veor	q6, q6, q9
+	vld1.8	{q12-q13}, [r0]!
+	veor	q4, q4, q10
+	veor	q2, q2, q11
+	vld1.8	{q14-q15}, [r0]!
+	veor	q7, q7, q12
+	vst1.8	{q0-q1}, [r1]!	@ write output
+	veor	q3, q3, q13
+	vst1.8	{q6}, [r1]!
+	veor	q5, q5, q14
+	vst1.8	{q4}, [r1]!
+	vst1.8	{q2}, [r1]!
+	vst1.8	{q7}, [r1]!
+	vst1.8	{q3}, [r1]!
+	vst1.8	{q5}, [r1]!
+
+	b	.Lcbc_dec_loop
+
+.Lcbc_dec_loop_finish:
+	adds	r2, r2, #8
+	beq	.Lcbc_dec_done
+
+	vld1.8	{q0}, [r0]!		@ load input
+	cmp	r2, #2
+	blo	.Lcbc_dec_one
+	vld1.8	{q1}, [r0]!
+#ifndef	BSAES_ASM_EXTENDED_KEY
+	mov	r4, sp			@ pass the key
+#else
+	add	r4, r3, #248
+#endif
+	mov	r5, r10
+	vstmia	r9, {q15}			@ put aside IV
+	beq	.Lcbc_dec_two
+	vld1.8	{q2}, [r0]!
+	cmp	r2, #4
+	blo	.Lcbc_dec_three
+	vld1.8	{q3}, [r0]!
+	beq	.Lcbc_dec_four
+	vld1.8	{q4}, [r0]!
+	cmp	r2, #6
+	blo	.Lcbc_dec_five
+	vld1.8	{q5}, [r0]!
+	beq	.Lcbc_dec_six
+	vld1.8	{q6}, [r0]!
+	sub	r0, r0, #0x70
+
+	bl	_bsaes_decrypt8
+
+	vldmia	r9, {q14}			@ reload IV
+	vld1.8	{q8-q9}, [r0]!	@ reload input
+	veor	q0, q0, q14	@ ^= IV
+	vld1.8	{q10-q11}, [r0]!
+	veor	q1, q1, q8
+	veor	q6, q6, q9
+	vld1.8	{q12-q13}, [r0]!
+	veor	q4, q4, q10
+	veor	q2, q2, q11
+	vld1.8	{q15}, [r0]!
+	veor	q7, q7, q12
+	vst1.8	{q0-q1}, [r1]!	@ write output
+	veor	q3, q3, q13
+	vst1.8	{q6}, [r1]!
+	vst1.8	{q4}, [r1]!
+	vst1.8	{q2}, [r1]!
+	vst1.8	{q7}, [r1]!
+	vst1.8	{q3}, [r1]!
+	b	.Lcbc_dec_done
+.align	4
+.Lcbc_dec_six:
+	sub	r0, r0, #0x60
+	bl	_bsaes_decrypt8
+	vldmia	r9,{q14}			@ reload IV
+	vld1.8	{q8-q9}, [r0]!	@ reload input
+	veor	q0, q0, q14	@ ^= IV
+	vld1.8	{q10-q11}, [r0]!
+	veor	q1, q1, q8
+	veor	q6, q6, q9
+	vld1.8	{q12}, [r0]!
+	veor	q4, q4, q10
+	veor	q2, q2, q11
+	vld1.8	{q15}, [r0]!
+	veor	q7, q7, q12
+	vst1.8	{q0-q1}, [r1]!	@ write output
+	vst1.8	{q6}, [r1]!
+	vst1.8	{q4}, [r1]!
+	vst1.8	{q2}, [r1]!
+	vst1.8	{q7}, [r1]!
+	b	.Lcbc_dec_done
+.align	4
+.Lcbc_dec_five:
+	sub	r0, r0, #0x50
+	bl	_bsaes_decrypt8
+	vldmia	r9, {q14}			@ reload IV
+	vld1.8	{q8-q9}, [r0]!	@ reload input
+	veor	q0, q0, q14	@ ^= IV
+	vld1.8	{q10-q11}, [r0]!
+	veor	q1, q1, q8
+	veor	q6, q6, q9
+	vld1.8	{q15}, [r0]!
+	veor	q4, q4, q10
+	vst1.8	{q0-q1}, [r1]!	@ write output
+	veor	q2, q2, q11
+	vst1.8	{q6}, [r1]!
+	vst1.8	{q4}, [r1]!
+	vst1.8	{q2}, [r1]!
+	b	.Lcbc_dec_done
+.align	4
+.Lcbc_dec_four:
+	sub	r0, r0, #0x40
+	bl	_bsaes_decrypt8
+	vldmia	r9, {q14}			@ reload IV
+	vld1.8	{q8-q9}, [r0]!	@ reload input
+	veor	q0, q0, q14	@ ^= IV
+	vld1.8	{q10}, [r0]!
+	veor	q1, q1, q8
+	veor	q6, q6, q9
+	vld1.8	{q15}, [r0]!
+	veor	q4, q4, q10
+	vst1.8	{q0-q1}, [r1]!	@ write output
+	vst1.8	{q6}, [r1]!
+	vst1.8	{q4}, [r1]!
+	b	.Lcbc_dec_done
+.align	4
+.Lcbc_dec_three:
+	sub	r0, r0, #0x30
+	bl	_bsaes_decrypt8
+	vldmia	r9, {q14}			@ reload IV
+	vld1.8	{q8-q9}, [r0]!	@ reload input
+	veor	q0, q0, q14	@ ^= IV
+	vld1.8	{q15}, [r0]!
+	veor	q1, q1, q8
+	veor	q6, q6, q9
+	vst1.8	{q0-q1}, [r1]!	@ write output
+	vst1.8	{q6}, [r1]!
+	b	.Lcbc_dec_done
+.align	4
+.Lcbc_dec_two:
+	sub	r0, r0, #0x20
+	bl	_bsaes_decrypt8
+	vldmia	r9, {q14}			@ reload IV
+	vld1.8	{q8}, [r0]!		@ reload input
+	veor	q0, q0, q14	@ ^= IV
+	vld1.8	{q15}, [r0]!		@ reload input
+	veor	q1, q1, q8
+	vst1.8	{q0-q1}, [r1]!	@ write output
+	b	.Lcbc_dec_done
+.align	4
+.Lcbc_dec_one:
+	sub	r0, r0, #0x10
+	mov	r10, r1			@ save original out pointer
+	mov	r1, r9			@ use the iv scratch space as out buffer
+	mov	r2, r3
+	vmov	q4,q15		@ just in case ensure that IV
+	vmov	q5,q0			@ and input are preserved
+	bl	AES_decrypt
+	vld1.8	{q0}, [r9,:64]		@ load result
+	veor	q0, q0, q4	@ ^= IV
+	vmov	q15, q5		@ q5 holds input
+	vst1.8	{q0}, [r10]		@ 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		sp!, {q0-q1}
+	cmp		sp, r9
+	bne		.Lcbc_dec_bzero
+#endif
+
+	mov	sp, r9
+	add	sp, #0x10			@ add sp,r9,#0x10 is no good for thumb
+	vst1.8	{q15}, [r8]		@ return IV
+	VFP_ABI_POP
+	ldmia	sp!, {r4-r10, pc}
+.size	bsaes_cbc_encrypt,.-bsaes_cbc_encrypt
+.extern	AES_encrypt
+.global	bsaes_ctr32_encrypt_blocks
+.type	bsaes_ctr32_encrypt_blocks,%function
+.align	5
+bsaes_ctr32_encrypt_blocks:
+	cmp	r2, #8			@ use plain AES for
+	blo	.Lctr_enc_short			@ small sizes
+
+	mov	ip, sp
+	stmdb	sp!, {r4-r10, lr}
+	VFP_ABI_PUSH
+	ldr	r8, [ip]			@ ctr is 1st arg on the stack
+	sub	sp, sp, #0x10			@ scratch space to carry over the ctr
+	mov	r9, sp				@ save sp
+
+	ldr	r10, [r3, #240]		@ get # of rounds
+#ifndef	BSAES_ASM_EXTENDED_KEY
+	@ allocate the key schedule on the stack
+	sub	r12, sp, r10, lsl#7		@ 128 bytes per inner round key
+	add	r12, #96			@ size of bit-sliced key schedule
+
+	@ populate the key schedule
+	mov	r4, r3			@ pass key
+	mov	r5, r10			@ pass # of rounds
+	mov	sp, r12				@ sp is sp
+	bl	_bsaes_key_convert
+	veor	q7,q7,q15	@ fix up last round key
+	vstmia	r12, {q7}			@ save last round key
+
+	vld1.8	{q0}, [r8]		@ load counter
+	add	r8, r6, #.LREVM0SR-.LM0	@ borrow r8
+	vldmia	sp, {q4}		@ load round0 key
+#else
+	ldr	r12, [r3, #244]
+	eors	r12, #1
+	beq	0f
+
+	@ populate the key schedule
+	str	r12, [r3, #244]
+	mov	r4, r3			@ pass key
+	mov	r5, r10			@ pass # of rounds
+	add	r12, r3, #248			@ pass key schedule
+	bl	_bsaes_key_convert
+	veor	q7,q7,q15	@ fix up last round key
+	vstmia	r12, {q7}			@ save last round key
+
+.align	2
+0:	add	r12, r3, #248
+	vld1.8	{q0}, [r8]		@ load counter
+	adrl	r8, .LREVM0SR			@ borrow r8
+	vldmia	r12, {q4}			@ load round0 key
+	sub	sp, #0x10			@ place for adjusted round0 key
+#endif
+
+	vmov.i32	q8,#1		@ compose 1<<96
+	veor		q9,q9,q9
+	vrev32.8	q0,q0
+	vext.8		q8,q9,q8,#4
+	vrev32.8	q4,q4
+	vadd.u32	q9,q8,q8	@ compose 2<<96
+	vstmia	sp, {q4}		@ save adjusted round0 key
+	b	.Lctr_enc_loop
+
+.align	4
+.Lctr_enc_loop:
+	vadd.u32	q10, q8, q9	@ compose 3<<96
+	vadd.u32	q1, q0, q8	@ +1
+	vadd.u32	q2, q0, q9	@ +2
+	vadd.u32	q3, q0, q10	@ +3
+	vadd.u32	q4, q1, q10
+	vadd.u32	q5, q2, q10
+	vadd.u32	q6, q3, q10
+	vadd.u32	q7, q4, q10
+	vadd.u32	q10, q5, q10	@ next counter
+
+	@ Borrow prologue from _bsaes_encrypt8 to use the opportunity
+	@ to flip byte order in 32-bit counter
+
+	vldmia		sp, {q9}		@ load round0 key
+#ifndef	BSAES_ASM_EXTENDED_KEY
+	add		r4, sp, #0x10		@ pass next round key
+#else
+	add		r4, r3, #264
+#endif
+	vldmia		r8, {q8}			@ .LREVM0SR
+	mov		r5, r10			@ pass rounds
+	vstmia		r9, {q10}			@ save next counter
+	sub		r6, r8, #.LREVM0SR-.LSR	@ pass constants
+
+	bl		_bsaes_encrypt8_alt
+
+	subs		r2, r2, #8
+	blo		.Lctr_enc_loop_done
+
+	vld1.8		{q8-q9}, [r0]!	@ load input
+	vld1.8		{q10-q11}, [r0]!
+	veor		q0, q8
+	veor		q1, q9
+	vld1.8		{q12-q13}, [r0]!
+	veor		q4, q10
+	veor		q6, q11
+	vld1.8		{q14-q15}, [r0]!
+	veor		q3, q12
+	vst1.8		{q0-q1}, [r1]!	@ write output
+	veor		q7, q13
+	veor		q2, q14
+	vst1.8		{q4}, [r1]!
+	veor		q5, q15
+	vst1.8		{q6}, [r1]!
+	vmov.i32	q8, #1			@ compose 1<<96
+	vst1.8		{q3}, [r1]!
+	veor		q9, q9, q9
+	vst1.8		{q7}, [r1]!
+	vext.8		q8, q9, q8, #4
+	vst1.8		{q2}, [r1]!
+	vadd.u32	q9,q8,q8		@ compose 2<<96
+	vst1.8		{q5}, [r1]!
+	vldmia		r9, {q0}			@ load counter
+
+	bne		.Lctr_enc_loop
+	b		.Lctr_enc_done
+
+.align	4
+.Lctr_enc_loop_done:
+	add		r2, r2, #8
+	vld1.8		{q8}, [r0]!	@ load input
+	veor		q0, q8
+	vst1.8		{q0}, [r1]!	@ write output
+	cmp		r2, #2
+	blo		.Lctr_enc_done
+	vld1.8		{q9}, [r0]!
+	veor		q1, q9
+	vst1.8		{q1}, [r1]!
+	beq		.Lctr_enc_done
+	vld1.8		{q10}, [r0]!
+	veor		q4, q10
+	vst1.8		{q4}, [r1]!
+	cmp		r2, #4
+	blo		.Lctr_enc_done
+	vld1.8		{q11}, [r0]!
+	veor		q6, q11
+	vst1.8		{q6}, [r1]!
+	beq		.Lctr_enc_done
+	vld1.8		{q12}, [r0]!
+	veor		q3, q12
+	vst1.8		{q3}, [r1]!
+	cmp		r2, #6
+	blo		.Lctr_enc_done
+	vld1.8		{q13}, [r0]!
+	veor		q7, q13
+	vst1.8		{q7}, [r1]!
+	beq		.Lctr_enc_done
+	vld1.8		{q14}, [r0]
+	veor		q2, q14
+	vst1.8		{q2}, [r1]!
+
+.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		sp!, {q0-q1}
+	cmp		sp, r9
+	bne		.Lctr_enc_bzero
+#else
+	vstmia		sp, {q0-q1}
+#endif
+
+	mov	sp, r9
+	add	sp, #0x10		@ add sp,r9,#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, r0		@ copy arguments
+	mov	r5, r1
+	mov	r6, r2
+	mov	r7, r3
+	ldr	r8, [ip, #12]		@ load counter LSW
+	vld1.8	{q1}, [ip]		@ load whole counter value
+#ifdef __ARMEL__
+	rev	r8, r8
+#endif
+	sub	sp, sp, #0x10
+	vst1.8	{q1}, [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	{q0}, [r4]!	@ load input
+	vld1.8	{q1}, [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	q0,q0,q1
+	vst1.8	{q0}, [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
+.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 r3
+
+	mov	r7, r0
+	mov	r8, r1
+	mov	r9, r2
+	mov	r10, 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	r1, [r10, #240]		@ get # of rounds
+	mov	r3, r6
+#ifndef	BSAES_ASM_EXTENDED_KEY
+	@ allocate the key schedule on the stack
+	sub	r12, sp, r1, lsl#7		@ 128 bytes per inner round key
+	@ add	r12, #96			@ size of bit-sliced key schedule
+	sub	r12, #48			@ place for tweak[9]
+
+	@ populate the key schedule
+	mov	r4, r10			@ pass key
+	mov	r5, r1			@ pass # of rounds
+	mov	sp, r12
+	add	r12, #0x90			@ pass key schedule
+	bl	_bsaes_key_convert
+	veor	q7, q7, q15	@ fix up last round key
+	vstmia	r12, {q7}			@ save last round key
+#else
+	ldr	r12, [r10, #244]
+	eors	r12, #1
+	beq	0f
+
+	str	r12, [r10, #244]
+	mov	r4, r10			@ pass key
+	mov	r5, r1			@ pass # of rounds
+	add	r12, r10, #248			@ pass key schedule
+	bl	_bsaes_key_convert
+	veor	q7, q7, q15	@ fix up last round key
+	vstmia	r12, {q7}
+
+.align	2
+0:	sub	sp, #0x90			@ place for tweak[9]
+#endif
+
+	vld1.8	{q8}, [r0]			@ initial tweak
+	adr	r2, .Lxts_magic
+
+	subs	r9, #0x80
+	blo	.Lxts_enc_short
+	b	.Lxts_enc_loop
+
+.align	4
+.Lxts_enc_loop:
+	vldmia		r2, {q5}	@ load XTS magic
+	vshr.s64	q6, q8, #63
+	mov		r0, sp
+	vand		q6, q6, q5
+	vadd.u64	q9, q8, q8
+	vst1.64		{q8}, [r0,:128]!
+	vswp		d13,d12
+	vshr.s64	q7, q9, #63
+	veor		q9, q9, q6
+	vand		q7, q7, q5
+	vadd.u64	q10, q9, q9
+	vst1.64		{q9}, [r0,:128]!
+	vswp		d15,d14
+	vshr.s64	q6, q10, #63
+	veor		q10, q10, q7
+	vand		q6, q6, q5
+	vld1.8		{q0}, [r7]!
+	vadd.u64	q11, q10, q10
+	vst1.64		{q10}, [r0,:128]!
+	vswp		d13,d12
+	vshr.s64	q7, q11, #63
+	veor		q11, q11, q6
+	vand		q7, q7, q5
+	vld1.8		{q1}, [r7]!
+	veor		q0, q0, q8
+	vadd.u64	q12, q11, q11
+	vst1.64		{q11}, [r0,:128]!
+	vswp		d15,d14
+	vshr.s64	q6, q12, #63
+	veor		q12, q12, q7
+	vand		q6, q6, q5
+	vld1.8		{q2}, [r7]!
+	veor		q1, q1, q9
+	vadd.u64	q13, q12, q12
+	vst1.64		{q12}, [r0,:128]!
+	vswp		d13,d12
+	vshr.s64	q7, q13, #63
+	veor		q13, q13, q6
+	vand		q7, q7, q5
+	vld1.8		{q3}, [r7]!
+	veor		q2, q2, q10
+	vadd.u64	q14, q13, q13
+	vst1.64		{q13}, [r0,:128]!
+	vswp		d15,d14
+	vshr.s64	q6, q14, #63
+	veor		q14, q14, q7
+	vand		q6, q6, q5
+	vld1.8		{q4}, [r7]!
+	veor		q3, q3, q11
+	vadd.u64	q15, q14, q14
+	vst1.64		{q14}, [r0,:128]!
+	vswp		d13,d12
+	vshr.s64	q7, q15, #63
+	veor		q15, q15, q6
+	vand		q7, q7, q5
+	vld1.8		{q5}, [r7]!
+	veor		q4, q4, q12
+	vadd.u64	q8, q15, q15
+	vst1.64		{q15}, [r0,:128]!
+	vswp		d15,d14
+	veor		q8, q8, q7
+	vst1.64		{q8}, [r0,:128]		@ next round tweak
+
+	vld1.8		{q6-q7}, [r7]!
+	veor		q5, q5, q13
+#ifndef	BSAES_ASM_EXTENDED_KEY
+	add		r4, sp, #0x90			@ pass key schedule
+#else
+	add		r4, r10, #248			@ pass key schedule
+#endif
+	veor		q6, q6, q14
+	mov		r5, r1			@ pass rounds
+	veor		q7, q7, q15
+	mov		r0, sp
+
+	bl		_bsaes_encrypt8
+
+	vld1.64		{q8-q9}, [r0,:128]!
+	vld1.64		{q10-q11}, [r0,:128]!
+	veor		q0, q0, q8
+	vld1.64		{q12-q13}, [r0,:128]!
+	veor		q1, q1, q9
+	veor		q8, q4, q10
+	vst1.8		{q0-q1}, [r8]!
+	veor		q9, q6, q11
+	vld1.64		{q14-q15}, [r0,:128]!
+	veor		q10, q3, q12
+	vst1.8		{q8-q9}, [r8]!
+	veor		q11, q7, q13
+	veor		q12, q2, q14
+	vst1.8		{q10-q11}, [r8]!
+	veor		q13, q5, q15
+	vst1.8		{q12-q13}, [r8]!
+
+	vld1.64		{q8}, [r0,:128]		@ next round tweak
+
+	subs		r9, #0x80
+	bpl		.Lxts_enc_loop
+
+.Lxts_enc_short:
+	adds		r9, #0x70
+	bmi		.Lxts_enc_done
+
+	vldmia		r2, {q5}	@ load XTS magic
+	vshr.s64	q7, q8, #63
+	mov		r0, sp
+	vand		q7, q7, q5
+	vadd.u64	q9, q8, q8
+	vst1.64		{q8}, [r0,:128]!
+	vswp		d15,d14
+	vshr.s64	q6, q9, #63
+	veor		q9, q9, q7
+	vand		q6, q6, q5
+	vadd.u64	q10, q9, q9
+	vst1.64		{q9}, [r0,:128]!
+	vswp		d13,d12
+	vshr.s64	q7, q10, #63
+	veor		q10, q10, q6
+	vand		q7, q7, q5
+	vld1.8		{q0}, [r7]!
+	subs		r9, #0x10
+	bmi		.Lxts_enc_1
+	vadd.u64	q11, q10, q10
+	vst1.64		{q10}, [r0,:128]!
+	vswp		d15,d14
+	vshr.s64	q6, q11, #63
+	veor		q11, q11, q7
+	vand		q6, q6, q5
+	vld1.8		{q1}, [r7]!
+	subs		r9, #0x10
+	bmi		.Lxts_enc_2
+	veor		q0, q0, q8
+	vadd.u64	q12, q11, q11
+	vst1.64		{q11}, [r0,:128]!
+	vswp		d13,d12
+	vshr.s64	q7, q12, #63
+	veor		q12, q12, q6
+	vand		q7, q7, q5
+	vld1.8		{q2}, [r7]!
+	subs		r9, #0x10
+	bmi		.Lxts_enc_3
+	veor		q1, q1, q9
+	vadd.u64	q13, q12, q12
+	vst1.64		{q12}, [r0,:128]!
+	vswp		d15,d14
+	vshr.s64	q6, q13, #63
+	veor		q13, q13, q7
+	vand		q6, q6, q5
+	vld1.8		{q3}, [r7]!
+	subs		r9, #0x10
+	bmi		.Lxts_enc_4
+	veor		q2, q2, q10
+	vadd.u64	q14, q13, q13
+	vst1.64		{q13}, [r0,:128]!
+	vswp		d13,d12
+	vshr.s64	q7, q14, #63
+	veor		q14, q14, q6
+	vand		q7, q7, q5
+	vld1.8		{q4}, [r7]!
+	subs		r9, #0x10
+	bmi		.Lxts_enc_5
+	veor		q3, q3, q11
+	vadd.u64	q15, q14, q14
+	vst1.64		{q14}, [r0,:128]!
+	vswp		d15,d14
+	vshr.s64	q6, q15, #63
+	veor		q15, q15, q7
+	vand		q6, q6, q5
+	vld1.8		{q5}, [r7]!
+	subs		r9, #0x10
+	bmi		.Lxts_enc_6
+	veor		q4, q4, q12
+	sub		r9, #0x10
+	vst1.64		{q15}, [r0,:128]		@ next round tweak
+
+	vld1.8		{q6}, [r7]!
+	veor		q5, q5, q13
+#ifndef	BSAES_ASM_EXTENDED_KEY
+	add		r4, sp, #0x90			@ pass key schedule
+#else
+	add		r4, r10, #248			@ pass key schedule
+#endif
+	veor		q6, q6, q14
+	mov		r5, r1			@ pass rounds
+	mov		r0, sp
+
+	bl		_bsaes_encrypt8
+
+	vld1.64		{q8-q9}, [r0,:128]!
+	vld1.64		{q10-q11}, [r0,:128]!
+	veor		q0, q0, q8
+	vld1.64		{q12-q13}, [r0,:128]!
+	veor		q1, q1, q9
+	veor		q8, q4, q10
+	vst1.8		{q0-q1}, [r8]!
+	veor		q9, q6, q11
+	vld1.64		{q14}, [r0,:128]!
+	veor		q10, q3, q12
+	vst1.8		{q8-q9}, [r8]!
+	veor		q11, q7, q13
+	veor		q12, q2, q14
+	vst1.8		{q10-q11}, [r8]!
+	vst1.8		{q12}, [r8]!
+
+	vld1.64		{q8}, [r0,:128]		@ next round tweak
+	b		.Lxts_enc_done
+.align	4
+.Lxts_enc_6:
+	vst1.64		{q14}, [r0,:128]		@ next round tweak
+
+	veor		q4, q4, q12
+#ifndef	BSAES_ASM_EXTENDED_KEY
+	add		r4, sp, #0x90			@ pass key schedule
+#else
+	add		r4, r10, #248			@ pass key schedule
+#endif
+	veor		q5, q5, q13
+	mov		r5, r1			@ pass rounds
+	mov		r0, sp
+
+	bl		_bsaes_encrypt8
+
+	vld1.64		{q8-q9}, [r0,:128]!
+	vld1.64		{q10-q11}, [r0,:128]!
+	veor		q0, q0, q8
+	vld1.64		{q12-q13}, [r0,:128]!
+	veor		q1, q1, q9
+	veor		q8, q4, q10
+	vst1.8		{q0-q1}, [r8]!
+	veor		q9, q6, q11
+	veor		q10, q3, q12
+	vst1.8		{q8-q9}, [r8]!
+	veor		q11, q7, q13
+	vst1.8		{q10-q11}, [r8]!
+
+	vld1.64		{q8}, [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		{q13}, [r0,:128]		@ next round tweak
+
+	veor		q3, q3, q11
+#ifndef	BSAES_ASM_EXTENDED_KEY
+	add		r4, sp, #0x90			@ pass key schedule
+#else
+	add		r4, r10, #248			@ pass key schedule
+#endif
+	veor		q4, q4, q12
+	mov		r5, r1			@ pass rounds
+	mov		r0, sp
+
+	bl		_bsaes_encrypt8
+
+	vld1.64		{q8-q9}, [r0,:128]!
+	vld1.64		{q10-q11}, [r0,:128]!
+	veor		q0, q0, q8
+	vld1.64		{q12}, [r0,:128]!
+	veor		q1, q1, q9
+	veor		q8, q4, q10
+	vst1.8		{q0-q1}, [r8]!
+	veor		q9, q6, q11
+	veor		q10, q3, q12
+	vst1.8		{q8-q9}, [r8]!
+	vst1.8		{q10}, [r8]!
+
+	vld1.64		{q8}, [r0,:128]		@ next round tweak
+	b		.Lxts_enc_done
+.align	4
+.Lxts_enc_4:
+	vst1.64		{q12}, [r0,:128]		@ next round tweak
+
+	veor		q2, q2, q10
+#ifndef	BSAES_ASM_EXTENDED_KEY
+	add		r4, sp, #0x90			@ pass key schedule
+#else
+	add		r4, r10, #248			@ pass key schedule
+#endif
+	veor		q3, q3, q11
+	mov		r5, r1			@ pass rounds
+	mov		r0, sp
+
+	bl		_bsaes_encrypt8
+
+	vld1.64		{q8-q9}, [r0,:128]!
+	vld1.64		{q10-q11}, [r0,:128]!
+	veor		q0, q0, q8
+	veor		q1, q1, q9
+	veor		q8, q4, q10
+	vst1.8		{q0-q1}, [r8]!
+	veor		q9, q6, q11
+	vst1.8		{q8-q9}, [r8]!
+
+	vld1.64		{q8}, [r0,:128]		@ next round tweak
+	b		.Lxts_enc_done
+.align	4
+.Lxts_enc_3:
+	vst1.64		{q11}, [r0,:128]		@ next round tweak
+
+	veor		q1, q1, q9
+#ifndef	BSAES_ASM_EXTENDED_KEY
+	add		r4, sp, #0x90			@ pass key schedule
+#else
+	add		r4, r10, #248			@ pass key schedule
+#endif
+	veor		q2, q2, q10
+	mov		r5, r1			@ pass rounds
+	mov		r0, sp
+
+	bl		_bsaes_encrypt8
+
+	vld1.64		{q8-q9}, [r0,:128]!
+	vld1.64		{q10}, [r0,:128]!
+	veor		q0, q0, q8
+	veor		q1, q1, q9
+	veor		q8, q4, q10
+	vst1.8		{q0-q1}, [r8]!
+	vst1.8		{q8}, [r8]!
+
+	vld1.64		{q8}, [r0,:128]		@ next round tweak
+	b		.Lxts_enc_done
+.align	4
+.Lxts_enc_2:
+	vst1.64		{q10}, [r0,:128]		@ next round tweak
+
+	veor		q0, q0, q8
+#ifndef	BSAES_ASM_EXTENDED_KEY
+	add		r4, sp, #0x90			@ pass key schedule
+#else
+	add		r4, r10, #248			@ pass key schedule
+#endif
+	veor		q1, q1, q9
+	mov		r5, r1			@ pass rounds
+	mov		r0, sp
+
+	bl		_bsaes_encrypt8
+
+	vld1.64		{q8-q9}, [r0,:128]!
+	veor		q0, q0, q8
+	veor		q1, q1, q9
+	vst1.8		{q0-q1}, [r8]!
+
+	vld1.64		{q8}, [r0,:128]		@ next round tweak
+	b		.Lxts_enc_done
+.align	4
+.Lxts_enc_1:
+	mov		r0, sp
+	veor		q0, q8
+	mov		r1, sp
+	vst1.8		{q0}, [sp,:128]
+	mov		r2, r10
+	mov		r4, r3				@ preserve fp
+
+	bl		AES_encrypt
+
+	vld1.8		{q0}, [sp,:128]
+	veor		q0, q0, q8
+	vst1.8		{q0}, [r8]!
+	mov		r3, r4
+
+	vmov		q8, q9		@ next round tweak
+
+.Lxts_enc_done:
+#ifndef	XTS_CHAIN_TWEAK
+	adds		r9, #0x10
+	beq		.Lxts_enc_ret
+	sub		r6, r8, #0x10
+
+.Lxts_enc_steal:
+	ldrb		r0, [r7], #1
+	ldrb		r1, [r8, #-0x10]
+	strb		r0, [r8, #-0x10]
+	strb		r1, [r8], #1
+
+	subs		r9, #1
+	bhi		.Lxts_enc_steal
+
+	vld1.8		{q0}, [r6]
+	mov		r0, sp
+	veor		q0, q0, q8
+	mov		r1, sp
+	vst1.8		{q0}, [sp,:128]
+	mov		r2, r10
+	mov		r4, r3			@ preserve fp
+
+	bl		AES_encrypt
+
+	vld1.8		{q0}, [sp,:128]
+	veor		q0, q0, q8
+	vst1.8		{q0}, [r6]
+	mov		r3, r4
+#endif
+
+.Lxts_enc_ret:
+	bic		r0, r3, #0xf
+	vmov.i32	q0, #0
+	vmov.i32	q1, #0
+#ifdef	XTS_CHAIN_TWEAK
+	ldr		r1, [r3, #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, r3
+#ifdef	XTS_CHAIN_TWEAK
+	vst1.8		{q8}, [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 r3
+
+	mov	r7, r0
+	mov	r8, r1
+	mov	r9, r2
+	mov	r10, 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	r1, [r10, #240]		@ get # of rounds
+	mov	r3, r6
+#ifndef	BSAES_ASM_EXTENDED_KEY
+	@ allocate the key schedule on the stack
+	sub	r12, sp, r1, lsl#7		@ 128 bytes per inner round key
+	@ add	r12, #96			@ size of bit-sliced key schedule
+	sub	r12, #48			@ place for tweak[9]
+
+	@ populate the key schedule
+	mov	r4, r10			@ pass key
+	mov	r5, r1			@ pass # of rounds
+	mov	sp, r12
+	add	r12, #0x90			@ pass key schedule
+	bl	_bsaes_key_convert
+	add	r4, sp, #0x90
+	vldmia	r4, {q6}
+	vstmia	r12,  {q15}		@ save last round key
+	veor	q7, q7, q6	@ fix up round 0 key
+	vstmia	r4, {q7}
+#else
+	ldr	r12, [r10, #244]
+	eors	r12, #1
+	beq	0f
+
+	str	r12, [r10, #244]
+	mov	r4, r10			@ pass key
+	mov	r5, r1			@ pass # of rounds
+	add	r12, r10, #248			@ pass key schedule
+	bl	_bsaes_key_convert
+	add	r4, r10, #248
+	vldmia	r4, {q6}
+	vstmia	r12,  {q15}		@ save last round key
+	veor	q7, q7, q6	@ fix up round 0 key
+	vstmia	r4, {q7}
+
+.align	2
+0:	sub	sp, #0x90			@ place for tweak[9]
+#endif
+	vld1.8	{q8}, [r0]			@ initial tweak
+	adr	r2, .Lxts_magic
+
+#ifndef	XTS_CHAIN_TWEAK
+	tst	r9, #0xf			@ if not multiple of 16
+	it	ne				@ Thumb2 thing, sanity check in ARM
+	subne	r9, #0x10			@ subtract another 16 bytes
+#endif
+	subs	r9, #0x80
+
+	blo	.Lxts_dec_short
+	b	.Lxts_dec_loop
+
+.align	4
+.Lxts_dec_loop:
+	vldmia		r2, {q5}	@ load XTS magic
+	vshr.s64	q6, q8, #63
+	mov		r0, sp
+	vand		q6, q6, q5
+	vadd.u64	q9, q8, q8
+	vst1.64		{q8}, [r0,:128]!
+	vswp		d13,d12
+	vshr.s64	q7, q9, #63
+	veor		q9, q9, q6
+	vand		q7, q7, q5
+	vadd.u64	q10, q9, q9
+	vst1.64		{q9}, [r0,:128]!
+	vswp		d15,d14
+	vshr.s64	q6, q10, #63
+	veor		q10, q10, q7
+	vand		q6, q6, q5
+	vld1.8		{q0}, [r7]!
+	vadd.u64	q11, q10, q10
+	vst1.64		{q10}, [r0,:128]!
+	vswp		d13,d12
+	vshr.s64	q7, q11, #63
+	veor		q11, q11, q6
+	vand		q7, q7, q5
+	vld1.8		{q1}, [r7]!
+	veor		q0, q0, q8
+	vadd.u64	q12, q11, q11
+	vst1.64		{q11}, [r0,:128]!
+	vswp		d15,d14
+	vshr.s64	q6, q12, #63
+	veor		q12, q12, q7
+	vand		q6, q6, q5
+	vld1.8		{q2}, [r7]!
+	veor		q1, q1, q9
+	vadd.u64	q13, q12, q12
+	vst1.64		{q12}, [r0,:128]!
+	vswp		d13,d12
+	vshr.s64	q7, q13, #63
+	veor		q13, q13, q6
+	vand		q7, q7, q5
+	vld1.8		{q3}, [r7]!
+	veor		q2, q2, q10
+	vadd.u64	q14, q13, q13
+	vst1.64		{q13}, [r0,:128]!
+	vswp		d15,d14
+	vshr.s64	q6, q14, #63
+	veor		q14, q14, q7
+	vand		q6, q6, q5
+	vld1.8		{q4}, [r7]!
+	veor		q3, q3, q11
+	vadd.u64	q15, q14, q14
+	vst1.64		{q14}, [r0,:128]!
+	vswp		d13,d12
+	vshr.s64	q7, q15, #63
+	veor		q15, q15, q6
+	vand		q7, q7, q5
+	vld1.8		{q5}, [r7]!
+	veor		q4, q4, q12
+	vadd.u64	q8, q15, q15
+	vst1.64		{q15}, [r0,:128]!
+	vswp		d15,d14
+	veor		q8, q8, q7
+	vst1.64		{q8}, [r0,:128]		@ next round tweak
+
+	vld1.8		{q6-q7}, [r7]!
+	veor		q5, q5, q13
+#ifndef	BSAES_ASM_EXTENDED_KEY
+	add		r4, sp, #0x90			@ pass key schedule
+#else
+	add		r4, r10, #248			@ pass key schedule
+#endif
+	veor		q6, q6, q14
+	mov		r5, r1			@ pass rounds
+	veor		q7, q7, q15
+	mov		r0, sp
+
+	bl		_bsaes_decrypt8
+
+	vld1.64		{q8-q9}, [r0,:128]!
+	vld1.64		{q10-q11}, [r0,:128]!
+	veor		q0, q0, q8
+	vld1.64		{q12-q13}, [r0,:128]!
+	veor		q1, q1, q9
+	veor		q8, q6, q10
+	vst1.8		{q0-q1}, [r8]!
+	veor		q9, q4, q11
+	vld1.64		{q14-q15}, [r0,:128]!
+	veor		q10, q2, q12
+	vst1.8		{q8-q9}, [r8]!
+	veor		q11, q7, q13
+	veor		q12, q3, q14
+	vst1.8		{q10-q11}, [r8]!
+	veor		q13, q5, q15
+	vst1.8		{q12-q13}, [r8]!
+
+	vld1.64		{q8}, [r0,:128]		@ next round tweak
+
+	subs		r9, #0x80
+	bpl		.Lxts_dec_loop
+
+.Lxts_dec_short:
+	adds		r9, #0x70
+	bmi		.Lxts_dec_done
+
+	vldmia		r2, {q5}	@ load XTS magic
+	vshr.s64	q7, q8, #63
+	mov		r0, sp
+	vand		q7, q7, q5
+	vadd.u64	q9, q8, q8
+	vst1.64		{q8}, [r0,:128]!
+	vswp		d15,d14
+	vshr.s64	q6, q9, #63
+	veor		q9, q9, q7
+	vand		q6, q6, q5
+	vadd.u64	q10, q9, q9
+	vst1.64		{q9}, [r0,:128]!
+	vswp		d13,d12
+	vshr.s64	q7, q10, #63
+	veor		q10, q10, q6
+	vand		q7, q7, q5
+	vld1.8		{q0}, [r7]!
+	subs		r9, #0x10
+	bmi		.Lxts_dec_1
+	vadd.u64	q11, q10, q10
+	vst1.64		{q10}, [r0,:128]!
+	vswp		d15,d14
+	vshr.s64	q6, q11, #63
+	veor		q11, q11, q7
+	vand		q6, q6, q5
+	vld1.8		{q1}, [r7]!
+	subs		r9, #0x10
+	bmi		.Lxts_dec_2
+	veor		q0, q0, q8
+	vadd.u64	q12, q11, q11
+	vst1.64		{q11}, [r0,:128]!
+	vswp		d13,d12
+	vshr.s64	q7, q12, #63
+	veor		q12, q12, q6
+	vand		q7, q7, q5
+	vld1.8		{q2}, [r7]!
+	subs		r9, #0x10
+	bmi		.Lxts_dec_3
+	veor		q1, q1, q9
+	vadd.u64	q13, q12, q12
+	vst1.64		{q12}, [r0,:128]!
+	vswp		d15,d14
+	vshr.s64	q6, q13, #63
+	veor		q13, q13, q7
+	vand		q6, q6, q5
+	vld1.8		{q3}, [r7]!
+	subs		r9, #0x10
+	bmi		.Lxts_dec_4
+	veor		q2, q2, q10
+	vadd.u64	q14, q13, q13
+	vst1.64		{q13}, [r0,:128]!
+	vswp		d13,d12
+	vshr.s64	q7, q14, #63
+	veor		q14, q14, q6
+	vand		q7, q7, q5
+	vld1.8		{q4}, [r7]!
+	subs		r9, #0x10
+	bmi		.Lxts_dec_5
+	veor		q3, q3, q11
+	vadd.u64	q15, q14, q14
+	vst1.64		{q14}, [r0,:128]!
+	vswp		d15,d14
+	vshr.s64	q6, q15, #63
+	veor		q15, q15, q7
+	vand		q6, q6, q5
+	vld1.8		{q5}, [r7]!
+	subs		r9, #0x10
+	bmi		.Lxts_dec_6
+	veor		q4, q4, q12
+	sub		r9, #0x10
+	vst1.64		{q15}, [r0,:128]		@ next round tweak
+
+	vld1.8		{q6}, [r7]!
+	veor		q5, q5, q13
+#ifndef	BSAES_ASM_EXTENDED_KEY
+	add		r4, sp, #0x90			@ pass key schedule
+#else
+	add		r4, r10, #248			@ pass key schedule
+#endif
+	veor		q6, q6, q14
+	mov		r5, r1			@ pass rounds
+	mov		r0, sp
+
+	bl		_bsaes_decrypt8
+
+	vld1.64		{q8-q9}, [r0,:128]!
+	vld1.64		{q10-q11}, [r0,:128]!
+	veor		q0, q0, q8
+	vld1.64		{q12-q13}, [r0,:128]!
+	veor		q1, q1, q9
+	veor		q8, q6, q10
+	vst1.8		{q0-q1}, [r8]!
+	veor		q9, q4, q11
+	vld1.64		{q14}, [r0,:128]!
+	veor		q10, q2, q12
+	vst1.8		{q8-q9}, [r8]!
+	veor		q11, q7, q13
+	veor		q12, q3, q14
+	vst1.8		{q10-q11}, [r8]!
+	vst1.8		{q12}, [r8]!
+
+	vld1.64		{q8}, [r0,:128]		@ next round tweak
+	b		.Lxts_dec_done
+.align	4
+.Lxts_dec_6:
+	vst1.64		{q14}, [r0,:128]		@ next round tweak
+
+	veor		q4, q4, q12
+#ifndef	BSAES_ASM_EXTENDED_KEY
+	add		r4, sp, #0x90			@ pass key schedule
+#else
+	add		r4, r10, #248			@ pass key schedule
+#endif
+	veor		q5, q5, q13
+	mov		r5, r1			@ pass rounds
+	mov		r0, sp
+
+	bl		_bsaes_decrypt8
+
+	vld1.64		{q8-q9}, [r0,:128]!
+	vld1.64		{q10-q11}, [r0,:128]!
+	veor		q0, q0, q8
+	vld1.64		{q12-q13}, [r0,:128]!
+	veor		q1, q1, q9
+	veor		q8, q6, q10
+	vst1.8		{q0-q1}, [r8]!
+	veor		q9, q4, q11
+	veor		q10, q2, q12
+	vst1.8		{q8-q9}, [r8]!
+	veor		q11, q7, q13
+	vst1.8		{q10-q11}, [r8]!
+
+	vld1.64		{q8}, [r0,:128]		@ next round tweak
+	b		.Lxts_dec_done
+.align	4
+.Lxts_dec_5:
+	vst1.64		{q13}, [r0,:128]		@ next round tweak
+
+	veor		q3, q3, q11
+#ifndef	BSAES_ASM_EXTENDED_KEY
+	add		r4, sp, #0x90			@ pass key schedule
+#else
+	add		r4, r10, #248			@ pass key schedule
+#endif
+	veor		q4, q4, q12
+	mov		r5, r1			@ pass rounds
+	mov		r0, sp
+
+	bl		_bsaes_decrypt8
+
+	vld1.64		{q8-q9}, [r0,:128]!
+	vld1.64		{q10-q11}, [r0,:128]!
+	veor		q0, q0, q8
+	vld1.64		{q12}, [r0,:128]!
+	veor		q1, q1, q9
+	veor		q8, q6, q10
+	vst1.8		{q0-q1}, [r8]!
+	veor		q9, q4, q11
+	veor		q10, q2, q12
+	vst1.8		{q8-q9}, [r8]!
+	vst1.8		{q10}, [r8]!
+
+	vld1.64		{q8}, [r0,:128]		@ next round tweak
+	b		.Lxts_dec_done
+.align	4
+.Lxts_dec_4:
+	vst1.64		{q12}, [r0,:128]		@ next round tweak
+
+	veor		q2, q2, q10
+#ifndef	BSAES_ASM_EXTENDED_KEY
+	add		r4, sp, #0x90			@ pass key schedule
+#else
+	add		r4, r10, #248			@ pass key schedule
+#endif
+	veor		q3, q3, q11
+	mov		r5, r1			@ pass rounds
+	mov		r0, sp
+
+	bl		_bsaes_decrypt8
+
+	vld1.64		{q8-q9}, [r0,:128]!
+	vld1.64		{q10-q11}, [r0,:128]!
+	veor		q0, q0, q8
+	veor		q1, q1, q9
+	veor		q8, q6, q10
+	vst1.8		{q0-q1}, [r8]!
+	veor		q9, q4, q11
+	vst1.8		{q8-q9}, [r8]!
+
+	vld1.64		{q8}, [r0,:128]		@ next round tweak
+	b		.Lxts_dec_done
+.align	4
+.Lxts_dec_3:
+	vst1.64		{q11}, [r0,:128]		@ next round tweak
+
+	veor		q1, q1, q9
+#ifndef	BSAES_ASM_EXTENDED_KEY
+	add		r4, sp, #0x90			@ pass key schedule
+#else
+	add		r4, r10, #248			@ pass key schedule
+#endif
+	veor		q2, q2, q10
+	mov		r5, r1			@ pass rounds
+	mov		r0, sp
+
+	bl		_bsaes_decrypt8
+
+	vld1.64		{q8-q9}, [r0,:128]!
+	vld1.64		{q10}, [r0,:128]!
+	veor		q0, q0, q8
+	veor		q1, q1, q9
+	veor		q8, q6, q10
+	vst1.8		{q0-q1}, [r8]!
+	vst1.8		{q8}, [r8]!
+
+	vld1.64		{q8}, [r0,:128]		@ next round tweak
+	b		.Lxts_dec_done
+.align	4
+.Lxts_dec_2:
+	vst1.64		{q10}, [r0,:128]		@ next round tweak
+
+	veor		q0, q0, q8
+#ifndef	BSAES_ASM_EXTENDED_KEY
+	add		r4, sp, #0x90			@ pass key schedule
+#else
+	add		r4, r10, #248			@ pass key schedule
+#endif
+	veor		q1, q1, q9
+	mov		r5, r1			@ pass rounds
+	mov		r0, sp
+
+	bl		_bsaes_decrypt8
+
+	vld1.64		{q8-q9}, [r0,:128]!
+	veor		q0, q0, q8
+	veor		q1, q1, q9
+	vst1.8		{q0-q1}, [r8]!
+
+	vld1.64		{q8}, [r0,:128]		@ next round tweak
+	b		.Lxts_dec_done
+.align	4
+.Lxts_dec_1:
+	mov		r0, sp
+	veor		q0, q8
+	mov		r1, sp
+	vst1.8		{q0}, [sp,:128]
+	mov		r2, r10
+	mov		r4, r3				@ preserve fp
+	mov		r5, r2			@ preserve magic
+
+	bl		AES_decrypt
+
+	vld1.8		{q0}, [sp,:128]
+	veor		q0, q0, q8
+	vst1.8		{q0}, [r8]!
+	mov		r3, r4
+	mov		r2, r5
+
+	vmov		q8, q9		@ next round tweak
+
+.Lxts_dec_done:
+#ifndef	XTS_CHAIN_TWEAK
+	adds		r9, #0x10
+	beq		.Lxts_dec_ret
+
+	@ calculate one round of extra tweak for the stolen ciphertext
+	vldmia		r2, {q5}
+	vshr.s64	q6, q8, #63
+	vand		q6, q6, q5
+	vadd.u64	q9, q8, q8
+	vswp		d13,d12
+	veor		q9, q9, q6
+
+	@ perform the final decryption with the last tweak value
+	vld1.8		{q0}, [r7]!
+	mov		r0, sp
+	veor		q0, q0, q9
+	mov		r1, sp
+	vst1.8		{q0}, [sp,:128]
+	mov		r2, r10
+	mov		r4, r3			@ preserve fp
+
+	bl		AES_decrypt
+
+	vld1.8		{q0}, [sp,:128]
+	veor		q0, q0, q9
+	vst1.8		{q0}, [r8]
+
+	mov		r6, r8
+.Lxts_dec_steal:
+	ldrb		r1, [r8]
+	ldrb		r0, [r7], #1
+	strb		r1, [r8, #0x10]
+	strb		r0, [r8], #1
+
+	subs		r9, #1
+	bhi		.Lxts_dec_steal
+
+	vld1.8		{q0}, [r6]
+	mov		r0, sp
+	veor		q0, q8
+	mov		r1, sp
+	vst1.8		{q0}, [sp,:128]
+	mov		r2, r10
+
+	bl		AES_decrypt
+
+	vld1.8		{q0}, [sp,:128]
+	veor		q0, q0, q8
+	vst1.8		{q0}, [r6]
+	mov		r3, r4
+#endif
+
+.Lxts_dec_ret:
+	bic		r0, r3, #0xf
+	vmov.i32	q0, #0
+	vmov.i32	q1, #0
+#ifdef	XTS_CHAIN_TWEAK
+	ldr		r1, [r3, #0x20+VFP_ABI_FRAME]	@ chain tweak
+#endif
+.Lxts_dec_bzero:				@ wipe key schedule [if any]
+	vstmia		sp!, {q0-q1}
+	cmp		sp, r0
+	bne		.Lxts_dec_bzero
+
+	mov		sp, r3
+#ifdef	XTS_CHAIN_TWEAK
+	vst1.8		{q8}, [r1]
+#endif
+	VFP_ABI_POP
+	ldmia		sp!, {r4-r10, pc}	@ return
+
+.size	bsaes_xts_decrypt,.-bsaes_xts_decrypt
+#endif
diff --git a/arch/arm/crypto/aesbs-glue.c b/arch/arm/crypto/aesbs-glue.c
new file mode 100644
index 000000000..6d6852986
--- /dev/null
+++ b/arch/arm/crypto/aesbs-glue.c
@@ -0,0 +1,437 @@
+/*
+ * linux/arch/arm/crypto/aesbs-glue.c - glue code for NEON bit sliced AES
+ *
+ * Copyright (C) 2013 Linaro Ltd <ard.biesheuvel@linaro.org>
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2 as
+ * published by the Free Software Foundation.
+ */
+
+#include <asm/neon.h>
+#include <crypto/aes.h>
+#include <crypto/ablk_helper.h>
+#include <crypto/algapi.h>
+#include <linux/module.h>
+
+#include "aes_glue.h"
+
+#define BIT_SLICED_KEY_MAXSIZE	(128 * (AES_MAXNR - 1) + 2 * AES_BLOCK_SIZE)
+
+struct BS_KEY {
+	struct AES_KEY	rk;
+	int		converted;
+	u8 __aligned(8)	bs[BIT_SLICED_KEY_MAXSIZE];
+} __aligned(8);
+
+asmlinkage void bsaes_enc_key_convert(u8 out[], struct AES_KEY const *in);
+asmlinkage void bsaes_dec_key_convert(u8 out[], struct AES_KEY const *in);
+
+asmlinkage void bsaes_cbc_encrypt(u8 const in[], u8 out[], u32 bytes,
+				  struct BS_KEY *key, u8 iv[]);
+
+asmlinkage void bsaes_ctr32_encrypt_blocks(u8 const in[], u8 out[], u32 blocks,
+					   struct BS_KEY *key, u8 const iv[]);
+
+asmlinkage void bsaes_xts_encrypt(u8 const in[], u8 out[], u32 bytes,
+				  struct BS_KEY *key, u8 tweak[]);
+
+asmlinkage void bsaes_xts_decrypt(u8 const in[], u8 out[], u32 bytes,
+				  struct BS_KEY *key, u8 tweak[]);
+
+struct aesbs_cbc_ctx {
+	struct AES_KEY	enc;
+	struct BS_KEY	dec;
+};
+
+struct aesbs_ctr_ctx {
+	struct BS_KEY	enc;
+};
+
+struct aesbs_xts_ctx {
+	struct BS_KEY	enc;
+	struct BS_KEY	dec;
+	struct AES_KEY	twkey;
+};
+
+static int aesbs_cbc_set_key(struct crypto_tfm *tfm, const u8 *in_key,
+			     unsigned int key_len)
+{
+	struct aesbs_cbc_ctx *ctx = crypto_tfm_ctx(tfm);
+	int bits = key_len * 8;
+
+	if (private_AES_set_encrypt_key(in_key, bits, &ctx->enc)) {
+		tfm->crt_flags |= CRYPTO_TFM_RES_BAD_KEY_LEN;
+		return -EINVAL;
+	}
+	ctx->dec.rk = ctx->enc;
+	private_AES_set_decrypt_key(in_key, bits, &ctx->dec.rk);
+	ctx->dec.converted = 0;
+	return 0;
+}
+
+static int aesbs_ctr_set_key(struct crypto_tfm *tfm, const u8 *in_key,
+			     unsigned int key_len)
+{
+	struct aesbs_ctr_ctx *ctx = crypto_tfm_ctx(tfm);
+	int bits = key_len * 8;
+
+	if (private_AES_set_encrypt_key(in_key, bits, &ctx->enc.rk)) {
+		tfm->crt_flags |= CRYPTO_TFM_RES_BAD_KEY_LEN;
+		return -EINVAL;
+	}
+	ctx->enc.converted = 0;
+	return 0;
+}
+
+static int aesbs_xts_set_key(struct crypto_tfm *tfm, const u8 *in_key,
+			     unsigned int key_len)
+{
+	struct aesbs_xts_ctx *ctx = crypto_tfm_ctx(tfm);
+	int bits = key_len * 4;
+
+	if (private_AES_set_encrypt_key(in_key, bits, &ctx->enc.rk)) {
+		tfm->crt_flags |= CRYPTO_TFM_RES_BAD_KEY_LEN;
+		return -EINVAL;
+	}
+	ctx->dec.rk = ctx->enc.rk;
+	private_AES_set_decrypt_key(in_key, bits, &ctx->dec.rk);
+	private_AES_set_encrypt_key(in_key + key_len / 2, bits, &ctx->twkey);
+	ctx->enc.converted = ctx->dec.converted = 0;
+	return 0;
+}
+
+static int aesbs_cbc_encrypt(struct blkcipher_desc *desc,
+			     struct scatterlist *dst,
+			     struct scatterlist *src, unsigned int nbytes)
+{
+	struct aesbs_cbc_ctx *ctx = crypto_blkcipher_ctx(desc->tfm);
+	struct blkcipher_walk walk;
+	int err;
+
+	blkcipher_walk_init(&walk, dst, src, nbytes);
+	err = blkcipher_walk_virt(desc, &walk);
+
+	while (walk.nbytes) {
+		u32 blocks = walk.nbytes / AES_BLOCK_SIZE;
+		u8 *src = walk.src.virt.addr;
+
+		if (walk.dst.virt.addr == walk.src.virt.addr) {
+			u8 *iv = walk.iv;
+
+			do {
+				crypto_xor(src, iv, AES_BLOCK_SIZE);
+				AES_encrypt(src, src, &ctx->enc);
+				iv = src;
+				src += AES_BLOCK_SIZE;
+			} while (--blocks);
+			memcpy(walk.iv, iv, AES_BLOCK_SIZE);
+		} else {
+			u8 *dst = walk.dst.virt.addr;
+
+			do {
+				crypto_xor(walk.iv, src, AES_BLOCK_SIZE);
+				AES_encrypt(walk.iv, dst, &ctx->enc);
+				memcpy(walk.iv, dst, AES_BLOCK_SIZE);
+				src += AES_BLOCK_SIZE;
+				dst += AES_BLOCK_SIZE;
+			} while (--blocks);
+		}
+		err = blkcipher_walk_done(desc, &walk, walk.nbytes % AES_BLOCK_SIZE);
+	}
+	return err;
+}
+
+static int aesbs_cbc_decrypt(struct blkcipher_desc *desc,
+			     struct scatterlist *dst,
+			     struct scatterlist *src, unsigned int nbytes)
+{
+	struct aesbs_cbc_ctx *ctx = crypto_blkcipher_ctx(desc->tfm);
+	struct blkcipher_walk walk;
+	int err;
+
+	blkcipher_walk_init(&walk, dst, src, nbytes);
+	err = blkcipher_walk_virt_block(desc, &walk, 8 * AES_BLOCK_SIZE);
+
+	while ((walk.nbytes / AES_BLOCK_SIZE) >= 8) {
+		kernel_neon_begin();
+		bsaes_cbc_encrypt(walk.src.virt.addr, walk.dst.virt.addr,
+				  walk.nbytes, &ctx->dec, walk.iv);
+		kernel_neon_end();
+		err = blkcipher_walk_done(desc, &walk, walk.nbytes % AES_BLOCK_SIZE);
+	}
+	while (walk.nbytes) {
+		u32 blocks = walk.nbytes / AES_BLOCK_SIZE;
+		u8 *dst = walk.dst.virt.addr;
+		u8 *src = walk.src.virt.addr;
+		u8 bk[2][AES_BLOCK_SIZE];
+		u8 *iv = walk.iv;
+
+		do {
+			if (walk.dst.virt.addr == walk.src.virt.addr)
+				memcpy(bk[blocks & 1], src, AES_BLOCK_SIZE);
+
+			AES_decrypt(src, dst, &ctx->dec.rk);
+			crypto_xor(dst, iv, AES_BLOCK_SIZE);
+
+			if (walk.dst.virt.addr == walk.src.virt.addr)
+				iv = bk[blocks & 1];
+			else
+				iv = src;
+
+			dst += AES_BLOCK_SIZE;
+			src += AES_BLOCK_SIZE;
+		} while (--blocks);
+		err = blkcipher_walk_done(desc, &walk, walk.nbytes % AES_BLOCK_SIZE);
+	}
+	return err;
+}
+
+static void inc_be128_ctr(__be32 ctr[], u32 addend)
+{
+	int i;
+
+	for (i = 3; i >= 0; i--, addend = 1) {
+		u32 n = be32_to_cpu(ctr[i]) + addend;
+
+		ctr[i] = cpu_to_be32(n);
+		if (n >= addend)
+			break;
+	}
+}
+
+static int aesbs_ctr_encrypt(struct blkcipher_desc *desc,
+			     struct scatterlist *dst, struct scatterlist *src,
+			     unsigned int nbytes)
+{
+	struct aesbs_ctr_ctx *ctx = crypto_blkcipher_ctx(desc->tfm);
+	struct blkcipher_walk walk;
+	u32 blocks;
+	int err;
+
+	blkcipher_walk_init(&walk, dst, src, nbytes);
+	err = blkcipher_walk_virt_block(desc, &walk, 8 * AES_BLOCK_SIZE);
+
+	while ((blocks = walk.nbytes / AES_BLOCK_SIZE)) {
+		u32 tail = walk.nbytes % AES_BLOCK_SIZE;
+		__be32 *ctr = (__be32 *)walk.iv;
+		u32 headroom = UINT_MAX - be32_to_cpu(ctr[3]);
+
+		/* avoid 32 bit counter overflow in the NEON code */
+		if (unlikely(headroom < blocks)) {
+			blocks = headroom + 1;
+			tail = walk.nbytes - blocks * AES_BLOCK_SIZE;
+		}
+		kernel_neon_begin();
+		bsaes_ctr32_encrypt_blocks(walk.src.virt.addr,
+					   walk.dst.virt.addr, blocks,
+					   &ctx->enc, walk.iv);
+		kernel_neon_end();
+		inc_be128_ctr(ctr, blocks);
+
+		nbytes -= blocks * AES_BLOCK_SIZE;
+		if (nbytes && nbytes == tail && nbytes <= AES_BLOCK_SIZE)
+			break;
+
+		err = blkcipher_walk_done(desc, &walk, tail);
+	}
+	if (walk.nbytes) {
+		u8 *tdst = walk.dst.virt.addr + blocks * AES_BLOCK_SIZE;
+		u8 *tsrc = walk.src.virt.addr + blocks * AES_BLOCK_SIZE;
+		u8 ks[AES_BLOCK_SIZE];
+
+		AES_encrypt(walk.iv, ks, &ctx->enc.rk);
+		if (tdst != tsrc)
+			memcpy(tdst, tsrc, nbytes);
+		crypto_xor(tdst, ks, nbytes);
+		err = blkcipher_walk_done(desc, &walk, 0);
+	}
+	return err;
+}
+
+static int aesbs_xts_encrypt(struct blkcipher_desc *desc,
+			     struct scatterlist *dst,
+			     struct scatterlist *src, unsigned int nbytes)
+{
+	struct aesbs_xts_ctx *ctx = crypto_blkcipher_ctx(desc->tfm);
+	struct blkcipher_walk walk;
+	int err;
+
+	blkcipher_walk_init(&walk, dst, src, nbytes);
+	err = blkcipher_walk_virt_block(desc, &walk, 8 * AES_BLOCK_SIZE);
+
+	/* generate the initial tweak */
+	AES_encrypt(walk.iv, walk.iv, &ctx->twkey);
+
+	while (walk.nbytes) {
+		kernel_neon_begin();
+		bsaes_xts_encrypt(walk.src.virt.addr, walk.dst.virt.addr,
+				  walk.nbytes, &ctx->enc, walk.iv);
+		kernel_neon_end();
+		err = blkcipher_walk_done(desc, &walk, walk.nbytes % AES_BLOCK_SIZE);
+	}
+	return err;
+}
+
+static int aesbs_xts_decrypt(struct blkcipher_desc *desc,
+			     struct scatterlist *dst,
+			     struct scatterlist *src, unsigned int nbytes)
+{
+	struct aesbs_xts_ctx *ctx = crypto_blkcipher_ctx(desc->tfm);
+	struct blkcipher_walk walk;
+	int err;
+
+	blkcipher_walk_init(&walk, dst, src, nbytes);
+	err = blkcipher_walk_virt_block(desc, &walk, 8 * AES_BLOCK_SIZE);
+
+	/* generate the initial tweak */
+	AES_encrypt(walk.iv, walk.iv, &ctx->twkey);
+
+	while (walk.nbytes) {
+		kernel_neon_begin();
+		bsaes_xts_decrypt(walk.src.virt.addr, walk.dst.virt.addr,
+				  walk.nbytes, &ctx->dec, walk.iv);
+		kernel_neon_end();
+		err = blkcipher_walk_done(desc, &walk, walk.nbytes % AES_BLOCK_SIZE);
+	}
+	return err;
+}
+
+static struct crypto_alg aesbs_algs[] = { {
+	.cra_name		= "__cbc-aes-neonbs",
+	.cra_driver_name	= "__driver-cbc-aes-neonbs",
+	.cra_priority		= 0,
+	.cra_flags		= CRYPTO_ALG_TYPE_BLKCIPHER |
+				  CRYPTO_ALG_INTERNAL,
+	.cra_blocksize		= AES_BLOCK_SIZE,
+	.cra_ctxsize		= sizeof(struct aesbs_cbc_ctx),
+	.cra_alignmask		= 7,
+	.cra_type		= &crypto_blkcipher_type,
+	.cra_module		= THIS_MODULE,
+	.cra_blkcipher = {
+		.min_keysize	= AES_MIN_KEY_SIZE,
+		.max_keysize	= AES_MAX_KEY_SIZE,
+		.ivsize		= AES_BLOCK_SIZE,
+		.setkey		= aesbs_cbc_set_key,
+		.encrypt	= aesbs_cbc_encrypt,
+		.decrypt	= aesbs_cbc_decrypt,
+	},
+}, {
+	.cra_name		= "__ctr-aes-neonbs",
+	.cra_driver_name	= "__driver-ctr-aes-neonbs",
+	.cra_priority		= 0,
+	.cra_flags		= CRYPTO_ALG_TYPE_BLKCIPHER |
+				  CRYPTO_ALG_INTERNAL,
+	.cra_blocksize		= 1,
+	.cra_ctxsize		= sizeof(struct aesbs_ctr_ctx),
+	.cra_alignmask		= 7,
+	.cra_type		= &crypto_blkcipher_type,
+	.cra_module		= THIS_MODULE,
+	.cra_blkcipher = {
+		.min_keysize	= AES_MIN_KEY_SIZE,
+		.max_keysize	= AES_MAX_KEY_SIZE,
+		.ivsize		= AES_BLOCK_SIZE,
+		.setkey		= aesbs_ctr_set_key,
+		.encrypt	= aesbs_ctr_encrypt,
+		.decrypt	= aesbs_ctr_encrypt,
+	},
+}, {
+	.cra_name		= "__xts-aes-neonbs",
+	.cra_driver_name	= "__driver-xts-aes-neonbs",
+	.cra_priority		= 0,
+	.cra_flags		= CRYPTO_ALG_TYPE_BLKCIPHER |
+				  CRYPTO_ALG_INTERNAL,
+	.cra_blocksize		= AES_BLOCK_SIZE,
+	.cra_ctxsize		= sizeof(struct aesbs_xts_ctx),
+	.cra_alignmask		= 7,
+	.cra_type		= &crypto_blkcipher_type,
+	.cra_module		= THIS_MODULE,
+	.cra_blkcipher = {
+		.min_keysize	= 2 * AES_MIN_KEY_SIZE,
+		.max_keysize	= 2 * AES_MAX_KEY_SIZE,
+		.ivsize		= AES_BLOCK_SIZE,
+		.setkey		= aesbs_xts_set_key,
+		.encrypt	= aesbs_xts_encrypt,
+		.decrypt	= aesbs_xts_decrypt,
+	},
+}, {
+	.cra_name		= "cbc(aes)",
+	.cra_driver_name	= "cbc-aes-neonbs",
+	.cra_priority		= 300,
+	.cra_flags		= CRYPTO_ALG_TYPE_ABLKCIPHER|CRYPTO_ALG_ASYNC,
+	.cra_blocksize		= AES_BLOCK_SIZE,
+	.cra_ctxsize		= sizeof(struct async_helper_ctx),
+	.cra_alignmask		= 7,
+	.cra_type		= &crypto_ablkcipher_type,
+	.cra_module		= THIS_MODULE,
+	.cra_init		= ablk_init,
+	.cra_exit		= ablk_exit,
+	.cra_ablkcipher = {
+		.min_keysize	= AES_MIN_KEY_SIZE,
+		.max_keysize	= AES_MAX_KEY_SIZE,
+		.ivsize		= AES_BLOCK_SIZE,
+		.setkey		= ablk_set_key,
+		.encrypt	= __ablk_encrypt,
+		.decrypt	= ablk_decrypt,
+	}
+}, {
+	.cra_name		= "ctr(aes)",
+	.cra_driver_name	= "ctr-aes-neonbs",
+	.cra_priority		= 300,
+	.cra_flags		= CRYPTO_ALG_TYPE_ABLKCIPHER|CRYPTO_ALG_ASYNC,
+	.cra_blocksize		= 1,
+	.cra_ctxsize		= sizeof(struct async_helper_ctx),
+	.cra_alignmask		= 7,
+	.cra_type		= &crypto_ablkcipher_type,
+	.cra_module		= THIS_MODULE,
+	.cra_init		= ablk_init,
+	.cra_exit		= ablk_exit,
+	.cra_ablkcipher = {
+		.min_keysize	= AES_MIN_KEY_SIZE,
+		.max_keysize	= AES_MAX_KEY_SIZE,
+		.ivsize		= AES_BLOCK_SIZE,
+		.setkey		= ablk_set_key,
+		.encrypt	= ablk_encrypt,
+		.decrypt	= ablk_decrypt,
+	}
+}, {
+	.cra_name		= "xts(aes)",
+	.cra_driver_name	= "xts-aes-neonbs",
+	.cra_priority		= 300,
+	.cra_flags		= CRYPTO_ALG_TYPE_ABLKCIPHER|CRYPTO_ALG_ASYNC,
+	.cra_blocksize		= AES_BLOCK_SIZE,
+	.cra_ctxsize		= sizeof(struct async_helper_ctx),
+	.cra_alignmask		= 7,
+	.cra_type		= &crypto_ablkcipher_type,
+	.cra_module		= THIS_MODULE,
+	.cra_init		= ablk_init,
+	.cra_exit		= ablk_exit,
+	.cra_ablkcipher = {
+		.min_keysize	= 2 * AES_MIN_KEY_SIZE,
+		.max_keysize	= 2 * AES_MAX_KEY_SIZE,
+		.ivsize		= AES_BLOCK_SIZE,
+		.setkey		= ablk_set_key,
+		.encrypt	= ablk_encrypt,
+		.decrypt	= ablk_decrypt,
+	}
+} };
+
+static int __init aesbs_mod_init(void)
+{
+	if (!cpu_has_neon())
+		return -ENODEV;
+
+	return crypto_register_algs(aesbs_algs, ARRAY_SIZE(aesbs_algs));
+}
+
+static void __exit aesbs_mod_exit(void)
+{
+	crypto_unregister_algs(aesbs_algs, ARRAY_SIZE(aesbs_algs));
+}
+
+module_init(aesbs_mod_init);
+module_exit(aesbs_mod_exit);
+
+MODULE_DESCRIPTION("Bit sliced AES in CBC/CTR/XTS modes using NEON");
+MODULE_AUTHOR("Ard Biesheuvel <ard.biesheuvel@linaro.org>");
+MODULE_LICENSE("GPL");
diff --git a/arch/arm/crypto/bsaes-armv7.pl b/arch/arm/crypto/bsaes-armv7.pl
new file mode 100644
index 000000000..a4d3856e7
--- /dev/null
+++ b/arch/arm/crypto/bsaes-armv7.pl
@@ -0,0 +1,2471 @@
+#!/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__
+# define __ARM_MAX_ARCH__ 7
+#endif
+
+#ifdef __thumb__
+# define adrl adr
+#endif
+
+#if __ARM_MAX_ARCH__>=7
+.arch	armv7-a
+.fpu	neon
+
+.text
+.syntax	unified 	@ ARMv7-capable assembler is expected to handle this
+#ifdef __thumb2__
+.thumb
+#else
+.code   32
+#endif
+
+.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
+
+#ifndef	XTS_CHAIN_TWEAK
+	tst	$len, #0xf			@ if not multiple of 16
+	it	ne				@ Thumb2 thing, sanity check in ARM
+	subne	$len, #0x10			@ subtract another 16 bytes
+#endif
+	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, $fp, #0xf
+	vmov.i32	q0, #0
+	vmov.i32	q1, #0
+#ifdef	XTS_CHAIN_TWEAK
+	ldr		r1, [$fp, #0x20+VFP_ABI_FRAME]	@ chain tweak
+#endif
+.Lxts_dec_bzero:				@ wipe key schedule [if any]
+	vstmia		sp!, {q0-q1}
+	cmp		sp, r0
+	bne		.Lxts_dec_bzero
+
+	mov		sp, $fp
+#ifdef	XTS_CHAIN_TWEAK
+	vst1.8		{@XMM[8]}, [r1]
+#endif
+	VFP_ABI_POP
+	ldmia		sp!, {r4-r10, pc}	@ return
+
+.size	bsaes_xts_decrypt,.-bsaes_xts_decrypt
+___
+}
+$code.=<<___;
+#endif
+___
+
+$code =~ s/\`([^\`]*)\`/eval($1)/gem;
+
+open SELF,$0;
+while(<SELF>) {
+	next if (/^#!/);
+        last if (!s/^#/@/ and !/^$/);
+        print;
+}
+close SELF;
+
+print $code;
+
+close STDOUT;
diff --git a/arch/arm/crypto/ghash-ce-core.S b/arch/arm/crypto/ghash-ce-core.S
new file mode 100644
index 000000000..f6ab8bcc9
--- /dev/null
+++ b/arch/arm/crypto/ghash-ce-core.S
@@ -0,0 +1,94 @@
+/*
+ * Accelerated GHASH implementation with ARMv8 vmull.p64 instructions.
+ *
+ * Copyright (C) 2015 Linaro Ltd. <ard.biesheuvel@linaro.org>
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License version 2 as published
+ * by the Free Software Foundation.
+ */
+
+#include <linux/linkage.h>
+#include <asm/assembler.h>
+
+	SHASH		.req	q0
+	SHASH2		.req	q1
+	T1		.req	q2
+	T2		.req	q3
+	MASK		.req	q4
+	XL		.req	q5
+	XM		.req	q6
+	XH		.req	q7
+	IN1		.req	q7
+
+	SHASH_L		.req	d0
+	SHASH_H		.req	d1
+	SHASH2_L	.req	d2
+	T1_L		.req	d4
+	MASK_L		.req	d8
+	XL_L		.req	d10
+	XL_H		.req	d11
+	XM_L		.req	d12
+	XM_H		.req	d13
+	XH_L		.req	d14
+
+	.text
+	.fpu		crypto-neon-fp-armv8
+
+	/*
+	 * void pmull_ghash_update(int blocks, u64 dg[], const char *src,
+	 *			   struct ghash_key const *k, const char *head)
+	 */
+ENTRY(pmull_ghash_update)
+	vld1.64		{SHASH}, [r3]
+	vld1.64		{XL}, [r1]
+	vmov.i8		MASK, #0xe1
+	vext.8		SHASH2, SHASH, SHASH, #8
+	vshl.u64	MASK, MASK, #57
+	veor		SHASH2, SHASH2, SHASH
+
+	/* do the head block first, if supplied */
+	ldr		ip, [sp]
+	teq		ip, #0
+	beq		0f
+	vld1.64		{T1}, [ip]
+	teq		r0, #0
+	b		1f
+
+0:	vld1.64		{T1}, [r2]!
+	subs		r0, r0, #1
+
+1:	/* multiply XL by SHASH in GF(2^128) */
+#ifndef CONFIG_CPU_BIG_ENDIAN
+	vrev64.8	T1, T1
+#endif
+	vext.8		T2, XL, XL, #8
+	vext.8		IN1, T1, T1, #8
+	veor		T1, T1, T2
+	veor		XL, XL, IN1
+
+	vmull.p64	XH, SHASH_H, XL_H		@ a1 * b1
+	veor		T1, T1, XL
+	vmull.p64	XL, SHASH_L, XL_L		@ a0 * b0
+	vmull.p64	XM, SHASH2_L, T1_L		@ (a1 + a0)(b1 + b0)
+
+	vext.8		T1, XL, XH, #8
+	veor		T2, XL, XH
+	veor		XM, XM, T1
+	veor		XM, XM, T2
+	vmull.p64	T2, XL_L, MASK_L
+
+	vmov		XH_L, XM_H
+	vmov		XM_H, XL_L
+
+	veor		XL, XM, T2
+	vext.8		T2, XL, XL, #8
+	vmull.p64	XL, XL_L, MASK_L
+	veor		T2, T2, XH
+	veor		XL, XL, T2
+
+	bne		0b
+
+	vst1.64		{XL}, [r1]
+	bx		lr
+ENDPROC(pmull_ghash_update)
diff --git a/arch/arm/crypto/ghash-ce-glue.c b/arch/arm/crypto/ghash-ce-glue.c
new file mode 100644
index 000000000..03a39fe29
--- /dev/null
+++ b/arch/arm/crypto/ghash-ce-glue.c
@@ -0,0 +1,320 @@
+/*
+ * Accelerated GHASH implementation with ARMv8 vmull.p64 instructions.
+ *
+ * Copyright (C) 2015 Linaro Ltd. <ard.biesheuvel@linaro.org>
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License version 2 as published
+ * by the Free Software Foundation.
+ */
+
+#include <asm/hwcap.h>
+#include <asm/neon.h>
+#include <asm/simd.h>
+#include <asm/unaligned.h>
+#include <crypto/cryptd.h>
+#include <crypto/internal/hash.h>
+#include <crypto/gf128mul.h>
+#include <linux/crypto.h>
+#include <linux/module.h>
+
+MODULE_DESCRIPTION("GHASH secure hash using ARMv8 Crypto Extensions");
+MODULE_AUTHOR("Ard Biesheuvel <ard.biesheuvel@linaro.org>");
+MODULE_LICENSE("GPL v2");
+
+#define GHASH_BLOCK_SIZE	16
+#define GHASH_DIGEST_SIZE	16
+
+struct ghash_key {
+	u64	a;
+	u64	b;
+};
+
+struct ghash_desc_ctx {
+	u64 digest[GHASH_DIGEST_SIZE/sizeof(u64)];
+	u8 buf[GHASH_BLOCK_SIZE];
+	u32 count;
+};
+
+struct ghash_async_ctx {
+	struct cryptd_ahash *cryptd_tfm;
+};
+
+asmlinkage void pmull_ghash_update(int blocks, u64 dg[], const char *src,
+				   struct ghash_key const *k, const char *head);
+
+static int ghash_init(struct shash_desc *desc)
+{
+	struct ghash_desc_ctx *ctx = shash_desc_ctx(desc);
+
+	*ctx = (struct ghash_desc_ctx){};
+	return 0;
+}
+
+static int ghash_update(struct shash_desc *desc, const u8 *src,
+			unsigned int len)
+{
+	struct ghash_desc_ctx *ctx = shash_desc_ctx(desc);
+	unsigned int partial = ctx->count % GHASH_BLOCK_SIZE;
+
+	ctx->count += len;
+
+	if ((partial + len) >= GHASH_BLOCK_SIZE) {
+		struct ghash_key *key = crypto_shash_ctx(desc->tfm);
+		int blocks;
+
+		if (partial) {
+			int p = GHASH_BLOCK_SIZE - partial;
+
+			memcpy(ctx->buf + partial, src, p);
+			src += p;
+			len -= p;
+		}
+
+		blocks = len / GHASH_BLOCK_SIZE;
+		len %= GHASH_BLOCK_SIZE;
+
+		kernel_neon_begin();
+		pmull_ghash_update(blocks, ctx->digest, src, key,
+				   partial ? ctx->buf : NULL);
+		kernel_neon_end();
+		src += blocks * GHASH_BLOCK_SIZE;
+		partial = 0;
+	}
+	if (len)
+		memcpy(ctx->buf + partial, src, len);
+	return 0;
+}
+
+static int ghash_final(struct shash_desc *desc, u8 *dst)
+{
+	struct ghash_desc_ctx *ctx = shash_desc_ctx(desc);
+	unsigned int partial = ctx->count % GHASH_BLOCK_SIZE;
+
+	if (partial) {
+		struct ghash_key *key = crypto_shash_ctx(desc->tfm);
+
+		memset(ctx->buf + partial, 0, GHASH_BLOCK_SIZE - partial);
+		kernel_neon_begin();
+		pmull_ghash_update(1, ctx->digest, ctx->buf, key, NULL);
+		kernel_neon_end();
+	}
+	put_unaligned_be64(ctx->digest[1], dst);
+	put_unaligned_be64(ctx->digest[0], dst + 8);
+
+	*ctx = (struct ghash_desc_ctx){};
+	return 0;
+}
+
+static int ghash_setkey(struct crypto_shash *tfm,
+			const u8 *inkey, unsigned int keylen)
+{
+	struct ghash_key *key = crypto_shash_ctx(tfm);
+	u64 a, b;
+
+	if (keylen != GHASH_BLOCK_SIZE) {
+		crypto_shash_set_flags(tfm, CRYPTO_TFM_RES_BAD_KEY_LEN);
+		return -EINVAL;
+	}
+
+	/* perform multiplication by 'x' in GF(2^128) */
+	b = get_unaligned_be64(inkey);
+	a = get_unaligned_be64(inkey + 8);
+
+	key->a = (a << 1) | (b >> 63);
+	key->b = (b << 1) | (a >> 63);
+
+	if (b >> 63)
+		key->b ^= 0xc200000000000000UL;
+
+	return 0;
+}
+
+static struct shash_alg ghash_alg = {
+	.digestsize		= GHASH_DIGEST_SIZE,
+	.init			= ghash_init,
+	.update			= ghash_update,
+	.final			= ghash_final,
+	.setkey			= ghash_setkey,
+	.descsize		= sizeof(struct ghash_desc_ctx),
+	.base			= {
+		.cra_name	= "ghash",
+		.cra_driver_name = "__driver-ghash-ce",
+		.cra_priority	= 0,
+		.cra_flags	= CRYPTO_ALG_TYPE_SHASH | CRYPTO_ALG_INTERNAL,
+		.cra_blocksize	= GHASH_BLOCK_SIZE,
+		.cra_ctxsize	= sizeof(struct ghash_key),
+		.cra_module	= THIS_MODULE,
+	},
+};
+
+static int ghash_async_init(struct ahash_request *req)
+{
+	struct crypto_ahash *tfm = crypto_ahash_reqtfm(req);
+	struct ghash_async_ctx *ctx = crypto_ahash_ctx(tfm);
+	struct ahash_request *cryptd_req = ahash_request_ctx(req);
+	struct cryptd_ahash *cryptd_tfm = ctx->cryptd_tfm;
+
+	if (!may_use_simd()) {
+		memcpy(cryptd_req, req, sizeof(*req));
+		ahash_request_set_tfm(cryptd_req, &cryptd_tfm->base);
+		return crypto_ahash_init(cryptd_req);
+	} else {
+		struct shash_desc *desc = cryptd_shash_desc(cryptd_req);
+		struct crypto_shash *child = cryptd_ahash_child(cryptd_tfm);
+
+		desc->tfm = child;
+		desc->flags = req->base.flags;
+		return crypto_shash_init(desc);
+	}
+}
+
+static int ghash_async_update(struct ahash_request *req)
+{
+	struct ahash_request *cryptd_req = ahash_request_ctx(req);
+
+	if (!may_use_simd()) {
+		struct crypto_ahash *tfm = crypto_ahash_reqtfm(req);
+		struct ghash_async_ctx *ctx = crypto_ahash_ctx(tfm);
+		struct cryptd_ahash *cryptd_tfm = ctx->cryptd_tfm;
+
+		memcpy(cryptd_req, req, sizeof(*req));
+		ahash_request_set_tfm(cryptd_req, &cryptd_tfm->base);
+		return crypto_ahash_update(cryptd_req);
+	} else {
+		struct shash_desc *desc = cryptd_shash_desc(cryptd_req);
+		return shash_ahash_update(req, desc);
+	}
+}
+
+static int ghash_async_final(struct ahash_request *req)
+{
+	struct ahash_request *cryptd_req = ahash_request_ctx(req);
+
+	if (!may_use_simd()) {
+		struct crypto_ahash *tfm = crypto_ahash_reqtfm(req);
+		struct ghash_async_ctx *ctx = crypto_ahash_ctx(tfm);
+		struct cryptd_ahash *cryptd_tfm = ctx->cryptd_tfm;
+
+		memcpy(cryptd_req, req, sizeof(*req));
+		ahash_request_set_tfm(cryptd_req, &cryptd_tfm->base);
+		return crypto_ahash_final(cryptd_req);
+	} else {
+		struct shash_desc *desc = cryptd_shash_desc(cryptd_req);
+		return crypto_shash_final(desc, req->result);
+	}
+}
+
+static int ghash_async_digest(struct ahash_request *req)
+{
+	struct crypto_ahash *tfm = crypto_ahash_reqtfm(req);
+	struct ghash_async_ctx *ctx = crypto_ahash_ctx(tfm);
+	struct ahash_request *cryptd_req = ahash_request_ctx(req);
+	struct cryptd_ahash *cryptd_tfm = ctx->cryptd_tfm;
+
+	if (!may_use_simd()) {
+		memcpy(cryptd_req, req, sizeof(*req));
+		ahash_request_set_tfm(cryptd_req, &cryptd_tfm->base);
+		return crypto_ahash_digest(cryptd_req);
+	} else {
+		struct shash_desc *desc = cryptd_shash_desc(cryptd_req);
+		struct crypto_shash *child = cryptd_ahash_child(cryptd_tfm);
+
+		desc->tfm = child;
+		desc->flags = req->base.flags;
+		return shash_ahash_digest(req, desc);
+	}
+}
+
+static int ghash_async_setkey(struct crypto_ahash *tfm, const u8 *key,
+			      unsigned int keylen)
+{
+	struct ghash_async_ctx *ctx = crypto_ahash_ctx(tfm);
+	struct crypto_ahash *child = &ctx->cryptd_tfm->base;
+	int err;
+
+	crypto_ahash_clear_flags(child, CRYPTO_TFM_REQ_MASK);
+	crypto_ahash_set_flags(child, crypto_ahash_get_flags(tfm)
+			       & CRYPTO_TFM_REQ_MASK);
+	err = crypto_ahash_setkey(child, key, keylen);
+	crypto_ahash_set_flags(tfm, crypto_ahash_get_flags(child)
+			       & CRYPTO_TFM_RES_MASK);
+
+	return err;
+}
+
+static int ghash_async_init_tfm(struct crypto_tfm *tfm)
+{
+	struct cryptd_ahash *cryptd_tfm;
+	struct ghash_async_ctx *ctx = crypto_tfm_ctx(tfm);
+
+	cryptd_tfm = cryptd_alloc_ahash("__driver-ghash-ce",
+					CRYPTO_ALG_INTERNAL,
+					CRYPTO_ALG_INTERNAL);
+	if (IS_ERR(cryptd_tfm))
+		return PTR_ERR(cryptd_tfm);
+	ctx->cryptd_tfm = cryptd_tfm;
+	crypto_ahash_set_reqsize(__crypto_ahash_cast(tfm),
+				 sizeof(struct ahash_request) +
+				 crypto_ahash_reqsize(&cryptd_tfm->base));
+
+	return 0;
+}
+
+static void ghash_async_exit_tfm(struct crypto_tfm *tfm)
+{
+	struct ghash_async_ctx *ctx = crypto_tfm_ctx(tfm);
+
+	cryptd_free_ahash(ctx->cryptd_tfm);
+}
+
+static struct ahash_alg ghash_async_alg = {
+	.init			= ghash_async_init,
+	.update			= ghash_async_update,
+	.final			= ghash_async_final,
+	.setkey			= ghash_async_setkey,
+	.digest			= ghash_async_digest,
+	.halg.digestsize	= GHASH_DIGEST_SIZE,
+	.halg.base		= {
+		.cra_name	= "ghash",
+		.cra_driver_name = "ghash-ce",
+		.cra_priority	= 300,
+		.cra_flags	= CRYPTO_ALG_TYPE_AHASH | CRYPTO_ALG_ASYNC,
+		.cra_blocksize	= GHASH_BLOCK_SIZE,
+		.cra_type	= &crypto_ahash_type,
+		.cra_ctxsize	= sizeof(struct ghash_async_ctx),
+		.cra_module	= THIS_MODULE,
+		.cra_init	= ghash_async_init_tfm,
+		.cra_exit	= ghash_async_exit_tfm,
+	},
+};
+
+static int __init ghash_ce_mod_init(void)
+{
+	int err;
+
+	if (!(elf_hwcap2 & HWCAP2_PMULL))
+		return -ENODEV;
+
+	err = crypto_register_shash(&ghash_alg);
+	if (err)
+		return err;
+	err = crypto_register_ahash(&ghash_async_alg);
+	if (err)
+		goto err_shash;
+
+	return 0;
+
+err_shash:
+	crypto_unregister_shash(&ghash_alg);
+	return err;
+}
+
+static void __exit ghash_ce_mod_exit(void)
+{
+	crypto_unregister_ahash(&ghash_async_alg);
+	crypto_unregister_shash(&ghash_alg);
+}
+
+module_init(ghash_ce_mod_init);
+module_exit(ghash_ce_mod_exit);
diff --git a/arch/arm/crypto/sha1-armv4-large.S b/arch/arm/crypto/sha1-armv4-large.S
new file mode 100644
index 000000000..99207c45e
--- /dev/null
+++ b/arch/arm/crypto/sha1-armv4-large.S
@@ -0,0 +1,497 @@
+#define __ARM_ARCH__ __LINUX_ARM_ARCH__
+@ ====================================================================
+@ Written by Andy Polyakov <appro@fy.chalmers.se> 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/.
+@ ====================================================================
+
+@ sha1_block procedure for ARMv4.
+@
+@ January 2007.
+
+@ Size/performance trade-off
+@ ====================================================================
+@ impl		size in bytes	comp cycles[*]	measured performance
+@ ====================================================================
+@ thumb		304		3212		4420
+@ armv4-small	392/+29%	1958/+64%	2250/+96%
+@ armv4-compact	740/+89%	1552/+26%	1840/+22%
+@ armv4-large	1420/+92%	1307/+19%	1370/+34%[***]
+@ full unroll	~5100/+260%	~1260/+4%	~1300/+5%
+@ ====================================================================
+@ thumb		= same as 'small' but in Thumb instructions[**] and
+@		  with recurring code in two private functions;
+@ small		= detached Xload/update, loops are folded;
+@ compact	= detached Xload/update, 5x unroll;
+@ large		= interleaved Xload/update, 5x unroll;
+@ full unroll	= interleaved Xload/update, full unroll, estimated[!];
+@
+@ [*]	Manually counted instructions in "grand" loop body. Measured
+@	performance is affected by prologue and epilogue overhead,
+@	i-cache availability, branch penalties, etc.
+@ [**]	While each Thumb instruction is twice smaller, they are not as
+@	diverse as ARM ones: e.g., there are only two arithmetic
+@	instructions with 3 arguments, no [fixed] rotate, addressing
+@	modes are limited. As result it takes more instructions to do
+@	the same job in Thumb, therefore the code is never twice as
+@	small and always slower.
+@ [***]	which is also ~35% better than compiler generated code. Dual-
+@	issue Cortex A8 core was measured to process input block in
+@	~990 cycles.
+
+@ August 2010.
+@
+@ Rescheduling for dual-issue pipeline resulted in 13% improvement on
+@ Cortex A8 core and in absolute terms ~870 cycles per input block
+@ [or 13.6 cycles per byte].
+
+@ February 2011.
+@
+@ Profiler-assisted and platform-specific optimization resulted in 10%
+@ improvement on Cortex A8 core and 12.2 cycles per byte.
+
+#include <linux/linkage.h>
+
+.text
+
+.align	2
+ENTRY(sha1_block_data_order)
+	stmdb	sp!,{r4-r12,lr}
+	add	r2,r1,r2,lsl#6	@ r2 to point at the end of r1
+	ldmia	r0,{r3,r4,r5,r6,r7}
+.Lloop:
+	ldr	r8,.LK_00_19
+	mov	r14,sp
+	sub	sp,sp,#15*4
+	mov	r5,r5,ror#30
+	mov	r6,r6,ror#30
+	mov	r7,r7,ror#30		@ [6]
+.L_00_15:
+#if __ARM_ARCH__<7
+	ldrb	r10,[r1,#2]
+	ldrb	r9,[r1,#3]
+	ldrb	r11,[r1,#1]
+	add	r7,r8,r7,ror#2			@ E+=K_00_19
+	ldrb	r12,[r1],#4
+	orr	r9,r9,r10,lsl#8
+	eor	r10,r5,r6			@ F_xx_xx
+	orr	r9,r9,r11,lsl#16
+	add	r7,r7,r3,ror#27			@ E+=ROR(A,27)
+	orr	r9,r9,r12,lsl#24
+#else
+	ldr	r9,[r1],#4			@ handles unaligned
+	add	r7,r8,r7,ror#2			@ E+=K_00_19
+	eor	r10,r5,r6			@ F_xx_xx
+	add	r7,r7,r3,ror#27			@ E+=ROR(A,27)
+#ifdef __ARMEL__
+	rev	r9,r9				@ byte swap
+#endif
+#endif
+	and	r10,r4,r10,ror#2
+	add	r7,r7,r9			@ E+=X[i]
+	eor	r10,r10,r6,ror#2		@ F_00_19(B,C,D)
+	str	r9,[r14,#-4]!
+	add	r7,r7,r10			@ E+=F_00_19(B,C,D)
+#if __ARM_ARCH__<7
+	ldrb	r10,[r1,#2]
+	ldrb	r9,[r1,#3]
+	ldrb	r11,[r1,#1]
+	add	r6,r8,r6,ror#2			@ E+=K_00_19
+	ldrb	r12,[r1],#4
+	orr	r9,r9,r10,lsl#8
+	eor	r10,r4,r5			@ F_xx_xx
+	orr	r9,r9,r11,lsl#16
+	add	r6,r6,r7,ror#27			@ E+=ROR(A,27)
+	orr	r9,r9,r12,lsl#24
+#else
+	ldr	r9,[r1],#4			@ handles unaligned
+	add	r6,r8,r6,ror#2			@ E+=K_00_19
+	eor	r10,r4,r5			@ F_xx_xx
+	add	r6,r6,r7,ror#27			@ E+=ROR(A,27)
+#ifdef __ARMEL__
+	rev	r9,r9				@ byte swap
+#endif
+#endif
+	and	r10,r3,r10,ror#2
+	add	r6,r6,r9			@ E+=X[i]
+	eor	r10,r10,r5,ror#2		@ F_00_19(B,C,D)
+	str	r9,[r14,#-4]!
+	add	r6,r6,r10			@ E+=F_00_19(B,C,D)
+#if __ARM_ARCH__<7
+	ldrb	r10,[r1,#2]
+	ldrb	r9,[r1,#3]
+	ldrb	r11,[r1,#1]
+	add	r5,r8,r5,ror#2			@ E+=K_00_19
+	ldrb	r12,[r1],#4
+	orr	r9,r9,r10,lsl#8
+	eor	r10,r3,r4			@ F_xx_xx
+	orr	r9,r9,r11,lsl#16
+	add	r5,r5,r6,ror#27			@ E+=ROR(A,27)
+	orr	r9,r9,r12,lsl#24
+#else
+	ldr	r9,[r1],#4			@ handles unaligned
+	add	r5,r8,r5,ror#2			@ E+=K_00_19
+	eor	r10,r3,r4			@ F_xx_xx
+	add	r5,r5,r6,ror#27			@ E+=ROR(A,27)
+#ifdef __ARMEL__
+	rev	r9,r9				@ byte swap
+#endif
+#endif
+	and	r10,r7,r10,ror#2
+	add	r5,r5,r9			@ E+=X[i]
+	eor	r10,r10,r4,ror#2		@ F_00_19(B,C,D)
+	str	r9,[r14,#-4]!
+	add	r5,r5,r10			@ E+=F_00_19(B,C,D)
+#if __ARM_ARCH__<7
+	ldrb	r10,[r1,#2]
+	ldrb	r9,[r1,#3]
+	ldrb	r11,[r1,#1]
+	add	r4,r8,r4,ror#2			@ E+=K_00_19
+	ldrb	r12,[r1],#4
+	orr	r9,r9,r10,lsl#8
+	eor	r10,r7,r3			@ F_xx_xx
+	orr	r9,r9,r11,lsl#16
+	add	r4,r4,r5,ror#27			@ E+=ROR(A,27)
+	orr	r9,r9,r12,lsl#24
+#else
+	ldr	r9,[r1],#4			@ handles unaligned
+	add	r4,r8,r4,ror#2			@ E+=K_00_19
+	eor	r10,r7,r3			@ F_xx_xx
+	add	r4,r4,r5,ror#27			@ E+=ROR(A,27)
+#ifdef __ARMEL__
+	rev	r9,r9				@ byte swap
+#endif
+#endif
+	and	r10,r6,r10,ror#2
+	add	r4,r4,r9			@ E+=X[i]
+	eor	r10,r10,r3,ror#2		@ F_00_19(B,C,D)
+	str	r9,[r14,#-4]!
+	add	r4,r4,r10			@ E+=F_00_19(B,C,D)
+#if __ARM_ARCH__<7
+	ldrb	r10,[r1,#2]
+	ldrb	r9,[r1,#3]
+	ldrb	r11,[r1,#1]
+	add	r3,r8,r3,ror#2			@ E+=K_00_19
+	ldrb	r12,[r1],#4
+	orr	r9,r9,r10,lsl#8
+	eor	r10,r6,r7			@ F_xx_xx
+	orr	r9,r9,r11,lsl#16
+	add	r3,r3,r4,ror#27			@ E+=ROR(A,27)
+	orr	r9,r9,r12,lsl#24
+#else
+	ldr	r9,[r1],#4			@ handles unaligned
+	add	r3,r8,r3,ror#2			@ E+=K_00_19
+	eor	r10,r6,r7			@ F_xx_xx
+	add	r3,r3,r4,ror#27			@ E+=ROR(A,27)
+#ifdef __ARMEL__
+	rev	r9,r9				@ byte swap
+#endif
+#endif
+	and	r10,r5,r10,ror#2
+	add	r3,r3,r9			@ E+=X[i]
+	eor	r10,r10,r7,ror#2		@ F_00_19(B,C,D)
+	str	r9,[r14,#-4]!
+	add	r3,r3,r10			@ E+=F_00_19(B,C,D)
+	cmp	r14,sp
+	bne	.L_00_15		@ [((11+4)*5+2)*3]
+	sub	sp,sp,#25*4
+#if __ARM_ARCH__<7
+	ldrb	r10,[r1,#2]
+	ldrb	r9,[r1,#3]
+	ldrb	r11,[r1,#1]
+	add	r7,r8,r7,ror#2			@ E+=K_00_19
+	ldrb	r12,[r1],#4
+	orr	r9,r9,r10,lsl#8
+	eor	r10,r5,r6			@ F_xx_xx
+	orr	r9,r9,r11,lsl#16
+	add	r7,r7,r3,ror#27			@ E+=ROR(A,27)
+	orr	r9,r9,r12,lsl#24
+#else
+	ldr	r9,[r1],#4			@ handles unaligned
+	add	r7,r8,r7,ror#2			@ E+=K_00_19
+	eor	r10,r5,r6			@ F_xx_xx
+	add	r7,r7,r3,ror#27			@ E+=ROR(A,27)
+#ifdef __ARMEL__
+	rev	r9,r9				@ byte swap
+#endif
+#endif
+	and	r10,r4,r10,ror#2
+	add	r7,r7,r9			@ E+=X[i]
+	eor	r10,r10,r6,ror#2		@ F_00_19(B,C,D)
+	str	r9,[r14,#-4]!
+	add	r7,r7,r10			@ E+=F_00_19(B,C,D)
+	ldr	r9,[r14,#15*4]
+	ldr	r10,[r14,#13*4]
+	ldr	r11,[r14,#7*4]
+	add	r6,r8,r6,ror#2			@ E+=K_xx_xx
+	ldr	r12,[r14,#2*4]
+	eor	r9,r9,r10
+	eor	r11,r11,r12			@ 1 cycle stall
+	eor	r10,r4,r5			@ F_xx_xx
+	mov	r9,r9,ror#31
+	add	r6,r6,r7,ror#27			@ E+=ROR(A,27)
+	eor	r9,r9,r11,ror#31
+	str	r9,[r14,#-4]!
+	and r10,r3,r10,ror#2					@ F_xx_xx
+						@ F_xx_xx
+	add	r6,r6,r9			@ E+=X[i]
+	eor	r10,r10,r5,ror#2		@ F_00_19(B,C,D)
+	add	r6,r6,r10			@ E+=F_00_19(B,C,D)
+	ldr	r9,[r14,#15*4]
+	ldr	r10,[r14,#13*4]
+	ldr	r11,[r14,#7*4]
+	add	r5,r8,r5,ror#2			@ E+=K_xx_xx
+	ldr	r12,[r14,#2*4]
+	eor	r9,r9,r10
+	eor	r11,r11,r12			@ 1 cycle stall
+	eor	r10,r3,r4			@ F_xx_xx
+	mov	r9,r9,ror#31
+	add	r5,r5,r6,ror#27			@ E+=ROR(A,27)
+	eor	r9,r9,r11,ror#31
+	str	r9,[r14,#-4]!
+	and r10,r7,r10,ror#2					@ F_xx_xx
+						@ F_xx_xx
+	add	r5,r5,r9			@ E+=X[i]
+	eor	r10,r10,r4,ror#2		@ F_00_19(B,C,D)
+	add	r5,r5,r10			@ E+=F_00_19(B,C,D)
+	ldr	r9,[r14,#15*4]
+	ldr	r10,[r14,#13*4]
+	ldr	r11,[r14,#7*4]
+	add	r4,r8,r4,ror#2			@ E+=K_xx_xx
+	ldr	r12,[r14,#2*4]
+	eor	r9,r9,r10
+	eor	r11,r11,r12			@ 1 cycle stall
+	eor	r10,r7,r3			@ F_xx_xx
+	mov	r9,r9,ror#31
+	add	r4,r4,r5,ror#27			@ E+=ROR(A,27)
+	eor	r9,r9,r11,ror#31
+	str	r9,[r14,#-4]!
+	and r10,r6,r10,ror#2					@ F_xx_xx
+						@ F_xx_xx
+	add	r4,r4,r9			@ E+=X[i]
+	eor	r10,r10,r3,ror#2		@ F_00_19(B,C,D)
+	add	r4,r4,r10			@ E+=F_00_19(B,C,D)
+	ldr	r9,[r14,#15*4]
+	ldr	r10,[r14,#13*4]
+	ldr	r11,[r14,#7*4]
+	add	r3,r8,r3,ror#2			@ E+=K_xx_xx
+	ldr	r12,[r14,#2*4]
+	eor	r9,r9,r10
+	eor	r11,r11,r12			@ 1 cycle stall
+	eor	r10,r6,r7			@ F_xx_xx
+	mov	r9,r9,ror#31
+	add	r3,r3,r4,ror#27			@ E+=ROR(A,27)
+	eor	r9,r9,r11,ror#31
+	str	r9,[r14,#-4]!
+	and r10,r5,r10,ror#2					@ F_xx_xx
+						@ F_xx_xx
+	add	r3,r3,r9			@ E+=X[i]
+	eor	r10,r10,r7,ror#2		@ F_00_19(B,C,D)
+	add	r3,r3,r10			@ E+=F_00_19(B,C,D)
+
+	ldr	r8,.LK_20_39		@ [+15+16*4]
+	cmn	sp,#0			@ [+3], clear carry to denote 20_39
+.L_20_39_or_60_79:
+	ldr	r9,[r14,#15*4]
+	ldr	r10,[r14,#13*4]
+	ldr	r11,[r14,#7*4]
+	add	r7,r8,r7,ror#2			@ E+=K_xx_xx
+	ldr	r12,[r14,#2*4]
+	eor	r9,r9,r10
+	eor	r11,r11,r12			@ 1 cycle stall
+	eor	r10,r5,r6			@ F_xx_xx
+	mov	r9,r9,ror#31
+	add	r7,r7,r3,ror#27			@ E+=ROR(A,27)
+	eor	r9,r9,r11,ror#31
+	str	r9,[r14,#-4]!
+	eor r10,r4,r10,ror#2					@ F_xx_xx
+						@ F_xx_xx
+	add	r7,r7,r9			@ E+=X[i]
+	add	r7,r7,r10			@ E+=F_20_39(B,C,D)
+	ldr	r9,[r14,#15*4]
+	ldr	r10,[r14,#13*4]
+	ldr	r11,[r14,#7*4]
+	add	r6,r8,r6,ror#2			@ E+=K_xx_xx
+	ldr	r12,[r14,#2*4]
+	eor	r9,r9,r10
+	eor	r11,r11,r12			@ 1 cycle stall
+	eor	r10,r4,r5			@ F_xx_xx
+	mov	r9,r9,ror#31
+	add	r6,r6,r7,ror#27			@ E+=ROR(A,27)
+	eor	r9,r9,r11,ror#31
+	str	r9,[r14,#-4]!
+	eor r10,r3,r10,ror#2					@ F_xx_xx
+						@ F_xx_xx
+	add	r6,r6,r9			@ E+=X[i]
+	add	r6,r6,r10			@ E+=F_20_39(B,C,D)
+	ldr	r9,[r14,#15*4]
+	ldr	r10,[r14,#13*4]
+	ldr	r11,[r14,#7*4]
+	add	r5,r8,r5,ror#2			@ E+=K_xx_xx
+	ldr	r12,[r14,#2*4]
+	eor	r9,r9,r10
+	eor	r11,r11,r12			@ 1 cycle stall
+	eor	r10,r3,r4			@ F_xx_xx
+	mov	r9,r9,ror#31
+	add	r5,r5,r6,ror#27			@ E+=ROR(A,27)
+	eor	r9,r9,r11,ror#31
+	str	r9,[r14,#-4]!
+	eor r10,r7,r10,ror#2					@ F_xx_xx
+						@ F_xx_xx
+	add	r5,r5,r9			@ E+=X[i]
+	add	r5,r5,r10			@ E+=F_20_39(B,C,D)
+	ldr	r9,[r14,#15*4]
+	ldr	r10,[r14,#13*4]
+	ldr	r11,[r14,#7*4]
+	add	r4,r8,r4,ror#2			@ E+=K_xx_xx
+	ldr	r12,[r14,#2*4]
+	eor	r9,r9,r10
+	eor	r11,r11,r12			@ 1 cycle stall
+	eor	r10,r7,r3			@ F_xx_xx
+	mov	r9,r9,ror#31
+	add	r4,r4,r5,ror#27			@ E+=ROR(A,27)
+	eor	r9,r9,r11,ror#31
+	str	r9,[r14,#-4]!
+	eor r10,r6,r10,ror#2					@ F_xx_xx
+						@ F_xx_xx
+	add	r4,r4,r9			@ E+=X[i]
+	add	r4,r4,r10			@ E+=F_20_39(B,C,D)
+	ldr	r9,[r14,#15*4]
+	ldr	r10,[r14,#13*4]
+	ldr	r11,[r14,#7*4]
+	add	r3,r8,r3,ror#2			@ E+=K_xx_xx
+	ldr	r12,[r14,#2*4]
+	eor	r9,r9,r10
+	eor	r11,r11,r12			@ 1 cycle stall
+	eor	r10,r6,r7			@ F_xx_xx
+	mov	r9,r9,ror#31
+	add	r3,r3,r4,ror#27			@ E+=ROR(A,27)
+	eor	r9,r9,r11,ror#31
+	str	r9,[r14,#-4]!
+	eor r10,r5,r10,ror#2					@ F_xx_xx
+						@ F_xx_xx
+	add	r3,r3,r9			@ E+=X[i]
+	add	r3,r3,r10			@ E+=F_20_39(B,C,D)
+ ARM(	teq	r14,sp		)	@ preserve carry
+ THUMB(	mov	r11,sp		)
+ THUMB(	teq	r14,r11		)	@ preserve carry
+	bne	.L_20_39_or_60_79	@ [+((12+3)*5+2)*4]
+	bcs	.L_done			@ [+((12+3)*5+2)*4], spare 300 bytes
+
+	ldr	r8,.LK_40_59
+	sub	sp,sp,#20*4		@ [+2]
+.L_40_59:
+	ldr	r9,[r14,#15*4]
+	ldr	r10,[r14,#13*4]
+	ldr	r11,[r14,#7*4]
+	add	r7,r8,r7,ror#2			@ E+=K_xx_xx
+	ldr	r12,[r14,#2*4]
+	eor	r9,r9,r10
+	eor	r11,r11,r12			@ 1 cycle stall
+	eor	r10,r5,r6			@ F_xx_xx
+	mov	r9,r9,ror#31
+	add	r7,r7,r3,ror#27			@ E+=ROR(A,27)
+	eor	r9,r9,r11,ror#31
+	str	r9,[r14,#-4]!
+	and r10,r4,r10,ror#2					@ F_xx_xx
+	and r11,r5,r6					@ F_xx_xx
+	add	r7,r7,r9			@ E+=X[i]
+	add	r7,r7,r10			@ E+=F_40_59(B,C,D)
+	add	r7,r7,r11,ror#2
+	ldr	r9,[r14,#15*4]
+	ldr	r10,[r14,#13*4]
+	ldr	r11,[r14,#7*4]
+	add	r6,r8,r6,ror#2			@ E+=K_xx_xx
+	ldr	r12,[r14,#2*4]
+	eor	r9,r9,r10
+	eor	r11,r11,r12			@ 1 cycle stall
+	eor	r10,r4,r5			@ F_xx_xx
+	mov	r9,r9,ror#31
+	add	r6,r6,r7,ror#27			@ E+=ROR(A,27)
+	eor	r9,r9,r11,ror#31
+	str	r9,[r14,#-4]!
+	and r10,r3,r10,ror#2					@ F_xx_xx
+	and r11,r4,r5					@ F_xx_xx
+	add	r6,r6,r9			@ E+=X[i]
+	add	r6,r6,r10			@ E+=F_40_59(B,C,D)
+	add	r6,r6,r11,ror#2
+	ldr	r9,[r14,#15*4]
+	ldr	r10,[r14,#13*4]
+	ldr	r11,[r14,#7*4]
+	add	r5,r8,r5,ror#2			@ E+=K_xx_xx
+	ldr	r12,[r14,#2*4]
+	eor	r9,r9,r10
+	eor	r11,r11,r12			@ 1 cycle stall
+	eor	r10,r3,r4			@ F_xx_xx
+	mov	r9,r9,ror#31
+	add	r5,r5,r6,ror#27			@ E+=ROR(A,27)
+	eor	r9,r9,r11,ror#31
+	str	r9,[r14,#-4]!
+	and r10,r7,r10,ror#2					@ F_xx_xx
+	and r11,r3,r4					@ F_xx_xx
+	add	r5,r5,r9			@ E+=X[i]
+	add	r5,r5,r10			@ E+=F_40_59(B,C,D)
+	add	r5,r5,r11,ror#2
+	ldr	r9,[r14,#15*4]
+	ldr	r10,[r14,#13*4]
+	ldr	r11,[r14,#7*4]
+	add	r4,r8,r4,ror#2			@ E+=K_xx_xx
+	ldr	r12,[r14,#2*4]
+	eor	r9,r9,r10
+	eor	r11,r11,r12			@ 1 cycle stall
+	eor	r10,r7,r3			@ F_xx_xx
+	mov	r9,r9,ror#31
+	add	r4,r4,r5,ror#27			@ E+=ROR(A,27)
+	eor	r9,r9,r11,ror#31
+	str	r9,[r14,#-4]!
+	and r10,r6,r10,ror#2					@ F_xx_xx
+	and r11,r7,r3					@ F_xx_xx
+	add	r4,r4,r9			@ E+=X[i]
+	add	r4,r4,r10			@ E+=F_40_59(B,C,D)
+	add	r4,r4,r11,ror#2
+	ldr	r9,[r14,#15*4]
+	ldr	r10,[r14,#13*4]
+	ldr	r11,[r14,#7*4]
+	add	r3,r8,r3,ror#2			@ E+=K_xx_xx
+	ldr	r12,[r14,#2*4]
+	eor	r9,r9,r10
+	eor	r11,r11,r12			@ 1 cycle stall
+	eor	r10,r6,r7			@ F_xx_xx
+	mov	r9,r9,ror#31
+	add	r3,r3,r4,ror#27			@ E+=ROR(A,27)
+	eor	r9,r9,r11,ror#31
+	str	r9,[r14,#-4]!
+	and r10,r5,r10,ror#2					@ F_xx_xx
+	and r11,r6,r7					@ F_xx_xx
+	add	r3,r3,r9			@ E+=X[i]
+	add	r3,r3,r10			@ E+=F_40_59(B,C,D)
+	add	r3,r3,r11,ror#2
+	cmp	r14,sp
+	bne	.L_40_59		@ [+((12+5)*5+2)*4]
+
+	ldr	r8,.LK_60_79
+	sub	sp,sp,#20*4
+	cmp	sp,#0			@ set carry to denote 60_79
+	b	.L_20_39_or_60_79	@ [+4], spare 300 bytes
+.L_done:
+	add	sp,sp,#80*4		@ "deallocate" stack frame
+	ldmia	r0,{r8,r9,r10,r11,r12}
+	add	r3,r8,r3
+	add	r4,r9,r4
+	add	r5,r10,r5,ror#2
+	add	r6,r11,r6,ror#2
+	add	r7,r12,r7,ror#2
+	stmia	r0,{r3,r4,r5,r6,r7}
+	teq	r1,r2
+	bne	.Lloop			@ [+18], total 1307
+
+	ldmia	sp!,{r4-r12,pc}
+.align	2
+.LK_00_19:	.word	0x5a827999
+.LK_20_39:	.word	0x6ed9eba1
+.LK_40_59:	.word	0x8f1bbcdc
+.LK_60_79:	.word	0xca62c1d6
+ENDPROC(sha1_block_data_order)
+.asciz	"SHA1 block transform for ARMv4, CRYPTOGAMS by <appro@openssl.org>"
+.align	2
diff --git a/arch/arm/crypto/sha1-armv7-neon.S b/arch/arm/crypto/sha1-armv7-neon.S
new file mode 100644
index 000000000..dcd01f3f0
--- /dev/null
+++ b/arch/arm/crypto/sha1-armv7-neon.S
@@ -0,0 +1,639 @@
+/* sha1-armv7-neon.S - ARM/NEON accelerated SHA-1 transform function
+ *
+ * Copyright © 2013-2014 Jussi Kivilinna <jussi.kivilinna@iki.fi>
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License as published by the Free
+ * Software Foundation; either version 2 of the License, or (at your option)
+ * any later version.
+ */
+
+#include <linux/linkage.h>
+#include <asm/assembler.h>
+
+.syntax unified
+.code   32
+.fpu neon
+
+.text
+
+
+/* Context structure */
+
+#define state_h0 0
+#define state_h1 4
+#define state_h2 8
+#define state_h3 12
+#define state_h4 16
+
+
+/* Constants */
+
+#define K1  0x5A827999
+#define K2  0x6ED9EBA1
+#define K3  0x8F1BBCDC
+#define K4  0xCA62C1D6
+.align 4
+.LK_VEC:
+.LK1:	.long K1, K1, K1, K1
+.LK2:	.long K2, K2, K2, K2
+.LK3:	.long K3, K3, K3, K3
+.LK4:	.long K4, K4, K4, K4
+
+
+/* Register macros */
+
+#define RSTATE r0
+#define RDATA r1
+#define RNBLKS r2
+#define ROLDSTACK r3
+#define RWK lr
+
+#define _a r4
+#define _b r5
+#define _c r6
+#define _d r7
+#define _e r8
+
+#define RT0 r9
+#define RT1 r10
+#define RT2 r11
+#define RT3 r12
+
+#define W0 q0
+#define W1 q7
+#define W2 q2
+#define W3 q3
+#define W4 q4
+#define W5 q6
+#define W6 q5
+#define W7 q1
+
+#define tmp0 q8
+#define tmp1 q9
+#define tmp2 q10
+#define tmp3 q11
+
+#define qK1 q12
+#define qK2 q13
+#define qK3 q14
+#define qK4 q15
+
+#ifdef CONFIG_CPU_BIG_ENDIAN
+#define ARM_LE(code...)
+#else
+#define ARM_LE(code...)		code
+#endif
+
+/* Round function macros. */
+
+#define WK_offs(i) (((i) & 15) * 4)
+
+#define _R_F1(a,b,c,d,e,i,pre1,pre2,pre3,i16,\
+	      W,W_m04,W_m08,W_m12,W_m16,W_m20,W_m24,W_m28) \
+	ldr RT3, [sp, WK_offs(i)]; \
+		pre1(i16,W,W_m04,W_m08,W_m12,W_m16,W_m20,W_m24,W_m28); \
+	bic RT0, d, b; \
+	add e, e, a, ror #(32 - 5); \
+	and RT1, c, b; \
+		pre2(i16,W,W_m04,W_m08,W_m12,W_m16,W_m20,W_m24,W_m28); \
+	add RT0, RT0, RT3; \
+	add e, e, RT1; \
+	ror b, #(32 - 30); \
+		pre3(i16,W,W_m04,W_m08,W_m12,W_m16,W_m20,W_m24,W_m28); \
+	add e, e, RT0;
+
+#define _R_F2(a,b,c,d,e,i,pre1,pre2,pre3,i16,\
+	      W,W_m04,W_m08,W_m12,W_m16,W_m20,W_m24,W_m28) \
+	ldr RT3, [sp, WK_offs(i)]; \
+		pre1(i16,W,W_m04,W_m08,W_m12,W_m16,W_m20,W_m24,W_m28); \
+	eor RT0, d, b; \
+	add e, e, a, ror #(32 - 5); \
+	eor RT0, RT0, c; \
+		pre2(i16,W,W_m04,W_m08,W_m12,W_m16,W_m20,W_m24,W_m28); \
+	add e, e, RT3; \
+	ror b, #(32 - 30); \
+		pre3(i16,W,W_m04,W_m08,W_m12,W_m16,W_m20,W_m24,W_m28); \
+	add e, e, RT0; \
+
+#define _R_F3(a,b,c,d,e,i,pre1,pre2,pre3,i16,\
+	      W,W_m04,W_m08,W_m12,W_m16,W_m20,W_m24,W_m28) \
+	ldr RT3, [sp, WK_offs(i)]; \
+		pre1(i16,W,W_m04,W_m08,W_m12,W_m16,W_m20,W_m24,W_m28); \
+	eor RT0, b, c; \
+	and RT1, b, c; \
+	add e, e, a, ror #(32 - 5); \
+		pre2(i16,W,W_m04,W_m08,W_m12,W_m16,W_m20,W_m24,W_m28); \
+	and RT0, RT0, d; \
+	add RT1, RT1, RT3; \
+	add e, e, RT0; \
+	ror b, #(32 - 30); \
+		pre3(i16,W,W_m04,W_m08,W_m12,W_m16,W_m20,W_m24,W_m28); \
+	add e, e, RT1;
+
+#define _R_F4(a,b,c,d,e,i,pre1,pre2,pre3,i16,\
+	      W,W_m04,W_m08,W_m12,W_m16,W_m20,W_m24,W_m28) \
+	_R_F2(a,b,c,d,e,i,pre1,pre2,pre3,i16,\
+	      W,W_m04,W_m08,W_m12,W_m16,W_m20,W_m24,W_m28)
+
+#define _R(a,b,c,d,e,f,i,pre1,pre2,pre3,i16,\
+           W,W_m04,W_m08,W_m12,W_m16,W_m20,W_m24,W_m28) \
+	_R_##f(a,b,c,d,e,i,pre1,pre2,pre3,i16,\
+	       W,W_m04,W_m08,W_m12,W_m16,W_m20,W_m24,W_m28)
+
+#define R(a,b,c,d,e,f,i) \
+	_R_##f(a,b,c,d,e,i,dummy,dummy,dummy,i16,\
+	       W,W_m04,W_m08,W_m12,W_m16,W_m20,W_m24,W_m28)
+
+#define dummy(...)
+
+
+/* Input expansion macros. */
+
+/********* Precalc macros for rounds 0-15 *************************************/
+
+#define W_PRECALC_00_15() \
+	add       RWK, sp, #(WK_offs(0));			\
+	\
+	vld1.32   {W0, W7}, [RDATA]!;				\
+ ARM_LE(vrev32.8  W0, W0;	)	/* big => little */	\
+	vld1.32   {W6, W5}, [RDATA]!;				\
+	vadd.u32  tmp0, W0, curK;				\
+ ARM_LE(vrev32.8  W7, W7;	)	/* big => little */	\
+ ARM_LE(vrev32.8  W6, W6;	)	/* big => little */	\
+	vadd.u32  tmp1, W7, curK;				\
+ ARM_LE(vrev32.8  W5, W5;	)	/* big => little */	\
+	vadd.u32  tmp2, W6, curK;				\
+	vst1.32   {tmp0, tmp1}, [RWK]!;				\
+	vadd.u32  tmp3, W5, curK;				\
+	vst1.32   {tmp2, tmp3}, [RWK];				\
+
+#define WPRECALC_00_15_0(i,W,W_m04,W_m08,W_m12,W_m16,W_m20,W_m24,W_m28) \
+	vld1.32   {W0, W7}, [RDATA]!;				\
+
+#define WPRECALC_00_15_1(i,W,W_m04,W_m08,W_m12,W_m16,W_m20,W_m24,W_m28) \
+	add       RWK, sp, #(WK_offs(0));			\
+
+#define WPRECALC_00_15_2(i,W,W_m04,W_m08,W_m12,W_m16,W_m20,W_m24,W_m28) \
+ ARM_LE(vrev32.8  W0, W0;	)	/* big => little */	\
+
+#define WPRECALC_00_15_3(i,W,W_m04,W_m08,W_m12,W_m16,W_m20,W_m24,W_m28) \
+	vld1.32   {W6, W5}, [RDATA]!;				\
+
+#define WPRECALC_00_15_4(i,W,W_m04,W_m08,W_m12,W_m16,W_m20,W_m24,W_m28) \
+	vadd.u32  tmp0, W0, curK;				\
+
+#define WPRECALC_00_15_5(i,W,W_m04,W_m08,W_m12,W_m16,W_m20,W_m24,W_m28) \
+ ARM_LE(vrev32.8  W7, W7;	)	/* big => little */	\
+
+#define WPRECALC_00_15_6(i,W,W_m04,W_m08,W_m12,W_m16,W_m20,W_m24,W_m28) \
+ ARM_LE(vrev32.8  W6, W6;	)	/* big => little */	\
+
+#define WPRECALC_00_15_7(i,W,W_m04,W_m08,W_m12,W_m16,W_m20,W_m24,W_m28) \
+	vadd.u32  tmp1, W7, curK;				\
+
+#define WPRECALC_00_15_8(i,W,W_m04,W_m08,W_m12,W_m16,W_m20,W_m24,W_m28) \
+ ARM_LE(vrev32.8  W5, W5;	)	/* big => little */	\
+
+#define WPRECALC_00_15_9(i,W,W_m04,W_m08,W_m12,W_m16,W_m20,W_m24,W_m28) \
+	vadd.u32  tmp2, W6, curK;				\
+
+#define WPRECALC_00_15_10(i,W,W_m04,W_m08,W_m12,W_m16,W_m20,W_m24,W_m28) \
+	vst1.32   {tmp0, tmp1}, [RWK]!;				\
+
+#define WPRECALC_00_15_11(i,W,W_m04,W_m08,W_m12,W_m16,W_m20,W_m24,W_m28) \
+	vadd.u32  tmp3, W5, curK;				\
+
+#define WPRECALC_00_15_12(i,W,W_m04,W_m08,W_m12,W_m16,W_m20,W_m24,W_m28) \
+	vst1.32   {tmp2, tmp3}, [RWK];				\
+
+
+/********* Precalc macros for rounds 16-31 ************************************/
+
+#define WPRECALC_16_31_0(i,W,W_m04,W_m08,W_m12,W_m16,W_m20,W_m24,W_m28) \
+	veor      tmp0, tmp0;			\
+	vext.8    W, W_m16, W_m12, #8;		\
+
+#define WPRECALC_16_31_1(i,W,W_m04,W_m08,W_m12,W_m16,W_m20,W_m24,W_m28) \
+	add       RWK, sp, #(WK_offs(i));	\
+	vext.8    tmp0, W_m04, tmp0, #4;	\
+
+#define WPRECALC_16_31_2(i,W,W_m04,W_m08,W_m12,W_m16,W_m20,W_m24,W_m28) \
+	veor      tmp0, tmp0, W_m16;		\
+	veor.32   W, W, W_m08;			\
+
+#define WPRECALC_16_31_3(i,W,W_m04,W_m08,W_m12,W_m16,W_m20,W_m24,W_m28) \
+	veor      tmp1, tmp1;			\
+	veor      W, W, tmp0;			\
+
+#define WPRECALC_16_31_4(i,W,W_m04,W_m08,W_m12,W_m16,W_m20,W_m24,W_m28) \
+	vshl.u32  tmp0, W, #1;			\
+
+#define WPRECALC_16_31_5(i,W,W_m04,W_m08,W_m12,W_m16,W_m20,W_m24,W_m28) \
+	vext.8    tmp1, tmp1, W, #(16-12);	\
+	vshr.u32  W, W, #31;			\
+
+#define WPRECALC_16_31_6(i,W,W_m04,W_m08,W_m12,W_m16,W_m20,W_m24,W_m28) \
+	vorr      tmp0, tmp0, W;		\
+	vshr.u32  W, tmp1, #30;			\
+
+#define WPRECALC_16_31_7(i,W,W_m04,W_m08,W_m12,W_m16,W_m20,W_m24,W_m28) \
+	vshl.u32  tmp1, tmp1, #2;		\
+
+#define WPRECALC_16_31_8(i,W,W_m04,W_m08,W_m12,W_m16,W_m20,W_m24,W_m28) \
+	veor      tmp0, tmp0, W;		\
+
+#define WPRECALC_16_31_9(i,W,W_m04,W_m08,W_m12,W_m16,W_m20,W_m24,W_m28) \
+	veor      W, tmp0, tmp1;		\
+
+#define WPRECALC_16_31_10(i,W,W_m04,W_m08,W_m12,W_m16,W_m20,W_m24,W_m28) \
+	vadd.u32  tmp0, W, curK;		\
+
+#define WPRECALC_16_31_11(i,W,W_m04,W_m08,W_m12,W_m16,W_m20,W_m24,W_m28) \
+	vst1.32   {tmp0}, [RWK];
+
+
+/********* Precalc macros for rounds 32-79 ************************************/
+
+#define WPRECALC_32_79_0(i,W,W_m04,W_m08,W_m12,W_m16,W_m20,W_m24,W_m28) \
+	veor W, W_m28; \
+
+#define WPRECALC_32_79_1(i,W,W_m04,W_m08,W_m12,W_m16,W_m20,W_m24,W_m28) \
+	vext.8 tmp0, W_m08, W_m04, #8; \
+
+#define WPRECALC_32_79_2(i,W,W_m04,W_m08,W_m12,W_m16,W_m20,W_m24,W_m28) \
+	veor W, W_m16; \
+
+#define WPRECALC_32_79_3(i,W,W_m04,W_m08,W_m12,W_m16,W_m20,W_m24,W_m28) \
+	veor W, tmp0; \
+
+#define WPRECALC_32_79_4(i,W,W_m04,W_m08,W_m12,W_m16,W_m20,W_m24,W_m28) \
+	add RWK, sp, #(WK_offs(i&~3)); \
+
+#define WPRECALC_32_79_5(i,W,W_m04,W_m08,W_m12,W_m16,W_m20,W_m24,W_m28) \
+	vshl.u32 tmp1, W, #2; \
+
+#define WPRECALC_32_79_6(i,W,W_m04,W_m08,W_m12,W_m16,W_m20,W_m24,W_m28) \
+	vshr.u32 tmp0, W, #30; \
+
+#define WPRECALC_32_79_7(i,W,W_m04,W_m08,W_m12,W_m16,W_m20,W_m24,W_m28) \
+	vorr W, tmp0, tmp1; \
+
+#define WPRECALC_32_79_8(i,W,W_m04,W_m08,W_m12,W_m16,W_m20,W_m24,W_m28) \
+	vadd.u32 tmp0, W, curK; \
+
+#define WPRECALC_32_79_9(i,W,W_m04,W_m08,W_m12,W_m16,W_m20,W_m24,W_m28) \
+	vst1.32 {tmp0}, [RWK];
+
+
+/*
+ * Transform nblks*64 bytes (nblks*16 32-bit words) at DATA.
+ *
+ * unsigned int
+ * sha1_transform_neon (void *ctx, const unsigned char *data,
+ *                      unsigned int nblks)
+ */
+.align 3
+ENTRY(sha1_transform_neon)
+  /* input:
+   *	r0: ctx, CTX
+   *	r1: data (64*nblks bytes)
+   *	r2: nblks
+   */
+
+  cmp RNBLKS, #0;
+  beq .Ldo_nothing;
+
+  push {r4-r12, lr};
+  /*vpush {q4-q7};*/
+
+  adr RT3, .LK_VEC;
+
+  mov ROLDSTACK, sp;
+
+  /* Align stack. */
+  sub RT0, sp, #(16*4);
+  and RT0, #(~(16-1));
+  mov sp, RT0;
+
+  vld1.32 {qK1-qK2}, [RT3]!; /* Load K1,K2 */
+
+  /* Get the values of the chaining variables. */
+  ldm RSTATE, {_a-_e};
+
+  vld1.32 {qK3-qK4}, [RT3]; /* Load K3,K4 */
+
+#undef curK
+#define curK qK1
+  /* Precalc 0-15. */
+  W_PRECALC_00_15();
+
+.Loop:
+  /* Transform 0-15 + Precalc 16-31. */
+  _R( _a, _b, _c, _d, _e, F1,  0,
+      WPRECALC_16_31_0, WPRECALC_16_31_1, WPRECALC_16_31_2, 16,
+      W4, W5, W6, W7, W0, _, _, _ );
+  _R( _e, _a, _b, _c, _d, F1,  1,
+      WPRECALC_16_31_3, WPRECALC_16_31_4, WPRECALC_16_31_5, 16,
+      W4, W5, W6, W7, W0, _, _, _ );
+  _R( _d, _e, _a, _b, _c, F1,  2,
+      WPRECALC_16_31_6, WPRECALC_16_31_7, WPRECALC_16_31_8, 16,
+      W4, W5, W6, W7, W0, _, _, _ );
+  _R( _c, _d, _e, _a, _b, F1,  3,
+      WPRECALC_16_31_9, WPRECALC_16_31_10,WPRECALC_16_31_11,16,
+      W4, W5, W6, W7, W0, _, _, _ );
+
+#undef curK
+#define curK qK2
+  _R( _b, _c, _d, _e, _a, F1,  4,
+      WPRECALC_16_31_0, WPRECALC_16_31_1, WPRECALC_16_31_2, 20,
+      W3, W4, W5, W6, W7, _, _, _ );
+  _R( _a, _b, _c, _d, _e, F1,  5,
+      WPRECALC_16_31_3, WPRECALC_16_31_4, WPRECALC_16_31_5, 20,
+      W3, W4, W5, W6, W7, _, _, _ );
+  _R( _e, _a, _b, _c, _d, F1,  6,
+      WPRECALC_16_31_6, WPRECALC_16_31_7, WPRECALC_16_31_8, 20,
+      W3, W4, W5, W6, W7, _, _, _ );
+  _R( _d, _e, _a, _b, _c, F1,  7,
+      WPRECALC_16_31_9, WPRECALC_16_31_10,WPRECALC_16_31_11,20,
+      W3, W4, W5, W6, W7, _, _, _ );
+
+  _R( _c, _d, _e, _a, _b, F1,  8,
+      WPRECALC_16_31_0, WPRECALC_16_31_1, WPRECALC_16_31_2, 24,
+      W2, W3, W4, W5, W6, _, _, _ );
+  _R( _b, _c, _d, _e, _a, F1,  9,
+      WPRECALC_16_31_3, WPRECALC_16_31_4, WPRECALC_16_31_5, 24,
+      W2, W3, W4, W5, W6, _, _, _ );
+  _R( _a, _b, _c, _d, _e, F1, 10,
+      WPRECALC_16_31_6, WPRECALC_16_31_7, WPRECALC_16_31_8, 24,
+      W2, W3, W4, W5, W6, _, _, _ );
+  _R( _e, _a, _b, _c, _d, F1, 11,
+      WPRECALC_16_31_9, WPRECALC_16_31_10,WPRECALC_16_31_11,24,
+      W2, W3, W4, W5, W6, _, _, _ );
+
+  _R( _d, _e, _a, _b, _c, F1, 12,
+      WPRECALC_16_31_0, WPRECALC_16_31_1, WPRECALC_16_31_2, 28,
+      W1, W2, W3, W4, W5, _, _, _ );
+  _R( _c, _d, _e, _a, _b, F1, 13,
+      WPRECALC_16_31_3, WPRECALC_16_31_4, WPRECALC_16_31_5, 28,
+      W1, W2, W3, W4, W5, _, _, _ );
+  _R( _b, _c, _d, _e, _a, F1, 14,
+      WPRECALC_16_31_6, WPRECALC_16_31_7, WPRECALC_16_31_8, 28,
+      W1, W2, W3, W4, W5, _, _, _ );
+  _R( _a, _b, _c, _d, _e, F1, 15,
+      WPRECALC_16_31_9, WPRECALC_16_31_10,WPRECALC_16_31_11,28,
+      W1, W2, W3, W4, W5, _, _, _ );
+
+  /* Transform 16-63 + Precalc 32-79. */
+  _R( _e, _a, _b, _c, _d, F1, 16,
+      WPRECALC_32_79_0, WPRECALC_32_79_1, WPRECALC_32_79_2, 32,
+      W0, W1, W2, W3, W4, W5, W6, W7);
+  _R( _d, _e, _a, _b, _c, F1, 17,
+      WPRECALC_32_79_3, WPRECALC_32_79_4, WPRECALC_32_79_5, 32,
+      W0, W1, W2, W3, W4, W5, W6, W7);
+  _R( _c, _d, _e, _a, _b, F1, 18,
+      WPRECALC_32_79_6, dummy,            WPRECALC_32_79_7, 32,
+      W0, W1, W2, W3, W4, W5, W6, W7);
+  _R( _b, _c, _d, _e, _a, F1, 19,
+      WPRECALC_32_79_8, dummy,            WPRECALC_32_79_9, 32,
+      W0, W1, W2, W3, W4, W5, W6, W7);
+
+  _R( _a, _b, _c, _d, _e, F2, 20,
+      WPRECALC_32_79_0, WPRECALC_32_79_1, WPRECALC_32_79_2, 36,
+      W7, W0, W1, W2, W3, W4, W5, W6);
+  _R( _e, _a, _b, _c, _d, F2, 21,
+      WPRECALC_32_79_3, WPRECALC_32_79_4, WPRECALC_32_79_5, 36,
+      W7, W0, W1, W2, W3, W4, W5, W6);
+  _R( _d, _e, _a, _b, _c, F2, 22,
+      WPRECALC_32_79_6, dummy,            WPRECALC_32_79_7, 36,
+      W7, W0, W1, W2, W3, W4, W5, W6);
+  _R( _c, _d, _e, _a, _b, F2, 23,
+      WPRECALC_32_79_8, dummy,            WPRECALC_32_79_9, 36,
+      W7, W0, W1, W2, W3, W4, W5, W6);
+
+#undef curK
+#define curK qK3
+  _R( _b, _c, _d, _e, _a, F2, 24,
+      WPRECALC_32_79_0, WPRECALC_32_79_1, WPRECALC_32_79_2, 40,
+      W6, W7, W0, W1, W2, W3, W4, W5);
+  _R( _a, _b, _c, _d, _e, F2, 25,
+      WPRECALC_32_79_3, WPRECALC_32_79_4, WPRECALC_32_79_5, 40,
+      W6, W7, W0, W1, W2, W3, W4, W5);
+  _R( _e, _a, _b, _c, _d, F2, 26,
+      WPRECALC_32_79_6, dummy,            WPRECALC_32_79_7, 40,
+      W6, W7, W0, W1, W2, W3, W4, W5);
+  _R( _d, _e, _a, _b, _c, F2, 27,
+      WPRECALC_32_79_8, dummy,            WPRECALC_32_79_9, 40,
+      W6, W7, W0, W1, W2, W3, W4, W5);
+
+  _R( _c, _d, _e, _a, _b, F2, 28,
+      WPRECALC_32_79_0, WPRECALC_32_79_1, WPRECALC_32_79_2, 44,
+      W5, W6, W7, W0, W1, W2, W3, W4);
+  _R( _b, _c, _d, _e, _a, F2, 29,
+      WPRECALC_32_79_3, WPRECALC_32_79_4, WPRECALC_32_79_5, 44,
+      W5, W6, W7, W0, W1, W2, W3, W4);
+  _R( _a, _b, _c, _d, _e, F2, 30,
+      WPRECALC_32_79_6, dummy,            WPRECALC_32_79_7, 44,
+      W5, W6, W7, W0, W1, W2, W3, W4);
+  _R( _e, _a, _b, _c, _d, F2, 31,
+      WPRECALC_32_79_8, dummy,            WPRECALC_32_79_9, 44,
+      W5, W6, W7, W0, W1, W2, W3, W4);
+
+  _R( _d, _e, _a, _b, _c, F2, 32,
+      WPRECALC_32_79_0, WPRECALC_32_79_1, WPRECALC_32_79_2, 48,
+      W4, W5, W6, W7, W0, W1, W2, W3);
+  _R( _c, _d, _e, _a, _b, F2, 33,
+      WPRECALC_32_79_3, WPRECALC_32_79_4, WPRECALC_32_79_5, 48,
+      W4, W5, W6, W7, W0, W1, W2, W3);
+  _R( _b, _c, _d, _e, _a, F2, 34,
+      WPRECALC_32_79_6, dummy,            WPRECALC_32_79_7, 48,
+      W4, W5, W6, W7, W0, W1, W2, W3);
+  _R( _a, _b, _c, _d, _e, F2, 35,
+      WPRECALC_32_79_8, dummy,            WPRECALC_32_79_9, 48,
+      W4, W5, W6, W7, W0, W1, W2, W3);
+
+  _R( _e, _a, _b, _c, _d, F2, 36,
+      WPRECALC_32_79_0, WPRECALC_32_79_1, WPRECALC_32_79_2, 52,
+      W3, W4, W5, W6, W7, W0, W1, W2);
+  _R( _d, _e, _a, _b, _c, F2, 37,
+      WPRECALC_32_79_3, WPRECALC_32_79_4, WPRECALC_32_79_5, 52,
+      W3, W4, W5, W6, W7, W0, W1, W2);
+  _R( _c, _d, _e, _a, _b, F2, 38,
+      WPRECALC_32_79_6, dummy,            WPRECALC_32_79_7, 52,
+      W3, W4, W5, W6, W7, W0, W1, W2);
+  _R( _b, _c, _d, _e, _a, F2, 39,
+      WPRECALC_32_79_8, dummy,            WPRECALC_32_79_9, 52,
+      W3, W4, W5, W6, W7, W0, W1, W2);
+
+  _R( _a, _b, _c, _d, _e, F3, 40,
+      WPRECALC_32_79_0, WPRECALC_32_79_1, WPRECALC_32_79_2, 56,
+      W2, W3, W4, W5, W6, W7, W0, W1);
+  _R( _e, _a, _b, _c, _d, F3, 41,
+      WPRECALC_32_79_3, WPRECALC_32_79_4, WPRECALC_32_79_5, 56,
+      W2, W3, W4, W5, W6, W7, W0, W1);
+  _R( _d, _e, _a, _b, _c, F3, 42,
+      WPRECALC_32_79_6, dummy,            WPRECALC_32_79_7, 56,
+      W2, W3, W4, W5, W6, W7, W0, W1);
+  _R( _c, _d, _e, _a, _b, F3, 43,
+      WPRECALC_32_79_8, dummy,            WPRECALC_32_79_9, 56,
+      W2, W3, W4, W5, W6, W7, W0, W1);
+
+#undef curK
+#define curK qK4
+  _R( _b, _c, _d, _e, _a, F3, 44,
+      WPRECALC_32_79_0, WPRECALC_32_79_1, WPRECALC_32_79_2, 60,
+      W1, W2, W3, W4, W5, W6, W7, W0);
+  _R( _a, _b, _c, _d, _e, F3, 45,
+      WPRECALC_32_79_3, WPRECALC_32_79_4, WPRECALC_32_79_5, 60,
+      W1, W2, W3, W4, W5, W6, W7, W0);
+  _R( _e, _a, _b, _c, _d, F3, 46,
+      WPRECALC_32_79_6, dummy,            WPRECALC_32_79_7, 60,
+      W1, W2, W3, W4, W5, W6, W7, W0);
+  _R( _d, _e, _a, _b, _c, F3, 47,
+      WPRECALC_32_79_8, dummy,            WPRECALC_32_79_9, 60,
+      W1, W2, W3, W4, W5, W6, W7, W0);
+
+  _R( _c, _d, _e, _a, _b, F3, 48,
+      WPRECALC_32_79_0, WPRECALC_32_79_1, WPRECALC_32_79_2, 64,
+      W0, W1, W2, W3, W4, W5, W6, W7);
+  _R( _b, _c, _d, _e, _a, F3, 49,
+      WPRECALC_32_79_3, WPRECALC_32_79_4, WPRECALC_32_79_5, 64,
+      W0, W1, W2, W3, W4, W5, W6, W7);
+  _R( _a, _b, _c, _d, _e, F3, 50,
+      WPRECALC_32_79_6, dummy,            WPRECALC_32_79_7, 64,
+      W0, W1, W2, W3, W4, W5, W6, W7);
+  _R( _e, _a, _b, _c, _d, F3, 51,
+      WPRECALC_32_79_8, dummy,            WPRECALC_32_79_9, 64,
+      W0, W1, W2, W3, W4, W5, W6, W7);
+
+  _R( _d, _e, _a, _b, _c, F3, 52,
+      WPRECALC_32_79_0, WPRECALC_32_79_1, WPRECALC_32_79_2, 68,
+      W7, W0, W1, W2, W3, W4, W5, W6);
+  _R( _c, _d, _e, _a, _b, F3, 53,
+      WPRECALC_32_79_3, WPRECALC_32_79_4, WPRECALC_32_79_5, 68,
+      W7, W0, W1, W2, W3, W4, W5, W6);
+  _R( _b, _c, _d, _e, _a, F3, 54,
+      WPRECALC_32_79_6, dummy,            WPRECALC_32_79_7, 68,
+      W7, W0, W1, W2, W3, W4, W5, W6);
+  _R( _a, _b, _c, _d, _e, F3, 55,
+      WPRECALC_32_79_8, dummy,            WPRECALC_32_79_9, 68,
+      W7, W0, W1, W2, W3, W4, W5, W6);
+
+  _R( _e, _a, _b, _c, _d, F3, 56,
+      WPRECALC_32_79_0, WPRECALC_32_79_1, WPRECALC_32_79_2, 72,
+      W6, W7, W0, W1, W2, W3, W4, W5);
+  _R( _d, _e, _a, _b, _c, F3, 57,
+      WPRECALC_32_79_3, WPRECALC_32_79_4, WPRECALC_32_79_5, 72,
+      W6, W7, W0, W1, W2, W3, W4, W5);
+  _R( _c, _d, _e, _a, _b, F3, 58,
+      WPRECALC_32_79_6, dummy,            WPRECALC_32_79_7, 72,
+      W6, W7, W0, W1, W2, W3, W4, W5);
+  _R( _b, _c, _d, _e, _a, F3, 59,
+      WPRECALC_32_79_8, dummy,            WPRECALC_32_79_9, 72,
+      W6, W7, W0, W1, W2, W3, W4, W5);
+
+  subs RNBLKS, #1;
+
+  _R( _a, _b, _c, _d, _e, F4, 60,
+      WPRECALC_32_79_0, WPRECALC_32_79_1, WPRECALC_32_79_2, 76,
+      W5, W6, W7, W0, W1, W2, W3, W4);
+  _R( _e, _a, _b, _c, _d, F4, 61,
+      WPRECALC_32_79_3, WPRECALC_32_79_4, WPRECALC_32_79_5, 76,
+      W5, W6, W7, W0, W1, W2, W3, W4);
+  _R( _d, _e, _a, _b, _c, F4, 62,
+      WPRECALC_32_79_6, dummy,            WPRECALC_32_79_7, 76,
+      W5, W6, W7, W0, W1, W2, W3, W4);
+  _R( _c, _d, _e, _a, _b, F4, 63,
+      WPRECALC_32_79_8, dummy,            WPRECALC_32_79_9, 76,
+      W5, W6, W7, W0, W1, W2, W3, W4);
+
+  beq .Lend;
+
+  /* Transform 64-79 + Precalc 0-15 of next block. */
+#undef curK
+#define curK qK1
+  _R( _b, _c, _d, _e, _a, F4, 64,
+      WPRECALC_00_15_0, dummy, dummy, _, _, _, _, _, _, _, _, _ );
+  _R( _a, _b, _c, _d, _e, F4, 65,
+      WPRECALC_00_15_1, dummy, dummy, _, _, _, _, _, _, _, _, _ );
+  _R( _e, _a, _b, _c, _d, F4, 66,
+      WPRECALC_00_15_2, dummy, dummy, _, _, _, _, _, _, _, _, _ );
+  _R( _d, _e, _a, _b, _c, F4, 67,
+      WPRECALC_00_15_3, dummy, dummy, _, _, _, _, _, _, _, _, _ );
+
+  _R( _c, _d, _e, _a, _b, F4, 68,
+      dummy,            dummy, dummy, _, _, _, _, _, _, _, _, _ );
+  _R( _b, _c, _d, _e, _a, F4, 69,
+      dummy,            dummy, dummy, _, _, _, _, _, _, _, _, _ );
+  _R( _a, _b, _c, _d, _e, F4, 70,
+      WPRECALC_00_15_4, dummy, dummy, _, _, _, _, _, _, _, _, _ );
+  _R( _e, _a, _b, _c, _d, F4, 71,
+      WPRECALC_00_15_5, dummy, dummy, _, _, _, _, _, _, _, _, _ );
+
+  _R( _d, _e, _a, _b, _c, F4, 72,
+      dummy,            dummy, dummy, _, _, _, _, _, _, _, _, _ );
+  _R( _c, _d, _e, _a, _b, F4, 73,
+      dummy,            dummy, dummy, _, _, _, _, _, _, _, _, _ );
+  _R( _b, _c, _d, _e, _a, F4, 74,
+      WPRECALC_00_15_6, dummy, dummy, _, _, _, _, _, _, _, _, _ );
+  _R( _a, _b, _c, _d, _e, F4, 75,
+      WPRECALC_00_15_7, dummy, dummy, _, _, _, _, _, _, _, _, _ );
+
+  _R( _e, _a, _b, _c, _d, F4, 76,
+      WPRECALC_00_15_8, dummy, dummy, _, _, _, _, _, _, _, _, _ );
+  _R( _d, _e, _a, _b, _c, F4, 77,
+      WPRECALC_00_15_9, dummy, dummy, _, _, _, _, _, _, _, _, _ );
+  _R( _c, _d, _e, _a, _b, F4, 78,
+      WPRECALC_00_15_10, dummy, dummy, _, _, _, _, _, _, _, _, _ );
+  _R( _b, _c, _d, _e, _a, F4, 79,
+      WPRECALC_00_15_11, dummy, WPRECALC_00_15_12, _, _, _, _, _, _, _, _, _ );
+
+  /* Update the chaining variables. */
+  ldm RSTATE, {RT0-RT3};
+  add _a, RT0;
+  ldr RT0, [RSTATE, #state_h4];
+  add _b, RT1;
+  add _c, RT2;
+  add _d, RT3;
+  add _e, RT0;
+  stm RSTATE, {_a-_e};
+
+  b .Loop;
+
+.Lend:
+  /* Transform 64-79 */
+  R( _b, _c, _d, _e, _a, F4, 64 );
+  R( _a, _b, _c, _d, _e, F4, 65 );
+  R( _e, _a, _b, _c, _d, F4, 66 );
+  R( _d, _e, _a, _b, _c, F4, 67 );
+  R( _c, _d, _e, _a, _b, F4, 68 );
+  R( _b, _c, _d, _e, _a, F4, 69 );
+  R( _a, _b, _c, _d, _e, F4, 70 );
+  R( _e, _a, _b, _c, _d, F4, 71 );
+  R( _d, _e, _a, _b, _c, F4, 72 );
+  R( _c, _d, _e, _a, _b, F4, 73 );
+  R( _b, _c, _d, _e, _a, F4, 74 );
+  R( _a, _b, _c, _d, _e, F4, 75 );
+  R( _e, _a, _b, _c, _d, F4, 76 );
+  R( _d, _e, _a, _b, _c, F4, 77 );
+  R( _c, _d, _e, _a, _b, F4, 78 );
+  R( _b, _c, _d, _e, _a, F4, 79 );
+
+  mov sp, ROLDSTACK;
+
+  /* Update the chaining variables. */
+  ldm RSTATE, {RT0-RT3};
+  add _a, RT0;
+  ldr RT0, [RSTATE, #state_h4];
+  add _b, RT1;
+  add _c, RT2;
+  add _d, RT3;
+  /*vpop {q4-q7};*/
+  add _e, RT0;
+  stm RSTATE, {_a-_e};
+
+  pop {r4-r12, pc};
+
+.Ldo_nothing:
+  bx lr
+ENDPROC(sha1_transform_neon)
diff --git a/arch/arm/crypto/sha1-ce-core.S b/arch/arm/crypto/sha1-ce-core.S
new file mode 100644
index 000000000..b623f51cc
--- /dev/null
+++ b/arch/arm/crypto/sha1-ce-core.S
@@ -0,0 +1,125 @@
+/*
+ * sha1-ce-core.S - SHA-1 secure hash using ARMv8 Crypto Extensions
+ *
+ * Copyright (C) 2015 Linaro Ltd.
+ * Author: Ard Biesheuvel <ard.biesheuvel@linaro.org>
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2 as
+ * published by the Free Software Foundation.
+ */
+
+#include <linux/linkage.h>
+#include <asm/assembler.h>
+
+	.text
+	.fpu		crypto-neon-fp-armv8
+
+	k0		.req	q0
+	k1		.req	q1
+	k2		.req	q2
+	k3		.req	q3
+
+	ta0		.req	q4
+	ta1		.req	q5
+	tb0		.req	q5
+	tb1		.req	q4
+
+	dga		.req	q6
+	dgb		.req	q7
+	dgbs		.req	s28
+
+	dg0		.req	q12
+	dg1a0		.req	q13
+	dg1a1		.req	q14
+	dg1b0		.req	q14
+	dg1b1		.req	q13
+
+	.macro		add_only, op, ev, rc, s0, dg1
+	.ifnb		\s0
+	vadd.u32	tb\ev, q\s0, \rc
+	.endif
+	sha1h.32	dg1b\ev, dg0
+	.ifb		\dg1
+	sha1\op\().32	dg0, dg1a\ev, ta\ev
+	.else
+	sha1\op\().32	dg0, \dg1, ta\ev
+	.endif
+	.endm
+
+	.macro		add_update, op, ev, rc, s0, s1, s2, s3, dg1
+	sha1su0.32	q\s0, q\s1, q\s2
+	add_only	\op, \ev, \rc, \s1, \dg1
+	sha1su1.32	q\s0, q\s3
+	.endm
+
+	.align		6
+.Lsha1_rcon:
+	.word		0x5a827999, 0x5a827999, 0x5a827999, 0x5a827999
+	.word		0x6ed9eba1, 0x6ed9eba1, 0x6ed9eba1, 0x6ed9eba1
+	.word		0x8f1bbcdc, 0x8f1bbcdc, 0x8f1bbcdc, 0x8f1bbcdc
+	.word		0xca62c1d6, 0xca62c1d6, 0xca62c1d6, 0xca62c1d6
+
+	/*
+	 * void sha1_ce_transform(struct sha1_state *sst, u8 const *src,
+	 *			  int blocks);
+	 */
+ENTRY(sha1_ce_transform)
+	/* load round constants */
+	adr		ip, .Lsha1_rcon
+	vld1.32		{k0-k1}, [ip, :128]!
+	vld1.32		{k2-k3}, [ip, :128]
+
+	/* load state */
+	vld1.32		{dga}, [r0]
+	vldr		dgbs, [r0, #16]
+
+	/* load input */
+0:	vld1.32		{q8-q9}, [r1]!
+	vld1.32		{q10-q11}, [r1]!
+	subs		r2, r2, #1
+
+#ifndef CONFIG_CPU_BIG_ENDIAN
+	vrev32.8	q8, q8
+	vrev32.8	q9, q9
+	vrev32.8	q10, q10
+	vrev32.8	q11, q11
+#endif
+
+	vadd.u32	ta0, q8, k0
+	vmov		dg0, dga
+
+	add_update	c, 0, k0,  8,  9, 10, 11, dgb
+	add_update	c, 1, k0,  9, 10, 11,  8
+	add_update	c, 0, k0, 10, 11,  8,  9
+	add_update	c, 1, k0, 11,  8,  9, 10
+	add_update	c, 0, k1,  8,  9, 10, 11
+
+	add_update	p, 1, k1,  9, 10, 11,  8
+	add_update	p, 0, k1, 10, 11,  8,  9
+	add_update	p, 1, k1, 11,  8,  9, 10
+	add_update	p, 0, k1,  8,  9, 10, 11
+	add_update	p, 1, k2,  9, 10, 11,  8
+
+	add_update	m, 0, k2, 10, 11,  8,  9
+	add_update	m, 1, k2, 11,  8,  9, 10
+	add_update	m, 0, k2,  8,  9, 10, 11
+	add_update	m, 1, k2,  9, 10, 11,  8
+	add_update	m, 0, k3, 10, 11,  8,  9
+
+	add_update	p, 1, k3, 11,  8,  9, 10
+	add_only	p, 0, k3,  9
+	add_only	p, 1, k3, 10
+	add_only	p, 0, k3, 11
+	add_only	p, 1
+
+	/* update state */
+	vadd.u32	dga, dga, dg0
+	vadd.u32	dgb, dgb, dg1a0
+	bne		0b
+
+	/* store new state */
+	vst1.32		{dga}, [r0]
+	vstr		dgbs, [r0, #16]
+	bx		lr
+ENDPROC(sha1_ce_transform)
diff --git a/arch/arm/crypto/sha1-ce-glue.c b/arch/arm/crypto/sha1-ce-glue.c
new file mode 100644
index 000000000..80bc2fcd2
--- /dev/null
+++ b/arch/arm/crypto/sha1-ce-glue.c
@@ -0,0 +1,96 @@
+/*
+ * sha1-ce-glue.c - SHA-1 secure hash using ARMv8 Crypto Extensions
+ *
+ * Copyright (C) 2015 Linaro Ltd <ard.biesheuvel@linaro.org>
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2 as
+ * published by the Free Software Foundation.
+ */
+
+#include <crypto/internal/hash.h>
+#include <crypto/sha.h>
+#include <crypto/sha1_base.h>
+#include <linux/crypto.h>
+#include <linux/module.h>
+
+#include <asm/hwcap.h>
+#include <asm/neon.h>
+#include <asm/simd.h>
+
+#include "sha1.h"
+
+MODULE_DESCRIPTION("SHA1 secure hash using ARMv8 Crypto Extensions");
+MODULE_AUTHOR("Ard Biesheuvel <ard.biesheuvel@linaro.org>");
+MODULE_LICENSE("GPL v2");
+
+asmlinkage void sha1_ce_transform(struct sha1_state *sst, u8 const *src,
+				  int blocks);
+
+static int sha1_ce_update(struct shash_desc *desc, const u8 *data,
+			  unsigned int len)
+{
+	struct sha1_state *sctx = shash_desc_ctx(desc);
+
+	if (!may_use_simd() ||
+	    (sctx->count % SHA1_BLOCK_SIZE) + len < SHA1_BLOCK_SIZE)
+		return sha1_update_arm(desc, data, len);
+
+	kernel_neon_begin();
+	sha1_base_do_update(desc, data, len, sha1_ce_transform);
+	kernel_neon_end();
+
+	return 0;
+}
+
+static int sha1_ce_finup(struct shash_desc *desc, const u8 *data,
+			 unsigned int len, u8 *out)
+{
+	if (!may_use_simd())
+		return sha1_finup_arm(desc, data, len, out);
+
+	kernel_neon_begin();
+	if (len)
+		sha1_base_do_update(desc, data, len, sha1_ce_transform);
+	sha1_base_do_finalize(desc, sha1_ce_transform);
+	kernel_neon_end();
+
+	return sha1_base_finish(desc, out);
+}
+
+static int sha1_ce_final(struct shash_desc *desc, u8 *out)
+{
+	return sha1_ce_finup(desc, NULL, 0, out);
+}
+
+static struct shash_alg alg = {
+	.init			= sha1_base_init,
+	.update			= sha1_ce_update,
+	.final			= sha1_ce_final,
+	.finup			= sha1_ce_finup,
+	.descsize		= sizeof(struct sha1_state),
+	.digestsize		= SHA1_DIGEST_SIZE,
+	.base			= {
+		.cra_name		= "sha1",
+		.cra_driver_name	= "sha1-ce",
+		.cra_priority		= 200,
+		.cra_flags		= CRYPTO_ALG_TYPE_SHASH,
+		.cra_blocksize		= SHA1_BLOCK_SIZE,
+		.cra_module		= THIS_MODULE,
+	}
+};
+
+static int __init sha1_ce_mod_init(void)
+{
+	if (!(elf_hwcap2 & HWCAP2_SHA1))
+		return -ENODEV;
+	return crypto_register_shash(&alg);
+}
+
+static void __exit sha1_ce_mod_fini(void)
+{
+	crypto_unregister_shash(&alg);
+}
+
+module_init(sha1_ce_mod_init);
+module_exit(sha1_ce_mod_fini);
diff --git a/arch/arm/crypto/sha1.h b/arch/arm/crypto/sha1.h
new file mode 100644
index 000000000..ffd8bd08b
--- /dev/null
+++ b/arch/arm/crypto/sha1.h
@@ -0,0 +1,13 @@
+#ifndef ASM_ARM_CRYPTO_SHA1_H
+#define ASM_ARM_CRYPTO_SHA1_H
+
+#include <linux/crypto.h>
+#include <crypto/sha.h>
+
+extern int sha1_update_arm(struct shash_desc *desc, const u8 *data,
+			   unsigned int len);
+
+extern int sha1_finup_arm(struct shash_desc *desc, const u8 *data,
+			   unsigned int len, u8 *out);
+
+#endif
diff --git a/arch/arm/crypto/sha1_glue.c b/arch/arm/crypto/sha1_glue.c
new file mode 100644
index 000000000..6fc73bf87
--- /dev/null
+++ b/arch/arm/crypto/sha1_glue.c
@@ -0,0 +1,95 @@
+/*
+ * Cryptographic API.
+ * Glue code for the SHA1 Secure Hash Algorithm assembler implementation
+ *
+ * This file is based on sha1_generic.c and sha1_ssse3_glue.c
+ *
+ * Copyright (c) Alan Smithee.
+ * Copyright (c) Andrew McDonald <andrew@mcdonald.org.uk>
+ * Copyright (c) Jean-Francois Dive <jef@linuxbe.org>
+ * Copyright (c) Mathias Krause <minipli@googlemail.com>
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License as published by the Free
+ * Software Foundation; either version 2 of the License, or (at your option)
+ * any later version.
+ *
+ */
+
+#include <crypto/internal/hash.h>
+#include <linux/init.h>
+#include <linux/module.h>
+#include <linux/cryptohash.h>
+#include <linux/types.h>
+#include <crypto/sha.h>
+#include <crypto/sha1_base.h>
+#include <asm/byteorder.h>
+
+#include "sha1.h"
+
+asmlinkage void sha1_block_data_order(u32 *digest,
+		const unsigned char *data, unsigned int rounds);
+
+int sha1_update_arm(struct shash_desc *desc, const u8 *data,
+		    unsigned int len)
+{
+	/* make sure casting to sha1_block_fn() is safe */
+	BUILD_BUG_ON(offsetof(struct sha1_state, state) != 0);
+
+	return sha1_base_do_update(desc, data, len,
+				   (sha1_block_fn *)sha1_block_data_order);
+}
+EXPORT_SYMBOL_GPL(sha1_update_arm);
+
+static int sha1_final(struct shash_desc *desc, u8 *out)
+{
+	sha1_base_do_finalize(desc, (sha1_block_fn *)sha1_block_data_order);
+	return sha1_base_finish(desc, out);
+}
+
+int sha1_finup_arm(struct shash_desc *desc, const u8 *data,
+		   unsigned int len, u8 *out)
+{
+	sha1_base_do_update(desc, data, len,
+			    (sha1_block_fn *)sha1_block_data_order);
+	return sha1_final(desc, out);
+}
+EXPORT_SYMBOL_GPL(sha1_finup_arm);
+
+static struct shash_alg alg = {
+	.digestsize	=	SHA1_DIGEST_SIZE,
+	.init		=	sha1_base_init,
+	.update		=	sha1_update_arm,
+	.final		=	sha1_final,
+	.finup		=	sha1_finup_arm,
+	.descsize	=	sizeof(struct sha1_state),
+	.base		=	{
+		.cra_name	=	"sha1",
+		.cra_driver_name=	"sha1-asm",
+		.cra_priority	=	150,
+		.cra_flags	=	CRYPTO_ALG_TYPE_SHASH,
+		.cra_blocksize	=	SHA1_BLOCK_SIZE,
+		.cra_module	=	THIS_MODULE,
+	}
+};
+
+
+static int __init sha1_mod_init(void)
+{
+	return crypto_register_shash(&alg);
+}
+
+
+static void __exit sha1_mod_fini(void)
+{
+	crypto_unregister_shash(&alg);
+}
+
+
+module_init(sha1_mod_init);
+module_exit(sha1_mod_fini);
+
+MODULE_LICENSE("GPL");
+MODULE_DESCRIPTION("SHA1 Secure Hash Algorithm (ARM)");
+MODULE_ALIAS_CRYPTO("sha1");
+MODULE_AUTHOR("David McCullough <ucdevel@gmail.com>");
diff --git a/arch/arm/crypto/sha1_neon_glue.c b/arch/arm/crypto/sha1_neon_glue.c
new file mode 100644
index 000000000..4e22f122f
--- /dev/null
+++ b/arch/arm/crypto/sha1_neon_glue.c
@@ -0,0 +1,110 @@
+/*
+ * Glue code for the SHA1 Secure Hash Algorithm assembler implementation using
+ * ARM NEON instructions.
+ *
+ * Copyright © 2014 Jussi Kivilinna <jussi.kivilinna@iki.fi>
+ *
+ * This file is based on sha1_generic.c and sha1_ssse3_glue.c:
+ *  Copyright (c) Alan Smithee.
+ *  Copyright (c) Andrew McDonald <andrew@mcdonald.org.uk>
+ *  Copyright (c) Jean-Francois Dive <jef@linuxbe.org>
+ *  Copyright (c) Mathias Krause <minipli@googlemail.com>
+ *  Copyright (c) Chandramouli Narayanan <mouli@linux.intel.com>
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License as published by the Free
+ * Software Foundation; either version 2 of the License, or (at your option)
+ * any later version.
+ *
+ */
+
+#include <crypto/internal/hash.h>
+#include <linux/init.h>
+#include <linux/module.h>
+#include <linux/mm.h>
+#include <linux/cryptohash.h>
+#include <linux/types.h>
+#include <crypto/sha.h>
+#include <crypto/sha1_base.h>
+#include <asm/neon.h>
+#include <asm/simd.h>
+
+#include "sha1.h"
+
+asmlinkage void sha1_transform_neon(void *state_h, const char *data,
+				    unsigned int rounds);
+
+static int sha1_neon_update(struct shash_desc *desc, const u8 *data,
+			  unsigned int len)
+{
+	struct sha1_state *sctx = shash_desc_ctx(desc);
+
+	if (!may_use_simd() ||
+	    (sctx->count % SHA1_BLOCK_SIZE) + len < SHA1_BLOCK_SIZE)
+		return sha1_update_arm(desc, data, len);
+
+	kernel_neon_begin();
+	sha1_base_do_update(desc, data, len,
+			    (sha1_block_fn *)sha1_transform_neon);
+	kernel_neon_end();
+
+	return 0;
+}
+
+static int sha1_neon_finup(struct shash_desc *desc, const u8 *data,
+			   unsigned int len, u8 *out)
+{
+	if (!may_use_simd())
+		return sha1_finup_arm(desc, data, len, out);
+
+	kernel_neon_begin();
+	if (len)
+		sha1_base_do_update(desc, data, len,
+				    (sha1_block_fn *)sha1_transform_neon);
+	sha1_base_do_finalize(desc, (sha1_block_fn *)sha1_transform_neon);
+	kernel_neon_end();
+
+	return sha1_base_finish(desc, out);
+}
+
+static int sha1_neon_final(struct shash_desc *desc, u8 *out)
+{
+	return sha1_neon_finup(desc, NULL, 0, out);
+}
+
+static struct shash_alg alg = {
+	.digestsize	=	SHA1_DIGEST_SIZE,
+	.init		=	sha1_base_init,
+	.update		=	sha1_neon_update,
+	.final		=	sha1_neon_final,
+	.finup		=	sha1_neon_finup,
+	.descsize	=	sizeof(struct sha1_state),
+	.base		=	{
+		.cra_name		= "sha1",
+		.cra_driver_name	= "sha1-neon",
+		.cra_priority		= 250,
+		.cra_flags		= CRYPTO_ALG_TYPE_SHASH,
+		.cra_blocksize		= SHA1_BLOCK_SIZE,
+		.cra_module		= THIS_MODULE,
+	}
+};
+
+static int __init sha1_neon_mod_init(void)
+{
+	if (!cpu_has_neon())
+		return -ENODEV;
+
+	return crypto_register_shash(&alg);
+}
+
+static void __exit sha1_neon_mod_fini(void)
+{
+	crypto_unregister_shash(&alg);
+}
+
+module_init(sha1_neon_mod_init);
+module_exit(sha1_neon_mod_fini);
+
+MODULE_LICENSE("GPL");
+MODULE_DESCRIPTION("SHA1 Secure Hash Algorithm, NEON accelerated");
+MODULE_ALIAS_CRYPTO("sha1");
diff --git a/arch/arm/crypto/sha2-ce-core.S b/arch/arm/crypto/sha2-ce-core.S
new file mode 100644
index 000000000..87ec11a5f
--- /dev/null
+++ b/arch/arm/crypto/sha2-ce-core.S
@@ -0,0 +1,125 @@
+/*
+ * sha2-ce-core.S - SHA-224/256 secure hash using ARMv8 Crypto Extensions
+ *
+ * Copyright (C) 2015 Linaro Ltd.
+ * Author: Ard Biesheuvel <ard.biesheuvel@linaro.org>
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2 as
+ * published by the Free Software Foundation.
+ */
+
+#include <linux/linkage.h>
+#include <asm/assembler.h>
+
+	.text
+	.fpu		crypto-neon-fp-armv8
+
+	k0		.req	q7
+	k1		.req	q8
+	rk		.req	r3
+
+	ta0		.req	q9
+	ta1		.req	q10
+	tb0		.req	q10
+	tb1		.req	q9
+
+	dga		.req	q11
+	dgb		.req	q12
+
+	dg0		.req	q13
+	dg1		.req	q14
+	dg2		.req	q15
+
+	.macro		add_only, ev, s0
+	vmov		dg2, dg0
+	.ifnb		\s0
+	vld1.32		{k\ev}, [rk, :128]!
+	.endif
+	sha256h.32	dg0, dg1, tb\ev
+	sha256h2.32	dg1, dg2, tb\ev
+	.ifnb		\s0
+	vadd.u32	ta\ev, q\s0, k\ev
+	.endif
+	.endm
+
+	.macro		add_update, ev, s0, s1, s2, s3
+	sha256su0.32	q\s0, q\s1
+	add_only	\ev, \s1
+	sha256su1.32	q\s0, q\s2, q\s3
+	.endm
+
+	.align		6
+.Lsha256_rcon:
+	.word		0x428a2f98, 0x71374491, 0xb5c0fbcf, 0xe9b5dba5
+	.word		0x3956c25b, 0x59f111f1, 0x923f82a4, 0xab1c5ed5
+	.word		0xd807aa98, 0x12835b01, 0x243185be, 0x550c7dc3
+	.word		0x72be5d74, 0x80deb1fe, 0x9bdc06a7, 0xc19bf174
+	.word		0xe49b69c1, 0xefbe4786, 0x0fc19dc6, 0x240ca1cc
+	.word		0x2de92c6f, 0x4a7484aa, 0x5cb0a9dc, 0x76f988da
+	.word		0x983e5152, 0xa831c66d, 0xb00327c8, 0xbf597fc7
+	.word		0xc6e00bf3, 0xd5a79147, 0x06ca6351, 0x14292967
+	.word		0x27b70a85, 0x2e1b2138, 0x4d2c6dfc, 0x53380d13
+	.word		0x650a7354, 0x766a0abb, 0x81c2c92e, 0x92722c85
+	.word		0xa2bfe8a1, 0xa81a664b, 0xc24b8b70, 0xc76c51a3
+	.word		0xd192e819, 0xd6990624, 0xf40e3585, 0x106aa070
+	.word		0x19a4c116, 0x1e376c08, 0x2748774c, 0x34b0bcb5
+	.word		0x391c0cb3, 0x4ed8aa4a, 0x5b9cca4f, 0x682e6ff3
+	.word		0x748f82ee, 0x78a5636f, 0x84c87814, 0x8cc70208
+	.word		0x90befffa, 0xa4506ceb, 0xbef9a3f7, 0xc67178f2
+
+	/*
+	 * void sha2_ce_transform(struct sha256_state *sst, u8 const *src,
+				  int blocks);
+	 */
+ENTRY(sha2_ce_transform)
+	/* load state */
+	vld1.32		{dga-dgb}, [r0]
+
+	/* load input */
+0:	vld1.32		{q0-q1}, [r1]!
+	vld1.32		{q2-q3}, [r1]!
+	subs		r2, r2, #1
+
+#ifndef CONFIG_CPU_BIG_ENDIAN
+	vrev32.8	q0, q0
+	vrev32.8	q1, q1
+	vrev32.8	q2, q2
+	vrev32.8	q3, q3
+#endif
+
+	/* load first round constant */
+	adr		rk, .Lsha256_rcon
+	vld1.32		{k0}, [rk, :128]!
+
+	vadd.u32	ta0, q0, k0
+	vmov		dg0, dga
+	vmov		dg1, dgb
+
+	add_update	1, 0, 1, 2, 3
+	add_update	0, 1, 2, 3, 0
+	add_update	1, 2, 3, 0, 1
+	add_update	0, 3, 0, 1, 2
+	add_update	1, 0, 1, 2, 3
+	add_update	0, 1, 2, 3, 0
+	add_update	1, 2, 3, 0, 1
+	add_update	0, 3, 0, 1, 2
+	add_update	1, 0, 1, 2, 3
+	add_update	0, 1, 2, 3, 0
+	add_update	1, 2, 3, 0, 1
+	add_update	0, 3, 0, 1, 2
+
+	add_only	1, 1
+	add_only	0, 2
+	add_only	1, 3
+	add_only	0
+
+	/* update state */
+	vadd.u32	dga, dga, dg0
+	vadd.u32	dgb, dgb, dg1
+	bne		0b
+
+	/* store new state */
+	vst1.32		{dga-dgb}, [r0]
+	bx		lr
+ENDPROC(sha2_ce_transform)
diff --git a/arch/arm/crypto/sha2-ce-glue.c b/arch/arm/crypto/sha2-ce-glue.c
new file mode 100644
index 000000000..0755b2d65
--- /dev/null
+++ b/arch/arm/crypto/sha2-ce-glue.c
@@ -0,0 +1,114 @@
+/*
+ * sha2-ce-glue.c - SHA-224/SHA-256 using ARMv8 Crypto Extensions
+ *
+ * Copyright (C) 2015 Linaro Ltd <ard.biesheuvel@linaro.org>
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2 as
+ * published by the Free Software Foundation.
+ */
+
+#include <crypto/internal/hash.h>
+#include <crypto/sha.h>
+#include <crypto/sha256_base.h>
+#include <linux/crypto.h>
+#include <linux/module.h>
+
+#include <asm/hwcap.h>
+#include <asm/simd.h>
+#include <asm/neon.h>
+#include <asm/unaligned.h>
+
+#include "sha256_glue.h"
+
+MODULE_DESCRIPTION("SHA-224/SHA-256 secure hash using ARMv8 Crypto Extensions");
+MODULE_AUTHOR("Ard Biesheuvel <ard.biesheuvel@linaro.org>");
+MODULE_LICENSE("GPL v2");
+
+asmlinkage void sha2_ce_transform(struct sha256_state *sst, u8 const *src,
+				  int blocks);
+
+static int sha2_ce_update(struct shash_desc *desc, const u8 *data,
+			  unsigned int len)
+{
+	struct sha256_state *sctx = shash_desc_ctx(desc);
+
+	if (!may_use_simd() ||
+	    (sctx->count % SHA256_BLOCK_SIZE) + len < SHA256_BLOCK_SIZE)
+		return crypto_sha256_arm_update(desc, data, len);
+
+	kernel_neon_begin();
+	sha256_base_do_update(desc, data, len,
+			      (sha256_block_fn *)sha2_ce_transform);
+	kernel_neon_end();
+
+	return 0;
+}
+
+static int sha2_ce_finup(struct shash_desc *desc, const u8 *data,
+			 unsigned int len, u8 *out)
+{
+	if (!may_use_simd())
+		return crypto_sha256_arm_finup(desc, data, len, out);
+
+	kernel_neon_begin();
+	if (len)
+		sha256_base_do_update(desc, data, len,
+				      (sha256_block_fn *)sha2_ce_transform);
+	sha256_base_do_finalize(desc, (sha256_block_fn *)sha2_ce_transform);
+	kernel_neon_end();
+
+	return sha256_base_finish(desc, out);
+}
+
+static int sha2_ce_final(struct shash_desc *desc, u8 *out)
+{
+	return sha2_ce_finup(desc, NULL, 0, out);
+}
+
+static struct shash_alg algs[] = { {
+	.init			= sha224_base_init,
+	.update			= sha2_ce_update,
+	.final			= sha2_ce_final,
+	.finup			= sha2_ce_finup,
+	.descsize		= sizeof(struct sha256_state),
+	.digestsize		= SHA224_DIGEST_SIZE,
+	.base			= {
+		.cra_name		= "sha224",
+		.cra_driver_name	= "sha224-ce",
+		.cra_priority		= 300,
+		.cra_flags		= CRYPTO_ALG_TYPE_SHASH,
+		.cra_blocksize		= SHA256_BLOCK_SIZE,
+		.cra_module		= THIS_MODULE,
+	}
+}, {
+	.init			= sha256_base_init,
+	.update			= sha2_ce_update,
+	.final			= sha2_ce_final,
+	.finup			= sha2_ce_finup,
+	.descsize		= sizeof(struct sha256_state),
+	.digestsize		= SHA256_DIGEST_SIZE,
+	.base			= {
+		.cra_name		= "sha256",
+		.cra_driver_name	= "sha256-ce",
+		.cra_priority		= 300,
+		.cra_flags		= CRYPTO_ALG_TYPE_SHASH,
+		.cra_blocksize		= SHA256_BLOCK_SIZE,
+		.cra_module		= THIS_MODULE,
+	}
+} };
+
+static int __init sha2_ce_mod_init(void)
+{
+	if (!(elf_hwcap2 & HWCAP2_SHA2))
+		return -ENODEV;
+	return crypto_register_shashes(algs, ARRAY_SIZE(algs));
+}
+
+static void __exit sha2_ce_mod_fini(void)
+{
+	crypto_unregister_shashes(algs, ARRAY_SIZE(algs));
+}
+
+module_init(sha2_ce_mod_init);
+module_exit(sha2_ce_mod_fini);
diff --git a/arch/arm/crypto/sha256-armv4.pl b/arch/arm/crypto/sha256-armv4.pl
new file mode 100644
index 000000000..fac0533ea
--- /dev/null
+++ b/arch/arm/crypto/sha256-armv4.pl
@@ -0,0 +1,716 @@
+#!/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/.
+#
+# Permission to use under GPL terms is granted.
+# ====================================================================
+
+# SHA256 block procedure for ARMv4. May 2007.
+
+# Performance is ~2x better than gcc 3.4 generated code and in "abso-
+# lute" terms is ~2250 cycles per 64-byte block or ~35 cycles per
+# byte [on single-issue Xscale PXA250 core].
+
+# July 2010.
+#
+# Rescheduling for dual-issue pipeline resulted in 22% improvement on
+# Cortex A8 core and ~20 cycles per processed byte.
+
+# February 2011.
+#
+# Profiler-assisted and platform-specific optimization resulted in 16%
+# improvement on Cortex A8 core and ~15.4 cycles per processed byte.
+
+# September 2013.
+#
+# Add NEON implementation. On Cortex A8 it was measured to process one
+# byte in 12.5 cycles or 23% faster than integer-only code. Snapdragon
+# S4 does it in 12.5 cycles too, but it's 50% faster than integer-only
+# code (meaning that latter performs sub-optimally, nothing was done
+# about it).
+
+# May 2014.
+#
+# Add ARMv8 code path performing at 2.0 cpb on Apple A7.
+
+while (($output=shift) && ($output!~/^\w[\w\-]*\.\w+$/)) {}
+open STDOUT,">$output";
+
+$ctx="r0";	$t0="r0";
+$inp="r1";	$t4="r1";
+$len="r2";	$t1="r2";
+$T1="r3";	$t3="r3";
+$A="r4";
+$B="r5";
+$C="r6";
+$D="r7";
+$E="r8";
+$F="r9";
+$G="r10";
+$H="r11";
+@V=($A,$B,$C,$D,$E,$F,$G,$H);
+$t2="r12";
+$Ktbl="r14";
+
+@Sigma0=( 2,13,22);
+@Sigma1=( 6,11,25);
+@sigma0=( 7,18, 3);
+@sigma1=(17,19,10);
+
+sub BODY_00_15 {
+my ($i,$a,$b,$c,$d,$e,$f,$g,$h) = @_;
+
+$code.=<<___ if ($i<16);
+#if __ARM_ARCH__>=7
+	@ ldr	$t1,[$inp],#4			@ $i
+# if $i==15
+	str	$inp,[sp,#17*4]			@ make room for $t4
+# endif
+	eor	$t0,$e,$e,ror#`$Sigma1[1]-$Sigma1[0]`
+	add	$a,$a,$t2			@ h+=Maj(a,b,c) from the past
+	eor	$t0,$t0,$e,ror#`$Sigma1[2]-$Sigma1[0]`	@ Sigma1(e)
+# ifndef __ARMEB__
+	rev	$t1,$t1
+# endif
+#else
+	@ ldrb	$t1,[$inp,#3]			@ $i
+	add	$a,$a,$t2			@ h+=Maj(a,b,c) from the past
+	ldrb	$t2,[$inp,#2]
+	ldrb	$t0,[$inp,#1]
+	orr	$t1,$t1,$t2,lsl#8
+	ldrb	$t2,[$inp],#4
+	orr	$t1,$t1,$t0,lsl#16
+# if $i==15
+	str	$inp,[sp,#17*4]			@ make room for $t4
+# endif
+	eor	$t0,$e,$e,ror#`$Sigma1[1]-$Sigma1[0]`
+	orr	$t1,$t1,$t2,lsl#24
+	eor	$t0,$t0,$e,ror#`$Sigma1[2]-$Sigma1[0]`	@ Sigma1(e)
+#endif
+___
+$code.=<<___;
+	ldr	$t2,[$Ktbl],#4			@ *K256++
+	add	$h,$h,$t1			@ h+=X[i]
+	str	$t1,[sp,#`$i%16`*4]
+	eor	$t1,$f,$g
+	add	$h,$h,$t0,ror#$Sigma1[0]	@ h+=Sigma1(e)
+	and	$t1,$t1,$e
+	add	$h,$h,$t2			@ h+=K256[i]
+	eor	$t1,$t1,$g			@ Ch(e,f,g)
+	eor	$t0,$a,$a,ror#`$Sigma0[1]-$Sigma0[0]`
+	add	$h,$h,$t1			@ h+=Ch(e,f,g)
+#if $i==31
+	and	$t2,$t2,#0xff
+	cmp	$t2,#0xf2			@ done?
+#endif
+#if $i<15
+# if __ARM_ARCH__>=7
+	ldr	$t1,[$inp],#4			@ prefetch
+# else
+	ldrb	$t1,[$inp,#3]
+# endif
+	eor	$t2,$a,$b			@ a^b, b^c in next round
+#else
+	ldr	$t1,[sp,#`($i+2)%16`*4]		@ from future BODY_16_xx
+	eor	$t2,$a,$b			@ a^b, b^c in next round
+	ldr	$t4,[sp,#`($i+15)%16`*4]	@ from future BODY_16_xx
+#endif
+	eor	$t0,$t0,$a,ror#`$Sigma0[2]-$Sigma0[0]`	@ Sigma0(a)
+	and	$t3,$t3,$t2			@ (b^c)&=(a^b)
+	add	$d,$d,$h			@ d+=h
+	eor	$t3,$t3,$b			@ Maj(a,b,c)
+	add	$h,$h,$t0,ror#$Sigma0[0]	@ h+=Sigma0(a)
+	@ add	$h,$h,$t3			@ h+=Maj(a,b,c)
+___
+	($t2,$t3)=($t3,$t2);
+}
+
+sub BODY_16_XX {
+my ($i,$a,$b,$c,$d,$e,$f,$g,$h) = @_;
+
+$code.=<<___;
+	@ ldr	$t1,[sp,#`($i+1)%16`*4]		@ $i
+	@ ldr	$t4,[sp,#`($i+14)%16`*4]
+	mov	$t0,$t1,ror#$sigma0[0]
+	add	$a,$a,$t2			@ h+=Maj(a,b,c) from the past
+	mov	$t2,$t4,ror#$sigma1[0]
+	eor	$t0,$t0,$t1,ror#$sigma0[1]
+	eor	$t2,$t2,$t4,ror#$sigma1[1]
+	eor	$t0,$t0,$t1,lsr#$sigma0[2]	@ sigma0(X[i+1])
+	ldr	$t1,[sp,#`($i+0)%16`*4]
+	eor	$t2,$t2,$t4,lsr#$sigma1[2]	@ sigma1(X[i+14])
+	ldr	$t4,[sp,#`($i+9)%16`*4]
+
+	add	$t2,$t2,$t0
+	eor	$t0,$e,$e,ror#`$Sigma1[1]-$Sigma1[0]`	@ from BODY_00_15
+	add	$t1,$t1,$t2
+	eor	$t0,$t0,$e,ror#`$Sigma1[2]-$Sigma1[0]`	@ Sigma1(e)
+	add	$t1,$t1,$t4			@ X[i]
+___
+	&BODY_00_15(@_);
+}
+
+$code=<<___;
+#ifndef __KERNEL__
+# include "arm_arch.h"
+#else
+# define __ARM_ARCH__ __LINUX_ARM_ARCH__
+# define __ARM_MAX_ARCH__ 7
+#endif
+
+.text
+#if __ARM_ARCH__<7
+.code	32
+#else
+.syntax unified
+# ifdef __thumb2__
+#  define adrl adr
+.thumb
+# else
+.code   32
+# endif
+#endif
+
+.type	K256,%object
+.align	5
+K256:
+.word	0x428a2f98,0x71374491,0xb5c0fbcf,0xe9b5dba5
+.word	0x3956c25b,0x59f111f1,0x923f82a4,0xab1c5ed5
+.word	0xd807aa98,0x12835b01,0x243185be,0x550c7dc3
+.word	0x72be5d74,0x80deb1fe,0x9bdc06a7,0xc19bf174
+.word	0xe49b69c1,0xefbe4786,0x0fc19dc6,0x240ca1cc
+.word	0x2de92c6f,0x4a7484aa,0x5cb0a9dc,0x76f988da
+.word	0x983e5152,0xa831c66d,0xb00327c8,0xbf597fc7
+.word	0xc6e00bf3,0xd5a79147,0x06ca6351,0x14292967
+.word	0x27b70a85,0x2e1b2138,0x4d2c6dfc,0x53380d13
+.word	0x650a7354,0x766a0abb,0x81c2c92e,0x92722c85
+.word	0xa2bfe8a1,0xa81a664b,0xc24b8b70,0xc76c51a3
+.word	0xd192e819,0xd6990624,0xf40e3585,0x106aa070
+.word	0x19a4c116,0x1e376c08,0x2748774c,0x34b0bcb5
+.word	0x391c0cb3,0x4ed8aa4a,0x5b9cca4f,0x682e6ff3
+.word	0x748f82ee,0x78a5636f,0x84c87814,0x8cc70208
+.word	0x90befffa,0xa4506ceb,0xbef9a3f7,0xc67178f2
+.size	K256,.-K256
+.word	0				@ terminator
+#if __ARM_MAX_ARCH__>=7 && !defined(__KERNEL__)
+.LOPENSSL_armcap:
+.word	OPENSSL_armcap_P-sha256_block_data_order
+#endif
+.align	5
+
+.global	sha256_block_data_order
+.type	sha256_block_data_order,%function
+sha256_block_data_order:
+#if __ARM_ARCH__<7
+	sub	r3,pc,#8		@ sha256_block_data_order
+#else
+	adr	r3,sha256_block_data_order
+#endif
+#if __ARM_MAX_ARCH__>=7 && !defined(__KERNEL__)
+	ldr	r12,.LOPENSSL_armcap
+	ldr	r12,[r3,r12]		@ OPENSSL_armcap_P
+	tst	r12,#ARMV8_SHA256
+	bne	.LARMv8
+	tst	r12,#ARMV7_NEON
+	bne	.LNEON
+#endif
+	add	$len,$inp,$len,lsl#6	@ len to point at the end of inp
+	stmdb	sp!,{$ctx,$inp,$len,r4-r11,lr}
+	ldmia	$ctx,{$A,$B,$C,$D,$E,$F,$G,$H}
+	sub	$Ktbl,r3,#256+32	@ K256
+	sub	sp,sp,#16*4		@ alloca(X[16])
+.Loop:
+# if __ARM_ARCH__>=7
+	ldr	$t1,[$inp],#4
+# else
+	ldrb	$t1,[$inp,#3]
+# endif
+	eor	$t3,$B,$C		@ magic
+	eor	$t2,$t2,$t2
+___
+for($i=0;$i<16;$i++)	{ &BODY_00_15($i,@V); unshift(@V,pop(@V)); }
+$code.=".Lrounds_16_xx:\n";
+for (;$i<32;$i++)	{ &BODY_16_XX($i,@V); unshift(@V,pop(@V)); }
+$code.=<<___;
+#if __ARM_ARCH__>=7
+	ite	eq			@ Thumb2 thing, sanity check in ARM
+#endif
+	ldreq	$t3,[sp,#16*4]		@ pull ctx
+	bne	.Lrounds_16_xx
+
+	add	$A,$A,$t2		@ h+=Maj(a,b,c) from the past
+	ldr	$t0,[$t3,#0]
+	ldr	$t1,[$t3,#4]
+	ldr	$t2,[$t3,#8]
+	add	$A,$A,$t0
+	ldr	$t0,[$t3,#12]
+	add	$B,$B,$t1
+	ldr	$t1,[$t3,#16]
+	add	$C,$C,$t2
+	ldr	$t2,[$t3,#20]
+	add	$D,$D,$t0
+	ldr	$t0,[$t3,#24]
+	add	$E,$E,$t1
+	ldr	$t1,[$t3,#28]
+	add	$F,$F,$t2
+	ldr	$inp,[sp,#17*4]		@ pull inp
+	ldr	$t2,[sp,#18*4]		@ pull inp+len
+	add	$G,$G,$t0
+	add	$H,$H,$t1
+	stmia	$t3,{$A,$B,$C,$D,$E,$F,$G,$H}
+	cmp	$inp,$t2
+	sub	$Ktbl,$Ktbl,#256	@ rewind Ktbl
+	bne	.Loop
+
+	add	sp,sp,#`16+3`*4	@ destroy frame
+#if __ARM_ARCH__>=5
+	ldmia	sp!,{r4-r11,pc}
+#else
+	ldmia	sp!,{r4-r11,lr}
+	tst	lr,#1
+	moveq	pc,lr			@ be binary compatible with V4, yet
+	bx	lr			@ interoperable with Thumb ISA:-)
+#endif
+.size	sha256_block_data_order,.-sha256_block_data_order
+___
+######################################################################
+# NEON stuff
+#
+{{{
+my @X=map("q$_",(0..3));
+my ($T0,$T1,$T2,$T3,$T4,$T5)=("q8","q9","q10","q11","d24","d25");
+my $Xfer=$t4;
+my $j=0;
+
+sub Dlo()   { shift=~m|q([1]?[0-9])|?"d".($1*2):"";     }
+sub Dhi()   { shift=~m|q([1]?[0-9])|?"d".($1*2+1):"";   }
+
+sub AUTOLOAD()          # thunk [simplified] x86-style perlasm
+{ my $opcode = $AUTOLOAD; $opcode =~ s/.*:://; $opcode =~ s/_/\./;
+  my $arg = pop;
+    $arg = "#$arg" if ($arg*1 eq $arg);
+    $code .= "\t$opcode\t".join(',',@_,$arg)."\n";
+}
+
+sub Xupdate()
+{ use integer;
+  my $body = shift;
+  my @insns = (&$body,&$body,&$body,&$body);
+  my ($a,$b,$c,$d,$e,$f,$g,$h);
+
+	&vext_8		($T0,@X[0],@X[1],4);	# X[1..4]
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	&vext_8		($T1,@X[2],@X[3],4);	# X[9..12]
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	&vshr_u32	($T2,$T0,$sigma0[0]);
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	&vadd_i32	(@X[0],@X[0],$T1);	# X[0..3] += X[9..12]
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	&vshr_u32	($T1,$T0,$sigma0[2]);
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	&vsli_32	($T2,$T0,32-$sigma0[0]);
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	&vshr_u32	($T3,$T0,$sigma0[1]);
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	&veor		($T1,$T1,$T2);
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	&vsli_32	($T3,$T0,32-$sigma0[1]);
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	  &vshr_u32	($T4,&Dhi(@X[3]),$sigma1[0]);
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	&veor		($T1,$T1,$T3);		# sigma0(X[1..4])
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	  &vsli_32	($T4,&Dhi(@X[3]),32-$sigma1[0]);
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	  &vshr_u32	($T5,&Dhi(@X[3]),$sigma1[2]);
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	&vadd_i32	(@X[0],@X[0],$T1);	# X[0..3] += sigma0(X[1..4])
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	  &veor		($T5,$T5,$T4);
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	  &vshr_u32	($T4,&Dhi(@X[3]),$sigma1[1]);
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	  &vsli_32	($T4,&Dhi(@X[3]),32-$sigma1[1]);
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	  &veor		($T5,$T5,$T4);		# sigma1(X[14..15])
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	&vadd_i32	(&Dlo(@X[0]),&Dlo(@X[0]),$T5);# X[0..1] += sigma1(X[14..15])
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	  &vshr_u32	($T4,&Dlo(@X[0]),$sigma1[0]);
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	  &vsli_32	($T4,&Dlo(@X[0]),32-$sigma1[0]);
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	  &vshr_u32	($T5,&Dlo(@X[0]),$sigma1[2]);
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	  &veor		($T5,$T5,$T4);
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	  &vshr_u32	($T4,&Dlo(@X[0]),$sigma1[1]);
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	&vld1_32	("{$T0}","[$Ktbl,:128]!");
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	  &vsli_32	($T4,&Dlo(@X[0]),32-$sigma1[1]);
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	  &veor		($T5,$T5,$T4);		# sigma1(X[16..17])
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	&vadd_i32	(&Dhi(@X[0]),&Dhi(@X[0]),$T5);# X[2..3] += sigma1(X[16..17])
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	&vadd_i32	($T0,$T0,@X[0]);
+	 while($#insns>=2) { eval(shift(@insns)); }
+	&vst1_32	("{$T0}","[$Xfer,:128]!");
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+
+	push(@X,shift(@X));		# "rotate" X[]
+}
+
+sub Xpreload()
+{ use integer;
+  my $body = shift;
+  my @insns = (&$body,&$body,&$body,&$body);
+  my ($a,$b,$c,$d,$e,$f,$g,$h);
+
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	&vld1_32	("{$T0}","[$Ktbl,:128]!");
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	&vrev32_8	(@X[0],@X[0]);
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	&vadd_i32	($T0,$T0,@X[0]);
+	 foreach (@insns) { eval; }	# remaining instructions
+	&vst1_32	("{$T0}","[$Xfer,:128]!");
+
+	push(@X,shift(@X));		# "rotate" X[]
+}
+
+sub body_00_15 () {
+	(
+	'($a,$b,$c,$d,$e,$f,$g,$h)=@V;'.
+	'&add	($h,$h,$t1)',			# h+=X[i]+K[i]
+	'&eor	($t1,$f,$g)',
+	'&eor	($t0,$e,$e,"ror#".($Sigma1[1]-$Sigma1[0]))',
+	'&add	($a,$a,$t2)',			# h+=Maj(a,b,c) from the past
+	'&and	($t1,$t1,$e)',
+	'&eor	($t2,$t0,$e,"ror#".($Sigma1[2]-$Sigma1[0]))',	# Sigma1(e)
+	'&eor	($t0,$a,$a,"ror#".($Sigma0[1]-$Sigma0[0]))',
+	'&eor	($t1,$t1,$g)',			# Ch(e,f,g)
+	'&add	($h,$h,$t2,"ror#$Sigma1[0]")',	# h+=Sigma1(e)
+	'&eor	($t2,$a,$b)',			# a^b, b^c in next round
+	'&eor	($t0,$t0,$a,"ror#".($Sigma0[2]-$Sigma0[0]))',	# Sigma0(a)
+	'&add	($h,$h,$t1)',			# h+=Ch(e,f,g)
+	'&ldr	($t1,sprintf "[sp,#%d]",4*(($j+1)&15))	if (($j&15)!=15);'.
+	'&ldr	($t1,"[$Ktbl]")				if ($j==15);'.
+	'&ldr	($t1,"[sp,#64]")			if ($j==31)',
+	'&and	($t3,$t3,$t2)',			# (b^c)&=(a^b)
+	'&add	($d,$d,$h)',			# d+=h
+	'&add	($h,$h,$t0,"ror#$Sigma0[0]");'.	# h+=Sigma0(a)
+	'&eor	($t3,$t3,$b)',			# Maj(a,b,c)
+	'$j++;	unshift(@V,pop(@V)); ($t2,$t3)=($t3,$t2);'
+	)
+}
+
+$code.=<<___;
+#if __ARM_MAX_ARCH__>=7
+.arch	armv7-a
+.fpu	neon
+
+.global	sha256_block_data_order_neon
+.type	sha256_block_data_order_neon,%function
+.align	4
+sha256_block_data_order_neon:
+.LNEON:
+	stmdb	sp!,{r4-r12,lr}
+
+	sub	$H,sp,#16*4+16
+	adrl	$Ktbl,K256
+	bic	$H,$H,#15		@ align for 128-bit stores
+	mov	$t2,sp
+	mov	sp,$H			@ alloca
+	add	$len,$inp,$len,lsl#6	@ len to point at the end of inp
+
+	vld1.8		{@X[0]},[$inp]!
+	vld1.8		{@X[1]},[$inp]!
+	vld1.8		{@X[2]},[$inp]!
+	vld1.8		{@X[3]},[$inp]!
+	vld1.32		{$T0},[$Ktbl,:128]!
+	vld1.32		{$T1},[$Ktbl,:128]!
+	vld1.32		{$T2},[$Ktbl,:128]!
+	vld1.32		{$T3},[$Ktbl,:128]!
+	vrev32.8	@X[0],@X[0]		@ yes, even on
+	str		$ctx,[sp,#64]
+	vrev32.8	@X[1],@X[1]		@ big-endian
+	str		$inp,[sp,#68]
+	mov		$Xfer,sp
+	vrev32.8	@X[2],@X[2]
+	str		$len,[sp,#72]
+	vrev32.8	@X[3],@X[3]
+	str		$t2,[sp,#76]		@ save original sp
+	vadd.i32	$T0,$T0,@X[0]
+	vadd.i32	$T1,$T1,@X[1]
+	vst1.32		{$T0},[$Xfer,:128]!
+	vadd.i32	$T2,$T2,@X[2]
+	vst1.32		{$T1},[$Xfer,:128]!
+	vadd.i32	$T3,$T3,@X[3]
+	vst1.32		{$T2},[$Xfer,:128]!
+	vst1.32		{$T3},[$Xfer,:128]!
+
+	ldmia		$ctx,{$A-$H}
+	sub		$Xfer,$Xfer,#64
+	ldr		$t1,[sp,#0]
+	eor		$t2,$t2,$t2
+	eor		$t3,$B,$C
+	b		.L_00_48
+
+.align	4
+.L_00_48:
+___
+	&Xupdate(\&body_00_15);
+	&Xupdate(\&body_00_15);
+	&Xupdate(\&body_00_15);
+	&Xupdate(\&body_00_15);
+$code.=<<___;
+	teq	$t1,#0				@ check for K256 terminator
+	ldr	$t1,[sp,#0]
+	sub	$Xfer,$Xfer,#64
+	bne	.L_00_48
+
+	ldr		$inp,[sp,#68]
+	ldr		$t0,[sp,#72]
+	sub		$Ktbl,$Ktbl,#256	@ rewind $Ktbl
+	teq		$inp,$t0
+	it		eq
+	subeq		$inp,$inp,#64		@ avoid SEGV
+	vld1.8		{@X[0]},[$inp]!		@ load next input block
+	vld1.8		{@X[1]},[$inp]!
+	vld1.8		{@X[2]},[$inp]!
+	vld1.8		{@X[3]},[$inp]!
+	it		ne
+	strne		$inp,[sp,#68]
+	mov		$Xfer,sp
+___
+	&Xpreload(\&body_00_15);
+	&Xpreload(\&body_00_15);
+	&Xpreload(\&body_00_15);
+	&Xpreload(\&body_00_15);
+$code.=<<___;
+	ldr	$t0,[$t1,#0]
+	add	$A,$A,$t2			@ h+=Maj(a,b,c) from the past
+	ldr	$t2,[$t1,#4]
+	ldr	$t3,[$t1,#8]
+	ldr	$t4,[$t1,#12]
+	add	$A,$A,$t0			@ accumulate
+	ldr	$t0,[$t1,#16]
+	add	$B,$B,$t2
+	ldr	$t2,[$t1,#20]
+	add	$C,$C,$t3
+	ldr	$t3,[$t1,#24]
+	add	$D,$D,$t4
+	ldr	$t4,[$t1,#28]
+	add	$E,$E,$t0
+	str	$A,[$t1],#4
+	add	$F,$F,$t2
+	str	$B,[$t1],#4
+	add	$G,$G,$t3
+	str	$C,[$t1],#4
+	add	$H,$H,$t4
+	str	$D,[$t1],#4
+	stmia	$t1,{$E-$H}
+
+	ittte	ne
+	movne	$Xfer,sp
+	ldrne	$t1,[sp,#0]
+	eorne	$t2,$t2,$t2
+	ldreq	sp,[sp,#76]			@ restore original sp
+	itt	ne
+	eorne	$t3,$B,$C
+	bne	.L_00_48
+
+	ldmia	sp!,{r4-r12,pc}
+.size	sha256_block_data_order_neon,.-sha256_block_data_order_neon
+#endif
+___
+}}}
+######################################################################
+# ARMv8 stuff
+#
+{{{
+my ($ABCD,$EFGH,$abcd)=map("q$_",(0..2));
+my @MSG=map("q$_",(8..11));
+my ($W0,$W1,$ABCD_SAVE,$EFGH_SAVE)=map("q$_",(12..15));
+my $Ktbl="r3";
+
+$code.=<<___;
+#if __ARM_MAX_ARCH__>=7 && !defined(__KERNEL__)
+
+# ifdef __thumb2__
+#  define INST(a,b,c,d)	.byte	c,d|0xc,a,b
+# else
+#  define INST(a,b,c,d)	.byte	a,b,c,d
+# endif
+
+.type	sha256_block_data_order_armv8,%function
+.align	5
+sha256_block_data_order_armv8:
+.LARMv8:
+	vld1.32	{$ABCD,$EFGH},[$ctx]
+# ifdef __thumb2__
+	adr	$Ktbl,.LARMv8
+	sub	$Ktbl,$Ktbl,#.LARMv8-K256
+# else
+	adrl	$Ktbl,K256
+# endif
+	add	$len,$inp,$len,lsl#6	@ len to point at the end of inp
+
+.Loop_v8:
+	vld1.8		{@MSG[0]-@MSG[1]},[$inp]!
+	vld1.8		{@MSG[2]-@MSG[3]},[$inp]!
+	vld1.32		{$W0},[$Ktbl]!
+	vrev32.8	@MSG[0],@MSG[0]
+	vrev32.8	@MSG[1],@MSG[1]
+	vrev32.8	@MSG[2],@MSG[2]
+	vrev32.8	@MSG[3],@MSG[3]
+	vmov		$ABCD_SAVE,$ABCD	@ offload
+	vmov		$EFGH_SAVE,$EFGH
+	teq		$inp,$len
+___
+for($i=0;$i<12;$i++) {
+$code.=<<___;
+	vld1.32		{$W1},[$Ktbl]!
+	vadd.i32	$W0,$W0,@MSG[0]
+	sha256su0	@MSG[0],@MSG[1]
+	vmov		$abcd,$ABCD
+	sha256h		$ABCD,$EFGH,$W0
+	sha256h2	$EFGH,$abcd,$W0
+	sha256su1	@MSG[0],@MSG[2],@MSG[3]
+___
+	($W0,$W1)=($W1,$W0);	push(@MSG,shift(@MSG));
+}
+$code.=<<___;
+	vld1.32		{$W1},[$Ktbl]!
+	vadd.i32	$W0,$W0,@MSG[0]
+	vmov		$abcd,$ABCD
+	sha256h		$ABCD,$EFGH,$W0
+	sha256h2	$EFGH,$abcd,$W0
+
+	vld1.32		{$W0},[$Ktbl]!
+	vadd.i32	$W1,$W1,@MSG[1]
+	vmov		$abcd,$ABCD
+	sha256h		$ABCD,$EFGH,$W1
+	sha256h2	$EFGH,$abcd,$W1
+
+	vld1.32		{$W1},[$Ktbl]
+	vadd.i32	$W0,$W0,@MSG[2]
+	sub		$Ktbl,$Ktbl,#256-16	@ rewind
+	vmov		$abcd,$ABCD
+	sha256h		$ABCD,$EFGH,$W0
+	sha256h2	$EFGH,$abcd,$W0
+
+	vadd.i32	$W1,$W1,@MSG[3]
+	vmov		$abcd,$ABCD
+	sha256h		$ABCD,$EFGH,$W1
+	sha256h2	$EFGH,$abcd,$W1
+
+	vadd.i32	$ABCD,$ABCD,$ABCD_SAVE
+	vadd.i32	$EFGH,$EFGH,$EFGH_SAVE
+	it		ne
+	bne		.Loop_v8
+
+	vst1.32		{$ABCD,$EFGH},[$ctx]
+
+	ret		@ bx lr
+.size	sha256_block_data_order_armv8,.-sha256_block_data_order_armv8
+#endif
+___
+}}}
+$code.=<<___;
+.asciz  "SHA256 block transform for ARMv4/NEON/ARMv8, CRYPTOGAMS by <appro\@openssl.org>"
+.align	2
+#if __ARM_MAX_ARCH__>=7 && !defined(__KERNEL__)
+.comm   OPENSSL_armcap_P,4,4
+#endif
+___
+
+open SELF,$0;
+while(<SELF>) {
+	next if (/^#!/);
+	last if (!s/^#/@/ and !/^$/);
+	print;
+}
+close SELF;
+
+{   my  %opcode = (
+	"sha256h"	=> 0xf3000c40,	"sha256h2"	=> 0xf3100c40,
+	"sha256su0"	=> 0xf3ba03c0,	"sha256su1"	=> 0xf3200c40	);
+
+    sub unsha256 {
+	my ($mnemonic,$arg)=@_;
+
+	if ($arg =~ m/q([0-9]+)(?:,\s*q([0-9]+))?,\s*q([0-9]+)/o) {
+	    my $word = $opcode{$mnemonic}|(($1&7)<<13)|(($1&8)<<19)
+					 |(($2&7)<<17)|(($2&8)<<4)
+					 |(($3&7)<<1) |(($3&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 "INST(0x%02x,0x%02x,0x%02x,0x%02x)\t@ %s %s",
+			$word&0xff,($word>>8)&0xff,
+			($word>>16)&0xff,($word>>24)&0xff,
+			$mnemonic,$arg;
+	}
+    }
+}
+
+foreach (split($/,$code)) {
+
+	s/\`([^\`]*)\`/eval $1/geo;
+
+	s/\b(sha256\w+)\s+(q.*)/unsha256($1,$2)/geo;
+
+	s/\bret\b/bx	lr/go		or
+	s/\bbx\s+lr\b/.word\t0xe12fff1e/go;	# make it possible to compile with -march=armv4
+
+	print $_,"\n";
+}
+
+close STDOUT; # enforce flush
diff --git a/arch/arm/crypto/sha256-core.S_shipped b/arch/arm/crypto/sha256-core.S_shipped
new file mode 100644
index 000000000..555a1a8ee
--- /dev/null
+++ b/arch/arm/crypto/sha256-core.S_shipped
@@ -0,0 +1,2808 @@
+
+@ ====================================================================
+@ 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/.
+@
+@ Permission to use under GPL terms is granted.
+@ ====================================================================
+
+@ SHA256 block procedure for ARMv4. May 2007.
+
+@ Performance is ~2x better than gcc 3.4 generated code and in "abso-
+@ lute" terms is ~2250 cycles per 64-byte block or ~35 cycles per
+@ byte [on single-issue Xscale PXA250 core].
+
+@ July 2010.
+@
+@ Rescheduling for dual-issue pipeline resulted in 22% improvement on
+@ Cortex A8 core and ~20 cycles per processed byte.
+
+@ February 2011.
+@
+@ Profiler-assisted and platform-specific optimization resulted in 16%
+@ improvement on Cortex A8 core and ~15.4 cycles per processed byte.
+
+@ September 2013.
+@
+@ Add NEON implementation. On Cortex A8 it was measured to process one
+@ byte in 12.5 cycles or 23% faster than integer-only code. Snapdragon
+@ S4 does it in 12.5 cycles too, but it's 50% faster than integer-only
+@ code (meaning that latter performs sub-optimally, nothing was done
+@ about it).
+
+@ May 2014.
+@
+@ Add ARMv8 code path performing at 2.0 cpb on Apple A7.
+
+#ifndef __KERNEL__
+# include "arm_arch.h"
+#else
+# define __ARM_ARCH__ __LINUX_ARM_ARCH__
+# define __ARM_MAX_ARCH__ 7
+#endif
+
+.text
+#if __ARM_ARCH__<7
+.code	32
+#else
+.syntax unified
+# ifdef __thumb2__
+#  define adrl adr
+.thumb
+# else
+.code   32
+# endif
+#endif
+
+.type	K256,%object
+.align	5
+K256:
+.word	0x428a2f98,0x71374491,0xb5c0fbcf,0xe9b5dba5
+.word	0x3956c25b,0x59f111f1,0x923f82a4,0xab1c5ed5
+.word	0xd807aa98,0x12835b01,0x243185be,0x550c7dc3
+.word	0x72be5d74,0x80deb1fe,0x9bdc06a7,0xc19bf174
+.word	0xe49b69c1,0xefbe4786,0x0fc19dc6,0x240ca1cc
+.word	0x2de92c6f,0x4a7484aa,0x5cb0a9dc,0x76f988da
+.word	0x983e5152,0xa831c66d,0xb00327c8,0xbf597fc7
+.word	0xc6e00bf3,0xd5a79147,0x06ca6351,0x14292967
+.word	0x27b70a85,0x2e1b2138,0x4d2c6dfc,0x53380d13
+.word	0x650a7354,0x766a0abb,0x81c2c92e,0x92722c85
+.word	0xa2bfe8a1,0xa81a664b,0xc24b8b70,0xc76c51a3
+.word	0xd192e819,0xd6990624,0xf40e3585,0x106aa070
+.word	0x19a4c116,0x1e376c08,0x2748774c,0x34b0bcb5
+.word	0x391c0cb3,0x4ed8aa4a,0x5b9cca4f,0x682e6ff3
+.word	0x748f82ee,0x78a5636f,0x84c87814,0x8cc70208
+.word	0x90befffa,0xa4506ceb,0xbef9a3f7,0xc67178f2
+.size	K256,.-K256
+.word	0				@ terminator
+#if __ARM_MAX_ARCH__>=7 && !defined(__KERNEL__)
+.LOPENSSL_armcap:
+.word	OPENSSL_armcap_P-sha256_block_data_order
+#endif
+.align	5
+
+.global	sha256_block_data_order
+.type	sha256_block_data_order,%function
+sha256_block_data_order:
+#if __ARM_ARCH__<7
+	sub	r3,pc,#8		@ sha256_block_data_order
+#else
+	adr	r3,sha256_block_data_order
+#endif
+#if __ARM_MAX_ARCH__>=7 && !defined(__KERNEL__)
+	ldr	r12,.LOPENSSL_armcap
+	ldr	r12,[r3,r12]		@ OPENSSL_armcap_P
+	tst	r12,#ARMV8_SHA256
+	bne	.LARMv8
+	tst	r12,#ARMV7_NEON
+	bne	.LNEON
+#endif
+	add	r2,r1,r2,lsl#6	@ len to point at the end of inp
+	stmdb	sp!,{r0,r1,r2,r4-r11,lr}
+	ldmia	r0,{r4,r5,r6,r7,r8,r9,r10,r11}
+	sub	r14,r3,#256+32	@ K256
+	sub	sp,sp,#16*4		@ alloca(X[16])
+.Loop:
+# if __ARM_ARCH__>=7
+	ldr	r2,[r1],#4
+# else
+	ldrb	r2,[r1,#3]
+# endif
+	eor	r3,r5,r6		@ magic
+	eor	r12,r12,r12
+#if __ARM_ARCH__>=7
+	@ ldr	r2,[r1],#4			@ 0
+# if 0==15
+	str	r1,[sp,#17*4]			@ make room for r1
+# endif
+	eor	r0,r8,r8,ror#5
+	add	r4,r4,r12			@ h+=Maj(a,b,c) from the past
+	eor	r0,r0,r8,ror#19	@ Sigma1(e)
+# ifndef __ARMEB__
+	rev	r2,r2
+# endif
+#else
+	@ ldrb	r2,[r1,#3]			@ 0
+	add	r4,r4,r12			@ h+=Maj(a,b,c) from the past
+	ldrb	r12,[r1,#2]
+	ldrb	r0,[r1,#1]
+	orr	r2,r2,r12,lsl#8
+	ldrb	r12,[r1],#4
+	orr	r2,r2,r0,lsl#16
+# if 0==15
+	str	r1,[sp,#17*4]			@ make room for r1
+# endif
+	eor	r0,r8,r8,ror#5
+	orr	r2,r2,r12,lsl#24
+	eor	r0,r0,r8,ror#19	@ Sigma1(e)
+#endif
+	ldr	r12,[r14],#4			@ *K256++
+	add	r11,r11,r2			@ h+=X[i]
+	str	r2,[sp,#0*4]
+	eor	r2,r9,r10
+	add	r11,r11,r0,ror#6	@ h+=Sigma1(e)
+	and	r2,r2,r8
+	add	r11,r11,r12			@ h+=K256[i]
+	eor	r2,r2,r10			@ Ch(e,f,g)
+	eor	r0,r4,r4,ror#11
+	add	r11,r11,r2			@ h+=Ch(e,f,g)
+#if 0==31
+	and	r12,r12,#0xff
+	cmp	r12,#0xf2			@ done?
+#endif
+#if 0<15
+# if __ARM_ARCH__>=7
+	ldr	r2,[r1],#4			@ prefetch
+# else
+	ldrb	r2,[r1,#3]
+# endif
+	eor	r12,r4,r5			@ a^b, b^c in next round
+#else
+	ldr	r2,[sp,#2*4]		@ from future BODY_16_xx
+	eor	r12,r4,r5			@ a^b, b^c in next round
+	ldr	r1,[sp,#15*4]	@ from future BODY_16_xx
+#endif
+	eor	r0,r0,r4,ror#20	@ Sigma0(a)
+	and	r3,r3,r12			@ (b^c)&=(a^b)
+	add	r7,r7,r11			@ d+=h
+	eor	r3,r3,r5			@ Maj(a,b,c)
+	add	r11,r11,r0,ror#2	@ h+=Sigma0(a)
+	@ add	r11,r11,r3			@ h+=Maj(a,b,c)
+#if __ARM_ARCH__>=7
+	@ ldr	r2,[r1],#4			@ 1
+# if 1==15
+	str	r1,[sp,#17*4]			@ make room for r1
+# endif
+	eor	r0,r7,r7,ror#5
+	add	r11,r11,r3			@ h+=Maj(a,b,c) from the past
+	eor	r0,r0,r7,ror#19	@ Sigma1(e)
+# ifndef __ARMEB__
+	rev	r2,r2
+# endif
+#else
+	@ ldrb	r2,[r1,#3]			@ 1
+	add	r11,r11,r3			@ h+=Maj(a,b,c) from the past
+	ldrb	r3,[r1,#2]
+	ldrb	r0,[r1,#1]
+	orr	r2,r2,r3,lsl#8
+	ldrb	r3,[r1],#4
+	orr	r2,r2,r0,lsl#16
+# if 1==15
+	str	r1,[sp,#17*4]			@ make room for r1
+# endif
+	eor	r0,r7,r7,ror#5
+	orr	r2,r2,r3,lsl#24
+	eor	r0,r0,r7,ror#19	@ Sigma1(e)
+#endif
+	ldr	r3,[r14],#4			@ *K256++
+	add	r10,r10,r2			@ h+=X[i]
+	str	r2,[sp,#1*4]
+	eor	r2,r8,r9
+	add	r10,r10,r0,ror#6	@ h+=Sigma1(e)
+	and	r2,r2,r7
+	add	r10,r10,r3			@ h+=K256[i]
+	eor	r2,r2,r9			@ Ch(e,f,g)
+	eor	r0,r11,r11,ror#11
+	add	r10,r10,r2			@ h+=Ch(e,f,g)
+#if 1==31
+	and	r3,r3,#0xff
+	cmp	r3,#0xf2			@ done?
+#endif
+#if 1<15
+# if __ARM_ARCH__>=7
+	ldr	r2,[r1],#4			@ prefetch
+# else
+	ldrb	r2,[r1,#3]
+# endif
+	eor	r3,r11,r4			@ a^b, b^c in next round
+#else
+	ldr	r2,[sp,#3*4]		@ from future BODY_16_xx
+	eor	r3,r11,r4			@ a^b, b^c in next round
+	ldr	r1,[sp,#0*4]	@ from future BODY_16_xx
+#endif
+	eor	r0,r0,r11,ror#20	@ Sigma0(a)
+	and	r12,r12,r3			@ (b^c)&=(a^b)
+	add	r6,r6,r10			@ d+=h
+	eor	r12,r12,r4			@ Maj(a,b,c)
+	add	r10,r10,r0,ror#2	@ h+=Sigma0(a)
+	@ add	r10,r10,r12			@ h+=Maj(a,b,c)
+#if __ARM_ARCH__>=7
+	@ ldr	r2,[r1],#4			@ 2
+# if 2==15
+	str	r1,[sp,#17*4]			@ make room for r1
+# endif
+	eor	r0,r6,r6,ror#5
+	add	r10,r10,r12			@ h+=Maj(a,b,c) from the past
+	eor	r0,r0,r6,ror#19	@ Sigma1(e)
+# ifndef __ARMEB__
+	rev	r2,r2
+# endif
+#else
+	@ ldrb	r2,[r1,#3]			@ 2
+	add	r10,r10,r12			@ h+=Maj(a,b,c) from the past
+	ldrb	r12,[r1,#2]
+	ldrb	r0,[r1,#1]
+	orr	r2,r2,r12,lsl#8
+	ldrb	r12,[r1],#4
+	orr	r2,r2,r0,lsl#16
+# if 2==15
+	str	r1,[sp,#17*4]			@ make room for r1
+# endif
+	eor	r0,r6,r6,ror#5
+	orr	r2,r2,r12,lsl#24
+	eor	r0,r0,r6,ror#19	@ Sigma1(e)
+#endif
+	ldr	r12,[r14],#4			@ *K256++
+	add	r9,r9,r2			@ h+=X[i]
+	str	r2,[sp,#2*4]
+	eor	r2,r7,r8
+	add	r9,r9,r0,ror#6	@ h+=Sigma1(e)
+	and	r2,r2,r6
+	add	r9,r9,r12			@ h+=K256[i]
+	eor	r2,r2,r8			@ Ch(e,f,g)
+	eor	r0,r10,r10,ror#11
+	add	r9,r9,r2			@ h+=Ch(e,f,g)
+#if 2==31
+	and	r12,r12,#0xff
+	cmp	r12,#0xf2			@ done?
+#endif
+#if 2<15
+# if __ARM_ARCH__>=7
+	ldr	r2,[r1],#4			@ prefetch
+# else
+	ldrb	r2,[r1,#3]
+# endif
+	eor	r12,r10,r11			@ a^b, b^c in next round
+#else
+	ldr	r2,[sp,#4*4]		@ from future BODY_16_xx
+	eor	r12,r10,r11			@ a^b, b^c in next round
+	ldr	r1,[sp,#1*4]	@ from future BODY_16_xx
+#endif
+	eor	r0,r0,r10,ror#20	@ Sigma0(a)
+	and	r3,r3,r12			@ (b^c)&=(a^b)
+	add	r5,r5,r9			@ d+=h
+	eor	r3,r3,r11			@ Maj(a,b,c)
+	add	r9,r9,r0,ror#2	@ h+=Sigma0(a)
+	@ add	r9,r9,r3			@ h+=Maj(a,b,c)
+#if __ARM_ARCH__>=7
+	@ ldr	r2,[r1],#4			@ 3
+# if 3==15
+	str	r1,[sp,#17*4]			@ make room for r1
+# endif
+	eor	r0,r5,r5,ror#5
+	add	r9,r9,r3			@ h+=Maj(a,b,c) from the past
+	eor	r0,r0,r5,ror#19	@ Sigma1(e)
+# ifndef __ARMEB__
+	rev	r2,r2
+# endif
+#else
+	@ ldrb	r2,[r1,#3]			@ 3
+	add	r9,r9,r3			@ h+=Maj(a,b,c) from the past
+	ldrb	r3,[r1,#2]
+	ldrb	r0,[r1,#1]
+	orr	r2,r2,r3,lsl#8
+	ldrb	r3,[r1],#4
+	orr	r2,r2,r0,lsl#16
+# if 3==15
+	str	r1,[sp,#17*4]			@ make room for r1
+# endif
+	eor	r0,r5,r5,ror#5
+	orr	r2,r2,r3,lsl#24
+	eor	r0,r0,r5,ror#19	@ Sigma1(e)
+#endif
+	ldr	r3,[r14],#4			@ *K256++
+	add	r8,r8,r2			@ h+=X[i]
+	str	r2,[sp,#3*4]
+	eor	r2,r6,r7
+	add	r8,r8,r0,ror#6	@ h+=Sigma1(e)
+	and	r2,r2,r5
+	add	r8,r8,r3			@ h+=K256[i]
+	eor	r2,r2,r7			@ Ch(e,f,g)
+	eor	r0,r9,r9,ror#11
+	add	r8,r8,r2			@ h+=Ch(e,f,g)
+#if 3==31
+	and	r3,r3,#0xff
+	cmp	r3,#0xf2			@ done?
+#endif
+#if 3<15
+# if __ARM_ARCH__>=7
+	ldr	r2,[r1],#4			@ prefetch
+# else
+	ldrb	r2,[r1,#3]
+# endif
+	eor	r3,r9,r10			@ a^b, b^c in next round
+#else
+	ldr	r2,[sp,#5*4]		@ from future BODY_16_xx
+	eor	r3,r9,r10			@ a^b, b^c in next round
+	ldr	r1,[sp,#2*4]	@ from future BODY_16_xx
+#endif
+	eor	r0,r0,r9,ror#20	@ Sigma0(a)
+	and	r12,r12,r3			@ (b^c)&=(a^b)
+	add	r4,r4,r8			@ d+=h
+	eor	r12,r12,r10			@ Maj(a,b,c)
+	add	r8,r8,r0,ror#2	@ h+=Sigma0(a)
+	@ add	r8,r8,r12			@ h+=Maj(a,b,c)
+#if __ARM_ARCH__>=7
+	@ ldr	r2,[r1],#4			@ 4
+# if 4==15
+	str	r1,[sp,#17*4]			@ make room for r1
+# endif
+	eor	r0,r4,r4,ror#5
+	add	r8,r8,r12			@ h+=Maj(a,b,c) from the past
+	eor	r0,r0,r4,ror#19	@ Sigma1(e)
+# ifndef __ARMEB__
+	rev	r2,r2
+# endif
+#else
+	@ ldrb	r2,[r1,#3]			@ 4
+	add	r8,r8,r12			@ h+=Maj(a,b,c) from the past
+	ldrb	r12,[r1,#2]
+	ldrb	r0,[r1,#1]
+	orr	r2,r2,r12,lsl#8
+	ldrb	r12,[r1],#4
+	orr	r2,r2,r0,lsl#16
+# if 4==15
+	str	r1,[sp,#17*4]			@ make room for r1
+# endif
+	eor	r0,r4,r4,ror#5
+	orr	r2,r2,r12,lsl#24
+	eor	r0,r0,r4,ror#19	@ Sigma1(e)
+#endif
+	ldr	r12,[r14],#4			@ *K256++
+	add	r7,r7,r2			@ h+=X[i]
+	str	r2,[sp,#4*4]
+	eor	r2,r5,r6
+	add	r7,r7,r0,ror#6	@ h+=Sigma1(e)
+	and	r2,r2,r4
+	add	r7,r7,r12			@ h+=K256[i]
+	eor	r2,r2,r6			@ Ch(e,f,g)
+	eor	r0,r8,r8,ror#11
+	add	r7,r7,r2			@ h+=Ch(e,f,g)
+#if 4==31
+	and	r12,r12,#0xff
+	cmp	r12,#0xf2			@ done?
+#endif
+#if 4<15
+# if __ARM_ARCH__>=7
+	ldr	r2,[r1],#4			@ prefetch
+# else
+	ldrb	r2,[r1,#3]
+# endif
+	eor	r12,r8,r9			@ a^b, b^c in next round
+#else
+	ldr	r2,[sp,#6*4]		@ from future BODY_16_xx
+	eor	r12,r8,r9			@ a^b, b^c in next round
+	ldr	r1,[sp,#3*4]	@ from future BODY_16_xx
+#endif
+	eor	r0,r0,r8,ror#20	@ Sigma0(a)
+	and	r3,r3,r12			@ (b^c)&=(a^b)
+	add	r11,r11,r7			@ d+=h
+	eor	r3,r3,r9			@ Maj(a,b,c)
+	add	r7,r7,r0,ror#2	@ h+=Sigma0(a)
+	@ add	r7,r7,r3			@ h+=Maj(a,b,c)
+#if __ARM_ARCH__>=7
+	@ ldr	r2,[r1],#4			@ 5
+# if 5==15
+	str	r1,[sp,#17*4]			@ make room for r1
+# endif
+	eor	r0,r11,r11,ror#5
+	add	r7,r7,r3			@ h+=Maj(a,b,c) from the past
+	eor	r0,r0,r11,ror#19	@ Sigma1(e)
+# ifndef __ARMEB__
+	rev	r2,r2
+# endif
+#else
+	@ ldrb	r2,[r1,#3]			@ 5
+	add	r7,r7,r3			@ h+=Maj(a,b,c) from the past
+	ldrb	r3,[r1,#2]
+	ldrb	r0,[r1,#1]
+	orr	r2,r2,r3,lsl#8
+	ldrb	r3,[r1],#4
+	orr	r2,r2,r0,lsl#16
+# if 5==15
+	str	r1,[sp,#17*4]			@ make room for r1
+# endif
+	eor	r0,r11,r11,ror#5
+	orr	r2,r2,r3,lsl#24
+	eor	r0,r0,r11,ror#19	@ Sigma1(e)
+#endif
+	ldr	r3,[r14],#4			@ *K256++
+	add	r6,r6,r2			@ h+=X[i]
+	str	r2,[sp,#5*4]
+	eor	r2,r4,r5
+	add	r6,r6,r0,ror#6	@ h+=Sigma1(e)
+	and	r2,r2,r11
+	add	r6,r6,r3			@ h+=K256[i]
+	eor	r2,r2,r5			@ Ch(e,f,g)
+	eor	r0,r7,r7,ror#11
+	add	r6,r6,r2			@ h+=Ch(e,f,g)
+#if 5==31
+	and	r3,r3,#0xff
+	cmp	r3,#0xf2			@ done?
+#endif
+#if 5<15
+# if __ARM_ARCH__>=7
+	ldr	r2,[r1],#4			@ prefetch
+# else
+	ldrb	r2,[r1,#3]
+# endif
+	eor	r3,r7,r8			@ a^b, b^c in next round
+#else
+	ldr	r2,[sp,#7*4]		@ from future BODY_16_xx
+	eor	r3,r7,r8			@ a^b, b^c in next round
+	ldr	r1,[sp,#4*4]	@ from future BODY_16_xx
+#endif
+	eor	r0,r0,r7,ror#20	@ Sigma0(a)
+	and	r12,r12,r3			@ (b^c)&=(a^b)
+	add	r10,r10,r6			@ d+=h
+	eor	r12,r12,r8			@ Maj(a,b,c)
+	add	r6,r6,r0,ror#2	@ h+=Sigma0(a)
+	@ add	r6,r6,r12			@ h+=Maj(a,b,c)
+#if __ARM_ARCH__>=7
+	@ ldr	r2,[r1],#4			@ 6
+# if 6==15
+	str	r1,[sp,#17*4]			@ make room for r1
+# endif
+	eor	r0,r10,r10,ror#5
+	add	r6,r6,r12			@ h+=Maj(a,b,c) from the past
+	eor	r0,r0,r10,ror#19	@ Sigma1(e)
+# ifndef __ARMEB__
+	rev	r2,r2
+# endif
+#else
+	@ ldrb	r2,[r1,#3]			@ 6
+	add	r6,r6,r12			@ h+=Maj(a,b,c) from the past
+	ldrb	r12,[r1,#2]
+	ldrb	r0,[r1,#1]
+	orr	r2,r2,r12,lsl#8
+	ldrb	r12,[r1],#4
+	orr	r2,r2,r0,lsl#16
+# if 6==15
+	str	r1,[sp,#17*4]			@ make room for r1
+# endif
+	eor	r0,r10,r10,ror#5
+	orr	r2,r2,r12,lsl#24
+	eor	r0,r0,r10,ror#19	@ Sigma1(e)
+#endif
+	ldr	r12,[r14],#4			@ *K256++
+	add	r5,r5,r2			@ h+=X[i]
+	str	r2,[sp,#6*4]
+	eor	r2,r11,r4
+	add	r5,r5,r0,ror#6	@ h+=Sigma1(e)
+	and	r2,r2,r10
+	add	r5,r5,r12			@ h+=K256[i]
+	eor	r2,r2,r4			@ Ch(e,f,g)
+	eor	r0,r6,r6,ror#11
+	add	r5,r5,r2			@ h+=Ch(e,f,g)
+#if 6==31
+	and	r12,r12,#0xff
+	cmp	r12,#0xf2			@ done?
+#endif
+#if 6<15
+# if __ARM_ARCH__>=7
+	ldr	r2,[r1],#4			@ prefetch
+# else
+	ldrb	r2,[r1,#3]
+# endif
+	eor	r12,r6,r7			@ a^b, b^c in next round
+#else
+	ldr	r2,[sp,#8*4]		@ from future BODY_16_xx
+	eor	r12,r6,r7			@ a^b, b^c in next round
+	ldr	r1,[sp,#5*4]	@ from future BODY_16_xx
+#endif
+	eor	r0,r0,r6,ror#20	@ Sigma0(a)
+	and	r3,r3,r12			@ (b^c)&=(a^b)
+	add	r9,r9,r5			@ d+=h
+	eor	r3,r3,r7			@ Maj(a,b,c)
+	add	r5,r5,r0,ror#2	@ h+=Sigma0(a)
+	@ add	r5,r5,r3			@ h+=Maj(a,b,c)
+#if __ARM_ARCH__>=7
+	@ ldr	r2,[r1],#4			@ 7
+# if 7==15
+	str	r1,[sp,#17*4]			@ make room for r1
+# endif
+	eor	r0,r9,r9,ror#5
+	add	r5,r5,r3			@ h+=Maj(a,b,c) from the past
+	eor	r0,r0,r9,ror#19	@ Sigma1(e)
+# ifndef __ARMEB__
+	rev	r2,r2
+# endif
+#else
+	@ ldrb	r2,[r1,#3]			@ 7
+	add	r5,r5,r3			@ h+=Maj(a,b,c) from the past
+	ldrb	r3,[r1,#2]
+	ldrb	r0,[r1,#1]
+	orr	r2,r2,r3,lsl#8
+	ldrb	r3,[r1],#4
+	orr	r2,r2,r0,lsl#16
+# if 7==15
+	str	r1,[sp,#17*4]			@ make room for r1
+# endif
+	eor	r0,r9,r9,ror#5
+	orr	r2,r2,r3,lsl#24
+	eor	r0,r0,r9,ror#19	@ Sigma1(e)
+#endif
+	ldr	r3,[r14],#4			@ *K256++
+	add	r4,r4,r2			@ h+=X[i]
+	str	r2,[sp,#7*4]
+	eor	r2,r10,r11
+	add	r4,r4,r0,ror#6	@ h+=Sigma1(e)
+	and	r2,r2,r9
+	add	r4,r4,r3			@ h+=K256[i]
+	eor	r2,r2,r11			@ Ch(e,f,g)
+	eor	r0,r5,r5,ror#11
+	add	r4,r4,r2			@ h+=Ch(e,f,g)
+#if 7==31
+	and	r3,r3,#0xff
+	cmp	r3,#0xf2			@ done?
+#endif
+#if 7<15
+# if __ARM_ARCH__>=7
+	ldr	r2,[r1],#4			@ prefetch
+# else
+	ldrb	r2,[r1,#3]
+# endif
+	eor	r3,r5,r6			@ a^b, b^c in next round
+#else
+	ldr	r2,[sp,#9*4]		@ from future BODY_16_xx
+	eor	r3,r5,r6			@ a^b, b^c in next round
+	ldr	r1,[sp,#6*4]	@ from future BODY_16_xx
+#endif
+	eor	r0,r0,r5,ror#20	@ Sigma0(a)
+	and	r12,r12,r3			@ (b^c)&=(a^b)
+	add	r8,r8,r4			@ d+=h
+	eor	r12,r12,r6			@ Maj(a,b,c)
+	add	r4,r4,r0,ror#2	@ h+=Sigma0(a)
+	@ add	r4,r4,r12			@ h+=Maj(a,b,c)
+#if __ARM_ARCH__>=7
+	@ ldr	r2,[r1],#4			@ 8
+# if 8==15
+	str	r1,[sp,#17*4]			@ make room for r1
+# endif
+	eor	r0,r8,r8,ror#5
+	add	r4,r4,r12			@ h+=Maj(a,b,c) from the past
+	eor	r0,r0,r8,ror#19	@ Sigma1(e)
+# ifndef __ARMEB__
+	rev	r2,r2
+# endif
+#else
+	@ ldrb	r2,[r1,#3]			@ 8
+	add	r4,r4,r12			@ h+=Maj(a,b,c) from the past
+	ldrb	r12,[r1,#2]
+	ldrb	r0,[r1,#1]
+	orr	r2,r2,r12,lsl#8
+	ldrb	r12,[r1],#4
+	orr	r2,r2,r0,lsl#16
+# if 8==15
+	str	r1,[sp,#17*4]			@ make room for r1
+# endif
+	eor	r0,r8,r8,ror#5
+	orr	r2,r2,r12,lsl#24
+	eor	r0,r0,r8,ror#19	@ Sigma1(e)
+#endif
+	ldr	r12,[r14],#4			@ *K256++
+	add	r11,r11,r2			@ h+=X[i]
+	str	r2,[sp,#8*4]
+	eor	r2,r9,r10
+	add	r11,r11,r0,ror#6	@ h+=Sigma1(e)
+	and	r2,r2,r8
+	add	r11,r11,r12			@ h+=K256[i]
+	eor	r2,r2,r10			@ Ch(e,f,g)
+	eor	r0,r4,r4,ror#11
+	add	r11,r11,r2			@ h+=Ch(e,f,g)
+#if 8==31
+	and	r12,r12,#0xff
+	cmp	r12,#0xf2			@ done?
+#endif
+#if 8<15
+# if __ARM_ARCH__>=7
+	ldr	r2,[r1],#4			@ prefetch
+# else
+	ldrb	r2,[r1,#3]
+# endif
+	eor	r12,r4,r5			@ a^b, b^c in next round
+#else
+	ldr	r2,[sp,#10*4]		@ from future BODY_16_xx
+	eor	r12,r4,r5			@ a^b, b^c in next round
+	ldr	r1,[sp,#7*4]	@ from future BODY_16_xx
+#endif
+	eor	r0,r0,r4,ror#20	@ Sigma0(a)
+	and	r3,r3,r12			@ (b^c)&=(a^b)
+	add	r7,r7,r11			@ d+=h
+	eor	r3,r3,r5			@ Maj(a,b,c)
+	add	r11,r11,r0,ror#2	@ h+=Sigma0(a)
+	@ add	r11,r11,r3			@ h+=Maj(a,b,c)
+#if __ARM_ARCH__>=7
+	@ ldr	r2,[r1],#4			@ 9
+# if 9==15
+	str	r1,[sp,#17*4]			@ make room for r1
+# endif
+	eor	r0,r7,r7,ror#5
+	add	r11,r11,r3			@ h+=Maj(a,b,c) from the past
+	eor	r0,r0,r7,ror#19	@ Sigma1(e)
+# ifndef __ARMEB__
+	rev	r2,r2
+# endif
+#else
+	@ ldrb	r2,[r1,#3]			@ 9
+	add	r11,r11,r3			@ h+=Maj(a,b,c) from the past
+	ldrb	r3,[r1,#2]
+	ldrb	r0,[r1,#1]
+	orr	r2,r2,r3,lsl#8
+	ldrb	r3,[r1],#4
+	orr	r2,r2,r0,lsl#16
+# if 9==15
+	str	r1,[sp,#17*4]			@ make room for r1
+# endif
+	eor	r0,r7,r7,ror#5
+	orr	r2,r2,r3,lsl#24
+	eor	r0,r0,r7,ror#19	@ Sigma1(e)
+#endif
+	ldr	r3,[r14],#4			@ *K256++
+	add	r10,r10,r2			@ h+=X[i]
+	str	r2,[sp,#9*4]
+	eor	r2,r8,r9
+	add	r10,r10,r0,ror#6	@ h+=Sigma1(e)
+	and	r2,r2,r7
+	add	r10,r10,r3			@ h+=K256[i]
+	eor	r2,r2,r9			@ Ch(e,f,g)
+	eor	r0,r11,r11,ror#11
+	add	r10,r10,r2			@ h+=Ch(e,f,g)
+#if 9==31
+	and	r3,r3,#0xff
+	cmp	r3,#0xf2			@ done?
+#endif
+#if 9<15
+# if __ARM_ARCH__>=7
+	ldr	r2,[r1],#4			@ prefetch
+# else
+	ldrb	r2,[r1,#3]
+# endif
+	eor	r3,r11,r4			@ a^b, b^c in next round
+#else
+	ldr	r2,[sp,#11*4]		@ from future BODY_16_xx
+	eor	r3,r11,r4			@ a^b, b^c in next round
+	ldr	r1,[sp,#8*4]	@ from future BODY_16_xx
+#endif
+	eor	r0,r0,r11,ror#20	@ Sigma0(a)
+	and	r12,r12,r3			@ (b^c)&=(a^b)
+	add	r6,r6,r10			@ d+=h
+	eor	r12,r12,r4			@ Maj(a,b,c)
+	add	r10,r10,r0,ror#2	@ h+=Sigma0(a)
+	@ add	r10,r10,r12			@ h+=Maj(a,b,c)
+#if __ARM_ARCH__>=7
+	@ ldr	r2,[r1],#4			@ 10
+# if 10==15
+	str	r1,[sp,#17*4]			@ make room for r1
+# endif
+	eor	r0,r6,r6,ror#5
+	add	r10,r10,r12			@ h+=Maj(a,b,c) from the past
+	eor	r0,r0,r6,ror#19	@ Sigma1(e)
+# ifndef __ARMEB__
+	rev	r2,r2
+# endif
+#else
+	@ ldrb	r2,[r1,#3]			@ 10
+	add	r10,r10,r12			@ h+=Maj(a,b,c) from the past
+	ldrb	r12,[r1,#2]
+	ldrb	r0,[r1,#1]
+	orr	r2,r2,r12,lsl#8
+	ldrb	r12,[r1],#4
+	orr	r2,r2,r0,lsl#16
+# if 10==15
+	str	r1,[sp,#17*4]			@ make room for r1
+# endif
+	eor	r0,r6,r6,ror#5
+	orr	r2,r2,r12,lsl#24
+	eor	r0,r0,r6,ror#19	@ Sigma1(e)
+#endif
+	ldr	r12,[r14],#4			@ *K256++
+	add	r9,r9,r2			@ h+=X[i]
+	str	r2,[sp,#10*4]
+	eor	r2,r7,r8
+	add	r9,r9,r0,ror#6	@ h+=Sigma1(e)
+	and	r2,r2,r6
+	add	r9,r9,r12			@ h+=K256[i]
+	eor	r2,r2,r8			@ Ch(e,f,g)
+	eor	r0,r10,r10,ror#11
+	add	r9,r9,r2			@ h+=Ch(e,f,g)
+#if 10==31
+	and	r12,r12,#0xff
+	cmp	r12,#0xf2			@ done?
+#endif
+#if 10<15
+# if __ARM_ARCH__>=7
+	ldr	r2,[r1],#4			@ prefetch
+# else
+	ldrb	r2,[r1,#3]
+# endif
+	eor	r12,r10,r11			@ a^b, b^c in next round
+#else
+	ldr	r2,[sp,#12*4]		@ from future BODY_16_xx
+	eor	r12,r10,r11			@ a^b, b^c in next round
+	ldr	r1,[sp,#9*4]	@ from future BODY_16_xx
+#endif
+	eor	r0,r0,r10,ror#20	@ Sigma0(a)
+	and	r3,r3,r12			@ (b^c)&=(a^b)
+	add	r5,r5,r9			@ d+=h
+	eor	r3,r3,r11			@ Maj(a,b,c)
+	add	r9,r9,r0,ror#2	@ h+=Sigma0(a)
+	@ add	r9,r9,r3			@ h+=Maj(a,b,c)
+#if __ARM_ARCH__>=7
+	@ ldr	r2,[r1],#4			@ 11
+# if 11==15
+	str	r1,[sp,#17*4]			@ make room for r1
+# endif
+	eor	r0,r5,r5,ror#5
+	add	r9,r9,r3			@ h+=Maj(a,b,c) from the past
+	eor	r0,r0,r5,ror#19	@ Sigma1(e)
+# ifndef __ARMEB__
+	rev	r2,r2
+# endif
+#else
+	@ ldrb	r2,[r1,#3]			@ 11
+	add	r9,r9,r3			@ h+=Maj(a,b,c) from the past
+	ldrb	r3,[r1,#2]
+	ldrb	r0,[r1,#1]
+	orr	r2,r2,r3,lsl#8
+	ldrb	r3,[r1],#4
+	orr	r2,r2,r0,lsl#16
+# if 11==15
+	str	r1,[sp,#17*4]			@ make room for r1
+# endif
+	eor	r0,r5,r5,ror#5
+	orr	r2,r2,r3,lsl#24
+	eor	r0,r0,r5,ror#19	@ Sigma1(e)
+#endif
+	ldr	r3,[r14],#4			@ *K256++
+	add	r8,r8,r2			@ h+=X[i]
+	str	r2,[sp,#11*4]
+	eor	r2,r6,r7
+	add	r8,r8,r0,ror#6	@ h+=Sigma1(e)
+	and	r2,r2,r5
+	add	r8,r8,r3			@ h+=K256[i]
+	eor	r2,r2,r7			@ Ch(e,f,g)
+	eor	r0,r9,r9,ror#11
+	add	r8,r8,r2			@ h+=Ch(e,f,g)
+#if 11==31
+	and	r3,r3,#0xff
+	cmp	r3,#0xf2			@ done?
+#endif
+#if 11<15
+# if __ARM_ARCH__>=7
+	ldr	r2,[r1],#4			@ prefetch
+# else
+	ldrb	r2,[r1,#3]
+# endif
+	eor	r3,r9,r10			@ a^b, b^c in next round
+#else
+	ldr	r2,[sp,#13*4]		@ from future BODY_16_xx
+	eor	r3,r9,r10			@ a^b, b^c in next round
+	ldr	r1,[sp,#10*4]	@ from future BODY_16_xx
+#endif
+	eor	r0,r0,r9,ror#20	@ Sigma0(a)
+	and	r12,r12,r3			@ (b^c)&=(a^b)
+	add	r4,r4,r8			@ d+=h
+	eor	r12,r12,r10			@ Maj(a,b,c)
+	add	r8,r8,r0,ror#2	@ h+=Sigma0(a)
+	@ add	r8,r8,r12			@ h+=Maj(a,b,c)
+#if __ARM_ARCH__>=7
+	@ ldr	r2,[r1],#4			@ 12
+# if 12==15
+	str	r1,[sp,#17*4]			@ make room for r1
+# endif
+	eor	r0,r4,r4,ror#5
+	add	r8,r8,r12			@ h+=Maj(a,b,c) from the past
+	eor	r0,r0,r4,ror#19	@ Sigma1(e)
+# ifndef __ARMEB__
+	rev	r2,r2
+# endif
+#else
+	@ ldrb	r2,[r1,#3]			@ 12
+	add	r8,r8,r12			@ h+=Maj(a,b,c) from the past
+	ldrb	r12,[r1,#2]
+	ldrb	r0,[r1,#1]
+	orr	r2,r2,r12,lsl#8
+	ldrb	r12,[r1],#4
+	orr	r2,r2,r0,lsl#16
+# if 12==15
+	str	r1,[sp,#17*4]			@ make room for r1
+# endif
+	eor	r0,r4,r4,ror#5
+	orr	r2,r2,r12,lsl#24
+	eor	r0,r0,r4,ror#19	@ Sigma1(e)
+#endif
+	ldr	r12,[r14],#4			@ *K256++
+	add	r7,r7,r2			@ h+=X[i]
+	str	r2,[sp,#12*4]
+	eor	r2,r5,r6
+	add	r7,r7,r0,ror#6	@ h+=Sigma1(e)
+	and	r2,r2,r4
+	add	r7,r7,r12			@ h+=K256[i]
+	eor	r2,r2,r6			@ Ch(e,f,g)
+	eor	r0,r8,r8,ror#11
+	add	r7,r7,r2			@ h+=Ch(e,f,g)
+#if 12==31
+	and	r12,r12,#0xff
+	cmp	r12,#0xf2			@ done?
+#endif
+#if 12<15
+# if __ARM_ARCH__>=7
+	ldr	r2,[r1],#4			@ prefetch
+# else
+	ldrb	r2,[r1,#3]
+# endif
+	eor	r12,r8,r9			@ a^b, b^c in next round
+#else
+	ldr	r2,[sp,#14*4]		@ from future BODY_16_xx
+	eor	r12,r8,r9			@ a^b, b^c in next round
+	ldr	r1,[sp,#11*4]	@ from future BODY_16_xx
+#endif
+	eor	r0,r0,r8,ror#20	@ Sigma0(a)
+	and	r3,r3,r12			@ (b^c)&=(a^b)
+	add	r11,r11,r7			@ d+=h
+	eor	r3,r3,r9			@ Maj(a,b,c)
+	add	r7,r7,r0,ror#2	@ h+=Sigma0(a)
+	@ add	r7,r7,r3			@ h+=Maj(a,b,c)
+#if __ARM_ARCH__>=7
+	@ ldr	r2,[r1],#4			@ 13
+# if 13==15
+	str	r1,[sp,#17*4]			@ make room for r1
+# endif
+	eor	r0,r11,r11,ror#5
+	add	r7,r7,r3			@ h+=Maj(a,b,c) from the past
+	eor	r0,r0,r11,ror#19	@ Sigma1(e)
+# ifndef __ARMEB__
+	rev	r2,r2
+# endif
+#else
+	@ ldrb	r2,[r1,#3]			@ 13
+	add	r7,r7,r3			@ h+=Maj(a,b,c) from the past
+	ldrb	r3,[r1,#2]
+	ldrb	r0,[r1,#1]
+	orr	r2,r2,r3,lsl#8
+	ldrb	r3,[r1],#4
+	orr	r2,r2,r0,lsl#16
+# if 13==15
+	str	r1,[sp,#17*4]			@ make room for r1
+# endif
+	eor	r0,r11,r11,ror#5
+	orr	r2,r2,r3,lsl#24
+	eor	r0,r0,r11,ror#19	@ Sigma1(e)
+#endif
+	ldr	r3,[r14],#4			@ *K256++
+	add	r6,r6,r2			@ h+=X[i]
+	str	r2,[sp,#13*4]
+	eor	r2,r4,r5
+	add	r6,r6,r0,ror#6	@ h+=Sigma1(e)
+	and	r2,r2,r11
+	add	r6,r6,r3			@ h+=K256[i]
+	eor	r2,r2,r5			@ Ch(e,f,g)
+	eor	r0,r7,r7,ror#11
+	add	r6,r6,r2			@ h+=Ch(e,f,g)
+#if 13==31
+	and	r3,r3,#0xff
+	cmp	r3,#0xf2			@ done?
+#endif
+#if 13<15
+# if __ARM_ARCH__>=7
+	ldr	r2,[r1],#4			@ prefetch
+# else
+	ldrb	r2,[r1,#3]
+# endif
+	eor	r3,r7,r8			@ a^b, b^c in next round
+#else
+	ldr	r2,[sp,#15*4]		@ from future BODY_16_xx
+	eor	r3,r7,r8			@ a^b, b^c in next round
+	ldr	r1,[sp,#12*4]	@ from future BODY_16_xx
+#endif
+	eor	r0,r0,r7,ror#20	@ Sigma0(a)
+	and	r12,r12,r3			@ (b^c)&=(a^b)
+	add	r10,r10,r6			@ d+=h
+	eor	r12,r12,r8			@ Maj(a,b,c)
+	add	r6,r6,r0,ror#2	@ h+=Sigma0(a)
+	@ add	r6,r6,r12			@ h+=Maj(a,b,c)
+#if __ARM_ARCH__>=7
+	@ ldr	r2,[r1],#4			@ 14
+# if 14==15
+	str	r1,[sp,#17*4]			@ make room for r1
+# endif
+	eor	r0,r10,r10,ror#5
+	add	r6,r6,r12			@ h+=Maj(a,b,c) from the past
+	eor	r0,r0,r10,ror#19	@ Sigma1(e)
+# ifndef __ARMEB__
+	rev	r2,r2
+# endif
+#else
+	@ ldrb	r2,[r1,#3]			@ 14
+	add	r6,r6,r12			@ h+=Maj(a,b,c) from the past
+	ldrb	r12,[r1,#2]
+	ldrb	r0,[r1,#1]
+	orr	r2,r2,r12,lsl#8
+	ldrb	r12,[r1],#4
+	orr	r2,r2,r0,lsl#16
+# if 14==15
+	str	r1,[sp,#17*4]			@ make room for r1
+# endif
+	eor	r0,r10,r10,ror#5
+	orr	r2,r2,r12,lsl#24
+	eor	r0,r0,r10,ror#19	@ Sigma1(e)
+#endif
+	ldr	r12,[r14],#4			@ *K256++
+	add	r5,r5,r2			@ h+=X[i]
+	str	r2,[sp,#14*4]
+	eor	r2,r11,r4
+	add	r5,r5,r0,ror#6	@ h+=Sigma1(e)
+	and	r2,r2,r10
+	add	r5,r5,r12			@ h+=K256[i]
+	eor	r2,r2,r4			@ Ch(e,f,g)
+	eor	r0,r6,r6,ror#11
+	add	r5,r5,r2			@ h+=Ch(e,f,g)
+#if 14==31
+	and	r12,r12,#0xff
+	cmp	r12,#0xf2			@ done?
+#endif
+#if 14<15
+# if __ARM_ARCH__>=7
+	ldr	r2,[r1],#4			@ prefetch
+# else
+	ldrb	r2,[r1,#3]
+# endif
+	eor	r12,r6,r7			@ a^b, b^c in next round
+#else
+	ldr	r2,[sp,#0*4]		@ from future BODY_16_xx
+	eor	r12,r6,r7			@ a^b, b^c in next round
+	ldr	r1,[sp,#13*4]	@ from future BODY_16_xx
+#endif
+	eor	r0,r0,r6,ror#20	@ Sigma0(a)
+	and	r3,r3,r12			@ (b^c)&=(a^b)
+	add	r9,r9,r5			@ d+=h
+	eor	r3,r3,r7			@ Maj(a,b,c)
+	add	r5,r5,r0,ror#2	@ h+=Sigma0(a)
+	@ add	r5,r5,r3			@ h+=Maj(a,b,c)
+#if __ARM_ARCH__>=7
+	@ ldr	r2,[r1],#4			@ 15
+# if 15==15
+	str	r1,[sp,#17*4]			@ make room for r1
+# endif
+	eor	r0,r9,r9,ror#5
+	add	r5,r5,r3			@ h+=Maj(a,b,c) from the past
+	eor	r0,r0,r9,ror#19	@ Sigma1(e)
+# ifndef __ARMEB__
+	rev	r2,r2
+# endif
+#else
+	@ ldrb	r2,[r1,#3]			@ 15
+	add	r5,r5,r3			@ h+=Maj(a,b,c) from the past
+	ldrb	r3,[r1,#2]
+	ldrb	r0,[r1,#1]
+	orr	r2,r2,r3,lsl#8
+	ldrb	r3,[r1],#4
+	orr	r2,r2,r0,lsl#16
+# if 15==15
+	str	r1,[sp,#17*4]			@ make room for r1
+# endif
+	eor	r0,r9,r9,ror#5
+	orr	r2,r2,r3,lsl#24
+	eor	r0,r0,r9,ror#19	@ Sigma1(e)
+#endif
+	ldr	r3,[r14],#4			@ *K256++
+	add	r4,r4,r2			@ h+=X[i]
+	str	r2,[sp,#15*4]
+	eor	r2,r10,r11
+	add	r4,r4,r0,ror#6	@ h+=Sigma1(e)
+	and	r2,r2,r9
+	add	r4,r4,r3			@ h+=K256[i]
+	eor	r2,r2,r11			@ Ch(e,f,g)
+	eor	r0,r5,r5,ror#11
+	add	r4,r4,r2			@ h+=Ch(e,f,g)
+#if 15==31
+	and	r3,r3,#0xff
+	cmp	r3,#0xf2			@ done?
+#endif
+#if 15<15
+# if __ARM_ARCH__>=7
+	ldr	r2,[r1],#4			@ prefetch
+# else
+	ldrb	r2,[r1,#3]
+# endif
+	eor	r3,r5,r6			@ a^b, b^c in next round
+#else
+	ldr	r2,[sp,#1*4]		@ from future BODY_16_xx
+	eor	r3,r5,r6			@ a^b, b^c in next round
+	ldr	r1,[sp,#14*4]	@ from future BODY_16_xx
+#endif
+	eor	r0,r0,r5,ror#20	@ Sigma0(a)
+	and	r12,r12,r3			@ (b^c)&=(a^b)
+	add	r8,r8,r4			@ d+=h
+	eor	r12,r12,r6			@ Maj(a,b,c)
+	add	r4,r4,r0,ror#2	@ h+=Sigma0(a)
+	@ add	r4,r4,r12			@ h+=Maj(a,b,c)
+.Lrounds_16_xx:
+	@ ldr	r2,[sp,#1*4]		@ 16
+	@ ldr	r1,[sp,#14*4]
+	mov	r0,r2,ror#7
+	add	r4,r4,r12			@ h+=Maj(a,b,c) from the past
+	mov	r12,r1,ror#17
+	eor	r0,r0,r2,ror#18
+	eor	r12,r12,r1,ror#19
+	eor	r0,r0,r2,lsr#3	@ sigma0(X[i+1])
+	ldr	r2,[sp,#0*4]
+	eor	r12,r12,r1,lsr#10	@ sigma1(X[i+14])
+	ldr	r1,[sp,#9*4]
+
+	add	r12,r12,r0
+	eor	r0,r8,r8,ror#5	@ from BODY_00_15
+	add	r2,r2,r12
+	eor	r0,r0,r8,ror#19	@ Sigma1(e)
+	add	r2,r2,r1			@ X[i]
+	ldr	r12,[r14],#4			@ *K256++
+	add	r11,r11,r2			@ h+=X[i]
+	str	r2,[sp,#0*4]
+	eor	r2,r9,r10
+	add	r11,r11,r0,ror#6	@ h+=Sigma1(e)
+	and	r2,r2,r8
+	add	r11,r11,r12			@ h+=K256[i]
+	eor	r2,r2,r10			@ Ch(e,f,g)
+	eor	r0,r4,r4,ror#11
+	add	r11,r11,r2			@ h+=Ch(e,f,g)
+#if 16==31
+	and	r12,r12,#0xff
+	cmp	r12,#0xf2			@ done?
+#endif
+#if 16<15
+# if __ARM_ARCH__>=7
+	ldr	r2,[r1],#4			@ prefetch
+# else
+	ldrb	r2,[r1,#3]
+# endif
+	eor	r12,r4,r5			@ a^b, b^c in next round
+#else
+	ldr	r2,[sp,#2*4]		@ from future BODY_16_xx
+	eor	r12,r4,r5			@ a^b, b^c in next round
+	ldr	r1,[sp,#15*4]	@ from future BODY_16_xx
+#endif
+	eor	r0,r0,r4,ror#20	@ Sigma0(a)
+	and	r3,r3,r12			@ (b^c)&=(a^b)
+	add	r7,r7,r11			@ d+=h
+	eor	r3,r3,r5			@ Maj(a,b,c)
+	add	r11,r11,r0,ror#2	@ h+=Sigma0(a)
+	@ add	r11,r11,r3			@ h+=Maj(a,b,c)
+	@ ldr	r2,[sp,#2*4]		@ 17
+	@ ldr	r1,[sp,#15*4]
+	mov	r0,r2,ror#7
+	add	r11,r11,r3			@ h+=Maj(a,b,c) from the past
+	mov	r3,r1,ror#17
+	eor	r0,r0,r2,ror#18
+	eor	r3,r3,r1,ror#19
+	eor	r0,r0,r2,lsr#3	@ sigma0(X[i+1])
+	ldr	r2,[sp,#1*4]
+	eor	r3,r3,r1,lsr#10	@ sigma1(X[i+14])
+	ldr	r1,[sp,#10*4]
+
+	add	r3,r3,r0
+	eor	r0,r7,r7,ror#5	@ from BODY_00_15
+	add	r2,r2,r3
+	eor	r0,r0,r7,ror#19	@ Sigma1(e)
+	add	r2,r2,r1			@ X[i]
+	ldr	r3,[r14],#4			@ *K256++
+	add	r10,r10,r2			@ h+=X[i]
+	str	r2,[sp,#1*4]
+	eor	r2,r8,r9
+	add	r10,r10,r0,ror#6	@ h+=Sigma1(e)
+	and	r2,r2,r7
+	add	r10,r10,r3			@ h+=K256[i]
+	eor	r2,r2,r9			@ Ch(e,f,g)
+	eor	r0,r11,r11,ror#11
+	add	r10,r10,r2			@ h+=Ch(e,f,g)
+#if 17==31
+	and	r3,r3,#0xff
+	cmp	r3,#0xf2			@ done?
+#endif
+#if 17<15
+# if __ARM_ARCH__>=7
+	ldr	r2,[r1],#4			@ prefetch
+# else
+	ldrb	r2,[r1,#3]
+# endif
+	eor	r3,r11,r4			@ a^b, b^c in next round
+#else
+	ldr	r2,[sp,#3*4]		@ from future BODY_16_xx
+	eor	r3,r11,r4			@ a^b, b^c in next round
+	ldr	r1,[sp,#0*4]	@ from future BODY_16_xx
+#endif
+	eor	r0,r0,r11,ror#20	@ Sigma0(a)
+	and	r12,r12,r3			@ (b^c)&=(a^b)
+	add	r6,r6,r10			@ d+=h
+	eor	r12,r12,r4			@ Maj(a,b,c)
+	add	r10,r10,r0,ror#2	@ h+=Sigma0(a)
+	@ add	r10,r10,r12			@ h+=Maj(a,b,c)
+	@ ldr	r2,[sp,#3*4]		@ 18
+	@ ldr	r1,[sp,#0*4]
+	mov	r0,r2,ror#7
+	add	r10,r10,r12			@ h+=Maj(a,b,c) from the past
+	mov	r12,r1,ror#17
+	eor	r0,r0,r2,ror#18
+	eor	r12,r12,r1,ror#19
+	eor	r0,r0,r2,lsr#3	@ sigma0(X[i+1])
+	ldr	r2,[sp,#2*4]
+	eor	r12,r12,r1,lsr#10	@ sigma1(X[i+14])
+	ldr	r1,[sp,#11*4]
+
+	add	r12,r12,r0
+	eor	r0,r6,r6,ror#5	@ from BODY_00_15
+	add	r2,r2,r12
+	eor	r0,r0,r6,ror#19	@ Sigma1(e)
+	add	r2,r2,r1			@ X[i]
+	ldr	r12,[r14],#4			@ *K256++
+	add	r9,r9,r2			@ h+=X[i]
+	str	r2,[sp,#2*4]
+	eor	r2,r7,r8
+	add	r9,r9,r0,ror#6	@ h+=Sigma1(e)
+	and	r2,r2,r6
+	add	r9,r9,r12			@ h+=K256[i]
+	eor	r2,r2,r8			@ Ch(e,f,g)
+	eor	r0,r10,r10,ror#11
+	add	r9,r9,r2			@ h+=Ch(e,f,g)
+#if 18==31
+	and	r12,r12,#0xff
+	cmp	r12,#0xf2			@ done?
+#endif
+#if 18<15
+# if __ARM_ARCH__>=7
+	ldr	r2,[r1],#4			@ prefetch
+# else
+	ldrb	r2,[r1,#3]
+# endif
+	eor	r12,r10,r11			@ a^b, b^c in next round
+#else
+	ldr	r2,[sp,#4*4]		@ from future BODY_16_xx
+	eor	r12,r10,r11			@ a^b, b^c in next round
+	ldr	r1,[sp,#1*4]	@ from future BODY_16_xx
+#endif
+	eor	r0,r0,r10,ror#20	@ Sigma0(a)
+	and	r3,r3,r12			@ (b^c)&=(a^b)
+	add	r5,r5,r9			@ d+=h
+	eor	r3,r3,r11			@ Maj(a,b,c)
+	add	r9,r9,r0,ror#2	@ h+=Sigma0(a)
+	@ add	r9,r9,r3			@ h+=Maj(a,b,c)
+	@ ldr	r2,[sp,#4*4]		@ 19
+	@ ldr	r1,[sp,#1*4]
+	mov	r0,r2,ror#7
+	add	r9,r9,r3			@ h+=Maj(a,b,c) from the past
+	mov	r3,r1,ror#17
+	eor	r0,r0,r2,ror#18
+	eor	r3,r3,r1,ror#19
+	eor	r0,r0,r2,lsr#3	@ sigma0(X[i+1])
+	ldr	r2,[sp,#3*4]
+	eor	r3,r3,r1,lsr#10	@ sigma1(X[i+14])
+	ldr	r1,[sp,#12*4]
+
+	add	r3,r3,r0
+	eor	r0,r5,r5,ror#5	@ from BODY_00_15
+	add	r2,r2,r3
+	eor	r0,r0,r5,ror#19	@ Sigma1(e)
+	add	r2,r2,r1			@ X[i]
+	ldr	r3,[r14],#4			@ *K256++
+	add	r8,r8,r2			@ h+=X[i]
+	str	r2,[sp,#3*4]
+	eor	r2,r6,r7
+	add	r8,r8,r0,ror#6	@ h+=Sigma1(e)
+	and	r2,r2,r5
+	add	r8,r8,r3			@ h+=K256[i]
+	eor	r2,r2,r7			@ Ch(e,f,g)
+	eor	r0,r9,r9,ror#11
+	add	r8,r8,r2			@ h+=Ch(e,f,g)
+#if 19==31
+	and	r3,r3,#0xff
+	cmp	r3,#0xf2			@ done?
+#endif
+#if 19<15
+# if __ARM_ARCH__>=7
+	ldr	r2,[r1],#4			@ prefetch
+# else
+	ldrb	r2,[r1,#3]
+# endif
+	eor	r3,r9,r10			@ a^b, b^c in next round
+#else
+	ldr	r2,[sp,#5*4]		@ from future BODY_16_xx
+	eor	r3,r9,r10			@ a^b, b^c in next round
+	ldr	r1,[sp,#2*4]	@ from future BODY_16_xx
+#endif
+	eor	r0,r0,r9,ror#20	@ Sigma0(a)
+	and	r12,r12,r3			@ (b^c)&=(a^b)
+	add	r4,r4,r8			@ d+=h
+	eor	r12,r12,r10			@ Maj(a,b,c)
+	add	r8,r8,r0,ror#2	@ h+=Sigma0(a)
+	@ add	r8,r8,r12			@ h+=Maj(a,b,c)
+	@ ldr	r2,[sp,#5*4]		@ 20
+	@ ldr	r1,[sp,#2*4]
+	mov	r0,r2,ror#7
+	add	r8,r8,r12			@ h+=Maj(a,b,c) from the past
+	mov	r12,r1,ror#17
+	eor	r0,r0,r2,ror#18
+	eor	r12,r12,r1,ror#19
+	eor	r0,r0,r2,lsr#3	@ sigma0(X[i+1])
+	ldr	r2,[sp,#4*4]
+	eor	r12,r12,r1,lsr#10	@ sigma1(X[i+14])
+	ldr	r1,[sp,#13*4]
+
+	add	r12,r12,r0
+	eor	r0,r4,r4,ror#5	@ from BODY_00_15
+	add	r2,r2,r12
+	eor	r0,r0,r4,ror#19	@ Sigma1(e)
+	add	r2,r2,r1			@ X[i]
+	ldr	r12,[r14],#4			@ *K256++
+	add	r7,r7,r2			@ h+=X[i]
+	str	r2,[sp,#4*4]
+	eor	r2,r5,r6
+	add	r7,r7,r0,ror#6	@ h+=Sigma1(e)
+	and	r2,r2,r4
+	add	r7,r7,r12			@ h+=K256[i]
+	eor	r2,r2,r6			@ Ch(e,f,g)
+	eor	r0,r8,r8,ror#11
+	add	r7,r7,r2			@ h+=Ch(e,f,g)
+#if 20==31
+	and	r12,r12,#0xff
+	cmp	r12,#0xf2			@ done?
+#endif
+#if 20<15
+# if __ARM_ARCH__>=7
+	ldr	r2,[r1],#4			@ prefetch
+# else
+	ldrb	r2,[r1,#3]
+# endif
+	eor	r12,r8,r9			@ a^b, b^c in next round
+#else
+	ldr	r2,[sp,#6*4]		@ from future BODY_16_xx
+	eor	r12,r8,r9			@ a^b, b^c in next round
+	ldr	r1,[sp,#3*4]	@ from future BODY_16_xx
+#endif
+	eor	r0,r0,r8,ror#20	@ Sigma0(a)
+	and	r3,r3,r12			@ (b^c)&=(a^b)
+	add	r11,r11,r7			@ d+=h
+	eor	r3,r3,r9			@ Maj(a,b,c)
+	add	r7,r7,r0,ror#2	@ h+=Sigma0(a)
+	@ add	r7,r7,r3			@ h+=Maj(a,b,c)
+	@ ldr	r2,[sp,#6*4]		@ 21
+	@ ldr	r1,[sp,#3*4]
+	mov	r0,r2,ror#7
+	add	r7,r7,r3			@ h+=Maj(a,b,c) from the past
+	mov	r3,r1,ror#17
+	eor	r0,r0,r2,ror#18
+	eor	r3,r3,r1,ror#19
+	eor	r0,r0,r2,lsr#3	@ sigma0(X[i+1])
+	ldr	r2,[sp,#5*4]
+	eor	r3,r3,r1,lsr#10	@ sigma1(X[i+14])
+	ldr	r1,[sp,#14*4]
+
+	add	r3,r3,r0
+	eor	r0,r11,r11,ror#5	@ from BODY_00_15
+	add	r2,r2,r3
+	eor	r0,r0,r11,ror#19	@ Sigma1(e)
+	add	r2,r2,r1			@ X[i]
+	ldr	r3,[r14],#4			@ *K256++
+	add	r6,r6,r2			@ h+=X[i]
+	str	r2,[sp,#5*4]
+	eor	r2,r4,r5
+	add	r6,r6,r0,ror#6	@ h+=Sigma1(e)
+	and	r2,r2,r11
+	add	r6,r6,r3			@ h+=K256[i]
+	eor	r2,r2,r5			@ Ch(e,f,g)
+	eor	r0,r7,r7,ror#11
+	add	r6,r6,r2			@ h+=Ch(e,f,g)
+#if 21==31
+	and	r3,r3,#0xff
+	cmp	r3,#0xf2			@ done?
+#endif
+#if 21<15
+# if __ARM_ARCH__>=7
+	ldr	r2,[r1],#4			@ prefetch
+# else
+	ldrb	r2,[r1,#3]
+# endif
+	eor	r3,r7,r8			@ a^b, b^c in next round
+#else
+	ldr	r2,[sp,#7*4]		@ from future BODY_16_xx
+	eor	r3,r7,r8			@ a^b, b^c in next round
+	ldr	r1,[sp,#4*4]	@ from future BODY_16_xx
+#endif
+	eor	r0,r0,r7,ror#20	@ Sigma0(a)
+	and	r12,r12,r3			@ (b^c)&=(a^b)
+	add	r10,r10,r6			@ d+=h
+	eor	r12,r12,r8			@ Maj(a,b,c)
+	add	r6,r6,r0,ror#2	@ h+=Sigma0(a)
+	@ add	r6,r6,r12			@ h+=Maj(a,b,c)
+	@ ldr	r2,[sp,#7*4]		@ 22
+	@ ldr	r1,[sp,#4*4]
+	mov	r0,r2,ror#7
+	add	r6,r6,r12			@ h+=Maj(a,b,c) from the past
+	mov	r12,r1,ror#17
+	eor	r0,r0,r2,ror#18
+	eor	r12,r12,r1,ror#19
+	eor	r0,r0,r2,lsr#3	@ sigma0(X[i+1])
+	ldr	r2,[sp,#6*4]
+	eor	r12,r12,r1,lsr#10	@ sigma1(X[i+14])
+	ldr	r1,[sp,#15*4]
+
+	add	r12,r12,r0
+	eor	r0,r10,r10,ror#5	@ from BODY_00_15
+	add	r2,r2,r12
+	eor	r0,r0,r10,ror#19	@ Sigma1(e)
+	add	r2,r2,r1			@ X[i]
+	ldr	r12,[r14],#4			@ *K256++
+	add	r5,r5,r2			@ h+=X[i]
+	str	r2,[sp,#6*4]
+	eor	r2,r11,r4
+	add	r5,r5,r0,ror#6	@ h+=Sigma1(e)
+	and	r2,r2,r10
+	add	r5,r5,r12			@ h+=K256[i]
+	eor	r2,r2,r4			@ Ch(e,f,g)
+	eor	r0,r6,r6,ror#11
+	add	r5,r5,r2			@ h+=Ch(e,f,g)
+#if 22==31
+	and	r12,r12,#0xff
+	cmp	r12,#0xf2			@ done?
+#endif
+#if 22<15
+# if __ARM_ARCH__>=7
+	ldr	r2,[r1],#4			@ prefetch
+# else
+	ldrb	r2,[r1,#3]
+# endif
+	eor	r12,r6,r7			@ a^b, b^c in next round
+#else
+	ldr	r2,[sp,#8*4]		@ from future BODY_16_xx
+	eor	r12,r6,r7			@ a^b, b^c in next round
+	ldr	r1,[sp,#5*4]	@ from future BODY_16_xx
+#endif
+	eor	r0,r0,r6,ror#20	@ Sigma0(a)
+	and	r3,r3,r12			@ (b^c)&=(a^b)
+	add	r9,r9,r5			@ d+=h
+	eor	r3,r3,r7			@ Maj(a,b,c)
+	add	r5,r5,r0,ror#2	@ h+=Sigma0(a)
+	@ add	r5,r5,r3			@ h+=Maj(a,b,c)
+	@ ldr	r2,[sp,#8*4]		@ 23
+	@ ldr	r1,[sp,#5*4]
+	mov	r0,r2,ror#7
+	add	r5,r5,r3			@ h+=Maj(a,b,c) from the past
+	mov	r3,r1,ror#17
+	eor	r0,r0,r2,ror#18
+	eor	r3,r3,r1,ror#19
+	eor	r0,r0,r2,lsr#3	@ sigma0(X[i+1])
+	ldr	r2,[sp,#7*4]
+	eor	r3,r3,r1,lsr#10	@ sigma1(X[i+14])
+	ldr	r1,[sp,#0*4]
+
+	add	r3,r3,r0
+	eor	r0,r9,r9,ror#5	@ from BODY_00_15
+	add	r2,r2,r3
+	eor	r0,r0,r9,ror#19	@ Sigma1(e)
+	add	r2,r2,r1			@ X[i]
+	ldr	r3,[r14],#4			@ *K256++
+	add	r4,r4,r2			@ h+=X[i]
+	str	r2,[sp,#7*4]
+	eor	r2,r10,r11
+	add	r4,r4,r0,ror#6	@ h+=Sigma1(e)
+	and	r2,r2,r9
+	add	r4,r4,r3			@ h+=K256[i]
+	eor	r2,r2,r11			@ Ch(e,f,g)
+	eor	r0,r5,r5,ror#11
+	add	r4,r4,r2			@ h+=Ch(e,f,g)
+#if 23==31
+	and	r3,r3,#0xff
+	cmp	r3,#0xf2			@ done?
+#endif
+#if 23<15
+# if __ARM_ARCH__>=7
+	ldr	r2,[r1],#4			@ prefetch
+# else
+	ldrb	r2,[r1,#3]
+# endif
+	eor	r3,r5,r6			@ a^b, b^c in next round
+#else
+	ldr	r2,[sp,#9*4]		@ from future BODY_16_xx
+	eor	r3,r5,r6			@ a^b, b^c in next round
+	ldr	r1,[sp,#6*4]	@ from future BODY_16_xx
+#endif
+	eor	r0,r0,r5,ror#20	@ Sigma0(a)
+	and	r12,r12,r3			@ (b^c)&=(a^b)
+	add	r8,r8,r4			@ d+=h
+	eor	r12,r12,r6			@ Maj(a,b,c)
+	add	r4,r4,r0,ror#2	@ h+=Sigma0(a)
+	@ add	r4,r4,r12			@ h+=Maj(a,b,c)
+	@ ldr	r2,[sp,#9*4]		@ 24
+	@ ldr	r1,[sp,#6*4]
+	mov	r0,r2,ror#7
+	add	r4,r4,r12			@ h+=Maj(a,b,c) from the past
+	mov	r12,r1,ror#17
+	eor	r0,r0,r2,ror#18
+	eor	r12,r12,r1,ror#19
+	eor	r0,r0,r2,lsr#3	@ sigma0(X[i+1])
+	ldr	r2,[sp,#8*4]
+	eor	r12,r12,r1,lsr#10	@ sigma1(X[i+14])
+	ldr	r1,[sp,#1*4]
+
+	add	r12,r12,r0
+	eor	r0,r8,r8,ror#5	@ from BODY_00_15
+	add	r2,r2,r12
+	eor	r0,r0,r8,ror#19	@ Sigma1(e)
+	add	r2,r2,r1			@ X[i]
+	ldr	r12,[r14],#4			@ *K256++
+	add	r11,r11,r2			@ h+=X[i]
+	str	r2,[sp,#8*4]
+	eor	r2,r9,r10
+	add	r11,r11,r0,ror#6	@ h+=Sigma1(e)
+	and	r2,r2,r8
+	add	r11,r11,r12			@ h+=K256[i]
+	eor	r2,r2,r10			@ Ch(e,f,g)
+	eor	r0,r4,r4,ror#11
+	add	r11,r11,r2			@ h+=Ch(e,f,g)
+#if 24==31
+	and	r12,r12,#0xff
+	cmp	r12,#0xf2			@ done?
+#endif
+#if 24<15
+# if __ARM_ARCH__>=7
+	ldr	r2,[r1],#4			@ prefetch
+# else
+	ldrb	r2,[r1,#3]
+# endif
+	eor	r12,r4,r5			@ a^b, b^c in next round
+#else
+	ldr	r2,[sp,#10*4]		@ from future BODY_16_xx
+	eor	r12,r4,r5			@ a^b, b^c in next round
+	ldr	r1,[sp,#7*4]	@ from future BODY_16_xx
+#endif
+	eor	r0,r0,r4,ror#20	@ Sigma0(a)
+	and	r3,r3,r12			@ (b^c)&=(a^b)
+	add	r7,r7,r11			@ d+=h
+	eor	r3,r3,r5			@ Maj(a,b,c)
+	add	r11,r11,r0,ror#2	@ h+=Sigma0(a)
+	@ add	r11,r11,r3			@ h+=Maj(a,b,c)
+	@ ldr	r2,[sp,#10*4]		@ 25
+	@ ldr	r1,[sp,#7*4]
+	mov	r0,r2,ror#7
+	add	r11,r11,r3			@ h+=Maj(a,b,c) from the past
+	mov	r3,r1,ror#17
+	eor	r0,r0,r2,ror#18
+	eor	r3,r3,r1,ror#19
+	eor	r0,r0,r2,lsr#3	@ sigma0(X[i+1])
+	ldr	r2,[sp,#9*4]
+	eor	r3,r3,r1,lsr#10	@ sigma1(X[i+14])
+	ldr	r1,[sp,#2*4]
+
+	add	r3,r3,r0
+	eor	r0,r7,r7,ror#5	@ from BODY_00_15
+	add	r2,r2,r3
+	eor	r0,r0,r7,ror#19	@ Sigma1(e)
+	add	r2,r2,r1			@ X[i]
+	ldr	r3,[r14],#4			@ *K256++
+	add	r10,r10,r2			@ h+=X[i]
+	str	r2,[sp,#9*4]
+	eor	r2,r8,r9
+	add	r10,r10,r0,ror#6	@ h+=Sigma1(e)
+	and	r2,r2,r7
+	add	r10,r10,r3			@ h+=K256[i]
+	eor	r2,r2,r9			@ Ch(e,f,g)
+	eor	r0,r11,r11,ror#11
+	add	r10,r10,r2			@ h+=Ch(e,f,g)
+#if 25==31
+	and	r3,r3,#0xff
+	cmp	r3,#0xf2			@ done?
+#endif
+#if 25<15
+# if __ARM_ARCH__>=7
+	ldr	r2,[r1],#4			@ prefetch
+# else
+	ldrb	r2,[r1,#3]
+# endif
+	eor	r3,r11,r4			@ a^b, b^c in next round
+#else
+	ldr	r2,[sp,#11*4]		@ from future BODY_16_xx
+	eor	r3,r11,r4			@ a^b, b^c in next round
+	ldr	r1,[sp,#8*4]	@ from future BODY_16_xx
+#endif
+	eor	r0,r0,r11,ror#20	@ Sigma0(a)
+	and	r12,r12,r3			@ (b^c)&=(a^b)
+	add	r6,r6,r10			@ d+=h
+	eor	r12,r12,r4			@ Maj(a,b,c)
+	add	r10,r10,r0,ror#2	@ h+=Sigma0(a)
+	@ add	r10,r10,r12			@ h+=Maj(a,b,c)
+	@ ldr	r2,[sp,#11*4]		@ 26
+	@ ldr	r1,[sp,#8*4]
+	mov	r0,r2,ror#7
+	add	r10,r10,r12			@ h+=Maj(a,b,c) from the past
+	mov	r12,r1,ror#17
+	eor	r0,r0,r2,ror#18
+	eor	r12,r12,r1,ror#19
+	eor	r0,r0,r2,lsr#3	@ sigma0(X[i+1])
+	ldr	r2,[sp,#10*4]
+	eor	r12,r12,r1,lsr#10	@ sigma1(X[i+14])
+	ldr	r1,[sp,#3*4]
+
+	add	r12,r12,r0
+	eor	r0,r6,r6,ror#5	@ from BODY_00_15
+	add	r2,r2,r12
+	eor	r0,r0,r6,ror#19	@ Sigma1(e)
+	add	r2,r2,r1			@ X[i]
+	ldr	r12,[r14],#4			@ *K256++
+	add	r9,r9,r2			@ h+=X[i]
+	str	r2,[sp,#10*4]
+	eor	r2,r7,r8
+	add	r9,r9,r0,ror#6	@ h+=Sigma1(e)
+	and	r2,r2,r6
+	add	r9,r9,r12			@ h+=K256[i]
+	eor	r2,r2,r8			@ Ch(e,f,g)
+	eor	r0,r10,r10,ror#11
+	add	r9,r9,r2			@ h+=Ch(e,f,g)
+#if 26==31
+	and	r12,r12,#0xff
+	cmp	r12,#0xf2			@ done?
+#endif
+#if 26<15
+# if __ARM_ARCH__>=7
+	ldr	r2,[r1],#4			@ prefetch
+# else
+	ldrb	r2,[r1,#3]
+# endif
+	eor	r12,r10,r11			@ a^b, b^c in next round
+#else
+	ldr	r2,[sp,#12*4]		@ from future BODY_16_xx
+	eor	r12,r10,r11			@ a^b, b^c in next round
+	ldr	r1,[sp,#9*4]	@ from future BODY_16_xx
+#endif
+	eor	r0,r0,r10,ror#20	@ Sigma0(a)
+	and	r3,r3,r12			@ (b^c)&=(a^b)
+	add	r5,r5,r9			@ d+=h
+	eor	r3,r3,r11			@ Maj(a,b,c)
+	add	r9,r9,r0,ror#2	@ h+=Sigma0(a)
+	@ add	r9,r9,r3			@ h+=Maj(a,b,c)
+	@ ldr	r2,[sp,#12*4]		@ 27
+	@ ldr	r1,[sp,#9*4]
+	mov	r0,r2,ror#7
+	add	r9,r9,r3			@ h+=Maj(a,b,c) from the past
+	mov	r3,r1,ror#17
+	eor	r0,r0,r2,ror#18
+	eor	r3,r3,r1,ror#19
+	eor	r0,r0,r2,lsr#3	@ sigma0(X[i+1])
+	ldr	r2,[sp,#11*4]
+	eor	r3,r3,r1,lsr#10	@ sigma1(X[i+14])
+	ldr	r1,[sp,#4*4]
+
+	add	r3,r3,r0
+	eor	r0,r5,r5,ror#5	@ from BODY_00_15
+	add	r2,r2,r3
+	eor	r0,r0,r5,ror#19	@ Sigma1(e)
+	add	r2,r2,r1			@ X[i]
+	ldr	r3,[r14],#4			@ *K256++
+	add	r8,r8,r2			@ h+=X[i]
+	str	r2,[sp,#11*4]
+	eor	r2,r6,r7
+	add	r8,r8,r0,ror#6	@ h+=Sigma1(e)
+	and	r2,r2,r5
+	add	r8,r8,r3			@ h+=K256[i]
+	eor	r2,r2,r7			@ Ch(e,f,g)
+	eor	r0,r9,r9,ror#11
+	add	r8,r8,r2			@ h+=Ch(e,f,g)
+#if 27==31
+	and	r3,r3,#0xff
+	cmp	r3,#0xf2			@ done?
+#endif
+#if 27<15
+# if __ARM_ARCH__>=7
+	ldr	r2,[r1],#4			@ prefetch
+# else
+	ldrb	r2,[r1,#3]
+# endif
+	eor	r3,r9,r10			@ a^b, b^c in next round
+#else
+	ldr	r2,[sp,#13*4]		@ from future BODY_16_xx
+	eor	r3,r9,r10			@ a^b, b^c in next round
+	ldr	r1,[sp,#10*4]	@ from future BODY_16_xx
+#endif
+	eor	r0,r0,r9,ror#20	@ Sigma0(a)
+	and	r12,r12,r3			@ (b^c)&=(a^b)
+	add	r4,r4,r8			@ d+=h
+	eor	r12,r12,r10			@ Maj(a,b,c)
+	add	r8,r8,r0,ror#2	@ h+=Sigma0(a)
+	@ add	r8,r8,r12			@ h+=Maj(a,b,c)
+	@ ldr	r2,[sp,#13*4]		@ 28
+	@ ldr	r1,[sp,#10*4]
+	mov	r0,r2,ror#7
+	add	r8,r8,r12			@ h+=Maj(a,b,c) from the past
+	mov	r12,r1,ror#17
+	eor	r0,r0,r2,ror#18
+	eor	r12,r12,r1,ror#19
+	eor	r0,r0,r2,lsr#3	@ sigma0(X[i+1])
+	ldr	r2,[sp,#12*4]
+	eor	r12,r12,r1,lsr#10	@ sigma1(X[i+14])
+	ldr	r1,[sp,#5*4]
+
+	add	r12,r12,r0
+	eor	r0,r4,r4,ror#5	@ from BODY_00_15
+	add	r2,r2,r12
+	eor	r0,r0,r4,ror#19	@ Sigma1(e)
+	add	r2,r2,r1			@ X[i]
+	ldr	r12,[r14],#4			@ *K256++
+	add	r7,r7,r2			@ h+=X[i]
+	str	r2,[sp,#12*4]
+	eor	r2,r5,r6
+	add	r7,r7,r0,ror#6	@ h+=Sigma1(e)
+	and	r2,r2,r4
+	add	r7,r7,r12			@ h+=K256[i]
+	eor	r2,r2,r6			@ Ch(e,f,g)
+	eor	r0,r8,r8,ror#11
+	add	r7,r7,r2			@ h+=Ch(e,f,g)
+#if 28==31
+	and	r12,r12,#0xff
+	cmp	r12,#0xf2			@ done?
+#endif
+#if 28<15
+# if __ARM_ARCH__>=7
+	ldr	r2,[r1],#4			@ prefetch
+# else
+	ldrb	r2,[r1,#3]
+# endif
+	eor	r12,r8,r9			@ a^b, b^c in next round
+#else
+	ldr	r2,[sp,#14*4]		@ from future BODY_16_xx
+	eor	r12,r8,r9			@ a^b, b^c in next round
+	ldr	r1,[sp,#11*4]	@ from future BODY_16_xx
+#endif
+	eor	r0,r0,r8,ror#20	@ Sigma0(a)
+	and	r3,r3,r12			@ (b^c)&=(a^b)
+	add	r11,r11,r7			@ d+=h
+	eor	r3,r3,r9			@ Maj(a,b,c)
+	add	r7,r7,r0,ror#2	@ h+=Sigma0(a)
+	@ add	r7,r7,r3			@ h+=Maj(a,b,c)
+	@ ldr	r2,[sp,#14*4]		@ 29
+	@ ldr	r1,[sp,#11*4]
+	mov	r0,r2,ror#7
+	add	r7,r7,r3			@ h+=Maj(a,b,c) from the past
+	mov	r3,r1,ror#17
+	eor	r0,r0,r2,ror#18
+	eor	r3,r3,r1,ror#19
+	eor	r0,r0,r2,lsr#3	@ sigma0(X[i+1])
+	ldr	r2,[sp,#13*4]
+	eor	r3,r3,r1,lsr#10	@ sigma1(X[i+14])
+	ldr	r1,[sp,#6*4]
+
+	add	r3,r3,r0
+	eor	r0,r11,r11,ror#5	@ from BODY_00_15
+	add	r2,r2,r3
+	eor	r0,r0,r11,ror#19	@ Sigma1(e)
+	add	r2,r2,r1			@ X[i]
+	ldr	r3,[r14],#4			@ *K256++
+	add	r6,r6,r2			@ h+=X[i]
+	str	r2,[sp,#13*4]
+	eor	r2,r4,r5
+	add	r6,r6,r0,ror#6	@ h+=Sigma1(e)
+	and	r2,r2,r11
+	add	r6,r6,r3			@ h+=K256[i]
+	eor	r2,r2,r5			@ Ch(e,f,g)
+	eor	r0,r7,r7,ror#11
+	add	r6,r6,r2			@ h+=Ch(e,f,g)
+#if 29==31
+	and	r3,r3,#0xff
+	cmp	r3,#0xf2			@ done?
+#endif
+#if 29<15
+# if __ARM_ARCH__>=7
+	ldr	r2,[r1],#4			@ prefetch
+# else
+	ldrb	r2,[r1,#3]
+# endif
+	eor	r3,r7,r8			@ a^b, b^c in next round
+#else
+	ldr	r2,[sp,#15*4]		@ from future BODY_16_xx
+	eor	r3,r7,r8			@ a^b, b^c in next round
+	ldr	r1,[sp,#12*4]	@ from future BODY_16_xx
+#endif
+	eor	r0,r0,r7,ror#20	@ Sigma0(a)
+	and	r12,r12,r3			@ (b^c)&=(a^b)
+	add	r10,r10,r6			@ d+=h
+	eor	r12,r12,r8			@ Maj(a,b,c)
+	add	r6,r6,r0,ror#2	@ h+=Sigma0(a)
+	@ add	r6,r6,r12			@ h+=Maj(a,b,c)
+	@ ldr	r2,[sp,#15*4]		@ 30
+	@ ldr	r1,[sp,#12*4]
+	mov	r0,r2,ror#7
+	add	r6,r6,r12			@ h+=Maj(a,b,c) from the past
+	mov	r12,r1,ror#17
+	eor	r0,r0,r2,ror#18
+	eor	r12,r12,r1,ror#19
+	eor	r0,r0,r2,lsr#3	@ sigma0(X[i+1])
+	ldr	r2,[sp,#14*4]
+	eor	r12,r12,r1,lsr#10	@ sigma1(X[i+14])
+	ldr	r1,[sp,#7*4]
+
+	add	r12,r12,r0
+	eor	r0,r10,r10,ror#5	@ from BODY_00_15
+	add	r2,r2,r12
+	eor	r0,r0,r10,ror#19	@ Sigma1(e)
+	add	r2,r2,r1			@ X[i]
+	ldr	r12,[r14],#4			@ *K256++
+	add	r5,r5,r2			@ h+=X[i]
+	str	r2,[sp,#14*4]
+	eor	r2,r11,r4
+	add	r5,r5,r0,ror#6	@ h+=Sigma1(e)
+	and	r2,r2,r10
+	add	r5,r5,r12			@ h+=K256[i]
+	eor	r2,r2,r4			@ Ch(e,f,g)
+	eor	r0,r6,r6,ror#11
+	add	r5,r5,r2			@ h+=Ch(e,f,g)
+#if 30==31
+	and	r12,r12,#0xff
+	cmp	r12,#0xf2			@ done?
+#endif
+#if 30<15
+# if __ARM_ARCH__>=7
+	ldr	r2,[r1],#4			@ prefetch
+# else
+	ldrb	r2,[r1,#3]
+# endif
+	eor	r12,r6,r7			@ a^b, b^c in next round
+#else
+	ldr	r2,[sp,#0*4]		@ from future BODY_16_xx
+	eor	r12,r6,r7			@ a^b, b^c in next round
+	ldr	r1,[sp,#13*4]	@ from future BODY_16_xx
+#endif
+	eor	r0,r0,r6,ror#20	@ Sigma0(a)
+	and	r3,r3,r12			@ (b^c)&=(a^b)
+	add	r9,r9,r5			@ d+=h
+	eor	r3,r3,r7			@ Maj(a,b,c)
+	add	r5,r5,r0,ror#2	@ h+=Sigma0(a)
+	@ add	r5,r5,r3			@ h+=Maj(a,b,c)
+	@ ldr	r2,[sp,#0*4]		@ 31
+	@ ldr	r1,[sp,#13*4]
+	mov	r0,r2,ror#7
+	add	r5,r5,r3			@ h+=Maj(a,b,c) from the past
+	mov	r3,r1,ror#17
+	eor	r0,r0,r2,ror#18
+	eor	r3,r3,r1,ror#19
+	eor	r0,r0,r2,lsr#3	@ sigma0(X[i+1])
+	ldr	r2,[sp,#15*4]
+	eor	r3,r3,r1,lsr#10	@ sigma1(X[i+14])
+	ldr	r1,[sp,#8*4]
+
+	add	r3,r3,r0
+	eor	r0,r9,r9,ror#5	@ from BODY_00_15
+	add	r2,r2,r3
+	eor	r0,r0,r9,ror#19	@ Sigma1(e)
+	add	r2,r2,r1			@ X[i]
+	ldr	r3,[r14],#4			@ *K256++
+	add	r4,r4,r2			@ h+=X[i]
+	str	r2,[sp,#15*4]
+	eor	r2,r10,r11
+	add	r4,r4,r0,ror#6	@ h+=Sigma1(e)
+	and	r2,r2,r9
+	add	r4,r4,r3			@ h+=K256[i]
+	eor	r2,r2,r11			@ Ch(e,f,g)
+	eor	r0,r5,r5,ror#11
+	add	r4,r4,r2			@ h+=Ch(e,f,g)
+#if 31==31
+	and	r3,r3,#0xff
+	cmp	r3,#0xf2			@ done?
+#endif
+#if 31<15
+# if __ARM_ARCH__>=7
+	ldr	r2,[r1],#4			@ prefetch
+# else
+	ldrb	r2,[r1,#3]
+# endif
+	eor	r3,r5,r6			@ a^b, b^c in next round
+#else
+	ldr	r2,[sp,#1*4]		@ from future BODY_16_xx
+	eor	r3,r5,r6			@ a^b, b^c in next round
+	ldr	r1,[sp,#14*4]	@ from future BODY_16_xx
+#endif
+	eor	r0,r0,r5,ror#20	@ Sigma0(a)
+	and	r12,r12,r3			@ (b^c)&=(a^b)
+	add	r8,r8,r4			@ d+=h
+	eor	r12,r12,r6			@ Maj(a,b,c)
+	add	r4,r4,r0,ror#2	@ h+=Sigma0(a)
+	@ add	r4,r4,r12			@ h+=Maj(a,b,c)
+#if __ARM_ARCH__>=7
+	ite	eq			@ Thumb2 thing, sanity check in ARM
+#endif
+	ldreq	r3,[sp,#16*4]		@ pull ctx
+	bne	.Lrounds_16_xx
+
+	add	r4,r4,r12		@ h+=Maj(a,b,c) from the past
+	ldr	r0,[r3,#0]
+	ldr	r2,[r3,#4]
+	ldr	r12,[r3,#8]
+	add	r4,r4,r0
+	ldr	r0,[r3,#12]
+	add	r5,r5,r2
+	ldr	r2,[r3,#16]
+	add	r6,r6,r12
+	ldr	r12,[r3,#20]
+	add	r7,r7,r0
+	ldr	r0,[r3,#24]
+	add	r8,r8,r2
+	ldr	r2,[r3,#28]
+	add	r9,r9,r12
+	ldr	r1,[sp,#17*4]		@ pull inp
+	ldr	r12,[sp,#18*4]		@ pull inp+len
+	add	r10,r10,r0
+	add	r11,r11,r2
+	stmia	r3,{r4,r5,r6,r7,r8,r9,r10,r11}
+	cmp	r1,r12
+	sub	r14,r14,#256	@ rewind Ktbl
+	bne	.Loop
+
+	add	sp,sp,#19*4	@ destroy frame
+#if __ARM_ARCH__>=5
+	ldmia	sp!,{r4-r11,pc}
+#else
+	ldmia	sp!,{r4-r11,lr}
+	tst	lr,#1
+	moveq	pc,lr			@ be binary compatible with V4, yet
+	.word	0xe12fff1e			@ interoperable with Thumb ISA:-)
+#endif
+.size	sha256_block_data_order,.-sha256_block_data_order
+#if __ARM_MAX_ARCH__>=7
+.arch	armv7-a
+.fpu	neon
+
+.global	sha256_block_data_order_neon
+.type	sha256_block_data_order_neon,%function
+.align	4
+sha256_block_data_order_neon:
+.LNEON:
+	stmdb	sp!,{r4-r12,lr}
+
+	sub	r11,sp,#16*4+16
+	adrl	r14,K256
+	bic	r11,r11,#15		@ align for 128-bit stores
+	mov	r12,sp
+	mov	sp,r11			@ alloca
+	add	r2,r1,r2,lsl#6	@ len to point at the end of inp
+
+	vld1.8		{q0},[r1]!
+	vld1.8		{q1},[r1]!
+	vld1.8		{q2},[r1]!
+	vld1.8		{q3},[r1]!
+	vld1.32		{q8},[r14,:128]!
+	vld1.32		{q9},[r14,:128]!
+	vld1.32		{q10},[r14,:128]!
+	vld1.32		{q11},[r14,:128]!
+	vrev32.8	q0,q0		@ yes, even on
+	str		r0,[sp,#64]
+	vrev32.8	q1,q1		@ big-endian
+	str		r1,[sp,#68]
+	mov		r1,sp
+	vrev32.8	q2,q2
+	str		r2,[sp,#72]
+	vrev32.8	q3,q3
+	str		r12,[sp,#76]		@ save original sp
+	vadd.i32	q8,q8,q0
+	vadd.i32	q9,q9,q1
+	vst1.32		{q8},[r1,:128]!
+	vadd.i32	q10,q10,q2
+	vst1.32		{q9},[r1,:128]!
+	vadd.i32	q11,q11,q3
+	vst1.32		{q10},[r1,:128]!
+	vst1.32		{q11},[r1,:128]!
+
+	ldmia		r0,{r4-r11}
+	sub		r1,r1,#64
+	ldr		r2,[sp,#0]
+	eor		r12,r12,r12
+	eor		r3,r5,r6
+	b		.L_00_48
+
+.align	4
+.L_00_48:
+	vext.8	q8,q0,q1,#4
+	add	r11,r11,r2
+	eor	r2,r9,r10
+	eor	r0,r8,r8,ror#5
+	vext.8	q9,q2,q3,#4
+	add	r4,r4,r12
+	and	r2,r2,r8
+	eor	r12,r0,r8,ror#19
+	vshr.u32	q10,q8,#7
+	eor	r0,r4,r4,ror#11
+	eor	r2,r2,r10
+	vadd.i32	q0,q0,q9
+	add	r11,r11,r12,ror#6
+	eor	r12,r4,r5
+	vshr.u32	q9,q8,#3
+	eor	r0,r0,r4,ror#20
+	add	r11,r11,r2
+	vsli.32	q10,q8,#25
+	ldr	r2,[sp,#4]
+	and	r3,r3,r12
+	vshr.u32	q11,q8,#18
+	add	r7,r7,r11
+	add	r11,r11,r0,ror#2
+	eor	r3,r3,r5
+	veor	q9,q9,q10
+	add	r10,r10,r2
+	vsli.32	q11,q8,#14
+	eor	r2,r8,r9
+	eor	r0,r7,r7,ror#5
+	vshr.u32	d24,d7,#17
+	add	r11,r11,r3
+	and	r2,r2,r7
+	veor	q9,q9,q11
+	eor	r3,r0,r7,ror#19
+	eor	r0,r11,r11,ror#11
+	vsli.32	d24,d7,#15
+	eor	r2,r2,r9
+	add	r10,r10,r3,ror#6
+	vshr.u32	d25,d7,#10
+	eor	r3,r11,r4
+	eor	r0,r0,r11,ror#20
+	vadd.i32	q0,q0,q9
+	add	r10,r10,r2
+	ldr	r2,[sp,#8]
+	veor	d25,d25,d24
+	and	r12,r12,r3
+	add	r6,r6,r10
+	vshr.u32	d24,d7,#19
+	add	r10,r10,r0,ror#2
+	eor	r12,r12,r4
+	vsli.32	d24,d7,#13
+	add	r9,r9,r2
+	eor	r2,r7,r8
+	veor	d25,d25,d24
+	eor	r0,r6,r6,ror#5
+	add	r10,r10,r12
+	vadd.i32	d0,d0,d25
+	and	r2,r2,r6
+	eor	r12,r0,r6,ror#19
+	vshr.u32	d24,d0,#17
+	eor	r0,r10,r10,ror#11
+	eor	r2,r2,r8
+	vsli.32	d24,d0,#15
+	add	r9,r9,r12,ror#6
+	eor	r12,r10,r11
+	vshr.u32	d25,d0,#10
+	eor	r0,r0,r10,ror#20
+	add	r9,r9,r2
+	veor	d25,d25,d24
+	ldr	r2,[sp,#12]
+	and	r3,r3,r12
+	vshr.u32	d24,d0,#19
+	add	r5,r5,r9
+	add	r9,r9,r0,ror#2
+	eor	r3,r3,r11
+	vld1.32	{q8},[r14,:128]!
+	add	r8,r8,r2
+	vsli.32	d24,d0,#13
+	eor	r2,r6,r7
+	eor	r0,r5,r5,ror#5
+	veor	d25,d25,d24
+	add	r9,r9,r3
+	and	r2,r2,r5
+	vadd.i32	d1,d1,d25
+	eor	r3,r0,r5,ror#19
+	eor	r0,r9,r9,ror#11
+	vadd.i32	q8,q8,q0
+	eor	r2,r2,r7
+	add	r8,r8,r3,ror#6
+	eor	r3,r9,r10
+	eor	r0,r0,r9,ror#20
+	add	r8,r8,r2
+	ldr	r2,[sp,#16]
+	and	r12,r12,r3
+	add	r4,r4,r8
+	vst1.32	{q8},[r1,:128]!
+	add	r8,r8,r0,ror#2
+	eor	r12,r12,r10
+	vext.8	q8,q1,q2,#4
+	add	r7,r7,r2
+	eor	r2,r5,r6
+	eor	r0,r4,r4,ror#5
+	vext.8	q9,q3,q0,#4
+	add	r8,r8,r12
+	and	r2,r2,r4
+	eor	r12,r0,r4,ror#19
+	vshr.u32	q10,q8,#7
+	eor	r0,r8,r8,ror#11
+	eor	r2,r2,r6
+	vadd.i32	q1,q1,q9
+	add	r7,r7,r12,ror#6
+	eor	r12,r8,r9
+	vshr.u32	q9,q8,#3
+	eor	r0,r0,r8,ror#20
+	add	r7,r7,r2
+	vsli.32	q10,q8,#25
+	ldr	r2,[sp,#20]
+	and	r3,r3,r12
+	vshr.u32	q11,q8,#18
+	add	r11,r11,r7
+	add	r7,r7,r0,ror#2
+	eor	r3,r3,r9
+	veor	q9,q9,q10
+	add	r6,r6,r2
+	vsli.32	q11,q8,#14
+	eor	r2,r4,r5
+	eor	r0,r11,r11,ror#5
+	vshr.u32	d24,d1,#17
+	add	r7,r7,r3
+	and	r2,r2,r11
+	veor	q9,q9,q11
+	eor	r3,r0,r11,ror#19
+	eor	r0,r7,r7,ror#11
+	vsli.32	d24,d1,#15
+	eor	r2,r2,r5
+	add	r6,r6,r3,ror#6
+	vshr.u32	d25,d1,#10
+	eor	r3,r7,r8
+	eor	r0,r0,r7,ror#20
+	vadd.i32	q1,q1,q9
+	add	r6,r6,r2
+	ldr	r2,[sp,#24]
+	veor	d25,d25,d24
+	and	r12,r12,r3
+	add	r10,r10,r6
+	vshr.u32	d24,d1,#19
+	add	r6,r6,r0,ror#2
+	eor	r12,r12,r8
+	vsli.32	d24,d1,#13
+	add	r5,r5,r2
+	eor	r2,r11,r4
+	veor	d25,d25,d24
+	eor	r0,r10,r10,ror#5
+	add	r6,r6,r12
+	vadd.i32	d2,d2,d25
+	and	r2,r2,r10
+	eor	r12,r0,r10,ror#19
+	vshr.u32	d24,d2,#17
+	eor	r0,r6,r6,ror#11
+	eor	r2,r2,r4
+	vsli.32	d24,d2,#15
+	add	r5,r5,r12,ror#6
+	eor	r12,r6,r7
+	vshr.u32	d25,d2,#10
+	eor	r0,r0,r6,ror#20
+	add	r5,r5,r2
+	veor	d25,d25,d24
+	ldr	r2,[sp,#28]
+	and	r3,r3,r12
+	vshr.u32	d24,d2,#19
+	add	r9,r9,r5
+	add	r5,r5,r0,ror#2
+	eor	r3,r3,r7
+	vld1.32	{q8},[r14,:128]!
+	add	r4,r4,r2
+	vsli.32	d24,d2,#13
+	eor	r2,r10,r11
+	eor	r0,r9,r9,ror#5
+	veor	d25,d25,d24
+	add	r5,r5,r3
+	and	r2,r2,r9
+	vadd.i32	d3,d3,d25
+	eor	r3,r0,r9,ror#19
+	eor	r0,r5,r5,ror#11
+	vadd.i32	q8,q8,q1
+	eor	r2,r2,r11
+	add	r4,r4,r3,ror#6
+	eor	r3,r5,r6
+	eor	r0,r0,r5,ror#20
+	add	r4,r4,r2
+	ldr	r2,[sp,#32]
+	and	r12,r12,r3
+	add	r8,r8,r4
+	vst1.32	{q8},[r1,:128]!
+	add	r4,r4,r0,ror#2
+	eor	r12,r12,r6
+	vext.8	q8,q2,q3,#4
+	add	r11,r11,r2
+	eor	r2,r9,r10
+	eor	r0,r8,r8,ror#5
+	vext.8	q9,q0,q1,#4
+	add	r4,r4,r12
+	and	r2,r2,r8
+	eor	r12,r0,r8,ror#19
+	vshr.u32	q10,q8,#7
+	eor	r0,r4,r4,ror#11
+	eor	r2,r2,r10
+	vadd.i32	q2,q2,q9
+	add	r11,r11,r12,ror#6
+	eor	r12,r4,r5
+	vshr.u32	q9,q8,#3
+	eor	r0,r0,r4,ror#20
+	add	r11,r11,r2
+	vsli.32	q10,q8,#25
+	ldr	r2,[sp,#36]
+	and	r3,r3,r12
+	vshr.u32	q11,q8,#18
+	add	r7,r7,r11
+	add	r11,r11,r0,ror#2
+	eor	r3,r3,r5
+	veor	q9,q9,q10
+	add	r10,r10,r2
+	vsli.32	q11,q8,#14
+	eor	r2,r8,r9
+	eor	r0,r7,r7,ror#5
+	vshr.u32	d24,d3,#17
+	add	r11,r11,r3
+	and	r2,r2,r7
+	veor	q9,q9,q11
+	eor	r3,r0,r7,ror#19
+	eor	r0,r11,r11,ror#11
+	vsli.32	d24,d3,#15
+	eor	r2,r2,r9
+	add	r10,r10,r3,ror#6
+	vshr.u32	d25,d3,#10
+	eor	r3,r11,r4
+	eor	r0,r0,r11,ror#20
+	vadd.i32	q2,q2,q9
+	add	r10,r10,r2
+	ldr	r2,[sp,#40]
+	veor	d25,d25,d24
+	and	r12,r12,r3
+	add	r6,r6,r10
+	vshr.u32	d24,d3,#19
+	add	r10,r10,r0,ror#2
+	eor	r12,r12,r4
+	vsli.32	d24,d3,#13
+	add	r9,r9,r2
+	eor	r2,r7,r8
+	veor	d25,d25,d24
+	eor	r0,r6,r6,ror#5
+	add	r10,r10,r12
+	vadd.i32	d4,d4,d25
+	and	r2,r2,r6
+	eor	r12,r0,r6,ror#19
+	vshr.u32	d24,d4,#17
+	eor	r0,r10,r10,ror#11
+	eor	r2,r2,r8
+	vsli.32	d24,d4,#15
+	add	r9,r9,r12,ror#6
+	eor	r12,r10,r11
+	vshr.u32	d25,d4,#10
+	eor	r0,r0,r10,ror#20
+	add	r9,r9,r2
+	veor	d25,d25,d24
+	ldr	r2,[sp,#44]
+	and	r3,r3,r12
+	vshr.u32	d24,d4,#19
+	add	r5,r5,r9
+	add	r9,r9,r0,ror#2
+	eor	r3,r3,r11
+	vld1.32	{q8},[r14,:128]!
+	add	r8,r8,r2
+	vsli.32	d24,d4,#13
+	eor	r2,r6,r7
+	eor	r0,r5,r5,ror#5
+	veor	d25,d25,d24
+	add	r9,r9,r3
+	and	r2,r2,r5
+	vadd.i32	d5,d5,d25
+	eor	r3,r0,r5,ror#19
+	eor	r0,r9,r9,ror#11
+	vadd.i32	q8,q8,q2
+	eor	r2,r2,r7
+	add	r8,r8,r3,ror#6
+	eor	r3,r9,r10
+	eor	r0,r0,r9,ror#20
+	add	r8,r8,r2
+	ldr	r2,[sp,#48]
+	and	r12,r12,r3
+	add	r4,r4,r8
+	vst1.32	{q8},[r1,:128]!
+	add	r8,r8,r0,ror#2
+	eor	r12,r12,r10
+	vext.8	q8,q3,q0,#4
+	add	r7,r7,r2
+	eor	r2,r5,r6
+	eor	r0,r4,r4,ror#5
+	vext.8	q9,q1,q2,#4
+	add	r8,r8,r12
+	and	r2,r2,r4
+	eor	r12,r0,r4,ror#19
+	vshr.u32	q10,q8,#7
+	eor	r0,r8,r8,ror#11
+	eor	r2,r2,r6
+	vadd.i32	q3,q3,q9
+	add	r7,r7,r12,ror#6
+	eor	r12,r8,r9
+	vshr.u32	q9,q8,#3
+	eor	r0,r0,r8,ror#20
+	add	r7,r7,r2
+	vsli.32	q10,q8,#25
+	ldr	r2,[sp,#52]
+	and	r3,r3,r12
+	vshr.u32	q11,q8,#18
+	add	r11,r11,r7
+	add	r7,r7,r0,ror#2
+	eor	r3,r3,r9
+	veor	q9,q9,q10
+	add	r6,r6,r2
+	vsli.32	q11,q8,#14
+	eor	r2,r4,r5
+	eor	r0,r11,r11,ror#5
+	vshr.u32	d24,d5,#17
+	add	r7,r7,r3
+	and	r2,r2,r11
+	veor	q9,q9,q11
+	eor	r3,r0,r11,ror#19
+	eor	r0,r7,r7,ror#11
+	vsli.32	d24,d5,#15
+	eor	r2,r2,r5
+	add	r6,r6,r3,ror#6
+	vshr.u32	d25,d5,#10
+	eor	r3,r7,r8
+	eor	r0,r0,r7,ror#20
+	vadd.i32	q3,q3,q9
+	add	r6,r6,r2
+	ldr	r2,[sp,#56]
+	veor	d25,d25,d24
+	and	r12,r12,r3
+	add	r10,r10,r6
+	vshr.u32	d24,d5,#19
+	add	r6,r6,r0,ror#2
+	eor	r12,r12,r8
+	vsli.32	d24,d5,#13
+	add	r5,r5,r2
+	eor	r2,r11,r4
+	veor	d25,d25,d24
+	eor	r0,r10,r10,ror#5
+	add	r6,r6,r12
+	vadd.i32	d6,d6,d25
+	and	r2,r2,r10
+	eor	r12,r0,r10,ror#19
+	vshr.u32	d24,d6,#17
+	eor	r0,r6,r6,ror#11
+	eor	r2,r2,r4
+	vsli.32	d24,d6,#15
+	add	r5,r5,r12,ror#6
+	eor	r12,r6,r7
+	vshr.u32	d25,d6,#10
+	eor	r0,r0,r6,ror#20
+	add	r5,r5,r2
+	veor	d25,d25,d24
+	ldr	r2,[sp,#60]
+	and	r3,r3,r12
+	vshr.u32	d24,d6,#19
+	add	r9,r9,r5
+	add	r5,r5,r0,ror#2
+	eor	r3,r3,r7
+	vld1.32	{q8},[r14,:128]!
+	add	r4,r4,r2
+	vsli.32	d24,d6,#13
+	eor	r2,r10,r11
+	eor	r0,r9,r9,ror#5
+	veor	d25,d25,d24
+	add	r5,r5,r3
+	and	r2,r2,r9
+	vadd.i32	d7,d7,d25
+	eor	r3,r0,r9,ror#19
+	eor	r0,r5,r5,ror#11
+	vadd.i32	q8,q8,q3
+	eor	r2,r2,r11
+	add	r4,r4,r3,ror#6
+	eor	r3,r5,r6
+	eor	r0,r0,r5,ror#20
+	add	r4,r4,r2
+	ldr	r2,[r14]
+	and	r12,r12,r3
+	add	r8,r8,r4
+	vst1.32	{q8},[r1,:128]!
+	add	r4,r4,r0,ror#2
+	eor	r12,r12,r6
+	teq	r2,#0				@ check for K256 terminator
+	ldr	r2,[sp,#0]
+	sub	r1,r1,#64
+	bne	.L_00_48
+
+	ldr		r1,[sp,#68]
+	ldr		r0,[sp,#72]
+	sub		r14,r14,#256	@ rewind r14
+	teq		r1,r0
+	it		eq
+	subeq		r1,r1,#64		@ avoid SEGV
+	vld1.8		{q0},[r1]!		@ load next input block
+	vld1.8		{q1},[r1]!
+	vld1.8		{q2},[r1]!
+	vld1.8		{q3},[r1]!
+	it		ne
+	strne		r1,[sp,#68]
+	mov		r1,sp
+	add	r11,r11,r2
+	eor	r2,r9,r10
+	eor	r0,r8,r8,ror#5
+	add	r4,r4,r12
+	vld1.32	{q8},[r14,:128]!
+	and	r2,r2,r8
+	eor	r12,r0,r8,ror#19
+	eor	r0,r4,r4,ror#11
+	eor	r2,r2,r10
+	vrev32.8	q0,q0
+	add	r11,r11,r12,ror#6
+	eor	r12,r4,r5
+	eor	r0,r0,r4,ror#20
+	add	r11,r11,r2
+	vadd.i32	q8,q8,q0
+	ldr	r2,[sp,#4]
+	and	r3,r3,r12
+	add	r7,r7,r11
+	add	r11,r11,r0,ror#2
+	eor	r3,r3,r5
+	add	r10,r10,r2
+	eor	r2,r8,r9
+	eor	r0,r7,r7,ror#5
+	add	r11,r11,r3
+	and	r2,r2,r7
+	eor	r3,r0,r7,ror#19
+	eor	r0,r11,r11,ror#11
+	eor	r2,r2,r9
+	add	r10,r10,r3,ror#6
+	eor	r3,r11,r4
+	eor	r0,r0,r11,ror#20
+	add	r10,r10,r2
+	ldr	r2,[sp,#8]
+	and	r12,r12,r3
+	add	r6,r6,r10
+	add	r10,r10,r0,ror#2
+	eor	r12,r12,r4
+	add	r9,r9,r2
+	eor	r2,r7,r8
+	eor	r0,r6,r6,ror#5
+	add	r10,r10,r12
+	and	r2,r2,r6
+	eor	r12,r0,r6,ror#19
+	eor	r0,r10,r10,ror#11
+	eor	r2,r2,r8
+	add	r9,r9,r12,ror#6
+	eor	r12,r10,r11
+	eor	r0,r0,r10,ror#20
+	add	r9,r9,r2
+	ldr	r2,[sp,#12]
+	and	r3,r3,r12
+	add	r5,r5,r9
+	add	r9,r9,r0,ror#2
+	eor	r3,r3,r11
+	add	r8,r8,r2
+	eor	r2,r6,r7
+	eor	r0,r5,r5,ror#5
+	add	r9,r9,r3
+	and	r2,r2,r5
+	eor	r3,r0,r5,ror#19
+	eor	r0,r9,r9,ror#11
+	eor	r2,r2,r7
+	add	r8,r8,r3,ror#6
+	eor	r3,r9,r10
+	eor	r0,r0,r9,ror#20
+	add	r8,r8,r2
+	ldr	r2,[sp,#16]
+	and	r12,r12,r3
+	add	r4,r4,r8
+	add	r8,r8,r0,ror#2
+	eor	r12,r12,r10
+	vst1.32	{q8},[r1,:128]!
+	add	r7,r7,r2
+	eor	r2,r5,r6
+	eor	r0,r4,r4,ror#5
+	add	r8,r8,r12
+	vld1.32	{q8},[r14,:128]!
+	and	r2,r2,r4
+	eor	r12,r0,r4,ror#19
+	eor	r0,r8,r8,ror#11
+	eor	r2,r2,r6
+	vrev32.8	q1,q1
+	add	r7,r7,r12,ror#6
+	eor	r12,r8,r9
+	eor	r0,r0,r8,ror#20
+	add	r7,r7,r2
+	vadd.i32	q8,q8,q1
+	ldr	r2,[sp,#20]
+	and	r3,r3,r12
+	add	r11,r11,r7
+	add	r7,r7,r0,ror#2
+	eor	r3,r3,r9
+	add	r6,r6,r2
+	eor	r2,r4,r5
+	eor	r0,r11,r11,ror#5
+	add	r7,r7,r3
+	and	r2,r2,r11
+	eor	r3,r0,r11,ror#19
+	eor	r0,r7,r7,ror#11
+	eor	r2,r2,r5
+	add	r6,r6,r3,ror#6
+	eor	r3,r7,r8
+	eor	r0,r0,r7,ror#20
+	add	r6,r6,r2
+	ldr	r2,[sp,#24]
+	and	r12,r12,r3
+	add	r10,r10,r6
+	add	r6,r6,r0,ror#2
+	eor	r12,r12,r8
+	add	r5,r5,r2
+	eor	r2,r11,r4
+	eor	r0,r10,r10,ror#5
+	add	r6,r6,r12
+	and	r2,r2,r10
+	eor	r12,r0,r10,ror#19
+	eor	r0,r6,r6,ror#11
+	eor	r2,r2,r4
+	add	r5,r5,r12,ror#6
+	eor	r12,r6,r7
+	eor	r0,r0,r6,ror#20
+	add	r5,r5,r2
+	ldr	r2,[sp,#28]
+	and	r3,r3,r12
+	add	r9,r9,r5
+	add	r5,r5,r0,ror#2
+	eor	r3,r3,r7
+	add	r4,r4,r2
+	eor	r2,r10,r11
+	eor	r0,r9,r9,ror#5
+	add	r5,r5,r3
+	and	r2,r2,r9
+	eor	r3,r0,r9,ror#19
+	eor	r0,r5,r5,ror#11
+	eor	r2,r2,r11
+	add	r4,r4,r3,ror#6
+	eor	r3,r5,r6
+	eor	r0,r0,r5,ror#20
+	add	r4,r4,r2
+	ldr	r2,[sp,#32]
+	and	r12,r12,r3
+	add	r8,r8,r4
+	add	r4,r4,r0,ror#2
+	eor	r12,r12,r6
+	vst1.32	{q8},[r1,:128]!
+	add	r11,r11,r2
+	eor	r2,r9,r10
+	eor	r0,r8,r8,ror#5
+	add	r4,r4,r12
+	vld1.32	{q8},[r14,:128]!
+	and	r2,r2,r8
+	eor	r12,r0,r8,ror#19
+	eor	r0,r4,r4,ror#11
+	eor	r2,r2,r10
+	vrev32.8	q2,q2
+	add	r11,r11,r12,ror#6
+	eor	r12,r4,r5
+	eor	r0,r0,r4,ror#20
+	add	r11,r11,r2
+	vadd.i32	q8,q8,q2
+	ldr	r2,[sp,#36]
+	and	r3,r3,r12
+	add	r7,r7,r11
+	add	r11,r11,r0,ror#2
+	eor	r3,r3,r5
+	add	r10,r10,r2
+	eor	r2,r8,r9
+	eor	r0,r7,r7,ror#5
+	add	r11,r11,r3
+	and	r2,r2,r7
+	eor	r3,r0,r7,ror#19
+	eor	r0,r11,r11,ror#11
+	eor	r2,r2,r9
+	add	r10,r10,r3,ror#6
+	eor	r3,r11,r4
+	eor	r0,r0,r11,ror#20
+	add	r10,r10,r2
+	ldr	r2,[sp,#40]
+	and	r12,r12,r3
+	add	r6,r6,r10
+	add	r10,r10,r0,ror#2
+	eor	r12,r12,r4
+	add	r9,r9,r2
+	eor	r2,r7,r8
+	eor	r0,r6,r6,ror#5
+	add	r10,r10,r12
+	and	r2,r2,r6
+	eor	r12,r0,r6,ror#19
+	eor	r0,r10,r10,ror#11
+	eor	r2,r2,r8
+	add	r9,r9,r12,ror#6
+	eor	r12,r10,r11
+	eor	r0,r0,r10,ror#20
+	add	r9,r9,r2
+	ldr	r2,[sp,#44]
+	and	r3,r3,r12
+	add	r5,r5,r9
+	add	r9,r9,r0,ror#2
+	eor	r3,r3,r11
+	add	r8,r8,r2
+	eor	r2,r6,r7
+	eor	r0,r5,r5,ror#5
+	add	r9,r9,r3
+	and	r2,r2,r5
+	eor	r3,r0,r5,ror#19
+	eor	r0,r9,r9,ror#11
+	eor	r2,r2,r7
+	add	r8,r8,r3,ror#6
+	eor	r3,r9,r10
+	eor	r0,r0,r9,ror#20
+	add	r8,r8,r2
+	ldr	r2,[sp,#48]
+	and	r12,r12,r3
+	add	r4,r4,r8
+	add	r8,r8,r0,ror#2
+	eor	r12,r12,r10
+	vst1.32	{q8},[r1,:128]!
+	add	r7,r7,r2
+	eor	r2,r5,r6
+	eor	r0,r4,r4,ror#5
+	add	r8,r8,r12
+	vld1.32	{q8},[r14,:128]!
+	and	r2,r2,r4
+	eor	r12,r0,r4,ror#19
+	eor	r0,r8,r8,ror#11
+	eor	r2,r2,r6
+	vrev32.8	q3,q3
+	add	r7,r7,r12,ror#6
+	eor	r12,r8,r9
+	eor	r0,r0,r8,ror#20
+	add	r7,r7,r2
+	vadd.i32	q8,q8,q3
+	ldr	r2,[sp,#52]
+	and	r3,r3,r12
+	add	r11,r11,r7
+	add	r7,r7,r0,ror#2
+	eor	r3,r3,r9
+	add	r6,r6,r2
+	eor	r2,r4,r5
+	eor	r0,r11,r11,ror#5
+	add	r7,r7,r3
+	and	r2,r2,r11
+	eor	r3,r0,r11,ror#19
+	eor	r0,r7,r7,ror#11
+	eor	r2,r2,r5
+	add	r6,r6,r3,ror#6
+	eor	r3,r7,r8
+	eor	r0,r0,r7,ror#20
+	add	r6,r6,r2
+	ldr	r2,[sp,#56]
+	and	r12,r12,r3
+	add	r10,r10,r6
+	add	r6,r6,r0,ror#2
+	eor	r12,r12,r8
+	add	r5,r5,r2
+	eor	r2,r11,r4
+	eor	r0,r10,r10,ror#5
+	add	r6,r6,r12
+	and	r2,r2,r10
+	eor	r12,r0,r10,ror#19
+	eor	r0,r6,r6,ror#11
+	eor	r2,r2,r4
+	add	r5,r5,r12,ror#6
+	eor	r12,r6,r7
+	eor	r0,r0,r6,ror#20
+	add	r5,r5,r2
+	ldr	r2,[sp,#60]
+	and	r3,r3,r12
+	add	r9,r9,r5
+	add	r5,r5,r0,ror#2
+	eor	r3,r3,r7
+	add	r4,r4,r2
+	eor	r2,r10,r11
+	eor	r0,r9,r9,ror#5
+	add	r5,r5,r3
+	and	r2,r2,r9
+	eor	r3,r0,r9,ror#19
+	eor	r0,r5,r5,ror#11
+	eor	r2,r2,r11
+	add	r4,r4,r3,ror#6
+	eor	r3,r5,r6
+	eor	r0,r0,r5,ror#20
+	add	r4,r4,r2
+	ldr	r2,[sp,#64]
+	and	r12,r12,r3
+	add	r8,r8,r4
+	add	r4,r4,r0,ror#2
+	eor	r12,r12,r6
+	vst1.32	{q8},[r1,:128]!
+	ldr	r0,[r2,#0]
+	add	r4,r4,r12			@ h+=Maj(a,b,c) from the past
+	ldr	r12,[r2,#4]
+	ldr	r3,[r2,#8]
+	ldr	r1,[r2,#12]
+	add	r4,r4,r0			@ accumulate
+	ldr	r0,[r2,#16]
+	add	r5,r5,r12
+	ldr	r12,[r2,#20]
+	add	r6,r6,r3
+	ldr	r3,[r2,#24]
+	add	r7,r7,r1
+	ldr	r1,[r2,#28]
+	add	r8,r8,r0
+	str	r4,[r2],#4
+	add	r9,r9,r12
+	str	r5,[r2],#4
+	add	r10,r10,r3
+	str	r6,[r2],#4
+	add	r11,r11,r1
+	str	r7,[r2],#4
+	stmia	r2,{r8-r11}
+
+	ittte	ne
+	movne	r1,sp
+	ldrne	r2,[sp,#0]
+	eorne	r12,r12,r12
+	ldreq	sp,[sp,#76]			@ restore original sp
+	itt	ne
+	eorne	r3,r5,r6
+	bne	.L_00_48
+
+	ldmia	sp!,{r4-r12,pc}
+.size	sha256_block_data_order_neon,.-sha256_block_data_order_neon
+#endif
+#if __ARM_MAX_ARCH__>=7 && !defined(__KERNEL__)
+
+# ifdef __thumb2__
+#  define INST(a,b,c,d)	.byte	c,d|0xc,a,b
+# else
+#  define INST(a,b,c,d)	.byte	a,b,c,d
+# endif
+
+.type	sha256_block_data_order_armv8,%function
+.align	5
+sha256_block_data_order_armv8:
+.LARMv8:
+	vld1.32	{q0,q1},[r0]
+# ifdef __thumb2__
+	adr	r3,.LARMv8
+	sub	r3,r3,#.LARMv8-K256
+# else
+	adrl	r3,K256
+# endif
+	add	r2,r1,r2,lsl#6	@ len to point at the end of inp
+
+.Loop_v8:
+	vld1.8		{q8-q9},[r1]!
+	vld1.8		{q10-q11},[r1]!
+	vld1.32		{q12},[r3]!
+	vrev32.8	q8,q8
+	vrev32.8	q9,q9
+	vrev32.8	q10,q10
+	vrev32.8	q11,q11
+	vmov		q14,q0	@ offload
+	vmov		q15,q1
+	teq		r1,r2
+	vld1.32		{q13},[r3]!
+	vadd.i32	q12,q12,q8
+	INST(0xe2,0x03,0xfa,0xf3)	@ sha256su0 q8,q9
+	vmov		q2,q0
+	INST(0x68,0x0c,0x02,0xf3)	@ sha256h q0,q1,q12
+	INST(0x68,0x2c,0x14,0xf3)	@ sha256h2 q1,q2,q12
+	INST(0xe6,0x0c,0x64,0xf3)	@ sha256su1 q8,q10,q11
+	vld1.32		{q12},[r3]!
+	vadd.i32	q13,q13,q9
+	INST(0xe4,0x23,0xfa,0xf3)	@ sha256su0 q9,q10
+	vmov		q2,q0
+	INST(0x6a,0x0c,0x02,0xf3)	@ sha256h q0,q1,q13
+	INST(0x6a,0x2c,0x14,0xf3)	@ sha256h2 q1,q2,q13
+	INST(0xe0,0x2c,0x66,0xf3)	@ sha256su1 q9,q11,q8
+	vld1.32		{q13},[r3]!
+	vadd.i32	q12,q12,q10
+	INST(0xe6,0x43,0xfa,0xf3)	@ sha256su0 q10,q11
+	vmov		q2,q0
+	INST(0x68,0x0c,0x02,0xf3)	@ sha256h q0,q1,q12
+	INST(0x68,0x2c,0x14,0xf3)	@ sha256h2 q1,q2,q12
+	INST(0xe2,0x4c,0x60,0xf3)	@ sha256su1 q10,q8,q9
+	vld1.32		{q12},[r3]!
+	vadd.i32	q13,q13,q11
+	INST(0xe0,0x63,0xfa,0xf3)	@ sha256su0 q11,q8
+	vmov		q2,q0
+	INST(0x6a,0x0c,0x02,0xf3)	@ sha256h q0,q1,q13
+	INST(0x6a,0x2c,0x14,0xf3)	@ sha256h2 q1,q2,q13
+	INST(0xe4,0x6c,0x62,0xf3)	@ sha256su1 q11,q9,q10
+	vld1.32		{q13},[r3]!
+	vadd.i32	q12,q12,q8
+	INST(0xe2,0x03,0xfa,0xf3)	@ sha256su0 q8,q9
+	vmov		q2,q0
+	INST(0x68,0x0c,0x02,0xf3)	@ sha256h q0,q1,q12
+	INST(0x68,0x2c,0x14,0xf3)	@ sha256h2 q1,q2,q12
+	INST(0xe6,0x0c,0x64,0xf3)	@ sha256su1 q8,q10,q11
+	vld1.32		{q12},[r3]!
+	vadd.i32	q13,q13,q9
+	INST(0xe4,0x23,0xfa,0xf3)	@ sha256su0 q9,q10
+	vmov		q2,q0
+	INST(0x6a,0x0c,0x02,0xf3)	@ sha256h q0,q1,q13
+	INST(0x6a,0x2c,0x14,0xf3)	@ sha256h2 q1,q2,q13
+	INST(0xe0,0x2c,0x66,0xf3)	@ sha256su1 q9,q11,q8
+	vld1.32		{q13},[r3]!
+	vadd.i32	q12,q12,q10
+	INST(0xe6,0x43,0xfa,0xf3)	@ sha256su0 q10,q11
+	vmov		q2,q0
+	INST(0x68,0x0c,0x02,0xf3)	@ sha256h q0,q1,q12
+	INST(0x68,0x2c,0x14,0xf3)	@ sha256h2 q1,q2,q12
+	INST(0xe2,0x4c,0x60,0xf3)	@ sha256su1 q10,q8,q9
+	vld1.32		{q12},[r3]!
+	vadd.i32	q13,q13,q11
+	INST(0xe0,0x63,0xfa,0xf3)	@ sha256su0 q11,q8
+	vmov		q2,q0
+	INST(0x6a,0x0c,0x02,0xf3)	@ sha256h q0,q1,q13
+	INST(0x6a,0x2c,0x14,0xf3)	@ sha256h2 q1,q2,q13
+	INST(0xe4,0x6c,0x62,0xf3)	@ sha256su1 q11,q9,q10
+	vld1.32		{q13},[r3]!
+	vadd.i32	q12,q12,q8
+	INST(0xe2,0x03,0xfa,0xf3)	@ sha256su0 q8,q9
+	vmov		q2,q0
+	INST(0x68,0x0c,0x02,0xf3)	@ sha256h q0,q1,q12
+	INST(0x68,0x2c,0x14,0xf3)	@ sha256h2 q1,q2,q12
+	INST(0xe6,0x0c,0x64,0xf3)	@ sha256su1 q8,q10,q11
+	vld1.32		{q12},[r3]!
+	vadd.i32	q13,q13,q9
+	INST(0xe4,0x23,0xfa,0xf3)	@ sha256su0 q9,q10
+	vmov		q2,q0
+	INST(0x6a,0x0c,0x02,0xf3)	@ sha256h q0,q1,q13
+	INST(0x6a,0x2c,0x14,0xf3)	@ sha256h2 q1,q2,q13
+	INST(0xe0,0x2c,0x66,0xf3)	@ sha256su1 q9,q11,q8
+	vld1.32		{q13},[r3]!
+	vadd.i32	q12,q12,q10
+	INST(0xe6,0x43,0xfa,0xf3)	@ sha256su0 q10,q11
+	vmov		q2,q0
+	INST(0x68,0x0c,0x02,0xf3)	@ sha256h q0,q1,q12
+	INST(0x68,0x2c,0x14,0xf3)	@ sha256h2 q1,q2,q12
+	INST(0xe2,0x4c,0x60,0xf3)	@ sha256su1 q10,q8,q9
+	vld1.32		{q12},[r3]!
+	vadd.i32	q13,q13,q11
+	INST(0xe0,0x63,0xfa,0xf3)	@ sha256su0 q11,q8
+	vmov		q2,q0
+	INST(0x6a,0x0c,0x02,0xf3)	@ sha256h q0,q1,q13
+	INST(0x6a,0x2c,0x14,0xf3)	@ sha256h2 q1,q2,q13
+	INST(0xe4,0x6c,0x62,0xf3)	@ sha256su1 q11,q9,q10
+	vld1.32		{q13},[r3]!
+	vadd.i32	q12,q12,q8
+	vmov		q2,q0
+	INST(0x68,0x0c,0x02,0xf3)	@ sha256h q0,q1,q12
+	INST(0x68,0x2c,0x14,0xf3)	@ sha256h2 q1,q2,q12
+
+	vld1.32		{q12},[r3]!
+	vadd.i32	q13,q13,q9
+	vmov		q2,q0
+	INST(0x6a,0x0c,0x02,0xf3)	@ sha256h q0,q1,q13
+	INST(0x6a,0x2c,0x14,0xf3)	@ sha256h2 q1,q2,q13
+
+	vld1.32		{q13},[r3]
+	vadd.i32	q12,q12,q10
+	sub		r3,r3,#256-16	@ rewind
+	vmov		q2,q0
+	INST(0x68,0x0c,0x02,0xf3)	@ sha256h q0,q1,q12
+	INST(0x68,0x2c,0x14,0xf3)	@ sha256h2 q1,q2,q12
+
+	vadd.i32	q13,q13,q11
+	vmov		q2,q0
+	INST(0x6a,0x0c,0x02,0xf3)	@ sha256h q0,q1,q13
+	INST(0x6a,0x2c,0x14,0xf3)	@ sha256h2 q1,q2,q13
+
+	vadd.i32	q0,q0,q14
+	vadd.i32	q1,q1,q15
+	it		ne
+	bne		.Loop_v8
+
+	vst1.32		{q0,q1},[r0]
+
+	bx	lr		@ bx lr
+.size	sha256_block_data_order_armv8,.-sha256_block_data_order_armv8
+#endif
+.asciz  "SHA256 block transform for ARMv4/NEON/ARMv8, CRYPTOGAMS by <appro@openssl.org>"
+.align	2
+#if __ARM_MAX_ARCH__>=7 && !defined(__KERNEL__)
+.comm   OPENSSL_armcap_P,4,4
+#endif
diff --git a/arch/arm/crypto/sha256_glue.c b/arch/arm/crypto/sha256_glue.c
new file mode 100644
index 000000000..a84e869ef
--- /dev/null
+++ b/arch/arm/crypto/sha256_glue.c
@@ -0,0 +1,128 @@
+/*
+ * Glue code for the SHA256 Secure Hash Algorithm assembly implementation
+ * using optimized ARM assembler and NEON instructions.
+ *
+ * Copyright © 2015 Google Inc.
+ *
+ * This file is based on sha256_ssse3_glue.c:
+ *   Copyright (C) 2013 Intel Corporation
+ *   Author: Tim Chen <tim.c.chen@linux.intel.com>
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License as published by the Free
+ * Software Foundation; either version 2 of the License, or (at your option)
+ * any later version.
+ *
+ */
+
+#include <crypto/internal/hash.h>
+#include <linux/crypto.h>
+#include <linux/init.h>
+#include <linux/module.h>
+#include <linux/mm.h>
+#include <linux/cryptohash.h>
+#include <linux/types.h>
+#include <linux/string.h>
+#include <crypto/sha.h>
+#include <crypto/sha256_base.h>
+#include <asm/simd.h>
+#include <asm/neon.h>
+
+#include "sha256_glue.h"
+
+asmlinkage void sha256_block_data_order(u32 *digest, const void *data,
+					unsigned int num_blks);
+
+int crypto_sha256_arm_update(struct shash_desc *desc, const u8 *data,
+			     unsigned int len)
+{
+	/* make sure casting to sha256_block_fn() is safe */
+	BUILD_BUG_ON(offsetof(struct sha256_state, state) != 0);
+
+	return sha256_base_do_update(desc, data, len,
+				(sha256_block_fn *)sha256_block_data_order);
+}
+EXPORT_SYMBOL(crypto_sha256_arm_update);
+
+static int sha256_final(struct shash_desc *desc, u8 *out)
+{
+	sha256_base_do_finalize(desc,
+				(sha256_block_fn *)sha256_block_data_order);
+	return sha256_base_finish(desc, out);
+}
+
+int crypto_sha256_arm_finup(struct shash_desc *desc, const u8 *data,
+			    unsigned int len, u8 *out)
+{
+	sha256_base_do_update(desc, data, len,
+			      (sha256_block_fn *)sha256_block_data_order);
+	return sha256_final(desc, out);
+}
+EXPORT_SYMBOL(crypto_sha256_arm_finup);
+
+static struct shash_alg algs[] = { {
+	.digestsize	=	SHA256_DIGEST_SIZE,
+	.init		=	sha256_base_init,
+	.update		=	crypto_sha256_arm_update,
+	.final		=	sha256_final,
+	.finup		=	crypto_sha256_arm_finup,
+	.descsize	=	sizeof(struct sha256_state),
+	.base		=	{
+		.cra_name	=	"sha256",
+		.cra_driver_name =	"sha256-asm",
+		.cra_priority	=	150,
+		.cra_flags	=	CRYPTO_ALG_TYPE_SHASH,
+		.cra_blocksize	=	SHA256_BLOCK_SIZE,
+		.cra_module	=	THIS_MODULE,
+	}
+}, {
+	.digestsize	=	SHA224_DIGEST_SIZE,
+	.init		=	sha224_base_init,
+	.update		=	crypto_sha256_arm_update,
+	.final		=	sha256_final,
+	.finup		=	crypto_sha256_arm_finup,
+	.descsize	=	sizeof(struct sha256_state),
+	.base		=	{
+		.cra_name	=	"sha224",
+		.cra_driver_name =	"sha224-asm",
+		.cra_priority	=	150,
+		.cra_flags	=	CRYPTO_ALG_TYPE_SHASH,
+		.cra_blocksize	=	SHA224_BLOCK_SIZE,
+		.cra_module	=	THIS_MODULE,
+	}
+} };
+
+static int __init sha256_mod_init(void)
+{
+	int res = crypto_register_shashes(algs, ARRAY_SIZE(algs));
+
+	if (res < 0)
+		return res;
+
+	if (IS_ENABLED(CONFIG_KERNEL_MODE_NEON) && cpu_has_neon()) {
+		res = crypto_register_shashes(sha256_neon_algs,
+					      ARRAY_SIZE(sha256_neon_algs));
+
+		if (res < 0)
+			crypto_unregister_shashes(algs, ARRAY_SIZE(algs));
+	}
+
+	return res;
+}
+
+static void __exit sha256_mod_fini(void)
+{
+	crypto_unregister_shashes(algs, ARRAY_SIZE(algs));
+
+	if (IS_ENABLED(CONFIG_KERNEL_MODE_NEON) && cpu_has_neon())
+		crypto_unregister_shashes(sha256_neon_algs,
+					  ARRAY_SIZE(sha256_neon_algs));
+}
+
+module_init(sha256_mod_init);
+module_exit(sha256_mod_fini);
+
+MODULE_LICENSE("GPL");
+MODULE_DESCRIPTION("SHA256 Secure Hash Algorithm (ARM), including NEON");
+
+MODULE_ALIAS_CRYPTO("sha256");
diff --git a/arch/arm/crypto/sha256_glue.h b/arch/arm/crypto/sha256_glue.h
new file mode 100644
index 000000000..7cf0bf786
--- /dev/null
+++ b/arch/arm/crypto/sha256_glue.h
@@ -0,0 +1,14 @@
+#ifndef _CRYPTO_SHA256_GLUE_H
+#define _CRYPTO_SHA256_GLUE_H
+
+#include <linux/crypto.h>
+
+extern struct shash_alg sha256_neon_algs[2];
+
+int crypto_sha256_arm_update(struct shash_desc *desc, const u8 *data,
+			     unsigned int len);
+
+int crypto_sha256_arm_finup(struct shash_desc *desc, const u8 *data,
+			    unsigned int len, u8 *hash);
+
+#endif /* _CRYPTO_SHA256_GLUE_H */
diff --git a/arch/arm/crypto/sha256_neon_glue.c b/arch/arm/crypto/sha256_neon_glue.c
new file mode 100644
index 000000000..39ccd6588
--- /dev/null
+++ b/arch/arm/crypto/sha256_neon_glue.c
@@ -0,0 +1,101 @@
+/*
+ * Glue code for the SHA256 Secure Hash Algorithm assembly implementation
+ * using NEON instructions.
+ *
+ * Copyright © 2015 Google Inc.
+ *
+ * This file is based on sha512_neon_glue.c:
+ *   Copyright © 2014 Jussi Kivilinna <jussi.kivilinna@iki.fi>
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License as published by the Free
+ * Software Foundation; either version 2 of the License, or (at your option)
+ * any later version.
+ *
+ */
+
+#include <crypto/internal/hash.h>
+#include <linux/cryptohash.h>
+#include <linux/types.h>
+#include <linux/string.h>
+#include <crypto/sha.h>
+#include <crypto/sha256_base.h>
+#include <asm/byteorder.h>
+#include <asm/simd.h>
+#include <asm/neon.h>
+
+#include "sha256_glue.h"
+
+asmlinkage void sha256_block_data_order_neon(u32 *digest, const void *data,
+					     unsigned int num_blks);
+
+static int sha256_update(struct shash_desc *desc, const u8 *data,
+			 unsigned int len)
+{
+	struct sha256_state *sctx = shash_desc_ctx(desc);
+
+	if (!may_use_simd() ||
+	    (sctx->count % SHA256_BLOCK_SIZE) + len < SHA256_BLOCK_SIZE)
+		return crypto_sha256_arm_update(desc, data, len);
+
+	kernel_neon_begin();
+	sha256_base_do_update(desc, data, len,
+			(sha256_block_fn *)sha256_block_data_order_neon);
+	kernel_neon_end();
+
+	return 0;
+}
+
+static int sha256_finup(struct shash_desc *desc, const u8 *data,
+			unsigned int len, u8 *out)
+{
+	if (!may_use_simd())
+		return crypto_sha256_arm_finup(desc, data, len, out);
+
+	kernel_neon_begin();
+	if (len)
+		sha256_base_do_update(desc, data, len,
+			(sha256_block_fn *)sha256_block_data_order_neon);
+	sha256_base_do_finalize(desc,
+			(sha256_block_fn *)sha256_block_data_order_neon);
+	kernel_neon_end();
+
+	return sha256_base_finish(desc, out);
+}
+
+static int sha256_final(struct shash_desc *desc, u8 *out)
+{
+	return sha256_finup(desc, NULL, 0, out);
+}
+
+struct shash_alg sha256_neon_algs[] = { {
+	.digestsize	=	SHA256_DIGEST_SIZE,
+	.init		=	sha256_base_init,
+	.update		=	sha256_update,
+	.final		=	sha256_final,
+	.finup		=	sha256_finup,
+	.descsize	=	sizeof(struct sha256_state),
+	.base		=	{
+		.cra_name	=	"sha256",
+		.cra_driver_name =	"sha256-neon",
+		.cra_priority	=	250,
+		.cra_flags	=	CRYPTO_ALG_TYPE_SHASH,
+		.cra_blocksize	=	SHA256_BLOCK_SIZE,
+		.cra_module	=	THIS_MODULE,
+	}
+}, {
+	.digestsize	=	SHA224_DIGEST_SIZE,
+	.init		=	sha224_base_init,
+	.update		=	sha256_update,
+	.final		=	sha256_final,
+	.finup		=	sha256_finup,
+	.descsize	=	sizeof(struct sha256_state),
+	.base		=	{
+		.cra_name	=	"sha224",
+		.cra_driver_name =	"sha224-neon",
+		.cra_priority	=	250,
+		.cra_flags	=	CRYPTO_ALG_TYPE_SHASH,
+		.cra_blocksize	=	SHA224_BLOCK_SIZE,
+		.cra_module	=	THIS_MODULE,
+	}
+} };
diff --git a/arch/arm/crypto/sha512-armv7-neon.S b/arch/arm/crypto/sha512-armv7-neon.S
new file mode 100644
index 000000000..fe99472e5
--- /dev/null
+++ b/arch/arm/crypto/sha512-armv7-neon.S
@@ -0,0 +1,455 @@
+/* sha512-armv7-neon.S  -  ARM/NEON assembly implementation of SHA-512 transform
+ *
+ * Copyright © 2013-2014 Jussi Kivilinna <jussi.kivilinna@iki.fi>
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License as published by the Free
+ * Software Foundation; either version 2 of the License, or (at your option)
+ * any later version.
+ */
+
+#include <linux/linkage.h>
+
+
+.syntax unified
+.code   32
+.fpu neon
+
+.text
+
+/* structure of SHA512_CONTEXT */
+#define hd_a 0
+#define hd_b ((hd_a) + 8)
+#define hd_c ((hd_b) + 8)
+#define hd_d ((hd_c) + 8)
+#define hd_e ((hd_d) + 8)
+#define hd_f ((hd_e) + 8)
+#define hd_g ((hd_f) + 8)
+
+/* register macros */
+#define RK %r2
+
+#define RA d0
+#define RB d1
+#define RC d2
+#define RD d3
+#define RE d4
+#define RF d5
+#define RG d6
+#define RH d7
+
+#define RT0 d8
+#define RT1 d9
+#define RT2 d10
+#define RT3 d11
+#define RT4 d12
+#define RT5 d13
+#define RT6 d14
+#define RT7 d15
+
+#define RT01q q4
+#define RT23q q5
+#define RT45q q6
+#define RT67q q7
+
+#define RW0 d16
+#define RW1 d17
+#define RW2 d18
+#define RW3 d19
+#define RW4 d20
+#define RW5 d21
+#define RW6 d22
+#define RW7 d23
+#define RW8 d24
+#define RW9 d25
+#define RW10 d26
+#define RW11 d27
+#define RW12 d28
+#define RW13 d29
+#define RW14 d30
+#define RW15 d31
+
+#define RW01q q8
+#define RW23q q9
+#define RW45q q10
+#define RW67q q11
+#define RW89q q12
+#define RW1011q q13
+#define RW1213q q14
+#define RW1415q q15
+
+/***********************************************************************
+ * ARM assembly implementation of sha512 transform
+ ***********************************************************************/
+#define rounds2_0_63(ra, rb, rc, rd, re, rf, rg, rh, rw0, rw1, rw01q, rw2, \
+                     rw23q, rw1415q, rw9, rw10, interleave_op, arg1) \
+	/* t1 = h + Sum1 (e) + Ch (e, f, g) + k[t] + w[t]; */ \
+	vshr.u64 RT2, re, #14; \
+	vshl.u64 RT3, re, #64 - 14; \
+	interleave_op(arg1); \
+	vshr.u64 RT4, re, #18; \
+	vshl.u64 RT5, re, #64 - 18; \
+	vld1.64 {RT0}, [RK]!; \
+	veor.64 RT23q, RT23q, RT45q; \
+	vshr.u64 RT4, re, #41; \
+	vshl.u64 RT5, re, #64 - 41; \
+	vadd.u64 RT0, RT0, rw0; \
+	veor.64 RT23q, RT23q, RT45q; \
+	vmov.64 RT7, re; \
+	veor.64 RT1, RT2, RT3; \
+	vbsl.64 RT7, rf, rg; \
+	\
+	vadd.u64 RT1, RT1, rh; \
+	vshr.u64 RT2, ra, #28; \
+	vshl.u64 RT3, ra, #64 - 28; \
+	vadd.u64 RT1, RT1, RT0; \
+	vshr.u64 RT4, ra, #34; \
+	vshl.u64 RT5, ra, #64 - 34; \
+	vadd.u64 RT1, RT1, RT7; \
+	\
+	/* h = Sum0 (a) + Maj (a, b, c); */ \
+	veor.64 RT23q, RT23q, RT45q; \
+	vshr.u64 RT4, ra, #39; \
+	vshl.u64 RT5, ra, #64 - 39; \
+	veor.64 RT0, ra, rb; \
+	veor.64 RT23q, RT23q, RT45q; \
+	vbsl.64 RT0, rc, rb; \
+	vadd.u64 rd, rd, RT1; /* d+=t1; */ \
+	veor.64 rh, RT2, RT3; \
+	\
+	/* t1 = g + Sum1 (d) + Ch (d, e, f) + k[t] + w[t]; */ \
+	vshr.u64 RT2, rd, #14; \
+	vshl.u64 RT3, rd, #64 - 14; \
+	vadd.u64 rh, rh, RT0; \
+	vshr.u64 RT4, rd, #18; \
+	vshl.u64 RT5, rd, #64 - 18; \
+	vadd.u64 rh, rh, RT1; /* h+=t1; */ \
+	vld1.64 {RT0}, [RK]!; \
+	veor.64 RT23q, RT23q, RT45q; \
+	vshr.u64 RT4, rd, #41; \
+	vshl.u64 RT5, rd, #64 - 41; \
+	vadd.u64 RT0, RT0, rw1; \
+	veor.64 RT23q, RT23q, RT45q; \
+	vmov.64 RT7, rd; \
+	veor.64 RT1, RT2, RT3; \
+	vbsl.64 RT7, re, rf; \
+	\
+	vadd.u64 RT1, RT1, rg; \
+	vshr.u64 RT2, rh, #28; \
+	vshl.u64 RT3, rh, #64 - 28; \
+	vadd.u64 RT1, RT1, RT0; \
+	vshr.u64 RT4, rh, #34; \
+	vshl.u64 RT5, rh, #64 - 34; \
+	vadd.u64 RT1, RT1, RT7; \
+	\
+	/* g = Sum0 (h) + Maj (h, a, b); */ \
+	veor.64 RT23q, RT23q, RT45q; \
+	vshr.u64 RT4, rh, #39; \
+	vshl.u64 RT5, rh, #64 - 39; \
+	veor.64 RT0, rh, ra; \
+	veor.64 RT23q, RT23q, RT45q; \
+	vbsl.64 RT0, rb, ra; \
+	vadd.u64 rc, rc, RT1; /* c+=t1; */ \
+	veor.64 rg, RT2, RT3; \
+	\
+	/* w[0] += S1 (w[14]) + w[9] + S0 (w[1]); */ \
+	/* w[1] += S1 (w[15]) + w[10] + S0 (w[2]); */ \
+	\
+	/**** S0(w[1:2]) */ \
+	\
+	/* w[0:1] += w[9:10] */ \
+	/* RT23q = rw1:rw2 */ \
+	vext.u64 RT23q, rw01q, rw23q, #1; \
+	vadd.u64 rw0, rw9; \
+	vadd.u64 rg, rg, RT0; \
+	vadd.u64 rw1, rw10;\
+	vadd.u64 rg, rg, RT1; /* g+=t1; */ \
+	\
+	vshr.u64 RT45q, RT23q, #1; \
+	vshl.u64 RT67q, RT23q, #64 - 1; \
+	vshr.u64 RT01q, RT23q, #8; \
+	veor.u64 RT45q, RT45q, RT67q; \
+	vshl.u64 RT67q, RT23q, #64 - 8; \
+	veor.u64 RT45q, RT45q, RT01q; \
+	vshr.u64 RT01q, RT23q, #7; \
+	veor.u64 RT45q, RT45q, RT67q; \
+	\
+	/**** S1(w[14:15]) */ \
+	vshr.u64 RT23q, rw1415q, #6; \
+	veor.u64 RT01q, RT01q, RT45q; \
+	vshr.u64 RT45q, rw1415q, #19; \
+	vshl.u64 RT67q, rw1415q, #64 - 19; \
+	veor.u64 RT23q, RT23q, RT45q; \
+	vshr.u64 RT45q, rw1415q, #61; \
+	veor.u64 RT23q, RT23q, RT67q; \
+	vshl.u64 RT67q, rw1415q, #64 - 61; \
+	veor.u64 RT23q, RT23q, RT45q; \
+	vadd.u64 rw01q, RT01q; /* w[0:1] += S(w[1:2]) */ \
+	veor.u64 RT01q, RT23q, RT67q;
+#define vadd_RT01q(rw01q) \
+	/* w[0:1] += S(w[14:15]) */ \
+	vadd.u64 rw01q, RT01q;
+
+#define dummy(_) /*_*/
+
+#define rounds2_64_79(ra, rb, rc, rd, re, rf, rg, rh, rw0, rw1, \
+	              interleave_op1, arg1, interleave_op2, arg2) \
+	/* t1 = h + Sum1 (e) + Ch (e, f, g) + k[t] + w[t]; */ \
+	vshr.u64 RT2, re, #14; \
+	vshl.u64 RT3, re, #64 - 14; \
+	interleave_op1(arg1); \
+	vshr.u64 RT4, re, #18; \
+	vshl.u64 RT5, re, #64 - 18; \
+	interleave_op2(arg2); \
+	vld1.64 {RT0}, [RK]!; \
+	veor.64 RT23q, RT23q, RT45q; \
+	vshr.u64 RT4, re, #41; \
+	vshl.u64 RT5, re, #64 - 41; \
+	vadd.u64 RT0, RT0, rw0; \
+	veor.64 RT23q, RT23q, RT45q; \
+	vmov.64 RT7, re; \
+	veor.64 RT1, RT2, RT3; \
+	vbsl.64 RT7, rf, rg; \
+	\
+	vadd.u64 RT1, RT1, rh; \
+	vshr.u64 RT2, ra, #28; \
+	vshl.u64 RT3, ra, #64 - 28; \
+	vadd.u64 RT1, RT1, RT0; \
+	vshr.u64 RT4, ra, #34; \
+	vshl.u64 RT5, ra, #64 - 34; \
+	vadd.u64 RT1, RT1, RT7; \
+	\
+	/* h = Sum0 (a) + Maj (a, b, c); */ \
+	veor.64 RT23q, RT23q, RT45q; \
+	vshr.u64 RT4, ra, #39; \
+	vshl.u64 RT5, ra, #64 - 39; \
+	veor.64 RT0, ra, rb; \
+	veor.64 RT23q, RT23q, RT45q; \
+	vbsl.64 RT0, rc, rb; \
+	vadd.u64 rd, rd, RT1; /* d+=t1; */ \
+	veor.64 rh, RT2, RT3; \
+	\
+	/* t1 = g + Sum1 (d) + Ch (d, e, f) + k[t] + w[t]; */ \
+	vshr.u64 RT2, rd, #14; \
+	vshl.u64 RT3, rd, #64 - 14; \
+	vadd.u64 rh, rh, RT0; \
+	vshr.u64 RT4, rd, #18; \
+	vshl.u64 RT5, rd, #64 - 18; \
+	vadd.u64 rh, rh, RT1; /* h+=t1; */ \
+	vld1.64 {RT0}, [RK]!; \
+	veor.64 RT23q, RT23q, RT45q; \
+	vshr.u64 RT4, rd, #41; \
+	vshl.u64 RT5, rd, #64 - 41; \
+	vadd.u64 RT0, RT0, rw1; \
+	veor.64 RT23q, RT23q, RT45q; \
+	vmov.64 RT7, rd; \
+	veor.64 RT1, RT2, RT3; \
+	vbsl.64 RT7, re, rf; \
+	\
+	vadd.u64 RT1, RT1, rg; \
+	vshr.u64 RT2, rh, #28; \
+	vshl.u64 RT3, rh, #64 - 28; \
+	vadd.u64 RT1, RT1, RT0; \
+	vshr.u64 RT4, rh, #34; \
+	vshl.u64 RT5, rh, #64 - 34; \
+	vadd.u64 RT1, RT1, RT7; \
+	\
+	/* g = Sum0 (h) + Maj (h, a, b); */ \
+	veor.64 RT23q, RT23q, RT45q; \
+	vshr.u64 RT4, rh, #39; \
+	vshl.u64 RT5, rh, #64 - 39; \
+	veor.64 RT0, rh, ra; \
+	veor.64 RT23q, RT23q, RT45q; \
+	vbsl.64 RT0, rb, ra; \
+	vadd.u64 rc, rc, RT1; /* c+=t1; */ \
+	veor.64 rg, RT2, RT3;
+#define vadd_rg_RT0(rg) \
+	vadd.u64 rg, rg, RT0;
+#define vadd_rg_RT1(rg) \
+	vadd.u64 rg, rg, RT1; /* g+=t1; */
+
+.align 3
+ENTRY(sha512_transform_neon)
+	/* Input:
+	 *	%r0: SHA512_CONTEXT
+	 *	%r1: data
+	 *	%r2: u64 k[] constants
+	 *	%r3: nblks
+	 */
+	push {%lr};
+
+	mov %lr, #0;
+
+	/* Load context to d0-d7 */
+	vld1.64 {RA-RD}, [%r0]!;
+	vld1.64 {RE-RH}, [%r0];
+	sub %r0, #(4*8);
+
+	/* Load input to w[16], d16-d31 */
+	/* NOTE: Assumes that on ARMv7 unaligned accesses are always allowed. */
+	vld1.64 {RW0-RW3}, [%r1]!;
+	vld1.64 {RW4-RW7}, [%r1]!;
+	vld1.64 {RW8-RW11}, [%r1]!;
+	vld1.64 {RW12-RW15}, [%r1]!;
+#ifdef __ARMEL__
+	/* byteswap */
+	vrev64.8 RW01q, RW01q;
+	vrev64.8 RW23q, RW23q;
+	vrev64.8 RW45q, RW45q;
+	vrev64.8 RW67q, RW67q;
+	vrev64.8 RW89q, RW89q;
+	vrev64.8 RW1011q, RW1011q;
+	vrev64.8 RW1213q, RW1213q;
+	vrev64.8 RW1415q, RW1415q;
+#endif
+
+	/* EABI says that d8-d15 must be preserved by callee. */
+	/*vpush {RT0-RT7};*/
+
+.Loop:
+	rounds2_0_63(RA, RB, RC, RD, RE, RF, RG, RH, RW0, RW1, RW01q, RW2,
+		     RW23q, RW1415q, RW9, RW10, dummy, _);
+	b .Lenter_rounds;
+
+.Loop_rounds:
+	rounds2_0_63(RA, RB, RC, RD, RE, RF, RG, RH, RW0, RW1, RW01q, RW2,
+		     RW23q, RW1415q, RW9, RW10, vadd_RT01q, RW1415q);
+.Lenter_rounds:
+	rounds2_0_63(RG, RH, RA, RB, RC, RD, RE, RF, RW2, RW3, RW23q, RW4,
+		     RW45q, RW01q, RW11, RW12, vadd_RT01q, RW01q);
+	rounds2_0_63(RE, RF, RG, RH, RA, RB, RC, RD, RW4, RW5, RW45q, RW6,
+		     RW67q, RW23q, RW13, RW14, vadd_RT01q, RW23q);
+	rounds2_0_63(RC, RD, RE, RF, RG, RH, RA, RB, RW6, RW7, RW67q, RW8,
+		     RW89q, RW45q, RW15, RW0, vadd_RT01q, RW45q);
+	rounds2_0_63(RA, RB, RC, RD, RE, RF, RG, RH, RW8, RW9, RW89q, RW10,
+		     RW1011q, RW67q, RW1, RW2, vadd_RT01q, RW67q);
+	rounds2_0_63(RG, RH, RA, RB, RC, RD, RE, RF, RW10, RW11, RW1011q, RW12,
+		     RW1213q, RW89q, RW3, RW4, vadd_RT01q, RW89q);
+	add %lr, #16;
+	rounds2_0_63(RE, RF, RG, RH, RA, RB, RC, RD, RW12, RW13, RW1213q, RW14,
+		     RW1415q, RW1011q, RW5, RW6, vadd_RT01q, RW1011q);
+	cmp %lr, #64;
+	rounds2_0_63(RC, RD, RE, RF, RG, RH, RA, RB, RW14, RW15, RW1415q, RW0,
+		     RW01q, RW1213q, RW7, RW8, vadd_RT01q, RW1213q);
+	bne .Loop_rounds;
+
+	subs %r3, #1;
+
+	rounds2_64_79(RA, RB, RC, RD, RE, RF, RG, RH, RW0, RW1,
+		      vadd_RT01q, RW1415q, dummy, _);
+	rounds2_64_79(RG, RH, RA, RB, RC, RD, RE, RF, RW2, RW3,
+		      vadd_rg_RT0, RG, vadd_rg_RT1, RG);
+	beq .Lhandle_tail;
+	vld1.64 {RW0-RW3}, [%r1]!;
+	rounds2_64_79(RE, RF, RG, RH, RA, RB, RC, RD, RW4, RW5,
+		      vadd_rg_RT0, RE, vadd_rg_RT1, RE);
+	rounds2_64_79(RC, RD, RE, RF, RG, RH, RA, RB, RW6, RW7,
+		      vadd_rg_RT0, RC, vadd_rg_RT1, RC);
+#ifdef __ARMEL__
+	vrev64.8 RW01q, RW01q;
+	vrev64.8 RW23q, RW23q;
+#endif
+	vld1.64 {RW4-RW7}, [%r1]!;
+	rounds2_64_79(RA, RB, RC, RD, RE, RF, RG, RH, RW8, RW9,
+		      vadd_rg_RT0, RA, vadd_rg_RT1, RA);
+	rounds2_64_79(RG, RH, RA, RB, RC, RD, RE, RF, RW10, RW11,
+		      vadd_rg_RT0, RG, vadd_rg_RT1, RG);
+#ifdef __ARMEL__
+	vrev64.8 RW45q, RW45q;
+	vrev64.8 RW67q, RW67q;
+#endif
+	vld1.64 {RW8-RW11}, [%r1]!;
+	rounds2_64_79(RE, RF, RG, RH, RA, RB, RC, RD, RW12, RW13,
+		      vadd_rg_RT0, RE, vadd_rg_RT1, RE);
+	rounds2_64_79(RC, RD, RE, RF, RG, RH, RA, RB, RW14, RW15,
+		      vadd_rg_RT0, RC, vadd_rg_RT1, RC);
+#ifdef __ARMEL__
+	vrev64.8 RW89q, RW89q;
+	vrev64.8 RW1011q, RW1011q;
+#endif
+	vld1.64 {RW12-RW15}, [%r1]!;
+	vadd_rg_RT0(RA);
+	vadd_rg_RT1(RA);
+
+	/* Load context */
+	vld1.64 {RT0-RT3}, [%r0]!;
+	vld1.64 {RT4-RT7}, [%r0];
+	sub %r0, #(4*8);
+
+#ifdef __ARMEL__
+	vrev64.8 RW1213q, RW1213q;
+	vrev64.8 RW1415q, RW1415q;
+#endif
+
+	vadd.u64 RA, RT0;
+	vadd.u64 RB, RT1;
+	vadd.u64 RC, RT2;
+	vadd.u64 RD, RT3;
+	vadd.u64 RE, RT4;
+	vadd.u64 RF, RT5;
+	vadd.u64 RG, RT6;
+	vadd.u64 RH, RT7;
+
+	/* Store the first half of context */
+	vst1.64 {RA-RD}, [%r0]!;
+	sub RK, $(8*80);
+	vst1.64 {RE-RH}, [%r0]; /* Store the last half of context */
+	mov %lr, #0;
+	sub %r0, #(4*8);
+
+	b .Loop;
+
+.Lhandle_tail:
+	rounds2_64_79(RE, RF, RG, RH, RA, RB, RC, RD, RW4, RW5,
+		      vadd_rg_RT0, RE, vadd_rg_RT1, RE);
+	rounds2_64_79(RC, RD, RE, RF, RG, RH, RA, RB, RW6, RW7,
+		      vadd_rg_RT0, RC, vadd_rg_RT1, RC);
+	rounds2_64_79(RA, RB, RC, RD, RE, RF, RG, RH, RW8, RW9,
+		      vadd_rg_RT0, RA, vadd_rg_RT1, RA);
+	rounds2_64_79(RG, RH, RA, RB, RC, RD, RE, RF, RW10, RW11,
+		      vadd_rg_RT0, RG, vadd_rg_RT1, RG);
+	rounds2_64_79(RE, RF, RG, RH, RA, RB, RC, RD, RW12, RW13,
+		      vadd_rg_RT0, RE, vadd_rg_RT1, RE);
+	rounds2_64_79(RC, RD, RE, RF, RG, RH, RA, RB, RW14, RW15,
+		      vadd_rg_RT0, RC, vadd_rg_RT1, RC);
+
+	/* Load context to d16-d23 */
+	vld1.64 {RW0-RW3}, [%r0]!;
+	vadd_rg_RT0(RA);
+	vld1.64 {RW4-RW7}, [%r0];
+	vadd_rg_RT1(RA);
+	sub %r0, #(4*8);
+
+	vadd.u64 RA, RW0;
+	vadd.u64 RB, RW1;
+	vadd.u64 RC, RW2;
+	vadd.u64 RD, RW3;
+	vadd.u64 RE, RW4;
+	vadd.u64 RF, RW5;
+	vadd.u64 RG, RW6;
+	vadd.u64 RH, RW7;
+
+	/* Store the first half of context */
+	vst1.64 {RA-RD}, [%r0]!;
+
+	/* Clear used registers */
+	/* d16-d31 */
+	veor.u64 RW01q, RW01q;
+	veor.u64 RW23q, RW23q;
+	veor.u64 RW45q, RW45q;
+	veor.u64 RW67q, RW67q;
+	vst1.64 {RE-RH}, [%r0]; /* Store the last half of context */
+	veor.u64 RW89q, RW89q;
+	veor.u64 RW1011q, RW1011q;
+	veor.u64 RW1213q, RW1213q;
+	veor.u64 RW1415q, RW1415q;
+	/* d8-d15 */
+	/*vpop {RT0-RT7};*/
+	/* d0-d7 (q0-q3) */
+	veor.u64 %q0, %q0;
+	veor.u64 %q1, %q1;
+	veor.u64 %q2, %q2;
+	veor.u64 %q3, %q3;
+
+	pop {%pc};
+ENDPROC(sha512_transform_neon)
diff --git a/arch/arm/crypto/sha512_neon_glue.c b/arch/arm/crypto/sha512_neon_glue.c
new file mode 100644
index 000000000..b124dce83
--- /dev/null
+++ b/arch/arm/crypto/sha512_neon_glue.c
@@ -0,0 +1,305 @@
+/*
+ * Glue code for the SHA512 Secure Hash Algorithm assembly implementation
+ * using NEON instructions.
+ *
+ * Copyright © 2014 Jussi Kivilinna <jussi.kivilinna@iki.fi>
+ *
+ * This file is based on sha512_ssse3_glue.c:
+ *   Copyright (C) 2013 Intel Corporation
+ *   Author: Tim Chen <tim.c.chen@linux.intel.com>
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License as published by the Free
+ * Software Foundation; either version 2 of the License, or (at your option)
+ * any later version.
+ *
+ */
+
+#include <crypto/internal/hash.h>
+#include <linux/init.h>
+#include <linux/module.h>
+#include <linux/mm.h>
+#include <linux/cryptohash.h>
+#include <linux/types.h>
+#include <linux/string.h>
+#include <crypto/sha.h>
+#include <asm/byteorder.h>
+#include <asm/simd.h>
+#include <asm/neon.h>
+
+
+static const u64 sha512_k[] = {
+	0x428a2f98d728ae22ULL, 0x7137449123ef65cdULL,
+	0xb5c0fbcfec4d3b2fULL, 0xe9b5dba58189dbbcULL,
+	0x3956c25bf348b538ULL, 0x59f111f1b605d019ULL,
+	0x923f82a4af194f9bULL, 0xab1c5ed5da6d8118ULL,
+	0xd807aa98a3030242ULL, 0x12835b0145706fbeULL,
+	0x243185be4ee4b28cULL, 0x550c7dc3d5ffb4e2ULL,
+	0x72be5d74f27b896fULL, 0x80deb1fe3b1696b1ULL,
+	0x9bdc06a725c71235ULL, 0xc19bf174cf692694ULL,
+	0xe49b69c19ef14ad2ULL, 0xefbe4786384f25e3ULL,
+	0x0fc19dc68b8cd5b5ULL, 0x240ca1cc77ac9c65ULL,
+	0x2de92c6f592b0275ULL, 0x4a7484aa6ea6e483ULL,
+	0x5cb0a9dcbd41fbd4ULL, 0x76f988da831153b5ULL,
+	0x983e5152ee66dfabULL, 0xa831c66d2db43210ULL,
+	0xb00327c898fb213fULL, 0xbf597fc7beef0ee4ULL,
+	0xc6e00bf33da88fc2ULL, 0xd5a79147930aa725ULL,
+	0x06ca6351e003826fULL, 0x142929670a0e6e70ULL,
+	0x27b70a8546d22ffcULL, 0x2e1b21385c26c926ULL,
+	0x4d2c6dfc5ac42aedULL, 0x53380d139d95b3dfULL,
+	0x650a73548baf63deULL, 0x766a0abb3c77b2a8ULL,
+	0x81c2c92e47edaee6ULL, 0x92722c851482353bULL,
+	0xa2bfe8a14cf10364ULL, 0xa81a664bbc423001ULL,
+	0xc24b8b70d0f89791ULL, 0xc76c51a30654be30ULL,
+	0xd192e819d6ef5218ULL, 0xd69906245565a910ULL,
+	0xf40e35855771202aULL, 0x106aa07032bbd1b8ULL,
+	0x19a4c116b8d2d0c8ULL, 0x1e376c085141ab53ULL,
+	0x2748774cdf8eeb99ULL, 0x34b0bcb5e19b48a8ULL,
+	0x391c0cb3c5c95a63ULL, 0x4ed8aa4ae3418acbULL,
+	0x5b9cca4f7763e373ULL, 0x682e6ff3d6b2b8a3ULL,
+	0x748f82ee5defb2fcULL, 0x78a5636f43172f60ULL,
+	0x84c87814a1f0ab72ULL, 0x8cc702081a6439ecULL,
+	0x90befffa23631e28ULL, 0xa4506cebde82bde9ULL,
+	0xbef9a3f7b2c67915ULL, 0xc67178f2e372532bULL,
+	0xca273eceea26619cULL, 0xd186b8c721c0c207ULL,
+	0xeada7dd6cde0eb1eULL, 0xf57d4f7fee6ed178ULL,
+	0x06f067aa72176fbaULL, 0x0a637dc5a2c898a6ULL,
+	0x113f9804bef90daeULL, 0x1b710b35131c471bULL,
+	0x28db77f523047d84ULL, 0x32caab7b40c72493ULL,
+	0x3c9ebe0a15c9bebcULL, 0x431d67c49c100d4cULL,
+	0x4cc5d4becb3e42b6ULL, 0x597f299cfc657e2aULL,
+	0x5fcb6fab3ad6faecULL, 0x6c44198c4a475817ULL
+};
+
+
+asmlinkage void sha512_transform_neon(u64 *digest, const void *data,
+				      const u64 k[], unsigned int num_blks);
+
+
+static int sha512_neon_init(struct shash_desc *desc)
+{
+	struct sha512_state *sctx = shash_desc_ctx(desc);
+
+	sctx->state[0] = SHA512_H0;
+	sctx->state[1] = SHA512_H1;
+	sctx->state[2] = SHA512_H2;
+	sctx->state[3] = SHA512_H3;
+	sctx->state[4] = SHA512_H4;
+	sctx->state[5] = SHA512_H5;
+	sctx->state[6] = SHA512_H6;
+	sctx->state[7] = SHA512_H7;
+	sctx->count[0] = sctx->count[1] = 0;
+
+	return 0;
+}
+
+static int __sha512_neon_update(struct shash_desc *desc, const u8 *data,
+				unsigned int len, unsigned int partial)
+{
+	struct sha512_state *sctx = shash_desc_ctx(desc);
+	unsigned int done = 0;
+
+	sctx->count[0] += len;
+	if (sctx->count[0] < len)
+		sctx->count[1]++;
+
+	if (partial) {
+		done = SHA512_BLOCK_SIZE - partial;
+		memcpy(sctx->buf + partial, data, done);
+		sha512_transform_neon(sctx->state, sctx->buf, sha512_k, 1);
+	}
+
+	if (len - done >= SHA512_BLOCK_SIZE) {
+		const unsigned int rounds = (len - done) / SHA512_BLOCK_SIZE;
+
+		sha512_transform_neon(sctx->state, data + done, sha512_k,
+				      rounds);
+
+		done += rounds * SHA512_BLOCK_SIZE;
+	}
+
+	memcpy(sctx->buf, data + done, len - done);
+
+	return 0;
+}
+
+static int sha512_neon_update(struct shash_desc *desc, const u8 *data,
+			     unsigned int len)
+{
+	struct sha512_state *sctx = shash_desc_ctx(desc);
+	unsigned int partial = sctx->count[0] % SHA512_BLOCK_SIZE;
+	int res;
+
+	/* Handle the fast case right here */
+	if (partial + len < SHA512_BLOCK_SIZE) {
+		sctx->count[0] += len;
+		if (sctx->count[0] < len)
+			sctx->count[1]++;
+		memcpy(sctx->buf + partial, data, len);
+
+		return 0;
+	}
+
+	if (!may_use_simd()) {
+		res = crypto_sha512_update(desc, data, len);
+	} else {
+		kernel_neon_begin();
+		res = __sha512_neon_update(desc, data, len, partial);
+		kernel_neon_end();
+	}
+
+	return res;
+}
+
+
+/* Add padding and return the message digest. */
+static int sha512_neon_final(struct shash_desc *desc, u8 *out)
+{
+	struct sha512_state *sctx = shash_desc_ctx(desc);
+	unsigned int i, index, padlen;
+	__be64 *dst = (__be64 *)out;
+	__be64 bits[2];
+	static const u8 padding[SHA512_BLOCK_SIZE] = { 0x80, };
+
+	/* save number of bits */
+	bits[1] = cpu_to_be64(sctx->count[0] << 3);
+	bits[0] = cpu_to_be64(sctx->count[1] << 3 | sctx->count[0] >> 61);
+
+	/* Pad out to 112 mod 128 and append length */
+	index = sctx->count[0] & 0x7f;
+	padlen = (index < 112) ? (112 - index) : ((128+112) - index);
+
+	if (!may_use_simd()) {
+		crypto_sha512_update(desc, padding, padlen);
+		crypto_sha512_update(desc, (const u8 *)&bits, sizeof(bits));
+	} else {
+		kernel_neon_begin();
+		/* We need to fill a whole block for __sha512_neon_update() */
+		if (padlen <= 112) {
+			sctx->count[0] += padlen;
+			if (sctx->count[0] < padlen)
+				sctx->count[1]++;
+			memcpy(sctx->buf + index, padding, padlen);
+		} else {
+			__sha512_neon_update(desc, padding, padlen, index);
+		}
+		__sha512_neon_update(desc, (const u8 *)&bits,
+					sizeof(bits), 112);
+		kernel_neon_end();
+	}
+
+	/* Store state in digest */
+	for (i = 0; i < 8; i++)
+		dst[i] = cpu_to_be64(sctx->state[i]);
+
+	/* Wipe context */
+	memset(sctx, 0, sizeof(*sctx));
+
+	return 0;
+}
+
+static int sha512_neon_export(struct shash_desc *desc, void *out)
+{
+	struct sha512_state *sctx = shash_desc_ctx(desc);
+
+	memcpy(out, sctx, sizeof(*sctx));
+
+	return 0;
+}
+
+static int sha512_neon_import(struct shash_desc *desc, const void *in)
+{
+	struct sha512_state *sctx = shash_desc_ctx(desc);
+
+	memcpy(sctx, in, sizeof(*sctx));
+
+	return 0;
+}
+
+static int sha384_neon_init(struct shash_desc *desc)
+{
+	struct sha512_state *sctx = shash_desc_ctx(desc);
+
+	sctx->state[0] = SHA384_H0;
+	sctx->state[1] = SHA384_H1;
+	sctx->state[2] = SHA384_H2;
+	sctx->state[3] = SHA384_H3;
+	sctx->state[4] = SHA384_H4;
+	sctx->state[5] = SHA384_H5;
+	sctx->state[6] = SHA384_H6;
+	sctx->state[7] = SHA384_H7;
+
+	sctx->count[0] = sctx->count[1] = 0;
+
+	return 0;
+}
+
+static int sha384_neon_final(struct shash_desc *desc, u8 *hash)
+{
+	u8 D[SHA512_DIGEST_SIZE];
+
+	sha512_neon_final(desc, D);
+
+	memcpy(hash, D, SHA384_DIGEST_SIZE);
+	memzero_explicit(D, SHA512_DIGEST_SIZE);
+
+	return 0;
+}
+
+static struct shash_alg algs[] = { {
+	.digestsize	=	SHA512_DIGEST_SIZE,
+	.init		=	sha512_neon_init,
+	.update		=	sha512_neon_update,
+	.final		=	sha512_neon_final,
+	.export		=	sha512_neon_export,
+	.import		=	sha512_neon_import,
+	.descsize	=	sizeof(struct sha512_state),
+	.statesize	=	sizeof(struct sha512_state),
+	.base		=	{
+		.cra_name	=	"sha512",
+		.cra_driver_name =	"sha512-neon",
+		.cra_priority	=	250,
+		.cra_flags	=	CRYPTO_ALG_TYPE_SHASH,
+		.cra_blocksize	=	SHA512_BLOCK_SIZE,
+		.cra_module	=	THIS_MODULE,
+	}
+},  {
+	.digestsize	=	SHA384_DIGEST_SIZE,
+	.init		=	sha384_neon_init,
+	.update		=	sha512_neon_update,
+	.final		=	sha384_neon_final,
+	.export		=	sha512_neon_export,
+	.import		=	sha512_neon_import,
+	.descsize	=	sizeof(struct sha512_state),
+	.statesize	=	sizeof(struct sha512_state),
+	.base		=	{
+		.cra_name	=	"sha384",
+		.cra_driver_name =	"sha384-neon",
+		.cra_priority	=	250,
+		.cra_flags	=	CRYPTO_ALG_TYPE_SHASH,
+		.cra_blocksize	=	SHA384_BLOCK_SIZE,
+		.cra_module	=	THIS_MODULE,
+	}
+} };
+
+static int __init sha512_neon_mod_init(void)
+{
+	if (!cpu_has_neon())
+		return -ENODEV;
+
+	return crypto_register_shashes(algs, ARRAY_SIZE(algs));
+}
+
+static void __exit sha512_neon_mod_fini(void)
+{
+	crypto_unregister_shashes(algs, ARRAY_SIZE(algs));
+}
+
+module_init(sha512_neon_mod_init);
+module_exit(sha512_neon_mod_fini);
+
+MODULE_LICENSE("GPL");
+MODULE_DESCRIPTION("SHA512 Secure Hash Algorithm, NEON accelerated");
+
+MODULE_ALIAS_CRYPTO("sha512");
+MODULE_ALIAS_CRYPTO("sha384");