[PATCH 1/1] arm/crypto: Add optimized AES and SHA1 routines
From: David McCullough
Date: Mon Aug 06 2012 - 22:02:18 EST
Add assembler versions of AES and SHA1 for ARM platforms. This has provided
up to a 50% improvement in IPsec/TCP throughout for tunnels using AES128/SHA1.
Platform CPU SPeed Endian Before (bps) After (bps) Improvement
IXP425 533 MHz big 11217042 15566294 ~38%
KS8695 166 MHz little 3828549 5795373 ~51%
Signed-off-by: David McCullough <ucdevel@xxxxxxxxx>
---
diff -Npur linux-3.5.orig/arch/arm/crypto/aes-armv4.S linux-3.5/arch/arm/crypto/aes-armv4.S
--- linux-3.5.orig/arch/arm/crypto/aes-armv4.S 1970-01-01 10:00:00.000000000 +1000
+++ linux-3.5/arch/arm/crypto/aes-armv4.S 2012-08-07 11:11:31.671359005 +1000
@@ -0,0 +1,1112 @@
+#define __ARM_ARCH__ __LINUX_ARM_ARCH__
+@ ====================================================================
+@ Written by Andy Polyakov <appro@xxxxxxxxxxxxxx> 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.
+
+.text
+.code 32
+
+.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) {
+.global AES_encrypt
+.type AES_encrypt,%function
+.align 5
+AES_encrypt:
+ sub r3,pc,#8 @ 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
+#if __ARM_ARCH__>=5
+ ldmia sp!,{r4-r12,pc}
+#else
+ ldmia sp!,{r4-r12,lr}
+ tst lr,#1
+ moveq pc,lr @ be binary compatible with V4, yet
+ .word 0xe12fff1e @ interoperable with Thumb ISA:-)
+#endif
+.size AES_encrypt,.-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
+
+.global private_AES_set_encrypt_key
+.type private_AES_set_encrypt_key,%function
+.align 5
+private_AES_set_encrypt_key:
+_armv4_AES_set_encrypt_key:
+ sub r3,pc,#8 @ 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: tst lr,#1
+ moveq pc,lr @ be binary compatible with V4, yet
+ .word 0xe12fff1e @ interoperable with Thumb ISA:-)
+.size private_AES_set_encrypt_key,.-private_AES_set_encrypt_key
+
+.global private_AES_set_decrypt_key
+.type private_AES_set_decrypt_key,%function
+.align 5
+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
+#if __ARM_ARCH__>=5
+ ldmia sp!,{r4-r12,pc}
+#else
+ ldmia sp!,{r4-r12,lr}
+ tst lr,#1
+ moveq pc,lr @ be binary compatible with V4, yet
+ .word 0xe12fff1e @ interoperable with Thumb ISA:-)
+#endif
+.size private_AES_set_decrypt_key,.-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) {
+.global AES_decrypt
+.type AES_decrypt,%function
+.align 5
+AES_decrypt:
+ sub r3,pc,#8 @ 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
+#if __ARM_ARCH__>=5
+ ldmia sp!,{r4-r12,pc}
+#else
+ ldmia sp!,{r4-r12,lr}
+ tst lr,#1
+ moveq pc,lr @ be binary compatible with V4, yet
+ .word 0xe12fff1e @ interoperable with Thumb ISA:-)
+#endif
+.size AES_decrypt,.-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]
+ ldrb r1,[r10,r1,lsr#24] @ Td4[s1>>24]
+ 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
+
+ ldrb r2,[r10,r2,lsr#24] @ Td4[s2>>24]
+ 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]
+ ldrb r3,[r10,r3,lsr#24] @ Td4[s3>>24]
+ 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@xxxxxxxxxxx>"
+.align 2
diff -Npur linux-3.5.orig/arch/arm/crypto/aes_glue.c linux-3.5/arch/arm/crypto/aes_glue.c
--- linux-3.5.orig/arch/arm/crypto/aes_glue.c 1970-01-01 10:00:00.000000000 +1000
+++ linux-3.5/arch/arm/crypto/aes_glue.c 2012-08-07 10:54:27.911412788 +1000
@@ -0,0 +1,108 @@
+/*
+ * Glue Code for the asm optimized version of the AES Cipher Algorithm
+ */
+
+#include <linux/module.h>
+#include <linux/crypto.h>
+#include <crypto/aes.h>
+
+#define AES_MAXNR 14
+
+typedef struct {
+ unsigned int rd_key[4 *(AES_MAXNR + 1)];
+ int rounds;
+} AES_KEY;
+
+struct AES_CTX {
+ AES_KEY enc_key;
+ AES_KEY dec_key;
+};
+
+asmlinkage void AES_encrypt(const u8 *in, u8 *out, AES_KEY *ctx);
+asmlinkage void AES_decrypt(const u8 *in, u8 *out, AES_KEY *ctx);
+asmlinkage int private_AES_set_decrypt_key(const unsigned char *userKey, const int bits, AES_KEY *key);
+asmlinkage int private_AES_set_encrypt_key(const unsigned char *userKey, const int bits, AES_KEY *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("aes");
+MODULE_ALIAS("aes-asm");
+MODULE_AUTHOR("David McCullough <ucdevel@xxxxxxxxx>");
diff -Npur linux-3.5.orig/arch/arm/crypto/Makefile linux-3.5/arch/arm/crypto/Makefile
--- linux-3.5.orig/arch/arm/crypto/Makefile 1970-01-01 10:00:00.000000000 +1000
+++ linux-3.5/arch/arm/crypto/Makefile 2012-08-07 10:52:09.720762289 +1000
@@ -0,0 +1,10 @@
+#
+# Arch-specific CryptoAPI modules.
+#
+
+obj-$(CONFIG_CRYPTO_AES_ARM) += aes-arm.o
+obj-$(CONFIG_CRYPTO_SHA1_ARM) += sha1-arm.o
+
+aes-arm-y := aes-armv4.o aes_glue.o
+sha1-arm-y := sha1-armv4-large.o sha1_glue.o
+
diff -Npur linux-3.5.orig/arch/arm/crypto/sha1-armv4-large.S linux-3.5/arch/arm/crypto/sha1-armv4-large.S
--- linux-3.5.orig/arch/arm/crypto/sha1-armv4-large.S 1970-01-01 10:00:00.000000000 +1000
+++ linux-3.5/arch/arm/crypto/sha1-armv4-large.S 2012-08-07 11:11:31.671359005 +1000
@@ -0,0 +1,503 @@
+#define __ARM_ARCH__ __LINUX_ARM_ARCH__
+@ ====================================================================
+@ Written by Andy Polyakov <appro@xxxxxxxxxxxxxx> 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.
+
+.text
+
+.global sha1_block_data_order
+.type sha1_block_data_order,%function
+
+.align 2
+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)
+ teq r14,sp
+ bne .L_00_15 @ [((11+4)*5+2)*3]
+#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]
+ sub sp,sp,#25*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)
+ teq r14,sp @ 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
+ teq 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
+
+#if __ARM_ARCH__>=5
+ ldmia sp!,{r4-r12,pc}
+#else
+ ldmia sp!,{r4-r12,lr}
+ tst lr,#1
+ moveq pc,lr @ be binary compatible with V4, yet
+ .word 0xe12fff1e @ interoperable with Thumb ISA:-)
+#endif
+.align 2
+.LK_00_19: .word 0x5a827999
+.LK_20_39: .word 0x6ed9eba1
+.LK_40_59: .word 0x8f1bbcdc
+.LK_60_79: .word 0xca62c1d6
+.size sha1_block_data_order,.-sha1_block_data_order
+.asciz "SHA1 block transform for ARMv4, CRYPTOGAMS by <appro@xxxxxxxxxxx>"
+.align 2
diff -Npur linux-3.5.orig/arch/arm/crypto/sha1_glue.c linux-3.5/arch/arm/crypto/sha1_glue.c
--- linux-3.5.orig/arch/arm/crypto/sha1_glue.c 1970-01-01 10:00:00.000000000 +1000
+++ linux-3.5/arch/arm/crypto/sha1_glue.c 2012-08-07 10:54:49.300739952 +1000
@@ -0,0 +1,179 @@
+/*
+ * 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@xxxxxxxxxxxxxxx>
+ * Copyright (c) Jean-Francois Dive <jef@xxxxxxxxxxx>
+ * Copyright (c) Mathias Krause <minipli@xxxxxxxxxxxxxx>
+ *
+ * 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 <asm/byteorder.h>
+
+struct SHA1_CTX {
+ uint32_t h0,h1,h2,h3,h4;
+ u64 count;
+ u8 data[SHA1_BLOCK_SIZE];
+};
+
+asmlinkage void sha1_block_data_order(struct SHA1_CTX *digest,
+ const unsigned char *data, unsigned int rounds);
+
+
+static int sha1_init(struct shash_desc *desc)
+{
+ struct SHA1_CTX *sctx = shash_desc_ctx(desc);
+ memset(sctx, 0, sizeof(*sctx));
+ sctx->h0 = SHA1_H0;
+ sctx->h1 = SHA1_H1;
+ sctx->h2 = SHA1_H2;
+ sctx->h3 = SHA1_H3;
+ sctx->h4 = SHA1_H4;
+ return 0;
+}
+
+
+static int __sha1_update(struct SHA1_CTX *sctx, const u8 *data,
+ unsigned int len, unsigned int partial)
+{
+ unsigned int done = 0;
+
+ sctx->count += len;
+
+ if (partial) {
+ done = SHA1_BLOCK_SIZE - partial;
+ memcpy(sctx->data + partial, data, done);
+ sha1_block_data_order(sctx, sctx->data, 1);
+ }
+
+ if (len - done >= SHA1_BLOCK_SIZE) {
+ const unsigned int rounds = (len - done) / SHA1_BLOCK_SIZE;
+ sha1_block_data_order(sctx, data + done, rounds);
+ done += rounds * SHA1_BLOCK_SIZE;
+ }
+
+ memcpy(sctx->data, data + done, len - done);
+ return 0;
+}
+
+
+static int sha1_update(struct shash_desc *desc, const u8 *data,
+ unsigned int len)
+{
+ struct SHA1_CTX *sctx = shash_desc_ctx(desc);
+ unsigned int partial = sctx->count % SHA1_BLOCK_SIZE;
+ int res;
+
+ /* Handle the fast case right here */
+ if (partial + len < SHA1_BLOCK_SIZE) {
+ sctx->count += len;
+ memcpy(sctx->data + partial, data, len);
+ return 0;
+ }
+ res = __sha1_update(sctx, data, len, partial);
+ return res;
+}
+
+
+/* Add padding and return the message digest. */
+static int sha1_final(struct shash_desc *desc, u8 *out)
+{
+ struct SHA1_CTX *sctx = shash_desc_ctx(desc);
+ unsigned int i, index, padlen;
+ __be32 *dst = (__be32 *)out;
+ __be64 bits;
+ static const u8 padding[SHA1_BLOCK_SIZE] = { 0x80, };
+
+ bits = cpu_to_be64(sctx->count << 3);
+
+ /* Pad out to 56 mod 64 and append length */
+ index = sctx->count % SHA1_BLOCK_SIZE;
+ padlen = (index < 56) ? (56 - index) : ((SHA1_BLOCK_SIZE+56) - index);
+ /* We need to fill a whole block for __sha1_update() */
+ if (padlen <= 56) {
+ sctx->count += padlen;
+ memcpy(sctx->data + index, padding, padlen);
+ } else {
+ __sha1_update(sctx, padding, padlen, index);
+ }
+ __sha1_update(sctx, (const u8 *)&bits, sizeof(bits), 56);
+
+ /* Store state in digest */
+ for (i = 0; i < 5; i++)
+ dst[i] = cpu_to_be32(((u32 *)sctx)[i]);
+
+ /* Wipe context */
+ memset(sctx, 0, sizeof(*sctx));
+ return 0;
+}
+
+
+static int sha1_export(struct shash_desc *desc, void *out)
+{
+ struct SHA1_CTX *sctx = shash_desc_ctx(desc);
+ memcpy(out, sctx, sizeof(*sctx));
+ return 0;
+}
+
+
+static int sha1_import(struct shash_desc *desc, const void *in)
+{
+ struct SHA1_CTX *sctx = shash_desc_ctx(desc);
+ memcpy(sctx, in, sizeof(*sctx));
+ return 0;
+}
+
+
+static struct shash_alg alg = {
+ .digestsize = SHA1_DIGEST_SIZE,
+ .init = sha1_init,
+ .update = sha1_update,
+ .final = sha1_final,
+ .export = sha1_export,
+ .import = sha1_import,
+ .descsize = sizeof(struct SHA1_CTX),
+ .statesize = sizeof(struct SHA1_CTX),
+ .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("sha1");
+MODULE_AUTHOR("David McCullough <ucdevel@xxxxxxxxx>");
diff -Npur linux-3.5.orig/arch/arm/Makefile linux-3.5/arch/arm/Makefile
--- linux-3.5.orig/arch/arm/Makefile 2012-07-22 06:58:29.000000000 +1000
+++ linux-3.5/arch/arm/Makefile 2012-08-07 10:52:09.720762289 +1000
@@ -250,6 +250,7 @@ core-$(CONFIG_VFP) += arch/arm/vfp/
# If we have a machine-specific directory, then include it in the build.
core-y += arch/arm/kernel/ arch/arm/mm/ arch/arm/common/
core-y += arch/arm/net/
+core-y += arch/arm/crypto/
core-y += $(machdirs) $(platdirs)
drivers-$(CONFIG_OPROFILE) += arch/arm/oprofile/
diff -Npur linux-3.5.orig/crypto/Kconfig linux-3.5/crypto/Kconfig
--- linux-3.5.orig/crypto/Kconfig 2012-07-22 06:58:29.000000000 +1000
+++ linux-3.5/crypto/Kconfig 2012-08-07 10:52:09.720762289 +1000
@@ -423,6 +423,15 @@ config CRYPTO_SHA1_SSSE3
using Supplemental SSE3 (SSSE3) instructions or Advanced Vector
Extensions (AVX), when available.
+config CRYPTO_SHA1_ARM
+ tristate "SHA1 digest algorithm (ARM-asm)"
+ depends on ARM
+ 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_SHA256
tristate "SHA224 and SHA256 digest algorithm"
select CRYPTO_HASH
@@ -577,6 +586,30 @@ config CRYPTO_AES_NI_INTEL
ECB, CBC, LRW, PCBC, XTS. The 64 bit version has additional
acceleration for CTR.
+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_ANUBIS
tristate "Anubis cipher algorithm"
select CRYPTO_ALGAPI
--
David McCullough, david_mccullough@xxxxxxxxxx, Ph:+61 734352815
McAfee - SnapGear http://www.mcafee.com http://www.uCdot.org
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/