diff options
Diffstat (limited to 'arch/arm')
54 files changed, 12157 insertions, 99 deletions
diff --git a/arch/arm/Kconfig b/arch/arm/Kconfig index 9422f298f32f..1a86004f3f27 100644 --- a/arch/arm/Kconfig +++ b/arch/arm/Kconfig @@ -1958,6 +1958,15 @@ config XEN help Say Y if you want to run Linux in a Virtual Machine on Xen on ARM. +config ARM_FLUSH_CONSOLE_ON_RESTART + bool "Force flush the console on restart" + help + If the console is locked while the system is rebooted, the messages + in the temporary logbuffer would not have propogated to all the + console drivers. This option forces the console lock to be + released if it failed to be acquired, which will cause all the + pending messages to be flushed. + endmenu menu "Boot options" @@ -1987,6 +1996,21 @@ config DEPRECATED_PARAM_STRUCT This was deprecated in 2001 and announced to live on for 5 years. Some old boot loaders still use this way. +config BUILD_ARM_APPENDED_DTB_IMAGE + bool "Build a concatenated zImage/dtb by default" + depends on OF + help + Enabling this option will cause a concatenated zImage and DTB to + be built by default (instead of a standalone zImage.) The image + will built in arch/arm/boot/zImage-dtb.<dtb name> + +config BUILD_ARM_APPENDED_DTB_IMAGE_NAME + string "Default dtb name" + depends on BUILD_ARM_APPENDED_DTB_IMAGE + help + name of the dtb to append when building a concatenated + zImage/dtb. + # Compressed boot loader in ROM. Yes, we really want to ask about # TEXT and BSS so we preserve their values in the config files. config ZBOOT_ROM_TEXT @@ -2336,6 +2360,13 @@ config NEON Say Y to include support code for NEON, the ARMv7 Advanced SIMD Extension. +config KERNEL_MODE_NEON + bool "Support for NEON in kernel mode" + default n + depends on NEON + help + Say Y to include support for NEON in kernel mode. + endmenu menu "Userspace binary formats" diff --git a/arch/arm/Kconfig.debug b/arch/arm/Kconfig.debug index 5fdb6dbc5f89..4d1793474c9c 100644 --- a/arch/arm/Kconfig.debug +++ b/arch/arm/Kconfig.debug @@ -63,6 +63,27 @@ config DEBUG_USER 8 - SIGSEGV faults 16 - SIGBUS faults +config DEBUG_RODATA + bool "Write protect kernel text section" + default n + depends on DEBUG_KERNEL && MMU + ---help--- + Mark the kernel text section as write-protected in the pagetables, + in order to catch accidental (and incorrect) writes to such const + data. This will cause the size of the kernel, plus up to 4MB, to + be mapped as pages instead of sections, which will increase TLB + pressure. + If in doubt, say "N". + +config DEBUG_RODATA_TEST + bool "Testcase for the DEBUG_RODATA feature" + depends on DEBUG_RODATA + default n + ---help--- + This option enables a testcase for the DEBUG_RODATA + feature. + If in doubt, say "N" + # These options are only for real kernel hackers who want to get their hands dirty. config DEBUG_LL bool "Kernel low-level debugging functions (read help!)" diff --git a/arch/arm/Makefile b/arch/arm/Makefile index 70bc19e2274f..9d36200374f0 100644 --- a/arch/arm/Makefile +++ b/arch/arm/Makefile @@ -265,6 +265,8 @@ libs-y := arch/arm/lib/ $(libs-y) # Default target when executing plain make ifeq ($(CONFIG_XIP_KERNEL),y) KBUILD_IMAGE := xipImage +else ifeq ($(CONFIG_BUILD_ARM_APPENDED_DTB_IMAGE),y) +KBUILD_IMAGE := zImage-dtb.$(CONFIG_BUILD_ARM_APPENDED_DTB_IMAGE_NAME) else KBUILD_IMAGE := zImage endif diff --git a/arch/arm/boot/Makefile b/arch/arm/boot/Makefile index 84aa2caf07ed..085bb96493a3 100644 --- a/arch/arm/boot/Makefile +++ b/arch/arm/boot/Makefile @@ -14,6 +14,7 @@ ifneq ($(MACHINE),) include $(srctree)/$(MACHINE)/Makefile.boot endif +include $(srctree)/arch/arm/boot/dts/Makefile # Note: the following conditions must always be true: # ZRELADDR == virt_to_phys(PAGE_OFFSET + TEXT_OFFSET) diff --git a/arch/arm/boot/compressed/head.S b/arch/arm/boot/compressed/head.S index f6e34be012ff..a8264aa9b03a 100644 --- a/arch/arm/boot/compressed/head.S +++ b/arch/arm/boot/compressed/head.S @@ -714,6 +714,8 @@ __armv7_mmu_cache_on: bic r6, r6, #1 << 31 @ 32-bit translation system bic r6, r6, #3 << 0 @ use only ttbr0 mcrne p15, 0, r3, c2, c0, 0 @ load page table pointer + mcrne p15, 0, r0, c8, c7, 0 @ flush I,D TLBs + mcr p15, 0, r0, c7, c5, 4 @ ISB mcrne p15, 0, r1, c3, c0, 0 @ load domain access control mcrne p15, 0, r6, c2, c0, 2 @ load ttb control #endif diff --git a/arch/arm/common/Kconfig b/arch/arm/common/Kconfig index 9353184d730d..ce01364a96e3 100644 --- a/arch/arm/common/Kconfig +++ b/arch/arm/common/Kconfig @@ -17,3 +17,7 @@ config SHARP_PARAM config SHARP_SCOOP bool + +config FIQ_GLUE + bool + select FIQ diff --git a/arch/arm/common/Makefile b/arch/arm/common/Makefile index 462cd580fc2d..505c479202b6 100644 --- a/arch/arm/common/Makefile +++ b/arch/arm/common/Makefile @@ -4,6 +4,7 @@ obj-y += firmware.o +obj-$(CONFIG_FIQ_GLUE) += fiq_glue.o fiq_glue_setup.o obj-$(CONFIG_ICST) += icst.o obj-$(CONFIG_SA1111) += sa1111.o obj-$(CONFIG_PCI_HOST_VIA82C505) += via82c505.o diff --git a/arch/arm/common/fiq_glue.S b/arch/arm/common/fiq_glue.S new file mode 100644 index 000000000000..24b42cec4813 --- /dev/null +++ b/arch/arm/common/fiq_glue.S @@ -0,0 +1,118 @@ +/* + * Copyright (C) 2008 Google, Inc. + * + * This software is licensed under the terms of the GNU General Public + * License version 2, as published by the Free Software Foundation, and + * may be copied, distributed, and modified under those terms. + * + * This program is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + * GNU General Public License for more details. + * + */ + +#include <linux/linkage.h> +#include <asm/assembler.h> + + .text + + .global fiq_glue_end + + /* fiq stack: r0-r15,cpsr,spsr of interrupted mode */ + +ENTRY(fiq_glue) + /* store pc, cpsr from previous mode, reserve space for spsr */ + mrs r12, spsr + sub lr, lr, #4 + subs r10, #1 + bne nested_fiq + + str r12, [sp, #-8]! + str lr, [sp, #-4]! + + /* store r8-r14 from previous mode */ + sub sp, sp, #(7 * 4) + stmia sp, {r8-r14}^ + nop + + /* store r0-r7 from previous mode */ + stmfd sp!, {r0-r7} + + /* setup func(data,regs) arguments */ + mov r0, r9 + mov r1, sp + mov r3, r8 + + mov r7, sp + + /* Get sp and lr from non-user modes */ + and r4, r12, #MODE_MASK + cmp r4, #USR_MODE + beq fiq_from_usr_mode + + mov r7, sp + orr r4, r4, #(PSR_I_BIT | PSR_F_BIT) + msr cpsr_c, r4 + str sp, [r7, #(4 * 13)] + str lr, [r7, #(4 * 14)] + mrs r5, spsr + str r5, [r7, #(4 * 17)] + + cmp r4, #(SVC_MODE | PSR_I_BIT | PSR_F_BIT) + /* use fiq stack if we reenter this mode */ + subne sp, r7, #(4 * 3) + +fiq_from_usr_mode: + msr cpsr_c, #(SVC_MODE | PSR_I_BIT | PSR_F_BIT) + mov r2, sp + sub sp, r7, #12 + stmfd sp!, {r2, ip, lr} + /* call func(data,regs) */ + blx r3 + ldmfd sp, {r2, ip, lr} + mov sp, r2 + + /* restore/discard saved state */ + cmp r4, #USR_MODE + beq fiq_from_usr_mode_exit + + msr cpsr_c, r4 + ldr sp, [r7, #(4 * 13)] + ldr lr, [r7, #(4 * 14)] + msr spsr_cxsf, r5 + +fiq_from_usr_mode_exit: + msr cpsr_c, #(FIQ_MODE | PSR_I_BIT | PSR_F_BIT) + + ldmfd sp!, {r0-r7} + ldr lr, [sp, #(4 * 7)] + ldr r12, [sp, #(4 * 8)] + add sp, sp, #(10 * 4) +exit_fiq: + msr spsr_cxsf, r12 + add r10, #1 + cmp r11, #0 + moveqs pc, lr + bx r11 /* jump to custom fiq return function */ + +nested_fiq: + orr r12, r12, #(PSR_F_BIT) + b exit_fiq + +fiq_glue_end: + +ENTRY(fiq_glue_setup) /* func, data, sp, smc call number */ + stmfd sp!, {r4} + mrs r4, cpsr + msr cpsr_c, #(FIQ_MODE | PSR_I_BIT | PSR_F_BIT) + movs r8, r0 + mov r9, r1 + mov sp, r2 + mov r11, r3 + moveq r10, #0 + movne r10, #1 + msr cpsr_c, r4 + ldmfd sp!, {r4} + bx lr + diff --git a/arch/arm/common/fiq_glue_setup.c b/arch/arm/common/fiq_glue_setup.c new file mode 100644 index 000000000000..8cb1b611c6d5 --- /dev/null +++ b/arch/arm/common/fiq_glue_setup.c @@ -0,0 +1,147 @@ +/* + * Copyright (C) 2010 Google, Inc. + * + * This software is licensed under the terms of the GNU General Public + * License version 2, as published by the Free Software Foundation, and + * may be copied, distributed, and modified under those terms. + * + * This program is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + * GNU General Public License for more details. + */ + +#include <linux/kernel.h> +#include <linux/percpu.h> +#include <linux/slab.h> +#include <asm/fiq.h> +#include <asm/fiq_glue.h> + +extern unsigned char fiq_glue, fiq_glue_end; +extern void fiq_glue_setup(void *func, void *data, void *sp, + fiq_return_handler_t fiq_return_handler); + +static struct fiq_handler fiq_debbuger_fiq_handler = { + .name = "fiq_glue", +}; +DEFINE_PER_CPU(void *, fiq_stack); +static struct fiq_glue_handler *current_handler; +static fiq_return_handler_t fiq_return_handler; +static DEFINE_MUTEX(fiq_glue_lock); + +static void fiq_glue_setup_helper(void *info) +{ + struct fiq_glue_handler *handler = info; + fiq_glue_setup(handler->fiq, handler, + __get_cpu_var(fiq_stack) + THREAD_START_SP, + fiq_return_handler); +} + +int fiq_glue_register_handler(struct fiq_glue_handler *handler) +{ + int ret; + int cpu; + + if (!handler || !handler->fiq) + return -EINVAL; + + mutex_lock(&fiq_glue_lock); + if (fiq_stack) { + ret = -EBUSY; + goto err_busy; + } + + for_each_possible_cpu(cpu) { + void *stack; + stack = (void *)__get_free_pages(GFP_KERNEL, THREAD_SIZE_ORDER); + if (WARN_ON(!stack)) { + ret = -ENOMEM; + goto err_alloc_fiq_stack; + } + per_cpu(fiq_stack, cpu) = stack; + } + + ret = claim_fiq(&fiq_debbuger_fiq_handler); + if (WARN_ON(ret)) + goto err_claim_fiq; + + current_handler = handler; + on_each_cpu(fiq_glue_setup_helper, handler, true); + set_fiq_handler(&fiq_glue, &fiq_glue_end - &fiq_glue); + + mutex_unlock(&fiq_glue_lock); + return 0; + +err_claim_fiq: +err_alloc_fiq_stack: + for_each_possible_cpu(cpu) { + __free_pages(per_cpu(fiq_stack, cpu), THREAD_SIZE_ORDER); + per_cpu(fiq_stack, cpu) = NULL; + } +err_busy: + mutex_unlock(&fiq_glue_lock); + return ret; +} + +static void fiq_glue_update_return_handler(void (*fiq_return)(void)) +{ + fiq_return_handler = fiq_return; + if (current_handler) + on_each_cpu(fiq_glue_setup_helper, current_handler, true); +} + +int fiq_glue_set_return_handler(void (*fiq_return)(void)) +{ + int ret; + + mutex_lock(&fiq_glue_lock); + if (fiq_return_handler) { + ret = -EBUSY; + goto err_busy; + } + fiq_glue_update_return_handler(fiq_return); + ret = 0; +err_busy: + mutex_unlock(&fiq_glue_lock); + + return ret; +} +EXPORT_SYMBOL(fiq_glue_set_return_handler); + +int fiq_glue_clear_return_handler(void (*fiq_return)(void)) +{ + int ret; + + mutex_lock(&fiq_glue_lock); + if (WARN_ON(fiq_return_handler != fiq_return)) { + ret = -EINVAL; + goto err_inval; + } + fiq_glue_update_return_handler(NULL); + ret = 0; +err_inval: + mutex_unlock(&fiq_glue_lock); + + return ret; +} +EXPORT_SYMBOL(fiq_glue_clear_return_handler); + +/** + * fiq_glue_resume - Restore fiqs after suspend or low power idle states + * + * This must be called before calling local_fiq_enable after returning from a + * power state where the fiq mode registers were lost. If a driver provided + * a resume hook when it registered the handler it will be called. + */ + +void fiq_glue_resume(void) +{ + if (!current_handler) + return; + fiq_glue_setup(current_handler->fiq, current_handler, + __get_cpu_var(fiq_stack) + THREAD_START_SP, + fiq_return_handler); + if (current_handler->resume) + current_handler->resume(current_handler); +} + diff --git a/arch/arm/crypto/.gitignore b/arch/arm/crypto/.gitignore new file mode 100644 index 000000000000..6231d36b3635 --- /dev/null +++ b/arch/arm/crypto/.gitignore @@ -0,0 +1 @@ +aesbs-core.S diff --git a/arch/arm/crypto/Makefile b/arch/arm/crypto/Makefile index a2c83851bc90..2cee53b27238 100644 --- a/arch/arm/crypto/Makefile +++ b/arch/arm/crypto/Makefile @@ -3,7 +3,27 @@ # 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-$(CONFIG_CRYPTO_SHA512_ARM_NEON) += sha512-arm-neon.o -aes-arm-y := aes-armv4.o aes_glue.o -sha1-arm-y := sha1-armv4-large.o sha1_glue.o +aes-arm-y := aes-armv4.o aes_glue.o +aes-arm-bs-y := aesbs-core.o aesbs-glue.o +sha1-arm-y := sha1-armv4-large.o sha1_glue.o +sha1-arm-neon-y := sha1-armv7-neon.o sha1_neon_glue.o +sha256-arm-neon-$(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 + +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_glue.c b/arch/arm/crypto/aes_glue.c index e73ec2ab1316..0409b8f89782 100644 --- a/arch/arm/crypto/aes_glue.c +++ b/arch/arm/crypto/aes_glue.c @@ -6,22 +6,12 @@ #include <linux/crypto.h> #include <crypto/aes.h> -#define AES_MAXNR 14 +#include "aes_glue.h" -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); +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) { @@ -81,7 +71,7 @@ static struct crypto_alg aes_alg = { .cipher = { .cia_min_keysize = AES_MIN_KEY_SIZE, .cia_max_keysize = AES_MAX_KEY_SIZE, - .cia_setkey = aes_set_key, + .cia_setkey = aes_set_key, .cia_encrypt = aes_encrypt, .cia_decrypt = aes_decrypt } diff --git a/arch/arm/crypto/aes_glue.h b/arch/arm/crypto/aes_glue.h new file mode 100644 index 000000000000..cca3e51eb606 --- /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 000000000000..71e5fc7cfb18 --- /dev/null +++ b/arch/arm/crypto/aesbs-core.S_shipped @@ -0,0 +1,2544 @@ + +@ ==================================================================== +@ 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__ 7 +#endif + +#ifdef __thumb__ +# define adrl adr +#endif + +#if __ARM_ARCH__>=7 +.text +.syntax unified @ ARMv7-capable assembler is expected to handle this +#ifdef __thumb2__ +.thumb +#else +.code 32 +#endif + +.fpu neon + +.type _bsaes_decrypt8,%function +.align 4 +_bsaes_decrypt8: + adr 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 + + tst r9, #0xf @ if not multiple of 16 + it ne @ Thumb2 thing, sanity check in ARM + subne r9, #0x10 @ subtract another 16 bytes + 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 000000000000..15468fbbdea3 --- /dev/null +++ b/arch/arm/crypto/aesbs-glue.c @@ -0,0 +1,434 @@ +/* + * 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, + .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, + .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, + .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 000000000000..be068db960ee --- /dev/null +++ b/arch/arm/crypto/bsaes-armv7.pl @@ -0,0 +1,2467 @@ +#!/usr/bin/env perl + +# ==================================================================== +# Written by Andy Polyakov <appro@openssl.org> for the OpenSSL +# project. The module is, however, dual licensed under OpenSSL and +# CRYPTOGAMS licenses depending on where you obtain it. For further +# details see http://www.openssl.org/~appro/cryptogams/. +# +# Specific modes and adaptation for Linux kernel by Ard Biesheuvel +# <ard.biesheuvel@linaro.org>. Permission to use under GPL terms is +# granted. +# ==================================================================== + +# Bit-sliced AES for ARM NEON +# +# February 2012. +# +# This implementation is direct adaptation of bsaes-x86_64 module for +# ARM NEON. Except that this module is endian-neutral [in sense that +# it can be compiled for either endianness] by courtesy of vld1.8's +# neutrality. Initial version doesn't implement interface to OpenSSL, +# only low-level primitives and unsupported entry points, just enough +# to collect performance results, which for Cortex-A8 core are: +# +# encrypt 19.5 cycles per byte processed with 128-bit key +# decrypt 22.1 cycles per byte processed with 128-bit key +# key conv. 440 cycles per 128-bit key/0.18 of 8x block +# +# Snapdragon S4 encrypts byte in 17.6 cycles and decrypts in 19.7, +# which is [much] worse than anticipated (for further details see +# http://www.openssl.org/~appro/Snapdragon-S4.html). +# +# Cortex-A15 manages in 14.2/16.1 cycles [when integer-only code +# manages in 20.0 cycles]. +# +# When comparing to x86_64 results keep in mind that NEON unit is +# [mostly] single-issue and thus can't [fully] benefit from +# instruction-level parallelism. And when comparing to aes-armv4 +# results keep in mind key schedule conversion overhead (see +# bsaes-x86_64.pl for further details)... +# +# <appro@openssl.org> + +# April-August 2013 +# +# Add CBC, CTR and XTS subroutines, adapt for kernel use. +# +# <ard.biesheuvel@linaro.org> + +while (($output=shift) && ($output!~/^\w[\w\-]*\.\w+$/)) {} +open STDOUT,">$output"; + +my ($inp,$out,$len,$key)=("r0","r1","r2","r3"); +my @XMM=map("q$_",(0..15)); + +{ +my ($key,$rounds,$const)=("r4","r5","r6"); + +sub Dlo() { shift=~m|q([1]?[0-9])|?"d".($1*2):""; } +sub Dhi() { shift=~m|q([1]?[0-9])|?"d".($1*2+1):""; } + +sub Sbox { +# input in lsb > [b0, b1, b2, b3, b4, b5, b6, b7] < msb +# output in lsb > [b0, b1, b4, b6, b3, b7, b2, b5] < msb +my @b=@_[0..7]; +my @t=@_[8..11]; +my @s=@_[12..15]; + &InBasisChange (@b); + &Inv_GF256 (@b[6,5,0,3,7,1,4,2],@t,@s); + &OutBasisChange (@b[7,1,4,2,6,5,0,3]); +} + +sub InBasisChange { +# input in lsb > [b0, b1, b2, b3, b4, b5, b6, b7] < msb +# output in lsb > [b6, b5, b0, b3, b7, b1, b4, b2] < msb +my @b=@_[0..7]; +$code.=<<___; + veor @b[2], @b[2], @b[1] + veor @b[5], @b[5], @b[6] + veor @b[3], @b[3], @b[0] + veor @b[6], @b[6], @b[2] + veor @b[5], @b[5], @b[0] + + veor @b[6], @b[6], @b[3] + veor @b[3], @b[3], @b[7] + veor @b[7], @b[7], @b[5] + veor @b[3], @b[3], @b[4] + veor @b[4], @b[4], @b[5] + + veor @b[2], @b[2], @b[7] + veor @b[3], @b[3], @b[1] + veor @b[1], @b[1], @b[5] +___ +} + +sub OutBasisChange { +# input in lsb > [b0, b1, b2, b3, b4, b5, b6, b7] < msb +# output in lsb > [b6, b1, b2, b4, b7, b0, b3, b5] < msb +my @b=@_[0..7]; +$code.=<<___; + veor @b[0], @b[0], @b[6] + veor @b[1], @b[1], @b[4] + veor @b[4], @b[4], @b[6] + veor @b[2], @b[2], @b[0] + veor @b[6], @b[6], @b[1] + + veor @b[1], @b[1], @b[5] + veor @b[5], @b[5], @b[3] + veor @b[3], @b[3], @b[7] + veor @b[7], @b[7], @b[5] + veor @b[2], @b[2], @b[5] + + veor @b[4], @b[4], @b[7] +___ +} + +sub InvSbox { +# input in lsb > [b0, b1, b2, b3, b4, b5, b6, b7] < msb +# output in lsb > [b0, b1, b6, b4, b2, b7, b3, b5] < msb +my @b=@_[0..7]; +my @t=@_[8..11]; +my @s=@_[12..15]; + &InvInBasisChange (@b); + &Inv_GF256 (@b[5,1,2,6,3,7,0,4],@t,@s); + &InvOutBasisChange (@b[3,7,0,4,5,1,2,6]); +} + +sub InvInBasisChange { # OutBasisChange in reverse (with twist) +my @b=@_[5,1,2,6,3,7,0,4]; +$code.=<<___ + veor @b[1], @b[1], @b[7] + veor @b[4], @b[4], @b[7] + + veor @b[7], @b[7], @b[5] + veor @b[1], @b[1], @b[3] + veor @b[2], @b[2], @b[5] + veor @b[3], @b[3], @b[7] + + veor @b[6], @b[6], @b[1] + veor @b[2], @b[2], @b[0] + veor @b[5], @b[5], @b[3] + veor @b[4], @b[4], @b[6] + veor @b[0], @b[0], @b[6] + veor @b[1], @b[1], @b[4] +___ +} + +sub InvOutBasisChange { # InBasisChange in reverse +my @b=@_[2,5,7,3,6,1,0,4]; +$code.=<<___; + veor @b[1], @b[1], @b[5] + veor @b[2], @b[2], @b[7] + + veor @b[3], @b[3], @b[1] + veor @b[4], @b[4], @b[5] + veor @b[7], @b[7], @b[5] + veor @b[3], @b[3], @b[4] + veor @b[5], @b[5], @b[0] + veor @b[3], @b[3], @b[7] + veor @b[6], @b[6], @b[2] + veor @b[2], @b[2], @b[1] + veor @b[6], @b[6], @b[3] + + veor @b[3], @b[3], @b[0] + veor @b[5], @b[5], @b[6] +___ +} + +sub Mul_GF4 { +#;************************************************************* +#;* Mul_GF4: Input x0-x1,y0-y1 Output x0-x1 Temp t0 (8) * +#;************************************************************* +my ($x0,$x1,$y0,$y1,$t0,$t1)=@_; +$code.=<<___; + veor $t0, $y0, $y1 + vand $t0, $t0, $x0 + veor $x0, $x0, $x1 + vand $t1, $x1, $y0 + vand $x0, $x0, $y1 + veor $x1, $t1, $t0 + veor $x0, $x0, $t1 +___ +} + +sub Mul_GF4_N { # not used, see next subroutine +# multiply and scale by N +my ($x0,$x1,$y0,$y1,$t0)=@_; +$code.=<<___; + veor $t0, $y0, $y1 + vand $t0, $t0, $x0 + veor $x0, $x0, $x1 + vand $x1, $x1, $y0 + vand $x0, $x0, $y1 + veor $x1, $x1, $x0 + veor $x0, $x0, $t0 +___ +} + +sub Mul_GF4_N_GF4 { +# interleaved Mul_GF4_N and Mul_GF4 +my ($x0,$x1,$y0,$y1,$t0, + $x2,$x3,$y2,$y3,$t1)=@_; +$code.=<<___; + veor $t0, $y0, $y1 + veor $t1, $y2, $y3 + vand $t0, $t0, $x0 + vand $t1, $t1, $x2 + veor $x0, $x0, $x1 + veor $x2, $x2, $x3 + vand $x1, $x1, $y0 + vand $x3, $x3, $y2 + vand $x0, $x0, $y1 + vand $x2, $x2, $y3 + veor $x1, $x1, $x0 + veor $x2, $x2, $x3 + veor $x0, $x0, $t0 + veor $x3, $x3, $t1 +___ +} +sub Mul_GF16_2 { +my @x=@_[0..7]; +my @y=@_[8..11]; +my @t=@_[12..15]; +$code.=<<___; + veor @t[0], @x[0], @x[2] + veor @t[1], @x[1], @x[3] +___ + &Mul_GF4 (@x[0], @x[1], @y[0], @y[1], @t[2..3]); +$code.=<<___; + veor @y[0], @y[0], @y[2] + veor @y[1], @y[1], @y[3] +___ + Mul_GF4_N_GF4 (@t[0], @t[1], @y[0], @y[1], @t[3], + @x[2], @x[3], @y[2], @y[3], @t[2]); +$code.=<<___; + veor @x[0], @x[0], @t[0] + veor @x[2], @x[2], @t[0] + veor @x[1], @x[1], @t[1] + veor @x[3], @x[3], @t[1] + + veor @t[0], @x[4], @x[6] + veor @t[1], @x[5], @x[7] +___ + &Mul_GF4_N_GF4 (@t[0], @t[1], @y[0], @y[1], @t[3], + @x[6], @x[7], @y[2], @y[3], @t[2]); +$code.=<<___; + veor @y[0], @y[0], @y[2] + veor @y[1], @y[1], @y[3] +___ + &Mul_GF4 (@x[4], @x[5], @y[0], @y[1], @t[2..3]); +$code.=<<___; + veor @x[4], @x[4], @t[0] + veor @x[6], @x[6], @t[0] + veor @x[5], @x[5], @t[1] + veor @x[7], @x[7], @t[1] +___ +} +sub Inv_GF256 { +#;******************************************************************** +#;* Inv_GF256: Input x0-x7 Output x0-x7 Temp t0-t3,s0-s3 (144) * +#;******************************************************************** +my @x=@_[0..7]; +my @t=@_[8..11]; +my @s=@_[12..15]; +# direct optimizations from hardware +$code.=<<___; + veor @t[3], @x[4], @x[6] + veor @t[2], @x[5], @x[7] + veor @t[1], @x[1], @x[3] + veor @s[1], @x[7], @x[6] + vmov @t[0], @t[2] + veor @s[0], @x[0], @x[2] + + vorr @t[2], @t[2], @t[1] + veor @s[3], @t[3], @t[0] + vand @s[2], @t[3], @s[0] + vorr @t[3], @t[3], @s[0] + veor @s[0], @s[0], @t[1] + vand @t[0], @t[0], @t[1] + veor @t[1], @x[3], @x[2] + vand @s[3], @s[3], @s[0] + vand @s[1], @s[1], @t[1] + veor @t[1], @x[4], @x[5] + veor @s[0], @x[1], @x[0] + veor @t[3], @t[3], @s[1] + veor @t[2], @t[2], @s[1] + vand @s[1], @t[1], @s[0] + vorr @t[1], @t[1], @s[0] + veor @t[3], @t[3], @s[3] + veor @t[0], @t[0], @s[1] + veor @t[2], @t[2], @s[2] + veor @t[1], @t[1], @s[3] + veor @t[0], @t[0], @s[2] + vand @s[0], @x[7], @x[3] + veor @t[1], @t[1], @s[2] + vand @s[1], @x[6], @x[2] + vand @s[2], @x[5], @x[1] + vorr @s[3], @x[4], @x[0] + veor @t[3], @t[3], @s[0] + veor @t[1], @t[1], @s[2] + veor @t[0], @t[0], @s[3] + veor @t[2], @t[2], @s[1] + + @ Inv_GF16 \t0, \t1, \t2, \t3, \s0, \s1, \s2, \s3 + + @ new smaller inversion + + vand @s[2], @t[3], @t[1] + vmov @s[0], @t[0] + + veor @s[1], @t[2], @s[2] + veor @s[3], @t[0], @s[2] + veor @s[2], @t[0], @s[2] @ @s[2]=@s[3] + + vbsl @s[1], @t[1], @t[0] + vbsl @s[3], @t[3], @t[2] + veor @t[3], @t[3], @t[2] + + vbsl @s[0], @s[1], @s[2] + vbsl @t[0], @s[2], @s[1] + + vand @s[2], @s[0], @s[3] + veor @t[1], @t[1], @t[0] + + veor @s[2], @s[2], @t[3] +___ +# output in s3, s2, s1, t1 + +# Mul_GF16_2 \x0, \x1, \x2, \x3, \x4, \x5, \x6, \x7, \t2, \t3, \t0, \t1, \s0, \s1, \s2, \s3 + +# Mul_GF16_2 \x0, \x1, \x2, \x3, \x4, \x5, \x6, \x7, \s3, \s2, \s1, \t1, \s0, \t0, \t2, \t3 + &Mul_GF16_2(@x,@s[3,2,1],@t[1],@s[0],@t[0,2,3]); + +### output msb > [x3,x2,x1,x0,x7,x6,x5,x4] < lsb +} + +# AES linear components + +sub ShiftRows { +my @x=@_[0..7]; +my @t=@_[8..11]; +my $mask=pop; +$code.=<<___; + vldmia $key!, {@t[0]-@t[3]} + veor @t[0], @t[0], @x[0] + veor @t[1], @t[1], @x[1] + vtbl.8 `&Dlo(@x[0])`, {@t[0]}, `&Dlo($mask)` + vtbl.8 `&Dhi(@x[0])`, {@t[0]}, `&Dhi($mask)` + vldmia $key!, {@t[0]} + veor @t[2], @t[2], @x[2] + vtbl.8 `&Dlo(@x[1])`, {@t[1]}, `&Dlo($mask)` + vtbl.8 `&Dhi(@x[1])`, {@t[1]}, `&Dhi($mask)` + vldmia $key!, {@t[1]} + veor @t[3], @t[3], @x[3] + vtbl.8 `&Dlo(@x[2])`, {@t[2]}, `&Dlo($mask)` + vtbl.8 `&Dhi(@x[2])`, {@t[2]}, `&Dhi($mask)` + vldmia $key!, {@t[2]} + vtbl.8 `&Dlo(@x[3])`, {@t[3]}, `&Dlo($mask)` + vtbl.8 `&Dhi(@x[3])`, {@t[3]}, `&Dhi($mask)` + vldmia $key!, {@t[3]} + veor @t[0], @t[0], @x[4] + veor @t[1], @t[1], @x[5] + vtbl.8 `&Dlo(@x[4])`, {@t[0]}, `&Dlo($mask)` + vtbl.8 `&Dhi(@x[4])`, {@t[0]}, `&Dhi($mask)` + veor @t[2], @t[2], @x[6] + vtbl.8 `&Dlo(@x[5])`, {@t[1]}, `&Dlo($mask)` + vtbl.8 `&Dhi(@x[5])`, {@t[1]}, `&Dhi($mask)` + veor @t[3], @t[3], @x[7] + vtbl.8 `&Dlo(@x[6])`, {@t[2]}, `&Dlo($mask)` + vtbl.8 `&Dhi(@x[6])`, {@t[2]}, `&Dhi($mask)` + vtbl.8 `&Dlo(@x[7])`, {@t[3]}, `&Dlo($mask)` + vtbl.8 `&Dhi(@x[7])`, {@t[3]}, `&Dhi($mask)` +___ +} + +sub MixColumns { +# modified to emit output in order suitable for feeding back to aesenc[last] +my @x=@_[0..7]; +my @t=@_[8..15]; +my $inv=@_[16]; # optional +$code.=<<___; + vext.8 @t[0], @x[0], @x[0], #12 @ x0 <<< 32 + vext.8 @t[1], @x[1], @x[1], #12 + veor @x[0], @x[0], @t[0] @ x0 ^ (x0 <<< 32) + vext.8 @t[2], @x[2], @x[2], #12 + veor @x[1], @x[1], @t[1] + vext.8 @t[3], @x[3], @x[3], #12 + veor @x[2], @x[2], @t[2] + vext.8 @t[4], @x[4], @x[4], #12 + veor @x[3], @x[3], @t[3] + vext.8 @t[5], @x[5], @x[5], #12 + veor @x[4], @x[4], @t[4] + vext.8 @t[6], @x[6], @x[6], #12 + veor @x[5], @x[5], @t[5] + vext.8 @t[7], @x[7], @x[7], #12 + veor @x[6], @x[6], @t[6] + + veor @t[1], @t[1], @x[0] + veor @x[7], @x[7], @t[7] + vext.8 @x[0], @x[0], @x[0], #8 @ (x0 ^ (x0 <<< 32)) <<< 64) + veor @t[2], @t[2], @x[1] + veor @t[0], @t[0], @x[7] + veor @t[1], @t[1], @x[7] + vext.8 @x[1], @x[1], @x[1], #8 + veor @t[5], @t[5], @x[4] + veor @x[0], @x[0], @t[0] + veor @t[6], @t[6], @x[5] + veor @x[1], @x[1], @t[1] + vext.8 @t[0], @x[4], @x[4], #8 + veor @t[4], @t[4], @x[3] + vext.8 @t[1], @x[5], @x[5], #8 + veor @t[7], @t[7], @x[6] + vext.8 @x[4], @x[3], @x[3], #8 + veor @t[3], @t[3], @x[2] + vext.8 @x[5], @x[7], @x[7], #8 + veor @t[4], @t[4], @x[7] + vext.8 @x[3], @x[6], @x[6], #8 + veor @t[3], @t[3], @x[7] + vext.8 @x[6], @x[2], @x[2], #8 + veor @x[7], @t[1], @t[5] +___ +$code.=<<___ if (!$inv); + veor @x[2], @t[0], @t[4] + veor @x[4], @x[4], @t[3] + veor @x[5], @x[5], @t[7] + veor @x[3], @x[3], @t[6] + @ vmov @x[2], @t[0] + veor @x[6], @x[6], @t[2] + @ vmov @x[7], @t[1] +___ +$code.=<<___ if ($inv); + veor @t[3], @t[3], @x[4] + veor @x[5], @x[5], @t[7] + veor @x[2], @x[3], @t[6] + veor @x[3], @t[0], @t[4] + veor @x[4], @x[6], @t[2] + vmov @x[6], @t[3] + @ vmov @x[7], @t[1] +___ +} + +sub InvMixColumns_orig { +my @x=@_[0..7]; +my @t=@_[8..15]; + +$code.=<<___; + @ multiplication by 0x0e + vext.8 @t[7], @x[7], @x[7], #12 + vmov @t[2], @x[2] + veor @x[2], @x[2], @x[5] @ 2 5 + veor @x[7], @x[7], @x[5] @ 7 5 + vext.8 @t[0], @x[0], @x[0], #12 + vmov @t[5], @x[5] + veor @x[5], @x[5], @x[0] @ 5 0 [1] + veor @x[0], @x[0], @x[1] @ 0 1 + vext.8 @t[1], @x[1], @x[1], #12 + veor @x[1], @x[1], @x[2] @ 1 25 + veor @x[0], @x[0], @x[6] @ 01 6 [2] + vext.8 @t[3], @x[3], @x[3], #12 + veor @x[1], @x[1], @x[3] @ 125 3 [4] + veor @x[2], @x[2], @x[0] @ 25 016 [3] + veor @x[3], @x[3], @x[7] @ 3 75 + veor @x[7], @x[7], @x[6] @ 75 6 [0] + vext.8 @t[6], @x[6], @x[6], #12 + vmov @t[4], @x[4] + veor @x[6], @x[6], @x[4] @ 6 4 + veor @x[4], @x[4], @x[3] @ 4 375 [6] + veor @x[3], @x[3], @x[7] @ 375 756=36 + veor @x[6], @x[6], @t[5] @ 64 5 [7] + veor @x[3], @x[3], @t[2] @ 36 2 + vext.8 @t[5], @t[5], @t[5], #12 + veor @x[3], @x[3], @t[4] @ 362 4 [5] +___ + my @y = @x[7,5,0,2,1,3,4,6]; +$code.=<<___; + @ multiplication by 0x0b + veor @y[1], @y[1], @y[0] + veor @y[0], @y[0], @t[0] + vext.8 @t[2], @t[2], @t[2], #12 + veor @y[1], @y[1], @t[1] + veor @y[0], @y[0], @t[5] + vext.8 @t[4], @t[4], @t[4], #12 + veor @y[1], @y[1], @t[6] + veor @y[0], @y[0], @t[7] + veor @t[7], @t[7], @t[6] @ clobber t[7] + + veor @y[3], @y[3], @t[0] + veor @y[1], @y[1], @y[0] + vext.8 @t[0], @t[0], @t[0], #12 + veor @y[2], @y[2], @t[1] + veor @y[4], @y[4], @t[1] + vext.8 @t[1], @t[1], @t[1], #12 + veor @y[2], @y[2], @t[2] + veor @y[3], @y[3], @t[2] + veor @y[5], @y[5], @t[2] + veor @y[2], @y[2], @t[7] + vext.8 @t[2], @t[2], @t[2], #12 + veor @y[3], @y[3], @t[3] + veor @y[6], @y[6], @t[3] + veor @y[4], @y[4], @t[3] + veor @y[7], @y[7], @t[4] + vext.8 @t[3], @t[3], @t[3], #12 + veor @y[5], @y[5], @t[4] + veor @y[7], @y[7], @t[7] + veor @t[7], @t[7], @t[5] @ clobber t[7] even more + veor @y[3], @y[3], @t[5] + veor @y[4], @y[4], @t[4] + + veor @y[5], @y[5], @t[7] + vext.8 @t[4], @t[4], @t[4], #12 + veor @y[6], @y[6], @t[7] + veor @y[4], @y[4], @t[7] + + veor @t[7], @t[7], @t[5] + vext.8 @t[5], @t[5], @t[5], #12 + + @ multiplication by 0x0d + veor @y[4], @y[4], @y[7] + veor @t[7], @t[7], @t[6] @ restore t[7] + veor @y[7], @y[7], @t[4] + vext.8 @t[6], @t[6], @t[6], #12 + veor @y[2], @y[2], @t[0] + veor @y[7], @y[7], @t[5] + vext.8 @t[7], @t[7], @t[7], #12 + veor @y[2], @y[2], @t[2] + + veor @y[3], @y[3], @y[1] + veor @y[1], @y[1], @t[1] + veor @y[0], @y[0], @t[0] + veor @y[3], @y[3], @t[0] + veor @y[1], @y[1], @t[5] + veor @y[0], @y[0], @t[5] + vext.8 @t[0], @t[0], @t[0], #12 + veor @y[1], @y[1], @t[7] + veor @y[0], @y[0], @t[6] + veor @y[3], @y[3], @y[1] + veor @y[4], @y[4], @t[1] + vext.8 @t[1], @t[1], @t[1], #12 + + veor @y[7], @y[7], @t[7] + veor @y[4], @y[4], @t[2] + veor @y[5], @y[5], @t[2] + veor @y[2], @y[2], @t[6] + veor @t[6], @t[6], @t[3] @ clobber t[6] + vext.8 @t[2], @t[2], @t[2], #12 + veor @y[4], @y[4], @y[7] + veor @y[3], @y[3], @t[6] + + veor @y[6], @y[6], @t[6] + veor @y[5], @y[5], @t[5] + vext.8 @t[5], @t[5], @t[5], #12 + veor @y[6], @y[6], @t[4] + vext.8 @t[4], @t[4], @t[4], #12 + veor @y[5], @y[5], @t[6] + veor @y[6], @y[6], @t[7] + vext.8 @t[7], @t[7], @t[7], #12 + veor @t[6], @t[6], @t[3] @ restore t[6] + vext.8 @t[3], @t[3], @t[3], #12 + + @ multiplication by 0x09 + veor @y[4], @y[4], @y[1] + veor @t[1], @t[1], @y[1] @ t[1]=y[1] + veor @t[0], @t[0], @t[5] @ clobber t[0] + vext.8 @t[6], @t[6], @t[6], #12 + veor @t[1], @t[1], @t[5] + veor @y[3], @y[3], @t[0] + veor @t[0], @t[0], @y[0] @ t[0]=y[0] + veor @t[1], @t[1], @t[6] + veor @t[6], @t[6], @t[7] @ clobber t[6] + veor @y[4], @y[4], @t[1] + veor @y[7], @y[7], @t[4] + veor @y[6], @y[6], @t[3] + veor @y[5], @y[5], @t[2] + veor @t[4], @t[4], @y[4] @ t[4]=y[4] + veor @t[3], @t[3], @y[3] @ t[3]=y[3] + veor @t[5], @t[5], @y[5] @ t[5]=y[5] + veor @t[2], @t[2], @y[2] @ t[2]=y[2] + veor @t[3], @t[3], @t[7] + veor @XMM[5], @t[5], @t[6] + veor @XMM[6], @t[6], @y[6] @ t[6]=y[6] + veor @XMM[2], @t[2], @t[6] + veor @XMM[7], @t[7], @y[7] @ t[7]=y[7] + + vmov @XMM[0], @t[0] + vmov @XMM[1], @t[1] + @ vmov @XMM[2], @t[2] + vmov @XMM[3], @t[3] + vmov @XMM[4], @t[4] + @ vmov @XMM[5], @t[5] + @ vmov @XMM[6], @t[6] + @ vmov @XMM[7], @t[7] +___ +} + +sub InvMixColumns { +my @x=@_[0..7]; +my @t=@_[8..15]; + +# Thanks to Jussi Kivilinna for providing pointer to +# +# | 0e 0b 0d 09 | | 02 03 01 01 | | 05 00 04 00 | +# | 09 0e 0b 0d | = | 01 02 03 01 | x | 00 05 00 04 | +# | 0d 09 0e 0b | | 01 01 02 03 | | 04 00 05 00 | +# | 0b 0d 09 0e | | 03 01 01 02 | | 00 04 00 05 | + +$code.=<<___; + @ multiplication by 0x05-0x00-0x04-0x00 + vext.8 @t[0], @x[0], @x[0], #8 + vext.8 @t[6], @x[6], @x[6], #8 + vext.8 @t[7], @x[7], @x[7], #8 + veor @t[0], @t[0], @x[0] + vext.8 @t[1], @x[1], @x[1], #8 + veor @t[6], @t[6], @x[6] + vext.8 @t[2], @x[2], @x[2], #8 + veor @t[7], @t[7], @x[7] + vext.8 @t[3], @x[3], @x[3], #8 + veor @t[1], @t[1], @x[1] + vext.8 @t[4], @x[4], @x[4], #8 + veor @t[2], @t[2], @x[2] + vext.8 @t[5], @x[5], @x[5], #8 + veor @t[3], @t[3], @x[3] + veor @t[4], @t[4], @x[4] + veor @t[5], @t[5], @x[5] + + veor @x[0], @x[0], @t[6] + veor @x[1], @x[1], @t[6] + veor @x[2], @x[2], @t[0] + veor @x[4], @x[4], @t[2] + veor @x[3], @x[3], @t[1] + veor @x[1], @x[1], @t[7] + veor @x[2], @x[2], @t[7] + veor @x[4], @x[4], @t[6] + veor @x[5], @x[5], @t[3] + veor @x[3], @x[3], @t[6] + veor @x[6], @x[6], @t[4] + veor @x[4], @x[4], @t[7] + veor @x[5], @x[5], @t[7] + veor @x[7], @x[7], @t[5] +___ + &MixColumns (@x,@t,1); # flipped 2<->3 and 4<->6 +} + +sub swapmove { +my ($a,$b,$n,$mask,$t)=@_; +$code.=<<___; + vshr.u64 $t, $b, #$n + veor $t, $t, $a + vand $t, $t, $mask + veor $a, $a, $t + vshl.u64 $t, $t, #$n + veor $b, $b, $t +___ +} +sub swapmove2x { +my ($a0,$b0,$a1,$b1,$n,$mask,$t0,$t1)=@_; +$code.=<<___; + vshr.u64 $t0, $b0, #$n + vshr.u64 $t1, $b1, #$n + veor $t0, $t0, $a0 + veor $t1, $t1, $a1 + vand $t0, $t0, $mask + vand $t1, $t1, $mask + veor $a0, $a0, $t0 + vshl.u64 $t0, $t0, #$n + veor $a1, $a1, $t1 + vshl.u64 $t1, $t1, #$n + veor $b0, $b0, $t0 + veor $b1, $b1, $t1 +___ +} + +sub bitslice { +my @x=reverse(@_[0..7]); +my ($t0,$t1,$t2,$t3)=@_[8..11]; +$code.=<<___; + vmov.i8 $t0,#0x55 @ compose .LBS0 + vmov.i8 $t1,#0x33 @ compose .LBS1 +___ + &swapmove2x(@x[0,1,2,3],1,$t0,$t2,$t3); + &swapmove2x(@x[4,5,6,7],1,$t0,$t2,$t3); +$code.=<<___; + vmov.i8 $t0,#0x0f @ compose .LBS2 +___ + &swapmove2x(@x[0,2,1,3],2,$t1,$t2,$t3); + &swapmove2x(@x[4,6,5,7],2,$t1,$t2,$t3); + + &swapmove2x(@x[0,4,1,5],4,$t0,$t2,$t3); + &swapmove2x(@x[2,6,3,7],4,$t0,$t2,$t3); +} + +$code.=<<___; +#ifndef __KERNEL__ +# include "arm_arch.h" + +# define VFP_ABI_PUSH vstmdb sp!,{d8-d15} +# define VFP_ABI_POP vldmia sp!,{d8-d15} +# define VFP_ABI_FRAME 0x40 +#else +# define VFP_ABI_PUSH +# define VFP_ABI_POP +# define VFP_ABI_FRAME 0 +# define BSAES_ASM_EXTENDED_KEY +# define XTS_CHAIN_TWEAK +# define __ARM_ARCH__ 7 +#endif + +#ifdef __thumb__ +# define adrl adr +#endif + +#if __ARM_ARCH__>=7 +.text +.syntax unified @ ARMv7-capable assembler is expected to handle this +#ifdef __thumb2__ +.thumb +#else +.code 32 +#endif + +.fpu neon + +.type _bsaes_decrypt8,%function +.align 4 +_bsaes_decrypt8: + adr $const,_bsaes_decrypt8 + vldmia $key!, {@XMM[9]} @ round 0 key + add $const,$const,#.LM0ISR-_bsaes_decrypt8 + + vldmia $const!, {@XMM[8]} @ .LM0ISR + veor @XMM[10], @XMM[0], @XMM[9] @ xor with round0 key + veor @XMM[11], @XMM[1], @XMM[9] + vtbl.8 `&Dlo(@XMM[0])`, {@XMM[10]}, `&Dlo(@XMM[8])` + vtbl.8 `&Dhi(@XMM[0])`, {@XMM[10]}, `&Dhi(@XMM[8])` + veor @XMM[12], @XMM[2], @XMM[9] + vtbl.8 `&Dlo(@XMM[1])`, {@XMM[11]}, `&Dlo(@XMM[8])` + vtbl.8 `&Dhi(@XMM[1])`, {@XMM[11]}, `&Dhi(@XMM[8])` + veor @XMM[13], @XMM[3], @XMM[9] + vtbl.8 `&Dlo(@XMM[2])`, {@XMM[12]}, `&Dlo(@XMM[8])` + vtbl.8 `&Dhi(@XMM[2])`, {@XMM[12]}, `&Dhi(@XMM[8])` + veor @XMM[14], @XMM[4], @XMM[9] + vtbl.8 `&Dlo(@XMM[3])`, {@XMM[13]}, `&Dlo(@XMM[8])` + vtbl.8 `&Dhi(@XMM[3])`, {@XMM[13]}, `&Dhi(@XMM[8])` + veor @XMM[15], @XMM[5], @XMM[9] + vtbl.8 `&Dlo(@XMM[4])`, {@XMM[14]}, `&Dlo(@XMM[8])` + vtbl.8 `&Dhi(@XMM[4])`, {@XMM[14]}, `&Dhi(@XMM[8])` + veor @XMM[10], @XMM[6], @XMM[9] + vtbl.8 `&Dlo(@XMM[5])`, {@XMM[15]}, `&Dlo(@XMM[8])` + vtbl.8 `&Dhi(@XMM[5])`, {@XMM[15]}, `&Dhi(@XMM[8])` + veor @XMM[11], @XMM[7], @XMM[9] + vtbl.8 `&Dlo(@XMM[6])`, {@XMM[10]}, `&Dlo(@XMM[8])` + vtbl.8 `&Dhi(@XMM[6])`, {@XMM[10]}, `&Dhi(@XMM[8])` + vtbl.8 `&Dlo(@XMM[7])`, {@XMM[11]}, `&Dlo(@XMM[8])` + vtbl.8 `&Dhi(@XMM[7])`, {@XMM[11]}, `&Dhi(@XMM[8])` +___ + &bitslice (@XMM[0..7, 8..11]); +$code.=<<___; + sub $rounds,$rounds,#1 + b .Ldec_sbox +.align 4 +.Ldec_loop: +___ + &ShiftRows (@XMM[0..7, 8..12]); +$code.=".Ldec_sbox:\n"; + &InvSbox (@XMM[0..7, 8..15]); +$code.=<<___; + subs $rounds,$rounds,#1 + bcc .Ldec_done +___ + &InvMixColumns (@XMM[0,1,6,4,2,7,3,5, 8..15]); +$code.=<<___; + vldmia $const, {@XMM[12]} @ .LISR + ite eq @ Thumb2 thing, sanity check in ARM + addeq $const,$const,#0x10 + bne .Ldec_loop + vldmia $const, {@XMM[12]} @ .LISRM0 + b .Ldec_loop +.align 4 +.Ldec_done: +___ + &bitslice (@XMM[0,1,6,4,2,7,3,5, 8..11]); +$code.=<<___; + vldmia $key, {@XMM[8]} @ last round key + veor @XMM[6], @XMM[6], @XMM[8] + veor @XMM[4], @XMM[4], @XMM[8] + veor @XMM[2], @XMM[2], @XMM[8] + veor @XMM[7], @XMM[7], @XMM[8] + veor @XMM[3], @XMM[3], @XMM[8] + veor @XMM[5], @XMM[5], @XMM[8] + veor @XMM[0], @XMM[0], @XMM[8] + veor @XMM[1], @XMM[1], @XMM[8] + bx lr +.size _bsaes_decrypt8,.-_bsaes_decrypt8 + +.type _bsaes_const,%object +.align 6 +_bsaes_const: +.LM0ISR: @ InvShiftRows constants + .quad 0x0a0e0206070b0f03, 0x0004080c0d010509 +.LISR: + .quad 0x0504070602010003, 0x0f0e0d0c080b0a09 +.LISRM0: + .quad 0x01040b0e0205080f, 0x0306090c00070a0d +.LM0SR: @ ShiftRows constants + .quad 0x0a0e02060f03070b, 0x0004080c05090d01 +.LSR: + .quad 0x0504070600030201, 0x0f0e0d0c0a09080b +.LSRM0: + .quad 0x0304090e00050a0f, 0x01060b0c0207080d +.LM0: + .quad 0x02060a0e03070b0f, 0x0004080c0105090d +.LREVM0SR: + .quad 0x090d01050c000408, 0x03070b0f060a0e02 +.asciz "Bit-sliced AES for NEON, CRYPTOGAMS by <appro\@openssl.org>" +.align 6 +.size _bsaes_const,.-_bsaes_const + +.type _bsaes_encrypt8,%function +.align 4 +_bsaes_encrypt8: + adr $const,_bsaes_encrypt8 + vldmia $key!, {@XMM[9]} @ round 0 key + sub $const,$const,#_bsaes_encrypt8-.LM0SR + + vldmia $const!, {@XMM[8]} @ .LM0SR +_bsaes_encrypt8_alt: + veor @XMM[10], @XMM[0], @XMM[9] @ xor with round0 key + veor @XMM[11], @XMM[1], @XMM[9] + vtbl.8 `&Dlo(@XMM[0])`, {@XMM[10]}, `&Dlo(@XMM[8])` + vtbl.8 `&Dhi(@XMM[0])`, {@XMM[10]}, `&Dhi(@XMM[8])` + veor @XMM[12], @XMM[2], @XMM[9] + vtbl.8 `&Dlo(@XMM[1])`, {@XMM[11]}, `&Dlo(@XMM[8])` + vtbl.8 `&Dhi(@XMM[1])`, {@XMM[11]}, `&Dhi(@XMM[8])` + veor @XMM[13], @XMM[3], @XMM[9] + vtbl.8 `&Dlo(@XMM[2])`, {@XMM[12]}, `&Dlo(@XMM[8])` + vtbl.8 `&Dhi(@XMM[2])`, {@XMM[12]}, `&Dhi(@XMM[8])` + veor @XMM[14], @XMM[4], @XMM[9] + vtbl.8 `&Dlo(@XMM[3])`, {@XMM[13]}, `&Dlo(@XMM[8])` + vtbl.8 `&Dhi(@XMM[3])`, {@XMM[13]}, `&Dhi(@XMM[8])` + veor @XMM[15], @XMM[5], @XMM[9] + vtbl.8 `&Dlo(@XMM[4])`, {@XMM[14]}, `&Dlo(@XMM[8])` + vtbl.8 `&Dhi(@XMM[4])`, {@XMM[14]}, `&Dhi(@XMM[8])` + veor @XMM[10], @XMM[6], @XMM[9] + vtbl.8 `&Dlo(@XMM[5])`, {@XMM[15]}, `&Dlo(@XMM[8])` + vtbl.8 `&Dhi(@XMM[5])`, {@XMM[15]}, `&Dhi(@XMM[8])` + veor @XMM[11], @XMM[7], @XMM[9] + vtbl.8 `&Dlo(@XMM[6])`, {@XMM[10]}, `&Dlo(@XMM[8])` + vtbl.8 `&Dhi(@XMM[6])`, {@XMM[10]}, `&Dhi(@XMM[8])` + vtbl.8 `&Dlo(@XMM[7])`, {@XMM[11]}, `&Dlo(@XMM[8])` + vtbl.8 `&Dhi(@XMM[7])`, {@XMM[11]}, `&Dhi(@XMM[8])` +_bsaes_encrypt8_bitslice: +___ + &bitslice (@XMM[0..7, 8..11]); +$code.=<<___; + sub $rounds,$rounds,#1 + b .Lenc_sbox +.align 4 +.Lenc_loop: +___ + &ShiftRows (@XMM[0..7, 8..12]); +$code.=".Lenc_sbox:\n"; + &Sbox (@XMM[0..7, 8..15]); +$code.=<<___; + subs $rounds,$rounds,#1 + bcc .Lenc_done +___ + &MixColumns (@XMM[0,1,4,6,3,7,2,5, 8..15]); +$code.=<<___; + vldmia $const, {@XMM[12]} @ .LSR + ite eq @ Thumb2 thing, samity check in ARM + addeq $const,$const,#0x10 + bne .Lenc_loop + vldmia $const, {@XMM[12]} @ .LSRM0 + b .Lenc_loop +.align 4 +.Lenc_done: +___ + # output in lsb > [t0, t1, t4, t6, t3, t7, t2, t5] < msb + &bitslice (@XMM[0,1,4,6,3,7,2,5, 8..11]); +$code.=<<___; + vldmia $key, {@XMM[8]} @ last round key + veor @XMM[4], @XMM[4], @XMM[8] + veor @XMM[6], @XMM[6], @XMM[8] + veor @XMM[3], @XMM[3], @XMM[8] + veor @XMM[7], @XMM[7], @XMM[8] + veor @XMM[2], @XMM[2], @XMM[8] + veor @XMM[5], @XMM[5], @XMM[8] + veor @XMM[0], @XMM[0], @XMM[8] + veor @XMM[1], @XMM[1], @XMM[8] + bx lr +.size _bsaes_encrypt8,.-_bsaes_encrypt8 +___ +} +{ +my ($out,$inp,$rounds,$const)=("r12","r4","r5","r6"); + +sub bitslice_key { +my @x=reverse(@_[0..7]); +my ($bs0,$bs1,$bs2,$t2,$t3)=@_[8..12]; + + &swapmove (@x[0,1],1,$bs0,$t2,$t3); +$code.=<<___; + @ &swapmove(@x[2,3],1,$t0,$t2,$t3); + vmov @x[2], @x[0] + vmov @x[3], @x[1] +___ + #&swapmove2x(@x[4,5,6,7],1,$t0,$t2,$t3); + + &swapmove2x (@x[0,2,1,3],2,$bs1,$t2,$t3); +$code.=<<___; + @ &swapmove2x(@x[4,6,5,7],2,$t1,$t2,$t3); + vmov @x[4], @x[0] + vmov @x[6], @x[2] + vmov @x[5], @x[1] + vmov @x[7], @x[3] +___ + &swapmove2x (@x[0,4,1,5],4,$bs2,$t2,$t3); + &swapmove2x (@x[2,6,3,7],4,$bs2,$t2,$t3); +} + +$code.=<<___; +.type _bsaes_key_convert,%function +.align 4 +_bsaes_key_convert: + adr $const,_bsaes_key_convert + vld1.8 {@XMM[7]}, [$inp]! @ load round 0 key + sub $const,$const,#_bsaes_key_convert-.LM0 + vld1.8 {@XMM[15]}, [$inp]! @ load round 1 key + + vmov.i8 @XMM[8], #0x01 @ bit masks + vmov.i8 @XMM[9], #0x02 + vmov.i8 @XMM[10], #0x04 + vmov.i8 @XMM[11], #0x08 + vmov.i8 @XMM[12], #0x10 + vmov.i8 @XMM[13], #0x20 + vldmia $const, {@XMM[14]} @ .LM0 + +#ifdef __ARMEL__ + vrev32.8 @XMM[7], @XMM[7] + vrev32.8 @XMM[15], @XMM[15] +#endif + sub $rounds,$rounds,#1 + vstmia $out!, {@XMM[7]} @ save round 0 key + b .Lkey_loop + +.align 4 +.Lkey_loop: + vtbl.8 `&Dlo(@XMM[7])`,{@XMM[15]},`&Dlo(@XMM[14])` + vtbl.8 `&Dhi(@XMM[7])`,{@XMM[15]},`&Dhi(@XMM[14])` + vmov.i8 @XMM[6], #0x40 + vmov.i8 @XMM[15], #0x80 + + vtst.8 @XMM[0], @XMM[7], @XMM[8] + vtst.8 @XMM[1], @XMM[7], @XMM[9] + vtst.8 @XMM[2], @XMM[7], @XMM[10] + vtst.8 @XMM[3], @XMM[7], @XMM[11] + vtst.8 @XMM[4], @XMM[7], @XMM[12] + vtst.8 @XMM[5], @XMM[7], @XMM[13] + vtst.8 @XMM[6], @XMM[7], @XMM[6] + vtst.8 @XMM[7], @XMM[7], @XMM[15] + vld1.8 {@XMM[15]}, [$inp]! @ load next round key + vmvn @XMM[0], @XMM[0] @ "pnot" + vmvn @XMM[1], @XMM[1] + vmvn @XMM[5], @XMM[5] + vmvn @XMM[6], @XMM[6] +#ifdef __ARMEL__ + vrev32.8 @XMM[15], @XMM[15] +#endif + subs $rounds,$rounds,#1 + vstmia $out!,{@XMM[0]-@XMM[7]} @ write bit-sliced round key + bne .Lkey_loop + + vmov.i8 @XMM[7],#0x63 @ compose .L63 + @ don't save last round key + bx lr +.size _bsaes_key_convert,.-_bsaes_key_convert +___ +} + +if (0) { # following four functions are unsupported interface + # used for benchmarking... +$code.=<<___; +.globl bsaes_enc_key_convert +.type bsaes_enc_key_convert,%function +.align 4 +bsaes_enc_key_convert: + stmdb sp!,{r4-r6,lr} + vstmdb sp!,{d8-d15} @ ABI specification says so + + ldr r5,[$inp,#240] @ pass rounds + mov r4,$inp @ pass key + mov r12,$out @ pass key schedule + bl _bsaes_key_convert + veor @XMM[7],@XMM[7],@XMM[15] @ fix up last round key + vstmia r12, {@XMM[7]} @ save last round key + + vldmia sp!,{d8-d15} + ldmia sp!,{r4-r6,pc} +.size bsaes_enc_key_convert,.-bsaes_enc_key_convert + +.globl bsaes_encrypt_128 +.type bsaes_encrypt_128,%function +.align 4 +bsaes_encrypt_128: + stmdb sp!,{r4-r6,lr} + vstmdb sp!,{d8-d15} @ ABI specification says so +.Lenc128_loop: + vld1.8 {@XMM[0]-@XMM[1]}, [$inp]! @ load input + vld1.8 {@XMM[2]-@XMM[3]}, [$inp]! + mov r4,$key @ pass the key + vld1.8 {@XMM[4]-@XMM[5]}, [$inp]! + mov r5,#10 @ pass rounds + vld1.8 {@XMM[6]-@XMM[7]}, [$inp]! + + bl _bsaes_encrypt8 + + vst1.8 {@XMM[0]-@XMM[1]}, [$out]! @ write output + vst1.8 {@XMM[4]}, [$out]! + vst1.8 {@XMM[6]}, [$out]! + vst1.8 {@XMM[3]}, [$out]! + vst1.8 {@XMM[7]}, [$out]! + vst1.8 {@XMM[2]}, [$out]! + subs $len,$len,#0x80 + vst1.8 {@XMM[5]}, [$out]! + bhi .Lenc128_loop + + vldmia sp!,{d8-d15} + ldmia sp!,{r4-r6,pc} +.size bsaes_encrypt_128,.-bsaes_encrypt_128 + +.globl bsaes_dec_key_convert +.type bsaes_dec_key_convert,%function +.align 4 +bsaes_dec_key_convert: + stmdb sp!,{r4-r6,lr} + vstmdb sp!,{d8-d15} @ ABI specification says so + + ldr r5,[$inp,#240] @ pass rounds + mov r4,$inp @ pass key + mov r12,$out @ pass key schedule + bl _bsaes_key_convert + vldmia $out, {@XMM[6]} + vstmia r12, {@XMM[15]} @ save last round key + veor @XMM[7], @XMM[7], @XMM[6] @ fix up round 0 key + vstmia $out, {@XMM[7]} + + vldmia sp!,{d8-d15} + ldmia sp!,{r4-r6,pc} +.size bsaes_dec_key_convert,.-bsaes_dec_key_convert + +.globl bsaes_decrypt_128 +.type bsaes_decrypt_128,%function +.align 4 +bsaes_decrypt_128: + stmdb sp!,{r4-r6,lr} + vstmdb sp!,{d8-d15} @ ABI specification says so +.Ldec128_loop: + vld1.8 {@XMM[0]-@XMM[1]}, [$inp]! @ load input + vld1.8 {@XMM[2]-@XMM[3]}, [$inp]! + mov r4,$key @ pass the key + vld1.8 {@XMM[4]-@XMM[5]}, [$inp]! + mov r5,#10 @ pass rounds + vld1.8 {@XMM[6]-@XMM[7]}, [$inp]! + + bl _bsaes_decrypt8 + + vst1.8 {@XMM[0]-@XMM[1]}, [$out]! @ write output + vst1.8 {@XMM[6]}, [$out]! + vst1.8 {@XMM[4]}, [$out]! + vst1.8 {@XMM[2]}, [$out]! + vst1.8 {@XMM[7]}, [$out]! + vst1.8 {@XMM[3]}, [$out]! + subs $len,$len,#0x80 + vst1.8 {@XMM[5]}, [$out]! + bhi .Ldec128_loop + + vldmia sp!,{d8-d15} + ldmia sp!,{r4-r6,pc} +.size bsaes_decrypt_128,.-bsaes_decrypt_128 +___ +} +{ +my ($inp,$out,$len,$key, $ivp,$fp,$rounds)=map("r$_",(0..3,8..10)); +my ($keysched)=("sp"); + +$code.=<<___; +.extern AES_cbc_encrypt +.extern AES_decrypt + +.global bsaes_cbc_encrypt +.type bsaes_cbc_encrypt,%function +.align 5 +bsaes_cbc_encrypt: +#ifndef __KERNEL__ + cmp $len, #128 +#ifndef __thumb__ + blo AES_cbc_encrypt +#else + bhs 1f + b AES_cbc_encrypt +1: +#endif +#endif + + @ it is up to the caller to make sure we are called with enc == 0 + + mov ip, sp + stmdb sp!, {r4-r10, lr} + VFP_ABI_PUSH + ldr $ivp, [ip] @ IV is 1st arg on the stack + mov $len, $len, lsr#4 @ len in 16 byte blocks + sub sp, #0x10 @ scratch space to carry over the IV + mov $fp, sp @ save sp + + ldr $rounds, [$key, #240] @ get # of rounds +#ifndef BSAES_ASM_EXTENDED_KEY + @ allocate the key schedule on the stack + sub r12, sp, $rounds, lsl#7 @ 128 bytes per inner round key + add r12, #`128-32` @ sifze of bit-slices key schedule + + @ populate the key schedule + mov r4, $key @ pass key + mov r5, $rounds @ pass # of rounds + mov sp, r12 @ sp is $keysched + bl _bsaes_key_convert + vldmia $keysched, {@XMM[6]} + vstmia r12, {@XMM[15]} @ save last round key + veor @XMM[7], @XMM[7], @XMM[6] @ fix up round 0 key + vstmia $keysched, {@XMM[7]} +#else + ldr r12, [$key, #244] + eors r12, #1 + beq 0f + + @ populate the key schedule + str r12, [$key, #244] + mov r4, $key @ pass key + mov r5, $rounds @ pass # of rounds + add r12, $key, #248 @ pass key schedule + bl _bsaes_key_convert + add r4, $key, #248 + vldmia r4, {@XMM[6]} + vstmia r12, {@XMM[15]} @ save last round key + veor @XMM[7], @XMM[7], @XMM[6] @ fix up round 0 key + vstmia r4, {@XMM[7]} + +.align 2 +0: +#endif + + vld1.8 {@XMM[15]}, [$ivp] @ load IV + b .Lcbc_dec_loop + +.align 4 +.Lcbc_dec_loop: + subs $len, $len, #0x8 + bmi .Lcbc_dec_loop_finish + + vld1.8 {@XMM[0]-@XMM[1]}, [$inp]! @ load input + vld1.8 {@XMM[2]-@XMM[3]}, [$inp]! +#ifndef BSAES_ASM_EXTENDED_KEY + mov r4, $keysched @ pass the key +#else + add r4, $key, #248 +#endif + vld1.8 {@XMM[4]-@XMM[5]}, [$inp]! + mov r5, $rounds + vld1.8 {@XMM[6]-@XMM[7]}, [$inp] + sub $inp, $inp, #0x60 + vstmia $fp, {@XMM[15]} @ put aside IV + + bl _bsaes_decrypt8 + + vldmia $fp, {@XMM[14]} @ reload IV + vld1.8 {@XMM[8]-@XMM[9]}, [$inp]! @ reload input + veor @XMM[0], @XMM[0], @XMM[14] @ ^= IV + vld1.8 {@XMM[10]-@XMM[11]}, [$inp]! + veor @XMM[1], @XMM[1], @XMM[8] + veor @XMM[6], @XMM[6], @XMM[9] + vld1.8 {@XMM[12]-@XMM[13]}, [$inp]! + veor @XMM[4], @XMM[4], @XMM[10] + veor @XMM[2], @XMM[2], @XMM[11] + vld1.8 {@XMM[14]-@XMM[15]}, [$inp]! + veor @XMM[7], @XMM[7], @XMM[12] + vst1.8 {@XMM[0]-@XMM[1]}, [$out]! @ write output + veor @XMM[3], @XMM[3], @XMM[13] + vst1.8 {@XMM[6]}, [$out]! + veor @XMM[5], @XMM[5], @XMM[14] + vst1.8 {@XMM[4]}, [$out]! + vst1.8 {@XMM[2]}, [$out]! + vst1.8 {@XMM[7]}, [$out]! + vst1.8 {@XMM[3]}, [$out]! + vst1.8 {@XMM[5]}, [$out]! + + b .Lcbc_dec_loop + +.Lcbc_dec_loop_finish: + adds $len, $len, #8 + beq .Lcbc_dec_done + + vld1.8 {@XMM[0]}, [$inp]! @ load input + cmp $len, #2 + blo .Lcbc_dec_one + vld1.8 {@XMM[1]}, [$inp]! +#ifndef BSAES_ASM_EXTENDED_KEY + mov r4, $keysched @ pass the key +#else + add r4, $key, #248 +#endif + mov r5, $rounds + vstmia $fp, {@XMM[15]} @ put aside IV + beq .Lcbc_dec_two + vld1.8 {@XMM[2]}, [$inp]! + cmp $len, #4 + blo .Lcbc_dec_three + vld1.8 {@XMM[3]}, [$inp]! + beq .Lcbc_dec_four + vld1.8 {@XMM[4]}, [$inp]! + cmp $len, #6 + blo .Lcbc_dec_five + vld1.8 {@XMM[5]}, [$inp]! + beq .Lcbc_dec_six + vld1.8 {@XMM[6]}, [$inp]! + sub $inp, $inp, #0x70 + + bl _bsaes_decrypt8 + + vldmia $fp, {@XMM[14]} @ reload IV + vld1.8 {@XMM[8]-@XMM[9]}, [$inp]! @ reload input + veor @XMM[0], @XMM[0], @XMM[14] @ ^= IV + vld1.8 {@XMM[10]-@XMM[11]}, [$inp]! + veor @XMM[1], @XMM[1], @XMM[8] + veor @XMM[6], @XMM[6], @XMM[9] + vld1.8 {@XMM[12]-@XMM[13]}, [$inp]! + veor @XMM[4], @XMM[4], @XMM[10] + veor @XMM[2], @XMM[2], @XMM[11] + vld1.8 {@XMM[15]}, [$inp]! + veor @XMM[7], @XMM[7], @XMM[12] + vst1.8 {@XMM[0]-@XMM[1]}, [$out]! @ write output + veor @XMM[3], @XMM[3], @XMM[13] + vst1.8 {@XMM[6]}, [$out]! + vst1.8 {@XMM[4]}, [$out]! + vst1.8 {@XMM[2]}, [$out]! + vst1.8 {@XMM[7]}, [$out]! + vst1.8 {@XMM[3]}, [$out]! + b .Lcbc_dec_done +.align 4 +.Lcbc_dec_six: + sub $inp, $inp, #0x60 + bl _bsaes_decrypt8 + vldmia $fp,{@XMM[14]} @ reload IV + vld1.8 {@XMM[8]-@XMM[9]}, [$inp]! @ reload input + veor @XMM[0], @XMM[0], @XMM[14] @ ^= IV + vld1.8 {@XMM[10]-@XMM[11]}, [$inp]! + veor @XMM[1], @XMM[1], @XMM[8] + veor @XMM[6], @XMM[6], @XMM[9] + vld1.8 {@XMM[12]}, [$inp]! + veor @XMM[4], @XMM[4], @XMM[10] + veor @XMM[2], @XMM[2], @XMM[11] + vld1.8 {@XMM[15]}, [$inp]! + veor @XMM[7], @XMM[7], @XMM[12] + vst1.8 {@XMM[0]-@XMM[1]}, [$out]! @ write output + vst1.8 {@XMM[6]}, [$out]! + vst1.8 {@XMM[4]}, [$out]! + vst1.8 {@XMM[2]}, [$out]! + vst1.8 {@XMM[7]}, [$out]! + b .Lcbc_dec_done +.align 4 +.Lcbc_dec_five: + sub $inp, $inp, #0x50 + bl _bsaes_decrypt8 + vldmia $fp, {@XMM[14]} @ reload IV + vld1.8 {@XMM[8]-@XMM[9]}, [$inp]! @ reload input + veor @XMM[0], @XMM[0], @XMM[14] @ ^= IV + vld1.8 {@XMM[10]-@XMM[11]}, [$inp]! + veor @XMM[1], @XMM[1], @XMM[8] + veor @XMM[6], @XMM[6], @XMM[9] + vld1.8 {@XMM[15]}, [$inp]! + veor @XMM[4], @XMM[4], @XMM[10] + vst1.8 {@XMM[0]-@XMM[1]}, [$out]! @ write output + veor @XMM[2], @XMM[2], @XMM[11] + vst1.8 {@XMM[6]}, [$out]! + vst1.8 {@XMM[4]}, [$out]! + vst1.8 {@XMM[2]}, [$out]! + b .Lcbc_dec_done +.align 4 +.Lcbc_dec_four: + sub $inp, $inp, #0x40 + bl _bsaes_decrypt8 + vldmia $fp, {@XMM[14]} @ reload IV + vld1.8 {@XMM[8]-@XMM[9]}, [$inp]! @ reload input + veor @XMM[0], @XMM[0], @XMM[14] @ ^= IV + vld1.8 {@XMM[10]}, [$inp]! + veor @XMM[1], @XMM[1], @XMM[8] + veor @XMM[6], @XMM[6], @XMM[9] + vld1.8 {@XMM[15]}, [$inp]! + veor @XMM[4], @XMM[4], @XMM[10] + vst1.8 {@XMM[0]-@XMM[1]}, [$out]! @ write output + vst1.8 {@XMM[6]}, [$out]! + vst1.8 {@XMM[4]}, [$out]! + b .Lcbc_dec_done +.align 4 +.Lcbc_dec_three: + sub $inp, $inp, #0x30 + bl _bsaes_decrypt8 + vldmia $fp, {@XMM[14]} @ reload IV + vld1.8 {@XMM[8]-@XMM[9]}, [$inp]! @ reload input + veor @XMM[0], @XMM[0], @XMM[14] @ ^= IV + vld1.8 {@XMM[15]}, [$inp]! + veor @XMM[1], @XMM[1], @XMM[8] + veor @XMM[6], @XMM[6], @XMM[9] + vst1.8 {@XMM[0]-@XMM[1]}, [$out]! @ write output + vst1.8 {@XMM[6]}, [$out]! + b .Lcbc_dec_done +.align 4 +.Lcbc_dec_two: + sub $inp, $inp, #0x20 + bl _bsaes_decrypt8 + vldmia $fp, {@XMM[14]} @ reload IV + vld1.8 {@XMM[8]}, [$inp]! @ reload input + veor @XMM[0], @XMM[0], @XMM[14] @ ^= IV + vld1.8 {@XMM[15]}, [$inp]! @ reload input + veor @XMM[1], @XMM[1], @XMM[8] + vst1.8 {@XMM[0]-@XMM[1]}, [$out]! @ write output + b .Lcbc_dec_done +.align 4 +.Lcbc_dec_one: + sub $inp, $inp, #0x10 + mov $rounds, $out @ save original out pointer + mov $out, $fp @ use the iv scratch space as out buffer + mov r2, $key + vmov @XMM[4],@XMM[15] @ just in case ensure that IV + vmov @XMM[5],@XMM[0] @ and input are preserved + bl AES_decrypt + vld1.8 {@XMM[0]}, [$fp,:64] @ load result + veor @XMM[0], @XMM[0], @XMM[4] @ ^= IV + vmov @XMM[15], @XMM[5] @ @XMM[5] holds input + vst1.8 {@XMM[0]}, [$rounds] @ write output + +.Lcbc_dec_done: +#ifndef BSAES_ASM_EXTENDED_KEY + vmov.i32 q0, #0 + vmov.i32 q1, #0 +.Lcbc_dec_bzero: @ wipe key schedule [if any] + vstmia $keysched!, {q0-q1} + cmp $keysched, $fp + bne .Lcbc_dec_bzero +#endif + + mov sp, $fp + add sp, #0x10 @ add sp,$fp,#0x10 is no good for thumb + vst1.8 {@XMM[15]}, [$ivp] @ return IV + VFP_ABI_POP + ldmia sp!, {r4-r10, pc} +.size bsaes_cbc_encrypt,.-bsaes_cbc_encrypt +___ +} +{ +my ($inp,$out,$len,$key, $ctr,$fp,$rounds)=(map("r$_",(0..3,8..10))); +my $const = "r6"; # shared with _bsaes_encrypt8_alt +my $keysched = "sp"; + +$code.=<<___; +.extern AES_encrypt +.global bsaes_ctr32_encrypt_blocks +.type bsaes_ctr32_encrypt_blocks,%function +.align 5 +bsaes_ctr32_encrypt_blocks: + cmp $len, #8 @ use plain AES for + blo .Lctr_enc_short @ small sizes + + mov ip, sp + stmdb sp!, {r4-r10, lr} + VFP_ABI_PUSH + ldr $ctr, [ip] @ ctr is 1st arg on the stack + sub sp, sp, #0x10 @ scratch space to carry over the ctr + mov $fp, sp @ save sp + + ldr $rounds, [$key, #240] @ get # of rounds +#ifndef BSAES_ASM_EXTENDED_KEY + @ allocate the key schedule on the stack + sub r12, sp, $rounds, lsl#7 @ 128 bytes per inner round key + add r12, #`128-32` @ size of bit-sliced key schedule + + @ populate the key schedule + mov r4, $key @ pass key + mov r5, $rounds @ pass # of rounds + mov sp, r12 @ sp is $keysched + bl _bsaes_key_convert + veor @XMM[7],@XMM[7],@XMM[15] @ fix up last round key + vstmia r12, {@XMM[7]} @ save last round key + + vld1.8 {@XMM[0]}, [$ctr] @ load counter + add $ctr, $const, #.LREVM0SR-.LM0 @ borrow $ctr + vldmia $keysched, {@XMM[4]} @ load round0 key +#else + ldr r12, [$key, #244] + eors r12, #1 + beq 0f + + @ populate the key schedule + str r12, [$key, #244] + mov r4, $key @ pass key + mov r5, $rounds @ pass # of rounds + add r12, $key, #248 @ pass key schedule + bl _bsaes_key_convert + veor @XMM[7],@XMM[7],@XMM[15] @ fix up last round key + vstmia r12, {@XMM[7]} @ save last round key + +.align 2 +0: add r12, $key, #248 + vld1.8 {@XMM[0]}, [$ctr] @ load counter + adrl $ctr, .LREVM0SR @ borrow $ctr + vldmia r12, {@XMM[4]} @ load round0 key + sub sp, #0x10 @ place for adjusted round0 key +#endif + + vmov.i32 @XMM[8],#1 @ compose 1<<96 + veor @XMM[9],@XMM[9],@XMM[9] + vrev32.8 @XMM[0],@XMM[0] + vext.8 @XMM[8],@XMM[9],@XMM[8],#4 + vrev32.8 @XMM[4],@XMM[4] + vadd.u32 @XMM[9],@XMM[8],@XMM[8] @ compose 2<<96 + vstmia $keysched, {@XMM[4]} @ save adjusted round0 key + b .Lctr_enc_loop + +.align 4 +.Lctr_enc_loop: + vadd.u32 @XMM[10], @XMM[8], @XMM[9] @ compose 3<<96 + vadd.u32 @XMM[1], @XMM[0], @XMM[8] @ +1 + vadd.u32 @XMM[2], @XMM[0], @XMM[9] @ +2 + vadd.u32 @XMM[3], @XMM[0], @XMM[10] @ +3 + vadd.u32 @XMM[4], @XMM[1], @XMM[10] + vadd.u32 @XMM[5], @XMM[2], @XMM[10] + vadd.u32 @XMM[6], @XMM[3], @XMM[10] + vadd.u32 @XMM[7], @XMM[4], @XMM[10] + vadd.u32 @XMM[10], @XMM[5], @XMM[10] @ next counter + + @ Borrow prologue from _bsaes_encrypt8 to use the opportunity + @ to flip byte order in 32-bit counter + + vldmia $keysched, {@XMM[9]} @ load round0 key +#ifndef BSAES_ASM_EXTENDED_KEY + add r4, $keysched, #0x10 @ pass next round key +#else + add r4, $key, #`248+16` +#endif + vldmia $ctr, {@XMM[8]} @ .LREVM0SR + mov r5, $rounds @ pass rounds + vstmia $fp, {@XMM[10]} @ save next counter + sub $const, $ctr, #.LREVM0SR-.LSR @ pass constants + + bl _bsaes_encrypt8_alt + + subs $len, $len, #8 + blo .Lctr_enc_loop_done + + vld1.8 {@XMM[8]-@XMM[9]}, [$inp]! @ load input + vld1.8 {@XMM[10]-@XMM[11]}, [$inp]! + veor @XMM[0], @XMM[8] + veor @XMM[1], @XMM[9] + vld1.8 {@XMM[12]-@XMM[13]}, [$inp]! + veor @XMM[4], @XMM[10] + veor @XMM[6], @XMM[11] + vld1.8 {@XMM[14]-@XMM[15]}, [$inp]! + veor @XMM[3], @XMM[12] + vst1.8 {@XMM[0]-@XMM[1]}, [$out]! @ write output + veor @XMM[7], @XMM[13] + veor @XMM[2], @XMM[14] + vst1.8 {@XMM[4]}, [$out]! + veor @XMM[5], @XMM[15] + vst1.8 {@XMM[6]}, [$out]! + vmov.i32 @XMM[8], #1 @ compose 1<<96 + vst1.8 {@XMM[3]}, [$out]! + veor @XMM[9], @XMM[9], @XMM[9] + vst1.8 {@XMM[7]}, [$out]! + vext.8 @XMM[8], @XMM[9], @XMM[8], #4 + vst1.8 {@XMM[2]}, [$out]! + vadd.u32 @XMM[9],@XMM[8],@XMM[8] @ compose 2<<96 + vst1.8 {@XMM[5]}, [$out]! + vldmia $fp, {@XMM[0]} @ load counter + + bne .Lctr_enc_loop + b .Lctr_enc_done + +.align 4 +.Lctr_enc_loop_done: + add $len, $len, #8 + vld1.8 {@XMM[8]}, [$inp]! @ load input + veor @XMM[0], @XMM[8] + vst1.8 {@XMM[0]}, [$out]! @ write output + cmp $len, #2 + blo .Lctr_enc_done + vld1.8 {@XMM[9]}, [$inp]! + veor @XMM[1], @XMM[9] + vst1.8 {@XMM[1]}, [$out]! + beq .Lctr_enc_done + vld1.8 {@XMM[10]}, [$inp]! + veor @XMM[4], @XMM[10] + vst1.8 {@XMM[4]}, [$out]! + cmp $len, #4 + blo .Lctr_enc_done + vld1.8 {@XMM[11]}, [$inp]! + veor @XMM[6], @XMM[11] + vst1.8 {@XMM[6]}, [$out]! + beq .Lctr_enc_done + vld1.8 {@XMM[12]}, [$inp]! + veor @XMM[3], @XMM[12] + vst1.8 {@XMM[3]}, [$out]! + cmp $len, #6 + blo .Lctr_enc_done + vld1.8 {@XMM[13]}, [$inp]! + veor @XMM[7], @XMM[13] + vst1.8 {@XMM[7]}, [$out]! + beq .Lctr_enc_done + vld1.8 {@XMM[14]}, [$inp] + veor @XMM[2], @XMM[14] + vst1.8 {@XMM[2]}, [$out]! + +.Lctr_enc_done: + vmov.i32 q0, #0 + vmov.i32 q1, #0 +#ifndef BSAES_ASM_EXTENDED_KEY +.Lctr_enc_bzero: @ wipe key schedule [if any] + vstmia $keysched!, {q0-q1} + cmp $keysched, $fp + bne .Lctr_enc_bzero +#else + vstmia $keysched, {q0-q1} +#endif + + mov sp, $fp + add sp, #0x10 @ add sp,$fp,#0x10 is no good for thumb + VFP_ABI_POP + ldmia sp!, {r4-r10, pc} @ return + +.align 4 +.Lctr_enc_short: + ldr ip, [sp] @ ctr pointer is passed on stack + stmdb sp!, {r4-r8, lr} + + mov r4, $inp @ copy arguments + mov r5, $out + mov r6, $len + mov r7, $key + ldr r8, [ip, #12] @ load counter LSW + vld1.8 {@XMM[1]}, [ip] @ load whole counter value +#ifdef __ARMEL__ + rev r8, r8 +#endif + sub sp, sp, #0x10 + vst1.8 {@XMM[1]}, [sp,:64] @ copy counter value + sub sp, sp, #0x10 + +.Lctr_enc_short_loop: + add r0, sp, #0x10 @ input counter value + mov r1, sp @ output on the stack + mov r2, r7 @ key + + bl AES_encrypt + + vld1.8 {@XMM[0]}, [r4]! @ load input + vld1.8 {@XMM[1]}, [sp,:64] @ load encrypted counter + add r8, r8, #1 +#ifdef __ARMEL__ + rev r0, r8 + str r0, [sp, #0x1c] @ next counter value +#else + str r8, [sp, #0x1c] @ next counter value +#endif + veor @XMM[0],@XMM[0],@XMM[1] + vst1.8 {@XMM[0]}, [r5]! @ store output + subs r6, r6, #1 + bne .Lctr_enc_short_loop + + vmov.i32 q0, #0 + vmov.i32 q1, #0 + vstmia sp!, {q0-q1} + + ldmia sp!, {r4-r8, pc} +.size bsaes_ctr32_encrypt_blocks,.-bsaes_ctr32_encrypt_blocks +___ +} +{ +###################################################################### +# void bsaes_xts_[en|de]crypt(const char *inp,char *out,size_t len, +# const AES_KEY *key1, const AES_KEY *key2, +# const unsigned char iv[16]); +# +my ($inp,$out,$len,$key,$rounds,$magic,$fp)=(map("r$_",(7..10,1..3))); +my $const="r6"; # returned by _bsaes_key_convert +my $twmask=@XMM[5]; +my @T=@XMM[6..7]; + +$code.=<<___; +.globl bsaes_xts_encrypt +.type bsaes_xts_encrypt,%function +.align 4 +bsaes_xts_encrypt: + mov ip, sp + stmdb sp!, {r4-r10, lr} @ 0x20 + VFP_ABI_PUSH + mov r6, sp @ future $fp + + mov $inp, r0 + mov $out, r1 + mov $len, r2 + mov $key, r3 + + sub r0, sp, #0x10 @ 0x10 + bic r0, #0xf @ align at 16 bytes + mov sp, r0 + +#ifdef XTS_CHAIN_TWEAK + ldr r0, [ip] @ pointer to input tweak +#else + @ generate initial tweak + ldr r0, [ip, #4] @ iv[] + mov r1, sp + ldr r2, [ip, #0] @ key2 + bl AES_encrypt + mov r0,sp @ pointer to initial tweak +#endif + + ldr $rounds, [$key, #240] @ get # of rounds + mov $fp, r6 +#ifndef BSAES_ASM_EXTENDED_KEY + @ allocate the key schedule on the stack + sub r12, sp, $rounds, lsl#7 @ 128 bytes per inner round key + @ add r12, #`128-32` @ size of bit-sliced key schedule + sub r12, #`32+16` @ place for tweak[9] + + @ populate the key schedule + mov r4, $key @ pass key + mov r5, $rounds @ pass # of rounds + mov sp, r12 + add r12, #0x90 @ pass key schedule + bl _bsaes_key_convert + veor @XMM[7], @XMM[7], @XMM[15] @ fix up last round key + vstmia r12, {@XMM[7]} @ save last round key +#else + ldr r12, [$key, #244] + eors r12, #1 + beq 0f + + str r12, [$key, #244] + mov r4, $key @ pass key + mov r5, $rounds @ pass # of rounds + add r12, $key, #248 @ pass key schedule + bl _bsaes_key_convert + veor @XMM[7], @XMM[7], @XMM[15] @ fix up last round key + vstmia r12, {@XMM[7]} + +.align 2 +0: sub sp, #0x90 @ place for tweak[9] +#endif + + vld1.8 {@XMM[8]}, [r0] @ initial tweak + adr $magic, .Lxts_magic + + subs $len, #0x80 + blo .Lxts_enc_short + b .Lxts_enc_loop + +.align 4 +.Lxts_enc_loop: + vldmia $magic, {$twmask} @ load XTS magic + vshr.s64 @T[0], @XMM[8], #63 + mov r0, sp + vand @T[0], @T[0], $twmask +___ +for($i=9;$i<16;$i++) { +$code.=<<___; + vadd.u64 @XMM[$i], @XMM[$i-1], @XMM[$i-1] + vst1.64 {@XMM[$i-1]}, [r0,:128]! + vswp `&Dhi("@T[0]")`,`&Dlo("@T[0]")` + vshr.s64 @T[1], @XMM[$i], #63 + veor @XMM[$i], @XMM[$i], @T[0] + vand @T[1], @T[1], $twmask +___ + @T=reverse(@T); + +$code.=<<___ if ($i>=10); + vld1.8 {@XMM[$i-10]}, [$inp]! +___ +$code.=<<___ if ($i>=11); + veor @XMM[$i-11], @XMM[$i-11], @XMM[$i-3] +___ +} +$code.=<<___; + vadd.u64 @XMM[8], @XMM[15], @XMM[15] + vst1.64 {@XMM[15]}, [r0,:128]! + vswp `&Dhi("@T[0]")`,`&Dlo("@T[0]")` + veor @XMM[8], @XMM[8], @T[0] + vst1.64 {@XMM[8]}, [r0,:128] @ next round tweak + + vld1.8 {@XMM[6]-@XMM[7]}, [$inp]! + veor @XMM[5], @XMM[5], @XMM[13] +#ifndef BSAES_ASM_EXTENDED_KEY + add r4, sp, #0x90 @ pass key schedule +#else + add r4, $key, #248 @ pass key schedule +#endif + veor @XMM[6], @XMM[6], @XMM[14] + mov r5, $rounds @ pass rounds + veor @XMM[7], @XMM[7], @XMM[15] + mov r0, sp + + bl _bsaes_encrypt8 + + vld1.64 {@XMM[ 8]-@XMM[ 9]}, [r0,:128]! + vld1.64 {@XMM[10]-@XMM[11]}, [r0,:128]! + veor @XMM[0], @XMM[0], @XMM[ 8] + vld1.64 {@XMM[12]-@XMM[13]}, [r0,:128]! + veor @XMM[1], @XMM[1], @XMM[ 9] + veor @XMM[8], @XMM[4], @XMM[10] + vst1.8 {@XMM[0]-@XMM[1]}, [$out]! + veor @XMM[9], @XMM[6], @XMM[11] + vld1.64 {@XMM[14]-@XMM[15]}, [r0,:128]! + veor @XMM[10], @XMM[3], @XMM[12] + vst1.8 {@XMM[8]-@XMM[9]}, [$out]! + veor @XMM[11], @XMM[7], @XMM[13] + veor @XMM[12], @XMM[2], @XMM[14] + vst1.8 {@XMM[10]-@XMM[11]}, [$out]! + veor @XMM[13], @XMM[5], @XMM[15] + vst1.8 {@XMM[12]-@XMM[13]}, [$out]! + + vld1.64 {@XMM[8]}, [r0,:128] @ next round tweak + + subs $len, #0x80 + bpl .Lxts_enc_loop + +.Lxts_enc_short: + adds $len, #0x70 + bmi .Lxts_enc_done + + vldmia $magic, {$twmask} @ load XTS magic + vshr.s64 @T[0], @XMM[8], #63 + mov r0, sp + vand @T[0], @T[0], $twmask +___ +for($i=9;$i<16;$i++) { +$code.=<<___; + vadd.u64 @XMM[$i], @XMM[$i-1], @XMM[$i-1] + vst1.64 {@XMM[$i-1]}, [r0,:128]! + vswp `&Dhi("@T[0]")`,`&Dlo("@T[0]")` + vshr.s64 @T[1], @XMM[$i], #63 + veor @XMM[$i], @XMM[$i], @T[0] + vand @T[1], @T[1], $twmask +___ + @T=reverse(@T); + +$code.=<<___ if ($i>=10); + vld1.8 {@XMM[$i-10]}, [$inp]! + subs $len, #0x10 + bmi .Lxts_enc_`$i-9` +___ +$code.=<<___ if ($i>=11); + veor @XMM[$i-11], @XMM[$i-11], @XMM[$i-3] +___ +} +$code.=<<___; + sub $len, #0x10 + vst1.64 {@XMM[15]}, [r0,:128] @ next round tweak + + vld1.8 {@XMM[6]}, [$inp]! + veor @XMM[5], @XMM[5], @XMM[13] +#ifndef BSAES_ASM_EXTENDED_KEY + add r4, sp, #0x90 @ pass key schedule +#else + add r4, $key, #248 @ pass key schedule +#endif + veor @XMM[6], @XMM[6], @XMM[14] + mov r5, $rounds @ pass rounds + mov r0, sp + + bl _bsaes_encrypt8 + + vld1.64 {@XMM[ 8]-@XMM[ 9]}, [r0,:128]! + vld1.64 {@XMM[10]-@XMM[11]}, [r0,:128]! + veor @XMM[0], @XMM[0], @XMM[ 8] + vld1.64 {@XMM[12]-@XMM[13]}, [r0,:128]! + veor @XMM[1], @XMM[1], @XMM[ 9] + veor @XMM[8], @XMM[4], @XMM[10] + vst1.8 {@XMM[0]-@XMM[1]}, [$out]! + veor @XMM[9], @XMM[6], @XMM[11] + vld1.64 {@XMM[14]}, [r0,:128]! + veor @XMM[10], @XMM[3], @XMM[12] + vst1.8 {@XMM[8]-@XMM[9]}, [$out]! + veor @XMM[11], @XMM[7], @XMM[13] + veor @XMM[12], @XMM[2], @XMM[14] + vst1.8 {@XMM[10]-@XMM[11]}, [$out]! + vst1.8 {@XMM[12]}, [$out]! + + vld1.64 {@XMM[8]}, [r0,:128] @ next round tweak + b .Lxts_enc_done +.align 4 +.Lxts_enc_6: + vst1.64 {@XMM[14]}, [r0,:128] @ next round tweak + + veor @XMM[4], @XMM[4], @XMM[12] +#ifndef BSAES_ASM_EXTENDED_KEY + add r4, sp, #0x90 @ pass key schedule +#else + add r4, $key, #248 @ pass key schedule +#endif + veor @XMM[5], @XMM[5], @XMM[13] + mov r5, $rounds @ pass rounds + mov r0, sp + + bl _bsaes_encrypt8 + + vld1.64 {@XMM[ 8]-@XMM[ 9]}, [r0,:128]! + vld1.64 {@XMM[10]-@XMM[11]}, [r0,:128]! + veor @XMM[0], @XMM[0], @XMM[ 8] + vld1.64 {@XMM[12]-@XMM[13]}, [r0,:128]! + veor @XMM[1], @XMM[1], @XMM[ 9] + veor @XMM[8], @XMM[4], @XMM[10] + vst1.8 {@XMM[0]-@XMM[1]}, [$out]! + veor @XMM[9], @XMM[6], @XMM[11] + veor @XMM[10], @XMM[3], @XMM[12] + vst1.8 {@XMM[8]-@XMM[9]}, [$out]! + veor @XMM[11], @XMM[7], @XMM[13] + vst1.8 {@XMM[10]-@XMM[11]}, [$out]! + + vld1.64 {@XMM[8]}, [r0,:128] @ next round tweak + b .Lxts_enc_done + +@ put this in range for both ARM and Thumb mode adr instructions +.align 5 +.Lxts_magic: + .quad 1, 0x87 + +.align 5 +.Lxts_enc_5: + vst1.64 {@XMM[13]}, [r0,:128] @ next round tweak + + veor @XMM[3], @XMM[3], @XMM[11] +#ifndef BSAES_ASM_EXTENDED_KEY + add r4, sp, #0x90 @ pass key schedule +#else + add r4, $key, #248 @ pass key schedule +#endif + veor @XMM[4], @XMM[4], @XMM[12] + mov r5, $rounds @ pass rounds + mov r0, sp + + bl _bsaes_encrypt8 + + vld1.64 {@XMM[ 8]-@XMM[ 9]}, [r0,:128]! + vld1.64 {@XMM[10]-@XMM[11]}, [r0,:128]! + veor @XMM[0], @XMM[0], @XMM[ 8] + vld1.64 {@XMM[12]}, [r0,:128]! + veor @XMM[1], @XMM[1], @XMM[ 9] + veor @XMM[8], @XMM[4], @XMM[10] + vst1.8 {@XMM[0]-@XMM[1]}, [$out]! + veor @XMM[9], @XMM[6], @XMM[11] + veor @XMM[10], @XMM[3], @XMM[12] + vst1.8 {@XMM[8]-@XMM[9]}, [$out]! + vst1.8 {@XMM[10]}, [$out]! + + vld1.64 {@XMM[8]}, [r0,:128] @ next round tweak + b .Lxts_enc_done +.align 4 +.Lxts_enc_4: + vst1.64 {@XMM[12]}, [r0,:128] @ next round tweak + + veor @XMM[2], @XMM[2], @XMM[10] +#ifndef BSAES_ASM_EXTENDED_KEY + add r4, sp, #0x90 @ pass key schedule +#else + add r4, $key, #248 @ pass key schedule +#endif + veor @XMM[3], @XMM[3], @XMM[11] + mov r5, $rounds @ pass rounds + mov r0, sp + + bl _bsaes_encrypt8 + + vld1.64 {@XMM[ 8]-@XMM[ 9]}, [r0,:128]! + vld1.64 {@XMM[10]-@XMM[11]}, [r0,:128]! + veor @XMM[0], @XMM[0], @XMM[ 8] + veor @XMM[1], @XMM[1], @XMM[ 9] + veor @XMM[8], @XMM[4], @XMM[10] + vst1.8 {@XMM[0]-@XMM[1]}, [$out]! + veor @XMM[9], @XMM[6], @XMM[11] + vst1.8 {@XMM[8]-@XMM[9]}, [$out]! + + vld1.64 {@XMM[8]}, [r0,:128] @ next round tweak + b .Lxts_enc_done +.align 4 +.Lxts_enc_3: + vst1.64 {@XMM[11]}, [r0,:128] @ next round tweak + + veor @XMM[1], @XMM[1], @XMM[9] +#ifndef BSAES_ASM_EXTENDED_KEY + add r4, sp, #0x90 @ pass key schedule +#else + add r4, $key, #248 @ pass key schedule +#endif + veor @XMM[2], @XMM[2], @XMM[10] + mov r5, $rounds @ pass rounds + mov r0, sp + + bl _bsaes_encrypt8 + + vld1.64 {@XMM[8]-@XMM[9]}, [r0,:128]! + vld1.64 {@XMM[10]}, [r0,:128]! + veor @XMM[0], @XMM[0], @XMM[ 8] + veor @XMM[1], @XMM[1], @XMM[ 9] + veor @XMM[8], @XMM[4], @XMM[10] + vst1.8 {@XMM[0]-@XMM[1]}, [$out]! + vst1.8 {@XMM[8]}, [$out]! + + vld1.64 {@XMM[8]}, [r0,:128] @ next round tweak + b .Lxts_enc_done +.align 4 +.Lxts_enc_2: + vst1.64 {@XMM[10]}, [r0,:128] @ next round tweak + + veor @XMM[0], @XMM[0], @XMM[8] +#ifndef BSAES_ASM_EXTENDED_KEY + add r4, sp, #0x90 @ pass key schedule +#else + add r4, $key, #248 @ pass key schedule +#endif + veor @XMM[1], @XMM[1], @XMM[9] + mov r5, $rounds @ pass rounds + mov r0, sp + + bl _bsaes_encrypt8 + + vld1.64 {@XMM[8]-@XMM[9]}, [r0,:128]! + veor @XMM[0], @XMM[0], @XMM[ 8] + veor @XMM[1], @XMM[1], @XMM[ 9] + vst1.8 {@XMM[0]-@XMM[1]}, [$out]! + + vld1.64 {@XMM[8]}, [r0,:128] @ next round tweak + b .Lxts_enc_done +.align 4 +.Lxts_enc_1: + mov r0, sp + veor @XMM[0], @XMM[8] + mov r1, sp + vst1.8 {@XMM[0]}, [sp,:128] + mov r2, $key + mov r4, $fp @ preserve fp + + bl AES_encrypt + + vld1.8 {@XMM[0]}, [sp,:128] + veor @XMM[0], @XMM[0], @XMM[8] + vst1.8 {@XMM[0]}, [$out]! + mov $fp, r4 + + vmov @XMM[8], @XMM[9] @ next round tweak + +.Lxts_enc_done: +#ifndef XTS_CHAIN_TWEAK + adds $len, #0x10 + beq .Lxts_enc_ret + sub r6, $out, #0x10 + +.Lxts_enc_steal: + ldrb r0, [$inp], #1 + ldrb r1, [$out, #-0x10] + strb r0, [$out, #-0x10] + strb r1, [$out], #1 + + subs $len, #1 + bhi .Lxts_enc_steal + + vld1.8 {@XMM[0]}, [r6] + mov r0, sp + veor @XMM[0], @XMM[0], @XMM[8] + mov r1, sp + vst1.8 {@XMM[0]}, [sp,:128] + mov r2, $key + mov r4, $fp @ preserve fp + + bl AES_encrypt + + vld1.8 {@XMM[0]}, [sp,:128] + veor @XMM[0], @XMM[0], @XMM[8] + vst1.8 {@XMM[0]}, [r6] + mov $fp, r4 +#endif + +.Lxts_enc_ret: + bic r0, $fp, #0xf + vmov.i32 q0, #0 + vmov.i32 q1, #0 +#ifdef XTS_CHAIN_TWEAK + ldr r1, [$fp, #0x20+VFP_ABI_FRAME] @ chain tweak +#endif +.Lxts_enc_bzero: @ wipe key schedule [if any] + vstmia sp!, {q0-q1} + cmp sp, r0 + bne .Lxts_enc_bzero + + mov sp, $fp +#ifdef XTS_CHAIN_TWEAK + vst1.8 {@XMM[8]}, [r1] +#endif + VFP_ABI_POP + ldmia sp!, {r4-r10, pc} @ return + +.size bsaes_xts_encrypt,.-bsaes_xts_encrypt + +.globl bsaes_xts_decrypt +.type bsaes_xts_decrypt,%function +.align 4 +bsaes_xts_decrypt: + mov ip, sp + stmdb sp!, {r4-r10, lr} @ 0x20 + VFP_ABI_PUSH + mov r6, sp @ future $fp + + mov $inp, r0 + mov $out, r1 + mov $len, r2 + mov $key, r3 + + sub r0, sp, #0x10 @ 0x10 + bic r0, #0xf @ align at 16 bytes + mov sp, r0 + +#ifdef XTS_CHAIN_TWEAK + ldr r0, [ip] @ pointer to input tweak +#else + @ generate initial tweak + ldr r0, [ip, #4] @ iv[] + mov r1, sp + ldr r2, [ip, #0] @ key2 + bl AES_encrypt + mov r0, sp @ pointer to initial tweak +#endif + + ldr $rounds, [$key, #240] @ get # of rounds + mov $fp, r6 +#ifndef BSAES_ASM_EXTENDED_KEY + @ allocate the key schedule on the stack + sub r12, sp, $rounds, lsl#7 @ 128 bytes per inner round key + @ add r12, #`128-32` @ size of bit-sliced key schedule + sub r12, #`32+16` @ place for tweak[9] + + @ populate the key schedule + mov r4, $key @ pass key + mov r5, $rounds @ pass # of rounds + mov sp, r12 + add r12, #0x90 @ pass key schedule + bl _bsaes_key_convert + add r4, sp, #0x90 + vldmia r4, {@XMM[6]} + vstmia r12, {@XMM[15]} @ save last round key + veor @XMM[7], @XMM[7], @XMM[6] @ fix up round 0 key + vstmia r4, {@XMM[7]} +#else + ldr r12, [$key, #244] + eors r12, #1 + beq 0f + + str r12, [$key, #244] + mov r4, $key @ pass key + mov r5, $rounds @ pass # of rounds + add r12, $key, #248 @ pass key schedule + bl _bsaes_key_convert + add r4, $key, #248 + vldmia r4, {@XMM[6]} + vstmia r12, {@XMM[15]} @ save last round key + veor @XMM[7], @XMM[7], @XMM[6] @ fix up round 0 key + vstmia r4, {@XMM[7]} + +.align 2 +0: sub sp, #0x90 @ place for tweak[9] +#endif + vld1.8 {@XMM[8]}, [r0] @ initial tweak + adr $magic, .Lxts_magic + + tst $len, #0xf @ if not multiple of 16 + it ne @ Thumb2 thing, sanity check in ARM + subne $len, #0x10 @ subtract another 16 bytes + subs $len, #0x80 + + blo .Lxts_dec_short + b .Lxts_dec_loop + +.align 4 +.Lxts_dec_loop: + vldmia $magic, {$twmask} @ load XTS magic + vshr.s64 @T[0], @XMM[8], #63 + mov r0, sp + vand @T[0], @T[0], $twmask +___ +for($i=9;$i<16;$i++) { +$code.=<<___; + vadd.u64 @XMM[$i], @XMM[$i-1], @XMM[$i-1] + vst1.64 {@XMM[$i-1]}, [r0,:128]! + vswp `&Dhi("@T[0]")`,`&Dlo("@T[0]")` + vshr.s64 @T[1], @XMM[$i], #63 + veor @XMM[$i], @XMM[$i], @T[0] + vand @T[1], @T[1], $twmask +___ + @T=reverse(@T); + +$code.=<<___ if ($i>=10); + vld1.8 {@XMM[$i-10]}, [$inp]! +___ +$code.=<<___ if ($i>=11); + veor @XMM[$i-11], @XMM[$i-11], @XMM[$i-3] +___ +} +$code.=<<___; + vadd.u64 @XMM[8], @XMM[15], @XMM[15] + vst1.64 {@XMM[15]}, [r0,:128]! + vswp `&Dhi("@T[0]")`,`&Dlo("@T[0]")` + veor @XMM[8], @XMM[8], @T[0] + vst1.64 {@XMM[8]}, [r0,:128] @ next round tweak + + vld1.8 {@XMM[6]-@XMM[7]}, [$inp]! + veor @XMM[5], @XMM[5], @XMM[13] +#ifndef BSAES_ASM_EXTENDED_KEY + add r4, sp, #0x90 @ pass key schedule +#else + add r4, $key, #248 @ pass key schedule +#endif + veor @XMM[6], @XMM[6], @XMM[14] + mov r5, $rounds @ pass rounds + veor @XMM[7], @XMM[7], @XMM[15] + mov r0, sp + + bl _bsaes_decrypt8 + + vld1.64 {@XMM[ 8]-@XMM[ 9]}, [r0,:128]! + vld1.64 {@XMM[10]-@XMM[11]}, [r0,:128]! + veor @XMM[0], @XMM[0], @XMM[ 8] + vld1.64 {@XMM[12]-@XMM[13]}, [r0,:128]! + veor @XMM[1], @XMM[1], @XMM[ 9] + veor @XMM[8], @XMM[6], @XMM[10] + vst1.8 {@XMM[0]-@XMM[1]}, [$out]! + veor @XMM[9], @XMM[4], @XMM[11] + vld1.64 {@XMM[14]-@XMM[15]}, [r0,:128]! + veor @XMM[10], @XMM[2], @XMM[12] + vst1.8 {@XMM[8]-@XMM[9]}, [$out]! + veor @XMM[11], @XMM[7], @XMM[13] + veor @XMM[12], @XMM[3], @XMM[14] + vst1.8 {@XMM[10]-@XMM[11]}, [$out]! + veor @XMM[13], @XMM[5], @XMM[15] + vst1.8 {@XMM[12]-@XMM[13]}, [$out]! + + vld1.64 {@XMM[8]}, [r0,:128] @ next round tweak + + subs $len, #0x80 + bpl .Lxts_dec_loop + +.Lxts_dec_short: + adds $len, #0x70 + bmi .Lxts_dec_done + + vldmia $magic, {$twmask} @ load XTS magic + vshr.s64 @T[0], @XMM[8], #63 + mov r0, sp + vand @T[0], @T[0], $twmask +___ +for($i=9;$i<16;$i++) { +$code.=<<___; + vadd.u64 @XMM[$i], @XMM[$i-1], @XMM[$i-1] + vst1.64 {@XMM[$i-1]}, [r0,:128]! + vswp `&Dhi("@T[0]")`,`&Dlo("@T[0]")` + vshr.s64 @T[1], @XMM[$i], #63 + veor @XMM[$i], @XMM[$i], @T[0] + vand @T[1], @T[1], $twmask +___ + @T=reverse(@T); + +$code.=<<___ if ($i>=10); + vld1.8 {@XMM[$i-10]}, [$inp]! + subs $len, #0x10 + bmi .Lxts_dec_`$i-9` +___ +$code.=<<___ if ($i>=11); + veor @XMM[$i-11], @XMM[$i-11], @XMM[$i-3] +___ +} +$code.=<<___; + sub $len, #0x10 + vst1.64 {@XMM[15]}, [r0,:128] @ next round tweak + + vld1.8 {@XMM[6]}, [$inp]! + veor @XMM[5], @XMM[5], @XMM[13] +#ifndef BSAES_ASM_EXTENDED_KEY + add r4, sp, #0x90 @ pass key schedule +#else + add r4, $key, #248 @ pass key schedule +#endif + veor @XMM[6], @XMM[6], @XMM[14] + mov r5, $rounds @ pass rounds + mov r0, sp + + bl _bsaes_decrypt8 + + vld1.64 {@XMM[ 8]-@XMM[ 9]}, [r0,:128]! + vld1.64 {@XMM[10]-@XMM[11]}, [r0,:128]! + veor @XMM[0], @XMM[0], @XMM[ 8] + vld1.64 {@XMM[12]-@XMM[13]}, [r0,:128]! + veor @XMM[1], @XMM[1], @XMM[ 9] + veor @XMM[8], @XMM[6], @XMM[10] + vst1.8 {@XMM[0]-@XMM[1]}, [$out]! + veor @XMM[9], @XMM[4], @XMM[11] + vld1.64 {@XMM[14]}, [r0,:128]! + veor @XMM[10], @XMM[2], @XMM[12] + vst1.8 {@XMM[8]-@XMM[9]}, [$out]! + veor @XMM[11], @XMM[7], @XMM[13] + veor @XMM[12], @XMM[3], @XMM[14] + vst1.8 {@XMM[10]-@XMM[11]}, [$out]! + vst1.8 {@XMM[12]}, [$out]! + + vld1.64 {@XMM[8]}, [r0,:128] @ next round tweak + b .Lxts_dec_done +.align 4 +.Lxts_dec_6: + vst1.64 {@XMM[14]}, [r0,:128] @ next round tweak + + veor @XMM[4], @XMM[4], @XMM[12] +#ifndef BSAES_ASM_EXTENDED_KEY + add r4, sp, #0x90 @ pass key schedule +#else + add r4, $key, #248 @ pass key schedule +#endif + veor @XMM[5], @XMM[5], @XMM[13] + mov r5, $rounds @ pass rounds + mov r0, sp + + bl _bsaes_decrypt8 + + vld1.64 {@XMM[ 8]-@XMM[ 9]}, [r0,:128]! + vld1.64 {@XMM[10]-@XMM[11]}, [r0,:128]! + veor @XMM[0], @XMM[0], @XMM[ 8] + vld1.64 {@XMM[12]-@XMM[13]}, [r0,:128]! + veor @XMM[1], @XMM[1], @XMM[ 9] + veor @XMM[8], @XMM[6], @XMM[10] + vst1.8 {@XMM[0]-@XMM[1]}, [$out]! + veor @XMM[9], @XMM[4], @XMM[11] + veor @XMM[10], @XMM[2], @XMM[12] + vst1.8 {@XMM[8]-@XMM[9]}, [$out]! + veor @XMM[11], @XMM[7], @XMM[13] + vst1.8 {@XMM[10]-@XMM[11]}, [$out]! + + vld1.64 {@XMM[8]}, [r0,:128] @ next round tweak + b .Lxts_dec_done +.align 4 +.Lxts_dec_5: + vst1.64 {@XMM[13]}, [r0,:128] @ next round tweak + + veor @XMM[3], @XMM[3], @XMM[11] +#ifndef BSAES_ASM_EXTENDED_KEY + add r4, sp, #0x90 @ pass key schedule +#else + add r4, $key, #248 @ pass key schedule +#endif + veor @XMM[4], @XMM[4], @XMM[12] + mov r5, $rounds @ pass rounds + mov r0, sp + + bl _bsaes_decrypt8 + + vld1.64 {@XMM[ 8]-@XMM[ 9]}, [r0,:128]! + vld1.64 {@XMM[10]-@XMM[11]}, [r0,:128]! + veor @XMM[0], @XMM[0], @XMM[ 8] + vld1.64 {@XMM[12]}, [r0,:128]! + veor @XMM[1], @XMM[1], @XMM[ 9] + veor @XMM[8], @XMM[6], @XMM[10] + vst1.8 {@XMM[0]-@XMM[1]}, [$out]! + veor @XMM[9], @XMM[4], @XMM[11] + veor @XMM[10], @XMM[2], @XMM[12] + vst1.8 {@XMM[8]-@XMM[9]}, [$out]! + vst1.8 {@XMM[10]}, [$out]! + + vld1.64 {@XMM[8]}, [r0,:128] @ next round tweak + b .Lxts_dec_done +.align 4 +.Lxts_dec_4: + vst1.64 {@XMM[12]}, [r0,:128] @ next round tweak + + veor @XMM[2], @XMM[2], @XMM[10] +#ifndef BSAES_ASM_EXTENDED_KEY + add r4, sp, #0x90 @ pass key schedule +#else + add r4, $key, #248 @ pass key schedule +#endif + veor @XMM[3], @XMM[3], @XMM[11] + mov r5, $rounds @ pass rounds + mov r0, sp + + bl _bsaes_decrypt8 + + vld1.64 {@XMM[ 8]-@XMM[ 9]}, [r0,:128]! + vld1.64 {@XMM[10]-@XMM[11]}, [r0,:128]! + veor @XMM[0], @XMM[0], @XMM[ 8] + veor @XMM[1], @XMM[1], @XMM[ 9] + veor @XMM[8], @XMM[6], @XMM[10] + vst1.8 {@XMM[0]-@XMM[1]}, [$out]! + veor @XMM[9], @XMM[4], @XMM[11] + vst1.8 {@XMM[8]-@XMM[9]}, [$out]! + + vld1.64 {@XMM[8]}, [r0,:128] @ next round tweak + b .Lxts_dec_done +.align 4 +.Lxts_dec_3: + vst1.64 {@XMM[11]}, [r0,:128] @ next round tweak + + veor @XMM[1], @XMM[1], @XMM[9] +#ifndef BSAES_ASM_EXTENDED_KEY + add r4, sp, #0x90 @ pass key schedule +#else + add r4, $key, #248 @ pass key schedule +#endif + veor @XMM[2], @XMM[2], @XMM[10] + mov r5, $rounds @ pass rounds + mov r0, sp + + bl _bsaes_decrypt8 + + vld1.64 {@XMM[8]-@XMM[9]}, [r0,:128]! + vld1.64 {@XMM[10]}, [r0,:128]! + veor @XMM[0], @XMM[0], @XMM[ 8] + veor @XMM[1], @XMM[1], @XMM[ 9] + veor @XMM[8], @XMM[6], @XMM[10] + vst1.8 {@XMM[0]-@XMM[1]}, [$out]! + vst1.8 {@XMM[8]}, [$out]! + + vld1.64 {@XMM[8]}, [r0,:128] @ next round tweak + b .Lxts_dec_done +.align 4 +.Lxts_dec_2: + vst1.64 {@XMM[10]}, [r0,:128] @ next round tweak + + veor @XMM[0], @XMM[0], @XMM[8] +#ifndef BSAES_ASM_EXTENDED_KEY + add r4, sp, #0x90 @ pass key schedule +#else + add r4, $key, #248 @ pass key schedule +#endif + veor @XMM[1], @XMM[1], @XMM[9] + mov r5, $rounds @ pass rounds + mov r0, sp + + bl _bsaes_decrypt8 + + vld1.64 {@XMM[8]-@XMM[9]}, [r0,:128]! + veor @XMM[0], @XMM[0], @XMM[ 8] + veor @XMM[1], @XMM[1], @XMM[ 9] + vst1.8 {@XMM[0]-@XMM[1]}, [$out]! + + vld1.64 {@XMM[8]}, [r0,:128] @ next round tweak + b .Lxts_dec_done +.align 4 +.Lxts_dec_1: + mov r0, sp + veor @XMM[0], @XMM[8] + mov r1, sp + vst1.8 {@XMM[0]}, [sp,:128] + mov r2, $key + mov r4, $fp @ preserve fp + mov r5, $magic @ preserve magic + + bl AES_decrypt + + vld1.8 {@XMM[0]}, [sp,:128] + veor @XMM[0], @XMM[0], @XMM[8] + vst1.8 {@XMM[0]}, [$out]! + mov $fp, r4 + mov $magic, r5 + + vmov @XMM[8], @XMM[9] @ next round tweak + +.Lxts_dec_done: +#ifndef XTS_CHAIN_TWEAK + adds $len, #0x10 + beq .Lxts_dec_ret + + @ calculate one round of extra tweak for the stolen ciphertext + vldmia $magic, {$twmask} + vshr.s64 @XMM[6], @XMM[8], #63 + vand @XMM[6], @XMM[6], $twmask + vadd.u64 @XMM[9], @XMM[8], @XMM[8] + vswp `&Dhi("@XMM[6]")`,`&Dlo("@XMM[6]")` + veor @XMM[9], @XMM[9], @XMM[6] + + @ perform the final decryption with the last tweak value + vld1.8 {@XMM[0]}, [$inp]! + mov r0, sp + veor @XMM[0], @XMM[0], @XMM[9] + mov r1, sp + vst1.8 {@XMM[0]}, [sp,:128] + mov r2, $key + mov r4, $fp @ preserve fp + + bl AES_decrypt + + vld1.8 {@XMM[0]}, [sp,:128] + veor @XMM[0], @XMM[0], @XMM[9] + vst1.8 {@XMM[0]}, [$out] + + mov r6, $out +.Lxts_dec_steal: + ldrb r1, [$out] + ldrb r0, [$inp], #1 + strb r1, [$out, #0x10] + strb r0, [$out], #1 + + subs $len, #1 + bhi .Lxts_dec_steal + + vld1.8 {@XMM[0]}, [r6] + mov r0, sp + veor @XMM[0], @XMM[8] + mov r1, sp + vst1.8 {@XMM[0]}, [sp,:128] + mov r2, $key + + bl AES_decrypt + + vld1.8 {@XMM[0]}, [sp,:128] + veor @XMM[0], @XMM[0], @XMM[8] + vst1.8 {@XMM[0]}, [r6] + mov $fp, r4 +#endif + +.Lxts_dec_ret: + bic r0, $fp, #0xf + vmov.i32 q0, #0 + vmov.i32 q1, #0 +#ifdef XTS_CHAIN_TWEAK + ldr r1, [$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/sha1-armv7-neon.S b/arch/arm/crypto/sha1-armv7-neon.S new file mode 100644 index 000000000000..50013c0e2864 --- /dev/null +++ b/arch/arm/crypto/sha1-armv7-neon.S @@ -0,0 +1,634 @@ +/* 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> + + +.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 q1 +#define W2 q2 +#define W3 q3 +#define W4 q4 +#define W5 q5 +#define W6 q6 +#define W7 q7 + +#define tmp0 q8 +#define tmp1 q9 +#define tmp2 q10 +#define tmp3 q11 + +#define qK1 q12 +#define qK2 q13 +#define qK3 q14 +#define qK4 q15 + + +/* 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 {tmp0, tmp1}, [RDATA]!; \ + vrev32.8 W0, tmp0; /* big => little */ \ + vld1.32 {tmp2, tmp3}, [RDATA]!; \ + vadd.u32 tmp0, W0, curK; \ + vrev32.8 W7, tmp1; /* big => little */ \ + vrev32.8 W6, tmp2; /* big => little */ \ + vadd.u32 tmp1, W7, curK; \ + vrev32.8 W5, tmp3; /* 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 {tmp0, tmp1}, [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) \ + vrev32.8 W0, tmp0; /* 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 {tmp2, tmp3}, [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) \ + vrev32.8 W7, tmp1; /* big => little */ \ + +#define WPRECALC_00_15_6(i,W,W_m04,W_m08,W_m12,W_m16,W_m20,W_m24,W_m28) \ + vrev32.8 W6, tmp2; /* 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) \ + vrev32.8 W5, tmp3; /* 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_glue.c b/arch/arm/crypto/sha1_glue.c index ace4cd67464c..e31b0440c613 100644 --- a/arch/arm/crypto/sha1_glue.c +++ b/arch/arm/crypto/sha1_glue.c @@ -23,32 +23,27 @@ #include <linux/types.h> #include <crypto/sha.h> #include <asm/byteorder.h> +#include <asm/crypto/sha1.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, +asmlinkage void sha1_block_data_order(u32 *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; + struct sha1_state *sctx = shash_desc_ctx(desc); + + *sctx = (struct sha1_state){ + .state = { SHA1_H0, SHA1_H1, SHA1_H2, SHA1_H3, SHA1_H4 }, + }; + return 0; } -static int __sha1_update(struct SHA1_CTX *sctx, const u8 *data, - unsigned int len, unsigned int partial) +static int __sha1_update(struct sha1_state *sctx, const u8 *data, + unsigned int len, unsigned int partial) { unsigned int done = 0; @@ -56,43 +51,44 @@ static int __sha1_update(struct SHA1_CTX *sctx, const u8 *data, if (partial) { done = SHA1_BLOCK_SIZE - partial; - memcpy(sctx->data + partial, data, done); - sha1_block_data_order(sctx, sctx->data, 1); + memcpy(sctx->buffer + partial, data, done); + sha1_block_data_order(sctx->state, sctx->buffer, 1); } if (len - done >= SHA1_BLOCK_SIZE) { const unsigned int rounds = (len - done) / SHA1_BLOCK_SIZE; - sha1_block_data_order(sctx, data + done, rounds); + sha1_block_data_order(sctx->state, data + done, rounds); done += rounds * SHA1_BLOCK_SIZE; } - memcpy(sctx->data, data + done, len - done); + memcpy(sctx->buffer, data + done, len - done); return 0; } -static int sha1_update(struct shash_desc *desc, const u8 *data, - unsigned int len) +int sha1_update_arm(struct shash_desc *desc, const u8 *data, + unsigned int len) { - struct SHA1_CTX *sctx = shash_desc_ctx(desc); + struct sha1_state *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); + memcpy(sctx->buffer + partial, data, len); return 0; } res = __sha1_update(sctx, data, len, partial); return res; } +EXPORT_SYMBOL_GPL(sha1_update_arm); /* 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); + struct sha1_state *sctx = shash_desc_ctx(desc); unsigned int i, index, padlen; __be32 *dst = (__be32 *)out; __be64 bits; @@ -106,7 +102,7 @@ static int sha1_final(struct shash_desc *desc, u8 *out) /* We need to fill a whole block for __sha1_update() */ if (padlen <= 56) { sctx->count += padlen; - memcpy(sctx->data + index, padding, padlen); + memcpy(sctx->buffer + index, padding, padlen); } else { __sha1_update(sctx, padding, padlen, index); } @@ -114,7 +110,7 @@ static int sha1_final(struct shash_desc *desc, u8 *out) /* Store state in digest */ for (i = 0; i < 5; i++) - dst[i] = cpu_to_be32(((u32 *)sctx)[i]); + dst[i] = cpu_to_be32(sctx->state[i]); /* Wipe context */ memset(sctx, 0, sizeof(*sctx)); @@ -124,7 +120,7 @@ static int sha1_final(struct shash_desc *desc, u8 *out) static int sha1_export(struct shash_desc *desc, void *out) { - struct SHA1_CTX *sctx = shash_desc_ctx(desc); + struct sha1_state *sctx = shash_desc_ctx(desc); memcpy(out, sctx, sizeof(*sctx)); return 0; } @@ -132,7 +128,7 @@ static int sha1_export(struct shash_desc *desc, void *out) static int sha1_import(struct shash_desc *desc, const void *in) { - struct SHA1_CTX *sctx = shash_desc_ctx(desc); + struct sha1_state *sctx = shash_desc_ctx(desc); memcpy(sctx, in, sizeof(*sctx)); return 0; } @@ -141,12 +137,12 @@ static int sha1_import(struct shash_desc *desc, const void *in) static struct shash_alg alg = { .digestsize = SHA1_DIGEST_SIZE, .init = sha1_init, - .update = sha1_update, + .update = sha1_update_arm, .final = sha1_final, .export = sha1_export, .import = sha1_import, - .descsize = sizeof(struct SHA1_CTX), - .statesize = sizeof(struct SHA1_CTX), + .descsize = sizeof(struct sha1_state), + .statesize = sizeof(struct sha1_state), .base = { .cra_name = "sha1", .cra_driver_name= "sha1-asm", diff --git a/arch/arm/crypto/sha1_neon_glue.c b/arch/arm/crypto/sha1_neon_glue.c new file mode 100644 index 000000000000..6f1b411b1d55 --- /dev/null +++ b/arch/arm/crypto/sha1_neon_glue.c @@ -0,0 +1,197 @@ +/* + * 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 <asm/byteorder.h> +#include <asm/neon.h> +#include <asm/simd.h> +#include <asm/crypto/sha1.h> + + +asmlinkage void sha1_transform_neon(void *state_h, const char *data, + unsigned int rounds); + + +static int sha1_neon_init(struct shash_desc *desc) +{ + struct sha1_state *sctx = shash_desc_ctx(desc); + + *sctx = (struct sha1_state){ + .state = { SHA1_H0, SHA1_H1, SHA1_H2, SHA1_H3, SHA1_H4 }, + }; + + return 0; +} + +static int __sha1_neon_update(struct shash_desc *desc, const u8 *data, + unsigned int len, unsigned int partial) +{ + struct sha1_state *sctx = shash_desc_ctx(desc); + unsigned int done = 0; + + sctx->count += len; + + if (partial) { + done = SHA1_BLOCK_SIZE - partial; + memcpy(sctx->buffer + partial, data, done); + sha1_transform_neon(sctx->state, sctx->buffer, 1); + } + + if (len - done >= SHA1_BLOCK_SIZE) { + const unsigned int rounds = (len - done) / SHA1_BLOCK_SIZE; + + sha1_transform_neon(sctx->state, data + done, rounds); + done += rounds * SHA1_BLOCK_SIZE; + } + + memcpy(sctx->buffer, data + done, len - done); + + return 0; +} + +static int sha1_neon_update(struct shash_desc *desc, const u8 *data, + unsigned int len) +{ + struct sha1_state *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->buffer + partial, data, len); + + return 0; + } + + if (!may_use_simd()) { + res = sha1_update_arm(desc, data, len); + } else { + kernel_neon_begin(); + res = __sha1_neon_update(desc, data, len, partial); + kernel_neon_end(); + } + + return res; +} + + +/* Add padding and return the message digest. */ +static int sha1_neon_final(struct shash_desc *desc, u8 *out) +{ + struct sha1_state *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); + if (!may_use_simd()) { + sha1_update_arm(desc, padding, padlen); + sha1_update_arm(desc, (const u8 *)&bits, sizeof(bits)); + } else { + kernel_neon_begin(); + /* We need to fill a whole block for __sha1_neon_update() */ + if (padlen <= 56) { + sctx->count += padlen; + memcpy(sctx->buffer + index, padding, padlen); + } else { + __sha1_neon_update(desc, padding, padlen, index); + } + __sha1_neon_update(desc, (const u8 *)&bits, sizeof(bits), 56); + kernel_neon_end(); + } + + /* Store state in digest */ + for (i = 0; i < 5; i++) + dst[i] = cpu_to_be32(sctx->state[i]); + + /* Wipe context */ + memset(sctx, 0, sizeof(*sctx)); + + return 0; +} + +static int sha1_neon_export(struct shash_desc *desc, void *out) +{ + struct sha1_state *sctx = shash_desc_ctx(desc); + + memcpy(out, sctx, sizeof(*sctx)); + + return 0; +} + +static int sha1_neon_import(struct shash_desc *desc, const void *in) +{ + struct sha1_state *sctx = shash_desc_ctx(desc); + + memcpy(sctx, in, sizeof(*sctx)); + + return 0; +} + +static struct shash_alg alg = { + .digestsize = SHA1_DIGEST_SIZE, + .init = sha1_neon_init, + .update = sha1_neon_update, + .final = sha1_neon_final, + .export = sha1_neon_export, + .import = sha1_neon_import, + .descsize = sizeof(struct sha1_state), + .statesize = 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("sha1"); diff --git a/arch/arm/crypto/sha256-armv4.pl b/arch/arm/crypto/sha256-armv4.pl new file mode 100644 index 000000000000..fac0533ea633 --- /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 000000000000..555a1a8eec90 --- /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 000000000000..bb03482e53e2 --- /dev/null +++ b/arch/arm/crypto/sha256_glue.c @@ -0,0 +1,246 @@ +/* + * 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 <asm/byteorder.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 sha256_init(struct shash_desc *desc) +{ + struct sha256_state *sctx = shash_desc_ctx(desc); + + sctx->state[0] = SHA256_H0; + sctx->state[1] = SHA256_H1; + sctx->state[2] = SHA256_H2; + sctx->state[3] = SHA256_H3; + sctx->state[4] = SHA256_H4; + sctx->state[5] = SHA256_H5; + sctx->state[6] = SHA256_H6; + sctx->state[7] = SHA256_H7; + sctx->count = 0; + + return 0; +} + +int sha224_init(struct shash_desc *desc) +{ + struct sha256_state *sctx = shash_desc_ctx(desc); + + sctx->state[0] = SHA224_H0; + sctx->state[1] = SHA224_H1; + sctx->state[2] = SHA224_H2; + sctx->state[3] = SHA224_H3; + sctx->state[4] = SHA224_H4; + sctx->state[5] = SHA224_H5; + sctx->state[6] = SHA224_H6; + sctx->state[7] = SHA224_H7; + sctx->count = 0; + + return 0; +} + +int __sha256_update(struct shash_desc *desc, const u8 *data, unsigned int len, + unsigned int partial) +{ + struct sha256_state *sctx = shash_desc_ctx(desc); + unsigned int done = 0; + + sctx->count += len; + + if (partial) { + done = SHA256_BLOCK_SIZE - partial; + memcpy(sctx->buf + partial, data, done); + sha256_block_data_order(sctx->state, sctx->buf, 1); + } + + if (len - done >= SHA256_BLOCK_SIZE) { + const unsigned int rounds = (len - done) / SHA256_BLOCK_SIZE; + + sha256_block_data_order(sctx->state, data + done, rounds); + done += rounds * SHA256_BLOCK_SIZE; + } + + memcpy(sctx->buf, data + done, len - done); + + return 0; +} + +int sha256_update(struct shash_desc *desc, const u8 *data, unsigned int len) +{ + struct sha256_state *sctx = shash_desc_ctx(desc); + unsigned int partial = sctx->count % SHA256_BLOCK_SIZE; + + /* Handle the fast case right here */ + if (partial + len < SHA256_BLOCK_SIZE) { + sctx->count += len; + memcpy(sctx->buf + partial, data, len); + + return 0; + } + + return __sha256_update(desc, data, len, partial); +} + +/* Add padding and return the message digest. */ +static int sha256_final(struct shash_desc *desc, u8 *out) +{ + struct sha256_state *sctx = shash_desc_ctx(desc); + unsigned int i, index, padlen; + __be32 *dst = (__be32 *)out; + __be64 bits; + static const u8 padding[SHA256_BLOCK_SIZE] = { 0x80, }; + + /* save number of bits */ + bits = cpu_to_be64(sctx->count << 3); + + /* Pad out to 56 mod 64 and append length */ + index = sctx->count % SHA256_BLOCK_SIZE; + padlen = (index < 56) ? (56 - index) : ((SHA256_BLOCK_SIZE+56)-index); + + /* We need to fill a whole block for __sha256_update */ + if (padlen <= 56) { + sctx->count += padlen; + memcpy(sctx->buf + index, padding, padlen); + } else { + __sha256_update(desc, padding, padlen, index); + } + __sha256_update(desc, (const u8 *)&bits, sizeof(bits), 56); + + /* Store state in digest */ + for (i = 0; i < 8; i++) + dst[i] = cpu_to_be32(sctx->state[i]); + + /* Wipe context */ + memset(sctx, 0, sizeof(*sctx)); + + return 0; +} + +static int sha224_final(struct shash_desc *desc, u8 *out) +{ + u8 D[SHA256_DIGEST_SIZE]; + + sha256_final(desc, D); + + memcpy(out, D, SHA224_DIGEST_SIZE); + memset(D, 0, SHA256_DIGEST_SIZE); + + return 0; +} + +int sha256_export(struct shash_desc *desc, void *out) +{ + struct sha256_state *sctx = shash_desc_ctx(desc); + + memcpy(out, sctx, sizeof(*sctx)); + + return 0; +} + +int sha256_import(struct shash_desc *desc, const void *in) +{ + struct sha256_state *sctx = shash_desc_ctx(desc); + + memcpy(sctx, in, sizeof(*sctx)); + + return 0; +} + +static struct shash_alg algs[] = { { + .digestsize = SHA256_DIGEST_SIZE, + .init = sha256_init, + .update = sha256_update, + .final = sha256_final, + .export = sha256_export, + .import = sha256_import, + .descsize = sizeof(struct sha256_state), + .statesize = 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_init, + .update = sha256_update, + .final = sha224_final, + .export = sha256_export, + .import = sha256_import, + .descsize = sizeof(struct sha256_state), + .statesize = 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 000000000000..0312f4ffe8cc --- /dev/null +++ b/arch/arm/crypto/sha256_glue.h @@ -0,0 +1,23 @@ +#ifndef _CRYPTO_SHA256_GLUE_H +#define _CRYPTO_SHA256_GLUE_H + +#include <linux/crypto.h> +#include <crypto/sha.h> + +extern struct shash_alg sha256_neon_algs[2]; + +extern int sha256_init(struct shash_desc *desc); + +extern int sha224_init(struct shash_desc *desc); + +extern int __sha256_update(struct shash_desc *desc, const u8 *data, + unsigned int len, unsigned int partial); + +extern int sha256_update(struct shash_desc *desc, const u8 *data, + unsigned int len); + +extern int sha256_export(struct shash_desc *desc, void *out); + +extern int sha256_import(struct shash_desc *desc, const void *in); + +#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 000000000000..3ff0a7f1d092 --- /dev/null +++ b/arch/arm/crypto/sha256_neon_glue.c @@ -0,0 +1,172 @@ +/* + * 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 <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_neon_update(struct shash_desc *desc, const u8 *data, + unsigned int len, unsigned int partial) +{ + struct sha256_state *sctx = shash_desc_ctx(desc); + unsigned int done = 0; + + sctx->count += len; + + if (partial) { + done = SHA256_BLOCK_SIZE - partial; + memcpy(sctx->buf + partial, data, done); + sha256_block_data_order_neon(sctx->state, sctx->buf, 1); + } + + if (len - done >= SHA256_BLOCK_SIZE) { + const unsigned int rounds = (len - done) / SHA256_BLOCK_SIZE; + + sha256_block_data_order_neon(sctx->state, data + done, rounds); + done += rounds * SHA256_BLOCK_SIZE; + } + + memcpy(sctx->buf, data + done, len - done); + + return 0; +} + +static int sha256_neon_update(struct shash_desc *desc, const u8 *data, + unsigned int len) +{ + struct sha256_state *sctx = shash_desc_ctx(desc); + unsigned int partial = sctx->count % SHA256_BLOCK_SIZE; + int res; + + /* Handle the fast case right here */ + if (partial + len < SHA256_BLOCK_SIZE) { + sctx->count += len; + memcpy(sctx->buf + partial, data, len); + + return 0; + } + + if (!may_use_simd()) { + res = __sha256_update(desc, data, len, partial); + } else { + kernel_neon_begin(); + res = __sha256_neon_update(desc, data, len, partial); + kernel_neon_end(); + } + + return res; +} + +/* Add padding and return the message digest. */ +static int sha256_neon_final(struct shash_desc *desc, u8 *out) +{ + struct sha256_state *sctx = shash_desc_ctx(desc); + unsigned int i, index, padlen; + __be32 *dst = (__be32 *)out; + __be64 bits; + static const u8 padding[SHA256_BLOCK_SIZE] = { 0x80, }; + + /* save number of bits */ + bits = cpu_to_be64(sctx->count << 3); + + /* Pad out to 56 mod 64 and append length */ + index = sctx->count % SHA256_BLOCK_SIZE; + padlen = (index < 56) ? (56 - index) : ((SHA256_BLOCK_SIZE+56)-index); + + if (!may_use_simd()) { + sha256_update(desc, padding, padlen); + sha256_update(desc, (const u8 *)&bits, sizeof(bits)); + } else { + kernel_neon_begin(); + /* We need to fill a whole block for __sha256_neon_update() */ + if (padlen <= 56) { + sctx->count += padlen; + memcpy(sctx->buf + index, padding, padlen); + } else { + __sha256_neon_update(desc, padding, padlen, index); + } + __sha256_neon_update(desc, (const u8 *)&bits, + sizeof(bits), 56); + kernel_neon_end(); + } + + /* Store state in digest */ + for (i = 0; i < 8; i++) + dst[i] = cpu_to_be32(sctx->state[i]); + + /* Wipe context */ + memset(sctx, 0, sizeof(*sctx)); + + return 0; +} + +static int sha224_neon_final(struct shash_desc *desc, u8 *out) +{ + u8 D[SHA256_DIGEST_SIZE]; + + sha256_neon_final(desc, D); + + memcpy(out, D, SHA224_DIGEST_SIZE); + memset(D, 0, SHA256_DIGEST_SIZE); + + return 0; +} + +struct shash_alg sha256_neon_algs[] = { { + .digestsize = SHA256_DIGEST_SIZE, + .init = sha256_init, + .update = sha256_neon_update, + .final = sha256_neon_final, + .export = sha256_export, + .import = sha256_import, + .descsize = sizeof(struct sha256_state), + .statesize = 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_init, + .update = sha256_neon_update, + .final = sha224_neon_final, + .export = sha256_export, + .import = sha256_import, + .descsize = sizeof(struct sha256_state), + .statesize = 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 000000000000..fe99472e507c --- /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 000000000000..0d2758ff5e12 --- /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); + memset(D, 0, 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("sha512"); +MODULE_ALIAS("sha384"); diff --git a/arch/arm/include/asm/Kbuild b/arch/arm/include/asm/Kbuild index d3db39860b9c..6577b8aeb711 100644 --- a/arch/arm/include/asm/Kbuild +++ b/arch/arm/include/asm/Kbuild @@ -24,6 +24,7 @@ generic-y += sembuf.h generic-y += serial.h generic-y += shmbuf.h generic-y += siginfo.h +generic-y += simd.h generic-y += sizes.h generic-y += socket.h generic-y += sockios.h diff --git a/arch/arm/include/asm/barrier.h b/arch/arm/include/asm/barrier.h index 60f15e274e6d..2f59f7443396 100644 --- a/arch/arm/include/asm/barrier.h +++ b/arch/arm/include/asm/barrier.h @@ -59,6 +59,21 @@ #define smp_wmb() dmb(ishst) #endif +#define smp_store_release(p, v) \ +do { \ + compiletime_assert_atomic_type(*p); \ + smp_mb(); \ + ACCESS_ONCE(*p) = (v); \ +} while (0) + +#define smp_load_acquire(p) \ +({ \ + typeof(*p) ___p1 = ACCESS_ONCE(*p); \ + compiletime_assert_atomic_type(*p); \ + smp_mb(); \ + ___p1; \ +}) + #define read_barrier_depends() do { } while(0) #define smp_read_barrier_depends() do { } while(0) diff --git a/arch/arm/include/asm/cacheflush.h b/arch/arm/include/asm/cacheflush.h index 2059f019bef4..455e6637c881 100644 --- a/arch/arm/include/asm/cacheflush.h +++ b/arch/arm/include/asm/cacheflush.h @@ -16,6 +16,7 @@ #include <asm/shmparam.h> #include <asm/cachetype.h> #include <asm/outercache.h> +#include <asm/rodata.h> #define CACHE_COLOUR(vaddr) ((vaddr & (SHMLBA - 1)) >> PAGE_SHIFT) diff --git a/arch/arm/include/asm/crypto/sha1.h b/arch/arm/include/asm/crypto/sha1.h new file mode 100644 index 000000000000..75e6a417416b --- /dev/null +++ b/arch/arm/include/asm/crypto/sha1.h @@ -0,0 +1,10 @@ +#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); + +#endif diff --git a/arch/arm/include/asm/fiq_glue.h b/arch/arm/include/asm/fiq_glue.h new file mode 100644 index 000000000000..a9e244f9f197 --- /dev/null +++ b/arch/arm/include/asm/fiq_glue.h @@ -0,0 +1,33 @@ +/* + * Copyright (C) 2010 Google, Inc. + * + * This software is licensed under the terms of the GNU General Public + * License version 2, as published by the Free Software Foundation, and + * may be copied, distributed, and modified under those terms. + * + * This program is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + * GNU General Public License for more details. + */ + +#ifndef __ASM_FIQ_GLUE_H +#define __ASM_FIQ_GLUE_H + +struct fiq_glue_handler { + void (*fiq)(struct fiq_glue_handler *h, void *regs, void *svc_sp); + void (*resume)(struct fiq_glue_handler *h); +}; +typedef void (*fiq_return_handler_t)(void); + +int fiq_glue_register_handler(struct fiq_glue_handler *handler); +int fiq_glue_set_return_handler(fiq_return_handler_t fiq_return); +int fiq_glue_clear_return_handler(fiq_return_handler_t fiq_return); + +#ifdef CONFIG_FIQ_GLUE +void fiq_glue_resume(void); +#else +static inline void fiq_glue_resume(void) {} +#endif + +#endif diff --git a/arch/arm/include/asm/hardirq.h b/arch/arm/include/asm/hardirq.h index 3d7351c844aa..fe3ea776dc34 100644 --- a/arch/arm/include/asm/hardirq.h +++ b/arch/arm/include/asm/hardirq.h @@ -5,7 +5,7 @@ #include <linux/threads.h> #include <asm/irq.h> -#define NR_IPI 7 +#define NR_IPI 8 typedef struct { unsigned int __softirq_pending; diff --git a/arch/arm/include/asm/hardware/cache-l2x0.h b/arch/arm/include/asm/hardware/cache-l2x0.h index 3b2c40b5bfa2..0ca0f5a7c84b 100644 --- a/arch/arm/include/asm/hardware/cache-l2x0.h +++ b/arch/arm/include/asm/hardware/cache-l2x0.h @@ -66,6 +66,7 @@ #define L2X0_STNDBY_MODE_EN (1 << 0) /* Registers shifts and masks */ +#define L2X0_CACHE_ID_REV_MASK (0x3f) #define L2X0_CACHE_ID_PART_MASK (0xf << 6) #define L2X0_CACHE_ID_PART_L210 (1 << 6) #define L2X0_CACHE_ID_PART_L310 (3 << 6) @@ -106,6 +107,8 @@ #define L2X0_WAY_SIZE_SHIFT 3 +#define REV_PL310_R2P0 4 + #ifndef __ASSEMBLY__ extern void __init l2x0_init(void __iomem *base, u32 aux_val, u32 aux_mask); #if defined(CONFIG_CACHE_L2X0) && defined(CONFIG_OF) diff --git a/arch/arm/include/asm/irq.h b/arch/arm/include/asm/irq.h index 53c15dec7af6..809203a4b71b 100644 --- a/arch/arm/include/asm/irq.h +++ b/arch/arm/include/asm/irq.h @@ -35,6 +35,9 @@ extern void (*handle_arch_irq)(struct pt_regs *); extern void set_handle_irq(void (*handle_irq)(struct pt_regs *)); #endif +void arch_trigger_all_cpu_backtrace(void); +#define arch_trigger_all_cpu_backtrace arch_trigger_all_cpu_backtrace + #endif #endif diff --git a/arch/arm/include/asm/mach/mmc.h b/arch/arm/include/asm/mach/mmc.h new file mode 100644 index 000000000000..bca864ac945f --- /dev/null +++ b/arch/arm/include/asm/mach/mmc.h @@ -0,0 +1,28 @@ +/* + * arch/arm/include/asm/mach/mmc.h + */ +#ifndef ASMARM_MACH_MMC_H +#define ASMARM_MACH_MMC_H + +#include <linux/mmc/host.h> +#include <linux/mmc/card.h> +#include <linux/mmc/sdio_func.h> + +struct embedded_sdio_data { + struct sdio_cis cis; + struct sdio_cccr cccr; + struct sdio_embedded_func *funcs; + int num_funcs; +}; + +struct mmc_platform_data { + unsigned int ocr_mask; /* available voltages */ + int built_in; /* built-in device flag */ + int card_present; /* card detect state */ + u32 (*translate_vdd)(struct device *, unsigned int); + unsigned int (*status)(struct device *); + struct embedded_sdio_data *embedded_sdio; + int (*register_status_notify)(void (*callback)(int card_present, void *dev_id), void *dev_id); +}; + +#endif diff --git a/arch/arm/include/asm/neon.h b/arch/arm/include/asm/neon.h new file mode 100644 index 000000000000..8f730fe70093 --- /dev/null +++ b/arch/arm/include/asm/neon.h @@ -0,0 +1,36 @@ +/* + * linux/arch/arm/include/asm/neon.h + * + * 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/hwcap.h> + +#define cpu_has_neon() (!!(elf_hwcap & HWCAP_NEON)) + +#ifdef __ARM_NEON__ + +/* + * If you are affected by the BUILD_BUG below, it probably means that you are + * using NEON code /and/ calling the kernel_neon_begin() function from the same + * compilation unit. To prevent issues that may arise from GCC reordering or + * generating(1) NEON instructions outside of these begin/end functions, the + * only supported way of using NEON code in the kernel is by isolating it in a + * separate compilation unit, and calling it from another unit from inside a + * kernel_neon_begin/kernel_neon_end pair. + * + * (1) Current GCC (4.7) might generate NEON instructions at O3 level if + * -mpfu=neon is set. + */ + +#define kernel_neon_begin() \ + BUILD_BUG_ON_MSG(1, "kernel_neon_begin() called from NEON code") + +#else +void kernel_neon_begin(void); +#endif +void kernel_neon_end(void); diff --git a/arch/arm/include/asm/rodata.h b/arch/arm/include/asm/rodata.h new file mode 100644 index 000000000000..8c8add87bbc5 --- /dev/null +++ b/arch/arm/include/asm/rodata.h @@ -0,0 +1,32 @@ +/* + * arch/arm/include/asm/rodata.h + * + * Copyright (C) 2011 Google, Inc. + * + * Author: Colin Cross <ccross@android.com> + * + * 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. + */ +#ifndef _ASMARM_RODATA_H +#define _ASMARM_RODATA_H + +#ifndef __ASSEMBLY__ + +#ifdef CONFIG_DEBUG_RODATA + +int set_memory_rw(unsigned long virt, int numpages); +int set_memory_ro(unsigned long virt, int numpages); + +void mark_rodata_ro(void); +void set_kernel_text_rw(void); +void set_kernel_text_ro(void); +#else +static inline void set_kernel_text_rw(void) { } +static inline void set_kernel_text_ro(void) { } +#endif + +#endif + +#endif diff --git a/arch/arm/include/asm/smp.h b/arch/arm/include/asm/smp.h index 610ccf33f5e7..67a18a5ed9fa 100644 --- a/arch/arm/include/asm/smp.h +++ b/arch/arm/include/asm/smp.h @@ -82,6 +82,7 @@ extern void arch_send_call_function_ipi_mask(const struct cpumask *mask); extern void arch_send_wakeup_ipi_mask(const struct cpumask *mask); extern int register_ipi_completion(struct completion *completion, int cpu); +extern void smp_send_all_cpu_backtrace(void); struct smp_operations { #ifdef CONFIG_SMP diff --git a/arch/arm/include/asm/syscall.h b/arch/arm/include/asm/syscall.h index 73ddd7239b33..ed805f1d3785 100644 --- a/arch/arm/include/asm/syscall.h +++ b/arch/arm/include/asm/syscall.h @@ -103,8 +103,7 @@ static inline void syscall_set_arguments(struct task_struct *task, memcpy(®s->ARM_r0 + i, args, n * sizeof(args[0])); } -static inline int syscall_get_arch(struct task_struct *task, - struct pt_regs *regs) +static inline int syscall_get_arch(void) { /* ARM tasks don't change audit architectures on the fly. */ return AUDIT_ARCH_ARM; diff --git a/arch/arm/include/asm/unistd.h b/arch/arm/include/asm/unistd.h index cbd61977c996..43876245fc57 100644 --- a/arch/arm/include/asm/unistd.h +++ b/arch/arm/include/asm/unistd.h @@ -15,7 +15,7 @@ #include <uapi/asm/unistd.h> -#define __NR_syscalls (380) +#define __NR_syscalls (384) #define __ARM_NR_cmpxchg (__ARM_NR_BASE+0x00fff0) #define __ARCH_WANT_STAT64 diff --git a/arch/arm/include/uapi/asm/unistd.h b/arch/arm/include/uapi/asm/unistd.h index af33b44990ed..17407c92c0da 100644 --- a/arch/arm/include/uapi/asm/unistd.h +++ b/arch/arm/include/uapi/asm/unistd.h @@ -406,6 +406,12 @@ #define __NR_process_vm_writev (__NR_SYSCALL_BASE+377) #define __NR_kcmp (__NR_SYSCALL_BASE+378) #define __NR_finit_module (__NR_SYSCALL_BASE+379) +/* Reserve for later +#define __NR_sched_setattr (__NR_SYSCALL_BASE+380) +#define __NR_sched_getattr (__NR_SYSCALL_BASE+381) +#define __NR_renameat2 (__NR_SYSCALL_BASE+382) +*/ +#define __NR_seccomp (__NR_SYSCALL_BASE+383) /* * This may need to be greater than __NR_last_syscall+1 in order to diff --git a/arch/arm/kernel/calls.S b/arch/arm/kernel/calls.S index c6ca7e376773..725f844926ea 100644 --- a/arch/arm/kernel/calls.S +++ b/arch/arm/kernel/calls.S @@ -389,6 +389,11 @@ CALL(sys_process_vm_writev) CALL(sys_kcmp) CALL(sys_finit_module) +/* 380 */ CALL(sys_ni_syscall) /* reserved sys_sched_setattr */ + CALL(sys_ni_syscall) /* reserved sys_sched_getattr */ + CALL(sys_ni_syscall) /* reserved sys_renameat2 */ + CALL(sys_seccomp) + #ifndef syscalls_counted .equ syscalls_padding, ((NR_syscalls + 3) & ~3) - NR_syscalls #define syscalls_counted diff --git a/arch/arm/kernel/ftrace.c b/arch/arm/kernel/ftrace.c index 34e56647dcee..6a740a93f4bb 100644 --- a/arch/arm/kernel/ftrace.c +++ b/arch/arm/kernel/ftrace.c @@ -13,6 +13,7 @@ */ #include <linux/ftrace.h> +#include <linux/module.h> #include <linux/uaccess.h> #include <asm/cacheflush.h> @@ -63,6 +64,20 @@ static unsigned long adjust_address(struct dyn_ftrace *rec, unsigned long addr) } #endif +int ftrace_arch_code_modify_prepare(void) +{ + set_kernel_text_rw(); + set_all_modules_text_rw(); + return 0; +} + +int ftrace_arch_code_modify_post_process(void) +{ + set_all_modules_text_ro(); + set_kernel_text_ro(); + return 0; +} + static unsigned long ftrace_call_replace(unsigned long pc, unsigned long addr) { return arm_gen_branch_link(pc, addr); diff --git a/arch/arm/kernel/kgdb.c b/arch/arm/kernel/kgdb.c index 778c2f7024ff..b321c8fbb87d 100644 --- a/arch/arm/kernel/kgdb.c +++ b/arch/arm/kernel/kgdb.c @@ -144,6 +144,8 @@ int kgdb_arch_handle_exception(int exception_vector, int signo, static int kgdb_brk_fn(struct pt_regs *regs, unsigned int instr) { + if (user_mode(regs)) + return -1; kgdb_handle_exception(1, SIGTRAP, 0, regs); return 0; @@ -151,6 +153,8 @@ static int kgdb_brk_fn(struct pt_regs *regs, unsigned int instr) static int kgdb_compiled_brk_fn(struct pt_regs *regs, unsigned int instr) { + if (user_mode(regs)) + return -1; compiled_break = 1; kgdb_handle_exception(1, SIGTRAP, 0, regs); diff --git a/arch/arm/kernel/process.c b/arch/arm/kernel/process.c index ac4c2e5e17e4..978002e5b406 100644 --- a/arch/arm/kernel/process.c +++ b/arch/arm/kernel/process.c @@ -32,6 +32,7 @@ #include <linux/hw_breakpoint.h> #include <linux/cpuidle.h> #include <linux/leds.h> +#include <linux/console.h> #include <asm/cacheflush.h> #include <asm/idmap.h> @@ -57,9 +58,46 @@ static const char *isa_modes[] = { "ARM" , "Thumb" , "Jazelle", "ThumbEE" }; +#ifdef CONFIG_SMP +void arch_trigger_all_cpu_backtrace(void) +{ + smp_send_all_cpu_backtrace(); +} +#else +void arch_trigger_all_cpu_backtrace(void) +{ + dump_stack(); +} +#endif + extern void call_with_stack(void (*fn)(void *), void *arg, void *sp); typedef void (*phys_reset_t)(unsigned long); +#ifdef CONFIG_ARM_FLUSH_CONSOLE_ON_RESTART +void arm_machine_flush_console(void) +{ + printk("\n"); + pr_emerg("Restarting %s\n", linux_banner); + if (console_trylock()) { + console_unlock(); + return; + } + + mdelay(50); + + local_irq_disable(); + if (!console_trylock()) + pr_emerg("arm_restart: Console was locked! Busting\n"); + else + pr_emerg("arm_restart: Console was locked!\n"); + console_unlock(); +} +#else +void arm_machine_flush_console(void) +{ +} +#endif + /* * A temporary stack to use for CPU reset. This is static so that we * don't clobber it with the identity mapping. When running with this @@ -147,6 +185,7 @@ void arch_cpu_idle_prepare(void) void arch_cpu_idle_enter(void) { + idle_notifier_call_chain(IDLE_START); ledtrig_cpu(CPU_LED_IDLE_START); #ifdef CONFIG_PL310_ERRATA_769419 wmb(); @@ -156,6 +195,7 @@ void arch_cpu_idle_enter(void) void arch_cpu_idle_exit(void) { ledtrig_cpu(CPU_LED_IDLE_END); + idle_notifier_call_chain(IDLE_END); } #ifdef CONFIG_HOTPLUG_CPU @@ -195,6 +235,16 @@ __setup("reboot=", reboot_setup); */ void machine_shutdown(void) { +#ifdef CONFIG_SMP + /* + * Disable preemption so we're guaranteed to + * run to power off or reboot and prevent + * the possibility of switching to another + * thread that might wind up blocking on + * one of the stopped CPUs. + */ + preempt_disable(); +#endif disable_nonboot_cpus(); } @@ -240,6 +290,10 @@ void machine_restart(char *cmd) { smp_send_stop(); + /* Flush the console to make sure all the relevant messages make it + * out to the console drivers */ + arm_machine_flush_console(); + arm_pm_restart(reboot_mode, cmd); /* Give a grace period for failure to restart of 1s */ @@ -251,6 +305,77 @@ void machine_restart(char *cmd) while (1); } +/* + * dump a block of kernel memory from around the given address + */ +static void show_data(unsigned long addr, int nbytes, const char *name) +{ + int i, j; + int nlines; + u32 *p; + + /* + * don't attempt to dump non-kernel addresses or + * values that are probably just small negative numbers + */ + if (addr < PAGE_OFFSET || addr > -256UL) + return; + + printk("\n%s: %#lx:\n", name, addr); + + /* + * round address down to a 32 bit boundary + * and always dump a multiple of 32 bytes + */ + p = (u32 *)(addr & ~(sizeof(u32) - 1)); + nbytes += (addr & (sizeof(u32) - 1)); + nlines = (nbytes + 31) / 32; + + + for (i = 0; i < nlines; i++) { + /* + * just display low 16 bits of address to keep + * each line of the dump < 80 characters + */ + printk("%04lx ", (unsigned long)p & 0xffff); + for (j = 0; j < 8; j++) { + u32 data; + if (probe_kernel_address(p, data)) { + printk(" ********"); + } else { + printk(" %08x", data); + } + ++p; + } + printk("\n"); + } +} + +static void show_extra_register_data(struct pt_regs *regs, int nbytes) +{ + mm_segment_t fs; + + fs = get_fs(); + set_fs(KERNEL_DS); + show_data(regs->ARM_pc - nbytes, nbytes * 2, "PC"); + show_data(regs->ARM_lr - nbytes, nbytes * 2, "LR"); + show_data(regs->ARM_sp - nbytes, nbytes * 2, "SP"); + show_data(regs->ARM_ip - nbytes, nbytes * 2, "IP"); + show_data(regs->ARM_fp - nbytes, nbytes * 2, "FP"); + show_data(regs->ARM_r0 - nbytes, nbytes * 2, "R0"); + show_data(regs->ARM_r1 - nbytes, nbytes * 2, "R1"); + show_data(regs->ARM_r2 - nbytes, nbytes * 2, "R2"); + show_data(regs->ARM_r3 - nbytes, nbytes * 2, "R3"); + show_data(regs->ARM_r4 - nbytes, nbytes * 2, "R4"); + show_data(regs->ARM_r5 - nbytes, nbytes * 2, "R5"); + show_data(regs->ARM_r6 - nbytes, nbytes * 2, "R6"); + show_data(regs->ARM_r7 - nbytes, nbytes * 2, "R7"); + show_data(regs->ARM_r8 - nbytes, nbytes * 2, "R8"); + show_data(regs->ARM_r9 - nbytes, nbytes * 2, "R9"); + show_data(regs->ARM_r10 - nbytes, nbytes * 2, "R10"); + set_fs(fs); +} + void __show_regs(struct pt_regs *regs) { unsigned long flags; @@ -307,6 +432,8 @@ void __show_regs(struct pt_regs *regs) printk("Control: %08x%s\n", ctrl, buf); } #endif + + show_extra_register_data(regs, 128); } void show_regs(struct pt_regs * regs) diff --git a/arch/arm/kernel/ptrace.c b/arch/arm/kernel/ptrace.c index 03deeffd9f6d..394424b25254 100644 --- a/arch/arm/kernel/ptrace.c +++ b/arch/arm/kernel/ptrace.c @@ -916,7 +916,7 @@ enum ptrace_syscall_dir { PTRACE_SYSCALL_EXIT, }; -static int tracehook_report_syscall(struct pt_regs *regs, +static void tracehook_report_syscall(struct pt_regs *regs, enum ptrace_syscall_dir dir) { unsigned long ip; @@ -934,7 +934,6 @@ static int tracehook_report_syscall(struct pt_regs *regs, current_thread_info()->syscall = -1; regs->ARM_ip = ip; - return current_thread_info()->syscall; } asmlinkage int syscall_trace_enter(struct pt_regs *regs, int scno) @@ -946,7 +945,9 @@ asmlinkage int syscall_trace_enter(struct pt_regs *regs, int scno) return -1; if (test_thread_flag(TIF_SYSCALL_TRACE)) - scno = tracehook_report_syscall(regs, PTRACE_SYSCALL_ENTER); + tracehook_report_syscall(regs, PTRACE_SYSCALL_ENTER); + + scno = current_thread_info()->syscall; if (test_thread_flag(TIF_SYSCALL_TRACEPOINT)) trace_sys_enter(regs, scno); diff --git a/arch/arm/kernel/smp.c b/arch/arm/kernel/smp.c index d6e3d1ca4ddf..f2724e475b96 100644 --- a/arch/arm/kernel/smp.c +++ b/arch/arm/kernel/smp.c @@ -70,6 +70,7 @@ enum ipi_msg_type { IPI_CALL_FUNC_SINGLE, IPI_CPU_STOP, IPI_COMPLETION, + IPI_CPU_BACKTRACE, }; static DECLARE_COMPLETION(cpu_running); @@ -468,6 +469,7 @@ static const char *ipi_types[NR_IPI] = { S(IPI_CALL_FUNC_SINGLE, "Single function call interrupts"), S(IPI_CPU_STOP, "CPU stop interrupts"), S(IPI_COMPLETION, "completion interrupts"), + S(IPI_CPU_BACKTRACE, "CPU backtrace"), }; void show_ipi_list(struct seq_file *p, int prec) @@ -606,6 +608,58 @@ static void ipi_complete(unsigned int cpu) complete(per_cpu(cpu_completion, cpu)); } +static cpumask_t backtrace_mask; +static DEFINE_RAW_SPINLOCK(backtrace_lock); + +/* "in progress" flag of arch_trigger_all_cpu_backtrace */ +static unsigned long backtrace_flag; + +void smp_send_all_cpu_backtrace(void) +{ + unsigned int this_cpu = smp_processor_id(); + int i; + + if (test_and_set_bit(0, &backtrace_flag)) + /* + * If there is already a trigger_all_cpu_backtrace() in progress + * (backtrace_flag == 1), don't output double cpu dump infos. + */ + return; + + cpumask_copy(&backtrace_mask, cpu_online_mask); + cpu_clear(this_cpu, backtrace_mask); + + pr_info("Backtrace for cpu %d (current):\n", this_cpu); + dump_stack(); + + pr_info("\nsending IPI to all other CPUs:\n"); + smp_cross_call(&backtrace_mask, IPI_CPU_BACKTRACE); + + /* Wait for up to 10 seconds for all other CPUs to do the backtrace */ + for (i = 0; i < 10 * 1000; i++) { + if (cpumask_empty(&backtrace_mask)) + break; + mdelay(1); + } + + clear_bit(0, &backtrace_flag); + smp_mb__after_clear_bit(); +} + +/* + * ipi_cpu_backtrace - handle IPI from smp_send_all_cpu_backtrace() + */ +static void ipi_cpu_backtrace(unsigned int cpu, struct pt_regs *regs) +{ + if (cpu_isset(cpu, backtrace_mask)) { + raw_spin_lock(&backtrace_lock); + pr_warning("IPI backtrace for cpu %d\n", cpu); + show_regs(regs); + raw_spin_unlock(&backtrace_lock); + cpu_clear(cpu, backtrace_mask); + } +} + /* * Main handler for inter-processor interrupts */ @@ -663,6 +717,10 @@ void handle_IPI(int ipinr, struct pt_regs *regs) irq_exit(); break; + case IPI_CPU_BACKTRACE: + ipi_cpu_backtrace(cpu, regs); + break; + default: printk(KERN_CRIT "CPU%u: Unknown IPI message 0x%x\n", cpu, ipinr); diff --git a/arch/arm/mm/Makefile b/arch/arm/mm/Makefile index 224a9cc09877..2d861a210ac5 100644 --- a/arch/arm/mm/Makefile +++ b/arch/arm/mm/Makefile @@ -7,6 +7,7 @@ obj-y := dma-mapping.o extable.o fault.o init.o \ obj-$(CONFIG_MMU) += fault-armv.o flush.o idmap.o ioremap.o \ mmap.o pgd.o mmu.o +obj-$(CONFIG_DEBUG_RODATA) += rodata.o ifneq ($(CONFIG_MMU),y) obj-y += nommu.o diff --git a/arch/arm/mm/cache-l2x0.c b/arch/arm/mm/cache-l2x0.c index c465faca51b0..90a130f98acf 100644 --- a/arch/arm/mm/cache-l2x0.c +++ b/arch/arm/mm/cache-l2x0.c @@ -33,6 +33,9 @@ static void __iomem *l2x0_base; static DEFINE_RAW_SPINLOCK(l2x0_lock); static u32 l2x0_way_mask; /* Bitmask of active ways */ static u32 l2x0_size; +static u32 l2x0_cache_id; +static unsigned int l2x0_sets; +static unsigned int l2x0_ways; static unsigned long sync_reg_offset = L2X0_CACHE_SYNC; /* Aurora don't have the cache ID register available, so we have to @@ -49,6 +52,13 @@ struct l2x0_of_data { static bool of_init = false; +static inline bool is_pl310_rev(int rev) +{ + return (l2x0_cache_id & + (L2X0_CACHE_ID_PART_MASK | L2X0_CACHE_ID_REV_MASK)) == + (L2X0_CACHE_ID_PART_L310 | rev); +} + static inline void cache_wait_way(void __iomem *reg, unsigned long mask) { /* wait for cache operation by line or way to complete */ @@ -137,6 +147,23 @@ static void l2x0_cache_sync(void) raw_spin_unlock_irqrestore(&l2x0_lock, flags); } +#ifdef CONFIG_PL310_ERRATA_727915 +static void l2x0_for_each_set_way(void __iomem *reg) +{ + int set; + int way; + unsigned long flags; + + for (way = 0; way < l2x0_ways; way++) { + raw_spin_lock_irqsave(&l2x0_lock, flags); + for (set = 0; set < l2x0_sets; set++) + writel_relaxed((way << 28) | (set << 5), reg); + cache_sync(); + raw_spin_unlock_irqrestore(&l2x0_lock, flags); + } +} +#endif + static void __l2x0_flush_all(void) { debug_writel(0x03); @@ -150,6 +177,13 @@ static void l2x0_flush_all(void) { unsigned long flags; +#ifdef CONFIG_PL310_ERRATA_727915 + if (is_pl310_rev(REV_PL310_R2P0)) { + l2x0_for_each_set_way(l2x0_base + L2X0_CLEAN_INV_LINE_IDX); + return; + } +#endif + /* clean all ways */ raw_spin_lock_irqsave(&l2x0_lock, flags); __l2x0_flush_all(); @@ -160,11 +194,20 @@ static void l2x0_clean_all(void) { unsigned long flags; +#ifdef CONFIG_PL310_ERRATA_727915 + if (is_pl310_rev(REV_PL310_R2P0)) { + l2x0_for_each_set_way(l2x0_base + L2X0_CLEAN_LINE_IDX); + return; + } +#endif + /* clean all ways */ raw_spin_lock_irqsave(&l2x0_lock, flags); + debug_writel(0x03); writel_relaxed(l2x0_way_mask, l2x0_base + L2X0_CLEAN_WAY); cache_wait_way(l2x0_base + L2X0_CLEAN_WAY, l2x0_way_mask); cache_sync(); + debug_writel(0x00); raw_spin_unlock_irqrestore(&l2x0_lock, flags); } @@ -323,65 +366,64 @@ static void l2x0_unlock(u32 cache_id) void __init l2x0_init(void __iomem *base, u32 aux_val, u32 aux_mask) { u32 aux; - u32 cache_id; u32 way_size = 0; - int ways; int way_size_shift = L2X0_WAY_SIZE_SHIFT; const char *type; l2x0_base = base; if (cache_id_part_number_from_dt) - cache_id = cache_id_part_number_from_dt; + l2x0_cache_id = cache_id_part_number_from_dt; else - cache_id = readl_relaxed(l2x0_base + L2X0_CACHE_ID); + l2x0_cache_id = readl_relaxed(l2x0_base + L2X0_CACHE_ID); aux = readl_relaxed(l2x0_base + L2X0_AUX_CTRL); aux &= aux_mask; aux |= aux_val; /* Determine the number of ways */ - switch (cache_id & L2X0_CACHE_ID_PART_MASK) { + switch (l2x0_cache_id & L2X0_CACHE_ID_PART_MASK) { case L2X0_CACHE_ID_PART_L310: if (aux & (1 << 16)) - ways = 16; + l2x0_ways = 16; else - ways = 8; + l2x0_ways = 8; type = "L310"; #ifdef CONFIG_PL310_ERRATA_753970 /* Unmapped register. */ sync_reg_offset = L2X0_DUMMY_REG; #endif - if ((cache_id & L2X0_CACHE_ID_RTL_MASK) <= L2X0_CACHE_ID_RTL_R3P0) + if ((l2x0_cache_id & L2X0_CACHE_ID_RTL_MASK) <= L2X0_CACHE_ID_RTL_R3P0) outer_cache.set_debug = pl310_set_debug; break; case L2X0_CACHE_ID_PART_L210: - ways = (aux >> 13) & 0xf; + l2x0_ways = (aux >> 13) & 0xf; type = "L210"; break; case AURORA_CACHE_ID: sync_reg_offset = AURORA_SYNC_REG; - ways = (aux >> 13) & 0xf; - ways = 2 << ((ways + 1) >> 2); + l2x0_ways = (aux >> 13) & 0xf; + l2x0_ways = 2 << ((l2x0_ways + 1) >> 2); way_size_shift = AURORA_WAY_SIZE_SHIFT; type = "Aurora"; break; default: /* Assume unknown chips have 8 ways */ - ways = 8; + l2x0_ways = 8; type = "L2x0 series"; break; } - l2x0_way_mask = (1 << ways) - 1; + l2x0_way_mask = (1 << l2x0_ways) - 1; /* * L2 cache Size = Way size * Number of ways */ way_size = (aux & L2X0_AUX_CTRL_WAY_SIZE_MASK) >> 17; - way_size = 1 << (way_size + way_size_shift); + way_size = SZ_1K << (way_size + way_size_shift); - l2x0_size = ways * way_size * SZ_1K; + l2x0_size = l2x0_ways * way_size; + l2x0_sets = way_size / CACHE_LINE_SIZE; /* * Check if l2x0 controller is already enabled. @@ -390,7 +432,7 @@ void __init l2x0_init(void __iomem *base, u32 aux_val, u32 aux_mask) */ if (!(readl_relaxed(l2x0_base + L2X0_CTRL) & L2X0_CTRL_EN)) { /* Make sure that I&D is not locked down when starting */ - l2x0_unlock(cache_id); + l2x0_unlock(l2x0_cache_id); /* l2x0 controller is disabled */ writel_relaxed(aux, l2x0_base + L2X0_AUX_CTRL); @@ -419,7 +461,7 @@ void __init l2x0_init(void __iomem *base, u32 aux_val, u32 aux_mask) printk(KERN_INFO "%s cache controller enabled\n", type); printk(KERN_INFO "l2x0: %d ways, CACHE_ID 0x%08x, AUX_CTRL 0x%08x, Cache size: %d B\n", - ways, cache_id, aux, l2x0_size); + l2x0_ways, l2x0_cache_id, aux, l2x0_size); } #ifdef CONFIG_OF diff --git a/arch/arm/mm/cache-v6.S b/arch/arm/mm/cache-v6.S index d8fd4d4bd3d4..7a3d3d8d98d7 100644 --- a/arch/arm/mm/cache-v6.S +++ b/arch/arm/mm/cache-v6.S @@ -270,6 +270,11 @@ v6_dma_clean_range: * - end - virtual end address of region */ ENTRY(v6_dma_flush_range) +#ifdef CONFIG_CACHE_FLUSH_RANGE_LIMIT + sub r2, r1, r0 + cmp r2, #CONFIG_CACHE_FLUSH_RANGE_LIMIT + bhi v6_dma_flush_dcache_all +#endif #ifdef CONFIG_DMA_CACHE_RWFO ldrb r2, [r0] @ read for ownership strb r2, [r0] @ write for ownership @@ -292,6 +297,18 @@ ENTRY(v6_dma_flush_range) mcr p15, 0, r0, c7, c10, 4 @ drain write buffer mov pc, lr +#ifdef CONFIG_CACHE_FLUSH_RANGE_LIMIT +v6_dma_flush_dcache_all: + mov r0, #0 +#ifdef HARVARD_CACHE + mcr p15, 0, r0, c7, c14, 0 @ D cache clean+invalidate +#else + mcr p15, 0, r0, c7, c15, 0 @ Cache clean+invalidate +#endif + mcr p15, 0, r0, c7, c10, 4 @ drain write buffer + mov pc, lr +#endif + /* * dma_map_area(start, size, dir) * - start - kernel virtual start address diff --git a/arch/arm/mm/fault.c b/arch/arm/mm/fault.c index 54fcddafec15..9820ad4b80c0 100644 --- a/arch/arm/mm/fault.c +++ b/arch/arm/mm/fault.c @@ -274,10 +274,10 @@ do_page_fault(unsigned long addr, unsigned int fsr, struct pt_regs *regs) local_irq_enable(); /* - * If we're in an interrupt or have no user + * If we're in an interrupt, or have no irqs, or have no user * context, we must not take the fault.. */ - if (in_atomic() || !mm) + if (in_atomic() || irqs_disabled() || !mm) goto no_context; if (user_mode(regs)) diff --git a/arch/arm/mm/mmu.c b/arch/arm/mm/mmu.c index 4c7d5cddef35..a4e5cd5c3067 100644 --- a/arch/arm/mm/mmu.c +++ b/arch/arm/mm/mmu.c @@ -605,11 +605,25 @@ static void __init *early_alloc(unsigned long sz) return early_alloc_aligned(sz, sz); } -static pte_t * __init early_pte_alloc(pmd_t *pmd, unsigned long addr, unsigned long prot) +static pte_t * __init early_pte_alloc(pmd_t *pmd) +{ + if (pmd_none(*pmd) || pmd_bad(*pmd)) + return early_alloc(PTE_HWTABLE_OFF + PTE_HWTABLE_SIZE); + return pmd_page_vaddr(*pmd); +} + +static void __init early_pte_install(pmd_t *pmd, pte_t *pte, unsigned long prot) +{ + __pmd_populate(pmd, __pa(pte), prot); + BUG_ON(pmd_bad(*pmd)); +} + +static pte_t * __init early_pte_alloc_and_install(pmd_t *pmd, + unsigned long addr, unsigned long prot) { if (pmd_none(*pmd)) { - pte_t *pte = early_alloc(PTE_HWTABLE_OFF + PTE_HWTABLE_SIZE); - __pmd_populate(pmd, __pa(pte), prot); + pte_t *pte = early_pte_alloc(pmd); + early_pte_install(pmd, pte, prot); } BUG_ON(pmd_bad(*pmd)); return pte_offset_kernel(pmd, addr); @@ -619,11 +633,17 @@ static void __init alloc_init_pte(pmd_t *pmd, unsigned long addr, unsigned long end, unsigned long pfn, const struct mem_type *type) { - pte_t *pte = early_pte_alloc(pmd, addr, type->prot_l1); + pte_t *start_pte = early_pte_alloc(pmd); + pte_t *pte = start_pte + pte_index(addr); + + /* If replacing a section mapping, the whole section must be replaced */ + BUG_ON(!pmd_none(*pmd) && pmd_bad(*pmd) && ((addr | end) & ~PMD_MASK)); + do { set_pte_ext(pte, pfn_pte(pfn, __pgprot(type->prot_pte)), 0); pfn++; } while (pte++, addr += PAGE_SIZE, addr != end); + early_pte_install(pmd, start_pte, type->prot_l1); } static void __init __map_init_section(pmd_t *pmd, unsigned long addr, @@ -655,7 +675,8 @@ static void __init __map_init_section(pmd_t *pmd, unsigned long addr, static void __init alloc_init_pmd(pud_t *pud, unsigned long addr, unsigned long end, phys_addr_t phys, - const struct mem_type *type) + const struct mem_type *type, + bool force_pages) { pmd_t *pmd = pmd_offset(pud, addr); unsigned long next; @@ -672,7 +693,8 @@ static void __init alloc_init_pmd(pud_t *pud, unsigned long addr, * aligned to a section boundary. */ if (type->prot_sect && - ((addr | next | phys) & ~SECTION_MASK) == 0) { + ((addr | next | phys) & ~SECTION_MASK) == 0 && + !force_pages) { __map_init_section(pmd, addr, next, phys, type); } else { alloc_init_pte(pmd, addr, next, @@ -686,14 +708,15 @@ static void __init alloc_init_pmd(pud_t *pud, unsigned long addr, static void __init alloc_init_pud(pgd_t *pgd, unsigned long addr, unsigned long end, phys_addr_t phys, - const struct mem_type *type) + const struct mem_type *type, + bool force_pages) { pud_t *pud = pud_offset(pgd, addr); unsigned long next; do { next = pud_addr_end(addr, end); - alloc_init_pmd(pud, addr, next, phys, type); + alloc_init_pmd(pud, addr, next, phys, type, force_pages); phys += next - addr; } while (pud++, addr = next, addr != end); } @@ -767,7 +790,7 @@ static void __init create_36bit_mapping(struct map_desc *md, * offsets, and we take full advantage of sections and * supersections. */ -static void __init create_mapping(struct map_desc *md) +static void __init create_mapping(struct map_desc *md, bool force_pages) { unsigned long addr, length, end; phys_addr_t phys; @@ -817,7 +840,7 @@ static void __init create_mapping(struct map_desc *md) do { unsigned long next = pgd_addr_end(addr, end); - alloc_init_pud(pgd, addr, next, phys, type); + alloc_init_pud(pgd, addr, next, phys, type, force_pages); phys += next - addr; addr = next; @@ -839,7 +862,7 @@ void __init iotable_init(struct map_desc *io_desc, int nr) svm = early_alloc_aligned(sizeof(*svm) * nr, __alignof__(*svm)); for (md = io_desc; nr; md++, nr--) { - create_mapping(md); + create_mapping(md, false); vm = &svm->vm; vm->addr = (void *)(md->virtual & PAGE_MASK); @@ -960,7 +983,7 @@ void __init debug_ll_io_init(void) map.virtual &= PAGE_MASK; map.length = PAGE_SIZE; map.type = MT_DEVICE; - create_mapping(&map); + create_mapping(&map, false); } #endif @@ -1005,6 +1028,28 @@ void __init sanity_check_meminfo(void) struct membank *bank = &meminfo.bank[j]; *bank = meminfo.bank[i]; +#ifdef CONFIG_SPARSEMEM + if (pfn_to_section_nr(bank_pfn_start(bank)) != + pfn_to_section_nr(bank_pfn_end(bank) - 1)) { + phys_addr_t sz; + unsigned long start_pfn = bank_pfn_start(bank); + unsigned long end_pfn = SECTION_ALIGN_UP(start_pfn + 1); + sz = ((phys_addr_t)(end_pfn - start_pfn) << PAGE_SHIFT); + + if (meminfo.nr_banks >= NR_BANKS) { + pr_crit("NR_BANKS too low, ignoring %lld bytes of memory\n", + (unsigned long long)(bank->size - sz)); + } else { + memmove(bank + 1, bank, + (meminfo.nr_banks - i) * sizeof(*bank)); + meminfo.nr_banks++; + bank[1].size -= sz; + bank[1].start = __pfn_to_phys(end_pfn); + } + bank->size = sz; + } +#endif + if (bank->start > ULONG_MAX) highmem = 1; @@ -1202,7 +1247,7 @@ static void __init devicemaps_init(struct machine_desc *mdesc) map.virtual = MODULES_VADDR; map.length = ((unsigned long)_etext - map.virtual + ~SECTION_MASK) & SECTION_MASK; map.type = MT_ROM; - create_mapping(&map); + create_mapping(&map, false); #endif /* @@ -1213,14 +1258,14 @@ static void __init devicemaps_init(struct machine_desc *mdesc) map.virtual = FLUSH_BASE; map.length = SZ_1M; map.type = MT_CACHECLEAN; - create_mapping(&map); + create_mapping(&map, false); #endif #ifdef FLUSH_BASE_MINICACHE map.pfn = __phys_to_pfn(FLUSH_BASE_PHYS + SZ_1M); map.virtual = FLUSH_BASE_MINICACHE; map.length = SZ_1M; map.type = MT_MINICLEAN; - create_mapping(&map); + create_mapping(&map, false); #endif /* @@ -1236,13 +1281,13 @@ static void __init devicemaps_init(struct machine_desc *mdesc) #else map.type = MT_LOW_VECTORS; #endif - create_mapping(&map); + create_mapping(&map, false); if (!vectors_high()) { map.virtual = 0; map.length = PAGE_SIZE * 2; map.type = MT_LOW_VECTORS; - create_mapping(&map); + create_mapping(&map, false); } /* Now create a kernel read-only mapping */ @@ -1250,7 +1295,7 @@ static void __init devicemaps_init(struct machine_desc *mdesc) map.virtual = 0xffff0000 + PAGE_SIZE; map.length = PAGE_SIZE; map.type = MT_LOW_VECTORS; - create_mapping(&map); + create_mapping(&map, false); /* * Ask the machine support to map in the statically mapped devices. @@ -1275,20 +1320,23 @@ static void __init devicemaps_init(struct machine_desc *mdesc) static void __init kmap_init(void) { #ifdef CONFIG_HIGHMEM - pkmap_page_table = early_pte_alloc(pmd_off_k(PKMAP_BASE), + pkmap_page_table = early_pte_alloc_and_install(pmd_off_k(PKMAP_BASE), PKMAP_BASE, _PAGE_KERNEL_TABLE); #endif } + static void __init map_lowmem(void) { struct memblock_region *reg; + phys_addr_t start; + phys_addr_t end; + struct map_desc map; /* Map all the lowmem memory banks. */ for_each_memblock(memory, reg) { - phys_addr_t start = reg->base; - phys_addr_t end = start + reg->size; - struct map_desc map; + start = reg->base; + end = start + reg->size; if (end > arm_lowmem_limit) end = arm_lowmem_limit; @@ -1300,8 +1348,20 @@ static void __init map_lowmem(void) map.length = end - start; map.type = MT_MEMORY; - create_mapping(&map); + create_mapping(&map, false); } + +#ifdef CONFIG_DEBUG_RODATA + start = __pa(_stext) & PMD_MASK; + end = ALIGN(__pa(__end_rodata), PMD_SIZE); + + map.pfn = __phys_to_pfn(start); + map.virtual = __phys_to_virt(start); + map.length = end - start; + map.type = MT_MEMORY; + + create_mapping(&map, true); +#endif } /* diff --git a/arch/arm/mm/rodata.c b/arch/arm/mm/rodata.c new file mode 100644 index 000000000000..9a8eb841c428 --- /dev/null +++ b/arch/arm/mm/rodata.c @@ -0,0 +1,159 @@ +/* + * linux/arch/arm/mm/rodata.c + * + * Copyright (C) 2011 Google, Inc. + * + * Author: Colin Cross <ccross@android.com> + * + * Based on x86 implementation in arch/x86/mm/init_32.c + * + * 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/kernel.h> +#include <linux/mm.h> +#include <linux/module.h> + +#include <asm/cache.h> +#include <asm/pgtable.h> +#include <asm/rodata.h> +#include <asm/sections.h> +#include <asm/tlbflush.h> + +#include "mm.h" + +static int kernel_set_to_readonly __read_mostly; + +#ifdef CONFIG_DEBUG_RODATA_TEST +static const int rodata_test_data = 0xC3; + +static noinline void rodata_test(void) +{ + int result; + + pr_info("%s: attempting to write to read-only section:\n", __func__); + + if (*(volatile int *)&rodata_test_data != 0xC3) { + pr_err("read only data changed before test\n"); + return; + } + + /* + * Attempt to to write to rodata_test_data, trapping the expected + * data abort. If the trap executed, result will be 1. If it didn't, + * result will be 0xFF. + */ + asm volatile( + "0: str %[zero], [%[rodata_test_data]]\n" + " mov %[result], #0xFF\n" + " b 2f\n" + "1: mov %[result], #1\n" + "2:\n" + + /* Exception fixup - if store at label 0 faults, jumps to 1 */ + ".pushsection __ex_table, \"a\"\n" + " .long 0b, 1b\n" + ".popsection\n" + + : [result] "=r" (result) + : [rodata_test_data] "r" (&rodata_test_data), [zero] "r" (0) + : "memory" + ); + + if (result == 1) + pr_info("write to read-only section trapped, success\n"); + else + pr_err("write to read-only section NOT trapped, test failed\n"); + + if (*(volatile int *)&rodata_test_data != 0xC3) + pr_err("read only data changed during write\n"); +} +#else +static inline void rodata_test(void) { } +#endif + +static int set_page_attributes(unsigned long virt, int numpages, + pte_t (*f)(pte_t)) +{ + pmd_t *pmd; + pte_t *pte; + unsigned long start = virt; + unsigned long end = virt + (numpages << PAGE_SHIFT); + unsigned long pmd_end; + + while (virt < end) { + pmd = pmd_off_k(virt); + pmd_end = min(ALIGN(virt + 1, PMD_SIZE), end); + + if ((pmd_val(*pmd) & PMD_TYPE_MASK) != PMD_TYPE_TABLE) { + pr_err("%s: pmd %p=%08lx for %08lx not page table\n", + __func__, pmd, pmd_val(*pmd), virt); + virt = pmd_end; + continue; + } + + while (virt < pmd_end) { + pte = pte_offset_kernel(pmd, virt); + set_pte_ext(pte, f(*pte), 0); + virt += PAGE_SIZE; + } + } + + flush_tlb_kernel_range(start, end); + + return 0; +} + +int set_memory_ro(unsigned long virt, int numpages) +{ + return set_page_attributes(virt, numpages, pte_wrprotect); +} +EXPORT_SYMBOL(set_memory_ro); + +int set_memory_rw(unsigned long virt, int numpages) +{ + return set_page_attributes(virt, numpages, pte_mkwrite); +} +EXPORT_SYMBOL(set_memory_rw); + +void set_kernel_text_rw(void) +{ + unsigned long start = PAGE_ALIGN((unsigned long)_text); + unsigned long size = PAGE_ALIGN((unsigned long)__end_rodata) - start; + + if (!kernel_set_to_readonly) + return; + + pr_debug("Set kernel text: %lx - %lx to read-write\n", + start, start + size); + + set_memory_rw(start, size >> PAGE_SHIFT); +} + +void set_kernel_text_ro(void) +{ + unsigned long start = PAGE_ALIGN((unsigned long)_text); + unsigned long size = PAGE_ALIGN((unsigned long)__end_rodata) - start; + + if (!kernel_set_to_readonly) + return; + + pr_info_once("Write protecting the kernel text section %lx - %lx\n", + start, start + size); + + pr_debug("Set kernel text: %lx - %lx to read only\n", + start, start + size); + + set_memory_ro(start, size >> PAGE_SHIFT); +} + +void mark_rodata_ro(void) +{ + kernel_set_to_readonly = 1; + + set_kernel_text_ro(); + + rodata_test(); +} diff --git a/arch/arm/vfp/vfpmodule.c b/arch/arm/vfp/vfpmodule.c index 5dfbb0b8e7f4..452fb3ad68aa 100644 --- a/arch/arm/vfp/vfpmodule.c +++ b/arch/arm/vfp/vfpmodule.c @@ -20,6 +20,7 @@ #include <linux/init.h> #include <linux/uaccess.h> #include <linux/user.h> +#include <linux/export.h> #include <asm/cp15.h> #include <asm/cputype.h> @@ -648,6 +649,52 @@ static int vfp_hotplug(struct notifier_block *b, unsigned long action, return NOTIFY_OK; } +#ifdef CONFIG_KERNEL_MODE_NEON + +/* + * Kernel-side NEON support functions + */ +void kernel_neon_begin(void) +{ + struct thread_info *thread = current_thread_info(); + unsigned int cpu; + u32 fpexc; + + /* + * Kernel mode NEON is only allowed outside of interrupt context + * with preemption disabled. This will make sure that the kernel + * mode NEON register contents never need to be preserved. + */ + BUG_ON(in_interrupt()); + cpu = get_cpu(); + + fpexc = fmrx(FPEXC) | FPEXC_EN; + fmxr(FPEXC, fpexc); + + /* + * Save the userland NEON/VFP state. Under UP, + * the owner could be a task other than 'current' + */ + if (vfp_state_in_hw(cpu, thread)) + vfp_save_state(&thread->vfpstate, fpexc); +#ifndef CONFIG_SMP + else if (vfp_current_hw_state[cpu] != NULL) + vfp_save_state(vfp_current_hw_state[cpu], fpexc); +#endif + vfp_current_hw_state[cpu] = NULL; +} +EXPORT_SYMBOL(kernel_neon_begin); + +void kernel_neon_end(void) +{ + /* Disable the NEON/VFP unit. */ + fmxr(FPEXC, fmrx(FPEXC) & ~FPEXC_EN); + put_cpu(); +} +EXPORT_SYMBOL(kernel_neon_end); + +#endif /* CONFIG_KERNEL_MODE_NEON */ + /* * VFP support code initialisation. */ @@ -731,4 +778,4 @@ static int __init vfp_init(void) return 0; } -late_initcall(vfp_init); +core_initcall(vfp_init); |