Commit b36d8c09 authored by Ard Biesheuvel's avatar Ard Biesheuvel Committed by Herbert Xu

crypto: arm/chacha - remove dependency on generic ChaCha driver

Instead of falling back to the generic ChaCha skcipher driver for
non-SIMD cases, use a fast scalar implementation for ARM authored
by Eric Biggers. This removes the module dependency on chacha-generic
altogether, which also simplifies things when we expose the ChaCha
library interface from this module.
Signed-off-by: default avatarArd Biesheuvel <ardb@kernel.org>
Signed-off-by: default avatarHerbert Xu <herbert@gondor.apana.org.au>
parent 29621d09
...@@ -128,10 +128,8 @@ config CRYPTO_CRC32_ARM_CE ...@@ -128,10 +128,8 @@ config CRYPTO_CRC32_ARM_CE
select CRYPTO_HASH select CRYPTO_HASH
config CRYPTO_CHACHA20_NEON config CRYPTO_CHACHA20_NEON
tristate "NEON accelerated ChaCha stream cipher algorithms" tristate "NEON and scalar accelerated ChaCha stream cipher algorithms"
depends on KERNEL_MODE_NEON
select CRYPTO_SKCIPHER select CRYPTO_SKCIPHER
select CRYPTO_CHACHA20
config CRYPTO_NHPOLY1305_NEON config CRYPTO_NHPOLY1305_NEON
tristate "NEON accelerated NHPoly1305 hash function (for Adiantum)" tristate "NEON accelerated NHPoly1305 hash function (for Adiantum)"
......
...@@ -33,7 +33,8 @@ aes-arm-ce-y := aes-ce-core.o aes-ce-glue.o ...@@ -33,7 +33,8 @@ aes-arm-ce-y := aes-ce-core.o aes-ce-glue.o
ghash-arm-ce-y := ghash-ce-core.o ghash-ce-glue.o ghash-arm-ce-y := ghash-ce-core.o ghash-ce-glue.o
crct10dif-arm-ce-y := crct10dif-ce-core.o crct10dif-ce-glue.o crct10dif-arm-ce-y := crct10dif-ce-core.o crct10dif-ce-glue.o
crc32-arm-ce-y:= crc32-ce-core.o crc32-ce-glue.o crc32-arm-ce-y:= crc32-ce-core.o crc32-ce-glue.o
chacha-neon-y := chacha-neon-core.o chacha-neon-glue.o chacha-neon-y := chacha-scalar-core.o chacha-glue.o
chacha-neon-$(CONFIG_KERNEL_MODE_NEON) += chacha-neon-core.o
nhpoly1305-neon-y := nh-neon-core.o nhpoly1305-neon-glue.o nhpoly1305-neon-y := nh-neon-core.o nhpoly1305-neon-glue.o
ifdef REGENERATE_ARM_CRYPTO ifdef REGENERATE_ARM_CRYPTO
......
// SPDX-License-Identifier: GPL-2.0
/* /*
* ARM NEON accelerated ChaCha and XChaCha stream ciphers, * ARM NEON accelerated ChaCha and XChaCha stream ciphers,
* including ChaCha20 (RFC7539) * including ChaCha20 (RFC7539)
* *
* Copyright (C) 2016 Linaro, Ltd. <ard.biesheuvel@linaro.org> * Copyright (C) 2016-2019 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.
*
* Based on:
* ChaCha20 256-bit cipher algorithm, RFC7539, SIMD glue code
*
* Copyright (C) 2015 Martin Willi * Copyright (C) 2015 Martin Willi
*
* 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/algapi.h> #include <crypto/algapi.h>
...@@ -26,6 +14,7 @@ ...@@ -26,6 +14,7 @@
#include <linux/kernel.h> #include <linux/kernel.h>
#include <linux/module.h> #include <linux/module.h>
#include <asm/cputype.h>
#include <asm/hwcap.h> #include <asm/hwcap.h>
#include <asm/neon.h> #include <asm/neon.h>
#include <asm/simd.h> #include <asm/simd.h>
...@@ -34,8 +23,17 @@ asmlinkage void chacha_block_xor_neon(const u32 *state, u8 *dst, const u8 *src, ...@@ -34,8 +23,17 @@ asmlinkage void chacha_block_xor_neon(const u32 *state, u8 *dst, const u8 *src,
int nrounds); int nrounds);
asmlinkage void chacha_4block_xor_neon(const u32 *state, u8 *dst, const u8 *src, asmlinkage void chacha_4block_xor_neon(const u32 *state, u8 *dst, const u8 *src,
int nrounds); int nrounds);
asmlinkage void hchacha_block_arm(const u32 *state, u32 *out, int nrounds);
asmlinkage void hchacha_block_neon(const u32 *state, u32 *out, int nrounds); asmlinkage void hchacha_block_neon(const u32 *state, u32 *out, int nrounds);
asmlinkage void chacha_doarm(u8 *dst, const u8 *src, unsigned int bytes,
const u32 *state, int nrounds);
static inline bool neon_usable(void)
{
return crypto_simd_usable();
}
static void chacha_doneon(u32 *state, u8 *dst, const u8 *src, static void chacha_doneon(u32 *state, u8 *dst, const u8 *src,
unsigned int bytes, int nrounds) unsigned int bytes, int nrounds)
{ {
...@@ -62,8 +60,9 @@ static void chacha_doneon(u32 *state, u8 *dst, const u8 *src, ...@@ -62,8 +60,9 @@ static void chacha_doneon(u32 *state, u8 *dst, const u8 *src,
} }
} }
static int chacha_neon_stream_xor(struct skcipher_request *req, static int chacha_stream_xor(struct skcipher_request *req,
const struct chacha_ctx *ctx, const u8 *iv) const struct chacha_ctx *ctx, const u8 *iv,
bool neon)
{ {
struct skcipher_walk walk; struct skcipher_walk walk;
u32 state[16]; u32 state[16];
...@@ -71,7 +70,7 @@ static int chacha_neon_stream_xor(struct skcipher_request *req, ...@@ -71,7 +70,7 @@ static int chacha_neon_stream_xor(struct skcipher_request *req,
err = skcipher_walk_virt(&walk, req, false); err = skcipher_walk_virt(&walk, req, false);
crypto_chacha_init(state, ctx, iv); chacha_init_generic(state, ctx->key, iv);
while (walk.nbytes > 0) { while (walk.nbytes > 0) {
unsigned int nbytes = walk.nbytes; unsigned int nbytes = walk.nbytes;
...@@ -79,28 +78,41 @@ static int chacha_neon_stream_xor(struct skcipher_request *req, ...@@ -79,28 +78,41 @@ static int chacha_neon_stream_xor(struct skcipher_request *req,
if (nbytes < walk.total) if (nbytes < walk.total)
nbytes = round_down(nbytes, walk.stride); nbytes = round_down(nbytes, walk.stride);
if (!neon) {
chacha_doarm(walk.dst.virt.addr, walk.src.virt.addr,
nbytes, state, ctx->nrounds);
state[12] += DIV_ROUND_UP(nbytes, CHACHA_BLOCK_SIZE);
} else {
kernel_neon_begin(); kernel_neon_begin();
chacha_doneon(state, walk.dst.virt.addr, walk.src.virt.addr, chacha_doneon(state, walk.dst.virt.addr,
nbytes, ctx->nrounds); walk.src.virt.addr, nbytes, ctx->nrounds);
kernel_neon_end(); kernel_neon_end();
}
err = skcipher_walk_done(&walk, walk.nbytes - nbytes); err = skcipher_walk_done(&walk, walk.nbytes - nbytes);
} }
return err; return err;
} }
static int chacha_neon(struct skcipher_request *req) static int do_chacha(struct skcipher_request *req, bool neon)
{ {
struct crypto_skcipher *tfm = crypto_skcipher_reqtfm(req); struct crypto_skcipher *tfm = crypto_skcipher_reqtfm(req);
struct chacha_ctx *ctx = crypto_skcipher_ctx(tfm); struct chacha_ctx *ctx = crypto_skcipher_ctx(tfm);
if (req->cryptlen <= CHACHA_BLOCK_SIZE || !crypto_simd_usable()) return chacha_stream_xor(req, ctx, req->iv, neon);
return crypto_chacha_crypt(req); }
return chacha_neon_stream_xor(req, ctx, req->iv); static int chacha_arm(struct skcipher_request *req)
{
return do_chacha(req, false);
} }
static int xchacha_neon(struct skcipher_request *req) static int chacha_neon(struct skcipher_request *req)
{
return do_chacha(req, neon_usable());
}
static int do_xchacha(struct skcipher_request *req, bool neon)
{ {
struct crypto_skcipher *tfm = crypto_skcipher_reqtfm(req); struct crypto_skcipher *tfm = crypto_skcipher_reqtfm(req);
struct chacha_ctx *ctx = crypto_skcipher_ctx(tfm); struct chacha_ctx *ctx = crypto_skcipher_ctx(tfm);
...@@ -108,22 +120,82 @@ static int xchacha_neon(struct skcipher_request *req) ...@@ -108,22 +120,82 @@ static int xchacha_neon(struct skcipher_request *req)
u32 state[16]; u32 state[16];
u8 real_iv[16]; u8 real_iv[16];
if (req->cryptlen <= CHACHA_BLOCK_SIZE || !crypto_simd_usable()) chacha_init_generic(state, ctx->key, req->iv);
return crypto_xchacha_crypt(req);
crypto_chacha_init(state, ctx, req->iv);
if (!neon) {
hchacha_block_arm(state, subctx.key, ctx->nrounds);
} else {
kernel_neon_begin(); kernel_neon_begin();
hchacha_block_neon(state, subctx.key, ctx->nrounds); hchacha_block_neon(state, subctx.key, ctx->nrounds);
kernel_neon_end(); kernel_neon_end();
}
subctx.nrounds = ctx->nrounds; subctx.nrounds = ctx->nrounds;
memcpy(&real_iv[0], req->iv + 24, 8); memcpy(&real_iv[0], req->iv + 24, 8);
memcpy(&real_iv[8], req->iv + 16, 8); memcpy(&real_iv[8], req->iv + 16, 8);
return chacha_neon_stream_xor(req, &subctx, real_iv); return chacha_stream_xor(req, &subctx, real_iv, neon);
} }
static struct skcipher_alg algs[] = { static int xchacha_arm(struct skcipher_request *req)
{
return do_xchacha(req, false);
}
static int xchacha_neon(struct skcipher_request *req)
{
return do_xchacha(req, neon_usable());
}
static struct skcipher_alg arm_algs[] = {
{
.base.cra_name = "chacha20",
.base.cra_driver_name = "chacha20-arm",
.base.cra_priority = 200,
.base.cra_blocksize = 1,
.base.cra_ctxsize = sizeof(struct chacha_ctx),
.base.cra_module = THIS_MODULE,
.min_keysize = CHACHA_KEY_SIZE,
.max_keysize = CHACHA_KEY_SIZE,
.ivsize = CHACHA_IV_SIZE,
.chunksize = CHACHA_BLOCK_SIZE,
.setkey = chacha20_setkey,
.encrypt = chacha_arm,
.decrypt = chacha_arm,
}, {
.base.cra_name = "xchacha20",
.base.cra_driver_name = "xchacha20-arm",
.base.cra_priority = 200,
.base.cra_blocksize = 1,
.base.cra_ctxsize = sizeof(struct chacha_ctx),
.base.cra_module = THIS_MODULE,
.min_keysize = CHACHA_KEY_SIZE,
.max_keysize = CHACHA_KEY_SIZE,
.ivsize = XCHACHA_IV_SIZE,
.chunksize = CHACHA_BLOCK_SIZE,
.setkey = chacha20_setkey,
.encrypt = xchacha_arm,
.decrypt = xchacha_arm,
}, {
.base.cra_name = "xchacha12",
.base.cra_driver_name = "xchacha12-arm",
.base.cra_priority = 200,
.base.cra_blocksize = 1,
.base.cra_ctxsize = sizeof(struct chacha_ctx),
.base.cra_module = THIS_MODULE,
.min_keysize = CHACHA_KEY_SIZE,
.max_keysize = CHACHA_KEY_SIZE,
.ivsize = XCHACHA_IV_SIZE,
.chunksize = CHACHA_BLOCK_SIZE,
.setkey = chacha12_setkey,
.encrypt = xchacha_arm,
.decrypt = xchacha_arm,
},
};
static struct skcipher_alg neon_algs[] = {
{ {
.base.cra_name = "chacha20", .base.cra_name = "chacha20",
.base.cra_driver_name = "chacha20-neon", .base.cra_driver_name = "chacha20-neon",
...@@ -137,7 +209,7 @@ static struct skcipher_alg algs[] = { ...@@ -137,7 +209,7 @@ static struct skcipher_alg algs[] = {
.ivsize = CHACHA_IV_SIZE, .ivsize = CHACHA_IV_SIZE,
.chunksize = CHACHA_BLOCK_SIZE, .chunksize = CHACHA_BLOCK_SIZE,
.walksize = 4 * CHACHA_BLOCK_SIZE, .walksize = 4 * CHACHA_BLOCK_SIZE,
.setkey = crypto_chacha20_setkey, .setkey = chacha20_setkey,
.encrypt = chacha_neon, .encrypt = chacha_neon,
.decrypt = chacha_neon, .decrypt = chacha_neon,
}, { }, {
...@@ -153,7 +225,7 @@ static struct skcipher_alg algs[] = { ...@@ -153,7 +225,7 @@ static struct skcipher_alg algs[] = {
.ivsize = XCHACHA_IV_SIZE, .ivsize = XCHACHA_IV_SIZE,
.chunksize = CHACHA_BLOCK_SIZE, .chunksize = CHACHA_BLOCK_SIZE,
.walksize = 4 * CHACHA_BLOCK_SIZE, .walksize = 4 * CHACHA_BLOCK_SIZE,
.setkey = crypto_chacha20_setkey, .setkey = chacha20_setkey,
.encrypt = xchacha_neon, .encrypt = xchacha_neon,
.decrypt = xchacha_neon, .decrypt = xchacha_neon,
}, { }, {
...@@ -169,7 +241,7 @@ static struct skcipher_alg algs[] = { ...@@ -169,7 +241,7 @@ static struct skcipher_alg algs[] = {
.ivsize = XCHACHA_IV_SIZE, .ivsize = XCHACHA_IV_SIZE,
.chunksize = CHACHA_BLOCK_SIZE, .chunksize = CHACHA_BLOCK_SIZE,
.walksize = 4 * CHACHA_BLOCK_SIZE, .walksize = 4 * CHACHA_BLOCK_SIZE,
.setkey = crypto_chacha12_setkey, .setkey = chacha12_setkey,
.encrypt = xchacha_neon, .encrypt = xchacha_neon,
.decrypt = xchacha_neon, .decrypt = xchacha_neon,
} }
...@@ -177,26 +249,56 @@ static struct skcipher_alg algs[] = { ...@@ -177,26 +249,56 @@ static struct skcipher_alg algs[] = {
static int __init chacha_simd_mod_init(void) static int __init chacha_simd_mod_init(void)
{ {
if (!(elf_hwcap & HWCAP_NEON)) int err;
return -ENODEV;
return crypto_register_skciphers(algs, ARRAY_SIZE(algs)); err = crypto_register_skciphers(arm_algs, ARRAY_SIZE(arm_algs));
if (err)
return err;
if (IS_ENABLED(CONFIG_KERNEL_MODE_NEON) && (elf_hwcap & HWCAP_NEON)) {
int i;
switch (read_cpuid_part()) {
case ARM_CPU_PART_CORTEX_A7:
case ARM_CPU_PART_CORTEX_A5:
/*
* The Cortex-A7 and Cortex-A5 do not perform well with
* the NEON implementation but do incredibly with the
* scalar one and use less power.
*/
for (i = 0; i < ARRAY_SIZE(neon_algs); i++)
neon_algs[i].base.cra_priority = 0;
break;
}
err = crypto_register_skciphers(neon_algs, ARRAY_SIZE(neon_algs));
if (err)
crypto_unregister_skciphers(arm_algs, ARRAY_SIZE(arm_algs));
}
return err;
} }
static void __exit chacha_simd_mod_fini(void) static void __exit chacha_simd_mod_fini(void)
{ {
crypto_unregister_skciphers(algs, ARRAY_SIZE(algs)); crypto_unregister_skciphers(arm_algs, ARRAY_SIZE(arm_algs));
if (IS_ENABLED(CONFIG_KERNEL_MODE_NEON) && (elf_hwcap & HWCAP_NEON))
crypto_unregister_skciphers(neon_algs, ARRAY_SIZE(neon_algs));
} }
module_init(chacha_simd_mod_init); module_init(chacha_simd_mod_init);
module_exit(chacha_simd_mod_fini); module_exit(chacha_simd_mod_fini);
MODULE_DESCRIPTION("ChaCha and XChaCha stream ciphers (NEON accelerated)"); MODULE_DESCRIPTION("ChaCha and XChaCha stream ciphers (scalar and NEON accelerated)");
MODULE_AUTHOR("Ard Biesheuvel <ard.biesheuvel@linaro.org>"); MODULE_AUTHOR("Ard Biesheuvel <ard.biesheuvel@linaro.org>");
MODULE_LICENSE("GPL v2"); MODULE_LICENSE("GPL v2");
MODULE_ALIAS_CRYPTO("chacha20"); MODULE_ALIAS_CRYPTO("chacha20");
MODULE_ALIAS_CRYPTO("chacha20-neon"); MODULE_ALIAS_CRYPTO("chacha20-arm");
MODULE_ALIAS_CRYPTO("xchacha20"); MODULE_ALIAS_CRYPTO("xchacha20");
MODULE_ALIAS_CRYPTO("xchacha20-neon"); MODULE_ALIAS_CRYPTO("xchacha20-arm");
MODULE_ALIAS_CRYPTO("xchacha12"); MODULE_ALIAS_CRYPTO("xchacha12");
MODULE_ALIAS_CRYPTO("xchacha12-arm");
#ifdef CONFIG_KERNEL_MODE_NEON
MODULE_ALIAS_CRYPTO("chacha20-neon");
MODULE_ALIAS_CRYPTO("xchacha20-neon");
MODULE_ALIAS_CRYPTO("xchacha12-neon"); MODULE_ALIAS_CRYPTO("xchacha12-neon");
#endif
...@@ -41,14 +41,6 @@ ...@@ -41,14 +41,6 @@
X14 .req r12 X14 .req r12
X15 .req r14 X15 .req r14
.Lexpand_32byte_k:
// "expand 32-byte k"
.word 0x61707865, 0x3320646e, 0x79622d32, 0x6b206574
#ifdef __thumb2__
# define adrl adr
#endif
.macro __rev out, in, t0, t1, t2 .macro __rev out, in, t0, t1, t2
.if __LINUX_ARM_ARCH__ >= 6 .if __LINUX_ARM_ARCH__ >= 6
rev \out, \in rev \out, \in
...@@ -391,61 +383,65 @@ ...@@ -391,61 +383,65 @@
.endm // _chacha .endm // _chacha
/* /*
* void chacha20_arm(u8 *out, const u8 *in, size_t len, const u32 key[8], * void chacha_doarm(u8 *dst, const u8 *src, unsigned int bytes,
* const u32 iv[4]); * const u32 *state, int nrounds);
*/ */
ENTRY(chacha20_arm) ENTRY(chacha_doarm)
cmp r2, #0 // len == 0? cmp r2, #0 // len == 0?
reteq lr reteq lr
ldr ip, [sp]
cmp ip, #12
push {r0-r2,r4-r11,lr} push {r0-r2,r4-r11,lr}
// Push state x0-x15 onto stack. // Push state x0-x15 onto stack.
// Also store an extra copy of x10-x11 just before the state. // Also store an extra copy of x10-x11 just before the state.
ldr r4, [sp, #48] // iv add X12, r3, #48
mov r0, sp ldm X12, {X12,X13,X14,X15}
sub sp, #80 push {X12,X13,X14,X15}
sub sp, sp, #64
// iv: x12-x15
ldm r4, {X12,X13,X14,X15}
stmdb r0!, {X12,X13,X14,X15}
// key: x4-x11 __ldrd X8_X10, X9_X11, r3, 40
__ldrd X8_X10, X9_X11, r3, 24
__strd X8_X10, X9_X11, sp, 8 __strd X8_X10, X9_X11, sp, 8
stmdb r0!, {X8_X10, X9_X11} __strd X8_X10, X9_X11, sp, 56
ldm r3, {X4-X9_X11} ldm r3, {X0-X9_X11}
stmdb r0!, {X4-X9_X11}
// constants: x0-x3
adrl X3, .Lexpand_32byte_k
ldm X3, {X0-X3}
__strd X0, X1, sp, 16 __strd X0, X1, sp, 16
__strd X2, X3, sp, 24 __strd X2, X3, sp, 24
__strd X4, X5, sp, 32
__strd X6, X7, sp, 40
__strd X8_X10, X9_X11, sp, 48
beq 1f
_chacha 20 _chacha 20
add sp, #76 0: add sp, #76
pop {r4-r11, pc} pop {r4-r11, pc}
ENDPROC(chacha20_arm)
1: _chacha 12
b 0b
ENDPROC(chacha_doarm)
/* /*
* void hchacha20_arm(const u32 state[16], u32 out[8]); * void hchacha_block_arm(const u32 state[16], u32 out[8], int nrounds);
*/ */
ENTRY(hchacha20_arm) ENTRY(hchacha_block_arm)
push {r1,r4-r11,lr} push {r1,r4-r11,lr}
cmp r2, #12 // ChaCha12 ?
mov r14, r0 mov r14, r0
ldmia r14!, {r0-r11} // load x0-x11 ldmia r14!, {r0-r11} // load x0-x11
push {r10-r11} // store x10-x11 to stack push {r10-r11} // store x10-x11 to stack
ldm r14, {r10-r12,r14} // load x12-x15 ldm r14, {r10-r12,r14} // load x12-x15
sub sp, #8 sub sp, #8
beq 1f
_chacha_permute 20 _chacha_permute 20
// Skip over (unused0-unused1, x10-x11) // Skip over (unused0-unused1, x10-x11)
add sp, #16 0: add sp, #16
// Fix up rotations of x12-x15 // Fix up rotations of x12-x15
ror X12, X12, #drot ror X12, X12, #drot
...@@ -458,4 +454,7 @@ ENTRY(hchacha20_arm) ...@@ -458,4 +454,7 @@ ENTRY(hchacha20_arm)
stm r4, {X0,X1,X2,X3,X12,X13,X14,X15} stm r4, {X0,X1,X2,X3,X12,X13,X14,X15}
pop {r4-r11,pc} pop {r4-r11,pc}
ENDPROC(hchacha20_arm)
1: _chacha_permute 12
b 0b
ENDPROC(hchacha_block_arm)
/* /*
* ARM NEON accelerated ChaCha and XChaCha stream ciphers, * ARM NEON and scalar accelerated ChaCha and XChaCha stream ciphers,
* including ChaCha20 (RFC7539) * including ChaCha20 (RFC7539)
* *
* Copyright (C) 2016 - 2017 Linaro, Ltd. <ard.biesheuvel@linaro.org> * Copyright (C) 2016 - 2017 Linaro, Ltd. <ard.biesheuvel@linaro.org>
......
Markdown is supported
0%
or
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment