From 3265c4babe93832167cb148083a0544548c23e6a Mon Sep 17 00:00:00 2001 From: Kim Phillips Date: Fri, 6 Mar 2015 18:46:21 -0600 Subject: crypto: powerpc - move files to fix build error The current cryptodev-2.6 tree commits: d9850fc529ef ("crypto: powerpc/sha1 - kernel config") 50ba29aaa7b0 ("crypto: powerpc/sha1 - glue") failed to properly place files under arch/powerpc/crypto, which leads to build errors: make[1]: *** No rule to make target 'arch/powerpc/crypto/sha1-spe-asm.o', needed by 'arch/powerpc/crypto/sha1-ppc-spe.o'. Stop. make[1]: *** No rule to make target 'arch/powerpc/crypto/sha1_spe_glue.o', needed by 'arch/powerpc/crypto/sha1-ppc-spe.o'. Stop. Makefile:947: recipe for target 'arch/powerpc/crypto' failed Move the two sha1 spe files under crypto/, and whilst there, rename other powerpc crypto files with underscores to use dashes for consistency. Cc: Markus Stockhausen Signed-off-by: Kim Phillips Signed-off-by: Herbert Xu --- arch/powerpc/crypto/Makefile | 8 +- arch/powerpc/crypto/aes-spe-glue.c | 512 ++++++++++++++++++++++++++++++++++ arch/powerpc/crypto/aes_spe_glue.c | 512 ---------------------------------- arch/powerpc/crypto/md5-glue.c | 165 +++++++++++ arch/powerpc/crypto/md5_glue.c | 165 ----------- arch/powerpc/crypto/sha1-spe-asm.S | 299 ++++++++++++++++++++ arch/powerpc/crypto/sha1-spe-glue.c | 210 ++++++++++++++ arch/powerpc/crypto/sha256-spe-glue.c | 275 ++++++++++++++++++ arch/powerpc/crypto/sha256_spe_glue.c | 275 ------------------ arch/powerpc/sha1-spe-asm.S | 299 -------------------- arch/powerpc/sha1_spe_glue.c | 210 -------------- 11 files changed, 1465 insertions(+), 1465 deletions(-) create mode 100644 arch/powerpc/crypto/aes-spe-glue.c delete mode 100644 arch/powerpc/crypto/aes_spe_glue.c create mode 100644 arch/powerpc/crypto/md5-glue.c delete mode 100644 arch/powerpc/crypto/md5_glue.c create mode 100644 arch/powerpc/crypto/sha1-spe-asm.S create mode 100644 arch/powerpc/crypto/sha1-spe-glue.c create mode 100644 arch/powerpc/crypto/sha256-spe-glue.c delete mode 100644 arch/powerpc/crypto/sha256_spe_glue.c delete mode 100644 arch/powerpc/sha1-spe-asm.S delete mode 100644 arch/powerpc/sha1_spe_glue.c diff --git a/arch/powerpc/crypto/Makefile b/arch/powerpc/crypto/Makefile index c6b25cba3a0c..9c221b69c181 100644 --- a/arch/powerpc/crypto/Makefile +++ b/arch/powerpc/crypto/Makefile @@ -10,8 +10,8 @@ obj-$(CONFIG_CRYPTO_SHA1_PPC) += sha1-powerpc.o obj-$(CONFIG_CRYPTO_SHA1_PPC_SPE) += sha1-ppc-spe.o obj-$(CONFIG_CRYPTO_SHA256_PPC_SPE) += sha256-ppc-spe.o -aes-ppc-spe-y := aes-spe-core.o aes-spe-keys.o aes-tab-4k.o aes-spe-modes.o aes_spe_glue.o -md5-ppc-y := md5-asm.o md5_glue.o +aes-ppc-spe-y := aes-spe-core.o aes-spe-keys.o aes-tab-4k.o aes-spe-modes.o aes-spe-glue.o +md5-ppc-y := md5-asm.o md5-glue.o sha1-powerpc-y := sha1-powerpc-asm.o sha1.o -sha1-ppc-spe-y := sha1-spe-asm.o sha1_spe_glue.o -sha256-ppc-spe-y := sha256-spe-asm.o sha256_spe_glue.o +sha1-ppc-spe-y := sha1-spe-asm.o sha1-spe-glue.o +sha256-ppc-spe-y := sha256-spe-asm.o sha256-spe-glue.o diff --git a/arch/powerpc/crypto/aes-spe-glue.c b/arch/powerpc/crypto/aes-spe-glue.c new file mode 100644 index 000000000000..bd5e63f72ad4 --- /dev/null +++ b/arch/powerpc/crypto/aes-spe-glue.c @@ -0,0 +1,512 @@ +/* + * Glue code for AES implementation for SPE instructions (PPC) + * + * Based on generic implementation. The assembler module takes care + * about the SPE registers so it can run from interrupt context. + * + * Copyright (c) 2015 Markus Stockhausen + * + * 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 +#include +#include +#include +#include +#include +#include +#include +#include + +/* + * MAX_BYTES defines the number of bytes that are allowed to be processed + * between preempt_disable() and preempt_enable(). e500 cores can issue two + * instructions per clock cycle using one 32/64 bit unit (SU1) and one 32 + * bit unit (SU2). One of these can be a memory access that is executed via + * a single load and store unit (LSU). XTS-AES-256 takes ~780 operations per + * 16 byte block block or 25 cycles per byte. Thus 768 bytes of input data + * will need an estimated maximum of 20,000 cycles. Headroom for cache misses + * included. Even with the low end model clocked at 667 MHz this equals to a + * critical time window of less than 30us. The value has been choosen to + * process a 512 byte disk block in one or a large 1400 bytes IPsec network + * packet in two runs. + * + */ +#define MAX_BYTES 768 + +struct ppc_aes_ctx { + u32 key_enc[AES_MAX_KEYLENGTH_U32]; + u32 key_dec[AES_MAX_KEYLENGTH_U32]; + u32 rounds; +}; + +struct ppc_xts_ctx { + u32 key_enc[AES_MAX_KEYLENGTH_U32]; + u32 key_dec[AES_MAX_KEYLENGTH_U32]; + u32 key_twk[AES_MAX_KEYLENGTH_U32]; + u32 rounds; +}; + +extern void ppc_encrypt_aes(u8 *out, const u8 *in, u32 *key_enc, u32 rounds); +extern void ppc_decrypt_aes(u8 *out, const u8 *in, u32 *key_dec, u32 rounds); +extern void ppc_encrypt_ecb(u8 *out, const u8 *in, u32 *key_enc, u32 rounds, + u32 bytes); +extern void ppc_decrypt_ecb(u8 *out, const u8 *in, u32 *key_dec, u32 rounds, + u32 bytes); +extern void ppc_encrypt_cbc(u8 *out, const u8 *in, u32 *key_enc, u32 rounds, + u32 bytes, u8 *iv); +extern void ppc_decrypt_cbc(u8 *out, const u8 *in, u32 *key_dec, u32 rounds, + u32 bytes, u8 *iv); +extern void ppc_crypt_ctr (u8 *out, const u8 *in, u32 *key_enc, u32 rounds, + u32 bytes, u8 *iv); +extern void ppc_encrypt_xts(u8 *out, const u8 *in, u32 *key_enc, u32 rounds, + u32 bytes, u8 *iv, u32 *key_twk); +extern void ppc_decrypt_xts(u8 *out, const u8 *in, u32 *key_dec, u32 rounds, + u32 bytes, u8 *iv, u32 *key_twk); + +extern void ppc_expand_key_128(u32 *key_enc, const u8 *key); +extern void ppc_expand_key_192(u32 *key_enc, const u8 *key); +extern void ppc_expand_key_256(u32 *key_enc, const u8 *key); + +extern void ppc_generate_decrypt_key(u32 *key_dec,u32 *key_enc, + unsigned int key_len); + +static void spe_begin(void) +{ + /* disable preemption and save users SPE registers if required */ + preempt_disable(); + enable_kernel_spe(); +} + +static void spe_end(void) +{ + /* reenable preemption */ + preempt_enable(); +} + +static int ppc_aes_setkey(struct crypto_tfm *tfm, const u8 *in_key, + unsigned int key_len) +{ + struct ppc_aes_ctx *ctx = crypto_tfm_ctx(tfm); + + if (key_len != AES_KEYSIZE_128 && + key_len != AES_KEYSIZE_192 && + key_len != AES_KEYSIZE_256) { + tfm->crt_flags |= CRYPTO_TFM_RES_BAD_KEY_LEN; + return -EINVAL; + } + + switch (key_len) { + case AES_KEYSIZE_128: + ctx->rounds = 4; + ppc_expand_key_128(ctx->key_enc, in_key); + break; + case AES_KEYSIZE_192: + ctx->rounds = 5; + ppc_expand_key_192(ctx->key_enc, in_key); + break; + case AES_KEYSIZE_256: + ctx->rounds = 6; + ppc_expand_key_256(ctx->key_enc, in_key); + break; + } + + ppc_generate_decrypt_key(ctx->key_dec, ctx->key_enc, key_len); + + return 0; +} + +static int ppc_xts_setkey(struct crypto_tfm *tfm, const u8 *in_key, + unsigned int key_len) +{ + struct ppc_xts_ctx *ctx = crypto_tfm_ctx(tfm); + + key_len >>= 1; + + if (key_len != AES_KEYSIZE_128 && + key_len != AES_KEYSIZE_192 && + key_len != AES_KEYSIZE_256) { + tfm->crt_flags |= CRYPTO_TFM_RES_BAD_KEY_LEN; + return -EINVAL; + } + + switch (key_len) { + case AES_KEYSIZE_128: + ctx->rounds = 4; + ppc_expand_key_128(ctx->key_enc, in_key); + ppc_expand_key_128(ctx->key_twk, in_key + AES_KEYSIZE_128); + break; + case AES_KEYSIZE_192: + ctx->rounds = 5; + ppc_expand_key_192(ctx->key_enc, in_key); + ppc_expand_key_192(ctx->key_twk, in_key + AES_KEYSIZE_192); + break; + case AES_KEYSIZE_256: + ctx->rounds = 6; + ppc_expand_key_256(ctx->key_enc, in_key); + ppc_expand_key_256(ctx->key_twk, in_key + AES_KEYSIZE_256); + break; + } + + ppc_generate_decrypt_key(ctx->key_dec, ctx->key_enc, key_len); + + return 0; +} + +static void ppc_aes_encrypt(struct crypto_tfm *tfm, u8 *out, const u8 *in) +{ + struct ppc_aes_ctx *ctx = crypto_tfm_ctx(tfm); + + spe_begin(); + ppc_encrypt_aes(out, in, ctx->key_enc, ctx->rounds); + spe_end(); +} + +static void ppc_aes_decrypt(struct crypto_tfm *tfm, u8 *out, const u8 *in) +{ + struct ppc_aes_ctx *ctx = crypto_tfm_ctx(tfm); + + spe_begin(); + ppc_decrypt_aes(out, in, ctx->key_dec, ctx->rounds); + spe_end(); +} + +static int ppc_ecb_encrypt(struct blkcipher_desc *desc, struct scatterlist *dst, + struct scatterlist *src, unsigned int nbytes) +{ + struct ppc_aes_ctx *ctx = crypto_blkcipher_ctx(desc->tfm); + struct blkcipher_walk walk; + unsigned int ubytes; + int err; + + desc->flags &= ~CRYPTO_TFM_REQ_MAY_SLEEP; + blkcipher_walk_init(&walk, dst, src, nbytes); + err = blkcipher_walk_virt(desc, &walk); + + while ((nbytes = walk.nbytes)) { + ubytes = nbytes > MAX_BYTES ? + nbytes - MAX_BYTES : nbytes & (AES_BLOCK_SIZE - 1); + nbytes -= ubytes; + + spe_begin(); + ppc_encrypt_ecb(walk.dst.virt.addr, walk.src.virt.addr, + ctx->key_enc, ctx->rounds, nbytes); + spe_end(); + + err = blkcipher_walk_done(desc, &walk, ubytes); + } + + return err; +} + +static int ppc_ecb_decrypt(struct blkcipher_desc *desc, struct scatterlist *dst, + struct scatterlist *src, unsigned int nbytes) +{ + struct ppc_aes_ctx *ctx = crypto_blkcipher_ctx(desc->tfm); + struct blkcipher_walk walk; + unsigned int ubytes; + int err; + + desc->flags &= ~CRYPTO_TFM_REQ_MAY_SLEEP; + blkcipher_walk_init(&walk, dst, src, nbytes); + err = blkcipher_walk_virt(desc, &walk); + + while ((nbytes = walk.nbytes)) { + ubytes = nbytes > MAX_BYTES ? + nbytes - MAX_BYTES : nbytes & (AES_BLOCK_SIZE - 1); + nbytes -= ubytes; + + spe_begin(); + ppc_decrypt_ecb(walk.dst.virt.addr, walk.src.virt.addr, + ctx->key_dec, ctx->rounds, nbytes); + spe_end(); + + err = blkcipher_walk_done(desc, &walk, ubytes); + } + + return err; +} + +static int ppc_cbc_encrypt(struct blkcipher_desc *desc, struct scatterlist *dst, + struct scatterlist *src, unsigned int nbytes) +{ + struct ppc_aes_ctx *ctx = crypto_blkcipher_ctx(desc->tfm); + struct blkcipher_walk walk; + unsigned int ubytes; + int err; + + desc->flags &= ~CRYPTO_TFM_REQ_MAY_SLEEP; + blkcipher_walk_init(&walk, dst, src, nbytes); + err = blkcipher_walk_virt(desc, &walk); + + while ((nbytes = walk.nbytes)) { + ubytes = nbytes > MAX_BYTES ? + nbytes - MAX_BYTES : nbytes & (AES_BLOCK_SIZE - 1); + nbytes -= ubytes; + + spe_begin(); + ppc_encrypt_cbc(walk.dst.virt.addr, walk.src.virt.addr, + ctx->key_enc, ctx->rounds, nbytes, walk.iv); + spe_end(); + + err = blkcipher_walk_done(desc, &walk, ubytes); + } + + return err; +} + +static int ppc_cbc_decrypt(struct blkcipher_desc *desc, struct scatterlist *dst, + struct scatterlist *src, unsigned int nbytes) +{ + struct ppc_aes_ctx *ctx = crypto_blkcipher_ctx(desc->tfm); + struct blkcipher_walk walk; + unsigned int ubytes; + int err; + + desc->flags &= ~CRYPTO_TFM_REQ_MAY_SLEEP; + blkcipher_walk_init(&walk, dst, src, nbytes); + err = blkcipher_walk_virt(desc, &walk); + + while ((nbytes = walk.nbytes)) { + ubytes = nbytes > MAX_BYTES ? + nbytes - MAX_BYTES : nbytes & (AES_BLOCK_SIZE - 1); + nbytes -= ubytes; + + spe_begin(); + ppc_decrypt_cbc(walk.dst.virt.addr, walk.src.virt.addr, + ctx->key_dec, ctx->rounds, nbytes, walk.iv); + spe_end(); + + err = blkcipher_walk_done(desc, &walk, ubytes); + } + + return err; +} + +static int ppc_ctr_crypt(struct blkcipher_desc *desc, struct scatterlist *dst, + struct scatterlist *src, unsigned int nbytes) +{ + struct ppc_aes_ctx *ctx = crypto_blkcipher_ctx(desc->tfm); + struct blkcipher_walk walk; + unsigned int pbytes, ubytes; + int err; + + desc->flags &= ~CRYPTO_TFM_REQ_MAY_SLEEP; + blkcipher_walk_init(&walk, dst, src, nbytes); + err = blkcipher_walk_virt_block(desc, &walk, AES_BLOCK_SIZE); + + while ((pbytes = walk.nbytes)) { + pbytes = pbytes > MAX_BYTES ? MAX_BYTES : pbytes; + pbytes = pbytes == nbytes ? + nbytes : pbytes & ~(AES_BLOCK_SIZE - 1); + ubytes = walk.nbytes - pbytes; + + spe_begin(); + ppc_crypt_ctr(walk.dst.virt.addr, walk.src.virt.addr, + ctx->key_enc, ctx->rounds, pbytes , walk.iv); + spe_end(); + + nbytes -= pbytes; + err = blkcipher_walk_done(desc, &walk, ubytes); + } + + return err; +} + +static int ppc_xts_encrypt(struct blkcipher_desc *desc, struct scatterlist *dst, + struct scatterlist *src, unsigned int nbytes) +{ + struct ppc_xts_ctx *ctx = crypto_blkcipher_ctx(desc->tfm); + struct blkcipher_walk walk; + unsigned int ubytes; + int err; + u32 *twk; + + desc->flags &= ~CRYPTO_TFM_REQ_MAY_SLEEP; + blkcipher_walk_init(&walk, dst, src, nbytes); + err = blkcipher_walk_virt(desc, &walk); + twk = ctx->key_twk; + + while ((nbytes = walk.nbytes)) { + ubytes = nbytes > MAX_BYTES ? + nbytes - MAX_BYTES : nbytes & (AES_BLOCK_SIZE - 1); + nbytes -= ubytes; + + spe_begin(); + ppc_encrypt_xts(walk.dst.virt.addr, walk.src.virt.addr, + ctx->key_enc, ctx->rounds, nbytes, walk.iv, twk); + spe_end(); + + twk = NULL; + err = blkcipher_walk_done(desc, &walk, ubytes); + } + + return err; +} + +static int ppc_xts_decrypt(struct blkcipher_desc *desc, struct scatterlist *dst, + struct scatterlist *src, unsigned int nbytes) +{ + struct ppc_xts_ctx *ctx = crypto_blkcipher_ctx(desc->tfm); + struct blkcipher_walk walk; + unsigned int ubytes; + int err; + u32 *twk; + + desc->flags &= ~CRYPTO_TFM_REQ_MAY_SLEEP; + blkcipher_walk_init(&walk, dst, src, nbytes); + err = blkcipher_walk_virt(desc, &walk); + twk = ctx->key_twk; + + while ((nbytes = walk.nbytes)) { + ubytes = nbytes > MAX_BYTES ? + nbytes - MAX_BYTES : nbytes & (AES_BLOCK_SIZE - 1); + nbytes -= ubytes; + + spe_begin(); + ppc_decrypt_xts(walk.dst.virt.addr, walk.src.virt.addr, + ctx->key_dec, ctx->rounds, nbytes, walk.iv, twk); + spe_end(); + + twk = NULL; + err = blkcipher_walk_done(desc, &walk, ubytes); + } + + return err; +} + +/* + * Algorithm definitions. Disabling alignment (cra_alignmask=0) was chosen + * because the e500 platform can handle unaligned reads/writes very efficently. + * This improves IPsec thoughput by another few percent. Additionally we assume + * that AES context is always aligned to at least 8 bytes because it is created + * with kmalloc() in the crypto infrastructure + * + */ +static struct crypto_alg aes_algs[] = { { + .cra_name = "aes", + .cra_driver_name = "aes-ppc-spe", + .cra_priority = 300, + .cra_flags = CRYPTO_ALG_TYPE_CIPHER, + .cra_blocksize = AES_BLOCK_SIZE, + .cra_ctxsize = sizeof(struct ppc_aes_ctx), + .cra_alignmask = 0, + .cra_module = THIS_MODULE, + .cra_u = { + .cipher = { + .cia_min_keysize = AES_MIN_KEY_SIZE, + .cia_max_keysize = AES_MAX_KEY_SIZE, + .cia_setkey = ppc_aes_setkey, + .cia_encrypt = ppc_aes_encrypt, + .cia_decrypt = ppc_aes_decrypt + } + } +}, { + .cra_name = "ecb(aes)", + .cra_driver_name = "ecb-ppc-spe", + .cra_priority = 300, + .cra_flags = CRYPTO_ALG_TYPE_BLKCIPHER, + .cra_blocksize = AES_BLOCK_SIZE, + .cra_ctxsize = sizeof(struct ppc_aes_ctx), + .cra_alignmask = 0, + .cra_type = &crypto_blkcipher_type, + .cra_module = THIS_MODULE, + .cra_u = { + .blkcipher = { + .min_keysize = AES_MIN_KEY_SIZE, + .max_keysize = AES_MAX_KEY_SIZE, + .ivsize = AES_BLOCK_SIZE, + .setkey = ppc_aes_setkey, + .encrypt = ppc_ecb_encrypt, + .decrypt = ppc_ecb_decrypt, + } + } +}, { + .cra_name = "cbc(aes)", + .cra_driver_name = "cbc-ppc-spe", + .cra_priority = 300, + .cra_flags = CRYPTO_ALG_TYPE_BLKCIPHER, + .cra_blocksize = AES_BLOCK_SIZE, + .cra_ctxsize = sizeof(struct ppc_aes_ctx), + .cra_alignmask = 0, + .cra_type = &crypto_blkcipher_type, + .cra_module = THIS_MODULE, + .cra_u = { + .blkcipher = { + .min_keysize = AES_MIN_KEY_SIZE, + .max_keysize = AES_MAX_KEY_SIZE, + .ivsize = AES_BLOCK_SIZE, + .setkey = ppc_aes_setkey, + .encrypt = ppc_cbc_encrypt, + .decrypt = ppc_cbc_decrypt, + } + } +}, { + .cra_name = "ctr(aes)", + .cra_driver_name = "ctr-ppc-spe", + .cra_priority = 300, + .cra_flags = CRYPTO_ALG_TYPE_BLKCIPHER, + .cra_blocksize = 1, + .cra_ctxsize = sizeof(struct ppc_aes_ctx), + .cra_alignmask = 0, + .cra_type = &crypto_blkcipher_type, + .cra_module = THIS_MODULE, + .cra_u = { + .blkcipher = { + .min_keysize = AES_MIN_KEY_SIZE, + .max_keysize = AES_MAX_KEY_SIZE, + .ivsize = AES_BLOCK_SIZE, + .setkey = ppc_aes_setkey, + .encrypt = ppc_ctr_crypt, + .decrypt = ppc_ctr_crypt, + } + } +}, { + .cra_name = "xts(aes)", + .cra_driver_name = "xts-ppc-spe", + .cra_priority = 300, + .cra_flags = CRYPTO_ALG_TYPE_BLKCIPHER, + .cra_blocksize = AES_BLOCK_SIZE, + .cra_ctxsize = sizeof(struct ppc_xts_ctx), + .cra_alignmask = 0, + .cra_type = &crypto_blkcipher_type, + .cra_module = THIS_MODULE, + .cra_u = { + .blkcipher = { + .min_keysize = AES_MIN_KEY_SIZE * 2, + .max_keysize = AES_MAX_KEY_SIZE * 2, + .ivsize = AES_BLOCK_SIZE, + .setkey = ppc_xts_setkey, + .encrypt = ppc_xts_encrypt, + .decrypt = ppc_xts_decrypt, + } + } +} }; + +static int __init ppc_aes_mod_init(void) +{ + return crypto_register_algs(aes_algs, ARRAY_SIZE(aes_algs)); +} + +static void __exit ppc_aes_mod_fini(void) +{ + crypto_unregister_algs(aes_algs, ARRAY_SIZE(aes_algs)); +} + +module_init(ppc_aes_mod_init); +module_exit(ppc_aes_mod_fini); + +MODULE_LICENSE("GPL"); +MODULE_DESCRIPTION("AES-ECB/CBC/CTR/XTS, SPE optimized"); + +MODULE_ALIAS_CRYPTO("aes"); +MODULE_ALIAS_CRYPTO("ecb(aes)"); +MODULE_ALIAS_CRYPTO("cbc(aes)"); +MODULE_ALIAS_CRYPTO("ctr(aes)"); +MODULE_ALIAS_CRYPTO("xts(aes)"); +MODULE_ALIAS_CRYPTO("aes-ppc-spe"); diff --git a/arch/powerpc/crypto/aes_spe_glue.c b/arch/powerpc/crypto/aes_spe_glue.c deleted file mode 100644 index bd5e63f72ad4..000000000000 --- a/arch/powerpc/crypto/aes_spe_glue.c +++ /dev/null @@ -1,512 +0,0 @@ -/* - * Glue code for AES implementation for SPE instructions (PPC) - * - * Based on generic implementation. The assembler module takes care - * about the SPE registers so it can run from interrupt context. - * - * Copyright (c) 2015 Markus Stockhausen - * - * 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 -#include -#include -#include -#include -#include -#include -#include -#include - -/* - * MAX_BYTES defines the number of bytes that are allowed to be processed - * between preempt_disable() and preempt_enable(). e500 cores can issue two - * instructions per clock cycle using one 32/64 bit unit (SU1) and one 32 - * bit unit (SU2). One of these can be a memory access that is executed via - * a single load and store unit (LSU). XTS-AES-256 takes ~780 operations per - * 16 byte block block or 25 cycles per byte. Thus 768 bytes of input data - * will need an estimated maximum of 20,000 cycles. Headroom for cache misses - * included. Even with the low end model clocked at 667 MHz this equals to a - * critical time window of less than 30us. The value has been choosen to - * process a 512 byte disk block in one or a large 1400 bytes IPsec network - * packet in two runs. - * - */ -#define MAX_BYTES 768 - -struct ppc_aes_ctx { - u32 key_enc[AES_MAX_KEYLENGTH_U32]; - u32 key_dec[AES_MAX_KEYLENGTH_U32]; - u32 rounds; -}; - -struct ppc_xts_ctx { - u32 key_enc[AES_MAX_KEYLENGTH_U32]; - u32 key_dec[AES_MAX_KEYLENGTH_U32]; - u32 key_twk[AES_MAX_KEYLENGTH_U32]; - u32 rounds; -}; - -extern void ppc_encrypt_aes(u8 *out, const u8 *in, u32 *key_enc, u32 rounds); -extern void ppc_decrypt_aes(u8 *out, const u8 *in, u32 *key_dec, u32 rounds); -extern void ppc_encrypt_ecb(u8 *out, const u8 *in, u32 *key_enc, u32 rounds, - u32 bytes); -extern void ppc_decrypt_ecb(u8 *out, const u8 *in, u32 *key_dec, u32 rounds, - u32 bytes); -extern void ppc_encrypt_cbc(u8 *out, const u8 *in, u32 *key_enc, u32 rounds, - u32 bytes, u8 *iv); -extern void ppc_decrypt_cbc(u8 *out, const u8 *in, u32 *key_dec, u32 rounds, - u32 bytes, u8 *iv); -extern void ppc_crypt_ctr (u8 *out, const u8 *in, u32 *key_enc, u32 rounds, - u32 bytes, u8 *iv); -extern void ppc_encrypt_xts(u8 *out, const u8 *in, u32 *key_enc, u32 rounds, - u32 bytes, u8 *iv, u32 *key_twk); -extern void ppc_decrypt_xts(u8 *out, const u8 *in, u32 *key_dec, u32 rounds, - u32 bytes, u8 *iv, u32 *key_twk); - -extern void ppc_expand_key_128(u32 *key_enc, const u8 *key); -extern void ppc_expand_key_192(u32 *key_enc, const u8 *key); -extern void ppc_expand_key_256(u32 *key_enc, const u8 *key); - -extern void ppc_generate_decrypt_key(u32 *key_dec,u32 *key_enc, - unsigned int key_len); - -static void spe_begin(void) -{ - /* disable preemption and save users SPE registers if required */ - preempt_disable(); - enable_kernel_spe(); -} - -static void spe_end(void) -{ - /* reenable preemption */ - preempt_enable(); -} - -static int ppc_aes_setkey(struct crypto_tfm *tfm, const u8 *in_key, - unsigned int key_len) -{ - struct ppc_aes_ctx *ctx = crypto_tfm_ctx(tfm); - - if (key_len != AES_KEYSIZE_128 && - key_len != AES_KEYSIZE_192 && - key_len != AES_KEYSIZE_256) { - tfm->crt_flags |= CRYPTO_TFM_RES_BAD_KEY_LEN; - return -EINVAL; - } - - switch (key_len) { - case AES_KEYSIZE_128: - ctx->rounds = 4; - ppc_expand_key_128(ctx->key_enc, in_key); - break; - case AES_KEYSIZE_192: - ctx->rounds = 5; - ppc_expand_key_192(ctx->key_enc, in_key); - break; - case AES_KEYSIZE_256: - ctx->rounds = 6; - ppc_expand_key_256(ctx->key_enc, in_key); - break; - } - - ppc_generate_decrypt_key(ctx->key_dec, ctx->key_enc, key_len); - - return 0; -} - -static int ppc_xts_setkey(struct crypto_tfm *tfm, const u8 *in_key, - unsigned int key_len) -{ - struct ppc_xts_ctx *ctx = crypto_tfm_ctx(tfm); - - key_len >>= 1; - - if (key_len != AES_KEYSIZE_128 && - key_len != AES_KEYSIZE_192 && - key_len != AES_KEYSIZE_256) { - tfm->crt_flags |= CRYPTO_TFM_RES_BAD_KEY_LEN; - return -EINVAL; - } - - switch (key_len) { - case AES_KEYSIZE_128: - ctx->rounds = 4; - ppc_expand_key_128(ctx->key_enc, in_key); - ppc_expand_key_128(ctx->key_twk, in_key + AES_KEYSIZE_128); - break; - case AES_KEYSIZE_192: - ctx->rounds = 5; - ppc_expand_key_192(ctx->key_enc, in_key); - ppc_expand_key_192(ctx->key_twk, in_key + AES_KEYSIZE_192); - break; - case AES_KEYSIZE_256: - ctx->rounds = 6; - ppc_expand_key_256(ctx->key_enc, in_key); - ppc_expand_key_256(ctx->key_twk, in_key + AES_KEYSIZE_256); - break; - } - - ppc_generate_decrypt_key(ctx->key_dec, ctx->key_enc, key_len); - - return 0; -} - -static void ppc_aes_encrypt(struct crypto_tfm *tfm, u8 *out, const u8 *in) -{ - struct ppc_aes_ctx *ctx = crypto_tfm_ctx(tfm); - - spe_begin(); - ppc_encrypt_aes(out, in, ctx->key_enc, ctx->rounds); - spe_end(); -} - -static void ppc_aes_decrypt(struct crypto_tfm *tfm, u8 *out, const u8 *in) -{ - struct ppc_aes_ctx *ctx = crypto_tfm_ctx(tfm); - - spe_begin(); - ppc_decrypt_aes(out, in, ctx->key_dec, ctx->rounds); - spe_end(); -} - -static int ppc_ecb_encrypt(struct blkcipher_desc *desc, struct scatterlist *dst, - struct scatterlist *src, unsigned int nbytes) -{ - struct ppc_aes_ctx *ctx = crypto_blkcipher_ctx(desc->tfm); - struct blkcipher_walk walk; - unsigned int ubytes; - int err; - - desc->flags &= ~CRYPTO_TFM_REQ_MAY_SLEEP; - blkcipher_walk_init(&walk, dst, src, nbytes); - err = blkcipher_walk_virt(desc, &walk); - - while ((nbytes = walk.nbytes)) { - ubytes = nbytes > MAX_BYTES ? - nbytes - MAX_BYTES : nbytes & (AES_BLOCK_SIZE - 1); - nbytes -= ubytes; - - spe_begin(); - ppc_encrypt_ecb(walk.dst.virt.addr, walk.src.virt.addr, - ctx->key_enc, ctx->rounds, nbytes); - spe_end(); - - err = blkcipher_walk_done(desc, &walk, ubytes); - } - - return err; -} - -static int ppc_ecb_decrypt(struct blkcipher_desc *desc, struct scatterlist *dst, - struct scatterlist *src, unsigned int nbytes) -{ - struct ppc_aes_ctx *ctx = crypto_blkcipher_ctx(desc->tfm); - struct blkcipher_walk walk; - unsigned int ubytes; - int err; - - desc->flags &= ~CRYPTO_TFM_REQ_MAY_SLEEP; - blkcipher_walk_init(&walk, dst, src, nbytes); - err = blkcipher_walk_virt(desc, &walk); - - while ((nbytes = walk.nbytes)) { - ubytes = nbytes > MAX_BYTES ? - nbytes - MAX_BYTES : nbytes & (AES_BLOCK_SIZE - 1); - nbytes -= ubytes; - - spe_begin(); - ppc_decrypt_ecb(walk.dst.virt.addr, walk.src.virt.addr, - ctx->key_dec, ctx->rounds, nbytes); - spe_end(); - - err = blkcipher_walk_done(desc, &walk, ubytes); - } - - return err; -} - -static int ppc_cbc_encrypt(struct blkcipher_desc *desc, struct scatterlist *dst, - struct scatterlist *src, unsigned int nbytes) -{ - struct ppc_aes_ctx *ctx = crypto_blkcipher_ctx(desc->tfm); - struct blkcipher_walk walk; - unsigned int ubytes; - int err; - - desc->flags &= ~CRYPTO_TFM_REQ_MAY_SLEEP; - blkcipher_walk_init(&walk, dst, src, nbytes); - err = blkcipher_walk_virt(desc, &walk); - - while ((nbytes = walk.nbytes)) { - ubytes = nbytes > MAX_BYTES ? - nbytes - MAX_BYTES : nbytes & (AES_BLOCK_SIZE - 1); - nbytes -= ubytes; - - spe_begin(); - ppc_encrypt_cbc(walk.dst.virt.addr, walk.src.virt.addr, - ctx->key_enc, ctx->rounds, nbytes, walk.iv); - spe_end(); - - err = blkcipher_walk_done(desc, &walk, ubytes); - } - - return err; -} - -static int ppc_cbc_decrypt(struct blkcipher_desc *desc, struct scatterlist *dst, - struct scatterlist *src, unsigned int nbytes) -{ - struct ppc_aes_ctx *ctx = crypto_blkcipher_ctx(desc->tfm); - struct blkcipher_walk walk; - unsigned int ubytes; - int err; - - desc->flags &= ~CRYPTO_TFM_REQ_MAY_SLEEP; - blkcipher_walk_init(&walk, dst, src, nbytes); - err = blkcipher_walk_virt(desc, &walk); - - while ((nbytes = walk.nbytes)) { - ubytes = nbytes > MAX_BYTES ? - nbytes - MAX_BYTES : nbytes & (AES_BLOCK_SIZE - 1); - nbytes -= ubytes; - - spe_begin(); - ppc_decrypt_cbc(walk.dst.virt.addr, walk.src.virt.addr, - ctx->key_dec, ctx->rounds, nbytes, walk.iv); - spe_end(); - - err = blkcipher_walk_done(desc, &walk, ubytes); - } - - return err; -} - -static int ppc_ctr_crypt(struct blkcipher_desc *desc, struct scatterlist *dst, - struct scatterlist *src, unsigned int nbytes) -{ - struct ppc_aes_ctx *ctx = crypto_blkcipher_ctx(desc->tfm); - struct blkcipher_walk walk; - unsigned int pbytes, ubytes; - int err; - - desc->flags &= ~CRYPTO_TFM_REQ_MAY_SLEEP; - blkcipher_walk_init(&walk, dst, src, nbytes); - err = blkcipher_walk_virt_block(desc, &walk, AES_BLOCK_SIZE); - - while ((pbytes = walk.nbytes)) { - pbytes = pbytes > MAX_BYTES ? MAX_BYTES : pbytes; - pbytes = pbytes == nbytes ? - nbytes : pbytes & ~(AES_BLOCK_SIZE - 1); - ubytes = walk.nbytes - pbytes; - - spe_begin(); - ppc_crypt_ctr(walk.dst.virt.addr, walk.src.virt.addr, - ctx->key_enc, ctx->rounds, pbytes , walk.iv); - spe_end(); - - nbytes -= pbytes; - err = blkcipher_walk_done(desc, &walk, ubytes); - } - - return err; -} - -static int ppc_xts_encrypt(struct blkcipher_desc *desc, struct scatterlist *dst, - struct scatterlist *src, unsigned int nbytes) -{ - struct ppc_xts_ctx *ctx = crypto_blkcipher_ctx(desc->tfm); - struct blkcipher_walk walk; - unsigned int ubytes; - int err; - u32 *twk; - - desc->flags &= ~CRYPTO_TFM_REQ_MAY_SLEEP; - blkcipher_walk_init(&walk, dst, src, nbytes); - err = blkcipher_walk_virt(desc, &walk); - twk = ctx->key_twk; - - while ((nbytes = walk.nbytes)) { - ubytes = nbytes > MAX_BYTES ? - nbytes - MAX_BYTES : nbytes & (AES_BLOCK_SIZE - 1); - nbytes -= ubytes; - - spe_begin(); - ppc_encrypt_xts(walk.dst.virt.addr, walk.src.virt.addr, - ctx->key_enc, ctx->rounds, nbytes, walk.iv, twk); - spe_end(); - - twk = NULL; - err = blkcipher_walk_done(desc, &walk, ubytes); - } - - return err; -} - -static int ppc_xts_decrypt(struct blkcipher_desc *desc, struct scatterlist *dst, - struct scatterlist *src, unsigned int nbytes) -{ - struct ppc_xts_ctx *ctx = crypto_blkcipher_ctx(desc->tfm); - struct blkcipher_walk walk; - unsigned int ubytes; - int err; - u32 *twk; - - desc->flags &= ~CRYPTO_TFM_REQ_MAY_SLEEP; - blkcipher_walk_init(&walk, dst, src, nbytes); - err = blkcipher_walk_virt(desc, &walk); - twk = ctx->key_twk; - - while ((nbytes = walk.nbytes)) { - ubytes = nbytes > MAX_BYTES ? - nbytes - MAX_BYTES : nbytes & (AES_BLOCK_SIZE - 1); - nbytes -= ubytes; - - spe_begin(); - ppc_decrypt_xts(walk.dst.virt.addr, walk.src.virt.addr, - ctx->key_dec, ctx->rounds, nbytes, walk.iv, twk); - spe_end(); - - twk = NULL; - err = blkcipher_walk_done(desc, &walk, ubytes); - } - - return err; -} - -/* - * Algorithm definitions. Disabling alignment (cra_alignmask=0) was chosen - * because the e500 platform can handle unaligned reads/writes very efficently. - * This improves IPsec thoughput by another few percent. Additionally we assume - * that AES context is always aligned to at least 8 bytes because it is created - * with kmalloc() in the crypto infrastructure - * - */ -static struct crypto_alg aes_algs[] = { { - .cra_name = "aes", - .cra_driver_name = "aes-ppc-spe", - .cra_priority = 300, - .cra_flags = CRYPTO_ALG_TYPE_CIPHER, - .cra_blocksize = AES_BLOCK_SIZE, - .cra_ctxsize = sizeof(struct ppc_aes_ctx), - .cra_alignmask = 0, - .cra_module = THIS_MODULE, - .cra_u = { - .cipher = { - .cia_min_keysize = AES_MIN_KEY_SIZE, - .cia_max_keysize = AES_MAX_KEY_SIZE, - .cia_setkey = ppc_aes_setkey, - .cia_encrypt = ppc_aes_encrypt, - .cia_decrypt = ppc_aes_decrypt - } - } -}, { - .cra_name = "ecb(aes)", - .cra_driver_name = "ecb-ppc-spe", - .cra_priority = 300, - .cra_flags = CRYPTO_ALG_TYPE_BLKCIPHER, - .cra_blocksize = AES_BLOCK_SIZE, - .cra_ctxsize = sizeof(struct ppc_aes_ctx), - .cra_alignmask = 0, - .cra_type = &crypto_blkcipher_type, - .cra_module = THIS_MODULE, - .cra_u = { - .blkcipher = { - .min_keysize = AES_MIN_KEY_SIZE, - .max_keysize = AES_MAX_KEY_SIZE, - .ivsize = AES_BLOCK_SIZE, - .setkey = ppc_aes_setkey, - .encrypt = ppc_ecb_encrypt, - .decrypt = ppc_ecb_decrypt, - } - } -}, { - .cra_name = "cbc(aes)", - .cra_driver_name = "cbc-ppc-spe", - .cra_priority = 300, - .cra_flags = CRYPTO_ALG_TYPE_BLKCIPHER, - .cra_blocksize = AES_BLOCK_SIZE, - .cra_ctxsize = sizeof(struct ppc_aes_ctx), - .cra_alignmask = 0, - .cra_type = &crypto_blkcipher_type, - .cra_module = THIS_MODULE, - .cra_u = { - .blkcipher = { - .min_keysize = AES_MIN_KEY_SIZE, - .max_keysize = AES_MAX_KEY_SIZE, - .ivsize = AES_BLOCK_SIZE, - .setkey = ppc_aes_setkey, - .encrypt = ppc_cbc_encrypt, - .decrypt = ppc_cbc_decrypt, - } - } -}, { - .cra_name = "ctr(aes)", - .cra_driver_name = "ctr-ppc-spe", - .cra_priority = 300, - .cra_flags = CRYPTO_ALG_TYPE_BLKCIPHER, - .cra_blocksize = 1, - .cra_ctxsize = sizeof(struct ppc_aes_ctx), - .cra_alignmask = 0, - .cra_type = &crypto_blkcipher_type, - .cra_module = THIS_MODULE, - .cra_u = { - .blkcipher = { - .min_keysize = AES_MIN_KEY_SIZE, - .max_keysize = AES_MAX_KEY_SIZE, - .ivsize = AES_BLOCK_SIZE, - .setkey = ppc_aes_setkey, - .encrypt = ppc_ctr_crypt, - .decrypt = ppc_ctr_crypt, - } - } -}, { - .cra_name = "xts(aes)", - .cra_driver_name = "xts-ppc-spe", - .cra_priority = 300, - .cra_flags = CRYPTO_ALG_TYPE_BLKCIPHER, - .cra_blocksize = AES_BLOCK_SIZE, - .cra_ctxsize = sizeof(struct ppc_xts_ctx), - .cra_alignmask = 0, - .cra_type = &crypto_blkcipher_type, - .cra_module = THIS_MODULE, - .cra_u = { - .blkcipher = { - .min_keysize = AES_MIN_KEY_SIZE * 2, - .max_keysize = AES_MAX_KEY_SIZE * 2, - .ivsize = AES_BLOCK_SIZE, - .setkey = ppc_xts_setkey, - .encrypt = ppc_xts_encrypt, - .decrypt = ppc_xts_decrypt, - } - } -} }; - -static int __init ppc_aes_mod_init(void) -{ - return crypto_register_algs(aes_algs, ARRAY_SIZE(aes_algs)); -} - -static void __exit ppc_aes_mod_fini(void) -{ - crypto_unregister_algs(aes_algs, ARRAY_SIZE(aes_algs)); -} - -module_init(ppc_aes_mod_init); -module_exit(ppc_aes_mod_fini); - -MODULE_LICENSE("GPL"); -MODULE_DESCRIPTION("AES-ECB/CBC/CTR/XTS, SPE optimized"); - -MODULE_ALIAS_CRYPTO("aes"); -MODULE_ALIAS_CRYPTO("ecb(aes)"); -MODULE_ALIAS_CRYPTO("cbc(aes)"); -MODULE_ALIAS_CRYPTO("ctr(aes)"); -MODULE_ALIAS_CRYPTO("xts(aes)"); -MODULE_ALIAS_CRYPTO("aes-ppc-spe"); diff --git a/arch/powerpc/crypto/md5-glue.c b/arch/powerpc/crypto/md5-glue.c new file mode 100644 index 000000000000..452fb4dc575f --- /dev/null +++ b/arch/powerpc/crypto/md5-glue.c @@ -0,0 +1,165 @@ +/* + * Glue code for MD5 implementation for PPC assembler + * + * Based on generic implementation. + * + * Copyright (c) 2015 Markus Stockhausen + * + * 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 +#include +#include +#include +#include +#include +#include +#include + +extern void ppc_md5_transform(u32 *state, const u8 *src, u32 blocks); + +static inline void ppc_md5_clear_context(struct md5_state *sctx) +{ + int count = sizeof(struct md5_state) >> 2; + u32 *ptr = (u32 *)sctx; + + /* make sure we can clear the fast way */ + BUILD_BUG_ON(sizeof(struct md5_state) % 4); + do { *ptr++ = 0; } while (--count); +} + +static int ppc_md5_init(struct shash_desc *desc) +{ + struct md5_state *sctx = shash_desc_ctx(desc); + + sctx->hash[0] = 0x67452301; + sctx->hash[1] = 0xefcdab89; + sctx->hash[2] = 0x98badcfe; + sctx->hash[3] = 0x10325476; + sctx->byte_count = 0; + + return 0; +} + +static int ppc_md5_update(struct shash_desc *desc, const u8 *data, + unsigned int len) +{ + struct md5_state *sctx = shash_desc_ctx(desc); + const unsigned int offset = sctx->byte_count & 0x3f; + unsigned int avail = 64 - offset; + const u8 *src = data; + + sctx->byte_count += len; + + if (avail > len) { + memcpy((char *)sctx->block + offset, src, len); + return 0; + } + + if (offset) { + memcpy((char *)sctx->block + offset, src, avail); + ppc_md5_transform(sctx->hash, (const u8 *)sctx->block, 1); + len -= avail; + src += avail; + } + + if (len > 63) { + ppc_md5_transform(sctx->hash, src, len >> 6); + src += len & ~0x3f; + len &= 0x3f; + } + + memcpy((char *)sctx->block, src, len); + return 0; +} + +static int ppc_md5_final(struct shash_desc *desc, u8 *out) +{ + struct md5_state *sctx = shash_desc_ctx(desc); + const unsigned int offset = sctx->byte_count & 0x3f; + const u8 *src = (const u8 *)sctx->block; + u8 *p = (u8 *)src + offset; + int padlen = 55 - offset; + __le64 *pbits = (__le64 *)((char *)sctx->block + 56); + __le32 *dst = (__le32 *)out; + + *p++ = 0x80; + + if (padlen < 0) { + memset(p, 0x00, padlen + sizeof (u64)); + ppc_md5_transform(sctx->hash, src, 1); + p = (char *)sctx->block; + padlen = 56; + } + + memset(p, 0, padlen); + *pbits = cpu_to_le64(sctx->byte_count << 3); + ppc_md5_transform(sctx->hash, src, 1); + + dst[0] = cpu_to_le32(sctx->hash[0]); + dst[1] = cpu_to_le32(sctx->hash[1]); + dst[2] = cpu_to_le32(sctx->hash[2]); + dst[3] = cpu_to_le32(sctx->hash[3]); + + ppc_md5_clear_context(sctx); + return 0; +} + +static int ppc_md5_export(struct shash_desc *desc, void *out) +{ + struct md5_state *sctx = shash_desc_ctx(desc); + + memcpy(out, sctx, sizeof(*sctx)); + return 0; +} + +static int ppc_md5_import(struct shash_desc *desc, const void *in) +{ + struct md5_state *sctx = shash_desc_ctx(desc); + + memcpy(sctx, in, sizeof(*sctx)); + return 0; +} + +static struct shash_alg alg = { + .digestsize = MD5_DIGEST_SIZE, + .init = ppc_md5_init, + .update = ppc_md5_update, + .final = ppc_md5_final, + .export = ppc_md5_export, + .import = ppc_md5_import, + .descsize = sizeof(struct md5_state), + .statesize = sizeof(struct md5_state), + .base = { + .cra_name = "md5", + .cra_driver_name= "md5-ppc", + .cra_priority = 200, + .cra_flags = CRYPTO_ALG_TYPE_SHASH, + .cra_blocksize = MD5_HMAC_BLOCK_SIZE, + .cra_module = THIS_MODULE, + } +}; + +static int __init ppc_md5_mod_init(void) +{ + return crypto_register_shash(&alg); +} + +static void __exit ppc_md5_mod_fini(void) +{ + crypto_unregister_shash(&alg); +} + +module_init(ppc_md5_mod_init); +module_exit(ppc_md5_mod_fini); + +MODULE_LICENSE("GPL"); +MODULE_DESCRIPTION("MD5 Secure Hash Algorithm, PPC assembler"); + +MODULE_ALIAS_CRYPTO("md5"); +MODULE_ALIAS_CRYPTO("md5-ppc"); diff --git a/arch/powerpc/crypto/md5_glue.c b/arch/powerpc/crypto/md5_glue.c deleted file mode 100644 index 452fb4dc575f..000000000000 --- a/arch/powerpc/crypto/md5_glue.c +++ /dev/null @@ -1,165 +0,0 @@ -/* - * Glue code for MD5 implementation for PPC assembler - * - * Based on generic implementation. - * - * Copyright (c) 2015 Markus Stockhausen - * - * 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 -#include -#include -#include -#include -#include -#include -#include - -extern void ppc_md5_transform(u32 *state, const u8 *src, u32 blocks); - -static inline void ppc_md5_clear_context(struct md5_state *sctx) -{ - int count = sizeof(struct md5_state) >> 2; - u32 *ptr = (u32 *)sctx; - - /* make sure we can clear the fast way */ - BUILD_BUG_ON(sizeof(struct md5_state) % 4); - do { *ptr++ = 0; } while (--count); -} - -static int ppc_md5_init(struct shash_desc *desc) -{ - struct md5_state *sctx = shash_desc_ctx(desc); - - sctx->hash[0] = 0x67452301; - sctx->hash[1] = 0xefcdab89; - sctx->hash[2] = 0x98badcfe; - sctx->hash[3] = 0x10325476; - sctx->byte_count = 0; - - return 0; -} - -static int ppc_md5_update(struct shash_desc *desc, const u8 *data, - unsigned int len) -{ - struct md5_state *sctx = shash_desc_ctx(desc); - const unsigned int offset = sctx->byte_count & 0x3f; - unsigned int avail = 64 - offset; - const u8 *src = data; - - sctx->byte_count += len; - - if (avail > len) { - memcpy((char *)sctx->block + offset, src, len); - return 0; - } - - if (offset) { - memcpy((char *)sctx->block + offset, src, avail); - ppc_md5_transform(sctx->hash, (const u8 *)sctx->block, 1); - len -= avail; - src += avail; - } - - if (len > 63) { - ppc_md5_transform(sctx->hash, src, len >> 6); - src += len & ~0x3f; - len &= 0x3f; - } - - memcpy((char *)sctx->block, src, len); - return 0; -} - -static int ppc_md5_final(struct shash_desc *desc, u8 *out) -{ - struct md5_state *sctx = shash_desc_ctx(desc); - const unsigned int offset = sctx->byte_count & 0x3f; - const u8 *src = (const u8 *)sctx->block; - u8 *p = (u8 *)src + offset; - int padlen = 55 - offset; - __le64 *pbits = (__le64 *)((char *)sctx->block + 56); - __le32 *dst = (__le32 *)out; - - *p++ = 0x80; - - if (padlen < 0) { - memset(p, 0x00, padlen + sizeof (u64)); - ppc_md5_transform(sctx->hash, src, 1); - p = (char *)sctx->block; - padlen = 56; - } - - memset(p, 0, padlen); - *pbits = cpu_to_le64(sctx->byte_count << 3); - ppc_md5_transform(sctx->hash, src, 1); - - dst[0] = cpu_to_le32(sctx->hash[0]); - dst[1] = cpu_to_le32(sctx->hash[1]); - dst[2] = cpu_to_le32(sctx->hash[2]); - dst[3] = cpu_to_le32(sctx->hash[3]); - - ppc_md5_clear_context(sctx); - return 0; -} - -static int ppc_md5_export(struct shash_desc *desc, void *out) -{ - struct md5_state *sctx = shash_desc_ctx(desc); - - memcpy(out, sctx, sizeof(*sctx)); - return 0; -} - -static int ppc_md5_import(struct shash_desc *desc, const void *in) -{ - struct md5_state *sctx = shash_desc_ctx(desc); - - memcpy(sctx, in, sizeof(*sctx)); - return 0; -} - -static struct shash_alg alg = { - .digestsize = MD5_DIGEST_SIZE, - .init = ppc_md5_init, - .update = ppc_md5_update, - .final = ppc_md5_final, - .export = ppc_md5_export, - .import = ppc_md5_import, - .descsize = sizeof(struct md5_state), - .statesize = sizeof(struct md5_state), - .base = { - .cra_name = "md5", - .cra_driver_name= "md5-ppc", - .cra_priority = 200, - .cra_flags = CRYPTO_ALG_TYPE_SHASH, - .cra_blocksize = MD5_HMAC_BLOCK_SIZE, - .cra_module = THIS_MODULE, - } -}; - -static int __init ppc_md5_mod_init(void) -{ - return crypto_register_shash(&alg); -} - -static void __exit ppc_md5_mod_fini(void) -{ - crypto_unregister_shash(&alg); -} - -module_init(ppc_md5_mod_init); -module_exit(ppc_md5_mod_fini); - -MODULE_LICENSE("GPL"); -MODULE_DESCRIPTION("MD5 Secure Hash Algorithm, PPC assembler"); - -MODULE_ALIAS_CRYPTO("md5"); -MODULE_ALIAS_CRYPTO("md5-ppc"); diff --git a/arch/powerpc/crypto/sha1-spe-asm.S b/arch/powerpc/crypto/sha1-spe-asm.S new file mode 100644 index 000000000000..fcb6cf002889 --- /dev/null +++ b/arch/powerpc/crypto/sha1-spe-asm.S @@ -0,0 +1,299 @@ +/* + * Fast SHA-1 implementation for SPE instruction set (PPC) + * + * This code makes use of the SPE SIMD instruction set as defined in + * http://cache.freescale.com/files/32bit/doc/ref_manual/SPEPIM.pdf + * Implementation is based on optimization guide notes from + * http://cache.freescale.com/files/32bit/doc/app_note/AN2665.pdf + * + * Copyright (c) 2015 Markus Stockhausen + * + * 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 +#include + +#define rHP r3 /* pointer to hash value */ +#define rWP r4 /* pointer to input */ +#define rKP r5 /* pointer to constants */ + +#define rW0 r14 /* 64 bit round words */ +#define rW1 r15 +#define rW2 r16 +#define rW3 r17 +#define rW4 r18 +#define rW5 r19 +#define rW6 r20 +#define rW7 r21 + +#define rH0 r6 /* 32 bit hash values */ +#define rH1 r7 +#define rH2 r8 +#define rH3 r9 +#define rH4 r10 + +#define rT0 r22 /* 64 bit temporary */ +#define rT1 r0 /* 32 bit temporaries */ +#define rT2 r11 +#define rT3 r12 + +#define rK r23 /* 64 bit constant in volatile register */ + +#define LOAD_K01 + +#define LOAD_K11 \ + evlwwsplat rK,0(rKP); + +#define LOAD_K21 \ + evlwwsplat rK,4(rKP); + +#define LOAD_K31 \ + evlwwsplat rK,8(rKP); + +#define LOAD_K41 \ + evlwwsplat rK,12(rKP); + +#define INITIALIZE \ + stwu r1,-128(r1); /* create stack frame */ \ + evstdw r14,8(r1); /* We must save non volatile */ \ + evstdw r15,16(r1); /* registers. Take the chance */ \ + evstdw r16,24(r1); /* and save the SPE part too */ \ + evstdw r17,32(r1); \ + evstdw r18,40(r1); \ + evstdw r19,48(r1); \ + evstdw r20,56(r1); \ + evstdw r21,64(r1); \ + evstdw r22,72(r1); \ + evstdw r23,80(r1); + + +#define FINALIZE \ + evldw r14,8(r1); /* restore SPE registers */ \ + evldw r15,16(r1); \ + evldw r16,24(r1); \ + evldw r17,32(r1); \ + evldw r18,40(r1); \ + evldw r19,48(r1); \ + evldw r20,56(r1); \ + evldw r21,64(r1); \ + evldw r22,72(r1); \ + evldw r23,80(r1); \ + xor r0,r0,r0; \ + stw r0,8(r1); /* Delete sensitive data */ \ + stw r0,16(r1); /* that we might have pushed */ \ + stw r0,24(r1); /* from other context that runs */ \ + stw r0,32(r1); /* the same code. Assume that */ \ + stw r0,40(r1); /* the lower part of the GPRs */ \ + stw r0,48(r1); /* were already overwritten on */ \ + stw r0,56(r1); /* the way down to here */ \ + stw r0,64(r1); \ + stw r0,72(r1); \ + stw r0,80(r1); \ + addi r1,r1,128; /* cleanup stack frame */ + +#ifdef __BIG_ENDIAN__ +#define LOAD_DATA(reg, off) \ + lwz reg,off(rWP); /* load data */ +#define NEXT_BLOCK \ + addi rWP,rWP,64; /* increment per block */ +#else +#define LOAD_DATA(reg, off) \ + lwbrx reg,0,rWP; /* load data */ \ + addi rWP,rWP,4; /* increment per word */ +#define NEXT_BLOCK /* nothing to do */ +#endif + +#define R_00_15(a, b, c, d, e, w0, w1, k, off) \ + LOAD_DATA(w0, off) /* 1: W */ \ + and rT2,b,c; /* 1: F' = B and C */ \ + LOAD_K##k##1 \ + andc rT1,d,b; /* 1: F" = ~B and D */ \ + rotrwi rT0,a,27; /* 1: A' = A rotl 5 */ \ + or rT2,rT2,rT1; /* 1: F = F' or F" */ \ + add e,e,rT0; /* 1: E = E + A' */ \ + rotrwi b,b,2; /* 1: B = B rotl 30 */ \ + add e,e,w0; /* 1: E = E + W */ \ + LOAD_DATA(w1, off+4) /* 2: W */ \ + add e,e,rT2; /* 1: E = E + F */ \ + and rT1,a,b; /* 2: F' = B and C */ \ + add e,e,rK; /* 1: E = E + K */ \ + andc rT2,c,a; /* 2: F" = ~B and D */ \ + add d,d,rK; /* 2: E = E + K */ \ + or rT2,rT2,rT1; /* 2: F = F' or F" */ \ + rotrwi rT0,e,27; /* 2: A' = A rotl 5 */ \ + add d,d,w1; /* 2: E = E + W */ \ + rotrwi a,a,2; /* 2: B = B rotl 30 */ \ + add d,d,rT0; /* 2: E = E + A' */ \ + evmergelo w1,w1,w0; /* mix W[0]/W[1] */ \ + add d,d,rT2 /* 2: E = E + F */ + +#define R_16_19(a, b, c, d, e, w0, w1, w4, w6, w7, k) \ + and rT2,b,c; /* 1: F' = B and C */ \ + evmergelohi rT0,w7,w6; /* W[-3] */ \ + andc rT1,d,b; /* 1: F" = ~B and D */ \ + evxor w0,w0,rT0; /* W = W[-16] xor W[-3] */ \ + or rT1,rT1,rT2; /* 1: F = F' or F" */ \ + evxor w0,w0,w4; /* W = W xor W[-8] */ \ + add e,e,rT1; /* 1: E = E + F */ \ + evxor w0,w0,w1; /* W = W xor W[-14] */ \ + rotrwi rT2,a,27; /* 1: A' = A rotl 5 */ \ + evrlwi w0,w0,1; /* W = W rotl 1 */ \ + add e,e,rT2; /* 1: E = E + A' */ \ + evaddw rT0,w0,rK; /* WK = W + K */ \ + rotrwi b,b,2; /* 1: B = B rotl 30 */ \ + LOAD_K##k##1 \ + evmergehi rT1,rT1,rT0; /* WK1/WK2 */ \ + add e,e,rT0; /* 1: E = E + WK */ \ + add d,d,rT1; /* 2: E = E + WK */ \ + and rT2,a,b; /* 2: F' = B and C */ \ + andc rT1,c,a; /* 2: F" = ~B and D */ \ + rotrwi rT0,e,27; /* 2: A' = A rotl 5 */ \ + or rT1,rT1,rT2; /* 2: F = F' or F" */ \ + add d,d,rT0; /* 2: E = E + A' */ \ + rotrwi a,a,2; /* 2: B = B rotl 30 */ \ + add d,d,rT1 /* 2: E = E + F */ + +#define R_20_39(a, b, c, d, e, w0, w1, w4, w6, w7, k) \ + evmergelohi rT0,w7,w6; /* W[-3] */ \ + xor rT2,b,c; /* 1: F' = B xor C */ \ + evxor w0,w0,rT0; /* W = W[-16] xor W[-3] */ \ + xor rT2,rT2,d; /* 1: F = F' xor D */ \ + evxor w0,w0,w4; /* W = W xor W[-8] */ \ + add e,e,rT2; /* 1: E = E + F */ \ + evxor w0,w0,w1; /* W = W xor W[-14] */ \ + rotrwi rT2,a,27; /* 1: A' = A rotl 5 */ \ + evrlwi w0,w0,1; /* W = W rotl 1 */ \ + add e,e,rT2; /* 1: E = E + A' */ \ + evaddw rT0,w0,rK; /* WK = W + K */ \ + rotrwi b,b,2; /* 1: B = B rotl 30 */ \ + LOAD_K##k##1 \ + evmergehi rT1,rT1,rT0; /* WK1/WK2 */ \ + add e,e,rT0; /* 1: E = E + WK */ \ + xor rT2,a,b; /* 2: F' = B xor C */ \ + add d,d,rT1; /* 2: E = E + WK */ \ + xor rT2,rT2,c; /* 2: F = F' xor D */ \ + rotrwi rT0,e,27; /* 2: A' = A rotl 5 */ \ + add d,d,rT2; /* 2: E = E + F */ \ + rotrwi a,a,2; /* 2: B = B rotl 30 */ \ + add d,d,rT0 /* 2: E = E + A' */ + +#define R_40_59(a, b, c, d, e, w0, w1, w4, w6, w7, k) \ + and rT2,b,c; /* 1: F' = B and C */ \ + evmergelohi rT0,w7,w6; /* W[-3] */ \ + or rT1,b,c; /* 1: F" = B or C */ \ + evxor w0,w0,rT0; /* W = W[-16] xor W[-3] */ \ + and rT1,d,rT1; /* 1: F" = F" and D */ \ + evxor w0,w0,w4; /* W = W xor W[-8] */ \ + or rT2,rT2,rT1; /* 1: F = F' or F" */ \ + evxor w0,w0,w1; /* W = W xor W[-14] */ \ + add e,e,rT2; /* 1: E = E + F */ \ + evrlwi w0,w0,1; /* W = W rotl 1 */ \ + rotrwi rT2,a,27; /* 1: A' = A rotl 5 */ \ + evaddw rT0,w0,rK; /* WK = W + K */ \ + add e,e,rT2; /* 1: E = E + A' */ \ + LOAD_K##k##1 \ + evmergehi rT1,rT1,rT0; /* WK1/WK2 */ \ + rotrwi b,b,2; /* 1: B = B rotl 30 */ \ + add e,e,rT0; /* 1: E = E + WK */ \ + and rT2,a,b; /* 2: F' = B and C */ \ + or rT0,a,b; /* 2: F" = B or C */ \ + add d,d,rT1; /* 2: E = E + WK */ \ + and rT0,c,rT0; /* 2: F" = F" and D */ \ + rotrwi a,a,2; /* 2: B = B rotl 30 */ \ + or rT2,rT2,rT0; /* 2: F = F' or F" */ \ + rotrwi rT0,e,27; /* 2: A' = A rotl 5 */ \ + add d,d,rT2; /* 2: E = E + F */ \ + add d,d,rT0 /* 2: E = E + A' */ + +#define R_60_79(a, b, c, d, e, w0, w1, w4, w6, w7, k) \ + R_20_39(a, b, c, d, e, w0, w1, w4, w6, w7, k) + +_GLOBAL(ppc_spe_sha1_transform) + INITIALIZE + + lwz rH0,0(rHP) + lwz rH1,4(rHP) + mtctr r5 + lwz rH2,8(rHP) + lis rKP,PPC_SPE_SHA1_K@h + lwz rH3,12(rHP) + ori rKP,rKP,PPC_SPE_SHA1_K@l + lwz rH4,16(rHP) + +ppc_spe_sha1_main: + R_00_15(rH0, rH1, rH2, rH3, rH4, rW1, rW0, 1, 0) + R_00_15(rH3, rH4, rH0, rH1, rH2, rW2, rW1, 0, 8) + R_00_15(rH1, rH2, rH3, rH4, rH0, rW3, rW2, 0, 16) + R_00_15(rH4, rH0, rH1, rH2, rH3, rW4, rW3, 0, 24) + R_00_15(rH2, rH3, rH4, rH0, rH1, rW5, rW4, 0, 32) + R_00_15(rH0, rH1, rH2, rH3, rH4, rW6, rW5, 0, 40) + R_00_15(rH3, rH4, rH0, rH1, rH2, rT3, rW6, 0, 48) + R_00_15(rH1, rH2, rH3, rH4, rH0, rT3, rW7, 0, 56) + + R_16_19(rH4, rH0, rH1, rH2, rH3, rW0, rW1, rW4, rW6, rW7, 0) + R_16_19(rH2, rH3, rH4, rH0, rH1, rW1, rW2, rW5, rW7, rW0, 2) + + R_20_39(rH0, rH1, rH2, rH3, rH4, rW2, rW3, rW6, rW0, rW1, 0) + R_20_39(rH3, rH4, rH0, rH1, rH2, rW3, rW4, rW7, rW1, rW2, 0) + R_20_39(rH1, rH2, rH3, rH4, rH0, rW4, rW5, rW0, rW2, rW3, 0) + R_20_39(rH4, rH0, rH1, rH2, rH3, rW5, rW6, rW1, rW3, rW4, 0) + R_20_39(rH2, rH3, rH4, rH0, rH1, rW6, rW7, rW2, rW4, rW5, 0) + R_20_39(rH0, rH1, rH2, rH3, rH4, rW7, rW0, rW3, rW5, rW6, 0) + R_20_39(rH3, rH4, rH0, rH1, rH2, rW0, rW1, rW4, rW6, rW7, 0) + R_20_39(rH1, rH2, rH3, rH4, rH0, rW1, rW2, rW5, rW7, rW0, 0) + R_20_39(rH4, rH0, rH1, rH2, rH3, rW2, rW3, rW6, rW0, rW1, 0) + R_20_39(rH2, rH3, rH4, rH0, rH1, rW3, rW4, rW7, rW1, rW2, 3) + + R_40_59(rH0, rH1, rH2, rH3, rH4, rW4, rW5, rW0, rW2, rW3, 0) + R_40_59(rH3, rH4, rH0, rH1, rH2, rW5, rW6, rW1, rW3, rW4, 0) + R_40_59(rH1, rH2, rH3, rH4, rH0, rW6, rW7, rW2, rW4, rW5, 0) + R_40_59(rH4, rH0, rH1, rH2, rH3, rW7, rW0, rW3, rW5, rW6, 0) + R_40_59(rH2, rH3, rH4, rH0, rH1, rW0, rW1, rW4, rW6, rW7, 0) + R_40_59(rH0, rH1, rH2, rH3, rH4, rW1, rW2, rW5, rW7, rW0, 0) + R_40_59(rH3, rH4, rH0, rH1, rH2, rW2, rW3, rW6, rW0, rW1, 0) + R_40_59(rH1, rH2, rH3, rH4, rH0, rW3, rW4, rW7, rW1, rW2, 0) + R_40_59(rH4, rH0, rH1, rH2, rH3, rW4, rW5, rW0, rW2, rW3, 0) + R_40_59(rH2, rH3, rH4, rH0, rH1, rW5, rW6, rW1, rW3, rW4, 4) + + R_60_79(rH0, rH1, rH2, rH3, rH4, rW6, rW7, rW2, rW4, rW5, 0) + R_60_79(rH3, rH4, rH0, rH1, rH2, rW7, rW0, rW3, rW5, rW6, 0) + R_60_79(rH1, rH2, rH3, rH4, rH0, rW0, rW1, rW4, rW6, rW7, 0) + R_60_79(rH4, rH0, rH1, rH2, rH3, rW1, rW2, rW5, rW7, rW0, 0) + R_60_79(rH2, rH3, rH4, rH0, rH1, rW2, rW3, rW6, rW0, rW1, 0) + R_60_79(rH0, rH1, rH2, rH3, rH4, rW3, rW4, rW7, rW1, rW2, 0) + R_60_79(rH3, rH4, rH0, rH1, rH2, rW4, rW5, rW0, rW2, rW3, 0) + lwz rT3,0(rHP) + R_60_79(rH1, rH2, rH3, rH4, rH0, rW5, rW6, rW1, rW3, rW4, 0) + lwz rW1,4(rHP) + R_60_79(rH4, rH0, rH1, rH2, rH3, rW6, rW7, rW2, rW4, rW5, 0) + lwz rW2,8(rHP) + R_60_79(rH2, rH3, rH4, rH0, rH1, rW7, rW0, rW3, rW5, rW6, 0) + lwz rW3,12(rHP) + NEXT_BLOCK + lwz rW4,16(rHP) + + add rH0,rH0,rT3 + stw rH0,0(rHP) + add rH1,rH1,rW1 + stw rH1,4(rHP) + add rH2,rH2,rW2 + stw rH2,8(rHP) + add rH3,rH3,rW3 + stw rH3,12(rHP) + add rH4,rH4,rW4 + stw rH4,16(rHP) + + bdnz ppc_spe_sha1_main + + FINALIZE + blr + +.data +.align 4 +PPC_SPE_SHA1_K: + .long 0x5A827999,0x6ED9EBA1,0x8F1BBCDC,0xCA62C1D6 diff --git a/arch/powerpc/crypto/sha1-spe-glue.c b/arch/powerpc/crypto/sha1-spe-glue.c new file mode 100644 index 000000000000..3e1d22212521 --- /dev/null +++ b/arch/powerpc/crypto/sha1-spe-glue.c @@ -0,0 +1,210 @@ +/* + * Glue code for SHA-1 implementation for SPE instructions (PPC) + * + * Based on generic implementation. + * + * Copyright (c) 2015 Markus Stockhausen + * + * 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 +#include +#include +#include +#include +#include +#include +#include +#include +#include + +/* + * MAX_BYTES defines the number of bytes that are allowed to be processed + * between preempt_disable() and preempt_enable(). SHA1 takes ~1000 + * operations per 64 bytes. e500 cores can issue two arithmetic instructions + * per clock cycle using one 32/64 bit unit (SU1) and one 32 bit unit (SU2). + * Thus 2KB of input data will need an estimated maximum of 18,000 cycles. + * Headroom for cache misses included. Even with the low end model clocked + * at 667 MHz this equals to a critical time window of less than 27us. + * + */ +#define MAX_BYTES 2048 + +extern void ppc_spe_sha1_transform(u32 *state, const u8 *src, u32 blocks); + +static void spe_begin(void) +{ + /* We just start SPE operations and will save SPE registers later. */ + preempt_disable(); + enable_kernel_spe(); +} + +static void spe_end(void) +{ + /* reenable preemption */ + preempt_enable(); +} + +static inline void ppc_sha1_clear_context(struct sha1_state *sctx) +{ + int count = sizeof(struct sha1_state) >> 2; + u32 *ptr = (u32 *)sctx; + + /* make sure we can clear the fast way */ + BUILD_BUG_ON(sizeof(struct sha1_state) % 4); + do { *ptr++ = 0; } while (--count); +} + +static int ppc_spe_sha1_init(struct shash_desc *desc) +{ + struct sha1_state *sctx = shash_desc_ctx(desc); + + sctx->state[0] = SHA1_H0; + sctx->state[1] = SHA1_H1; + sctx->state[2] = SHA1_H2; + sctx->state[3] = SHA1_H3; + sctx->state[4] = SHA1_H4; + sctx->count = 0; + + return 0; +} + +static int ppc_spe_sha1_update(struct shash_desc *desc, const u8 *data, + unsigned int len) +{ + struct sha1_state *sctx = shash_desc_ctx(desc); + const unsigned int offset = sctx->count & 0x3f; + const unsigned int avail = 64 - offset; + unsigned int bytes; + const u8 *src = data; + + if (avail > len) { + sctx->count += len; + memcpy((char *)sctx->buffer + offset, src, len); + return 0; + } + + sctx->count += len; + + if (offset) { + memcpy((char *)sctx->buffer + offset, src, avail); + + spe_begin(); + ppc_spe_sha1_transform(sctx->state, (const u8 *)sctx->buffer, 1); + spe_end(); + + len -= avail; + src += avail; + } + + while (len > 63) { + bytes = (len > MAX_BYTES) ? MAX_BYTES : len; + bytes = bytes & ~0x3f; + + spe_begin(); + ppc_spe_sha1_transform(sctx->state, src, bytes >> 6); + spe_end(); + + src += bytes; + len -= bytes; + }; + + memcpy((char *)sctx->buffer, src, len); + return 0; +} + +static int ppc_spe_sha1_final(struct shash_desc *desc, u8 *out) +{ + struct sha1_state *sctx = shash_desc_ctx(desc); + const unsigned int offset = sctx->count & 0x3f; + char *p = (char *)sctx->buffer + offset; + int padlen; + __be64 *pbits = (__be64 *)(((char *)&sctx->buffer) + 56); + __be32 *dst = (__be32 *)out; + + padlen = 55 - offset; + *p++ = 0x80; + + spe_begin(); + + if (padlen < 0) { + memset(p, 0x00, padlen + sizeof (u64)); + ppc_spe_sha1_transform(sctx->state, sctx->buffer, 1); + p = (char *)sctx->buffer; + padlen = 56; + } + + memset(p, 0, padlen); + *pbits = cpu_to_be64(sctx->count << 3); + ppc_spe_sha1_transform(sctx->state, sctx->buffer, 1); + + spe_end(); + + dst[0] = cpu_to_be32(sctx->state[0]); + dst[1] = cpu_to_be32(sctx->state[1]); + dst[2] = cpu_to_be32(sctx->state[2]); + dst[3] = cpu_to_be32(sctx->state[3]); + dst[4] = cpu_to_be32(sctx->state[4]); + + ppc_sha1_clear_context(sctx); + return 0; +} + +static int ppc_spe_sha1_export(struct shash_desc *desc, void *out) +{ + struct sha1_state *sctx = shash_desc_ctx(desc); + + memcpy(out, sctx, sizeof(*sctx)); + return 0; +} + +static int ppc_spe_sha1_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 = ppc_spe_sha1_init, + .update = ppc_spe_sha1_update, + .final = ppc_spe_sha1_final, + .export = ppc_spe_sha1_export, + .import = ppc_spe_sha1_import, + .descsize = sizeof(struct sha1_state), + .statesize = sizeof(struct sha1_state), + .base = { + .cra_name = "sha1", + .cra_driver_name= "sha1-ppc-spe", + .cra_priority = 300, + .cra_flags = CRYPTO_ALG_TYPE_SHASH, + .cra_blocksize = SHA1_BLOCK_SIZE, + .cra_module = THIS_MODULE, + } +}; + +static int __init ppc_spe_sha1_mod_init(void) +{ + return crypto_register_shash(&alg); +} + +static void __exit ppc_spe_sha1_mod_fini(void) +{ + crypto_unregister_shash(&alg); +} + +module_init(ppc_spe_sha1_mod_init); +module_exit(ppc_spe_sha1_mod_fini); + +MODULE_LICENSE("GPL"); +MODULE_DESCRIPTION("SHA1 Secure Hash Algorithm, SPE optimized"); + +MODULE_ALIAS_CRYPTO("sha1"); +MODULE_ALIAS_CRYPTO("sha1-ppc-spe"); diff --git a/arch/powerpc/crypto/sha256-spe-glue.c b/arch/powerpc/crypto/sha256-spe-glue.c new file mode 100644 index 000000000000..f4a616fe1a82 --- /dev/null +++ b/arch/powerpc/crypto/sha256-spe-glue.c @@ -0,0 +1,275 @@ +/* + * Glue code for SHA-256 implementation for SPE instructions (PPC) + * + * Based on generic implementation. The assembler module takes care + * about the SPE registers so it can run from interrupt context. + * + * Copyright (c) 2015 Markus Stockhausen + * + * 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 +#include +#include +#include +#include +#include +#include +#include +#include +#include + +/* + * MAX_BYTES defines the number of bytes that are allowed to be processed + * between preempt_disable() and preempt_enable(). SHA256 takes ~2,000 + * operations per 64 bytes. e500 cores can issue two arithmetic instructions + * per clock cycle using one 32/64 bit unit (SU1) and one 32 bit unit (SU2). + * Thus 1KB of input data will need an estimated maximum of 18,000 cycles. + * Headroom for cache misses included. Even with the low end model clocked + * at 667 MHz this equals to a critical time window of less than 27us. + * + */ +#define MAX_BYTES 1024 + +extern void ppc_spe_sha256_transform(u32 *state, const u8 *src, u32 blocks); + +static void spe_begin(void) +{ + /* We just start SPE operations and will save SPE registers later. */ + preempt_disable(); + enable_kernel_spe(); +} + +static void spe_end(void) +{ + /* reenable preemption */ + preempt_enable(); +} + +static inline void ppc_sha256_clear_context(struct sha256_state *sctx) +{ + int count = sizeof(struct sha256_state) >> 2; + u32 *ptr = (u32 *)sctx; + + /* make sure we can clear the fast way */ + BUILD_BUG_ON(sizeof(struct sha256_state) % 4); + do { *ptr++ = 0; } while (--count); +} + +static int ppc_spe_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; +} + +static int ppc_spe_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; +} + +static int ppc_spe_sha256_update(struct shash_desc *desc, const u8 *data, + unsigned int len) +{ + struct sha256_state *sctx = shash_desc_ctx(desc); + const unsigned int offset = sctx->count & 0x3f; + const unsigned int avail = 64 - offset; + unsigned int bytes; + const u8 *src = data; + + if (avail > len) { + sctx->count += len; + memcpy((char *)sctx->buf + offset, src, len); + return 0; + } + + sctx->count += len; + + if (offset) { + memcpy((char *)sctx->buf + offset, src, avail); + + spe_begin(); + ppc_spe_sha256_transform(sctx->state, (const u8 *)sctx->buf, 1); + spe_end(); + + len -= avail; + src += avail; + } + + while (len > 63) { + /* cut input data into smaller blocks */ + bytes = (len > MAX_BYTES) ? MAX_BYTES : len; + bytes = bytes & ~0x3f; + + spe_begin(); + ppc_spe_sha256_transform(sctx->state, src, bytes >> 6); + spe_end(); + + src += bytes; + len -= bytes; + }; + + memcpy((char *)sctx->buf, src, len); + return 0; +} + +static int ppc_spe_sha256_final(struct shash_desc *desc, u8 *out) +{ + struct sha256_state *sctx = shash_desc_ctx(desc); + const unsigned int offset = sctx->count & 0x3f; + char *p = (char *)sctx->buf + offset; + int padlen; + __be64 *pbits = (__be64 *)(((char *)&sctx->buf) + 56); + __be32 *dst = (__be32 *)out; + + padlen = 55 - offset; + *p++ = 0x80; + + spe_begin(); + + if (padlen < 0) { + memset(p, 0x00, padlen + sizeof (u64)); + ppc_spe_sha256_transform(sctx->state, sctx->buf, 1); + p = (char *)sctx->buf; + padlen = 56; + } + + memset(p, 0, padlen); + *pbits = cpu_to_be64(sctx->count << 3); + ppc_spe_sha256_transform(sctx->state, sctx->buf, 1); + + spe_end(); + + dst[0] = cpu_to_be32(sctx->state[0]); + dst[1] = cpu_to_be32(sctx->state[1]); + dst[2] = cpu_to_be32(sctx->state[2]); + dst[3] = cpu_to_be32(sctx->state[3]); + dst[4] = cpu_to_be32(sctx->state[4]); + dst[5] = cpu_to_be32(sctx->state[5]); + dst[6] = cpu_to_be32(sctx->state[6]); + dst[7] = cpu_to_be32(sctx->state[7]); + + ppc_sha256_clear_context(sctx); + return 0; +} + +static int ppc_spe_sha224_final(struct shash_desc *desc, u8 *out) +{ + u32 D[SHA256_DIGEST_SIZE >> 2]; + __be32 *dst = (__be32 *)out; + + ppc_spe_sha256_final(desc, (u8 *)D); + + /* avoid bytewise memcpy */ + dst[0] = D[0]; + dst[1] = D[1]; + dst[2] = D[2]; + dst[3] = D[3]; + dst[4] = D[4]; + dst[5] = D[5]; + dst[6] = D[6]; + + /* clear sensitive data */ + memzero_explicit(D, SHA256_DIGEST_SIZE); + return 0; +} + +static int ppc_spe_sha256_export(struct shash_desc *desc, void *out) +{ + struct sha256_state *sctx = shash_desc_ctx(desc); + + memcpy(out, sctx, sizeof(*sctx)); + return 0; +} + +static int ppc_spe_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[2] = { { + .digestsize = SHA256_DIGEST_SIZE, + .init = ppc_spe_sha256_init, + .update = ppc_spe_sha256_update, + .final = ppc_spe_sha256_final, + .export = ppc_spe_sha256_export, + .import = ppc_spe_sha256_import, + .descsize = sizeof(struct sha256_state), + .statesize = sizeof(struct sha256_state), + .base = { + .cra_name = "sha256", + .cra_driver_name= "sha256-ppc-spe", + .cra_priority = 300, + .cra_flags = CRYPTO_ALG_TYPE_SHASH, + .cra_blocksize = SHA256_BLOCK_SIZE, + .cra_module = THIS_MODULE, + } +}, { + .digestsize = SHA224_DIGEST_SIZE, + .init = ppc_spe_sha224_init, + .update = ppc_spe_sha256_update, + .final = ppc_spe_sha224_final, + .export = ppc_spe_sha256_export, + .import = ppc_spe_sha256_import, + .descsize = sizeof(struct sha256_state), + .statesize = sizeof(struct sha256_state), + .base = { + .cra_name = "sha224", + .cra_driver_name= "sha224-ppc-spe", + .cra_priority = 300, + .cra_flags = CRYPTO_ALG_TYPE_SHASH, + .cra_blocksize = SHA224_BLOCK_SIZE, + .cra_module = THIS_MODULE, + } +} }; + +static int __init ppc_spe_sha256_mod_init(void) +{ + return crypto_register_shashes(algs, ARRAY_SIZE(algs)); +} + +static void __exit ppc_spe_sha256_mod_fini(void) +{ + crypto_unregister_shashes(algs, ARRAY_SIZE(algs)); +} + +module_init(ppc_spe_sha256_mod_init); +module_exit(ppc_spe_sha256_mod_fini); + +MODULE_LICENSE("GPL"); +MODULE_DESCRIPTION("SHA-224 and SHA-256 Secure Hash Algorithm, SPE optimized"); + +MODULE_ALIAS_CRYPTO("sha224"); +MODULE_ALIAS_CRYPTO("sha224-ppc-spe"); +MODULE_ALIAS_CRYPTO("sha256"); +MODULE_ALIAS_CRYPTO("sha256-ppc-spe"); diff --git a/arch/powerpc/crypto/sha256_spe_glue.c b/arch/powerpc/crypto/sha256_spe_glue.c deleted file mode 100644 index f4a616fe1a82..000000000000 --- a/arch/powerpc/crypto/sha256_spe_glue.c +++ /dev/null @@ -1,275 +0,0 @@ -/* - * Glue code for SHA-256 implementation for SPE instructions (PPC) - * - * Based on generic implementation. The assembler module takes care - * about the SPE registers so it can run from interrupt context. - * - * Copyright (c) 2015 Markus Stockhausen - * - * 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 -#include -#include -#include -#include -#include -#include -#include -#include -#include - -/* - * MAX_BYTES defines the number of bytes that are allowed to be processed - * between preempt_disable() and preempt_enable(). SHA256 takes ~2,000 - * operations per 64 bytes. e500 cores can issue two arithmetic instructions - * per clock cycle using one 32/64 bit unit (SU1) and one 32 bit unit (SU2). - * Thus 1KB of input data will need an estimated maximum of 18,000 cycles. - * Headroom for cache misses included. Even with the low end model clocked - * at 667 MHz this equals to a critical time window of less than 27us. - * - */ -#define MAX_BYTES 1024 - -extern void ppc_spe_sha256_transform(u32 *state, const u8 *src, u32 blocks); - -static void spe_begin(void) -{ - /* We just start SPE operations and will save SPE registers later. */ - preempt_disable(); - enable_kernel_spe(); -} - -static void spe_end(void) -{ - /* reenable preemption */ - preempt_enable(); -} - -static inline void ppc_sha256_clear_context(struct sha256_state *sctx) -{ - int count = sizeof(struct sha256_state) >> 2; - u32 *ptr = (u32 *)sctx; - - /* make sure we can clear the fast way */ - BUILD_BUG_ON(sizeof(struct sha256_state) % 4); - do { *ptr++ = 0; } while (--count); -} - -static int ppc_spe_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; -} - -static int ppc_spe_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; -} - -static int ppc_spe_sha256_update(struct shash_desc *desc, const u8 *data, - unsigned int len) -{ - struct sha256_state *sctx = shash_desc_ctx(desc); - const unsigned int offset = sctx->count & 0x3f; - const unsigned int avail = 64 - offset; - unsigned int bytes; - const u8 *src = data; - - if (avail > len) { - sctx->count += len; - memcpy((char *)sctx->buf + offset, src, len); - return 0; - } - - sctx->count += len; - - if (offset) { - memcpy((char *)sctx->buf + offset, src, avail); - - spe_begin(); - ppc_spe_sha256_transform(sctx->state, (const u8 *)sctx->buf, 1); - spe_end(); - - len -= avail; - src += avail; - } - - while (len > 63) { - /* cut input data into smaller blocks */ - bytes = (len > MAX_BYTES) ? MAX_BYTES : len; - bytes = bytes & ~0x3f; - - spe_begin(); - ppc_spe_sha256_transform(sctx->state, src, bytes >> 6); - spe_end(); - - src += bytes; - len -= bytes; - }; - - memcpy((char *)sctx->buf, src, len); - return 0; -} - -static int ppc_spe_sha256_final(struct shash_desc *desc, u8 *out) -{ - struct sha256_state *sctx = shash_desc_ctx(desc); - const unsigned int offset = sctx->count & 0x3f; - char *p = (char *)sctx->buf + offset; - int padlen; - __be64 *pbits = (__be64 *)(((char *)&sctx->buf) + 56); - __be32 *dst = (__be32 *)out; - - padlen = 55 - offset; - *p++ = 0x80; - - spe_begin(); - - if (padlen < 0) { - memset(p, 0x00, padlen + sizeof (u64)); - ppc_spe_sha256_transform(sctx->state, sctx->buf, 1); - p = (char *)sctx->buf; - padlen = 56; - } - - memset(p, 0, padlen); - *pbits = cpu_to_be64(sctx->count << 3); - ppc_spe_sha256_transform(sctx->state, sctx->buf, 1); - - spe_end(); - - dst[0] = cpu_to_be32(sctx->state[0]); - dst[1] = cpu_to_be32(sctx->state[1]); - dst[2] = cpu_to_be32(sctx->state[2]); - dst[3] = cpu_to_be32(sctx->state[3]); - dst[4] = cpu_to_be32(sctx->state[4]); - dst[5] = cpu_to_be32(sctx->state[5]); - dst[6] = cpu_to_be32(sctx->state[6]); - dst[7] = cpu_to_be32(sctx->state[7]); - - ppc_sha256_clear_context(sctx); - return 0; -} - -static int ppc_spe_sha224_final(struct shash_desc *desc, u8 *out) -{ - u32 D[SHA256_DIGEST_SIZE >> 2]; - __be32 *dst = (__be32 *)out; - - ppc_spe_sha256_final(desc, (u8 *)D); - - /* avoid bytewise memcpy */ - dst[0] = D[0]; - dst[1] = D[1]; - dst[2] = D[2]; - dst[3] = D[3]; - dst[4] = D[4]; - dst[5] = D[5]; - dst[6] = D[6]; - - /* clear sensitive data */ - memzero_explicit(D, SHA256_DIGEST_SIZE); - return 0; -} - -static int ppc_spe_sha256_export(struct shash_desc *desc, void *out) -{ - struct sha256_state *sctx = shash_desc_ctx(desc); - - memcpy(out, sctx, sizeof(*sctx)); - return 0; -} - -static int ppc_spe_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[2] = { { - .digestsize = SHA256_DIGEST_SIZE, - .init = ppc_spe_sha256_init, - .update = ppc_spe_sha256_update, - .final = ppc_spe_sha256_final, - .export = ppc_spe_sha256_export, - .import = ppc_spe_sha256_import, - .descsize = sizeof(struct sha256_state), - .statesize = sizeof(struct sha256_state), - .base = { - .cra_name = "sha256", - .cra_driver_name= "sha256-ppc-spe", - .cra_priority = 300, - .cra_flags = CRYPTO_ALG_TYPE_SHASH, - .cra_blocksize = SHA256_BLOCK_SIZE, - .cra_module = THIS_MODULE, - } -}, { - .digestsize = SHA224_DIGEST_SIZE, - .init = ppc_spe_sha224_init, - .update = ppc_spe_sha256_update, - .final = ppc_spe_sha224_final, - .export = ppc_spe_sha256_export, - .import = ppc_spe_sha256_import, - .descsize = sizeof(struct sha256_state), - .statesize = sizeof(struct sha256_state), - .base = { - .cra_name = "sha224", - .cra_driver_name= "sha224-ppc-spe", - .cra_priority = 300, - .cra_flags = CRYPTO_ALG_TYPE_SHASH, - .cra_blocksize = SHA224_BLOCK_SIZE, - .cra_module = THIS_MODULE, - } -} }; - -static int __init ppc_spe_sha256_mod_init(void) -{ - return crypto_register_shashes(algs, ARRAY_SIZE(algs)); -} - -static void __exit ppc_spe_sha256_mod_fini(void) -{ - crypto_unregister_shashes(algs, ARRAY_SIZE(algs)); -} - -module_init(ppc_spe_sha256_mod_init); -module_exit(ppc_spe_sha256_mod_fini); - -MODULE_LICENSE("GPL"); -MODULE_DESCRIPTION("SHA-224 and SHA-256 Secure Hash Algorithm, SPE optimized"); - -MODULE_ALIAS_CRYPTO("sha224"); -MODULE_ALIAS_CRYPTO("sha224-ppc-spe"); -MODULE_ALIAS_CRYPTO("sha256"); -MODULE_ALIAS_CRYPTO("sha256-ppc-spe"); diff --git a/arch/powerpc/sha1-spe-asm.S b/arch/powerpc/sha1-spe-asm.S deleted file mode 100644 index fcb6cf002889..000000000000 --- a/arch/powerpc/sha1-spe-asm.S +++ /dev/null @@ -1,299 +0,0 @@ -/* - * Fast SHA-1 implementation for SPE instruction set (PPC) - * - * This code makes use of the SPE SIMD instruction set as defined in - * http://cache.freescale.com/files/32bit/doc/ref_manual/SPEPIM.pdf - * Implementation is based on optimization guide notes from - * http://cache.freescale.com/files/32bit/doc/app_note/AN2665.pdf - * - * Copyright (c) 2015 Markus Stockhausen - * - * 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 -#include - -#define rHP r3 /* pointer to hash value */ -#define rWP r4 /* pointer to input */ -#define rKP r5 /* pointer to constants */ - -#define rW0 r14 /* 64 bit round words */ -#define rW1 r15 -#define rW2 r16 -#define rW3 r17 -#define rW4 r18 -#define rW5 r19 -#define rW6 r20 -#define rW7 r21 - -#define rH0 r6 /* 32 bit hash values */ -#define rH1 r7 -#define rH2 r8 -#define rH3 r9 -#define rH4 r10 - -#define rT0 r22 /* 64 bit temporary */ -#define rT1 r0 /* 32 bit temporaries */ -#define rT2 r11 -#define rT3 r12 - -#define rK r23 /* 64 bit constant in volatile register */ - -#define LOAD_K01 - -#define LOAD_K11 \ - evlwwsplat rK,0(rKP); - -#define LOAD_K21 \ - evlwwsplat rK,4(rKP); - -#define LOAD_K31 \ - evlwwsplat rK,8(rKP); - -#define LOAD_K41 \ - evlwwsplat rK,12(rKP); - -#define INITIALIZE \ - stwu r1,-128(r1); /* create stack frame */ \ - evstdw r14,8(r1); /* We must save non volatile */ \ - evstdw r15,16(r1); /* registers. Take the chance */ \ - evstdw r16,24(r1); /* and save the SPE part too */ \ - evstdw r17,32(r1); \ - evstdw r18,40(r1); \ - evstdw r19,48(r1); \ - evstdw r20,56(r1); \ - evstdw r21,64(r1); \ - evstdw r22,72(r1); \ - evstdw r23,80(r1); - - -#define FINALIZE \ - evldw r14,8(r1); /* restore SPE registers */ \ - evldw r15,16(r1); \ - evldw r16,24(r1); \ - evldw r17,32(r1); \ - evldw r18,40(r1); \ - evldw r19,48(r1); \ - evldw r20,56(r1); \ - evldw r21,64(r1); \ - evldw r22,72(r1); \ - evldw r23,80(r1); \ - xor r0,r0,r0; \ - stw r0,8(r1); /* Delete sensitive data */ \ - stw r0,16(r1); /* that we might have pushed */ \ - stw r0,24(r1); /* from other context that runs */ \ - stw r0,32(r1); /* the same code. Assume that */ \ - stw r0,40(r1); /* the lower part of the GPRs */ \ - stw r0,48(r1); /* were already overwritten on */ \ - stw r0,56(r1); /* the way down to here */ \ - stw r0,64(r1); \ - stw r0,72(r1); \ - stw r0,80(r1); \ - addi r1,r1,128; /* cleanup stack frame */ - -#ifdef __BIG_ENDIAN__ -#define LOAD_DATA(reg, off) \ - lwz reg,off(rWP); /* load data */ -#define NEXT_BLOCK \ - addi rWP,rWP,64; /* increment per block */ -#else -#define LOAD_DATA(reg, off) \ - lwbrx reg,0,rWP; /* load data */ \ - addi rWP,rWP,4; /* increment per word */ -#define NEXT_BLOCK /* nothing to do */ -#endif - -#define R_00_15(a, b, c, d, e, w0, w1, k, off) \ - LOAD_DATA(w0, off) /* 1: W */ \ - and rT2,b,c; /* 1: F' = B and C */ \ - LOAD_K##k##1 \ - andc rT1,d,b; /* 1: F" = ~B and D */ \ - rotrwi rT0,a,27; /* 1: A' = A rotl 5 */ \ - or rT2,rT2,rT1; /* 1: F = F' or F" */ \ - add e,e,rT0; /* 1: E = E + A' */ \ - rotrwi b,b,2; /* 1: B = B rotl 30 */ \ - add e,e,w0; /* 1: E = E + W */ \ - LOAD_DATA(w1, off+4) /* 2: W */ \ - add e,e,rT2; /* 1: E = E + F */ \ - and rT1,a,b; /* 2: F' = B and C */ \ - add e,e,rK; /* 1: E = E + K */ \ - andc rT2,c,a; /* 2: F" = ~B and D */ \ - add d,d,rK; /* 2: E = E + K */ \ - or rT2,rT2,rT1; /* 2: F = F' or F" */ \ - rotrwi rT0,e,27; /* 2: A' = A rotl 5 */ \ - add d,d,w1; /* 2: E = E + W */ \ - rotrwi a,a,2; /* 2: B = B rotl 30 */ \ - add d,d,rT0; /* 2: E = E + A' */ \ - evmergelo w1,w1,w0; /* mix W[0]/W[1] */ \ - add d,d,rT2 /* 2: E = E + F */ - -#define R_16_19(a, b, c, d, e, w0, w1, w4, w6, w7, k) \ - and rT2,b,c; /* 1: F' = B and C */ \ - evmergelohi rT0,w7,w6; /* W[-3] */ \ - andc rT1,d,b; /* 1: F" = ~B and D */ \ - evxor w0,w0,rT0; /* W = W[-16] xor W[-3] */ \ - or rT1,rT1,rT2; /* 1: F = F' or F" */ \ - evxor w0,w0,w4; /* W = W xor W[-8] */ \ - add e,e,rT1; /* 1: E = E + F */ \ - evxor w0,w0,w1; /* W = W xor W[-14] */ \ - rotrwi rT2,a,27; /* 1: A' = A rotl 5 */ \ - evrlwi w0,w0,1; /* W = W rotl 1 */ \ - add e,e,rT2; /* 1: E = E + A' */ \ - evaddw rT0,w0,rK; /* WK = W + K */ \ - rotrwi b,b,2; /* 1: B = B rotl 30 */ \ - LOAD_K##k##1 \ - evmergehi rT1,rT1,rT0; /* WK1/WK2 */ \ - add e,e,rT0; /* 1: E = E + WK */ \ - add d,d,rT1; /* 2: E = E + WK */ \ - and rT2,a,b; /* 2: F' = B and C */ \ - andc rT1,c,a; /* 2: F" = ~B and D */ \ - rotrwi rT0,e,27; /* 2: A' = A rotl 5 */ \ - or rT1,rT1,rT2; /* 2: F = F' or F" */ \ - add d,d,rT0; /* 2: E = E + A' */ \ - rotrwi a,a,2; /* 2: B = B rotl 30 */ \ - add d,d,rT1 /* 2: E = E + F */ - -#define R_20_39(a, b, c, d, e, w0, w1, w4, w6, w7, k) \ - evmergelohi rT0,w7,w6; /* W[-3] */ \ - xor rT2,b,c; /* 1: F' = B xor C */ \ - evxor w0,w0,rT0; /* W = W[-16] xor W[-3] */ \ - xor rT2,rT2,d; /* 1: F = F' xor D */ \ - evxor w0,w0,w4; /* W = W xor W[-8] */ \ - add e,e,rT2; /* 1: E = E + F */ \ - evxor w0,w0,w1; /* W = W xor W[-14] */ \ - rotrwi rT2,a,27; /* 1: A' = A rotl 5 */ \ - evrlwi w0,w0,1; /* W = W rotl 1 */ \ - add e,e,rT2; /* 1: E = E + A' */ \ - evaddw rT0,w0,rK; /* WK = W + K */ \ - rotrwi b,b,2; /* 1: B = B rotl 30 */ \ - LOAD_K##k##1 \ - evmergehi rT1,rT1,rT0; /* WK1/WK2 */ \ - add e,e,rT0; /* 1: E = E + WK */ \ - xor rT2,a,b; /* 2: F' = B xor C */ \ - add d,d,rT1; /* 2: E = E + WK */ \ - xor rT2,rT2,c; /* 2: F = F' xor D */ \ - rotrwi rT0,e,27; /* 2: A' = A rotl 5 */ \ - add d,d,rT2; /* 2: E = E + F */ \ - rotrwi a,a,2; /* 2: B = B rotl 30 */ \ - add d,d,rT0 /* 2: E = E + A' */ - -#define R_40_59(a, b, c, d, e, w0, w1, w4, w6, w7, k) \ - and rT2,b,c; /* 1: F' = B and C */ \ - evmergelohi rT0,w7,w6; /* W[-3] */ \ - or rT1,b,c; /* 1: F" = B or C */ \ - evxor w0,w0,rT0; /* W = W[-16] xor W[-3] */ \ - and rT1,d,rT1; /* 1: F" = F" and D */ \ - evxor w0,w0,w4; /* W = W xor W[-8] */ \ - or rT2,rT2,rT1; /* 1: F = F' or F" */ \ - evxor w0,w0,w1; /* W = W xor W[-14] */ \ - add e,e,rT2; /* 1: E = E + F */ \ - evrlwi w0,w0,1; /* W = W rotl 1 */ \ - rotrwi rT2,a,27; /* 1: A' = A rotl 5 */ \ - evaddw rT0,w0,rK; /* WK = W + K */ \ - add e,e,rT2; /* 1: E = E + A' */ \ - LOAD_K##k##1 \ - evmergehi rT1,rT1,rT0; /* WK1/WK2 */ \ - rotrwi b,b,2; /* 1: B = B rotl 30 */ \ - add e,e,rT0; /* 1: E = E + WK */ \ - and rT2,a,b; /* 2: F' = B and C */ \ - or rT0,a,b; /* 2: F" = B or C */ \ - add d,d,rT1; /* 2: E = E + WK */ \ - and rT0,c,rT0; /* 2: F" = F" and D */ \ - rotrwi a,a,2; /* 2: B = B rotl 30 */ \ - or rT2,rT2,rT0; /* 2: F = F' or F" */ \ - rotrwi rT0,e,27; /* 2: A' = A rotl 5 */ \ - add d,d,rT2; /* 2: E = E + F */ \ - add d,d,rT0 /* 2: E = E + A' */ - -#define R_60_79(a, b, c, d, e, w0, w1, w4, w6, w7, k) \ - R_20_39(a, b, c, d, e, w0, w1, w4, w6, w7, k) - -_GLOBAL(ppc_spe_sha1_transform) - INITIALIZE - - lwz rH0,0(rHP) - lwz rH1,4(rHP) - mtctr r5 - lwz rH2,8(rHP) - lis rKP,PPC_SPE_SHA1_K@h - lwz rH3,12(rHP) - ori rKP,rKP,PPC_SPE_SHA1_K@l - lwz rH4,16(rHP) - -ppc_spe_sha1_main: - R_00_15(rH0, rH1, rH2, rH3, rH4, rW1, rW0, 1, 0) - R_00_15(rH3, rH4, rH0, rH1, rH2, rW2, rW1, 0, 8) - R_00_15(rH1, rH2, rH3, rH4, rH0, rW3, rW2, 0, 16) - R_00_15(rH4, rH0, rH1, rH2, rH3, rW4, rW3, 0, 24) - R_00_15(rH2, rH3, rH4, rH0, rH1, rW5, rW4, 0, 32) - R_00_15(rH0, rH1, rH2, rH3, rH4, rW6, rW5, 0, 40) - R_00_15(rH3, rH4, rH0, rH1, rH2, rT3, rW6, 0, 48) - R_00_15(rH1, rH2, rH3, rH4, rH0, rT3, rW7, 0, 56) - - R_16_19(rH4, rH0, rH1, rH2, rH3, rW0, rW1, rW4, rW6, rW7, 0) - R_16_19(rH2, rH3, rH4, rH0, rH1, rW1, rW2, rW5, rW7, rW0, 2) - - R_20_39(rH0, rH1, rH2, rH3, rH4, rW2, rW3, rW6, rW0, rW1, 0) - R_20_39(rH3, rH4, rH0, rH1, rH2, rW3, rW4, rW7, rW1, rW2, 0) - R_20_39(rH1, rH2, rH3, rH4, rH0, rW4, rW5, rW0, rW2, rW3, 0) - R_20_39(rH4, rH0, rH1, rH2, rH3, rW5, rW6, rW1, rW3, rW4, 0) - R_20_39(rH2, rH3, rH4, rH0, rH1, rW6, rW7, rW2, rW4, rW5, 0) - R_20_39(rH0, rH1, rH2, rH3, rH4, rW7, rW0, rW3, rW5, rW6, 0) - R_20_39(rH3, rH4, rH0, rH1, rH2, rW0, rW1, rW4, rW6, rW7, 0) - R_20_39(rH1, rH2, rH3, rH4, rH0, rW1, rW2, rW5, rW7, rW0, 0) - R_20_39(rH4, rH0, rH1, rH2, rH3, rW2, rW3, rW6, rW0, rW1, 0) - R_20_39(rH2, rH3, rH4, rH0, rH1, rW3, rW4, rW7, rW1, rW2, 3) - - R_40_59(rH0, rH1, rH2, rH3, rH4, rW4, rW5, rW0, rW2, rW3, 0) - R_40_59(rH3, rH4, rH0, rH1, rH2, rW5, rW6, rW1, rW3, rW4, 0) - R_40_59(rH1, rH2, rH3, rH4, rH0, rW6, rW7, rW2, rW4, rW5, 0) - R_40_59(rH4, rH0, rH1, rH2, rH3, rW7, rW0, rW3, rW5, rW6, 0) - R_40_59(rH2, rH3, rH4, rH0, rH1, rW0, rW1, rW4, rW6, rW7, 0) - R_40_59(rH0, rH1, rH2, rH3, rH4, rW1, rW2, rW5, rW7, rW0, 0) - R_40_59(rH3, rH4, rH0, rH1, rH2, rW2, rW3, rW6, rW0, rW1, 0) - R_40_59(rH1, rH2, rH3, rH4, rH0, rW3, rW4, rW7, rW1, rW2, 0) - R_40_59(rH4, rH0, rH1, rH2, rH3, rW4, rW5, rW0, rW2, rW3, 0) - R_40_59(rH2, rH3, rH4, rH0, rH1, rW5, rW6, rW1, rW3, rW4, 4) - - R_60_79(rH0, rH1, rH2, rH3, rH4, rW6, rW7, rW2, rW4, rW5, 0) - R_60_79(rH3, rH4, rH0, rH1, rH2, rW7, rW0, rW3, rW5, rW6, 0) - R_60_79(rH1, rH2, rH3, rH4, rH0, rW0, rW1, rW4, rW6, rW7, 0) - R_60_79(rH4, rH0, rH1, rH2, rH3, rW1, rW2, rW5, rW7, rW0, 0) - R_60_79(rH2, rH3, rH4, rH0, rH1, rW2, rW3, rW6, rW0, rW1, 0) - R_60_79(rH0, rH1, rH2, rH3, rH4, rW3, rW4, rW7, rW1, rW2, 0) - R_60_79(rH3, rH4, rH0, rH1, rH2, rW4, rW5, rW0, rW2, rW3, 0) - lwz rT3,0(rHP) - R_60_79(rH1, rH2, rH3, rH4, rH0, rW5, rW6, rW1, rW3, rW4, 0) - lwz rW1,4(rHP) - R_60_79(rH4, rH0, rH1, rH2, rH3, rW6, rW7, rW2, rW4, rW5, 0) - lwz rW2,8(rHP) - R_60_79(rH2, rH3, rH4, rH0, rH1, rW7, rW0, rW3, rW5, rW6, 0) - lwz rW3,12(rHP) - NEXT_BLOCK - lwz rW4,16(rHP) - - add rH0,rH0,rT3 - stw rH0,0(rHP) - add rH1,rH1,rW1 - stw rH1,4(rHP) - add rH2,rH2,rW2 - stw rH2,8(rHP) - add rH3,rH3,rW3 - stw rH3,12(rHP) - add rH4,rH4,rW4 - stw rH4,16(rHP) - - bdnz ppc_spe_sha1_main - - FINALIZE - blr - -.data -.align 4 -PPC_SPE_SHA1_K: - .long 0x5A827999,0x6ED9EBA1,0x8F1BBCDC,0xCA62C1D6 diff --git a/arch/powerpc/sha1_spe_glue.c b/arch/powerpc/sha1_spe_glue.c deleted file mode 100644 index 3e1d22212521..000000000000 --- a/arch/powerpc/sha1_spe_glue.c +++ /dev/null @@ -1,210 +0,0 @@ -/* - * Glue code for SHA-1 implementation for SPE instructions (PPC) - * - * Based on generic implementation. - * - * Copyright (c) 2015 Markus Stockhausen - * - * 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 -#include -#include -#include -#include -#include -#include -#include -#include -#include - -/* - * MAX_BYTES defines the number of bytes that are allowed to be processed - * between preempt_disable() and preempt_enable(). SHA1 takes ~1000 - * operations per 64 bytes. e500 cores can issue two arithmetic instructions - * per clock cycle using one 32/64 bit unit (SU1) and one 32 bit unit (SU2). - * Thus 2KB of input data will need an estimated maximum of 18,000 cycles. - * Headroom for cache misses included. Even with the low end model clocked - * at 667 MHz this equals to a critical time window of less than 27us. - * - */ -#define MAX_BYTES 2048 - -extern void ppc_spe_sha1_transform(u32 *state, const u8 *src, u32 blocks); - -static void spe_begin(void) -{ - /* We just start SPE operations and will save SPE registers later. */ - preempt_disable(); - enable_kernel_spe(); -} - -static void spe_end(void) -{ - /* reenable preemption */ - preempt_enable(); -} - -static inline void ppc_sha1_clear_context(struct sha1_state *sctx) -{ - int count = sizeof(struct sha1_state) >> 2; - u32 *ptr = (u32 *)sctx; - - /* make sure we can clear the fast way */ - BUILD_BUG_ON(sizeof(struct sha1_state) % 4); - do { *ptr++ = 0; } while (--count); -} - -static int ppc_spe_sha1_init(struct shash_desc *desc) -{ - struct sha1_state *sctx = shash_desc_ctx(desc); - - sctx->state[0] = SHA1_H0; - sctx->state[1] = SHA1_H1; - sctx->state[2] = SHA1_H2; - sctx->state[3] = SHA1_H3; - sctx->state[4] = SHA1_H4; - sctx->count = 0; - - return 0; -} - -static int ppc_spe_sha1_update(struct shash_desc *desc, const u8 *data, - unsigned int len) -{ - struct sha1_state *sctx = shash_desc_ctx(desc); - const unsigned int offset = sctx->count & 0x3f; - const unsigned int avail = 64 - offset; - unsigned int bytes; - const u8 *src = data; - - if (avail > len) { - sctx->count += len; - memcpy((char *)sctx->buffer + offset, src, len); - return 0; - } - - sctx->count += len; - - if (offset) { - memcpy((char *)sctx->buffer + offset, src, avail); - - spe_begin(); - ppc_spe_sha1_transform(sctx->state, (const u8 *)sctx->buffer, 1); - spe_end(); - - len -= avail; - src += avail; - } - - while (len > 63) { - bytes = (len > MAX_BYTES) ? MAX_BYTES : len; - bytes = bytes & ~0x3f; - - spe_begin(); - ppc_spe_sha1_transform(sctx->state, src, bytes >> 6); - spe_end(); - - src += bytes; - len -= bytes; - }; - - memcpy((char *)sctx->buffer, src, len); - return 0; -} - -static int ppc_spe_sha1_final(struct shash_desc *desc, u8 *out) -{ - struct sha1_state *sctx = shash_desc_ctx(desc); - const unsigned int offset = sctx->count & 0x3f; - char *p = (char *)sctx->buffer + offset; - int padlen; - __be64 *pbits = (__be64 *)(((char *)&sctx->buffer) + 56); - __be32 *dst = (__be32 *)out; - - padlen = 55 - offset; - *p++ = 0x80; - - spe_begin(); - - if (padlen < 0) { - memset(p, 0x00, padlen + sizeof (u64)); - ppc_spe_sha1_transform(sctx->state, sctx->buffer, 1); - p = (char *)sctx->buffer; - padlen = 56; - } - - memset(p, 0, padlen); - *pbits = cpu_to_be64(sctx->count << 3); - ppc_spe_sha1_transform(sctx->state, sctx->buffer, 1); - - spe_end(); - - dst[0] = cpu_to_be32(sctx->state[0]); - dst[1] = cpu_to_be32(sctx->state[1]); - dst[2] = cpu_to_be32(sctx->state[2]); - dst[3] = cpu_to_be32(sctx->state[3]); - dst[4] = cpu_to_be32(sctx->state[4]); - - ppc_sha1_clear_context(sctx); - return 0; -} - -static int ppc_spe_sha1_export(struct shash_desc *desc, void *out) -{ - struct sha1_state *sctx = shash_desc_ctx(desc); - - memcpy(out, sctx, sizeof(*sctx)); - return 0; -} - -static int ppc_spe_sha1_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 = ppc_spe_sha1_init, - .update = ppc_spe_sha1_update, - .final = ppc_spe_sha1_final, - .export = ppc_spe_sha1_export, - .import = ppc_spe_sha1_import, - .descsize = sizeof(struct sha1_state), - .statesize = sizeof(struct sha1_state), - .base = { - .cra_name = "sha1", - .cra_driver_name= "sha1-ppc-spe", - .cra_priority = 300, - .cra_flags = CRYPTO_ALG_TYPE_SHASH, - .cra_blocksize = SHA1_BLOCK_SIZE, - .cra_module = THIS_MODULE, - } -}; - -static int __init ppc_spe_sha1_mod_init(void) -{ - return crypto_register_shash(&alg); -} - -static void __exit ppc_spe_sha1_mod_fini(void) -{ - crypto_unregister_shash(&alg); -} - -module_init(ppc_spe_sha1_mod_init); -module_exit(ppc_spe_sha1_mod_fini); - -MODULE_LICENSE("GPL"); -MODULE_DESCRIPTION("SHA1 Secure Hash Algorithm, SPE optimized"); - -MODULE_ALIAS_CRYPTO("sha1"); -MODULE_ALIAS_CRYPTO("sha1-ppc-spe"); -- cgit v1.2.3