2019-05-27 06:55:01 +00:00
|
|
|
/* SPDX-License-Identifier: GPL-2.0-or-later */
|
2006-08-21 11:08:13 +00:00
|
|
|
/*
|
|
|
|
* Cryptographic API for algorithms (i.e., low-level API).
|
|
|
|
*
|
|
|
|
* Copyright (c) 2006 Herbert Xu <herbert@gondor.apana.org.au>
|
|
|
|
*/
|
|
|
|
#ifndef _CRYPTO_ALGAPI_H
|
|
|
|
#define _CRYPTO_ALGAPI_H
|
|
|
|
|
2021-12-10 14:30:09 +00:00
|
|
|
#include <linux/align.h>
|
2022-12-05 04:31:44 +00:00
|
|
|
#include <linux/cache.h>
|
2006-08-21 11:08:13 +00:00
|
|
|
#include <linux/crypto.h>
|
2021-12-10 14:30:09 +00:00
|
|
|
#include <linux/kconfig.h>
|
2007-04-16 10:48:54 +00:00
|
|
|
#include <linux/list.h>
|
2021-12-10 14:30:09 +00:00
|
|
|
#include <linux/types.h>
|
2006-08-21 11:08:13 +00:00
|
|
|
|
2022-02-23 07:07:01 +00:00
|
|
|
#include <asm/unaligned.h>
|
|
|
|
|
2018-04-09 13:54:46 +00:00
|
|
|
/*
|
|
|
|
* Maximum values for blocksize and alignmask, used to allocate
|
|
|
|
* static buffers that are big enough for any combination of
|
2018-08-07 21:18:40 +00:00
|
|
|
* algs and architectures. Ciphers have a lower maximum size.
|
2018-04-09 13:54:46 +00:00
|
|
|
*/
|
2018-08-07 21:18:40 +00:00
|
|
|
#define MAX_ALGAPI_BLOCKSIZE 160
|
2022-11-25 04:36:39 +00:00
|
|
|
#define MAX_ALGAPI_ALIGNMASK 127
|
2018-04-09 13:54:46 +00:00
|
|
|
#define MAX_CIPHER_BLOCKSIZE 16
|
|
|
|
#define MAX_CIPHER_ALIGNMASK 15
|
|
|
|
|
2022-11-25 04:36:31 +00:00
|
|
|
#ifdef ARCH_DMA_MINALIGN
|
|
|
|
#define CRYPTO_DMA_ALIGN ARCH_DMA_MINALIGN
|
|
|
|
#else
|
|
|
|
#define CRYPTO_DMA_ALIGN CRYPTO_MINALIGN
|
|
|
|
#endif
|
|
|
|
|
|
|
|
#define CRYPTO_DMA_PADDING ((CRYPTO_DMA_ALIGN - 1) & ~(CRYPTO_MINALIGN - 1))
|
|
|
|
|
2015-05-11 09:48:12 +00:00
|
|
|
struct crypto_aead;
|
2015-07-08 23:17:15 +00:00
|
|
|
struct crypto_instance;
|
2006-08-06 11:16:34 +00:00
|
|
|
struct module;
|
2021-12-10 14:30:09 +00:00
|
|
|
struct notifier_block;
|
2007-01-01 07:37:02 +00:00
|
|
|
struct rtattr;
|
2006-08-21 14:06:54 +00:00
|
|
|
struct seq_file;
|
2020-08-19 11:58:20 +00:00
|
|
|
struct sk_buff;
|
2006-08-21 14:06:54 +00:00
|
|
|
|
|
|
|
struct crypto_type {
|
2007-01-24 09:50:26 +00:00
|
|
|
unsigned int (*ctxsize)(struct crypto_alg *alg, u32 type, u32 mask);
|
2009-07-13 12:46:25 +00:00
|
|
|
unsigned int (*extsize)(struct crypto_alg *alg);
|
2007-01-24 09:50:26 +00:00
|
|
|
int (*init)(struct crypto_tfm *tfm, u32 type, u32 mask);
|
2009-07-13 12:46:25 +00:00
|
|
|
int (*init_tfm)(struct crypto_tfm *tfm);
|
2006-08-21 14:06:54 +00:00
|
|
|
void (*show)(struct seq_file *m, struct crypto_alg *alg);
|
2011-09-27 05:24:29 +00:00
|
|
|
int (*report)(struct sk_buff *skb, struct crypto_alg *alg);
|
2015-07-08 23:17:15 +00:00
|
|
|
void (*free)(struct crypto_instance *inst);
|
2008-09-20 21:52:53 +00:00
|
|
|
|
|
|
|
unsigned int type;
|
|
|
|
unsigned int maskclear;
|
|
|
|
unsigned int maskset;
|
|
|
|
unsigned int tfmsize;
|
2006-08-21 14:06:54 +00:00
|
|
|
};
|
2006-08-06 11:16:34 +00:00
|
|
|
|
|
|
|
struct crypto_instance {
|
|
|
|
struct crypto_alg alg;
|
|
|
|
|
|
|
|
struct crypto_template *tmpl;
|
2019-12-18 07:53:01 +00:00
|
|
|
|
|
|
|
union {
|
|
|
|
/* Node in list of instances after registration. */
|
|
|
|
struct hlist_node list;
|
|
|
|
/* List of attached spawns before registration. */
|
|
|
|
struct crypto_spawn *spawns;
|
|
|
|
};
|
2006-08-06 11:16:34 +00:00
|
|
|
|
|
|
|
void *__ctx[] CRYPTO_MINALIGN_ATTR;
|
|
|
|
};
|
|
|
|
|
|
|
|
struct crypto_template {
|
|
|
|
struct list_head list;
|
|
|
|
struct hlist_head instances;
|
|
|
|
struct module *module;
|
|
|
|
|
2009-07-07 04:30:33 +00:00
|
|
|
int (*create)(struct crypto_template *tmpl, struct rtattr **tb);
|
2006-08-06 11:16:34 +00:00
|
|
|
|
|
|
|
char name[CRYPTO_MAX_ALG_NAME];
|
|
|
|
};
|
|
|
|
|
2006-09-21 01:39:29 +00:00
|
|
|
struct crypto_spawn {
|
|
|
|
struct list_head list;
|
|
|
|
struct crypto_alg *alg;
|
2019-12-18 07:53:01 +00:00
|
|
|
union {
|
|
|
|
/* Back pointer to instance after registration.*/
|
|
|
|
struct crypto_instance *inst;
|
|
|
|
/* Spawn list pointer prior to registration. */
|
|
|
|
struct crypto_spawn *next;
|
|
|
|
};
|
2009-07-08 07:55:52 +00:00
|
|
|
const struct crypto_type *frontend;
|
2007-04-08 11:31:36 +00:00
|
|
|
u32 mask;
|
2019-12-07 14:15:17 +00:00
|
|
|
bool dead;
|
2019-12-18 07:53:01 +00:00
|
|
|
bool registered;
|
2006-09-21 01:39:29 +00:00
|
|
|
};
|
|
|
|
|
2007-04-16 10:48:54 +00:00
|
|
|
struct crypto_queue {
|
|
|
|
struct list_head list;
|
|
|
|
struct list_head *backlog;
|
|
|
|
|
|
|
|
unsigned int qlen;
|
|
|
|
unsigned int max_qlen;
|
|
|
|
};
|
|
|
|
|
2006-08-12 11:56:17 +00:00
|
|
|
struct scatter_walk {
|
|
|
|
struct scatterlist *sg;
|
|
|
|
unsigned int offset;
|
|
|
|
};
|
|
|
|
|
2021-06-17 07:28:10 +00:00
|
|
|
struct crypto_attr_alg {
|
|
|
|
char name[CRYPTO_MAX_ALG_NAME];
|
|
|
|
};
|
|
|
|
|
|
|
|
struct crypto_attr_type {
|
|
|
|
u32 type;
|
|
|
|
u32 mask;
|
|
|
|
};
|
|
|
|
|
2006-09-21 01:44:08 +00:00
|
|
|
void crypto_mod_put(struct crypto_alg *alg);
|
|
|
|
|
2006-08-06 11:16:34 +00:00
|
|
|
int crypto_register_template(struct crypto_template *tmpl);
|
2019-01-18 05:58:11 +00:00
|
|
|
int crypto_register_templates(struct crypto_template *tmpls, int count);
|
2006-08-06 11:16:34 +00:00
|
|
|
void crypto_unregister_template(struct crypto_template *tmpl);
|
2019-01-18 05:58:11 +00:00
|
|
|
void crypto_unregister_templates(struct crypto_template *tmpls, int count);
|
2006-08-06 11:16:34 +00:00
|
|
|
struct crypto_template *crypto_lookup_template(const char *name);
|
|
|
|
|
2009-07-14 10:45:45 +00:00
|
|
|
int crypto_register_instance(struct crypto_template *tmpl,
|
|
|
|
struct crypto_instance *inst);
|
2019-12-15 23:51:19 +00:00
|
|
|
void crypto_unregister_instance(struct crypto_instance *inst);
|
2009-07-14 10:45:45 +00:00
|
|
|
|
2020-01-03 03:58:48 +00:00
|
|
|
int crypto_grab_spawn(struct crypto_spawn *spawn, struct crypto_instance *inst,
|
|
|
|
const char *name, u32 type, u32 mask);
|
2006-09-21 01:39:29 +00:00
|
|
|
void crypto_drop_spawn(struct crypto_spawn *spawn);
|
2006-12-16 23:05:58 +00:00
|
|
|
struct crypto_tfm *crypto_spawn_tfm(struct crypto_spawn *spawn, u32 type,
|
|
|
|
u32 mask);
|
2009-07-08 07:55:52 +00:00
|
|
|
void *crypto_spawn_tfm2(struct crypto_spawn *spawn);
|
2006-09-21 01:39:29 +00:00
|
|
|
|
2007-01-01 07:37:02 +00:00
|
|
|
struct crypto_attr_type *crypto_get_attr_type(struct rtattr **tb);
|
2020-07-10 06:20:38 +00:00
|
|
|
int crypto_check_attr_type(struct rtattr **tb, u32 type, u32 *mask_ret);
|
2007-12-07 12:18:17 +00:00
|
|
|
const char *crypto_attr_alg_name(struct rtattr *rta);
|
2016-06-29 10:04:13 +00:00
|
|
|
int crypto_inst_setname(struct crypto_instance *inst, const char *name,
|
|
|
|
struct crypto_alg *alg);
|
2006-08-06 13:10:45 +00:00
|
|
|
|
2007-04-16 10:48:54 +00:00
|
|
|
void crypto_init_queue(struct crypto_queue *queue, unsigned int max_qlen);
|
|
|
|
int crypto_enqueue_request(struct crypto_queue *queue,
|
|
|
|
struct crypto_async_request *request);
|
2020-04-28 15:49:03 +00:00
|
|
|
void crypto_enqueue_request_head(struct crypto_queue *queue,
|
|
|
|
struct crypto_async_request *request);
|
2007-04-16 10:48:54 +00:00
|
|
|
struct crypto_async_request *crypto_dequeue_request(struct crypto_queue *queue);
|
2016-01-26 12:25:38 +00:00
|
|
|
static inline unsigned int crypto_queue_len(struct crypto_queue *queue)
|
|
|
|
{
|
|
|
|
return queue->qlen;
|
|
|
|
}
|
2007-04-16 10:48:54 +00:00
|
|
|
|
2007-11-20 09:26:06 +00:00
|
|
|
void crypto_inc(u8 *a, unsigned int size);
|
2017-07-24 10:28:03 +00:00
|
|
|
void __crypto_xor(u8 *dst, const u8 *src1, const u8 *src2, unsigned int size);
|
crypto: algapi - make crypto_xor() and crypto_inc() alignment agnostic
Instead of unconditionally forcing 4 byte alignment for all generic
chaining modes that rely on crypto_xor() or crypto_inc() (which may
result in unnecessary copying of data when the underlying hardware
can perform unaligned accesses efficiently), make those functions
deal with unaligned input explicitly, but only if the Kconfig symbol
HAVE_EFFICIENT_UNALIGNED_ACCESS is set. This will allow us to drop
the alignmasks from the CBC, CMAC, CTR, CTS, PCBC and SEQIV drivers.
For crypto_inc(), this simply involves making the 4-byte stride
conditional on HAVE_EFFICIENT_UNALIGNED_ACCESS being set, given that
it typically operates on 16 byte buffers.
For crypto_xor(), an algorithm is implemented that simply runs through
the input using the largest strides possible if unaligned accesses are
allowed. If they are not, an optimal sequence of memory accesses is
emitted that takes the relative alignment of the input buffers into
account, e.g., if the relative misalignment of dst and src is 4 bytes,
the entire xor operation will be completed using 4 byte loads and stores
(modulo unaligned bits at the start and end). Note that all expressions
involving misalign are simply eliminated by the compiler when
HAVE_EFFICIENT_UNALIGNED_ACCESS is defined.
Signed-off-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2017-02-05 10:06:12 +00:00
|
|
|
|
|
|
|
static inline void crypto_xor(u8 *dst, const u8 *src, unsigned int size)
|
|
|
|
{
|
|
|
|
if (IS_ENABLED(CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS) &&
|
|
|
|
__builtin_constant_p(size) &&
|
|
|
|
(size % sizeof(unsigned long)) == 0) {
|
|
|
|
unsigned long *d = (unsigned long *)dst;
|
|
|
|
unsigned long *s = (unsigned long *)src;
|
2022-02-23 07:07:01 +00:00
|
|
|
unsigned long l;
|
crypto: algapi - make crypto_xor() and crypto_inc() alignment agnostic
Instead of unconditionally forcing 4 byte alignment for all generic
chaining modes that rely on crypto_xor() or crypto_inc() (which may
result in unnecessary copying of data when the underlying hardware
can perform unaligned accesses efficiently), make those functions
deal with unaligned input explicitly, but only if the Kconfig symbol
HAVE_EFFICIENT_UNALIGNED_ACCESS is set. This will allow us to drop
the alignmasks from the CBC, CMAC, CTR, CTS, PCBC and SEQIV drivers.
For crypto_inc(), this simply involves making the 4-byte stride
conditional on HAVE_EFFICIENT_UNALIGNED_ACCESS being set, given that
it typically operates on 16 byte buffers.
For crypto_xor(), an algorithm is implemented that simply runs through
the input using the largest strides possible if unaligned accesses are
allowed. If they are not, an optimal sequence of memory accesses is
emitted that takes the relative alignment of the input buffers into
account, e.g., if the relative misalignment of dst and src is 4 bytes,
the entire xor operation will be completed using 4 byte loads and stores
(modulo unaligned bits at the start and end). Note that all expressions
involving misalign are simply eliminated by the compiler when
HAVE_EFFICIENT_UNALIGNED_ACCESS is defined.
Signed-off-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2017-02-05 10:06:12 +00:00
|
|
|
|
|
|
|
while (size > 0) {
|
2022-02-23 07:07:01 +00:00
|
|
|
l = get_unaligned(d) ^ get_unaligned(s++);
|
|
|
|
put_unaligned(l, d++);
|
crypto: algapi - make crypto_xor() and crypto_inc() alignment agnostic
Instead of unconditionally forcing 4 byte alignment for all generic
chaining modes that rely on crypto_xor() or crypto_inc() (which may
result in unnecessary copying of data when the underlying hardware
can perform unaligned accesses efficiently), make those functions
deal with unaligned input explicitly, but only if the Kconfig symbol
HAVE_EFFICIENT_UNALIGNED_ACCESS is set. This will allow us to drop
the alignmasks from the CBC, CMAC, CTR, CTS, PCBC and SEQIV drivers.
For crypto_inc(), this simply involves making the 4-byte stride
conditional on HAVE_EFFICIENT_UNALIGNED_ACCESS being set, given that
it typically operates on 16 byte buffers.
For crypto_xor(), an algorithm is implemented that simply runs through
the input using the largest strides possible if unaligned accesses are
allowed. If they are not, an optimal sequence of memory accesses is
emitted that takes the relative alignment of the input buffers into
account, e.g., if the relative misalignment of dst and src is 4 bytes,
the entire xor operation will be completed using 4 byte loads and stores
(modulo unaligned bits at the start and end). Note that all expressions
involving misalign are simply eliminated by the compiler when
HAVE_EFFICIENT_UNALIGNED_ACCESS is defined.
Signed-off-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2017-02-05 10:06:12 +00:00
|
|
|
size -= sizeof(unsigned long);
|
|
|
|
}
|
|
|
|
} else {
|
2017-07-24 10:28:03 +00:00
|
|
|
__crypto_xor(dst, dst, src, size);
|
crypto: algapi - make crypto_xor() and crypto_inc() alignment agnostic
Instead of unconditionally forcing 4 byte alignment for all generic
chaining modes that rely on crypto_xor() or crypto_inc() (which may
result in unnecessary copying of data when the underlying hardware
can perform unaligned accesses efficiently), make those functions
deal with unaligned input explicitly, but only if the Kconfig symbol
HAVE_EFFICIENT_UNALIGNED_ACCESS is set. This will allow us to drop
the alignmasks from the CBC, CMAC, CTR, CTS, PCBC and SEQIV drivers.
For crypto_inc(), this simply involves making the 4-byte stride
conditional on HAVE_EFFICIENT_UNALIGNED_ACCESS being set, given that
it typically operates on 16 byte buffers.
For crypto_xor(), an algorithm is implemented that simply runs through
the input using the largest strides possible if unaligned accesses are
allowed. If they are not, an optimal sequence of memory accesses is
emitted that takes the relative alignment of the input buffers into
account, e.g., if the relative misalignment of dst and src is 4 bytes,
the entire xor operation will be completed using 4 byte loads and stores
(modulo unaligned bits at the start and end). Note that all expressions
involving misalign are simply eliminated by the compiler when
HAVE_EFFICIENT_UNALIGNED_ACCESS is defined.
Signed-off-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2017-02-05 10:06:12 +00:00
|
|
|
}
|
|
|
|
}
|
2007-11-20 09:26:06 +00:00
|
|
|
|
crypto: algapi - make crypto_xor() take separate dst and src arguments
There are quite a number of occurrences in the kernel of the pattern
if (dst != src)
memcpy(dst, src, walk.total % AES_BLOCK_SIZE);
crypto_xor(dst, final, walk.total % AES_BLOCK_SIZE);
or
crypto_xor(keystream, src, nbytes);
memcpy(dst, keystream, nbytes);
where crypto_xor() is preceded or followed by a memcpy() invocation
that is only there because crypto_xor() uses its output parameter as
one of the inputs. To avoid having to add new instances of this pattern
in the arm64 code, which will be refactored to implement non-SIMD
fallbacks, add an alternative implementation called crypto_xor_cpy(),
taking separate input and output arguments. This removes the need for
the separate memcpy().
Signed-off-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2017-07-24 10:28:04 +00:00
|
|
|
static inline void crypto_xor_cpy(u8 *dst, const u8 *src1, const u8 *src2,
|
|
|
|
unsigned int size)
|
|
|
|
{
|
|
|
|
if (IS_ENABLED(CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS) &&
|
|
|
|
__builtin_constant_p(size) &&
|
|
|
|
(size % sizeof(unsigned long)) == 0) {
|
|
|
|
unsigned long *d = (unsigned long *)dst;
|
|
|
|
unsigned long *s1 = (unsigned long *)src1;
|
|
|
|
unsigned long *s2 = (unsigned long *)src2;
|
2022-02-23 07:07:01 +00:00
|
|
|
unsigned long l;
|
crypto: algapi - make crypto_xor() take separate dst and src arguments
There are quite a number of occurrences in the kernel of the pattern
if (dst != src)
memcpy(dst, src, walk.total % AES_BLOCK_SIZE);
crypto_xor(dst, final, walk.total % AES_BLOCK_SIZE);
or
crypto_xor(keystream, src, nbytes);
memcpy(dst, keystream, nbytes);
where crypto_xor() is preceded or followed by a memcpy() invocation
that is only there because crypto_xor() uses its output parameter as
one of the inputs. To avoid having to add new instances of this pattern
in the arm64 code, which will be refactored to implement non-SIMD
fallbacks, add an alternative implementation called crypto_xor_cpy(),
taking separate input and output arguments. This removes the need for
the separate memcpy().
Signed-off-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2017-07-24 10:28:04 +00:00
|
|
|
|
|
|
|
while (size > 0) {
|
2022-02-23 07:07:01 +00:00
|
|
|
l = get_unaligned(s1++) ^ get_unaligned(s2++);
|
|
|
|
put_unaligned(l, d++);
|
crypto: algapi - make crypto_xor() take separate dst and src arguments
There are quite a number of occurrences in the kernel of the pattern
if (dst != src)
memcpy(dst, src, walk.total % AES_BLOCK_SIZE);
crypto_xor(dst, final, walk.total % AES_BLOCK_SIZE);
or
crypto_xor(keystream, src, nbytes);
memcpy(dst, keystream, nbytes);
where crypto_xor() is preceded or followed by a memcpy() invocation
that is only there because crypto_xor() uses its output parameter as
one of the inputs. To avoid having to add new instances of this pattern
in the arm64 code, which will be refactored to implement non-SIMD
fallbacks, add an alternative implementation called crypto_xor_cpy(),
taking separate input and output arguments. This removes the need for
the separate memcpy().
Signed-off-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2017-07-24 10:28:04 +00:00
|
|
|
size -= sizeof(unsigned long);
|
|
|
|
}
|
|
|
|
} else {
|
|
|
|
__crypto_xor(dst, src1, src2, size);
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2022-11-25 04:36:31 +00:00
|
|
|
static inline void *crypto_tfm_ctx(struct crypto_tfm *tfm)
|
|
|
|
{
|
|
|
|
return tfm->__crt_ctx;
|
|
|
|
}
|
|
|
|
|
|
|
|
static inline void *crypto_tfm_ctx_align(struct crypto_tfm *tfm,
|
|
|
|
unsigned int align)
|
|
|
|
{
|
|
|
|
if (align <= crypto_tfm_ctx_alignment())
|
|
|
|
align = 1;
|
|
|
|
|
|
|
|
return PTR_ALIGN(crypto_tfm_ctx(tfm), align);
|
|
|
|
}
|
|
|
|
|
2006-08-21 14:07:53 +00:00
|
|
|
static inline void *crypto_tfm_ctx_aligned(struct crypto_tfm *tfm)
|
|
|
|
{
|
2022-11-25 04:36:31 +00:00
|
|
|
return crypto_tfm_ctx_align(tfm, crypto_tfm_alg_alignmask(tfm) + 1);
|
|
|
|
}
|
|
|
|
|
|
|
|
static inline unsigned int crypto_dma_align(void)
|
|
|
|
{
|
|
|
|
return CRYPTO_DMA_ALIGN;
|
|
|
|
}
|
|
|
|
|
|
|
|
static inline unsigned int crypto_dma_padding(void)
|
|
|
|
{
|
|
|
|
return (crypto_dma_align() - 1) & ~(crypto_tfm_ctx_alignment() - 1);
|
|
|
|
}
|
|
|
|
|
|
|
|
static inline void *crypto_tfm_ctx_dma(struct crypto_tfm *tfm)
|
|
|
|
{
|
|
|
|
return crypto_tfm_ctx_align(tfm, crypto_dma_align());
|
2006-08-21 14:07:53 +00:00
|
|
|
}
|
|
|
|
|
2007-04-16 10:49:20 +00:00
|
|
|
static inline struct crypto_instance *crypto_tfm_alg_instance(
|
|
|
|
struct crypto_tfm *tfm)
|
|
|
|
{
|
|
|
|
return container_of(tfm->__crt_alg, struct crypto_instance, alg);
|
|
|
|
}
|
|
|
|
|
2006-08-06 11:16:34 +00:00
|
|
|
static inline void *crypto_instance_ctx(struct crypto_instance *inst)
|
|
|
|
{
|
|
|
|
return inst->__ctx;
|
|
|
|
}
|
|
|
|
|
2007-04-16 10:48:54 +00:00
|
|
|
static inline struct crypto_async_request *crypto_get_backlog(
|
|
|
|
struct crypto_queue *queue)
|
|
|
|
{
|
|
|
|
return queue->backlog == &queue->list ? NULL :
|
|
|
|
container_of(queue->backlog, struct crypto_async_request, list);
|
|
|
|
}
|
|
|
|
|
2020-07-10 06:20:38 +00:00
|
|
|
static inline u32 crypto_requires_off(struct crypto_attr_type *algt, u32 off)
|
2017-02-26 04:22:35 +00:00
|
|
|
{
|
2020-07-10 06:20:38 +00:00
|
|
|
return (algt->type ^ off) & algt->mask & off;
|
2017-02-26 04:22:35 +00:00
|
|
|
}
|
|
|
|
|
2007-12-17 12:07:31 +00:00
|
|
|
/*
|
2020-07-10 06:20:38 +00:00
|
|
|
* When an algorithm uses another algorithm (e.g., if it's an instance of a
|
|
|
|
* template), these are the flags that should always be set on the "outer"
|
|
|
|
* algorithm if any "inner" algorithm has them set.
|
2007-12-17 12:07:31 +00:00
|
|
|
*/
|
2020-07-10 06:20:39 +00:00
|
|
|
#define CRYPTO_ALG_INHERITED_FLAGS \
|
2020-07-10 06:20:40 +00:00
|
|
|
(CRYPTO_ALG_ASYNC | CRYPTO_ALG_NEED_FALLBACK | \
|
|
|
|
CRYPTO_ALG_ALLOCATES_MEMORY)
|
2020-07-10 06:20:38 +00:00
|
|
|
|
|
|
|
/*
|
|
|
|
* Given the type and mask that specify the flags restrictions on a template
|
|
|
|
* instance being created, return the mask that should be passed to
|
|
|
|
* crypto_grab_*() (along with type=0) to honor any request the user made to
|
|
|
|
* have any of the CRYPTO_ALG_INHERITED_FLAGS clear.
|
|
|
|
*/
|
|
|
|
static inline u32 crypto_algt_inherited_mask(struct crypto_attr_type *algt)
|
2007-12-17 12:07:31 +00:00
|
|
|
{
|
2020-07-10 06:20:38 +00:00
|
|
|
return crypto_requires_off(algt, CRYPTO_ALG_INHERITED_FLAGS);
|
2007-12-17 12:07:31 +00:00
|
|
|
}
|
|
|
|
|
crypto: crypto_memneq - add equality testing of memory regions w/o timing leaks
When comparing MAC hashes, AEAD authentication tags, or other hash
values in the context of authentication or integrity checking, it
is important not to leak timing information to a potential attacker,
i.e. when communication happens over a network.
Bytewise memory comparisons (such as memcmp) are usually optimized so
that they return a nonzero value as soon as a mismatch is found. E.g,
on x86_64/i5 for 512 bytes this can be ~50 cyc for a full mismatch
and up to ~850 cyc for a full match (cold). This early-return behavior
can leak timing information as a side channel, allowing an attacker to
iteratively guess the correct result.
This patch adds a new method crypto_memneq ("memory not equal to each
other") to the crypto API that compares memory areas of the same length
in roughly "constant time" (cache misses could change the timing, but
since they don't reveal information about the content of the strings
being compared, they are effectively benign). Iow, best and worst case
behaviour take the same amount of time to complete (in contrast to
memcmp).
Note that crypto_memneq (unlike memcmp) can only be used to test for
equality or inequality, NOT for lexicographical order. This, however,
is not an issue for its use-cases within the crypto API.
We tried to locate all of the places in the crypto API where memcmp was
being used for authentication or integrity checking, and convert them
over to crypto_memneq.
crypto_memneq is declared noinline, placed in its own source file,
and compiled with optimizations that might increase code size disabled
("Os") because a smart compiler (or LTO) might notice that the return
value is always compared against zero/nonzero, and might then
reintroduce the same early-return optimization that we are trying to
avoid.
Using #pragma or __attribute__ optimization annotations of the code
for disabling optimization was avoided as it seems to be considered
broken or unmaintained for long time in GCC [1]. Therefore, we work
around that by specifying the compile flag for memneq.o directly in
the Makefile. We found that this seems to be most appropriate.
As we use ("Os"), this patch also provides a loop-free "fast-path" for
frequently used 16 byte digests. Similarly to kernel library string
functions, leave an option for future even further optimized architecture
specific assembler implementations.
This was a joint work of James Yonan and Daniel Borkmann. Also thanks
for feedback from Florian Weimer on this and earlier proposals [2].
[1] http://gcc.gnu.org/ml/gcc/2012-07/msg00211.html
[2] https://lkml.org/lkml/2013/2/10/131
Signed-off-by: James Yonan <james@openvpn.net>
Signed-off-by: Daniel Borkmann <dborkman@redhat.com>
Cc: Florian Weimer <fw@deneb.enyo.de>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2013-09-26 08:20:39 +00:00
|
|
|
noinline unsigned long __crypto_memneq(const void *a, const void *b, size_t size);
|
|
|
|
|
|
|
|
/**
|
|
|
|
* crypto_memneq - Compare two areas of memory without leaking
|
|
|
|
* timing information.
|
|
|
|
*
|
|
|
|
* @a: One area of memory
|
|
|
|
* @b: Another area of memory
|
|
|
|
* @size: The size of the area.
|
|
|
|
*
|
|
|
|
* Returns 0 when data is equal, 1 otherwise.
|
|
|
|
*/
|
|
|
|
static inline int crypto_memneq(const void *a, const void *b, size_t size)
|
|
|
|
{
|
|
|
|
return __crypto_memneq(a, b, size) != 0UL ? 1 : 0;
|
|
|
|
}
|
2006-08-21 11:08:13 +00:00
|
|
|
|
2018-08-30 15:00:14 +00:00
|
|
|
int crypto_register_notifier(struct notifier_block *nb);
|
|
|
|
int crypto_unregister_notifier(struct notifier_block *nb);
|
|
|
|
|
|
|
|
/* Crypto notification events. */
|
|
|
|
enum {
|
|
|
|
CRYPTO_MSG_ALG_REQUEST,
|
|
|
|
CRYPTO_MSG_ALG_REGISTER,
|
|
|
|
CRYPTO_MSG_ALG_LOADED,
|
|
|
|
};
|
|
|
|
|
2023-01-31 08:01:45 +00:00
|
|
|
static inline void crypto_request_complete(struct crypto_async_request *req,
|
|
|
|
int err)
|
|
|
|
{
|
|
|
|
crypto_completion_t complete = req->complete;
|
|
|
|
complete(req, err);
|
|
|
|
}
|
|
|
|
|
crypto: crypto_memneq - add equality testing of memory regions w/o timing leaks
When comparing MAC hashes, AEAD authentication tags, or other hash
values in the context of authentication or integrity checking, it
is important not to leak timing information to a potential attacker,
i.e. when communication happens over a network.
Bytewise memory comparisons (such as memcmp) are usually optimized so
that they return a nonzero value as soon as a mismatch is found. E.g,
on x86_64/i5 for 512 bytes this can be ~50 cyc for a full mismatch
and up to ~850 cyc for a full match (cold). This early-return behavior
can leak timing information as a side channel, allowing an attacker to
iteratively guess the correct result.
This patch adds a new method crypto_memneq ("memory not equal to each
other") to the crypto API that compares memory areas of the same length
in roughly "constant time" (cache misses could change the timing, but
since they don't reveal information about the content of the strings
being compared, they are effectively benign). Iow, best and worst case
behaviour take the same amount of time to complete (in contrast to
memcmp).
Note that crypto_memneq (unlike memcmp) can only be used to test for
equality or inequality, NOT for lexicographical order. This, however,
is not an issue for its use-cases within the crypto API.
We tried to locate all of the places in the crypto API where memcmp was
being used for authentication or integrity checking, and convert them
over to crypto_memneq.
crypto_memneq is declared noinline, placed in its own source file,
and compiled with optimizations that might increase code size disabled
("Os") because a smart compiler (or LTO) might notice that the return
value is always compared against zero/nonzero, and might then
reintroduce the same early-return optimization that we are trying to
avoid.
Using #pragma or __attribute__ optimization annotations of the code
for disabling optimization was avoided as it seems to be considered
broken or unmaintained for long time in GCC [1]. Therefore, we work
around that by specifying the compile flag for memneq.o directly in
the Makefile. We found that this seems to be most appropriate.
As we use ("Os"), this patch also provides a loop-free "fast-path" for
frequently used 16 byte digests. Similarly to kernel library string
functions, leave an option for future even further optimized architecture
specific assembler implementations.
This was a joint work of James Yonan and Daniel Borkmann. Also thanks
for feedback from Florian Weimer on this and earlier proposals [2].
[1] http://gcc.gnu.org/ml/gcc/2012-07/msg00211.html
[2] https://lkml.org/lkml/2013/2/10/131
Signed-off-by: James Yonan <james@openvpn.net>
Signed-off-by: Daniel Borkmann <dborkman@redhat.com>
Cc: Florian Weimer <fw@deneb.enyo.de>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2013-09-26 08:20:39 +00:00
|
|
|
#endif /* _CRYPTO_ALGAPI_H */
|