openssl/crypto/rand/rand_unix.c

659 lines
18 KiB
C
Raw Normal View History

/*
* Copyright 1995-2018 The OpenSSL Project Authors. All Rights Reserved.
*
* Licensed under the OpenSSL license (the "License"). You may not use
* this file except in compliance with the License. You can obtain a copy
* in the file LICENSE in the source distribution or at
* https://www.openssl.org/source/license.html
*/
#ifndef _GNU_SOURCE
# define _GNU_SOURCE
#endif
#include "e_os.h"
#include <stdio.h>
#include "internal/cryptlib.h"
#include <openssl/rand.h>
#include "rand_lcl.h"
#include "internal/rand_int.h"
#include <stdio.h>
#include "internal/dso.h"
#if defined(__linux)
# include <sys/syscall.h>
#endif
#if defined(__FreeBSD__)
# include <sys/types.h>
# include <sys/sysctl.h>
# include <sys/param.h>
#endif
#if defined(__OpenBSD__) || defined(__NetBSD__)
# include <sys/param.h>
#endif
#if defined(OPENSSL_SYS_UNIX) || defined(__DJGPP__)
DRBG: implement a get_nonce() callback Fixes #5849 In pull request #5503 a fallback was added which adds a random nonce of security_strength/2 bits if no nonce callback is provided. This change raised the entropy requirements form 256 to 384 bit, which can cause problems on some platforms (e.g. VMS, see issue #5849). The requirements for the nonce are given in section 8.6.7 of NIST SP 800-90Ar1: A nonce may be required in the construction of a seed during instantiation in order to provide a security cushion to block certain attacks. The nonce shall be either: a) A value with at least (security_strength/2) bits of entropy, or b) A value that is expected to repeat no more often than a (security_strength/2)-bit random string would be expected to repeat. Each nonce shall be unique to the cryptographic module in which instantiation is performed, but need not be secret. When used, the nonce shall be considered to be a critical security parameter. This commit implements a nonce of type b) in order to lower the entropy requirements during instantiation back to 256 bits. The formulation "shall be unique to the cryptographic module" above implies that the nonce needs to be unique among (with high probability) among all DRBG instances in "space" and "time". We try to achieve this goal by creating a nonce of the following form nonce = app-specific-data || high-resolution-utc-timestamp || counter Where || denotes concatenation. The application specific data can be something like the process or group id of the application. A utc timestamp is used because it increases monotonically, provided the system time is synchronized. This approach may not be perfect yet for a FIPS evaluation, but it should be good enough for the moment. This commit also harmonizes the implementation of the get_nonce() and the get_additional_data() callbacks and moves the platform specific parts from rand_lib.c into rand_unix.c, rand_win.c, and rand_vms.c. Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/5920)
2018-04-10 08:22:52 +00:00
# include <sys/types.h>
# include <sys/stat.h>
# include <fcntl.h>
DRBG: implement a get_nonce() callback Fixes #5849 In pull request #5503 a fallback was added which adds a random nonce of security_strength/2 bits if no nonce callback is provided. This change raised the entropy requirements form 256 to 384 bit, which can cause problems on some platforms (e.g. VMS, see issue #5849). The requirements for the nonce are given in section 8.6.7 of NIST SP 800-90Ar1: A nonce may be required in the construction of a seed during instantiation in order to provide a security cushion to block certain attacks. The nonce shall be either: a) A value with at least (security_strength/2) bits of entropy, or b) A value that is expected to repeat no more often than a (security_strength/2)-bit random string would be expected to repeat. Each nonce shall be unique to the cryptographic module in which instantiation is performed, but need not be secret. When used, the nonce shall be considered to be a critical security parameter. This commit implements a nonce of type b) in order to lower the entropy requirements during instantiation back to 256 bits. The formulation "shall be unique to the cryptographic module" above implies that the nonce needs to be unique among (with high probability) among all DRBG instances in "space" and "time". We try to achieve this goal by creating a nonce of the following form nonce = app-specific-data || high-resolution-utc-timestamp || counter Where || denotes concatenation. The application specific data can be something like the process or group id of the application. A utc timestamp is used because it increases monotonically, provided the system time is synchronized. This approach may not be perfect yet for a FIPS evaluation, but it should be good enough for the moment. This commit also harmonizes the implementation of the get_nonce() and the get_additional_data() callbacks and moves the platform specific parts from rand_lib.c into rand_unix.c, rand_win.c, and rand_vms.c. Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/5920)
2018-04-10 08:22:52 +00:00
# include <unistd.h>
# include <sys/time.h>
static uint64_t get_time_stamp(void);
static uint64_t get_timer_bits(void);
/* Macro to convert two thirty two bit values into a sixty four bit one */
# define TWO32TO64(a, b) ((((uint64_t)(a)) << 32) + (b))
/*
* Check for the existence and support of POSIX timers. The standard
* says that the _POSIX_TIMERS macro will have a positive value if they
* are available.
*
* However, we want an additional constraint: that the timer support does
* not require an extra library dependency. Early versions of glibc
* require -lrt to be specified on the link line to access the timers,
* so this needs to be checked for.
*
* It is worse because some libraries define __GLIBC__ but don't
* support the version testing macro (e.g. uClibc). This means
* an extra check is needed.
*
* The final condition is:
* "have posix timers and either not glibc or glibc without -lrt"
*
* The nested #if sequences are required to avoid using a parameterised
* macro that might be undefined.
*/
# undef OSSL_POSIX_TIMER_OKAY
# if defined(_POSIX_TIMERS) && _POSIX_TIMERS > 0
# if defined(__GLIBC__)
# if defined(__GLIBC_PREREQ)
# if __GLIBC_PREREQ(2, 17)
# define OSSL_POSIX_TIMER_OKAY
# endif
# endif
# else
# define OSSL_POSIX_TIMER_OKAY
# endif
# endif
#endif /* defined(OPENSSL_SYS_UNIX) || defined(__DJGPP__) */
int syscall_random(void *buf, size_t buflen);
Fix reseeding issues of the public RAND_DRBG Reseeding is handled very differently by the classic RAND_METHOD API and the new RAND_DRBG api. These differences led to some problems when the new RAND_DRBG was made the default OpenSSL RNG. In particular, RAND_add() did not work as expected anymore. These issues are discussed on the thread '[openssl-dev] Plea for a new public OpenSSL RNG API' and in Pull Request #4328. This commit fixes the mentioned issues, introducing the following changes: - Replace the fixed size RAND_BYTES_BUFFER by a new RAND_POOL API which facilitates collecting entropy by the get_entropy() callback. - Don't use RAND_poll()/RAND_add() for collecting entropy from the get_entropy() callback anymore. Instead, replace RAND_poll() by RAND_POOL_acquire_entropy(). - Add a new function rand_drbg_restart() which tries to get the DRBG in an instantiated state by all means, regardless of the current state (uninstantiated, error, ...) the DRBG is in. If the caller provides entropy or additional input, it will be used for reseeding. - Restore the original documented behaviour of RAND_add() and RAND_poll() (namely to reseed the DRBG immediately) by a new implementation based on rand_drbg_restart(). - Add automatic error recovery from temporary failures of the entropy source to RAND_DRBG_generate() using the rand_drbg_restart() function. Reviewed-by: Paul Dale <paul.dale@oracle.com> Reviewed-by: Kurt Roeckx <kurt@roeckx.be> Reviewed-by: Rich Salz <rsalz@openssl.org> Reviewed-by: Ben Kaduk <kaduk@mit.edu> (Merged from https://github.com/openssl/openssl/pull/4328)
2017-08-31 21:16:22 +00:00
#if (defined(OPENSSL_SYS_VXWORKS) || defined(OPENSSL_SYS_UEFI)) && \
!defined(OPENSSL_RAND_SEED_NONE)
Fix reseeding issues of the public RAND_DRBG Reseeding is handled very differently by the classic RAND_METHOD API and the new RAND_DRBG api. These differences led to some problems when the new RAND_DRBG was made the default OpenSSL RNG. In particular, RAND_add() did not work as expected anymore. These issues are discussed on the thread '[openssl-dev] Plea for a new public OpenSSL RNG API' and in Pull Request #4328. This commit fixes the mentioned issues, introducing the following changes: - Replace the fixed size RAND_BYTES_BUFFER by a new RAND_POOL API which facilitates collecting entropy by the get_entropy() callback. - Don't use RAND_poll()/RAND_add() for collecting entropy from the get_entropy() callback anymore. Instead, replace RAND_poll() by RAND_POOL_acquire_entropy(). - Add a new function rand_drbg_restart() which tries to get the DRBG in an instantiated state by all means, regardless of the current state (uninstantiated, error, ...) the DRBG is in. If the caller provides entropy or additional input, it will be used for reseeding. - Restore the original documented behaviour of RAND_add() and RAND_poll() (namely to reseed the DRBG immediately) by a new implementation based on rand_drbg_restart(). - Add automatic error recovery from temporary failures of the entropy source to RAND_DRBG_generate() using the rand_drbg_restart() function. Reviewed-by: Paul Dale <paul.dale@oracle.com> Reviewed-by: Kurt Roeckx <kurt@roeckx.be> Reviewed-by: Rich Salz <rsalz@openssl.org> Reviewed-by: Ben Kaduk <kaduk@mit.edu> (Merged from https://github.com/openssl/openssl/pull/4328)
2017-08-31 21:16:22 +00:00
# error "UEFI and VXWorks only support seeding NONE"
#endif
#if !(defined(OPENSSL_SYS_WINDOWS) || defined(OPENSSL_SYS_WIN32) \
|| defined(OPENSSL_SYS_VMS) || defined(OPENSSL_SYS_VXWORKS) \
|| defined(OPENSSL_SYS_UEFI))
# if defined(OPENSSL_SYS_VOS)
# ifndef OPENSSL_RAND_SEED_OS
# error "Unsupported seeding method configured; must be os"
# endif
# if defined(OPENSSL_SYS_VOS_HPPA) && defined(OPENSSL_SYS_VOS_IA32)
# error "Unsupported HP-PA and IA32 at the same time."
# endif
# if !defined(OPENSSL_SYS_VOS_HPPA) && !defined(OPENSSL_SYS_VOS_IA32)
# error "Must have one of HP-PA or IA32"
# endif
/*
* The following algorithm repeatedly samples the real-time clock (RTC) to
* generate a sequence of unpredictable data. The algorithm relies upon the
* uneven execution speed of the code (due to factors such as cache misses,
* interrupts, bus activity, and scheduling) and upon the rather large
* relative difference between the speed of the clock and the rate at which
* it can be read. If it is ported to an environment where execution speed
* is more constant or where the RTC ticks at a much slower rate, or the
* clock can be read with fewer instructions, it is likely that the results
* would be far more predictable. This should only be used for legacy
* platforms.
*
Fix reseeding issues of the public RAND_DRBG Reseeding is handled very differently by the classic RAND_METHOD API and the new RAND_DRBG api. These differences led to some problems when the new RAND_DRBG was made the default OpenSSL RNG. In particular, RAND_add() did not work as expected anymore. These issues are discussed on the thread '[openssl-dev] Plea for a new public OpenSSL RNG API' and in Pull Request #4328. This commit fixes the mentioned issues, introducing the following changes: - Replace the fixed size RAND_BYTES_BUFFER by a new RAND_POOL API which facilitates collecting entropy by the get_entropy() callback. - Don't use RAND_poll()/RAND_add() for collecting entropy from the get_entropy() callback anymore. Instead, replace RAND_poll() by RAND_POOL_acquire_entropy(). - Add a new function rand_drbg_restart() which tries to get the DRBG in an instantiated state by all means, regardless of the current state (uninstantiated, error, ...) the DRBG is in. If the caller provides entropy or additional input, it will be used for reseeding. - Restore the original documented behaviour of RAND_add() and RAND_poll() (namely to reseed the DRBG immediately) by a new implementation based on rand_drbg_restart(). - Add automatic error recovery from temporary failures of the entropy source to RAND_DRBG_generate() using the rand_drbg_restart() function. Reviewed-by: Paul Dale <paul.dale@oracle.com> Reviewed-by: Kurt Roeckx <kurt@roeckx.be> Reviewed-by: Rich Salz <rsalz@openssl.org> Reviewed-by: Ben Kaduk <kaduk@mit.edu> (Merged from https://github.com/openssl/openssl/pull/4328)
2017-08-31 21:16:22 +00:00
* As a precaution, we assume only 2 bits of entropy per byte.
*/
size_t rand_pool_acquire_entropy(RAND_POOL *pool)
2009-04-07 16:33:26 +00:00
{
short int code;
int i, k;
Fix reseeding issues of the public RAND_DRBG Reseeding is handled very differently by the classic RAND_METHOD API and the new RAND_DRBG api. These differences led to some problems when the new RAND_DRBG was made the default OpenSSL RNG. In particular, RAND_add() did not work as expected anymore. These issues are discussed on the thread '[openssl-dev] Plea for a new public OpenSSL RNG API' and in Pull Request #4328. This commit fixes the mentioned issues, introducing the following changes: - Replace the fixed size RAND_BYTES_BUFFER by a new RAND_POOL API which facilitates collecting entropy by the get_entropy() callback. - Don't use RAND_poll()/RAND_add() for collecting entropy from the get_entropy() callback anymore. Instead, replace RAND_poll() by RAND_POOL_acquire_entropy(). - Add a new function rand_drbg_restart() which tries to get the DRBG in an instantiated state by all means, regardless of the current state (uninstantiated, error, ...) the DRBG is in. If the caller provides entropy or additional input, it will be used for reseeding. - Restore the original documented behaviour of RAND_add() and RAND_poll() (namely to reseed the DRBG immediately) by a new implementation based on rand_drbg_restart(). - Add automatic error recovery from temporary failures of the entropy source to RAND_DRBG_generate() using the rand_drbg_restart() function. Reviewed-by: Paul Dale <paul.dale@oracle.com> Reviewed-by: Kurt Roeckx <kurt@roeckx.be> Reviewed-by: Rich Salz <rsalz@openssl.org> Reviewed-by: Ben Kaduk <kaduk@mit.edu> (Merged from https://github.com/openssl/openssl/pull/4328)
2017-08-31 21:16:22 +00:00
size_t bytes_needed;
struct timespec ts;
unsigned char v;
# ifdef OPENSSL_SYS_VOS_HPPA
long duration;
extern void s$sleep(long *_duration, short int *_code);
# else
long long duration;
extern void s$sleep2(long long *_duration, short int *_code);
# endif
bytes_needed = rand_pool_bytes_needed(pool, 4 /*entropy_factor*/);
Fix reseeding issues of the public RAND_DRBG Reseeding is handled very differently by the classic RAND_METHOD API and the new RAND_DRBG api. These differences led to some problems when the new RAND_DRBG was made the default OpenSSL RNG. In particular, RAND_add() did not work as expected anymore. These issues are discussed on the thread '[openssl-dev] Plea for a new public OpenSSL RNG API' and in Pull Request #4328. This commit fixes the mentioned issues, introducing the following changes: - Replace the fixed size RAND_BYTES_BUFFER by a new RAND_POOL API which facilitates collecting entropy by the get_entropy() callback. - Don't use RAND_poll()/RAND_add() for collecting entropy from the get_entropy() callback anymore. Instead, replace RAND_poll() by RAND_POOL_acquire_entropy(). - Add a new function rand_drbg_restart() which tries to get the DRBG in an instantiated state by all means, regardless of the current state (uninstantiated, error, ...) the DRBG is in. If the caller provides entropy or additional input, it will be used for reseeding. - Restore the original documented behaviour of RAND_add() and RAND_poll() (namely to reseed the DRBG immediately) by a new implementation based on rand_drbg_restart(). - Add automatic error recovery from temporary failures of the entropy source to RAND_DRBG_generate() using the rand_drbg_restart() function. Reviewed-by: Paul Dale <paul.dale@oracle.com> Reviewed-by: Kurt Roeckx <kurt@roeckx.be> Reviewed-by: Rich Salz <rsalz@openssl.org> Reviewed-by: Ben Kaduk <kaduk@mit.edu> (Merged from https://github.com/openssl/openssl/pull/4328)
2017-08-31 21:16:22 +00:00
for (i = 0; i < bytes_needed; i++) {
/*
* burn some cpu; hope for interrupts, cache collisions, bus
* interference, etc.
*/
for (k = 0; k < 99; k++)
ts.tv_nsec = random();
# ifdef OPENSSL_SYS_VOS_HPPA
/* sleep for 1/1024 of a second (976 us). */
duration = 1;
s$sleep(&duration, &code);
# else
/* sleep for 1/65536 of a second (15 us). */
duration = 1;
s$sleep2(&duration, &code);
# endif
/* Get wall clock time, take 8 bits. */
clock_gettime(CLOCK_REALTIME, &ts);
v = (unsigned char)(ts.tv_nsec & 0xFF);
rand_pool_add(pool, arg, &v, sizeof(v) , 2);
}
return rand_pool_entropy_available(pool);
2009-04-07 16:33:26 +00:00
}
void rand_pool_cleanup(void)
{
}
void rand_pool_keep_random_devices_open(int keep)
{
}
# else
# if defined(OPENSSL_RAND_SEED_EGD) && \
(defined(OPENSSL_NO_EGD) || !defined(DEVRANDOM_EGD))
# error "Seeding uses EGD but EGD is turned off or no device given"
# endif
# if defined(OPENSSL_RAND_SEED_DEVRANDOM) && !defined(DEVRANDOM)
# error "Seeding uses urandom but DEVRANDOM is not configured"
# endif
# if defined(OPENSSL_RAND_SEED_OS)
# if !defined(DEVRANDOM)
# error "OS seeding requires DEVRANDOM to be configured"
# endif
# define OPENSSL_RAND_SEED_GETRANDOM
# define OPENSSL_RAND_SEED_DEVRANDOM
# endif
# if defined(OPENSSL_RAND_SEED_LIBRANDOM)
# error "librandom not (yet) supported"
# endif
# if (defined(__FreeBSD__) || defined(__NetBSD__)) && defined(KERN_ARND)
/*
* sysctl_random(): Use sysctl() to read a random number from the kernel
* Returns the size on success, 0 on failure.
*/
static size_t sysctl_random(char *buf, size_t buflen)
{
int mib[2];
size_t done = 0;
size_t len;
/*
* On FreeBSD old implementations returned longs, newer versions support
* variable sizes up to 256 byte. The code below would not work properly
* when the sysctl returns long and we want to request something not a
* multiple of longs, which should never be the case.
*/
if (!ossl_assert(buflen % sizeof(long) == 0))
return 0;
/*
* On NetBSD before 4.0 KERN_ARND was an alias for KERN_URND, and only
* filled in an int, leaving the rest uninitialized. Since NetBSD 4.0
* it returns a variable number of bytes with the current version supporting
* up to 256 bytes.
* Just return an error on older NetBSD versions.
*/
#if defined(__NetBSD__) && __NetBSD_Version__ < 400000000
return 0;
#endif
mib[0] = CTL_KERN;
mib[1] = KERN_ARND;
do {
len = buflen;
if (sysctl(mib, 2, buf, &len, NULL, 0) == -1)
return done;
done += len;
buf += len;
buflen -= len;
} while (buflen > 0);
return done;
}
# endif
/*
* syscall_random(): Try to get random data using a system call
* returns the number of bytes returned in buf, or <= 0 on error.
*/
int syscall_random(void *buf, size_t buflen)
{
/*
* Do runtime detection to find getentropy().
*
* Known OSs that should support this:
* - Darwin since 16 (OSX 10.12, IOS 10.0).
* - Solaris since 11.3
* - OpenBSD since 5.6
* - Linux since 3.17 with glibc 2.25
* - FreeBSD since 12.0 (1200061)
*/
# if defined(__GNUC__) && __GNUC__>=2 && defined(__ELF__) && !defined(__hpux)
extern int getentropy(void *bufer, size_t length) __attribute__((weak));
if (getentropy != NULL)
return getentropy(buf, buflen) == 0 ? buflen : 0;
# else
union {
void *p;
int (*f)(void *buffer, size_t length);
} p_getentropy;
/*
* We could cache the result of the lookup, but we normally don't
* call this function often.
*/
ERR_set_mark();
p_getentropy.p = DSO_global_lookup("getentropy");
ERR_pop_to_mark();
if (p_getentropy.p != NULL)
return p_getentropy.f(buf, buflen) == 0 ? buflen : 0;
# endif
/* Linux supports this since version 3.17 */
# if defined(__linux) && defined(SYS_getrandom)
return (int)syscall(SYS_getrandom, buf, buflen, 0);
# endif
# if (defined(__FreeBSD__) || defined(__NetBSD__)) && defined(KERN_ARND)
return (int)sysctl_random(buf, buflen);
# endif
return -1;
}
#if !defined(OPENSSL_RAND_SEED_NONE) && defined(OPENSSL_RAND_SEED_DEVRANDOM)
static const char *random_device_paths[] = { DEVRANDOM };
static struct random_device {
int fd;
dev_t dev;
ino_t ino;
mode_t mode;
dev_t rdev;
} random_devices[OSSL_NELEM(random_device_paths)];
static int keep_random_devices_open = 1;
/*
* Verify that the file descriptor associated with the random source is
* still valid. The rationale for doing this is the fact that it is not
* uncommon for daemons to close all open file handles when daemonizing.
* So the handle might have been closed or even reused for opening
* another file.
*/
static int check_random_device(struct random_device * rd)
{
struct stat st;
return rd->fd != -1
&& fstat(rd->fd, &st) != -1
&& rd->dev == st.st_dev
&& rd->ino == st.st_ino
&& ((rd->mode ^ st.st_mode) & ~(S_IRWXU | S_IRWXG | S_IRWXO)) == 0
&& rd->rdev == st.st_rdev;
}
/*
* Open a random device if required and return its file descriptor or -1 on error
*/
static int get_random_device(size_t n)
{
struct stat st;
struct random_device * rd = &random_devices[n];
/* reuse existing file descriptor if it is (still) valid */
if (check_random_device(rd))
return rd->fd;
/* open the random device ... */
if ((rd->fd = open(random_device_paths[n], O_RDONLY)) == -1)
return rd->fd;
/* ... and cache its relevant stat(2) data */
if (fstat(rd->fd, &st) != -1) {
rd->dev = st.st_dev;
rd->ino = st.st_ino;
rd->mode = st.st_mode;
rd->rdev = st.st_rdev;
} else {
close(rd->fd);
rd->fd = -1;
}
return rd->fd;
}
/*
* Close a random device making sure it is a random device
*/
static void close_random_device(size_t n)
{
struct random_device * rd = &random_devices[n];
if (check_random_device(rd))
close(rd->fd);
rd->fd = -1;
}
static void open_random_devices(void)
{
size_t i;
for (i = 0; i < OSSL_NELEM(random_devices); i++)
(void)get_random_device(i);
}
int rand_pool_init(void)
{
size_t i;
for (i = 0; i < OSSL_NELEM(random_devices); i++)
random_devices[i].fd = -1;
open_random_devices();
return 1;
}
void rand_pool_cleanup(void)
{
size_t i;
for (i = 0; i < OSSL_NELEM(random_devices); i++)
close_random_device(i);
}
void rand_pool_keep_random_devices_open(int keep)
{
if (keep)
open_random_devices();
else
rand_pool_cleanup();
keep_random_devices_open = keep;
}
# else /* defined(OPENSSL_RAND_SEED_NONE)
* || !defined(OPENSSL_RAND_SEED_DEVRANDOM)
*/
int rand_pool_init(void)
{
return 1;
}
void rand_pool_cleanup(void)
{
}
void rand_pool_keep_random_devices_open(int keep)
{
}
# endif /* !defined(OPENSSL_RAND_SEED_NONE)
* && defined(OPENSSL_RAND_SEED_DEVRANDOM)
*/
/*
Fix reseeding issues of the public RAND_DRBG Reseeding is handled very differently by the classic RAND_METHOD API and the new RAND_DRBG api. These differences led to some problems when the new RAND_DRBG was made the default OpenSSL RNG. In particular, RAND_add() did not work as expected anymore. These issues are discussed on the thread '[openssl-dev] Plea for a new public OpenSSL RNG API' and in Pull Request #4328. This commit fixes the mentioned issues, introducing the following changes: - Replace the fixed size RAND_BYTES_BUFFER by a new RAND_POOL API which facilitates collecting entropy by the get_entropy() callback. - Don't use RAND_poll()/RAND_add() for collecting entropy from the get_entropy() callback anymore. Instead, replace RAND_poll() by RAND_POOL_acquire_entropy(). - Add a new function rand_drbg_restart() which tries to get the DRBG in an instantiated state by all means, regardless of the current state (uninstantiated, error, ...) the DRBG is in. If the caller provides entropy or additional input, it will be used for reseeding. - Restore the original documented behaviour of RAND_add() and RAND_poll() (namely to reseed the DRBG immediately) by a new implementation based on rand_drbg_restart(). - Add automatic error recovery from temporary failures of the entropy source to RAND_DRBG_generate() using the rand_drbg_restart() function. Reviewed-by: Paul Dale <paul.dale@oracle.com> Reviewed-by: Kurt Roeckx <kurt@roeckx.be> Reviewed-by: Rich Salz <rsalz@openssl.org> Reviewed-by: Ben Kaduk <kaduk@mit.edu> (Merged from https://github.com/openssl/openssl/pull/4328)
2017-08-31 21:16:22 +00:00
* Try the various seeding methods in turn, exit when successful.
*
* TODO(DRBG): If more than one entropy source is available, is it
* preferable to stop as soon as enough entropy has been collected
* (as favored by @rsalz) or should one rather be defensive and add
* more entropy than requested and/or from different sources?
*
* Currently, the user can select multiple entropy sources in the
* configure step, yet in practice only the first available source
* will be used. A more flexible solution has been requested, but
* currently it is not clear how this can be achieved without
* overengineering the problem. There are many parameters which
* could be taken into account when selecting the order and amount
* of input from the different entropy sources (trust, quality,
* possibility of blocking).
*/
size_t rand_pool_acquire_entropy(RAND_POOL *pool)
{
# ifdef OPENSSL_RAND_SEED_NONE
return rand_pool_entropy_available(pool);
# else
Fix reseeding issues of the public RAND_DRBG Reseeding is handled very differently by the classic RAND_METHOD API and the new RAND_DRBG api. These differences led to some problems when the new RAND_DRBG was made the default OpenSSL RNG. In particular, RAND_add() did not work as expected anymore. These issues are discussed on the thread '[openssl-dev] Plea for a new public OpenSSL RNG API' and in Pull Request #4328. This commit fixes the mentioned issues, introducing the following changes: - Replace the fixed size RAND_BYTES_BUFFER by a new RAND_POOL API which facilitates collecting entropy by the get_entropy() callback. - Don't use RAND_poll()/RAND_add() for collecting entropy from the get_entropy() callback anymore. Instead, replace RAND_poll() by RAND_POOL_acquire_entropy(). - Add a new function rand_drbg_restart() which tries to get the DRBG in an instantiated state by all means, regardless of the current state (uninstantiated, error, ...) the DRBG is in. If the caller provides entropy or additional input, it will be used for reseeding. - Restore the original documented behaviour of RAND_add() and RAND_poll() (namely to reseed the DRBG immediately) by a new implementation based on rand_drbg_restart(). - Add automatic error recovery from temporary failures of the entropy source to RAND_DRBG_generate() using the rand_drbg_restart() function. Reviewed-by: Paul Dale <paul.dale@oracle.com> Reviewed-by: Kurt Roeckx <kurt@roeckx.be> Reviewed-by: Rich Salz <rsalz@openssl.org> Reviewed-by: Ben Kaduk <kaduk@mit.edu> (Merged from https://github.com/openssl/openssl/pull/4328)
2017-08-31 21:16:22 +00:00
size_t bytes_needed;
size_t entropy_available = 0;
unsigned char *buffer;
# ifdef OPENSSL_RAND_SEED_GETRANDOM
bytes_needed = rand_pool_bytes_needed(pool, 1 /*entropy_factor*/);
buffer = rand_pool_add_begin(pool, bytes_needed);
Fix reseeding issues of the public RAND_DRBG Reseeding is handled very differently by the classic RAND_METHOD API and the new RAND_DRBG api. These differences led to some problems when the new RAND_DRBG was made the default OpenSSL RNG. In particular, RAND_add() did not work as expected anymore. These issues are discussed on the thread '[openssl-dev] Plea for a new public OpenSSL RNG API' and in Pull Request #4328. This commit fixes the mentioned issues, introducing the following changes: - Replace the fixed size RAND_BYTES_BUFFER by a new RAND_POOL API which facilitates collecting entropy by the get_entropy() callback. - Don't use RAND_poll()/RAND_add() for collecting entropy from the get_entropy() callback anymore. Instead, replace RAND_poll() by RAND_POOL_acquire_entropy(). - Add a new function rand_drbg_restart() which tries to get the DRBG in an instantiated state by all means, regardless of the current state (uninstantiated, error, ...) the DRBG is in. If the caller provides entropy or additional input, it will be used for reseeding. - Restore the original documented behaviour of RAND_add() and RAND_poll() (namely to reseed the DRBG immediately) by a new implementation based on rand_drbg_restart(). - Add automatic error recovery from temporary failures of the entropy source to RAND_DRBG_generate() using the rand_drbg_restart() function. Reviewed-by: Paul Dale <paul.dale@oracle.com> Reviewed-by: Kurt Roeckx <kurt@roeckx.be> Reviewed-by: Rich Salz <rsalz@openssl.org> Reviewed-by: Ben Kaduk <kaduk@mit.edu> (Merged from https://github.com/openssl/openssl/pull/4328)
2017-08-31 21:16:22 +00:00
if (buffer != NULL) {
size_t bytes = 0;
if (syscall_random(buffer, bytes_needed) == (int)bytes_needed)
Fix reseeding issues of the public RAND_DRBG Reseeding is handled very differently by the classic RAND_METHOD API and the new RAND_DRBG api. These differences led to some problems when the new RAND_DRBG was made the default OpenSSL RNG. In particular, RAND_add() did not work as expected anymore. These issues are discussed on the thread '[openssl-dev] Plea for a new public OpenSSL RNG API' and in Pull Request #4328. This commit fixes the mentioned issues, introducing the following changes: - Replace the fixed size RAND_BYTES_BUFFER by a new RAND_POOL API which facilitates collecting entropy by the get_entropy() callback. - Don't use RAND_poll()/RAND_add() for collecting entropy from the get_entropy() callback anymore. Instead, replace RAND_poll() by RAND_POOL_acquire_entropy(). - Add a new function rand_drbg_restart() which tries to get the DRBG in an instantiated state by all means, regardless of the current state (uninstantiated, error, ...) the DRBG is in. If the caller provides entropy or additional input, it will be used for reseeding. - Restore the original documented behaviour of RAND_add() and RAND_poll() (namely to reseed the DRBG immediately) by a new implementation based on rand_drbg_restart(). - Add automatic error recovery from temporary failures of the entropy source to RAND_DRBG_generate() using the rand_drbg_restart() function. Reviewed-by: Paul Dale <paul.dale@oracle.com> Reviewed-by: Kurt Roeckx <kurt@roeckx.be> Reviewed-by: Rich Salz <rsalz@openssl.org> Reviewed-by: Ben Kaduk <kaduk@mit.edu> (Merged from https://github.com/openssl/openssl/pull/4328)
2017-08-31 21:16:22 +00:00
bytes = bytes_needed;
rand_pool_add_end(pool, bytes, 8 * bytes);
entropy_available = rand_pool_entropy_available(pool);
}
Fix reseeding issues of the public RAND_DRBG Reseeding is handled very differently by the classic RAND_METHOD API and the new RAND_DRBG api. These differences led to some problems when the new RAND_DRBG was made the default OpenSSL RNG. In particular, RAND_add() did not work as expected anymore. These issues are discussed on the thread '[openssl-dev] Plea for a new public OpenSSL RNG API' and in Pull Request #4328. This commit fixes the mentioned issues, introducing the following changes: - Replace the fixed size RAND_BYTES_BUFFER by a new RAND_POOL API which facilitates collecting entropy by the get_entropy() callback. - Don't use RAND_poll()/RAND_add() for collecting entropy from the get_entropy() callback anymore. Instead, replace RAND_poll() by RAND_POOL_acquire_entropy(). - Add a new function rand_drbg_restart() which tries to get the DRBG in an instantiated state by all means, regardless of the current state (uninstantiated, error, ...) the DRBG is in. If the caller provides entropy or additional input, it will be used for reseeding. - Restore the original documented behaviour of RAND_add() and RAND_poll() (namely to reseed the DRBG immediately) by a new implementation based on rand_drbg_restart(). - Add automatic error recovery from temporary failures of the entropy source to RAND_DRBG_generate() using the rand_drbg_restart() function. Reviewed-by: Paul Dale <paul.dale@oracle.com> Reviewed-by: Kurt Roeckx <kurt@roeckx.be> Reviewed-by: Rich Salz <rsalz@openssl.org> Reviewed-by: Ben Kaduk <kaduk@mit.edu> (Merged from https://github.com/openssl/openssl/pull/4328)
2017-08-31 21:16:22 +00:00
if (entropy_available > 0)
return entropy_available;
# endif
# if defined(OPENSSL_RAND_SEED_LIBRANDOM)
{
/* Not yet implemented. */
}
# endif
# ifdef OPENSSL_RAND_SEED_DEVRANDOM
bytes_needed = rand_pool_bytes_needed(pool, 1 /*entropy_factor*/);
{
size_t i;
for (i = 0; bytes_needed > 0 && i < OSSL_NELEM(random_device_paths); i++) {
const int fd = get_random_device(i);
if (fd == -1)
continue;
buffer = rand_pool_add_begin(pool, bytes_needed);
Fix reseeding issues of the public RAND_DRBG Reseeding is handled very differently by the classic RAND_METHOD API and the new RAND_DRBG api. These differences led to some problems when the new RAND_DRBG was made the default OpenSSL RNG. In particular, RAND_add() did not work as expected anymore. These issues are discussed on the thread '[openssl-dev] Plea for a new public OpenSSL RNG API' and in Pull Request #4328. This commit fixes the mentioned issues, introducing the following changes: - Replace the fixed size RAND_BYTES_BUFFER by a new RAND_POOL API which facilitates collecting entropy by the get_entropy() callback. - Don't use RAND_poll()/RAND_add() for collecting entropy from the get_entropy() callback anymore. Instead, replace RAND_poll() by RAND_POOL_acquire_entropy(). - Add a new function rand_drbg_restart() which tries to get the DRBG in an instantiated state by all means, regardless of the current state (uninstantiated, error, ...) the DRBG is in. If the caller provides entropy or additional input, it will be used for reseeding. - Restore the original documented behaviour of RAND_add() and RAND_poll() (namely to reseed the DRBG immediately) by a new implementation based on rand_drbg_restart(). - Add automatic error recovery from temporary failures of the entropy source to RAND_DRBG_generate() using the rand_drbg_restart() function. Reviewed-by: Paul Dale <paul.dale@oracle.com> Reviewed-by: Kurt Roeckx <kurt@roeckx.be> Reviewed-by: Rich Salz <rsalz@openssl.org> Reviewed-by: Ben Kaduk <kaduk@mit.edu> (Merged from https://github.com/openssl/openssl/pull/4328)
2017-08-31 21:16:22 +00:00
if (buffer != NULL) {
const ssize_t n = read(fd, buffer, bytes_needed);
Fix reseeding issues of the public RAND_DRBG Reseeding is handled very differently by the classic RAND_METHOD API and the new RAND_DRBG api. These differences led to some problems when the new RAND_DRBG was made the default OpenSSL RNG. In particular, RAND_add() did not work as expected anymore. These issues are discussed on the thread '[openssl-dev] Plea for a new public OpenSSL RNG API' and in Pull Request #4328. This commit fixes the mentioned issues, introducing the following changes: - Replace the fixed size RAND_BYTES_BUFFER by a new RAND_POOL API which facilitates collecting entropy by the get_entropy() callback. - Don't use RAND_poll()/RAND_add() for collecting entropy from the get_entropy() callback anymore. Instead, replace RAND_poll() by RAND_POOL_acquire_entropy(). - Add a new function rand_drbg_restart() which tries to get the DRBG in an instantiated state by all means, regardless of the current state (uninstantiated, error, ...) the DRBG is in. If the caller provides entropy or additional input, it will be used for reseeding. - Restore the original documented behaviour of RAND_add() and RAND_poll() (namely to reseed the DRBG immediately) by a new implementation based on rand_drbg_restart(). - Add automatic error recovery from temporary failures of the entropy source to RAND_DRBG_generate() using the rand_drbg_restart() function. Reviewed-by: Paul Dale <paul.dale@oracle.com> Reviewed-by: Kurt Roeckx <kurt@roeckx.be> Reviewed-by: Rich Salz <rsalz@openssl.org> Reviewed-by: Ben Kaduk <kaduk@mit.edu> (Merged from https://github.com/openssl/openssl/pull/4328)
2017-08-31 21:16:22 +00:00
if (n <= 0) {
close_random_device(i);
continue;
}
rand_pool_add_end(pool, n, 8 * n);
}
if (!keep_random_devices_open)
close_random_device(i);
Fix reseeding issues of the public RAND_DRBG Reseeding is handled very differently by the classic RAND_METHOD API and the new RAND_DRBG api. These differences led to some problems when the new RAND_DRBG was made the default OpenSSL RNG. In particular, RAND_add() did not work as expected anymore. These issues are discussed on the thread '[openssl-dev] Plea for a new public OpenSSL RNG API' and in Pull Request #4328. This commit fixes the mentioned issues, introducing the following changes: - Replace the fixed size RAND_BYTES_BUFFER by a new RAND_POOL API which facilitates collecting entropy by the get_entropy() callback. - Don't use RAND_poll()/RAND_add() for collecting entropy from the get_entropy() callback anymore. Instead, replace RAND_poll() by RAND_POOL_acquire_entropy(). - Add a new function rand_drbg_restart() which tries to get the DRBG in an instantiated state by all means, regardless of the current state (uninstantiated, error, ...) the DRBG is in. If the caller provides entropy or additional input, it will be used for reseeding. - Restore the original documented behaviour of RAND_add() and RAND_poll() (namely to reseed the DRBG immediately) by a new implementation based on rand_drbg_restart(). - Add automatic error recovery from temporary failures of the entropy source to RAND_DRBG_generate() using the rand_drbg_restart() function. Reviewed-by: Paul Dale <paul.dale@oracle.com> Reviewed-by: Kurt Roeckx <kurt@roeckx.be> Reviewed-by: Rich Salz <rsalz@openssl.org> Reviewed-by: Ben Kaduk <kaduk@mit.edu> (Merged from https://github.com/openssl/openssl/pull/4328)
2017-08-31 21:16:22 +00:00
bytes_needed = rand_pool_bytes_needed(pool, 1 /*entropy_factor*/);
}
entropy_available = rand_pool_entropy_available(pool);
if (entropy_available > 0)
return entropy_available;
}
# endif
# ifdef OPENSSL_RAND_SEED_RDTSC
Fix reseeding issues of the public RAND_DRBG Reseeding is handled very differently by the classic RAND_METHOD API and the new RAND_DRBG api. These differences led to some problems when the new RAND_DRBG was made the default OpenSSL RNG. In particular, RAND_add() did not work as expected anymore. These issues are discussed on the thread '[openssl-dev] Plea for a new public OpenSSL RNG API' and in Pull Request #4328. This commit fixes the mentioned issues, introducing the following changes: - Replace the fixed size RAND_BYTES_BUFFER by a new RAND_POOL API which facilitates collecting entropy by the get_entropy() callback. - Don't use RAND_poll()/RAND_add() for collecting entropy from the get_entropy() callback anymore. Instead, replace RAND_poll() by RAND_POOL_acquire_entropy(). - Add a new function rand_drbg_restart() which tries to get the DRBG in an instantiated state by all means, regardless of the current state (uninstantiated, error, ...) the DRBG is in. If the caller provides entropy or additional input, it will be used for reseeding. - Restore the original documented behaviour of RAND_add() and RAND_poll() (namely to reseed the DRBG immediately) by a new implementation based on rand_drbg_restart(). - Add automatic error recovery from temporary failures of the entropy source to RAND_DRBG_generate() using the rand_drbg_restart() function. Reviewed-by: Paul Dale <paul.dale@oracle.com> Reviewed-by: Kurt Roeckx <kurt@roeckx.be> Reviewed-by: Rich Salz <rsalz@openssl.org> Reviewed-by: Ben Kaduk <kaduk@mit.edu> (Merged from https://github.com/openssl/openssl/pull/4328)
2017-08-31 21:16:22 +00:00
entropy_available = rand_acquire_entropy_from_tsc(pool);
if (entropy_available > 0)
return entropy_available;
# endif
# ifdef OPENSSL_RAND_SEED_RDCPU
Fix reseeding issues of the public RAND_DRBG Reseeding is handled very differently by the classic RAND_METHOD API and the new RAND_DRBG api. These differences led to some problems when the new RAND_DRBG was made the default OpenSSL RNG. In particular, RAND_add() did not work as expected anymore. These issues are discussed on the thread '[openssl-dev] Plea for a new public OpenSSL RNG API' and in Pull Request #4328. This commit fixes the mentioned issues, introducing the following changes: - Replace the fixed size RAND_BYTES_BUFFER by a new RAND_POOL API which facilitates collecting entropy by the get_entropy() callback. - Don't use RAND_poll()/RAND_add() for collecting entropy from the get_entropy() callback anymore. Instead, replace RAND_poll() by RAND_POOL_acquire_entropy(). - Add a new function rand_drbg_restart() which tries to get the DRBG in an instantiated state by all means, regardless of the current state (uninstantiated, error, ...) the DRBG is in. If the caller provides entropy or additional input, it will be used for reseeding. - Restore the original documented behaviour of RAND_add() and RAND_poll() (namely to reseed the DRBG immediately) by a new implementation based on rand_drbg_restart(). - Add automatic error recovery from temporary failures of the entropy source to RAND_DRBG_generate() using the rand_drbg_restart() function. Reviewed-by: Paul Dale <paul.dale@oracle.com> Reviewed-by: Kurt Roeckx <kurt@roeckx.be> Reviewed-by: Rich Salz <rsalz@openssl.org> Reviewed-by: Ben Kaduk <kaduk@mit.edu> (Merged from https://github.com/openssl/openssl/pull/4328)
2017-08-31 21:16:22 +00:00
entropy_available = rand_acquire_entropy_from_cpu(pool);
if (entropy_available > 0)
return entropy_available;
# endif
# ifdef OPENSSL_RAND_SEED_EGD
bytes_needed = rand_pool_bytes_needed(pool, 1 /*entropy_factor*/);
Fix reseeding issues of the public RAND_DRBG Reseeding is handled very differently by the classic RAND_METHOD API and the new RAND_DRBG api. These differences led to some problems when the new RAND_DRBG was made the default OpenSSL RNG. In particular, RAND_add() did not work as expected anymore. These issues are discussed on the thread '[openssl-dev] Plea for a new public OpenSSL RNG API' and in Pull Request #4328. This commit fixes the mentioned issues, introducing the following changes: - Replace the fixed size RAND_BYTES_BUFFER by a new RAND_POOL API which facilitates collecting entropy by the get_entropy() callback. - Don't use RAND_poll()/RAND_add() for collecting entropy from the get_entropy() callback anymore. Instead, replace RAND_poll() by RAND_POOL_acquire_entropy(). - Add a new function rand_drbg_restart() which tries to get the DRBG in an instantiated state by all means, regardless of the current state (uninstantiated, error, ...) the DRBG is in. If the caller provides entropy or additional input, it will be used for reseeding. - Restore the original documented behaviour of RAND_add() and RAND_poll() (namely to reseed the DRBG immediately) by a new implementation based on rand_drbg_restart(). - Add automatic error recovery from temporary failures of the entropy source to RAND_DRBG_generate() using the rand_drbg_restart() function. Reviewed-by: Paul Dale <paul.dale@oracle.com> Reviewed-by: Kurt Roeckx <kurt@roeckx.be> Reviewed-by: Rich Salz <rsalz@openssl.org> Reviewed-by: Ben Kaduk <kaduk@mit.edu> (Merged from https://github.com/openssl/openssl/pull/4328)
2017-08-31 21:16:22 +00:00
if (bytes_needed > 0) {
static const char *paths[] = { DEVRANDOM_EGD, NULL };
int i;
for (i = 0; paths[i] != NULL; i++) {
buffer = rand_pool_add_begin(pool, bytes_needed);
Fix reseeding issues of the public RAND_DRBG Reseeding is handled very differently by the classic RAND_METHOD API and the new RAND_DRBG api. These differences led to some problems when the new RAND_DRBG was made the default OpenSSL RNG. In particular, RAND_add() did not work as expected anymore. These issues are discussed on the thread '[openssl-dev] Plea for a new public OpenSSL RNG API' and in Pull Request #4328. This commit fixes the mentioned issues, introducing the following changes: - Replace the fixed size RAND_BYTES_BUFFER by a new RAND_POOL API which facilitates collecting entropy by the get_entropy() callback. - Don't use RAND_poll()/RAND_add() for collecting entropy from the get_entropy() callback anymore. Instead, replace RAND_poll() by RAND_POOL_acquire_entropy(). - Add a new function rand_drbg_restart() which tries to get the DRBG in an instantiated state by all means, regardless of the current state (uninstantiated, error, ...) the DRBG is in. If the caller provides entropy or additional input, it will be used for reseeding. - Restore the original documented behaviour of RAND_add() and RAND_poll() (namely to reseed the DRBG immediately) by a new implementation based on rand_drbg_restart(). - Add automatic error recovery from temporary failures of the entropy source to RAND_DRBG_generate() using the rand_drbg_restart() function. Reviewed-by: Paul Dale <paul.dale@oracle.com> Reviewed-by: Kurt Roeckx <kurt@roeckx.be> Reviewed-by: Rich Salz <rsalz@openssl.org> Reviewed-by: Ben Kaduk <kaduk@mit.edu> (Merged from https://github.com/openssl/openssl/pull/4328)
2017-08-31 21:16:22 +00:00
if (buffer != NULL) {
size_t bytes = 0;
int num = RAND_query_egd_bytes(paths[i],
buffer, (int)bytes_needed);
if (num == (int)bytes_needed)
bytes = bytes_needed;
rand_pool_add_end(pool, bytes, 8 * bytes);
entropy_available = rand_pool_entropy_available(pool);
}
Fix reseeding issues of the public RAND_DRBG Reseeding is handled very differently by the classic RAND_METHOD API and the new RAND_DRBG api. These differences led to some problems when the new RAND_DRBG was made the default OpenSSL RNG. In particular, RAND_add() did not work as expected anymore. These issues are discussed on the thread '[openssl-dev] Plea for a new public OpenSSL RNG API' and in Pull Request #4328. This commit fixes the mentioned issues, introducing the following changes: - Replace the fixed size RAND_BYTES_BUFFER by a new RAND_POOL API which facilitates collecting entropy by the get_entropy() callback. - Don't use RAND_poll()/RAND_add() for collecting entropy from the get_entropy() callback anymore. Instead, replace RAND_poll() by RAND_POOL_acquire_entropy(). - Add a new function rand_drbg_restart() which tries to get the DRBG in an instantiated state by all means, regardless of the current state (uninstantiated, error, ...) the DRBG is in. If the caller provides entropy or additional input, it will be used for reseeding. - Restore the original documented behaviour of RAND_add() and RAND_poll() (namely to reseed the DRBG immediately) by a new implementation based on rand_drbg_restart(). - Add automatic error recovery from temporary failures of the entropy source to RAND_DRBG_generate() using the rand_drbg_restart() function. Reviewed-by: Paul Dale <paul.dale@oracle.com> Reviewed-by: Kurt Roeckx <kurt@roeckx.be> Reviewed-by: Rich Salz <rsalz@openssl.org> Reviewed-by: Ben Kaduk <kaduk@mit.edu> (Merged from https://github.com/openssl/openssl/pull/4328)
2017-08-31 21:16:22 +00:00
if (entropy_available > 0)
return entropy_available;
}
}
# endif
return rand_pool_entropy_available(pool);
# endif
}
# endif
DRBG: implement a get_nonce() callback Fixes #5849 In pull request #5503 a fallback was added which adds a random nonce of security_strength/2 bits if no nonce callback is provided. This change raised the entropy requirements form 256 to 384 bit, which can cause problems on some platforms (e.g. VMS, see issue #5849). The requirements for the nonce are given in section 8.6.7 of NIST SP 800-90Ar1: A nonce may be required in the construction of a seed during instantiation in order to provide a security cushion to block certain attacks. The nonce shall be either: a) A value with at least (security_strength/2) bits of entropy, or b) A value that is expected to repeat no more often than a (security_strength/2)-bit random string would be expected to repeat. Each nonce shall be unique to the cryptographic module in which instantiation is performed, but need not be secret. When used, the nonce shall be considered to be a critical security parameter. This commit implements a nonce of type b) in order to lower the entropy requirements during instantiation back to 256 bits. The formulation "shall be unique to the cryptographic module" above implies that the nonce needs to be unique among (with high probability) among all DRBG instances in "space" and "time". We try to achieve this goal by creating a nonce of the following form nonce = app-specific-data || high-resolution-utc-timestamp || counter Where || denotes concatenation. The application specific data can be something like the process or group id of the application. A utc timestamp is used because it increases monotonically, provided the system time is synchronized. This approach may not be perfect yet for a FIPS evaluation, but it should be good enough for the moment. This commit also harmonizes the implementation of the get_nonce() and the get_additional_data() callbacks and moves the platform specific parts from rand_lib.c into rand_unix.c, rand_win.c, and rand_vms.c. Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/5920)
2018-04-10 08:22:52 +00:00
#endif
#if defined(OPENSSL_SYS_UNIX) || defined(__DJGPP__)
DRBG: implement a get_nonce() callback Fixes #5849 In pull request #5503 a fallback was added which adds a random nonce of security_strength/2 bits if no nonce callback is provided. This change raised the entropy requirements form 256 to 384 bit, which can cause problems on some platforms (e.g. VMS, see issue #5849). The requirements for the nonce are given in section 8.6.7 of NIST SP 800-90Ar1: A nonce may be required in the construction of a seed during instantiation in order to provide a security cushion to block certain attacks. The nonce shall be either: a) A value with at least (security_strength/2) bits of entropy, or b) A value that is expected to repeat no more often than a (security_strength/2)-bit random string would be expected to repeat. Each nonce shall be unique to the cryptographic module in which instantiation is performed, but need not be secret. When used, the nonce shall be considered to be a critical security parameter. This commit implements a nonce of type b) in order to lower the entropy requirements during instantiation back to 256 bits. The formulation "shall be unique to the cryptographic module" above implies that the nonce needs to be unique among (with high probability) among all DRBG instances in "space" and "time". We try to achieve this goal by creating a nonce of the following form nonce = app-specific-data || high-resolution-utc-timestamp || counter Where || denotes concatenation. The application specific data can be something like the process or group id of the application. A utc timestamp is used because it increases monotonically, provided the system time is synchronized. This approach may not be perfect yet for a FIPS evaluation, but it should be good enough for the moment. This commit also harmonizes the implementation of the get_nonce() and the get_additional_data() callbacks and moves the platform specific parts from rand_lib.c into rand_unix.c, rand_win.c, and rand_vms.c. Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/5920)
2018-04-10 08:22:52 +00:00
int rand_pool_add_nonce_data(RAND_POOL *pool)
{
struct {
pid_t pid;
CRYPTO_THREAD_ID tid;
uint64_t time;
} data = { 0 };
/*
* Add process id, thread id, and a high resolution timestamp to
* ensure that the nonce is unique whith high probability for
* different process instances.
*/
data.pid = getpid();
data.tid = CRYPTO_THREAD_get_current_id();
data.time = get_time_stamp();
return rand_pool_add(pool, (unsigned char *)&data, sizeof(data), 0);
}
int rand_pool_add_additional_data(RAND_POOL *pool)
{
struct {
CRYPTO_THREAD_ID tid;
uint64_t time;
} data = { 0 };
/*
* Add some noise from the thread id and a high resolution timer.
* The thread id adds a little randomness if the drbg is accessed
* concurrently (which is the case for the <master> drbg).
*/
data.tid = CRYPTO_THREAD_get_current_id();
data.time = get_timer_bits();
return rand_pool_add(pool, (unsigned char *)&data, sizeof(data), 0);
}
/*
* Get the current time with the highest possible resolution
*
* The time stamp is added to the nonce, so it is optimized for not repeating.
* The current time is ideal for this purpose, provided the computer's clock
* is synchronized.
*/
static uint64_t get_time_stamp(void)
{
# if defined(OSSL_POSIX_TIMER_OKAY)
{
struct timespec ts;
if (clock_gettime(CLOCK_REALTIME, &ts) == 0)
return TWO32TO64(ts.tv_sec, ts.tv_nsec);
}
# endif
# if defined(__unix__) \
|| (defined(_POSIX_C_SOURCE) && _POSIX_C_SOURCE >= 200112L)
{
struct timeval tv;
if (gettimeofday(&tv, NULL) == 0)
return TWO32TO64(tv.tv_sec, tv.tv_usec);
}
# endif
return time(NULL);
}
/*
* Get an arbitrary timer value of the highest possible resolution
*
* The timer value is added as random noise to the additional data,
* which is not considered a trusted entropy sourec, so any result
* is acceptable.
*/
static uint64_t get_timer_bits(void)
{
uint64_t res = OPENSSL_rdtsc();
if (res != 0)
return res;
# if defined(__sun) || defined(__hpux)
return gethrtime();
# elif defined(_AIX)
{
timebasestruct_t t;
read_wall_time(&t, TIMEBASE_SZ);
return TWO32TO64(t.tb_high, t.tb_low);
}
# elif defined(OSSL_POSIX_TIMER_OKAY)
{
struct timespec ts;
# ifdef CLOCK_BOOTTIME
# define CLOCK_TYPE CLOCK_BOOTTIME
# elif defined(_POSIX_MONOTONIC_CLOCK)
# define CLOCK_TYPE CLOCK_MONOTONIC
# else
# define CLOCK_TYPE CLOCK_REALTIME
# endif
if (clock_gettime(CLOCK_TYPE, &ts) == 0)
return TWO32TO64(ts.tv_sec, ts.tv_nsec);
}
# endif
# if defined(__unix__) \
|| (defined(_POSIX_C_SOURCE) && _POSIX_C_SOURCE >= 200112L)
{
struct timeval tv;
DRBG: implement a get_nonce() callback Fixes #5849 In pull request #5503 a fallback was added which adds a random nonce of security_strength/2 bits if no nonce callback is provided. This change raised the entropy requirements form 256 to 384 bit, which can cause problems on some platforms (e.g. VMS, see issue #5849). The requirements for the nonce are given in section 8.6.7 of NIST SP 800-90Ar1: A nonce may be required in the construction of a seed during instantiation in order to provide a security cushion to block certain attacks. The nonce shall be either: a) A value with at least (security_strength/2) bits of entropy, or b) A value that is expected to repeat no more often than a (security_strength/2)-bit random string would be expected to repeat. Each nonce shall be unique to the cryptographic module in which instantiation is performed, but need not be secret. When used, the nonce shall be considered to be a critical security parameter. This commit implements a nonce of type b) in order to lower the entropy requirements during instantiation back to 256 bits. The formulation "shall be unique to the cryptographic module" above implies that the nonce needs to be unique among (with high probability) among all DRBG instances in "space" and "time". We try to achieve this goal by creating a nonce of the following form nonce = app-specific-data || high-resolution-utc-timestamp || counter Where || denotes concatenation. The application specific data can be something like the process or group id of the application. A utc timestamp is used because it increases monotonically, provided the system time is synchronized. This approach may not be perfect yet for a FIPS evaluation, but it should be good enough for the moment. This commit also harmonizes the implementation of the get_nonce() and the get_additional_data() callbacks and moves the platform specific parts from rand_lib.c into rand_unix.c, rand_win.c, and rand_vms.c. Reviewed-by: Richard Levitte <levitte@openssl.org> (Merged from https://github.com/openssl/openssl/pull/5920)
2018-04-10 08:22:52 +00:00
if (gettimeofday(&tv, NULL) == 0)
return TWO32TO64(tv.tv_sec, tv.tv_usec);
}
# endif
return time(NULL);
}
#endif /* defined(OPENSSL_SYS_UNIX) || defined(__DJGPP__) */