|Special device files|
In Unix-like operating systems, /dev/random is a special file that serves as a blocking pseudorandom number generator. It allows access to environmental noise collected from device drivers and other sources. Not all operating systems implement the same semantics for /dev/random.
Random number generation from kernel space was implemented for the first time for Linux in 1994 by Theodore Ts'o. The implementation uses secure hashes rather than ciphers, to avoid legal restrictions that were in place when the generator was originally designed. The implementation was also designed with the assumption that any given hash or cipher might eventually be found to be weak, and so the design is durable in the face of any such weaknesses. Fast recovery from pool compromise is not considered a requirement, because the requirements for pool compromise are sufficient for much easier and more direct attacks on unrelated parts of the operating system.
In this implementation, the generator keeps an estimate of the number of bits of noise in the entropy pool. From this entropy pool random numbers are created. When read, the
/dev/random device will only return random bytes within the estimated number of bits of noise in the entropy pool.
/dev/random should be suitable for uses that need very high quality randomness such as one-time pad or key generation. When the entropy pool is empty, reads from
/dev/random will block until additional environmental noise is gathered. The intent is to serve as a cryptographically secure pseudorandom number generator, delivering output with entropy as large as possible. This is suggested by the authors for use in generating cryptographic keys for high-value or long-term protection.
A counterpart to
/dev/random is /dev/urandom ("unlimited"/non-blocking random source) which reuses the internal pool to produce more pseudo-random bits. This means that the call will not block, but the output may contain less entropy than the corresponding read from
/dev/urandom is still intended as a pseudorandom number generator suitable for most cryptographic purposes, some people claim
/dev/urandom as not recommended[who?] for the generation of long-term cryptographic keys. However this is in general not the case because once the entropy pool is unpredictable it doesn't leak security by a reduced number of bits.
It is also possible to write to
/dev/random. This allows any user to mix random data into the pool. Non-random data is harmless, because only a privileged user can issue the ioctl needed to increase the entropy estimate. The current amount of entropy and the size of the Linux kernel entropy pool are available in
/proc/sys/kernel/random/, which can be displayed by the command
Gutterman, Pinkas, & Reinman in March 2006 published a detailed cryptographic analysis of the Linux random number generator in which they describe several weaknesses. Perhaps the most severe issue they report is with embedded or Live CD systems, such as routers and diskless clients, for which the bootup state is predictable and the available supply of entropy from the environment may be limited. For a system with non-volatile memory, they recommend saving some state from the RNG at shutdown so that it can be included in the RNG state on the next reboot. In the case of a router for which network traffic represents the primary available source of entropy, they note that saving state across reboots "would require potential attackers to either eavesdrop on all network traffic" from when the router is first put into service, or obtain direct access to the router's internal state. This issue, they note, is particularly critical in the case of a wireless router whose network traffic can be captured from a distance, and which may be using the RNG to generate keys for data encryption.
With Linux kernel 3.16 and newer, the kernel itself mixes data from hardware random number generators into
/dev/random on a sliding scale based on the definable entropy estimation quality of the HWRNG. This means that no userspace daemon, such as
rng-tools is needed to do that job. With Linux kernel 3.17+, the VirtIO RNG was modified to have a default quality defined above 0, and as such, is currently the only HWRNG mixed into
/dev/random by default.
The entropy pool can be improved by programs like
randomsound etc. With
rng-tools, hardware random number generators like Entropy Key, etc. can write to
/dev/random. The programs
ent can test these random number generators.
In January 2014, Daniel J. Bernstein published a critique of how Linux mixes different sources of entropy. He outlines an attack in which one source of entropy capable of monitoring the other sources of entropy could modify its output to nullify the randomness of the other sources of entropy. Consider the function H(x,y,z) where H is a hash function and x, y, and z are sources of entropy with z being the output of a CPU based malicious HRNG Z:
- Z generates a random value of r.
- Z computes H(x,y,r).
- If the output of H(x,y,r) is equal to the desired value, output r as z.
- Else, repeat starting at 1.
Bernstein estimated that an attacker would need to repeat H(x,y,r) 16 times to compromise DSA and ECDSA. This is possible because Linux reseeds H on an ongoing basis instead of using a single high quality seed.
Is there any serious argument that adding new entropy all the time is a good thing? The Linux /dev/urandom manual page claims that without new entropy the user is "theoretically vulnerable to a cryptographic attack", but (as I've mentioned in various venues) this is a ludicrous argument—how can anyone simultaneously believe that
- we can't figure out how to deterministically expand one 256-bit secret into an endless stream of unpredictable keys (this is what we need from urandom), but
- we can figure out how to use a single key to safely encrypt many messages (this is what we need from SSL, PGP, etc.)?
The FreeBSD operating system implements provides
/dev/urandom for compatibility but the behavior is very different from that of Linux. On FreeBSD,
/dev/urandom is just a link to
/dev/random and only blocks until properly seeded. FreeBSD's PRNG (Fortuna) reseeds regularly but does not attempt to estimate entropy. On a system with small amount of network and disk activity, reseeding is done after a fraction of a second.
While entropy pool based methods are completely secure if implemented correctly, if they overestimate their entropy they may become less secure than well-seeded PRNGs. In some cases an attacker may have a considerable amount of control over the entropy; for example a diskless server may get almost all of it from the network, rendering it potentially vulnerable to man-in-the-middle attacks.
Since OpenBSD 5.1 (May 1, 2012) /dev/random and /dev/arandom use an algorithm based on RC4 but renamed, for licensing purposes, ARC4. While random number generation here uses system entropy gathered in several ways, the ARC4 algorithm provides a fail-safe, ensuring that a rapid and high quality pseudo-random number stream is provided even on a low entropy pool. The system automatically uses hardware random number generators (such as those provided on some Intel PCI hubs) if they are available, through the OpenBSD Cryptographic Framework.
As of OpenBSD 5.5 (May 1, 2014), the arc4random() call used for OpenBSD's random devices no longer uses ARC4, but ChaCha20 (arc4random name might be reconsidered as A Replacement Call for Random). NetBSD's implementation of the legacy arc4random() API has also been switched over to ChaCha20 as well.
OS X and iOS
Other operating systems
/dev/urandom are also available on Solaris, NetBSD, Tru64 UNIX 5.1B, AIX 5.2 and HP-UX 11i v2. As with FreeBSD, AIX implements its own Yarrow-based design, however AIX uses considerably fewer entropy sources than the standard
/dev/random implementation and stops refilling the pool when it thinks it contains enough entropy.
In Windows NT, similar functionality is delivered by
ksecdd.sys, but reading the special file
\Device\KsecDD does not work as in UNIX. The documented methods to generate cryptographically random bytes are CryptGenRandom and RtlGenRandom.
While DOS doesn't naturally provide such functionality, there is an open-source third-party driver called
noise.sys, which functions similarly in that it creates two devices,
URANDOM$, which are also accessible as
/DEV/URANDOM$, that programs can access for random data.
The Linux emulator Cygwin on Windows provide implementations of both
/dev/urandom, which can be used in scripts and programs.
EGD as an alternative
A software program called EGD (entropy gathering daemon) is a common alternative for Unix systems that do not support the /dev/random device. It is a user-space daemon, which provides high-quality cryptographic random data. Some cryptographic software such as OpenSSL, GNU Privacy Guard, and the Apache HTTP Server support using EGD when a /dev/random device is not available. OpenSSL disabled support for the EGD daemon by default in OpenSSL 1.1.0; applications should check for support using the OPENSSL_NO_EGD preprocessor macro.
EGD, gathers random entropy from various sources, process it to remove bias and improve cryptographic quality, and then makes it available over a Unix domain socket (with /dev/egd-pool being a common choice) or over a TCP socket. The entropy gathering usually entails periodically forking subprocesses to query attributes of the system that are likely to be frequently changing and unpredictable, such as monitoring CPU, I/O, and network usage as well as the contents of various log files and temporary directories.
The alternative PRNGD, is a compatible pseudo-random source.
- command 0: query the amount of entropy currently available. The EGD daemon returns a 4-byte number in big-endian format representing the number of random bytes that can currently be satisfied without delay.
- command 1: get random bytes, no blocking. The second byte in the request tells EGD how many random bytes of output it should return, from 1 to 255. If EGD does not have enough entropy to immediately satisfy the request, then fewer bytes, or perhaps no bytes, may be returned. The first octet of the reply indicates how many additional bytes, those containing the random data, immediately follow in the reply.
- command 2: get random bytes, blocking. The second byte tells EGD how many random bytes of output it should return. If EGD does not have enough entropy, it will wait until it has gathered enough before responding. Unlike command 1, the reply starts immediately with the random bytes rather than a length octet, as the total length of returned data will not vary from the amount requested.
- command 3: update entropy. This command allows the client to provide additional entropy to be added to EGD's internal pool. The next two bytes, interpreted as a 16-bit big-endian integer indicate how many bits of randomness the caller is claiming to be supplying. The fourth byte indicates how many additional bytes of source data follow in the request. The EGD daemon may mix in the received entropy and will return nothing back.
- Entropy-supplying system calls
- Standard streams
- Hardware random number generator
- Cryptographically secure pseudorandom number generator
- Fortuna algorithm
- CryptGenRandom—the Microsoft Windows API's CSPRNG
- Torvalds, Linus (2005-04-16). "Linux Kernel drivers/char/random.c comment documentation @ 1da177e4". Retrieved 2014-07-22.
- Lloyd, Jack (2008-12-09). "On Syllable's /dev/random". Archived from the original on 2009-04-29. Retrieved 2013-07-03.
- "/dev/random". Everything2. 2003-06-08. Archived from the original on 2009-04-29. Retrieved 2013-07-03.
- Linux Programmer's Manual – Special Files –
- "/dev/random and /dev/urandom implementation in Linux 1.3.39, function
random_read_unlimited". 1995-11-04. Retrieved 2013-11-21.
- Gutterman, Zvi; Pinkas, Benny; Reinman, Tzachy (2006-03-06). "Analysis of the Linux Random Number Generator" (PDF). Archived (PDF) from the original on 2008-10-03. Retrieved 2013-07-03.
- "Cryptography Users Guide". Texas Instruments. 2013-06-04. Retrieved 2013-07-03.
- man-pages/random.4 at revision 9dc53e71c24ab77d682dffbd204f94211161905c, line 68
- "Manual Pages: arc4random(3)". 2013-10-22. Retrieved 2014-06-23.
- deraadt, ed. (2014-07-21). "libc/crypt/arc4random.c". BSD Cross Reference, OpenBSD src/lib/. Retrieved 2015-01-13.
ChaCha based random number generator for OpenBSD.
- riastradh, ed. (2014-11-16). "libc/gen/arc4random.c". BSD Cross Reference, NetBSD src/lib/. Retrieved 2015-01-13.
Legacy arc4random(3) API from OpenBSD reimplemented using the ChaCha20 PRF, with per-thread state.
- Darwin and OS X Kernel Interfaces Manual –
- Moffat, Darren (2013-09-12). "Solaris Random Number Generation". Retrieved 2016-05-12.
- NetBSD Kernel Interfaces Manual –
- "random(4)". 1999-09-19. Retrieved 2013-07-03.
- "random and urandom Devices". pSeries and AIX Information Center. 2010-03-15. Archived from the original on 2013-07-11. Retrieved 2013-07-03.
- "HP-UX Strong Random Number Generator". 2004-07-23. Retrieved 2013-07-03.
- Roberts, Iain (2003-04-25). "AIX 5.2 /dev/random and /dev/urandom devices". lists.gnupg.org. Archived from the original on 2012-02-22. Retrieved 2013-07-03.
- "Doug Kaufman's Web Site - DOS ports". 2006-11-02. Retrieved 2013-07-03.
- Warner, Brian (2002-07-25). "EGD: The Entropy Gathering Daemon". Retrieved 2013-07-03.
- Jänicke, Lutz (2007-01-12). "PRNGD - Pseudo Random Number Generator Daemon". Retrieved 2013-07-03.
- Biege, Thomas (2006-11-06). "Analysis of a strong Pseudo Random Number Generator by anatomizing Linux' Random Number Device" (PDF).
- Hühn, Thomas (2014). "Myths about /dev/urandom".