The old code didn't produce very good random numbers. It had a bias.
And that was from performing some unnecessary floating-point
calculations that overcomplicate the problem.
Cc: "Jason A. Donenfeld" <Jason@zx2c4.com>
Cc: Cristian Rodríguez <crrodriguez@opensuse.org>
Cc: Adhemerval Zanella <adhemerval.zanella@linaro.org>
Cc: Björn Esser <besser82@fedoraproject.org>
Cc: Yann Droneaud <ydroneaud@opteya.com>
Cc: Joseph Myers <joseph@codesourcery.com>
Cc: Sam James <sam@gentoo.org>
Signed-off-by: Alejandro Colomar <alx@kernel.org>
A set of APIs similar to arc4random(3) is complex enough to deserve its
own file.
Cc: "Jason A. Donenfeld" <Jason@zx2c4.com>
Cc: Cristian Rodríguez <crrodriguez@opensuse.org>
Cc: Adhemerval Zanella <adhemerval.zanella@linaro.org>
Cc: Björn Esser <besser82@fedoraproject.org>
Cc: Yann Droneaud <ydroneaud@opteya.com>
Cc: Joseph Myers <joseph@codesourcery.com>
Signed-off-by: Alejandro Colomar <alx@kernel.org>
arc4random(3) returns a number.
arc4random_buf(3) fills a buffer.
arc4random_uniform(3) returns a number less than a bound.
and I'd add a hypothetical one which we use:
*_interval() should return a number within the interval [min, max].
In reality, the function being called csrand() in this patch is not
really cryptographically secure, since it had a bias, but a subsequent
patch will fix that.
Signed-off-by: Alejandro Colomar <alx@kernel.org>
We were always casting the result to u_long. Better just use that type
in the function. Since we're returning u_long, it makes sense to also
specify the input as u_long. In fact, that'll help for doing bitwise
operations inside this function.
Signed-off-by: Alejandro Colomar <alx@kernel.org>
I have plans to split this function in smaller functions that implement
bits of this functionallity, to simplify the implementation. So, let's
use names that distinguish them.
This one produces a number within an interval, so make that clear. Also
make clear that the function produces cryptographically-secure numbers.
Signed-off-by: Alejandro Colomar <alx@kernel.org>
arc4random(3) without kernel support is unsafe, as it can't know when to
drop the buffer. Since we depend on libbsd since recently, we have
arc4random(3) functions always available, and thus, this code would have
always called arc4random_buf(3bsd), which is unsafe. Put it after some
better alternatives, at least until in a decade or so all systems have a
recent enough glibc.
glibc implements arc4random(3) safely, since it's just a wrapper around
getrandom(2).
Link: <https://inbox.sourceware.org/libc-alpha/20220722122137.3270666-1-adhemerval.zanella@linaro.org/>
Link: <https://inbox.sourceware.org/libc-alpha/5c29df04-6283-9eee-6648-215b52cfa26b@cs.ucla.edu/T/>
Cc: Cristian Rodríguez <crrodriguez@opensuse.org>
Cc: Adhemerval Zanella <adhemerval.zanella@linaro.org>
Cc: Guillem Jover <guillem@hadrons.org>
Cc: Björn Esser <besser82@fedoraproject.org>
Reviewed-by: "Jason A. Donenfeld" <Jason@zx2c4.com>
Signed-off-by: Alejandro Colomar <alx@kernel.org>
salt.c:102:22: warning: type qualifiers ignored on function return type [-Wignored-qualifiers]
102 | static /*@observer@*/const unsigned long SHA_get_salt_rounds (/*@null@*/int *prefered_rounds);
| ^~~~~
salt.c:110:22: warning: type qualifiers ignored on function return type [-Wignored-qualifiers]
110 | static /*@observer@*/const unsigned long YESCRYPT_get_salt_cost (/*@null@*/int *prefered_cost);
| ^~~~~
subordinateio.c:160:8: warning: type qualifiers ignored on function return type [-Wignored-qualifiers]
160 | static const bool range_exists(struct commonio_db *db, const char *owner)
| ^~~~~
On systems with Linux kernel < 3.17, getentropy() and getrandom() may
exist but return ENOSYS. Use /dev/urandom as a fallback to avoid a hard
requirement on Linux kernel version.
Fixes#512.
Signed-off-by: Xi Ruoyao <xry111@xry111.site>
In order to remove some of the FIXMEs it was necessary to change the
code and call getulong() instead of getlong().
Signed-off-by: Iker Pedrosa <ipedrosa@redhat.com>
If SHA_CRYPT_MIN_ROUNDS and SHA_CRYPT_MAX_ROUNDS are both unspecified,
use SHA_ROUNDS_DEFAULT.
Previously, the code fell through, calling shadow_random(-1, -1). This
ultimately set rounds = (unsigned long) -1, which ends up being a very
large number! This then got capped to SHA_ROUNDS_MAX later in the
function.
The new behavior matches BCRYPT_get_salt_rounds().
Bug: https://bugs.gentoo.org/808195
Fixes: https://github.com/shadow-maint/shadow/issues/393
Most Linux distributions, including Fedora and RHEL 8, are shipping
with libxcrypt >= 4.0.
Since that version of libxcrypt the provided family of crypt_gensalt()
functions are able to use automatic entropy drawn from secure system
ressources, like arc4random(), getentropy() or getrandom().
Anyways, the settings generated by crypt_gensalt() are always
guaranteed to works with the crypt() function.
Using crypt_gensalt() is also needed to make proper use of newer
hashing methods, like yescrypt, provided by libxcrypt.
Signed-off-by: Björn Esser <besser82@fedoraproject.org>
In a previous commit we introduced /dev/urandom as a source to obtain
random bytes from. This may not be available on all systems, or when
operating inside of a chroot.
Almost all systems provide functions to obtain random bytes from
secure system ressources. Thus we should prefer to use these, and
fall back to /dev/urandom, if there is no such function present, as
a last resort.
Signed-off-by: Björn Esser <besser82@fedoraproject.org>
Using the random() function to obtain pseudo-random bytes
for generating salt strings is considered to be dangerous.
See CWE-327.
We really should use a more reliable source for obtaining
pseudo-random bytes like /dev/urandom.
Fixes#376.
Signed-off-by: Björn Esser <besser82@fedoraproject.org>
In the previous commit we refactored the functions converting the
rounds number into a string for use with the crypt() function, to
not require any static buffer anymore.
Add some clarifying comments about how the minimum required buffer
length is computed inside of these functions.
Signed-off-by: Björn Esser <besser82@fedoraproject.org>
* Move all pre-processor defines to the top of the file.
* Unify the gensalt() function to be useable for all supported
hash methods.
* Drop the gensalt_{b,yes}crypt() functions in favor of the
previous change.
* Refactor the functions converting the rounds number into
a string for use with the crypt() function, to not require
any static buffer anymore.
* Clarify the comment about how crypt_make_salt() chooses the used
hash method from the settings in the login.defs file.
* Use memset() to fill static buffers with zero before using them.
* Use a fixed amount of 16 random base64-chars for the
sha{256,512}crypt hash methods, which is effectively still less
than the recommendation from NIST (>= 128 bits), but the maximum
those methods can effectively use (approx. 90 bits).
* Rename ROUNDS_{MIN,MAX} to SHA_ROUNDS_{MIN,MAX}.
* Bugfixes in the logic of setting rounds in BCRYPT_salt_rounds().
* Likewise for YESCRYPT_salt_cost().
* Fix formatting and white-space errors.
Signed-off-by: Björn Esser <besser82@fedoraproject.org>
The corresponding functions for the other hash methods all take
a pointer to an integer value as the only paramater, so this
particular function should do so as well.
Signed-off-by: Björn Esser <besser82@fedoraproject.org>
Closes#325
Add a new subid_init() function which can be used to specify the
stream on which error messages should be printed. (If you want to
get fancy you can redirect that to memory :) If subid_init() is
not called, use stderr. If NULL is passed, then /dev/null will
be used.
This patch also fixes up the 'Prog', which previously had to be
defined by any program linking against libsubid. Now, by default
in libsubid it will show (subid). Once subid_init() is called,
it will use the first variable passed to subid_init().
Signed-off-by: Serge Hallyn <serge@hallyn.com>
* libmisc/salt.c (shadow_random): Use long instead of size_t.
Compatibility with size_t is easier to check since it's used for
smaller numbers (salt size).
* libmisc/salt.c: random() max value is 2^31-1 (same as RAND_MAX
on GNU). As it is not clear whether on some systems the max value
can exceed this number and whether some systems have max values
which would be lower, we take this into account when defining the
salt size and number of rounds for SHA encrypted passwords. Higher
values are favored.
boolean.
* libmisc/basename.c: Avoid implicit conversion of pointer to
boolean.
* libmisc/basename.c, lib/prototypes.h (Basename): Return a
constant string.
* libmisc/basename.c, libmisc/obscure.c, lib/prototypes.h,
libmisc/xmalloc.c, libmisc/getdate.h, libmisc/system.c,
libmisc/getgr_nam_gid.c, libmisc/failure.c, libmisc/valid.c: Add
splint annotations.
* libmisc/chowndir.c: Avoid memory leak.
* libmisc/chowndir.c: Do not check *printf/*puts return value.
* libmisc/chowntty.c: Avoid implicit conversion between integer
types.
* libmisc/obscure.c: Return a bool when possible instead of int.
* libmisc/shell.c: Do not check *printf/*puts return value.
* libmisc/shell.c: Do not check execle return value.
* libmisc/setupenv.c: Avoid implicit conversion between integer
types.
* libmisc/xmalloc.c: size should not be zero to avoid returning
NULL pointers.
* libmisc/hushed.c: Do not check *printf/*puts return value.
* libmisc/system.c: Avoid implicit conversion of integer to
boolean. safe_system last argument is a boolean.
* libmisc/system.c: Check return value of dup2.
* libmisc/system.c: Do not check *printf/*puts return value.
* libmisc/system.c: Do not check execve return value.
* libmisc/salt.c: Do not check *printf/*puts return value.
* libmisc/loginprompt.c: Do not check gethostname return value.
* libmisc/find_new_gid.c, libmisc/find_new_uid.c: Do not check
gr_rewind/pw_rewind return value.
* libmisc/ttytype.c: Limit the number of parsed characters in the
sscanf format.
* libmisc/ttytype.c: Test if a type was really read.
* libmisc/sub.c: Do not check *printf/*puts return value.
* libmisc/sub.c: Avoid implicit conversion of integer to boolean.
* src/userdel.c: Fix typo in comment.
* src/userdel.c: Avoid implicit conversion of boolean to integer.
* src/userdel.c: safe_system last argument is a boolean.
* src/newusers.c: Avoid implicit conversion of boolean to integer.
* src/newusers.c: Avoid implicit conversion of integer to boolean.
* src/usermod.c: Add brackets.
* src/usermod.c: Avoid implicit conversion of characters or
integers to booleans.
* src/vipw.c: Avoid implicit conversion of integer to boolean.
* src/su.c: Avoid implicit conversion of integer to boolean.
* src/su.c: Add brackets.
* src/useradd.c: Avoid implicit conversion of characters or
integers to booleans.
* libmisc/audit_help.c: Include prototypes.h to get the prototype
of audit_help_open.
* libmisc/salt.c: Use booleans instead of negating integers.
* src/passwd.c: Declare the check_selinux_access prototype and
avoid name clashes (change_user -> changed_user; change_uid ->
changed_uid; access -> requested_access)
- generation of SHA encrypted passwords (chpasswd, gpasswd, newusers,
chgpasswd; and also passwd if configured without PAM support).
The number of rounds and number of salt bytes was fixed to their lower
allowed values (resp. configurable and 8), hence voiding some of the
advantages of this encryption method. Dictionary attacks with
precomputed tables were easier than expected, but still harder than with
the MD5 (or DES) methods.
* NEWS, libmisc/salt.c (SHA_salt_size): Seed the RNG, and fix a
overflow. These caused the SHA salt size to always be 8 bytes,
instead of being in the 8-16 range. Thanks to Peter Vrabec
pvrabec@redhat.com for noticing.
* NEWS, libmisc/salt.c (SHA_salt_rounds): Seed the RNG with
seedRNG instead of srand, and fix the same overflow. This caused
the number of rounds to always be the smallest one.
Files with no license use the default 3-clauses BSD license. The copyright
were mostly not recorded; they were updated according to the Changelog.
"Julianne Frances Haugh and contributors" changed to "copyright holders
and contributors".
microseconds to avoid having the same salt for different passwords
generated in the same second. This permits to avoid using the same salt
for different passwords in newusers.
Daubert for the patch.
* libmisc/salt.c: Include <stdio.h>, needed for stderr and printf
functions.
* lib/encrypt.c: Include <stdio.h>, needed for perror, stderr and
printf functions
* src/usermod.c: sgr_locked exists only if SHADOWGRP is defined.
* src/chgpasswd.c: Only check is the gshadow file exists if
SHADOWGRP is defined.
SHA_salt_size(), and SHA_salt_rounds().
* libmisc/salt.c: l64a() and gensalt() are static.
* libmisc/salt.c: The `meth' parameter of crypt_make_salt() is a
const. (ditto for the method variable).
* libmisc/salt.c: SHA_salt_rounds returns a const string.
* libmisc/salt.c: Avoid warnings with cast of random() to double.
* libmisc/salt.c: Replace rand() by random().
default. Keeping the feature enabled is safe. Disabling it permits
to disable the references to the SHA256 and SHA512 password
encryption algorithms from the usage help and manuals (in addition
to the support for these algorithms in the code).
* libmisc/obscure.c, libmisc/salt.c, src/newusers.c,
src/chpasswd.c, src/chgpasswd.c, src/passwd.c: ENCRYPT_METHOD is
always supported in login.defs. Remove the ENCRYPTMETHOD_SELECT
preprocessor condition.
* libmisc/obscure.c, libmisc/salt.c, src/newusers.c,
src/chpasswd.c, src/chgpasswd.c, src/passwd.c: Disable SHA256 and
SHA512 if USE_SHA_CRYPT is not defined (this corresponds to a
subset of the ENCRYPTMETHOD_SELECT sections).
SHA256, and SHA512 exactly (not only the first 3/6 chars).
* libmisc/salt.c (SHA_salt_rounds): Set rounds to the specified
prefered_rounds value, if specified.
* src/gpasswd.c, libmisc/salt.c: Fix compilation warnings (use
size_t for lengths).
* src/chpasswd.c, src/chgpasswd.c: Add missing parenthesis.
