Commit Graph

24 Commits

Author SHA1 Message Date
Dr. Matthias St. Pierre
849529257c drbg: ensure fork-safety without using a pthread_atfork handler
When the new OpenSSL CSPRNG was introduced in version 1.1.1,
it was announced in the release notes that it would be fork-safe,
which the old CSPRNG hadn't been.

The fork-safety was implemented using a fork count, which was
incremented by a pthread_atfork handler. Initially, this handler
was enabled by default. Unfortunately, the default behaviour
had to be changed for other reasons in commit b5319bdbd0, so
the new OpenSSL CSPRNG failed to keep its promise.

This commit restores the fork-safety using a different approach.
It replaces the fork count by a fork id, which coincides with
the process id on UNIX-like operating systems and is zero on other
operating systems. It is used to detect when an automatic reseed
after a fork is necessary.

To prevent a future regression, it also adds a test to verify that
the child reseeds after fork.

CVE-2019-1549

Reviewed-by: Paul Dale <paul.dale@oracle.com>
Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/9832)
2019-09-11 11:22:18 +02:00
Vladimir Kotal
d6dda392c1 use native atomic increment function on Solaris
Reviewed-by: Paul Dale <paul.dale@oracle.com>
Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/9215)
2019-08-09 13:16:41 +01:00
Matt Caswell
3593266d1c Make core code available within the FIPS module
Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/8728)
2019-05-23 11:02:04 +01:00
Richard Levitte
0e9725bcb9 Following the license change, modify the boilerplates in crypto/
[skip ci]

Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/7827)
2018-12-06 15:32:17 +01:00
Andy Polyakov
d2b863643d crypto/threads_*: remove CRYPTO_atomic_{read|write}.
CRYPTO_atomic_read was added with intention to read statistics counters,
but readings are effectively indistinguishable from regular load (even
in non-lock-free case). This is because you can get out-dated value in
both cases. CRYPTO_atomic_write was added for symmetry and was never used.

Reviewed-by: Kurt Roeckx <kurt@roeckx.be>
(Merged from https://github.com/openssl/openssl/pull/6883)
2018-08-17 12:40:39 +02:00
Richard Levitte
28428130db Update copyright year
Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/5990)
2018-04-17 15:18:40 +02:00
Rich Salz
7de2b9c4af Set error code if alloc returns NULL
Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/5886)
2018-04-05 15:13:55 -04:00
Pauli
f49452c297 Return a value from atomic read on Windows.
Use a read lock when reading using pthreads.

Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/4517)
2017-10-11 09:47:54 +10:00
Pauli
30ff41beab Add atomic write call
Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/4414)
2017-10-10 08:45:53 +10:00
Pauli
94683b7acb Add a CRYPTO_atomic_read call which allows an int variable to be read
in an atomic fashion.

Reviewed-by: Rich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/4414)
2017-10-10 08:45:52 +10:00
Rich Salz
b842fcbb37 Put thread-fork-init inside a run-once guard
Thanks to Christian Heimes for pointing this out.

Reviewed-by: Andy Polyakov <appro@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/4191)
2017-08-18 11:48:35 -04:00
Rich Salz
5f8dd0f849 Add missing include of cryptlib.h
Also use "" not <> for all include cryptlib

Reviewed-by: Andy Polyakov <appro@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/4082)
2017-08-06 17:06:19 -04:00
Rich Salz
2915fe19a6 Add fork handlers, based on pthread_atfork
Only for Unix platforms

Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3754)
2017-06-29 16:19:41 -04:00
Dmitry Kostjuchenko
bc1dba2095 Grouped data declarations [skip ci]
Reviewed-by: Rich Salz <rsalz@openssl.org>
Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/1981)
2017-02-03 13:48:44 +01:00
Dmitry Kostjuchenko
5d5eed4456 Removed tab spaces.
Reviewed-by: Rich Salz <rsalz@openssl.org>
Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/1981)
2017-02-03 13:48:44 +01:00
Dmitry Kostjuchenko
ec93a2921f Corrections according the review comments.
Updated indentations according project rules, renamed file-local define to the shorter version - USE_RWLOCK, fixed declaration after the if statement in CRYPTO_THREAD_lock_new().
Reviewed-by: Rich Salz <rsalz@openssl.org>
Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/1981)
2017-02-03 13:48:44 +01:00
Dmitry Kostjuchenko
2accf3f7e0 Compile fix on platforms with missing pthread_rwlock_t.
Fix compilation on platforms with missing pthread_rwlock_t implementation by replacing it with pthread_mutex_t. An example of such platform can be Android OS 2.0 - 2.1, API level 5 (Eclair), Android NDK platform - android-5 where pthread_rwlock_t is not implemented and is missing in pthread.h.

In case of missing pthread_rwlock_t implementation CRYPTO_RWLOCK will work as exclusive lock in write-only mode of pthread_rwlock_t lock.

The implementation based on pthread_mutex_t must be using PTHREAD_MUTEX_RECURSIVE mode to be compatible with recursive behavior of pthread_rwlock_rdlock.
Reviewed-by: Rich Salz <rsalz@openssl.org>
Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/1981)
2017-02-03 13:48:44 +01:00
Richard Levitte
1beca67688 CRYPTO_atomic_add(): check that the object is lock free
If not, fall back to our own code, using the given mutex

Reviewed-by: Andy Polyakov <appro@openssl.org>
2016-08-24 14:37:48 +01:00
Richard Levitte
11fc6c7611 CRYPTO_atomic_add(): use acquire release memory order rather than relaxed
For increments, the relaxed model is fine.  For decrements, it's
recommended to use the acquire release model.  We therefore go for the
latter.

Reviewed-by: Andy Polyakov <appro@openssl.org>
2016-08-24 14:37:40 +01:00
Richard Levitte
cb4b54c23b Check for __GNUC__ to use GNU C atomic buildins
Note: we trust any other compiler that fully implements GNU extension
to define __GNUC__

RT#4642

Reviewed-by: Andy Polyakov <appro@openssl.org>
2016-08-24 14:37:30 +01:00
Rich Salz
b1322259d9 Copyright consolidation 09/10
Reviewed-by: Richard Levitte <levitte@openssl.org>
2016-05-17 14:53:16 -04:00
Viktor Dukhovni
5c4328f04f Fold threads.h into crypto.h making API public
Document thread-safe lock creation

Reviewed-by: Richard Levitte <levitte@openssl.org>
2016-05-16 12:16:26 -04:00
FdaSilvaYY
0b2fc9286f GH773: Possible leak on CRYPTO_THREAD_lock_new failure
Signed-off-by: Rich Salz <rsalz@akamai.com>
Reviewed-by: Matt Caswell <matt@openssl.org>
2016-03-03 19:47:01 -05:00
Alessandro Ghedini
71a04cfca0 Implement new multi-threading API
Reviewed-by: Rich Salz <rsalz@openssl.org>
Reviewed-by: Matt Caswell <matt@openssl.org>
2016-02-26 10:00:36 +00:00