The current retirement code for rcu qp's has a race condition,
which can cause use-after-free errors, but only if more than
3 QPs are allocated, which is not the default configuration.
This fixes an oversight in commit 5949918f9afa ("Rework and
simplify RCU code")
Reviewed-by: Neil Horman <nhorman@openssl.org>
Reviewed-by: Tomas Mraz <tomas@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/26952)
(cherry picked from commit 6e7be995fd7fd24d38b95982cf90f801ac045743)
This test was disabled due to "Stochastic failures in
the RCU test on MACOSX" by #23967, which sounds like an
issue that is probably fixed now.
Reviewed-by: Neil Horman <nhorman@openssl.org>
Reviewed-by: Paul Dale <ppzgs1@gmail.com>
Reviewed-by: Tomas Mraz <tomas@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/26834)
(cherry picked from commit 4d16d2f40a135830ee667e1588c75293a7b23480)
Reviewed-by: Neil Horman <nhorman@openssl.org>
Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/25664)
(cherry picked from commit 420d5d6294449527f4dd986b4fed8681bd4ae8fb)
Fix cases where `int` argument was passed instead of `size_t`.
CLA: trivial
Reviewed-by: Richard Levitte <levitte@openssl.org>
Reviewed-by: Matt Caswell <matt@openssl.org>
Reviewed-by: Tomas Mraz <tomas@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/25857)
(cherry picked from commit ccaa754b5f66cc50d8ecbac48b38268e2acd715e)
Reviewed-by: Paul Dale <ppzgs1@gmail.com>
Reviewed-by: Matt Caswell <matt@openssl.org>
Reviewed-by: Tomas Mraz <tomas@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/24630)
Currently, rcu has a global bit of data, the CRYPTO_THREAD_LOCAL object
to store per thread data. This works in some cases, but fails in FIPS,
becuase it contains its own copy of the global key.
So
1) Make the rcu_thr_key a per-context variable, and force
ossl_rcu_lock_new to be context aware
2) Store a pointer to the context in the lock object
3) Use the context to get the global thread key on read/write lock
4) Use ossl_thread_start_init to properly register a cleanup on thread
exit
5) Fix up missed calls to OSSL_thread_stop() in our tests
(cherry picked from commit 24d16d3a1915a06a2130385a87de9a37fc09c4b9)
Fixes#24581
Reviewed-by: Paul Dale <ppzgs1@gmail.com>
Reviewed-by: Matt Caswell <matt@openssl.org>
Reviewed-by: Tomas Mraz <tomas@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/24630)
Coverity caught the following issues:
1591477
1591475
1591473
1591470
all of which are simmilar, in that they catch potential divide by zero
in double values. It can't actually happen since the the threads which
increment these counters don't exit until they reach non-zero values,
but its easy to add the checks, so lets do that to ensure that we don't
change something in the future that causes it.
Reviewed-by: Tomas Mraz <tomas@openssl.org>
Reviewed-by: Paul Dale <pauli@openssl.org>
Reviewed-by: Paul Dale <ppzgs1@gmail.com>
(Merged from https://github.com/openssl/openssl/pull/24337)
Add checks for the return value of CRYPTO_THREAD_lock_new() in order to avoid Null pointer dereference.
Fixes: 5f8b812931 ("Add locking to atomic operations in rw/rcu tests")
Fixes: d0e1a0ae70 ("RCU lock implementation")
Fixes: 71a04cfca0 ("Implement new multi-threading API")
Signed-off-by: Jiasheng Jiang <jiasheng@purdue.edu>
Reviewed-by: Neil Horman <nhorman@openssl.org>
Reviewed-by: Tomas Mraz <tomas@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/24313)
(cherry picked from commit 327261c076b8468382e1effea14d79446cc22b4d)
I neglected to add locks to the calls to CRYPTO_atomic_add in these
test, which on newer compilers is fine, as atomic operations are
defined. However on older compilers the __ATOMIC_ACQ_REL definition is
missing causing these function to be implemented using an rwlock, which
when NULL causes the locks to fail.
Fix this my creating the lock and using them appropriately
Fixes#24000
Reviewed-by: Matt Caswell <matt@openssl.org>
Reviewed-by: Paul Dale <ppzgs1@gmail.com>
(Merged from https://github.com/openssl/openssl/pull/24001)
(cherry picked from commit 5f8b812931e5da24df08913c05ff8e4f4494f014)
Stochastic failures in the RCU test on MACOSX are occuring. Due to beta
release, disabling this test on MACOSX until post 3.3 release
Reviewed-by: Tomas Mraz <tomas@openssl.org>
Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/23967)
(cherry picked from commit 1967539e212c17139dc810096da987c8100b1ba2)
Its possible in some conditions for the rw/rcu torture tests to wrap the
counter, leading to false positive failures, make them 64 bits to avoid
this
Reviewed-by: Shane Lontis <shane.lontis@oracle.com>
Reviewed-by: Tomas Mraz <tomas@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/23724)
(cherry picked from commit b50c174ee3b11f916285046d52574ba653745083)
Introduce an RCU lock implementation as an alternative locking mechanism
to openssl. The api is documented in the ossl_rcu.pod
file
Read side implementaiton is comparable to that of RWLOCKS:
ossl_rcu_read_lock(lock);
<
critical section in which data can be accessed via
ossl_derefrence
>
ossl_rcu_read_unlock(lock);
Write side implementation is:
ossl_rcu_write_lock(lock);
<
critical section in which data can be updated via
ossl_assign_pointer
and stale data can optionally be scheduled for removal
via ossl_rcu_call
>
ossl_rcu_write_unlock(lock);
...
ossl_synchronize_rcu(lock);
ossl_rcu_call fixup
Reviewed-by: Hugo Landau <hlandau@openssl.org>
Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/22729)
The new decoder optimisation has some locking involved so we confirm that
reading a PEM file from multiple threads works as expected.
Reviewed-by: Tomas Mraz <tomas@openssl.org>
Reviewed-by: Paul Dale <pauli@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/21426)
Signed-off-by: Čestmír Kalina <ckalina@redhat.com>
Reviewed-by: Tomas Mraz <tomas@openssl.org>
Reviewed-by: Paul Dale <pauli@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/19473)
While POSIX threads are cancellable and may be asynchronously cancelled,
their cancellation is not guaranteed by the POSIX standard.
test_thread_noreturn, which simulates a long-running possibly
unresponsive thread:
THREAD #1 THREAD #2
LOCK L1
SPAWN #2
LOCK L1
On MacOS, cancelling such thread only queues cancellation request, but
the following pthread_join hangs.
Replace this implementation by an unbounded sequence of sleeps instead.
Signed-off-by: Čestmír Kalina <ckalina@redhat.com>
Reviewed-by: Hugo Landau <hlandau@openssl.org>
Reviewed-by: Tomas Mraz <tomas@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/19433)
Multiple concurrent joins with a running thread suffer from a race
condition that allows concurrent join calls to perform concurrent arch
specific join calls, which is UB on POSIX, or to concurrently execute
join and terminate calls.
As soon as a thread T1 exists, one of the threads that joins with T1
is selected to perform the join, the remaining ones await completion.
Once completed, the remaining calls immediately return. If the join
failed, another thread is selected to attempt the join operation.
Forcefully terminating a thread that is in the process of joining
another thread is not supported.
Common code from thread_posix and thread_win was refactored to use
common wrapper that handles synchronization.
Signed-off-by: Čestmír Kalina <ckalina@redhat.com>
Reviewed-by: Hugo Landau <hlandau@openssl.org>
Reviewed-by: Tomas Mraz <tomas@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/19433)
Some primitives are designed to be used in a multi-threaded environment,
if supported, e.g., Argon2.
This patch adds support for preemptive threading and basic synchronization
primitives for platforms compliant with POSIX threads or Windows CRT.
Native functions are wrapped to provide a common (internal) API.
Threading support can be disabled at compile time. If enabled, threading
is disabled by default and needs to be explicitly enabled by the user.
Thread enablement requires an explicit limit on the number of threads that
OpenSSL may spawn (non-negative integer/infinity). The limit may be changed.
Signed-off-by: Čestmír Kalina <ckalina@redhat.com>
Reviewed-by: Hugo Landau <hlandau@openssl.org>
Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/12255)
Fixes#18226.
Reviewed-by: Matt Caswell <matt@openssl.org>
Reviewed-by: Paul Dale <pauli@openssl.org>
Reviewed-by: Tomas Mraz <tomas@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/18331)
Not all platforms support tsan operations, those that don't need to have an
alternative locking path.
Fixes#17447
Reviewed-by: Bernd Edlinger <bernd.edlinger@hotmail.de>
(Merged from https://github.com/openssl/openssl/pull/17479)
Run more threads and load the legacy provider (which uses a child lib ctx)
in order to hit more possible thread failures.
Reviewed-by: Tomas Mraz <tomas@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/16980)
Reviewed-by: Tomas Mraz <tomas@openssl.org>
Reviewed-by: Richard Levitte <levitte@openssl.org>
Reviewed-by: Paul Dale <pauli@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/15708)
The FIPS provider leaks a RAND if the POST is run at initialisation time.
This test case reliably reproduces this event.
Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/15278)
Force the thread test to use the configuration file via a command line arg.
Use the test library support for libctx creation.
Fixes#15243
Reviewed-by: Shane Lontis <shane.lontis@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/15256)
Add EVP_PKEY_gen(), EVP_PKEY_Q_gen(), EVP_RSA_gen(), and EVP_EC_gen().
Also export auxiliary function OSSL_EC_curve_nid2name()
and improve deprecation info on RSA and EC key generation/management functions.
Reviewed-by: Shane Lontis <shane.lontis@oracle.com>
(Merged from https://github.com/openssl/openssl/pull/14695)
Check that we don't see any threading issues when loading/unloading a
provider from multiple threads.
Reviewed-by: Paul Dale <pauli@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/15010)
Most of these were already deprecated but a few have been missed. This
commit corrects that.
Fixes#14303Fixes#14317
Reviewed-by: Richard Levitte <levitte@openssl.org>
Reviewed-by: Shane Lontis <shane.lontis@oracle.com>
Reviewed-by: Paul Dale <pauli@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/14319)
Reviewed-by: Richard Levitte <levitte@openssl.org>
Reviewed-by: Shane Lontis <shane.lontis@oracle.com>
Reviewed-by: Paul Dale <pauli@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/14319)
If we don't synchronize properly in the core provider code, and build
with a thread sanitizer, this should cause a crash.
Reviewed-by: Paul Dale <pauli@openssl.org>
Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/14354)
EVP_PKEYs may be shared across mutliple threads. For example this is
common for users of libssl who provide a single EVP_PKEY private key for
an SSL_CTX, which is then shared between multiple threads for each SSL
object.
Reviewed-by: Paul Dale <pauli@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/13987)
Issue #13682 suggests that doing a simple fetch from multi-threads may
result in issues so we add a test for that.
Reviewed-by: Tomas Mraz <tmraz@fedoraproject.org>
(Merged from https://github.com/openssl/openssl/pull/13660)
that needed test_main now works using the same infrastructure as tests that used
register_tests.
This meant:
* renaming register_tests to setup_tests and giving it a success/failure return.
* renaming the init_test function to setup_test_framework.
* renaming the finish_test function to pulldown_test_framework.
* adding a user provided global_init function that runs before the test frame
work is initialised. It returns a failure indication that stops the stest.
* adding helper functions that permit tests to access their command line args.
* spliting the BIO initialisation and finalisation out from the test setup and
teardown.
* hiding some of the now test internal functions.
* fix the comments in testutil.h
Reviewed-by: Richard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/3953)
