crypto/threads_pthread.c

/*
 * Copyright 2016-2025 The OpenSSL Project Authors. All Rights Reserved.
 *
 * Licensed under the Apache License 2.0 (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
 */

/* We need to use the OPENSSL_fork_*() deprecated APIs */
#define OPENSSL_SUPPRESS_DEPRECATED

#if !defined(__GNUC__) || !defined(__ATOMIC_ACQ_REL) || defined(BROKEN_CLANG_ATOMICS) || defined(OPENSSL_NO_STDIO)
/*
 * we only enable REPORT_RWLOCK_CONTENTION on clang/gcc when we have
 * atomics available.  We do this because we need to use an atomic to track
 * when we can close the log file.  We could use the CRYPTO_atomic_ api
 * but that requires lock creation which gets us into a bad recursive loop
 * when we try to initialize the file pointer
 */
#ifdef REPORT_RWLOCK_CONTENTION
#warning "RWLOCK CONTENTION REPORTING NOT SUPPORTED, Disabling"
#undef REPORT_RWLOCK_CONTENTION
#endif
#endif

#ifdef REPORT_RWLOCK_CONTENTION
#define _GNU_SOURCE
#include <execinfo.h>
#include <unistd.h>
#endif

#include <openssl/crypto.h>
#include <crypto/cryptlib.h>
#include <crypto/sparse_array.h>
#include "internal/cryptlib.h"
#include "internal/threads_common.h"
#include "internal/rcu.h"
#ifdef REPORT_RWLOCK_CONTENTION
#include <fcntl.h>
#include <stdbool.h>
#include <sys/syscall.h>
#include <sys/uio.h>
#include "internal/time.h"
#endif
#include "rcu_internal.h"

#if defined(__clang__) && defined(__has_feature)
#if __has_feature(thread_sanitizer)
#define __SANITIZE_THREAD__
#endif
#endif

#if defined(__SANITIZE_THREAD__)
#include <sanitizer/tsan_interface.h>
#define TSAN_FAKE_UNLOCK(x)          \
    __tsan_mutex_pre_unlock((x), 0); \
    __tsan_mutex_post_unlock((x), 0)

#define TSAN_FAKE_LOCK(x)          \
    __tsan_mutex_pre_lock((x), 0); \
    __tsan_mutex_post_lock((x), 0, 0)
#else
#define TSAN_FAKE_UNLOCK(x)
#define TSAN_FAKE_LOCK(x)
#endif

#if defined(__sun)
#include <atomic.h>
#endif

#if defined(__apple_build_version__) && __apple_build_version__ < 6000000
/*
 * OS/X 10.7 and 10.8 had a weird version of clang which has __ATOMIC_ACQUIRE and
 * __ATOMIC_ACQ_REL but which expects only one parameter for __atomic_is_lock_free()
 * rather than two which has signature __atomic_is_lock_free(sizeof(_Atomic(T))).
 * All of this makes impossible to use __atomic_is_lock_free here.
 *
 * See: https://github.com/llvm/llvm-project/commit/a4c2602b714e6c6edb98164550a5ae829b2de760
 */
#define BROKEN_CLANG_ATOMICS
#endif

#if defined(OPENSSL_THREADS) && !defined(CRYPTO_TDEBUG) && !defined(OPENSSL_SYS_WINDOWS)

#if defined(OPENSSL_SYS_UNIX)
#include <sys/types.h>
#include <unistd.h>
#endif

#include <assert.h>

/*
 * The Non-Stop KLT thread model currently seems broken in its rwlock
 * implementation
 * Likewise is there a problem with the glibc implementation on riscv.
 */
#if defined(PTHREAD_RWLOCK_INITIALIZER) && !defined(_KLT_MODEL_) \
    && !defined(__riscv)
#define USE_RWLOCK
#endif

/*
 * For all GNU/clang atomic builtins, we also need fallbacks, to cover all
 * other compilers.

 * Unfortunately, we can't do that with some "generic type", because there's no
 * guarantee that the chosen generic type is large enough to cover all cases.
 * Therefore, we implement fallbacks for each applicable type, with composed
 * names that include the type they handle.
 *
 * (an anecdote: we previously tried to use |void *| as the generic type, with
 * the thought that the pointer itself is the largest type.  However, this is
 * not true on 32-bit pointer platforms, as a |uint64_t| is twice as large)
 *
 * All applicable ATOMIC_ macros take the intended type as first parameter, so
 * they can map to the correct fallback function.  In the GNU/clang case, that
 * parameter is simply ignored.
 */

/*
 * Internal types used with the ATOMIC_ macros, to make it possible to compose
 * fallback function names.
 */
typedef void *pvoid;

#if defined(__GNUC__) && defined(__ATOMIC_ACQUIRE) && !defined(BROKEN_CLANG_ATOMICS) \
    && !defined(USE_ATOMIC_FALLBACKS)
#define ATOMIC_LOAD_N(t, p, o) __atomic_load_n(p, o)
#define ATOMIC_STORE_N(t, p, v, o) __atomic_store_n(p, v, o)
#define ATOMIC_STORE(t, p, v, o) __atomic_store(p, v, o)
#define ATOMIC_ADD_FETCH(p, v, o) __atomic_add_fetch(p, v, o)
#define ATOMIC_SUB_FETCH(p, v, o) __atomic_sub_fetch(p, v, o)
#else
static pthread_mutex_t atomic_sim_lock = PTHREAD_MUTEX_INITIALIZER;

#define IMPL_fallback_atomic_load_n(t)                    \
    static ossl_inline t fallback_atomic_load_n_##t(t *p) \
    {                                                     \
        t ret;                                            \
                                                          \
        pthread_mutex_lock(&atomic_sim_lock);             \
        ret = *p;                                         \
        pthread_mutex_unlock(&atomic_sim_lock);           \
        return ret;                                       \
    }
IMPL_fallback_atomic_load_n(uint32_t)
    IMPL_fallback_atomic_load_n(uint64_t)
        IMPL_fallback_atomic_load_n(pvoid)

#define ATOMIC_LOAD_N(t, p, o) fallback_atomic_load_n_##t(p)

#define IMPL_fallback_atomic_store_n(t)                         \
    static ossl_inline t fallback_atomic_store_n_##t(t *p, t v) \
    {                                                           \
        t ret;                                                  \
                                                                \
        pthread_mutex_lock(&atomic_sim_lock);                   \
        ret = *p;                                               \
        *p = v;                                                 \
        pthread_mutex_unlock(&atomic_sim_lock);                 \
        return ret;                                             \
    }
            IMPL_fallback_atomic_store_n(uint32_t)

#define ATOMIC_STORE_N(t, p, v, o) fallback_atomic_store_n_##t(p, v)

#define IMPL_fallback_atomic_store(t)                             \
    static ossl_inline void fallback_atomic_store_##t(t *p, t *v) \
    {                                                             \
        pthread_mutex_lock(&atomic_sim_lock);                     \
        *p = *v;                                                  \
        pthread_mutex_unlock(&atomic_sim_lock);                   \
    }
                IMPL_fallback_atomic_store(pvoid)

#define ATOMIC_STORE(t, p, v, o) fallback_atomic_store_##t(p, v)

    /*
     * The fallbacks that follow don't need any per type implementation, as
     * they are designed for uint64_t only.  If there comes a time when multiple
     * types need to be covered, it's relatively easy to refactor them the same
     * way as the fallbacks above.
     */

    static ossl_inline uint64_t fallback_atomic_add_fetch(uint64_t *p, uint64_t v)
{
    uint64_t ret;

    pthread_mutex_lock(&atomic_sim_lock);
    *p += v;
    ret = *p;
    pthread_mutex_unlock(&atomic_sim_lock);
    return ret;
}

#define ATOMIC_ADD_FETCH(p, v, o) fallback_atomic_add_fetch(p, v)

static ossl_inline uint64_t fallback_atomic_sub_fetch(uint64_t *p, uint64_t v)
{
    uint64_t ret;

    pthread_mutex_lock(&atomic_sim_lock);
    *p -= v;
    ret = *p;
    pthread_mutex_unlock(&atomic_sim_lock);
    return ret;
}

#define ATOMIC_SUB_FETCH(p, v, o) fallback_atomic_sub_fetch(p, v)
#endif

/*
 * This is the core of an rcu lock. It tracks the readers and writers for the
 * current quiescence point for a given lock. Users is the 64 bit value that
 * stores the READERS/ID as defined above
 *
 */
struct rcu_qp {
    uint64_t users;
};

struct thread_qp {
    struct rcu_qp *qp;
    unsigned int depth;
    CRYPTO_RCU_LOCK *lock;
};

#define MAX_QPS 10
/*
 * This is the per thread tracking data
 * that is assigned to each thread participating
 * in an rcu qp
 *
 * qp points to the qp that it last acquired
 *
 */
struct rcu_thr_data {
    struct thread_qp thread_qps[MAX_QPS];
};

/*
 * This is the internal version of a CRYPTO_RCU_LOCK
 * it is cast from CRYPTO_RCU_LOCK
 */
struct rcu_lock_st {
    /* Callbacks to call for next ossl_synchronize_rcu */
    struct rcu_cb_item *cb_items;

    /* The context we are being created against */
    OSSL_LIB_CTX *ctx;

    /* Array of quiescent points for synchronization */
    struct rcu_qp *qp_group;

    /* rcu generation counter for in-order retirement */
    uint32_t id_ctr;

    /* Number of elements in qp_group array */
    uint32_t group_count;

    /* Index of the current qp in the qp_group array */
    uint32_t reader_idx;

    /* value of the next id_ctr value to be retired */
    uint32_t next_to_retire;

    /* index of the next free rcu_qp in the qp_group */
    uint32_t current_alloc_idx;

    /* number of qp's in qp_group array currently being retired */
    uint32_t writers_alloced;

    /* lock protecting write side operations */
    pthread_mutex_t write_lock;

    /* lock protecting updates to writers_alloced/current_alloc_idx */
    pthread_mutex_t alloc_lock;

    /* signal to wake threads waiting on alloc_lock */
    pthread_cond_t alloc_signal;

    /* lock to enforce in-order retirement */
    pthread_mutex_t prior_lock;

    /* signal to wake threads waiting on prior_lock */
    pthread_cond_t prior_signal;
};

/* Read side acquisition of the current qp */
static struct rcu_qp *get_hold_current_qp(struct rcu_lock_st *lock)
{
    uint32_t qp_idx;

    /* get the current qp index */
    for (;;) {
        qp_idx = ATOMIC_LOAD_N(uint32_t, &lock->reader_idx, __ATOMIC_RELAXED);

        /*
         * Notes on use of __ATOMIC_ACQUIRE
         * We need to ensure the following:
         * 1) That subsequent operations aren't optimized by hoisting them above
         * this operation.  Specifically, we don't want the below re-load of
         * qp_idx to get optimized away
         * 2) We want to ensure that any updating of reader_idx on the write side
         * of the lock is flushed from a local cpu cache so that we see any
         * updates prior to the load.  This is a non-issue on cache coherent
         * systems like x86, but is relevant on other arches
         */
        ATOMIC_ADD_FETCH(&lock->qp_group[qp_idx].users, (uint64_t)1,
            __ATOMIC_ACQUIRE);

        /* if the idx hasn't changed, we're good, else try again */
        if (qp_idx == ATOMIC_LOAD_N(uint32_t, &lock->reader_idx, __ATOMIC_ACQUIRE))
            break;

        ATOMIC_SUB_FETCH(&lock->qp_group[qp_idx].users, (uint64_t)1,
            __ATOMIC_RELAXED);
    }

    return &lock->qp_group[qp_idx];
}

static void ossl_rcu_free_local_data(void *arg)
{
    OSSL_LIB_CTX *ctx = arg;
    struct rcu_thr_data *data = CRYPTO_THREAD_get_local_ex(CRYPTO_THREAD_LOCAL_RCU_KEY, ctx);

    CRYPTO_THREAD_set_local_ex(CRYPTO_THREAD_LOCAL_RCU_KEY, ctx, NULL);
    OPENSSL_free(data);
}

int ossl_rcu_read_lock(CRYPTO_RCU_LOCK *lock)
{
    struct rcu_thr_data *data;
    int i, available_qp = -1;

    /*
     * we're going to access current_qp here so ask the
     * processor to fetch it
     */
    data = CRYPTO_THREAD_get_local_ex(CRYPTO_THREAD_LOCAL_RCU_KEY, lock->ctx);

    if (data == NULL) {
        data = OPENSSL_zalloc(sizeof(*data));
        if (data == NULL)
            return 0;

        if (!CRYPTO_THREAD_set_local_ex(CRYPTO_THREAD_LOCAL_RCU_KEY, lock->ctx, data)) {
            OPENSSL_free(data);
            return 0;
        }
        if (!ossl_init_thread_start(NULL, lock->ctx, ossl_rcu_free_local_data)) {
            OPENSSL_free(data);
            CRYPTO_THREAD_set_local_ex(CRYPTO_THREAD_LOCAL_RCU_KEY, lock->ctx, NULL);
            return 0;
        }
    }

    for (i = 0; i < MAX_QPS; i++) {
        if (data->thread_qps[i].qp == NULL && available_qp == -1)
            available_qp = i;
        /* If we have a hold on this lock already, we're good */
        if (data->thread_qps[i].lock == lock) {
            data->thread_qps[i].depth++;
            return 1;
        }
    }

    /*
     * if we get here, then we don't have a hold on this lock yet
     */
    assert(available_qp != -1);

    data->thread_qps[available_qp].qp = get_hold_current_qp(lock);
    data->thread_qps[available_qp].depth = 1;
    data->thread_qps[available_qp].lock = lock;
    return 1;
}

void ossl_rcu_read_unlock(CRYPTO_RCU_LOCK *lock)
{
    int i;
    struct rcu_thr_data *data = CRYPTO_THREAD_get_local_ex(CRYPTO_THREAD_LOCAL_RCU_KEY, lock->ctx);
    uint64_t ret;

    assert(data != NULL);

    for (i = 0; i < MAX_QPS; i++) {
        if (data->thread_qps[i].lock == lock) {
            /*
             * we have to use __ATOMIC_RELEASE here
             * to ensure that all preceding read instructions complete
             * before the decrement is visible to ossl_synchronize_rcu
             */
            data->thread_qps[i].depth--;
            if (data->thread_qps[i].depth == 0) {
                ret = ATOMIC_SUB_FETCH(&data->thread_qps[i].qp->users,
                    (uint64_t)1, __ATOMIC_RELEASE);
                OPENSSL_assert(ret != UINT64_MAX);
                data->thread_qps[i].qp = NULL;
                data->thread_qps[i].lock = NULL;
            }
            return;
        }
    }
    /*
     * If we get here, we're trying to unlock a lock that we never acquired -
     * that's fatal.
     */
    assert(0);
}

/*
 * Write side allocation routine to get the current qp
 * and replace it with a new one
 */
static struct rcu_qp *update_qp(CRYPTO_RCU_LOCK *lock, uint32_t *curr_id)
{
    uint32_t current_idx;

    pthread_mutex_lock(&lock->alloc_lock);

    /*
     * we need at least one qp to be available with one
     * left over, so that readers can start working on
     * one that isn't yet being waited on
     */
    while (lock->group_count - lock->writers_alloced < 2)
        /* we have to wait for one to be free */
        pthread_cond_wait(&lock->alloc_signal, &lock->alloc_lock);

    current_idx = lock->current_alloc_idx;

    /* Allocate the qp */
    lock->writers_alloced++;

    /* increment the allocation index */
    lock->current_alloc_idx = (lock->current_alloc_idx + 1) % lock->group_count;

    *curr_id = lock->id_ctr;
    lock->id_ctr++;

    /*
     * make the current state of everything visible by this release
     * when get_hold_current_qp acquires the next qp
     */
    ATOMIC_STORE_N(uint32_t, &lock->reader_idx, lock->current_alloc_idx,
        __ATOMIC_RELEASE);

    /*
     * this should make sure that the new value of reader_idx is visible in
     * get_hold_current_qp, directly after incrementing the users count
     */
    ATOMIC_ADD_FETCH(&lock->qp_group[current_idx].users, (uint64_t)0,
        __ATOMIC_RELEASE);

    /* wake up any waiters */
    pthread_cond_signal(&lock->alloc_signal);
    pthread_mutex_unlock(&lock->alloc_lock);
    return &lock->qp_group[current_idx];
}

static void retire_qp(CRYPTO_RCU_LOCK *lock, struct rcu_qp *qp)
{
    pthread_mutex_lock(&lock->alloc_lock);
    lock->writers_alloced--;
    pthread_cond_signal(&lock->alloc_signal);
    pthread_mutex_unlock(&lock->alloc_lock);
}

static struct rcu_qp *allocate_new_qp_group(CRYPTO_RCU_LOCK *lock,
    uint32_t count)
{
    struct rcu_qp *new = OPENSSL_calloc(count, sizeof(*new));

    lock->group_count = count;
    return new;
}

void ossl_rcu_write_lock(CRYPTO_RCU_LOCK *lock)
{
    pthread_mutex_lock(&lock->write_lock);
    TSAN_FAKE_UNLOCK(&lock->write_lock);
}

void ossl_rcu_write_unlock(CRYPTO_RCU_LOCK *lock)
{
    TSAN_FAKE_LOCK(&lock->write_lock);
    pthread_mutex_unlock(&lock->write_lock);
}

void ossl_synchronize_rcu(CRYPTO_RCU_LOCK *lock)
{
    struct rcu_qp *qp;
    uint64_t count;
    uint32_t curr_id;
    struct rcu_cb_item *cb_items, *tmpcb;

    pthread_mutex_lock(&lock->write_lock);
    cb_items = lock->cb_items;
    lock->cb_items = NULL;
    pthread_mutex_unlock(&lock->write_lock);

    qp = update_qp(lock, &curr_id);

    /* retire in order */
    pthread_mutex_lock(&lock->prior_lock);
    while (lock->next_to_retire != curr_id)
        pthread_cond_wait(&lock->prior_signal, &lock->prior_lock);

    /*
     * wait for the reader count to reach zero
     * Note the use of __ATOMIC_ACQUIRE here to ensure that any
     * prior __ATOMIC_RELEASE write operation in ossl_rcu_read_unlock
     * is visible prior to our read
     * however this is likely just necessary to silence a tsan warning
     * because the read side should not do any write operation
     * outside the atomic itself
     */
    do {
        count = ATOMIC_LOAD_N(uint64_t, &qp->users, __ATOMIC_ACQUIRE);
    } while (count != (uint64_t)0);

    lock->next_to_retire++;
    pthread_cond_broadcast(&lock->prior_signal);
    pthread_mutex_unlock(&lock->prior_lock);

    retire_qp(lock, qp);

    /* handle any callbacks that we have */
    while (cb_items != NULL) {
        tmpcb = cb_items;
        cb_items = cb_items->next;
        tmpcb->fn(tmpcb->data);
        OPENSSL_free(tmpcb);
    }
}

/*
 * Note: This call assumes its made under the protection of
 * ossl_rcu_write_lock
 */
int ossl_rcu_call(CRYPTO_RCU_LOCK *lock, rcu_cb_fn cb, void *data)
{
    struct rcu_cb_item *new = OPENSSL_zalloc(sizeof(*new));

    if (new == NULL)
        return 0;

    new->data = data;
    new->fn = cb;

    new->next = lock->cb_items;
    lock->cb_items = new;

    return 1;
}

void *ossl_rcu_uptr_deref(void **p)
{
    return ATOMIC_LOAD_N(pvoid, p, __ATOMIC_ACQUIRE);
}

void ossl_rcu_assign_uptr(void **p, void **v)
{
    ATOMIC_STORE(pvoid, p, v, __ATOMIC_RELEASE);
}

CRYPTO_RCU_LOCK *ossl_rcu_lock_new(int num_writers, OSSL_LIB_CTX *ctx)
{
    struct rcu_lock_st *new;

    /*
     * We need a minimum of 2 qp's
     */
    if (num_writers < 2)
        num_writers = 2;

    ctx = ossl_lib_ctx_get_concrete(ctx);
    if (ctx == NULL)
        return 0;

    new = OPENSSL_zalloc(sizeof(*new));
    if (new == NULL)
        return NULL;

    new->ctx = ctx;
    pthread_mutex_init(&new->write_lock, NULL);
    pthread_mutex_init(&new->prior_lock, NULL);
    pthread_mutex_init(&new->alloc_lock, NULL);
    pthread_cond_init(&new->prior_signal, NULL);
    pthread_cond_init(&new->alloc_signal, NULL);

    new->qp_group = allocate_new_qp_group(new, num_writers);
    if (new->qp_group == NULL) {
        OPENSSL_free(new);
        new = NULL;
    }

    return new;
}

void ossl_rcu_lock_free(CRYPTO_RCU_LOCK *lock)
{
    struct rcu_lock_st *rlock = (struct rcu_lock_st *)lock;

    if (lock == NULL)
        return;

    /* make sure we're synchronized */
    ossl_synchronize_rcu(rlock);

    OPENSSL_free(rlock->qp_group);
    /* There should only be a single qp left now */
    OPENSSL_free(rlock);
}

#ifdef REPORT_RWLOCK_CONTENTION
/*
 * Normally we would use a BIO here to do this, but we create locks during
 * library initialization, and creating a bio too early, creates a recursive set
 * of stack calls that leads us to call CRYPTO_thread_run_once while currently
 * executing the init routine for various run_once functions, which leads to
 * deadlock.  Avoid that by just using a FILE pointer.  Also note that we
 * directly use a pthread_mutex_t to protect access from multiple threads
 * to the contention log file.  We do this because we want to avoid use
 * of the CRYPTO_THREAD api so as to prevent recursive blocking reports.
 */
static CRYPTO_ONCE init_contention_data_flag = CRYPTO_ONCE_STATIC_INIT;
pthread_mutex_t log_lock = PTHREAD_MUTEX_INITIALIZER;
CRYPTO_THREAD_LOCAL thread_contention_data;

struct stack_info {
    unsigned int nptrs;
    int write;
    OSSL_TIME start;
    OSSL_TIME duration;
    char **strings;
};

#define STACKS_COUNT 32
#define BT_BUF_SIZE 1024
struct stack_traces {
    int fd;
    int lock_depth;
    size_t idx;
    struct stack_info stacks[STACKS_COUNT];
};

/* The glibc gettid() definition presents only since 2.30. */
static ossl_inline pid_t get_tid(void)
{
#ifdef OPENSSL_SYS_MACOSX
    /*
     * MACOS has the gettid call, but it does something completely different
     * here than on other unixes.  Specifically it returns the uid of the calling thread
     * (if set), or -1.  We need to use a MACOS specific call to get the thread id here
     */
    uint64_t tid;

    pthread_threadid_np(NULL, &tid);
    return (pid_t)tid;
#else
    return syscall(SYS_gettid);
#endif
}

#ifdef FIPS_MODULE
#define FIPS_SFX "-fips"
#else
#define FIPS_SFX ""
#endif
static void *init_contention_data(void)
{
    struct stack_traces *traces;
    char fname_fmt[] = "lock-contention-log" FIPS_SFX ".%d.txt";
    char fname[sizeof(fname_fmt) + sizeof(int) * 3];

    traces = OPENSSL_zalloc(sizeof(struct stack_traces));

    snprintf(fname, sizeof(fname), fname_fmt, get_tid());

    traces->fd = open(fname, O_WRONLY | O_APPEND | O_CLOEXEC | O_CREAT, 0600);

    return traces;
}

static void destroy_contention_data(void *data)
{
    struct stack_traces *st = data;

    close(st->fd);
    OPENSSL_free(data);
}

static void init_contention_data_once(void)
{
    /*
     * Create a thread local key here to store our list of stack traces
     * to be printed when we unlock the lock we are holding
     */
    CRYPTO_THREAD_init_local(&thread_contention_data, destroy_contention_data);
    return;
}

static struct stack_traces *get_stack_traces(bool init)
{
    struct stack_traces *traces = CRYPTO_THREAD_get_local(&thread_contention_data);

    if (!traces && init) {
        traces = init_contention_data();
        CRYPTO_THREAD_set_local(&thread_contention_data, traces);
    }

    return traces;
}

static void print_stack_traces(struct stack_traces *traces)
{
    unsigned int j;
    struct iovec *iov;
    int iovcnt;

    while (traces != NULL && traces->idx >= 1) {
        traces->idx--;
        dprintf(traces->fd,
            "lock blocked on %s for %zu usec at time %zu tid %d\n",
            traces->stacks[traces->idx].write == 1 ? "WRITE" : "READ",
            ossl_time2us(traces->stacks[traces->idx].duration),
            ossl_time2us(traces->stacks[traces->idx].start),
            get_tid());
        if (traces->stacks[traces->idx].strings != NULL) {
            static const char lf = '\n';

            iovcnt = traces->stacks[traces->idx].nptrs * 2 + 1;
            iov = alloca(iovcnt * sizeof(*iov));
            for (j = 0; j < traces->stacks[traces->idx].nptrs; j++) {
                iov[2 * j].iov_base = traces->stacks[traces->idx].strings[j];
                iov[2 * j].iov_len = strlen(traces->stacks[traces->idx].strings[j]);
                iov[2 * j + 1].iov_base = (char *)&lf;
                iov[2 * j + 1].iov_len = 1;
            }
            iov[traces->stacks[traces->idx].nptrs * 2].iov_base = (char *)&lf;
            iov[traces->stacks[traces->idx].nptrs * 2].iov_len = 1;
        } else {
            static const char no_bt[] = "No stack trace available\n\n";

            iovcnt = 1;
            iov = alloca(iovcnt * sizeof(*iov));
            iov[0].iov_base = (char *)no_bt;
            iov[0].iov_len = sizeof(no_bt) - 1;
        }
        writev(traces->fd, iov, iovcnt);
        free(traces->stacks[traces->idx].strings);
    }
}

static ossl_inline void ossl_init_rwlock_contention_data(void)
{
    CRYPTO_THREAD_run_once(&init_contention_data_flag, init_contention_data_once);
}

static int record_lock_contention(pthread_rwlock_t *lock,
    struct stack_traces *traces, bool write)
{
    void *buffer[BT_BUF_SIZE];
    OSSL_TIME start, end;
    int ret;

    start = ossl_time_now();
    ret = (write ? pthread_rwlock_wrlock : pthread_rwlock_rdlock)(lock);
    if (ret)
        return ret;
    end = ossl_time_now();
    traces->stacks[traces->idx].nptrs = backtrace(buffer, BT_BUF_SIZE);
    traces->stacks[traces->idx].strings = backtrace_symbols(buffer,
        traces->stacks[traces->idx].nptrs);
    traces->stacks[traces->idx].duration = ossl_time_subtract(end, start);
    traces->stacks[traces->idx].start = start;
    traces->stacks[traces->idx].write = write;
    traces->idx++;
    if (traces->idx >= STACKS_COUNT) {
        fprintf(stderr, "STACK RECORD OVERFLOW!\n");
        print_stack_traces(traces);
    }

    return 0;
}

static ossl_inline int ossl_rwlock_rdlock(pthread_rwlock_t *lock)
{
    struct stack_traces *traces = get_stack_traces(true);

    if (ossl_unlikely(traces == NULL))
        return ENOMEM;

    traces->lock_depth++;
    if (pthread_rwlock_tryrdlock(lock)) {
        int ret = record_lock_contention(lock, traces, false);

        if (ret)
            traces->lock_depth--;

        return ret;
    }

    return 0;
}

static ossl_inline int ossl_rwlock_wrlock(pthread_rwlock_t *lock)
{
    struct stack_traces *traces = get_stack_traces(true);

    if (ossl_unlikely(traces == NULL))
        return ENOMEM;

    traces->lock_depth++;
    if (pthread_rwlock_trywrlock(lock)) {
        int ret = record_lock_contention(lock, traces, true);

        if (ret)
            traces->lock_depth--;

        return ret;
    }

    return 0;
}

static ossl_inline int ossl_rwlock_unlock(pthread_rwlock_t *lock)
{
    int ret;

    ret = pthread_rwlock_unlock(lock);
    if (ret)
        return ret;

    {
        struct stack_traces *traces = get_stack_traces(false);

        if (traces != NULL) {
            traces->lock_depth--;
            assert(traces->lock_depth >= 0);
            if (traces->lock_depth == 0)
                print_stack_traces(traces);
        }
    }

    return 0;
}

#else /* !REPORT_RWLOCK_CONTENTION */

static ossl_inline void ossl_init_rwlock_contention_data(void)
{
}

static ossl_inline int ossl_rwlock_rdlock(pthread_rwlock_t *rwlock)
{
    return pthread_rwlock_rdlock(rwlock);
}

static ossl_inline int ossl_rwlock_wrlock(pthread_rwlock_t *rwlock)
{
    return pthread_rwlock_wrlock(rwlock);
}

static ossl_inline int ossl_rwlock_unlock(pthread_rwlock_t *rwlock)
{
    return pthread_rwlock_unlock(rwlock);
}
#endif /* REPORT_RWLOCK_CONTENTION */

CRYPTO_RWLOCK *CRYPTO_THREAD_lock_new(void)
{
#ifdef USE_RWLOCK
    CRYPTO_RWLOCK *lock;

    ossl_init_rwlock_contention_data();

    if ((lock = OPENSSL_zalloc(sizeof(pthread_rwlock_t))) == NULL)
        /* Don't set error, to avoid recursion blowup. */
        return NULL;

    if (pthread_rwlock_init(lock, NULL) != 0) {
        OPENSSL_free(lock);
        return NULL;
    }
#else
    pthread_mutexattr_t attr;
    CRYPTO_RWLOCK *lock;

    if ((lock = OPENSSL_zalloc(sizeof(pthread_mutex_t))) == NULL)
        /* Don't set error, to avoid recursion blowup. */
        return NULL;

    /*
     * We don't use recursive mutexes, but try to catch errors if we do.
     */
    pthread_mutexattr_init(&attr);
#if !defined(__TANDEM) && !defined(_SPT_MODEL_)
#if !defined(NDEBUG) && !defined(OPENSSL_NO_MUTEX_ERRORCHECK)
    pthread_mutexattr_settype(&attr, PTHREAD_MUTEX_ERRORCHECK);
#endif
#else
    /* The SPT Thread Library does not define MUTEX attributes. */
#endif

    if (pthread_mutex_init(lock, &attr) != 0) {
        pthread_mutexattr_destroy(&attr);
        OPENSSL_free(lock);
        return NULL;
    }

    pthread_mutexattr_destroy(&attr);
#endif

    return lock;
}

__owur int CRYPTO_THREAD_read_lock(CRYPTO_RWLOCK *lock)
{
#ifdef USE_RWLOCK
    if (!ossl_assert(ossl_rwlock_rdlock(lock) == 0))
        return 0;
#else
    if (pthread_mutex_lock(lock) != 0) {
        assert(errno != EDEADLK && errno != EBUSY);
        return 0;
    }
#endif

    return 1;
}

__owur int CRYPTO_THREAD_write_lock(CRYPTO_RWLOCK *lock)
{
#ifdef USE_RWLOCK
    if (!ossl_assert(ossl_rwlock_wrlock(lock) == 0))
        return 0;
#else
    if (pthread_mutex_lock(lock) != 0) {
        assert(errno != EDEADLK && errno != EBUSY);
        return 0;
    }
#endif

    return 1;
}

int CRYPTO_THREAD_unlock(CRYPTO_RWLOCK *lock)
{
#ifdef USE_RWLOCK
    if (ossl_rwlock_unlock(lock) != 0)
        return 0;
#else
    if (pthread_mutex_unlock(lock) != 0) {
        assert(errno != EPERM);
        return 0;
    }
#endif

    return 1;
}

void CRYPTO_THREAD_lock_free(CRYPTO_RWLOCK *lock)
{
    if (lock == NULL)
        return;

#ifdef USE_RWLOCK
    pthread_rwlock_destroy(lock);
#else
    pthread_mutex_destroy(lock);
#endif
    OPENSSL_free(lock);

    return;
}

int CRYPTO_THREAD_run_once(CRYPTO_ONCE *once, void (*init)(void))
{
    if (ossl_unlikely(pthread_once(once, init) != 0))
        return 0;

    return 1;
}

int CRYPTO_THREAD_init_local(CRYPTO_THREAD_LOCAL *key, void (*cleanup)(void *))
{
    if (pthread_key_create(key, cleanup) != 0)
        return 0;

    return 1;
}

void *CRYPTO_THREAD_get_local(CRYPTO_THREAD_LOCAL *key)
{
    return pthread_getspecific(*key);
}

int CRYPTO_THREAD_set_local(CRYPTO_THREAD_LOCAL *key, void *val)
{
    if (pthread_setspecific(*key, val) != 0)
        return 0;

    return 1;
}

int CRYPTO_THREAD_cleanup_local(CRYPTO_THREAD_LOCAL *key)
{
    if (pthread_key_delete(*key) != 0)
        return 0;

    return 1;
}

CRYPTO_THREAD_ID CRYPTO_THREAD_get_current_id(void)
{
    return pthread_self();
}

int CRYPTO_THREAD_compare_id(CRYPTO_THREAD_ID a, CRYPTO_THREAD_ID b)
{
    return pthread_equal(a, b);
}

int CRYPTO_atomic_add(int *val, int amount, int *ret, CRYPTO_RWLOCK *lock)
{
#if defined(__GNUC__) && defined(__ATOMIC_ACQ_REL) && !defined(BROKEN_CLANG_ATOMICS)
    if (__atomic_is_lock_free(sizeof(*val), val)) {
        *ret = __atomic_add_fetch(val, amount, __ATOMIC_ACQ_REL);
        return 1;
    }
#elif defined(__sun) && (defined(__SunOS_5_10) || defined(__SunOS_5_11))
    /* This will work for all future Solaris versions. */
    if (ret != NULL) {
        *ret = atomic_add_int_nv((volatile unsigned int *)val, amount);
        return 1;
    }
#endif
    if (lock == NULL || !CRYPTO_THREAD_write_lock(lock))
        return 0;

    *val += amount;
    *ret = *val;

    if (!CRYPTO_THREAD_unlock(lock))
        return 0;

    return 1;
}

int CRYPTO_atomic_add64(uint64_t *val, uint64_t op, uint64_t *ret,
    CRYPTO_RWLOCK *lock)
{
#if defined(__GNUC__) && defined(__ATOMIC_ACQ_REL) && !defined(BROKEN_CLANG_ATOMICS)
    if (__atomic_is_lock_free(sizeof(*val), val)) {
        *ret = __atomic_add_fetch(val, op, __ATOMIC_ACQ_REL);
        return 1;
    }
#elif defined(__sun) && (defined(__SunOS_5_10) || defined(__SunOS_5_11))
    /* This will work for all future Solaris versions. */
    if (ret != NULL) {
        *ret = atomic_add_64_nv(val, op);
        return 1;
    }
#endif
    if (lock == NULL || !CRYPTO_THREAD_write_lock(lock))
        return 0;
    *val += op;
    *ret = *val;

    if (!CRYPTO_THREAD_unlock(lock))
        return 0;

    return 1;
}

int CRYPTO_atomic_and(uint64_t *val, uint64_t op, uint64_t *ret,
    CRYPTO_RWLOCK *lock)
{
#if defined(__GNUC__) && defined(__ATOMIC_ACQ_REL) && !defined(BROKEN_CLANG_ATOMICS)
    if (__atomic_is_lock_free(sizeof(*val), val)) {
        *ret = __atomic_and_fetch(val, op, __ATOMIC_ACQ_REL);
        return 1;
    }
#elif defined(__sun) && (defined(__SunOS_5_10) || defined(__SunOS_5_11))
    /* This will work for all future Solaris versions. */
    if (ret != NULL) {
        *ret = atomic_and_64_nv(val, op);
        return 1;
    }
#endif
    if (lock == NULL || !CRYPTO_THREAD_write_lock(lock))
        return 0;
    *val &= op;
    *ret = *val;

    if (!CRYPTO_THREAD_unlock(lock))
        return 0;

    return 1;
}

int CRYPTO_atomic_or(uint64_t *val, uint64_t op, uint64_t *ret,
    CRYPTO_RWLOCK *lock)
{
#if defined(__GNUC__) && defined(__ATOMIC_ACQ_REL) && !defined(BROKEN_CLANG_ATOMICS)
    if (__atomic_is_lock_free(sizeof(*val), val)) {
        *ret = __atomic_or_fetch(val, op, __ATOMIC_ACQ_REL);
        return 1;
    }
#elif defined(__sun) && (defined(__SunOS_5_10) || defined(__SunOS_5_11))
    /* This will work for all future Solaris versions. */
    if (ret != NULL) {
        *ret = atomic_or_64_nv(val, op);
        return 1;
    }
#endif
    if (lock == NULL || !CRYPTO_THREAD_write_lock(lock))
        return 0;
    *val |= op;
    *ret = *val;

    if (!CRYPTO_THREAD_unlock(lock))
        return 0;

    return 1;
}

int CRYPTO_atomic_load(uint64_t *val, uint64_t *ret, CRYPTO_RWLOCK *lock)
{
#if defined(__GNUC__) && defined(__ATOMIC_ACQ_REL) && !defined(BROKEN_CLANG_ATOMICS)
    if (__atomic_is_lock_free(sizeof(*val), val)) {
        __atomic_load(val, ret, __ATOMIC_ACQUIRE);
        return 1;
    }
#elif defined(__sun) && (defined(__SunOS_5_10) || defined(__SunOS_5_11))
    /* This will work for all future Solaris versions. */
    if (ret != NULL) {
        *ret = atomic_or_64_nv(val, 0);
        return 1;
    }
#endif
    if (lock == NULL || !CRYPTO_THREAD_read_lock(lock))
        return 0;
    *ret = *val;
    if (!CRYPTO_THREAD_unlock(lock))
        return 0;

    return 1;
}

int CRYPTO_atomic_store(uint64_t *dst, uint64_t val, CRYPTO_RWLOCK *lock)
{
#if defined(__GNUC__) && defined(__ATOMIC_ACQ_REL) && !defined(BROKEN_CLANG_ATOMICS)
    if (__atomic_is_lock_free(sizeof(*dst), dst)) {
        __atomic_store(dst, &val, __ATOMIC_RELEASE);
        return 1;
    }
#elif defined(__sun) && (defined(__SunOS_5_10) || defined(__SunOS_5_11))
    /* This will work for all future Solaris versions. */
    if (dst != NULL) {
        atomic_swap_64(dst, val);
        return 1;
    }
#endif
    if (lock == NULL || !CRYPTO_THREAD_write_lock(lock))
        return 0;
    *dst = val;
    if (!CRYPTO_THREAD_unlock(lock))
        return 0;

    return 1;
}

int CRYPTO_atomic_load_int(int *val, int *ret, CRYPTO_RWLOCK *lock)
{
#if defined(__GNUC__) && defined(__ATOMIC_ACQ_REL) && !defined(BROKEN_CLANG_ATOMICS)
    if (__atomic_is_lock_free(sizeof(*val), val)) {
        __atomic_load(val, ret, __ATOMIC_ACQUIRE);
        return 1;
    }
#elif defined(__sun) && (defined(__SunOS_5_10) || defined(__SunOS_5_11))
    /* This will work for all future Solaris versions. */
    if (ret != NULL) {
        *ret = (int)atomic_or_uint_nv((unsigned int *)val, 0);
        return 1;
    }
#endif
    if (lock == NULL || !CRYPTO_THREAD_read_lock(lock))
        return 0;
    *ret = *val;
    if (!CRYPTO_THREAD_unlock(lock))
        return 0;

    return 1;
}

#ifndef FIPS_MODULE
int openssl_init_fork_handlers(void)
{
    return 1;
}
#endif /* FIPS_MODULE */

int openssl_get_fork_id(void)
{
    return getpid();
}
#endif