diff options
Diffstat (limited to 'deps/jemalloc/src')
-rw-r--r-- | deps/jemalloc/src/background_thread.c | 880 | ||||
-rw-r--r-- | deps/jemalloc/src/extent_dss.c | 269 | ||||
-rw-r--r-- | deps/jemalloc/src/extent_mmap.c | 42 | ||||
-rw-r--r-- | deps/jemalloc/src/hooks.c | 12 | ||||
-rw-r--r-- | deps/jemalloc/src/jemalloc_cpp.cpp | 132 | ||||
-rw-r--r-- | deps/jemalloc/src/large.c | 371 | ||||
-rw-r--r-- | deps/jemalloc/src/malloc_io.c | 689 | ||||
-rw-r--r-- | deps/jemalloc/src/mutex_pool.c | 18 | ||||
-rw-r--r-- | deps/jemalloc/src/nstime.c | 170 | ||||
-rw-r--r-- | deps/jemalloc/src/pages.c | 423 | ||||
-rw-r--r-- | deps/jemalloc/src/prng.c | 3 | ||||
-rw-r--r-- | deps/jemalloc/src/spin.c | 4 | ||||
-rw-r--r-- | deps/jemalloc/src/sz.c | 106 | ||||
-rw-r--r-- | deps/jemalloc/src/ticker.c | 3 | ||||
-rw-r--r-- | deps/jemalloc/src/witness.c | 100 |
15 files changed, 3222 insertions, 0 deletions
diff --git a/deps/jemalloc/src/background_thread.c b/deps/jemalloc/src/background_thread.c new file mode 100644 index 0000000000..eb30eb5b42 --- /dev/null +++ b/deps/jemalloc/src/background_thread.c @@ -0,0 +1,880 @@ +#define JEMALLOC_BACKGROUND_THREAD_C_ +#include "jemalloc/internal/jemalloc_preamble.h" +#include "jemalloc/internal/jemalloc_internal_includes.h" + +#include "jemalloc/internal/assert.h" + +/******************************************************************************/ +/* Data. */ + +/* This option should be opt-in only. */ +#define BACKGROUND_THREAD_DEFAULT false +/* Read-only after initialization. */ +bool opt_background_thread = BACKGROUND_THREAD_DEFAULT; + +/* Used for thread creation, termination and stats. */ +malloc_mutex_t background_thread_lock; +/* Indicates global state. Atomic because decay reads this w/o locking. */ +atomic_b_t background_thread_enabled_state; +size_t n_background_threads; +/* Thread info per-index. */ +background_thread_info_t *background_thread_info; + +/* False if no necessary runtime support. */ +bool can_enable_background_thread; + +/******************************************************************************/ + +#ifdef JEMALLOC_PTHREAD_CREATE_WRAPPER +#include <dlfcn.h> + +static int (*pthread_create_fptr)(pthread_t *__restrict, const pthread_attr_t *, + void *(*)(void *), void *__restrict); +static pthread_once_t once_control = PTHREAD_ONCE_INIT; + +static void +pthread_create_wrapper_once(void) { +#ifdef JEMALLOC_LAZY_LOCK + isthreaded = true; +#endif +} + +int +pthread_create_wrapper(pthread_t *__restrict thread, const pthread_attr_t *attr, + void *(*start_routine)(void *), void *__restrict arg) { + pthread_once(&once_control, pthread_create_wrapper_once); + + return pthread_create_fptr(thread, attr, start_routine, arg); +} +#endif /* JEMALLOC_PTHREAD_CREATE_WRAPPER */ + +#ifndef JEMALLOC_BACKGROUND_THREAD +#define NOT_REACHED { not_reached(); } +bool background_thread_create(tsd_t *tsd, unsigned arena_ind) NOT_REACHED +bool background_threads_enable(tsd_t *tsd) NOT_REACHED +bool background_threads_disable(tsd_t *tsd) NOT_REACHED +void background_thread_interval_check(tsdn_t *tsdn, arena_t *arena, + arena_decay_t *decay, size_t npages_new) NOT_REACHED +void background_thread_prefork0(tsdn_t *tsdn) NOT_REACHED +void background_thread_prefork1(tsdn_t *tsdn) NOT_REACHED +void background_thread_postfork_parent(tsdn_t *tsdn) NOT_REACHED +void background_thread_postfork_child(tsdn_t *tsdn) NOT_REACHED +bool background_thread_stats_read(tsdn_t *tsdn, + background_thread_stats_t *stats) NOT_REACHED +void background_thread_ctl_init(tsdn_t *tsdn) NOT_REACHED +#undef NOT_REACHED +#else + +static bool background_thread_enabled_at_fork; + +static void +background_thread_info_init(tsdn_t *tsdn, background_thread_info_t *info) { + background_thread_wakeup_time_set(tsdn, info, 0); + info->npages_to_purge_new = 0; + if (config_stats) { + info->tot_n_runs = 0; + nstime_init(&info->tot_sleep_time, 0); + } +} + +static inline bool +set_current_thread_affinity(UNUSED int cpu) { +#if defined(JEMALLOC_HAVE_SCHED_SETAFFINITY) + cpu_set_t cpuset; + CPU_ZERO(&cpuset); + CPU_SET(cpu, &cpuset); + int ret = sched_setaffinity(0, sizeof(cpu_set_t), &cpuset); + + return (ret != 0); +#else + return false; +#endif +} + +/* Threshold for determining when to wake up the background thread. */ +#define BACKGROUND_THREAD_NPAGES_THRESHOLD UINT64_C(1024) +#define BILLION UINT64_C(1000000000) +/* Minimal sleep interval 100 ms. */ +#define BACKGROUND_THREAD_MIN_INTERVAL_NS (BILLION / 10) + +static inline size_t +decay_npurge_after_interval(arena_decay_t *decay, size_t interval) { + size_t i; + uint64_t sum = 0; + for (i = 0; i < interval; i++) { + sum += decay->backlog[i] * h_steps[i]; + } + for (; i < SMOOTHSTEP_NSTEPS; i++) { + sum += decay->backlog[i] * (h_steps[i] - h_steps[i - interval]); + } + + return (size_t)(sum >> SMOOTHSTEP_BFP); +} + +static uint64_t +arena_decay_compute_purge_interval_impl(tsdn_t *tsdn, arena_decay_t *decay, + extents_t *extents) { + if (malloc_mutex_trylock(tsdn, &decay->mtx)) { + /* Use minimal interval if decay is contended. */ + return BACKGROUND_THREAD_MIN_INTERVAL_NS; + } + + uint64_t interval; + ssize_t decay_time = atomic_load_zd(&decay->time_ms, ATOMIC_RELAXED); + if (decay_time <= 0) { + /* Purging is eagerly done or disabled currently. */ + interval = BACKGROUND_THREAD_INDEFINITE_SLEEP; + goto label_done; + } + + uint64_t decay_interval_ns = nstime_ns(&decay->interval); + assert(decay_interval_ns > 0); + size_t npages = extents_npages_get(extents); + if (npages == 0) { + unsigned i; + for (i = 0; i < SMOOTHSTEP_NSTEPS; i++) { + if (decay->backlog[i] > 0) { + break; + } + } + if (i == SMOOTHSTEP_NSTEPS) { + /* No dirty pages recorded. Sleep indefinitely. */ + interval = BACKGROUND_THREAD_INDEFINITE_SLEEP; + goto label_done; + } + } + if (npages <= BACKGROUND_THREAD_NPAGES_THRESHOLD) { + /* Use max interval. */ + interval = decay_interval_ns * SMOOTHSTEP_NSTEPS; + goto label_done; + } + + size_t lb = BACKGROUND_THREAD_MIN_INTERVAL_NS / decay_interval_ns; + size_t ub = SMOOTHSTEP_NSTEPS; + /* Minimal 2 intervals to ensure reaching next epoch deadline. */ + lb = (lb < 2) ? 2 : lb; + if ((decay_interval_ns * ub <= BACKGROUND_THREAD_MIN_INTERVAL_NS) || + (lb + 2 > ub)) { + interval = BACKGROUND_THREAD_MIN_INTERVAL_NS; + goto label_done; + } + + assert(lb + 2 <= ub); + size_t npurge_lb, npurge_ub; + npurge_lb = decay_npurge_after_interval(decay, lb); + if (npurge_lb > BACKGROUND_THREAD_NPAGES_THRESHOLD) { + interval = decay_interval_ns * lb; + goto label_done; + } + npurge_ub = decay_npurge_after_interval(decay, ub); + if (npurge_ub < BACKGROUND_THREAD_NPAGES_THRESHOLD) { + interval = decay_interval_ns * ub; + goto label_done; + } + + unsigned n_search = 0; + size_t target, npurge; + while ((npurge_lb + BACKGROUND_THREAD_NPAGES_THRESHOLD < npurge_ub) + && (lb + 2 < ub)) { + target = (lb + ub) / 2; + npurge = decay_npurge_after_interval(decay, target); + if (npurge > BACKGROUND_THREAD_NPAGES_THRESHOLD) { + ub = target; + npurge_ub = npurge; + } else { + lb = target; + npurge_lb = npurge; + } + assert(n_search++ < lg_floor(SMOOTHSTEP_NSTEPS) + 1); + } + interval = decay_interval_ns * (ub + lb) / 2; +label_done: + interval = (interval < BACKGROUND_THREAD_MIN_INTERVAL_NS) ? + BACKGROUND_THREAD_MIN_INTERVAL_NS : interval; + malloc_mutex_unlock(tsdn, &decay->mtx); + + return interval; +} + +/* Compute purge interval for background threads. */ +static uint64_t +arena_decay_compute_purge_interval(tsdn_t *tsdn, arena_t *arena) { + uint64_t i1, i2; + i1 = arena_decay_compute_purge_interval_impl(tsdn, &arena->decay_dirty, + &arena->extents_dirty); + if (i1 == BACKGROUND_THREAD_MIN_INTERVAL_NS) { + return i1; + } + i2 = arena_decay_compute_purge_interval_impl(tsdn, &arena->decay_muzzy, + &arena->extents_muzzy); + + return i1 < i2 ? i1 : i2; +} + +static void +background_thread_sleep(tsdn_t *tsdn, background_thread_info_t *info, + uint64_t interval) { + if (config_stats) { + info->tot_n_runs++; + } + info->npages_to_purge_new = 0; + + struct timeval tv; + /* Specific clock required by timedwait. */ + gettimeofday(&tv, NULL); + nstime_t before_sleep; + nstime_init2(&before_sleep, tv.tv_sec, tv.tv_usec * 1000); + + int ret; + if (interval == BACKGROUND_THREAD_INDEFINITE_SLEEP) { + assert(background_thread_indefinite_sleep(info)); + ret = pthread_cond_wait(&info->cond, &info->mtx.lock); + assert(ret == 0); + } else { + assert(interval >= BACKGROUND_THREAD_MIN_INTERVAL_NS && + interval <= BACKGROUND_THREAD_INDEFINITE_SLEEP); + /* We need malloc clock (can be different from tv). */ + nstime_t next_wakeup; + nstime_init(&next_wakeup, 0); + nstime_update(&next_wakeup); + nstime_iadd(&next_wakeup, interval); + assert(nstime_ns(&next_wakeup) < + BACKGROUND_THREAD_INDEFINITE_SLEEP); + background_thread_wakeup_time_set(tsdn, info, + nstime_ns(&next_wakeup)); + + nstime_t ts_wakeup; + nstime_copy(&ts_wakeup, &before_sleep); + nstime_iadd(&ts_wakeup, interval); + struct timespec ts; + ts.tv_sec = (size_t)nstime_sec(&ts_wakeup); + ts.tv_nsec = (size_t)nstime_nsec(&ts_wakeup); + + assert(!background_thread_indefinite_sleep(info)); + ret = pthread_cond_timedwait(&info->cond, &info->mtx.lock, &ts); + assert(ret == ETIMEDOUT || ret == 0); + background_thread_wakeup_time_set(tsdn, info, + BACKGROUND_THREAD_INDEFINITE_SLEEP); + } + if (config_stats) { + gettimeofday(&tv, NULL); + nstime_t after_sleep; + nstime_init2(&after_sleep, tv.tv_sec, tv.tv_usec * 1000); + if (nstime_compare(&after_sleep, &before_sleep) > 0) { + nstime_subtract(&after_sleep, &before_sleep); + nstime_add(&info->tot_sleep_time, &after_sleep); + } + } +} + +static bool +background_thread_pause_check(tsdn_t *tsdn, background_thread_info_t *info) { + if (unlikely(info->state == background_thread_paused)) { + malloc_mutex_unlock(tsdn, &info->mtx); + /* Wait on global lock to update status. */ + malloc_mutex_lock(tsdn, &background_thread_lock); + malloc_mutex_unlock(tsdn, &background_thread_lock); + malloc_mutex_lock(tsdn, &info->mtx); + return true; + } + + return false; +} + +static inline void +background_work_sleep_once(tsdn_t *tsdn, background_thread_info_t *info, unsigned ind) { + uint64_t min_interval = BACKGROUND_THREAD_INDEFINITE_SLEEP; + unsigned narenas = narenas_total_get(); + + for (unsigned i = ind; i < narenas; i += ncpus) { + arena_t *arena = arena_get(tsdn, i, false); + if (!arena) { + continue; + } + arena_decay(tsdn, arena, true, false); + if (min_interval == BACKGROUND_THREAD_MIN_INTERVAL_NS) { + /* Min interval will be used. */ + continue; + } + uint64_t interval = arena_decay_compute_purge_interval(tsdn, + arena); + assert(interval >= BACKGROUND_THREAD_MIN_INTERVAL_NS); + if (min_interval > interval) { + min_interval = interval; + } + } + background_thread_sleep(tsdn, info, min_interval); +} + +static bool +background_threads_disable_single(tsd_t *tsd, background_thread_info_t *info) { + if (info == &background_thread_info[0]) { + malloc_mutex_assert_owner(tsd_tsdn(tsd), + &background_thread_lock); + } else { + malloc_mutex_assert_not_owner(tsd_tsdn(tsd), + &background_thread_lock); + } + + pre_reentrancy(tsd, NULL); + malloc_mutex_lock(tsd_tsdn(tsd), &info->mtx); + bool has_thread; + assert(info->state != background_thread_paused); + if (info->state == background_thread_started) { + has_thread = true; + info->state = background_thread_stopped; + pthread_cond_signal(&info->cond); + } else { + has_thread = false; + } + malloc_mutex_unlock(tsd_tsdn(tsd), &info->mtx); + + if (!has_thread) { + post_reentrancy(tsd); + return false; + } + void *ret; + if (pthread_join(info->thread, &ret)) { + post_reentrancy(tsd); + return true; + } + assert(ret == NULL); + n_background_threads--; + post_reentrancy(tsd); + + return false; +} + +static void *background_thread_entry(void *ind_arg); + +static int +background_thread_create_signals_masked(pthread_t *thread, + const pthread_attr_t *attr, void *(*start_routine)(void *), void *arg) { + /* + * Mask signals during thread creation so that the thread inherits + * an empty signal set. + */ + sigset_t set; + sigfillset(&set); + sigset_t oldset; + int mask_err = pthread_sigmask(SIG_SETMASK, &set, &oldset); + if (mask_err != 0) { + return mask_err; + } + int create_err = pthread_create_wrapper(thread, attr, start_routine, + arg); + /* + * Restore the signal mask. Failure to restore the signal mask here + * changes program behavior. + */ + int restore_err = pthread_sigmask(SIG_SETMASK, &oldset, NULL); + if (restore_err != 0) { + malloc_printf("<jemalloc>: background thread creation " + "failed (%d), and signal mask restoration failed " + "(%d)\n", create_err, restore_err); + if (opt_abort) { + abort(); + } + } + return create_err; +} + +static void +check_background_thread_creation(tsd_t *tsd, unsigned *n_created, + bool *created_threads) { + if (likely(*n_created == n_background_threads)) { + return; + } + + malloc_mutex_unlock(tsd_tsdn(tsd), &background_thread_info[0].mtx); +label_restart: + malloc_mutex_lock(tsd_tsdn(tsd), &background_thread_lock); + for (unsigned i = 1; i < ncpus; i++) { + if (created_threads[i]) { + continue; + } + background_thread_info_t *info = &background_thread_info[i]; + malloc_mutex_lock(tsd_tsdn(tsd), &info->mtx); + assert(info->state != background_thread_paused); + bool create = (info->state == background_thread_started); + malloc_mutex_unlock(tsd_tsdn(tsd), &info->mtx); + if (!create) { + continue; + } + + /* + * To avoid deadlock with prefork handlers (which waits for the + * mutex held here), unlock before calling pthread_create(). + */ + malloc_mutex_unlock(tsd_tsdn(tsd), &background_thread_lock); + + pre_reentrancy(tsd, NULL); + int err = background_thread_create_signals_masked(&info->thread, + NULL, background_thread_entry, (void *)(uintptr_t)i); + post_reentrancy(tsd); + + if (err == 0) { + (*n_created)++; + created_threads[i] = true; + } else { + malloc_printf("<jemalloc>: background thread " + "creation failed (%d)\n", err); + if (opt_abort) { + abort(); + } + } + /* Restart since we unlocked. */ + goto label_restart; + } + malloc_mutex_lock(tsd_tsdn(tsd), &background_thread_info[0].mtx); + malloc_mutex_unlock(tsd_tsdn(tsd), &background_thread_lock); +} + +static void +background_thread0_work(tsd_t *tsd) { + /* Thread0 is also responsible for launching / terminating threads. */ + VARIABLE_ARRAY(bool, created_threads, ncpus); + unsigned i; + for (i = 1; i < ncpus; i++) { + created_threads[i] = false; + } + /* Start working, and create more threads when asked. */ + unsigned n_created = 1; + while (background_thread_info[0].state != background_thread_stopped) { + if (background_thread_pause_check(tsd_tsdn(tsd), + &background_thread_info[0])) { + continue; + } + check_background_thread_creation(tsd, &n_created, + (bool *)&created_threads); + background_work_sleep_once(tsd_tsdn(tsd), + &background_thread_info[0], 0); + } + + /* + * Shut down other threads at exit. Note that the ctl thread is holding + * the global background_thread mutex (and is waiting) for us. + */ + assert(!background_thread_enabled()); + for (i = 1; i < ncpus; i++) { + background_thread_info_t *info = &background_thread_info[i]; + assert(info->state != background_thread_paused); + if (created_threads[i]) { + background_threads_disable_single(tsd, info); + } else { + malloc_mutex_lock(tsd_tsdn(tsd), &info->mtx); + /* Clear in case the thread wasn't created. */ + info->state = background_thread_stopped; + malloc_mutex_unlock(tsd_tsdn(tsd), &info->mtx); + } + } + background_thread_info[0].state = background_thread_stopped; + assert(n_background_threads == 1); +} + +static void +background_work(tsd_t *tsd, unsigned ind) { + background_thread_info_t *info = &background_thread_info[ind]; + + malloc_mutex_lock(tsd_tsdn(tsd), &info->mtx); + background_thread_wakeup_time_set(tsd_tsdn(tsd), info, + BACKGROUND_THREAD_INDEFINITE_SLEEP); + if (ind == 0) { + background_thread0_work(tsd); + } else { + while (info->state != background_thread_stopped) { + if (background_thread_pause_check(tsd_tsdn(tsd), + info)) { + continue; + } + background_work_sleep_once(tsd_tsdn(tsd), info, ind); + } + } + assert(info->state == background_thread_stopped); + background_thread_wakeup_time_set(tsd_tsdn(tsd), info, 0); + malloc_mutex_unlock(tsd_tsdn(tsd), &info->mtx); +} + +static void * +background_thread_entry(void *ind_arg) { + unsigned thread_ind = (unsigned)(uintptr_t)ind_arg; + assert(thread_ind < ncpus); +#ifdef JEMALLOC_HAVE_PTHREAD_SETNAME_NP + pthread_setname_np(pthread_self(), "jemalloc_bg_thd"); +#endif + if (opt_percpu_arena != percpu_arena_disabled) { + set_current_thread_affinity((int)thread_ind); + } + /* + * Start periodic background work. We use internal tsd which avoids + * side effects, for example triggering new arena creation (which in + * turn triggers another background thread creation). + */ + background_work(tsd_internal_fetch(), thread_ind); + assert(pthread_equal(pthread_self(), + background_thread_info[thread_ind].thread)); + + return NULL; +} + +static void +background_thread_init(tsd_t *tsd, background_thread_info_t *info) { + malloc_mutex_assert_owner(tsd_tsdn(tsd), &background_thread_lock); + info->state = background_thread_started; + background_thread_info_init(tsd_tsdn(tsd), info); + n_background_threads++; +} + +/* Create a new background thread if needed. */ +bool +background_thread_create(tsd_t *tsd, unsigned arena_ind) { + assert(have_background_thread); + malloc_mutex_assert_owner(tsd_tsdn(tsd), &background_thread_lock); + + /* We create at most NCPUs threads. */ + size_t thread_ind = arena_ind % ncpus; + background_thread_info_t *info = &background_thread_info[thread_ind]; + + bool need_new_thread; + malloc_mutex_lock(tsd_tsdn(tsd), &info->mtx); + need_new_thread = background_thread_enabled() && + (info->state == background_thread_stopped); + if (need_new_thread) { + background_thread_init(tsd, info); + } + malloc_mutex_unlock(tsd_tsdn(tsd), &info->mtx); + if (!need_new_thread) { + return false; + } + if (arena_ind != 0) { + /* Threads are created asynchronously by Thread 0. */ + background_thread_info_t *t0 = &background_thread_info[0]; + malloc_mutex_lock(tsd_tsdn(tsd), &t0->mtx); + assert(t0->state == background_thread_started); + pthread_cond_signal(&t0->cond); + malloc_mutex_unlock(tsd_tsdn(tsd), &t0->mtx); + + return false; + } + + pre_reentrancy(tsd, NULL); + /* + * To avoid complications (besides reentrancy), create internal + * background threads with the underlying pthread_create. + */ + int err = background_thread_create_signals_masked(&info->thread, NULL, + background_thread_entry, (void *)thread_ind); + post_reentrancy(tsd); + + if (err != 0) { + malloc_printf("<jemalloc>: arena 0 background thread creation " + "failed (%d)\n", err); + malloc_mutex_lock(tsd_tsdn(tsd), &info->mtx); + info->state = background_thread_stopped; + n_background_threads--; + malloc_mutex_unlock(tsd_tsdn(tsd), &info->mtx); + + return true; + } + + return false; +} + +bool +background_threads_enable(tsd_t *tsd) { + assert(n_background_threads == 0); + assert(background_thread_enabled()); + malloc_mutex_assert_owner(tsd_tsdn(tsd), &background_thread_lock); + + VARIABLE_ARRAY(bool, marked, ncpus); + unsigned i, nmarked; + for (i = 0; i < ncpus; i++) { + marked[i] = false; + } + nmarked = 0; + /* Mark the threads we need to create for thread 0. */ + unsigned n = narenas_total_get(); + for (i = 1; i < n; i++) { + if (marked[i % ncpus] || + arena_get(tsd_tsdn(tsd), i, false) == NULL) { + continue; + } + background_thread_info_t *info = &background_thread_info[i]; + malloc_mutex_lock(tsd_tsdn(tsd), &info->mtx); + assert(info->state == background_thread_stopped); + background_thread_init(tsd, info); + malloc_mutex_unlock(tsd_tsdn(tsd), &info->mtx); + marked[i % ncpus] = true; + if (++nmarked == ncpus) { + break; + } + } + + return background_thread_create(tsd, 0); +} + +bool +background_threads_disable(tsd_t *tsd) { + assert(!background_thread_enabled()); + malloc_mutex_assert_owner(tsd_tsdn(tsd), &background_thread_lock); + + /* Thread 0 will be responsible for terminating other threads. */ + if (background_threads_disable_single(tsd, + &background_thread_info[0])) { + return true; + } + assert(n_background_threads == 0); + + return false; +} + +/* Check if we need to signal the background thread early. */ +void +background_thread_interval_check(tsdn_t *tsdn, arena_t *arena, + arena_decay_t *decay, size_t npages_new) { + background_thread_info_t *info = arena_background_thread_info_get( + arena); + if (malloc_mutex_trylock(tsdn, &info->mtx)) { + /* + * Background thread may hold the mutex for a long period of + * time. We'd like to avoid the variance on application + * threads. So keep this non-blocking, and leave the work to a + * future epoch. + */ + return; + } + + if (info->state != background_thread_started) { + goto label_done; + } + if (malloc_mutex_trylock(tsdn, &decay->mtx)) { + goto label_done; + } + + ssize_t decay_time = atomic_load_zd(&decay->time_ms, ATOMIC_RELAXED); + if (decay_time <= 0) { + /* Purging is eagerly done or disabled currently. */ + goto label_done_unlock2; + } + uint64_t decay_interval_ns = nstime_ns(&decay->interval); + assert(decay_interval_ns > 0); + + nstime_t diff; + nstime_init(&diff, background_thread_wakeup_time_get(info)); + if (nstime_compare(&diff, &decay->epoch) <= 0) { + goto label_done_unlock2; + } + nstime_subtract(&diff, &decay->epoch); + if (nstime_ns(&diff) < BACKGROUND_THREAD_MIN_INTERVAL_NS) { + goto label_done_unlock2; + } + + if (npages_new > 0) { + size_t n_epoch = (size_t)(nstime_ns(&diff) / decay_interval_ns); + /* + * Compute how many new pages we would need to purge by the next + * wakeup, which is used to determine if we should signal the + * background thread. + */ + uint64_t npurge_new; + if (n_epoch >= SMOOTHSTEP_NSTEPS) { + npurge_new = npages_new; + } else { + uint64_t h_steps_max = h_steps[SMOOTHSTEP_NSTEPS - 1]; + assert(h_steps_max >= + h_steps[SMOOTHSTEP_NSTEPS - 1 - n_epoch]); + npurge_new = npages_new * (h_steps_max - + h_steps[SMOOTHSTEP_NSTEPS - 1 - n_epoch]); + npurge_new >>= SMOOTHSTEP_BFP; + } + info->npages_to_purge_new += npurge_new; + } + + bool should_signal; + if (info->npages_to_purge_new > BACKGROUND_THREAD_NPAGES_THRESHOLD) { + should_signal = true; + } else if (unlikely(background_thread_indefinite_sleep(info)) && + (extents_npages_get(&arena->extents_dirty) > 0 || + extents_npages_get(&arena->extents_muzzy) > 0 || + info->npages_to_purge_new > 0)) { + should_signal = true; + } else { + should_signal = false; + } + + if (should_signal) { + info->npages_to_purge_new = 0; + pthread_cond_signal(&info->cond); + } +label_done_unlock2: + malloc_mutex_unlock(tsdn, &decay->mtx); +label_done: + malloc_mutex_unlock(tsdn, &info->mtx); +} + +void +background_thread_prefork0(tsdn_t *tsdn) { + malloc_mutex_prefork(tsdn, &background_thread_lock); + background_thread_enabled_at_fork = background_thread_enabled(); +} + +void +background_thread_prefork1(tsdn_t *tsdn) { + for (unsigned i = 0; i < ncpus; i++) { + malloc_mutex_prefork(tsdn, &background_thread_info[i].mtx); + } +} + +void +background_thread_postfork_parent(tsdn_t *tsdn) { + for (unsigned i = 0; i < ncpus; i++) { + malloc_mutex_postfork_parent(tsdn, + &background_thread_info[i].mtx); + } + malloc_mutex_postfork_parent(tsdn, &background_thread_lock); +} + +void +background_thread_postfork_child(tsdn_t *tsdn) { + for (unsigned i = 0; i < ncpus; i++) { + malloc_mutex_postfork_child(tsdn, + &background_thread_info[i].mtx); + } + malloc_mutex_postfork_child(tsdn, &background_thread_lock); + if (!background_thread_enabled_at_fork) { + return; + } + + /* Clear background_thread state (reset to disabled for child). */ + malloc_mutex_lock(tsdn, &background_thread_lock); + n_background_threads = 0; + background_thread_enabled_set(tsdn, false); + for (unsigned i = 0; i < ncpus; i++) { + background_thread_info_t *info = &background_thread_info[i]; + malloc_mutex_lock(tsdn, &info->mtx); + info->state = background_thread_stopped; + int ret = pthread_cond_init(&info->cond, NULL); + assert(ret == 0); + background_thread_info_init(tsdn, info); + malloc_mutex_unlock(tsdn, &info->mtx); + } + malloc_mutex_unlock(tsdn, &background_thread_lock); +} + +bool +background_thread_stats_read(tsdn_t *tsdn, background_thread_stats_t *stats) { + assert(config_stats); + malloc_mutex_lock(tsdn, &background_thread_lock); + if (!background_thread_enabled()) { + malloc_mutex_unlock(tsdn, &background_thread_lock); + return true; + } + + stats->num_threads = n_background_threads; + uint64_t num_runs = 0; + nstime_init(&stats->run_interval, 0); + for (unsigned i = 0; i < ncpus; i++) { + background_thread_info_t *info = &background_thread_info[i]; + malloc_mutex_lock(tsdn, &info->mtx); + if (info->state != background_thread_stopped) { + num_runs += info->tot_n_runs; + nstime_add(&stats->run_interval, &info->tot_sleep_time); + } + malloc_mutex_unlock(tsdn, &info->mtx); + } + stats->num_runs = num_runs; + if (num_runs > 0) { + nstime_idivide(&stats->run_interval, num_runs); + } + malloc_mutex_unlock(tsdn, &background_thread_lock); + + return false; +} + +#undef BACKGROUND_THREAD_NPAGES_THRESHOLD +#undef BILLION +#undef BACKGROUND_THREAD_MIN_INTERVAL_NS + +/* + * When lazy lock is enabled, we need to make sure setting isthreaded before + * taking any background_thread locks. This is called early in ctl (instead of + * wait for the pthread_create calls to trigger) because the mutex is required + * before creating background threads. + */ +void +background_thread_ctl_init(tsdn_t *tsdn) { + malloc_mutex_assert_not_owner(tsdn, &background_thread_lock); +#ifdef JEMALLOC_PTHREAD_CREATE_WRAPPER + pthread_once(&once_control, pthread_create_wrapper_once); +#endif +} + +#endif /* defined(JEMALLOC_BACKGROUND_THREAD) */ + +bool +background_thread_boot0(void) { + if (!have_background_thread && opt_background_thread) { + malloc_printf("<jemalloc>: option background_thread currently " + "supports pthread only\n"); + return true; + } + +#ifdef JEMALLOC_PTHREAD_CREATE_WRAPPER + pthread_create_fptr = dlsym(RTLD_NEXT, "pthread_create"); + if (pthread_create_fptr == NULL) { + can_enable_background_thread = false; + if (config_lazy_lock || opt_background_thread) { + malloc_write("<jemalloc>: Error in dlsym(RTLD_NEXT, " + "\"pthread_create\")\n"); + abort(); + } + } else { + can_enable_background_thread = true; + } +#endif + return false; +} + +bool +background_thread_boot1(tsdn_t *tsdn) { +#ifdef JEMALLOC_BACKGROUND_THREAD + assert(have_background_thread); + assert(narenas_total_get() > 0); + + background_thread_enabled_set(tsdn, opt_background_thread); + if (malloc_mutex_init(&background_thread_lock, + "background_thread_global", + WITNESS_RANK_BACKGROUND_THREAD_GLOBAL, + malloc_mutex_rank_exclusive)) { + return true; + } + if (opt_background_thread) { + background_thread_ctl_init(tsdn); + } + + background_thread_info = (background_thread_info_t *)base_alloc(tsdn, + b0get(), ncpus * sizeof(background_thread_info_t), CACHELINE); + if (background_thread_info == NULL) { + return true; + } + + for (unsigned i = 0; i < ncpus; i++) { + background_thread_info_t *info = &background_thread_info[i]; + /* Thread mutex is rank_inclusive because of thread0. */ + if (malloc_mutex_init(&info->mtx, "background_thread", + WITNESS_RANK_BACKGROUND_THREAD, + malloc_mutex_address_ordered)) { + return true; + } + if (pthread_cond_init(&info->cond, NULL)) { + return true; + } + malloc_mutex_lock(tsdn, &info->mtx); + info->state = background_thread_stopped; + background_thread_info_init(tsdn, info); + malloc_mutex_unlock(tsdn, &info->mtx); + } +#endif + + return false; +} diff --git a/deps/jemalloc/src/extent_dss.c b/deps/jemalloc/src/extent_dss.c new file mode 100644 index 0000000000..e72da95870 --- /dev/null +++ b/deps/jemalloc/src/extent_dss.c @@ -0,0 +1,269 @@ +#define JEMALLOC_EXTENT_DSS_C_ +#include "jemalloc/internal/jemalloc_preamble.h" +#include "jemalloc/internal/jemalloc_internal_includes.h" + +#include "jemalloc/internal/assert.h" +#include "jemalloc/internal/extent_dss.h" +#include "jemalloc/internal/spin.h" + +/******************************************************************************/ +/* Data. */ + +const char *opt_dss = DSS_DEFAULT; + +const char *dss_prec_names[] = { + "disabled", + "primary", + "secondary", + "N/A" +}; + +/* + * Current dss precedence default, used when creating new arenas. NB: This is + * stored as unsigned rather than dss_prec_t because in principle there's no + * guarantee that sizeof(dss_prec_t) is the same as sizeof(unsigned), and we use + * atomic operations to synchronize the setting. + */ +static atomic_u_t dss_prec_default = ATOMIC_INIT( + (unsigned)DSS_PREC_DEFAULT); + +/* Base address of the DSS. */ +static void *dss_base; +/* Atomic boolean indicating whether a thread is currently extending DSS. */ +static atomic_b_t dss_extending; +/* Atomic boolean indicating whether the DSS is exhausted. */ +static atomic_b_t dss_exhausted; +/* Atomic current upper limit on DSS addresses. */ +static atomic_p_t dss_max; + +/******************************************************************************/ + +static void * +extent_dss_sbrk(intptr_t increment) { +#ifdef JEMALLOC_DSS + return sbrk(increment); +#else + not_implemented(); + return NULL; +#endif +} + +dss_prec_t +extent_dss_prec_get(void) { + dss_prec_t ret; + + if (!have_dss) { + return dss_prec_disabled; + } + ret = (dss_prec_t)atomic_load_u(&dss_prec_default, ATOMIC_ACQUIRE); + return ret; +} + +bool +extent_dss_prec_set(dss_prec_t dss_prec) { + if (!have_dss) { + return (dss_prec != dss_prec_disabled); + } + atomic_store_u(&dss_prec_default, (unsigned)dss_prec, ATOMIC_RELEASE); + return false; +} + +static void +extent_dss_extending_start(void) { + spin_t spinner = SPIN_INITIALIZER; + while (true) { + bool expected = false; + if (atomic_compare_exchange_weak_b(&dss_extending, &expected, + true, ATOMIC_ACQ_REL, ATOMIC_RELAXED)) { + break; + } + spin_adaptive(&spinner); + } +} + +static void +extent_dss_extending_finish(void) { + assert(atomic_load_b(&dss_extending, ATOMIC_RELAXED)); + + atomic_store_b(&dss_extending, false, ATOMIC_RELEASE); +} + +static void * +extent_dss_max_update(void *new_addr) { + /* + * Get the current end of the DSS as max_cur and assure that dss_max is + * up to date. + */ + void *max_cur = extent_dss_sbrk(0); + if (max_cur == (void *)-1) { + return NULL; + } + atomic_store_p(&dss_max, max_cur, ATOMIC_RELEASE); + /* Fixed new_addr can only be supported if it is at the edge of DSS. */ + if (new_addr != NULL && max_cur != new_addr) { + return NULL; + } + return max_cur; +} + +void * +extent_alloc_dss(tsdn_t *tsdn, arena_t *arena, void *new_addr, size_t size, + size_t alignment, bool *zero, bool *commit) { + extent_t *gap; + + cassert(have_dss); + assert(size > 0); + assert(alignment > 0); + + /* + * sbrk() uses a signed increment argument, so take care not to + * interpret a large allocation request as a negative increment. + */ + if ((intptr_t)size < 0) { + return NULL; + } + + gap = extent_alloc(tsdn, arena); + if (gap == NULL) { + return NULL; + } + + extent_dss_extending_start(); + if (!atomic_load_b(&dss_exhausted, ATOMIC_ACQUIRE)) { + /* + * The loop is necessary to recover from races with other + * threads that are using the DSS for something other than + * malloc. + */ + while (true) { + void *max_cur = extent_dss_max_update(new_addr); + if (max_cur == NULL) { + goto label_oom; + } + + /* + * Compute how much page-aligned gap space (if any) is + * necessary to satisfy alignment. This space can be + * recycled for later use. + */ + void *gap_addr_page = (void *)(PAGE_CEILING( + (uintptr_t)max_cur)); + void *ret = (void *)ALIGNMENT_CEILING( + (uintptr_t)gap_addr_page, alignment); + size_t gap_size_page = (uintptr_t)ret - + (uintptr_t)gap_addr_page; + if (gap_size_page != 0) { + extent_init(gap, arena, gap_addr_page, + gap_size_page, false, NSIZES, + arena_extent_sn_next(arena), + extent_state_active, false, true); + } + /* + * Compute the address just past the end of the desired + * allocation space. + */ + void *dss_next = (void *)((uintptr_t)ret + size); + if ((uintptr_t)ret < (uintptr_t)max_cur || + (uintptr_t)dss_next < (uintptr_t)max_cur) { + goto label_oom; /* Wrap-around. */ + } + /* Compute the increment, including subpage bytes. */ + void *gap_addr_subpage = max_cur; + size_t gap_size_subpage = (uintptr_t)ret - + (uintptr_t)gap_addr_subpage; + intptr_t incr = gap_size_subpage + size; + + assert((uintptr_t)max_cur + incr == (uintptr_t)ret + + size); + + /* Try to allocate. */ + void *dss_prev = extent_dss_sbrk(incr); + if (dss_prev == max_cur) { + /* Success. */ + atomic_store_p(&dss_max, dss_next, + ATOMIC_RELEASE); + extent_dss_extending_finish(); + + if (gap_size_page != 0) { + extent_dalloc_gap(tsdn, arena, gap); + } else { + extent_dalloc(tsdn, arena, gap); + } + if (!*commit) { + *commit = pages_decommit(ret, size); + } + if (*zero && *commit) { + extent_hooks_t *extent_hooks = + EXTENT_HOOKS_INITIALIZER; + extent_t extent; + + extent_init(&extent, arena, ret, size, + size, false, NSIZES, + extent_state_active, false, true); + if (extent_purge_forced_wrapper(tsdn, + arena, &extent_hooks, &extent, 0, + size)) { + memset(ret, 0, size); + } + } + return ret; + } + /* + * Failure, whether due to OOM or a race with a raw + * sbrk() call from outside the allocator. + */ + if (dss_prev == (void *)-1) { + /* OOM. */ + atomic_store_b(&dss_exhausted, true, + ATOMIC_RELEASE); + goto label_oom; + } + } + } +label_oom: + extent_dss_extending_finish(); + extent_dalloc(tsdn, arena, gap); + return NULL; +} + +static bool +extent_in_dss_helper(void *addr, void *max) { + return ((uintptr_t)addr >= (uintptr_t)dss_base && (uintptr_t)addr < + (uintptr_t)max); +} + +bool +extent_in_dss(void *addr) { + cassert(have_dss); + + return extent_in_dss_helper(addr, atomic_load_p(&dss_max, + ATOMIC_ACQUIRE)); +} + +bool +extent_dss_mergeable(void *addr_a, void *addr_b) { + void *max; + + cassert(have_dss); + + if ((uintptr_t)addr_a < (uintptr_t)dss_base && (uintptr_t)addr_b < + (uintptr_t)dss_base) { + return true; + } + + max = atomic_load_p(&dss_max, ATOMIC_ACQUIRE); + return (extent_in_dss_helper(addr_a, max) == + extent_in_dss_helper(addr_b, max)); +} + +void +extent_dss_boot(void) { + cassert(have_dss); + + dss_base = extent_dss_sbrk(0); + atomic_store_b(&dss_extending, false, ATOMIC_RELAXED); + atomic_store_b(&dss_exhausted, dss_base == (void *)-1, ATOMIC_RELAXED); + atomic_store_p(&dss_max, dss_base, ATOMIC_RELAXED); +} + +/******************************************************************************/ diff --git a/deps/jemalloc/src/extent_mmap.c b/deps/jemalloc/src/extent_mmap.c new file mode 100644 index 0000000000..8d607dc803 --- /dev/null +++ b/deps/jemalloc/src/extent_mmap.c @@ -0,0 +1,42 @@ +#define JEMALLOC_EXTENT_MMAP_C_ +#include "jemalloc/internal/jemalloc_preamble.h" +#include "jemalloc/internal/jemalloc_internal_includes.h" + +#include "jemalloc/internal/assert.h" +#include "jemalloc/internal/extent_mmap.h" + +/******************************************************************************/ +/* Data. */ + +bool opt_retain = +#ifdef JEMALLOC_RETAIN + true +#else + false +#endif + ; + +/******************************************************************************/ + +void * +extent_alloc_mmap(void *new_addr, size_t size, size_t alignment, bool *zero, + bool *commit) { + void *ret = pages_map(new_addr, size, ALIGNMENT_CEILING(alignment, + PAGE), commit); + if (ret == NULL) { + return NULL; + } + assert(ret != NULL); + if (*commit) { + *zero = true; + } + return ret; +} + +bool +extent_dalloc_mmap(void *addr, size_t size) { + if (!opt_retain) { + pages_unmap(addr, size); + } + return opt_retain; +} diff --git a/deps/jemalloc/src/hooks.c b/deps/jemalloc/src/hooks.c new file mode 100644 index 0000000000..6266ecd47f --- /dev/null +++ b/deps/jemalloc/src/hooks.c @@ -0,0 +1,12 @@ +#include "jemalloc/internal/jemalloc_preamble.h" + +/* + * The hooks are a little bit screwy -- they're not genuinely exported in the + * sense that we want them available to end-users, but we do want them visible + * from outside the generated library, so that we can use them in test code. + */ +JEMALLOC_EXPORT +void (*hooks_arena_new_hook)() = NULL; + +JEMALLOC_EXPORT +void (*hooks_libc_hook)() = NULL; diff --git a/deps/jemalloc/src/jemalloc_cpp.cpp b/deps/jemalloc/src/jemalloc_cpp.cpp new file mode 100644 index 0000000000..844ab398a7 --- /dev/null +++ b/deps/jemalloc/src/jemalloc_cpp.cpp @@ -0,0 +1,132 @@ +#include <mutex> +#include <new> + +#define JEMALLOC_CPP_CPP_ +#ifdef __cplusplus +extern "C" { +#endif + +#include "jemalloc/internal/jemalloc_preamble.h" +#include "jemalloc/internal/jemalloc_internal_includes.h" + +#ifdef __cplusplus +} +#endif + +// All operators in this file are exported. + +// Possibly alias hidden versions of malloc and sdallocx to avoid an extra plt +// thunk? +// +// extern __typeof (sdallocx) sdallocx_int +// __attribute ((alias ("sdallocx"), +// visibility ("hidden"))); +// +// ... but it needs to work with jemalloc namespaces. + +void *operator new(std::size_t size); +void *operator new[](std::size_t size); +void *operator new(std::size_t size, const std::nothrow_t &) noexcept; +void *operator new[](std::size_t size, const std::nothrow_t &) noexcept; +void operator delete(void *ptr) noexcept; +void operator delete[](void *ptr) noexcept; +void operator delete(void *ptr, const std::nothrow_t &) noexcept; +void operator delete[](void *ptr, const std::nothrow_t &) noexcept; + +#if __cpp_sized_deallocation >= 201309 +/* C++14's sized-delete operators. */ +void operator delete(void *ptr, std::size_t size) noexcept; +void operator delete[](void *ptr, std::size_t size) noexcept; +#endif + +template <bool IsNoExcept> +void * +newImpl(std::size_t size) noexcept(IsNoExcept) { + void *ptr = je_malloc(size); + if (likely(ptr != nullptr)) + return ptr; + + while (ptr == nullptr) { + std::new_handler handler; + // GCC-4.8 and clang 4.0 do not have std::get_new_handler. + { + static std::mutex mtx; + std::lock_guard<std::mutex> lock(mtx); + + handler = std::set_new_handler(nullptr); + std::set_new_handler(handler); + } + if (handler == nullptr) + break; + + try { + handler(); + } catch (const std::bad_alloc &) { + break; + } + + ptr = je_malloc(size); + } + + if (ptr == nullptr && !IsNoExcept) + std::__throw_bad_alloc(); + return ptr; +} + +void * +operator new(std::size_t size) { + return newImpl<false>(size); +} + +void * +operator new[](std::size_t size) { + return newImpl<false>(size); +} + +void * +operator new(std::size_t size, const std::nothrow_t &) noexcept { + return newImpl<true>(size); +} + +void * +operator new[](std::size_t size, const std::nothrow_t &) noexcept { + return newImpl<true>(size); +} + +void +operator delete(void *ptr) noexcept { + je_free(ptr); +} + +void +operator delete[](void *ptr) noexcept { + je_free(ptr); +} + +void +operator delete(void *ptr, const std::nothrow_t &) noexcept { + je_free(ptr); +} + +void operator delete[](void *ptr, const std::nothrow_t &) noexcept { + je_free(ptr); +} + +#if __cpp_sized_deallocation >= 201309 + +void +operator delete(void *ptr, std::size_t size) noexcept { + if (unlikely(ptr == nullptr)) { + return; + } + je_sdallocx(ptr, size, /*flags=*/0); +} + +void operator delete[](void *ptr, std::size_t size) noexcept { + if (unlikely(ptr == nullptr)) { + return; + } + je_sdallocx(ptr, size, /*flags=*/0); +} + +#endif // __cpp_sized_deallocation diff --git a/deps/jemalloc/src/large.c b/deps/jemalloc/src/large.c new file mode 100644 index 0000000000..27a2c67987 --- /dev/null +++ b/deps/jemalloc/src/large.c @@ -0,0 +1,371 @@ +#define JEMALLOC_LARGE_C_ +#include "jemalloc/internal/jemalloc_preamble.h" +#include "jemalloc/internal/jemalloc_internal_includes.h" + +#include "jemalloc/internal/assert.h" +#include "jemalloc/internal/extent_mmap.h" +#include "jemalloc/internal/mutex.h" +#include "jemalloc/internal/rtree.h" +#include "jemalloc/internal/util.h" + +/******************************************************************************/ + +void * +large_malloc(tsdn_t *tsdn, arena_t *arena, size_t usize, bool zero) { + assert(usize == sz_s2u(usize)); + + return large_palloc(tsdn, arena, usize, CACHELINE, zero); +} + +void * +large_palloc(tsdn_t *tsdn, arena_t *arena, size_t usize, size_t alignment, + bool zero) { + size_t ausize; + extent_t *extent; + bool is_zeroed; + UNUSED bool idump JEMALLOC_CC_SILENCE_INIT(false); + + assert(!tsdn_null(tsdn) || arena != NULL); + + ausize = sz_sa2u(usize, alignment); + if (unlikely(ausize == 0 || ausize > LARGE_MAXCLASS)) { + return NULL; + } + + if (config_fill && unlikely(opt_zero)) { + zero = true; + } + /* + * Copy zero into is_zeroed and pass the copy when allocating the + * extent, so that it is possible to make correct junk/zero fill + * decisions below, even if is_zeroed ends up true when zero is false. + */ + is_zeroed = zero; + if (likely(!tsdn_null(tsdn))) { + arena = arena_choose(tsdn_tsd(tsdn), arena); + } + if (unlikely(arena == NULL) || (extent = arena_extent_alloc_large(tsdn, + arena, usize, alignment, &is_zeroed)) == NULL) { + return NULL; + } + + /* See comments in arena_bin_slabs_full_insert(). */ + if (!arena_is_auto(arena)) { + /* Insert extent into large. */ + malloc_mutex_lock(tsdn, &arena->large_mtx); + extent_list_append(&arena->large, extent); + malloc_mutex_unlock(tsdn, &arena->large_mtx); + } + if (config_prof && arena_prof_accum(tsdn, arena, usize)) { + prof_idump(tsdn); + } + + if (zero) { + assert(is_zeroed); + } else if (config_fill && unlikely(opt_junk_alloc)) { + memset(extent_addr_get(extent), JEMALLOC_ALLOC_JUNK, + extent_usize_get(extent)); + } + + arena_decay_tick(tsdn, arena); + return extent_addr_get(extent); +} + +static void +large_dalloc_junk_impl(void *ptr, size_t size) { + memset(ptr, JEMALLOC_FREE_JUNK, size); +} +large_dalloc_junk_t *JET_MUTABLE large_dalloc_junk = large_dalloc_junk_impl; + +static void +large_dalloc_maybe_junk_impl(void *ptr, size_t size) { + if (config_fill && have_dss && unlikely(opt_junk_free)) { + /* + * Only bother junk filling if the extent isn't about to be + * unmapped. + */ + if (opt_retain || (have_dss && extent_in_dss(ptr))) { + large_dalloc_junk(ptr, size); + } + } +} +large_dalloc_maybe_junk_t *JET_MUTABLE large_dalloc_maybe_junk = + large_dalloc_maybe_junk_impl; + +static bool +large_ralloc_no_move_shrink(tsdn_t *tsdn, extent_t *extent, size_t usize) { + arena_t *arena = extent_arena_get(extent); + size_t oldusize = extent_usize_get(extent); + extent_hooks_t *extent_hooks = extent_hooks_get(arena); + size_t diff = extent_size_get(extent) - (usize + sz_large_pad); + + assert(oldusize > usize); + + if (extent_hooks->split == NULL) { + return true; + } + + /* Split excess pages. */ + if (diff != 0) { + extent_t *trail = extent_split_wrapper(tsdn, arena, + &extent_hooks, extent, usize + sz_large_pad, + sz_size2index(usize), false, diff, NSIZES, false); + if (trail == NULL) { + return true; + } + + if (config_fill && unlikely(opt_junk_free)) { + large_dalloc_maybe_junk(extent_addr_get(trail), + extent_size_get(trail)); + } + + arena_extents_dirty_dalloc(tsdn, arena, &extent_hooks, trail); + } + + arena_extent_ralloc_large_shrink(tsdn, arena, extent, oldusize); + + return false; +} + +static bool +large_ralloc_no_move_expand(tsdn_t *tsdn, extent_t *extent, size_t usize, + bool zero) { + arena_t *arena = extent_arena_get(extent); + size_t oldusize = extent_usize_get(extent); + extent_hooks_t *extent_hooks = extent_hooks_get(arena); + size_t trailsize = usize - oldusize; + + if (extent_hooks->merge == NULL) { + return true; + } + + if (config_fill && unlikely(opt_zero)) { + zero = true; + } + /* + * Copy zero into is_zeroed_trail and pass the copy when allocating the + * extent, so that it is possible to make correct junk/zero fill + * decisions below, even if is_zeroed_trail ends up true when zero is + * false. + */ + bool is_zeroed_trail = zero; + bool commit = true; + extent_t *trail; + bool new_mapping; + if ((trail = extents_alloc(tsdn, arena, &extent_hooks, + &arena->extents_dirty, extent_past_get(extent), trailsize, 0, + CACHELINE, false, NSIZES, &is_zeroed_trail, &commit)) != NULL + || (trail = extents_alloc(tsdn, arena, &extent_hooks, + &arena->extents_muzzy, extent_past_get(extent), trailsize, 0, + CACHELINE, false, NSIZES, &is_zeroed_trail, &commit)) != NULL) { + if (config_stats) { + new_mapping = false; + } + } else { + if ((trail = extent_alloc_wrapper(tsdn, arena, &extent_hooks, + extent_past_get(extent), trailsize, 0, CACHELINE, false, + NSIZES, &is_zeroed_trail, &commit)) == NULL) { + return true; + } + if (config_stats) { + new_mapping = true; + } + } + + if (extent_merge_wrapper(tsdn, arena, &extent_hooks, extent, trail)) { + extent_dalloc_wrapper(tsdn, arena, &extent_hooks, trail); + return true; + } + rtree_ctx_t rtree_ctx_fallback; + rtree_ctx_t *rtree_ctx = tsdn_rtree_ctx(tsdn, &rtree_ctx_fallback); + szind_t szind = sz_size2index(usize); + extent_szind_set(extent, szind); + rtree_szind_slab_update(tsdn, &extents_rtree, rtree_ctx, + (uintptr_t)extent_addr_get(extent), szind, false); + + if (config_stats && new_mapping) { + arena_stats_mapped_add(tsdn, &arena->stats, trailsize); + } + + if (zero) { + if (config_cache_oblivious) { + /* + * Zero the trailing bytes of the original allocation's + * last page, since they are in an indeterminate state. + * There will always be trailing bytes, because ptr's + * offset from the beginning of the extent is a multiple + * of CACHELINE in [0 .. PAGE). + */ + void *zbase = (void *) + ((uintptr_t)extent_addr_get(extent) + oldusize); + void *zpast = PAGE_ADDR2BASE((void *)((uintptr_t)zbase + + PAGE)); + size_t nzero = (uintptr_t)zpast - (uintptr_t)zbase; + assert(nzero > 0); + memset(zbase, 0, nzero); + } + assert(is_zeroed_trail); + } else if (config_fill && unlikely(opt_junk_alloc)) { + memset((void *)((uintptr_t)extent_addr_get(extent) + oldusize), + JEMALLOC_ALLOC_JUNK, usize - oldusize); + } + + arena_extent_ralloc_large_expand(tsdn, arena, extent, oldusize); + + return false; +} + +bool +large_ralloc_no_move(tsdn_t *tsdn, extent_t *extent, size_t usize_min, + size_t usize_max, bool zero) { + size_t oldusize = extent_usize_get(extent); + + /* The following should have been caught by callers. */ + assert(usize_min > 0 && usize_max <= LARGE_MAXCLASS); + /* Both allocation sizes must be large to avoid a move. */ + assert(oldusize >= LARGE_MINCLASS && usize_max >= LARGE_MINCLASS); + + if (usize_max > oldusize) { + /* Attempt to expand the allocation in-place. */ + if (!large_ralloc_no_move_expand(tsdn, extent, usize_max, + zero)) { + arena_decay_tick(tsdn, extent_arena_get(extent)); + return false; + } + /* Try again, this time with usize_min. */ + if (usize_min < usize_max && usize_min > oldusize && + large_ralloc_no_move_expand(tsdn, extent, usize_min, + zero)) { + arena_decay_tick(tsdn, extent_arena_get(extent)); + return false; + } + } + + /* + * Avoid moving the allocation if the existing extent size accommodates + * the new size. + */ + if (oldusize >= usize_min && oldusize <= usize_max) { + arena_decay_tick(tsdn, extent_arena_get(extent)); + return false; + } + + /* Attempt to shrink the allocation in-place. */ + if (oldusize > usize_max) { + if (!large_ralloc_no_move_shrink(tsdn, extent, usize_max)) { + arena_decay_tick(tsdn, extent_arena_get(extent)); + return false; + } + } + return true; +} + +static void * +large_ralloc_move_helper(tsdn_t *tsdn, arena_t *arena, size_t usize, + size_t alignment, bool zero) { + if (alignment <= CACHELINE) { + return large_malloc(tsdn, arena, usize, zero); + } + return large_palloc(tsdn, arena, usize, alignment, zero); +} + +void * +large_ralloc(tsdn_t *tsdn, arena_t *arena, extent_t *extent, size_t usize, + size_t alignment, bool zero, tcache_t *tcache) { + size_t oldusize = extent_usize_get(extent); + + /* The following should have been caught by callers. */ + assert(usize > 0 && usize <= LARGE_MAXCLASS); + /* Both allocation sizes must be large to avoid a move. */ + assert(oldusize >= LARGE_MINCLASS && usize >= LARGE_MINCLASS); + + /* Try to avoid moving the allocation. */ + if (!large_ralloc_no_move(tsdn, extent, usize, usize, zero)) { + return extent_addr_get(extent); + } + + /* + * usize and old size are different enough that we need to use a + * different size class. In that case, fall back to allocating new + * space and copying. + */ + void *ret = large_ralloc_move_helper(tsdn, arena, usize, alignment, + zero); + if (ret == NULL) { + return NULL; + } + + size_t copysize = (usize < oldusize) ? usize : oldusize; + memcpy(ret, extent_addr_get(extent), copysize); + isdalloct(tsdn, extent_addr_get(extent), oldusize, tcache, NULL, true); + return ret; +} + +/* + * junked_locked indicates whether the extent's data have been junk-filled, and + * whether the arena's large_mtx is currently held. + */ +static void +large_dalloc_prep_impl(tsdn_t *tsdn, arena_t *arena, extent_t *extent, + bool junked_locked) { + if (!junked_locked) { + /* See comments in arena_bin_slabs_full_insert(). */ + if (!arena_is_auto(arena)) { + malloc_mutex_lock(tsdn, &arena->large_mtx); + extent_list_remove(&arena->large, extent); + malloc_mutex_unlock(tsdn, &arena->large_mtx); + } + large_dalloc_maybe_junk(extent_addr_get(extent), + extent_usize_get(extent)); + } else { + malloc_mutex_assert_owner(tsdn, &arena->large_mtx); + if (!arena_is_auto(arena)) { + extent_list_remove(&arena->large, extent); + } + } + arena_extent_dalloc_large_prep(tsdn, arena, extent); +} + +static void +large_dalloc_finish_impl(tsdn_t *tsdn, arena_t *arena, extent_t *extent) { + extent_hooks_t *extent_hooks = EXTENT_HOOKS_INITIALIZER; + arena_extents_dirty_dalloc(tsdn, arena, &extent_hooks, extent); +} + +void +large_dalloc_prep_junked_locked(tsdn_t *tsdn, extent_t *extent) { + large_dalloc_prep_impl(tsdn, extent_arena_get(extent), extent, true); +} + +void +large_dalloc_finish(tsdn_t *tsdn, extent_t *extent) { + large_dalloc_finish_impl(tsdn, extent_arena_get(extent), extent); +} + +void +large_dalloc(tsdn_t *tsdn, extent_t *extent) { + arena_t *arena = extent_arena_get(extent); + large_dalloc_prep_impl(tsdn, arena, extent, false); + large_dalloc_finish_impl(tsdn, arena, extent); + arena_decay_tick(tsdn, arena); +} + +size_t +large_salloc(tsdn_t *tsdn, const extent_t *extent) { + return extent_usize_get(extent); +} + +prof_tctx_t * +large_prof_tctx_get(tsdn_t *tsdn, const extent_t *extent) { + return extent_prof_tctx_get(extent); +} + +void +large_prof_tctx_set(tsdn_t *tsdn, extent_t *extent, prof_tctx_t *tctx) { + extent_prof_tctx_set(extent, tctx); +} + +void +large_prof_tctx_reset(tsdn_t *tsdn, extent_t *extent) { + large_prof_tctx_set(tsdn, extent, (prof_tctx_t *)(uintptr_t)1U); +} diff --git a/deps/jemalloc/src/malloc_io.c b/deps/jemalloc/src/malloc_io.c new file mode 100644 index 0000000000..6b99afcd3f --- /dev/null +++ b/deps/jemalloc/src/malloc_io.c @@ -0,0 +1,689 @@ +#define JEMALLOC_MALLOC_IO_C_ +#include "jemalloc/internal/jemalloc_preamble.h" +#include "jemalloc/internal/jemalloc_internal_includes.h" + +#include "jemalloc/internal/malloc_io.h" +#include "jemalloc/internal/util.h" + +#ifdef assert +# undef assert +#endif +#ifdef not_reached +# undef not_reached +#endif +#ifdef not_implemented +# undef not_implemented +#endif +#ifdef assert_not_implemented +# undef assert_not_implemented +#endif + +/* + * Define simple versions of assertion macros that won't recurse in case + * of assertion failures in malloc_*printf(). + */ +#define assert(e) do { \ + if (config_debug && !(e)) { \ + malloc_write("<jemalloc>: Failed assertion\n"); \ + abort(); \ + } \ +} while (0) + +#define not_reached() do { \ + if (config_debug) { \ + malloc_write("<jemalloc>: Unreachable code reached\n"); \ + abort(); \ + } \ + unreachable(); \ +} while (0) + +#define not_implemented() do { \ + if (config_debug) { \ + malloc_write("<jemalloc>: Not implemented\n"); \ + abort(); \ + } \ +} while (0) + +#define assert_not_implemented(e) do { \ + if (unlikely(config_debug && !(e))) { \ + not_implemented(); \ + } \ +} while (0) + +/******************************************************************************/ +/* Function prototypes for non-inline static functions. */ + +static void wrtmessage(void *cbopaque, const char *s); +#define U2S_BUFSIZE ((1U << (LG_SIZEOF_INTMAX_T + 3)) + 1) +static char *u2s(uintmax_t x, unsigned base, bool uppercase, char *s, + size_t *slen_p); +#define D2S_BUFSIZE (1 + U2S_BUFSIZE) +static char *d2s(intmax_t x, char sign, char *s, size_t *slen_p); +#define O2S_BUFSIZE (1 + U2S_BUFSIZE) +static char *o2s(uintmax_t x, bool alt_form, char *s, size_t *slen_p); +#define X2S_BUFSIZE (2 + U2S_BUFSIZE) +static char *x2s(uintmax_t x, bool alt_form, bool uppercase, char *s, + size_t *slen_p); + +/******************************************************************************/ + +/* malloc_message() setup. */ +static void +wrtmessage(void *cbopaque, const char *s) { +#if defined(JEMALLOC_USE_SYSCALL) && defined(SYS_write) + /* + * Use syscall(2) rather than write(2) when possible in order to avoid + * the possibility of memory allocation within libc. This is necessary + * on FreeBSD; most operating systems do not have this problem though. + * + * syscall() returns long or int, depending on platform, so capture the + * unused result in the widest plausible type to avoid compiler + * warnings. + */ + UNUSED long result = syscall(SYS_write, STDERR_FILENO, s, strlen(s)); +#else + UNUSED ssize_t result = write(STDERR_FILENO, s, strlen(s)); +#endif +} + +JEMALLOC_EXPORT void (*je_malloc_message)(void *, const char *s); + +/* + * Wrapper around malloc_message() that avoids the need for + * je_malloc_message(...) throughout the code. + */ +void +malloc_write(const char *s) { + if (je_malloc_message != NULL) { + je_malloc_message(NULL, s); + } else { + wrtmessage(NULL, s); + } +} + +/* + * glibc provides a non-standard strerror_r() when _GNU_SOURCE is defined, so + * provide a wrapper. + */ +int +buferror(int err, char *buf, size_t buflen) { +#ifdef _WIN32 + FormatMessageA(FORMAT_MESSAGE_FROM_SYSTEM, NULL, err, 0, + (LPSTR)buf, (DWORD)buflen, NULL); + return 0; +#elif defined(__GLIBC__) && defined(_GNU_SOURCE) + char *b = strerror_r(err, buf, buflen); + if (b != buf) { + strncpy(buf, b, buflen); + buf[buflen-1] = '\0'; + } + return 0; +#else + return strerror_r(err, buf, buflen); +#endif +} + +uintmax_t +malloc_strtoumax(const char *restrict nptr, char **restrict endptr, int base) { + uintmax_t ret, digit; + unsigned b; + bool neg; + const char *p, *ns; + + p = nptr; + if (base < 0 || base == 1 || base > 36) { + ns = p; + set_errno(EINVAL); + ret = UINTMAX_MAX; + goto label_return; + } + b = base; + + /* Swallow leading whitespace and get sign, if any. */ + neg = false; + while (true) { + switch (*p) { + case '\t': case '\n': case '\v': case '\f': case '\r': case ' ': + p++; + break; + case '-': + neg = true; + /* Fall through. */ + case '+': + p++; + /* Fall through. */ + default: + goto label_prefix; + } + } + + /* Get prefix, if any. */ + label_prefix: + /* + * Note where the first non-whitespace/sign character is so that it is + * possible to tell whether any digits are consumed (e.g., " 0" vs. + * " -x"). + */ + ns = p; + if (*p == '0') { + switch (p[1]) { + case '0': case '1': case '2': case '3': case '4': case '5': + case '6': case '7': + if (b == 0) { + b = 8; + } + if (b == 8) { + p++; + } + break; + case 'X': case 'x': + switch (p[2]) { + case '0': case '1': case '2': case '3': case '4': + case '5': case '6': case '7': case '8': case '9': + case 'A': case 'B': case 'C': case 'D': case 'E': + case 'F': + case 'a': case 'b': case 'c': case 'd': case 'e': + case 'f': + if (b == 0) { + b = 16; + } + if (b == 16) { + p += 2; + } + break; + default: + break; + } + break; + default: + p++; + ret = 0; + goto label_return; + } + } + if (b == 0) { + b = 10; + } + + /* Convert. */ + ret = 0; + while ((*p >= '0' && *p <= '9' && (digit = *p - '0') < b) + || (*p >= 'A' && *p <= 'Z' && (digit = 10 + *p - 'A') < b) + || (*p >= 'a' && *p <= 'z' && (digit = 10 + *p - 'a') < b)) { + uintmax_t pret = ret; + ret *= b; + ret += digit; + if (ret < pret) { + /* Overflow. */ + set_errno(ERANGE); + ret = UINTMAX_MAX; + goto label_return; + } + p++; + } + if (neg) { + ret = (uintmax_t)(-((intmax_t)ret)); + } + + if (p == ns) { + /* No conversion performed. */ + set_errno(EINVAL); + ret = UINTMAX_MAX; + goto label_return; + } + +label_return: + if (endptr != NULL) { + if (p == ns) { + /* No characters were converted. */ + *endptr = (char *)nptr; + } else { + *endptr = (char *)p; + } + } + return ret; +} + +static char * +u2s(uintmax_t x, unsigned base, bool uppercase, char *s, size_t *slen_p) { + unsigned i; + + i = U2S_BUFSIZE - 1; + s[i] = '\0'; + switch (base) { + case 10: + do { + i--; + s[i] = "0123456789"[x % (uint64_t)10]; + x /= (uint64_t)10; + } while (x > 0); + break; + case 16: { + const char *digits = (uppercase) + ? "0123456789ABCDEF" + : "0123456789abcdef"; + + do { + i--; + s[i] = digits[x & 0xf]; + x >>= 4; + } while (x > 0); + break; + } default: { + const char *digits = (uppercase) + ? "0123456789ABCDEFGHIJKLMNOPQRSTUVWXYZ" + : "0123456789abcdefghijklmnopqrstuvwxyz"; + + assert(base >= 2 && base <= 36); + do { + i--; + s[i] = digits[x % (uint64_t)base]; + x /= (uint64_t)base; + } while (x > 0); + }} + + *slen_p = U2S_BUFSIZE - 1 - i; + return &s[i]; +} + +static char * +d2s(intmax_t x, char sign, char *s, size_t *slen_p) { + bool neg; + + if ((neg = (x < 0))) { + x = -x; + } + s = u2s(x, 10, false, s, slen_p); + if (neg) { + sign = '-'; + } + switch (sign) { + case '-': + if (!neg) { + break; + } + /* Fall through. */ + case ' ': + case '+': + s--; + (*slen_p)++; + *s = sign; + break; + default: not_reached(); + } + return s; +} + +static char * +o2s(uintmax_t x, bool alt_form, char *s, size_t *slen_p) { + s = u2s(x, 8, false, s, slen_p); + if (alt_form && *s != '0') { + s--; + (*slen_p)++; + *s = '0'; + } + return s; +} + +static char * +x2s(uintmax_t x, bool alt_form, bool uppercase, char *s, size_t *slen_p) { + s = u2s(x, 16, uppercase, s, slen_p); + if (alt_form) { + s -= 2; + (*slen_p) += 2; + memcpy(s, uppercase ? "0X" : "0x", 2); + } + return s; +} + +size_t +malloc_vsnprintf(char *str, size_t size, const char *format, va_list ap) { + size_t i; + const char *f; + +#define APPEND_C(c) do { \ + if (i < size) { \ + str[i] = (c); \ + } \ + i++; \ +} while (0) +#define APPEND_S(s, slen) do { \ + if (i < size) { \ + size_t cpylen = (slen <= size - i) ? slen : size - i; \ + memcpy(&str[i], s, cpylen); \ + } \ + i += slen; \ +} while (0) +#define APPEND_PADDED_S(s, slen, width, left_justify) do { \ + /* Left padding. */ \ + size_t pad_len = (width == -1) ? 0 : ((slen < (size_t)width) ? \ + (size_t)width - slen : 0); \ + if (!left_justify && pad_len != 0) { \ + size_t j; \ + for (j = 0; j < pad_len; j++) { \ + APPEND_C(' '); \ + } \ + } \ + /* Value. */ \ + APPEND_S(s, slen); \ + /* Right padding. */ \ + if (left_justify && pad_len != 0) { \ + size_t j; \ + for (j = 0; j < pad_len; j++) { \ + APPEND_C(' '); \ + } \ + } \ +} while (0) +#define GET_ARG_NUMERIC(val, len) do { \ + switch (len) { \ + case '?': \ + val = va_arg(ap, int); \ + break; \ + case '?' | 0x80: \ + val = va_arg(ap, unsigned int); \ + break; \ + case 'l': \ + val = va_arg(ap, long); \ + break; \ + case 'l' | 0x80: \ + val = va_arg(ap, unsigned long); \ + break; \ + case 'q': \ + val = va_arg(ap, long long); \ + break; \ + case 'q' | 0x80: \ + val = va_arg(ap, unsigned long long); \ + break; \ + case 'j': \ + val = va_arg(ap, intmax_t); \ + break; \ + case 'j' | 0x80: \ + val = va_arg(ap, uintmax_t); \ + break; \ + case 't': \ + val = va_arg(ap, ptrdiff_t); \ + break; \ + case 'z': \ + val = va_arg(ap, ssize_t); \ + break; \ + case 'z' | 0x80: \ + val = va_arg(ap, size_t); \ + break; \ + case 'p': /* Synthetic; used for %p. */ \ + val = va_arg(ap, uintptr_t); \ + break; \ + default: \ + not_reached(); \ + val = 0; \ + } \ +} while (0) + + i = 0; + f = format; + while (true) { + switch (*f) { + case '\0': goto label_out; + case '%': { + bool alt_form = false; + bool left_justify = false; + bool plus_space = false; + bool plus_plus = false; + int prec = -1; + int width = -1; + unsigned char len = '?'; + char *s; + size_t slen; + + f++; + /* Flags. */ + while (true) { + switch (*f) { + case '#': + assert(!alt_form); + alt_form = true; + break; + case '-': + assert(!left_justify); + left_justify = true; + break; + case ' ': + assert(!plus_space); + plus_space = true; + break; + case '+': + assert(!plus_plus); + plus_plus = true; + break; + default: goto label_width; + } + f++; + } + /* Width. */ + label_width: + switch (*f) { + case '*': + width = va_arg(ap, int); + f++; + if (width < 0) { + left_justify = true; + width = -width; + } + break; + case '0': case '1': case '2': case '3': case '4': + case '5': case '6': case '7': case '8': case '9': { + uintmax_t uwidth; + set_errno(0); + uwidth = malloc_strtoumax(f, (char **)&f, 10); + assert(uwidth != UINTMAX_MAX || get_errno() != + ERANGE); + width = (int)uwidth; + break; + } default: + break; + } + /* Width/precision separator. */ + if (*f == '.') { + f++; + } else { + goto label_length; + } + /* Precision. */ + switch (*f) { + case '*': + prec = va_arg(ap, int); + f++; + break; + case '0': case '1': case '2': case '3': case '4': + case '5': case '6': case '7': case '8': case '9': { + uintmax_t uprec; + set_errno(0); + uprec = malloc_strtoumax(f, (char **)&f, 10); + assert(uprec != UINTMAX_MAX || get_errno() != + ERANGE); + prec = (int)uprec; + break; + } + default: break; + } + /* Length. */ + label_length: + switch (*f) { + case 'l': + f++; + if (*f == 'l') { + len = 'q'; + f++; + } else { + len = 'l'; + } + break; + case 'q': case 'j': case 't': case 'z': + len = *f; + f++; + break; + default: break; + } + /* Conversion specifier. */ + switch (*f) { + case '%': + /* %% */ + APPEND_C(*f); + f++; + break; + case 'd': case 'i': { + intmax_t val JEMALLOC_CC_SILENCE_INIT(0); + char buf[D2S_BUFSIZE]; + + GET_ARG_NUMERIC(val, len); + s = d2s(val, (plus_plus ? '+' : (plus_space ? + ' ' : '-')), buf, &slen); + APPEND_PADDED_S(s, slen, width, left_justify); + f++; + break; + } case 'o': { + uintmax_t val JEMALLOC_CC_SILENCE_INIT(0); + char buf[O2S_BUFSIZE]; + + GET_ARG_NUMERIC(val, len | 0x80); + s = o2s(val, alt_form, buf, &slen); + APPEND_PADDED_S(s, slen, width, left_justify); + f++; + break; + } case 'u': { + uintmax_t val JEMALLOC_CC_SILENCE_INIT(0); + char buf[U2S_BUFSIZE]; + + GET_ARG_NUMERIC(val, len | 0x80); + s = u2s(val, 10, false, buf, &slen); + APPEND_PADDED_S(s, slen, width, left_justify); + f++; + break; + } case 'x': case 'X': { + uintmax_t val JEMALLOC_CC_SILENCE_INIT(0); + char buf[X2S_BUFSIZE]; + + GET_ARG_NUMERIC(val, len | 0x80); + s = x2s(val, alt_form, *f == 'X', buf, &slen); + APPEND_PADDED_S(s, slen, width, left_justify); + f++; + break; + } case 'c': { + unsigned char val; + char buf[2]; + + assert(len == '?' || len == 'l'); + assert_not_implemented(len != 'l'); + val = va_arg(ap, int); + buf[0] = val; + buf[1] = '\0'; + APPEND_PADDED_S(buf, 1, width, left_justify); + f++; + break; + } case 's': + assert(len == '?' || len == 'l'); + assert_not_implemented(len != 'l'); + s = va_arg(ap, char *); + slen = (prec < 0) ? strlen(s) : (size_t)prec; + APPEND_PADDED_S(s, slen, width, left_justify); + f++; + break; + case 'p': { + uintmax_t val; + char buf[X2S_BUFSIZE]; + + GET_ARG_NUMERIC(val, 'p'); + s = x2s(val, true, false, buf, &slen); + APPEND_PADDED_S(s, slen, width, left_justify); + f++; + break; + } default: not_reached(); + } + break; + } default: { + APPEND_C(*f); + f++; + break; + }} + } + label_out: + if (i < size) { + str[i] = '\0'; + } else { + str[size - 1] = '\0'; + } + +#undef APPEND_C +#undef APPEND_S +#undef APPEND_PADDED_S +#undef GET_ARG_NUMERIC + return i; +} + +JEMALLOC_FORMAT_PRINTF(3, 4) +size_t +malloc_snprintf(char *str, size_t size, const char *format, ...) { + size_t ret; + va_list ap; + + va_start(ap, format); + ret = malloc_vsnprintf(str, size, format, ap); + va_end(ap); + + return ret; +} + +void +malloc_vcprintf(void (*write_cb)(void *, const char *), void *cbopaque, + const char *format, va_list ap) { + char buf[MALLOC_PRINTF_BUFSIZE]; + + if (write_cb == NULL) { + /* + * The caller did not provide an alternate write_cb callback + * function, so use the default one. malloc_write() is an + * inline function, so use malloc_message() directly here. + */ + write_cb = (je_malloc_message != NULL) ? je_malloc_message : + wrtmessage; + cbopaque = NULL; + } + + malloc_vsnprintf(buf, sizeof(buf), format, ap); + write_cb(cbopaque, buf); +} + +/* + * Print to a callback function in such a way as to (hopefully) avoid memory + * allocation. + */ +JEMALLOC_FORMAT_PRINTF(3, 4) +void +malloc_cprintf(void (*write_cb)(void *, const char *), void *cbopaque, + const char *format, ...) { + va_list ap; + + va_start(ap, format); + malloc_vcprintf(write_cb, cbopaque, format, ap); + va_end(ap); +} + +/* Print to stderr in such a way as to avoid memory allocation. */ +JEMALLOC_FORMAT_PRINTF(1, 2) +void +malloc_printf(const char *format, ...) { + va_list ap; + + va_start(ap, format); + malloc_vcprintf(NULL, NULL, format, ap); + va_end(ap); +} + +/* + * Restore normal assertion macros, in order to make it possible to compile all + * C files as a single concatenation. + */ +#undef assert +#undef not_reached +#undef not_implemented +#undef assert_not_implemented +#include "jemalloc/internal/assert.h" diff --git a/deps/jemalloc/src/mutex_pool.c b/deps/jemalloc/src/mutex_pool.c new file mode 100644 index 0000000000..f24d10e44a --- /dev/null +++ b/deps/jemalloc/src/mutex_pool.c @@ -0,0 +1,18 @@ +#define JEMALLOC_MUTEX_POOL_C_ + +#include "jemalloc/internal/jemalloc_preamble.h" +#include "jemalloc/internal/jemalloc_internal_includes.h" + +#include "jemalloc/internal/mutex.h" +#include "jemalloc/internal/mutex_pool.h" + +bool +mutex_pool_init(mutex_pool_t *pool, const char *name, witness_rank_t rank) { + for (int i = 0; i < MUTEX_POOL_SIZE; ++i) { + if (malloc_mutex_init(&pool->mutexes[i], name, rank, + malloc_mutex_address_ordered)) { + return true; + } + } + return false; +} diff --git a/deps/jemalloc/src/nstime.c b/deps/jemalloc/src/nstime.c new file mode 100644 index 0000000000..71db353965 --- /dev/null +++ b/deps/jemalloc/src/nstime.c @@ -0,0 +1,170 @@ +#include "jemalloc/internal/jemalloc_preamble.h" +#include "jemalloc/internal/jemalloc_internal_includes.h" + +#include "jemalloc/internal/nstime.h" + +#include "jemalloc/internal/assert.h" + +#define BILLION UINT64_C(1000000000) +#define MILLION UINT64_C(1000000) + +void +nstime_init(nstime_t *time, uint64_t ns) { + time->ns = ns; +} + +void +nstime_init2(nstime_t *time, uint64_t sec, uint64_t nsec) { + time->ns = sec * BILLION + nsec; +} + +uint64_t +nstime_ns(const nstime_t *time) { + return time->ns; +} + +uint64_t +nstime_msec(const nstime_t *time) { + return time->ns / MILLION; +} + +uint64_t +nstime_sec(const nstime_t *time) { + return time->ns / BILLION; +} + +uint64_t +nstime_nsec(const nstime_t *time) { + return time->ns % BILLION; +} + +void +nstime_copy(nstime_t *time, const nstime_t *source) { + *time = *source; +} + +int +nstime_compare(const nstime_t *a, const nstime_t *b) { + return (a->ns > b->ns) - (a->ns < b->ns); +} + +void +nstime_add(nstime_t *time, const nstime_t *addend) { + assert(UINT64_MAX - time->ns >= addend->ns); + + time->ns += addend->ns; +} + +void +nstime_iadd(nstime_t *time, uint64_t addend) { + assert(UINT64_MAX - time->ns >= addend); + + time->ns += addend; +} + +void +nstime_subtract(nstime_t *time, const nstime_t *subtrahend) { + assert(nstime_compare(time, subtrahend) >= 0); + + time->ns -= subtrahend->ns; +} + +void +nstime_isubtract(nstime_t *time, uint64_t subtrahend) { + assert(time->ns >= subtrahend); + + time->ns -= subtrahend; +} + +void +nstime_imultiply(nstime_t *time, uint64_t multiplier) { + assert((((time->ns | multiplier) & (UINT64_MAX << (sizeof(uint64_t) << + 2))) == 0) || ((time->ns * multiplier) / multiplier == time->ns)); + + time->ns *= multiplier; +} + +void +nstime_idivide(nstime_t *time, uint64_t divisor) { + assert(divisor != 0); + + time->ns /= divisor; +} + +uint64_t +nstime_divide(const nstime_t *time, const nstime_t *divisor) { + assert(divisor->ns != 0); + + return time->ns / divisor->ns; +} + +#ifdef _WIN32 +# define NSTIME_MONOTONIC true +static void +nstime_get(nstime_t *time) { + FILETIME ft; + uint64_t ticks_100ns; + + GetSystemTimeAsFileTime(&ft); + ticks_100ns = (((uint64_t)ft.dwHighDateTime) << 32) | ft.dwLowDateTime; + + nstime_init(time, ticks_100ns * 100); +} +#elif defined(JEMALLOC_HAVE_CLOCK_MONOTONIC_COARSE) +# define NSTIME_MONOTONIC true +static void +nstime_get(nstime_t *time) { + struct timespec ts; + + clock_gettime(CLOCK_MONOTONIC_COARSE, &ts); + nstime_init2(time, ts.tv_sec, ts.tv_nsec); +} +#elif defined(JEMALLOC_HAVE_CLOCK_MONOTONIC) +# define NSTIME_MONOTONIC true +static void +nstime_get(nstime_t *time) { + struct timespec ts; + + clock_gettime(CLOCK_MONOTONIC, &ts); + nstime_init2(time, ts.tv_sec, ts.tv_nsec); +} +#elif defined(JEMALLOC_HAVE_MACH_ABSOLUTE_TIME) +# define NSTIME_MONOTONIC true +static void +nstime_get(nstime_t *time) { + nstime_init(time, mach_absolute_time()); +} +#else +# define NSTIME_MONOTONIC false +static void +nstime_get(nstime_t *time) { + struct timeval tv; + + gettimeofday(&tv, NULL); + nstime_init2(time, tv.tv_sec, tv.tv_usec * 1000); +} +#endif + +static bool +nstime_monotonic_impl(void) { + return NSTIME_MONOTONIC; +#undef NSTIME_MONOTONIC +} +nstime_monotonic_t *JET_MUTABLE nstime_monotonic = nstime_monotonic_impl; + +static bool +nstime_update_impl(nstime_t *time) { + nstime_t old_time; + + nstime_copy(&old_time, time); + nstime_get(time); + + /* Handle non-monotonic clocks. */ + if (unlikely(nstime_compare(&old_time, time) > 0)) { + nstime_copy(time, &old_time); + return true; + } + + return false; +} +nstime_update_t *JET_MUTABLE nstime_update = nstime_update_impl; diff --git a/deps/jemalloc/src/pages.c b/deps/jemalloc/src/pages.c new file mode 100644 index 0000000000..6f2ba5669b --- /dev/null +++ b/deps/jemalloc/src/pages.c @@ -0,0 +1,423 @@ +#define JEMALLOC_PAGES_C_ +#include "jemalloc/internal/jemalloc_preamble.h" + +#include "jemalloc/internal/pages.h" + +#include "jemalloc/internal/jemalloc_internal_includes.h" + +#include "jemalloc/internal/assert.h" +#include "jemalloc/internal/malloc_io.h" + +#ifdef JEMALLOC_SYSCTL_VM_OVERCOMMIT +#include <sys/sysctl.h> +#endif + +/******************************************************************************/ +/* Data. */ + +/* Actual operating system page size, detected during bootstrap, <= PAGE. */ +static size_t os_page; + +#ifndef _WIN32 +# define PAGES_PROT_COMMIT (PROT_READ | PROT_WRITE) +# define PAGES_PROT_DECOMMIT (PROT_NONE) +static int mmap_flags; +#endif +static bool os_overcommits; + +/******************************************************************************/ +/* + * Function prototypes for static functions that are referenced prior to + * definition. + */ + +static void os_pages_unmap(void *addr, size_t size); + +/******************************************************************************/ + +static void * +os_pages_map(void *addr, size_t size, size_t alignment, bool *commit) { + assert(ALIGNMENT_ADDR2BASE(addr, os_page) == addr); + assert(ALIGNMENT_CEILING(size, os_page) == size); + assert(size != 0); + + if (os_overcommits) { + *commit = true; + } + + void *ret; +#ifdef _WIN32 + /* + * If VirtualAlloc can't allocate at the given address when one is + * given, it fails and returns NULL. + */ + ret = VirtualAlloc(addr, size, MEM_RESERVE | (*commit ? MEM_COMMIT : 0), + PAGE_READWRITE); +#else + /* + * We don't use MAP_FIXED here, because it can cause the *replacement* + * of existing mappings, and we only want to create new mappings. + */ + { + int prot = *commit ? PAGES_PROT_COMMIT : PAGES_PROT_DECOMMIT; + + ret = mmap(addr, size, prot, mmap_flags, -1, 0); + } + assert(ret != NULL); + + if (ret == MAP_FAILED) { + ret = NULL; + } else if (addr != NULL && ret != addr) { + /* + * We succeeded in mapping memory, but not in the right place. + */ + os_pages_unmap(ret, size); + ret = NULL; + } +#endif + assert(ret == NULL || (addr == NULL && ret != addr) || (addr != NULL && + ret == addr)); + return ret; +} + +static void * +os_pages_trim(void *addr, size_t alloc_size, size_t leadsize, size_t size, + bool *commit) { + void *ret = (void *)((uintptr_t)addr + leadsize); + + assert(alloc_size >= leadsize + size); +#ifdef _WIN32 + os_pages_unmap(addr, alloc_size); + void *new_addr = os_pages_map(ret, size, PAGE, commit); + if (new_addr == ret) { + return ret; + } + if (new_addr != NULL) { + os_pages_unmap(new_addr, size); + } + return NULL; +#else + size_t trailsize = alloc_size - leadsize - size; + + if (leadsize != 0) { + os_pages_unmap(addr, leadsize); + } + if (trailsize != 0) { + os_pages_unmap((void *)((uintptr_t)ret + size), trailsize); + } + return ret; +#endif +} + +static void +os_pages_unmap(void *addr, size_t size) { + assert(ALIGNMENT_ADDR2BASE(addr, os_page) == addr); + assert(ALIGNMENT_CEILING(size, os_page) == size); + +#ifdef _WIN32 + if (VirtualFree(addr, 0, MEM_RELEASE) == 0) +#else + if (munmap(addr, size) == -1) +#endif + { + char buf[BUFERROR_BUF]; + + buferror(get_errno(), buf, sizeof(buf)); + malloc_printf("<jemalloc>: Error in " +#ifdef _WIN32 + "VirtualFree" +#else + "munmap" +#endif + "(): %s\n", buf); + if (opt_abort) { + abort(); + } + } +} + +static void * +pages_map_slow(size_t size, size_t alignment, bool *commit) { + size_t alloc_size = size + alignment - os_page; + /* Beware size_t wrap-around. */ + if (alloc_size < size) { + return NULL; + } + + void *ret; + do { + void *pages = os_pages_map(NULL, alloc_size, alignment, commit); + if (pages == NULL) { + return NULL; + } + size_t leadsize = ALIGNMENT_CEILING((uintptr_t)pages, alignment) + - (uintptr_t)pages; + ret = os_pages_trim(pages, alloc_size, leadsize, size, commit); + } while (ret == NULL); + + assert(ret != NULL); + assert(PAGE_ADDR2BASE(ret) == ret); + return ret; +} + +void * +pages_map(void *addr, size_t size, size_t alignment, bool *commit) { + assert(alignment >= PAGE); + assert(ALIGNMENT_ADDR2BASE(addr, alignment) == addr); + + /* + * Ideally, there would be a way to specify alignment to mmap() (like + * NetBSD has), but in the absence of such a feature, we have to work + * hard to efficiently create aligned mappings. The reliable, but + * slow method is to create a mapping that is over-sized, then trim the + * excess. However, that always results in one or two calls to + * os_pages_unmap(), and it can leave holes in the process's virtual + * memory map if memory grows downward. + * + * Optimistically try mapping precisely the right amount before falling + * back to the slow method, with the expectation that the optimistic + * approach works most of the time. + */ + + void *ret = os_pages_map(addr, size, os_page, commit); + if (ret == NULL || ret == addr) { + return ret; + } + assert(addr == NULL); + if (ALIGNMENT_ADDR2OFFSET(ret, alignment) != 0) { + os_pages_unmap(ret, size); + return pages_map_slow(size, alignment, commit); + } + + assert(PAGE_ADDR2BASE(ret) == ret); + return ret; +} + +void +pages_unmap(void *addr, size_t size) { + assert(PAGE_ADDR2BASE(addr) == addr); + assert(PAGE_CEILING(size) == size); + + os_pages_unmap(addr, size); +} + +static bool +pages_commit_impl(void *addr, size_t size, bool commit) { + assert(PAGE_ADDR2BASE(addr) == addr); + assert(PAGE_CEILING(size) == size); + + if (os_overcommits) { + return true; + } + +#ifdef _WIN32 + return (commit ? (addr != VirtualAlloc(addr, size, MEM_COMMIT, + PAGE_READWRITE)) : (!VirtualFree(addr, size, MEM_DECOMMIT))); +#else + { + int prot = commit ? PAGES_PROT_COMMIT : PAGES_PROT_DECOMMIT; + void *result = mmap(addr, size, prot, mmap_flags | MAP_FIXED, + -1, 0); + if (result == MAP_FAILED) { + return true; + } + if (result != addr) { + /* + * We succeeded in mapping memory, but not in the right + * place. + */ + os_pages_unmap(result, size); + return true; + } + return false; + } +#endif +} + +bool +pages_commit(void *addr, size_t size) { + return pages_commit_impl(addr, size, true); +} + +bool +pages_decommit(void *addr, size_t size) { + return pages_commit_impl(addr, size, false); +} + +bool +pages_purge_lazy(void *addr, size_t size) { + assert(PAGE_ADDR2BASE(addr) == addr); + assert(PAGE_CEILING(size) == size); + + if (!pages_can_purge_lazy) { + return true; + } + +#ifdef _WIN32 + VirtualAlloc(addr, size, MEM_RESET, PAGE_READWRITE); + return false; +#elif defined(JEMALLOC_PURGE_MADVISE_FREE) && \ + !defined(PAGES_CAN_PURGE_LAZY) + return (madvise(addr, size, MADV_FREE) != 0); +#elif defined(JEMALLOC_PURGE_MADVISE_DONTNEED) && \ + !defined(JEMALLOC_PURGE_MADVISE_DONTNEED_ZEROS) + return (madvise(addr, size, MADV_DONTNEED) != 0); +#else + not_reached(); +#endif +} + +bool +pages_purge_forced(void *addr, size_t size) { + assert(PAGE_ADDR2BASE(addr) == addr); + assert(PAGE_CEILING(size) == size); + + if (!pages_can_purge_forced) { + return true; + } + +#if defined(JEMALLOC_PURGE_MADVISE_DONTNEED) && \ + defined(JEMALLOC_PURGE_MADVISE_DONTNEED_ZEROS) + return (madvise(addr, size, MADV_DONTNEED) != 0); +#elif defined(JEMALLOC_MAPS_COALESCE) + /* Try to overlay a new demand-zeroed mapping. */ + return pages_commit(addr, size); +#else + not_reached(); +#endif +} + +bool +pages_huge(void *addr, size_t size) { + assert(HUGEPAGE_ADDR2BASE(addr) == addr); + assert(HUGEPAGE_CEILING(size) == size); + +#ifdef JEMALLOC_THP + return (madvise(addr, size, MADV_HUGEPAGE) != 0); +#else + return true; +#endif +} + +bool +pages_nohuge(void *addr, size_t size) { + assert(HUGEPAGE_ADDR2BASE(addr) == addr); + assert(HUGEPAGE_CEILING(size) == size); + +#ifdef JEMALLOC_THP + return (madvise(addr, size, MADV_NOHUGEPAGE) != 0); +#else + return false; +#endif +} + +static size_t +os_page_detect(void) { +#ifdef _WIN32 + SYSTEM_INFO si; + GetSystemInfo(&si); + return si.dwPageSize; +#else + long result = sysconf(_SC_PAGESIZE); + if (result == -1) { + return LG_PAGE; + } + return (size_t)result; +#endif +} + +#ifdef JEMALLOC_SYSCTL_VM_OVERCOMMIT +static bool +os_overcommits_sysctl(void) { + int vm_overcommit; + size_t sz; + + sz = sizeof(vm_overcommit); + if (sysctlbyname("vm.overcommit", &vm_overcommit, &sz, NULL, 0) != 0) { + return false; /* Error. */ + } + + return ((vm_overcommit & 0x3) == 0); +} +#endif + +#ifdef JEMALLOC_PROC_SYS_VM_OVERCOMMIT_MEMORY +/* + * Use syscall(2) rather than {open,read,close}(2) when possible to avoid + * reentry during bootstrapping if another library has interposed system call + * wrappers. + */ +static bool +os_overcommits_proc(void) { + int fd; + char buf[1]; + ssize_t nread; + +#if defined(JEMALLOC_USE_SYSCALL) && defined(SYS_open) + fd = (int)syscall(SYS_open, "/proc/sys/vm/overcommit_memory", O_RDONLY | + O_CLOEXEC); +#elif defined(JEMALLOC_USE_SYSCALL) && defined(SYS_openat) + fd = (int)syscall(SYS_openat, + AT_FDCWD, "/proc/sys/vm/overcommit_memory", O_RDONLY | O_CLOEXEC); +#else + fd = open("/proc/sys/vm/overcommit_memory", O_RDONLY | O_CLOEXEC); +#endif + if (fd == -1) { + return false; /* Error. */ + } + +#if defined(JEMALLOC_USE_SYSCALL) && defined(SYS_read) + nread = (ssize_t)syscall(SYS_read, fd, &buf, sizeof(buf)); +#else + nread = read(fd, &buf, sizeof(buf)); +#endif + +#if defined(JEMALLOC_USE_SYSCALL) && defined(SYS_close) + syscall(SYS_close, fd); +#else + close(fd); +#endif + + if (nread < 1) { + return false; /* Error. */ + } + /* + * /proc/sys/vm/overcommit_memory meanings: + * 0: Heuristic overcommit. + * 1: Always overcommit. + * 2: Never overcommit. + */ + return (buf[0] == '0' || buf[0] == '1'); +} +#endif + +bool +pages_boot(void) { + os_page = os_page_detect(); + if (os_page > PAGE) { + malloc_write("<jemalloc>: Unsupported system page size\n"); + if (opt_abort) { + abort(); + } + return true; + } + +#ifndef _WIN32 + mmap_flags = MAP_PRIVATE | MAP_ANON; +#endif + +#ifdef JEMALLOC_SYSCTL_VM_OVERCOMMIT + os_overcommits = os_overcommits_sysctl(); +#elif defined(JEMALLOC_PROC_SYS_VM_OVERCOMMIT_MEMORY) + os_overcommits = os_overcommits_proc(); +# ifdef MAP_NORESERVE + if (os_overcommits) { + mmap_flags |= MAP_NORESERVE; + } +# endif +#else + os_overcommits = false; +#endif + + return false; +} diff --git a/deps/jemalloc/src/prng.c b/deps/jemalloc/src/prng.c new file mode 100644 index 0000000000..83c04bf9b5 --- /dev/null +++ b/deps/jemalloc/src/prng.c @@ -0,0 +1,3 @@ +#define JEMALLOC_PRNG_C_ +#include "jemalloc/internal/jemalloc_preamble.h" +#include "jemalloc/internal/jemalloc_internal_includes.h" diff --git a/deps/jemalloc/src/spin.c b/deps/jemalloc/src/spin.c new file mode 100644 index 0000000000..24372c26c9 --- /dev/null +++ b/deps/jemalloc/src/spin.c @@ -0,0 +1,4 @@ +#define JEMALLOC_SPIN_C_ +#include "jemalloc/internal/jemalloc_preamble.h" + +#include "jemalloc/internal/spin.h" diff --git a/deps/jemalloc/src/sz.c b/deps/jemalloc/src/sz.c new file mode 100644 index 0000000000..0986615f71 --- /dev/null +++ b/deps/jemalloc/src/sz.c @@ -0,0 +1,106 @@ +#include "jemalloc/internal/jemalloc_preamble.h" +#include "jemalloc/internal/sz.h" + +JEMALLOC_ALIGNED(CACHELINE) +const size_t sz_pind2sz_tab[NPSIZES+1] = { +#define PSZ_yes(lg_grp, ndelta, lg_delta) \ + (((ZU(1)<<lg_grp) + (ZU(ndelta)<<lg_delta))), +#define PSZ_no(lg_grp, ndelta, lg_delta) +#define SC(index, lg_grp, lg_delta, ndelta, psz, bin, pgs, lg_delta_lookup) \ + PSZ_##psz(lg_grp, ndelta, lg_delta) + SIZE_CLASSES +#undef PSZ_yes +#undef PSZ_no +#undef SC + (LARGE_MAXCLASS + PAGE) +}; + +JEMALLOC_ALIGNED(CACHELINE) +const size_t sz_index2size_tab[NSIZES] = { +#define SC(index, lg_grp, lg_delta, ndelta, psz, bin, pgs, lg_delta_lookup) \ + ((ZU(1)<<lg_grp) + (ZU(ndelta)<<lg_delta)), + SIZE_CLASSES +#undef SC +}; + +JEMALLOC_ALIGNED(CACHELINE) +const uint8_t sz_size2index_tab[] = { +#if LG_TINY_MIN == 0 +#warning "Dangerous LG_TINY_MIN" +#define S2B_0(i) i, +#elif LG_TINY_MIN == 1 +#warning "Dangerous LG_TINY_MIN" +#define S2B_1(i) i, +#elif LG_TINY_MIN == 2 +#warning "Dangerous LG_TINY_MIN" +#define S2B_2(i) i, +#elif LG_TINY_MIN == 3 +#define S2B_3(i) i, +#elif LG_TINY_MIN == 4 +#define S2B_4(i) i, +#elif LG_TINY_MIN == 5 +#define S2B_5(i) i, +#elif LG_TINY_MIN == 6 +#define S2B_6(i) i, +#elif LG_TINY_MIN == 7 +#define S2B_7(i) i, +#elif LG_TINY_MIN == 8 +#define S2B_8(i) i, +#elif LG_TINY_MIN == 9 +#define S2B_9(i) i, +#elif LG_TINY_MIN == 10 +#define S2B_10(i) i, +#elif LG_TINY_MIN == 11 +#define S2B_11(i) i, +#else +#error "Unsupported LG_TINY_MIN" +#endif +#if LG_TINY_MIN < 1 +#define S2B_1(i) S2B_0(i) S2B_0(i) +#endif +#if LG_TINY_MIN < 2 +#define S2B_2(i) S2B_1(i) S2B_1(i) +#endif +#if LG_TINY_MIN < 3 +#define S2B_3(i) S2B_2(i) S2B_2(i) +#endif +#if LG_TINY_MIN < 4 +#define S2B_4(i) S2B_3(i) S2B_3(i) +#endif +#if LG_TINY_MIN < 5 +#define S2B_5(i) S2B_4(i) S2B_4(i) +#endif +#if LG_TINY_MIN < 6 +#define S2B_6(i) S2B_5(i) S2B_5(i) +#endif +#if LG_TINY_MIN < 7 +#define S2B_7(i) S2B_6(i) S2B_6(i) +#endif +#if LG_TINY_MIN < 8 +#define S2B_8(i) S2B_7(i) S2B_7(i) +#endif +#if LG_TINY_MIN < 9 +#define S2B_9(i) S2B_8(i) S2B_8(i) +#endif +#if LG_TINY_MIN < 10 +#define S2B_10(i) S2B_9(i) S2B_9(i) +#endif +#if LG_TINY_MIN < 11 +#define S2B_11(i) S2B_10(i) S2B_10(i) +#endif +#define S2B_no(i) +#define SC(index, lg_grp, lg_delta, ndelta, psz, bin, pgs, lg_delta_lookup) \ + S2B_##lg_delta_lookup(index) + SIZE_CLASSES +#undef S2B_3 +#undef S2B_4 +#undef S2B_5 +#undef S2B_6 +#undef S2B_7 +#undef S2B_8 +#undef S2B_9 +#undef S2B_10 +#undef S2B_11 +#undef S2B_no +#undef SC +}; diff --git a/deps/jemalloc/src/ticker.c b/deps/jemalloc/src/ticker.c new file mode 100644 index 0000000000..d7b8cd26c0 --- /dev/null +++ b/deps/jemalloc/src/ticker.c @@ -0,0 +1,3 @@ +#define JEMALLOC_TICKER_C_ +#include "jemalloc/internal/jemalloc_preamble.h" +#include "jemalloc/internal/jemalloc_internal_includes.h" diff --git a/deps/jemalloc/src/witness.c b/deps/jemalloc/src/witness.c new file mode 100644 index 0000000000..f42b72ad1a --- /dev/null +++ b/deps/jemalloc/src/witness.c @@ -0,0 +1,100 @@ +#define JEMALLOC_WITNESS_C_ +#include "jemalloc/internal/jemalloc_preamble.h" +#include "jemalloc/internal/jemalloc_internal_includes.h" + +#include "jemalloc/internal/assert.h" +#include "jemalloc/internal/malloc_io.h" + +void +witness_init(witness_t *witness, const char *name, witness_rank_t rank, + witness_comp_t *comp, void *opaque) { + witness->name = name; + witness->rank = rank; + witness->comp = comp; + witness->opaque = opaque; +} + +static void +witness_lock_error_impl(const witness_list_t *witnesses, + const witness_t *witness) { + witness_t *w; + + malloc_printf("<jemalloc>: Lock rank order reversal:"); + ql_foreach(w, witnesses, link) { + malloc_printf(" %s(%u)", w->name, w->rank); + } + malloc_printf(" %s(%u)\n", witness->name, witness->rank); + abort(); +} +witness_lock_error_t *JET_MUTABLE witness_lock_error = witness_lock_error_impl; + +static void +witness_owner_error_impl(const witness_t *witness) { + malloc_printf("<jemalloc>: Should own %s(%u)\n", witness->name, + witness->rank); + abort(); +} +witness_owner_error_t *JET_MUTABLE witness_owner_error = + witness_owner_error_impl; + +static void +witness_not_owner_error_impl(const witness_t *witness) { + malloc_printf("<jemalloc>: Should not own %s(%u)\n", witness->name, + witness->rank); + abort(); +} +witness_not_owner_error_t *JET_MUTABLE witness_not_owner_error = + witness_not_owner_error_impl; + +static void +witness_depth_error_impl(const witness_list_t *witnesses, + witness_rank_t rank_inclusive, unsigned depth) { + witness_t *w; + + malloc_printf("<jemalloc>: Should own %u lock%s of rank >= %u:", depth, + (depth != 1) ? "s" : "", rank_inclusive); + ql_foreach(w, witnesses, link) { + malloc_printf(" %s(%u)", w->name, w->rank); + } + malloc_printf("\n"); + abort(); +} +witness_depth_error_t *JET_MUTABLE witness_depth_error = + witness_depth_error_impl; + +void +witnesses_cleanup(witness_tsd_t *witness_tsd) { + witness_assert_lockless(witness_tsd_tsdn(witness_tsd)); + + /* Do nothing. */ +} + +void +witness_prefork(witness_tsd_t *witness_tsd) { + if (!config_debug) { + return; + } + witness_tsd->forking = true; +} + +void +witness_postfork_parent(witness_tsd_t *witness_tsd) { + if (!config_debug) { + return; + } + witness_tsd->forking = false; +} + +void +witness_postfork_child(witness_tsd_t *witness_tsd) { + if (!config_debug) { + return; + } +#ifndef JEMALLOC_MUTEX_INIT_CB + witness_list_t *witnesses; + + witnesses = &witness_tsd->witnesses; + ql_new(witnesses); +#endif + witness_tsd->forking = false; +} |