Added TCMalloc and JEMalloc projects

BeefRT/JEMalloc/include/jemalloc/internal/activity_callback.h

@@ -0,0 +1,23 @@
+#ifndef JEMALLOC_INTERNAL_ACTIVITY_CALLBACK_H
+#define JEMALLOC_INTERNAL_ACTIVITY_CALLBACK_H
+
+/*
+ * The callback to be executed "periodically", in response to some amount of
+ * allocator activity.
+ *
+ * This callback need not be computing any sort of peak (although that's the
+ * intended first use case), but we drive it from the peak counter, so it's
+ * keeps things tidy to keep it here.
+ *
+ * The calls to this thunk get driven by the peak_event module.
+ */
+#define ACTIVITY_CALLBACK_THUNK_INITIALIZER {NULL, NULL}
+typedef void (*activity_callback_t)(void *uctx, uint64_t allocated,
+    uint64_t deallocated);
+typedef struct activity_callback_thunk_s activity_callback_thunk_t;
+struct activity_callback_thunk_s {
+	activity_callback_t callback;
+	void *uctx;
+};
+
+#endif /* JEMALLOC_INTERNAL_ACTIVITY_CALLBACK_H */

BeefRT/JEMalloc/include/jemalloc/internal/arena_externs.h

@@ -0,0 +1,121 @@
+#ifndef JEMALLOC_INTERNAL_ARENA_EXTERNS_H
+#define JEMALLOC_INTERNAL_ARENA_EXTERNS_H
+
+#include "jemalloc/internal/bin.h"
+#include "jemalloc/internal/div.h"
+#include "jemalloc/internal/extent_dss.h"
+#include "jemalloc/internal/hook.h"
+#include "jemalloc/internal/pages.h"
+#include "jemalloc/internal/stats.h"
+
+/*
+ * When the amount of pages to be purged exceeds this amount, deferred purge
+ * should happen.
+ */
+#define ARENA_DEFERRED_PURGE_NPAGES_THRESHOLD UINT64_C(1024)
+
+extern ssize_t opt_dirty_decay_ms;
+extern ssize_t opt_muzzy_decay_ms;
+
+extern percpu_arena_mode_t opt_percpu_arena;
+extern const char *percpu_arena_mode_names[];
+
+extern div_info_t arena_binind_div_info[SC_NBINS];
+
+extern malloc_mutex_t arenas_lock;
+extern emap_t arena_emap_global;
+
+extern size_t opt_oversize_threshold;
+extern size_t oversize_threshold;
+
+/*
+ * arena_bin_offsets[binind] is the offset of the first bin shard for size class
+ * binind.
+ */
+extern uint32_t arena_bin_offsets[SC_NBINS];
+
+void arena_basic_stats_merge(tsdn_t *tsdn, arena_t *arena,
+    unsigned *nthreads, const char **dss, ssize_t *dirty_decay_ms,
+    ssize_t *muzzy_decay_ms, size_t *nactive, size_t *ndirty, size_t *nmuzzy);
+void arena_stats_merge(tsdn_t *tsdn, arena_t *arena, unsigned *nthreads,
+    const char **dss, ssize_t *dirty_decay_ms, ssize_t *muzzy_decay_ms,
+    size_t *nactive, size_t *ndirty, size_t *nmuzzy, arena_stats_t *astats,
+    bin_stats_data_t *bstats, arena_stats_large_t *lstats,
+    pac_estats_t *estats, hpa_shard_stats_t *hpastats, sec_stats_t *secstats);
+void arena_handle_deferred_work(tsdn_t *tsdn, arena_t *arena);
+edata_t *arena_extent_alloc_large(tsdn_t *tsdn, arena_t *arena,
+    size_t usize, size_t alignment, bool zero);
+void arena_extent_dalloc_large_prep(tsdn_t *tsdn, arena_t *arena,
+    edata_t *edata);
+void arena_extent_ralloc_large_shrink(tsdn_t *tsdn, arena_t *arena,
+    edata_t *edata, size_t oldsize);
+void arena_extent_ralloc_large_expand(tsdn_t *tsdn, arena_t *arena,
+    edata_t *edata, size_t oldsize);
+bool arena_decay_ms_set(tsdn_t *tsdn, arena_t *arena, extent_state_t state,
+    ssize_t decay_ms);
+ssize_t arena_decay_ms_get(arena_t *arena, extent_state_t state);
+void arena_decay(tsdn_t *tsdn, arena_t *arena, bool is_background_thread,
+    bool all);
+uint64_t arena_time_until_deferred(tsdn_t *tsdn, arena_t *arena);
+void arena_do_deferred_work(tsdn_t *tsdn, arena_t *arena);
+void arena_reset(tsd_t *tsd, arena_t *arena);
+void arena_destroy(tsd_t *tsd, arena_t *arena);
+void arena_cache_bin_fill_small(tsdn_t *tsdn, arena_t *arena,
+    cache_bin_t *cache_bin, cache_bin_info_t *cache_bin_info, szind_t binind,
+    const unsigned nfill);
+
+void *arena_malloc_hard(tsdn_t *tsdn, arena_t *arena, size_t size,
+    szind_t ind, bool zero);
+void *arena_palloc(tsdn_t *tsdn, arena_t *arena, size_t usize,
+    size_t alignment, bool zero, tcache_t *tcache);
+void arena_prof_promote(tsdn_t *tsdn, void *ptr, size_t usize);
+void arena_dalloc_promoted(tsdn_t *tsdn, void *ptr, tcache_t *tcache,
+    bool slow_path);
+void arena_slab_dalloc(tsdn_t *tsdn, arena_t *arena, edata_t *slab);
+
+void arena_dalloc_bin_locked_handle_newly_empty(tsdn_t *tsdn, arena_t *arena,
+    edata_t *slab, bin_t *bin);
+void arena_dalloc_bin_locked_handle_newly_nonempty(tsdn_t *tsdn, arena_t *arena,
+    edata_t *slab, bin_t *bin);
+void arena_dalloc_small(tsdn_t *tsdn, void *ptr);
+bool arena_ralloc_no_move(tsdn_t *tsdn, void *ptr, size_t oldsize, size_t size,
+    size_t extra, bool zero, size_t *newsize);
+void *arena_ralloc(tsdn_t *tsdn, arena_t *arena, void *ptr, size_t oldsize,
+    size_t size, size_t alignment, bool zero, tcache_t *tcache,
+    hook_ralloc_args_t *hook_args);
+dss_prec_t arena_dss_prec_get(arena_t *arena);
+ehooks_t *arena_get_ehooks(arena_t *arena);
+extent_hooks_t *arena_set_extent_hooks(tsd_t *tsd, arena_t *arena,
+    extent_hooks_t *extent_hooks);
+bool arena_dss_prec_set(arena_t *arena, dss_prec_t dss_prec);
+ssize_t arena_dirty_decay_ms_default_get(void);
+bool arena_dirty_decay_ms_default_set(ssize_t decay_ms);
+ssize_t arena_muzzy_decay_ms_default_get(void);
+bool arena_muzzy_decay_ms_default_set(ssize_t decay_ms);
+bool arena_retain_grow_limit_get_set(tsd_t *tsd, arena_t *arena,
+    size_t *old_limit, size_t *new_limit);
+unsigned arena_nthreads_get(arena_t *arena, bool internal);
+void arena_nthreads_inc(arena_t *arena, bool internal);
+void arena_nthreads_dec(arena_t *arena, bool internal);
+arena_t *arena_new(tsdn_t *tsdn, unsigned ind, const arena_config_t *config);
+bool arena_init_huge(void);
+bool arena_is_huge(unsigned arena_ind);
+arena_t *arena_choose_huge(tsd_t *tsd);
+bin_t *arena_bin_choose(tsdn_t *tsdn, arena_t *arena, szind_t binind,
+    unsigned *binshard);
+size_t arena_fill_small_fresh(tsdn_t *tsdn, arena_t *arena, szind_t binind,
+    void **ptrs, size_t nfill, bool zero);
+bool arena_boot(sc_data_t *sc_data, base_t *base, bool hpa);
+void arena_prefork0(tsdn_t *tsdn, arena_t *arena);
+void arena_prefork1(tsdn_t *tsdn, arena_t *arena);
+void arena_prefork2(tsdn_t *tsdn, arena_t *arena);
+void arena_prefork3(tsdn_t *tsdn, arena_t *arena);
+void arena_prefork4(tsdn_t *tsdn, arena_t *arena);
+void arena_prefork5(tsdn_t *tsdn, arena_t *arena);
+void arena_prefork6(tsdn_t *tsdn, arena_t *arena);
+void arena_prefork7(tsdn_t *tsdn, arena_t *arena);
+void arena_prefork8(tsdn_t *tsdn, arena_t *arena);
+void arena_postfork_parent(tsdn_t *tsdn, arena_t *arena);
+void arena_postfork_child(tsdn_t *tsdn, arena_t *arena);
+
+#endif /* JEMALLOC_INTERNAL_ARENA_EXTERNS_H */

BeefRT/JEMalloc/include/jemalloc/internal/arena_inlines_a.h

@@ -0,0 +1,24 @@
+#ifndef JEMALLOC_INTERNAL_ARENA_INLINES_A_H
+#define JEMALLOC_INTERNAL_ARENA_INLINES_A_H
+
+static inline unsigned
+arena_ind_get(const arena_t *arena) {
+	return arena->ind;
+}
+
+static inline void
+arena_internal_add(arena_t *arena, size_t size) {
+	atomic_fetch_add_zu(&arena->stats.internal, size, ATOMIC_RELAXED);
+}
+
+static inline void
+arena_internal_sub(arena_t *arena, size_t size) {
+	atomic_fetch_sub_zu(&arena->stats.internal, size, ATOMIC_RELAXED);
+}
+
+static inline size_t
+arena_internal_get(arena_t *arena) {
+	return atomic_load_zu(&arena->stats.internal, ATOMIC_RELAXED);
+}
+
+#endif /* JEMALLOC_INTERNAL_ARENA_INLINES_A_H */

BeefRT/JEMalloc/include/jemalloc/internal/arena_inlines_b.h

@@ -0,0 +1,550 @@
+#ifndef JEMALLOC_INTERNAL_ARENA_INLINES_B_H
+#define JEMALLOC_INTERNAL_ARENA_INLINES_B_H
+
+#include "jemalloc/internal/div.h"
+#include "jemalloc/internal/emap.h"
+#include "jemalloc/internal/jemalloc_internal_types.h"
+#include "jemalloc/internal/mutex.h"
+#include "jemalloc/internal/rtree.h"
+#include "jemalloc/internal/safety_check.h"
+#include "jemalloc/internal/sc.h"
+#include "jemalloc/internal/sz.h"
+#include "jemalloc/internal/ticker.h"
+
+static inline arena_t *
+arena_get_from_edata(edata_t *edata) {
+	return (arena_t *)atomic_load_p(&arenas[edata_arena_ind_get(edata)],
+	    ATOMIC_RELAXED);
+}
+
+JEMALLOC_ALWAYS_INLINE arena_t *
+arena_choose_maybe_huge(tsd_t *tsd, arena_t *arena, size_t size) {
+	if (arena != NULL) {
+		return arena;
+	}
+
+	/*
+	 * For huge allocations, use the dedicated huge arena if both are true:
+	 * 1) is using auto arena selection (i.e. arena == NULL), and 2) the
+	 * thread is not assigned to a manual arena.
+	 */
+	if (unlikely(size >= oversize_threshold)) {
+		arena_t *tsd_arena = tsd_arena_get(tsd);
+		if (tsd_arena == NULL || arena_is_auto(tsd_arena)) {
+			return arena_choose_huge(tsd);
+		}
+	}
+
+	return arena_choose(tsd, NULL);
+}
+
+JEMALLOC_ALWAYS_INLINE void
+arena_prof_info_get(tsd_t *tsd, const void *ptr, emap_alloc_ctx_t *alloc_ctx,
+    prof_info_t *prof_info, bool reset_recent) {
+	cassert(config_prof);
+	assert(ptr != NULL);
+	assert(prof_info != NULL);
+
+	edata_t *edata = NULL;
+	bool is_slab;
+
+	/* Static check. */
+	if (alloc_ctx == NULL) {
+		edata = emap_edata_lookup(tsd_tsdn(tsd), &arena_emap_global,
+		    ptr);
+		is_slab = edata_slab_get(edata);
+	} else if (unlikely(!(is_slab = alloc_ctx->slab))) {
+		edata = emap_edata_lookup(tsd_tsdn(tsd), &arena_emap_global,
+		    ptr);
+	}
+
+	if (unlikely(!is_slab)) {
+		/* edata must have been initialized at this point. */
+		assert(edata != NULL);
+		large_prof_info_get(tsd, edata, prof_info, reset_recent);
+	} else {
+		prof_info->alloc_tctx = (prof_tctx_t *)(uintptr_t)1U;
+		/*
+		 * No need to set other fields in prof_info; they will never be
+		 * accessed if (uintptr_t)alloc_tctx == (uintptr_t)1U.
+		 */
+	}
+}
+
+JEMALLOC_ALWAYS_INLINE void
+arena_prof_tctx_reset(tsd_t *tsd, const void *ptr,
+    emap_alloc_ctx_t *alloc_ctx) {
+	cassert(config_prof);
+	assert(ptr != NULL);
+
+	/* Static check. */
+	if (alloc_ctx == NULL) {
+		edata_t *edata = emap_edata_lookup(tsd_tsdn(tsd),
+		    &arena_emap_global, ptr);
+		if (unlikely(!edata_slab_get(edata))) {
+			large_prof_tctx_reset(edata);
+		}
+	} else {
+		if (unlikely(!alloc_ctx->slab)) {
+			edata_t *edata = emap_edata_lookup(tsd_tsdn(tsd),
+			    &arena_emap_global, ptr);
+			large_prof_tctx_reset(edata);
+		}
+	}
+}
+
+JEMALLOC_ALWAYS_INLINE void
+arena_prof_tctx_reset_sampled(tsd_t *tsd, const void *ptr) {
+	cassert(config_prof);
+	assert(ptr != NULL);
+
+	edata_t *edata = emap_edata_lookup(tsd_tsdn(tsd), &arena_emap_global,
+	    ptr);
+	assert(!edata_slab_get(edata));
+
+	large_prof_tctx_reset(edata);
+}
+
+JEMALLOC_ALWAYS_INLINE void
+arena_prof_info_set(tsd_t *tsd, edata_t *edata, prof_tctx_t *tctx,
+    size_t size) {
+	cassert(config_prof);
+
+	assert(!edata_slab_get(edata));
+	large_prof_info_set(edata, tctx, size);
+}
+
+JEMALLOC_ALWAYS_INLINE void
+arena_decay_ticks(tsdn_t *tsdn, arena_t *arena, unsigned nticks) {
+	if (unlikely(tsdn_null(tsdn))) {
+		return;
+	}
+	tsd_t *tsd = tsdn_tsd(tsdn);
+	/*
+	 * We use the ticker_geom_t to avoid having per-arena state in the tsd.
+	 * Instead of having a countdown-until-decay timer running for every
+	 * arena in every thread, we flip a coin once per tick, whose
+	 * probability of coming up heads is 1/nticks; this is effectively the
+	 * operation of the ticker_geom_t.  Each arena has the same chance of a
+	 * coinflip coming up heads (1/ARENA_DECAY_NTICKS_PER_UPDATE), so we can
+	 * use a single ticker for all of them.
+	 */
+	ticker_geom_t *decay_ticker = tsd_arena_decay_tickerp_get(tsd);
+	uint64_t *prng_state = tsd_prng_statep_get(tsd);
+	if (unlikely(ticker_geom_ticks(decay_ticker, prng_state, nticks))) {
+		arena_decay(tsdn, arena, false, false);
+	}
+}
+
+JEMALLOC_ALWAYS_INLINE void
+arena_decay_tick(tsdn_t *tsdn, arena_t *arena) {
+	arena_decay_ticks(tsdn, arena, 1);
+}
+
+JEMALLOC_ALWAYS_INLINE void *
+arena_malloc(tsdn_t *tsdn, arena_t *arena, size_t size, szind_t ind, bool zero,
+    tcache_t *tcache, bool slow_path) {
+	assert(!tsdn_null(tsdn) || tcache == NULL);
+
+	if (likely(tcache != NULL)) {
+		if (likely(size <= SC_SMALL_MAXCLASS)) {
+			return tcache_alloc_small(tsdn_tsd(tsdn), arena,
+			    tcache, size, ind, zero, slow_path);
+		}
+		if (likely(size <= tcache_maxclass)) {
+			return tcache_alloc_large(tsdn_tsd(tsdn), arena,
+			    tcache, size, ind, zero, slow_path);
+		}
+		/* (size > tcache_maxclass) case falls through. */
+		assert(size > tcache_maxclass);
+	}
+
+	return arena_malloc_hard(tsdn, arena, size, ind, zero);
+}
+
+JEMALLOC_ALWAYS_INLINE arena_t *
+arena_aalloc(tsdn_t *tsdn, const void *ptr) {
+	edata_t *edata = emap_edata_lookup(tsdn, &arena_emap_global, ptr);
+	unsigned arena_ind = edata_arena_ind_get(edata);
+	return (arena_t *)atomic_load_p(&arenas[arena_ind], ATOMIC_RELAXED);
+}
+
+JEMALLOC_ALWAYS_INLINE size_t
+arena_salloc(tsdn_t *tsdn, const void *ptr) {
+	assert(ptr != NULL);
+	emap_alloc_ctx_t alloc_ctx;
+	emap_alloc_ctx_lookup(tsdn, &arena_emap_global, ptr, &alloc_ctx);
+	assert(alloc_ctx.szind != SC_NSIZES);
+
+	return sz_index2size(alloc_ctx.szind);
+}
+
+JEMALLOC_ALWAYS_INLINE size_t
+arena_vsalloc(tsdn_t *tsdn, const void *ptr) {
+	/*
+	 * Return 0 if ptr is not within an extent managed by jemalloc.  This
+	 * function has two extra costs relative to isalloc():
+	 * - The rtree calls cannot claim to be dependent lookups, which induces
+	 *   rtree lookup load dependencies.
+	 * - The lookup may fail, so there is an extra branch to check for
+	 *   failure.
+	 */
+
+	emap_full_alloc_ctx_t full_alloc_ctx;
+	bool missing = emap_full_alloc_ctx_try_lookup(tsdn, &arena_emap_global,
+	    ptr, &full_alloc_ctx);
+	if (missing) {
+		return 0;
+	}
+
+	if (full_alloc_ctx.edata == NULL) {
+		return 0;
+	}
+	assert(edata_state_get(full_alloc_ctx.edata) == extent_state_active);
+	/* Only slab members should be looked up via interior pointers. */
+	assert(edata_addr_get(full_alloc_ctx.edata) == ptr
+	    || edata_slab_get(full_alloc_ctx.edata));
+
+	assert(full_alloc_ctx.szind != SC_NSIZES);
+
+	return sz_index2size(full_alloc_ctx.szind);
+}
+
+JEMALLOC_ALWAYS_INLINE bool
+large_dalloc_safety_checks(edata_t *edata, void *ptr, szind_t szind) {
+	if (!config_opt_safety_checks) {
+		return false;
+	}
+
+	/*
+	 * Eagerly detect double free and sized dealloc bugs for large sizes.
+	 * The cost is low enough (as edata will be accessed anyway) to be
+	 * enabled all the time.
+	 */
+	if (unlikely(edata == NULL ||
+	    edata_state_get(edata) != extent_state_active)) {
+		safety_check_fail("Invalid deallocation detected: "
+		    "pages being freed (%p) not currently active, "
+		    "possibly caused by double free bugs.",
+		    (uintptr_t)edata_addr_get(edata));
+		return true;
+	}
+	size_t input_size = sz_index2size(szind);
+	if (unlikely(input_size != edata_usize_get(edata))) {
+		safety_check_fail_sized_dealloc(/* current_dealloc */ true, ptr,
+		    /* true_size */ edata_usize_get(edata), input_size);
+		return true;
+	}
+
+	return false;
+}
+
+static inline void
+arena_dalloc_large_no_tcache(tsdn_t *tsdn, void *ptr, szind_t szind) {
+	if (config_prof && unlikely(szind < SC_NBINS)) {
+		arena_dalloc_promoted(tsdn, ptr, NULL, true);
+	} else {
+		edata_t *edata = emap_edata_lookup(tsdn, &arena_emap_global,
+		    ptr);
+		if (large_dalloc_safety_checks(edata, ptr, szind)) {
+			/* See the comment in isfree. */
+			return;
+		}
+		large_dalloc(tsdn, edata);
+	}
+}
+
+static inline void
+arena_dalloc_no_tcache(tsdn_t *tsdn, void *ptr) {
+	assert(ptr != NULL);
+
+	emap_alloc_ctx_t alloc_ctx;
+	emap_alloc_ctx_lookup(tsdn, &arena_emap_global, ptr, &alloc_ctx);
+
+	if (config_debug) {
+		edata_t *edata = emap_edata_lookup(tsdn, &arena_emap_global,
+		    ptr);
+		assert(alloc_ctx.szind == edata_szind_get(edata));
+		assert(alloc_ctx.szind < SC_NSIZES);
+		assert(alloc_ctx.slab == edata_slab_get(edata));
+	}
+
+	if (likely(alloc_ctx.slab)) {
+		/* Small allocation. */
+		arena_dalloc_small(tsdn, ptr);
+	} else {
+		arena_dalloc_large_no_tcache(tsdn, ptr, alloc_ctx.szind);
+	}
+}
+
+JEMALLOC_ALWAYS_INLINE void
+arena_dalloc_large(tsdn_t *tsdn, void *ptr, tcache_t *tcache, szind_t szind,
+    bool slow_path) {
+	if (szind < nhbins) {
+		if (config_prof && unlikely(szind < SC_NBINS)) {
+			arena_dalloc_promoted(tsdn, ptr, tcache, slow_path);
+		} else {
+			tcache_dalloc_large(tsdn_tsd(tsdn), tcache, ptr, szind,
+			    slow_path);
+		}
+	} else {
+		edata_t *edata = emap_edata_lookup(tsdn, &arena_emap_global,
+		    ptr);
+		if (large_dalloc_safety_checks(edata, ptr, szind)) {
+			/* See the comment in isfree. */
+			return;
+		}
+		large_dalloc(tsdn, edata);
+	}
+}
+
+JEMALLOC_ALWAYS_INLINE void
+arena_dalloc(tsdn_t *tsdn, void *ptr, tcache_t *tcache,
+    emap_alloc_ctx_t *caller_alloc_ctx, bool slow_path) {
+	assert(!tsdn_null(tsdn) || tcache == NULL);
+	assert(ptr != NULL);
+
+	if (unlikely(tcache == NULL)) {
+		arena_dalloc_no_tcache(tsdn, ptr);
+		return;
+	}
+
+	emap_alloc_ctx_t alloc_ctx;
+	if (caller_alloc_ctx != NULL) {
+		alloc_ctx = *caller_alloc_ctx;
+	} else {
+		util_assume(!tsdn_null(tsdn));
+		emap_alloc_ctx_lookup(tsdn, &arena_emap_global, ptr,
+		    &alloc_ctx);
+	}
+
+	if (config_debug) {
+		edata_t *edata = emap_edata_lookup(tsdn, &arena_emap_global,
+		    ptr);
+		assert(alloc_ctx.szind == edata_szind_get(edata));
+		assert(alloc_ctx.szind < SC_NSIZES);
+		assert(alloc_ctx.slab == edata_slab_get(edata));
+	}
+
+	if (likely(alloc_ctx.slab)) {
+		/* Small allocation. */
+		tcache_dalloc_small(tsdn_tsd(tsdn), tcache, ptr,
+		    alloc_ctx.szind, slow_path);
+	} else {
+		arena_dalloc_large(tsdn, ptr, tcache, alloc_ctx.szind,
+		    slow_path);
+	}
+}
+
+static inline void
+arena_sdalloc_no_tcache(tsdn_t *tsdn, void *ptr, size_t size) {
+	assert(ptr != NULL);
+	assert(size <= SC_LARGE_MAXCLASS);
+
+	emap_alloc_ctx_t alloc_ctx;
+	if (!config_prof || !opt_prof) {
+		/*
+		 * There is no risk of being confused by a promoted sampled
+		 * object, so base szind and slab on the given size.
+		 */
+		alloc_ctx.szind = sz_size2index(size);
+		alloc_ctx.slab = (alloc_ctx.szind < SC_NBINS);
+	}
+
+	if ((config_prof && opt_prof) || config_debug) {
+		emap_alloc_ctx_lookup(tsdn, &arena_emap_global, ptr,
+		    &alloc_ctx);
+
+		assert(alloc_ctx.szind == sz_size2index(size));
+		assert((config_prof && opt_prof)
+		    || alloc_ctx.slab == (alloc_ctx.szind < SC_NBINS));
+
+		if (config_debug) {
+			edata_t *edata = emap_edata_lookup(tsdn,
+			    &arena_emap_global, ptr);
+			assert(alloc_ctx.szind == edata_szind_get(edata));
+			assert(alloc_ctx.slab == edata_slab_get(edata));
+		}
+	}
+
+	if (likely(alloc_ctx.slab)) {
+		/* Small allocation. */
+		arena_dalloc_small(tsdn, ptr);
+	} else {
+		arena_dalloc_large_no_tcache(tsdn, ptr, alloc_ctx.szind);
+	}
+}
+
+JEMALLOC_ALWAYS_INLINE void
+arena_sdalloc(tsdn_t *tsdn, void *ptr, size_t size, tcache_t *tcache,
+    emap_alloc_ctx_t *caller_alloc_ctx, bool slow_path) {
+	assert(!tsdn_null(tsdn) || tcache == NULL);
+	assert(ptr != NULL);
+	assert(size <= SC_LARGE_MAXCLASS);
+
+	if (unlikely(tcache == NULL)) {
+		arena_sdalloc_no_tcache(tsdn, ptr, size);
+		return;
+	}
+
+	emap_alloc_ctx_t alloc_ctx;
+	if (config_prof && opt_prof) {
+		if (caller_alloc_ctx == NULL) {
+			/* Uncommon case and should be a static check. */
+			emap_alloc_ctx_lookup(tsdn, &arena_emap_global, ptr,
+			    &alloc_ctx);
+			assert(alloc_ctx.szind == sz_size2index(size));
+		} else {
+			alloc_ctx = *caller_alloc_ctx;
+		}
+	} else {
+		/*
+		 * There is no risk of being confused by a promoted sampled
+		 * object, so base szind and slab on the given size.
+		 */
+		alloc_ctx.szind = sz_size2index(size);
+		alloc_ctx.slab = (alloc_ctx.szind < SC_NBINS);
+	}
+
+	if (config_debug) {
+		edata_t *edata = emap_edata_lookup(tsdn, &arena_emap_global,
+		    ptr);
+		assert(alloc_ctx.szind == edata_szind_get(edata));
+		assert(alloc_ctx.slab == edata_slab_get(edata));
+	}
+
+	if (likely(alloc_ctx.slab)) {
+		/* Small allocation. */
+		tcache_dalloc_small(tsdn_tsd(tsdn), tcache, ptr,
+		    alloc_ctx.szind, slow_path);
+	} else {
+		arena_dalloc_large(tsdn, ptr, tcache, alloc_ctx.szind,
+		    slow_path);
+	}
+}
+
+static inline void
+arena_cache_oblivious_randomize(tsdn_t *tsdn, arena_t *arena, edata_t *edata,
+    size_t alignment) {
+	assert(edata_base_get(edata) == edata_addr_get(edata));
+
+	if (alignment < PAGE) {
+		unsigned lg_range = LG_PAGE -
+		    lg_floor(CACHELINE_CEILING(alignment));
+		size_t r;
+		if (!tsdn_null(tsdn)) {
+			tsd_t *tsd = tsdn_tsd(tsdn);
+			r = (size_t)prng_lg_range_u64(
+			    tsd_prng_statep_get(tsd), lg_range);
+		} else {
+			uint64_t stack_value = (uint64_t)(uintptr_t)&r;
+			r = (size_t)prng_lg_range_u64(&stack_value, lg_range);
+		}
+		uintptr_t random_offset = ((uintptr_t)r) << (LG_PAGE -
+		    lg_range);
+		edata->e_addr = (void *)((uintptr_t)edata->e_addr +
+		    random_offset);
+		assert(ALIGNMENT_ADDR2BASE(edata->e_addr, alignment) ==
+		    edata->e_addr);
+	}
+}
+
+/*
+ * The dalloc bin info contains just the information that the common paths need
+ * during tcache flushes.  By force-inlining these paths, and using local copies
+ * of data (so that the compiler knows it's constant), we avoid a whole bunch of
+ * redundant loads and stores by leaving this information in registers.
+ */
+typedef struct arena_dalloc_bin_locked_info_s arena_dalloc_bin_locked_info_t;
+struct arena_dalloc_bin_locked_info_s {
+	div_info_t div_info;
+	uint32_t nregs;
+	uint64_t ndalloc;
+};
+
+JEMALLOC_ALWAYS_INLINE size_t
+arena_slab_regind(arena_dalloc_bin_locked_info_t *info, szind_t binind,
+    edata_t *slab, const void *ptr) {
+	size_t diff, regind;
+
+	/* Freeing a pointer outside the slab can cause assertion failure. */
+	assert((uintptr_t)ptr >= (uintptr_t)edata_addr_get(slab));
+	assert((uintptr_t)ptr < (uintptr_t)edata_past_get(slab));
+	/* Freeing an interior pointer can cause assertion failure. */
+	assert(((uintptr_t)ptr - (uintptr_t)edata_addr_get(slab)) %
+	    (uintptr_t)bin_infos[binind].reg_size == 0);
+
+	diff = (size_t)((uintptr_t)ptr - (uintptr_t)edata_addr_get(slab));
+
+	/* Avoid doing division with a variable divisor. */
+	regind = div_compute(&info->div_info, diff);
+
+	assert(regind < bin_infos[binind].nregs);
+
+	return regind;
+}
+
+JEMALLOC_ALWAYS_INLINE void
+arena_dalloc_bin_locked_begin(arena_dalloc_bin_locked_info_t *info,
+    szind_t binind) {
+	info->div_info = arena_binind_div_info[binind];
+	info->nregs = bin_infos[binind].nregs;
+	info->ndalloc = 0;
+}
+
+/*
+ * Does the deallocation work associated with freeing a single pointer (a
+ * "step") in between a arena_dalloc_bin_locked begin and end call.
+ *
+ * Returns true if arena_slab_dalloc must be called on slab.  Doesn't do
+ * stats updates, which happen during finish (this lets running counts get left
+ * in a register).
+ */
+JEMALLOC_ALWAYS_INLINE bool
+arena_dalloc_bin_locked_step(tsdn_t *tsdn, arena_t *arena, bin_t *bin,
+    arena_dalloc_bin_locked_info_t *info, szind_t binind, edata_t *slab,
+    void *ptr) {
+	const bin_info_t *bin_info = &bin_infos[binind];
+	size_t regind = arena_slab_regind(info, binind, slab, ptr);
+	slab_data_t *slab_data = edata_slab_data_get(slab);
+
+	assert(edata_nfree_get(slab) < bin_info->nregs);
+	/* Freeing an unallocated pointer can cause assertion failure. */
+	assert(bitmap_get(slab_data->bitmap, &bin_info->bitmap_info, regind));
+
+	bitmap_unset(slab_data->bitmap, &bin_info->bitmap_info, regind);
+	edata_nfree_inc(slab);
+
+	if (config_stats) {
+		info->ndalloc++;
+	}
+
+	unsigned nfree = edata_nfree_get(slab);
+	if (nfree == bin_info->nregs) {
+		arena_dalloc_bin_locked_handle_newly_empty(tsdn, arena, slab,
+		    bin);
+		return true;
+	} else if (nfree == 1 && slab != bin->slabcur) {
+		arena_dalloc_bin_locked_handle_newly_nonempty(tsdn, arena, slab,
+		    bin);
+	}
+	return false;
+}
+
+JEMALLOC_ALWAYS_INLINE void
+arena_dalloc_bin_locked_finish(tsdn_t *tsdn, arena_t *arena, bin_t *bin,
+    arena_dalloc_bin_locked_info_t *info) {
+	if (config_stats) {
+		bin->stats.ndalloc += info->ndalloc;
+		assert(bin->stats.curregs >= (size_t)info->ndalloc);
+		bin->stats.curregs -= (size_t)info->ndalloc;
+	}
+}
+
+static inline bin_t *
+arena_get_bin(arena_t *arena, szind_t binind, unsigned binshard) {
+	bin_t *shard0 = (bin_t *)((uintptr_t)arena + arena_bin_offsets[binind]);
+	return shard0 + binshard;
+}
+
+#endif /* JEMALLOC_INTERNAL_ARENA_INLINES_B_H */

BeefRT/JEMalloc/include/jemalloc/internal/arena_stats.h

@@ -0,0 +1,114 @@
+#ifndef JEMALLOC_INTERNAL_ARENA_STATS_H
+#define JEMALLOC_INTERNAL_ARENA_STATS_H
+
+#include "jemalloc/internal/atomic.h"
+#include "jemalloc/internal/lockedint.h"
+#include "jemalloc/internal/mutex.h"
+#include "jemalloc/internal/mutex_prof.h"
+#include "jemalloc/internal/pa.h"
+#include "jemalloc/internal/sc.h"
+
+JEMALLOC_DIAGNOSTIC_DISABLE_SPURIOUS
+
+typedef struct arena_stats_large_s arena_stats_large_t;
+struct arena_stats_large_s {
+	/*
+	 * Total number of allocation/deallocation requests served directly by
+	 * the arena.
+	 */
+	locked_u64_t	nmalloc;
+	locked_u64_t	ndalloc;
+
+	/*
+	 * Number of allocation requests that correspond to this size class.
+	 * This includes requests served by tcache, though tcache only
+	 * periodically merges into this counter.
+	 */
+	locked_u64_t	nrequests; /* Partially derived. */
+	/*
+	 * Number of tcache fills / flushes for large (similarly, periodically
+	 * merged).  Note that there is no large tcache batch-fill currently
+	 * (i.e. only fill 1 at a time); however flush may be batched.
+	 */
+	locked_u64_t	nfills; /* Partially derived. */
+	locked_u64_t	nflushes; /* Partially derived. */
+
+	/* Current number of allocations of this size class. */
+	size_t		curlextents; /* Derived. */
+};
+
+/*
+ * Arena stats.  Note that fields marked "derived" are not directly maintained
+ * within the arena code; rather their values are derived during stats merge
+ * requests.
+ */
+typedef struct arena_stats_s arena_stats_t;
+struct arena_stats_s {
+	LOCKEDINT_MTX_DECLARE(mtx)
+
+	/*
+	 * resident includes the base stats -- that's why it lives here and not
+	 * in pa_shard_stats_t.
+	 */
+	size_t			base; /* Derived. */
+	size_t			resident; /* Derived. */
+	size_t			metadata_thp; /* Derived. */
+	size_t			mapped; /* Derived. */
+
+	atomic_zu_t		internal;
+
+	size_t			allocated_large; /* Derived. */
+	uint64_t		nmalloc_large; /* Derived. */
+	uint64_t		ndalloc_large; /* Derived. */
+	uint64_t		nfills_large; /* Derived. */
+	uint64_t		nflushes_large; /* Derived. */
+	uint64_t		nrequests_large; /* Derived. */
+
+	/*
+	 * The stats logically owned by the pa_shard in the same arena.  This
+	 * lives here only because it's convenient for the purposes of the ctl
+	 * module -- it only knows about the single arena_stats.
+	 */
+	pa_shard_stats_t	pa_shard_stats;
+
+	/* Number of bytes cached in tcache associated with this arena. */
+	size_t			tcache_bytes; /* Derived. */
+	size_t			tcache_stashed_bytes; /* Derived. */
+
+	mutex_prof_data_t mutex_prof_data[mutex_prof_num_arena_mutexes];
+
+	/* One element for each large size class. */
+	arena_stats_large_t	lstats[SC_NSIZES - SC_NBINS];
+
+	/* Arena uptime. */
+	nstime_t		uptime;
+};
+
+static inline bool
+arena_stats_init(tsdn_t *tsdn, arena_stats_t *arena_stats) {
+	if (config_debug) {
+		for (size_t i = 0; i < sizeof(arena_stats_t); i++) {
+			assert(((char *)arena_stats)[i] == 0);
+		}
+	}
+	if (LOCKEDINT_MTX_INIT(arena_stats->mtx, "arena_stats",
+	    WITNESS_RANK_ARENA_STATS, malloc_mutex_rank_exclusive)) {
+		return true;
+	}
+	/* Memory is zeroed, so there is no need to clear stats. */
+	return false;
+}
+
+static inline void
+arena_stats_large_flush_nrequests_add(tsdn_t *tsdn, arena_stats_t *arena_stats,
+    szind_t szind, uint64_t nrequests) {
+	LOCKEDINT_MTX_LOCK(tsdn, arena_stats->mtx);
+	arena_stats_large_t *lstats = &arena_stats->lstats[szind - SC_NBINS];
+	locked_inc_u64(tsdn, LOCKEDINT_MTX(arena_stats->mtx),
+	    &lstats->nrequests, nrequests);
+	locked_inc_u64(tsdn, LOCKEDINT_MTX(arena_stats->mtx),
+	    &lstats->nflushes, 1);
+	LOCKEDINT_MTX_UNLOCK(tsdn, arena_stats->mtx);
+}
+
+#endif /* JEMALLOC_INTERNAL_ARENA_STATS_H */

BeefRT/JEMalloc/include/jemalloc/internal/arena_structs.h

@@ -0,0 +1,101 @@
+#ifndef JEMALLOC_INTERNAL_ARENA_STRUCTS_H
+#define JEMALLOC_INTERNAL_ARENA_STRUCTS_H
+
+#include "jemalloc/internal/arena_stats.h"
+#include "jemalloc/internal/atomic.h"
+#include "jemalloc/internal/bin.h"
+#include "jemalloc/internal/bitmap.h"
+#include "jemalloc/internal/counter.h"
+#include "jemalloc/internal/ecache.h"
+#include "jemalloc/internal/edata_cache.h"
+#include "jemalloc/internal/extent_dss.h"
+#include "jemalloc/internal/jemalloc_internal_types.h"
+#include "jemalloc/internal/mutex.h"
+#include "jemalloc/internal/nstime.h"
+#include "jemalloc/internal/pa.h"
+#include "jemalloc/internal/ql.h"
+#include "jemalloc/internal/sc.h"
+#include "jemalloc/internal/ticker.h"
+
+struct arena_s {
+	/*
+	 * Number of threads currently assigned to this arena.  Each thread has
+	 * two distinct assignments, one for application-serving allocation, and
+	 * the other for internal metadata allocation.  Internal metadata must
+	 * not be allocated from arenas explicitly created via the arenas.create
+	 * mallctl, because the arena.<i>.reset mallctl indiscriminately
+	 * discards all allocations for the affected arena.
+	 *
+	 *   0: Application allocation.
+	 *   1: Internal metadata allocation.
+	 *
+	 * Synchronization: atomic.
+	 */
+	atomic_u_t		nthreads[2];
+
+	/* Next bin shard for binding new threads. Synchronization: atomic. */
+	atomic_u_t		binshard_next;
+
+	/*
+	 * When percpu_arena is enabled, to amortize the cost of reading /
+	 * updating the current CPU id, track the most recent thread accessing
+	 * this arena, and only read CPU if there is a mismatch.
+	 */
+	tsdn_t		*last_thd;
+
+	/* Synchronization: internal. */
+	arena_stats_t		stats;
+
+	/*
+	 * Lists of tcaches and cache_bin_array_descriptors for extant threads
+	 * associated with this arena.  Stats from these are merged
+	 * incrementally, and at exit if opt_stats_print is enabled.
+	 *
+	 * Synchronization: tcache_ql_mtx.
+	 */
+	ql_head(tcache_slow_t)			tcache_ql;
+	ql_head(cache_bin_array_descriptor_t)	cache_bin_array_descriptor_ql;
+	malloc_mutex_t				tcache_ql_mtx;
+
+	/*
+	 * Represents a dss_prec_t, but atomically.
+	 *
+	 * Synchronization: atomic.
+	 */
+	atomic_u_t		dss_prec;
+
+	/*
+	 * Extant large allocations.
+	 *
+	 * Synchronization: large_mtx.
+	 */
+	edata_list_active_t	large;
+	/* Synchronizes all large allocation/update/deallocation. */
+	malloc_mutex_t		large_mtx;
+
+	/* The page-level allocator shard this arena uses. */
+	pa_shard_t		pa_shard;
+
+	/*
+	 * A cached copy of base->ind.  This can get accessed on hot paths;
+	 * looking it up in base requires an extra pointer hop / cache miss.
+	 */
+	unsigned ind;
+
+	/*
+	 * Base allocator, from which arena metadata are allocated.
+	 *
+	 * Synchronization: internal.
+	 */
+	base_t			*base;
+	/* Used to determine uptime.  Read-only after initialization. */
+	nstime_t		create_time;
+
+	/*
+	 * The arena is allocated alongside its bins; really this is a
+	 * dynamically sized array determined by the binshard settings.
+	 */
+	bin_t			bins[0];
+};
+
+#endif /* JEMALLOC_INTERNAL_ARENA_STRUCTS_H */

BeefRT/JEMalloc/include/jemalloc/internal/arena_types.h

@@ -0,0 +1,58 @@
+#ifndef JEMALLOC_INTERNAL_ARENA_TYPES_H
+#define JEMALLOC_INTERNAL_ARENA_TYPES_H
+
+#include "jemalloc/internal/sc.h"
+
+/* Default decay times in milliseconds. */
+#define DIRTY_DECAY_MS_DEFAULT	ZD(10 * 1000)
+#define MUZZY_DECAY_MS_DEFAULT	(0)
+/* Number of event ticks between time checks. */
+#define ARENA_DECAY_NTICKS_PER_UPDATE	1000
+
+typedef struct arena_decay_s arena_decay_t;
+typedef struct arena_s arena_t;
+
+typedef enum {
+	percpu_arena_mode_names_base   = 0, /* Used for options processing. */
+
+	/*
+	 * *_uninit are used only during bootstrapping, and must correspond
+	 * to initialized variant plus percpu_arena_mode_enabled_base.
+	 */
+	percpu_arena_uninit            = 0,
+	per_phycpu_arena_uninit        = 1,
+
+	/* All non-disabled modes must come after percpu_arena_disabled. */
+	percpu_arena_disabled          = 2,
+
+	percpu_arena_mode_names_limit  = 3, /* Used for options processing. */
+	percpu_arena_mode_enabled_base = 3,
+
+	percpu_arena                   = 3,
+	per_phycpu_arena               = 4  /* Hyper threads share arena. */
+} percpu_arena_mode_t;
+
+#define PERCPU_ARENA_ENABLED(m)	((m) >= percpu_arena_mode_enabled_base)
+#define PERCPU_ARENA_DEFAULT	percpu_arena_disabled
+
+/*
+ * When allocation_size >= oversize_threshold, use the dedicated huge arena
+ * (unless have explicitly spicified arena index).  0 disables the feature.
+ */
+#define OVERSIZE_THRESHOLD_DEFAULT (8 << 20)
+
+struct arena_config_s {
+	/* extent hooks to be used for the arena */
+	extent_hooks_t *extent_hooks;
+
+	/*
+	 * Use extent hooks for metadata (base) allocations when true.
+	 */
+	bool metadata_use_hooks;
+};
+
+typedef struct arena_config_s arena_config_t;
+
+extern const arena_config_t arena_config_default;
+
+#endif /* JEMALLOC_INTERNAL_ARENA_TYPES_H */

BeefRT/JEMalloc/include/jemalloc/internal/assert.h

@@ -0,0 +1,56 @@
+#include "jemalloc/internal/malloc_io.h"
+#include "jemalloc/internal/util.h"
+
+/*
+ * Define a custom assert() in order to reduce the chances of deadlock during
+ * assertion failure.
+ */
+#ifndef assert
+#define assert(e) do {							\
+	if (unlikely(config_debug && !(e))) {				\
+		malloc_printf(						\
+		    "<jemalloc>: %s:%d: Failed assertion: \"%s\"\n",	\
+		    __FILE__, __LINE__, #e);				\
+		abort();						\
+	}								\
+} while (0)
+#endif
+
+#ifndef not_reached
+#define not_reached() do {						\
+	if (config_debug) {						\
+		malloc_printf(						\
+		    "<jemalloc>: %s:%d: Unreachable code reached\n",	\
+		    __FILE__, __LINE__);				\
+		abort();						\
+	}								\
+	unreachable();							\
+} while (0)
+#endif
+
+#ifndef not_implemented
+#define not_implemented() do {						\
+	if (config_debug) {						\
+		malloc_printf("<jemalloc>: %s:%d: Not implemented\n",	\
+		    __FILE__, __LINE__);				\
+		abort();						\
+	}								\
+} while (0)
+#endif
+
+#ifndef assert_not_implemented
+#define assert_not_implemented(e) do {					\
+	if (unlikely(config_debug && !(e))) {				\
+		not_implemented();					\
+	}								\
+} while (0)
+#endif
+
+/* Use to assert a particular configuration, e.g., cassert(config_debug). */
+#ifndef cassert
+#define cassert(c) do {							\
+	if (unlikely(!(c))) {						\
+		not_reached();						\
+	}								\
+} while (0)
+#endif

BeefRT/JEMalloc/include/jemalloc/internal/atomic.h

@@ -0,0 +1,107 @@
+#ifndef JEMALLOC_INTERNAL_ATOMIC_H
+#define JEMALLOC_INTERNAL_ATOMIC_H
+
+#define ATOMIC_INLINE JEMALLOC_ALWAYS_INLINE
+
+#define JEMALLOC_U8_ATOMICS
+#if defined(JEMALLOC_GCC_ATOMIC_ATOMICS)
+#  include "jemalloc/internal/atomic_gcc_atomic.h"
+#  if !defined(JEMALLOC_GCC_U8_ATOMIC_ATOMICS)
+#    undef JEMALLOC_U8_ATOMICS
+#  endif
+#elif defined(JEMALLOC_GCC_SYNC_ATOMICS)
+#  include "jemalloc/internal/atomic_gcc_sync.h"
+#  if !defined(JEMALLOC_GCC_U8_SYNC_ATOMICS)
+#    undef JEMALLOC_U8_ATOMICS
+#  endif
+#elif defined(_MSC_VER)
+#  include "jemalloc/internal/atomic_msvc.h"
+#elif defined(JEMALLOC_C11_ATOMICS)
+#  include "jemalloc/internal/atomic_c11.h"
+#else
+#  error "Don't have atomics implemented on this platform."
+#endif
+
+/*
+ * This header gives more or less a backport of C11 atomics. The user can write
+ * JEMALLOC_GENERATE_ATOMICS(type, short_type, lg_sizeof_type); to generate
+ * counterparts of the C11 atomic functions for type, as so:
+ *   JEMALLOC_GENERATE_ATOMICS(int *, pi, 3);
+ * and then write things like:
+ *   int *some_ptr;
+ *   atomic_pi_t atomic_ptr_to_int;
+ *   atomic_store_pi(&atomic_ptr_to_int, some_ptr, ATOMIC_RELAXED);
+ *   int *prev_value = atomic_exchange_pi(&ptr_to_int, NULL, ATOMIC_ACQ_REL);
+ *   assert(some_ptr == prev_value);
+ * and expect things to work in the obvious way.
+ *
+ * Also included (with naming differences to avoid conflicts with the standard
+ * library):
+ *   atomic_fence(atomic_memory_order_t) (mimics C11's atomic_thread_fence).
+ *   ATOMIC_INIT (mimics C11's ATOMIC_VAR_INIT).
+ */
+
+/*
+ * Pure convenience, so that we don't have to type "atomic_memory_order_"
+ * quite so often.
+ */
+#define ATOMIC_RELAXED atomic_memory_order_relaxed
+#define ATOMIC_ACQUIRE atomic_memory_order_acquire
+#define ATOMIC_RELEASE atomic_memory_order_release
+#define ATOMIC_ACQ_REL atomic_memory_order_acq_rel
+#define ATOMIC_SEQ_CST atomic_memory_order_seq_cst
+
+/*
+ * Another convenience -- simple atomic helper functions.
+ */
+#define JEMALLOC_GENERATE_EXPANDED_INT_ATOMICS(type, short_type,	\
+    lg_size)								\
+    JEMALLOC_GENERATE_INT_ATOMICS(type, short_type, lg_size)		\
+    ATOMIC_INLINE void							\
+    atomic_load_add_store_##short_type(atomic_##short_type##_t *a,	\
+	type inc) {							\
+	    type oldval = atomic_load_##short_type(a, ATOMIC_RELAXED);	\
+	    type newval = oldval + inc;					\
+	    atomic_store_##short_type(a, newval, ATOMIC_RELAXED);	\
+	}								\
+    ATOMIC_INLINE void							\
+    atomic_load_sub_store_##short_type(atomic_##short_type##_t *a,	\
+	type inc) {							\
+	    type oldval = atomic_load_##short_type(a, ATOMIC_RELAXED);	\
+	    type newval = oldval - inc;					\
+	    atomic_store_##short_type(a, newval, ATOMIC_RELAXED);	\
+	}
+
+/*
+ * Not all platforms have 64-bit atomics.  If we do, this #define exposes that
+ * fact.
+ */
+#if (LG_SIZEOF_PTR == 3 || LG_SIZEOF_INT == 3)
+#  define JEMALLOC_ATOMIC_U64
+#endif
+
+JEMALLOC_GENERATE_ATOMICS(void *, p, LG_SIZEOF_PTR)
+
+/*
+ * There's no actual guarantee that sizeof(bool) == 1, but it's true on the only
+ * platform that actually needs to know the size, MSVC.
+ */
+JEMALLOC_GENERATE_ATOMICS(bool, b, 0)
+
+JEMALLOC_GENERATE_EXPANDED_INT_ATOMICS(unsigned, u, LG_SIZEOF_INT)
+
+JEMALLOC_GENERATE_EXPANDED_INT_ATOMICS(size_t, zu, LG_SIZEOF_PTR)
+
+JEMALLOC_GENERATE_EXPANDED_INT_ATOMICS(ssize_t, zd, LG_SIZEOF_PTR)
+
+JEMALLOC_GENERATE_EXPANDED_INT_ATOMICS(uint8_t, u8, 0)
+
+JEMALLOC_GENERATE_EXPANDED_INT_ATOMICS(uint32_t, u32, 2)
+
+#ifdef JEMALLOC_ATOMIC_U64
+JEMALLOC_GENERATE_EXPANDED_INT_ATOMICS(uint64_t, u64, 3)
+#endif
+
+#undef ATOMIC_INLINE
+
+#endif /* JEMALLOC_INTERNAL_ATOMIC_H */

BeefRT/JEMalloc/include/jemalloc/internal/atomic_c11.h

@@ -0,0 +1,97 @@
+#ifndef JEMALLOC_INTERNAL_ATOMIC_C11_H
+#define JEMALLOC_INTERNAL_ATOMIC_C11_H
+
+#include <stdatomic.h>
+
+#define ATOMIC_INIT(...) ATOMIC_VAR_INIT(__VA_ARGS__)
+
+#define atomic_memory_order_t memory_order
+#define atomic_memory_order_relaxed memory_order_relaxed
+#define atomic_memory_order_acquire memory_order_acquire
+#define atomic_memory_order_release memory_order_release
+#define atomic_memory_order_acq_rel memory_order_acq_rel
+#define atomic_memory_order_seq_cst memory_order_seq_cst
+
+#define atomic_fence atomic_thread_fence
+
+#define JEMALLOC_GENERATE_ATOMICS(type, short_type,			\
+    /* unused */ lg_size)						\
+typedef _Atomic(type) atomic_##short_type##_t;				\
+									\
+ATOMIC_INLINE type							\
+atomic_load_##short_type(const atomic_##short_type##_t *a,		\
+    atomic_memory_order_t mo) {						\
+	/*								\
+	 * A strict interpretation of the C standard prevents		\
+	 * atomic_load from taking a const argument, but it's		\
+	 * convenient for our purposes. This cast is a workaround.	\
+	 */								\
+	atomic_##short_type##_t* a_nonconst =				\
+	    (atomic_##short_type##_t*)a;				\
+	return atomic_load_explicit(a_nonconst, mo);			\
+}									\
+									\
+ATOMIC_INLINE void							\
+atomic_store_##short_type(atomic_##short_type##_t *a,			\
+    type val, atomic_memory_order_t mo) {				\
+	atomic_store_explicit(a, val, mo);				\
+}									\
+									\
+ATOMIC_INLINE type							\
+atomic_exchange_##short_type(atomic_##short_type##_t *a, type val,	\
+    atomic_memory_order_t mo) {						\
+	return atomic_exchange_explicit(a, val, mo);			\
+}									\
+									\
+ATOMIC_INLINE bool							\
+atomic_compare_exchange_weak_##short_type(atomic_##short_type##_t *a,	\
+    type *expected, type desired, atomic_memory_order_t success_mo,	\
+    atomic_memory_order_t failure_mo) {					\
+	return atomic_compare_exchange_weak_explicit(a, expected,	\
+	    desired, success_mo, failure_mo);				\
+}									\
+									\
+ATOMIC_INLINE bool							\
+atomic_compare_exchange_strong_##short_type(atomic_##short_type##_t *a,	\
+    type *expected, type desired, atomic_memory_order_t success_mo,	\
+    atomic_memory_order_t failure_mo) {					\
+	return atomic_compare_exchange_strong_explicit(a, expected,	\
+	    desired, success_mo, failure_mo);				\
+}
+
+/*
+ * Integral types have some special operations available that non-integral ones
+ * lack.
+ */
+#define JEMALLOC_GENERATE_INT_ATOMICS(type, short_type, 		\
+    /* unused */ lg_size)						\
+JEMALLOC_GENERATE_ATOMICS(type, short_type, /* unused */ lg_size)	\
+									\
+ATOMIC_INLINE type							\
+atomic_fetch_add_##short_type(atomic_##short_type##_t *a,		\
+    type val, atomic_memory_order_t mo) {				\
+	return atomic_fetch_add_explicit(a, val, mo);			\
+}									\
+									\
+ATOMIC_INLINE type							\
+atomic_fetch_sub_##short_type(atomic_##short_type##_t *a,		\
+    type val, atomic_memory_order_t mo) {				\
+	return atomic_fetch_sub_explicit(a, val, mo);			\
+}									\
+ATOMIC_INLINE type							\
+atomic_fetch_and_##short_type(atomic_##short_type##_t *a,		\
+    type val, atomic_memory_order_t mo) {				\
+	return atomic_fetch_and_explicit(a, val, mo);			\
+}									\
+ATOMIC_INLINE type							\
+atomic_fetch_or_##short_type(atomic_##short_type##_t *a,		\
+    type val, atomic_memory_order_t mo) {				\
+	return atomic_fetch_or_explicit(a, val, mo);			\
+}									\
+ATOMIC_INLINE type							\
+atomic_fetch_xor_##short_type(atomic_##short_type##_t *a,		\
+    type val, atomic_memory_order_t mo) {				\
+	return atomic_fetch_xor_explicit(a, val, mo);			\
+}
+
+#endif /* JEMALLOC_INTERNAL_ATOMIC_C11_H */

BeefRT/JEMalloc/include/jemalloc/internal/atomic_gcc_atomic.h

@@ -0,0 +1,129 @@
+#ifndef JEMALLOC_INTERNAL_ATOMIC_GCC_ATOMIC_H
+#define JEMALLOC_INTERNAL_ATOMIC_GCC_ATOMIC_H
+
+#include "jemalloc/internal/assert.h"
+
+#define ATOMIC_INIT(...) {__VA_ARGS__}
+
+typedef enum {
+	atomic_memory_order_relaxed,
+	atomic_memory_order_acquire,
+	atomic_memory_order_release,
+	atomic_memory_order_acq_rel,
+	atomic_memory_order_seq_cst
+} atomic_memory_order_t;
+
+ATOMIC_INLINE int
+atomic_enum_to_builtin(atomic_memory_order_t mo) {
+	switch (mo) {
+	case atomic_memory_order_relaxed:
+		return __ATOMIC_RELAXED;
+	case atomic_memory_order_acquire:
+		return __ATOMIC_ACQUIRE;
+	case atomic_memory_order_release:
+		return __ATOMIC_RELEASE;
+	case atomic_memory_order_acq_rel:
+		return __ATOMIC_ACQ_REL;
+	case atomic_memory_order_seq_cst:
+		return __ATOMIC_SEQ_CST;
+	}
+	/* Can't happen; the switch is exhaustive. */
+	not_reached();
+}
+
+ATOMIC_INLINE void
+atomic_fence(atomic_memory_order_t mo) {
+	__atomic_thread_fence(atomic_enum_to_builtin(mo));
+}
+
+#define JEMALLOC_GENERATE_ATOMICS(type, short_type,			\
+    /* unused */ lg_size)						\
+typedef struct {							\
+	type repr;							\
+} atomic_##short_type##_t;						\
+									\
+ATOMIC_INLINE type							\
+atomic_load_##short_type(const atomic_##short_type##_t *a,		\
+    atomic_memory_order_t mo) {						\
+	type result;							\
+	__atomic_load(&a->repr, &result, atomic_enum_to_builtin(mo));	\
+	return result;							\
+}									\
+									\
+ATOMIC_INLINE void							\
+atomic_store_##short_type(atomic_##short_type##_t *a, type val,		\
+    atomic_memory_order_t mo) {						\
+	__atomic_store(&a->repr, &val, atomic_enum_to_builtin(mo));	\
+}									\
+									\
+ATOMIC_INLINE type							\
+atomic_exchange_##short_type(atomic_##short_type##_t *a, type val,	\
+    atomic_memory_order_t mo) {						\
+	type result;							\
+	__atomic_exchange(&a->repr, &val, &result,			\
+	    atomic_enum_to_builtin(mo));				\
+	return result;							\
+}									\
+									\
+ATOMIC_INLINE bool							\
+atomic_compare_exchange_weak_##short_type(atomic_##short_type##_t *a,	\
+    UNUSED type *expected, type desired,				\
+    atomic_memory_order_t success_mo,					\
+    atomic_memory_order_t failure_mo) {					\
+	return __atomic_compare_exchange(&a->repr, expected, &desired,	\
+	    true, atomic_enum_to_builtin(success_mo),			\
+	    atomic_enum_to_builtin(failure_mo));			\
+}									\
+									\
+ATOMIC_INLINE bool							\
+atomic_compare_exchange_strong_##short_type(atomic_##short_type##_t *a,	\
+    UNUSED type *expected, type desired,				\
+    atomic_memory_order_t success_mo,					\
+    atomic_memory_order_t failure_mo) {					\
+	return __atomic_compare_exchange(&a->repr, expected, &desired,	\
+	    false,							\
+	    atomic_enum_to_builtin(success_mo),				\
+	    atomic_enum_to_builtin(failure_mo));			\
+}
+
+
+#define JEMALLOC_GENERATE_INT_ATOMICS(type, short_type,			\
+    /* unused */ lg_size)						\
+JEMALLOC_GENERATE_ATOMICS(type, short_type, /* unused */ lg_size)	\
+									\
+ATOMIC_INLINE type							\
+atomic_fetch_add_##short_type(atomic_##short_type##_t *a, type val,	\
+    atomic_memory_order_t mo) {						\
+	return __atomic_fetch_add(&a->repr, val,			\
+	    atomic_enum_to_builtin(mo));				\
+}									\
+									\
+ATOMIC_INLINE type							\
+atomic_fetch_sub_##short_type(atomic_##short_type##_t *a, type val,	\
+    atomic_memory_order_t mo) {						\
+	return __atomic_fetch_sub(&a->repr, val,			\
+	    atomic_enum_to_builtin(mo));				\
+}									\
+									\
+ATOMIC_INLINE type							\
+atomic_fetch_and_##short_type(atomic_##short_type##_t *a, type val,	\
+    atomic_memory_order_t mo) {						\
+	return __atomic_fetch_and(&a->repr, val,			\
+	    atomic_enum_to_builtin(mo));				\
+}									\
+									\
+ATOMIC_INLINE type							\
+atomic_fetch_or_##short_type(atomic_##short_type##_t *a, type val,	\
+    atomic_memory_order_t mo) {						\
+	return __atomic_fetch_or(&a->repr, val,				\
+	    atomic_enum_to_builtin(mo));				\
+}									\
+									\
+ATOMIC_INLINE type							\
+atomic_fetch_xor_##short_type(atomic_##short_type##_t *a, type val,	\
+    atomic_memory_order_t mo) {						\
+	return __atomic_fetch_xor(&a->repr, val,			\
+	    atomic_enum_to_builtin(mo));				\
+}
+
+#endif /* JEMALLOC_INTERNAL_ATOMIC_GCC_ATOMIC_H */

BeefRT/JEMalloc/include/jemalloc/internal/atomic_gcc_sync.h

@@ -0,0 +1,195 @@
+#ifndef JEMALLOC_INTERNAL_ATOMIC_GCC_SYNC_H
+#define JEMALLOC_INTERNAL_ATOMIC_GCC_SYNC_H
+
+#define ATOMIC_INIT(...) {__VA_ARGS__}
+
+typedef enum {
+	atomic_memory_order_relaxed,
+	atomic_memory_order_acquire,
+	atomic_memory_order_release,
+	atomic_memory_order_acq_rel,
+	atomic_memory_order_seq_cst
+} atomic_memory_order_t;
+
+ATOMIC_INLINE void
+atomic_fence(atomic_memory_order_t mo) {
+	/* Easy cases first: no barrier, and full barrier. */
+	if (mo == atomic_memory_order_relaxed) {
+		asm volatile("" ::: "memory");
+		return;
+	}
+	if (mo == atomic_memory_order_seq_cst) {
+		asm volatile("" ::: "memory");
+		__sync_synchronize();
+		asm volatile("" ::: "memory");
+		return;
+	}
+	asm volatile("" ::: "memory");
+#  if defined(__i386__) || defined(__x86_64__)
+	/* This is implicit on x86. */
+#  elif defined(__ppc64__)
+	asm volatile("lwsync");
+#  elif defined(__ppc__)
+	asm volatile("sync");
+#  elif defined(__sparc__) && defined(__arch64__)
+	if (mo == atomic_memory_order_acquire) {
+		asm volatile("membar #LoadLoad | #LoadStore");
+	} else if (mo == atomic_memory_order_release) {
+		asm volatile("membar #LoadStore | #StoreStore");
+	} else {
+		asm volatile("membar #LoadLoad | #LoadStore | #StoreStore");
+	}
+#  else
+	__sync_synchronize();
+#  endif
+	asm volatile("" ::: "memory");
+}
+
+/*
+ * A correct implementation of seq_cst loads and stores on weakly ordered
+ * architectures could do either of the following:
+ *   1. store() is weak-fence -> store -> strong fence, load() is load ->
+ *      strong-fence.
+ *   2. store() is strong-fence -> store, load() is strong-fence -> load ->
+ *      weak-fence.
+ * The tricky thing is, load() and store() above can be the load or store
+ * portions of a gcc __sync builtin, so we have to follow GCC's lead, which
+ * means going with strategy 2.
+ * On strongly ordered architectures, the natural strategy is to stick a strong
+ * fence after seq_cst stores, and have naked loads.  So we want the strong
+ * fences in different places on different architectures.
+ * atomic_pre_sc_load_fence and atomic_post_sc_store_fence allow us to
+ * accomplish this.
+ */
+
+ATOMIC_INLINE void
+atomic_pre_sc_load_fence() {
+#  if defined(__i386__) || defined(__x86_64__) ||			\
+    (defined(__sparc__) && defined(__arch64__))
+	atomic_fence(atomic_memory_order_relaxed);
+#  else
+	atomic_fence(atomic_memory_order_seq_cst);
+#  endif
+}
+
+ATOMIC_INLINE void
+atomic_post_sc_store_fence() {
+#  if defined(__i386__) || defined(__x86_64__) ||			\
+    (defined(__sparc__) && defined(__arch64__))
+	atomic_fence(atomic_memory_order_seq_cst);
+#  else
+	atomic_fence(atomic_memory_order_relaxed);
+#  endif
+
+}
+
+#define JEMALLOC_GENERATE_ATOMICS(type, short_type,			\
+    /* unused */ lg_size)						\
+typedef struct {							\
+	type volatile repr;						\
+} atomic_##short_type##_t;						\
+									\
+ATOMIC_INLINE type							\
+atomic_load_##short_type(const atomic_##short_type##_t *a,		\
+    atomic_memory_order_t mo) {						\
+	if (mo == atomic_memory_order_seq_cst) {			\
+		atomic_pre_sc_load_fence();				\
+	}								\
+	type result = a->repr;						\
+	if (mo != atomic_memory_order_relaxed) {			\
+		atomic_fence(atomic_memory_order_acquire);		\
+	}								\
+	return result;							\
+}									\
+									\
+ATOMIC_INLINE void							\
+atomic_store_##short_type(atomic_##short_type##_t *a,			\
+    type val, atomic_memory_order_t mo) {				\
+	if (mo != atomic_memory_order_relaxed) {			\
+		atomic_fence(atomic_memory_order_release);		\
+	}								\
+	a->repr = val;							\
+	if (mo == atomic_memory_order_seq_cst) {			\
+		atomic_post_sc_store_fence();				\
+	}								\
+}									\
+									\
+ATOMIC_INLINE type							\
+atomic_exchange_##short_type(atomic_##short_type##_t *a, type val, \
+    atomic_memory_order_t mo) {                  					 \
+	/*								\
+	 * Because of FreeBSD, we care about gcc 4.2, which doesn't have\
+	 * an atomic exchange builtin.  We fake it with a CAS loop.	\
+	 */								\
+	while (true) {							\
+		type old = a->repr;					\
+		if (__sync_bool_compare_and_swap(&a->repr, old, val)) {	\
+			return old;					\
+		}							\
+	}								\
+}									\
+									\
+ATOMIC_INLINE bool							\
+atomic_compare_exchange_weak_##short_type(atomic_##short_type##_t *a,	\
+    type *expected, type desired,                                     \
+    atomic_memory_order_t success_mo,                          \
+    atomic_memory_order_t failure_mo) {				                \
+	type prev = __sync_val_compare_and_swap(&a->repr, *expected,	\
+	    desired);							\
+	if (prev == *expected) {					\
+		return true;						\
+	} else {							\
+		*expected = prev;					\
+		return false;						\
+	}								\
+}									\
+ATOMIC_INLINE bool							\
+atomic_compare_exchange_strong_##short_type(atomic_##short_type##_t *a,	\
+    type *expected, type desired,                                       \
+    atomic_memory_order_t success_mo,                            \
+    atomic_memory_order_t failure_mo) {                          \
+	type prev = __sync_val_compare_and_swap(&a->repr, *expected,	\
+	    desired);							\
+	if (prev == *expected) {					\
+		return true;						\
+	} else {							\
+		*expected = prev;					\
+		return false;						\
+	}								\
+}
+
+#define JEMALLOC_GENERATE_INT_ATOMICS(type, short_type,			\
+    /* unused */ lg_size)						\
+JEMALLOC_GENERATE_ATOMICS(type, short_type, /* unused */ lg_size)	\
+									\
+ATOMIC_INLINE type							\
+atomic_fetch_add_##short_type(atomic_##short_type##_t *a, type val,	\
+    atomic_memory_order_t mo) {						\
+	return __sync_fetch_and_add(&a->repr, val);			\
+}									\
+									\
+ATOMIC_INLINE type							\
+atomic_fetch_sub_##short_type(atomic_##short_type##_t *a, type val,	\
+    atomic_memory_order_t mo) {						\
+	return __sync_fetch_and_sub(&a->repr, val);			\
+}									\
+									\
+ATOMIC_INLINE type							\
+atomic_fetch_and_##short_type(atomic_##short_type##_t *a, type val,	\
+    atomic_memory_order_t mo) {						\
+	return __sync_fetch_and_and(&a->repr, val);			\
+}									\
+									\
+ATOMIC_INLINE type							\
+atomic_fetch_or_##short_type(atomic_##short_type##_t *a, type val,	\
+    atomic_memory_order_t mo) {						\
+	return __sync_fetch_and_or(&a->repr, val);			\
+}									\
+									\
+ATOMIC_INLINE type							\
+atomic_fetch_xor_##short_type(atomic_##short_type##_t *a, type val,	\
+    atomic_memory_order_t mo) {						\
+	return __sync_fetch_and_xor(&a->repr, val);			\
+}
+
+#endif /* JEMALLOC_INTERNAL_ATOMIC_GCC_SYNC_H */

BeefRT/JEMalloc/include/jemalloc/internal/atomic_msvc.h

@@ -0,0 +1,158 @@
+#ifndef JEMALLOC_INTERNAL_ATOMIC_MSVC_H
+#define JEMALLOC_INTERNAL_ATOMIC_MSVC_H
+
+#define ATOMIC_INIT(...) {__VA_ARGS__}
+
+typedef enum {
+	atomic_memory_order_relaxed,
+	atomic_memory_order_acquire,
+	atomic_memory_order_release,
+	atomic_memory_order_acq_rel,
+	atomic_memory_order_seq_cst
+} atomic_memory_order_t;
+
+typedef char atomic_repr_0_t;
+typedef short atomic_repr_1_t;
+typedef long atomic_repr_2_t;
+typedef __int64 atomic_repr_3_t;
+
+ATOMIC_INLINE void
+atomic_fence(atomic_memory_order_t mo) {
+	_ReadWriteBarrier();
+#  if defined(_M_ARM) || defined(_M_ARM64)
+	/* ARM needs a barrier for everything but relaxed. */
+	if (mo != atomic_memory_order_relaxed) {
+		MemoryBarrier();
+	}
+#  elif defined(_M_IX86) || defined (_M_X64)
+	/* x86 needs a barrier only for seq_cst. */
+	if (mo == atomic_memory_order_seq_cst) {
+		MemoryBarrier();
+	}
+#  else
+#  error "Don't know how to create atomics for this platform for MSVC."
+#  endif
+	_ReadWriteBarrier();
+}
+
+#define ATOMIC_INTERLOCKED_REPR(lg_size) atomic_repr_ ## lg_size ## _t
+
+#define ATOMIC_CONCAT(a, b) ATOMIC_RAW_CONCAT(a, b)
+#define ATOMIC_RAW_CONCAT(a, b) a ## b
+
+#define ATOMIC_INTERLOCKED_NAME(base_name, lg_size) ATOMIC_CONCAT(	\
+    base_name, ATOMIC_INTERLOCKED_SUFFIX(lg_size))
+
+#define ATOMIC_INTERLOCKED_SUFFIX(lg_size)				\
+    ATOMIC_CONCAT(ATOMIC_INTERLOCKED_SUFFIX_, lg_size)
+
+#define ATOMIC_INTERLOCKED_SUFFIX_0 8
+#define ATOMIC_INTERLOCKED_SUFFIX_1 16
+#define ATOMIC_INTERLOCKED_SUFFIX_2
+#define ATOMIC_INTERLOCKED_SUFFIX_3 64
+
+#define JEMALLOC_GENERATE_ATOMICS(type, short_type, lg_size)		\
+typedef struct {							\
+	ATOMIC_INTERLOCKED_REPR(lg_size) repr;				\
+} atomic_##short_type##_t;						\
+									\
+ATOMIC_INLINE type							\
+atomic_load_##short_type(const atomic_##short_type##_t *a,		\
+    atomic_memory_order_t mo) {						\
+	ATOMIC_INTERLOCKED_REPR(lg_size) ret = a->repr;			\
+	if (mo != atomic_memory_order_relaxed) {			\
+		atomic_fence(atomic_memory_order_acquire);		\
+	}								\
+	return (type) ret;						\
+}									\
+									\
+ATOMIC_INLINE void							\
+atomic_store_##short_type(atomic_##short_type##_t *a,			\
+    type val, atomic_memory_order_t mo) {				\
+	if (mo != atomic_memory_order_relaxed) {			\
+		atomic_fence(atomic_memory_order_release);		\
+	}								\
+	a->repr = (ATOMIC_INTERLOCKED_REPR(lg_size)) val;		\
+	if (mo == atomic_memory_order_seq_cst) {			\
+		atomic_fence(atomic_memory_order_seq_cst);		\
+	}								\
+}									\
+									\
+ATOMIC_INLINE type							\
+atomic_exchange_##short_type(atomic_##short_type##_t *a, type val,	\
+    atomic_memory_order_t mo) {						\
+	return (type)ATOMIC_INTERLOCKED_NAME(_InterlockedExchange,	\
+	    lg_size)(&a->repr, (ATOMIC_INTERLOCKED_REPR(lg_size))val);	\
+}									\
+									\
+ATOMIC_INLINE bool							\
+atomic_compare_exchange_weak_##short_type(atomic_##short_type##_t *a,	\
+    type *expected, type desired, atomic_memory_order_t success_mo,	\
+    atomic_memory_order_t failure_mo) {					\
+	ATOMIC_INTERLOCKED_REPR(lg_size) e =				\
+	    (ATOMIC_INTERLOCKED_REPR(lg_size))*expected;		\
+	ATOMIC_INTERLOCKED_REPR(lg_size) d =				\
+	    (ATOMIC_INTERLOCKED_REPR(lg_size))desired;			\
+	ATOMIC_INTERLOCKED_REPR(lg_size) old =				\
+	    ATOMIC_INTERLOCKED_NAME(_InterlockedCompareExchange, 	\
+		lg_size)(&a->repr, d, e);				\
+	if (old == e) {							\
+		return true;						\
+	} else {							\
+		*expected = (type)old;					\
+		return false;						\
+	}								\
+}									\
+									\
+ATOMIC_INLINE bool							\
+atomic_compare_exchange_strong_##short_type(atomic_##short_type##_t *a,	\
+    type *expected, type desired, atomic_memory_order_t success_mo,	\
+    atomic_memory_order_t failure_mo) {					\
+	/* We implement the weak version with strong semantics. */	\
+	return atomic_compare_exchange_weak_##short_type(a, expected,	\
+	    desired, success_mo, failure_mo);				\
+}
+
+
+#define JEMALLOC_GENERATE_INT_ATOMICS(type, short_type, lg_size)	\
+JEMALLOC_GENERATE_ATOMICS(type, short_type, lg_size)			\
+									\
+ATOMIC_INLINE type							\
+atomic_fetch_add_##short_type(atomic_##short_type##_t *a,		\
+    type val, atomic_memory_order_t mo) {				\
+	return (type)ATOMIC_INTERLOCKED_NAME(_InterlockedExchangeAdd,	\
+	    lg_size)(&a->repr, (ATOMIC_INTERLOCKED_REPR(lg_size))val);	\
+}									\
+									\
+ATOMIC_INLINE type							\
+atomic_fetch_sub_##short_type(atomic_##short_type##_t *a,		\
+    type val, atomic_memory_order_t mo) {				\
+	/*								\
+	 * MSVC warns on negation of unsigned operands, but for us it	\
+	 * gives exactly the right semantics (MAX_TYPE + 1 - operand).	\
+	 */								\
+	__pragma(warning(push))						\
+	__pragma(warning(disable: 4146))				\
+	return atomic_fetch_add_##short_type(a, -val, mo);		\
+	__pragma(warning(pop))						\
+}									\
+ATOMIC_INLINE type							\
+atomic_fetch_and_##short_type(atomic_##short_type##_t *a,		\
+    type val, atomic_memory_order_t mo) {				\
+	return (type)ATOMIC_INTERLOCKED_NAME(_InterlockedAnd, lg_size)(	\
+	    &a->repr, (ATOMIC_INTERLOCKED_REPR(lg_size))val);		\
+}									\
+ATOMIC_INLINE type							\
+atomic_fetch_or_##short_type(atomic_##short_type##_t *a,		\
+    type val, atomic_memory_order_t mo) {				\
+	return (type)ATOMIC_INTERLOCKED_NAME(_InterlockedOr, lg_size)(	\
+	    &a->repr, (ATOMIC_INTERLOCKED_REPR(lg_size))val);		\
+}									\
+ATOMIC_INLINE type							\
+atomic_fetch_xor_##short_type(atomic_##short_type##_t *a,		\
+    type val, atomic_memory_order_t mo) {				\
+	return (type)ATOMIC_INTERLOCKED_NAME(_InterlockedXor, lg_size)(	\
+	    &a->repr, (ATOMIC_INTERLOCKED_REPR(lg_size))val);		\
+}
+
+#endif /* JEMALLOC_INTERNAL_ATOMIC_MSVC_H */

BeefRT/JEMalloc/include/jemalloc/internal/background_thread_externs.h

@@ -0,0 +1,33 @@
+#ifndef JEMALLOC_INTERNAL_BACKGROUND_THREAD_EXTERNS_H
+#define JEMALLOC_INTERNAL_BACKGROUND_THREAD_EXTERNS_H
+
+extern bool opt_background_thread;
+extern size_t opt_max_background_threads;
+extern malloc_mutex_t background_thread_lock;
+extern atomic_b_t background_thread_enabled_state;
+extern size_t n_background_threads;
+extern size_t max_background_threads;
+extern background_thread_info_t *background_thread_info;
+
+bool background_thread_create(tsd_t *tsd, unsigned arena_ind);
+bool background_threads_enable(tsd_t *tsd);
+bool background_threads_disable(tsd_t *tsd);
+bool background_thread_is_started(background_thread_info_t* info);
+void background_thread_wakeup_early(background_thread_info_t *info,
+    nstime_t *remaining_sleep);
+void background_thread_prefork0(tsdn_t *tsdn);
+void background_thread_prefork1(tsdn_t *tsdn);
+void background_thread_postfork_parent(tsdn_t *tsdn);
+void background_thread_postfork_child(tsdn_t *tsdn);
+bool background_thread_stats_read(tsdn_t *tsdn,
+    background_thread_stats_t *stats);
+void background_thread_ctl_init(tsdn_t *tsdn);
+
+#ifdef JEMALLOC_PTHREAD_CREATE_WRAPPER
+extern int pthread_create_wrapper(pthread_t *__restrict, const pthread_attr_t *,
+    void *(*)(void *), void *__restrict);
+#endif
+bool background_thread_boot0(void);
+bool background_thread_boot1(tsdn_t *tsdn, base_t *base);
+
+#endif /* JEMALLOC_INTERNAL_BACKGROUND_THREAD_EXTERNS_H */

BeefRT/JEMalloc/include/jemalloc/internal/background_thread_inlines.h

@@ -0,0 +1,48 @@
+#ifndef JEMALLOC_INTERNAL_BACKGROUND_THREAD_INLINES_H
+#define JEMALLOC_INTERNAL_BACKGROUND_THREAD_INLINES_H
+
+JEMALLOC_ALWAYS_INLINE bool
+background_thread_enabled(void) {
+	return atomic_load_b(&background_thread_enabled_state, ATOMIC_RELAXED);
+}
+
+JEMALLOC_ALWAYS_INLINE void
+background_thread_enabled_set(tsdn_t *tsdn, bool state) {
+	malloc_mutex_assert_owner(tsdn, &background_thread_lock);
+	atomic_store_b(&background_thread_enabled_state, state, ATOMIC_RELAXED);
+}
+
+JEMALLOC_ALWAYS_INLINE background_thread_info_t *
+arena_background_thread_info_get(arena_t *arena) {
+	unsigned arena_ind = arena_ind_get(arena);
+	return &background_thread_info[arena_ind % max_background_threads];
+}
+
+JEMALLOC_ALWAYS_INLINE background_thread_info_t *
+background_thread_info_get(size_t ind) {
+	return &background_thread_info[ind % max_background_threads];
+}
+
+JEMALLOC_ALWAYS_INLINE uint64_t
+background_thread_wakeup_time_get(background_thread_info_t *info) {
+	uint64_t next_wakeup = nstime_ns(&info->next_wakeup);
+	assert(atomic_load_b(&info->indefinite_sleep, ATOMIC_ACQUIRE) ==
+	    (next_wakeup == BACKGROUND_THREAD_INDEFINITE_SLEEP));
+	return next_wakeup;
+}
+
+JEMALLOC_ALWAYS_INLINE void
+background_thread_wakeup_time_set(tsdn_t *tsdn, background_thread_info_t *info,
+    uint64_t wakeup_time) {
+	malloc_mutex_assert_owner(tsdn, &info->mtx);
+	atomic_store_b(&info->indefinite_sleep,
+	    wakeup_time == BACKGROUND_THREAD_INDEFINITE_SLEEP, ATOMIC_RELEASE);
+	nstime_init(&info->next_wakeup, wakeup_time);
+}
+
+JEMALLOC_ALWAYS_INLINE bool
+background_thread_indefinite_sleep(background_thread_info_t *info) {
+	return atomic_load_b(&info->indefinite_sleep, ATOMIC_ACQUIRE);
+}
+
+#endif /* JEMALLOC_INTERNAL_BACKGROUND_THREAD_INLINES_H */

BeefRT/JEMalloc/include/jemalloc/internal/background_thread_structs.h

@@ -0,0 +1,66 @@
+#ifndef JEMALLOC_INTERNAL_BACKGROUND_THREAD_STRUCTS_H
+#define JEMALLOC_INTERNAL_BACKGROUND_THREAD_STRUCTS_H
+
+/* This file really combines "structs" and "types", but only transitionally. */
+
+#if defined(JEMALLOC_BACKGROUND_THREAD) || defined(JEMALLOC_LAZY_LOCK)
+#  define JEMALLOC_PTHREAD_CREATE_WRAPPER
+#endif
+
+#define BACKGROUND_THREAD_INDEFINITE_SLEEP UINT64_MAX
+#define MAX_BACKGROUND_THREAD_LIMIT MALLOCX_ARENA_LIMIT
+#define DEFAULT_NUM_BACKGROUND_THREAD 4
+
+/*
+ * These exist only as a transitional state.  Eventually, deferral should be
+ * part of the PAI, and each implementation can indicate wait times with more
+ * specificity.
+ */
+#define BACKGROUND_THREAD_HPA_INTERVAL_MAX_UNINITIALIZED (-2)
+#define BACKGROUND_THREAD_HPA_INTERVAL_MAX_DEFAULT_WHEN_ENABLED 5000
+
+#define BACKGROUND_THREAD_DEFERRED_MIN UINT64_C(0)
+#define BACKGROUND_THREAD_DEFERRED_MAX UINT64_MAX
+
+typedef enum {
+	background_thread_stopped,
+	background_thread_started,
+	/* Thread waits on the global lock when paused (for arena_reset). */
+	background_thread_paused,
+} background_thread_state_t;
+
+struct background_thread_info_s {
+#ifdef JEMALLOC_BACKGROUND_THREAD
+	/* Background thread is pthread specific. */
+	pthread_t		thread;
+	pthread_cond_t		cond;
+#endif
+	malloc_mutex_t		mtx;
+	background_thread_state_t	state;
+	/* When true, it means no wakeup scheduled. */
+	atomic_b_t		indefinite_sleep;
+	/* Next scheduled wakeup time (absolute time in ns). */
+	nstime_t		next_wakeup;
+	/*
+	 *  Since the last background thread run, newly added number of pages
+	 *  that need to be purged by the next wakeup.  This is adjusted on
+	 *  epoch advance, and is used to determine whether we should signal the
+	 *  background thread to wake up earlier.
+	 */
+	size_t			npages_to_purge_new;
+	/* Stats: total number of runs since started. */
+	uint64_t		tot_n_runs;
+	/* Stats: total sleep time since started. */
+	nstime_t		tot_sleep_time;
+};
+typedef struct background_thread_info_s background_thread_info_t;
+
+struct background_thread_stats_s {
+	size_t num_threads;
+	uint64_t num_runs;
+	nstime_t run_interval;
+	mutex_prof_data_t max_counter_per_bg_thd;
+};
+typedef struct background_thread_stats_s background_thread_stats_t;
+
+#endif /* JEMALLOC_INTERNAL_BACKGROUND_THREAD_STRUCTS_H */

BeefRT/JEMalloc/include/jemalloc/internal/base.h

@@ -0,0 +1,110 @@
+#ifndef JEMALLOC_INTERNAL_BASE_H
+#define JEMALLOC_INTERNAL_BASE_H
+
+#include "jemalloc/internal/edata.h"
+#include "jemalloc/internal/ehooks.h"
+#include "jemalloc/internal/mutex.h"
+
+enum metadata_thp_mode_e {
+	metadata_thp_disabled   = 0,
+	/*
+	 * Lazily enable hugepage for metadata. To avoid high RSS caused by THP
+	 * + low usage arena (i.e. THP becomes a significant percentage), the
+	 * "auto" option only starts using THP after a base allocator used up
+	 * the first THP region.  Starting from the second hugepage (in a single
+	 * arena), "auto" behaves the same as "always", i.e. madvise hugepage
+	 * right away.
+	 */
+	metadata_thp_auto       = 1,
+	metadata_thp_always     = 2,
+	metadata_thp_mode_limit = 3
+};
+typedef enum metadata_thp_mode_e metadata_thp_mode_t;
+
+#define METADATA_THP_DEFAULT metadata_thp_disabled
+extern metadata_thp_mode_t opt_metadata_thp;
+extern const char *metadata_thp_mode_names[];
+
+
+/* Embedded at the beginning of every block of base-managed virtual memory. */
+typedef struct base_block_s base_block_t;
+struct base_block_s {
+	/* Total size of block's virtual memory mapping. */
+	size_t size;
+
+	/* Next block in list of base's blocks. */
+	base_block_t *next;
+
+	/* Tracks unused trailing space. */
+	edata_t edata;
+};
+
+typedef struct base_s base_t;
+struct base_s {
+	/*
+	 * User-configurable extent hook functions.
+	 */
+	ehooks_t ehooks;
+
+	/*
+	 * User-configurable extent hook functions for metadata allocations.
+	 */
+	ehooks_t ehooks_base;
+
+	/* Protects base_alloc() and base_stats_get() operations. */
+	malloc_mutex_t mtx;
+
+	/* Using THP when true (metadata_thp auto mode). */
+	bool auto_thp_switched;
+	/*
+	 * Most recent size class in the series of increasingly large base
+	 * extents.  Logarithmic spacing between subsequent allocations ensures
+	 * that the total number of distinct mappings remains small.
+	 */
+	pszind_t pind_last;
+
+	/* Serial number generation state. */
+	size_t extent_sn_next;
+
+	/* Chain of all blocks associated with base. */
+	base_block_t *blocks;
+
+	/* Heap of extents that track unused trailing space within blocks. */
+	edata_heap_t avail[SC_NSIZES];
+
+	/* Stats, only maintained if config_stats. */
+	size_t allocated;
+	size_t resident;
+	size_t mapped;
+	/* Number of THP regions touched. */
+	size_t n_thp;
+};
+
+static inline unsigned
+base_ind_get(const base_t *base) {
+	return ehooks_ind_get(&base->ehooks);
+}
+
+static inline bool
+metadata_thp_enabled(void) {
+	return (opt_metadata_thp != metadata_thp_disabled);
+}
+
+base_t *b0get(void);
+base_t *base_new(tsdn_t *tsdn, unsigned ind,
+    const extent_hooks_t *extent_hooks, bool metadata_use_hooks);
+void base_delete(tsdn_t *tsdn, base_t *base);
+ehooks_t *base_ehooks_get(base_t *base);
+ehooks_t *base_ehooks_get_for_metadata(base_t *base);
+extent_hooks_t *base_extent_hooks_set(base_t *base,
+    extent_hooks_t *extent_hooks);
+void *base_alloc(tsdn_t *tsdn, base_t *base, size_t size, size_t alignment);
+edata_t *base_alloc_edata(tsdn_t *tsdn, base_t *base);
+void base_stats_get(tsdn_t *tsdn, base_t *base, size_t *allocated,
+    size_t *resident, size_t *mapped, size_t *n_thp);
+void base_prefork(tsdn_t *tsdn, base_t *base);
+void base_postfork_parent(tsdn_t *tsdn, base_t *base);
+void base_postfork_child(tsdn_t *tsdn, base_t *base);
+bool base_boot(tsdn_t *tsdn);
+
+#endif /* JEMALLOC_INTERNAL_BASE_H */

BeefRT/JEMalloc/include/jemalloc/internal/bin.h

@@ -0,0 +1,82 @@
+#ifndef JEMALLOC_INTERNAL_BIN_H
+#define JEMALLOC_INTERNAL_BIN_H
+
+#include "jemalloc/internal/bin_stats.h"
+#include "jemalloc/internal/bin_types.h"
+#include "jemalloc/internal/edata.h"
+#include "jemalloc/internal/mutex.h"
+#include "jemalloc/internal/sc.h"
+
+/*
+ * A bin contains a set of extents that are currently being used for slab
+ * allocations.
+ */
+typedef struct bin_s bin_t;
+struct bin_s {
+	/* All operations on bin_t fields require lock ownership. */
+	malloc_mutex_t		lock;
+
+	/*
+	 * Bin statistics.  These get touched every time the lock is acquired,
+	 * so put them close by in the hopes of getting some cache locality.
+	 */
+	bin_stats_t	stats;
+
+	/*
+	 * Current slab being used to service allocations of this bin's size
+	 * class.  slabcur is independent of slabs_{nonfull,full}; whenever
+	 * slabcur is reassigned, the previous slab must be deallocated or
+	 * inserted into slabs_{nonfull,full}.
+	 */
+	edata_t			*slabcur;
+
+	/*
+	 * Heap of non-full slabs.  This heap is used to assure that new
+	 * allocations come from the non-full slab that is oldest/lowest in
+	 * memory.
+	 */
+	edata_heap_t		slabs_nonfull;
+
+	/* List used to track full slabs. */
+	edata_list_active_t	slabs_full;
+};
+
+/* A set of sharded bins of the same size class. */
+typedef struct bins_s bins_t;
+struct bins_s {
+	/* Sharded bins.  Dynamically sized. */
+	bin_t *bin_shards;
+};
+
+void bin_shard_sizes_boot(unsigned bin_shards[SC_NBINS]);
+bool bin_update_shard_size(unsigned bin_shards[SC_NBINS], size_t start_size,
+    size_t end_size, size_t nshards);
+
+/* Initializes a bin to empty.  Returns true on error. */
+bool bin_init(bin_t *bin);
+
+/* Forking. */
+void bin_prefork(tsdn_t *tsdn, bin_t *bin);
+void bin_postfork_parent(tsdn_t *tsdn, bin_t *bin);
+void bin_postfork_child(tsdn_t *tsdn, bin_t *bin);
+
+/* Stats. */
+static inline void
+bin_stats_merge(tsdn_t *tsdn, bin_stats_data_t *dst_bin_stats, bin_t *bin) {
+	malloc_mutex_lock(tsdn, &bin->lock);
+	malloc_mutex_prof_accum(tsdn, &dst_bin_stats->mutex_data, &bin->lock);
+	bin_stats_t *stats = &dst_bin_stats->stats_data;
+	stats->nmalloc += bin->stats.nmalloc;
+	stats->ndalloc += bin->stats.ndalloc;
+	stats->nrequests += bin->stats.nrequests;
+	stats->curregs += bin->stats.curregs;
+	stats->nfills += bin->stats.nfills;
+	stats->nflushes += bin->stats.nflushes;
+	stats->nslabs += bin->stats.nslabs;
+	stats->reslabs += bin->stats.reslabs;
+	stats->curslabs += bin->stats.curslabs;
+	stats->nonfull_slabs += bin->stats.nonfull_slabs;
+	malloc_mutex_unlock(tsdn, &bin->lock);
+}
+
+#endif /* JEMALLOC_INTERNAL_BIN_H */

BeefRT/JEMalloc/include/jemalloc/internal/bin_info.h

@@ -0,0 +1,50 @@
+#ifndef JEMALLOC_INTERNAL_BIN_INFO_H
+#define JEMALLOC_INTERNAL_BIN_INFO_H
+
+#include "jemalloc/internal/bitmap.h"
+
+/*
+ * Read-only information associated with each element of arena_t's bins array
+ * is stored separately, partly to reduce memory usage (only one copy, rather
+ * than one per arena), but mainly to avoid false cacheline sharing.
+ *
+ * Each slab has the following layout:
+ *
+ *   /--------------------\
+ *   | region 0           |
+ *   |--------------------|
+ *   | region 1           |
+ *   |--------------------|
+ *   | ...                |
+ *   | ...                |
+ *   | ...                |
+ *   |--------------------|
+ *   | region nregs-1     |
+ *   \--------------------/
+ */
+typedef struct bin_info_s bin_info_t;
+struct bin_info_s {
+	/* Size of regions in a slab for this bin's size class. */
+	size_t			reg_size;
+
+	/* Total size of a slab for this bin's size class. */
+	size_t			slab_size;
+
+	/* Total number of regions in a slab for this bin's size class. */
+	uint32_t		nregs;
+
+	/* Number of sharded bins in each arena for this size class. */
+	uint32_t		n_shards;
+
+	/*
+	 * Metadata used to manipulate bitmaps for slabs associated with this
+	 * bin.
+	 */
+	bitmap_info_t		bitmap_info;
+};
+
+extern bin_info_t bin_infos[SC_NBINS];
+
+void bin_info_boot(sc_data_t *sc_data, unsigned bin_shard_sizes[SC_NBINS]);
+
+#endif /* JEMALLOC_INTERNAL_BIN_INFO_H */

BeefRT/JEMalloc/include/jemalloc/internal/bin_stats.h

@@ -0,0 +1,57 @@
+#ifndef JEMALLOC_INTERNAL_BIN_STATS_H
+#define JEMALLOC_INTERNAL_BIN_STATS_H
+
+#include "jemalloc/internal/mutex_prof.h"
+
+typedef struct bin_stats_s bin_stats_t;
+struct bin_stats_s {
+	/*
+	 * Total number of allocation/deallocation requests served directly by
+	 * the bin.  Note that tcache may allocate an object, then recycle it
+	 * many times, resulting many increments to nrequests, but only one
+	 * each to nmalloc and ndalloc.
+	 */
+	uint64_t	nmalloc;
+	uint64_t	ndalloc;
+
+	/*
+	 * Number of allocation requests that correspond to the size of this
+	 * bin.  This includes requests served by tcache, though tcache only
+	 * periodically merges into this counter.
+	 */
+	uint64_t	nrequests;
+
+	/*
+	 * Current number of regions of this size class, including regions
+	 * currently cached by tcache.
+	 */
+	size_t		curregs;
+
+	/* Number of tcache fills from this bin. */
+	uint64_t	nfills;
+
+	/* Number of tcache flushes to this bin. */
+	uint64_t	nflushes;
+
+	/* Total number of slabs created for this bin's size class. */
+	uint64_t	nslabs;
+
+	/*
+	 * Total number of slabs reused by extracting them from the slabs heap
+	 * for this bin's size class.
+	 */
+	uint64_t	reslabs;
+
+	/* Current number of slabs in this bin. */
+	size_t		curslabs;
+
+	/* Current size of nonfull slabs heap in this bin. */
+	size_t		nonfull_slabs;
+};
+
+typedef struct bin_stats_data_s bin_stats_data_t;
+struct bin_stats_data_s {
+	bin_stats_t stats_data;
+	mutex_prof_data_t mutex_data;
+};
+#endif /* JEMALLOC_INTERNAL_BIN_STATS_H */

BeefRT/JEMalloc/include/jemalloc/internal/bin_types.h

@@ -0,0 +1,17 @@
+#ifndef JEMALLOC_INTERNAL_BIN_TYPES_H
+#define JEMALLOC_INTERNAL_BIN_TYPES_H
+
+#include "jemalloc/internal/sc.h"
+
+#define BIN_SHARDS_MAX (1 << EDATA_BITS_BINSHARD_WIDTH)
+#define N_BIN_SHARDS_DEFAULT 1
+
+/* Used in TSD static initializer only. Real init in arena_bind(). */
+#define TSD_BINSHARDS_ZERO_INITIALIZER {{UINT8_MAX}}
+
+typedef struct tsd_binshards_s tsd_binshards_t;
+struct tsd_binshards_s {
+	uint8_t binshard[SC_NBINS];
+};
+
+#endif /* JEMALLOC_INTERNAL_BIN_TYPES_H */

BeefRT/JEMalloc/include/jemalloc/internal/bit_util.h

@@ -0,0 +1,422 @@
+#ifndef JEMALLOC_INTERNAL_BIT_UTIL_H
+#define JEMALLOC_INTERNAL_BIT_UTIL_H
+
+#include "jemalloc/internal/assert.h"
+
+/* Sanity check. */
+#if !defined(JEMALLOC_INTERNAL_FFSLL) || !defined(JEMALLOC_INTERNAL_FFSL) \
+    || !defined(JEMALLOC_INTERNAL_FFS)
+#  error JEMALLOC_INTERNAL_FFS{,L,LL} should have been defined by configure
+#endif
+
+/*
+ * Unlike the builtins and posix ffs functions, our ffs requires a non-zero
+ * input, and returns the position of the lowest bit set (as opposed to the
+ * posix versions, which return 1 larger than that position and use a return
+ * value of zero as a sentinel.  This tends to simplify logic in callers, and
+ * allows for consistency with the builtins we build fls on top of.
+ */
+static inline unsigned
+ffs_llu(unsigned long long x) {
+	util_assume(x != 0);
+	return JEMALLOC_INTERNAL_FFSLL(x) - 1;
+}
+
+static inline unsigned
+ffs_lu(unsigned long x) {
+	util_assume(x != 0);
+	return JEMALLOC_INTERNAL_FFSL(x) - 1;
+}
+
+static inline unsigned
+ffs_u(unsigned x) {
+	util_assume(x != 0);
+	return JEMALLOC_INTERNAL_FFS(x) - 1;
+}
+
+#define DO_FLS_SLOW(x, suffix) do {					\
+	util_assume(x != 0);						\
+	x |= (x >> 1);							\
+	x |= (x >> 2);							\
+	x |= (x >> 4);							\
+	x |= (x >> 8);							\
+	x |= (x >> 16);							\
+	if (sizeof(x) > 4) {						\
+		/*							\
+		 * If sizeof(x) is 4, then the expression "x >> 32"	\
+		 * will generate compiler warnings even if the code	\
+		 * never executes.  This circumvents the warning, and	\
+		 * gets compiled out in optimized builds.		\
+		 */							\
+		int constant_32 = sizeof(x) * 4;			\
+		x |= (x >> constant_32);				\
+	}								\
+	x++;								\
+	if (x == 0) {							\
+		return 8 * sizeof(x) - 1;				\
+	}								\
+	return ffs_##suffix(x) - 1;					\
+} while(0)
+
+static inline unsigned
+fls_llu_slow(unsigned long long x) {
+	DO_FLS_SLOW(x, llu);
+}
+
+static inline unsigned
+fls_lu_slow(unsigned long x) {
+	DO_FLS_SLOW(x, lu);
+}
+
+static inline unsigned
+fls_u_slow(unsigned x) {
+	DO_FLS_SLOW(x, u);
+}
+
+#undef DO_FLS_SLOW
+
+#ifdef JEMALLOC_HAVE_BUILTIN_CLZ
+static inline unsigned
+fls_llu(unsigned long long x) {
+	util_assume(x != 0);
+	/*
+	 * Note that the xor here is more naturally written as subtraction; the
+	 * last bit set is the number of bits in the type minus the number of
+	 * leading zero bits.  But GCC implements that as:
+	 *    bsr     edi, edi
+	 *    mov     eax, 31
+	 *    xor     edi, 31
+	 *    sub     eax, edi
+	 * If we write it as xor instead, then we get
+	 *    bsr     eax, edi
+	 * as desired.
+	 */
+	return (8 * sizeof(x) - 1) ^ __builtin_clzll(x);
+}
+
+static inline unsigned
+fls_lu(unsigned long x) {
+	util_assume(x != 0);
+	return (8 * sizeof(x) - 1) ^ __builtin_clzl(x);
+}
+
+static inline unsigned
+fls_u(unsigned x) {
+	util_assume(x != 0);
+	return (8 * sizeof(x) - 1) ^ __builtin_clz(x);
+}
+#elif defined(_MSC_VER)
+
+#if LG_SIZEOF_PTR == 3
+#define DO_BSR64(bit, x) _BitScanReverse64(&bit, x)
+#else
+/*
+ * This never actually runs; we're just dodging a compiler error for the
+ * never-taken branch where sizeof(void *) == 8.
+ */
+#define DO_BSR64(bit, x) bit = 0; unreachable()
+#endif
+
+#define DO_FLS(x) do {							\
+	if (x == 0) {							\
+		return 8 * sizeof(x);					\
+	}								\
+	unsigned long bit;						\
+	if (sizeof(x) == 4) {						\
+		_BitScanReverse(&bit, (unsigned)x);			\
+		return (unsigned)bit;					\
+	}								\
+	if (sizeof(x) == 8 && sizeof(void *) == 8) {			\
+		DO_BSR64(bit, x);					\
+		return (unsigned)bit;					\
+	}								\
+	if (sizeof(x) == 8 && sizeof(void *) == 4) {			\
+		/* Dodge a compiler warning, as above. */		\
+		int constant_32 = sizeof(x) * 4;			\
+		if (_BitScanReverse(&bit,				\
+		    (unsigned)(x >> constant_32))) {			\
+			return 32 + (unsigned)bit;			\
+		} else {						\
+			_BitScanReverse(&bit, (unsigned)x);		\
+			return (unsigned)bit;				\
+		}							\
+	}								\
+	unreachable();							\
+} while (0)
+
+static inline unsigned
+fls_llu(unsigned long long x) {
+	DO_FLS(x);
+}
+
+static inline unsigned
+fls_lu(unsigned long x) {
+	DO_FLS(x);
+}
+
+static inline unsigned
+fls_u(unsigned x) {
+	DO_FLS(x);
+}
+
+#undef DO_FLS
+#undef DO_BSR64
+#else
+
+static inline unsigned
+fls_llu(unsigned long long x) {
+	return fls_llu_slow(x);
+}
+
+static inline unsigned
+fls_lu(unsigned long x) {
+	return fls_lu_slow(x);
+}
+
+static inline unsigned
+fls_u(unsigned x) {
+	return fls_u_slow(x);
+}
+#endif
+
+#if LG_SIZEOF_LONG_LONG > 3
+#  error "Haven't implemented popcount for 16-byte ints."
+#endif
+
+#define DO_POPCOUNT(x, type) do {					\
+	/*								\
+	 * Algorithm from an old AMD optimization reference manual.	\
+	 * We're putting a little bit more work than you might expect	\
+	 * into the no-instrinsic case, since we only support the	\
+	 * GCC intrinsics spelling of popcount (for now).  Detecting	\
+	 * whether or not the popcount builtin is actually useable in	\
+	 * MSVC is nontrivial.						\
+	 */								\
+									\
+	type bmul = (type)0x0101010101010101ULL;			\
+									\
+	/*								\
+	 * Replace each 2 bits with the sideways sum of the original	\
+	 * values.  0x5 = 0b0101.					\
+	 *								\
+	 * You might expect this to be:					\
+	 *   x = (x & 0x55...) + ((x >> 1) & 0x55...).			\
+	 * That costs an extra mask relative to this, though.		\
+	 */								\
+	x = x - ((x >> 1) & (0x55U * bmul));				\
+	/* Replace each 4 bits with their sideays sum.  0x3 = 0b0011. */\
+	x = (x & (bmul * 0x33U)) + ((x >> 2) & (bmul * 0x33U));		\
+	/*								\
+	 * Replace each 8 bits with their sideways sum.  Note that we	\
+	 * can't overflow within each 4-bit sum here, so we can skip	\
+	 * the initial mask.						\
+	 */								\
+	x = (x + (x >> 4)) & (bmul * 0x0FU);				\
+	/*								\
+	 * None of the partial sums in this multiplication (viewed in	\
+	 * base-256) can overflow into the next digit.  So the least	\
+	 * significant byte of the product will be the least		\
+	 * significant byte of the original value, the second least	\
+	 * significant byte will be the sum of the two least		\
+	 * significant bytes of the original value, and so on.		\
+	 * Importantly, the high byte will be the byte-wise sum of all	\
+	 * the bytes of the original value.				\
+	 */								\
+	x = x * bmul;							\
+	x >>= ((sizeof(x) - 1) * 8);					\
+	return (unsigned)x;						\
+} while(0)
+
+static inline unsigned
+popcount_u_slow(unsigned bitmap) {
+	DO_POPCOUNT(bitmap, unsigned);
+}
+
+static inline unsigned
+popcount_lu_slow(unsigned long bitmap) {
+	DO_POPCOUNT(bitmap, unsigned long);
+}
+
+static inline unsigned
+popcount_llu_slow(unsigned long long bitmap) {
+	DO_POPCOUNT(bitmap, unsigned long long);
+}
+
+#undef DO_POPCOUNT
+
+static inline unsigned
+popcount_u(unsigned bitmap) {
+#ifdef JEMALLOC_INTERNAL_POPCOUNT
+	return JEMALLOC_INTERNAL_POPCOUNT(bitmap);
+#else
+	return popcount_u_slow(bitmap);
+#endif
+}
+
+static inline unsigned
+popcount_lu(unsigned long bitmap) {
+#ifdef JEMALLOC_INTERNAL_POPCOUNTL
+	return JEMALLOC_INTERNAL_POPCOUNTL(bitmap);
+#else
+	return popcount_lu_slow(bitmap);
+#endif
+}
+
+static inline unsigned
+popcount_llu(unsigned long long bitmap) {
+#ifdef JEMALLOC_INTERNAL_POPCOUNTLL
+	return JEMALLOC_INTERNAL_POPCOUNTLL(bitmap);
+#else
+	return popcount_llu_slow(bitmap);
+#endif
+}
+
+/*
+ * Clears first unset bit in bitmap, and returns
+ * place of bit.  bitmap *must not* be 0.
+ */
+
+static inline size_t
+cfs_lu(unsigned long* bitmap) {
+	util_assume(*bitmap != 0);
+	size_t bit = ffs_lu(*bitmap);
+	*bitmap ^= ZU(1) << bit;
+	return bit;
+}
+
+static inline unsigned
+ffs_zu(size_t x) {
+#if LG_SIZEOF_PTR == LG_SIZEOF_INT
+	return ffs_u(x);
+#elif LG_SIZEOF_PTR == LG_SIZEOF_LONG
+	return ffs_lu(x);
+#elif LG_SIZEOF_PTR == LG_SIZEOF_LONG_LONG
+	return ffs_llu(x);
+#else
+#error No implementation for size_t ffs()
+#endif
+}
+
+static inline unsigned
+fls_zu(size_t x) {
+#if LG_SIZEOF_PTR == LG_SIZEOF_INT
+	return fls_u(x);
+#elif LG_SIZEOF_PTR == LG_SIZEOF_LONG
+	return fls_lu(x);
+#elif LG_SIZEOF_PTR == LG_SIZEOF_LONG_LONG
+	return fls_llu(x);
+#else
+#error No implementation for size_t fls()
+#endif
+}
+
+
+static inline unsigned
+ffs_u64(uint64_t x) {
+#if LG_SIZEOF_LONG == 3
+	return ffs_lu(x);
+#elif LG_SIZEOF_LONG_LONG == 3
+	return ffs_llu(x);
+#else
+#error No implementation for 64-bit ffs()
+#endif
+}
+
+static inline unsigned
+fls_u64(uint64_t x) {
+#if LG_SIZEOF_LONG == 3
+	return fls_lu(x);
+#elif LG_SIZEOF_LONG_LONG == 3
+	return fls_llu(x);
+#else
+#error No implementation for 64-bit fls()
+#endif
+}
+
+static inline unsigned
+ffs_u32(uint32_t x) {
+#if LG_SIZEOF_INT == 2
+	return ffs_u(x);
+#else
+#error No implementation for 32-bit ffs()
+#endif
+	return ffs_u(x);
+}
+
+static inline unsigned
+fls_u32(uint32_t x) {
+#if LG_SIZEOF_INT == 2
+	return fls_u(x);
+#else
+#error No implementation for 32-bit fls()
+#endif
+	return fls_u(x);
+}
+
+static inline uint64_t
+pow2_ceil_u64(uint64_t x) {
+	if (unlikely(x <= 1)) {
+		return x;
+	}
+	size_t msb_on_index = fls_u64(x - 1);
+	/*
+	 * Range-check; it's on the callers to ensure that the result of this
+	 * call won't overflow.
+	 */
+	assert(msb_on_index < 63);
+	return 1ULL << (msb_on_index + 1);
+}
+
+static inline uint32_t
+pow2_ceil_u32(uint32_t x) {
+	if (unlikely(x <= 1)) {
+	    return x;
+	}
+	size_t msb_on_index = fls_u32(x - 1);
+	/* As above. */
+	assert(msb_on_index < 31);
+	return 1U << (msb_on_index + 1);
+}
+
+/* Compute the smallest power of 2 that is >= x. */
+static inline size_t
+pow2_ceil_zu(size_t x) {
+#if (LG_SIZEOF_PTR == 3)
+	return pow2_ceil_u64(x);
+#else
+	return pow2_ceil_u32(x);
+#endif
+}
+
+static inline unsigned
+lg_floor(size_t x) {
+	util_assume(x != 0);
+#if (LG_SIZEOF_PTR == 3)
+	return fls_u64(x);
+#else
+	return fls_u32(x);
+#endif
+}
+
+static inline unsigned
+lg_ceil(size_t x) {
+	return lg_floor(x) + ((x & (x - 1)) == 0 ? 0 : 1);
+}
+
+/* A compile-time version of lg_floor and lg_ceil. */
+#define LG_FLOOR_1(x) 0
+#define LG_FLOOR_2(x) (x < (1ULL << 1) ? LG_FLOOR_1(x) : 1 + LG_FLOOR_1(x >> 1))
+#define LG_FLOOR_4(x) (x < (1ULL << 2) ? LG_FLOOR_2(x) : 2 + LG_FLOOR_2(x >> 2))
+#define LG_FLOOR_8(x) (x < (1ULL << 4) ? LG_FLOOR_4(x) : 4 + LG_FLOOR_4(x >> 4))
+#define LG_FLOOR_16(x) (x < (1ULL << 8) ? LG_FLOOR_8(x) : 8 + LG_FLOOR_8(x >> 8))
+#define LG_FLOOR_32(x) (x < (1ULL << 16) ? LG_FLOOR_16(x) : 16 + LG_FLOOR_16(x >> 16))
+#define LG_FLOOR_64(x) (x < (1ULL << 32) ? LG_FLOOR_32(x) : 32 + LG_FLOOR_32(x >> 32))
+#if LG_SIZEOF_PTR == 2
+#  define LG_FLOOR(x) LG_FLOOR_32((x))
+#else
+#  define LG_FLOOR(x) LG_FLOOR_64((x))
+#endif
+
+#define LG_CEIL(x) (LG_FLOOR(x) + (((x) & ((x) - 1)) == 0 ? 0 : 1))
+
+#endif /* JEMALLOC_INTERNAL_BIT_UTIL_H */

BeefRT/JEMalloc/include/jemalloc/internal/bitmap.h

@@ -0,0 +1,368 @@
+#ifndef JEMALLOC_INTERNAL_BITMAP_H
+#define JEMALLOC_INTERNAL_BITMAP_H
+
+#include "jemalloc/internal/bit_util.h"
+#include "jemalloc/internal/sc.h"
+
+typedef unsigned long bitmap_t;
+#define LG_SIZEOF_BITMAP	LG_SIZEOF_LONG
+
+/* Maximum bitmap bit count is 2^LG_BITMAP_MAXBITS. */
+#if SC_LG_SLAB_MAXREGS > LG_CEIL(SC_NSIZES)
+/* Maximum bitmap bit count is determined by maximum regions per slab. */
+#  define LG_BITMAP_MAXBITS	SC_LG_SLAB_MAXREGS
+#else
+/* Maximum bitmap bit count is determined by number of extent size classes. */
+#  define LG_BITMAP_MAXBITS	LG_CEIL(SC_NSIZES)
+#endif
+#define BITMAP_MAXBITS		(ZU(1) << LG_BITMAP_MAXBITS)
+
+/* Number of bits per group. */
+#define LG_BITMAP_GROUP_NBITS		(LG_SIZEOF_BITMAP + 3)
+#define BITMAP_GROUP_NBITS		(1U << LG_BITMAP_GROUP_NBITS)
+#define BITMAP_GROUP_NBITS_MASK		(BITMAP_GROUP_NBITS-1)
+
+/*
+ * Do some analysis on how big the bitmap is before we use a tree.  For a brute
+ * force linear search, if we would have to call ffs_lu() more than 2^3 times,
+ * use a tree instead.
+ */
+#if LG_BITMAP_MAXBITS - LG_BITMAP_GROUP_NBITS > 3
+#  define BITMAP_USE_TREE
+#endif
+
+/* Number of groups required to store a given number of bits. */
+#define BITMAP_BITS2GROUPS(nbits)					\
+    (((nbits) + BITMAP_GROUP_NBITS_MASK) >> LG_BITMAP_GROUP_NBITS)
+
+/*
+ * Number of groups required at a particular level for a given number of bits.
+ */
+#define BITMAP_GROUPS_L0(nbits)						\
+    BITMAP_BITS2GROUPS(nbits)
+#define BITMAP_GROUPS_L1(nbits)						\
+    BITMAP_BITS2GROUPS(BITMAP_BITS2GROUPS(nbits))
+#define BITMAP_GROUPS_L2(nbits)						\
+    BITMAP_BITS2GROUPS(BITMAP_BITS2GROUPS(BITMAP_BITS2GROUPS((nbits))))
+#define BITMAP_GROUPS_L3(nbits)						\
+    BITMAP_BITS2GROUPS(BITMAP_BITS2GROUPS(BITMAP_BITS2GROUPS(		\
+	BITMAP_BITS2GROUPS((nbits)))))
+#define BITMAP_GROUPS_L4(nbits)						\
+    BITMAP_BITS2GROUPS(BITMAP_BITS2GROUPS(BITMAP_BITS2GROUPS(		\
+	BITMAP_BITS2GROUPS(BITMAP_BITS2GROUPS((nbits))))))
+
+/*
+ * Assuming the number of levels, number of groups required for a given number
+ * of bits.
+ */
+#define BITMAP_GROUPS_1_LEVEL(nbits)					\
+    BITMAP_GROUPS_L0(nbits)
+#define BITMAP_GROUPS_2_LEVEL(nbits)					\
+    (BITMAP_GROUPS_1_LEVEL(nbits) + BITMAP_GROUPS_L1(nbits))
+#define BITMAP_GROUPS_3_LEVEL(nbits)					\
+    (BITMAP_GROUPS_2_LEVEL(nbits) + BITMAP_GROUPS_L2(nbits))
+#define BITMAP_GROUPS_4_LEVEL(nbits)					\
+    (BITMAP_GROUPS_3_LEVEL(nbits) + BITMAP_GROUPS_L3(nbits))
+#define BITMAP_GROUPS_5_LEVEL(nbits)					\
+    (BITMAP_GROUPS_4_LEVEL(nbits) + BITMAP_GROUPS_L4(nbits))
+
+/*
+ * Maximum number of groups required to support LG_BITMAP_MAXBITS.
+ */
+#ifdef BITMAP_USE_TREE
+
+#if LG_BITMAP_MAXBITS <= LG_BITMAP_GROUP_NBITS
+#  define BITMAP_GROUPS(nbits)	BITMAP_GROUPS_1_LEVEL(nbits)
+#  define BITMAP_GROUPS_MAX	BITMAP_GROUPS_1_LEVEL(BITMAP_MAXBITS)
+#elif LG_BITMAP_MAXBITS <= LG_BITMAP_GROUP_NBITS * 2
+#  define BITMAP_GROUPS(nbits)	BITMAP_GROUPS_2_LEVEL(nbits)
+#  define BITMAP_GROUPS_MAX	BITMAP_GROUPS_2_LEVEL(BITMAP_MAXBITS)
+#elif LG_BITMAP_MAXBITS <= LG_BITMAP_GROUP_NBITS * 3
+#  define BITMAP_GROUPS(nbits)	BITMAP_GROUPS_3_LEVEL(nbits)
+#  define BITMAP_GROUPS_MAX	BITMAP_GROUPS_3_LEVEL(BITMAP_MAXBITS)
+#elif LG_BITMAP_MAXBITS <= LG_BITMAP_GROUP_NBITS * 4
+#  define BITMAP_GROUPS(nbits)	BITMAP_GROUPS_4_LEVEL(nbits)
+#  define BITMAP_GROUPS_MAX	BITMAP_GROUPS_4_LEVEL(BITMAP_MAXBITS)
+#elif LG_BITMAP_MAXBITS <= LG_BITMAP_GROUP_NBITS * 5
+#  define BITMAP_GROUPS(nbits)	BITMAP_GROUPS_5_LEVEL(nbits)
+#  define BITMAP_GROUPS_MAX	BITMAP_GROUPS_5_LEVEL(BITMAP_MAXBITS)
+#else
+#  error "Unsupported bitmap size"
+#endif
+
+/*
+ * Maximum number of levels possible.  This could be statically computed based
+ * on LG_BITMAP_MAXBITS:
+ *
+ * #define BITMAP_MAX_LEVELS \
+ *     (LG_BITMAP_MAXBITS / LG_SIZEOF_BITMAP) \
+ *     + !!(LG_BITMAP_MAXBITS % LG_SIZEOF_BITMAP)
+ *
+ * However, that would not allow the generic BITMAP_INFO_INITIALIZER() macro, so
+ * instead hardcode BITMAP_MAX_LEVELS to the largest number supported by the
+ * various cascading macros.  The only additional cost this incurs is some
+ * unused trailing entries in bitmap_info_t structures; the bitmaps themselves
+ * are not impacted.
+ */
+#define BITMAP_MAX_LEVELS	5
+
+#define BITMAP_INFO_INITIALIZER(nbits) {				\
+	/* nbits. */							\
+	nbits,								\
+	/* nlevels. */							\
+	(BITMAP_GROUPS_L0(nbits) > BITMAP_GROUPS_L1(nbits)) +		\
+	    (BITMAP_GROUPS_L1(nbits) > BITMAP_GROUPS_L2(nbits)) +	\
+	    (BITMAP_GROUPS_L2(nbits) > BITMAP_GROUPS_L3(nbits)) +	\
+	    (BITMAP_GROUPS_L3(nbits) > BITMAP_GROUPS_L4(nbits)) + 1,	\
+	/* levels. */							\
+	{								\
+		{0},							\
+		{BITMAP_GROUPS_L0(nbits)},				\
+		{BITMAP_GROUPS_L1(nbits) + BITMAP_GROUPS_L0(nbits)},	\
+		{BITMAP_GROUPS_L2(nbits) + BITMAP_GROUPS_L1(nbits) +	\
+		    BITMAP_GROUPS_L0(nbits)},				\
+		{BITMAP_GROUPS_L3(nbits) + BITMAP_GROUPS_L2(nbits) +	\
+		    BITMAP_GROUPS_L1(nbits) + BITMAP_GROUPS_L0(nbits)},	\
+		{BITMAP_GROUPS_L4(nbits) + BITMAP_GROUPS_L3(nbits) +	\
+		     BITMAP_GROUPS_L2(nbits) + BITMAP_GROUPS_L1(nbits)	\
+		     + BITMAP_GROUPS_L0(nbits)}				\
+	}								\
+}
+
+#else /* BITMAP_USE_TREE */
+
+#define BITMAP_GROUPS(nbits)	BITMAP_BITS2GROUPS(nbits)
+#define BITMAP_GROUPS_MAX	BITMAP_BITS2GROUPS(BITMAP_MAXBITS)
+
+#define BITMAP_INFO_INITIALIZER(nbits) {				\
+	/* nbits. */							\
+	nbits,								\
+	/* ngroups. */							\
+	BITMAP_BITS2GROUPS(nbits)					\
+}
+
+#endif /* BITMAP_USE_TREE */
+
+typedef struct bitmap_level_s {
+	/* Offset of this level's groups within the array of groups. */
+	size_t group_offset;
+} bitmap_level_t;
+
+typedef struct bitmap_info_s {
+	/* Logical number of bits in bitmap (stored at bottom level). */
+	size_t nbits;
+
+#ifdef BITMAP_USE_TREE
+	/* Number of levels necessary for nbits. */
+	unsigned nlevels;
+
+	/*
+	 * Only the first (nlevels+1) elements are used, and levels are ordered
+	 * bottom to top (e.g. the bottom level is stored in levels[0]).
+	 */
+	bitmap_level_t levels[BITMAP_MAX_LEVELS+1];
+#else /* BITMAP_USE_TREE */
+	/* Number of groups necessary for nbits. */
+	size_t ngroups;
+#endif /* BITMAP_USE_TREE */
+} bitmap_info_t;
+
+void bitmap_info_init(bitmap_info_t *binfo, size_t nbits);
+void bitmap_init(bitmap_t *bitmap, const bitmap_info_t *binfo, bool fill);
+size_t bitmap_size(const bitmap_info_t *binfo);
+
+static inline bool
+bitmap_full(bitmap_t *bitmap, const bitmap_info_t *binfo) {
+#ifdef BITMAP_USE_TREE
+	size_t rgoff = binfo->levels[binfo->nlevels].group_offset - 1;
+	bitmap_t rg = bitmap[rgoff];
+	/* The bitmap is full iff the root group is 0. */
+	return (rg == 0);
+#else
+	size_t i;
+
+	for (i = 0; i < binfo->ngroups; i++) {
+		if (bitmap[i] != 0) {
+			return false;
+		}
+	}
+	return true;
+#endif
+}
+
+static inline bool
+bitmap_get(bitmap_t *bitmap, const bitmap_info_t *binfo, size_t bit) {
+	size_t goff;
+	bitmap_t g;
+
+	assert(bit < binfo->nbits);
+	goff = bit >> LG_BITMAP_GROUP_NBITS;
+	g = bitmap[goff];
+	return !(g & (ZU(1) << (bit & BITMAP_GROUP_NBITS_MASK)));
+}
+
+static inline void
+bitmap_set(bitmap_t *bitmap, const bitmap_info_t *binfo, size_t bit) {
+	size_t goff;
+	bitmap_t *gp;
+	bitmap_t g;
+
+	assert(bit < binfo->nbits);
+	assert(!bitmap_get(bitmap, binfo, bit));
+	goff = bit >> LG_BITMAP_GROUP_NBITS;
+	gp = &bitmap[goff];
+	g = *gp;
+	assert(g & (ZU(1) << (bit & BITMAP_GROUP_NBITS_MASK)));
+	g ^= ZU(1) << (bit & BITMAP_GROUP_NBITS_MASK);
+	*gp = g;
+	assert(bitmap_get(bitmap, binfo, bit));
+#ifdef BITMAP_USE_TREE
+	/* Propagate group state transitions up the tree. */
+	if (g == 0) {
+		unsigned i;
+		for (i = 1; i < binfo->nlevels; i++) {
+			bit = goff;
+			goff = bit >> LG_BITMAP_GROUP_NBITS;
+			gp = &bitmap[binfo->levels[i].group_offset + goff];
+			g = *gp;
+			assert(g & (ZU(1) << (bit & BITMAP_GROUP_NBITS_MASK)));
+			g ^= ZU(1) << (bit & BITMAP_GROUP_NBITS_MASK);
+			*gp = g;
+			if (g != 0) {
+				break;
+			}
+		}
+	}
+#endif
+}
+
+/* ffu: find first unset >= bit. */
+static inline size_t
+bitmap_ffu(const bitmap_t *bitmap, const bitmap_info_t *binfo, size_t min_bit) {
+	assert(min_bit < binfo->nbits);
+
+#ifdef BITMAP_USE_TREE
+	size_t bit = 0;
+	for (unsigned level = binfo->nlevels; level--;) {
+		size_t lg_bits_per_group = (LG_BITMAP_GROUP_NBITS * (level +
+		    1));
+		bitmap_t group = bitmap[binfo->levels[level].group_offset + (bit
+		    >> lg_bits_per_group)];
+		unsigned group_nmask = (unsigned)(((min_bit > bit) ? (min_bit -
+		    bit) : 0) >> (lg_bits_per_group - LG_BITMAP_GROUP_NBITS));
+		assert(group_nmask <= BITMAP_GROUP_NBITS);
+		bitmap_t group_mask = ~((1LU << group_nmask) - 1);
+		bitmap_t group_masked = group & group_mask;
+		if (group_masked == 0LU) {
+			if (group == 0LU) {
+				return binfo->nbits;
+			}
+			/*
+			 * min_bit was preceded by one or more unset bits in
+			 * this group, but there are no other unset bits in this
+			 * group.  Try again starting at the first bit of the
+			 * next sibling.  This will recurse at most once per
+			 * non-root level.
+			 */
+			size_t sib_base = bit + (ZU(1) << lg_bits_per_group);
+			assert(sib_base > min_bit);
+			assert(sib_base > bit);
+			if (sib_base >= binfo->nbits) {
+				return binfo->nbits;
+			}
+			return bitmap_ffu(bitmap, binfo, sib_base);
+		}
+		bit += ((size_t)ffs_lu(group_masked)) <<
+		    (lg_bits_per_group - LG_BITMAP_GROUP_NBITS);
+	}
+	assert(bit >= min_bit);
+	assert(bit < binfo->nbits);
+	return bit;
+#else
+	size_t i = min_bit >> LG_BITMAP_GROUP_NBITS;
+	bitmap_t g = bitmap[i] & ~((1LU << (min_bit & BITMAP_GROUP_NBITS_MASK))
+	    - 1);
+	size_t bit;
+	do {
+		if (g != 0) {
+			bit = ffs_lu(g);
+			return (i << LG_BITMAP_GROUP_NBITS) + bit;
+		}
+		i++;
+		g = bitmap[i];
+	} while (i < binfo->ngroups);
+	return binfo->nbits;
+#endif
+}
+
+/* sfu: set first unset. */
+static inline size_t
+bitmap_sfu(bitmap_t *bitmap, const bitmap_info_t *binfo) {
+	size_t bit;
+	bitmap_t g;
+	unsigned i;
+
+	assert(!bitmap_full(bitmap, binfo));
+
+#ifdef BITMAP_USE_TREE
+	i = binfo->nlevels - 1;
+	g = bitmap[binfo->levels[i].group_offset];
+	bit = ffs_lu(g);
+	while (i > 0) {
+		i--;
+		g = bitmap[binfo->levels[i].group_offset + bit];
+		bit = (bit << LG_BITMAP_GROUP_NBITS) + ffs_lu(g);
+	}
+#else
+	i = 0;
+	g = bitmap[0];
+	while (g == 0) {
+		i++;
+		g = bitmap[i];
+	}
+	bit = (i << LG_BITMAP_GROUP_NBITS) + ffs_lu(g);
+#endif
+	bitmap_set(bitmap, binfo, bit);
+	return bit;
+}
+
+static inline void
+bitmap_unset(bitmap_t *bitmap, const bitmap_info_t *binfo, size_t bit) {
+	size_t goff;
+	bitmap_t *gp;
+	bitmap_t g;
+	UNUSED bool propagate;
+
+	assert(bit < binfo->nbits);
+	assert(bitmap_get(bitmap, binfo, bit));
+	goff = bit >> LG_BITMAP_GROUP_NBITS;
+	gp = &bitmap[goff];
+	g = *gp;
+	propagate = (g == 0);
+	assert((g & (ZU(1) << (bit & BITMAP_GROUP_NBITS_MASK))) == 0);
+	g ^= ZU(1) << (bit & BITMAP_GROUP_NBITS_MASK);
+	*gp = g;
+	assert(!bitmap_get(bitmap, binfo, bit));
+#ifdef BITMAP_USE_TREE
+	/* Propagate group state transitions up the tree. */
+	if (propagate) {
+		unsigned i;
+		for (i = 1; i < binfo->nlevels; i++) {
+			bit = goff;
+			goff = bit >> LG_BITMAP_GROUP_NBITS;
+			gp = &bitmap[binfo->levels[i].group_offset + goff];
+			g = *gp;
+			propagate = (g == 0);
+			assert((g & (ZU(1) << (bit & BITMAP_GROUP_NBITS_MASK)))
+			    == 0);
+			g ^= ZU(1) << (bit & BITMAP_GROUP_NBITS_MASK);
+			*gp = g;
+			if (!propagate) {
+				break;
+			}
+		}
+	}
+#endif /* BITMAP_USE_TREE */
+}
+
+#endif /* JEMALLOC_INTERNAL_BITMAP_H */

BeefRT/JEMalloc/include/jemalloc/internal/buf_writer.h

@@ -0,0 +1,32 @@
+#ifndef JEMALLOC_INTERNAL_BUF_WRITER_H
+#define JEMALLOC_INTERNAL_BUF_WRITER_H
+
+/*
+ * Note: when using the buffered writer, cbopaque is passed to write_cb only
+ * when the buffer is flushed.  It would make a difference if cbopaque points
+ * to something that's changing for each write_cb call, or something that
+ * affects write_cb in a way dependent on the content of the output string.
+ * However, the most typical usage case in practice is that cbopaque points to
+ * some "option like" content for the write_cb, so it doesn't matter.
+ */
+
+typedef struct {
+	write_cb_t *write_cb;
+	void *cbopaque;
+	char *buf;
+	size_t buf_size;
+	size_t buf_end;
+	bool internal_buf;
+} buf_writer_t;
+
+bool buf_writer_init(tsdn_t *tsdn, buf_writer_t *buf_writer,
+    write_cb_t *write_cb, void *cbopaque, char *buf, size_t buf_len);
+void buf_writer_flush(buf_writer_t *buf_writer);
+write_cb_t buf_writer_cb;
+void buf_writer_terminate(tsdn_t *tsdn, buf_writer_t *buf_writer);
+
+typedef ssize_t (read_cb_t)(void *read_cbopaque, void *buf, size_t limit);
+void buf_writer_pipe(buf_writer_t *buf_writer, read_cb_t *read_cb,
+    void *read_cbopaque);
+
+#endif /* JEMALLOC_INTERNAL_BUF_WRITER_H */

BeefRT/JEMalloc/include/jemalloc/internal/cache_bin.h

@@ -0,0 +1,670 @@
+#ifndef JEMALLOC_INTERNAL_CACHE_BIN_H
+#define JEMALLOC_INTERNAL_CACHE_BIN_H
+
+#include "jemalloc/internal/ql.h"
+#include "jemalloc/internal/sz.h"
+
+/*
+ * The cache_bins are the mechanism that the tcache and the arena use to
+ * communicate.  The tcache fills from and flushes to the arena by passing a
+ * cache_bin_t to fill/flush.  When the arena needs to pull stats from the
+ * tcaches associated with it, it does so by iterating over its
+ * cache_bin_array_descriptor_t objects and reading out per-bin stats it
+ * contains.  This makes it so that the arena need not know about the existence
+ * of the tcache at all.
+ */
+
+/*
+ * The size in bytes of each cache bin stack.  We also use this to indicate
+ * *counts* of individual objects.
+ */
+typedef uint16_t cache_bin_sz_t;
+
+/*
+ * Leave a noticeable mark pattern on the cache bin stack boundaries, in case a
+ * bug starts leaking those.  Make it look like the junk pattern but be distinct
+ * from it.
+ */
+static const uintptr_t cache_bin_preceding_junk =
+    (uintptr_t)0x7a7a7a7a7a7a7a7aULL;
+/* Note: a7 vs. 7a above -- this tells you which pointer leaked. */
+static const uintptr_t cache_bin_trailing_junk =
+    (uintptr_t)0xa7a7a7a7a7a7a7a7ULL;
+
+/*
+ * That implies the following value, for the maximum number of items in any
+ * individual bin.  The cache bins track their bounds looking just at the low
+ * bits of a pointer, compared against a cache_bin_sz_t.  So that's
+ *   1 << (sizeof(cache_bin_sz_t) * 8)
+ * bytes spread across pointer sized objects to get the maximum.
+ */
+#define CACHE_BIN_NCACHED_MAX (((size_t)1 << sizeof(cache_bin_sz_t) * 8) \
+    / sizeof(void *) - 1)
+
+/*
+ * This lives inside the cache_bin (for locality reasons), and is initialized
+ * alongside it, but is otherwise not modified by any cache bin operations.
+ * It's logically public and maintained by its callers.
+ */
+typedef struct cache_bin_stats_s cache_bin_stats_t;
+struct cache_bin_stats_s {
+	/*
+	 * Number of allocation requests that corresponded to the size of this
+	 * bin.
+	 */
+	uint64_t nrequests;
+};
+
+/*
+ * Read-only information associated with each element of tcache_t's tbins array
+ * is stored separately, mainly to reduce memory usage.
+ */
+typedef struct cache_bin_info_s cache_bin_info_t;
+struct cache_bin_info_s {
+	cache_bin_sz_t ncached_max;
+};
+
+/*
+ * Responsible for caching allocations associated with a single size.
+ *
+ * Several pointers are used to track the stack.  To save on metadata bytes,
+ * only the stack_head is a full sized pointer (which is dereferenced on the
+ * fastpath), while the others store only the low 16 bits -- this is correct
+ * because a single stack never takes more space than 2^16 bytes, and at the
+ * same time only equality checks are performed on the low bits.
+ *
+ * (low addr)                                                  (high addr)
+ * |------stashed------|------available------|------cached-----|
+ * ^                   ^                     ^                 ^
+ * low_bound(derived)  low_bits_full         stack_head        low_bits_empty
+ */
+typedef struct cache_bin_s cache_bin_t;
+struct cache_bin_s {
+	/*
+	 * The stack grows down.  Whenever the bin is nonempty, the head points
+	 * to an array entry containing a valid allocation.  When it is empty,
+	 * the head points to one element past the owned array.
+	 */
+	void **stack_head;
+	/*
+	 * cur_ptr and stats are both modified frequently.  Let's keep them
+	 * close so that they have a higher chance of being on the same
+	 * cacheline, thus less write-backs.
+	 */
+	cache_bin_stats_t tstats;
+
+	/*
+	 * The low bits of the address of the first item in the stack that
+	 * hasn't been used since the last GC, to track the low water mark (min
+	 * # of cached items).
+	 *
+	 * Since the stack grows down, this is a higher address than
+	 * low_bits_full.
+	 */
+	uint16_t low_bits_low_water;
+
+	/*
+	 * The low bits of the value that stack_head will take on when the array
+	 * is full (of cached & stashed items).  But remember that stack_head
+	 * always points to a valid item when the array is nonempty -- this is
+	 * in the array.
+	 *
+	 * Recall that since the stack grows down, this is the lowest available
+	 * address in the array for caching.  Only adjusted when stashing items.
+	 */
+	uint16_t low_bits_full;
+
+	/*
+	 * The low bits of the value that stack_head will take on when the array
+	 * is empty.
+	 *
+	 * The stack grows down -- this is one past the highest address in the
+	 * array.  Immutable after initialization.
+	 */
+	uint16_t low_bits_empty;
+};
+
+/*
+ * The cache_bins live inside the tcache, but the arena (by design) isn't
+ * supposed to know much about tcache internals.  To let the arena iterate over
+ * associated bins, we keep (with the tcache) a linked list of
+ * cache_bin_array_descriptor_ts that tell the arena how to find the bins.
+ */
+typedef struct cache_bin_array_descriptor_s cache_bin_array_descriptor_t;
+struct cache_bin_array_descriptor_s {
+	/*
+	 * The arena keeps a list of the cache bins associated with it, for
+	 * stats collection.
+	 */
+	ql_elm(cache_bin_array_descriptor_t) link;
+	/* Pointers to the tcache bins. */
+	cache_bin_t *bins;
+};
+
+static inline void
+cache_bin_array_descriptor_init(cache_bin_array_descriptor_t *descriptor,
+    cache_bin_t *bins) {
+	ql_elm_new(descriptor, link);
+	descriptor->bins = bins;
+}
+
+JEMALLOC_ALWAYS_INLINE bool
+cache_bin_nonfast_aligned(const void *ptr) {
+	if (!config_uaf_detection) {
+		return false;
+	}
+	/*
+	 * Currently we use alignment to decide which pointer to junk & stash on
+	 * dealloc (for catching use-after-free).  In some common cases a
+	 * page-aligned check is needed already (sdalloc w/ config_prof), so we
+	 * are getting it more or less for free -- no added instructions on
+	 * free_fastpath.
+	 *
+	 * Another way of deciding which pointer to sample, is adding another
+	 * thread_event to pick one every N bytes.  That also adds no cost on
+	 * the fastpath, however it will tend to pick large allocations which is
+	 * not the desired behavior.
+	 */
+	return ((uintptr_t)ptr & san_cache_bin_nonfast_mask) == 0;
+}
+
+/* Returns ncached_max: Upper limit on ncached. */
+static inline cache_bin_sz_t
+cache_bin_info_ncached_max(cache_bin_info_t *info) {
+	return info->ncached_max;
+}
+
+/*
+ * Internal.
+ *
+ * Asserts that the pointer associated with earlier is <= the one associated
+ * with later.
+ */
+static inline void
+cache_bin_assert_earlier(cache_bin_t *bin, uint16_t earlier, uint16_t later) {
+	if (earlier > later) {
+		assert(bin->low_bits_full > bin->low_bits_empty);
+	}
+}
+
+/*
+ * Internal.
+ *
+ * Does difference calculations that handle wraparound correctly.  Earlier must
+ * be associated with the position earlier in memory.
+ */
+static inline uint16_t
+cache_bin_diff(cache_bin_t *bin, uint16_t earlier, uint16_t later, bool racy) {
+	/*
+	 * When it's racy, bin->low_bits_full can be modified concurrently. It
+	 * can cross the uint16_t max value and become less than
+	 * bin->low_bits_empty at the time of the check.
+	 */
+	if (!racy) {
+		cache_bin_assert_earlier(bin, earlier, later);
+	}
+	return later - earlier;
+}
+
+/*
+ * Number of items currently cached in the bin, without checking ncached_max.
+ * We require specifying whether or not the request is racy or not (i.e. whether
+ * or not concurrent modifications are possible).
+ */
+static inline cache_bin_sz_t
+cache_bin_ncached_get_internal(cache_bin_t *bin, bool racy) {
+	cache_bin_sz_t diff = cache_bin_diff(bin,
+	    (uint16_t)(uintptr_t)bin->stack_head, bin->low_bits_empty, racy);
+	cache_bin_sz_t n = diff / sizeof(void *);
+	/*
+	 * We have undefined behavior here; if this function is called from the
+	 * arena stats updating code, then stack_head could change from the
+	 * first line to the next one.  Morally, these loads should be atomic,
+	 * but compilers won't currently generate comparisons with in-memory
+	 * operands against atomics, and these variables get accessed on the
+	 * fast paths.  This should still be "safe" in the sense of generating
+	 * the correct assembly for the foreseeable future, though.
+	 */
+	assert(n == 0 || *(bin->stack_head) != NULL || racy);
+	return n;
+}
+
+/*
+ * Number of items currently cached in the bin, with checking ncached_max.  The
+ * caller must know that no concurrent modification of the cache_bin is
+ * possible.
+ */
+static inline cache_bin_sz_t
+cache_bin_ncached_get_local(cache_bin_t *bin, cache_bin_info_t *info) {
+	cache_bin_sz_t n = cache_bin_ncached_get_internal(bin,
+	    /* racy */ false);
+	assert(n <= cache_bin_info_ncached_max(info));
+	return n;
+}
+
+/*
+ * Internal.
+ *
+ * A pointer to the position one past the end of the backing array.
+ *
+ * Do not call if racy, because both 'bin->stack_head' and 'bin->low_bits_full'
+ * are subject to concurrent modifications.
+ */
+static inline void **
+cache_bin_empty_position_get(cache_bin_t *bin) {
+	cache_bin_sz_t diff = cache_bin_diff(bin,
+	    (uint16_t)(uintptr_t)bin->stack_head, bin->low_bits_empty,
+	    /* racy */ false);
+	uintptr_t empty_bits = (uintptr_t)bin->stack_head + diff;
+	void **ret = (void **)empty_bits;
+
+	assert(ret >= bin->stack_head);
+
+	return ret;
+}
+
+/*
+ * Internal.
+ *
+ * Calculates low bits of the lower bound of the usable cache bin's range (see
+ * cache_bin_t visual representation above).
+ *
+ * No values are concurrently modified, so should be safe to read in a
+ * multithreaded environment. Currently concurrent access happens only during
+ * arena statistics collection.
+ */
+static inline uint16_t
+cache_bin_low_bits_low_bound_get(cache_bin_t *bin, cache_bin_info_t *info) {
+	return (uint16_t)bin->low_bits_empty -
+	    info->ncached_max * sizeof(void *);
+}
+
+/*
+ * Internal.
+ *
+ * A pointer to the position with the lowest address of the backing array.
+ */
+static inline void **
+cache_bin_low_bound_get(cache_bin_t *bin, cache_bin_info_t *info) {
+	cache_bin_sz_t ncached_max = cache_bin_info_ncached_max(info);
+	void **ret = cache_bin_empty_position_get(bin) - ncached_max;
+	assert(ret <= bin->stack_head);
+
+	return ret;
+}
+
+/*
+ * As the name implies.  This is important since it's not correct to try to
+ * batch fill a nonempty cache bin.
+ */
+static inline void
+cache_bin_assert_empty(cache_bin_t *bin, cache_bin_info_t *info) {
+	assert(cache_bin_ncached_get_local(bin, info) == 0);
+	assert(cache_bin_empty_position_get(bin) == bin->stack_head);
+}
+
+/*
+ * Get low water, but without any of the correctness checking we do for the
+ * caller-usable version, if we are temporarily breaking invariants (like
+ * ncached >= low_water during flush).
+ */
+static inline cache_bin_sz_t
+cache_bin_low_water_get_internal(cache_bin_t *bin) {
+	return cache_bin_diff(bin, bin->low_bits_low_water,
+	    bin->low_bits_empty, /* racy */ false) / sizeof(void *);
+}
+
+/* Returns the numeric value of low water in [0, ncached]. */
+static inline cache_bin_sz_t
+cache_bin_low_water_get(cache_bin_t *bin, cache_bin_info_t *info) {
+	cache_bin_sz_t low_water = cache_bin_low_water_get_internal(bin);
+	assert(low_water <= cache_bin_info_ncached_max(info));
+	assert(low_water <= cache_bin_ncached_get_local(bin, info));
+
+	cache_bin_assert_earlier(bin, (uint16_t)(uintptr_t)bin->stack_head,
+	    bin->low_bits_low_water);
+
+	return low_water;
+}
+
+/*
+ * Indicates that the current cache bin position should be the low water mark
+ * going forward.
+ */
+static inline void
+cache_bin_low_water_set(cache_bin_t *bin) {
+	bin->low_bits_low_water = (uint16_t)(uintptr_t)bin->stack_head;
+}
+
+static inline void
+cache_bin_low_water_adjust(cache_bin_t *bin) {
+	if (cache_bin_ncached_get_internal(bin, /* racy */ false)
+	    < cache_bin_low_water_get_internal(bin)) {
+		cache_bin_low_water_set(bin);
+	}
+}
+
+JEMALLOC_ALWAYS_INLINE void *
+cache_bin_alloc_impl(cache_bin_t *bin, bool *success, bool adjust_low_water) {
+	/*
+	 * success (instead of ret) should be checked upon the return of this
+	 * function.  We avoid checking (ret == NULL) because there is never a
+	 * null stored on the avail stack (which is unknown to the compiler),
+	 * and eagerly checking ret would cause pipeline stall (waiting for the
+	 * cacheline).
+	 */
+
+	/*
+	 * This may read from the empty position; however the loaded value won't
+	 * be used.  It's safe because the stack has one more slot reserved.
+	 */
+	void *ret = *bin->stack_head;
+	uint16_t low_bits = (uint16_t)(uintptr_t)bin->stack_head;
+	void **new_head = bin->stack_head + 1;
+
+	/*
+	 * Note that the low water mark is at most empty; if we pass this check,
+	 * we know we're non-empty.
+	 */
+	if (likely(low_bits != bin->low_bits_low_water)) {
+		bin->stack_head = new_head;
+		*success = true;
+		return ret;
+	}
+	if (!adjust_low_water) {
+		*success = false;
+		return NULL;
+	}
+	/*
+	 * In the fast-path case where we call alloc_easy and then alloc, the
+	 * previous checking and computation is optimized away -- we didn't
+	 * actually commit any of our operations.
+	 */
+	if (likely(low_bits != bin->low_bits_empty)) {
+		bin->stack_head = new_head;
+		bin->low_bits_low_water = (uint16_t)(uintptr_t)new_head;
+		*success = true;
+		return ret;
+	}
+	*success = false;
+	return NULL;
+}
+
+/*
+ * Allocate an item out of the bin, failing if we're at the low-water mark.
+ */
+JEMALLOC_ALWAYS_INLINE void *
+cache_bin_alloc_easy(cache_bin_t *bin, bool *success) {
+	/* We don't look at info if we're not adjusting low-water. */
+	return cache_bin_alloc_impl(bin, success, false);
+}
+
+/*
+ * Allocate an item out of the bin, even if we're currently at the low-water
+ * mark (and failing only if the bin is empty).
+ */
+JEMALLOC_ALWAYS_INLINE void *
+cache_bin_alloc(cache_bin_t *bin, bool *success) {
+	return cache_bin_alloc_impl(bin, success, true);
+}
+
+JEMALLOC_ALWAYS_INLINE cache_bin_sz_t
+cache_bin_alloc_batch(cache_bin_t *bin, size_t num, void **out) {
+	cache_bin_sz_t n = cache_bin_ncached_get_internal(bin,
+	    /* racy */ false);
+	if (n > num) {
+		n = (cache_bin_sz_t)num;
+	}
+	memcpy(out, bin->stack_head, n * sizeof(void *));
+	bin->stack_head += n;
+	cache_bin_low_water_adjust(bin);
+
+	return n;
+}
+
+JEMALLOC_ALWAYS_INLINE bool
+cache_bin_full(cache_bin_t *bin) {
+	return ((uint16_t)(uintptr_t)bin->stack_head == bin->low_bits_full);
+}
+
+/*
+ * Free an object into the given bin.  Fails only if the bin is full.
+ */
+JEMALLOC_ALWAYS_INLINE bool
+cache_bin_dalloc_easy(cache_bin_t *bin, void *ptr) {
+	if (unlikely(cache_bin_full(bin))) {
+		return false;
+	}
+
+	bin->stack_head--;
+	*bin->stack_head = ptr;
+	cache_bin_assert_earlier(bin, bin->low_bits_full,
+	    (uint16_t)(uintptr_t)bin->stack_head);
+
+	return true;
+}
+
+/* Returns false if failed to stash (i.e. bin is full). */
+JEMALLOC_ALWAYS_INLINE bool
+cache_bin_stash(cache_bin_t *bin, void *ptr) {
+	if (cache_bin_full(bin)) {
+		return false;
+	}
+
+	/* Stash at the full position, in the [full, head) range. */
+	uint16_t low_bits_head = (uint16_t)(uintptr_t)bin->stack_head;
+	/* Wraparound handled as well. */
+	uint16_t diff = cache_bin_diff(bin, bin->low_bits_full, low_bits_head,
+	    /* racy */ false);
+	*(void **)((uintptr_t)bin->stack_head - diff) = ptr;
+
+	assert(!cache_bin_full(bin));
+	bin->low_bits_full += sizeof(void *);
+	cache_bin_assert_earlier(bin, bin->low_bits_full, low_bits_head);
+
+	return true;
+}
+
+/*
+ * Get the number of stashed pointers.
+ *
+ * When called from a thread not owning the TLS (i.e. racy = true), it's
+ * important to keep in mind that 'bin->stack_head' and 'bin->low_bits_full' can
+ * be modified concurrently and almost none assertions about their values can be
+ * made.
+ */
+JEMALLOC_ALWAYS_INLINE cache_bin_sz_t
+cache_bin_nstashed_get_internal(cache_bin_t *bin, cache_bin_info_t *info,
+    bool racy) {
+	cache_bin_sz_t ncached_max = cache_bin_info_ncached_max(info);
+	uint16_t low_bits_low_bound = cache_bin_low_bits_low_bound_get(bin,
+	    info);
+
+	cache_bin_sz_t n = cache_bin_diff(bin, low_bits_low_bound,
+	    bin->low_bits_full, racy) / sizeof(void *);
+	assert(n <= ncached_max);
+
+	if (!racy) {
+		/* Below are for assertions only. */
+		void **low_bound = cache_bin_low_bound_get(bin, info);
+
+		assert((uint16_t)(uintptr_t)low_bound == low_bits_low_bound);
+		void *stashed = *(low_bound + n - 1);
+		bool aligned = cache_bin_nonfast_aligned(stashed);
+#ifdef JEMALLOC_JET
+		/* Allow arbitrary pointers to be stashed in tests. */
+		aligned = true;
+#endif
+		assert(n == 0 || (stashed != NULL && aligned));
+	}
+
+	return n;
+}
+
+JEMALLOC_ALWAYS_INLINE cache_bin_sz_t
+cache_bin_nstashed_get_local(cache_bin_t *bin, cache_bin_info_t *info) {
+	cache_bin_sz_t n = cache_bin_nstashed_get_internal(bin, info,
+	    /* racy */ false);
+	assert(n <= cache_bin_info_ncached_max(info));
+	return n;
+}
+
+/*
+ * Obtain a racy view of the number of items currently in the cache bin, in the
+ * presence of possible concurrent modifications.
+ */
+static inline void
+cache_bin_nitems_get_remote(cache_bin_t *bin, cache_bin_info_t *info,
+    cache_bin_sz_t *ncached, cache_bin_sz_t *nstashed) {
+	cache_bin_sz_t n = cache_bin_ncached_get_internal(bin, /* racy */ true);
+	assert(n <= cache_bin_info_ncached_max(info));
+	*ncached = n;
+
+	n = cache_bin_nstashed_get_internal(bin, info, /* racy */ true);
+	assert(n <= cache_bin_info_ncached_max(info));
+	*nstashed = n;
+	/* Note that cannot assert ncached + nstashed <= ncached_max (racy). */
+}
+
+/*
+ * Filling and flushing are done in batch, on arrays of void *s.  For filling,
+ * the arrays go forward, and can be accessed with ordinary array arithmetic.
+ * For flushing, we work from the end backwards, and so need to use special
+ * accessors that invert the usual ordering.
+ *
+ * This is important for maintaining first-fit; the arena code fills with
+ * earliest objects first, and so those are the ones we should return first for
+ * cache_bin_alloc calls.  When flushing, we should flush the objects that we
+ * wish to return later; those at the end of the array.  This is better for the
+ * first-fit heuristic as well as for cache locality; the most recently freed
+ * objects are the ones most likely to still be in cache.
+ *
+ * This all sounds very hand-wavey and theoretical, but reverting the ordering
+ * on one or the other pathway leads to measurable slowdowns.
+ */
+
+typedef struct cache_bin_ptr_array_s cache_bin_ptr_array_t;
+struct cache_bin_ptr_array_s {
+	cache_bin_sz_t n;
+	void **ptr;
+};
+
+/*
+ * Declare a cache_bin_ptr_array_t sufficient for nval items.
+ *
+ * In the current implementation, this could be just part of a
+ * cache_bin_ptr_array_init_... call, since we reuse the cache bin stack memory.
+ * Indirecting behind a macro, though, means experimenting with linked-list
+ * representations is easy (since they'll require an alloca in the calling
+ * frame).
+ */
+#define CACHE_BIN_PTR_ARRAY_DECLARE(name, nval)				\
+    cache_bin_ptr_array_t name;						\
+    name.n = (nval)
+
+/*
+ * Start a fill.  The bin must be empty, and This must be followed by a
+ * finish_fill call before doing any alloc/dalloc operations on the bin.
+ */
+static inline void
+cache_bin_init_ptr_array_for_fill(cache_bin_t *bin, cache_bin_info_t *info,
+    cache_bin_ptr_array_t *arr, cache_bin_sz_t nfill) {
+	cache_bin_assert_empty(bin, info);
+	arr->ptr = cache_bin_empty_position_get(bin) - nfill;
+}
+
+/*
+ * While nfill in cache_bin_init_ptr_array_for_fill is the number we *intend* to
+ * fill, nfilled here is the number we actually filled (which may be less, in
+ * case of OOM.
+ */
+static inline void
+cache_bin_finish_fill(cache_bin_t *bin, cache_bin_info_t *info,
+    cache_bin_ptr_array_t *arr, cache_bin_sz_t nfilled) {
+	cache_bin_assert_empty(bin, info);
+	void **empty_position = cache_bin_empty_position_get(bin);
+	if (nfilled < arr->n) {
+		memmove(empty_position - nfilled, empty_position - arr->n,
+		    nfilled * sizeof(void *));
+	}
+	bin->stack_head = empty_position - nfilled;
+}
+
+/*
+ * Same deal, but with flush.  Unlike fill (which can fail), the user must flush
+ * everything we give them.
+ */
+static inline void
+cache_bin_init_ptr_array_for_flush(cache_bin_t *bin, cache_bin_info_t *info,
+    cache_bin_ptr_array_t *arr, cache_bin_sz_t nflush) {
+	arr->ptr = cache_bin_empty_position_get(bin) - nflush;
+	assert(cache_bin_ncached_get_local(bin, info) == 0
+	    || *arr->ptr != NULL);
+}
+
+static inline void
+cache_bin_finish_flush(cache_bin_t *bin, cache_bin_info_t *info,
+    cache_bin_ptr_array_t *arr, cache_bin_sz_t nflushed) {
+	unsigned rem = cache_bin_ncached_get_local(bin, info) - nflushed;
+	memmove(bin->stack_head + nflushed, bin->stack_head,
+	    rem * sizeof(void *));
+	bin->stack_head = bin->stack_head + nflushed;
+	cache_bin_low_water_adjust(bin);
+}
+
+static inline void
+cache_bin_init_ptr_array_for_stashed(cache_bin_t *bin, szind_t binind,
+    cache_bin_info_t *info, cache_bin_ptr_array_t *arr,
+    cache_bin_sz_t nstashed) {
+	assert(nstashed > 0);
+	assert(cache_bin_nstashed_get_local(bin, info) == nstashed);
+
+	void **low_bound = cache_bin_low_bound_get(bin, info);
+	arr->ptr = low_bound;
+	assert(*arr->ptr != NULL);
+}
+
+static inline void
+cache_bin_finish_flush_stashed(cache_bin_t *bin, cache_bin_info_t *info) {
+	void **low_bound = cache_bin_low_bound_get(bin, info);
+
+	/* Reset the bin local full position. */
+	bin->low_bits_full = (uint16_t)(uintptr_t)low_bound;
+	assert(cache_bin_nstashed_get_local(bin, info) == 0);
+}
+
+/*
+ * Initialize a cache_bin_info to represent up to the given number of items in
+ * the cache_bins it is associated with.
+ */
+void cache_bin_info_init(cache_bin_info_t *bin_info,
+    cache_bin_sz_t ncached_max);
+/*
+ * Given an array of initialized cache_bin_info_ts, determine how big an
+ * allocation is required to initialize a full set of cache_bin_ts.
+ */
+void cache_bin_info_compute_alloc(cache_bin_info_t *infos, szind_t ninfos,
+    size_t *size, size_t *alignment);
+
+/*
+ * Actually initialize some cache bins.  Callers should allocate the backing
+ * memory indicated by a call to cache_bin_compute_alloc.  They should then
+ * preincrement, call init once for each bin and info, and then call
+ * cache_bin_postincrement.  *alloc_cur will then point immediately past the end
+ * of the allocation.
+ */
+void cache_bin_preincrement(cache_bin_info_t *infos, szind_t ninfos,
+    void *alloc, size_t *cur_offset);
+void cache_bin_postincrement(cache_bin_info_t *infos, szind_t ninfos,
+    void *alloc, size_t *cur_offset);
+void cache_bin_init(cache_bin_t *bin, cache_bin_info_t *info, void *alloc,
+    size_t *cur_offset);
+
+/*
+ * If a cache bin was zero initialized (either because it lives in static or
+ * thread-local storage, or was memset to 0), this function indicates whether or
+ * not cache_bin_init was called on it.
+ */
+bool cache_bin_still_zero_initialized(cache_bin_t *bin);
+
+#endif /* JEMALLOC_INTERNAL_CACHE_BIN_H */

BeefRT/JEMalloc/include/jemalloc/internal/ckh.h

@@ -0,0 +1,101 @@
+#ifndef JEMALLOC_INTERNAL_CKH_H
+#define JEMALLOC_INTERNAL_CKH_H
+
+#include "jemalloc/internal/tsd.h"
+
+/* Cuckoo hashing implementation.  Skip to the end for the interface. */
+
+/******************************************************************************/
+/* INTERNAL DEFINITIONS -- IGNORE */
+/******************************************************************************/
+
+/* Maintain counters used to get an idea of performance. */
+/* #define CKH_COUNT */
+/* Print counter values in ckh_delete() (requires CKH_COUNT). */
+/* #define CKH_VERBOSE */
+
+/*
+ * There are 2^LG_CKH_BUCKET_CELLS cells in each hash table bucket.  Try to fit
+ * one bucket per L1 cache line.
+ */
+#define LG_CKH_BUCKET_CELLS (LG_CACHELINE - LG_SIZEOF_PTR - 1)
+
+/* Typedefs to allow easy function pointer passing. */
+typedef void ckh_hash_t (const void *, size_t[2]);
+typedef bool ckh_keycomp_t (const void *, const void *);
+
+/* Hash table cell. */
+typedef struct {
+	const void *key;
+	const void *data;
+} ckhc_t;
+
+/* The hash table itself. */
+typedef struct {
+#ifdef CKH_COUNT
+	/* Counters used to get an idea of performance. */
+	uint64_t ngrows;
+	uint64_t nshrinks;
+	uint64_t nshrinkfails;
+	uint64_t ninserts;
+	uint64_t nrelocs;
+#endif
+
+	/* Used for pseudo-random number generation. */
+	uint64_t prng_state;
+
+	/* Total number of items. */
+	size_t count;
+
+	/*
+	 * Minimum and current number of hash table buckets.  There are
+	 * 2^LG_CKH_BUCKET_CELLS cells per bucket.
+	 */
+	unsigned lg_minbuckets;
+	unsigned lg_curbuckets;
+
+	/* Hash and comparison functions. */
+	ckh_hash_t *hash;
+	ckh_keycomp_t *keycomp;
+
+	/* Hash table with 2^lg_curbuckets buckets. */
+	ckhc_t *tab;
+} ckh_t;
+
+/******************************************************************************/
+/* BEGIN PUBLIC API */
+/******************************************************************************/
+
+/* Lifetime management.  Minitems is the initial capacity. */
+bool ckh_new(tsd_t *tsd, ckh_t *ckh, size_t minitems, ckh_hash_t *hash,
+    ckh_keycomp_t *keycomp);
+void ckh_delete(tsd_t *tsd, ckh_t *ckh);
+
+/* Get the number of elements in the set. */
+size_t ckh_count(ckh_t *ckh);
+
+/*
+ * To iterate over the elements in the table, initialize *tabind to 0 and call
+ * this function until it returns true.  Each call that returns false will
+ * update *key and *data to the next element in the table, assuming the pointers
+ * are non-NULL.
+ */
+bool ckh_iter(ckh_t *ckh, size_t *tabind, void **key, void **data);
+
+/*
+ * Basic hash table operations -- insert, removal, lookup.  For ckh_remove and
+ * ckh_search, key or data can be NULL.  The hash-table only stores pointers to
+ * the key and value, and doesn't do any lifetime management.
+ */
+bool ckh_insert(tsd_t *tsd, ckh_t *ckh, const void *key, const void *data);
+bool ckh_remove(tsd_t *tsd, ckh_t *ckh, const void *searchkey, void **key,
+    void **data);
+bool ckh_search(ckh_t *ckh, const void *searchkey, void **key, void **data);
+
+/* Some useful hash and comparison functions for strings and pointers. */
+void ckh_string_hash(const void *key, size_t r_hash[2]);
+bool ckh_string_keycomp(const void *k1, const void *k2);
+void ckh_pointer_hash(const void *key, size_t r_hash[2]);
+bool ckh_pointer_keycomp(const void *k1, const void *k2);
+
+#endif /* JEMALLOC_INTERNAL_CKH_H */

BeefRT/JEMalloc/include/jemalloc/internal/counter.h

@@ -0,0 +1,34 @@
+#ifndef JEMALLOC_INTERNAL_COUNTER_H
+#define JEMALLOC_INTERNAL_COUNTER_H
+
+#include "jemalloc/internal/mutex.h"
+
+typedef struct counter_accum_s {
+	LOCKEDINT_MTX_DECLARE(mtx)
+	locked_u64_t accumbytes;
+	uint64_t interval;
+} counter_accum_t;
+
+JEMALLOC_ALWAYS_INLINE bool
+counter_accum(tsdn_t *tsdn, counter_accum_t *counter, uint64_t bytes) {
+	uint64_t interval = counter->interval;
+	assert(interval > 0);
+	LOCKEDINT_MTX_LOCK(tsdn, counter->mtx);
+	/*
+	 * If the event moves fast enough (and/or if the event handling is slow
+	 * enough), extreme overflow can cause counter trigger coalescing.
+	 * This is an intentional mechanism that avoids rate-limiting
+	 * allocation.
+	 */
+	bool overflow = locked_inc_mod_u64(tsdn, LOCKEDINT_MTX(counter->mtx),
+	    &counter->accumbytes, bytes, interval);
+	LOCKEDINT_MTX_UNLOCK(tsdn, counter->mtx);
+	return overflow;
+}
+
+bool counter_accum_init(counter_accum_t *counter, uint64_t interval);
+void counter_prefork(tsdn_t *tsdn, counter_accum_t *counter);
+void counter_postfork_parent(tsdn_t *tsdn, counter_accum_t *counter);
+void counter_postfork_child(tsdn_t *tsdn, counter_accum_t *counter);
+
+#endif /* JEMALLOC_INTERNAL_COUNTER_H */

BeefRT/JEMalloc/include/jemalloc/internal/ctl.h

@@ -0,0 +1,159 @@
+#ifndef JEMALLOC_INTERNAL_CTL_H
+#define JEMALLOC_INTERNAL_CTL_H
+
+#include "jemalloc/internal/jemalloc_internal_types.h"
+#include "jemalloc/internal/malloc_io.h"
+#include "jemalloc/internal/mutex_prof.h"
+#include "jemalloc/internal/ql.h"
+#include "jemalloc/internal/sc.h"
+#include "jemalloc/internal/stats.h"
+
+/* Maximum ctl tree depth. */
+#define CTL_MAX_DEPTH	7
+
+typedef struct ctl_node_s {
+	bool named;
+} ctl_node_t;
+
+typedef struct ctl_named_node_s {
+	ctl_node_t node;
+	const char *name;
+	/* If (nchildren == 0), this is a terminal node. */
+	size_t nchildren;
+	const ctl_node_t *children;
+	int (*ctl)(tsd_t *, const size_t *, size_t, void *, size_t *, void *,
+	    size_t);
+} ctl_named_node_t;
+
+typedef struct ctl_indexed_node_s {
+	struct ctl_node_s node;
+	const ctl_named_node_t *(*index)(tsdn_t *, const size_t *, size_t,
+	    size_t);
+} ctl_indexed_node_t;
+
+typedef struct ctl_arena_stats_s {
+	arena_stats_t astats;
+
+	/* Aggregate stats for small size classes, based on bin stats. */
+	size_t allocated_small;
+	uint64_t nmalloc_small;
+	uint64_t ndalloc_small;
+	uint64_t nrequests_small;
+	uint64_t nfills_small;
+	uint64_t nflushes_small;
+
+	bin_stats_data_t bstats[SC_NBINS];
+	arena_stats_large_t lstats[SC_NSIZES - SC_NBINS];
+	pac_estats_t estats[SC_NPSIZES];
+	hpa_shard_stats_t hpastats;
+	sec_stats_t secstats;
+} ctl_arena_stats_t;
+
+typedef struct ctl_stats_s {
+	size_t allocated;
+	size_t active;
+	size_t metadata;
+	size_t metadata_thp;
+	size_t resident;
+	size_t mapped;
+	size_t retained;
+
+	background_thread_stats_t background_thread;
+	mutex_prof_data_t mutex_prof_data[mutex_prof_num_global_mutexes];
+} ctl_stats_t;
+
+typedef struct ctl_arena_s ctl_arena_t;
+struct ctl_arena_s {
+	unsigned arena_ind;
+	bool initialized;
+	ql_elm(ctl_arena_t) destroyed_link;
+
+	/* Basic stats, supported even if !config_stats. */
+	unsigned nthreads;
+	const char *dss;
+	ssize_t dirty_decay_ms;
+	ssize_t muzzy_decay_ms;
+	size_t pactive;
+	size_t pdirty;
+	size_t pmuzzy;
+
+	/* NULL if !config_stats. */
+	ctl_arena_stats_t *astats;
+};
+
+typedef struct ctl_arenas_s {
+	uint64_t epoch;
+	unsigned narenas;
+	ql_head(ctl_arena_t) destroyed;
+
+	/*
+	 * Element 0 corresponds to merged stats for extant arenas (accessed via
+	 * MALLCTL_ARENAS_ALL), element 1 corresponds to merged stats for
+	 * destroyed arenas (accessed via MALLCTL_ARENAS_DESTROYED), and the
+	 * remaining MALLOCX_ARENA_LIMIT elements correspond to arenas.
+	 */
+	ctl_arena_t *arenas[2 + MALLOCX_ARENA_LIMIT];
+} ctl_arenas_t;
+
+int ctl_byname(tsd_t *tsd, const char *name, void *oldp, size_t *oldlenp,
+    void *newp, size_t newlen);
+int ctl_nametomib(tsd_t *tsd, const char *name, size_t *mibp, size_t *miblenp);
+int ctl_bymib(tsd_t *tsd, const size_t *mib, size_t miblen, void *oldp,
+    size_t *oldlenp, void *newp, size_t newlen);
+int ctl_mibnametomib(tsd_t *tsd, size_t *mib, size_t miblen, const char *name,
+    size_t *miblenp);
+int ctl_bymibname(tsd_t *tsd, size_t *mib, size_t miblen, const char *name,
+    size_t *miblenp, void *oldp, size_t *oldlenp, void *newp, size_t newlen);
+bool ctl_boot(void);
+void ctl_prefork(tsdn_t *tsdn);
+void ctl_postfork_parent(tsdn_t *tsdn);
+void ctl_postfork_child(tsdn_t *tsdn);
+void ctl_mtx_assert_held(tsdn_t *tsdn);
+
+#define xmallctl(name, oldp, oldlenp, newp, newlen) do {		\
+	if (je_mallctl(name, oldp, oldlenp, newp, newlen)		\
+	    != 0) {							\
+		malloc_printf(						\
+		    "<jemalloc>: Failure in xmallctl(\"%s\", ...)\n",	\
+		    name);						\
+		abort();						\
+	}								\
+} while (0)
+
+#define xmallctlnametomib(name, mibp, miblenp) do {			\
+	if (je_mallctlnametomib(name, mibp, miblenp) != 0) {		\
+		malloc_printf("<jemalloc>: Failure in "			\
+		    "xmallctlnametomib(\"%s\", ...)\n", name);		\
+		abort();						\
+	}								\
+} while (0)
+
+#define xmallctlbymib(mib, miblen, oldp, oldlenp, newp, newlen) do {	\
+	if (je_mallctlbymib(mib, miblen, oldp, oldlenp, newp,		\
+	    newlen) != 0) {						\
+		malloc_write(						\
+		    "<jemalloc>: Failure in xmallctlbymib()\n");	\
+		abort();						\
+	}								\
+} while (0)
+
+#define xmallctlmibnametomib(mib, miblen, name, miblenp) do {		\
+	if (ctl_mibnametomib(tsd_fetch(), mib, miblen, name, miblenp)	\
+	    != 0) {							\
+		malloc_write(						\
+		    "<jemalloc>: Failure in ctl_mibnametomib()\n");	\
+		abort();						\
+	}								\
+} while (0)
+
+#define xmallctlbymibname(mib, miblen, name, miblenp, oldp, oldlenp,	\
+    newp, newlen) do {							\
+	if (ctl_bymibname(tsd_fetch(), mib, miblen, name, miblenp,	\
+	    oldp, oldlenp, newp, newlen) != 0) {			\
+		malloc_write(						\
+		    "<jemalloc>: Failure in ctl_bymibname()\n");	\
+		abort();						\
+	}								\
+} while (0)
+
+#endif /* JEMALLOC_INTERNAL_CTL_H */

BeefRT/JEMalloc/include/jemalloc/internal/decay.h

@@ -0,0 +1,186 @@
+#ifndef JEMALLOC_INTERNAL_DECAY_H
+#define JEMALLOC_INTERNAL_DECAY_H
+
+#include "jemalloc/internal/smoothstep.h"
+
+#define DECAY_UNBOUNDED_TIME_TO_PURGE ((uint64_t)-1)
+
+/*
+ * The decay_t computes the number of pages we should purge at any given time.
+ * Page allocators inform a decay object when pages enter a decay-able state
+ * (i.e. dirty or muzzy), and query it to determine how many pages should be
+ * purged at any given time.
+ *
+ * This is mostly a single-threaded data structure and doesn't care about
+ * synchronization at all; it's the caller's responsibility to manage their
+ * synchronization on their own.  There are two exceptions:
+ * 1) It's OK to racily call decay_ms_read (i.e. just the simplest state query).
+ * 2) The mtx and purging fields live (and are initialized) here, but are
+ *    logically owned by the page allocator.  This is just a convenience (since
+ *    those fields would be duplicated for both the dirty and muzzy states
+ *    otherwise).
+ */
+typedef struct decay_s decay_t;
+struct decay_s {
+	/* Synchronizes all non-atomic fields. */
+	malloc_mutex_t mtx;
+	/*
+	 * True if a thread is currently purging the extents associated with
+	 * this decay structure.
+	 */
+	bool purging;
+	/*
+	 * Approximate time in milliseconds from the creation of a set of unused
+	 * dirty pages until an equivalent set of unused dirty pages is purged
+	 * and/or reused.
+	 */
+	atomic_zd_t time_ms;
+	/* time / SMOOTHSTEP_NSTEPS. */
+	nstime_t interval;
+	/*
+	 * Time at which the current decay interval logically started.  We do
+	 * not actually advance to a new epoch until sometime after it starts
+	 * because of scheduling and computation delays, and it is even possible
+	 * to completely skip epochs.  In all cases, during epoch advancement we
+	 * merge all relevant activity into the most recently recorded epoch.
+	 */
+	nstime_t epoch;
+	/* Deadline randomness generator. */
+	uint64_t jitter_state;
+	/*
+	 * Deadline for current epoch.  This is the sum of interval and per
+	 * epoch jitter which is a uniform random variable in [0..interval).
+	 * Epochs always advance by precise multiples of interval, but we
+	 * randomize the deadline to reduce the likelihood of arenas purging in
+	 * lockstep.
+	 */
+	nstime_t deadline;
+	/*
+	 * The number of pages we cap ourselves at in the current epoch, per
+	 * decay policies.  Updated on an epoch change.  After an epoch change,
+	 * the caller should take steps to try to purge down to this amount.
+	 */
+	size_t npages_limit;
+	/*
+	 * Number of unpurged pages at beginning of current epoch.  During epoch
+	 * advancement we use the delta between arena->decay_*.nunpurged and
+	 * ecache_npages_get(&arena->ecache_*) to determine how many dirty pages,
+	 * if any, were generated.
+	 */
+	size_t nunpurged;
+	/*
+	 * Trailing log of how many unused dirty pages were generated during
+	 * each of the past SMOOTHSTEP_NSTEPS decay epochs, where the last
+	 * element is the most recent epoch.  Corresponding epoch times are
+	 * relative to epoch.
+	 *
+	 * Updated only on epoch advance, triggered by
+	 * decay_maybe_advance_epoch, below.
+	 */
+	size_t backlog[SMOOTHSTEP_NSTEPS];
+
+	/* Peak number of pages in associated extents.  Used for debug only. */
+	uint64_t ceil_npages;
+};
+
+/*
+ * The current decay time setting.  This is the only public access to a decay_t
+ * that's allowed without holding mtx.
+ */
+static inline ssize_t
+decay_ms_read(const decay_t *decay) {
+	return atomic_load_zd(&decay->time_ms, ATOMIC_RELAXED);
+}
+
+/*
+ * See the comment on the struct field -- the limit on pages we should allow in
+ * this decay state this epoch.
+ */
+static inline size_t
+decay_npages_limit_get(const decay_t *decay) {
+	return decay->npages_limit;
+}
+
+/* How many unused dirty pages were generated during the last epoch. */
+static inline size_t
+decay_epoch_npages_delta(const decay_t *decay) {
+	return decay->backlog[SMOOTHSTEP_NSTEPS - 1];
+}
+
+/*
+ * Current epoch duration, in nanoseconds.  Given that new epochs are started
+ * somewhat haphazardly, this is not necessarily exactly the time between any
+ * two calls to decay_maybe_advance_epoch; see the comments on fields in the
+ * decay_t.
+ */
+static inline uint64_t
+decay_epoch_duration_ns(const decay_t *decay) {
+	return nstime_ns(&decay->interval);
+}
+
+static inline bool
+decay_immediately(const decay_t *decay) {
+	ssize_t decay_ms = decay_ms_read(decay);
+	return decay_ms == 0;
+}
+
+static inline bool
+decay_disabled(const decay_t *decay) {
+	ssize_t decay_ms = decay_ms_read(decay);
+	return decay_ms < 0;
+}
+
+/* Returns true if decay is enabled and done gradually. */
+static inline bool
+decay_gradually(const decay_t *decay) {
+	ssize_t decay_ms = decay_ms_read(decay);
+	return decay_ms > 0;
+}
+
+/*
+ * Returns true if the passed in decay time setting is valid.
+ * < -1 : invalid
+ * -1   : never decay
+ *  0   : decay immediately
+ *  > 0 : some positive decay time, up to a maximum allowed value of
+ *  NSTIME_SEC_MAX * 1000, which corresponds to decaying somewhere in the early
+ *  27th century.  By that time, we expect to have implemented alternate purging
+ *  strategies.
+ */
+bool decay_ms_valid(ssize_t decay_ms);
+
+/*
+ * As a precondition, the decay_t must be zeroed out (as if with memset).
+ *
+ * Returns true on error.
+ */
+bool decay_init(decay_t *decay, nstime_t *cur_time, ssize_t decay_ms);
+
+/*
+ * Given an already-initialized decay_t, reinitialize it with the given decay
+ * time.  The decay_t must have previously been initialized (and should not then
+ * be zeroed).
+ */
+void decay_reinit(decay_t *decay, nstime_t *cur_time, ssize_t decay_ms);
+
+/*
+ * Compute how many of 'npages_new' pages we would need to purge in 'time'.
+ */
+uint64_t decay_npages_purge_in(decay_t *decay, nstime_t *time,
+    size_t npages_new);
+
+/* Returns true if the epoch advanced and there are pages to purge. */
+bool decay_maybe_advance_epoch(decay_t *decay, nstime_t *new_time,
+    size_t current_npages);
+
+/*
+ * Calculates wait time until a number of pages in the interval
+ * [0.5 * npages_threshold .. 1.5 * npages_threshold] should be purged.
+ *
+ * Returns number of nanoseconds or DECAY_UNBOUNDED_TIME_TO_PURGE in case of
+ * indefinite wait.
+ */
+uint64_t decay_ns_until_purge(decay_t *decay, size_t npages_current,
+    uint64_t npages_threshold);
+
+#endif /* JEMALLOC_INTERNAL_DECAY_H */

BeefRT/JEMalloc/include/jemalloc/internal/div.h

@@ -0,0 +1,41 @@
+#ifndef JEMALLOC_INTERNAL_DIV_H
+#define JEMALLOC_INTERNAL_DIV_H
+
+#include "jemalloc/internal/assert.h"
+
+/*
+ * This module does the division that computes the index of a region in a slab,
+ * given its offset relative to the base.
+ * That is, given a divisor d, an n = i * d (all integers), we'll return i.
+ * We do some pre-computation to do this more quickly than a CPU division
+ * instruction.
+ * We bound n < 2^32, and don't support dividing by one.
+ */
+
+typedef struct div_info_s div_info_t;
+struct div_info_s {
+	uint32_t magic;
+#ifdef JEMALLOC_DEBUG
+	size_t d;
+#endif
+};
+
+void div_init(div_info_t *div_info, size_t divisor);
+
+static inline size_t
+div_compute(div_info_t *div_info, size_t n) {
+	assert(n <= (uint32_t)-1);
+	/*
+	 * This generates, e.g. mov; imul; shr on x86-64. On a 32-bit machine,
+	 * the compilers I tried were all smart enough to turn this into the
+	 * appropriate "get the high 32 bits of the result of a multiply" (e.g.
+	 * mul; mov edx eax; on x86, umull on arm, etc.).
+	 */
+	size_t i = ((uint64_t)n * (uint64_t)div_info->magic) >> 32;
+#ifdef JEMALLOC_DEBUG
+	assert(i * div_info->d == n);
+#endif
+	return i;
+}
+
+#endif /* JEMALLOC_INTERNAL_DIV_H */

BeefRT/JEMalloc/include/jemalloc/internal/ecache.h

@@ -0,0 +1,55 @@
+#ifndef JEMALLOC_INTERNAL_ECACHE_H
+#define JEMALLOC_INTERNAL_ECACHE_H
+
+#include "jemalloc/internal/eset.h"
+#include "jemalloc/internal/san.h"
+#include "jemalloc/internal/mutex.h"
+
+typedef struct ecache_s ecache_t;
+struct ecache_s {
+	malloc_mutex_t mtx;
+	eset_t eset;
+	eset_t guarded_eset;
+	/* All stored extents must be in the same state. */
+	extent_state_t state;
+	/* The index of the ehooks the ecache is associated with. */
+	unsigned ind;
+	/*
+	 * If true, delay coalescing until eviction; otherwise coalesce during
+	 * deallocation.
+	 */
+	bool delay_coalesce;
+};
+
+static inline size_t
+ecache_npages_get(ecache_t *ecache) {
+	return eset_npages_get(&ecache->eset) +
+	    eset_npages_get(&ecache->guarded_eset);
+}
+
+/* Get the number of extents in the given page size index. */
+static inline size_t
+ecache_nextents_get(ecache_t *ecache, pszind_t ind) {
+	return eset_nextents_get(&ecache->eset, ind) +
+	    eset_nextents_get(&ecache->guarded_eset, ind);
+}
+
+/* Get the sum total bytes of the extents in the given page size index. */
+static inline size_t
+ecache_nbytes_get(ecache_t *ecache, pszind_t ind) {
+	return eset_nbytes_get(&ecache->eset, ind) +
+	    eset_nbytes_get(&ecache->guarded_eset, ind);
+}
+
+static inline unsigned
+ecache_ind_get(ecache_t *ecache) {
+	return ecache->ind;
+}
+
+bool ecache_init(tsdn_t *tsdn, ecache_t *ecache, extent_state_t state,
+    unsigned ind, bool delay_coalesce);
+void ecache_prefork(tsdn_t *tsdn, ecache_t *ecache);
+void ecache_postfork_parent(tsdn_t *tsdn, ecache_t *ecache);
+void ecache_postfork_child(tsdn_t *tsdn, ecache_t *ecache);
+
+#endif /* JEMALLOC_INTERNAL_ECACHE_H */

BeefRT/JEMalloc/include/jemalloc/internal/edata.h

@@ -0,0 +1,698 @@
+#ifndef JEMALLOC_INTERNAL_EDATA_H
+#define JEMALLOC_INTERNAL_EDATA_H
+
+#include "jemalloc/internal/atomic.h"
+#include "jemalloc/internal/bin_info.h"
+#include "jemalloc/internal/bit_util.h"
+#include "jemalloc/internal/hpdata.h"
+#include "jemalloc/internal/nstime.h"
+#include "jemalloc/internal/ph.h"
+#include "jemalloc/internal/ql.h"
+#include "jemalloc/internal/sc.h"
+#include "jemalloc/internal/slab_data.h"
+#include "jemalloc/internal/sz.h"
+#include "jemalloc/internal/typed_list.h"
+
+/*
+ * sizeof(edata_t) is 128 bytes on 64-bit architectures.  Ensure the alignment
+ * to free up the low bits in the rtree leaf.
+ */
+#define EDATA_ALIGNMENT 128
+
+enum extent_state_e {
+	extent_state_active   = 0,
+	extent_state_dirty    = 1,
+	extent_state_muzzy    = 2,
+	extent_state_retained = 3,
+	extent_state_transition = 4, /* States below are intermediate. */
+	extent_state_merging = 5,
+	extent_state_max = 5 /* Sanity checking only. */
+};
+typedef enum extent_state_e extent_state_t;
+
+enum extent_head_state_e {
+	EXTENT_NOT_HEAD,
+	EXTENT_IS_HEAD   /* See comments in ehooks_default_merge_impl(). */
+};
+typedef enum extent_head_state_e extent_head_state_t;
+
+/*
+ * Which implementation of the page allocator interface, (PAI, defined in
+ * pai.h) owns the given extent?
+ */
+enum extent_pai_e {
+	EXTENT_PAI_PAC = 0,
+	EXTENT_PAI_HPA = 1
+};
+typedef enum extent_pai_e extent_pai_t;
+
+struct e_prof_info_s {
+	/* Time when this was allocated. */
+	nstime_t	e_prof_alloc_time;
+	/* Allocation request size. */
+	size_t		e_prof_alloc_size;
+	/* Points to a prof_tctx_t. */
+	atomic_p_t	e_prof_tctx;
+	/*
+	 * Points to a prof_recent_t for the allocation; NULL
+	 * means the recent allocation record no longer exists.
+	 * Protected by prof_recent_alloc_mtx.
+	 */
+	atomic_p_t	e_prof_recent_alloc;
+};
+typedef struct e_prof_info_s e_prof_info_t;
+
+/*
+ * The information about a particular edata that lives in an emap.  Space is
+ * more precious there (the information, plus the edata pointer, has to live in
+ * a 64-bit word if we want to enable a packed representation.
+ *
+ * There are two things that are special about the information here:
+ * - It's quicker to access.  You have one fewer pointer hop, since finding the
+ *   edata_t associated with an item always requires accessing the rtree leaf in
+ *   which this data is stored.
+ * - It can be read unsynchronized, and without worrying about lifetime issues.
+ */
+typedef struct edata_map_info_s edata_map_info_t;
+struct edata_map_info_s {
+	bool slab;
+	szind_t szind;
+};
+
+typedef struct edata_cmp_summary_s edata_cmp_summary_t;
+struct edata_cmp_summary_s {
+	uint64_t sn;
+	uintptr_t addr;
+};
+
+/* Extent (span of pages).  Use accessor functions for e_* fields. */
+typedef struct edata_s edata_t;
+ph_structs(edata_avail, edata_t);
+ph_structs(edata_heap, edata_t);
+struct edata_s {
+	/*
+	 * Bitfield containing several fields:
+	 *
+	 * a: arena_ind
+	 * b: slab
+	 * c: committed
+	 * p: pai
+	 * z: zeroed
+	 * g: guarded
+	 * t: state
+	 * i: szind
+	 * f: nfree
+	 * s: bin_shard
+	 *
+	 * 00000000 ... 0000ssss ssffffff ffffiiii iiiitttg zpcbaaaa aaaaaaaa
+	 *
+	 * arena_ind: Arena from which this extent came, or all 1 bits if
+	 *            unassociated.
+	 *
+	 * slab: The slab flag indicates whether the extent is used for a slab
+	 *       of small regions.  This helps differentiate small size classes,
+	 *       and it indicates whether interior pointers can be looked up via
+	 *       iealloc().
+	 *
+	 * committed: The committed flag indicates whether physical memory is
+	 *            committed to the extent, whether explicitly or implicitly
+	 *            as on a system that overcommits and satisfies physical
+	 *            memory needs on demand via soft page faults.
+	 *
+	 * pai: The pai flag is an extent_pai_t.
+	 *
+	 * zeroed: The zeroed flag is used by extent recycling code to track
+	 *         whether memory is zero-filled.
+	 *
+	 * guarded: The guarded flag is use by the sanitizer to track whether
+	 *          the extent has page guards around it.
+	 *
+	 * state: The state flag is an extent_state_t.
+	 *
+	 * szind: The szind flag indicates usable size class index for
+	 *        allocations residing in this extent, regardless of whether the
+	 *        extent is a slab.  Extent size and usable size often differ
+	 *        even for non-slabs, either due to sz_large_pad or promotion of
+	 *        sampled small regions.
+	 *
+	 * nfree: Number of free regions in slab.
+	 *
+	 * bin_shard: the shard of the bin from which this extent came.
+	 */
+	uint64_t		e_bits;
+#define MASK(CURRENT_FIELD_WIDTH, CURRENT_FIELD_SHIFT) ((((((uint64_t)0x1U) << (CURRENT_FIELD_WIDTH)) - 1)) << (CURRENT_FIELD_SHIFT))
+
+#define EDATA_BITS_ARENA_WIDTH  MALLOCX_ARENA_BITS
+#define EDATA_BITS_ARENA_SHIFT  0
+#define EDATA_BITS_ARENA_MASK  MASK(EDATA_BITS_ARENA_WIDTH, EDATA_BITS_ARENA_SHIFT)
+
+#define EDATA_BITS_SLAB_WIDTH  1
+#define EDATA_BITS_SLAB_SHIFT  (EDATA_BITS_ARENA_WIDTH + EDATA_BITS_ARENA_SHIFT)
+#define EDATA_BITS_SLAB_MASK  MASK(EDATA_BITS_SLAB_WIDTH, EDATA_BITS_SLAB_SHIFT)
+
+#define EDATA_BITS_COMMITTED_WIDTH  1
+#define EDATA_BITS_COMMITTED_SHIFT  (EDATA_BITS_SLAB_WIDTH + EDATA_BITS_SLAB_SHIFT)
+#define EDATA_BITS_COMMITTED_MASK  MASK(EDATA_BITS_COMMITTED_WIDTH, EDATA_BITS_COMMITTED_SHIFT)
+
+#define EDATA_BITS_PAI_WIDTH  1
+#define EDATA_BITS_PAI_SHIFT  (EDATA_BITS_COMMITTED_WIDTH + EDATA_BITS_COMMITTED_SHIFT)
+#define EDATA_BITS_PAI_MASK  MASK(EDATA_BITS_PAI_WIDTH, EDATA_BITS_PAI_SHIFT)
+
+#define EDATA_BITS_ZEROED_WIDTH  1
+#define EDATA_BITS_ZEROED_SHIFT  (EDATA_BITS_PAI_WIDTH + EDATA_BITS_PAI_SHIFT)
+#define EDATA_BITS_ZEROED_MASK  MASK(EDATA_BITS_ZEROED_WIDTH, EDATA_BITS_ZEROED_SHIFT)
+
+#define EDATA_BITS_GUARDED_WIDTH  1
+#define EDATA_BITS_GUARDED_SHIFT  (EDATA_BITS_ZEROED_WIDTH + EDATA_BITS_ZEROED_SHIFT)
+#define EDATA_BITS_GUARDED_MASK  MASK(EDATA_BITS_GUARDED_WIDTH, EDATA_BITS_GUARDED_SHIFT)
+
+#define EDATA_BITS_STATE_WIDTH  3
+#define EDATA_BITS_STATE_SHIFT  (EDATA_BITS_GUARDED_WIDTH + EDATA_BITS_GUARDED_SHIFT)
+#define EDATA_BITS_STATE_MASK  MASK(EDATA_BITS_STATE_WIDTH, EDATA_BITS_STATE_SHIFT)
+
+#define EDATA_BITS_SZIND_WIDTH  LG_CEIL(SC_NSIZES)
+#define EDATA_BITS_SZIND_SHIFT  (EDATA_BITS_STATE_WIDTH + EDATA_BITS_STATE_SHIFT)
+#define EDATA_BITS_SZIND_MASK  MASK(EDATA_BITS_SZIND_WIDTH, EDATA_BITS_SZIND_SHIFT)
+
+#define EDATA_BITS_NFREE_WIDTH  (SC_LG_SLAB_MAXREGS + 1)
+#define EDATA_BITS_NFREE_SHIFT  (EDATA_BITS_SZIND_WIDTH + EDATA_BITS_SZIND_SHIFT)
+#define EDATA_BITS_NFREE_MASK  MASK(EDATA_BITS_NFREE_WIDTH, EDATA_BITS_NFREE_SHIFT)
+
+#define EDATA_BITS_BINSHARD_WIDTH  6
+#define EDATA_BITS_BINSHARD_SHIFT  (EDATA_BITS_NFREE_WIDTH + EDATA_BITS_NFREE_SHIFT)
+#define EDATA_BITS_BINSHARD_MASK  MASK(EDATA_BITS_BINSHARD_WIDTH, EDATA_BITS_BINSHARD_SHIFT)
+
+#define EDATA_BITS_IS_HEAD_WIDTH 1
+#define EDATA_BITS_IS_HEAD_SHIFT  (EDATA_BITS_BINSHARD_WIDTH + EDATA_BITS_BINSHARD_SHIFT)
+#define EDATA_BITS_IS_HEAD_MASK  MASK(EDATA_BITS_IS_HEAD_WIDTH, EDATA_BITS_IS_HEAD_SHIFT)
+
+	/* Pointer to the extent that this structure is responsible for. */
+	void			*e_addr;
+
+	union {
+		/*
+		 * Extent size and serial number associated with the extent
+		 * structure (different than the serial number for the extent at
+		 * e_addr).
+		 *
+		 * ssssssss [...] ssssssss ssssnnnn nnnnnnnn
+		 */
+		size_t			e_size_esn;
+	#define EDATA_SIZE_MASK	((size_t)~(PAGE-1))
+	#define EDATA_ESN_MASK		((size_t)PAGE-1)
+		/* Base extent size, which may not be a multiple of PAGE. */
+		size_t			e_bsize;
+	};
+
+	/*
+	 * If this edata is a user allocation from an HPA, it comes out of some
+	 * pageslab (we don't yet support huegpage allocations that don't fit
+	 * into pageslabs).  This tracks it.
+	 */
+	hpdata_t *e_ps;
+
+	/*
+	 * Serial number.  These are not necessarily unique; splitting an extent
+	 * results in two extents with the same serial number.
+	 */
+	uint64_t e_sn;
+
+	union {
+		/*
+		 * List linkage used when the edata_t is active; either in
+		 * arena's large allocations or bin_t's slabs_full.
+		 */
+		ql_elm(edata_t)	ql_link_active;
+		/*
+		 * Pairing heap linkage.  Used whenever the extent is inactive
+		 * (in the page allocators), or when it is active and in
+		 * slabs_nonfull, or when the edata_t is unassociated with an
+		 * extent and sitting in an edata_cache.
+		 */
+		union {
+			edata_heap_link_t heap_link;
+			edata_avail_link_t avail_link;
+		};
+	};
+
+	union {
+		/*
+		 * List linkage used when the extent is inactive:
+		 * - Stashed dirty extents
+		 * - Ecache LRU functionality.
+		 */
+		ql_elm(edata_t) ql_link_inactive;
+		/* Small region slab metadata. */
+		slab_data_t	e_slab_data;
+
+		/* Profiling data, used for large objects. */
+		e_prof_info_t	e_prof_info;
+	};
+};
+
+TYPED_LIST(edata_list_active, edata_t, ql_link_active)
+TYPED_LIST(edata_list_inactive, edata_t, ql_link_inactive)
+
+static inline unsigned
+edata_arena_ind_get(const edata_t *edata) {
+	unsigned arena_ind = (unsigned)((edata->e_bits &
+	    EDATA_BITS_ARENA_MASK) >> EDATA_BITS_ARENA_SHIFT);
+	assert(arena_ind < MALLOCX_ARENA_LIMIT);
+
+	return arena_ind;
+}
+
+static inline szind_t
+edata_szind_get_maybe_invalid(const edata_t *edata) {
+	szind_t szind = (szind_t)((edata->e_bits & EDATA_BITS_SZIND_MASK) >>
+	    EDATA_BITS_SZIND_SHIFT);
+	assert(szind <= SC_NSIZES);
+	return szind;
+}
+
+static inline szind_t
+edata_szind_get(const edata_t *edata) {
+	szind_t szind = edata_szind_get_maybe_invalid(edata);
+	assert(szind < SC_NSIZES); /* Never call when "invalid". */
+	return szind;
+}
+
+static inline size_t
+edata_usize_get(const edata_t *edata) {
+	return sz_index2size(edata_szind_get(edata));
+}
+
+static inline unsigned
+edata_binshard_get(const edata_t *edata) {
+	unsigned binshard = (unsigned)((edata->e_bits &
+	    EDATA_BITS_BINSHARD_MASK) >> EDATA_BITS_BINSHARD_SHIFT);
+	assert(binshard < bin_infos[edata_szind_get(edata)].n_shards);
+	return binshard;
+}
+
+static inline uint64_t
+edata_sn_get(const edata_t *edata) {
+	return edata->e_sn;
+}
+
+static inline extent_state_t
+edata_state_get(const edata_t *edata) {
+	return (extent_state_t)((edata->e_bits & EDATA_BITS_STATE_MASK) >>
+	    EDATA_BITS_STATE_SHIFT);
+}
+
+static inline bool
+edata_guarded_get(const edata_t *edata) {
+	return (bool)((edata->e_bits & EDATA_BITS_GUARDED_MASK) >>
+	    EDATA_BITS_GUARDED_SHIFT);
+}
+
+static inline bool
+edata_zeroed_get(const edata_t *edata) {
+	return (bool)((edata->e_bits & EDATA_BITS_ZEROED_MASK) >>
+	    EDATA_BITS_ZEROED_SHIFT);
+}
+
+static inline bool
+edata_committed_get(const edata_t *edata) {
+	return (bool)((edata->e_bits & EDATA_BITS_COMMITTED_MASK) >>
+	    EDATA_BITS_COMMITTED_SHIFT);
+}
+
+static inline extent_pai_t
+edata_pai_get(const edata_t *edata) {
+	return (extent_pai_t)((edata->e_bits & EDATA_BITS_PAI_MASK) >>
+	    EDATA_BITS_PAI_SHIFT);
+}
+
+static inline bool
+edata_slab_get(const edata_t *edata) {
+	return (bool)((edata->e_bits & EDATA_BITS_SLAB_MASK) >>
+	    EDATA_BITS_SLAB_SHIFT);
+}
+
+static inline unsigned
+edata_nfree_get(const edata_t *edata) {
+	assert(edata_slab_get(edata));
+	return (unsigned)((edata->e_bits & EDATA_BITS_NFREE_MASK) >>
+	    EDATA_BITS_NFREE_SHIFT);
+}
+
+static inline void *
+edata_base_get(const edata_t *edata) {
+	assert(edata->e_addr == PAGE_ADDR2BASE(edata->e_addr) ||
+	    !edata_slab_get(edata));
+	return PAGE_ADDR2BASE(edata->e_addr);
+}
+
+static inline void *
+edata_addr_get(const edata_t *edata) {
+	assert(edata->e_addr == PAGE_ADDR2BASE(edata->e_addr) ||
+	    !edata_slab_get(edata));
+	return edata->e_addr;
+}
+
+static inline size_t
+edata_size_get(const edata_t *edata) {
+	return (edata->e_size_esn & EDATA_SIZE_MASK);
+}
+
+static inline size_t
+edata_esn_get(const edata_t *edata) {
+	return (edata->e_size_esn & EDATA_ESN_MASK);
+}
+
+static inline size_t
+edata_bsize_get(const edata_t *edata) {
+	return edata->e_bsize;
+}
+
+static inline hpdata_t *
+edata_ps_get(const edata_t *edata) {
+	assert(edata_pai_get(edata) == EXTENT_PAI_HPA);
+	return edata->e_ps;
+}
+
+static inline void *
+edata_before_get(const edata_t *edata) {
+	return (void *)((uintptr_t)edata_base_get(edata) - PAGE);
+}
+
+static inline void *
+edata_last_get(const edata_t *edata) {
+	return (void *)((uintptr_t)edata_base_get(edata) +
+	    edata_size_get(edata) - PAGE);
+}
+
+static inline void *
+edata_past_get(const edata_t *edata) {
+	return (void *)((uintptr_t)edata_base_get(edata) +
+	    edata_size_get(edata));
+}
+
+static inline slab_data_t *
+edata_slab_data_get(edata_t *edata) {
+	assert(edata_slab_get(edata));
+	return &edata->e_slab_data;
+}
+
+static inline const slab_data_t *
+edata_slab_data_get_const(const edata_t *edata) {
+	assert(edata_slab_get(edata));
+	return &edata->e_slab_data;
+}
+
+static inline prof_tctx_t *
+edata_prof_tctx_get(const edata_t *edata) {
+	return (prof_tctx_t *)atomic_load_p(&edata->e_prof_info.e_prof_tctx,
+	    ATOMIC_ACQUIRE);
+}
+
+static inline const nstime_t *
+edata_prof_alloc_time_get(const edata_t *edata) {
+	return &edata->e_prof_info.e_prof_alloc_time;
+}
+
+static inline size_t
+edata_prof_alloc_size_get(const edata_t *edata) {
+	return edata->e_prof_info.e_prof_alloc_size;
+}
+
+static inline prof_recent_t *
+edata_prof_recent_alloc_get_dont_call_directly(const edata_t *edata) {
+	return (prof_recent_t *)atomic_load_p(
+	    &edata->e_prof_info.e_prof_recent_alloc, ATOMIC_RELAXED);
+}
+
+static inline void
+edata_arena_ind_set(edata_t *edata, unsigned arena_ind) {
+	edata->e_bits = (edata->e_bits & ~EDATA_BITS_ARENA_MASK) |
+	    ((uint64_t)arena_ind << EDATA_BITS_ARENA_SHIFT);
+}
+
+static inline void
+edata_binshard_set(edata_t *edata, unsigned binshard) {
+	/* The assertion assumes szind is set already. */
+	assert(binshard < bin_infos[edata_szind_get(edata)].n_shards);
+	edata->e_bits = (edata->e_bits & ~EDATA_BITS_BINSHARD_MASK) |
+	    ((uint64_t)binshard << EDATA_BITS_BINSHARD_SHIFT);
+}
+
+static inline void
+edata_addr_set(edata_t *edata, void *addr) {
+	edata->e_addr = addr;
+}
+
+static inline void
+edata_size_set(edata_t *edata, size_t size) {
+	assert((size & ~EDATA_SIZE_MASK) == 0);
+	edata->e_size_esn = size | (edata->e_size_esn & ~EDATA_SIZE_MASK);
+}
+
+static inline void
+edata_esn_set(edata_t *edata, size_t esn) {
+	edata->e_size_esn = (edata->e_size_esn & ~EDATA_ESN_MASK) | (esn &
+	    EDATA_ESN_MASK);
+}
+
+static inline void
+edata_bsize_set(edata_t *edata, size_t bsize) {
+	edata->e_bsize = bsize;
+}
+
+static inline void
+edata_ps_set(edata_t *edata, hpdata_t *ps) {
+	assert(edata_pai_get(edata) == EXTENT_PAI_HPA);
+	edata->e_ps = ps;
+}
+
+static inline void
+edata_szind_set(edata_t *edata, szind_t szind) {
+	assert(szind <= SC_NSIZES); /* SC_NSIZES means "invalid". */
+	edata->e_bits = (edata->e_bits & ~EDATA_BITS_SZIND_MASK) |
+	    ((uint64_t)szind << EDATA_BITS_SZIND_SHIFT);
+}
+
+static inline void
+edata_nfree_set(edata_t *edata, unsigned nfree) {
+	assert(edata_slab_get(edata));
+	edata->e_bits = (edata->e_bits & ~EDATA_BITS_NFREE_MASK) |
+	    ((uint64_t)nfree << EDATA_BITS_NFREE_SHIFT);
+}
+
+static inline void
+edata_nfree_binshard_set(edata_t *edata, unsigned nfree, unsigned binshard) {
+	/* The assertion assumes szind is set already. */
+	assert(binshard < bin_infos[edata_szind_get(edata)].n_shards);
+	edata->e_bits = (edata->e_bits &
+	    (~EDATA_BITS_NFREE_MASK & ~EDATA_BITS_BINSHARD_MASK)) |
+	    ((uint64_t)binshard << EDATA_BITS_BINSHARD_SHIFT) |
+	    ((uint64_t)nfree << EDATA_BITS_NFREE_SHIFT);
+}
+
+static inline void
+edata_nfree_inc(edata_t *edata) {
+	assert(edata_slab_get(edata));
+	edata->e_bits += ((uint64_t)1U << EDATA_BITS_NFREE_SHIFT);
+}
+
+static inline void
+edata_nfree_dec(edata_t *edata) {
+	assert(edata_slab_get(edata));
+	edata->e_bits -= ((uint64_t)1U << EDATA_BITS_NFREE_SHIFT);
+}
+
+static inline void
+edata_nfree_sub(edata_t *edata, uint64_t n) {
+	assert(edata_slab_get(edata));
+	edata->e_bits -= (n << EDATA_BITS_NFREE_SHIFT);
+}
+
+static inline void
+edata_sn_set(edata_t *edata, uint64_t sn) {
+	edata->e_sn = sn;
+}
+
+static inline void
+edata_state_set(edata_t *edata, extent_state_t state) {
+	edata->e_bits = (edata->e_bits & ~EDATA_BITS_STATE_MASK) |
+	    ((uint64_t)state << EDATA_BITS_STATE_SHIFT);
+}
+
+static inline void
+edata_guarded_set(edata_t *edata, bool guarded) {
+	edata->e_bits = (edata->e_bits & ~EDATA_BITS_GUARDED_MASK) |
+	    ((uint64_t)guarded << EDATA_BITS_GUARDED_SHIFT);
+}
+
+static inline void
+edata_zeroed_set(edata_t *edata, bool zeroed) {
+	edata->e_bits = (edata->e_bits & ~EDATA_BITS_ZEROED_MASK) |
+	    ((uint64_t)zeroed << EDATA_BITS_ZEROED_SHIFT);
+}
+
+static inline void
+edata_committed_set(edata_t *edata, bool committed) {
+	edata->e_bits = (edata->e_bits & ~EDATA_BITS_COMMITTED_MASK) |
+	    ((uint64_t)committed << EDATA_BITS_COMMITTED_SHIFT);
+}
+
+static inline void
+edata_pai_set(edata_t *edata, extent_pai_t pai) {
+	edata->e_bits = (edata->e_bits & ~EDATA_BITS_PAI_MASK) |
+	    ((uint64_t)pai << EDATA_BITS_PAI_SHIFT);
+}
+
+static inline void
+edata_slab_set(edata_t *edata, bool slab) {
+	edata->e_bits = (edata->e_bits & ~EDATA_BITS_SLAB_MASK) |
+	    ((uint64_t)slab << EDATA_BITS_SLAB_SHIFT);
+}
+
+static inline void
+edata_prof_tctx_set(edata_t *edata, prof_tctx_t *tctx) {
+	atomic_store_p(&edata->e_prof_info.e_prof_tctx, tctx, ATOMIC_RELEASE);
+}
+
+static inline void
+edata_prof_alloc_time_set(edata_t *edata, nstime_t *t) {
+	nstime_copy(&edata->e_prof_info.e_prof_alloc_time, t);
+}
+
+static inline void
+edata_prof_alloc_size_set(edata_t *edata, size_t size) {
+	edata->e_prof_info.e_prof_alloc_size = size;
+}
+
+static inline void
+edata_prof_recent_alloc_set_dont_call_directly(edata_t *edata,
+    prof_recent_t *recent_alloc) {
+	atomic_store_p(&edata->e_prof_info.e_prof_recent_alloc, recent_alloc,
+	    ATOMIC_RELAXED);
+}
+
+static inline bool
+edata_is_head_get(edata_t *edata) {
+	return (bool)((edata->e_bits & EDATA_BITS_IS_HEAD_MASK) >>
+	    EDATA_BITS_IS_HEAD_SHIFT);
+}
+
+static inline void
+edata_is_head_set(edata_t *edata, bool is_head) {
+	edata->e_bits = (edata->e_bits & ~EDATA_BITS_IS_HEAD_MASK) |
+	    ((uint64_t)is_head << EDATA_BITS_IS_HEAD_SHIFT);
+}
+
+static inline bool
+edata_state_in_transition(extent_state_t state) {
+	return state >= extent_state_transition;
+}
+
+/*
+ * Because this function is implemented as a sequence of bitfield modifications,
+ * even though each individual bit is properly initialized, we technically read
+ * uninitialized data within it.  This is mostly fine, since most callers get
+ * their edatas from zeroing sources, but callers who make stack edata_ts need
+ * to manually zero them.
+ */
+static inline void
+edata_init(edata_t *edata, unsigned arena_ind, void *addr, size_t size,
+    bool slab, szind_t szind, uint64_t sn, extent_state_t state, bool zeroed,
+    bool committed, extent_pai_t pai, extent_head_state_t is_head) {
+	assert(addr == PAGE_ADDR2BASE(addr) || !slab);
+
+	edata_arena_ind_set(edata, arena_ind);
+	edata_addr_set(edata, addr);
+	edata_size_set(edata, size);
+	edata_slab_set(edata, slab);
+	edata_szind_set(edata, szind);
+	edata_sn_set(edata, sn);
+	edata_state_set(edata, state);
+	edata_guarded_set(edata, false);
+	edata_zeroed_set(edata, zeroed);
+	edata_committed_set(edata, committed);
+	edata_pai_set(edata, pai);
+	edata_is_head_set(edata, is_head == EXTENT_IS_HEAD);
+	if (config_prof) {
+		edata_prof_tctx_set(edata, NULL);
+	}
+}
+
+static inline void
+edata_binit(edata_t *edata, void *addr, size_t bsize, uint64_t sn) {
+	edata_arena_ind_set(edata, (1U << MALLOCX_ARENA_BITS) - 1);
+	edata_addr_set(edata, addr);
+	edata_bsize_set(edata, bsize);
+	edata_slab_set(edata, false);
+	edata_szind_set(edata, SC_NSIZES);
+	edata_sn_set(edata, sn);
+	edata_state_set(edata, extent_state_active);
+	edata_guarded_set(edata, false);
+	edata_zeroed_set(edata, true);
+	edata_committed_set(edata, true);
+	/*
+	 * This isn't strictly true, but base allocated extents never get
+	 * deallocated and can't be looked up in the emap, but no sense in
+	 * wasting a state bit to encode this fact.
+	 */
+	edata_pai_set(edata, EXTENT_PAI_PAC);
+}
+
+static inline int
+edata_esn_comp(const edata_t *a, const edata_t *b) {
+	size_t a_esn = edata_esn_get(a);
+	size_t b_esn = edata_esn_get(b);
+
+	return (a_esn > b_esn) - (a_esn < b_esn);
+}
+
+static inline int
+edata_ead_comp(const edata_t *a, const edata_t *b) {
+	uintptr_t a_eaddr = (uintptr_t)a;
+	uintptr_t b_eaddr = (uintptr_t)b;
+
+	return (a_eaddr > b_eaddr) - (a_eaddr < b_eaddr);
+}
+
+static inline edata_cmp_summary_t
+edata_cmp_summary_get(const edata_t *edata) {
+	return (edata_cmp_summary_t){edata_sn_get(edata),
+		(uintptr_t)edata_addr_get(edata)};
+}
+
+static inline int
+edata_cmp_summary_comp(edata_cmp_summary_t a, edata_cmp_summary_t b) {
+	int ret;
+	ret = (a.sn > b.sn) - (a.sn < b.sn);
+	if (ret != 0) {
+		return ret;
+	}
+	ret = (a.addr > b.addr) - (a.addr < b.addr);
+	return ret;
+}
+
+static inline int
+edata_snad_comp(const edata_t *a, const edata_t *b) {
+	edata_cmp_summary_t a_cmp = edata_cmp_summary_get(a);
+	edata_cmp_summary_t b_cmp = edata_cmp_summary_get(b);
+
+	return edata_cmp_summary_comp(a_cmp, b_cmp);
+}
+
+static inline int
+edata_esnead_comp(const edata_t *a, const edata_t *b) {
+	int ret;
+
+	ret = edata_esn_comp(a, b);
+	if (ret != 0) {
+		return ret;
+	}
+
+	ret = edata_ead_comp(a, b);
+	return ret;
+}
+
+ph_proto(, edata_avail, edata_t)
+ph_proto(, edata_heap, edata_t)
+
+#endif /* JEMALLOC_INTERNAL_EDATA_H */

BeefRT/JEMalloc/include/jemalloc/internal/edata_cache.h

@@ -0,0 +1,49 @@
+#ifndef JEMALLOC_INTERNAL_EDATA_CACHE_H
+#define JEMALLOC_INTERNAL_EDATA_CACHE_H
+
+#include "jemalloc/internal/base.h"
+
+/* For tests only. */
+#define EDATA_CACHE_FAST_FILL 4
+
+/*
+ * A cache of edata_t structures allocated via base_alloc_edata (as opposed to
+ * the underlying extents they describe).  The contents of returned edata_t
+ * objects are garbage and cannot be relied upon.
+ */
+
+typedef struct edata_cache_s edata_cache_t;
+struct edata_cache_s {
+	edata_avail_t avail;
+	atomic_zu_t count;
+	malloc_mutex_t mtx;
+	base_t *base;
+};
+
+bool edata_cache_init(edata_cache_t *edata_cache, base_t *base);
+edata_t *edata_cache_get(tsdn_t *tsdn, edata_cache_t *edata_cache);
+void edata_cache_put(tsdn_t *tsdn, edata_cache_t *edata_cache, edata_t *edata);
+
+void edata_cache_prefork(tsdn_t *tsdn, edata_cache_t *edata_cache);
+void edata_cache_postfork_parent(tsdn_t *tsdn, edata_cache_t *edata_cache);
+void edata_cache_postfork_child(tsdn_t *tsdn, edata_cache_t *edata_cache);
+
+/*
+ * An edata_cache_small is like an edata_cache, but it relies on external
+ * synchronization and avoids first-fit strategies.
+ */
+
+typedef struct edata_cache_fast_s edata_cache_fast_t;
+struct edata_cache_fast_s {
+	edata_list_inactive_t list;
+	edata_cache_t *fallback;
+	bool disabled;
+};
+
+void edata_cache_fast_init(edata_cache_fast_t *ecs, edata_cache_t *fallback);
+edata_t *edata_cache_fast_get(tsdn_t *tsdn, edata_cache_fast_t *ecs);
+void edata_cache_fast_put(tsdn_t *tsdn, edata_cache_fast_t *ecs,
+    edata_t *edata);
+void edata_cache_fast_disable(tsdn_t *tsdn, edata_cache_fast_t *ecs);
+
+#endif /* JEMALLOC_INTERNAL_EDATA_CACHE_H */

BeefRT/JEMalloc/include/jemalloc/internal/ehooks.h

@@ -0,0 +1,412 @@
+#ifndef JEMALLOC_INTERNAL_EHOOKS_H
+#define JEMALLOC_INTERNAL_EHOOKS_H
+
+#include "jemalloc/internal/atomic.h"
+#include "jemalloc/internal/extent_mmap.h"
+
+/*
+ * This module is the internal interface to the extent hooks (both
+ * user-specified and external).  Eventually, this will give us the flexibility
+ * to use multiple different versions of user-visible extent-hook APIs under a
+ * single user interface.
+ *
+ * Current API expansions (not available to anyone but the default hooks yet):
+ *   - Head state tracking.  Hooks can decide whether or not to merge two
+ *     extents based on whether or not one of them is the head (i.e. was
+ *     allocated on its own).  The later extent loses its "head" status.
+ */
+
+extern const extent_hooks_t ehooks_default_extent_hooks;
+
+typedef struct ehooks_s ehooks_t;
+struct ehooks_s {
+	/*
+	 * The user-visible id that goes with the ehooks (i.e. that of the base
+	 * they're a part of, the associated arena's index within the arenas
+	 * array).
+	 */
+	unsigned ind;
+	/* Logically an extent_hooks_t *. */
+	atomic_p_t ptr;
+};
+
+extern const extent_hooks_t ehooks_default_extent_hooks;
+
+/*
+ * These are not really part of the public API.  Each hook has a fast-path for
+ * the default-hooks case that can avoid various small inefficiencies:
+ *   - Forgetting tsd and then calling tsd_get within the hook.
+ *   - Getting more state than necessary out of the extent_t.
+ *   - Doing arena_ind -> arena -> arena_ind lookups.
+ * By making the calls to these functions visible to the compiler, it can move
+ * those extra bits of computation down below the fast-paths where they get ignored.
+ */
+void *ehooks_default_alloc_impl(tsdn_t *tsdn, void *new_addr, size_t size,
+    size_t alignment, bool *zero, bool *commit, unsigned arena_ind);
+bool ehooks_default_dalloc_impl(void *addr, size_t size);
+void ehooks_default_destroy_impl(void *addr, size_t size);
+bool ehooks_default_commit_impl(void *addr, size_t offset, size_t length);
+bool ehooks_default_decommit_impl(void *addr, size_t offset, size_t length);
+#ifdef PAGES_CAN_PURGE_LAZY
+bool ehooks_default_purge_lazy_impl(void *addr, size_t offset, size_t length);
+#endif
+#ifdef PAGES_CAN_PURGE_FORCED
+bool ehooks_default_purge_forced_impl(void *addr, size_t offset, size_t length);
+#endif
+bool ehooks_default_split_impl();
+/*
+ * Merge is the only default extent hook we declare -- see the comment in
+ * ehooks_merge.
+ */
+bool ehooks_default_merge(extent_hooks_t *extent_hooks, void *addr_a,
+    size_t size_a, void *addr_b, size_t size_b, bool committed,
+    unsigned arena_ind);
+bool ehooks_default_merge_impl(tsdn_t *tsdn, void *addr_a, void *addr_b);
+void ehooks_default_zero_impl(void *addr, size_t size);
+void ehooks_default_guard_impl(void *guard1, void *guard2);
+void ehooks_default_unguard_impl(void *guard1, void *guard2);
+
+/*
+ * We don't officially support reentrancy from wtihin the extent hooks.  But
+ * various people who sit within throwing distance of the jemalloc team want
+ * that functionality in certain limited cases.  The default reentrancy guards
+ * assert that we're not reentrant from a0 (since it's the bootstrap arena,
+ * where reentrant allocations would be redirected), which we would incorrectly
+ * trigger in cases where a0 has extent hooks (those hooks themselves can't be
+ * reentrant, then, but there are reasonable uses for such functionality, like
+ * putting internal metadata on hugepages).  Therefore, we use the raw
+ * reentrancy guards.
+ *
+ * Eventually, we need to think more carefully about whether and where we
+ * support allocating from within extent hooks (and what that means for things
+ * like profiling, stats collection, etc.), and document what the guarantee is.
+ */
+static inline void
+ehooks_pre_reentrancy(tsdn_t *tsdn) {
+	tsd_t *tsd = tsdn_null(tsdn) ? tsd_fetch() : tsdn_tsd(tsdn);
+	tsd_pre_reentrancy_raw(tsd);
+}
+
+static inline void
+ehooks_post_reentrancy(tsdn_t *tsdn) {
+	tsd_t *tsd = tsdn_null(tsdn) ? tsd_fetch() : tsdn_tsd(tsdn);
+	tsd_post_reentrancy_raw(tsd);
+}
+
+/* Beginning of the public API. */
+void ehooks_init(ehooks_t *ehooks, extent_hooks_t *extent_hooks, unsigned ind);
+
+static inline unsigned
+ehooks_ind_get(const ehooks_t *ehooks) {
+	return ehooks->ind;
+}
+
+static inline void
+ehooks_set_extent_hooks_ptr(ehooks_t *ehooks, extent_hooks_t *extent_hooks) {
+	atomic_store_p(&ehooks->ptr, extent_hooks, ATOMIC_RELEASE);
+}
+
+static inline extent_hooks_t *
+ehooks_get_extent_hooks_ptr(ehooks_t *ehooks) {
+	return (extent_hooks_t *)atomic_load_p(&ehooks->ptr, ATOMIC_ACQUIRE);
+}
+
+static inline bool
+ehooks_are_default(ehooks_t *ehooks) {
+	return ehooks_get_extent_hooks_ptr(ehooks) ==
+	    &ehooks_default_extent_hooks;
+}
+
+/*
+ * In some cases, a caller needs to allocate resources before attempting to call
+ * a hook.  If that hook is doomed to fail, this is wasteful.  We therefore
+ * include some checks for such cases.
+ */
+static inline bool
+ehooks_dalloc_will_fail(ehooks_t *ehooks) {
+	if (ehooks_are_default(ehooks)) {
+		return opt_retain;
+	} else {
+		return ehooks_get_extent_hooks_ptr(ehooks)->dalloc == NULL;
+	}
+}
+
+static inline bool
+ehooks_split_will_fail(ehooks_t *ehooks) {
+	return ehooks_get_extent_hooks_ptr(ehooks)->split == NULL;
+}
+
+static inline bool
+ehooks_merge_will_fail(ehooks_t *ehooks) {
+	return ehooks_get_extent_hooks_ptr(ehooks)->merge == NULL;
+}
+
+static inline bool
+ehooks_guard_will_fail(ehooks_t *ehooks) {
+	/*
+	 * Before the guard hooks are officially introduced, limit the use to
+	 * the default hooks only.
+	 */
+	return !ehooks_are_default(ehooks);
+}
+
+/*
+ * Some hooks are required to return zeroed memory in certain situations.  In
+ * debug mode, we do some heuristic checks that they did what they were supposed
+ * to.
+ *
+ * This isn't really ehooks-specific (i.e. anyone can check for zeroed memory).
+ * But incorrect zero information indicates an ehook bug.
+ */
+static inline void
+ehooks_debug_zero_check(void *addr, size_t size) {
+	assert(((uintptr_t)addr & PAGE_MASK) == 0);
+	assert((size & PAGE_MASK) == 0);
+	assert(size > 0);
+	if (config_debug) {
+		/* Check the whole first page. */
+		size_t *p = (size_t *)addr;
+		for (size_t i = 0; i < PAGE / sizeof(size_t); i++) {
+			assert(p[i] == 0);
+		}
+		/*
+		 * And 4 spots within.  There's a tradeoff here; the larger
+		 * this number, the more likely it is that we'll catch a bug
+		 * where ehooks return a sparsely non-zero range.  But
+		 * increasing the number of checks also increases the number of
+		 * page faults in debug mode.  FreeBSD does much of their
+		 * day-to-day development work in debug mode, so we don't want
+		 * even the debug builds to be too slow.
+		 */
+		const size_t nchecks = 4;
+		assert(PAGE >= sizeof(size_t) * nchecks);
+		for (size_t i = 0; i < nchecks; ++i) {
+			assert(p[i * (size / sizeof(size_t) / nchecks)] == 0);
+		}
+	}
+}
+
+
+static inline void *
+ehooks_alloc(tsdn_t *tsdn, ehooks_t *ehooks, void *new_addr, size_t size,
+    size_t alignment, bool *zero, bool *commit) {
+	bool orig_zero = *zero;
+	void *ret;
+	extent_hooks_t *extent_hooks = ehooks_get_extent_hooks_ptr(ehooks);
+	if (extent_hooks == &ehooks_default_extent_hooks) {
+		ret = ehooks_default_alloc_impl(tsdn, new_addr, size,
+		    alignment, zero, commit, ehooks_ind_get(ehooks));
+	} else {
+		ehooks_pre_reentrancy(tsdn);
+		ret = extent_hooks->alloc(extent_hooks, new_addr, size,
+		    alignment, zero, commit, ehooks_ind_get(ehooks));
+		ehooks_post_reentrancy(tsdn);
+	}
+	assert(new_addr == NULL || ret == NULL || new_addr == ret);
+	assert(!orig_zero || *zero);
+	if (*zero && ret != NULL) {
+		ehooks_debug_zero_check(ret, size);
+	}
+	return ret;
+}
+
+static inline bool
+ehooks_dalloc(tsdn_t *tsdn, ehooks_t *ehooks, void *addr, size_t size,
+    bool committed) {
+	extent_hooks_t *extent_hooks = ehooks_get_extent_hooks_ptr(ehooks);
+	if (extent_hooks == &ehooks_default_extent_hooks) {
+		return ehooks_default_dalloc_impl(addr, size);
+	} else if (extent_hooks->dalloc == NULL) {
+		return true;
+	} else {
+		ehooks_pre_reentrancy(tsdn);
+		bool err = extent_hooks->dalloc(extent_hooks, addr, size,
+		    committed, ehooks_ind_get(ehooks));
+		ehooks_post_reentrancy(tsdn);
+		return err;
+	}
+}
+
+static inline void
+ehooks_destroy(tsdn_t *tsdn, ehooks_t *ehooks, void *addr, size_t size,
+    bool committed) {
+	extent_hooks_t *extent_hooks = ehooks_get_extent_hooks_ptr(ehooks);
+	if (extent_hooks == &ehooks_default_extent_hooks) {
+		ehooks_default_destroy_impl(addr, size);
+	} else if (extent_hooks->destroy == NULL) {
+		/* Do nothing. */
+	} else {
+		ehooks_pre_reentrancy(tsdn);
+		extent_hooks->destroy(extent_hooks, addr, size, committed,
+		    ehooks_ind_get(ehooks));
+		ehooks_post_reentrancy(tsdn);
+	}
+}
+
+static inline bool
+ehooks_commit(tsdn_t *tsdn, ehooks_t *ehooks, void *addr, size_t size,
+    size_t offset, size_t length) {
+	extent_hooks_t *extent_hooks = ehooks_get_extent_hooks_ptr(ehooks);
+	bool err;
+	if (extent_hooks == &ehooks_default_extent_hooks) {
+		err = ehooks_default_commit_impl(addr, offset, length);
+	} else if (extent_hooks->commit == NULL) {
+		err = true;
+	} else {
+		ehooks_pre_reentrancy(tsdn);
+		err = extent_hooks->commit(extent_hooks, addr, size,
+		    offset, length, ehooks_ind_get(ehooks));
+		ehooks_post_reentrancy(tsdn);
+	}
+	if (!err) {
+		ehooks_debug_zero_check(addr, size);
+	}
+	return err;
+}
+
+static inline bool
+ehooks_decommit(tsdn_t *tsdn, ehooks_t *ehooks, void *addr, size_t size,
+    size_t offset, size_t length) {
+	extent_hooks_t *extent_hooks = ehooks_get_extent_hooks_ptr(ehooks);
+	if (extent_hooks == &ehooks_default_extent_hooks) {
+		return ehooks_default_decommit_impl(addr, offset, length);
+	} else if (extent_hooks->decommit == NULL) {
+		return true;
+	} else {
+		ehooks_pre_reentrancy(tsdn);
+		bool err = extent_hooks->decommit(extent_hooks, addr, size,
+		    offset, length, ehooks_ind_get(ehooks));
+		ehooks_post_reentrancy(tsdn);
+		return err;
+	}
+}
+
+static inline bool
+ehooks_purge_lazy(tsdn_t *tsdn, ehooks_t *ehooks, void *addr, size_t size,
+    size_t offset, size_t length) {
+	extent_hooks_t *extent_hooks = ehooks_get_extent_hooks_ptr(ehooks);
+#ifdef PAGES_CAN_PURGE_LAZY
+	if (extent_hooks == &ehooks_default_extent_hooks) {
+		return ehooks_default_purge_lazy_impl(addr, offset, length);
+	}
+#endif
+	if (extent_hooks->purge_lazy == NULL) {
+		return true;
+	} else {
+		ehooks_pre_reentrancy(tsdn);
+		bool err = extent_hooks->purge_lazy(extent_hooks, addr, size,
+		    offset, length, ehooks_ind_get(ehooks));
+		ehooks_post_reentrancy(tsdn);
+		return err;
+	}
+}
+
+static inline bool
+ehooks_purge_forced(tsdn_t *tsdn, ehooks_t *ehooks, void *addr, size_t size,
+    size_t offset, size_t length) {
+	extent_hooks_t *extent_hooks = ehooks_get_extent_hooks_ptr(ehooks);
+	/*
+	 * It would be correct to have a ehooks_debug_zero_check call at the end
+	 * of this function; purge_forced is required to zero.  But checking
+	 * would touch the page in question, which may have performance
+	 * consequences (imagine the hooks are using hugepages, with a global
+	 * zero page off).  Even in debug mode, it's usually a good idea to
+	 * avoid cases that can dramatically increase memory consumption.
+	 */
+#ifdef PAGES_CAN_PURGE_FORCED
+	if (extent_hooks == &ehooks_default_extent_hooks) {
+		return ehooks_default_purge_forced_impl(addr, offset, length);
+	}
+#endif
+	if (extent_hooks->purge_forced == NULL) {
+		return true;
+	} else {
+		ehooks_pre_reentrancy(tsdn);
+		bool err = extent_hooks->purge_forced(extent_hooks, addr, size,
+		    offset, length, ehooks_ind_get(ehooks));
+		ehooks_post_reentrancy(tsdn);
+		return err;
+	}
+}
+
+static inline bool
+ehooks_split(tsdn_t *tsdn, ehooks_t *ehooks, void *addr, size_t size,
+    size_t size_a, size_t size_b, bool committed) {
+	extent_hooks_t *extent_hooks = ehooks_get_extent_hooks_ptr(ehooks);
+	if (ehooks_are_default(ehooks)) {
+		return ehooks_default_split_impl();
+	} else if (extent_hooks->split == NULL) {
+		return true;
+	} else {
+		ehooks_pre_reentrancy(tsdn);
+		bool err = extent_hooks->split(extent_hooks, addr, size, size_a,
+		    size_b, committed, ehooks_ind_get(ehooks));
+		ehooks_post_reentrancy(tsdn);
+		return err;
+	}
+}
+
+static inline bool
+ehooks_merge(tsdn_t *tsdn, ehooks_t *ehooks, void *addr_a, size_t size_a,
+    void *addr_b, size_t size_b, bool committed) {
+	extent_hooks_t *extent_hooks = ehooks_get_extent_hooks_ptr(ehooks);
+	if (extent_hooks == &ehooks_default_extent_hooks) {
+		return ehooks_default_merge_impl(tsdn, addr_a, addr_b);
+	} else if (extent_hooks->merge == NULL) {
+		return true;
+	} else {
+		ehooks_pre_reentrancy(tsdn);
+		bool err = extent_hooks->merge(extent_hooks, addr_a, size_a,
+		    addr_b, size_b, committed, ehooks_ind_get(ehooks));
+		ehooks_post_reentrancy(tsdn);
+		return err;
+	}
+}
+
+static inline void
+ehooks_zero(tsdn_t *tsdn, ehooks_t *ehooks, void *addr, size_t size) {
+	extent_hooks_t *extent_hooks = ehooks_get_extent_hooks_ptr(ehooks);
+	if (extent_hooks == &ehooks_default_extent_hooks) {
+		ehooks_default_zero_impl(addr, size);
+	} else {
+		/*
+		 * It would be correct to try using the user-provided purge
+		 * hooks (since they are required to have zeroed the extent if
+		 * they indicate success), but we don't necessarily know their
+		 * cost.  We'll be conservative and use memset.
+		 */
+		memset(addr, 0, size);
+	}
+}
+
+static inline bool
+ehooks_guard(tsdn_t *tsdn, ehooks_t *ehooks, void *guard1, void *guard2) {
+	bool err;
+	extent_hooks_t *extent_hooks = ehooks_get_extent_hooks_ptr(ehooks);
+
+	if (extent_hooks == &ehooks_default_extent_hooks) {
+		ehooks_default_guard_impl(guard1, guard2);
+		err = false;
+	} else {
+		err = true;
+	}
+
+	return err;
+}
+
+static inline bool
+ehooks_unguard(tsdn_t *tsdn, ehooks_t *ehooks, void *guard1, void *guard2) {
+	bool err;
+	extent_hooks_t *extent_hooks = ehooks_get_extent_hooks_ptr(ehooks);
+
+	if (extent_hooks == &ehooks_default_extent_hooks) {
+		ehooks_default_unguard_impl(guard1, guard2);
+		err = false;
+	} else {
+		err = true;
+	}
+
+	return err;
+}
+
+#endif /* JEMALLOC_INTERNAL_EHOOKS_H */

BeefRT/JEMalloc/include/jemalloc/internal/emap.h

@@ -0,0 +1,357 @@
+#ifndef JEMALLOC_INTERNAL_EMAP_H
+#define JEMALLOC_INTERNAL_EMAP_H
+
+#include "jemalloc/internal/base.h"
+#include "jemalloc/internal/rtree.h"
+
+/*
+ * Note: Ends without at semicolon, so that
+ *     EMAP_DECLARE_RTREE_CTX;
+ * in uses will avoid empty-statement warnings.
+ */
+#define EMAP_DECLARE_RTREE_CTX						\
+    rtree_ctx_t rtree_ctx_fallback;					\
+    rtree_ctx_t *rtree_ctx = tsdn_rtree_ctx(tsdn, &rtree_ctx_fallback)
+
+typedef struct emap_s emap_t;
+struct emap_s {
+	rtree_t rtree;
+};
+
+/* Used to pass rtree lookup context down the path. */
+typedef struct emap_alloc_ctx_t emap_alloc_ctx_t;
+struct emap_alloc_ctx_t {
+	szind_t szind;
+	bool slab;
+};
+
+typedef struct emap_full_alloc_ctx_s emap_full_alloc_ctx_t;
+struct emap_full_alloc_ctx_s {
+	szind_t szind;
+	bool slab;
+	edata_t *edata;
+};
+
+bool emap_init(emap_t *emap, base_t *base, bool zeroed);
+
+void emap_remap(tsdn_t *tsdn, emap_t *emap, edata_t *edata, szind_t szind,
+    bool slab);
+
+void emap_update_edata_state(tsdn_t *tsdn, emap_t *emap, edata_t *edata,
+    extent_state_t state);
+
+/*
+ * The two acquire functions below allow accessing neighbor edatas, if it's safe
+ * and valid to do so (i.e. from the same arena, of the same state, etc.).  This
+ * is necessary because the ecache locks are state based, and only protect
+ * edatas with the same state.  Therefore the neighbor edata's state needs to be
+ * verified first, before chasing the edata pointer.  The returned edata will be
+ * in an acquired state, meaning other threads will be prevented from accessing
+ * it, even if technically the edata can still be discovered from the rtree.
+ *
+ * This means, at any moment when holding pointers to edata, either one of the
+ * state based locks is held (and the edatas are all of the protected state), or
+ * the edatas are in an acquired state (e.g. in active or merging state).  The
+ * acquire operation itself (changing the edata to an acquired state) is done
+ * under the state locks.
+ */
+edata_t *emap_try_acquire_edata_neighbor(tsdn_t *tsdn, emap_t *emap,
+    edata_t *edata, extent_pai_t pai, extent_state_t expected_state,
+    bool forward);
+edata_t *emap_try_acquire_edata_neighbor_expand(tsdn_t *tsdn, emap_t *emap,
+    edata_t *edata, extent_pai_t pai, extent_state_t expected_state);
+void emap_release_edata(tsdn_t *tsdn, emap_t *emap, edata_t *edata,
+    extent_state_t new_state);
+
+/*
+ * Associate the given edata with its beginning and end address, setting the
+ * szind and slab info appropriately.
+ * Returns true on error (i.e. resource exhaustion).
+ */
+bool emap_register_boundary(tsdn_t *tsdn, emap_t *emap, edata_t *edata,
+    szind_t szind, bool slab);
+
+/*
+ * Does the same thing, but with the interior of the range, for slab
+ * allocations.
+ *
+ * You might wonder why we don't just have a single emap_register function that
+ * does both depending on the value of 'slab'.  The answer is twofold:
+ * - As a practical matter, in places like the extract->split->commit pathway,
+ *   we defer the interior operation until we're sure that the commit won't fail
+ *   (but we have to register the split boundaries there).
+ * - In general, we're trying to move to a world where the page-specific
+ *   allocator doesn't know as much about how the pages it allocates will be
+ *   used, and passing a 'slab' parameter everywhere makes that more
+ *   complicated.
+ *
+ * Unlike the boundary version, this function can't fail; this is because slabs
+ * can't get big enough to touch a new page that neither of the boundaries
+ * touched, so no allocation is necessary to fill the interior once the boundary
+ * has been touched.
+ */
+void emap_register_interior(tsdn_t *tsdn, emap_t *emap, edata_t *edata,
+    szind_t szind);
+
+void emap_deregister_boundary(tsdn_t *tsdn, emap_t *emap, edata_t *edata);
+void emap_deregister_interior(tsdn_t *tsdn, emap_t *emap, edata_t *edata);
+
+typedef struct emap_prepare_s emap_prepare_t;
+struct emap_prepare_s {
+	rtree_leaf_elm_t *lead_elm_a;
+	rtree_leaf_elm_t *lead_elm_b;
+	rtree_leaf_elm_t *trail_elm_a;
+	rtree_leaf_elm_t *trail_elm_b;
+};
+
+/**
+ * These functions the emap metadata management for merging, splitting, and
+ * reusing extents.  In particular, they set the boundary mappings from
+ * addresses to edatas.  If the result is going to be used as a slab, you
+ * still need to call emap_register_interior on it, though.
+ *
+ * Remap simply changes the szind and slab status of an extent's boundary
+ * mappings.  If the extent is not a slab, it doesn't bother with updating the
+ * end mapping (since lookups only occur in the interior of an extent for
+ * slabs).  Since the szind and slab status only make sense for active extents,
+ * this should only be called while activating or deactivating an extent.
+ *
+ * Split and merge have a "prepare" and a "commit" portion.  The prepare portion
+ * does the operations that can be done without exclusive access to the extent
+ * in question, while the commit variant requires exclusive access to maintain
+ * the emap invariants.  The only function that can fail is emap_split_prepare,
+ * and it returns true on failure (at which point the caller shouldn't commit).
+ *
+ * In all cases, "lead" refers to the lower-addressed extent, and trail to the
+ * higher-addressed one.  It's the caller's responsibility to set the edata
+ * state appropriately.
+ */
+bool emap_split_prepare(tsdn_t *tsdn, emap_t *emap, emap_prepare_t *prepare,
+    edata_t *edata, size_t size_a, edata_t *trail, size_t size_b);
+void emap_split_commit(tsdn_t *tsdn, emap_t *emap, emap_prepare_t *prepare,
+    edata_t *lead, size_t size_a, edata_t *trail, size_t size_b);
+void emap_merge_prepare(tsdn_t *tsdn, emap_t *emap, emap_prepare_t *prepare,
+    edata_t *lead, edata_t *trail);
+void emap_merge_commit(tsdn_t *tsdn, emap_t *emap, emap_prepare_t *prepare,
+    edata_t *lead, edata_t *trail);
+
+/* Assert that the emap's view of the given edata matches the edata's view. */
+void emap_do_assert_mapped(tsdn_t *tsdn, emap_t *emap, edata_t *edata);
+static inline void
+emap_assert_mapped(tsdn_t *tsdn, emap_t *emap, edata_t *edata) {
+	if (config_debug) {
+		emap_do_assert_mapped(tsdn, emap, edata);
+	}
+}
+
+/* Assert that the given edata isn't in the map. */
+void emap_do_assert_not_mapped(tsdn_t *tsdn, emap_t *emap, edata_t *edata);
+static inline void
+emap_assert_not_mapped(tsdn_t *tsdn, emap_t *emap, edata_t *edata) {
+	if (config_debug) {
+		emap_do_assert_not_mapped(tsdn, emap, edata);
+	}
+}
+
+JEMALLOC_ALWAYS_INLINE bool
+emap_edata_in_transition(tsdn_t *tsdn, emap_t *emap, edata_t *edata) {
+	assert(config_debug);
+	emap_assert_mapped(tsdn, emap, edata);
+
+	EMAP_DECLARE_RTREE_CTX;
+	rtree_contents_t contents = rtree_read(tsdn, &emap->rtree, rtree_ctx,
+	    (uintptr_t)edata_base_get(edata));
+
+	return edata_state_in_transition(contents.metadata.state);
+}
+
+JEMALLOC_ALWAYS_INLINE bool
+emap_edata_is_acquired(tsdn_t *tsdn, emap_t *emap, edata_t *edata) {
+	if (!config_debug) {
+		/* For assertions only. */
+		return false;
+	}
+
+	/*
+	 * The edata is considered acquired if no other threads will attempt to
+	 * read / write any fields from it.  This includes a few cases:
+	 *
+	 * 1) edata not hooked into emap yet -- This implies the edata just got
+	 * allocated or initialized.
+	 *
+	 * 2) in an active or transition state -- In both cases, the edata can
+	 * be discovered from the emap, however the state tracked in the rtree
+	 * will prevent other threads from accessing the actual edata.
+	 */
+	EMAP_DECLARE_RTREE_CTX;
+	rtree_leaf_elm_t *elm = rtree_leaf_elm_lookup(tsdn, &emap->rtree,
+	    rtree_ctx, (uintptr_t)edata_base_get(edata), /* dependent */ true,
+	    /* init_missing */ false);
+	if (elm == NULL) {
+		return true;
+	}
+	rtree_contents_t contents = rtree_leaf_elm_read(tsdn, &emap->rtree, elm,
+	    /* dependent */ true);
+	if (contents.edata == NULL ||
+	    contents.metadata.state == extent_state_active ||
+	    edata_state_in_transition(contents.metadata.state)) {
+		return true;
+	}
+
+	return false;
+}
+
+JEMALLOC_ALWAYS_INLINE void
+extent_assert_can_coalesce(const edata_t *inner, const edata_t *outer) {
+	assert(edata_arena_ind_get(inner) == edata_arena_ind_get(outer));
+	assert(edata_pai_get(inner) == edata_pai_get(outer));
+	assert(edata_committed_get(inner) == edata_committed_get(outer));
+	assert(edata_state_get(inner) == extent_state_active);
+	assert(edata_state_get(outer) == extent_state_merging);
+	assert(!edata_guarded_get(inner) && !edata_guarded_get(outer));
+	assert(edata_base_get(inner) == edata_past_get(outer) ||
+	    edata_base_get(outer) == edata_past_get(inner));
+}
+
+JEMALLOC_ALWAYS_INLINE void
+extent_assert_can_expand(const edata_t *original, const edata_t *expand) {
+	assert(edata_arena_ind_get(original) == edata_arena_ind_get(expand));
+	assert(edata_pai_get(original) == edata_pai_get(expand));
+	assert(edata_state_get(original) == extent_state_active);
+	assert(edata_state_get(expand) == extent_state_merging);
+	assert(edata_past_get(original) == edata_base_get(expand));
+}
+
+JEMALLOC_ALWAYS_INLINE edata_t *
+emap_edata_lookup(tsdn_t *tsdn, emap_t *emap, const void *ptr) {
+	EMAP_DECLARE_RTREE_CTX;
+
+	return rtree_read(tsdn, &emap->rtree, rtree_ctx, (uintptr_t)ptr).edata;
+}
+
+/* Fills in alloc_ctx with the info in the map. */
+JEMALLOC_ALWAYS_INLINE void
+emap_alloc_ctx_lookup(tsdn_t *tsdn, emap_t *emap, const void *ptr,
+    emap_alloc_ctx_t *alloc_ctx) {
+	EMAP_DECLARE_RTREE_CTX;
+
+	rtree_metadata_t metadata = rtree_metadata_read(tsdn, &emap->rtree,
+	    rtree_ctx, (uintptr_t)ptr);
+	alloc_ctx->szind = metadata.szind;
+	alloc_ctx->slab = metadata.slab;
+}
+
+/* The pointer must be mapped. */
+JEMALLOC_ALWAYS_INLINE void
+emap_full_alloc_ctx_lookup(tsdn_t *tsdn, emap_t *emap, const void *ptr,
+    emap_full_alloc_ctx_t *full_alloc_ctx) {
+	EMAP_DECLARE_RTREE_CTX;
+
+	rtree_contents_t contents = rtree_read(tsdn, &emap->rtree, rtree_ctx,
+	    (uintptr_t)ptr);
+	full_alloc_ctx->edata = contents.edata;
+	full_alloc_ctx->szind = contents.metadata.szind;
+	full_alloc_ctx->slab = contents.metadata.slab;
+}
+
+/*
+ * The pointer is allowed to not be mapped.
+ *
+ * Returns true when the pointer is not present.
+ */
+JEMALLOC_ALWAYS_INLINE bool
+emap_full_alloc_ctx_try_lookup(tsdn_t *tsdn, emap_t *emap, const void *ptr,
+    emap_full_alloc_ctx_t *full_alloc_ctx) {
+	EMAP_DECLARE_RTREE_CTX;
+
+	rtree_contents_t contents;
+	bool err = rtree_read_independent(tsdn, &emap->rtree, rtree_ctx,
+	    (uintptr_t)ptr, &contents);
+	if (err) {
+		return true;
+	}
+	full_alloc_ctx->edata = contents.edata;
+	full_alloc_ctx->szind = contents.metadata.szind;
+	full_alloc_ctx->slab = contents.metadata.slab;
+	return false;
+}
+
+/*
+ * Only used on the fastpath of free.  Returns true when cannot be fulfilled by
+ * fast path, e.g. when the metadata key is not cached.
+ */
+JEMALLOC_ALWAYS_INLINE bool
+emap_alloc_ctx_try_lookup_fast(tsd_t *tsd, emap_t *emap, const void *ptr,
+    emap_alloc_ctx_t *alloc_ctx) {
+	/* Use the unsafe getter since this may gets called during exit. */
+	rtree_ctx_t *rtree_ctx = tsd_rtree_ctxp_get_unsafe(tsd);
+
+	rtree_metadata_t metadata;
+	bool err = rtree_metadata_try_read_fast(tsd_tsdn(tsd), &emap->rtree,
+	    rtree_ctx, (uintptr_t)ptr, &metadata);
+	if (err) {
+		return true;
+	}
+	alloc_ctx->szind = metadata.szind;
+	alloc_ctx->slab = metadata.slab;
+	return false;
+}
+
+/*
+ * We want to do batch lookups out of the cache bins, which use
+ * cache_bin_ptr_array_get to access the i'th element of the bin (since they
+ * invert usual ordering in deciding what to flush).  This lets the emap avoid
+ * caring about its caller's ordering.
+ */
+typedef const void *(*emap_ptr_getter)(void *ctx, size_t ind);
+/*
+ * This allows size-checking assertions, which we can only do while we're in the
+ * process of edata lookups.
+ */
+typedef void (*emap_metadata_visitor)(void *ctx, emap_full_alloc_ctx_t *alloc_ctx);
+
+typedef union emap_batch_lookup_result_u emap_batch_lookup_result_t;
+union emap_batch_lookup_result_u {
+	edata_t *edata;
+	rtree_leaf_elm_t *rtree_leaf;
+};
+
+JEMALLOC_ALWAYS_INLINE void
+emap_edata_lookup_batch(tsd_t *tsd, emap_t *emap, size_t nptrs,
+    emap_ptr_getter ptr_getter, void *ptr_getter_ctx,
+    emap_metadata_visitor metadata_visitor, void *metadata_visitor_ctx,
+    emap_batch_lookup_result_t *result) {
+	/* Avoids null-checking tsdn in the loop below. */
+	util_assume(tsd != NULL);
+	rtree_ctx_t *rtree_ctx = tsd_rtree_ctxp_get(tsd);
+
+	for (size_t i = 0; i < nptrs; i++) {
+		const void *ptr = ptr_getter(ptr_getter_ctx, i);
+		/*
+		 * Reuse the edatas array as a temp buffer, lying a little about
+		 * the types.
+		 */
+		result[i].rtree_leaf = rtree_leaf_elm_lookup(tsd_tsdn(tsd),
+		    &emap->rtree, rtree_ctx, (uintptr_t)ptr,
+		    /* dependent */ true, /* init_missing */ false);
+	}
+
+	for (size_t i = 0; i < nptrs; i++) {
+		rtree_leaf_elm_t *elm = result[i].rtree_leaf;
+		rtree_contents_t contents = rtree_leaf_elm_read(tsd_tsdn(tsd),
+		    &emap->rtree, elm, /* dependent */ true);
+		result[i].edata = contents.edata;
+		emap_full_alloc_ctx_t alloc_ctx;
+		/*
+		 * Not all these fields are read in practice by the metadata
+		 * visitor.  But the compiler can easily optimize away the ones
+		 * that aren't, so no sense in being incomplete.
+		 */
+		alloc_ctx.szind = contents.metadata.szind;
+		alloc_ctx.slab = contents.metadata.slab;
+		alloc_ctx.edata = contents.edata;
+		metadata_visitor(metadata_visitor_ctx, &alloc_ctx);
+	}
+}
+
+#endif /* JEMALLOC_INTERNAL_EMAP_H */

BeefRT/JEMalloc/include/jemalloc/internal/emitter.h

@@ -0,0 +1,510 @@
+#ifndef JEMALLOC_INTERNAL_EMITTER_H
+#define JEMALLOC_INTERNAL_EMITTER_H
+
+#include "jemalloc/internal/ql.h"
+
+typedef enum emitter_output_e emitter_output_t;
+enum emitter_output_e {
+	emitter_output_json,
+	emitter_output_json_compact,
+	emitter_output_table
+};
+
+typedef enum emitter_justify_e emitter_justify_t;
+enum emitter_justify_e {
+	emitter_justify_left,
+	emitter_justify_right,
+	/* Not for users; just to pass to internal functions. */
+	emitter_justify_none
+};
+
+typedef enum emitter_type_e emitter_type_t;
+enum emitter_type_e {
+	emitter_type_bool,
+	emitter_type_int,
+	emitter_type_int64,
+	emitter_type_unsigned,
+	emitter_type_uint32,
+	emitter_type_uint64,
+	emitter_type_size,
+	emitter_type_ssize,
+	emitter_type_string,
+	/*
+	 * A title is a column title in a table; it's just a string, but it's
+	 * not quoted.
+	 */
+	emitter_type_title,
+};
+
+typedef struct emitter_col_s emitter_col_t;
+struct emitter_col_s {
+	/* Filled in by the user. */
+	emitter_justify_t justify;
+	int width;
+	emitter_type_t type;
+	union {
+		bool bool_val;
+		int int_val;
+		unsigned unsigned_val;
+		uint32_t uint32_val;
+		uint32_t uint32_t_val;
+		uint64_t uint64_val;
+		uint64_t uint64_t_val;
+		size_t size_val;
+		ssize_t ssize_val;
+		const char *str_val;
+	};
+
+	/* Filled in by initialization. */
+	ql_elm(emitter_col_t) link;
+};
+
+typedef struct emitter_row_s emitter_row_t;
+struct emitter_row_s {
+	ql_head(emitter_col_t) cols;
+};
+
+typedef struct emitter_s emitter_t;
+struct emitter_s {
+	emitter_output_t output;
+	/* The output information. */
+	write_cb_t *write_cb;
+	void *cbopaque;
+	int nesting_depth;
+	/* True if we've already emitted a value at the given depth. */
+	bool item_at_depth;
+	/* True if we emitted a key and will emit corresponding value next. */
+	bool emitted_key;
+};
+
+static inline bool
+emitter_outputs_json(emitter_t *emitter) {
+	return emitter->output == emitter_output_json ||
+	    emitter->output == emitter_output_json_compact;
+}
+
+/* Internal convenience function.  Write to the emitter the given string. */
+JEMALLOC_FORMAT_PRINTF(2, 3)
+static inline void
+emitter_printf(emitter_t *emitter, const char *format, ...) {
+	va_list ap;
+
+	va_start(ap, format);
+	malloc_vcprintf(emitter->write_cb, emitter->cbopaque, format, ap);
+	va_end(ap);
+}
+
+static inline const char * JEMALLOC_FORMAT_ARG(3)
+emitter_gen_fmt(char *out_fmt, size_t out_size, const char *fmt_specifier,
+    emitter_justify_t justify, int width) {
+	size_t written;
+	fmt_specifier++;
+	if (justify == emitter_justify_none) {
+		written = malloc_snprintf(out_fmt, out_size,
+		    "%%%s", fmt_specifier);
+	} else if (justify == emitter_justify_left) {
+		written = malloc_snprintf(out_fmt, out_size,
+		    "%%-%d%s", width, fmt_specifier);
+	} else {
+		written = malloc_snprintf(out_fmt, out_size,
+		    "%%%d%s", width, fmt_specifier);
+	}
+	/* Only happens in case of bad format string, which *we* choose. */
+	assert(written <  out_size);
+	return out_fmt;
+}
+
+/*
+ * Internal.  Emit the given value type in the relevant encoding (so that the
+ * bool true gets mapped to json "true", but the string "true" gets mapped to
+ * json "\"true\"", for instance.
+ *
+ * Width is ignored if justify is emitter_justify_none.
+ */
+static inline void
+emitter_print_value(emitter_t *emitter, emitter_justify_t justify, int width,
+    emitter_type_t value_type, const void *value) {
+	size_t str_written;
+#define BUF_SIZE 256
+#define FMT_SIZE 10
+	/*
+	 * We dynamically generate a format string to emit, to let us use the
+	 * snprintf machinery.  This is kinda hacky, but gets the job done
+	 * quickly without having to think about the various snprintf edge
+	 * cases.
+	 */
+	char fmt[FMT_SIZE];
+	char buf[BUF_SIZE];
+
+#define EMIT_SIMPLE(type, format)					\
+	emitter_printf(emitter,						\
+	    emitter_gen_fmt(fmt, FMT_SIZE, format, justify, width),	\
+	    *(const type *)value);
+
+	switch (value_type) {
+	case emitter_type_bool:
+		emitter_printf(emitter,
+		    emitter_gen_fmt(fmt, FMT_SIZE, "%s", justify, width),
+		    *(const bool *)value ?  "true" : "false");
+		break;
+	case emitter_type_int:
+		EMIT_SIMPLE(int, "%d")
+		break;
+	case emitter_type_int64:
+		EMIT_SIMPLE(int64_t, "%" FMTd64)
+		break;
+	case emitter_type_unsigned:
+		EMIT_SIMPLE(unsigned, "%u")
+		break;
+	case emitter_type_ssize:
+		EMIT_SIMPLE(ssize_t, "%zd")
+		break;
+	case emitter_type_size:
+		EMIT_SIMPLE(size_t, "%zu")
+		break;
+	case emitter_type_string:
+		str_written = malloc_snprintf(buf, BUF_SIZE, "\"%s\"",
+		    *(const char *const *)value);
+		/*
+		 * We control the strings we output; we shouldn't get anything
+		 * anywhere near the fmt size.
+		 */
+		assert(str_written < BUF_SIZE);
+		emitter_printf(emitter,
+		    emitter_gen_fmt(fmt, FMT_SIZE, "%s", justify, width), buf);
+		break;
+	case emitter_type_uint32:
+		EMIT_SIMPLE(uint32_t, "%" FMTu32)
+		break;
+	case emitter_type_uint64:
+		EMIT_SIMPLE(uint64_t, "%" FMTu64)
+		break;
+	case emitter_type_title:
+		EMIT_SIMPLE(char *const, "%s");
+		break;
+	default:
+		unreachable();
+	}
+#undef BUF_SIZE
+#undef FMT_SIZE
+}
+
+
+/* Internal functions.  In json mode, tracks nesting state. */
+static inline void
+emitter_nest_inc(emitter_t *emitter) {
+	emitter->nesting_depth++;
+	emitter->item_at_depth = false;
+}
+
+static inline void
+emitter_nest_dec(emitter_t *emitter) {
+	emitter->nesting_depth--;
+	emitter->item_at_depth = true;
+}
+
+static inline void
+emitter_indent(emitter_t *emitter) {
+	int amount = emitter->nesting_depth;
+	const char *indent_str;
+	assert(emitter->output != emitter_output_json_compact);
+	if (emitter->output == emitter_output_json) {
+		indent_str = "\t";
+	} else {
+		amount *= 2;
+		indent_str = " ";
+	}
+	for (int i = 0; i < amount; i++) {
+		emitter_printf(emitter, "%s", indent_str);
+	}
+}
+
+static inline void
+emitter_json_key_prefix(emitter_t *emitter) {
+	assert(emitter_outputs_json(emitter));
+	if (emitter->emitted_key) {
+		emitter->emitted_key = false;
+		return;
+	}
+	if (emitter->item_at_depth) {
+		emitter_printf(emitter, ",");
+	}
+	if (emitter->output != emitter_output_json_compact) {
+		emitter_printf(emitter, "\n");
+		emitter_indent(emitter);
+	}
+}
+
+/******************************************************************************/
+/* Public functions for emitter_t. */
+
+static inline void
+emitter_init(emitter_t *emitter, emitter_output_t emitter_output,
+    write_cb_t *write_cb, void *cbopaque) {
+	emitter->output = emitter_output;
+	emitter->write_cb = write_cb;
+	emitter->cbopaque = cbopaque;
+	emitter->item_at_depth = false;
+	emitter->emitted_key = false;
+	emitter->nesting_depth = 0;
+}
+
+/******************************************************************************/
+/* JSON public API. */
+
+/*
+ * Emits a key (e.g. as appears in an object). The next json entity emitted will
+ * be the corresponding value.
+ */
+static inline void
+emitter_json_key(emitter_t *emitter, const char *json_key) {
+	if (emitter_outputs_json(emitter)) {
+		emitter_json_key_prefix(emitter);
+		emitter_printf(emitter, "\"%s\":%s", json_key,
+		    emitter->output == emitter_output_json_compact ? "" : " ");
+		emitter->emitted_key = true;
+	}
+}
+
+static inline void
+emitter_json_value(emitter_t *emitter, emitter_type_t value_type,
+    const void *value) {
+	if (emitter_outputs_json(emitter)) {
+		emitter_json_key_prefix(emitter);
+		emitter_print_value(emitter, emitter_justify_none, -1,
+		    value_type, value);
+		emitter->item_at_depth = true;
+	}
+}
+
+/* Shorthand for calling emitter_json_key and then emitter_json_value. */
+static inline void
+emitter_json_kv(emitter_t *emitter, const char *json_key,
+    emitter_type_t value_type, const void *value) {
+	emitter_json_key(emitter, json_key);
+	emitter_json_value(emitter, value_type, value);
+}
+
+static inline void
+emitter_json_array_begin(emitter_t *emitter) {
+	if (emitter_outputs_json(emitter)) {
+		emitter_json_key_prefix(emitter);
+		emitter_printf(emitter, "[");
+		emitter_nest_inc(emitter);
+	}
+}
+
+/* Shorthand for calling emitter_json_key and then emitter_json_array_begin. */
+static inline void
+emitter_json_array_kv_begin(emitter_t *emitter, const char *json_key) {
+	emitter_json_key(emitter, json_key);
+	emitter_json_array_begin(emitter);
+}
+
+static inline void
+emitter_json_array_end(emitter_t *emitter) {
+	if (emitter_outputs_json(emitter)) {
+		assert(emitter->nesting_depth > 0);
+		emitter_nest_dec(emitter);
+		if (emitter->output != emitter_output_json_compact) {
+			emitter_printf(emitter, "\n");
+			emitter_indent(emitter);
+		}
+		emitter_printf(emitter, "]");
+	}
+}
+
+static inline void
+emitter_json_object_begin(emitter_t *emitter) {
+	if (emitter_outputs_json(emitter)) {
+		emitter_json_key_prefix(emitter);
+		emitter_printf(emitter, "{");
+		emitter_nest_inc(emitter);
+	}
+}
+
+/* Shorthand for calling emitter_json_key and then emitter_json_object_begin. */
+static inline void
+emitter_json_object_kv_begin(emitter_t *emitter, const char *json_key) {
+	emitter_json_key(emitter, json_key);
+	emitter_json_object_begin(emitter);
+}
+
+static inline void
+emitter_json_object_end(emitter_t *emitter) {
+	if (emitter_outputs_json(emitter)) {
+		assert(emitter->nesting_depth > 0);
+		emitter_nest_dec(emitter);
+		if (emitter->output != emitter_output_json_compact) {
+			emitter_printf(emitter, "\n");
+			emitter_indent(emitter);
+		}
+		emitter_printf(emitter, "}");
+	}
+}
+
+
+/******************************************************************************/
+/* Table public API. */
+
+static inline void
+emitter_table_dict_begin(emitter_t *emitter, const char *table_key) {
+	if (emitter->output == emitter_output_table) {
+		emitter_indent(emitter);
+		emitter_printf(emitter, "%s\n", table_key);
+		emitter_nest_inc(emitter);
+	}
+}
+
+static inline void
+emitter_table_dict_end(emitter_t *emitter) {
+	if (emitter->output == emitter_output_table) {
+		emitter_nest_dec(emitter);
+	}
+}
+
+static inline void
+emitter_table_kv_note(emitter_t *emitter, const char *table_key,
+    emitter_type_t value_type, const void *value,
+    const char *table_note_key, emitter_type_t table_note_value_type,
+    const void *table_note_value) {
+	if (emitter->output == emitter_output_table) {
+		emitter_indent(emitter);
+		emitter_printf(emitter, "%s: ", table_key);
+		emitter_print_value(emitter, emitter_justify_none, -1,
+		    value_type, value);
+		if (table_note_key != NULL) {
+			emitter_printf(emitter, " (%s: ", table_note_key);
+			emitter_print_value(emitter, emitter_justify_none, -1,
+			    table_note_value_type, table_note_value);
+			emitter_printf(emitter, ")");
+		}
+		emitter_printf(emitter, "\n");
+	}
+	emitter->item_at_depth = true;
+}
+
+static inline void
+emitter_table_kv(emitter_t *emitter, const char *table_key,
+    emitter_type_t value_type, const void *value) {
+	emitter_table_kv_note(emitter, table_key, value_type, value, NULL,
+	    emitter_type_bool, NULL);
+}
+
+
+/* Write to the emitter the given string, but only in table mode. */
+JEMALLOC_FORMAT_PRINTF(2, 3)
+static inline void
+emitter_table_printf(emitter_t *emitter, const char *format, ...) {
+	if (emitter->output == emitter_output_table) {
+		va_list ap;
+		va_start(ap, format);
+		malloc_vcprintf(emitter->write_cb, emitter->cbopaque, format, ap);
+		va_end(ap);
+	}
+}
+
+static inline void
+emitter_table_row(emitter_t *emitter, emitter_row_t *row) {
+	if (emitter->output != emitter_output_table) {
+		return;
+	}
+	emitter_col_t *col;
+	ql_foreach(col, &row->cols, link) {
+		emitter_print_value(emitter, col->justify, col->width,
+		    col->type, (const void *)&col->bool_val);
+	}
+	emitter_table_printf(emitter, "\n");
+}
+
+static inline void
+emitter_row_init(emitter_row_t *row) {
+	ql_new(&row->cols);
+}
+
+static inline void
+emitter_col_init(emitter_col_t *col, emitter_row_t *row) {
+	ql_elm_new(col, link);
+	ql_tail_insert(&row->cols, col, link);
+}
+
+
+/******************************************************************************/
+/*
+ * Generalized public API. Emits using either JSON or table, according to
+ * settings in the emitter_t. */
+
+/*
+ * Note emits a different kv pair as well, but only in table mode.  Omits the
+ * note if table_note_key is NULL.
+ */
+static inline void
+emitter_kv_note(emitter_t *emitter, const char *json_key, const char *table_key,
+    emitter_type_t value_type, const void *value,
+    const char *table_note_key, emitter_type_t table_note_value_type,
+    const void *table_note_value) {
+	if (emitter_outputs_json(emitter)) {
+		emitter_json_key(emitter, json_key);
+		emitter_json_value(emitter, value_type, value);
+	} else {
+		emitter_table_kv_note(emitter, table_key, value_type, value,
+		    table_note_key, table_note_value_type, table_note_value);
+	}
+	emitter->item_at_depth = true;
+}
+
+static inline void
+emitter_kv(emitter_t *emitter, const char *json_key, const char *table_key,
+    emitter_type_t value_type, const void *value) {
+	emitter_kv_note(emitter, json_key, table_key, value_type, value, NULL,
+	    emitter_type_bool, NULL);
+}
+
+static inline void
+emitter_dict_begin(emitter_t *emitter, const char *json_key,
+    const char *table_header) {
+	if (emitter_outputs_json(emitter)) {
+		emitter_json_key(emitter, json_key);
+		emitter_json_object_begin(emitter);
+	} else {
+		emitter_table_dict_begin(emitter, table_header);
+	}
+}
+
+static inline void
+emitter_dict_end(emitter_t *emitter) {
+	if (emitter_outputs_json(emitter)) {
+		emitter_json_object_end(emitter);
+	} else {
+		emitter_table_dict_end(emitter);
+	}
+}
+
+static inline void
+emitter_begin(emitter_t *emitter) {
+	if (emitter_outputs_json(emitter)) {
+		assert(emitter->nesting_depth == 0);
+		emitter_printf(emitter, "{");
+		emitter_nest_inc(emitter);
+	} else {
+		/*
+		 * This guarantees that we always call write_cb at least once.
+		 * This is useful if some invariant is established by each call
+		 * to write_cb, but doesn't hold initially: e.g., some buffer
+		 * holds a null-terminated string.
+		 */
+		emitter_printf(emitter, "%s", "");
+	}
+}
+
+static inline void
+emitter_end(emitter_t *emitter) {
+	if (emitter_outputs_json(emitter)) {
+		assert(emitter->nesting_depth == 1);
+		emitter_nest_dec(emitter);
+		emitter_printf(emitter, "%s", emitter->output ==
+		    emitter_output_json_compact ? "}" : "\n}\n");
+	}
+}
+
+#endif /* JEMALLOC_INTERNAL_EMITTER_H */

BeefRT/JEMalloc/include/jemalloc/internal/eset.h

@@ -0,0 +1,77 @@
+#ifndef JEMALLOC_INTERNAL_ESET_H
+#define JEMALLOC_INTERNAL_ESET_H
+
+#include "jemalloc/internal/atomic.h"
+#include "jemalloc/internal/fb.h"
+#include "jemalloc/internal/edata.h"
+#include "jemalloc/internal/mutex.h"
+
+/*
+ * An eset ("extent set") is a quantized collection of extents, with built-in
+ * LRU queue.
+ *
+ * This class is not thread-safe; synchronization must be done externally if
+ * there are mutating operations.  One exception is the stats counters, which
+ * may be read without any locking.
+ */
+
+typedef struct eset_bin_s eset_bin_t;
+struct eset_bin_s {
+	edata_heap_t heap;
+	/*
+	 * We do first-fit across multiple size classes.  If we compared against
+	 * the min element in each heap directly, we'd take a cache miss per
+	 * extent we looked at.  If we co-locate the edata summaries, we only
+	 * take a miss on the edata we're actually going to return (which is
+	 * inevitable anyways).
+	 */
+	edata_cmp_summary_t heap_min;
+};
+
+typedef struct eset_bin_stats_s eset_bin_stats_t;
+struct eset_bin_stats_s {
+	atomic_zu_t nextents;
+	atomic_zu_t nbytes;
+};
+
+typedef struct eset_s eset_t;
+struct eset_s {
+	/* Bitmap for which set bits correspond to non-empty heaps. */
+	fb_group_t bitmap[FB_NGROUPS(SC_NPSIZES + 1)];
+
+	/* Quantized per size class heaps of extents. */
+	eset_bin_t bins[SC_NPSIZES + 1];
+
+	eset_bin_stats_t bin_stats[SC_NPSIZES + 1];
+
+	/* LRU of all extents in heaps. */
+	edata_list_inactive_t lru;
+
+	/* Page sum for all extents in heaps. */
+	atomic_zu_t npages;
+
+	/*
+	 * A duplication of the data in the containing ecache.  We use this only
+	 * for assertions on the states of the passed-in extents.
+	 */
+	extent_state_t state;
+};
+
+void eset_init(eset_t *eset, extent_state_t state);
+
+size_t eset_npages_get(eset_t *eset);
+/* Get the number of extents in the given page size index. */
+size_t eset_nextents_get(eset_t *eset, pszind_t ind);
+/* Get the sum total bytes of the extents in the given page size index. */
+size_t eset_nbytes_get(eset_t *eset, pszind_t ind);
+
+void eset_insert(eset_t *eset, edata_t *edata);
+void eset_remove(eset_t *eset, edata_t *edata);
+/*
+ * Select an extent from this eset of the given size and alignment.  Returns
+ * null if no such item could be found.
+ */
+edata_t *eset_fit(eset_t *eset, size_t esize, size_t alignment, bool exact_only,
+    unsigned lg_max_fit);
+
+#endif /* JEMALLOC_INTERNAL_ESET_H */

BeefRT/JEMalloc/include/jemalloc/internal/exp_grow.h

@@ -0,0 +1,50 @@
+#ifndef JEMALLOC_INTERNAL_EXP_GROW_H
+#define JEMALLOC_INTERNAL_EXP_GROW_H
+
+typedef struct exp_grow_s exp_grow_t;
+struct exp_grow_s {
+	/*
+	 * Next extent size class in a growing series to use when satisfying a
+	 * request via the extent hooks (only if opt_retain).  This limits the
+	 * number of disjoint virtual memory ranges so that extent merging can
+	 * be effective even if multiple arenas' extent allocation requests are
+	 * highly interleaved.
+	 *
+	 * retain_grow_limit is the max allowed size ind to expand (unless the
+	 * required size is greater).  Default is no limit, and controlled
+	 * through mallctl only.
+	 */
+	pszind_t next;
+	pszind_t limit;
+};
+
+static inline bool
+exp_grow_size_prepare(exp_grow_t *exp_grow, size_t alloc_size_min,
+    size_t *r_alloc_size, pszind_t *r_skip) {
+	*r_skip = 0;
+	*r_alloc_size = sz_pind2sz(exp_grow->next + *r_skip);
+	while (*r_alloc_size < alloc_size_min) {
+		(*r_skip)++;
+		if (exp_grow->next + *r_skip  >=
+		    sz_psz2ind(SC_LARGE_MAXCLASS)) {
+			/* Outside legal range. */
+			return true;
+		}
+		*r_alloc_size = sz_pind2sz(exp_grow->next + *r_skip);
+	}
+	return false;
+}
+
+static inline void
+exp_grow_size_commit(exp_grow_t *exp_grow, pszind_t skip) {
+	if (exp_grow->next + skip + 1 <= exp_grow->limit) {
+		exp_grow->next += skip + 1;
+	} else {
+		exp_grow->next = exp_grow->limit;
+	}
+
+}
+
+void exp_grow_init(exp_grow_t *exp_grow);
+
+#endif /* JEMALLOC_INTERNAL_EXP_GROW_H */

BeefRT/JEMalloc/include/jemalloc/internal/extent.h

@@ -0,0 +1,137 @@
+#ifndef JEMALLOC_INTERNAL_EXTENT_H
+#define JEMALLOC_INTERNAL_EXTENT_H
+
+#include "jemalloc/internal/ecache.h"
+#include "jemalloc/internal/ehooks.h"
+#include "jemalloc/internal/ph.h"
+#include "jemalloc/internal/rtree.h"
+
+/*
+ * This module contains the page-level allocator.  It chooses the addresses that
+ * allocations requested by other modules will inhabit, and updates the global
+ * metadata to reflect allocation/deallocation/purging decisions.
+ */
+
+/*
+ * When reuse (and split) an active extent, (1U << opt_lg_extent_max_active_fit)
+ * is the max ratio between the size of the active extent and the new extent.
+ */
+#define LG_EXTENT_MAX_ACTIVE_FIT_DEFAULT 6
+extern size_t opt_lg_extent_max_active_fit;
+
+edata_t *ecache_alloc(tsdn_t *tsdn, pac_t *pac, ehooks_t *ehooks,
+    ecache_t *ecache, edata_t *expand_edata, size_t size, size_t alignment,
+    bool zero, bool guarded);
+edata_t *ecache_alloc_grow(tsdn_t *tsdn, pac_t *pac, ehooks_t *ehooks,
+    ecache_t *ecache, edata_t *expand_edata, size_t size, size_t alignment,
+    bool zero, bool guarded);
+void ecache_dalloc(tsdn_t *tsdn, pac_t *pac, ehooks_t *ehooks,
+    ecache_t *ecache, edata_t *edata);
+edata_t *ecache_evict(tsdn_t *tsdn, pac_t *pac, ehooks_t *ehooks,
+    ecache_t *ecache, size_t npages_min);
+
+void extent_gdump_add(tsdn_t *tsdn, const edata_t *edata);
+void extent_record(tsdn_t *tsdn, pac_t *pac, ehooks_t *ehooks, ecache_t *ecache,
+    edata_t *edata);
+void extent_dalloc_gap(tsdn_t *tsdn, pac_t *pac, ehooks_t *ehooks,
+    edata_t *edata);
+edata_t *extent_alloc_wrapper(tsdn_t *tsdn, pac_t *pac, ehooks_t *ehooks,
+    void *new_addr, size_t size, size_t alignment, bool zero, bool *commit,
+    bool growing_retained);
+void extent_dalloc_wrapper(tsdn_t *tsdn, pac_t *pac, ehooks_t *ehooks,
+    edata_t *edata);
+void extent_destroy_wrapper(tsdn_t *tsdn, pac_t *pac, ehooks_t *ehooks,
+    edata_t *edata);
+bool extent_commit_wrapper(tsdn_t *tsdn, ehooks_t *ehooks, edata_t *edata,
+    size_t offset, size_t length);
+bool extent_decommit_wrapper(tsdn_t *tsdn, ehooks_t *ehooks, edata_t *edata,
+    size_t offset, size_t length);
+bool extent_purge_lazy_wrapper(tsdn_t *tsdn, ehooks_t *ehooks, edata_t *edata,
+    size_t offset, size_t length);
+bool extent_purge_forced_wrapper(tsdn_t *tsdn, ehooks_t *ehooks, edata_t *edata,
+    size_t offset, size_t length);
+edata_t *extent_split_wrapper(tsdn_t *tsdn, pac_t *pac,
+    ehooks_t *ehooks, edata_t *edata, size_t size_a, size_t size_b,
+    bool holding_core_locks);
+bool extent_merge_wrapper(tsdn_t *tsdn, pac_t *pac, ehooks_t *ehooks,
+    edata_t *a, edata_t *b);
+bool extent_commit_zero(tsdn_t *tsdn, ehooks_t *ehooks, edata_t *edata,
+    bool commit, bool zero, bool growing_retained);
+size_t extent_sn_next(pac_t *pac);
+bool extent_boot(void);
+
+JEMALLOC_ALWAYS_INLINE bool
+extent_neighbor_head_state_mergeable(bool edata_is_head,
+    bool neighbor_is_head, bool forward) {
+	/*
+	 * Head states checking: disallow merging if the higher addr extent is a
+	 * head extent.  This helps preserve first-fit, and more importantly
+	 * makes sure no merge across arenas.
+	 */
+	if (forward) {
+		if (neighbor_is_head) {
+			return false;
+		}
+	} else {
+		if (edata_is_head) {
+			return false;
+		}
+	}
+	return true;
+}
+
+JEMALLOC_ALWAYS_INLINE bool
+extent_can_acquire_neighbor(edata_t *edata, rtree_contents_t contents,
+    extent_pai_t pai, extent_state_t expected_state, bool forward,
+    bool expanding) {
+	edata_t *neighbor = contents.edata;
+	if (neighbor == NULL) {
+		return false;
+	}
+	/* It's not safe to access *neighbor yet; must verify states first. */
+	bool neighbor_is_head = contents.metadata.is_head;
+	if (!extent_neighbor_head_state_mergeable(edata_is_head_get(edata),
+	    neighbor_is_head, forward)) {
+		return false;
+	}
+	extent_state_t neighbor_state = contents.metadata.state;
+	if (pai == EXTENT_PAI_PAC) {
+		if (neighbor_state != expected_state) {
+			return false;
+		}
+		/* From this point, it's safe to access *neighbor. */
+		if (!expanding && (edata_committed_get(edata) !=
+		    edata_committed_get(neighbor))) {
+			/*
+			 * Some platforms (e.g. Windows) require an explicit
+			 * commit step (and writing to uncommitted memory is not
+			 * allowed).
+			 */
+			return false;
+		}
+	} else {
+		if (neighbor_state == extent_state_active) {
+			return false;
+		}
+		/* From this point, it's safe to access *neighbor. */
+	}
+
+	assert(edata_pai_get(edata) == pai);
+	if (edata_pai_get(neighbor) != pai) {
+		return false;
+	}
+	if (opt_retain) {
+		assert(edata_arena_ind_get(edata) ==
+		    edata_arena_ind_get(neighbor));
+	} else {
+		if (edata_arena_ind_get(edata) !=
+		    edata_arena_ind_get(neighbor)) {
+			return false;
+		}
+	}
+	assert(!edata_guarded_get(edata) && !edata_guarded_get(neighbor));
+
+	return true;
+}
+
+#endif /* JEMALLOC_INTERNAL_EXTENT_H */

BeefRT/JEMalloc/include/jemalloc/internal/extent_dss.h

@@ -0,0 +1,26 @@
+#ifndef JEMALLOC_INTERNAL_EXTENT_DSS_H
+#define JEMALLOC_INTERNAL_EXTENT_DSS_H
+
+typedef enum {
+	dss_prec_disabled  = 0,
+	dss_prec_primary   = 1,
+	dss_prec_secondary = 2,
+
+	dss_prec_limit     = 3
+} dss_prec_t;
+#define DSS_PREC_DEFAULT dss_prec_secondary
+#define DSS_DEFAULT "secondary"
+
+extern const char *dss_prec_names[];
+
+extern const char *opt_dss;
+
+dss_prec_t extent_dss_prec_get(void);
+bool extent_dss_prec_set(dss_prec_t dss_prec);
+void *extent_alloc_dss(tsdn_t *tsdn, arena_t *arena, void *new_addr,
+    size_t size, size_t alignment, bool *zero, bool *commit);
+bool extent_in_dss(void *addr);
+bool extent_dss_mergeable(void *addr_a, void *addr_b);
+void extent_dss_boot(void);
+
+#endif /* JEMALLOC_INTERNAL_EXTENT_DSS_H */

BeefRT/JEMalloc/include/jemalloc/internal/extent_mmap.h

@@ -0,0 +1,10 @@
+#ifndef JEMALLOC_INTERNAL_EXTENT_MMAP_EXTERNS_H
+#define JEMALLOC_INTERNAL_EXTENT_MMAP_EXTERNS_H
+
+extern bool opt_retain;
+
+void *extent_alloc_mmap(void *new_addr, size_t size, size_t alignment,
+    bool *zero, bool *commit);
+bool extent_dalloc_mmap(void *addr, size_t size);
+
+#endif /* JEMALLOC_INTERNAL_EXTENT_MMAP_EXTERNS_H */

BeefRT/JEMalloc/include/jemalloc/internal/fb.h

@@ -0,0 +1,373 @@
+#ifndef JEMALLOC_INTERNAL_FB_H
+#define JEMALLOC_INTERNAL_FB_H
+
+/*
+ * The flat bitmap module.  This has a larger API relative to the bitmap module
+ * (supporting things like backwards searches, and searching for both set and
+ * unset bits), at the cost of slower operations for very large bitmaps.
+ *
+ * Initialized flat bitmaps start at all-zeros (all bits unset).
+ */
+
+typedef unsigned long fb_group_t;
+#define FB_GROUP_BITS (ZU(1) << (LG_SIZEOF_LONG + 3))
+#define FB_NGROUPS(nbits) ((nbits) / FB_GROUP_BITS \
+    + ((nbits) % FB_GROUP_BITS == 0 ? 0 : 1))
+
+static inline void
+fb_init(fb_group_t *fb, size_t nbits) {
+	size_t ngroups = FB_NGROUPS(nbits);
+	memset(fb, 0, ngroups * sizeof(fb_group_t));
+}
+
+static inline bool
+fb_empty(fb_group_t *fb, size_t nbits) {
+	size_t ngroups = FB_NGROUPS(nbits);
+	for (size_t i = 0; i < ngroups; i++) {
+		if (fb[i] != 0) {
+			return false;
+		}
+	}
+	return true;
+}
+
+static inline bool
+fb_full(fb_group_t *fb, size_t nbits) {
+	size_t ngroups = FB_NGROUPS(nbits);
+	size_t trailing_bits = nbits % FB_GROUP_BITS;
+	size_t limit = (trailing_bits == 0 ? ngroups : ngroups - 1);
+	for (size_t i = 0; i < limit; i++) {
+		if (fb[i] != ~(fb_group_t)0) {
+			return false;
+		}
+	}
+	if (trailing_bits == 0) {
+		return true;
+	}
+	return fb[ngroups - 1] == ((fb_group_t)1 << trailing_bits) - 1;
+}
+
+static inline bool
+fb_get(fb_group_t *fb, size_t nbits, size_t bit) {
+	assert(bit < nbits);
+	size_t group_ind = bit / FB_GROUP_BITS;
+	size_t bit_ind = bit % FB_GROUP_BITS;
+	return (bool)(fb[group_ind] & ((fb_group_t)1 << bit_ind));
+}
+
+static inline void
+fb_set(fb_group_t *fb, size_t nbits, size_t bit) {
+	assert(bit < nbits);
+	size_t group_ind = bit / FB_GROUP_BITS;
+	size_t bit_ind = bit % FB_GROUP_BITS;
+	fb[group_ind] |= ((fb_group_t)1 << bit_ind);
+}
+
+static inline void
+fb_unset(fb_group_t *fb, size_t nbits, size_t bit) {
+	assert(bit < nbits);
+	size_t group_ind = bit / FB_GROUP_BITS;
+	size_t bit_ind = bit % FB_GROUP_BITS;
+	fb[group_ind] &= ~((fb_group_t)1 << bit_ind);
+}
+
+
+/*
+ * Some implementation details.  This visitation function lets us apply a group
+ * visitor to each group in the bitmap (potentially modifying it).  The mask
+ * indicates which bits are logically part of the visitation.
+ */
+typedef void (*fb_group_visitor_t)(void *ctx, fb_group_t *fb, fb_group_t mask);
+JEMALLOC_ALWAYS_INLINE void
+fb_visit_impl(fb_group_t *fb, size_t nbits, fb_group_visitor_t visit, void *ctx,
+    size_t start, size_t cnt) {
+	assert(cnt > 0);
+	assert(start + cnt <= nbits);
+	size_t group_ind = start / FB_GROUP_BITS;
+	size_t start_bit_ind = start % FB_GROUP_BITS;
+	/*
+	 * The first group is special; it's the only one we don't start writing
+	 * to from bit 0.
+	 */
+	size_t first_group_cnt = (start_bit_ind + cnt > FB_GROUP_BITS
+		? FB_GROUP_BITS - start_bit_ind : cnt);
+	/*
+	 * We can basically split affected words into:
+	 *   - The first group, where we touch only the high bits
+	 *   - The last group, where we touch only the low bits
+	 *   - The middle, where we set all the bits to the same thing.
+	 * We treat each case individually.  The last two could be merged, but
+	 * this can lead to bad codegen for those middle words.
+	 */
+	/* First group */
+	fb_group_t mask = ((~(fb_group_t)0)
+	    >> (FB_GROUP_BITS - first_group_cnt))
+	    << start_bit_ind;
+	visit(ctx, &fb[group_ind], mask);
+
+	cnt -= first_group_cnt;
+	group_ind++;
+	/* Middle groups */
+	while (cnt > FB_GROUP_BITS) {
+		visit(ctx, &fb[group_ind], ~(fb_group_t)0);
+		cnt -= FB_GROUP_BITS;
+		group_ind++;
+	}
+	/* Last group */
+	if (cnt != 0) {
+		mask = (~(fb_group_t)0) >> (FB_GROUP_BITS - cnt);
+		visit(ctx, &fb[group_ind], mask);
+	}
+}
+
+JEMALLOC_ALWAYS_INLINE void
+fb_assign_visitor(void *ctx, fb_group_t *fb, fb_group_t mask) {
+	bool val = *(bool *)ctx;
+	if (val) {
+		*fb |= mask;
+	} else {
+		*fb &= ~mask;
+	}
+}
+
+/* Sets the cnt bits starting at position start.  Must not have a 0 count. */
+static inline void
+fb_set_range(fb_group_t *fb, size_t nbits, size_t start, size_t cnt) {
+	bool val = true;
+	fb_visit_impl(fb, nbits, &fb_assign_visitor, &val, start, cnt);
+}
+
+/* Unsets the cnt bits starting at position start.  Must not have a 0 count. */
+static inline void
+fb_unset_range(fb_group_t *fb, size_t nbits, size_t start, size_t cnt) {
+	bool val = false;
+	fb_visit_impl(fb, nbits, &fb_assign_visitor, &val, start, cnt);
+}
+
+JEMALLOC_ALWAYS_INLINE void
+fb_scount_visitor(void *ctx, fb_group_t *fb, fb_group_t mask) {
+	size_t *scount = (size_t *)ctx;
+	*scount += popcount_lu(*fb & mask);
+}
+
+/* Finds the number of set bit in the of length cnt starting at start. */
+JEMALLOC_ALWAYS_INLINE size_t
+fb_scount(fb_group_t *fb, size_t nbits, size_t start, size_t cnt) {
+	size_t scount = 0;
+	fb_visit_impl(fb, nbits, &fb_scount_visitor, &scount, start, cnt);
+	return scount;
+}
+
+/* Finds the number of unset bit in the of length cnt starting at start. */
+JEMALLOC_ALWAYS_INLINE size_t
+fb_ucount(fb_group_t *fb, size_t nbits, size_t start, size_t cnt) {
+	size_t scount = fb_scount(fb, nbits, start, cnt);
+	return cnt - scount;
+}
+
+/*
+ * An implementation detail; find the first bit at position >= min_bit with the
+ * value val.
+ *
+ * Returns the number of bits in the bitmap if no such bit exists.
+ */
+JEMALLOC_ALWAYS_INLINE ssize_t
+fb_find_impl(fb_group_t *fb, size_t nbits, size_t start, bool val,
+    bool forward) {
+	assert(start < nbits);
+	size_t ngroups = FB_NGROUPS(nbits);
+	ssize_t group_ind = start / FB_GROUP_BITS;
+	size_t bit_ind = start % FB_GROUP_BITS;
+
+	fb_group_t maybe_invert = (val ? 0 : (fb_group_t)-1);
+
+	fb_group_t group = fb[group_ind];
+	group ^= maybe_invert;
+	if (forward) {
+		/* Only keep ones in bits bit_ind and above. */
+		group &= ~((1LU << bit_ind) - 1);
+	} else {
+		/*
+		 * Only keep ones in bits bit_ind and below.  You might more
+		 * naturally express this as (1 << (bit_ind + 1)) - 1, but
+		 * that shifts by an invalid amount if bit_ind is one less than
+		 * FB_GROUP_BITS.
+		 */
+		group &= ((2LU << bit_ind) - 1);
+	}
+	ssize_t group_ind_bound = forward ? (ssize_t)ngroups : -1;
+	while (group == 0) {
+		group_ind += forward ? 1 : -1;
+		if (group_ind == group_ind_bound) {
+			return forward ? (ssize_t)nbits : (ssize_t)-1;
+		}
+		group = fb[group_ind];
+		group ^= maybe_invert;
+	}
+	assert(group != 0);
+	size_t bit = forward ? ffs_lu(group) : fls_lu(group);
+	size_t pos = group_ind * FB_GROUP_BITS + bit;
+	/*
+	 * The high bits of a partially filled last group are zeros, so if we're
+	 * looking for zeros we don't want to report an invalid result.
+	 */
+	if (forward && !val && pos > nbits) {
+		return nbits;
+	}
+	return pos;
+}
+
+/*
+ * Find the first set bit in the bitmap with an index >= min_bit.  Returns the
+ * number of bits in the bitmap if no such bit exists.
+ */
+static inline size_t
+fb_ffu(fb_group_t *fb, size_t nbits, size_t min_bit) {
+	return (size_t)fb_find_impl(fb, nbits, min_bit, /* val */ false,
+	    /* forward */ true);
+}
+
+/* The same, but looks for an unset bit. */
+static inline size_t
+fb_ffs(fb_group_t *fb, size_t nbits, size_t min_bit) {
+	return (size_t)fb_find_impl(fb, nbits, min_bit, /* val */ true,
+	    /* forward */ true);
+}
+
+/*
+ * Find the last set bit in the bitmap with an index <= max_bit.  Returns -1 if
+ * no such bit exists.
+ */
+static inline ssize_t
+fb_flu(fb_group_t *fb, size_t nbits, size_t max_bit) {
+	return fb_find_impl(fb, nbits, max_bit, /* val */ false,
+	    /* forward */ false);
+}
+
+static inline ssize_t
+fb_fls(fb_group_t *fb, size_t nbits, size_t max_bit) {
+	return fb_find_impl(fb, nbits, max_bit, /* val */ true,
+	    /* forward */ false);
+}
+
+/* Returns whether or not we found a range. */
+JEMALLOC_ALWAYS_INLINE bool
+fb_iter_range_impl(fb_group_t *fb, size_t nbits, size_t start, size_t *r_begin,
+    size_t *r_len, bool val, bool forward) {
+	assert(start < nbits);
+	ssize_t next_range_begin = fb_find_impl(fb, nbits, start, val, forward);
+	if ((forward && next_range_begin == (ssize_t)nbits)
+	    || (!forward && next_range_begin == (ssize_t)-1)) {
+		return false;
+	}
+	/* Half open range; the set bits are [begin, end). */
+	ssize_t next_range_end = fb_find_impl(fb, nbits, next_range_begin, !val,
+	    forward);
+	if (forward) {
+		*r_begin = next_range_begin;
+		*r_len = next_range_end - next_range_begin;
+	} else {
+		*r_begin = next_range_end + 1;
+		*r_len = next_range_begin - next_range_end;
+	}
+	return true;
+}
+
+/*
+ * Used to iterate through ranges of set bits.
+ *
+ * Tries to find the next contiguous sequence of set bits with a first index >=
+ * start.  If one exists, puts the earliest bit of the range in *r_begin, its
+ * length in *r_len, and returns true.  Otherwise, returns false (without
+ * touching *r_begin or *r_end).
+ */
+static inline bool
+fb_srange_iter(fb_group_t *fb, size_t nbits, size_t start, size_t *r_begin,
+    size_t *r_len) {
+	return fb_iter_range_impl(fb, nbits, start, r_begin, r_len,
+	    /* val */ true, /* forward */ true);
+}
+
+/*
+ * The same as fb_srange_iter, but searches backwards from start rather than
+ * forwards.  (The position returned is still the earliest bit in the range).
+ */
+static inline bool
+fb_srange_riter(fb_group_t *fb, size_t nbits, size_t start, size_t *r_begin,
+    size_t *r_len) {
+	return fb_iter_range_impl(fb, nbits, start, r_begin, r_len,
+	    /* val */ true, /* forward */ false);
+}
+
+/* Similar to fb_srange_iter, but searches for unset bits. */
+static inline bool
+fb_urange_iter(fb_group_t *fb, size_t nbits, size_t start, size_t *r_begin,
+    size_t *r_len) {
+	return fb_iter_range_impl(fb, nbits, start, r_begin, r_len,
+	    /* val */ false, /* forward */ true);
+}
+
+/* Similar to fb_srange_riter, but searches for unset bits. */
+static inline bool
+fb_urange_riter(fb_group_t *fb, size_t nbits, size_t start, size_t *r_begin,
+    size_t *r_len) {
+	return fb_iter_range_impl(fb, nbits, start, r_begin, r_len,
+	    /* val */ false, /* forward */ false);
+}
+
+JEMALLOC_ALWAYS_INLINE size_t
+fb_range_longest_impl(fb_group_t *fb, size_t nbits, bool val) {
+	size_t begin = 0;
+	size_t longest_len = 0;
+	size_t len = 0;
+	while (begin < nbits && fb_iter_range_impl(fb, nbits, begin, &begin,
+	    &len, val, /* forward */ true)) {
+		if (len > longest_len) {
+			longest_len = len;
+		}
+		begin += len;
+	}
+	return longest_len;
+}
+
+static inline size_t
+fb_srange_longest(fb_group_t *fb, size_t nbits) {
+	return fb_range_longest_impl(fb, nbits, /* val */ true);
+}
+
+static inline size_t
+fb_urange_longest(fb_group_t *fb, size_t nbits) {
+	return fb_range_longest_impl(fb, nbits, /* val */ false);
+}
+
+/*
+ * Initializes each bit of dst with the bitwise-AND of the corresponding bits of
+ * src1 and src2.  All bitmaps must be the same size.
+ */
+static inline void
+fb_bit_and(fb_group_t *dst, fb_group_t *src1, fb_group_t *src2, size_t nbits) {
+	size_t ngroups = FB_NGROUPS(nbits);
+	for (size_t i = 0; i < ngroups; i++) {
+		dst[i] = src1[i] & src2[i];
+	}
+}
+
+/* Like fb_bit_and, but with bitwise-OR. */
+static inline void
+fb_bit_or(fb_group_t *dst, fb_group_t *src1, fb_group_t *src2, size_t nbits) {
+	size_t ngroups = FB_NGROUPS(nbits);
+	for (size_t i = 0; i < ngroups; i++) {
+		dst[i] = src1[i] | src2[i];
+	}
+}
+
+/* Initializes dst bit i to the negation of source bit i. */
+static inline void
+fb_bit_not(fb_group_t *dst, fb_group_t *src, size_t nbits) {
+	size_t ngroups = FB_NGROUPS(nbits);
+	for (size_t i = 0; i < ngroups; i++) {
+		dst[i] = ~src[i];
+	}
+}
+
+#endif /* JEMALLOC_INTERNAL_FB_H */

BeefRT/JEMalloc/include/jemalloc/internal/fxp.h

@@ -0,0 +1,126 @@
+#ifndef JEMALLOC_INTERNAL_FXP_H
+#define JEMALLOC_INTERNAL_FXP_H
+
+/*
+ * A simple fixed-point math implementation, supporting only unsigned values
+ * (with overflow being an error).
+ *
+ * It's not in general safe to use floating point in core code, because various
+ * libc implementations we get linked against can assume that malloc won't touch
+ * floating point state and call it with an unusual calling convention.
+ */
+
+/*
+ * High 16 bits are the integer part, low 16 are the fractional part.  Or
+ * equivalently, repr == 2**16 * val, where we use "val" to refer to the
+ * (imaginary) fractional representation of the true value.
+ *
+ * We pick a uint32_t here since it's convenient in some places to
+ * double the representation size (i.e. multiplication and division use
+ * 64-bit integer types), and a uint64_t is the largest type we're
+ * certain is available.
+ */
+typedef uint32_t fxp_t;
+#define FXP_INIT_INT(x) ((x) << 16)
+#define FXP_INIT_PERCENT(pct) (((pct) << 16) / 100)
+
+/*
+ * Amount of precision used in parsing and printing numbers.  The integer bound
+ * is simply because the integer part of the number gets 16 bits, and so is
+ * bounded by 65536.
+ *
+ * We use a lot of precision for the fractional part, even though most of it
+ * gets rounded off; this lets us get exact values for the important special
+ * case where the denominator is a small power of 2 (for instance,
+ * 1/512 == 0.001953125 is exactly representable even with only 16 bits of
+ * fractional precision).  We need to left-shift by 16 before dividing by
+ * 10**precision, so we pick precision to be floor(log(2**48)) = 14.
+ */
+#define FXP_INTEGER_PART_DIGITS 5
+#define FXP_FRACTIONAL_PART_DIGITS 14
+
+/*
+ * In addition to the integer and fractional parts of the number, we need to
+ * include a null character and (possibly) a decimal point.
+ */
+#define FXP_BUF_SIZE (FXP_INTEGER_PART_DIGITS + FXP_FRACTIONAL_PART_DIGITS + 2)
+
+static inline fxp_t
+fxp_add(fxp_t a, fxp_t b) {
+	return a + b;
+}
+
+static inline fxp_t
+fxp_sub(fxp_t a, fxp_t b) {
+	assert(a >= b);
+	return a - b;
+}
+
+static inline fxp_t
+fxp_mul(fxp_t a, fxp_t b) {
+	uint64_t unshifted = (uint64_t)a * (uint64_t)b;
+	/*
+	 * Unshifted is (a.val * 2**16) * (b.val * 2**16)
+	 *   == (a.val * b.val) * 2**32, but we want
+	 * (a.val * b.val) * 2 ** 16.
+	 */
+	return (uint32_t)(unshifted >> 16);
+}
+
+static inline fxp_t
+fxp_div(fxp_t a, fxp_t b) {
+	assert(b != 0);
+	uint64_t unshifted = ((uint64_t)a << 32) / (uint64_t)b;
+	/*
+	 * Unshifted is (a.val * 2**16) * (2**32) / (b.val * 2**16)
+	 *   == (a.val / b.val) * (2 ** 32), which again corresponds to a right
+	 *   shift of 16.
+	 */
+	return (uint32_t)(unshifted >> 16);
+}
+
+static inline uint32_t
+fxp_round_down(fxp_t a) {
+	return a >> 16;
+}
+
+static inline uint32_t
+fxp_round_nearest(fxp_t a) {
+	uint32_t fractional_part = (a  & ((1U << 16) - 1));
+	uint32_t increment = (uint32_t)(fractional_part >= (1U << 15));
+	return (a >> 16) + increment;
+}
+
+/*
+ * Approximately computes x * frac, without the size limitations that would be
+ * imposed by converting u to an fxp_t.
+ */
+static inline size_t
+fxp_mul_frac(size_t x_orig, fxp_t frac) {
+	assert(frac <= (1U << 16));
+	/*
+	 * Work around an over-enthusiastic warning about type limits below (on
+	 * 32-bit platforms, a size_t is always less than 1ULL << 48).
+	 */
+	uint64_t x = (uint64_t)x_orig;
+	/*
+	 * If we can guarantee no overflow, multiply first before shifting, to
+	 * preserve some precision.  Otherwise, shift first and then multiply.
+	 * In the latter case, we only lose the low 16 bits of a 48-bit number,
+	 * so we're still accurate to within 1/2**32.
+	 */
+	if (x < (1ULL << 48)) {
+		return (size_t)((x * frac) >> 16);
+	} else {
+		return (size_t)((x >> 16) * (uint64_t)frac);
+	}
+}
+
+/*
+ * Returns true on error.  Otherwise, returns false and updates *ptr to point to
+ * the first character not parsed (because it wasn't a digit).
+ */
+bool fxp_parse(fxp_t *a, const char *ptr, char **end);
+void fxp_print(fxp_t a, char buf[FXP_BUF_SIZE]);
+
+#endif /* JEMALLOC_INTERNAL_FXP_H */

BeefRT/JEMalloc/include/jemalloc/internal/hash.h

@@ -0,0 +1,320 @@
+#ifndef JEMALLOC_INTERNAL_HASH_H
+#define JEMALLOC_INTERNAL_HASH_H
+
+#include "jemalloc/internal/assert.h"
+
+/*
+ * The following hash function is based on MurmurHash3, placed into the public
+ * domain by Austin Appleby.  See https://github.com/aappleby/smhasher for
+ * details.
+ */
+
+/******************************************************************************/
+/* Internal implementation. */
+static inline uint32_t
+hash_rotl_32(uint32_t x, int8_t r) {
+	return ((x << r) | (x >> (32 - r)));
+}
+
+static inline uint64_t
+hash_rotl_64(uint64_t x, int8_t r) {
+	return ((x << r) | (x >> (64 - r)));
+}
+
+static inline uint32_t
+hash_get_block_32(const uint32_t *p, int i) {
+	/* Handle unaligned read. */
+	if (unlikely((uintptr_t)p & (sizeof(uint32_t)-1)) != 0) {
+		uint32_t ret;
+
+		memcpy(&ret, (uint8_t *)(p + i), sizeof(uint32_t));
+		return ret;
+	}
+
+	return p[i];
+}
+
+static inline uint64_t
+hash_get_block_64(const uint64_t *p, int i) {
+	/* Handle unaligned read. */
+	if (unlikely((uintptr_t)p & (sizeof(uint64_t)-1)) != 0) {
+		uint64_t ret;
+
+		memcpy(&ret, (uint8_t *)(p + i), sizeof(uint64_t));
+		return ret;
+	}
+
+	return p[i];
+}
+
+static inline uint32_t
+hash_fmix_32(uint32_t h) {
+	h ^= h >> 16;
+	h *= 0x85ebca6b;
+	h ^= h >> 13;
+	h *= 0xc2b2ae35;
+	h ^= h >> 16;
+
+	return h;
+}
+
+static inline uint64_t
+hash_fmix_64(uint64_t k) {
+	k ^= k >> 33;
+	k *= KQU(0xff51afd7ed558ccd);
+	k ^= k >> 33;
+	k *= KQU(0xc4ceb9fe1a85ec53);
+	k ^= k >> 33;
+
+	return k;
+}
+
+static inline uint32_t
+hash_x86_32(const void *key, int len, uint32_t seed) {
+	const uint8_t *data = (const uint8_t *) key;
+	const int nblocks = len / 4;
+
+	uint32_t h1 = seed;
+
+	const uint32_t c1 = 0xcc9e2d51;
+	const uint32_t c2 = 0x1b873593;
+
+	/* body */
+	{
+		const uint32_t *blocks = (const uint32_t *) (data + nblocks*4);
+		int i;
+
+		for (i = -nblocks; i; i++) {
+			uint32_t k1 = hash_get_block_32(blocks, i);
+
+			k1 *= c1;
+			k1 = hash_rotl_32(k1, 15);
+			k1 *= c2;
+
+			h1 ^= k1;
+			h1 = hash_rotl_32(h1, 13);
+			h1 = h1*5 + 0xe6546b64;
+		}
+	}
+
+	/* tail */
+	{
+		const uint8_t *tail = (const uint8_t *) (data + nblocks*4);
+
+		uint32_t k1 = 0;
+
+		switch (len & 3) {
+		case 3: k1 ^= tail[2] << 16; JEMALLOC_FALLTHROUGH;
+		case 2: k1 ^= tail[1] << 8; JEMALLOC_FALLTHROUGH;
+		case 1: k1 ^= tail[0]; k1 *= c1; k1 = hash_rotl_32(k1, 15);
+			k1 *= c2; h1 ^= k1;
+		}
+	}
+
+	/* finalization */
+	h1 ^= len;
+
+	h1 = hash_fmix_32(h1);
+
+	return h1;
+}
+
+static inline void
+hash_x86_128(const void *key, const int len, uint32_t seed,
+    uint64_t r_out[2]) {
+	const uint8_t * data = (const uint8_t *) key;
+	const int nblocks = len / 16;
+
+	uint32_t h1 = seed;
+	uint32_t h2 = seed;
+	uint32_t h3 = seed;
+	uint32_t h4 = seed;
+
+	const uint32_t c1 = 0x239b961b;
+	const uint32_t c2 = 0xab0e9789;
+	const uint32_t c3 = 0x38b34ae5;
+	const uint32_t c4 = 0xa1e38b93;
+
+	/* body */
+	{
+		const uint32_t *blocks = (const uint32_t *) (data + nblocks*16);
+		int i;
+
+		for (i = -nblocks; i; i++) {
+			uint32_t k1 = hash_get_block_32(blocks, i*4 + 0);
+			uint32_t k2 = hash_get_block_32(blocks, i*4 + 1);
+			uint32_t k3 = hash_get_block_32(blocks, i*4 + 2);
+			uint32_t k4 = hash_get_block_32(blocks, i*4 + 3);
+
+			k1 *= c1; k1 = hash_rotl_32(k1, 15); k1 *= c2; h1 ^= k1;
+
+			h1 = hash_rotl_32(h1, 19); h1 += h2;
+			h1 = h1*5 + 0x561ccd1b;
+
+			k2 *= c2; k2 = hash_rotl_32(k2, 16); k2 *= c3; h2 ^= k2;
+
+			h2 = hash_rotl_32(h2, 17); h2 += h3;
+			h2 = h2*5 + 0x0bcaa747;
+
+			k3 *= c3; k3 = hash_rotl_32(k3, 17); k3 *= c4; h3 ^= k3;
+
+			h3 = hash_rotl_32(h3, 15); h3 += h4;
+			h3 = h3*5 + 0x96cd1c35;
+
+			k4 *= c4; k4 = hash_rotl_32(k4, 18); k4 *= c1; h4 ^= k4;
+
+			h4 = hash_rotl_32(h4, 13); h4 += h1;
+			h4 = h4*5 + 0x32ac3b17;
+		}
+	}
+
+	/* tail */
+	{
+		const uint8_t *tail = (const uint8_t *) (data + nblocks*16);
+		uint32_t k1 = 0;
+		uint32_t k2 = 0;
+		uint32_t k3 = 0;
+		uint32_t k4 = 0;
+
+		switch (len & 15) {
+		case 15: k4 ^= tail[14] << 16; JEMALLOC_FALLTHROUGH;
+		case 14: k4 ^= tail[13] << 8; JEMALLOC_FALLTHROUGH;
+		case 13: k4 ^= tail[12] << 0;
+			k4 *= c4; k4 = hash_rotl_32(k4, 18); k4 *= c1; h4 ^= k4;
+			JEMALLOC_FALLTHROUGH;
+		case 12: k3 ^= (uint32_t) tail[11] << 24; JEMALLOC_FALLTHROUGH;
+		case 11: k3 ^= tail[10] << 16; JEMALLOC_FALLTHROUGH;
+		case 10: k3 ^= tail[ 9] << 8; JEMALLOC_FALLTHROUGH;
+		case  9: k3 ^= tail[ 8] << 0;
+			k3 *= c3; k3 = hash_rotl_32(k3, 17); k3 *= c4; h3 ^= k3;
+			JEMALLOC_FALLTHROUGH;
+		case  8: k2 ^= (uint32_t) tail[ 7] << 24; JEMALLOC_FALLTHROUGH;
+		case  7: k2 ^= tail[ 6] << 16; JEMALLOC_FALLTHROUGH;
+		case  6: k2 ^= tail[ 5] << 8; JEMALLOC_FALLTHROUGH;
+		case  5: k2 ^= tail[ 4] << 0;
+			k2 *= c2; k2 = hash_rotl_32(k2, 16); k2 *= c3; h2 ^= k2;
+			JEMALLOC_FALLTHROUGH;
+		case  4: k1 ^= (uint32_t) tail[ 3] << 24; JEMALLOC_FALLTHROUGH;
+		case  3: k1 ^= tail[ 2] << 16; JEMALLOC_FALLTHROUGH;
+		case  2: k1 ^= tail[ 1] << 8; JEMALLOC_FALLTHROUGH;
+		case  1: k1 ^= tail[ 0] << 0;
+			k1 *= c1; k1 = hash_rotl_32(k1, 15); k1 *= c2; h1 ^= k1;
+			break;
+		}
+	}
+
+	/* finalization */
+	h1 ^= len; h2 ^= len; h3 ^= len; h4 ^= len;
+
+	h1 += h2; h1 += h3; h1 += h4;
+	h2 += h1; h3 += h1; h4 += h1;
+
+	h1 = hash_fmix_32(h1);
+	h2 = hash_fmix_32(h2);
+	h3 = hash_fmix_32(h3);
+	h4 = hash_fmix_32(h4);
+
+	h1 += h2; h1 += h3; h1 += h4;
+	h2 += h1; h3 += h1; h4 += h1;
+
+	r_out[0] = (((uint64_t) h2) << 32) | h1;
+	r_out[1] = (((uint64_t) h4) << 32) | h3;
+}
+
+static inline void
+hash_x64_128(const void *key, const int len, const uint32_t seed,
+    uint64_t r_out[2]) {
+	const uint8_t *data = (const uint8_t *) key;
+	const int nblocks = len / 16;
+
+	uint64_t h1 = seed;
+	uint64_t h2 = seed;
+
+	const uint64_t c1 = KQU(0x87c37b91114253d5);
+	const uint64_t c2 = KQU(0x4cf5ad432745937f);
+
+	/* body */
+	{
+		const uint64_t *blocks = (const uint64_t *) (data);
+		int i;
+
+		for (i = 0; i < nblocks; i++) {
+			uint64_t k1 = hash_get_block_64(blocks, i*2 + 0);
+			uint64_t k2 = hash_get_block_64(blocks, i*2 + 1);
+
+			k1 *= c1; k1 = hash_rotl_64(k1, 31); k1 *= c2; h1 ^= k1;
+
+			h1 = hash_rotl_64(h1, 27); h1 += h2;
+			h1 = h1*5 + 0x52dce729;
+
+			k2 *= c2; k2 = hash_rotl_64(k2, 33); k2 *= c1; h2 ^= k2;
+
+			h2 = hash_rotl_64(h2, 31); h2 += h1;
+			h2 = h2*5 + 0x38495ab5;
+		}
+	}
+
+	/* tail */
+	{
+		const uint8_t *tail = (const uint8_t*)(data + nblocks*16);
+		uint64_t k1 = 0;
+		uint64_t k2 = 0;
+
+		switch (len & 15) {
+		case 15: k2 ^= ((uint64_t)(tail[14])) << 48; JEMALLOC_FALLTHROUGH;
+		case 14: k2 ^= ((uint64_t)(tail[13])) << 40; JEMALLOC_FALLTHROUGH;
+		case 13: k2 ^= ((uint64_t)(tail[12])) << 32; JEMALLOC_FALLTHROUGH;
+		case 12: k2 ^= ((uint64_t)(tail[11])) << 24; JEMALLOC_FALLTHROUGH;
+		case 11: k2 ^= ((uint64_t)(tail[10])) << 16; JEMALLOC_FALLTHROUGH;
+		case 10: k2 ^= ((uint64_t)(tail[ 9])) << 8;  JEMALLOC_FALLTHROUGH;
+		case  9: k2 ^= ((uint64_t)(tail[ 8])) << 0;
+			k2 *= c2; k2 = hash_rotl_64(k2, 33); k2 *= c1; h2 ^= k2;
+			JEMALLOC_FALLTHROUGH;
+		case  8: k1 ^= ((uint64_t)(tail[ 7])) << 56; JEMALLOC_FALLTHROUGH;
+		case  7: k1 ^= ((uint64_t)(tail[ 6])) << 48; JEMALLOC_FALLTHROUGH;
+		case  6: k1 ^= ((uint64_t)(tail[ 5])) << 40; JEMALLOC_FALLTHROUGH;
+		case  5: k1 ^= ((uint64_t)(tail[ 4])) << 32; JEMALLOC_FALLTHROUGH;
+		case  4: k1 ^= ((uint64_t)(tail[ 3])) << 24; JEMALLOC_FALLTHROUGH;
+		case  3: k1 ^= ((uint64_t)(tail[ 2])) << 16; JEMALLOC_FALLTHROUGH;
+		case  2: k1 ^= ((uint64_t)(tail[ 1])) << 8;  JEMALLOC_FALLTHROUGH;
+		case  1: k1 ^= ((uint64_t)(tail[ 0])) << 0;
+			k1 *= c1; k1 = hash_rotl_64(k1, 31); k1 *= c2; h1 ^= k1;
+			break;
+		}
+	}
+
+	/* finalization */
+	h1 ^= len; h2 ^= len;
+
+	h1 += h2;
+	h2 += h1;
+
+	h1 = hash_fmix_64(h1);
+	h2 = hash_fmix_64(h2);
+
+	h1 += h2;
+	h2 += h1;
+
+	r_out[0] = h1;
+	r_out[1] = h2;
+}
+
+/******************************************************************************/
+/* API. */
+static inline void
+hash(const void *key, size_t len, const uint32_t seed, size_t r_hash[2]) {
+	assert(len <= INT_MAX); /* Unfortunate implementation limitation. */
+
+#if (LG_SIZEOF_PTR == 3 && !defined(JEMALLOC_BIG_ENDIAN))
+	hash_x64_128(key, (int)len, seed, (uint64_t *)r_hash);
+#else
+	{
+		uint64_t hashes[2];
+		hash_x86_128(key, (int)len, seed, hashes);
+		r_hash[0] = (size_t)hashes[0];
+		r_hash[1] = (size_t)hashes[1];
+	}
+#endif
+}
+
+#endif /* JEMALLOC_INTERNAL_HASH_H */

BeefRT/JEMalloc/include/jemalloc/internal/hook.h

@@ -0,0 +1,163 @@
+#ifndef JEMALLOC_INTERNAL_HOOK_H
+#define JEMALLOC_INTERNAL_HOOK_H
+
+#include "jemalloc/internal/tsd.h"
+
+/*
+ * This API is *extremely* experimental, and may get ripped out, changed in API-
+ * and ABI-incompatible ways, be insufficiently or incorrectly documented, etc.
+ *
+ * It allows hooking the stateful parts of the API to see changes as they
+ * happen.
+ *
+ * Allocation hooks are called after the allocation is done, free hooks are
+ * called before the free is done, and expand hooks are called after the
+ * allocation is expanded.
+ *
+ * For realloc and rallocx, if the expansion happens in place, the expansion
+ * hook is called.  If it is moved, then the alloc hook is called on the new
+ * location, and then the free hook is called on the old location (i.e. both
+ * hooks are invoked in between the alloc and the dalloc).
+ *
+ * If we return NULL from OOM, then usize might not be trustworthy.  Calling
+ * realloc(NULL, size) only calls the alloc hook, and calling realloc(ptr, 0)
+ * only calls the free hook.  (Calling realloc(NULL, 0) is treated as malloc(0),
+ * and only calls the alloc hook).
+ *
+ * Reentrancy:
+ *   Reentrancy is guarded against from within the hook implementation.  If you
+ *   call allocator functions from within a hook, the hooks will not be invoked
+ *   again.
+ * Threading:
+ *   The installation of a hook synchronizes with all its uses.  If you can
+ *   prove the installation of a hook happens-before a jemalloc entry point,
+ *   then the hook will get invoked (unless there's a racing removal).
+ *
+ *   Hook insertion appears to be atomic at a per-thread level (i.e. if a thread
+ *   allocates and has the alloc hook invoked, then a subsequent free on the
+ *   same thread will also have the free hook invoked).
+ *
+ *   The *removal* of a hook does *not* block until all threads are done with
+ *   the hook.  Hook authors have to be resilient to this, and need some
+ *   out-of-band mechanism for cleaning up any dynamically allocated memory
+ *   associated with their hook.
+ * Ordering:
+ *   Order of hook execution is unspecified, and may be different than insertion
+ *   order.
+ */
+
+#define HOOK_MAX 4
+
+enum hook_alloc_e {
+	hook_alloc_malloc,
+	hook_alloc_posix_memalign,
+	hook_alloc_aligned_alloc,
+	hook_alloc_calloc,
+	hook_alloc_memalign,
+	hook_alloc_valloc,
+	hook_alloc_mallocx,
+
+	/* The reallocating functions have both alloc and dalloc variants */
+	hook_alloc_realloc,
+	hook_alloc_rallocx,
+};
+/*
+ * We put the enum typedef after the enum, since this file may get included by
+ * jemalloc_cpp.cpp, and C++ disallows enum forward declarations.
+ */
+typedef enum hook_alloc_e hook_alloc_t;
+
+enum hook_dalloc_e {
+	hook_dalloc_free,
+	hook_dalloc_dallocx,
+	hook_dalloc_sdallocx,
+
+	/*
+	 * The dalloc halves of reallocation (not called if in-place expansion
+	 * happens).
+	 */
+	hook_dalloc_realloc,
+	hook_dalloc_rallocx,
+};
+typedef enum hook_dalloc_e hook_dalloc_t;
+
+
+enum hook_expand_e {
+	hook_expand_realloc,
+	hook_expand_rallocx,
+	hook_expand_xallocx,
+};
+typedef enum hook_expand_e hook_expand_t;
+
+typedef void (*hook_alloc)(
+    void *extra, hook_alloc_t type, void *result, uintptr_t result_raw,
+    uintptr_t args_raw[3]);
+
+typedef void (*hook_dalloc)(
+    void *extra, hook_dalloc_t type, void *address, uintptr_t args_raw[3]);
+
+typedef void (*hook_expand)(
+    void *extra, hook_expand_t type, void *address, size_t old_usize,
+    size_t new_usize, uintptr_t result_raw, uintptr_t args_raw[4]);
+
+typedef struct hooks_s hooks_t;
+struct hooks_s {
+	hook_alloc alloc_hook;
+	hook_dalloc dalloc_hook;
+	hook_expand expand_hook;
+	void *extra;
+};
+
+/*
+ * Begin implementation details; everything above this point might one day live
+ * in a public API.  Everything below this point never will.
+ */
+
+/*
+ * The realloc pathways haven't gotten any refactoring love in a while, and it's
+ * fairly difficult to pass information from the entry point to the hooks.  We
+ * put the informaiton the hooks will need into a struct to encapsulate
+ * everything.
+ *
+ * Much of these pathways are force-inlined, so that the compiler can avoid
+ * materializing this struct until we hit an extern arena function.  For fairly
+ * goofy reasons, *many* of the realloc paths hit an extern arena function.
+ * These paths are cold enough that it doesn't matter; eventually, we should
+ * rewrite the realloc code to make the expand-in-place and the
+ * free-then-realloc paths more orthogonal, at which point we don't need to
+ * spread the hook logic all over the place.
+ */
+typedef struct hook_ralloc_args_s hook_ralloc_args_t;
+struct hook_ralloc_args_s {
+	/* I.e. as opposed to rallocx. */
+	bool is_realloc;
+	/*
+	 * The expand hook takes 4 arguments, even if only 3 are actually used;
+	 * we add an extra one in case the user decides to memcpy without
+	 * looking too closely at the hooked function.
+	 */
+	uintptr_t args[4];
+};
+
+/*
+ * Returns an opaque handle to be used when removing the hook.  NULL means that
+ * we couldn't install the hook.
+ */
+bool hook_boot();
+
+void *hook_install(tsdn_t *tsdn, hooks_t *hooks);
+/* Uninstalls the hook with the handle previously returned from hook_install. */
+void hook_remove(tsdn_t *tsdn, void *opaque);
+
+/* Hooks */
+
+void hook_invoke_alloc(hook_alloc_t type, void *result, uintptr_t result_raw,
+    uintptr_t args_raw[3]);
+
+void hook_invoke_dalloc(hook_dalloc_t type, void *address,
+    uintptr_t args_raw[3]);
+
+void hook_invoke_expand(hook_expand_t type, void *address, size_t old_usize,
+    size_t new_usize, uintptr_t result_raw, uintptr_t args_raw[4]);
+
+#endif /* JEMALLOC_INTERNAL_HOOK_H */

BeefRT/JEMalloc/include/jemalloc/internal/hpa.h

@@ -0,0 +1,182 @@
+#ifndef JEMALLOC_INTERNAL_HPA_H
+#define JEMALLOC_INTERNAL_HPA_H
+
+#include "jemalloc/internal/exp_grow.h"
+#include "jemalloc/internal/hpa_hooks.h"
+#include "jemalloc/internal/hpa_opts.h"
+#include "jemalloc/internal/pai.h"
+#include "jemalloc/internal/psset.h"
+
+typedef struct hpa_central_s hpa_central_t;
+struct hpa_central_s {
+	/*
+	 * The mutex guarding most of the operations on the central data
+	 * structure.
+	 */
+	malloc_mutex_t mtx;
+	/*
+	 * Guards expansion of eden.  We separate this from the regular mutex so
+	 * that cheaper operations can still continue while we're doing the OS
+	 * call.
+	 */
+	malloc_mutex_t grow_mtx;
+	/*
+	 * Either NULL (if empty), or some integer multiple of a
+	 * hugepage-aligned number of hugepages.  We carve them off one at a
+	 * time to satisfy new pageslab requests.
+	 *
+	 * Guarded by grow_mtx.
+	 */
+	void *eden;
+	size_t eden_len;
+	/* Source for metadata. */
+	base_t *base;
+	/* Number of grow operations done on this hpa_central_t. */
+	uint64_t age_counter;
+
+	/* The HPA hooks. */
+	hpa_hooks_t hooks;
+};
+
+typedef struct hpa_shard_nonderived_stats_s hpa_shard_nonderived_stats_t;
+struct hpa_shard_nonderived_stats_s {
+	/*
+	 * The number of times we've purged within a hugepage.
+	 *
+	 * Guarded by mtx.
+	 */
+	uint64_t npurge_passes;
+	/*
+	 * The number of individual purge calls we perform (which should always
+	 * be bigger than npurge_passes, since each pass purges at least one
+	 * extent within a hugepage.
+	 *
+	 * Guarded by mtx.
+	 */
+	uint64_t npurges;
+
+	/*
+	 * The number of times we've hugified a pageslab.
+	 *
+	 * Guarded by mtx.
+	 */
+	uint64_t nhugifies;
+	/*
+	 * The number of times we've dehugified a pageslab.
+	 *
+	 * Guarded by mtx.
+	 */
+	uint64_t ndehugifies;
+};
+
+/* Completely derived; only used by CTL. */
+typedef struct hpa_shard_stats_s hpa_shard_stats_t;
+struct hpa_shard_stats_s {
+	psset_stats_t psset_stats;
+	hpa_shard_nonderived_stats_t nonderived_stats;
+};
+
+typedef struct hpa_shard_s hpa_shard_t;
+struct hpa_shard_s {
+	/*
+	 * pai must be the first member; we cast from a pointer to it to a
+	 * pointer to the hpa_shard_t.
+	 */
+	pai_t pai;
+
+	/* The central allocator we get our hugepages from. */
+	hpa_central_t *central;
+	/* Protects most of this shard's state. */
+	malloc_mutex_t mtx;
+	/*
+	 * Guards the shard's access to the central allocator (preventing
+	 * multiple threads operating on this shard from accessing the central
+	 * allocator).
+	 */
+	malloc_mutex_t grow_mtx;
+	/* The base metadata allocator. */
+	base_t *base;
+
+	/*
+	 * This edata cache is the one we use when allocating a small extent
+	 * from a pageslab.  The pageslab itself comes from the centralized
+	 * allocator, and so will use its edata_cache.
+	 */
+	edata_cache_fast_t ecf;
+
+	psset_t psset;
+
+	/*
+	 * How many grow operations have occurred.
+	 *
+	 * Guarded by grow_mtx.
+	 */
+	uint64_t age_counter;
+
+	/* The arena ind we're associated with. */
+	unsigned ind;
+
+	/*
+	 * Our emap.  This is just a cache of the emap pointer in the associated
+	 * hpa_central.
+	 */
+	emap_t *emap;
+
+	/* The configuration choices for this hpa shard. */
+	hpa_shard_opts_t opts;
+
+	/*
+	 * How many pages have we started but not yet finished purging in this
+	 * hpa shard.
+	 */
+	size_t npending_purge;
+
+	/*
+	 * Those stats which are copied directly into the CTL-centric hpa shard
+	 * stats.
+	 */
+	hpa_shard_nonderived_stats_t stats;
+
+	/*
+	 * Last time we performed purge on this shard.
+	 */
+	nstime_t last_purge;
+};
+
+/*
+ * Whether or not the HPA can be used given the current configuration.  This is
+ * is not necessarily a guarantee that it backs its allocations by hugepages,
+ * just that it can function properly given the system it's running on.
+ */
+bool hpa_supported();
+bool hpa_central_init(hpa_central_t *central, base_t *base, const hpa_hooks_t *hooks);
+bool hpa_shard_init(hpa_shard_t *shard, hpa_central_t *central, emap_t *emap,
+    base_t *base, edata_cache_t *edata_cache, unsigned ind,
+    const hpa_shard_opts_t *opts);
+
+void hpa_shard_stats_accum(hpa_shard_stats_t *dst, hpa_shard_stats_t *src);
+void hpa_shard_stats_merge(tsdn_t *tsdn, hpa_shard_t *shard,
+    hpa_shard_stats_t *dst);
+
+/*
+ * Notify the shard that we won't use it for allocations much longer.  Due to
+ * the possibility of races, we don't actually prevent allocations; just flush
+ * and disable the embedded edata_cache_small.
+ */
+void hpa_shard_disable(tsdn_t *tsdn, hpa_shard_t *shard);
+void hpa_shard_destroy(tsdn_t *tsdn, hpa_shard_t *shard);
+
+void hpa_shard_set_deferral_allowed(tsdn_t *tsdn, hpa_shard_t *shard,
+    bool deferral_allowed);
+void hpa_shard_do_deferred_work(tsdn_t *tsdn, hpa_shard_t *shard);
+
+/*
+ * We share the fork ordering with the PA and arena prefork handling; that's why
+ * these are 3 and 4 rather than 0 and 1.
+ */
+void hpa_shard_prefork3(tsdn_t *tsdn, hpa_shard_t *shard);
+void hpa_shard_prefork4(tsdn_t *tsdn, hpa_shard_t *shard);
+void hpa_shard_postfork_parent(tsdn_t *tsdn, hpa_shard_t *shard);
+void hpa_shard_postfork_child(tsdn_t *tsdn, hpa_shard_t *shard);
+
+#endif /* JEMALLOC_INTERNAL_HPA_H */

BeefRT/JEMalloc/include/jemalloc/internal/hpa_hooks.h

@@ -0,0 +1,17 @@
+#ifndef JEMALLOC_INTERNAL_HPA_HOOKS_H
+#define JEMALLOC_INTERNAL_HPA_HOOKS_H
+
+typedef struct hpa_hooks_s hpa_hooks_t;
+struct hpa_hooks_s {
+	void *(*map)(size_t size);
+	void (*unmap)(void *ptr, size_t size);
+	void (*purge)(void *ptr, size_t size);
+	void (*hugify)(void *ptr, size_t size);
+	void (*dehugify)(void *ptr, size_t size);
+	void (*curtime)(nstime_t *r_time, bool first_reading);
+	uint64_t (*ms_since)(nstime_t *r_time);
+};
+
+extern hpa_hooks_t hpa_hooks_default;
+
+#endif /* JEMALLOC_INTERNAL_HPA_HOOKS_H */

BeefRT/JEMalloc/include/jemalloc/internal/hpa_opts.h

@@ -0,0 +1,74 @@
+#ifndef JEMALLOC_INTERNAL_HPA_OPTS_H
+#define JEMALLOC_INTERNAL_HPA_OPTS_H
+
+#include "jemalloc/internal/fxp.h"
+
+/*
+ * This file is morally part of hpa.h, but is split out for header-ordering
+ * reasons.
+ */
+
+typedef struct hpa_shard_opts_s hpa_shard_opts_t;
+struct hpa_shard_opts_s {
+	/*
+	 * The largest size we'll allocate out of the shard.  For those
+	 * allocations refused, the caller (in practice, the PA module) will
+	 * fall back to the more general (for now) PAC, which can always handle
+	 * any allocation request.
+	 */
+	size_t slab_max_alloc;
+
+	/*
+	 * When the number of active bytes in a hugepage is >=
+	 * hugification_threshold, we force hugify it.
+	 */
+	size_t hugification_threshold;
+
+	/*
+	 * The HPA purges whenever the number of pages exceeds dirty_mult *
+	 * active_pages.  This may be set to (fxp_t)-1 to disable purging.
+	 */
+	fxp_t dirty_mult;
+
+	/*
+	 * Whether or not the PAI methods are allowed to defer work to a
+	 * subsequent hpa_shard_do_deferred_work() call.  Practically, this
+	 * corresponds to background threads being enabled.  We track this
+	 * ourselves for encapsulation purposes.
+	 */
+	bool deferral_allowed;
+
+	/*
+	 * How long a hugepage has to be a hugification candidate before it will
+	 * actually get hugified.
+	 */
+	uint64_t hugify_delay_ms;
+
+	/*
+	 * Minimum amount of time between purges.
+	 */
+	uint64_t min_purge_interval_ms;
+};
+
+#define HPA_SHARD_OPTS_DEFAULT {					\
+	/* slab_max_alloc */						\
+	64 * 1024,							\
+	/* hugification_threshold */					\
+	HUGEPAGE * 95 / 100,						\
+	/* dirty_mult */						\
+	FXP_INIT_PERCENT(25),						\
+	/*								\
+	 * deferral_allowed						\
+	 * 								\
+	 * Really, this is always set by the arena during creation	\
+	 * or by an hpa_shard_set_deferral_allowed call, so the value	\
+	 * we put here doesn't matter.					\
+	 */								\
+	false,								\
+	/* hugify_delay_ms */						\
+	10 * 1000,							\
+	/* min_purge_interval_ms */					\
+	5 * 1000							\
+}
+
+#endif /* JEMALLOC_INTERNAL_HPA_OPTS_H */

BeefRT/JEMalloc/include/jemalloc/internal/hpdata.h

@@ -0,0 +1,413 @@
+#ifndef JEMALLOC_INTERNAL_HPDATA_H
+#define JEMALLOC_INTERNAL_HPDATA_H
+
+#include "jemalloc/internal/fb.h"
+#include "jemalloc/internal/ph.h"
+#include "jemalloc/internal/ql.h"
+#include "jemalloc/internal/typed_list.h"
+
+/*
+ * The metadata representation we use for extents in hugepages.  While the PAC
+ * uses the edata_t to represent both active and inactive extents, the HP only
+ * uses the edata_t for active ones; instead, inactive extent state is tracked
+ * within hpdata associated with the enclosing hugepage-sized, hugepage-aligned
+ * region of virtual address space.
+ *
+ * An hpdata need not be "truly" backed by a hugepage (which is not necessarily
+ * an observable property of any given region of address space).  It's just
+ * hugepage-sized and hugepage-aligned; it's *potentially* huge.
+ */
+typedef struct hpdata_s hpdata_t;
+ph_structs(hpdata_age_heap, hpdata_t);
+struct hpdata_s {
+	/*
+	 * We likewise follow the edata convention of mangling names and forcing
+	 * the use of accessors -- this lets us add some consistency checks on
+	 * access.
+	 */
+
+	/*
+	 * The address of the hugepage in question.  This can't be named h_addr,
+	 * since that conflicts with a macro defined in Windows headers.
+	 */
+	void *h_address;
+	/* Its age (measured in psset operations). */
+	uint64_t h_age;
+	/* Whether or not we think the hugepage is mapped that way by the OS. */
+	bool h_huge;
+
+	/*
+	 * For some properties, we keep parallel sets of bools; h_foo_allowed
+	 * and h_in_psset_foo_container.  This is a decoupling mechanism to
+	 * avoid bothering the hpa (which manages policies) from the psset
+	 * (which is the mechanism used to enforce those policies).  This allows
+	 * all the container management logic to live in one place, without the
+	 * HPA needing to know or care how that happens.
+	 */
+
+	/*
+	 * Whether or not the hpdata is allowed to be used to serve allocations,
+	 * and whether or not the psset is currently tracking it as such.
+	 */
+	bool h_alloc_allowed;
+	bool h_in_psset_alloc_container;
+
+	/*
+	 * The same, but with purging.  There's no corresponding
+	 * h_in_psset_purge_container, because the psset (currently) always
+	 * removes hpdatas from their containers during updates (to implement
+	 * LRU for purging).
+	 */
+	bool h_purge_allowed;
+
+	/* And with hugifying. */
+	bool h_hugify_allowed;
+	/* When we became a hugification candidate. */
+	nstime_t h_time_hugify_allowed;
+	bool h_in_psset_hugify_container;
+
+	/* Whether or not a purge or hugify is currently happening. */
+	bool h_mid_purge;
+	bool h_mid_hugify;
+
+	/*
+	 * Whether or not the hpdata is being updated in the psset (i.e. if
+	 * there has been a psset_update_begin call issued without a matching
+	 * psset_update_end call).  Eventually this will expand to other types
+	 * of updates.
+	 */
+	bool h_updating;
+
+	/* Whether or not the hpdata is in a psset. */
+	bool h_in_psset;
+
+	union {
+		/* When nonempty (and also nonfull), used by the psset bins. */
+		hpdata_age_heap_link_t age_link;
+		/*
+		 * When empty (or not corresponding to any hugepage), list
+		 * linkage.
+		 */
+		ql_elm(hpdata_t) ql_link_empty;
+	};
+
+	/*
+	 * Linkage for the psset to track candidates for purging and hugifying.
+	 */
+	ql_elm(hpdata_t) ql_link_purge;
+	ql_elm(hpdata_t) ql_link_hugify;
+
+	/* The length of the largest contiguous sequence of inactive pages. */
+	size_t h_longest_free_range;
+
+	/* Number of active pages. */
+	size_t h_nactive;
+
+	/* A bitmap with bits set in the active pages. */
+	fb_group_t active_pages[FB_NGROUPS(HUGEPAGE_PAGES)];
+
+	/*
+	 * Number of dirty or active pages, and a bitmap tracking them.  One
+	 * way to think of this is as which pages are dirty from the OS's
+	 * perspective.
+	 */
+	size_t h_ntouched;
+
+	/* The touched pages (using the same definition as above). */
+	fb_group_t touched_pages[FB_NGROUPS(HUGEPAGE_PAGES)];
+};
+
+TYPED_LIST(hpdata_empty_list, hpdata_t, ql_link_empty)
+TYPED_LIST(hpdata_purge_list, hpdata_t, ql_link_purge)
+TYPED_LIST(hpdata_hugify_list, hpdata_t, ql_link_hugify)
+
+ph_proto(, hpdata_age_heap, hpdata_t);
+
+static inline void *
+hpdata_addr_get(const hpdata_t *hpdata) {
+	return hpdata->h_address;
+}
+
+static inline void
+hpdata_addr_set(hpdata_t *hpdata, void *addr) {
+	assert(HUGEPAGE_ADDR2BASE(addr) == addr);
+	hpdata->h_address = addr;
+}
+
+static inline uint64_t
+hpdata_age_get(const hpdata_t *hpdata) {
+	return hpdata->h_age;
+}
+
+static inline void
+hpdata_age_set(hpdata_t *hpdata, uint64_t age) {
+	hpdata->h_age = age;
+}
+
+static inline bool
+hpdata_huge_get(const hpdata_t *hpdata) {
+	return hpdata->h_huge;
+}
+
+static inline bool
+hpdata_alloc_allowed_get(const hpdata_t *hpdata) {
+	return hpdata->h_alloc_allowed;
+}
+
+static inline void
+hpdata_alloc_allowed_set(hpdata_t *hpdata, bool alloc_allowed) {
+	hpdata->h_alloc_allowed = alloc_allowed;
+}
+
+static inline bool
+hpdata_in_psset_alloc_container_get(const hpdata_t *hpdata) {
+	return hpdata->h_in_psset_alloc_container;
+}
+
+static inline void
+hpdata_in_psset_alloc_container_set(hpdata_t *hpdata, bool in_container) {
+	assert(in_container != hpdata->h_in_psset_alloc_container);
+	hpdata->h_in_psset_alloc_container = in_container;
+}
+
+static inline bool
+hpdata_purge_allowed_get(const hpdata_t *hpdata) {
+	return hpdata->h_purge_allowed;
+}
+
+static inline void
+hpdata_purge_allowed_set(hpdata_t *hpdata, bool purge_allowed) {
+       assert(purge_allowed == false || !hpdata->h_mid_purge);
+       hpdata->h_purge_allowed = purge_allowed;
+}
+
+static inline bool
+hpdata_hugify_allowed_get(const hpdata_t *hpdata) {
+	return hpdata->h_hugify_allowed;
+}
+
+static inline void
+hpdata_allow_hugify(hpdata_t *hpdata, nstime_t now) {
+	assert(!hpdata->h_mid_hugify);
+	hpdata->h_hugify_allowed = true;
+	hpdata->h_time_hugify_allowed = now;
+}
+
+static inline nstime_t
+hpdata_time_hugify_allowed(hpdata_t *hpdata) {
+	return hpdata->h_time_hugify_allowed;
+}
+
+static inline void
+hpdata_disallow_hugify(hpdata_t *hpdata) {
+	hpdata->h_hugify_allowed = false;
+}
+
+static inline bool
+hpdata_in_psset_hugify_container_get(const hpdata_t *hpdata) {
+	return hpdata->h_in_psset_hugify_container;
+}
+
+static inline void
+hpdata_in_psset_hugify_container_set(hpdata_t *hpdata, bool in_container) {
+	assert(in_container != hpdata->h_in_psset_hugify_container);
+	hpdata->h_in_psset_hugify_container = in_container;
+}
+
+static inline bool
+hpdata_mid_purge_get(const hpdata_t *hpdata) {
+	return hpdata->h_mid_purge;
+}
+
+static inline void
+hpdata_mid_purge_set(hpdata_t *hpdata, bool mid_purge) {
+	assert(mid_purge != hpdata->h_mid_purge);
+	hpdata->h_mid_purge = mid_purge;
+}
+
+static inline bool
+hpdata_mid_hugify_get(const hpdata_t *hpdata) {
+	return hpdata->h_mid_hugify;
+}
+
+static inline void
+hpdata_mid_hugify_set(hpdata_t *hpdata, bool mid_hugify) {
+	assert(mid_hugify != hpdata->h_mid_hugify);
+	hpdata->h_mid_hugify = mid_hugify;
+}
+
+static inline bool
+hpdata_changing_state_get(const hpdata_t *hpdata) {
+	return hpdata->h_mid_purge || hpdata->h_mid_hugify;
+}
+
+
+static inline bool
+hpdata_updating_get(const hpdata_t *hpdata) {
+	return hpdata->h_updating;
+}
+
+static inline void
+hpdata_updating_set(hpdata_t *hpdata, bool updating) {
+	assert(updating != hpdata->h_updating);
+	hpdata->h_updating = updating;
+}
+
+static inline bool
+hpdata_in_psset_get(const hpdata_t *hpdata) {
+	return hpdata->h_in_psset;
+}
+
+static inline void
+hpdata_in_psset_set(hpdata_t *hpdata, bool in_psset) {
+	assert(in_psset != hpdata->h_in_psset);
+	hpdata->h_in_psset = in_psset;
+}
+
+static inline size_t
+hpdata_longest_free_range_get(const hpdata_t *hpdata) {
+	return hpdata->h_longest_free_range;
+}
+
+static inline void
+hpdata_longest_free_range_set(hpdata_t *hpdata, size_t longest_free_range) {
+	assert(longest_free_range <= HUGEPAGE_PAGES);
+	hpdata->h_longest_free_range = longest_free_range;
+}
+
+static inline size_t
+hpdata_nactive_get(hpdata_t *hpdata) {
+	return hpdata->h_nactive;
+}
+
+static inline size_t
+hpdata_ntouched_get(hpdata_t *hpdata) {
+	return hpdata->h_ntouched;
+}
+
+static inline size_t
+hpdata_ndirty_get(hpdata_t *hpdata) {
+	return hpdata->h_ntouched - hpdata->h_nactive;
+}
+
+static inline size_t
+hpdata_nretained_get(hpdata_t *hpdata) {
+	return HUGEPAGE_PAGES - hpdata->h_ntouched;
+}
+
+static inline void
+hpdata_assert_empty(hpdata_t *hpdata) {
+	assert(fb_empty(hpdata->active_pages, HUGEPAGE_PAGES));
+	assert(hpdata->h_nactive == 0);
+}
+
+/*
+ * Only used in tests, and in hpdata_assert_consistent, below.  Verifies some
+ * consistency properties of the hpdata (e.g. that cached counts of page stats
+ * match computed ones).
+ */
+static inline bool
+hpdata_consistent(hpdata_t *hpdata) {
+	if(fb_urange_longest(hpdata->active_pages, HUGEPAGE_PAGES)
+	    != hpdata_longest_free_range_get(hpdata)) {
+		return false;
+	}
+	if (fb_scount(hpdata->active_pages, HUGEPAGE_PAGES, 0, HUGEPAGE_PAGES)
+	    != hpdata->h_nactive) {
+		return false;
+	}
+	if (fb_scount(hpdata->touched_pages, HUGEPAGE_PAGES, 0, HUGEPAGE_PAGES)
+	    != hpdata->h_ntouched) {
+		return false;
+	}
+	if (hpdata->h_ntouched < hpdata->h_nactive) {
+		return false;
+	}
+	if (hpdata->h_huge && hpdata->h_ntouched != HUGEPAGE_PAGES) {
+		return false;
+	}
+	if (hpdata_changing_state_get(hpdata)
+	    && ((hpdata->h_purge_allowed) || hpdata->h_hugify_allowed)) {
+		return false;
+	}
+	if (hpdata_hugify_allowed_get(hpdata)
+	    != hpdata_in_psset_hugify_container_get(hpdata)) {
+		return false;
+	}
+	return true;
+}
+
+static inline void
+hpdata_assert_consistent(hpdata_t *hpdata) {
+	assert(hpdata_consistent(hpdata));
+}
+
+static inline bool
+hpdata_empty(hpdata_t *hpdata) {
+	return hpdata->h_nactive == 0;
+}
+
+static inline bool
+hpdata_full(hpdata_t *hpdata) {
+	return hpdata->h_nactive == HUGEPAGE_PAGES;
+}
+
+void hpdata_init(hpdata_t *hpdata, void *addr, uint64_t age);
+
+/*
+ * Given an hpdata which can serve an allocation request, pick and reserve an
+ * offset within that allocation.
+ */
+void *hpdata_reserve_alloc(hpdata_t *hpdata, size_t sz);
+void hpdata_unreserve(hpdata_t *hpdata, void *begin, size_t sz);
+
+/*
+ * The hpdata_purge_prepare_t allows grabbing the metadata required to purge
+ * subranges of a hugepage while holding a lock, drop the lock during the actual
+ * purging of them, and reacquire it to update the metadata again.
+ */
+typedef struct hpdata_purge_state_s hpdata_purge_state_t;
+struct hpdata_purge_state_s {
+	size_t npurged;
+	size_t ndirty_to_purge;
+	fb_group_t to_purge[FB_NGROUPS(HUGEPAGE_PAGES)];
+	size_t next_purge_search_begin;
+};
+
+/*
+ * Initializes purge state.  The access to hpdata must be externally
+ * synchronized with other hpdata_* calls.
+ *
+ * You can tell whether or not a thread is purging or hugifying a given hpdata
+ * via hpdata_changing_state_get(hpdata).  Racing hugification or purging
+ * operations aren't allowed.
+ *
+ * Once you begin purging, you have to follow through and call hpdata_purge_next
+ * until you're done, and then end.  Allocating out of an hpdata undergoing
+ * purging is not allowed.
+ *
+ * Returns the number of dirty pages that will be purged.
+ */
+size_t hpdata_purge_begin(hpdata_t *hpdata, hpdata_purge_state_t *purge_state);
+
+/*
+ * If there are more extents to purge, sets *r_purge_addr and *r_purge_size to
+ * true, and returns true.  Otherwise, returns false to indicate that we're
+ * done.
+ *
+ * This requires exclusive access to the purge state, but *not* to the hpdata.
+ * In particular, unreserve calls are allowed while purging (i.e. you can dalloc
+ * into one part of the hpdata while purging a different part).
+ */
+bool hpdata_purge_next(hpdata_t *hpdata, hpdata_purge_state_t *purge_state,
+    void **r_purge_addr, size_t *r_purge_size);
+/*
+ * Updates the hpdata metadata after all purging is done.  Needs external
+ * synchronization.
+ */
+void hpdata_purge_end(hpdata_t *hpdata, hpdata_purge_state_t *purge_state);
+
+void hpdata_hugify(hpdata_t *hpdata);
+void hpdata_dehugify(hpdata_t *hpdata);
+
+#endif /* JEMALLOC_INTERNAL_HPDATA_H */

BeefRT/JEMalloc/include/jemalloc/internal/inspect.h

@@ -0,0 +1,40 @@
+#ifndef JEMALLOC_INTERNAL_INSPECT_H
+#define JEMALLOC_INTERNAL_INSPECT_H
+
+/*
+ * This module contains the heap introspection capabilities.  For now they are
+ * exposed purely through mallctl APIs in the experimental namespace, but this
+ * may change over time.
+ */
+
+/*
+ * The following two structs are for experimental purposes. See
+ * experimental_utilization_query_ctl and
+ * experimental_utilization_batch_query_ctl in src/ctl.c.
+ */
+typedef struct inspect_extent_util_stats_s inspect_extent_util_stats_t;
+struct inspect_extent_util_stats_s {
+	size_t nfree;
+	size_t nregs;
+	size_t size;
+};
+
+typedef struct inspect_extent_util_stats_verbose_s
+    inspect_extent_util_stats_verbose_t;
+
+struct inspect_extent_util_stats_verbose_s {
+	void *slabcur_addr;
+	size_t nfree;
+	size_t nregs;
+	size_t size;
+	size_t bin_nfree;
+	size_t bin_nregs;
+};
+
+void inspect_extent_util_stats_get(tsdn_t *tsdn, const void *ptr,
+    size_t *nfree, size_t *nregs, size_t *size);
+void inspect_extent_util_stats_verbose_get(tsdn_t *tsdn, const void *ptr,
+    size_t *nfree, size_t *nregs, size_t *size,
+    size_t *bin_nfree, size_t *bin_nregs, void **slabcur_addr);
+
+#endif /* JEMALLOC_INTERNAL_INSPECT_H */

BeefRT/JEMalloc/include/jemalloc/internal/jemalloc_internal_decls.h

@@ -0,0 +1,108 @@
+#ifndef JEMALLOC_INTERNAL_DECLS_H
+#define JEMALLOC_INTERNAL_DECLS_H
+
+#include <math.h>
+#ifdef _WIN32
+#  include <windows.h>
+#  include "msvc_compat/windows_extra.h"
+#  include "msvc_compat/strings.h"
+#  ifdef _WIN64
+#    if LG_VADDR <= 32
+#      error Generate the headers using x64 vcargs
+#    endif
+#  else
+#    if LG_VADDR > 32
+#      undef LG_VADDR
+#      define LG_VADDR 32
+#    endif
+#  endif
+#else
+#  include <sys/param.h>
+#  include <sys/mman.h>
+#  if !defined(__pnacl__) && !defined(__native_client__)
+#    include <sys/syscall.h>
+#    if !defined(SYS_write) && defined(__NR_write)
+#      define SYS_write __NR_write
+#    endif
+#    if defined(SYS_open) && defined(__aarch64__)
+       /* Android headers may define SYS_open to __NR_open even though
+        * __NR_open may not exist on AArch64 (superseded by __NR_openat). */
+#      undef SYS_open
+#    endif
+#    include <sys/uio.h>
+#  endif
+#  include <pthread.h>
+#  if defined(__FreeBSD__) || defined(__DragonFly__)
+#  include <pthread_np.h>
+#  include <sched.h>
+#  if defined(__FreeBSD__)
+#    define cpu_set_t cpuset_t
+#  endif
+#  endif
+#  include <signal.h>
+#  ifdef JEMALLOC_OS_UNFAIR_LOCK
+#    include <os/lock.h>
+#  endif
+#  ifdef JEMALLOC_GLIBC_MALLOC_HOOK
+#    include <sched.h>
+#  endif
+#  include <errno.h>
+#  include <sys/time.h>
+#  include <time.h>
+#  ifdef JEMALLOC_HAVE_MACH_ABSOLUTE_TIME
+#    include <mach/mach_time.h>
+#  endif
+#endif
+#include <sys/types.h>
+
+#include <limits.h>
+#ifndef SIZE_T_MAX
+#  define SIZE_T_MAX	SIZE_MAX
+#endif
+#ifndef SSIZE_MAX
+#  define SSIZE_MAX	((ssize_t)(SIZE_T_MAX >> 1))
+#endif
+#include <stdarg.h>
+#include <stdbool.h>
+#include <stdio.h>
+#include <stdlib.h>
+#include <stdint.h>
+#include <stddef.h>
+#ifndef offsetof
+#  define offsetof(type, member)	((size_t)&(((type *)NULL)->member))
+#endif
+#include <string.h>
+#include <strings.h>
+#include <ctype.h>
+#ifdef _MSC_VER
+#  include <io.h>
+typedef intptr_t ssize_t;
+#  define PATH_MAX 1024
+#  define STDERR_FILENO 2
+#  define __func__ __FUNCTION__
+#  ifdef JEMALLOC_HAS_RESTRICT
+#    define restrict __restrict
+#  endif
+/* Disable warnings about deprecated system functions. */
+#  pragma warning(disable: 4996)
+#if _MSC_VER < 1800
+static int
+isblank(int c) {
+	return (c == '\t' || c == ' ');
+}
+#endif
+#else
+#  include <unistd.h>
+#endif
+#include <fcntl.h>
+
+/*
+ * The Win32 midl compiler has #define small char; we don't use midl, but
+ * "small" is a nice identifier to have available when talking about size
+ * classes.
+ */
+#ifdef small
+#  undef small
+#endif
+
+#endif /* JEMALLOC_INTERNAL_H */

BeefRT/JEMalloc/include/jemalloc/internal/jemalloc_internal_defs.h

@@ -0,0 +1,428 @@
+/* include/jemalloc/internal/jemalloc_internal_defs.h.  Generated from jemalloc_internal_defs.h.in by configure.  */
+#ifndef JEMALLOC_INTERNAL_DEFS_H_
+#define JEMALLOC_INTERNAL_DEFS_H_
+/*
+ * If JEMALLOC_PREFIX is defined via --with-jemalloc-prefix, it will cause all
+ * public APIs to be prefixed.  This makes it possible, with some care, to use
+ * multiple allocators simultaneously.
+ */
+#define JEMALLOC_PREFIX "je_"
+#define JEMALLOC_CPREFIX "JE_"
+
+/*
+ * Define overrides for non-standard allocator-related functions if they are
+ * present on the system.
+ */
+/* #undef JEMALLOC_OVERRIDE___LIBC_CALLOC */
+/* #undef JEMALLOC_OVERRIDE___LIBC_FREE */
+/* #undef JEMALLOC_OVERRIDE___LIBC_MALLOC */
+/* #undef JEMALLOC_OVERRIDE___LIBC_MEMALIGN */
+/* #undef JEMALLOC_OVERRIDE___LIBC_REALLOC */
+/* #undef JEMALLOC_OVERRIDE___LIBC_VALLOC */
+/* #undef JEMALLOC_OVERRIDE___POSIX_MEMALIGN */
+
+/*
+ * JEMALLOC_PRIVATE_NAMESPACE is used as a prefix for all library-private APIs.
+ * For shared libraries, symbol visibility mechanisms prevent these symbols
+ * from being exported, but for static libraries, naming collisions are a real
+ * possibility.
+ */
+#define JEMALLOC_PRIVATE_NAMESPACE je_
+
+/*
+ * Hyper-threaded CPUs may need a special instruction inside spin loops in
+ * order to yield to another virtual CPU.
+ */
+#define CPU_SPINWAIT _mm_pause()
+/* 1 if CPU_SPINWAIT is defined, 0 otherwise. */
+#define HAVE_CPU_SPINWAIT 1
+
+/*
+ * Number of significant bits in virtual addresses.  This may be less than the
+ * total number of bits in a pointer, e.g. on x64, for which the uppermost 16
+ * bits are the same as bit 47.
+ */
+#define LG_VADDR 48
+
+/* Defined if C11 atomics are available. */
+/* #undef JEMALLOC_C11_ATOMICS */
+
+/* Defined if GCC __atomic atomics are available. */
+/* #undef JEMALLOC_GCC_ATOMIC_ATOMICS */
+/* and the 8-bit variant support. */
+/* #undef JEMALLOC_GCC_U8_ATOMIC_ATOMICS */
+
+/* Defined if GCC __sync atomics are available. */
+/* #undef JEMALLOC_GCC_SYNC_ATOMICS */
+/* and the 8-bit variant support. */
+/* #undef JEMALLOC_GCC_U8_SYNC_ATOMICS */
+
+/*
+ * Defined if __builtin_clz() and __builtin_clzl() are available.
+ */
+/* #undef JEMALLOC_HAVE_BUILTIN_CLZ */
+
+/*
+ * Defined if os_unfair_lock_*() functions are available, as provided by Darwin.
+ */
+/* #undef JEMALLOC_OS_UNFAIR_LOCK */
+
+/* Defined if syscall(2) is usable. */
+/* #undef JEMALLOC_USE_SYSCALL */
+
+/*
+ * Defined if secure_getenv(3) is available.
+ */
+/* #undef JEMALLOC_HAVE_SECURE_GETENV */
+
+/*
+ * Defined if issetugid(2) is available.
+ */
+/* #undef JEMALLOC_HAVE_ISSETUGID */
+
+/* Defined if pthread_atfork(3) is available. */
+/* #undef JEMALLOC_HAVE_PTHREAD_ATFORK */
+
+/* Defined if pthread_setname_np(3) is available. */
+/* #undef JEMALLOC_HAVE_PTHREAD_SETNAME_NP */
+
+/* Defined if pthread_getname_np(3) is available. */
+/* #undef JEMALLOC_HAVE_PTHREAD_GETNAME_NP */
+
+/* Defined if pthread_get_name_np(3) is available. */
+/* #undef JEMALLOC_HAVE_PTHREAD_GET_NAME_NP */
+
+/*
+ * Defined if clock_gettime(CLOCK_MONOTONIC_COARSE, ...) is available.
+ */
+/* #undef JEMALLOC_HAVE_CLOCK_MONOTONIC_COARSE */
+
+/*
+ * Defined if clock_gettime(CLOCK_MONOTONIC, ...) is available.
+ */
+/* #undef JEMALLOC_HAVE_CLOCK_MONOTONIC */
+
+/*
+ * Defined if mach_absolute_time() is available.
+ */
+/* #undef JEMALLOC_HAVE_MACH_ABSOLUTE_TIME */
+
+/*
+ * Defined if clock_gettime(CLOCK_REALTIME, ...) is available.
+ */
+/* #undef JEMALLOC_HAVE_CLOCK_REALTIME */
+
+/*
+ * Defined if _malloc_thread_cleanup() exists.  At least in the case of
+ * FreeBSD, pthread_key_create() allocates, which if used during malloc
+ * bootstrapping will cause recursion into the pthreads library.  Therefore, if
+ * _malloc_thread_cleanup() exists, use it as the basis for thread cleanup in
+ * malloc_tsd.
+ */
+/* #undef JEMALLOC_MALLOC_THREAD_CLEANUP */
+
+/*
+ * Defined if threaded initialization is known to be safe on this platform.
+ * Among other things, it must be possible to initialize a mutex without
+ * triggering allocation in order for threaded allocation to be safe.
+ */
+/* #undef JEMALLOC_THREADED_INIT */
+
+/*
+ * Defined if the pthreads implementation defines
+ * _pthread_mutex_init_calloc_cb(), in which case the function is used in order
+ * to avoid recursive allocation during mutex initialization.
+ */
+/* #undef JEMALLOC_MUTEX_INIT_CB */
+
+/* Non-empty if the tls_model attribute is supported. */
+#define JEMALLOC_TLS_MODEL 
+
+/*
+ * JEMALLOC_DEBUG enables assertions and other sanity checks, and disables
+ * inline functions.
+ */
+/* #undef JEMALLOC_DEBUG */
+
+/* JEMALLOC_STATS enables statistics calculation. */
+#define JEMALLOC_STATS 
+
+/* JEMALLOC_EXPERIMENTAL_SMALLOCX_API enables experimental smallocx API. */
+/* #undef JEMALLOC_EXPERIMENTAL_SMALLOCX_API */
+
+/* JEMALLOC_PROF enables allocation profiling. */
+/* #undef JEMALLOC_PROF */
+
+/* Use libunwind for profile backtracing if defined. */
+/* #undef JEMALLOC_PROF_LIBUNWIND */
+
+/* Use libgcc for profile backtracing if defined. */
+/* #undef JEMALLOC_PROF_LIBGCC */
+
+/* Use gcc intrinsics for profile backtracing if defined. */
+/* #undef JEMALLOC_PROF_GCC */
+
+/*
+ * JEMALLOC_DSS enables use of sbrk(2) to allocate extents from the data storage
+ * segment (DSS).
+ */
+/* #undef JEMALLOC_DSS */
+
+/* Support memory filling (junk/zero). */
+#define JEMALLOC_FILL 
+
+/* Support utrace(2)-based tracing. */
+/* #undef JEMALLOC_UTRACE */
+
+/* Support utrace(2)-based tracing (label based signature). */
+/* #undef JEMALLOC_UTRACE_LABEL */
+
+/* Support optional abort() on OOM. */
+/* #undef JEMALLOC_XMALLOC */
+
+/* Support lazy locking (avoid locking unless a second thread is launched). */
+/* #undef JEMALLOC_LAZY_LOCK */
+
+/*
+ * Minimum allocation alignment is 2^LG_QUANTUM bytes (ignoring tiny size
+ * classes).
+ */
+/* #undef LG_QUANTUM */
+
+/* One page is 2^LG_PAGE bytes. */
+#define LG_PAGE 12
+
+/* Maximum number of regions in a slab. */
+/* #undef CONFIG_LG_SLAB_MAXREGS */
+
+/*
+ * One huge page is 2^LG_HUGEPAGE bytes.  Note that this is defined even if the
+ * system does not explicitly support huge pages; system calls that require
+ * explicit huge page support are separately configured.
+ */
+#define LG_HUGEPAGE 21
+
+/*
+ * If defined, adjacent virtual memory mappings with identical attributes
+ * automatically coalesce, and they fragment when changes are made to subranges.
+ * This is the normal order of things for mmap()/munmap(), but on Windows
+ * VirtualAlloc()/VirtualFree() operations must be precisely matched, i.e.
+ * mappings do *not* coalesce/fragment.
+ */
+/* #undef JEMALLOC_MAPS_COALESCE */
+
+/*
+ * If defined, retain memory for later reuse by default rather than using e.g.
+ * munmap() to unmap freed extents.  This is enabled on 64-bit Linux because
+ * common sequences of mmap()/munmap() calls will cause virtual memory map
+ * holes.
+ */
+/* #undef JEMALLOC_RETAIN */
+
+/* TLS is used to map arenas and magazine caches to threads. */
+/* #undef JEMALLOC_TLS */
+
+/*
+ * Used to mark unreachable code to quiet "end of non-void" compiler warnings.
+ * Don't use this directly; instead use unreachable() from util.h
+ */
+#define JEMALLOC_INTERNAL_UNREACHABLE abort
+
+/*
+ * ffs*() functions to use for bitmapping.  Don't use these directly; instead,
+ * use ffs_*() from util.h.
+ */
+#define JEMALLOC_INTERNAL_FFSLL ffsll
+#define JEMALLOC_INTERNAL_FFSL ffsl
+#define JEMALLOC_INTERNAL_FFS ffs
+
+/*
+ * popcount*() functions to use for bitmapping.
+ */
+/* #undef JEMALLOC_INTERNAL_POPCOUNTL */
+/* #undef JEMALLOC_INTERNAL_POPCOUNT */
+
+/*
+ * If defined, explicitly attempt to more uniformly distribute large allocation
+ * pointer alignments across all cache indices.
+ */
+#define JEMALLOC_CACHE_OBLIVIOUS 
+
+/*
+ * If defined, enable logging facilities.  We make this a configure option to
+ * avoid taking extra branches everywhere.
+ */
+/* #undef JEMALLOC_LOG */
+
+/*
+ * If defined, use readlinkat() (instead of readlink()) to follow
+ * /etc/malloc_conf.
+ */
+/* #undef JEMALLOC_READLINKAT */
+
+/*
+ * Darwin (OS X) uses zones to work around Mach-O symbol override shortcomings.
+ */
+/* #undef JEMALLOC_ZONE */
+
+/*
+ * Methods for determining whether the OS overcommits.
+ * JEMALLOC_PROC_SYS_VM_OVERCOMMIT_MEMORY: Linux's
+ *                                         /proc/sys/vm.overcommit_memory file.
+ * JEMALLOC_SYSCTL_VM_OVERCOMMIT: FreeBSD's vm.overcommit sysctl.
+ */
+/* #undef JEMALLOC_SYSCTL_VM_OVERCOMMIT */
+/* #undef JEMALLOC_PROC_SYS_VM_OVERCOMMIT_MEMORY */
+
+/* Defined if madvise(2) is available. */
+/* #undef JEMALLOC_HAVE_MADVISE */
+
+/*
+ * Defined if transparent huge pages are supported via the MADV_[NO]HUGEPAGE
+ * arguments to madvise(2).
+ */
+/* #undef JEMALLOC_HAVE_MADVISE_HUGE */
+
+/*
+ * Methods for purging unused pages differ between operating systems.
+ *
+ *   madvise(..., MADV_FREE) : This marks pages as being unused, such that they
+ *                             will be discarded rather than swapped out.
+ *   madvise(..., MADV_DONTNEED) : If JEMALLOC_PURGE_MADVISE_DONTNEED_ZEROS is
+ *                                 defined, this immediately discards pages,
+ *                                 such that new pages will be demand-zeroed if
+ *                                 the address region is later touched;
+ *                                 otherwise this behaves similarly to
+ *                                 MADV_FREE, though typically with higher
+ *                                 system overhead.
+ */
+/* #undef JEMALLOC_PURGE_MADVISE_FREE */
+/* #undef JEMALLOC_PURGE_MADVISE_DONTNEED */
+/* #undef JEMALLOC_PURGE_MADVISE_DONTNEED_ZEROS */
+
+/* Defined if madvise(2) is available but MADV_FREE is not (x86 Linux only). */
+/* #undef JEMALLOC_DEFINE_MADVISE_FREE */
+
+/*
+ * Defined if MADV_DO[NT]DUMP is supported as an argument to madvise.
+ */
+/* #undef JEMALLOC_MADVISE_DONTDUMP */
+
+/*
+ * Defined if MADV_[NO]CORE is supported as an argument to madvise.
+ */
+/* #undef JEMALLOC_MADVISE_NOCORE */
+
+/* Defined if mprotect(2) is available. */
+/* #undef JEMALLOC_HAVE_MPROTECT */
+
+/*
+ * Defined if transparent huge pages (THPs) are supported via the
+ * MADV_[NO]HUGEPAGE arguments to madvise(2), and THP support is enabled.
+ */
+/* #undef JEMALLOC_THP */
+
+/* Defined if posix_madvise is available. */
+/* #undef JEMALLOC_HAVE_POSIX_MADVISE */
+
+/*
+ * Method for purging unused pages using posix_madvise.
+ *
+ *   posix_madvise(..., POSIX_MADV_DONTNEED)
+ */
+/* #undef JEMALLOC_PURGE_POSIX_MADVISE_DONTNEED */
+/* #undef JEMALLOC_PURGE_POSIX_MADVISE_DONTNEED_ZEROS */
+
+/*
+ * Defined if memcntl page admin call is supported
+ */
+/* #undef JEMALLOC_HAVE_MEMCNTL */
+
+/*
+ * Defined if malloc_size is supported
+ */
+/* #undef JEMALLOC_HAVE_MALLOC_SIZE */
+
+/* Define if operating system has alloca.h header. */
+/* #undef JEMALLOC_HAS_ALLOCA_H */
+
+/* C99 restrict keyword supported. */
+/* #undef JEMALLOC_HAS_RESTRICT */
+
+/* For use by hash code. */
+/* #undef JEMALLOC_BIG_ENDIAN */
+
+/* sizeof(int) == 2^LG_SIZEOF_INT. */
+#define LG_SIZEOF_INT 2
+
+/* sizeof(long) == 2^LG_SIZEOF_LONG. */
+#define LG_SIZEOF_LONG 2
+
+/* sizeof(long long) == 2^LG_SIZEOF_LONG_LONG. */
+#define LG_SIZEOF_LONG_LONG 3
+
+/* sizeof(intmax_t) == 2^LG_SIZEOF_INTMAX_T. */
+#define LG_SIZEOF_INTMAX_T 3
+
+/* glibc malloc hooks (__malloc_hook, __realloc_hook, __free_hook). */
+/* #undef JEMALLOC_GLIBC_MALLOC_HOOK */
+
+/* glibc memalign hook. */
+/* #undef JEMALLOC_GLIBC_MEMALIGN_HOOK */
+
+/* pthread support */
+/* #undef JEMALLOC_HAVE_PTHREAD */
+
+/* dlsym() support */
+/* #undef JEMALLOC_HAVE_DLSYM */
+
+/* Adaptive mutex support in pthreads. */
+/* #undef JEMALLOC_HAVE_PTHREAD_MUTEX_ADAPTIVE_NP */
+
+/* GNU specific sched_getcpu support */
+/* #undef JEMALLOC_HAVE_SCHED_GETCPU */
+
+/* GNU specific sched_setaffinity support */
+/* #undef JEMALLOC_HAVE_SCHED_SETAFFINITY */
+
+/*
+ * If defined, all the features necessary for background threads are present.
+ */
+/* #undef JEMALLOC_BACKGROUND_THREAD */
+
+/*
+ * If defined, jemalloc symbols are not exported (doesn't work when
+ * JEMALLOC_PREFIX is not defined).
+ */
+/* #undef JEMALLOC_EXPORT */
+
+/* config.malloc_conf options string. */
+#define JEMALLOC_CONFIG_MALLOC_CONF ""
+
+/* If defined, jemalloc takes the malloc/free/etc. symbol names. */
+/* #undef JEMALLOC_IS_MALLOC */
+
+/*
+ * Defined if strerror_r returns char * if _GNU_SOURCE is defined.
+ */
+/* #undef JEMALLOC_STRERROR_R_RETURNS_CHAR_WITH_GNU_SOURCE */
+
+/* Performs additional safety checks when defined. */
+/* #undef JEMALLOC_OPT_SAFETY_CHECKS */
+
+/* Is C++ support being built? */
+/* #undef JEMALLOC_ENABLE_CXX */
+
+/* Performs additional size checks when defined. */
+/* #undef JEMALLOC_OPT_SIZE_CHECKS */
+
+/* Allows sampled junk and stash for checking use-after-free when defined. */
+/* #undef JEMALLOC_UAF_DETECTION */
+
+/* Darwin VM_MAKE_TAG support */
+/* #undef JEMALLOC_HAVE_VM_MAKE_TAG */
+
+/* If defined, realloc(ptr, 0) defaults to "free" instead of "alloc". */
+#define JEMALLOC_ZERO_REALLOC_DEFAULT_FREE 
+
+#endif /* JEMALLOC_INTERNAL_DEFS_H_ */

BeefRT/JEMalloc/include/jemalloc/internal/jemalloc_internal_defs.h.in

@@ -0,0 +1,427 @@
+#ifndef JEMALLOC_INTERNAL_DEFS_H_
+#define JEMALLOC_INTERNAL_DEFS_H_
+/*
+ * If JEMALLOC_PREFIX is defined via --with-jemalloc-prefix, it will cause all
+ * public APIs to be prefixed.  This makes it possible, with some care, to use
+ * multiple allocators simultaneously.
+ */
+#undef JEMALLOC_PREFIX
+#undef JEMALLOC_CPREFIX
+
+/*
+ * Define overrides for non-standard allocator-related functions if they are
+ * present on the system.
+ */
+#undef JEMALLOC_OVERRIDE___LIBC_CALLOC
+#undef JEMALLOC_OVERRIDE___LIBC_FREE
+#undef JEMALLOC_OVERRIDE___LIBC_MALLOC
+#undef JEMALLOC_OVERRIDE___LIBC_MEMALIGN
+#undef JEMALLOC_OVERRIDE___LIBC_REALLOC
+#undef JEMALLOC_OVERRIDE___LIBC_VALLOC
+#undef JEMALLOC_OVERRIDE___POSIX_MEMALIGN
+
+/*
+ * JEMALLOC_PRIVATE_NAMESPACE is used as a prefix for all library-private APIs.
+ * For shared libraries, symbol visibility mechanisms prevent these symbols
+ * from being exported, but for static libraries, naming collisions are a real
+ * possibility.
+ */
+#undef JEMALLOC_PRIVATE_NAMESPACE
+
+/*
+ * Hyper-threaded CPUs may need a special instruction inside spin loops in
+ * order to yield to another virtual CPU.
+ */
+#undef CPU_SPINWAIT
+/* 1 if CPU_SPINWAIT is defined, 0 otherwise. */
+#undef HAVE_CPU_SPINWAIT
+
+/*
+ * Number of significant bits in virtual addresses.  This may be less than the
+ * total number of bits in a pointer, e.g. on x64, for which the uppermost 16
+ * bits are the same as bit 47.
+ */
+#undef LG_VADDR
+
+/* Defined if C11 atomics are available. */
+#undef JEMALLOC_C11_ATOMICS
+
+/* Defined if GCC __atomic atomics are available. */
+#undef JEMALLOC_GCC_ATOMIC_ATOMICS
+/* and the 8-bit variant support. */
+#undef JEMALLOC_GCC_U8_ATOMIC_ATOMICS
+
+/* Defined if GCC __sync atomics are available. */
+#undef JEMALLOC_GCC_SYNC_ATOMICS
+/* and the 8-bit variant support. */
+#undef JEMALLOC_GCC_U8_SYNC_ATOMICS
+
+/*
+ * Defined if __builtin_clz() and __builtin_clzl() are available.
+ */
+#undef JEMALLOC_HAVE_BUILTIN_CLZ
+
+/*
+ * Defined if os_unfair_lock_*() functions are available, as provided by Darwin.
+ */
+#undef JEMALLOC_OS_UNFAIR_LOCK
+
+/* Defined if syscall(2) is usable. */
+#undef JEMALLOC_USE_SYSCALL
+
+/*
+ * Defined if secure_getenv(3) is available.
+ */
+#undef JEMALLOC_HAVE_SECURE_GETENV
+
+/*
+ * Defined if issetugid(2) is available.
+ */
+#undef JEMALLOC_HAVE_ISSETUGID
+
+/* Defined if pthread_atfork(3) is available. */
+#undef JEMALLOC_HAVE_PTHREAD_ATFORK
+
+/* Defined if pthread_setname_np(3) is available. */
+#undef JEMALLOC_HAVE_PTHREAD_SETNAME_NP
+
+/* Defined if pthread_getname_np(3) is available. */
+#undef JEMALLOC_HAVE_PTHREAD_GETNAME_NP
+
+/* Defined if pthread_get_name_np(3) is available. */
+#undef JEMALLOC_HAVE_PTHREAD_GET_NAME_NP
+
+/*
+ * Defined if clock_gettime(CLOCK_MONOTONIC_COARSE, ...) is available.
+ */
+#undef JEMALLOC_HAVE_CLOCK_MONOTONIC_COARSE
+
+/*
+ * Defined if clock_gettime(CLOCK_MONOTONIC, ...) is available.
+ */
+#undef JEMALLOC_HAVE_CLOCK_MONOTONIC
+
+/*
+ * Defined if mach_absolute_time() is available.
+ */
+#undef JEMALLOC_HAVE_MACH_ABSOLUTE_TIME
+
+/*
+ * Defined if clock_gettime(CLOCK_REALTIME, ...) is available.
+ */
+#undef JEMALLOC_HAVE_CLOCK_REALTIME
+
+/*
+ * Defined if _malloc_thread_cleanup() exists.  At least in the case of
+ * FreeBSD, pthread_key_create() allocates, which if used during malloc
+ * bootstrapping will cause recursion into the pthreads library.  Therefore, if
+ * _malloc_thread_cleanup() exists, use it as the basis for thread cleanup in
+ * malloc_tsd.
+ */
+#undef JEMALLOC_MALLOC_THREAD_CLEANUP
+
+/*
+ * Defined if threaded initialization is known to be safe on this platform.
+ * Among other things, it must be possible to initialize a mutex without
+ * triggering allocation in order for threaded allocation to be safe.
+ */
+#undef JEMALLOC_THREADED_INIT
+
+/*
+ * Defined if the pthreads implementation defines
+ * _pthread_mutex_init_calloc_cb(), in which case the function is used in order
+ * to avoid recursive allocation during mutex initialization.
+ */
+#undef JEMALLOC_MUTEX_INIT_CB
+
+/* Non-empty if the tls_model attribute is supported. */
+#undef JEMALLOC_TLS_MODEL
+
+/*
+ * JEMALLOC_DEBUG enables assertions and other sanity checks, and disables
+ * inline functions.
+ */
+#undef JEMALLOC_DEBUG
+
+/* JEMALLOC_STATS enables statistics calculation. */
+#undef JEMALLOC_STATS
+
+/* JEMALLOC_EXPERIMENTAL_SMALLOCX_API enables experimental smallocx API. */
+#undef JEMALLOC_EXPERIMENTAL_SMALLOCX_API
+
+/* JEMALLOC_PROF enables allocation profiling. */
+#undef JEMALLOC_PROF
+
+/* Use libunwind for profile backtracing if defined. */
+#undef JEMALLOC_PROF_LIBUNWIND
+
+/* Use libgcc for profile backtracing if defined. */
+#undef JEMALLOC_PROF_LIBGCC
+
+/* Use gcc intrinsics for profile backtracing if defined. */
+#undef JEMALLOC_PROF_GCC
+
+/*
+ * JEMALLOC_DSS enables use of sbrk(2) to allocate extents from the data storage
+ * segment (DSS).
+ */
+#undef JEMALLOC_DSS
+
+/* Support memory filling (junk/zero). */
+#undef JEMALLOC_FILL
+
+/* Support utrace(2)-based tracing. */
+#undef JEMALLOC_UTRACE
+
+/* Support utrace(2)-based tracing (label based signature). */
+#undef JEMALLOC_UTRACE_LABEL
+
+/* Support optional abort() on OOM. */
+#undef JEMALLOC_XMALLOC
+
+/* Support lazy locking (avoid locking unless a second thread is launched). */
+#undef JEMALLOC_LAZY_LOCK
+
+/*
+ * Minimum allocation alignment is 2^LG_QUANTUM bytes (ignoring tiny size
+ * classes).
+ */
+#undef LG_QUANTUM
+
+/* One page is 2^LG_PAGE bytes. */
+#undef LG_PAGE
+
+/* Maximum number of regions in a slab. */
+#undef CONFIG_LG_SLAB_MAXREGS
+
+/*
+ * One huge page is 2^LG_HUGEPAGE bytes.  Note that this is defined even if the
+ * system does not explicitly support huge pages; system calls that require
+ * explicit huge page support are separately configured.
+ */
+#undef LG_HUGEPAGE
+
+/*
+ * If defined, adjacent virtual memory mappings with identical attributes
+ * automatically coalesce, and they fragment when changes are made to subranges.
+ * This is the normal order of things for mmap()/munmap(), but on Windows
+ * VirtualAlloc()/VirtualFree() operations must be precisely matched, i.e.
+ * mappings do *not* coalesce/fragment.
+ */
+#undef JEMALLOC_MAPS_COALESCE
+
+/*
+ * If defined, retain memory for later reuse by default rather than using e.g.
+ * munmap() to unmap freed extents.  This is enabled on 64-bit Linux because
+ * common sequences of mmap()/munmap() calls will cause virtual memory map
+ * holes.
+ */
+#undef JEMALLOC_RETAIN
+
+/* TLS is used to map arenas and magazine caches to threads. */
+#undef JEMALLOC_TLS
+
+/*
+ * Used to mark unreachable code to quiet "end of non-void" compiler warnings.
+ * Don't use this directly; instead use unreachable() from util.h
+ */
+#undef JEMALLOC_INTERNAL_UNREACHABLE
+
+/*
+ * ffs*() functions to use for bitmapping.  Don't use these directly; instead,
+ * use ffs_*() from util.h.
+ */
+#undef JEMALLOC_INTERNAL_FFSLL
+#undef JEMALLOC_INTERNAL_FFSL
+#undef JEMALLOC_INTERNAL_FFS
+
+/*
+ * popcount*() functions to use for bitmapping.
+ */
+#undef JEMALLOC_INTERNAL_POPCOUNTL
+#undef JEMALLOC_INTERNAL_POPCOUNT
+
+/*
+ * If defined, explicitly attempt to more uniformly distribute large allocation
+ * pointer alignments across all cache indices.
+ */
+#undef JEMALLOC_CACHE_OBLIVIOUS
+
+/*
+ * If defined, enable logging facilities.  We make this a configure option to
+ * avoid taking extra branches everywhere.
+ */
+#undef JEMALLOC_LOG
+
+/*
+ * If defined, use readlinkat() (instead of readlink()) to follow
+ * /etc/malloc_conf.
+ */
+#undef JEMALLOC_READLINKAT
+
+/*
+ * Darwin (OS X) uses zones to work around Mach-O symbol override shortcomings.
+ */
+#undef JEMALLOC_ZONE
+
+/*
+ * Methods for determining whether the OS overcommits.
+ * JEMALLOC_PROC_SYS_VM_OVERCOMMIT_MEMORY: Linux's
+ *                                         /proc/sys/vm.overcommit_memory file.
+ * JEMALLOC_SYSCTL_VM_OVERCOMMIT: FreeBSD's vm.overcommit sysctl.
+ */
+#undef JEMALLOC_SYSCTL_VM_OVERCOMMIT
+#undef JEMALLOC_PROC_SYS_VM_OVERCOMMIT_MEMORY
+
+/* Defined if madvise(2) is available. */
+#undef JEMALLOC_HAVE_MADVISE
+
+/*
+ * Defined if transparent huge pages are supported via the MADV_[NO]HUGEPAGE
+ * arguments to madvise(2).
+ */
+#undef JEMALLOC_HAVE_MADVISE_HUGE
+
+/*
+ * Methods for purging unused pages differ between operating systems.
+ *
+ *   madvise(..., MADV_FREE) : This marks pages as being unused, such that they
+ *                             will be discarded rather than swapped out.
+ *   madvise(..., MADV_DONTNEED) : If JEMALLOC_PURGE_MADVISE_DONTNEED_ZEROS is
+ *                                 defined, this immediately discards pages,
+ *                                 such that new pages will be demand-zeroed if
+ *                                 the address region is later touched;
+ *                                 otherwise this behaves similarly to
+ *                                 MADV_FREE, though typically with higher
+ *                                 system overhead.
+ */
+#undef JEMALLOC_PURGE_MADVISE_FREE
+#undef JEMALLOC_PURGE_MADVISE_DONTNEED
+#undef JEMALLOC_PURGE_MADVISE_DONTNEED_ZEROS
+
+/* Defined if madvise(2) is available but MADV_FREE is not (x86 Linux only). */
+#undef JEMALLOC_DEFINE_MADVISE_FREE
+
+/*
+ * Defined if MADV_DO[NT]DUMP is supported as an argument to madvise.
+ */
+#undef JEMALLOC_MADVISE_DONTDUMP
+
+/*
+ * Defined if MADV_[NO]CORE is supported as an argument to madvise.
+ */
+#undef JEMALLOC_MADVISE_NOCORE
+
+/* Defined if mprotect(2) is available. */
+#undef JEMALLOC_HAVE_MPROTECT
+
+/*
+ * Defined if transparent huge pages (THPs) are supported via the
+ * MADV_[NO]HUGEPAGE arguments to madvise(2), and THP support is enabled.
+ */
+#undef JEMALLOC_THP
+
+/* Defined if posix_madvise is available. */
+#undef JEMALLOC_HAVE_POSIX_MADVISE
+
+/*
+ * Method for purging unused pages using posix_madvise.
+ *
+ *   posix_madvise(..., POSIX_MADV_DONTNEED)
+ */
+#undef JEMALLOC_PURGE_POSIX_MADVISE_DONTNEED
+#undef JEMALLOC_PURGE_POSIX_MADVISE_DONTNEED_ZEROS
+
+/*
+ * Defined if memcntl page admin call is supported
+ */
+#undef JEMALLOC_HAVE_MEMCNTL
+
+/*
+ * Defined if malloc_size is supported
+ */
+#undef JEMALLOC_HAVE_MALLOC_SIZE
+
+/* Define if operating system has alloca.h header. */
+#undef JEMALLOC_HAS_ALLOCA_H
+
+/* C99 restrict keyword supported. */
+#undef JEMALLOC_HAS_RESTRICT
+
+/* For use by hash code. */
+#undef JEMALLOC_BIG_ENDIAN
+
+/* sizeof(int) == 2^LG_SIZEOF_INT. */
+#undef LG_SIZEOF_INT
+
+/* sizeof(long) == 2^LG_SIZEOF_LONG. */
+#undef LG_SIZEOF_LONG
+
+/* sizeof(long long) == 2^LG_SIZEOF_LONG_LONG. */
+#undef LG_SIZEOF_LONG_LONG
+
+/* sizeof(intmax_t) == 2^LG_SIZEOF_INTMAX_T. */
+#undef LG_SIZEOF_INTMAX_T
+
+/* glibc malloc hooks (__malloc_hook, __realloc_hook, __free_hook). */
+#undef JEMALLOC_GLIBC_MALLOC_HOOK
+
+/* glibc memalign hook. */
+#undef JEMALLOC_GLIBC_MEMALIGN_HOOK
+
+/* pthread support */
+#undef JEMALLOC_HAVE_PTHREAD
+
+/* dlsym() support */
+#undef JEMALLOC_HAVE_DLSYM
+
+/* Adaptive mutex support in pthreads. */
+#undef JEMALLOC_HAVE_PTHREAD_MUTEX_ADAPTIVE_NP
+
+/* GNU specific sched_getcpu support */
+#undef JEMALLOC_HAVE_SCHED_GETCPU
+
+/* GNU specific sched_setaffinity support */
+#undef JEMALLOC_HAVE_SCHED_SETAFFINITY
+
+/*
+ * If defined, all the features necessary for background threads are present.
+ */
+#undef JEMALLOC_BACKGROUND_THREAD
+
+/*
+ * If defined, jemalloc symbols are not exported (doesn't work when
+ * JEMALLOC_PREFIX is not defined).
+ */
+#undef JEMALLOC_EXPORT
+
+/* config.malloc_conf options string. */
+#undef JEMALLOC_CONFIG_MALLOC_CONF
+
+/* If defined, jemalloc takes the malloc/free/etc. symbol names. */
+#undef JEMALLOC_IS_MALLOC
+
+/*
+ * Defined if strerror_r returns char * if _GNU_SOURCE is defined.
+ */
+#undef JEMALLOC_STRERROR_R_RETURNS_CHAR_WITH_GNU_SOURCE
+
+/* Performs additional safety checks when defined. */
+#undef JEMALLOC_OPT_SAFETY_CHECKS
+
+/* Is C++ support being built? */
+#undef JEMALLOC_ENABLE_CXX
+
+/* Performs additional size checks when defined. */
+#undef JEMALLOC_OPT_SIZE_CHECKS
+
+/* Allows sampled junk and stash for checking use-after-free when defined. */
+#undef JEMALLOC_UAF_DETECTION
+
+/* Darwin VM_MAKE_TAG support */
+#undef JEMALLOC_HAVE_VM_MAKE_TAG
+
+/* If defined, realloc(ptr, 0) defaults to "free" instead of "alloc". */
+#undef JEMALLOC_ZERO_REALLOC_DEFAULT_FREE
+
+#endif /* JEMALLOC_INTERNAL_DEFS_H_ */

BeefRT/JEMalloc/include/jemalloc/internal/jemalloc_internal_externs.h

@@ -0,0 +1,75 @@
+#ifndef JEMALLOC_INTERNAL_EXTERNS_H
+#define JEMALLOC_INTERNAL_EXTERNS_H
+
+#include "jemalloc/internal/atomic.h"
+#include "jemalloc/internal/hpa_opts.h"
+#include "jemalloc/internal/sec_opts.h"
+#include "jemalloc/internal/tsd_types.h"
+#include "jemalloc/internal/nstime.h"
+
+/* TSD checks this to set thread local slow state accordingly. */
+extern bool malloc_slow;
+
+/* Run-time options. */
+extern bool opt_abort;
+extern bool opt_abort_conf;
+extern bool opt_trust_madvise;
+extern bool opt_confirm_conf;
+extern bool opt_hpa;
+extern hpa_shard_opts_t opt_hpa_opts;
+extern sec_opts_t opt_hpa_sec_opts;
+
+extern const char *opt_junk;
+extern bool opt_junk_alloc;
+extern bool opt_junk_free;
+extern void (*junk_free_callback)(void *ptr, size_t size);
+extern void (*junk_alloc_callback)(void *ptr, size_t size);
+extern bool opt_utrace;
+extern bool opt_xmalloc;
+extern bool opt_experimental_infallible_new;
+extern bool opt_zero;
+extern unsigned opt_narenas;
+extern zero_realloc_action_t opt_zero_realloc_action;
+extern malloc_init_t malloc_init_state;
+extern const char *zero_realloc_mode_names[];
+extern atomic_zu_t zero_realloc_count;
+extern bool opt_cache_oblivious;
+
+/* Escape free-fastpath when ptr & mask == 0 (for sanitization purpose). */
+extern uintptr_t san_cache_bin_nonfast_mask;
+
+/* Number of CPUs. */
+extern unsigned ncpus;
+
+/* Number of arenas used for automatic multiplexing of threads and arenas. */
+extern unsigned narenas_auto;
+
+/* Base index for manual arenas. */
+extern unsigned manual_arena_base;
+
+/*
+ * Arenas that are used to service external requests.  Not all elements of the
+ * arenas array are necessarily used; arenas are created lazily as needed.
+ */
+extern atomic_p_t arenas[];
+
+void *a0malloc(size_t size);
+void a0dalloc(void *ptr);
+void *bootstrap_malloc(size_t size);
+void *bootstrap_calloc(size_t num, size_t size);
+void bootstrap_free(void *ptr);
+void arena_set(unsigned ind, arena_t *arena);
+unsigned narenas_total_get(void);
+arena_t *arena_init(tsdn_t *tsdn, unsigned ind, const arena_config_t *config);
+arena_t *arena_choose_hard(tsd_t *tsd, bool internal);
+void arena_migrate(tsd_t *tsd, arena_t *oldarena, arena_t *newarena);
+void iarena_cleanup(tsd_t *tsd);
+void arena_cleanup(tsd_t *tsd);
+size_t batch_alloc(void **ptrs, size_t num, size_t size, int flags);
+void jemalloc_prefork(void);
+void jemalloc_postfork_parent(void);
+void jemalloc_postfork_child(void);
+void je_sdallocx_noflags(void *ptr, size_t size);
+void *malloc_default(size_t size);
+
+#endif /* JEMALLOC_INTERNAL_EXTERNS_H */

BeefRT/JEMalloc/include/jemalloc/internal/jemalloc_internal_includes.h

@@ -0,0 +1,84 @@
+#ifndef JEMALLOC_INTERNAL_INCLUDES_H
+#define JEMALLOC_INTERNAL_INCLUDES_H
+
+/*
+ * jemalloc can conceptually be broken into components (arena, tcache, etc.),
+ * but there are circular dependencies that cannot be broken without
+ * substantial performance degradation.
+ *
+ * Historically, we dealt with this by each header into four sections (types,
+ * structs, externs, and inlines), and included each header file multiple times
+ * in this file, picking out the portion we want on each pass using the
+ * following #defines:
+ *   JEMALLOC_H_TYPES   : Preprocessor-defined constants and pseudo-opaque data
+ *                        types.
+ *   JEMALLOC_H_STRUCTS : Data structures.
+ *   JEMALLOC_H_EXTERNS : Extern data declarations and function prototypes.
+ *   JEMALLOC_H_INLINES : Inline functions.
+ *
+ * We're moving toward a world in which the dependencies are explicit; each file
+ * will #include the headers it depends on (rather than relying on them being
+ * implicitly available via this file including every header file in the
+ * project).
+ *
+ * We're now in an intermediate state: we've broken up the header files to avoid
+ * having to include each one multiple times, but have not yet moved the
+ * dependency information into the header files (i.e. we still rely on the
+ * ordering in this file to ensure all a header's dependencies are available in
+ * its translation unit).  Each component is now broken up into multiple header
+ * files, corresponding to the sections above (e.g. instead of "foo.h", we now
+ * have "foo_types.h", "foo_structs.h", "foo_externs.h", "foo_inlines.h").
+ *
+ * Those files which have been converted to explicitly include their
+ * inter-component dependencies are now in the initial HERMETIC HEADERS
+ * section.  All headers may still rely on jemalloc_preamble.h (which, by fiat,
+ * must be included first in every translation unit) for system headers and
+ * global jemalloc definitions, however.
+ */
+
+/******************************************************************************/
+/* TYPES */
+/******************************************************************************/
+
+#include "jemalloc/internal/arena_types.h"
+#include "jemalloc/internal/tcache_types.h"
+#include "jemalloc/internal/prof_types.h"
+
+/******************************************************************************/
+/* STRUCTS */
+/******************************************************************************/
+
+#include "jemalloc/internal/prof_structs.h"
+#include "jemalloc/internal/arena_structs.h"
+#include "jemalloc/internal/tcache_structs.h"
+#include "jemalloc/internal/background_thread_structs.h"
+
+/******************************************************************************/
+/* EXTERNS */
+/******************************************************************************/
+
+#include "jemalloc/internal/jemalloc_internal_externs.h"
+#include "jemalloc/internal/arena_externs.h"
+#include "jemalloc/internal/large_externs.h"
+#include "jemalloc/internal/tcache_externs.h"
+#include "jemalloc/internal/prof_externs.h"
+#include "jemalloc/internal/background_thread_externs.h"
+
+/******************************************************************************/
+/* INLINES */
+/******************************************************************************/
+
+#include "jemalloc/internal/jemalloc_internal_inlines_a.h"
+/*
+ * Include portions of arena code interleaved with tcache code in order to
+ * resolve circular dependencies.
+ */
+#include "jemalloc/internal/arena_inlines_a.h"
+#include "jemalloc/internal/jemalloc_internal_inlines_b.h"
+#include "jemalloc/internal/tcache_inlines.h"
+#include "jemalloc/internal/arena_inlines_b.h"
+#include "jemalloc/internal/jemalloc_internal_inlines_c.h"
+#include "jemalloc/internal/prof_inlines.h"
+#include "jemalloc/internal/background_thread_inlines.h"
+
+#endif /* JEMALLOC_INTERNAL_INCLUDES_H */

BeefRT/JEMalloc/include/jemalloc/internal/jemalloc_internal_inlines_a.h

@@ -0,0 +1,122 @@
+#ifndef JEMALLOC_INTERNAL_INLINES_A_H
+#define JEMALLOC_INTERNAL_INLINES_A_H
+
+#include "jemalloc/internal/atomic.h"
+#include "jemalloc/internal/bit_util.h"
+#include "jemalloc/internal/jemalloc_internal_types.h"
+#include "jemalloc/internal/sc.h"
+#include "jemalloc/internal/ticker.h"
+
+JEMALLOC_ALWAYS_INLINE malloc_cpuid_t
+malloc_getcpu(void) {
+	assert(have_percpu_arena);
+#if defined(_WIN32)
+	return GetCurrentProcessorNumber();
+#elif defined(JEMALLOC_HAVE_SCHED_GETCPU)
+	return (malloc_cpuid_t)sched_getcpu();
+#else
+	not_reached();
+	return -1;
+#endif
+}
+
+/* Return the chosen arena index based on current cpu. */
+JEMALLOC_ALWAYS_INLINE unsigned
+percpu_arena_choose(void) {
+	assert(have_percpu_arena && PERCPU_ARENA_ENABLED(opt_percpu_arena));
+
+	malloc_cpuid_t cpuid = malloc_getcpu();
+	assert(cpuid >= 0);
+
+	unsigned arena_ind;
+	if ((opt_percpu_arena == percpu_arena) || ((unsigned)cpuid < ncpus /
+	    2)) {
+		arena_ind = cpuid;
+	} else {
+		assert(opt_percpu_arena == per_phycpu_arena);
+		/* Hyper threads on the same physical CPU share arena. */
+		arena_ind = cpuid - ncpus / 2;
+	}
+
+	return arena_ind;
+}
+
+/* Return the limit of percpu auto arena range, i.e. arenas[0...ind_limit). */
+JEMALLOC_ALWAYS_INLINE unsigned
+percpu_arena_ind_limit(percpu_arena_mode_t mode) {
+	assert(have_percpu_arena && PERCPU_ARENA_ENABLED(mode));
+	if (mode == per_phycpu_arena && ncpus > 1) {
+		if (ncpus % 2) {
+			/* This likely means a misconfig. */
+			return ncpus / 2 + 1;
+		}
+		return ncpus / 2;
+	} else {
+		return ncpus;
+	}
+}
+
+static inline arena_t *
+arena_get(tsdn_t *tsdn, unsigned ind, bool init_if_missing) {
+	arena_t *ret;
+
+	assert(ind < MALLOCX_ARENA_LIMIT);
+
+	ret = (arena_t *)atomic_load_p(&arenas[ind], ATOMIC_ACQUIRE);
+	if (unlikely(ret == NULL)) {
+		if (init_if_missing) {
+			ret = arena_init(tsdn, ind, &arena_config_default);
+		}
+	}
+	return ret;
+}
+
+JEMALLOC_ALWAYS_INLINE bool
+tcache_available(tsd_t *tsd) {
+	/*
+	 * Thread specific auto tcache might be unavailable if: 1) during tcache
+	 * initialization, or 2) disabled through thread.tcache.enabled mallctl
+	 * or config options.  This check covers all cases.
+	 */
+	if (likely(tsd_tcache_enabled_get(tsd))) {
+		/* Associated arena == NULL implies tcache init in progress. */
+		if (config_debug && tsd_tcache_slowp_get(tsd)->arena != NULL) {
+			tcache_assert_initialized(tsd_tcachep_get(tsd));
+		}
+		return true;
+	}
+
+	return false;
+}
+
+JEMALLOC_ALWAYS_INLINE tcache_t *
+tcache_get(tsd_t *tsd) {
+	if (!tcache_available(tsd)) {
+		return NULL;
+	}
+
+	return tsd_tcachep_get(tsd);
+}
+
+JEMALLOC_ALWAYS_INLINE tcache_slow_t *
+tcache_slow_get(tsd_t *tsd) {
+	if (!tcache_available(tsd)) {
+		return NULL;
+	}
+
+	return tsd_tcache_slowp_get(tsd);
+}
+
+static inline void
+pre_reentrancy(tsd_t *tsd, arena_t *arena) {
+	/* arena is the current context.  Reentry from a0 is not allowed. */
+	assert(arena != arena_get(tsd_tsdn(tsd), 0, false));
+	tsd_pre_reentrancy_raw(tsd);
+}
+
+static inline void
+post_reentrancy(tsd_t *tsd) {
+	tsd_post_reentrancy_raw(tsd);
+}
+
+#endif /* JEMALLOC_INTERNAL_INLINES_A_H */

BeefRT/JEMalloc/include/jemalloc/internal/jemalloc_internal_inlines_b.h

@@ -0,0 +1,103 @@
+#ifndef JEMALLOC_INTERNAL_INLINES_B_H
+#define JEMALLOC_INTERNAL_INLINES_B_H
+
+#include "jemalloc/internal/extent.h"
+
+static inline void
+percpu_arena_update(tsd_t *tsd, unsigned cpu) {
+	assert(have_percpu_arena);
+	arena_t *oldarena = tsd_arena_get(tsd);
+	assert(oldarena != NULL);
+	unsigned oldind = arena_ind_get(oldarena);
+
+	if (oldind != cpu) {
+		unsigned newind = cpu;
+		arena_t *newarena = arena_get(tsd_tsdn(tsd), newind, true);
+		assert(newarena != NULL);
+
+		/* Set new arena/tcache associations. */
+		arena_migrate(tsd, oldarena, newarena);
+		tcache_t *tcache = tcache_get(tsd);
+		if (tcache != NULL) {
+			tcache_slow_t *tcache_slow = tsd_tcache_slowp_get(tsd);
+			tcache_arena_reassociate(tsd_tsdn(tsd), tcache_slow,
+			    tcache, newarena);
+		}
+	}
+}
+
+
+/* Choose an arena based on a per-thread value. */
+static inline arena_t *
+arena_choose_impl(tsd_t *tsd, arena_t *arena, bool internal) {
+	arena_t *ret;
+
+	if (arena != NULL) {
+		return arena;
+	}
+
+	/* During reentrancy, arena 0 is the safest bet. */
+	if (unlikely(tsd_reentrancy_level_get(tsd) > 0)) {
+		return arena_get(tsd_tsdn(tsd), 0, true);
+	}
+
+	ret = internal ? tsd_iarena_get(tsd) : tsd_arena_get(tsd);
+	if (unlikely(ret == NULL)) {
+		ret = arena_choose_hard(tsd, internal);
+		assert(ret);
+		if (tcache_available(tsd)) {
+			tcache_slow_t *tcache_slow = tsd_tcache_slowp_get(tsd);
+			tcache_t *tcache = tsd_tcachep_get(tsd);
+			if (tcache_slow->arena != NULL) {
+				/* See comments in tsd_tcache_data_init().*/
+				assert(tcache_slow->arena ==
+				    arena_get(tsd_tsdn(tsd), 0, false));
+				if (tcache_slow->arena != ret) {
+					tcache_arena_reassociate(tsd_tsdn(tsd),
+					    tcache_slow, tcache, ret);
+				}
+			} else {
+				tcache_arena_associate(tsd_tsdn(tsd),
+				    tcache_slow, tcache, ret);
+			}
+		}
+	}
+
+	/*
+	 * Note that for percpu arena, if the current arena is outside of the
+	 * auto percpu arena range, (i.e. thread is assigned to a manually
+	 * managed arena), then percpu arena is skipped.
+	 */
+	if (have_percpu_arena && PERCPU_ARENA_ENABLED(opt_percpu_arena) &&
+	    !internal && (arena_ind_get(ret) <
+	    percpu_arena_ind_limit(opt_percpu_arena)) && (ret->last_thd !=
+	    tsd_tsdn(tsd))) {
+		unsigned ind = percpu_arena_choose();
+		if (arena_ind_get(ret) != ind) {
+			percpu_arena_update(tsd, ind);
+			ret = tsd_arena_get(tsd);
+		}
+		ret->last_thd = tsd_tsdn(tsd);
+	}
+
+	return ret;
+}
+
+static inline arena_t *
+arena_choose(tsd_t *tsd, arena_t *arena) {
+	return arena_choose_impl(tsd, arena, false);
+}
+
+static inline arena_t *
+arena_ichoose(tsd_t *tsd, arena_t *arena) {
+	return arena_choose_impl(tsd, arena, true);
+}
+
+static inline bool
+arena_is_auto(arena_t *arena) {
+	assert(narenas_auto > 0);
+
+	return (arena_ind_get(arena) < manual_arena_base);
+}
+
+#endif /* JEMALLOC_INTERNAL_INLINES_B_H */

BeefRT/JEMalloc/include/jemalloc/internal/jemalloc_internal_inlines_c.h

@@ -0,0 +1,340 @@
+#ifndef JEMALLOC_INTERNAL_INLINES_C_H
+#define JEMALLOC_INTERNAL_INLINES_C_H
+
+#include "jemalloc/internal/hook.h"
+#include "jemalloc/internal/jemalloc_internal_types.h"
+#include "jemalloc/internal/log.h"
+#include "jemalloc/internal/sz.h"
+#include "jemalloc/internal/thread_event.h"
+#include "jemalloc/internal/witness.h"
+
+/*
+ * Translating the names of the 'i' functions:
+ *   Abbreviations used in the first part of the function name (before
+ *   alloc/dalloc) describe what that function accomplishes:
+ *     a: arena (query)
+ *     s: size (query, or sized deallocation)
+ *     e: extent (query)
+ *     p: aligned (allocates)
+ *     vs: size (query, without knowing that the pointer is into the heap)
+ *     r: rallocx implementation
+ *     x: xallocx implementation
+ *   Abbreviations used in the second part of the function name (after
+ *   alloc/dalloc) describe the arguments it takes
+ *     z: whether to return zeroed memory
+ *     t: accepts a tcache_t * parameter
+ *     m: accepts an arena_t * parameter
+ */
+
+JEMALLOC_ALWAYS_INLINE arena_t *
+iaalloc(tsdn_t *tsdn, const void *ptr) {
+	assert(ptr != NULL);
+
+	return arena_aalloc(tsdn, ptr);
+}
+
+JEMALLOC_ALWAYS_INLINE size_t
+isalloc(tsdn_t *tsdn, const void *ptr) {
+	assert(ptr != NULL);
+
+	return arena_salloc(tsdn, ptr);
+}
+
+JEMALLOC_ALWAYS_INLINE void *
+iallocztm(tsdn_t *tsdn, size_t size, szind_t ind, bool zero, tcache_t *tcache,
+    bool is_internal, arena_t *arena, bool slow_path) {
+	void *ret;
+
+	assert(!is_internal || tcache == NULL);
+	assert(!is_internal || arena == NULL || arena_is_auto(arena));
+	if (!tsdn_null(tsdn) && tsd_reentrancy_level_get(tsdn_tsd(tsdn)) == 0) {
+		witness_assert_depth_to_rank(tsdn_witness_tsdp_get(tsdn),
+		    WITNESS_RANK_CORE, 0);
+	}
+
+	ret = arena_malloc(tsdn, arena, size, ind, zero, tcache, slow_path);
+	if (config_stats && is_internal && likely(ret != NULL)) {
+		arena_internal_add(iaalloc(tsdn, ret), isalloc(tsdn, ret));
+	}
+	return ret;
+}
+
+JEMALLOC_ALWAYS_INLINE void *
+ialloc(tsd_t *tsd, size_t size, szind_t ind, bool zero, bool slow_path) {
+	return iallocztm(tsd_tsdn(tsd), size, ind, zero, tcache_get(tsd), false,
+	    NULL, slow_path);
+}
+
+JEMALLOC_ALWAYS_INLINE void *
+ipallocztm(tsdn_t *tsdn, size_t usize, size_t alignment, bool zero,
+    tcache_t *tcache, bool is_internal, arena_t *arena) {
+	void *ret;
+
+	assert(usize != 0);
+	assert(usize == sz_sa2u(usize, alignment));
+	assert(!is_internal || tcache == NULL);
+	assert(!is_internal || arena == NULL || arena_is_auto(arena));
+	witness_assert_depth_to_rank(tsdn_witness_tsdp_get(tsdn),
+	    WITNESS_RANK_CORE, 0);
+
+	ret = arena_palloc(tsdn, arena, usize, alignment, zero, tcache);
+	assert(ALIGNMENT_ADDR2BASE(ret, alignment) == ret);
+	if (config_stats && is_internal && likely(ret != NULL)) {
+		arena_internal_add(iaalloc(tsdn, ret), isalloc(tsdn, ret));
+	}
+	return ret;
+}
+
+JEMALLOC_ALWAYS_INLINE void *
+ipalloct(tsdn_t *tsdn, size_t usize, size_t alignment, bool zero,
+    tcache_t *tcache, arena_t *arena) {
+	return ipallocztm(tsdn, usize, alignment, zero, tcache, false, arena);
+}
+
+JEMALLOC_ALWAYS_INLINE void *
+ipalloc(tsd_t *tsd, size_t usize, size_t alignment, bool zero) {
+	return ipallocztm(tsd_tsdn(tsd), usize, alignment, zero,
+	    tcache_get(tsd), false, NULL);
+}
+
+JEMALLOC_ALWAYS_INLINE size_t
+ivsalloc(tsdn_t *tsdn, const void *ptr) {
+	return arena_vsalloc(tsdn, ptr);
+}
+
+JEMALLOC_ALWAYS_INLINE void
+idalloctm(tsdn_t *tsdn, void *ptr, tcache_t *tcache,
+    emap_alloc_ctx_t *alloc_ctx, bool is_internal, bool slow_path) {
+	assert(ptr != NULL);
+	assert(!is_internal || tcache == NULL);
+	assert(!is_internal || arena_is_auto(iaalloc(tsdn, ptr)));
+	witness_assert_depth_to_rank(tsdn_witness_tsdp_get(tsdn),
+	    WITNESS_RANK_CORE, 0);
+	if (config_stats && is_internal) {
+		arena_internal_sub(iaalloc(tsdn, ptr), isalloc(tsdn, ptr));
+	}
+	if (!is_internal && !tsdn_null(tsdn) &&
+	    tsd_reentrancy_level_get(tsdn_tsd(tsdn)) != 0) {
+		assert(tcache == NULL);
+	}
+	arena_dalloc(tsdn, ptr, tcache, alloc_ctx, slow_path);
+}
+
+JEMALLOC_ALWAYS_INLINE void
+idalloc(tsd_t *tsd, void *ptr) {
+	idalloctm(tsd_tsdn(tsd), ptr, tcache_get(tsd), NULL, false, true);
+}
+
+JEMALLOC_ALWAYS_INLINE void
+isdalloct(tsdn_t *tsdn, void *ptr, size_t size, tcache_t *tcache,
+    emap_alloc_ctx_t *alloc_ctx, bool slow_path) {
+	witness_assert_depth_to_rank(tsdn_witness_tsdp_get(tsdn),
+	    WITNESS_RANK_CORE, 0);
+	arena_sdalloc(tsdn, ptr, size, tcache, alloc_ctx, slow_path);
+}
+
+JEMALLOC_ALWAYS_INLINE void *
+iralloct_realign(tsdn_t *tsdn, void *ptr, size_t oldsize, size_t size,
+    size_t alignment, bool zero, tcache_t *tcache, arena_t *arena,
+    hook_ralloc_args_t *hook_args) {
+	witness_assert_depth_to_rank(tsdn_witness_tsdp_get(tsdn),
+	    WITNESS_RANK_CORE, 0);
+	void *p;
+	size_t usize, copysize;
+
+	usize = sz_sa2u(size, alignment);
+	if (unlikely(usize == 0 || usize > SC_LARGE_MAXCLASS)) {
+		return NULL;
+	}
+	p = ipalloct(tsdn, usize, alignment, zero, tcache, arena);
+	if (p == NULL) {
+		return NULL;
+	}
+	/*
+	 * Copy at most size bytes (not size+extra), since the caller has no
+	 * expectation that the extra bytes will be reliably preserved.
+	 */
+	copysize = (size < oldsize) ? size : oldsize;
+	memcpy(p, ptr, copysize);
+	hook_invoke_alloc(hook_args->is_realloc
+	    ? hook_alloc_realloc : hook_alloc_rallocx, p, (uintptr_t)p,
+	    hook_args->args);
+	hook_invoke_dalloc(hook_args->is_realloc
+	    ? hook_dalloc_realloc : hook_dalloc_rallocx, ptr, hook_args->args);
+	isdalloct(tsdn, ptr, oldsize, tcache, NULL, true);
+	return p;
+}
+
+/*
+ * is_realloc threads through the knowledge of whether or not this call comes
+ * from je_realloc (as opposed to je_rallocx); this ensures that we pass the
+ * correct entry point into any hooks.
+ * Note that these functions are all force-inlined, so no actual bool gets
+ * passed-around anywhere.
+ */
+JEMALLOC_ALWAYS_INLINE void *
+iralloct(tsdn_t *tsdn, void *ptr, size_t oldsize, size_t size, size_t alignment,
+    bool zero, tcache_t *tcache, arena_t *arena, hook_ralloc_args_t *hook_args)
+{
+	assert(ptr != NULL);
+	assert(size != 0);
+	witness_assert_depth_to_rank(tsdn_witness_tsdp_get(tsdn),
+	    WITNESS_RANK_CORE, 0);
+
+	if (alignment != 0 && ((uintptr_t)ptr & ((uintptr_t)alignment-1))
+	    != 0) {
+		/*
+		 * Existing object alignment is inadequate; allocate new space
+		 * and copy.
+		 */
+		return iralloct_realign(tsdn, ptr, oldsize, size, alignment,
+		    zero, tcache, arena, hook_args);
+	}
+
+	return arena_ralloc(tsdn, arena, ptr, oldsize, size, alignment, zero,
+	    tcache, hook_args);
+}
+
+JEMALLOC_ALWAYS_INLINE void *
+iralloc(tsd_t *tsd, void *ptr, size_t oldsize, size_t size, size_t alignment,
+    bool zero, hook_ralloc_args_t *hook_args) {
+	return iralloct(tsd_tsdn(tsd), ptr, oldsize, size, alignment, zero,
+	    tcache_get(tsd), NULL, hook_args);
+}
+
+JEMALLOC_ALWAYS_INLINE bool
+ixalloc(tsdn_t *tsdn, void *ptr, size_t oldsize, size_t size, size_t extra,
+    size_t alignment, bool zero, size_t *newsize) {
+	assert(ptr != NULL);
+	assert(size != 0);
+	witness_assert_depth_to_rank(tsdn_witness_tsdp_get(tsdn),
+	    WITNESS_RANK_CORE, 0);
+
+	if (alignment != 0 && ((uintptr_t)ptr & ((uintptr_t)alignment-1))
+	    != 0) {
+		/* Existing object alignment is inadequate. */
+		*newsize = oldsize;
+		return true;
+	}
+
+	return arena_ralloc_no_move(tsdn, ptr, oldsize, size, extra, zero,
+	    newsize);
+}
+
+JEMALLOC_ALWAYS_INLINE void
+fastpath_success_finish(tsd_t *tsd, uint64_t allocated_after,
+    cache_bin_t *bin, void *ret) {
+	thread_allocated_set(tsd, allocated_after);
+	if (config_stats) {
+		bin->tstats.nrequests++;
+	}
+
+	LOG("core.malloc.exit", "result: %p", ret);
+}
+
+JEMALLOC_ALWAYS_INLINE bool
+malloc_initialized(void) {
+	return (malloc_init_state == malloc_init_initialized);
+}
+
+/*
+ * malloc() fastpath.  Included here so that we can inline it into operator new;
+ * function call overhead there is non-negligible as a fraction of total CPU in
+ * allocation-heavy C++ programs.  We take the fallback alloc to allow malloc
+ * (which can return NULL) to differ in its behavior from operator new (which
+ * can't).  It matches the signature of malloc / operator new so that we can
+ * tail-call the fallback allocator, allowing us to avoid setting up the call
+ * frame in the common case.
+ *
+ * Fastpath assumes size <= SC_LOOKUP_MAXCLASS, and that we hit
+ * tcache.  If either of these is false, we tail-call to the slowpath,
+ * malloc_default().  Tail-calling is used to avoid any caller-saved
+ * registers.
+ *
+ * fastpath supports ticker and profiling, both of which will also
+ * tail-call to the slowpath if they fire.
+ */
+JEMALLOC_ALWAYS_INLINE void *
+imalloc_fastpath(size_t size, void *(fallback_alloc)(size_t)) {
+	LOG("core.malloc.entry", "size: %zu", size);
+	if (tsd_get_allocates() && unlikely(!malloc_initialized())) {
+		return fallback_alloc(size);
+	}
+
+	tsd_t *tsd = tsd_get(false);
+	if (unlikely((size > SC_LOOKUP_MAXCLASS) || tsd == NULL)) {
+		return fallback_alloc(size);
+	}
+	/*
+	 * The code below till the branch checking the next_event threshold may
+	 * execute before malloc_init(), in which case the threshold is 0 to
+	 * trigger slow path and initialization.
+	 *
+	 * Note that when uninitialized, only the fast-path variants of the sz /
+	 * tsd facilities may be called.
+	 */
+	szind_t ind;
+	/*
+	 * The thread_allocated counter in tsd serves as a general purpose
+	 * accumulator for bytes of allocation to trigger different types of
+	 * events.  usize is always needed to advance thread_allocated, though
+	 * it's not always needed in the core allocation logic.
+	 */
+	size_t usize;
+	sz_size2index_usize_fastpath(size, &ind, &usize);
+	/* Fast path relies on size being a bin. */
+	assert(ind < SC_NBINS);
+	assert((SC_LOOKUP_MAXCLASS < SC_SMALL_MAXCLASS) &&
+	    (size <= SC_SMALL_MAXCLASS));
+
+	uint64_t allocated, threshold;
+	te_malloc_fastpath_ctx(tsd, &allocated, &threshold);
+	uint64_t allocated_after = allocated + usize;
+	/*
+	 * The ind and usize might be uninitialized (or partially) before
+	 * malloc_init().  The assertions check for: 1) full correctness (usize
+	 * & ind) when initialized; and 2) guaranteed slow-path (threshold == 0)
+	 * when !initialized.
+	 */
+	if (!malloc_initialized()) {
+		assert(threshold == 0);
+	} else {
+		assert(ind == sz_size2index(size));
+		assert(usize > 0 && usize == sz_index2size(ind));
+	}
+	/*
+	 * Check for events and tsd non-nominal (fast_threshold will be set to
+	 * 0) in a single branch.
+	 */
+	if (unlikely(allocated_after >= threshold)) {
+		return fallback_alloc(size);
+	}
+	assert(tsd_fast(tsd));
+
+	tcache_t *tcache = tsd_tcachep_get(tsd);
+	assert(tcache == tcache_get(tsd));
+	cache_bin_t *bin = &tcache->bins[ind];
+	bool tcache_success;
+	void *ret;
+
+	/*
+	 * We split up the code this way so that redundant low-water
+	 * computation doesn't happen on the (more common) case in which we
+	 * don't touch the low water mark.  The compiler won't do this
+	 * duplication on its own.
+	 */
+	ret = cache_bin_alloc_easy(bin, &tcache_success);
+	if (tcache_success) {
+		fastpath_success_finish(tsd, allocated_after, bin, ret);
+		return ret;
+	}
+	ret = cache_bin_alloc(bin, &tcache_success);
+	if (tcache_success) {
+		fastpath_success_finish(tsd, allocated_after, bin, ret);
+		return ret;
+	}
+
+	return fallback_alloc(size);
+}
+
+#endif /* JEMALLOC_INTERNAL_INLINES_C_H */

BeefRT/JEMalloc/include/jemalloc/internal/jemalloc_internal_macros.h

@@ -0,0 +1,111 @@
+#ifndef JEMALLOC_INTERNAL_MACROS_H
+#define JEMALLOC_INTERNAL_MACROS_H
+
+#ifdef JEMALLOC_DEBUG
+#  define JEMALLOC_ALWAYS_INLINE static inline
+#else
+#  ifdef _MSC_VER
+#    define JEMALLOC_ALWAYS_INLINE static __forceinline
+#  else
+#    define JEMALLOC_ALWAYS_INLINE JEMALLOC_ATTR(always_inline) static inline
+#  endif
+#endif
+#ifdef _MSC_VER
+#  define inline _inline
+#endif
+
+#define UNUSED JEMALLOC_ATTR(unused)
+
+#define ZU(z)	((size_t)z)
+#define ZD(z)	((ssize_t)z)
+#define QU(q)	((uint64_t)q)
+#define QD(q)	((int64_t)q)
+
+#define KZU(z)	ZU(z##ULL)
+#define KZD(z)	ZD(z##LL)
+#define KQU(q)	QU(q##ULL)
+#define KQD(q)	QI(q##LL)
+
+#ifndef __DECONST
+#  define	__DECONST(type, var)	((type)(uintptr_t)(const void *)(var))
+#endif
+
+#if !defined(JEMALLOC_HAS_RESTRICT) || defined(__cplusplus)
+#  define restrict
+#endif
+
+/* Various function pointers are static and immutable except during testing. */
+#ifdef JEMALLOC_JET
+#  define JET_MUTABLE
+#else
+#  define JET_MUTABLE const
+#endif
+
+#define JEMALLOC_VA_ARGS_HEAD(head, ...) head
+#define JEMALLOC_VA_ARGS_TAIL(head, ...) __VA_ARGS__
+
+/* Diagnostic suppression macros */
+#if defined(_MSC_VER) && !defined(__clang__)
+#  define JEMALLOC_DIAGNOSTIC_PUSH __pragma(warning(push))
+#  define JEMALLOC_DIAGNOSTIC_POP __pragma(warning(pop))
+#  define JEMALLOC_DIAGNOSTIC_IGNORE(W) __pragma(warning(disable:W))
+#  define JEMALLOC_DIAGNOSTIC_IGNORE_MISSING_STRUCT_FIELD_INITIALIZERS
+#  define JEMALLOC_DIAGNOSTIC_IGNORE_TYPE_LIMITS
+#  define JEMALLOC_DIAGNOSTIC_IGNORE_ALLOC_SIZE_LARGER_THAN
+#  define JEMALLOC_DIAGNOSTIC_DISABLE_SPURIOUS
+/* #pragma GCC diagnostic first appeared in gcc 4.6. */
+#elif (defined(__GNUC__) && ((__GNUC__ > 4) || ((__GNUC__ == 4) && \
+  (__GNUC_MINOR__ > 5)))) || defined(__clang__)
+/*
+ * The JEMALLOC_PRAGMA__ macro is an implementation detail of the GCC and Clang
+ * diagnostic suppression macros and should not be used anywhere else.
+ */
+#  define JEMALLOC_PRAGMA__(X) _Pragma(#X)
+#  define JEMALLOC_DIAGNOSTIC_PUSH JEMALLOC_PRAGMA__(GCC diagnostic push)
+#  define JEMALLOC_DIAGNOSTIC_POP JEMALLOC_PRAGMA__(GCC diagnostic pop)
+#  define JEMALLOC_DIAGNOSTIC_IGNORE(W) \
+     JEMALLOC_PRAGMA__(GCC diagnostic ignored W)
+
+/*
+ * The -Wmissing-field-initializers warning is buggy in GCC versions < 5.1 and
+ * all clang versions up to version 7 (currently trunk, unreleased).  This macro
+ * suppresses the warning for the affected compiler versions only.
+ */
+#  if ((defined(__GNUC__) && !defined(__clang__)) && (__GNUC__ < 5)) || \
+     defined(__clang__)
+#    define JEMALLOC_DIAGNOSTIC_IGNORE_MISSING_STRUCT_FIELD_INITIALIZERS  \
+          JEMALLOC_DIAGNOSTIC_IGNORE("-Wmissing-field-initializers")
+#  else
+#    define JEMALLOC_DIAGNOSTIC_IGNORE_MISSING_STRUCT_FIELD_INITIALIZERS
+#  endif
+
+#  define JEMALLOC_DIAGNOSTIC_IGNORE_TYPE_LIMITS  \
+     JEMALLOC_DIAGNOSTIC_IGNORE("-Wtype-limits")
+#  define JEMALLOC_DIAGNOSTIC_IGNORE_UNUSED_PARAMETER \
+     JEMALLOC_DIAGNOSTIC_IGNORE("-Wunused-parameter")
+#  if defined(__GNUC__) && !defined(__clang__) && (__GNUC__ >= 7)
+#    define JEMALLOC_DIAGNOSTIC_IGNORE_ALLOC_SIZE_LARGER_THAN \
+       JEMALLOC_DIAGNOSTIC_IGNORE("-Walloc-size-larger-than=")
+#  else
+#    define JEMALLOC_DIAGNOSTIC_IGNORE_ALLOC_SIZE_LARGER_THAN
+#  endif
+#  define JEMALLOC_DIAGNOSTIC_DISABLE_SPURIOUS \
+  JEMALLOC_DIAGNOSTIC_PUSH \
+  JEMALLOC_DIAGNOSTIC_IGNORE_UNUSED_PARAMETER
+#else
+#  define JEMALLOC_DIAGNOSTIC_PUSH
+#  define JEMALLOC_DIAGNOSTIC_POP
+#  define JEMALLOC_DIAGNOSTIC_IGNORE(W)
+#  define JEMALLOC_DIAGNOSTIC_IGNORE_MISSING_STRUCT_FIELD_INITIALIZERS
+#  define JEMALLOC_DIAGNOSTIC_IGNORE_TYPE_LIMITS
+#  define JEMALLOC_DIAGNOSTIC_IGNORE_ALLOC_SIZE_LARGER_THAN
+#  define JEMALLOC_DIAGNOSTIC_DISABLE_SPURIOUS
+#endif
+
+/*
+ * Disables spurious diagnostics for all headers.  Since these headers are not
+ * included by users directly, it does not affect their diagnostic settings.
+ */
+JEMALLOC_DIAGNOSTIC_DISABLE_SPURIOUS
+
+#endif /* JEMALLOC_INTERNAL_MACROS_H */

BeefRT/JEMalloc/include/jemalloc/internal/jemalloc_internal_types.h

@@ -0,0 +1,130 @@
+#ifndef JEMALLOC_INTERNAL_TYPES_H
+#define JEMALLOC_INTERNAL_TYPES_H
+
+#include "jemalloc/internal/quantum.h"
+
+/* Processor / core id type. */
+typedef int malloc_cpuid_t;
+
+/* When realloc(non-null-ptr, 0) is called, what happens? */
+enum zero_realloc_action_e {
+	/* Realloc(ptr, 0) is free(ptr); return malloc(0); */
+	zero_realloc_action_alloc = 0,
+	/* Realloc(ptr, 0) is free(ptr); */
+	zero_realloc_action_free = 1,
+	/* Realloc(ptr, 0) aborts. */
+	zero_realloc_action_abort = 2
+};
+typedef enum zero_realloc_action_e zero_realloc_action_t;
+
+/* Signature of write callback. */
+typedef void (write_cb_t)(void *, const char *);
+
+enum malloc_init_e {
+	malloc_init_uninitialized	= 3,
+	malloc_init_a0_initialized	= 2,
+	malloc_init_recursible		= 1,
+	malloc_init_initialized		= 0 /* Common case --> jnz. */
+};
+typedef enum malloc_init_e malloc_init_t;
+
+/*
+ * Flags bits:
+ *
+ * a: arena
+ * t: tcache
+ * 0: unused
+ * z: zero
+ * n: alignment
+ *
+ * aaaaaaaa aaaatttt tttttttt 0znnnnnn
+ */
+#define MALLOCX_ARENA_BITS	12
+#define MALLOCX_TCACHE_BITS	12
+#define MALLOCX_LG_ALIGN_BITS	6
+#define MALLOCX_ARENA_SHIFT	20
+#define MALLOCX_TCACHE_SHIFT	8
+#define MALLOCX_ARENA_MASK \
+    (((1 << MALLOCX_ARENA_BITS) - 1) << MALLOCX_ARENA_SHIFT)
+/* NB: Arena index bias decreases the maximum number of arenas by 1. */
+#define MALLOCX_ARENA_LIMIT	((1 << MALLOCX_ARENA_BITS) - 1)
+#define MALLOCX_TCACHE_MASK \
+    (((1 << MALLOCX_TCACHE_BITS) - 1) << MALLOCX_TCACHE_SHIFT)
+#define MALLOCX_TCACHE_MAX	((1 << MALLOCX_TCACHE_BITS) - 3)
+#define MALLOCX_LG_ALIGN_MASK	((1 << MALLOCX_LG_ALIGN_BITS) - 1)
+/* Use MALLOCX_ALIGN_GET() if alignment may not be specified in flags. */
+#define MALLOCX_ALIGN_GET_SPECIFIED(flags)				\
+    (ZU(1) << (flags & MALLOCX_LG_ALIGN_MASK))
+#define MALLOCX_ALIGN_GET(flags)					\
+    (MALLOCX_ALIGN_GET_SPECIFIED(flags) & (SIZE_T_MAX-1))
+#define MALLOCX_ZERO_GET(flags)						\
+    ((bool)(flags & MALLOCX_ZERO))
+
+#define MALLOCX_TCACHE_GET(flags)					\
+    (((unsigned)((flags & MALLOCX_TCACHE_MASK) >> MALLOCX_TCACHE_SHIFT)) - 2)
+#define MALLOCX_ARENA_GET(flags)					\
+    (((unsigned)(((unsigned)flags) >> MALLOCX_ARENA_SHIFT)) - 1)
+
+/* Smallest size class to support. */
+#define TINY_MIN		(1U << LG_TINY_MIN)
+
+#define LONG			((size_t)(1U << LG_SIZEOF_LONG))
+#define LONG_MASK		(LONG - 1)
+
+/* Return the smallest long multiple that is >= a. */
+#define LONG_CEILING(a)							\
+	(((a) + LONG_MASK) & ~LONG_MASK)
+
+#define SIZEOF_PTR		(1U << LG_SIZEOF_PTR)
+#define PTR_MASK		(SIZEOF_PTR - 1)
+
+/* Return the smallest (void *) multiple that is >= a. */
+#define PTR_CEILING(a)							\
+	(((a) + PTR_MASK) & ~PTR_MASK)
+
+/*
+ * Maximum size of L1 cache line.  This is used to avoid cache line aliasing.
+ * In addition, this controls the spacing of cacheline-spaced size classes.
+ *
+ * CACHELINE cannot be based on LG_CACHELINE because __declspec(align()) can
+ * only handle raw constants.
+ */
+#define LG_CACHELINE		6
+#define CACHELINE		64
+#define CACHELINE_MASK		(CACHELINE - 1)
+
+/* Return the smallest cacheline multiple that is >= s. */
+#define CACHELINE_CEILING(s)						\
+	(((s) + CACHELINE_MASK) & ~CACHELINE_MASK)
+
+/* Return the nearest aligned address at or below a. */
+#define ALIGNMENT_ADDR2BASE(a, alignment)				\
+	((void *)((uintptr_t)(a) & ((~(alignment)) + 1)))
+
+/* Return the offset between a and the nearest aligned address at or below a. */
+#define ALIGNMENT_ADDR2OFFSET(a, alignment)				\
+	((size_t)((uintptr_t)(a) & (alignment - 1)))
+
+/* Return the smallest alignment multiple that is >= s. */
+#define ALIGNMENT_CEILING(s, alignment)					\
+	(((s) + (alignment - 1)) & ((~(alignment)) + 1))
+
+/* Declare a variable-length array. */
+#if __STDC_VERSION__ < 199901L
+#  ifdef _MSC_VER
+#    include <malloc.h>
+#    define alloca _alloca
+#  else
+#    ifdef JEMALLOC_HAS_ALLOCA_H
+#      include <alloca.h>
+#    else
+#      include <stdlib.h>
+#    endif
+#  endif
+#  define VARIABLE_ARRAY(type, name, count) \
+	type *name = alloca(sizeof(type) * (count))
+#else
+#  define VARIABLE_ARRAY(type, name, count) type name[(count)]
+#endif
+
+#endif /* JEMALLOC_INTERNAL_TYPES_H */

BeefRT/JEMalloc/include/jemalloc/internal/jemalloc_preamble.h

@@ -0,0 +1,263 @@
+#ifndef JEMALLOC_PREAMBLE_H
+#define JEMALLOC_PREAMBLE_H
+
+#include "jemalloc_internal_defs.h"
+#include "jemalloc/internal/jemalloc_internal_decls.h"
+
+#if defined(JEMALLOC_UTRACE) || defined(JEMALLOC_UTRACE_LABEL)
+#include <sys/ktrace.h>
+#  if defined(JEMALLOC_UTRACE)
+#    define UTRACE_CALL(p, l) utrace(p, l)
+#  else
+#    define UTRACE_CALL(p, l) utrace("jemalloc_process", p, l)
+#    define JEMALLOC_UTRACE
+#  endif
+#endif
+
+#define JEMALLOC_NO_DEMANGLE
+#ifdef JEMALLOC_JET
+#  undef JEMALLOC_IS_MALLOC
+#  define JEMALLOC_N(n) jet_##n
+#  include "jemalloc/internal/public_namespace.h"
+#  define JEMALLOC_NO_RENAME
+#  include "../jemalloc.h"
+#  undef JEMALLOC_NO_RENAME
+#else
+#  define JEMALLOC_N(n) je_##n
+#  include "../jemalloc.h"
+#endif
+
+#if defined(JEMALLOC_OSATOMIC)
+#include <libkern/OSAtomic.h>
+#endif
+
+#ifdef JEMALLOC_ZONE
+#include <mach/mach_error.h>
+#include <mach/mach_init.h>
+#include <mach/vm_map.h>
+#endif
+
+#include "jemalloc/internal/jemalloc_internal_macros.h"
+
+/*
+ * Note that the ordering matters here; the hook itself is name-mangled.  We
+ * want the inclusion of hooks to happen early, so that we hook as much as
+ * possible.
+ */
+#ifndef JEMALLOC_NO_PRIVATE_NAMESPACE
+#  ifndef JEMALLOC_JET
+#    include "jemalloc/internal/private_namespace.h"
+#  else
+#    include "jemalloc/internal/private_namespace_jet.h"
+#  endif
+#endif
+#include "jemalloc/internal/test_hooks.h"
+
+#ifdef JEMALLOC_DEFINE_MADVISE_FREE
+#  define JEMALLOC_MADV_FREE 8
+#endif
+
+static const bool config_debug =
+#ifdef JEMALLOC_DEBUG
+    true
+#else
+    false
+#endif
+    ;
+static const bool have_dss =
+#ifdef JEMALLOC_DSS
+    true
+#else
+    false
+#endif
+    ;
+static const bool have_madvise_huge =
+#ifdef JEMALLOC_HAVE_MADVISE_HUGE
+    true
+#else
+    false
+#endif
+    ;
+static const bool config_fill =
+#ifdef JEMALLOC_FILL
+    true
+#else
+    false
+#endif
+    ;
+static const bool config_lazy_lock =
+#ifdef JEMALLOC_LAZY_LOCK
+    true
+#else
+    false
+#endif
+    ;
+static const char * const config_malloc_conf = JEMALLOC_CONFIG_MALLOC_CONF;
+static const bool config_prof =
+#ifdef JEMALLOC_PROF
+    true
+#else
+    false
+#endif
+    ;
+static const bool config_prof_libgcc =
+#ifdef JEMALLOC_PROF_LIBGCC
+    true
+#else
+    false
+#endif
+    ;
+static const bool config_prof_libunwind =
+#ifdef JEMALLOC_PROF_LIBUNWIND
+    true
+#else
+    false
+#endif
+    ;
+static const bool maps_coalesce =
+#ifdef JEMALLOC_MAPS_COALESCE
+    true
+#else
+    false
+#endif
+    ;
+static const bool config_stats =
+#ifdef JEMALLOC_STATS
+    true
+#else
+    false
+#endif
+    ;
+static const bool config_tls =
+#ifdef JEMALLOC_TLS
+    true
+#else
+    false
+#endif
+    ;
+static const bool config_utrace =
+#ifdef JEMALLOC_UTRACE
+    true
+#else
+    false
+#endif
+    ;
+static const bool config_xmalloc =
+#ifdef JEMALLOC_XMALLOC
+    true
+#else
+    false
+#endif
+    ;
+static const bool config_cache_oblivious =
+#ifdef JEMALLOC_CACHE_OBLIVIOUS
+    true
+#else
+    false
+#endif
+    ;
+/*
+ * Undocumented, for jemalloc development use only at the moment.  See the note
+ * in jemalloc/internal/log.h.
+ */
+static const bool config_log =
+#ifdef JEMALLOC_LOG
+    true
+#else
+    false
+#endif
+    ;
+/*
+ * Are extra safety checks enabled; things like checking the size of sized
+ * deallocations, double-frees, etc.
+ */
+static const bool config_opt_safety_checks =
+#ifdef JEMALLOC_OPT_SAFETY_CHECKS
+    true
+#elif defined(JEMALLOC_DEBUG)
+    /*
+     * This lets us only guard safety checks by one flag instead of two; fast
+     * checks can guard solely by config_opt_safety_checks and run in debug mode
+     * too.
+     */
+    true
+#else
+    false
+#endif
+    ;
+
+/*
+ * Extra debugging of sized deallocations too onerous to be included in the
+ * general safety checks.
+ */
+static const bool config_opt_size_checks =
+#if defined(JEMALLOC_OPT_SIZE_CHECKS) || defined(JEMALLOC_DEBUG)
+    true
+#else
+    false
+#endif
+    ;
+
+static const bool config_uaf_detection =
+#if defined(JEMALLOC_UAF_DETECTION) || defined(JEMALLOC_DEBUG)
+    true
+#else
+    false
+#endif
+    ;
+
+/* Whether or not the C++ extensions are enabled. */
+static const bool config_enable_cxx =
+#ifdef JEMALLOC_ENABLE_CXX
+    true
+#else
+    false
+#endif
+;
+
+#if defined(_WIN32) || defined(JEMALLOC_HAVE_SCHED_GETCPU)
+/* Currently percpu_arena depends on sched_getcpu. */
+#define JEMALLOC_PERCPU_ARENA
+#endif
+static const bool have_percpu_arena =
+#ifdef JEMALLOC_PERCPU_ARENA
+    true
+#else
+    false
+#endif
+    ;
+/*
+ * Undocumented, and not recommended; the application should take full
+ * responsibility for tracking provenance.
+ */
+static const bool force_ivsalloc =
+#ifdef JEMALLOC_FORCE_IVSALLOC
+    true
+#else
+    false
+#endif
+    ;
+static const bool have_background_thread =
+#ifdef JEMALLOC_BACKGROUND_THREAD
+    true
+#else
+    false
+#endif
+    ;
+static const bool config_high_res_timer =
+#ifdef JEMALLOC_HAVE_CLOCK_REALTIME
+    true
+#else
+    false
+#endif
+    ;
+
+static const bool have_memcntl =
+#ifdef JEMALLOC_HAVE_MEMCNTL
+    true
+#else
+    false
+#endif
+    ;
+
+#endif /* JEMALLOC_PREAMBLE_H */

BeefRT/JEMalloc/include/jemalloc/internal/jemalloc_preamble.h.in

@@ -0,0 +1,263 @@
+#ifndef JEMALLOC_PREAMBLE_H
+#define JEMALLOC_PREAMBLE_H
+
+#include "jemalloc_internal_defs.h"
+#include "jemalloc/internal/jemalloc_internal_decls.h"
+
+#if defined(JEMALLOC_UTRACE) || defined(JEMALLOC_UTRACE_LABEL)
+#include <sys/ktrace.h>
+#  if defined(JEMALLOC_UTRACE)
+#    define UTRACE_CALL(p, l) utrace(p, l)
+#  else
+#    define UTRACE_CALL(p, l) utrace("jemalloc_process", p, l)
+#    define JEMALLOC_UTRACE
+#  endif
+#endif
+
+#define JEMALLOC_NO_DEMANGLE
+#ifdef JEMALLOC_JET
+#  undef JEMALLOC_IS_MALLOC
+#  define JEMALLOC_N(n) jet_##n
+#  include "jemalloc/internal/public_namespace.h"
+#  define JEMALLOC_NO_RENAME
+#  include "../jemalloc@install_suffix@.h"
+#  undef JEMALLOC_NO_RENAME
+#else
+#  define JEMALLOC_N(n) @private_namespace@##n
+#  include "../jemalloc@install_suffix@.h"
+#endif
+
+#if defined(JEMALLOC_OSATOMIC)
+#include <libkern/OSAtomic.h>
+#endif
+
+#ifdef JEMALLOC_ZONE
+#include <mach/mach_error.h>
+#include <mach/mach_init.h>
+#include <mach/vm_map.h>
+#endif
+
+#include "jemalloc/internal/jemalloc_internal_macros.h"
+
+/*
+ * Note that the ordering matters here; the hook itself is name-mangled.  We
+ * want the inclusion of hooks to happen early, so that we hook as much as
+ * possible.
+ */
+#ifndef JEMALLOC_NO_PRIVATE_NAMESPACE
+#  ifndef JEMALLOC_JET
+#    include "jemalloc/internal/private_namespace.h"
+#  else
+#    include "jemalloc/internal/private_namespace_jet.h"
+#  endif
+#endif
+#include "jemalloc/internal/test_hooks.h"
+
+#ifdef JEMALLOC_DEFINE_MADVISE_FREE
+#  define JEMALLOC_MADV_FREE 8
+#endif
+
+static const bool config_debug =
+#ifdef JEMALLOC_DEBUG
+    true
+#else
+    false
+#endif
+    ;
+static const bool have_dss =
+#ifdef JEMALLOC_DSS
+    true
+#else
+    false
+#endif
+    ;
+static const bool have_madvise_huge =
+#ifdef JEMALLOC_HAVE_MADVISE_HUGE
+    true
+#else
+    false
+#endif
+    ;
+static const bool config_fill =
+#ifdef JEMALLOC_FILL
+    true
+#else
+    false
+#endif
+    ;
+static const bool config_lazy_lock =
+#ifdef JEMALLOC_LAZY_LOCK
+    true
+#else
+    false
+#endif
+    ;
+static const char * const config_malloc_conf = JEMALLOC_CONFIG_MALLOC_CONF;
+static const bool config_prof =
+#ifdef JEMALLOC_PROF
+    true
+#else
+    false
+#endif
+    ;
+static const bool config_prof_libgcc =
+#ifdef JEMALLOC_PROF_LIBGCC
+    true
+#else
+    false
+#endif
+    ;
+static const bool config_prof_libunwind =
+#ifdef JEMALLOC_PROF_LIBUNWIND
+    true
+#else
+    false
+#endif
+    ;
+static const bool maps_coalesce =
+#ifdef JEMALLOC_MAPS_COALESCE
+    true
+#else
+    false
+#endif
+    ;
+static const bool config_stats =
+#ifdef JEMALLOC_STATS
+    true
+#else
+    false
+#endif
+    ;
+static const bool config_tls =
+#ifdef JEMALLOC_TLS
+    true
+#else
+    false
+#endif
+    ;
+static const bool config_utrace =
+#ifdef JEMALLOC_UTRACE
+    true
+#else
+    false
+#endif
+    ;
+static const bool config_xmalloc =
+#ifdef JEMALLOC_XMALLOC
+    true
+#else
+    false
+#endif
+    ;
+static const bool config_cache_oblivious =
+#ifdef JEMALLOC_CACHE_OBLIVIOUS
+    true
+#else
+    false
+#endif
+    ;
+/*
+ * Undocumented, for jemalloc development use only at the moment.  See the note
+ * in jemalloc/internal/log.h.
+ */
+static const bool config_log =
+#ifdef JEMALLOC_LOG
+    true
+#else
+    false
+#endif
+    ;
+/*
+ * Are extra safety checks enabled; things like checking the size of sized
+ * deallocations, double-frees, etc.
+ */
+static const bool config_opt_safety_checks =
+#ifdef JEMALLOC_OPT_SAFETY_CHECKS
+    true
+#elif defined(JEMALLOC_DEBUG)
+    /*
+     * This lets us only guard safety checks by one flag instead of two; fast
+     * checks can guard solely by config_opt_safety_checks and run in debug mode
+     * too.
+     */
+    true
+#else
+    false
+#endif
+    ;
+
+/*
+ * Extra debugging of sized deallocations too onerous to be included in the
+ * general safety checks.
+ */
+static const bool config_opt_size_checks =
+#if defined(JEMALLOC_OPT_SIZE_CHECKS) || defined(JEMALLOC_DEBUG)
+    true
+#else
+    false
+#endif
+    ;
+
+static const bool config_uaf_detection =
+#if defined(JEMALLOC_UAF_DETECTION) || defined(JEMALLOC_DEBUG)
+    true
+#else
+    false
+#endif
+    ;
+
+/* Whether or not the C++ extensions are enabled. */
+static const bool config_enable_cxx =
+#ifdef JEMALLOC_ENABLE_CXX
+    true
+#else
+    false
+#endif
+;
+
+#if defined(_WIN32) || defined(JEMALLOC_HAVE_SCHED_GETCPU)
+/* Currently percpu_arena depends on sched_getcpu. */
+#define JEMALLOC_PERCPU_ARENA
+#endif
+static const bool have_percpu_arena =
+#ifdef JEMALLOC_PERCPU_ARENA
+    true
+#else
+    false
+#endif
+    ;
+/*
+ * Undocumented, and not recommended; the application should take full
+ * responsibility for tracking provenance.
+ */
+static const bool force_ivsalloc =
+#ifdef JEMALLOC_FORCE_IVSALLOC
+    true
+#else
+    false
+#endif
+    ;
+static const bool have_background_thread =
+#ifdef JEMALLOC_BACKGROUND_THREAD
+    true
+#else
+    false
+#endif
+    ;
+static const bool config_high_res_timer =
+#ifdef JEMALLOC_HAVE_CLOCK_REALTIME
+    true
+#else
+    false
+#endif
+    ;
+
+static const bool have_memcntl =
+#ifdef JEMALLOC_HAVE_MEMCNTL
+    true
+#else
+    false
+#endif
+    ;
+
+#endif /* JEMALLOC_PREAMBLE_H */

BeefRT/JEMalloc/include/jemalloc/internal/large_externs.h

@@ -0,0 +1,24 @@
+#ifndef JEMALLOC_INTERNAL_LARGE_EXTERNS_H
+#define JEMALLOC_INTERNAL_LARGE_EXTERNS_H
+
+#include "jemalloc/internal/hook.h"
+
+void *large_malloc(tsdn_t *tsdn, arena_t *arena, size_t usize, bool zero);
+void *large_palloc(tsdn_t *tsdn, arena_t *arena, size_t usize, size_t alignment,
+    bool zero);
+bool large_ralloc_no_move(tsdn_t *tsdn, edata_t *edata, size_t usize_min,
+    size_t usize_max, bool zero);
+void *large_ralloc(tsdn_t *tsdn, arena_t *arena, void *ptr, size_t usize,
+    size_t alignment, bool zero, tcache_t *tcache,
+    hook_ralloc_args_t *hook_args);
+
+void large_dalloc_prep_locked(tsdn_t *tsdn, edata_t *edata);
+void large_dalloc_finish(tsdn_t *tsdn, edata_t *edata);
+void large_dalloc(tsdn_t *tsdn, edata_t *edata);
+size_t large_salloc(tsdn_t *tsdn, const edata_t *edata);
+void large_prof_info_get(tsd_t *tsd, edata_t *edata, prof_info_t *prof_info,
+    bool reset_recent);
+void large_prof_tctx_reset(edata_t *edata);
+void large_prof_info_set(edata_t *edata, prof_tctx_t *tctx, size_t size);
+
+#endif /* JEMALLOC_INTERNAL_LARGE_EXTERNS_H */

BeefRT/JEMalloc/include/jemalloc/internal/lockedint.h

@@ -0,0 +1,204 @@
+#ifndef JEMALLOC_INTERNAL_LOCKEDINT_H
+#define JEMALLOC_INTERNAL_LOCKEDINT_H
+
+/*
+ * In those architectures that support 64-bit atomics, we use atomic updates for
+ * our 64-bit values.  Otherwise, we use a plain uint64_t and synchronize
+ * externally.
+ */
+
+typedef struct locked_u64_s locked_u64_t;
+#ifdef JEMALLOC_ATOMIC_U64
+struct locked_u64_s {
+	atomic_u64_t val;
+};
+#else
+/* Must hold the associated mutex. */
+struct locked_u64_s {
+	uint64_t val;
+};
+#endif
+
+typedef struct locked_zu_s locked_zu_t;
+struct locked_zu_s {
+	atomic_zu_t val;
+};
+
+#ifndef JEMALLOC_ATOMIC_U64
+#  define LOCKEDINT_MTX_DECLARE(name) malloc_mutex_t name;
+#  define LOCKEDINT_MTX_INIT(mu, name, rank, rank_mode)			\
+    malloc_mutex_init(&(mu), name, rank, rank_mode)
+#  define LOCKEDINT_MTX(mtx) (&(mtx))
+#  define LOCKEDINT_MTX_LOCK(tsdn, mu) malloc_mutex_lock(tsdn, &(mu))
+#  define LOCKEDINT_MTX_UNLOCK(tsdn, mu) malloc_mutex_unlock(tsdn, &(mu))
+#  define LOCKEDINT_MTX_PREFORK(tsdn, mu) malloc_mutex_prefork(tsdn, &(mu))
+#  define LOCKEDINT_MTX_POSTFORK_PARENT(tsdn, mu)			\
+    malloc_mutex_postfork_parent(tsdn, &(mu))
+#  define LOCKEDINT_MTX_POSTFORK_CHILD(tsdn, mu)			\
+    malloc_mutex_postfork_child(tsdn, &(mu))
+#else
+#  define LOCKEDINT_MTX_DECLARE(name)
+#  define LOCKEDINT_MTX(mtx) NULL
+#  define LOCKEDINT_MTX_INIT(mu, name, rank, rank_mode) false
+#  define LOCKEDINT_MTX_LOCK(tsdn, mu)
+#  define LOCKEDINT_MTX_UNLOCK(tsdn, mu)
+#  define LOCKEDINT_MTX_PREFORK(tsdn, mu)
+#  define LOCKEDINT_MTX_POSTFORK_PARENT(tsdn, mu)
+#  define LOCKEDINT_MTX_POSTFORK_CHILD(tsdn, mu)
+#endif
+
+#ifdef JEMALLOC_ATOMIC_U64
+#  define LOCKEDINT_MTX_ASSERT_INTERNAL(tsdn, mtx) assert((mtx) == NULL)
+#else
+#  define LOCKEDINT_MTX_ASSERT_INTERNAL(tsdn, mtx)			\
+    malloc_mutex_assert_owner(tsdn, (mtx))
+#endif
+
+static inline uint64_t
+locked_read_u64(tsdn_t *tsdn, malloc_mutex_t *mtx, locked_u64_t *p) {
+	LOCKEDINT_MTX_ASSERT_INTERNAL(tsdn, mtx);
+#ifdef JEMALLOC_ATOMIC_U64
+	return atomic_load_u64(&p->val, ATOMIC_RELAXED);
+#else
+	return p->val;
+#endif
+}
+
+static inline void
+locked_inc_u64(tsdn_t *tsdn, malloc_mutex_t *mtx, locked_u64_t *p,
+    uint64_t x) {
+	LOCKEDINT_MTX_ASSERT_INTERNAL(tsdn, mtx);
+#ifdef JEMALLOC_ATOMIC_U64
+	atomic_fetch_add_u64(&p->val, x, ATOMIC_RELAXED);
+#else
+	p->val += x;
+#endif
+}
+
+static inline void
+locked_dec_u64(tsdn_t *tsdn, malloc_mutex_t *mtx, locked_u64_t *p,
+    uint64_t x) {
+	LOCKEDINT_MTX_ASSERT_INTERNAL(tsdn, mtx);
+#ifdef JEMALLOC_ATOMIC_U64
+	uint64_t r = atomic_fetch_sub_u64(&p->val, x, ATOMIC_RELAXED);
+	assert(r - x <= r);
+#else
+	p->val -= x;
+	assert(p->val + x >= p->val);
+#endif
+}
+
+/* Increment and take modulus.  Returns whether the modulo made any change.  */
+static inline bool
+locked_inc_mod_u64(tsdn_t *tsdn, malloc_mutex_t *mtx, locked_u64_t *p,
+    const uint64_t x, const uint64_t modulus) {
+	LOCKEDINT_MTX_ASSERT_INTERNAL(tsdn, mtx);
+	uint64_t before, after;
+	bool overflow;
+#ifdef JEMALLOC_ATOMIC_U64
+	before = atomic_load_u64(&p->val, ATOMIC_RELAXED);
+	do {
+		after = before + x;
+		assert(after >= before);
+		overflow = (after >= modulus);
+		if (overflow) {
+			after %= modulus;
+		}
+	} while (!atomic_compare_exchange_weak_u64(&p->val, &before, after,
+	    ATOMIC_RELAXED, ATOMIC_RELAXED));
+#else
+	before = p->val;
+	after = before + x;
+	overflow = (after >= modulus);
+	if (overflow) {
+		after %= modulus;
+	}
+	p->val = after;
+#endif
+	return overflow;
+}
+
+/*
+ * Non-atomically sets *dst += src.  *dst needs external synchronization.
+ * This lets us avoid the cost of a fetch_add when its unnecessary (note that
+ * the types here are atomic).
+ */
+static inline void
+locked_inc_u64_unsynchronized(locked_u64_t *dst, uint64_t src) {
+#ifdef JEMALLOC_ATOMIC_U64
+	uint64_t cur_dst = atomic_load_u64(&dst->val, ATOMIC_RELAXED);
+	atomic_store_u64(&dst->val, src + cur_dst, ATOMIC_RELAXED);
+#else
+	dst->val += src;
+#endif
+}
+
+static inline uint64_t
+locked_read_u64_unsynchronized(locked_u64_t *p) {
+#ifdef JEMALLOC_ATOMIC_U64
+	return atomic_load_u64(&p->val, ATOMIC_RELAXED);
+#else
+	return p->val;
+#endif
+}
+
+static inline void
+locked_init_u64_unsynchronized(locked_u64_t *p, uint64_t x) {
+#ifdef JEMALLOC_ATOMIC_U64
+	atomic_store_u64(&p->val, x, ATOMIC_RELAXED);
+#else
+	p->val = x;
+#endif
+}
+
+static inline size_t
+locked_read_zu(tsdn_t *tsdn, malloc_mutex_t *mtx, locked_zu_t *p) {
+	LOCKEDINT_MTX_ASSERT_INTERNAL(tsdn, mtx);
+#ifdef JEMALLOC_ATOMIC_U64
+	return atomic_load_zu(&p->val, ATOMIC_RELAXED);
+#else
+	return atomic_load_zu(&p->val, ATOMIC_RELAXED);
+#endif
+}
+
+static inline void
+locked_inc_zu(tsdn_t *tsdn, malloc_mutex_t *mtx, locked_zu_t *p,
+    size_t x) {
+	LOCKEDINT_MTX_ASSERT_INTERNAL(tsdn, mtx);
+#ifdef JEMALLOC_ATOMIC_U64
+	atomic_fetch_add_zu(&p->val, x, ATOMIC_RELAXED);
+#else
+	size_t cur = atomic_load_zu(&p->val, ATOMIC_RELAXED);
+	atomic_store_zu(&p->val, cur + x, ATOMIC_RELAXED);
+#endif
+}
+
+static inline void
+locked_dec_zu(tsdn_t *tsdn, malloc_mutex_t *mtx, locked_zu_t *p,
+    size_t x) {
+	LOCKEDINT_MTX_ASSERT_INTERNAL(tsdn, mtx);
+#ifdef JEMALLOC_ATOMIC_U64
+	size_t r = atomic_fetch_sub_zu(&p->val, x, ATOMIC_RELAXED);
+	assert(r - x <= r);
+#else
+	size_t cur = atomic_load_zu(&p->val, ATOMIC_RELAXED);
+	atomic_store_zu(&p->val, cur - x, ATOMIC_RELAXED);
+#endif
+}
+
+/* Like the _u64 variant, needs an externally synchronized *dst. */
+static inline void
+locked_inc_zu_unsynchronized(locked_zu_t *dst, size_t src) {
+	size_t cur_dst = atomic_load_zu(&dst->val, ATOMIC_RELAXED);
+	atomic_store_zu(&dst->val, src + cur_dst, ATOMIC_RELAXED);
+}
+
+/*
+ * Unlike the _u64 variant, this is safe to call unconditionally.
+ */
+static inline size_t
+locked_read_atomic_zu(locked_zu_t *p) {
+	return atomic_load_zu(&p->val, ATOMIC_RELAXED);
+}
+
+#endif /* JEMALLOC_INTERNAL_LOCKEDINT_H */

BeefRT/JEMalloc/include/jemalloc/internal/log.h

@@ -0,0 +1,115 @@
+#ifndef JEMALLOC_INTERNAL_LOG_H
+#define JEMALLOC_INTERNAL_LOG_H
+
+#include "jemalloc/internal/atomic.h"
+#include "jemalloc/internal/malloc_io.h"
+#include "jemalloc/internal/mutex.h"
+
+#ifdef JEMALLOC_LOG
+#  define JEMALLOC_LOG_VAR_BUFSIZE 1000
+#else
+#  define JEMALLOC_LOG_VAR_BUFSIZE 1
+#endif
+
+#define JEMALLOC_LOG_BUFSIZE 4096
+
+/*
+ * The log malloc_conf option is a '|'-delimited list of log_var name segments
+ * which should be logged.  The names are themselves hierarchical, with '.' as
+ * the delimiter (a "segment" is just a prefix in the log namespace).  So, if
+ * you have:
+ *
+ * log("arena", "log msg for arena"); // 1
+ * log("arena.a", "log msg for arena.a"); // 2
+ * log("arena.b", "log msg for arena.b"); // 3
+ * log("arena.a.a", "log msg for arena.a.a"); // 4
+ * log("extent.a", "log msg for extent.a"); // 5
+ * log("extent.b", "log msg for extent.b"); // 6
+ *
+ * And your malloc_conf option is "log=arena.a|extent", then lines 2, 4, 5, and
+ * 6 will print at runtime.  You can enable logging from all log vars by
+ * writing "log=.".
+ *
+ * None of this should be regarded as a stable API for right now.  It's intended
+ * as a debugging interface, to let us keep around some of our printf-debugging
+ * statements.
+ */
+
+extern char log_var_names[JEMALLOC_LOG_VAR_BUFSIZE];
+extern atomic_b_t log_init_done;
+
+typedef struct log_var_s log_var_t;
+struct log_var_s {
+	/*
+	 * Lowest bit is "inited", second lowest is "enabled".  Putting them in
+	 * a single word lets us avoid any fences on weak architectures.
+	 */
+	atomic_u_t state;
+	const char *name;
+};
+
+#define LOG_NOT_INITIALIZED 0U
+#define LOG_INITIALIZED_NOT_ENABLED 1U
+#define LOG_ENABLED 2U
+
+#define LOG_VAR_INIT(name_str) {ATOMIC_INIT(LOG_NOT_INITIALIZED), name_str}
+
+/*
+ * Returns the value we should assume for state (which is not necessarily
+ * accurate; if logging is done before logging has finished initializing, then
+ * we default to doing the safe thing by logging everything).
+ */
+unsigned log_var_update_state(log_var_t *log_var);
+
+/* We factor out the metadata management to allow us to test more easily. */
+#define log_do_begin(log_var)						\
+if (config_log) {							\
+	unsigned log_state = atomic_load_u(&(log_var).state,		\
+	    ATOMIC_RELAXED);						\
+	if (unlikely(log_state == LOG_NOT_INITIALIZED)) {		\
+		log_state = log_var_update_state(&(log_var));		\
+		assert(log_state != LOG_NOT_INITIALIZED);		\
+	}								\
+	if (log_state == LOG_ENABLED) {					\
+		{
+			/* User code executes here. */
+#define log_do_end(log_var)						\
+		}							\
+	}								\
+}
+
+/*
+ * MSVC has some preprocessor bugs in its expansion of __VA_ARGS__ during
+ * preprocessing.  To work around this, we take all potential extra arguments in
+ * a var-args functions.  Since a varargs macro needs at least one argument in
+ * the "...", we accept the format string there, and require that the first
+ * argument in this "..." is a const char *.
+ */
+static inline void
+log_impl_varargs(const char *name, ...) {
+	char buf[JEMALLOC_LOG_BUFSIZE];
+	va_list ap;
+
+	va_start(ap, name);
+	const char *format = va_arg(ap, const char *);
+	size_t dst_offset = 0;
+	dst_offset += malloc_snprintf(buf, JEMALLOC_LOG_BUFSIZE, "%s: ", name);
+	dst_offset += malloc_vsnprintf(buf + dst_offset,
+	    JEMALLOC_LOG_BUFSIZE - dst_offset, format, ap);
+	dst_offset += malloc_snprintf(buf + dst_offset,
+	    JEMALLOC_LOG_BUFSIZE - dst_offset, "\n");
+	va_end(ap);
+
+	malloc_write(buf);
+}
+
+/* Call as log("log.var.str", "format_string %d", arg_for_format_string); */
+#define LOG(log_var_str, ...)						\
+do {									\
+	static log_var_t log_var = LOG_VAR_INIT(log_var_str);		\
+	log_do_begin(log_var)						\
+		log_impl_varargs((log_var).name, __VA_ARGS__);		\
+	log_do_end(log_var)						\
+} while (0)
+
+#endif /* JEMALLOC_INTERNAL_LOG_H */

BeefRT/JEMalloc/include/jemalloc/internal/malloc_io.h

@@ -0,0 +1,105 @@
+#ifndef JEMALLOC_INTERNAL_MALLOC_IO_H
+#define JEMALLOC_INTERNAL_MALLOC_IO_H
+
+#include "jemalloc/internal/jemalloc_internal_types.h"
+
+#ifdef _WIN32
+#  ifdef _WIN64
+#    define FMT64_PREFIX "ll"
+#    define FMTPTR_PREFIX "ll"
+#  else
+#    define FMT64_PREFIX "ll"
+#    define FMTPTR_PREFIX ""
+#  endif
+#  define FMTd32 "d"
+#  define FMTu32 "u"
+#  define FMTx32 "x"
+#  define FMTd64 FMT64_PREFIX "d"
+#  define FMTu64 FMT64_PREFIX "u"
+#  define FMTx64 FMT64_PREFIX "x"
+#  define FMTdPTR FMTPTR_PREFIX "d"
+#  define FMTuPTR FMTPTR_PREFIX "u"
+#  define FMTxPTR FMTPTR_PREFIX "x"
+#else
+#  include <inttypes.h>
+#  define FMTd32 PRId32
+#  define FMTu32 PRIu32
+#  define FMTx32 PRIx32
+#  define FMTd64 PRId64
+#  define FMTu64 PRIu64
+#  define FMTx64 PRIx64
+#  define FMTdPTR PRIdPTR
+#  define FMTuPTR PRIuPTR
+#  define FMTxPTR PRIxPTR
+#endif
+
+/* Size of stack-allocated buffer passed to buferror(). */
+#define BUFERROR_BUF		64
+
+/*
+ * Size of stack-allocated buffer used by malloc_{,v,vc}printf().  This must be
+ * large enough for all possible uses within jemalloc.
+ */
+#define MALLOC_PRINTF_BUFSIZE	4096
+
+write_cb_t wrtmessage;
+int buferror(int err, char *buf, size_t buflen);
+uintmax_t malloc_strtoumax(const char *restrict nptr, char **restrict endptr,
+    int base);
+void malloc_write(const char *s);
+
+/*
+ * malloc_vsnprintf() supports a subset of snprintf(3) that avoids floating
+ * point math.
+ */
+size_t malloc_vsnprintf(char *str, size_t size, const char *format,
+    va_list ap);
+size_t malloc_snprintf(char *str, size_t size, const char *format, ...)
+    JEMALLOC_FORMAT_PRINTF(3, 4);
+/*
+ * The caller can set write_cb to null to choose to print with the
+ * je_malloc_message hook.
+ */
+void malloc_vcprintf(write_cb_t *write_cb, void *cbopaque, const char *format,
+    va_list ap);
+void malloc_cprintf(write_cb_t *write_cb, void *cbopaque, const char *format,
+    ...) JEMALLOC_FORMAT_PRINTF(3, 4);
+void malloc_printf(const char *format, ...) JEMALLOC_FORMAT_PRINTF(1, 2);
+
+static inline ssize_t
+malloc_write_fd(int fd, const void *buf, size_t count) {
+#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
+	 * result in the widest plausible type to avoid compiler warnings.
+	 */
+	long result = syscall(SYS_write, fd, buf, count);
+#else
+	ssize_t result = (ssize_t)write(fd, buf,
+#ifdef _WIN32
+	    (unsigned int)
+#endif
+	    count);
+#endif
+	return (ssize_t)result;
+}
+
+static inline ssize_t
+malloc_read_fd(int fd, void *buf, size_t count) {
+#if defined(JEMALLOC_USE_SYSCALL) && defined(SYS_read)
+	long result = syscall(SYS_read, fd, buf, count);
+#else
+	ssize_t result = read(fd, buf,
+#ifdef _WIN32
+	    (unsigned int)
+#endif
+	    count);
+#endif
+	return (ssize_t)result;
+}
+
+#endif /* JEMALLOC_INTERNAL_MALLOC_IO_H */

BeefRT/JEMalloc/include/jemalloc/internal/mpsc_queue.h

@@ -0,0 +1,134 @@
+#ifndef JEMALLOC_INTERNAL_MPSC_QUEUE_H
+#define JEMALLOC_INTERNAL_MPSC_QUEUE_H
+
+#include "jemalloc/internal/atomic.h"
+
+/*
+ * A concurrent implementation of a multi-producer, single-consumer queue.  It
+ * supports three concurrent operations:
+ * - Push
+ * - Push batch
+ * - Pop batch
+ *
+ * These operations are all lock-free.
+ *
+ * The implementation is the simple two-stack queue built on a Treiber stack.
+ * It's not terribly efficient, but this isn't expected to go into anywhere with
+ * hot code.  In fact, we don't really even need queue semantics in any
+ * anticipated use cases; we could get away with just the stack.  But this way
+ * lets us frame the API in terms of the existing list types, which is a nice
+ * convenience.  We can save on cache misses by introducing our own (parallel)
+ * single-linked list type here, and dropping FIFO semantics, if we need this to
+ * get faster.  Since we're currently providing queue semantics though, we use
+ * the prev field in the link rather than the next field for Treiber-stack
+ * linkage, so that we can preserve order for bash-pushed lists (recall that the
+ * two-stack tricks reverses orders in the lock-free first stack).
+ */
+
+#define mpsc_queue(a_type)						\
+struct {								\
+	atomic_p_t tail;						\
+}
+
+#define mpsc_queue_proto(a_attr, a_prefix, a_queue_type, a_type,	\
+    a_list_type)							\
+/* Initialize a queue. */						\
+a_attr void								\
+a_prefix##new(a_queue_type *queue);					\
+/* Insert all items in src into the queue, clearing src. */		\
+a_attr void								\
+a_prefix##push_batch(a_queue_type *queue, a_list_type *src);		\
+/* Insert node into the queue. */					\
+a_attr void								\
+a_prefix##push(a_queue_type *queue, a_type *node);			\
+/*									\
+ * Pop all items in the queue into the list at dst.  dst should already	\
+ * be initialized (and may contain existing items, which then remain	\
+ * in dst).								\
+ */									\
+a_attr void								\
+a_prefix##pop_batch(a_queue_type *queue, a_list_type *dst);
+
+#define mpsc_queue_gen(a_attr, a_prefix, a_queue_type, a_type,		\
+    a_list_type, a_link)						\
+a_attr void								\
+a_prefix##new(a_queue_type *queue) {					\
+	atomic_store_p(&queue->tail, NULL, ATOMIC_RELAXED);		\
+}									\
+a_attr void								\
+a_prefix##push_batch(a_queue_type *queue, a_list_type *src) {		\
+	/*								\
+	 * Reuse the ql list next field as the Treiber stack next	\
+	 * field.							\
+	 */								\
+	a_type *first = ql_first(src);					\
+	a_type *last = ql_last(src, a_link);				\
+	void* cur_tail = atomic_load_p(&queue->tail, ATOMIC_RELAXED);	\
+	do {								\
+		/*							\
+		 * Note that this breaks the queue ring structure;	\
+		 * it's not a ring any more!				\
+		 */							\
+		first->a_link.qre_prev = cur_tail;			\
+		/*							\
+		 * Note: the upcoming CAS doesn't need an atomic; every	\
+		 * push only needs to synchronize with the next pop,	\
+		 * which we get from the release sequence rules.	\
+		 */							\
+	} while (!atomic_compare_exchange_weak_p(&queue->tail,		\
+	    &cur_tail, last, ATOMIC_RELEASE, ATOMIC_RELAXED));		\
+	ql_new(src);							\
+}									\
+a_attr void								\
+a_prefix##push(a_queue_type *queue, a_type *node) {			\
+	ql_elm_new(node, a_link);					\
+	a_list_type list;						\
+	ql_new(&list);							\
+	ql_head_insert(&list, node, a_link);				\
+	a_prefix##push_batch(queue, &list);				\
+}									\
+a_attr void								\
+a_prefix##pop_batch(a_queue_type *queue, a_list_type *dst) {		\
+	a_type *tail = atomic_load_p(&queue->tail, ATOMIC_RELAXED);	\
+	if (tail == NULL) {						\
+		/*							\
+		 * In the common special case where there are no	\
+		 * pending elements, bail early without a costly RMW.	\
+		 */							\
+		return;							\
+	}								\
+	tail = atomic_exchange_p(&queue->tail, NULL, ATOMIC_ACQUIRE);	\
+	/*								\
+	 * It's a single-consumer queue, so if cur started non-NULL,	\
+	 * it'd better stay non-NULL.					\
+	 */								\
+	assert(tail != NULL);						\
+	/*								\
+	 * We iterate through the stack and both fix up the link	\
+	 * structure (stack insertion broke the list requirement that	\
+	 * the list be circularly linked).  It's just as efficient at	\
+	 * this point to make the queue a "real" queue, so do that as	\
+	 * well.							\
+	 * If this ever gets to be a hot spot, we can omit this fixup	\
+	 * and make the queue a bag (i.e. not necessarily ordered), but	\
+	 * that would mean jettisoning the existing list API as the 	\
+	 * batch pushing/popping interface.				\
+	 */								\
+	a_list_type reversed;						\
+	ql_new(&reversed);						\
+	while (tail != NULL) {						\
+		/*							\
+		 * Pop an item off the stack, prepend it onto the list	\
+		 * (reversing the order).  Recall that we use the	\
+		 * list prev field as the Treiber stack next field to	\
+		 * preserve order of batch-pushed items when reversed.	\
+		 */							\
+		a_type *next = tail->a_link.qre_prev;			\
+		ql_elm_new(tail, a_link);				\
+		ql_head_insert(&reversed, tail, a_link);		\
+		tail = next;						\
+	}								\
+	ql_concat(dst, &reversed, a_link);				\
+}
+
+#endif /* JEMALLOC_INTERNAL_MPSC_QUEUE_H */

BeefRT/JEMalloc/include/jemalloc/internal/mutex.h

@@ -0,0 +1,319 @@
+#ifndef JEMALLOC_INTERNAL_MUTEX_H
+#define JEMALLOC_INTERNAL_MUTEX_H
+
+#include "jemalloc/internal/atomic.h"
+#include "jemalloc/internal/mutex_prof.h"
+#include "jemalloc/internal/tsd.h"
+#include "jemalloc/internal/witness.h"
+
+extern int64_t opt_mutex_max_spin;
+
+typedef enum {
+	/* Can only acquire one mutex of a given witness rank at a time. */
+	malloc_mutex_rank_exclusive,
+	/*
+	 * Can acquire multiple mutexes of the same witness rank, but in
+	 * address-ascending order only.
+	 */
+	malloc_mutex_address_ordered
+} malloc_mutex_lock_order_t;
+
+typedef struct malloc_mutex_s malloc_mutex_t;
+struct malloc_mutex_s {
+	union {
+		struct {
+			/*
+			 * prof_data is defined first to reduce cacheline
+			 * bouncing: the data is not touched by the mutex holder
+			 * during unlocking, while might be modified by
+			 * contenders.  Having it before the mutex itself could
+			 * avoid prefetching a modified cacheline (for the
+			 * unlocking thread).
+			 */
+			mutex_prof_data_t	prof_data;
+#ifdef _WIN32
+#  if _WIN32_WINNT >= 0x0600
+			SRWLOCK         	lock;
+#  else
+			CRITICAL_SECTION	lock;
+#  endif
+#elif (defined(JEMALLOC_OS_UNFAIR_LOCK))
+			os_unfair_lock		lock;
+#elif (defined(JEMALLOC_MUTEX_INIT_CB))
+			pthread_mutex_t		lock;
+			malloc_mutex_t		*postponed_next;
+#else
+			pthread_mutex_t		lock;
+#endif
+			/*
+			 * Hint flag to avoid exclusive cache line contention
+			 * during spin waiting
+			 */
+			atomic_b_t		locked;
+		};
+		/*
+		 * We only touch witness when configured w/ debug.  However we
+		 * keep the field in a union when !debug so that we don't have
+		 * to pollute the code base with #ifdefs, while avoid paying the
+		 * memory cost.
+		 */
+#if !defined(JEMALLOC_DEBUG)
+		witness_t			witness;
+		malloc_mutex_lock_order_t	lock_order;
+#endif
+	};
+
+#if defined(JEMALLOC_DEBUG)
+	witness_t			witness;
+	malloc_mutex_lock_order_t	lock_order;
+#endif
+};
+
+#ifdef _WIN32
+#  if _WIN32_WINNT >= 0x0600
+#    define MALLOC_MUTEX_LOCK(m)    AcquireSRWLockExclusive(&(m)->lock)
+#    define MALLOC_MUTEX_UNLOCK(m)  ReleaseSRWLockExclusive(&(m)->lock)
+#    define MALLOC_MUTEX_TRYLOCK(m) (!TryAcquireSRWLockExclusive(&(m)->lock))
+#  else
+#    define MALLOC_MUTEX_LOCK(m)    EnterCriticalSection(&(m)->lock)
+#    define MALLOC_MUTEX_UNLOCK(m)  LeaveCriticalSection(&(m)->lock)
+#    define MALLOC_MUTEX_TRYLOCK(m) (!TryEnterCriticalSection(&(m)->lock))
+#  endif
+#elif (defined(JEMALLOC_OS_UNFAIR_LOCK))
+#    define MALLOC_MUTEX_LOCK(m)    os_unfair_lock_lock(&(m)->lock)
+#    define MALLOC_MUTEX_UNLOCK(m)  os_unfair_lock_unlock(&(m)->lock)
+#    define MALLOC_MUTEX_TRYLOCK(m) (!os_unfair_lock_trylock(&(m)->lock))
+#else
+#    define MALLOC_MUTEX_LOCK(m)    pthread_mutex_lock(&(m)->lock)
+#    define MALLOC_MUTEX_UNLOCK(m)  pthread_mutex_unlock(&(m)->lock)
+#    define MALLOC_MUTEX_TRYLOCK(m) (pthread_mutex_trylock(&(m)->lock) != 0)
+#endif
+
+#define LOCK_PROF_DATA_INITIALIZER					\
+    {NSTIME_ZERO_INITIALIZER, NSTIME_ZERO_INITIALIZER, 0, 0, 0,		\
+	    ATOMIC_INIT(0), 0, NULL, 0}
+
+#ifdef _WIN32
+#  define MALLOC_MUTEX_INITIALIZER
+#elif (defined(JEMALLOC_OS_UNFAIR_LOCK))
+#  if defined(JEMALLOC_DEBUG)
+#    define MALLOC_MUTEX_INITIALIZER					\
+  {{{LOCK_PROF_DATA_INITIALIZER, OS_UNFAIR_LOCK_INIT, ATOMIC_INIT(false)}}, \
+         WITNESS_INITIALIZER("mutex", WITNESS_RANK_OMIT), 0}
+#  else
+#    define MALLOC_MUTEX_INITIALIZER                      \
+  {{{LOCK_PROF_DATA_INITIALIZER, OS_UNFAIR_LOCK_INIT, ATOMIC_INIT(false)}},  \
+      WITNESS_INITIALIZER("mutex", WITNESS_RANK_OMIT)}
+#  endif
+#elif (defined(JEMALLOC_MUTEX_INIT_CB))
+#  if (defined(JEMALLOC_DEBUG))
+#     define MALLOC_MUTEX_INITIALIZER					\
+      {{{LOCK_PROF_DATA_INITIALIZER, PTHREAD_MUTEX_INITIALIZER, NULL, ATOMIC_INIT(false)}},	\
+           WITNESS_INITIALIZER("mutex", WITNESS_RANK_OMIT), 0}
+#  else
+#     define MALLOC_MUTEX_INITIALIZER					\
+      {{{LOCK_PROF_DATA_INITIALIZER, PTHREAD_MUTEX_INITIALIZER, NULL, ATOMIC_INIT(false)}},	\
+           WITNESS_INITIALIZER("mutex", WITNESS_RANK_OMIT)}
+#  endif
+
+#else
+#    define MALLOC_MUTEX_TYPE PTHREAD_MUTEX_DEFAULT
+#  if defined(JEMALLOC_DEBUG)
+#    define MALLOC_MUTEX_INITIALIZER					\
+     {{{LOCK_PROF_DATA_INITIALIZER, PTHREAD_MUTEX_INITIALIZER, ATOMIC_INIT(false)}}, \
+           WITNESS_INITIALIZER("mutex", WITNESS_RANK_OMIT), 0}
+#  else
+#    define MALLOC_MUTEX_INITIALIZER                          \
+     {{{LOCK_PROF_DATA_INITIALIZER, PTHREAD_MUTEX_INITIALIZER, ATOMIC_INIT(false)}},	\
+      WITNESS_INITIALIZER("mutex", WITNESS_RANK_OMIT)}
+#  endif
+#endif
+
+#ifdef JEMALLOC_LAZY_LOCK
+extern bool isthreaded;
+#else
+#  undef isthreaded /* Undo private_namespace.h definition. */
+#  define isthreaded true
+#endif
+
+bool malloc_mutex_init(malloc_mutex_t *mutex, const char *name,
+    witness_rank_t rank, malloc_mutex_lock_order_t lock_order);
+void malloc_mutex_prefork(tsdn_t *tsdn, malloc_mutex_t *mutex);
+void malloc_mutex_postfork_parent(tsdn_t *tsdn, malloc_mutex_t *mutex);
+void malloc_mutex_postfork_child(tsdn_t *tsdn, malloc_mutex_t *mutex);
+bool malloc_mutex_boot(void);
+void malloc_mutex_prof_data_reset(tsdn_t *tsdn, malloc_mutex_t *mutex);
+
+void malloc_mutex_lock_slow(malloc_mutex_t *mutex);
+
+static inline void
+malloc_mutex_lock_final(malloc_mutex_t *mutex) {
+	MALLOC_MUTEX_LOCK(mutex);
+	atomic_store_b(&mutex->locked, true, ATOMIC_RELAXED);
+}
+
+static inline bool
+malloc_mutex_trylock_final(malloc_mutex_t *mutex) {
+	return MALLOC_MUTEX_TRYLOCK(mutex);
+}
+
+static inline void
+mutex_owner_stats_update(tsdn_t *tsdn, malloc_mutex_t *mutex) {
+	if (config_stats) {
+		mutex_prof_data_t *data = &mutex->prof_data;
+		data->n_lock_ops++;
+		if (data->prev_owner != tsdn) {
+			data->prev_owner = tsdn;
+			data->n_owner_switches++;
+		}
+	}
+}
+
+/* Trylock: return false if the lock is successfully acquired. */
+static inline bool
+malloc_mutex_trylock(tsdn_t *tsdn, malloc_mutex_t *mutex) {
+	witness_assert_not_owner(tsdn_witness_tsdp_get(tsdn), &mutex->witness);
+	if (isthreaded) {
+		if (malloc_mutex_trylock_final(mutex)) {
+			atomic_store_b(&mutex->locked, true, ATOMIC_RELAXED);
+			return true;
+		}
+		mutex_owner_stats_update(tsdn, mutex);
+	}
+	witness_lock(tsdn_witness_tsdp_get(tsdn), &mutex->witness);
+
+	return false;
+}
+
+/* Aggregate lock prof data. */
+static inline void
+malloc_mutex_prof_merge(mutex_prof_data_t *sum, mutex_prof_data_t *data) {
+	nstime_add(&sum->tot_wait_time, &data->tot_wait_time);
+	if (nstime_compare(&sum->max_wait_time, &data->max_wait_time) < 0) {
+		nstime_copy(&sum->max_wait_time, &data->max_wait_time);
+	}
+
+	sum->n_wait_times += data->n_wait_times;
+	sum->n_spin_acquired += data->n_spin_acquired;
+
+	if (sum->max_n_thds < data->max_n_thds) {
+		sum->max_n_thds = data->max_n_thds;
+	}
+	uint32_t cur_n_waiting_thds = atomic_load_u32(&sum->n_waiting_thds,
+	    ATOMIC_RELAXED);
+	uint32_t new_n_waiting_thds = cur_n_waiting_thds + atomic_load_u32(
+	    &data->n_waiting_thds, ATOMIC_RELAXED);
+	atomic_store_u32(&sum->n_waiting_thds, new_n_waiting_thds,
+	    ATOMIC_RELAXED);
+	sum->n_owner_switches += data->n_owner_switches;
+	sum->n_lock_ops += data->n_lock_ops;
+}
+
+static inline void
+malloc_mutex_lock(tsdn_t *tsdn, malloc_mutex_t *mutex) {
+	witness_assert_not_owner(tsdn_witness_tsdp_get(tsdn), &mutex->witness);
+	if (isthreaded) {
+		if (malloc_mutex_trylock_final(mutex)) {
+			malloc_mutex_lock_slow(mutex);
+			atomic_store_b(&mutex->locked, true, ATOMIC_RELAXED);
+		}
+		mutex_owner_stats_update(tsdn, mutex);
+	}
+	witness_lock(tsdn_witness_tsdp_get(tsdn), &mutex->witness);
+}
+
+static inline void
+malloc_mutex_unlock(tsdn_t *tsdn, malloc_mutex_t *mutex) {
+	atomic_store_b(&mutex->locked, false, ATOMIC_RELAXED);
+	witness_unlock(tsdn_witness_tsdp_get(tsdn), &mutex->witness);
+	if (isthreaded) {
+		MALLOC_MUTEX_UNLOCK(mutex);
+	}
+}
+
+static inline void
+malloc_mutex_assert_owner(tsdn_t *tsdn, malloc_mutex_t *mutex) {
+	witness_assert_owner(tsdn_witness_tsdp_get(tsdn), &mutex->witness);
+}
+
+static inline void
+malloc_mutex_assert_not_owner(tsdn_t *tsdn, malloc_mutex_t *mutex) {
+	witness_assert_not_owner(tsdn_witness_tsdp_get(tsdn), &mutex->witness);
+}
+
+static inline void
+malloc_mutex_prof_copy(mutex_prof_data_t *dst, mutex_prof_data_t *source) {
+	/*
+	 * Not *really* allowed (we shouldn't be doing non-atomic loads of
+	 * atomic data), but the mutex protection makes this safe, and writing
+	 * a member-for-member copy is tedious for this situation.
+	 */
+	*dst = *source;
+	/* n_wait_thds is not reported (modified w/o locking). */
+	atomic_store_u32(&dst->n_waiting_thds, 0, ATOMIC_RELAXED);
+}
+
+/* Copy the prof data from mutex for processing. */
+static inline void
+malloc_mutex_prof_read(tsdn_t *tsdn, mutex_prof_data_t *data,
+    malloc_mutex_t *mutex) {
+	/* Can only read holding the mutex. */
+	malloc_mutex_assert_owner(tsdn, mutex);
+	malloc_mutex_prof_copy(data, &mutex->prof_data);
+}
+
+static inline void
+malloc_mutex_prof_accum(tsdn_t *tsdn, mutex_prof_data_t *data,
+    malloc_mutex_t *mutex) {
+	mutex_prof_data_t *source = &mutex->prof_data;
+	/* Can only read holding the mutex. */
+	malloc_mutex_assert_owner(tsdn, mutex);
+
+	nstime_add(&data->tot_wait_time, &source->tot_wait_time);
+	if (nstime_compare(&source->max_wait_time, &data->max_wait_time) > 0) {
+		nstime_copy(&data->max_wait_time, &source->max_wait_time);
+	}
+	data->n_wait_times += source->n_wait_times;
+	data->n_spin_acquired += source->n_spin_acquired;
+	if (data->max_n_thds < source->max_n_thds) {
+		data->max_n_thds = source->max_n_thds;
+	}
+	/* n_wait_thds is not reported. */
+	atomic_store_u32(&data->n_waiting_thds, 0, ATOMIC_RELAXED);
+	data->n_owner_switches += source->n_owner_switches;
+	data->n_lock_ops += source->n_lock_ops;
+}
+
+/* Compare the prof data and update to the maximum. */
+static inline void
+malloc_mutex_prof_max_update(tsdn_t *tsdn, mutex_prof_data_t *data,
+    malloc_mutex_t *mutex) {
+	mutex_prof_data_t *source = &mutex->prof_data;
+	/* Can only read holding the mutex. */
+	malloc_mutex_assert_owner(tsdn, mutex);
+
+	if (nstime_compare(&source->tot_wait_time, &data->tot_wait_time) > 0) {
+		nstime_copy(&data->tot_wait_time, &source->tot_wait_time);
+	}
+	if (nstime_compare(&source->max_wait_time, &data->max_wait_time) > 0) {
+		nstime_copy(&data->max_wait_time, &source->max_wait_time);
+	}
+	if (source->n_wait_times > data->n_wait_times) {
+		data->n_wait_times = source->n_wait_times;
+	}
+	if (source->n_spin_acquired > data->n_spin_acquired) {
+		data->n_spin_acquired = source->n_spin_acquired;
+	}
+	if (source->max_n_thds > data->max_n_thds) {
+		data->max_n_thds = source->max_n_thds;
+	}
+	if (source->n_owner_switches > data->n_owner_switches) {
+		data->n_owner_switches = source->n_owner_switches;
+	}
+	if (source->n_lock_ops > data->n_lock_ops) {
+		data->n_lock_ops = source->n_lock_ops;
+	}
+	/* n_wait_thds is not reported. */
+}
+
+#endif /* JEMALLOC_INTERNAL_MUTEX_H */

BeefRT/JEMalloc/include/jemalloc/internal/mutex_prof.h

@@ -0,0 +1,117 @@
+#ifndef JEMALLOC_INTERNAL_MUTEX_PROF_H
+#define JEMALLOC_INTERNAL_MUTEX_PROF_H
+
+#include "jemalloc/internal/atomic.h"
+#include "jemalloc/internal/nstime.h"
+#include "jemalloc/internal/tsd_types.h"
+
+#define MUTEX_PROF_GLOBAL_MUTEXES					\
+    OP(background_thread)						\
+    OP(max_per_bg_thd)							\
+    OP(ctl)								\
+    OP(prof)								\
+    OP(prof_thds_data)							\
+    OP(prof_dump)							\
+    OP(prof_recent_alloc)						\
+    OP(prof_recent_dump)						\
+    OP(prof_stats)
+
+typedef enum {
+#define OP(mtx) global_prof_mutex_##mtx,
+	MUTEX_PROF_GLOBAL_MUTEXES
+#undef OP
+	mutex_prof_num_global_mutexes
+} mutex_prof_global_ind_t;
+
+#define MUTEX_PROF_ARENA_MUTEXES					\
+    OP(large)								\
+    OP(extent_avail)							\
+    OP(extents_dirty)							\
+    OP(extents_muzzy)							\
+    OP(extents_retained)						\
+    OP(decay_dirty)							\
+    OP(decay_muzzy)							\
+    OP(base)								\
+    OP(tcache_list)							\
+    OP(hpa_shard)							\
+    OP(hpa_shard_grow)							\
+    OP(hpa_sec)
+
+typedef enum {
+#define OP(mtx) arena_prof_mutex_##mtx,
+	MUTEX_PROF_ARENA_MUTEXES
+#undef OP
+	mutex_prof_num_arena_mutexes
+} mutex_prof_arena_ind_t;
+
+/*
+ * The forth parameter is a boolean value that is true for derived rate counters
+ * and false for real ones.
+ */
+#define MUTEX_PROF_UINT64_COUNTERS					\
+    OP(num_ops, uint64_t, "n_lock_ops", false, num_ops)					\
+    OP(num_ops_ps, uint64_t, "(#/sec)", true, num_ops)				\
+    OP(num_wait, uint64_t, "n_waiting", false, num_wait)				\
+    OP(num_wait_ps, uint64_t, "(#/sec)", true, num_wait)				\
+    OP(num_spin_acq, uint64_t, "n_spin_acq", false, num_spin_acq)			\
+    OP(num_spin_acq_ps, uint64_t, "(#/sec)", true, num_spin_acq)			\
+    OP(num_owner_switch, uint64_t, "n_owner_switch", false, num_owner_switch)		\
+    OP(num_owner_switch_ps, uint64_t, "(#/sec)", true, num_owner_switch)	\
+    OP(total_wait_time, uint64_t, "total_wait_ns", false, total_wait_time)		\
+    OP(total_wait_time_ps, uint64_t, "(#/sec)", true, total_wait_time)		\
+    OP(max_wait_time, uint64_t, "max_wait_ns", false, max_wait_time)
+
+#define MUTEX_PROF_UINT32_COUNTERS					\
+    OP(max_num_thds, uint32_t, "max_n_thds", false, max_num_thds)
+
+#define MUTEX_PROF_COUNTERS						\
+		MUTEX_PROF_UINT64_COUNTERS				\
+		MUTEX_PROF_UINT32_COUNTERS
+
+#define OP(counter, type, human, derived, base_counter) mutex_counter_##counter,
+
+#define COUNTER_ENUM(counter_list, t)					\
+		typedef enum {						\
+			counter_list					\
+			mutex_prof_num_##t##_counters			\
+		} mutex_prof_##t##_counter_ind_t;
+
+COUNTER_ENUM(MUTEX_PROF_UINT64_COUNTERS, uint64_t)
+COUNTER_ENUM(MUTEX_PROF_UINT32_COUNTERS, uint32_t)
+
+#undef COUNTER_ENUM
+#undef OP
+
+typedef struct {
+	/*
+	 * Counters touched on the slow path, i.e. when there is lock
+	 * contention.  We update them once we have the lock.
+	 */
+	/* Total time (in nano seconds) spent waiting on this mutex. */
+	nstime_t		tot_wait_time;
+	/* Max time (in nano seconds) spent on a single lock operation. */
+	nstime_t		max_wait_time;
+	/* # of times have to wait for this mutex (after spinning). */
+	uint64_t		n_wait_times;
+	/* # of times acquired the mutex through local spinning. */
+	uint64_t		n_spin_acquired;
+	/* Max # of threads waiting for the mutex at the same time. */
+	uint32_t		max_n_thds;
+	/* Current # of threads waiting on the lock.  Atomic synced. */
+	atomic_u32_t		n_waiting_thds;
+
+	/*
+	 * Data touched on the fast path.  These are modified right after we
+	 * grab the lock, so it's placed closest to the end (i.e. right before
+	 * the lock) so that we have a higher chance of them being on the same
+	 * cacheline.
+	 */
+	/* # of times the mutex holder is different than the previous one. */
+	uint64_t		n_owner_switches;
+	/* Previous mutex holder, to facilitate n_owner_switches. */
+	tsdn_t			*prev_owner;
+	/* # of lock() operations in total. */
+	uint64_t		n_lock_ops;
+} mutex_prof_data_t;
+
+#endif /* JEMALLOC_INTERNAL_MUTEX_PROF_H */

BeefRT/JEMalloc/include/jemalloc/internal/nstime.h

@@ -0,0 +1,73 @@
+#ifndef JEMALLOC_INTERNAL_NSTIME_H
+#define JEMALLOC_INTERNAL_NSTIME_H
+
+/* Maximum supported number of seconds (~584 years). */
+#define NSTIME_SEC_MAX KQU(18446744072)
+
+#define NSTIME_MAGIC ((uint32_t)0xb8a9ce37)
+#ifdef JEMALLOC_DEBUG
+#  define NSTIME_ZERO_INITIALIZER {0, NSTIME_MAGIC}
+#else
+#  define NSTIME_ZERO_INITIALIZER {0}
+#endif
+
+typedef struct {
+	uint64_t ns;
+#ifdef JEMALLOC_DEBUG
+	uint32_t magic; /* Tracks if initialized. */
+#endif
+} nstime_t;
+
+static const nstime_t nstime_zero = NSTIME_ZERO_INITIALIZER;
+
+void nstime_init(nstime_t *time, uint64_t ns);
+void nstime_init2(nstime_t *time, uint64_t sec, uint64_t nsec);
+uint64_t nstime_ns(const nstime_t *time);
+uint64_t nstime_sec(const nstime_t *time);
+uint64_t nstime_msec(const nstime_t *time);
+uint64_t nstime_nsec(const nstime_t *time);
+void nstime_copy(nstime_t *time, const nstime_t *source);
+int nstime_compare(const nstime_t *a, const nstime_t *b);
+void nstime_add(nstime_t *time, const nstime_t *addend);
+void nstime_iadd(nstime_t *time, uint64_t addend);
+void nstime_subtract(nstime_t *time, const nstime_t *subtrahend);
+void nstime_isubtract(nstime_t *time, uint64_t subtrahend);
+void nstime_imultiply(nstime_t *time, uint64_t multiplier);
+void nstime_idivide(nstime_t *time, uint64_t divisor);
+uint64_t nstime_divide(const nstime_t *time, const nstime_t *divisor);
+uint64_t nstime_ns_since(const nstime_t *past);
+
+typedef bool (nstime_monotonic_t)(void);
+extern nstime_monotonic_t *JET_MUTABLE nstime_monotonic;
+
+typedef void (nstime_update_t)(nstime_t *);
+extern nstime_update_t *JET_MUTABLE nstime_update;
+
+typedef void (nstime_prof_update_t)(nstime_t *);
+extern nstime_prof_update_t *JET_MUTABLE nstime_prof_update;
+
+void nstime_init_update(nstime_t *time);
+void nstime_prof_init_update(nstime_t *time);
+
+enum prof_time_res_e {
+	prof_time_res_default = 0,
+	prof_time_res_high = 1
+};
+typedef enum prof_time_res_e prof_time_res_t;
+
+extern prof_time_res_t opt_prof_time_res;
+extern const char *prof_time_res_mode_names[];
+
+JEMALLOC_ALWAYS_INLINE void
+nstime_init_zero(nstime_t *time) {
+	nstime_copy(time, &nstime_zero);
+}
+
+JEMALLOC_ALWAYS_INLINE bool
+nstime_equals_zero(nstime_t *time) {
+	int diff = nstime_compare(time, &nstime_zero);
+	assert(diff >= 0);
+	return diff == 0;
+}
+
+#endif /* JEMALLOC_INTERNAL_NSTIME_H */

BeefRT/JEMalloc/include/jemalloc/internal/pa.h

@@ -0,0 +1,243 @@
+#ifndef JEMALLOC_INTERNAL_PA_H
+#define JEMALLOC_INTERNAL_PA_H
+
+#include "jemalloc/internal/base.h"
+#include "jemalloc/internal/decay.h"
+#include "jemalloc/internal/ecache.h"
+#include "jemalloc/internal/edata_cache.h"
+#include "jemalloc/internal/emap.h"
+#include "jemalloc/internal/hpa.h"
+#include "jemalloc/internal/lockedint.h"
+#include "jemalloc/internal/pac.h"
+#include "jemalloc/internal/pai.h"
+#include "jemalloc/internal/sec.h"
+
+/*
+ * The page allocator; responsible for acquiring pages of memory for
+ * allocations.  It picks the implementation of the page allocator interface
+ * (i.e. a pai_t) to handle a given page-level allocation request.  For now, the
+ * only such implementation is the PAC code ("page allocator classic"), but
+ * others will be coming soon.
+ */
+
+typedef struct pa_central_s pa_central_t;
+struct pa_central_s {
+	hpa_central_t hpa;
+};
+
+/*
+ * The stats for a particular pa_shard.  Because of the way the ctl module
+ * handles stats epoch data collection (it has its own arena_stats, and merges
+ * the stats from each arena into it), this needs to live in the arena_stats_t;
+ * hence we define it here and let the pa_shard have a pointer (rather than the
+ * more natural approach of just embedding it in the pa_shard itself).
+ *
+ * We follow the arena_stats_t approach of marking the derived fields.  These
+ * are the ones that are not maintained on their own; instead, their values are
+ * derived during those stats merges.
+ */
+typedef struct pa_shard_stats_s pa_shard_stats_t;
+struct pa_shard_stats_s {
+	/* Number of edata_t structs allocated by base, but not being used. */
+	size_t edata_avail; /* Derived. */
+	/*
+	 * Stats specific to the PAC.  For now, these are the only stats that
+	 * exist, but there will eventually be other page allocators.  Things
+	 * like edata_avail make sense in a cross-PA sense, but things like
+	 * npurges don't.
+	 */
+	pac_stats_t pac_stats;
+};
+
+/*
+ * The local allocator handle.  Keeps the state necessary to satisfy page-sized
+ * allocations.
+ *
+ * The contents are mostly internal to the PA module.  The key exception is that
+ * arena decay code is allowed to grab pointers to the dirty and muzzy ecaches
+ * decay_ts, for a couple of queries, passing them back to a PA function, or
+ * acquiring decay.mtx and looking at decay.purging.  The reasoning is that,
+ * while PA decides what and how to purge, the arena code decides when and where
+ * (e.g. on what thread).  It's allowed to use the presence of another purger to
+ * decide.
+ * (The background thread code also touches some other decay internals, but
+ * that's not fundamental; its' just an artifact of a partial refactoring, and
+ * its accesses could be straightforwardly moved inside the decay module).
+ */
+typedef struct pa_shard_s pa_shard_t;
+struct pa_shard_s {
+	/* The central PA this shard is associated with. */
+	pa_central_t *central;
+
+	/*
+	 * Number of pages in active extents.
+	 *
+	 * Synchronization: atomic.
+	 */
+	atomic_zu_t nactive;
+
+	/*
+	 * Whether or not we should prefer the hugepage allocator.  Atomic since
+	 * it may be concurrently modified by a thread setting extent hooks.
+	 * Note that we still may do HPA operations in this arena; if use_hpa is
+	 * changed from true to false, we'll free back to the hugepage allocator
+	 * for those allocations.
+	 */
+	atomic_b_t use_hpa;
+
+	/*
+	 * If we never used the HPA to begin with, it wasn't initialized, and so
+	 * we shouldn't try to e.g. acquire its mutexes during fork.  This
+	 * tracks that knowledge.
+	 */
+	bool ever_used_hpa;
+
+	/* Allocates from a PAC. */
+	pac_t pac;
+
+	/*
+	 * We place a small extent cache in front of the HPA, since we intend
+	 * these configurations to use many fewer arenas, and therefore have a
+	 * higher risk of hot locks.
+	 */
+	sec_t hpa_sec;
+	hpa_shard_t hpa_shard;
+
+	/* The source of edata_t objects. */
+	edata_cache_t edata_cache;
+
+	unsigned ind;
+
+	malloc_mutex_t *stats_mtx;
+	pa_shard_stats_t *stats;
+
+	/* The emap this shard is tied to. */
+	emap_t *emap;
+
+	/* The base from which we get the ehooks and allocate metadat. */
+	base_t *base;
+};
+
+static inline bool
+pa_shard_dont_decay_muzzy(pa_shard_t *shard) {
+	return ecache_npages_get(&shard->pac.ecache_muzzy) == 0 &&
+	    pac_decay_ms_get(&shard->pac, extent_state_muzzy) <= 0;
+}
+
+static inline ehooks_t *
+pa_shard_ehooks_get(pa_shard_t *shard) {
+	return base_ehooks_get(shard->base);
+}
+
+/* Returns true on error. */
+bool pa_central_init(pa_central_t *central, base_t *base, bool hpa,
+    hpa_hooks_t *hpa_hooks);
+
+/* Returns true on error. */
+bool pa_shard_init(tsdn_t *tsdn, pa_shard_t *shard, pa_central_t *central,
+    emap_t *emap, base_t *base, unsigned ind, pa_shard_stats_t *stats,
+    malloc_mutex_t *stats_mtx, nstime_t *cur_time, size_t oversize_threshold,
+    ssize_t dirty_decay_ms, ssize_t muzzy_decay_ms);
+
+/*
+ * This isn't exposed to users; we allow late enablement of the HPA shard so
+ * that we can boot without worrying about the HPA, then turn it on in a0.
+ */
+bool pa_shard_enable_hpa(tsdn_t *tsdn, pa_shard_t *shard,
+    const hpa_shard_opts_t *hpa_opts, const sec_opts_t *hpa_sec_opts);
+
+/*
+ * We stop using the HPA when custom extent hooks are installed, but still
+ * redirect deallocations to it.
+ */
+void pa_shard_disable_hpa(tsdn_t *tsdn, pa_shard_t *shard);
+
+/*
+ * This does the PA-specific parts of arena reset (i.e. freeing all active
+ * allocations).
+ */
+void pa_shard_reset(tsdn_t *tsdn, pa_shard_t *shard);
+
+/*
+ * Destroy all the remaining retained extents.  Should only be called after
+ * decaying all active, dirty, and muzzy extents to the retained state, as the
+ * last step in destroying the shard.
+ */
+void pa_shard_destroy(tsdn_t *tsdn, pa_shard_t *shard);
+
+/* Gets an edata for the given allocation. */
+edata_t *pa_alloc(tsdn_t *tsdn, pa_shard_t *shard, size_t size,
+    size_t alignment, bool slab, szind_t szind, bool zero, bool guarded,
+    bool *deferred_work_generated);
+/* Returns true on error, in which case nothing changed. */
+bool pa_expand(tsdn_t *tsdn, pa_shard_t *shard, edata_t *edata, size_t old_size,
+    size_t new_size, szind_t szind, bool zero, bool *deferred_work_generated);
+/*
+ * The same.  Sets *generated_dirty to true if we produced new dirty pages, and
+ * false otherwise.
+ */
+bool pa_shrink(tsdn_t *tsdn, pa_shard_t *shard, edata_t *edata, size_t old_size,
+    size_t new_size, szind_t szind, bool *deferred_work_generated);
+/*
+ * Frees the given edata back to the pa.  Sets *generated_dirty if we produced
+ * new dirty pages (well, we always set it for now; but this need not be the
+ * case).
+ * (We could make generated_dirty the return value of course, but this is more
+ * consistent with the shrink pathway and our error codes here).
+ */
+void pa_dalloc(tsdn_t *tsdn, pa_shard_t *shard, edata_t *edata,
+    bool *deferred_work_generated);
+bool pa_decay_ms_set(tsdn_t *tsdn, pa_shard_t *shard, extent_state_t state,
+    ssize_t decay_ms, pac_purge_eagerness_t eagerness);
+ssize_t pa_decay_ms_get(pa_shard_t *shard, extent_state_t state);
+
+/*
+ * Do deferred work on this PA shard.
+ *
+ * Morally, this should do both PAC decay and the HPA deferred work.  For now,
+ * though, the arena, background thread, and PAC modules are tightly interwoven
+ * in a way that's tricky to extricate, so we only do the HPA-specific parts.
+ */
+void pa_shard_set_deferral_allowed(tsdn_t *tsdn, pa_shard_t *shard,
+    bool deferral_allowed);
+void pa_shard_do_deferred_work(tsdn_t *tsdn, pa_shard_t *shard);
+void pa_shard_try_deferred_work(tsdn_t *tsdn, pa_shard_t *shard);
+uint64_t pa_shard_time_until_deferred_work(tsdn_t *tsdn, pa_shard_t *shard);
+
+/******************************************************************************/
+/*
+ * Various bits of "boring" functionality that are still part of this module,
+ * but that we relegate to pa_extra.c, to keep the core logic in pa.c as
+ * readable as possible.
+ */
+
+/*
+ * These fork phases are synchronized with the arena fork phase numbering to
+ * make it easy to keep straight. That's why there's no prefork1.
+ */
+void pa_shard_prefork0(tsdn_t *tsdn, pa_shard_t *shard);
+void pa_shard_prefork2(tsdn_t *tsdn, pa_shard_t *shard);
+void pa_shard_prefork3(tsdn_t *tsdn, pa_shard_t *shard);
+void pa_shard_prefork4(tsdn_t *tsdn, pa_shard_t *shard);
+void pa_shard_prefork5(tsdn_t *tsdn, pa_shard_t *shard);
+void pa_shard_postfork_parent(tsdn_t *tsdn, pa_shard_t *shard);
+void pa_shard_postfork_child(tsdn_t *tsdn, pa_shard_t *shard);
+
+void pa_shard_basic_stats_merge(pa_shard_t *shard, size_t *nactive,
+    size_t *ndirty, size_t *nmuzzy);
+
+void pa_shard_stats_merge(tsdn_t *tsdn, pa_shard_t *shard,
+    pa_shard_stats_t *pa_shard_stats_out, pac_estats_t *estats_out,
+    hpa_shard_stats_t *hpa_stats_out, sec_stats_t *sec_stats_out,
+    size_t *resident);
+
+/*
+ * Reads the PA-owned mutex stats into the output stats array, at the
+ * appropriate positions.  Morally, these stats should really live in
+ * pa_shard_stats_t, but the indices are sort of baked into the various mutex
+ * prof macros.  This would be a good thing to do at some point.
+ */
+void pa_shard_mtx_stats_read(tsdn_t *tsdn, pa_shard_t *shard,
+    mutex_prof_data_t mutex_prof_data[mutex_prof_num_arena_mutexes]);
+
+#endif /* JEMALLOC_INTERNAL_PA_H */

BeefRT/JEMalloc/include/jemalloc/internal/pac.h

@@ -0,0 +1,179 @@
+#ifndef JEMALLOC_INTERNAL_PAC_H
+#define JEMALLOC_INTERNAL_PAC_H
+
+#include "jemalloc/internal/exp_grow.h"
+#include "jemalloc/internal/pai.h"
+#include "san_bump.h"
+
+
+/*
+ * Page allocator classic; an implementation of the PAI interface that:
+ * - Can be used for arenas with custom extent hooks.
+ * - Can always satisfy any allocation request (including highly-fragmentary
+ *   ones).
+ * - Can use efficient OS-level zeroing primitives for demand-filled pages.
+ */
+
+/* How "eager" decay/purging should be. */
+enum pac_purge_eagerness_e {
+	PAC_PURGE_ALWAYS,
+	PAC_PURGE_NEVER,
+	PAC_PURGE_ON_EPOCH_ADVANCE
+};
+typedef enum pac_purge_eagerness_e pac_purge_eagerness_t;
+
+typedef struct pac_decay_stats_s pac_decay_stats_t;
+struct pac_decay_stats_s {
+	/* Total number of purge sweeps. */
+	locked_u64_t npurge;
+	/* Total number of madvise calls made. */
+	locked_u64_t nmadvise;
+	/* Total number of pages purged. */
+	locked_u64_t purged;
+};
+
+typedef struct pac_estats_s pac_estats_t;
+struct pac_estats_s {
+	/*
+	 * Stats for a given index in the range [0, SC_NPSIZES] in the various
+	 * ecache_ts.
+	 * We track both bytes and # of extents: two extents in the same bucket
+	 * may have different sizes if adjacent size classes differ by more than
+	 * a page, so bytes cannot always be derived from # of extents.
+	 */
+	size_t ndirty;
+	size_t dirty_bytes;
+	size_t nmuzzy;
+	size_t muzzy_bytes;
+	size_t nretained;
+	size_t retained_bytes;
+};
+
+typedef struct pac_stats_s pac_stats_t;
+struct pac_stats_s {
+	pac_decay_stats_t decay_dirty;
+	pac_decay_stats_t decay_muzzy;
+
+	/*
+	 * Number of unused virtual memory bytes currently retained.  Retained
+	 * bytes are technically mapped (though always decommitted or purged),
+	 * but they are excluded from the mapped statistic (above).
+	 */
+	size_t retained; /* Derived. */
+
+	/*
+	 * Number of bytes currently mapped, excluding retained memory (and any
+	 * base-allocated memory, which is tracked by the arena stats).
+	 *
+	 * We name this "pac_mapped" to avoid confusion with the arena_stats
+	 * "mapped".
+	 */
+	atomic_zu_t pac_mapped;
+
+	/* VM space had to be leaked (undocumented).  Normally 0. */
+	atomic_zu_t abandoned_vm;
+};
+
+typedef struct pac_s pac_t;
+struct pac_s {
+	/*
+	 * Must be the first member (we convert it to a PAC given only a
+	 * pointer).  The handle to the allocation interface.
+	 */
+	pai_t pai;
+	/*
+	 * Collections of extents that were previously allocated.  These are
+	 * used when allocating extents, in an attempt to re-use address space.
+	 *
+	 * Synchronization: internal.
+	 */
+	ecache_t ecache_dirty;
+	ecache_t ecache_muzzy;
+	ecache_t ecache_retained;
+
+	base_t *base;
+	emap_t *emap;
+	edata_cache_t *edata_cache;
+
+	/* The grow info for the retained ecache. */
+	exp_grow_t exp_grow;
+	malloc_mutex_t grow_mtx;
+
+	/* Special allocator for guarded frequently reused extents. */
+	san_bump_alloc_t sba;
+
+	/* How large extents should be before getting auto-purged. */
+	atomic_zu_t oversize_threshold;
+
+	/*
+	 * Decay-based purging state, responsible for scheduling extent state
+	 * transitions.
+	 *
+	 * Synchronization: via the internal mutex.
+	 */
+	decay_t decay_dirty; /* dirty --> muzzy */
+	decay_t decay_muzzy; /* muzzy --> retained */
+
+	malloc_mutex_t *stats_mtx;
+	pac_stats_t *stats;
+
+	/* Extent serial number generator state. */
+	atomic_zu_t extent_sn_next;
+};
+
+bool pac_init(tsdn_t *tsdn, pac_t *pac, base_t *base, emap_t *emap,
+    edata_cache_t *edata_cache, nstime_t *cur_time, size_t oversize_threshold,
+    ssize_t dirty_decay_ms, ssize_t muzzy_decay_ms, pac_stats_t *pac_stats,
+    malloc_mutex_t *stats_mtx);
+
+static inline size_t
+pac_mapped(pac_t *pac) {
+	return atomic_load_zu(&pac->stats->pac_mapped, ATOMIC_RELAXED);
+}
+
+static inline ehooks_t *
+pac_ehooks_get(pac_t *pac) {
+	return base_ehooks_get(pac->base);
+}
+
+/*
+ * All purging functions require holding decay->mtx.  This is one of the few
+ * places external modules are allowed to peek inside pa_shard_t internals.
+ */
+
+/*
+ * Decays the number of pages currently in the ecache.  This might not leave the
+ * ecache empty if other threads are inserting dirty objects into it
+ * concurrently with the call.
+ */
+void pac_decay_all(tsdn_t *tsdn, pac_t *pac, decay_t *decay,
+    pac_decay_stats_t *decay_stats, ecache_t *ecache, bool fully_decay);
+/*
+ * Updates decay settings for the current time, and conditionally purges in
+ * response (depending on decay_purge_setting).  Returns whether or not the
+ * epoch advanced.
+ */
+bool pac_maybe_decay_purge(tsdn_t *tsdn, pac_t *pac, decay_t *decay,
+    pac_decay_stats_t *decay_stats, ecache_t *ecache,
+    pac_purge_eagerness_t eagerness);
+
+/*
+ * Gets / sets the maximum amount that we'll grow an arena down the
+ * grow-retained pathways (unless forced to by an allocaction request).
+ *
+ * Set new_limit to NULL if it's just a query, or old_limit to NULL if you don't
+ * care about the previous value.
+ *
+ * Returns true on error (if the new limit is not valid).
+ */
+bool pac_retain_grow_limit_get_set(tsdn_t *tsdn, pac_t *pac, size_t *old_limit,
+    size_t *new_limit);
+
+bool pac_decay_ms_set(tsdn_t *tsdn, pac_t *pac, extent_state_t state,
+    ssize_t decay_ms, pac_purge_eagerness_t eagerness);
+ssize_t pac_decay_ms_get(pac_t *pac, extent_state_t state);
+
+void pac_reset(tsdn_t *tsdn, pac_t *pac);
+void pac_destroy(tsdn_t *tsdn, pac_t *pac);
+
+#endif /* JEMALLOC_INTERNAL_PAC_H */

BeefRT/JEMalloc/include/jemalloc/internal/pages.h

@@ -0,0 +1,119 @@
+#ifndef JEMALLOC_INTERNAL_PAGES_EXTERNS_H
+#define JEMALLOC_INTERNAL_PAGES_EXTERNS_H
+
+/* Page size.  LG_PAGE is determined by the configure script. */
+#ifdef PAGE_MASK
+#  undef PAGE_MASK
+#endif
+#define PAGE		((size_t)(1U << LG_PAGE))
+#define PAGE_MASK	((size_t)(PAGE - 1))
+/* Return the page base address for the page containing address a. */
+#define PAGE_ADDR2BASE(a)						\
+	((void *)((uintptr_t)(a) & ~PAGE_MASK))
+/* Return the smallest pagesize multiple that is >= s. */
+#define PAGE_CEILING(s)							\
+	(((s) + PAGE_MASK) & ~PAGE_MASK)
+/* Return the largest pagesize multiple that is <=s. */
+#define PAGE_FLOOR(s) 							\
+	((s) & ~PAGE_MASK)
+
+/* Huge page size.  LG_HUGEPAGE is determined by the configure script. */
+#define HUGEPAGE	((size_t)(1U << LG_HUGEPAGE))
+#define HUGEPAGE_MASK	((size_t)(HUGEPAGE - 1))
+
+#if LG_HUGEPAGE != 0
+#  define HUGEPAGE_PAGES (HUGEPAGE / PAGE)
+#else
+/*
+ * It's convenient to define arrays (or bitmaps) of HUGEPAGE_PAGES lengths.  If
+ * we can't autodetect the hugepage size, it gets treated as 0, in which case
+ * we'll trigger a compiler error in those arrays.  Avoid this case by ensuring
+ * that this value is at least 1.  (We won't ever run in this degraded state;
+ * hpa_supported() returns false in this case.
+ */
+#  define HUGEPAGE_PAGES 1
+#endif
+
+/* Return the huge page base address for the huge page containing address a. */
+#define HUGEPAGE_ADDR2BASE(a)						\
+	((void *)((uintptr_t)(a) & ~HUGEPAGE_MASK))
+/* Return the smallest pagesize multiple that is >= s. */
+#define HUGEPAGE_CEILING(s)						\
+	(((s) + HUGEPAGE_MASK) & ~HUGEPAGE_MASK)
+
+/* PAGES_CAN_PURGE_LAZY is defined if lazy purging is supported. */
+#if defined(_WIN32) || defined(JEMALLOC_PURGE_MADVISE_FREE)
+#  define PAGES_CAN_PURGE_LAZY
+#endif
+/*
+ * PAGES_CAN_PURGE_FORCED is defined if forced purging is supported.
+ *
+ * The only supported way to hard-purge on Windows is to decommit and then
+ * re-commit, but doing so is racy, and if re-commit fails it's a pain to
+ * propagate the "poisoned" memory state.  Since we typically decommit as the
+ * next step after purging on Windows anyway, there's no point in adding such
+ * complexity.
+ */
+#if !defined(_WIN32) && ((defined(JEMALLOC_PURGE_MADVISE_DONTNEED) && \
+    defined(JEMALLOC_PURGE_MADVISE_DONTNEED_ZEROS)) || \
+    defined(JEMALLOC_MAPS_COALESCE))
+#  define PAGES_CAN_PURGE_FORCED
+#endif
+
+static const bool pages_can_purge_lazy =
+#ifdef PAGES_CAN_PURGE_LAZY
+    true
+#else
+    false
+#endif
+    ;
+static const bool pages_can_purge_forced =
+#ifdef PAGES_CAN_PURGE_FORCED
+    true
+#else
+    false
+#endif
+    ;
+
+#if defined(JEMALLOC_HAVE_MADVISE_HUGE) || defined(JEMALLOC_HAVE_MEMCNTL)
+#  define PAGES_CAN_HUGIFY
+#endif
+
+static const bool pages_can_hugify =
+#ifdef PAGES_CAN_HUGIFY
+    true
+#else
+    false
+#endif
+    ;
+
+typedef enum {
+	thp_mode_default       = 0, /* Do not change hugepage settings. */
+	thp_mode_always        = 1, /* Always set MADV_HUGEPAGE. */
+	thp_mode_never         = 2, /* Always set MADV_NOHUGEPAGE. */
+
+	thp_mode_names_limit   = 3, /* Used for option processing. */
+	thp_mode_not_supported = 3  /* No THP support detected. */
+} thp_mode_t;
+
+#define THP_MODE_DEFAULT thp_mode_default
+extern thp_mode_t opt_thp;
+extern thp_mode_t init_system_thp_mode; /* Initial system wide state. */
+extern const char *thp_mode_names[];
+
+void *pages_map(void *addr, size_t size, size_t alignment, bool *commit);
+void pages_unmap(void *addr, size_t size);
+bool pages_commit(void *addr, size_t size);
+bool pages_decommit(void *addr, size_t size);
+bool pages_purge_lazy(void *addr, size_t size);
+bool pages_purge_forced(void *addr, size_t size);
+bool pages_huge(void *addr, size_t size);
+bool pages_nohuge(void *addr, size_t size);
+bool pages_dontdump(void *addr, size_t size);
+bool pages_dodump(void *addr, size_t size);
+bool pages_boot(void);
+void pages_set_thp_state (void *ptr, size_t size);
+void pages_mark_guards(void *head, void *tail);
+void pages_unmark_guards(void *head, void *tail);
+
+#endif /* JEMALLOC_INTERNAL_PAGES_EXTERNS_H */

BeefRT/JEMalloc/include/jemalloc/internal/pai.h

@@ -0,0 +1,95 @@
+#ifndef JEMALLOC_INTERNAL_PAI_H
+#define JEMALLOC_INTERNAL_PAI_H
+
+/* An interface for page allocation. */
+
+typedef struct pai_s pai_t;
+struct pai_s {
+	/* Returns NULL on failure. */
+	edata_t *(*alloc)(tsdn_t *tsdn, pai_t *self, size_t size,
+	    size_t alignment, bool zero, bool guarded, bool frequent_reuse,
+	    bool *deferred_work_generated);
+	/*
+	 * Returns the number of extents added to the list (which may be fewer
+	 * than requested, in case of OOM).  The list should already be
+	 * initialized.  The only alignment guarantee is page-alignment, and
+	 * the results are not necessarily zeroed.
+	 */
+	size_t (*alloc_batch)(tsdn_t *tsdn, pai_t *self, size_t size,
+	    size_t nallocs, edata_list_active_t *results,
+	    bool *deferred_work_generated);
+	bool (*expand)(tsdn_t *tsdn, pai_t *self, edata_t *edata,
+	    size_t old_size, size_t new_size, bool zero,
+	    bool *deferred_work_generated);
+	bool (*shrink)(tsdn_t *tsdn, pai_t *self, edata_t *edata,
+	    size_t old_size, size_t new_size, bool *deferred_work_generated);
+	void (*dalloc)(tsdn_t *tsdn, pai_t *self, edata_t *edata,
+	    bool *deferred_work_generated);
+	/* This function empties out list as a side-effect of being called. */
+	void (*dalloc_batch)(tsdn_t *tsdn, pai_t *self,
+	    edata_list_active_t *list, bool *deferred_work_generated);
+	uint64_t (*time_until_deferred_work)(tsdn_t *tsdn, pai_t *self);
+};
+
+/*
+ * These are just simple convenience functions to avoid having to reference the
+ * same pai_t twice on every invocation.
+ */
+
+static inline edata_t *
+pai_alloc(tsdn_t *tsdn, pai_t *self, size_t size, size_t alignment,
+    bool zero, bool guarded, bool frequent_reuse,
+    bool *deferred_work_generated) {
+	return self->alloc(tsdn, self, size, alignment, zero, guarded,
+	    frequent_reuse, deferred_work_generated);
+}
+
+static inline size_t
+pai_alloc_batch(tsdn_t *tsdn, pai_t *self, size_t size, size_t nallocs,
+    edata_list_active_t *results, bool *deferred_work_generated) {
+	return self->alloc_batch(tsdn, self, size, nallocs, results,
+	    deferred_work_generated);
+}
+
+static inline bool
+pai_expand(tsdn_t *tsdn, pai_t *self, edata_t *edata, size_t old_size,
+    size_t new_size, bool zero, bool *deferred_work_generated) {
+	return self->expand(tsdn, self, edata, old_size, new_size, zero,
+	    deferred_work_generated);
+}
+
+static inline bool
+pai_shrink(tsdn_t *tsdn, pai_t *self, edata_t *edata, size_t old_size,
+    size_t new_size, bool *deferred_work_generated) {
+	return self->shrink(tsdn, self, edata, old_size, new_size,
+	    deferred_work_generated);
+}
+
+static inline void
+pai_dalloc(tsdn_t *tsdn, pai_t *self, edata_t *edata,
+    bool *deferred_work_generated) {
+	self->dalloc(tsdn, self, edata, deferred_work_generated);
+}
+
+static inline void
+pai_dalloc_batch(tsdn_t *tsdn, pai_t *self, edata_list_active_t *list,
+    bool *deferred_work_generated) {
+	self->dalloc_batch(tsdn, self, list, deferred_work_generated);
+}
+
+static inline uint64_t
+pai_time_until_deferred_work(tsdn_t *tsdn, pai_t *self) {
+	return self->time_until_deferred_work(tsdn, self);
+}
+
+/*
+ * An implementation of batch allocation that simply calls alloc once for
+ * each item in the list.
+ */
+size_t pai_alloc_batch_default(tsdn_t *tsdn, pai_t *self, size_t size,
+    size_t nallocs, edata_list_active_t *results, bool *deferred_work_generated);
+/* Ditto, for dalloc. */
+void pai_dalloc_batch_default(tsdn_t *tsdn, pai_t *self,
+    edata_list_active_t *list, bool *deferred_work_generated);
+
+#endif /* JEMALLOC_INTERNAL_PAI_H */

BeefRT/JEMalloc/include/jemalloc/internal/peak.h

@@ -0,0 +1,37 @@
+#ifndef JEMALLOC_INTERNAL_PEAK_H
+#define JEMALLOC_INTERNAL_PEAK_H
+
+typedef struct peak_s peak_t;
+struct peak_s {
+	/* The highest recorded peak value, after adjustment (see below). */
+	uint64_t cur_max;
+	/*
+	 * The difference between alloc and dalloc at the last set_zero call;
+	 * this lets us cancel out the appropriate amount of excess.
+	 */
+	uint64_t adjustment;
+};
+
+#define PEAK_INITIALIZER {0, 0}
+
+static inline uint64_t
+peak_max(peak_t *peak) {
+	return peak->cur_max;
+}
+
+static inline void
+peak_update(peak_t *peak, uint64_t alloc, uint64_t dalloc) {
+	int64_t candidate_max = (int64_t)(alloc - dalloc - peak->adjustment);
+	if (candidate_max > (int64_t)peak->cur_max) {
+		peak->cur_max = candidate_max;
+	}
+}
+
+/* Resets the counter to zero; all peaks are now relative to this point. */
+static inline void
+peak_set_zero(peak_t *peak, uint64_t alloc, uint64_t dalloc) {
+	peak->cur_max = 0;
+	peak->adjustment = alloc - dalloc;
+}
+
+#endif /* JEMALLOC_INTERNAL_PEAK_H */

BeefRT/JEMalloc/include/jemalloc/internal/peak_event.h

@@ -0,0 +1,24 @@
+#ifndef JEMALLOC_INTERNAL_PEAK_EVENT_H
+#define JEMALLOC_INTERNAL_PEAK_EVENT_H
+
+/*
+ * While peak.h contains the simple helper struct that tracks state, this
+ * contains the allocator tie-ins (and knows about tsd, the event module, etc.).
+ */
+
+/* Update the peak with current tsd state. */
+void peak_event_update(tsd_t *tsd);
+/* Set current state to zero. */
+void peak_event_zero(tsd_t *tsd);
+uint64_t peak_event_max(tsd_t *tsd);
+
+/* Manual hooks. */
+/* The activity-triggered hooks. */
+uint64_t peak_alloc_new_event_wait(tsd_t *tsd);
+uint64_t peak_alloc_postponed_event_wait(tsd_t *tsd);
+void peak_alloc_event_handler(tsd_t *tsd, uint64_t elapsed);
+uint64_t peak_dalloc_new_event_wait(tsd_t *tsd);
+uint64_t peak_dalloc_postponed_event_wait(tsd_t *tsd);
+void peak_dalloc_event_handler(tsd_t *tsd, uint64_t elapsed);
+
+#endif /* JEMALLOC_INTERNAL_PEAK_EVENT_H */

BeefRT/JEMalloc/include/jemalloc/internal/ph.h

@@ -0,0 +1,520 @@
+#ifndef JEMALLOC_INTERNAL_PH_H
+#define JEMALLOC_INTERNAL_PH_H
+
+/*
+ * A Pairing Heap implementation.
+ *
+ * "The Pairing Heap: A New Form of Self-Adjusting Heap"
+ * https://www.cs.cmu.edu/~sleator/papers/pairing-heaps.pdf
+ *
+ * With auxiliary twopass list, described in a follow on paper.
+ *
+ * "Pairing Heaps: Experiments and Analysis"
+ * http://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.106.2988&rep=rep1&type=pdf
+ *
+ *******************************************************************************
+ *
+ * We include a non-obvious optimization:
+ * - First, we introduce a new pop-and-link operation; pop the two most
+ *   recently-inserted items off the aux-list, link them, and push the resulting
+ *   heap.
+ * - We maintain a count of the number of insertions since the last time we
+ *   merged the aux-list (i.e. via first() or remove_first()).  After N inserts,
+ *   we do ffs(N) pop-and-link operations.
+ *
+ * One way to think of this is that we're progressively building up a tree in
+ * the aux-list, rather than a linked-list (think of the series of merges that
+ * will be performed as the aux-count grows).
+ *
+ * There's a couple reasons we benefit from this:
+ * - Ordinarily, after N insertions, the aux-list is of size N.  With our
+ *   strategy, it's of size O(log(N)).  So we decrease the worst-case time of
+ *   first() calls, and reduce the average cost of remove_min calls.  Since
+ *   these almost always occur while holding a lock, we practically reduce the
+ *   frequency of unusually long hold times.
+ * - This moves the bulk of the work of merging the aux-list onto the threads
+ *   that are inserting into the heap.  In some common scenarios, insertions
+ *   happen in bulk, from a single thread (think tcache flushing; we potentially
+ *   move many slabs from slabs_full to slabs_nonfull).  All the nodes in this
+ *   case are in the inserting threads cache, and linking them is very cheap
+ *   (cache misses dominate linking cost).  Without this optimization, linking
+ *   happens on the next call to remove_first.  Since that remove_first call
+ *   likely happens on a different thread (or at least, after the cache has
+ *   gotten cold if done on the same thread), deferring linking trades cheap
+ *   link operations now for expensive ones later.
+ *
+ * The ffs trick keeps amortized insert cost at constant time.  Similar
+ * strategies based on periodically sorting the list after a batch of operations
+ * perform worse than this in practice, even with various fancy tricks; they
+ * all took amortized complexity of an insert from O(1) to O(log(n)).
+ */
+
+typedef int (*ph_cmp_t)(void *, void *);
+
+/* Node structure. */
+typedef struct phn_link_s phn_link_t;
+struct phn_link_s {
+	void *prev;
+	void *next;
+	void *lchild;
+};
+
+typedef struct ph_s ph_t;
+struct ph_s {
+	void *root;
+	/*
+	 * Inserts done since the last aux-list merge.  This is not necessarily
+	 * the size of the aux-list, since it's possible that removals have
+	 * happened since, and we don't track whether or not those removals are
+	 * from the aux list.
+	 */
+	size_t auxcount;
+};
+
+JEMALLOC_ALWAYS_INLINE phn_link_t *
+phn_link_get(void *phn, size_t offset) {
+	return (phn_link_t *)(((uintptr_t)phn) + offset);
+}
+
+JEMALLOC_ALWAYS_INLINE void
+phn_link_init(void *phn, size_t offset) {
+	phn_link_get(phn, offset)->prev = NULL;
+	phn_link_get(phn, offset)->next = NULL;
+	phn_link_get(phn, offset)->lchild = NULL;
+}
+
+/* Internal utility helpers. */
+JEMALLOC_ALWAYS_INLINE void *
+phn_lchild_get(void *phn, size_t offset) {
+	return phn_link_get(phn, offset)->lchild;
+}
+
+JEMALLOC_ALWAYS_INLINE void
+phn_lchild_set(void *phn, void *lchild, size_t offset) {
+	phn_link_get(phn, offset)->lchild = lchild;
+}
+
+JEMALLOC_ALWAYS_INLINE void *
+phn_next_get(void *phn, size_t offset) {
+	return phn_link_get(phn, offset)->next;
+}
+
+JEMALLOC_ALWAYS_INLINE void
+phn_next_set(void *phn, void *next, size_t offset) {
+	phn_link_get(phn, offset)->next = next;
+}
+
+JEMALLOC_ALWAYS_INLINE void *
+phn_prev_get(void *phn, size_t offset) {
+	return phn_link_get(phn, offset)->prev;
+}
+
+JEMALLOC_ALWAYS_INLINE void
+phn_prev_set(void *phn, void *prev, size_t offset) {
+	phn_link_get(phn, offset)->prev = prev;
+}
+
+JEMALLOC_ALWAYS_INLINE void
+phn_merge_ordered(void *phn0, void *phn1, size_t offset,
+    ph_cmp_t cmp) {
+	void *phn0child;
+
+	assert(phn0 != NULL);
+	assert(phn1 != NULL);
+	assert(cmp(phn0, phn1) <= 0);
+
+	phn_prev_set(phn1, phn0, offset);
+	phn0child = phn_lchild_get(phn0, offset);
+	phn_next_set(phn1, phn0child, offset);
+	if (phn0child != NULL) {
+		phn_prev_set(phn0child, phn1, offset);
+	}
+	phn_lchild_set(phn0, phn1, offset);
+}
+
+JEMALLOC_ALWAYS_INLINE void *
+phn_merge(void *phn0, void *phn1, size_t offset, ph_cmp_t cmp) {
+	void *result;
+	if (phn0 == NULL) {
+		result = phn1;
+	} else if (phn1 == NULL) {
+		result = phn0;
+	} else if (cmp(phn0, phn1) < 0) {
+		phn_merge_ordered(phn0, phn1, offset, cmp);
+		result = phn0;
+	} else {
+		phn_merge_ordered(phn1, phn0, offset, cmp);
+		result = phn1;
+	}
+	return result;
+}
+
+JEMALLOC_ALWAYS_INLINE void *
+phn_merge_siblings(void *phn, size_t offset, ph_cmp_t cmp) {
+	void *head = NULL;
+	void *tail = NULL;
+	void *phn0 = phn;
+	void *phn1 = phn_next_get(phn0, offset);
+
+	/*
+	 * Multipass merge, wherein the first two elements of a FIFO
+	 * are repeatedly merged, and each result is appended to the
+	 * singly linked FIFO, until the FIFO contains only a single
+	 * element.  We start with a sibling list but no reference to
+	 * its tail, so we do a single pass over the sibling list to
+	 * populate the FIFO.
+	 */
+	if (phn1 != NULL) {
+		void *phnrest = phn_next_get(phn1, offset);
+		if (phnrest != NULL) {
+			phn_prev_set(phnrest, NULL, offset);
+		}
+		phn_prev_set(phn0, NULL, offset);
+		phn_next_set(phn0, NULL, offset);
+		phn_prev_set(phn1, NULL, offset);
+		phn_next_set(phn1, NULL, offset);
+		phn0 = phn_merge(phn0, phn1, offset, cmp);
+		head = tail = phn0;
+		phn0 = phnrest;
+		while (phn0 != NULL) {
+			phn1 = phn_next_get(phn0, offset);
+			if (phn1 != NULL) {
+				phnrest = phn_next_get(phn1, offset);
+				if (phnrest != NULL) {
+					phn_prev_set(phnrest, NULL, offset);
+				}
+				phn_prev_set(phn0, NULL, offset);
+				phn_next_set(phn0, NULL, offset);
+				phn_prev_set(phn1, NULL, offset);
+				phn_next_set(phn1, NULL, offset);
+				phn0 = phn_merge(phn0, phn1, offset, cmp);
+				phn_next_set(tail, phn0, offset);
+				tail = phn0;
+				phn0 = phnrest;
+			} else {
+				phn_next_set(tail, phn0, offset);
+				tail = phn0;
+				phn0 = NULL;
+			}
+		}
+		phn0 = head;
+		phn1 = phn_next_get(phn0, offset);
+		if (phn1 != NULL) {
+			while (true) {
+				head = phn_next_get(phn1, offset);
+				assert(phn_prev_get(phn0, offset) == NULL);
+				phn_next_set(phn0, NULL, offset);
+				assert(phn_prev_get(phn1, offset) == NULL);
+				phn_next_set(phn1, NULL, offset);
+				phn0 = phn_merge(phn0, phn1, offset, cmp);
+				if (head == NULL) {
+					break;
+				}
+				phn_next_set(tail, phn0, offset);
+				tail = phn0;
+				phn0 = head;
+				phn1 = phn_next_get(phn0, offset);
+			}
+		}
+	}
+	return phn0;
+}
+
+JEMALLOC_ALWAYS_INLINE void
+ph_merge_aux(ph_t *ph, size_t offset, ph_cmp_t cmp) {
+	ph->auxcount = 0;
+	void *phn = phn_next_get(ph->root, offset);
+	if (phn != NULL) {
+		phn_prev_set(ph->root, NULL, offset);
+		phn_next_set(ph->root, NULL, offset);
+		phn_prev_set(phn, NULL, offset);
+		phn = phn_merge_siblings(phn, offset, cmp);
+		assert(phn_next_get(phn, offset) == NULL);
+		ph->root = phn_merge(ph->root, phn, offset, cmp);
+	}
+}
+
+JEMALLOC_ALWAYS_INLINE void *
+ph_merge_children(void *phn, size_t offset, ph_cmp_t cmp) {
+	void *result;
+	void *lchild = phn_lchild_get(phn, offset);
+	if (lchild == NULL) {
+		result = NULL;
+	} else {
+		result = phn_merge_siblings(lchild, offset, cmp);
+	}
+	return result;
+}
+
+JEMALLOC_ALWAYS_INLINE void
+ph_new(ph_t *ph) {
+	ph->root = NULL;
+	ph->auxcount = 0;
+}
+
+JEMALLOC_ALWAYS_INLINE bool
+ph_empty(ph_t *ph) {
+	return ph->root == NULL;
+}
+
+JEMALLOC_ALWAYS_INLINE void *
+ph_first(ph_t *ph, size_t offset, ph_cmp_t cmp) {
+	if (ph->root == NULL) {
+		return NULL;
+	}
+	ph_merge_aux(ph, offset, cmp);
+	return ph->root;
+}
+
+JEMALLOC_ALWAYS_INLINE void *
+ph_any(ph_t *ph, size_t offset) {
+	if (ph->root == NULL) {
+		return NULL;
+	}
+	void *aux = phn_next_get(ph->root, offset);
+	if (aux != NULL) {
+		return aux;
+	}
+	return ph->root;
+}
+
+/* Returns true if we should stop trying to merge. */
+JEMALLOC_ALWAYS_INLINE bool
+ph_try_aux_merge_pair(ph_t *ph, size_t offset, ph_cmp_t cmp) {
+	assert(ph->root != NULL);
+	void *phn0 = phn_next_get(ph->root, offset);
+	if (phn0 == NULL) {
+		return true;
+	}
+	void *phn1 = phn_next_get(phn0, offset);
+	if (phn1 == NULL) {
+		return true;
+	}
+	void *next_phn1 = phn_next_get(phn1, offset);
+	phn_next_set(phn0, NULL, offset);
+	phn_prev_set(phn0, NULL, offset);
+	phn_next_set(phn1, NULL, offset);
+	phn_prev_set(phn1, NULL, offset);
+	phn0 = phn_merge(phn0, phn1, offset, cmp);
+	phn_next_set(phn0, next_phn1, offset);
+	if (next_phn1 != NULL) {
+		phn_prev_set(next_phn1, phn0, offset);
+	}
+	phn_next_set(ph->root, phn0, offset);
+	phn_prev_set(phn0, ph->root, offset);
+	return next_phn1 == NULL;
+}
+
+JEMALLOC_ALWAYS_INLINE void
+ph_insert(ph_t *ph, void *phn, size_t offset, ph_cmp_t cmp) {
+	phn_link_init(phn, offset);
+
+	/*
+	 * Treat the root as an aux list during insertion, and lazily merge
+	 * during a_prefix##remove_first().  For elements that are inserted,
+	 * then removed via a_prefix##remove() before the aux list is ever
+	 * processed, this makes insert/remove constant-time, whereas eager
+	 * merging would make insert O(log n).
+	 */
+	if (ph->root == NULL) {
+		ph->root = phn;
+	} else {
+		/*
+		 * As a special case, check to see if we can replace the root.
+		 * This is practically common in some important cases, and lets
+		 * us defer some insertions (hopefully, until the point where
+		 * some of the items in the aux list have been removed, savings
+		 * us from linking them at all).
+		 */
+		if (cmp(phn, ph->root) < 0) {
+			phn_lchild_set(phn, ph->root, offset);
+			phn_prev_set(ph->root, phn, offset);
+			ph->root = phn;
+			ph->auxcount = 0;
+			return;
+		}
+		ph->auxcount++;
+		phn_next_set(phn, phn_next_get(ph->root, offset), offset);
+		if (phn_next_get(ph->root, offset) != NULL) {
+			phn_prev_set(phn_next_get(ph->root, offset), phn,
+			    offset);
+		}
+		phn_prev_set(phn, ph->root, offset);
+		phn_next_set(ph->root, phn, offset);
+	}
+	if (ph->auxcount > 1) {
+		unsigned nmerges = ffs_zu(ph->auxcount - 1);
+		bool done = false;
+		for (unsigned i = 0; i < nmerges && !done; i++) {
+			done = ph_try_aux_merge_pair(ph, offset, cmp);
+		}
+	}
+}
+
+JEMALLOC_ALWAYS_INLINE void *
+ph_remove_first(ph_t *ph, size_t offset, ph_cmp_t cmp) {
+	void *ret;
+
+	if (ph->root == NULL) {
+		return NULL;
+	}
+	ph_merge_aux(ph, offset, cmp);
+	ret = ph->root;
+	ph->root = ph_merge_children(ph->root, offset, cmp);
+
+	return ret;
+
+}
+
+JEMALLOC_ALWAYS_INLINE void
+ph_remove(ph_t *ph, void *phn, size_t offset, ph_cmp_t cmp) {
+	void *replace;
+	void *parent;
+
+	if (ph->root == phn) {
+		/*
+		 * We can delete from aux list without merging it, but we need
+		 * to merge if we are dealing with the root node and it has
+		 * children.
+		 */
+		if (phn_lchild_get(phn, offset) == NULL) {
+			ph->root = phn_next_get(phn, offset);
+			if (ph->root != NULL) {
+				phn_prev_set(ph->root, NULL, offset);
+			}
+			return;
+		}
+		ph_merge_aux(ph, offset, cmp);
+		if (ph->root == phn) {
+			ph->root = ph_merge_children(ph->root, offset, cmp);
+			return;
+		}
+	}
+
+	/* Get parent (if phn is leftmost child) before mutating. */
+	if ((parent = phn_prev_get(phn, offset)) != NULL) {
+		if (phn_lchild_get(parent, offset) != phn) {
+			parent = NULL;
+		}
+	}
+	/* Find a possible replacement node, and link to parent. */
+	replace = ph_merge_children(phn, offset, cmp);
+	/* Set next/prev for sibling linked list. */
+	if (replace != NULL) {
+		if (parent != NULL) {
+			phn_prev_set(replace, parent, offset);
+			phn_lchild_set(parent, replace, offset);
+		} else {
+			phn_prev_set(replace, phn_prev_get(phn, offset),
+			    offset);
+			if (phn_prev_get(phn, offset) != NULL) {
+				phn_next_set(phn_prev_get(phn, offset), replace,
+				    offset);
+			}
+		}
+		phn_next_set(replace, phn_next_get(phn, offset), offset);
+		if (phn_next_get(phn, offset) != NULL) {
+			phn_prev_set(phn_next_get(phn, offset), replace,
+			    offset);
+		}
+	} else {
+		if (parent != NULL) {
+			void *next = phn_next_get(phn, offset);
+			phn_lchild_set(parent, next, offset);
+			if (next != NULL) {
+				phn_prev_set(next, parent, offset);
+			}
+		} else {
+			assert(phn_prev_get(phn, offset) != NULL);
+			phn_next_set(
+			    phn_prev_get(phn, offset),
+			    phn_next_get(phn, offset), offset);
+		}
+		if (phn_next_get(phn, offset) != NULL) {
+			phn_prev_set(
+			    phn_next_get(phn, offset),
+			    phn_prev_get(phn, offset), offset);
+		}
+	}
+}
+
+#define ph_structs(a_prefix, a_type)					\
+typedef struct {							\
+	phn_link_t link;						\
+} a_prefix##_link_t;							\
+									\
+typedef struct {							\
+	ph_t ph;							\
+} a_prefix##_t;
+
+/*
+ * The ph_proto() macro generates function prototypes that correspond to the
+ * functions generated by an equivalently parameterized call to ph_gen().
+ */
+#define ph_proto(a_attr, a_prefix, a_type)				\
+									\
+a_attr void a_prefix##_new(a_prefix##_t *ph);				\
+a_attr bool a_prefix##_empty(a_prefix##_t *ph);				\
+a_attr a_type *a_prefix##_first(a_prefix##_t *ph);			\
+a_attr a_type *a_prefix##_any(a_prefix##_t *ph);			\
+a_attr void a_prefix##_insert(a_prefix##_t *ph, a_type *phn);		\
+a_attr a_type *a_prefix##_remove_first(a_prefix##_t *ph);		\
+a_attr void a_prefix##_remove(a_prefix##_t *ph, a_type *phn);		\
+a_attr a_type *a_prefix##_remove_any(a_prefix##_t *ph);
+
+/* The ph_gen() macro generates a type-specific pairing heap implementation. */
+#define ph_gen(a_attr, a_prefix, a_type, a_field, a_cmp)		\
+JEMALLOC_ALWAYS_INLINE int						\
+a_prefix##_ph_cmp(void *a, void *b) {					\
+	return a_cmp((a_type *)a, (a_type *)b);				\
+}									\
+									\
+a_attr void								\
+a_prefix##_new(a_prefix##_t *ph) {					\
+	ph_new(&ph->ph);						\
+}									\
+									\
+a_attr bool								\
+a_prefix##_empty(a_prefix##_t *ph) {					\
+	return ph_empty(&ph->ph);					\
+}									\
+									\
+a_attr a_type *								\
+a_prefix##_first(a_prefix##_t *ph) {					\
+	return ph_first(&ph->ph, offsetof(a_type, a_field),		\
+	    &a_prefix##_ph_cmp);					\
+}									\
+									\
+a_attr a_type *								\
+a_prefix##_any(a_prefix##_t *ph) {					\
+	return ph_any(&ph->ph, offsetof(a_type, a_field));		\
+}									\
+									\
+a_attr void								\
+a_prefix##_insert(a_prefix##_t *ph, a_type *phn) {			\
+	ph_insert(&ph->ph, phn, offsetof(a_type, a_field),		\
+	    a_prefix##_ph_cmp);						\
+}									\
+									\
+a_attr a_type *								\
+a_prefix##_remove_first(a_prefix##_t *ph) {				\
+	return ph_remove_first(&ph->ph, offsetof(a_type, a_field),	\
+	    a_prefix##_ph_cmp);						\
+}									\
+									\
+a_attr void								\
+a_prefix##_remove(a_prefix##_t *ph, a_type *phn) {			\
+	ph_remove(&ph->ph, phn, offsetof(a_type, a_field),		\
+	    a_prefix##_ph_cmp);						\
+}									\
+									\
+a_attr a_type *								\
+a_prefix##_remove_any(a_prefix##_t *ph) {				\
+	a_type *ret = a_prefix##_any(ph);				\
+	if (ret != NULL) {						\
+		a_prefix##_remove(ph, ret);				\
+	}								\
+	return ret;							\
+}
+
+#endif /* JEMALLOC_INTERNAL_PH_H */

BeefRT/JEMalloc/include/jemalloc/internal/private_namespace.sh

@@ -0,0 +1,5 @@
+#!/bin/sh
+
+for symbol in `cat "$@"` ; do
+  echo "#define ${symbol} JEMALLOC_N(${symbol})"
+done

BeefRT/JEMalloc/include/jemalloc/internal/private_symbols.awk

@@ -0,0 +1,52 @@
+#!/usr/bin/env awk -f
+
+BEGIN {
+  sym_prefix = ""
+  split("\
+        je_aligned_alloc \
+        je_calloc \
+        je_dallocx \
+        je_free \
+        je_mallctl \
+        je_mallctlbymib \
+        je_mallctlnametomib \
+        je_malloc \
+        je_malloc_conf \
+        je_malloc_conf_2_conf_harder \
+        je_malloc_message \
+        je_malloc_stats_print \
+        je_malloc_usable_size \
+        je_mallocx \
+        je_smallocx_54eaed1d8b56b1aa528be3bdd1877e59c56fa90c \
+        je_nallocx \
+        je_posix_memalign \
+        je_rallocx \
+        je_realloc \
+        je_sallocx \
+        je_sdallocx \
+        je_xallocx \
+        tls_callback \
+        ", exported_symbol_names)
+  # Store exported symbol names as keys in exported_symbols.
+  for (i in exported_symbol_names) {
+    exported_symbols[exported_symbol_names[i]] = 1
+  }
+}
+
+# Process 'nm -a <c_source.o>' output.
+#
+# Handle lines like:
+#   0000000000000008 D opt_junk
+#   0000000000007574 T malloc_initialized
+(NF == 3 && $2 ~ /^[ABCDGRSTVW]$/ && !($3 in exported_symbols) && $3 ~ /^[A-Za-z0-9_]+$/) {
+  print substr($3, 1+length(sym_prefix), length($3)-length(sym_prefix))
+}
+
+# Process 'dumpbin /SYMBOLS <c_source.obj>' output.
+#
+# Handle lines like:
+#   353 00008098 SECT4  notype       External     | opt_junk
+#   3F1 00000000 SECT7  notype ()    External     | malloc_initialized
+($3 ~ /^SECT[0-9]+/ && $(NF-2) == "External" && !($NF in exported_symbols)) {
+  print $NF
+}

BeefRT/JEMalloc/include/jemalloc/internal/private_symbols.sh

@@ -0,0 +1,51 @@
+#!/bin/sh
+#
+# Generate private_symbols[_jet].awk.
+#
+# Usage: private_symbols.sh <sym_prefix> <sym>*
+#
+# <sym_prefix> is typically "" or "_".
+
+sym_prefix=$1
+shift
+
+cat <<EOF
+#!/usr/bin/env awk -f
+
+BEGIN {
+  sym_prefix = "${sym_prefix}"
+  split("\\
+EOF
+
+for public_sym in "$@" ; do
+  cat <<EOF
+        ${sym_prefix}${public_sym} \\
+EOF
+done
+
+cat <<"EOF"
+        ", exported_symbol_names)
+  # Store exported symbol names as keys in exported_symbols.
+  for (i in exported_symbol_names) {
+    exported_symbols[exported_symbol_names[i]] = 1
+  }
+}
+
+# Process 'nm -a <c_source.o>' output.
+#
+# Handle lines like:
+#   0000000000000008 D opt_junk
+#   0000000000007574 T malloc_initialized
+(NF == 3 && $2 ~ /^[ABCDGRSTVW]$/ && !($3 in exported_symbols) && $3 ~ /^[A-Za-z0-9_]+$/) {
+  print substr($3, 1+length(sym_prefix), length($3)-length(sym_prefix))
+}
+
+# Process 'dumpbin /SYMBOLS <c_source.obj>' output.
+#
+# Handle lines like:
+#   353 00008098 SECT4  notype       External     | opt_junk
+#   3F1 00000000 SECT7  notype ()    External     | malloc_initialized
+($3 ~ /^SECT[0-9]+/ && $(NF-2) == "External" && !($NF in exported_symbols)) {
+  print $NF
+}
+EOF

BeefRT/JEMalloc/include/jemalloc/internal/private_symbols_jet.awk

@@ -0,0 +1,52 @@
+#!/usr/bin/env awk -f
+
+BEGIN {
+  sym_prefix = ""
+  split("\
+        jet_aligned_alloc \
+        jet_calloc \
+        jet_dallocx \
+        jet_free \
+        jet_mallctl \
+        jet_mallctlbymib \
+        jet_mallctlnametomib \
+        jet_malloc \
+        jet_malloc_conf \
+        jet_malloc_conf_2_conf_harder \
+        jet_malloc_message \
+        jet_malloc_stats_print \
+        jet_malloc_usable_size \
+        jet_mallocx \
+        jet_smallocx_54eaed1d8b56b1aa528be3bdd1877e59c56fa90c \
+        jet_nallocx \
+        jet_posix_memalign \
+        jet_rallocx \
+        jet_realloc \
+        jet_sallocx \
+        jet_sdallocx \
+        jet_xallocx \
+        tls_callback \
+        ", exported_symbol_names)
+  # Store exported symbol names as keys in exported_symbols.
+  for (i in exported_symbol_names) {
+    exported_symbols[exported_symbol_names[i]] = 1
+  }
+}
+
+# Process 'nm -a <c_source.o>' output.
+#
+# Handle lines like:
+#   0000000000000008 D opt_junk
+#   0000000000007574 T malloc_initialized
+(NF == 3 && $2 ~ /^[ABCDGRSTVW]$/ && !($3 in exported_symbols) && $3 ~ /^[A-Za-z0-9_]+$/) {
+  print substr($3, 1+length(sym_prefix), length($3)-length(sym_prefix))
+}
+
+# Process 'dumpbin /SYMBOLS <c_source.obj>' output.
+#
+# Handle lines like:
+#   353 00008098 SECT4  notype       External     | opt_junk
+#   3F1 00000000 SECT7  notype ()    External     | malloc_initialized
+($3 ~ /^SECT[0-9]+/ && $(NF-2) == "External" && !($NF in exported_symbols)) {
+  print $NF
+}

BeefRT/JEMalloc/include/jemalloc/internal/prng.h

@@ -0,0 +1,168 @@
+#ifndef JEMALLOC_INTERNAL_PRNG_H
+#define JEMALLOC_INTERNAL_PRNG_H
+
+#include "jemalloc/internal/bit_util.h"
+
+/*
+ * Simple linear congruential pseudo-random number generator:
+ *
+ *   prng(y) = (a*x + c) % m
+ *
+ * where the following constants ensure maximal period:
+ *
+ *   a == Odd number (relatively prime to 2^n), and (a-1) is a multiple of 4.
+ *   c == Odd number (relatively prime to 2^n).
+ *   m == 2^32
+ *
+ * See Knuth's TAOCP 3rd Ed., Vol. 2, pg. 17 for details on these constraints.
+ *
+ * This choice of m has the disadvantage that the quality of the bits is
+ * proportional to bit position.  For example, the lowest bit has a cycle of 2,
+ * the next has a cycle of 4, etc.  For this reason, we prefer to use the upper
+ * bits.
+ */
+
+/******************************************************************************/
+/* INTERNAL DEFINITIONS -- IGNORE */
+/******************************************************************************/
+#define PRNG_A_32	UINT32_C(1103515241)
+#define PRNG_C_32	UINT32_C(12347)
+
+#define PRNG_A_64	UINT64_C(6364136223846793005)
+#define PRNG_C_64	UINT64_C(1442695040888963407)
+
+JEMALLOC_ALWAYS_INLINE uint32_t
+prng_state_next_u32(uint32_t state) {
+	return (state * PRNG_A_32) + PRNG_C_32;
+}
+
+JEMALLOC_ALWAYS_INLINE uint64_t
+prng_state_next_u64(uint64_t state) {
+	return (state * PRNG_A_64) + PRNG_C_64;
+}
+
+JEMALLOC_ALWAYS_INLINE size_t
+prng_state_next_zu(size_t state) {
+#if LG_SIZEOF_PTR == 2
+	return (state * PRNG_A_32) + PRNG_C_32;
+#elif LG_SIZEOF_PTR == 3
+	return (state * PRNG_A_64) + PRNG_C_64;
+#else
+#error Unsupported pointer size
+#endif
+}
+
+/******************************************************************************/
+/* BEGIN PUBLIC API */
+/******************************************************************************/
+
+/*
+ * The prng_lg_range functions give a uniform int in the half-open range [0,
+ * 2**lg_range).
+ */
+
+JEMALLOC_ALWAYS_INLINE uint32_t
+prng_lg_range_u32(uint32_t *state, unsigned lg_range) {
+	assert(lg_range > 0);
+	assert(lg_range <= 32);
+
+	*state = prng_state_next_u32(*state);
+	uint32_t ret = *state >> (32 - lg_range);
+
+	return ret;
+}
+
+JEMALLOC_ALWAYS_INLINE uint64_t
+prng_lg_range_u64(uint64_t *state, unsigned lg_range) {
+	assert(lg_range > 0);
+	assert(lg_range <= 64);
+
+	*state = prng_state_next_u64(*state);
+	uint64_t ret = *state >> (64 - lg_range);
+
+	return ret;
+}
+
+JEMALLOC_ALWAYS_INLINE size_t
+prng_lg_range_zu(size_t *state, unsigned lg_range) {
+	assert(lg_range > 0);
+	assert(lg_range <= ZU(1) << (3 + LG_SIZEOF_PTR));
+
+	*state = prng_state_next_zu(*state);
+	size_t ret = *state >> ((ZU(1) << (3 + LG_SIZEOF_PTR)) - lg_range);
+
+	return ret;
+}
+
+/*
+ * The prng_range functions behave like the prng_lg_range, but return a result
+ * in [0, range) instead of [0, 2**lg_range).
+ */
+
+JEMALLOC_ALWAYS_INLINE uint32_t
+prng_range_u32(uint32_t *state, uint32_t range) {
+	assert(range != 0);
+	/*
+	 * If range were 1, lg_range would be 0, so the shift in
+	 * prng_lg_range_u32 would be a shift of a 32-bit variable by 32 bits,
+	 * which is UB.  Just handle this case as a one-off.
+	 */
+	if (range == 1) {
+		return 0;
+	}
+
+	/* Compute the ceiling of lg(range). */
+	unsigned lg_range = ffs_u32(pow2_ceil_u32(range));
+
+	/* Generate a result in [0..range) via repeated trial. */
+	uint32_t ret;
+	do {
+		ret = prng_lg_range_u32(state, lg_range);
+	} while (ret >= range);
+
+	return ret;
+}
+
+JEMALLOC_ALWAYS_INLINE uint64_t
+prng_range_u64(uint64_t *state, uint64_t range) {
+	assert(range != 0);
+
+	/* See the note in prng_range_u32. */
+	if (range == 1) {
+		return 0;
+	}
+
+	/* Compute the ceiling of lg(range). */
+	unsigned lg_range = ffs_u64(pow2_ceil_u64(range));
+
+	/* Generate a result in [0..range) via repeated trial. */
+	uint64_t ret;
+	do {
+		ret = prng_lg_range_u64(state, lg_range);
+	} while (ret >= range);
+
+	return ret;
+}
+
+JEMALLOC_ALWAYS_INLINE size_t
+prng_range_zu(size_t *state, size_t range) {
+	assert(range != 0);
+
+	/* See the note in prng_range_u32. */
+	if (range == 1) {
+		return 0;
+	}
+
+	/* Compute the ceiling of lg(range). */
+	unsigned lg_range = ffs_u64(pow2_ceil_u64(range));
+
+	/* Generate a result in [0..range) via repeated trial. */
+	size_t ret;
+	do {
+		ret = prng_lg_range_zu(state, lg_range);
+	} while (ret >= range);
+
+	return ret;
+}
+
+#endif /* JEMALLOC_INTERNAL_PRNG_H */

BeefRT/JEMalloc/include/jemalloc/internal/prof_data.h

@@ -0,0 +1,37 @@
+#ifndef JEMALLOC_INTERNAL_PROF_DATA_H
+#define JEMALLOC_INTERNAL_PROF_DATA_H
+
+#include "jemalloc/internal/mutex.h"
+
+extern malloc_mutex_t bt2gctx_mtx;
+extern malloc_mutex_t tdatas_mtx;
+extern malloc_mutex_t prof_dump_mtx;
+
+extern malloc_mutex_t *gctx_locks;
+extern malloc_mutex_t *tdata_locks;
+
+extern size_t prof_unbiased_sz[PROF_SC_NSIZES];
+extern size_t prof_shifted_unbiased_cnt[PROF_SC_NSIZES];
+
+void prof_bt_hash(const void *key, size_t r_hash[2]);
+bool prof_bt_keycomp(const void *k1, const void *k2);
+
+bool prof_data_init(tsd_t *tsd);
+prof_tctx_t *prof_lookup(tsd_t *tsd, prof_bt_t *bt);
+char *prof_thread_name_alloc(tsd_t *tsd, const char *thread_name);
+int prof_thread_name_set_impl(tsd_t *tsd, const char *thread_name);
+void prof_unbias_map_init();
+void prof_dump_impl(tsd_t *tsd, write_cb_t *prof_dump_write, void *cbopaque,
+    prof_tdata_t *tdata, bool leakcheck);
+prof_tdata_t * prof_tdata_init_impl(tsd_t *tsd, uint64_t thr_uid,
+    uint64_t thr_discrim, char *thread_name, bool active);
+void prof_tdata_detach(tsd_t *tsd, prof_tdata_t *tdata);
+void prof_reset(tsd_t *tsd, size_t lg_sample);
+void prof_tctx_try_destroy(tsd_t *tsd, prof_tctx_t *tctx);
+
+/* Used in unit tests. */
+size_t prof_tdata_count(void);
+size_t prof_bt_count(void);
+void prof_cnt_all(prof_cnt_t *cnt_all);
+
+#endif /* JEMALLOC_INTERNAL_PROF_DATA_H */

BeefRT/JEMalloc/include/jemalloc/internal/prof_externs.h

@@ -0,0 +1,95 @@
+#ifndef JEMALLOC_INTERNAL_PROF_EXTERNS_H
+#define JEMALLOC_INTERNAL_PROF_EXTERNS_H
+
+#include "jemalloc/internal/mutex.h"
+#include "jemalloc/internal/prof_hook.h"
+
+extern bool opt_prof;
+extern bool opt_prof_active;
+extern bool opt_prof_thread_active_init;
+extern size_t opt_lg_prof_sample;    /* Mean bytes between samples. */
+extern ssize_t opt_lg_prof_interval; /* lg(prof_interval). */
+extern bool opt_prof_gdump;          /* High-water memory dumping. */
+extern bool opt_prof_final;          /* Final profile dumping. */
+extern bool opt_prof_leak;           /* Dump leak summary at exit. */
+extern bool opt_prof_leak_error;     /* Exit with error code if memory leaked */
+extern bool opt_prof_accum;          /* Report cumulative bytes. */
+extern bool opt_prof_log;            /* Turn logging on at boot. */
+extern char opt_prof_prefix[
+    /* Minimize memory bloat for non-prof builds. */
+#ifdef JEMALLOC_PROF
+    PATH_MAX +
+#endif
+    1];
+extern bool opt_prof_unbias;
+
+/* For recording recent allocations */
+extern ssize_t opt_prof_recent_alloc_max;
+
+/* Whether to use thread name provided by the system or by mallctl. */
+extern bool opt_prof_sys_thread_name;
+
+/* Whether to record per size class counts and request size totals. */
+extern bool opt_prof_stats;
+
+/* Accessed via prof_active_[gs]et{_unlocked,}(). */
+extern bool prof_active_state;
+
+/* Accessed via prof_gdump_[gs]et{_unlocked,}(). */
+extern bool prof_gdump_val;
+
+/* Profile dump interval, measured in bytes allocated. */
+extern uint64_t prof_interval;
+
+/*
+ * Initialized as opt_lg_prof_sample, and potentially modified during profiling
+ * resets.
+ */
+extern size_t lg_prof_sample;
+
+extern bool prof_booted;
+
+void prof_backtrace_hook_set(prof_backtrace_hook_t hook);
+prof_backtrace_hook_t prof_backtrace_hook_get();
+
+void prof_dump_hook_set(prof_dump_hook_t hook);
+prof_dump_hook_t prof_dump_hook_get();
+
+/* Functions only accessed in prof_inlines.h */
+prof_tdata_t *prof_tdata_init(tsd_t *tsd);
+prof_tdata_t *prof_tdata_reinit(tsd_t *tsd, prof_tdata_t *tdata);
+
+void prof_alloc_rollback(tsd_t *tsd, prof_tctx_t *tctx);
+void prof_malloc_sample_object(tsd_t *tsd, const void *ptr, size_t size,
+    size_t usize, prof_tctx_t *tctx);
+void prof_free_sampled_object(tsd_t *tsd, size_t usize, prof_info_t *prof_info);
+prof_tctx_t *prof_tctx_create(tsd_t *tsd);
+void prof_idump(tsdn_t *tsdn);
+bool prof_mdump(tsd_t *tsd, const char *filename);
+void prof_gdump(tsdn_t *tsdn);
+
+void prof_tdata_cleanup(tsd_t *tsd);
+bool prof_active_get(tsdn_t *tsdn);
+bool prof_active_set(tsdn_t *tsdn, bool active);
+const char *prof_thread_name_get(tsd_t *tsd);
+int prof_thread_name_set(tsd_t *tsd, const char *thread_name);
+bool prof_thread_active_get(tsd_t *tsd);
+bool prof_thread_active_set(tsd_t *tsd, bool active);
+bool prof_thread_active_init_get(tsdn_t *tsdn);
+bool prof_thread_active_init_set(tsdn_t *tsdn, bool active_init);
+bool prof_gdump_get(tsdn_t *tsdn);
+bool prof_gdump_set(tsdn_t *tsdn, bool active);
+void prof_boot0(void);
+void prof_boot1(void);
+bool prof_boot2(tsd_t *tsd, base_t *base);
+void prof_prefork0(tsdn_t *tsdn);
+void prof_prefork1(tsdn_t *tsdn);
+void prof_postfork_parent(tsdn_t *tsdn);
+void prof_postfork_child(tsdn_t *tsdn);
+
+/* Only accessed by thread event. */
+uint64_t prof_sample_new_event_wait(tsd_t *tsd);
+uint64_t prof_sample_postponed_event_wait(tsd_t *tsd);
+void prof_sample_event_handler(tsd_t *tsd, uint64_t elapsed);
+
+#endif /* JEMALLOC_INTERNAL_PROF_EXTERNS_H */

BeefRT/JEMalloc/include/jemalloc/internal/prof_hook.h

@@ -0,0 +1,21 @@
+#ifndef JEMALLOC_INTERNAL_PROF_HOOK_H
+#define JEMALLOC_INTERNAL_PROF_HOOK_H
+
+/*
+ * The hooks types of which are declared in this file are experimental and
+ * undocumented, thus the typedefs are located in an 'internal' header.
+ */
+
+/*
+ * A hook to mock out backtrace functionality.  This can be handy, since it's
+ * otherwise difficult to guarantee that two allocations are reported as coming
+ * from the exact same stack trace in the presence of an optimizing compiler.
+ */
+typedef void (*prof_backtrace_hook_t)(void **, unsigned *, unsigned);
+
+/*
+ * A callback hook that notifies about recently dumped heap profile.
+ */
+typedef void (*prof_dump_hook_t)(const char *filename);
+
+#endif /* JEMALLOC_INTERNAL_PROF_HOOK_H */

BeefRT/JEMalloc/include/jemalloc/internal/prof_inlines.h

@@ -0,0 +1,261 @@
+#ifndef JEMALLOC_INTERNAL_PROF_INLINES_H
+#define JEMALLOC_INTERNAL_PROF_INLINES_H
+
+#include "jemalloc/internal/safety_check.h"
+#include "jemalloc/internal/sz.h"
+#include "jemalloc/internal/thread_event.h"
+
+JEMALLOC_ALWAYS_INLINE void
+prof_active_assert() {
+	cassert(config_prof);
+	/*
+	 * If opt_prof is off, then prof_active must always be off, regardless
+	 * of whether prof_active_mtx is in effect or not.
+	 */
+	assert(opt_prof || !prof_active_state);
+}
+
+JEMALLOC_ALWAYS_INLINE bool
+prof_active_get_unlocked(void) {
+	prof_active_assert();
+	/*
+	 * Even if opt_prof is true, sampling can be temporarily disabled by
+	 * setting prof_active to false.  No locking is used when reading
+	 * prof_active in the fast path, so there are no guarantees regarding
+	 * how long it will take for all threads to notice state changes.
+	 */
+	return prof_active_state;
+}
+
+JEMALLOC_ALWAYS_INLINE bool
+prof_gdump_get_unlocked(void) {
+	/*
+	 * No locking is used when reading prof_gdump_val in the fast path, so
+	 * there are no guarantees regarding how long it will take for all
+	 * threads to notice state changes.
+	 */
+	return prof_gdump_val;
+}
+
+JEMALLOC_ALWAYS_INLINE prof_tdata_t *
+prof_tdata_get(tsd_t *tsd, bool create) {
+	prof_tdata_t *tdata;
+
+	cassert(config_prof);
+
+	tdata = tsd_prof_tdata_get(tsd);
+	if (create) {
+		assert(tsd_reentrancy_level_get(tsd) == 0);
+		if (unlikely(tdata == NULL)) {
+			if (tsd_nominal(tsd)) {
+				tdata = prof_tdata_init(tsd);
+				tsd_prof_tdata_set(tsd, tdata);
+			}
+		} else if (unlikely(tdata->expired)) {
+			tdata = prof_tdata_reinit(tsd, tdata);
+			tsd_prof_tdata_set(tsd, tdata);
+		}
+		assert(tdata == NULL || tdata->attached);
+	}
+
+	return tdata;
+}
+
+JEMALLOC_ALWAYS_INLINE void
+prof_info_get(tsd_t *tsd, const void *ptr, emap_alloc_ctx_t *alloc_ctx,
+    prof_info_t *prof_info) {
+	cassert(config_prof);
+	assert(ptr != NULL);
+	assert(prof_info != NULL);
+
+	arena_prof_info_get(tsd, ptr, alloc_ctx, prof_info, false);
+}
+
+JEMALLOC_ALWAYS_INLINE void
+prof_info_get_and_reset_recent(tsd_t *tsd, const void *ptr,
+    emap_alloc_ctx_t *alloc_ctx, prof_info_t *prof_info) {
+	cassert(config_prof);
+	assert(ptr != NULL);
+	assert(prof_info != NULL);
+
+	arena_prof_info_get(tsd, ptr, alloc_ctx, prof_info, true);
+}
+
+JEMALLOC_ALWAYS_INLINE void
+prof_tctx_reset(tsd_t *tsd, const void *ptr, emap_alloc_ctx_t *alloc_ctx) {
+	cassert(config_prof);
+	assert(ptr != NULL);
+
+	arena_prof_tctx_reset(tsd, ptr, alloc_ctx);
+}
+
+JEMALLOC_ALWAYS_INLINE void
+prof_tctx_reset_sampled(tsd_t *tsd, const void *ptr) {
+	cassert(config_prof);
+	assert(ptr != NULL);
+
+	arena_prof_tctx_reset_sampled(tsd, ptr);
+}
+
+JEMALLOC_ALWAYS_INLINE void
+prof_info_set(tsd_t *tsd, edata_t *edata, prof_tctx_t *tctx, size_t size) {
+	cassert(config_prof);
+	assert(edata != NULL);
+	assert((uintptr_t)tctx > (uintptr_t)1U);
+
+	arena_prof_info_set(tsd, edata, tctx, size);
+}
+
+JEMALLOC_ALWAYS_INLINE bool
+prof_sample_should_skip(tsd_t *tsd, bool sample_event) {
+	cassert(config_prof);
+
+	/* Fastpath: no need to load tdata */
+	if (likely(!sample_event)) {
+		return true;
+	}
+
+	/*
+	 * sample_event is always obtained from the thread event module, and
+	 * whenever it's true, it means that the thread event module has
+	 * already checked the reentrancy level.
+	 */
+	assert(tsd_reentrancy_level_get(tsd) == 0);
+
+	prof_tdata_t *tdata = prof_tdata_get(tsd, true);
+	if (unlikely(tdata == NULL)) {
+		return true;
+	}
+
+	return !tdata->active;
+}
+
+JEMALLOC_ALWAYS_INLINE prof_tctx_t *
+prof_alloc_prep(tsd_t *tsd, bool prof_active, bool sample_event) {
+	prof_tctx_t *ret;
+
+	if (!prof_active ||
+	    likely(prof_sample_should_skip(tsd, sample_event))) {
+		ret = (prof_tctx_t *)(uintptr_t)1U;
+	} else {
+		ret = prof_tctx_create(tsd);
+	}
+
+	return ret;
+}
+
+JEMALLOC_ALWAYS_INLINE void
+prof_malloc(tsd_t *tsd, const void *ptr, size_t size, size_t usize,
+    emap_alloc_ctx_t *alloc_ctx, prof_tctx_t *tctx) {
+	cassert(config_prof);
+	assert(ptr != NULL);
+	assert(usize == isalloc(tsd_tsdn(tsd), ptr));
+
+	if (unlikely((uintptr_t)tctx > (uintptr_t)1U)) {
+		prof_malloc_sample_object(tsd, ptr, size, usize, tctx);
+	} else {
+		prof_tctx_reset(tsd, ptr, alloc_ctx);
+	}
+}
+
+JEMALLOC_ALWAYS_INLINE void
+prof_realloc(tsd_t *tsd, const void *ptr, size_t size, size_t usize,
+    prof_tctx_t *tctx, bool prof_active, const void *old_ptr, size_t old_usize,
+    prof_info_t *old_prof_info, bool sample_event) {
+	bool sampled, old_sampled, moved;
+
+	cassert(config_prof);
+	assert(ptr != NULL || (uintptr_t)tctx <= (uintptr_t)1U);
+
+	if (prof_active && ptr != NULL) {
+		assert(usize == isalloc(tsd_tsdn(tsd), ptr));
+		if (prof_sample_should_skip(tsd, sample_event)) {
+			/*
+			 * Don't sample.  The usize passed to prof_alloc_prep()
+			 * was larger than what actually got allocated, so a
+			 * backtrace was captured for this allocation, even
+			 * though its actual usize was insufficient to cross the
+			 * sample threshold.
+			 */
+			prof_alloc_rollback(tsd, tctx);
+			tctx = (prof_tctx_t *)(uintptr_t)1U;
+		}
+	}
+
+	sampled = ((uintptr_t)tctx > (uintptr_t)1U);
+	old_sampled = ((uintptr_t)old_prof_info->alloc_tctx > (uintptr_t)1U);
+	moved = (ptr != old_ptr);
+
+	if (unlikely(sampled)) {
+		prof_malloc_sample_object(tsd, ptr, size, usize, tctx);
+	} else if (moved) {
+		prof_tctx_reset(tsd, ptr, NULL);
+	} else if (unlikely(old_sampled)) {
+		/*
+		 * prof_tctx_reset() would work for the !moved case as well,
+		 * but prof_tctx_reset_sampled() is slightly cheaper, and the
+		 * proper thing to do here in the presence of explicit
+		 * knowledge re: moved state.
+		 */
+		prof_tctx_reset_sampled(tsd, ptr);
+	} else {
+		prof_info_t prof_info;
+		prof_info_get(tsd, ptr, NULL, &prof_info);
+		assert((uintptr_t)prof_info.alloc_tctx == (uintptr_t)1U);
+	}
+
+	/*
+	 * The prof_free_sampled_object() call must come after the
+	 * prof_malloc_sample_object() call, because tctx and old_tctx may be
+	 * the same, in which case reversing the call order could cause the tctx
+	 * to be prematurely destroyed as a side effect of momentarily zeroed
+	 * counters.
+	 */
+	if (unlikely(old_sampled)) {
+		prof_free_sampled_object(tsd, old_usize, old_prof_info);
+	}
+}
+
+JEMALLOC_ALWAYS_INLINE size_t
+prof_sample_align(size_t orig_align) {
+	/*
+	 * Enforce page alignment, so that sampled allocations can be identified
+	 * w/o metadata lookup.
+	 */
+	assert(opt_prof);
+	return (opt_cache_oblivious && orig_align < PAGE) ? PAGE :
+	    orig_align;
+}
+
+JEMALLOC_ALWAYS_INLINE bool
+prof_sample_aligned(const void *ptr) {
+	return ((uintptr_t)ptr & PAGE_MASK) == 0;
+}
+
+JEMALLOC_ALWAYS_INLINE bool
+prof_sampled(tsd_t *tsd, const void *ptr) {
+	prof_info_t prof_info;
+	prof_info_get(tsd, ptr, NULL, &prof_info);
+	bool sampled = (uintptr_t)prof_info.alloc_tctx > (uintptr_t)1U;
+	if (sampled) {
+		assert(prof_sample_aligned(ptr));
+	}
+	return sampled;
+}
+
+JEMALLOC_ALWAYS_INLINE void
+prof_free(tsd_t *tsd, const void *ptr, size_t usize,
+    emap_alloc_ctx_t *alloc_ctx) {
+	prof_info_t prof_info;
+	prof_info_get_and_reset_recent(tsd, ptr, alloc_ctx, &prof_info);
+
+	cassert(config_prof);
+	assert(usize == isalloc(tsd_tsdn(tsd), ptr));
+
+	if (unlikely((uintptr_t)prof_info.alloc_tctx > (uintptr_t)1U)) {
+		assert(prof_sample_aligned(ptr));
+		prof_free_sampled_object(tsd, usize, &prof_info);
+	}
+}
+
+#endif /* JEMALLOC_INTERNAL_PROF_INLINES_H */

BeefRT/JEMalloc/include/jemalloc/internal/prof_log.h

@@ -0,0 +1,22 @@
+#ifndef JEMALLOC_INTERNAL_PROF_LOG_H
+#define JEMALLOC_INTERNAL_PROF_LOG_H
+
+#include "jemalloc/internal/mutex.h"
+
+extern malloc_mutex_t log_mtx;
+
+void prof_try_log(tsd_t *tsd, size_t usize, prof_info_t *prof_info);
+bool prof_log_init(tsd_t *tsdn);
+
+/* Used in unit tests. */
+size_t prof_log_bt_count(void);
+size_t prof_log_alloc_count(void);
+size_t prof_log_thr_count(void);
+bool prof_log_is_logging(void);
+bool prof_log_rep_check(void);
+void prof_log_dummy_set(bool new_value);
+
+bool prof_log_start(tsdn_t *tsdn, const char *filename);
+bool prof_log_stop(tsdn_t *tsdn);
+
+#endif /* JEMALLOC_INTERNAL_PROF_LOG_H */

BeefRT/JEMalloc/include/jemalloc/internal/prof_recent.h

@@ -0,0 +1,23 @@
+#ifndef JEMALLOC_INTERNAL_PROF_RECENT_H
+#define JEMALLOC_INTERNAL_PROF_RECENT_H
+
+extern malloc_mutex_t prof_recent_alloc_mtx;
+extern malloc_mutex_t prof_recent_dump_mtx;
+
+bool prof_recent_alloc_prepare(tsd_t *tsd, prof_tctx_t *tctx);
+void prof_recent_alloc(tsd_t *tsd, edata_t *edata, size_t size, size_t usize);
+void prof_recent_alloc_reset(tsd_t *tsd, edata_t *edata);
+bool prof_recent_init();
+void edata_prof_recent_alloc_init(edata_t *edata);
+
+/* Used in unit tests. */
+typedef ql_head(prof_recent_t) prof_recent_list_t;
+extern prof_recent_list_t prof_recent_alloc_list;
+edata_t *prof_recent_alloc_edata_get_no_lock_test(const prof_recent_t *node);
+prof_recent_t *edata_prof_recent_alloc_get_no_lock_test(const edata_t *edata);
+
+ssize_t prof_recent_alloc_max_ctl_read();
+ssize_t prof_recent_alloc_max_ctl_write(tsd_t *tsd, ssize_t max);
+void prof_recent_alloc_dump(tsd_t *tsd, write_cb_t *write_cb, void *cbopaque);
+
+#endif /* JEMALLOC_INTERNAL_PROF_RECENT_H */

BeefRT/JEMalloc/include/jemalloc/internal/prof_stats.h

@@ -0,0 +1,17 @@
+#ifndef JEMALLOC_INTERNAL_PROF_STATS_H
+#define JEMALLOC_INTERNAL_PROF_STATS_H
+
+typedef struct prof_stats_s prof_stats_t;
+struct prof_stats_s {
+	uint64_t req_sum;
+	uint64_t count;
+};
+
+extern malloc_mutex_t prof_stats_mtx;
+
+void prof_stats_inc(tsd_t *tsd, szind_t ind, size_t size);
+void prof_stats_dec(tsd_t *tsd, szind_t ind, size_t size);
+void prof_stats_get_live(tsd_t *tsd, szind_t ind, prof_stats_t *stats);
+void prof_stats_get_accum(tsd_t *tsd, szind_t ind, prof_stats_t *stats);
+
+#endif /* JEMALLOC_INTERNAL_PROF_STATS_H */

BeefRT/JEMalloc/include/jemalloc/internal/prof_structs.h

@@ -0,0 +1,221 @@
+#ifndef JEMALLOC_INTERNAL_PROF_STRUCTS_H
+#define JEMALLOC_INTERNAL_PROF_STRUCTS_H
+
+#include "jemalloc/internal/ckh.h"
+#include "jemalloc/internal/edata.h"
+#include "jemalloc/internal/mutex.h"
+#include "jemalloc/internal/prng.h"
+#include "jemalloc/internal/rb.h"
+
+struct prof_bt_s {
+	/* Backtrace, stored as len program counters. */
+	void		**vec;
+	unsigned	len;
+};
+
+#ifdef JEMALLOC_PROF_LIBGCC
+/* Data structure passed to libgcc _Unwind_Backtrace() callback functions. */
+typedef struct {
+	void 		**vec;
+	unsigned	*len;
+	unsigned	max;
+} prof_unwind_data_t;
+#endif
+
+struct prof_cnt_s {
+	/* Profiling counters. */
+	uint64_t	curobjs;
+	uint64_t	curobjs_shifted_unbiased;
+	uint64_t	curbytes;
+	uint64_t	curbytes_unbiased;
+	uint64_t	accumobjs;
+	uint64_t	accumobjs_shifted_unbiased;
+	uint64_t	accumbytes;
+	uint64_t	accumbytes_unbiased;
+};
+
+typedef enum {
+	prof_tctx_state_initializing,
+	prof_tctx_state_nominal,
+	prof_tctx_state_dumping,
+	prof_tctx_state_purgatory /* Dumper must finish destroying. */
+} prof_tctx_state_t;
+
+struct prof_tctx_s {
+	/* Thread data for thread that performed the allocation. */
+	prof_tdata_t		*tdata;
+
+	/*
+	 * Copy of tdata->thr_{uid,discrim}, necessary because tdata may be
+	 * defunct during teardown.
+	 */
+	uint64_t		thr_uid;
+	uint64_t		thr_discrim;
+
+	/*
+	 * Reference count of how many times this tctx object is referenced in
+	 * recent allocation / deallocation records, protected by tdata->lock.
+	 */
+	uint64_t		recent_count;
+
+	/* Profiling counters, protected by tdata->lock. */
+	prof_cnt_t		cnts;
+
+	/* Associated global context. */
+	prof_gctx_t		*gctx;
+
+	/*
+	 * UID that distinguishes multiple tctx's created by the same thread,
+	 * but coexisting in gctx->tctxs.  There are two ways that such
+	 * coexistence can occur:
+	 * - A dumper thread can cause a tctx to be retained in the purgatory
+	 *   state.
+	 * - Although a single "producer" thread must create all tctx's which
+	 *   share the same thr_uid, multiple "consumers" can each concurrently
+	 *   execute portions of prof_tctx_destroy().  prof_tctx_destroy() only
+	 *   gets called once each time cnts.cur{objs,bytes} drop to 0, but this
+	 *   threshold can be hit again before the first consumer finishes
+	 *   executing prof_tctx_destroy().
+	 */
+	uint64_t		tctx_uid;
+
+	/* Linkage into gctx's tctxs. */
+	rb_node(prof_tctx_t)	tctx_link;
+
+	/*
+	 * True during prof_alloc_prep()..prof_malloc_sample_object(), prevents
+	 * sample vs destroy race.
+	 */
+	bool			prepared;
+
+	/* Current dump-related state, protected by gctx->lock. */
+	prof_tctx_state_t	state;
+
+	/*
+	 * Copy of cnts snapshotted during early dump phase, protected by
+	 * dump_mtx.
+	 */
+	prof_cnt_t		dump_cnts;
+};
+typedef rb_tree(prof_tctx_t) prof_tctx_tree_t;
+
+struct prof_info_s {
+	/* Time when the allocation was made. */
+	nstime_t		alloc_time;
+	/* Points to the prof_tctx_t corresponding to the allocation. */
+	prof_tctx_t		*alloc_tctx;
+	/* Allocation request size. */
+	size_t			alloc_size;
+};
+
+struct prof_gctx_s {
+	/* Protects nlimbo, cnt_summed, and tctxs. */
+	malloc_mutex_t		*lock;
+
+	/*
+	 * Number of threads that currently cause this gctx to be in a state of
+	 * limbo due to one of:
+	 *   - Initializing this gctx.
+	 *   - Initializing per thread counters associated with this gctx.
+	 *   - Preparing to destroy this gctx.
+	 *   - Dumping a heap profile that includes this gctx.
+	 * nlimbo must be 1 (single destroyer) in order to safely destroy the
+	 * gctx.
+	 */
+	unsigned		nlimbo;
+
+	/*
+	 * Tree of profile counters, one for each thread that has allocated in
+	 * this context.
+	 */
+	prof_tctx_tree_t	tctxs;
+
+	/* Linkage for tree of contexts to be dumped. */
+	rb_node(prof_gctx_t)	dump_link;
+
+	/* Temporary storage for summation during dump. */
+	prof_cnt_t		cnt_summed;
+
+	/* Associated backtrace. */
+	prof_bt_t		bt;
+
+	/* Backtrace vector, variable size, referred to by bt. */
+	void			*vec[1];
+};
+typedef rb_tree(prof_gctx_t) prof_gctx_tree_t;
+
+struct prof_tdata_s {
+	malloc_mutex_t		*lock;
+
+	/* Monotonically increasing unique thread identifier. */
+	uint64_t		thr_uid;
+
+	/*
+	 * Monotonically increasing discriminator among tdata structures
+	 * associated with the same thr_uid.
+	 */
+	uint64_t		thr_discrim;
+
+	/* Included in heap profile dumps if non-NULL. */
+	char			*thread_name;
+
+	bool			attached;
+	bool			expired;
+
+	rb_node(prof_tdata_t)	tdata_link;
+
+	/*
+	 * Counter used to initialize prof_tctx_t's tctx_uid.  No locking is
+	 * necessary when incrementing this field, because only one thread ever
+	 * does so.
+	 */
+	uint64_t		tctx_uid_next;
+
+	/*
+	 * Hash of (prof_bt_t *)-->(prof_tctx_t *).  Each thread tracks
+	 * backtraces for which it has non-zero allocation/deallocation counters
+	 * associated with thread-specific prof_tctx_t objects.  Other threads
+	 * may write to prof_tctx_t contents when freeing associated objects.
+	 */
+	ckh_t			bt2tctx;
+
+	/* State used to avoid dumping while operating on prof internals. */
+	bool			enq;
+	bool			enq_idump;
+	bool			enq_gdump;
+
+	/*
+	 * Set to true during an early dump phase for tdata's which are
+	 * currently being dumped.  New threads' tdata's have this initialized
+	 * to false so that they aren't accidentally included in later dump
+	 * phases.
+	 */
+	bool			dumping;
+
+	/*
+	 * True if profiling is active for this tdata's thread
+	 * (thread.prof.active mallctl).
+	 */
+	bool			active;
+
+	/* Temporary storage for summation during dump. */
+	prof_cnt_t		cnt_summed;
+
+	/* Backtrace vector, used for calls to prof_backtrace(). */
+	void			*vec[PROF_BT_MAX];
+};
+typedef rb_tree(prof_tdata_t) prof_tdata_tree_t;
+
+struct prof_recent_s {
+	nstime_t alloc_time;
+	nstime_t dalloc_time;
+
+	ql_elm(prof_recent_t) link;
+	size_t size;
+	size_t usize;
+	atomic_p_t alloc_edata; /* NULL means allocation has been freed. */
+	prof_tctx_t *alloc_tctx;
+	prof_tctx_t *dalloc_tctx;
+};
+
+#endif /* JEMALLOC_INTERNAL_PROF_STRUCTS_H */

BeefRT/JEMalloc/include/jemalloc/internal/prof_sys.h

@@ -0,0 +1,30 @@
+#ifndef JEMALLOC_INTERNAL_PROF_SYS_H
+#define JEMALLOC_INTERNAL_PROF_SYS_H
+
+extern malloc_mutex_t prof_dump_filename_mtx;
+extern base_t *prof_base;
+
+void bt_init(prof_bt_t *bt, void **vec);
+void prof_backtrace(tsd_t *tsd, prof_bt_t *bt);
+void prof_hooks_init();
+void prof_unwind_init();
+void prof_sys_thread_name_fetch(tsd_t *tsd);
+int prof_getpid(void);
+void prof_get_default_filename(tsdn_t *tsdn, char *filename, uint64_t ind);
+bool prof_prefix_set(tsdn_t *tsdn, const char *prefix);
+void prof_fdump_impl(tsd_t *tsd);
+void prof_idump_impl(tsd_t *tsd);
+bool prof_mdump_impl(tsd_t *tsd, const char *filename);
+void prof_gdump_impl(tsd_t *tsd);
+
+/* Used in unit tests. */
+typedef int (prof_sys_thread_name_read_t)(char *buf, size_t limit);
+extern prof_sys_thread_name_read_t *JET_MUTABLE prof_sys_thread_name_read;
+typedef int (prof_dump_open_file_t)(const char *, int);
+extern prof_dump_open_file_t *JET_MUTABLE prof_dump_open_file;
+typedef ssize_t (prof_dump_write_file_t)(int, const void *, size_t);
+extern prof_dump_write_file_t *JET_MUTABLE prof_dump_write_file;
+typedef int (prof_dump_open_maps_t)();
+extern prof_dump_open_maps_t *JET_MUTABLE prof_dump_open_maps;
+
+#endif /* JEMALLOC_INTERNAL_PROF_SYS_H */

BeefRT/JEMalloc/include/jemalloc/internal/prof_types.h

@@ -0,0 +1,75 @@
+#ifndef JEMALLOC_INTERNAL_PROF_TYPES_H
+#define JEMALLOC_INTERNAL_PROF_TYPES_H
+
+typedef struct prof_bt_s prof_bt_t;
+typedef struct prof_cnt_s prof_cnt_t;
+typedef struct prof_tctx_s prof_tctx_t;
+typedef struct prof_info_s prof_info_t;
+typedef struct prof_gctx_s prof_gctx_t;
+typedef struct prof_tdata_s prof_tdata_t;
+typedef struct prof_recent_s prof_recent_t;
+
+/* Option defaults. */
+#ifdef JEMALLOC_PROF
+#  define PROF_PREFIX_DEFAULT		"jeprof"
+#else
+#  define PROF_PREFIX_DEFAULT		""
+#endif
+#define LG_PROF_SAMPLE_DEFAULT		19
+#define LG_PROF_INTERVAL_DEFAULT	-1
+
+/*
+ * Hard limit on stack backtrace depth.  The version of prof_backtrace() that
+ * is based on __builtin_return_address() necessarily has a hard-coded number
+ * of backtrace frame handlers, and should be kept in sync with this setting.
+ */
+#define PROF_BT_MAX			128
+
+/* Initial hash table size. */
+#define PROF_CKH_MINITEMS		64
+
+/* Size of memory buffer to use when writing dump files. */
+#ifndef JEMALLOC_PROF
+/* Minimize memory bloat for non-prof builds. */
+#  define PROF_DUMP_BUFSIZE		1
+#elif defined(JEMALLOC_DEBUG)
+/* Use a small buffer size in debug build, mainly to facilitate testing. */
+#  define PROF_DUMP_BUFSIZE		16
+#else
+#  define PROF_DUMP_BUFSIZE		65536
+#endif
+
+/* Size of size class related tables */
+#ifdef JEMALLOC_PROF
+#  define PROF_SC_NSIZES		SC_NSIZES
+#else
+/* Minimize memory bloat for non-prof builds. */
+#  define PROF_SC_NSIZES		1
+#endif
+
+/* Size of stack-allocated buffer used by prof_printf(). */
+#define PROF_PRINTF_BUFSIZE		128
+
+/*
+ * Number of mutexes shared among all gctx's.  No space is allocated for these
+ * unless profiling is enabled, so it's okay to over-provision.
+ */
+#define PROF_NCTX_LOCKS			1024
+
+/*
+ * Number of mutexes shared among all tdata's.  No space is allocated for these
+ * unless profiling is enabled, so it's okay to over-provision.
+ */
+#define PROF_NTDATA_LOCKS		256
+
+/* Minimize memory bloat for non-prof builds. */
+#ifdef JEMALLOC_PROF
+#define PROF_DUMP_FILENAME_LEN (PATH_MAX + 1)
+#else
+#define PROF_DUMP_FILENAME_LEN 1
+#endif
+
+/* Default number of recent allocations to record. */
+#define PROF_RECENT_ALLOC_MAX_DEFAULT 0
+
+#endif /* JEMALLOC_INTERNAL_PROF_TYPES_H */

BeefRT/JEMalloc/include/jemalloc/internal/psset.h

@@ -0,0 +1,131 @@
+#ifndef JEMALLOC_INTERNAL_PSSET_H
+#define JEMALLOC_INTERNAL_PSSET_H
+
+#include "jemalloc/internal/hpdata.h"
+
+/*
+ * A page-slab set.  What the eset is to PAC, the psset is to HPA.  It maintains
+ * a collection of page-slabs (the intent being that they are backed by
+ * hugepages, or at least could be), and handles allocation and deallocation
+ * requests.
+ */
+
+/*
+ * One more than the maximum pszind_t we will serve out of the HPA.
+ * Practically, we expect only the first few to be actually used.  This
+ * corresponds to a maximum size of of 512MB on systems with 4k pages and
+ * SC_NGROUP == 4, which is already an unreasonably large maximum.  Morally, you
+ * can think of this as being SC_NPSIZES, but there's no sense in wasting that
+ * much space in the arena, making bitmaps that much larger, etc.
+ */
+#define PSSET_NPSIZES 64
+
+/*
+ * We keep two purge lists per page size class; one for hugified hpdatas (at
+ * index 2*pszind), and one for the non-hugified hpdatas (at index 2*pszind +
+ * 1).  This lets us implement a preference for purging non-hugified hpdatas
+ * among similarly-dirty ones.
+ * We reserve the last two indices for empty slabs, in that case purging
+ * hugified ones (which are definitionally all waste) before non-hugified ones
+ * (i.e. reversing the order).
+ */
+#define PSSET_NPURGE_LISTS (2 * PSSET_NPSIZES)
+
+typedef struct psset_bin_stats_s psset_bin_stats_t;
+struct psset_bin_stats_s {
+	/* How many pageslabs are in this bin? */
+	size_t npageslabs;
+	/* Of them, how many pages are active? */
+	size_t nactive;
+	/* And how many are dirty? */
+	size_t ndirty;
+};
+
+typedef struct psset_stats_s psset_stats_t;
+struct psset_stats_s {
+	/*
+	 * The second index is huge stats; nonfull_slabs[pszind][0] contains
+	 * stats for the non-huge slabs in bucket pszind, while
+	 * nonfull_slabs[pszind][1] contains stats for the huge slabs.
+	 */
+	psset_bin_stats_t nonfull_slabs[PSSET_NPSIZES][2];
+
+	/*
+	 * Full slabs don't live in any edata heap, but we still track their
+	 * stats.
+	 */
+	psset_bin_stats_t full_slabs[2];
+
+	/* Empty slabs are similar. */
+	psset_bin_stats_t empty_slabs[2];
+};
+
+typedef struct psset_s psset_t;
+struct psset_s {
+	/*
+	 * The pageslabs, quantized by the size class of the largest contiguous
+	 * free run of pages in a pageslab.
+	 */
+	hpdata_age_heap_t pageslabs[PSSET_NPSIZES];
+	/* Bitmap for which set bits correspond to non-empty heaps. */
+	fb_group_t pageslab_bitmap[FB_NGROUPS(PSSET_NPSIZES)];
+	/*
+	 * The sum of all bin stats in stats.  This lets us quickly answer
+	 * queries for the number of dirty, active, and retained pages in the
+	 * entire set.
+	 */
+	psset_bin_stats_t merged_stats;
+	psset_stats_t stats;
+	/*
+	 * Slabs with no active allocations, but which are allowed to serve new
+	 * allocations.
+	 */
+	hpdata_empty_list_t empty;
+	/*
+	 * Slabs which are available to be purged, ordered by how much we want
+	 * to purge them (with later indices indicating slabs we want to purge
+	 * more).
+	 */
+	hpdata_purge_list_t to_purge[PSSET_NPURGE_LISTS];
+	/* Bitmap for which set bits correspond to non-empty purge lists. */
+	fb_group_t purge_bitmap[FB_NGROUPS(PSSET_NPURGE_LISTS)];
+	/* Slabs which are available to be hugified. */
+	hpdata_hugify_list_t to_hugify;
+};
+
+void psset_init(psset_t *psset);
+void psset_stats_accum(psset_stats_t *dst, psset_stats_t *src);
+
+/*
+ * Begin or end updating the given pageslab's metadata.  While the pageslab is
+ * being updated, it won't be returned from psset_fit calls.
+ */
+void psset_update_begin(psset_t *psset, hpdata_t *ps);
+void psset_update_end(psset_t *psset, hpdata_t *ps);
+
+/* Analogous to the eset_fit; pick a hpdata to serve the request. */
+hpdata_t *psset_pick_alloc(psset_t *psset, size_t size);
+/* Pick one to purge. */
+hpdata_t *psset_pick_purge(psset_t *psset);
+/* Pick one to hugify. */
+hpdata_t *psset_pick_hugify(psset_t *psset);
+
+void psset_insert(psset_t *psset, hpdata_t *ps);
+void psset_remove(psset_t *psset, hpdata_t *ps);
+
+static inline size_t
+psset_npageslabs(psset_t *psset) {
+	return psset->merged_stats.npageslabs;
+}
+
+static inline size_t
+psset_nactive(psset_t *psset) {
+	return psset->merged_stats.nactive;
+}
+
+static inline size_t
+psset_ndirty(psset_t *psset) {
+	return psset->merged_stats.ndirty;
+}
+
+#endif /* JEMALLOC_INTERNAL_PSSET_H */

BeefRT/JEMalloc/include/jemalloc/internal/public_namespace.h

@@ -0,0 +1,22 @@
+#define je_aligned_alloc JEMALLOC_N(aligned_alloc)
+#define je_calloc JEMALLOC_N(calloc)
+#define je_dallocx JEMALLOC_N(dallocx)
+#define je_free JEMALLOC_N(free)
+#define je_mallctl JEMALLOC_N(mallctl)
+#define je_mallctlbymib JEMALLOC_N(mallctlbymib)
+#define je_mallctlnametomib JEMALLOC_N(mallctlnametomib)
+#define je_malloc JEMALLOC_N(malloc)
+#define je_malloc_conf JEMALLOC_N(malloc_conf)
+#define je_malloc_conf_2_conf_harder JEMALLOC_N(malloc_conf_2_conf_harder)
+#define je_malloc_message JEMALLOC_N(malloc_message)
+#define je_malloc_stats_print JEMALLOC_N(malloc_stats_print)
+#define je_malloc_usable_size JEMALLOC_N(malloc_usable_size)
+#define je_mallocx JEMALLOC_N(mallocx)
+#define je_smallocx_54eaed1d8b56b1aa528be3bdd1877e59c56fa90c JEMALLOC_N(smallocx_54eaed1d8b56b1aa528be3bdd1877e59c56fa90c)
+#define je_nallocx JEMALLOC_N(nallocx)
+#define je_posix_memalign JEMALLOC_N(posix_memalign)
+#define je_rallocx JEMALLOC_N(rallocx)
+#define je_realloc JEMALLOC_N(realloc)
+#define je_sallocx JEMALLOC_N(sallocx)
+#define je_sdallocx JEMALLOC_N(sdallocx)
+#define je_xallocx JEMALLOC_N(xallocx)

BeefRT/JEMalloc/include/jemalloc/internal/public_namespace.sh

@@ -0,0 +1,6 @@
+#!/bin/sh
+
+for nm in `cat $1` ; do
+  n=`echo ${nm} |tr ':' ' ' |awk '{print $1}'`
+  echo "#define je_${n} JEMALLOC_N(${n})"
+done

BeefRT/JEMalloc/include/jemalloc/internal/public_symbols.txt

@@ -0,0 +1,22 @@
+aligned_alloc:je_aligned_alloc
+calloc:je_calloc
+dallocx:je_dallocx
+free:je_free
+mallctl:je_mallctl
+mallctlbymib:je_mallctlbymib
+mallctlnametomib:je_mallctlnametomib
+malloc:je_malloc
+malloc_conf:je_malloc_conf
+malloc_conf_2_conf_harder:je_malloc_conf_2_conf_harder
+malloc_message:je_malloc_message
+malloc_stats_print:je_malloc_stats_print
+malloc_usable_size:je_malloc_usable_size
+mallocx:je_mallocx
+smallocx_54eaed1d8b56b1aa528be3bdd1877e59c56fa90c:je_smallocx_54eaed1d8b56b1aa528be3bdd1877e59c56fa90c
+nallocx:je_nallocx
+posix_memalign:je_posix_memalign
+rallocx:je_rallocx
+realloc:je_realloc
+sallocx:je_sallocx
+sdallocx:je_sdallocx
+xallocx:je_xallocx

BeefRT/JEMalloc/include/jemalloc/internal/public_unnamespace.h

@@ -0,0 +1,22 @@
+#undef je_aligned_alloc
+#undef je_calloc
+#undef je_dallocx
+#undef je_free
+#undef je_mallctl
+#undef je_mallctlbymib
+#undef je_mallctlnametomib
+#undef je_malloc
+#undef je_malloc_conf
+#undef je_malloc_conf_2_conf_harder
+#undef je_malloc_message
+#undef je_malloc_stats_print
+#undef je_malloc_usable_size
+#undef je_mallocx
+#undef je_smallocx_54eaed1d8b56b1aa528be3bdd1877e59c56fa90c
+#undef je_nallocx
+#undef je_posix_memalign
+#undef je_rallocx
+#undef je_realloc
+#undef je_sallocx
+#undef je_sdallocx
+#undef je_xallocx

BeefRT/JEMalloc/include/jemalloc/internal/public_unnamespace.sh

@@ -0,0 +1,6 @@
+#!/bin/sh
+
+for nm in `cat $1` ; do
+  n=`echo ${nm} |tr ':' ' ' |awk '{print $1}'`
+  echo "#undef je_${n}"
+done

BeefRT/JEMalloc/include/jemalloc/internal/ql.h

@@ -0,0 +1,197 @@
+#ifndef JEMALLOC_INTERNAL_QL_H
+#define JEMALLOC_INTERNAL_QL_H
+
+#include "jemalloc/internal/qr.h"
+
+/*
+ * A linked-list implementation.
+ *
+ * This is built on top of the ring implementation, but that can be viewed as an
+ * implementation detail (i.e. trying to advance past the tail of the list
+ * doesn't wrap around).
+ *
+ * You define a struct like so:
+ * typedef strucy my_s my_t;
+ * struct my_s {
+ *   int data;
+ *   ql_elm(my_t) my_link;
+ * };
+ *
+ * // We wobble between "list" and "head" for this type; we're now mostly
+ * // heading towards "list".
+ * typedef ql_head(my_t) my_list_t;
+ *
+ * You then pass a my_list_t * for a_head arguments, a my_t * for a_elm
+ * arguments, the token "my_link" for a_field arguments, and the token "my_t"
+ * for a_type arguments.
+ */
+
+/* List definitions. */
+#define ql_head(a_type)							\
+struct {								\
+	a_type *qlh_first;						\
+}
+
+/* Static initializer for an empty list. */
+#define ql_head_initializer(a_head) {NULL}
+
+/* The field definition. */
+#define ql_elm(a_type)	qr(a_type)
+
+/* A pointer to the first element in the list, or NULL if the list is empty. */
+#define ql_first(a_head) ((a_head)->qlh_first)
+
+/* Dynamically initializes a list. */
+#define ql_new(a_head) do {						\
+	ql_first(a_head) = NULL;					\
+} while (0)
+
+/*
+ * Sets dest to be the contents of src (overwriting any elements there), leaving
+ * src empty.
+ */
+#define ql_move(a_head_dest, a_head_src) do {				\
+	ql_first(a_head_dest) = ql_first(a_head_src);			\
+	ql_new(a_head_src);						\
+} while (0)
+
+/* True if the list is empty, otherwise false. */
+#define ql_empty(a_head) (ql_first(a_head) == NULL)
+
+/*
+ * Initializes a ql_elm.  Must be called even if the field is about to be
+ * overwritten.
+ */
+#define ql_elm_new(a_elm, a_field) qr_new((a_elm), a_field)
+
+/*
+ * Obtains the last item in the list.
+ */
+#define ql_last(a_head, a_field)					\
+	(ql_empty(a_head) ? NULL : qr_prev(ql_first(a_head), a_field))
+
+/*
+ * Gets a pointer to the next/prev element in the list.  Trying to advance past
+ * the end or retreat before the beginning of the list returns NULL.
+ */
+#define ql_next(a_head, a_elm, a_field)					\
+	((ql_last(a_head, a_field) != (a_elm))				\
+	    ? qr_next((a_elm), a_field)	: NULL)
+#define ql_prev(a_head, a_elm, a_field)					\
+	((ql_first(a_head) != (a_elm)) ? qr_prev((a_elm), a_field)	\
+				       : NULL)
+
+/* Inserts a_elm before a_qlelm in the list. */
+#define ql_before_insert(a_head, a_qlelm, a_elm, a_field) do {		\
+	qr_before_insert((a_qlelm), (a_elm), a_field);			\
+	if (ql_first(a_head) == (a_qlelm)) {				\
+		ql_first(a_head) = (a_elm);				\
+	}								\
+} while (0)
+
+/* Inserts a_elm after a_qlelm in the list. */
+#define ql_after_insert(a_qlelm, a_elm, a_field)			\
+	qr_after_insert((a_qlelm), (a_elm), a_field)
+
+/* Inserts a_elm as the first item in the list. */
+#define ql_head_insert(a_head, a_elm, a_field) do {			\
+	if (!ql_empty(a_head)) {					\
+		qr_before_insert(ql_first(a_head), (a_elm), a_field);	\
+	}								\
+	ql_first(a_head) = (a_elm);					\
+} while (0)
+
+/* Inserts a_elm as the last item in the list. */
+#define ql_tail_insert(a_head, a_elm, a_field) do {			\
+	if (!ql_empty(a_head)) {					\
+		qr_before_insert(ql_first(a_head), (a_elm), a_field);	\
+	}								\
+	ql_first(a_head) = qr_next((a_elm), a_field);			\
+} while (0)
+
+/*
+ * Given lists a = [a_1, ..., a_n] and [b_1, ..., b_n], results in:
+ * a = [a1, ..., a_n, b_1, ..., b_n] and b = [].
+ */
+#define ql_concat(a_head_a, a_head_b, a_field) do {			\
+	if (ql_empty(a_head_a)) {					\
+		ql_move(a_head_a, a_head_b);				\
+	} else if (!ql_empty(a_head_b)) {				\
+		qr_meld(ql_first(a_head_a), ql_first(a_head_b),		\
+		    a_field);						\
+		ql_new(a_head_b);					\
+	}								\
+} while (0)
+
+/* Removes a_elm from the list. */
+#define ql_remove(a_head, a_elm, a_field) do {				\
+	if (ql_first(a_head) == (a_elm)) {				\
+		ql_first(a_head) = qr_next(ql_first(a_head), a_field);	\
+	}								\
+	if (ql_first(a_head) != (a_elm)) {				\
+		qr_remove((a_elm), a_field);				\
+	} else {							\
+		ql_new(a_head);						\
+	}								\
+} while (0)
+
+/* Removes the first item in the list. */
+#define ql_head_remove(a_head, a_type, a_field) do {			\
+	a_type *t = ql_first(a_head);					\
+	ql_remove((a_head), t, a_field);				\
+} while (0)
+
+/* Removes the last item in the list. */
+#define ql_tail_remove(a_head, a_type, a_field) do {			\
+	a_type *t = ql_last(a_head, a_field);				\
+	ql_remove((a_head), t, a_field);				\
+} while (0)
+
+/*
+ * Given a = [a_1, a_2, ..., a_n-1, a_n, a_n+1, ...],
+ * ql_split(a, a_n, b, some_field) results in
+ *   a = [a_1, a_2, ..., a_n-1]
+ * and replaces b's contents with:
+ *   b = [a_n, a_n+1, ...]
+ */
+#define ql_split(a_head_a, a_elm, a_head_b, a_field) do {		\
+	if (ql_first(a_head_a) == (a_elm)) {				\
+		ql_move(a_head_b, a_head_a);				\
+	} else {							\
+		qr_split(ql_first(a_head_a), (a_elm), a_field);		\
+		ql_first(a_head_b) = (a_elm);				\
+	}								\
+} while (0)
+
+/*
+ * An optimized version of:
+ *	a_type *t = ql_first(a_head);
+ *	ql_remove((a_head), t, a_field);
+ *	ql_tail_insert((a_head), t, a_field);
+ */
+#define ql_rotate(a_head, a_field) do {					\
+	ql_first(a_head) = qr_next(ql_first(a_head), a_field);		\
+} while (0)
+
+/*
+ * Helper macro to iterate over each element in a list in order, starting from
+ * the head (or in reverse order, starting from the tail).  The usage is
+ * (assuming my_t and my_list_t defined as above).
+ *
+ * int sum(my_list_t *list) {
+ *   int sum = 0;
+ *   my_t *iter;
+ *   ql_foreach(iter, list, link) {
+ *     sum += iter->data;
+ *   }
+ *   return sum;
+ * }
+ */
+
+#define ql_foreach(a_var, a_head, a_field)				\
+	qr_foreach((a_var), ql_first(a_head), a_field)
+
+#define ql_reverse_foreach(a_var, a_head, a_field)			\
+	qr_reverse_foreach((a_var), ql_first(a_head), a_field)
+
+#endif /* JEMALLOC_INTERNAL_QL_H */

BeefRT/JEMalloc/include/jemalloc/internal/qr.h

@@ -0,0 +1,140 @@
+#ifndef JEMALLOC_INTERNAL_QR_H
+#define JEMALLOC_INTERNAL_QR_H
+
+/*
+ * A ring implementation based on an embedded circular doubly-linked list.
+ *
+ * You define your struct like so:
+ *
+ * typedef struct my_s my_t;
+ * struct my_s {
+ *   int data;
+ *   qr(my_t) my_link;
+ * };
+ *
+ * And then pass a my_t * into macros for a_qr arguments, and the token
+ * "my_link" into a_field fields.
+ */
+
+/* Ring definitions. */
+#define qr(a_type)							\
+struct {								\
+	a_type	*qre_next;						\
+	a_type	*qre_prev;						\
+}
+
+/*
+ * Initialize a qr link.  Every link must be initialized before being used, even
+ * if that initialization is going to be immediately overwritten (say, by being
+ * passed into an insertion macro).
+ */
+#define qr_new(a_qr, a_field) do {					\
+	(a_qr)->a_field.qre_next = (a_qr);				\
+	(a_qr)->a_field.qre_prev = (a_qr);				\
+} while (0)
+
+/*
+ * Go forwards or backwards in the ring.  Note that (the ring being circular), this
+ * always succeeds -- you just keep looping around and around the ring if you
+ * chase pointers without end.
+ */
+#define qr_next(a_qr, a_field) ((a_qr)->a_field.qre_next)
+#define qr_prev(a_qr, a_field) ((a_qr)->a_field.qre_prev)
+
+/*
+ * Given two rings:
+ *    a -> a_1 -> ... -> a_n --
+ *    ^                       |
+ *    |------------------------
+ *
+ *    b -> b_1 -> ... -> b_n --
+ *    ^                       |
+ *    |------------------------
+ *
+ * Results in the ring:
+ *   a -> a_1 -> ... -> a_n -> b -> b_1 -> ... -> b_n --
+ *   ^                                                 |
+ *   |-------------------------------------------------|
+ *
+ * a_qr_a can directly be a qr_next() macro, but a_qr_b cannot.
+ */
+#define qr_meld(a_qr_a, a_qr_b, a_field) do {				\
+	(a_qr_b)->a_field.qre_prev->a_field.qre_next =			\
+	    (a_qr_a)->a_field.qre_prev;					\
+	(a_qr_a)->a_field.qre_prev = (a_qr_b)->a_field.qre_prev;	\
+	(a_qr_b)->a_field.qre_prev =					\
+	    (a_qr_b)->a_field.qre_prev->a_field.qre_next;		\
+	(a_qr_a)->a_field.qre_prev->a_field.qre_next = (a_qr_a);	\
+	(a_qr_b)->a_field.qre_prev->a_field.qre_next = (a_qr_b);	\
+} while (0)
+
+/*
+ * Logically, this is just a meld.  The intent, though, is that a_qrelm is a
+ * single-element ring, so that "before" has a more obvious interpretation than
+ * meld.
+ */
+#define qr_before_insert(a_qrelm, a_qr, a_field)			\
+	qr_meld((a_qrelm), (a_qr), a_field)
+
+/* Ditto, but inserting after rather than before. */
+#define qr_after_insert(a_qrelm, a_qr, a_field)				\
+	qr_before_insert(qr_next(a_qrelm, a_field), (a_qr), a_field)
+
+/*
+ * Inverts meld; given the ring:
+ *   a -> a_1 -> ... -> a_n -> b -> b_1 -> ... -> b_n --
+ *   ^                                                 |
+ *   |-------------------------------------------------|
+ *
+ * Results in two rings:
+ *    a -> a_1 -> ... -> a_n --
+ *    ^                       |
+ *    |------------------------
+ *
+ *    b -> b_1 -> ... -> b_n --
+ *    ^                       |
+ *    |------------------------
+ *
+ * qr_meld() and qr_split() are functionally equivalent, so there's no need to
+ * have two copies of the code.
+ */
+#define qr_split(a_qr_a, a_qr_b, a_field)				\
+	qr_meld((a_qr_a), (a_qr_b), a_field)
+
+/*
+ * Splits off a_qr from the rest of its ring, so that it becomes a
+ * single-element ring.
+ */
+#define qr_remove(a_qr, a_field)					\
+	qr_split(qr_next(a_qr, a_field), (a_qr), a_field)
+
+/*
+ * Helper macro to iterate over each element in a ring exactly once, starting
+ * with a_qr.  The usage is (assuming my_t defined as above):
+ *
+ * int sum(my_t *item) {
+ *   int sum = 0;
+ *   my_t *iter;
+ *   qr_foreach(iter, item, link) {
+ *     sum += iter->data;
+ *   }
+ *   return sum;
+ * }
+ */
+#define qr_foreach(var, a_qr, a_field)					\
+	for ((var) = (a_qr);						\
+	    (var) != NULL;						\
+	    (var) = (((var)->a_field.qre_next != (a_qr))		\
+	    ? (var)->a_field.qre_next : NULL))
+
+/*
+ * The same (and with the same usage) as qr_foreach, but in the opposite order,
+ * ending with a_qr.
+ */
+#define qr_reverse_foreach(var, a_qr, a_field)				\
+	for ((var) = ((a_qr) != NULL) ? qr_prev(a_qr, a_field) : NULL;	\
+	    (var) != NULL;						\
+	    (var) = (((var) != (a_qr))					\
+	    ? (var)->a_field.qre_prev : NULL))
+
+#endif /* JEMALLOC_INTERNAL_QR_H */