Unverified Commit c4b4b6c0 authored by Oran Agra's avatar Oran Agra Committed by GitHub
Browse files

Merge pull request #9623 from yoav-steinberg/upgrade_jemalloc_5.2.1

Upgraded to jemalloc 5.2.1 from 5.1.0.
Cherry picked all relevant fixes (by diffing our 5.1.0 to upstream 5.10 and finding relevant commits).
Details of what was done:

[cherry-picked] fd7d51c3 2021-05-03 Resolve nonsense static analysis warnings (Oran Agra)
[cherry-picked] 448c435b 2020-09-29 Fix compilation warnings in Lua and jemalloc dependencies (#7785) (YoongHM)
[skipped - already in upstream] 9216b96b 2020-09-21 Fix compilation warning in jemalloc's malloc_vsnprintf (#7789) (YoongHM)
[cherry-picked] 88d71f47 2020-05-20 fix a rare active defrag edge case bug leading to stagnation (Oran Agra)
[skipped - already in upstream] 2fec7d9c 2019-05-30 Jemalloc: Avoid blocking on background thread lock for stats.
[cherry-picked] 920158ec 2018-07-11 Active defrag fixes for 32bit builds (again) (Oran Agra)
[cherry-picked] e8099cab 2018-06-26 add defrag hint support into jemalloc 5 (Oran Agra)
[re-done] 4e729fcd 2018-05-24 Generate configure for Jemalloc. (antirez)

Additionally had to do this:
7727cc2 2021-10-10 Fix defrag to support sharded bins in arena (added in v5.2.1) (Yoav Steinberg)

When reviewing please look at all except the first commit which is just replacing 5.1.0 with 5.2.1 sources.
Also I think we should merge this without squashing to preserve the changes we did to to jemalloc.
parents 276b460e 85737e67
#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 */
#ifndef JEMALLOC_INTERNAL_HOOKS_H
#define JEMALLOC_INTERNAL_HOOKS_H
extern JEMALLOC_EXPORT void (*hooks_arena_new_hook)();
extern JEMALLOC_EXPORT void (*hooks_libc_hook)();
#define JEMALLOC_HOOK(fn, hook) ((void)(hook != NULL && (hook(), 0)), fn)
#define open JEMALLOC_HOOK(open, hooks_libc_hook)
#define read JEMALLOC_HOOK(read, hooks_libc_hook)
#define write JEMALLOC_HOOK(write, hooks_libc_hook)
#define readlink JEMALLOC_HOOK(readlink, hooks_libc_hook)
#define close JEMALLOC_HOOK(close, hooks_libc_hook)
#define creat JEMALLOC_HOOK(creat, hooks_libc_hook)
#define secure_getenv JEMALLOC_HOOK(secure_getenv, hooks_libc_hook)
/* Note that this is undef'd and re-define'd in src/prof.c. */
#define _Unwind_Backtrace JEMALLOC_HOOK(_Unwind_Backtrace, hooks_libc_hook)
#endif /* JEMALLOC_INTERNAL_HOOKS_H */
......@@ -31,6 +31,9 @@
# include <sys/uio.h>
# endif
# include <pthread.h>
# ifdef __FreeBSD__
# include <pthread_np.h>
# endif
# include <signal.h>
# ifdef JEMALLOC_OS_UNFAIR_LOCK
# include <os/lock.h>
......
......@@ -48,25 +48,13 @@
/* 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
/*
* Defined if __sync_add_and_fetch(uint32_t *, uint32_t) and
* __sync_sub_and_fetch(uint32_t *, uint32_t) are available, despite
* __GCC_HAVE_SYNC_COMPARE_AND_SWAP_4 not being defined (which means the
* functions are defined in libgcc instead of being inlines).
*/
#undef JE_FORCE_SYNC_COMPARE_AND_SWAP_4
/*
* Defined if __sync_add_and_fetch(uint64_t *, uint64_t) and
* __sync_sub_and_fetch(uint64_t *, uint64_t) are available, despite
* __GCC_HAVE_SYNC_COMPARE_AND_SWAP_8 not being defined (which means the
* functions are defined in libgcc instead of being inlines).
*/
#undef JE_FORCE_SYNC_COMPARE_AND_SWAP_8
/* and the 8-bit variant support. */
#undef JEMALLOC_GCC_U8_SYNC_ATOMICS
/*
* Defined if __builtin_clz() and __builtin_clzl() are available.
......@@ -78,12 +66,6 @@
*/
#undef JEMALLOC_OS_UNFAIR_LOCK
/*
* Defined if OSSpin*() functions are available, as provided by Darwin, and
* documented in the spinlock(3) manual page.
*/
#undef JEMALLOC_OSSPIN
/* Defined if syscall(2) is usable. */
#undef JEMALLOC_USE_SYSCALL
......@@ -153,6 +135,9 @@
/* 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
......@@ -233,6 +218,12 @@
#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.
......@@ -245,6 +236,12 @@
*/
#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.
*/
......@@ -363,4 +360,7 @@
*/
#undef JEMALLOC_STRERROR_R_RETURNS_CHAR_WITH_GNU_SOURCE
/* Performs additional safety checks when defined. */
#undef JEMALLOC_OPT_SAFETY_CHECKS
#endif /* JEMALLOC_INTERNAL_DEFS_H_ */
......@@ -2,7 +2,6 @@
#define JEMALLOC_INTERNAL_EXTERNS_H
#include "jemalloc/internal/atomic.h"
#include "jemalloc/internal/size_classes.h"
#include "jemalloc/internal/tsd_types.h"
/* TSD checks this to set thread local slow state accordingly. */
......@@ -11,6 +10,7 @@ extern bool malloc_slow;
/* Run-time options. */
extern bool opt_abort;
extern bool opt_abort_conf;
extern bool opt_confirm_conf;
extern const char *opt_junk;
extern bool opt_junk_alloc;
extern bool opt_junk_free;
......@@ -25,6 +25,9 @@ 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.
......@@ -49,5 +52,6 @@ void jemalloc_prefork(void);
void jemalloc_postfork_parent(void);
void jemalloc_postfork_child(void);
bool malloc_initialized(void);
void je_sdallocx_noflags(void *ptr, size_t size);
#endif /* JEMALLOC_INTERNAL_EXTERNS_H */
......@@ -4,13 +4,15 @@
#include "jemalloc/internal/atomic.h"
#include "jemalloc/internal/bit_util.h"
#include "jemalloc/internal/jemalloc_internal_types.h"
#include "jemalloc/internal/size_classes.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(JEMALLOC_HAVE_SCHED_GETCPU)
#if defined(_WIN32)
return GetCurrentProcessorNumber();
#elif defined(JEMALLOC_HAVE_SCHED_GETCPU)
return (malloc_cpuid_t)sched_getcpu();
#else
not_reached();
......@@ -108,14 +110,14 @@ decay_ticker_get(tsd_t *tsd, unsigned ind) {
JEMALLOC_ALWAYS_INLINE cache_bin_t *
tcache_small_bin_get(tcache_t *tcache, szind_t binind) {
assert(binind < NBINS);
assert(binind < SC_NBINS);
return &tcache->bins_small[binind];
}
JEMALLOC_ALWAYS_INLINE cache_bin_t *
tcache_large_bin_get(tcache_t *tcache, szind_t binind) {
assert(binind >= NBINS &&binind < nhbins);
return &tcache->bins_large[binind - NBINS];
assert(binind >= SC_NBINS &&binind < nhbins);
return &tcache->bins_large[binind - SC_NBINS];
}
JEMALLOC_ALWAYS_INLINE bool
......@@ -156,7 +158,7 @@ pre_reentrancy(tsd_t *tsd, arena_t *arena) {
if (fast) {
/* Prepare slow path for reentrancy. */
tsd_slow_update(tsd);
assert(tsd->state == tsd_state_nominal_slow);
assert(tsd_state_get(tsd) == tsd_state_nominal_slow);
}
}
......
......@@ -71,7 +71,8 @@ arena_ichoose(tsd_t *tsd, arena_t *arena) {
static inline bool
arena_is_auto(arena_t *arena) {
assert(narenas_auto > 0);
return (arena_ind_get(arena) < narenas_auto);
return (arena_ind_get(arena) < manual_arena_base);
}
JEMALLOC_ALWAYS_INLINE extent_t *
......
#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/sz.h"
#include "jemalloc/internal/witness.h"
......@@ -42,7 +43,6 @@ 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(size != 0);
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) {
......@@ -133,31 +133,20 @@ isdalloct(tsdn_t *tsdn, void *ptr, size_t size, tcache_t *tcache,
JEMALLOC_ALWAYS_INLINE void *
iralloct_realign(tsdn_t *tsdn, void *ptr, size_t oldsize, size_t size,
size_t extra, size_t alignment, bool zero, tcache_t *tcache,
arena_t *arena) {
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 + extra, alignment);
if (unlikely(usize == 0 || usize > LARGE_MAXCLASS)) {
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) {
if (extra == 0) {
return NULL;
}
/* Try again, without extra this time. */
usize = sz_sa2u(size, alignment);
if (unlikely(usize == 0 || usize > LARGE_MAXCLASS)) {
return NULL;
}
p = ipalloct(tsdn, usize, alignment, zero, tcache, arena);
if (p == NULL) {
return NULL;
}
return NULL;
}
/*
* Copy at most size bytes (not size+extra), since the caller has no
......@@ -165,13 +154,26 @@ iralloct_realign(tsdn_t *tsdn, void *ptr, size_t oldsize, size_t size,
*/
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) {
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),
......@@ -183,24 +185,24 @@ iralloct(tsdn_t *tsdn, void *ptr, size_t oldsize, size_t size, size_t alignment,
* Existing object alignment is inadequate; allocate new space
* and copy.
*/
return iralloct_realign(tsdn, ptr, oldsize, size, 0, alignment,
zero, tcache, arena);
return iralloct_realign(tsdn, ptr, oldsize, size, alignment,
zero, tcache, arena, hook_args);
}
return arena_ralloc(tsdn, arena, ptr, oldsize, size, alignment, zero,
tcache);
tcache, hook_args);
}
JEMALLOC_ALWAYS_INLINE void *
iralloc(tsd_t *tsd, void *ptr, size_t oldsize, size_t size, size_t alignment,
bool zero) {
bool zero, hook_ralloc_args_t *hook_args) {
return iralloct(tsd_tsdn(tsd), ptr, oldsize, size, alignment, zero,
tcache_get(tsd), NULL);
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 alignment, bool zero, size_t *newsize) {
assert(ptr != NULL);
assert(size != 0);
witness_assert_depth_to_rank(tsdn_witness_tsdp_get(tsdn),
......@@ -209,10 +211,12 @@ ixalloc(tsdn_t *tsdn, void *ptr, size_t oldsize, size_t size, size_t extra,
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);
return arena_ralloc_no_move(tsdn, ptr, oldsize, size, extra, zero,
newsize);
}
JEMALLOC_ALWAYS_INLINE int
......@@ -228,7 +232,8 @@ iget_defrag_hint(tsdn_t *tsdn, void* ptr) {
extent_t *slab = iealloc(tsdn, ptr);
arena_t *arena = extent_arena_get(slab);
szind_t binind = extent_szind_get(slab);
bin_t *bin = &arena->bins[binind];
unsigned binshard = extent_binshard_get(slab);
bin_t *bin = &arena->bins[binind].bin_shards[binshard];
malloc_mutex_lock(tsdn, &bin->lock);
/* don't bother moving allocations from the slab currently used for new allocations */
if (slab != bin->slabcur) {
......
......@@ -30,7 +30,7 @@
# define restrict
#endif
/* Various function pointers are statick and immutable except during testing. */
/* Various function pointers are static and immutable except during testing. */
#ifdef JEMALLOC_JET
# define JET_MUTABLE
#else
......@@ -40,4 +40,75 @@
#define JEMALLOC_VA_ARGS_HEAD(head, ...) head
#define JEMALLOC_VA_ARGS_TAIL(head, ...) __VA_ARGS__
#if (defined(__GNUC__) || defined(__GNUG__)) && !defined(__clang__) \
&& defined(JEMALLOC_HAVE_ATTR) && (__GNUC__ >= 7)
#define JEMALLOC_FALLTHROUGH JEMALLOC_ATTR(fallthrough);
#else
#define JEMALLOC_FALLTHROUGH /* falls through */
#endif
/* 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 */
#ifndef JEMALLOC_INTERNAL_TYPES_H
#define JEMALLOC_INTERNAL_TYPES_H
#include "jemalloc/internal/quantum.h"
/* Page size index type. */
typedef unsigned pszind_t;
......@@ -50,79 +52,6 @@ typedef int malloc_cpuid_t;
/* Smallest size class to support. */
#define TINY_MIN (1U << LG_TINY_MIN)
/*
* Minimum allocation alignment is 2^LG_QUANTUM bytes (ignoring tiny size
* classes).
*/
#ifndef LG_QUANTUM
# if (defined(__i386__) || defined(_M_IX86))
# define LG_QUANTUM 4
# endif
# ifdef __ia64__
# define LG_QUANTUM 4
# endif
# ifdef __alpha__
# define LG_QUANTUM 4
# endif
# if (defined(__sparc64__) || defined(__sparcv9) || defined(__sparc_v9__))
# define LG_QUANTUM 4
# endif
# if (defined(__amd64__) || defined(__x86_64__) || defined(_M_X64))
# define LG_QUANTUM 4
# endif
# ifdef __arm__
# define LG_QUANTUM 3
# endif
# ifdef __aarch64__
# define LG_QUANTUM 4
# endif
# ifdef __hppa__
# define LG_QUANTUM 4
# endif
# ifdef __m68k__
# define LG_QUANTUM 3
# endif
# ifdef __mips__
# define LG_QUANTUM 3
# endif
# ifdef __nios2__
# define LG_QUANTUM 3
# endif
# ifdef __or1k__
# define LG_QUANTUM 3
# endif
# ifdef __powerpc__
# define LG_QUANTUM 4
# endif
# if defined(__riscv) || defined(__riscv__)
# define LG_QUANTUM 4
# endif
# ifdef __s390__
# define LG_QUANTUM 4
# endif
# if (defined (__SH3E__) || defined(__SH4_SINGLE__) || defined(__SH4__) || \
defined(__SH4_SINGLE_ONLY__))
# define LG_QUANTUM 4
# endif
# ifdef __tile__
# define LG_QUANTUM 4
# endif
# ifdef __le32__
# define LG_QUANTUM 4
# endif
# ifndef LG_QUANTUM
# error "Unknown minimum alignment for architecture; specify via "
"--with-lg-quantum"
# endif
#endif
#define QUANTUM ((size_t)(1U << LG_QUANTUM))
#define QUANTUM_MASK (QUANTUM - 1)
/* Return the smallest quantum multiple that is >= a. */
#define QUANTUM_CEILING(a) \
(((a) + QUANTUM_MASK) & ~QUANTUM_MASK)
#define LONG ((size_t)(1U << LG_SIZEOF_LONG))
#define LONG_MASK (LONG - 1)
......
......@@ -21,7 +21,7 @@
# include "../jemalloc@install_suffix@.h"
#endif
#if (defined(JEMALLOC_OSATOMIC) || defined(JEMALLOC_OSSPIN))
#if defined(JEMALLOC_OSATOMIC)
#include <libkern/OSAtomic.h>
#endif
......@@ -45,7 +45,7 @@
# include "jemalloc/internal/private_namespace_jet.h"
# endif
#endif
#include "jemalloc/internal/hooks.h"
#include "jemalloc/internal/test_hooks.h"
#ifdef JEMALLOC_DEFINE_MADVISE_FREE
# define JEMALLOC_MADV_FREE 8
......@@ -161,7 +161,26 @@ static const bool config_log =
false
#endif
;
#ifdef JEMALLOC_HAVE_SCHED_GETCPU
/*
* 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
;
#if defined(_WIN32) || defined(JEMALLOC_HAVE_SCHED_GETCPU)
/* Currently percpu_arena depends on sched_getcpu. */
#define JEMALLOC_PERCPU_ARENA
#endif
......
#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, extent_t *extent, size_t usize_min,
size_t usize_max, bool zero);
void *large_ralloc(tsdn_t *tsdn, arena_t *arena, extent_t *extent, size_t usize,
size_t alignment, bool zero, tcache_t *tcache);
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);
typedef void (large_dalloc_junk_t)(void *, size_t);
extern large_dalloc_junk_t *JET_MUTABLE large_dalloc_junk;
......@@ -23,4 +26,7 @@ prof_tctx_t *large_prof_tctx_get(tsdn_t *tsdn, const extent_t *extent);
void large_prof_tctx_set(tsdn_t *tsdn, extent_t *extent, prof_tctx_t *tctx);
void large_prof_tctx_reset(tsdn_t *tsdn, extent_t *extent);
nstime_t large_prof_alloc_time_get(const extent_t *extent);
void large_prof_alloc_time_set(extent_t *extent, nstime_t time);
#endif /* JEMALLOC_INTERNAL_LARGE_EXTERNS_H */
......@@ -54,7 +54,7 @@ size_t malloc_vsnprintf(char *str, size_t size, const char *format,
size_t malloc_snprintf(char *str, size_t size, const char *format, ...)
JEMALLOC_FORMAT_PRINTF(3, 4);
/*
* The caller can set write_cb and cbopaque to null to choose to print with the
* The caller can set write_cb to null to choose to print with the
* je_malloc_message hook.
*/
void malloc_vcprintf(void (*write_cb)(void *, const char *), void *cbopaque,
......
......@@ -37,14 +37,17 @@ struct malloc_mutex_s {
# endif
#elif (defined(JEMALLOC_OS_UNFAIR_LOCK))
os_unfair_lock lock;
#elif (defined(JEMALLOC_OSSPIN))
OSSpinLock 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
......@@ -84,10 +87,6 @@ struct malloc_mutex_s {
# 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))
#elif (defined(JEMALLOC_OSSPIN))
# define MALLOC_MUTEX_LOCK(m) OSSpinLockLock(&(m)->lock)
# define MALLOC_MUTEX_UNLOCK(m) OSSpinLockUnlock(&(m)->lock)
# define MALLOC_MUTEX_TRYLOCK(m) (!OSSpinLockTry(&(m)->lock))
#else
# define MALLOC_MUTEX_LOCK(m) pthread_mutex_lock(&(m)->lock)
# define MALLOC_MUTEX_UNLOCK(m) pthread_mutex_unlock(&(m)->lock)
......@@ -101,22 +100,37 @@ struct malloc_mutex_s {
#ifdef _WIN32
# define MALLOC_MUTEX_INITIALIZER
#elif (defined(JEMALLOC_OS_UNFAIR_LOCK))
# define MALLOC_MUTEX_INITIALIZER \
{{{LOCK_PROF_DATA_INITIALIZER, OS_UNFAIR_LOCK_INIT}}, \
WITNESS_INITIALIZER("mutex", WITNESS_RANK_OMIT)}
#elif (defined(JEMALLOC_OSSPIN))
# define MALLOC_MUTEX_INITIALIZER \
{{{LOCK_PROF_DATA_INITIALIZER, 0}}, \
# 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))
# define MALLOC_MUTEX_INITIALIZER \
{{{LOCK_PROF_DATA_INITIALIZER, PTHREAD_MUTEX_INITIALIZER, NULL}}, \
WITNESS_INITIALIZER("mutex", WITNESS_RANK_OMIT)}
# 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}}, \
WITNESS_INITIALIZER("mutex", WITNESS_RANK_OMIT)}
{{{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
......@@ -139,6 +153,7 @@ 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
......@@ -164,6 +179,7 @@ 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);
......@@ -203,6 +219,7 @@ malloc_mutex_lock(tsdn_t *tsdn, malloc_mutex_t *mutex) {
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);
}
......@@ -211,6 +228,7 @@ malloc_mutex_lock(tsdn_t *tsdn, malloc_mutex_t *mutex) {
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);
......@@ -245,4 +263,26 @@ malloc_mutex_prof_read(tsdn_t *tsdn, mutex_prof_data_t *data,
atomic_store_u32(&data->n_waiting_thds, 0, ATOMIC_RELAXED);
}
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;
}
#endif /* JEMALLOC_INTERNAL_MUTEX_H */
......@@ -35,22 +35,31 @@ typedef enum {
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") \
OP(num_wait, uint64_t, "n_waiting") \
OP(num_spin_acq, uint64_t, "n_spin_acq") \
OP(num_owner_switch, uint64_t, "n_owner_switch") \
OP(total_wait_time, uint64_t, "total_wait_ns") \
OP(max_wait_time, uint64_t, "max_wait_ns")
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")
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) mutex_counter_##counter,
#define OP(counter, type, human, derived, base_counter) mutex_counter_##counter,
#define COUNTER_ENUM(counter_list, t) \
typedef enum { \
......
......@@ -14,6 +14,7 @@ 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_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
......@@ -45,7 +46,8 @@ extern size_t lg_prof_sample;
void prof_alloc_rollback(tsd_t *tsd, prof_tctx_t *tctx, bool updated);
void prof_malloc_sample_object(tsdn_t *tsdn, const void *ptr, size_t usize,
prof_tctx_t *tctx);
void prof_free_sampled_object(tsd_t *tsd, size_t usize, prof_tctx_t *tctx);
void prof_free_sampled_object(tsd_t *tsd, const void *ptr, size_t usize,
prof_tctx_t *tctx);
void bt_init(prof_bt_t *bt, void **vec);
void prof_backtrace(prof_bt_t *bt);
prof_tctx_t *prof_lookup(tsd_t *tsd, prof_bt_t *bt);
......@@ -89,4 +91,15 @@ void prof_postfork_parent(tsdn_t *tsdn);
void prof_postfork_child(tsdn_t *tsdn);
void prof_sample_threshold_update(prof_tdata_t *tdata);
bool prof_log_start(tsdn_t *tsdn, const char *filename);
bool prof_log_stop(tsdn_t *tsdn);
#ifdef JEMALLOC_JET
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);
#endif
#endif /* JEMALLOC_INTERNAL_PROF_EXTERNS_H */
......@@ -4,7 +4,8 @@
#include "jemalloc/internal/mutex.h"
static inline bool
prof_accum_add(tsdn_t *tsdn, prof_accum_t *prof_accum, uint64_t accumbytes) {
prof_accum_add(tsdn_t *tsdn, prof_accum_t *prof_accum,
uint64_t accumbytes) {
cassert(config_prof);
bool overflow;
......@@ -42,7 +43,8 @@ prof_accum_add(tsdn_t *tsdn, prof_accum_t *prof_accum, uint64_t accumbytes) {
}
static inline void
prof_accum_cancel(tsdn_t *tsdn, prof_accum_t *prof_accum, size_t usize) {
prof_accum_cancel(tsdn_t *tsdn, prof_accum_t *prof_accum,
size_t usize) {
cassert(config_prof);
/*
......@@ -55,15 +57,15 @@ prof_accum_cancel(tsdn_t *tsdn, prof_accum_t *prof_accum, size_t usize) {
#ifdef JEMALLOC_ATOMIC_U64
a0 = atomic_load_u64(&prof_accum->accumbytes, ATOMIC_RELAXED);
do {
a1 = (a0 >= LARGE_MINCLASS - usize) ? a0 - (LARGE_MINCLASS -
usize) : 0;
a1 = (a0 >= SC_LARGE_MINCLASS - usize)
? a0 - (SC_LARGE_MINCLASS - usize) : 0;
} while (!atomic_compare_exchange_weak_u64(&prof_accum->accumbytes, &a0,
a1, ATOMIC_RELAXED, ATOMIC_RELAXED));
#else
malloc_mutex_lock(tsdn, &prof_accum->mtx);
a0 = prof_accum->accumbytes;
a1 = (a0 >= LARGE_MINCLASS - usize) ? a0 - (LARGE_MINCLASS - usize) :
0;
a1 = (a0 >= SC_LARGE_MINCLASS - usize)
? a0 - (SC_LARGE_MINCLASS - usize) : 0;
prof_accum->accumbytes = a1;
malloc_mutex_unlock(tsdn, &prof_accum->mtx);
#endif
......
#ifndef JEMALLOC_INTERNAL_PROF_INLINES_B_H
#define JEMALLOC_INTERNAL_PROF_INLINES_B_H
#include "jemalloc/internal/safety_check.h"
#include "jemalloc/internal/sz.h"
JEMALLOC_ALWAYS_INLINE bool
......@@ -61,13 +62,54 @@ prof_tctx_reset(tsdn_t *tsdn, const void *ptr, prof_tctx_t *tctx) {
arena_prof_tctx_reset(tsdn, ptr, tctx);
}
JEMALLOC_ALWAYS_INLINE nstime_t
prof_alloc_time_get(tsdn_t *tsdn, const void *ptr, alloc_ctx_t *alloc_ctx) {
cassert(config_prof);
assert(ptr != NULL);
return arena_prof_alloc_time_get(tsdn, ptr, alloc_ctx);
}
JEMALLOC_ALWAYS_INLINE void
prof_alloc_time_set(tsdn_t *tsdn, const void *ptr, alloc_ctx_t *alloc_ctx,
nstime_t t) {
cassert(config_prof);
assert(ptr != NULL);
arena_prof_alloc_time_set(tsdn, ptr, alloc_ctx, t);
}
JEMALLOC_ALWAYS_INLINE bool
prof_sample_check(tsd_t *tsd, size_t usize, bool update) {
ssize_t check = update ? 0 : usize;
int64_t bytes_until_sample = tsd_bytes_until_sample_get(tsd);
if (update) {
bytes_until_sample -= usize;
if (tsd_nominal(tsd)) {
tsd_bytes_until_sample_set(tsd, bytes_until_sample);
}
}
if (likely(bytes_until_sample >= check)) {
return true;
}
return false;
}
JEMALLOC_ALWAYS_INLINE bool
prof_sample_accum_update(tsd_t *tsd, size_t usize, bool update,
prof_tdata_t **tdata_out) {
prof_tdata_t **tdata_out) {
prof_tdata_t *tdata;
cassert(config_prof);
/* Fastpath: no need to load tdata */
if (likely(prof_sample_check(tsd, usize, update))) {
return true;
}
bool booted = tsd_prof_tdata_get(tsd);
tdata = prof_tdata_get(tsd, true);
if (unlikely((uintptr_t)tdata <= (uintptr_t)PROF_TDATA_STATE_MAX)) {
tdata = NULL;
......@@ -81,21 +123,23 @@ prof_sample_accum_update(tsd_t *tsd, size_t usize, bool update,
return true;
}
if (likely(tdata->bytes_until_sample >= usize)) {
if (update) {
tdata->bytes_until_sample -= usize;
}
/*
* If this was the first creation of tdata, then
* prof_tdata_get() reset bytes_until_sample, so decrement and
* check it again
*/
if (!booted && prof_sample_check(tsd, usize, update)) {
return true;
} else {
if (tsd_reentrancy_level_get(tsd) > 0) {
return true;
}
/* Compute new sample threshold. */
if (update) {
prof_sample_threshold_update(tdata);
}
return !tdata->active;
}
if (tsd_reentrancy_level_get(tsd) > 0) {
return true;
}
/* Compute new sample threshold. */
if (update) {
prof_sample_threshold_update(tdata);
}
return !tdata->active;
}
JEMALLOC_ALWAYS_INLINE prof_tctx_t *
......@@ -187,7 +231,7 @@ prof_realloc(tsd_t *tsd, const void *ptr, size_t usize, prof_tctx_t *tctx,
* counters.
*/
if (unlikely(old_sampled)) {
prof_free_sampled_object(tsd, old_usize, old_tctx);
prof_free_sampled_object(tsd, ptr, old_usize, old_tctx);
}
}
......@@ -199,7 +243,7 @@ prof_free(tsd_t *tsd, const void *ptr, size_t usize, alloc_ctx_t *alloc_ctx) {
assert(usize == isalloc(tsd_tsdn(tsd), ptr));
if (unlikely((uintptr_t)tctx > (uintptr_t)1U)) {
prof_free_sampled_object(tsd, usize, tctx);
prof_free_sampled_object(tsd, ptr, usize, tctx);
}
}
......
......@@ -169,7 +169,6 @@ struct prof_tdata_s {
/* Sampling state. */
uint64_t prng_state;
uint64_t bytes_until_sample;
/* State used to avoid dumping while operating on prof internals. */
bool enq;
......
#ifndef JEMALLOC_INTERNAL_QUANTUM_H
#define JEMALLOC_INTERNAL_QUANTUM_H
/*
* Minimum allocation alignment is 2^LG_QUANTUM bytes (ignoring tiny size
* classes).
*/
#ifndef LG_QUANTUM
# if (defined(__i386__) || defined(_M_IX86))
# define LG_QUANTUM 4
# endif
# ifdef __ia64__
# define LG_QUANTUM 4
# endif
# ifdef __alpha__
# define LG_QUANTUM 4
# endif
# if (defined(__sparc64__) || defined(__sparcv9) || defined(__sparc_v9__))
# define LG_QUANTUM 4
# endif
# if (defined(__amd64__) || defined(__x86_64__) || defined(_M_X64))
# define LG_QUANTUM 4
# endif
# ifdef __arm__
# define LG_QUANTUM 3
# endif
# ifdef __aarch64__
# define LG_QUANTUM 4
# endif
# ifdef __hppa__
# define LG_QUANTUM 4
# endif
# ifdef __m68k__
# define LG_QUANTUM 3
# endif
# ifdef __mips__
# define LG_QUANTUM 3
# endif
# ifdef __nios2__
# define LG_QUANTUM 3
# endif
# ifdef __or1k__
# define LG_QUANTUM 3
# endif
# ifdef __powerpc__
# define LG_QUANTUM 4
# endif
# if defined(__riscv) || defined(__riscv__)
# define LG_QUANTUM 4
# endif
# ifdef __s390__
# define LG_QUANTUM 4
# endif
# if (defined (__SH3E__) || defined(__SH4_SINGLE__) || defined(__SH4__) || \
defined(__SH4_SINGLE_ONLY__))
# define LG_QUANTUM 4
# endif
# ifdef __tile__
# define LG_QUANTUM 4
# endif
# ifdef __le32__
# define LG_QUANTUM 4
# endif
# ifndef LG_QUANTUM
# error "Unknown minimum alignment for architecture; specify via "
"--with-lg-quantum"
# endif
#endif
#define QUANTUM ((size_t)(1U << LG_QUANTUM))
#define QUANTUM_MASK (QUANTUM - 1)
/* Return the smallest quantum multiple that is >= a. */
#define QUANTUM_CEILING(a) \
(((a) + QUANTUM_MASK) & ~QUANTUM_MASK)
#endif /* JEMALLOC_INTERNAL_QUANTUM_H */
Markdown is supported
0% or .
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment