project-base/server-skynet-source-3rd-jemalloc
2
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity
4094b7c03f
server-skynet-source-3rd-je.../src/prof.c

1154 lines
27 KiB
C
Raw Normal View History

Replace tabs following #define with spaces. This resolves #564.
2017-01-20 13:41:41 +08:00
#define JEMALLOC_PROF_C_
Header refactoring: Split up jemalloc_internal.h This is a biggy. jemalloc_internal.h has been doing multiple jobs for a while now: - The source of system-wide definitions. - The catch-all include file. - The module header file for jemalloc.c This commit splits up this functionality. The system-wide definitions responsibility has moved to jemalloc_preamble.h. The catch-all include file is now jemalloc_internal_includes.h. The module headers for jemalloc.c are now in jemalloc_internal_[externs|inlines|types].h, just as they are for the other modules.
2017-04-11 09:17:55 +08:00
#include "jemalloc/internal/jemalloc_preamble.h"
#include "jemalloc/internal/jemalloc_internal_includes.h"
Add "prof.dump_prefix" to override filename prefixes for dumps.
2019-09-10 11:18:41 +08:00
#include "jemalloc/internal/ctl.h"
Header refactoring: move assert.h out of the catch-all
2017-04-12 05:43:12 +08:00
#include "jemalloc/internal/assert.h"
Header refactoring: unify and de-catchall mutex module
2017-05-24 03:28:19 +08:00
#include "jemalloc/internal/mutex.h"
Header refactoring: move malloc_io.h out of the catchall
2017-04-12 04:06:31 +08:00
Refactor profiling Refactored core profiling codebase into two logical parts: (a) `prof_data.c`: core internal data structure managing & dumping; (b) `prof.c`: mutexes & outward-facing APIs. Some internal functions had to be exposed out, but there are not that many of them if the modularization is (hopefully) clean enough.
2019-07-18 06:52:50 +08:00
/*
* This file implements the profiling "APIs" needed by other parts of jemalloc,
* and also manages the relevant "operational" data, mainly options and mutexes;
* the core profiling data structures are encapsulated in prof_data.c.
*/
Implement allocation profiling and leack checking. Add the --enable-prof and --enable-prof-libunwind configure options. Add the B/b, F/f, I/i, L/l, and U/u JEMALLOC_OPTIONS. Interval-based profile dump triggering is not yet implemented. Add supporting generic code: * Add memory barriers. * Add prn (LCG PRNG). * Add hash (Murmur hash function). * Add ckh (cuckoo hash tables).
2010-02-11 02:37:56 +08:00
/******************************************************************************/
#ifdef JEMALLOC_PROF_LIBUNWIND
Replace tabs following #define with spaces. This resolves #564.
2017-01-20 13:41:41 +08:00
#define UNW_LOCAL_ONLY
Implement allocation profiling and leack checking. Add the --enable-prof and --enable-prof-libunwind configure options. Add the B/b, F/f, I/i, L/l, and U/u JEMALLOC_OPTIONS. Interval-based profile dump triggering is not yet implemented. Add supporting generic code: * Add memory barriers. * Add prn (LCG PRNG). * Add hash (Murmur hash function). * Add ckh (cuckoo hash tables).
2010-02-11 02:37:56 +08:00
#include <libunwind.h>
#endif
Improve backtracing-related configuration. Clean up configuration for backtracing when profiling is enabled, and document the configuration logic in INSTALL. Disable libgcc-based backtracing except on x64 (where it is known to work). Add the --disable-prof-gcc option.
2011-03-16 13:23:12 +08:00
#ifdef JEMALLOC_PROF_LIBGCC
Add hooking functionality This allows us to hook chosen functions and do interesting things there (in particular: reentrancy checking).
2017-03-29 08:30:54 +08:00
/*
* We have a circular dependency -- jemalloc_internal.h tells us if we should
* use libgcc's unwinding functionality, but after we've included that, we've
* already hooked _Unwind_Backtrace. We'll temporarily disable hooking.
*/
#undef _Unwind_Backtrace
Improve backtracing-related configuration. Clean up configuration for backtracing when profiling is enabled, and document the configuration logic in INSTALL. Disable libgcc-based backtracing except on x64 (where it is known to work). Add the --disable-prof-gcc option.
2011-03-16 13:23:12 +08:00
#include <unwind.h>
Rename hooks module to test_hooks. "Hooks" is really the best name for the module that will contain the publicly exposed hooks. So lets rename the current "hooks" module (that hook external dependencies, for reentrancy testing) to "test_hooks".
2018-04-10 09:09:34 +08:00
#define _Unwind_Backtrace JEMALLOC_HOOK(_Unwind_Backtrace, test_hooks_libc_hook)
Improve backtracing-related configuration. Clean up configuration for backtracing when profiling is enabled, and document the configuration logic in INSTALL. Disable libgcc-based backtracing except on x64 (where it is known to work). Add the --disable-prof-gcc option.
2011-03-16 13:23:12 +08:00
#endif
Implement allocation profiling and leack checking. Add the --enable-prof and --enable-prof-libunwind configure options. Add the B/b, F/f, I/i, L/l, and U/u JEMALLOC_OPTIONS. Interval-based profile dump triggering is not yet implemented. Add supporting generic code: * Add memory barriers. * Add prn (LCG PRNG). * Add hash (Murmur hash function). * Add ckh (cuckoo hash tables).
2010-02-11 02:37:56 +08:00
/******************************************************************************/
/* Data. */
bool opt_prof = false;
Add sampling activation/deactivation control. Add the E/e options to control whether the application starts with sampling active/inactive (secondary control to F/f). Add the prof.active mallctl so that the application can activate/deactivate sampling on the fly.
2010-04-01 09:43:24 +08:00
bool opt_prof_active = true;
Implement/test/fix prof-related mallctl's. Implement/test/fix the opt.prof_thread_active_init, prof.thread_active_init, and thread.prof.active mallctl's. Test/fix the thread.prof.name mallctl. Refactor opt_prof_active to be read-only and move mutable state into the prof_active variable. Stop leaning on ctl-related locking for protection.
2014-10-04 14:25:30 +08:00
bool opt_prof_thread_active_init = true;
Implement sampling for heap profiling.
2010-03-02 12:15:26 +08:00
size_t opt_lg_prof_sample = LG_PROF_SAMPLE_DEFAULT;
Make interval-triggered profile dumping optional. Make it possible to disable interval-triggered profile dumping, even if profiling is enabled. This is useful if the user only wants a single dump at exit, or if the application manually triggers profile dumps.
2010-04-01 08:35:51 +08:00
ssize_t opt_lg_prof_interval = LG_PROF_INTERVAL_DEFAULT;
Replace JEMALLOC_OPTIONS with MALLOC_CONF. Replace the single-character run-time flags with key/value pairs, which can be set via the malloc_conf global, /etc/malloc.conf, and the MALLOC_CONF environment variable. Replace the JEMALLOC_PROF_PREFIX environment variable with the "opt.prof_prefix" option. Replace umax2s() with u2s().
2010-10-24 09:37:06 +08:00
bool opt_prof_gdump = false;
Avoid atexit(3) when possible, disable prof_final by default. atexit(3) can deadlock internally during its own initialization if jemalloc calls atexit() during jemalloc initialization. Mitigate the impact by restructuring prof initialization to avoid calling atexit() unless the registered function will actually dump a final heap profile. Additionally, disable prof_final by default so that this land mine is opt-in rather than opt-out. This resolves #144.
2014-10-09 08:57:19 +08:00
bool opt_prof_final = false;
Implement allocation profiling and leack checking. Add the --enable-prof and --enable-prof-libunwind configure options. Add the B/b, F/f, I/i, L/l, and U/u JEMALLOC_OPTIONS. Interval-based profile dump triggering is not yet implemented. Add supporting generic code: * Add memory barriers. * Add prn (LCG PRNG). * Add hash (Murmur hash function). * Add ckh (cuckoo hash tables).
2010-02-11 02:37:56 +08:00
bool opt_prof_leak = false;
Update prof defaults to match common usage. Change the "opt.lg_prof_sample" default from 0 to 19 (1 B to 512 KiB). Change the "opt.prof_accum" default from true to false. Add the "opt.prof_final" mallctl, so that "opt.prof_prefix" need not be abused to disable final profile dumping.
2012-04-18 07:39:33 +08:00
bool opt_prof_accum = false;
Add "prof.dump_prefix" to override filename prefixes for dumps.
2019-09-10 11:18:41 +08:00
char opt_prof_prefix[PROF_DUMP_FILENAME_LEN];
Implement allocation profiling and leack checking. Add the --enable-prof and --enable-prof-libunwind configure options. Add the B/b, F/f, I/i, L/l, and U/u JEMALLOC_OPTIONS. Interval-based profile dump triggering is not yet implemented. Add supporting generic code: * Add memory barriers. * Add prn (LCG PRNG). * Add hash (Murmur hash function). * Add ckh (cuckoo hash tables).
2010-02-11 02:37:56 +08:00
Implement/test/fix prof-related mallctl's. Implement/test/fix the opt.prof_thread_active_init, prof.thread_active_init, and thread.prof.active mallctl's. Test/fix the thread.prof.name mallctl. Refactor opt_prof_active to be read-only and move mutable state into the prof_active variable. Stop leaning on ctl-related locking for protection.
2014-10-04 14:25:30 +08:00
/*
* Initialized as opt_prof_active, and accessed via
* prof_active_[gs]et{_unlocked,}().
*/
bool prof_active;
static malloc_mutex_t prof_active_mtx;
/*
* Initialized as opt_prof_thread_active_init, and accessed via
* prof_thread_active_init_[gs]et().
*/
static bool prof_thread_active_init;
static malloc_mutex_t prof_thread_active_init_mtx;
Implement the prof.gdump mallctl. This feature makes it possible to toggle the gdump feature on/off during program execution, whereas the the opt.prof_dump mallctl value can only be set during program startup. This resolves #72.
2015-01-26 13:16:57 +08:00
/*
* Initialized as opt_prof_gdump, and accessed via
* prof_gdump_[gs]et{_unlocked,}().
*/
bool prof_gdump_val;
static malloc_mutex_t prof_gdump_mtx;
Avoid arena_prof_accum()-related locking when possible. Refactor arena_prof_accum() and its callers to avoid arena locking when prof_interval is 0 (as when profiling is disabled). Reported by Ben Maurer.
2012-11-14 04:56:27 +08:00
uint64_t prof_interval = 0;
Implement interval-based heap profile dumping. Add mallctl interfaces for profiling parameters. Fix a file descriptor leak in heap profile dumping.
2010-02-12 05:19:21 +08:00
Implement per thread heap profiling. Rename data structures (prof_thr_cnt_t-->prof_tctx_t, prof_ctx_t-->prof_gctx_t), and convert to storing a prof_tctx_t for sampled objects. Convert PROF_ALLOC_PREP() to prof_alloc_prep(), since precise backtrace depth within jemalloc functions is no longer an issue (pprof prunes irrelevant frames). Implement mallctl's: - prof.reset implements full sample data reset, and optional change of sample interval. - prof.lg_sample reads the current sample interval (opt.lg_prof_sample was the permanent source of truth prior to prof.reset). - thread.prof.name provides naming capability for threads within heap profile dumps. - thread.prof.active makes it possible to activate/deactivate heap profiling for individual threads. Modify the heap dump files to contain per thread heap profile data. This change is incompatible with the existing pprof, which will require enhancements to read and process the enriched data.
2014-08-19 07:22:13 +08:00
size_t lg_prof_sample;
Remove ephemeral mutexes. Remove ephemeral mutexes from the prof machinery, and remove malloc_mutex_destroy(). This simplifies mutex management on systems that call malloc()/free() inside pthread_mutex_{create,destroy}(). Add atomic_*_u() for operation on unsigned values. Fix prof_printf() to call malloc_vsnprintf() rather than malloc_snprintf().
2012-03-24 09:05:51 +08:00
/*
Implement per thread heap profiling. Rename data structures (prof_thr_cnt_t-->prof_tctx_t, prof_ctx_t-->prof_gctx_t), and convert to storing a prof_tctx_t for sampled objects. Convert PROF_ALLOC_PREP() to prof_alloc_prep(), since precise backtrace depth within jemalloc functions is no longer an issue (pprof prunes irrelevant frames). Implement mallctl's: - prof.reset implements full sample data reset, and optional change of sample interval. - prof.lg_sample reads the current sample interval (opt.lg_prof_sample was the permanent source of truth prior to prof.reset). - thread.prof.name provides naming capability for threads within heap profile dumps. - thread.prof.active makes it possible to activate/deactivate heap profiling for individual threads. Modify the heap dump files to contain per thread heap profile data. This change is incompatible with the existing pprof, which will require enhancements to read and process the enriched data.
2014-08-19 07:22:13 +08:00
* Table of mutexes that are shared among gctx's. These are leaf locks, so
* there is no problem with using them for more than one gctx at the same time.
* The primary motivation for this sharing though is that gctx's are ephemeral,
Remove ephemeral mutexes. Remove ephemeral mutexes from the prof machinery, and remove malloc_mutex_destroy(). This simplifies mutex management on systems that call malloc()/free() inside pthread_mutex_{create,destroy}(). Add atomic_*_u() for operation on unsigned values. Fix prof_printf() to call malloc_vsnprintf() rather than malloc_snprintf().
2012-03-24 09:05:51 +08:00
* and destroying mutexes causes complications for systems that allocate when
* creating/destroying mutexes.
*/
Implement per thread heap profiling. Rename data structures (prof_thr_cnt_t-->prof_tctx_t, prof_ctx_t-->prof_gctx_t), and convert to storing a prof_tctx_t for sampled objects. Convert PROF_ALLOC_PREP() to prof_alloc_prep(), since precise backtrace depth within jemalloc functions is no longer an issue (pprof prunes irrelevant frames). Implement mallctl's: - prof.reset implements full sample data reset, and optional change of sample interval. - prof.lg_sample reads the current sample interval (opt.lg_prof_sample was the permanent source of truth prior to prof.reset). - thread.prof.name provides naming capability for threads within heap profile dumps. - thread.prof.active makes it possible to activate/deactivate heap profiling for individual threads. Modify the heap dump files to contain per thread heap profile data. This change is incompatible with the existing pprof, which will require enhancements to read and process the enriched data.
2014-08-19 07:22:13 +08:00
static malloc_mutex_t *gctx_locks;
Make prof's cum_gctx a C11-style atomic
2017-04-05 09:36:45 +08:00
static atomic_u_t cum_gctxs; /* Atomic counter. */
Remove ephemeral mutexes. Remove ephemeral mutexes from the prof machinery, and remove malloc_mutex_destroy(). This simplifies mutex management on systems that call malloc()/free() inside pthread_mutex_{create,destroy}(). Add atomic_*_u() for operation on unsigned values. Fix prof_printf() to call malloc_vsnprintf() rather than malloc_snprintf().
2012-03-24 09:05:51 +08:00
Increase default backtrace depth from 4 to 128. Increase the default backtrace depth, because shallow backtraces tend to result in confusing pprof output graphs.
2010-10-03 13:38:14 +08:00
/*
Implement per thread heap profiling. Rename data structures (prof_thr_cnt_t-->prof_tctx_t, prof_ctx_t-->prof_gctx_t), and convert to storing a prof_tctx_t for sampled objects. Convert PROF_ALLOC_PREP() to prof_alloc_prep(), since precise backtrace depth within jemalloc functions is no longer an issue (pprof prunes irrelevant frames). Implement mallctl's: - prof.reset implements full sample data reset, and optional change of sample interval. - prof.lg_sample reads the current sample interval (opt.lg_prof_sample was the permanent source of truth prior to prof.reset). - thread.prof.name provides naming capability for threads within heap profile dumps. - thread.prof.active makes it possible to activate/deactivate heap profiling for individual threads. Modify the heap dump files to contain per thread heap profile data. This change is incompatible with the existing pprof, which will require enhancements to read and process the enriched data.
2014-08-19 07:22:13 +08:00
* Table of mutexes that are shared among tdata's. No operations require
* holding multiple tdata locks, so there is no problem with using them for more
* than one tdata at the same time, even though a gctx lock may be acquired
* while holding a tdata lock.
*/
static malloc_mutex_t *tdata_locks;
Added lock profiling and output for global locks (ctl, prof and base).
2017-03-12 12:28:31 +08:00
/* Non static to enable profiling. */
malloc_mutex_t bt2gctx_mtx;
Implement per thread heap profiling. Rename data structures (prof_thr_cnt_t-->prof_tctx_t, prof_ctx_t-->prof_gctx_t), and convert to storing a prof_tctx_t for sampled objects. Convert PROF_ALLOC_PREP() to prof_alloc_prep(), since precise backtrace depth within jemalloc functions is no longer an issue (pprof prunes irrelevant frames). Implement mallctl's: - prof.reset implements full sample data reset, and optional change of sample interval. - prof.lg_sample reads the current sample interval (opt.lg_prof_sample was the permanent source of truth prior to prof.reset). - thread.prof.name provides naming capability for threads within heap profile dumps. - thread.prof.active makes it possible to activate/deactivate heap profiling for individual threads. Modify the heap dump files to contain per thread heap profile data. This change is incompatible with the existing pprof, which will require enhancements to read and process the enriched data.
2014-08-19 07:22:13 +08:00
Refactor profiling Refactored core profiling codebase into two logical parts: (a) `prof_data.c`: core internal data structure managing & dumping; (b) `prof.c`: mutexes & outward-facing APIs. Some internal functions had to be exposed out, but there are not that many of them if the modularization is (hopefully) clean enough.
2019-07-18 06:52:50 +08:00
malloc_mutex_t tdatas_mtx;
Implement per thread heap profiling. Rename data structures (prof_thr_cnt_t-->prof_tctx_t, prof_ctx_t-->prof_gctx_t), and convert to storing a prof_tctx_t for sampled objects. Convert PROF_ALLOC_PREP() to prof_alloc_prep(), since precise backtrace depth within jemalloc functions is no longer an issue (pprof prunes irrelevant frames). Implement mallctl's: - prof.reset implements full sample data reset, and optional change of sample interval. - prof.lg_sample reads the current sample interval (opt.lg_prof_sample was the permanent source of truth prior to prof.reset). - thread.prof.name provides naming capability for threads within heap profile dumps. - thread.prof.active makes it possible to activate/deactivate heap profiling for individual threads. Modify the heap dump files to contain per thread heap profile data. This change is incompatible with the existing pprof, which will require enhancements to read and process the enriched data.
2014-08-19 07:22:13 +08:00
static uint64_t next_thr_uid;
Fix prof regressions. Don't use atomic_add_uint64(), because it isn't available on 32-bit platforms. Fix forking support functions to manage all prof-related mutexes. These regressions were introduced by 602c8e0971160e4b85b08b16cf8a2375aa24bc04 (Implement per thread heap profiling.), which did not make it into any releases prior to these fixes.
2014-09-12 09:06:30 +08:00
static malloc_mutex_t next_thr_uid_mtx;
Implement sampling for heap profiling.
2010-03-02 12:15:26 +08:00
Rename "prof_dump_seq_mtx" to "prof_dump_filename_mtx".
2019-09-10 11:04:18 +08:00
static malloc_mutex_t prof_dump_filename_mtx;
Implement allocation profiling and leack checking. Add the --enable-prof and --enable-prof-libunwind configure options. Add the B/b, F/f, I/i, L/l, and U/u JEMALLOC_OPTIONS. Interval-based profile dump triggering is not yet implemented. Add supporting generic code: * Add memory barriers. * Add prn (LCG PRNG). * Add hash (Murmur hash function). * Add ckh (cuckoo hash tables).
2010-02-11 02:37:56 +08:00
static uint64_t prof_dump_seq;
static uint64_t prof_dump_iseq;
static uint64_t prof_dump_mseq;
static uint64_t prof_dump_useq;
Refactor profiling Refactored core profiling codebase into two logical parts: (a) `prof_data.c`: core internal data structure managing & dumping; (b) `prof.c`: mutexes & outward-facing APIs. Some internal functions had to be exposed out, but there are not that many of them if the modularization is (hopefully) clean enough.
2019-07-18 06:52:50 +08:00
malloc_mutex_t prof_dump_mtx;
Add "prof.dump_prefix" to override filename prefixes for dumps.
2019-09-10 11:18:41 +08:00
static char *prof_dump_prefix = NULL;
Implement allocation profiling and leack checking. Add the --enable-prof and --enable-prof-libunwind configure options. Add the B/b, F/f, I/i, L/l, and U/u JEMALLOC_OPTIONS. Interval-based profile dump triggering is not yet implemented. Add supporting generic code: * Add memory barriers. * Add prn (LCG PRNG). * Add hash (Murmur hash function). * Add ckh (cuckoo hash tables).
2010-02-11 02:37:56 +08:00
/* Do not dump any profiles until bootstrapping is complete. */
Refactor prof log Prof logging is conceptually seperate from core profiling, so split it out as a module of its own. There are a few internal functions that had to be exposed but I think it is a fair trade-off.
2019-07-13 07:37:37 +08:00
bool prof_booted = false;
Implement allocation profiling and leack checking. Add the --enable-prof and --enable-prof-libunwind configure options. Add the B/b, F/f, I/i, L/l, and U/u JEMALLOC_OPTIONS. Interval-based profile dump triggering is not yet implemented. Add supporting generic code: * Add memory barriers. * Add prn (LCG PRNG). * Add hash (Murmur hash function). * Add ckh (cuckoo hash tables).
2010-02-11 02:37:56 +08:00
/******************************************************************************/
Implement per thread heap profiling. Rename data structures (prof_thr_cnt_t-->prof_tctx_t, prof_ctx_t-->prof_gctx_t), and convert to storing a prof_tctx_t for sampled objects. Convert PROF_ALLOC_PREP() to prof_alloc_prep(), since precise backtrace depth within jemalloc functions is no longer an issue (pprof prunes irrelevant frames). Implement mallctl's: - prof.reset implements full sample data reset, and optional change of sample interval. - prof.lg_sample reads the current sample interval (opt.lg_prof_sample was the permanent source of truth prior to prof.reset). - thread.prof.name provides naming capability for threads within heap profile dumps. - thread.prof.active makes it possible to activate/deactivate heap profiling for individual threads. Modify the heap dump files to contain per thread heap profile data. This change is incompatible with the existing pprof, which will require enhancements to read and process the enriched data.
2014-08-19 07:22:13 +08:00
Refactor profiling Refactored core profiling codebase into two logical parts: (a) `prof_data.c`: core internal data structure managing & dumping; (b) `prof.c`: mutexes & outward-facing APIs. Some internal functions had to be exposed out, but there are not that many of them if the modularization is (hopefully) clean enough.
2019-07-18 06:52:50 +08:00
static bool
prof_tctx_should_destroy(tsdn_t *tsdn, prof_tctx_t *tctx) {
malloc_mutex_assert_owner(tsdn, tctx->tdata->lock);
Implement per thread heap profiling. Rename data structures (prof_thr_cnt_t-->prof_tctx_t, prof_ctx_t-->prof_gctx_t), and convert to storing a prof_tctx_t for sampled objects. Convert PROF_ALLOC_PREP() to prof_alloc_prep(), since precise backtrace depth within jemalloc functions is no longer an issue (pprof prunes irrelevant frames). Implement mallctl's: - prof.reset implements full sample data reset, and optional change of sample interval. - prof.lg_sample reads the current sample interval (opt.lg_prof_sample was the permanent source of truth prior to prof.reset). - thread.prof.name provides naming capability for threads within heap profile dumps. - thread.prof.active makes it possible to activate/deactivate heap profiling for individual threads. Modify the heap dump files to contain per thread heap profile data. This change is incompatible with the existing pprof, which will require enhancements to read and process the enriched data.
2014-08-19 07:22:13 +08:00
Refactor profiling Refactored core profiling codebase into two logical parts: (a) `prof_data.c`: core internal data structure managing & dumping; (b) `prof.c`: mutexes & outward-facing APIs. Some internal functions had to be exposed out, but there are not that many of them if the modularization is (hopefully) clean enough.
2019-07-18 06:52:50 +08:00
if (opt_prof_accum) {
return false;
Fix a heap profiling regression. Fix prof_tctx_comp() to incorporate tctx state into the comparison. During a dump it is possible for both a purgatory tctx and an otherwise equivalent nominal tctx to reside in the tree at the same time. This regression was introduced by 602c8e0971160e4b85b08b16cf8a2375aa24bc04 (Implement per thread heap profiling.).
2015-03-13 07:25:18 +08:00
}
Refactor profiling Refactored core profiling codebase into two logical parts: (a) `prof_data.c`: core internal data structure managing & dumping; (b) `prof.c`: mutexes & outward-facing APIs. Some internal functions had to be exposed out, but there are not that many of them if the modularization is (hopefully) clean enough.
2019-07-18 06:52:50 +08:00
if (tctx->cnts.curobjs != 0) {
return false;
Update brace style. Add braces around single-line blocks, and remove line breaks before function-opening braces. This resolves #537.
2017-01-16 08:56:30 +08:00
}
Refactor profiling Refactored core profiling codebase into two logical parts: (a) `prof_data.c`: core internal data structure managing & dumping; (b) `prof.c`: mutexes & outward-facing APIs. Some internal functions had to be exposed out, but there are not that many of them if the modularization is (hopefully) clean enough.
2019-07-18 06:52:50 +08:00
if (tctx->prepared) {
return false;
Test prof.reset mallctl and fix numerous discovered bugs.
2014-10-03 14:01:10 +08:00
}
Refactor profiling Refactored core profiling codebase into two logical parts: (a) `prof_data.c`: core internal data structure managing & dumping; (b) `prof.c`: mutexes & outward-facing APIs. Some internal functions had to be exposed out, but there are not that many of them if the modularization is (hopefully) clean enough.
2019-07-18 06:52:50 +08:00
return true;
Implement per thread heap profiling. Rename data structures (prof_thr_cnt_t-->prof_tctx_t, prof_ctx_t-->prof_gctx_t), and convert to storing a prof_tctx_t for sampled objects. Convert PROF_ALLOC_PREP() to prof_alloc_prep(), since precise backtrace depth within jemalloc functions is no longer an issue (pprof prunes irrelevant frames). Implement mallctl's: - prof.reset implements full sample data reset, and optional change of sample interval. - prof.lg_sample reads the current sample interval (opt.lg_prof_sample was the permanent source of truth prior to prof.reset). - thread.prof.name provides naming capability for threads within heap profile dumps. - thread.prof.active makes it possible to activate/deactivate heap profiling for individual threads. Modify the heap dump files to contain per thread heap profile data. This change is incompatible with the existing pprof, which will require enhancements to read and process the enriched data.
2014-08-19 07:22:13 +08:00
}
Fix a profile sampling race. Fix a profile sampling race that was due to preparing to sample, yet doing nothing to assure that the context remains valid until the stats are updated. These regressions were caused by 602c8e0971160e4b85b08b16cf8a2375aa24bc04 (Implement per thread heap profiling.), which did not make it into any releases prior to these fixes.
2014-09-10 10:37:26 +08:00
void
Update brace style. Add braces around single-line blocks, and remove line breaks before function-opening braces. This resolves #537.
2017-01-16 08:56:30 +08:00
prof_alloc_rollback(tsd_t *tsd, prof_tctx_t *tctx, bool updated) {
Fix a profile sampling race. Fix a profile sampling race that was due to preparing to sample, yet doing nothing to assure that the context remains valid until the stats are updated. These regressions were caused by 602c8e0971160e4b85b08b16cf8a2375aa24bc04 (Implement per thread heap profiling.), which did not make it into any releases prior to these fixes.
2014-09-10 10:37:26 +08:00
cassert(config_prof);
Suppress tdata creation in reentrancy This change suppresses tdata initialization and prof sample threshold update in interrupting malloc calls. Interrupting calls have no need for tdata. Delaying tdata creation aligns better with our lazy tdata creation principle, and it also helps us gain control back from interrupting calls more quickly and reduces any risk of delegating tdata creation to an interrupting call.
2019-10-04 04:01:12 +08:00
if (tsd_reentrancy_level_get(tsd) > 0) {
assert((uintptr_t)tctx == (uintptr_t)1U);
return;
}
prof_tdata_t *tdata;
Fix a profile sampling race. Fix a profile sampling race that was due to preparing to sample, yet doing nothing to assure that the context remains valid until the stats are updated. These regressions were caused by 602c8e0971160e4b85b08b16cf8a2375aa24bc04 (Implement per thread heap profiling.), which did not make it into any releases prior to these fixes.
2014-09-10 10:37:26 +08:00
if (updated) {
/*
* Compute a new sample threshold. This isn't very important in
* practice, because this function is rarely executed, so the
* potential for sample bias is minimal except in contrived
* programs.
*/
Convert all tsd variables to reside in a single tsd structure.
2014-09-23 12:09:23 +08:00
tdata = prof_tdata_get(tsd, true);
Update brace style. Add braces around single-line blocks, and remove line breaks before function-opening braces. This resolves #537.
2017-01-16 08:56:30 +08:00
if (tdata != NULL) {
Fix prof_alloc_rollback(). Fix prof_alloc_rollback() to read tdata from thread-specific data rather than dereferencing a potentially invalid tctx.
2015-09-18 05:47:39 +08:00
prof_sample_threshold_update(tdata);
Update brace style. Add braces around single-line blocks, and remove line breaks before function-opening braces. This resolves #537.
2017-01-16 08:56:30 +08:00
}
Fix a profile sampling race. Fix a profile sampling race that was due to preparing to sample, yet doing nothing to assure that the context remains valid until the stats are updated. These regressions were caused by 602c8e0971160e4b85b08b16cf8a2375aa24bc04 (Implement per thread heap profiling.), which did not make it into any releases prior to these fixes.
2014-09-10 10:37:26 +08:00
}
if ((uintptr_t)tctx > (uintptr_t)1U) {
Resolve bootstrapping issues when embedded in FreeBSD libc. b2c0d6322d2307458ae2b28545f8a5c9903d7ef5 (Add witness, a simple online locking validator.) caused a broad propagation of tsd throughout the internal API, but tsd_fetch() was designed to fail prior to tsd bootstrapping. Fix this by splitting tsd_t into non-nullable tsd_t and nullable tsdn_t, and modifying all internal APIs that do not critically rely on tsd to take nullable pointers. Furthermore, add the tsd_booted_get() function so that tsdn_fetch() can probe whether tsd bootstrapping is complete and return NULL if not. All dangerous conversions of nullable pointers are tsdn_tsd() calls that assert-fail on invalid conversion.
2016-05-11 13:21:10 +08:00
malloc_mutex_lock(tsd_tsdn(tsd), tctx->tdata->lock);
Fix a profile sampling race. Fix a profile sampling race that was due to preparing to sample, yet doing nothing to assure that the context remains valid until the stats are updated. These regressions were caused by 602c8e0971160e4b85b08b16cf8a2375aa24bc04 (Implement per thread heap profiling.), which did not make it into any releases prior to these fixes.
2014-09-10 10:37:26 +08:00
tctx->prepared = false;
Update brace style. Add braces around single-line blocks, and remove line breaks before function-opening braces. This resolves #537.
2017-01-16 08:56:30 +08:00
if (prof_tctx_should_destroy(tsd_tsdn(tsd), tctx)) {
Convert all tsd variables to reside in a single tsd structure.
2014-09-23 12:09:23 +08:00
prof_tctx_destroy(tsd, tctx);
Update brace style. Add braces around single-line blocks, and remove line breaks before function-opening braces. This resolves #537.
2017-01-16 08:56:30 +08:00
} else {
Resolve bootstrapping issues when embedded in FreeBSD libc. b2c0d6322d2307458ae2b28545f8a5c9903d7ef5 (Add witness, a simple online locking validator.) caused a broad propagation of tsd throughout the internal API, but tsd_fetch() was designed to fail prior to tsd bootstrapping. Fix this by splitting tsd_t into non-nullable tsd_t and nullable tsdn_t, and modifying all internal APIs that do not critically rely on tsd to take nullable pointers. Furthermore, add the tsd_booted_get() function so that tsdn_fetch() can probe whether tsd bootstrapping is complete and return NULL if not. All dangerous conversions of nullable pointers are tsdn_tsd() calls that assert-fail on invalid conversion.
2016-05-11 13:21:10 +08:00
malloc_mutex_unlock(tsd_tsdn(tsd), tctx->tdata->lock);
Update brace style. Add braces around single-line blocks, and remove line breaks before function-opening braces. This resolves #537.
2017-01-16 08:56:30 +08:00
}
Fix a profile sampling race. Fix a profile sampling race that was due to preparing to sample, yet doing nothing to assure that the context remains valid until the stats are updated. These regressions were caused by 602c8e0971160e4b85b08b16cf8a2375aa24bc04 (Implement per thread heap profiling.), which did not make it into any releases prior to these fixes.
2014-09-10 10:37:26 +08:00
}
}
Implement per thread heap profiling. Rename data structures (prof_thr_cnt_t-->prof_tctx_t, prof_ctx_t-->prof_gctx_t), and convert to storing a prof_tctx_t for sampled objects. Convert PROF_ALLOC_PREP() to prof_alloc_prep(), since precise backtrace depth within jemalloc functions is no longer an issue (pprof prunes irrelevant frames). Implement mallctl's: - prof.reset implements full sample data reset, and optional change of sample interval. - prof.lg_sample reads the current sample interval (opt.lg_prof_sample was the permanent source of truth prior to prof.reset). - thread.prof.name provides naming capability for threads within heap profile dumps. - thread.prof.active makes it possible to activate/deactivate heap profiling for individual threads. Modify the heap dump files to contain per thread heap profile data. This change is incompatible with the existing pprof, which will require enhancements to read and process the enriched data.
2014-08-19 07:22:13 +08:00
void
Push down iealloc() calls. Call iealloc() as deep into call chains as possible without causing redundant calls.
2017-03-21 02:00:07 +08:00
prof_malloc_sample_object(tsdn_t *tsdn, const void *ptr, size_t usize,
prof_tctx_t *tctx) {
Pass alloc_ctx down profiling path. With this change, when profiling is enabled, we avoid doing redundant rtree lookups. Also changed dalloc_atx_t to alloc_atx_t, as it's now used on allocation path as well (to speed up profiling).
2017-04-12 09:13:10 +08:00
prof_tctx_set(tsdn, ptr, usize, NULL, tctx);
Implement per thread heap profiling. Rename data structures (prof_thr_cnt_t-->prof_tctx_t, prof_ctx_t-->prof_gctx_t), and convert to storing a prof_tctx_t for sampled objects. Convert PROF_ALLOC_PREP() to prof_alloc_prep(), since precise backtrace depth within jemalloc functions is no longer an issue (pprof prunes irrelevant frames). Implement mallctl's: - prof.reset implements full sample data reset, and optional change of sample interval. - prof.lg_sample reads the current sample interval (opt.lg_prof_sample was the permanent source of truth prior to prof.reset). - thread.prof.name provides naming capability for threads within heap profile dumps. - thread.prof.active makes it possible to activate/deactivate heap profiling for individual threads. Modify the heap dump files to contain per thread heap profile data. This change is incompatible with the existing pprof, which will require enhancements to read and process the enriched data.
2014-08-19 07:22:13 +08:00
Add logging for sampled allocations - prof_opt_log flag starts logging automatically at runtime - prof_log_{start,stop} mallctl for manual control
2018-07-06 01:56:33 +08:00
/* Get the current time and set this in the extent_t. We'll read this
* when free() is called. */
nstime_t t = NSTIME_ZERO_INITIALIZER;
nstime_update(&t);
prof_alloc_time_set(tsdn, ptr, NULL, t);
Resolve bootstrapping issues when embedded in FreeBSD libc. b2c0d6322d2307458ae2b28545f8a5c9903d7ef5 (Add witness, a simple online locking validator.) caused a broad propagation of tsd throughout the internal API, but tsd_fetch() was designed to fail prior to tsd bootstrapping. Fix this by splitting tsd_t into non-nullable tsd_t and nullable tsdn_t, and modifying all internal APIs that do not critically rely on tsd to take nullable pointers. Furthermore, add the tsd_booted_get() function so that tsdn_fetch() can probe whether tsd bootstrapping is complete and return NULL if not. All dangerous conversions of nullable pointers are tsdn_tsd() calls that assert-fail on invalid conversion.
2016-05-11 13:21:10 +08:00
malloc_mutex_lock(tsdn, tctx->tdata->lock);
Implement per thread heap profiling. Rename data structures (prof_thr_cnt_t-->prof_tctx_t, prof_ctx_t-->prof_gctx_t), and convert to storing a prof_tctx_t for sampled objects. Convert PROF_ALLOC_PREP() to prof_alloc_prep(), since precise backtrace depth within jemalloc functions is no longer an issue (pprof prunes irrelevant frames). Implement mallctl's: - prof.reset implements full sample data reset, and optional change of sample interval. - prof.lg_sample reads the current sample interval (opt.lg_prof_sample was the permanent source of truth prior to prof.reset). - thread.prof.name provides naming capability for threads within heap profile dumps. - thread.prof.active makes it possible to activate/deactivate heap profiling for individual threads. Modify the heap dump files to contain per thread heap profile data. This change is incompatible with the existing pprof, which will require enhancements to read and process the enriched data.
2014-08-19 07:22:13 +08:00
tctx->cnts.curobjs++;
tctx->cnts.curbytes += usize;
if (opt_prof_accum) {
tctx->cnts.accumobjs++;
tctx->cnts.accumbytes += usize;
}
Fix a profile sampling race. Fix a profile sampling race that was due to preparing to sample, yet doing nothing to assure that the context remains valid until the stats are updated. These regressions were caused by 602c8e0971160e4b85b08b16cf8a2375aa24bc04 (Implement per thread heap profiling.), which did not make it into any releases prior to these fixes.
2014-09-10 10:37:26 +08:00
tctx->prepared = false;
Resolve bootstrapping issues when embedded in FreeBSD libc. b2c0d6322d2307458ae2b28545f8a5c9903d7ef5 (Add witness, a simple online locking validator.) caused a broad propagation of tsd throughout the internal API, but tsd_fetch() was designed to fail prior to tsd bootstrapping. Fix this by splitting tsd_t into non-nullable tsd_t and nullable tsdn_t, and modifying all internal APIs that do not critically rely on tsd to take nullable pointers. Furthermore, add the tsd_booted_get() function so that tsdn_fetch() can probe whether tsd bootstrapping is complete and return NULL if not. All dangerous conversions of nullable pointers are tsdn_tsd() calls that assert-fail on invalid conversion.
2016-05-11 13:21:10 +08:00
malloc_mutex_unlock(tsdn, tctx->tdata->lock);
Implement per thread heap profiling. Rename data structures (prof_thr_cnt_t-->prof_tctx_t, prof_ctx_t-->prof_gctx_t), and convert to storing a prof_tctx_t for sampled objects. Convert PROF_ALLOC_PREP() to prof_alloc_prep(), since precise backtrace depth within jemalloc functions is no longer an issue (pprof prunes irrelevant frames). Implement mallctl's: - prof.reset implements full sample data reset, and optional change of sample interval. - prof.lg_sample reads the current sample interval (opt.lg_prof_sample was the permanent source of truth prior to prof.reset). - thread.prof.name provides naming capability for threads within heap profile dumps. - thread.prof.active makes it possible to activate/deactivate heap profiling for individual threads. Modify the heap dump files to contain per thread heap profile data. This change is incompatible with the existing pprof, which will require enhancements to read and process the enriched data.
2014-08-19 07:22:13 +08:00
}
void
Safety checks: Add a redzoning feature.
2019-03-23 03:53:11 +08:00
prof_free_sampled_object(tsd_t *tsd, const void *ptr, size_t usize,
Add logging for sampled allocations - prof_opt_log flag starts logging automatically at runtime - prof_log_{start,stop} mallctl for manual control
2018-07-06 01:56:33 +08:00
prof_tctx_t *tctx) {
Resolve bootstrapping issues when embedded in FreeBSD libc. b2c0d6322d2307458ae2b28545f8a5c9903d7ef5 (Add witness, a simple online locking validator.) caused a broad propagation of tsd throughout the internal API, but tsd_fetch() was designed to fail prior to tsd bootstrapping. Fix this by splitting tsd_t into non-nullable tsd_t and nullable tsdn_t, and modifying all internal APIs that do not critically rely on tsd to take nullable pointers. Furthermore, add the tsd_booted_get() function so that tsdn_fetch() can probe whether tsd bootstrapping is complete and return NULL if not. All dangerous conversions of nullable pointers are tsdn_tsd() calls that assert-fail on invalid conversion.
2016-05-11 13:21:10 +08:00
malloc_mutex_lock(tsd_tsdn(tsd), tctx->tdata->lock);
Add logging for sampled allocations - prof_opt_log flag starts logging automatically at runtime - prof_log_{start,stop} mallctl for manual control
2018-07-06 01:56:33 +08:00
Implement per thread heap profiling. Rename data structures (prof_thr_cnt_t-->prof_tctx_t, prof_ctx_t-->prof_gctx_t), and convert to storing a prof_tctx_t for sampled objects. Convert PROF_ALLOC_PREP() to prof_alloc_prep(), since precise backtrace depth within jemalloc functions is no longer an issue (pprof prunes irrelevant frames). Implement mallctl's: - prof.reset implements full sample data reset, and optional change of sample interval. - prof.lg_sample reads the current sample interval (opt.lg_prof_sample was the permanent source of truth prior to prof.reset). - thread.prof.name provides naming capability for threads within heap profile dumps. - thread.prof.active makes it possible to activate/deactivate heap profiling for individual threads. Modify the heap dump files to contain per thread heap profile data. This change is incompatible with the existing pprof, which will require enhancements to read and process the enriched data.
2014-08-19 07:22:13 +08:00
assert(tctx->cnts.curobjs > 0);
assert(tctx->cnts.curbytes >= usize);
tctx->cnts.curobjs--;
tctx->cnts.curbytes -= usize;
Add logging for sampled allocations - prof_opt_log flag starts logging automatically at runtime - prof_log_{start,stop} mallctl for manual control
2018-07-06 01:56:33 +08:00
prof_try_log(tsd, ptr, usize, tctx);
Update brace style. Add braces around single-line blocks, and remove line breaks before function-opening braces. This resolves #537.
2017-01-16 08:56:30 +08:00
if (prof_tctx_should_destroy(tsd_tsdn(tsd), tctx)) {
Convert all tsd variables to reside in a single tsd structure.
2014-09-23 12:09:23 +08:00
prof_tctx_destroy(tsd, tctx);
Update brace style. Add braces around single-line blocks, and remove line breaks before function-opening braces. This resolves #537.
2017-01-16 08:56:30 +08:00
} else {
Resolve bootstrapping issues when embedded in FreeBSD libc. b2c0d6322d2307458ae2b28545f8a5c9903d7ef5 (Add witness, a simple online locking validator.) caused a broad propagation of tsd throughout the internal API, but tsd_fetch() was designed to fail prior to tsd bootstrapping. Fix this by splitting tsd_t into non-nullable tsd_t and nullable tsdn_t, and modifying all internal APIs that do not critically rely on tsd to take nullable pointers. Furthermore, add the tsd_booted_get() function so that tsdn_fetch() can probe whether tsd bootstrapping is complete and return NULL if not. All dangerous conversions of nullable pointers are tsdn_tsd() calls that assert-fail on invalid conversion.
2016-05-11 13:21:10 +08:00
malloc_mutex_unlock(tsd_tsdn(tsd), tctx->tdata->lock);
Update brace style. Add braces around single-line blocks, and remove line breaks before function-opening braces. This resolves #537.
2017-01-16 08:56:30 +08:00
}
Implement per thread heap profiling. Rename data structures (prof_thr_cnt_t-->prof_tctx_t, prof_ctx_t-->prof_gctx_t), and convert to storing a prof_tctx_t for sampled objects. Convert PROF_ALLOC_PREP() to prof_alloc_prep(), since precise backtrace depth within jemalloc functions is no longer an issue (pprof prunes irrelevant frames). Implement mallctl's: - prof.reset implements full sample data reset, and optional change of sample interval. - prof.lg_sample reads the current sample interval (opt.lg_prof_sample was the permanent source of truth prior to prof.reset). - thread.prof.name provides naming capability for threads within heap profile dumps. - thread.prof.active makes it possible to activate/deactivate heap profiling for individual threads. Modify the heap dump files to contain per thread heap profile data. This change is incompatible with the existing pprof, which will require enhancements to read and process the enriched data.
2014-08-19 07:22:13 +08:00
}
Dump heap profile backtraces in a stable order. Also iterate over per thread stats in a stable order, which prepares the way for stable ordering of per thread heap profile dumps.
2014-08-17 03:58:55 +08:00
Inline the fast path for heap sampling. Inline the heap sampling code that is executed for every allocation event (regardless of whether a sample is taken). Combine all prof TLS data into a single data structure, in order to reduce the TLS lookup volume.
2010-10-21 10:05:59 +08:00
void
Update brace style. Add braces around single-line blocks, and remove line breaks before function-opening braces. This resolves #537.
2017-01-16 08:56:30 +08:00
bt_init(prof_bt_t *bt, void **vec) {
Reduce cpp conditional logic complexity. Convert configuration-related cpp conditional logic to use static constant variables, e.g.: #ifdef JEMALLOC_DEBUG [...] #endif becomes: if (config_debug) { [...] } The advantage is clearer, more concise code. The main disadvantage is that data structures no longer have conditionally defined fields, so they pay the cost of all fields regardless of whether they are used. In practice, this is only a minor concern; config_stats will go away in an upcoming change, and config_prof is the only other major feature that depends on more than a few special-purpose fields.
2012-02-11 12:22:09 +08:00
cassert(config_prof);
Implement allocation profiling and leack checking. Add the --enable-prof and --enable-prof-libunwind configure options. Add the B/b, F/f, I/i, L/l, and U/u JEMALLOC_OPTIONS. Interval-based profile dump triggering is not yet implemented. Add supporting generic code: * Add memory barriers. * Add prn (LCG PRNG). * Add hash (Murmur hash function). * Add ckh (cuckoo hash tables).
2010-02-11 02:37:56 +08:00
bt->vec = vec;
bt->len = 0;
}
Improve backtracing-related configuration. Clean up configuration for backtracing when profiling is enabled, and document the configuration logic in INSTALL. Disable libgcc-based backtracing except on x64 (where it is known to work). Add the --disable-prof-gcc option.
2011-03-16 13:23:12 +08:00
#ifdef JEMALLOC_PROF_LIBUNWIND
Fix prof_backtrace() reentrancy level
2019-09-04 08:11:06 +08:00
static void
prof_backtrace_impl(prof_bt_t *bt) {
Simplify backtracing. Simplify backtracing to not ignore any frames, and compensate for this in pprof in order to increase flexibility with respect to function-based refactoring even in the presence of non-deterministic inlining. Modify pprof to blacklist all jemalloc allocation entry points including non-standard ones like mallocx(), and ignore all allocator-internal frames. Prior to this change, pprof excluded the specifically blacklisted functions from backtraces, but it left allocator-internal frames intact.
2014-04-23 09:41:15 +08:00
int nframes;
Reduce cpp conditional logic complexity. Convert configuration-related cpp conditional logic to use static constant variables, e.g.: #ifdef JEMALLOC_DEBUG [...] #endif becomes: if (config_debug) { [...] } The advantage is clearer, more concise code. The main disadvantage is that data structures no longer have conditionally defined fields, so they pay the cost of all fields regardless of whether they are used. In practice, this is only a minor concern; config_stats will go away in an upcoming change, and config_prof is the only other major feature that depends on more than a few special-purpose fields.
2012-02-11 12:22:09 +08:00
cassert(config_prof);
Implement allocation profiling and leack checking. Add the --enable-prof and --enable-prof-libunwind configure options. Add the B/b, F/f, I/i, L/l, and U/u JEMALLOC_OPTIONS. Interval-based profile dump triggering is not yet implemented. Add supporting generic code: * Add memory barriers. * Add prn (LCG PRNG). * Add hash (Murmur hash function). * Add ckh (cuckoo hash tables).
2010-02-11 02:37:56 +08:00
assert(bt->len == 0);
assert(bt->vec != NULL);
Simplify backtracing. Simplify backtracing to not ignore any frames, and compensate for this in pprof in order to increase flexibility with respect to function-based refactoring even in the presence of non-deterministic inlining. Modify pprof to blacklist all jemalloc allocation entry points including non-standard ones like mallocx(), and ignore all allocator-internal frames. Prior to this change, pprof excluded the specifically blacklisted functions from backtraces, but it left allocator-internal frames intact.
2014-04-23 09:41:15 +08:00
nframes = unw_backtrace(bt->vec, PROF_BT_MAX);
Update brace style. Add braces around single-line blocks, and remove line breaks before function-opening braces. This resolves #537.
2017-01-16 08:56:30 +08:00
if (nframes <= 0) {
prof_backtrace: use unw_backtrace unw_backtrace: - does internal per-thread caching - doesn't acquire an internal lock
2014-04-22 11:52:35 +08:00
return;
Update brace style. Add braces around single-line blocks, and remove line breaks before function-opening braces. This resolves #537.
2017-01-16 08:56:30 +08:00
}
Simplify backtracing. Simplify backtracing to not ignore any frames, and compensate for this in pprof in order to increase flexibility with respect to function-based refactoring even in the presence of non-deterministic inlining. Modify pprof to blacklist all jemalloc allocation entry points including non-standard ones like mallocx(), and ignore all allocator-internal frames. Prior to this change, pprof excluded the specifically blacklisted functions from backtraces, but it left allocator-internal frames intact.
2014-04-23 09:41:15 +08:00
bt->len = nframes;
Implement allocation profiling and leack checking. Add the --enable-prof and --enable-prof-libunwind configure options. Add the B/b, F/f, I/i, L/l, and U/u JEMALLOC_OPTIONS. Interval-based profile dump triggering is not yet implemented. Add supporting generic code: * Add memory barriers. * Add prn (LCG PRNG). * Add hash (Murmur hash function). * Add ckh (cuckoo hash tables).
2010-02-11 02:37:56 +08:00
}
Reduce cpp conditional logic complexity. Convert configuration-related cpp conditional logic to use static constant variables, e.g.: #ifdef JEMALLOC_DEBUG [...] #endif becomes: if (config_debug) { [...] } The advantage is clearer, more concise code. The main disadvantage is that data structures no longer have conditionally defined fields, so they pay the cost of all fields regardless of whether they are used. In practice, this is only a minor concern; config_stats will go away in an upcoming change, and config_prof is the only other major feature that depends on more than a few special-purpose fields.
2012-02-11 12:22:09 +08:00
#elif (defined(JEMALLOC_PROF_LIBGCC))
Improve backtracing-related configuration. Clean up configuration for backtracing when profiling is enabled, and document the configuration logic in INSTALL. Disable libgcc-based backtracing except on x64 (where it is known to work). Add the --disable-prof-gcc option.
2011-03-16 13:23:12 +08:00
static _Unwind_Reason_Code
Update brace style. Add braces around single-line blocks, and remove line breaks before function-opening braces. This resolves #537.
2017-01-16 08:56:30 +08:00
prof_unwind_init_callback(struct _Unwind_Context *context, void *arg) {
Reduce cpp conditional logic complexity. Convert configuration-related cpp conditional logic to use static constant variables, e.g.: #ifdef JEMALLOC_DEBUG [...] #endif becomes: if (config_debug) { [...] } The advantage is clearer, more concise code. The main disadvantage is that data structures no longer have conditionally defined fields, so they pay the cost of all fields regardless of whether they are used. In practice, this is only a minor concern; config_stats will go away in an upcoming change, and config_prof is the only other major feature that depends on more than a few special-purpose fields.
2012-02-11 12:22:09 +08:00
cassert(config_prof);
Remove extraneous parens around return arguments. This resolves #540.
2017-01-20 10:15:45 +08:00
return _URC_NO_REASON;
Improve backtracing-related configuration. Clean up configuration for backtracing when profiling is enabled, and document the configuration logic in INSTALL. Disable libgcc-based backtracing except on x64 (where it is known to work). Add the --disable-prof-gcc option.
2011-03-16 13:23:12 +08:00
}
static _Unwind_Reason_Code
Update brace style. Add braces around single-line blocks, and remove line breaks before function-opening braces. This resolves #537.
2017-01-16 08:56:30 +08:00
prof_unwind_callback(struct _Unwind_Context *context, void *arg) {
Improve backtracing-related configuration. Clean up configuration for backtracing when profiling is enabled, and document the configuration logic in INSTALL. Disable libgcc-based backtracing except on x64 (where it is known to work). Add the --disable-prof-gcc option.
2011-03-16 13:23:12 +08:00
prof_unwind_data_t *data = (prof_unwind_data_t *)arg;
Simplify backtracing. Simplify backtracing to not ignore any frames, and compensate for this in pprof in order to increase flexibility with respect to function-based refactoring even in the presence of non-deterministic inlining. Modify pprof to blacklist all jemalloc allocation entry points including non-standard ones like mallocx(), and ignore all allocator-internal frames. Prior to this change, pprof excluded the specifically blacklisted functions from backtraces, but it left allocator-internal frames intact.
2014-04-23 09:41:15 +08:00
void *ip;
Improve backtracing-related configuration. Clean up configuration for backtracing when profiling is enabled, and document the configuration logic in INSTALL. Disable libgcc-based backtracing except on x64 (where it is known to work). Add the --disable-prof-gcc option.
2011-03-16 13:23:12 +08:00
Reduce cpp conditional logic complexity. Convert configuration-related cpp conditional logic to use static constant variables, e.g.: #ifdef JEMALLOC_DEBUG [...] #endif becomes: if (config_debug) { [...] } The advantage is clearer, more concise code. The main disadvantage is that data structures no longer have conditionally defined fields, so they pay the cost of all fields regardless of whether they are used. In practice, this is only a minor concern; config_stats will go away in an upcoming change, and config_prof is the only other major feature that depends on more than a few special-purpose fields.
2012-02-11 12:22:09 +08:00
cassert(config_prof);
Simplify backtracing. Simplify backtracing to not ignore any frames, and compensate for this in pprof in order to increase flexibility with respect to function-based refactoring even in the presence of non-deterministic inlining. Modify pprof to blacklist all jemalloc allocation entry points including non-standard ones like mallocx(), and ignore all allocator-internal frames. Prior to this change, pprof excluded the specifically blacklisted functions from backtraces, but it left allocator-internal frames intact.
2014-04-23 09:41:15 +08:00
ip = (void *)_Unwind_GetIP(context);
Update brace style. Add braces around single-line blocks, and remove line breaks before function-opening braces. This resolves #537.
2017-01-16 08:56:30 +08:00
if (ip == NULL) {
Remove extraneous parens around return arguments. This resolves #540.
2017-01-20 10:15:45 +08:00
return _URC_END_OF_STACK;
Update brace style. Add braces around single-line blocks, and remove line breaks before function-opening braces. This resolves #537.
2017-01-16 08:56:30 +08:00
}
Simplify backtracing. Simplify backtracing to not ignore any frames, and compensate for this in pprof in order to increase flexibility with respect to function-based refactoring even in the presence of non-deterministic inlining. Modify pprof to blacklist all jemalloc allocation entry points including non-standard ones like mallocx(), and ignore all allocator-internal frames. Prior to this change, pprof excluded the specifically blacklisted functions from backtraces, but it left allocator-internal frames intact.
2014-04-23 09:41:15 +08:00
data->bt->vec[data->bt->len] = ip;
data->bt->len++;
Update brace style. Add braces around single-line blocks, and remove line breaks before function-opening braces. This resolves #537.
2017-01-16 08:56:30 +08:00
if (data->bt->len == data->max) {
Remove extraneous parens around return arguments. This resolves #540.
2017-01-20 10:15:45 +08:00
return _URC_END_OF_STACK;
Update brace style. Add braces around single-line blocks, and remove line breaks before function-opening braces. This resolves #537.
2017-01-16 08:56:30 +08:00
}
Improve backtracing-related configuration. Clean up configuration for backtracing when profiling is enabled, and document the configuration logic in INSTALL. Disable libgcc-based backtracing except on x64 (where it is known to work). Add the --disable-prof-gcc option.
2011-03-16 13:23:12 +08:00
Remove extraneous parens around return arguments. This resolves #540.
2017-01-20 10:15:45 +08:00
return _URC_NO_REASON;
Improve backtracing-related configuration. Clean up configuration for backtracing when profiling is enabled, and document the configuration logic in INSTALL. Disable libgcc-based backtracing except on x64 (where it is known to work). Add the --disable-prof-gcc option.
2011-03-16 13:23:12 +08:00
}
Fix prof_backtrace() reentrancy level
2019-09-04 08:11:06 +08:00
static void
prof_backtrace_impl(prof_bt_t *bt) {
Simplify backtracing. Simplify backtracing to not ignore any frames, and compensate for this in pprof in order to increase flexibility with respect to function-based refactoring even in the presence of non-deterministic inlining. Modify pprof to blacklist all jemalloc allocation entry points including non-standard ones like mallocx(), and ignore all allocator-internal frames. Prior to this change, pprof excluded the specifically blacklisted functions from backtraces, but it left allocator-internal frames intact.
2014-04-23 09:41:15 +08:00
prof_unwind_data_t data = {bt, PROF_BT_MAX};
Improve backtracing-related configuration. Clean up configuration for backtracing when profiling is enabled, and document the configuration logic in INSTALL. Disable libgcc-based backtracing except on x64 (where it is known to work). Add the --disable-prof-gcc option.
2011-03-16 13:23:12 +08:00
Reduce cpp conditional logic complexity. Convert configuration-related cpp conditional logic to use static constant variables, e.g.: #ifdef JEMALLOC_DEBUG [...] #endif becomes: if (config_debug) { [...] } The advantage is clearer, more concise code. The main disadvantage is that data structures no longer have conditionally defined fields, so they pay the cost of all fields regardless of whether they are used. In practice, this is only a minor concern; config_stats will go away in an upcoming change, and config_prof is the only other major feature that depends on more than a few special-purpose fields.
2012-02-11 12:22:09 +08:00
cassert(config_prof);
Improve backtracing-related configuration. Clean up configuration for backtracing when profiling is enabled, and document the configuration logic in INSTALL. Disable libgcc-based backtracing except on x64 (where it is known to work). Add the --disable-prof-gcc option.
2011-03-16 13:23:12 +08:00
_Unwind_Backtrace(prof_unwind_callback, &data);
}
Reduce cpp conditional logic complexity. Convert configuration-related cpp conditional logic to use static constant variables, e.g.: #ifdef JEMALLOC_DEBUG [...] #endif becomes: if (config_debug) { [...] } The advantage is clearer, more concise code. The main disadvantage is that data structures no longer have conditionally defined fields, so they pay the cost of all fields regardless of whether they are used. In practice, this is only a minor concern; config_stats will go away in an upcoming change, and config_prof is the only other major feature that depends on more than a few special-purpose fields.
2012-02-11 12:22:09 +08:00
#elif (defined(JEMALLOC_PROF_GCC))
Fix prof_backtrace() reentrancy level
2019-09-04 08:11:06 +08:00
static void
prof_backtrace_impl(prof_bt_t *bt) {
Replace tabs following #define with spaces. This resolves #564.
2017-01-20 13:41:41 +08:00
#define BT_FRAME(i) \
Simplify backtracing. Simplify backtracing to not ignore any frames, and compensate for this in pprof in order to increase flexibility with respect to function-based refactoring even in the presence of non-deterministic inlining. Modify pprof to blacklist all jemalloc allocation entry points including non-standard ones like mallocx(), and ignore all allocator-internal frames. Prior to this change, pprof excluded the specifically blacklisted functions from backtraces, but it left allocator-internal frames intact.
2014-04-23 09:41:15 +08:00
if ((i) < PROF_BT_MAX) { \
Implement allocation profiling and leack checking. Add the --enable-prof and --enable-prof-libunwind configure options. Add the B/b, F/f, I/i, L/l, and U/u JEMALLOC_OPTIONS. Interval-based profile dump triggering is not yet implemented. Add supporting generic code: * Add memory barriers. * Add prn (LCG PRNG). * Add hash (Murmur hash function). * Add ckh (cuckoo hash tables).
2010-02-11 02:37:56 +08:00
void *p; \
Update brace style. Add braces around single-line blocks, and remove line breaks before function-opening braces. This resolves #537.
2017-01-16 08:56:30 +08:00
if (__builtin_frame_address(i) == 0) { \
Various heap profiling improvements. Add the --disable-prof-libgcc configure option, and add backtracing based on libgcc, which is used by default. Fix a bug in hash(). Fix various configuration-dependent compilation errors.
2010-02-11 10:15:53 +08:00
return; \
Update brace style. Add braces around single-line blocks, and remove line breaks before function-opening braces. This resolves #537.
2017-01-16 08:56:30 +08:00
} \
Implement allocation profiling and leack checking. Add the --enable-prof and --enable-prof-libunwind configure options. Add the B/b, F/f, I/i, L/l, and U/u JEMALLOC_OPTIONS. Interval-based profile dump triggering is not yet implemented. Add supporting generic code: * Add memory barriers. * Add prn (LCG PRNG). * Add hash (Murmur hash function). * Add ckh (cuckoo hash tables).
2010-02-11 02:37:56 +08:00
p = __builtin_return_address(i); \
Update brace style. Add braces around single-line blocks, and remove line breaks before function-opening braces. This resolves #537.
2017-01-16 08:56:30 +08:00
if (p == NULL) { \
Various heap profiling improvements. Add the --disable-prof-libgcc configure option, and add backtracing based on libgcc, which is used by default. Fix a bug in hash(). Fix various configuration-dependent compilation errors.
2010-02-11 10:15:53 +08:00
return; \
Update brace style. Add braces around single-line blocks, and remove line breaks before function-opening braces. This resolves #537.
2017-01-16 08:56:30 +08:00
} \
Simplify backtracing. Simplify backtracing to not ignore any frames, and compensate for this in pprof in order to increase flexibility with respect to function-based refactoring even in the presence of non-deterministic inlining. Modify pprof to blacklist all jemalloc allocation entry points including non-standard ones like mallocx(), and ignore all allocator-internal frames. Prior to this change, pprof excluded the specifically blacklisted functions from backtraces, but it left allocator-internal frames intact.
2014-04-23 09:41:15 +08:00
bt->vec[(i)] = p; \
bt->len = (i) + 1; \
Update brace style. Add braces around single-line blocks, and remove line breaks before function-opening braces. This resolves #537.
2017-01-16 08:56:30 +08:00
} else { \
return; \
}
Implement allocation profiling and leack checking. Add the --enable-prof and --enable-prof-libunwind configure options. Add the B/b, F/f, I/i, L/l, and U/u JEMALLOC_OPTIONS. Interval-based profile dump triggering is not yet implemented. Add supporting generic code: * Add memory barriers. * Add prn (LCG PRNG). * Add hash (Murmur hash function). * Add ckh (cuckoo hash tables).
2010-02-11 02:37:56 +08:00
Reduce cpp conditional logic complexity. Convert configuration-related cpp conditional logic to use static constant variables, e.g.: #ifdef JEMALLOC_DEBUG [...] #endif becomes: if (config_debug) { [...] } The advantage is clearer, more concise code. The main disadvantage is that data structures no longer have conditionally defined fields, so they pay the cost of all fields regardless of whether they are used. In practice, this is only a minor concern; config_stats will go away in an upcoming change, and config_prof is the only other major feature that depends on more than a few special-purpose fields.
2012-02-11 12:22:09 +08:00
cassert(config_prof);
Implement allocation profiling and leack checking. Add the --enable-prof and --enable-prof-libunwind configure options. Add the B/b, F/f, I/i, L/l, and U/u JEMALLOC_OPTIONS. Interval-based profile dump triggering is not yet implemented. Add supporting generic code: * Add memory barriers. * Add prn (LCG PRNG). * Add hash (Murmur hash function). * Add ckh (cuckoo hash tables).
2010-02-11 02:37:56 +08:00
BT_FRAME(0)
BT_FRAME(1)
BT_FRAME(2)
BT_FRAME(3)
BT_FRAME(4)
BT_FRAME(5)
BT_FRAME(6)
BT_FRAME(7)
BT_FRAME(8)
BT_FRAME(9)
BT_FRAME(10)
BT_FRAME(11)
BT_FRAME(12)
BT_FRAME(13)
BT_FRAME(14)
BT_FRAME(15)
BT_FRAME(16)
BT_FRAME(17)
BT_FRAME(18)
BT_FRAME(19)
BT_FRAME(20)
BT_FRAME(21)
BT_FRAME(22)
BT_FRAME(23)
BT_FRAME(24)
BT_FRAME(25)
BT_FRAME(26)
BT_FRAME(27)
BT_FRAME(28)
BT_FRAME(29)
BT_FRAME(30)
BT_FRAME(31)
BT_FRAME(32)
BT_FRAME(33)
BT_FRAME(34)
BT_FRAME(35)
BT_FRAME(36)
BT_FRAME(37)
BT_FRAME(38)
BT_FRAME(39)
BT_FRAME(40)
BT_FRAME(41)
BT_FRAME(42)
BT_FRAME(43)
BT_FRAME(44)
BT_FRAME(45)
BT_FRAME(46)
BT_FRAME(47)
BT_FRAME(48)
BT_FRAME(49)
BT_FRAME(50)
BT_FRAME(51)
BT_FRAME(52)
BT_FRAME(53)
BT_FRAME(54)
BT_FRAME(55)
BT_FRAME(56)
BT_FRAME(57)
BT_FRAME(58)
BT_FRAME(59)
BT_FRAME(60)
BT_FRAME(61)
BT_FRAME(62)
BT_FRAME(63)
BT_FRAME(64)
BT_FRAME(65)
BT_FRAME(66)
BT_FRAME(67)
BT_FRAME(68)
BT_FRAME(69)
BT_FRAME(70)
BT_FRAME(71)
BT_FRAME(72)
BT_FRAME(73)
BT_FRAME(74)
BT_FRAME(75)
BT_FRAME(76)
BT_FRAME(77)
BT_FRAME(78)
BT_FRAME(79)
BT_FRAME(80)
BT_FRAME(81)
BT_FRAME(82)
BT_FRAME(83)
BT_FRAME(84)
BT_FRAME(85)
BT_FRAME(86)
BT_FRAME(87)
BT_FRAME(88)
BT_FRAME(89)
BT_FRAME(90)
BT_FRAME(91)
BT_FRAME(92)
BT_FRAME(93)
BT_FRAME(94)
BT_FRAME(95)
BT_FRAME(96)
BT_FRAME(97)
BT_FRAME(98)
BT_FRAME(99)
BT_FRAME(100)
BT_FRAME(101)
BT_FRAME(102)
BT_FRAME(103)
BT_FRAME(104)
BT_FRAME(105)
BT_FRAME(106)
BT_FRAME(107)
BT_FRAME(108)
BT_FRAME(109)
BT_FRAME(110)
BT_FRAME(111)
BT_FRAME(112)
BT_FRAME(113)
BT_FRAME(114)
BT_FRAME(115)
BT_FRAME(116)
BT_FRAME(117)
BT_FRAME(118)
BT_FRAME(119)
BT_FRAME(120)
BT_FRAME(121)
BT_FRAME(122)
BT_FRAME(123)
BT_FRAME(124)
BT_FRAME(125)
BT_FRAME(126)
BT_FRAME(127)
#undef BT_FRAME
}
Reduce cpp conditional logic complexity. Convert configuration-related cpp conditional logic to use static constant variables, e.g.: #ifdef JEMALLOC_DEBUG [...] #endif becomes: if (config_debug) { [...] } The advantage is clearer, more concise code. The main disadvantage is that data structures no longer have conditionally defined fields, so they pay the cost of all fields regardless of whether they are used. In practice, this is only a minor concern; config_stats will go away in an upcoming change, and config_prof is the only other major feature that depends on more than a few special-purpose fields.
2012-02-11 12:22:09 +08:00
#else
Fix prof_backtrace() reentrancy level
2019-09-04 08:11:06 +08:00
static void
prof_backtrace_impl(prof_bt_t *bt) {
Reduce cpp conditional logic complexity. Convert configuration-related cpp conditional logic to use static constant variables, e.g.: #ifdef JEMALLOC_DEBUG [...] #endif becomes: if (config_debug) { [...] } The advantage is clearer, more concise code. The main disadvantage is that data structures no longer have conditionally defined fields, so they pay the cost of all fields regardless of whether they are used. In practice, this is only a minor concern; config_stats will go away in an upcoming change, and config_prof is the only other major feature that depends on more than a few special-purpose fields.
2012-02-11 12:22:09 +08:00
cassert(config_prof);
Prefer not_reached() over assert(false) where appropriate.
2013-10-22 05:56:27 +08:00
not_reached();
Reduce cpp conditional logic complexity. Convert configuration-related cpp conditional logic to use static constant variables, e.g.: #ifdef JEMALLOC_DEBUG [...] #endif becomes: if (config_debug) { [...] } The advantage is clearer, more concise code. The main disadvantage is that data structures no longer have conditionally defined fields, so they pay the cost of all fields regardless of whether they are used. In practice, this is only a minor concern; config_stats will go away in an upcoming change, and config_prof is the only other major feature that depends on more than a few special-purpose fields.
2012-02-11 12:22:09 +08:00
}
Implement allocation profiling and leack checking. Add the --enable-prof and --enable-prof-libunwind configure options. Add the B/b, F/f, I/i, L/l, and U/u JEMALLOC_OPTIONS. Interval-based profile dump triggering is not yet implemented. Add supporting generic code: * Add memory barriers. * Add prn (LCG PRNG). * Add hash (Murmur hash function). * Add ckh (cuckoo hash tables).
2010-02-11 02:37:56 +08:00
#endif
Fix prof_backtrace() reentrancy level
2019-09-04 08:11:06 +08:00
void
Pass tsd down to prof_backtrace()
2019-09-05 00:24:34 +08:00
prof_backtrace(tsd_t *tsd, prof_bt_t *bt) {
Fix prof_backtrace() reentrancy level
2019-09-04 08:11:06 +08:00
cassert(config_prof);
pre_reentrancy(tsd, NULL);
prof_backtrace_impl(bt);
post_reentrancy(tsd);
}
Refactor profiling Refactored core profiling codebase into two logical parts: (a) `prof_data.c`: core internal data structure managing & dumping; (b) `prof.c`: mutexes & outward-facing APIs. Some internal functions had to be exposed out, but there are not that many of them if the modularization is (hopefully) clean enough.
2019-07-18 06:52:50 +08:00
malloc_mutex_t *
Update brace style. Add braces around single-line blocks, and remove line breaks before function-opening braces. This resolves #537.
2017-01-16 08:56:30 +08:00
prof_gctx_mutex_choose(void) {
Make prof's cum_gctx a C11-style atomic
2017-04-05 09:36:45 +08:00
unsigned ngctxs = atomic_fetch_add_u(&cum_gctxs, 1, ATOMIC_RELAXED);
Implement per thread heap profiling. Rename data structures (prof_thr_cnt_t-->prof_tctx_t, prof_ctx_t-->prof_gctx_t), and convert to storing a prof_tctx_t for sampled objects. Convert PROF_ALLOC_PREP() to prof_alloc_prep(), since precise backtrace depth within jemalloc functions is no longer an issue (pprof prunes irrelevant frames). Implement mallctl's: - prof.reset implements full sample data reset, and optional change of sample interval. - prof.lg_sample reads the current sample interval (opt.lg_prof_sample was the permanent source of truth prior to prof.reset). - thread.prof.name provides naming capability for threads within heap profile dumps. - thread.prof.active makes it possible to activate/deactivate heap profiling for individual threads. Modify the heap dump files to contain per thread heap profile data. This change is incompatible with the existing pprof, which will require enhancements to read and process the enriched data.
2014-08-19 07:22:13 +08:00
Remove extraneous parens around return arguments. This resolves #540.
2017-01-20 10:15:45 +08:00
return &gctx_locks[(ngctxs - 1) % PROF_NCTX_LOCKS];
Implement per thread heap profiling. Rename data structures (prof_thr_cnt_t-->prof_tctx_t, prof_ctx_t-->prof_gctx_t), and convert to storing a prof_tctx_t for sampled objects. Convert PROF_ALLOC_PREP() to prof_alloc_prep(), since precise backtrace depth within jemalloc functions is no longer an issue (pprof prunes irrelevant frames). Implement mallctl's: - prof.reset implements full sample data reset, and optional change of sample interval. - prof.lg_sample reads the current sample interval (opt.lg_prof_sample was the permanent source of truth prior to prof.reset). - thread.prof.name provides naming capability for threads within heap profile dumps. - thread.prof.active makes it possible to activate/deactivate heap profiling for individual threads. Modify the heap dump files to contain per thread heap profile data. This change is incompatible with the existing pprof, which will require enhancements to read and process the enriched data.
2014-08-19 07:22:13 +08:00
}
Refactor profiling Refactored core profiling codebase into two logical parts: (a) `prof_data.c`: core internal data structure managing & dumping; (b) `prof.c`: mutexes & outward-facing APIs. Some internal functions had to be exposed out, but there are not that many of them if the modularization is (hopefully) clean enough.
2019-07-18 06:52:50 +08:00
malloc_mutex_t *
Update brace style. Add braces around single-line blocks, and remove line breaks before function-opening braces. This resolves #537.
2017-01-16 08:56:30 +08:00
prof_tdata_mutex_choose(uint64_t thr_uid) {
Remove extraneous parens around return arguments. This resolves #540.
2017-01-20 10:15:45 +08:00
return &tdata_locks[thr_uid % PROF_NTDATA_LOCKS];
Refactor prof_dump() to reduce contention. Refactor prof_dump() to use a two pass algorithm, and prof_leave() prior to the second pass. This avoids write(2) system calls while holding critical prof resources. Fix prof_dump() to close the dump file descriptor for all relevant error paths. Minimize the size of prof-related static buffers when prof is disabled. This saves roughly 65 KiB of application memory for non-prof builds. Refactor prof_ctx_init() out of prof_lookup_global().
2014-01-17 05:23:56 +08:00
}
Scale leak report summary according to sampling probability. This makes the numbers reported in the leak report summary closely match those reported by jeprof. This resolves #356.
2016-05-05 03:14:36 +08:00
/*
* The bodies of this function and prof_leakcheck() are compiled out unless heap
* profiling is enabled, so that it is possible to compile jemalloc with
* floating point support completely disabled. Avoiding floating point code is
* important on memory-constrained systems, but it also enables a workaround for
* versions of glibc that don't properly save/restore floating point registers
* during dynamic lazy symbol loading (which internally calls into whatever
* malloc implementation happens to be integrated into the application). Note
* that some compilers (e.g. gcc 4.8) may use floating point registers for fast
* memory moves, so jemalloc must be compiled with such optimizations disabled
* (e.g.
* -mno-sse) in order for the workaround to be complete.
*/
refactor profiling. only use a bytes till next sample variable.
2014-04-16 04:47:13 +08:00
void
Update brace style. Add braces around single-line blocks, and remove line breaks before function-opening braces. This resolves #537.
2017-01-16 08:56:30 +08:00
prof_sample_threshold_update(prof_tdata_t *tdata) {
refactor profiling. only use a bytes till next sample variable.
2014-04-16 04:47:13 +08:00
#ifdef JEMALLOC_PROF
Update brace style. Add braces around single-line blocks, and remove line breaks before function-opening braces. This resolves #537.
2017-01-16 08:56:30 +08:00
if (!config_prof) {
refactor profiling. only use a bytes till next sample variable.
2014-04-16 04:47:13 +08:00
return;
Update brace style. Add braces around single-line blocks, and remove line breaks before function-opening braces. This resolves #537.
2017-01-16 08:56:30 +08:00
}
refactor profiling. only use a bytes till next sample variable.
2014-04-16 04:47:13 +08:00
Implement per thread heap profiling. Rename data structures (prof_thr_cnt_t-->prof_tctx_t, prof_ctx_t-->prof_gctx_t), and convert to storing a prof_tctx_t for sampled objects. Convert PROF_ALLOC_PREP() to prof_alloc_prep(), since precise backtrace depth within jemalloc functions is no longer an issue (pprof prunes irrelevant frames). Implement mallctl's: - prof.reset implements full sample data reset, and optional change of sample interval. - prof.lg_sample reads the current sample interval (opt.lg_prof_sample was the permanent source of truth prior to prof.reset). - thread.prof.name provides naming capability for threads within heap profile dumps. - thread.prof.active makes it possible to activate/deactivate heap profiling for individual threads. Modify the heap dump files to contain per thread heap profile data. This change is incompatible with the existing pprof, which will require enhancements to read and process the enriched data.
2014-08-19 07:22:13 +08:00
if (lg_prof_sample == 0) {
move bytes until sample to tsd. Fastpath allocation does not need to load tdata now, avoiding several branches.
2018-10-10 01:59:02 +08:00
tsd_bytes_until_sample_set(tsd_fetch(), 0);
refactor profiling. only use a bytes till next sample variable.
2014-04-16 04:47:13 +08:00
return;
}
/*
Implement per thread heap profiling. Rename data structures (prof_thr_cnt_t-->prof_tctx_t, prof_ctx_t-->prof_gctx_t), and convert to storing a prof_tctx_t for sampled objects. Convert PROF_ALLOC_PREP() to prof_alloc_prep(), since precise backtrace depth within jemalloc functions is no longer an issue (pprof prunes irrelevant frames). Implement mallctl's: - prof.reset implements full sample data reset, and optional change of sample interval. - prof.lg_sample reads the current sample interval (opt.lg_prof_sample was the permanent source of truth prior to prof.reset). - thread.prof.name provides naming capability for threads within heap profile dumps. - thread.prof.active makes it possible to activate/deactivate heap profiling for individual threads. Modify the heap dump files to contain per thread heap profile data. This change is incompatible with the existing pprof, which will require enhancements to read and process the enriched data.
2014-08-19 07:22:13 +08:00
* Compute sample interval as a geometrically distributed random
* variable with mean (2^lg_prof_sample).
refactor profiling. only use a bytes till next sample variable.
2014-04-16 04:47:13 +08:00
*
Implement per thread heap profiling. Rename data structures (prof_thr_cnt_t-->prof_tctx_t, prof_ctx_t-->prof_gctx_t), and convert to storing a prof_tctx_t for sampled objects. Convert PROF_ALLOC_PREP() to prof_alloc_prep(), since precise backtrace depth within jemalloc functions is no longer an issue (pprof prunes irrelevant frames). Implement mallctl's: - prof.reset implements full sample data reset, and optional change of sample interval. - prof.lg_sample reads the current sample interval (opt.lg_prof_sample was the permanent source of truth prior to prof.reset). - thread.prof.name provides naming capability for threads within heap profile dumps. - thread.prof.active makes it possible to activate/deactivate heap profiling for individual threads. Modify the heap dump files to contain per thread heap profile data. This change is incompatible with the existing pprof, which will require enhancements to read and process the enriched data.
2014-08-19 07:22:13 +08:00
* __ __
* | log(u) | 1
* tdata->bytes_until_sample = | -------- |, where p = ---------------
* | log(1-p) | lg_prof_sample
* 2
refactor profiling. only use a bytes till next sample variable.
2014-04-16 04:47:13 +08:00
*
* For more information on the math, see:
*
* Non-Uniform Random Variate Generation
* Luc Devroye
* Springer-Verlag, New York, 1986
* pp 500
* (http://luc.devroye.org/rnbookindex.html)
*/
move bytes until sample to tsd. Fastpath allocation does not need to load tdata now, avoiding several branches.
2018-10-10 01:59:02 +08:00
uint64_t r = prng_lg_range_u64(&tdata->prng_state, 53);
double u = (double)r * (1.0/9007199254740992.0L);
uint64_t bytes_until_sample = (uint64_t)(log(u) /
Implement per thread heap profiling. Rename data structures (prof_thr_cnt_t-->prof_tctx_t, prof_ctx_t-->prof_gctx_t), and convert to storing a prof_tctx_t for sampled objects. Convert PROF_ALLOC_PREP() to prof_alloc_prep(), since precise backtrace depth within jemalloc functions is no longer an issue (pprof prunes irrelevant frames). Implement mallctl's: - prof.reset implements full sample data reset, and optional change of sample interval. - prof.lg_sample reads the current sample interval (opt.lg_prof_sample was the permanent source of truth prior to prof.reset). - thread.prof.name provides naming capability for threads within heap profile dumps. - thread.prof.active makes it possible to activate/deactivate heap profiling for individual threads. Modify the heap dump files to contain per thread heap profile data. This change is incompatible with the existing pprof, which will require enhancements to read and process the enriched data.
2014-08-19 07:22:13 +08:00
log(1.0 - (1.0 / (double)((uint64_t)1U << lg_prof_sample))))
refactor profiling. only use a bytes till next sample variable.
2014-04-16 04:47:13 +08:00
+ (uint64_t)1U;
restrict bytes_until_sample to int64_t. This allows optimal asm generation of sub bytes_until_sample, usize; je; for x86 arch. Subtraction is unconditional, and only flags are checked for the jump, no extra compare is necessary. This also reduces register pressure.
2018-10-10 02:25:36 +08:00
if (bytes_until_sample > SSIZE_MAX) {
bytes_until_sample = SSIZE_MAX;
}
move bytes until sample to tsd. Fastpath allocation does not need to load tdata now, avoiding several branches.
2018-10-10 01:59:02 +08:00
tsd_bytes_until_sample_set(tsd_fetch(), bytes_until_sample);
refactor profiling. only use a bytes till next sample variable.
2014-04-16 04:47:13 +08:00
#endif
}
Refactor profiling Refactored core profiling codebase into two logical parts: (a) `prof_data.c`: core internal data structure managing & dumping; (b) `prof.c`: mutexes & outward-facing APIs. Some internal functions had to be exposed out, but there are not that many of them if the modularization is (hopefully) clean enough.
2019-07-18 06:52:50 +08:00
int
prof_getpid(void) {
#ifdef _WIN32
return GetCurrentProcessId();
#else
return getpid();
Revert "Refactor profiling" This reverts commit 0b462407ae84a62b3c097f0e9f18df487a47d9a7.
2019-07-30 05:09:20 +08:00
#endif
Refactor profiling Refactored core profiling codebase into two logical parts: (a) `prof_data.c`: core internal data structure managing & dumping; (b) `prof.c`: mutexes & outward-facing APIs. Some internal functions had to be exposed out, but there are not that many of them if the modularization is (hopefully) clean enough.
2019-07-18 06:52:50 +08:00
}
Refactor prof_dump() to reduce contention. Refactor prof_dump() to use a two pass algorithm, and prof_leave() prior to the second pass. This avoids write(2) system calls while holding critical prof resources. Fix prof_dump() to close the dump file descriptor for all relevant error paths. Minimize the size of prof-related static buffers when prof is disabled. This saves roughly 65 KiB of application memory for non-prof builds. Refactor prof_ctx_init() out of prof_lookup_global().
2014-01-17 05:23:56 +08:00
Add "prof.dump_prefix" to override filename prefixes for dumps.
2019-09-10 11:18:41 +08:00
static const char *
prof_dump_prefix_get(tsdn_t* tsdn) {
malloc_mutex_assert_owner(tsdn, &prof_dump_filename_mtx);
return prof_dump_prefix == NULL ? opt_prof_prefix : prof_dump_prefix;
}
static bool
prof_dump_prefix_is_empty(tsdn_t *tsdn) {
malloc_mutex_lock(tsdn, &prof_dump_filename_mtx);
bool ret = (prof_dump_prefix_get(tsdn)[0] == '\0');
malloc_mutex_unlock(tsdn, &prof_dump_filename_mtx);
return ret;
}
Refactor profiling Refactored core profiling codebase into two logical parts: (a) `prof_data.c`: core internal data structure managing & dumping; (b) `prof.c`: mutexes & outward-facing APIs. Some internal functions had to be exposed out, but there are not that many of them if the modularization is (hopefully) clean enough.
2019-07-18 06:52:50 +08:00
#define DUMP_FILENAME_BUFSIZE (PATH_MAX + 1)
#define VSEQ_INVALID UINT64_C(0xffffffffffffffff)
static void
Add "prof.dump_prefix" to override filename prefixes for dumps.
2019-09-10 11:18:41 +08:00
prof_dump_filename(tsd_t *tsd, char *filename, char v, uint64_t vseq) {
Refactor profiling Refactored core profiling codebase into two logical parts: (a) `prof_data.c`: core internal data structure managing & dumping; (b) `prof.c`: mutexes & outward-facing APIs. Some internal functions had to be exposed out, but there are not that many of them if the modularization is (hopefully) clean enough.
2019-07-18 06:52:50 +08:00
cassert(config_prof);
Reduce cpp conditional logic complexity. Convert configuration-related cpp conditional logic to use static constant variables, e.g.: #ifdef JEMALLOC_DEBUG [...] #endif becomes: if (config_debug) { [...] } The advantage is clearer, more concise code. The main disadvantage is that data structures no longer have conditionally defined fields, so they pay the cost of all fields regardless of whether they are used. In practice, this is only a minor concern; config_stats will go away in an upcoming change, and config_prof is the only other major feature that depends on more than a few special-purpose fields.
2012-02-11 12:22:09 +08:00
Add "prof.dump_prefix" to override filename prefixes for dumps.
2019-09-10 11:18:41 +08:00
assert(tsd_reentrancy_level_get(tsd) == 0);
const char *prof_prefix = prof_dump_prefix_get(tsd_tsdn(tsd));
Refactor profiling Refactored core profiling codebase into two logical parts: (a) `prof_data.c`: core internal data structure managing & dumping; (b) `prof.c`: mutexes & outward-facing APIs. Some internal functions had to be exposed out, but there are not that many of them if the modularization is (hopefully) clean enough.
2019-07-18 06:52:50 +08:00
if (vseq != VSEQ_INVALID) {
/* "<prefix>.<pid>.<seq>.v<vseq>.heap" */
malloc_snprintf(filename, DUMP_FILENAME_BUFSIZE,
"%s.%d.%"FMTu64".%c%"FMTu64".heap",
Add "prof.dump_prefix" to override filename prefixes for dumps.
2019-09-10 11:18:41 +08:00
prof_prefix, prof_getpid(), prof_dump_seq, v, vseq);
Refactor profiling Refactored core profiling codebase into two logical parts: (a) `prof_data.c`: core internal data structure managing & dumping; (b) `prof.c`: mutexes & outward-facing APIs. Some internal functions had to be exposed out, but there are not that many of them if the modularization is (hopefully) clean enough.
2019-07-18 06:52:50 +08:00
} else {
/* "<prefix>.<pid>.<seq>.<v>.heap" */
malloc_snprintf(filename, DUMP_FILENAME_BUFSIZE,
"%s.%d.%"FMTu64".%c.heap",
Add "prof.dump_prefix" to override filename prefixes for dumps.
2019-09-10 11:18:41 +08:00
prof_prefix, prof_getpid(), prof_dump_seq, v);
Implement allocation profiling and leack checking. Add the --enable-prof and --enable-prof-libunwind configure options. Add the B/b, F/f, I/i, L/l, and U/u JEMALLOC_OPTIONS. Interval-based profile dump triggering is not yet implemented. Add supporting generic code: * Add memory barriers. * Add prn (LCG PRNG). * Add hash (Murmur hash function). * Add ckh (cuckoo hash tables).
2010-02-11 02:37:56 +08:00
}
Refactor profiling Refactored core profiling codebase into two logical parts: (a) `prof_data.c`: core internal data structure managing & dumping; (b) `prof.c`: mutexes & outward-facing APIs. Some internal functions had to be exposed out, but there are not that many of them if the modularization is (hopefully) clean enough.
2019-07-18 06:52:50 +08:00
prof_dump_seq++;
Implement allocation profiling and leack checking. Add the --enable-prof and --enable-prof-libunwind configure options. Add the B/b, F/f, I/i, L/l, and U/u JEMALLOC_OPTIONS. Interval-based profile dump triggering is not yet implemented. Add supporting generic code: * Add memory barriers. * Add prn (LCG PRNG). * Add hash (Murmur hash function). * Add ckh (cuckoo hash tables).
2010-02-11 02:37:56 +08:00
}
Add "prof.dump_prefix" to override filename prefixes for dumps.
2019-09-10 11:18:41 +08:00
void
prof_get_default_filename(tsdn_t *tsdn, char *filename, uint64_t ind) {
malloc_mutex_lock(tsdn, &prof_dump_filename_mtx);
malloc_snprintf(filename, PROF_DUMP_FILENAME_LEN,
"%s.%d.%"FMTu64".json", prof_dump_prefix_get(tsdn), prof_getpid(),
ind);
malloc_mutex_unlock(tsdn, &prof_dump_filename_mtx);
}
Refactor profiling Refactored core profiling codebase into two logical parts: (a) `prof_data.c`: core internal data structure managing & dumping; (b) `prof.c`: mutexes & outward-facing APIs. Some internal functions had to be exposed out, but there are not that many of them if the modularization is (hopefully) clean enough.
2019-07-18 06:52:50 +08:00
static void
prof_fdump(void) {
tsd_t *tsd;
char filename[DUMP_FILENAME_BUFSIZE];
Add witness, a simple online locking validator. This resolves #358.
2016-04-14 14:36:15 +08:00
Refactor profiling Refactored core profiling codebase into two logical parts: (a) `prof_data.c`: core internal data structure managing & dumping; (b) `prof.c`: mutexes & outward-facing APIs. Some internal functions had to be exposed out, but there are not that many of them if the modularization is (hopefully) clean enough.
2019-07-18 06:52:50 +08:00
cassert(config_prof);
Refactor profiling Refactored core profiling codebase into two logical parts: (a) `prof_data.c`: core internal data structure managing & dumping; (b) `prof.c`: mutexes & outward-facing APIs. Some internal functions had to be exposed out, but there are not that many of them if the modularization is (hopefully) clean enough.
2019-07-18 06:52:50 +08:00
assert(opt_prof_final);
Fix a heap profiling regression. Add the prof_tctx_state_destroying transitionary state to fix a race between a thread destroying a tctx and another thread creating a new equivalent tctx. This regression was introduced by 602c8e0971160e4b85b08b16cf8a2375aa24bc04 (Implement per thread heap profiling.).
2015-03-15 05:01:35 +08:00
Refactor profiling Refactored core profiling codebase into two logical parts: (a) `prof_data.c`: core internal data structure managing & dumping; (b) `prof.c`: mutexes & outward-facing APIs. Some internal functions had to be exposed out, but there are not that many of them if the modularization is (hopefully) clean enough.
2019-07-18 06:52:50 +08:00
if (!prof_booted) {
return;
Implement interval-based heap profile dumping. Add mallctl interfaces for profiling parameters. Fix a file descriptor leak in heap profile dumping.
2010-02-12 05:19:21 +08:00
}
Refactor profiling Refactored core profiling codebase into two logical parts: (a) `prof_data.c`: core internal data structure managing & dumping; (b) `prof.c`: mutexes & outward-facing APIs. Some internal functions had to be exposed out, but there are not that many of them if the modularization is (hopefully) clean enough.
2019-07-18 06:52:50 +08:00
tsd = tsd_fetch();
assert(tsd_reentrancy_level_get(tsd) == 0);
Add "prof.dump_prefix" to override filename prefixes for dumps.
2019-09-10 11:18:41 +08:00
assert(!prof_dump_prefix_is_empty(tsd_tsdn(tsd)));
Revert "Refactor profiling" This reverts commit 0b462407ae84a62b3c097f0e9f18df487a47d9a7.
2019-07-30 05:09:20 +08:00
Rename "prof_dump_seq_mtx" to "prof_dump_filename_mtx".
2019-09-10 11:04:18 +08:00
malloc_mutex_lock(tsd_tsdn(tsd), &prof_dump_filename_mtx);
Add "prof.dump_prefix" to override filename prefixes for dumps.
2019-09-10 11:18:41 +08:00
prof_dump_filename(tsd, filename, 'f', VSEQ_INVALID);
Rename "prof_dump_seq_mtx" to "prof_dump_filename_mtx".
2019-09-10 11:04:18 +08:00
malloc_mutex_unlock(tsd_tsdn(tsd), &prof_dump_filename_mtx);
Refactor profiling Refactored core profiling codebase into two logical parts: (a) `prof_data.c`: core internal data structure managing & dumping; (b) `prof.c`: mutexes & outward-facing APIs. Some internal functions had to be exposed out, but there are not that many of them if the modularization is (hopefully) clean enough.
2019-07-18 06:52:50 +08:00
prof_dump(tsd, false, filename, opt_prof_leak);
Revert "Refactor profiling" This reverts commit 0b462407ae84a62b3c097f0e9f18df487a47d9a7.
2019-07-30 05:09:20 +08:00
}
Refactor profiling Refactored core profiling codebase into two logical parts: (a) `prof_data.c`: core internal data structure managing & dumping; (b) `prof.c`: mutexes & outward-facing APIs. Some internal functions had to be exposed out, but there are not that many of them if the modularization is (hopefully) clean enough.
2019-07-18 06:52:50 +08:00
bool
prof_accum_init(tsdn_t *tsdn, prof_accum_t *prof_accum) {
Revert "Refactor profiling" This reverts commit 0b462407ae84a62b3c097f0e9f18df487a47d9a7.
2019-07-30 05:09:20 +08:00
cassert(config_prof);
Refactor profiling Refactored core profiling codebase into two logical parts: (a) `prof_data.c`: core internal data structure managing & dumping; (b) `prof.c`: mutexes & outward-facing APIs. Some internal functions had to be exposed out, but there are not that many of them if the modularization is (hopefully) clean enough.
2019-07-18 06:52:50 +08:00
#ifndef JEMALLOC_ATOMIC_U64
if (malloc_mutex_init(&prof_accum->mtx, "prof_accum",
WITNESS_RANK_PROF_ACCUM, malloc_mutex_rank_exclusive)) {
return true;
Revert "Refactor profiling" This reverts commit 0b462407ae84a62b3c097f0e9f18df487a47d9a7.
2019-07-30 05:09:20 +08:00
}
Refactor profiling Refactored core profiling codebase into two logical parts: (a) `prof_data.c`: core internal data structure managing & dumping; (b) `prof.c`: mutexes & outward-facing APIs. Some internal functions had to be exposed out, but there are not that many of them if the modularization is (hopefully) clean enough.
2019-07-18 06:52:50 +08:00
prof_accum->accumbytes = 0;
#else
atomic_store_u64(&prof_accum->accumbytes, 0, ATOMIC_RELAXED);
#endif
Revert "Refactor profiling" This reverts commit 0b462407ae84a62b3c097f0e9f18df487a47d9a7.
2019-07-30 05:09:20 +08:00
return false;
}
Add "prof.dump_prefix" to override filename prefixes for dumps.
2019-09-10 11:18:41 +08:00
bool
prof_dump_prefix_set(tsdn_t *tsdn, const char *prefix) {
cassert(config_prof);
ctl_mtx_assert_held(tsdn);
malloc_mutex_lock(tsdn, &prof_dump_filename_mtx);
if (prof_dump_prefix == NULL) {
malloc_mutex_unlock(tsdn, &prof_dump_filename_mtx);
/* Everything is still guarded by ctl_mtx. */
char *buffer = base_alloc(tsdn, b0get(), PROF_DUMP_FILENAME_LEN,
QUANTUM);
if (buffer == NULL) {
return true;
}
malloc_mutex_lock(tsdn, &prof_dump_filename_mtx);
prof_dump_prefix = buffer;
}
assert(prof_dump_prefix != NULL);
strncpy(prof_dump_prefix, prefix, PROF_DUMP_FILENAME_LEN - 1);
prof_dump_prefix[PROF_DUMP_FILENAME_LEN - 1] = '\0';
malloc_mutex_unlock(tsdn, &prof_dump_filename_mtx);
return false;
}
Refactor profiling Refactored core profiling codebase into two logical parts: (a) `prof_data.c`: core internal data structure managing & dumping; (b) `prof.c`: mutexes & outward-facing APIs. Some internal functions had to be exposed out, but there are not that many of them if the modularization is (hopefully) clean enough.
2019-07-18 06:52:50 +08:00
void
prof_idump(tsdn_t *tsdn) {
tsd_t *tsd;
prof_tdata_t *tdata;
Revert "Refactor profiling" This reverts commit 0b462407ae84a62b3c097f0e9f18df487a47d9a7.
2019-07-30 05:09:20 +08:00
Refactor profiling Refactored core profiling codebase into two logical parts: (a) `prof_data.c`: core internal data structure managing & dumping; (b) `prof.c`: mutexes & outward-facing APIs. Some internal functions had to be exposed out, but there are not that many of them if the modularization is (hopefully) clean enough.
2019-07-18 06:52:50 +08:00
cassert(config_prof);
Revert "Refactor profiling" This reverts commit 0b462407ae84a62b3c097f0e9f18df487a47d9a7.
2019-07-30 05:09:20 +08:00
Refactor profiling Refactored core profiling codebase into two logical parts: (a) `prof_data.c`: core internal data structure managing & dumping; (b) `prof.c`: mutexes & outward-facing APIs. Some internal functions had to be exposed out, but there are not that many of them if the modularization is (hopefully) clean enough.
2019-07-18 06:52:50 +08:00
if (!prof_booted || tsdn_null(tsdn) || !prof_active_get_unlocked()) {
Revert "Refactor profiling" This reverts commit 0b462407ae84a62b3c097f0e9f18df487a47d9a7.
2019-07-30 05:09:20 +08:00
return;
}
tsd = tsdn_tsd(tsdn);
if (tsd_reentrancy_level_get(tsd) > 0) {
return;
}
tdata = prof_tdata_get(tsd, false);
if (tdata == NULL) {
return;
}
if (tdata->enq) {
tdata->enq_idump = true;
return;
}
Add "prof.dump_prefix" to override filename prefixes for dumps.
2019-09-10 11:18:41 +08:00
malloc_mutex_lock(tsd_tsdn(tsd), &prof_dump_filename_mtx);
if (prof_dump_prefix_get(tsd_tsdn(tsd))[0] == '\0') {
Rename "prof_dump_seq_mtx" to "prof_dump_filename_mtx".
2019-09-10 11:04:18 +08:00
malloc_mutex_unlock(tsd_tsdn(tsd), &prof_dump_filename_mtx);
Add "prof.dump_prefix" to override filename prefixes for dumps.
2019-09-10 11:18:41 +08:00
return;
Revert "Refactor profiling" This reverts commit 0b462407ae84a62b3c097f0e9f18df487a47d9a7.
2019-07-30 05:09:20 +08:00
}
Add "prof.dump_prefix" to override filename prefixes for dumps.
2019-09-10 11:18:41 +08:00
char filename[PATH_MAX + 1];
prof_dump_filename(tsd, filename, 'i', prof_dump_iseq);
prof_dump_iseq++;
malloc_mutex_unlock(tsd_tsdn(tsd), &prof_dump_filename_mtx);
prof_dump(tsd, false, filename, false);
Revert "Refactor profiling" This reverts commit 0b462407ae84a62b3c097f0e9f18df487a47d9a7.
2019-07-30 05:09:20 +08:00
}
bool
prof_mdump(tsd_t *tsd, const char *filename) {
cassert(config_prof);
assert(tsd_reentrancy_level_get(tsd) == 0);
if (!opt_prof || !prof_booted) {
Remove extraneous parens around return arguments. This resolves #540.
2017-01-20 10:15:45 +08:00
return true;
Update brace style. Add braces around single-line blocks, and remove line breaks before function-opening braces. This resolves #537.
2017-01-16 08:56:30 +08:00
}
Avoid prof_dump during reentrancy.
2017-04-25 09:14:57 +08:00
char filename_buf[DUMP_FILENAME_BUFSIZE];
Allow prof.dump mallctl to specify filename.
2010-03-03 03:57:30 +08:00
if (filename == NULL) {
/* No filename specified, so automatically generate one. */
Add "prof.dump_prefix" to override filename prefixes for dumps.
2019-09-10 11:18:41 +08:00
malloc_mutex_lock(tsd_tsdn(tsd), &prof_dump_filename_mtx);
if (prof_dump_prefix_get(tsd_tsdn(tsd))[0] == '\0') {
malloc_mutex_unlock(tsd_tsdn(tsd), &prof_dump_filename_mtx);
Remove extraneous parens around return arguments. This resolves #540.
2017-01-20 10:15:45 +08:00
return true;
Update brace style. Add braces around single-line blocks, and remove line breaks before function-opening braces. This resolves #537.
2017-01-16 08:56:30 +08:00
}
Add "prof.dump_prefix" to override filename prefixes for dumps.
2019-09-10 11:18:41 +08:00
prof_dump_filename(tsd, filename_buf, 'm', prof_dump_mseq);
Allow prof.dump mallctl to specify filename.
2010-03-03 03:57:30 +08:00
prof_dump_mseq++;
Rename "prof_dump_seq_mtx" to "prof_dump_filename_mtx".
2019-09-10 11:04:18 +08:00
malloc_mutex_unlock(tsd_tsdn(tsd), &prof_dump_filename_mtx);
Allow prof.dump mallctl to specify filename.
2010-03-03 03:57:30 +08:00
filename = filename_buf;
}
Remove extraneous parens around return arguments. This resolves #540.
2017-01-20 10:15:45 +08:00
return prof_dump(tsd, true, filename, false);
Implement allocation profiling and leack checking. Add the --enable-prof and --enable-prof-libunwind configure options. Add the B/b, F/f, I/i, L/l, and U/u JEMALLOC_OPTIONS. Interval-based profile dump triggering is not yet implemented. Add supporting generic code: * Add memory barriers. * Add prn (LCG PRNG). * Add hash (Murmur hash function). * Add ckh (cuckoo hash tables).
2010-02-11 02:37:56 +08:00
}
void
Update brace style. Add braces around single-line blocks, and remove line breaks before function-opening braces. This resolves #537.
2017-01-16 08:56:30 +08:00
prof_gdump(tsdn_t *tsdn) {
Resolve bootstrapping issues when embedded in FreeBSD libc. b2c0d6322d2307458ae2b28545f8a5c9903d7ef5 (Add witness, a simple online locking validator.) caused a broad propagation of tsd throughout the internal API, but tsd_fetch() was designed to fail prior to tsd bootstrapping. Fix this by splitting tsd_t into non-nullable tsd_t and nullable tsdn_t, and modifying all internal APIs that do not critically rely on tsd to take nullable pointers. Furthermore, add the tsd_booted_get() function so that tsdn_fetch() can probe whether tsd bootstrapping is complete and return NULL if not. All dangerous conversions of nullable pointers are tsdn_tsd() calls that assert-fail on invalid conversion.
2016-05-11 13:21:10 +08:00
tsd_t *tsd;
Implement per thread heap profiling. Rename data structures (prof_thr_cnt_t-->prof_tctx_t, prof_ctx_t-->prof_gctx_t), and convert to storing a prof_tctx_t for sampled objects. Convert PROF_ALLOC_PREP() to prof_alloc_prep(), since precise backtrace depth within jemalloc functions is no longer an issue (pprof prunes irrelevant frames). Implement mallctl's: - prof.reset implements full sample data reset, and optional change of sample interval. - prof.lg_sample reads the current sample interval (opt.lg_prof_sample was the permanent source of truth prior to prof.reset). - thread.prof.name provides naming capability for threads within heap profile dumps. - thread.prof.active makes it possible to activate/deactivate heap profiling for individual threads. Modify the heap dump files to contain per thread heap profile data. This change is incompatible with the existing pprof, which will require enhancements to read and process the enriched data.
2014-08-19 07:22:13 +08:00
prof_tdata_t *tdata;
Implement allocation profiling and leack checking. Add the --enable-prof and --enable-prof-libunwind configure options. Add the B/b, F/f, I/i, L/l, and U/u JEMALLOC_OPTIONS. Interval-based profile dump triggering is not yet implemented. Add supporting generic code: * Add memory barriers. * Add prn (LCG PRNG). * Add hash (Murmur hash function). * Add ckh (cuckoo hash tables).
2010-02-11 02:37:56 +08:00
Reduce cpp conditional logic complexity. Convert configuration-related cpp conditional logic to use static constant variables, e.g.: #ifdef JEMALLOC_DEBUG [...] #endif becomes: if (config_debug) { [...] } The advantage is clearer, more concise code. The main disadvantage is that data structures no longer have conditionally defined fields, so they pay the cost of all fields regardless of whether they are used. In practice, this is only a minor concern; config_stats will go away in an upcoming change, and config_prof is the only other major feature that depends on more than a few special-purpose fields.
2012-02-11 12:22:09 +08:00
cassert(config_prof);
Control idump and gdump with prof_active.
2018-04-07 04:45:37 +08:00
if (!prof_booted || tsdn_null(tsdn) || !prof_active_get_unlocked()) {
Implement allocation profiling and leack checking. Add the --enable-prof and --enable-prof-libunwind configure options. Add the B/b, F/f, I/i, L/l, and U/u JEMALLOC_OPTIONS. Interval-based profile dump triggering is not yet implemented. Add supporting generic code: * Add memory barriers. * Add prn (LCG PRNG). * Add hash (Murmur hash function). * Add ckh (cuckoo hash tables).
2010-02-11 02:37:56 +08:00
return;
Update brace style. Add braces around single-line blocks, and remove line breaks before function-opening braces. This resolves #537.
2017-01-16 08:56:30 +08:00
}
Resolve bootstrapping issues when embedded in FreeBSD libc. b2c0d6322d2307458ae2b28545f8a5c9903d7ef5 (Add witness, a simple online locking validator.) caused a broad propagation of tsd throughout the internal API, but tsd_fetch() was designed to fail prior to tsd bootstrapping. Fix this by splitting tsd_t into non-nullable tsd_t and nullable tsdn_t, and modifying all internal APIs that do not critically rely on tsd to take nullable pointers. Furthermore, add the tsd_booted_get() function so that tsdn_fetch() can probe whether tsd bootstrapping is complete and return NULL if not. All dangerous conversions of nullable pointers are tsdn_tsd() calls that assert-fail on invalid conversion.
2016-05-11 13:21:10 +08:00
tsd = tsdn_tsd(tsdn);
Header refactoring: tsd - cleanup and dependency breaking. This removes the tsd macros (which are used only for tsd_t in real builds). We break up the circular dependencies involving tsd. We also move all tsd access through getters and setters. This allows us to assert that we only touch data when tsd is in a valid state. We simplify the usages of the x macro trick, removing all the customizability (get/set, init, cleanup), moving the lifetime logic to tsd_init and tsd_cleanup. This lets us make initialization order independent of order within tsd_t.
2017-04-27 09:37:44 +08:00
if (tsd_reentrancy_level_get(tsd) > 0) {
Avoid prof_dump during reentrancy.
2017-04-25 09:14:57 +08:00
return;
}
Convert all tsd variables to reside in a single tsd structure.
2014-09-23 12:09:23 +08:00
tdata = prof_tdata_get(tsd, false);
Update brace style. Add braces around single-line blocks, and remove line breaks before function-opening braces. This resolves #537.
2017-01-16 08:56:30 +08:00
if (tdata == NULL) {
Fix heap profiling bugs. Fix a potential deadlock that could occur during interval- and growth-triggered heap profile dumps. Fix an off-by-one heap profile statistics bug that could be observed in interval- and growth-triggered heap profiles. Fix heap profile dump filename sequence numbers (regression during conversion to malloc_snprintf()).
2012-04-23 07:00:11 +08:00
return;
Update brace style. Add braces around single-line blocks, and remove line breaks before function-opening braces. This resolves #537.
2017-01-16 08:56:30 +08:00
}
Implement per thread heap profiling. Rename data structures (prof_thr_cnt_t-->prof_tctx_t, prof_ctx_t-->prof_gctx_t), and convert to storing a prof_tctx_t for sampled objects. Convert PROF_ALLOC_PREP() to prof_alloc_prep(), since precise backtrace depth within jemalloc functions is no longer an issue (pprof prunes irrelevant frames). Implement mallctl's: - prof.reset implements full sample data reset, and optional change of sample interval. - prof.lg_sample reads the current sample interval (opt.lg_prof_sample was the permanent source of truth prior to prof.reset). - thread.prof.name provides naming capability for threads within heap profile dumps. - thread.prof.active makes it possible to activate/deactivate heap profiling for individual threads. Modify the heap dump files to contain per thread heap profile data. This change is incompatible with the existing pprof, which will require enhancements to read and process the enriched data.
2014-08-19 07:22:13 +08:00
if (tdata->enq) {
tdata->enq_gdump = true;
Implement allocation profiling and leack checking. Add the --enable-prof and --enable-prof-libunwind configure options. Add the B/b, F/f, I/i, L/l, and U/u JEMALLOC_OPTIONS. Interval-based profile dump triggering is not yet implemented. Add supporting generic code: * Add memory barriers. * Add prn (LCG PRNG). * Add hash (Murmur hash function). * Add ckh (cuckoo hash tables).
2010-02-11 02:37:56 +08:00
return;
}
Add "prof.dump_prefix" to override filename prefixes for dumps.
2019-09-10 11:18:41 +08:00
malloc_mutex_lock(tsdn, &prof_dump_filename_mtx);
if (prof_dump_prefix_get(tsdn)[0] == '\0') {
Rename "prof_dump_seq_mtx" to "prof_dump_filename_mtx".
2019-09-10 11:04:18 +08:00
malloc_mutex_unlock(tsdn, &prof_dump_filename_mtx);
Add "prof.dump_prefix" to override filename prefixes for dumps.
2019-09-10 11:18:41 +08:00
return;
Replace JEMALLOC_OPTIONS with MALLOC_CONF. Replace the single-character run-time flags with key/value pairs, which can be set via the malloc_conf global, /etc/malloc.conf, and the MALLOC_CONF environment variable. Replace the JEMALLOC_PROF_PREFIX environment variable with the "opt.prof_prefix" option. Replace umax2s() with u2s().
2010-10-24 09:37:06 +08:00
}
Add "prof.dump_prefix" to override filename prefixes for dumps.
2019-09-10 11:18:41 +08:00
char filename[DUMP_FILENAME_BUFSIZE];
prof_dump_filename(tsd, filename, 'u', prof_dump_useq);
prof_dump_useq++;
malloc_mutex_unlock(tsdn, &prof_dump_filename_mtx);
prof_dump(tsd, false, filename, false);
Implement allocation profiling and leack checking. Add the --enable-prof and --enable-prof-libunwind configure options. Add the B/b, F/f, I/i, L/l, and U/u JEMALLOC_OPTIONS. Interval-based profile dump triggering is not yet implemented. Add supporting generic code: * Add memory barriers. * Add prn (LCG PRNG). * Add hash (Murmur hash function). * Add ckh (cuckoo hash tables).
2010-02-11 02:37:56 +08:00
}
Get rid of most of the various inline macros.
2017-04-22 00:37:34 +08:00
static uint64_t
Update brace style. Add braces around single-line blocks, and remove line breaks before function-opening braces. This resolves #537.
2017-01-16 08:56:30 +08:00
prof_thr_uid_alloc(tsdn_t *tsdn) {
Fix prof regressions. Don't use atomic_add_uint64(), because it isn't available on 32-bit platforms. Fix forking support functions to manage all prof-related mutexes. These regressions were introduced by 602c8e0971160e4b85b08b16cf8a2375aa24bc04 (Implement per thread heap profiling.), which did not make it into any releases prior to these fixes.
2014-09-12 09:06:30 +08:00
uint64_t thr_uid;
Implement per thread heap profiling. Rename data structures (prof_thr_cnt_t-->prof_tctx_t, prof_ctx_t-->prof_gctx_t), and convert to storing a prof_tctx_t for sampled objects. Convert PROF_ALLOC_PREP() to prof_alloc_prep(), since precise backtrace depth within jemalloc functions is no longer an issue (pprof prunes irrelevant frames). Implement mallctl's: - prof.reset implements full sample data reset, and optional change of sample interval. - prof.lg_sample reads the current sample interval (opt.lg_prof_sample was the permanent source of truth prior to prof.reset). - thread.prof.name provides naming capability for threads within heap profile dumps. - thread.prof.active makes it possible to activate/deactivate heap profiling for individual threads. Modify the heap dump files to contain per thread heap profile data. This change is incompatible with the existing pprof, which will require enhancements to read and process the enriched data.
2014-08-19 07:22:13 +08:00
Resolve bootstrapping issues when embedded in FreeBSD libc. b2c0d6322d2307458ae2b28545f8a5c9903d7ef5 (Add witness, a simple online locking validator.) caused a broad propagation of tsd throughout the internal API, but tsd_fetch() was designed to fail prior to tsd bootstrapping. Fix this by splitting tsd_t into non-nullable tsd_t and nullable tsdn_t, and modifying all internal APIs that do not critically rely on tsd to take nullable pointers. Furthermore, add the tsd_booted_get() function so that tsdn_fetch() can probe whether tsd bootstrapping is complete and return NULL if not. All dangerous conversions of nullable pointers are tsdn_tsd() calls that assert-fail on invalid conversion.
2016-05-11 13:21:10 +08:00
malloc_mutex_lock(tsdn, &next_thr_uid_mtx);
Fix prof regressions. Don't use atomic_add_uint64(), because it isn't available on 32-bit platforms. Fix forking support functions to manage all prof-related mutexes. These regressions were introduced by 602c8e0971160e4b85b08b16cf8a2375aa24bc04 (Implement per thread heap profiling.), which did not make it into any releases prior to these fixes.
2014-09-12 09:06:30 +08:00
thr_uid = next_thr_uid;
next_thr_uid++;
Resolve bootstrapping issues when embedded in FreeBSD libc. b2c0d6322d2307458ae2b28545f8a5c9903d7ef5 (Add witness, a simple online locking validator.) caused a broad propagation of tsd throughout the internal API, but tsd_fetch() was designed to fail prior to tsd bootstrapping. Fix this by splitting tsd_t into non-nullable tsd_t and nullable tsdn_t, and modifying all internal APIs that do not critically rely on tsd to take nullable pointers. Furthermore, add the tsd_booted_get() function so that tsdn_fetch() can probe whether tsd bootstrapping is complete and return NULL if not. All dangerous conversions of nullable pointers are tsdn_tsd() calls that assert-fail on invalid conversion.
2016-05-11 13:21:10 +08:00
malloc_mutex_unlock(tsdn, &next_thr_uid_mtx);
Fix prof regressions. Don't use atomic_add_uint64(), because it isn't available on 32-bit platforms. Fix forking support functions to manage all prof-related mutexes. These regressions were introduced by 602c8e0971160e4b85b08b16cf8a2375aa24bc04 (Implement per thread heap profiling.), which did not make it into any releases prior to these fixes.
2014-09-12 09:06:30 +08:00
Remove extraneous parens around return arguments. This resolves #540.
2017-01-20 10:15:45 +08:00
return thr_uid;
Implement per thread heap profiling. Rename data structures (prof_thr_cnt_t-->prof_tctx_t, prof_ctx_t-->prof_gctx_t), and convert to storing a prof_tctx_t for sampled objects. Convert PROF_ALLOC_PREP() to prof_alloc_prep(), since precise backtrace depth within jemalloc functions is no longer an issue (pprof prunes irrelevant frames). Implement mallctl's: - prof.reset implements full sample data reset, and optional change of sample interval. - prof.lg_sample reads the current sample interval (opt.lg_prof_sample was the permanent source of truth prior to prof.reset). - thread.prof.name provides naming capability for threads within heap profile dumps. - thread.prof.active makes it possible to activate/deactivate heap profiling for individual threads. Modify the heap dump files to contain per thread heap profile data. This change is incompatible with the existing pprof, which will require enhancements to read and process the enriched data.
2014-08-19 07:22:13 +08:00
}
prof_tdata_t *
Update brace style. Add braces around single-line blocks, and remove line breaks before function-opening braces. This resolves #537.
2017-01-16 08:56:30 +08:00
prof_tdata_init(tsd_t *tsd) {
Remove extraneous parens around return arguments. This resolves #540.
2017-01-20 10:15:45 +08:00
return prof_tdata_init_impl(tsd, prof_thr_uid_alloc(tsd_tsdn(tsd)), 0,
NULL, prof_thread_active_init_get(tsd_tsdn(tsd)));
Implement per thread heap profiling. Rename data structures (prof_thr_cnt_t-->prof_tctx_t, prof_ctx_t-->prof_gctx_t), and convert to storing a prof_tctx_t for sampled objects. Convert PROF_ALLOC_PREP() to prof_alloc_prep(), since precise backtrace depth within jemalloc functions is no longer an issue (pprof prunes irrelevant frames). Implement mallctl's: - prof.reset implements full sample data reset, and optional change of sample interval. - prof.lg_sample reads the current sample interval (opt.lg_prof_sample was the permanent source of truth prior to prof.reset). - thread.prof.name provides naming capability for threads within heap profile dumps. - thread.prof.active makes it possible to activate/deactivate heap profiling for individual threads. Modify the heap dump files to contain per thread heap profile data. This change is incompatible with the existing pprof, which will require enhancements to read and process the enriched data.
2014-08-19 07:22:13 +08:00
}
Refactor profiling Refactored core profiling codebase into two logical parts: (a) `prof_data.c`: core internal data structure managing & dumping; (b) `prof.c`: mutexes & outward-facing APIs. Some internal functions had to be exposed out, but there are not that many of them if the modularization is (hopefully) clean enough.
2019-07-18 06:52:50 +08:00
static char *
prof_thread_name_alloc(tsdn_t *tsdn, const char *thread_name) {
char *ret;
size_t size;
Revert "Refactor profiling" This reverts commit 0b462407ae84a62b3c097f0e9f18df487a47d9a7.
2019-07-30 05:09:20 +08:00
Refactor profiling Refactored core profiling codebase into two logical parts: (a) `prof_data.c`: core internal data structure managing & dumping; (b) `prof.c`: mutexes & outward-facing APIs. Some internal functions had to be exposed out, but there are not that many of them if the modularization is (hopefully) clean enough.
2019-07-18 06:52:50 +08:00
if (thread_name == NULL) {
return NULL;
Update brace style. Add braces around single-line blocks, and remove line breaks before function-opening braces. This resolves #537.
2017-01-16 08:56:30 +08:00
}
Revert "Refactor profiling" This reverts commit 0b462407ae84a62b3c097f0e9f18df487a47d9a7.
2019-07-30 05:09:20 +08:00
Refactor profiling Refactored core profiling codebase into two logical parts: (a) `prof_data.c`: core internal data structure managing & dumping; (b) `prof.c`: mutexes & outward-facing APIs. Some internal functions had to be exposed out, but there are not that many of them if the modularization is (hopefully) clean enough.
2019-07-18 06:52:50 +08:00
size = strlen(thread_name) + 1;
if (size == 1) {
return "";
Revert "Refactor profiling" This reverts commit 0b462407ae84a62b3c097f0e9f18df487a47d9a7.
2019-07-30 05:09:20 +08:00
}
Refactor profiling Refactored core profiling codebase into two logical parts: (a) `prof_data.c`: core internal data structure managing & dumping; (b) `prof.c`: mutexes & outward-facing APIs. Some internal functions had to be exposed out, but there are not that many of them if the modularization is (hopefully) clean enough.
2019-07-18 06:52:50 +08:00
ret = iallocztm(tsdn, size, sz_size2index(size), false, NULL, true,
arena_get(TSDN_NULL, 0, true), true);
if (ret == NULL) {
return NULL;
Update brace style. Add braces around single-line blocks, and remove line breaks before function-opening braces. This resolves #537.
2017-01-16 08:56:30 +08:00
}
Refactor profiling Refactored core profiling codebase into two logical parts: (a) `prof_data.c`: core internal data structure managing & dumping; (b) `prof.c`: mutexes & outward-facing APIs. Some internal functions had to be exposed out, but there are not that many of them if the modularization is (hopefully) clean enough.
2019-07-18 06:52:50 +08:00
memcpy(ret, thread_name, size);
return ret;
Implement per thread heap profiling. Rename data structures (prof_thr_cnt_t-->prof_tctx_t, prof_ctx_t-->prof_gctx_t), and convert to storing a prof_tctx_t for sampled objects. Convert PROF_ALLOC_PREP() to prof_alloc_prep(), since precise backtrace depth within jemalloc functions is no longer an issue (pprof prunes irrelevant frames). Implement mallctl's: - prof.reset implements full sample data reset, and optional change of sample interval. - prof.lg_sample reads the current sample interval (opt.lg_prof_sample was the permanent source of truth prior to prof.reset). - thread.prof.name provides naming capability for threads within heap profile dumps. - thread.prof.active makes it possible to activate/deactivate heap profiling for individual threads. Modify the heap dump files to contain per thread heap profile data. This change is incompatible with the existing pprof, which will require enhancements to read and process the enriched data.
2014-08-19 07:22:13 +08:00
}
Inline the fast path for heap sampling. Inline the heap sampling code that is executed for every allocation event (regardless of whether a sample is taken). Combine all prof TLS data into a single data structure, in order to reduce the TLS lookup volume.
2010-10-21 10:05:59 +08:00
Test prof.reset mallctl and fix numerous discovered bugs.
2014-10-03 14:01:10 +08:00
prof_tdata_t *
Update brace style. Add braces around single-line blocks, and remove line breaks before function-opening braces. This resolves #537.
2017-01-16 08:56:30 +08:00
prof_tdata_reinit(tsd_t *tsd, prof_tdata_t *tdata) {
Test prof.reset mallctl and fix numerous discovered bugs.
2014-10-03 14:01:10 +08:00
uint64_t thr_uid = tdata->thr_uid;
uint64_t thr_discrim = tdata->thr_discrim + 1;
Implement/test/fix prof-related mallctl's. Implement/test/fix the opt.prof_thread_active_init, prof.thread_active_init, and thread.prof.active mallctl's. Test/fix the thread.prof.name mallctl. Refactor opt_prof_active to be read-only and move mutable state into the prof_active variable. Stop leaning on ctl-related locking for protection.
2014-10-04 14:25:30 +08:00
char *thread_name = (tdata->thread_name != NULL) ?
Resolve bootstrapping issues when embedded in FreeBSD libc. b2c0d6322d2307458ae2b28545f8a5c9903d7ef5 (Add witness, a simple online locking validator.) caused a broad propagation of tsd throughout the internal API, but tsd_fetch() was designed to fail prior to tsd bootstrapping. Fix this by splitting tsd_t into non-nullable tsd_t and nullable tsdn_t, and modifying all internal APIs that do not critically rely on tsd to take nullable pointers. Furthermore, add the tsd_booted_get() function so that tsdn_fetch() can probe whether tsd bootstrapping is complete and return NULL if not. All dangerous conversions of nullable pointers are tsdn_tsd() calls that assert-fail on invalid conversion.
2016-05-11 13:21:10 +08:00
prof_thread_name_alloc(tsd_tsdn(tsd), tdata->thread_name) : NULL;
Implement/test/fix prof-related mallctl's. Implement/test/fix the opt.prof_thread_active_init, prof.thread_active_init, and thread.prof.active mallctl's. Test/fix the thread.prof.name mallctl. Refactor opt_prof_active to be read-only and move mutable state into the prof_active variable. Stop leaning on ctl-related locking for protection.
2014-10-04 14:25:30 +08:00
bool active = tdata->active;
Fix heap profiling bugs. Fix a potential deadlock that could occur during interval- and growth-triggered heap profile dumps. Fix an off-by-one heap profile statistics bug that could be observed in interval- and growth-triggered heap profiles. Fix heap profile dump filename sequence numbers (regression during conversion to malloc_snprintf()).
2012-04-23 07:00:11 +08:00
Test prof.reset mallctl and fix numerous discovered bugs.
2014-10-03 14:01:10 +08:00
prof_tdata_detach(tsd, tdata);
Remove extraneous parens around return arguments. This resolves #540.
2017-01-20 10:15:45 +08:00
return prof_tdata_init_impl(tsd, thr_uid, thr_discrim, thread_name,
active);
Implement per thread heap profiling. Rename data structures (prof_thr_cnt_t-->prof_tctx_t, prof_ctx_t-->prof_gctx_t), and convert to storing a prof_tctx_t for sampled objects. Convert PROF_ALLOC_PREP() to prof_alloc_prep(), since precise backtrace depth within jemalloc functions is no longer an issue (pprof prunes irrelevant frames). Implement mallctl's: - prof.reset implements full sample data reset, and optional change of sample interval. - prof.lg_sample reads the current sample interval (opt.lg_prof_sample was the permanent source of truth prior to prof.reset). - thread.prof.name provides naming capability for threads within heap profile dumps. - thread.prof.active makes it possible to activate/deactivate heap profiling for individual threads. Modify the heap dump files to contain per thread heap profile data. This change is incompatible with the existing pprof, which will require enhancements to read and process the enriched data.
2014-08-19 07:22:13 +08:00
}
Inline the fast path for heap sampling. Inline the heap sampling code that is executed for every allocation event (regardless of whether a sample is taken). Combine all prof TLS data into a single data structure, in order to reduce the TLS lookup volume.
2010-10-21 10:05:59 +08:00
Implement tsd. Implement tsd, which is a TLS/TSD abstraction that uses one or both internally. Modify bootstrapping such that no tsd's are utilized until allocation is safe. Remove malloc_[v]tprintf(), and use malloc_snprintf() instead. Fix %p argument size handling in malloc_vsnprintf(). Fix a long-standing statistics-related bug in the "thread.arena" mallctl that could cause crashes due to linked list corruption.
2012-03-22 09:33:03 +08:00
void
Update brace style. Add braces around single-line blocks, and remove line breaks before function-opening braces. This resolves #537.
2017-01-16 08:56:30 +08:00
prof_tdata_cleanup(tsd_t *tsd) {
Convert all tsd variables to reside in a single tsd structure.
2014-09-23 12:09:23 +08:00
prof_tdata_t *tdata;
Implement allocation profiling and leack checking. Add the --enable-prof and --enable-prof-libunwind configure options. Add the B/b, F/f, I/i, L/l, and U/u JEMALLOC_OPTIONS. Interval-based profile dump triggering is not yet implemented. Add supporting generic code: * Add memory barriers. * Add prn (LCG PRNG). * Add hash (Murmur hash function). * Add ckh (cuckoo hash tables).
2010-02-11 02:37:56 +08:00
Update brace style. Add braces around single-line blocks, and remove line breaks before function-opening braces. This resolves #537.
2017-01-16 08:56:30 +08:00
if (!config_prof) {
Convert all tsd variables to reside in a single tsd structure.
2014-09-23 12:09:23 +08:00
return;
Update brace style. Add braces around single-line blocks, and remove line breaks before function-opening braces. This resolves #537.
2017-01-16 08:56:30 +08:00
}
Reduce cpp conditional logic complexity. Convert configuration-related cpp conditional logic to use static constant variables, e.g.: #ifdef JEMALLOC_DEBUG [...] #endif becomes: if (config_debug) { [...] } The advantage is clearer, more concise code. The main disadvantage is that data structures no longer have conditionally defined fields, so they pay the cost of all fields regardless of whether they are used. In practice, this is only a minor concern; config_stats will go away in an upcoming change, and config_prof is the only other major feature that depends on more than a few special-purpose fields.
2012-02-11 12:22:09 +08:00
Convert all tsd variables to reside in a single tsd structure.
2014-09-23 12:09:23 +08:00
tdata = tsd_prof_tdata_get(tsd);
Update brace style. Add braces around single-line blocks, and remove line breaks before function-opening braces. This resolves #537.
2017-01-16 08:56:30 +08:00
if (tdata != NULL) {
Convert all tsd variables to reside in a single tsd structure.
2014-09-23 12:09:23 +08:00
prof_tdata_detach(tsd, tdata);
Update brace style. Add braces around single-line blocks, and remove line breaks before function-opening braces. This resolves #537.
2017-01-16 08:56:30 +08:00
}
Increase default backtrace depth from 4 to 128. Increase the default backtrace depth, because shallow backtraces tend to result in confusing pprof output graphs.
2010-10-03 13:38:14 +08:00
}
Implement/test/fix prof-related mallctl's. Implement/test/fix the opt.prof_thread_active_init, prof.thread_active_init, and thread.prof.active mallctl's. Test/fix the thread.prof.name mallctl. Refactor opt_prof_active to be read-only and move mutable state into the prof_active variable. Stop leaning on ctl-related locking for protection.
2014-10-04 14:25:30 +08:00
bool
Update brace style. Add braces around single-line blocks, and remove line breaks before function-opening braces. This resolves #537.
2017-01-16 08:56:30 +08:00
prof_active_get(tsdn_t *tsdn) {
Implement/test/fix prof-related mallctl's. Implement/test/fix the opt.prof_thread_active_init, prof.thread_active_init, and thread.prof.active mallctl's. Test/fix the thread.prof.name mallctl. Refactor opt_prof_active to be read-only and move mutable state into the prof_active variable. Stop leaning on ctl-related locking for protection.
2014-10-04 14:25:30 +08:00
bool prof_active_current;
Guard prof_active reset by opt_prof Set `prof_active` to read-only when `opt_prof` is turned off.
2019-08-28 05:42:14 +08:00
prof_active_assert();
Resolve bootstrapping issues when embedded in FreeBSD libc. b2c0d6322d2307458ae2b28545f8a5c9903d7ef5 (Add witness, a simple online locking validator.) caused a broad propagation of tsd throughout the internal API, but tsd_fetch() was designed to fail prior to tsd bootstrapping. Fix this by splitting tsd_t into non-nullable tsd_t and nullable tsdn_t, and modifying all internal APIs that do not critically rely on tsd to take nullable pointers. Furthermore, add the tsd_booted_get() function so that tsdn_fetch() can probe whether tsd bootstrapping is complete and return NULL if not. All dangerous conversions of nullable pointers are tsdn_tsd() calls that assert-fail on invalid conversion.
2016-05-11 13:21:10 +08:00
malloc_mutex_lock(tsdn, &prof_active_mtx);
Implement/test/fix prof-related mallctl's. Implement/test/fix the opt.prof_thread_active_init, prof.thread_active_init, and thread.prof.active mallctl's. Test/fix the thread.prof.name mallctl. Refactor opt_prof_active to be read-only and move mutable state into the prof_active variable. Stop leaning on ctl-related locking for protection.
2014-10-04 14:25:30 +08:00
prof_active_current = prof_active;
Resolve bootstrapping issues when embedded in FreeBSD libc. b2c0d6322d2307458ae2b28545f8a5c9903d7ef5 (Add witness, a simple online locking validator.) caused a broad propagation of tsd throughout the internal API, but tsd_fetch() was designed to fail prior to tsd bootstrapping. Fix this by splitting tsd_t into non-nullable tsd_t and nullable tsdn_t, and modifying all internal APIs that do not critically rely on tsd to take nullable pointers. Furthermore, add the tsd_booted_get() function so that tsdn_fetch() can probe whether tsd bootstrapping is complete and return NULL if not. All dangerous conversions of nullable pointers are tsdn_tsd() calls that assert-fail on invalid conversion.
2016-05-11 13:21:10 +08:00
malloc_mutex_unlock(tsdn, &prof_active_mtx);
Remove extraneous parens around return arguments. This resolves #540.
2017-01-20 10:15:45 +08:00
return prof_active_current;
Implement/test/fix prof-related mallctl's. Implement/test/fix the opt.prof_thread_active_init, prof.thread_active_init, and thread.prof.active mallctl's. Test/fix the thread.prof.name mallctl. Refactor opt_prof_active to be read-only and move mutable state into the prof_active variable. Stop leaning on ctl-related locking for protection.
2014-10-04 14:25:30 +08:00
}
bool
Update brace style. Add braces around single-line blocks, and remove line breaks before function-opening braces. This resolves #537.
2017-01-16 08:56:30 +08:00
prof_active_set(tsdn_t *tsdn, bool active) {
Implement/test/fix prof-related mallctl's. Implement/test/fix the opt.prof_thread_active_init, prof.thread_active_init, and thread.prof.active mallctl's. Test/fix the thread.prof.name mallctl. Refactor opt_prof_active to be read-only and move mutable state into the prof_active variable. Stop leaning on ctl-related locking for protection.
2014-10-04 14:25:30 +08:00
bool prof_active_old;
Guard prof_active reset by opt_prof Set `prof_active` to read-only when `opt_prof` is turned off.
2019-08-28 05:42:14 +08:00
prof_active_assert();
Resolve bootstrapping issues when embedded in FreeBSD libc. b2c0d6322d2307458ae2b28545f8a5c9903d7ef5 (Add witness, a simple online locking validator.) caused a broad propagation of tsd throughout the internal API, but tsd_fetch() was designed to fail prior to tsd bootstrapping. Fix this by splitting tsd_t into non-nullable tsd_t and nullable tsdn_t, and modifying all internal APIs that do not critically rely on tsd to take nullable pointers. Furthermore, add the tsd_booted_get() function so that tsdn_fetch() can probe whether tsd bootstrapping is complete and return NULL if not. All dangerous conversions of nullable pointers are tsdn_tsd() calls that assert-fail on invalid conversion.
2016-05-11 13:21:10 +08:00
malloc_mutex_lock(tsdn, &prof_active_mtx);
Implement/test/fix prof-related mallctl's. Implement/test/fix the opt.prof_thread_active_init, prof.thread_active_init, and thread.prof.active mallctl's. Test/fix the thread.prof.name mallctl. Refactor opt_prof_active to be read-only and move mutable state into the prof_active variable. Stop leaning on ctl-related locking for protection.
2014-10-04 14:25:30 +08:00
prof_active_old = prof_active;
prof_active = active;
Resolve bootstrapping issues when embedded in FreeBSD libc. b2c0d6322d2307458ae2b28545f8a5c9903d7ef5 (Add witness, a simple online locking validator.) caused a broad propagation of tsd throughout the internal API, but tsd_fetch() was designed to fail prior to tsd bootstrapping. Fix this by splitting tsd_t into non-nullable tsd_t and nullable tsdn_t, and modifying all internal APIs that do not critically rely on tsd to take nullable pointers. Furthermore, add the tsd_booted_get() function so that tsdn_fetch() can probe whether tsd bootstrapping is complete and return NULL if not. All dangerous conversions of nullable pointers are tsdn_tsd() calls that assert-fail on invalid conversion.
2016-05-11 13:21:10 +08:00
malloc_mutex_unlock(tsdn, &prof_active_mtx);
Guard prof_active reset by opt_prof Set `prof_active` to read-only when `opt_prof` is turned off.
2019-08-28 05:42:14 +08:00
prof_active_assert();
Remove extraneous parens around return arguments. This resolves #540.
2017-01-20 10:15:45 +08:00
return prof_active_old;
Implement/test/fix prof-related mallctl's. Implement/test/fix the opt.prof_thread_active_init, prof.thread_active_init, and thread.prof.active mallctl's. Test/fix the thread.prof.name mallctl. Refactor opt_prof_active to be read-only and move mutable state into the prof_active variable. Stop leaning on ctl-related locking for protection.
2014-10-04 14:25:30 +08:00
}
Implement per thread heap profiling. Rename data structures (prof_thr_cnt_t-->prof_tctx_t, prof_ctx_t-->prof_gctx_t), and convert to storing a prof_tctx_t for sampled objects. Convert PROF_ALLOC_PREP() to prof_alloc_prep(), since precise backtrace depth within jemalloc functions is no longer an issue (pprof prunes irrelevant frames). Implement mallctl's: - prof.reset implements full sample data reset, and optional change of sample interval. - prof.lg_sample reads the current sample interval (opt.lg_prof_sample was the permanent source of truth prior to prof.reset). - thread.prof.name provides naming capability for threads within heap profile dumps. - thread.prof.active makes it possible to activate/deactivate heap profiling for individual threads. Modify the heap dump files to contain per thread heap profile data. This change is incompatible with the existing pprof, which will require enhancements to read and process the enriched data.
2014-08-19 07:22:13 +08:00
const char *
Update brace style. Add braces around single-line blocks, and remove line breaks before function-opening braces. This resolves #537.
2017-01-16 08:56:30 +08:00
prof_thread_name_get(tsd_t *tsd) {
Suppress tdata creation in reentrancy This change suppresses tdata initialization and prof sample threshold update in interrupting malloc calls. Interrupting calls have no need for tdata. Delaying tdata creation aligns better with our lazy tdata creation principle, and it also helps us gain control back from interrupting calls more quickly and reduces any risk of delegating tdata creation to an interrupting call.
2019-10-04 04:01:12 +08:00
assert(tsd_reentrancy_level_get(tsd) == 0);
Convert all tsd variables to reside in a single tsd structure.
2014-09-23 12:09:23 +08:00
prof_tdata_t *tdata;
tdata = prof_tdata_get(tsd, true);
Update brace style. Add braces around single-line blocks, and remove line breaks before function-opening braces. This resolves #537.
2017-01-16 08:56:30 +08:00
if (tdata == NULL) {
Remove extraneous parens around return arguments. This resolves #540.
2017-01-20 10:15:45 +08:00
return "";
Update brace style. Add braces around single-line blocks, and remove line breaks before function-opening braces. This resolves #537.
2017-01-16 08:56:30 +08:00
}
Implement/test/fix prof-related mallctl's. Implement/test/fix the opt.prof_thread_active_init, prof.thread_active_init, and thread.prof.active mallctl's. Test/fix the thread.prof.name mallctl. Refactor opt_prof_active to be read-only and move mutable state into the prof_active variable. Stop leaning on ctl-related locking for protection.
2014-10-04 14:25:30 +08:00
return (tdata->thread_name != NULL ? tdata->thread_name : "");
}
int
Update brace style. Add braces around single-line blocks, and remove line breaks before function-opening braces. This resolves #537.
2017-01-16 08:56:30 +08:00
prof_thread_name_set(tsd_t *tsd, const char *thread_name) {
Suppress tdata creation in reentrancy This change suppresses tdata initialization and prof sample threshold update in interrupting malloc calls. Interrupting calls have no need for tdata. Delaying tdata creation aligns better with our lazy tdata creation principle, and it also helps us gain control back from interrupting calls more quickly and reduces any risk of delegating tdata creation to an interrupting call.
2019-10-04 04:01:12 +08:00
assert(tsd_reentrancy_level_get(tsd) == 0);
Implement per thread heap profiling. Rename data structures (prof_thr_cnt_t-->prof_tctx_t, prof_ctx_t-->prof_gctx_t), and convert to storing a prof_tctx_t for sampled objects. Convert PROF_ALLOC_PREP() to prof_alloc_prep(), since precise backtrace depth within jemalloc functions is no longer an issue (pprof prunes irrelevant frames). Implement mallctl's: - prof.reset implements full sample data reset, and optional change of sample interval. - prof.lg_sample reads the current sample interval (opt.lg_prof_sample was the permanent source of truth prior to prof.reset). - thread.prof.name provides naming capability for threads within heap profile dumps. - thread.prof.active makes it possible to activate/deactivate heap profiling for individual threads. Modify the heap dump files to contain per thread heap profile data. This change is incompatible with the existing pprof, which will require enhancements to read and process the enriched data.
2014-08-19 07:22:13 +08:00
prof_tdata_t *tdata;
Implement/test/fix prof-related mallctl's. Implement/test/fix the opt.prof_thread_active_init, prof.thread_active_init, and thread.prof.active mallctl's. Test/fix the thread.prof.name mallctl. Refactor opt_prof_active to be read-only and move mutable state into the prof_active variable. Stop leaning on ctl-related locking for protection.
2014-10-04 14:25:30 +08:00
unsigned i;
Implement per thread heap profiling. Rename data structures (prof_thr_cnt_t-->prof_tctx_t, prof_ctx_t-->prof_gctx_t), and convert to storing a prof_tctx_t for sampled objects. Convert PROF_ALLOC_PREP() to prof_alloc_prep(), since precise backtrace depth within jemalloc functions is no longer an issue (pprof prunes irrelevant frames). Implement mallctl's: - prof.reset implements full sample data reset, and optional change of sample interval. - prof.lg_sample reads the current sample interval (opt.lg_prof_sample was the permanent source of truth prior to prof.reset). - thread.prof.name provides naming capability for threads within heap profile dumps. - thread.prof.active makes it possible to activate/deactivate heap profiling for individual threads. Modify the heap dump files to contain per thread heap profile data. This change is incompatible with the existing pprof, which will require enhancements to read and process the enriched data.
2014-08-19 07:22:13 +08:00
char *s;
Convert all tsd variables to reside in a single tsd structure.
2014-09-23 12:09:23 +08:00
tdata = prof_tdata_get(tsd, true);
Update brace style. Add braces around single-line blocks, and remove line breaks before function-opening braces. This resolves #537.
2017-01-16 08:56:30 +08:00
if (tdata == NULL) {
Remove extraneous parens around return arguments. This resolves #540.
2017-01-20 10:15:45 +08:00
return EAGAIN;
Update brace style. Add braces around single-line blocks, and remove line breaks before function-opening braces. This resolves #537.
2017-01-16 08:56:30 +08:00
}
Implement/test/fix prof-related mallctl's. Implement/test/fix the opt.prof_thread_active_init, prof.thread_active_init, and thread.prof.active mallctl's. Test/fix the thread.prof.name mallctl. Refactor opt_prof_active to be read-only and move mutable state into the prof_active variable. Stop leaning on ctl-related locking for protection.
2014-10-04 14:25:30 +08:00
/* Validate input. */
Update brace style. Add braces around single-line blocks, and remove line breaks before function-opening braces. This resolves #537.
2017-01-16 08:56:30 +08:00
if (thread_name == NULL) {
Remove extraneous parens around return arguments. This resolves #540.
2017-01-20 10:15:45 +08:00
return EFAULT;
Update brace style. Add braces around single-line blocks, and remove line breaks before function-opening braces. This resolves #537.
2017-01-16 08:56:30 +08:00
}
Implement/test/fix prof-related mallctl's. Implement/test/fix the opt.prof_thread_active_init, prof.thread_active_init, and thread.prof.active mallctl's. Test/fix the thread.prof.name mallctl. Refactor opt_prof_active to be read-only and move mutable state into the prof_active variable. Stop leaning on ctl-related locking for protection.
2014-10-04 14:25:30 +08:00
for (i = 0; thread_name[i] != '\0'; i++) {
char c = thread_name[i];
Update brace style. Add braces around single-line blocks, and remove line breaks before function-opening braces. This resolves #537.
2017-01-16 08:56:30 +08:00
if (!isgraph(c) && !isblank(c)) {
Remove extraneous parens around return arguments. This resolves #540.
2017-01-20 10:15:45 +08:00
return EFAULT;
Update brace style. Add braces around single-line blocks, and remove line breaks before function-opening braces. This resolves #537.
2017-01-16 08:56:30 +08:00
}
Implement/test/fix prof-related mallctl's. Implement/test/fix the opt.prof_thread_active_init, prof.thread_active_init, and thread.prof.active mallctl's. Test/fix the thread.prof.name mallctl. Refactor opt_prof_active to be read-only and move mutable state into the prof_active variable. Stop leaning on ctl-related locking for protection.
2014-10-04 14:25:30 +08:00
}
Implement per thread heap profiling. Rename data structures (prof_thr_cnt_t-->prof_tctx_t, prof_ctx_t-->prof_gctx_t), and convert to storing a prof_tctx_t for sampled objects. Convert PROF_ALLOC_PREP() to prof_alloc_prep(), since precise backtrace depth within jemalloc functions is no longer an issue (pprof prunes irrelevant frames). Implement mallctl's: - prof.reset implements full sample data reset, and optional change of sample interval. - prof.lg_sample reads the current sample interval (opt.lg_prof_sample was the permanent source of truth prior to prof.reset). - thread.prof.name provides naming capability for threads within heap profile dumps. - thread.prof.active makes it possible to activate/deactivate heap profiling for individual threads. Modify the heap dump files to contain per thread heap profile data. This change is incompatible with the existing pprof, which will require enhancements to read and process the enriched data.
2014-08-19 07:22:13 +08:00
Resolve bootstrapping issues when embedded in FreeBSD libc. b2c0d6322d2307458ae2b28545f8a5c9903d7ef5 (Add witness, a simple online locking validator.) caused a broad propagation of tsd throughout the internal API, but tsd_fetch() was designed to fail prior to tsd bootstrapping. Fix this by splitting tsd_t into non-nullable tsd_t and nullable tsdn_t, and modifying all internal APIs that do not critically rely on tsd to take nullable pointers. Furthermore, add the tsd_booted_get() function so that tsdn_fetch() can probe whether tsd bootstrapping is complete and return NULL if not. All dangerous conversions of nullable pointers are tsdn_tsd() calls that assert-fail on invalid conversion.
2016-05-11 13:21:10 +08:00
s = prof_thread_name_alloc(tsd_tsdn(tsd), thread_name);
Update brace style. Add braces around single-line blocks, and remove line breaks before function-opening braces. This resolves #537.
2017-01-16 08:56:30 +08:00
if (s == NULL) {
Remove extraneous parens around return arguments. This resolves #540.
2017-01-20 10:15:45 +08:00
return EAGAIN;
Update brace style. Add braces around single-line blocks, and remove line breaks before function-opening braces. This resolves #537.
2017-01-16 08:56:30 +08:00
}
Implement per thread heap profiling. Rename data structures (prof_thr_cnt_t-->prof_tctx_t, prof_ctx_t-->prof_gctx_t), and convert to storing a prof_tctx_t for sampled objects. Convert PROF_ALLOC_PREP() to prof_alloc_prep(), since precise backtrace depth within jemalloc functions is no longer an issue (pprof prunes irrelevant frames). Implement mallctl's: - prof.reset implements full sample data reset, and optional change of sample interval. - prof.lg_sample reads the current sample interval (opt.lg_prof_sample was the permanent source of truth prior to prof.reset). - thread.prof.name provides naming capability for threads within heap profile dumps. - thread.prof.active makes it possible to activate/deactivate heap profiling for individual threads. Modify the heap dump files to contain per thread heap profile data. This change is incompatible with the existing pprof, which will require enhancements to read and process the enriched data.
2014-08-19 07:22:13 +08:00
Implement/test/fix prof-related mallctl's. Implement/test/fix the opt.prof_thread_active_init, prof.thread_active_init, and thread.prof.active mallctl's. Test/fix the thread.prof.name mallctl. Refactor opt_prof_active to be read-only and move mutable state into the prof_active variable. Stop leaning on ctl-related locking for protection.
2014-10-04 14:25:30 +08:00
if (tdata->thread_name != NULL) {
Pass dealloc_ctx down free() fast path. This gets rid of the redundent rtree lookup down fast path.
2017-04-08 05:12:30 +08:00
idalloctm(tsd_tsdn(tsd), tdata->thread_name, NULL, NULL, true,
true);
Implement/test/fix prof-related mallctl's. Implement/test/fix the opt.prof_thread_active_init, prof.thread_active_init, and thread.prof.active mallctl's. Test/fix the thread.prof.name mallctl. Refactor opt_prof_active to be read-only and move mutable state into the prof_active variable. Stop leaning on ctl-related locking for protection.
2014-10-04 14:25:30 +08:00
tdata->thread_name = NULL;
}
Update brace style. Add braces around single-line blocks, and remove line breaks before function-opening braces. This resolves #537.
2017-01-16 08:56:30 +08:00
if (strlen(s) > 0) {
Implement/test/fix prof-related mallctl's. Implement/test/fix the opt.prof_thread_active_init, prof.thread_active_init, and thread.prof.active mallctl's. Test/fix the thread.prof.name mallctl. Refactor opt_prof_active to be read-only and move mutable state into the prof_active variable. Stop leaning on ctl-related locking for protection.
2014-10-04 14:25:30 +08:00
tdata->thread_name = s;
Update brace style. Add braces around single-line blocks, and remove line breaks before function-opening braces. This resolves #537.
2017-01-16 08:56:30 +08:00
}
Remove extraneous parens around return arguments. This resolves #540.
2017-01-20 10:15:45 +08:00
return 0;
Implement per thread heap profiling. Rename data structures (prof_thr_cnt_t-->prof_tctx_t, prof_ctx_t-->prof_gctx_t), and convert to storing a prof_tctx_t for sampled objects. Convert PROF_ALLOC_PREP() to prof_alloc_prep(), since precise backtrace depth within jemalloc functions is no longer an issue (pprof prunes irrelevant frames). Implement mallctl's: - prof.reset implements full sample data reset, and optional change of sample interval. - prof.lg_sample reads the current sample interval (opt.lg_prof_sample was the permanent source of truth prior to prof.reset). - thread.prof.name provides naming capability for threads within heap profile dumps. - thread.prof.active makes it possible to activate/deactivate heap profiling for individual threads. Modify the heap dump files to contain per thread heap profile data. This change is incompatible with the existing pprof, which will require enhancements to read and process the enriched data.
2014-08-19 07:22:13 +08:00
}
bool
Update brace style. Add braces around single-line blocks, and remove line breaks before function-opening braces. This resolves #537.
2017-01-16 08:56:30 +08:00
prof_thread_active_get(tsd_t *tsd) {
Suppress tdata creation in reentrancy This change suppresses tdata initialization and prof sample threshold update in interrupting malloc calls. Interrupting calls have no need for tdata. Delaying tdata creation aligns better with our lazy tdata creation principle, and it also helps us gain control back from interrupting calls more quickly and reduces any risk of delegating tdata creation to an interrupting call.
2019-10-04 04:01:12 +08:00
assert(tsd_reentrancy_level_get(tsd) == 0);
Convert all tsd variables to reside in a single tsd structure.
2014-09-23 12:09:23 +08:00
prof_tdata_t *tdata;
tdata = prof_tdata_get(tsd, true);
Update brace style. Add braces around single-line blocks, and remove line breaks before function-opening braces. This resolves #537.
2017-01-16 08:56:30 +08:00
if (tdata == NULL) {
Remove extraneous parens around return arguments. This resolves #540.
2017-01-20 10:15:45 +08:00
return false;
Update brace style. Add braces around single-line blocks, and remove line breaks before function-opening braces. This resolves #537.
2017-01-16 08:56:30 +08:00
}
Remove extraneous parens around return arguments. This resolves #540.
2017-01-20 10:15:45 +08:00
return tdata->active;
Implement per thread heap profiling. Rename data structures (prof_thr_cnt_t-->prof_tctx_t, prof_ctx_t-->prof_gctx_t), and convert to storing a prof_tctx_t for sampled objects. Convert PROF_ALLOC_PREP() to prof_alloc_prep(), since precise backtrace depth within jemalloc functions is no longer an issue (pprof prunes irrelevant frames). Implement mallctl's: - prof.reset implements full sample data reset, and optional change of sample interval. - prof.lg_sample reads the current sample interval (opt.lg_prof_sample was the permanent source of truth prior to prof.reset). - thread.prof.name provides naming capability for threads within heap profile dumps. - thread.prof.active makes it possible to activate/deactivate heap profiling for individual threads. Modify the heap dump files to contain per thread heap profile data. This change is incompatible with the existing pprof, which will require enhancements to read and process the enriched data.
2014-08-19 07:22:13 +08:00
}
bool
Update brace style. Add braces around single-line blocks, and remove line breaks before function-opening braces. This resolves #537.
2017-01-16 08:56:30 +08:00
prof_thread_active_set(tsd_t *tsd, bool active) {
Suppress tdata creation in reentrancy This change suppresses tdata initialization and prof sample threshold update in interrupting malloc calls. Interrupting calls have no need for tdata. Delaying tdata creation aligns better with our lazy tdata creation principle, and it also helps us gain control back from interrupting calls more quickly and reduces any risk of delegating tdata creation to an interrupting call.
2019-10-04 04:01:12 +08:00
assert(tsd_reentrancy_level_get(tsd) == 0);
Implement per thread heap profiling. Rename data structures (prof_thr_cnt_t-->prof_tctx_t, prof_ctx_t-->prof_gctx_t), and convert to storing a prof_tctx_t for sampled objects. Convert PROF_ALLOC_PREP() to prof_alloc_prep(), since precise backtrace depth within jemalloc functions is no longer an issue (pprof prunes irrelevant frames). Implement mallctl's: - prof.reset implements full sample data reset, and optional change of sample interval. - prof.lg_sample reads the current sample interval (opt.lg_prof_sample was the permanent source of truth prior to prof.reset). - thread.prof.name provides naming capability for threads within heap profile dumps. - thread.prof.active makes it possible to activate/deactivate heap profiling for individual threads. Modify the heap dump files to contain per thread heap profile data. This change is incompatible with the existing pprof, which will require enhancements to read and process the enriched data.
2014-08-19 07:22:13 +08:00
prof_tdata_t *tdata;
Convert all tsd variables to reside in a single tsd structure.
2014-09-23 12:09:23 +08:00
tdata = prof_tdata_get(tsd, true);
Update brace style. Add braces around single-line blocks, and remove line breaks before function-opening braces. This resolves #537.
2017-01-16 08:56:30 +08:00
if (tdata == NULL) {
Remove extraneous parens around return arguments. This resolves #540.
2017-01-20 10:15:45 +08:00
return true;
Update brace style. Add braces around single-line blocks, and remove line breaks before function-opening braces. This resolves #537.
2017-01-16 08:56:30 +08:00
}
Implement per thread heap profiling. Rename data structures (prof_thr_cnt_t-->prof_tctx_t, prof_ctx_t-->prof_gctx_t), and convert to storing a prof_tctx_t for sampled objects. Convert PROF_ALLOC_PREP() to prof_alloc_prep(), since precise backtrace depth within jemalloc functions is no longer an issue (pprof prunes irrelevant frames). Implement mallctl's: - prof.reset implements full sample data reset, and optional change of sample interval. - prof.lg_sample reads the current sample interval (opt.lg_prof_sample was the permanent source of truth prior to prof.reset). - thread.prof.name provides naming capability for threads within heap profile dumps. - thread.prof.active makes it possible to activate/deactivate heap profiling for individual threads. Modify the heap dump files to contain per thread heap profile data. This change is incompatible with the existing pprof, which will require enhancements to read and process the enriched data.
2014-08-19 07:22:13 +08:00
tdata->active = active;
Remove extraneous parens around return arguments. This resolves #540.
2017-01-20 10:15:45 +08:00
return false;
Implement per thread heap profiling. Rename data structures (prof_thr_cnt_t-->prof_tctx_t, prof_ctx_t-->prof_gctx_t), and convert to storing a prof_tctx_t for sampled objects. Convert PROF_ALLOC_PREP() to prof_alloc_prep(), since precise backtrace depth within jemalloc functions is no longer an issue (pprof prunes irrelevant frames). Implement mallctl's: - prof.reset implements full sample data reset, and optional change of sample interval. - prof.lg_sample reads the current sample interval (opt.lg_prof_sample was the permanent source of truth prior to prof.reset). - thread.prof.name provides naming capability for threads within heap profile dumps. - thread.prof.active makes it possible to activate/deactivate heap profiling for individual threads. Modify the heap dump files to contain per thread heap profile data. This change is incompatible with the existing pprof, which will require enhancements to read and process the enriched data.
2014-08-19 07:22:13 +08:00
}
Implement/test/fix prof-related mallctl's. Implement/test/fix the opt.prof_thread_active_init, prof.thread_active_init, and thread.prof.active mallctl's. Test/fix the thread.prof.name mallctl. Refactor opt_prof_active to be read-only and move mutable state into the prof_active variable. Stop leaning on ctl-related locking for protection.
2014-10-04 14:25:30 +08:00
bool
Update brace style. Add braces around single-line blocks, and remove line breaks before function-opening braces. This resolves #537.
2017-01-16 08:56:30 +08:00
prof_thread_active_init_get(tsdn_t *tsdn) {
Implement/test/fix prof-related mallctl's. Implement/test/fix the opt.prof_thread_active_init, prof.thread_active_init, and thread.prof.active mallctl's. Test/fix the thread.prof.name mallctl. Refactor opt_prof_active to be read-only and move mutable state into the prof_active variable. Stop leaning on ctl-related locking for protection.
2014-10-04 14:25:30 +08:00
bool active_init;
Resolve bootstrapping issues when embedded in FreeBSD libc. b2c0d6322d2307458ae2b28545f8a5c9903d7ef5 (Add witness, a simple online locking validator.) caused a broad propagation of tsd throughout the internal API, but tsd_fetch() was designed to fail prior to tsd bootstrapping. Fix this by splitting tsd_t into non-nullable tsd_t and nullable tsdn_t, and modifying all internal APIs that do not critically rely on tsd to take nullable pointers. Furthermore, add the tsd_booted_get() function so that tsdn_fetch() can probe whether tsd bootstrapping is complete and return NULL if not. All dangerous conversions of nullable pointers are tsdn_tsd() calls that assert-fail on invalid conversion.
2016-05-11 13:21:10 +08:00
malloc_mutex_lock(tsdn, &prof_thread_active_init_mtx);
Implement/test/fix prof-related mallctl's. Implement/test/fix the opt.prof_thread_active_init, prof.thread_active_init, and thread.prof.active mallctl's. Test/fix the thread.prof.name mallctl. Refactor opt_prof_active to be read-only and move mutable state into the prof_active variable. Stop leaning on ctl-related locking for protection.
2014-10-04 14:25:30 +08:00
active_init = prof_thread_active_init;
Resolve bootstrapping issues when embedded in FreeBSD libc. b2c0d6322d2307458ae2b28545f8a5c9903d7ef5 (Add witness, a simple online locking validator.) caused a broad propagation of tsd throughout the internal API, but tsd_fetch() was designed to fail prior to tsd bootstrapping. Fix this by splitting tsd_t into non-nullable tsd_t and nullable tsdn_t, and modifying all internal APIs that do not critically rely on tsd to take nullable pointers. Furthermore, add the tsd_booted_get() function so that tsdn_fetch() can probe whether tsd bootstrapping is complete and return NULL if not. All dangerous conversions of nullable pointers are tsdn_tsd() calls that assert-fail on invalid conversion.
2016-05-11 13:21:10 +08:00
malloc_mutex_unlock(tsdn, &prof_thread_active_init_mtx);
Remove extraneous parens around return arguments. This resolves #540.
2017-01-20 10:15:45 +08:00
return active_init;
Implement/test/fix prof-related mallctl's. Implement/test/fix the opt.prof_thread_active_init, prof.thread_active_init, and thread.prof.active mallctl's. Test/fix the thread.prof.name mallctl. Refactor opt_prof_active to be read-only and move mutable state into the prof_active variable. Stop leaning on ctl-related locking for protection.
2014-10-04 14:25:30 +08:00
}
bool
Update brace style. Add braces around single-line blocks, and remove line breaks before function-opening braces. This resolves #537.
2017-01-16 08:56:30 +08:00
prof_thread_active_init_set(tsdn_t *tsdn, bool active_init) {
Implement/test/fix prof-related mallctl's. Implement/test/fix the opt.prof_thread_active_init, prof.thread_active_init, and thread.prof.active mallctl's. Test/fix the thread.prof.name mallctl. Refactor opt_prof_active to be read-only and move mutable state into the prof_active variable. Stop leaning on ctl-related locking for protection.
2014-10-04 14:25:30 +08:00
bool active_init_old;
Resolve bootstrapping issues when embedded in FreeBSD libc. b2c0d6322d2307458ae2b28545f8a5c9903d7ef5 (Add witness, a simple online locking validator.) caused a broad propagation of tsd throughout the internal API, but tsd_fetch() was designed to fail prior to tsd bootstrapping. Fix this by splitting tsd_t into non-nullable tsd_t and nullable tsdn_t, and modifying all internal APIs that do not critically rely on tsd to take nullable pointers. Furthermore, add the tsd_booted_get() function so that tsdn_fetch() can probe whether tsd bootstrapping is complete and return NULL if not. All dangerous conversions of nullable pointers are tsdn_tsd() calls that assert-fail on invalid conversion.
2016-05-11 13:21:10 +08:00
malloc_mutex_lock(tsdn, &prof_thread_active_init_mtx);
Implement/test/fix prof-related mallctl's. Implement/test/fix the opt.prof_thread_active_init, prof.thread_active_init, and thread.prof.active mallctl's. Test/fix the thread.prof.name mallctl. Refactor opt_prof_active to be read-only and move mutable state into the prof_active variable. Stop leaning on ctl-related locking for protection.
2014-10-04 14:25:30 +08:00
active_init_old = prof_thread_active_init;
prof_thread_active_init = active_init;
Resolve bootstrapping issues when embedded in FreeBSD libc. b2c0d6322d2307458ae2b28545f8a5c9903d7ef5 (Add witness, a simple online locking validator.) caused a broad propagation of tsd throughout the internal API, but tsd_fetch() was designed to fail prior to tsd bootstrapping. Fix this by splitting tsd_t into non-nullable tsd_t and nullable tsdn_t, and modifying all internal APIs that do not critically rely on tsd to take nullable pointers. Furthermore, add the tsd_booted_get() function so that tsdn_fetch() can probe whether tsd bootstrapping is complete and return NULL if not. All dangerous conversions of nullable pointers are tsdn_tsd() calls that assert-fail on invalid conversion.
2016-05-11 13:21:10 +08:00
malloc_mutex_unlock(tsdn, &prof_thread_active_init_mtx);
Remove extraneous parens around return arguments. This resolves #540.
2017-01-20 10:15:45 +08:00
return active_init_old;
Implement/test/fix prof-related mallctl's. Implement/test/fix the opt.prof_thread_active_init, prof.thread_active_init, and thread.prof.active mallctl's. Test/fix the thread.prof.name mallctl. Refactor opt_prof_active to be read-only and move mutable state into the prof_active variable. Stop leaning on ctl-related locking for protection.
2014-10-04 14:25:30 +08:00
}
Implement the prof.gdump mallctl. This feature makes it possible to toggle the gdump feature on/off during program execution, whereas the the opt.prof_dump mallctl value can only be set during program startup. This resolves #72.
2015-01-26 13:16:57 +08:00
bool
Update brace style. Add braces around single-line blocks, and remove line breaks before function-opening braces. This resolves #537.
2017-01-16 08:56:30 +08:00
prof_gdump_get(tsdn_t *tsdn) {
Implement the prof.gdump mallctl. This feature makes it possible to toggle the gdump feature on/off during program execution, whereas the the opt.prof_dump mallctl value can only be set during program startup. This resolves #72.
2015-01-26 13:16:57 +08:00
bool prof_gdump_current;
Resolve bootstrapping issues when embedded in FreeBSD libc. b2c0d6322d2307458ae2b28545f8a5c9903d7ef5 (Add witness, a simple online locking validator.) caused a broad propagation of tsd throughout the internal API, but tsd_fetch() was designed to fail prior to tsd bootstrapping. Fix this by splitting tsd_t into non-nullable tsd_t and nullable tsdn_t, and modifying all internal APIs that do not critically rely on tsd to take nullable pointers. Furthermore, add the tsd_booted_get() function so that tsdn_fetch() can probe whether tsd bootstrapping is complete and return NULL if not. All dangerous conversions of nullable pointers are tsdn_tsd() calls that assert-fail on invalid conversion.
2016-05-11 13:21:10 +08:00
malloc_mutex_lock(tsdn, &prof_gdump_mtx);
Implement the prof.gdump mallctl. This feature makes it possible to toggle the gdump feature on/off during program execution, whereas the the opt.prof_dump mallctl value can only be set during program startup. This resolves #72.
2015-01-26 13:16:57 +08:00
prof_gdump_current = prof_gdump_val;
Resolve bootstrapping issues when embedded in FreeBSD libc. b2c0d6322d2307458ae2b28545f8a5c9903d7ef5 (Add witness, a simple online locking validator.) caused a broad propagation of tsd throughout the internal API, but tsd_fetch() was designed to fail prior to tsd bootstrapping. Fix this by splitting tsd_t into non-nullable tsd_t and nullable tsdn_t, and modifying all internal APIs that do not critically rely on tsd to take nullable pointers. Furthermore, add the tsd_booted_get() function so that tsdn_fetch() can probe whether tsd bootstrapping is complete and return NULL if not. All dangerous conversions of nullable pointers are tsdn_tsd() calls that assert-fail on invalid conversion.
2016-05-11 13:21:10 +08:00
malloc_mutex_unlock(tsdn, &prof_gdump_mtx);
Remove extraneous parens around return arguments. This resolves #540.
2017-01-20 10:15:45 +08:00
return prof_gdump_current;
Implement the prof.gdump mallctl. This feature makes it possible to toggle the gdump feature on/off during program execution, whereas the the opt.prof_dump mallctl value can only be set during program startup. This resolves #72.
2015-01-26 13:16:57 +08:00
}
bool
Update brace style. Add braces around single-line blocks, and remove line breaks before function-opening braces. This resolves #537.
2017-01-16 08:56:30 +08:00
prof_gdump_set(tsdn_t *tsdn, bool gdump) {
Implement the prof.gdump mallctl. This feature makes it possible to toggle the gdump feature on/off during program execution, whereas the the opt.prof_dump mallctl value can only be set during program startup. This resolves #72.
2015-01-26 13:16:57 +08:00
bool prof_gdump_old;
Resolve bootstrapping issues when embedded in FreeBSD libc. b2c0d6322d2307458ae2b28545f8a5c9903d7ef5 (Add witness, a simple online locking validator.) caused a broad propagation of tsd throughout the internal API, but tsd_fetch() was designed to fail prior to tsd bootstrapping. Fix this by splitting tsd_t into non-nullable tsd_t and nullable tsdn_t, and modifying all internal APIs that do not critically rely on tsd to take nullable pointers. Furthermore, add the tsd_booted_get() function so that tsdn_fetch() can probe whether tsd bootstrapping is complete and return NULL if not. All dangerous conversions of nullable pointers are tsdn_tsd() calls that assert-fail on invalid conversion.
2016-05-11 13:21:10 +08:00
malloc_mutex_lock(tsdn, &prof_gdump_mtx);
Implement the prof.gdump mallctl. This feature makes it possible to toggle the gdump feature on/off during program execution, whereas the the opt.prof_dump mallctl value can only be set during program startup. This resolves #72.
2015-01-26 13:16:57 +08:00
prof_gdump_old = prof_gdump_val;
prof_gdump_val = gdump;
Resolve bootstrapping issues when embedded in FreeBSD libc. b2c0d6322d2307458ae2b28545f8a5c9903d7ef5 (Add witness, a simple online locking validator.) caused a broad propagation of tsd throughout the internal API, but tsd_fetch() was designed to fail prior to tsd bootstrapping. Fix this by splitting tsd_t into non-nullable tsd_t and nullable tsdn_t, and modifying all internal APIs that do not critically rely on tsd to take nullable pointers. Furthermore, add the tsd_booted_get() function so that tsdn_fetch() can probe whether tsd bootstrapping is complete and return NULL if not. All dangerous conversions of nullable pointers are tsdn_tsd() calls that assert-fail on invalid conversion.
2016-05-11 13:21:10 +08:00
malloc_mutex_unlock(tsdn, &prof_gdump_mtx);
Remove extraneous parens around return arguments. This resolves #540.
2017-01-20 10:15:45 +08:00
return prof_gdump_old;
Implement the prof.gdump mallctl. This feature makes it possible to toggle the gdump feature on/off during program execution, whereas the the opt.prof_dump mallctl value can only be set during program startup. This resolves #72.
2015-01-26 13:16:57 +08:00
}
Implement allocation profiling and leack checking. Add the --enable-prof and --enable-prof-libunwind configure options. Add the B/b, F/f, I/i, L/l, and U/u JEMALLOC_OPTIONS. Interval-based profile dump triggering is not yet implemented. Add supporting generic code: * Add memory barriers. * Add prn (LCG PRNG). * Add hash (Murmur hash function). * Add ckh (cuckoo hash tables).
2010-02-11 02:37:56 +08:00
void
Update brace style. Add braces around single-line blocks, and remove line breaks before function-opening braces. This resolves #537.
2017-01-16 08:56:30 +08:00
prof_boot0(void) {
Reduce cpp conditional logic complexity. Convert configuration-related cpp conditional logic to use static constant variables, e.g.: #ifdef JEMALLOC_DEBUG [...] #endif becomes: if (config_debug) { [...] } The advantage is clearer, more concise code. The main disadvantage is that data structures no longer have conditionally defined fields, so they pay the cost of all fields regardless of whether they are used. In practice, this is only a minor concern; config_stats will go away in an upcoming change, and config_prof is the only other major feature that depends on more than a few special-purpose fields.
2012-02-11 12:22:09 +08:00
cassert(config_prof);
Replace JEMALLOC_OPTIONS with MALLOC_CONF. Replace the single-character run-time flags with key/value pairs, which can be set via the malloc_conf global, /etc/malloc.conf, and the MALLOC_CONF environment variable. Replace the JEMALLOC_PROF_PREFIX environment variable with the "opt.prof_prefix" option. Replace umax2s() with u2s().
2010-10-24 09:37:06 +08:00
memcpy(opt_prof_prefix, PROF_PREFIX_DEFAULT,
sizeof(PROF_PREFIX_DEFAULT));
}
void
Update brace style. Add braces around single-line blocks, and remove line breaks before function-opening braces. This resolves #537.
2017-01-16 08:56:30 +08:00
prof_boot1(void) {
Reduce cpp conditional logic complexity. Convert configuration-related cpp conditional logic to use static constant variables, e.g.: #ifdef JEMALLOC_DEBUG [...] #endif becomes: if (config_debug) { [...] } The advantage is clearer, more concise code. The main disadvantage is that data structures no longer have conditionally defined fields, so they pay the cost of all fields regardless of whether they are used. In practice, this is only a minor concern; config_stats will go away in an upcoming change, and config_prof is the only other major feature that depends on more than a few special-purpose fields.
2012-02-11 12:22:09 +08:00
cassert(config_prof);
Implement allocation profiling and leack checking. Add the --enable-prof and --enable-prof-libunwind configure options. Add the B/b, F/f, I/i, L/l, and U/u JEMALLOC_OPTIONS. Interval-based profile dump triggering is not yet implemented. Add supporting generic code: * Add memory barriers. * Add prn (LCG PRNG). * Add hash (Murmur hash function). * Add ckh (cuckoo hash tables).
2010-02-11 02:37:56 +08:00
/*
Remove support for non-prof-promote heap profiling metadata. Make promotion of sampled small objects to large objects mandatory, so that profiling metadata can always be stored in the chunk map, rather than requiring one pointer per small region in each small-region page run. In practice the non-prof-promote code was only useful when using jemalloc to track all objects and report them as leaks at program exit. However, Valgrind is at least as good a tool for this particular use case. Furthermore, the non-prof-promote code is getting in the way of some optimizations that will make heap profiling much cheaper for the predominant use case (sampling a small representative proportion of all allocations).
2014-04-12 05:24:51 +08:00
* opt_prof must be in its final state before any arenas are
* initialized, so this function must be executed early.
Implement allocation profiling and leack checking. Add the --enable-prof and --enable-prof-libunwind configure options. Add the B/b, F/f, I/i, L/l, and U/u JEMALLOC_OPTIONS. Interval-based profile dump triggering is not yet implemented. Add supporting generic code: * Add memory barriers. * Add prn (LCG PRNG). * Add hash (Murmur hash function). * Add ckh (cuckoo hash tables).
2010-02-11 02:37:56 +08:00
*/
Convert to uniform style: cond == false --> !cond
2014-10-04 01:16:09 +08:00
if (opt_prof_leak && !opt_prof) {
Implement allocation profiling and leack checking. Add the --enable-prof and --enable-prof-libunwind configure options. Add the B/b, F/f, I/i, L/l, and U/u JEMALLOC_OPTIONS. Interval-based profile dump triggering is not yet implemented. Add supporting generic code: * Add memory barriers. * Add prn (LCG PRNG). * Add hash (Murmur hash function). * Add ckh (cuckoo hash tables).
2010-02-11 02:37:56 +08:00
/*
* Enable opt_prof, but in such a way that profiles are never
* automatically dumped.
*/
opt_prof = true;
Replace JEMALLOC_OPTIONS with MALLOC_CONF. Replace the single-character run-time flags with key/value pairs, which can be set via the malloc_conf global, /etc/malloc.conf, and the MALLOC_CONF environment variable. Replace the JEMALLOC_PROF_PREFIX environment variable with the "opt.prof_prefix" option. Replace umax2s() with u2s().
2010-10-24 09:37:06 +08:00
opt_prof_gdump = false;
Make interval-triggered profile dumping optional. Make it possible to disable interval-triggered profile dumping, even if profiling is enabled. This is useful if the user only wants a single dump at exit, or if the application manually triggers profile dumps.
2010-04-01 08:35:51 +08:00
} else if (opt_prof) {
if (opt_lg_prof_interval >= 0) {
prof_interval = (((uint64_t)1U) <<
opt_lg_prof_interval);
Avoid arena_prof_accum()-related locking when possible. Refactor arena_prof_accum() and its callers to avoid arena locking when prof_interval is 0 (as when profiling is disabled). Reported by Ben Maurer.
2012-11-14 04:56:27 +08:00
}
Make interval-triggered profile dumping optional. Make it possible to disable interval-triggered profile dumping, even if profiling is enabled. This is useful if the user only wants a single dump at exit, or if the application manually triggers profile dumps.
2010-04-01 08:35:51 +08:00
}
Implement allocation profiling and leack checking. Add the --enable-prof and --enable-prof-libunwind configure options. Add the B/b, F/f, I/i, L/l, and U/u JEMALLOC_OPTIONS. Interval-based profile dump triggering is not yet implemented. Add supporting generic code: * Add memory barriers. * Add prn (LCG PRNG). * Add hash (Murmur hash function). * Add ckh (cuckoo hash tables).
2010-02-11 02:37:56 +08:00
}
bool
Update brace style. Add braces around single-line blocks, and remove line breaks before function-opening braces. This resolves #537.
2017-01-16 08:56:30 +08:00
prof_boot2(tsd_t *tsd) {
Reduce cpp conditional logic complexity. Convert configuration-related cpp conditional logic to use static constant variables, e.g.: #ifdef JEMALLOC_DEBUG [...] #endif becomes: if (config_debug) { [...] } The advantage is clearer, more concise code. The main disadvantage is that data structures no longer have conditionally defined fields, so they pay the cost of all fields regardless of whether they are used. In practice, this is only a minor concern; config_stats will go away in an upcoming change, and config_prof is the only other major feature that depends on more than a few special-purpose fields.
2012-02-11 12:22:09 +08:00
cassert(config_prof);
Implement allocation profiling and leack checking. Add the --enable-prof and --enable-prof-libunwind configure options. Add the B/b, F/f, I/i, L/l, and U/u JEMALLOC_OPTIONS. Interval-based profile dump triggering is not yet implemented. Add supporting generic code: * Add memory barriers. * Add prn (LCG PRNG). * Add hash (Murmur hash function). * Add ckh (cuckoo hash tables).
2010-02-11 02:37:56 +08:00
if (opt_prof) {
Remove ephemeral mutexes. Remove ephemeral mutexes from the prof machinery, and remove malloc_mutex_destroy(). This simplifies mutex management on systems that call malloc()/free() inside pthread_mutex_{create,destroy}(). Add atomic_*_u() for operation on unsigned values. Fix prof_printf() to call malloc_vsnprintf() rather than malloc_snprintf().
2012-03-24 09:05:51 +08:00
unsigned i;
Implement per thread heap profiling. Rename data structures (prof_thr_cnt_t-->prof_tctx_t, prof_ctx_t-->prof_gctx_t), and convert to storing a prof_tctx_t for sampled objects. Convert PROF_ALLOC_PREP() to prof_alloc_prep(), since precise backtrace depth within jemalloc functions is no longer an issue (pprof prunes irrelevant frames). Implement mallctl's: - prof.reset implements full sample data reset, and optional change of sample interval. - prof.lg_sample reads the current sample interval (opt.lg_prof_sample was the permanent source of truth prior to prof.reset). - thread.prof.name provides naming capability for threads within heap profile dumps. - thread.prof.active makes it possible to activate/deactivate heap profiling for individual threads. Modify the heap dump files to contain per thread heap profile data. This change is incompatible with the existing pprof, which will require enhancements to read and process the enriched data.
2014-08-19 07:22:13 +08:00
lg_prof_sample = opt_lg_prof_sample;
Implement/test/fix prof-related mallctl's. Implement/test/fix the opt.prof_thread_active_init, prof.thread_active_init, and thread.prof.active mallctl's. Test/fix the thread.prof.name mallctl. Refactor opt_prof_active to be read-only and move mutable state into the prof_active variable. Stop leaning on ctl-related locking for protection.
2014-10-04 14:25:30 +08:00
prof_active = opt_prof_active;
Add witness, a simple online locking validator. This resolves #358.
2016-04-14 14:36:15 +08:00
if (malloc_mutex_init(&prof_active_mtx, "prof_active",
Allow mutexes to take a lock ordering enum at construction. This lets us specify whether and how mutexes of the same rank are allowed to be acquired. Currently, we only allow two polices (only a single mutex at a given rank at a time, and mutexes acquired in ascending order), but we can plausibly allow more (e.g. the "release uncontended mutexes before blocking").
2017-05-16 06:38:15 +08:00
WITNESS_RANK_PROF_ACTIVE, malloc_mutex_rank_exclusive)) {
Remove extraneous parens around return arguments. This resolves #540.
2017-01-20 10:15:45 +08:00
return true;
Update brace style. Add braces around single-line blocks, and remove line breaks before function-opening braces. This resolves #537.
2017-01-16 08:56:30 +08:00
}
Implement/test/fix prof-related mallctl's. Implement/test/fix the opt.prof_thread_active_init, prof.thread_active_init, and thread.prof.active mallctl's. Test/fix the thread.prof.name mallctl. Refactor opt_prof_active to be read-only and move mutable state into the prof_active variable. Stop leaning on ctl-related locking for protection.
2014-10-04 14:25:30 +08:00
Implement the prof.gdump mallctl. This feature makes it possible to toggle the gdump feature on/off during program execution, whereas the the opt.prof_dump mallctl value can only be set during program startup. This resolves #72.
2015-01-26 13:16:57 +08:00
prof_gdump_val = opt_prof_gdump;
Add witness, a simple online locking validator. This resolves #358.
2016-04-14 14:36:15 +08:00
if (malloc_mutex_init(&prof_gdump_mtx, "prof_gdump",
Allow mutexes to take a lock ordering enum at construction. This lets us specify whether and how mutexes of the same rank are allowed to be acquired. Currently, we only allow two polices (only a single mutex at a given rank at a time, and mutexes acquired in ascending order), but we can plausibly allow more (e.g. the "release uncontended mutexes before blocking").
2017-05-16 06:38:15 +08:00
WITNESS_RANK_PROF_GDUMP, malloc_mutex_rank_exclusive)) {
Remove extraneous parens around return arguments. This resolves #540.
2017-01-20 10:15:45 +08:00
return true;
Update brace style. Add braces around single-line blocks, and remove line breaks before function-opening braces. This resolves #537.
2017-01-16 08:56:30 +08:00
}
Implement the prof.gdump mallctl. This feature makes it possible to toggle the gdump feature on/off during program execution, whereas the the opt.prof_dump mallctl value can only be set during program startup. This resolves #72.
2015-01-26 13:16:57 +08:00
Implement/test/fix prof-related mallctl's. Implement/test/fix the opt.prof_thread_active_init, prof.thread_active_init, and thread.prof.active mallctl's. Test/fix the thread.prof.name mallctl. Refactor opt_prof_active to be read-only and move mutable state into the prof_active variable. Stop leaning on ctl-related locking for protection.
2014-10-04 14:25:30 +08:00
prof_thread_active_init = opt_prof_thread_active_init;
Add witness, a simple online locking validator. This resolves #358.
2016-04-14 14:36:15 +08:00
if (malloc_mutex_init(&prof_thread_active_init_mtx,
"prof_thread_active_init",
Allow mutexes to take a lock ordering enum at construction. This lets us specify whether and how mutexes of the same rank are allowed to be acquired. Currently, we only allow two polices (only a single mutex at a given rank at a time, and mutexes acquired in ascending order), but we can plausibly allow more (e.g. the "release uncontended mutexes before blocking").
2017-05-16 06:38:15 +08:00
WITNESS_RANK_PROF_THREAD_ACTIVE_INIT,
malloc_mutex_rank_exclusive)) {
Remove extraneous parens around return arguments. This resolves #540.
2017-01-20 10:15:45 +08:00
return true;
Update brace style. Add braces around single-line blocks, and remove line breaks before function-opening braces. This resolves #537.
2017-01-16 08:56:30 +08:00
}
Implement/test/fix prof-related mallctl's. Implement/test/fix the opt.prof_thread_active_init, prof.thread_active_init, and thread.prof.active mallctl's. Test/fix the thread.prof.name mallctl. Refactor opt_prof_active to be read-only and move mutable state into the prof_active variable. Stop leaning on ctl-related locking for protection.
2014-10-04 14:25:30 +08:00
Refactor profiling Refactored core profiling codebase into two logical parts: (a) `prof_data.c`: core internal data structure managing & dumping; (b) `prof.c`: mutexes & outward-facing APIs. Some internal functions had to be exposed out, but there are not that many of them if the modularization is (hopefully) clean enough.
2019-07-18 06:52:50 +08:00
if (prof_data_init(tsd)) {
Remove extraneous parens around return arguments. This resolves #540.
2017-01-20 10:15:45 +08:00
return true;
Update brace style. Add braces around single-line blocks, and remove line breaks before function-opening braces. This resolves #537.
2017-01-16 08:56:30 +08:00
}
Refactor profiling Refactored core profiling codebase into two logical parts: (a) `prof_data.c`: core internal data structure managing & dumping; (b) `prof.c`: mutexes & outward-facing APIs. Some internal functions had to be exposed out, but there are not that many of them if the modularization is (hopefully) clean enough.
2019-07-18 06:52:50 +08:00
Add witness, a simple online locking validator. This resolves #358.
2016-04-14 14:36:15 +08:00
if (malloc_mutex_init(&bt2gctx_mtx, "prof_bt2gctx",
Allow mutexes to take a lock ordering enum at construction. This lets us specify whether and how mutexes of the same rank are allowed to be acquired. Currently, we only allow two polices (only a single mutex at a given rank at a time, and mutexes acquired in ascending order), but we can plausibly allow more (e.g. the "release uncontended mutexes before blocking").
2017-05-16 06:38:15 +08:00
WITNESS_RANK_PROF_BT2GCTX, malloc_mutex_rank_exclusive)) {
Remove extraneous parens around return arguments. This resolves #540.
2017-01-20 10:15:45 +08:00
return true;
Update brace style. Add braces around single-line blocks, and remove line breaks before function-opening braces. This resolves #537.
2017-01-16 08:56:30 +08:00
}
Implement allocation profiling and leack checking. Add the --enable-prof and --enable-prof-libunwind configure options. Add the B/b, F/f, I/i, L/l, and U/u JEMALLOC_OPTIONS. Interval-based profile dump triggering is not yet implemented. Add supporting generic code: * Add memory barriers. * Add prn (LCG PRNG). * Add hash (Murmur hash function). * Add ckh (cuckoo hash tables).
2010-02-11 02:37:56 +08:00
Add witness, a simple online locking validator. This resolves #358.
2016-04-14 14:36:15 +08:00
if (malloc_mutex_init(&tdatas_mtx, "prof_tdatas",
Allow mutexes to take a lock ordering enum at construction. This lets us specify whether and how mutexes of the same rank are allowed to be acquired. Currently, we only allow two polices (only a single mutex at a given rank at a time, and mutexes acquired in ascending order), but we can plausibly allow more (e.g. the "release uncontended mutexes before blocking").
2017-05-16 06:38:15 +08:00
WITNESS_RANK_PROF_TDATAS, malloc_mutex_rank_exclusive)) {
Remove extraneous parens around return arguments. This resolves #540.
2017-01-20 10:15:45 +08:00
return true;
Update brace style. Add braces around single-line blocks, and remove line breaks before function-opening braces. This resolves #537.
2017-01-16 08:56:30 +08:00
}
Implement per thread heap profiling. Rename data structures (prof_thr_cnt_t-->prof_tctx_t, prof_ctx_t-->prof_gctx_t), and convert to storing a prof_tctx_t for sampled objects. Convert PROF_ALLOC_PREP() to prof_alloc_prep(), since precise backtrace depth within jemalloc functions is no longer an issue (pprof prunes irrelevant frames). Implement mallctl's: - prof.reset implements full sample data reset, and optional change of sample interval. - prof.lg_sample reads the current sample interval (opt.lg_prof_sample was the permanent source of truth prior to prof.reset). - thread.prof.name provides naming capability for threads within heap profile dumps. - thread.prof.active makes it possible to activate/deactivate heap profiling for individual threads. Modify the heap dump files to contain per thread heap profile data. This change is incompatible with the existing pprof, which will require enhancements to read and process the enriched data.
2014-08-19 07:22:13 +08:00
next_thr_uid = 0;
Add witness, a simple online locking validator. This resolves #358.
2016-04-14 14:36:15 +08:00
if (malloc_mutex_init(&next_thr_uid_mtx, "prof_next_thr_uid",
Allow mutexes to take a lock ordering enum at construction. This lets us specify whether and how mutexes of the same rank are allowed to be acquired. Currently, we only allow two polices (only a single mutex at a given rank at a time, and mutexes acquired in ascending order), but we can plausibly allow more (e.g. the "release uncontended mutexes before blocking").
2017-05-16 06:38:15 +08:00
WITNESS_RANK_PROF_NEXT_THR_UID, malloc_mutex_rank_exclusive)) {
Remove extraneous parens around return arguments. This resolves #540.
2017-01-20 10:15:45 +08:00
return true;
Update brace style. Add braces around single-line blocks, and remove line breaks before function-opening braces. This resolves #537.
2017-01-16 08:56:30 +08:00
}
Implement per thread heap profiling. Rename data structures (prof_thr_cnt_t-->prof_tctx_t, prof_ctx_t-->prof_gctx_t), and convert to storing a prof_tctx_t for sampled objects. Convert PROF_ALLOC_PREP() to prof_alloc_prep(), since precise backtrace depth within jemalloc functions is no longer an issue (pprof prunes irrelevant frames). Implement mallctl's: - prof.reset implements full sample data reset, and optional change of sample interval. - prof.lg_sample reads the current sample interval (opt.lg_prof_sample was the permanent source of truth prior to prof.reset). - thread.prof.name provides naming capability for threads within heap profile dumps. - thread.prof.active makes it possible to activate/deactivate heap profiling for individual threads. Modify the heap dump files to contain per thread heap profile data. This change is incompatible with the existing pprof, which will require enhancements to read and process the enriched data.
2014-08-19 07:22:13 +08:00
Rename "prof_dump_seq_mtx" to "prof_dump_filename_mtx".
2019-09-10 11:04:18 +08:00
if (malloc_mutex_init(&prof_dump_filename_mtx, "prof_dump_filename",
WITNESS_RANK_PROF_DUMP_FILENAME, malloc_mutex_rank_exclusive)) {
Remove extraneous parens around return arguments. This resolves #540.
2017-01-20 10:15:45 +08:00
return true;
Update brace style. Add braces around single-line blocks, and remove line breaks before function-opening braces. This resolves #537.
2017-01-16 08:56:30 +08:00
}
Add witness, a simple online locking validator. This resolves #358.
2016-04-14 14:36:15 +08:00
if (malloc_mutex_init(&prof_dump_mtx, "prof_dump",
Allow mutexes to take a lock ordering enum at construction. This lets us specify whether and how mutexes of the same rank are allowed to be acquired. Currently, we only allow two polices (only a single mutex at a given rank at a time, and mutexes acquired in ascending order), but we can plausibly allow more (e.g. the "release uncontended mutexes before blocking").
2017-05-16 06:38:15 +08:00
WITNESS_RANK_PROF_DUMP, malloc_mutex_rank_exclusive)) {
Remove extraneous parens around return arguments. This resolves #540.
2017-01-20 10:15:45 +08:00
return true;
Update brace style. Add braces around single-line blocks, and remove line breaks before function-opening braces. This resolves #537.
2017-01-16 08:56:30 +08:00
}
Implement allocation profiling and leack checking. Add the --enable-prof and --enable-prof-libunwind configure options. Add the B/b, F/f, I/i, L/l, and U/u JEMALLOC_OPTIONS. Interval-based profile dump triggering is not yet implemented. Add supporting generic code: * Add memory barriers. * Add prn (LCG PRNG). * Add hash (Murmur hash function). * Add ckh (cuckoo hash tables).
2010-02-11 02:37:56 +08:00
Avoid atexit(3) when possible, disable prof_final by default. atexit(3) can deadlock internally during its own initialization if jemalloc calls atexit() during jemalloc initialization. Mitigate the impact by restructuring prof initialization to avoid calling atexit() unless the registered function will actually dump a final heap profile. Additionally, disable prof_final by default so that this land mine is opt-in rather than opt-out. This resolves #144.
2014-10-09 08:57:19 +08:00
if (opt_prof_final && opt_prof_prefix[0] != '\0' &&
atexit(prof_fdump) != 0) {
Simplify malloc_message(). Rather than passing four strings to malloc_message(), malloc_write4(), and all the functions that use them, only pass one string.
2010-03-04 09:45:38 +08:00
malloc_write("<jemalloc>: Error in atexit()\n");
Update brace style. Add braces around single-line blocks, and remove line breaks before function-opening braces. This resolves #537.
2017-01-16 08:56:30 +08:00
if (opt_abort) {
Implement allocation profiling and leack checking. Add the --enable-prof and --enable-prof-libunwind configure options. Add the B/b, F/f, I/i, L/l, and U/u JEMALLOC_OPTIONS. Interval-based profile dump triggering is not yet implemented. Add supporting generic code: * Add memory barriers. * Add prn (LCG PRNG). * Add hash (Murmur hash function). * Add ckh (cuckoo hash tables).
2010-02-11 02:37:56 +08:00
abort();
Update brace style. Add braces around single-line blocks, and remove line breaks before function-opening braces. This resolves #537.
2017-01-16 08:56:30 +08:00
}
Implement allocation profiling and leack checking. Add the --enable-prof and --enable-prof-libunwind configure options. Add the B/b, F/f, I/i, L/l, and U/u JEMALLOC_OPTIONS. Interval-based profile dump triggering is not yet implemented. Add supporting generic code: * Add memory barriers. * Add prn (LCG PRNG). * Add hash (Murmur hash function). * Add ckh (cuckoo hash tables).
2010-02-11 02:37:56 +08:00
}
Remove ephemeral mutexes. Remove ephemeral mutexes from the prof machinery, and remove malloc_mutex_destroy(). This simplifies mutex management on systems that call malloc()/free() inside pthread_mutex_{create,destroy}(). Add atomic_*_u() for operation on unsigned values. Fix prof_printf() to call malloc_vsnprintf() rather than malloc_snprintf().
2012-03-24 09:05:51 +08:00
Refactor prof log Prof logging is conceptually seperate from core profiling, so split it out as a module of its own. There are a few internal functions that had to be exposed but I think it is a fair trade-off.
2019-07-13 07:37:37 +08:00
if (prof_log_init(tsd)) {
Add logging for sampled allocations - prof_opt_log flag starts logging automatically at runtime - prof_log_{start,stop} mallctl for manual control
2018-07-06 01:56:33 +08:00
return true;
}
Do not (recursively) allocate within tsd_fetch(). Refactor tsd so that tsdn_fetch() does not trigger allocation, since allocation could cause infinite recursion. This resolves #458.
2016-10-21 14:59:12 +08:00
gctx_locks = (malloc_mutex_t *)base_alloc(tsd_tsdn(tsd),
Implement per arena base allocators. Add/rename related mallctls: - Add stats.arenas.<i>.base . - Rename stats.arenas.<i>.metadata to stats.arenas.<i>.internal . - Add stats.arenas.<i>.resident . Modify the arenas.extend mallctl to take an optional (extent_hooks_t *) argument so that it is possible for all base allocations to be serviced by the specified extent hooks. This resolves #463.
2016-12-23 06:39:10 +08:00
b0get(), PROF_NCTX_LOCKS * sizeof(malloc_mutex_t),
CACHELINE);
Update brace style. Add braces around single-line blocks, and remove line breaks before function-opening braces. This resolves #537.
2017-01-16 08:56:30 +08:00
if (gctx_locks == NULL) {
Remove extraneous parens around return arguments. This resolves #540.
2017-01-20 10:15:45 +08:00
return true;
Update brace style. Add braces around single-line blocks, and remove line breaks before function-opening braces. This resolves #537.
2017-01-16 08:56:30 +08:00
}
Remove ephemeral mutexes. Remove ephemeral mutexes from the prof machinery, and remove malloc_mutex_destroy(). This simplifies mutex management on systems that call malloc()/free() inside pthread_mutex_{create,destroy}(). Add atomic_*_u() for operation on unsigned values. Fix prof_printf() to call malloc_vsnprintf() rather than malloc_snprintf().
2012-03-24 09:05:51 +08:00
for (i = 0; i < PROF_NCTX_LOCKS; i++) {
Add witness, a simple online locking validator. This resolves #358.
2016-04-14 14:36:15 +08:00
if (malloc_mutex_init(&gctx_locks[i], "prof_gctx",
Allow mutexes to take a lock ordering enum at construction. This lets us specify whether and how mutexes of the same rank are allowed to be acquired. Currently, we only allow two polices (only a single mutex at a given rank at a time, and mutexes acquired in ascending order), but we can plausibly allow more (e.g. the "release uncontended mutexes before blocking").
2017-05-16 06:38:15 +08:00
WITNESS_RANK_PROF_GCTX,
malloc_mutex_rank_exclusive)) {
Remove extraneous parens around return arguments. This resolves #540.
2017-01-20 10:15:45 +08:00
return true;
Update brace style. Add braces around single-line blocks, and remove line breaks before function-opening braces. This resolves #537.
2017-01-16 08:56:30 +08:00
}
Implement per thread heap profiling. Rename data structures (prof_thr_cnt_t-->prof_tctx_t, prof_ctx_t-->prof_gctx_t), and convert to storing a prof_tctx_t for sampled objects. Convert PROF_ALLOC_PREP() to prof_alloc_prep(), since precise backtrace depth within jemalloc functions is no longer an issue (pprof prunes irrelevant frames). Implement mallctl's: - prof.reset implements full sample data reset, and optional change of sample interval. - prof.lg_sample reads the current sample interval (opt.lg_prof_sample was the permanent source of truth prior to prof.reset). - thread.prof.name provides naming capability for threads within heap profile dumps. - thread.prof.active makes it possible to activate/deactivate heap profiling for individual threads. Modify the heap dump files to contain per thread heap profile data. This change is incompatible with the existing pprof, which will require enhancements to read and process the enriched data.
2014-08-19 07:22:13 +08:00
}
Do not (recursively) allocate within tsd_fetch(). Refactor tsd so that tsdn_fetch() does not trigger allocation, since allocation could cause infinite recursion. This resolves #458.
2016-10-21 14:59:12 +08:00
tdata_locks = (malloc_mutex_t *)base_alloc(tsd_tsdn(tsd),
Implement per arena base allocators. Add/rename related mallctls: - Add stats.arenas.<i>.base . - Rename stats.arenas.<i>.metadata to stats.arenas.<i>.internal . - Add stats.arenas.<i>.resident . Modify the arenas.extend mallctl to take an optional (extent_hooks_t *) argument so that it is possible for all base allocations to be serviced by the specified extent hooks. This resolves #463.
2016-12-23 06:39:10 +08:00
b0get(), PROF_NTDATA_LOCKS * sizeof(malloc_mutex_t),
CACHELINE);
Update brace style. Add braces around single-line blocks, and remove line breaks before function-opening braces. This resolves #537.
2017-01-16 08:56:30 +08:00
if (tdata_locks == NULL) {
Remove extraneous parens around return arguments. This resolves #540.
2017-01-20 10:15:45 +08:00
return true;
Update brace style. Add braces around single-line blocks, and remove line breaks before function-opening braces. This resolves #537.
2017-01-16 08:56:30 +08:00
}
Implement per thread heap profiling. Rename data structures (prof_thr_cnt_t-->prof_tctx_t, prof_ctx_t-->prof_gctx_t), and convert to storing a prof_tctx_t for sampled objects. Convert PROF_ALLOC_PREP() to prof_alloc_prep(), since precise backtrace depth within jemalloc functions is no longer an issue (pprof prunes irrelevant frames). Implement mallctl's: - prof.reset implements full sample data reset, and optional change of sample interval. - prof.lg_sample reads the current sample interval (opt.lg_prof_sample was the permanent source of truth prior to prof.reset). - thread.prof.name provides naming capability for threads within heap profile dumps. - thread.prof.active makes it possible to activate/deactivate heap profiling for individual threads. Modify the heap dump files to contain per thread heap profile data. This change is incompatible with the existing pprof, which will require enhancements to read and process the enriched data.
2014-08-19 07:22:13 +08:00
for (i = 0; i < PROF_NTDATA_LOCKS; i++) {
Add witness, a simple online locking validator. This resolves #358.
2016-04-14 14:36:15 +08:00
if (malloc_mutex_init(&tdata_locks[i], "prof_tdata",
Allow mutexes to take a lock ordering enum at construction. This lets us specify whether and how mutexes of the same rank are allowed to be acquired. Currently, we only allow two polices (only a single mutex at a given rank at a time, and mutexes acquired in ascending order), but we can plausibly allow more (e.g. the "release uncontended mutexes before blocking").
2017-05-16 06:38:15 +08:00
WITNESS_RANK_PROF_TDATA,
malloc_mutex_rank_exclusive)) {
Remove extraneous parens around return arguments. This resolves #540.
2017-01-20 10:15:45 +08:00
return true;
Update brace style. Add braces around single-line blocks, and remove line breaks before function-opening braces. This resolves #537.
2017-01-16 08:56:30 +08:00
}
Remove ephemeral mutexes. Remove ephemeral mutexes from the prof machinery, and remove malloc_mutex_destroy(). This simplifies mutex management on systems that call malloc()/free() inside pthread_mutex_{create,destroy}(). Add atomic_*_u() for operation on unsigned values. Fix prof_printf() to call malloc_vsnprintf() rather than malloc_snprintf().
2012-03-24 09:05:51 +08:00
}
Various heap profiling improvements. Add the --disable-prof-libgcc configure option, and add backtracing based on libgcc, which is used by default. Fix a bug in hash(). Fix various configuration-dependent compilation errors.
2010-02-11 10:15:53 +08:00
#ifdef JEMALLOC_PROF_LIBGCC
Guard libgcc unwind init with opt_prof. Only triggers libgcc unwind init when prof is enabled. This helps workaround some bootstrapping issues.
2019-02-20 07:58:13 +08:00
/*
* Cause the backtracing machinery to allocate its internal
* state before enabling profiling.
*/
_Unwind_Backtrace(prof_unwind_init_callback, NULL);
Various heap profiling improvements. Add the --disable-prof-libgcc configure option, and add backtracing based on libgcc, which is used by default. Fix a bug in hash(). Fix various configuration-dependent compilation errors.
2010-02-11 10:15:53 +08:00
#endif
Guard libgcc unwind init with opt_prof. Only triggers libgcc unwind init when prof is enabled. This helps workaround some bootstrapping issues.
2019-02-20 07:58:13 +08:00
}
Implement allocation profiling and leack checking. Add the --enable-prof and --enable-prof-libunwind configure options. Add the B/b, F/f, I/i, L/l, and U/u JEMALLOC_OPTIONS. Interval-based profile dump triggering is not yet implemented. Add supporting generic code: * Add memory barriers. * Add prn (LCG PRNG). * Add hash (Murmur hash function). * Add ckh (cuckoo hash tables).
2010-02-11 02:37:56 +08:00
prof_booted = true;
Remove extraneous parens around return arguments. This resolves #540.
2017-01-20 10:15:45 +08:00
return false;
Implement allocation profiling and leack checking. Add the --enable-prof and --enable-prof-libunwind configure options. Add the B/b, F/f, I/i, L/l, and U/u JEMALLOC_OPTIONS. Interval-based profile dump triggering is not yet implemented. Add supporting generic code: * Add memory barriers. * Add prn (LCG PRNG). * Add hash (Murmur hash function). * Add ckh (cuckoo hash tables).
2010-02-11 02:37:56 +08:00
}
Fix fork(2)-related deadlocks. Add a library constructor for jemalloc that initializes the allocator. This fixes a race that could occur if threads were created by the main thread prior to any memory allocation, followed by fork(2), and then memory allocation in the child process. Fix the prefork/postfork functions to acquire/release the ctl, prof, and rtree mutexes. This fixes various fork() child process deadlocks, but one possible deadlock remains (intentionally) unaddressed: prof backtracing can acquire runtime library mutexes, so deadlock is still possible if heap profiling is enabled during fork(). This deadlock is known to be a real issue in at least the case of libgcc-based backtracing. Reported by tfengjun.
2012-10-10 05:46:22 +08:00
void
Update brace style. Add braces around single-line blocks, and remove line breaks before function-opening braces. This resolves #537.
2017-01-16 08:56:30 +08:00
prof_prefork0(tsdn_t *tsdn) {
Conditionalize prof fork handling on config_prof. This allows the compiler to completely remove dead code.
2017-01-30 09:35:57 +08:00
if (config_prof && opt_prof) {
Fix fork(2)-related deadlocks. Add a library constructor for jemalloc that initializes the allocator. This fixes a race that could occur if threads were created by the main thread prior to any memory allocation, followed by fork(2), and then memory allocation in the child process. Fix the prefork/postfork functions to acquire/release the ctl, prof, and rtree mutexes. This fixes various fork() child process deadlocks, but one possible deadlock remains (intentionally) unaddressed: prof backtracing can acquire runtime library mutexes, so deadlock is still possible if heap profiling is enabled during fork(). This deadlock is known to be a real issue in at least the case of libgcc-based backtracing. Reported by tfengjun.
2012-10-10 05:46:22 +08:00
unsigned i;
Resolve bootstrapping issues when embedded in FreeBSD libc. b2c0d6322d2307458ae2b28545f8a5c9903d7ef5 (Add witness, a simple online locking validator.) caused a broad propagation of tsd throughout the internal API, but tsd_fetch() was designed to fail prior to tsd bootstrapping. Fix this by splitting tsd_t into non-nullable tsd_t and nullable tsdn_t, and modifying all internal APIs that do not critically rely on tsd to take nullable pointers. Furthermore, add the tsd_booted_get() function so that tsdn_fetch() can probe whether tsd bootstrapping is complete and return NULL if not. All dangerous conversions of nullable pointers are tsdn_tsd() calls that assert-fail on invalid conversion.
2016-05-11 13:21:10 +08:00
malloc_mutex_prefork(tsdn, &prof_dump_mtx);
malloc_mutex_prefork(tsdn, &bt2gctx_mtx);
malloc_mutex_prefork(tsdn, &tdatas_mtx);
Update brace style. Add braces around single-line blocks, and remove line breaks before function-opening braces. This resolves #537.
2017-01-16 08:56:30 +08:00
for (i = 0; i < PROF_NTDATA_LOCKS; i++) {
Resolve bootstrapping issues when embedded in FreeBSD libc. b2c0d6322d2307458ae2b28545f8a5c9903d7ef5 (Add witness, a simple online locking validator.) caused a broad propagation of tsd throughout the internal API, but tsd_fetch() was designed to fail prior to tsd bootstrapping. Fix this by splitting tsd_t into non-nullable tsd_t and nullable tsdn_t, and modifying all internal APIs that do not critically rely on tsd to take nullable pointers. Furthermore, add the tsd_booted_get() function so that tsdn_fetch() can probe whether tsd bootstrapping is complete and return NULL if not. All dangerous conversions of nullable pointers are tsdn_tsd() calls that assert-fail on invalid conversion.
2016-05-11 13:21:10 +08:00
malloc_mutex_prefork(tsdn, &tdata_locks[i]);
Update brace style. Add braces around single-line blocks, and remove line breaks before function-opening braces. This resolves #537.
2017-01-16 08:56:30 +08:00
}
for (i = 0; i < PROF_NCTX_LOCKS; i++) {
Resolve bootstrapping issues when embedded in FreeBSD libc. b2c0d6322d2307458ae2b28545f8a5c9903d7ef5 (Add witness, a simple online locking validator.) caused a broad propagation of tsd throughout the internal API, but tsd_fetch() was designed to fail prior to tsd bootstrapping. Fix this by splitting tsd_t into non-nullable tsd_t and nullable tsdn_t, and modifying all internal APIs that do not critically rely on tsd to take nullable pointers. Furthermore, add the tsd_booted_get() function so that tsdn_fetch() can probe whether tsd bootstrapping is complete and return NULL if not. All dangerous conversions of nullable pointers are tsdn_tsd() calls that assert-fail on invalid conversion.
2016-05-11 13:21:10 +08:00
malloc_mutex_prefork(tsdn, &gctx_locks[i]);
Update brace style. Add braces around single-line blocks, and remove line breaks before function-opening braces. This resolves #537.
2017-01-16 08:56:30 +08:00
}
Fix fork()-related lock rank ordering reversals.
2016-04-26 14:14:40 +08:00
}
}
void
Update brace style. Add braces around single-line blocks, and remove line breaks before function-opening braces. This resolves #537.
2017-01-16 08:56:30 +08:00
prof_prefork1(tsdn_t *tsdn) {
Conditionalize prof fork handling on config_prof. This allows the compiler to completely remove dead code.
2017-01-30 09:35:57 +08:00
if (config_prof && opt_prof) {
Resolve bootstrapping issues when embedded in FreeBSD libc. b2c0d6322d2307458ae2b28545f8a5c9903d7ef5 (Add witness, a simple online locking validator.) caused a broad propagation of tsd throughout the internal API, but tsd_fetch() was designed to fail prior to tsd bootstrapping. Fix this by splitting tsd_t into non-nullable tsd_t and nullable tsdn_t, and modifying all internal APIs that do not critically rely on tsd to take nullable pointers. Furthermore, add the tsd_booted_get() function so that tsdn_fetch() can probe whether tsd bootstrapping is complete and return NULL if not. All dangerous conversions of nullable pointers are tsdn_tsd() calls that assert-fail on invalid conversion.
2016-05-11 13:21:10 +08:00
malloc_mutex_prefork(tsdn, &prof_active_mtx);
Rename "prof_dump_seq_mtx" to "prof_dump_filename_mtx".
2019-09-10 11:04:18 +08:00
malloc_mutex_prefork(tsdn, &prof_dump_filename_mtx);
Resolve bootstrapping issues when embedded in FreeBSD libc. b2c0d6322d2307458ae2b28545f8a5c9903d7ef5 (Add witness, a simple online locking validator.) caused a broad propagation of tsd throughout the internal API, but tsd_fetch() was designed to fail prior to tsd bootstrapping. Fix this by splitting tsd_t into non-nullable tsd_t and nullable tsdn_t, and modifying all internal APIs that do not critically rely on tsd to take nullable pointers. Furthermore, add the tsd_booted_get() function so that tsdn_fetch() can probe whether tsd bootstrapping is complete and return NULL if not. All dangerous conversions of nullable pointers are tsdn_tsd() calls that assert-fail on invalid conversion.
2016-05-11 13:21:10 +08:00
malloc_mutex_prefork(tsdn, &prof_gdump_mtx);
malloc_mutex_prefork(tsdn, &next_thr_uid_mtx);
malloc_mutex_prefork(tsdn, &prof_thread_active_init_mtx);
Fix fork(2)-related deadlocks. Add a library constructor for jemalloc that initializes the allocator. This fixes a race that could occur if threads were created by the main thread prior to any memory allocation, followed by fork(2), and then memory allocation in the child process. Fix the prefork/postfork functions to acquire/release the ctl, prof, and rtree mutexes. This fixes various fork() child process deadlocks, but one possible deadlock remains (intentionally) unaddressed: prof backtracing can acquire runtime library mutexes, so deadlock is still possible if heap profiling is enabled during fork(). This deadlock is known to be a real issue in at least the case of libgcc-based backtracing. Reported by tfengjun.
2012-10-10 05:46:22 +08:00
}
}
void
Update brace style. Add braces around single-line blocks, and remove line breaks before function-opening braces. This resolves #537.
2017-01-16 08:56:30 +08:00
prof_postfork_parent(tsdn_t *tsdn) {
Conditionalize prof fork handling on config_prof. This allows the compiler to completely remove dead code.
2017-01-30 09:35:57 +08:00
if (config_prof && opt_prof) {
Fix fork(2)-related deadlocks. Add a library constructor for jemalloc that initializes the allocator. This fixes a race that could occur if threads were created by the main thread prior to any memory allocation, followed by fork(2), and then memory allocation in the child process. Fix the prefork/postfork functions to acquire/release the ctl, prof, and rtree mutexes. This fixes various fork() child process deadlocks, but one possible deadlock remains (intentionally) unaddressed: prof backtracing can acquire runtime library mutexes, so deadlock is still possible if heap profiling is enabled during fork(). This deadlock is known to be a real issue in at least the case of libgcc-based backtracing. Reported by tfengjun.
2012-10-10 05:46:22 +08:00
unsigned i;
Resolve bootstrapping issues when embedded in FreeBSD libc. b2c0d6322d2307458ae2b28545f8a5c9903d7ef5 (Add witness, a simple online locking validator.) caused a broad propagation of tsd throughout the internal API, but tsd_fetch() was designed to fail prior to tsd bootstrapping. Fix this by splitting tsd_t into non-nullable tsd_t and nullable tsdn_t, and modifying all internal APIs that do not critically rely on tsd to take nullable pointers. Furthermore, add the tsd_booted_get() function so that tsdn_fetch() can probe whether tsd bootstrapping is complete and return NULL if not. All dangerous conversions of nullable pointers are tsdn_tsd() calls that assert-fail on invalid conversion.
2016-05-11 13:21:10 +08:00
malloc_mutex_postfork_parent(tsdn,
&prof_thread_active_init_mtx);
malloc_mutex_postfork_parent(tsdn, &next_thr_uid_mtx);
malloc_mutex_postfork_parent(tsdn, &prof_gdump_mtx);
Rename "prof_dump_seq_mtx" to "prof_dump_filename_mtx".
2019-09-10 11:04:18 +08:00
malloc_mutex_postfork_parent(tsdn, &prof_dump_filename_mtx);
Resolve bootstrapping issues when embedded in FreeBSD libc. b2c0d6322d2307458ae2b28545f8a5c9903d7ef5 (Add witness, a simple online locking validator.) caused a broad propagation of tsd throughout the internal API, but tsd_fetch() was designed to fail prior to tsd bootstrapping. Fix this by splitting tsd_t into non-nullable tsd_t and nullable tsdn_t, and modifying all internal APIs that do not critically rely on tsd to take nullable pointers. Furthermore, add the tsd_booted_get() function so that tsdn_fetch() can probe whether tsd bootstrapping is complete and return NULL if not. All dangerous conversions of nullable pointers are tsdn_tsd() calls that assert-fail on invalid conversion.
2016-05-11 13:21:10 +08:00
malloc_mutex_postfork_parent(tsdn, &prof_active_mtx);
Update brace style. Add braces around single-line blocks, and remove line breaks before function-opening braces. This resolves #537.
2017-01-16 08:56:30 +08:00
for (i = 0; i < PROF_NCTX_LOCKS; i++) {
Resolve bootstrapping issues when embedded in FreeBSD libc. b2c0d6322d2307458ae2b28545f8a5c9903d7ef5 (Add witness, a simple online locking validator.) caused a broad propagation of tsd throughout the internal API, but tsd_fetch() was designed to fail prior to tsd bootstrapping. Fix this by splitting tsd_t into non-nullable tsd_t and nullable tsdn_t, and modifying all internal APIs that do not critically rely on tsd to take nullable pointers. Furthermore, add the tsd_booted_get() function so that tsdn_fetch() can probe whether tsd bootstrapping is complete and return NULL if not. All dangerous conversions of nullable pointers are tsdn_tsd() calls that assert-fail on invalid conversion.
2016-05-11 13:21:10 +08:00
malloc_mutex_postfork_parent(tsdn, &gctx_locks[i]);
Update brace style. Add braces around single-line blocks, and remove line breaks before function-opening braces. This resolves #537.
2017-01-16 08:56:30 +08:00
}
for (i = 0; i < PROF_NTDATA_LOCKS; i++) {
Resolve bootstrapping issues when embedded in FreeBSD libc. b2c0d6322d2307458ae2b28545f8a5c9903d7ef5 (Add witness, a simple online locking validator.) caused a broad propagation of tsd throughout the internal API, but tsd_fetch() was designed to fail prior to tsd bootstrapping. Fix this by splitting tsd_t into non-nullable tsd_t and nullable tsdn_t, and modifying all internal APIs that do not critically rely on tsd to take nullable pointers. Furthermore, add the tsd_booted_get() function so that tsdn_fetch() can probe whether tsd bootstrapping is complete and return NULL if not. All dangerous conversions of nullable pointers are tsdn_tsd() calls that assert-fail on invalid conversion.
2016-05-11 13:21:10 +08:00
malloc_mutex_postfork_parent(tsdn, &tdata_locks[i]);
Update brace style. Add braces around single-line blocks, and remove line breaks before function-opening braces. This resolves #537.
2017-01-16 08:56:30 +08:00
}
Resolve bootstrapping issues when embedded in FreeBSD libc. b2c0d6322d2307458ae2b28545f8a5c9903d7ef5 (Add witness, a simple online locking validator.) caused a broad propagation of tsd throughout the internal API, but tsd_fetch() was designed to fail prior to tsd bootstrapping. Fix this by splitting tsd_t into non-nullable tsd_t and nullable tsdn_t, and modifying all internal APIs that do not critically rely on tsd to take nullable pointers. Furthermore, add the tsd_booted_get() function so that tsdn_fetch() can probe whether tsd bootstrapping is complete and return NULL if not. All dangerous conversions of nullable pointers are tsdn_tsd() calls that assert-fail on invalid conversion.
2016-05-11 13:21:10 +08:00
malloc_mutex_postfork_parent(tsdn, &tdatas_mtx);
malloc_mutex_postfork_parent(tsdn, &bt2gctx_mtx);
malloc_mutex_postfork_parent(tsdn, &prof_dump_mtx);
Fix fork(2)-related deadlocks. Add a library constructor for jemalloc that initializes the allocator. This fixes a race that could occur if threads were created by the main thread prior to any memory allocation, followed by fork(2), and then memory allocation in the child process. Fix the prefork/postfork functions to acquire/release the ctl, prof, and rtree mutexes. This fixes various fork() child process deadlocks, but one possible deadlock remains (intentionally) unaddressed: prof backtracing can acquire runtime library mutexes, so deadlock is still possible if heap profiling is enabled during fork(). This deadlock is known to be a real issue in at least the case of libgcc-based backtracing. Reported by tfengjun.
2012-10-10 05:46:22 +08:00
}
}
void
Update brace style. Add braces around single-line blocks, and remove line breaks before function-opening braces. This resolves #537.
2017-01-16 08:56:30 +08:00
prof_postfork_child(tsdn_t *tsdn) {
Conditionalize prof fork handling on config_prof. This allows the compiler to completely remove dead code.
2017-01-30 09:35:57 +08:00
if (config_prof && opt_prof) {
Fix fork(2)-related deadlocks. Add a library constructor for jemalloc that initializes the allocator. This fixes a race that could occur if threads were created by the main thread prior to any memory allocation, followed by fork(2), and then memory allocation in the child process. Fix the prefork/postfork functions to acquire/release the ctl, prof, and rtree mutexes. This fixes various fork() child process deadlocks, but one possible deadlock remains (intentionally) unaddressed: prof backtracing can acquire runtime library mutexes, so deadlock is still possible if heap profiling is enabled during fork(). This deadlock is known to be a real issue in at least the case of libgcc-based backtracing. Reported by tfengjun.
2012-10-10 05:46:22 +08:00
unsigned i;
Resolve bootstrapping issues when embedded in FreeBSD libc. b2c0d6322d2307458ae2b28545f8a5c9903d7ef5 (Add witness, a simple online locking validator.) caused a broad propagation of tsd throughout the internal API, but tsd_fetch() was designed to fail prior to tsd bootstrapping. Fix this by splitting tsd_t into non-nullable tsd_t and nullable tsdn_t, and modifying all internal APIs that do not critically rely on tsd to take nullable pointers. Furthermore, add the tsd_booted_get() function so that tsdn_fetch() can probe whether tsd bootstrapping is complete and return NULL if not. All dangerous conversions of nullable pointers are tsdn_tsd() calls that assert-fail on invalid conversion.
2016-05-11 13:21:10 +08:00
malloc_mutex_postfork_child(tsdn, &prof_thread_active_init_mtx);
malloc_mutex_postfork_child(tsdn, &next_thr_uid_mtx);
malloc_mutex_postfork_child(tsdn, &prof_gdump_mtx);
Rename "prof_dump_seq_mtx" to "prof_dump_filename_mtx".
2019-09-10 11:04:18 +08:00
malloc_mutex_postfork_child(tsdn, &prof_dump_filename_mtx);
Resolve bootstrapping issues when embedded in FreeBSD libc. b2c0d6322d2307458ae2b28545f8a5c9903d7ef5 (Add witness, a simple online locking validator.) caused a broad propagation of tsd throughout the internal API, but tsd_fetch() was designed to fail prior to tsd bootstrapping. Fix this by splitting tsd_t into non-nullable tsd_t and nullable tsdn_t, and modifying all internal APIs that do not critically rely on tsd to take nullable pointers. Furthermore, add the tsd_booted_get() function so that tsdn_fetch() can probe whether tsd bootstrapping is complete and return NULL if not. All dangerous conversions of nullable pointers are tsdn_tsd() calls that assert-fail on invalid conversion.
2016-05-11 13:21:10 +08:00
malloc_mutex_postfork_child(tsdn, &prof_active_mtx);
Update brace style. Add braces around single-line blocks, and remove line breaks before function-opening braces. This resolves #537.
2017-01-16 08:56:30 +08:00
for (i = 0; i < PROF_NCTX_LOCKS; i++) {
Resolve bootstrapping issues when embedded in FreeBSD libc. b2c0d6322d2307458ae2b28545f8a5c9903d7ef5 (Add witness, a simple online locking validator.) caused a broad propagation of tsd throughout the internal API, but tsd_fetch() was designed to fail prior to tsd bootstrapping. Fix this by splitting tsd_t into non-nullable tsd_t and nullable tsdn_t, and modifying all internal APIs that do not critically rely on tsd to take nullable pointers. Furthermore, add the tsd_booted_get() function so that tsdn_fetch() can probe whether tsd bootstrapping is complete and return NULL if not. All dangerous conversions of nullable pointers are tsdn_tsd() calls that assert-fail on invalid conversion.
2016-05-11 13:21:10 +08:00
malloc_mutex_postfork_child(tsdn, &gctx_locks[i]);
Update brace style. Add braces around single-line blocks, and remove line breaks before function-opening braces. This resolves #537.
2017-01-16 08:56:30 +08:00
}
for (i = 0; i < PROF_NTDATA_LOCKS; i++) {
Resolve bootstrapping issues when embedded in FreeBSD libc. b2c0d6322d2307458ae2b28545f8a5c9903d7ef5 (Add witness, a simple online locking validator.) caused a broad propagation of tsd throughout the internal API, but tsd_fetch() was designed to fail prior to tsd bootstrapping. Fix this by splitting tsd_t into non-nullable tsd_t and nullable tsdn_t, and modifying all internal APIs that do not critically rely on tsd to take nullable pointers. Furthermore, add the tsd_booted_get() function so that tsdn_fetch() can probe whether tsd bootstrapping is complete and return NULL if not. All dangerous conversions of nullable pointers are tsdn_tsd() calls that assert-fail on invalid conversion.
2016-05-11 13:21:10 +08:00
malloc_mutex_postfork_child(tsdn, &tdata_locks[i]);
Update brace style. Add braces around single-line blocks, and remove line breaks before function-opening braces. This resolves #537.
2017-01-16 08:56:30 +08:00
}
Resolve bootstrapping issues when embedded in FreeBSD libc. b2c0d6322d2307458ae2b28545f8a5c9903d7ef5 (Add witness, a simple online locking validator.) caused a broad propagation of tsd throughout the internal API, but tsd_fetch() was designed to fail prior to tsd bootstrapping. Fix this by splitting tsd_t into non-nullable tsd_t and nullable tsdn_t, and modifying all internal APIs that do not critically rely on tsd to take nullable pointers. Furthermore, add the tsd_booted_get() function so that tsdn_fetch() can probe whether tsd bootstrapping is complete and return NULL if not. All dangerous conversions of nullable pointers are tsdn_tsd() calls that assert-fail on invalid conversion.
2016-05-11 13:21:10 +08:00
malloc_mutex_postfork_child(tsdn, &tdatas_mtx);
malloc_mutex_postfork_child(tsdn, &bt2gctx_mtx);
malloc_mutex_postfork_child(tsdn, &prof_dump_mtx);
Fix fork(2)-related deadlocks. Add a library constructor for jemalloc that initializes the allocator. This fixes a race that could occur if threads were created by the main thread prior to any memory allocation, followed by fork(2), and then memory allocation in the child process. Fix the prefork/postfork functions to acquire/release the ctl, prof, and rtree mutexes. This fixes various fork() child process deadlocks, but one possible deadlock remains (intentionally) unaddressed: prof backtracing can acquire runtime library mutexes, so deadlock is still possible if heap profiling is enabled during fork(). This deadlock is known to be a real issue in at least the case of libgcc-based backtracing. Reported by tfengjun.
2012-10-10 05:46:22 +08:00
}
}
Implement allocation profiling and leack checking. Add the --enable-prof and --enable-prof-libunwind configure options. Add the B/b, F/f, I/i, L/l, and U/u JEMALLOC_OPTIONS. Interval-based profile dump triggering is not yet implemented. Add supporting generic code: * Add memory barriers. * Add prn (LCG PRNG). * Add hash (Murmur hash function). * Add ckh (cuckoo hash tables).
2010-02-11 02:37:56 +08:00
/******************************************************************************/
Reference in New Issue Copy Permalink