RepoMirrors/haproxy
1
0
Fork 0
mirror of http://git.haproxy.org/git/haproxy.git/ synced 2025-01-19 04:00:46 +00:00
Code Issues Packages Projects Releases Wiki Activity
f7b7c3e2f2
haproxy/include/common/hathreads.h

873 lines
26 KiB
C
Raw Normal View History

MEDIUM: threads: Add hathreads header file This file contains all functions and macros used to deal with concurrency in HAProxy. It contains all high-level function to do atomic operation (HA_ATOMIC_*). Note, for now, we rely on "__atomic" GCC builtins to do atomic operation. So HAProxy can be compiled with the thread support iff these builtins are available. It also contains wrappers around plocks to use spin or read/write locks. These wrappers are used to abstract the internal representation of the locking system and to add information to help debugging, when compiled with suitable options. To add extra info on locks, you need to add DEBUG=-DDEBUG_THREAD or DEBUG=-DDEBUG_FULL compilation option. In addition to timing info on locks, we keep info on where a lock was acquired the last time (function name, file and line). There are also the thread id and a flag to know if it is still locked or not. This will be useful to debug deadlocks.
2017-10-12 14:09:09 +00:00
/*
* include/common/hathreads.h
* definitions, macros and inline functions about threads.
*
* Copyright (C) 2017 Christopher Fauet - cfaulet@haproxy.com
*
* This library is free software; you can redistribute it and/or
* modify it under the terms of the GNU Lesser General Public
* License as published by the Free Software Foundation, version 2.1
* exclusively.
*
* This library is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
* Lesser General Public License for more details.
*
* You should have received a copy of the GNU Lesser General Public
* License along with this library; if not, write to the Free Software
* Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
*/
#ifndef _COMMON_HATHREADS_H
#define _COMMON_HATHREADS_H
#include <common/config.h>
#define MAX_THREADS_MASK ((unsigned long)-1)
extern THREAD_LOCAL unsigned int tid; /* The thread id */
BUG/MINOR: threads: tid_bit must be a unsigned long This is specific to threads, no backport is needed.
2017-11-14 09:16:04 +00:00
extern THREAD_LOCAL unsigned long tid_bit; /* The bit corresponding to the thread id */
MEDIUM: threads: Add hathreads header file This file contains all functions and macros used to deal with concurrency in HAProxy. It contains all high-level function to do atomic operation (HA_ATOMIC_*). Note, for now, we rely on "__atomic" GCC builtins to do atomic operation. So HAProxy can be compiled with the thread support iff these builtins are available. It also contains wrappers around plocks to use spin or read/write locks. These wrappers are used to abstract the internal representation of the locking system and to add information to help debugging, when compiled with suitable options. To add extra info on locks, you need to add DEBUG=-DDEBUG_THREAD or DEBUG=-DDEBUG_FULL compilation option. In addition to timing info on locks, we keep info on where a lock was acquired the last time (function name, file and line). There are also the thread id and a flag to know if it is still locked or not. This will be useful to debug deadlocks.
2017-10-12 14:09:09 +00:00
#ifndef USE_THREAD
MINOR: threads: add a MAX_THREADS define instead of LONGBITS This one allows not to inflate some structures when threads are disabled. Now struct global is 1.4 kB instead of 33 kB. Should be backported to 1.8 for ease of backporting of upcoming patches.
2018-01-20 17:19:22 +00:00
#define MAX_THREADS 1
MINOR: threads: Use __decl_hathreads to declare locks This macro should be used to declare variables or struct members depending on the USE_THREAD compile option. It avoids the encapsulation of such declarations between #ifdef/#endif. It is used to declare all lock variables.
2017-11-13 09:34:01 +00:00
#define __decl_hathreads(decl)
MEDIUM: threads: Add hathreads header file This file contains all functions and macros used to deal with concurrency in HAProxy. It contains all high-level function to do atomic operation (HA_ATOMIC_*). Note, for now, we rely on "__atomic" GCC builtins to do atomic operation. So HAProxy can be compiled with the thread support iff these builtins are available. It also contains wrappers around plocks to use spin or read/write locks. These wrappers are used to abstract the internal representation of the locking system and to add information to help debugging, when compiled with suitable options. To add extra info on locks, you need to add DEBUG=-DDEBUG_THREAD or DEBUG=-DDEBUG_FULL compilation option. In addition to timing info on locks, we keep info on where a lock was acquired the last time (function name, file and line). There are also the thread id and a flag to know if it is still locked or not. This will be useful to debug deadlocks.
2017-10-12 14:09:09 +00:00
#define HA_ATOMIC_CAS(val, old, new) ({((*val) == (*old)) ? (*(val) = (new) , 1) : (*(old) = *(val), 0);})
#define HA_ATOMIC_ADD(val, i) ({*(val) += (i);})
#define HA_ATOMIC_SUB(val, i) ({*(val) -= (i);})
#define HA_ATOMIC_AND(val, flags) ({*(val) &= (flags);})
#define HA_ATOMIC_OR(val, flags) ({*(val) |= (flags);})
#define HA_ATOMIC_XCHG(val, new) \
({ \
typeof(*(val)) __old = *(val); \
*(val) = new; \
__old; \
})
MINOR: threads: add test and set/reset operations This just adds a set of naive bts/btr operations based on OR/AND. Later it could rely on pl_bts/btr to use arch-specific versions if needed.
2018-01-25 16:43:58 +00:00
#define HA_ATOMIC_BTS(val, bit) \
({ \
typeof((val)) __p = (val); \
typeof(*__p) __b = (1UL << (bit)); \
typeof(*__p) __t = *__p & __b; \
if (!__t) \
*__p |= __b; \
__t; \
})
#define HA_ATOMIC_BTR(val, bit) \
({ \
typeof((val)) __p = (val); \
typeof(*__p) __b = (1UL << (bit)); \
typeof(*__p) __t = *__p & __b; \
if (__t) \
*__p &= ~__b; \
__t; \
})
MEDIUM: threads: Add hathreads header file This file contains all functions and macros used to deal with concurrency in HAProxy. It contains all high-level function to do atomic operation (HA_ATOMIC_*). Note, for now, we rely on "__atomic" GCC builtins to do atomic operation. So HAProxy can be compiled with the thread support iff these builtins are available. It also contains wrappers around plocks to use spin or read/write locks. These wrappers are used to abstract the internal representation of the locking system and to add information to help debugging, when compiled with suitable options. To add extra info on locks, you need to add DEBUG=-DDEBUG_THREAD or DEBUG=-DDEBUG_FULL compilation option. In addition to timing info on locks, we keep info on where a lock was acquired the last time (function name, file and line). There are also the thread id and a flag to know if it is still locked or not. This will be useful to debug deadlocks.
2017-10-12 14:09:09 +00:00
#define HA_ATOMIC_STORE(val, new) ({*(val) = new;})
#define HA_ATOMIC_UPDATE_MAX(val, new) \
({ \
typeof(*(val)) __new = (new); \
\
if (*(val) < __new) \
*(val) = __new; \
*(val); \
})
#define HA_ATOMIC_UPDATE_MIN(val, new) \
({ \
typeof(*(val)) __new = (new); \
\
if (*(val) > __new) \
*(val) = __new; \
*(val); \
})
MINOR: threads: add a portable barrier for threads and non-threads HA_BARRIER() is just a simple memory barrier to prevent the compiler from reordering our code.
2017-10-31 17:00:20 +00:00
#define HA_BARRIER() do { } while (0)
MEDIUM: threads: Adds a set of functions to handle sync-point A sync-point is a protected area where you have the warranty that no concurrency access is possible. It is implementated as a thread barrier to enter in the sync-point and another one to exit from it. Inside the sync-point, all threads that must do some syncrhonous processing will be called one after the other while all other threads will wait. All threads will then exit from the sync-point at the same time. A sync-point will be evaluated only when necessary because it is a costly operation. To limit the waiting time of each threads, we must have a mechanism to wakeup all threads. This is done with a pipe shared by all threads. By writting in this pipe, we will interrupt all threads blocked on a poller. The pipe is then flushed before exiting from the sync-point.
2017-10-19 09:59:15 +00:00
#define THREAD_SYNC_INIT(m) do { /* do nothing */ } while(0)
#define THREAD_SYNC_ENABLE() do { /* do nothing */ } while(0)
#define THREAD_WANT_SYNC() do { /* do nothing */ } while(0)
#define THREAD_ENTER_SYNC() do { /* do nothing */ } while(0)
#define THREAD_EXIT_SYNC() do { /* do nothing */ } while(0)
#define THREAD_NO_SYNC() ({ 0; })
#define THREAD_NEED_SYNC() ({ 1; })
BUILD: threads: Rename SPIN/RWLOCK macros using HA_ prefix This remove any name conflicts, especially on Solaris.
2017-11-07 09:42:54 +00:00
#define HA_SPIN_INIT(l) do { /* do nothing */ } while(0)
#define HA_SPIN_DESTROY(l) do { /* do nothing */ } while(0)
#define HA_SPIN_LOCK(lbl, l) do { /* do nothing */ } while(0)
#define HA_SPIN_TRYLOCK(lbl, l) ({ 0; })
#define HA_SPIN_UNLOCK(lbl, l) do { /* do nothing */ } while(0)
#define HA_RWLOCK_INIT(l) do { /* do nothing */ } while(0)
#define HA_RWLOCK_DESTROY(l) do { /* do nothing */ } while(0)
#define HA_RWLOCK_WRLOCK(lbl, l) do { /* do nothing */ } while(0)
#define HA_RWLOCK_TRYWRLOCK(lbl, l) ({ 0; })
#define HA_RWLOCK_WRUNLOCK(lbl, l) do { /* do nothing */ } while(0)
#define HA_RWLOCK_RDLOCK(lbl, l) do { /* do nothing */ } while(0)
#define HA_RWLOCK_TRYRDLOCK(lbl, l) ({ 0; })
#define HA_RWLOCK_RDUNLOCK(lbl, l) do { /* do nothing */ } while(0)
MEDIUM: threads: Add hathreads header file This file contains all functions and macros used to deal with concurrency in HAProxy. It contains all high-level function to do atomic operation (HA_ATOMIC_*). Note, for now, we rely on "__atomic" GCC builtins to do atomic operation. So HAProxy can be compiled with the thread support iff these builtins are available. It also contains wrappers around plocks to use spin or read/write locks. These wrappers are used to abstract the internal representation of the locking system and to add information to help debugging, when compiled with suitable options. To add extra info on locks, you need to add DEBUG=-DDEBUG_THREAD or DEBUG=-DDEBUG_FULL compilation option. In addition to timing info on locks, we keep info on where a lock was acquired the last time (function name, file and line). There are also the thread id and a flag to know if it is still locked or not. This will be useful to debug deadlocks.
2017-10-12 14:09:09 +00:00
MINOR: threads: Introduce double-width CAS on x86_64 and arm. Introduce double-width compare-and-swap on arches that support it, right now x86_64, arm, and aarch64. Also introduce functions to do memory barriers.
2017-12-21 16:13:05 +00:00
static inline void __ha_barrier_load(void)
{
}
static inline void __ha_barrier_store(void)
{
}
static inline void __ha_barrier_full(void)
{
}
MEDIUM: threads: Add hathreads header file This file contains all functions and macros used to deal with concurrency in HAProxy. It contains all high-level function to do atomic operation (HA_ATOMIC_*). Note, for now, we rely on "__atomic" GCC builtins to do atomic operation. So HAProxy can be compiled with the thread support iff these builtins are available. It also contains wrappers around plocks to use spin or read/write locks. These wrappers are used to abstract the internal representation of the locking system and to add information to help debugging, when compiled with suitable options. To add extra info on locks, you need to add DEBUG=-DDEBUG_THREAD or DEBUG=-DDEBUG_FULL compilation option. In addition to timing info on locks, we keep info on where a lock was acquired the last time (function name, file and line). There are also the thread id and a flag to know if it is still locked or not. This will be useful to debug deadlocks.
2017-10-12 14:09:09 +00:00
#else /* USE_THREAD */
#include <stdio.h>
#include <stdlib.h>
#include <string.h>
#include <pthread.h>
#include <import/plock.h>
MINOR: threads: add a MAX_THREADS define instead of LONGBITS This one allows not to inflate some structures when threads are disabled. Now struct global is 1.4 kB instead of 33 kB. Should be backported to 1.8 for ease of backporting of upcoming patches.
2018-01-20 17:19:22 +00:00
#define MAX_THREADS LONGBITS
MINOR: threads: Use __decl_hathreads to declare locks This macro should be used to declare variables or struct members depending on the USE_THREAD compile option. It avoids the encapsulation of such declarations between #ifdef/#endif. It is used to declare all lock variables.
2017-11-13 09:34:01 +00:00
#define __decl_hathreads(decl) decl
MEDIUM: threads: Add hathreads header file This file contains all functions and macros used to deal with concurrency in HAProxy. It contains all high-level function to do atomic operation (HA_ATOMIC_*). Note, for now, we rely on "__atomic" GCC builtins to do atomic operation. So HAProxy can be compiled with the thread support iff these builtins are available. It also contains wrappers around plocks to use spin or read/write locks. These wrappers are used to abstract the internal representation of the locking system and to add information to help debugging, when compiled with suitable options. To add extra info on locks, you need to add DEBUG=-DDEBUG_THREAD or DEBUG=-DDEBUG_FULL compilation option. In addition to timing info on locks, we keep info on where a lock was acquired the last time (function name, file and line). There are also the thread id and a flag to know if it is still locked or not. This will be useful to debug deadlocks.
2017-10-12 14:09:09 +00:00
/* TODO: thread: For now, we rely on GCC builtins but it could be a good idea to
* have a header file regrouping all functions dealing with threads. */
MINOR: hathreads: add support for gcc < 4.7 Till now the use of __atomic_* gcc builtins required gcc >= 4.7. Since some supported and quite common operating systems like CentOS 6 still come with older versions (4.4) and the mapping to the older builtins is reasonably simple, let's implement it. This code is only used for gcc < 4.7. It has been quickly tested on a machine using gcc 4.4.4 and provided expected results. This patch should be backported to 1.8.
2018-01-04 17:49:31 +00:00
BUILD/MINOR: ancient gcc versions atomic fix Commit 1a69af6d3892fe1946bb8babb3044d2d26afd46e introduced code for atomic prior to 4.7. Unfortunately clang uses as well those constants which is misleading.
2018-01-11 14:20:43 +00:00
#if defined(__GNUC__) && (__GNUC__ < 4 || __GNUC__ == 4 && __GNUC_MINOR__ < 7) && !defined(__clang__)
MINOR: hathreads: add support for gcc < 4.7 Till now the use of __atomic_* gcc builtins required gcc >= 4.7. Since some supported and quite common operating systems like CentOS 6 still come with older versions (4.4) and the mapping to the older builtins is reasonably simple, let's implement it. This code is only used for gcc < 4.7. It has been quickly tested on a machine using gcc 4.4.4 and provided expected results. This patch should be backported to 1.8.
2018-01-04 17:49:31 +00:00
/* gcc < 4.7 */
#define HA_ATOMIC_ADD(val, i) __sync_add_and_fetch(val, i)
#define HA_ATOMIC_SUB(val, i) __sync_sub_and_fetch(val, i)
#define HA_ATOMIC_AND(val, flags) __sync_and_and_fetch(val, flags)
#define HA_ATOMIC_OR(val, flags) __sync_or_and_fetch(val, flags)
/* the CAS is a bit complicated. The older API doesn't support returning the
* value and the swap's result at the same time. So here we take what looks
* like the safest route, consisting in using the boolean version guaranteeing
* that the operation was performed or not, and we snoop a previous value. If
* the compare succeeds, we return. If it fails, we return the previous value,
* but only if it differs from the expected one. If it's the same it's a race
* thus we try again to avoid confusing a possibly sensitive caller.
*/
#define HA_ATOMIC_CAS(val, old, new) \
({ \
typeof((val)) __val = (val); \
typeof((old)) __oldp = (old); \
typeof(*(old)) __oldv; \
typeof((new)) __new = (new); \
int __ret; \
do { \
__oldv = *__val; \
__ret = __sync_bool_compare_and_swap(__val, *__oldp, __new); \
} while (!__ret && *__oldp == __oldv); \
if (!__ret) \
*__oldp = __oldv; \
__ret; \
})
#define HA_ATOMIC_XCHG(val, new) \
({ \
typeof((val)) __val = (val); \
typeof(*(val)) __old; \
typeof((new)) __new = (new); \
do { __old = *__val; \
} while (!__sync_bool_compare_and_swap(__val, __old, __new)); \
__old; \
})
MINOR: threads: add test and set/reset operations This just adds a set of naive bts/btr operations based on OR/AND. Later it could rely on pl_bts/btr to use arch-specific versions if needed.
2018-01-25 16:43:58 +00:00
#define HA_ATOMIC_BTS(val, bit) \
({ \
typeof(*(val)) __b = (1UL << (bit)); \
__sync_fetch_and_or((val), __b) & __b; \
})
#define HA_ATOMIC_BTR(val, bit) \
({ \
typeof(*(val)) __b = (1UL << (bit)); \
__sync_fetch_and_and((val), ~__b) & __b; \
})
MINOR: hathreads: add support for gcc < 4.7 Till now the use of __atomic_* gcc builtins required gcc >= 4.7. Since some supported and quite common operating systems like CentOS 6 still come with older versions (4.4) and the mapping to the older builtins is reasonably simple, let's implement it. This code is only used for gcc < 4.7. It has been quickly tested on a machine using gcc 4.4.4 and provided expected results. This patch should be backported to 1.8.
2018-01-04 17:49:31 +00:00
#define HA_ATOMIC_STORE(val, new) \
({ \
typeof((val)) __val = (val); \
typeof(*(val)) __old; \
typeof((new)) __new = (new); \
do { __old = *__val; \
} while (!__sync_bool_compare_and_swap(__val, __old, __new)); \
})
#else
/* gcc >= 4.7 */
MEDIUM: threads: Add hathreads header file This file contains all functions and macros used to deal with concurrency in HAProxy. It contains all high-level function to do atomic operation (HA_ATOMIC_*). Note, for now, we rely on "__atomic" GCC builtins to do atomic operation. So HAProxy can be compiled with the thread support iff these builtins are available. It also contains wrappers around plocks to use spin or read/write locks. These wrappers are used to abstract the internal representation of the locking system and to add information to help debugging, when compiled with suitable options. To add extra info on locks, you need to add DEBUG=-DDEBUG_THREAD or DEBUG=-DDEBUG_FULL compilation option. In addition to timing info on locks, we keep info on where a lock was acquired the last time (function name, file and line). There are also the thread id and a flag to know if it is still locked or not. This will be useful to debug deadlocks.
2017-10-12 14:09:09 +00:00
#define HA_ATOMIC_CAS(val, old, new) __atomic_compare_exchange_n(val, old, new, 0, 0, 0)
#define HA_ATOMIC_ADD(val, i) __atomic_add_fetch(val, i, 0)
#define HA_ATOMIC_SUB(val, i) __atomic_sub_fetch(val, i, 0)
#define HA_ATOMIC_AND(val, flags) __atomic_and_fetch(val, flags, 0)
#define HA_ATOMIC_OR(val, flags) __atomic_or_fetch(val, flags, 0)
MINOR: threads: add test and set/reset operations This just adds a set of naive bts/btr operations based on OR/AND. Later it could rely on pl_bts/btr to use arch-specific versions if needed.
2018-01-25 16:43:58 +00:00
#define HA_ATOMIC_BTS(val, bit) \
({ \
typeof(*(val)) __b = (1UL << (bit)); \
__sync_fetch_and_or((val), __b) & __b; \
})
#define HA_ATOMIC_BTR(val, bit) \
({ \
typeof(*(val)) __b = (1UL << (bit)); \
__sync_fetch_and_and((val), ~__b) & __b; \
})
MEDIUM: threads: Add hathreads header file This file contains all functions and macros used to deal with concurrency in HAProxy. It contains all high-level function to do atomic operation (HA_ATOMIC_*). Note, for now, we rely on "__atomic" GCC builtins to do atomic operation. So HAProxy can be compiled with the thread support iff these builtins are available. It also contains wrappers around plocks to use spin or read/write locks. These wrappers are used to abstract the internal representation of the locking system and to add information to help debugging, when compiled with suitable options. To add extra info on locks, you need to add DEBUG=-DDEBUG_THREAD or DEBUG=-DDEBUG_FULL compilation option. In addition to timing info on locks, we keep info on where a lock was acquired the last time (function name, file and line). There are also the thread id and a flag to know if it is still locked or not. This will be useful to debug deadlocks.
2017-10-12 14:09:09 +00:00
#define HA_ATOMIC_XCHG(val, new) __atomic_exchange_n(val, new, 0)
#define HA_ATOMIC_STORE(val, new) __atomic_store_n(val, new, 0)
MINOR: hathreads: add support for gcc < 4.7 Till now the use of __atomic_* gcc builtins required gcc >= 4.7. Since some supported and quite common operating systems like CentOS 6 still come with older versions (4.4) and the mapping to the older builtins is reasonably simple, let's implement it. This code is only used for gcc < 4.7. It has been quickly tested on a machine using gcc 4.4.4 and provided expected results. This patch should be backported to 1.8.
2018-01-04 17:49:31 +00:00
#endif
MEDIUM: threads: Add hathreads header file This file contains all functions and macros used to deal with concurrency in HAProxy. It contains all high-level function to do atomic operation (HA_ATOMIC_*). Note, for now, we rely on "__atomic" GCC builtins to do atomic operation. So HAProxy can be compiled with the thread support iff these builtins are available. It also contains wrappers around plocks to use spin or read/write locks. These wrappers are used to abstract the internal representation of the locking system and to add information to help debugging, when compiled with suitable options. To add extra info on locks, you need to add DEBUG=-DDEBUG_THREAD or DEBUG=-DDEBUG_FULL compilation option. In addition to timing info on locks, we keep info on where a lock was acquired the last time (function name, file and line). There are also the thread id and a flag to know if it is still locked or not. This will be useful to debug deadlocks.
2017-10-12 14:09:09 +00:00
#define HA_ATOMIC_UPDATE_MAX(val, new) \
({ \
typeof(*(val)) __old = *(val); \
typeof(*(val)) __new = (new); \
\
while (__old < __new && !HA_ATOMIC_CAS(val, &__old, __new)); \
(*val); \
})
#define HA_ATOMIC_UPDATE_MIN(val, new) \
({ \
typeof((*val)) __old = *(val); \
typeof((*val)) __new = (new); \
\
while (__old > __new && !HA_ATOMIC_CAS(val, &__old, __new)); \
(*val); \
})
MINOR: threads: add a portable barrier for threads and non-threads HA_BARRIER() is just a simple memory barrier to prevent the compiler from reordering our code.
2017-10-31 17:00:20 +00:00
#define HA_BARRIER() pl_barrier()
MEDIUM: threads: Adds a set of functions to handle sync-point A sync-point is a protected area where you have the warranty that no concurrency access is possible. It is implementated as a thread barrier to enter in the sync-point and another one to exit from it. Inside the sync-point, all threads that must do some syncrhonous processing will be called one after the other while all other threads will wait. All threads will then exit from the sync-point at the same time. A sync-point will be evaluated only when necessary because it is a costly operation. To limit the waiting time of each threads, we must have a mechanism to wakeup all threads. This is done with a pipe shared by all threads. By writting in this pipe, we will interrupt all threads blocked on a poller. The pipe is then flushed before exiting from the sync-point.
2017-10-19 09:59:15 +00:00
#define THREAD_SYNC_INIT(m) thread_sync_init(m)
#define THREAD_SYNC_ENABLE() thread_sync_enable()
#define THREAD_WANT_SYNC() thread_want_sync()
#define THREAD_ENTER_SYNC() thread_enter_sync()
#define THREAD_EXIT_SYNC() thread_exit_sync()
#define THREAD_NO_SYNC() thread_no_sync()
#define THREAD_NEED_SYNC() thread_need_sync()
int thread_sync_init(unsigned long mask);
void thread_sync_enable(void);
void thread_want_sync(void);
void thread_enter_sync(void);
void thread_exit_sync(void);
int thread_no_sync(void);
int thread_need_sync(void);
MEDIUM: threads: Add hathreads header file This file contains all functions and macros used to deal with concurrency in HAProxy. It contains all high-level function to do atomic operation (HA_ATOMIC_*). Note, for now, we rely on "__atomic" GCC builtins to do atomic operation. So HAProxy can be compiled with the thread support iff these builtins are available. It also contains wrappers around plocks to use spin or read/write locks. These wrappers are used to abstract the internal representation of the locking system and to add information to help debugging, when compiled with suitable options. To add extra info on locks, you need to add DEBUG=-DDEBUG_THREAD or DEBUG=-DDEBUG_FULL compilation option. In addition to timing info on locks, we keep info on where a lock was acquired the last time (function name, file and line). There are also the thread id and a flag to know if it is still locked or not. This will be useful to debug deadlocks.
2017-10-12 14:09:09 +00:00
#if defined(DEBUG_THREAD) || defined(DEBUG_FULL)
BUG/MINOR: threads: Update labels array because of changes in lock_label enum Recent changes to the enum were not synchronized with the lock debugging code. Now we use a switch/case instead of an array so that the compiler throws a warning if there is any inconsistency. To be backported to 1.8 (at least to add the START entry).
2018-01-30 10:04:29 +00:00
/* WARNING!!! if you update this enum, please also keep lock_label() up to date below */
MEDIUM: threads: Add hathreads header file This file contains all functions and macros used to deal with concurrency in HAProxy. It contains all high-level function to do atomic operation (HA_ATOMIC_*). Note, for now, we rely on "__atomic" GCC builtins to do atomic operation. So HAProxy can be compiled with the thread support iff these builtins are available. It also contains wrappers around plocks to use spin or read/write locks. These wrappers are used to abstract the internal representation of the locking system and to add information to help debugging, when compiled with suitable options. To add extra info on locks, you need to add DEBUG=-DDEBUG_THREAD or DEBUG=-DDEBUG_FULL compilation option. In addition to timing info on locks, we keep info on where a lock was acquired the last time (function name, file and line). There are also the thread id and a flag to know if it is still locked or not. This will be useful to debug deadlocks.
2017-10-12 14:09:09 +00:00
enum lock_label {
MEDIUM: threads: Adds a set of functions to handle sync-point A sync-point is a protected area where you have the warranty that no concurrency access is possible. It is implementated as a thread barrier to enter in the sync-point and another one to exit from it. Inside the sync-point, all threads that must do some syncrhonous processing will be called one after the other while all other threads will wait. All threads will then exit from the sync-point at the same time. A sync-point will be evaluated only when necessary because it is a costly operation. To limit the waiting time of each threads, we must have a mechanism to wakeup all threads. This is done with a pipe shared by all threads. By writting in this pipe, we will interrupt all threads blocked on a poller. The pipe is then flushed before exiting from the sync-point.
2017-10-19 09:59:15 +00:00
THREAD_SYNC_LOCK = 0,
MAJOR: threads/fd: Make fd stuffs thread-safe Many changes have been made to do so. First, the fd_updt array, where all pending FDs for polling are stored, is now a thread-local array. Then 3 locks have been added to protect, respectively, the fdtab array, the fd_cache array and poll information. In addition, a lock for each entry in the fdtab array has been added to protect all accesses to a specific FD or its information. For pollers, according to the poller, the way to manage the concurrency is different. There is a poller loop on each thread. So the set of monitored FDs may need to be protected. epoll and kqueue are thread-safe per-se, so there few things to do to protect these pollers. This is not possible with select and poll, so there is no sharing between the threads. The poller on each thread is independant from others. Finally, per-thread init/deinit functions are used for each pollers and for FD part for manage thread-local ressources. Now, you must be carefull when a FD is created during the HAProxy startup. All update on the FD state must be made in the threads context and never before their creation. This is mandatory because fd_updt array is thread-local and initialized only for threads. Because there is no pollers for the main one, this array remains uninitialized in this context. For this reason, listeners are now enabled in run_thread_poll_loop function, just like the worker pipe.
2017-05-29 08:40:41 +00:00
FD_LOCK,
MAJOR: threads/task: handle multithread on task scheduler 2 global locks have been added to protect, respectively, the run queue and the wait queue. And a process mask has been added on each task. Like for FDs, this mask is used to know which threads are allowed to process a task. For many tasks, all threads are granted. And this must be your first intension when you create a new task, else you have a good reason to make a task sticky on some threads. This is then the responsibility to the process callback to lock what have to be locked in the task context. Nevertheless, all tasks linked to a session must be sticky on the thread creating the session. It is important that I/O handlers processing session FDs and these tasks run on the same thread to avoid conflicts.
2017-09-27 12:59:38 +00:00
TASK_RQ_LOCK,
TASK_WQ_LOCK,
MEDIUM: threads/pool: Make pool thread-safe by locking all access to a pool A lock has been added for each memory pool. It is used to protect the pool during allocations and releases. It is also used when pool info are dumped.
2017-08-29 07:52:38 +00:00
POOL_LOCK,
MEDIUM: threads/listeners: Make listeners thread-safe First, we use atomic operations to update jobs/totalconn/actconn variables, listener's nbconn variable and listener's counters. Then we add a lock on listeners to protect access to their information. And finally, listener queues (global and per proxy) are also protected by a lock. Here, because access to these queues are unusal, we use the same lock for all queues instead of a global one for the global queue and a lock per proxy for others.
2017-05-30 13:36:50 +00:00
LISTENER_LOCK,
LISTENER_QUEUE_LOCK,
MEDIUM: threads/proxy: Add a lock per proxy and atomically update proxy vars Now, each proxy contains a lock that must be used when necessary to protect it. Moreover, all proxy's counters are now updated using atomic operations.
2017-06-02 13:33:24 +00:00
PROXY_LOCK,
MEDIUM: threads/server: Add a lock per server and atomically update server vars The server's lock is use, among other things, to lock acces to the active connection list of a server.
2017-06-08 12:04:45 +00:00
SERVER_LOCK,
MINOR: threads/server: Add a lock to deal with insert in updates_servers list This list is used to save changes on the servers state. So when serveral threads are used, it must be locked. The changes are then applied in the sync-point. To do so, servers_update_status has be moved in the sync-point. So this is useless to lock it at this step because the sync-point is a protected area by iteself.
2017-10-16 10:00:40 +00:00
UPDATED_SERVERS_LOCK,
MEDIUM: threads/lb: Make LB algorithms (lb_*.c) thread-safe A lock for LB parameters has been added inside the proxy structure and atomic operations have been used to update server variables releated to lb. The only significant change is about lb_map. Because the servers status are updated in the sync-point, we can call recalc_server_map function synchronously in map_set_server_status_up/down function.
2017-06-09 12:17:53 +00:00
LBPRM_LOCK,
MEDIUM: threads/signal: Add a lock to make signals thread-safe A global lock has been added to protect the signal processing. So when a signal it triggered, only one thread will catch it.
2017-05-30 13:34:30 +00:00
SIGNALS_LOCK,
MEDIUM: threads/stick-tables: handle multithreads on stick tables The stick table API was slightly reworked: A global spin lock on stick table was added to perform lookup and insert in a thread safe way. The handling of refcount on entries is now handled directly by stick tables functions under protection of this lock and was removed from the code of callers. The "stktable_store" function is no more externalized and users should now use "stktable_set_entry" in any case of insertion. This last one performs a lookup followed by a store if not found. So the code using "stktable_store" was re-worked. Lookup, and set_entry functions automatically increase the refcount of the returned/stored entry. The function "sticktable_touch" was renamed "sticktable_touch_local" and is now able to decrease the refcount if last arg is set to true. It is allowing to release the entry without taking the lock twice. A new function "sticktable_touch_remote" is now used to insert entries coming from remote peers at the right place in the update tree. The code of peer update was re-worked to use this new function. This function is also able to decrease the refcount if wanted. The function "stksess_kill" also handle a parameter to decrease the refcount on the entry. A read/write lock is added on each entry to protect the data content updates of the entry.
2017-06-13 17:37:32 +00:00
STK_TABLE_LOCK,
STK_SESS_LOCK,
MAJOR: threads/applet: Handle multithreading for applets A global lock has been added to protect accesses to the list of active applets. A process mask has also been added on each applet. Like for FDs and tasks, it is used to know which threads are allowed to process an applet. Because applets are, most of time, linked to a session, it should be sticky on the same thread. But in all cases, it is the responsibility of the applet handler to lock what have to be protected in the applet context.
2017-06-19 10:38:55 +00:00
APPLETS_LOCK,
MAJOR: threads/peers: Make peers thread safe A lock is used to protect accesses to a peer structure. A the lock is taken in the applet handler when the peer is identified and released living the applet handler. In the scheduling task for peers section, the lock is taken for every listed peer and released at the end of the process task function. The peer 'force shutdown' function was also re-worked.
2017-06-19 15:46:37 +00:00
PEER_LOCK,
MAJOR: threads/buffer: Make buffer wait queue thread safe Adds a global lock to protect the buffer wait queue.
2017-06-21 13:42:52 +00:00
BUF_WQ_LOCK,
MEDIUM: threads/stream: Make streams list thread safe Adds a global lock to protect the full streams list used to dump sessions on stats socket.
2017-06-30 14:23:45 +00:00
STRMS_LOCK,
MAJOR: threads/ssl: Make SSL part thread-safe First, OpenSSL is now initialized to be thread-safe. This is done by setting 2 callbacks. The first one is ssl_locking_function. It handles the locks and unlocks. The second one is ssl_id_function. It returns the current thread id. During the init step, we create as much as R/W locks as needed, ie the number returned by CRYPTO_num_locks function. Next, The reusable SSL session in the server context is now thread-local. Shctx is now also initialized if HAProxy is started with several threads. And finally, a global lock has been added to protect the LRU cache used to store generated certificates. The function ssl_sock_get_generated_cert is now deprecated because the retrieved certificate can be removed by another threads in same time. Instead, a new function has been added, ssl_sock_assign_generated_cert. It must be used to search a certificate in the cache and set it immediatly if found.
2017-06-15 14:37:39 +00:00
SSL_LOCK,
SSL_GEN_CERTS_LOCK,
MAJOR: threads/map: Make acls/maps thread safe locks have been added in pat_ref and pattern_expr structures to protect all accesses to an instance of on of them. Moreover, a global lock has been added to protect the LRU cache used for pattern matching. Patterns are now duplicated after a successfull matching, to avoid modification by other threads when the result is used. Finally, the function reloading a pattern list has been modified to be thread-safe.
2017-07-03 09:34:05 +00:00
PATREF_LOCK,
PATEXP_LOCK,
PATLRU_LOCK,
MEDIUM: thread/vars: Make vars thread-safe A RW lock has been added to the vars structure to protect each list of variables. And a global RW lock is used to protect registered names. When a varibable is fetched, we duplicate sample data because the variable could be modified by another thread.
2017-07-24 14:30:34 +00:00
VARS_LOCK,
MEDIUM: threads/compression: Make HTTP compression thread-safe
2017-07-25 09:07:15 +00:00
COMP_POOL_LOCK,
MEDIUM: threads/lua: Add locks around the Lua execution parts. Note that the Lua processing is not really thread safe. It provides heavy system which consists to add our own lock function in the Lua code and recompile the library. This system will probably not accepted by maintainers of various distribs. Our main excution point of the Lua is the function lua_resume(). A quick looking on the Lua sources displays a lua_lock() a the start of function and a lua_unlock() at the end of the function. So I conclude that the Lua thread safe mode just perform a mutex around all execution. So I prefer to do this in the HAProxy code, it will be easier for distro maintainers. Note that the HAProxy lua functions rounded by the macro SET_SAFE_LJMP and RESET_SAFE_LJMP manipulates the Lua stack, so it will be careful to set mutex around these functions.
2017-07-12 09:41:21 +00:00
LUA_LOCK,
MEDIUM: threads/tasks: Add lock around notifications This patch add lock around some notification calls
2017-07-16 22:14:07 +00:00
NOTIF_LOCK,
MEDIUM: thread/spoe: Make the SPOE thread-safe Because there is not migration mechanism yet, all runtime information about an SPOE agent are thread-local and async exchanges with agents are disabled when we have serveral threads. Howerver, pipelining is still available. So for now, the thread part of the SPOE is pretty simple.
2017-09-25 12:48:02 +00:00
SPOE_APPLET_LOCK,
MEDIUM: thread/dns: Make DNS thread-safe
2017-10-04 14:17:58 +00:00
DNS_LOCK,
MINOR: threads/checks: Add a lock to protect the pid list used by external checks
2017-10-20 13:40:23 +00:00
PID_LIST_LOCK,
MINOR: threads/mailers: Add a lock to protect queues of email alerts
2017-10-23 13:54:24 +00:00
EMAIL_ALERTS_LOCK,
BUG/MEDIUM: splice/threads: pipe reuse list was not protected. The list is now protected using a global spinlock.
2017-11-07 10:19:48 +00:00
PIPES_LOCK,
BUG/MEDIUM: threads/mworker: fix a race on startup Marc Fournier reported an interesting case when using threads with the master-worker mode : sometimes, a listener would have its FD closed during startup. Sometimes it could even be health checks seeing this. What happens is that after the threads are created, and the pollers enabled on each threads, the master-worker pipe is registered, and at the same time a close() is performed on the write side of this pipe since the children must not use it. But since this is replicated in every thread, what happens is that the first thread closes the pipe, thus releases the FD, and the next thread starting a listener in parallel gets this FD reassigned. Then another thread closes the FD again, which this time corresponds to the listener. It can also happen with the health check sockets if they're started early enough. This patch splits the mworker_pipe_register() function in two, so that the close() of the write side of the FD is performed very early after the fork() and long before threads are created (we don't need to delay it anyway). Only the pipe registration is done in the threaded code since it is important that the pollers are properly allocated for this. The mworker_pipe_register() function now takes care of registering the pipe only once, and this is guaranteed by a new surrounding lock. The call to protocol_enable_all() looks fragile in theory since it scans the list of proxies and their listeners, though in practice all threads scan the same list and take the same locks for each listener so it's not possible that any of them escapes the process and finishes before all listeners are started. And the operation is idempotent. This fix must be backported to 1.8. Thanks to Marc for providing very detailed traces clearly showing the problem.
2018-01-23 18:01:49 +00:00
START_LOCK,
BUG/MINOR: ssl/threads: Make management of the TLS ticket keys files thread-safe A TLS ticket keys file can be updated on the CLI and used in same time. So we need to protect it to be sure all accesses are thread-safe. Because updates are infrequent, a R/W lock has been used. This patch must be backported in 1.8
2018-02-16 10:23:49 +00:00
TLSKEYS_REF_LOCK,
BUG/MAJOR: threads/queue: Fix thread-safety issues on the queues management The management of the servers and the proxies queues was not thread-safe at all. First, the accesses to <strm>->pend_pos were not protected. So it was possible to release it on a thread (for instance because the stream is released) and to use it in same time on another one (because we redispatch pending connections for a server). Then, the accesses to stream's information (flags and target) from anywhere is forbidden. To be safe, The stream's state must always be updated in the context of process_stream. So to fix these issues, the queue module has been refactored. A lock has been added in the pendconn structure. And now, when we try to dequeue a pending connection, we start by unlinking it from the server/proxy queue and we wake up the stream. Then, it is the stream reponsibility to really dequeue it (or release it). This way, we are sure that only the stream can create and release its <pend_pos> field. However, be careful. This new implementation should be thread-safe (hopefully...). But it is not optimal and in some situations, it could be really slower in multi-threaded mode than in single-threaded one. The problem is that, when we try to dequeue pending connections, we process it from the older one to the newer one independently to the thread's affinity. So we need to wait the other threads' wakeup to really process them. If threads are blocked in the poller, this will add a significant latency. This problem happens when maxconn values are very low. This patch must be backported in 1.8.
2018-03-14 15:18:06 +00:00
PENDCONN_LOCK,
MEDIUM: threads: Adds a set of functions to handle sync-point A sync-point is a protected area where you have the warranty that no concurrency access is possible. It is implementated as a thread barrier to enter in the sync-point and another one to exit from it. Inside the sync-point, all threads that must do some syncrhonous processing will be called one after the other while all other threads will wait. All threads will then exit from the sync-point at the same time. A sync-point will be evaluated only when necessary because it is a costly operation. To limit the waiting time of each threads, we must have a mechanism to wakeup all threads. This is done with a pipe shared by all threads. By writting in this pipe, we will interrupt all threads blocked on a poller. The pipe is then flushed before exiting from the sync-point.
2017-10-19 09:59:15 +00:00
LOCK_LABELS
MEDIUM: threads: Add hathreads header file This file contains all functions and macros used to deal with concurrency in HAProxy. It contains all high-level function to do atomic operation (HA_ATOMIC_*). Note, for now, we rely on "__atomic" GCC builtins to do atomic operation. So HAProxy can be compiled with the thread support iff these builtins are available. It also contains wrappers around plocks to use spin or read/write locks. These wrappers are used to abstract the internal representation of the locking system and to add information to help debugging, when compiled with suitable options. To add extra info on locks, you need to add DEBUG=-DDEBUG_THREAD or DEBUG=-DDEBUG_FULL compilation option. In addition to timing info on locks, we keep info on where a lock was acquired the last time (function name, file and line). There are also the thread id and a flag to know if it is still locked or not. This will be useful to debug deadlocks.
2017-10-12 14:09:09 +00:00
};
struct lock_stat {
uint64_t nsec_wait_for_write;
uint64_t nsec_wait_for_read;
uint64_t num_write_locked;
uint64_t num_write_unlocked;
uint64_t num_read_locked;
uint64_t num_read_unlocked;
};
extern struct lock_stat lock_stats[LOCK_LABELS];
#define __HA_SPINLOCK_T unsigned long
#define __SPIN_INIT(l) ({ (*l) = 0; })
#define __SPIN_DESTROY(l) ({ (*l) = 0; })
MINOR: threads: use faster locks for the spin locks The spin locks used to rely on W locks, which involve a loop waiting for readers to leave, and this doesn't happen here. It's more efficient to use S locks instead, which are also mutually exclusive and do not have this loop. This saves one test per spinlock and a few tens of bytes allowing certain functions to be inlined.
2017-11-06 00:03:26 +00:00
#define __SPIN_LOCK(l) pl_take_s(l)
#define __SPIN_TRYLOCK(l) !pl_try_s(l)
#define __SPIN_UNLOCK(l) pl_drop_s(l)
MEDIUM: threads: Add hathreads header file This file contains all functions and macros used to deal with concurrency in HAProxy. It contains all high-level function to do atomic operation (HA_ATOMIC_*). Note, for now, we rely on "__atomic" GCC builtins to do atomic operation. So HAProxy can be compiled with the thread support iff these builtins are available. It also contains wrappers around plocks to use spin or read/write locks. These wrappers are used to abstract the internal representation of the locking system and to add information to help debugging, when compiled with suitable options. To add extra info on locks, you need to add DEBUG=-DDEBUG_THREAD or DEBUG=-DDEBUG_FULL compilation option. In addition to timing info on locks, we keep info on where a lock was acquired the last time (function name, file and line). There are also the thread id and a flag to know if it is still locked or not. This will be useful to debug deadlocks.
2017-10-12 14:09:09 +00:00
#define __HA_RWLOCK_T unsigned long
#define __RWLOCK_INIT(l) ({ (*l) = 0; })
#define __RWLOCK_DESTROY(l) ({ (*l) = 0; })
#define __RWLOCK_WRLOCK(l) pl_take_w(l)
#define __RWLOCK_TRYWRLOCK(l) !pl_try_w(l)
#define __RWLOCK_WRUNLOCK(l) pl_drop_w(l)
#define __RWLOCK_RDLOCK(l) pl_take_r(l)
#define __RWLOCK_TRYRDLOCK(l) !pl_try_r(l)
#define __RWLOCK_RDUNLOCK(l) pl_drop_r(l)
#define HA_SPINLOCK_T struct ha_spinlock
BUILD: threads: Rename SPIN/RWLOCK macros using HA_ prefix This remove any name conflicts, especially on Solaris.
2017-11-07 09:42:54 +00:00
#define HA_SPIN_INIT(l) __spin_init(l)
#define HA_SPIN_DESTROY(l) __spin_destroy(l)
MEDIUM: threads: Add hathreads header file This file contains all functions and macros used to deal with concurrency in HAProxy. It contains all high-level function to do atomic operation (HA_ATOMIC_*). Note, for now, we rely on "__atomic" GCC builtins to do atomic operation. So HAProxy can be compiled with the thread support iff these builtins are available. It also contains wrappers around plocks to use spin or read/write locks. These wrappers are used to abstract the internal representation of the locking system and to add information to help debugging, when compiled with suitable options. To add extra info on locks, you need to add DEBUG=-DDEBUG_THREAD or DEBUG=-DDEBUG_FULL compilation option. In addition to timing info on locks, we keep info on where a lock was acquired the last time (function name, file and line). There are also the thread id and a flag to know if it is still locked or not. This will be useful to debug deadlocks.
2017-10-12 14:09:09 +00:00
BUILD: threads: Rename SPIN/RWLOCK macros using HA_ prefix This remove any name conflicts, especially on Solaris.
2017-11-07 09:42:54 +00:00
#define HA_SPIN_LOCK(lbl, l) __spin_lock(lbl, l, __func__, __FILE__, __LINE__)
#define HA_SPIN_TRYLOCK(lbl, l) __spin_trylock(lbl, l, __func__, __FILE__, __LINE__)
#define HA_SPIN_UNLOCK(lbl, l) __spin_unlock(lbl, l, __func__, __FILE__, __LINE__)
MEDIUM: threads: Add hathreads header file This file contains all functions and macros used to deal with concurrency in HAProxy. It contains all high-level function to do atomic operation (HA_ATOMIC_*). Note, for now, we rely on "__atomic" GCC builtins to do atomic operation. So HAProxy can be compiled with the thread support iff these builtins are available. It also contains wrappers around plocks to use spin or read/write locks. These wrappers are used to abstract the internal representation of the locking system and to add information to help debugging, when compiled with suitable options. To add extra info on locks, you need to add DEBUG=-DDEBUG_THREAD or DEBUG=-DDEBUG_FULL compilation option. In addition to timing info on locks, we keep info on where a lock was acquired the last time (function name, file and line). There are also the thread id and a flag to know if it is still locked or not. This will be useful to debug deadlocks.
2017-10-12 14:09:09 +00:00
#define HA_RWLOCK_T struct ha_rwlock
BUILD: threads: Rename SPIN/RWLOCK macros using HA_ prefix This remove any name conflicts, especially on Solaris.
2017-11-07 09:42:54 +00:00
#define HA_RWLOCK_INIT(l) __ha_rwlock_init((l))
#define HA_RWLOCK_DESTROY(l) __ha_rwlock_destroy((l))
#define HA_RWLOCK_WRLOCK(lbl,l) __ha_rwlock_wrlock(lbl, l, __func__, __FILE__, __LINE__)
#define HA_RWLOCK_TRYWRLOCK(lbl,l) __ha_rwlock_trywrlock(lbl, l, __func__, __FILE__, __LINE__)
#define HA_RWLOCK_WRUNLOCK(lbl,l) __ha_rwlock_wrunlock(lbl, l, __func__, __FILE__, __LINE__)
#define HA_RWLOCK_RDLOCK(lbl,l) __ha_rwlock_rdlock(lbl, l)
#define HA_RWLOCK_TRYRDLOCK(lbl,l) __ha_rwlock_tryrdlock(lbl, l)
#define HA_RWLOCK_RDUNLOCK(lbl,l) __ha_rwlock_rdunlock(lbl, l)
MEDIUM: threads: Add hathreads header file This file contains all functions and macros used to deal with concurrency in HAProxy. It contains all high-level function to do atomic operation (HA_ATOMIC_*). Note, for now, we rely on "__atomic" GCC builtins to do atomic operation. So HAProxy can be compiled with the thread support iff these builtins are available. It also contains wrappers around plocks to use spin or read/write locks. These wrappers are used to abstract the internal representation of the locking system and to add information to help debugging, when compiled with suitable options. To add extra info on locks, you need to add DEBUG=-DDEBUG_THREAD or DEBUG=-DDEBUG_FULL compilation option. In addition to timing info on locks, we keep info on where a lock was acquired the last time (function name, file and line). There are also the thread id and a flag to know if it is still locked or not. This will be useful to debug deadlocks.
2017-10-12 14:09:09 +00:00
struct ha_spinlock {
__HA_SPINLOCK_T lock;
struct {
unsigned long owner; /* a bit is set to 1 << tid for the lock owner */
unsigned long waiters; /* a bit is set to 1 << tid for waiting threads */
struct {
const char *function;
const char *file;
int line;
} last_location; /* location of the last owner */
} info;
};
struct ha_rwlock {
__HA_RWLOCK_T lock;
struct {
unsigned long cur_writer; /* a bit is set to 1 << tid for the lock owner */
unsigned long wait_writers; /* a bit is set to 1 << tid for waiting writers */
unsigned long cur_readers; /* a bit is set to 1 << tid for current readers */
unsigned long wait_readers; /* a bit is set to 1 << tid for waiting waiters */
struct {
const char *function;
const char *file;
int line;
} last_location; /* location of the last write owner */
} info;
};
BUG/MINOR: threads: Update labels array because of changes in lock_label enum Recent changes to the enum were not synchronized with the lock debugging code. Now we use a switch/case instead of an array so that the compiler throws a warning if there is any inconsistency. To be backported to 1.8 (at least to add the START entry).
2018-01-30 10:04:29 +00:00
static inline const char *lock_label(enum lock_label label)
{
switch (label) {
case THREAD_SYNC_LOCK: return "THREAD_SYNC";
case FD_LOCK: return "FD";
case TASK_RQ_LOCK: return "TASK_RQ";
case TASK_WQ_LOCK: return "TASK_WQ";
case POOL_LOCK: return "POOL";
case LISTENER_LOCK: return "LISTENER";
case LISTENER_QUEUE_LOCK: return "LISTENER_QUEUE";
case PROXY_LOCK: return "PROXY";
case SERVER_LOCK: return "SERVER";
case UPDATED_SERVERS_LOCK: return "UPDATED_SERVERS";
case LBPRM_LOCK: return "LBPRM";
case SIGNALS_LOCK: return "SIGNALS";
case STK_TABLE_LOCK: return "STK_TABLE";
case STK_SESS_LOCK: return "STK_SESS";
case APPLETS_LOCK: return "APPLETS";
case PEER_LOCK: return "PEER";
case BUF_WQ_LOCK: return "BUF_WQ";
case STRMS_LOCK: return "STRMS";
case SSL_LOCK: return "SSL";
case SSL_GEN_CERTS_LOCK: return "SSL_GEN_CERTS";
case PATREF_LOCK: return "PATREF";
case PATEXP_LOCK: return "PATEXP";
case PATLRU_LOCK: return "PATLRU";
case VARS_LOCK: return "VARS";
case COMP_POOL_LOCK: return "COMP_POOL";
case LUA_LOCK: return "LUA";
case NOTIF_LOCK: return "NOTIF";
case SPOE_APPLET_LOCK: return "SPOE_APPLET";
case DNS_LOCK: return "DNS";
case PID_LIST_LOCK: return "PID_LIST";
case EMAIL_ALERTS_LOCK: return "EMAIL_ALERTS";
case PIPES_LOCK: return "PIPES";
case START_LOCK: return "START";
BUG/MINOR: ssl/threads: Make management of the TLS ticket keys files thread-safe A TLS ticket keys file can be updated on the CLI and used in same time. So we need to protect it to be sure all accesses are thread-safe. Because updates are infrequent, a R/W lock has been used. This patch must be backported in 1.8
2018-02-16 10:23:49 +00:00
case TLSKEYS_REF_LOCK: return "TLSKEYS_REF";
BUG/MAJOR: threads/queue: Fix thread-safety issues on the queues management The management of the servers and the proxies queues was not thread-safe at all. First, the accesses to <strm>->pend_pos were not protected. So it was possible to release it on a thread (for instance because the stream is released) and to use it in same time on another one (because we redispatch pending connections for a server). Then, the accesses to stream's information (flags and target) from anywhere is forbidden. To be safe, The stream's state must always be updated in the context of process_stream. So to fix these issues, the queue module has been refactored. A lock has been added in the pendconn structure. And now, when we try to dequeue a pending connection, we start by unlinking it from the server/proxy queue and we wake up the stream. Then, it is the stream reponsibility to really dequeue it (or release it). This way, we are sure that only the stream can create and release its <pend_pos> field. However, be careful. This new implementation should be thread-safe (hopefully...). But it is not optimal and in some situations, it could be really slower in multi-threaded mode than in single-threaded one. The problem is that, when we try to dequeue pending connections, we process it from the older one to the newer one independently to the thread's affinity. So we need to wait the other threads' wakeup to really process them. If threads are blocked in the poller, this will add a significant latency. This problem happens when maxconn values are very low. This patch must be backported in 1.8.
2018-03-14 15:18:06 +00:00
case PENDCONN_LOCK: return "PENDCONN";
BUG/MINOR: threads: Update labels array because of changes in lock_label enum Recent changes to the enum were not synchronized with the lock debugging code. Now we use a switch/case instead of an array so that the compiler throws a warning if there is any inconsistency. To be backported to 1.8 (at least to add the START entry).
2018-01-30 10:04:29 +00:00
case LOCK_LABELS: break; /* keep compiler happy */
};
/* only way to come here is consecutive to an internal bug */
abort();
}
MEDIUM: threads: Add hathreads header file This file contains all functions and macros used to deal with concurrency in HAProxy. It contains all high-level function to do atomic operation (HA_ATOMIC_*). Note, for now, we rely on "__atomic" GCC builtins to do atomic operation. So HAProxy can be compiled with the thread support iff these builtins are available. It also contains wrappers around plocks to use spin or read/write locks. These wrappers are used to abstract the internal representation of the locking system and to add information to help debugging, when compiled with suitable options. To add extra info on locks, you need to add DEBUG=-DDEBUG_THREAD or DEBUG=-DDEBUG_FULL compilation option. In addition to timing info on locks, we keep info on where a lock was acquired the last time (function name, file and line). There are also the thread id and a flag to know if it is still locked or not. This will be useful to debug deadlocks.
2017-10-12 14:09:09 +00:00
static inline void show_lock_stats()
{
int lbl;
for (lbl = 0; lbl < LOCK_LABELS; lbl++) {
fprintf(stderr,
"Stats about Lock %s: \n"
"\t # write lock : %lu\n"
"\t # write unlock: %lu (%ld)\n"
"\t # wait time for write : %.3f msec\n"
"\t # wait time for write/lock: %.3f nsec\n"
"\t # read lock : %lu\n"
"\t # read unlock : %lu (%ld)\n"
"\t # wait time for read : %.3f msec\n"
"\t # wait time for read/lock : %.3f nsec\n",
BUG/MINOR: threads: Update labels array because of changes in lock_label enum Recent changes to the enum were not synchronized with the lock debugging code. Now we use a switch/case instead of an array so that the compiler throws a warning if there is any inconsistency. To be backported to 1.8 (at least to add the START entry).
2018-01-30 10:04:29 +00:00
lock_label(lbl),
MEDIUM: threads: Add hathreads header file This file contains all functions and macros used to deal with concurrency in HAProxy. It contains all high-level function to do atomic operation (HA_ATOMIC_*). Note, for now, we rely on "__atomic" GCC builtins to do atomic operation. So HAProxy can be compiled with the thread support iff these builtins are available. It also contains wrappers around plocks to use spin or read/write locks. These wrappers are used to abstract the internal representation of the locking system and to add information to help debugging, when compiled with suitable options. To add extra info on locks, you need to add DEBUG=-DDEBUG_THREAD or DEBUG=-DDEBUG_FULL compilation option. In addition to timing info on locks, we keep info on where a lock was acquired the last time (function name, file and line). There are also the thread id and a flag to know if it is still locked or not. This will be useful to debug deadlocks.
2017-10-12 14:09:09 +00:00
lock_stats[lbl].num_write_locked,
lock_stats[lbl].num_write_unlocked,
lock_stats[lbl].num_write_unlocked - lock_stats[lbl].num_write_locked,
(double)lock_stats[lbl].nsec_wait_for_write / 1000000.0,
lock_stats[lbl].num_write_locked ? ((double)lock_stats[lbl].nsec_wait_for_write / (double)lock_stats[lbl].num_write_locked) : 0,
lock_stats[lbl].num_read_locked,
lock_stats[lbl].num_read_unlocked,
lock_stats[lbl].num_read_unlocked - lock_stats[lbl].num_read_locked,
(double)lock_stats[lbl].nsec_wait_for_read / 1000000.0,
lock_stats[lbl].num_read_locked ? ((double)lock_stats[lbl].nsec_wait_for_read / (double)lock_stats[lbl].num_read_locked) : 0);
}
}
/* Following functions are used to collect some stats about locks. We wrap
* pthread functions to known how much time we wait in a lock. */
static uint64_t nsec_now(void) {
struct timespec ts;
clock_gettime(CLOCK_MONOTONIC, &ts);
return ((uint64_t) ts.tv_sec * 1000000000ULL +
(uint64_t) ts.tv_nsec);
}
static inline void __ha_rwlock_init(struct ha_rwlock *l)
{
memset(l, 0, sizeof(struct ha_rwlock));
__RWLOCK_INIT(&l->lock);
}
static inline void __ha_rwlock_destroy(struct ha_rwlock *l)
{
__RWLOCK_DESTROY(&l->lock);
memset(l, 0, sizeof(struct ha_rwlock));
}
static inline void __ha_rwlock_wrlock(enum lock_label lbl, struct ha_rwlock *l,
const char *func, const char *file, int line)
{
uint64_t start_time;
if (unlikely(l->info.cur_writer & tid_bit)) {
/* the thread is already owning the lock for write */
abort();
}
if (unlikely(l->info.cur_readers & tid_bit)) {
/* the thread is already owning the lock for read */
abort();
}
HA_ATOMIC_OR(&l->info.wait_writers, tid_bit);
start_time = nsec_now();
__RWLOCK_WRLOCK(&l->lock);
HA_ATOMIC_ADD(&lock_stats[lbl].nsec_wait_for_write, (nsec_now() - start_time));
HA_ATOMIC_ADD(&lock_stats[lbl].num_write_locked, 1);
l->info.cur_writer = tid_bit;
l->info.last_location.function = func;
l->info.last_location.file = file;
l->info.last_location.line = line;
HA_ATOMIC_AND(&l->info.wait_writers, ~tid_bit);
}
static inline int __ha_rwlock_trywrlock(enum lock_label lbl, struct ha_rwlock *l,
const char *func, const char *file, int line)
{
uint64_t start_time;
int r;
if (unlikely(l->info.cur_writer & tid_bit)) {
/* the thread is already owning the lock for write */
abort();
}
if (unlikely(l->info.cur_readers & tid_bit)) {
/* the thread is already owning the lock for read */
abort();
}
/* We set waiting writer because trywrlock could wait for readers to quit */
HA_ATOMIC_OR(&l->info.wait_writers, tid_bit);
start_time = nsec_now();
r = __RWLOCK_TRYWRLOCK(&l->lock);
HA_ATOMIC_ADD(&lock_stats[lbl].nsec_wait_for_write, (nsec_now() - start_time));
if (unlikely(r)) {
HA_ATOMIC_AND(&l->info.wait_writers, ~tid_bit);
return r;
}
HA_ATOMIC_ADD(&lock_stats[lbl].num_write_locked, 1);
l->info.cur_writer = tid_bit;
l->info.last_location.function = func;
l->info.last_location.file = file;
l->info.last_location.line = line;
HA_ATOMIC_AND(&l->info.wait_writers, ~tid_bit);
return 0;
}
static inline void __ha_rwlock_wrunlock(enum lock_label lbl,struct ha_rwlock *l,
const char *func, const char *file, int line)
{
if (unlikely(!(l->info.cur_writer & tid_bit))) {
/* the thread is not owning the lock for write */
abort();
}
l->info.cur_writer = 0;
l->info.last_location.function = func;
l->info.last_location.file = file;
l->info.last_location.line = line;
__RWLOCK_WRUNLOCK(&l->lock);
HA_ATOMIC_ADD(&lock_stats[lbl].num_write_unlocked, 1);
}
static inline void __ha_rwlock_rdlock(enum lock_label lbl,struct ha_rwlock *l)
{
uint64_t start_time;
if (unlikely(l->info.cur_writer & tid_bit)) {
/* the thread is already owning the lock for write */
abort();
}
if (unlikely(l->info.cur_readers & tid_bit)) {
/* the thread is already owning the lock for read */
abort();
}
HA_ATOMIC_OR(&l->info.wait_readers, tid_bit);
start_time = nsec_now();
__RWLOCK_RDLOCK(&l->lock);
HA_ATOMIC_ADD(&lock_stats[lbl].nsec_wait_for_read, (nsec_now() - start_time));
HA_ATOMIC_ADD(&lock_stats[lbl].num_read_locked, 1);
HA_ATOMIC_OR(&l->info.cur_readers, tid_bit);
HA_ATOMIC_AND(&l->info.wait_readers, ~tid_bit);
}
static inline int __ha_rwlock_tryrdlock(enum lock_label lbl,struct ha_rwlock *l)
{
int r;
if (unlikely(l->info.cur_writer & tid_bit)) {
/* the thread is already owning the lock for write */
abort();
}
if (unlikely(l->info.cur_readers & tid_bit)) {
/* the thread is already owning the lock for read */
abort();
}
/* try read should never wait */
r = __RWLOCK_TRYRDLOCK(&l->lock);
if (unlikely(r))
return r;
HA_ATOMIC_ADD(&lock_stats[lbl].num_read_locked, 1);
HA_ATOMIC_OR(&l->info.cur_readers, tid_bit);
return 0;
}
static inline void __ha_rwlock_rdunlock(enum lock_label lbl,struct ha_rwlock *l)
{
if (unlikely(!(l->info.cur_readers & tid_bit))) {
/* the thread is not owning the lock for read */
abort();
}
HA_ATOMIC_AND(&l->info.cur_readers, ~tid_bit);
__RWLOCK_RDUNLOCK(&l->lock);
HA_ATOMIC_ADD(&lock_stats[lbl].num_read_unlocked, 1);
}
static inline void __spin_init(struct ha_spinlock *l)
{
memset(l, 0, sizeof(struct ha_spinlock));
__SPIN_INIT(&l->lock);
}
static inline void __spin_destroy(struct ha_spinlock *l)
{
__SPIN_DESTROY(&l->lock);
memset(l, 0, sizeof(struct ha_spinlock));
}
static inline void __spin_lock(enum lock_label lbl, struct ha_spinlock *l,
const char *func, const char *file, int line)
{
uint64_t start_time;
if (unlikely(l->info.owner & tid_bit)) {
/* the thread is already owning the lock */
abort();
}
HA_ATOMIC_OR(&l->info.waiters, tid_bit);
start_time = nsec_now();
__SPIN_LOCK(&l->lock);
HA_ATOMIC_ADD(&lock_stats[lbl].nsec_wait_for_write, (nsec_now() - start_time));
HA_ATOMIC_ADD(&lock_stats[lbl].num_write_locked, 1);
l->info.owner = tid_bit;
l->info.last_location.function = func;
l->info.last_location.file = file;
l->info.last_location.line = line;
HA_ATOMIC_AND(&l->info.waiters, ~tid_bit);
}
static inline int __spin_trylock(enum lock_label lbl, struct ha_spinlock *l,
const char *func, const char *file, int line)
{
int r;
if (unlikely(l->info.owner & tid_bit)) {
/* the thread is already owning the lock */
abort();
}
/* try read should never wait */
r = __SPIN_TRYLOCK(&l->lock);
if (unlikely(r))
return r;
HA_ATOMIC_ADD(&lock_stats[lbl].num_write_locked, 1);
l->info.owner = tid_bit;
l->info.last_location.function = func;
l->info.last_location.file = file;
l->info.last_location.line = line;
return 0;
}
static inline void __spin_unlock(enum lock_label lbl, struct ha_spinlock *l,
const char *func, const char *file, int line)
{
if (unlikely(!(l->info.owner & tid_bit))) {
/* the thread is not owning the lock */
abort();
}
l->info.owner = 0;
l->info.last_location.function = func;
l->info.last_location.file = file;
l->info.last_location.line = line;
BUG/MINOR: thread: fix a typo in the debug code __spin_unlock() used to call RWLOCK_WRUNLOCK() to unlock in the debug code. It's harmless as they happen to be identical.
2017-11-02 15:26:02 +00:00
__SPIN_UNLOCK(&l->lock);
MEDIUM: threads: Add hathreads header file This file contains all functions and macros used to deal with concurrency in HAProxy. It contains all high-level function to do atomic operation (HA_ATOMIC_*). Note, for now, we rely on "__atomic" GCC builtins to do atomic operation. So HAProxy can be compiled with the thread support iff these builtins are available. It also contains wrappers around plocks to use spin or read/write locks. These wrappers are used to abstract the internal representation of the locking system and to add information to help debugging, when compiled with suitable options. To add extra info on locks, you need to add DEBUG=-DDEBUG_THREAD or DEBUG=-DDEBUG_FULL compilation option. In addition to timing info on locks, we keep info on where a lock was acquired the last time (function name, file and line). There are also the thread id and a flag to know if it is still locked or not. This will be useful to debug deadlocks.
2017-10-12 14:09:09 +00:00
HA_ATOMIC_ADD(&lock_stats[lbl].num_write_unlocked, 1);
}
#else /* DEBUG_THREAD */
#define HA_SPINLOCK_T unsigned long
BUILD: threads: Rename SPIN/RWLOCK macros using HA_ prefix This remove any name conflicts, especially on Solaris.
2017-11-07 09:42:54 +00:00
#define HA_SPIN_INIT(l) ({ (*l) = 0; })
#define HA_SPIN_DESTROY(l) ({ (*l) = 0; })
#define HA_SPIN_LOCK(lbl, l) pl_take_s(l)
#define HA_SPIN_TRYLOCK(lbl, l) !pl_try_s(l)
#define HA_SPIN_UNLOCK(lbl, l) pl_drop_s(l)
MEDIUM: threads: Add hathreads header file This file contains all functions and macros used to deal with concurrency in HAProxy. It contains all high-level function to do atomic operation (HA_ATOMIC_*). Note, for now, we rely on "__atomic" GCC builtins to do atomic operation. So HAProxy can be compiled with the thread support iff these builtins are available. It also contains wrappers around plocks to use spin or read/write locks. These wrappers are used to abstract the internal representation of the locking system and to add information to help debugging, when compiled with suitable options. To add extra info on locks, you need to add DEBUG=-DDEBUG_THREAD or DEBUG=-DDEBUG_FULL compilation option. In addition to timing info on locks, we keep info on where a lock was acquired the last time (function name, file and line). There are also the thread id and a flag to know if it is still locked or not. This will be useful to debug deadlocks.
2017-10-12 14:09:09 +00:00
#define HA_RWLOCK_T unsigned long
BUILD: threads: Rename SPIN/RWLOCK macros using HA_ prefix This remove any name conflicts, especially on Solaris.
2017-11-07 09:42:54 +00:00
#define HA_RWLOCK_INIT(l) ({ (*l) = 0; })
#define HA_RWLOCK_DESTROY(l) ({ (*l) = 0; })
#define HA_RWLOCK_WRLOCK(lbl,l) pl_take_w(l)
#define HA_RWLOCK_TRYWRLOCK(lbl,l) !pl_try_w(l)
#define HA_RWLOCK_WRUNLOCK(lbl,l) pl_drop_w(l)
#define HA_RWLOCK_RDLOCK(lbl,l) pl_take_r(l)
#define HA_RWLOCK_TRYRDLOCK(lbl,l) !pl_try_r(l)
#define HA_RWLOCK_RDUNLOCK(lbl,l) pl_drop_r(l)
MEDIUM: threads: Add hathreads header file This file contains all functions and macros used to deal with concurrency in HAProxy. It contains all high-level function to do atomic operation (HA_ATOMIC_*). Note, for now, we rely on "__atomic" GCC builtins to do atomic operation. So HAProxy can be compiled with the thread support iff these builtins are available. It also contains wrappers around plocks to use spin or read/write locks. These wrappers are used to abstract the internal representation of the locking system and to add information to help debugging, when compiled with suitable options. To add extra info on locks, you need to add DEBUG=-DDEBUG_THREAD or DEBUG=-DDEBUG_FULL compilation option. In addition to timing info on locks, we keep info on where a lock was acquired the last time (function name, file and line). There are also the thread id and a flag to know if it is still locked or not. This will be useful to debug deadlocks.
2017-10-12 14:09:09 +00:00
#endif /* DEBUG_THREAD */
MINOR: threads: Introduce double-width CAS on x86_64 and arm. Introduce double-width compare-and-swap on arches that support it, right now x86_64, arm, and aarch64. Also introduce functions to do memory barriers.
2017-12-21 16:13:05 +00:00
#ifdef __x86_64__
#define HA_HAVE_CAS_DW 1
#define HA_CAS_IS_8B
static __inline int
__ha_cas_dw(void *target, void *compare, const void *set)
{
char ret;
__asm __volatile("lock cmpxchg16b %0; setz %3"
: "+m" (*(void **)target),
"=a" (((void **)compare)[0]),
"=d" (((void **)compare)[1]),
"=q" (ret)
: "a" (((void **)compare)[0]),
"d" (((void **)compare)[1]),
"b" (((const void **)set)[0]),
"c" (((const void **)set)[1])
: "memory", "cc");
return (ret);
}
static __inline void
__ha_barrier_load(void)
{
__asm __volatile("lfence" ::: "memory");
}
static __inline void
__ha_barrier_store(void)
{
__asm __volatile("sfence" ::: "memory");
}
static __inline void
__ha_barrier_full(void)
{
__asm __volatile("mfence" ::: "memory");
}
#elif defined(__arm__) && (defined(__ARM_ARCH_7__) || defined(__ARM_ARCH_7A__))
#define HA_HAVE_CAS_DW 1
static __inline void
__ha_barrier_load(void)
{
__asm __volatile("dmb" ::: "memory");
}
static __inline void
__ha_barrier_store(void)
{
__asm __volatile("dsb" ::: "memory");
}
static __inline void
__ha_barrier_full(void)
{
__asm __volatile("dmb" ::: "memory");
}
BUG/MEDIUM: threads: fix the double CAS implementation for ARMv7 Commit f61f0cb ("MINOR: threads: Introduce double-width CAS on x86_64 and arm.") introduced the double CAS. But the ARMv7 version is bogus, it uses the value of the pointers instead of dereferencing them. When lucky, it simply doesn't build due to impossible registers combinations. Otherwise it will immediately crash at run time when facing traffic. No backport is needed, this bug was introduced in 1.9-dev.
2018-02-14 13:16:28 +00:00
static __inline int __ha_cas_dw(void *target, void *compare, const void *set)
MINOR: threads: Introduce double-width CAS on x86_64 and arm. Introduce double-width compare-and-swap on arches that support it, right now x86_64, arm, and aarch64. Also introduce functions to do memory barriers.
2017-12-21 16:13:05 +00:00
{
uint64_t previous;
int tmp;
__asm __volatile("1:"
"ldrexd %0, [%4];"
"cmp %Q0, %Q2;"
"ittt eq;"
"cmpeq %R0, %R2;"
"strexdeq %1, %3, [%4];"
"cmpeq %1, #1;"
"beq 1b;"
: "=&r" (previous), "=&r" (tmp)
BUG/MEDIUM: threads: fix the double CAS implementation for ARMv7 Commit f61f0cb ("MINOR: threads: Introduce double-width CAS on x86_64 and arm.") introduced the double CAS. But the ARMv7 version is bogus, it uses the value of the pointers instead of dereferencing them. When lucky, it simply doesn't build due to impossible registers combinations. Otherwise it will immediately crash at run time when facing traffic. No backport is needed, this bug was introduced in 1.9-dev.
2018-02-14 13:16:28 +00:00
: "r" (*(uint64_t *)compare), "r" (*(uint64_t *)set), "r" (target)
MINOR: threads: Introduce double-width CAS on x86_64 and arm. Introduce double-width compare-and-swap on arches that support it, right now x86_64, arm, and aarch64. Also introduce functions to do memory barriers.
2017-12-21 16:13:05 +00:00
: "memory", "cc");
tmp = (previous == *(uint64_t *)compare);
*(uint64_t *)compare = previous;
return (tmp);
}
#elif defined (__aarch64__)
#define HA_HAVE_CAS_DW 1
#define HA_CAS_IS_8B
static __inline void
__ha_barrier_load(void)
{
__asm __volatile("dmb ishld" ::: "memory");
}
static __inline void
__ha_barrier_store(void)
{
__asm __volatile("dmb ishst" ::: "memory");
}
static __inline void
__ha_barrier_full(void)
{
__asm __volatile("dmb ish" ::: "memory");
}
static __inline int __ha_cas_dw(void *target, void *compare, void *set)
{
void *value[2];
uint64_t tmp1, tmp2;
__asm__ __volatile__("1:"
"ldxp %0, %1, [%4];"
"mov %2, %0;"
"mov %3, %1;"
"eor %0, %0, %5;"
"eor %1, %1, %6;"
"orr %1, %0, %1;"
"mov %w0, #0;"
"cbnz %1, 2f;"
"stxp %w0, %7, %8, [%4];"
"cbnz %w0, 1b;"
"mov %w0, #1;"
"2:"
: "=&r" (tmp1), "=&r" (tmp2), "=&r" (value[0]), "=&r" (value[1])
: "r" (target), "r" (((void **)(compare))[0]), "r" (((void **)(compare))[1]), "r" (((void **)(set))[0]), "r" (((void **)(set))[1])
: "cc", "memory");
memcpy(compare, &value, sizeof(value));
return (tmp1);
}
#else
#define __ha_barrier_load __sync_synchronize
#define __ha_barrier_store __sync_synchronize
#define __ha_barrier_full __sync_synchronize
#endif
MEDIUM: threads: Add hathreads header file This file contains all functions and macros used to deal with concurrency in HAProxy. It contains all high-level function to do atomic operation (HA_ATOMIC_*). Note, for now, we rely on "__atomic" GCC builtins to do atomic operation. So HAProxy can be compiled with the thread support iff these builtins are available. It also contains wrappers around plocks to use spin or read/write locks. These wrappers are used to abstract the internal representation of the locking system and to add information to help debugging, when compiled with suitable options. To add extra info on locks, you need to add DEBUG=-DDEBUG_THREAD or DEBUG=-DDEBUG_FULL compilation option. In addition to timing info on locks, we keep info on where a lock was acquired the last time (function name, file and line). There are also the thread id and a flag to know if it is still locked or not. This will be useful to debug deadlocks.
2017-10-12 14:09:09 +00:00
#endif /* USE_THREAD */
MINOR: threads: Introduce double-width CAS on x86_64 and arm. Introduce double-width compare-and-swap on arches that support it, right now x86_64, arm, and aarch64. Also introduce functions to do memory barriers.
2017-12-21 16:13:05 +00:00
static inline void __ha_compiler_barrier(void)
{
__asm __volatile("" ::: "memory");
}
MINOR: cli/threads: make "show fd" report thread_sync_io_handler instead of "unknown" The output was confusing when the sync point's dummy handler was shown. This patch should be backported to 1.8 to help with troubleshooting.
2018-03-28 16:06:47 +00:00
/* Dummy I/O handler used by the sync pipe.*/
void thread_sync_io_handler(int fd);
MEDIUM: threads: Add hathreads header file This file contains all functions and macros used to deal with concurrency in HAProxy. It contains all high-level function to do atomic operation (HA_ATOMIC_*). Note, for now, we rely on "__atomic" GCC builtins to do atomic operation. So HAProxy can be compiled with the thread support iff these builtins are available. It also contains wrappers around plocks to use spin or read/write locks. These wrappers are used to abstract the internal representation of the locking system and to add information to help debugging, when compiled with suitable options. To add extra info on locks, you need to add DEBUG=-DDEBUG_THREAD or DEBUG=-DDEBUG_FULL compilation option. In addition to timing info on locks, we keep info on where a lock was acquired the last time (function name, file and line). There are also the thread id and a flag to know if it is still locked or not. This will be useful to debug deadlocks.
2017-10-12 14:09:09 +00:00
#endif /* _COMMON_HATHREADS_H */
Reference in New Issue Copy Permalink