mirrors/haproxy
1
0
Fork 0
mirror of https://git.haproxy.org/git/haproxy.git/ synced 2026-05-04 12:41:00 +02:00
Code Issues Projects Releases Wiki Activity

CLEANUP: atomic: add an explicit _FETCH variant for add/sub/and/or

Browse Source 
Currently our atomic ops return a value but it's never known whether
the fetch is done before or after the operation, which causes some
confusion each time the value is desired. Let's create an explicit
variant of these operations suffixed with _FETCH to explicitly mention
that the fetch occurs after the operation, and make use of it at the
few call places.
This commit is contained in:
Willy Tarreau 2021-04-06 11:44:07 +02:00
parent 6756d95a8e
commit 1db427399c
14 changed files with 60 additions and 20 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

40
include/haproxy/atomic.h
View File

@ -77,6 +77,12 @@
#define HA_ATOMIC_OR(val, flags) ({*(val) |= (flags);})
#define HA_ATOMIC_ADD(val, i) ({*(val) += (i);})
#define HA_ATOMIC_SUB(val, i) ({*(val) -= (i);})
#define HA_ATOMIC_AND_FETCH(val, flags) ({ *(val) &= (flags); })
#define HA_ATOMIC_OR_FETCH(val, flags) ({ *(val) |= (flags); })
#define HA_ATOMIC_ADD_FETCH(val, i) ({ *(val) += (i); })
#define HA_ATOMIC_SUB_FETCH(val, i) ({ *(val) -= (i); })
#define HA_ATOMIC_XADD(val, i) \
({ \
typeof((val)) __p_xadd = (val); \
@ -196,6 +202,12 @@
#define HA_ATOMIC_OR(val, flags) __sync_or_and_fetch(val, flags)
#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_FETCH(val, flags) __sync_and_and_fetch(val, flags)
#define HA_ATOMIC_OR_FETCH(val, flags) __sync_or_and_fetch(val, flags)
#define HA_ATOMIC_ADD_FETCH(val, i) __sync_add_and_fetch(val, i)
#define HA_ATOMIC_SUB_FETCH(val, i) __sync_sub_and_fetch(val, i)
#define HA_ATOMIC_XADD(val, i) __sync_fetch_and_add(val, i)
#define HA_ATOMIC_BTS(val, bit) \
@ -271,6 +283,12 @@
#define HA_ATOMIC_OR(val, flags) __atomic_or_fetch(val, flags, __ATOMIC_SEQ_CST)
#define HA_ATOMIC_ADD(val, i) __atomic_add_fetch(val, i, __ATOMIC_SEQ_CST)
#define HA_ATOMIC_SUB(val, i) __atomic_sub_fetch(val, i, __ATOMIC_SEQ_CST)
#define HA_ATOMIC_AND_FETCH(val, flags) __atomic_and_fetch(val, flags, __ATOMIC_SEQ_CST)
#define HA_ATOMIC_OR_FETCH(val, flags) __atomic_or_fetch(val, flags, __ATOMIC_SEQ_CST)
#define HA_ATOMIC_ADD_FETCH(val, i) __atomic_add_fetch(val, i, __ATOMIC_SEQ_CST)
#define HA_ATOMIC_SUB_FETCH(val, i) __atomic_sub_fetch(val, i, __ATOMIC_SEQ_CST)
#define HA_ATOMIC_XADD(val, i) __atomic_fetch_add(val, i, __ATOMIC_SEQ_CST)
#define HA_ATOMIC_BTS(val, bit) \
@ -326,6 +344,12 @@
#define _HA_ATOMIC_OR(val, flags) __atomic_or_fetch(val, flags, __ATOMIC_RELAXED)
#define _HA_ATOMIC_ADD(val, i) __atomic_add_fetch(val, i, __ATOMIC_RELAXED)
#define _HA_ATOMIC_SUB(val, i) __atomic_sub_fetch(val, i, __ATOMIC_RELAXED)
#define _HA_ATOMIC_AND_FETCH(val, flags) __atomic_and_fetch(val, flags, __ATOMIC_RELAXED)
#define _HA_ATOMIC_OR_FETCH(val, flags) __atomic_or_fetch(val, flags, __ATOMIC_RELAXED)
#define _HA_ATOMIC_ADD_FETCH(val, i) __atomic_add_fetch(val, i, __ATOMIC_RELAXED)
#define _HA_ATOMIC_SUB_FETCH(val, i) __atomic_sub_fetch(val, i, __ATOMIC_RELAXED)
#define _HA_ATOMIC_XADD(val, i) __atomic_fetch_add(val, i, __ATOMIC_RELAXED)
#define _HA_ATOMIC_CAS(val, old, new) __atomic_compare_exchange_n(val, old, new, 0, __ATOMIC_RELAXED, __ATOMIC_RELAXED)
/* warning, n is a pointer to the double value for dwcas */
@ -634,6 +658,10 @@ static inline void __ha_compiler_barrier(void)
#define _HA_ATOMIC_ADD HA_ATOMIC_ADD
#endif /* !_HA_ATOMIC_ADD */
#ifndef _HA_ATOMIC_ADD_FETCH
#define _HA_ATOMIC_ADD_FETCH HA_ATOMIC_ADD_FETCH
#endif /* !_HA_ATOMIC_ADD_FETCH */
#ifndef _HA_ATOMIC_XADD
#define _HA_ATOMIC_XADD HA_ATOMIC_XADD
#endif /* !_HA_ATOMIC_SUB */
@ -642,14 +670,26 @@ static inline void __ha_compiler_barrier(void)
#define _HA_ATOMIC_SUB HA_ATOMIC_SUB
#endif /* !_HA_ATOMIC_SUB */
#ifndef _HA_ATOMIC_SUB_FETCH
#define _HA_ATOMIC_SUB_FETCH HA_ATOMIC_SUB_FETCH
#endif /* !_HA_ATOMIC_SUB_FETCH */
#ifndef _HA_ATOMIC_AND
#define _HA_ATOMIC_AND HA_ATOMIC_AND
#endif /* !_HA_ATOMIC_AND */
#ifndef _HA_ATOMIC_AND_FETCH
#define _HA_ATOMIC_AND_FETCH HA_ATOMIC_AND_FETCH
#endif /* !_HA_ATOMIC_AND_FETCH */
#ifndef _HA_ATOMIC_OR
#define _HA_ATOMIC_OR HA_ATOMIC_OR
#endif /* !_HA_ATOMIC_OR */
#ifndef _HA_ATOMIC_OR_FETCH
#define _HA_ATOMIC_OR_FETCH HA_ATOMIC_OR_FETCH
#endif /* !_HA_ATOMIC_OR_FETCH */
#ifndef _HA_ATOMIC_XCHG
#define _HA_ATOMIC_XCHG HA_ATOMIC_XCHG
#endif /* !_HA_ATOMIC_XCHG */

4
include/haproxy/fd.h
View File

@ -126,7 +126,7 @@ static inline void done_update_polling(int fd)
{
unsigned long update_mask;
update_mask = _HA_ATOMIC_AND(&fdtab[fd].update_mask, ~tid_bit);
update_mask = _HA_ATOMIC_AND_FETCH(&fdtab[fd].update_mask, ~tid_bit);
while ((update_mask & all_threads_mask)== 0) {
/* If we were the last one that had to update that entry, remove it from the list */
fd_rm_from_fd_list(&update_list, fd, offsetof(struct fdtab, update));
@ -346,7 +346,7 @@ static inline int fd_set_running(int fd)
*/
static inline long fd_clr_running(int fd)
{
return _HA_ATOMIC_AND(&fdtab[fd].running_mask, ~tid_bit);
return _HA_ATOMIC_AND_FETCH(&fdtab[fd].running_mask, ~tid_bit);
}
/* Update events seen for FD <fd> and its state if needed. This should be

8
include/haproxy/freq_ctr.h
View File

@ -51,7 +51,7 @@ static inline unsigned int update_freq_ctr(struct freq_ctr *ctr, unsigned int in
do {
now_tmp = global_now >> 32;
if (curr_sec == (now_tmp & 0x7fffffff))
return _HA_ATOMIC_ADD(&ctr->curr_ctr, inc);
return _HA_ATOMIC_ADD_FETCH(&ctr->curr_ctr, inc);
/* remove the bit, used for the lock */
curr_sec &= 0x7fffffff;
@ -72,7 +72,7 @@ static inline unsigned int update_freq_ctr(struct freq_ctr *ctr, unsigned int in
/* release the lock and update the time in case of rotate. */
_HA_ATOMIC_STORE(&ctr->curr_sec, curr_sec & 0x7fffffff);
return _HA_ATOMIC_ADD(&ctr->curr_ctr, inc);
return _HA_ATOMIC_ADD_FETCH(&ctr->curr_ctr, inc);
}
/* Update a frequency counter by <inc> incremental units. It is automatically
@ -90,7 +90,7 @@ static inline unsigned int update_freq_ctr_period(struct freq_ctr_period *ctr,
do {
now_ms_tmp = global_now_ms;
if (now_ms_tmp - curr_tick < period)
return _HA_ATOMIC_ADD(&ctr->curr_ctr, inc);
return _HA_ATOMIC_ADD_FETCH(&ctr->curr_ctr, inc);
/* remove the bit, used for the lock */
curr_tick &= ~1;
@ -112,7 +112,7 @@ static inline unsigned int update_freq_ctr_period(struct freq_ctr_period *ctr,
/* release the lock and update the time in case of rotate. */
_HA_ATOMIC_STORE(&ctr->curr_tick, curr_tick);
return _HA_ATOMIC_ADD(&ctr->curr_ctr, inc);
return _HA_ATOMIC_ADD_FETCH(&ctr->curr_ctr, inc);
}
/* Read a frequency counter taking history into account for missing time in

2
include/haproxy/pattern.h
View File

@ -204,7 +204,7 @@ void pat_ref_reload(struct pat_ref *ref, struct pat_ref *replace);
*/
static inline unsigned int pat_ref_newgen(struct pat_ref *ref)
{
return HA_ATOMIC_ADD(&ref->next_gen, 1);
return HA_ATOMIC_ADD_FETCH(&ref->next_gen, 1);
}
/* Give up a previously assigned generation number. By doing this the caller

4
include/haproxy/server.h
View File

@ -252,7 +252,7 @@ static inline void srv_use_conn(struct server *srv, struct connection *conn)
{
unsigned int curr;
curr = _HA_ATOMIC_ADD(&srv->curr_used_conns, 1);
curr = _HA_ATOMIC_ADD_FETCH(&srv->curr_used_conns, 1);
/* It's ok not to do that atomically, we don't need an
* exact max.
@ -318,7 +318,7 @@ static inline int srv_add_to_idle_list(struct server *srv, struct connection *co
!conn->mux->used_streams(conn) && conn->mux->avail_streams(conn)) {
int retadd;
retadd = _HA_ATOMIC_ADD(&srv->curr_idle_conns, 1);
retadd = _HA_ATOMIC_ADD_FETCH(&srv->curr_idle_conns, 1);
if (retadd > srv->max_idle_conns) {
_HA_ATOMIC_SUB(&srv->curr_idle_conns, 1);
return 0;

2
include/haproxy/task.h
View File

@ -207,7 +207,7 @@ static inline void _task_wakeup(struct task *t, unsigned int f, const char *file
{
unsigned int state;
state = _HA_ATOMIC_OR(&t->state, f);
state = _HA_ATOMIC_OR_FETCH(&t->state, f);
while (!(state & (TASK_RUNNING | TASK_QUEUED))) {
if (_HA_ATOMIC_CAS(&t->state, &state, state | TASK_QUEUED)) {
#ifdef DEBUG_TASK

2
src/backend.c
View File

@ -1703,7 +1703,7 @@ skip_reuse:
int count;
s->flags |= SF_CURR_SESS;
count = _HA_ATOMIC_ADD(&srv->cur_sess, 1);
count = _HA_ATOMIC_ADD_FETCH(&srv->cur_sess, 1);
HA_ATOMIC_UPDATE_MAX(&srv->counters.cur_sess_max, count);
if (s->be->lbprm.server_take_conn)
s->be->lbprm.server_take_conn(srv, 0);

2
src/dict.c
View File

@ -116,7 +116,7 @@ void dict_entry_unref(struct dict *d, struct dict_entry *de)
if (!de)
return;
if (HA_ATOMIC_SUB(&de->refcount, 1) != 0)
if (HA_ATOMIC_SUB_FETCH(&de->refcount, 1) != 0)
return;
HA_RWLOCK_WRLOCK(DICT_LOCK, &d->rwlock);

4
src/fd.c
View File

@ -374,7 +374,7 @@ int fd_takeover(int fd, void *expected_owner)
int ret = -1;
#ifndef HA_HAVE_CAS_DW
if (_HA_ATOMIC_OR(&fdtab[fd].running_mask, tid_bit) == tid_bit) {
if (_HA_ATOMIC_OR_FETCH(&fdtab[fd].running_mask, tid_bit) == tid_bit) {
HA_RWLOCK_WRLOCK(OTHER_LOCK, &fd_mig_lock);
if (fdtab[fd].owner == expected_owner) {
fdtab[fd].thread_mask = tid_bit;
@ -388,7 +388,7 @@ int fd_takeover(int fd, void *expected_owner)
new_masks[0] = new_masks[1] = tid_bit;
old_masks[0] = _HA_ATOMIC_OR(&fdtab[fd].running_mask, tid_bit);
old_masks[0] = _HA_ATOMIC_OR_FETCH(&fdtab[fd].running_mask, tid_bit);
old_masks[1] = fdtab[fd].thread_mask;
/* protect ourself against a delete then an insert for the same fd,

2
src/haproxy.c
View File

@ -2385,7 +2385,7 @@ void run_poll_loop()
int i;
if (stopping) {
if (_HA_ATOMIC_OR(&stopping_thread_mask, tid_bit) == tid_bit) {
if (_HA_ATOMIC_OR_FETCH(&stopping_thread_mask, tid_bit) == tid_bit) {
/* notify all threads that stopping was just set */
for (i = 0; i < global.nbthread; i++)
if (((all_threads_mask & ~stopping_thread_mask) >> i) & 1)

2
src/proxy.c
View File

@ -2126,7 +2126,7 @@ int stream_set_backend(struct stream *s, struct proxy *be)
s->be = be;
HA_ATOMIC_UPDATE_MAX(&be->be_counters.conn_max,
HA_ATOMIC_ADD(&be->beconn, 1));
HA_ATOMIC_ADD_FETCH(&be->beconn, 1));
proxy_inc_be_ctr(be);
/* assign new parameters to the stream from the new backend */

4
src/queue.c
View File

@ -400,7 +400,7 @@ struct pendconn *pendconn_add(struct stream *strm)
if (srv) {
unsigned int old_max, new_max;
new_max = _HA_ATOMIC_ADD(&srv->nbpend, 1);
new_max = _HA_ATOMIC_ADD_FETCH(&srv->nbpend, 1);
old_max = srv->counters.nbpend_max;
while (new_max > old_max) {
if (likely(_HA_ATOMIC_CAS(&srv->counters.nbpend_max, &old_max, new_max)))
@ -416,7 +416,7 @@ struct pendconn *pendconn_add(struct stream *strm)
else {
unsigned int old_max, new_max;
new_max = _HA_ATOMIC_ADD(&px->nbpend, 1);
new_max = _HA_ATOMIC_ADD_FETCH(&px->nbpend, 1);
old_max = px->be_counters.nbpend_max;
while (new_max > old_max) {
if (likely(_HA_ATOMIC_CAS(&px->be_counters.nbpend_max, &old_max, new_max)))

2
src/ssl_sock.c
View File

@ -1943,7 +1943,7 @@ ssl_sock_do_create_cert(const char *servername, struct bind_conf *bind_conf, SSL
* number */
if (X509_set_version(newcrt, 2L) != 1)
goto mkcert_error;
ASN1_INTEGER_set(X509_get_serialNumber(newcrt), _HA_ATOMIC_ADD(&ssl_ctx_serial, 1));
ASN1_INTEGER_set(X509_get_serialNumber(newcrt), _HA_ATOMIC_ADD_FETCH(&ssl_ctx_serial, 1));
/* Set duration for the certificate */
if (!X509_gmtime_adj(X509_getm_notBefore(newcrt), (long)-60*60*24) ||

2
src/task.c
View File

@ -585,7 +585,7 @@ unsigned int run_tasks_from_lists(unsigned int budgets[])
HA_ATOMIC_ADD(&profile_entry->cpu_time, cpu);
}
state = _HA_ATOMIC_AND(&t->state, ~TASK_RUNNING);
state = _HA_ATOMIC_AND_FETCH(&t->state, ~TASK_RUNNING);
if (unlikely(state & TASK_KILLED)) {
task_unlink_wq(t);
__task_free(t);