Timer hash table for debug builds

pull/12663/head
Sree Kuchibhotla 7 years ago
parent c2261f21e1
commit 96d23c3f08
  1. 143
      src/core/lib/iomgr/timer_generic.c

@ -79,67 +79,121 @@ static timer_shard g_shards[NUM_SHARDS];
* Access to this is protected by g_shared_mutables.mu */
static timer_shard *g_shard_queue[NUM_SHARDS];
#ifndef NDEBUG
/* == Hash table for duplicate timer detection == */
#define NUM_HASH_BUCKETS 1000
#define NUM_SLOTS_PER_BUCKET 5
static gpr_mu g_hash_mu;
static grpc_timer *g_timer_hash[1000][5] = {{NULL, NULL}};
#define NUM_SLOTS_PER_BUCKET 30
static void init_timer_hash() { gpr_mu_init(&g_hash_mu); }
static gpr_mu g_hash_mu[NUM_HASH_BUCKETS]; /* One mutex per bucket */
static grpc_timer *g_timer_ht[NUM_HASH_BUCKETS]
[NUM_SLOTS_PER_BUCKET] = {{NULL, NULL}};
static bool is_timer_present(grpc_timer *t) {
static void init_timer_ht() {
for (int i = 0; i < NUM_HASH_BUCKETS; i++) {
gpr_mu_init(&g_hash_mu[i]);
}
}
static bool is_in_ht(grpc_timer *t) {
size_t i = GPR_HASH_POINTER(t, NUM_HASH_BUCKETS);
bool is_found = false;
gpr_mu_lock(&g_hash_mu);
gpr_mu_lock(&g_hash_mu[i]);
for (int j = 0; j < NUM_SLOTS_PER_BUCKET; j++) {
if (g_timer_hash[i][j] == t) {
if (g_timer_ht[i][j] == t) {
is_found = true;
break;
}
}
gpr_mu_unlock(&g_hash_mu);
gpr_mu_unlock(&g_hash_mu[i]);
return is_found;
}
static void check_and_add_timer(grpc_timer *t) {
static void add_to_ht(grpc_timer *t) {
size_t i = GPR_HASH_POINTER(t, NUM_HASH_BUCKETS);
bool added = false;
gpr_mu_lock(&g_hash_mu);
gpr_mu_lock(&g_hash_mu[i]);
for (int j = 0; j < NUM_SLOTS_PER_BUCKET; j++) {
if (g_timer_hash[i][j] == NULL) {
g_timer_hash[i][j] = t;
if (g_timer_ht[i][j] == NULL) {
g_timer_ht[i][j] = t;
added = true;
break;
} else if (g_timer_hash[i][j] == t) {
gpr_log(GPR_ERROR, "*** DUPLICATE TIMER BEING ADDED (%p) **", (void *)t);
} else if (g_timer_ht[i][j] == t) {
grpc_closure *c = t->closure;
gpr_log(GPR_ERROR,
"** Duplicate timer (%p) being added. Closure: (%p), created at: "
"(%s:%d), scheduled at: (%s:%d) **",
t, c, c->file_created, c->line_created, c->file_initiated,
c->line_initiated);
abort();
}
}
gpr_mu_unlock(&g_hash_mu);
gpr_mu_unlock(&g_hash_mu[i]);
if (!added) {
gpr_log(GPR_ERROR, "** NOT ENOUGH BUCKETS **");
gpr_log(GPR_ERROR,
"** Not enough slots in the timer hash table. Please increase "
"NUM_SLOTS_PER_BUCKET **");
abort();
}
}
static void remove_timer(grpc_timer *t) {
static void remove_from_ht(grpc_timer *t) {
size_t i = GPR_HASH_POINTER(t, NUM_HASH_BUCKETS);
bool removed = false;
gpr_mu_lock(&g_hash_mu);
gpr_mu_lock(&g_hash_mu[i]);
for (int j = 0; j < NUM_SLOTS_PER_BUCKET; j++) {
if (g_timer_hash[i][j] == t) {
g_timer_hash[i][j] = 0;
if (g_timer_ht[i][j] == t) {
g_timer_ht[i][j] = 0;
removed = true;
break;
}
}
gpr_mu_unlock(&g_hash_mu);
gpr_mu_unlock(&g_hash_mu[i]);
if (!removed) {
gpr_log(GPR_ERROR, "*** Unable to remove %p. BUG! **", (void *)t);
grpc_closure *c = t->closure;
gpr_log(GPR_ERROR,
"** Removing timer (%p) that is not added to hash table. Closure "
"(%p), created at: (%s:%d), scheduled at: (%s:%d) **",
t, c, c->file_created, c->line_created, c->file_initiated,
c->line_initiated);
abort();
}
}
/* If a timer is added to a timer shard (either heap or a list), it cannot
* be pending. A timer is added to hash table only-if it is added to the
* timer shard.
* Therefore, if timer->pending is false, it cannot be in hash table */
static void validate_non_pending_timer(grpc_timer *t) {
if (!t->pending && is_in_ht(t)) {
grpc_closure *c = t->closure;
gpr_log(GPR_ERROR,
"** gpr_timer_cancel() called on a non-pending timer (%p) which "
"is in the hash table. Closure: (%p), created at: (%s:%d), "
"scheduled at: (%s:%d) **",
t, c, c->file_created, c->line_created, c->file_initiated,
c->line_initiated);
abort();
}
}
#define INIT_TIMER_HASH_TABLE() init_timer_ht()
#define ADD_TO_HASH_TABLE(t) add_to_ht((t))
#define REMOVE_FROM_HASH_TABLE(t) remove_from_ht((t))
#define VALIDATE_NON_PENDING_TIMER(t) validate_non_pending_timer((t))
#else
#define INIT_TIMER_HASH_TABLE()
#define ADD_TO_HASH_TABLE(t)
#define REMOVE_FROM_HASH_TABLE(t)
#define VALIDATE_NON_PENDING_TIMER(t)
#endif
/* Thread local variable that stores the deadline of the next timer the thread
* has last-seen. This is an optimization to prevent the thread from checking
* shared_mutables.min_timer (which requires acquiring shared_mutables.mu lock,
@ -237,7 +291,7 @@ void grpc_timer_list_init(gpr_timespec now) {
g_shard_queue[i] = shard;
}
init_timer_hash();
INIT_TIMER_HASH_TABLE();
}
void grpc_timer_list_shutdown(grpc_exec_ctx *exec_ctx) {
@ -309,9 +363,8 @@ void grpc_timer_init(grpc_exec_ctx *exec_ctx, grpc_timer *timer,
gpr_atm deadline_atm = timer->deadline = timespec_to_atm_round_up(deadline);
if (GRPC_TRACER_ON(grpc_timer_trace)) {
gpr_log(GPR_DEBUG,
"TIMER %p: SET %" PRId64 ".%09d [%" PRIdPTR "] now %" PRId64
".%09d [%" PRIdPTR "] call %p[%p]",
gpr_log(GPR_DEBUG, "TIMER %p: SET %" PRId64 ".%09d [%" PRIdPTR
"] now %" PRId64 ".%09d [%" PRIdPTR "] call %p[%p]",
timer, deadline.tv_sec, deadline.tv_nsec, deadline_atm, now.tv_sec,
now.tv_nsec, timespec_to_atm_round_down(now), closure, closure->cb);
}
@ -337,8 +390,7 @@ void grpc_timer_init(grpc_exec_ctx *exec_ctx, grpc_timer *timer,
grpc_time_averaged_stats_add_sample(&shard->stats,
ts_to_dbl(gpr_time_sub(deadline, now)));
/** TODO: sreek - CHECK HERE AND ADD **/
check_and_add_timer(timer);
ADD_TO_HASH_TABLE(timer);
if (deadline_atm < shard->queue_deadline_cap) {
is_first_timer = grpc_timer_heap_add(&shard->heap, timer);
@ -347,9 +399,8 @@ void grpc_timer_init(grpc_exec_ctx *exec_ctx, grpc_timer *timer,
list_join(&shard->list, timer);
}
if (GRPC_TRACER_ON(grpc_timer_trace)) {
gpr_log(GPR_DEBUG,
" .. add to shard %d with queue_deadline_cap=%" PRIdPTR
" => is_first_timer=%s",
gpr_log(GPR_DEBUG, " .. add to shard %d with queue_deadline_cap=%" PRIdPTR
" => is_first_timer=%s",
(int)(shard - g_shards), shard->queue_deadline_cap,
is_first_timer ? "true" : "false");
}
@ -404,8 +455,7 @@ void grpc_timer_cancel(grpc_exec_ctx *exec_ctx, grpc_timer *timer) {
}
if (timer->pending) {
/* TODO: sreek - Remove the timer here */
remove_timer(timer);
REMOVE_FROM_HASH_TABLE(timer);
GRPC_CLOSURE_SCHED(exec_ctx, timer->closure, GRPC_ERROR_CANCELLED);
timer->pending = false;
@ -415,12 +465,7 @@ void grpc_timer_cancel(grpc_exec_ctx *exec_ctx, grpc_timer *timer) {
grpc_timer_heap_remove(&shard->heap, timer);
}
} else {
if (is_timer_present(timer)) {
gpr_log(GPR_ERROR,
"** gpr_timer_cancel called on a non-pending timer! %p",
(void *)timer);
abort();
}
VALIDATE_NON_PENDING_TIMER(timer);
}
gpr_mu_unlock(&shard->mu);
}
@ -452,18 +497,6 @@ static int refill_heap(timer_shard *shard, gpr_atm now) {
for (timer = shard->list.next; timer != &shard->list; timer = next) {
next = timer->next;
#ifndef NDEBUG
if (next == timer && next != &shard->list) {
grpc_closure *c = timer->closure;
gpr_log(GPR_ERROR,
"We have a problem!!!! - timer %p closure: %p, created-at: "
"[%s,%d], scheduled-at: [%s, %d]",
timer, c, c->file_initiated, c->line_created, c->file_initiated,
c->line_initiated);
abort();
}
#endif
if (timer->deadline < shard->queue_deadline_cap) {
if (GRPC_TRACER_ON(grpc_timer_check_trace)) {
gpr_log(GPR_DEBUG, " .. add timer with deadline %" PRIdPTR " to heap",
@ -516,8 +549,7 @@ static size_t pop_timers(grpc_exec_ctx *exec_ctx, timer_shard *shard,
grpc_timer *timer;
gpr_mu_lock(&shard->mu);
while ((timer = pop_one(shard, now))) {
/* TODO: sreek: Remove timer here */
remove_timer(timer);
REMOVE_FROM_HASH_TABLE(timer);
GRPC_CLOSURE_SCHED(exec_ctx, timer->closure, GRPC_ERROR_REF(error));
n++;
}
@ -630,9 +662,8 @@ grpc_timer_check_result grpc_timer_check(grpc_exec_ctx *exec_ctx,
gpr_asprintf(&next_str, "%" PRId64 ".%09d [%" PRIdPTR "]", next->tv_sec,
next->tv_nsec, timespec_to_atm_round_down(*next));
}
gpr_log(GPR_DEBUG,
"TIMER CHECK BEGIN: now=%" PRId64 ".%09d [%" PRIdPTR
"] next=%s tls_min=%" PRIdPTR " glob_min=%" PRIdPTR,
gpr_log(GPR_DEBUG, "TIMER CHECK BEGIN: now=%" PRId64 ".%09d [%" PRIdPTR
"] next=%s tls_min=%" PRIdPTR " glob_min=%" PRIdPTR,
now.tv_sec, now.tv_nsec, now_atm, next_str,
gpr_tls_get(&g_last_seen_min_timer),
gpr_atm_no_barrier_load(&g_shared_mutables.min_timer));

Loading…
Cancel
Save