From 09a9f9867bf0c7fd793f46f1e639ad2042175721 Mon Sep 17 00:00:00 2001 From: "Paul E. McKenney" Date: Sat, 18 Apr 2009 02:43:32 -0400 Subject: [PATCH] Support kernels with broken signal delivery Some Linux kernels seems to have problems sending _all_ the signals. Use a workaround which re-sends the signal when we wait for it to be executed. Signed-off-by: Mathieu Desnoyers --- urcu.c | 82 +++++++++++++++++++++++++++++++++++++++------------------- 1 file changed, 55 insertions(+), 27 deletions(-) diff --git a/urcu.c b/urcu.c index 475b4d1..38abf7d 100644 --- a/urcu.c +++ b/urcu.c @@ -14,6 +14,7 @@ #include #include #include +#include #include "urcu.h" @@ -39,6 +40,7 @@ long __thread urcu_active_readers; struct reader_registry { pthread_t tid; long *urcu_active_readers; + char *need_mb; }; #ifdef DEBUG_YIELD @@ -47,24 +49,34 @@ unsigned int __thread rand_yield; #endif static struct reader_registry *registry; +static char __thread need_mb; static int num_readers, alloc_readers; -#ifndef DEBUG_FULL_MB -static int sig_done; -#endif void internal_urcu_lock(void) { -#if 0 int ret; - /* Mutex sleeping does not play well with busy-waiting loop. */ + +#ifndef DISTRUST_SIGNALS_EXTREME ret = pthread_mutex_lock(&urcu_mutex); if (ret) { perror("Error in pthread mutex lock"); exit(-1); } -#endif - while (pthread_mutex_trylock(&urcu_mutex) != 0) - cpu_relax(); +#else /* #ifndef DISTRUST_SIGNALS_EXTREME */ + while ((ret = pthread_mutex_trylock(&urcu_mutex)) != 0) { + if (ret != EBUSY && ret != EINTR) { + printf("ret = %d, errno = %d\n", ret, errno); + perror("Error in pthread mutex lock"); + exit(-1); + } + if (need_mb) { + smp_mb(); + need_mb = 0; + smp_mb(); + } + poll(NULL,0,10); + } +#endif /* #else #ifndef DISTRUST_SIGNALS_EXTREME */ } void internal_urcu_unlock(void) @@ -87,7 +99,7 @@ static void switch_next_urcu_qparity(void) } #ifdef DEBUG_FULL_MB -static void force_mb_single_thread(pthread_t tid) +static void force_mb_single_thread(struct reader_registry *index) { smp_mb(); } @@ -98,26 +110,27 @@ static void force_mb_all_threads(void) } #else -static void force_mb_single_thread(pthread_t tid) +static void force_mb_single_thread(struct reader_registry *index) { assert(registry); - sig_done = 0; /* * pthread_kill has a smp_mb(). But beware, we assume it performs * a cache flush on architectures with non-coherent cache. Let's play * safe and don't assume anything : we use smp_mc() to make sure the * cache flush is enforced. - * smp_mb(); write sig_done before sending the signals */ - smp_mc(); /* write sig_done before sending the signals */ - pthread_kill(tid, SIGURCU); + *index->need_mb = 1; + smp_mc(); /* write ->need_mb before sending the signals */ + pthread_kill(index->tid, SIGURCU); + smp_mb(); /* * Wait for sighandler (and thus mb()) to execute on every thread. * BUSY-LOOP. */ - while (LOAD_SHARED(sig_done) < 1) - cpu_relax(); - smp_mb(); /* read sig_done before ending the barrier */ + while (*index->need_mb) { + poll(NULL, 0, 1); + } + smp_mb(); /* read ->need_mb before ending the barrier */ } static void force_mb_all_threads(void) @@ -129,24 +142,37 @@ static void force_mb_all_threads(void) */ if (!registry) return; - sig_done = 0; /* * pthread_kill has a smp_mb(). But beware, we assume it performs * a cache flush on architectures with non-coherent cache. Let's play * safe and don't assume anything : we use smp_mc() to make sure the * cache flush is enforced. - * smp_mb(); write sig_done before sending the signals */ - smp_mc(); /* write sig_done before sending the signals */ - for (index = registry; index < registry + num_readers; index++) + for (index = registry; index < registry + num_readers; index++) { + *index->need_mb = 1; + smp_mc(); /* write need_mb before sending the signal */ pthread_kill(index->tid, SIGURCU); + } /* * Wait for sighandler (and thus mb()) to execute on every thread. - * BUSY-LOOP. + * + * Note that the pthread_kill() will never be executed on systems + * that correctly deliver signals in a timely manner. However, it + * is not uncommon for kernels to have bugs that can result in + * lost or unduly delayed signals. + * + * If you are seeing the below pthread_kill() executing much at + * all, we suggest testing the underlying kernel and filing the + * relevant bug report. For Linux kernels, we recommend getting + * the Linux Test Project (LTP). */ - while (LOAD_SHARED(sig_done) < num_readers) - cpu_relax(); - smp_mb(); /* read sig_done before ending the barrier */ + for (index = registry; index < registry + num_readers; index++) { + while (*index->need_mb) { + pthread_kill(index->tid, SIGURCU); + poll(NULL, 0, 1); + } + } + smp_mb(); /* read ->need_mb before ending the barrier */ } #endif @@ -167,7 +193,7 @@ void wait_for_quiescent_state(void) */ while (rcu_old_gp_ongoing(index->urcu_active_readers)) { if (wait_loops++ == KICK_READER_LOOPS) { - force_mb_single_thread(index->tid); + force_mb_single_thread(index); wait_loops = 0; } else { cpu_relax(); @@ -254,6 +280,7 @@ void urcu_add_reader(pthread_t id) registry[num_readers].tid = id; /* reference to the TLS of _this_ reader thread. */ registry[num_readers].urcu_active_readers = &urcu_active_readers; + registry[num_readers].need_mb = &need_mb; num_readers++; } @@ -303,7 +330,8 @@ void sigurcu_handler(int signo, siginfo_t *siginfo, void *context) * executed on. */ smp_mb(); - atomic_inc(&sig_done); + need_mb = 0; + smp_mb(); } void __attribute__((constructor)) urcu_init(void) -- 2.34.1