r15251@tombo: nickm | 2008-04-22 11:59:46 -0400

On platforms using pthreads, allow a thread to acquire a lock it already holds.  This is crucial for logging: otherwise any log message thrown from inside the logging process (especially from control.c) will deadlock.  Win32 CriticalSections are already recursive.  Bug spotted by nwf.  Bugfix on 0.2.0.16-alpha.  Backport candidate. I hope this is portable.


svn:r14406
This commit is contained in:
Nick Mathewson 2008-04-22 15:59:59 +00:00
parent 0cbecbe050
commit 299014b2c7
2 changed files with 9 additions and 1 deletions

View File

@ -24,6 +24,9 @@ Changes in version 0.2.1.1-alpha - 2008-??-??
most smartlists hold around 8-12 elements tops.
- Avoid allocating extra space when computing consensuses on
64-bit platforms. Bug spotted by aakova.
- Use recursive pthread mutexes in order to avoid deadlock when
logging debug-level messages to a controller. Bug spotted by
nwf, bugfix on 0.2.0.16-alpha.
o Minor features:
- Allow separate log levels to be configured for different logging

View File

@ -1735,13 +1735,14 @@ tor_get_thread_id(void)
struct tor_mutex_t {
pthread_mutex_t mutex;
};
static pthread_mutexattr_t attr_reentrant;
/** Allocate and return new lock. */
tor_mutex_t *
tor_mutex_new(void)
{
int err;
tor_mutex_t *mutex = tor_malloc_zero(sizeof(tor_mutex_t));
err = pthread_mutex_init(&mutex->mutex, NULL);
err = pthread_mutex_init(&mutex->mutex, &attr_reentrant);
if (PREDICT_UNLIKELY(err)) {
log_err(LD_GENERAL, "Error %d creating a mutex.", err);
tor_fragile_assert();
@ -1857,6 +1858,10 @@ tor_cond_signal_all(tor_cond_t *cond)
void
tor_threads_init(void)
{
#ifdef USE_PTHREADS
pthread_mutexattr_init(&attr_reentrant);
pthread_mutexattr_settype(&attr_reentrant, PTHREAD_MUTEX_RECURSIVE);
#endif
}
#elif defined(USE_WIN32_THREADS)
#if 0