generic-library/bionic
1
0
Fork 0
Code Issues Pull Requests Releases Wiki Activity

Add memory barriers to pthread_once.

Browse Source 
The implementation was using a double-checked locking approach that
could break on SMP.

In addition to the barriers I also switched to a volatile pointer.  I
don't think this will matter unless gcc can conclude that _normal_lock
can't affect *once_control, but I figured it was better to be safe.
(It seems to have no impact whatsoever on the generated code.)

Bug 3022795.

Change-Id: Ib91da25d57ff5bee4288526e39d457153ef6aacd
This commit is contained in:
Andy McFadden 2010-09-23 12:30:12 -07:00
parent 8939c1e1cc
commit b1c9cc2f2d
1 changed files with 7 additions and 3 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

10
libc/bionic/pthread.c
View File

@ -1886,12 +1886,16 @@ int pthread_getcpuclockid(pthread_t tid, clockid_t *clockid)
int pthread_once( pthread_once_t* once_control, void (*init_routine)(void) )
{
static pthread_mutex_t once_lock = PTHREAD_RECURSIVE_MUTEX_INITIALIZER;
volatile pthread_once_t* ocptr = once_control;
if (*once_control == PTHREAD_ONCE_INIT) {
pthread_once_t tmp = *ocptr;
ANDROID_MEMBAR_FULL();
if (tmp == PTHREAD_ONCE_INIT) {
pthread_mutex_lock( &once_lock );
if (*once_control == PTHREAD_ONCE_INIT) {
if (*ocptr == PTHREAD_ONCE_INIT) {
(*init_routine)();
*once_control = ~PTHREAD_ONCE_INIT;
ANDROID_MEMBAR_FULL();
*ocptr = ~PTHREAD_ONCE_INIT;
}
pthread_mutex_unlock( &once_lock );
}