99 *
1010 *
1111 * IDENTIFICATION
12- * $Header: /cvsroot/pgsql/src/backend/storage/lmgr/s_lock.c,v 1.14 2003/08/04 15:28:33 tgl Exp $
12+ * $Header: /cvsroot/pgsql/src/backend/storage/lmgr/s_lock.c,v 1.15 2003/08/06 16:43:43 tgl Exp $
1313 *
1414 *-------------------------------------------------------------------------
1515 */
@@ -31,7 +31,7 @@ s_lock_stuck(volatile slock_t *lock, const char *file, int line)
3131fprintf (stderr ,
3232"\nStuck spinlock (%p) detected at %s:%d.\n" ,
3333lock ,file ,line );
34- abort ( );
34+ exit ( 1 );
3535#else
3636elog (PANIC ,"stuck spinlock (%p) detected at %s:%d" ,
3737lock ,file ,line );
@@ -45,35 +45,68 @@ s_lock_stuck(volatile slock_t *lock, const char *file, int line)
4545void
4646s_lock (volatile slock_t * lock ,const char * file ,int line )
4747{
48- unsigned spins = 0 ;
49- unsigned delays = 0 ;
50- struct timeval delay ;
51-
5248/*
5349 * We loop tightly for awhile, then delay using select() and try
5450 * again. Preferably, "awhile" should be a small multiple of the
5551 * maximum time we expect a spinlock to be held. 100 iterations seems
56- * about right.
52+ * about right. In most multi-CPU scenarios, the spinlock is probably
53+ * held by a process on another CPU and will be released before we
54+ * finish 100 iterations. However, on a uniprocessor, the tight loop
55+ * is just a waste of cycles, so don't iterate thousands of times.
56+ *
57+ * Once we do decide to block, we use randomly increasing select() delays.
58+ * The first delay is 10 msec, then the delay randomly increases to about
59+ * one second, after which we reset to 10 msec and start again. The idea
60+ * here is that in the presence of heavy contention we need to increase
61+ * the delay, else the spinlock holder may never get to run and release
62+ * the lock. (Consider situation where spinlock holder has been nice'd
63+ * down in priority by the scheduler --- it will not get scheduled until
64+ * all would-be acquirers are sleeping, so if we always use a 10-msec
65+ * sleep, there is a real possibility of starvation.) But we can't just
66+ * clamp the delay to an upper bound, else it would take a long time to
67+ * make a reasonable number of tries.
5768 *
58- * We use a 10 millisec select delay because that is the lower limit on
59- * many platforms.The timeout is figured on this delay only, and so
60- * the nominal 1 minute is a lower bound.
69+ * We time out and declare error after NUM_DELAYS delays (thus, exactly
70+ * that many tries). With the given settings, this will usually take
71+ * 3 or so minutes. It seems better to fix the total number of tries (and
72+ * thus the probability of unintended failure) than to fix the total time
73+ * spent.
74+ *
75+ * The select() delays are measured in centiseconds (0.01 sec) because
76+ * 10 msec is a common resolution limit at the OS level.
6177 */
6278#define SPINS_PER_DELAY 100
63- #define DELAY_MSEC 10
64- #define TIMEOUT_MSEC (60 * 1000)
79+ #define NUM_DELAYS 1000
80+ #define MIN_DELAY_CSEC 1
81+ #define MAX_DELAY_CSEC 100
82+
83+ int spins = 0 ;
84+ int delays = 0 ;
85+ int cur_delay = MIN_DELAY_CSEC ;
86+ struct timeval delay ;
6587
6688while (TAS (lock ))
6789{
6890if (++ spins > SPINS_PER_DELAY )
6991{
70- if (++ delays > ( TIMEOUT_MSEC / DELAY_MSEC ) )
92+ if (++ delays > NUM_DELAYS )
7193s_lock_stuck (lock ,file ,line );
7294
73- delay .tv_sec = 0 ;
74- delay .tv_usec = DELAY_MSEC * 1000 ;
95+ delay .tv_sec = cur_delay / 100 ;
96+ delay .tv_usec = ( cur_delay % 100 ) * 10000 ;
7597(void )select (0 ,NULL ,NULL ,NULL ,& delay );
7698
99+ #if defined(S_LOCK_TEST )
100+ fprintf (stdout ,"*" );fflush (stdout );
101+ #endif
102+
103+ /* increase delay by a random fraction between 1X and 2X */
104+ cur_delay += (int ) (cur_delay *
105+ (((double )random ()) / ((double )MAX_RANDOM_VALUE ))+ 0.5 );
106+ /* wrap back to minimum delay when max is exceeded */
107+ if (cur_delay > MAX_DELAY_CSEC )
108+ cur_delay = MIN_DELAY_CSEC ;
109+
77110spins = 0 ;
78111}
79112}
@@ -217,6 +250,8 @@ volatile slock_t test_lock;
217250int
218251main ()
219252{
253+ srandom ((unsignedint )time (NULL ));
254+
220255S_INIT_LOCK (& test_lock );
221256
222257if (!S_LOCK_FREE (& test_lock ))
@@ -249,9 +284,9 @@ main()
249284return 1 ;
250285}
251286
252- printf ("S_LOCK_TEST: this willhang for a minute or so and then abort \n" );
253- printf (" with a 'stuck spinlock' message if S_LOCK() \n" );
254- printf (" and TAS() are working.\n" );
287+ printf ("S_LOCK_TEST: this willprint %d stars and then\n" , NUM_DELAYS );
288+ printf ("exit with a 'stuck spinlock' message\n" );
289+ printf ("if S_LOCK() and TAS() are working.\n" );
255290fflush (stdout );
256291
257292s_lock (& test_lock ,__FILE__ ,__LINE__ );