Import wiredtiger: e1bcc30da91eedd0b17cebb725cc7e607ffa2340 from branch mongodb-3.6

ref: 48a3cbc17f..e1bcc30da9
for: 3.5.4

WT-2790       Fix a text case false positive in test_sweep01
WT-2909       Create automatable test verifying checkpoint integrity after errors
WT-3088       bug: Don't evict a page with refs visible to readers after a split
WT-3097       Race on reconfigure or shutdown can lead to waiting for statistics log server
WT-3111       util_create() doesnt free memory assigned to "uri"
WT-3113       Add a verbose mode to dump the cache when eviction is stuck
WT-3115       Change the dhandle lock to a read/write lock
WT-3120       Fix ordering problem in connection_close for filesystem loaded in an extension
WT-3135       search_near() for index with custom collator
WT-3137       Hang in __log_slot_join/__log_slot_switch_internal
WT-3139       Enhance wtperf to support periodic table scans
WT-3144       bug fix: random cursor returns not-found when descending to an empty page
WT-3148       Improve eviction efficiency with many small trees
WT-3149       Change eviction to start new walks from a random place in the tree
WT-3150       Reduce impact of checkpoints on eviction server
WT-3152       Convert table lock from a spinlock to a read write lock
WT-3156       Assertion in log_write fires after write failure
WT-3157       checkpoint/transaction integrity issue when writes fail.
WT-3159       Incorrect key for index containing multiple variable sized entries
WT-3161       checkpoint hang after write failure injection.
WT-3164       Ensure all relevant btree fields are reset on checkpoint error
WT-3170       Clear the eviction walk point while populating from a tree
WT-3173       Add runtime detection for s390x CRC32 hardware support
WT-3174       Coverity/lint cleanup
WT-3175       New hang in internal page split
WT-3179       Test bug: clang sanitizer failure in fail_fs
WT-3180       Fault injection tests should only run as "long" tests and should not create core files
WT-3184       Problem duplicating index cursor with custom collator
WT-3186       Fix error path and panic detection in logging loops

1 files changed, 22 insertions, 6 deletions

diff --git a/src/third_party/wiredtiger/src/os_posix/os_mtx_cond.c b/src/third_party/wiredtiger/src/os_posix/os_mtx_cond.c
index be8b1abda31..a5ee78f9e3e 100644
--- a/src/third_party/wiredtiger/src/os_posix/os_mtx_cond.c
+++ b/src/third_party/wiredtiger/src/os_posix/os_mtx_cond.c
@@ -13,8 +13,7 @@
  *	Allocate and initialize a condition variable.
  */
 int
-__wt_cond_alloc(WT_SESSION_IMPL *session,
-    const char *name, bool is_signalled, WT_CONDVAR **condp)
+__wt_cond_alloc(WT_SESSION_IMPL *session, const char *name, WT_CONDVAR **condp)
 {
 	WT_CONDVAR *cond;
 	WT_DECL_RET;
@@ -27,7 +26,7 @@ __wt_cond_alloc(WT_SESSION_IMPL *session,
 	WT_ERR(pthread_cond_init(&cond->cond, NULL));
 
 	cond->name = name;
-	cond->waiters = is_signalled ? -1 : 0;
+	cond->waiters = 0;
 
 	*condp = cond;
 	return (0);
@@ -42,8 +41,8 @@ err:	__wt_free(session, cond);
  * out period expires, let the caller know.
  */
 void
-__wt_cond_wait_signal(
-    WT_SESSION_IMPL *session, WT_CONDVAR *cond, uint64_t usecs, bool *signalled)
+__wt_cond_wait_signal(WT_SESSION_IMPL *session, WT_CONDVAR *cond,
+    uint64_t usecs, bool (*run_func)(WT_SESSION_IMPL *), bool *signalled)
 {
 	struct timespec ts;
 	WT_DECL_RET;
@@ -62,6 +61,23 @@ __wt_cond_wait_signal(
 	WT_ERR(pthread_mutex_lock(&cond->mtx));
 	locked = true;
 
+	/*
+	 * It's possible to race with threads waking us up. That's not a problem
+	 * if there are multiple wakeups because the next wakeup will get us, or
+	 * if we're only pausing for a short period. It's a problem if there's
+	 * only a single wakeup, our waker is likely waiting for us to exit.
+	 * After acquiring the mutex (so we're guaranteed to be awakened by any
+	 * future wakeup call), optionally check if we're OK to keep running.
+	 * This won't ensure our caller won't just loop and call us again, but
+	 * at least it's not our fault.
+	 *
+	 * Assert we're not waiting longer than a second if not checking the
+	 * run status.
+	 */
+	WT_ASSERT(session, run_func != NULL || usecs <= WT_MILLION);
+	if (run_func != NULL && !run_func(session))
+		goto skipping;
+
 	if (usecs > 0) {
 		__wt_epoch(session, &ts);
 		ts.tv_sec += (time_t)
@@ -81,7 +97,7 @@ __wt_cond_wait_signal(
 	    ret == ETIME ||
 #endif
 	    ret == ETIMEDOUT) {
-		*signalled = false;
+skipping:	*signalled = false;
 		ret = 0;
 	}