postgres.git / summary / log / commit / refs

commit    59d1e2b95a826869e2789ffe01e9e552148eefde
Author:   Robert Haas <rhaas@postgresql.org>
Date:     Thu Dec 21 14:09:04 2017 +0000
First released in: 11.0, 10.2

    Cancel CV sleep during subtransaction abort.
    
    Generally, error recovery paths that need to do things like
    LWLockReleaseAll and pgstat_report_wait_end also need to call
    ConditionVariableCancelSleep, but AbortSubTransaction was missed.
    
    Since subtransaction abort might destroy up the DSM segment that
    contains the ConditionVariable stored in cv_sleep_target, this
    can result in a crash for anything using condition variables.
    
    Reported and diagnosed by Andres Freund.
    
    Discussion: http://postgr.es/m/20171221110048.rxk6464azzl5t2fi@alap3.anarazel.de


 src/backend/access/transam/xact.c | 3 +++
 1 file changed, 3 insertions(+)

diff --git a/src/backend/access/transam/xact.c b/src/backend/access/transam/xact.c
index e93d740b210..d430e662e67 100644
--- a/src/backend/access/transam/xact.c
+++ b/src/backend/access/transam/xact.c
@@ -4717,6 +4717,9 @@ AbortSubTransaction(void)
 	/* Reset WAL record construction state */
 	XLogResetInsertion();
 
+	/* Cancel condition variable sleep */
+	ConditionVariableCancelSleep();
+
 	/*
 	 * Also clean up any open wait for lock, since the lock manager will choke
 	 * if we try to wait for another lock before doing this.


[parent: 1804284042e6]