From: Willy Tarreau <w@1wt.eu>
Date: Wed, 13 Jul 2022 06:59:39 +0000 (+0200)
Subject: BUG/MEDIUM: debug: fix possible hang when multiple threads dump at once
X-Git-Tag: v2.7-dev2~64
X-Git-Url: http://git.ipfire.org/cgi-bin/gitweb.cgi?a=commitdiff_plain;h=672972604;p=thirdparty%2Fhaproxy.git

BUG/MEDIUM: debug: fix possible hang when multiple threads dump at once

A bug in the thread dumper was introduced by commit 00c27b50c ("MEDIUM:
debug: make the thread dumper not rely on a thread mask anymore"). If
two or more threads try to trigger a thread dump exactly at the same
time, the second one may loop indefinitely trying to set the value to 1
while the other ones will wait for it to finish dumping before leaving.
This is a consequence of a logic change using thread numbers instead of
a thread mask, as threads do not need to see all other ones there anymore.

No backport is needed, this is only for 2.7.
---

diff --git a/src/debug.c b/src/debug.c
index 8d6210828f..b793e3ec03 100644
--- a/src/debug.c
+++ b/src/debug.c
@@ -1313,12 +1313,11 @@ void ha_thread_dump_all_to_trash()
 {
 	unsigned int old;
 
-	while (1) {
-		old = 0;
-		if (HA_ATOMIC_CAS(&thread_dump_state, &old, 1))
-			break;
-		ha_thread_relax();
-	}
+	/* initiate a dump starting from first thread. Use a CAS
+	 * so that we don't wait if we're not the first one.
+	 */
+	old = 0;
+	HA_ATOMIC_CAS(&thread_dump_state, &old, 1);
 
 	thread_dump_buffer = &trash;
 	thread_dump_tid = tid;