From: Willy Tarreau <w@1wt.eu>
Date: Wed, 24 Feb 2021 10:13:59 +0000 (+0100)
Subject: BUG/MINOR: proxy: wake up all threads when sending the hard-stop signal
X-Git-Tag: v2.4-dev10~69
X-Git-Url: http://git.ipfire.org/cgi-bin/gitweb.cgi?a=commitdiff_plain;h=0d03825b93cc59a289e838105f9d83d53ccdfc8b;p=thirdparty%2Fhaproxy.git

BUG/MINOR: proxy: wake up all threads when sending the hard-stop signal

The hard-stop event didn't wake threads up. In the past it wasn't an issue
as the poll timeout was limited to 1 second, but since commit 4f59d3861
("MINOR: time: increase the minimum wakeup interval to 60s") it has become
a problem because old processes can remain live for up to one minute after
the hard-stop-after delay. Let's just wake them up.

This may be backported to older releases, though before 2.4 the extra
delay was only one second.
---

diff --git a/src/proxy.c b/src/proxy.c
index 3e3d7cc23e..ea7fc7f6f6 100644
--- a/src/proxy.c
+++ b/src/proxy.c
@@ -1612,11 +1612,15 @@ struct task *hard_stop(struct task *t, void *context, unsigned short state)
 {
 	struct proxy *p;
 	struct stream *s;
+	int thr;
 
 	if (killed) {
 		ha_warning("Some tasks resisted to hard-stop, exiting now.\n");
 		send_log(NULL, LOG_WARNING, "Some tasks resisted to hard-stop, exiting now.\n");
 		killed = 2;
+		for (thr = 0; thr < global.nbthread; thr++)
+			if (((all_threads_mask & ~tid_bit) >> thr) & 1)
+				wake_thread(thr);
 		t->expire = TICK_ETERNITY;
 		return t;
 	}