All of lore.kernel.org
 help / color / mirror / Atom feed
From: mwilck@suse.com
To: Christophe Varoqui <christophe.varoqui@opensvc.com>,
	Benjamin Marzinski <bmarzins@redhat.com>
Cc: dm-devel@redhat.com, Martin Wilck <mwilck@suse.com>
Subject: [PATCH 06/19] multipathd: generalize and fix wait_for_state_change_if()
Date: Wed, 16 Sep 2020 17:37:05 +0200	[thread overview]
Message-ID: <20200916153718.582-7-mwilck@suse.com> (raw)
In-Reply-To: <20200916153718.582-1-mwilck@suse.com>

From: Martin Wilck <mwilck@suse.com>

It's unlikely but not impossible that other threads change the state
while we're waiting, and if we grab the lock again, it's still not
what we wanted. We need to continue waiting until either the condition
is met, or time timeout expired.

Moreover, generalize this code so that it can also be used in
set_config_state().

Signed-off-by: Martin Wilck <mwilck@suse.com>
---
 multipathd/main.c | 25 ++++++++++++++++++-------
 1 file changed, 18 insertions(+), 7 deletions(-)

diff --git a/multipathd/main.c b/multipathd/main.c
index 451d68c..49809e0 100644
--- a/multipathd/main.c
+++ b/multipathd/main.c
@@ -223,6 +223,23 @@ static void config_cleanup(__attribute__((unused)) void *arg)
 	pthread_mutex_unlock(&config_lock);
 }
 
+#define __wait_for_state_change(condition, ms)				\
+	({								\
+		struct timespec tmo;					\
+		int rc = 0;						\
+									\
+		if (condition) {					\
+			get_monotonic_time(&tmo);			\
+			tmo.tv_nsec += (ms) * 1000 * 1000;		\
+			normalize_timespec(&tmo);			\
+			do						\
+				rc = pthread_cond_timedwait(		\
+					&config_cond, &config_lock, &tmo); \
+			while (rc == 0 && (condition));			\
+		}							\
+		rc;							\
+	})
+
 /*
  * If the current status is @oldstate, wait for at most @ms milliseconds
  * for the state to change, and return the new state, which may still be
@@ -232,20 +249,14 @@ enum daemon_status wait_for_state_change_if(enum daemon_status oldstate,
 					    unsigned long ms)
 {
 	enum daemon_status st;
-	struct timespec tmo;
 
 	if (oldstate == DAEMON_SHUTDOWN)
 		return DAEMON_SHUTDOWN;
 
 	pthread_mutex_lock(&config_lock);
 	pthread_cleanup_push(config_cleanup, NULL);
+	__wait_for_state_change(running_state == oldstate, ms);
 	st = running_state;
-	if (st == oldstate && clock_gettime(CLOCK_MONOTONIC, &tmo) == 0) {
-		tmo.tv_nsec += ms * 1000 * 1000;
-		normalize_timespec(&tmo);
-		(void)pthread_cond_timedwait(&config_cond, &config_lock, &tmo);
-		st = running_state;
-	}
 	pthread_cleanup_pop(1);
 	return st;
 }
-- 
2.28.0

  parent reply	other threads:[~2020-09-16 15:37 UTC|newest]

Thread overview: 31+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2020-09-16 15:36 [PATCH 00/19] multipath-tools: shutdown, libdevmapper races, globals mwilck
2020-09-16 15:37 ` [PATCH 01/19] multipathd: allow shutdown during configure() mwilck
2020-09-16 15:37 ` [PATCH 02/19] multipathd: avoid sending "READY=1" to systemd on early shutdown mwilck
2020-09-16 15:37 ` [PATCH 03/19] multipathd: send "STOPPING=1" to systemd on shutdown mwilck
2020-09-16 15:37 ` [PATCH 04/19] multipathd: send "RELOADING=1" to systemd on DAEMON_CONFIGURE state mwilck
2020-09-16 15:37 ` [PATCH 05/19] multipathd: use volatile qualifier for running_state mwilck
2020-09-16 15:37 ` mwilck [this message]
2020-09-16 15:37 ` [PATCH 07/19] multipathd: set_config_state(): avoid code duplication mwilck
2020-09-19 19:01   ` Benjamin Marzinski
2020-09-16 15:37 ` [PATCH 08/19] multipathd: cancel threads early during shutdown mwilck
2020-09-16 15:37 ` [PATCH 09/19] multipath-tools: don't call dm_lib_release() any more mwilck
2020-09-16 15:37 ` [PATCH 10/19] libmultipath: devmapper: refactor libdm version determination mwilck
2020-09-19 22:14   ` Benjamin Marzinski
2020-09-21  8:35     ` Martin Wilck
2020-09-16 15:37 ` [PATCH 11/19] libmultipath: protect racy libdevmapper calls with a mutex mwilck
2020-09-16 15:37 ` [PATCH 12/19] libmultipath: constify file argument in config parser mwilck
2020-09-16 15:37 ` [PATCH 13/19] libmultipath: provide defaults for {get, put}_multipath_config mwilck
2020-09-21 19:08   ` Benjamin Marzinski
2020-09-21 19:42     ` Martin Wilck
2020-09-16 15:37 ` [PATCH 14/19] libmpathpersist: allow using libmultipath " mwilck
2020-09-16 15:37 ` [PATCH 15/19] multipath: use {get_put}_multipath_config from libmultipath mwilck
2020-09-16 15:37 ` [PATCH 16/19] mpathpersist: use {get, put}_multipath_config() " mwilck
2020-09-16 15:37 ` [PATCH 17/19] libmultipath: add udev and logsink symbols mwilck
2020-09-21 20:10   ` Benjamin Marzinski
2020-09-23  8:16     ` Martin Wilck
2020-09-23 16:05       ` Benjamin Marzinski
2020-09-16 15:37 ` [PATCH 18/19] multipath: remove logsink and udev mwilck
2020-09-16 15:37 ` [PATCH 19/19] mpathpersist: " mwilck
2020-09-21 20:15   ` Benjamin Marzinski
2020-09-22 11:32     ` Martin Wilck
2020-09-21 20:20 ` [PATCH 00/19] multipath-tools: shutdown, libdevmapper races, globals Benjamin Marzinski

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=20200916153718.582-7-mwilck@suse.com \
    --to=mwilck@suse.com \
    --cc=bmarzins@redhat.com \
    --cc=christophe.varoqui@opensvc.com \
    --cc=dm-devel@redhat.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.