All of lore.kernel.org
 help / color / mirror / Atom feed
From: tang.junhui@zte.com.cn
To: christophe.varoqui@opensvc.com, hare@suse.de,
	bmarzins@redhat.com, mwilck@suse.com, bart.vanassche@sandisk.com
Cc: zhang.kai16@zte.com.cn, dm-devel@redhat.com,
	tang.junhui@zte.com.cn, tang.wenjun3@zte.com.cn
Subject: [PATCH 08/12] libmultipath: wait one seconds for more uevents in uevent_listen() in uevents burst situations
Date: Tue, 27 Dec 2016 16:03:25 +0800	[thread overview]
Message-ID: <1482825809-9528-9-git-send-email-tang.junhui@zte.com.cn> (raw)
In-Reply-To: <1482825809-9528-1-git-send-email-tang.junhui@zte.com.cn>

From: tang.junhui <tang.junhui@zte.com.cn>

The more uevents are merged, the higher efficiency program will performs.
So, do not process uevents after receiving immediately in uevents burst
situations, but continue wait 1 seconds for more uevents except that too
much uevents (2048) have already been received or too much time eclipse
(60 seconds).

Change-Id: I763d491540e8114a81d12d603281540a81502742
Signed-off-by: tang.junhui <tang.junhui@zte.com.cn>
---
 libmultipath/uevent.c | 35 +++++++++++++++++++++++++++++++++--
 1 file changed, 33 insertions(+), 2 deletions(-)

diff --git a/libmultipath/uevent.c b/libmultipath/uevent.c
index cc10d65..b0b05e9 100644
--- a/libmultipath/uevent.c
+++ b/libmultipath/uevent.c
@@ -39,6 +39,7 @@
 #include <linux/netlink.h>
 #include <pthread.h>
 #include <sys/mman.h>
+#include <sys/time.h>
 #include <libudev.h>
 #include <errno.h>
 
@@ -490,11 +491,37 @@ struct uevent *uevent_from_udev_device(struct udev_device *dev)
 	return uev;
 }
 
+bool uevent_burst(struct timeval *start_time, int events)
+{
+	struct timeval diff_time, end_time;
+	unsigned long speed;
+	unsigned long eclipse_ms;
+
+	gettimeofday(&end_time, NULL);
+	timersub(&end_time, start_time, &diff_time);
+
+	eclipse_ms = diff_time.tv_sec * 1000 + diff_time.tv_usec / 1000;
+	if (eclipse_ms == 0)
+		return true;
+	/* max wait 60s */
+	if (eclipse_ms > 60*1000) {
+		condlog(1, "burst continued =%lu ms, stoped", eclipse_ms);
+		return false;
+	}
+
+	speed = (events * 1000) / eclipse_ms;
+	if (speed > 10)
+		return true;
+
+	return false;
+}
+
 int uevent_listen(struct udev *udev)
 {
 	int err = 2;
 	struct udev_monitor *monitor = NULL;
 	int fd, socket_flags, events;
+	struct timeval start_time;
 	int need_failback = 1;
 	int timeout = 30;
 	LIST_HEAD(uevlisten_tmp);
@@ -548,6 +575,7 @@ int uevent_listen(struct udev *udev)
 	}
 
 	events = 0;
+	gettimeofday(&start_time, NULL);
 	while (1) {
 		struct uevent *uev;
 		struct udev_device *dev;
@@ -562,7 +590,8 @@ int uevent_listen(struct udev *udev)
 		errno = 0;
 		fdcount = poll(&ev_poll, 1, poll_timeout);
 		if (fdcount && ev_poll.revents & POLLIN) {
-			timeout = 0;
+			timeout = uevent_burst(&start_time, events + 1) ? 1:0;
+
 			dev = udev_monitor_receive_device(monitor);
 			if (!dev) {
 				condlog(0, "failed getting udev device");
@@ -578,7 +607,8 @@ int uevent_listen(struct udev *udev)
 			}
 			list_add_tail(&uev->node, &uevlisten_tmp);
 			events++;
-			continue;
+			if(events < 2048)
+				continue;
 		}
 		if (fdcount < 0) {
 			if (errno == EINTR)
@@ -600,6 +630,7 @@ int uevent_listen(struct udev *udev)
 			pthread_mutex_unlock(uevq_lockp);
 			events = 0;
 		}
+		gettimeofday(&start_time, NULL);
 		timeout = 30;
 	}
 	need_failback = 0;
-- 
2.8.1.windows.1

  parent reply	other threads:[~2016-12-27  8:03 UTC|newest]

Thread overview: 34+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2016-12-27  8:03 [PATCH 00/12] multipath-tools: improve processing efficiency for addition and deletion of multipath devices tang.junhui
2016-12-27  8:03 ` [PATCH 01/12] libmultipath: add wwid for "struct uevent" to record wwid of uevent tang.junhui
2017-01-03 22:02   ` Benjamin Marzinski
2017-01-04  6:56     ` tang.junhui
2017-01-04 18:14       ` Benjamin Marzinski
2017-01-04 20:33         ` Martin Wilck
2017-01-05  3:00           ` Benjamin Marzinski
2017-01-05  3:10         ` tang.junhui
2017-01-05 17:36           ` Benjamin Marzinski
2017-01-06  0:59             ` tang.junhui
2017-01-06 16:02               ` Benjamin Marzinski
2017-01-09  7:22                 ` tang.junhui
2016-12-27  8:03 ` [PATCH 02/12] libmultipath: add merge_node for "struct uevent" to record nodes of merged uevents tang.junhui
2016-12-27  8:03 ` [PATCH 03/12] libmultipath: add two list iteration macros tang.junhui
2016-12-27  8:03 ` [PATCH 04/12] multipathd: add need_do_map to indicate whether need calling domap() in ev_add_path() tang.junhui
2016-12-27  8:03 ` [PATCH 05/12] multipathd: add need_do_map to indicate whether need calling domap() in ev_remove_path() tang.junhui
2016-12-27  8:03 ` [PATCH 06/12] multipathd: move uev_discard() to uevent.c and change its name to uevent_can_discard() tang.junhui
2016-12-27  8:03 ` [PATCH 07/12] multipathd: move calling filter_devnode() from uev_trigger() " tang.junhui
2016-12-27  8:03 ` tang.junhui [this message]
2016-12-28 20:25   ` [PATCH 08/12] libmultipath: wait one seconds for more uevents in uevent_listen() in uevents burst situations Martin Wilck
2016-12-29  0:48     ` tang.junhui
2017-01-03 22:31   ` Benjamin Marzinski
2017-01-04  7:32     ` tang.junhui
2016-12-27  8:03 ` [PATCH 09/12] multipathd: merge uevents before proccessing tang.junhui
2017-01-04  0:30   ` Benjamin Marzinski
2017-01-04  3:29     ` tang.junhui
2016-12-27  8:03 ` [PATCH 10/12] libmultipath: filter " tang.junhui
2017-01-04  1:21   ` Benjamin Marzinski
2017-01-04  2:03     ` tang.junhui
2016-12-27  8:03 ` [PATCH 11/12] multipathd: proccess merged uevents tang.junhui
2017-01-04  1:03   ` Benjamin Marzinski
2017-01-04  1:54     ` tang.junhui
2016-12-27  8:03 ` [PATCH 12/12] libmultipath: use existing wwid when wwid has already been existed in uevent tang.junhui
2016-12-29  1:54 [PATCH 08/12] libmultipath: wait one seconds for more uevents in uevent_listen() in uevents burst situations tang.junhui

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=1482825809-9528-9-git-send-email-tang.junhui@zte.com.cn \
    --to=tang.junhui@zte.com.cn \
    --cc=bart.vanassche@sandisk.com \
    --cc=bmarzins@redhat.com \
    --cc=christophe.varoqui@opensvc.com \
    --cc=dm-devel@redhat.com \
    --cc=hare@suse.de \
    --cc=mwilck@suse.com \
    --cc=tang.wenjun3@zte.com.cn \
    --cc=zhang.kai16@zte.com.cn \
    /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.