DPDK patches and discussions
 help / color / mirror / Atom feed
* [dpdk-dev] [PATCH] eal: close MP socket during cleanup
@ 2018-12-19  7:41 Qi Zhang
  2018-12-19  8:46 ` Thomas Monjalon
  2018-12-20 12:51 ` [dpdk-dev] [PATCH v2] " Qi Zhang
  0 siblings, 2 replies; 5+ messages in thread
From: Qi Zhang @ 2018-12-19  7:41 UTC (permalink / raw)
  To: anatoly.burakov; +Cc: thomas, dev, ferruh.yigit, Qi Zhang, stable

When secondary process quit, the mp_socket* file still exist, that
cause rte_mp_request_sync fail when try to send message on a floating
socket.

The patch fix the issue by introduce a function rte_mp_channel_fini.
This function will be called by rte_eal_cleanup and it will close the
mp socket and delete the mp_socket* file.

Fixes: bacaa2754017 ("eal: add channel for multi-process communication")
Cc: stable@dpdk.org

Signed-off-by: Qi Zhang <qi.z.zhang@intel.com>
---
 lib/librte_eal/common/eal_common_proc.c | 22 +++++++++++++++++++++-
 lib/librte_eal/common/eal_private.h     |  6 ++++++
 lib/librte_eal/linuxapp/eal/eal.c       |  1 +
 3 files changed, 28 insertions(+), 1 deletion(-)

diff --git a/lib/librte_eal/common/eal_common_proc.c b/lib/librte_eal/common/eal_common_proc.c
index 6b876590a..ff19bea15 100644
--- a/lib/librte_eal/common/eal_common_proc.c
+++ b/lib/librte_eal/common/eal_common_proc.c
@@ -37,6 +37,7 @@ static int mp_fd = -1;
 static char mp_filter[PATH_MAX];   /* Filter for secondary process sockets */
 static char mp_dir_path[PATH_MAX]; /* The directory path for all mp sockets */
 static pthread_mutex_t mp_mutex_action = PTHREAD_MUTEX_INITIALIZER;
+static char peer_name[PATH_MAX];
 
 struct action_entry {
 	TAILQ_ENTRY(action_entry) next;
@@ -511,9 +512,9 @@ async_reply_handle(void *arg)
 static int
 open_socket_fd(void)
 {
-	char peer_name[PATH_MAX] = {0};
 	struct sockaddr_un un;
 
+	peer_name[0] = '\0';
 	if (rte_eal_process_type() == RTE_PROC_SECONDARY)
 		snprintf(peer_name, sizeof(peer_name),
 				"%d_%"PRIx64, getpid(), rte_rdtsc());
@@ -542,6 +543,19 @@ open_socket_fd(void)
 	return mp_fd;
 }
 
+static void
+close_socket_fd(void)
+{
+	char path[PATH_MAX];
+
+	if (mp_fd < 0)
+		return;
+
+	close(mp_fd);
+	create_socket_path(peer_name, path, sizeof(path));
+	unlink(path);
+}
+
 int
 rte_mp_channel_init(void)
 {
@@ -602,6 +616,12 @@ rte_mp_channel_init(void)
 	return 0;
 }
 
+void
+rte_mp_channel_fini(void)
+{
+	close_socket_fd();
+}
+
 /**
  * Return -1, as fail to send message and it's caused by the local side.
  * Return 0, as fail to send message and it's caused by the remote side.
diff --git a/lib/librte_eal/common/eal_private.h b/lib/librte_eal/common/eal_private.h
index 442c6dc48..10b6857f6 100644
--- a/lib/librte_eal/common/eal_private.h
+++ b/lib/librte_eal/common/eal_private.h
@@ -259,6 +259,12 @@ struct rte_bus *rte_bus_find_by_device_name(const char *str);
 int rte_mp_channel_init(void);
 
 /**
+ * Primary/secondary communication cleanup.
+ */
+
+void rte_mp_channel_fini(void);
+
+/**
  * @internal
  * Parse a device string and store its information in an
  * rte_devargs structure.
diff --git a/lib/librte_eal/linuxapp/eal/eal.c b/lib/librte_eal/linuxapp/eal/eal.c
index d252c8591..52df05e8d 100644
--- a/lib/librte_eal/linuxapp/eal/eal.c
+++ b/lib/librte_eal/linuxapp/eal/eal.c
@@ -1229,6 +1229,7 @@ rte_eal_cleanup(void)
 	if (rte_eal_process_type() == RTE_PROC_PRIMARY)
 		rte_memseg_walk(mark_freeable, NULL);
 	rte_service_finalize();
+	rte_mp_channel_fini();
 	return 0;
 }
 
-- 
2.13.6

^ permalink raw reply	[flat|nested] 5+ messages in thread

end of thread, other threads:[~2018-12-21  0:18 UTC | newest]

Thread overview: 5+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2018-12-19  7:41 [dpdk-dev] [PATCH] eal: close MP socket during cleanup Qi Zhang
2018-12-19  8:46 ` Thomas Monjalon
2018-12-20 12:30   ` Zhang, Qi Z
2018-12-20 12:51 ` [dpdk-dev] [PATCH v2] " Qi Zhang
2018-12-21  0:18   ` Thomas Monjalon

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).