[lustre-devel] [PATCH 1/8] staging: lustre: lnet: simplify lnet_eq_wait_locked

James Simmons jsimmons at infradead.org
Sun Jun 24 14:53:45 PDT 2018


We can simplify the code by taking advantage of the behavior
of schedule_timeout_interruptible(). Instead of testing if
tms is less than zero we can pass in a signed long that
schedule_timeout_interruptible is expecting and for the case
of no timeout we can pass in MAX_SCHEDULE_TIMEOUT.

Signed-off-by: James Simmons <uja.ornl at yahoo.com>
WC-bug-id: https://jira.whamcloud.com/browse/LU-9019
Reviewed-on: https://review.whamcloud.com/23147
Reviewed-by: Doug Oucharek <dougso at me.com>
Reviewed-by: Dmitry Eremin <dmitry.eremin at intel.com>
Reviewed-by: Oleg Drokin <green at whamcloud.com>
Signed-off-by: James Simmons <jsimmons at infradead.org>
---
 drivers/staging/lustre/include/linux/lnet/api.h    |  2 +-
 .../lustre/include/uapi/linux/lnet/lnet-types.h    |  2 --
 drivers/staging/lustre/lnet/lnet/api-ni.c          | 38 ++++++++++++++--------
 drivers/staging/lustre/lnet/lnet/lib-eq.c          | 26 +++++----------
 4 files changed, 33 insertions(+), 35 deletions(-)

diff --git a/drivers/staging/lustre/include/linux/lnet/api.h b/drivers/staging/lustre/include/linux/lnet/api.h
index dae2e4f..2242921 100644
--- a/drivers/staging/lustre/include/linux/lnet/api.h
+++ b/drivers/staging/lustre/include/linux/lnet/api.h
@@ -168,7 +168,7 @@ int LNetEQAlloc(unsigned int       count_in,
 
 int LNetEQPoll(struct lnet_handle_eq *eventqs_in,
 	       int		 neq_in,
-	       int		 timeout_ms,
+	       signed long	 timeout,
 	       int		 interruptible,
 	       struct lnet_event *event_out,
 	       int		*which_eq_out);
diff --git a/drivers/staging/lustre/include/uapi/linux/lnet/lnet-types.h b/drivers/staging/lustre/include/uapi/linux/lnet/lnet-types.h
index 1be9b7a..14715a8 100644
--- a/drivers/staging/lustre/include/uapi/linux/lnet/lnet-types.h
+++ b/drivers/staging/lustre/include/uapi/linux/lnet/lnet-types.h
@@ -77,8 +77,6 @@
 #define LNET_PID_USERFLAG 0x80000000 /* set in userspace peers */
 #define LNET_PID_LUSTRE	  12345
 
-#define LNET_TIME_FOREVER (-1)
-
 /* how an LNET NID encodes net:address */
 /** extract the address part of an lnet_nid_t */
 
diff --git a/drivers/staging/lustre/lnet/lnet/api-ni.c b/drivers/staging/lustre/lnet/lnet/api-ni.c
index f9ed697..878c92c 100644
--- a/drivers/staging/lustre/lnet/lnet/api-ni.c
+++ b/drivers/staging/lustre/lnet/lnet/api-ni.c
@@ -59,7 +59,7 @@
 module_param(rnet_htable_size, int, 0444);
 MODULE_PARM_DESC(rnet_htable_size, "size of remote network hash table");
 
-static int lnet_ping(struct lnet_process_id id, int timeout_ms,
+static int lnet_ping(struct lnet_process_id id, signed long timeout,
 		     struct lnet_process_id __user *ids, int n_ids);
 
 static char *
@@ -2052,17 +2052,29 @@ void lnet_lib_exit(void)
 	case IOC_LIBCFS_LNET_FAULT:
 		return lnet_fault_ctl(data->ioc_flags, data);
 
-	case IOC_LIBCFS_PING:
+	case IOC_LIBCFS_PING: {
+		signed long timeout;
+
 		id.nid = data->ioc_nid;
 		id.pid = data->ioc_u32[0];
-		rc = lnet_ping(id, data->ioc_u32[1], /* timeout */
-			       data->ioc_pbuf1,
+
+		/* Don't block longer than 2 minutes */
+		if (data->ioc_u32[1] > 120 * MSEC_PER_SEC)
+			return -EINVAL;
+
+		/* If timestamp is negative then disable timeout */
+		if ((s32)data->ioc_u32[1] < 0)
+			timeout = MAX_SCHEDULE_TIMEOUT;
+		else
+			timeout = msecs_to_jiffies(data->ioc_u32[1]);
+
+		rc = lnet_ping(id, timeout, data->ioc_pbuf1,
 			       data->ioc_plen1 / sizeof(struct lnet_process_id));
 		if (rc < 0)
 			return rc;
 		data->ioc_count = rc;
 		return 0;
-
+	}
 	default:
 		ni = lnet_net2ni(data->ioc_net);
 		if (!ni)
@@ -2126,7 +2138,7 @@ void LNetDebugPeer(struct lnet_process_id id)
 }
 EXPORT_SYMBOL(LNetGetId);
 
-static int lnet_ping(struct lnet_process_id id, int timeout_ms,
+static int lnet_ping(struct lnet_process_id id, signed long timeout,
 		     struct lnet_process_id __user *ids, int n_ids)
 {
 	struct lnet_handle_eq eqh;
@@ -2136,7 +2148,7 @@ static int lnet_ping(struct lnet_process_id id, int timeout_ms,
 	int which;
 	int unlinked = 0;
 	int replied = 0;
-	const int a_long_time = 60000; /* mS */
+	const signed long a_long_time = msecs_to_jiffies(60 * MSEC_PER_SEC);
 	int infosz;
 	struct lnet_ping_info *info;
 	struct lnet_process_id tmpid;
@@ -2147,10 +2159,8 @@ static int lnet_ping(struct lnet_process_id id, int timeout_ms,
 
 	infosz = offsetof(struct lnet_ping_info, pi_ni[n_ids]);
 
-	if (n_ids <= 0 ||
-	    id.nid == LNET_NID_ANY ||
-	    timeout_ms > 500000 ||	      /* arbitrary limit! */
-	    n_ids > 20)			 /* arbitrary limit! */
+	/* n_ids limit is arbitrary */
+	if (n_ids <= 0 || n_ids > 20 || id.nid == LNET_NID_ANY)
 		return -EINVAL;
 
 	if (id.pid == LNET_PID_ANY)
@@ -2194,13 +2204,13 @@ static int lnet_ping(struct lnet_process_id id, int timeout_ms,
 
 		/* NB must wait for the UNLINK event below... */
 		unlinked = 1;
-		timeout_ms = a_long_time;
+		timeout = a_long_time;
 	}
 
 	do {
 		/* MUST block for unlink to complete */
 
-		rc2 = LNetEQPoll(&eqh, 1, timeout_ms, !unlinked,
+		rc2 = LNetEQPoll(&eqh, 1, timeout, !unlinked,
 				 &event, &which);
 
 		CDEBUG(D_NET, "poll %d(%d %d)%s\n", rc2,
@@ -2222,7 +2232,7 @@ static int lnet_ping(struct lnet_process_id id, int timeout_ms,
 				LNetMDUnlink(mdh);
 				/* No assertion (racing with network) */
 				unlinked = 1;
-				timeout_ms = a_long_time;
+				timeout = a_long_time;
 			} else if (!rc2) {
 				/* timed out waiting for unlink */
 				CWARN("ping %s: late network completion\n",
diff --git a/drivers/staging/lustre/lnet/lnet/lib-eq.c b/drivers/staging/lustre/lnet/lnet/lib-eq.c
index c78e703..366ad8a 100644
--- a/drivers/staging/lustre/lnet/lnet/lib-eq.c
+++ b/drivers/staging/lustre/lnet/lnet/lib-eq.c
@@ -308,13 +308,12 @@
  */
 
 static int
-lnet_eq_wait_locked(int *timeout_ms, long state)
+lnet_eq_wait_locked(signed long *timeout, long state)
 __must_hold(&the_lnet.ln_eq_wait_lock)
 {
-	int tms = *timeout_ms;
+	signed long tms = *timeout;
 	int wait;
 	wait_queue_entry_t wl;
-	unsigned long now;
 
 	if (!tms)
 		return -ENXIO; /* don't want to wait and no new event */
@@ -325,18 +324,9 @@
 
 	lnet_eq_wait_unlock();
 
-	if (tms < 0) {
-		schedule();
-	} else {
-		now = jiffies;
-		schedule_timeout(msecs_to_jiffies(tms));
-		tms -= jiffies_to_msecs(jiffies - now);
-		if (tms < 0) /* no more wait but may have new event */
-			tms = 0;
-	}
-
+	tms = schedule_timeout_interruptible(tms);
 	wait = tms; /* might need to call here again */
-	*timeout_ms = tms;
+	*timeout = tms;
 
 	lnet_eq_wait_lock();
 	remove_wait_queue(&the_lnet.ln_eq_waitq, &wl);
@@ -356,8 +346,8 @@
  * fixed period, or block indefinitely.
  *
  * \param eventqs,neq An array of EQ handles, and size of the array.
- * \param timeout_ms Time in milliseconds to wait for an event to occur on
- * one of the EQs. The constant LNET_TIME_FOREVER can be used to indicate an
+ * \param timeout Time in jiffies to wait for an event to occur on
+ * one of the EQs. The constant MAX_SCHEDULE_TIMEOUT can be used to indicate an
  * infinite timeout.
  * \param interruptible, if true, use TASK_INTERRUPTIBLE, else TASK_NOLOAD
  * \param event,which On successful return (1 or -EOVERFLOW), \a event will
@@ -372,7 +362,7 @@
  * \retval -ENOENT    If there's an invalid handle in \a eventqs.
  */
 int
-LNetEQPoll(struct lnet_handle_eq *eventqs, int neq, int timeout_ms,
+LNetEQPoll(struct lnet_handle_eq *eventqs, int neq, signed long timeout,
 	   int interruptible,
 	   struct lnet_event *event, int *which)
 {
@@ -414,7 +404,7 @@
 		 *  0 : don't want to wait anymore, but might have new event
 		 *      so need to call dequeue again
 		 */
-		wait = lnet_eq_wait_locked(&timeout_ms,
+		wait = lnet_eq_wait_locked(&timeout,
 					   interruptible ? TASK_INTERRUPTIBLE
 					   : TASK_NOLOAD);
 		if (wait < 0) /* no new event */
-- 
1.8.3.1



More information about the lustre-devel mailing list