ZTS: Fix zpool_reopen_001_pos (DEBUG)

Update the vdev_disk_open() retry logic to use a specified number
of milliseconds to be more robust.  Additionally, log both the
time waited and requested timeout to the internal log for debugging.

Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov>

module/os/linux/zfs/vdev_disk.c

@@ -39,6 +39,13 @@
 
 static void *zfs_vdev_holder = VDEV_HOLDER;
 
+/*
+ * Wait up to zfs_vdev_open_timeout_ms milliseconds before determining the
+ * device is missing. The missing path may be transient since the links
+ * can be briefly removed and recreated in response to udev events.
+ */
+static unsigned zfs_vdev_open_timeout_ms = 500;
+
 /* size of the "reserved" partition, in blocks */
 #define	EFI_MIN_RESV_SIZE	(16 * 1024)
 
@@ -146,8 +153,7 @@ vdev_disk_open(vdev_t *v, uint64_t *psize, uint64_t *max_psize,
 {
 	struct block_device *bdev;
 	fmode_t mode = vdev_bdev_mode(spa_mode(v->vdev_spa));
-	int count = 0, block_size;
-	int bdev_retry_count = 50;
+	hrtime_t timeout = MSEC2NSEC(zfs_vdev_open_timeout_ms);
 	vdev_disk_t *vd;
 
 	/* Must have a pathname and it must be absolute. */
@@ -162,7 +168,7 @@ vdev_disk_open(vdev_t *v, uint64_t *psize, uint64_t *max_psize,
 	 * partition force re-scanning the partition table while closed
 	 * in order to get an accurate updated block device size.  Then
 	 * since udev may need to recreate the device links increase the
-	 * open retry count before reporting the device as unavailable.
+	 * open retry timeout before reporting the device as unavailable.
 	 */
 	vd = v->vdev_tsd;
 	if (vd) {
@@ -188,8 +194,10 @@ vdev_disk_open(vdev_t *v, uint64_t *psize, uint64_t *max_psize,
 			if (!IS_ERR(bdev)) {
 				int error = vdev_bdev_reread_part(bdev);
 				blkdev_put(bdev, mode | FMODE_EXCL);
-				if (error == 0)
-					bdev_retry_count = 100;
+				if (error == 0) {
+					timeout = MSEC2NSEC(
+					    zfs_vdev_open_timeout_ms * 2);
+				}
 			}
 		}
 	} else {
@@ -222,21 +230,23 @@ vdev_disk_open(vdev_t *v, uint64_t *psize, uint64_t *max_psize,
 	 * and it is reasonable to sleep and retry before giving up.  In
 	 * practice delays have been observed to be on the order of 100ms.
 	 */
+	hrtime_t start = gethrtime();
 	bdev = ERR_PTR(-ENXIO);
-	while (IS_ERR(bdev) && count < bdev_retry_count) {
+	while (IS_ERR(bdev) && ((gethrtime() - start) < timeout)) {
 		bdev = blkdev_get_by_path(v->vdev_path, mode | FMODE_EXCL,
 		    zfs_vdev_holder);
 		if (unlikely(PTR_ERR(bdev) == -ENOENT)) {
 			schedule_timeout(MSEC_TO_TICK(10));
-			count++;
 		} else if (IS_ERR(bdev)) {
 			break;
 		}
 	}
 
 	if (IS_ERR(bdev)) {
 		int error = -PTR_ERR(bdev);
-		vdev_dbgmsg(v, "open error=%d count=%d", error, count);
+		vdev_dbgmsg(v, "open error=%d timeout=%llu/%llu", error,
+		    (u_longlong_t)(gethrtime() - start),
+		    (u_longlong_t)timeout);
 		vd->vd_bdev = NULL;
 		v->vdev_tsd = vd;
 		rw_exit(&vd->vd_lock);
@@ -247,10 +257,15 @@ vdev_disk_open(vdev_t *v, uint64_t *psize, uint64_t *max_psize,
 		rw_exit(&vd->vd_lock);
 	}
 
+	/* DEBUG: To be removed in the final version. */
+	vdev_dbgmsg(v, "open timeout=%llu/%llu",
+	    (u_longlong_t)(gethrtime() - start),
+	    (u_longlong_t)timeout);
+
 	struct request_queue *q = bdev_get_queue(vd->vd_bdev);
 
 	/*  Determine the physical block size */
-	block_size = bdev_physical_block_size(vd->vd_bdev);
+	int block_size = bdev_physical_block_size(vd->vd_bdev);
 
 	/* Clear the nowritecache bit, causes vdev_reopen() to try again. */
 	v->vdev_nowritecache = B_FALSE;