zpool reopen should detect expanded devices

Update bdev_capacity to have wholedisk vdevs query the
size of the underlying block device (correcting for the size
of the efi parition and partition alignment) and therefore detect
expanded space.

Correct vdev_get_stats_ex so that the expandsize is aligned
to metaslab size and new space is only reported if it is large
enough for a new metaslab.

Reviewed by: Don Brady <[email protected]>
Reviewed-by: Brian Behlendorf <[email protected]>
Reviewed by: George Wilson <[email protected]>
Reviewed-by: Matthew Ahrens <[email protected]>
Reviewed by: John Wren Kennedy <[email protected]>
Signed-off-by: sara hartse <[email protected]>
External-issue: LX-165
Closes #7546 
Issue #7582 

6 files changed, 107 insertions, 42 deletions

diff --git a/include/sys/vdev_disk.h b/include/sys/vdev_disk.h
index 15570b105..b8a32b316 100644
--- a/include/sys/vdev_disk.h
+++ b/include/sys/vdev_disk.h
@@ -23,11 +23,23 @@
  * Produced at Lawrence Livermore National Laboratory (cf, DISCLAIMER).
  * Written by Brian Behlendorf <[email protected]>.
  * LLNL-CODE-403049.
+ * Copyright (c) 2018 by Delphix. All rights reserved.
  */
 
 #ifndef _SYS_VDEV_DISK_H
 #define	_SYS_VDEV_DISK_H
 
+/*
+ * Don't start the slice at the default block of 34; many storage
+ * devices will use a stripe width of 128k, other vendors prefer a 1m
+ * alignment.  It is best to play it safe and ensure a 1m alignment
+ * given 512B blocks.  When the block size is larger by a power of 2
+ * we will still be 1m aligned.  Some devices are sensitive to the
+ * partition ending alignment as well.
+ */
+#define	NEW_START_BLOCK		2048
+#define	PARTITION_END_ALIGNMENT	2048
+
 #ifdef _KERNEL
 #include <sys/vdev.h>
 
diff --git a/lib/libefi/rdwr_efi.c b/lib/libefi/rdwr_efi.c
index 455fe527d..4b78cc58e 100644
--- a/lib/libefi/rdwr_efi.c
+++ b/lib/libefi/rdwr_efi.c
@@ -22,6 +22,7 @@
 /*
  * Copyright (c) 2002, 2010, Oracle and/or its affiliates. All rights reserved.
  * Copyright 2012 Nexenta Systems, Inc.  All rights reserved.
+ * Copyright (c) 2018 by Delphix. All rights reserved.
  */
 
 #include <stdio.h>
@@ -1154,7 +1155,7 @@ efi_use_whole_disk(int fd)
 
 	/*
 	 * Find the last physically non-zero partition.
-	 * This is the reserved partition.
+	 * This should be the reserved partition.
 	 */
 	for (i = 0; i < efi_label->efi_nparts; i ++) {
 		if (resv_start < efi_label->efi_parts[i].p_start) {
@@ -1164,6 +1165,23 @@ efi_use_whole_disk(int fd)
 	}
 
 	/*
+	 * Verify that we've found the reserved partition by checking
+	 * that it looks the way it did when we created it in zpool_label_disk.
+	 * If we've found the incorrect partition, then we know that this
+	 * device was reformatted and no longer is soley used by ZFS.
+	 */
+	if ((efi_label->efi_parts[resv_index].p_size != EFI_MIN_RESV_SIZE) ||
+	    (efi_label->efi_parts[resv_index].p_tag != V_RESERVED) ||
+	    (resv_index != 8)) {
+		if (efi_debug) {
+			(void) fprintf(stderr,
+			    "efi_use_whole_disk: wholedisk not available\n");
+		}
+		efi_free(efi_label);
+		return (VT_ENOSPC);
+	}
+
+	/*
 	 * Find the last physically non-zero partition before that.
 	 * This is the data partition.
 	 */
diff --git a/lib/libzfs/libzfs_pool.c b/lib/libzfs/libzfs_pool.c
index d082a5f66..935b42993 100644
--- a/lib/libzfs/libzfs_pool.c
+++ b/lib/libzfs/libzfs_pool.c
@@ -22,7 +22,7 @@
 /*
  * Copyright 2015 Nexenta Systems, Inc.  All rights reserved.
  * Copyright (c) 2005, 2010, Oracle and/or its affiliates. All rights reserved.
- * Copyright (c) 2011, 2017 by Delphix. All rights reserved.
+ * Copyright (c) 2011, 2018 by Delphix. All rights reserved.
  * Copyright 2016 Igor Kozhukhov <[email protected]>
  * Copyright (c) 2017 Datto Inc.
  * Copyright (c) 2017 Open-E, Inc. All Rights Reserved.
@@ -44,6 +44,7 @@
 #include <sys/systeminfo.h>
 #include <sys/vtoc.h>
 #include <sys/zfs_ioctl.h>
+#include <sys/vdev_disk.h>
 #include <dlfcn.h>
 
 #include "zfs_namecheck.h"
@@ -926,17 +927,6 @@ zpool_prop_get_feature(zpool_handle_t *zhp, const char *propname, char *buf,
 }
 
 /*
- * Don't start the slice at the default block of 34; many storage
- * devices will use a stripe width of 128k, other vendors prefer a 1m
- * alignment.  It is best to play it safe and ensure a 1m alignment
- * given 512B blocks.  When the block size is larger by a power of 2
- * we will still be 1m aligned.  Some devices are sensitive to the
- * partition ending alignment as well.
- */
-#define	NEW_START_BLOCK		2048
-#define	PARTITION_END_ALIGNMENT	2048
-
-/*
  * Validate the given pool name, optionally putting an extended error message in
  * 'buf'.
  */
diff --git a/module/zfs/vdev.c b/module/zfs/vdev.c
index 388be3617..5b67e5f5f 100644
--- a/module/zfs/vdev.c
+++ b/module/zfs/vdev.c
@@ -21,7 +21,7 @@
 
 /*
  * Copyright (c) 2005, 2010, Oracle and/or its affiliates. All rights reserved.
- * Copyright (c) 2011, 2015 by Delphix. All rights reserved.
+ * Copyright (c) 2011, 2018 by Delphix. All rights reserved.
  * Copyright 2017 Nexenta Systems, Inc.
  * Copyright (c) 2014 Integros [integros.com]
  * Copyright 2016 Toomas Soome <[email protected]>
@@ -3493,7 +3493,6 @@ vdev_get_stats_ex(vdev_t *vd, vdev_stat_t *vs, vdev_stat_ex_t *vsx)
 			    vd->vdev_max_asize - vd->vdev_asize,
 			    1ULL << tvd->vdev_ms_shift);
 		}
-		vs->vs_esize = vd->vdev_max_asize - vd->vdev_asize;
 		if (vd->vdev_aux == NULL && vd == vd->vdev_top &&
 		    vdev_is_concrete(vd)) {
 			vs->vs_fragmentation = vd->vdev_mg->mg_fragmentation;
diff --git a/module/zfs/vdev_disk.c b/module/zfs/vdev_disk.c
index 5cdfd960c..11744c1c1 100644
--- a/module/zfs/vdev_disk.c
+++ b/module/zfs/vdev_disk.c
@@ -23,7 +23,7 @@
  * Produced at Lawrence Livermore National Laboratory (cf, DISCLAIMER).
  * Rewritten for Linux by Brian Behlendorf <[email protected]>.
  * LLNL-CODE-403049.
- * Copyright (c) 2012, 2016 by Delphix. All rights reserved.
+ * Copyright (c) 2012, 2018 by Delphix. All rights reserved.
  */
 
 #include <sys/zfs_context.h>
@@ -34,10 +34,14 @@
 #include <sys/fs/zfs.h>
 #include <sys/zio.h>
 #include <linux/mod_compat.h>
+#include <linux/msdos_fs.h>
 
 char *zfs_vdev_scheduler = VDEV_SCHEDULER;
 static void *zfs_vdev_holder = VDEV_HOLDER;
 
+/* size of the "reserved" partition, in blocks */
+#define	EFI_MIN_RESV_SIZE	(16 * 1024)
+
 /*
  * Virtual device vector for disks.
  */
@@ -81,17 +85,39 @@ vdev_bdev_mode(int smode)
 }
 #endif /* HAVE_OPEN_BDEV_EXCLUSIVE */
 
+/* The capacity (in bytes) of a bdev that is available to be used by a vdev */
 static uint64_t
-bdev_capacity(struct block_device *bdev)
+bdev_capacity(struct block_device *bdev, boolean_t wholedisk)
 {
 	struct hd_struct *part = bdev->bd_part;
+	uint64_t sectors = get_capacity(bdev->bd_disk);
+	/* If there are no paritions, return the entire device capacity */
+	if (part == NULL)
+		return (sectors << SECTOR_BITS);
 
-	/* The partition capacity referenced by the block device */
-	if (part)
-		return (part->nr_sects << 9);
-
-	/* Otherwise assume the full device capacity */
-	return (get_capacity(bdev->bd_disk) << 9);
+	/*
+	 * If there are partitions, decide if we are using a `wholedisk`
+	 * layout (composed of part1 and part9) or just a single partition.
+	 */
+	if (wholedisk) {
+		/* Verify the expected device layout */
+		ASSERT3P(bdev, !=, bdev->bd_contains);
+		/*
+		 * Sectors used by the EFI partition (part9) as well as
+		 * partion alignment.
+		 */
+		uint64_t used = EFI_MIN_RESV_SIZE + NEW_START_BLOCK +
+		    PARTITION_END_ALIGNMENT;
+
+		/* Space available to the vdev, i.e. the size of part1 */
+		if (sectors <= used)
+			return (0);
+		uint64_t available = sectors - used;
+		return (available << SECTOR_BITS);
+	} else {
+		/* The partition capacity referenced by the block device */
+		return (part->nr_sects << SECTOR_BITS);
+	}
 }
 
 static void
@@ -330,9 +356,7 @@ skip_open:
 	v->vdev_nonrot = blk_queue_nonrot(bdev_get_queue(vd->vd_bdev));
 
 	/* Physical volume size in bytes */
-	*psize = bdev_capacity(vd->vd_bdev);
-
-	/* TODO: report possible expansion size */
+	*psize = bdev_capacity(vd->vd_bdev, v->vdev_wholedisk);
 	*max_psize = *psize;
 
 	/* Based on the minimum sector size set the block size */
diff --git a/tests/zfs-tests/tests/functional/cli_root/zpool_expand/zpool_expand_002_pos.ksh b/tests/zfs-tests/tests/functional/cli_root/zpool_expand/zpool_expand_002_pos.ksh
index d578ae602..66b6969db 100755
--- a/tests/zfs-tests/tests/functional/cli_root/zpool_expand/zpool_expand_002_pos.ksh
+++ b/tests/zfs-tests/tests/functional/cli_root/zpool_expand/zpool_expand_002_pos.ksh
@@ -26,7 +26,7 @@
 #
 
 #
-# Copyright (c) 2012, 2016 by Delphix. All rights reserved.
+# Copyright (c) 2012, 2018 by Delphix. All rights reserved.
 # Copyright (c) 2017 Lawrence Livermore National Security, LLC.
 #
 
@@ -43,8 +43,9 @@
 # 1) Create 3 files
 # 2) Create a pool backed by the files
 # 3) Expand the files' size with truncate
-# 4) Use zpool online -e to online the vdevs
-# 5) Check that the pool size was expanded
+# 4) Use zpool reopen to check the expandsize
+# 5) Use zpool online -e to online the vdevs
+# 6) Check that the pool size was expanded
 #
 
 verify_runnable "global"
@@ -64,8 +65,8 @@ log_onexit cleanup
 
 log_assert "zpool can expand after zpool online -e zvol vdevs on LUN expansion"
 
-
 for type in " " mirror raidz raidz2; do
+	# Initialize the file devices and the pool
 	for i in 1 2 3; do
 		log_must truncate -s $org_size ${TEMPFILE}.$i
 	done
@@ -80,13 +81,35 @@ for type in " " mirror raidz raidz2; do
 		    "$autoexp"
 	fi
 	typeset prev_size=$(get_pool_prop size $TESTPOOL1)
-	typeset zfs_prev_size=$(zfs get -p avail $TESTPOOL1 | tail -1 | \
-	    awk '{print $3}')
+	typeset zfs_prev_size=$(get_prop avail $TESTPOOL1)
 
+	# Increase the size of the file devices
 	for i in 1 2 3; do
 		log_must truncate -s $exp_size ${TEMPFILE}.$i
 	done
 
+	# Reopen the pool and check that the `expandsize` property is set
+	log_must zpool reopen $TESTPOOL1
+	typeset zpool_expandsize=$(get_pool_prop expandsize $TESTPOOL1)
+
+	if [[ $type == "mirror" ]]; then
+		typeset expected_zpool_expandsize=$(($exp_size-$org_size))
+	else
+		typeset expected_zpool_expandsize=$((3*($exp_size-$org_size)))
+	fi
+
+	if [[ "$zpool_expandsize" = "-" ]]; then
+		log_fail "pool $TESTPOOL1 did not detect any " \
+		    "expandsize after reopen"
+	fi
+
+	if [[ $zpool_expandsize -ne $expected_zpool_expandsize ]]; then
+		log_fail "pool $TESTPOOL1 did not detect correct " \
+		    "expandsize after reopen: found $zpool_expandsize," \
+		    "expected $expected_zpool_expandsize"
+	fi
+
+	# Online the devices to add the new space to the pool
 	for i in 1 2 3; do
 		log_must zpool online -e $TESTPOOL1 ${TEMPFILE}.$i
 	done
@@ -96,8 +119,7 @@ for type in " " mirror raidz raidz2; do
 	sync
 
 	typeset expand_size=$(get_pool_prop size $TESTPOOL1)
-	typeset zfs_expand_size=$(zfs get -p avail $TESTPOOL1 | tail -1 | \
-	    awk '{print $3}')
+	typeset zfs_expand_size=$(get_prop avail $TESTPOOL1)
 	log_note "$TESTPOOL1 $type has previous size: $prev_size and " \
 	    "expanded size: $expand_size"
 
@@ -112,8 +134,8 @@ for type in " " mirror raidz raidz2; do
 			    grep "(+${expansion_size}" | wc -l)
 
 			if [[ $size_addition -ne $i ]]; then
-				log_fail "pool $TESTPOOL1 is not autoexpand " \
-				    "after LUN expansion"
+				log_fail "pool $TESTPOOL1 did not expand " \
+				    "after LUN expansion and zpool online -e"
 			fi
 		elif [[ $type == "mirror" ]]; then
 			typeset expansion_size=$(($exp_size-$org_size))
@@ -123,8 +145,8 @@ for type in " " mirror raidz raidz2; do
 			    grep "(+${expansion_size})" >/dev/null 2>&1
 
 			if [[ $? -ne 0 ]]; then
-				log_fail "pool $TESTPOOL1 is not autoexpand " \
-				    "after LUN expansion"
+				log_fail "pool $TESTPOOL1 did not expand " \
+				    "after LUN expansion and zpool online -e"
 			fi
 		else
 			typeset expansion_size=$((3*($exp_size-$org_size)))
@@ -134,13 +156,13 @@ for type in " " mirror raidz raidz2; do
 			    grep "(+${expansion_size})" >/dev/null 2>&1
 
 			if [[ $? -ne 0 ]] ; then
-				log_fail "pool $TESTPOOL1 is not autoexpand " \
-				    "after LUN expansion"
+				log_fail "pool $TESTPOOL1 did not expand " \
+				    "after LUN expansion and zpool online -e"
 			fi
 		fi
 	else
-		log_fail "pool $TESTPOOL1 is not autoexpanded after LUN " \
-		    "expansion"
+		log_fail "pool $TESTPOOL1 did not expand after LUN expansion " \
+		    "and zpool online -e"
 	fi
 	log_must zpool destroy $TESTPOOL1
 done