MFC r260138: MFV r242733:

3306 zdb should be able to issue reads in parallel
3321 'zpool reopen' command should be documented in the man page
and help message

illumos/illumos-gate@31d7e8fa33fae995f558673adb22641b5aa8b6e1

FreeBSD porting notes: the kernel part of this changeset depends
on Solaris buf(9S) interfaces and are not really applicable for
our use.  vdev_disk.c is patched as-is to reduce diverge from
upstream, but vdev_file.c is left intact.

8 files changed, 230 insertions, 59 deletions

diff --git a/cddl/contrib/opensolaris/cmd/zdb/zdb.8 b/cddl/contrib/opensolaris/cmd/zdb/zdb.8
index e036b96..fbf1fce 100644
--- a/cddl/contrib/opensolaris/cmd/zdb/zdb.8
+++ b/cddl/contrib/opensolaris/cmd/zdb/zdb.8
@@ -14,11 +14,12 @@
 .\"
 .\" Copyright 2012, Richard Lowe.
 .\" Copyright (c) 2012, Marcelo Araujo <araujo@FreeBSD.org>.
+.\" Copyright (c) 2012 by Delphix. All rights reserved.
 .\" All Rights Reserved.
 .\"
 .\" $FreeBSD$
 .\"
-.Dd May 10, 2012
+.Dd December 31, 2013
 .Dt ZDB 8
 .Os
 .Sh NAME
@@ -29,27 +30,35 @@
 .Op Fl CumdibcsDvhLXFPA
 .Op Fl e Op Fl p Ar path...
 .Op Fl t Ar txg
+.Op Fl U Ar cache
+.Op Fl M Ar inflight I/Os
 .Ar poolname
 .Op Ar object ...
 .Nm
 .Op Fl divPA
 .Op Fl e Op Fl p Ar path...
+.Op Fl U Ar cache
 .Ar dataset
 .Op Ar object ...
 .Nm
 .Fl m Op Fl LXFPA
 .Op Fl t Ar txg
 .Op Fl e Op Fl p Ar path...
+.Op Fl U Ar cache
 .Ar poolname
 .Nm
 .Fl R Op Fl A
 .Op Fl e Op Fl p Ar path...
+.Op Fl U Ar cache
+.Ar poolname
 .Ar poolname
 .Ar vdev Ns : Ns Ar offset Ns : Ns Ar size Ns Op Ns : Ns Ar flags
 .Nm
 .Fl S
 .Op Fl AP
 .Op Fl e Op Fl p Ar path...
+.Op Fl U Ar cache
+.Ar poolname
 .Ar poolname
 .Nm
 .Fl l
@@ -205,6 +214,11 @@ flag specifies the path under which devices are to be searched.
 .It Fl F
 Attempt to make an unreadable pool readable by trying progressively older
 transactions.
+.It Fl M Ar inflight I/Os
+Limit the number of outstanding checksum I/Os to the specified value.
+The default value is 200. This option affects the performance of the
+.Fl c
+option.
 .It Fl P
 Print numbers in an unscaled form more amenable to parsing, eg. 1000000 rather
 than 1M.
@@ -218,9 +232,7 @@ options for a means to see the available uberblocks and their associated
 transaction numbers.
 .It Fl U Ar cachefile
 Use a cache file other than
-.Pa /etc/zfs/zpool.cache .
-This option is only valid with
-.Fl C
+.Pa /boot/zfs/zpool.cache .
 .It Fl v
 Enable verbosity.
 Specify multiple times for increased verbosity.
diff --git a/cddl/contrib/opensolaris/cmd/zdb/zdb.c b/cddl/contrib/opensolaris/cmd/zdb/zdb.c
index 02a6265..9294cd6 100644
--- a/cddl/contrib/opensolaris/cmd/zdb/zdb.c
+++ b/cddl/contrib/opensolaris/cmd/zdb/zdb.c
@@ -89,6 +89,7 @@ extern void dump_intent_log(zilog_t *);
 uint64_t *zopt_object = NULL;
 int zopt_objects = 0;
 libzfs_handle_t *g_zfs;
+uint64_t max_inflight = 200;
 
 /*
  * These libumem hooks provide a reasonable set of defaults for the allocator's
@@ -110,16 +111,17 @@ static void
 usage(void)
 {
 	(void) fprintf(stderr,
-            "Usage: %s [-CumdibcsDvhLXFPA] [-t txg] [-e [-p path...]]"
-            "poolname [object...]\n"
-            "       %s [-divPA] [-e -p path...] dataset [object...]\n"
-            "       %s -m [-LXFPA] [-t txg] [-e [-p path...]]"
-            "poolname [vdev [metaslab...]]\n"
-            "       %s -R [-A] [-e [-p path...]] poolname "
-            "vdev:offset:size[:flags]\n"
-            "       %s -S [-PA] [-e [-p path...]] poolname\n"
-            "       %s -l [-uA] device\n"
-            "       %s -C [-A] [-U config]\n\n",
+	    "Usage: %s [-CumdibcsDvhLXFPA] [-t txg] [-e [-p path...]] "
+	    "[-U config] [-M inflight I/Os] poolname [object...]\n"
+	    "       %s [-divPA] [-e -p path...] [-U config] dataset "
+	    "[object...]\n"
+	    "       %s -m [-LXFPA] [-t txg] [-e [-p path...]] [-U config] "
+	    "poolname [vdev [metaslab...]]\n"
+	    "       %s -R [-A] [-e [-p path...]] poolname "
+	    "vdev:offset:size[:flags]\n"
+	    "       %s -S [-PA] [-e [-p path...]] [-U config] poolname\n"
+	    "       %s -l [-uA] device\n"
+	    "       %s -C [-A] [-U config]\n\n",
 	    cmdname, cmdname, cmdname, cmdname, cmdname, cmdname, cmdname);
 
 	(void) fprintf(stderr, "    Dataset name must include at least one "
@@ -164,6 +166,8 @@ usage(void)
 	(void) fprintf(stderr, "	-P print numbers in parseable form\n");
 	(void) fprintf(stderr, "        -t <txg> -- highest txg to use when "
 	    "searching for uberblocks\n");
+	(void) fprintf(stderr, "        -M <number of inflight I/Os> -- "
+	    "specify the maximum number of checksumming I/Os [default is 200]");
 	(void) fprintf(stderr, "Specify an option more than once (e.g. -bb) "
 	    "to make only that option verbose\n");
 	(void) fprintf(stderr, "Default is to dump everything non-verbosely\n");
@@ -2154,6 +2158,47 @@ zdb_count_block(zdb_cb_t *zcb, zilog_t *zilog, const blkptr_t *bp,
 	    bp, NULL, NULL, ZIO_FLAG_CANFAIL)), ==, 0);
 }
 
+/* ARGSUSED */
+static void
+zdb_blkptr_done(zio_t *zio)
+{
+	spa_t *spa = zio->io_spa;
+	blkptr_t *bp = zio->io_bp;
+	int ioerr = zio->io_error;
+	zdb_cb_t *zcb = zio->io_private;
+	zbookmark_t *zb = &zio->io_bookmark;
+
+	zio_data_buf_free(zio->io_data, zio->io_size);
+
+	mutex_enter(&spa->spa_scrub_lock);
+	spa->spa_scrub_inflight--;
+	cv_broadcast(&spa->spa_scrub_io_cv);
+
+	if (ioerr && !(zio->io_flags & ZIO_FLAG_SPECULATIVE)) {
+		char blkbuf[BP_SPRINTF_LEN];
+
+		zcb->zcb_haderrors = 1;
+		zcb->zcb_errors[ioerr]++;
+
+		if (dump_opt['b'] >= 2)
+			sprintf_blkptr(blkbuf, bp);
+		else
+			blkbuf[0] = '\0';
+
+		(void) printf("zdb_blkptr_cb: "
+		    "Got error %d reading "
+		    "<%llu, %llu, %lld, %llx> %s -- skipping\n",
+		    ioerr,
+		    (u_longlong_t)zb->zb_objset,
+		    (u_longlong_t)zb->zb_object,
+		    (u_longlong_t)zb->zb_level,
+		    (u_longlong_t)zb->zb_blkid,
+		    blkbuf);
+	}
+	mutex_exit(&spa->spa_scrub_lock);
+}
+
+/* ARGSUSED */
 static int
 zdb_blkptr_cb(spa_t *spa, zilog_t *zilog, const blkptr_t *bp,
     const zbookmark_t *zb, const dnode_phys_t *dnp, void *arg)
@@ -2174,38 +2219,22 @@ zdb_blkptr_cb(spa_t *spa, zilog_t *zilog, const blkptr_t *bp,
 	is_metadata = (BP_GET_LEVEL(bp) != 0 || DMU_OT_IS_METADATA(type));
 
 	if (dump_opt['c'] > 1 || (dump_opt['c'] && is_metadata)) {
-		int ioerr;
 		size_t size = BP_GET_PSIZE(bp);
-		void *data = malloc(size);
+		void *data = zio_data_buf_alloc(size);
 		int flags = ZIO_FLAG_CANFAIL | ZIO_FLAG_SCRUB | ZIO_FLAG_RAW;
 
 		/* If it's an intent log block, failure is expected. */
 		if (zb->zb_level == ZB_ZIL_LEVEL)
 			flags |= ZIO_FLAG_SPECULATIVE;
 
-		ioerr = zio_wait(zio_read(NULL, spa, bp, data, size,
-		    NULL, NULL, ZIO_PRIORITY_ASYNC_READ, flags, zb));
+		mutex_enter(&spa->spa_scrub_lock);
+		while (spa->spa_scrub_inflight > max_inflight)
+			cv_wait(&spa->spa_scrub_io_cv, &spa->spa_scrub_lock);
+		spa->spa_scrub_inflight++;
+		mutex_exit(&spa->spa_scrub_lock);
 
-		free(data);
-		if (ioerr && !(flags & ZIO_FLAG_SPECULATIVE)) {
-			zcb->zcb_haderrors = 1;
-			zcb->zcb_errors[ioerr]++;
-
-			if (dump_opt['b'] >= 2)
-				sprintf_blkptr(blkbuf, bp);
-			else
-				blkbuf[0] = '\0';
-
-			(void) printf("zdb_blkptr_cb: "
-			    "Got error %d reading "
-			    "<%llu, %llu, %lld, %llx> %s -- skipping\n",
-			    ioerr,
-			    (u_longlong_t)zb->zb_objset,
-			    (u_longlong_t)zb->zb_object,
-			    (u_longlong_t)zb->zb_level,
-			    (u_longlong_t)zb->zb_blkid,
-			    blkbuf);
-		}
+		zio_nowait(zio_read(NULL, spa, bp, data, size,
+		    zdb_blkptr_done, zcb, ZIO_PRIORITY_ASYNC_READ, flags, zb));
 	}
 
 	zcb->zcb_readfails = 0;
@@ -2433,6 +2462,18 @@ dump_block_stats(spa_t *spa)
 	zcb.zcb_start = zcb.zcb_lastprint = gethrtime();
 	zcb.zcb_haderrors |= traverse_pool(spa, 0, flags, zdb_blkptr_cb, &zcb);
 
+	/*
+	 * If we've traversed the data blocks then we need to wait for those
+	 * I/Os to complete. We leverage "The Godfather" zio to wait on
+	 * all async I/Os to complete.
+	 */
+	if (dump_opt['c']) {
+		(void) zio_wait(spa->spa_async_zio_root);
+		spa->spa_async_zio_root = zio_root(spa, NULL, NULL,
+		    ZIO_FLAG_CANFAIL | ZIO_FLAG_SPECULATIVE |
+		    ZIO_FLAG_GODFATHER);
+	}
+
 	if (zcb.zcb_haderrors) {
 		(void) printf("\nError counts:\n\n");
 		(void) printf("\t%5s  %s\n", "errno", "count");
@@ -3203,7 +3244,7 @@ main(int argc, char **argv)
 
 	dprintf_setup(&argc, argv);
 
-	while ((c = getopt(argc, argv, "bcdhilmsuCDRSAFLXevp:t:U:P")) != -1) {
+	while ((c = getopt(argc, argv, "bcdhilmM:suCDRSAFLXevp:t:U:P")) != -1) {
 		switch (c) {
 		case 'b':
 		case 'c':
@@ -3232,6 +3273,15 @@ main(int argc, char **argv)
 		case 'v':
 			verbose++;
 			break;
+		case 'M':
+			max_inflight = strtoull(optarg, NULL, 0);
+			if (max_inflight == 0) {
+				(void) fprintf(stderr, "maximum number "
+				    "of inflight I/Os must be greater "
+				    "than 0\n");
+				usage();
+			}
+			break;
 		case 'p':
 			if (searchdirs == NULL) {
 				searchdirs = umem_alloc(sizeof (char *),
diff --git a/cddl/contrib/opensolaris/cmd/zpool/zpool.8 b/cddl/contrib/opensolaris/cmd/zpool/zpool.8
index ebd22c5..0f4f4ea 100644
--- a/cddl/contrib/opensolaris/cmd/zpool/zpool.8
+++ b/cddl/contrib/opensolaris/cmd/zpool/zpool.8
@@ -25,7 +25,7 @@
 .\"
 .\" $FreeBSD$
 .\"
-.Dd March 14, 2013
+.Dd December 31, 2013
 .Dt ZPOOL 8
 .Os
 .Sh NAME
@@ -141,6 +141,9 @@
 .Cm remove
 .Ar pool device ...
 .Nm
+.Cm reopen
+.Ar pool
+.Nm
 .Cm replace
 .Op Fl f
 .Ar pool device
@@ -1431,6 +1434,13 @@ command. Non-redundant and
 devices cannot be removed from a pool.
 .It Xo
 .Nm
+.Cm reopen
+.Ar pool
+.Xc
+.Pp
+Reopen all the vdevs associated with the pool.
+.It Xo
+.Nm
 .Cm replace
 .Op Fl f
 .Ar pool device
diff --git a/cddl/contrib/opensolaris/cmd/zpool/zpool_main.c b/cddl/contrib/opensolaris/cmd/zpool/zpool_main.c
index 09ab280..dd3b2eb 100644
--- a/cddl/contrib/opensolaris/cmd/zpool/zpool_main.c
+++ b/cddl/contrib/opensolaris/cmd/zpool/zpool_main.c
@@ -248,7 +248,7 @@ get_usage(zpool_help_t idx) {
 	case HELP_REMOVE:
 		return (gettext("\tremove <pool> <device> ...\n"));
 	case HELP_REOPEN:
-		return (""); /* Undocumented command */
+		return (gettext("\treopen <pool>\n"));
 	case HELP_SCRUB:
 		return (gettext("\tscrub [-s] <pool> ...\n"));
 	case HELP_STATUS:
@@ -3720,22 +3720,37 @@ zpool_do_reguid(int argc, char **argv)
  * zpool reopen <pool>
  *
  * Reopen the pool so that the kernel can update the sizes of all vdevs.
- *
- * NOTE: This command is currently undocumented.  If the command is ever
- * exposed then the appropriate usage() messages will need to be made.
  */
 int
 zpool_do_reopen(int argc, char **argv)
 {
+	int c;
 	int ret = 0;
 	zpool_handle_t *zhp;
 	char *pool;
 
+	/* check options */
+	while ((c = getopt(argc, argv, "")) != -1) {
+		switch (c) {
+		case '?':
+			(void) fprintf(stderr, gettext("invalid option '%c'\n"),
+			    optopt);
+			usage(B_FALSE);
+		}
+	}
+
 	argc--;
 	argv++;
 
-	if (argc != 1)
-		return (2);
+	if (argc < 1) {
+		(void) fprintf(stderr, gettext("missing pool name\n"));
+		usage(B_FALSE);
+	}
+
+	if (argc > 1) {
+		(void) fprintf(stderr, gettext("too many arguments\n"));
+		usage(B_FALSE);
+	}
 
 	pool = argv[0];
 	if ((zhp = zpool_open_canfail(g_zfs, pool)) == NULL)
diff --git a/cddl/contrib/opensolaris/lib/libzpool/common/kernel.c b/cddl/contrib/opensolaris/lib/libzpool/common/kernel.c
index c4e2d2e..19fbf4f 100644
--- a/cddl/contrib/opensolaris/lib/libzpool/common/kernel.c
+++ b/cddl/contrib/opensolaris/lib/libzpool/common/kernel.c
@@ -661,7 +661,7 @@ __dprintf(const char *file, const char *func, int line, const char *fmt, ...)
 		if (dprintf_find_string("pid"))
 			(void) printf("%d ", getpid());
 		if (dprintf_find_string("tid"))
-			(void) printf("%u ", thr_self());
+			(void) printf("%ul ", thr_self());
 #if 0
 		if (dprintf_find_string("cpu"))
 			(void) printf("%u ", getcpuid());
@@ -1125,3 +1125,50 @@ zvol_create_minors(const char *name)
 	return (0);
 }
 #endif
+
+#ifdef illumos
+void
+bioinit(buf_t *bp)
+{
+	bzero(bp, sizeof (buf_t));
+}
+
+void
+biodone(buf_t *bp)
+{
+	if (bp->b_iodone != NULL) {
+		(*(bp->b_iodone))(bp);
+		return;
+	}
+	ASSERT((bp->b_flags & B_DONE) == 0);
+	bp->b_flags |= B_DONE;
+}
+
+void
+bioerror(buf_t *bp, int error)
+{
+	ASSERT(bp != NULL);
+	ASSERT(error >= 0);
+
+	if (error != 0) {
+		bp->b_flags |= B_ERROR;
+	} else {
+		bp->b_flags &= ~B_ERROR;
+	}
+	bp->b_error = error;
+}
+
+
+int
+geterror(struct buf *bp)
+{
+	int error = 0;
+
+	if (bp->b_flags & B_ERROR) {
+		error = bp->b_error;
+		if (!error)
+			error = EIO;
+	}
+	return (error);
+}
+#endif
diff --git a/cddl/contrib/opensolaris/lib/libzpool/common/sys/zfs_context.h b/cddl/contrib/opensolaris/lib/libzpool/common/sys/zfs_context.h
index 471756d..b5562c7 100644
--- a/cddl/contrib/opensolaris/lib/libzpool/common/sys/zfs_context.h
+++ b/cddl/contrib/opensolaris/lib/libzpool/common/sys/zfs_context.h
@@ -778,6 +778,38 @@ extern void cyclic_remove(cyclic_id_t);
 extern int cyclic_reprogram(cyclic_id_t, hrtime_t);
 #endif	/* illumos */
 
+#ifdef illumos
+/*
+ * Buf structure
+ */
+#define	B_BUSY		0x0001
+#define	B_DONE		0x0002
+#define	B_ERROR		0x0004
+#define	B_READ		0x0040	/* read when I/O occurs */
+#define	B_WRITE		0x0100	/* non-read pseudo-flag */
+
+typedef struct buf {
+	int	b_flags;
+	size_t b_bcount;
+	union {
+		caddr_t b_addr;
+	} b_un;
+
+	lldaddr_t	_b_blkno;
+#define	b_lblkno	_b_blkno._f
+	size_t	b_resid;
+	size_t	b_bufsize;
+	int	(*b_iodone)(struct buf *);
+	int	b_error;
+	void	*b_private;
+} buf_t;
+
+extern void bioinit(buf_t *);
+extern void biodone(buf_t *);
+extern void bioerror(buf_t *, int);
+extern int geterror(buf_t *);
+#endif
+
 #ifdef	__cplusplus
 }
 #endif
diff --git a/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/sys/vdev_impl.h b/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/sys/vdev_impl.h
index 307e557..518ebc4 100644
--- a/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/sys/vdev_impl.h
+++ b/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/sys/vdev_impl.h
@@ -359,6 +359,16 @@ extern void vdev_set_min_asize(vdev_t *vd);
 /* zdb uses this tunable, so it must be declared here to make lint happy. */
 extern int zfs_vdev_cache_size;
 
+#ifdef illumos
+/*
+ * The vdev_buf_t is used to translate between zio_t and buf_t, and back again.
+ */
+typedef struct vdev_buf {
+	buf_t	vb_buf;		/* buffer that describes the io */
+	zio_t	*vb_io;		/* pointer back to the original zio_t */
+} vdev_buf_t;
+#endif
+
 #ifdef	__cplusplus
 }
 #endif
diff --git a/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/vdev_disk.c b/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/vdev_disk.c
index 3d11966..2785646 100644
--- a/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/vdev_disk.c
+++ b/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/vdev_disk.c
@@ -42,11 +42,6 @@
 
 extern ldi_ident_t zfs_li;
 
-typedef struct vdev_disk_buf {
-	buf_t	vdb_buf;
-	zio_t	*vdb_io;
-} vdev_disk_buf_t;
-
 static void
 vdev_disk_hold(vdev_t *vd)
 {
@@ -483,8 +478,8 @@ vdev_disk_ldi_physio(ldi_handle_t vd_lh, caddr_t data,
 static void
 vdev_disk_io_intr(buf_t *bp)
 {
-	vdev_disk_buf_t *vdb = (vdev_disk_buf_t *)bp;
-	zio_t *zio = vdb->vdb_io;
+	vdev_buf_t *vb = (vdev_buf_t *)bp;
+	zio_t *zio = vb->vb_io;
 
 	/*
 	 * The rest of the zio stack only deals with EIO, ECKSUM, and ENXIO.
@@ -496,7 +491,7 @@ vdev_disk_io_intr(buf_t *bp)
 	if (zio->io_error == 0 && bp->b_resid != 0)
 		zio->io_error = SET_ERROR(EIO);
 
-	kmem_free(vdb, sizeof (vdev_disk_buf_t));
+	kmem_free(vb, sizeof (vdev_buf_t));
 
 	zio_interrupt(zio);
 }
@@ -527,7 +522,7 @@ vdev_disk_io_start(zio_t *zio)
 {
 	vdev_t *vd = zio->io_vd;
 	vdev_disk_t *dvd = vd->vdev_tsd;
-	vdev_disk_buf_t *vdb;
+	vdev_buf_t *vb;
 	struct dk_callback *dkc;
 	buf_t *bp;
 	int error;
@@ -591,10 +586,10 @@ vdev_disk_io_start(zio_t *zio)
 		return (ZIO_PIPELINE_CONTINUE);
 	}
 
-	vdb = kmem_alloc(sizeof (vdev_disk_buf_t), KM_SLEEP);
+	vb = kmem_alloc(sizeof (vdev_buf_t), KM_SLEEP);
 
-	vdb->vdb_io = zio;
-	bp = &vdb->vdb_buf;
+	vb->vb_io = zio;
+	bp = &vb->vb_buf;
 
 	bioinit(bp);
 	bp->b_flags = B_BUSY | B_NOCACHE |