Correct some issues with zfs boot.

 - Teach it to read gang blocks. (essentially untested)
   If you see "ZFS: gang block detected!", please let
   me know, so we can either remove the printf if it
   works, or fix it if it doesn't.

 - If multiple partitions exist on a disk, probe them all.
   We also need to reset dsk->start to 0 to read the right
   sector here.

 - With GPT, we can have 128 partitions.

 - If the bootfs property has ever been set on a pool
   it seems that it never goes away.  zpool won't allow
   you to add to the pool with the bootfs property set.
   However, if you clear the property back to default
   we end up getting 0 for the object number and read
   a bogus block pointer and fail to boot.

 - Fix some error printfs. The printf in the loader is
   only capable of c,s and u formats.

 - Teach printf how to display %llu

Reviewed by:	dfr, jhb
MFC after:	2 weeks

4 files changed, 93 insertions, 23 deletions

diff --git a/sys/boot/i386/zfsboot/zfsboot.c b/sys/boot/i386/zfsboot/zfsboot.c
index a654393..c0dedd7 100644
--- a/sys/boot/i386/zfsboot/zfsboot.c
+++ b/sys/boot/i386/zfsboot/zfsboot.c
@@ -474,6 +474,7 @@ probe_drive(struct dsk *dsk, spa_t **spap)
     slba = hdr.hdr_lba_table;
     elba = slba + hdr.hdr_entries / entries_per_sec;
     while (slba < elba) {
+	dsk->start = 0;
 	if (drvread(dsk, sec, slba, 1))
 	    return;
 	for (part = 0; part < entries_per_sec; part++) {
@@ -494,7 +495,6 @@ probe_drive(struct dsk *dsk, spa_t **spap)
 		     */
 		    dsk = copy_dsk(dsk);
 		}
-		break;
 	    }
 	}
 	slba++;
@@ -857,12 +857,13 @@ static void
 printf(const char *fmt,...)
 {
     va_list ap;
-    char buf[10];
+    char buf[20];
     char *s;
-    unsigned u;
+    unsigned long long u;
     int c;
     int minus;
     int prec;
+    int l;
     int len;
     int pad;
 
@@ -871,6 +872,7 @@ printf(const char *fmt,...)
 	if (c == '%') {
 	    minus = 0;
 	    prec = 0;
+	    l = 0;
 	nextfmt:
 	    c = *fmt++;
 	    switch (c) {
@@ -892,6 +894,9 @@ printf(const char *fmt,...)
 	    case 'c':
 		putchar(va_arg(ap, int));
 		continue;
+	    case 'l':
+		l++;
+		goto nextfmt;
 	    case 's':
 		s = va_arg(ap, char *);
 		if (prec) {
@@ -914,7 +919,17 @@ printf(const char *fmt,...)
 		}
 		continue;
 	    case 'u':
-		u = va_arg(ap, unsigned);
+		switch (l) {
+		case 2:
+		    u = va_arg(ap, unsigned long long);
+		    break;
+		case 1:
+		    u = va_arg(ap, unsigned long);
+		    break;
+		default:
+		    u = va_arg(ap, unsigned);
+		    break;
+		}
 		s = buf;
 		do
 		    *s++ = '0' + u % 10U;
diff --git a/sys/boot/zfs/zfs.c b/sys/boot/zfs/zfs.c
index 9784ef9..52df773 100644
--- a/sys/boot/zfs/zfs.c
+++ b/sys/boot/zfs/zfs.c
@@ -100,7 +100,7 @@ zfs_open(const char *upath, struct open_file *f)
 	f->f_fsdata = (void *)fp;
 
 	if (spa->spa_root_objset.os_type != DMU_OST_ZFS) {
-		printf("Unexpected object set type %lld\n",
+		printf("Unexpected object set type %llu\n",
 		    spa->spa_root_objset.os_type);
 		rc = EIO;
 		goto out;
@@ -413,7 +413,7 @@ zfs_dev_init(void)
 		if (vdev_probe(vdev_read, (void*) (uintptr_t) fd, 0))
 			close(fd);
 
-		for (slice = 1; slice <= 4; slice++) {
+		for (slice = 1; slice <= 128; slice++) {
 			sprintf(devname, "disk%dp%d:", unit, slice);
 			fd = open(devname, O_RDONLY);
 			if (fd == -1) {
diff --git a/sys/boot/zfs/zfsimpl.c b/sys/boot/zfs/zfsimpl.c
index ff567a4..497fd7c 100644
--- a/sys/boot/zfs/zfsimpl.c
+++ b/sys/boot/zfs/zfsimpl.c
@@ -53,6 +53,8 @@ static char *zfs_temp_buf, *zfs_temp_end, *zfs_temp_ptr;
 
 #define TEMP_SIZE	(1*SPA_MAXBLOCKSIZE)
 
+static int zio_read(spa_t *spa, const blkptr_t *bp, void *buf);
+
 static void
 zfs_init(void)
 {
@@ -897,6 +899,33 @@ ilog2(int n)
 }
 
 static int
+zio_read_gang(spa_t *spa, const blkptr_t *bp, const dva_t *dva, void *buf)
+{
+	zio_gbh_phys_t zio_gb;
+	vdev_t *vdev;
+	int vdevid;
+	off_t offset;
+	int i;
+
+	vdevid = DVA_GET_VDEV(dva);
+	offset = DVA_GET_OFFSET(dva);
+	STAILQ_FOREACH(vdev, &spa->spa_vdevs, v_childlink)
+		if (vdev->v_id == vdevid)
+			break;
+	if (!vdev || !vdev->v_read)
+		return (EIO);
+	if (vdev->v_read(vdev, bp, &zio_gb, offset, SPA_GANGBLOCKSIZE))
+		return (EIO);
+
+	for (i = 0; i < SPA_GBH_NBLKPTRS; i++) {
+		if (zio_read(spa, &zio_gb.zg_blkptr[i], buf))
+			return (EIO);
+	}
+ 
+	return (0);
+}
+
+static int
 zio_read(spa_t *spa, const blkptr_t *bp, void *buf)
 {
 	int cpfunc = BP_GET_COMPRESS(bp);
@@ -920,20 +949,27 @@ zio_read(spa_t *spa, const blkptr_t *bp, void *buf)
 		if (!dva->dva_word[0] && !dva->dva_word[1])
 			continue;
 
-		vdevid = DVA_GET_VDEV(dva);
-		offset = DVA_GET_OFFSET(dva);
-		STAILQ_FOREACH(vdev, &spa->spa_vdevs, v_childlink)
-			if (vdev->v_id == vdevid)
-				break;
-		if (!vdev || !vdev->v_read)
-			continue;
-		if (vdev->v_read(vdev, bp, pbuf, offset, psize))
-			continue;
+		if (DVA_GET_GANG(dva)) {
+			printf("ZFS: gang block detected!\n");
+			if (zio_read_gang(spa, bp, dva, buf))
+				return (EIO); 
+		} else {
+			vdevid = DVA_GET_VDEV(dva);
+			offset = DVA_GET_OFFSET(dva);
+			STAILQ_FOREACH(vdev, &spa->spa_vdevs, v_childlink)
+				if (vdev->v_id == vdevid)
+					break;
+			if (!vdev || !vdev->v_read) {
+				continue;
+			}
+			if (vdev->v_read(vdev, bp, pbuf, offset, psize))
+				continue;
 
-		if (cpfunc != ZIO_COMPRESS_OFF) {
-			if (zio_decompress_data(cpfunc, pbuf, psize,
-				buf, lsize))
-				return (EIO);
+			if (cpfunc != ZIO_COMPRESS_OFF) {
+				if (zio_decompress_data(cpfunc, pbuf, psize,
+				    buf, lsize))
+					return (EIO);
+			}
 		}
 
 		return (0);
@@ -1331,13 +1367,13 @@ zfs_mount_dataset(spa_t *spa, uint64_t objnum, objset_phys_t *objset)
 	dsl_dataset_phys_t *ds;
 
 	if (objset_get_dnode(spa, &spa->spa_mos, objnum, &dataset)) {
-		printf("ZFS: can't find dataset %lld\n", objnum);
+		printf("ZFS: can't find dataset %llu\n", objnum);
 		return (EIO);
 	}
 
 	ds = (dsl_dataset_phys_t *) &dataset.dn_bonus;
 	if (zio_read(spa, &ds->ds_bp, objset)) {
-		printf("ZFS: can't read object set for dataset %lld\n", objnum);
+		printf("ZFS: can't read object set for dataset %llu\n", objnum);
 		return (EIO);
 	}
 
@@ -1367,7 +1403,8 @@ zfs_mount_root(spa_t *spa, objset_phys_t *objset)
 	 */
 	if (zap_lookup(spa, &dir, DMU_POOL_PROPS, &props) == 0
 	     && objset_get_dnode(spa, &spa->spa_mos, props, &propdir) == 0
-	     && zap_lookup(spa, &propdir, "bootfs", &bootfs) == 0)
+	     && zap_lookup(spa, &propdir, "bootfs", &bootfs) == 0
+	     && bootfs != 0)
 		return zfs_mount_dataset(spa, bootfs, objset);
 
 	/*
@@ -1425,7 +1462,7 @@ zfs_lookup(spa_t *spa, const char *upath, dnode_phys_t *dnode)
 	int symlinks_followed = 0;
 
 	if (spa->spa_root_objset.os_type != DMU_OST_ZFS) {
-		printf("ZFS: unexpected object set type %lld\n",
+		printf("ZFS: unexpected object set type %llu\n",
 		       spa->spa_root_objset.os_type);
 		return (EIO);
 	}
diff --git a/sys/cddl/boot/zfs/zfsimpl.h b/sys/cddl/boot/zfs/zfsimpl.h
index a0b7b72..688bb5c 100644
--- a/sys/cddl/boot/zfs/zfsimpl.h
+++ b/sys/cddl/boot/zfs/zfsimpl.h
@@ -374,6 +374,24 @@ typedef struct vdev_label {
 #define	VDEV_LABEL_END_SIZE	(2 * sizeof (vdev_label_t))
 #define	VDEV_LABELS		4
 
+/*
+ * Gang block headers are self-checksumming and contain an array
+ * of block pointers.
+ */
+#define SPA_GANGBLOCKSIZE	SPA_MINBLOCKSIZE
+#define SPA_GBH_NBLKPTRS	((SPA_GANGBLOCKSIZE - \
+	sizeof (zio_block_tail_t)) / sizeof (blkptr_t))
+#define SPA_GBH_FILLER		((SPA_GANGBLOCKSIZE - \
+	sizeof (zio_block_tail_t) - \
+	(SPA_GBH_NBLKPTRS * sizeof (blkptr_t))) /\
+	sizeof (uint64_t))
+
+typedef struct zio_gbh {
+	blkptr_t		zg_blkptr[SPA_GBH_NBLKPTRS];
+	uint64_t		zg_filler[SPA_GBH_FILLER];
+	zio_block_tail_t	zg_tail;
+} zio_gbh_phys_t;
+
 enum zio_checksum {
 	ZIO_CHECKSUM_INHERIT = 0,
 	ZIO_CHECKSUM_ON,