Merge branch 'upstream-0.6.2'

META

@@ -1,7 +1,7 @@
 Meta:         1
 Name:         zfs
 Branch:       1.0
-Version:      0.6.1
+Version:      0.6.2
 Release:      1
 Release-Tags: relext
 License:      CDDL

cmd/zfs/zfs_main.c

@@ -66,6 +66,7 @@
 #include "zfs_iter.h"
 #include "zfs_util.h"
 #include "zfs_comutil.h"
+#include "libzfs_impl.h"
 
 libzfs_handle_t *g_zfs;
 
@@ -2288,10 +2289,8 @@ userspace_cb(void *arg, const char *domain, uid_t rid, uint64_t space)
 			if (!cb->cb_sid2posix) {
 				e = directory_name_from_sid(NULL, sid, &name,
 				    &classes);
-				if (e != NULL) {
+				if (e != NULL)
 					directory_error_free(e);
-					return (1);
-				}
 				if (name == NULL)
 					name = sid;
 			}
@@ -2543,7 +2542,7 @@ zfs_do_userspace(int argc, char **argv)
 	boolean_t prtnum = B_FALSE;
 	boolean_t parsable = B_FALSE;
 	boolean_t sid2posix = B_FALSE;
-	int error = 0;
+	int ret = 0;
 	int c;
 	zfs_sort_column_t *sortcol = NULL;
 	int types = USTYPE_PSX_USR | USTYPE_SMB_USR;
@@ -2688,18 +2687,19 @@ zfs_do_userspace(int argc, char **argv)
 		    !(types & (USTYPE_PSX_GRP | USTYPE_SMB_GRP))))
 			continue;
 		cb.cb_prop = p;
-		error = zfs_userspace(zhp, p, userspace_cb, &cb);
-		if (error)
-			break;
+		if ((ret = zfs_userspace(zhp, p, userspace_cb, &cb)) != 0)
+			return (ret);
 	}
 
 	/* Sort the list */
+	if ((node = uu_avl_first(avl_tree)) == NULL)
+		return (0);
+
 	us_populated = B_TRUE;
+
 	listpool = uu_list_pool_create("tmplist", sizeof (us_node_t),
 	    offsetof(us_node_t, usn_listnode), NULL, UU_DEFAULT);
 	list = uu_list_create(listpool, NULL, UU_DEFAULT);
-
-	node = uu_avl_first(avl_tree);
 	uu_list_node_init(node, &node->usn_listnode, listpool);
 
 	while (node != NULL) {
@@ -2740,7 +2740,7 @@ zfs_do_userspace(int argc, char **argv)
 	uu_avl_destroy(avl_tree);
 	uu_avl_pool_destroy(avl_pool);
 
-	return (error);
+	return (ret);
 }
 
 /*
@@ -6342,12 +6342,6 @@ main(int argc, char **argv)
 
 	opterr = 0;
 
-	if ((mnttab_file = fopen(MNTTAB, "r")) == NULL) {
-		(void) fprintf(stderr, gettext("internal error: unable to "
-		    "open %s\n"), MNTTAB);
-		return (1);
-	}
-
 	/*
 	 * Make sure the user has specified some command.
 	 */
@@ -6386,6 +6380,8 @@ main(int argc, char **argv)
 	if ((g_zfs = libzfs_init()) == NULL)
 		return (1);
 
+	mnttab_file = g_zfs->libzfs_mnttab;
+
 	zpool_set_history_str("zfs", argc, argv, history_str);
 	verify(zpool_stage_history(g_zfs, history_str) == 0);
 
@@ -6410,8 +6406,6 @@ main(int argc, char **argv)
 	}
 	libzfs_fini(g_zfs);
 
-	(void) fclose(mnttab_file);
-
 	/*
 	 * The 'ZFS_ABORT' environment variable causes us to dump core on exit
 	 * for the purposes of running ::findleaks.

cmd/zinject/translate.c

@@ -115,7 +115,11 @@ parse_pathname(const char *inpath, char *dataset, char *relpath,
 		return (-1);
 	}
 
+#ifdef HAVE_SETMNTENT
+	if ((fp = setmntent(MNTTAB, "r")) == NULL) {
+#else
 	if ((fp = fopen(MNTTAB, "r")) == NULL) {
+#endif
 		(void) fprintf(stderr, "cannot open /etc/mtab\n");
 		return (-1);
 	}

cmd/zpool/zpool_main.c

@@ -241,7 +241,7 @@ get_usage(zpool_help_t idx) {
 	case HELP_LABELCLEAR:
 		return (gettext("\tlabelclear [-f] <vdev>\n"));
 	case HELP_LIST:
-		return (gettext("\tlist [-H] [-o property[,...]] "
+		return (gettext("\tlist [-Hv] [-o property[,...]] "
 		    "[-T d|u] [pool] ... [interval [count]]\n"));
 	case HELP_OFFLINE:
 		return (gettext("\toffline [-t] <pool> <device> ...\n"));
@@ -3060,7 +3060,7 @@ int
 zpool_do_list(int argc, char **argv)
 {
 	int c;
-	int ret;
+	int ret = 0;
 	list_cbdata_t cb = { 0 };
 	static char default_props[] =
 	    "name,size,allocated,free,capacity,dedupratio,"

cmd/zpool/zpool_vdev.c

@@ -67,6 +67,8 @@
 #include <libintl.h>
 #include <libnvpair.h>
 #include <limits.h>
+//#include <scsi/scsi.h>
+//#include <scsi/sg.h>
 #include <stdio.h>
 #include <string.h>
 #include <unistd.h>
@@ -83,6 +85,7 @@
 #endif /* HAVE_LIBBLKID */
 
 #include "zpool_util.h"
+#include <sys/zfs_context.h>
 
 /*
  * For any given vdev specification, we can have multiple errors.  The
@@ -92,6 +95,142 @@
 boolean_t error_seen;
 boolean_t is_force;
 
+typedef struct vdev_disk_db_entry
+{
+	char id[24];
+	int sector_size;
+} vdev_disk_db_entry_t;
+
+/*
+ * Database of block devices that lie about physical sector sizes.  The
+ * identification string must be precisely 24 characters to avoid false
+ * negatives
+ */
+static vdev_disk_db_entry_t vdev_disk_database[] = {
+	{"ATA     Corsair Force 3 ", 8192},
+	{"ATA     INTEL SSDSA2CT04", 8192},
+	{"ATA     INTEL SSDSA2CW16", 8192},
+	{"ATA     INTEL SSDSC2CT18", 8192},
+	{"ATA     INTEL SSDSC2CW12", 8192},
+	{"ATA     KINGSTON SH100S3", 8192},
+	{"ATA     M4-CT064M4SSD2  ", 8192},
+	{"ATA     M4-CT128M4SSD2  ", 8192},
+	{"ATA     M4-CT256M4SSD2  ", 8192},
+	{"ATA     M4-CT512M4SSD2  ", 8192},
+	{"ATA     OCZ-AGILITY2    ", 8192},
+	{"ATA     OCZ-VERTEX2 3.5 ", 8192},
+	{"ATA     OCZ-VERTEX3     ", 8192},
+	{"ATA     OCZ-VERTEX3 LT  ", 8192},
+	{"ATA     OCZ-VERTEX3 MI  ", 8192},
+	{"ATA     SAMSUNG SSD 830 ", 8192},
+	{"ATA     Samsung SSD 840 ", 8192},
+	{"ATA     INTEL SSDSA2M040", 4096},
+	{"ATA     INTEL SSDSA2M080", 4096},
+	{"ATA     INTEL SSDSA2M160", 4096},
+	/* Imported from Open Solaris*/
+	{"ATA     MARVELL SD88SA02", 4096},
+	/* Advanced format Hard drives */
+	{"ATA     Hitachi HDS5C303", 4096},
+	{"ATA     SAMSUNG HD204UI ", 4096},
+	{"ATA     ST2000DL004 HD20", 4096},
+	{"ATA     WDC WD10EARS-00M", 4096},
+	{"ATA     WDC WD10EARS-00S", 4096},
+	{"ATA     WDC WD10EARS-00Z", 4096},
+	{"ATA     WDC WD15EARS-00M", 4096},
+	{"ATA     WDC WD15EARS-00S", 4096},
+	{"ATA     WDC WD15EARS-00Z", 4096},
+	{"ATA     WDC WD20EARS-00M", 4096},
+	{"ATA     WDC WD20EARS-00S", 4096},
+	{"ATA     WDC WD20EARS-00Z", 4096},
+	/* Virtual disks: Assume zvols with default volblocksize */
+#if 0
+	{"ATA     QEMU HARDDISK   ", 8192},
+	{"IET     VIRTUAL-DISK    ", 8192},
+	{"OI      COMSTAR         ", 8192},
+#endif
+};
+
+static const int vdev_disk_database_size =
+	sizeof (vdev_disk_database) / sizeof (vdev_disk_database[0]);
+
+#define	INQ_REPLY_LEN	96
+#define	INQ_CMD_LEN	6
+
+#ifdef __LINUX__
+static boolean_t
+check_sector_size_database(char *path, int *sector_size)
+{
+	unsigned char inq_buff[INQ_REPLY_LEN];
+	unsigned char sense_buffer[32];
+	unsigned char inq_cmd_blk[INQ_CMD_LEN] =
+	    {INQUIRY, 0, 0, 0, INQ_REPLY_LEN, 0};
+	sg_io_hdr_t io_hdr;
+	int error;
+	int fd;
+	int i;
+
+	/* Prepare INQUIRY command */
+	memset(&io_hdr, 0, sizeof(sg_io_hdr_t));
+	io_hdr.interface_id = 'S';
+	io_hdr.cmd_len = sizeof(inq_cmd_blk);
+	io_hdr.mx_sb_len = sizeof(sense_buffer);
+	io_hdr.dxfer_direction = SG_DXFER_FROM_DEV;
+	io_hdr.dxfer_len = INQ_REPLY_LEN;
+	io_hdr.dxferp = inq_buff;
+	io_hdr.cmdp = inq_cmd_blk;
+	io_hdr.sbp = sense_buffer;
+	io_hdr.timeout = 10;        /* 10 milliseconds is ample time */
+
+	if ((fd = open(path, O_RDONLY|O_DIRECT)) < 0)
+		return (B_FALSE);
+
+	error = ioctl(fd, SG_IO, (unsigned long) &io_hdr);
+
+	(void) close(fd);
+
+	if (error < 0)
+		return (B_FALSE);
+
+	if ((io_hdr.info & SG_INFO_OK_MASK) != SG_INFO_OK)
+		return (B_FALSE);
+
+	for (i = 0; i < vdev_disk_database_size; i++) {
+		if (memcmp(inq_buff + 8, vdev_disk_database[i].id, 24))
+			continue;
+
+		*sector_size = vdev_disk_database[i].sector_size;
+		return (B_TRUE);
+	}
+
+	return (B_FALSE);
+}
+#endif
+
+
+#ifdef __APPLE__
+// FIXME
+static boolean_t
+check_sector_size_database(char *path, int *sector_size)
+{
+    char diskname[256] = {0};
+    int i, error = 0;
+
+	if (error < 0)
+		return (B_FALSE);
+
+	for (i = 0; i < vdev_disk_database_size; i++) {
+		if (memcmp(diskname, vdev_disk_database[i].id, 24))
+			continue;
+
+		*sector_size = vdev_disk_database[i].sector_size;
+		return (B_TRUE);
+	}
+
+	return (B_FALSE);
+}
+#endif
+
+
 /*PRINTFLIKE1*/
 static void
 vdev_error(const char *fmt, ...)
@@ -460,6 +599,7 @@ make_leaf_vdev(nvlist_t *props, const char *arg, uint64_t is_log)
 	nvlist_t *vdev = NULL;
 	char *type = NULL;
 	boolean_t wholedisk = B_FALSE;
+	uint64_t ashift = 0;
 	int err;
 
 	/*
@@ -545,19 +685,31 @@ make_leaf_vdev(nvlist_t *props, const char *arg, uint64_t is_log)
 		verify(nvlist_add_uint64(vdev, ZPOOL_CONFIG_WHOLE_DISK,
 		    (uint64_t)wholedisk) == 0);
 
+	/*
+	 * Override defaults if custom properties are provided.
+	 */
 	if (props != NULL) {
-		uint64_t ashift = 0;
 		char *value = NULL;
 
 		if (nvlist_lookup_string(props,
 		    zpool_prop_to_name(ZPOOL_PROP_ASHIFT), &value) == 0)
 			zfs_nicestrtonum(NULL, value, &ashift);
+	}
+
+	/*
+	 * If the device is known to incorrectly report its physical sector
+	 * size explicitly provide the known correct value.
+	 */
+	if (ashift == 0) {
+		int sector_size;
 
-		if (ashift > 0)
-			verify(nvlist_add_uint64(vdev, ZPOOL_CONFIG_ASHIFT,
-			    ashift) == 0);
+		if (check_sector_size_database(path, &sector_size) == B_TRUE)
+			ashift = highbit(sector_size) - 1;
 	}
 
+	if (ashift > 0)
+		nvlist_add_uint64(vdev, ZPOOL_CONFIG_ASHIFT, ashift);
+
 	return (vdev);
 }
 

config/kernel-shrink.m4

@@ -28,6 +28,45 @@ AC_DEFUN([ZFS_AC_KERNEL_SHRINK], [
 	])
 ])
 
+dnl #
+dnl # 3.3 API change
+dnl # The super_block structure was changed to use an hlist_node instead
+dnl # of a list_head for the .s_instance linkage.
+dnl #
+dnl # This was done in part to resolve a race in the iterate_supers_type()
+dnl # function which was introduced in Linux 3.0 kernel.  The iterator
+dnl # was supposed to provide a safe way to call an arbitrary function on
+dnl # all super blocks of a specific type.  Unfortunately, because a
+dnl # list_head was used it was possible for iterate_supers_type() to
+dnl # get stuck spinning a super block which was just deactivated.
+dnl #
+dnl # This can occur because when the list head is removed from the
+dnl # fs_supers list it is reinitialized to point to itself.  If the
+dnl # iterate_supers_type() function happened to be processing the
+dnl # removed list_head it will get stuck spinning on that list_head.
+dnl #
+dnl # To resolve the issue for existing 3.0 - 3.2 kernels we detect when
+dnl # a list_head is used.  Then to prevent the spinning from occurring
+dnl # the .next pointer is set to the fs_supers list_head which ensures
+dnl # the iterate_supers_type() function will always terminate.
+dnl #
+AC_DEFUN([ZFS_AC_KERNEL_S_INSTANCES_LIST_HEAD], [
+	AC_MSG_CHECKING([whether super_block has s_instances list_head])
+	ZFS_LINUX_TRY_COMPILE([
+		#include <linux/fs.h>
+	],[
+		struct super_block sb __attribute__ ((unused));
+
+		INIT_LIST_HEAD(&sb.s_instances);
+	],[
+		AC_MSG_RESULT(yes)
+		AC_DEFINE(HAVE_S_INSTANCES_LIST_HEAD, 1,
+		    [struct super_block has s_instances list_head])
+	],[
+		AC_MSG_RESULT(no)
+	])
+])
+
 AC_DEFUN([ZFS_AC_KERNEL_NR_CACHED_OBJECTS], [
 	AC_MSG_CHECKING([whether sops->nr_cached_objects() exists])
 	ZFS_LINUX_TRY_COMPILE([