thr3ads.net - Btrfs devel - [PATCH 0/4] add super-recover tool to recover bad supers [Sep 2013]

If this information is useful, please help other people find it:
Share via:

Wang Shilong

2013-Sep-18 08:27 UTC

[PATCH 0/4] add super-recover tool to recover bad supers

Until now if one of device''s first superblock is corrupt,btrfs will
fail to mount. Luckily, btrfs have at least two superblocks for
every disk.

In theory, if silent corrupting happens when we are writting
superblocks into disk, we must hold at least one good superblock.

One side effect is that user must gurantee that the disk must be
a btrfs disk. Otherwise, this tool may destroy other fs.(This is
also reason why btrfs only use first superblock in every disk to mount)

This little program will scan devices and find good superblocks
(max generation, checksum matched). And use this superblock to
recover bad superblocks.The process is:

Step1: scan btrfs devices
Step2: find good supers
Step3: open ctree
Step4: writ all supers

We pass the latest good supers into open_ctree() and open_ctree()
will recalucate every superblock''s dev_item by searching chunk tree.

Notice: This patchset based on David''s dev/rescue branch

Wang Shilong (4):
  Btrfs-progs: do not run ioctls in check_mounted_where()
  Btrfs-progs: pass flag to control whether run ioctl in
    btrfs_scan_for_fsid()
  Btrfs-progs: move ask_user() to utils.c
  Btrfs-progs: add super-recover to recover bad supers

 Makefile          |   2 +-
 btrfs-find-root.c |   2 +-
 chunk-recover.c   |  20 +---
 cmds-rescue.c     |  55 ++++++++++
 commands.h        |   2 +
 disk-io.c         |  37 +++++--
 disk-io.h         |   5 +-
 super-recover.c   | 323 ++++++++++++++++++++++++++++++++++++++++++++++++++++++
 utils.c           |  20 +++-
 utils.h           |   1 +
 10 files changed, 437 insertions(+), 30 deletions(-)
 create mode 100644 super-recover.c

-- 
1.8.3.1

--
To unsubscribe from this list: send the line "unsubscribe linux-btrfs"
in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Wang Shilong

2013-Sep-18 08:27 UTC

head link

[PATCH 1/4] Btrfs-progs: do not run ioctls in check_mounted_where()

We don''t need to run ioctls when checking whether btrfs
has mounted somewhere.

Signed-off-by: Wang Shilong <wangsl.fnst@cn.fujitsu.com>
---
 utils.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/utils.c b/utils.c
index 134bf80..e3738d3 100644
--- a/utils.c
+++ b/utils.c
@@ -956,7 +956,7 @@ int check_mounted_where(int fd, const char *file, char
*where, int size,
 
 	/* scan other devices */
 	if (is_btrfs && total_devs > 1) {
-		if((ret = btrfs_scan_for_fsid(1)))
+		if((ret = btrfs_scan_for_fsid(0)))
 			return ret;
 	}
 
-- 
1.8.3.1

--
To unsubscribe from this list: send the line "unsubscribe linux-btrfs"
in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Wang Shilong

2013-Sep-18 08:27 UTC

head link

[PATCH 2/4] Btrfs-progs: pass flag to control whether run ioctl in btrfs_scan_for_fsid()

If some fatal superblocks are damaged, running ioctl will return failure,
in this case, we should avoid run ioctl.

Signed-off-by: Wang Shilong <wangsl.fnst@cn.fujitsu.com>
---
 btrfs-find-root.c | 2 +-
 chunk-recover.c   | 2 +-
 disk-io.c         | 6 +++---
 disk-io.h         | 3 ++-
 4 files changed, 7 insertions(+), 6 deletions(-)

diff --git a/btrfs-find-root.c b/btrfs-find-root.c
index b48c800..7572f52 100644
--- a/btrfs-find-root.c
+++ b/btrfs-find-root.c
@@ -82,7 +82,7 @@ static struct btrfs_root *open_ctree_broken(int fd, const char
*device)
 		return NULL;
 	}
 
-	ret = btrfs_scan_fs_devices(fd, device, &fs_devices, 0);
+	ret = btrfs_scan_fs_devices(fd, device, &fs_devices, 0, 1);
 	if (ret)
 		goto out;
 
diff --git a/chunk-recover.c b/chunk-recover.c
index 45da9a6..9af4887 100644
--- a/chunk-recover.c
+++ b/chunk-recover.c
@@ -1291,7 +1291,7 @@ static int recover_prepare(struct recover_control *rc,
char *path)
 		goto fail_free_sb;
 	}
 
-	ret = btrfs_scan_fs_devices(fd, path, &fs_devices, 0);
+	ret = btrfs_scan_fs_devices(fd, path, &fs_devices, 0, 1);
 	if (ret)
 		goto fail_free_sb;
 
diff --git a/disk-io.c b/disk-io.c
index bc673c2..6281569 100644
--- a/disk-io.c
+++ b/disk-io.c
@@ -911,7 +911,7 @@ void btrfs_cleanup_all_caches(struct btrfs_fs_info *fs_info)
 
 int btrfs_scan_fs_devices(int fd, const char *path,
 			  struct btrfs_fs_devices **fs_devices,
-			  u64 sb_bytenr)
+			  u64 sb_bytenr, int run_ioctl)
 {
 	u64 total_devs;
 	int ret;
@@ -926,7 +926,7 @@ int btrfs_scan_fs_devices(int fd, const char *path,
 	}
 
 	if (total_devs != 1) {
-		ret = btrfs_scan_for_fsid(1);
+		ret = btrfs_scan_for_fsid(run_ioctl);
 		if (ret)
 			return ret;
 	}
@@ -1005,7 +1005,7 @@ static struct btrfs_fs_info *__open_ctree_fd(int fp, const
char *path,
 	if (restore)
 		fs_info->on_restoring = 1;
 
-	ret = btrfs_scan_fs_devices(fp, path, &fs_devices, sb_bytenr);
+	ret = btrfs_scan_fs_devices(fp, path, &fs_devices, sb_bytenr, 1);
 	if (ret)
 		goto out;
 
diff --git a/disk-io.h b/disk-io.h
index b61eb43..7219c2e 100644
--- a/disk-io.h
+++ b/disk-io.h
@@ -57,7 +57,8 @@ int btrfs_setup_all_roots(struct btrfs_fs_info *fs_info,
 void btrfs_release_all_roots(struct btrfs_fs_info *fs_info);
 void btrfs_cleanup_all_caches(struct btrfs_fs_info *fs_info);
 int btrfs_scan_fs_devices(int fd, const char *path,
-			  struct btrfs_fs_devices **fs_devices, u64 sb_bytenr);
+			  struct btrfs_fs_devices **fs_devices, u64 sb_bytenr,
+			  int run_ioctl);
 int btrfs_setup_chunk_tree_and_device_map(struct btrfs_fs_info *fs_info);
 
 struct btrfs_root *open_ctree(const char *filename, u64 sb_bytenr, int writes);
-- 
1.8.3.1

--
To unsubscribe from this list: send the line "unsubscribe linux-btrfs"
in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Wang Shilong

2013-Sep-18 08:27 UTC

head link

[PATCH 3/4] Btrfs-progs: move ask_user() to utils.c

Signed-off-by: Wang Shilong <wangsl.fnst@cn.fujitsu.com>
---
 chunk-recover.c | 18 ------------------
 utils.c         | 18 ++++++++++++++++++
 utils.h         |  1 +
 3 files changed, 19 insertions(+), 18 deletions(-)

diff --git a/chunk-recover.c b/chunk-recover.c
index 9af4887..e44ca72 100644
--- a/chunk-recover.c
+++ b/chunk-recover.c
@@ -1307,24 +1307,6 @@ fail_close_fd:
 	return ret;
 }
 
-/*
- * This reads a line from the stdin and only returns non-zero if the
- * first whitespace delimited token is a case insensitive match with yes
- * or y.
- */
-static int ask_user(char *question)
-{
-	char buf[30] = {0,};
-	char *saveptr = NULL;
-	char *answer;
-
-	printf("%s [y/N]: ", question);
-
-	return fgets(buf, sizeof(buf) - 1, stdin) &&
-	       (answer = strtok_r(buf, " \t\n\r", &saveptr)) &&
-	       (!strcasecmp(answer, "yes") || !strcasecmp(answer,
"y"));
-}
-
 static int btrfs_get_device_extents(u64 chunk_object,
 				    struct list_head *orphan_devexts,
 				    struct list_head *ret_list)
diff --git a/utils.c b/utils.c
index e3738d3..2bfd59a 100644
--- a/utils.c
+++ b/utils.c
@@ -1955,3 +1955,21 @@ int is_vol_small(char *file)
 		return 0;
 	}
 }
+
+/*
+ * This reads a line from the stdin and only returns non-zero if the
+ * first whitespace delimited token is a case insensitive match with yes
+ * or y.
+ */
+int ask_user(char *question)
+{
+	char buf[30] = {0,};
+	char *saveptr = NULL;
+	char *answer;
+
+	printf("%s [y/N]: ", question);
+
+	return fgets(buf, sizeof(buf) - 1, stdin) &&
+	       (answer = strtok_r(buf, " \t\n\r", &saveptr)) &&
+	       (!strcasecmp(answer, "yes") || !strcasecmp(answer,
"y"));
+}
diff --git a/utils.h b/utils.h
index 6c2553a..2cb9d9b 100644
--- a/utils.h
+++ b/utils.h
@@ -80,4 +80,5 @@ int test_num_disk_vs_raid(u64 metadata_profile, u64
data_profile,
 int is_vol_small(char *file);
 int csum_tree_block(struct btrfs_root *root, struct extent_buffer *buf,
 			   int verify);
+int ask_user(char *question);
 #endif
-- 
1.8.3.1

--
To unsubscribe from this list: send the line "unsubscribe linux-btrfs"
in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Wang Shilong

2013-Sep-18 08:27 UTC

head link

[PATCH 4/4] Btrfs-progs: add super-recover to recover bad supers

Until now if one of device''s first superblock is corrupt,btrfs
will fail to mount. Luckily, btrfs have at least two superblocks
for every disk.

In theory, if silent corrupting happens when we are writting
superblocks into disk, we must hold at least one good superblock.

One side effect is that user must gurantee that the disk must be
a btrfs disk. Otherwise, this tool may destroy other fs.(This is also
reason why btrfs only use first superblock in every disk to mount)

This little program will try to correct bad superblocks from
good superblocks with max generation.

However, users should gurantee the disk passed is a btrfs disk,
otherwise, the tool may destroy other fs.

There will be five kinds of return values:

0: all supers are valid, no need to recover
1: usage or syntax error
2: recover all bad superblocks successfully
3: fail to recover bad superblocks
4: abort to recover bad superblocks

Signed-off-by: Wang Shilong <wangsl.fnst@cn.fujitsu.com>
---
 Makefile        |   2 +-
 cmds-rescue.c   |  55 ++++++++++
 commands.h      |   2 +
 disk-io.c       |  33 +++++-
 disk-io.h       |   2 +
 super-recover.c | 323 ++++++++++++++++++++++++++++++++++++++++++++++++++++++++
 6 files changed, 411 insertions(+), 6 deletions(-)
 create mode 100644 super-recover.c

diff --git a/Makefile b/Makefile
index bd32121..3acb3b8 100644
--- a/Makefile
+++ b/Makefile
@@ -14,7 +14,7 @@ cmds_objects = cmds-subvolume.o cmds-filesystem.o
cmds-device.o cmds-scrub.o \
 	       cmds-inspect.o cmds-balance.o cmds-send.o cmds-receive.o \
 	       cmds-quota.o cmds-qgroup.o cmds-replace.o cmds-check.o \
 	       cmds-restore.o cmds-dedup.o cmds-rescue.o \
-	       chunk-recover.o
+	       chunk-recover.o super-recover.o
 libbtrfs_objects = send-stream.o send-utils.o rbtree.o btrfs-list.o crc32c.o \
 		   uuid-tree.o
 libbtrfs_headers = send-stream.h send-utils.h send.h rbtree.h btrfs-list.h \
diff --git a/cmds-rescue.c b/cmds-rescue.c
index cfbc198..e18eb98 100644
--- a/cmds-rescue.c
+++ b/cmds-rescue.c
@@ -28,6 +28,7 @@ static const char * const rescue_cmd_group_usage[] = {
 };
 
 int btrfs_recover_chunk_tree(char *path, int verbose, int yes);
+int btrfs_recover_superblocks(char *path, int verbose, int yes);
 
 const char * const cmd_chunk_recover_usage[] = {
 	"btrfs rescue chunk-recover [options] <device>",
@@ -39,6 +40,15 @@ const char * const cmd_chunk_recover_usage[] = {
 	NULL
 };
 
+const char * const cmd_super_recover_usage[] = {
+	"btrfs rescue super-recover [options] <device>",
+	"Recover bad superblocks from good copies",
+	"",
+	"-y	Assume an answer of `yes'' to all questions",
+	"-v	Verbose mode",
+	NULL
+};
+
 int cmd_chunk_recover(int argc, char *argv[])
 {
 	int ret = 0;
@@ -87,9 +97,54 @@ int cmd_chunk_recover(int argc, char *argv[])
 	return ret;
 }
 
+/*
+ * return codes:
+ *   0 : All superblocks are valid, no need to recover
+ *   1 : Usage or syntax error
+ *   2 : Recover all bad superblocks successfully
+ *   3 : Fail to Recover bad supeblocks
+ *   4 : Abort to recover bad superblocks
+ */
+int cmd_super_recover(int argc, char **argv)
+{
+	int ret;
+	int verbose = 0;
+	int yes = 0;
+	char *dname;
+
+	while (1) {
+		int c = getopt(argc, argv, "vy");
+		if (c < 0)
+			break;
+		switch (c) {
+		case ''v'':
+			verbose = 1;
+			break;
+		case ''y'':
+			yes = 1;
+			break;
+		default:
+			usage(cmd_super_recover_usage);
+		}
+	}
+	argc = argc - optind;
+	if (argc != 1)
+		usage(cmd_super_recover_usage);
+
+	dname = argv[optind];
+	ret = check_mounted(dname);
+	if (ret) {
+		fprintf(stderr, "the device is busy\n");
+		return 1;
+	}
+	ret = btrfs_recover_superblocks(dname, verbose, yes);
+	return ret;
+}
+
 const struct cmd_group rescue_cmd_group = {
 	rescue_cmd_group_usage, NULL, {
 		{ "chunk-recover", cmd_chunk_recover, cmd_chunk_recover_usage,
NULL, 0},
+		{ "super-recover", cmd_super_recover, cmd_super_recover_usage,
NULL, 0},
 		{ 0, 0, 0, 0, 0 }
 	}
 };
diff --git a/commands.h b/commands.h
index d25b8c1..24a5bd8 100644
--- a/commands.h
+++ b/commands.h
@@ -97,6 +97,7 @@ extern const char * const cmd_send_usage[];
 extern const char * const cmd_receive_usage[];
 extern const char * const cmd_check_usage[];
 extern const char * const cmd_chunk_recover_usage[];
+extern const char * const cmd_super_recover_usage[];
 extern const char * const cmd_restore_usage[];
 extern const char * const cmd_rescue_usage[];
 
@@ -107,6 +108,7 @@ int cmd_device(int argc, char **argv);
 int cmd_scrub(int argc, char **argv);
 int cmd_check(int argc, char **argv);
 int cmd_chunk_recover(int argc, char **argv);
+int cmd_super_recover(int argc, char **argv);
 int cmd_inspect(int argc, char **argv);
 int cmd_send(int argc, char **argv);
 int cmd_receive(int argc, char **argv);
diff --git a/disk-io.c b/disk-io.c
index 6281569..de89fa3 100644
--- a/disk-io.c
+++ b/disk-io.c
@@ -982,7 +982,8 @@ int btrfs_setup_chunk_tree_and_device_map(struct
btrfs_fs_info *fs_info)
 static struct btrfs_fs_info *__open_ctree_fd(int fp, const char *path,
 					     u64 sb_bytenr,
 					     u64 root_tree_bytenr, int writes,
-					     int partial, int restore)
+					     int partial, int restore,
+					     int run_ioctl)
 {
 	struct btrfs_fs_info *fs_info;
 	struct btrfs_super_block *disk_super;
@@ -1005,7 +1006,8 @@ static struct btrfs_fs_info *__open_ctree_fd(int fp, const
char *path,
 	if (restore)
 		fs_info->on_restoring = 1;
 
-	ret = btrfs_scan_fs_devices(fp, path, &fs_devices, sb_bytenr, 1);
+	ret = btrfs_scan_fs_devices(fp, path, &fs_devices, sb_bytenr,
+				    run_ioctl);
 	if (ret)
 		goto out;
 
@@ -1078,7 +1080,7 @@ struct btrfs_fs_info *open_ctree_fs_info_restore(const
char *filename,
 		return NULL;
 	}
 	info = __open_ctree_fd(fp, filename, sb_bytenr, root_tree_bytenr,
-			       writes, partial, restore);
+			       writes, partial, restore, 1);
 	close(fp);
 	return info;
 }
@@ -1100,11 +1102,32 @@ struct btrfs_fs_info *open_ctree_fs_info(const char
*filename,
 		return NULL;
 	}
 	info = __open_ctree_fd(fp, filename, sb_bytenr, root_tree_bytenr,
-			       writes, partial, 0);
+			       writes, partial, 0, 1);
 	close(fp);
 	return info;
 }
 
+struct btrfs_root *open_ctree_with_broken_super(const char *filename,
+					u64 sb_bytenr, int writes)
+{
+	int fp;
+	struct btrfs_fs_info *info;
+	int flags = O_CREAT | O_RDWR;
+
+	if (!writes)
+		flags = O_RDONLY;
+
+	fp = open(filename, flags, 0600);
+	if (fp < 0) {
+		fprintf(stderr, "Could not open %s\n", filename);
+		return NULL;
+	}
+	info = __open_ctree_fd(fp, filename, sb_bytenr, 0,
+			       writes, 0, 0, 0);
+	close(fp);
+	return info->fs_root;
+}
+
 struct btrfs_root *open_ctree(const char *filename, u64 sb_bytenr, int writes)
 {
 	struct btrfs_fs_info *info;
@@ -1119,7 +1142,7 @@ struct btrfs_root *open_ctree_fd(int fp, const char *path,
u64 sb_bytenr,
 				 int writes)
 {
 	struct btrfs_fs_info *info;
-	info = __open_ctree_fd(fp, path, sb_bytenr, 0, writes, 0, 0);
+	info = __open_ctree_fd(fp, path, sb_bytenr, 0, writes, 0, 0, 1);
 	if (!info)
 		return NULL;
 	return info->fs_root;
diff --git a/disk-io.h b/disk-io.h
index 7219c2e..681ff79 100644
--- a/disk-io.h
+++ b/disk-io.h
@@ -70,6 +70,8 @@ struct btrfs_fs_info *open_ctree_fs_info_restore(const char
*filename,
 struct btrfs_fs_info *open_ctree_fs_info(const char *filename,
 					 u64 sb_bytenr, u64 root_tree_bytenr,
 					 int writes, int partial);
+struct btrfs_root *open_ctree_with_broken_super(const char *filename,
+					u64 sb_bytenr, int writes);
 int close_ctree(struct btrfs_root *root);
 int write_all_supers(struct btrfs_root *root);
 int write_ctree_super(struct btrfs_trans_handle *trans,
diff --git a/super-recover.c b/super-recover.c
new file mode 100644
index 0000000..8c9c5a2
--- /dev/null
+++ b/super-recover.c
@@ -0,0 +1,323 @@
+/*
+ * Copyright (C) 2013 Fujitsu.  All rights reserved.
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public
+ * License v2 as published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
+ * General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public
+ * License along with this program; if not, write to the
+ * Free Software Foundation, Inc., 59 Temple Place - Suite 330,
+ * Boston, MA 021110-1307, USA.
+ */
+
+#define _XOPEN_SOURCE 500
+#define _GNU_SOURCE 1
+
+#include <stdio.h>
+#include <stdlib.h>
+#include <fcntl.h>
+#include <sys/stat.h>
+#include <uuid/uuid.h>
+#include <errno.h>
+#include <unistd.h>
+#include <ctype.h>
+#include <getopt.h>
+
+#include "kerncompat.h"
+#include "ctree.h"
+#include "disk-io.h"
+#include "list.h"
+#include "utils.h"
+#include "crc32c.h"
+#include "volumes.h"
+#include "version.h"
+#include "transaction.h"
+
+#include "commands.h"
+
+struct btrfs_recover_superblock {
+	struct btrfs_fs_devices *fs_devices;
+
+	struct list_head good_supers;
+	struct list_head bad_supers;
+
+	u64 max_generation;
+	int recover_flag;
+};
+
+struct super_block_record {
+	struct list_head list;
+
+	char *device_name;
+	struct btrfs_super_block sb;
+
+	u64 bytenr;
+};
+
+static
+void init_recover_superblock(struct btrfs_recover_superblock *recover)
+{
+	INIT_LIST_HEAD(&recover->good_supers);
+	INIT_LIST_HEAD(&recover->bad_supers);
+
+	recover->fs_devices = NULL;
+	recover->max_generation = 0;
+	recover->recover_flag = 0;
+}
+
+static
+void free_recover_superblock(struct btrfs_recover_superblock *recover)
+{
+	struct btrfs_device *device;
+	struct super_block_record *record;
+
+	if (!recover->fs_devices)
+		return;
+
+	while (!list_empty(&recover->fs_devices->devices)) {
+		device = list_entry(recover->fs_devices->devices.next,
+				struct btrfs_device, dev_list);
+		list_del_init(&device->dev_list);
+		free(device->name);
+		free(device);
+	}
+	free(recover->fs_devices);
+
+	while (!list_empty(&recover->good_supers)) {
+		record = list_entry(recover->good_supers.next,
+				struct super_block_record, list);
+		list_del_init(&record->list);
+		free(record->device_name);
+		free(record);
+	}
+
+	while (!list_empty(&recover->bad_supers)) {
+		record = list_entry(recover->bad_supers.next,
+				struct super_block_record, list);
+		list_del_init(&record->list);
+		free(record->device_name);
+		free(record);
+	}
+}
+
+static int check_super(u64 bytenr, struct btrfs_super_block *sb)
+{
+	int csum_size = btrfs_super_csum_size(sb);
+	char result[csum_size];
+	u32 crc = ~(u32)0;
+
+	if (btrfs_super_bytenr(sb) != bytenr)
+		return 0;
+	if (sb->magic != cpu_to_le64(BTRFS_MAGIC))
+		return 0;
+
+	crc = btrfs_csum_data(NULL, (char *)sb + BTRFS_CSUM_SIZE,
+			crc, BTRFS_SUPER_INFO_SIZE - BTRFS_CSUM_SIZE);
+	btrfs_csum_final(crc, result);
+
+	return !memcmp(sb, &result, csum_size);
+}
+
+static int add_superblock_record(struct btrfs_super_block *sb, char *fname,
+			u64 bytenr, struct list_head *head)
+{
+	struct super_block_record *record;
+
+	record = malloc(sizeof(struct super_block_record));
+	if (!record)
+		return -ENOMEM;
+
+	record->device_name = strdup(fname);
+	if (!record->device_name) {
+		free(record);
+		return -ENOMEM;
+	}
+	memcpy(&record->sb, sb, sizeof(*sb));
+	record->bytenr = bytenr;
+	list_add_tail(&record->list, head);
+
+	return 0;
+}
+
+static int
+read_dev_supers(char *filename, struct btrfs_recover_superblock *recover)
+{
+	int i, ret, fd;
+	u8 buf[BTRFS_SUPER_INFO_SIZE];
+	u64 max_gen, bytenr;
+
+	struct btrfs_super_block *sb = (struct btrfs_super_block *)buf;
+
+	fd = open(filename, O_RDONLY, 0666);
+	if (fd < 0)
+		return -errno;
+
+	for (i = 0; i < BTRFS_SUPER_MIRROR_MAX; i++) {
+		bytenr = btrfs_sb_offset(i);
+		ret = pread64(fd, buf, sizeof(buf), bytenr);
+		if (ret < sizeof(buf)) {
+			close(fd);
+			return -errno;
+		}
+		ret = check_super(bytenr, sb);
+		if (ret) {
+			ret = add_superblock_record(sb, filename, bytenr,
+							&recover->good_supers);
+			if (ret)
+				return ret;
+			max_gen = btrfs_super_generation(sb);
+			if (max_gen > recover->max_generation)
+				recover->max_generation = max_gen;
+		} else {
+			ret = add_superblock_record(sb, filename, bytenr,
+						&recover->bad_supers);
+			if (ret)
+				return ret;
+		}
+	}
+	close(fd);
+	return ret;
+}
+
+static int read_fs_supers(struct btrfs_recover_superblock *recover)
+{
+	struct super_block_record *record;
+	struct super_block_record *next_record;
+	struct btrfs_device *dev;
+	int ret;
+	u64 gen;
+
+	list_for_each_entry(dev, &recover->fs_devices->devices,
+				dev_list) {
+		ret = read_dev_supers(dev->name, recover);
+		if (ret)
+			return ret;
+	}
+	list_for_each_entry_safe(record, next_record,
+			&recover->good_supers, list) {
+		gen = btrfs_super_generation(&record->sb);
+		if (gen < recover->max_generation)
+			list_move_tail(&record->list, &recover->bad_supers);
+	}
+
+	return 0;
+}
+
+static struct super_block_record *recover_get_good_super(
+				struct btrfs_recover_superblock *recover)
+{
+	struct super_block_record *record;
+	record = list_entry(recover->good_supers.next,
+				struct super_block_record, list);
+	return record;
+}
+
+static void print_all_devices(struct list_head *devices)
+{
+	struct btrfs_device *dev;
+
+	printf("All Devices:\n");
+	list_for_each_entry(dev, devices, dev_list) {
+		printf("\t");
+		printf("Device: id = %llu, name = %s\n",
+			dev->devid, dev->name);
+	}
+	printf("\n");
+}
+
+static void print_super_info(struct super_block_record *record)
+{
+	printf("\t\tdevice name = %s\n", record->device_name);
+	printf("\t\tsuperblock bytenr = %llu\n", record->bytenr);
+}
+
+static void print_all_supers(struct btrfs_recover_superblock *recover)
+{
+	struct super_block_record *record;
+
+	printf("\t[All good supers]:\n");
+	list_for_each_entry(record, &recover->good_supers, list) {
+		print_super_info(record);
+		printf("\n");
+	}
+
+	printf("\t[All bad supers]:\n");
+	list_for_each_entry(record, &recover->bad_supers, list) {
+		print_super_info(record);
+		printf("\n");
+	}
+	printf("\n");
+}
+
+int btrfs_recover_superblocks(const char *dname,
+			int verbose, int yes)
+{
+	int fd, ret;
+	struct btrfs_recover_superblock recover;
+	struct super_block_record *record;
+	struct btrfs_root *root = NULL;
+
+	fd = open(dname, O_RDONLY);
+	if (fd < 0) {
+		fprintf(stderr, "open %s error\n", dname);
+		return 1;
+	}
+	init_recover_superblock(&recover);
+
+	ret = btrfs_scan_fs_devices(fd, dname, &recover.fs_devices, 0, 0);
+	if (ret) {
+		ret = 1;
+		goto no_recover;
+	}
+
+	if (verbose)
+		print_all_devices(&recover.fs_devices->devices);
+
+	ret = read_fs_supers(&recover);
+	if (ret) {
+		ret = 1;
+		goto out;
+	}
+	if (verbose) {
+		printf("Before Recovering:\n");
+		print_all_supers(&recover);
+	}
+
+	if (list_empty(&recover.bad_supers)) {
+		printf("All supers are valid, no need to recover\n");
+		goto no_recover;
+	}
+	if (!yes) {
+		ret = ask_user("Make sure this is a btrfs disk otherwise the tool will
destroy other fs, Are you sure?");
+		if (!ret) {
+			ret = 4;
+			goto no_recover;
+		}
+	}
+	record = recover_get_good_super(&recover);
+	root = open_ctree_with_broken_super(record->device_name,
+					record->bytenr, O_RDWR);
+	BUG_ON(!root);
+
+	root->fs_info->super_bytenr = BTRFS_SUPER_INFO_OFFSET;
+	ret = write_all_supers(root);
+	if (!ret) {
+		ret = 2;
+		printf("Recovering bad superblocks successful\n");
+	} else {
+		ret = 3;
+		printf("Failed to recover bad superblocks\n");
+	}
+out:
+	close_ctree(root);
+no_recover:
+	free_recover_superblock(&recover);
+	return ret;
+}
+
-- 
1.8.3.1

--
To unsubscribe from this list: send the line "unsubscribe linux-btrfs"
in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

David Sterba

2013-Sep-18 15:13 UTC

head link

Re: [PATCH 4/4] Btrfs-progs: add super-recover to recover bad supers

On Wed, Sep 18, 2013 at 04:27:36PM +0800, Wang Shilong
wrote:> --- a/disk-io.c
> +++ b/disk-io.c
> @@ -982,7 +982,8 @@ int btrfs_setup_chunk_tree_and_device_map(struct
btrfs_fs_info *fs_info)
>  static struct btrfs_fs_info *__open_ctree_fd(int fp, const char *path,
>  					     u64 sb_bytenr,
>  					     u64 root_tree_bytenr, int writes,
> -					     int partial, int restore)
> +					     int partial, int restore,
> +					     int run_ioctl)
Yet another parameter to __open_ctree_fd, this really needs a cleanup.
(Later is fine.)
> +	root = open_ctree_with_broken_super(record->device_name,
> +					record->bytenr, O_RDWR);
The last parameter is ''writes'', I think it''s 0/1, not
the open() flags.
> +	BUG_ON(!root);
Is it a fatal error to BUG_ON here? Please add proper error handling.

thanks,
david
--
To unsubscribe from this list: send the line "unsubscribe linux-btrfs"
in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

David Sterba

2013-Sep-18 15:18 UTC

head link

Re: [PATCH 0/4] add super-recover tool to recover bad supers

On Wed, Sep 18, 2013 at 04:27:32PM +0800, Wang Shilong
wrote:> Until now if one of device''s first superblock is corrupt,btrfs
will
> fail to mount. Luckily, btrfs have at least two superblocks for
> every disk.
> 
> This little program will scan devices and find good superblocks
> (max generation, checksum matched). And use this superblock to
> recover bad superblocks.The process is:
> 
> Step1: scan btrfs devices
> Step2: find good supers
> Step3: open ctree
> Step4: writ all supers
Looks great! I''ll pull that into integration even with the BUG_ON in
the
4/4 patch and will replace it later.

david
--
To unsubscribe from this list: send the line "unsubscribe linux-btrfs"
in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Wang Shilong

2013-Sep-18 15:31 UTC

head link

Re: [PATCH 4/4] Btrfs-progs: add super-recover to recover bad supers

Hello David,
> On Wed, Sep 18, 2013 at 04:27:36PM +0800, Wang Shilong wrote:
>> --- a/disk-io.c
>> +++ b/disk-io.c
>> @@ -982,7 +982,8 @@ int btrfs_setup_chunk_tree_and_device_map(struct
btrfs_fs_info *fs_info)
>> static struct btrfs_fs_info *__open_ctree_fd(int fp, const char *path,
>> 					     u64 sb_bytenr,
>> 					     u64 root_tree_bytenr, int writes,
>> -					     int partial, int restore)
>> +					     int partial, int restore,
>> +					     int run_ioctl)
> 
> Yet another parameter to __open_ctree_fd, this really needs a cleanup.
> (Later is fine.)
I have considered about this, __open_ctree_fd() is a *static* helper, if i
don''t
add another parameter to it. I have to move out of __open_ctree_fd()  into
open_ctree_with_broken_super() which also seems  not good!

> 
>> +	root = open_ctree_with_broken_super(record->device_name,
>> +					record->bytenr, O_RDWR);
> 
> The last parameter is ''writes'', I think it''s
0/1, not the open() flags.
Yeah,  will fix it.
> 
>> +	BUG_ON(!root);
> 
> Is it a fatal error to BUG_ON here? Please add proper error handling.
Yes, BUG_ON() is not really friendly, let me see if i can handle it more
gracefully.

Thanks,
Wang> 
> thanks,
> david
> --
> To unsubscribe from this list: send the line "unsubscribe
linux-btrfs" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html
--
To unsubscribe from this list: send the line "unsubscribe linux-btrfs"
in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

David Sterba

2013-Sep-18 16:31 UTC

head link

Re: [PATCH 4/4] Btrfs-progs: add super-recover to recover bad supers

On Wed, Sep 18, 2013 at 11:31:16PM +0800, Wang Shilong
wrote:> >> static struct btrfs_fs_info *__open_ctree_fd(int fp, const char
*path,
> >> 					     u64 sb_bytenr,
> >> 					     u64 root_tree_bytenr, int writes,
> >> -					     int partial, int restore)
> >> +					     int partial, int restore,
> >> +					     int run_ioctl)
> > 
> > Yet another parameter to __open_ctree_fd, this really needs a cleanup.
> > (Later is fine.)
> 
> I have considered about this, __open_ctree_fd() is a *static* helper, if i
don''t
> add another parameter to it. I have to move out of __open_ctree_fd()  into
> open_ctree_with_broken_super() which also seems  not good!
The idea is that all open_ctree* variants will take one parameter that
accumulates what''s currently passed via
writes/partial/restore/run_ioctl:

#define OPEN_CTREE_PARTIAL 	(1 << 0ULL)
#define OPEN_CTREE_WRITES	(1 << 1ULL)
etc ...

static struct btrfs_fs_info *__open_ctree_fd(int fp, const char *path,
					     u64 sb_bytenr,
					     u64 root_tree_bytenr,
					     unsigned open_flags)

and the callers updated accordingly. The calls then look like

- open_ctree_fs_info_restore(target, 0, 0, 0, 1);
+ open_ctree_fs_info_restore(target, 0, 0, OPEN_CTREE_PARTIAL);
--
To unsubscribe from this list: send the line "unsubscribe linux-btrfs"
in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Wang Shilong

2013-Sep-21 01:16 UTC

head link

Re: [PATCH 4/4] Btrfs-progs: add super-recover to recover bad supers

Hello David,
> On Wed, Sep 18, 2013 at 11:31:16PM +0800, Wang Shilong wrote:
>>>> static struct btrfs_fs_info *__open_ctree_fd(int fp, const char
*path,
>>>> 					     u64 sb_bytenr,
>>>> 					     u64 root_tree_bytenr, int writes,
>>>> -					     int partial, int restore)
>>>> +					     int partial, int restore,
>>>> +					     int run_ioctl)
>>> 
>>> Yet another parameter to __open_ctree_fd, this really needs a
cleanup.
>>> (Later is fine.)
>> 
>> I have considered about this, __open_ctree_fd() is a *static* helper,
if i don''t
>> add another parameter to it. I have to move out of __open_ctree_fd() 
into
>> open_ctree_with_broken_super() which also seems  not good!
> 
> The idea is that all open_ctree* variants will take one parameter that
> accumulates what''s currently passed via
writes/partial/restore/run_ioctl:
> 
> #define OPEN_CTREE_PARTIAL 	(1 << 0ULL)
> #define OPEN_CTREE_WRITES	(1 << 1ULL)
> etc ...
> 
> static struct btrfs_fs_info *__open_ctree_fd(int fp, const char *path,
> 					     u64 sb_bytenr,
> 					     u64 root_tree_bytenr,
> 					     unsigned open_flags)
> 
> and the callers updated accordingly. The calls then look like
> 
> - open_ctree_fs_info_restore(target, 0, 0, 0, 1);
> + open_ctree_fs_info_restore(target, 0, 0, OPEN_CTREE_PARTIAL);
Good idea, i think this should be another patch. 

Thanks,
Wang

--
To unsubscribe from this list: send the line "unsubscribe linux-btrfs"
in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

David Sterba

2013-Sep-23 12:10 UTC

head link

[offlist] Re: [PATCH 4/4] Btrfs-progs: add super-recover to recover bad supers

On Sat, Sep 21, 2013 at 09:16:25AM +0800, Wang Shilong
wrote:> > - open_ctree_fs_info_restore(target, 0, 0, 0, 1);
> > + open_ctree_fs_info_restore(target, 0, 0, OPEN_CTREE_PARTIAL);
> 
> Good idea, i think this should be another patch. 
yes
--
To unsubscribe from this list: send the line "unsubscribe linux-btrfs"
in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Btrfs devel - Sep 2013 - [PATCH 0/4] add super-recover tool to recover bad supers

[PATCH 0/4] add super-recover tool to recover bad supers

[PATCH 1/4] Btrfs-progs: do not run ioctls in check_mounted_where()

[PATCH 2/4] Btrfs-progs: pass flag to control whether run ioctl in btrfs_scan_for_fsid()

[PATCH 3/4] Btrfs-progs: move ask_user() to utils.c

[PATCH 4/4] Btrfs-progs: add super-recover to recover bad supers

Re: [PATCH 4/4] Btrfs-progs: add super-recover to recover bad supers

Re: [PATCH 0/4] add super-recover tool to recover bad supers

Re: [PATCH 4/4] Btrfs-progs: add super-recover to recover bad supers

Re: [PATCH 4/4] Btrfs-progs: add super-recover to recover bad supers

Re: [PATCH 4/4] Btrfs-progs: add super-recover to recover bad supers

[offlist] Re: [PATCH 4/4] Btrfs-progs: add super-recover to recover bad supers