linux-btrfs.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: Anand Jain <anand.jain@oracle.com>
To: linux-btrfs@vger.kernel.org
Subject: [PATCH v2 6/7] btrfs: verify checksum for all devices in mount context
Date: Fri, 30 Mar 2018 08:09:22 +0800	[thread overview]
Message-ID: <20180330000924.11148-7-anand.jain@oracle.com> (raw)
In-Reply-To: <20180330000924.11148-1-anand.jain@oracle.com>

During mount context, we aren't verifying the superblock checksum
for all the devices, instead, we verify it only for the
struct btrfs_fs_device::latest_bdev. This patch fixes it by moving
the checksum verification code from the function open_ctree() into
the function btrfs_read_dev_one_super().

By doing this now we are verifying the superblock checksum in the
mount-context, device-replace and, device-delete context. The
device-replace and device-delete call-chain is as show below after
this patch.

delete-device/replace:
btrfs_rm_device() || btrfs_dev_replace_by_ioctl()
|_btrfs_find_device_by_devspec()
  |_btrfs_find_device_missing_or_by_path()
    |_btrfs_find_device_by_path()
      |_btrfs_get_bdev_and_sb()
        |_btrfs_read_dev_super()
          |_btrfs_read_dev_one_super()
	    |_btrfs_check_super_csum()

Test case:
 Before:
   mkfs.btrfs -fq -draid1 -mraid1 /dev/sdb /dev/sdc
   dd if=/dev/urandom of=/dev/sdc ibs=1 obs=1 count=1 seek=64K
   mount /dev/sdb /btrfs <-- success as kernel does not check csum
			     for non-btrfs_fs_devices::latest_bdev.
 After:
   mkfs.btrfs -fq -draid1 -mraid1 /dev/sdb /dev/sdc
   dd if=/dev/urandom of=/dev/sdc ibs=1 obs=1 count=1 seek=64K
   mount /dev/sdb /btrfs
     mount: mount /dev/sdc on /btrfs failed: Structure needs cleaning
   mount -o degraded /dev/sdc /btrfs
     mount: mount /dev/sdc on /btrfs failed: Structure needs cleaning

 So the current recovery step is to fix the primary superblock, by
 using the btrfs-progs cli.

Signed-off-by: Anand Jain <anand.jain@oracle.com>
Reviewed-by: Nikolay Borisov <nborisov@suse.com>
---
v1->v2:
 git commit log update. With the call-chain (which I believe will
  go away in the long term, we need the uuid from the userland not
  the disk-path). And add test case.
 Check err < 0 explicitly and drop check for "else if (err == -EUCLEAN)"
  v1:
	err = btrfs_check_super_csum(bh->b_data);
	if (err) {
		if (err == -EINVAL)
			pr_err("BTRFS error (device %pg): unsupported checksum algorithm",
				bdev);
		else if (err == -EUCLEAN)
			pr_err("BTRFS error (device %pg): superblock checksum mismatch",
				bdev);
		brelse(bh);
		return err;
	}

 fs/btrfs/disk-io.c | 33 +++++++++++++++++----------------
 1 file changed, 17 insertions(+), 16 deletions(-)

diff --git a/fs/btrfs/disk-io.c b/fs/btrfs/disk-io.c
index 37b0cf489fdd..1eabcc4bf8d2 100644
--- a/fs/btrfs/disk-io.c
+++ b/fs/btrfs/disk-io.c
@@ -2565,22 +2565,6 @@ int open_ctree(struct super_block *sb,
 	}
 
 	/*
-	 * We want to check superblock checksum, the type is stored inside.
-	 * Pass the whole disk block of size BTRFS_SUPER_INFO_SIZE (4k).
-	 */
-	err = btrfs_check_super_csum(bh->b_data);
-	if (err) {
-		if (err == -EINVAL)
-			pr_err("BTRFS error (device %pg): unsupported checksum algorithm",
-				fs_devices->latest_bdev);
-		else if (err == -EUCLEAN)
-			pr_err("BTRFS error (device %pg): superblock checksum mismatch",
-				fs_devices->latest_bdev);
-		brelse(bh);
-		goto fail_alloc;
-	}
-
-	/*
 	 * super_copy is zeroed at allocation time and we never touch the
 	 * following bytes up to INFO_SIZE, the checksum is calculated from
 	 * the whole block of INFO_SIZE
@@ -3126,6 +3110,7 @@ int btrfs_read_dev_one_super(struct block_device *bdev, int copy_num,
 	struct buffer_head *bh;
 	struct btrfs_super_block *super;
 	u64 bytenr;
+	int err;
 
 	bytenr = btrfs_sb_offset(copy_num);
 	if (bytenr + BTRFS_SUPER_INFO_SIZE >= i_size_read(bdev->bd_inode))
@@ -3146,6 +3131,22 @@ int btrfs_read_dev_one_super(struct block_device *bdev, int copy_num,
 		return -EINVAL;
 	}
 
+	/*
+	 * Check the superblock checksum, the type is stored inside.
+	 * Pass the whole disk block of size BTRFS_SUPER_INFO_SIZE (4k).
+	 */
+	err = btrfs_check_super_csum(bh->b_data);
+	if (err < 0) {
+		if (err == -EINVAL)
+			pr_err("BTRFS error (device %pg): unsupported checksum algorithm",
+				bdev);
+		else if (err == -EUCLEAN)
+			pr_err("BTRFS error (device %pg): superblock checksum mismatch",
+				bdev);
+		brelse(bh);
+		return err;
+	}
+
 	*bh_ret = bh;
 	return 0;
 }
-- 
2.7.0


  parent reply	other threads:[~2018-03-30  0:07 UTC|newest]

Thread overview: 11+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2018-03-30  0:09 [PATCH v4 0/7] Superblock read and verify cleanups Anand Jain
2018-03-30  0:09 ` [PATCH 1/7] btrfs: cleanup btrfs_check_super_csum() for better code flow Anand Jain
2018-03-30  0:09 ` [PATCH v4 2/7] btrfs: return required error from btrfs_check_super_csum Anand Jain
2018-03-30  0:09 ` [PATCH v2 3/7] btrfs: cleanup btrfs_read_disk_super() to return std error Anand Jain
2018-03-30  0:09 ` [PATCH v2 4/7] btrfs: check if the fsid in the primary sb and copy sb are same Anand Jain
2018-03-30  0:09 ` [PATCH v4 5/7] btrfs: verify superblock checksum during scan Anand Jain
2018-03-30  0:09 ` Anand Jain [this message]
2018-03-30  0:09 ` [PATCH 7/7] btrfs: drop the redundant invalidate_bdev() Anand Jain
2018-04-10 14:48 ` [PATCH v4 0/7] Superblock read and verify cleanups David Sterba
2018-04-12  6:43   ` Anand Jain
  -- strict thread matches above, loose matches on Subject: below --
2018-03-29 10:36 [PATCH v3 0/9] " Anand Jain
2018-03-29 10:37 ` [PATCH v2 6/7] btrfs: verify checksum for all devices in mount context Anand Jain

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=20180330000924.11148-7-anand.jain@oracle.com \
    --to=anand.jain@oracle.com \
    --cc=linux-btrfs@vger.kernel.org \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).