diff mbox

[v2] ext3: Add journal error check into ext3_rename()

Message ID AANLkTi=Ve03g026yb_x1VMDcuq4nrppuRcovPvDwHEc8@mail.gmail.com
State Not Applicable, archived
Headers show

Commit Message

Amir Goldstein Nov. 23, 2010, 7:58 a.m. UTC
On Tue, Nov 23, 2010 at 6:30 AM, Namhyung Kim <namhyung@gmail.com> wrote:
> Check return value of ext3_journal_get_write_access() and
> ext3_journal_dirty_metadata().
>
> Signed-off-by: Namhyung Kim <namhyung@gmail.com>
> ---
>  fs/ext3/namei.c |   19 +++++++++++++++----
>  1 files changed, 15 insertions(+), 4 deletions(-)
>
> diff --git a/fs/ext3/namei.c b/fs/ext3/namei.c
> index 672cea1..c8415e7 100644
> --- a/fs/ext3/namei.c
> +++ b/fs/ext3/namei.c
> @@ -2371,7 +2371,9 @@ static int ext3_rename (struct inode * old_dir, struct dentry *old_dentry,
>                        goto end_rename;
>        } else {
>                BUFFER_TRACE(new_bh, "get write access");
> -               ext3_journal_get_write_access(handle, new_bh);
> +               retval = ext3_journal_get_write_access(handle, new_bh);
> +               if (retval)
> +                       goto journal_error;
>                new_de->inode = cpu_to_le32(old_inode->i_ino);
>                if (EXT3_HAS_INCOMPAT_FEATURE(new_dir->i_sb,
>                                              EXT3_FEATURE_INCOMPAT_FILETYPE))
> @@ -2380,7 +2382,9 @@ static int ext3_rename (struct inode * old_dir, struct dentry *old_dentry,
>                new_dir->i_ctime = new_dir->i_mtime = CURRENT_TIME_SEC;
>                ext3_mark_inode_dirty(handle, new_dir);
>                BUFFER_TRACE(new_bh, "call ext3_journal_dirty_metadata");
> -               ext3_journal_dirty_metadata(handle, new_bh);
> +               retval = ext3_journal_dirty_metadata(handle, new_bh);
> +               if (retval)
> +                       goto journal_error;
>                brelse(new_bh);
>                new_bh = NULL;
>        }
> @@ -2429,10 +2433,17 @@ static int ext3_rename (struct inode * old_dir, struct dentry *old_dentry,
>        ext3_update_dx_flag(old_dir);
>        if (dir_bh) {
>                BUFFER_TRACE(dir_bh, "get_write_access");
> -               ext3_journal_get_write_access(handle, dir_bh);
> +               retval = ext3_journal_get_write_access(handle, dir_bh);
> +               if (retval)
> +                       goto journal_error;
>                PARENT_INO(dir_bh->b_data) = cpu_to_le32(new_dir->i_ino);
>                BUFFER_TRACE(dir_bh, "call ext3_journal_dirty_metadata");
> -               ext3_journal_dirty_metadata(handle, dir_bh);
> +               retval = ext3_journal_dirty_metadata(handle, dir_bh);
> +               if (retval) {
> +journal_error:
> +                       ext3_std_error(new_dir->i_sb, retval);
> +                       goto end_rename;
> +               }


Hi Kim,

A better way to handle this situation is to get_write_access much sooner,
when things can still be rolled back properly.
Please find below a patch I am using in next3 to fix some un-handled
journal errors
and see if you can/want to use any of it as reference for your patches.

Amir.


Some places in Ext3 original code don't check for return value of
JBD functions.  Check for snapshot/journal errors in those places.

Signed-off-by: Amir Goldstein <amir73il@users.sf.net>

--------------------------------------------------------------------------------

@@ -2337,6 +2350,10 @@ static int ext3_rename (struct inode *
        if (!new_inode && new_dir!=old_dir &&
                new_dir->i_nlink >= EXT3_LINK_MAX)
            goto end_rename;
+       BUFFER_TRACE(dir_bh, "get_write_access");
+       retval = ext3_journal_get_write_access(handle, dir_bh);
+       if (retval)
+           goto end_rename;
    }
    if (!new_bh) {
        retval = ext3_add_entry (handle, new_dentry, old_inode);
@@ -2344,7 +2361,9 @@ static int ext3_rename (struct inode *
            goto end_rename;
    } else {
        BUFFER_TRACE(new_bh, "get write access");
-       ext3_journal_get_write_access(handle, new_bh);
+       retval = ext3_journal_get_write_access(handle, new_bh);
+       if (retval)
+           goto end_rename;
        new_de->inode = cpu_to_le32(old_inode->i_ino);
        if (EXT3_HAS_INCOMPAT_FEATURE(new_dir->i_sb,
                          EXT3_FEATURE_INCOMPAT_FILETYPE))
@@ -2401,8 +2420,6 @@ static int ext3_rename (struct inode *
    old_dir->i_ctime = old_dir->i_mtime = CURRENT_TIME_SEC;
    ext3_update_dx_flag(old_dir);
    if (dir_bh) {
-       BUFFER_TRACE(dir_bh, "get_write_access");
-       ext3_journal_get_write_access(handle, dir_bh);
        PARENT_INO(dir_bh->b_data) = cpu_to_le32(new_dir->i_ino);
        BUFFER_TRACE(dir_bh, "call ext3_journal_dirty_metadata");
        ext3_journal_dirty_metadata(handle, dir_bh);
--
To unsubscribe from this list: send the line "unsubscribe linux-ext4" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Comments

Jan Kara Nov. 23, 2010, 12:57 p.m. UTC | #1
Hello Amir,

On Tue 23-11-10 09:58:58, Amir Goldstein wrote:
> A better way to handle this situation is to get_write_access much sooner,
> when things can still be rolled back properly.
> Please find below a patch I am using in next3 to fix some un-handled
> journal errors
> and see if you can/want to use any of it as reference for your patches.
  While generally it is certainly desirable to detect errors early and bail
out without doing harm, in this particular case I don't see a big point
in complicating the code since the only realistic failure case is that the
journal is aborted, we have to run fsck anyway, and we are not causing any
unfixable damage... So I'll take the patch as is.

									Honza

> Some places in Ext3 original code don't check for return value of
> JBD functions.  Check for snapshot/journal errors in those places.
> 
> Signed-off-by: Amir Goldstein <amir73il@users.sf.net>
> 
> --------------------------------------------------------------------------------
> diff -Nuarp a/fs/ext3/balloc.c b/fs/ext3/balloc.c
> --- a/fs/ext3/balloc.c  2010-11-23 09:38:13.395922811 +0200
> +++ b/fs/ext3/balloc.c  2010-11-23 09:38:13.185923236 +0200
> @@ -674,7 +674,9 @@ do_more:
> 
>     /* We dirtied the bitmap block */
>     BUFFER_TRACE(bitmap_bh, "dirtied bitmap block");
> -   err = ext3_journal_dirty_metadata(handle, bitmap_bh);
> +   ret = ext3_journal_dirty_metadata(handle, bitmap_bh);
> +   if (!err)
> +       err = ret;
> 
>     /* And the group descriptor block */
>     BUFFER_TRACE(gd_bh, "dirtied group descriptor block");
> 
> diff -Nuarp a/fs/ext3/inode.c b/fs/ext3/inode.c
> --- a/fs/ext3/inode.c   2010-11-23 09:38:13.405923214 +0200
> +++ b/fs/ext3/inode.c   2010-11-23 09:38:13.195922747 +0200
> @@ -2362,6 +2362,9 @@ static void ext3_clear_blocks(handle_t
>         if (bh) {
>             BUFFER_TRACE(bh, "retaking write access");
>             ext3_journal_get_write_access_inode(handle, inode, bh);
> +           /* we may have lost write_access on bh */
> +           if (is_handle_aborted(handle))
> +               return;
>         }
>     }
> 
> @@ -2444,6 +2447,9 @@ static void ext3_free_data(handle_t *ha
>                 ext3_clear_blocks(handle, inode, this_bh,
>                           block_to_free,
>                           count, block_to_free_p, p);
> +               /* we may have lost write_access on this_bh */
> +               if (is_handle_aborted(handle))
> +                   return;
>                 block_to_free = nr;
>                 block_to_free_p = p;
>                 count = 1;
> @@ -2454,6 +2460,9 @@ static void ext3_free_data(handle_t *ha
>     if (count > 0)
>         ext3_clear_blocks(handle, inode, this_bh, block_to_free,
>                   count, block_to_free_p, p);
> +   /* we may have lost write_access on this_bh */
> +   if (is_handle_aborted(handle))
> +       return;
> 
>     if (this_bh) {
>         BUFFER_TRACE(this_bh, "call ext3_journal_dirty_metadata");
> 
> diff -Nuarp a/fs/ext3/namei.c b/fs/ext3/namei.c
> --- a/fs/ext3/namei.c   2010-11-23 09:38:13.415922664 +0200
> +++ b/fs/ext3/namei.c   2010-11-23 09:38:13.205923227 +0200
> @@ -1649,8 +1649,12 @@ static int ext3_delete_entry (handle_t
>         if (!ext3_check_dir_entry("ext3_delete_entry", dir, de, bh, i))
>             return -EIO;
>         if (de == de_del)  {
> +           int err;
> +
>             BUFFER_TRACE(bh, "get_write_access");
> -           ext3_journal_get_write_access(handle, bh);
> +           err = ext3_journal_get_write_access(handle, bh);
> +           if (err)
> +               return err;
>             if (pde)
>                 pde->rec_len = ext3_rec_len_to_disk(
>                     ext3_rec_len_from_disk(pde->rec_len) +
> @@ -1797,7 +1801,16 @@ retry:
>         goto out_stop;
>     }
>     BUFFER_TRACE(dir_block, "get_write_access");
> -   ext3_journal_get_write_access(handle, dir_block);
> +   err = ext3_journal_get_write_access(handle, dir_block);
> +   if (err) {
> +       drop_nlink(inode); /* is this nlink == 0? */
> +       unlock_new_inode(inode);
> +       /* no need to check for errors - we failed anyway */
> +       (void) ext3_mark_inode_dirty(handle, inode);
> +       iput(inode);
> +       brelse(dir_block);
> +       goto out_stop;
> +   }
>     de = (struct ext3_dir_entry_2 *) dir_block->b_data;
>     de->inode = cpu_to_le32(inode->i_ino);
>     de->name_len = 1;
> 
> @@ -2337,6 +2350,10 @@ static int ext3_rename (struct inode *
>         if (!new_inode && new_dir!=old_dir &&
>                 new_dir->i_nlink >= EXT3_LINK_MAX)
>             goto end_rename;
> +       BUFFER_TRACE(dir_bh, "get_write_access");
> +       retval = ext3_journal_get_write_access(handle, dir_bh);
> +       if (retval)
> +           goto end_rename;
>     }
>     if (!new_bh) {
>         retval = ext3_add_entry (handle, new_dentry, old_inode);
> @@ -2344,7 +2361,9 @@ static int ext3_rename (struct inode *
>             goto end_rename;
>     } else {
>         BUFFER_TRACE(new_bh, "get write access");
> -       ext3_journal_get_write_access(handle, new_bh);
> +       retval = ext3_journal_get_write_access(handle, new_bh);
> +       if (retval)
> +           goto end_rename;
>         new_de->inode = cpu_to_le32(old_inode->i_ino);
>         if (EXT3_HAS_INCOMPAT_FEATURE(new_dir->i_sb,
>                           EXT3_FEATURE_INCOMPAT_FILETYPE))
> @@ -2401,8 +2420,6 @@ static int ext3_rename (struct inode *
>     old_dir->i_ctime = old_dir->i_mtime = CURRENT_TIME_SEC;
>     ext3_update_dx_flag(old_dir);
>     if (dir_bh) {
> -       BUFFER_TRACE(dir_bh, "get_write_access");
> -       ext3_journal_get_write_access(handle, dir_bh);
>         PARENT_INO(dir_bh->b_data) = cpu_to_le32(new_dir->i_ino);
>         BUFFER_TRACE(dir_bh, "call ext3_journal_dirty_metadata");
>         ext3_journal_dirty_metadata(handle, dir_bh);
diff mbox

Patch

diff -Nuarp a/fs/ext3/balloc.c b/fs/ext3/balloc.c
--- a/fs/ext3/balloc.c  2010-11-23 09:38:13.395922811 +0200
+++ b/fs/ext3/balloc.c  2010-11-23 09:38:13.185923236 +0200
@@ -674,7 +674,9 @@  do_more:

    /* We dirtied the bitmap block */
    BUFFER_TRACE(bitmap_bh, "dirtied bitmap block");
-   err = ext3_journal_dirty_metadata(handle, bitmap_bh);
+   ret = ext3_journal_dirty_metadata(handle, bitmap_bh);
+   if (!err)
+       err = ret;

    /* And the group descriptor block */
    BUFFER_TRACE(gd_bh, "dirtied group descriptor block");

diff -Nuarp a/fs/ext3/inode.c b/fs/ext3/inode.c
--- a/fs/ext3/inode.c   2010-11-23 09:38:13.405923214 +0200
+++ b/fs/ext3/inode.c   2010-11-23 09:38:13.195922747 +0200
@@ -2362,6 +2362,9 @@  static void ext3_clear_blocks(handle_t
        if (bh) {
            BUFFER_TRACE(bh, "retaking write access");
            ext3_journal_get_write_access_inode(handle, inode, bh);
+           /* we may have lost write_access on bh */
+           if (is_handle_aborted(handle))
+               return;
        }
    }

@@ -2444,6 +2447,9 @@  static void ext3_free_data(handle_t *ha
                ext3_clear_blocks(handle, inode, this_bh,
                          block_to_free,
                          count, block_to_free_p, p);
+               /* we may have lost write_access on this_bh */
+               if (is_handle_aborted(handle))
+                   return;
                block_to_free = nr;
                block_to_free_p = p;
                count = 1;
@@ -2454,6 +2460,9 @@  static void ext3_free_data(handle_t *ha
    if (count > 0)
        ext3_clear_blocks(handle, inode, this_bh, block_to_free,
                  count, block_to_free_p, p);
+   /* we may have lost write_access on this_bh */
+   if (is_handle_aborted(handle))
+       return;

    if (this_bh) {
        BUFFER_TRACE(this_bh, "call ext3_journal_dirty_metadata");

diff -Nuarp a/fs/ext3/namei.c b/fs/ext3/namei.c
--- a/fs/ext3/namei.c   2010-11-23 09:38:13.415922664 +0200
+++ b/fs/ext3/namei.c   2010-11-23 09:38:13.205923227 +0200
@@ -1649,8 +1649,12 @@  static int ext3_delete_entry (handle_t
        if (!ext3_check_dir_entry("ext3_delete_entry", dir, de, bh, i))
            return -EIO;
        if (de == de_del)  {
+           int err;
+
            BUFFER_TRACE(bh, "get_write_access");
-           ext3_journal_get_write_access(handle, bh);
+           err = ext3_journal_get_write_access(handle, bh);
+           if (err)
+               return err;
            if (pde)
                pde->rec_len = ext3_rec_len_to_disk(
                    ext3_rec_len_from_disk(pde->rec_len) +
@@ -1797,7 +1801,16 @@  retry:
        goto out_stop;
    }
    BUFFER_TRACE(dir_block, "get_write_access");
-   ext3_journal_get_write_access(handle, dir_block);
+   err = ext3_journal_get_write_access(handle, dir_block);
+   if (err) {
+       drop_nlink(inode); /* is this nlink == 0? */
+       unlock_new_inode(inode);
+       /* no need to check for errors - we failed anyway */
+       (void) ext3_mark_inode_dirty(handle, inode);
+       iput(inode);
+       brelse(dir_block);
+       goto out_stop;
+   }
    de = (struct ext3_dir_entry_2 *) dir_block->b_data;
    de->inode = cpu_to_le32(inode->i_ino);
    de->name_len = 1;