When pass1 finds blocks that are mapped to multiple files, it will
print every duplicated block. If there are long sequences of
duplicate blocks (e.g. the e_pblk field is wrong in an extent), this
can cause a gigantic flood of output when a range could convey the
same information. Therefore, teach pass1b to print ranges when
possible.
Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
In ext2fs_extent_set_bmap() and ext2fs_punch_extent(), fix the parents
when altering either end of an extent so that the parent nodes reflect
the added mapping.
There's a slight complication to using fix_parents: if there are two
mappings to an lblk in the tree, the value of handle->path->curr can
point to either extent afterwards), which is documented in a comment.
Some additional color commentary from Darrick:
In the _set_bmap() case, I noticed that the "remapping last block in
extent" case would produce symptoms if we are trying to remap a
block from "extent" to "next_extent", and the two extents are
pointed to by different index nodes. _extent_replace(...,
next_extent) updates e_lblk in the leaf extent, but because there's
no _extent_fix_parents() call, the index nodes never get updated.
In the _punch_extent() case, we conclude that we need to split an
extent into two pieces since we're punching out the middle. If the
extent is the last extent in the block, the second extent will be
inserted into a new leaf node block. Without _fix_parents(), the
index node doesn't seem to get updated.
Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
In ext2fs_extent_free(), h(andle)->max_depth is used as a loop
conditional variable to free all the h->path[].buf pointers. However,
ext2fs_extent_delete() sets max_depth = 0 if we've removed everything
from the extent tree, which causes a subsequent _free() to leak some
buf pointers. max_depth can be re-incremented when splitting extent
nodes, but there's no guarantee that it'll reach the old value before
the free.
Therefore, remember the size of h->paths[] separately, and use that
when freeing the extent handle.
Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
Fix a few minor bugs that cppcheck complained about.
Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
In ext2fs_block_alloc_stats_range(), the quantity "-inuse * n" is
calculated as a signed 32-bit quantity. Unfortunately, gcc (4.6.3 on
Ubuntu 12.04) doesn't sign-extend this quantity to fill the blk64_t
parameter that ext2fs_free_blocks_count_add() wants, so the end result
is that the superblock gets a ridiculously huge free block count.
Changing the declaration of 'n' to blk64_t seems to fix this.
Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
Fix a number of things that cppcheck complains about. Most of these
are minor resource leaks and forgotten declarations.
Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
When we're looking for directory blocks for the inode remapping step,
we need to include inline_data directories in the remap process.
Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
In ext2fs_inline_data_dir_iterate(), we must be very careful to undo
any modifications we make to the dir_context pointer passed in by the
caller, because it's entirely possible that the caller will still want
to do something with the ctx or something inside.
Specifically, ext2fs_dblist_dir_iterate() wants to be able to free
ctx->buf, and it reuses the ctx for multiple dblist entries. That
means that assigning ctx->buf will cause weird crashes at the end of
dir_iterate().
Since we're being careful with ctx, we might as well handle adding the
INLINE_DATA flag to ctx->flags for ext2fs_process_dir_block, since the
dblist caller forgets to unset the flag before reusing the ctx.
This fixes some crashes and valgrind complaints in resize2fs, and is
necessary for the next patch, which fixes resize2fs not to corrupt
inline_data filesystems.
Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
When expanding an inline data inode, it's possible that the reduction
in the size of the EA structures causes the freeing of the EA block,
which changes the inode. If this happens, the local version of the
inode that ext2fs_inline_data_expand was modifying will be out of sync
with what's on the disk. This local copy gets written out to disk
after a block allocation, at which point it's possible that the inode
EA block and logical block zero point to the same physical block,
which is bad news.
Therefore, write the local copy to disk before removing the inline
data EA, and reread it afterwards.
Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
When checking inline data blocks, always zero pctx->errcode because
otherwise a previous error condition could leak through and "cause" a
fatal block iteration failure. I found this by corrupting an xattr
block on an inline_data inode and fsck aborted when I tried to repair
it.
Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
i_blocks covers the number of blocks allocated to an inode for data,
extents, and ACL blocks. Since it's possible for a file to have a
separate ACL block and inline data, we must be careful when expanding
an inline data file to adjust, not set, the value of i_blocks.
Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
ext2fs_inline_data_set() tries to ensure that there is sufficient free
space in the inode to store the inline data. Unfortunately, it gets
the check wrong -- ext2fs_xattr_inode_max_size() returns the amount of
unused bytes in the EA area, and _data_set() doesn't factor in the
size of the existing inline data. Therefore, a strict rewrite of an
N-byte inlinedata with another N-byte inlinedata fails.
Fix the code to do the size check correctly.
Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
If a directory's contents are stored entirely inside the inode,
there's no index to rebuild and no dirblock checksum to recompute.
As far as I know these are the only two reasons to call dir rehash.
Therefore, we can move on to the next dir instead of what we do right
now, which is try to iterate the dir blocks (which of course fails due
to the inline_data iflag being set) and then flood stdout with useless
messages that aren't even failures.
Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
Fix a typo that we didn't notice because all the world's an x86. :-)
Reported-by: jon ernst <jonernst07@gmail.com>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
Enhance debugfs to be able to display and modify extended attributes, and
create some simple tests for the extended attribute editing functions.
Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
Add magic number checking to the extended attribute editing handle;
move inline data to the head of the attribute list when writing so
that inline data ends up in the inode area; and always zero the
attribute space before writing to ensure that we can delete the last
xattr.
Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
When calling populate_fs, the map for hardlink detection is not
cleaned up between populate_fs invocations, which could lead to
unexpected results if anyone calls populate_fs twice in the same
client program). This doesn't happen right now, but we might as well
clean it up.
The detctor fails if the external directory crosses mountpoints,
so fix that too.
Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
Since create_inode.c is shared between debugfs and mke2fs, don't
spread debugfs internal state into mke2fs.
Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
Fix a couple of small style issues in the create_inode files.
Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
Reviewed-by: Andreas Dilger <adilger@dilger.ca>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
do_write_internal returns errno when ext2 library calls fail; since
errno only reflects the outcome of the last C library call, this will
result in confused callers. Eliminate the naked return since
this results in an undefined return value.
Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
Update the manual for the -d option
Signed-off-by: Robert Yang <liezhi.yang@windriver.com>
Reviewed-by: Darren Hart <dvhart@linux.intel.com>
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
* Use the functions in misc/create_inode.c, and remove the duplicated
code.
* The CREATE_INODE_DEPS in the debugfs/Makefile.in is used for recording
create_inode.o's depends which is from misc/Makefile.in, we have to
recompile create_inode.o because we need it to print more messages when
it is used by debugfs, just like we recompile e2freefrag.o, but it seems
that the e2freefrag.o's depends in debugfs/Makefile.in is incorrect, it
would not rebuild when its depends (e.g.: lib/config.h) is changed,
which would cause unexpected errors. Make duplicated code in
debugfs/Makefile.in and misc/Makefile.in is not a good idea, maybe we'd
better define CREATE_INODE_DEPS in the top Makefile, I'd like to send
another patch and fix the e2freefrag if you are fine with it.
Signed-off-by: Robert Yang <liezhi.yang@windriver.com>
Reviewed-by: Darren Hart <dvhart@linux.intel.com>
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
Create the inode and save the native inode number when we meet the hard
link (st_nlink > 1) at the first time, use ext2fs_link() to link the
name to the target inode number when we meet the same native inode
number again.
This algorithm is referred this from the genext2fs.
Signed-off-by: Robert Yang <liezhi.yang@windriver.com>
Reviewed-by: Darren Hart <dvhart@linux.intel.com>
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
This option is used for adding the files from a given directory (the
root-directory) to the filesystem, it is similiar to genext2fs, but
genext2fs doesn't fully support ext4.
Signed-off-by: Robert Yang <liezhi.yang@windriver.com>
Reviewed-by: Darren Hart <dvhart@linux.intel.com>
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
Set the uid, gid, mode and time for inode.
Signed-off-by: Robert Yang <liezhi.yang@windriver.com>
Reviewed-by: Darren Hart <dvhart@linux.intel.com>
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
The do_mkdir_internal() is used for making dir on the target fs, most of
the code are from debugfs/debugfs.c, the debugfs/debugfs.c will be
modified to use this function.
Signed-off-by: Robert Yang <liezhi.yang@windriver.com>
Reviewed-by: Darren Hart <dvhart@linux.intel.com>
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
The do_write_internal() is used for copying file from native fs to
target, most of the code are from debugfs/debugfs.c, the
debugfs/debugfs.c will be modified to use this function.
Signed-off-by: Robert Yang <liezhi.yang@windriver.com>
Reviewed-by: Darren Hart <dvhart@linux.intel.com>
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
The do_symlink_internal() is used for creating symlinks, most of the
code are from debugfs/debugfs.c, the debugfs/debugfs.c will be modified
to use this function.
Signed-off-by: Robert Yang <liezhi.yang@windriver.com>
Reviewed-by: Darren Hart <dvhart@linux.intel.com>
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
The do_mknod_internal() is used for creating special file which is
block, character and fifo, most of the code are from debugfs/debugfs.c,
the debugfs/debugfs.c will be modified to use this function.
Signed-off-by: Robert Yang <liezhi.yang@windriver.com>
Reviewed-by: Darren Hart <dvhart@linux.intel.com>
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
Use opendir() and readdir() to read the native directory, then use
lstat() to identify the file type and call the corresponding function to
add the file to the filesystem, call the populate_fs() recursively if it
is a directory.
NOTE: the libext2fs can't create the socket file.
Signed-off-by: Robert Yang <liezhi.yang@windriver.com>
Reviewed-by: Darren Hart <dvhart@linux.intel.com>
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
We will add a -d option which will be used for adding the files from a
given directory to the filesystem, it is similiar to genext2fs, but
genext2fs doesn't fully support ext4.
* We already have the basic operations in debugfs:
- Copy regular file
- Create directory
- Create symlink
- Create special file
We will move these operations into create_inode.h and create_inode.c,
then let both mke2fs and debugfs use them.
* What we need to do are:
- Copy the given directory recursively, this will be done by the
populate_fs()
- Set the owner, mode and other informations
- Handle the hard links
TODO:
- The libext2fs can't create the socket file (S_IFSOCK), do we have a
plan to support it ?
Signed-off-by: Robert Yang <liezhi.yang@windriver.com>
Reviewed-by: Darren Hart <dvhart@linux.intel.com>
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
Since auto_64-bit_support is on by default, resize_inode feature will
be disabled when creating a >16T ext4 according to mke2fs.conf(5).
This should also be done when making ext4 with "-O 64bit" to enable
64bit feature explicitly. Otherwise online resize to enlarge a
over-16T fs to larger would fail.
[root@localhost resize]# truncate -s 50t fs.img
[root@localhost resize]# losetup /dev/loop0 fs.img
[root@localhost resize]# mkfs -t ext4 -O 64bit /dev/loop0 30t
[root@localhost resize]# mount /dev/loop0 mnt
[root@localhost resize]# resize2fs /dev/loop0
resize2fs 1.42.7 (21-Jan-2013)
Filesystem at /dev/loop0 is mounted on /root/resize/mnt; on-line resizing required
old_desc_blocks = 3840, new_desc_blocks = 6400
resize2fs: Invalid argument While checking for on-line resizing support
And dmesg shows
[688378.442623] EXT4-fs (loop0): resizing filesystem from 6710886400 to 13421772800 blocks
[688378.443216] EXT4-fs warning (device loop0): verify_reserved_gdb:700: reserved GDT 3201 missing grp 177147 (5804756097)
[688378.443222] EXT4-fs (loop0): resized filesystem to 8858370048
[688378.528451] EXT4-fs warning (device loop0): ext4_group_extend:1710: can't shrink FS - resize aborted
With this fix resize2fs could do the online enlarge correctly.
Signed-off-by: Eryu Guan <guaneryu@gmail.com>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
There is no reason to use e2fsck_global_ctx in
e2fsck_set_bitmap_type(), since we can get the context structure from
fs->priv_data.
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
The e2fsck_global_ctx varible was only being set if HAVE_SIGNAL_H is
defined. There are systems, such as Android, where this is not true.
This was causing e2fsck_set_bitmap_type() to seg fault since
e2fsck_global_ctx was not NULL.
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
Reported-by: JP Abgrall <jpa@google.com>
In this unit test, we will test the interface of inline data and make
sure it is fine.
Signed-off-by: Zheng Liu <wenqing.lz@taobao.com>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
Currently we have already exported inode cache flush and free functions
for users. This commit exports inode cache creation function. Later
we will use this function to initialize inode cache and do some unit
tests for inline data.
Signed-off-by: Zheng Liu <wenqing.lz@taobao.com>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
In this test, inode flag is some random data, and after we apply inline
data patch set we should need to handle it.
Signed-off-by: Zheng Liu <wenqing.lz@taobao.com>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
In e2fsck_expand_directory() we don't handle a dir with inline data
because when this function is called the directory inode shouldn't
contains inline data.
Signed-off-by: Zheng Liu <wenqing.lz@taobao.com>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
Now inline_data doesn't depend on ext_attr. Hence we don't need to do
this sanity check. But if the inode size is too small (128 bytes),
inline_data will be useless because we couldn't save data in ibody
extented attribute. So we need to report this error.
Signed-off-by: Zheng Liu <wenqing.lz@taobao.com>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>