Currently, there exists a race between delayed allocated writes and
the writeback when bigalloc feature is in use. The race was because we
wanted to determine what blocks in a cluster are under delayed
allocation and we were using buffer_delayed(bh) check for it. But, the
writeback codepath clears this bit without any synchronization which
resulted in a race and an ext4 warning similar to:
EXT4-fs (ram1): ext4_da_update_reserve_space: ino 13, used 1 with only 0
reserved data blocks
The race existed in two places.
(1) between ext4_find_delalloc_range() and ext4_map_blocks() when called from
writeback code path.
(2) between ext4_find_delalloc_range() and ext4_da_get_block_prep() (where
buffer_delayed(bh) is set.
To fix (1), this patch introduces a new buffer_head state bit -
BH_Da_Mapped. This bit is set under the protection of
EXT4_I(inode)->i_data_sem when we have actually mapped the delayed
allocated blocks during the writeout time. We can now reliably check
for this bit inside ext4_find_delalloc_range() to determine whether
the reservation for the blocks have already been claimed or not.
To fix (2), it was necessary to set buffer_delay(bh) under the
protection of i_data_sem. So, I extracted the very beginning of
ext4_map_blocks into a new function - ext4_da_map_blocks() - and
performed the required setting of bh_delay bit and the quota
reservation under the protection of i_data_sem. These two fixes makes
the checking of buffer_delay(bh) and buffer_da_mapped(bh) consistent,
thus removing the race.
Tested: I was able to reproduce the problem by running 'dd' and
'fsync' in parallel. Also, xfstests sometimes used to reproduce this
race. After the fix both my test and xfstests were successful and no
race (warning message) was observed.
Google-Bug-Id: 4997027
Signed-off-by: Aditya Kali <adityakali@google.com>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
This patch adds some tracepoints in ext4/extents.c and updates a tracepoint in
ext4/inode.c.
Tested: Built and ran the kernel and verified that these tracepoints work.
Also ran xfstests.
Signed-off-by: Aditya Kali <adityakali@google.com>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
Rename the function so it is more clear what is going on. Also rename
the various variables so it's clearer what's happening.
Also fix a missing blocks to cluster conversion when reading the
number of reserved blocks for root.
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
This function really claims a number of free clusters, not blocks, so
rename it so it's clearer what's going on.
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
This function really returns the number of clusters after initializing
an uninitalized block bitmap has been initialized.
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
This function really counts the free clusters reported in the block
group descriptors, so rename it to reduce confusion.
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
The field bg_free_blocks_count_{lo,high} in the block group
descriptor has been repurposed to hold the number of free clusters for
bigalloc functions. So rename the functions so it makes it easier to
read and audit the block allocation and block freeing code.
Note: at this point in bigalloc development we doesn't support
online resize, so this also makes it really obvious all of the places
we need to fix up to add support for online resize.
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
With bigalloc changes, the i_blocks value was not correctly set (it was still
set to number of blocks being used, but in case of bigalloc, we want i_blocks
to represent the number of clusters being used). Since the quota subsystem sets
the i_blocks value, this patch fixes the quota accounting and makes sure that
the i_blocks value is set correctly.
Signed-off-by: Aditya Kali <adityakali@google.com>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
The default group preallocation size had been previously set to 512
blocks/clusters, regardless of the block/cluster size. This is
probably too big for large cluster sizes. So adjust the default so
that it is 2 megabytes or 32 clusters, whichever is larger.
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
Convert the free_blocks to be free_clusters to make the final revised
bigalloc changes easier to read/understand.
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
Convert the percpu counters s_dirtyblocks_counter and
s_freeblocks_counter in struct ext4_super_info to be
s_dirtyclusters_counter and s_freeclusters_counter.
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
When we are truncating (as opposed unlinking) a file, we need to worry
about partial truncates of a file, especially in the light of sparse
files. The changes here make sure that arbitrary truncates of sparse
files works correctly. Yeah, it's messy.
Note that these functions will need to be revisted when the punch
ioctl is integrated --- in fact this commit will probably have merge
conflicts with the punch changes which Allison Henders and the IBM LTC
have been working on. I will need to fix this up when either patch
hits mainline.
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
If we need to allocate a new block in ext4_ext_map_blocks(), the
function needs to see if the cluster has already been allocated.
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
The ext4_free_blocks() function now has two new flags that indicate
whether a partial cluster at the beginning or the end of the block
extents should be freed or not. That will be up the caller (i.e.,
truncate), who can figure out whether partial clusters at the
beginning or the end of a block range can be freed.
We also have to update the ext4_mb_free_metadata() and
release_blocks_on_commit() machinery to be cluster-based, since it is
used by ext4_free_blocks().
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
In most of mballoc.c, we do everything in units of clusters, since the
block allocation bitmaps and buddy bitmaps are all denominated in
clusters. The one place where we do deal with absolute block numbers
is in the code that handles the preallocation regions, since in the
case of inode-based preallocation regions, the start of the
preallocation region can't be relative to the beginning of the group.
So this adds a bit of complexity, where pa_pstart and pa_lstart are
block numbers, while pa_free, pa_len, and fe_len are denominated in
units of clusters.
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
Certain parts of the ext4 code base, primarily in mballoc.c, use a
block group number and offset from the beginning of the block group.
This offset is invariably used to index into the allocation bitmap, so
change the offset to be denominated in units of clusters.
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
The function ext4_free_blocks_after_init() used to be a #define of
ext4_init_block_bitmap(). This actually made it difficult to
understand how the function worked, and made it hard make changes to
support clusters. So as an initial cleanup, I've separated out the
functionality of initializing block bitmap from calculating the number
of free blocks in the new block group.
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
This makes it easier to understand how ext4_init_block_bitmap() works,
and it will assist when we split out ext4_free_blocks_after_init() in
the next commit.
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
Change the places in fs/ext4/mballoc.c where EXT4_BLOCKS_PER_GROUP are
used to indicate the number of bits in a block bitmap (which is really
a cluster allocation bitmap in bigalloc file systems). There are
still some places in the ext4 codebase where usage of
EXT4_BLOCKS_PER_GROUP needs to be audited/fixed, in code paths that
aren't used given the initial restricted assumptions for bigalloc.
These will need to be fixed before we can relax those restrictions.
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
At least initially if the bigalloc feature is enabled, we will not
support non-extent mapped inodes, online resizing, online defrag, or
the FITRIM ioctl. This simplifies the initial implementation.
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
This adds supports for bigalloc file systems. It teaches the mount
code just enough about bigalloc superblock fields that it will mount
the file system without freaking out that the number of blocks per
group is too big.
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
The del_gendisk() function uninitializes the disk-specific data
structures, including the bdi structure, without telling anyone
else. Once this happens, any attempt to call mark_buffer_dirty()
(for example, by ext4_commit_super), will cause a kernel OOPS.
Fix this for now until we can fix things in an architecturally correct
way.
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
While running extended fsx tests to verify the preceeding patches,
a similar bug was also found in the write operation
When ever a write operation begins or ends in a hole,
or extends EOF, the partial page contained in the hole
or beyond EOF needs to be zeroed out.
To correct this the new ext4_discard_partial_page_buffers_no_lock
routine is used to zero out the partial page, but only for buffer
heads that are already unmapped.
Signed-off-by: Allison Henderson <achender@linux.vnet.ibm.com>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
While running extended fsx tests to verify the first
two patches, a similar bug was also found in the
truncate operation.
This bug happens because the truncate routine only zeros
the unblock aligned portion of the last page. This means
that the block aligned portions of the page appearing after
i_size are left unzeroed, and the buffer heads still mapped.
This bug is corrected by using ext4_discard_partial_page_buffers
in the truncate routine to zero the partial page and unmap
the buffer headers.
Signed-off-by: Allison Henderson <achender@linux.vnet.ibm.com>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
In delayed allocation mode, it's important to only call
ext4_jbd2_file_inode when the file has been extended. This is
necessary to avoid a race which first got introduced in commit
678aaf481, but which was made much more common with the introduction
of the "punch hole" functionality. (Especially when dioread_nolock
was enabled; when I could reliably reproduce this problem with
xfstests #74.)
The race is this: If while trying to writeback a delayed allocation
inode, there is a need to map delalloc blocks, and we run out of space
in the journal, *and* at the same time the inode is already on the
committing transaction's t_inode_list (because for example while doing
the punch hole operation, ext4_jbd2_file_inode() is called), then the
commit operation will wait for the inode to finish all of its pending
writebacks by calling filemap_fdatawait(), but since that inode has
one or more pages with the PageWriteback flag set, the commit
operation will wait forever, and the so the writeback of the inode can
never take place, and the kjournald thread and the writeback thread
end up waiting for each other --- forever.
It's important at this point to recall why an inode is placed on the
t_inode_list; it is to provide the data=ordered guarantees that we
don't end up exposing stale data. In the case where we are truncating
or punching a hole in the inode, there is no possibility that stale
data could be exposed in the first place, so we don't need to put the
inode on the t_inode_list!
The right long-term fix is to get rid of data=ordered mode altogether,
and only update the extent tree or indirect blocks after the data has
been written. Until then, this change will also avoid some
unnecessary waiting in the commit operation.
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
Cc: Allison Henderson <achender@linux.vnet.ibm.com>
Cc: Jan Kara <jack@suse.cz>
Add debugging information in case jbd2_journal_dirty_metadata() is
called with a buffer_head which didn't have
jbd2_journal_get_write_access() called on it, or if the journal_head
has the wrong transaction in it. In addition, return an error code.
This won't change anything for ocfs2, which will BUG_ON() the non-zero
exit code.
For ext4, the caller of this function is ext4_handle_dirty_metadata(),
and on seeing a non-zero return code, will call __ext4_journal_stop(),
which will print the function and line number of the (buggy) calling
function and abort the journal. This will allow us to recover instead
of bug halting, which is better from a robustness and reliability
point of view.
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
If the user explicitly specifies conflicting mount options for
delalloc or dioread_nolock and data=journal, fail the mount, instead
of printing a warning and continuing (since many user's won't look at
dmesg and notice the warning).
Also, print a single warning that data=journal implies that delayed
allocation is not on by default (since it's not supported), and
furthermore that O_DIRECT is not supported. Improve the text in
Documentation/filesystems/ext4.txt so this is clear there as well.
Similarly, if the dioread_nolock mount option is specified when the
file system block size != PAGE_SIZE, fail the mount instead of
printing a warning message and ignoring the mount option.
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
This patch fixes a second punch hole bug found by xfstests 127.
This bug happens because punch hole needs to flush the pages
of the hole to avoid race conditions. But if the end of the
hole is in the same page as i_size, the buffer heads beyond
i_size need to be unmapped and the page needs to be zeroed
after it is flushed.
To correct this, the new ext4_discard_partial_page_buffers
routine is used to zero and unmap the partial page
beyond i_size if the end of the hole appears in the same
page as i_size.
The code has also been optimized to set the end of the hole
to the page after i_size if the specified hole exceeds i_size,
and the code that flushes the pages has been simplified.
Signed-off-by: Allison Henderson <achender@linux.vnet.ibm.com>
This patch addresses a bug found by xfstests 75, 112, 127
when blocksize = 1k
This bug happens because the punch hole code only zeros
out non block aligned regions of the page. This means that if the
blocks are smaller than a page, then the block aligned regions of
the page inside the hole are left un-zeroed, and their buffer heads
are still mapped. This bug is corrected by using
ext4_discard_partial_page_buffers to properly zero the partial page
at the head and tail of the hole, and unmap the corresponding buffer
heads
This patch also addresses a bug reported by Lukas while working on a
new patch to add discard support for loop devices using punch hole.
The bug happened because of the first and last block number
needed to be cast to a larger data type before calculating the
byte offset, but since now we only need the byte offsets of the
pages, we no longer even need to be calculating the byte offsets
of the blocks. The code to do the block offset calculations is
removed in this patch.
Signed-off-by: Allison Henderson <achender@linux.vnet.ibm.com>
This patch adds two new routines: ext4_discard_partial_page_buffers
and ext4_discard_partial_page_buffers_no_lock.
The ext4_discard_partial_page_buffers routine is a wrapper
function to ext4_discard_partial_page_buffers_no_lock.
The wrapper function locks the page and passes it to
ext4_discard_partial_page_buffers_no_lock.
Calling functions that already have the page locked can call
ext4_discard_partial_page_buffers_no_lock directly.
The ext4_discard_partial_page_buffers_no_lock function
zeros a specified range in a page, and unmaps the
corresponding buffer heads. Only block aligned regions of the
page will have their buffer heads unmapped. Unblock aligned regions
will be mapped if needed so that they can be updated with the
partial zero out. This function is meant to
be used to update a page and its buffer heads to be zeroed
and unmapped when the corresponding blocks have been released
or will be released.
This routine is used in the following scenarios:
* A hole is punched and the non page aligned regions
of the head and tail of the hole need to be discarded
* The file is truncated and the partial page beyond EOF needs
to be discarded
* The end of a hole is in the same page as EOF. After the
page is flushed, the partial page beyond EOF needs to be
discarded.
* A write operation begins or ends inside a hole and the partial
page appearing before or after the write needs to be discarded
* A write operation extends EOF and the partial page beyond EOF
needs to be discarded
This function takes a flag EXT4_DISCARD_PARTIAL_PG_ZERO_UNMAPPED
which is used when a write operation begins or ends in a hole.
When the EXT4_DISCARD_PARTIAL_PG_ZERO_UNMAPPED flag is used, only
buffer heads that are already unmapped will have the corresponding
regions of the page zeroed.
Signed-off-by: Allison Henderson <achender@linux.vnet.ibm.com>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
ext4_dx_add_entry manipulates bh2 and frames[0].bh, which are two buffer_heads
that point to directory blocks assigned to the directory inode. However, the
function calls ext4_handle_dirty_metadata with the inode of the file that's
being added to the directory, not the directory inode itself. Therefore,
correct the code to dirty the directory buffers with the directory inode, not
the file inode.
Signed-off-by: Darrick J. Wong <djwong@us.ibm.com>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
Cc: stable@kernel.org
ext4_mkdir calls ext4_handle_dirty_metadata with dir_block and the inode "dir".
Unfortunately, dir_block belongs to the newly created directory (which is
"inode"), not the parent directory (which is "dir"). Fix the incorrect
association.
Signed-off-by: Darrick J. Wong <djwong@us.ibm.com>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
Cc: stable@kernel.org
When ext4_rename performs a directory rename (move), dir_bh is a
buffer that is modified to update the '..' link in the directory being
moved (old_inode). However, ext4_handle_dirty_metadata is called with
the old parent directory inode (old_dir) and dir_bh, which is
incorrect because dir_bh does not belong to the parent inode. Fix
this error.
Signed-off-by: Darrick J. Wong <djwong@us.ibm.com>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
Cc: stable@kernel.org
Currently attempts to open a file with O_DIRECT in data=journal mode
causes the open to fail with -EINVAL. This makes it very hard to test
data=journal mode. So we will let the open succeed, but then always
fall back to O_DSYNC buffered writes.
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
This doesn't make much sense, and it exposes a bug in the kernel where
attempts to create a new file in an append-only directory using
O_CREAT will fail (but still leave a zero-length file). This was
discovered when xfstests #79 was generalized so it could run on all
file systems.
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
Cc:stable@kernel.org
The i_mutex lock and flush_completed_IO() added by commit 2581fdc810
in ext4_evict_inode() causes lockdep complaining about potential
deadlock in several places. In most/all of these LOCKDEP complaints
it looks like it's a false positive, since many of the potential
circular locking cases can't take place by the time the
ext4_evict_inode() is called; but since at the very least it may mask
real problems, we need to address this.
This change removes the flush_completed_IO() and i_mutex lock in
ext4_evict_inode(). Instead, we take a different approach to resolve
the software lockup that commit 2581fdc810 intends to fix. Rather
than having ext4-dio-unwritten thread wait for grabing the i_mutex
lock of an inode, we use mutex_trylock() instead, and simply requeue
the work item if we fail to grab the inode's i_mutex lock.
This should speed up work queue processing in general and also
prevents the following deadlock scenario: During page fault,
shrink_icache_memory is called that in turn evicts another inode B.
Inode B has some pending io_end work so it calls ext4_ioend_wait()
that waits for inode B's i_ioend_count to become zero. However, inode
B's ioend work was queued behind some of inode A's ioend work on the
same cpu's ext4-dio-unwritten workqueue. As the ext4-dio-unwritten
thread on that cpu is processing inode A's ioend work, it tries to
grab inode A's i_mutex lock. Since the i_mutex lock of inode A is
still hold before the page fault happened, we enter a deadlock.
Signed-off-by: Jiaying Zhang <jiayingz@google.com>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
Add a new REQ_PRIO to let requests preempt others in the cfq I/O schedule,
and lave REQ_META purely for marking requests as metadata in blktrace.
All existing callers of REQ_META except for XFS are updated to also
set REQ_PRIO for now.
Signed-off-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Namhyung Kim <namhyung@gmail.com>
Signed-off-by: Jens Axboe <jaxboe@fusionio.com>
Replace all occurnanced of the undocumented READ_META with READ | REQ_META
and remove the unused WRITE_META define.
Signed-off-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Jens Axboe <jaxboe@fusionio.com>
* 'for_linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tytso/ext4:
ext4: flush any pending end_io requests before DIO reads w/dioread_nolock
ext4: fix nomblk_io_submit option so it correctly converts uninit blocks
ext4: Resolve the hang of direct i/o read in handling EXT4_IO_END_UNWRITTEN.
ext4: call ext4_ioend_wait and ext4_flush_completed_IO in ext4_evict_inode
ext4: Fix ext4_should_writeback_data() for no-journal mode
There is a race between ext4 buffer write and direct_IO read with
dioread_nolock mount option enabled. The problem is that we clear
PageWriteback flag during end_io time but will do
uninitialized-to-initialized extent conversion later with dioread_nolock.
If an O_direct read request comes in during this period, ext4 will return
zero instead of the recently written data.
This patch checks whether there are any pending uninitialized-to-initialized
extent conversion requests before doing O_direct read to close the race.
Note that this is just a bandaid fix. The fundamental issue is that we
clear PageWriteback flag before we really complete an IO, which is
problem-prone. To fix the fundamental issue, we may need to implement an
extent tree cache that we can use to look up pending to-be-converted extents.
Signed-off-by: Jiaying Zhang <jiayingz@google.com>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
Cc: stable@kernel.org
Bug discovered by Jan Kara:
Finally, commit 1449032be1 returned back
the old IO submission code but apparently it forgot to return the old
handling of uninitialized buffers so we unconditionnaly call
block_write_full_page() without specifying end_io function. So AFAICS
we never convert unwritten extents to written in some cases. For
example when I mount the fs as: mount -t ext4 -o
nomblk_io_submit,dioread_nolock /dev/ubdb /mnt and do
int fd = open(argv[1], O_RDWR | O_CREAT | O_TRUNC, 0600);
char buf[1024];
memset(buf, 'a', sizeof(buf));
fallocate(fd, 0, 0, 16384);
write(fd, buf, sizeof(buf));
I get a file full of zeros (after remounting the filesystem so that
pagecache is dropped) instead of seeing the first KB contain 'a's.
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
Cc: stable@kernel.org
EXT4_IO_END_UNWRITTEN flag set and the increase of i_aiodio_unwritten
should be done simultaneously since ext4_end_io_nolock always clear
the flag and decrease the counter in the same time.
We don't increase i_aiodio_unwritten when setting
EXT4_IO_END_UNWRITTEN so it will go nagative and causes some process
to wait forever.
Part of the patch came from Eric in his e-mail, but it doesn't fix the
problem met by Michael actually.
http://marc.info/?l=linux-ext4&m=131316851417460&w=2
Reported-and-Tested-by: Michael Tokarev<mjt@tls.msk.ru>
Signed-off-by: Eric Sandeen <sandeen@redhat.com>
Signed-off-by: Tao Ma <boyu.mt@taobao.com>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
Cc: stable@kernel.org
Flush inode's i_completed_io_list before calling ext4_io_wait to
prevent the following deadlock scenario: A page fault happens while
some process is writing inode A. During page fault,
shrink_icache_memory is called that in turn evicts another inode
B. Inode B has some pending io_end work so it calls ext4_ioend_wait()
that waits for inode B's i_ioend_count to become zero. However, inode
B's ioend work was queued behind some of inode A's ioend work on the
same cpu's ext4-dio-unwritten workqueue. As the ext4-dio-unwritten
thread on that cpu is processing inode A's ioend work, it tries to
grab inode A's i_mutex lock. Since the i_mutex lock of inode A is
still hold before the page fault happened, we enter a deadlock.
Also moves ext4_flush_completed_IO and ext4_ioend_wait from
ext4_destroy_inode() to ext4_evict_inode(). During inode deleteion,
ext4_evict_inode() is called before ext4_destroy_inode() and in
ext4_evict_inode(), we may call ext4_truncate() without holding
i_mutex lock. As a result, there is a race between flush_completed_IO
that is called from ext4_ext_truncate() and ext4_end_io_work, which
may cause corruption on an io_end structure. This change moves
ext4_flush_completed_IO and ext4_ioend_wait from ext4_destroy_inode()
to ext4_evict_inode() to resolve the race between ext4_truncate() and
ext4_end_io_work during inode deletion.
Signed-off-by: Jiaying Zhang <jiayingz@google.com>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
Cc: stable@kernel.org
ext4_should_writeback_data() had an incorrect sequence of
tests to determine if it should return 0 or 1: in
particular, even in no-journal mode, 0 was being returned
for a non-regular-file inode.
This meant that, in non-journal mode, we would use
ext4_journalled_aops for directories, symlinks, and other
non-regular files. However, calling journalled aop
callbacks when there is no valid handle, can cause problems.
This would cause a kernel crash with Jan Kara's commit
2d859db3e4 ("ext4: fix data corruption in inodes with
journalled data"), because we now dereference 'handle' in
ext4_journalled_write_end().
I also added BUG_ONs to check for a valid handle in the
obviously journal-only aops callbacks.
I tested this running xfstests with a scratch device in
these modes:
- no-journal
- data=ordered
- data=writeback
- data=journal
All work fine; the data=journal run has many failures and a
crash in xfstests 074, but this is no different from a
vanilla kernel.
Signed-off-by: Curt Wohlgemuth <curtw@google.com>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
Cc: stable@kernel.org
Commit df5e622340 ("ext4: fix deadlock in ext4_symlink() in ENOSPC
conditions") recalculated the number of credits needed for a long
symlink, in the process of splitting it into two transactions. However,
the first credit calculation under-counted because if selinux is
enabled, credits are needed to create the selinux xattr as well.
Overrunning the reservation will result in an OOPS in
jbd2_journal_dirty_metadata() due to this assert:
J_ASSERT_JH(jh, handle->h_buffer_credits > 0);
Fix this by increasing the reservation size.
Signed-off-by: Eric Sandeen <sandeen@redhat.com>
Reviewed-by: Jan Kara <jack@suse.cz>
Acked-by: "Theodore Ts'o" <tytso@mit.edu>
Cc: stable@kernel.org
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Commit 9933fc0i (ext4: introduce ext4_kvmalloc(), ext4_kzalloc(), and
ext4_kvfree()) intruduced wrappers around k*alloc/vmalloc but introduced
a typo for ext4_kzalloc() by not using kzalloc() but kmalloc().
Signed-off-by: Mathias Krause <minipli@googlemail.com>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
* 'for_linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tytso/ext4: (60 commits)
ext4: prevent memory leaks from ext4_mb_init_backend() on error path
ext4: use EXT4_BAD_INO for buddy cache to avoid colliding with valid inode #
ext4: use ext4_msg() instead of printk in mballoc
ext4: use ext4_kvzalloc()/ext4_kvmalloc() for s_group_desc and s_group_info
ext4: introduce ext4_kvmalloc(), ext4_kzalloc(), and ext4_kvfree()
ext4: use the correct error exit path in ext4_init_inode_table()
ext4: add missing kfree() on error return path in add_new_gdb()
ext4: change umode_t in tracepoint headers to be an explicit __u16
ext4: fix races in ext4_sync_parent()
ext4: Fix overflow caused by missing cast in ext4_fallocate()
ext4: add action of moving index in ext4_ext_rm_idx for Punch Hole
ext4: simplify parameters of reserve_backup_gdb()
ext4: simplify parameters of add_new_gdb()
ext4: remove lock_buffer in bclean() and setup_new_group_blocks()
ext4: simplify journal handling in setup_new_group_blocks()
ext4: let setup_new_group_blocks() set multiple bits at a time
ext4: fix a typo in ext4_group_extend()
ext4: let ext4_group_add_blocks() handle 0 blocks quickly
ext4: let ext4_group_add_blocks() return an error code
ext4: rename ext4_add_groupblocks() to ext4_group_add_blocks()
...
Fix up conflict in fs/ext4/inode.c: commit aacfc19c62 ("fs: simplify
the blockdev_direct_IO prototype") had changed the ext4_ind_direct_IO()
function for the new simplified calling convention, while commit
dae1e52cb1 ("ext4: move ext4_ind_* functions from inode.c to
indirect.c") moved the function to another file.
In ext4_mb_init(), if the s_locality_group allocation fails it will
currently cause the allocations made in ext4_mb_init_backend() to
be leaked. Moving the ext4_mb_init_backend() allocation after the
s_locality_group allocation avoids that problem.
Signed-off-by: Yu Jian <yujian@whamcloud.com>
Signed-off-by: Andreas Dilger <adilger@whamcloud.com>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
Introduce new helper functions which try kmalloc, and then fall back
to vmalloc if necessary, and use them for allocating and deallocating
s_flex_groups.
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
This patch lets ext4_init_inode_table() handle errors right.
ext4_init_inode_table() should down_write() alloc_sem which
has been up_write()ed and stop the started journal handle.
Signed-off-by: Yongqiang Yang <xiaoqiangnk@gmail.com>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
We added some more error handling in b40971426a "ext4: add error
checking to calls to ext4_handle_dirty_metadata()". But we need to
call kfree() as well to avoid a memory leak.
Signed-off-by: Dan Carpenter <error27@gmail.com>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
Fix problems if fsync() races against a rename of a parent directory
as pointed out by Al Viro in his own inimitable way:
>While we are at it, could somebody please explain what the hell is ext4
>doing in
>static int ext4_sync_parent(struct inode *inode)
>{
> struct writeback_control wbc;
> struct dentry *dentry = NULL;
> int ret = 0;
>
> while (inode && ext4_test_inode_state(inode, EXT4_STATE_NEWENTRY)) {
> ext4_clear_inode_state(inode, EXT4_STATE_NEWENTRY);
> dentry = list_entry(inode->i_dentry.next,
> struct dentry, d_alias);
> if (!dentry || !dentry->d_parent || !dentry->d_parent->d_inode)
> break;
> inode = dentry->d_parent->d_inode;
> ret = sync_mapping_buffers(inode->i_mapping);
> ...
>Note that dentry obviously can't be NULL there. dentry->d_parent is never
>NULL. And dentry->d_parent would better not be negative, for crying out
>loud! What's worse, there's no guarantees that dentry->d_parent will
>remain our parent over that sync_mapping_buffers() *and* that inode won't
>just be freed under us (after rename() and memory pressure leading to
>eviction of what used to be our dentry->d_parent)......
Reported-by: Al Viro <viro@ZenIV.linux.org.uk>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
The logical block number in map.l_blk is a __u32, and so before we
shift it left, by the block size, we neeed cast it to a 64-bit size.
Otherwise i_size can be corrupted on an ENOSPC.
# df -T /mnt/mp1
Filesystem Type 1K-blocks Used Available Use% Mounted on
/dev/sda6 ext4 9843276 153056 9190200 2% /mnt/mp1
# fallocate -o 0 -l 2199023251456 /mnt/mp1/testfile
fallocate: /mnt/mp1/testfile: fallocate failed: No space left on device
# stat /mnt/mp1/testfile
File: `/mnt/mp1/testfile'
Size: 4293656576 Blocks: 19380440 IO Block: 4096 regular file
Device: 806h/2054d Inode: 12 Links: 1
Access: (0644/-rw-r--r--) Uid: ( 0/ root) Gid: ( 0/ root)
Access: 2011-07-25 13:01:31.414490496 +0900
Modify: 2011-07-25 13:01:31.414490496 +0900
Change: 2011-07-25 13:01:31.454490495 +0900
Signed-off-by: Utako Kusaka <u-kusaka@wm.jp.nec.com>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
--
fs/ext4/extents.c | 2 +-
1 files changed, 1 insertions(+), 1 deletions(-)
The old function ext4_ext_rm_idx is used only for truncate case
because it just remove last index in extent-index-block. When punching
hole, it usually needed to remove "middle" index, therefore we must
move indexes which after it forward.
(I create a file with 1 depth extent tree and punch hole in the middle
of it, the last index in index-block strangly gone, so I find out this
bug)
Signed-off-by: Robin Dong <sanbai@taobao.com>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
The reserve_backup_gdb() function only needs the block group number;
there's no need to pass a pointer to struct ext4_new_group_data to it.
Signed-off-by: Yongqiang Yang <xiaoqiangnk@gmail.com>
add_new_gdb() only needs the block group number; there is no need to
pass a pointer to struct ext4_new_group_data to add_new_gdb().
Instead of filling in a pointer the struct buffer_head in
add_new_gdb(), it's simpler to have the caller fetch it from the
s_group_desc[] array.
[Fixed error path to handle the case where struct buffer_head *primary
hasn't been set yet. -- Ted]
Signed-off-by: Yongqiang Yang <xiaoqiangnk@gmail.com>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
There is no need to lock the buffers since no one else should be
touching these buffers besides the file system.
Signed-off-by: Yongqiang Yang <xiaoqiangnk@gmail.com>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
This patch simplifies journal handling in setup_new_group_blocks().
In previous code, block bitmap is modified everywhere in
setup_new_group_blocks(), ext4_get_write_access() in
extend_or_restart_transaction() is used to guarantee that the block
bitmap stays in the new handle, this makes things complicated.
The previous commit changed things so that the modifications on the
block bitmap are batched and done by ext4_set_bits() at the end of the
for loop. This allows us to simplify things.
Signed-off-by: Yongqiang Yang <xiaoqiangnk@gmail.com>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
Rename mb_set_bits() to ext4_set_bits() and make it a global function
so that setup_new_group_blocks() can use it.
Signed-off-by: Yongqiang Yang <xiaoqiangnk@gmail.com>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
If ext4_group_add_blocks() is called with 0 block, make it return 0
without doing any extra work.
Signed-off-by: Yongqiang Yang <xiaoqiangnk@gmail.com>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
This patch lets ext4_group_add_blocks() return an error code if it
fails, so that upper functions can handle error correctly.
Signed-off-by: Yongqiang Yang <xiaoqiangnk@gmail.com>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
A filesystem with errors is not allowed to being resized, otherwise,
it is easy to destroy the filesystem.
Signed-off-by: Yongqiang Yang <xiaoqiangnk@gmail.com>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
Before this patch, parallel resizers are allowed and protected by a
mutex lock, actually, there is no need to support parallel resizer, so
this patch prevents parallel resizers by atmoic bit ops, like
lock_page() and unlock_page() do.
To do this, the patch removed the mutex lock s_resize_lock from struct
ext4_sb_info and added a unsigned long field named s_resize_flags
which inidicates if there is a resizer.
Signed-off-by: Yongqiang Yang <xiaoqiangnk@gmail.com>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
When journalling data for an inode (either because it is a symlink or
because the filesystem is mounted in data=journal mode), ext4_evict_inode()
can discard unwritten data by calling truncate_inode_pages(). This is
because we don't mark the buffer / page dirty when journalling data but only
add the buffer to the running transaction and thus mm does not know there
are still unwritten data.
Fix the problem by carefully tracking transaction containing inode's data,
committing this transaction, and writing uncheckpointed buffers when inode
should be reaped.
Signed-off-by: Jan Kara <jack@suse.cz>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
Replace the ->check_acl method with a ->get_acl method that simply reads an
ACL from disk after having a cache miss. This means we can replace the ACL
checking boilerplate code with a single implementation in namei.c.
Signed-off-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
new helper: posix_acl_create(&acl, gfp, mode_p). Replaces acl with
modified clone, on failure releases acl and replaces with NULL.
Returns 0 or -ve on error. All callers of posix_acl_create_masq()
switched.
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
new helper: posix_acl_chmod(&acl, gfp, mode). Replaces acl with modified
clone or with NULL if that has failed; returns 0 or -ve on error. All
callers of posix_acl_chmod_masq() switched to that - they'd been doing
exactly the same thing.
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
This moves logic for checking the cached ACL values from low-level
filesystems into generic code. The end result is a streamlined ACL
check that doesn't need to load the inode->i_op->check_acl pointer at
all for the common cached case.
The filesystems also don't need to check for a non-blocking RCU walk
case in their acl_check() functions, because that is all handled at a
VFS layer.
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
The debug message in ext4_ext_insert_extent before moving extent
is incorrect (the "from xx to xx").
Signed-off-by: Robin Dong <sanbai@taobao.com>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
The argument "inode" in function ext4_ext_next_allocated_block looks useless,
so clean it.
Signed-off-by: Robin Dong <sanbai@taobao.com>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
ac_repeats isn't referenced in the mballoc code. So remove it.
Signed-off-by: Tao Ma <boyu.mt@taobao.com>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
In ext4_mb_release, we use s_mb_buddies_generated++. Although
the output is OK, but I don't think we need this extra ++.
Signed-off-by: Tao Ma <boyu.mt@taobao.com>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
ext4_mb_load_buddy() calls ext4_get_group_info() for setting both
"grp" and "e4b->bd_info", but it could do "e4b->bd_info = grp".
Reported-by: Andreas Dilger <adilger@whamcloud.com>
Signed-off-by: Tao Ma <boyu.mt@taobao.com>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
Btrfs needs to be able to control how filemap_write_and_wait_range() is called
in fsync to make it less of a painful operation, so push down taking i_mutex and
the calling of filemap_write_and_wait() down into the ->fsync() handlers. Some
file systems can drop taking the i_mutex altogether it seems, like ext3 and
ocfs2. For correctness sake I just pushed everything down in all cases to make
sure that we keep the current behavior the same for everybody, and then each
individual fs maintainer can make up their mind about what to do from there.
Thanks,
Acked-by: Jan Kara <jack@suse.cz>
Signed-off-by: Josef Bacik <josef@redhat.com>
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
Since Ext4 has its own lseek we need to make sure it handles
SEEK_HOLE/SEEK_DATA. For now just do the same thing that is done in the generic
case, somebody else can come along and make it do fancy things later. Thanks,
Signed-off-by: Josef Bacik <josef@redhat.com>
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
For filesystems that delay their end_io processing we should keep our
i_dio_count until the the processing is done. Enable this by moving
the inode_dio_done call to the end_io handler if one exist. Note that
the actual move to the workqueue for ext4 and XFS is not done in
this patch yet, but left to the filesystem maintainers. At least
for XFS it's not needed yet either as XFS has an internal equivalent
to i_dio_count.
Signed-off-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
Simple filesystems always pass inode->i_sb_bdev as the block device
argument, and never need a end_io handler. Let's simply things for
them and for my grepping activity by dropping these arguments. The
only thing not falling into that scheme is ext4, which passes and
end_io handler without needing special flags (yet), but given how
messy the direct I/O code there is use of __blockdev_direct_IO
in one instead of two out of three cases isn't going to make a large
difference anyway.
Signed-off-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
Let filesystems handle waiting for direct I/O requests themselves instead
of doing it beforehand. This means filesystem-specific locks to prevent
new dio referenes from appearing can be held. This is important to allow
generalizing i_dio_count to non-DIO_LOCKING filesystems.
Signed-off-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
Rewrite ext4_page_mkwrite() to use __block_page_mkwrite() helper. This
removes the need of using i_alloc_sem to avoid races with truncate which
seems to be the wrong locking order according to lock ordering documented in
mm/rmap.c. Also calling ext4_da_write_begin() as used by the old code seems to
be problematic because we can decide to flush delay-allocated blocks which
will acquire s_umount semaphore - again creating unpleasant lock dependency
if not directly a deadlock.
Also add a check for frozen filesystem so that we don't busyloop in page fault
when the filesystem is frozen.
Signed-off-by: Jan Kara <jack@suse.cz>
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
This patch changes the security_inode_init_security API by adding a
filesystem specific callback to write security extended attributes.
This change is in preparation for supporting the initialization of
multiple LSM xattrs and the EVM xattr. Initially the callback function
walks an array of xattrs, writing each xattr separately, but could be
optimized to write multiple xattrs at once.
For existing security_inode_init_security() calls, which have not yet
been converted to use the new callback function, such as those in
reiserfs and ocfs2, this patch defines security_old_inode_init_security().
Signed-off-by: Mimi Zohar <zohar@us.ibm.com>
If eh_entries is equal to (or greater than) eh_max, the operation of
inserting new extent_idx will make number of entries overflow.
So check eh_entries before inserting the new extent_idx.
Although there is no bug case according the code (function
ext4_ext_insert_index is called by ext4_ext_split and ext4_ext_split
is called only if the index block has free space), the right logic
should be "lookup the capacity before insertion".
Signed-off-by: Robin Dong <sanbai@taobao.com>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
This patch avoids an extraneous lookup of the extent cache
in ext4_ext_map_blocks() when the flag
EXT4_GET_BLOCKS_PUNCH_OUT_EXT is absent.
The existing logic was performing the lookup but not making
use of the result. The patch simply reverses the order of evaluation
in the condition.
Since ext4_ext_in_cache() does not initialize newex on misses, bypassing
its invocation does not introduce any new issue in this regard.
Signed-off-by: Robin Dong <sanbai@taobao.com>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
Reviewed-by: Lukas Czerner <lczerner@redhat.com>
Reviewed-by: Eric Gouriou <egouriou@google.com>
This patch removes the extra parameter in ext4_ext_remove_space()
which is no longer needed.
Signed-off-by: Allison Henderson <achender@linux.vnet.ibm.com>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
This patch optimizes the punch hole operation by skipping the
tree walking code that is used by truncate. Since punch hole
is done through map blocks, the path to the extent is already
known in this function, so we do not need to look it up again.
Signed-off-by: Allison Henderson <achender@linux.vnet.ibm.com>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
If the stripe width was set to 1, then this patch will ignore
that stripe width and ext4 will act as if the stripe width
were 0 with respect to optimizing allocations.
Signed-off-by: Dan Ehrenberg <dehrenberg@google.com>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
Previously, if a stripe width was provided, then it would be used
as the preallocation granularity, with no santiy checking and no
way to override this. Now, mb_prealloc_size defaults to the smallest
multiple of stripe size that is greater than or equal to the old
default mb_prealloc_size, and this can be overridden with the sysfs
interface.
Signed-off-by: Dan Ehrenberg <dehrenberg@google.com>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
Compilation of ext4/namei.c brought up an error and warning messages
when compiled with -DDX_DEBUG
Signed-off-by: Bernd Schubert <bernd.schubert@itwm.fraunhofer.de>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
The comment from Al Viro about possible race in the ext4_orphan_add() is
not justified. There is no race possible as we always have either i_mutex
locked, or the inode can not be referenced from outside hence the
J_ASSERS should not be hit from the reason described in comment.
This commit replaces it with notion that we are holding i_mutex so it
should not be possible for i_nlink to be changed while waiting for
s_orphan_lock.
Signed-off-by: Lukas Czerner <lczerner@redhat.com>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
If we meet with an error in ext4_mb_add_groupinfo, we kfree
sbi->s_group_info[group >> EXT4_DESC_PER_BLOCK_BITS(sb)], but fail to
reset it to NULL. So the caller ext4_mb_init_backend will try to kfree
it again and causes a double free. So fix it by resetting it to NULL.
Some typo in comments of mballoc.c are also changed.
Signed-off-by: Tao Ma <boyu.mt@taobao.com>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
In ext4_groupinfo_create_slab, we create ext4_groupinfo_caches within
ext4_grpinfo_slab_create_mutex, but set it outside the lock, and there
does exist some case that we may create it twice and causes a memory
leak. So set it before we call mutex_unlock.
Signed-off-by: Tao Ma <boyu.mt@taobao.com>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
Optimize ext4_ext_insert_extent() by avoiding
ext4_ext_next_leaf_block() when the result is not used/needed.
Signed-off-by: Robin Dong <sanbai@taobao.com>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
If eh->eh_entries is smaller than eh->eh_max, the routine will
go to the "repeat" and then go to "has_space" directlly ,
since argument "depth" and "eh" are not even changed.
Therefore, goto "has_space" directly and remove redundant "repeat" tag.
Signed-off-by: Robin Dong <sanbai@taobao.com>
at ext4_trim_all_free() comment, there is no longer an @e4b parameter,
instead it is @group.
Reported-by: Andreas Dilger <adilger@dilger.ca>
Signed-off-by: Tao Ma <boyu.mt@taobao.com>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
In ext4, when FITRIM is called every time, we iterate all the
groups and do trim one by one. It is a bit time wasting if the
group has been trimmed and there is no change since the last
trim.
So this patch adds a new flag in ext4_group_info->bb_state to
indicate that the group has been trimmed, and it will be cleared
if some blocks is freed(in release_blocks_on_commit). Another
trim_minlen is added in ext4_sb_info to record the last minlen
we use to trim the volume, so that if the caller provide a small
one, we will go on the trim regardless of the bb_state.
A simple test with my intel x25m ssd:
df -h shows:
/dev/sdb1 40G 21G 17G 56% /mnt/ext4
Block size: 4096
run the FITRIM with the following parameter:
range.start = 0;
range.len = UINT64_MAX;
range.minlen = 1048576;
without the patch:
[root@boyu-tm linux-2.6]# time ./ftrim /mnt/ext4/a
real 0m5.505s
user 0m0.000s
sys 0m1.224s
[root@boyu-tm linux-2.6]# time ./ftrim /mnt/ext4/a
real 0m5.359s
user 0m0.000s
sys 0m1.178s
[root@boyu-tm linux-2.6]# time ./ftrim /mnt/ext4/a
real 0m5.228s
user 0m0.000s
sys 0m1.151s
with the patch:
[root@boyu-tm linux-2.6]# time ./ftrim /mnt/ext4/a
real 0m5.625s
user 0m0.000s
sys 0m1.269s
[root@boyu-tm linux-2.6]# time ./ftrim /mnt/ext4/a
real 0m0.002s
user 0m0.000s
sys 0m0.001s
[root@boyu-tm linux-2.6]# time ./ftrim /mnt/ext4/a
real 0m0.002s
user 0m0.000s
sys 0m0.001s
A big improvement for the 2nd and 3rd run.
Even after I delete some big image files, it is still much
faster than iterating the whole disk.
[root@boyu-tm test]# time ./ftrim /mnt/ext4/a
real 0m1.217s
user 0m0.000s
sys 0m0.196s
Cc: Lukas Czerner <lczerner@redhat.com>
Reviewed-by: Andreas Dilger <adilger.kernel@dilger.ca>
Signed-off-by: Tao Ma <boyu.mt@taobao.com>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
When we trim some free blocks in a group of ext4, we need to
calculate the free blocks properly and check whether there are
enough freed blocks left for us to trim. Current solution will
only calculate free spaces if they are large for a trim which
isn't appropriate.
Let us see a small example:
a group has 1.5M free which are 300k, 300k, 300k, 300k, 300k.
And minblocks is 1M. With current solution, we have to iterate
the whole group since these 300k will never be subtracted from
1.5M. But actually we should exit after we find the first 2
free spaces since the left 3 chunks only sum up to 900K if we
subtract the first 600K although they can't be trimed.
Reviewed-by: Andreas Dilger <adilger@dilger.ca>
Signed-off-by: Tao Ma <boyu.mt@taobao.com>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
In 0f0a25b, we adjust 'len' with s_first_data_block - start, but
it could underflow in case blocksize=1K, fstrim_range.len=512 and
fstrim_range.start = 0. In this case, when we run the code:
len -= first_data_blk - start; len will be underflow to -1ULL.
In the end, although we are safe that last_group check later will limit
the trim to the whole volume, but that isn't what the user really want.
So this patch fix it. It also adds the check for 'start' like ext3 so that
we can break immediately if the start is invalid.
Cc: Lukas Czerner <lczerner@redhat.com>
Signed-off-by: Tao Ma <boyu.mt@taobao.com>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
Upon corrupted inode or disk failures, we may fail after we already
allocate some blocks from the inode or take some blocks from the
inode's preallocation list, but before we successfully insert the
corresponding extent to the extent tree. In this case, we should free
any allocated blocks and discard the inode's preallocated blocks
because the entries in the inode's preallocation list may be in an
inconsistent state.
Signed-off-by: Jiaying Zhang <jiayingz@google.com>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
Cc: stable@kernel.org
The current implementation of ext4_free_blocks() always calls
dquot_free_block This looks quite sensible in the most cases: blocks
to be freed are associated with inode and were accounted in quota and
i_blocks some time ago.
However, there is a case when blocks to free were not accounted by the
time calling ext4_free_blocks() yet:
1. delalloc is on, write_begin pre-allocated some space in quota
2. write-back happens, ext4 allocates some blocks in ext4_ext_map_blocks()
3. then ext4_ext_map_blocks() gets an error (e.g. ENOSPC) from
ext4_ext_insert_extent() and calls ext4_free_blocks().
In this scenario, ext4_free_blocks() calls dquot_free_block() who, in
turn, decrements i_blocks for blocks which were not accounted yet (due
to delalloc) After clean umount, e2fsck reports something like:
> Inode 21, i_blocks is 5080, should be 5128. Fix<y>?
because i_blocks was erroneously decremented as explained above.
The patch fixes the problem by passing the new flag
EXT4_FREE_BLOCKS_NO_QUOT_UPDATE to ext4_free_blocks(), to request
that the dquot_free_block() call be skipped.
Signed-off-by: Maxim Patlasov <maxim.patlasov@gmail.com>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
Cc: stable@kernel.org
These days, bio_alloc() is guaranteed to never fail (as long as nvecs
is less than BIO_MAX_PAGES), so we don't need the loop around the
struct bio allocation.
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
I found that ext4_ext_find_goal() and ext4_find_near()
share the same code for returning a coloured start block
based on i_block_group.
We can refactor this into a common function so that they
don't diverge in the future.
Thanks to adilger for suggesting the new function name.
Signed-off-by: Eric Sandeen <sandeen@redhat.com>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
This patch moves functions from inode.c to indirect.c.
The moved functions are ext4_ind_* functions and their helpers.
Functions called from inode.c are declared extern.
Signed-off-by: Amir Goldstein <amir73il@users.sf.net>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
Move two functions that will be needed by the indirect functions to be
moved to indirect.c as well as inode.c to truncate.h as inline
functions, so that we can avoid having duplicate copies of the
function (which can be a maintenance problem) without having to expose
them as globally functions.
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
In preparation for moving the indirect functions to a separate file,
move __ext4_check_blockref() to block_validity.c and rename it to
ext4_check_blockref() which is exported as globally visible function.
Also, rename the cpp macro ext4_check_inode_blockref() to
ext4_ind_check_inode(), to make it clear that it is only valid for use
with non-extent mapped inodes.
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
We are going to move all ext4_ind_* functions to indirect.c.
Before we do that, let's rename 2 functions called ext4_indirect_*
to ext4_ind_*, to keep to the naming convention.
Signed-off-by: Amir Goldstein <amir73il@users.sf.net>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
We are about to move all indirect inode functions to a new file.
Before we do that, let's split ext4_ind_truncate() out of ext4_truncate()
leaving only generic code in the latter, so we will be able to move
ext4_ind_truncate() to the new file.
Signed-off-by: Amir Goldstein <amir73il@users.sf.net>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
In function ext4_ext_insert_index when eh_entries of curp is
bigger than eh_max, error messages will be printed out, but the content
is about logical and ei_block, that's incorret.
Signed-off-by: Robin Dong <sanbai@taobao.com>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
sync(2) is performed in two stages: the WB_SYNC_NONE sync and the
WB_SYNC_ALL sync. Identify the first stage with .tagged_writepages and
do livelock prevention for it, too.
Jan's commit f446daaea9 ("mm: implement writeback livelock avoidance
using page tagging") is a partial fix in that it only fixed the
WB_SYNC_ALL phase livelock.
Although ext4 is tested to no longer livelock with commit f446daaea9,
it may due to some "redirty_tail() after pages_skipped" effect which
is by no means a guarantee for _all_ the file systems.
Note that writeback_inodes_sb() is called by not only sync(), they are
treated the same because the other callers also need livelock prevention.
Impact: It changes the order in which pages/inodes are synced to disk.
Now in the WB_SYNC_NONE stage, it won't proceed to write the next inode
until finished with the current inode.
Acked-by: Jan Kara <jack@suse.cz>
CC: Dave Chinner <david@fromorbit.com>
Signed-off-by: Wu Fengguang <fengguang.wu@intel.com>
While creating fixed tracepoints for ext3, basically by porting them
from ext4, I found a lot of useless retyping, wrong type usage, useless
variable passing and other inconsistencies in the ext4 fixed tracepoint
code.
This patch cleans the fixed tracepoint code for ext4 and also simplify
some of them.
Signed-off-by: Lukas Czerner <lczerner@redhat.com>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
Currently we are not marking the extent as the last one
(FIEMAP_EXTENT_LAST) if there is a hole at the end of the file. This is
because we just do not check for it right now and continue searching for
next extent. But at the point we hit the hole at the end of the file, it
is too late.
This commit adds check for the allocated block in subsequent extent and
if there is no more extents (block = EXT_MAX_BLOCKS) just flag the
current one as the last one.
This behaviour has been spotted unintentionally by 252 xfstest, when the
test hangs out, because of wrong loop condition. However on other
filesystems (like xfs) it will exit anyway, because we notice the last
extent flag and exit.
With this patch xfstest 252 does not hang anymore, ext4 fiemap
implementation still reports bad extent type in some cases, however
this seems to be different issue.
Signed-off-by: Lukas Czerner <lczerner@redhat.com>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
Kazuya Mio reported that he was able to hit BUG_ON(next == lblock)
in ext4_ext_put_gap_in_cache() while creating a sparse file in extent
format and fill the tail of file up to its end. We will hit the BUG_ON
when we write the last block (2^32-1) into the sparse file.
The root cause of the problem lies in the fact that we specifically set
s_maxbytes so that block at s_maxbytes fit into on-disk extent format,
which is 32 bit long. However, we are not storing start and end block
number, but rather start block number and length in blocks. It means
that in order to cover extent from 0 to EXT_MAX_BLOCK we need
EXT_MAX_BLOCK+1 to fit into len (because we counting block 0 as well) -
and it does not.
The only way to fix it without changing the meaning of the struct
ext4_extent members is, as Kazuya Mio suggested, to lower s_maxbytes
by one fs block so we can cover the whole extent we can get by the
on-disk extent format.
Also in many places EXT_MAX_BLOCK is used as length instead of maximum
logical block number as the name suggests, it is all a bit messy. So
this commit renames it to EXT_MAX_BLOCKS and change its usage in some
places to actually be maximum number of blocks in the extent.
The bug which this commit fixes can be reproduced as follows:
dd if=/dev/zero of=/mnt/mp1/file bs=<blocksize> count=1 seek=$((2**32-2))
sync
dd if=/dev/zero of=/mnt/mp1/file bs=<blocksize> count=1 seek=$((2**32-1))
Reported-by: Kazuya Mio <k-mio@sx.jp.nec.com>
Signed-off-by: Lukas Czerner <lczerner@redhat.com>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
metadata is not parameter of ext4_free_blocks() any more.
Signed-off-by: Yongqiang Yang <xiaoqiangnk@gmail.com>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
Tell the filesystem if we just updated timestamp (I_DIRTY_SYNC) or
anything else, so that the filesystem can track internally if it
needs to push out a transaction for fdatasync or not.
This is just the prototype change with no user for it yet. I plan
to push large XFS changes for the next merge window, and getting
this trivial infrastructure in this window would help a lot to avoid
tree interdependencies.
Also remove incorrect comments that ->dirty_inode can't block. That
has been changed a long time ago, and many implementations rely on it.
Signed-off-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/djm/tmem:
xen: cleancache shim to Xen Transcendent Memory
ocfs2: add cleancache support
ext4: add cleancache support
btrfs: add cleancache support
ext3: add cleancache support
mm/fs: add hooks to support cleancache
mm: cleancache core ops functions and config
fs: add field to superblock to support cleancache
mm/fs: cleancache documentation
Fix up trivial conflict in fs/btrfs/extent_io.c due to includes
This seventh patch of eight in this cleancache series "opts-in"
cleancache for ext4. Filesystems must explicitly enable cleancache
by calling cleancache_init_fs anytime an instance of the filesystem
is mounted. For ext4, all other cleancache hooks are in
the VFS layer including the matching cleancache_flush_fs
hook which must be called on unmount.
Details and a FAQ can be found in Documentation/vm/cleancache.txt
[v6-v8: no changes]
[v5: jeremy@goop.org: simplify init hook and any future fs init changes]
Signed-off-by: Dan Magenheimer <dan.magenheimer@oracle.com>
Reviewed-by: Jeremy Fitzhardinge <jeremy@goop.org>
Reviewed-by: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com>
Acked-by: Andreas Dilger <adilger@sun.com>
Cc: Ted Ts'o <tytso@mit.edu>
Cc: Andrew Morton <akpm@linux-foundation.org>
Cc: Al Viro <viro@ZenIV.linux.org.uk>
Cc: Matthew Wilcox <matthew@wil.cx>
Cc: Nick Piggin <npiggin@kernel.dk>
Cc: Mel Gorman <mel@csn.ul.ie>
Cc: Rik Van Riel <riel@redhat.com>
Cc: Jan Beulich <JBeulich@novell.com>
Cc: Chris Mason <chris.mason@oracle.com>
Cc: Mark Fasheh <mfasheh@suse.com>
Cc: Joel Becker <joel.becker@oracle.com>
Cc: Nitin Gupta <ngupta@vflare.org>
Make ext4_ext_split() get extents to be moved by calculating in a statement
instead of counting in a loop.
Signed-off-by: Yongqiang Yang <xiaoqiangnk@gmail.com>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
Trivial conversion. Fixup one error handling case calling vmtruncate()
and remove ->truncate callback. We also fix a bug that IS_IMMUTABLE and
IS_APPEND files could not be truncated during failed writes. In fact, the
test can be completely removed as upper layers do necessary permission
checks for truncate in do_sys_[f]truncate() and may_open() anyway.
Signed-off-by: Jan Kara <jack@suse.cz>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
Currently, an fallocate request of size slightly larger than a power of
2 is turned into two block requests, each a power of 2, with the extra
blocks pre-allocated for future use. When an application calls
fallocate, it already has an idea about how large the file may grow so
there is usually little benefit to reserve extra blocks on the
preallocation list. This reduces disk fragmentation.
Tested: fsstress. Also verified manually that fallocat'ed files are
contiguously laid out with this change (whereas without it they begin at
power-of-2 boundaries, leaving blocks in between). CPU usage of
fallocate is not appreciably higher. In a tight fallocate loop, CPU
usage hovers between 5%-8% with this change, and 5%-7% without it.
Using a simulated file system aging program which the file system to
70%, the percentage of free extents larger than 8MB (as measured by
e2freefrag) increased from 38.8% without this change, to 69.4% with
this change.
Signed-off-by: Vivek Haldar <haldar@google.com>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
This patch adds new routines: "ext4_punch_hole" "ext4_ext_punch_hole"
and "ext4_ext_check_cache"
fallocate has been modified to call ext4_punch_hole when the punch hole
flag is passed. At the moment, we only support punching holes in
extents, so this routine is pretty much a wrapper for the ext4_ext_punch_hole
routine.
The ext4_ext_punch_hole routine first completes all outstanding writes
with the associated pages, and then releases them. The unblock
aligned data is zeroed, and all blocks in between are punched out.
The ext4_ext_check_cache routine is very similar to ext4_ext_in_cache
except it accepts a ext4_ext_cache parameter instead of a ext4_extent
parameter. This routine is used by ext4_ext_punch_hole to check and
see if a block in a hole that has been cached. The ext4_ext_cache
parameter is necessary because the members ext4_extent structure are
not large enough to hold a 32 bit value. The existing
ext4_ext_in_cache routine has become a wrapper to this new function.
[ext4 punch hole patch series 5/5 v7]
Signed-off-by: Allison Henderson <achender@us.ibm.com>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
Reviewed-by: Mingming Cao <cmm@us.ibm.com>
This patch adds a new flag to ext4_map_blocks() that specifies the
given range of blocks should be punched out. Extents are first
converted to uninitialized extents before they are punched
out. Because punching a hole may require that the extent be split, it
is possible that the splitting may need more blocks than are
available. To deal with this, use of reserved blocks are enabled to
allow the split to proceed.
The routine then returns the number of blocks successfully
punched out.
[ext4 punch hole patch series 4/5 v7]
Signed-off-by: Allison Henderson <achender@us.ibm.com>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
Reviewed-by: Mingming Cao <cmm@us.ibm.com>
This patch modifies the truncate routines to support hole punching
Below is a brief summary of the patches changes:
- Added end param to ext_ext4_rm_leaf
This function has been modified to accept an end parameter
which enables it to punch holes in leafs instead of just
truncating them.
- Implemented the "remove head" case in the ext_remove_blocks routine
This routine is used by ext_ext4_rm_leaf to remove the tail
of an extent during a truncate. The new ext_ext4_rm_leaf
routine will now also use it to remove the head of an extent in the
case that the hole covers a region of blocks at the beginning
of an extent.
- Added "end" param to ext4_ext_remove_space routine
This function has been modified to accept a stop parameter, which
is passed through to ext4_ext_rm_leaf.
[ext4 punch hole patch series 3/5 v6]
Signed-off-by: Allison Henderson <achender@us.ibm.com>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
This patch modifies the existing ext4_block_truncate_page() function
which was used by the truncate code path, and which zeroes out block
unaligned data, by adding a new length parameter, and renames it to
ext4_block_zero_page_rage(). This function can now be used to zero out the
head of a block, the tail of a block, or the middle
of a block.
The ext4_block_truncate_page() function is now a wrapper to
ext4_block_zero_page_range().
[ext4 punch hole patch series 2/5 v7]
Signed-off-by: Allison Henderson <achender@us.ibm.com>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
Reviewed-by: Mingming Cao <cmm@us.ibm.com>
This patch adds an allocation request flag to the ext4_has_free_blocks
function which enables the use of reserved blocks. This will allow a
punch hole to proceed even if the disk is full. Punching a hole may
require additional blocks to first split the extents.
Because ext4_has_free_blocks is a low level function, the flag needs
to be passed down through several functions listed below:
ext4_ext_insert_extent
ext4_ext_create_new_leaf
ext4_ext_grow_indepth
ext4_ext_split
ext4_ext_new_meta_block
ext4_mb_new_blocks
ext4_claim_free_blocks
ext4_has_free_blocks
[ext4 punch hole patch series 1/5 v7]
Signed-off-by: Allison Henderson <achender@us.ibm.com>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
Reviewed-by: Mingming Cao <cmm@us.ibm.com>
I am working on patch to add quota as a built-in feature for ext4
filesystem. The implementation is based on the design given at
https://ext4.wiki.kernel.org/index.php/Design_For_1st_Class_Quota_in_Ext4.
This patch reserves the inode numbers 3 and 4 for quota purposes and
also reserves EXT4_FEATURE_RO_COMPAT_QUOTA feature code.
Signed-off-by: Aditya Kali <adityakali@google.com>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
Prevent an ext4 filesystem from being mounted multiple times.
A sequence number is stored on disk and is periodically updated (every 5
seconds by default) by a mounted filesystem.
At mount time, we now wait for s_mmp_update_interval seconds to make sure
that the MMP sequence does not change.
In case of failure, the nodename, bdevname and the time at which the MMP
block was last updated is displayed.
Signed-off-by: Andreas Dilger <adilger@whamcloud.com>
Signed-off-by: Johann Lombardi <johann@whamcloud.com>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
I found the issue that the number of free blocks went negative.
# stat -f /mnt/mp1/
File: "/mnt/mp1/"
ID: e175ccb83a872efe Namelen: 255 Type: ext2/ext3
Block size: 4096 Fundamental block size: 4096
Blocks: Total: 258022 Free: -15 Available: -13122
Inodes: Total: 65536 Free: 63029
f_bfree in struct statfs will go negative when the filesystem has
few free blocks. Because the number of dirty blocks is bigger than
the number of free blocks in the following two cases.
CASE 1:
ext4_da_writepages
mpage_da_map_and_submit
ext4_map_blocks
ext4_ext_map_blocks
ext4_mb_new_blocks
ext4_mb_diskspace_used
percpu_counter_sub(&sbi->s_freeblocks_counter, ac->ac_b_ex.fe_len);
<--- interrupt statfs systemcall --->
ext4_da_update_reserve_space
percpu_counter_sub(&sbi->s_dirtyblocks_counter,
used + ei->i_allocated_meta_blocks);
CASE 2:
ext4_write_begin
__block_write_begin
ext4_map_blocks
ext4_ext_map_blocks
ext4_mb_new_blocks
ext4_mb_diskspace_used
percpu_counter_sub(&sbi->s_freeblocks_counter, ac->ac_b_ex.fe_len);
<--- interrupt statfs systemcall --->
percpu_counter_sub(&sbi->s_dirtyblocks_counter, reserv_blks);
To avoid the issue, this patch ensures that f_bfree is non-negative.
Signed-off-by: Kazuya Mio <k-mio@sx.jp.nec.com>
We should protect reading bd_info->bb_first_free with the group lock
because otherwise we might miss some free blocks. This is not a big deal
at all, but the change to do right thing is really simple, so lets do
that.
Signed-off-by: Lukas Czerner <lczerner@redhat.com>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
Currently we are loading buddy ext4_mb_load_buddy() for every block
group we are going through in ext4_trim_fs() in many cases just to find
out that there is not enough space to be bothered with. As Amir Goldstein
suggested we can use bb_free information directly from ext4_group_info.
This commit removes ext4_mb_load_buddy() from ext4_trim_fs() and rather
get the ext4_group_info via ext4_get_group_info() and use the bb_free
information directly from that. This avoids unnecessary call to load
buddy in the case the group does not have enough free space to trim.
Loading buddy is now moved to ext4_trim_all_free().
Tested by me with xfstests 251.
Signed-off-by: Lukas Czerner <lczerner@redhat.com>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
jbd2_log_start_commit() returns 1 only when we really start a
transaction. But we also need to wait for a transaction when the
commit is already running. Fix this problem by waiting for
transaction commit unconditionally (which is just a quick check if the
transaction is already committed).
Also we have to be more careful with sending of a barrier because when
transaction is being committed in parallel to ext4_sync_file()
running, we cannot be sure that the barrier the journalling code sends
happens after we wrote all the data for fsync (note that not every
data writeout needs to trigger metadata changes thus commit of some
metadata changes can be running while other data is still written
out). So use jbd2_will_send_data_barrier() helper to detect the common
cases when we can be sure barrier will be issued by the commit code
and issue the barrier ourselves in the remaining cases.
Reported-by: Edward Goggin <egoggin@vmware.com>
Signed-off-by: Jan Kara <jack@suse.cz>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
To get delayed-extent information, ext4_ext_fiemap_cb() looks up
pagecache, it thus collects information starting from a page's
head block.
If blocksize < pagesize, the beginning blocks of a page may lies
before the request range. So ext4_ext_fiemap_cb() should proceed
ignoring them, because they has been handled before. If no mapped
buffer in the range is found in the 1st page, we need to look up
the 2nd page, otherwise delayed-extents after a hole will be ignored.
Without this patch, xfstests 225 will hung on ext4 with 1K block.
Reported-by: Amir Goldstein <amir73il@users.sourceforge.net>
Signed-off-by: Yongqiang Yang <xiaoqiangnk@gmail.com>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
In commit c8d46e41 (ext4: Add flag to files with blocks intentionally
past EOF), if the EOFBLOCKS_FL flag is set, we call ext4_truncate()
before calling vmtruncate(). This caused any allocated but unwritten
blocks created by calling fallocate() with the FALLOC_FL_KEEP_SIZE
flag to be dropped. This was done to make to make sure that
EOFBLOCKS_FL would not be cleared while still leaving blocks past
i_size allocated. This was not necessary, since ext4_truncate()
guarantees that blocks past i_size will be dropped, even in the case
where truncate() has increased i_size before calling ext4_truncate().
So fix this by removing the EOFBLOCKS_FL special case treatment in
ext4_setattr(). In addition, use truncate_setsize() followed by a
call to ext4_truncate() instead of using vmtruncate(). This is more
efficient since it skips the call to inode_newsize_ok(), which has
been checked already by inode_change_ok(). This is also in a win in
the case where EOFBLOCKS_FL is set since it avoids calling
ext4_truncate() twice.
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
ext4_ext_truncate() should not invoke up_write(&EXT4_I(inode)->i_data_sem)
when ext4_orphan_add() returns an error, as it hasn't performed a
down_write() yet. This trivial patch fixes this by moving the up_write()
invocation above the out_stop label.
Signed-off-by: Eric Gouriou <egouriou@google.com>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
