On an SMP ARM system running ext4, I've received a report that the
first J_ASSERT in jbd2_journal_commit_transaction has been triggering:
J_ASSERT(journal->j_running_transaction != NULL);
While investigating possible causes for this problem, I noticed that
__jbd2_log_start_commit() is getting called with j_state_lock only
read-locked, in spite of the fact that it's possible for it might
j_commit_request. Fix this by grabbing the necessary information so
we can test to see if we need to start a new transaction before
dropping the read lock, and then calling jbd2_log_start_commit() which
will grab the write lock.
Signed-off-by: "Theodore Ts'o" <[email protected]>
Commit 40389687 moved a call to ext4_forget() out of
ext4_free_branches and let ext4_free_blocks() handle calling
bforget(). But that change unfortunately did not replace the call to
ext4_forget() with brelse(), which was needed to drop the in-use count
of the indirect block's buffer head, which lead to a memory leak when
deleting files that used indirect blocks. Fix this.
Thanks to Hugh Dickins for pointing this out.
Cc: [email protected]
Signed-off-by: "Theodore Ts'o" <[email protected]>
This function was never implemented, except for a BUG_ON which was
tripping when ext4 is run without a journal. The problem is that
although the comment asserts that "truncate (which is the only way to
free block) discards all preallocations", ext4_free_blocks() is also
called in various error recovery paths when blocks have been
allocated, but for various reasons, we were not able to use those data
blocks (for example, because we ran out of memory while trying to
manipulate the extent tree, or some other similar situation).
In addition to the fact that this function isn't implemented except
for the incorrect BUG_ON, the single caller of this function,
ext4_free_blocks(), doesn't use it all if the journal is enabled.
So remove the (stub) function entirely for now. If we decide it's
better to add it back, it's only going to be useful with a relatively
large number of code changes anyway.
Google-Bug-Id: 3236408
Cc: Jiaying Zhang <[email protected]>
Signed-off-by: "Theodore Ts'o" <[email protected]>
Call ext4_std_error() in various places when we can't bail out
cleanly, so the file system can be marked as in error.
Signed-off-by: "Theodore Ts'o" <[email protected]>
Replace the jbd2_inode structure (which is 48 bytes) with a pointer
and only allocate the jbd2_inode when it is needed --- that is, when
the file system has a journal present and the inode has been opened
for writing. This allows us to further slim down the ext4_inode_info
structure.
Signed-off-by: "Theodore Ts'o" <[email protected]>
We can store the dynamic inode state flags in the high bits of
EXT4_I(inode)->i_flags, and eliminate i_state_flags. This saves 8
bytes from the size of ext4_inode_info structure, which when
multiplied by the number of the number of in the inode cache, can save
a lot of memory.
Signed-off-by: "Theodore Ts'o" <[email protected]>
By reordering the elements in the ext4_inode_info structure, we can
reduce the padding needed on an x86_64 system by 16 bytes.
Signed-off-by: "Theodore Ts'o" <[email protected]>
We can encode the ec_type information by using ee_len == 0 to denote
EXT4_EXT_CACHE_NO, ee_start == 0 to denote EXT4_EXT_CACHE_GAP, and if
neither is true, then the cache type must be EXT4_EXT_CACHE_EXTENT.
This allows us to reduce the size of ext4_ext_inode by another 8
bytes. (ec_type is 4 bytes, plus another 4 bytes of padding)
Signed-off-by: "Theodore Ts'o" <[email protected]>
This fixes a number of places where we used sector_t instead of
ext4_lblk_t for logical blocks, which for ext4 are still 32-bit data
types. No point wasting space in the ext4_inode_info structure, and
requiring 64-bit arithmetic on 32-bit systems, when it isn't
necessary.
Signed-off-by: "Theodore Ts'o" <[email protected]>
Remove the short element i_delalloc_reserved_flag from the
ext4_inode_info structure and replace it a new bit in i_state_flags.
Since we have an ext4_inode_info for every ext4 inode cached in the
inode cache, any savings we can produce here is a very good thing from
a memory utilization perspective.
Signed-off-by: "Theodore Ts'o" <[email protected]>
Any time you see code that tries to add error codes together, you
should want to claw your eyes out...
Signed-off-by: "Theodore Ts'o" <[email protected]>
Use the search_dirblock() in ext3_dx_find_entry(). It makes the code
easier to read, and it takes advantage of common code. It also saves
100 bytes or so of text space.
Signed-off-by: "Theodore Ts'o" <[email protected]>
Cc: Brad Spengler <[email protected]>
Signed-off-by: Jan Kara <[email protected]>
If the first htree directory is missing '.' or '..' but is otherwise a
valid directory, and we do a lookup for '.' or '..', it's possible to
dereference an uninitialized memory pointer in ext3_htree_next_block().
Avoid this.
We avoid this by moving the special case from ext3_dx_find_entry() to
ext3_find_entry(); this also means we can optimize ext3_find_entry()
slightly when NFS looks up "..".
Thanks to Brad Spengler for pointing a Clang warning that led me to
look more closely at this code. The warning was harmless, but it was
useful in pointing out code that was too ugly to live. This warning was
also reported by Roman Borisov.
Signed-off-by: "Theodore Ts'o" <[email protected]>
Cc: Jan Kara <[email protected]>
Cc: Brad Spengler <[email protected]>
Signed-off-by: Jan Kara <[email protected]>
Create a linux-headers-$KVER.deb package which can be used to build
external modules without having the source tree around.
Signed-off-by: "Theodore Ts'o" <[email protected]>
Signed-off-by: Michal Marek <[email protected]>
https://bugzilla.kernel.org/show_bug.cgi?id=25352
This regression was caused by commit a31437b85: "ext4: use
sb_issue_zeroout in setup_new_group_blocks", by accidentally dropping
the code which reserved the block group descriptor and inode table
blocks.
Signed-off-by: "Theodore Ts'o" <[email protected]>
This function gets called a lot for large directories, and the answer
is almost always "no, no, there's no problem". This means using
unlikely() is a good thing.
Signed-off-by: "Theodore Ts'o" <[email protected]>
'buffer_head' should be 'journal_head'
This is a port of a patch which Namhyung Kim <[email protected]> made
to fs/jbd to jbd2.
Signed-off-by: "Theodore Ts'o" <[email protected]>
We had an open-coded version of printk_ratelimited(); use the provided
abstraction to make the code cleaner and easier to understand.
Based on a similar patch for fs/jbd from Namhyung Kim <[email protected]>
Signed-off-by: "Theodore Ts'o" <[email protected]>
This fixes up some broken argument descriptions that Namhyung Kim had
originally submitted for ext3. This fixes the comments that were
still applicable in ext4.
Signed-off-by: "Theodore Ts'o" <[email protected]>
Change clear_opt() and set_opt() to take a superblock pointer instead
of a pointer to EXT4_SB(sb)->s_mount_opt. This makes it easier for us
to support a second mount option field.
Signed-off-by: "Theodore Ts'o" <[email protected]>
Jon Nelson has found a test case which causes postgresql to fail with
the error:
psql:t.sql:4: ERROR: invalid page header in block 38269 of relation base/16384/16581
Under memory pressure, it looks like part of a file can end up getting
replaced by zero's. Until we can figure out the cause, we'll roll
back the change and use block_write_full_page() instead of
ext4_bio_write_page(). The new, more efficient writing function can
be used via the mount option mblk_io_submit, so we can test and fix
the new page I/O code.
To reproduce the problem, install postgres 8.4 or 9.0, and pin enough
memory such that the system just at the end of triggering writeback
before running the following sql script:
begin;
create temporary table foo as select x as a, ARRAY[x] as b FROM
generate_series(1, 10000000 ) AS x;
create index foo_a_idx on foo (a);
create index foo_b_idx on foo USING GIN (b);
rollback;
If the temporary table is created on a hard drive partition which is
encrypted using dm_crypt, then under memory pressure, approximately
30-40% of the time, pgsql will issue the above failure.
This patch should fix this problem, and the problem will come back if
the file system is mounted with the mblk_io_submit mount option.
Reported-by: Jon Nelson <[email protected]>
Signed-off-by: "Theodore Ts'o" <[email protected]>
Commit 5c521830cf (ext4: Support discard requests when running in
no-journal mode) attempts to add sb_issue_discard() for data blocks
(in data=writeback mode) and in no-journal mode. Unfortunately, this
no longer works, because in commit dd3932eddf (block: remove
BLKDEV_IFL_WAIT), sb_issue_discard() only presents a synchronous
interface, and there are times when we call ext4_free_blocks() when we
are are holding a spinlock, or are otherwise in an atomic context.
For now, I've removed the call to sb_issue_discard() to prevent a
deadlock or (if spinlock debugging is enabled) failures like this:
BUG: scheduling while atomic: rc.sysinit/1376/0x00000002
Pid: 1376, comm: rc.sysinit Not tainted 2.6.36-ARCH #1
Call Trace:
[<ffffffff810397ce>] __schedule_bug+0x5e/0x70
[<ffffffff81403110>] schedule+0x950/0xa70
[<ffffffff81060bad>] ? insert_work+0x7d/0x90
[<ffffffff81060fbd>] ? queue_work_on+0x1d/0x30
[<ffffffff81061127>] ? queue_work+0x37/0x60
[<ffffffff8140377d>] schedule_timeout+0x21d/0x360
[<ffffffff812031c3>] ? generic_make_request+0x2c3/0x540
[<ffffffff81402680>] wait_for_common+0xc0/0x150
[<ffffffff81041490>] ? default_wake_function+0x0/0x10
[<ffffffff812034bc>] ? submit_bio+0x7c/0x100
[<ffffffff810680a0>] ? wake_bit_function+0x0/0x40
[<ffffffff814027b8>] wait_for_completion+0x18/0x20
[<ffffffff8120a969>] blkdev_issue_discard+0x1b9/0x210
[<ffffffff811ba03e>] ext4_free_blocks+0x68e/0xb60
[<ffffffff811b1650>] ? __ext4_handle_dirty_metadata+0x110/0x120
[<ffffffff811b098c>] ext4_ext_truncate+0x8cc/0xa70
[<ffffffff810d713e>] ? pagevec_lookup+0x1e/0x30
[<ffffffff81191618>] ext4_truncate+0x178/0x5d0
[<ffffffff810eacbb>] ? unmap_mapping_range+0xab/0x280
[<ffffffff810d8976>] vmtruncate+0x56/0x70
[<ffffffff811925cb>] ext4_setattr+0x14b/0x460
[<ffffffff811319e4>] notify_change+0x194/0x380
[<ffffffff81117f80>] do_truncate+0x60/0x90
[<ffffffff811e08fa>] ? security_inode_permission+0x1a/0x20
[<ffffffff811eaec1>] ? tomoyo_path_truncate+0x11/0x20
[<ffffffff81127539>] do_last+0x5d9/0x770
[<ffffffff811278bd>] do_filp_open+0x1ed/0x680
[<ffffffff8140644f>] ? page_fault+0x1f/0x30
[<ffffffff81132bfc>] ? alloc_fd+0xec/0x140
[<ffffffff81118db1>] do_sys_open+0x61/0x120
[<ffffffff81118e8b>] sys_open+0x1b/0x20
[<ffffffff81002e6b>] system_call_fastpath+0x16/0x1b
https://bugzilla.kernel.org/show_bug.cgi?id=22302
Reported-by: Mathias Burén <[email protected]>
Signed-off-by: "Theodore Ts'o" <[email protected]>
Cc: [email protected]
Use an atomic_t and make sure we don't free the structure while we
might still be submitting I/O for that page.
Signed-off-by: "Theodore Ts'o" <[email protected]>
The following BUG can occur when an inode which is getting freed when
it still has dirty pages outstanding, and it gets deleted (in this
because it was the target of a rename). In ordered mode, we need to
make sure the data pages are written just in case we crash before the
rename (or unlink) is committed. If the inode is being freed then
when we try to igrab the inode, we end up tripping the BUG_ON at
fs/ext4/page-io.c:146.
To solve this problem, we need to keep track of the number of io
callbacks which are pending, and avoid destroying the inode until they
have all been completed. That way we don't have to bump the inode
count to keep the inode from being destroyed; an approach which
doesn't work because the count could have already been dropped down to
zero before the inode writeback has started (at which point we're not
allowed to bump the count back up to 1, since it's already started
getting freed).
Thanks to Dave Chinner for suggesting this approach, which is also
used by XFS.
kernel BUG at /scratch_space/linux-2.6/fs/ext4/page-io.c:146!
Call Trace:
[<ffffffff811075b1>] ext4_bio_write_page+0x172/0x307
[<ffffffff811033a7>] mpage_da_submit_io+0x2f9/0x37b
[<ffffffff811068d7>] mpage_da_map_and_submit+0x2cc/0x2e2
[<ffffffff811069b3>] mpage_add_bh_to_extent+0xc6/0xd5
[<ffffffff81106c66>] write_cache_pages_da+0x2a4/0x3ac
[<ffffffff81107044>] ext4_da_writepages+0x2d6/0x44d
[<ffffffff81087910>] do_writepages+0x1c/0x25
[<ffffffff810810a4>] __filemap_fdatawrite_range+0x4b/0x4d
[<ffffffff810815f5>] filemap_fdatawrite_range+0xe/0x10
[<ffffffff81122a2e>] jbd2_journal_begin_ordered_truncate+0x7b/0xa2
[<ffffffff8110615d>] ext4_evict_inode+0x57/0x24c
[<ffffffff810c14a3>] evict+0x22/0x92
[<ffffffff810c1a3d>] iput+0x212/0x249
[<ffffffff810bdf16>] dentry_iput+0xa1/0xb9
[<ffffffff810bdf6b>] d_kill+0x3d/0x5d
[<ffffffff810be613>] dput+0x13a/0x147
[<ffffffff810b990d>] sys_renameat+0x1b5/0x258
[<ffffffff81145f71>] ? _atomic_dec_and_lock+0x2d/0x4c
[<ffffffff810b2950>] ? cp_new_stat+0xde/0xea
[<ffffffff810b29c1>] ? sys_newlstat+0x2d/0x38
[<ffffffff810b99c6>] sys_rename+0x16/0x18
[<ffffffff81002a2b>] system_call_fastpath+0x16/0x1b
Reported-by: Nick Bowler <[email protected]>
Signed-off-by: "Theodore Ts'o" <[email protected]>
Tested-by: Nick Bowler <[email protected]>
We now initialize the percpu counters before replaying the journal,
but after the journal, we recalculate the global counters, to deal
with the possibility of the per-blockgroup counts getting updated by
the journal replay.
Signed-off-by: "Theodore Ts'o" <[email protected]>
Newer GCC's reported the following build warning:
fs/ext4/super.c: In function 'ext4_lazyinit_thread':
fs/ext4/super.c:2702: warning: 'ret' may be used uninitialized in this function
Fix it by removing the need for the ret variable in the first place.
Signed-off-by: "Lukas Czerner" <[email protected]>
Reported-by: "Stefan Richter" <[email protected]>
Signed-off-by: "Theodore Ts'o" <[email protected]>
Linus noted, and complained to me, that doing while lots of "git diff"'s
of kernel sources, these spinlocks were responsible for 27% of the
spinlock cost on his two-processor system as reported by perf.
Git was doing lots of parallel stats, and this was putting a lot of
pressure on ext4_getattr(). A spinlock to protect a single
memory-to-memory copy is pointless, so remove it.
Signed-off-by: "Theodore Ts'o" <[email protected]>
Signed-off-by: Linus Torvalds <[email protected]>
We need to make check if a page does not have buffes by checking
page_has_buffers(page) before calling page_buffers(page) in
ext4_writepage(). Otherwise page_buffers() could throw a BUG_ON.
Thanks also to Markus Trippelsdorf and Avinash Kurup who also reported
the problem.
Signed-off-by: "Theodore Ts'o" <[email protected]>
Reported-by: Sedat Dilek <[email protected]>
Tested-by: Sedat Dilek <[email protected]>
Unfortunately perf can't deal with anything other than direct structure
accesses in the TP_printk() section. It will drop dead when it sees
jbd2_dev_to_name() in the "print fmt" section of the tracepoint.
Addresses-Google-Bug: 3138508
Signed-off-by: "Theodore Ts'o" <[email protected]>
Cleanup namespace leaks from fs/ext4 and the inline trivial functions
ext4_{ext,idx}_pblock() and ext4_{ext,idx}_store_pblock() since the
code size actually shrinks when we make these functions inline,
they're so trivial.
Signed-off-by: "Theodore Ts'o" <[email protected]>
These functions have no need to be exported beyond file context.
No functions needed to be moved for this commit; just some function
declarations changed to be static and removed from header files.
(A similar patch was submitted by Eric Sandeen, but I wanted to handle
code movement in separate patches to make sure code changes didn't
accidentally get dropped.)
Signed-off-by: Eric Sandeen <[email protected]>
Signed-off-by: "Theodore Ts'o" <[email protected]>
Commit 84061e0 fixed an accounting bug only to introduce the
possibility of a kernel OOPS if the journal has a non-zero j_errno
field indicating that the file system had detected a fs inconsistency.
After the journal replay, if the journal superblock indicates that the
file system has an error, this indication is transfered to the file
system and then ext4_commit_super() is called to write this to the
disk.
But since the percpu counters are now initialized after the journal
replay, the call to ext4_commit_super() will cause a kernel oops since
it needs to use the percpu counters the ext4 superblock structure.
The fix is to skip setting the ext4 free block and free inode fields
if the percpu counter has not been set.
Thanks to Ken Sumrall for reporting and analyzing the root causes of
this bug.
Addresses-Google-Bug: #3054080
Signed-off-by: "Theodore Ts'o" <[email protected]>
Call the block I/O layer directly instad of going through the buffer
layer. This should give us much better performance and scalability,
as well as lowering our CPU utilization when doing buffered writeback.
Signed-off-by: "Theodore Ts'o" <[email protected]>
This massively simplifies the ext4_da_writepages() code path by
completely removing mpage_put_bnr_bhs(), which is almost 100 lines of
code iterating over a set of pages using pagevec_lookup(), and folds
that functionality into mpage_da_submit_io()'s existing
pagevec_lookup() loop.
Signed-off-by: "Theodore Ts'o" <[email protected]>
Expand the call:
if (walk_page_buffers(NULL, page_bufs, 0, len, NULL,
ext4_bh_delay_or_unwritten))
goto redirty_page
into mpage_da_submit_io().
This will allow us to merge in mpage_put_bnr_to_bhs() in the next
patch.
Signed-off-by: "Theodore Ts'o" <[email protected]>
As a prepratory step to switching to bio_submit, inline
ext4_writepage() into mpage_da_submit() and then simplify things a
bit. This makes it clearer what mpage_da_submit needs to do.
Also, move the ClearPageChecked(page) call into
__ext4_journalled_writepage(), as a minor bit of cleanup refactoring.
This also allows us to pull i_size_read() and
ext4_should_journal_data() out of the loop, which should be a very
minor CPU savings.
Signed-off-by: "Theodore Ts'o" <[email protected]>
The actual code in ext4_writepage() is unnecessarily convoluted.
Simplify it so it is easier to understand, but otherwise logically
equivalent.
Signed-off-by: "Theodore Ts'o" <[email protected]>
Eventually we need to completely reorganize the ext4 writepage
callpath, but for now, we simplify things a little by calling
mpage_da_submit_io() from mpage_da_map_blocks(), since all of the
places where we call mpage_da_map_blocks() it is followed up by a call
to mpage_da_submit_io().
We're also a wee bit better with respect to error handling, but there
are still a number of issues where it's not clear what the right thing
is to do with ext4 functions deep in the writeback codepath fails.
Signed-off-by: "Theodore Ts'o" <[email protected]>
Also remove the SLAB_RECLAIM_ACCOUNT flag from the system zone kmem
cache. This slab tends to be fairly static, so it shouldn't be marked
as likely to have free pages that can be reclaimed.
Signed-off-by: "Theodore Ts'o" <[email protected]>
Use the search_dirblock() in ext4_dx_find_entry(). It makes the code
easier to read, and it takes advantage of common code. It also saves
100 bytes or so of text space.
Signed-off-by: "Theodore Ts'o" <[email protected]>
Cc: Brad Spengler <[email protected]>
If the first block of htree directory is missing '.' or '..' but is
otherwise a valid directory, and we do a lookup for '.' or '..', it's
possible to dereference an uninitialized memory pointer in
ext4_htree_next_block().
We avoid this by moving the special case from ext4_dx_find_entry() to
ext4_find_entry(); this also means we can optimize ext4_find_entry()
slightly when NFS looks up "..".
Thanks to Brad Spengler for pointing a Clang warning that led me to
look more closely at this code. The warning was harmless, but it was
useful in pointing out code that was too ugly to live. This warning was
also reported by Roman Borisov.
Signed-off-by: "Theodore Ts'o" <[email protected]>
Cc: Brad Spengler <[email protected]>
An attempt to modify the file system during the call to
jbd2_destroy_journal() can lead to a system lockup. So add some
checking to make it much more obvious when this happens to and to
determine where the offending code is located.
Signed-off-by: "Theodore Ts'o" <[email protected]>
It turns out we have several problems with how EOFBLOCKS_FL is
handled. First of all, there was a fencepost error where we were not
clearing the EOFBLOCKS_FL when fill in the last uninitialized block,
but rather when we allocate the next block _after_ the uninitalized
block. Secondly we were not testing to see if we needed to clear the
EOFBLOCKS_FL when writing to the file O_DIRECT or when were converting
an uninitialized block (which is the most common case).
Google-Bug-Id: 2928259
Signed-off-by: "Theodore Ts'o" <[email protected]>
This should remove the last exclusive lock from start_this_handle(),
so that we should now be able to start multiple transactions at the
same time on large SMP systems.
Signed-off-by: "Theodore Ts'o" <[email protected]>
Lockstat reports have shown that j_state_lock is a major source of
lock contention, especially on systems with more than 4 CPU cores. So
change it to be a read/write spinlock.
Signed-off-by: "Theodore Ts'o" <[email protected]>
By using an atomic_t for t_updates and t_outstanding credits, this
should allow us to not need to take transaction t_handle_lock in
jbd2_journal_stop().
Signed-off-by: "Theodore Ts'o" <[email protected]>
Allow mount options to be stored in the superblock. Also add default
mount option bits for nobarrier, block_validity, discard, and nodelalloc.
Signed-off-by: "Theodore Ts'o" <[email protected]>
There were some error paths in ext4_delete_inode() which was not
dropping the inode from the orphan list. This could lead to a BUG_ON
on umount when the orphan list is discovered to be non-empty.
Signed-off-by: "Theodore Ts'o" <[email protected]>
There are some drivers which may not set bdev->bd_dev. So make sure
it is non-NULL before dereferencing it.
Google-Bug-Id: 1773557
Signed-off-by: "Theodore Ts'o" <[email protected]>
If the user attempts to make a non-extent-mapped file to be too large,
return EFBIG, but don't call ext4_std_err() which will end up marking
the file system as containing an error.
Thanks to Toshiyuki Okajima-san at Fujitsu for pointing this out.
Signed-off-by: "Theodore Ts'o" <[email protected]>
__GFP_NOFAIL is going away, so add our own retry loop. Also add
jbd2__journal_start() and jbd2__journal_restart() which take a gfp
mask, so that file systems can optionally (re)start transaction
handles using GFP_KERNEL. If they do this, then they need to be
prepared to handle receiving an PTR_ERR(-ENOMEM) error, and be ready
to reflect that error up to userspace.
Signed-off-by: "Theodore Ts'o" <[email protected]>
This allows us to grab any file system error messages by scraping
/var/log/messages. This will make it easy for us to do error analysis
across the very large number of machines as we deploy ext4 across the
fleet.
Signed-off-by: "Theodore Ts'o" <[email protected]>
Save number of file system errors, and the time function name, line
number, block number, and inode number of the first and most recent
errors reported on the file system in the superblock.
Signed-off-by: "Theodore Ts'o" <[email protected]>
Also start passing the line number to ext4_check_dir since we're going
to need it in upcoming patch.
Signed-off-by: "Theodore Ts'o" <[email protected]>
We don't need to set s_dirt in most of the ext4 code when journaling
is enabled. In ext3/4 some of the summary statistics for # of free
inodes, blocks, and directories are calculated from the per-block
group statistics when the file system is mounted or unmounted. As a
result the superblock doesn't have to be updated, either via the
journal or by setting s_dirt. There are a few exceptions, most
notably when resizing the file system, where the superblock needs to
be modified --- and in that case it should be done as a journalled
operation if possible, and s_dirt set only in no-journal mode.
This patch will optimize out some unneeded disk writes when using ext4
with a journal.
Signed-off-by: "Theodore Ts'o" <[email protected]>
Dan Roseberg has reported a problem with the MOVE_EXT ioctl. If the
donor file is an append-only file, we should not allow the operation
to proceed, lest we end up overwriting the contents of an append-only
file.
Signed-off-by: "Theodore Ts'o" <[email protected]>
Cc: Dan Rosenberg <[email protected]>
Dimitry Monakhov discovered an edge case where it was possible for the
EXT4_EOFBLOCKS_FL flag could get cleared unnecessarily. This is true;
I have a test case that can be exercised via downloading and
decompressing the file:
wget ftp://ftp.kernel.org/pub/linux/kernel/people/tytso/ext4-testcases/eofblocks-fl-test-case.img.bz2
bunzip2 eofblocks-fl-test-case.img
dd if=/dev/zero of=eofblocks-fl-test-case.img bs=1k seek=17925 bs=1k count=1 conv=notrunc
However, triggering it in real life is highly unlikely since it
requires an extremely fragmented sparse file with a hole in exactly
the right place in the extent tree. (It actually took quite a bit of
work to generate this test case.) Still, it's nice to get even
extreme corner cases to be correct, so this patch makes sure that we
don't clear the EXT4_EOFBLOCKS_FL incorrectly even in this corner
case.
Signed-off-by: "Theodore Ts'o" <[email protected]>
If the EOFBLOCK_FL flag is set when it should not be and the inode is
zero length, then eh_entries is zero, and ex is NULL, so dereferencing
ex to print ex->ee_block causes a kernel OOPS in
ext4_ext_map_blocks().
On top of that, the error message which is printed isn't very helpful.
So we fix this by printing something more explanatory which doesn't
involve trying to print ex->ee_block.
Addresses-Google-Bug: #2655740
Signed-off-by: "Theodore Ts'o" <[email protected]>
EXT4_ERROR_INODE() tends to provide better error information and in a
more consistent format. Some errors were not even identifying the inode
or directory which was corrupted, which made them not very useful.
Addresses-Google-Bug: #2507977
Signed-off-by: "Theodore Ts'o" <[email protected]>
This saves a huge amount of stack space by avoiding unnecesary struct
buffer_head's from being allocated on the stack.
In addition, to make the code easier to understand, collapse and
refactor ext4_get_block(), ext4_get_block_write(),
noalloc_get_block_write(), into a single function.
Signed-off-by: "Theodore Ts'o" <[email protected]>
Jack up ext4_get_blocks() and add a new function, ext4_map_blocks()
which uses a much smaller structure, struct ext4_map_blocks which is
20 bytes, as opposed to a struct buffer_head, which nearly 5 times
bigger on an x86_64 machine. By switching things to use
ext4_map_blocks(), we can save stack space by using ext4_map_blocks()
since we can avoid allocating a struct buffer_head on the stack.
Signed-off-by: "Theodore Ts'o" <[email protected]>
Make a copy of write_cache_pages() for the benefit of
ext4_da_writepages(). This allows us to simplify the code some, and
will allow us to further customize the code in future patches.
There are some nasty hacks in write_cache_pages(), which Linus has
(correctly) characterized as vile. I've just copied it into
write_cache_pages_da(), without trying to clean those bits up lest I
break something in the ext4's delalloc implementation, which is a bit
fragile right now. This will allow Dave Chinner to clean up
write_cache_pages() in mm/page-writeback.c, without worrying about
breaking ext4. Eventually write_cache_pages_da() will go away when I
rewrite ext4's delayed allocation and create a general
ext4_writepages() which is used for all of ext4's writeback. Until
now this is the lowest risk way to clean up the core
write_cache_pages() function.
Signed-off-by: "Theodore Ts'o" <[email protected]>
Cc: Dave Chinner <[email protected]>
One of the most contended locks in the jbd2 layer is j_state_lock when
running dbench. This is especially true if using the real-time kernel
with its "sleeping spinlocks" patch that replaces spinlocks with
priority inheriting mutexes --- but it also shows up on large SMP
benchmarks.
Thanks to John Stultz for pointing this out.
Reviewed by Mingming Cao and Jan Kara.
Signed-off-by: "Theodore Ts'o" <[email protected]>
At several places we modify EXT4_I(inode)->i_state without holding
i_mutex (ext4_release_file, ext4_bmap, ext4_journalled_writepage,
ext4_do_update_inode, ...). These modifications are racy and we can
lose updates to i_state. So convert handling of i_state to use bitops
which are atomic.
Cc: Jan Kara <[email protected]>
Signed-off-by: "Theodore Ts'o" <[email protected]>
In the past, ext4_calc_metadata_amount(), and its sub-functions
ext4_ext_calc_metadata_amount() and ext4_indirect_calc_metadata_amount()
badly over-estimated the number of metadata blocks that might be
required for delayed allocation blocks. This didn't matter as much
when functions which managed the reserved metadata blocks were more
aggressive about dropping reserved metadata blocks as delayed
allocation blocks were written, but unfortunately they were too
aggressive. This was fixed in commit 0637c6f, but as a result the
over-estimation by ext4_calc_metadata_amount() would lead to reserving
2-3 times the number of pending delayed allocation blocks as
potentially required metadata blocks. So if there are 1 megabytes of
blocks which have been not yet been allocation, up to 3 megabytes of
space would get reserved out of the user's quota and from the file
system free space pool until all of the inode's data blocks have been
allocated.
This commit addresses this problem by much more accurately estimating
the number of metadata blocks that will be required. It will still
somewhat over-estimate the number of blocks needed, since it must make
a worst case estimate not knowing which physical blocks will be
needed, but it is much more accurate than before.
Signed-off-by: "Theodore Ts'o" <[email protected]>
Commit 0637c6f had a typo which caused the reserved metadata blocks to
not be released correctly. Fix this.
Signed-off-by: "Theodore Ts'o" <[email protected]>
As reported in Kernel Bugzilla #14936, commit d21cd8f triggered a BUG
in the function ext4_da_update_reserve_space() found in
fs/ext4/inode.c. The root cause of this BUG() was caused by the fact
that ext4_calc_metadata_amount() can severely over-estimate how many
metadata blocks will be needed, especially when using direct
block-mapped files.
In addition, it can also badly *under* estimate how much space is
needed, since ext4_calc_metadata_amount() assumes that the blocks are
contiguous, and this is not always true. If the application is
writing blocks to a sparse file, the number of metadata blocks
necessary can be severly underestimated by the functions
ext4_da_reserve_space(), ext4_da_update_reserve_space() and
ext4_da_release_space(). This was the cause of the dq_claim_space
reports found on kerneloops.org.
Unfortunately, doing this right means that we need to massively
over-estimate the amount of free space needed. So in some cases we
may need to force the inode to be written to disk asynchronously in
to avoid spurious quota failures.
http://bugzilla.kernel.org/show_bug.cgi?id=14936
Signed-off-by: "Theodore Ts'o" <[email protected]>
This is a bit complicated because we are trying to optimize when we
send barriers to the fs data disk. We could just throw in an extra
barrier to the data disk whenever we send a barrier to the journal
disk, but that's not always strictly necessary.
We only need to send a barrier during a commit when there are data
blocks which are must be written out due to an inode written in
ordered mode, or if fsync() depends on the commit to force data blocks
to disk. Finally, before we drop transactions from the beginning of
the journal during a checkpoint operation, we need to guarantee that
any blocks that were flushed out to the data disk are firmly on the
rust platter before we drop the transaction from the journal.
Thanks to Oleg Drokin for pointing out this flaw in ext3/ext4.
Signed-off-by: "Theodore Ts'o" <[email protected]>
Add module aliases for ext2 and ext3 when CONFIG_EXT4_USE_FOR_EXT23 is
set. This makes the existing user-space stuff like mkinitrd working
as is.
Signed-off-by: Takashi Iwai <[email protected]>
Signed-off-by: "Theodore Ts'o" <[email protected]>
Theodore Ts'o