Add the facility for ext4_forget() to be called from
ext4_free_blocks().  This simplifies the code in a large number of
places, and centralizes most of the work of calling ext4_forget() into
a single place.

Also fix a bug in the extents migration code; it wasn't calling
ext4_forget() when releasing the indirect blocks during the
conversion.  As a result, if the system cashed during or shortly after
the extents migration, and the released indirect blocks get reused as
data blocks, the journal replay would corrupt the data blocks.  With
this new patch, fixing this bug was as simple as adding the
EXT4_FREE_BLOCKS_FORGET flags to the call to ext4_free_blocks().
Signed-off-by: N"Theodore Ts'o" <tytso@mit.edu>
Cc: "Aneesh Kumar K.V" <aneesh.kumar@linux.vnet.ibm.com>

ext4_mb_free_blocks() is only called by ext4_free_blocks(), and the
latter function doesn't really do much.  So merge the two functions
together, such that ext4_free_blocks() is now found in
fs/ext4/mballoc.c.  This saves about 200 bytes of compiled text space.
Signed-off-by: N"Theodore Ts'o" <tytso@mit.edu>

It is anticipated that when sb_issue_discard starts doing
real work on trim-capable devices, we may see issues.  Make
this mount-time optional, and default it to off until we know
that things are working out OK.
Signed-off-by: NEric Sandeen <sandeen@redhat.com>
Signed-off-by: N"Theodore Ts'o" <tytso@mit.edu>

The /proc/fs/ext4/<dev>/mb_history was maintained manually, and had a
number of problems: it required a largish amount of memory to be
allocated for each ext4 filesystem, and the s_mb_history_lock
introduced a CPU contention problem.  

By ripping out the mb_history code and replacing it with ftrace
tracepoints, and we get more functionality: timestamps, event
filtering, the ability to correlate mballoc history with other ext4
tracepoints, etc.
Signed-off-by: N"Theodore Ts'o" <tytso@mit.edu>

There are a number of kernel printk's which are printed when an ext4
filesystem is mounted and unmounted.  Disable them to economize space
in the system logs.  In addition, disabling the mballoc stats by
default saves a number of unneeded atomic operations for every block
allocation or deallocation.
Signed-off-by: N"Theodore Ts'o" <tytso@mit.edu>

The hueristic was designed to avoid using locality group preallocation
when writing the last segment of a closed file.  Fix it by move
setting size to the maximum of size and isize until after we check
whether size == isize.
Signed-off-by: N"Theodore Ts'o" <tytso@mit.edu>

Today, the ext4 allocator will happily allocate blocks past
2^32 for indirect-block files, which results in the block
numbers getting truncated, and corruption ensues.

This patch limits such allocations to < 2^32, and adds
BUG_ONs if we do get blocks larger than that.

This should address RH Bug 519471, ext4 bitmap allocator 
must limit blocks to < 2^32

* ext4_find_goal() is modified to choose a goal < UINT_MAX,
  so that our starting point is in an acceptable range.

* ext4_xattr_block_set() is modified such that the goal block
  is < UINT_MAX, as above.

* ext4_mb_regular_allocator() is modified so that the group
  search does not continue into groups which are too high

* ext4_mb_use_preallocated() has a check that we don't use
  preallocated space which is too far out

* ext4_alloc_blocks() and ext4_xattr_block_set() add some BUG_ONs

No attempt has been made to limit inode locations to < 2^32,
so we may wind up with blocks far from their inodes.  Doing
this much already will lead to some odd ENOSPC issues when the
"lower 32" gets full, and further restricting inodes could
make that even weirder.

For high inodes, choosing a goal of the original, % UINT_MAX,
may be a bit odd, but then we're in an odd situation anyway,
and I don't know of a better heuristic.
Signed-off-by: NEric Sandeen <sandeen@redhat.com>
Signed-off-by: N"Theodore Ts'o" <tytso@mit.edu>

We don't need to take the alloc_sem lock when we are adding new
groups, since mballoc won't see the new group added until we bump
sbi->s_groups_count.
Signed-off-by: N"Theodore Ts'o" <tytso@mit.edu>
Signed-off-by: NAneesh Kumar K.V <aneesh.kumar@linux.vnet.ibm.com>

We should check for need init flag with the group's alloc_sem held, to
make sure while we are loading the buddy cache and holding a reference
to it, a file system resize can't add new blocks to same group.

The patch also drops the need init flag check in
ext4_mb_regular_allocator() because doing the check without holding
alloc_sem is racy.
Signed-off-by: N"Theodore Ts'o" <tytso@mit.edu>
Signed-off-by: NAneesh Kumar K.V <aneesh.kumar@linux.vnet.ibm.com>

This moves the function around so that it can be called from
ext4_mb_load_buddy().
Signed-off-by: NAneesh Kumar K.V <aneesh.kumar@linux.vnet.ibm.com>
Signed-off-by: N"Theodore Ts'o" <tytso@mit.edu>

Signed-off-by: NTobias Klauser <tklauser@distanz.ch>
Signed-off-by: N"Theodore Ts'o" <tytso@mit.edu>

unsigned  short is potentially too small to track blocks within
a group; today it is safe due to restrictions in e2fsprogs but
we have _lo / _hi bits for group blocks with the intent to go
up to 32 bits, so clean this up now.

There are many more places where we use unsigned/int/unsigned int
to contain a group block but this should at least fix all the
short types.

I added a few comments to the struct ext4_group_info definition
as well.
Signed-off-by: NEric Sandeen <sandeen@redhat.com>
Signed-off-by: N"Theodore Ts'o" <tytso@mit.edu>

Precursor to changing some types; to keep things in sync, it 
seems better to allocate/memset based on the size of the 
variables we are using rather than on some disconnected 
basic type like "unsigned short"
Signed-off-by: NEric Sandeen <sandeen@redhat.com>

While reading through some of the mballoc code it seems that a couple
spots in the size normalization function could be streamlined.

The test for non-overlapping PAs can be or'd for the start & end
conditions, and the tests for adjacent PAs can be else-if'd - 
it's essentially independently testing:

	if (A + B <= C)
		...
	if (A > C)
		...

These cannot both be true so it seems like the else-if might
be slightly more efficient and/or informative.
Signed-off-by: NEric Sandeen <sandeen@redhat.com>
Signed-off-by: N"Theodore Ts'o" <tytso@mit.edu>

ext4_mb_update_group_info is only called in one place, and it's
extremely simple.  There's no reason to have it in a separate function
in a separate file as far as I can tell, it just obfuscates what's
really going on.

Perhaps it was intended to keep the grp->bb_* manipulation local to
mballoc.c but we're already accessing other grp-> fields in balloc.c
directly so this seems ok.
Signed-off-by: NEric Sandeen <sandeen@redhat.com>
Signed-off-by: N"Theodore Ts'o" <tytso@mit.edu>

Currently the group preallocation code tries to find a large (512)
free block from which to do per-cpu group allocation for small files.
The problem with this scheme is that it leaves the filesystem horribly
fragmented.  In the worst case, if the filesystem is unmounted and
remounted (after a system shutdown, for example) we forget the fact
that wee were using a particular (now-partially filled) 512 block
extent.  So the next time we try to allocate space for a small file,
we will find *another* completely free 512 block chunk to allocate
small files.  Given that there are 32,768 blocks in a block group,
after 64 iterations of "mount, write one 4k file in a directory,
unmount", the block group will have 64 files, each separated by 511
blocks, and the block group will no longer have any free 512
completely free chunks of blocks for group preallocation space.

So if we try to allocate blocks for a file that has been closed, such
that we know the final size of the file, and the filesystem is not
busy, avoid using group preallocation.
Signed-off-by: N"Theodore Ts'o" <tytso@mit.edu>

The logic around sbi->s_mb_last_group and sbi->s_mb_last_start was all
screwed up.  These fields were getting unconditionally all the time,
set even when stream allocation had not taken place, and if they were
being used when the file was smaller than s_mb_stream_request, which
is when the allocation should _not_ be doing stream allocation.

Fix this by determining whether or not we stream allocation should
take place once, in ext4_mb_group_or_file(), and setting a flag which
gets used in ext4_mb_regular_allocator() and ext4_mb_use_best_found().
This simplifies the code and assures that we are consistently using
(or not using) the stream allocation logic.
Signed-off-by: N"Theodore Ts'o" <tytso@mit.edu>

Displaying the flags in base 16 makes it easier to see which flags
have been set.
Signed-off-by: N"Theodore Ts'o" <tytso@mit.edu>

Allow mballoc debugging to be enabled at run-time instead of just at
compile time.
Signed-off-by: N"Theodore Ts'o" <tytso@mit.edu>

Pavel Roskin pointed out that kmemcheck indicated that
ext4_mb_store_history() was accessing uninitialized values of
ac->ac_tail and ac->ac_buddy leading to garbage in the mballoc
history.  Fix this by initializing the entire structure to all zeros
first.

Also, two fields were getting doubly initialized by the caller of
ext4_mb_initialize_context, so remove them for efficiency's sake.
Signed-off-by: N"Theodore Ts'o" <tytso@mit.edu>

When MB_DEBUG is enabled, we get some compile warnings because
ext4_group_t is unsigned int.  This patch fixes them.

Signed-off-by Akira Fujita <a-fujita@rs.jp.nec.com>
Signed-off-by: N"Theodore Ts'o" <tytso@mit.edu>

Signed-off-by: NJoe Perches <joe@perches.com>
Signed-off-by: N"Theodore Ts'o" <tytso@mit.edu>

The allocation of the ext4_group_info array was moved to a new
function ext4_mb_add_group_info() in commit 5f21b0e6 so that online
resize would use a common (and correct) codepath.  Unfortunately, the
call to the new ext4_mb_add_group_info() function was added without
removing the code which originally allocated the array.  This caused a
memory leak each time an ext4 filesystem was mounted.

The fix is simple; remove the code that did the original allocation,
since it is no longer needed.
Reported-by: NCatalin Marinas <catalin.marinas@arm.com>
Tested-by: NCatalin Marinas <catalin.marinas@arm.com>
Signed-off-by: NAneesh Kumar K.V <aneesh.kumar@linux.vnet.ibm.com>
Signed-off-by: N"Theodore Ts'o" <tytso@mit.edu>

The function ext4_mb_free_blocks() was using an "unsigned long" to
pass a block number; this will cause 64-bit block numbers to get
truncated on x86 and other 32-bit platforms.
Signed-off-by: N"Theodore Ts'o" <tytso@mit.edu>
Reviewed-by: NEric Sandeen <sandeen@redhat.com>

The ext4 module uses rcu_call() thus it should use rcu_barrier()on
module unload.

The kmem cache ext4_pspace_cachep is sometimes free'ed using
call_rcu() callbacks.  Thus, we must wait for completion of call_rcu()
before doing kmem_cache_destroy().
Signed-off-by: NJesper Dangaard Brouer <hawk@comx.dk>
Signed-off-by: N"Theodore Ts'o" <tytso@mit.edu>

Ted noticed a stack-deep callchain through
writepages->ext4_mb_regular_allocator->ext4_mb_init_cache->submit_bh ...

With all the static functions in mballoc.c, gcc helpfully
inlines for us, and we get something like this:

ext4_mb_regular_allocator	(232 bytes stack)
	ext4_mb_init_cache	(232 bytes stack)
		submit_bh	(starts 464 deeper)

the 2 ext4 functions here get several others inlined; by telling
gcc not to inline them, we can save stack space for when we
head off into submit_bh land and associated block layer callchains.
The following noinlined functions are only called once, so this
won't impact any other callchains:

ext4_mb_regular_allocator 			(104) (was 232)
	ext4_mb_find_by_goal			 (56) (noinlined)
	ext4_mb_init_group			 (24) (noinlined)
		ext4_mb_init_cache		(136) (was 232)
			ext4_mb_generate_buddy	 (88) (noinlined)
			ext4_mb_generate_from_pa (40) (noinlined)
			submit_bh
	ext4_mb_simple_scan_group		 (24) (noinlined)
	ext4_mb_scan_aligned			 (56) (noinlined)
	ext4_mb_complex_scan_group		 (40) (noinlined)
	ext4_mb_try_best_found			 (24) (noinlined)

now when we head off into submit_bh() we're only 264 bytes deeper
in stack than when we entered ext4_mb_regular_allocator()
(vs. 464 bytes before).  Every 200 bytes helps.  :)
Signed-off-by: NEric Sandeen <sandeen@redhat.com>
Signed-off-by: N"Theodore Ts'o" <tytso@mit.edu>

To catch filesystem bugs or corruption which could lead to the
filesystem getting severly damaged, this patch adds a facility for
tracking all of the filesystem metadata blocks by contiguous regions
in a red-black tree.  This allows quick searching of the tree to
locate extents which might overlap with filesystem metadata blocks.

This facility is also used by the multi-block allocator to assure that
it is not allocating blocks out of the system zone, as well as by the
routines used when reading indirect blocks and extents information
from disk to make sure their contents are valid.
Signed-off-by: N"Theodore Ts'o" <tytso@mit.edu>

On UP systems without DEBUG_SPINLOCK, ext4_is_group_locked always fails
which triggers a BUG_ON() call.
This patch fixes it by using assert_spin_locked instead.
Signed-off-by: NVincent Minet <vincent@vincent-minet.net>
Signed-off-by: N"Theodore Ts'o" <tytso@mit.edu>

We have sb_bgl_lock() and ext4_group_info.bb_state
bit spinlock to protech group information. The later is only
used within mballoc code. Consolidate them to use sb_bgl_lock().
This makes the mballoc.c code much simpler and also avoid
confusion with two locks protecting same info.
Signed-off-by: NAneesh Kumar K.V <aneesh.kumar@linux.vnet.ibm.com>
Signed-off-by: N"Theodore Ts'o" <tytso@mit.edu>

In memory-constrained systems with many partitions, the ~68K for each
partition for the mb_history buffer can be excessive.

This patch adds a new mount option, mb_history_length, as well as a
way of setting the default via a module parameter (or via a sysfs
parameter in /sys/module/ext4/parameter/default_mb_history_length).
If the mb_history_length is set to zero, the mb_history facility is
disabled entirely.
Signed-off-by: NCurt Wohlgemuth <curtw@google.com>
Signed-off-by: N"Theodore Ts'o" <tytso@mit.edu>

By avoiding the use of not-yet-used block groups (i.e., block groups
with the BLOCK_UNINIT flag), mballoc had a tendency to create large
files with large non-contiguous gaps.  In addition avoiding the use of
new block groups had a tendency to push regular file data into the
first block group in a flex_bg group, which slows down the speed of
e2fsck pass 2, since it has a tendency to seek much more.  For
example:

               Before Patch                       After Patch
              Time in seconds                   Time in seconds
            Real /  User/  Sys   MB/s      Real /  User/  Sys    MB/s
Pass 1      8.52 / 2.21 / 0.46  20.43      8.84 / 4.97 / 1.11   19.68
Pass 2     21.16 / 1.02 / 1.86  11.30      6.54 / 1.77 / 1.78   36.39
Pass 3      0.01 / 0.00 / 0.00 139.00      0.01 / 0.01 / 0.00  128.90
Pass 4      0.16 / 0.15 / 0.00   0.00      0.17 / 0.17 / 0.00    0.00
Pass 5      2.52 / 1.99 / 0.09   0.79      2.31 / 1.78 / 0.06    0.86
Total      32.40 / 5.11 / 2.49  12.81     17.99 / 8.75 / 2.98   23.01

This was on a sample 80 gig root filesystem which was approximately
50% full.  Note the improved e2fsck pass 2 performance, by over a
factor of 3, due to a decreased number of seeks.  (The total amount of
I/O in pass 2 was unchanged; the layout of the directory blocks was
simply much better from e2fsck's's perspective.)

Other changes as a result of this patch on this sample filesystem:

                             Before Patch    After Patch
# of non-contig files           762             779
# of non-contig directories     571             570
# of BLOCK_UNINIT bg's          307             293
# of INODE_UNINIT bg's          503             503

Out of 640 block groups, of which 333 were in use, this patch caused
an extra 14 block groups to be utilized.  The number of non-contiguous
files did go up slightly, but when measured against the 99.9% of the
files (603,154) which were contiguously allocated, this is pretty
insignificant.
Signed-off-by: N"Theodore Ts'o" <tytso@mit.edu>
Signed-off-by: NAndreas Dilger <adilger@sun.com>

Signed-off-by: N"Theodore Ts'o" <tytso@mit.edu>

Ext4's on-line resizing adds a new block group and then, only at the
last step adjusts s_groups_count.  However, it's possible on SMP
systems that another CPU could see the updated the s_group_count and
not see the newly initialized data structures for the just-added block
group.  For this reason, it's important to insert a SMP read barrier
after reading s_groups_count and before reading any (for example) the
new block group descriptors allowed by the increased value of
s_groups_count.

Unfortunately, we rather blatently violate this locking protocol
documented in fs/ext4/resize.c.  Fortunately, (1) on-line resizes
happen relatively rarely, and (2) it seems rare that the filesystem
code will immediately try to use just-added block group before any
memory ordering issues resolve themselves.  So apparently problems
here are relatively hard to hit, since ext3 has been vulnerable to the
same issue for years with no one apparently complaining.
Signed-off-by: N"Theodore Ts'o" <tytso@mit.edu>

Smatch (http://repo.or.cz/w/smatch.git/) complains about the locking in
ext4_mb_add_n_trim() from fs/ext4/mballoc.c

  4438          list_for_each_entry_rcu(tmp_pa, &lg->lg_prealloc_list[order],
  4439                                                  pa_inode_list) {
  4440                  spin_lock(&tmp_pa->pa_lock);
  4441                  if (tmp_pa->pa_deleted) {
  4442                          spin_unlock(&pa->pa_lock);
  4443                          continue;
  4444                  }

Brown paper bag time...
Reported-by: NDan Carpenter <error27@gmail.com>
Reviewed-by: NEric Sandeen <sandeen@redhat.com>
Reviewed-by: NAneesh Kumar K.V <aneesh.kumar@gmail.com>
Signed-off-by: N"Theodore Ts'o" <tytso@mit.edu>
Cc: stable@kernel.org

This was found by smatch (http://repo.or.cz/w/smatch.git/)
Signed-off-by: NDan Carpenter <error27@gmail.com>
Signed-off-by: N"Theodore Ts'o" <tytso@mit.edu>
Cc: stable@kernel.org

Impact: code cleanup

This patch rename pa_linear to pa_type and add MB_INODE_PA
and MB_GROUP_PA to indicate inode and group prealloc space.
Signed-off-by: NAneesh Kumar K.V <aneesh.kumar@linux.vnet.ibm.com>
Reviewed-by: NEric Sandeen <sandeen@redhat.com>
Signed-off-by: N"Theodore Ts'o" <tytso@mit.edu>

Use lowercase names of quota functions instead of old uppercase ones.
Signed-off-by: NJan Kara <jack@suse.cz>
Acked-by: NMingming Cao <cmm@us.ibm.com>
CC: linux-ext4@vger.kernel.org

Uses quota reservation/claim/release to handle quota properly for delayed
allocation in the three steps: 1) quotas are reserved when data being copied
to cache when block allocation is defered 2) when new blocks are allocated.
reserved quotas are converted to the real allocated quota, 2) over-booked
quotas for metadata blocks are released back.
Signed-off-by: NMingming Cao <cmm@us.ibm.com>
Acked-by: N"Theodore Ts'o" <tytso@mit.edu>
Signed-off-by: NJan Kara <jack@suse.cz>

This is for Red Hat bug 490026: EXT4 panic, list corruption in
ext4_mb_new_inode_pa

ext4_lock_group(sb, group) is supposed to protect this list for
each group, and a common code flow to remove an album is like
this:

    ext4_get_group_no_and_offset(sb, pa->pa_pstart, &grp, NULL);
    ext4_lock_group(sb, grp);
    list_del(&pa->pa_group_list);
    ext4_unlock_group(sb, grp);

so it's critical that we get the right group number back for
this prealloc context, to lock the right group (the one 
associated with this pa) and prevent concurrent list manipulation.

however, ext4_mb_put_pa() passes in (pa->pa_pstart - 1) with a 
comment, "-1 is to protect from crossing allocation group".

This makes sense for the group_pa, where pa_pstart is advanced
by the length which has been used (in ext4_mb_release_context()),
and when the entire length has been used, pa_pstart has been
advanced to the first block of the next group.

However, for inode_pa, pa_pstart is never advanced; it's just
set once to the first block in the group and not moved after
that.  So in this case, if we subtract one in ext4_mb_put_pa(),
we are actually locking the *previous* group, and opening the
race with the other threads which do not subtract off the extra
block.
Signed-off-by: NEric Sandeen <sandeen@redhat.com>
Signed-off-by: N"Theodore Ts'o" <tytso@mit.edu>

Thiemo Nagel reported that:

# dd if=/dev/zero of=image.ext4 bs=1M count=2
# mkfs.ext4 -v -F -b 1024 -m 0 -g 512 -G 4 -I 128 -N 1 \
  -O large_file,dir_index,flex_bg,extent,sparse_super image.ext4
# mount -o loop image.ext4 mnt/
# dd if=/dev/zero of=mnt/file

oopsed, with a BUG_ON in ext4_mb_normalize_request because
size == EXT4_BLOCKS_PER_GROUP

It appears to me (esp. after talking to Andreas) that the BUG_ON
is bogus; a request of exactly EXT4_BLOCKS_PER_GROUP should
be allowed, though larger sizes do indicate a problem.

Fix that an another (apparently rare) codepath with a similar check.
Reported-by: NThiemo Nagel <thiemo.nagel@ph.tum.de>
Signed-off-by: NEric Sandeen <sandeen@redhat.com>
Signed-off-by: N"Theodore Ts'o" <tytso@mit.edu>