Currently cluster allocation is always trying to find a victim chain (a
chian has most space), and this may lead to poor performance because of
discontiguous allocation in some scenarios.
Our test case is block size 4k, cluster size 1M and mount option with
localalloc=2048 (2G), since a gd is 32256M (about 31.5G) and a localalloc
window is only 2G, creating 50G file will result in 2G from gd0, 2G from
gd1, ...
One way to improve performance is enlarge localalloc window size (max
31104M), but this will make end user feel that about 30G is suddenly
"missing", and localalloc currently do not support steal, which means one
node cannot use another node's localalloc even it is not used in fact. So
using the last gd to record the allocation and continues with the gd if it
has enough space for a localalloc window can make the allocation as more
contiguous as possible.
Our test result is below (evaluated in IOPS), which is using iometer
running in VM, dynamic vhd virtual disk stored in ocfs2.
IO model Original After Improved(%)
16K60%Write100%Random 703 876 24.59%
8K90%Write100%Random 735 827 12.59%
4K100%Write100%Random 859 915 6.52%
4K100%Read100%Random 2092 2600 24.30%
Signed-off-by: Joseph Qi <joseph.qi@huawei.com>
Tested-by: Norton Zhu <norton.zhu@huawei.com>
Cc: Mark Fasheh <mfasheh@suse.de>
Cc: Joel Becker <jlbec@evilplan.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
NULL check before kfree is redundant and so clean them up.
Signed-off-by: Joseph Qi <joseph.qi@huawei.com>
Reviewed-by: Mark Fasheh <mfasheh@suse.de>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
These uses sometimes do and sometimes don't have '\n' terminations. Make
the uses consistently use '\n' terminations and remove the newline from
the functions.
Miscellanea:
o Coalesce formats
o Realign arguments
Signed-off-by: Joe Perches <joe@perches.com>
Reviewed-by: Mark Fasheh <mfasheh@suse.de>
Cc: Joel Becker <jlbec@evilplan.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Caveat: This may return -EROFS for a read case, which seems wrong. This
is happening even without this patch series though. Should we convert
EROFS to EIO?
Signed-off-by: Goldwyn Rodrigues <rgoldwyn@suse.com>
Cc: Mark Fasheh <mfasheh@suse.com>
Cc: Joel Becker <jlbec@evilplan.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
ocfs2_block_group_clear_bits will clear bits in block group bitmap.
Once it succeeds but fails in the following step, it will cause block
group bitmap mismatch the corresponding count recorded in dinode.
So rollback the cleared bits if error occurs.
Signed-off-by: Joseph Qi <joseph.qi@huawei.com>
Cc: Mark Fasheh <mfasheh@suse.com>
Cc: Joel Becker <jlbec@evilplan.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
In ocfs2_info_handle_freeinode() and ocfs2_test_inode_bit() func, after
calls ocfs2_get_system_file_inode() to get inode ref, if calls
ocfs2_info_scan_inode_alloc() or ocfs2_inode_lock() failed, we should
iput inode alloc to avoid leaking the inode.
Signed-off-by: jiangyiwen <jiangyiwen@huawei.com>
Reviewed-by: Joseph Qi <joseph.qi@huawei.com>
Cc: Mark Fasheh <mfasheh@suse.de>
Cc: Joel Becker <jlbec@evilplan.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
After updating alloc_dinode counts in ocfs2_alloc_dinode_update_counts(),
if ocfs2_alloc_dinode_update_bitmap() failed, there is a rare case that
some space may be lost.
So, roll back alloc_dinode counts when ocfs2_block_group_set_bits()
failed.
[akpm@linux-foundation.org: coding-style fixes]
Signed-off-by: Younger Liu <younger.liucn@gmail.com>
Reviewed-by: Mark Fasheh <mfasheh@suse.de>
Cc: Joel Becker <jlbec@evilplan.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Ensure that ocfs2_update_inode_fsync_trans() is called any time we touch
an inode in a given transaction. This is a follow-on to the previous
patch to reduce lock contention and deadlocking during an fsync
operation.
Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
Cc: Mark Fasheh <mfasheh@suse.de>
Cc: Joel Becker <jlbec@evilplan.org>
Cc: Wengang <wen.gang.wang@oracle.com>
Cc: Greg Marsden <greg.marsden@oracle.com>
Cc: Srinivas Eeda <srinivas.eeda@oracle.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
ocfs2_alloc_dinode_update_counts() and ocfs2_block_group_set_bits() are
already provided in suballoc.c. So, the same functions in
move_extents.c are not needed any more.
Declare the functions in suballoc.h and remove redundant functions in
move_extents.c.
Signed-off-by: Younger Liu <liuyiyang@hisense.com>
Cc: Younger Liu <younger.liucn@gmail.com>
Cc: Mark Fasheh <mfasheh@suse.com>
Cc: Joel Becker <jlbec@evilplan.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
The only reason for sb_getblk() failing is if it can't allocate the
buffer_head. So return ENOMEM instead when it fails.
[joseph.qi@huawei.com: ocfs2_symlink_get_block() and ocfs2_read_blocks_sync() and ocfs2_read_blocks() need the same change]
Signed-off-by: Rui Xiang <rui.xiang@huawei.com>
Reviewed-by: Jie Liu <jeff.liu@oracle.com>
Reviewed-by: Mark Fasheh <mfasheh@suse.de>
Cc: Joel Becker <jlbec@evilplan.org>
Cc: Joseph Qi <joseph.qi@huawei.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Jie Liu <jeff.liu@oracle.com>
Cc: Joel Becker <jlbec@evilplan.org>
Cc: Mark Fasheh <mfasheh@suse.com>
Cc: Sunil Mushran <sunil.mushran@gmail.com>
Cc: Younger Liu <younger.liu@huawei.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
In ocfs2_relink_block_group(), we roll back all those changes if notify
intent to modify buffers for metadata update failed even if the relevant
buffer has not yet been modified/got dirty at that point, that are not
quite right because of:
- None buffer has been modified/dirty if failed to call
ocfs2_journal_access_gd() against the previous block group buffer
- Only the previous block group buffer has got dirty if failed to call
ocfs2_journal_access_gd() against the block group buffer
- There is no need to roll back the change for file entry buffer at all
Those problems will not cause anything wrong but unnecessary. This
patch fix them and kill the useless bg_ptr variable as well.
Signed-off-by: Jie Liu <jeff.liu@oracle.com>
Cc: Younger Liu <younger.liu@huawei.com>
Cc: Sunil Mushran <sunil.mushran@gmail.com>
Cc: Mark Fasheh <mfasheh@suse.com>
Cc: Joel Becker <jlbec@evilplan.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
ocfs2_block_group_alloc_discontig() disables chain relink by setting
ac->ac_allow_chain_relink = 0 because it grabs clusters from multiple
cluster groups.
It doesn't keep the credits for all chain relink,but
ocfs2_claim_suballoc_bits overrides this in this call trace:
ocfs2_block_group_claim_bits()->ocfs2_claim_clusters()->
__ocfs2_claim_clusters()->ocfs2_claim_suballoc_bits()
ocfs2_claim_suballoc_bits set ac->ac_allow_chain_relink = 1; then call
ocfs2_search_chain() one time and disable it again, and then we run out
of credits.
Fix is to allow relink by default and disable it in
ocfs2_block_group_alloc_discontig.
Without this patch, End-users will run into a crash due to run out of
credits, backtrace like this:
RIP: 0010:[<ffffffffa0808b14>] [<ffffffffa0808b14>]
jbd2_journal_dirty_metadata+0x164/0x170 [jbd2]
RSP: 0018:ffff8801b919b5b8 EFLAGS: 00010246
RAX: 0000000000000000 RBX: ffff88022139ddc0 RCX: ffff880159f652d0
RDX: ffff880178aa3000 RSI: ffff880159f652d0 RDI: ffff880087f09bf8
RBP: ffff8801b919b5e8 R08: 0000000000000000 R09: 0000000000000000
R10: 0000000000001e00 R11: 00000000000150b0 R12: ffff880159f652d0
R13: ffff8801a0cae908 R14: ffff880087f09bf8 R15: ffff88018d177800
FS: 00007fc9b0b6b6e0(0000) GS:ffff88022fd40000(0000) knlGS:0000000000000000
CS: 0010 DS: 0000 ES: 0000 CR0: 000000008005003b
CR2: 000000000040819c CR3: 0000000184017000 CR4: 00000000000006e0
DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000
DR3: 0000000000000000 DR6: 00000000ffff0ff0 DR7: 0000000000000400
Process dd (pid: 9945, threadinfo ffff8801b919a000, task ffff880149a264c0)
Call Trace:
ocfs2_journal_dirty+0x2f/0x70 [ocfs2]
ocfs2_relink_block_group+0x111/0x480 [ocfs2]
ocfs2_search_chain+0x455/0x9a0 [ocfs2]
...
Signed-off-by: Xiaowei.Hu <xiaowei.hu@oracle.com>
Reviewed-by: Srinivas Eeda <srinivas.eeda@oracle.com>
Cc: Mark Fasheh <mfasheh@suse.com>
Cc: Joel Becker <jlbec@evilplan.org>
Cc: <stable@vger.kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Since all 4 files, localalloc.c, suballoc.c, alloc.c and
resize.c, which use DISK_ALLOC are changed to trace events,
Remove masklog DISK_ALLOC totally.
Signed-off-by: Tao Ma <boyu.mt@taobao.com>
mlog_exit is used to record the exit status of a function.
But because it is added in so many functions, if we enable it,
the system logs get filled up quickly and cause too much I/O.
So actually no one can open it for a production system or even
for a test.
This patch just try to remove it or change it. So:
1. if all the error paths already use mlog_errno, it is just removed.
Otherwise, it will be replaced by mlog_errno.
2. if it is used to print some return value, it is replaced with
mlog(0,...).
mlog_exit_ptr is changed to mlog(0.
All those mlog(0,...) will be replaced with trace events later.
Signed-off-by: Tao Ma <boyu.mt@taobao.com>
ENTRY is used to record the entry of a function.
But because it is added in so many functions, if we enable it,
the system logs get filled up quickly and cause too much I/O.
So actually no one can open it for a production system or even
for a test.
So for mlog_entry_void, we just remove it.
for mlog_entry(...), we replace it with mlog(0,...), and they
will be replace by trace event later.
Signed-off-by: Tao Ma <boyu.mt@taobao.com>
This patch adds a safe check to ensure bg_free_bits_count doesn't exceed
bg_bits in a group descriptor. This is to avoid on disk corruption that was
seen recently.
debugfs: group <52803072>
Group Chain: 179 Parent Inode: 11 Generation: 2959379682
CRC32: 00000000 ECC: 0000
## Block# Total Used Free Contig Size
0 52803072 32256 4294965350 34202 18207 4032
......
Signed-off-by: Srinivas Eeda <srinivas.eeda@oracle.com>
Signed-off-by: Joel Becker <joel.becker@oracle.com>
e_leaf_clusters is a le16, so use cpu_to_le16 instead
of cpu_to_le32.
What's more, we change 'clusters' to unsigned int to
signify that the size of 'clusters' isn't important here.
Signed-off-by: Tao Ma <tao.ma@oracle.com>
Signed-off-by: Joel Becker <joel.becker@oracle.com>
This allows code which needs to know the eventual block number of an inode
but can't allocate it yet due to transaction or lock ordering. For example,
ocfs2_create_inode_in_orphan() currently gives a junk blkno for preparation
of the orphan dir because it can't yet know where the actual inode is placed
- that code is actually in ocfs2_mknod_locked. This is a problem when the
orphan dirs are indexed as the junk inode number will create an index entry
which goes unused (and fails the later removal from the orphan dir). Now
with these interfaces, ocfs2_create_inode_in_orphan() can run the block
group search (and get back the inode block number) *before* any actual
allocation occurs.
Signed-off-by: Mark Fasheh <mfasheh@suse.com>
Signed-off-by: Tao Ma <tao.ma@oracle.com>
ocfs2_search_chain() makes the same updates as
ocfs2_alloc_dinode_update_counts to the alloc inode. Instead of open coding
the bitmap update, use our helper function.
Signed-off-by: Mark Fasheh <mfasheh@suse.com>
Signed-off-by: Tao Ma <tao.ma@oracle.com>
We were setting ac->ac_last_group in ocfs2_claim_suballoc_bits from
res->sr_bg_blkno. Unfortunately, res->sr_bg_blkno is going to be zero under
normal (non-fragmented) circumstances. The discontig block group patches
effectively turned off that feature. Fix this by correctly calculating what
the next group hint should be.
Acked-by: Tao Ma <tao.ma@oracle.com>
Signed-off-by: Mark Fasheh <mfasheh@suse.com>
Tested-by: Goldwyn Rodrigues <rgoldwyn@suse.de>
Signed-off-by: Tao Ma <tao.ma@oracle.com>
We have added discontig block group now, and now an inode
can be allocated in an discontig block group. So get
it in ocfs2_get_suballoc_slot_bit.
The old ocfs2_test_suballoc_bit gets group block no
from the allocation inode which is wrong. Fix it by
passing the right group.
Acked-by: Mark Fasheh <mfasheh@suse.com>
Signed-off-by: Tao Ma <tao.ma@oracle.com>
In ocfs2_block_group_alloc, we set c_blkno by bg->bg_blkno.
But actually bg->bg_blkno is already changed to little endian
in ocfs2_block_group_fill. So remove the extra cpu_to_le64.
Reported-by: Marcos Matsunaga <Marcos.Matsunaga@oracle.com>
Signed-off-by: Tao Ma <tao.ma@oracle.com>
Signed-off-by: Joel Becker <joel.becker@oracle.com>
ocfs2_block_group_claim_bits() is never called with min_bits=0, but we
shouldn't leave status undefined if it ever is.
Signed-off-by: Joel Becker <joel.becker@oracle.com>
ac_last_group is used to record the last block group we
used during allocation. But the initialization process
only calls ocfs2_which_suballoc_group and fails to
use suballoc_loc properly. So let us do it.
Another function ocfs2_test_suballoc_bit also needs fix.
I have searched all the callers of ocfs2_which_suballoc_group,
and all the callers notices suballoc_loc now.
Signed-off-by: Tao Ma <tao.ma@oracle.com>
In case the block we are going to free is allocated from
a discontiguous block group, we have to use suballoc_loc
to be the right group.
Signed-off-by: Tao Ma <tao.ma@oracle.com>
ocfs2_group_bitmap_size has to handle the case when the
volume don't have discontiguous block group support. So
pass the feature_incompat in and check it.
Signed-off-by: Tao Ma <tao.ma@oracle.com>
The fixes include:
1. some endian problems.
2. we should use bit/bpc in ocfs2_block_group_grow_discontig to
allocate clusters.
3. set num_clusters properly in __ocfs2_claim_clusters.
4. change name from ocfs2_supports_discontig_bh to
ocfs2_supports_discontig_bg.
Signed-off-by: Tao Ma <tao.ma@oracle.com>
Rather than extending the transaction every time we add an extent to a
discontiguous block group, we grab enough credits to fill the extent
list up front. This means we can free the bits in the same transaction
if we end up not getting enough space.
Signed-off-by: Joel Becker <joel.becker@oracle.com>
Get the suballoc_loc from ocfs2_claim_new_inode() or
ocfs2_claim_metadata(). Store it on the appropriate field of the block
we just allocated.
Signed-off-by: Joel Becker <joel.becker@oracle.com>
Rather than calculating the resulting block number, return it on the
ocfs2_suballoc_result structure. This way we can calculate block
numbers for discontiguous block groups.
Cluster groups keep doing it the old way.
Signed-off-by: Joel Becker <joel.becker@oracle.com>
They all take an ocfs2_alloc_context, which has the allocation inode.
Signed-off-by: Joel Becker <joel.becker@oracle.com>
Signed-off-by: Tao Ma <tao.ma@oracle.com>
A discontiguous block group can find a range of free bits that straddle
more than one region of its space. Callers can't handle that, so we
trim the returned bits until they fit within one region.
Only cluster allocations ask for min_bits>1. Discontiguous block groups
are only for block allocations. So min_bits doesn't matter here.
Signed-off-by: Joel Becker <joel.becker@oracle.com>
We're going to be adding more info to a suballocator allocation. Rather
than growing every function in the chain, let's pass a result structure
around.
Signed-off-by: Joel Becker <joel.becker@oracle.com>
Signed-off-by: Tao Ma <tao.ma@oracle.com>
If we cannot get a contiguous region for a block group, allocate a
discontiguous one when the filesystem supports it.
Signed-off-by: Joel Becker <joel.becker@oracle.com>
Signed-off-by: Tao Ma <tao.ma@oracle.com>
Defines the OCFS2_FEATURE_INCOMPAT_DISCONTIG_BG feature bit and modifies
struct ocfs2_group_desc for the feature.
Signed-off-by: Joel Becker <joel.becker@oracle.com>
Signed-off-by: Tao Ma <tao.ma@oracle.com>
Inodes are always allocated from the global bitmap now so we don't need this
any more. Also, the existing implementation bounces reservations around
needlessly.
Signed-off-by: Mark Fasheh <mfasheh@suse.com>
Otherwise, the need for a very large contiguous allocation tends to
wreak havoc on many inode allocation reservations on the local alloc, thus
ruining any chances for contiguousness.
Signed-off-by: Mark Fasheh <mfasheh@suse.com>
Use the reservations system for unindexed dir tree allocations. We don't
bother with the indexed tree as reads from it are mostly random anyway.
Directory reservations are marked seperately, to allow the reservations code
a chance to optimize their window sizes. This patch allocates only 8 bits
for directory windows as they generally are not expected to grow as quickly
as file data. Future improvements to dir window sizing can trivially be
made.
Signed-off-by: Mark Fasheh <mfasheh@suse.com>
jbd[2]_journal_dirty_metadata() only returns 0. It's been returning 0
since before the kernel moved to git. There is no point in checking
this error.
ocfs2_journal_dirty() has been faithfully returning the status since the
beginning. All over ocfs2, we have blocks of code checking this can't
fail status. In the past few years, we've tried to avoid adding these
checks, because they are pointless. But anyone who looks at our code
assumes they are needed.
Finally, ocfs2_journal_dirty() is made a void function. All error
checking is removed from other files. We'll BUG_ON() the status of
jbd2_journal_dirty_metadata() just in case they change it someday. They
won't.
Signed-off-by: Joel Becker <joel.becker@oracle.com>
When the local alloc file changes windows, unused bits are freed back to the
global bitmap. By defnition, those bits can not be in use by any file. Also,
the local alloc will never have been able to allocate those bits if they
were part of a previous truncate. Therefore it makes sense that we should
clear unused local alloc bits in the undo buffer so that they can be used
immediatly.
[ Modified to call it ocfs2_release_clusters() -- Joel ]
Signed-off-by: Mark Fasheh <mfasheh@suse.com>
Signed-off-by: Joel Becker <joel.becker@oracle.com>
In ocfs2_validate_gd_parent, we check bg_chain against the
cl_next_free_rec of the dinode. Actually in resize, we have
the chance of bg_chain == cl_next_free_rec. So add some
additional condition check for it.
I also rename paramter "clean_error" to "resize", since the
old one is not clearly enough to indicate that we should only
meet with this case in resize.
btw, the correpsonding bug is
http://oss.oracle.com/bugzilla/show_bug.cgi?id=1230.
Signed-off-by: Tao Ma <tao.ma@oracle.com>
Signed-off-by: Joel Becker <joel.becker@oracle.com>
This patch add extent block (metadata) stealing mechanism for
extent allocation. This mechanism is same as the inode stealing.
if no room in slot specific extent_alloc, we will try to
allocate extent block from the next slot.
Signed-off-by: Tiger Yang <tiger.yang@oracle.com>
Acked-by: Tao Ma <tao.ma@oracle.com>
Signed-off-by: Joel Becker <joel.becker@oracle.com>
