linux-toradex.git/fs/f2fs/node.c, branch v3.14.34

mm: non-atomically mark page accessed during page cache allocation where possible

2015-01-30T01:40:52+00:00

commit 2457aec63745e235bcafb7ef312b182d8682f0fc upstream.

aops->write_begin may allocate a new page and make it visible only to have
mark_page_accessed called almost immediately after.  Once the page is
visible the atomic operations are necessary which is noticable overhead
when writing to an in-memory filesystem like tmpfs but should also be
noticable with fast storage.  The objective of the patch is to initialse
the accessed information with non-atomic operations before the page is
visible.

The bulk of filesystems directly or indirectly use
grab_cache_page_write_begin or find_or_create_page for the initial
allocation of a page cache page.  This patch adds an init_page_accessed()
helper which behaves like the first call to mark_page_accessed() but may
called before the page is visible and can be done non-atomically.

The primary APIs of concern in this care are the following and are used
by most filesystems.

	find_get_page
	find_lock_page
	find_or_create_page
	grab_cache_page_nowait
	grab_cache_page_write_begin

All of them are very similar in detail to the patch creates a core helper
pagecache_get_page() which takes a flags parameter that affects its
behavior such as whether the page should be marked accessed or not.  Then
old API is preserved but is basically a thin wrapper around this core
function.

Each of the filesystems are then updated to avoid calling
mark_page_accessed when it is known that the VM interfaces have already
done the job.  There is a slight snag in that the timing of the
mark_page_accessed() has now changed so in rare cases it's possible a page
gets to the end of the LRU as PageReferenced where as previously it might
have been repromoted.  This is expected to be rare but it's worth the
filesystem people thinking about it in case they see a problem with the
timing change.  It is also the case that some filesystems may be marking
pages accessed that previously did not but it makes sense that filesystems
have consistent behaviour in this regard.

The test case used to evaulate this is a simple dd of a large file done
multiple times with the file deleted on each iterations.  The size of the
file is 1/10th physical memory to avoid dirty page balancing.  In the
async case it will be possible that the workload completes without even
hitting the disk and will have variable results but highlight the impact
of mark_page_accessed for async IO.  The sync results are expected to be
more stable.  The exception is tmpfs where the normal case is for the "IO"
to not hit the disk.

The test machine was single socket and UMA to avoid any scheduling or NUMA
artifacts.  Throughput and wall times are presented for sync IO, only wall
times are shown for async as the granularity reported by dd and the
variability is unsuitable for comparison.  As async results were variable
do to writback timings, I'm only reporting the maximum figures.  The sync
results were stable enough to make the mean and stddev uninteresting.

The performance results are reported based on a run with no profiling.
Profile data is based on a separate run with oprofile running.

async dd
                                    3.15.0-rc3            3.15.0-rc3
                                       vanilla           accessed-v2
ext3    Max      elapsed     13.9900 (  0.00%)     11.5900 ( 17.16%)
tmpfs	Max      elapsed      0.5100 (  0.00%)      0.4900 (  3.92%)
btrfs   Max      elapsed     12.8100 (  0.00%)     12.7800 (  0.23%)
ext4	Max      elapsed     18.6000 (  0.00%)     13.3400 ( 28.28%)
xfs	Max      elapsed     12.5600 (  0.00%)      2.0900 ( 83.36%)

The XFS figure is a bit strange as it managed to avoid a worst case by
sheer luck but the average figures looked reasonable.

        samples percentage
ext3       86107    0.9783  vmlinux-3.15.0-rc4-vanilla        mark_page_accessed
ext3       23833    0.2710  vmlinux-3.15.0-rc4-accessed-v3r25 mark_page_accessed
ext3        5036    0.0573  vmlinux-3.15.0-rc4-accessed-v3r25 init_page_accessed
ext4       64566    0.8961  vmlinux-3.15.0-rc4-vanilla        mark_page_accessed
ext4        5322    0.0713  vmlinux-3.15.0-rc4-accessed-v3r25 mark_page_accessed
ext4        2869    0.0384  vmlinux-3.15.0-rc4-accessed-v3r25 init_page_accessed
xfs        62126    1.7675  vmlinux-3.15.0-rc4-vanilla        mark_page_accessed
xfs         1904    0.0554  vmlinux-3.15.0-rc4-accessed-v3r25 init_page_accessed
xfs          103    0.0030  vmlinux-3.15.0-rc4-accessed-v3r25 mark_page_accessed
btrfs      10655    0.1338  vmlinux-3.15.0-rc4-vanilla        mark_page_accessed
btrfs       2020    0.0273  vmlinux-3.15.0-rc4-accessed-v3r25 init_page_accessed
btrfs        587    0.0079  vmlinux-3.15.0-rc4-accessed-v3r25 mark_page_accessed
tmpfs      59562    3.2628  vmlinux-3.15.0-rc4-vanilla        mark_page_accessed
tmpfs       1210    0.0696  vmlinux-3.15.0-rc4-accessed-v3r25 init_page_accessed
tmpfs         94    0.0054  vmlinux-3.15.0-rc4-accessed-v3r25 mark_page_accessed

[akpm@linux-foundation.org: don't run init_page_accessed() against an uninitialised pointer]
Signed-off-by: Mel Gorman 
Cc: Johannes Weiner 
Cc: Vlastimil Babka 
Cc: Jan Kara 
Cc: Michal Hocko 
Cc: Hugh Dickins 
Cc: Dave Hansen 
Cc: Theodore Ts'o 
Cc: "Paul E. McKenney" 
Cc: Oleg Nesterov 
Cc: Rik van Riel 
Cc: Peter Zijlstra 
Tested-by: Prabhakar Lad 
Signed-off-by: Andrew Morton 
Signed-off-by: Linus Torvalds 
Signed-off-by: Mel Gorman 
Signed-off-by: Greg Kroah-Hartman

f2fs: drop obsolete node page when it is truncated

2014-01-22T23:04:21+00:00

If a node page is trucated, we'd better drop the page in the node_inode's page
cache for better memory footprint.

Signed-off-by: Jaegeuk Kim

f2fs: introduce NODE_MAPPING for code consistency

2014-01-22T09:41:08+00:00

This patch adds NODE_MAPPING which is similar as META_MAPPING introduced by
Gu Zheng.

Cc: Gu Zheng 
Signed-off-by: Jaegeuk Kim

f2fs: add help function META_MAPPING

2014-01-22T09:41:07+00:00

Introduce help function META_MAPPING() to get the cache meta blocks'
address space.

Signed-off-by: Gu Zheng 
Signed-off-by: Jaegeuk Kim

f2fs: clean checkpatch warnings

2014-01-20T01:27:12+00:00

Fixed a variety of trivial checkpatch warnings.  The only delta should
be some minor formatting on log strings that were split / too long.

Signed-off-by: Chris Fries 
Signed-off-by: Jaegeuk Kim

f2fs: improve write performance under frequent fsync calls

2014-01-08T02:16:20+00:00

When considering a bunch of data writes with very frequent fsync calls, we
are able to think the following performance regression.

N: Node IO, D: Data IO, IO scheduler: cfq

Issue    pending IOs
	 D1 D2 D3 D4
 D1         D2 D3 D4 N1
 D2            D3 D4 N1 N2
 N1            D3 D4 N2 D1
 --> N1 can be selected by cfq becase of the same priority of N and D.
     Then D3 and D4 would be delayed, resuling in performance degradation.

So, when processing the fsync call, it'd better give higher priority to data IOs
than node IOs by assigning WRITE and WRITE_SYNC respectively.
This patch improves the random wirte performance with frequent fsync calls by up
to 10%.

Signed-off-by: Jaegeuk Kim

f2fs: fix truncate_partial_nodes bug

2014-01-06T07:42:21+00:00

The truncate_partial_nodes puts pages incorrectly in the following two cases.
Note that the value for argc 'depth' can only be 2 or 3.
Please see truncate_inode_blocks() and truncate_partial_nodes().

1) An err is occurred in the first 'for' loop
  When err is occurred with depth = 2, pages[0] is invalid, so this page doesn't
  need to be put. There is no problem, however, when depth is 3, it doesn't put
  the pages correctly where pages[0] is valid and pages[1] is invalid.
  In this case, depth is set to 2 (ref to statemnt depth = i + 1), and then
  'goto fail'.
  In label 'fail', for (i = depth - 3; i >= 0; i--) cannot meet the condition
  because i = -1, so pages[0] cann't be put.

2) An err happened in the second 'for' loop
  Now we've got pages[0] with depth = 2, or we've got pages[0] and pages[1]
  with depth = 3. When an err is detected, we need 'goto fail' to put such
  the pages.
  When depth is 2, in label 'fail', for (i = depth - 3; i >= 0; i--) cann't
  meet the condition because i = -1, so pages[0] cann't be put.
  When depth is 3, in label 'fail', for (i = depth - 3; i >= 0; i--) can
  only put pages[0], pages[1] also cann't be put.

Note that 'depth' has been changed before first 'goto fail' (ref to statemnt
depth = i + 1), so passing this modified 'depth' to the tracepoint,
trace_f2fs_truncate_partial_nodes, is also incorrect.

Signed-off-by: Shifei Ge 
[Jaegeuk Kim: modify the description and fix one bug]
Signed-off-by: Jaegeuk Kim

f2fs: introduce F2FS_INODE macro to get f2fs_inode

2013-12-26T11:32:48+00:00

This patch introduces F2FS_INODE that returns struct f2fs_inode * from the inode
page.
By using this macro, we can remove unnecessary casting codes like below.

   struct f2fs_inode *ri = &F2FS_NODE(inode_page)->i;
-> struct f2fs_inode *ri = F2FS_INODE(inode_page);

Reviewed-by: Chao Yu 
Signed-off-by: Jaegeuk Kim

f2fs: update several comments

2013-12-23T01:26:03+00:00

Update several comments:
1. use f2fs_{un}lock_op install of mutex_{un}lock_op.
2. update comment of get_data_block().
3. update description of node offset.

Signed-off-by: Chao Yu 
Signed-off-by: Jaegeuk Kim

f2fs: remove the rw_flag domain from f2fs_io_info

2013-12-23T01:18:07+00:00

When using the f2fs_io_info in the low level, we still need to merge the
rw and rw_flag, so use the rw to hold all the io flags directly,
and remove the rw_flag field.

ps.It is based on the previous patch:
f2fs: move all the bio initialization into __bio_alloc

Signed-off-by: Gu Zheng 
Signed-off-by: Jaegeuk Kim