Signed-off-by: NFabian Frederick <fabf@skynet.be>
Suggested-By: NVyacheslav Dubeyko <slava@dubeyko.com>
Signed-off-by: NAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

Some function declarations in hfsplus_fs.h were with argument names,
some without, and some were mixed.  This patch adds argument names
everywhere, sorts function in order they go in .c files, and moves
hfs_part_find() to a proper section.

Auto-formatting and sorting was done with:
cfunctions *.c | indent -linux | sed "s| \* | \*|"
Signed-off-by: NSergei Antonov <saproj@gmail.com>
Cc: Vyacheslav Dubeyko <slava@dubeyko.com>
Cc: Hin-Tak Leung <htl10@users.sourceforge.net>
Signed-off-by: NAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

Replace while blocksize;shift by ilog2
Signed-off-by: NFabian Frederick <fabf@skynet.be>
Cc: Vyacheslav Dubeyko <slava@dubeyko.com>
Cc: Joe Perches <joe@perches.com>
Signed-off-by: NAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

Zero newly allocated extents in the catalog tree if volume attributes
tell us to.  Not doing so we risk getting the "unused node is not
erased" error.  See kHFSUnusedNodeFix flag in Apple's source code for
reference.

There was a previous commit clearing the node when it is freed: commit
899bed05 ("hfsplus: fix issue with unzeroed unused b-tree nodes").
But it did not handle newly allocated extents (this patch fixes it).
And it zeroed nodes in all trees unconditionally which is an overkill.

This patch adds a condition and also switches to 'tree->node_size' as a
simpler method of getting the length to zero.
Signed-off-by: NSergei Antonov <saproj@gmail.com>
Cc: Anton Altaparmakov <aia21@cam.ac.uk>
Cc: Al Viro <viro@zeniv.linux.org.uk>
Cc: Christoph Hellwig <hch@infradead.org>
Cc: Vyacheslav Dubeyko <slava@dubeyko.com>
Cc: Hin-Tak Leung <htl10@users.sourceforge.net>
Cc: Kyle Laracey <kalaracey@gmail.com>
Signed-off-by: NAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

Also add * before function comments (it was not detected by kernel-doc)
Signed-off-by: NFabian Frederick <fabf@skynet.be>
Cc: Vyacheslav Dubeyko <slava@dubeyko.com>
Signed-off-by: NAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

Replace seq_printf where possible
Signed-off-by: NFabian Frederick <fabf@skynet.be>
Cc: Vyacheslav Dubeyko <slava@dubeyko.com>
Signed-off-by: NAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

Also fixes some pr_ formats
Signed-off-by: NFabian Frederick <fabf@skynet.be>
Cc: Vyacheslav Dubeyko <slava@dubeyko.com>
Signed-off-by: NAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

hfsplus_readdir() incorrectly returned DT_REG for symbolic links and
special files.  Return DT_REG, DT_LNK, DT_FIFO, DT_CHR, DT_BLK, DT_SOCK,
or DT_UNKNOWN according to mode field in catalog record.  Programs
relying on information from readdir will now work correctly with HFS+.
Signed-off-by: NSergei Antonov <saproj@gmail.com>
Cc: Anton Altaparmakov <aia21@cam.ac.uk>
Cc: Al Viro <viro@zeniv.linux.org.uk>
Cc: Christoph Hellwig <hch@infradead.org>
Cc: Vyacheslav Dubeyko <slava@dubeyko.com>
Cc: Hin-Tak Leung <htl10@users.sourceforge.net>
Signed-off-by: NAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

The directory/file catalog b-tree equivalent, hfsplus_build_key_uni(),
is used by hfsplus_find_cat() for internal referencing between catalog
records.  There is no corresponding usage for attributes - attribute
records do not refer to one another.
Signed-off-by: NHin-Tak Leung <htl10@users.sourceforge.net>
Cc: Sougata Santra <sougata@tuxera.com>
Cc: Anton Altaparmakov <anton@tuxera.com>
Cc: Vyacheslav Dubeyko <slava@dubeyko.com>
Cc: Al Viro <viro@zeniv.linux.org.uk>
Cc: Christoph Hellwig <hch@infradead.org>
Signed-off-by: NAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

HFSPLUS_ATTR_MAX_STRLEN (=127) is the limit of attribute names for the
number of unicode character (UTF-16BE) storable in the HFS+ file system.
Almost all the current usage of it is wrong, in relation to NLS to
on-disk conversion.

Except for one use calling hfsplus_asc2uni (which should stay the same)
and its uses in calling hfsplus_uni2asc (which was corrected in the
earlier patch in this series concerning usage of hfsplus_uni2asc), all
the other uses are of the forms:

- char buffer[size]

- bound check: "if (namespace_adjusted_input_length > size) return failure;"

Conversion between on-disk unicode representation and NLS char strings
(in whichever direction) always needs to accommodate the worst-case NLS
conversion, so all char buffers of that size need to have a
NLS_MAX_CHARSET_SIZE x .

The bound checks are all wrong, since they compare nls_length derived
from strlen() to a unicode length limit.

It turns out that all the bound-checks do is to protect
hfsplus_asc2uni(), which can fail if the input is too large.

There is only one usage of it as far as attributes are concerned, in
hfsplus_attr_build_key().  It is in turn used by hfsplus_find_attr(),
hfsplus_create_attr(), hfsplus_delete_attr().  Thus making sure that
errors from hfsplus_asc2uni() is caught in hfsplus_attr_build_key() and
propagated is sufficient to replace all the bound checks.

Unpropagated errors from hfsplus_asc2uni() in the file catalog code was
addressed recently in an independent patch "hfsplus: fix longname
handling" by Sougata Santra.

Before this patch, trying to set a 55 CJK character (in a UTF-8 locale,
> 127/3=42) attribute plus user prefix fails with:

    $ setfattr -n user.`cat testing-string` -v `cat testing-string` \
        testing-string
    setfattr: testing-string: Operation not supported

and retrieving a stored long attributes is particular ugly(!):

    find /mnt/* -type f -exec getfattr -d {} \;
    getfattr: /mnt/testing-string: Input/output error

with console log:
    [268008.389781] hfsplus: unicode conversion failed

After the patch, both of the above works.

FYI, the test attribute string is prepared with:

echo -e -n \
"\xe9\x80\x99\xe6\x98\xaf\xe4\xb8\x80\xe5\x80\x8b\xe9\x9d\x9e\xe5" \
"\xb8\xb8\xe6\xbc\xab\xe9\x95\xb7\xe8\x80\x8c\xe6\xa5\xb5\xe5\x85" \
"\xb6\xe4\xb9\x8f\xe5\x91\xb3\xe5\x92\x8c\xe7\x9b\xb8\xe7\x95\xb6" \
"\xe7\x84\xa1\xe8\xb6\xa3\xe3\x80\x81\xe4\xbb\xa5\xe5\x8f\x8a\xe7" \
"\x84\xa1\xe7\x94\xa8\xe7\x9a\x84\xe3\x80\x81\xe5\x86\x8d\xe5\x8a" \
"\xa0\xe4\xb8\x8a\xe6\xaf\xab\xe7\x84\xa1\xe6\x84\x8f\xe7\xbe\xa9" \
"\xe7\x9a\x84\xe6\x93\xb4\xe5\xb1\x95\xe5\xb1\xac\xe6\x80\xa7\xef" \
"\xbc\x8c\xe8\x80\x8c\xe5\x85\xb6\xe5\x94\xaf\xe4\xb8\x80\xe5\x89" \
"\xb5\xe5\xbb\xba\xe7\x9b\xae\xe7\x9a\x84\xe5\x83\x85\xe6\x98\xaf" \
"\xe7\x82\xba\xe4\xba\x86\xe6\xb8\xac\xe8\xa9\xa6\xe4\xbd\x9c\xe7" \
"\x94\xa8\xe3\x80\x82" | tr -d ' '

(= "pointlessly long attribute for testing", elaborate Chinese in
UTF-8 enoding).

However, it is not possible to set double the size (110 + 5 is still
under 127) in a UTF-8 locale:

    $setfattr -n user.`cat testing-string testing-string` -v \
        `cat testing-string testing-string` testing-string
    setfattr: testing-string: Numerical result out of range

110 CJK char in UTF-8 is 330 bytes - the generic get/set attribute
system call code in linux/fs/xattr.c imposes a 255 byte limit.  One can
use a combination of iconv to encode content, changing terminal locale
for viewing, and an nls=cp932/cp936/cp949/cp950 mount option to fully
use 127-unicode attribute in a double-byte locale.

Also, as an additional information, it is possible to (mis-)use unicode
half-width/full-width forms (U+FFxx) to write attributes which looks
like english but not actually ascii.

Thanks Anton Altaparmakov for reviewing the earlier ideas behind this
change.

[akpm@linux-foundation.org: fix build]
[akpm@linux-foundation.org: fix build]
Signed-off-by: NHin-Tak Leung <htl10@users.sourceforge.net>
Cc: Anton Altaparmakov <anton@tuxera.com>
Cc: Vyacheslav Dubeyko <slava@dubeyko.com>
Cc: Al Viro <viro@zeniv.linux.org.uk>
Cc: Christoph Hellwig <hch@infradead.org>
Cc: Sougata Santra <sougata@tuxera.com>
Signed-off-by: NAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

This is a series of 3 patches which corrects issues in HFS+ concerning
the use of non-english file names and attributes.  Names and attributes
are stored internally as UTF-16 units up to a fixed maximum size, and
convert to and from user-representation by NLS.  The code incorrectly
assume that NLS string lengths are equal to unicode lengths, which is
only true for English ascii usage.

This patch (of 3):

The HFS Plus Volume Format specification (TN1150) states that file names
are stored internally as a maximum of 255 unicode characters, as defined
by The Unicode Standard, Version 2.0 [Unicode, Inc.  ISBN
0-201-48345-9].  File names are converted by the NLS system on Linux
before presented to the user.

255 CJK characters converts to UTF-8 with 1 unicode character to up to 3
bytes, and to GB18030 with 1 unicode character to up to 4 bytes.  Thus,
trying in a UTF-8 locale to list files with names of more than 85 CJK
characters results in:

    $ ls /mnt
    ls: reading directory /mnt: File name too long

The receiving buffer to hfsplus_uni2asc() needs to be 255 x
NLS_MAX_CHARSET_SIZE bytes, not 255 bytes as the code has always been.

Similar consideration applies to attributes, which are stored internally
as a maximum of 127 UTF-16BE units.  See XNU source for an up-to-date
reference on attributes.

Strictly speaking, the maximum value of NLS_MAX_CHARSET_SIZE = 6 is not
attainable in the case of conversion to UTF-8, as going beyond 3 bytes
requires the use of surrogate pairs, i.e.  consuming two input units.

Thanks Anton Altaparmakov for reviewing an earlier version of this
change.

This patch fixes all callers of hfsplus_uni2asc(), and also enables the
use of long non-English file names in HFS+.  The getting and setting,
and general usage of long non-English attributes requires further
forthcoming work, in the following patches of this series.

[akpm@linux-foundation.org: fix build]
Signed-off-by: NHin-Tak Leung <htl10@users.sourceforge.net>
Reviewed-by: NAnton Altaparmakov <anton@tuxera.com>
Cc: Vyacheslav Dubeyko <slava@dubeyko.com>
Cc: Al Viro <viro@zeniv.linux.org.uk>
Cc: Christoph Hellwig <hch@infradead.org>
Cc: Sougata Santra <sougata@tuxera.com>
Signed-off-by: NAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

Replace all function names by __func__ in pr_foo()
Signed-off-by: NFabian Frederick <fabf@skynet.be>
Cc: Jan Harkes <jaharkes@cs.cmu.edu>
Cc: Joe Perches <joe@perches.com>
Signed-off-by: NAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

- Add pr_fmt based on module name.

- Remove Coda: coda: from pr_foo()
Signed-off-by: NFabian Frederick <fabf@skynet.be>
Cc: Jan Harkes <jaharkes@cs.cmu.edu>
Cc: Joe Perches <joe@perches.com>
Signed-off-by: NAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

No level printk converted to pr_warn or pr_info
Signed-off-by: NFabian Frederick <fabf@skynet.be>
Cc: Jan Harkes <jaharkes@cs.cmu.edu>
Cc: Joe Perches <joe@perches.com>
Signed-off-by: NAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

Fix some comment errors.
Signed-off-by: NFabian Frederick <fabf@skynet.be>
Cc: Randy Dunlap <rdunlap@infradead.org>
Signed-off-by: NAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

sector_t is unsigned.
Signed-off-by: NFabian Frederick <fabf@skynet.be>
Signed-off-by: NAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

- strncpy + end of string assignement replaced by strlcpy

- Fix endif };

- Fix typo
Signed-off-by: NFabian Frederick <fabf@skynet.be>
Signed-off-by: NAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

strncpy + end of string assignment replaced by strlcpy
Signed-off-by: NFabian Frederick <fabf@skynet.be>
Signed-off-by: NAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

Replace seq_printf where possible.
Signed-off-by: NFabian Frederick <fabf@skynet.be>
Cc: Sage Weil <sage@inktank.com>
Signed-off-by: NAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

Update the last pr_warning callsites in fs branch
Signed-off-by: NFabian Frederick <fabf@skynet.be>
Cc: Sage Weil <sage@inktank.com>
Signed-off-by: NAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

autofs_dev_ioctl_init is only called by __init init_autofs4_fs
Signed-off-by: NFabian Frederick <fabf@skynet.be>
Acked-by: NIan Kent <raven@themaw.net>
Signed-off-by: NAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

Remove obsolete simple_strtoul in ncp_getopt
Signed-off-by: NFabian Frederick <fabf@skynet.be>
Cc: Petr Vandrovec <petr@vandrovec.name>
Signed-off-by: NAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

old_reloc() is only used in this file, make it static.
Signed-off-by: NAxel Lin <axel.lin@ingics.com>
Signed-off-by: NAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

Fix coccinelle warnings.
Signed-off-by: NFabian Frederick <fabf@skynet.be>
Signed-off-by: NAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

All KERN_DEBUG callsites being under #ifdef DEBUG we can safely convert
everything to pr_debug without changing current behaviour.

Remove #ifdef DEBUG around pr_debugs only (suggested by Joe Perches)
Signed-off-by: NFabian Frederick <fabf@skynet.be>
Cc: Joe Perches <joe@perches.com>
Signed-off-by: NAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

Also uniformize function arguments.
Signed-off-by: NFabian Frederick <fabf@skynet.be>
Cc: Joe Perches <joe@perches.com>
Signed-off-by: NAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

Convert all except KERN_DEBUG
(pr_debug doesn't work the same as printk(KERN_DEBUG and requires
special check)
Signed-off-by: NFabian Frederick <fabf@skynet.be>
Cc: Joe Perches <joe@perches.com>
Signed-off-by: NAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

Fixing 2 typo in function comments.
Signed-off-by: NFabian Frederick <fabf@skynet.be>
Cc: Al Viro <viro@zeniv.linux.org.uk>
Cc: "J. Bruce Fields" <bfields@redhat.com>
Signed-off-by: NAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

...like other filesystems.
Signed-off-by: NFabian Frederick <fabf@skynet.be>
Cc: Matthew Garrett <matthew.garrett@nebula.com>
Signed-off-by: NAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

We remove the call to grab_super_passive in call to super_cache_count.
This becomes a scalability bottleneck as multiple threads are trying to do
memory reclamation, e.g.  when we are doing large amount of file read and
page cache is under pressure.  The cached objects quickly got reclaimed
down to 0 and we are aborting the cache_scan() reclaim.  But counting
creates a log jam acquiring the sb_lock.

We are holding the shrinker_rwsem which ensures the safety of call to
list_lru_count_node() and s_op->nr_cached_objects.  The shrinker is
unregistered now before ->kill_sb() so the operation is safe when we are
doing unmount.

The impact will depend heavily on the machine and the workload but for a
small machine using postmark tuned to use 4xRAM size the results were

                                  3.15.0-rc5            3.15.0-rc5
                                     vanilla         shrinker-v1r1
Ops/sec Transactions         21.00 (  0.00%)       24.00 ( 14.29%)
Ops/sec FilesCreate          39.00 (  0.00%)       44.00 ( 12.82%)
Ops/sec CreateTransact       10.00 (  0.00%)       12.00 ( 20.00%)
Ops/sec FilesDeleted       6202.00 (  0.00%)     6202.00 (  0.00%)
Ops/sec DeleteTransact       11.00 (  0.00%)       12.00 (  9.09%)
Ops/sec DataRead/MB          25.97 (  0.00%)       29.10 ( 12.05%)
Ops/sec DataWrite/MB         49.99 (  0.00%)       56.02 ( 12.06%)

ffsb running in a configuration that is meant to simulate a mail server showed

                                 3.15.0-rc5             3.15.0-rc5
                                    vanilla          shrinker-v1r1
Ops/sec readall           9402.63 (  0.00%)      9567.97 (  1.76%)
Ops/sec create            4695.45 (  0.00%)      4735.00 (  0.84%)
Ops/sec delete             173.72 (  0.00%)       179.83 (  3.52%)
Ops/sec Transactions     14271.80 (  0.00%)     14482.81 (  1.48%)
Ops/sec Read                37.00 (  0.00%)        37.60 (  1.62%)
Ops/sec Write               18.20 (  0.00%)        18.30 (  0.55%)
Signed-off-by: NTim Chen <tim.c.chen@linux.intel.com>
Signed-off-by: NMel Gorman <mgorman@suse.de>
Cc: Johannes Weiner <hannes@cmpxchg.org>
Cc: Hugh Dickins <hughd@google.com>
Cc: Dave Chinner <david@fromorbit.com>
Tested-by: NYuanhan Liu <yuanhan.liu@linux.intel.com>
Cc: Bob Liu <bob.liu@oracle.com>
Cc: Jan Kara <jack@suse.cz>
Acked-by: NRik van Riel <riel@redhat.com>
Cc: Al Viro <viro@zeniv.linux.org.uk>
Signed-off-by: NAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

This series is aimed at regressions noticed during reclaim activity.  The
first two patches are shrinker patches that were posted ages ago but never
merged for reasons that are unclear to me.  I'm posting them again to see
if there was a reason they were dropped or if they just got lost.  Dave?
Time?  The last patch adjusts proportional reclaim.  Yuanhan Liu, can you
retest the vm scalability test cases on a larger machine?  Hugh, does this
work for you on the memcg test cases?

Based on ext4, I get the following results but unfortunately my larger
test machines are all unavailable so this is based on a relatively small
machine.

postmark
                                  3.15.0-rc5            3.15.0-rc5
                                     vanilla       proportion-v1r4
Ops/sec Transactions         21.00 (  0.00%)       25.00 ( 19.05%)
Ops/sec FilesCreate          39.00 (  0.00%)       45.00 ( 15.38%)
Ops/sec CreateTransact       10.00 (  0.00%)       12.00 ( 20.00%)
Ops/sec FilesDeleted       6202.00 (  0.00%)     6202.00 (  0.00%)
Ops/sec DeleteTransact       11.00 (  0.00%)       12.00 (  9.09%)
Ops/sec DataRead/MB          25.97 (  0.00%)       30.02 ( 15.59%)
Ops/sec DataWrite/MB         49.99 (  0.00%)       57.78 ( 15.58%)

ffsb (mail server simulator)
                                 3.15.0-rc5             3.15.0-rc5
                                    vanilla        proportion-v1r4
Ops/sec readall           9402.63 (  0.00%)      9805.74 (  4.29%)
Ops/sec create            4695.45 (  0.00%)      4781.39 (  1.83%)
Ops/sec delete             173.72 (  0.00%)       177.23 (  2.02%)
Ops/sec Transactions     14271.80 (  0.00%)     14764.37 (  3.45%)
Ops/sec Read                37.00 (  0.00%)        38.50 (  4.05%)
Ops/sec Write               18.20 (  0.00%)        18.50 (  1.65%)

dd of a large file
                                3.15.0-rc5            3.15.0-rc5
                                   vanilla       proportion-v1r4
WallTime DownloadTar       75.00 (  0.00%)       61.00 ( 18.67%)
WallTime DD               423.00 (  0.00%)      401.00 (  5.20%)
WallTime Delete             2.00 (  0.00%)        5.00 (-150.00%)

stutter (times mmap latency during large amounts of IO)

                            3.15.0-rc5            3.15.0-rc5
                               vanilla       proportion-v1r4
Unit >5ms Delays  80252.0000 (  0.00%)  81523.0000 ( -1.58%)
Unit Mmap min         8.2118 (  0.00%)      8.3206 ( -1.33%)
Unit Mmap mean       17.4614 (  0.00%)     17.2868 (  1.00%)
Unit Mmap stddev     24.9059 (  0.00%)     34.6771 (-39.23%)
Unit Mmap max      2811.6433 (  0.00%)   2645.1398 (  5.92%)
Unit Mmap 90%        20.5098 (  0.00%)     18.3105 ( 10.72%)
Unit Mmap 93%        22.9180 (  0.00%)     20.1751 ( 11.97%)
Unit Mmap 95%        25.2114 (  0.00%)     22.4988 ( 10.76%)
Unit Mmap 99%        46.1430 (  0.00%)     43.5952 (  5.52%)
Unit Ideal  Tput     85.2623 (  0.00%)     78.8906 (  7.47%)
Unit Tput min        44.0666 (  0.00%)     43.9609 (  0.24%)
Unit Tput mean       45.5646 (  0.00%)     45.2009 (  0.80%)
Unit Tput stddev      0.9318 (  0.00%)      1.1084 (-18.95%)
Unit Tput max        46.7375 (  0.00%)     46.7539 ( -0.04%)

This patch (of 3):

We will like to unregister the sb shrinker before ->kill_sb().  This will
allow cached objects to be counted without call to grab_super_passive() to
update ref count on sb.  We want to avoid locking during memory
reclamation especially when we are skipping the memory reclaim when we are
out of cached objects.

This is safe because grab_super_passive does a try-lock on the
sb->s_umount now, and so if we are in the unmount process, it won't ever
block.  That means what used to be a deadlock and races we were avoiding
by using grab_super_passive() is now:

        shrinker                        umount

        down_read(shrinker_rwsem)
                                        down_write(sb->s_umount)
                                        shrinker_unregister
                                          down_write(shrinker_rwsem)
                                            <blocks>
        grab_super_passive(sb)
          down_read_trylock(sb->s_umount)
            <fails>
        <shrinker aborts>
        ....
        <shrinkers finish running>
        up_read(shrinker_rwsem)
                                          <unblocks>
                                          <removes shrinker>
                                          up_write(shrinker_rwsem)
                                        ->kill_sb()
                                        ....

So it is safe to deregister the shrinker before ->kill_sb().
Signed-off-by: NTim Chen <tim.c.chen@linux.intel.com>
Signed-off-by: NMel Gorman <mgorman@suse.de>
Cc: Johannes Weiner <hannes@cmpxchg.org>
Cc: Hugh Dickins <hughd@google.com>
Cc: Dave Chinner <david@fromorbit.com>
Tested-by: NYuanhan Liu <yuanhan.liu@linux.intel.com>
Cc: Bob Liu <bob.liu@oracle.com>
Cc: Jan Kara <jack@suse.cz>
Acked-by: NRik van Riel <riel@redhat.com>
Cc: Al Viro <viro@zeniv.linux.org.uk>
Signed-off-by: NAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

Fix checkpatch warning:
WARNING: kfree(NULL) is safe this check is probably not required
Signed-off-by: NFabian Frederick <fabf@skynet.be>
Signed-off-by: NAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

...like other filesystems.
Signed-off-by: NFabian Frederick <fabf@skynet.be>
Signed-off-by: NAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

hugetlbfs_i_mmap_mutex_key is only used in inode.c
Signed-off-by: NFabian Frederick <fabf@skynet.be>
Signed-off-by: NAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

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: NMel Gorman <mgorman@suse.de>
Cc: Johannes Weiner <hannes@cmpxchg.org>
Cc: Vlastimil Babka <vbabka@suse.cz>
Cc: Jan Kara <jack@suse.cz>
Cc: Michal Hocko <mhocko@suse.cz>
Cc: Hugh Dickins <hughd@google.com>
Cc: Dave Hansen <dave.hansen@intel.com>
Cc: Theodore Ts'o <tytso@mit.edu>
Cc: "Paul E. McKenney" <paulmck@linux.vnet.ibm.com>
Cc: Oleg Nesterov <oleg@redhat.com>
Cc: Rik van Riel <riel@redhat.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Tested-by: NPrabhakar Lad <prabhakar.csengg@gmail.com>
Signed-off-by: NAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

Discarding buffers uses a bunch of atomic operations when discarding
buffers because ......  I can't think of a reason.  Use a cmpxchg loop to
clear all the necessary flags.  In most (all?) cases this will be a single
atomic operations.

[akpm@linux-foundation.org: move BUFFER_FLAGS_DISCARD into the .c file]
Signed-off-by: NMel Gorman <mgorman@suse.de>
Cc: Johannes Weiner <hannes@cmpxchg.org>
Cc: Vlastimil Babka <vbabka@suse.cz>
Cc: Jan Kara <jack@suse.cz>
Cc: Michal Hocko <mhocko@suse.cz>
Cc: Hugh Dickins <hughd@google.com>
Cc: Dave Hansen <dave.hansen@intel.com>
Cc: Theodore Ts'o <tytso@mit.edu>
Cc: "Paul E. McKenney" <paulmck@linux.vnet.ibm.com>
Cc: Oleg Nesterov <oleg@redhat.com>
Cc: Rik van Riel <riel@redhat.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Signed-off-by: NAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

cold is a bool, make it one.  Make the likely case the "if" part of the
block instead of the else as according to the optimisation manual this is
preferred.
Signed-off-by: NMel Gorman <mgorman@suse.de>
Acked-by: NRik van Riel <riel@redhat.com>
Cc: Johannes Weiner <hannes@cmpxchg.org>
Cc: Vlastimil Babka <vbabka@suse.cz>
Cc: Jan Kara <jack@suse.cz>
Cc: Michal Hocko <mhocko@suse.cz>
Cc: Hugh Dickins <hughd@google.com>
Cc: Dave Hansen <dave.hansen@intel.com>
Cc: Theodore Ts'o <tytso@mit.edu>
Cc: "Paul E. McKenney" <paulmck@linux.vnet.ibm.com>
Cc: Oleg Nesterov <oleg@redhat.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Signed-off-by: NAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

A block device driver may choose to provide a rw_page operation.  These
will be called when the filesystem is attempting to do page sized I/O to
page cache pages (ie not for direct I/O).  This does preclude I/Os that
are larger than page size, so this may only be a performance gain for
some devices.
Signed-off-by: NMatthew Wilcox <matthew.r.wilcox@intel.com>
Tested-by: NDheeraj Reddy <dheeraj.reddy@intel.com>
Cc: Dave Chinner <david@fromorbit.com>
Cc: Hugh Dickins <hughd@google.com>
Signed-off-by: NAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

page_endio() takes care of updating all the appropriate page flags once
I/O has finished to a page.  Switch to using mapping_set_error() instead
of setting AS_EIO directly; this will handle thin-provisioned devices
correctly.
Signed-off-by: NMatthew Wilcox <matthew.r.wilcox@intel.com>
Cc: Dave Chinner <david@fromorbit.com>
Cc: Dheeraj Reddy <dheeraj.reddy@intel.com>
Cc: Hugh Dickins <hughd@google.com>
Signed-off-by: NAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

__mpage_writepage() is over 200 lines long, has 20 local variables, four
goto labels and could desperately use simplification.  Splitting
clean_buffers() into a helper function improves matters a little,
removing 20+ lines from it.
Signed-off-by: NMatthew Wilcox <matthew.r.wilcox@intel.com>
Cc: Dave Chinner <david@fromorbit.com>
Cc: Dheeraj Reddy <dheeraj.reddy@intel.com>
Cc: Hugh Dickins <hughd@google.com>
Signed-off-by: NAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>