linux/fs
Mel Gorman e9c15badbb fs: Do not check if there is a fsnotify watcher on pseudo inodes
The kernel uses internal mounts created by kern_mount() and populated
with files with no lookup path by alloc_file_pseudo() for a variety of
reasons. An example of such a mount is for anonymous pipes. For pipes,
every vfs_write() regardless of filesystem, calls fsnotify_modify()
to notify of any changes which incurs a small amount of overhead in
fsnotify even when there are no watchers. It can also trigger for reads
and readv and writev, it was simply vfs_write() that was noticed first.

A patch is pending that reduces, but does not eliminate, the overhead of
fsnotify but for files that cannot be looked up via a path, even that
small overhead is unnecessary. The user API for all notification
subsystems (inotify, fanotify, ...) is based on the pathname and a dirfd
and proc entries appear to be the only visible representation of the
files. Proc does not have the same pathname as the internal entry and
the proc inode is not the same as the internal inode so even if fanotify
is used on a file under /proc/XX/fd, no useful events are notified.

This patch changes alloc_file_pseudo() to always opt out of fsnotify by
setting FMODE_NONOTIFY flag so that no check is made for fsnotify
watchers on pseudo files. This should be safe as the underlying helper
for the dentry is d_alloc_pseudo() which explicitly states that no
lookups are ever performed meaning that fanotify should have nothing
useful to attach to.

The test motivating this was "perf bench sched messaging --pipe". On
a single-socket machine using threads the difference of the patch was
as follows.

                              5.7.0                  5.7.0
                            vanilla        nofsnotify-v1r1
Amean     1       1.3837 (   0.00%)      1.3547 (   2.10%)
Amean     3       3.7360 (   0.00%)      3.6543 (   2.19%)
Amean     5       5.8130 (   0.00%)      5.7233 *   1.54%*
Amean     7       8.1490 (   0.00%)      7.9730 *   2.16%*
Amean     12     14.6843 (   0.00%)     14.1820 (   3.42%)
Amean     18     21.8840 (   0.00%)     21.7460 (   0.63%)
Amean     24     28.8697 (   0.00%)     29.1680 (  -1.03%)
Amean     30     36.0787 (   0.00%)     35.2640 *   2.26%*
Amean     32     38.0527 (   0.00%)     38.1223 (  -0.18%)

The difference is small but in some cases it's outside the noise so
while marginal, there is still some small benefit to ignoring fsnotify
for files allocated via alloc_file_pseudo() in some cases.

Link: https://lore.kernel.org/r/20200615121358.GF3183@techsingularity.net
Signed-off-by: Mel Gorman <mgorman@techsingularity.net>
Reviewed-by: Amir Goldstein <amir73il@gmail.com>
Signed-off-by: Jan Kara <jack@suse.cz>
2020-06-16 09:40:45 +02:00
..
9p 9p: read only once on O_NONBLOCK 2020-03-27 09:29:56 +00:00
adfs docs: filesystems: fix renamed references 2020-04-20 15:45:22 -06:00
affs docs: filesystems: fix renamed references 2020-04-20 15:45:22 -06:00
afs afs: Fix afs_store_data() to set mtime in new operation descriptor 2020-06-11 16:04:30 -07:00
autofs
befs
bfs docs: filesystems: fix renamed references 2020-04-20 15:45:22 -06:00
btrfs for-5.8-part2-tag 2020-06-14 09:47:25 -07:00
cachefiles A fair amount of stuff this time around, dominated by yet another massive 2020-06-01 15:45:27 -07:00
ceph ceph: skip checking caps when session reconnecting and releasing reqs 2020-06-01 13:22:53 +02:00
cifs smb3: Add debug message for new file creation with idsfromsid mount option 2020-06-12 16:31:06 -05:00
coda docs: filesystems: convert coda.txt to ReST 2020-05-05 09:22:21 -06:00
configfs A fair amount of stuff this time around, dominated by yet another massive 2020-06-01 15:45:27 -07:00
cramfs docs: filesystems: fix renamed references 2020-04-20 15:45:22 -06:00
crypto fscrypt updates for 5.8 2020-06-01 12:10:17 -07:00
debugfs Merge 5.7-rc3 into driver-core-next 2020-04-27 09:34:55 +02:00
devpts
dlm dlm for 5.8 2020-06-05 16:43:16 -07:00
ecryptfs A fair amount of stuff this time around, dominated by yet another massive 2020-06-01 15:45:27 -07:00
efivarfs
efs
erofs Changes since last update: 2020-06-02 20:16:55 -07:00
exfat Description for this pull request: 2020-06-09 11:24:59 -07:00
exportfs
ext2 mmap locking API: convert mmap_sem comments 2020-06-09 09:39:14 -07:00
ext4 This is the second round of ext4 commits for 5.8 merge window. It 2020-06-15 09:32:10 -07:00
f2fs f2fs-for-5.8-rc1 2020-06-09 11:28:59 -07:00
fat fat: improve the readahead for FAT entries 2020-06-04 19:06:25 -07:00
freevxfs
fscache Merge git://git.kernel.org/pub/scm/linux/kernel/git/netdev/net-next 2020-06-03 16:27:18 -07:00
fuse fuse update for 5.8 2020-06-09 15:48:24 -07:00
gfs2 Changes in gfs2: 2020-06-08 12:47:09 -07:00
hfs for-5.8/block-2020-06-01 2020-06-02 15:29:19 -07:00
hfsplus block: remove the error_sector argument to blkdev_issue_flush 2020-05-22 08:45:46 -06:00
hostfs hostfs: Use kasprintf() instead of fixed buffer formatting 2020-03-29 23:23:00 +02:00
hpfs hpfs: fix warning due to superfluous semicolon 2020-06-06 10:08:17 -07:00
hugetlbfs mmap locking API: convert mmap_sem API comments 2020-06-09 09:39:14 -07:00
iomap New code for 5.8: 2020-06-13 12:44:30 -07:00
isofs for-5.8/block-2020-06-01 2020-06-02 15:29:19 -07:00
jbd2 This is the second round of ext4 commits for 5.8 merge window. It 2020-06-15 09:32:10 -07:00
jffs2
jfs Replace zero-length array in JFS 2020-06-02 20:11:35 -07:00
kernfs mmap locking API: convert mmap_sem comments 2020-06-09 09:39:14 -07:00
lockd
minix
nfs NFS Client Updates for Linux 5.8 2020-06-11 12:22:41 -07:00
nfs_common
nfsd Highlights: 2020-06-11 10:33:13 -07:00
nilfs2 nilfs2: fix null pointer dereference at nilfs_segctor_do_construct() 2020-06-10 19:14:17 -07:00
nls treewide: replace '---help---' in Kconfig files with 'help' 2020-06-14 01:57:21 +09:00
notify treewide: replace '---help---' in Kconfig files with 'help' 2020-06-14 01:57:21 +09:00
ntfs Merge branch 'akpm' (patches from Andrew) 2020-06-02 12:21:36 -07:00
ocfs2 ocfs2: fix build failure when TCP/IP is disabled 2020-06-11 18:17:47 -07:00
omfs fs: convert mpage_readpages to mpage_readahead 2020-06-02 10:59:07 -07:00
openpromfs
orangefs orangefs: a conversion and a cleanup... 2020-06-05 16:44:36 -07:00
overlayfs overlayfs update for 5.8 2020-06-09 15:40:50 -07:00
proc Kbuild updates for v5.8 (2nd) 2020-06-13 13:29:16 -07:00
pstore Merge branch 'uaccess.__copy_from_user' of git://git.kernel.org/pub/scm/linux/kernel/git/viro/vfs 2020-06-01 16:18:46 -07:00
qnx4
qnx6 fs: convert mpage_readpages to mpage_readahead 2020-06-02 10:59:07 -07:00
quota sysctl: pass kernel pointers to ->proc_handler 2020-04-27 02:07:40 -04:00
ramfs
reiserfs \n 2020-06-04 13:53:10 -07:00
romfs treewide: replace '---help---' in Kconfig files with 'help' 2020-06-14 01:57:21 +09:00
squashfs Merge branch 'akpm' (patches from Andrew) 2020-06-02 12:21:36 -07:00
sysfs RDMA 5.8 merge window pull request 2020-06-05 14:05:57 -07:00
sysv docs: filesystems: fix renamed references 2020-04-20 15:45:22 -06:00
tracefs
ubifs mm: remove the pgprot argument to __vmalloc 2020-06-02 10:59:11 -07:00
udf for-5.8/block-2020-06-01 2020-06-02 15:29:19 -07:00
ufs
unicode .gitignore: add SPDX License Identifier 2020-03-25 11:50:48 +01:00
vboxsf vboxsf: don't use the source name in the bdi name 2020-05-07 08:45:47 -06:00
verity fs-verity: remove unnecessary extern keywords 2020-05-12 16:44:00 -07:00
xfs Fixes for 5.8: 2020-06-13 12:40:24 -07:00
zonefs zonefs changes for 5.8 2020-06-04 13:50:13 -07:00
aio.c kernel: move use_mm/unuse_mm to kthread.c 2020-06-10 19:14:18 -07:00
anon_inodes.c
attr.c
bad_inode.c fs: move the fiemap definitions out of fs.h 2020-06-03 23:16:55 -04:00
binfmt_aout.c exec: Rename flush_old_exec begin_new_exec 2020-05-07 16:55:47 -05:00
binfmt_elf_fdpic.c Merge branch 'uaccess.misc' of git://git.kernel.org/pub/scm/linux/kernel/git/viro/vfs 2020-06-10 16:02:54 -07:00
binfmt_elf.c Merge branch 'uaccess.misc' of git://git.kernel.org/pub/scm/linux/kernel/git/viro/vfs 2020-06-10 16:02:54 -07:00
binfmt_em86.c Merge branch 'akpm' (patches from Andrew) 2020-06-04 19:18:29 -07:00
binfmt_flat.c Merge branch 'uaccess.misc' of git://git.kernel.org/pub/scm/linux/kernel/git/viro/vfs 2020-06-10 16:02:54 -07:00
binfmt_misc.c Merge branch 'akpm' (patches from Andrew) 2020-06-04 19:18:29 -07:00
binfmt_script.c Merge branch 'akpm' (patches from Andrew) 2020-06-04 19:18:29 -07:00
block_dev.c for-5.8/drivers-2020-06-01 2020-06-02 15:37:03 -07:00
buffer.c fs/buffer.c: use attach/detach_page_private 2020-06-02 10:59:07 -07:00
char_dev.c vfs: allow unprivileged whiteout creation 2020-05-14 16:44:23 +02:00
compat_binfmt_elf.c Split the old READ_IMPLIES_EXEC workaround from executable PT_GNU_STACK 2020-06-05 13:45:21 -07:00
compat.c
coredump.c mmap locking API: convert mmap_sem comments 2020-06-09 09:39:14 -07:00
d_path.c
dax.c dax,iomap: Add helper dax_iomap_zero() to zero a range 2020-04-02 19:15:03 -07:00
dcache.c Merge git://git.kernel.org/pub/scm/linux/kernel/git/netdev/net-next 2020-06-03 16:27:18 -07:00
dcookies.c
direct-io.c for-5.8-part2-tag 2020-06-14 09:47:25 -07:00
drop_caches.c sysctl: pass kernel pointers to ->proc_handler 2020-04-27 02:07:40 -04:00
eventfd.c eventfd: convert to f_op->read_iter() 2020-05-06 22:33:43 -04:00
eventpoll.c epoll: call final ep_events_available() check under the lock 2020-05-14 10:00:35 -07:00
exec.c mmap locking API: convert mmap_sem comments 2020-06-09 09:39:14 -07:00
fcntl.c
fhandle.c
file_table.c fs: Do not check if there is a fsnotify watcher on pseudo inodes 2020-06-16 09:40:45 +02:00
file.c fix multiplication overflow in copy_fdtable() 2020-05-19 18:29:36 -04:00
filesystems.c fs/filesystems.c: downgrade user-reachable WARN_ONCE() to pr_warn_once() 2020-04-10 15:36:22 -07:00
fs_context.c vfs: don't parse "silent" option 2020-05-14 16:44:25 +02:00
fs_parser.c fs_parse: remove pr_notice() about each validation 2020-04-02 09:35:26 -07:00
fs_pin.c
fs_struct.c
fs_types.c
fs-writeback.c A lot of bug fixes and cleanups for ext4, including: 2020-06-05 16:19:28 -07:00
fsopen.c
inode.c AFS Changes 2020-06-05 16:26:36 -07:00
internal.h A lot of bug fixes and cleanups for ext4, including: 2020-06-05 16:19:28 -07:00
io_uring.c io_uring-5.8-2020-06-11 2020-06-11 16:10:08 -07:00
io-wq.c io_uring-5.8-2020-06-11 2020-06-11 16:10:08 -07:00
io-wq.h io_uring: avoid whole io_wq_work copy for requests completed inline 2020-06-10 17:58:46 -06:00
ioctl.c fs: remove the access_ok() check in ioctl_fiemap 2020-06-03 23:16:55 -04:00
Kconfig treewide: replace '---help---' in Kconfig files with 'help' 2020-06-14 01:57:21 +09:00
Kconfig.binfmt treewide: replace '---help---' in Kconfig files with 'help' 2020-06-14 01:57:21 +09:00
libfs.c block: remove the error_sector argument to blkdev_issue_flush 2020-05-22 08:45:46 -06:00
locks.c Highlights: 2020-06-11 10:33:13 -07:00
Makefile
mbcache.c
mount.h proc/mounts: add cursor 2020-05-14 16:44:24 +02:00
mpage.c fs: convert mpage_readpages to mpage_readahead 2020-06-02 10:59:07 -07:00
namei.c vfs: clean up posix_acl_permission() logic aroudn MAY_NOT_BLOCK 2020-06-08 11:04:19 -07:00
namespace.c Merge branch 'work.misc' of git://git.kernel.org/pub/scm/linux/kernel/git/viro/vfs 2020-06-10 16:09:11 -07:00
no-block.c
nsfs.c nsproxy: attach to namespaces via pidfds 2020-05-13 11:41:22 +02:00
open.c Merge branch 'akpm' (patches from Andrew) 2020-06-02 12:21:36 -07:00
pipe.c Notifications over pipes + Keyring notifications 2020-06-13 09:56:21 -07:00
pnode.c propagate_one(): mnt_set_mountpoint() needs mount_lock 2020-04-27 10:37:14 -04:00
pnode.h
posix_acl.c vfs: clean up posix_acl_permission() logic aroudn MAY_NOT_BLOCK 2020-06-08 11:04:19 -07:00
proc_namespace.c Merge branch 'proc-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/ebiederm/user-namespace 2020-06-04 13:54:34 -07:00
read_write.c powerpc: Add back __ARCH_WANT_SYS_LLSEEK macro 2020-04-03 00:09:59 +11:00
readdir.c readdir.c: get rid of the last __put_user(), drop now-useless access_ok() 2020-05-01 20:29:54 -04:00
select.c pselect6() and friends: take handling the combined 6th/7th args into helper 2020-05-29 19:10:42 -04:00
seq_file.c fs/seq_file.c: seq_read: Update pr_info_ratelimited 2020-06-04 19:06:25 -07:00
signalfd.c
splice.c Notifications over pipes + Keyring notifications 2020-06-13 09:56:21 -07:00
stack.c
stat.c New code for 5.8: 2020-06-02 19:45:12 -07:00
statfs.c
super.c Merge branch 'work.misc' of git://git.kernel.org/pub/scm/linux/kernel/git/viro/vfs 2020-06-10 16:09:11 -07:00
sync.c overlayfs update for 5.8 2020-06-09 15:40:50 -07:00
timerfd.c
userfaultfd.c mmap locking API: convert mmap_sem comments 2020-06-09 09:39:14 -07:00
utimes.c utimensat: AT_EMPTY_PATH support 2020-05-14 16:44:24 +02:00
xattr.c xattr: fix uninitialized out-param 2020-04-09 15:33:09 -04:00