linux/fs
Linus Torvalds d7852fbd0f access: avoid the RCU grace period for the temporary subjective credentials
It turns out that 'access()' (and 'faccessat()') can cause a lot of RCU
work because it installs a temporary credential that gets allocated and
freed for each system call.

The allocation and freeing overhead is mostly benign, but because
credentials can be accessed under the RCU read lock, the freeing
involves a RCU grace period.

Which is not a huge deal normally, but if you have a lot of access()
calls, this causes a fair amount of seconday damage: instead of having a
nice alloc/free patterns that hits in hot per-CPU slab caches, you have
all those delayed free's, and on big machines with hundreds of cores,
the RCU overhead can end up being enormous.

But it turns out that all of this is entirely unnecessary.  Exactly
because access() only installs the credential as the thread-local
subjective credential, the temporary cred pointer doesn't actually need
to be RCU free'd at all.  Once we're done using it, we can just free it
synchronously and avoid all the RCU overhead.

So add a 'non_rcu' flag to 'struct cred', which can be set by users that
know they only use it in non-RCU context (there are other potential
users for this).  We can make it a union with the rcu freeing list head
that we need for the RCU case, so this doesn't need any extra storage.

Note that this also makes 'get_current_cred()' clear the new non_rcu
flag, in case we have filesystems that take a long-term reference to the
cred and then expect the RCU delayed freeing afterwards.  It's not
entirely clear that this is required, but it makes for clear semantics:
the subjective cred remains non-RCU as long as you only access it
synchronously using the thread-local accessors, but you _can_ use it as
a generic cred if you want to.

It is possible that we should just remove the whole RCU markings for
->cred entirely.  Only ->real_cred is really supposed to be accessed
through RCU, and the long-term cred copies that nfs uses might want to
explicitly re-enable RCU freeing if required, rather than have
get_current_cred() do it implicitly.

But this is a "minimal semantic changes" change for the immediate
problem.

Acked-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Acked-by: Eric Dumazet <edumazet@google.com>
Acked-by: Paul E. McKenney <paulmck@linux.ibm.com>
Cc: Oleg Nesterov <oleg@redhat.com>
Cc: Jan Glauber <jglauber@marvell.com>
Cc: Jiri Kosina <jikos@kernel.org>
Cc: Jayachandran Chandrasekharan Nair <jnair@marvell.com>
Cc: Greg KH <greg@kroah.com>
Cc: Kees Cook <keescook@chromium.org>
Cc: David Howells <dhowells@redhat.com>
Cc: Miklos Szeredi <miklos@szeredi.hu>
Cc: Al Viro <viro@zeniv.linux.org.uk>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2019-07-24 10:12:09 -07:00
..
9p treewide: Replace GPLv2 boilerplate/reference with SPDX - rule 188 2019-05-30 11:29:21 -07:00
adfs treewide: Replace GPLv2 boilerplate/reference with SPDX - rule 500 2019-06-19 17:09:55 +02:00
affs treewide: Add SPDX license identifier - Makefile/Kconfig 2019-05-21 10:50:46 +02:00
afs AFS fixes 2019-06-28 08:34:12 +08:00
autofs treewide: Replace GPLv2 boilerplate/reference with SPDX - rule 83 2019-05-24 17:37:52 +02:00
befs treewide: Add SPDX license identifier - Makefile/Kconfig 2019-05-21 10:50:46 +02:00
bfs treewide: Add SPDX license identifier - Makefile/Kconfig 2019-05-21 10:50:46 +02:00
btrfs for-5.2-rc5-tag 2019-06-18 11:20:24 -07:00
cachefiles treewide: Replace GPLv2 boilerplate/reference with SPDX - rule 36 2019-05-24 17:27:11 +02:00
ceph ceph: fix ceph_mdsc_build_path to not stop on first component 2019-06-27 18:27:36 +02:00
cifs cifs: fix crash querying symlinks stored as reparse-points 2019-06-28 00:34:17 -05:00
coda treewide: Replace GPLv2 boilerplate/reference with SPDX - rule 152 2019-05-30 11:26:32 -07:00
configfs SPDX update for 5.2-rc3, round 1 2019-05-31 08:34:32 -07:00
cramfs treewide: Add SPDX license identifier - Makefile/Kconfig 2019-05-21 10:50:46 +02:00
crypto treewide: Add SPDX license identifier - Makefile/Kconfig 2019-05-21 10:50:46 +02:00
debugfs treewide: Add SPDX license identifier - Makefile/Kconfig 2019-05-21 10:50:46 +02:00
devpts treewide: Replace GPLv2 boilerplate/reference with SPDX - rule 83 2019-05-24 17:37:52 +02:00
dlm treewide: Replace GPLv2 boilerplate/reference with SPDX - rule 398 2019-06-05 17:37:12 +02:00
ecryptfs treewide: Replace GPLv2 boilerplate/reference with SPDX - rule 333 2019-06-05 17:37:06 +02:00
efivarfs treewide: Replace GPLv2 boilerplate/reference with SPDX - rule 500 2019-06-19 17:09:55 +02:00
efs treewide: Add SPDX license identifier - Makefile/Kconfig 2019-05-21 10:50:46 +02:00
exportfs treewide: Add SPDX license identifier - Makefile/Kconfig 2019-05-21 10:50:46 +02:00
ext2 treewide: Add SPDX license identifier - Makefile/Kconfig 2019-05-21 10:50:46 +02:00
ext4 Bug fixes (including a regression fix) for ext4. 2019-05-25 15:03:12 -07:00
f2fs treewide: Add SPDX license identifier - Makefile/Kconfig 2019-05-21 10:50:46 +02:00
fat treewide: Replace GPLv2 boilerplate/reference with SPDX - rule 282 2019-06-05 17:36:37 +02:00
freevxfs treewide: Add SPDX license identifier - Makefile/Kconfig 2019-05-21 10:50:46 +02:00
fscache treewide: Replace GPLv2 boilerplate/reference with SPDX - rule 152 2019-05-30 11:26:32 -07:00
fuse Revert "fuse: require /dev/fuse reads to have enough buffer capacity" 2019-06-11 13:35:22 +02:00
gfs2 Fix rounding error in gfs2_iomap_page_prepare 2019-06-14 17:27:12 -10:00
hfs treewide: Add SPDX license identifier - Makefile/Kconfig 2019-05-21 10:50:46 +02:00
hfsplus treewide: Add SPDX license identifier - Makefile/Kconfig 2019-05-21 10:50:46 +02:00
hostfs This pull request contains the following changes for UML: 2019-05-12 17:52:13 -04:00
hpfs treewide: Add SPDX license identifier - Makefile/Kconfig 2019-05-21 10:50:46 +02:00
hugetlbfs treewide: Add SPDX license identifier - Makefile/Kconfig 2019-05-21 10:50:46 +02:00
isofs treewide: Replace GPLv2 boilerplate/reference with SPDX - rule 142 2019-05-30 11:25:17 -07:00
jbd2 treewide: Add SPDX license identifier - Makefile/Kconfig 2019-05-21 10:50:46 +02:00
jffs2 treewide: Add SPDX license identifier - Makefile/Kconfig 2019-05-21 10:50:46 +02:00
jfs treewide: Replace GPLv2 boilerplate/reference with SPDX - rule 156 2019-05-30 11:26:35 -07:00
kernfs treewide: Replace GPLv2 boilerplate/reference with SPDX - rule 428 2019-06-05 17:37:16 +02:00
lockd Revert "lockd: Show pid of lockd for remote locks" 2019-05-31 09:43:26 -04:00
minix treewide: Add SPDX license identifier - Makefile/Kconfig 2019-05-21 10:50:46 +02:00
nfs NFS/flexfiles: Use the correct TCP timeout for flexfiles I/O 2019-06-28 11:48:52 -04:00
nfs_common treewide: Add SPDX license identifier - Makefile/Kconfig 2019-05-21 10:50:46 +02:00
nfsd nfsd: Fix overflow causing non-working mounts on 1 TB machines 2019-07-03 17:51:31 -04:00
nilfs2 treewide: Add SPDX license identifier - Makefile/Kconfig 2019-05-21 10:50:46 +02:00
nls treewide: Add SPDX license identifier - Makefile/Kconfig 2019-05-21 10:50:46 +02:00
notify fanotify: update connector fsid cache on add mark 2019-06-19 15:53:58 +02:00
ntfs treewide: Replace GPLv2 boilerplate/reference with SPDX - rule 97 2019-05-24 17:37:53 +02:00
ocfs2 fs/ocfs2: fix race in ocfs2_dentry_attach_lock() 2019-06-13 17:34:56 -10:00
omfs treewide: Replace GPLv2 boilerplate/reference with SPDX - rule 209 2019-05-30 11:29:53 -07:00
openpromfs treewide: Add SPDX license identifier - Makefile/Kconfig 2019-05-21 10:50:46 +02:00
orangefs treewide: Add SPDX license identifier - Makefile/Kconfig 2019-05-21 10:50:46 +02:00
overlayfs SPDX update for 5.2-rc6 2019-06-21 09:58:42 -07:00
proc fs/proc/array.c: allow reporting eip/esp for all coredumping threads 2019-06-29 16:43:44 +08:00
pstore SPDX update for 5.2-rc4 2019-06-08 12:52:42 -07:00
qnx4 treewide: Add SPDX license identifier - Makefile/Kconfig 2019-05-21 10:50:46 +02:00
qnx6 treewide: Add SPDX license identifier - Makefile/Kconfig 2019-05-21 10:50:46 +02:00
quota quota: fix a problem about transfer quota 2019-06-19 15:25:41 +02:00
ramfs treewide: Replace GPLv2 boilerplate/reference with SPDX - rule 152 2019-05-30 11:26:32 -07:00
reiserfs treewide: Add SPDX license identifier - Makefile/Kconfig 2019-05-21 10:50:46 +02:00
romfs treewide: Replace GPLv2 boilerplate/reference with SPDX - rule 152 2019-05-30 11:26:32 -07:00
squashfs treewide: Replace GPLv2 boilerplate/reference with SPDX - rule 499 2019-06-19 17:09:53 +02:00
sysfs treewide: Add SPDX license identifier - Makefile/Kconfig 2019-05-21 10:50:46 +02:00
sysv treewide: Add SPDX license identifier - Makefile/Kconfig 2019-05-21 10:50:46 +02:00
tracefs treewide: Replace GPLv2 boilerplate/reference with SPDX - rule 500 2019-06-19 17:09:55 +02:00
ubifs treewide: Replace GPLv2 boilerplate/reference with SPDX - rule 336 2019-06-05 17:37:07 +02:00
udf treewide: Add SPDX license identifier - Makefile/Kconfig 2019-05-21 10:50:46 +02:00
ufs treewide: Add SPDX license identifier - Makefile/Kconfig 2019-05-21 10:50:46 +02:00
unicode treewide: Replace GPLv2 boilerplate/reference with SPDX - rule 294 2019-06-05 17:36:38 +02:00
xfs block: return from __bio_try_merge_page if merging occured in the same page 2019-06-17 09:33:02 -06:00
aio.c signal: remove the wrong signal_pending() check in restore_user_sigmask() 2019-06-29 16:43:45 +08:00
anon_inodes.c treewide: Add SPDX license identifier for missed files 2019-05-21 10:50:45 +02:00
attr.c
bad_inode.c
binfmt_aout.c treewide: Add SPDX license identifier for more missed files 2019-05-21 10:50:45 +02:00
binfmt_elf_fdpic.c treewide: Replace GPLv2 boilerplate/reference with SPDX - rule 152 2019-05-30 11:26:32 -07:00
binfmt_elf.c treewide: Add SPDX license identifier for more missed files 2019-05-21 10:50:45 +02:00
binfmt_em86.c treewide: Add SPDX license identifier for more missed files 2019-05-21 10:50:45 +02:00
binfmt_flat.c fs/binfmt_flat.c: make load_flat_shared_library() work 2019-06-29 16:43:45 +08:00
binfmt_misc.c treewide: Add SPDX license identifier for more missed files 2019-05-21 10:50:45 +02:00
binfmt_script.c treewide: Add SPDX license identifier for more missed files 2019-05-21 10:50:45 +02:00
block_dev.c block: Don't revalidate bdev of hidden gendisk 2019-05-27 07:35:02 -06:00
buffer.c treewide: Add SPDX license identifier for missed files 2019-05-21 10:50:45 +02:00
char_dev.c chardev: update comment based on the code 2019-04-02 17:49:58 +02:00
compat_binfmt_elf.c treewide: Replace GPLv2 boilerplate/reference with SPDX - rule 193 2019-05-30 11:29:21 -07:00
compat_ioctl.c
compat.c treewide: Replace GPLv2 boilerplate/reference with SPDX - rule 500 2019-06-19 17:09:55 +02:00
coredump.c
d_path.c
dax.c dax fix v5.2-rc8 2019-07-05 11:32:11 +09:00
dcache.c treewide: Add SPDX license identifier for missed files 2019-05-21 10:50:45 +02:00
dcookies.c treewide: Add SPDX license identifier for missed files 2019-05-21 10:50:45 +02:00
direct-io.c treewide: Add SPDX license identifier for missed files 2019-05-21 10:50:45 +02:00
drop_caches.c fs/drop_caches.c: avoid softlockups in drop_pagecache_sb() 2019-02-01 15:46:24 -08:00
eventfd.c treewide: Add SPDX license identifier for missed files 2019-05-21 10:50:45 +02:00
eventpoll.c signal: remove the wrong signal_pending() check in restore_user_sigmask() 2019-06-29 16:43:45 +08:00
exec.c treewide: Add SPDX license identifier for missed files 2019-05-21 10:50:45 +02:00
fcntl.c fs: mark expected switch fall-throughs 2019-04-08 18:21:02 -05:00
fhandle.c
file_table.c treewide: Add SPDX license identifier for missed files 2019-05-21 10:50:45 +02:00
file.c io_uring-2019-03-06 2019-03-08 14:48:40 -08:00
filesystems.c vfs: Implement a filesystem superblock creation/configuration context 2019-02-28 03:29:26 -05:00
fs_context.c treewide: Replace GPLv2 boilerplate/reference with SPDX - rule 36 2019-05-24 17:27:11 +02:00
fs_parser.c treewide: Replace GPLv2 boilerplate/reference with SPDX - rule 36 2019-05-24 17:27:11 +02:00
fs_pin.c
fs_struct.c treewide: Add SPDX license identifier for missed files 2019-05-21 10:50:45 +02:00
fs_types.c fs: common implementation of file type 2019-01-21 17:48:13 +01:00
fs-writeback.c treewide: Add SPDX license identifier for missed files 2019-05-21 10:50:45 +02:00
fsopen.c treewide: Replace GPLv2 boilerplate/reference with SPDX - rule 36 2019-05-24 17:27:11 +02:00
inode.c mm: fix page cache convergence regression 2019-05-31 13:52:41 -04:00
internal.h treewide: Replace GPLv2 boilerplate/reference with SPDX - rule 152 2019-05-30 11:26:32 -07:00
io_uring.c Merge branch 'akpm' (patches from Andrew) 2019-06-29 17:11:01 +08:00
ioctl.c Remove 'type' argument from access_ok() function 2019-01-03 18:57:57 -08:00
iomap.c block: return from __bio_try_merge_page if merging occured in the same page 2019-06-17 09:33:02 -06:00
Kconfig fs: VALIDATE_FS_PARSER should default to n 2019-07-05 11:22:11 -04:00
Kconfig.binfmt treewide: Add SPDX license identifier - Makefile/Kconfig 2019-05-21 10:50:46 +02:00
libfs.c treewide: Add SPDX license identifier for missed files 2019-05-21 10:50:45 +02:00
locks.c treewide: Add SPDX license identifier for missed files 2019-05-21 10:50:45 +02:00
Makefile Add as a feature case-insensitive directories (the casefold feature) 2019-05-07 21:12:44 -07:00
mbcache.c treewide: Add SPDX license identifier for more missed files 2019-05-21 10:50:45 +02:00
mount.h saner handling of temporary namespaces 2019-01-30 17:44:07 -05:00
mpage.c block: remove the i argument to bio_for_each_segment_all 2019-04-30 09:26:13 -06:00
namei.c Clean up fscrypt's dcache revalidation support, and other 2019-05-07 21:28:04 -07:00
namespace.c vfs: move_mount: reject moving kernel internal mounts 2019-07-01 10:46:36 -04:00
no-block.c treewide: Replace GPLv2 boilerplate/reference with SPDX - rule 152 2019-05-30 11:26:32 -07:00
nsfs.c nsfs: unobfuscate 2019-04-09 19:20:57 -04:00
open.c access: avoid the RCU grace period for the temporary subjective credentials 2019-07-24 10:12:09 -07:00
pipe.c Merge branch 'page-refs' (page ref overflow) 2019-04-14 15:09:40 -07:00
pnode.c fs/namespace: fix unprivileged mount propagation 2019-06-17 17:36:09 -04:00
pnode.h treewide: Replace GPLv2 boilerplate/reference with SPDX - rule 209 2019-05-30 11:29:53 -07:00
posix_acl.c treewide: Add SPDX license identifier for missed files 2019-05-21 10:50:45 +02:00
proc_namespace.c
read_write.c vfs: pass ppos=NULL to .read()/.write() of FMODE_STREAM files 2019-05-06 17:46:52 +03:00
readdir.c Remove 'type' argument from access_ok() function 2019-01-03 18:57:57 -08:00
select.c signal: remove the wrong signal_pending() check in restore_user_sigmask() 2019-06-29 16:43:45 +08:00
seq_file.c fs: mark expected switch fall-throughs 2019-04-08 18:21:02 -05:00
signalfd.c fs: mark expected switch fall-throughs 2019-04-08 18:21:02 -05:00
splice.c treewide: Add SPDX license identifier for missed files 2019-05-21 10:50:45 +02:00
stack.c treewide: Add SPDX license identifier for missed files 2019-05-21 10:50:45 +02:00
stat.c fs: move generic stat response attr handling to vfs_getattr_nosec 2019-02-01 01:55:45 -05:00
statfs.c vfs: add vfs_get_fsid() helper 2019-02-07 16:38:35 +01:00
super.c [fix] get rid of checking for absent device name in vfs_get_tree() 2019-04-28 21:34:21 -04:00
sync.c fs/sync.c: sync_file_range(2) may use WB_SYNC_ALL writeback 2019-05-14 09:47:50 -07:00
timerfd.c y2038: syscalls: rename y2038 compat syscalls 2019-02-07 00:13:27 +01:00
userfaultfd.c fs/userfaultfd.c: disable irqs for fault_pending and event locks 2019-07-05 11:12:07 +09:00
utimes.c y2038: syscalls: rename y2038 compat syscalls 2019-02-07 00:13:27 +01:00
xattr.c treewide: Add SPDX license identifier for missed files 2019-05-21 10:50:45 +02:00