linux/fs
Xu Yang 4e527d5841 iomap: fault in smaller chunks for non-large folio mappings
Since commit (5d8edfb900 "iomap: Copy larger chunks from userspace"),
iomap will try to copy in larger chunks than PAGE_SIZE. However, if the
mapping doesn't support large folio, only one page of maximum 4KB will
be created and 4KB data will be writen to pagecache each time. Then,
next 4KB will be handled in next iteration. This will cause potential
write performance problem.

If chunk is 2MB, total 512 pages need to be handled finally. During this
period, fault_in_iov_iter_readable() is called to check iov_iter readable
validity. Since only 4KB will be handled each time, below address space
will be checked over and over again:

start         	end
-
buf,    	buf+2MB
buf+4KB, 	buf+2MB
buf+8KB, 	buf+2MB
...
buf+2044KB 	buf+2MB

Obviously the checking size is wrong since only 4KB will be handled each
time. So this will get a correct chunk to let iomap work well in non-large
folio case.

With this change, the write speed will be stable. Tested on ARM64 device.

Before:

 - dd if=/dev/zero of=/dev/sda bs=400K  count=10485  (334 MB/s)
 - dd if=/dev/zero of=/dev/sda bs=800K  count=5242   (278 MB/s)
 - dd if=/dev/zero of=/dev/sda bs=1600K count=2621   (204 MB/s)
 - dd if=/dev/zero of=/dev/sda bs=2200K count=1906   (170 MB/s)
 - dd if=/dev/zero of=/dev/sda bs=3000K count=1398   (150 MB/s)
 - dd if=/dev/zero of=/dev/sda bs=4500K count=932    (139 MB/s)

After:

 - dd if=/dev/zero of=/dev/sda bs=400K  count=10485  (339 MB/s)
 - dd if=/dev/zero of=/dev/sda bs=800K  count=5242   (330 MB/s)
 - dd if=/dev/zero of=/dev/sda bs=1600K count=2621   (332 MB/s)
 - dd if=/dev/zero of=/dev/sda bs=2200K count=1906   (333 MB/s)
 - dd if=/dev/zero of=/dev/sda bs=3000K count=1398   (333 MB/s)
 - dd if=/dev/zero of=/dev/sda bs=4500K count=932    (333 MB/s)

Fixes: 5d8edfb900 ("iomap: Copy larger chunks from userspace")
Cc: stable@vger.kernel.org
Reviewed-by: Darrick J. Wong <djwong@kernel.org>
Signed-off-by: Xu Yang <xu.yang_2@nxp.com>
Link: https://lore.kernel.org/r/20240521114939.2541461-2-xu.yang_2@nxp.com
Reviewed-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Matthew Wilcox (Oracle) <willy@infradead.org>
Signed-off-by: Christian Brauner <brauner@kernel.org>
2024-05-24 13:34:07 +02:00
..
9p netfs: Remove the old writeback code 2024-05-01 18:07:38 +01:00
adfs mm, slab: remove last vestiges of SLAB_MEM_SPREAD 2024-03-12 20:32:19 -07:00
affs affs: remove SLAB_MEM_SPREAD flag usage 2024-02-26 11:36:28 +01:00
afs vfs-6.10.netfs 2024-05-13 12:14:03 -07:00
autofs dcache stuff for this cycle 2024-01-11 20:11:35 -08:00
bcachefs bcachefs changes for 6.10-rc1 2024-05-19 13:45:48 -07:00
befs mm, slab: remove last vestiges of SLAB_MEM_SPREAD 2024-03-12 20:32:19 -07:00
bfs mm, slab: remove last vestiges of SLAB_MEM_SPREAD 2024-03-12 20:32:19 -07:00
btrfs Networking changes for 6.10. 2024-05-14 19:42:24 -07:00
cachefiles netfs, cachefiles: Implement helpers for new write code 2024-05-01 18:07:37 +01:00
ceph netfs: Switch to using unsigned long long rather than loff_t 2024-05-01 18:07:35 +01:00
coda mm, slab: remove last vestiges of SLAB_MEM_SPREAD 2024-03-12 20:32:19 -07:00
configfs
cramfs fs,block: yield devices early 2024-03-27 13:17:15 +01:00
crypto The usual shower of singleton fixes and minor series all over MM, 2024-05-19 09:21:03 -07:00
debugfs vfs: Convert debugfs to use the new mount API 2024-03-26 09:04:54 +01:00
devpts fs: Remove the now superfluous sentinel elements from ctl_table array 2023-12-28 04:57:57 -08:00
dlm dlm: return -ENOMEM if ls_recover_buf fails 2024-04-23 16:08:55 -05:00
ecryptfs hardening updates for 6.10-rc1 2024-05-13 14:14:05 -07:00
efivarfs efi: Clear up misconceptions about a maximum variable name size 2024-04-13 10:33:02 +02:00
efs efs: remove SLAB_MEM_SPREAD flag usage 2024-02-27 11:21:33 +01:00
erofs erofs: Zstandard compression support 2024-05-09 07:46:56 +08:00
exfat exfat: zero the reserved fields of file and stream extension dentries 2024-04-25 21:59:59 +09:00
exportfs fs: Create a generic is_dot_dotdot() utility 2024-01-23 10:58:56 -05:00
ext2 ext2: Remove LEGACY_DIRECT_IO dependency 2024-05-03 11:50:28 +02:00
ext4 Ext4 patches for the 6.10-rc1 merge window: 2024-05-18 14:11:54 -07:00
f2fs f2fs update for 6.10-rc1 2024-05-20 13:23:43 -07:00
fat fs: add kernel-doc comments to fat_parse_long() 2024-04-25 21:07:02 -07:00
freevxfs freevxfs: Convert freevxfs to the new mount API. 2024-03-26 09:04:53 +01:00
fuse virtiofs: include a newline in sysfs tag 2024-05-08 09:31:21 +02:00
gfs2 gfs2 updates 2024-05-14 17:35:22 -07:00
hfs hfs: really remove hfs_writepage 2023-12-29 11:58:34 -08:00
hfsplus hfsplus: refactor copy_name to not use strncpy 2024-04-24 16:55:28 -07:00
hostfs hostfs: use d_splice_alias() calling conventions to simplify failure exits 2023-12-21 12:51:00 -05:00
hpfs mm, slab: remove last vestiges of SLAB_MEM_SPREAD 2024-03-12 20:32:19 -07:00
hugetlbfs The usual shower of singleton fixes and minor series all over MM, 2024-05-19 09:21:03 -07:00
iomap iomap: fault in smaller chunks for non-large folio mappings 2024-05-24 13:34:07 +02:00
isofs isofs: Use *-y instead of *-objs in Makefile 2024-05-09 18:09:57 +02:00
jbd2 jbd2: add prefix 'jbd2' for 'shrink_type' 2024-05-09 10:52:31 -04:00
jffs2 jffs2: prevent xattr node from overflowing the eraseblock 2024-04-17 13:21:34 +02:00
jfs fs,block: yield devices early 2024-03-27 13:17:15 +01:00
kernfs kernfs: annotate different lockdep class for of->mutex of writable files 2024-04-14 06:55:46 -04:00
lockd lockd: host: Remove unnecessary statements'host = NULL;' 2024-05-06 09:07:20 -04:00
minix minix: convert minix to use the new mount api 2024-03-26 09:04:55 +01:00
netfs netfs: Fix AIO error handling when doing write-through 2024-05-24 13:34:06 +02:00
nfs The usual shower of singleton fixes and minor series all over MM, 2024-05-19 09:21:03 -07:00
nfs_common
nfsd \n 2024-05-20 12:31:43 -07:00
nilfs2 Mainly singleton patches, documented in their respective changelogs. 2024-05-19 14:02:03 -07:00
nls
notify Revert "fanotify: remove unneeded sub-zero check for unsigned value" 2024-05-20 12:43:58 -07:00
ntfs3 Merge git://git.kernel.org/pub/scm/linux/kernel/git/netdev/net 2024-05-02 12:06:25 -07:00
ocfs2 Mainly singleton patches, documented in their respective changelogs. 2024-05-19 14:02:03 -07:00
omfs
openpromfs openpromfs: finish conversion to the new mount API 2024-03-26 09:04:54 +01:00
orangefs orangefs: fix out-of-bounds fsid access 2024-05-14 17:44:14 -07:00
overlayfs integrity-v6.10 2024-05-15 08:43:02 -07:00
proc Mainly singleton patches, documented in their respective changelogs. 2024-05-19 14:02:03 -07:00
pstore pstore/zone: Don't clear memory twice 2024-03-09 12:33:22 -08:00
qnx4 mm, slab: remove last vestiges of SLAB_MEM_SPREAD 2024-03-12 20:32:19 -07:00
qnx6 qnx6: convert qnx6 to use the new mount api 2024-03-26 09:04:53 +01:00
quota quota: fix to propagate error of mark_dquot_dirty() to caller 2024-04-12 14:52:29 +02:00
ramfs mm: switch mm->get_unmapped_area() to a flag 2024-04-25 20:56:25 -07:00
reiserfs \n 2024-05-20 12:49:25 -07:00
romfs fs,block: yield devices early 2024-03-27 13:17:15 +01:00
smb cifs: fix data corruption in read after invalidate 2024-05-15 17:22:59 -05:00
squashfs Mainly singleton patches, documented in their respective changelogs. 2024-05-19 14:02:03 -07:00
sysfs fs: sysfs: Fix reference leak in sysfs_break_active_protection() 2024-04-11 15:16:48 +02:00
sysv sysv: remove SLAB_MEM_SPREAD flag usage 2024-02-27 11:21:31 +01:00
tracefs tracing cleanups for v6.10: 2024-05-17 18:34:27 -07:00
ubifs This pull request contains updates for UBI and UBIFS: 2024-03-21 15:09:29 -07:00
udf udf: Use a folio in udf_write_end() 2024-04-23 15:37:02 +02:00
ufs mm, slab: remove last vestiges of SLAB_MEM_SPREAD 2024-03-12 20:32:19 -07:00
unicode kbuild: use $(src) instead of $(srctree)/$(src) for source directory 2024-05-10 04:34:52 +09:00
vboxsf vboxsf: explicitly deny setlease attempts 2024-04-03 16:06:39 +02:00
verity fsverity: use register_sysctl_init() to avoid kmemleak warning 2024-05-03 08:30:58 -07:00
xfs New code for 6.10: 2024-05-20 12:55:12 -07:00
zonefs zonefs: Use str_plural() to fix Coccinelle warning 2024-04-10 07:23:47 +09:00
aio.c vfs-6.10.misc 2024-05-13 11:40:06 -07:00
anon_inodes.c fs: Create anon_inode_getfile_fmode() 2024-04-26 10:33:05 +02:00
attr.c lsm/stable-6.9 PR 20240312 2024-03-12 20:03:34 -07:00
backing-file.c fs: Use KMEM_CACHE instead of kmem_cache_create 2024-02-02 13:11:50 +01:00
bad_inode.c
binfmt_elf_fdpic.c binfmt_elf_fdpic: fix /proc/<pid>/auxv 2024-04-24 15:55:28 -07:00
binfmt_elf_test.c
binfmt_elf.c Mainly singleton patches, documented in their respective changelogs. 2024-05-19 14:02:03 -07:00
binfmt_flat.c
binfmt_misc.c execve updates for v6.7-rc1 2023-10-30 19:28:19 -10:00
binfmt_script.c
buffer.c buffer: improve bdev_getblk documentation 2024-05-05 17:53:40 -07:00
char_dev.c As usual, lots of singleton and doubleton patches all over the tree and 2023-11-02 20:53:31 -10:00
compat_binfmt_elf.c
coredump.c fs/coredump: Enable dynamic configuration of max file note size 2024-05-08 09:53:00 -07:00
d_path.c
dax.c dax: use huge_zero_folio 2024-04-25 20:56:20 -07:00
dcache.c fs/dcache: Re-use value stored to dentry->d_flags instead of re-reading 2024-04-09 12:29:03 +02:00
direct-io.c fs/direct-io: remove redundant assignment to variable retval 2024-04-11 10:21:24 +02:00
drop_caches.c
eventfd.c eventfd: strictly check the count parameter of eventfd_write to avoid inputting illegal strings 2024-02-08 10:12:26 +01:00
eventpoll.c epoll: be better about file lifetimes 2024-05-05 14:00:48 -07:00
exec.c The usual shower of singleton fixes and minor series all over MM, 2024-05-19 09:21:03 -07:00
fcntl.c fcntl: add F_DUPFD_QUERY fcntl() 2024-05-10 08:26:31 +02:00
fhandle.c fs: Annotate struct file_handle with __counted_by() and use struct_size() 2024-04-05 15:53:47 +02:00
file_table.c lsm/stable-6.9 PR 20240312 2024-03-12 20:03:34 -07:00
file.c file: remove __receive_fd() 2023-12-12 14:24:14 +01:00
filesystems.c
fs_context.c
fs_parser.c __fs_parse: Correct a documentation comment 2024-02-02 13:11:50 +01:00
fs_pin.c
fs_struct.c
fs_types.c
fs-writeback.c fs/writeback: remove unnecessary return in writeback_inodes_sb 2024-04-05 15:53:45 +02:00
fsopen.c
init.c
inode.c bcachefs updates for 6.9 2024-03-15 09:00:09 -07:00
internal.h pidfs: remove config option 2024-03-13 12:53:53 -07:00
ioctl.c fs/ioctl: Add a comment to keep the logic in sync with LSM policies 2024-05-13 06:58:35 +02:00
Kconfig - Sumanth Korikkar has taught s390 to allocate hotplug-time page frames 2024-03-14 17:43:30 -07:00
Kconfig.binfmt
kernel_read_file.c
libfs.c shmem: Fix shmem_rename2() 2024-04-17 13:49:44 +02:00
locks.c filelock: fix deadlock detection in POSIX locking 2024-02-20 09:53:33 +01:00
Makefile vfs-6.9.pidfd 2024-03-11 10:21:06 -07:00
mbcache.c vfs: remove SLAB_MEM_SPREAD flag usage 2024-02-27 11:21:31 +01:00
mnt_idmapping.c fs/mnt_idmapping.c: Return -EINVAL when no map is written 2024-02-08 10:12:37 +01:00
mount.h mounts: keep list of mounts in an rbtree 2023-11-18 14:56:16 +01:00
mpage.c block, fs: Restore the per-bio/request data lifetime fields 2024-02-06 14:31:05 +01:00
namei.c vfs-6.10.misc 2024-05-13 11:40:06 -07:00
namespace.c fs: relax mount_setattr() permission checks 2024-02-07 21:16:29 +01:00
nsfs.c pidfs: remove config option 2024-03-13 12:53:53 -07:00
open.c lsm/stable-6.9 PR 20240312 2024-03-12 20:03:34 -07:00
pidfs.c pidfs: remove config option 2024-03-13 12:53:53 -07:00
pipe.c fs/pipe: Convert to lockdep_cmp_fn 2024-02-02 13:11:49 +01:00
pnode.c mounts: keep list of mounts in an rbtree 2023-11-18 14:56:16 +01:00
pnode.h
posix_acl.c lsm/stable-6.9 PR 20240312 2024-03-12 20:03:34 -07:00
proc_namespace.c namespace: extract show_path() helper 2023-11-18 14:56:16 +01:00
read_write.c New code for 6.10: 2024-05-20 12:55:12 -07:00
readdir.c fsnotify: optionally pass access range in file permission hooks 2023-12-12 16:20:02 +01:00
remap_range.c vfs: export remap and write check helpers 2024-04-15 14:54:13 -07:00
select.c fs/select: rework stack allocation hack for clang 2024-02-20 09:23:52 +01:00
seq_file.c seq_file: Simplify __seq_puts() 2024-05-02 16:28:20 +02:00
signalfd.c signalfd: convert to ->read_iter() 2024-04-10 16:23:04 -06:00
splice.c fs: use splice_copy_file_range() inline helper 2023-12-12 16:20:02 +01:00
stack.c
stat.c statx: stx_subvol 2024-03-26 09:01:18 +01:00
statfs.c
super.c \n 2024-05-20 12:31:43 -07:00
sync.c
sysctls.c fs: Remove the now superfluous sentinel elements from ctl_table array 2023-12-28 04:57:57 -08:00
timerfd.c timerfd: convert to ->read_iter() 2024-04-10 16:23:02 -06:00
userfaultfd.c The usual shower of singleton fixes and minor series all over MM, 2024-05-19 09:21:03 -07:00
utimes.c
xattr.c evm: Move to LSM infrastructure 2024-02-15 23:43:47 -05:00