aha/fs
Jeff Moyer 23aee091d8 dio: don't zero out the pages array inside struct dio
Intel reported a performance regression caused by the following commit:

commit 848c4dd515
Author: Zach Brown <zach.brown@oracle.com>
Date:   Mon Aug 20 17:12:01 2007 -0700

    dio: zero struct dio with kzalloc instead of manually

    This patch uses kzalloc to zero all of struct dio rather than
    manually trying to track which fields we rely on being zero.  It
    passed aio+dio stress testing and some bug regression testing on
    ext3.

    This patch was introduced by Linus in the conversation that lead up
    to Badari's minimal fix to manually zero .map_bh.b_state in commit:

      6a648fa721

    It makes the code a bit smaller.  Maybe a couple fewer cachelines to
    load, if we're lucky:

       text    data     bss     dec     hex filename
    3285925  568506 1304616 5159047  4eb887 vmlinux
    3285797  568506 1304616 5158919  4eb807 vmlinux.patched

    I was unable to measure a stable difference in the number of cpu
    cycles spent in blockdev_direct_IO() when pushing aio+dio 256K reads
    at ~340MB/s.

    So the resulting intent of the patch isn't a performance gain but to
    avoid exposing ourselves to the risk of finding another field like
    .map_bh.b_state where we rely on zeroing but don't enforce it in the
    code.

Zach surmised that zeroing out the page array was what caused most of
the problem, and suggested the approach taken in the attached patch for
resolving the issue.  Intel re-tested with this patch and saw a 0.6%
performance gain (the original regression was 0.5%).

[akpm@linux-foundation.org: add comment]
Signed-off-by: Jeff Moyer <jmoyer@redhat.com>
Acked-by: Zach Brown <zach.brown@oracle.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2009-12-16 07:20:13 -08:00
..
9p 9p: fix build breakage introduced by FS-Cache 2009-12-01 07:35:11 -08:00
adfs adfs: remove redundant test on unsigned 2009-09-24 07:21:05 -07:00
affs
afs afs: remove manual O_SYNC handling 2009-12-10 15:02:50 +01:00
autofs trivial: remove unnecessary semicolons 2009-09-21 15:14:58 +02:00
autofs4 autofs4: always use lookup for lookup 2009-12-16 07:19:58 -08:00
befs fs: Make unload_nls() NULL pointer safe 2009-09-24 07:47:42 -04:00
bfs headers: smp_lock.h redux 2009-07-12 12:22:34 -07:00
btrfs task_struct: make journal_info conditional 2009-12-15 08:53:27 -08:00
cachefiles tree-wide: convert open calls to remove spaces to skip_spaces() lib function 2009-12-15 08:53:32 -08:00
cifs vfs: Implement proper O_SYNC semantics 2009-12-10 15:02:50 +01:00
coda sysctl: Drop & in front of every proc_handler. 2009-11-18 08:37:40 -08:00
configfs writeback: add name to backing_dev_info 2009-09-11 09:20:26 +02:00
cramfs
debugfs debugfs: fix create mutex racy fops and private data 2009-12-11 11:24:53 -08:00
devpts devpts_get_tty() should validate inode 2009-12-11 15:18:05 -08:00
dlm Merge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/teigland/dlm 2009-12-10 09:33:59 -08:00
ecryptfs ima: ecryptfs fix imbalance message 2009-10-08 11:31:38 -05:00
efs get rid of BKL in fs/efs 2009-06-17 00:36:36 -04:00
exofs exofs: Multi-device mirror support 2009-12-10 09:59:23 +02:00
exportfs
ext2 ext2: report metadata errors during fsync 2009-12-16 07:20:06 -08:00
ext3 ext3: PTR_ERR return of wrong pointer in setup_new_group_blocks() 2009-12-10 15:02:55 +01:00
ext4 tree-wide: convert open calls to remove spaces to skip_spaces() lib function 2009-12-15 08:53:32 -08:00
fat fatfs: use common time_to_tm in fat_time_unix2fat() 2009-12-16 07:20:06 -08:00
freevxfs headers: smp_lock.h redux 2009-07-12 12:22:34 -07:00
fscache FS-Cache: Avoid maybe-used-uninitialised warning on variable 2009-12-16 07:20:13 -08:00
fuse fuse: reject O_DIRECT flag also in fuse_create 2009-11-27 16:37:13 +01:00
gfs2 fs/gfs2/sys.c: use %pUB to print UUIDs 2009-12-15 08:53:33 -08:00
hfs hfs: fix a potential buffer overflow 2009-12-15 08:53:10 -08:00
hfsplus hfsplus: refuse to mount volumes larger than 2TB 2009-10-29 07:39:27 -07:00
hostfs hostfs: set maximum filesize in superblock for proper LFS support 2009-06-30 18:56:03 -07:00
hpfs hpfs: use bitmap_weight() 2009-12-16 07:20:06 -08:00
hppfs
hugetlbfs Merge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/viro/vfs-2.6 2009-09-24 08:32:11 -07:00
isofs zisofs: Implement reading of compressed files when PAGE_CACHE_SIZE > compress block size 2009-12-10 15:02:49 +01:00
jbd task_struct: make journal_info conditional 2009-12-15 08:53:27 -08:00
jbd2 task_struct: make journal_info conditional 2009-12-15 08:53:27 -08:00
jffs2 Merge branch 'for-next' into for-linus 2009-12-07 18:36:35 +01:00
jfs tree-wide: fix assorted typos all over the place 2009-12-04 15:39:55 +01:00
lockd sysctl: Drop & in front of every proc_handler. 2009-11-18 08:37:40 -08:00
minix V3 minixfs: add missing directory type checking 2009-09-23 07:39:57 -07:00
ncpfs tree-wide: fix assorted typos all over the place 2009-12-04 15:39:55 +01:00
nfs Merge git://git.linux-nfs.org/projects/trondmy/nfs-2.6 2009-12-14 10:00:24 -08:00
nfs_common
nfsd Fix memory corruption caused by nfsd readdir+ 2009-11-14 12:55:55 -08:00
nilfs2 task_struct: make journal_info conditional 2009-12-15 08:53:27 -08:00
nls Merge git://git.kernel.org/pub/scm/linux/kernel/git/hirofumi/fatfs-2.6 2009-09-30 09:31:14 -07:00
notify Merge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/jikos/trivial 2009-12-09 19:43:33 -08:00
ntfs Merge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/jikos/trivial 2009-12-09 19:43:33 -08:00
ocfs2 quota: Move definition of QFMT_OCFS2 to linux/quota.h 2009-12-10 15:02:53 +01:00
omfs tree-wide: fix assorted typos all over the place 2009-12-04 15:39:55 +01:00
openpromfs
partitions partitions: read whole sector with EFI GPT header 2009-11-23 09:29:58 +01:00
proc elf: kill USE_ELF_CORE_DUMP 2009-12-16 07:20:12 -08:00
qnx4 qnx4fs: add missing KERN_xxx to printk() calls 2009-11-09 09:40:57 +01:00
quota quota: Implement quota format with 64-bit space and inode limits 2009-12-10 15:02:54 +01:00
ramfs truncate: use new helpers 2009-09-24 08:41:47 -04:00
reiserfs reiserfs: don't compile procfs.o at all if no support 2009-12-16 07:20:06 -08:00
romfs ROMFS: fix length used with romfs_dev_strnlen() function 2009-10-11 11:33:56 -07:00
smbfs fs: Make unload_nls() NULL pointer safe 2009-09-24 07:47:42 -04:00
squashfs const: mark remaining super_operations const 2009-09-22 07:17:24 -07:00
sysfs sysfs: sysfs_setattr remove unnecessary permission check. 2009-12-11 11:24:54 -08:00
sysv get rid of BKL in fs/sysv 2009-06-17 00:36:37 -04:00
ubifs fs/ubifs: use %pUB to print UUIDs 2009-12-15 08:53:33 -08:00
udf udf: Avoid IO in udf_clear_inode 2009-12-14 21:40:04 +01:00
ufs ufs: NFS support 2009-12-16 07:20:06 -08:00
xfs Merge branch 'for-linus' of git://oss.sgi.com/xfs/xfs 2009-12-15 09:12:43 -08:00
aio.c aio: remove unused field 2009-12-16 07:20:13 -08:00
anon_inodes.c headers: remove sched.h from poll.h 2009-10-04 15:05:10 -07:00
attr.c truncate: new helpers 2009-09-24 08:41:47 -04:00
bad_inode.c
binfmt_aout.c
binfmt_elf.c elf: kill USE_ELF_CORE_DUMP 2009-12-16 07:20:12 -08:00
binfmt_elf_fdpic.c elf: kill USE_ELF_CORE_DUMP 2009-12-16 07:20:12 -08:00
binfmt_em86.c
binfmt_flat.c flat: use IS_ERR_VALUE() helper macro 2009-09-24 07:21:03 -07:00
binfmt_misc.c
binfmt_script.c
binfmt_som.c
bio-integrity.c block: Create bip slabs with embedded integrity vectors 2009-07-01 10:56:25 +02:00
bio.c Merge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/jikos/trivial 2009-12-09 19:43:33 -08:00
block_dev.c Merge branch 'for-linus' into for-2.6.33 2009-11-03 21:14:39 +01:00
buffer.c Merge branch 'writeback' of git://git.kernel.dk/linux-2.6-block 2009-09-25 09:27:30 -07:00
char_dev.c fs/char_dev.c: remove useless loop 2009-09-24 07:21:03 -07:00
compat.c x86, fs: Fix x86 procfs stack information for threads on 64-bit 2009-11-04 13:25:03 +01:00
compat_binfmt_elf.c
compat_ioctl.c md: move compat_ioctl handling into md.c 2009-12-14 12:51:41 +11:00
dcache.c sched: Pull up the might_sleep() check into cond_resched() 2009-07-18 15:51:44 +02:00
dcookies.c
direct-io.c dio: don't zero out the pages array inside struct dio 2009-12-16 07:20:13 -08:00
drop_caches.c sysctl: remove "struct file *" argument of ->proc_handler 2009-09-24 07:21:04 -07:00
eventfd.c anonfd: split interface into file creation and install 2009-09-23 07:39:29 -07:00
eventpoll.c sysctl: Drop & in front of every proc_handler. 2009-11-18 08:37:40 -08:00
exec.c procfs: allow threads to rename siblings via /proc/pid/tasks/tid/comm 2009-12-15 08:53:24 -08:00
fcntl.c fcntl: rename F_OWNER_GID to F_OWNER_PGRP 2009-11-17 17:40:33 -08:00
fifo.c
file.c headers: remove sched.h from interrupt.h 2009-10-11 11:20:58 -07:00
file_table.c LSM: imbed ima calls in the security hooks 2009-10-25 12:22:48 +08:00
filesystems.c
fs-writeback.c writeback: remove unused nonblocking and congestion checks 2009-12-03 13:54:25 +01:00
fs_struct.c
generic_acl.c
inode.c LSM: imbed ima calls in the security hooks 2009-10-25 12:22:48 +08:00
internal.h fs: fix overflow in sys_mount() for in-kernel calls 2009-09-24 08:40:15 -04:00
ioctl.c __generic_block_fiemap(): fix for files bigger than 4GB 2009-11-12 07:26:01 -08:00
ioprio.c
Kconfig task_struct: make journal_info conditional 2009-12-15 08:53:27 -08:00
Kconfig.binfmt
libfs.c libfs: return error code on failed attr set 2009-09-24 07:47:30 -04:00
locks.c const: make lock_manager_operations const 2009-09-22 07:17:25 -07:00
Makefile
mbcache.c
mpage.c
namei.c Merge branch 'for_linus' of git://git.kernel.org/pub/scm/linux/kernel/git/jack/linux-fs-2.6 2009-12-11 15:31:13 -08:00
namespace.c LSM: Pass original mount flags to security_sb_mount(). 2009-10-12 10:56:03 +11:00
nfsctl.c
no-block.c
open.c LSM: Move security_path_chmod()/security_path_chown() to after mutex_lock(). 2009-11-24 08:49:26 +11:00
pipe.c fs: pipe.c null pointer dereference 2009-10-22 08:11:44 +09:00
pnode.c
pnode.h
posix_acl.c
read_write.c sendfile(): check f_op.splice_write() rather than f_op.sendpage() 2009-11-04 09:09:52 +01:00
read_write.h
readdir.c
select.c headers: remove sched.h from poll.h 2009-10-04 15:05:10 -07:00
seq_file.c vfs: seq_file: add helpers for data filling 2009-09-24 07:47:35 -04:00
signalfd.c
splice.c sendfile(): check f_op.splice_write() rather than f_op.sendpage() 2009-11-04 09:09:52 +01:00
stack.c
stat.c
super.c freeze_bdev: grab active reference to frozen superblocks 2009-09-24 07:47:41 -04:00
sync.c kill wait_on_page_writeback_range 2009-12-10 15:02:50 +01:00
timerfd.c
utimes.c
xattr.c VFS: Factor out part of vfs_setxattr so it can be called from the SELinux hook for inode_setsecctx. 2009-09-10 10:11:22 +10:00
xattr_acl.c VFS: Use GFP_NOFS in posix_acl_from_xattr() 2009-12-03 11:48:07 +00:00