Rusty Russell [Sat, 13 Jun 2009 04:27:04 +0000 (22:27 -0600)]
lguest: clean up length-used value in example launcher
The "len" field in the used ring for virtio indicates the number of
bytes *written* to the buffer. This means the guest doesn't have to
zero the buffers in advance as it always knows the used length.
Erroneously, the console and network example code puts the length
*read* into that field. The guest ignores it, but it's wrong.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
Rusty Russell [Sat, 13 Jun 2009 04:27:03 +0000 (22:27 -0600)]
lguest: optimize by coding restore_flags and irq_enable in assembler.
The downside of the last patch which made restore_flags and irq_enable
check interrupts is that they are now too big to be patched directly
into the callsites, so the C versions are always used.
But the C versions go via PV_CALLEE_SAVE_REGS_THUNK which saves all
the registers. In fact, we don't need any registers in the fast path,
so we can do better than this if we actually code them in assembler.
The results are in the noise, but since it's about the same amount of
code, it's worth applying.
Rusty Russell [Sat, 13 Jun 2009 04:27:02 +0000 (22:27 -0600)]
lguest: improve interrupt handling, speed up stream networking
lguest never checked for pending interrupts when enabling interrupts, and
things still worked. However, it makes a significant difference to TCP
performance, so it's time we fixed it by introducing a pending_irq flag
and checking it on irq_restore and irq_enable.
These two routines are now too big to patch into the 8/10 bytes
patch space, so we drop that code.
Note: The high latency on interrupt delivery had a very curious
effect: once everything else was optimized, networking without GSO was
faster than networking with GSO, since more interrupts were sent and
hence a greater chance of one getting through to the Guest!
Note2: (Almost) Closing the same loophole for iret doesn't have any
measurable effect, so I'm leaving that patch for the moment.
Rusty Russell [Sat, 13 Jun 2009 04:27:02 +0000 (22:27 -0600)]
lguest: fix race in halt code
When the Guest does the LHCALL_HALT hypercall, we go to sleep, expecting
that a timer or the Waker will wake_up_process() us.
But we do it in a stupid way, leaving a classic missing wakeup race.
So split maybe_do_interrupt() into interrupt_pending() and
try_deliver_interrupt(), and check maybe_do_interrupt() and the
"break_out" flag before calling schedule.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
Rusty Russell [Sat, 13 Jun 2009 04:27:01 +0000 (22:27 -0600)]
lguest: remove invalid interrupt forcing logic.
20887611523e749d99cc7d64ff6c97d27529fbae (lguest: notify on empty) introduced
lguest support for the VIRTIO_F_NOTIFY_ON_EMPTY flag, but in fact it turned on
interrupts all the time.
Because we always process one buffer at a time, the inflight count is always 0
when call trigger_irq and so we always ignore VRING_AVAIL_F_NO_INTERRUPT from
the Guest.
It should be looking to see if there are more buffers in the Guest's queue:
if it's empty, then we force an interrupt.
This makes little difference, since we usually have an empty queue; but
that's the subject of another patch.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
Rusty Russell [Sat, 13 Jun 2009 04:27:01 +0000 (22:27 -0600)]
lguest: fix lguest wake on guest clock tick, or fd activity
The Launcher could be inside the Guest on another CPU; wake_up_process
will do nothing because it is "running". kick_process will knock it
back into our kernel in this case, otherwise we'll miss it until the
next guest exit.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
Rusty Russell [Sat, 13 Jun 2009 04:26:58 +0000 (22:26 -0600)]
lguest: be paranoid about guest playing with device descriptors.
We can't trust the values in the device descriptor table once the
guest has booted, so keep local copies. They could set them to
strange values then cause us to segv (they're 8 bit values, so they
can't make our pointers go too wild).
This becomes more important with the following patches which read them.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
Linus Torvalds [Fri, 12 Jun 2009 03:05:37 +0000 (20:05 -0700)]
Merge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/viro/vfs-2.6
* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/viro/vfs-2.6: (87 commits)
nilfs2: get rid of bd_mount_sem use from nilfs
nilfs2: correct exclusion control in nilfs_remount function
nilfs2: simplify remaining sget() use
nilfs2: get rid of sget use for checking if current mount is present
nilfs2: get rid of sget use for acquiring nilfs object
nilfs2: remove meaningless EBUSY case from nilfs_get_sb function
remove the call to ->write_super in __sync_filesystem
nilfs2: call nilfs2_write_super from nilfs2_sync_fs
jffs2: call jffs2_write_super from jffs2_sync_fs
ufs: add ->sync_fs
sysv: add ->sync_fs
hfsplus: add ->sync_fs
hfs: add ->sync_fs
fat: add ->sync_fs
ext2: add ->sync_fs
exofs: add ->sync_fs
bfs: add ->sync_fs
affs: add ->sync_fs
sanitize ->fsync() for affs
repair bfs_write_inode(), switch bfs to simple_fsync()
...
Linus Torvalds [Fri, 12 Jun 2009 03:05:08 +0000 (20:05 -0700)]
Merge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/gerg/m68knommu
* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/gerg/m68knommu:
m68knommu: remove unecessary include of thread_info.h in entry.S
m68knommu: enumerate INIT_THREAD fields properly
headers_check fix: m68k, swab.h
arch/m68knommu: Convert #ifdef DEBUG printk(KERN_DEBUG to pr_debug(
m68knommu: remove obsolete reset code
m68knommu: move CPU reset code for the 5272 ColdFire into its platform code
m68knommu: move CPU reset code for the 528x ColdFire into its platform code
m68knommu: move CPU reset code for the 527x ColdFire into its platform code
m68knommu: move CPU reset code for the 523x ColdFire into its platform code
m68knommu: move CPU reset code for the 520x ColdFire into its platform code
m68knommu: add CPU reset code for the 532x ColdFire
m68knommu: add CPU reset code for the 5249 ColdFire
m68knommu: add CPU reset code for the 5206e ColdFire
m68knommu: add CPU reset code for the 5206 ColdFire
m68knommu: add CPU reset code for the 5407 ColdFire
m68knommu: add CPU reset code for the 5307 ColdFire
m68knommu: merge system reset for code ColdFire 523x family
m68knommu: fix system reset for ColdFire 527x family
Ryusuke Konishi [Sun, 7 Jun 2009 16:39:33 +0000 (01:39 +0900)]
nilfs2: get rid of bd_mount_sem use from nilfs
This will remove every bd_mount_sem use in nilfs.
The intended exclusion control was replaced by the previous patch
("nilfs2: correct exclusion control in nilfs_remount function") for
nilfs_remount(), and this patch will replace remains with a new mutex
that this inserts in nilfs object.
Signed-off-by: Ryusuke Konishi <konishi.ryusuke@lab.ntt.co.jp> Cc: Christoph Hellwig <hch@infradead.org> Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
Ryusuke Konishi [Sun, 7 Jun 2009 16:39:32 +0000 (01:39 +0900)]
nilfs2: correct exclusion control in nilfs_remount function
nilfs_remount() changes mount state of a superblock instance. Even
though nilfs accesses other superblock instances during mount or
remount, the mount state was not properly protected in
nilfs_remount().
Moreover, nilfs_remount() has a lock order reversal problem;
nilfs_get_sb() holds:
1. sb->s_umount (locked by the caller in vfs)
2. bdev->bd_mount_sem
To avoid these problems, this patch divides a semaphore protecting
super block instances from nilfs->ns_sem, and applies it to the mount
state protection in nilfs_remount().
With this change, bd_mount_sem use is removed from nilfs_remount() and
the lock order reversal will be resolved. And the new rw-semaphore,
nilfs->ns_super_sem will properly protect the mount state except the
modification from nilfs_error function.
Signed-off-by: Ryusuke Konishi <konishi.ryusuke@lab.ntt.co.jp> Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
Ryusuke Konishi [Sun, 7 Jun 2009 16:39:31 +0000 (01:39 +0900)]
nilfs2: simplify remaining sget() use
This simplifies the test function passed on the remaining sget()
callsite in nilfs.
Instead of checking mount type (i.e. ro-mount/rw-mount/snapshot mount)
in the test function passed to sget(), this patch first looks up the
nilfs_sb_info struct which the given mount type matches, and then
acquires the super block instance holding the nilfs_sb_info.
Signed-off-by: Ryusuke Konishi <konishi.ryusuke@lab.ntt.co.jp> Cc: Al Viro <viro@zeniv.linux.org.uk> Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
Ryusuke Konishi [Sun, 7 Jun 2009 16:39:30 +0000 (01:39 +0900)]
nilfs2: get rid of sget use for checking if current mount is present
This stops using sget() for checking if an r/w-mount or an r/o-mount
exists on the device. This elimination uses a back pointer to the
current mount added to nilfs object.
Signed-off-by: Ryusuke Konishi <konishi.ryusuke@lab.ntt.co.jp> Cc: Al Viro <viro@zeniv.linux.org.uk> Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
Ryusuke Konishi [Sun, 7 Jun 2009 16:39:29 +0000 (01:39 +0900)]
nilfs2: get rid of sget use for acquiring nilfs object
This will change the way to obtain nilfs object in nilfs_get_sb()
function.
Previously, a preliminary sget() call was performed, and the nilfs
object was acquired from a super block instance found by the sget()
call.
This patch, instead, instroduces a new dedicated function
find_or_create_nilfs(); as the name implies, the function finds an
existent nilfs object from a global list or creates a new one if no
object is found on the device.
Signed-off-by: Ryusuke Konishi <konishi.ryusuke@lab.ntt.co.jp> Cc: Al Viro <viro@zeniv.linux.org.uk> Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
remove the call to ->write_super in __sync_filesystem
Now that all filesystems provide ->sync_fs methods we can change
__sync_filesystem to only call ->sync_fs.
This gives us a clear separation between periodic writeouts which
are driven by ->write_super and data integrity syncs that go
through ->sync_fs. (modulo file_fsync which is also going away)
Signed-off-by: Christoph Hellwig <hch@lst.de> Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
Add a ->sync_fs method for data integrity syncs. Factor out common code
between affs_put_super, affs_write_super and the new affs_sync_fs into
a helper.
Signed-off-by: Christoph Hellwig <hch@lst.de> Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
Al Viro [Mon, 8 Jun 2009 05:22:00 +0000 (01:22 -0400)]
sanitize ->fsync() for affs
unfortunately, for affs (especially for affs directories) we have
no real way to keep track of metadata ownership. So we have to
do more or less what file_fsync() does, but we do *not* need to
call write_super() there.
Al Viro [Sun, 7 Jun 2009 17:44:36 +0000 (13:44 -0400)]
Sanitize ->fsync() for FAT
* mark directory data blocks as assoc. metadata
* add new inode to deal with FAT, mark FAT blocks as assoc. metadata of that
* now ->fsync() is trivial both for files and directories
Al Viro [Sun, 7 Jun 2009 13:30:08 +0000 (09:30 -0400)]
Sanitize qnx4 fsync handling
* have directory operations use mark_buffer_dirty_inode(),
so that sync_mapping_buffers() would get those.
* make qnx4_write_inode() honour its last argument.
* get rid of insane copies of very ancient "walk the indirect blocks"
in qnx4/fsync - they never matched the actual fs layout and, fortunately,
never'd been called. Again, all this junk is not needed; ->fsync()
should just do sync_mapping_buffers + sync_inode (and if we implement
block allocation for qnx4, we'll need to use mark_buffer_dirty_inode()
for extent blocks)
Al Viro [Sun, 7 Jun 2009 18:56:44 +0000 (14:56 -0400)]
New helper - simple_fsync()
writes associated buffers, then does sync_inode() to write
the inode itself (and to make it clean). Depends on
->write_inode() honouring the second argument.
Nick Piggin [Thu, 28 May 2009 07:01:15 +0000 (09:01 +0200)]
fs: block_dump missing dentry locking
I think the block_dump output in __mark_inode_dirty is missing dentry locking.
Surely the i_dentry list can change any time, so we may not even *get* a
dentry there. If we do get one by chance, then it would appear to be able to
go away or get renamed at any time...
Nick Piggin [Tue, 2 Jun 2009 10:07:47 +0000 (12:07 +0200)]
fs: remove incorrect I_NEW warnings
Some filesystems can call in to sync an inode that is still in the
I_NEW state (eg. ext family, when mounted with -osync). This is OK
because the filesystem has sole access to the new inode, so it can
modify i_state without races (because no other thread should be
modifying it, by definition of I_NEW). Ie. a false positive, so
remove the warnings.
xfs: remove ->write_super and stop maintaining ->s_dirt
the write_super method is used for
(1) writing back the superblock periodically from pdflush
(2) called just before ->sync_fs for data integerity syncs
We don't need (1) because we have our own peridoc writeout through xfssyncd,
and we don't need (2) because xfs_fs_sync_fs performs a proper synchronous
superblock writeout after all other data and metadata has been written out.
Also remove ->s_dirt tracking as it's only used to decide when too call
->write_super.
Signed-off-by: Christoph Hellwig <hch@lst.de> Reviewed-by: Eric Sandeen <sandeen@sandeen.net> Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
Theodore Ts'o [Thu, 21 May 2009 20:01:00 +0000 (16:01 -0400)]
fs: Remove i_cindex from struct inode
The only user of the i_cindex element in the inode structure is used
is by the firewire drivers. As part of an attempt to slim down the
inode structure to save memory --- since a typical Linux system will
have hundreds of thousands if not millions of inodes cached, a
reduction in the size inode has high leverage.
The firewire driver does not need i_cindex in any fast path, so it's
simple enough to calculate when it is needed, instead of wasting space
in the inode structure.
Push down lock_super into ->write_super instances and remove it from the
caller.
Following filesystem don't need ->s_lock in ->write_super and are skipped:
* bfs, nilfs2 - no other uses of s_lock and have internal locks in
->write_super
* ext2 - uses BKL in ext2_write_super and has internal calls without s_lock
* reiserfs - no other uses of s_lock as has reiserfs_write_lock (BKL) in
->write_super
* xfs - no other uses of s_lock and uses internal lock (buffer lock on
superblock buffer) to serialize ->write_super. Also xfs_fs_write_super
is superflous and will go away in the next merge window
Signed-off-by: Christoph Hellwig <hch@lst.de> Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
Al Viro [Wed, 6 May 2009 14:43:07 +0000 (10:43 -0400)]
Push lock_super() into the ->remount_fs() of filesystems that care about it
Note that since we can't run into contention between remount_fs and write_super
(due to exclusion on s_umount), we have to care only about filesystems that
touch lock_super() on their own. Out of those ext3, ext4, hpfs, sysv and ufs
do need it; fat doesn't since its ->remount_fs() only accesses assign-once
data (basically, it's "we have no atime on directories and only have atime on
files for vfat; force nodiratime and possibly noatime into *flags").
Move BKL into ->put_super from the only caller. A couple of
filesystems had trivial enough ->put_super (only kfree and NULLing of
s_fs_info + stuff in there) to not get any locking: coda, cramfs, efs,
hugetlbfs, omfs, qnx4, shmem, all others got the full treatment. Most
of them probably don't need it, but I'd rather sort that out individually.
Preferably after all the other BKL pushdowns in that area.
[AV: original used to move lock_super() down as well; these changes are
removed since we don't do lock_super() at all in generic_shutdown_super()
now]
[AV: fuse, btrfs and xfs are known to need no damn BKL, exempt]
Signed-off-by: Christoph Hellwig <hch@lst.de> Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
Al Viro [Wed, 6 May 2009 02:10:44 +0000 (22:10 -0400)]
No need to do lock_super() for exclusion in generic_shutdown_super()
We can't run into contention on it. All other callers of lock_super()
either hold s_umount (and we have it exclusive) or hold an active
reference to superblock in question, which prevents the call of
generic_shutdown_super() while the reference is held. So we can
replace lock_super(s) with get_fs_excl() in generic_shutdown_super()
(and corresponding change for unlock_super(), of course).
Since ext4 expects s_lock held for its put_super, take lock_super()
into it. The rest of filesystems do not care at all.
enforce ->sync_fs is only called for rw superblock
Make sure a superblock really is writeable by checking MS_RDONLY
under s_umount. sync_filesystems needed some re-arragement for
that, but all but one sync_filesystem caller had the correct locking
already so that we could add that check there. cachefiles grew
s_umount locking.
I've also added a WARN_ON to sync_filesystem to assert this for
future callers.
Signed-off-by: Christoph Hellwig <hch@lst.de> Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
Merge the write_super helper into sync_super and move the check for
->write_super earlier so that we can avoid grabbing a reference to
a superblock that doesn't have it.
While we're at it also add a little comment documenting sync_supers.
Signed-off-by: Christoph Hellwig <hch@lst.de> Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
Alexey Dobriyan [Sun, 3 May 2009 23:32:03 +0000 (03:32 +0400)]
dcache: extrace and use d_unlinked()
d_unlinked() will be used in middle-term to ban checkpointing when opened
but unlinked file is detected, and in long term, to detect such situation
and special case on it.
Signed-off-by: Alexey Dobriyan <adobriyan@gmail.com> Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
remove ->write_super call in generic_shutdown_super
We just did a full fs writeout using sync_filesystem before, and if
that's not enough for the filesystem it can perform it's own writeout
in ->put_super, which many filesystems already do.
Move a call to foofs_write_super into every foofs_put_super for now to
guarantee identical behaviour until it's cleaned up by the individual
filesystem maintainers.
Exceptions:
- affs already has identical copy & pasted code at the beginning of
affs_put_super so no need to do it twice.
- xfs does the right thing without it and I have changes pending for
the xfs tree touching this are so I don't really need conflicts
here..
Signed-off-by: Christoph Hellwig <hch@lst.de> Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
Jan Kara [Mon, 27 Apr 2009 14:43:55 +0000 (16:43 +0200)]
quota: Introduce writeout_quota_sb() (version 4)
Introduce this function which just writes all the quota structures but
avoids all the syncing and cache pruning work to expose quota structures
to userspace. Use this function from __sync_filesystem when wait == 0.
Signed-off-by: Jan Kara <jack@suse.cz> Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
Currently the VFS calls vfs_dq_sync to sync out disk quotas for a given
superblock. This is a small wrapper around sync_dquots which for the
case of a non-NULL superblock is a small wrapper around quota_sync_sb.
Just make quota_sync_sb global (rename it to sync_quota_sb) and call it
directly. Also call it directly for those cases in quota.c that have a
superblock and leave sync_dquots purely an iterator over sync_quota_sb and
remove it's superblock argument.
To make this nicer move the check for the lack of a quota_sync method
from the callers into sync_quota_sb.
[folded build fix from Alexander Beregalov <a.beregalov@gmail.com>]
Signed-off-by: Christoph Hellwig <hch@lst.de> Signed-off-by: Jan Kara <jack@suse.cz> Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
Jan Kara [Mon, 27 Apr 2009 14:43:51 +0000 (16:43 +0200)]
vfs: Make sys_sync() use fsync_super() (version 4)
It is unnecessarily fragile to have two places (fsync_super() and do_sync())
doing data integrity sync of the filesystem. Alter __fsync_super() to
accommodate needs of both callers and use it. So after this patch
__fsync_super() is the only place where we gather all the calls needed to
properly send all data on a filesystem to disk.
Nice bonus is that we get a complete livelock avoidance and write_supers()
is now only used for periodic writeback of superblocks.
sync_blockdevs() introduced a couple of patches ago is gone now.
[build fixes folded]
Signed-off-by: Jan Kara <jack@suse.cz> Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
Jan Kara [Mon, 27 Apr 2009 14:43:50 +0000 (16:43 +0200)]
vfs: Make __fsync_super() a static function (version 4)
__fsync_super() does the same thing as fsync_super(). So change the only
caller to use fsync_super() and make __fsync_super() static. This removes
unnecessarily duplicated call to sync_blockdev() and prepares ground
for the changes to __fsync_super() in the following patches.
Signed-off-by: Jan Kara <jack@suse.cz> Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
Jan Kara [Mon, 27 Apr 2009 14:43:49 +0000 (16:43 +0200)]
vfs: Call ->sync_fs() even if s_dirt is 0 (version 4)
sync_filesystems() has a condition that if wait == 0 and s_dirt == 0, then
->sync_fs() isn't called. This does not really make much sence since s_dirt is
generally used by a filesystem to mean that ->write_super() needs to be called.
But ->sync_fs() does different things. I even suspect that some filesystems
(btrfs?) sets s_dirt just to fool this logic.
Signed-off-by: Jan Kara <jack@suse.cz> Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
Jan Kara [Mon, 27 Apr 2009 14:43:48 +0000 (16:43 +0200)]
vfs: Fix sys_sync() and fsync_super() reliability (version 4)
So far, do_sync() called:
sync_inodes(0);
sync_supers();
sync_filesystems(0);
sync_filesystems(1);
sync_inodes(1);
This ordering makes it kind of hard for filesystems as sync_inodes(0) need not
submit all the IO (for example it skips inodes with I_SYNC set) so e.g. forcing
transaction to disk in ->sync_fs() is not really enough. Therefore sys_sync has
not been completely reliable on some filesystems (ext3, ext4, reiserfs, ocfs2
and others are hit by this) when racing e.g. with background writeback. A
similar problem hits also other filesystems (e.g. ext2) because of
write_supers() being called before the sync_inodes(1).
Change the ordering of calls in do_sync() - this requires a new function
sync_blockdevs() to preserve the property that block devices are always synced
after write_super() / sync_fs() call.
The same issue is fixed in __fsync_super() function used on umount /
remount read-only.
[AV: build fixes]
Signed-off-by: Jan Kara <jack@suse.cz> Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
Remove the unused s_async_list in the superblock, a leftover of the
broken async inode deletion code that leaked into mainline. Having this
in the middle of the sync/unmount path is not helpful for the following
cleanups.
Signed-off-by: Christoph Hellwig <hch@lst.de> Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
This patch speeds up lmbench lat_mmap test by about another 2% after the
first patch.
Before:
avg = 462.286
std = 5.46106
After:
avg = 453.12
std = 9.58257
(50 runs of each, stddev gives a reasonable confidence)
It does this by introducing mnt_clone_write, which avoids some heavyweight
operations of mnt_want_write if called on a vfsmount which we know already
has a write count; and mnt_want_write_file, which can call mnt_clone_write
if the file is open for write.
After these two patches, mnt_want_write and mnt_drop_write go from 7% on
the profile down to 1.3% (including mnt_clone_write).
[AV: mnt_want_write_file() should take file alone and derive mnt from it;
not only all callers have that form, but that's the only mnt about which
we know that it's already held for write if file is opened for write]
Cc: Dave Hansen <haveblue@us.ibm.com> Signed-off-by: Nick Piggin <npiggin@suse.de> Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
This patch speeds up lmbench lat_mmap test by about 8%. lat_mmap is set up
basically to mmap a 64MB file on tmpfs, fault in its pages, then unmap it.
A microbenchmark yes, but it exercises some important paths in the mm.
Before:
avg = 501.9
std = 14.7773
After:
avg = 462.286
std = 5.46106
(50 runs of each, stddev gives a reasonable confidence, but there is quite
a bit of variation there still)
It does this by removing the complex per-cpu locking and counter-cache and
replaces it with a percpu counter in struct vfsmount. This makes the code
much simpler, and avoids spinlocks (although the msync is still pretty
costly, unfortunately). It results in about 900 bytes smaller code too. It
does increase the size of a vfsmount, however.
It should also give a speedup on large systems if CPUs are frequently operating
on different mounts (because the existing scheme has to operate on an atomic in
the struct vfsmount when switching between mounts). But I'm most interested in
the single threaded path performance for the moment.
[AV: minor cleanup]
Cc: Dave Hansen <haveblue@us.ibm.com> Signed-off-by: Nick Piggin <npiggin@suse.de> Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
Code Review / linux-2.6.git / log