fs: scale files_lock
Improve scalability of files_lock by adding per-cpu, per-sb files lists,
protected with an lglock. The lglock provides fast access to the per-cpu lists
to add and remove files. It also provides a snapshot of all the per-cpu lists
(although this is very slow).
One difficulty with this approach is that a file can be removed from the list
by another CPU. We must track which per-cpu list the file is on with a new
variale in the file struct (packed into a hole on 64-bit archs). Scalability
could suffer if files are frequently removed from different cpu's list.
However loads with frequent removal of files imply short interval between
adding and removing the files, and the scheduler attempts to avoid moving
processes too far away. Also, even in the case of cross-CPU removal, the
hardware has much more opportunity to parallelise cacheline transfers with N
cachelines than with 1.
A worst-case test of 1 CPU allocating files subsequently being freed by N CPUs
degenerates to contending on a single lock, which is no worse than before. When
more than one CPU are allocating files, even if they are always freed by
different CPUs, there will be more parallelism than the single-lock case.
Testing results:
On a 2 socket, 8 core opteron, I measure the number of times the lock is taken
to remove the file, the number of times it is removed by the same CPU that
added it, and the number of times it is removed by the same node that added it.
Booting: locks= 25049 cpu-hits= 23174 (92.5%) node-hits= 23945 (95.6%)
kbuild -j16 locks=2281913 cpu-hits=2208126 (96.8%) node-hits=2252674 (98.7%)
dbench 64 locks=4306582 cpu-hits=4287247 (99.6%) node-hits=4299527 (99.8%)
So a file is removed from the same CPU it was added by over 90% of the time.
It remains within the same node 95% of the time.
Tim Chen ran some numbers for a 64 thread Nehalem system performing a compile.
throughput
2.6.34-rc2 24.5
+patch 24.9
us sys idle IO wait (in %)
2.6.34-rc2 51.25 28.25 17.25 3.25
+patch 53.75 18.5 19 8.75
So significantly less CPU time spent in kernel code, higher idle time and
slightly higher throughput.
Single threaded performance difference was within the noise of microbenchmarks.
That is not to say penalty does not exist, the code is larger and more memory
accesses required so it will be slightly slower.
Cc: linux-kernel@vger.kernel.org
Cc: Tim Chen <tim.c.chen@linux.intel.com>
Cc: Andi Kleen <ak@linux.intel.com>
Signed-off-by: Nick Piggin <npiggin@kernel.dk>
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
fs: cleanup files_lock locking
Lock tty_files with a new spinlock, tty_files_lock; provide helpers to
manipulate the per-sb files list; unexport the files_lock spinlock.
Cc: linux-kernel@vger.kernel.org
Cc: Christoph Hellwig <hch@infradead.org>
Cc: Alan Cox <alan@lxorguk.ukuu.org.uk>
Acked-by: Andi Kleen <ak@linux.intel.com>
Acked-by: Greg Kroah-Hartman <gregkh@suse.de>
Signed-off-by: Nick Piggin <npiggin@kernel.dk>
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
This reverts commit 3bcf3860a4 (and the
accompanying commit c1e5c95402 "vfs/fsnotify: fsnotify_close can delay
the final work in fput" that was a horribly ugly hack to make it work at
all).
The 'struct file' approach not only causes that disgusting hack, it
somehow breaks pulseaudio, probably due to some other subtlety with
f_count handling.
Fix up various conflicts due to later fsnotify work.
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Improve the description of fget_light(), which is currently incorrect
about needing a prior refcnt (judging by the way it is actually used).
Signed-off-by: Tony Battersby <tonyb@cybernetics.com>
Cc: Al Viro <viro@zeniv.linux.org.uk>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
fanotify almost works like so:
user context calls fsnotify_* function with a struct file.
fsnotify takes a reference on the struct path
user context goes about it's buissiness
at some later point in time the fsnotify listener gets the struct path
fanotify listener calls dentry_open() to create a file which userspace can deal with
listener drops the reference on the struct path
at some later point the listener calls close() on it's new file
With the switch from struct path to struct file this presents a problem for
fput() and fsnotify_close(). fsnotify_close() is called when the filp has
already reached 0 and __fput() wants to do it's cleanup.
The solution presented here is a bit odd. If an event is created from a
struct file we take a reference on the file. We check however if the f_count
was already 0 and if so we take an EXTRA reference EVEN THOUGH IT WAS ZERO.
In __fput() (where we know the f_count hit 0 once) we check if the f_count is
non-zero and if so we drop that 'extra' ref and return without destroying the
file.
Signed-off-by: Eric Paris <eparis@redhat.com>
__aio_put_req() plays sick games with file refcount. What
it wants is fput() from atomic context; it's almost always
done with f_count > 1, so they only have to deal with delayed
work in rare cases when their reference happens to be the
last one. Current code decrements f_count and if it hasn't
hit 0, everything is fine. Otherwise it keeps a pointer
to struct file (with zero f_count!) around and has delayed
work do __fput() on it.
Better way to do it: use atomic_long_add_unless( , -1, 1)
instead of !atomic_long_dec_and_test(). IOW, decrement it
only if it's not the last reference, leave refcount alone
if it was. And use normal fput() in delayed work.
I've made that atomic_long_add_unless call a new helper -
fput_atomic(). Drops a reference to file if it's safe to
do in atomic (i.e. if that's not the last one), tells if
it had been able to do that. aio.c converted to it, __fput()
use is gone. req->ki_file *always* contributes to refcount
now. And __fput() became static.
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
We'll introduce FMODE_RANDOM which will be runtime modified. So protect
all runtime modification to f_mode with f_lock to avoid races.
Signed-off-by: Wu Fengguang <fengguang.wu@intel.com>
Cc: Al Viro <viro@zeniv.linux.org.uk>
Cc: Christoph Hellwig <hch@infradead.org>
Cc: Trond Myklebust <Trond.Myklebust@netapp.com>
Cc: Chuck Lever <chuck.lever@oracle.com>
Cc: <stable@kernel.org> [2.6.33.x]
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
When alloc_file() and init_file() were combined, the error handling of
mnt_clone_write() was taken into alloc_file() in a somewhat obfuscated
way. Since we don't use the error code for anything except warning,
we might as well warn directly without an extra variable.
Signed-off-by: Roland Dreier <rolandd@cisco.com>
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
Commit 3d1e4631 ("get rid of init_file()") removed the export of
alloc_file() -- possibly inadvertently, since that commit mainly
consisted of deleting the lines between the end of alloc_file() and
the start of the code in init_file().
There is in fact one modular use of alloc_file() in the tree, in
drivers/infiniband/core/uverbs_main.c, so re-add the export to fix:
ERROR: "alloc_file" [drivers/infiniband/core/ib_uverbs.ko] undefined!
when CONFIG_INFINIBAND_USER_ACCESS=m.
Cc: Al Viro <viro@zeniv.linux.org.uk>
Signed-off-by: Roland Dreier <rolandd@cisco.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
There are 2 groups of alloc_file() callers:
* ones that are followed by ima_counts_get
* ones giving non-regular files
So let's pull that ima_counts_get() into alloc_file();
it's a no-op in case of non-regular files.
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
All users outside of fs/ of get_empty_filp() have been removed. This patch
moves the definition from the include/ directory to internal.h so no new
users crop up and removes the EXPORT_SYMBOL. I'd love to see open intents
stop using it too, but that's a problem for another day and a smarter
developer!
Signed-off-by: Eric Paris <eparis@redhat.com>
Acked-by: Miklos Szeredi <miklos@szeredi.hu>
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
Based on discussions on LKML and LSM, where there are consecutive
security_ and ima_ calls in the vfs layer, move the ima_ calls to
the existing security_ hooks.
Signed-off-by: Mimi Zohar <zohar@us.ibm.com>
Signed-off-by: James Morris <jmorris@namei.org>
It's unused.
It isn't needed -- read or write flag is already passed and sysctl
shouldn't care about the rest.
It _was_ used in two places at arch/frv for some reason.
Signed-off-by: Alexey Dobriyan <adobriyan@gmail.com>
Cc: David Howells <dhowells@redhat.com>
Cc: "Eric W. Biederman" <ebiederm@xmission.com>
Cc: Al Viro <viro@zeniv.linux.org.uk>
Cc: Ralf Baechle <ralf@linux-mips.org>
Cc: Martin Schwidefsky <schwidefsky@de.ibm.com>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: "David S. Miller" <davem@davemloft.net>
Cc: James Morris <jmorris@namei.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
This function walks the s_files lock, and operates primarily on the
files in a superblock, so it better belongs here (eg. see also
fs_may_remount_ro).
[AV: ... and it shouldn't be static after that move]
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 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 lock moves out of the CONFIG_EPOLL ifdef and becomes f_lock. For now,
epoll remains the only user, but a future patch will use it to protect
f_flags as well.
Cc: Davide Libenzi <davidel@xmailserver.org>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Jonathan Corbet <corbet@lwn.net>
This patch replaces the generic integrity hooks, for which IMA registered
itself, with IMA integrity hooks in the appropriate places directly
in the fs directory.
Signed-off-by: Mimi Zohar <zohar@us.ibm.com>
Acked-by: Serge Hallyn <serue@us.ibm.com>
Signed-off-by: James Morris <jmorris@namei.org>
Instead of creating the "filp" kmem_cache in vfs_caches_init(),
we can do it a litle be later in files_init(), so that filp_cachep
is static to fs/file_table.c
Acked-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Signed-off-by: Eric Dumazet <dada1@cosmosbay.com>
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
Attach creds to file structs and discard f_uid/f_gid.
file_operations::open() methods (such as hppfs_open()) should use file->f_cred
rather than current_cred(). At the moment file->f_cred will be current_cred()
at this point.
Signed-off-by: David Howells <dhowells@redhat.com>
Reviewed-by: James Morris <jmorris@namei.org>
Signed-off-by: James Morris <jmorris@namei.org>
Wrap current->cred and a few other accessors to hide their actual
implementation.
Signed-off-by: David Howells <dhowells@redhat.com>
Acked-by: James Morris <jmorris@namei.org>
Acked-by: Serge Hallyn <serue@us.ibm.com>
Signed-off-by: James Morris <jmorris@namei.org>
Separate the task security context from task_struct. At this point, the
security data is temporarily embedded in the task_struct with two pointers
pointing to it.
Note that the Alpha arch is altered as it refers to (E)UID and (E)GID in
entry.S via asm-offsets.
With comment fixes Signed-off-by: Marc Dionne <marc.c.dionne@gmail.com>
Signed-off-by: David Howells <dhowells@redhat.com>
Acked-by: James Morris <jmorris@namei.org>
Acked-by: Serge Hallyn <serue@us.ibm.com>
Signed-off-by: James Morris <jmorris@namei.org>
As it is, all instances of ->release() for files that have ->fasync()
need to remember to evict file from fasync lists; forgetting that
creates a hole and we actually have a bunch that *does* forget.
So let's keep our lives simple - let __fput() check FASYNC in
file->f_flags and call ->fasync() there if it's been set. And lose that
crap in ->release() instances - leaving it there is still valid, but we
don't have to bother anymore.
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
make it atomic_long_t; while we are at it, get rid of useless checks in affs,
hfs and hpfs - ->open() always has it equal to 1, ->release() - to 0.
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
There have been a few oopses caused by 'struct file's with NULL f_vfsmnts.
There was also a set of potentially missed mnt_want_write()s from
dentry_open() calls.
This patch provides a very simple debugging framework to catch these kinds of
bugs. It will WARN_ON() them, but should stop us from having any oopses or
mnt_writer count imbalances.
I'm quite convinced that this is a good thing because it found bugs in the
stuff I was working on as soon as I wrote it.
[hch: made it conditional on a debug option.
But it's still a little bit too ugly]
[hch: merged forced remount r/o fix from Dave and akpm's fix for the fix]
Signed-off-by: Dave Hansen <haveblue@us.ibm.com>
Acked-by: Al Viro <viro@ZenIV.linux.org.uk>
Signed-off-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
This is the first really tricky patch in the series. It elevates the writer
count on a mount each time a non-special file is opened for write.
We used to do this in may_open(), but Miklos pointed out that __dentry_open()
is used as well to create filps. This will cover even those cases, while a
call in may_open() would not have.
There is also an elevated count around the vfs_create() call in open_namei().
See the comments for more details, but we need this to fix a 'create, remount,
fail r/w open()' race.
Some filesystems forego the use of normal vfs calls to create
struct files. Make sure that these users elevate the mnt
writer count because they will get __fput(), and we need
to make sure they're balanced.
Acked-by: Al Viro <viro@ZenIV.linux.org.uk>
Signed-off-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Dave Hansen <haveblue@us.ibm.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
If someone decides to demote a file from r/w to just
r/o, they can use this same code as __fput().
NFS does just that, and will use this in the next
patch.
AV: drop write access in __fput() only after we evict from file list.
Signed-off-by: Dave Hansen <haveblue@us.ibm.com>
Cc: Erez Zadok <ezk@cs.sunysb.edu>
Cc: Trond Myklebust <trond.myklebust@fys.uio.no>
Cc: "J Bruce Fields" <bfields@fieldses.org>
Acked-by: Al Viro <viro@ZenIV.linux.org.uk>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
Some new uses of get_empty_filp() have crept in; switched
to alloc_file() to make sure that pieces of initialization
won't be missing.
We really need to kill get_empty_filp().
[AV] fixed dentry leak on failure exit in anon_inode_getfd()
Cc: Erez Zadok <ezk@cs.sunysb.edu>
Cc: Trond Myklebust <trond.myklebust@fys.uio.no>
Cc: "J Bruce Fields" <bfields@fieldses.org>
Acked-by: Al Viro <viro@ZenIV.linux.org.uk>
Signed-off-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Dave Hansen <haveblue@us.ibm.com>
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
fs/file_table.c: use list_for_each_entry() instead of list_for_each()
in fs_may_remount_ro()
Signed-off-by: Matthias Kaehlcke <matthias.kaehlcke@gmail.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Why do we need r/o bind mounts?
This feature allows a read-only view into a read-write filesystem. In the
process of doing that, it also provides infrastructure for keeping track of
the number of writers to any given mount.
This has a number of uses. It allows chroots to have parts of filesystems
writable. It will be useful for containers in the future because users may
have root inside a container, but should not be allowed to write to
somefilesystems. This also replaces patches that vserver has had out of the
tree for several years.
It allows security enhancement by making sure that parts of your filesystem
read-only (such as when you don't trust your FTP server), when you don't want
to have entire new filesystems mounted, or when you want atime selectively
updated. I've been using the following script to test that the feature is
working as desired. It takes a directory and makes a regular bind and a r/o
bind mount of it. It then performs some normal filesystem operations on the
three directories, including ones that are expected to fail, like creating a
file on the r/o mount.
This patch:
Some filesystems forego the vfs and may_open() and create their own 'struct
file's.
This patch creates a couple of helper functions which can be used by these
filesystems, and will provide a unified place which the r/o bind mount code
may patch.
Also, rename an existing, static-scope init_file() to a less generic name.
Signed-off-by: Dave Hansen <haveblue@us.ibm.com>
Cc: Christoph Hellwig <hch@lst.de>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
s/percpu_counter_sum/&_positive/
Because its consitent with percpu_counter_read*
Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Remove includes of <linux/smp_lock.h> where it is not used/needed.
Suggested by Al Viro.
Builds cleanly on x86_64, i386, alpha, ia64, powerpc, sparc,
sparc64, and arm (all 59 defconfigs).
Signed-off-by: Randy Dunlap <randy.dunlap@oracle.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
This patch changes struct file to use struct path instead of having
independent pointers to struct dentry and struct vfsmount, and converts all
users of f_{dentry,vfsmnt} in fs/ to use f_path.{dentry,mnt}.
Additionally, it adds two #define's to make the transition easier for users of
the f_dentry and f_vfsmnt.
Signed-off-by: Josef "Jeff" Sipek <jsipek@cs.sunysb.edu>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
File handles can be requested to send sigio and sigurg to processes. By
tracking the destination processes using struct pid instead of pid_t we make
the interface safe from all potential pid wrap around problems.
Signed-off-by: Eric W. Biederman <ebiederm@xmission.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Move the i_cdev pointer in struct inode into a union.
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
The percpu counter data type are changed in this set of patches to support
more users like ext3 who need more than 32 bit to store the free blocks
total in the filesystem.
- Generic perpcu counters data type changes. The size of the global counter
and local counter were explictly specified using s64 and s32. The global
counter is changed from long to s64, while the local counter is changed from
long to s32, so we could avoid doing 64 bit update in most cases.
- Users of the percpu counters are updated to make use of the new
percpu_counter_init() routine now taking an additional parameter to allow
users to pass the initial value of the global counter.
Signed-off-by: Mingming Cao <cmm@us.ibm.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Eliminate a handful of cache references by keeping current in a register
instead of reloading (helps x86) and avoiding the overhead of a function
call. Inlining eventpoll_init_file() saves 24 bytes. Also reorder file
initialization to make writes occur more sequentially.
Signed-off-by: Benjamin LaHaise <bcrl@linux.intel.com>
Cc: Davide Libenzi <davidel@xmailserver.org>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
I have benchmarked this on an x86_64 NUMA system and see no significant
performance difference on kernbench. Tested on both x86_64 and powerpc.
The way we do file struct accounting is not very suitable for batched
freeing. For scalability reasons, file accounting was
constructor/destructor based. This meant that nr_files was decremented
only when the object was removed from the slab cache. This is susceptible
to slab fragmentation. With RCU based file structure, consequent batched
freeing and a test program like Serge's, we just speed this up and end up
with a very fragmented slab -
llm22:~ # cat /proc/sys/fs/file-nr
587730 0 758844
At the same time, I see only a 2000+ objects in filp cache. The following
patch I fixes this problem.
This patch changes the file counting by removing the filp_count_lock.
Instead we use a separate percpu counter, nr_files, for now and all
accesses to it are through get_nr_files() api. In the sysctl handler for
nr_files, we populate files_stat.nr_files before returning to user.
Counting files as an when they are created and destroyed (as opposed to
inside slab) allows us to correctly count open files with RCU.
Signed-off-by: Dipankar Sarma <dipankar@in.ibm.com>
Cc: "Paul E. McKenney" <paulmck@us.ibm.com>
Cc: "David S. Miller" <davem@davemloft.net>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
fs: Use <linux/capability.h> where capable() is used.
Signed-off-by: Randy Dunlap <rdunlap@xenotime.net>
Acked-by: Tim Schmielau <tim@physik3.uni-rostock.de>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Use atomic_inc_not_zero for rcu files instead of special case rcuref.
Signed-off-by: Nick Piggin <npiggin@suse.de>
Cc: "Paul E. McKenney" <paulmck@us.ibm.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
This patch renames struct kmem_cache_s to kmem_cache so we can start using
it instead of kmem_cache_t typedef.
Signed-off-by: Pekka Enberg <penberg@cs.helsinki.fi>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Nick Piggin