openwrt/staging/blogic.git
18 years ago[PATCH] RTC subsystem: EP93XX driver
Alessandro Zummo [Mon, 27 Mar 2006 09:16:45 +0000 (01:16 -0800)]
[PATCH] RTC subsystem: EP93XX driver

This patch adds a driver for the RTC embedded in the Cirrus Logic EP93XX
family of processors.

Signed-off-by: Alessandro Zummo <a.zummo@towertech.it>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] RTC subsystem: RS5C372 driver
Alessandro Zummo [Mon, 27 Mar 2006 09:16:45 +0000 (01:16 -0800)]
[PATCH] RTC subsystem: RS5C372 driver

RTC class aware driver for the Ricoh RS5C372 chip used, among others, on the
Synology DS101.

Signed-off-by: Alessandro Zummo <a.zummo@towertech.it>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] RTC subsystem: PCF8563 driver
Alessandro Zummo [Mon, 27 Mar 2006 09:16:44 +0000 (01:16 -0800)]
[PATCH] RTC subsystem: PCF8563 driver

An RTC class aware driver for the Philips PCF8563 RTC and Epson RTC8564 chips.

This chip is used on the Iomega NAS100D.

Signed-off-by: Alessandro Zummo <a.zummo@towertech.it>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] RTC subsystem: DS1672 driver
Alessandro Zummo [Mon, 27 Mar 2006 09:16:43 +0000 (01:16 -0800)]
[PATCH] RTC subsystem: DS1672 driver

Driver for the Dallas/Maxim DS1672 chip, found on the Loft
(http://www.giantshoulderinc.com).

Signed-off-by: Alessandro Zummo <a.zummo@towertech.it>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] RTC subsystem: test device/driver
Alessandro Zummo [Mon, 27 Mar 2006 09:16:42 +0000 (01:16 -0800)]
[PATCH] RTC subsystem: test device/driver

Interrupts can be generated by

echo "alarm|tick|update" >/sys/class/rtc/rtcX/device/irq

Signed-off-by: Alessandro Zummo <a.zummo@towertech.it>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] RTC subsystem: X1205 driver
Alessandro Zummo [Mon, 27 Mar 2006 09:16:42 +0000 (01:16 -0800)]
[PATCH] RTC subsystem: X1205 driver

A port of the existing x1205 driver under the new RTC subsystem.

It is actually under test within the NSLU2 project
(http://www.nslu2-linux.org) and it is working quite well.

It is the first driver under this new subsystem and should be used as a guide
to port other drivers.

Signed-off-by: Alessandro Zummo <a.zummo@towertech.it>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] RTC subsystem: dev interface
Alessandro Zummo [Mon, 27 Mar 2006 09:16:41 +0000 (01:16 -0800)]
[PATCH] RTC subsystem: dev interface

Add the dev interface to the RTC subsystem.

Each RTC will be available under /dev/rtcX .  A symlink from /dev/rtc0 to
/dev/rtc cab be obtained with the following udev rule:

KERNEL=="rtc0", SYMLINK+="rtc"

Signed-off-by: Alessandro Zummo <a.zummo@towertech.it>
Acked-by: Greg Kroah-Hartman <gregkh@suse.de>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] RTC subsystem: proc interface
Alessandro Zummo [Mon, 27 Mar 2006 09:16:40 +0000 (01:16 -0800)]
[PATCH] RTC subsystem: proc interface

Add the proc interface to the RTC subsystem.

The first RTC driver which registers with the class will be accessible by
/proc/driver/rtc .

This is required for compatibility with the standard RTC driver and to avoid
breaking any user space application which may erroneusly rely on this.

Signed-off-by: Alessandro Zummo <a.zummo@towertech.it>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] RTC subsystem: sysfs interface
Alessandro Zummo [Mon, 27 Mar 2006 09:16:39 +0000 (01:16 -0800)]
[PATCH] RTC subsystem: sysfs interface

This patch adds the sysfs interface to the RTC subsystem.

Each RTC client will have his own entry under /sys/classs/rtc/rtcN .

Within this entry some attributes are exported by the subsystem, like date and
time.

Signed-off-by: Alessandro Zummo <a.zummo@towertech.it>
Acked-by: Greg Kroah-Hartman <gregkh@suse.de>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] RTC subsystem: I2C driver ids
Alessandro Zummo [Mon, 27 Mar 2006 09:16:38 +0000 (01:16 -0800)]
[PATCH] RTC subsystem: I2C driver ids

This patch adds the I2C driver ids to i2c-id.h in preparation of the I2C
direct probing method.

This is kept separate so that it can be integrated to

Signed-off-by: Alessandro Zummo <a.zummo@towertech.it>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] RTC subsystem: I2C cleanup
Alessandro Zummo [Mon, 27 Mar 2006 09:16:37 +0000 (01:16 -0800)]
[PATCH] RTC subsystem: I2C cleanup

This patch, completely optional, removes from drivers/i2c/chips all the
drivers that are implemented in the new RTC subsystem.

It should be noted that none of the current driver is actually integrated,
i.e.  usable without further patches.

Signed-off-by: Alessandro Zummo <a.zummo@towertech.it>
Acked-by: Greg Kroah-Hartman <gregkh@suse.de>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] RTC subsystem: class
Alessandro Zummo [Mon, 27 Mar 2006 09:16:37 +0000 (01:16 -0800)]
[PATCH] RTC subsystem: class

Add the basic RTC subsystem infrastructure to the kernel.

rtc/class.c - registration facilities for RTC drivers
rtc/interface.c - kernel/rtc interface functions
rtc/hctosys.c - snippet of code that copies hw clock to sw clock
at bootup, if configured to do so.

Signed-off-by: Alessandro Zummo <a.zummo@towertech.it>
Acked-by: Greg Kroah-Hartman <gregkh@suse.de>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] RTC subsystem: ARM Integrator cleanup
Richard Purdie [Mon, 27 Mar 2006 09:16:36 +0000 (01:16 -0800)]
[PATCH] RTC subsystem: ARM Integrator cleanup

Fix some namespace conflicts between the RTC subsystem and the ARM Integrator
time functions.

Signed-off-by: Richard Purdie <rpurdie@rpsys.net>
Signed-off-by: Alessandro Zummo <a.zummo@towertech.it>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] RTC subsystem: ARM cleanup
Alessandro Zummo [Mon, 27 Mar 2006 09:16:35 +0000 (01:16 -0800)]
[PATCH] RTC subsystem: ARM cleanup

This patch removes from the ARM subsytem some of the rtc-related functions
that have been included in the RTC subsystem.  It also fixes some naming
collisions.

Signed-off-by: Alessandro Zummo <a.zummo@towertech.it>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] RTC Subsystem: library functions
Alessandro Zummo [Mon, 27 Mar 2006 09:16:34 +0000 (01:16 -0800)]
[PATCH] RTC Subsystem: library functions

RTC and date/time related functions.

Signed-off-by: Alessandro Zummo <a.zummo@towertech.it>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] mips: fixed collision of rtc function name
Yoichi Yuasa [Mon, 27 Mar 2006 09:16:33 +0000 (01:16 -0800)]
[PATCH] mips: fixed collision of rtc function name

Fix the collision of rtc function name.

Signed-off-by: Yoichi Yuasa <yoichi_yuasa@tripeaks.co.jp>
Cc: Alessandro Zummo <a.zummo@towertech.it>
Cc: Ralf Baechle <ralf@linux-mips.org>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] Notifier chain update: API changes
Alan Stern [Mon, 27 Mar 2006 09:16:30 +0000 (01:16 -0800)]
[PATCH] Notifier chain update: API changes

The kernel's implementation of notifier chains is unsafe.  There is no
protection against entries being added to or removed from a chain while the
chain is in use.  The issues were discussed in this thread:

    http://marc.theaimsgroup.com/?l=linux-kernel&m=113018709002036&w=2

We noticed that notifier chains in the kernel fall into two basic usage
classes:

"Blocking" chains are always called from a process context
and the callout routines are allowed to sleep;

"Atomic" chains can be called from an atomic context and
the callout routines are not allowed to sleep.

We decided to codify this distinction and make it part of the API.  Therefore
this set of patches introduces three new, parallel APIs: one for blocking
notifiers, one for atomic notifiers, and one for "raw" notifiers (which is
really just the old API under a new name).  New kinds of data structures are
used for the heads of the chains, and new routines are defined for
registration, unregistration, and calling a chain.  The three APIs are
explained in include/linux/notifier.h and their implementation is in
kernel/sys.c.

With atomic and blocking chains, the implementation guarantees that the chain
links will not be corrupted and that chain callers will not get messed up by
entries being added or removed.  For raw chains the implementation provides no
guarantees at all; users of this API must provide their own protections.  (The
idea was that situations may come up where the assumptions of the atomic and
blocking APIs are not appropriate, so it should be possible for users to
handle these things in their own way.)

There are some limitations, which should not be too hard to live with.  For
atomic/blocking chains, registration and unregistration must always be done in
a process context since the chain is protected by a mutex/rwsem.  Also, a
callout routine for a non-raw chain must not try to register or unregister
entries on its own chain.  (This did happen in a couple of places and the code
had to be changed to avoid it.)

Since atomic chains may be called from within an NMI handler, they cannot use
spinlocks for synchronization.  Instead we use RCU.  The overhead falls almost
entirely in the unregister routine, which is okay since unregistration is much
less frequent that calling a chain.

Here is the list of chains that we adjusted and their classifications.  None
of them use the raw API, so for the moment it is only a placeholder.

  ATOMIC CHAINS
  -------------
arch/i386/kernel/traps.c: i386die_chain
arch/ia64/kernel/traps.c: ia64die_chain
arch/powerpc/kernel/traps.c: powerpc_die_chain
arch/sparc64/kernel/traps.c: sparc64die_chain
arch/x86_64/kernel/traps.c: die_chain
drivers/char/ipmi/ipmi_si_intf.c: xaction_notifier_list
kernel/panic.c: panic_notifier_list
kernel/profile.c: task_free_notifier
net/bluetooth/hci_core.c: hci_notifier
net/ipv4/netfilter/ip_conntrack_core.c: ip_conntrack_chain
net/ipv4/netfilter/ip_conntrack_core.c: ip_conntrack_expect_chain
net/ipv6/addrconf.c: inet6addr_chain
net/netfilter/nf_conntrack_core.c: nf_conntrack_chain
net/netfilter/nf_conntrack_core.c: nf_conntrack_expect_chain
net/netlink/af_netlink.c: netlink_chain

  BLOCKING CHAINS
  ---------------
arch/powerpc/platforms/pseries/reconfig.c: pSeries_reconfig_chain
arch/s390/kernel/process.c: idle_chain
arch/x86_64/kernel/process.c idle_notifier
drivers/base/memory.c: memory_chain
drivers/cpufreq/cpufreq.c cpufreq_policy_notifier_list
drivers/cpufreq/cpufreq.c cpufreq_transition_notifier_list
drivers/macintosh/adb.c: adb_client_list
drivers/macintosh/via-pmu.c sleep_notifier_list
drivers/macintosh/via-pmu68k.c sleep_notifier_list
drivers/macintosh/windfarm_core.c wf_client_list
drivers/usb/core/notify.c usb_notifier_list
drivers/video/fbmem.c fb_notifier_list
kernel/cpu.c cpu_chain
kernel/module.c module_notify_list
kernel/profile.c munmap_notifier
kernel/profile.c task_exit_notifier
kernel/sys.c reboot_notifier_list
net/core/dev.c netdev_chain
net/decnet/dn_dev.c: dnaddr_chain
net/ipv4/devinet.c: inetaddr_chain

It's possible that some of these classifications are wrong.  If they are,
please let us know or submit a patch to fix them.  Note that any chain that
gets called very frequently should be atomic, because the rwsem read-locking
used for blocking chains is very likely to incur cache misses on SMP systems.
(However, if the chain's callout routines may sleep then the chain cannot be
atomic.)

The patch set was written by Alan Stern and Chandra Seetharaman, incorporating
material written by Keith Owens and suggestions from Paul McKenney and Andrew
Morton.

[jes@sgi.com: restructure the notifier chain initialization macros]
Signed-off-by: Alan Stern <stern@rowland.harvard.edu>
Signed-off-by: Chandra Seetharaman <sekharan@us.ibm.com>
Signed-off-by: Jes Sorensen <jes@sgi.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] lightweight robust futexes updates 2
Ingo Molnar [Mon, 27 Mar 2006 09:16:28 +0000 (01:16 -0800)]
[PATCH] lightweight robust futexes updates 2

futex.h updates:

- get rid of FUTEX_OWNER_PENDING - it's not used
- reduce ROBUST_LIST_LIMIT to a saner value

Signed-off-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] lightweight robust futexes updates
Ingo Molnar [Mon, 27 Mar 2006 09:16:27 +0000 (01:16 -0800)]
[PATCH] lightweight robust futexes updates

- fix: initialize the robust list(s) to NULL in copy_process.

- doc update

- cleanup: rename _inuser to _inatomic

- __user cleanups and other small cleanups

Signed-off-by: Ingo Molnar <mingo@elte.hu>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Arjan van de Ven <arjan@infradead.org>
Cc: Ulrich Drepper <drepper@redhat.com>
Cc: Andi Kleen <ak@muc.de>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] lightweight robust futexes: x86_64
Ingo Molnar [Mon, 27 Mar 2006 09:16:26 +0000 (01:16 -0800)]
[PATCH] lightweight robust futexes: x86_64

x86_64: add the futex_atomic_cmpxchg_inuser() assembly implementation, and
wire up the new syscalls.

Signed-off-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Arjan van de Ven <arjan@infradead.org>
Acked-by: Ulrich Drepper <drepper@redhat.com>
Cc: Andi Kleen <ak@muc.de>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] lightweight robust futexes: i386
Ingo Molnar [Mon, 27 Mar 2006 09:16:25 +0000 (01:16 -0800)]
[PATCH] lightweight robust futexes: i386

i386: add the futex_atomic_cmpxchg_inuser() assembly implementation, and wire
up the new syscalls.

Signed-off-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Arjan van de Ven <arjan@infradead.org>
Acked-by: Ulrich Drepper <drepper@redhat.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] lightweight robust futexes: compat
Ingo Molnar [Mon, 27 Mar 2006 09:16:24 +0000 (01:16 -0800)]
[PATCH] lightweight robust futexes: compat

32-bit syscall compatibility support.  (This patch also moves all futex
related compat functionality into kernel/futex_compat.c.)

Signed-off-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Arjan van de Ven <arjan@infradead.org>
Acked-by: Ulrich Drepper <drepper@redhat.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] lightweight robust futexes: docs
Ingo Molnar [Mon, 27 Mar 2006 09:16:23 +0000 (01:16 -0800)]
[PATCH] lightweight robust futexes: docs

Add robust-futex documentation.

Signed-off-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] lightweight robust futexes: core
Ingo Molnar [Mon, 27 Mar 2006 09:16:22 +0000 (01:16 -0800)]
[PATCH] lightweight robust futexes: core

Add the core infrastructure for robust futexes: structure definitions, the new
syscalls and the do_exit() based cleanup mechanism.

Signed-off-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Arjan van de Ven <arjan@infradead.org>
Acked-by: Ulrich Drepper <drepper@redhat.com>
Cc: Michael Kerrisk <mtk-manpages@gmx.net>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] lightweight robust futexes: arch defaults
Ingo Molnar [Mon, 27 Mar 2006 09:16:21 +0000 (01:16 -0800)]
[PATCH] lightweight robust futexes: arch defaults

This patchset provides a new (written from scratch) implementation of robust
futexes, called "lightweight robust futexes".  We believe this new
implementation is faster and simpler than the vma-based robust futex solutions
presented before, and we'd like this patchset to be adopted in the upstream
kernel.  This is version 1 of the patchset.

  Background
  ----------

What are robust futexes?  To answer that, we first need to understand what
futexes are: normal futexes are special types of locks that in the
noncontended case can be acquired/released from userspace without having to
enter the kernel.

A futex is in essence a user-space address, e.g.  a 32-bit lock variable
field.  If userspace notices contention (the lock is already owned and someone
else wants to grab it too) then the lock is marked with a value that says
"there's a waiter pending", and the sys_futex(FUTEX_WAIT) syscall is used to
wait for the other guy to release it.  The kernel creates a 'futex queue'
internally, so that it can later on match up the waiter with the waker -
without them having to know about each other.  When the owner thread releases
the futex, it notices (via the variable value) that there were waiter(s)
pending, and does the sys_futex(FUTEX_WAKE) syscall to wake them up.  Once all
waiters have taken and released the lock, the futex is again back to
'uncontended' state, and there's no in-kernel state associated with it.  The
kernel completely forgets that there ever was a futex at that address.  This
method makes futexes very lightweight and scalable.

"Robustness" is about dealing with crashes while holding a lock: if a process
exits prematurely while holding a pthread_mutex_t lock that is also shared
with some other process (e.g.  yum segfaults while holding a pthread_mutex_t,
or yum is kill -9-ed), then waiters for that lock need to be notified that the
last owner of the lock exited in some irregular way.

To solve such types of problems, "robust mutex" userspace APIs were created:
pthread_mutex_lock() returns an error value if the owner exits prematurely -
and the new owner can decide whether the data protected by the lock can be
recovered safely.

There is a big conceptual problem with futex based mutexes though: it is the
kernel that destroys the owner task (e.g.  due to a SEGFAULT), but the kernel
cannot help with the cleanup: if there is no 'futex queue' (and in most cases
there is none, futexes being fast lightweight locks) then the kernel has no
information to clean up after the held lock!  Userspace has no chance to clean
up after the lock either - userspace is the one that crashes, so it has no
opportunity to clean up.  Catch-22.

In practice, when e.g.  yum is kill -9-ed (or segfaults), a system reboot is
needed to release that futex based lock.  This is one of the leading
bugreports against yum.

To solve this problem, 'Robust Futex' patches were created and presented on
lkml: the one written by Todd Kneisel and David Singleton is the most advanced
at the moment.  These patches all tried to extend the futex abstraction by
registering futex-based locks in the kernel - and thus give the kernel a
chance to clean up.

E.g.  in David Singleton's robust-futex-6.patch, there are 3 new syscall
variants to sys_futex(): FUTEX_REGISTER, FUTEX_DEREGISTER and FUTEX_RECOVER.
The kernel attaches such robust futexes to vmas (via
vma->vm_file->f_mapping->robust_head), and at do_exit() time, all vmas are
searched to see whether they have a robust_head set.

Lots of work went into the vma-based robust-futex patch, and recently it has
improved significantly, but unfortunately it still has two fundamental
problems left:

 - they have quite complex locking and race scenarios.  The vma-based
   patches had been pending for years, but they are still not completely
   reliable.

 - they have to scan _every_ vma at sys_exit() time, per thread!

The second disadvantage is a real killer: pthread_exit() takes around 1
microsecond on Linux, but with thousands (or tens of thousands) of vmas every
pthread_exit() takes a millisecond or more, also totally destroying the CPU's
L1 and L2 caches!

This is very much noticeable even for normal process sys_exit_group() calls:
the kernel has to do the vma scanning unconditionally!  (this is because the
kernel has no knowledge about how many robust futexes there are to be cleaned
up, because a robust futex might have been registered in another task, and the
futex variable might have been simply mmap()-ed into this process's address
space).

This huge overhead forced the creation of CONFIG_FUTEX_ROBUST, but worse than
that: the overhead makes robust futexes impractical for any type of generic
Linux distribution.

So it became clear to us, something had to be done.  Last week, when Thomas
Gleixner tried to fix up the vma-based robust futex patch in the -rt tree, he
found a handful of new races and we were talking about it and were analyzing
the situation.  At that point a fundamentally different solution occured to
me.  This patchset (written in the past couple of days) implements that new
solution.  Be warned though - the patchset does things we normally dont do in
Linux, so some might find the approach disturbing.  Parental advice
recommended ;-)

  New approach to robust futexes
  ------------------------------

At the heart of this new approach there is a per-thread private list of robust
locks that userspace is holding (maintained by glibc) - which userspace list
is registered with the kernel via a new syscall [this registration happens at
most once per thread lifetime].  At do_exit() time, the kernel checks this
user-space list: are there any robust futex locks to be cleaned up?

In the common case, at do_exit() time, there is no list registered, so the
cost of robust futexes is just a simple current->robust_list != NULL
comparison.  If the thread has registered a list, then normally the list is
empty.  If the thread/process crashed or terminated in some incorrect way then
the list might be non-empty: in this case the kernel carefully walks the list
[not trusting it], and marks all locks that are owned by this thread with the
FUTEX_OWNER_DEAD bit, and wakes up one waiter (if any).

The list is guaranteed to be private and per-thread, so it's lockless.  There
is one race possible though: since adding to and removing from the list is
done after the futex is acquired by glibc, there is a few instructions window
for the thread (or process) to die there, leaving the futex hung.  To protect
against this possibility, userspace (glibc) also maintains a simple per-thread
'list_op_pending' field, to allow the kernel to clean up if the thread dies
after acquiring the lock, but just before it could have added itself to the
list.  Glibc sets this list_op_pending field before it tries to acquire the
futex, and clears it after the list-add (or list-remove) has finished.

That's all that is needed - all the rest of robust-futex cleanup is done in
userspace [just like with the previous patches].

Ulrich Drepper has implemented the necessary glibc support for this new
mechanism, which fully enables robust mutexes.  (Ulrich plans to commit these
changes to glibc-HEAD later today.)

Key differences of this userspace-list based approach, compared to the vma
based method:

 - it's much, much faster: at thread exit time, there's no need to loop
   over every vma (!), which the VM-based method has to do.  Only a very
   simple 'is the list empty' op is done.

 - no VM changes are needed - 'struct address_space' is left alone.

 - no registration of individual locks is needed: robust mutexes dont need
   any extra per-lock syscalls.  Robust mutexes thus become a very lightweight
   primitive - so they dont force the application designer to do a hard choice
   between performance and robustness - robust mutexes are just as fast.

 - no per-lock kernel allocation happens.

 - no resource limits are needed.

 - no kernel-space recovery call (FUTEX_RECOVER) is needed.

 - the implementation and the locking is "obvious", and there are no
   interactions with the VM.

  Performance
  -----------

I have benchmarked the time needed for the kernel to process a list of 1
million (!) held locks, using the new method [on a 2GHz CPU]:

 - with FUTEX_WAIT set [contended mutex]: 130 msecs
 - without FUTEX_WAIT set [uncontended mutex]: 30 msecs

I have also measured an approach where glibc does the lock notification [which
it currently does for !pshared robust mutexes], and that took 256 msecs -
clearly slower, due to the 1 million FUTEX_WAKE syscalls userspace had to do.

(1 million held locks are unheard of - we expect at most a handful of locks to
be held at a time.  Nevertheless it's nice to know that this approach scales
nicely.)

  Implementation details
  ----------------------

The patch adds two new syscalls: one to register the userspace list, and one
to query the registered list pointer:

 asmlinkage long
 sys_set_robust_list(struct robust_list_head __user *head,
                     size_t len);

 asmlinkage long
 sys_get_robust_list(int pid, struct robust_list_head __user **head_ptr,
                     size_t __user *len_ptr);

List registration is very fast: the pointer is simply stored in
current->robust_list.  [Note that in the future, if robust futexes become
widespread, we could extend sys_clone() to register a robust-list head for new
threads, without the need of another syscall.]

So there is virtually zero overhead for tasks not using robust futexes, and
even for robust futex users, there is only one extra syscall per thread
lifetime, and the cleanup operation, if it happens, is fast and
straightforward.  The kernel doesnt have any internal distinction between
robust and normal futexes.

If a futex is found to be held at exit time, the kernel sets the highest bit
of the futex word:

#define FUTEX_OWNER_DIED        0x40000000

and wakes up the next futex waiter (if any). User-space does the rest of
the cleanup.

Otherwise, robust futexes are acquired by glibc by putting the TID into the
futex field atomically.  Waiters set the FUTEX_WAITERS bit:

#define FUTEX_WAITERS           0x80000000

and the remaining bits are for the TID.

  Testing, architecture support
  -----------------------------

I've tested the new syscalls on x86 and x86_64, and have made sure the parsing
of the userspace list is robust [ ;-) ] even if the list is deliberately
corrupted.

i386 and x86_64 syscalls are wired up at the moment, and Ulrich has tested the
new glibc code (on x86_64 and i386), and it works for his robust-mutex
testcases.

All other architectures should build just fine too - but they wont have the
new syscalls yet.

Architectures need to implement the new futex_atomic_cmpxchg_inuser() inline
function before writing up the syscalls (that function returns -ENOSYS right
now).

This patch:

Add placeholder futex_atomic_cmpxchg_inuser() implementations to every
architecture that supports futexes.  It returns -ENOSYS.

Signed-off-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Arjan van de Ven <arjan@infradead.org>
Acked-by: Ulrich Drepper <drepper@redhat.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] mips: add ptr_to_compat()
Ingo Molnar [Mon, 27 Mar 2006 09:16:13 +0000 (01:16 -0800)]
[PATCH] mips: add ptr_to_compat()

Add ptr_to_compat() - needed by the new robust futex code.

Signed-off-by: Ingo Molnar <mingo@elte.hu>
Cc: Ralf Baechle <ralf@linux-mips.org>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] parisc: add ptr_to_compat()
Ingo Molnar [Mon, 27 Mar 2006 09:16:12 +0000 (01:16 -0800)]
[PATCH] parisc: add ptr_to_compat()

Add ptr_to_compat() to parisc - needed by the new robust futex code.

Signed-off-by: Ingo Molnar <mingo@elte.hu>
Cc: Kyle McMartin <kyle@mcmartin.ca>
Cc: Grant Grundler <iod00d@hp.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] s390: add ptr_to_compat()
Ingo Molnar [Mon, 27 Mar 2006 09:16:09 +0000 (01:16 -0800)]
[PATCH] s390: add ptr_to_compat()

Add ptr_to_compat() to s390 - needed by the new robust-futex code.

Signed-off-by: Ingo Molnar <mingo@elte.hu>
Cc: Heiko Carstens <heiko.carstens@de.ibm.com>
Cc: Martin Schwidefsky <schwidefsky@de.ibm.com>
untested. CHECKME: am i right about the 0x7fffffffUL masking?

Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] ia64: add ptr_to_compat()
Ingo Molnar [Mon, 27 Mar 2006 09:16:06 +0000 (01:16 -0800)]
[PATCH] ia64: add ptr_to_compat()

Add ptr_to_compat() to ia64 - needed by the robust-futex code.

Signed-off-by: Ingo Molnar <mingo@elte.hu>
Cc: "Luck, Tony" <tony.luck@intel.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] unify PFN_* macros
Dave Hansen [Mon, 27 Mar 2006 09:16:04 +0000 (01:16 -0800)]
[PATCH] unify PFN_* macros

Just about every architecture defines some macros to do operations on pfns.
 They're all virtually identical.  This patch consolidates all of them.

One minor glitch is that at least i386 uses them in a very skeletal header
file.  To keep away from #include dependency hell, I stuck the new
definitions in a new, isolated header.

Of all of the implementations, sh64 is the only one that varied by a bit.
It used some masks to ensure that any sign-extension got ripped away before
the arithmetic is done.  This has been posted to that sh64 maintainers and
the development list.

Compiles on x86, x86_64, ia64 and ppc64.

Signed-off-by: Dave Hansen <haveblue@us.ibm.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] uninline zone helpers
KAMEZAWA Hiroyuki [Mon, 27 Mar 2006 09:16:02 +0000 (01:16 -0800)]
[PATCH] uninline zone helpers

Helper functions for for_each_online_pgdat/for_each_zone look too big to be
inlined.  Speed of these helper macro itself is not very important.  (inner
loops are tend to do more work than this)

This patch make helper function to be out-of-lined.

inline out-of-line
.text   005c0680        005bf6a0

005c0680 - 005bf6a0 = FE0 = 4Kbytes.

Signed-off-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] for_each_online_pgdat: remove pgdat_list
KAMEZAWA Hiroyuki [Mon, 27 Mar 2006 09:16:01 +0000 (01:16 -0800)]
[PATCH] for_each_online_pgdat: remove pgdat_list

By using for_each_online_pgdat(), pgdat_list is not necessary now.  This patch
removes it.

Signed-off-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] for_each_online_pgdat: remove sorting pgdat
KAMEZAWA Hiroyuki [Mon, 27 Mar 2006 09:16:00 +0000 (01:16 -0800)]
[PATCH] for_each_online_pgdat: remove sorting pgdat

Because pgdat_list was linked to pgdat_list in *reverse* order, (By default)
some of arch has to sort it by themselves.

for_each_pgdat has gone..for_each_online_pgdat() uses node_online_map, which
doesn't need to be sorted.

This patch removes codes for sorting pgdat.

Signed-off-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] for_each_online_pgdat: renaming for_each_pgdat
KAMEZAWA Hiroyuki [Mon, 27 Mar 2006 09:15:59 +0000 (01:15 -0800)]
[PATCH] for_each_online_pgdat: renaming for_each_pgdat

Replace for_each_pgdat() with for_each_online_pgdat().

Signed-off-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] for_each_online_pgdat: for_each_bootmem
KAMEZAWA Hiroyuki [Mon, 27 Mar 2006 09:15:58 +0000 (01:15 -0800)]
[PATCH] for_each_online_pgdat: for_each_bootmem

Add a list_head to bootmem_data_t and make bootmems use it.  bootmem list is
sorted by node_boot_start.

Only nodes against which init_bootmem() is called are linked to the list.
(i386 allocates bootmem only from one node(0) not from all online nodes.)

A summary:
 1. for_each_online_pgdat() traverses all *online* nodes.
 2. alloc_bootmem() allocates memory only from initialized-for-bootmem nodes.

Signed-off-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] define for_each_online_pgdat
KAMEZAWA Hiroyuki [Mon, 27 Mar 2006 09:15:57 +0000 (01:15 -0800)]
[PATCH] define for_each_online_pgdat

This patch defines for_each_online_pgdat() as a replacement of
for_each_pgdat()

Now, online nodes are managed by node_online_map.  But for_each_pgdat()
uses pgdat_link to iterate over all nodes(pgdat).  This means management
structure for online pgdat is duplicated.

I think using node_online_map for for_each_pgdat() is simple and sane
rather ather than pgdat_link.  New macro is named as
for_each_online_pgdat().  Following patch will fix callers of
for_each_pgdat().

The bootmem allocater uses for_each_pgdat() before pgdat initialization.  I
don't think it's sane.  Following patch will fix it.

Signed-off-by: Yasunori Goto <y-goto@jp.fujitsu.com>
Signed-off-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] remove zone_mem_map
KAMEZAWA Hiroyuki [Mon, 27 Mar 2006 09:15:55 +0000 (01:15 -0800)]
[PATCH] remove zone_mem_map

This patch removes zone_mem_map.

pfn_to_page uses pgdat, page_to_pfn uses zone.  page_to_pfn can use pgdat
instead of zone, which is only one user of zone_mem_map.  By modifing it,
we can remove zone_mem_map.

Signed-off-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>
Cc: Dave Hansen <haveblue@us.ibm.com>
Cc: Christoph Lameter <christoph@lameter.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] unify pfn_to_page: ia64 pfn_to_page
KAMEZAWA Hiroyuki [Mon, 27 Mar 2006 09:15:53 +0000 (01:15 -0800)]
[PATCH] unify pfn_to_page: ia64 pfn_to_page

ia64 has special config CONFIG_VIRTUAL_MEM_MAP.
CONFIG_DISCONTIGMEM=y && CONFIG_VIRTUAL_MEM_MAP!=y is bug ?

Signed-off-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>
Cc: "Luck, Tony" <tony.luck@intel.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] unify pfn_to_page: xtensa pfn_to_page
KAMEZAWA Hiroyuki [Mon, 27 Mar 2006 09:15:52 +0000 (01:15 -0800)]
[PATCH] unify pfn_to_page: xtensa pfn_to_page

xtensa can use generic funcs.

Signed-off-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>
Cc: Chris Zankel <chris@zankel.net>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] unify pfn_to_page: v850 pfn_to_page
KAMEZAWA Hiroyuki [Mon, 27 Mar 2006 09:15:52 +0000 (01:15 -0800)]
[PATCH] unify pfn_to_page: v850 pfn_to_page

v850 can use generic funcs.

Signed-off-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>
Cc: Miles Bader <uclinux-v850@lsi.nec.co.jp>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] unify pfn_to_page: uml pfn_to_page
KAMEZAWA Hiroyuki [Mon, 27 Mar 2006 09:15:50 +0000 (01:15 -0800)]
[PATCH] unify pfn_to_page: uml pfn_to_page

UML can use generic funcs.

Signed-off-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>
Cc: Jeff Dike <jdike@addtoit.com>
Cc: Paolo 'Blaisorblade' Giarrusso <blaisorblade@yahoo.it>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] unify pfn_to_page: sparc pfn_to_page
KAMEZAWA Hiroyuki [Mon, 27 Mar 2006 09:15:48 +0000 (01:15 -0800)]
[PATCH] unify pfn_to_page: sparc pfn_to_page

sparc can use generic funcs.

Signed-off-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>
Cc: William Lee Irwin III <wli@holomorphy.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] unify pfn_to_page: sh64 pfn_to_page
KAMEZAWA Hiroyuki [Mon, 27 Mar 2006 09:15:47 +0000 (01:15 -0800)]
[PATCH] unify pfn_to_page: sh64 pfn_to_page

sh64 can use generic funcs.

Signed-off-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>
Cc: Paul Mundt <lethal@linux-sh.org>
Cc: Richard Curnow <rc@rc0.org.uk>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] unify pfn_to_page: sh pfn_to_page
KAMEZAWA Hiroyuki [Mon, 27 Mar 2006 09:15:46 +0000 (01:15 -0800)]
[PATCH] unify pfn_to_page: sh pfn_to_page

sh can use generic funcs.

Signed-off-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>
Cc: Paul Mundt <lethal@linux-sh.org>
Cc: Kazumoto Kojima <kkojima@rr.iij4u.or.jp>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] unify pfn_to_page: s390 pfn_to_page
KAMEZAWA Hiroyuki [Mon, 27 Mar 2006 09:15:45 +0000 (01:15 -0800)]
[PATCH] unify pfn_to_page: s390 pfn_to_page

s390 can use generic funcs.

Signed-off-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>
Cc: Heiko Carstens <heiko.carstens@de.ibm.com>
Cc: Martin Schwidefsky <schwidefsky@de.ibm.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] unify pfn_to_page: ppc pfn_to_page
KAMEZAWA Hiroyuki [Mon, 27 Mar 2006 09:15:44 +0000 (01:15 -0800)]
[PATCH] unify pfn_to_page: ppc pfn_to_page

PPC can use generic funcs.

Signed-off-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] unify pfn_to_page: parisc pfn_to_page
KAMEZAWA Hiroyuki [Mon, 27 Mar 2006 09:15:43 +0000 (01:15 -0800)]
[PATCH] unify pfn_to_page: parisc pfn_to_page

PARISC can use generic funcs.

Signed-off-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>
Cc: Kyle McMartin <kyle@mcmartin.ca>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] unify pfn_to_page: mips pfn_to_page
KAMEZAWA Hiroyuki [Mon, 27 Mar 2006 09:15:43 +0000 (01:15 -0800)]
[PATCH] unify pfn_to_page: mips pfn_to_page

MIPS can use generic funcs.

Signed-off-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>
Cc: Ralf Baechle <ralf@linux-mips.org>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] unify pfn_to_page: m32r pfn_to_page
KAMEZAWA Hiroyuki [Mon, 27 Mar 2006 09:15:42 +0000 (01:15 -0800)]
[PATCH] unify pfn_to_page: m32r pfn_to_page

m32r can use generic funcs.

Signed-off-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>
Cc: Hirokazu Takata <takata.hirokazu@renesas.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] unify pfn_to_page: h8300 pfn_to_page
KAMEZAWA Hiroyuki [Mon, 27 Mar 2006 09:15:41 +0000 (01:15 -0800)]
[PATCH] unify pfn_to_page: h8300 pfn_to_page

H8300 can use generic funcs.

Signed-off-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>
Cc: Yoshinori Sato <ysato@users.sourceforge.jp>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] unify pfn_to_page: FRV pfn_to_page
KAMEZAWA Hiroyuki [Mon, 27 Mar 2006 09:15:40 +0000 (01:15 -0800)]
[PATCH] unify pfn_to_page: FRV pfn_to_page

FRV can use generic funcs.

Signed-off-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>
Cc: David Howells <dhowells@redhat.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] unify pfn_to_page: cris pfn_to_page
KAMEZAWA Hiroyuki [Mon, 27 Mar 2006 09:15:39 +0000 (01:15 -0800)]
[PATCH] unify pfn_to_page: cris pfn_to_page

cris can use generic funcs.

Signed-off-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>
Cc: Mikael Starvik <starvik@axis.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] unify pfn_to_page: arm26 pfn_to_page
KAMEZAWA Hiroyuki [Mon, 27 Mar 2006 09:15:38 +0000 (01:15 -0800)]
[PATCH] unify pfn_to_page: arm26 pfn_to_page

arm26 can use generic funcs.

Signed-off-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>
Cc: Ian Molton <spyro@f2s.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] unify pfn_to_page: arm pfn_to_page
KAMEZAWA Hiroyuki [Mon, 27 Mar 2006 09:15:37 +0000 (01:15 -0800)]
[PATCH] unify pfn_to_page: arm pfn_to_page

ARM can use generic funcs.
PFN_TO_NID, LOCAL_MAP_NR are defined by sub-archs.

Signed-off-by: KAMEZAWA Hirotuki <kamezawa.hiroyu@jp.fujitsu.com>
Cc: Russell King <rmk@arm.linux.org.uk>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] unify pfn_to_page: alpha pfn_to_page
KAMEZAWA Hiroyuki [Mon, 27 Mar 2006 09:15:36 +0000 (01:15 -0800)]
[PATCH] unify pfn_to_page: alpha pfn_to_page

Alpha can use generic funcs.

Signed-off-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>
Cc: Richard Henderson <rth@twiddle.net>
Cc: Ivan Kokshaysky <ink@jurassic.park.msu.ru>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] unify pfn_to_page: powerpc pfn_to_page
KAMEZAWA Hiroyuki [Mon, 27 Mar 2006 09:15:35 +0000 (01:15 -0800)]
[PATCH] unify pfn_to_page: powerpc pfn_to_page

PowerPC can use generic ones.

Signed-off-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] unify pfn_to_page: x86_64 pfn_to_page
KAMEZAWA Hiroyuki [Mon, 27 Mar 2006 09:15:34 +0000 (01:15 -0800)]
[PATCH] unify pfn_to_page: x86_64 pfn_to_page

x86_64 can use generic funcs.
For DISCONTIGMEM, CONFIG_OUT_OF_LINE_PFN_TO_PAGE is selected.

Signed-off-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>
Cc: Andi Kleen <ak@muc.de>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] unify pfn_to_page: i386 pfn_to_page
KAMEZAWA Hiroyuki [Mon, 27 Mar 2006 09:15:33 +0000 (01:15 -0800)]
[PATCH] unify pfn_to_page: i386 pfn_to_page

i386 can use generic funcs.

Signed-off-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] unify pfn_to_page: generic functions
KAMEZAWA Hiroyuki [Mon, 27 Mar 2006 09:15:25 +0000 (01:15 -0800)]
[PATCH] unify pfn_to_page: generic functions

There are 3 memory models, FLATMEM, DISCONTIGMEM, SPARSEMEM.
Each arch has its own page_to_pfn(), pfn_to_page() for each models.
But most of them can use the same arithmetic.

This patch adds asm-generic/memory_model.h, which includes generic
page_to_pfn(), pfn_to_page() definitions for each memory model.

When CONFIG_OUT_OF_LINE_PFN_TO_PAGE=y, out-of-line functions are
used instead of macro. This is enabled by some archs and  reduces
text size.

Signed-off-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>
Cc: Hugh Dickins <hugh@veritas.com>
Cc: Andi Kleen <ak@muc.de>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Cc: Richard Henderson <rth@twiddle.net>
Cc: Ivan Kokshaysky <ink@jurassic.park.msu.ru>
Cc: Russell King <rmk@arm.linux.org.uk>
Cc: Ian Molton <spyro@f2s.com>
Cc: Mikael Starvik <starvik@axis.com>
Cc: David Howells <dhowells@redhat.com>
Cc: Yoshinori Sato <ysato@users.sourceforge.jp>
Cc: Hirokazu Takata <takata.hirokazu@renesas.com>
Cc: Ralf Baechle <ralf@linux-mips.org>
Cc: Kyle McMartin <kyle@mcmartin.ca>
Cc: Heiko Carstens <heiko.carstens@de.ibm.com>
Cc: Martin Schwidefsky <schwidefsky@de.ibm.com>
Cc: Paul Mundt <lethal@linux-sh.org>
Cc: Kazumoto Kojima <kkojima@rr.iij4u.or.jp>
Cc: Richard Curnow <rc@rc0.org.uk>
Cc: William Lee Irwin III <wli@holomorphy.com>
Cc: "David S. Miller" <davem@davemloft.net>
Cc: Jeff Dike <jdike@addtoit.com>
Cc: Paolo 'Blaisorblade' Giarrusso <blaisorblade@yahoo.it>
Cc: Miles Bader <uclinux-v850@lsi.nec.co.jp>
Cc: Chris Zankel <chris@zankel.net>
Cc: "Luck, Tony" <tony.luck@intel.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] x86: don't use cpuid.2 to determine cache info if cpuid.4 is supported
Shaohua Li [Mon, 27 Mar 2006 09:15:24 +0000 (01:15 -0800)]
[PATCH] x86: don't use cpuid.2 to determine cache info if cpuid.4 is supported

Don't use cpuid.2 to determine cache info if cpuid.4 is supported.  The
exception is P4 trace cache.  We always use cpuid.2 to get trace cache
under P4.

Signed-off-by: Shaohua Li <shaohua.li@intel.com>
Cc: Andi Kleen <ak@muc.de>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] sched: fix group power for allnodes_domains
Siddha, Suresh B [Mon, 27 Mar 2006 09:15:23 +0000 (01:15 -0800)]
[PATCH] sched: fix group power for allnodes_domains

Current sched groups power calculation for allnodes_domains is wrong.  We
should really be using cumulative power of the physical packages in that
group (similar to the calculation in node_domains)

Signed-off-by: Suresh Siddha <suresh.b.siddha@intel.com>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: Nick Piggin <nickpiggin@yahoo.com.au>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] sched: new sched domain for representing multi-core
Siddha, Suresh B [Mon, 27 Mar 2006 09:15:22 +0000 (01:15 -0800)]
[PATCH] sched: new sched domain for representing multi-core

Add a new sched domain for representing multi-core with shared caches
between cores.  Consider a dual package system, each package containing two
cores and with last level cache shared between cores with in a package.  If
there are two runnable processes, with this appended patch those two
processes will be scheduled on different packages.

On such systems, with this patch we have observed 8% perf improvement with
specJBB(2 warehouse) benchmark and 35% improvement with CFP2000 rate(with 2
users).

This new domain will come into play only on multi-core systems with shared
caches.  On other systems, this sched domain will be removed by domain
degeneration code.  This new domain can be also used for implementing power
savings policy (see OLS 2005 CMP kernel scheduler paper for more details..
I will post another patch for power savings policy soon)

Most of the arch/* file changes are for cpu_coregroup_map() implementation.

Signed-off-by: Suresh Siddha <suresh.b.siddha@intel.com>
Cc: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] Small schedule() optimization
Andreas Mohr [Mon, 27 Mar 2006 09:15:20 +0000 (01:15 -0800)]
[PATCH] Small schedule() optimization

small schedule() microoptimization.

Signed-off-by: Andreas Mohr <andi@lisas.de>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] sched: fix task interactivity calculation
Martin Andersson [Mon, 27 Mar 2006 09:15:18 +0000 (01:15 -0800)]
[PATCH] sched: fix task interactivity calculation

Is a truncation error in kernel/sched.c triggered when the nice value is
negative.  The affected code is used in the TASK_INTERACTIVE macro.

The code is:
#define SCALE(v1,v1_max,v2_max) \
(v1) * (v2_max) / (v1_max)

which is used in this way:
SCALE(TASK_NICE(p), 40, MAX_BONUS)

Comments in the code says:
  * This part scales the interactivity limit depending on niceness.
  *
  * We scale it linearly, offset by the INTERACTIVE_DELTA delta.
  * Here are a few examples of different nice levels:
  *
  *  TASK_INTERACTIVE(-20): [1,1,1,1,1,1,1,1,1,0,0]
  *  TASK_INTERACTIVE(-10): [1,1,1,1,1,1,1,0,0,0,0]
  *  TASK_INTERACTIVE(  0): [1,1,1,1,0,0,0,0,0,0,0]
  *  TASK_INTERACTIVE( 10): [1,1,0,0,0,0,0,0,0,0,0]
  *  TASK_INTERACTIVE( 19): [0,0,0,0,0,0,0,0,0,0,0]
  *
  * (the X axis represents the possible -5 ... 0 ... +5 dynamic
  *  priority range a task can explore, a value of '1' means the
  *  task is rated interactive.)

However, the current code does not scale it linearly and the result differs
from the given examples.  If the mathematical function "floor" is used when
the nice value is negative instead of the truncation one gets when using
integer division, the result conforms to the documentation.

Output of TASK_INTERACTIVE when using the kernel code:
nice    dynamic priorities
-20     1     1     1     1     1     1     1     1     1     0     0
-19     1     1     1     1     1     1     1     1     0     0     0
-18     1     1     1     1     1     1     1     1     0     0     0
-17     1     1     1     1     1     1     1     1     0     0     0
-16     1     1     1     1     1     1     1     1     0     0     0
-15     1     1     1     1     1     1     1     0     0     0     0
-14     1     1     1     1     1     1     1     0     0     0     0
-13     1     1     1     1     1     1     1     0     0     0     0
-12     1     1     1     1     1     1     1     0     0     0     0
-11     1     1     1     1     1     1     0     0     0     0     0
-10     1     1     1     1     1     1     0     0     0     0     0
  -9     1     1     1     1     1     1     0     0     0     0     0
  -8     1     1     1     1     1     1     0     0     0     0     0
  -7     1     1     1     1     1     0     0     0     0     0     0
  -6     1     1     1     1     1     0     0     0     0     0     0
  -5     1     1     1     1     1     0     0     0     0     0     0
  -4     1     1     1     1     1     0     0     0     0     0     0
  -3     1     1     1     1     0     0     0     0     0     0     0
  -2     1     1     1     1     0     0     0     0     0     0     0
  -1     1     1     1     1     0     0     0     0     0     0     0
  0      1     1     1     1     0     0     0     0     0     0     0
  1      1     1     1     1     0     0     0     0     0     0     0
  2      1     1     1     1     0     0     0     0     0     0     0
  3      1     1     1     1     0     0     0     0     0     0     0
  4      1     1     1     0     0     0     0     0     0     0     0
  5      1     1     1     0     0     0     0     0     0     0     0
  6      1     1     1     0     0     0     0     0     0     0     0
  7      1     1     1     0     0     0     0     0     0     0     0
  8      1     1     0     0     0     0     0     0     0     0     0
  9      1     1     0     0     0     0     0     0     0     0     0
10      1     1     0     0     0     0     0     0     0     0     0
11      1     1     0     0     0     0     0     0     0     0     0
12      1     0     0     0     0     0     0     0     0     0     0
13      1     0     0     0     0     0     0     0     0     0     0
14      1     0     0     0     0     0     0     0     0     0     0
15      1     0     0     0     0     0     0     0     0     0     0
16      0     0     0     0     0     0     0     0     0     0     0
17      0     0     0     0     0     0     0     0     0     0     0
18      0     0     0     0     0     0     0     0     0     0     0
19      0     0     0     0     0     0     0     0     0     0     0

Output of TASK_INTERACTIVE when using "floor"
nice    dynamic priorities
-20     1     1     1     1     1     1     1     1     1     0     0
-19     1     1     1     1     1     1     1     1     1     0     0
-18     1     1     1     1     1     1     1     1     1     0     0
-17     1     1     1     1     1     1     1     1     1     0     0
-16     1     1     1     1     1     1     1     1     0     0     0
-15     1     1     1     1     1     1     1     1     0     0     0
-14     1     1     1     1     1     1     1     1     0     0     0
-13     1     1     1     1     1     1     1     1     0     0     0
-12     1     1     1     1     1     1     1     0     0     0     0
-11     1     1     1     1     1     1     1     0     0     0     0
-10     1     1     1     1     1     1     1     0     0     0     0
  -9     1     1     1     1     1     1     1     0     0     0     0
  -8     1     1     1     1     1     1     0     0     0     0     0
  -7     1     1     1     1     1     1     0     0     0     0     0
  -6     1     1     1     1     1     1     0     0     0     0     0
  -5     1     1     1     1     1     1     0     0     0     0     0
  -4     1     1     1     1     1     0     0     0     0     0     0
  -3     1     1     1     1     1     0     0     0     0     0     0
  -2     1     1     1     1     1     0     0     0     0     0     0
  -1     1     1     1     1     1     0     0     0     0     0     0
   0     1     1     1     1     0     0     0     0     0     0     0
   1     1     1     1     1     0     0     0     0     0     0     0
   2     1     1     1     1     0     0     0     0     0     0     0
   3     1     1     1     1     0     0     0     0     0     0     0
   4     1     1     1     0     0     0     0     0     0     0     0
   5     1     1     1     0     0     0     0     0     0     0     0
   6     1     1     1     0     0     0     0     0     0     0     0
   7     1     1     1     0     0     0     0     0     0     0     0
   8     1     1     0     0     0     0     0     0     0     0     0
   9     1     1     0     0     0     0     0     0     0     0     0
  10     1     1     0     0     0     0     0     0     0     0     0
  11     1     1     0     0     0     0     0     0     0     0     0
  12     1     0     0     0     0     0     0     0     0     0     0
  13     1     0     0     0     0     0     0     0     0     0     0
  14     1     0     0     0     0     0     0     0     0     0     0
  15     1     0     0     0     0     0     0     0     0     0     0
  16     0     0     0     0     0     0     0     0     0     0     0
  17     0     0     0     0     0     0     0     0     0     0     0
  18     0     0     0     0     0     0     0     0     0     0     0
  19     0     0     0     0     0     0     0     0     0     0     0

Signed-off-by: Martin Andersson <martin.andersson@control.lth.se>
Acked-by: Ingo Molnar <mingo@elte.hu>
Cc: Nick Piggin <nickpiggin@yahoo.com.au>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Peter Williams <pwil3058@bigpond.net.au>
Cc: Con Kolivas <kernel@kolivas.org>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] knfsd: Update rpc-cache.txt to match recent changes
NeilBrown [Mon, 27 Mar 2006 09:15:12 +0000 (01:15 -0800)]
[PATCH] knfsd: Update rpc-cache.txt to match recent changes

Signed-off-by: Neil Brown <neilb@suse.de>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] knfsd: Tidy up unix_domain_find
NeilBrown [Mon, 27 Mar 2006 09:15:11 +0000 (01:15 -0800)]
[PATCH] knfsd: Tidy up unix_domain_find

We shouldn't really compare &new->h with anything when new ==NULL, and gather
three different if statements that all start

  if (rv ...

into one large if.

Signed-off-by: Neil Brown <neilb@suse.de>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] fs/nfsd/export.c,net/sunrpc/cache.c: make needlessly global code static
Adrian Bunk [Mon, 27 Mar 2006 09:15:10 +0000 (01:15 -0800)]
[PATCH] fs/nfsd/export.c,net/sunrpc/cache.c: make needlessly global code static

We can now make some code static.

Signed-off-by: Adrian Bunk <bunk@stusta.de>
Cc: Neil Brown <neilb@suse.de>
Cc: Trond Myklebust <trond.myklebust@fys.uio.no>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] knfsd: Convert sunrpc_cache to use krefs
NeilBrown [Mon, 27 Mar 2006 09:15:09 +0000 (01:15 -0800)]
[PATCH] knfsd: Convert sunrpc_cache to use krefs

.. it makes some of the code nicer.

Signed-off-by: Neil Brown <neilb@suse.de>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] knfsd: Unexport cache_fresh and fix a small race
NeilBrown [Mon, 27 Mar 2006 09:15:08 +0000 (01:15 -0800)]
[PATCH] knfsd: Unexport cache_fresh and fix a small race

Cache_fresh is now only used in cache.c, so unexport it.

Part of cache_fresh (setting CACHE_VALID) should really be done under the
lock, while part (calling cache_revisit_request etc) must be done outside the
lock.  So we split it up appropriately.

Signed-off-by: Neil Brown <neilb@suse.de>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] knfsd: Remove DefineCacheLookup
NeilBrown [Mon, 27 Mar 2006 09:15:07 +0000 (01:15 -0800)]
[PATCH] knfsd: Remove DefineCacheLookup

This has been replaced by more traditional code.

Signed-off-by: Neil Brown <neilb@suse.de>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] knfsd: An assortment of little fixes to the sunrpc cache code
NeilBrown [Mon, 27 Mar 2006 09:15:07 +0000 (01:15 -0800)]
[PATCH] knfsd: An assortment of little fixes to the sunrpc cache code

- in cache_check, h must be non-NULL as it has been de-referenced,
  so don't bother checking for NULL.

- When a cache-item is updated, we need to call cache_revisit_request to see
  if there is a pending request waiting for that item.  We were using
  a transition to CACHE_VALID to see if that was needed, however that is
  wrong as an expired entry will still be marked 'valid' (as the data is valid
  and will need to be released).  So instead use an off transition for
  CACHE_PENDING which is exactly the right thing to test.

- Add a little bit more debugging info.

Signed-off-by: Neil Brown <neilb@suse.de>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] knfsd: Use new cache code for name/id lookup caches
NeilBrown [Mon, 27 Mar 2006 09:15:06 +0000 (01:15 -0800)]
[PATCH] knfsd: Use new cache code for name/id lookup caches

Signed-off-by: Neil Brown <neilb@suse.de>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] knfsd: Use new cache code for rsc cache
NeilBrown [Mon, 27 Mar 2006 09:15:05 +0000 (01:15 -0800)]
[PATCH] knfsd: Use new cache code for rsc cache

Signed-off-by: Neil Brown <neilb@suse.de>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] knfsd: Use new sunrpc cache for rsi cache
NeilBrown [Mon, 27 Mar 2006 09:15:04 +0000 (01:15 -0800)]
[PATCH] knfsd: Use new sunrpc cache for rsi cache

Signed-off-by: Neil Brown <neilb@suse.de>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] knfsd: Use new cache_lookup for svc_expkey cache
NeilBrown [Mon, 27 Mar 2006 09:15:04 +0000 (01:15 -0800)]
[PATCH] knfsd: Use new cache_lookup for svc_expkey cache

Signed-off-by: Neil Brown <neilb@suse.de>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] knfsd: Use new cache_lookup for svc_export
NeilBrown [Mon, 27 Mar 2006 09:15:03 +0000 (01:15 -0800)]
[PATCH] knfsd: Use new cache_lookup for svc_export

Signed-off-by: Neil Brown <neilb@suse.de>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] knfsd: Convert ip_map cache to use the new lookup routine
NeilBrown [Mon, 27 Mar 2006 09:15:02 +0000 (01:15 -0800)]
[PATCH] knfsd: Convert ip_map cache to use the new lookup routine

Signed-off-by: Neil Brown <neilb@suse.de>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] knfsd: Create cache_lookup function instead of using a macro to declare one
NeilBrown [Mon, 27 Mar 2006 09:15:02 +0000 (01:15 -0800)]
[PATCH] knfsd: Create cache_lookup function instead of using a macro to declare one

The C++-like 'template' approach proves to be too ugly and hard to work with.

The old 'template' won't go away until all users are updated.

Signed-off-by: Neil Brown <neilb@suse.de>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] knfsd: Get rid of 'inplace' sunrpc caches
NeilBrown [Mon, 27 Mar 2006 09:15:01 +0000 (01:15 -0800)]
[PATCH] knfsd: Get rid of 'inplace' sunrpc caches

These were an unnecessary wart.  Also only have one 'DefineSimpleCache..'
instead of two.

Signed-off-by: Neil Brown <neilb@suse.de>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] knfsd: Break the hard linkage from svc_expkey to svc_export
NeilBrown [Mon, 27 Mar 2006 09:15:00 +0000 (01:15 -0800)]
[PATCH] knfsd: Break the hard linkage from svc_expkey to svc_export

Current svc_expkey holds a pointer to the svc_export structure, so updates to
that structure have to be in-place, which is a wart on the whole cache
infrastruct.  So we break that linkage and just do a second lookup.

If this became a performance issue, it would be possible to put a direct link
back in which was only used conditionally.  i.e.  when an object is replaced
in the cache, we set a flag in the old object.  When dereferencing the link
from svc_expkey, if the flag is set, we drop the reference and do a fresh
lookup.

Signed-off-by: Neil Brown <neilb@suse.de>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] knfsd: Change the store of auth_domains to not be a 'cache'
NeilBrown [Mon, 27 Mar 2006 09:14:59 +0000 (01:14 -0800)]
[PATCH] knfsd: Change the store of auth_domains to not be a 'cache'

The 'auth_domain's are simply handles on internal data structures.  They do
not cache information from user-space, and forcing them into the mold of a
'cache' misrepresents their true nature and causes confusion.

Signed-off-by: Neil Brown <neilb@suse.de>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] autofs4: atomic var underflow
Ian Kent [Mon, 27 Mar 2006 09:14:59 +0000 (01:14 -0800)]
[PATCH] autofs4: atomic var underflow

Fix accidental underflow of the atomic counter.

Signed-off-by: Ian Kent <raven@themaw.net>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] autofs4: follow_link missing functionality
Ian Kent [Mon, 27 Mar 2006 09:14:58 +0000 (01:14 -0800)]
[PATCH] autofs4: follow_link missing functionality

This functionality is also need for operation of autofs v5 direct mounts.

Signed-off-by: Ian Kent <raven@themaw.net>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] Remove redundant check from autofs4_put_super
Dave Jones [Mon, 27 Mar 2006 09:14:57 +0000 (01:14 -0800)]
[PATCH] Remove redundant check from autofs4_put_super

We have to have a valid sbi here, or we'd have oopsed already.  (There's a
dereference of sbi->catatonic a few lines above)

Coverity #740

Signed-off-by: Dave Jones <davej@redhat.com>
Cc: Ian Kent <raven@themaw.net>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] autofs4: change AUTOFS_TYP_* AUTOFS_TYPE_*
Ian Kent [Mon, 27 Mar 2006 09:14:56 +0000 (01:14 -0800)]
[PATCH] autofs4: change AUTOFS_TYP_* AUTOFS_TYPE_*

Signed-off-by: Ian Kent <raven@themaw.net>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] autofs4: add new packet type for v5 communications
Ian Kent [Mon, 27 Mar 2006 09:14:55 +0000 (01:14 -0800)]
[PATCH] autofs4: add new packet type for v5 communications

This patch define a new autofs packet for autofs v5 and updates the waitq.c
functions to handle the additional packet type.

Signed-off-by: Ian Kent <raven@themaw.net>
Cc: Al Viro <viro@ftp.linux.org.uk>
Cc: Christoph Hellwig <hch@lst.de>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] autofs4: add v5 expire logic
Ian Kent [Mon, 27 Mar 2006 09:14:55 +0000 (01:14 -0800)]
[PATCH] autofs4: add v5 expire logic

This patch adds expire logic for autofs direct mounts.

Signed-off-by: Ian Kent <raven@themaw.net>
Cc: Al Viro <viro@ftp.linux.org.uk>
Cc: Christoph Hellwig <hch@lst.de>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] autofs4: add v5 follow_link mount trigger method
Ian Kent [Mon, 27 Mar 2006 09:14:54 +0000 (01:14 -0800)]
[PATCH] autofs4: add v5 follow_link mount trigger method

This patch adds a follow_link inode method for the root of an autofs direct
mount trigger.  It also adds the corresponding mount options and updates the
show_mount method.

Signed-off-by: Ian Kent <raven@themaw.net>
Cc: Al Viro <viro@ftp.linux.org.uk>
Cc: Christoph Hellwig <hch@lst.de>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] autofs4: nameidata needs to be up to date for follow_link
Ian Kent [Mon, 27 Mar 2006 09:14:53 +0000 (01:14 -0800)]
[PATCH] autofs4: nameidata needs to be up to date for follow_link

In order to be able to trigger a mount using the follow_link inode method the
nameidata struct that is passed in needs to have the vfsmount of the autofs
trigger not its parent.

During a path walk if an autofs trigger is mounted on a dentry, when the
follow_link method is called, the nameidata struct contains the vfsmount and
mountpoint dentry of the parent mount while the dentry that is passed in is
the root of the autofs trigger mount.  I believe it is impossible to get the
vfsmount of the trigger mount, within the follow_link method, when only the
parent vfsmount and the root dentry of the trigger mount are known.

This patch updates the nameidata struct on entry to __do_follow_link if it
detects that it is out of date.  It moves the path_to_nameidata to above
__do_follow_link to facilitate calling it from there.  The dput_path is moved
as well as that seemed sensible.  No changes are made to these two functions.

Signed-off-by: Ian Kent <raven@themaw.net>
Cc: Al Viro <viro@ftp.linux.org.uk>
Cc: Christoph Hellwig <hch@lst.de>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] autofs4: increase module version
Ian Kent [Mon, 27 Mar 2006 09:14:52 +0000 (01:14 -0800)]
[PATCH] autofs4: increase module version

Update autofs4 version.

Signed-off-by: Ian Kent <raven@themaw.net>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] autofs4: change may_umount* functions to boolean
Ian Kent [Mon, 27 Mar 2006 09:14:51 +0000 (01:14 -0800)]
[PATCH] autofs4: change may_umount* functions to boolean

Change the functions may_umount and may_umount_tree to boolean functions to
aid code readability.

Signed-off-by: Ian Kent <raven@themaw.net>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] autofs4: rename simple_empty_nolock function
Ian Kent [Mon, 27 Mar 2006 09:14:50 +0000 (01:14 -0800)]
[PATCH] autofs4: rename simple_empty_nolock function

Rename the function simple_empty_nolock to __simple_empty in line with kernel
naming conventions.

Signed-off-by: Ian Kent <raven@themaw.net>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] autofs4: white space cleanup for waitq.c
Ian Kent [Mon, 27 Mar 2006 09:14:49 +0000 (01:14 -0800)]
[PATCH] autofs4: white space cleanup for waitq.c

Whitespace and formating changes to waitq code.

Signed-off-by: Ian Kent <raven@themaw.net>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] autofs4: add a show mount options for proc filesystem
Ian Kent [Mon, 27 Mar 2006 09:14:49 +0000 (01:14 -0800)]
[PATCH] autofs4: add a show mount options for proc filesystem

Add show_options method to display autofs4 mount options in the proc
filesystem.

Signed-off-by: Ian Kent <raven@themaw.net>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] autofs4: remove update_atime unused function
Ian Kent [Mon, 27 Mar 2006 09:14:48 +0000 (01:14 -0800)]
[PATCH] autofs4: remove update_atime unused function

Remove the update of i_atime from autofs4 in favour of having VFS update it.
i_atime is never used for expire in autofs4.

Signed-off-by: Ian Kent <raven@themaw.net>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] autofs4: expire mounts that hold no (extra) references only
Ian Kent [Mon, 27 Mar 2006 09:14:47 +0000 (01:14 -0800)]
[PATCH] autofs4: expire mounts that hold no (extra) references only

Alter the expire semantics that define how "busyness" is determined.
Currently a last_used counter is updated on every revalidate from processes
other than the mount owner process group.

This patch changes that so that an expire candidate is busy only if it has a
reference count greater than the expected minimum, such as when there is an
open file or working directory in use.

This method is the only way that busyness can be established for direct mounts
within the new implementation.  For consistency the expire semantic is made
the same for all mounts.

A side effect of the patch is that mounts which remain mounted unessessarily
in the presence of some GUI programs that scan the filesystem should now
expire.

Signed-off-by: Ian Kent <raven@themaw.net>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] autofs4: fix false negative return from expire
Ian Kent [Mon, 27 Mar 2006 09:14:46 +0000 (01:14 -0800)]
[PATCH] autofs4: fix false negative return from expire

Fix the case where an expire returns busy on a tree mount when it is in fact
not busy.  This case was overlooked when the patch to prevent the expiring
away of "scaffolding" directories for tree mounts was applied.

The problem arises when a tree of mounts is a member of a map with other keys.
 The current logic will not expire the tree if any other mount in the map is
busy.  The solution is to maintain a "minimum" use count for each autofs
dentry and compare this to the actual dentry usage count during expire.

Signed-off-by: Ian Kent <raven@themaw.net>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] autofs4: simplify expire tree traversal
Ian Kent [Mon, 27 Mar 2006 09:14:45 +0000 (01:14 -0800)]
[PATCH] autofs4: simplify expire tree traversal

Simplify the expire tree traversal code by using a function from namespace.c
to calculate the next entry in the top down tree traversals carried out during
the expire operation.

Signed-off-by: Ian Kent <raven@themaw.net>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] autofs4: expire code readability cleanup
Ian Kent [Mon, 27 Mar 2006 09:14:44 +0000 (01:14 -0800)]
[PATCH] autofs4: expire code readability cleanup

Change the names of the boolean functions autofs4_check_mount and
autofs4_check_tree to autofs4_mount_busy and autofs4_tree_busy respectively
and alters their return codes to suit in order to aid code readabilty.

A couple of white space cleanups are included as well.

Signed-off-by: Ian Kent <raven@themaw.net>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] autofs4: can't mount due to mount point dir not empty
Ian Kent [Mon, 27 Mar 2006 09:14:44 +0000 (01:14 -0800)]
[PATCH] autofs4: can't mount due to mount point dir not empty

Addresse a problem where stale dentrys stop mounts from happening.

When a mount point directory is pre-created and a non-existent entry within it
is requested a dentry ends up being created within the mount point directory
which stops future mounts.  The problem is solved by ignoring negative,
unhashed dentrys in the mount point d_subdirs list.

Additionally the apparent cacheing of -ENOENT returns from requests is
removed.  The test on d_time is a tautology and d_time is not initialised and
has an unexpected value.  In short it doesn't do what it's meant to.

The cacheing of failed requests to the daemon is important and will be
followed up later.

Signed-off-by: Ian Kent <raven@themaw.net>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>