sfrench/cifs-2.6.git
17 years agoInput: make input_register_handler() return error codes
Dmitry Torokhov [Thu, 14 Sep 2006 05:32:39 +0000 (01:32 -0400)]
Input: make input_register_handler() return error codes

Signed-off-by: Dmitry Torokhov <dtor@mail.ru>
17 years agoInput: remove cruft that was needed for transition to sysfs
Dmitry Torokhov [Thu, 14 Sep 2006 05:32:28 +0000 (01:32 -0400)]
Input: remove cruft that was needed for transition to sysfs

Signed-off-by: Dmitry Torokhov <dtor@mail.ru>
17 years agoInput: fix input module refcounting
Dmitry Torokhov [Thu, 14 Sep 2006 05:32:14 +0000 (01:32 -0400)]
Input: fix input module refcounting

Now that input_free_device is basically an alias for input_put_device
we need to acquire a reference to input module right when we allocate
device because input_dev_release releases reference to input module
unconditionally.

Signed-off-by: Dmitry Torokhov <dtor@mail.ru>
17 years agoInput: constify input core
Dmitry Torokhov [Thu, 14 Sep 2006 05:31:59 +0000 (01:31 -0400)]
Input: constify input core

Signed-off-by: Dmitry Torokhov <dtor@mail.ru>
17 years agoInput: libps2 - rearrange exports
Dmitry Torokhov [Thu, 14 Sep 2006 05:31:40 +0000 (01:31 -0400)]
Input: libps2 - rearrange exports

The new way is to mark function as exported right after its definition.

Signed-off-by: Dmitry Torokhov <dtor@mail.ru>
17 years agoInput: atkbd - support Microsoft Natural Elite Pro keyboards
Dmitry Torokhov [Thu, 14 Sep 2006 05:31:27 +0000 (01:31 -0400)]
Input: atkbd - support Microsoft Natural Elite Pro keyboards

Microsoft Natural Elite Pro keyboard produces unisual response to
the GET ID command - single byte 0xaa (normally keyboards produce
2-byte response). Fail GET ID command so atkbd gets a change to
do alternate probe.

Signed-off-by: Dmitry Torokhov <dtor@mail.ru>
17 years agoInput: i8042 - disable MUX mode on Toshiba Equium A110
Dmitry Torokhov [Thu, 14 Sep 2006 05:31:06 +0000 (01:31 -0400)]
Input: i8042 - disable MUX mode on Toshiba Equium A110

When keyboard controller is in active multiplexing mode ALPS touchpad
may get detected twice. Since the box does not have external PS/2
ports simply disabling MUX mode is safe solution.

Signed-off-by: Dmitry Torokhov <dtor@mail.ru>
17 years agoInput: i8042 - get rid of polling timer
Dmitry Torokhov [Mon, 11 Sep 2006 01:57:21 +0000 (21:57 -0400)]
Input: i8042 - get rid of polling timer

Remove polling timer that was used to detect keybord/mice hotplug and
register both IRQs right away instead of waiting for a driver to
attach to a port.

Signed-off-by: Dmitry Torokhov <dtor@mail.ru>
17 years agoInput: send key up events at disconnect
Dmitry Torokhov [Mon, 11 Sep 2006 01:56:06 +0000 (21:56 -0400)]
Input: send key up events at disconnect

Emit key up events for all pressed keys/buttons when disconnecting
an input device. Cures "stuck" enter key effect when unloading
keyboard module.

Signed-off-by: Dmitry Torokhov <dtor@mail.ru>
17 years agoInput: constify psmouse driver
Helge Deller [Mon, 11 Sep 2006 01:54:39 +0000 (21:54 -0400)]
Input: constify psmouse driver

Signed-off-by: Helge Deller <deller@gmx.de>
Signed-off-by: Dmitry Torokhov <dtor@mail.ru>
17 years agoInput: i8042 - add Amoi to the MUX blacklist
Cjacker Huang [Wed, 9 Aug 2006 03:38:22 +0000 (23:38 -0400)]
Input: i8042 - add Amoi to the MUX blacklist

ALPS touchpad on Amoi laptops (Amoi is a vendor in China) is not
detected when keyboard controller is in MUX mode, add to blacklist.

Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Dmitry Torokhov <dtor@mail.ru>
17 years agoInput: logips2pp - add sugnature 56 (Cordless MouseMan Wheel), cleanup
Helge Deller [Wed, 9 Aug 2006 03:37:18 +0000 (23:37 -0400)]
Input: logips2pp - add sugnature 56 (Cordless MouseMan Wheel), cleanup

Signed-off-by: Helge Deller <deller@gmx.de>
Signed-off-by: Dmitry Torokhov <dtor@mail.ru>
17 years agoInput: add driver for Touchwin serial touchscreens
Rick Koch [Sat, 5 Aug 2006 04:32:30 +0000 (00:32 -0400)]
Input: add driver for Touchwin serial touchscreens

Signed-off-by: Rick Koch <n1gp@hotmail.com>
Signed-off-by: Dmitry Torokhov <dtor@mail.ru>
17 years agoInput: add driver for Touchright serial touchscreens
Rick Koch [Sat, 5 Aug 2006 04:32:24 +0000 (00:32 -0400)]
Input: add driver for Touchright serial touchscreens

Signed-off-by: Rick Koch <n1gp@hotmail.com>
Signed-off-by: Dmitry Torokhov <dtor@mail.ru>
17 years agoInput: add driver for Penmount serial touchscreens
Rick Koch [Sat, 5 Aug 2006 04:32:18 +0000 (00:32 -0400)]
Input: add driver for Penmount serial touchscreens

Signed-off-by: Rick Koch <n1gp@hotmail.com>
Signed-off-by: Dmitry Torokhov <dtor@mail.ru>
17 years agoInput: elo - add support for non-pressure-sensitive touchscreens
Shaun Jackman [Sat, 5 Aug 2006 04:29:49 +0000 (00:29 -0400)]
Input: elo - add support for non-pressure-sensitive touchscreens

- Use the touch status bit rather than the pressure bits to
  distinguish a BTN_TOUCH event. Non-pressure-sensitive touchscreens
  always report full pressure
- Report ABS_PRESSURE information only if the touchscreen supports it

Signed-off-by: Shaun Jackman <sjackman@gmail.com>
Signed-off-by: Dmitry Torokhov <dtor@mail.ru>
17 years agoInput: elo - fix checksum calculation
Shaun Jackman [Sat, 5 Aug 2006 04:27:59 +0000 (00:27 -0400)]
Input: elo - fix checksum calculation

Fix 10-byte protocol checksum calculation and do not discard packet early
unless it is missing lead in byte.

Signed-off-by: Shaun Jackman <sjackman@gmail.com>
Signed-off-by: Dmitry Torokhov <dtor@mail.ru>
17 years agoInput: elo - handle input_register_device() failures
Dmitry Torokhov [Sat, 5 Aug 2006 04:27:00 +0000 (00:27 -0400)]
Input: elo - handle input_register_device() failures

Signed-off-by: Dmitry Torokhov <dtor@mail.ru>
17 years agoInput: hid - #if 0 the no longer used hid_find_field_by_usage()
Adrian Bunk [Sat, 5 Aug 2006 03:00:02 +0000 (23:00 -0400)]
Input: hid - #if 0 the no longer used hid_find_field_by_usage()

Signed-off-by: Adrian Bunk <bunk@stusta.de>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Dmitry Torokhov <dtor@mail.ru>
17 years agoInput: update the force feedback documentation
Anssi Hannula [Wed, 19 Jul 2006 05:44:22 +0000 (01:44 -0400)]
Input: update the force feedback documentation

Signed-off-by: Anssi Hannula <anssi.hannula@gmail.com>
Signed-off-by: Dmitry Torokhov <dtor@mail.ru>
17 years agoInput: add force feedback driver for PSX-style Zeroplus devices
Anssi Hannula [Wed, 19 Jul 2006 05:44:17 +0000 (01:44 -0400)]
Input: add force feedback driver for PSX-style Zeroplus devices

Signed-off-by: Anssi Hannula <anssi.hannula@gmail.com>
Signed-off-by: Dmitry Torokhov <dtor@mail.ru>
17 years agoInput: drop remnants of the old force-feedback interface
Anssi Hannula [Wed, 19 Jul 2006 05:44:08 +0000 (01:44 -0400)]
Input: drop remnants of the old force-feedback interface

Signed-off-by: Anssi Hannula <anssi.hannula@gmail.com>
Signed-off-by: Dmitry Torokhov <dtor@mail.ru>
17 years agoInput: uinput - switch to the new FF interface
Anssi Hannula [Wed, 19 Jul 2006 05:41:09 +0000 (01:41 -0400)]
Input: uinput - switch to the new FF interface

The userspace interface of the force feedback part is changed and
documentation in uinput.h is updated accordingly. MODULE_VERSION
is also incremented to reflect the revision.

Signed-off-by: Anssi Hannula <anssi.hannula@gmail.com>
Signed-off-by: Dmitry Torokhov <dtor@mail.ru>
17 years agoInput: use new FF interface in the HID force feedback drivers
Anssi Hannula [Wed, 19 Jul 2006 05:40:55 +0000 (01:40 -0400)]
Input: use new FF interface in the HID force feedback drivers

Signed-off-by: Anssi Hannula <anssi.hannula@gmail.com>
Signed-off-by: Dmitry Torokhov <dtor@mail.ru>
17 years agoInput: add force feedback driver for PID devices
Anssi Hannula [Wed, 19 Jul 2006 05:40:47 +0000 (01:40 -0400)]
Input: add force feedback driver for PID devices

This replaces the older PID driver which was never completed.

Signed-off-by: Anssi Hannula <anssi.hannula@gmail.com>
Signed-off-by: Dmitry Torokhov <dtor@mail.ru>
17 years agoInput: iforce - switch to the new FF interface
Anssi Hannula [Wed, 19 Jul 2006 05:40:39 +0000 (01:40 -0400)]
Input: iforce - switch to the new FF interface

Signed-off-by: Anssi Hannula <anssi.hannula@gmail.com>
Signed-off-by: Dmitry Torokhov <dtor@mail.ru>
17 years agoInput: unified force feedback support for memoryless devices
Anssi Hannula [Wed, 19 Jul 2006 05:40:30 +0000 (01:40 -0400)]
Input: unified force feedback support for memoryless devices

Consolidate core implementing memoryless devices in one module; added
support for gain and envelopes and periodic => rumble conversion.

Signed-off-by: Anssi Hannula <anssi.hannula@gmail.com>
Signed-off-by: Dmitry Torokhov <dtor@mail.ru>
17 years agoInput: implement new force feedback interface
Anssi Hannula [Wed, 19 Jul 2006 05:40:22 +0000 (01:40 -0400)]
Input: implement new force feedback interface

Implement a new force feedback interface, in which all non-driver-specific
operations are separated to a common module. This includes handling effect
type validations, locking, etc.

The effects are now file descriptor specific instead of the previous strange
half-process half-fd specific behaviour. The effect memory of devices is not
emptied if the root user opens and closes the device while another user is
using effects. This is a minor change and most likely no force feedback
aware programs are affected by this negatively.

Otherwise the userspace interface is left unaltered.

Signed-off-by: Anssi Hannula <anssi.hannula@gmail.com>
Signed-off-by: Dmitry Torokhov <dtor@mail.ru>
17 years agoInput: move fixp-arith.h to drivers/input
Anssi Hannula [Wed, 19 Jul 2006 05:40:14 +0000 (01:40 -0400)]
Input: move fixp-arith.h to drivers/input

Move fixp-arith.h from drivers/usb/input to drivers/input, as the part of
force feedback support that requires trigonometric functions is being moved
there.

Signed-off-by: Anssi Hannula <anssi.hannula@gmail.com>
Signed-off-by: Dmitry Torokhov <dtor@mail.ru>
17 years agoInput: rename input.ko into input-core.ko
Dmitry Torokhov [Wed, 19 Jul 2006 05:39:56 +0000 (01:39 -0400)]
Input: rename input.ko into input-core.ko

This will allow building input core module from several files
which is needed for the reworked force feedback support.

Signed-off-by: Dmitry Torokhov <dtor@mail.ru>
17 years agoInput: serio/gameport - check whether driver core calls succeeded
Randy Dunlap [Wed, 19 Jul 2006 05:14:55 +0000 (01:14 -0400)]
Input: serio/gameport - check whether driver core calls succeeded

Signed-off-by: Randy Dunlap <rdunlap@xenotime.net>
Signed-off-by: Dmitry Torokhov <dtor@mail.ru>
17 years agoInput: spaceball - make 4000FLX Lefty work
Nick Martin [Wed, 19 Jul 2006 05:14:44 +0000 (01:14 -0400)]
Input: spaceball - make 4000FLX Lefty work

Although the Spaceball 4000FLX Lefty is already supported by the
spaceball driver, it does not register properly due to SPACEBALL_MAX_ID
being set too low. Increment SPACEBALL_MAX_ID such that the 4000FLX Lefty
is properly recognized.

Signed-off-by: Nick Martin <nim+linux@nimlabs.org>
Signed-off-by: Dmitry Torokhov <dtor@mail.ru>
17 years agoInput: keyboard - simplify emulate_raw() implementation
Dmitry Torokhov [Wed, 19 Jul 2006 05:14:25 +0000 (01:14 -0400)]
Input: keyboard - simplify emulate_raw() implementation

Signed-off-by: Dmitry Torokhov <dtor@mail.ru>
17 years agoInput: keyboard - remove static variable and clean up initialization
Dmitry Torokhov [Wed, 19 Jul 2006 05:13:26 +0000 (01:13 -0400)]
Input: keyboard - remove static variable and clean up initialization

Signed-off-by: Dmitry Torokhov <dtor@mail.ru>
17 years agoInput: hiddev - use standard list implementation
Dmitry Torokhov [Wed, 19 Jul 2006 05:09:10 +0000 (01:09 -0400)]
Input: hiddev - use standard list implementation

Fixes Coverity #id 303

Signed-off-by: Dmitry Torokhov <dtor@mail.ru>
17 years agoInput: add missing handler->start() call
Dmitry Torokhov [Wed, 19 Jul 2006 05:08:51 +0000 (01:08 -0400)]
Input: add missing handler->start() call

The start() method need to be called every time we create
a new handle. This includes not only registering new devices
but also when registering new handlers.

Signed-off-by: Dmitry Torokhov <dtor@mail.ru>
17 years agoInput: HID - fix potential out-of-bound array access
Dmitry Torokhov [Sat, 15 Jul 2006 05:17:54 +0000 (01:17 -0400)]
Input: HID - fix potential out-of-bound array access

Fixes Coverity #id 978

Signed-off-by: Dmitry Torokhov <dtor@mail.ru>
17 years agoInput: fix list iteration in input_release_device()
Andrew Morton [Sat, 15 Jul 2006 05:17:38 +0000 (01:17 -0400)]
Input: fix list iteration in input_release_device()

Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Dmitry Torokhov <dtor@mail.ru>
17 years agoInput: iforce - add Trust Force Feedback Race Master support
Przemek Iskra [Sat, 15 Jul 2006 05:16:21 +0000 (01:16 -0400)]
Input: iforce - add Trust Force Feedback Race Master support

Signed-off-by: Vojtech Pavlik <vojtech@suse.cz>
Signed-off-by: Dmitry Torokhov <dtor@mail.ru>
17 years agoInput: iforce - check array bounds before accessing elements
Dmitry Torokhov [Fri, 7 Jul 2006 03:55:00 +0000 (23:55 -0400)]
Input: iforce - check array bounds before accessing elements

Fixes Coverity #id 864

Signed-off-by: Dmitry Torokhov <dtor@mail.ru>
17 years agoInput: libps2 - warn instead of oopsing when passed bad arguments
Dmitry Torokhov [Fri, 7 Jul 2006 03:54:48 +0000 (23:54 -0400)]
Input: libps2 - warn instead of oopsing when passed bad arguments

This is more user-friendly and also fixes Coverity #id 249

Signed-off-by: Dmitry Torokhov <dtor@mail.ru>
17 years agoInput: fm801-gp - fix use after free
Dmitry Torokhov [Fri, 7 Jul 2006 03:54:30 +0000 (23:54 -0400)]
Input: fm801-gp - fix use after free

Fixes Coverity #id 916

Signed-off-by: Dmitry Torokhov <dtor@mail.ru>
17 years agoInput: wistron - fix section reference mismatches
Andrew Morton [Thu, 6 Jul 2006 04:23:38 +0000 (00:23 -0400)]
Input: wistron - fix section reference mismatches

Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Dmitry Torokhov <dtor@mail.ru>
17 years agoInput: introduce input_inject_event() function
Dmitry Torokhov [Thu, 6 Jul 2006 04:22:43 +0000 (00:22 -0400)]
Input: introduce input_inject_event() function

Create input_inject_event() function which is to be used by input
handlers as opposed to input_event() which is reserved for drivers
implementing input devices. The difference is that if device is
"grabbed" by some process input_inject_event() will ignore events
unless sent from the handle that is currently owns the device.

Signed-off-by: Dmitry Torokhov <dtor@mail.ru>
17 years agoInput: add start() method to input handlers
Dmitry Torokhov [Thu, 6 Jul 2006 04:21:03 +0000 (00:21 -0400)]
Input: add start() method to input handlers

The new start() method is called immediately after connect() and also
when "grabbed" device is released by its owner. This will allow input
handlers to re-synchronize state of once-grabbed device with the rest
of devices.

Signed-off-by: Dmitry Torokhov <dtor@mail.ru>
17 years agoInput: remove accept method from input_dev
Dmitry Torokhov [Thu, 6 Jul 2006 04:20:47 +0000 (00:20 -0400)]
Input: remove accept method from input_dev

This method used to enforce exclusive access to iforce devices,
but presenlty there are no known users of this method.

Signed-off-by: Dmitry Torokhov <dtor@mail.ru>
17 years ago[PATCH] Makefile typo
Andreas Schwab [Wed, 5 Jul 2006 16:57:48 +0000 (18:57 +0200)]
[PATCH] Makefile typo

Fix a typo in the toplevel makefile.

Signed-off-by: Andreas Schwab <schwab@suse.de>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
17 years ago[PATCH] powerpc: implement missing jiffies64_to_cputime64()
David Woodhouse [Wed, 5 Jul 2006 10:24:26 +0000 (11:24 +0100)]
[PATCH] powerpc: implement missing jiffies64_to_cputime64()

asm-powerpc/cputime.h doesn't declare jiffies64_to_cputime64() or
cputime64_sub(), and due to CONFIG_VIRT_CPU_ACCOUNTING it's not picking
up the definition from asm-generic like x86-64 & friends do.

Cc: Dave Jones <davej@redhat.com>
Cc: Andrew Morton <akpm@osdl.org>
Signed-off-by: David Woodhouse <dwmw2@infradead.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
17 years ago[PATCH] powerpc: Fix loss of interrupts with MPIC
Benjamin Herrenschmidt [Wed, 5 Jul 2006 05:36:15 +0000 (15:36 +1000)]
[PATCH] powerpc: Fix loss of interrupts with MPIC

With the new interrupt rework, an interrupt "host" map() callback can be
called after the interrupt is already active.

It's called again for an already mapped interrupt to allow changing the
trigger setup, and currently this is not guarded with a test of wether
the interrupt is requested or not.

I plan to change some of this logic to be a bit less lenient against
random reconfiguring of live interrupts but just not yet.

The ported MPIC driver has a bug where when that happens, it will mask
the interrupt.  This changes it to preserve the previous masking of the
interrupt instead.

Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
17 years agoMerge master.kernel.org:/pub/scm/linux/kernel/git/davej/cpufreq
Linus Torvalds [Tue, 4 Jul 2006 21:00:26 +0000 (14:00 -0700)]
Merge /pub/scm/linux/kernel/git/davej/cpufreq

* master.kernel.org:/pub/scm/linux/kernel/git/davej/cpufreq:
  Move workqueue exports to where the functions are defined.
  [CPUFREQ] Misc cleanups in ondemand.
  [CPUFREQ] Make ondemand sampling per CPU and remove the mutex usage in sampling path.
  [CPUFREQ] Add queue_delayed_work_on() interface for workqueues.
  [CPUFREQ] Remove slowdown from ondemand sampling path.

17 years agoFix up headers_install wrt devfs removal
Linus Torvalds [Tue, 4 Jul 2006 21:00:06 +0000 (14:00 -0700)]
Fix up headers_install wrt devfs removal

No devfs_fs.h header any more..

Signed-off-by: Linus Torvalds <torvalds@osdl.org>
17 years agoMerge git://git.infradead.org/hdrinstall-2.6
Linus Torvalds [Tue, 4 Jul 2006 19:55:45 +0000 (12:55 -0700)]
Merge git://git.infradead.org/hdrinstall-2.6

* git://git.infradead.org/hdrinstall-2.6:
  Remove export of include/linux/isdn/tpam.h
  Remove <linux/i2c-id.h> and <linux/i2c-algo-ite.h> from userspace export
  Restrict headers exported to userspace for SPARC and SPARC64
  Add empty Kbuild files for 'make headers_install' in remaining arches.
  Add Kbuild file for Alpha 'make headers_install'
  Add Kbuild file for SPARC 'make headers_install'
  Add Kbuild file for IA64 'make headers_install'
  Add Kbuild file for S390 'make headers_install'
  Add Kbuild file for i386 'make headers_install'
  Add Kbuild file for x86_64 'make headers_install'
  Add Kbuild file for PowerPC 'make headers_install'
  Add generic Kbuild files for 'make headers_install'
  Basic implementation of 'make headers_check'
  Basic implementation of 'make headers_install'

17 years ago[PATCH] fix AB-BA deadlock inversion at cs46xx_dsp_remove_scb
Arjan van de Ven [Tue, 4 Jul 2006 10:07:22 +0000 (03:07 -0700)]
[PATCH] fix AB-BA deadlock inversion at cs46xx_dsp_remove_scb

There is a code sequence where the locking is substream->self_group.lock
-> ins->scbs[index].lock

substream->self_group.lock is interrupt safe, and taken from irq context
as well (trace is snipped for brevity)

so what can happen is

   cpu 0                    cpu 1
   user context user context

take ins->scbs[index].lock without disabling interrupts

   get substream->self_group.lock (irqsafe)
   try to get ins->scbs[index].lock (spins)

interrupt happens
try to get substream->self_group.lock (spins)

which is an obvious AB-BA deadlock

fix is to just take the lock with _irqsafe

Signed-off-by: Arjan van de Ven <arjan@linux.intel.com>
Cc: Jaroslav Kysela <perex@suse.cz>
Acked-by: Takashi Iwai <tiwai@suse.de>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
17 years ago[PATCH] mthca: initialize send and receive queue locks separately
Zach Brown [Tue, 4 Jul 2006 09:57:52 +0000 (02:57 -0700)]
[PATCH] mthca: initialize send and receive queue locks separately

mthca: initialize send and receive queue locks separately

lockdep identifies a lock by the call site of its initialization.  By
initializing the send and receive queue locks in mthca_wq_init() we confuse
lockdep.  It warns that that the ordered acquiry of both locks in
mthca_modify_qp() is recursive acquiry of one lock:

  =============================================
  [ INFO: possible recursive locking detected ]
  ---------------------------------------------
  modprobe/1192 is trying to acquire lock:
   (&wq->lock){....}, at: [<f892b4db>] mthca_modify_qp+0x60/0xa7b [ib_mthca]
  but task is already holding lock:
   (&wq->lock){....}, at: [<f892b4ce>] mthca_modify_qp+0x53/0xa7b [ib_mthca]

Initializing the locks separately in mthca_alloc_qp_common() stops the
warning and will let lockdep enforce proper ordering on paths that acquire
both locks.

Signed-off-by: Zach Brown <zach.brown@oracle.com>
Cc: Roland Dreier <rolandd@cisco.com>
Cc: Arjan van de Ven <arjan@linux.intel.com>
Cc: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
17 years ago[PATCH] Fix copying of pgdat array on each node for ia64 memory hotplug
Yasunori Goto [Tue, 4 Jul 2006 09:57:51 +0000 (02:57 -0700)]
[PATCH] Fix copying of pgdat array on each node for ia64 memory hotplug

I found a bug in memory hot-add code for ia64.

IA64's code has copies of pgdat's array on each node to reduce memory
access over crossing node.  This array is used by NODE_DATA() macro.  When
new node is hot-added, this pgdat's array should be updated and copied on
new node too.

However, I used for_each_online_node() in scatter_node_data() to copy
it. This meant its array is not copied on new node.
Because initialization of structures for new node was halfway,
so online_node_map couldn't be set at this time.

To copy arrays on new node, I changed it to check value of pgdat_list[]
which is source array of copies.  I tested this patch with my Memory Hotadd
emulation on Tiger4.  This patch is for 2.6.17-git20.

Signed-off-by: Yasunori Goto <y-goto@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>
17 years ago[PATCH] uclinux: fix proc_task()/get_proc-task() naming
Greg Ungerer [Tue, 4 Jul 2006 05:04:39 +0000 (15:04 +1000)]
[PATCH] uclinux: fix proc_task()/get_proc-task() naming

Fix changed name of proc_task() to get_proc_task().

Signed-off-by: Greg Ungerer <gerg@uclinux.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
17 years agoMerge branch 'release' of git://git.kernel.org/pub/scm/linux/kernel/git/lenb/linux...
Linus Torvalds [Tue, 4 Jul 2006 04:32:50 +0000 (21:32 -0700)]
Merge branch 'release' of git://git./linux/kernel/git/lenb/linux-acpi-2.6

* 'release' of git://git.kernel.org/pub/scm/linux/kernel/git/lenb/linux-acpi-2.6: (44 commits)
  ACPI: remove function tracing macros from drivers/acpi/*.c
  ACPI: add support for Smart Battery
  ACPI: handle battery notify event on broken BIOS
  ACPI: handle AC notify event on broken BIOS
  ACPI: asus_acpi: add S1N WLED control
  ACPI: asus_acpi: correct M6N/M6R display nodes
  ACPI: asus_acpi: add S1N WLED control
  ACPI: asus_acpi: rework model detection
  ACPI: asus_acpi: support L5D
  ACPI: asus_acpi: handle internal Bluetooth / support W5A
  ACPI: asus_acpi: support A4G
  ACPI: asus_acpi: support W3400N
  ACPI: asus_acpi: LED display support
  ACPI: asus_acpi: support A3G
  ACPI: asus_acpi: misc cleanups
  ACPI: video: Remove unneeded acpi_handle from driver.
  ACPI: thermal: Remove unneeded acpi_handle from driver.
  ACPI: power: Remove unneeded acpi_handle from driver.
  ACPI: pci_root: Remove unneeded acpi_handle from driver.
  ACPI: pci_link: Remove unneeded acpi_handle from driver.
  ...

17 years agoMerge branch 'devel' of master.kernel.org:/home/rmk/linux-2.6-serial
Linus Torvalds [Tue, 4 Jul 2006 04:31:05 +0000 (21:31 -0700)]
Merge branch 'devel' of /home/rmk/linux-2.6-serial

* 'devel' of master.kernel.org:/home/rmk/linux-2.6-serial:
  [SERIAL] Ensure 8250_pci quirks are not marked __devinit
  [SERIAL] Convert fifosize to an unsigned int

17 years agoMerge master.kernel.org:/home/rmk/linux-2.6-arm
Linus Torvalds [Tue, 4 Jul 2006 04:30:03 +0000 (21:30 -0700)]
Merge master.kernel.org:/home/rmk/linux-2.6-arm

* master.kernel.org:/home/rmk/linux-2.6-arm:
  [ARM] Fix lh7a40x_udc.c
  [ARM] Fix warning in consistent.c
  [ARM] Fix warnings in arch/arm/kernel/setup.c
  [ARM] Fix ecard.c resource warnings.
  [ARM] Fix ISA IRQ resources
  [ARM] Fix bad asm instruction in proc-arm925.S
  [ARM] More missing proc-macros.S includes
  [ARM] 3708/2: fix SMP build after section ioremap changes

17 years agoMerge git://git.infradead.org/mtd-2.6
Linus Torvalds [Tue, 4 Jul 2006 04:29:08 +0000 (21:29 -0700)]
Merge git://git.infradead.org/mtd-2.6

* git://git.infradead.org/mtd-2.6:
  [JFFS2][XATTR] Fix memory leak in POSIX-ACL support
  fs/jffs2/: make 2 functions static
  [MTD] NAND: Fix broken sharpsl driver
  [JFFS2][XATTR] Fix xd->refcnt race condition
  MTD: kernel-doc fixes + additions
  MTD: fix all kernel-doc warnings
  [MTD] DOC: Fixup read functions and do a little cleanup

17 years agoMerge master.kernel.org:/pub/scm/linux/kernel/git/davem/net-2.6
Linus Torvalds [Tue, 4 Jul 2006 04:28:14 +0000 (21:28 -0700)]
Merge /pub/scm/linux/kernel/git/davem/net-2.6

* master.kernel.org:/pub/scm/linux/kernel/git/davem/net-2.6: (27 commits)
  [Bluetooth] Add RFCOMM role switch support
  [Bluetooth] Allow disabling of credit based flow control
  [Bluetooth] Small cleanup of the L2CAP source code
  [Bluetooth] Use real devices for host controllers
  [Bluetooth] Add platform device for virtual and serial devices
  [Bluetooth] Add automatic sniff mode support
  [Bluetooth] Correct SCO buffer size on request
  [Bluetooth] Add suspend/resume support to the HCI USB driver
  [Bluetooth] Use raw mode for the Frontline sniffer device
  [BRIDGE]: br_dump_ifinfo index fix
  [ATM]: add+use poison defines
  [NET]: add+use poison defines
  [IOAT]: fix kernel-doc in source files
  [IOAT]: fix header file kernel-doc
  [TG3]: Add ipv6 TSO feature
  [IPV6]: Fix ipv6 GSO payload length
  [TIPC] Fixed sk_buff panic caused by tipc_link_bundle_buf (REVISED)
  [NET]: Verify gso_type too in gso_segment
  [IPVS]: Add sysctl documentation
  [ROSE]: Try all routes when establishing a ROSE connections.
  ...

17 years agoMerge master.kernel.org:/pub/scm/linux/kernel/git/jejb/scsi-misc-2.6
Linus Torvalds [Tue, 4 Jul 2006 04:27:18 +0000 (21:27 -0700)]
Merge /pub/scm/linux/kernel/git/jejb/scsi-misc-2.6

* master.kernel.org:/pub/scm/linux/kernel/git/jejb/scsi-misc-2.6: (57 commits)
  [SCSI] fix error handling in scsi_io_completion
  [SCSI] qla1280: fix section mismatch warnings
  [SCSI] mptsas: eliminate ghost devices
  [SCSI] qla2xxx: make some more functions static
  [SCSI] small whitespace cleanup for qlogic driver
  [SCSI] mptbase: mpt_interrupt should return IRQ_NONE
  [SCSI] mptsas: make two functions static
  [SCSI] sg.c: Fix bad error handling in
  [SCSI] 53c700: fix breakage caused by the autosense update
  [SCSI] iscsi: add async notification of session events
  [SCSI] iscsi: pass target nr to session creation
  [SCSI] iscsi: break up session creation into two stages
  [SCSI] iscsi: rm channel usage from iscsi
  [SCSI] iscsi: fix session refcouting
  [SCSI] iscsi: convert iscsi_tcp to new set/get param fns
  [SCSI] iscsi: convert iser to new set/get param fns
  [SCSI] iscsi: fixup set/get param functions
  [SCSI] iscsi: add target discvery event to transport class
  [SCSI] st: remove unused st_buffer.in_use
  [SCSI] atp870u: reduce huge stack usage
  ...

17 years agoMerge git://git.kernel.org/pub/scm/linux/kernel/git/sam/kbuild
Linus Torvalds [Tue, 4 Jul 2006 04:26:12 +0000 (21:26 -0700)]
Merge git://git./linux/kernel/git/sam/kbuild

* git://git.kernel.org/pub/scm/linux/kernel/git/sam/kbuild:
  kbuild: introduce utsrelease.h
  kbuild: explicit turn off gcc stack-protector

17 years ago[PATCH] revert "kthread: convert stop_machine into a kthread"
Andrew Morton [Tue, 4 Jul 2006 00:32:22 +0000 (17:32 -0700)]
[PATCH] revert "kthread: convert stop_machine into a kthread"

Jiri reports that the stop_machin kthread conversion caused his machine to
hang when suspending.  Hyperthreading is apparently involved.

I don't see why that would be and I can't reproduce it.  Revert to the 2.6.17
code.

Cc: "Serge E. Hallyn" <serue@us.ibm.com>
Cc: Jiri Slaby <jirislaby@gmail.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
17 years ago[PATCH] zfcp: fix incorrect usage of fsf_req_list_lock
Heiko Carstens [Tue, 4 Jul 2006 00:32:21 +0000 (17:32 -0700)]
[PATCH] zfcp: fix incorrect usage of fsf_req_list_lock

  =================================
  [ INFO: inconsistent lock state ]
  ---------------------------------
  inconsistent {in-hardirq-W} -> {hardirq-on-W} usage.
  swapper/0 [HC0[0]:SC1[1]:HE1:SE0] takes:
   (&adapter->fsf_req_list_lock){++..}, at: [<0000000000274486>] zfcp_qdio_reqid_check+0x46/0x178
  {in-hardirq-W} state was registered at:
    [<000000000005fb0c>] __lock_acquire+0xad8/0xed0
    [<00000000000604ae>] lock_acquire+0x9a/0xc8
    [<000000000035a326>] _spin_lock+0x4e/0x68
    [<0000000000274486>] zfcp_qdio_reqid_check+0x46/0x178
    [<000000000027469e>] zfcp_qdio_response_handler+0xe6/0x430
    [<0000000000219dd4>] tiqdio_thinint_handler+0xd20/0x213c
    [<000000000020229a>] do_adapter_IO+0xb2/0xc0
    [<0000000000206f32>] do_IRQ+0x136/0x16c
    [<0000000000020462>] io_no_vtime+0x16/0x1c
    [<0000000000019432>] cpu_idle+0x222/0x250
  irq event stamp: 129220
  hardirqs last  enabled at (129220): [<00000000000411e6>] tasklet_hi_action+0x5a/0x19c
  hardirqs last disabled at (129219): [<00000000000411c0>] tasklet_hi_action+0x34/0x19c
  softirqs last  enabled at (129212): [<0000000000040b62>] __do_softirq+0x13a/0x180
  softirqs last disabled at (129217): [<000000000001fd58>] do_softirq+0xec/0xf0

  other info that might help us debug this:
  no locks held by swapper/0.

  stack backtrace:
  00000000012bb670 0000000000000002 0000000000000000 00000000012bb780
         00000000012bb6e8 0000000000399122 0000000000399122 0000000000016b0a
         0000000000000000 0000000000000000 0000000000000000 00000000004660e8
         0000000000000000 000000000000000d 00000000012bb6e0 00000000012bb758
         0000000000368b90 0000000000016b0a 00000000012bb6e0 00000000012bb730
  Call Trace:
  ([<0000000000016a26>] show_trace+0x76/0xdc)
   [<0000000000016b2c>] show_stack+0xa0/0xd0
   [<0000000000016b8a>] dump_stack+0x2e/0x3c
   [<000000000005e3da>] print_usage_bug+0x27e/0x290
   [<000000000005ea9c>] mark_lock+0x6b0/0x6c0
   [<000000000005f33e>] __lock_acquire+0x30a/0xed0
   [<00000000000604ae>] lock_acquire+0x9a/0xc8
   [<000000000035a326>] _spin_lock+0x4e/0x68
   [<0000000000274486>] zfcp_qdio_reqid_check+0x46/0x178
   [<000000000027469e>] zfcp_qdio_response_handler+0xe6/0x430
   [<0000000000217bd2>] tiqdio_tl+0xd02/0x2120
   [<000000000004123a>] tasklet_hi_action+0xae/0x19c
   [<0000000000040ae4>] __do_softirq+0xbc/0x180
   [<000000000001fd58>] do_softirq+0xec/0xf0
   [<0000000000040c38>] irq_exit+0x90/0xa8
   [<0000000000206f40>] do_IRQ+0x144/0x16c
   [<0000000000020462>] io_no_vtime+0x16/0x1c
   [<0000000000019432>] cpu_idle+0x222/0x250
  ([<0000000000019416>] cpu_idle+0x206/0x250)
   [<000000000001405a>] rest_init+0x5a/0x68
   [<0000000000536998>] start_kernel+0x39c/0x3dc
   [<0000000000013046>] _stext+0x46/0x1000

Fix incorrect usage of fsf_req_list_lock. It's used in tasklet context
(irqs on) as well as in irq context. Therefore use the spin_lock_irqsave
variant to avoid deadlocks.

Acked-by: Andreas Herrmann <aherrman@de.ibm.com>
Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
17 years ago[PATCH] zfcp: fix incorrect usage of erp_lock
Heiko Carstens [Tue, 4 Jul 2006 00:32:20 +0000 (17:32 -0700)]
[PATCH] zfcp: fix incorrect usage of erp_lock

  =================================
  [ INFO: inconsistent lock state ]
  ---------------------------------
  inconsistent {hardirq-on-W} -> {in-hardirq-W} usage.
  swapper/0 [HC1[1]:SC0[0]:HE0:SE1] takes:
   (&adapter->erp_lock){+-..}, at: [<000000000026c7f8>] zfcp_erp_async_handler+0x3c/0x70
  {hardirq-on-W} state was registered at:
    [<000000000005f33e>] __lock_acquire+0x30a/0xed0
    [<00000000000604ae>] lock_acquire+0x9a/0xc8
    [<000000000035a7ae>] _write_lock+0x4e/0x68
    [<000000000026d822>] zfcp_erp_adapter_strategy_generic+0x286/0xd94
    [<000000000026fd72>] zfcp_erp_strategy_do_action+0x91e/0x1a94
    [<0000000000271a3a>] zfcp_erp_thread+0x21a/0x1568
    [<0000000000019096>] kernel_thread_starter+0x6/0xc
    [<0000000000019090>] kernel_thread_starter+0x0/0xc
  irq event stamp: 12078
  hardirqs last  enabled at (12077): [<0000000000019416>] cpu_idle+0x206/0x250
  hardirqs last disabled at (12078): [<0000000000020458>] io_no_vtime+0xc/0x1c
  softirqs last  enabled at (12072): [<0000000000040b62>] __do_softirq+0x13a/0x180
  softirqs last disabled at (12059): [<000000000001fd58>] do_softirq+0xec/0xf0

  other info that might help us debug this:
  no locks held by swapper/0.

  stack backtrace:
  00000000012bb648 0000000000000002 0000000000000000 00000000012bb758
         00000000012bb6c0 0000000000399122 0000000000399122 0000000000016b0a
         0000000000000000 0000000000000001 0000000000000000 00000000004660e8
         0000000000000000 000000000000000d 00000000012bb6b8 00000000012bb730
         0000000000368b90 0000000000016b0a 00000000012bb6b8 00000000012bb708
  Call Trace:
  ([<0000000000016a26>] show_trace+0x76/0xdc)
   [<0000000000016b2c>] show_stack+0xa0/0xd0
   [<0000000000016b8a>] dump_stack+0x2e/0x3c
   [<000000000005e3da>] print_usage_bug+0x27e/0x290
   [<000000000005e934>] mark_lock+0x548/0x6c0
   [<000000000005fb0c>] __lock_acquire+0xad8/0xed0
   [<00000000000604ae>] lock_acquire+0x9a/0xc8
   [<000000000035a662>] _write_lock_irqsave+0x62/0x80
   [<000000000026c7f8>] zfcp_erp_async_handler+0x3c/0x70
   [<0000000000279178>] zfcp_fsf_req_dispatch+0xd8/0x1fa8
   [<000000000027e538>] zfcp_fsf_req_complete+0x104/0xe4c
   [<0000000000274534>] zfcp_qdio_reqid_check+0xf4/0x178
   [<000000000027469e>] zfcp_qdio_response_handler+0xe6/0x430
   [<0000000000219dd4>] tiqdio_thinint_handler+0xd20/0x213c
   [<000000000020229a>] do_adapter_IO+0xb2/0xc0
   [<0000000000206f32>] do_IRQ+0x136/0x16c
   [<0000000000020462>] io_no_vtime+0x16/0x1c
   [<0000000000019432>] cpu_idle+0x222/0x250
  ([<0000000000019416>] cpu_idle+0x206/0x250)
   [<000000000001405a>] rest_init+0x5a/0x68
   [<0000000000536998>] start_kernel+0x39c/0x3dc
   [<0000000000013046>] _stext+0x46/0x1000

Fix incorrect usage of erp_lock. Using the write_lock() variant is wrong,
since this might lead to deadlocks.

Acked-by: Andreas Herrmann <aherrman@de.ibm.com>
Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
17 years ago[PATCH] Fix modular cpuid.ko
OGAWA Hirofumi [Tue, 4 Jul 2006 00:32:20 +0000 (17:32 -0700)]
[PATCH] Fix modular cpuid.ko

With recent change, if CONFIG_HOTPLUG_CPU is disabled,
register_cpu_notifier() is not exported.  And it breaked moduler msr/cpuid
(msr.c was already fixed).

We need to use register_hotcpu_notifier() now in module, instead of
register_cpu_notifier().

Signed-off-by: OGAWA Hirofumi <hirofumi@mail.parknet.co.jp>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
17 years ago[Bluetooth] Add RFCOMM role switch support
Marcel Holtmann [Mon, 3 Jul 2006 08:37:55 +0000 (10:37 +0200)]
[Bluetooth] Add RFCOMM role switch support

This patch adds the support for RFCOMM role switching before the
connection is fully established.

Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
17 years ago[Bluetooth] Allow disabling of credit based flow control
Marcel Holtmann [Mon, 3 Jul 2006 08:02:51 +0000 (10:02 +0200)]
[Bluetooth] Allow disabling of credit based flow control

This patch adds the module parameter disable_cfc which can be used to
disable the credit based flow control. The credit based flow control
was introduced with the Bluetooth 1.1 specification and devices can
negotiate its support, but for testing purpose it is helpful to allow
disabling of it.

Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
17 years ago[Bluetooth] Small cleanup of the L2CAP source code
Marcel Holtmann [Mon, 3 Jul 2006 08:02:46 +0000 (10:02 +0200)]
[Bluetooth] Small cleanup of the L2CAP source code

This patch is a small cleanup of the L2CAP source code. It makes some
coding style changes and moves some functions around to avoid forward
declarations.

Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
17 years ago[Bluetooth] Use real devices for host controllers
Marcel Holtmann [Mon, 3 Jul 2006 08:02:41 +0000 (10:02 +0200)]
[Bluetooth] Use real devices for host controllers

This patch converts the Bluetooth class devices into real devices. The
Bluetooth class is kept and the driver core provides the appropriate
symlinks for backward compatibility.

Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
17 years ago[Bluetooth] Add platform device for virtual and serial devices
Marcel Holtmann [Mon, 3 Jul 2006 08:02:37 +0000 (10:02 +0200)]
[Bluetooth] Add platform device for virtual and serial devices

This patch adds a generic Bluetooth platform device that can be used
as parent device by virtual and serial devices.

Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
17 years ago[Bluetooth] Add automatic sniff mode support
Marcel Holtmann [Mon, 3 Jul 2006 08:02:33 +0000 (10:02 +0200)]
[Bluetooth] Add automatic sniff mode support

This patch introduces the automatic sniff mode feature. This allows
the host to switch idle connections into sniff mode to safe power.

Signed-off-by: Ulisses Furquim <ulissesf@gmail.com>
Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
17 years ago[Bluetooth] Correct SCO buffer size on request
Marcel Holtmann [Mon, 3 Jul 2006 08:02:29 +0000 (10:02 +0200)]
[Bluetooth] Correct SCO buffer size on request

This patch introduces a quirk that allows the drivers to tell the host
to correct the SCO buffer size values.

Signed-off-by: Olivier Galibert <galibert@pobox.com>
Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
17 years ago[Bluetooth] Add suspend/resume support to the HCI USB driver
Marcel Holtmann [Mon, 3 Jul 2006 08:02:24 +0000 (10:02 +0200)]
[Bluetooth] Add suspend/resume support to the HCI USB driver

This patch implements the suspend/resume methods for the HCI USB
driver by killing all outstanding URBs on suspend, and re-issuing
them on resume.

Signed-off-by: Johannes Berg <johannes@sipsolutions.net>
Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
17 years ago[Bluetooth] Use raw mode for the Frontline sniffer device
Marcel Holtmann [Mon, 3 Jul 2006 08:02:18 +0000 (10:02 +0200)]
[Bluetooth] Use raw mode for the Frontline sniffer device

The Frontline sniffer device looks like a normal H:2 Bluetooth device,
but it is not and so mark it as raw mode device.

Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
17 years ago[BRIDGE]: br_dump_ifinfo index fix
Andrey Savochkin [Tue, 4 Jul 2006 02:50:14 +0000 (19:50 -0700)]
[BRIDGE]: br_dump_ifinfo index fix

Fix for inability of br_dump_ifinfo to handle non-zero start index:
loop index never increases when entered with non-zero start.
Spotted by Kirill Korotaev.

Signed-off-by: Andrey Savochkin <saw@swsoft.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
17 years ago[ATM]: add+use poison defines
Randy Dunlap [Tue, 4 Jul 2006 02:48:25 +0000 (19:48 -0700)]
[ATM]: add+use poison defines

ATM: add and use POISON define values.

Signed-off-by: Randy Dunlap <rdunlap@xenotime.net>
Signed-off-by: David S. Miller <davem@davemloft.net>
17 years ago[NET]: add+use poison defines
Randy Dunlap [Tue, 4 Jul 2006 02:47:27 +0000 (19:47 -0700)]
[NET]: add+use poison defines

Add and use poison defines in net/.

Signed-off-by: Randy Dunlap <rdunlap@xenotime.net>
Signed-off-by: David S. Miller <davem@davemloft.net>
17 years ago[IOAT]: fix kernel-doc in source files
Randy Dunlap [Tue, 4 Jul 2006 02:45:31 +0000 (19:45 -0700)]
[IOAT]: fix kernel-doc in source files

Fix kernel-doc warnings in drivers/dma/:
- use correct function & parameter names
- add descriptions where omitted

Signed-off-by: Randy Dunlap <rdunlap@xenotime.net>
Signed-off-by: David S. Miller <davem@davemloft.net>
17 years ago[IOAT]: fix header file kernel-doc
Randy Dunlap [Tue, 4 Jul 2006 02:44:51 +0000 (19:44 -0700)]
[IOAT]: fix header file kernel-doc

Fix kernel-doc problems in include/linux/dmaengine.h:
- add some fields/parameters
- expand some descriptions
- fix typos

Signed-off-by: Randy Dunlap <rdunlap@xenotime.net>
Signed-off-by: David S. Miller <davem@davemloft.net>
17 years ago[TG3]: Add ipv6 TSO feature
Michael Chan [Tue, 4 Jul 2006 02:42:14 +0000 (19:42 -0700)]
[TG3]: Add ipv6 TSO feature

Enable ipv6 TSO feature on chips that support it.

Update version to 3.61.

Signed-off-by: Michael Chan <mchan@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
17 years ago[IPV6]: Fix ipv6 GSO payload length
Michael Chan [Tue, 4 Jul 2006 02:41:11 +0000 (19:41 -0700)]
[IPV6]: Fix ipv6 GSO payload length

Fix ipv6 GSO payload length calculation.

The ipv6 payload length excludes the ipv6 base header length and so
must be subtracted.

Signed-off-by: Michael Chan <mchan@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
17 years ago[TIPC] Fixed sk_buff panic caused by tipc_link_bundle_buf (REVISED)
Allan Stephens [Tue, 4 Jul 2006 02:39:36 +0000 (19:39 -0700)]
[TIPC] Fixed sk_buff panic caused by tipc_link_bundle_buf (REVISED)

The recent change to direct inspection of bundle buffer tailroom did not
account for the possiblity of unrequested tailroom added by skb_alloc(),
thereby allowing a bundle to be created that exceeds the current link MTU.
An additional check now ensures that bundling works correctly no matter
if the bundle buffer is smaller, larger, or equal to the link MTU.

Signed-off-by: Allan Stephens <allan.stephens@windriver.com>
Signed-off-by: Per Liden <per.liden@ericsson.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
17 years ago[NET]: Verify gso_type too in gso_segment
Herbert Xu [Tue, 4 Jul 2006 02:38:35 +0000 (19:38 -0700)]
[NET]: Verify gso_type too in gso_segment

We don't want nasty Xen guests to pass a TCPv6 packet in with gso_type set
to TCPv4 or even UDP (or a packet that's both TCP and UDP).

Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Signed-off-by: David S. Miller <davem@davemloft.net>
17 years ago[IPVS]: Add sysctl documentation
Horms [Tue, 4 Jul 2006 02:35:40 +0000 (19:35 -0700)]
[IPVS]: Add sysctl documentation

* Derived from http://www.linuxvirtualserver.org/docs/sysctl.html, v1.4
  maintained by Wensong Zhang

* Adjusted preample to match ip-sysctl.txt

* Sorted options into alphabetical order

* Added expire_quiescent_template

* Removed timeout_* which are no longer present

* Incoporated doc/debug-levels.txt from IPVS source tree into
  description of ipvs_debug

* Minor spelling fixes

* Further editing more than welcome

Signed-Off-By: Horms <horms@verge.net.au>
Signed-off-by: David S. Miller <davem@davemloft.net>
17 years ago[ROSE]: Try all routes when establishing a ROSE connections.
Ralf Baechle [Tue, 4 Jul 2006 02:32:23 +0000 (19:32 -0700)]
[ROSE]: Try all routes when establishing a ROSE connections.

From Jean-Paul F6FBB

ROSE will only try to establish a route using the first route in its
routing table.  Fix to iterate through all additional routes if a
connection attempt has failed.

Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
17 years ago[NETROM]: Use socket helpers instead of direct fiddling with struct sock
Ralf Baechle DL5RB [Tue, 4 Jul 2006 02:31:14 +0000 (19:31 -0700)]
[NETROM]: Use socket helpers instead of direct fiddling with struct sock

Signed-off-by: Ralf Baechle DL5RB <ralf@linux-mips.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
17 years ago[AX.25]: Reference counting for AX.25 routes.
Ralf Baechle DL5RB [Tue, 4 Jul 2006 02:30:18 +0000 (19:30 -0700)]
[AX.25]: Reference counting for AX.25 routes.

In the past routes could be freed even though the were possibly in use ...

Signed-off-by: Ralf Baechle DL5RB <ralf@linux-mips.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
17 years ago[ROSE]: Fix dereference of skb pointer after free.
Ralf Baechle [Tue, 4 Jul 2006 02:29:15 +0000 (19:29 -0700)]
[ROSE]: Fix dereference of skb pointer after free.

If rose_route_frame return success we'll dereference a stale pointer.
Likely this is only going to result in bad statistics for the ROSE
interface.

This fixes coverity 946.

Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
17 years ago[IOAT]: Fix a warning in ioatdma
Benoit Boissinot [Tue, 4 Jul 2006 02:28:13 +0000 (19:28 -0700)]
[IOAT]: Fix a warning in ioatdma

drivers/dma/ioatdma.c: In function 'ioat_init_module':
drivers/dma/ioatdma.c:830: warning: control reaches end of non-void function

Signed-off-by: Benoit Boissinot <benoit.boissinot@ens-lyon.org>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
17 years ago[IOAT]: drivers/dma/iovlock.c: make num_pages_spanned() static
Adrian Bunk [Tue, 4 Jul 2006 02:27:20 +0000 (19:27 -0700)]
[IOAT]: drivers/dma/iovlock.c: make num_pages_spanned() static

This patch makes the needlessly global num_pages_spanned() static.

Signed-off-by: Adrian Bunk <bunk@stusta.de>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
17 years ago[AF_UNIX]: datagram getpeersec fix
Andrew Morton [Tue, 4 Jul 2006 02:26:15 +0000 (19:26 -0700)]
[AF_UNIX]: datagram getpeersec fix

The unix_get_peersec_dgram() stub should have been inlined so that it
disappears.

Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
17 years ago[IOAT]: fix sparse ulong warning
Randy Dunlap [Tue, 4 Jul 2006 02:24:19 +0000 (19:24 -0700)]
[IOAT]: fix sparse ulong warning

Fix sparse warning:
drivers/dma/ioatdma.c:444:32: warning: constant 0xFFFFFFFFFFFFFFC0 is so big it is unsigned long

Also needs a MAINTAINERS entry.

Signed-off-by: Randy Dunlap <rdunlap@xenotime.net>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
17 years agoMerge git://git.kernel.org/pub/scm/linux/kernel/git/paulus/powerpc
Linus Torvalds [Mon, 3 Jul 2006 22:28:34 +0000 (15:28 -0700)]
Merge git://git./linux/kernel/git/paulus/powerpc

* git://git.kernel.org/pub/scm/linux/kernel/git/paulus/powerpc:
  powerpc: add defconfig for Freescale MPC8349E-mITX board
  powerpc: Add base support for the Freescale MPC8349E-mITX eval board
  Documentation: correct values in MPC8548E SEC example node
  [POWERPC] Actually copy over i8259.c to arch/ppc/syslib this time
  [POWERPC] Add new interrupt mapping core and change platforms to use it
  [POWERPC] Copy i8259 code back to arch/ppc
  [POWERPC] New device-tree interrupt parsing code
  [POWERPC] Use the genirq framework
  [PATCH] genirq: Allow fasteoi handler to retrigger disabled interrupts
  [POWERPC] Update the SWIM3 (powermac) floppy driver
  [POWERPC] Fix error handling in detecting legacy serial ports
  [POWERPC] Fix booting on Momentum "Apache" board (a Maple derivative)
  [POWERPC] Fix various offb and BootX-related issues
  [POWERPC] Add a default config for 32-bit CHRP machines
  [POWERPC] fix implicit declaration on cell.
  [POWERPC] change get_property to return void *

17 years ago[PATCH] sched: cleanup, convert sched.c-internal typedefs to struct
Ingo Molnar [Mon, 3 Jul 2006 07:25:42 +0000 (00:25 -0700)]
[PATCH] sched: cleanup, convert sched.c-internal typedefs to struct

convert:

 - runqueue_t to 'struct rq'
 - prio_array_t to 'struct prio_array'
 - migration_req_t to 'struct migration_req'

I was the one who added these but they are both against the kernel coding
style and also were used inconsistently at places.  So just get rid of them at
once, now that we are flushing the scheduler patch-queue anyway.

Conversion was mostly scripted, the result was reviewed and all secondary
whitespace and style impact (if any) was fixed up by hand.

Signed-off-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
17 years ago[PATCH] sched: cleanup, remove task_t, convert to struct task_struct
Ingo Molnar [Mon, 3 Jul 2006 07:25:41 +0000 (00:25 -0700)]
[PATCH] sched: cleanup, remove task_t, convert to struct task_struct

cleanup: remove task_t and convert all the uses to struct task_struct. I
introduced it for the scheduler anno and it was a mistake.

Conversion was mostly scripted, the result was reviewed and all
secondary whitespace and style impact (if any) was fixed up by hand.

Signed-off-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
17 years ago[PATCH] sched: clean up fallout of recent changes
Ingo Molnar [Mon, 3 Jul 2006 07:25:40 +0000 (00:25 -0700)]
[PATCH] sched: clean up fallout of recent changes

Clean up some of the impact of recent (and not so recent) scheduler
changes:

 - turning macros into nice inline functions
 - sanitizing and unifying variable definitions
 - whitespace, style consistency, 80-lines, comment correctness, spelling
   and curly braces police

Due to the macro hell and variable placement simplifications there's even 26
bytes of .text saved:

   text    data     bss     dec     hex filename
  25510    4153     192   29855    749f sched.o.before
  25484    4153     192   29829    7485 sched.o.after

[akpm@osdl.org: build fix]
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
17 years ago[PATCH] lockdep: irqtrace subsystem, move account_system_vtime() calls into kernel...
Paul Mackerras [Mon, 3 Jul 2006 07:25:40 +0000 (00:25 -0700)]
[PATCH] lockdep: irqtrace subsystem, move account_system_vtime() calls into kernel/softirq.c

At the moment, powerpc and s390 have their own versions of do_softirq which
include local_bh_disable() and __local_bh_enable() calls.  They end up
calling __do_softirq (in kernel/softirq.c) which also does
local_bh_disable/enable.

Apparently the two levels of disable/enable trigger a warning from some
validation code that Ingo is working on, and he would like to see the outer
level removed.  But to do that, we have to move the account_system_vtime
calls that are currently in the arch do_softirq() implementations for
powerpc and s390 into the generic __do_softirq() (this is a no-op for other
archs because account_system_vtime is defined to be an empty inline
function on all other archs).  This patch does that.

Signed-off-by: Paul Mackerras <paulus@samba.org>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
17 years ago[PATCH] lockdep: annotate forcedeth.c disable_irq()
Ingo Molnar [Mon, 3 Jul 2006 07:25:39 +0000 (00:25 -0700)]
[PATCH] lockdep: annotate forcedeth.c disable_irq()

nv_do_nic_poll() is called from timer softirqs, which has interrupts enabled,
but np->lock might also be taken by some other interrupt context.

The driver does disable_irq() to get around this problem, so annotate the
disable_irq()/enable_irq() calls for lockdep.

Has no effect on non-lockdep kernels.

Signed-off-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Arjan van de Ven <arjan@linux.intel.com>
Cc: Ayaz Abdulla <aabdulla@nvidia.com>
Cc: Manfred Spraul <manfred@colorfullife.com>
Cc: Jeff Garzik <jeff@garzik.org>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>