10 years agokobject: Cleanup kobject_rename and !CONFIG_SYSFS
Eric W. Biederman [Fri, 4 Jul 2008 01:05:28 +0000]
kobject: Cleanup kobject_rename and !CONFIG_SYSFS

It finally dawned on me what the clean fix to sysfs_rename_dir
calling kobject_set_name is.  Move the work into kobject_rename
where it belongs.  The callers serialize us anyway so this is
safe.

Signed-off-by: Eric W. Biederman <ebiederm@xmission.com>
Acked-by: Tejun Heo <tj@kernel.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

10 years agokobject: Fix kobject_rename and !CONFIG_SYSFS
Eric W. Biederman [Thu, 8 May 2008 21:41:00 +0000]
kobject: Fix kobject_rename and !CONFIG_SYSFS

When looking at kobject_rename I found two bugs with
that exist when sysfs support is disabled in the kernel.

kobject_rename does not change the name on the kobject when
sysfs support is not compiled in.

kobject_rename without locking attempts to check the
validity of a rename operation, which the kobject layer
simply does not have the infrastructure to do.

This patch documents the previously unstated requirement of
kobject_rename that is the responsibility of the caller to
provide mutual exclusion and to be certain that the new_name
for the kobject is valid.

This patch modifies sysfs_rename_dir in !CONFIG_SYSFS case
to call kobject_set_name to actually change the kobject_name.

This patch removes the bogus and misleading check in kobject_rename
that attempts to see if a rename is valid.  The check is bogus
because we do not have the proper locking.  The check is misleading
because it looks like we can and do perform checking at the kobject
level that we don't.

Signed-off-by: Eric W. Biederman <ebiederm@xmission.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

10 years agosysfs: Make dir and name args to sysfs_notify() const
Trent Piepho [Thu, 25 Sep 2008 23:45:13 +0000]
sysfs: Make dir and name args to sysfs_notify() const

Because they can be, and because code like this produces a warning if
they're not:

struct device_attribute dev_attr;

sysfs_notify(&kobj, NULL, dev_attr.attr.name);

Signed-off-by: Trent Piepho <tpiepho@freescale.com>
CC: Neil Brown <neilb@suse.de>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

10 years agoplatform: add new device registration helper
Dmitry Baryshkov [Mon, 22 Sep 2008 21:41:40 +0000]
platform: add new device registration helper

Add a helper that registers simple platform_device w/o resources but with
parent and device data.

This is usefull to cleanup platform code from code that registers such
simple devices as leds-gpio, generic-bl, etc.

Signed-off-by: Dmitry Baryshkov <dbaryshkov@gmail.com>
Cc: David Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

10 years agosysfs: use ilookup5() instead of ilookup5_nowait()
Tejun Heo [Sat, 27 Sep 2008 22:48:08 +0000]
sysfs: use ilookup5() instead of ilookup5_nowait()

As inode creation is protected by sysfs_mutex, ilookup5_nowait()
always either fails to find at all or finds one which is fully
initialized, so using ilookup5_nowait() or ilookup5() doesn't make any
difference.  Switch to ilookup5() as it's planned to be removed.  This
change also makes lookup return value handling a bit simpler.

This change was suggested by Al Viro.

Signed-off-by: Tejun Heo <tj@kernel.org>
Cc: Al Viro <viro@hera.kernel.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

10 years agoPNP: create device attributes via default device attributes
Drew Moseley [Sat, 27 Sep 2008 23:31:35 +0000]
PNP: create device attributes via default device attributes

This creates the attributes before the uevent is sent.

Signed-off-by: Drew Moseley <dmoseley@mvista.com>
Acked-by: Kay Sievers <kay.sievers@vrfy.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

10 years agoDriver core: make bus_find_device_by_name() more robust
Peter Korsgaard [Tue, 23 Sep 2008 20:01:27 +0000]
Driver core: make bus_find_device_by_name() more robust

Use sysfs_streq() in bus_find_device_by_name() so trailing newlines
are ignored (E.G. in bind/unbind).

Signed-off-by: Peter Korsgaard <jacmet@sunsite.dk>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

10 years agousb: turn dev_warn+WARN_ON combos into dev_WARN
Arjan van de Ven [Sun, 21 Sep 2008 02:09:37 +0000]
usb: turn dev_warn+WARN_ON combos into dev_WARN

dev_WARN is both compacter and gives better debug information
than just a WARN_ON, since people and tools will copy the device
information message together with the WARN_ON in bug reports.

Signed-off-by: Arjan van de Ven <arjan@linux.intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

10 years agodebug: use dev_WARN() rather than WARN_ON() in device_pm_add()
Arjan van de Ven [Sun, 21 Sep 2008 02:09:00 +0000]
debug: use dev_WARN() rather than WARN_ON() in device_pm_add()

device_pm_add() has a WARN_ON that is showing relatively high on
kerneloops.org, but unfortunately the WARN_ON is less than useful
in that it doesn't print any information about what device is causing
the issue.

This patch fixes this by turning the WARN_ON() into the newly
introduces dev_WARN() which will print information about the
device in question.

Signed-off-by: Arjan van de Ven <arjan@linux.intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

10 years agodebug: Introduce a dev_WARN() function
Arjan van de Ven [Sun, 21 Sep 2008 02:08:39 +0000]
debug: Introduce a dev_WARN() function

in the line of dev_printk(), this patch introduces a dev_WARN() function,
that takes a struct device and then a printk format/args set of arguments.
Unlike dev_printk(), the effect is that of WARN() in that a full warning
message (including filename/line, module list, versions and a backtrace)
is printed in addition to the device name and the arguments.

Signed-off-by: Arjan van de Ven <arjan@linux.intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

10 years agosysfs: fix deadlock
Nick Piggin [Fri, 12 Sep 2008 09:24:11 +0000]
sysfs: fix deadlock

On Thu, Sep 11, 2008 at 10:27:10AM +0200, Ingo Molnar wrote:

> and it's working fine on most boxes. One testbox found this new locking
> scenario:
>
> PM: Adding info for No Bus:vcsa7
> EDAC DEBUG: MC0: i82860_check()
>
> =======================================================
> [ INFO: possible circular locking dependency detected ]
> 2.6.27-rc6-tip #1
> -------------------------------------------------------
> X/4873 is trying to acquire lock:
>  (&bb->mutex){--..}, at: [<c020ba20>] mmap+0x40/0xa0
>
> but task is already holding lock:
>  (&mm->mmap_sem){----}, at: [<c0125a1e>] sys_mmap2+0x8e/0xc0
>
> which lock already depends on the new lock.
>
>
> the existing dependency chain (in reverse order) is:
>
> -> #1 (&mm->mmap_sem){----}:
>        [<c017dc96>] validate_chain+0xa96/0xf50
>        [<c017ef2b>] __lock_acquire+0x2cb/0x5b0
>        [<c017f299>] lock_acquire+0x89/0xc0
>        [<c01aa8fb>] might_fault+0x6b/0x90
>        [<c040b618>] copy_to_user+0x38/0x60
>        [<c020bcfb>] read+0xfb/0x170
>        [<c01c09a5>] vfs_read+0x95/0x110
>        [<c01c1443>] sys_pread64+0x63/0x80
>        [<c012146f>] sysenter_do_call+0x12/0x43
>        [<ffffffff>] 0xffffffff
>
> -> #0 (&bb->mutex){--..}:
>        [<c017d8b7>] validate_chain+0x6b7/0xf50
>        [<c017ef2b>] __lock_acquire+0x2cb/0x5b0
>        [<c017f299>] lock_acquire+0x89/0xc0
>        [<c0d6f2ab>] __mutex_lock_common+0xab/0x3c0
>        [<c0d6f698>] mutex_lock_nested+0x38/0x50
>        [<c020ba20>] mmap+0x40/0xa0
>        [<c01b111e>] mmap_region+0x14e/0x450
>        [<c01b170f>] do_mmap_pgoff+0x2ef/0x310
>        [<c0125a3d>] sys_mmap2+0xad/0xc0
>        [<c012146f>] sysenter_do_call+0x12/0x43
>        [<ffffffff>] 0xffffffff
>
> other info that might help us debug this:
>
> 1 lock held by X/4873:
>  #0:  (&mm->mmap_sem){----}, at: [<c0125a1e>] sys_mmap2+0x8e/0xc0
>
> stack backtrace:
> Pid: 4873, comm: X Not tainted 2.6.27-rc6-tip #1
>  [<c017cd09>] print_circular_bug_tail+0x79/0xc0
>  [<c017d8b7>] validate_chain+0x6b7/0xf50
>  [<c017a5b5>] ? trace_hardirqs_off_caller+0x15/0xb0
>  [<c017ef2b>] __lock_acquire+0x2cb/0x5b0
>  [<c017f299>] lock_acquire+0x89/0xc0
>  [<c020ba20>] ? mmap+0x40/0xa0
>  [<c0d6f2ab>] __mutex_lock_common+0xab/0x3c0
>  [<c020ba20>] ? mmap+0x40/0xa0
>  [<c0d6f698>] mutex_lock_nested+0x38/0x50
>  [<c020ba20>] ? mmap+0x40/0xa0
>  [<c020ba20>] mmap+0x40/0xa0
>  [<c01b111e>] mmap_region+0x14e/0x450
>  [<c01afb88>] ? arch_get_unmapped_area_topdown+0xf8/0x160
>  [<c01b170f>] do_mmap_pgoff+0x2ef/0x310
>  [<c0125a3d>] sys_mmap2+0xad/0xc0
>  [<c012146f>] sysenter_do_call+0x12/0x43
>  [<c0120000>] ? __switch_to+0x130/0x220
>  =======================
> evbug.c: Event. Dev: input3, Type: 20, Code: 0, Value: 500
> warning: `sudo' uses deprecated v2 capabilities in a way that may be insecure.
>
> i've attached the config.
>
> at first sight it looks like a genuine bug in fs/sysfs/bin.c?

Yes, it is a real bug by the looks. bin.c takes bb->mutex under mmap_sem
when it is mmapped, and then does its copy_*_user under bb->mutex too.

Here is a basic fix for the sysfs lor.

From: Nick Piggin <npiggin@suse.de>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

10 years agodevice model: Do a quickcheck for driver binding before doing an expensive check
Arjan van de Ven [Sun, 14 Sep 2008 15:32:06 +0000]
device model: Do a quickcheck for driver binding before doing an expensive check

This patch adds a quick check for the driver<->device match before
taking the locks and doin gthe expensive checks. Taking the lock hurts
in asynchronous boot context where the device lock gets hit; one of the
init functions takes the lock and goes to do an expensive hardware init;
the other init functions walk the same PCI list and get stuck on the
lock as a result.

For the common case, we can know there's no chance whatsoever of a match
if the device isn't in the drivers ID table... so this patch does that
check as a best-effort-avoid-the-lock approach.

Bootcharts for before and after can be seen at
http://www.fenrus.org/before.svg
http://www.fenrus.org/after.svg

Note the long time "agp_ali_init" takes in the first graph; my laptop
doesn't even have an ALI chip in it!  (the bootgraphs look a bit
dissimilar, but that's the point, the first one has a bunch of arbitrary
delays in it that cause it to look very different)

This reduces my kernel boot time by about 20%

Signed-off-by: Arjan van de Ven <arjan@linux.intel.com>
Cc: Alan Stern <stern@rowland.harvard.edu>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

10 years agoDriver core: Fix cleanup in device_create_vargs().
Cornelia Huck [Wed, 3 Sep 2008 16:26:41 +0000]
Driver core: Fix cleanup in device_create_vargs().

If device_register() in device_create_vargs() fails, the device
must be cleaned up with put_device() (which is also fine on NULL)
instead of kfree().

Signed-off-by: Cornelia Huck <cornelia.huck@de.ibm.com>
Cc: stable <stable@kernel.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

10 years agoDriver core: Clarify device cleanup.
Cornelia Huck [Wed, 3 Sep 2008 16:26:40 +0000]
Driver core: Clarify device cleanup.

Make the comments on how to use device_initialize(), device_add()
and device_register() a bit clearer - in particular, explicitly
note that put_device() must be used once we tried to add the device
to the hierarchy.

Signed-off-by: Cornelia Huck <cornelia.huck@de.ibm.com>
Cc: stable <stable@kernel.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

10 years agodrivers/firmware/iscsi_ibft.c: make 3 functions static
Adrian Bunk [Tue, 12 Aug 2008 19:43:59 +0000]
drivers/firmware/iscsi_ibft.c: make 3 functions static

This patch makes the following needlessly global functions static:
- ibft_attr_show_initiator()
- ibft_attr_show_nic()
- ibft_attr_show_target()

Signed-off-by: Adrian Bunk <bunk@kernel.org>
Signed-off-by: Konrad Rzeszutek <ketuzsezr@darnok.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

10 years agosysfs: Support sysfs_notify from atomic context with new sysfs_notify_dirent
Neil Brown [Tue, 15 Jul 2008 22:58:04 +0000]
sysfs: Support sysfs_notify from atomic context with new sysfs_notify_dirent

Support sysfs_notify from atomic context with new sysfs_notify_dirent

sysfs_notify currently takes sysfs_mutex.
This means that it cannot be called in atomic context.
sysfs_mutex  is sometimes held over a malloc (sysfs_rename_dir)
so it can block on low memory.

In md I want to be able to notify on a sysfs attribute from
atomic context, and I don't want to block on low memory because I
could be in the writeout path for freeing memory.

So:
 - export the "sysfs_dirent" structure along with sysfs_get, sysfs_put
   and sysfs_get_dirent so I can get the sysfs_dirent that I want to
   notify on and hold it in an md structure.
 - split sysfs_notify_dirent out of sysfs_notify so the sysfs_dirent
   can be notified on with no blocking (just a spinlock).

Signed-off-by: Neil Brown <neilb@suse.de>
Acked-by: Tejun Heo <tj@kernel.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

10 years agodriver core: make struct platform_pm_ops static
Adrian Bunk [Mon, 21 Jul 2008 19:33:36 +0000]
driver core: make struct platform_pm_ops static

This patch makes the needlessly global struct platform_pm_ops static.

Signed-off-by: Adrian Bunk <bunk@kernel.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

10 years agodriver core: basic infrastructure for per-module dynamic debug messages
Jason Baron [Tue, 12 Aug 2008 20:46:19 +0000]
driver core: basic infrastructure for per-module dynamic debug messages

Base infrastructure to enable per-module debug messages.

I've introduced CONFIG_DYNAMIC_PRINTK_DEBUG, which when enabled centralizes
control of debugging statements on a per-module basis in one /proc file,
currently, <debugfs>/dynamic_printk/modules. When, CONFIG_DYNAMIC_PRINTK_DEBUG,
is not set, debugging statements can still be enabled as before, often by
defining 'DEBUG' for the proper compilation unit. Thus, this patch set has no
affect when CONFIG_DYNAMIC_PRINTK_DEBUG is not set.

The infrastructure currently ties into all pr_debug() and dev_dbg() calls. That
is, if CONFIG_DYNAMIC_PRINTK_DEBUG is set, all pr_debug() and dev_dbg() calls
can be dynamically enabled/disabled on a per-module basis.

Future plans include extending this functionality to subsystems, that define
their own debug levels and flags.

Usage:

Dynamic debugging is controlled by the debugfs file,
<debugfs>/dynamic_printk/modules. This file contains a list of the modules that
can be enabled. The format of the file is as follows:

<module_name> <enabled=0/1>
.
.
.

<module_name> : Name of the module in which the debug call resides
<enabled=0/1> : whether the messages are enabled or not

For example:

snd_hda_intel enabled=0
fixup enabled=1
driver enabled=0

Enable a module:

$echo "set enabled=1 <module_name>" > dynamic_printk/modules

Disable a module:

$echo "set enabled=0 <module_name>" > dynamic_printk/modules

Enable all modules:

$echo "set enabled=1 all" > dynamic_printk/modules

Disable all modules:

$echo "set enabled=0 all" > dynamic_printk/modules

Finally, passing "dynamic_printk" at the command line enables
debugging for all modules. This mode can be turned off via the above
disable command.

[gkh: minor cleanups and tweaks to make the build work quietly]

Signed-off-by: Jason Baron <jbaron@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

10 years agousb gadget: link fixes for network gadget
David Brownell [Tue, 19 Aug 2008 00:45:07 +0000]
usb gadget: link fixes for network gadget

Change how the Ethernet/RNDIS gadget driver builds:  don't
use separate compilation, since it works poorly when key
parts are library code (with init sections etc).  Instead
be as close as we can to "gcc --combine ...".

This is a bit more complicated than most of the others
because it had to resolve a few symbol collisions.

Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

10 years agousb gadget: link fixes for cdc composite gadget
David Brownell [Tue, 19 Aug 2008 00:43:56 +0000]
usb gadget: link fixes for cdc composite gadget

Change how the CDC Composite gadget driver builds:  don't
use separate compilation, since it works poorly when key
parts are library code (with init sections etc).  Instead
be as close as we can to "gcc --combine ...".

Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

10 years agousb gadget: link fixes for storage gadget
David Brownell [Tue, 19 Aug 2008 00:43:25 +0000]
usb gadget: link fixes for storage gadget

Change how the file storage gadget driver builds:  don't
use separate compilation, since it works poorly when key
parts are library code (with init sections etc).  Instead
be as close as we can to "gcc --combine ...".

Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

10 years agousb gadget: link fixes for printer gadget
David Brownell [Tue, 19 Aug 2008 00:42:49 +0000]
usb gadget: link fixes for printer gadget

Change how the printer gadget driver builds:  don't use
separate compilation, since it works poorly when key parts
are library code (with init sections etc).  Instead be as
close as we can to "gcc --combine ...".

Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

10 years agousb gadget: link fixes for MIDI gadget
David Brownell [Tue, 19 Aug 2008 00:42:04 +0000]
usb gadget: link fixes for MIDI gadget

Change how the MIDI gadget driver builds:  don't use separate
compilation, since it works poorly when key parts are library
code (with init sections etc).  Instead be as close as we can
to "gcc --combine ...".

Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

10 years agousb gadget: link fixes for gadget zero
David Brownell [Tue, 19 Aug 2008 00:41:31 +0000]
usb gadget: link fixes for gadget zero

Change how the Gadget Zero driver builds:  don't use
separate compilation, since it works poorly when key
parts are library code (with init sections etc).
Instead be as close as we can to "gcc --combine ...".

Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

10 years agousb gadget: link fixes for serial gadget
David Brownell [Tue, 19 Aug 2008 00:41:02 +0000]
usb gadget: link fixes for serial gadget

Change how the serial gadget driver builds:  don't use
separate compilation, since it works poorly when key parts
are library code (with init sections etc).  Instead be as
close as we can to "gcc --combine ...".

Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

10 years agodevice create: remove device_create_drvdata
Greg Kroah-Hartman [Tue, 22 Jul 2008 03:03:34 +0000]
device create: remove device_create_drvdata

Now that the tree is cleaned up, device_create_drvdata can be safely
removed.

Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

10 years agodevice create: ide: convert device_create_drvdata to device_create
Greg Kroah-Hartman [Tue, 22 Jul 2008 03:03:34 +0000]
device create: ide: convert device_create_drvdata to device_create

Now that device_create() has been audited, rename things back to the
original call to be sane.

Cc: Bartlomiej Zolnierkiewicz <bzolnier@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

10 years agodevice create: video: convert device_create_drvdata to device_create
Greg Kroah-Hartman [Tue, 22 Jul 2008 03:03:34 +0000]
device create: video: convert device_create_drvdata to device_create

Now that device_create() has been audited, rename things back to the
original call to be sane.

Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

10 years agodevice create: usb: convert device_create_drvdata to device_create
Greg Kroah-Hartman [Tue, 22 Jul 2008 03:03:34 +0000]
device create: usb: convert device_create_drvdata to device_create

Now that device_create() has been audited, rename things back to the
original call to be sane.

Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

10 years agodevice create: sound: convert device_create_drvdata to device_create
Greg Kroah-Hartman [Tue, 22 Jul 2008 03:03:34 +0000]
device create: sound: convert device_create_drvdata to device_create

Now that device_create() has been audited, rename things back to the
original call to be sane.

Cc: Jaroslav Kysela <perex@perex.cz>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

10 years agodevice create: scsi: convert device_create_drvdata to device_create
Greg Kroah-Hartman [Tue, 22 Jul 2008 03:03:34 +0000]
device create: scsi: convert device_create_drvdata to device_create

Now that device_create() has been audited, rename things back to the
original call to be sane.

Cc: James E.J. Bottomley <James.Bottomley@HansenPartnership.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

10 years agodevice create: s390: convert device_create_drvdata to device_create
Greg Kroah-Hartman [Tue, 22 Jul 2008 03:03:34 +0000]
device create: s390: convert device_create_drvdata to device_create

Now that device_create() has been audited, rename things back to the
original call to be sane.

Cc: Martin Schwidefsky <schwidefsky@de.ibm.com>
Cc: Heiko Carstens <heiko.carstens@de.ibm.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

10 years agodevice create: net: convert device_create_drvdata to device_create
Greg Kroah-Hartman [Tue, 22 Jul 2008 03:03:34 +0000]
device create: net: convert device_create_drvdata to device_create

Now that device_create() has been audited, rename things back to the
original call to be sane.

Cc: Jeff Garzik <jgarzik@pobox.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

10 years agodevice create: misc: convert device_create_drvdata to device_create
Greg Kroah-Hartman [Tue, 22 Jul 2008 03:03:34 +0000]
device create: misc: convert device_create_drvdata to device_create

Now that device_create() has been audited, rename things back to the
original call to be sane.

Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

10 years agodevice create: infiniband: convert device_create_drvdata to device_create
Greg Kroah-Hartman [Tue, 22 Jul 2008 03:03:34 +0000]
device create: infiniband: convert device_create_drvdata to device_create

Now that device_create() has been audited, rename things back to the
original call to be sane.

Cc: Roland Dreier <rolandd@cisco.com>
Cc: Sean Hefty <sean.hefty@intel.com>
Cc: Hal Rosenstock <hal.rosenstock@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

10 years agodevice create: ieee1394: convert device_create_drvdata to device_create
Greg Kroah-Hartman [Tue, 22 Jul 2008 03:03:34 +0000]
device create: ieee1394: convert device_create_drvdata to device_create

Now that device_create() has been audited, rename things back to the
original call to be sane.

Cc: Ben Collins <ben.collins@ubuntu.com>
Acked-by: Stefan Richter <stefanr@s5r6.in-berlin.de>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

10 years agodevice create: char: convert device_create_drvdata to device_create
Greg Kroah-Hartman [Tue, 22 Jul 2008 03:03:34 +0000]
device create: char: convert device_create_drvdata to device_create

Now that device_create() has been audited, rename things back to the
original call to be sane.

Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

10 years agodevice create: block: convert device_create_drvdata to device_create
Greg Kroah-Hartman [Tue, 22 Jul 2008 03:03:34 +0000]
device create: block: convert device_create_drvdata to device_create

Now that device_create() has been audited, rename things back to the
original call to be sane.

Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

10 years agosysfs: crash debugging
Andrew Morton [Fri, 24 Aug 2007 23:11:54 +0000]
sysfs: crash debugging

Print the name of the last-accessed sysfs file when we oops, to help track
down oopses which occur in sysfs store/read handlers.  Because these oopses
tend to not leave any trace of the offending code in the stack traces.

Cc: Kay Sievers <kay.sievers@vrfy.org>
Cc: Mathieu Desnoyers <mathieu.desnoyers@polymtl.ca>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

10 years agomodules: fix module "notes" kobject leak
Alexey Dobriyan [Tue, 23 Sep 2008 19:51:11 +0000]
modules: fix module "notes" kobject leak

Fix "notes" kobject leak

It happens every rmmod if KALLSYMS=y and SYSFS=y.

# modprobe foo

kobject: 'foo' (ffffffffa00743d0): kobject_add_internal: parent: 'module', set: 'module'
kobject: 'holders' (ffff88017e7c5770): kobject_add_internal: parent: 'foo', set: '<NULL>'
kobject: 'foo' (ffffffffa00743d0): kobject_uevent_env
kobject: 'foo' (ffffffffa00743d0): fill_kobj_path: path = '/module/foo'
kobject: 'notes' (ffff88017fa9b668): kobject_add_internal: parent: 'foo', set: '<NULL>'
  ^^^^^

# rmmod foo

kobject: 'holders' (ffff88017e7c5770): kobject_cleanup
kobject: 'holders' (ffff88017e7c5770): auto cleanup kobject_del
kobject: 'holders' (ffff88017e7c5770): calling ktype release
kobject: (ffff88017e7c5770): dynamic_kobj_release
kobject: 'holders': free name
kobject: 'foo' (ffffffffa00743d0): kobject_cleanup
kobject: 'foo' (ffffffffa00743d0): does not have a release() function, it is broken and must be fixed.
kobject: 'foo' (ffffffffa00743d0): auto cleanup 'remove' event
kobject: 'foo' (ffffffffa00743d0): kobject_uevent_env
kobject: 'foo' (ffffffffa00743d0): fill_kobj_path: path = '/module/foo'
kobject: 'foo' (ffffffffa00743d0): auto cleanup kobject_del
kobject: 'foo': free name

[whooops]

Signed-off-by: Alexey Dobriyan <adobriyan@gmail.com>
Cc: stable <stable@kernel.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

10 years agoMerge branch 'upstream' of git://ftp.linux-mips.org/pub/scm/upstream-linus
Linus Torvalds [Wed, 15 Oct 2008 17:22:21 +0000]
Merge branch 'upstream' of git://ftp.linux-mips.org/upstream-linus

* 'upstream' of git://ftp.linux-mips.org/pub/scm/upstream-linus:
  MIPS: Kill unused <asm/debug.h> inclusions
  MIPS: IP32: Add platform device for CMOS RTC; remove dead code
  RTC: M48T35: new RTC driver
  MIPS: IP27: Switch over to RTC class driver
  MIPS: DS1286: New RTC driver
  MIPS: IP22/28: Switch over to RTC class driver
  MIPS: PCI: Scan busses when they are registered
  MIPS: WGT634U: Add reset button support
  MIPS: BCM47xx: Use the new SSB GPIO API
  MIPS: BCM47xx: Remove references to BCM947XX
  MIPS: WGT634U: Add machine detection message
  MIPS: Align .data.cacheline_aligned based on CONFIG_MIPS_L1_CACHE_SHIFT
  MIPS: show_cpuinfo prints the type of the calling CPU
  MIPS: Fix wrong branch target in new spin_lock code.
  MIPS: Have a heart for a lonely, lost header file ...

10 years agotty: make sure that proc_clear_tty stores the cpu flags
Arjan van de Ven [Wed, 15 Oct 2008 09:52:34 +0000]
tty: make sure that proc_clear_tty stores the cpu flags

proc_clear_tty() gets called with interrupts off (while holding the task list
lock) from sys_setid. This means that it needs the _irqsave version of the
locking primitives.

Reported-by: Marcin Slusarz <marcin.slusarz@gmail.com>
Tested-by: Marcin Slusarz <marcin.slusarz@gmail.com>
Signed-off-by: Arjan van de Ven <arjan@linux.intel.com>
Signed-off-by: Alan Cox <alan@redhat.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

10 years agometronomefb: Fix warning when building 64bit
Alan Cox [Wed, 15 Oct 2008 09:50:32 +0000]
metronomefb: Fix warning when building 64bit

The metronome driver produces warnings when built on x86-64 as it assumes that
size_t is an int. Use %Zd instead.

Signed-off-by: Alan Cox <alan@redhat.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

10 years agoxfs: fix remount rw with unrecognized options
Christoph Hellwig [Sun, 12 Oct 2008 12:30:44 +0000]
xfs: fix remount rw with unrecognized options

When we skip unrecognized options in xfs_fs_remount we should just break
out of the switch and not return because otherwise we may skip clearing
the xfs-internal read-only flag.  This will only show up on some
operations like touch because most read-only checks are done by the VFS
which thinks this filesystem is r/w.  Eventually we should replace the
XFS read-only flag with a helper that always checks the VFS flag to make
sure they can never get out of sync.

Bug reported and fix verified by Marcel Beister on #xfs.
Bug fix verified by updated xfstests/189.

Signed-off-by: Christoph Hellwig <hch@lst.de>
Acked-by: Eric Sandeen <sandeen@sandeen.net>
Signed-off-by: Timothy Shimmin <tes@sgi.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

10 years agoMerge branch 'build_fix' of git://git.kernel.org/pub/scm/linux/kernel/git/mfasheh...
Linus Torvalds [Wed, 15 Oct 2008 15:08:57 +0000]
Merge branch 'build_fix' of git://git./linux/kernel/git/mfasheh/ocfs2

* 'build_fix' of git://git.kernel.org/pub/scm/linux/kernel/git/mfasheh/ocfs2:
  ocfs2: fix build error

10 years agoMerge branch 'next' of git://git.kernel.org/pub/scm/linux/kernel/git/benh/powerpc
Linus Torvalds [Wed, 15 Oct 2008 15:07:35 +0000]
Merge branch 'next' of git://git./linux/kernel/git/benh/powerpc

* 'next' of git://git.kernel.org/pub/scm/linux/kernel/git/benh/powerpc: (158 commits)
  powerpc: Fix CHRP PCI config access for indirect_pci
  powerpc/chrp: Fix detection of Python PCI host bridge on IBM CHRPs
  powerpc: Fix 32-bit SMP boot on CHRP
  powerpc: Fix link errors on 32-bit machines using legacy DMA
  powerpc/pci: Improve detection of unassigned bridge resources
  hvc_console: Fix free_irq in spinlocked section
  powerpc: Get USE_STRICT_MM_TYPECHECKS working again
  powerpc: Reflect the used arguments in machine_init() prototype
  powerpc: Fix DMA offset for non-coherent DMA
  powerpc: fix fsl_upm nand driver modular build
  powerpc/83xx: add NAND support for the MPC8360E-RDK boards
  powerpc: FPGA support for GE Fanuc SBC610
  i2c: MPC8349E-mITX Power Management and GPIO expander driver
  powerpc: reserve two DMA channels for audio in MPC8610 HPCD device tree
  powerpc: document the "fsl,ssi-dma-channel" compatible property
  powerpc: disable CHRP and PMAC support in various defconfigs
  OF: add fsl,mcu-mpc8349emitx to the exception list
  powerpc/83xx: add DS1374 RTC support for the MPC837xE-MDS boards
  powerpc: remove support for bootmem-allocated memory for the DIU driver
  powerpc: remove non-dependent load fsl_booke PTE_64BIT
  ...

10 years agoMIPS: Kill unused <asm/debug.h> inclusions
Shinya Kuribayashi [Sun, 12 Oct 2008 13:56:11 +0000]
MIPS: Kill unused <asm/debug.h> inclusions

Signed-off-by: Shinya Kuribayashi <skuribay@ruby.dti.ne.jp>
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>

10 years agoMIPS: IP32: Add platform device for CMOS RTC; remove dead code
Thomas Bogendoerfer [Mon, 22 Sep 2008 22:53:20 +0000]
MIPS: IP32: Add platform device for CMOS RTC; remove dead code

Signed-off-by: Thomas Bogendoerfer <tsbogend@alpha.franken.de>
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>

10 years agoRTC: M48T35: new RTC driver
Thomas Bogendoerfer [Tue, 14 Oct 2008 15:17:32 +0000]
RTC: M48T35: new RTC driver

This driver replaces the broken ip27-rtc driver in drivers/char and
gives back RTC support for SGI IP27 machines.

Signed-off-by: Thomas Bogendoerfer <tsbogend@alpha.franken.de>
Acked-by: Alessandro Zummo <alessandro.zummo@towertech.it>
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>

10 years agoMIPS: IP27: Switch over to RTC class driver
Thomas Bogendoerfer [Tue, 14 Oct 2008 15:17:28 +0000]
MIPS: IP27: Switch over to RTC class driver

This patchset removes some dead code and creates a platform device
for the RTC class driver.

Signed-off-by: Thomas Bogendoerfer <tsbogend@alpha.franken.de>
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>

10 years agoMIPS: DS1286: New RTC driver
Thomas Bogendoerfer [Tue, 14 Oct 2008 15:16:59 +0000]
MIPS: DS1286: New RTC driver

This driver replaces the broken DS1286 driver in drivers/char and gives back
RTC support for SGI IP22 and IP28 machines.

Signed-off-by: Thomas Bogendoerfer <tsbogend@alpha.franken.de>
Acked-by: Alessandro Zummo <alessandro.zummo@towertech.it>
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>

10 years agoMIPS: IP22/28: Switch over to RTC class driver
Thomas Bogendoerfer [Tue, 14 Oct 2008 15:16:55 +0000]
MIPS: IP22/28: Switch over to RTC class driver

This patchset removes some dead code and creates a platform device
for the RTC class driver.

Signed-off-by: Thomas Bogendoerfer <tsbogend@alpha.franken.de>
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>

10 years agoMIPS: PCI: Scan busses when they are registered
Aurelien Jarno [Tue, 14 Oct 2008 09:45:09 +0000]
MIPS: PCI: Scan busses when they are registered

The patch below changes register_pci_controller() such that controllers
being added after pcibios_init() has run are be scanned immediately.

This is needed for example by the BCM47xx PCI controller, which is located
on the SSB bus, which is now initialized after the PCI subsystem.

Signed-off-by: Aurelien Jarno <aurelien@aurel32.net>
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>

10 years agoMIPS: WGT634U: Add reset button support
Aurelien Jarno [Tue, 14 Oct 2008 09:44:43 +0000]
MIPS: WGT634U: Add reset button support

This patch adds support for the reset button of WGT634U machine, using
GPIO interrupts. Based on a patch from Michel Lespinasse.

Signed-off-by: Aurelien Jarno <aurelien@aurel32.net>
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>

10 years agoMIPS: BCM47xx: Use the new SSB GPIO API
Aurelien Jarno [Tue, 14 Oct 2008 09:44:26 +0000]
MIPS: BCM47xx: Use the new SSB GPIO API

This patch simplifies the BCM47xx GPIO code by using the new SSB GPIO
API, which does a lot things that were implemented directly in the
BCM47xx code.

Signed-off-by: Aurelien Jarno <aurelien@aurel32.net>
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>

10 years agoMIPS: BCM47xx: Remove references to BCM947XX
Aurelien Jarno [Tue, 14 Oct 2008 09:43:47 +0000]
MIPS: BCM47xx: Remove references to BCM947XX

This patch removes the remaining reference to the BCM947xx development
board codename.

Signed-off-by: Florian Fainelli <florian.fainelli@telecomint.eu>
Signed-off-by: Aurelien Jarno <aurelien@aurel32.net>
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>

10 years agoMIPS: WGT634U: Add machine detection message
Aurelien Jarno [Tue, 14 Oct 2008 09:42:10 +0000]
MIPS: WGT634U: Add machine detection message

This adds a printk message when a WGT634U machine is detected.

Signed-off-by: Aurelien Jarno <aurelien@aurel32.net>
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>

10 years agoMIPS: Align .data.cacheline_aligned based on CONFIG_MIPS_L1_CACHE_SHIFT
David Daney [Mon, 13 Oct 2008 23:36:13 +0000]
MIPS: Align .data.cacheline_aligned based on CONFIG_MIPS_L1_CACHE_SHIFT

Signed-off-by: David Daney <ddaney@caviumnetworks.com>
Signed-off-by: Tomaso Paoletti <tpaoletti@caviumnetworks.com>
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>

10 years agoMIPS: show_cpuinfo prints the type of the calling CPU
Johannes Dickgreber [Mon, 13 Oct 2008 17:36:21 +0000]
MIPS: show_cpuinfo prints the type of the calling CPU

It should print the type of the Nth processor.

Signed-off-by: Johannes Dickgreber <tanzy@gmx.de>
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>

10 years agoMIPS: Fix wrong branch target in new spin_lock code.
Johannes Dickgreber [Mon, 13 Oct 2008 17:33:32 +0000]
MIPS: Fix wrong branch target in new spin_lock code.

Signed-off-by: Johannes Dickgreber <tanzy@gmx.de>
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>

10 years agoMIPS: Have a heart for a lonely, lost header file ...
Ralf Baechle [Mon, 13 Oct 2008 12:41:55 +0000]
MIPS: Have a heart for a lonely, lost header file ...

... and move it to where all its brothers and sisters reside.  Requested by
Shane McDonald <mcdonald.shane@gmail.com>.

Signed-off-by: Ralf Baechle <ralf@linux-mips.org>

10 years agoocfs2: fix build error
Mark Fasheh [Wed, 15 Oct 2008 01:31:46 +0000]
ocfs2: fix build error

I merged the latest ocfs2_read_blocks() changes in xattr.c wrong. This makes
Ocfs2 compile again.

Signed-off-by: Mark Fasheh <mfasheh@suse.com>

10 years agoMerge commit 'origin'
Benjamin Herrenschmidt [Wed, 15 Oct 2008 00:31:54 +0000]
Merge commit 'origin'

Manual fixup of conflicts on:

arch/powerpc/include/asm/dcr-regs.h
drivers/net/ibm_newemac/core.h

10 years agoMerge branch 'i2c-for-linus' of git://jdelvare.pck.nerim.net/jdelvare-2.6
Linus Torvalds [Tue, 14 Oct 2008 23:53:02 +0000]
Merge branch 'i2c-for-linus' of git://jdelvare.pck.nerim.net/jdelvare-2.6

* 'i2c-for-linus' of git://jdelvare.pck.nerim.net/jdelvare-2.6:
  i2c-viapro: Add support for SMBus Process Call transactions
  i2c: Restore i2c_smbus_process_call function
  i2c: Do earlier driver model init
  i2c: Only build Tyan SMBus mux drivers on x86
  i2c: Guard against oopses from bad init sequences
  i2c: Document the implementation details of the /dev interface
  i2c: Improve dev-interface documentation
  i2c-parport-light: Don't register a platform device resource
  hwmon: (dme1737) Convert to a new-style i2c driver
  hwmon: (dme1737) Be less i2c-centric
  i2c/tps65010: Vibrator hookup to gpiolib
  i2c-viapro: Add VX800/VX820 support
  i2c: Renesas Highlander FPGA SMBus support
  i2c-pca-isa: Don't grab arbitrary resources
  i2c/isp1301_omap: Convert to a new-style i2c driver, part 2
  i2c/isp1301_omap: Convert to a new-style i2c driver, part 1

10 years agoMerge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/jikos/hid
Linus Torvalds [Tue, 14 Oct 2008 23:35:43 +0000]
Merge branch 'for-linus' of git://git./linux/kernel/git/jikos/hid

* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/jikos/hid: (55 commits)
  HID: build drivers for all quirky devices by default
  HID: add missing blacklist entry for Apple ATV ircontrol
  HID: add support for Bright ABNT2 brazilian device
  HID: Don't let Avermedia Radio FM800 be handled by usb hid drivers
  HID: fix numlock led on Dell device 0x413c/0x2105
  HID: remove warn() macro from usb hid drivers
  HID: remove info() macro from usb HID drivers
  HID: add appletv IR receiver quirk
  HID: fix a lockup regression when using force feedback on a PID device
  HID: hiddev.h: Fix example code.
  HID: hiddev.h: Fix mixed space and tabs in example code.
  HID: convert to dev_* prints
  HID: remove hid-ff
  HID: move zeroplus FF processing
  HID: move thrustmaster FF processing
  HID: move pantherlord FF processing
  HID: fix incorrent length condition in hidraw_write()
  HID: fix tty<->hid deadlock
  HID: ignore iBuddy devices
  HID: report descriptor fix for remaining MacBook JIS keyboards
  ...

10 years agoMerge branch 'upstream-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/mfashe...
Linus Torvalds [Tue, 14 Oct 2008 23:34:11 +0000]
Merge branch 'upstream-linus' of git://git./linux/kernel/git/mfasheh/ocfs2

* 'upstream-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/mfasheh/ocfs2: (56 commits)
  ocfs2: Make cached block reads the common case.
  ocfs2: Kill the last naked wait_on_buffer() for cached reads.
  ocfs2: Move ocfs2_bread() into dir.c
  ocfs2: Simplify ocfs2_read_block()
  ocfs2: Require an inode for ocfs2_read_block(s)().
  ocfs2: Separate out sync reads from ocfs2_read_blocks()
  ocfs2: Refactor xattr list and remove ocfs2_xattr_handler().
  ocfs2: Calculate EA hash only by its suffix.
  ocfs2: Move trusted and user attribute support into xattr.c
  ocfs2: Uninline ocfs2_xattr_name_hash()
  ocfs2: Don't check for NULL before brelse()
  ocfs2: use smaller counters in ocfs2_remove_xattr_clusters_from_cache
  ocfs2: Documentation update for user_xattr / nouser_xattr mount options
  ocfs2: make la_debug_mutex static
  ocfs2: Remove pointless !!
  ocfs2: Add empty bucket support in xattr.
  ocfs2/xattr.c: Fix a bug when inserting xattr.
  ocfs2: Add xattr mount option in ocfs2_show_options()
  ocfs2: Switch over to JBD2.
  ocfs2: Add the 'inode64' mount option.
  ...

10 years agortc-cmos: look for PNP RTC first, then for platform RTC
Bjorn Helgaas [Tue, 14 Oct 2008 23:01:59 +0000]
rtc-cmos: look for PNP RTC first, then for platform RTC

We shouldn't rely on "pnp_platform_devices" to tell us whether there
is a PNP RTC device.

I introduced "pnp_platform_devices", but I think it was a mistake.
All it tells us is whether we found any PNPBIOS or PNPACPI devices.
Many machines have some PNP devices, but do not describe the RTC
via PNP.  On those machines, we need to do the platform driver probe
to find the RTC.

We should just register the PNP driver and see whether it claims anything.
If we don't find a PNP RTC, fall back to the platform driver probe.

This (in conjunction with the arch/x86/kernel/rtc.c patch to add
a platform RTC device when PNP doesn't have one) should resolve
these issues:

    http://bugzilla.kernel.org/show_bug.cgi?id=11580
    https://bugzilla.redhat.com/show_bug.cgi?id=451188

Signed-off-by: Bjorn Helgaas <bjorn.helgaas@hp.com>
Acked-by: Rafael J. Wysocki <rjw@sisk.pl>
Acked-by: David Brownell <dbrownell@users.sourceforge.net>
Reported-by: Rik Theys <rik.theys@esat.kuleuven.be>
Reported-by: shr_msn@yahoo.com.tw
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

10 years agox86: register a platform RTC device if PNP doesn't describe it
Bjorn Helgaas [Tue, 14 Oct 2008 23:01:03 +0000]
x86: register a platform RTC device if PNP doesn't describe it

Most if not all x86 platforms have an RTC device, but sometimes the RTC
is not exposed as a PNP0b00/PNP0b01/PNP0b02 device in PNPBIOS or ACPI:

    http://bugzilla.kernel.org/show_bug.cgi?id=11580
    https://bugzilla.redhat.com/show_bug.cgi?id=451188

It's best if we can discover the RTC via PNP because then we know
which flavor of device it is, where it lives, and which IRQ it uses.

But if we can't, we should register a platform device using the
compiled-in RTC_PORT/RTC_IRQ resource assumptions.

Signed-off-by: Bjorn Helgaas <bjorn.helgaas@hp.com>
Acked-by: Rafael J. Wysocki <rjw@sisk.pl>
Acked-by: David Brownell <dbrownell@users.sourceforge.net>
Reported-by: Rik Theys <rik.theys@esat.kuleuven.be>
Reported-by: shr_msn@yahoo.com.tw
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

10 years agopowerpc: Fix CHRP PCI config access for indirect_pci
Benjamin Herrenschmidt [Mon, 13 Oct 2008 20:49:47 +0000]
powerpc: Fix CHRP PCI config access for indirect_pci

Recently, indirect_pci was changed to test if the bus number requested
is the one hanging straight off the PHB, then it substitutes the bus
number with another one contained in a new "self_busno" field of the
pci_controller structure.

However, this breaks CHRP which didn't initialize this new field, and
which relies on having the right bus number passed to the hardware.

This fixes it by initializing this variable properly for all CHRP bridges

Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>

10 years agopowerpc/chrp: Fix detection of Python PCI host bridge on IBM CHRPs
Benjamin Herrenschmidt [Mon, 13 Oct 2008 20:14:09 +0000]
powerpc/chrp: Fix detection of Python PCI host bridge on IBM CHRPs

The detection of the IBM "Python" PCI host bridge on IBM CHRP
machines such as old RS6000 was broken when we changed
of_device_is_compatible() from strncasecmp to strcasecmp (dropped
the "n" variant) due to the way IBM encodes the chip version.

We fix that by instead doing a match on the model property like
we do for others bridges in that file. It should be good enough
for those machines. If yours is still broken, let me know.

Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>

10 years agopowerpc: Fix 32-bit SMP boot on CHRP
Benjamin Herrenschmidt [Mon, 13 Oct 2008 18:38:48 +0000]
powerpc: Fix 32-bit SMP boot on CHRP

prom_init was changed to take a new argument, the address
where the kernel is loaded, which is now used to copy the
SMP spin loop down before use.

However, only head_64.S was adapted to pass this new value,
not head_32.S, thus breaking SMP boot on 32-bit SMP CHRP
machines.

Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>

10 years agopowerpc: Fix link errors on 32-bit machines using legacy DMA
Benjamin Herrenschmidt [Mon, 13 Oct 2008 17:51:46 +0000]
powerpc: Fix link errors on 32-bit machines using legacy DMA

The new merged DMA code will try to access isa_bridge_pcidev when
trying to DMA to/from legacy devices. This is however only defined
on 64-bit. Fixes this for now by adding the variable, even if it
stays NULL. In the long run, we'll make isa-bridge.c common to
32 and 64-bit.

Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>

10 years agopowerpc/pci: Improve detection of unassigned bridge resources
Benjamin Herrenschmidt [Mon, 13 Oct 2008 13:56:31 +0000]
powerpc/pci: Improve detection of unassigned bridge resources

When the powerpc PCI layer is not configured to re-assign everything,
it currently fails to detect that a PCI to PCI bridge has been left
unassigned by the firmware and tries to allocate resource for the
default window values in the bridge (0...X) (with the notable exception
of a hack we have in there that detects some Apple firmware unassigned
bridge resources).

This results in resource allocation failures, which are generally
fixed up later on but it causes scary warnings in the logs and we
have seen the fixup code fall over in some circumstances (a different
issue to fix as well).

This code improves that by providing a more complete & useful function
to intuit that a bridge was left unassigned by the firmware, and thus
force a full re-allocation by the PCI code without trying to allocate
the existing useless resources first.

The algorithm we use basically considers unassigned a window that
starts at 0 (PCI address) if the corresponding address space enable
bit is not set. In addition, for memory space, it considers such a
resource unassigned also if the host bridge isn't configured to
forward cycles to address 0 (ie, the resource basically overlaps
main memory).

This fixes a range of problems with things like Bare-Metal support
on pSeries machines, or attempt to use partial firmware PCI setup.

Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>

10 years agohvc_console: Fix free_irq in spinlocked section
Christian Borntraeger [Sun, 12 Oct 2008 21:51:31 +0000]
hvc_console: Fix free_irq in spinlocked section

    commit 611e097d7707741a336a0677d9d69bec40f29f3d
    Author: Christian Borntraeger <borntraeger@de.ibm.com>
    hvc_console: rework setup to replace irq functions with callbacks
    introduced a spinlock recursion problem. The notifier_del is
    called with a lock held, and in turns calls free_irq which then
    complains when manipulating procfs. This fixes it by moving the
    call to the notifier to outside of the locked section.

Signed-off-by: Christian Borntraeger<borntraeger@de.ibm.com>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>

10 years agortc-cmos: move wake setup from ACPI glue into RTC driver
Bjorn Helgaas [Tue, 14 Oct 2008 19:50:21 +0000]
rtc-cmos: move wake setup from ACPI glue into RTC driver

Move rtc_wake_setup() from drivers/acpi/glue.c into the RTC driver
in drivers/rtc/rtc-cmos.c.

This removes the ordering constraint between the module_init(acpi_rtc_init)
and the cmos_do_probe() code that depends on it.

Signed-off-by: Bjorn Helgaas <bjorn.helgaas@hp.com>
Acked-by: Rafael J. Wysocki <rjw@sisk.pl>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

10 years agoHID: build drivers for all quirky devices by default
Jiri Kosina [Tue, 14 Oct 2008 21:37:33 +0000]
HID: build drivers for all quirky devices by default

Once kernel configuration has CONFIG_HID turned on, let also all the
specialized drivers for quirky devices to be built (unless CONFIG_EMBEDDED is
specified), as usually users don't care that much which driver
gives them the functionality, but when they want generic support, they
probably want to have support for all the quirky devices as well.

Signed-off-by: Jiri Kosina <jkosina@suse.cz>

10 years agoHID: add missing blacklist entry for Apple ATV ircontrol
Jiri Kosina [Tue, 14 Oct 2008 20:45:40 +0000]
HID: add missing blacklist entry for Apple ATV ircontrol

This device is already handled by hid-apple driver, but the blacklist entry
was missing in generic driver.

Reported-by: Jiri Slaby <jslaby@suse.cz>
Signed-off-by: Jiri Kosina <jkosina@suse.cz>

10 years agoHID: add support for Bright ABNT2 brazilian device
Mauro Carvalho Chehab [Tue, 14 Oct 2008 20:41:43 +0000]
HID: add support for Bright ABNT2 brazilian device

This keyboard needs to reset the LEDS during probe.

Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>
Signed-off-by: Jiri Kosina <jkosina@suse.cz>

10 years agoHID: Don't let Avermedia Radio FM800 be handled by usb hid drivers
Mauro Carvalho Chehab [Tue, 14 Oct 2008 20:34:32 +0000]
HID: Don't let Avermedia Radio FM800 be handled by usb hid drivers

Based on an original patch from Alexey Klimov <klimov.linux@gmail.com>,
against kernel version 2.6.27.

This device is already handled by radio-mr800 driver, and we therefore
want usbhid not to touch it.

Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>
Signed-off-by: Jiri Kosina <jkosina@suse.cz>

10 years agoHID: fix numlock led on Dell device 0x413c/0x2105
Mauro Carvalho Chehab [Tue, 14 Oct 2008 13:08:47 +0000]
HID: fix numlock led on Dell device 0x413c/0x2105

This keyboard needs to re-sync numlock after probing.

Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>
Signed-off-by: Jiri Kosina <jkosina@suse.cz>

10 years agoHID: remove warn() macro from usb hid drivers
From: Greg Kroah-Hartman [Sat, 11 Oct 2008 22:25:51 +0000]
HID: remove warn() macro from usb hid drivers

USB should not be having it's own printk macros, so remove warn() and
use the system-wide standard of dev_warn() wherever possible.  In the
few places that will not work out, use a basic printk().

Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
Signed-off-by: Jiri Kosina <jkosina@suse.cz>

10 years agoHID: remove info() macro from usb HID drivers
Greg Kroah-Hartman [Sat, 11 Oct 2008 22:14:23 +0000]
HID: remove info() macro from usb HID drivers

USB should not be having it's own printk macros, so remove info() and
use the system-wide standard of dev_info() wherever possible.

Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
Signed-off-by: Jiri Kosina <jkosina@suse.cz>

10 years agoHID: add appletv IR receiver quirk
Peter Korsgaard [Mon, 6 Oct 2008 09:15:34 +0000]
HID: add appletv IR receiver quirk

Similar to the existing IRCONTROL4 handling

Signed-off-by: Peter Korsgaard <jacmet@sunsite.dk>
Signed-off-by: Jiri Kosina <jkosina@suse.cz>

10 years agoHID: fix a lockup regression when using force feedback on a PID device
Anssi Hannula [Sat, 4 Oct 2008 12:44:06 +0000]
HID: fix a lockup regression when using force feedback on a PID device

Commit 8006479c9b75fb6594a7b746af3d7f1fbb68f18f introduced a spinlock in
input_dev->event_lock, which is locked when handling input events.
However, the hid-pidff driver sleeps when handling events as it waits for
reports being sent to the device before changing the report contents
again.
This causes a system lockup when trying to use force feedback with a PID
device, a regression introduced in 2.6.24 and 2.6.23.15.

Fix it by extracting the raw report data from struct hid_report
immediately when hid_submit_report() is called, therefore allowing
drivers to change the contents of struct hid_report immediately without
affecting the already-queued transfer.

In hid-pidff, re-add the removed usbhid_wait_io() to
pidff_erase_effect() instead, to prevent a full report queue from causing
the submission to fail, thus not freeing up device memory.
pidff_erase_effect() is not called while dev->event_lock is held.

Signed-off-by: Anssi Hannula <anssi.hannula@gmail.com>
Signed-off-by: Jiri Kosina <jkosina@suse.cz>

10 years agoHID: hiddev.h: Fix example code.
Antonio Ospite [Thu, 2 Oct 2008 20:15:02 +0000]
HID: hiddev.h: Fix example code.

Fix hiddev.h example code.
To get the correct usage code, you need to set report_type and
report_id.

Signed-off-by: Antonio Ospite <ospite@studenti.unina.it>
Signed-off-by: Jiri Kosina <jkosina@suse.cz>

10 years agoHID: hiddev.h: Fix mixed space and tabs in example code.
Antonio Ospite [Thu, 2 Oct 2008 20:14:54 +0000]
HID: hiddev.h: Fix mixed space and tabs in example code.

Fix mixed space and tabs in example code.

Signed-off-by: Antonio Ospite <ospite@studenti.unina.it>
Signed-off-by: Jiri Kosina <jkosina@suse.cz>

10 years agoHID: convert to dev_* prints
Jiri Slaby [Thu, 18 Sep 2008 10:23:34 +0000]
HID: convert to dev_* prints

Since we have a real device bound to a driver, we may use struct
device for printing. Use dev_* functions instead of printks in
4 drivers.

Signed-off-by: Jiri Slaby <jirislaby@gmail.com>
Signed-off-by: Jiri Kosina <jkosina@suse.cz>

10 years agoHID: remove hid-ff
Jiri Slaby [Thu, 18 Sep 2008 10:23:33 +0000]
HID: remove hid-ff

hid-ff.c now calls only pidff (generic driver), the special ones are now
in separate drivers. Invoke pidff on all non-special directly.

Signed-off-by: Jiri Slaby <jirislaby@gmail.com>
Signed-off-by: Jiri Kosina <jkosina@suse.cz>

10 years agoHID: move zeroplus FF processing
Jiri Slaby [Thu, 18 Sep 2008 10:23:32 +0000]
HID: move zeroplus FF processing

Signed-off-by: Jiri Slaby <jirislaby@gmail.com>
Signed-off-by: Jiri Kosina <jkosina@suse.cz>

10 years agoHID: move thrustmaster FF processing
Jiri Slaby [Thu, 18 Sep 2008 10:23:31 +0000]
HID: move thrustmaster FF processing

Signed-off-by: Jiri Slaby <jirislaby@gmail.com>
Signed-off-by: Jiri Kosina <jkosina@suse.cz>

10 years agoHID: move pantherlord FF processing
Jiri Slaby [Thu, 18 Sep 2008 17:43:32 +0000]
HID: move pantherlord FF processing

Move the force feedback processing into a separate module.

[jkosina@suse.cz: fix Kconfig texts a little bit]

Signed-off-by: Jiri Slaby <jirislaby@gmail.com>
Signed-off-by: Jiri Kosina <jkosina@suse.cz>

10 years agoHID: fix incorrent length condition in hidraw_write()
Jiri Kosina [Wed, 17 Sep 2008 17:41:58 +0000]
HID: fix incorrent length condition in hidraw_write()

The bound check on the buffer length

if (count > HID_MIN_BUFFER_SIZE)

is of course incorrent, the proper check is

if (count > HID_MAX_BUFFER_SIZE)

Fix it.

Reported-by: Jerry Ryle <jerry@mindtribe.com>
Signed-off-by: Jiri Kosina <jkosina@suse.cz>

10 years agoHID: fix tty<->hid deadlock
Jiri Slaby [Mon, 8 Sep 2008 23:23:03 +0000]
HID: fix tty<->hid deadlock

hid_compat_load() runs on the default workqueue, it request_module(), it
execs modprobe, it exits, tty flushes default workqueue, it hangs, because
we are still in it.

Signed-off-by: Jiri Slaby <jirislaby@gmail.com>
Tested-by: <Valdis.Kletnieks@vt.edu>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Jiri Kosina <jkosina@suse.cz>

10 years agoHID: ignore iBuddy devices
Remi Cattiau [Mon, 8 Sep 2008 23:39:33 +0000]
HID: ignore iBuddy devices

iBuddy devices claim to be HID devices, but they are not.
Add them to the blacklist.

Signed-off-by: Remi Cattiau <remi@cattiau.com>
Signed-off-by: Jiri Kosina <jkosina@suse.cz>

10 years agoHID: report descriptor fix for remaining MacBook JIS keyboards
Tomoya Adachi [Thu, 4 Sep 2008 09:29:27 +0000]
HID: report descriptor fix for remaining MacBook JIS keyboards

This patch fixes a problem that MacBook JIS keyboard sends wrong report
descriptors. Although it has already been fixed in the first Core 2 Duo model,
it still remains in other models of MacBook.

Signed-off-by: Tomoya Adachi <adachi@il.is.s.u-tokyo.ac.jp>
Signed-off-by: Jiri Kosina <jkosina@suse.cz>

10 years agoHID: fix gyration build error
Randy Dunlap [Thu, 4 Sep 2008 08:55:00 +0000]
HID: fix gyration build error

Fix config symbol name in ifdef to fix build error:

ERROR: "hid_compat_gyration" [drivers/hid/hid-dummy.ko] undefined!

Signed-off-by: Randy Dunlap <randy.dunlap@oracle.com>
Acked-by: Jiri Slaby <jirislaby@gmail.com>
Signed-off-by: Jiri Kosina <jkosina@suse.cz>

10 years agoHID: Autocentering support for Logitech MOMO Racing Wheel
Sergey Belyashov [Tue, 2 Sep 2008 15:31:16 +0000]
HID: Autocentering support for Logitech MOMO Racing Wheel

Current kernel has no support for autocentering for Logitech wheels.  By
default autocentering enabled in wheel and constant effect does not work
properly. Using USB sniffer I found command which change autocentering
settings: 0xFE, 0x0D, 0x0R, 0x0L, 0x80, 0x00, 0x00, where R - clockwise force,
L - counter-clockwise (0x0-0xF, 0xC = 100%).

Signed-off-by: Sergey Belyashov <Sergey.Belyashov@gmail.com>
Signed-off-by: Jiri Kosina <jkosina@suse.cz>

10 years agoHID: fix grammo in HID_COMPAT Kconfig help text
Alex Chiang [Wed, 27 Aug 2008 11:36:18 +0000]
HID: fix grammo in HID_COMPAT Kconfig help text

The Kconfig option for HID_COMPAT should read "lose", not "loose".

Signed-off-by: Alex Chiang <achiang@hp.com>
Signed-off-by: Jiri Kosina <jkosina@suse.cz>

10 years agoHID: add support for Super Dual Box Pro USB PS2/PS2 adapter
Jiri Kosina [Wed, 27 Aug 2008 09:21:42 +0000]
HID: add support for Super Dual Box Pro USB PS2/PS2 adapter

This seems to be the very same device, as already supported Smartjoy
dual Plus, but with slightly different vendor ID. Let's support this
one too.

Reported-by: David Ashley <dashxdr@gmail.com>
Signed-off-by: Jiri Kosina <jkosina@suse.cz>

10 years agoHID: remove ignore quirk for MGE UPS devices
Richard Hughes [Wed, 27 Aug 2008 09:19:37 +0000]
HID: remove ignore quirk for MGE UPS devices

This patch reverts the change made four years ago here:
http://www.vg.kernel.org/pub/linux/kernel/people/gregkh/usb/2.4/usb-hid-2.4.25-pre7.patch

UPS's made by MGE can be used with usbhid just fine, and by removing the
ignore quirk allows them to be used with HAL so they just work when plugged
in, without needing to be manually configured.

With the ignore quirk in place a user would have to configure NUT before the
UPS could be used, as NUT uses it's own internal USB matching framework
to match against the USB devices, do low level control messages on the
device and then parse the HID tables all in userspace.

This is not needed, as allowing the device to be claimed as a usbhid device
allows it to be used like any other USB UPS device. The devices correctly
advertise the power device page which can be queried for the device state.

I assume the quirk was changed so that people using < libusb 0.1.8 could
still use NUT's internal HID code to manage the UPS.
libusb 0.1.8 was released quite some time ago: 2004-02-11.

This patch does not break NUT as in drivers/libusb.c the device is force
unbound from the kernel driver using usb_detach_kernel_driver_np () where
it can be controlled like normal.

[jkosina@suse.cz: adapt to the new hidbus code]

Signed-off-by: Richard Hughes <rhughes@redhat.com>
Signed-off-by: Jiri Kosina <jkosina@suse.cz>