GitHub/moto-9609/android_kernel_motorola_exynos9610.git
18 years ago[PATCH] pcmcia: remove get_socket callback
Dominik Brodowski [Sat, 12 Nov 2005 22:56:33 +0000 (23:56 +0100)]
[PATCH] pcmcia: remove get_socket callback

The .get_socket callback is never used by the PCMCIA core, therefore remove
it.

Signed-off-by: Dominik Brodowski <linux@dominikbrodowski.net>
18 years ago[PATCH] pcmcia: remove socket register_callback
Dominik Brodowski [Wed, 28 Sep 2005 17:41:55 +0000 (19:41 +0200)]
[PATCH] pcmcia: remove socket register_callback

Remove the register_callback declaration in struct pccard_operations as it is
unused.

Signed-off-by: Dominik Brodowski <linux@dominikbrodowski.net>
18 years ago[PATCH] yenta: optimize interrupt handler
Daniel Ritz [Tue, 20 Sep 2005 21:17:37 +0000 (14:17 -0700)]
[PATCH] yenta: optimize interrupt handler

Don't waste cpu time in yenta interrupt handler when the interrupt was for
another device.

Signed-off-by: Daniel Ritz <daniel.ritz@gmx.ch>
Signed-off-by: Dominik Brodowski <linux@dominikbrodowski.net>
18 years ago[PATCH] pcmcia: avoid macro usage in cistpl
Pavel Machek [Wed, 7 Sep 2005 23:00:26 +0000 (16:00 -0700)]
[PATCH] pcmcia: avoid macro usage in cistpl

Fix macro abuse in pcmcia.

Signed-off-by: Pavel Machek <pavel@suse.cz>
Signed-off-by: Dominik Brodowski <linux@dominikbrodowski.net>
18 years agoMerge git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/driver-2.6
Linus Torvalds [Thu, 5 Jan 2006 02:44:12 +0000 (18:44 -0800)]
Merge git://git./linux/kernel/git/gregkh/driver-2.6

Trivial manual merge fixup for usb_find_interface clashes.

18 years ago[PATCH] i386: "invalid operand" -> "invalid opcode"
Chuck Ebbert [Wed, 4 Jan 2006 03:36:14 +0000 (22:36 -0500)]
[PATCH] i386: "invalid operand" -> "invalid opcode"

According to the manual, INT 6 is "invalid opcode", not "invalid operand".

Signed-off-by: Chuck Ebbert <76306.1226@compuserve.com>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] Fix IXP4xx watchdog errata workaround
Deepak Saxena [Tue, 3 Jan 2006 20:50:30 +0000 (12:50 -0800)]
[PATCH] Fix IXP4xx watchdog errata workaround

The IXP4xx driver bails out on all A0 CPUs, but it should only do
so on IXP42x as IXP46x has functioning HW.

Signed-off-by: Deepak Saxena <dsaxena@plexity.net>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years agoMerge git://git.kernel.org/pub/scm/linux/kernel/git/perex/alsa
Linus Torvalds [Thu, 5 Jan 2006 00:38:36 +0000 (16:38 -0800)]
Merge git://git./linux/kernel/git/perex/alsa

18 years agoMerge git://git.kernel.org/pub/scm/linux/kernel/git/sam/kbuild
Linus Torvalds [Thu, 5 Jan 2006 00:36:52 +0000 (16:36 -0800)]
Merge git://git./linux/kernel/git/sam/kbuild

18 years agoMerge git://git.kernel.org/pub/scm/linux/kernel/git/bunk/trivial
Linus Torvalds [Thu, 5 Jan 2006 00:34:57 +0000 (16:34 -0800)]
Merge git://git./linux/kernel/git/bunk/trivial

18 years agoMerge branch 'upstream' of master.kernel.org:/pub/scm/linux/kernel/git/jgarzik/libata-dev
Linus Torvalds [Thu, 5 Jan 2006 00:32:33 +0000 (16:32 -0800)]
Merge branch 'upstream' of /linux/kernel/git/jgarzik/libata-dev

18 years agoMerge branch 'upstream-linus' of master.kernel.org:/pub/scm/linux/kernel/git/jgarzik...
Linus Torvalds [Thu, 5 Jan 2006 00:31:56 +0000 (16:31 -0800)]
Merge branch 'upstream-linus' of /linux/kernel/git/jgarzik/netdev-2.6

18 years agoMerge git://git.kernel.org/pub/scm/linux/kernel/git/jejb/scsi-misc-2.6
Linus Torvalds [Thu, 5 Jan 2006 00:30:12 +0000 (16:30 -0800)]
Merge git://git./linux/kernel/git/jejb/scsi-misc-2.6

18 years agoMerge git://git.kernel.org/pub/scm/linux/kernel/git/davem/net-2.6
Linus Torvalds [Thu, 5 Jan 2006 00:27:41 +0000 (16:27 -0800)]
Merge git://git./linux/kernel/git/davem/net-2.6

18 years agoMerge git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/usb-2.6
Linus Torvalds [Thu, 5 Jan 2006 00:25:44 +0000 (16:25 -0800)]
Merge git://git./linux/kernel/git/gregkh/usb-2.6

18 years agoMerge master.kernel.org:/pub/scm/linux/kernel/git/davej/cpufreq
Linus Torvalds [Thu, 5 Jan 2006 00:21:26 +0000 (16:21 -0800)]
Merge /pub/scm/linux/kernel/git/davej/cpufreq

18 years agoRelax the rw_verify_area() error checking.
Linus Torvalds [Thu, 5 Jan 2006 00:20:40 +0000 (16:20 -0800)]
Relax the rw_verify_area() error checking.

In particular, allow over-large read- or write-requests to be downgraded
to a more reasonable range, rather than considering them outright errors.

We want to protect lower layers from (the sadly all too common) overflow
conditions, but prefer to do so by chopping the requests up, rather than
just refusing them outright.

Cc: Peter Anvin <hpa@zytor.com>
Cc: Ulrich Drepper <drepper@redhat.com>
Cc: Andi Kleen <ak@suse.de>
Cc: Al Viro <viro@ftp.linux.org.uk>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] net: swich device attribute creation to default attrs
Kay Sievers [Mon, 19 Dec 2005 00:42:56 +0000 (01:42 +0100)]
[PATCH] net: swich device attribute creation to default attrs

Recent udev versions don't longer cover bad sysfs timing with built-in
logic. Explicit rules are required to do that. For net devices, the
following is needed:
  ACTION=="add", SUBSYSTEM=="net", WAIT_FOR_SYSFS="address"
to handle access to net device properties from an event handler without
races.

This patch changes the main net attributes to be created by the driver
core, which is done _before_ the event is sent out and will not require
the stat() loop of the WAIT_FOR_SYSFS key.

Signed-off-by: Kay Sievers <kay.sievers@suse.de>
Acked-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
18 years ago[PATCH] drivers/base/power/runtime.c: #if 0 dpm_set_power_state()
Adrian Bunk [Mon, 12 Dec 2005 09:31:03 +0000 (01:31 -0800)]
[PATCH] drivers/base/power/runtime.c: #if 0 dpm_set_power_state()

This patch #if 0's an unused global function.

Signed-off-by: Adrian Bunk <bunk@stusta.de>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
18 years ago[PATCH] sysfs: handle failures in sysfs_make_dirent
Steven Rostedt [Wed, 23 Nov 2005 14:15:44 +0000 (09:15 -0500)]
[PATCH] sysfs: handle failures in sysfs_make_dirent

I noticed that if sysfs_make_dirent fails to allocate the sd, then a
null will be passed to sysfs_put.

Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
18 years ago[PATCH] Driver core: Make block devices create the proper symlink name
Greg Kroah-Hartman [Tue, 13 Dec 2005 23:17:34 +0000 (15:17 -0800)]
[PATCH] Driver core: Make block devices create the proper symlink name

Block devices need to add the block device name to the symlink they put
in the device directory, otherwise multiple symlinks of the same name
can be created.  This matches the class system, which works the same
way, we just forgot to convert block at the same time.

Cc: Pete Zaitcev <zaitcev@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
18 years ago[PATCH] Driver core: only all userspace bind/unbind if CONFIG_HOTPLUG is enabled
Greg Kroah-Hartman [Tue, 13 Dec 2005 23:17:34 +0000 (15:17 -0800)]
[PATCH] Driver core: only all userspace bind/unbind if CONFIG_HOTPLUG is enabled

Thanks to drivers making their id tables __devinit, we can't allow
userspace to bind or unbind drivers from devices manually through sysfs.
So we only allow this if CONFIG_HOTPLUG is enabled.

Cc: Dmitry Torokhov <dmitry.torokhov@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
18 years ago[PATCH] Driver Core: Rearrange exports in platform.c
Dmitry Torokhov [Sat, 10 Dec 2005 06:36:28 +0000 (01:36 -0500)]
[PATCH] Driver Core: Rearrange exports in platform.c

Driver core: rearrange exports in platform.c

The new way is to specify export right after symbol definition.
Rearrange exports to follow new style to avoid mixing two styles
in one file.

Signed-off-by: Dmitry Torokhov <dtor@mail.ru>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
18 years ago[PATCH] Driver Core: Add platform_device_del()
Dmitry Torokhov [Sat, 10 Dec 2005 06:36:27 +0000 (01:36 -0500)]
[PATCH] Driver Core: Add platform_device_del()

Driver core: add platform_device_del function

Having platform_device_del90 allows more straightforward error
handling code in drivers registering platform devices.

Signed-off-by: Dmitry Torokhov <dtor@mail.ru>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
18 years ago[PATCH] Input: fix add modalias support build error
Rusty Russell [Sat, 10 Dec 2005 11:48:20 +0000 (22:48 +1100)]
[PATCH] Input: fix add modalias support build error

Fix build when scripts/mod/file2alias.c includes linux/input.h, which
tries to include /usr/include/linux/mod_devicetable.h:

 In file included from scripts/mod/file2alias.c:40:
 include/linux/input.h:21:35: linux/mod_devicetable.h: No such file or directory
 make[2]: *** [scripts/mod/file2alias.o] Error 1

Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
18 years ago[PATCH] Input: add modalias support
Rusty Russell [Wed, 7 Dec 2005 20:40:34 +0000 (21:40 +0100)]
[PATCH] Input: add modalias support

Here's the patch for modalias support for input classes.  It uses
comma-separated numbers, and doesn't describe all the potential keys (no
module currently cares, and that would make the strings huge).  The
changes to input.h are to move the definitions needed by file2alias
outside __KERNEL__.  I chose not to move those definitions to
mod_devicetable.h, because there are so many that it might break compile
of something else in the kernel.

The rest is fairly straightforward.

Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
CC: Kay Sievers <kay.sievers@vrfy.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
18 years ago[PATCH] ide: MODALIAS support for autoloading of ide-cd, ide-disk, ...
Kay Sievers [Mon, 12 Dec 2005 17:03:44 +0000 (18:03 +0100)]
[PATCH] ide: MODALIAS support for autoloading of ide-cd, ide-disk, ...

IDE: MODALIAS support for autoloading of ide-cd, ide-disk, ...

Add MODULE_ALIAS to IDE midlayer modules: ide-disk, ide-cd, ide-floppy and
ide-tape, to autoload these modules depending on the probed media type of
the IDE device.

It is used by udev and replaces the former agent shell script of the hotplug
package, which was required to lookup the media type in the proc filesystem.
Using proc was racy, cause the media file is created after the hotplug event
is sent out.

The module autoloading does not take any effect, until something like the
following udev rule is configured:
  SUBSYSTEM=="ide",  ACTION=="add", ENV{MODALIAS}=="?*", RUN+="/sbin/modprobe $env{MODALIAS}"

The module ide-scsi will not be autoloaded, cause it requires manual
configuration. It can't be, and never was supported for automatic setup in
the hotplug package. Adding a MODULE_ALIAS to ide-scsi for all supported
media types, would just lead to a default blacklist entry anyway.

  $ modinfo ide-disk
  filename:       /lib/modules/2.6.15-rc4-g1b0997f5/kernel/drivers/ide/ide-disk.ko
  description:    ATA DISK Driver
  alias:          ide:*m-disk*
  license:        GPL
  ...

  $ modprobe -vn ide:m-disk
  insmod /lib/modules/2.6.15-rc4-g1b0997f5/kernel/drivers/ide/ide-disk.ko

  $ cat /sys/bus/ide/devices/0.0/modalias
  ide:m-disk

It also adds attributes to the IDE device:
  $ tree /sys/bus/ide/devices/0.0/
  /sys/bus/ide/devices/0.0/
  |-- bus -> ../../../../../../../bus/ide
  |-- drivename
  |-- media
  |-- modalias
  |-- power
  |   |-- state
  |   `-- wakeup
  `-- uevent

  $ cat /sys/bus/ide/devices/0.0/{modalias,drivename,media}
  ide:m-disk
  hda
  disk

Signed-off-by: Kay Sievers <kay.sievers@suse.de>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
18 years ago[PATCH] kobject_uevent CONFIG_NET=n fix
akpm@osdl.org [Wed, 23 Nov 2005 07:36:13 +0000 (23:36 -0800)]
[PATCH] kobject_uevent CONFIG_NET=n fix

lib/lib.a(kobject_uevent.o)(.text+0x25f): In function `kobject_uevent':
: undefined reference to `__alloc_skb'
lib/lib.a(kobject_uevent.o)(.text+0x2a1): In function `kobject_uevent':
: undefined reference to `skb_over_panic'
lib/lib.a(kobject_uevent.o)(.text+0x31d): In function `kobject_uevent':
: undefined reference to `skb_over_panic'
lib/lib.a(kobject_uevent.o)(.text+0x356): In function `kobject_uevent':
: undefined reference to `netlink_broadcast'
lib/lib.a(kobject_uevent.o)(.init.text+0x9): In function `kobject_uevent_init':
: undefined reference to `netlink_kernel_create'
make: *** [.tmp_vmlinux1] Error 1

Netlink is unconditionally enabled if CONFIG_NET, so that's OK.

kobject_uevent.o is compiled even if !CONFIG_HOTPLUG, which is lazy.

Let's compound the sin.

Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
18 years ago[PATCH] Allow overlapping resources for platform devices
Kumar Gala [Mon, 28 Nov 2005 16:15:39 +0000 (10:15 -0600)]
[PATCH] Allow overlapping resources for platform devices

There are cases in which a device's memory mapped registers overlap
with another device's memory mapped registers.  On several PowerPC
devices this occurs for the MDIO bus, whose registers tended to overlap
with one of the ethernet controllers.

By switching from request_resource to insert_resource we can register
the MDIO bus as a proper platform device and not hack around how we
handle its memory mapped registers.

Signed-off-by: Kumar Gala <galak@kernel.crashing.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
18 years ago[PATCH] klist: Fix broken kref counting in find functions
Frank Pavlic [Sun, 27 Nov 2005 04:48:40 +0000 (20:48 -0800)]
[PATCH] klist: Fix broken kref counting in find functions

The klist reference counting in the find functions that use
klist_iter_init_node is broken.  If the function (for example
driver_find_device) is called with a NULL start object then everything is
fine, the first call to next_device()/klist_next increases the ref-count of
the first node on the list and does nothing for the start object which is
NULL.

If they are called with a valid start object then klist_next will decrement
the ref-count for the start object but nobody has incremented it.  Logical
place to fix this would be klist_iter_init_node because the function puts a
reference of the object into the klist_iter struct.

Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
Signed-off-by: Frank Pavlic <pavlic@de.ibm.com>
Cc: Patrick Mochel <mochel@digitalimplant.org>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
18 years ago[PATCH] Hold the device's parent's lock during probe and remove
Alan Stern [Thu, 17 Nov 2005 21:54:12 +0000 (16:54 -0500)]
[PATCH] Hold the device's parent's lock during probe and remove

This patch (as604) makes the driver core hold a device's parent's lock
as well as the device's lock during calls to the probe and remove
methods in a driver.  This facility is needed by USB device drivers,
owing to the peculiar way USB devices work:

A device provides multiple interfaces, and drivers are bound
to interfaces rather than to devices;

Nevertheless a reset, reset-configuration, suspend, or resume
affects the entire device and requires the caller to hold the
lock for the device, not just a lock for one of the interfaces.

Since a USB driver's probe method is always called with the interface
lock held, the locking order rules (always lock parent before child)
prevent these methods from acquiring the device lock.  The solution
provided here is to call all probe and remove methods, for all devices
(not just USB), with the parent lock already acquired.

Although currently only the USB subsystem requires these changes, people
have mentioned in prior discussion that the overhead of acquiring an
extra semaphore in all the prove/remove sequences is not overly large.

Up to now, the USB core has been using its own set of private
semaphores.  A followup patch will remove them, relying entirely on the
device semaphores provided by the driver core.

The code paths affected by this patch are:

device_add and device_del: The USB core already holds the parent
lock, so no actual change is needed.

driver_register and driver_unregister: The driver core will now
lock both the parent and the device before probing or removing.

driver_bind and driver_unbind (in sysfs): These routines will
now lock both the parent and the device before binding or
unbinding.

bus_rescan_devices: The helper routine will lock the parent
before probing a device.

I have not tested this patch for conflicts with other subsystems.  As
far as I can see, the only possibility of conflict would lie in the
bus_rescan_devices pathway, and it seems pretty remote.  Nevertheless,
it would be good for this to get a lot of testing in -mm.

Signed-off-by: Alan Stern <stern@rowland.harvard.edu>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
18 years ago[PATCH] driver kill hotplug word from sn and others fix
Paul Jackson [Sat, 26 Nov 2005 04:04:26 +0000 (20:04 -0800)]
[PATCH] driver kill hotplug word from sn and others fix

The first of these changes s/hotplug/uevent/ was needed to
compile sn2_defconfig (ia64/sn).  The other three files
changed are blind changes of all remaining bus_type.hotplug
references I could find to bus_type.uevent.

This patch attempts to finish similar changes made in the
gregkh-driver-kill-hotplug-word-from-driver-core Nov 22 patch.

Signed-off-by: Paul Jackson <pj@sgi.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
18 years ago[PATCH] HOTPLUG: always enable the .config option, unless EMBEDDED
Greg Kroah-Hartman [Wed, 16 Nov 2005 19:27:07 +0000 (11:27 -0800)]
[PATCH] HOTPLUG: always enable the .config option, unless EMBEDDED

With modules, dynamic /dev, and uevents, people really want
CONFIG_HOTPLUG to be enabled in their kernels.  If not, they can still
disable it, but it is discouraged.

Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
18 years ago[PATCH] driver core: replace "hotplug" by "uevent"
Kay Sievers [Wed, 16 Nov 2005 08:00:00 +0000 (09:00 +0100)]
[PATCH] driver core: replace "hotplug" by "uevent"

Leave the overloaded "hotplug" word to susbsystems which are handling
real devices. The driver core does not "plug" anything, it just exports
the state to userspace and generates events.

Signed-off-by: Kay Sievers <kay.sievers@suse.de>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
18 years ago[PATCH] merge kobject_uevent and kobject_hotplug
Kay Sievers [Fri, 11 Nov 2005 13:43:07 +0000 (14:43 +0100)]
[PATCH] merge kobject_uevent and kobject_hotplug

The distinction between hotplug and uevent does not make sense these
days, netlink events are the default.

udev depends entirely on netlink uevents. Only during early boot and
in initramfs, /sbin/hotplug is needed. So merge the two functions and
provide only one interface without all the options.

The netlink layer got a nice generic interface with named slots
recently, which is probably a better facility to plug events for
subsystem specific events.
Also the new poll() interface to /proc/mounts is a nicer way to
notify about changes than sending events through the core.
The uevents should only be used for driver core related requests to
userspace now.

Signed-off-by: Kay Sievers <kay.sievers@suse.de>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
18 years ago[PATCH] remove mount/umount uevents from superblock handling
Kay Sievers [Fri, 11 Nov 2005 05:09:55 +0000 (06:09 +0100)]
[PATCH] remove mount/umount uevents from superblock handling

The names of these events have been confusing from the beginning
on, as they have been more like claim/release events. We needed these
events for noticing HAL if storage devices have been mounted.

Thanks to Al, we have the proper solution now and can poll()
/proc/mounts instead to get notfied about mount tree changes.

Signed-off-by: Kay Sievers <kay.sievers@suse.de>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
18 years ago[PATCH] add uevent_helper control in /sys/kernel/
Kay Sievers [Fri, 11 Nov 2005 03:58:04 +0000 (04:58 +0100)]
[PATCH] add uevent_helper control in /sys/kernel/

This deprecates the /proc/sys/kernel/hotplug file, as all
this stuff should be in /sys some day, right? :)
In /sys/kernel/ we have now uevent_seqnum and uevent_helper.
The seqnum is no longer used by udev, as the version for this
kernel depends on netlink which events will never get
out-of-order.

Recent udev versions disable the /sbin/hotplug helper with
an init script, cause it leads to OOM on big boxes by running
hundreds of shells in parallel. It should be done now by:
  echo "" > /sys/kernel/uevent_helper

(Note that "-n" does not work, cause neighter proc nor sysfs
support truncate().)

Signed-off-by: Kay Sievers <kay.sievers@suse.de>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
18 years ago[PATCH] remove CONFIG_KOBJECT_UEVENT option
Kay Sievers [Fri, 11 Nov 2005 04:33:52 +0000 (05:33 +0100)]
[PATCH] remove CONFIG_KOBJECT_UEVENT option

It makes zero sense to have hotplug, but not the netlink
events enabled today. Remove this option and merge the
kobject_uevent.h header into the kobject.h header file.

Signed-off-by: Kay Sievers <kay.sievers@suse.de>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
18 years ago[PATCH] keep pnpbios usermod_helper away from hotplug_path[]
Kay Sievers [Fri, 11 Nov 2005 03:25:06 +0000 (04:25 +0100)]
[PATCH] keep pnpbios usermod_helper away from hotplug_path[]

These days we use udev to manage all kernel events. /proc/sys/kernel/hotplug
will usually be disabled by an init-script. pnpnbios is not integrated with
the driver core and should stay away from the now disabled /sbin/hotplug.

Set the helper to /sbin/phpbios, even when there is probably no current
user of this faciliy. If it's needed, it should definitely get proper driver
core integration instead of forking binaries from the kernel.

Signed-off-by: Kay Sievers <kay.sievers@suse.de>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
18 years ago[TCP] tcp_vegas: Fix slow start
Thomas Young [Wed, 4 Jan 2006 21:59:32 +0000 (13:59 -0800)]
[TCP] tcp_vegas: Fix slow start

Vegas' slow start was only adding one MSS per RTT rather than one for
every ack. Slow start behavior should now match Reno.

Signed-off-by: Thomas Young <tyo@ee.mu.oz.au>
Signed-off-by: David S. Miller <davem@davemloft.net>
18 years ago[NET]: More instruction checks fornet/core/filter.c
Kris Katterjohn [Wed, 4 Jan 2006 21:58:36 +0000 (13:58 -0800)]
[NET]: More instruction checks fornet/core/filter.c

Signed-off-by: Kris Katterjohn <kjak@users.sourceforge.net>
Signed-off-by: David S. Miller <davem@davemloft.net>
18 years ago[NETFILTER]: Use macro for spinlock_t/rwlock_t initializations/definition.
YOSHIFUJI Hideaki [Wed, 4 Jan 2006 21:56:54 +0000 (13:56 -0800)]
[NETFILTER]: Use macro for spinlock_t/rwlock_t initializations/definition.

Signed-off-by: YOSHIFUJI Hideaki <yoshfuji@linux-ipv6.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
18 years ago[IPV6]: Use macro for rwlock_t initialization.
YOSHIFUJI Hideaki [Wed, 4 Jan 2006 21:56:31 +0000 (13:56 -0800)]
[IPV6]: Use macro for rwlock_t initialization.

Signed-off-by: YOSHIFUJI Hideaki <yoshfuji@linux-ipv6.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
18 years ago[ECONET]: Use macro for spinlock_t definition.
YOSHIFUJI Hideaki [Wed, 4 Jan 2006 21:56:08 +0000 (13:56 -0800)]
[ECONET]: Use macro for spinlock_t definition.

Signed-off-by: YOSHIFUJI Hideaki <yoshfuji@linux-ipv6.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
18 years ago[PATCH] Fix pragma packing in ip2 driver
David Howells [Wed, 4 Jan 2006 18:01:03 +0000 (18:01 +0000)]
[PATCH] Fix pragma packing in ip2 driver

This fixes the pragma packing in the ip2 driver by popping the previous
setting rather than explicitly assuming that the correct setting is 4.

This also gets around a compiler bug in the FRV compiler when building
allmodconfig.

Signed-Off-By: David Howells <dhowells@redhat.com>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] USB: fix usb-skeleton limit resource usage patch.
Sam Bishop [Fri, 23 Dec 2005 00:11:02 +0000 (17:11 -0700)]
[PATCH] USB: fix usb-skeleton limit resource usage patch.

Prevents a compiler warning and uses down_interruptible() instead of down() in
process context.

Signed-off-by: Sam Bishop <sam@bishop.dhs.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
18 years ago[PATCH] USB: fix buffer size limiting in skeleton driver
Olav Kongas [Thu, 22 Dec 2005 10:44:52 +0000 (12:44 +0200)]
[PATCH] USB: fix buffer size limiting in skeleton driver

Fix buffer size limiting.

Signed-off-by: Olav Kongas <ok@artecdesign.ee>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
18 years ago[PATCH] USB: fix usb_find_interface for ppc64
Pete Zaitcev [Thu, 22 Dec 2005 01:24:54 +0000 (17:24 -0800)]
[PATCH] USB: fix usb_find_interface for ppc64

Fix usb_find_interface. You cannot case pointers to int and long on
a big-endian 64-bitter without consequences.

Signed-off-by: Pete Zaitcev <zaitcev@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
18 years ago[PATCH] USB: Export IEEE-1284 device id in sysfs for usblp devices
David Woodhouse [Fri, 23 Dec 2005 16:41:41 +0000 (16:41 +0000)]
[PATCH] USB: Export IEEE-1284 device id in sysfs for usblp devices

I looked at the userspace code which uses the LPIOC_GET_DEVICE_ID ioctl
and I almost went blind. Let's export it in sysfs instead, and just as a
string instead of with a big-endian length at the beginning of it.

This also prints the message about finding the printer _after_ we know
the minor device number it's going to have, rather than reporting all
printers as 'usblp0'.

Signed-off-by: David Woodhouse <dwmw2@infradead.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
18 years ago[PATCH] USB: ioctl compat for usblp.c
Pete Zaitcev [Thu, 22 Dec 2005 01:03:24 +0000 (17:03 -0800)]
[PATCH] USB: ioctl compat for usblp.c

From: David Woodhouse <dwmw2>

David has a G5 with a printer. I am quite surprised that nobody else noticed
this before. Linus has a G5. Hackers hate printing in general, maybe.

We do not use BKL anymore, because one of code paths had a sleeping call,
so we had to use a semaphore. I am sure it's safe to use unlocked_ioctl.

The new ioctls return long and retval is int. It looks completely fine to me.
We never want these extra bits, and the sign extension ought to work right.

Signed-off-by: Pete Zaitcev <zaitcev@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
--

18 years ago[PATCH] USB: zd1201: make sysfs device symlink
Nathan Lynch [Mon, 19 Dec 2005 05:41:38 +0000 (23:41 -0600)]
[PATCH] USB: zd1201: make sysfs device symlink

Noticed that my zd1201 adapter isn't "seen" by hal and NetworkManager.
The problem seems to be that unlike other network device drivers I
checked, zd1201 does not do a SET_NETDEV_DEV(), which makes it so a
"device" symlink is created under /sys/class/net/wlan0.

With the following patch the device symlink shows up, and now I am
happily using NetworkManager to control the adapter:

$ ls -l /sys/class/net/wlan0
total 0
-r--r--r--  1 root root 4096 Dec 18 13:42 address
-r--r--r--  1 root root 4096 Dec 18 13:42 addr_len
-r--r--r--  1 root root 4096 Dec 18 13:42 broadcast
-r--r--r--  1 root root 4096 Dec 18 13:42 carrier
lrwxrwxrwx  1 root root    0 Dec 18 13:42 device -> ../../../devices/pci0001:10/0001:10:1b.1/usb4/4-1
-r--r--r--  1 root root 4096 Dec 18 13:42 features

Signed-off-by: Nathan Lynch <ntl@pobox.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
18 years ago[PATCH] USB: Limiting of resource use in skeleton driver
Oliver Neukum [Wed, 21 Dec 2005 18:27:29 +0000 (19:27 +0100)]
[PATCH] USB: Limiting of resource use in skeleton driver

this introduces limits whose lack in the skeleton driver someone recently
complained about.

Signed-off-by: Oliver Neukum <oliver@neukum.name>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
18 years ago[PATCH] USB: replace __setup("nousb") with __module_param_call
Pete Zaitcev [Tue, 20 Dec 2005 22:15:04 +0000 (14:15 -0800)]
[PATCH] USB: replace __setup("nousb") with __module_param_call

Fedora users complain that passing "nousbstorage" to the installer causes
the rest of the USB support to disappear. The installer uses kernel command
line as a way to pass options through Syslinux. The problem stems from the
use of strncmp() in obsolete_checksetup().

I used __module_param_call() instead of module_param because I wanted to
preserve the old syntax in grub.conf, and it's the only macro which allows
to remove the prefix.

The fix is tested to accept the option "nousb" correctly now.

Signed-off-by: Pete Zaitcev <zaitcev@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
18 years ago[PATCH] USB: asix.c - Add Linksys USB200M Rev 2 ids
David Hollis [Mon, 19 Dec 2005 18:58:38 +0000 (13:58 -0500)]
[PATCH] USB: asix.c - Add Linksys USB200M Rev 2 ids

Attached patch adds device IDs for the Linksys USB200M Rev 2 device
which uses the AX88772 chipset.

Signed-off-by: David Hollis <dhollis@davehollis.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
18 years ago[PATCH] USB Storage: Fix unusual_devs.h order
Phil Dibowitz [Mon, 19 Dec 2005 05:30:02 +0000 (21:30 -0800)]
[PATCH] USB Storage: Fix unusual_devs.h order

Alan Stern pointed out there was an ordering issue in unusual_devs.h,
and this patch fixes it.

Signed-off-by: Phil Dibowitz <phil@ipom.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
18 years ago[PATCH] USB: ub 02 Removed unused variable
Daniel Marjamaki [Sat, 17 Dec 2005 10:38:46 +0000 (02:38 -0800)]
[PATCH] USB: ub 02 Removed unused variable

Removed an unused variable

Signed-off-by: Daniel Marjamaki <daniel.marjamaki@comhem.se>
Signed-off-by: Pete Zaitcev <zaitcev@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
18 years ago[PATCH] USB: ub 01 rename
Pete Zaitcev [Sat, 17 Dec 2005 10:34:12 +0000 (02:34 -0800)]
[PATCH] USB: ub 01 rename

Rename misleading UB_MINORS_PER_MAJOR into UB_PARTS_PER_LUN.

Signed-off-by: Pete Zaitcev <zaitcev@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
18 years ago[PATCH] USB: ub 00 implement retries and resets
Pete Zaitcev [Sat, 17 Dec 2005 10:16:43 +0000 (02:16 -0800)]
[PATCH] USB: ub 00 implement retries and resets

Implement command retries and resets in ub. It is advantageous for users
to know if their devices are getting bad. However, failing every I/O
is not practical if you have a external USB enclosure with a hard drive.

Signed-off-by: Pete Zaitcev <zaitcev@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
18 years ago[PATCH] USB: isp116x-hcd.c: Removed unused variable
Daniel Marjamäki [Tue, 13 Dec 2005 13:30:20 +0000 (15:30 +0200)]
[PATCH] USB: isp116x-hcd.c: Removed unused variable

Removed unused variable

Signed-off-by: Daniel Marjamäki <daniel.marjamaki@comhem.se>
Signed-off-by: Olav Kongas <ok@artecdesign.ee>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
18 years ago[PATCH] USB: Remove unneeded kmalloc() return value casts
Jesper Juhl [Sun, 11 Dec 2005 19:34:02 +0000 (20:34 +0100)]
[PATCH] USB: Remove unneeded kmalloc() return value casts

Remove kmalloc() return value casts that we don't need from
drivers/usb/*

Signed-off-by: Jesper Juhl <jesper.juhl@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
18 years ago[PATCH] USB: ftdi_sio: new IDs for Teratronik devices
Ian Abbott [Tue, 13 Dec 2005 16:18:47 +0000 (16:18 +0000)]
[PATCH] USB: ftdi_sio: new IDs for Teratronik devices

This patch adds vendor and product IDs to the ftdi_sio driver's device
ID table for two devices from teratronik.de.  The device IDs were
submitted by O. Wölfelschneider of Teratronik Elektronische Systeme
GmbH.

The charset of the patch is latin-1, same as the original files.

Please apply, thanks!  (I've tried to avoid a clash with Andrew Morton's
patch to add support for Posiflex PP-7700 printer to the same driver.)

Signed-off-by: Ian Abbott <abbotti@mev.co.uk>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
18 years ago[PATCH] USB: Support for Posiflex PP-7000 retail printer in Linux
Pavel Fedin [Fri, 9 Dec 2005 06:30:59 +0000 (09:30 +0300)]
[PATCH] USB: Support for Posiflex PP-7000 retail printer in Linux

This little patch adds recognition of Posiflex PP-7000 retail printer to
ftdo_sio module. The printer uses FT232BM bridge programmed with custom
VID/PID.  The patch posted to lkml and sf.net was for 2.6.11.1 kernel,
here is one reworked for 2.6.12.

Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
18 years ago[PATCH] USB: add driver for ATI/Philips USB RF remotes
Ville Syrjälä [Sat, 10 Dec 2005 18:30:54 +0000 (20:30 +0200)]
[PATCH] USB: add driver for ATI/Philips USB RF remotes

Summary: Driver for ATI/Philips USB RF remotes

This is a new input driver for ATI/Philips USB RF remotes (eg. ATI
Remote Wonder II).

Signed-off-by: Ville Syrjälä <syrjala@sci.fi>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
18 years ago[PATCH] USB: Use ARRAY_SIZE macro
Tobias Klauser [Sun, 11 Dec 2005 15:20:08 +0000 (16:20 +0100)]
[PATCH] USB: Use ARRAY_SIZE macro

Use ARRAY_SIZE macro instead of sizeof(x)/sizeof(x[0]) and remove
duplicates of ARRAY_SIZE. Some trailing whitespaces are also removed.

Patch is compile-tested on i386.

Signed-off-by: Tobias Klauser <tklauser@nuerscht.ch>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
18 years ago[PATCH] USB: fix local variable clash
Alan Stern [Tue, 13 Dec 2005 15:32:13 +0000 (10:32 -0500)]
[PATCH] USB: fix local variable clash

This patch (as621) fixes a local variable conflict I accidently
introduced into usb_set_configuration.

Signed-off-by: Alan Stern <stern@rowland.harvard.edu>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
18 years ago[PATCH] USB: usb-storage: add debug entry for REPORT LUNS
Paul Walmsley [Wed, 30 Nov 2005 21:57:45 +0000 (13:57 -0800)]
[PATCH] USB: usb-storage: add debug entry for REPORT LUNS

Bugs involving the REPORT LUNS SCSI-3 command are much easier to track
down if usb-storage displays the command's name, rather than "(Unknown
command)".

Signed-off-by: Paul Walmsley <paul@booyaka.com>
Cc: <mdharm-usb@one-eyed-alien.net>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
18 years ago[PATCH] USB Storage: update MAINTAINERS
Matthew Dharm [Mon, 5 Dec 2005 06:03:47 +0000 (22:03 -0800)]
[PATCH] USB Storage: update MAINTAINERS

Someone recently pointed out to me that the MAINTAINERS entry for
usb-storage was, perhaps, in need of changing.

Signed-off-by: Matthew Dharm <mdharm-usb@one-eyed-alien.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
18 years ago[PATCH] USB Storage: add alauda support
Matthew Dharm [Mon, 5 Dec 2005 06:02:44 +0000 (22:02 -0800)]
[PATCH] USB Storage: add alauda support

This patch adds another usb-storage subdriver, which supports two fairly
old dual-XD/SmartMedia reader-writers (USB1.1 devices).

This driver was written by Daniel Drake <dsd@gentoo.org> -- he notes
that he wrote this driver without specs, however a vendor-supplied GPL
driver for the previous generation of products ("sma03") did prove to be
quite useful, as did the sddr09 driver which also has to deal with
low-level physical block layout on SmartMedia.

The original patch has been reformed by me, as it clashed with the
libusual patches.

We really need to consolidate some of this common SmartMedia code, and
get together with the MTD guys to share it with them as well.

Signed-off-by: Daniel Drake <dsd@gentoo.org>
Signed-off-by: Matthew Dharm <mdharm-usb@one-eyed-alien.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
18 years ago[PATCH] USB Storage: more sddr09 cleanups
Matthew Dharm [Mon, 5 Dec 2005 05:59:45 +0000 (21:59 -0800)]
[PATCH] USB Storage: more sddr09 cleanups

This is the third of three patches to prepare the sddr09 subdriver for
conversion to the Sim-SCSI framework.  This patch (as596) moves the
computation of the LBA to the start of the read/write routines, so that
addresses completely beyond the end of the device can be detected and
reported differently from transfers that are partially within the
device's capacity.

Signed-off-by: Alan Stern <stern@rowland.harvard.edu>
Acked-by: Andries Brouwer <Andries.Brouwer@cwi.nl>
Signed-off-by: Matthew Dharm <mdharm-usb@one-eyed-alien.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
18 years ago[PATCH] USB Storage: sddr09 cleanups
Matthew Dharm [Mon, 5 Dec 2005 05:58:52 +0000 (21:58 -0800)]
[PATCH] USB Storage: sddr09 cleanups

This is the second of three patches to prepare the sddr09 subdriver for
conversion to the Sim-SCSI framework.  This patch (as595) updates the
code to use standard error values for return codes instead of our
special-purpose USB_STOR_TRANSPORT_... codes.  The reverse update is
then needed in the transport routine, but with the Sim-SCSI framework
that routine will go away.

Signed-off-by: Alan Stern <stern@rowland.harvard.edu>
Acked-by: Andries Brouwer <Andries.Brouwer@cwi.nl>
Signed-off-by: Matthew Dharm <mdharm-usb@one-eyed-alien.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
18 years ago[PATCH] USB Storage: cleanups of sddr09
Matthew Dharm [Mon, 5 Dec 2005 05:57:51 +0000 (21:57 -0800)]
[PATCH] USB Storage: cleanups of sddr09

This is the first of three patches to prepare the sddr09 subdriver for
conversion to the Sim-SCSI framework.  This patch (as594) straightens
out the initialization procedures and headers:

Some ugly code from usb.c was moved into sddr09.c.

Set-up of the private data structures was moved into the
initialization routine.

The connection between the "dpcm" version and the standalone
version was clarified.

A private declaration was moved from a header file into the
subdriver's .c file.

Signed-off-by: Alan Stern <stern@rowland.harvard.edu>
Acked-by: Andries Brouwer <Andries.Brouwer@cwi.nl>
Signed-off-by: Matthew Dharm <mdharm-usb@one-eyed-alien.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
18 years ago[PATCH] USB Storage: make OneTouch PM-aware
Matthew Dharm [Mon, 5 Dec 2005 05:56:51 +0000 (21:56 -0800)]
[PATCH] USB Storage: make OneTouch PM-aware

The OneTouch subdriver submits its own interrupt URB for notifications
about button presses.  Consequently it needs to know about suspend and
resume events, so it can cancel or restart the URB.

This patch (as593) adds a hook to struct us_data, to be used for
notifying subdrivers about Power Management events, and it implements
the hook in the OneTouch driver.

Signed-off-by: Alan Stern <stern@rowland.harvard.edu>
Signed-off-by: Nick Sillik <n.sillik@temple.edu>
Signed-off-by: Matthew Dharm <mdharm-usb@one-eyed-alien.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
18 years ago[PATCH] USB: Let usbmon collect less garbage
Pete Zaitcev [Sun, 4 Dec 2005 05:52:10 +0000 (21:52 -0800)]
[PATCH] USB: Let usbmon collect less garbage

Alan Stern pointed out that (in 2.6 kernel) one successful submission results
in one callback, even for ISO-out transfers. Thus, the silly check can be
removed from usbmon. This reduces the amount of garbage printed in case
of ISO and Interrupt transfers.

Signed-off-by: Pete Zaitcev <zaitcev@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
18 years ago[PATCH] USB: UHCI: edit some comments
Alan Stern [Wed, 30 Nov 2005 22:16:19 +0000 (17:16 -0500)]
[PATCH] USB: UHCI: edit some comments

This patch (as615b) edits a large number of comments in the uhci-hcd code,
mainly removing excess apostrophes.

Signed-off-by: Alan Stern <stern@rowland.harvard.edu>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
18 years ago[PATCH] USB: UHCI: change uhci_explen macro
Alan Stern [Wed, 30 Nov 2005 16:57:51 +0000 (11:57 -0500)]
[PATCH] USB: UHCI: change uhci_explen macro

This patch (as616) changed the uhci_explen macro in uhci-hcd.h so that
it now accepts the desired length, rather than length - 1 with special
handling for 0.  This also fixes a minor bug that would show up only
when a driver submits a 0-length bulk URB.

Signed-off-by: Alan Stern <stern@rowland.harvard.edu>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
18 years ago[PATCH] USB: ati_remote: use time_before() and friends
Marcelo Feitoza Parisi [Wed, 30 Nov 2005 21:50:39 +0000 (00:50 +0300)]
[PATCH] USB: ati_remote: use time_before() and friends

They deal with wrapping correctly and are nicer to read.

Signed-off-by: Marcelo Feitoza Parisi <marcelo@feitoza.com.br>
Signed-off-by: Alexey Dobriyan <adobriyan@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
18 years ago[PATCH] USB: pl2303_update_line_status data length fix
Horst Schirmeier [Thu, 28 Jul 2005 13:32:20 +0000 (15:32 +0200)]
[PATCH] USB: pl2303_update_line_status data length fix

Minimum data length must be UART_STATE + 1, as data[UART_STATE] is being
accessed for the new line_state. Although PL-2303 hardware is not
expected to send data with exactly UART_STATE length, this keeps it on
the safe side.

Signed-off-by: Horst Schirmeier <horst@schirmeier.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
18 years ago[PATCH] USB Storage: add unusual_devs entry for NIKON Coolpix 2000
fabien COSSE [Wed, 30 Nov 2005 09:16:00 +0000 (01:16 -0800)]
[PATCH] USB Storage: add unusual_devs entry for NIKON Coolpix 2000

This patch adds an unusual_devs.h entry for NIKON Coolpix 2000 camera
wich cause error: "Not Ready: Medium not present"
Works fine with th patched kernel...

Here are the informations in /proc/bus/usb/devices:

T:  Bus=02 Lev=01 Prnt=01 Port=00 Cnt=01 Dev#=  2 Spd=12  MxCh= 0
D:  Ver= 1.10 Cls=00(>ifc ) Sub=00 Prot=00 MxPS= 8 #Cfgs=  1
P:  Vendor=04b0 ProdID=0301 Rev= 0.10
S:  Manufacturer=NIKON
S:  Product=NIKON DSC E2000
C:* #Ifs= 1 Cfg#= 1 Atr=c0 MxPwr=  0mA
I:  If#= 0 Alt= 0 #EPs= 3 Cls=08(stor.) Sub=06 Prot=50
Driver=usb-storage
E:  Ad=01(O) Atr=02(Bulk) MxPS=  64 Ivl=0ms
E:  Ad=82(I) Atr=02(Bulk) MxPS=  64 Ivl=0ms
E:  Ad=83(I) Atr=03(Int.) MxPS=   8 Ivl=16ms

Signed-off-by: Fabien COSSE <fabien.cosse@wanadoo.fr>
Signed-off-by: Phil Dibowitz <phil@ipom.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
18 years ago[PATCH] USB: Correct ohci-pxa27x suspend/resume struct confusion
Richard Purdie [Mon, 28 Nov 2005 22:15:46 +0000 (22:15 +0000)]
[PATCH] USB: Correct ohci-pxa27x suspend/resume struct confusion

The device data in ohci-pxa27x is a struct hcd, not a struct ohci_hcd.
This correct the suspend/resume calls to account for this and adds some
code to invalidate the platform data when the module is removed.

Signed-off-by: Richard Purdie <rpurdie@rpsys.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
18 years ago[PATCH] USB: mark various usb tables const
Arjan van de Ven [Tue, 29 Nov 2005 08:43:42 +0000 (09:43 +0100)]
[PATCH] USB: mark various usb tables const

patch below marks various USB tables and variables as const so that they
end up in .rodata section and don't cacheline share with things that get
written to. For the non-array variables it also allows gcc to optimize
more.

Signed-off-by: Arjan van de Ven <arjan@infradead.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
18 years ago[PATCH] USB: drivers/usb/misc/sisusbvga/sisusb.c: remove dead code
Adrian Bunk [Tue, 29 Nov 2005 13:01:55 +0000 (14:01 +0100)]
[PATCH] USB: drivers/usb/misc/sisusbvga/sisusb.c: remove dead code

The Coverity checker found this dead code.

Signed-off-by: Adrian Bunk <bunk@stusta.de>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
18 years ago[PATCH] USB: Don't assume root-hub resume succeeds
Alan Stern [Tue, 29 Nov 2005 17:13:31 +0000 (12:13 -0500)]
[PATCH] USB: Don't assume root-hub resume succeeds

This patch (as614) makes a small change to the part of the hub driver
responsible for remote wakeup of root hubs.  When these wakeups occur
the driver is suspended, and in case the resume fails the driver should
remain suspended -- it shouldn't try to proceed with its normal
processing.

This will hardly ever matter in normal use, but it did crop up while I
was debugging a different problem.

Signed-off-by: Alan Stern <stern@rowland.harvard.edu>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
18 years ago[PATCH] USB Gadget: dummy_hcd: updates to hcd->state
Alan Stern [Tue, 29 Nov 2005 17:08:15 +0000 (12:08 -0500)]
[PATCH] USB Gadget: dummy_hcd: updates to hcd->state

This patch (as613) moves the updates to hcd->state in the dummy_hcd
driver to where they now belong.  It also uses the new
HC_FLAG_HW_ACCESSIBLE flag in a way that simulates a real PCI
controller, and it adds checks for attempts to resume the bus while the
controller is suspended or to suspend the controller while the bus is
active.

Signed-off-by: Alan Stern <stern@rowland.harvard.edu>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
18 years ago[PATCH] USB Gadget: file_storage: remove "volatile" declarations
Alan Stern [Tue, 29 Nov 2005 17:04:24 +0000 (12:04 -0500)]
[PATCH] USB Gadget: file_storage: remove "volatile" declarations

This patch (as612) removes the "volatile" declarations from the
file-storage gadget.  It turns out that they aren't needed for anything
much; adding a few memory barriers does a sufficient job.

The patch also removes a wait_queue.  Not much point having a queue when
only one task is ever going to be on it!

Signed-off-by: Alan Stern <stern@rowland.harvard.edu>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
18 years ago[PATCH] USB: don't allocate dma pools for PIO HCDs
Chris Humbert [Mon, 28 Nov 2005 17:29:23 +0000 (09:29 -0800)]
[PATCH] USB: don't allocate dma pools for PIO HCDs

USB: don't allocate dma pools for PIO HCDs

hcd_buffer_alloc() and hcd_buffer_free() have a similar dma_mask
check and revert to kmalloc()/kfree(), but hcd_buffer_create()
doesn't check dma_mask and allocates unused dma pools.

Signed-off-by: Chris Humbert <mahadri-kernel@drigon.com>
Acked-by: David Brownell <david-b@pacbell.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
18 years ago[PATCH] USB: input/touchkitusb: handle multiple packets
Daniel Ritz [Sun, 27 Nov 2005 21:23:38 +0000 (22:23 +0100)]
[PATCH] USB: input/touchkitusb: handle multiple packets

Some versions of the controller seem to put multiple report packet into a
single urb. also it can happen that a packet is split across multiple urbs.
unpatched you get a jumpy cursor on some screens.
the patch does:
- handle multiple packets per urb
- handle packets split across multiple urb
- check packet type
- cleanups

Signed-off-by: Daniel Ritz <daniel.ritz@gmx.ch>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
18 years ago[PATCH] USB: usbserial: race-condition fix.
Luiz Fernando Capitulino [Mon, 28 Nov 2005 21:16:07 +0000 (19:16 -0200)]
[PATCH] USB: usbserial: race-condition fix.

There is a race-condition in usb-serial driver that can be triggered if
a processes does 'port->tty->driver_data = NULL' in serial_close() while
other processes is in kernel-space about to call serial_ioctl() on the
same port.

This happens because a process can open the device while there is
another one closing it.

The patch below fixes that by adding a semaphore to ensure that no
process will open the device while another process is closing it.

Note that we can't use spinlocks here, since serial_open() and
serial_close() can sleep.

Signed-off-by: Luiz Capitulino <lcapitulino@mandriva.com.br>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
18 years ago[PATCH] USB: usbserial: Adds missing checks and bug fix.
Luiz Fernando Capitulino [Mon, 28 Nov 2005 21:16:05 +0000 (19:16 -0200)]
[PATCH] USB: usbserial: Adds missing checks and bug fix.

Checks if 'port' is NULL before using it in all tty operations, this
can avoid NULL pointer dereferences.

Signed-off-by: Luiz Capitulino <lcapitulino@mandriva.com.br>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
18 years ago[PATCH] USB: isp116x-hcd: minor cleanup
Olav Kongas [Fri, 25 Nov 2005 23:58:36 +0000 (01:58 +0200)]
[PATCH] USB: isp116x-hcd: minor cleanup

When going to suspend, there's no point in setting HC state in
host controller driver as USB core takes care of this.

Signed-off-by: Olav Kongas <ok@artecdesign.ee>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
18 years ago[PATCH] USB: Cleanups for usb gadget mass-storage
Pavel Machek [Sun, 27 Nov 2005 14:47:28 +0000 (09:47 -0500)]
[PATCH] USB: Cleanups for usb gadget mass-storage

Remove useless initalizers.

Signed-off-by: Pavel Machek <pavel@suse.cz>
Signed-off-by: Alan Stern <stern@rowland.harvard.edu>
18 years ago[PATCH] USB: Store port number in usb_device
Alan Stern [Wed, 23 Nov 2005 17:09:52 +0000 (12:09 -0500)]
[PATCH] USB: Store port number in usb_device

This patch (as610) adds a field to struct usb_device to store the device's
port number.  This allows us to remove several loops in the hub driver
(searching for a particular device among all the entries in the parent's
array of children).

Signed-off-by: Alan Stern <stern@rowland.harvard.edu>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
18 years ago[PATCH] USB: Consider power budget when choosing configuration
Alan Stern [Wed, 23 Nov 2005 17:03:12 +0000 (12:03 -0500)]
[PATCH] USB: Consider power budget when choosing configuration

This patch (as609) changes the way we keep track of power budgeting for
USB hubs and devices, and it updates the choose_configuration routine to
take this information into account.  (This is something we should have
been doing all along.)  A new field in struct usb_device holds the amount
of bus current available from the upstream port, and the usb_hub structure
keeps track of the current available for each downstream port.

Two new rules for configuration selection are added:

Don't select a self-powered configuration when only bus power
is available.

Don't select a configuration requiring more bus power than is
available.

However the first rule is #if-ed out, because I found that the internal
hub in my HP USB keyboard claims that its only configuration is
self-powered.  The rule would prevent the configuration from being chosen,
leaving the hub & keyboard unconfigured.  Since similar descriptor errors
may turn out to be fairly common, it seemed wise not to include a rule
that would break automatic configuration unnecessarily for such devices.

The second rule may also trigger unnecessarily, although this should be
less common.  More likely it will annoy people by sometimes failing to
accept configurations that should never have been chosen in the first
place.

The patch also changes usbcore's reaction when no configuration is
suitable.  Instead of raising an error and rejecting the device, now
the core will simply leave the device unconfigured.  People can always
work around such problems by installing configurations manually through
sysfs.

Signed-off-by: Alan Stern <stern@rowland.harvard.edu>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
18 years ago[PATCH] USB: One potential problem in gadget/serial.c
Fengwei Yin [Fri, 18 Nov 2005 02:35:36 +0000 (10:35 +0800)]
[PATCH] USB: One potential problem in gadget/serial.c

It looks like that the gs_serial module maybe sleep with spinlock in gs_close.
Sometimes, system hang when I remove the gs_serial module.

From: Fengwei Yin <xaityyy@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
18 years ago[PATCH] USB: small cleanups
Adrian Bunk [Sun, 20 Nov 2005 22:56:11 +0000 (23:56 +0100)]
[PATCH] USB: small cleanups

This patch contains the following cleanups:
- make needlessly global functions static
- every file should #include the headers containing the prototypes for
  it's global functions

Signed-off-by: Adrian Bunk <bunk@stusta.de>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
18 years ago[PATCH] USB: Fix locking for USB suspend/resume
Alan Stern [Mon, 21 Nov 2005 16:58:07 +0000 (11:58 -0500)]
[PATCH] USB: Fix locking for USB suspend/resume

The earlier USB locking updates didn't touch the suspend/resume
routines.  They need updating as well, since now the caller holds the
device semaphore.  This patch (as608) makes the necessary changes.  It
also adds a line to store the correct power state when a device is
resumed, something which was unaccountably missing.

Signed-off-by: Alan Stern <stern@rowland.harvard.edu>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
18 years ago[PATCH] USB: Disconnect children during hub unbind
Alan Stern [Fri, 18 Nov 2005 17:06:34 +0000 (12:06 -0500)]
[PATCH] USB: Disconnect children during hub unbind

This patch (as606b) is an updated version of my earlier patch to
disconnect children from a hub device when the hub driver is unbound.
Thanks to the changes in the driver core locking, we now know that the
entire hub device (and not just the interface) is locked whenever the
hub driver's disconnect method runs.  Hence it is safe to disconnect the
child device structures immediately instead of deferring the job.

The earlier version of the patch neglected to disable the hub's ports.
We don't want to forget that; otherwise we'd end up with live devices
using addresses that have been recycled.  This update adds the necessary
code.

Signed-off-by: Alan Stern <stern@rowland.harvard.edu>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
18 years ago[PATCH] USB: Remove USB private semaphore
Alan Stern [Thu, 17 Nov 2005 22:10:32 +0000 (17:10 -0500)]
[PATCH] USB: Remove USB private semaphore

This patch (as605) removes the private udev->serialize semaphore,
relying instead on the locking provided by the embedded struct device's
semaphore.  The changes are confined to the core, except that the
usb_trylock_device routine now uses the return convention of
down_trylock rather than down_read_trylock (they return opposite values
for no good reason).

A couple of other associated changes are included as well:

Now that we aren't concerned about HCDs that avoid using the
hcd glue layer, usb_disconnect no longer needs to acquire the
usb_bus_lock -- that can be done by usb_remove_hcd where it
belongs.

Devices aren't locked over the same scope of code in
usb_new_device and hub_port_connect_change as they used to be.
This shouldn't cause any trouble.

Along with the preceding driver core patch, this needs a lot of testing.

Signed-off-by: Alan Stern <stern@rowland.harvard.edu>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
18 years ago[PATCH] USB: remove .owner field from struct usb_driver
Greg Kroah-Hartman [Mon, 21 Nov 2005 22:53:03 +0000 (14:53 -0800)]
[PATCH] USB: remove .owner field from struct usb_driver

It is no longer needed, so let's remove it, saving a bit of memory.

Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
18 years ago[PATCH] USB: make registering a usb driver automatically set the module owner
Greg Kroah-Hartman [Mon, 21 Nov 2005 22:53:03 +0000 (14:53 -0800)]
[PATCH] USB: make registering a usb driver automatically set the module owner

This fixes the driver that forgot to set the module owner up.  Now we
can remove the unneeded pointer from the usb driver structure.  The idea
for how to do this was from Al Viro, who did this for the PCI drivers.

Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
18 years ago[PATCH] USB: allow usb drivers to disable dynamic ids
Greg Kroah-Hartman [Wed, 16 Nov 2005 21:41:28 +0000 (13:41 -0800)]
[PATCH] USB: allow usb drivers to disable dynamic ids

This lets drivers, like the usb-serial ones, disable the ability to add
ids from sysfs.

The usb-serial drivers are "odd" in that they are really usb-serial bus
drivers, not usb bus drivers, so the dynamic id logic will have to go
into the usb-serial bus core for those drivers to get that ability.

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