kernel/kernel-generic.git
14 years agoUSB: io_ti.c: don't return 0 if writing the download record failed
Roel Kluin [Tue, 20 Jul 2010 22:29:08 +0000 (15:29 -0700)]
USB: io_ti.c: don't return 0 if writing the download record failed

If the write download record failed we shouldn't return 0.

Signed-off-by: Roel Kluin <roel.kluin@gmail.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agoUSB: host: Remove dead CONFIG_ARCH_KARO
Christian Dietrich [Tue, 20 Jul 2010 07:08:01 +0000 (09:08 +0200)]
USB: host: Remove dead CONFIG_ARCH_KARO

CONFIG_ARCH_KARO doesn't exist in Kconfig and is never defined anywhere
else, therefore removing all references for it from the source code.

Signed-off-by: Christian Dietrich <qy03fugy@stud.informatik.uni-erlangen.de>
Acked-by: Ryan Mallon <ryan@bluewatersys.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agoUSB: serial: enable async suspend for usb serial port device
Ming Lei [Tue, 13 Jul 2010 15:56:24 +0000 (23:56 +0800)]
USB: serial: enable async suspend for usb serial port device

Usb serial port device is child of its usb interface device, so
we can enable async suspend of usb serial port device to speedup
system suspend.

Signed-off-by: Ming Lei <tom.leiming@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agoUSB: musb: forward debug mode feature to gadget
stephane duverger [Mon, 12 Jul 2010 16:43:47 +0000 (18:43 +0200)]
USB: musb: forward debug mode feature to gadget

This is a patch for the musb usb controller.

It allows forwarding of the debug mode feature to its gadget in order
to be able to act as an ehci debug device.

This patch has been tested on an IGEPv2 board running a 2.6.35-rc1
kernel.

Signed-off-by: Stephane Duverger <stephane.duverger@gmail.com>
Cc: Felipe Balbi <felipe.balbi@nokia.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agoUSB: EHCI Debug Port Device Gadget
stephane duverger [Mon, 12 Jul 2010 16:37:53 +0000 (18:37 +0200)]
USB: EHCI Debug Port Device Gadget

This is a patch that implements an USB EHCI Debug Device using the
Gadget API. This patch applies to a 2.6.35-rc3 kernel.

The gadget needs a compliant usb controller that forwards the
USB_DEVICE_DEBUG_MODE feature to its gadget.

The gadget provides two configuration modes, one that only printk() the
received data, and one that exposes a serial device to userland
(/dev/ttyGSxxx).

The gadget has been tested on an IGEPv2 board running a 2.6.35-rc1
kernel. The debug port was fed on the host side by a 2.6.34 kernel.

Signed-off-by: Stephane Duverger <stephane.duverger@gmail.com>
Cc: David Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agoUSB: gadget: Remove unnecessary casts of private_data
Joe Perches [Mon, 12 Jul 2010 20:50:11 +0000 (13:50 -0700)]
USB: gadget: Remove unnecessary casts of private_data

Signed-off-by: Joe Perches <joe@perches.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agoUSB: misc: Remove unnecessary casts of private_data
Joe Perches [Mon, 12 Jul 2010 20:50:12 +0000 (13:50 -0700)]
USB: misc: Remove unnecessary casts of private_data

Signed-off-by: Joe Perches <joe@perches.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agoUSB: usb-skeleton: Remove unnecessary casts of private_data
Joe Perches [Mon, 12 Jul 2010 20:50:13 +0000 (13:50 -0700)]
USB: usb-skeleton: Remove unnecessary casts of private_data

Signed-off-by: Joe Perches <joe@perches.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agoUSB: autoconvert trivial BKL users to private mutex
Arnd Bergmann [Sun, 11 Jul 2010 21:18:56 +0000 (23:18 +0200)]
USB: autoconvert trivial BKL users to private mutex

All these files use the big kernel lock in a trivial
way to serialize their private file operations,
typically resulting from an earlier semi-automatic
pushdown from VFS.

None of these drivers appears to want to lock against
other code, and they all use the BKL as the top-level
lock in their file operations, meaning that there
is no lock-order inversion problem.

Consequently, we can remove the BKL completely,
replacing it with a per-file mutex in every case.
Using a scripted approach means we can avoid
typos.

file=$1
name=$2
if grep -q lock_kernel ${file} ; then
    if grep -q 'include.*linux.mutex.h' ${file} ; then
            sed -i '/include.*<linux\/smp_lock.h>/d' ${file}
    else
            sed -i 's/include.*<linux\/smp_lock.h>.*$/include <linux\/mutex.h>/g' ${file}
    fi
    sed -i ${file} \
        -e "/^#include.*linux.mutex.h/,$ {
                1,/^\(static\|int\|long\)/ {
                     /^\(static\|int\|long\)/istatic DEFINE_MUTEX(${name}_mutex);

} }"  \
    -e "s/\(un\)*lock_kernel\>[ ]*()/mutex_\1lock(\&${name}_mutex)/g" \
    -e '/[      ]*cycle_kernel_lock();/d'
else
    sed -i -e '/include.*\<smp_lock.h\>/d' ${file}  \
                -e '/cycle_kernel_lock()/d'
fi

Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agousb: storage: freecom: Fixed several coding style issues.
Martin Enderleit [Sat, 10 Jul 2010 14:50:12 +0000 (16:50 +0200)]
usb: storage: freecom: Fixed several coding style issues.

Fixed several coding style issues in freecom.c.

Signed-off-by: Martin Enderleit <menderleit@gmail.com>
Cc: Matthew Dharm <mdharm-usb@one-eyed-alien.net>
Cc: Daniel Mack <daniel@caiaq.de>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agoUSB: musb: do not override DMA mode in channel program
Anand Gadiyar [Thu, 8 Jul 2010 11:04:55 +0000 (16:34 +0530)]
USB: musb: do not override DMA mode in channel program

There is no reason for the DMA channel program to override the
DMA mode passed down by its caller. Use the passed parameter
directly, and let the caller handle the decision on which mode
is to be used.

Signed-off-by: Anand Gadiyar <gadiyar@ti.com>
Acked-by: Felipe Balbi <felipe.balbi@nokia.com>
Signed-off-by: Ajay Kumar Gupta <ajay.gupta@ti.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agoUSB: musb: Kill board specific pinmux from driver file
Anand Gadiyar [Thu, 8 Jul 2010 11:04:54 +0000 (16:34 +0530)]
USB: musb: Kill board specific pinmux from driver file

This pin-muxing is best done in the board files. The driver should
not do this explicitly.

Also, this code causes a warning to be thrown when OMAP2430 and OMAP3/4
support are enabled in the same kernel.

Signed-off-by: Anand Gadiyar <gadiyar@ti.com>
Acked-by: Felipe Balbi <felipe.balbi@nokia.com>
Signed-off-by: Ajay Kumar Gupta <ajay.gupta@ti.com>
Acked-by: Tony Lindgren <tony@atomide.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agoUSB: ehci_omap: fix device detect issue with modules
Ajay Kumar Gupta [Thu, 8 Jul 2010 08:33:02 +0000 (14:03 +0530)]
USB: ehci_omap: fix device detect issue with modules

Currently devices don't get detected automatically if the ehci
module is inserted 2nd time onward. We need to disconnect and
reconnect the device for it to get detected and enumerated.

Resetting the USB PHY using PHY reset comamnd over ULPI fixes
this issue. Tested on OMAP3EVM.

Signed-off-by: Ajay Kumar Gupta <ajay.gupta@ti.com>
Acked-by: Felipe Balbi <felipe.balbi@nokia.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agoUSB: ulpi: fix compilation warning
Ajay Kumar Gupta [Thu, 8 Jul 2010 08:33:01 +0000 (14:03 +0530)]
USB: ulpi: fix compilation warning

Fixes below compilation warning from ulpi.h

include/linux/usb/ulpi.h:145:
        warning: 'struct otg_io_access_ops' declared inside parameter list
include/linux/usb/ulpi.h:145:
         warning: its scope is only this definition or declaration,
         which is probably not what you want

Signed-off-by: Ajay Kumar Gupta <ajay.gupta@ti.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agoUSB: musb: fix compilation warning in host only mode
Ajay Kumar Gupta [Thu, 8 Jul 2010 08:33:00 +0000 (14:03 +0530)]
USB: musb: fix compilation warning in host only mode

Fixes below compilation warning when host only configuration is
selected.
drivers/usb/musb/musb_core.c: In function 'musb_stage0_irq':
drivers/usb/musb/musb_core.c:711: warning: unused variable 'mbase'

Signed-off-by: Ajay Kumar Gupta <ajay.gupta@ti.com>
Acked-by: Felipe Balbi <felipe.balbi@nokia.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agoUSB: musb: use correct register widths in register dumps
Anand Gadiyar [Thu, 8 Jul 2010 08:32:59 +0000 (14:02 +0530)]
USB: musb: use correct register widths in register dumps

DMA_ADDR and DMA_COUNT are 32-bit registers, not 16-bit.

Marking them as 16-bit in the table causes only the lower
16-bits to be dumped and this is misleading.

Signed-off-by: Anand Gadiyar <gadiyar@ti.com>
Acked-by: Felipe Balbi <felipe.balbi@nokia.com>
Signed-off-by: Ajay Kumar Gupta <ajay.gupta@ti.com>
Cc: stable <stable@kernel.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agoUSB: core: hcd-pci: use for_each_pci_dev()
Kulikov Vasiliy [Sat, 3 Jul 2010 16:04:47 +0000 (20:04 +0400)]
USB: core: hcd-pci: use for_each_pci_dev()

Use for_each_pci_dev() to simplify the code.

Signed-off-by: Kulikov Vasiliy <segooon@gmail.com>
Acked-by: Alan Stern <stern@rowland.harvard.edu>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agoUSB: serial: io_ti: Don't return 0 if writing the download record failed
Roel Kluin [Thu, 1 Jul 2010 22:36:43 +0000 (00:36 +0200)]
USB: serial: io_ti: Don't return 0 if writing the download record failed

If the write download record failed we shouldn't return 0.

Signed-off-by: Roel Kluin <roel.kluin@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agoUSB: gadget: storage_common: comments updated
Michal Nazarewicz [Mon, 5 Jul 2010 14:38:04 +0000 (16:38 +0200)]
USB: gadget: storage_common: comments updated

Updated comment to describe why printing macros are needed even
thought they are copied form the composite.h.  Also, made multiline
comments follow the coding standard.

Signed-off-by: Michal Nazarewicz <m.nazarewicz@samsung.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agoUSB: gadget: compilation issue: missing TASK_INTERRUPTIBLE
stephane duverger [Tue, 29 Jun 2010 14:57:25 +0000 (16:57 +0200)]
USB: gadget: compilation issue: missing TASK_INTERRUPTIBLE

Here is the patch for the following issue:

drivers/usb/gadget/u_serial.c: In function ‘gs_start_tx’:
drivers/usb/gadget/u_serial.c:369: error: ‘TASK_INTERRUPTIBLE’ undeclared (first use in this function)
drivers/usb/gadget/u_serial.c:369: error: (Each undeclared identifier is reported only once
drivers/usb/gadget/u_serial.c:369: error: for each function it appears in.)
drivers/usb/gadget/u_serial.c: In function ‘gs_rx_push’:
drivers/usb/gadget/u_serial.c:546: error: ‘TASK_INTERRUPTIBLE’ undeclared (first use in this function)
drivers/usb/gadget/u_serial.c: In function ‘gs_close’:
drivers/usb/gadget/u_serial.c:857: error: ‘TASK_INTERRUPTIBLE’ undeclared (first use in this function)
drivers/usb/gadget/u_serial.c:857: error: implicit declaration of function ‘signal_pending’
drivers/usb/gadget/u_serial.c:857: error: implicit declaration of function ‘schedule_timeout’
drivers/usb/gadget/u_serial.c: In function ‘gserial_cleanup’:
drivers/usb/gadget/u_serial.c:1190: error: ‘TASK_UNINTERRUPTIBLE’ undeclared (first use in this function)
drivers/usb/gadget/u_serial.c:1190: error: implicit declaration of function ‘schedule’
drivers/usb/gadget/u_serial.c: In function ‘gserial_disconnect’:
drivers/usb/gadget/u_serial.c:1311: error: ‘TASK_INTERRUPTIBLE’ undeclared (first use in this function)

Signed-off-by: Stephane Duverger <stephane.duverger@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agoUSB: EHCI: fix NULL pointer dererence in HCDs that use HCD_LOCAL_MEM
Andrea Righi [Mon, 28 Jun 2010 14:56:45 +0000 (16:56 +0200)]
USB: EHCI: fix NULL pointer dererence in HCDs that use HCD_LOCAL_MEM

If we use the HCD_LOCAL_MEM flag and dma_declare_coherent_memory() to
enforce the host controller's local memory utilization we also need to
disable native scatter-gather support, otherwise hcd_alloc_coherent() in
map_urb_for_dma() is called with urb->transfer_buffer == NULL, that
triggers a NULL pointer dereference.

We can also consider to add a WARN_ON() and return an error code to
better catch this problem in the future.

At the moment no driver seems to hit this bug, so I should
consider this a low-priority fix.

Signed-off-by: Andrea Righi <arighi@develer.com>
Acked-by: Alan Stern <stern@rowland.harvard.edu>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agoUSB: xhci: trivial: use ARRAY_SIZE
Kulikov Vasiliy [Mon, 28 Jun 2010 11:55:46 +0000 (15:55 +0400)]
USB: xhci: trivial: use ARRAY_SIZE

Change sizeof(x) / sizeof(*x) to ARRAY_SIZE(x).

Signed-off-by: Kulikov Vasiliy <segooon@gmail.com>
Cc: Sarah Sharp <sarah.a.sharp@linux.intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agoUSB: drivers/usb/Makefile: conditionally descend to 'early'
Nicolas Kaiser [Sun, 27 Jun 2010 15:27:51 +0000 (17:27 +0200)]
USB: drivers/usb/Makefile: conditionally descend to 'early'

Don't descend to the EARLY_PRINTK_DBGP directory
unless it is actually used.

Signed-off-by: Nicolas Kaiser <nikai@nikai.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agoUSB: usblp: fixed switch, brace, whitespace and spacing coding style issues
Nicolas Kaiser [Fri, 25 Jun 2010 18:25:37 +0000 (20:25 +0200)]
USB: usblp: fixed switch, brace, whitespace and spacing coding style issues

Fixed switch, brace, whitespace and spacing coding style issues.

Signed-off-by: Nicolas Kaiser <nikai@nikai.net>
Acked-by: Pete Zaitcev <zaitcev@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agoUSB: add runtime PM for PCI-based host controllers
Alan Stern [Fri, 25 Jun 2010 18:02:57 +0000 (14:02 -0400)]
USB: add runtime PM for PCI-based host controllers

This patch (as1386) adds runtime-PM support for PCI-based USB host
controllers.  By default autosuspend is disallowed; the user must
enable it by writing "auto" to the controller's power/control sysfs
attribute.

Signed-off-by: Alan Stern <stern@rowland.harvard.edu>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agoUSB: UHCI: add support for Intel's wakeup flags
Alan Stern [Fri, 25 Jun 2010 18:02:49 +0000 (14:02 -0400)]
USB: UHCI: add support for Intel's wakeup flags

This patch (as1396) adds code to uhci-hcd to support the
vendor-specific wakeup settings found in Intel's ICHx hardware.  A
couple of unnecessary memory barriers are removed.  And the root hub
isn't put back into the "suspended" state if power was lost during a
system sleep -- there's not much point in doing so because the root hub
will be resumed shortly.

Signed-off-by: Alan Stern <stern@rowland.harvard.edu>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agoUSB: fix race between root-hub wakeup & controller suspend
Alan Stern [Fri, 25 Jun 2010 18:02:35 +0000 (14:02 -0400)]
USB: fix race between root-hub wakeup & controller suspend

This patch (as1395) adds code to hcd_pci_suspend() for handling wakeup
races.  This is another general race pattern, similar to the "open
vs. unregister" race we're all familiar with.  Here, the race is
between suspending a device and receiving a wakeup request from one of
the device's suspended children.

In particular, if a root-hub wakeup is requested at about the same
time as the corresponding USB controller is suspended, and if the
controller is enabled for wakeup, then the controller should either
fail to suspend or else wake right back up again.

During system sleep this won't happen very much, especially since host
controllers generally aren't enabled for wakeup during sleep.  However
it is definitely an issue for runtime PM.  Something like this will be
needed to prevent the controller from autosuspending while waiting for
a root-hub resume to take place.  (That is, in fact, the common case,
for which there is an extra test.)

Signed-off-by: Alan Stern <stern@rowland.harvard.edu>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agoUSB: controller resume should check the root hub
Alan Stern [Fri, 25 Jun 2010 18:02:24 +0000 (14:02 -0400)]
USB: controller resume should check the root hub

This patch (as1394) adds code to ehci-hcd, ohci-hcd, and uhci-hcd for
automatically resuming the root hub when the controller is resumed, if
the root hub has a wakeup request pending on some port.

During resume from system sleep this doesn't matter, because the root
hubs will naturally be resumed along with every other device in the
system.  However it _will_ matter for runtime PM: If the controller is
suspended and a remote wakeup request is received then the controller
will autoresume, but we need to ensure that the root hub also
autoresumes.  Otherwise the wakeup request would be ignored, the
controller would go back to sleep, and the cycle would repeat a large
number of times (I saw this happen before the patch was written).

Signed-off-by: Alan Stern <stern@rowland.harvard.edu>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agoUSB: add do_wakeup parameter for PCI HCD suspend
Alan Stern [Fri, 25 Jun 2010 18:02:14 +0000 (14:02 -0400)]
USB: add do_wakeup parameter for PCI HCD suspend

This patch (as1385) adds a "do_wakeup" parameter to the pci_suspend
method used by PCI-based host controller drivers.  ehci-hcd in
particular needs to know whether or not to enable wakeup when
suspending a controller.  Although that information is currently
available through device_may_wakeup(), when support is added for
runtime suspend this will no longer be true.

Signed-off-by: Alan Stern <stern@rowland.harvard.edu>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agoUSB: move PCI HCD resume routine
Alan Stern [Fri, 25 Jun 2010 18:02:03 +0000 (14:02 -0400)]
USB: move PCI HCD resume routine

This patch (as1384) moves the resume_common() routine in hcd-pci.c a
little higher in the source file to avoid forward references in an
upcoming patch.  It also replaces the "hibernated" argument with a
more general "event" argument, which will be useful when the routine
is called during a runtime resume.

Signed-off-by: Alan Stern <stern@rowland.harvard.edu>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agoUSB: refactor the powermac-specific ASIC clock code
Alan Stern [Fri, 25 Jun 2010 18:01:49 +0000 (14:01 -0400)]
USB: refactor the powermac-specific ASIC clock code

This patch (as1383) takes the powermac-specific code from the PCI HCD
glue layer and encapsulates it in its own subroutine.

Signed-off-by: Alan Stern <stern@rowland.harvard.edu>
Acked-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agoUSB: gadget: g_fs: code cleanup
Michal Nazarewicz [Fri, 25 Jun 2010 14:29:27 +0000 (16:29 +0200)]
USB: gadget: g_fs: code cleanup

This commit cleans the g_fs gadget hopefully making it more
readable.  This is achieved by usage of the usb_string_ids_tab()
function for batch string IDs registration as well as
generalising configuration so that a single routine is
used to add each configuration and bind interfaces.  As an
effect, the code is shorter and has fewer #ifdefs.

Moreover, in some circumstances previous code #defined
CONFIG_USB_FUNCTIONFS_GENERIC macro to prevent a situation
where gadget with no configurations is built.  This code removes
the #define form source code and achieves the same effect using
select in Kconfig.

This patch also changes wording and names of the Kconfig options.

Signed-off-by: Michal Nazarewicz <m.nazarewicz@samsung.com>
Signed-off-by: Kyungmin Park <kyungmin.park@samsung.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agoUSB: convert usb_hcd bitfields into atomic flags
Alan Stern [Tue, 22 Jun 2010 20:39:10 +0000 (16:39 -0400)]
USB: convert usb_hcd bitfields into atomic flags

This patch (as1393) converts several of the single-bit fields in
struct usb_hcd to atomic flags.  This is for safety's sake; not all
CPUs can update bitfield values atomically, and these flags are used
in multiple contexts.

The flag fields that are set only during registration or removal can
remain as they are, since non-atomic accesses at those times will not
cause any problems.

(Strictly speaking, the authorized_default flag should become atomic
as well.  I didn't bother with it because it gets changed only via
sysfs.  It can be done later, if anyone wants.)

Signed-off-by: Alan Stern <stern@rowland.harvard.edu>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agoUSB: gadget: section mismatch warning fixed
Michal Nazarewicz [Mon, 21 Jun 2010 11:57:04 +0000 (13:57 +0200)]
USB: gadget: section mismatch warning fixed

In may gadgets bind and bind like functions were in a init section
as they were only run during initialisation.  However, being
callback functions they were referenced from structures in “normal”
sections.  Changing the tag from “__init” to “__ref” fixes the
warnings.

Signed-off-by: Michal Nazarewicz <m.nazarewicz@samsung.com>
Cc: Kyungmin Park <kyungmin.park@samsung.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agoUSB: gadget: f_mass_storage: added eject callback
Michal Nazarewicz [Mon, 21 Jun 2010 11:57:09 +0000 (13:57 +0200)]
USB: gadget: f_mass_storage: added eject callback

Added pre_eject() and post_eject() callbacks which are
called before and after removable logical unit is ejected.
The first can prevent logical unit from being ejected.

This commit also changes the way callbacks are passed to
the function from gadget.  A fsg_operations structure has
been created which lists all callbacks -- this is passed
to the fsg_config.

This is important because it changes the way thread_exits()
callback is passed.

Signed-off-by: Michal Nazarewicz <m.nazarewicz@samsung.com>
Signed-off-by: Kyungmin Park <kyungmin.park@samsung.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agoUSB: gadget: composite: added disconnect callback
Michal Nazarewicz [Mon, 21 Jun 2010 11:57:08 +0000 (13:57 +0200)]
USB: gadget: composite: added disconnect callback

Added a disconnect() callback to composite devices which
is called by composite glue when its disconnect callback
is called by gadget.

Signed-off-by: Michal Nazarewicz <m.nazarewicz@samsung.com>
Signed-off-by: Kyungmin Park <kyungmin.park@samsung.com>
Acked-by: David Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agoUSB: gadget: g_multi: added documentation and INF files
Michal Nazarewicz [Mon, 21 Jun 2010 11:57:07 +0000 (13:57 +0200)]
USB: gadget: g_multi: added documentation and INF files

A short documentation of the g_multi driver along with INF
files for Windows XP SP3 are provided.

Signed-off-by: Michal Nazarewicz <m.nazarewicz@samsung.com>
Signed-off-by: Kyungmin Park <kyungmin.park@samsung.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agoUSB: gadget: g_serial: INF file updated
Michal Nazarewicz [Mon, 21 Jun 2010 11:57:06 +0000 (13:57 +0200)]
USB: gadget: g_serial: INF file updated

Updated the INF file for g_serial gadget.  It should work with
most recent Windows systems now.

Signed-off-by: Michal Nazarewicz <m.nazarewicz@samsung.com>
Signed-off-by: Kyungmin Park <kyungmin.park@samsung.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agoUSB: gadget: g_ether: updated INF file
Michal Nazarewicz [Mon, 21 Jun 2010 11:57:05 +0000 (13:57 +0200)]
USB: gadget: g_ether: updated INF file

Updated the INF file for the g_ether gadget.  It should work with most
recent Windows systems now.

Signed-off-by: Michal Nazarewicz <m.nazarewicz@samsung.com>
Signed-off-by: Kyungmin Park <kyungmin.park@samsung.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agoUSB: gadget: g_multi: code clean up and refactoring
Michal Nazarewicz [Mon, 21 Jun 2010 11:57:03 +0000 (13:57 +0200)]
USB: gadget: g_multi: code clean up and refactoring

The Multifunction Composite Gadget have been cleaned up
and refactored so hopefully it looks prettier and works
at least as good as before changes.

A Kconfig has also been fixed to make it impossible to build
FunctionFS gadget with no configurations.  With this patch, if
RNDIS is not chosen by the user CDC is force-selected.

Signed-off-by: Michal Nazarewicz <m.nazarewicz@samsung.com>
Signed-off-by: Kyungmin Park <kyungmin.park@samsung.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agoUSB: gadget: f_mass_storage: dead code removed
Michal Nazarewicz [Fri, 18 Jun 2010 13:59:43 +0000 (15:59 +0200)]
USB: gadget: f_mass_storage: dead code removed

The ep0req_name was never used in f_mass_storage hence it may
be safely removed from the code.  It was a leftover from File
Storage Gadget which used it for debug messages.

Signed-off-by: Michal Nazarewicz <m.nazarewicz@samsung.com>
Cc: David Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agoUSB: gadget: f_fs: use usb_string_ids_n()
Michal Nazarewicz [Wed, 16 Jun 2010 10:08:00 +0000 (12:08 +0200)]
USB: gadget: f_fs: use usb_string_ids_n()

Use usb_string_ids_n() function to simplify string ids
registeration.

Signed-off-by: Michal Nazarewicz <m.nazarewicz@samsung.com>
Signed-off-by: Kyungmin Park <kyungmin.park@samsung.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agoUSB: gadget: composite: usb_string_ids_*() functions added
Michal Nazarewicz [Wed, 16 Jun 2010 10:07:59 +0000 (12:07 +0200)]
USB: gadget: composite: usb_string_ids_*() functions added

usb_string_ids_tab() and usb_string_ids_n() functions added to
the composite framework.  The first accepts an array of
usb_string object and for each registeres a string id and the
second registeres a given number of ids and returns the first.

This may simplify string ids registration since gadgets and
composite functions won't have to call usb_string_id() several
times and each time check for errer status -- all this will be
done with a single call.

Signed-off-by: Michal Nazarewicz <m.nazarewicz@samsung.com>
Signed-off-by: Kyungmin Park <kyungmin.park@samsung.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agoUSB: gadget: f_fs: functionfs_add() renamed to functionfs_bind_config()
Michal Nazarewicz [Wed, 16 Jun 2010 10:07:58 +0000 (12:07 +0200)]
USB: gadget: f_fs: functionfs_add() renamed to functionfs_bind_config()

FunctionFS had a bit unique name for function used to add it
to USB configuration.  Renamed as to match naming convention
of other functions.

Signed-off-by: Michal Nazarewicz <m.nazarewicz@samsung.com>
Signed-off-by: Kyungmin Park <kyungmin.park@samsung.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agoUSB: gadget: f_mass_storage: fsg_add() renamed to fsg_bind_config()
Michal Nazarewicz [Wed, 16 Jun 2010 10:07:57 +0000 (12:07 +0200)]
USB: gadget: f_mass_storage: fsg_add() renamed to fsg_bind_config()

Mass Storage Function had a bit unique name for function
used to add it to USB configuration.  Renamed as to match
naming convention of other functions.

Signed-off-by: Michal Nazarewicz <m.nazarewicz@samsung.com>
Signed-off-by: Kyungmin Park <kyungmin.park@samsung.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agoUSB: gadget: g_mass_storage: static data instead of dynamic allocation
Michal Nazarewicz [Wed, 16 Jun 2010 10:07:56 +0000 (12:07 +0200)]
USB: gadget: g_mass_storage: static data instead of dynamic allocation

This patch changes msg_do_config() function so that it uses
a static object for a fsg_common structure instead of dynamically
allocated.  This is a micro-optimisation.

Signed-off-by: Michal Nazarewicz <m.nazarewicz@samsung.com>
Signed-off-by: Kyungmin Park <kyungmin.park@samsung.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agoUSB: mon: kill BKL usage
Arnd Bergmann [Tue, 1 Jun 2010 21:04:45 +0000 (23:04 +0200)]
USB: mon: kill BKL usage

compat_ioctl does not use the BKL, so I assume that
the native function does not need it either.

The open function is already protected by the
driver's mutex, the BKL is probably not needed
here either.

Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Cc: Alan Stern <stern@rowland.harvard.edu>
Cc: Pete Zaitcev <zaitcev@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agousb: gadget: Do not take BKL for gadget->ops->ioctl
Arnd Bergmann [Tue, 1 Jun 2010 21:04:44 +0000 (23:04 +0200)]
usb: gadget: Do not take BKL for gadget->ops->ioctl

There is no gadget driver in the tree that
actually implements the ioctl operation, so
obviously it is not necessary to hold the
BKL around the call.

Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Cc: David Brownell <dbrownell@users.sourceforge.net>
Cc: Michał Nazarewicz <m.nazarewicz@samsung.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agoUSB-BKL: Remove BKL use in uhci-debug
Andi Kleen [Tue, 1 Jun 2010 21:04:43 +0000 (23:04 +0200)]
USB-BKL: Remove BKL use in uhci-debug

BKL was not really needed, just came from earlier push downs.

The only part that's a bit dodgy is the lseek function. Would
need another lock or atomic access to fpos on 32bit?
Better to have a libfs lseek

Signed-off-by: Andi Kleen <ak@linux.intel.com>
Cc: Arnd Bergmann <arnd@arndb.de>
Cc: Alan Stern <stern@rowland.harvard.edu>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agoUSB-BKL: Remove BKL use for usb serial driver probing
Andi Kleen [Tue, 1 Jun 2010 21:04:42 +0000 (23:04 +0200)]
USB-BKL: Remove BKL use for usb serial driver probing

The usb serial driver initialization tried to use the BKL to stop
driver modules from unloading, but that didn't work anyways.

There was already some code to do proper try_module_get,
but it was conditional on having a new probe interface.
I checked all the low level drivers and they all have proper
.owner = THIS_MODULE, so it's ok to always use.

The other problem was the usb_serial_driver_list needing
protection by a lock. This was broken anyways because unregister
did not necessarily have the BKL.

I extended the extending table_lock mutex to protect this case too.

With these changes the BKL can be removed here.

Signed-off-by: Andi Kleen <ak@linux.intel.com>
Cc: Arnd Bergmann <arnd@arndb.de>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agoUSB-BKL: Convert usb_driver ioctl to unlocked_ioctl
Andi Kleen [Tue, 1 Jun 2010 21:04:41 +0000 (23:04 +0200)]
USB-BKL: Convert usb_driver ioctl to unlocked_ioctl

And audit all the users. None needed the BKL.  That was easy
because there was only very few around.

Tested with allmodconfig build on x86-64

Signed-off-by: Andi Kleen <ak@linux.intel.com>
Cc: Arnd Bergmann <arnd@arndb.de>
From: Andi Kleen <ak@linux.intel.com>

14 years agoUSB-BKL: Remove lock_kernel in usbfs update_sb()
Andi Kleen [Tue, 1 Jun 2010 21:04:40 +0000 (23:04 +0200)]
USB-BKL: Remove lock_kernel in usbfs update_sb()

The code this is attempting to lock against does not use the BKL,
so it's not needed.

Most likely this code is still broken/racy (Al Viro also thinks so),
but removing the BKL should not make it worse than before.

Signed-off-by: Andi Kleen <ak@linux.intel.com>
Cc: Arnd Bergmann <arnd@arndb.de>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agousb: atm: fixed spacing and indentation coding style issues
Nicolas Kaiser [Thu, 17 Jun 2010 09:55:49 +0000 (11:55 +0200)]
usb: atm: fixed spacing and indentation coding style issues

Fixed spacing and indentation coding style issues.

Signed-off-by: Nicolas Kaiser <nikai@nikai.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agoUSB: Add a serial number parameter to g_file_storage module
Yann Cantin [Sat, 5 Jun 2010 21:06:31 +0000 (23:06 +0200)]
USB: Add a serial number parameter to g_file_storage module

This patch add a serial number parameter to the g_file_storage
module. There's validity checks against the string passed to comply
with the specs.

Signed-off-by: Yann Cantin <yann.cantin@laposte.net>
Cc: Michał Nazarewicz <m.nazarewicz@samsung.com>
Cc: David Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agousb: conexant: fixed spacing and brace coding style issues
Nicolas Kaiser [Wed, 16 Jun 2010 16:56:05 +0000 (18:56 +0200)]
usb: conexant: fixed spacing and brace coding style issues

Fixed spacing and brace coding style issues.

Signed-off-by: Nicolas Kaiser <nikai@nikai.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agoUSB: gadget: langwell_udc.c: printk needs a (unsigned long long) cast for a dma_t
Joe Perches [Fri, 11 Jun 2010 02:20:43 +0000 (19:20 -0700)]
USB: gadget: langwell_udc.c: printk needs a (unsigned long long) cast for a dma_t

Signed-off-by: Joe Perches <joe@perches.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agoRevert "USB: Adding support for HTC Smartphones to ipaq"
Leann Ogasawara [Thu, 10 Jun 2010 22:49:24 +0000 (15:49 -0700)]
Revert "USB: Adding support for HTC Smartphones to ipaq"

ipaq already had this device id defined:

{ USB_DEVICE(0x0BB4, 0x00CF) }, /* HTC USB Modem */

Revert the commit which adds the duplicate entry.

This reverts commit 04cab1329336d4577d6638360c905e360934b425.

Originally-by: Ben Collins <ben.collins@canonical.com>
Signed-off-by: Leann Ogasawara <leann.ogasawara@canonical.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agoUSB: option: Remove duplicate AMOI_VENDOR_ID
Leann Ogasawara [Thu, 10 Jun 2010 21:51:51 +0000 (14:51 -0700)]
USB: option: Remove duplicate AMOI_VENDOR_ID

AMOI_VENDOR_ID is defined twice.  Remove the duplicate entry and move
the AMOI_PRODUCT_9508 definition to be grouped with the other AMOI
product definitions.

Originally-by: Ben Collins <ben.collins@ubuntu.com>
Signed-off-by: Leann Ogasawara <leann.ogasawara@canonical.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agoUSB: EHCI: EHCI 1.1 addendum: Enable Per-port change detect bits
Alek Du [Fri, 4 Jun 2010 07:47:56 +0000 (15:47 +0800)]
USB: EHCI: EHCI 1.1 addendum: Enable Per-port change detect bits

This patch will enable Per-port event feature defined in EHCI 1.1
addendum. This feature addresses an issue where HCD is currently
required to read and parse PORTSC for all enabled root hub ports. With
this patch, the overhead will be reduced.

Signed-off-by: Alek Du <alek.du@intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agoUSB: EHCI: EHCI 1.1 addendum: Basic LPM feature support
Alek Du [Fri, 4 Jun 2010 07:47:55 +0000 (15:47 +0800)]
USB: EHCI: EHCI 1.1 addendum: Basic LPM feature support

With this patch, the LPM capable EHCI host controller can put device
into L1 sleep state which is a mode that can enter/exit quickly, and
reduce power consumption.

Signed-off-by: Jacob Pan <jacob.jun.pan@intel.com>
Signed-off-by: Alek Du <alek.du@intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agoUSB: EHCI: EHCI 1.1 addendum: preparation
Alek Du [Fri, 4 Jun 2010 07:47:54 +0000 (15:47 +0800)]
USB: EHCI: EHCI 1.1 addendum: preparation

EHCI 1.1 addendum introduced several energy efficiency extensions for
EHCI USB host controllers:
1. LPM (link power management)
2. Per-port change
3. Shorter periodic frame list
4. Hardware prefetching

This patch is intended to define the HW bits and debug interface for
EHCI 1.1 addendum. The LPM and Per-port change patches will be sent out
after this patch.

Signed-off-by: Jacob Pan <jacob.jun.pan@intel.com>
Signed-off-by: Alek Du <alek.du@intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agousb: throw away custom hex digit methods
Andy Shevchenko [Tue, 15 Jun 2010 14:04:44 +0000 (17:04 +0300)]
usb: throw away custom hex digit methods

Recent kernel has common method to convert hex digit to its value.

Signed-off-by: Andy Shevchenko <ext-andriy.shevchenko@nokia.com>
Cc: David Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agoUSB: isd200.c: Remove unnecessary kmalloc cast
Joe Perches [Tue, 1 Jun 2010 03:23:19 +0000 (20:23 -0700)]
USB: isd200.c: Remove unnecessary kmalloc cast

Signed-off-by: Joe Perches <joe@perches.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agoUSB: host: Eliminate NULL dereference
Julia Lawall [Thu, 27 May 2010 16:10:08 +0000 (18:10 +0200)]
USB: host: Eliminate NULL dereference

The test above allows std to be NULL, so check that std is not NULL before
doing the dereference.

A simplified version of the semantic match that finds this problem is as
follows: (http://coccinelle.lip6.fr/)

// <smpl>
@r exists@
expression E,E1;
identifier f;
statement S1,S2,S3;
@@

if ((E == NULL && ...) || ...)
{
  ... when != if (...) S1 else S2
      when != E = E1
* E->f
  ... when any
}
else S3
// </smpl>

Signed-off-by: Julia Lawall <julia@diku.dk>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agoUSB: otg.h: Fix the mixup in parameters order.
Igor Grinberg [Thu, 27 May 2010 06:32:13 +0000 (09:32 +0300)]
USB: otg.h: Fix the mixup in parameters order.

otg_io_write() function does not follow the declaration of
struct otg_io_access_ops.

Signed-off-by: Igor Grinberg <grinberg@compulab.co.il>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agoUSB: core endpoint: Fix Coding Styles
csanchez@neurowork.net [Tue, 25 May 2010 15:53:17 +0000 (10:53 -0500)]
USB: core endpoint: Fix Coding Styles

Fixed coding styles in the core usb endpoint.

Signed-off-by: Carlos Sánchez Acosta <csanchez@neurowork.net>
Signed-off-by: Alejandro Sánchez Acosta <asanchez@neurowork.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agoUSB: core driver: Fix Coding Styles
csanchez@neurowork.net [Tue, 25 May 2010 15:38:22 +0000 (10:38 -0500)]
USB: core driver: Fix Coding Styles

Fixed coding styles in the core usb driver.

Signed-off-by: Carlos Sánchez Acosta <csanchez@neurowork.net>
Signed-off-by: Alejandro Sánchez Acosta <asanchez@neurowork.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agoUSB: Documentation/usb/linux.inf: Checkpatch cleanup
Andrea Gelmini [Sun, 23 May 2010 19:56:43 +0000 (21:56 +0200)]
USB: Documentation/usb/linux.inf: Checkpatch cleanup

Documentation/usb/linux.inf:66: ERROR: trailing whitespace
Documentation/usb/linux.inf:67: ERROR: trailing whitespace
Documentation/usb/linux.inf:98: ERROR: trailing whitespace
Documentation/usb/linux.inf:99: ERROR: trailing whitespace
Documentation/usb/linux.inf:115: ERROR: trailing whitespace
Documentation/usb/linux.inf:116: ERROR: trailing whitespace
Documentation/usb/linux.inf:120: ERROR: trailing whitespace
Documentation/usb/linux.inf:121: ERROR: trailing whitespace
Documentation/usb/linux.inf:122: ERROR: trailing whitespace
Documentation/usb/linux.inf:123: ERROR: trailing whitespace
Documentation/usb/linux.inf:144: ERROR: trailing whitespace
Documentation/usb/linux.inf:145: ERROR: trailing whitespace
Documentation/usb/linux.inf:146: ERROR: trailing whitespace
Documentation/usb/linux.inf:147: ERROR: trailing whitespace
Documentation/usb/linux.inf:158: ERROR: trailing whitespace

Signed-off-by: Andrea Gelmini <andrea.gelmini@gelma.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agoUSB: Documentation/usb/hotplug.txt: Checkpatch cleanup
Andrea Gelmini [Sun, 23 May 2010 19:56:42 +0000 (21:56 +0200)]
USB: Documentation/usb/hotplug.txt: Checkpatch cleanup

Documentation/usb/hotplug.txt:13: ERROR: trailing whitespace
Documentation/usb/hotplug.txt:87: ERROR: trailing whitespace

Signed-off-by: Andrea Gelmini <andrea.gelmini@gelma.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agoUSB: Documentation/usb/ehci.txt: Checkpatch cleanup
Andrea Gelmini [Sun, 23 May 2010 19:56:41 +0000 (21:56 +0200)]
USB: Documentation/usb/ehci.txt: Checkpatch cleanup

Documentation/usb/ehci.txt:12: ERROR: trailing whitespace

Signed-off-by: Andrea Gelmini <andrea.gelmini@gelma.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agoUSB: gadget: Use memdup_user
Julia Lawall [Sat, 22 May 2010 08:26:22 +0000 (10:26 +0200)]
USB: gadget: Use memdup_user

Use memdup_user when user data is immediately copied into the
allocated region.

The semantic patch that makes this change is as follows:
(http://coccinelle.lip6.fr/)

// <smpl>
@@
expression from,to,size,flag;
position p;
identifier l1,l2;
@@

-  to = \(kmalloc@p\|kzalloc@p\)(size,flag);
+  to = memdup_user(from,size);
   if (
-      to==NULL
+      IS_ERR(to)
                 || ...) {
   <+... when != goto l1;
-  -ENOMEM
+  PTR_ERR(to)
   ...+>
   }
-  if (copy_from_user(to, from, size) != 0) {
-    <+... when != goto l2;
-    -EFAULT
-    ...+>
-  }
// </smpl>

Signed-off-by: Julia Lawall <julia@diku.dk>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agoUSB: EHCI: make CONFIG_USB_EHCI_TT_NEWSCHED default to Y
Alan Stern [Fri, 21 May 2010 21:08:21 +0000 (17:08 -0400)]
USB: EHCI: make CONFIG_USB_EHCI_TT_NEWSCHED default to Y

This patch (as1382) changes the USB_EHCI_TT_NEWSCHED Kconfig option to
be non-experimental and to default to Y.  This option has existed for
a long time, and I have not heard any complaints concerning it.  By
contrast, several people have reported that their devices could be
made to work only by enabling the option.

The point of changing the default is to cause the option to be enabled
by distros that simply use the default settings for esoteric things
like this.

This change was motivated by Bugzilla #15649.

Signed-off-by: Alan Stern <stern@rowland.harvard.edu>
CC: David Brownell <david-b@pacbell.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agoUSB: iuu_phoenix: Add support for different baud rates.
James Courtier-Dutton [Fri, 21 May 2010 10:53:25 +0000 (11:53 +0100)]
USB: iuu_phoenix: Add support for different baud rates.

Signed-off-by: James Courtier-Dutton <James@superbug.co.uk>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agoUSB: xhci: Remove obsolete debugging printk.
Sarah Sharp [Mon, 7 Jun 2010 16:11:33 +0000 (09:11 -0700)]
USB: xhci: Remove obsolete debugging printk.

When code to manipulate the command register was refactored from
xhci_run() to xhci_start(), a debugging statement was left behind that no
longer applies.  Remove that statement.

Signed-off-by: Sarah Sharp <sarah.a.sharp@linux.intel.com>
Reported-by: Sergei Shtylyov <sshtylyov@mvista.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agoUSB: speedtouch: fixed more brace and spacing coding style issues
Nicolas Kaiser [Wed, 9 Jun 2010 18:43:03 +0000 (20:43 +0200)]
USB: speedtouch: fixed more brace and spacing coding style issues

Fixed spacing coding style issues.

Signed-off-by: Nicolas Kaiser <nikai@nikai.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agoUSB: speedtouch: fixed brace and spacing coding style issues
Nicolas Kaiser [Wed, 9 Jun 2010 18:22:03 +0000 (20:22 +0200)]
USB: speedtouch: fixed brace and spacing coding style issues

Fixed brace coding style issues.

Signed-off-by: Nicolas Kaiser <nikai@nikai.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agoUSB: UHCI: acquire spinlock before calling start_rh()
Alan Stern [Wed, 9 Jun 2010 21:34:39 +0000 (17:34 -0400)]
USB: UHCI: acquire spinlock before calling start_rh()

This patch (as1392) fixes a bug in uhci-hcd: The start_rh() routine is
supposed to be called with the private spinlock held.  If an IRQ comes
in at just the wrong time, the driver will think the controller has
died when in fact it simply hasn't start yet.

The patch also addresses some issues that may prevent an URB from
being unlinked after the controller has stopped.  This is an abnormal
occurrence (ordinarily the controller stops only when the entire bus
is suspended and hence there are no active URBs), so the pathways
haven't gotten much testing.  These two changes may be a little more
than is strictly necessary, but clearly they won't hurt.

Signed-off-by: Alan Stern <stern@rowland.harvard.edu>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agoUSB: add check to detect host controller hardware removal
Alan Stern [Wed, 9 Jun 2010 21:34:27 +0000 (17:34 -0400)]
USB: add check to detect host controller hardware removal

This patch (as1391) fixes a problem that can occur when USB host
controller hardware is hot-unplugged.  If no interrupts are generated
by the unplug then the HCD may not realize that the controller is
gone, and the subsequent unbind may hang waiting for interrupts that
never arrive.

The solution (for PCI-based controllers) is to call the HCD's
interrupt handler at the start of usb_hcd_pci_remove().  If the
hardware is gone, the handler will realize this when it tries to read
the controller's status register.

Signed-off-by: Alan Stern <stern@rowland.harvard.edu>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agoUSB: don't stop root-hub status polls too soon
Alan Stern [Wed, 9 Jun 2010 21:34:17 +0000 (17:34 -0400)]
USB: don't stop root-hub status polls too soon

This patch (as1390) fixes a problem that crops up when a UHCI host
controller is unbound from uhci-hcd while there are still some active
URBs.  The URBs have to be unlinked when the root hub is unregistered,
and uhci-hcd relies upon root-hub status polls as part of its
unlinking procedure.  But usb_hcd_poll_rh_status() won't make those
status calls if hcd->rh_registered is clear, and the flag is cleared
_before_ the unregistration takes place.

Since hcd->rh_registered is used for other things and needs to be
cleared early, the solution is to add a new flag (rh_pollable) and use
it instead.  It gets cleared _after_ the root hub is unregistered.

Now that the status polls don't end too soon, we have to make sure
they also don't occur too late -- after the root hub's usb_device
structure or the HCD's private structures are deallocated.  Therefore
the patch adds usb_get_device() and usb_put_device() calls to protect
the root hub structure, and it adds an extra del_timer_sync() to
prevent the root-hub timer from causing an unexpected status poll.

This additional complexity would not be needed if the HCD framework
had provided separate stop() and release() callbacks instead of just
stop().  This lack could be fixed at some future time (although it
would require changes to every host controller driver); when that
happens this patch won't be needed any more.

Signed-off-by: Alan Stern <stern@rowland.harvard.edu>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agoUSB: fix failure path in usb_add_hcd()
Alan Stern [Wed, 9 Jun 2010 21:34:05 +0000 (17:34 -0400)]
USB: fix failure path in usb_add_hcd()

This patch (as1389) fixes some errors in the failure pathway of
usb_add_hcd().  The actions it takes ought to be exactly the same as
those taken by usb_remove_hcd(), but they aren't.

In one case (removal of the usb_bus_attr_group), the two routines are
brought into agreement by changing usb_remove_hcd().  All the other
discrepancies are fixed by changing usb_add_hcd().

Signed-off-by: Alan Stern <stern@rowland.harvard.edu>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agoMerge branch 'writable_limits' of git://decibel.fi.muni.cz/~xslaby/linux
Linus Torvalds [Tue, 10 Aug 2010 19:07:51 +0000 (12:07 -0700)]
Merge branch 'writable_limits' of git://decibel.fi.muni.cz/~xslaby/linux

* 'writable_limits' of git://decibel.fi.muni.cz/~xslaby/linux:
  unistd: add __NR_prlimit64 syscall numbers
  rlimits: implement prlimit64 syscall
  rlimits: switch more rlimit syscalls to do_prlimit
  rlimits: redo do_setrlimit to more generic do_prlimit
  rlimits: add rlimit64 structure
  rlimits: do security check under task_lock
  rlimits: allow setrlimit to non-current tasks
  rlimits: split sys_setrlimit
  rlimits: selinux, do rlimits changes under task_lock
  rlimits: make sure ->rlim_max never grows in sys_setrlimit
  rlimits: add task_struct to update_rlimit_cpu
  rlimits: security, add task_struct to setrlimit

Fix up various system call number conflicts.  We not only added fanotify
system calls in the meantime, but asm-generic/unistd.h added a wait4
along with a range of reserved per-architecture system calls.

14 years agoMerge git://git.infradead.org/mtd-2.6
Linus Torvalds [Tue, 10 Aug 2010 18:49:21 +0000 (11:49 -0700)]
Merge git://git.infradead.org/mtd-2.6

* git://git.infradead.org/mtd-2.6: (79 commits)
  mtd: Remove obsolete <mtd/compatmac.h> include
  mtd: Update copyright notices
  jffs2: Update copyright notices
  mtd-physmap: add support users can assign the probe type in board files
  mtd: remove redwood map driver
  mxc_nand: Add v3 (i.MX51) Support
  mxc_nand: support 8bit ecc
  mxc_nand: fix correct_data function
  mxc_nand: add V1_V2 namespace to registers
  mxc_nand: factor out a check_int function
  mxc_nand: make some internally used functions overwriteable
  mxc_nand: rework get_dev_status
  mxc_nand: remove 0xe00 offset from registers
  mtd: denali: Add multi connected NAND support
  mtd: denali: Remove set_ecc_config function
  mtd: denali: Remove unuseful code in get_xx_nand_para functions
  mtd: denali: Remove device_info_tag structure
  mtd: m25p80: add support for the Winbond W25Q32 SPI flash chip
  mtd: m25p80: add support for the Intel/Numonyx {16,32,64}0S33B SPI flash chips
  mtd: m25p80: add support for the EON EN25P{32, 64} SPI flash chips
  ...

Fix up trivial conflicts in drivers/mtd/maps/{Kconfig,redwood.c} due to
redwood driver removal.

14 years agoMerge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/bcopeland...
Linus Torvalds [Tue, 10 Aug 2010 18:47:36 +0000 (11:47 -0700)]
Merge branch 'for-linus' of git://git./linux/kernel/git/bcopeland/omfs

* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/bcopeland/omfs:
  omfs: fix uninitialized variable warning
  omfs: sanity check cluster size
  omfs: refuse to mount if bitmap pointer is obviously wrong
  omfs: check bounds on block numbers before passing to sb_bread
  omfs: fix memory leak

14 years agoMerge branch 'for-linus' of git://git.infradead.org/users/eparis/notify
Linus Torvalds [Tue, 10 Aug 2010 18:39:13 +0000 (11:39 -0700)]
Merge branch 'for-linus' of git://git.infradead.org/users/eparis/notify

* 'for-linus' of git://git.infradead.org/users/eparis/notify: (132 commits)
  fanotify: use both marks when possible
  fsnotify: pass both the vfsmount mark and inode mark
  fsnotify: walk the inode and vfsmount lists simultaneously
  fsnotify: rework ignored mark flushing
  fsnotify: remove global fsnotify groups lists
  fsnotify: remove group->mask
  fsnotify: remove the global masks
  fsnotify: cleanup should_send_event
  fanotify: use the mark in handler functions
  audit: use the mark in handler functions
  dnotify: use the mark in handler functions
  inotify: use the mark in handler functions
  fsnotify: send fsnotify_mark to groups in event handling functions
  fsnotify: Exchange list heads instead of moving elements
  fsnotify: srcu to protect read side of inode and vfsmount locks
  fsnotify: use an explicit flag to indicate fsnotify_destroy_mark has been called
  fsnotify: use _rcu functions for mark list traversal
  fsnotify: place marks on object in order of group memory address
  vfs/fsnotify: fsnotify_close can delay the final work in fput
  fsnotify: store struct file not struct path
  ...

Fix up trivial delete/modify conflict in fs/notify/inotify/inotify.c.

14 years agoMerge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/viro/vfs-2.6
Linus Torvalds [Tue, 10 Aug 2010 18:26:52 +0000 (11:26 -0700)]
Merge branch 'for-linus' of git://git./linux/kernel/git/viro/vfs-2.6

* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/viro/vfs-2.6: (96 commits)
  no need for list_for_each_entry_safe()/resetting with superblock list
  Fix sget() race with failing mount
  vfs: don't hold s_umount over close_bdev_exclusive() call
  sysv: do not mark superblock dirty on remount
  sysv: do not mark superblock dirty on mount
  btrfs: remove junk sb_dirt change
  BFS: clean up the superblock usage
  AFFS: wait for sb synchronization when needed
  AFFS: clean up dirty flag usage
  cifs: truncate fallout
  mbcache: fix shrinker function return value
  mbcache: Remove unused features
  add f_flags to struct statfs(64)
  pass a struct path to vfs_statfs
  update VFS documentation for method changes.
  All filesystems that need invalidate_inode_buffers() are doing that explicitly
  convert remaining ->clear_inode() to ->evict_inode()
  Make ->drop_inode() just return whether inode needs to be dropped
  fs/inode.c:clear_inode() is gone
  fs/inode.c:evict() doesn't care about delete vs. non-delete paths now
  ...

Fix up trivial conflicts in fs/nilfs2/super.c

14 years agoMerge git://git.kernel.org/pub/scm/linux/kernel/git/davem/net-2.6
Linus Torvalds [Tue, 10 Aug 2010 04:05:52 +0000 (21:05 -0700)]
Merge git://git./linux/kernel/git/davem/net-2.6

* git://git.kernel.org/pub/scm/linux/kernel/git/davem/net-2.6: (59 commits)
  igbvf.txt: Add igbvf Documentation
  igb.txt: Add igb documentation
  e100/e1000*/igb*/ixgb*: Add missing read memory barrier
  ixgbe: fix build error with FCOE_CONFIG without DCB_CONFIG
  netxen: protect tx timeout recovery by rtnl lock
  isdn: gigaset: use after free
  isdn: gigaset: add missing unlock
  solos-pci: Fix race condition in tasklet RX handling
  pkt_sched: Fix sch_sfq vs tcf_bind_filter oops
  net: disable preemption before call smp_processor_id()
  tcp: no md5sig option size check bug
  iwlwifi: fix locking assertions
  iwlwifi: fix TX tracer
  isdn: fix information leak
  net: Fix napi_gro_frags vs netpoll path
  usbnet: remove noisy and hardly useful printk
  rtl8180: avoid potential NULL deref in rtl8180_beacon_work
  ath9k: Remove myself from the MAINTAINERS list
  libertas: scan before assocation if no BSSID was given
  libertas: fix association with some APs by using extended rates
  ...

14 years agoMerge git://git.kernel.org/pub/scm/linux/kernel/git/davem/ide-2.6
Linus Torvalds [Tue, 10 Aug 2010 04:05:33 +0000 (21:05 -0700)]
Merge git://git./linux/kernel/git/davem/ide-2.6

* git://git.kernel.org/pub/scm/linux/kernel/git/davem/ide-2.6:
  tx493xide: use min_t() macro instead of min()
  drivers/ide: Use memdup_user
  via82cxxx: fix typo for VT6415 PCIE PATA IDE Host Controller support.
  ide-cd: Do not access completed requests in the irq handler

14 years agoMerge git://git.kernel.org/pub/scm/linux/kernel/git/davem/sparc-2.6
Linus Torvalds [Tue, 10 Aug 2010 04:05:17 +0000 (21:05 -0700)]
Merge git://git./linux/kernel/git/davem/sparc-2.6

* git://git.kernel.org/pub/scm/linux/kernel/git/davem/sparc-2.6:
  sparc64: Update defconfig.
  sparc: Kill user copy check code.
  sparc64: Fix perf_arch_get_caller_regs().
  sparc64: Add missing ID to parport probing code.

14 years agoMerge branch 'rs485fix' of git://www.jni.nu/cris
Linus Torvalds [Tue, 10 Aug 2010 04:04:23 +0000 (21:04 -0700)]
Merge branch 'rs485fix' of git://www.jni.nu/cris

* 'rs485fix' of git://www.jni.nu/cris:
  CRIS: ioctl for getting RS485 information

14 years agoMerge branch 'merge' of git://git.kernel.org/pub/scm/linux/kernel/git/benh/powerpc
Linus Torvalds [Tue, 10 Aug 2010 04:02:42 +0000 (21:02 -0700)]
Merge branch 'merge' of git://git./linux/kernel/git/benh/powerpc

* 'merge' of git://git.kernel.org/pub/scm/linux/kernel/git/benh/powerpc:
  powerpc: fix build with make 3.82
  Revert "Input: appletouch - fix integer overflow issue"
  memblock: Fix memblock_is_region_reserved() to return a boolean
  powerpc: Trim defconfigs
  powerpc: fix i8042 module build error
  sound/soc: mpc5200_psc_ac97: Use gpio pins for cold reset
  powerpc/5200: add mpc5200_psc_ac97_gpio_reset

14 years agoMerge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/vapier...
Linus Torvalds [Tue, 10 Aug 2010 04:01:11 +0000 (21:01 -0700)]
Merge branch 'for-linus' of git://git./linux/kernel/git/vapier/blackfin

* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/vapier/blackfin: (32 commits)
  Blackfin: gpio: add a debounce stub
  Blackfin: update defconfigs
  Blackfin: remove CONFIG_MEM_GENERIC_BOARD
  Blackfin: dpmc: punt unnecessary RTC_ISTAT clearing
  Blackfin: unify rotary encoder bitmasks
  Blackfin: unify SDH/RSI bitmasks
  Blackfin: BF54x: tweak DMAC MMR naming to match other ports
  Blackfin: TWI: clean up the MMR names
  Blackfin: add EVT_OVERRIDE/IPRIO core MMR helpers
  Blackfin: add support for dynamic ftrace
  Blackfin: add support for LZO compressed kernels
  Blackfin: portmux: fix peripheral map overflow when requesting pins
  Blackfin: document SPI CS limitations with CPHA=0
  Blackfin: remove useless and outdated documentation
  Blackfin: BF51x/BF52x: support GPIO Hysteresis/Schmitt Trigger options
  Blackfin: gpio/portmux: clean up whitespace corruption
  Blackfin: make sure mmiowb inserts a write barrier with SSYNC
  Blackfin: fix DMA/cache bug when resuming from suspend to RAM
  Blackfin: BF51x: fix handling of PH8 (the "internal" SPI0SEL4 pin)
  Blackfin: add a GPIO_DEFAULT_BOOT_SPI_CS
  ...

14 years agoMerge branch 'next' of git://git.kernel.org/pub/scm/linux/kernel/git/djbw/async_tx
Linus Torvalds [Tue, 10 Aug 2010 04:00:07 +0000 (21:00 -0700)]
Merge branch 'next' of git://git./linux/kernel/git/djbw/async_tx

* 'next' of git://git.kernel.org/pub/scm/linux/kernel/git/djbw/async_tx: (30 commits)
  DMAENGINE: at_hdmac: locking fixlet
  DMAENGINE: pch_dma: kill another usage of __raw_{read|write}l
  dma: dmatest: fix potential sign bug
  ioat2: catch and recover from broken vtd configurations v6
  DMAENGINE: add runtime slave control to COH 901 318 v3
  DMAENGINE: add runtime slave config to DMA40 v3
  DMAENGINE: generic slave channel control v3
  dmaengine: Driver for Topcliff PCH DMA controller
  intel_mid: Add Mrst & Mfld DMA Drivers
  drivers/dma: Eliminate a NULL pointer dereference
  dma/timb_dma: compile warning on 32 bit
  DMAENGINE: ste_dma40: support older silicon
  DMAENGINE: ste_dma40: support disabling physical channels
  DMAENGINE: ste_dma40: no disabled phy channels on ux500
  DMAENGINE: ste_dma40: fix suspend bug
  DMAENGINE: ste_dma40: add DB8500 memcpy channels
  DMAENGINE: ste_dma40: no flow control on memcpy
  DMAENGINE: ste_dma40: arch updates for LCLA and LCPA
  DMAENGINE: ste_dma40: allocate LCLA dynamically
  DMAENGINE: ste_dma40: no premature stop
  ...

Fix up trivial conflicts in arch/arm/mach-ux500/devices-db8500.c

14 years agodrivers/char/vt.c:vc_do_resize(): local var `end' should be unsigned long
qiaochong [Tue, 10 Aug 2010 00:21:27 +0000 (17:21 -0700)]
drivers/char/vt.c:vc_do_resize(): local var `end' should be unsigned long

According include/linux/console_struct.h,vc_scr_end is unsigned long.

struct vc_data {
    unsigned short  vc_num;         /* Console number */
    unsigned int    vc_cols;        /* [#] Console size */
    unsigned int    vc_rows;
    unsigned int    vc_size_row;        /* Bytes per row */
    unsigned int    vc_scan_lines;      /* # of scan lines */
    unsigned long   vc_origin;      /* [!] Start of real screen */
    unsigned long   vc_scr_end;     /* [!] End of real screen */
    unsigned long   vc_visible_origin;  /* [!] Top of visible window */
    unsigned int    vc_top, vc_bottom;  /* Scrolling region */
    const struct consw *vc_sw;
    unsigned short  *vc_screenbuf;
...
}

Signed-off-by: qiaochong <qiaochong@loongson.cn>
Cc: Greg KH <greg@kroah.com>
Cc: Alan Cox <alan@lxorguk.ukuu.org.uk>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
14 years agodrivers/char/vt.c: fix vc->vc_origin on take_over_console()
qiaochong [Tue, 10 Aug 2010 00:21:23 +0000 (17:21 -0700)]
drivers/char/vt.c: fix vc->vc_origin on take_over_console()

kernel will die on some platform when switch from vga mode to framebuffer
mode.  The reason of this bug is that bind_con_driver reset vc->vc_origin
to (unsigned long)vc->vc_screenbuf.

On vgacon vc->vc_origin is not releated to vc->vc_screenbuf,if set
vc->vc_origin to vc->vc_screenbuf,kernel will die on vc_do_resize.

static int vc_do_resize(struct tty_struct *tty, struct tty_struct *real_tty,
        struct vc_data *vc, unsigned int cols, unsigned int lines)
{
    unsigned long old_origin, new_origin, new_scr_end, rlth, rrem, err = 0;
    unsigned int old_cols, old_rows, old_row_size, old_screen_size;
    unsigned int new_cols, new_rows, new_row_size, new_screen_size;
    unsigned int end, user;
...
        end = (old_rows > new_rows) ? old_origin +
            (old_row_size * new_rows) :
            vc->vc_scr_end;

...
/*
here for a test from vgacon to framebuffer:
old_origin=0x810814a0,end=0xb00b8fa0,vc->vc_origin=0x810814a0
the code bellow will copy memory from 0x810814a0 to 0xb00b8fa0,
this will cover kernel code,kernel died here.
*/

    while (old_origin < end) {

        scr_memcpyw((unsigned short *) new_origin,
                (unsigned short *) old_origin, rlth);
        if (rrem)
            scr_memsetw((void *)(new_origin + rlth),
                    vc->vc_video_erase_char, rrem);
        old_origin += old_row_size;
        new_origin += new_row_size;
    }

...
}

[akpm@linux-foundation.org: coding-style fixes]
Signed-off-by: qiaochong <qiaochong@loongson.cn>
Cc: Greg KH <greg@kroah.com>
Cc: Alan Cox <alan@lxorguk.ukuu.org.uk>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
14 years agorwsem: smaller wrappers around rwsem_down_failed_common
Michel Lespinasse [Tue, 10 Aug 2010 00:21:20 +0000 (17:21 -0700)]
rwsem: smaller wrappers around rwsem_down_failed_common

More code can be pushed from rwsem_down_read_failed and
rwsem_down_write_failed into rwsem_down_failed_common.

Following change adding down_read_critical infrastructure support also
enjoys having flags available in a register rather than having to fish it
out in the struct rwsem_waiter...

Signed-off-by: Michel Lespinasse <walken@google.com>
Acked-by: David Howells <dhowells@redhat.com>
Cc: Mike Waychison <mikew@google.com>
Cc: Suleiman Souhlal <suleiman@google.com>
Cc: Ying Han <yinghan@google.com>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: "H. Peter Anvin" <hpa@zytor.com>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
14 years agorwsem: wake queued readers when writer blocks on active read lock
Michel Lespinasse [Tue, 10 Aug 2010 00:21:19 +0000 (17:21 -0700)]
rwsem: wake queued readers when writer blocks on active read lock

This change addresses the following situation:

- Thread A acquires the rwsem for read
- Thread B tries to acquire the rwsem for write, notices there is already
  an active owner for the rwsem.
- Thread C tries to acquire the rwsem for read, notices that thread B already
  tried to acquire it.
- Thread C grabs the spinlock and queues itself on the wait queue.
- Thread B grabs the spinlock and queues itself behind C. At this point A is
  the only remaining active owner on the rwsem.

In this situation thread B could notice that it was the last active writer
on the rwsem, and decide to wake C to let it proceed in parallel with A
since they both only want the rwsem for read.

Signed-off-by: Michel Lespinasse <walken@google.com>
Acked-by: David Howells <dhowells@redhat.com>
Cc: Mike Waychison <mikew@google.com>
Cc: Suleiman Souhlal <suleiman@google.com>
Cc: Ying Han <yinghan@google.com>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: "H. Peter Anvin" <hpa@zytor.com>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
14 years agorwsem: let RWSEM_WAITING_BIAS represent any number of waiting threads
Michel Lespinasse [Tue, 10 Aug 2010 00:21:18 +0000 (17:21 -0700)]
rwsem: let RWSEM_WAITING_BIAS represent any number of waiting threads

Previously each waiting thread added a bias of RWSEM_WAITING_BIAS.  With
this change, the bias is added only once to indicate that the wait list is
non-empty.

This has a few nice properties which will be used in following changes:
- when the spinlock is held and the waiter list is known to be non-empty,
  count < RWSEM_WAITING_BIAS  <=>  there is an active writer on that sem
- count == RWSEM_WAITING_BIAS  <=>  there are waiting threads and no
                                     active readers/writers on that sem

Signed-off-by: Michel Lespinasse <walken@google.com>
Acked-by: David Howells <dhowells@redhat.com>
Cc: Mike Waychison <mikew@google.com>
Cc: Suleiman Souhlal <suleiman@google.com>
Cc: Ying Han <yinghan@google.com>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: "H. Peter Anvin" <hpa@zytor.com>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
14 years agorwsem: lighter active count checks when waking up readers
Michel Lespinasse [Tue, 10 Aug 2010 00:21:17 +0000 (17:21 -0700)]
rwsem: lighter active count checks when waking up readers

In __rwsem_do_wake(), we can skip the active count check unless we come
there from up_xxxx().  Also when checking the active count, it is not
actually necessary to increment it; this allows us to get rid of the read
side undo code and simplify the calculation of the final rwsem count
adjustment once we've counted the reader threads to wake.

The basic observation is the following.  When there are waiter threads on
a rwsem and the spinlock is held, other threads can only increment the
active count by trying to grab the rwsem in down_xxxx().  However
down_xxxx() will notice there are waiter threads and take the down_failed
path, blocking to acquire the spinlock on the way there.  Therefore, a
thread observing an active count of zero with waiters queued and the
spinlock held, is protected against other threads acquiring the rwsem
until it wakes the last waiter or releases the spinlock.

Signed-off-by: Michel Lespinasse <walken@google.com>
Acked-by: David Howells <dhowells@redhat.com>
Cc: Mike Waychison <mikew@google.com>
Cc: Suleiman Souhlal <suleiman@google.com>
Cc: Ying Han <yinghan@google.com>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: "H. Peter Anvin" <hpa@zytor.com>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
14 years agorwsem: fully separate code paths to wake writers vs readers
Michel Lespinasse [Tue, 10 Aug 2010 00:21:15 +0000 (17:21 -0700)]
rwsem: fully separate code paths to wake writers vs readers

This is in preparation for later changes in the series.

In __rwsem_do_wake(), the first queued waiter is checked first in order to
determine whether it's a writer or a reader.  The code paths diverge at
this point.  The code that checks and increments the rwsem active count is
duplicated on both sides - the point is that later changes in the series
will be able to independently modify both sides.

Signed-off-by: Michel Lespinasse <walken@google.com>
Acked-by: David Howells <dhowells@redhat.com>
Cc: Mike Waychison <mikew@google.com>
Cc: Suleiman Souhlal <suleiman@google.com>
Cc: Ying Han <yinghan@google.com>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: "H. Peter Anvin" <hpa@zytor.com>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
14 years agohwmon: f71882fg: add support for the Fintek F71808E
Giel van Schijndel [Tue, 10 Aug 2010 00:21:13 +0000 (17:21 -0700)]
hwmon: f71882fg: add support for the Fintek F71808E

Allow device probing to recognise the Fintek F71808E.

Sysfs interface:
 * Fan/pwm control is the same as for F71889FG
 * Temperature and voltage sensor handling is largely the same as for
   the F71889FG
  - Has one temperature sensor less (doesn't have temp3)
  - Misses one voltage sensor (doesn't have V6, thus in6_input refers to
    what in7_input refers for F71889FG)

For the purpose of the sysfs interface fxxxx_in_temp_attr[] is split up
such that it can largely be reused.

Signed-off-by: Giel van Schijndel <me@mortis.eu>
Cc: Jean Delvare <khali@linux-fr.org>
Cc: Hans de Goede <hdegoede@redhat.com>
Cc: Jonathan Cameron <jic23@cam.ac.uk>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>