GitHub/LineageOS/android_kernel_motorola_exynos9610.git
18 years ago[ARM] 3597/1: ixp4xx/nslu2: Board support for new LED subsystem
Rod Whitby [Thu, 22 Jun 2006 21:21:02 +0000 (22:21 +0100)]
[ARM] 3597/1: ixp4xx/nslu2: Board support for new LED subsystem

Patch from Rod Whitby

This patch implements NEW_LEDS support for the Linksys NSLU2.  The
NSLU2 has four LED indicators, which are the only form of output for
an unmodified device - there is no keyboard or display on an NSLU2.
For an NSLU2 which has been modified to bring out the serial port
console, it is important to register that device first separately, to
enable debugging of other device support.

Signed-off-by: John Bowler <jbowler@acm.org>
Signed-off-by: Rod Whitby <rod@whitby.id.au>
Signed-off-by: Deepak Saxena <dsaxena@plexity.net>
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
18 years ago[ARM] 3595/1: ixp4xx/nas100d: Board support for new LED subsystem
Rod Whitby [Thu, 22 Jun 2006 21:21:01 +0000 (22:21 +0100)]
[ARM] 3595/1: ixp4xx/nas100d: Board support for new LED subsystem

Patch from Rod Whitby

This patch implements NEW_LEDS support for the IOMega NAS100d.  The
NAS100d has three LED indicators, which are the only form of output
for an unmodified device - there is no keyboard or display on an
NAS100d.  For an NAS100d which has been modified to bring out the
serial port console, it is important to register that device first
separately, to enable debugging of other device support.

Signed-off-by: John Bowler <jbowler@acm.org>
Signed-off-by: Rod Whitby <rod@whitby.id.au>
Signed-off-by: Deepak Saxena <dsaxena@plexity.net>
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
18 years ago[ARM] 3626/1: ARM EABI: fix syscall restarting
Nicolas Pitre [Thu, 22 Jun 2006 21:18:45 +0000 (22:18 +0100)]
[ARM] 3626/1: ARM EABI: fix syscall restarting

Patch from Nicolas Pitre

The RESTARTBLOCK case currently store some code on the stack to invoke
sys_restart_syscall.  However this is ABI dependent and there is a
mismatch with the way __NR_restart_syscall gets defined when the kernel
is compiled for EABI.

There is also a long standing bug in the thumb case since with OABI the
__NR_restart_syscall value includes __NR_SYSCALL_BASE which should not
be the case for Thumb syscalls.

Credits to Yauheni Kaliuta <yauheni.kaliuta@gmail.com> for finding the
EABI bug.

Signed-off-by: Nicolas Pitre <nico@cam.org>
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
18 years ago[ARM] 3628/1: S3C24XX: add get_rate call to struct clk
Ben Dooks [Thu, 22 Jun 2006 21:18:21 +0000 (22:18 +0100)]
[ARM] 3628/1: S3C24XX: add get_rate call to struct clk

Patch from Ben Dooks

Add a get_rate call to allow an given clock
to over-ride the clk_get_rate() call.

This provides support for clocks which rely on
division of their parent to correctly report
their frequency when the parent can also change.

Signed-off-by: Ben Dooks <ben-linux@fluff.org>
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
18 years ago[ARM] 3627/1: S3C24XX: split s3c2410 clocks from core clocks
Ben Dooks [Thu, 22 Jun 2006 21:18:20 +0000 (22:18 +0100)]
[ARM] 3627/1: S3C24XX: split s3c2410 clocks from core clocks

Patch from Ben Dooks

Split the s3c2410 specific clocks from the core
clock code, as part of the work to support more
of the Samsung line of SoCs.

The patch does not use the sysdev mechanism as
the clocks are needed for the timer init, which
is very early in the kernel init sequence.

Signed-off-by: Ben Dooks <ben-linux@fluff.org>
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
18 years ago[ARM] 3613/1: S3C2410: Add sysdev and sysclass
Ben Dooks [Thu, 22 Jun 2006 21:18:13 +0000 (22:18 +0100)]
[ARM] 3613/1: S3C2410: Add sysdev and sysclass

Patch from Ben Dooks

The S3C2440 and S3C2442 both have their own sysdev
and sysclass for differentiating them from the
currently default S3C2410.

Add a sysdev for the S3C2410 as part of the work
to make the code be non-dependant on the S3C2410.

Signed-off-by: Ben Dooks <ben-linux@fluff.org>
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
18 years ago[ARM] 3624/1: Report true modem control line states
Uli Luckas [Thu, 22 Jun 2006 15:49:54 +0000 (16:49 +0100)]
[ARM] 3624/1: Report true modem control line states

Patch from Uli Luckas

This patch removes the fake return from serial_pxa_get_mctrl.

Signed-off-by: Uli Luckas <u.luckas@road-gmbh.de>
I just can't remember why this return was there.
Being in the first column clearly indicates it was meant to be removed.

Signed-off-by: Nicolas Pitre <nico@cam.org>
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
18 years ago[ARM] 3620/2: ixp23xx: add uengine loader support
Lennert Buytenhek [Thu, 22 Jun 2006 09:30:56 +0000 (10:30 +0100)]
[ARM] 3620/2: ixp23xx: add uengine loader support

Patch from Lennert Buytenhek

This patch allows the ixp2000 uengine loader that is already in the
tree to also be used on the ixp23xx.

Signed-off-by: Lennert Buytenhek <buytenh@wantstofly.org>
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
18 years ago[ARM] 3618/1: add defconfig for logicpd pxa270 card engine
Lennert Buytenhek [Thu, 22 Jun 2006 09:30:55 +0000 (10:30 +0100)]
[ARM] 3618/1: add defconfig for logicpd pxa270 card engine

Patch from Lennert Buytenhek

As it's slightly nontrivial to make it possible to build a single
kernel image for both the mainstone and the logicpd pxa270 card engine,
add a separate defconfig for the logicpd pxa270 card engine for now.

Signed-off-by: Lennert Buytenhek <buytenh@wantstofly.org>
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
18 years ago[ARM] 3617/1: ep93xx: fix slightly incorrect timer tick rate
Lennert Buytenhek [Thu, 22 Jun 2006 09:30:54 +0000 (10:30 +0100)]
[ARM] 3617/1: ep93xx: fix slightly incorrect timer tick rate

Patch from Lennert Buytenhek

The tick rate of timers 1-3 isn't exactly 508 kHz as some parts of the
relevant documentation claim, but more like 508.469 kHz (14.7456 MHz
divided by 29.)

Signed-off-by: Lennert Buytenhek <buytenh@wantstofly.org>
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
18 years ago[ARM] 3616/1: fix timer handler wrap logic for a number of platforms
Lennert Buytenhek [Thu, 22 Jun 2006 09:30:53 +0000 (10:30 +0100)]
[ARM] 3616/1: fix timer handler wrap logic for a number of platforms

Patch from Lennert Buytenhek

A couple of platforms aren't using the right comparison type in their
timer interrupt handlers (as we're comparing two wrapping timestamps,
we need a bmi/bpl-type comparison, not an unsigned comparison) -- this
patch fixes them up.

Signed-off-by: Lennert Buytenhek <buytenh@wantstofly.org>
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
18 years ago[ARM] 3615/1: ixp23xx: use platform devices for physmap flash
Lennert Buytenhek [Thu, 22 Jun 2006 09:30:52 +0000 (10:30 +0100)]
[ARM] 3615/1: ixp23xx: use platform devices for physmap flash

Patch from Lennert Buytenhek

Now that the physmap platform device rewrite is in, make the ixp23xx
boards use platform devices for physmap flash.

Signed-off-by: Lennert Buytenhek <buytenh@wantstofly.org>
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
18 years ago[ARM] 3614/1: ep93xx: use platform devices for physmap flash
Lennert Buytenhek [Thu, 22 Jun 2006 09:30:51 +0000 (10:30 +0100)]
[ARM] 3614/1: ep93xx: use platform devices for physmap flash

Patch from Lennert Buytenhek

Now that the physmap platform device rewrite is in, make the ep93xx
boards use platform devices for physmap flash.

Signed-off-by: Lennert Buytenhek <buytenh@wantstofly.org>
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
18 years ago[ARM] 3621/1: fix compilation breakage for pnx4008
Vitaly Wool [Thu, 22 Jun 2006 09:27:15 +0000 (10:27 +0100)]
[ARM] 3621/1: fix compilation breakage for pnx4008

Patch from Vitaly Wool

pnx4008_defconfig fails to build:

include/asm/hardware/debug-8250.S: Assembler messages:
include/asm/hardware/debug-8250.S:12: Error: Macro with this name was already defined.

This is due to senduart macro erroneously defined in include/asm-arm/arch-pnx4008/debug-macro.S. This patch removes it from that file.

Signed-off-by: Vitaly Wool <vwool@ru.mvista.com>
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
18 years ago[ARM] 3623/1: pnx4008: move GPIO-related defines to gpio.h
Vitaly Wool [Thu, 22 Jun 2006 09:26:21 +0000 (10:26 +0100)]
[ARM] 3623/1: pnx4008: move GPIO-related defines to gpio.h

Patch from Vitaly Wool

This patch moves GPIO-related defines and static inline funcs from include/asm-arm/arch-pnx4008/pm.h to include/asm-arm/arch-pnx4008/gpio.h.
Also, some more GPIO-related defines are added to include/asm-arm/arch-pnx4008/gpio.h as they are needed for the USB host driver (coming soon...)

Signed-off-by: Vitaly Wool <vwool@ru.mvista.com>
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
18 years ago[ARM] 3622/1: pnx4008: remove clk_use/clk_unuse
Vitaly Wool [Thu, 22 Jun 2006 09:26:20 +0000 (10:26 +0100)]
[ARM] 3622/1: pnx4008: remove clk_use/clk_unuse

Patch from Vitaly Wool

clk_use/clk_unuse functions are no longer needed, so removing those from arch/arm/mach-pnx4008/clock.c.
Also, the order of functions is rearranged a bit, to avoid forward declarations.

Signed-off-by: Vitaly Wool <vwool@ru.mvista.com>
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
18 years ago[ARM] Enable VFP to be built when non-VFP capable CPUs are selected
Russell King [Wed, 21 Jun 2006 12:51:41 +0000 (13:51 +0100)]
[ARM] Enable VFP to be built when non-VFP capable CPUs are selected

Since we pass flags to the compiler to control code generation based
on the least capable selected CPU, if we want to include VFP support,
we must tweak the assembler flags to allow the VFP instructions.
Moreover, we must not use the mrrc/mcrr versions since these will not
be recognised by the assembler.

We do not convert all instructions to the VFP-equivalent (yet) since
binutils appears to barf on "fmrx rn, fpinst" and doesn't provide any
other way (other than using the mrc equivalent) to encode this
instruction - which is rather a problem when you have a VFP
implementation which requires these instructions.

Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
18 years ago[ARM] Add thread_notify infrastructure
Russell King [Wed, 21 Jun 2006 12:31:52 +0000 (13:31 +0100)]
[ARM] Add thread_notify infrastructure

Some machine classes need to allow VFP support to be built into the
kernel, but still allow the kernel to run even though VFP isn't
present.  Unfortunately, the kernel hard-codes VFP instructions
into the thread switch, which prevents this being run-time selectable.

Solve this by introducing a notifier which things such as VFP can
hook into to be informed of events which affect the VFP subsystem
(eg, creation and destruction of threads, switches between threads.)

Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
18 years agoMerge git://oss.sgi.com:8090/xfs-2.6
Linus Torvalds [Thu, 22 Jun 2006 01:10:19 +0000 (18:10 -0700)]
Merge git://oss.sgi.com:8090/xfs-2.6

* git://oss.sgi.com:8090/xfs-2.6: (43 commits)
  [XFS] Remove files from the build that are now unused.
  [XFS] Fix a Makefile issue related to exports.o handling.
  [XFS] Remove version 1 directory code. Never functioned on Linux, just
  [XFS] Map EFSCORRUPTED to an actual error code, not just a made up one
  [XFS] Kill direct access to ->count in valusema(); all we ever use it for
  [XFS] Remove unneeded conditional code on NFS export interface related
  [XFS] Remove an incorrect use of unlikely() on a relatively likely code
  [XFS] Push some common code out of write path into core XFS code for
  [XFS] Remove unnecessary local from open_exec dmapi path.
  [XFS] Minor XFS documentation updates.
  [XFS] Fix broken const use inside local suffix_strtoul routine.
  [XFS] Fix nused counter.  It's currently getting set to -1 rather than
  [XFS] Fix mismerge of the fs_writable cleanup patch causing a freeze/thaw
  [XFS] Fix up debug code so that bulkstat wont generate thousands of
  [XFS] Remove unused parameter from di2xflags routine.
  [XFS] Cleanup a missed porting conversion, and freezing.
  [XFS] Resolve a namespace collision on remaining vtypes for FreeBSD
  [XFS] Resolve a namespace collision on vnode/vnodeops for FreeBSD porters.
  [XFS] Resolve a namespace collision on vfs/vfsops for FreeBSD porters.
  [XFS] statvfs component of directory/project quota support, code
  ...

18 years ago[PATCH] USB: only make /sys/class/usb show up when there is something in it
Greg Kroah-Hartman [Tue, 20 Jun 2006 22:14:07 +0000 (15:14 -0700)]
[PATCH] USB: only make /sys/class/usb show up when there is something in it

Now /sys/class/usb is dynamically created when we have something to put
in it, and removed when all devices go away.

Just trying to cut down on the clutter in sysfs...

Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
18 years ago[PATCH] USB: convert usb class devices to real devices
Greg Kroah-Hartman [Tue, 20 Jun 2006 20:09:50 +0000 (13:09 -0700)]
[PATCH] USB: convert usb class devices to real devices

Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
18 years ago[PATCH] USB: move usb_device_class class devices to be real devices
Greg Kroah-Hartman [Tue, 20 Jun 2006 06:59:31 +0000 (23:59 -0700)]
[PATCH] USB: move usb_device_class class devices to be real devices

This moves the usb class devices that control the usbfs nodes to show up
in the proper place in the larger device tree.

No userspace changes is needed, this is compatible due to the symlinks
generated by the driver core.

Cc: Kay Sievers <kay.sievers@vrfy.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
18 years ago[PATCH] USB: make endpoints real struct devices
Greg Kroah-Hartman [Wed, 14 Jun 2006 19:14:34 +0000 (12:14 -0700)]
[PATCH] USB: make endpoints real struct devices

This will allow for us to give endpoints a major/minor to create a
"usbfs2-like" way to access endpoints directly from userspace in an
easier manner than the current usbfs provides us.

Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
18 years ago[PATCH] USB: make usb_create_ep_files take a struct device
Greg Kroah-Hartman [Wed, 14 Jun 2006 19:14:34 +0000 (12:14 -0700)]
[PATCH] USB: make usb_create_ep_files take a struct device

Instead of a kobject, will make things easier in the future (don't know
what I was thinking when I did this originally...)

Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
18 years ago[PATCH] USB: move the endpoint specific sysfs code to it's own file
Greg Kroah-Hartman [Wed, 14 Jun 2006 19:14:34 +0000 (12:14 -0700)]
[PATCH] USB: move the endpoint specific sysfs code to it's own file

This makes it easier to modify in the future without touching anything else.

Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
18 years ago[PATCH] USB: move <linux/usb_input.h> to <linux/usb/input.h>
David Brownell [Tue, 13 Jun 2006 17:04:34 +0000 (10:04 -0700)]
[PATCH] USB: move <linux/usb_input.h> to <linux/usb/input.h>

Move <linux/usb_input.h> to <linux/usb/input.h> and remove some
redundant includes.

Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
18 years ago[PATCH] USB: move hardware-specific <linux/usb_*.h> to <linux/usb/*.h>
David Brownell [Tue, 13 Jun 2006 16:59:32 +0000 (09:59 -0700)]
[PATCH] USB: move hardware-specific <linux/usb_*.h> to <linux/usb/*.h>

This moves header files for controller-specific platform data
from <linux/usb_XXX.h> to <linux/usb/XXX.h> to start reducing
some clutter.

Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
18 years ago[PATCH] USB: move <linux/usb_cdc.h> to <linux/usb/cdc.h>
David Brownell [Tue, 13 Jun 2006 16:57:47 +0000 (09:57 -0700)]
[PATCH] USB: move <linux/usb_cdc.h> to <linux/usb/cdc.h>

This moves <linux/usb_cdc.h> to <linux/usb/cdc.h> to reduce some of the
clutter of usb header files.

Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
18 years ago[PATCH] USB: whitespace removal from usb/gadget/ether
David Brownell [Tue, 13 Jun 2006 16:54:40 +0000 (09:54 -0700)]
[PATCH] USB: whitespace removal from usb/gadget/ether

This removes extraneous whitespace from the Ethernet/RNDIS gadget driver.
It's all space-at-EOL, spaces-before-tabs, or tabs-then-spaces.

Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
18 years ago[PATCH] USB: unusual_devs entry for Nokia N80
Alan Stern [Tue, 13 Jun 2006 13:59:32 +0000 (09:59 -0400)]
[PATCH] USB: unusual_devs entry for Nokia N80

Here is a patch (as720) adding an unusual_devs entry for the Nokia N80
mobile phone.

Signed-off-by: Alan Stern <stern@rowland.harvard.edu>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
18 years ago[PATCH] usbserial: Fixes wrong return values.
Luiz Fernando N. Capitulino [Tue, 13 Jun 2006 01:46:20 +0000 (22:46 -0300)]
[PATCH] usbserial: Fixes wrong return values.

Some usbserial functions returns -EINVAL if the port doesn't exist or if
it's not opened. However, the right error code for such situations is
-ENODEV.

Signed-off-by: Luiz Fernando N. Capitulino <lcapitulino@mandriva.com.br>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
18 years ago[PATCH] usb gadget: update pxa2xx_udc.c driver to fully support IXP4xx platform
Milan Svoboda [Mon, 29 May 2006 10:34:00 +0000 (03:34 -0700)]
[PATCH] usb gadget: update pxa2xx_udc.c driver to fully support IXP4xx platform

This patch adds IXP465 into the list of known devices and
adds IXP425 and IXP465 to the list of devices that have cfr. This
is not described in the hardware documentation, but without
it driver won't work.

Workaround (#if 1) that seemed to get rid of lost
status irqs is disabled for IXP4XX as it caused freezes
during testing of control messages. No lost irqs are
visible on IXP4XX.

Driver survived tests running over night without any
visible problems.

Signed-off-by: Milan Svoboda <msvoboda@ra.rockwell.com>
Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
18 years ago[PATCH] usb gadget: fix compile errors
Milan Svoboda [Mon, 29 May 2006 10:34:00 +0000 (03:34 -0700)]
[PATCH] usb gadget: fix compile errors

This patch fixes compile errors when pxa2xx_udc is to be compiled
for ixp4xx platform.

Signed-off-by: Milan Svoboda <msvoboda@ra.rockwell.com>
Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
18 years ago[PATCH] usb gadget: allow drivers support speeds higher than full speed
Milan Svoboda [Mon, 29 May 2006 10:34:00 +0000 (03:34 -0700)]
[PATCH] usb gadget: allow drivers support speeds higher than full speed

This patch removes limitation which prevents use of drivers that support
speeds different that full speed.

Signed-off-by: Milan Svoboda <msvoboda@ra.rockwell.com>
Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
18 years ago[PATCH] USB: gadget-serial: do not save/restore IRQ flags in gs_close()
Franck Bui-Huu [Wed, 14 Jun 2006 08:47:18 +0000 (10:47 +0200)]
[PATCH] USB: gadget-serial: do not save/restore IRQ flags in gs_close()

As pointed out by David Brownell, we know that IRQs are never
blocked when calling gs_close function. So the save/restore
IRQ flags are pointless.

Signed-off-by: Franck Bui-Huu <vagabon.xyz@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
18 years ago[PATCH] USB: gadget-serial: fix a deadlock when closing the serial device
Franck Bui-Huu [Wed, 14 Jun 2006 08:29:21 +0000 (10:29 +0200)]
[PATCH] USB: gadget-serial: fix a deadlock when closing the serial device

When closing the device, the driver acquires/release twice the
port lock before/after waiting for the data to be completely
sent. Therefore it will dead lock.

This patch fixes it and also uses the generic scheduler services
for waiting for an event.

Signed-off-by: Franck Bui-Huu <vagabon.xyz@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
18 years ago[PATCH] USB Serial: clean tty fields on failed device open
Frank Gevaerts [Wed, 14 Jun 2006 13:52:05 +0000 (15:52 +0200)]
[PATCH] USB Serial: clean tty fields on failed device open

If either the driver's open() method or try_module_get() fails, we need to
set 'tty->driver_data' and 'port->tty' to NULL in serial_open(), otherwise
we'll get an OOPS in usb_device_disconnect() when the device is disconnected.

Signed-off-by: Frank Gevaerts <frank.gevaerts@fks.be>
Acked-by: Luiz Fernando N. Capitulino <lcapitulino@mandriva.com.br>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
18 years ago[PATCH] USB: trivial DEBUG message correction in gadget ether driver
Matt Reimer [Wed, 7 Jun 2006 18:46:13 +0000 (11:46 -0700)]
[PATCH] USB: trivial DEBUG message correction in gadget ether driver

Acked-by: David Brownell <david-b@pacbell.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
18 years ago[PATCH] USB: new driver for Cypress CY7C63xxx mirco controllers
Oliver Bock [Fri, 9 Jun 2006 22:42:19 +0000 (00:42 +0200)]
[PATCH] USB: new driver for Cypress CY7C63xxx mirco controllers

This is a new driver for the Cypress CY7C63xxx mirco controller series.
It currently supports the pre-programmed CYC63001A-PC by AK Modul-Bus
GmbH.  It's based on a kernel 2.4 driver (cyport) by Marcus Maul which I
ported to kernel 2.6 using sysfs. I intend to support more controllers
of this family (and more features) as soon as I get hold of the required
IDs etc. Please see the source code's header for more information.

Signed-off-by: Oliver Bock <o.bock@fh-wolfenbuettel.de>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
18 years ago[PATCH] USB: update usbmon.txt
Pete Zaitcev [Tue, 13 Jun 2006 03:09:39 +0000 (20:09 -0700)]
[PATCH] USB: update usbmon.txt

Fix up the documentation. Apparently, I left unedited copy-paste results
in examples. Also, Alan helped me to improve the most confusing parts.

Signed-off-by: Pete Zaitcev <zaitcev@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
18 years ago[PATCH] USB: implement error event in usbmon
Pete Zaitcev [Sat, 10 Jun 2006 05:03:32 +0000 (22:03 -0700)]
[PATCH] USB: implement error event in usbmon

Implement the "error" event in usbmon.

Signed-off-by: Pete Zaitcev <zaitcev@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
18 years ago[PATCH] USB: update usbmon, fix glued lines
Pete Zaitcev [Sat, 10 Jun 2006 03:10:10 +0000 (20:10 -0700)]
[PATCH] USB: update usbmon, fix glued lines

This update contains one bug fix: some lines can come out truncated,
because of the safety cutoff. This happened because I forgot to update
the size when status packets began to be printed.

The rest is:
 - Comments updates
 - Allow snooping with pkmap on x86_64, which is cache-coherent
 - Enlarge event buffers (certainly we can have a couple of pages)
 - Add event counter

First touch upon usbmon for 2.6.18.

Signed-off-by: Pete Zaitcev <zaitcev@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
18 years ago[PATCH] USB: io_edgeport touch-up
Pete Zaitcev [Wed, 7 Jun 2006 01:18:33 +0000 (18:18 -0700)]
[PATCH] USB: io_edgeport touch-up

Al Borchers suggested to do this in his review of my previous patch.
I guess that I skipped this initially because of my visceral dislike
of sizeof(data). But in this case it seems well localized.

Signed-off-by: Pete Zaitcev <zaitcev@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
18 years ago[PATCH] airprime.c: add Kyocera Wireless KPC650/Passport support
Timothy Sipples [Fri, 16 Jun 2006 11:42:59 +0000 (20:42 +0900)]
[PATCH] airprime.c: add Kyocera Wireless KPC650/Passport support

Adds vendor & product IDs to drivers/usb/serial/airprime.c for Kyocera
Wireless KPC650/Passport EV-DO/1xRTT PC Cards.

Signed-off-by: Timothy Sipples <Timothy.Sipples@us.ibm.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
18 years ago[PATCH] UHCI: Improve FSBR-off timing
Alan Stern [Mon, 5 Jun 2006 16:28:57 +0000 (12:28 -0400)]
[PATCH] UHCI: Improve FSBR-off timing

This patch (as707) improves the FSBR operation in uhci-hcd by turning it
off more quickly when it isn't needed.  FSBR puts a noticeable load on a
computer's PCI bus, so it should be disabled as soon as possible when it
isn't in use.  The patch leaves it running for only 10 ms after the last
URB stops using it, on the theory that this should be long enough for a
driver to submit another URB if it wants keep FSBR going.

Signed-off-by: Alan Stern <stern@rowland.harvard.edu>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
18 years ago[PATCH] UHCI: remove hc_inaccessible flag
Alan Stern [Mon, 5 Jun 2006 16:21:30 +0000 (12:21 -0400)]
[PATCH] UHCI: remove hc_inaccessible flag

This patch (as706) removes the private hc_inaccessible flag from
uhci-hcd.  It's not needed because it conveys exactly the same
information as the generic HCD_FLAG_HW_ACCESSIBLE bit.

In its place goes a new flag recording whether the controller is dead.
The new code allows a complete device reset to resurrect a dead
controller (although usbcore doesn't yet implement such a facility).

Signed-off-by: Alan Stern <stern@rowland.harvard.edu>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
18 years ago[PATCH] UHCI: various updates
David Brownell [Mon, 5 Jun 2006 16:16:39 +0000 (12:16 -0400)]
[PATCH] UHCI: various updates

This patch (as705) contains a small set of updates for uhci-hcd written
mostly by Dave Brownell:

  * Root hub suspend messages come out labeled as root hub messages;
    PCI messages should only come out when the pci device suspends.

  * Rename the reset() method to better match its init() role

  * Behave more like the other HCDs by returning -ESHUTDOWN for root-hub
    suspend/resume errors.

  * When an URB fails, associate the message with the usb device not
    the host controller (it still hides endpoint and direction)

From: David Brownell <david-b@pacbell.net>
Signed-off-by: Alan Stern <stern@rowland.harvard.edu>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
18 years ago[PATCH] usb-storage: unusual_devs entry for Nikon DSC D70s
Alan Stern [Tue, 6 Jun 2006 19:19:17 +0000 (15:19 -0400)]
[PATCH] usb-storage: unusual_devs entry for Nikon DSC D70s

This patch (as704) adds an unusual_devs entry for the Nikon DSC D70s,
which uses a different Product ID from the D70.  It also moves the entry
for the DSC E2000 up in the list, to preserve the numerical ordering.

Signed-off-by: Alan Stern <stern@rowland.harvard.edu>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
18 years ago[PATCH] USB: Add Apple MacBook product IDs to usbhid
Rene Rebe [Sun, 4 Jun 2006 09:39:09 +0000 (11:39 +0200)]
[PATCH] USB: Add Apple MacBook product IDs to usbhid

This adds the Apple MacBook product IDs for the Fn translation
to the usbhid.

Signed-off-by: Rene Rebe <rene@exactcode.de>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
18 years ago[PATCH] USB: EHCI on non-Au1200 build fix
Ralf Baechle [Sat, 3 Jun 2006 22:58:55 +0000 (23:58 +0100)]
[PATCH] USB: EHCI on non-Au1200 build fix

Including ehci-au1xxx.c on a non-Au1200 Alchemy only to have it throw
an error is stupid.

From: Ralf Baechle <ralf@linux-mips.org>
Acked-by: David Brownell <david-b@pacbell.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
18 years ago[PATCH] USB: free allocated memory on io_edgeport startup memory failure
Chris Lund [Sat, 3 Jun 2006 20:58:19 +0000 (13:58 -0700)]
[PATCH] USB: free allocated memory on io_edgeport startup memory failure

While an Edgeport is allocating individual port structures, if kmalloc
returns NULL, the serial structure is freed and -ENOMEM, but the ports
allocated before the failure are not freed.  This patch addresses that
condition.

Signed-off-by: Christopher Lund <docmax@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
18 years ago[PATCH] USB: new devices for the Option driver
Matthias Urlichs [Fri, 2 Jun 2006 09:48:56 +0000 (11:48 +0200)]
[PATCH] USB: new devices for the Option driver

This patch extends the "option" driver with a few more devices, some of
which are actually connected to USB the "right" way -- as opposed to
doing it via PCMCIA and OHCI.

Signed-Off-By: Matthias Urlichs <smurf@debian.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
18 years ago[PATCH] usb: drivers/usb/core/devio.c dereferences a userspace pointer
Philippe Retornaz [Fri, 2 Jun 2006 03:48:46 +0000 (20:48 -0700)]
[PATCH] usb: drivers/usb/core/devio.c dereferences a userspace pointer

See http://bugzilla.kernel.org/show_bug.cgi?id=6617.

This function dereference a __user pointer.

Signed-off-by: Philippe Retornaz <couriousous@mandriva.org>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
18 years ago[PATCH] usbcore: recovery from Set-Configuration failure
Alan Stern [Thu, 1 Jun 2006 17:59:16 +0000 (13:59 -0400)]
[PATCH] usbcore: recovery from Set-Configuration failure

This patch (as703) improves the error handling when a Set-Configuration
request fails.  The old interfaces are all unregistered before the
request is sent, and if the request fails then we don't know what config
the device is using.  So it makes no sense to leave actconfig pointing
to the old configuration with its invalid interfaces.

Signed-off-by: Alan Stern <stern@rowland.harvard.edu>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
18 years ago[PATCH] usbhid: use usb_reset_composite_device
Alan Stern [Thu, 1 Jun 2006 17:55:28 +0000 (13:55 -0400)]
[PATCH] usbhid: use usb_reset_composite_device

This patch (as702) makes usbhid use the new usb_reset_composite_device
API.  Now HID interfaces can coexist with other interfaces on the same
device, and a reset can safely be requested by any of the drivers.

Signed-off-by: Alan Stern <stern@rowland.harvard.edu>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
18 years ago[PATCH] usb-storage: use usb_reset_composite_device
Alan Stern [Thu, 1 Jun 2006 17:52:56 +0000 (13:52 -0400)]
[PATCH] usb-storage: use usb_reset_composite_device

This patch (as701) modifies usb-storage to take advantage of the new
usb_reset_composite_device() API.  Now we will be able to safely request
port resets even if other drivers are bound to a mass-storage device.

Signed-off-by: Alan Stern <stern@rowland.harvard.edu>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
18 years ago[PATCH] USB hub: use usb_reset_composite_device
Alan Stern [Thu, 1 Jun 2006 17:37:24 +0000 (13:37 -0400)]
[PATCH] USB hub: use usb_reset_composite_device

This patch (as700) modifies the hub driver to take advantage of the new
usb_reset_composite_device API.  The existing code had special-case
calls stuck into usb_reset_device, just before and after the reset.
With the new version there's no need for special-case stuff; it all
happens naturally in the form of pre_reset and post_reset notifications.

Signed-off-by: Alan Stern <stern@rowland.harvard.edu>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
18 years ago[PATCH] usbcore: port reset for composite devices
Alan Stern [Thu, 1 Jun 2006 17:33:42 +0000 (13:33 -0400)]
[PATCH] usbcore: port reset for composite devices

This patch (as699) adds usb_reset_composite_device(), a routine for
sending a USB port reset to a device with multiple interfaces owned by
different drivers.  Drivers are notified about impending and completed
resets through two new methods in the usb_driver structure.

The patch modifieds the usbfs ioctl code to make it use the new routine
instead of usb_reset_device().  Follow-up patches will modify the hub,
usb-storage, and usbhid drivers so they can utilize this new API.

Signed-off-by: Alan Stern <stern@rowland.harvard.edu>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
18 years ago[PATCH] USB: cdc_ether: recognize olympus r1000 (fix regression)
David Brownell [Wed, 31 May 2006 03:49:29 +0000 (20:49 -0700)]
[PATCH] USB: cdc_ether: recognize olympus r1000 (fix regression)

Earlier work splitting the "usbnet" driver out into a core plus driver
modules was missing a blacklist entry for the Olympus R-1000; it must
not use the CDC Ethernet driver, only the "zaurus" support works with
it.

Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
18 years ago[PATCH] USB: negative index in drivers/usb/host/isp116x-hcd.c
Eric Sesterhenn [Fri, 2 Jun 2006 03:48:45 +0000 (20:48 -0700)]
[PATCH] USB: negative index in drivers/usb/host/isp116x-hcd.c

From: Eric Sesterhenn <snakebyte@gmx.de>

This fixes coverity Bug #390.

With the following code

ret = ep->branch = balance(isp116x, ep->period, ep->load);
if (ret < 0)
goto fail;

the problem is that ret and balance are of the type int, and ep->branch is u16.
so the int balance() returns gets reduced to u16 and then converted to an int again,
which removes the sign. Maybe the following little c program can explain it better:

18 years ago[PATCH] USB: Add Sierra Wireless MC5720 ID to airprime.c
Jeremy Fitzhardinge [Wed, 31 May 2006 18:21:52 +0000 (11:21 -0700)]
[PATCH] USB: Add Sierra Wireless MC5720 ID to airprime.c

Recognize the Sierra Wireless MC5720.

Signed-off-by: Jeremy Fitzhardinge <jeremy@goop.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
18 years ago[PATCH] USB: Whiteheat: fix firmware spurious errors
Stuart MacDonald [Wed, 31 May 2006 17:28:40 +0000 (13:28 -0400)]
[PATCH] USB: Whiteheat: fix firmware spurious errors

Attached patch fixes spurious errors during firmware load.

Signed-off-by: Stuart MacDonald <stuartm@connecttech.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
18 years ago[PATCH] Driver for Apple Cinema Display
akpm@osdl.org [Fri, 26 May 2006 20:17:09 +0000 (13:17 -0700)]
[PATCH] Driver for Apple Cinema Display

This is a driver to control the brightness of an Apple Cinema Display over
USB.  It updates the local brightness value if the user presses a button on
the display.

Signed-off-by: Michael Hanselmann <linux-kernel@hansmi.ch>
Cc: Oliver Neukum <oliver@neukum.org>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
18 years ago[PATCH] USB: ftdi_sio: add support for Yost Engineering ServoCenter3.1
Ian Abbott [Tue, 30 May 2006 11:36:30 +0000 (12:36 +0100)]
[PATCH] USB: ftdi_sio: add support for Yost Engineering ServoCenter3.1

This patch adds support for Yost Engineering Inc's ServoCenter 3.1 USB
product to the ftdi_sio driver's device ID table.  The PID was supplied
by Aaron Prose of Yost Engineering on the ftdi-usb-sio-devel list.  The
PID 0xE050 matches the Windows INF files for this device.

Signed-off-by: Ian Abbott <abbotti@mev.co.uk>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
18 years ago[PATCH] usbhid: Remove unneeded blacklist entries
Alan Stern [Tue, 30 May 2006 14:21:14 +0000 (10:21 -0400)]
[PATCH] usbhid: Remove unneeded blacklist entries

Now that usbhid automatically applies HID_QUIRK_NOGET to keyboards and
mice, we no longer need the blacklist entries that were present for no
other purpose.  This patch (as698) removes them.

Signed-off-by: Alan Stern <stern@rowland.harvard.edu>
Acked-by: Vojtech Pavlik <vojtech@suse.cz>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
18 years ago[PATCH] usbcore: Fix broken RNDIS config selection
Giridhar Pemmasani [Tue, 30 May 2006 14:12:30 +0000 (10:12 -0400)]
[PATCH] usbcore: Fix broken RNDIS config selection

RNDIS devices don't get configured owing to a typo in
choose_configuration().  This patch from Giridhar Pemmasani fixes the
typo.

From: Giridhar Pemmasani <giri@lmc.cs.sunysb.edu>
Signed-off-by: Alan Stern <stern@rowland.harvard.edu>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
18 years ago[PATCH] USB: print message when device is rejected due to insufficient power
Daniel Drake [Fri, 26 May 2006 20:36:28 +0000 (21:36 +0100)]
[PATCH] USB: print message when device is rejected due to insufficient power

2.6.16 introduces USB power budgeting in the Linux kernel, and since then, a
fair number of users have observed that some of their devices no longer work in
unpowered hubs (this is not a bug, the devices claim that they need more than
100mA).

The very least we can do is print an informational message to the kernel log
when this happens, otherwise it is not at all clear why the device was not
accepted.

Signed-off-by: Daniel Drake <dsd@gentoo.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
18 years ago[PATCH] USB: more pegasus log spamming removed
David Brownell [Fri, 26 May 2006 17:17:03 +0000 (10:17 -0700)]
[PATCH] USB: more pegasus log spamming removed

Remove more log spamming from pegasus:  stop talking to the device once we
see ENODEV reported.  It may take a while before khubd notifies us.

Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
18 years ago[PATCH] ub: random cleanups
Pete Zaitcev [Fri, 26 May 2006 03:08:50 +0000 (20:08 -0700)]
[PATCH] ub: random cleanups

Remove some silly messages and cast in stone "temporary" messages which
we keep around. Also, I am hesitant to remove the initialization retries
without having the hardware to test (anyone who was at KS04 has a spare?)

Signed-off-by: Pete Zaitcev <zaitcev@yahoo.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
18 years ago[PATCH] ub: atomic add_disk
Pete Zaitcev [Fri, 26 May 2006 03:04:54 +0000 (20:04 -0700)]
[PATCH] ub: atomic add_disk

<zaitcev> I am taling about this: "if (disk->flags & GENHD_FL_UP) del_gendisk(disk);"
<zaitcev> If del_gendisk() undoes add_disk() like viro just said, why is it conditional?
<viro> huh?
<viro> add_disk() sets the damn flag
<zaitcev> So, I should not need to check ever
<viro> so the above is "if I've called add_disk(), call gendisk()"
<viro> which might be what you want, of course
<viro> but usually you know if you'd done add_disk() on that puppy anyway

In ub, nobody upstream should ever see half-constructed disks before
they were passed to add_disk. To that end, only add the struct lun to
the list on the path of no return. With that fix in place, we do
not need to test GENHD_FL_UP.

Signed-off-by: Pete Zaitcev <zaitcev@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
18 years ago[PATCH] USB: rmmod pl2303 after -28
Pete Zaitcev [Wed, 24 May 2006 18:04:04 +0000 (11:04 -0700)]
[PATCH] USB: rmmod pl2303 after -28

Wait for the scheduled work to finish before freeing memory, prevent oops.
http://bugzilla.kernel.org/show_bug.cgi?id=6596

Signed-off-by: Pete Zaitcev <zaitcev@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
18 years ago[PATCH] improved TT scheduling for EHCI
Dan Streetman [Wed, 24 May 2006 16:39:16 +0000 (09:39 -0700)]
[PATCH] improved TT scheduling for EHCI

This updates the EHCI driver by adding an improved scheduler for the
transaction translators, found in USB 2.0 hubs and used for low and
full speed devices.

 - adds periodic_tt_usecs() and some helper functions, which does
   the same thing that "periodic_usecs" does, except on the other
   side of the TT, i.e.  it calculates the low/fullspeed bandwidth
   usage instead of highspeed.

 - adds a tt_available() function which is the new implementation
   of what tt_no_collision() does ... while tt_no_collision() ensures
   that each TT handles only 1 periodic transfer at a time (a very
   pessimistic approach) this version instead tracks bandwidth and
   allows each TT to handle as many transfers as will fit on each TT's
   downstream bus (closer to best-case).

The new scheduler is selected by a config option, marked as EXPERIMENTAL
so it can be tested (and more broadly reviewed) for a while until it
seems safe to remove the original scheduler.

Signed-off-by: Dan Streetman <ddstreet@ieee.org>
Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
18 years ago[PATCH] usb-storage: get rid of the timer during URB submission
Franck Bui-Huu [Wed, 24 May 2006 14:57:28 +0000 (16:57 +0200)]
[PATCH] usb-storage: get rid of the timer during URB submission

This patch uses completion timeout instead of a timer to implement
a timeout when submitting an URB.

It also put the task in interruptible state instead of an
uninterruptible one while waiting for the completion.

Signed-off-by: Franck Bui-Huu <vagabon.xyz@gmail.com>
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: Syntax cleanup for pl2303 (trailing backslash)
Pete Zaitcev [Tue, 23 May 2006 05:05:17 +0000 (22:05 -0700)]
[PATCH] USB: Syntax cleanup for pl2303 (trailing backslash)

Remove the silly trailing backslash.

Signed-off-by: Pete Zaitcev <zaitcev@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
18 years ago[PATCH] USB: Improve Kconfig comment for mct_u232
Pete Zaitcev [Tue, 23 May 2006 05:02:32 +0000 (22:02 -0700)]
[PATCH] USB: Improve Kconfig comment for mct_u232

Add a couple of supported devices into the help message.

It's a long story... I promised this comment changed to a user long ago,
so I'd like to have that promise kept. In reality though, nobody is
likely to read this anyway.

Signed-off-by: Pete Zaitcev <zaitcev@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
18 years ago[PATCH] USB serial: encapsulate schedule_work, remove double-calling
Pete Zaitcev [Tue, 23 May 2006 04:58:49 +0000 (21:58 -0700)]
[PATCH] USB serial: encapsulate schedule_work, remove double-calling

I'm going to throw schedule_work away, it's retarded. But for starters,
let's have it encapsulated.

Also, generic and whiteheat were both calling usb_serial_port_softint
and scheduled work. Only one was necessary.

Signed-off-by: Pete Zaitcev <zaitcev@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
18 years ago[PATCH] usb: io_edgeport, cleanup to unicode handling
Pete Zaitcev [Tue, 23 May 2006 04:49:44 +0000 (21:49 -0700)]
[PATCH] usb: io_edgeport, cleanup to unicode handling

Clean up the unicode handling in io_edgeport. Make get_string size-limited.

Signed-off-by: Pete Zaitcev <zaitcev@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
18 years ago[PATCH] usbtest: report errors in iso tests
Alan Stern [Mon, 22 May 2006 20:47:13 +0000 (16:47 -0400)]
[PATCH] usbtest: report errors in iso tests

This patch (as693b) makes the usbtest driver report errors in the
isochronous bulk transfer tests instead of always returning 0.  As an
arbitrary cutoff, an error is returned if more than 10% of the packet
transfers fail.  It also stops a test immediately upon receiving an URB
submission error.

For a test harness, it's especially important to report when errors occur!

Signed-off-by: Alan Stern <stern@rowland.harvard.edu>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
18 years ago[PATCH] gadgetfs: fix memory leaks
Alan Stern [Mon, 22 May 2006 16:27:38 +0000 (12:27 -0400)]
[PATCH] gadgetfs: fix memory leaks

This patch (as692) fixes a few memory leaks in some unimportant error
pathways of the gadgetfs driver.

Signed-off-by: Alan Stern <stern@rowland.harvard.edu>
Acked-by: David Brownell <david-b@pacbell.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
18 years ago[PATCH] gadgetfs: fix AIO interface bugs
Alan Stern [Mon, 22 May 2006 16:26:31 +0000 (12:26 -0400)]
[PATCH] gadgetfs: fix AIO interface bugs

This patch (as691) fixes a few errors in the AIO interface for the
gadgetfs driver.  Now requests will complete properly instead of hanging.

Signed-off-by: Alan Stern <stern@rowland.harvard.edu>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
18 years ago[PATCH] USB: new cp2101 device
Vitja Makarov [Tue, 30 May 2006 20:40:06 +0000 (00:40 +0400)]
[PATCH] USB: new cp2101 device

By the way I have to ask you to add new (vid,pid) pair to cp2101 driver.

This device is argussoft's avr in-system programmer AS3M,
http://atmel.argussoft.ru/hard.htm
it's based on cp2101 chip and works pretty well with the linux driver.

It could be used with argussoft's `asisp1109.exe'
(http://atmel.argussoft.ru/download/software/as-tools.soft/asisp.zip)
tool run under wine.

Signed-off-by: Vitja Makarov <vitja.makarov@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
18 years ago[PATCH] USB: add usb_interrupt_msg() function for api completeness.
Greg Kroah-Hartman [Fri, 19 May 2006 20:20:20 +0000 (13:20 -0700)]
[PATCH] USB: add usb_interrupt_msg() function for api completeness.

Really just a wrapper around usb_bulk_msg() but now it's documented
much better.

Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
18 years ago[PATCH] Fix a deadlock in usbtest
Franck Bui-Huu [Mon, 15 May 2006 17:23:53 +0000 (19:23 +0200)]
[PATCH] Fix a deadlock in usbtest

ctrl_complete functions acquires ctx->lock and tries to unlink
all queued urbs in case of errors through usb_unlink_urb func.
In its turn usb_unlink_urb calls, through the hcd driver,
usb_hcd_giveback_urb which calls ctrl_complete again. At this
time, ctx->lock is already taken by the same function.

Signed-off-by: Franck Bui-Huu <vagabon.xyz@gmail.com>
Cc: David Brownell <david-b@pacbell.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
18 years ago[PATCH] UHCI: remove ISO TDs as they are used
Alan Stern [Fri, 19 May 2006 20:52:35 +0000 (16:52 -0400)]
[PATCH] UHCI: remove ISO TDs as they are used

This patch (as690) does the same thing for ISO TDs as as680 did for
non-ISO TDs: free them as they are used rather than all at once when an
URB is complete.  At the same time it fixes a minor buglet (I'm not
aware of it ever affecting anyone): An ISO TD should be retired when its
frame is over, regardless of whether or not the hardware has marked it
inactive.

Signed-off-by: Alan Stern <stern@rowland.harvard.edu>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
18 years ago[PATCH] UHCI: store the period in the queue header
Alan Stern [Fri, 19 May 2006 20:44:55 +0000 (16:44 -0400)]
[PATCH] UHCI: store the period in the queue header

This patch (as689) stores the period for periodic transfers (interrupt
and ISO) in the queue header.  This is necessary for proper bandwidth
tracking (not yet implemented).  It also makes the scheduling of ISO
transfers a bit more rigorous, with checks for out-of-bounds frame
numbers.

Signed-off-by: Alan Stern <stern@rowland.harvard.edu>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
18 years ago[PATCH] UHCI: fix race in ISO dequeuing
Alan Stern [Fri, 19 May 2006 20:39:52 +0000 (16:39 -0400)]
[PATCH] UHCI: fix race in ISO dequeuing

This patch (as688) fixes a small race in uhci-hcd.  Because ISO queues
aren't controlled by queue headers, they can't be unlinked.  Only
individual URBs can.  So whenever multiple ISO URBs are dequeued, it's
necessary to make sure the hardware is done with each one.  We can't
assume that dequeuing the first URB will suffice to unlink the entire
queue.

Signed-off-by: Alan Stern <stern@rowland.harvard.edu>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
18 years ago[PATCH] UHCI: use integer-sized frame numbers
Alan Stern [Fri, 19 May 2006 20:34:57 +0000 (16:34 -0400)]
[PATCH] UHCI: use integer-sized frame numbers

This patch (as687) changes uhci-hcd to keep track of frame numbers as
full-sized integers rather than 11-bit values.  This makes them a lot
easier to handle and makes it possible to schedule beyond a 2-second
window, should anyone ever want to do so.

Signed-off-by: Alan Stern <stern@rowland.harvard.edu>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
18 years ago[PATCH] USB: Allow high-bandwidth isochronous packets via usbfs
Micah Dowty [Fri, 19 May 2006 18:26:24 +0000 (11:26 -0700)]
[PATCH] USB: Allow high-bandwidth isochronous packets via usbfs

This patch increases an arbitrary limit on the size of
individual isochronous packets submitted via usbfs. The
limit is still arbitrary, but it's now large enough to
support the maximum packet size used by high-bandwidth
isochronous transfers.

Signed-off-by: Micah Dowty <micah@vmware.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
18 years ago[PATCH] USB: Remove 4088-byte limit on usbfs control URBs
Micah Dowty [Fri, 19 May 2006 18:20:11 +0000 (11:20 -0700)]
[PATCH] USB: Remove 4088-byte limit on usbfs control URBs

This patch removes the artificial 4088-byte limit that usbfs
currently places on Control transfers. The USB spec does not
specify a strict limit on the size of an entire control transfer.
It does, however, state that the data stage "follows the same
protocol rules as bulk transfers." (USB 2, 8.5.3)

The level of support for large control transfers in real host
controllers varies, but it's important to support at least 4K
transfers. Windows enforces a maximum control transfer size
of 4K, so there exists some hardware that requires a full 4096
byte data stage. Without this patch, we fall short of that by
8 bytes on architectures with a 4K page size, and it becomes
impossible to support such hardware with a user-space driver.

Since any limit placed on control transfers by usbfs would be
arbitrary, this patch replaces the PAGE_SIZE limit with the same
arbitrary limit used by bulk transfers.

Signed-off-by: Micah Dowty <micah@vmware.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
18 years ago[PATCH] USB: correct the USB info in Documentation/power/swsusp.txt
David Brownell [Wed, 17 May 2006 00:33:14 +0000 (17:33 -0700)]
[PATCH] USB: correct the USB info in Documentation/power/swsusp.txt

The swsusp.txt documentation harshes confusingly on USB, and this patch
addresses the issue.  It's harsh because it blames USB for some issues
that are generic to all drivers -- especially those supporting removable
media -- and it's confusing since it says that USB has the issue with
"suspend" not just swsusp ... while in reality, USB doesn't have the
issue when real system suspend states are used.

Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
Acked-by: Pavel Machek <pavel@ucw.cz>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
18 years ago[PATCH] UHCI: Work around old Intel bug
Alan Stern [Fri, 12 May 2006 15:41:59 +0000 (11:41 -0400)]
[PATCH] UHCI: Work around old Intel bug

Some old Intel UHCI controllers have a bug that has shown up in a few
systems (the PIIX3 "Neptune" chip set).  Until now there has not been
any simple way to work around the bug, but the lastest changes in
uhci-hcd have made it easy.  This patch (as684) adds the work-around.

Signed-off-by: Alan Stern <stern@rowland.harvard.edu>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
18 years ago[PATCH] UHCI: Reimplement FSBR
Alan Stern [Fri, 12 May 2006 15:35:45 +0000 (11:35 -0400)]
[PATCH] UHCI: Reimplement FSBR

This patch (as683) re-implements Full-Speed Bandwidth Reclamation (FSBR)
properly.  It keeps track of which endpoint queues have advanced, and
when none have advanced for a sufficiently long time, FSBR is turned
off.  The next TD on each of the non-moving queues is modified to
generate an interrupt on completion, so that FSBR can be re-enabled as
soon as the hardware starts to make some progress.

Signed-off-by: Alan Stern <stern@rowland.harvard.edu>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
18 years ago[PATCH] UHCI: Eliminate the TD-removal list
Alan Stern [Fri, 12 May 2006 15:29:04 +0000 (11:29 -0400)]
[PATCH] UHCI: Eliminate the TD-removal list

This patch (as682) gets rid of the TD-removal list in uhci-hcd.  It is
no longer needed because now TDs are not freed until we know the
hardware isn't using them.  It also simplifies the code for adding and
removing TDs to/from URBs.

Signed-off-by: Alan Stern <stern@rowland.harvard.edu>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
18 years ago[PATCH] UHCI: Move code for cleaning up unlinked URBs
Alan Stern [Fri, 12 May 2006 15:23:19 +0000 (11:23 -0400)]
[PATCH] UHCI: Move code for cleaning up unlinked URBs

This patch (as681) moves some code for cleaning up after unlinked URBs
out of the general completion pathway into the unlinking pathway.

Signed-off-by: Alan Stern <stern@rowland.harvard.edu>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
18 years ago[PATCH] UHCI: Remove non-iso TDs as they are used
Alan Stern [Fri, 12 May 2006 15:19:19 +0000 (11:19 -0400)]
[PATCH] UHCI: Remove non-iso TDs as they are used

This patch (as680) frees non-isochronous TDs as they are used, rather
than all at once when an URB is complete.  Although not a terribly
important change in itself, it opens the door to a later enhancement
that will reduce storage requirements by allocating only a limited
number of TDs at any time for each endpoint queue.

Signed-off-by: Alan Stern <stern@rowland.harvard.edu>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
18 years ago[PATCH] UHCI: Common result routine for Control/Bulk/Interrupt
Alan Stern [Fri, 12 May 2006 15:14:25 +0000 (11:14 -0400)]
[PATCH] UHCI: Common result routine for Control/Bulk/Interrupt

This patch (as679) combines the result routine for Control URBs with the
routine for Bulk/Interrupt URBs.  Along the way I eliminated the
debugging printouts for Control transfers unless the debugging level is
set higher than 1.  I also eliminated a long-unused (#ifdef'ed-out)
section that works around some buggy old APC BackUPS devices.

Signed-off-by: Alan Stern <stern@rowland.harvard.edu>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
18 years ago[PATCH] usbhid: automatically set HID_QUIRK_NOGET for keyboards and mice
Alan Stern [Mon, 15 May 2006 18:49:04 +0000 (14:49 -0400)]
[PATCH] usbhid: automatically set HID_QUIRK_NOGET for keyboards and mice

It seems to be relatively common for USB keyboards and mice to dislike
being polled for reports.  Since there's no need to poll a keyboard or
a mouse, this patch (as685) automatically sets the HID_QUIRK_NOGET flag
for devices that advertise themselves as either sort of device with boot
protocol support.

This won't cure all the problems since some devices don't support the
boot protocol, but it's simple and easy and it should fix quite a few
problems.

Signed-off-by: Alan Stern <stern@rowland.harvard.edu>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
18 years ago[PATCH] USB HID/HIDBP, INPUT DRIVERS: fix various usb/input/hid-input.c bugs that...
Bart Massey [Mon, 8 May 2006 21:40:13 +0000 (14:40 -0700)]
[PATCH] USB HID/HIDBP, INPUT DRIVERS: fix various usb/input/hid-input.c bugs that make Apple Mighty Mouse work poorly

Transposed lines of code in drivers/usb/input/hid-input.c causes the
capability bits for a new HID device to be set before quirks are applied
at configuration time.  When an HID event is then sent up to the input
layer, it may then be discarded as irrelevant because the wrong
capability bit is set.

Further, the quirks for the Apple Mighty Mouse are not quite right: the
horizontal scrolling needs its axis reversed, and the left and center
buttons are transposed.  Also, the mouse is labeled in the kernel with
its earlier name (I think) of Apple PowerMouse.

Steps to reproduce problem: Plug in an Apple Mighty Mouse.  Note that
horizontal scrolling doesn't work at all, and in fact doesn't generate
any input events on /dev/input/eventN.  Note also that pushing the
middle button performs the right button action, and vice versa.  Once
you have the horizontal scrolling working, note that it is backward WRT
both to vertical scrolling and to common sense.

This patch maybe should be broken up, as it does address two problems.
The transposed code in hidinput_configure_usage() probably creates bugs
beyond just the Mighty Mouse.  The rest of the patch renames POWERMOUSE
to MIGHTYMOUSE everywhere (which I *believe* is correct), fixes the
MIGHTYMOUSE quirk to swap the center and right mouse buttons, and adds a
new quirk HID_QUIRK_INVERT_HWHEEL also assigned to the MIGHTYMOUSE with
code in hidinput_hid_event() to implement it.

Signed-off-by: Bart Massey <bart@cs.pdx.edu>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
18 years ago[PATCH] USB: UHCI: fix obscure bug in enqueue()
Alan Stern [Fri, 5 May 2006 20:32:02 +0000 (16:32 -0400)]
[PATCH] USB: UHCI: fix obscure bug in enqueue()

This patch (as676) fixes a small bug in uhci-hcd's enqueue routine.  When
an URB is unlinked or gets an error and the completion handler queues
another URB for the same endpoint, the queue shouldn't be allowed to start
up again until the handler returns.  Not even if the new URB is the only
one on its queue.

Signed-off-by: Alan Stern <stern@rowland.harvard.edu>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
18 years ago[PATCH] USB: UHCI: store the endpoint type in the QH structure
Alan Stern [Fri, 5 May 2006 20:26:58 +0000 (16:26 -0400)]
[PATCH] USB: UHCI: store the endpoint type in the QH structure

This patch (as675) simplifies uhci-hcd slightly by storing each endpoint's
type in the corresponding Queue Header structure.

Signed-off-by: Alan Stern <stern@rowland.harvard.edu>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>