deliverable/linux.git
17 years agoMerge branch 'for-linus' of master.kernel.org:/pub/scm/linux/kernel/git/roland/infiniband
Linus Torvalds [Mon, 19 Feb 2007 21:00:26 +0000 (13:00 -0800)] 
Merge branch 'for-linus' of /linux/kernel/git/roland/infiniband

* 'for-linus' of master.kernel.org:/pub/scm/linux/kernel/git/roland/infiniband:
  IB/core: Set static rate in ib_init_ah_from_path()
  IB/ipath: Make ipath_map_sg() static
  IB/core: Fix sparse warnings about shadowed declarations
  RDMA/cma: Add multicast communication support
  IB/sa: Track multicast join/leave requests
  IPoIB: CM error handling thinko fix
  RDMA/cxgb3: Remove Open Grid Computing copyrights in iw_cxgb3 driver
  RDMA/cxgb3: Fail posts synchronously when in TERMINATE state
  RDMA/iwcm: iw_cm_id destruction race fixes
  IB/ehca: Change query_port() to return LINK_UP instead UNKNOWN
  IB/ehca: Allow en/disabling scaling code via module parameter
  IB/ehca: Fix race condition/locking issues in scaling code
  IB/ehca: Rework irq handler
  IPoIB: Only allow root to change between datagram and connected mode
  IB/mthca: Fix allocation of ICM chunks in coherent memory
  IB/mthca: Allow the QP state transition RESET->RESET

17 years agoMerge master.kernel.org:/pub/scm/linux/kernel/git/gregkh/pci-2.6
Linus Torvalds [Mon, 19 Feb 2007 20:59:55 +0000 (12:59 -0800)] 
Merge /pub/scm/linux/kernel/git/gregkh/pci-2.6

* master.kernel.org:/pub/scm/linux/kernel/git/gregkh/pci-2.6:
  PCI: Make PCI device numa-node attribute visible in sysfs
  PCI: add systems for automatic breadth-first device sorting
  PCI: PCI devices get assigned redundant IRQs
  PCI: Make CARDBUS_MEM_SIZE and CARDBUS_IO_SIZE boot options
  PCI: pci.txt fix __devexit() usage
  PCI/sysfs/kobject kernel-doc fixes

17 years agoMerge master.kernel.org:/pub/scm/linux/kernel/git/gregkh/driver-2.6
Linus Torvalds [Mon, 19 Feb 2007 20:59:26 +0000 (12:59 -0800)] 
Merge /pub/scm/linux/kernel/git/gregkh/driver-2.6

* master.kernel.org:/pub/scm/linux/kernel/git/gregkh/driver-2.6:
  sysfs: fix build errors: uevent with CONFIG_SYSFS=n
  pcmcia: some class_device fallout
  Driver core: device_add_attrs() cleanup
  debugfs: Remove misleading comments.
  debugfs: implement symbolic links
  Driver: remove redundant kobject_unregister checks
  kobject: kobj->k_name verification fix
  serial: Add PCMCIA IDs for Quatech DSP-100 dual RS232 adapter.
  Driver core: let request_module() send a /sys/modules/kmod/-uevent
  Driver.h copyright update

17 years agoMerge master.kernel.org:/pub/scm/linux/kernel/git/gregkh/usb-2.6
Linus Torvalds [Mon, 19 Feb 2007 20:59:04 +0000 (12:59 -0800)] 
Merge /pub/scm/linux/kernel/git/gregkh/usb-2.6

* master.kernel.org:/pub/scm/linux/kernel/git/gregkh/usb-2.6: (39 commits)
  USB: at91-ohci, handle extra at91sam9261 ahb clock
  USB: another id for cp2101 driver
  USB: ueagle-atm.c needs sched.h
  USB: at91_udc, shrink runtime footprint
  usbnet: add missing Kconfig for KC2190 cables
  usbnet: init fault (oops) cleanup, whitespace fixes
  usbnet: recognize SiteCom CN-124
  usb: Remove Airprime device from option.c
  USB: change __init to __devinit for isp116x_probe
  USB: ps3: don't call ps3_system_bus_driver_register on other platforms
  USB: hid-core.c: Removes GTCO CalComp Interwrite IPanel PIDs from blacklist
  USB: kernel-doc fixes
  USB: quirky device for cdc-acm
  USB: cdc-acm: fix incorrect throtteling, make set_control optional
  USB: unconfigure devices which have config 0
  USB: make usb_iso_packet_descriptor.status signed
  USB: fix g_serial small error
  USB: use __u32 rather than u32 in userspace ioctls in usbdevice_fs.h
  USB Storage: US_FL_IGNORE_RESIDUE needed for Aiptek MP3 Player
  USB: Fix misspelled "USBNET_MII" kernel config option.
  ...

17 years agoPull bugzilla-7897 into release branch
Len Brown [Sat, 17 Feb 2007 03:12:04 +0000 (22:12 -0500)] 
Pull bugzilla-7897 into release branch

17 years agoPull bugzilla-7887 into release branch
Len Brown [Sat, 17 Feb 2007 03:11:57 +0000 (22:11 -0500)] 
Pull bugzilla-7887 into release branch

17 years agoPull bugzilla-7570 into release branch
Len Brown [Sat, 17 Feb 2007 03:11:50 +0000 (22:11 -0500)] 
Pull bugzilla-7570 into release branch

17 years agoPull bugzilla-7200 into release branch
Len Brown [Sat, 17 Feb 2007 03:11:37 +0000 (22:11 -0500)] 
Pull bugzilla-7200 into release branch

17 years agoPull bugzilla-7122 into release branch
Len Brown [Sat, 17 Feb 2007 03:11:31 +0000 (22:11 -0500)] 
Pull bugzilla-7122 into release branch

17 years agoPull bugzilla-5534 into release branch
Len Brown [Sat, 17 Feb 2007 03:11:21 +0000 (22:11 -0500)] 
Pull bugzilla-5534 into release branch

17 years agoPull remove-hotkey into release branch
Len Brown [Sat, 17 Feb 2007 03:11:02 +0000 (22:11 -0500)] 
Pull remove-hotkey into release branch

17 years agoPull sony into release branch
Len Brown [Sat, 17 Feb 2007 03:10:55 +0000 (22:10 -0500)] 
Pull sony into release branch

17 years agoPull fluff into release branch
Len Brown [Sat, 17 Feb 2007 03:10:32 +0000 (22:10 -0500)] 
Pull fluff into release branch

Conflicts:

arch/x86_64/pci/mmconfig.c
drivers/acpi/bay.c

Signed-off-by: Len Brown <len.brown@intel.com>
17 years agoACPI: IA64: react to acpi_table_parse() return value change
John Keller [Fri, 16 Feb 2007 21:07:27 +0000 (15:07 -0600)] 
ACPI: IA64: react to acpi_table_parse() return value change

acpi_boot_init() is making a bad check on the return
status from acpi_table_parse(). acpi_table_parse() now
returns zero on success, one on failure.

Signed-off-by: Aaron Young <ayoung@sgi.com>
Signed-off-by: Len Brown <len.brown@intel.com>
17 years agoACPI: hotkey: remove driver, per feature-removal-schedule.txt
Len Brown [Sat, 10 Feb 2007 06:18:25 +0000 (01:18 -0500)] 
ACPI: hotkey: remove driver, per feature-removal-schedule.txt

Signed-off-by: Len Brown <len.brown@intel.com>
17 years agoPull misc-for-upstream into release branch
Len Brown [Fri, 16 Feb 2007 23:52:41 +0000 (18:52 -0500)] 
Pull misc-for-upstream into release branch

Conflicts:

drivers/usb/misc/appledisplay.c

Signed-off-by: Len Brown <len.brown@intel.com>
17 years agoACPI: bay: use IS_ERR for return of register_platform_device_simple
Henrique de Moraes Holschuh [Wed, 14 Feb 2007 12:55:00 +0000 (10:55 -0200)] 
ACPI: bay: use IS_ERR for return of register_platform_device_simple

register_platform_device_simple returns ERR_PTR(foo), so test it with
IS_ERR(foo).

Signed-off-by: Henrique de Moraes Holschuh <hmh@hmh.eng.br>
Signed-off-by: Len Brown <len.brown@intel.com>
17 years agoACPI: acpi_unload_table_id() always returns error
John Keller [Thu, 15 Feb 2007 20:08:30 +0000 (14:08 -0600)] 
ACPI: acpi_unload_table_id() always returns error

acpi_unload_table_id() is always returning an error status.
Also, once the matching table is found, don't bother looking
for another match.

Signed-off-by: John Keller <jpk@sgi.com>
Signed-off-by: Len Brown <len.brown@intel.com>
17 years agoUSB: at91-ohci, handle extra at91sam9261 ahb clock
Andrew Victor [Fri, 16 Feb 2007 18:18:58 +0000 (10:18 -0800)] 
USB: at91-ohci, handle extra at91sam9261 ahb clock

The AT91SAM9261 needs to activate an AHB clock (HCK0) to use the USB Host
controller. Previously clock.c would just enable it at startup, but now
all the unused clocks are automatically disabled.

Based on patch from Nicolas Ferre.

Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
17 years agoUSB: another id for cp2101 driver
Pozsar Balazs [Fri, 16 Feb 2007 11:47:49 +0000 (12:47 +0100)] 
USB: another id for cp2101 driver

This patch adds another usb id to the cp2101. It seems to work well.

Please apply, thanks.

Signed-off-by: Pozsar Balazs <pozsy@uhulinux.hu>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
17 years agoUSB: ueagle-atm.c needs sched.h
Randy Dunlap [Fri, 16 Feb 2007 09:47:33 +0000 (01:47 -0800)] 
USB: ueagle-atm.c needs sched.h

Driver needs sched.h for try_to_freeze().

Signed-off-by: Randy Dunlap <randy.dunlap@oracle.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
17 years agoUSB: at91_udc, shrink runtime footprint
David Brownell [Fri, 16 Feb 2007 02:47:17 +0000 (18:47 -0800)] 
USB: at91_udc, shrink runtime footprint

This is a runtime codespace shrink:  in most cases, platform devices should
put probe() should in the init section, and remove() in the exit section.
And I have no idea why the module init/exit routines were mismarked.
It also moves one function table into read-only data.

Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
17 years agousbnet: add missing Kconfig for KC2190 cables
David Brownell [Fri, 16 Feb 2007 05:32:31 +0000 (21:32 -0800)] 
usbnet: add missing Kconfig for KC2190 cables

Hmm, I noticed that support for one of the USB host-to-host cables
never got fully merged ... Kconfig wouldn't show it!  Fixed.

Also, changed the CDC Subset default to 'y' so that more of these
cables will work out-of-the-box.

Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
17 years agousbnet: init fault (oops) cleanup, whitespace fixes
David Brownell [Fri, 16 Feb 2007 02:52:30 +0000 (18:52 -0800)] 
usbnet: init fault (oops) cleanup, whitespace fixes

This cleans up some error handling paths in usbnet device probing;
one of them could cause oopsing, e.g. with some RNDIS devices.

It also removes some extraneous whitespace.

Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
17 years agousbnet: recognize SiteCom CN-124
David Brownell [Fri, 16 Feb 2007 02:50:01 +0000 (18:50 -0800)] 
usbnet: recognize SiteCom CN-124

Add Sitecom CN-124 device IDs to ALI M5632 usb host-to-host cable;
device IDs from Francois Barre <francois.barre@gmail.com>

Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
17 years agousb: Remove Airprime device from option.c
Kevin Lloyd [Thu, 15 Feb 2007 21:32:13 +0000 (13:32 -0800)] 
usb: Remove Airprime device from option.c

from: Kevin Lloyd <linux@sierrawireless.com>

This patch removes the Airprime 5220 device (branded as Audiovox) from
the option.c driver. This device is already supported by the sierra.c
driver.

This was based off of the option.c driver found in kernel 2.6.20-git11.

Signed-off-by: Kevin Lloyd <linux@sierrawireless.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
17 years agoUSB: change __init to __devinit for isp116x_probe
Prarit Bhargava [Fri, 9 Feb 2007 09:51:15 +0000 (01:51 -0800)] 
USB: change __init to __devinit for isp116x_probe

Change __init to __devinit for isp116x_probe.

Resolves MODPOST warning:

WARNING: drivers/usb/host/isp116x-hcd.o - Section mismatch: reference to
.init.text:isp116x_probe from .data.rel.local between 'isp116x_driver' (at
offset 0x0) and 'isp116x_hc_driver'

Signed-off-by: Prarit Bhargava <prarit@redhat.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
17 years agoUSB: ps3: don't call ps3_system_bus_driver_register on other platforms
Ishizaki Kou [Wed, 14 Feb 2007 07:04:17 +0000 (16:04 +0900)] 
USB: ps3: don't call ps3_system_bus_driver_register on other platforms

ps3_system_bus_driver_register is PS3 platform specific function.
On other platforms, it triggers WARN_ON in kref_get.

Signed-off-by: Kou Ishizaki <kou.ishizaki@toshiba.co.jp>
Acked-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Cc: David Brownell <dbrownell@users.sourceforge.net>
Cc: Geoff Levand <geoffrey.levand@am.sony.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
17 years agoUSB: hid-core.c: Removes GTCO CalComp Interwrite IPanel PIDs from blacklist
Jeremy Roberson [Wed, 14 Feb 2007 23:14:19 +0000 (16:14 -0700)] 
USB: hid-core.c: Removes GTCO CalComp Interwrite IPanel PIDs from blacklist

Removes our GTCO CalComp Interwrite IPanels from the hid-core.c blacklist
because the HID Driver properly handles them.

Signed-off-by: Jeremy A. Roberson <jroberson@gtcocalcomp.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
17 years agoUSB: kernel-doc fixes
Randy Dunlap [Sat, 10 Feb 2007 22:41:41 +0000 (14:41 -0800)] 
USB: kernel-doc fixes

Fix kernel-doc warnings and in USB core.

Signed-off-by: Randy Dunlap <randy.dunlap@oracle.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
17 years agoUSB: quirky device for cdc-acm
Oliver Neukum [Mon, 12 Feb 2007 07:50:03 +0000 (08:50 +0100)] 
USB: quirky device for cdc-acm

here's a quirklist entry reported by Stephen Murphy.

Signed-off-by: Oliver Neukum <oneukum@suse.de>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
17 years agoUSB: cdc-acm: fix incorrect throtteling, make set_control optional
Oliver Neukum [Mon, 12 Feb 2007 07:41:35 +0000 (08:41 +0100)] 
USB: cdc-acm: fix incorrect throtteling, make set_control optional

this is Joris' fixes reshuffelled and features renamed as David requested.

- acm_set_control is not mandatory, honour that
- throtteling is reset upon open
- throtteling is read consistently when processing input data

Signed-off-by: Joris van Rantwijk <jorispubl@xs4all.nl>
Signed-off-by: Oliver Neukum <oneukum@suse.de>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
17 years agoUSB: unconfigure devices which have config 0
Alan Stern [Thu, 8 Feb 2007 21:40:43 +0000 (16:40 -0500)] 
USB: unconfigure devices which have config 0

Some USB devices do have a configuration 0, in contravention of the
USB spec.  Normally 0 is supposed to indicate that a device is
unconfigured.

While we can't change what the device is doing, we can change usbcore.
This patch (as852) allows usb_set_configuration() to accept a config
value of -1 as indicating that the device should be unconfigured.  The
request actually sent to the device will still contain 0 as the value.
But even if the device does have a configuration 0, dev->actconfig
will be set to NULL and dev->state will be set to USB_STATE_ADDRESS.

Without some sort of special-case handling like this, there is no way
to unconfigure these non-compliant devices.

Signed-off-by: Alan Stern <stern@rowland.harvard.edu>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
17 years agoUSB: make usb_iso_packet_descriptor.status signed
Pete Zaitcev [Sun, 11 Feb 2007 21:56:13 +0000 (13:56 -0800)] 
USB: make usb_iso_packet_descriptor.status signed

The status in usb_iso_packet_descriptor should be signed, for the benefit
of someone who casts to a long or makes other benign misstep (the principle
of least surprise).

Signed-off-by: Pete Zaitcev <zaitcev@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
17 years agoUSB: fix g_serial small error
Bryan O'Donoghue [Sat, 27 Jan 2007 12:16:32 +0000 (12:16 +0000)] 
USB: fix g_serial small error

A SET_LINE_CODING control request should return a zero length packet
as an ACK to the host, during the status phase of a USB transaction.

The return value of gs_setup_class() is treated as the number of
bytes to write in the status phase of the control request, by
gs_setup(). For this case, the value returned by gs_setup_class should
be zero for SET_LINE_CODING but, right now, appears to be
sizeof(struct usb_cdc_line_coding).

However, if after doing the memcpy of the line coding descriptor we
set the variable "ret" to be zero, we should return the appropiate ZLP
to the host as an ACK in the status phase of the control request.
I've tested this out using Linux as both host and slave and confirmed
that the following small change fixes the spurious return of
sizeof(struct usb_cdc_line_coding)/wLength bytes in the status phase
of a USB_CDC_REQ_SET_LINE_CODING request. It's not a huge bug but, it
is worth fixing.

Signed-off-by: Bryan O'Donoghue <bodonoghue@codehermit.ie>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
17 years agoUSB: use __u32 rather than u32 in userspace ioctls in usbdevice_fs.h
Mike Frysinger [Tue, 6 Feb 2007 00:41:03 +0000 (16:41 -0800)] 
USB: use __u32 rather than u32 in userspace ioctls in usbdevice_fs.h

Use __u32 rather than u32 in userspace ioctl defines.

Signed-off-by: Mike Frysinger <vapier@gentoo.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
17 years agoUSB Storage: US_FL_IGNORE_RESIDUE needed for Aiptek MP3 Player
Dylan Taft [Tue, 6 Feb 2007 00:41:01 +0000 (16:41 -0800)] 
USB Storage: US_FL_IGNORE_RESIDUE needed for Aiptek MP3 Player

Device will not work as a mass storage device without US_FL_IGNORE_RESIDUE.

I bought this mp3 player that takes SD cards here

http://www.aiptek.com/Merchant2/merchant.mvc?Screen=PROD&Product_Code=AX4&Category_Code=MP3&Store_Code=AS

I can provide the errors in dmesg, if necessary, but this flag was
determined as necessary by doing a quick google on the errors that were
shown in dmesg.

Signed-off-by: Dylan Taft <d13f00l@gmail.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
17 years agoUSB: Fix misspelled "USBNET_MII" kernel config option.
Robert P. J. Day [Tue, 6 Feb 2007 00:41:02 +0000 (16:41 -0800)] 
USB: Fix misspelled "USBNET_MII" kernel config option.

Fix the misspelling of "USBNET_MII" to "USB_USBNET_MII".

Signed-off-by: Robert P. J. Day <rpjday@mindspring.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
17 years agoUSB: input: HID: add CIDC USB device to HID blacklist
Zheng XiaoJun [Tue, 6 Feb 2007 00:40:57 +0000 (16:40 -0800)] 
USB: input: HID: add CIDC USB device to HID blacklist

Add CIDC USB device to HID blacklist since it is actually a USB token
and has its own driver.

Signed-off-by: Zheng XiaoJun <zhengxiaojun@hsic.com.cn>
Cc: Dmitry Torokhov <dtor@mail.ru>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
17 years agoUSB: USB_RTL8150 must select MII
Adrian Bunk [Tue, 6 Feb 2007 00:40:57 +0000 (16:40 -0800)] 
USB: USB_RTL8150 must select MII

USB_RTL8150 must select MII to avoid link errors.

Stolen from a patch by Randy Dunlap.

Signed-off-by: Adrian Bunk <bunk@stusta.de>
Cc: "Randy.Dunlap" <rdunlap@xenotime.net>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
17 years agoUSB Storage: indistinguishable devices with broken and unbroken firmware
Oliver Neukum [Thu, 8 Feb 2007 08:04:48 +0000 (09:04 +0100)] 
USB Storage: indistinguishable devices with broken and unbroken firmware

there's a USB mass storage device which exists in two version. One
reports the correct size and the other does not. Apart from that they
are identical and cannot be told apart. Here's a heuristic based on the
empirical finding that drives have even sizes.

Signed-off-by: Oliver Neukum <oneukum@suse.de>
Acked-by: Alan Stern <stern@rowland.harvard.edu>
Acked-by: Matthew Dharm <mdharm-usb@one-eyed-alien.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
17 years agoUSB: fix autosuspend race in skeleton driver
Oliver Neukum [Thu, 8 Feb 2007 14:42:53 +0000 (15:42 +0100)] 
USB: fix autosuspend race in skeleton driver

as the skeleton driver was made ready for autosuspend a race condition
was introduced. The reference to get device must be gotten before the
autosuspend counter is upped, as this operation may sleep, dropping BKL.
Dropping BKL means that the pointer to the device may become invalid.
Here's the fix.

Signed-off-by: Oliver Neukum <oneukum@suse.de>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
17 years agoUSB: asix - Fix endian issues in asix_tx_fixup()
David Hollis [Mon, 5 Feb 2007 17:03:03 +0000 (12:03 -0500)] 
USB: asix - Fix endian issues in asix_tx_fixup()

The attached patch fixes endian issues in asix_tx_fixup() that prevented
AX88772 and AX88178 devices from working on big-endian systems.  With
the attached patch, all three chips are reported to work on big endian.

Signed-off-by: David Hollis <dhollis@davehollis.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
17 years agoUSB: fix concurrent buffer access in the hub driver
Alan Stern [Mon, 5 Feb 2007 14:56:15 +0000 (09:56 -0500)] 
USB: fix concurrent buffer access in the hub driver

This patch (as849) fixes a bug in the USB hub driver.  A single
pre-allocated buffer is used for all port status reads, but nothing
guarantees exclusive use of the buffer.  A mutex is added to provide
this guarantee.

Signed-off-by: Alan Stern <stern@rowland.harvard.edu>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
17 years agoUSB: Fix error cleanup path in airprime
Pete Zaitcev [Sat, 3 Feb 2007 07:13:14 +0000 (23:13 -0800)] 
USB: Fix error cleanup path in airprime

Fix up the error processing path: in usb_submit_urb failed, we forgot
to free buffers. Also, don't free buffers in read callback: less error
prone, 21 LOC less, no need to comment so much. N.B. write path is ok
to do kfree.

Signed-off-by: Pete Zaitcev <zaitcev@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
17 years agoUSB: descriptor structures have to be packed
Inaky Perez-Gonzalez [Sat, 3 Feb 2007 01:32:24 +0000 (17:32 -0800)] 
USB: descriptor structures have to be packed

usb: descriptor structures have to be packed

Many of the Wireless USB decriptors added to usb_ch9.h don't have the
__attribute__((packed)) tag, and thus, they don't reflect the wire
size. This patch fixes that.

Signed-off-by: Inaky Perez-Gonzalez <inaky@linux.intel.com>
Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
17 years agoEHCI: add debugging message to ehci_bus_suspend
Alan Stern [Thu, 1 Feb 2007 21:09:59 +0000 (16:09 -0500)] 
EHCI: add debugging message to ehci_bus_suspend

This patch (as848) adds a useful little debugging message to let us
know when ehci-hcd's bus_suspend method runs.  The other HCDs have
similar messages; now ehci-hcd doesn't need to feel left out.

Signed-off-by: Alan Stern <stern@rowland.harvard.edu>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
17 years agousbcore: small changes to hub driver's suspend method
Alan Stern [Thu, 1 Feb 2007 21:08:41 +0000 (16:08 -0500)] 
usbcore: small changes to hub driver's suspend method

This patch (as847) makes some small changes to the hub driver's
suspend method:

For root hubs, the status URB should be unlinked and other
activity stopped _before_ the bus_suspend method is called.

The test for hdev->bus being NULL has been removed, since
it can never succeed.

Signed-off-by: Alan Stern <stern@rowland.harvard.edu>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
17 years agoUSB: Fix apparent typo CONFIG_USB_CDCETHER.
Robert P. J. Day [Fri, 26 Jan 2007 00:34:52 +0000 (19:34 -0500)] 
USB: Fix apparent typo CONFIG_USB_CDCETHER.

Replace the apparent typo CONFIG_USB_CDCETHER with
CONFIG_USB_NET_CDCETHER.

Signed-off-by: Robert P. J. Day <rpjday@mindspring.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
17 years agoUSB: add flow control to usb-serial generic driver.
Joris van Rantwijk [Thu, 1 Feb 2007 19:08:18 +0000 (20:08 +0100)] 
USB: add flow control to usb-serial generic driver.

I added two fields to struct usb_serial_port to keep track of the
throttle state. Other usb-serial drivers typically use private data for
such things, but the generic driver can not really do that because some
of its code is also used by other drivers (which may have their own
private data needs).

As it is, I am not sure that this patch is useful in all scenarios.
It is certainly helpful for low-bandwidth devices that can hold their
data in response to throttling. But for devices that pump data in
real-time as fast as possible (webcam, A/D converter, etc), throttling
may actually cause more data loss.

From: Joris van Rantwijk <jorispubl@xs4all.nl>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
17 years agoUSB: unusual_devs update for Sony P990i phone
Alan Stern [Wed, 31 Jan 2007 15:57:55 +0000 (10:57 -0500)] 
USB: unusual_devs update for Sony P990i phone

This patch (as846) adds the IGNORE_RESIDUE flag to the unusual_devs
entry for Sony-Ericsson's P990i phone.

Signed-off-by: Alan Stern <stern@rowland.harvard.edu>
Signed-off-by: Phil Dibowitz <phil@ipom.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
17 years agoUSB: Teac HD-35PU patch to unusual_devs.h
Thomas Bächler [Fri, 26 Jan 2007 23:46:58 +0000 (00:46 +0100)] 
USB: Teac HD-35PU patch to unusual_devs.h

Hi, one of my users has two USB hard drives that need the following
patch, otherwise there are I/O errors similar to those here:
http://bugme.osdl.org/show_bug.cgi?id=3223

Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
17 years agoUSB: PL2303: Willcom WS002IN Support.
YOSHIFUJI Hideaki [Fri, 26 Jan 2007 13:51:38 +0000 (22:51 +0900)] 
USB: PL2303: Willcom WS002IN Support.

Signed-off-by: YOSHIFUJI Hideaki <yoshfuji@linux-ipv6.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
17 years agoUSB: fix needless failure under certain conditions
Oliver Neukum [Thu, 25 Jan 2007 10:22:24 +0000 (11:22 +0100)] 
USB: fix needless failure under certain conditions

in devices.c we have a piece of code for dealing with losing in a race.
If we indeed lose the race we don't care whether our own memory allocation
worked. The check for that is so early that we return early even if we
don't have to.

Signed-off-by: Oliver Neukum <oneukum@suse.de>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
17 years agoUSB: In init_endpoint_class, use PTR_ERR to obtain an errno value, not IS_ERR
Josh Triplett [Thu, 25 Jan 2007 09:32:34 +0000 (01:32 -0800)] 
USB: In init_endpoint_class, use PTR_ERR to obtain an errno value, not IS_ERR

init_endpoint_class calls class_create, and checks the result for an error
with IS_ERR; however, if true, it then returns the result of IS_ERR (a
boolean) rather than PTR_ERR (the actual errno).

Signed-off-by: Josh Triplett <josh@kernel.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
17 years agoUSB: Driver to charge USB blackberry devices
Greg Kroah-Hartman [Sat, 13 Jan 2007 18:57:42 +0000 (10:57 -0800)] 
USB: Driver to charge USB blackberry devices

A simple driver to turn on the charging capability of a USB BlackBerry
device when it is plugged into the machine.  It does not bind to the
device, so all userspace programs can still sync properly with it.

Note, if CONFIG_USB_SUSPEND is enabled, it can play havoc with this
device as the power to the port will be shut down.  This device id will
have to be added to the global blacklist table when it is created.

Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
17 years agoEHCI: turn off remote wakeup during shutdown
Alan Stern [Tue, 13 Feb 2007 19:55:27 +0000 (14:55 -0500)] 
EHCI: turn off remote wakeup during shutdown

This patch (as850b) disables remote wakeup (and everything else!) on
all EHCI ports when the shutdown() method is called.  If remote wakeup
is left active then some systems will reboot instead of powering off.
This fixes Bugzilla #7828.

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>
17 years agoIB/core: Set static rate in ib_init_ah_from_path()
Roland Dreier [Fri, 16 Feb 2007 23:31:24 +0000 (15:31 -0800)] 
IB/core: Set static rate in ib_init_ah_from_path()

The static rate from the path record should be put into the address
vector -- a long time ago the rate in the address attributes needed to
be a relative rate, which required more munging, but now that the
conversion from absolute to relative is done in the low-level driver,
it's easy for ib_init_ah_from_path() to put the absolute rate in.

Cc: Jason Gunthorpe <jgunthorpe@obsidianresearch.com>
Cc: Sean Hefty <sean.hefty@intel.com>
Signed-off-by: Roland Dreier <rolandd@cisco.com>
17 years agoPCI: Make PCI device numa-node attribute visible in sysfs
Brice Goglin [Sun, 28 Jan 2007 09:53:40 +0000 (10:53 +0100)] 
PCI: Make PCI device numa-node attribute visible in sysfs

Export the numa-node attribute of PCI devices in sysfs so that
user applications may choose where to be placed accordingly.

Signed-off-by: Brice Goglin <Brice.Goglin@ens-lyon.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
17 years agoPCI: add systems for automatic breadth-first device sorting
Andy Gospodarek [Tue, 6 Feb 2007 00:36:10 +0000 (16:36 -0800)] 
PCI: add systems for automatic breadth-first device sorting

This is an additional list of systems that exhibit the PCI device ordering
issue that prompted the following patch:

commit 6b4b78fed47e7380dfe9280b154e8b9bfcd4c86c
Author: Matt Domsch <Matt_Domsch@dell.com>
Date:   Fri Sep 29 15:23:23 2006 -0500

    PCI: optionally sort device lists breadth-first

Adding these systems to the list prevents the need for the additional
kernel command line argument.

Signed-off-by: Andy Gospodarek <andy@greyhouse.net>
Cc: Matt Domsch <Matt_Domsch@dell.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
17 years agoPCI: PCI devices get assigned redundant IRQs
Andreas Block [Tue, 6 Feb 2007 00:36:07 +0000 (16:36 -0800)] 
PCI: PCI devices get assigned redundant IRQs

I'm currently working on a port to a CPCI board with a MPC5200.  When
testing the PCI interrupt routing, I discovered the following: Even devices
which don't use interrupts (-> PCI Spec.: Interrupt Pin Register is zero),
get an interrupt assigned (this is at least true for most of the
PPC-targets I looked at).

The cause is pretty obvious in drivers/pci/setup-irq.c.  I guess at least
in an ideal world with correctly designed hardware, the code should rather
look as in the patch below.

Of course it doesn't hurt anybody to have an unuseable IRQ assigned to a
PCI-to-PCI-bridge (or something alike), but to me it seems a bit strange.
Please correct me, if I'm mislead.

The patch below is tested on the above mentioned CPCI-MPC5200 board and is
compiler tested with the latest git-repository kernel on x86.

Cc: "Eric W. Biederman" <ebiederm@xmission.com>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: Linus Torvalds <torvalds@osdl.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
17 years agoPCI: Make CARDBUS_MEM_SIZE and CARDBUS_IO_SIZE boot options
Atsushi Nemoto [Tue, 6 Feb 2007 00:36:06 +0000 (16:36 -0800)] 
PCI: Make CARDBUS_MEM_SIZE and CARDBUS_IO_SIZE boot options

CARDBUS_MEM_SIZE was increased to 64MB on 2.6.20-rc2, but larger size might
result in allocation failure for the reserving itself on some platforms
(for example typical 32bit MIPS).  Make it (and CARDBUS_IO_SIZE too)
customizable by "pci=" option for such platforms.

Signed-off-by: Atsushi Nemoto <anemo@mba.ocn.ne.jp>
Cc: Daniel Ritz <daniel.ritz@gmx.ch>
Cc: Ralf Baechle <ralf@linux-mips.org>
Cc: Ivan Kokshaysky <ink@jurassic.park.msu.ru>
Cc: Dominik Brodowski <linux@dominikbrodowski.net>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
17 years agoPCI: pci.txt fix __devexit() usage
Grant Grundler [Sun, 11 Feb 2007 07:04:04 +0000 (00:04 -0700)] 
PCI: pci.txt fix __devexit() usage

Marin Mitov <mitov@issp.bas.bg> spotted a brainfart where I had
failed to update copied text with *_remove and __devexit().

Marin made a good comment in his email to me:
| mydriver_probe() is _always_ executed, while mydriver_remove() is not.
| See: include/linux/init.h

Which says:
/* Functions marked as __devexit may be discarded at kernel link time, depending
   on config options.  Newer versions of binutils detect references from
   retained sections to discarded sections and flag an error.  Pointers to
   __devexit functions must use __devexit_p(function_name), the wrapper will
   insert either the function_name or NULL, depending on the config options.
 */

Signed-off-by: Grant Grundler <grundler@parisc-linux.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
17 years agoPCI/sysfs/kobject kernel-doc fixes
Randy Dunlap [Sat, 10 Feb 2007 22:41:56 +0000 (14:41 -0800)] 
PCI/sysfs/kobject kernel-doc fixes

Fix kernel-doc warnings in PCI, sysfs, and kobject files.

Signed-off-by: Randy Dunlap <randy.dunlap@oracle.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
17 years agosysfs: fix build errors: uevent with CONFIG_SYSFS=n
Randy Dunlap [Tue, 13 Feb 2007 23:19:06 +0000 (15:19 -0800)] 
sysfs: fix build errors: uevent with CONFIG_SYSFS=n

Fix source files to build with CONFIG_SYSFS=n.
module_subsys is not available.

SYSFS=n, MODULES=y: T:y
SYSFS=n, MODULES=n: T:y

SYSFS=y, MODULES=y: T:y
SYSFS=y, MODULES=n: T:y

Signed-off-by: Randy Dunlap <randy.dunlap@oracle.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
17 years agopcmcia: some class_device fallout
Alexey Dobriyan [Wed, 14 Feb 2007 06:39:27 +0000 (22:39 -0800)] 
pcmcia: some class_device fallout

As found on some arm defconfigs.

I only looked at how original patch changes things and other patches fix
compilation. ;-)

Signed-off-by: Alexey Dobriyan <adobriyan@gmail.com>
Cc: Dominik Brodowski <linux@dominikbrodowski.net>
Cc: Russell King <rmk@arm.linux.org.uk>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
17 years agoDriver core: device_add_attrs() cleanup
Andrew Morton [Wed, 14 Feb 2007 06:39:26 +0000 (22:39 -0800)] 
Driver core: device_add_attrs() cleanup

Clean up the coding in device_add_attrs() a bit.

Cc: Jean Delvare <khali@linux-fr.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
17 years agodebugfs: Remove misleading comments.
Cornelia Huck [Wed, 14 Feb 2007 06:57:47 +0000 (07:57 +0100)] 
debugfs: Remove misleading comments.

Just mention which error will be returned if debugfs is disabled. Callers
should be able to figure out themselves what they need to check.

Signed-off-by: Cornelia Huck <cornelia.huck@de.ibm.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
17 years agodebugfs: implement symbolic links
Peter Oberparleiter [Tue, 13 Feb 2007 11:13:54 +0000 (12:13 +0100)] 
debugfs: implement symbolic links

debugfs: implement symbolic links

Implement a new function debugfs_create_symlink() which can be used
to create symbolic links in debugfs. This function can be useful
for people moving functionality from /proc to debugfs (e.g. the
gcov-kernel patch).

Signed-off-by: Peter Oberparleiter <peter.oberparleiter@de.ibm.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
17 years agoDriver: remove redundant kobject_unregister checks
Mariusz Kozlowski [Wed, 14 Feb 2007 20:03:39 +0000 (21:03 +0100)] 
Driver: remove redundant kobject_unregister checks

Here is a patch that removes all redundant kobject_unregister argument checks.

Signed-off-by: Mariusz Kozlowski <m.kozlowski@tuxland.pl>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
17 years agokobject: kobj->k_name verification fix
Martin Stoilov [Tue, 6 Feb 2007 00:15:23 +0000 (16:15 -0800)] 
kobject: kobj->k_name verification fix

The function 'kobject_add' tries to verify the name of
a new kobject instance is properly set before continuing.
    if (!kobj->k_name)
        kobj->k_name = kobj->name;
    if (!kobj->k_name) {
        pr_debug("kobject attempted to be registered with no name!\n");
        WARN_ON(1);
        return -EINVAL;
    }
The statement:
    if (!kobj->k_name) {
        pr_debug("kobject attempted to be registered with no name!\n");
        WARN_ON(1);
        return -EINVAL;
    }
is useless the way it is right now, because it can never be true. I
think the
code was intended to be:
    if (!kobj->k_name)
        kobj->k_name = kobj->name;
    if (!*kobj->k_name) {
        pr_debug("kobject attempted to be registered with no name!\n");
        WARN_ON(1);
        return -EINVAL;
    }
because this would make sure the kobj->name buffer has something in it.
So the missing '*' is just a typo. Although, I would much prefer
expression like:
    if (*kobj->k_name == '\0') {
        pr_debug("kobject attempted to be registered with no name!\n");
        WARN_ON(1);
        return -EINVAL;
    }

because this would've made the intention clear, in this patch I just restore
the missing '*' without changing the coding style of the function.

Signed-off-by: Martin Stoilov <mstoilov@odesys.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
17 years agoserial: Add PCMCIA IDs for Quatech DSP-100 dual RS232 adapter.
Sergei Organov [Fri, 2 Feb 2007 17:43:12 +0000 (20:43 +0300)] 
serial: Add PCMCIA IDs for Quatech DSP-100 dual RS232 adapter.

Add PCMCIA IDs for Quatech DSP-100 dual RS232 adapter.

Signed-off-by: Sergei Organov <osv@javad.com>
Acked-by: Alan Cox <alan@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
17 years agoDriver core: let request_module() send a /sys/modules/kmod/-uevent
Kay Sievers [Fri, 2 Feb 2007 15:39:12 +0000 (16:39 +0100)] 
Driver core: let request_module() send a /sys/modules/kmod/-uevent

On recent systems, calls to /sbin/modprobe are handled by udev depending
on the kind of device the kernel has discovered. This patch creates an
uevent for the kernels internal request_module(), to let udev take control
over the request, instead of forking the binary directly by the kernel.
The direct execution of /sbin/modprobe can be disabled by setting:
  /sys/module/kmod/mod_request_helper (/proc/sys/kernel/modprobe)
to an empty string, the same way /proc/sys/kernel/hotplug is disabled on an
udev system.

Signed-off-by: Kay Sievers <kay.sievers@vrfy.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
17 years agoDriver.h copyright update
Greg Kroah-Hartman [Tue, 13 Feb 2007 06:33:06 +0000 (22:33 -0800)] 
Driver.h copyright update

It was pointed out that I had not updated my copyright on driver.h

Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
17 years agoIB/ipath: Make ipath_map_sg() static
Roland Dreier [Fri, 16 Feb 2007 22:58:08 +0000 (14:58 -0800)] 
IB/ipath: Make ipath_map_sg() static

Signed-off-by: Roland Dreier <rolandd@cisco.com>
17 years agoIB/core: Fix sparse warnings about shadowed declarations
Roland Dreier [Fri, 16 Feb 2007 22:41:14 +0000 (14:41 -0800)] 
IB/core: Fix sparse warnings about shadowed declarations

Change a couple of variable names to avoid sparse warnings about
symbols being shadowed.

Signed-off-by: Roland Dreier <rolandd@cisco.com>
17 years agoRDMA/cma: Add multicast communication support
Sean Hefty [Fri, 16 Feb 2007 01:00:18 +0000 (17:00 -0800)] 
RDMA/cma: Add multicast communication support

Extend rdma_cm to support multicast communication.  Multicast support
is added to the existing RDMA_PS_UDP port space, as well as a new
RDMA_PS_IPOIB port space.  The latter port space allows joining the
multicast groups used by IPoIB, which enables offloading IPoIB traffic
to a separate QP.  The port space determines the signature used in the
MGID when joining the group.  The newly added RDMA_PS_IPOIB also
allows for unicast operations, similar to RDMA_PS_UDP.

Supporting the RDMA_PS_IPOIB requires changing how UD QPs are initialized,
since we can no longer assume that the qkey is constant.  This requires
saving the Q_Key to use when attaching to a device, so that it is
available when creating the QP.  The Q_Key information is exported to
the user through the existing rdma_init_qp_attr() interface.

Multicast support is also exported to userspace through the rdma_ucm.

Signed-off-by: Roland Dreier <rolandd@cisco.com>
17 years agoIB/sa: Track multicast join/leave requests
Sean Hefty [Fri, 16 Feb 2007 01:00:17 +0000 (17:00 -0800)] 
IB/sa: Track multicast join/leave requests

The IB SA tracks multicast join/leave requests on a per port basis and
does not do any reference counting: if two users of the same port join
the same group, and one leaves that group, then the SA will remove the
port from the group even though there is one user who wants to stay a
member left.  Therefore, in order to support multiple users of the
same multicast group from the same port, we need to perform reference
counting locally.

To do this, add an multicast submodule to ib_sa to perform reference
counting of multicast join/leave operations.  Modify ib_ipoib (the
only in-kernel user of multicast) to use the new interface.

Signed-off-by: Roland Dreier <rolandd@cisco.com>
17 years agoIPoIB: CM error handling thinko fix
Michael S. Tsirkin [Thu, 15 Feb 2007 22:16:13 +0000 (00:16 +0200)] 
IPoIB: CM error handling thinko fix

ipoib_cm_alloc_rx_skb() might be called from IRQ context, so it must
use dev_kfree_skb_any(), not kfree_skb().

Signed-off-by: Michael S. Tsirkin <mst@mellanox.co.il>
Signed-off-by: Roland Dreier <rolandd@cisco.com>
17 years agoRDMA/cxgb3: Remove Open Grid Computing copyrights in iw_cxgb3 driver
Steve Wise [Thu, 15 Feb 2007 19:54:22 +0000 (13:54 -0600)] 
RDMA/cxgb3: Remove Open Grid Computing copyrights in iw_cxgb3 driver

Remove the Open Grid Computing copyright.  It shouldn't be there.

Signed-off-by: Steve Wise <swise@opengridcomputing.com>
Signed-off-by: Roland Dreier <rolandd@cisco.com>
17 years agoRDMA/cxgb3: Fail posts synchronously when in TERMINATE state
Steve Wise [Thu, 15 Feb 2007 14:49:02 +0000 (08:49 -0600)] 
RDMA/cxgb3: Fail posts synchronously when in TERMINATE state

For T3B devices, mark user QP in error once we transition
to TERMINATE.

Signed-off-by: Steve Wise <swise@opengridcomputing.com>
Signed-off-by: Roland Dreier <rolandd@cisco.com>
17 years agoRDMA/iwcm: iw_cm_id destruction race fixes
Steve Wise [Thu, 15 Feb 2007 14:09:36 +0000 (08:09 -0600)] 
RDMA/iwcm: iw_cm_id destruction race fixes

iwcm iw_cm_id destruction race condition fixes:

- iwcm_deref_id() always wakes up if there's another reference.
- clean up race condition in cm_work_handler().
- create static void free_cm_id() which deallocs the work entries and then
  kfrees the cm_id memory.  This reduces code replication.
- rem_ref() if this is the last reference -and- the IWCM owns freeing the
  cm_id, then free it.

Signed-off-by: Steve Wise <swise@opengridcomputing.com>
Signed-off-by: Tom Tucker <tom@opengridcomputing.com>
Acked-by: Krishna Kumar <krkumar2@in.ibm.com>
Signed-off-by: Roland Dreier <rolandd@cisco.com>
17 years agoIB/ehca: Change query_port() to return LINK_UP instead UNKNOWN
Hoang-Nam Nguyen [Thu, 15 Feb 2007 16:10:06 +0000 (17:10 +0100)] 
IB/ehca: Change query_port() to return LINK_UP instead UNKNOWN

Set the port phys state as returned from ehca_query_port() to LINK_UP.
ehca actually represents a logical HCA, whose phys/link state always
is LINK_UP.

Signed-off-by: Hoang-Nam Nguyen <hnguyen@de.ibm.com>
Signed-off-by: Roland Dreier <rolandd@cisco.com>
17 years agoIB/ehca: Allow en/disabling scaling code via module parameter
Hoang-Nam Nguyen [Thu, 15 Feb 2007 16:08:33 +0000 (17:08 +0100)] 
IB/ehca: Allow en/disabling scaling code via module parameter

Allow users to en/disable scaling code when loading ib_ehca module,
rather than requiring the module to be rebuilt to change the setting.

Signed-off-by: Hoang-Nam Nguyen <hnguyen@de.ibm.com>
Signed-off-by: Roland Dreier <rolandd@cisco.com>
17 years agoIB/ehca: Fix race condition/locking issues in scaling code
Hoang-Nam Nguyen [Thu, 15 Feb 2007 16:07:30 +0000 (17:07 +0100)] 
IB/ehca: Fix race condition/locking issues in scaling code

Fix a race condition in find_next_cpu_online() and some other locking
issues in ehca scaling code.

Signed-off-by: Hoang-Nam Nguyen <hnguyen@de.ibm.com>
Signed-off-by: Roland Dreier <rolandd@cisco.com>
17 years agoIB/ehca: Rework irq handler
Hoang-Nam Nguyen [Thu, 15 Feb 2007 16:06:33 +0000 (17:06 +0100)] 
IB/ehca: Rework irq handler

Rework ehca interrupt handling to avoid/reduce missed irq events.

Signed-off-by: Hoang-Nam Nguyen <hnguyen@de.ibm.com>
Signed-off-by: Roland Dreier <rolandd@cisco.com>
17 years agoIPoIB: Only allow root to change between datagram and connected mode
Roland Dreier [Fri, 16 Feb 2007 21:57:33 +0000 (13:57 -0800)] 
IPoIB: Only allow root to change between datagram and connected mode

Change the permissions of the "mode" sysfs attribute to be S_IWUSR
instead of S_IWUGO.

Signed-off-by: Roland Dreier <rolandd@cisco.com>
17 years agoIB/mthca: Fix allocation of ICM chunks in coherent memory
Roland Dreier [Fri, 16 Feb 2007 21:57:33 +0000 (13:57 -0800)] 
IB/mthca: Fix allocation of ICM chunks in coherent memory

The change to allow allocating ICM chunks from coherent memory did not
increment the count of sg entries properly, so a chunk that required
more than allocation would not be mapped properly by the HCA.

Fix this by adding the missing increment of chunk->nsg.

Signed-off-by: Roland Dreier <rolandd@cisco.com>
17 years agoIB/mthca: Allow the QP state transition RESET->RESET
Dotan Barak [Tue, 9 Jan 2007 09:53:07 +0000 (11:53 +0200)] 
IB/mthca: Allow the QP state transition RESET->RESET

RESET->RESET is an allowed QP state transition, so mthca should handle
it correctly, by just returning success without involving the firmware.

Signed-off-by: Dotan Barak <dotanb@mellanox.co.il>
Signed-off-by: Roland Dreier <rolandd@cisco.com>
17 years agoACPI: ec: add unlock in error path
Alexey Starikovskiy [Thu, 15 Feb 2007 20:16:18 +0000 (23:16 +0300)] 
ACPI: ec: add unlock in error path

Signed-off-by: Alexey Starikovskiy <alexey.y.starikovskiy@linux.intel.com>
Signed-off-by: Len Brown <len.brown@intel.com>
17 years agoMerge branch 'for-linus' of git://www.atmel.no/~hskinnemoen/linux/kernel/avr32
Linus Torvalds [Fri, 16 Feb 2007 16:19:44 +0000 (08:19 -0800)] 
Merge branch 'for-linus' of git://atmel.no/~hskinnemoen/linux/kernel/avr32

* 'for-linus' of git://www.atmel.no/~hskinnemoen/linux/kernel/avr32:
  [AVR32] Use per-controller spi_board_info structures
  [AVR32] Warn, don't BUG if clk_disable is called too many times
  [AVR32] Make sure all genclocks have a parent
  [AVR32] Remove unnecessary sys_nfsservctl conditional
  [AVR32] Wire up the SysV IPC calls properly
  [AVR32] Define ioremap_nocache, ioport_map and ioport_unmap
  [AVR32] Fix prototypes for __raw_writesb and friends

17 years agoMerge master.kernel.org:/pub/scm/linux/kernel/git/davej/agpgart
Linus Torvalds [Fri, 16 Feb 2007 16:17:04 +0000 (08:17 -0800)] 
Merge /pub/scm/linux/kernel/git/davej/agpgart

* master.kernel.org:/pub/scm/linux/kernel/git/davej/agpgart:
  [AGPGART] allow drm populated agp memory types cleanups
  [AGPGART] intel-agp: Use ARRAY_SIZE macro when appropriate
  [AGPGART] Add agp-type-to-mask-type method missing from some drivers.
  [AGPGART] Don't try to remap i810 registers on resume.
  [AGPGART] Allow drm-populated agp memory types
  [AGPGART] compat ioctl

17 years agoMerge master.kernel.org:/pub/scm/linux/kernel/git/davej/cpufreq
Linus Torvalds [Fri, 16 Feb 2007 16:16:01 +0000 (08:16 -0800)] 
Merge /pub/scm/linux/kernel/git/davej/cpufreq

* master.kernel.org:/pub/scm/linux/kernel/git/davej/cpufreq:
  [CPUFREQ] Longhaul - Redo Longhaul ver. 2
  [CPUFREQ] EPS - Correct 2nd brand test
  [CPUFREQ] Longhaul - Separate frequency and voltage transition
  [CPUFREQ] Longhaul - Models of Nehemiah
  [CPUFREQ] Whitespace fixup
  [CPUFREQ] Longhaul - Simplier minmult
  [CPUFREQ] CPU_FREQ_TABLE shouldn't be a def_tristate
  [CPUFREQ] ondemand governor use new cpufreq rwsem locking in work callback
  [CPUFREQ] ondemand governor restructure the work callback
  [CPUFREQ] Rewrite lock in cpufreq to eliminate cpufreq/hotplug related issues
  [CPUFREQ] Remove hotplug cpu crap
  [CPUFREQ] Enhanced PowerSaver driver
  [CPUFREQ] Longhaul - Add VT8235 support
  [CPUFREQ] Longhaul - Fix guess_fsb function
  [CPUFREQ] Longhaul - Remove duplicate tables
  [CPUFREQ] Longhaul - Introduce Nehemiah C
  [CPUFREQ] fix cpuinfo_cur_freq for CPU_HW_PSTATE
  [CPUFREQ] Longhaul - Remove "ignore_latency" option

17 years ago[PATCH] s3c2410fb: fix un-initialised dev field
Ben Dooks [Fri, 16 Feb 2007 09:28:42 +0000 (01:28 -0800)] 
[PATCH] s3c2410fb: fix un-initialised dev field

The current driver is not setting the dev field in the private data
structure, which can lead to an OOPS if the driver tries to report an
error.

Signed-off-by: Ben Dooks <ben-linux@fluff.org>
Cc: James Simmons <jsimmons@infradead.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
17 years ago[PATCH] ecryptfs: fix forgotten format specifier
Thomas Hisch [Fri, 16 Feb 2007 09:28:41 +0000 (01:28 -0800)] 
[PATCH] ecryptfs: fix forgotten format specifier

Add format specifier %d for uid in ecryptfs_printk

Signed-off-by: Thomas Hisch <t.hisch@gmail.com>
Cc: Michael Halcrow <mhalcrow@us.ibm.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
17 years ago[PATCH] eCryptfs: Reduce stack usage in ecryptfs_generate_key_packet_set()
Michael Halcrow [Fri, 16 Feb 2007 09:28:40 +0000 (01:28 -0800)] 
[PATCH] eCryptfs: Reduce stack usage in ecryptfs_generate_key_packet_set()

eCryptfs is gobbling a lot of stack in ecryptfs_generate_key_packet_set()
because it allocates a temporary memory-hungry ecryptfs_key_record struct.
This patch introduces a new kmem_cache for that struct and converts
ecryptfs_generate_key_packet_set() to use it.

Signed-off-by: Michael Halcrow <mhalcrow@us.ibm.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
17 years ago[PATCH] knfsd: stop NFSD writes from being broken into lots of little writes to files...
NeilBrown [Fri, 16 Feb 2007 09:28:38 +0000 (01:28 -0800)] 
[PATCH] knfsd: stop NFSD writes from being broken into lots of little writes to filesystem

When NFSD receives a write request, the data is typically in a number of
1448 byte segments and writev is used to collect them together.

Unfortunately, generic_file_buffered_write passes these to the filesystem
one at a time, so an e.g.  32K over-write becomes a series of partial-page
writes to each page, causing the filesystem to have to pre-read those pages
- wasted effort.

generic_file_buffered_write handles one segment of the vector at a time as
it has to pre-fault in each segment to avoid deadlocks.  When writing from
kernel-space (and nfsd does) this is not an issue, so
generic_file_buffered_write does not need to break and iovec from nfsd into
little pieces.

This patch avoids the splitting when  get_fs is KERNEL_DS as it is
from NFSd.

This issue was introduced by commit 6527c2bdf1f833cc18e8f42bd97973d583e4aa83

Acked-by: Nick Piggin <nickpiggin@yahoo.com.au>
Cc: Norman Weathers <norman.r.weathers@conocophillips.com>
Cc: Vladimir V. Saveliev <vs@namesys.com>
Signed-off-by: Neil Brown <neilb@suse.de>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
17 years ago[PATCH] knfsd: nfsd4: fix handling of directories without default ACLs
J. Bruce Fields [Fri, 16 Feb 2007 09:28:37 +0000 (01:28 -0800)] 
[PATCH] knfsd: nfsd4: fix handling of directories without default ACLs

When setting an ACL that lacks inheritable ACEs on a directory, we should set
a default ACL of zero length, not a default ACL with all bits denied.

Signed-off-by: "J. Bruce Fields" <bfields@citi.umich.edu>
Signed-off-by: Neil Brown <neilb@suse.de>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
17 years ago[PATCH] knfsd: nfsd4: acls: avoid unnecessary denies
J. Bruce Fields [Fri, 16 Feb 2007 09:28:36 +0000 (01:28 -0800)] 
[PATCH] knfsd: nfsd4: acls: avoid unnecessary denies

We're inserting deny's between some ACEs in order to enforce posix draft acl
semantics which prevent permissions from accumulating across entries in an
acl.

That's fine, but we're doing that by inserting a deny after *every* allow,
which is overkill.  We shouldn't be adding them in places where they actually
make no difference.

Also replaced some helper functions for creating acl entries; I prefer just
assigning directly to the struct fields--it takes a few more lines, but the
field names provide some documentation that I think makes the result easier
understand.

Signed-off-by: J. Bruce Fields <bfields@citi.umich.edu>
Signed-off-by: Neil Brown <neilb@suse.de>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
17 years ago[PATCH] knfsd: nfsd4: acls: don't return explicit mask
J. Bruce Fields [Fri, 16 Feb 2007 09:28:34 +0000 (01:28 -0800)] 
[PATCH] knfsd: nfsd4: acls: don't return explicit mask

Return just the effective permissions, and forget about the mask.  It isn't
worth the complexity.

WARNING: This breaks backwards compatibility with overly-picky nfsv4->posix
acl translation, as may has been included in some patched versions of libacl.
To our knowledge no such version was every distributed by anyone outside citi.

Signed-off-by: J. Bruce Fields <bfields@citi.umich.edu>
Signed-off-by: Neil Brown <neilb@suse.de>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
This page took 0.045098 seconds and 5 git commands to generate.