]> git.karo-electronics.de Git - linux-beck.git/log
linux-beck.git
10 years agoARM: shmobile: Initial r8a7791 and Koelsch multiplatform support
Magnus Damm [Wed, 6 Nov 2013 10:40:01 +0000 (19:40 +0900)]
ARM: shmobile: Initial r8a7791 and Koelsch multiplatform support

Add Koelsch and r8a7791 to CONFIG_SHMOBILE_MULTI. At this
point CCF is not yet supported so you cannot run this code
yet. For CCF support to happen several different components
are needed, and this is one simple portion that moves us
forward. Other patches need to build on top of this one.

Koelsch board support exists in 3 flavors:
1) SHMOBILE_MULTI, MACH_KOELSCH - board-koelsch-reference.c (CCF + DT)
2) SHMOBILE, MACH_KOELSCH_REFERENCE - board-koelsch-reference.c (DT)
3) SHMOBILE, MACH_KOELSCH - board-koelsch.c (legacy C code)

When CCF is done then 2) will be removed. When 1) includes same features
as 3) then 3) will be removed.

Signed-off-by: Magnus Damm <damm@opensource.se>
Signed-off-by: Simon Horman <horms+renesas@verge.net.au>
10 years agoARM: shmobile: koelsch: set .debounce_interval
Simon Horman [Tue, 29 Oct 2013 08:57:31 +0000 (17:57 +0900)]
ARM: shmobile: koelsch: set .debounce_interval

In R-Car GPIO hardware block, 'chattering removal' feature can be
enabled on GPIO-n-[3:0] pins, but it's not supported on rest of pins
GPIO-n-[31:4].

Set an appropriate debounce interval, instead.  We could confirm that
spurious/unnecessary GPIO interrupts are prevented by this settings.

Based on work for the lager board by Shinya Kuribayashi.

Cc: Shinya Kuribayashi <shinya.kuribayashi.px@renesas.com>
Signed-off-by: Simon Horman <horms+renesas@verge.net.au>
10 years agoARM: shmobile: lager: set .debounce_interval
Shinya Kuribayashi [Tue, 29 Oct 2013 08:57:30 +0000 (17:57 +0900)]
ARM: shmobile: lager: set .debounce_interval

In R-Car GPIO hardware block, 'chattering removal' feature can be
enabled on GPIO-n-[3:0] pins, but it's not supported on rest of pins
GPIO-n-[31:4].

Set an appropriate debounce interval, instead.  We could confirm that
spurious/unnecessary GPIO interrupts are prevented by this settings.

Signed-off-by: Shinya Kuribayashi <shinya.kuribayashi.px@renesas.com>
Acked-by: Magnus Damm <damm@opensource.se>
Signed-off-by: Simon Horman <horms+renesas@verge.net.au>
10 years agoARM: shmobile: bockw: add pin pull-up setting for SDHI
Kuninori Morimoto [Fri, 11 Oct 2013 06:35:34 +0000 (23:35 -0700)]
ARM: shmobile: bockw: add pin pull-up setting for SDHI

SDHI CD/WP pin needs pull-up

Signed-off-by: Kuninori Morimoto <kuninori.morimoto.gx@renesas.com>
Signed-off-by: Simon Horman <horms+renesas@verge.net.au>
10 years agoARM: shmobile: bockw: fixup FPGA ioremap area
Kuninori Morimoto [Fri, 11 Oct 2013 06:35:06 +0000 (23:35 -0700)]
ARM: shmobile: bockw: fixup FPGA ioremap area

Don't keep FPGA ioremap area.

Signed-off-by: Kuninori Morimoto <kuninori.morimoto.gx@renesas.com>
Signed-off-by: Simon Horman <horms+renesas@verge.net.au>
10 years agoARM: shmobile: r8a7791 Koelsch DT reference C bits
Magnus Damm [Tue, 8 Oct 2013 06:30:18 +0000 (15:30 +0900)]
ARM: shmobile: r8a7791 Koelsch DT reference C bits

Add DT reference support for the r8a7791 Koelsch board.

This board support file will be used together with common
clocks and multiplatform in the future.

Signed-off-by: Magnus Damm <damm@opensource.se>
Signed-off-by: Simon Horman <horms+renesas@verge.net.au>
10 years agoARM: shmobile: Add Koelsch SW2 support
Magnus Damm [Tue, 8 Oct 2013 03:40:08 +0000 (12:40 +0900)]
ARM: shmobile: Add Koelsch SW2 support

Enable Koelsch GPIO switch for GPIO input and IRQ testing purpose.

Signed-off-by: Magnus Damm <damm@opensource.se>
Signed-off-by: Simon Horman <horms+renesas@verge.net.au>
10 years agoARM: shmobile: Add Koelsch LED6, LED7 and LED8 support
Magnus Damm [Tue, 8 Oct 2013 03:39:59 +0000 (12:39 +0900)]
ARM: shmobile: Add Koelsch LED6, LED7 and LED8 support

Enable Koelsch LEDs for GPIO output testing purpose.

Signed-off-by: Magnus Damm <damm@opensource.se>
Signed-off-by: Simon Horman <horms+renesas@verge.net.au>
10 years agoARM: shmobile: Enable PFC/GPIO on the Koelsch board
Magnus Damm [Tue, 8 Oct 2013 03:39:49 +0000 (12:39 +0900)]
ARM: shmobile: Enable PFC/GPIO on the Koelsch board

Enable r8a7791 PFC and GPIO on the Koelsch board.

Signed-off-by: Magnus Damm <damm@opensource.se>
Signed-off-by: Simon Horman <horms+renesas@verge.net.au>
10 years agoARM: shmobile: bockw: use regulator for MMCIF
Kuninori Morimoto [Fri, 4 Oct 2013 01:30:55 +0000 (18:30 -0700)]
ARM: shmobile: bockw: use regulator for MMCIF

When regulators are used with MMC devices, sh_mmcif_plat_data::ocr
is not needed, they can be removed from platform data.
This patch adds v3.3 regulator settings,
and moved fixed-dummy regulator registration position

Signed-off-by: Kuninori Morimoto <kuninori.morimoto.gx@renesas.com>
Signed-off-by: Simon Horman <horms+renesas@verge.net.au>
10 years agoMerge tag 'renesas-cleanup-for-v3.14' into boards-base
Simon Horman [Tue, 10 Dec 2013 08:22:55 +0000 (17:22 +0900)]
Merge tag 'renesas-cleanup-for-v3.14' into boards-base

Renesas ARM based SoC cleanups for v3.14

* Tidy up clock table order for r7s72100, r8a7779, r8a7779, r8a7790,
  sh7372 and sh73a0 SoCs.
* Tidy up camera-rcar header for r8a7779 SoC
* Tidy up registration of VIN on r8a7779 SoC
* Tidy up PFC registration on r8a7790 SoC
* Correct typo in clocks for r8a7790 SoC
* Don't use named resources for IPMMU, I2C and TMU on sh73a0 SoC
* Don't use named resources for MMCIF and SDHI on ape6evm board
* Don't use named resources for MMCIF on lager board
* Tidy up device registration on lager board
* Tidy up headers for bockw board

10 years agoARM: shmobile: r8a7779: add HSPI clock support for DT
Simon Horman [Tue, 26 Nov 2013 07:47:10 +0000 (16:47 +0900)]
ARM: shmobile: r8a7779: add HSPI clock support for DT

Based on work for the r8a7778 SoC by Kuninori Morimoto.

Cc: Kuninori Morimoto <kuninori.morimoto.gx@renesas.com>
Signed-off-by: Simon Horman <horms+renesas@verge.net.au>
10 years agoARM: shmobile: r8a7740: add FSI clock support for DT
Kuninori Morimoto [Wed, 4 Dec 2013 01:28:27 +0000 (17:28 -0800)]
ARM: shmobile: r8a7740: add FSI clock support for DT

Signed-off-by: Kuninori Morimoto <kuninori.morimoto.gx@renesas.com>
Signed-off-by: Simon Horman <horms+renesas@verge.net.au>
10 years agoARM: shmobile: r8a7790: add SSI MSTP clocks
Kuninori Morimoto [Mon, 2 Dec 2013 02:17:18 +0000 (18:17 -0800)]
ARM: shmobile: r8a7790: add SSI MSTP clocks

Signed-off-by: Kuninori Morimoto <kuninori.morimoto.gx@renesas.com>
Signed-off-by: Simon Horman <horms+renesas@verge.net.au>
10 years agoARM: shmobile: r8a7778: add HPBIFx DMAEngine support
Kuninori Morimoto [Thu, 21 Nov 2013 07:25:32 +0000 (23:25 -0800)]
ARM: shmobile: r8a7778: add HPBIFx DMAEngine support

Signed-off-by: Kuninori Morimoto <kuninori.morimoto.gx@renesas.com>
Signed-off-by: Simon Horman <horms+renesas@verge.net.au>
10 years agoARM: shmobile: Select AUTO_ZRELADDR for EMEV2
Simon Horman [Tue, 10 Dec 2013 07:43:16 +0000 (16:43 +0900)]
ARM: shmobile: Select AUTO_ZRELADDR for EMEV2

Since ("ARM: shmobile: Remove legacy KZM9D board code")
It is now necessary for AUTO_ZRELADDR to be selected
in order for the kernel to build with kzm9d_defconfig.

Signed-off-by: Simon Horman <horms+renesas@verge.net.au>
10 years agoARM: shmobile: r8a7790: tidyup clock table order
Kuninori Morimoto [Tue, 19 Nov 2013 09:05:23 +0000 (01:05 -0800)]
ARM: shmobile: r8a7790: tidyup clock table order

SuperH lookups clock is using CLKDEV_CON/DEV/ICK_ID() macro
for a long term.
But in these days, the ICK clock is defined in random place.
This patch arranges it.

Signed-off-by: Kuninori Morimoto <kuninori.morimoto.gx@renesas.com>
Signed-off-by: Simon Horman <horms+renesas@verge.net.au>
10 years agoARM: shmobile: r8a7790: fixup I2C clock source
Kuninori Morimoto [Tue, 19 Nov 2013 09:04:35 +0000 (01:04 -0800)]
ARM: shmobile: r8a7790: fixup I2C clock source

I2C clock is based on P clock, not HP clock

Signed-off-by: Kuninori Morimoto <kuninori.morimoto.gx@renesas.com>
Signed-off-by: Simon Horman <horms+renesas@verge.net.au>
10 years agoARM: shmobile: r8a7790: care EXTAL divider settings
Kuninori Morimoto [Tue, 19 Nov 2013 09:04:20 +0000 (01:04 -0800)]
ARM: shmobile: r8a7790: care EXTAL divider settings

EXTAL clock frequency needs 1/2 when
(MD14, MD13, MD19) = (1, x, x).

Signed-off-by: Kuninori Morimoto <kuninori.morimoto.gx@renesas.com>
Signed-off-by: Simon Horman <horms+renesas@verge.net.au>
10 years agoARM: shmobile: Add r8a7791 clocks for thermal devices
Magnus Damm [Wed, 20 Nov 2013 08:00:05 +0000 (17:00 +0900)]
ARM: shmobile: Add r8a7791 clocks for thermal devices

Add the r8a7791 thermal device as legacy clocks.

Signed-off-by: Magnus Damm <damm@opensource.se>
Acked-by: Kuninori Morimoto <kuninori.morimoto.gx@renesas.com>
Signed-off-by: Simon Horman <horms+renesas@verge.net.au>
10 years agoARM: shmobile: Add r8a7791 thermal platform device
Magnus Damm [Wed, 20 Nov 2013 07:59:56 +0000 (16:59 +0900)]
ARM: shmobile: Add r8a7791 thermal platform device

Add a thermal platform device for the legacy case
on the r8a7791 SoC. This keeps the r8a7791 in sync
with the r8a7790 sister device.

Signed-off-by: Magnus Damm <damm@opensource.se>
Acked-by: Kuninori Morimoto <kuninori.morimoto.gx@renesas.com>
Signed-off-by: Simon Horman <horms+renesas@verge.net.au>
10 years agoARM: shmobile: Add r8a7790 clocks for thermal devices
Magnus Damm [Wed, 20 Nov 2013 07:59:39 +0000 (16:59 +0900)]
ARM: shmobile: Add r8a7790 clocks for thermal devices

Add the r8a7790 DT thermal device to the legacy clocks.

Signed-off-by: Magnus Damm <damm@opensource.se>
Acked-by: Kuninori Morimoto <kuninori.morimoto.gx@renesas.com>
Signed-off-by: Simon Horman <horms+renesas@verge.net.au>
10 years agoARM: Rename ARCH_SHMOBILE to ARCH_SHMOBILE_LEGACY
Laurent Pinchart [Sat, 9 Nov 2013 12:33:48 +0000 (13:33 +0100)]
ARM: Rename ARCH_SHMOBILE to ARCH_SHMOBILE_LEGACY

SH-Mobile platforms are transitioning from non-multiplatform to
multiplatform kernel. A new ARCH_SHMOBILE_MULTI configuration symbol has
been created to group all multiplatform-enabled SH-Mobile SoCs. The
existing ARCH_SHMOBILE configuration symbol groups SoCs that haven't
been converted yet.

This arrangement works fine for the arch/ code, but lots of drivers
needed on both ARCH_SHMOBILE and ARCH_SHMOBILE_MULTI depend on
ARCH_SHMOBILE only. In order to avoid changing them, rename
ARCH_SHMOBILE to ARCH_SHMOBILE_LEGACY, and create a new boolean
ARCH_SHMOBILE configuration symbol that is selected by both
ARCH_SHMOBILE_LEGACY and ARCH_SHMOBILE_MULTI.

Signed-off-by: Laurent Pinchart <laurent.pinchart+renesas@ideasonboard.com>
Acked-by: Magnus Damm <damm@opensource.se>
Signed-off-by: Simon Horman <horms+renesas@verge.net.au>
10 years agoARM: shmobile: r8a7791: Add DU and LVDS clocks
Laurent Pinchart [Wed, 13 Nov 2013 13:01:42 +0000 (14:01 +0100)]
ARM: shmobile: r8a7791: Add DU and LVDS clocks

The ZX parent clock isn't implemented yet, add it as well.

Signed-off-by: Laurent Pinchart <laurent.pinchart+renesas@ideasonboard.com>
Signed-off-by: Simon Horman <horms+renesas@verge.net.au>
10 years agoARM: shmobile: Select USE_OF on EMEV2
Magnus Damm [Fri, 8 Nov 2013 10:09:34 +0000 (19:09 +0900)]
ARM: shmobile: Select USE_OF on EMEV2

Now when the legacy KZM9D board code is gone, make sure
USE_OF is selected in case of the EMEV2 SoC.

Signed-off-by: Magnus Damm <damm@opensource.se>
Signed-off-by: Simon Horman <horms+renesas@verge.net.au>
10 years agoARM: shmobile: r8a7778: add HSPI clock support for DT
Kuninori Morimoto [Fri, 1 Nov 2013 01:22:39 +0000 (18:22 -0700)]
ARM: shmobile: r8a7778: add HSPI clock support for DT

Signed-off-by: Kuninori Morimoto <kuninori.morimoto.gx@renesas.com>
Signed-off-by: Simon Horman <horms+renesas@verge.net.au>
10 years agoARM: shmobile: Remove legacy platform devices from EMEV2 SoC code
Magnus Damm [Wed, 6 Nov 2013 23:21:38 +0000 (08:21 +0900)]
ARM: shmobile: Remove legacy platform devices from EMEV2 SoC code

Now when KZM9D legacy C board support code is
gone then remove emev2_add_standard_devices()
and all the platform devices from setup-emev2.c.

Signed-off-by: Magnus Damm <damm@opensource.se>
Signed-off-by: Simon Horman <horms+renesas@verge.net.au>
10 years agoARM: shmobile: Remove legacy KZM9D board code
Magnus Damm [Wed, 6 Nov 2013 23:21:29 +0000 (08:21 +0900)]
ARM: shmobile: Remove legacy KZM9D board code

Remove the C and platform device version of KZM9D.

The DT version of KZM9D board support can now instead
directly be used with SoC specific code in setup-emev2.c.

Signed-off-by: Magnus Damm <damm@opensource.se>
Signed-off-by: Simon Horman <horms+renesas@verge.net.au>
10 years agoARM: shmobile: Use ->init_late() in shared EMEV2 case
Magnus Damm [Wed, 6 Nov 2013 23:21:20 +0000 (08:21 +0900)]
ARM: shmobile: Use ->init_late() in shared EMEV2 case

Hook up shmobile_init_late() to enable various
code such as suspend-to-RAM and CPUIdle.

Signed-off-by: Magnus Damm <damm@opensource.se>
Signed-off-by: Simon Horman <horms+renesas@verge.net.au>
10 years agoARM: shmobile: Add shared EMEV2 code for ->init_machine()
Magnus Damm [Wed, 6 Nov 2013 23:21:11 +0000 (08:21 +0900)]
ARM: shmobile: Add shared EMEV2 code for ->init_machine()

Add a SoC specific function that initializes
clocks and starts DT probing in case of EMEV2.

This EMEV2 SoC support code may be built for
either legacy SHMOBILE or SMOBILE_MULTI.

The change allows us to support existing board
specific KZM9D DTB with these SoC specific
DT_MACHINE_START() callbacks.

Signed-off-by: Magnus Damm <damm@opensource.se>
Signed-off-by: Simon Horman <horms+renesas@verge.net.au>
10 years agoARM: shmobile: Enable MTU2 on r7s72100
Magnus Damm [Wed, 6 Nov 2013 10:43:32 +0000 (19:43 +0900)]
ARM: shmobile: Enable MTU2 on r7s72100

Add MTU2 as r7s72100 system timer.

Signed-off-by: Magnus Damm <damm@opensource.se>
Signed-off-by: Simon Horman <horms+renesas@verge.net.au>
10 years agoARM: shmobile: r8a7790: add QSPI support
Hiep Cao Minh [Tue, 22 Oct 2013 02:21:11 +0000 (11:21 +0900)]
ARM: shmobile: r8a7790: add QSPI support

Adds support for QSPI on the r8a7790.

Signed-off-by: Hiep Cao Minh <cm-hiep@jinso.co.jp>
Acked-by: Kuninori Morimoto <kuninori.morimoto.gx@renesas.com>
Signed-off-by: Simon Horman <horms+renesas@verge.net.au>
10 years agoARM: shmobile: r8a7790: Add USBHS clock support
Valentine Barshak [Wed, 9 Oct 2013 22:14:46 +0000 (02:14 +0400)]
ARM: shmobile: r8a7790: Add USBHS clock support

This adds USBHS clock support.

Signed-off-by: Valentine Barshak <valentine.barshak@cogentembedded.com>
Acked-by: Kuninori Morimoto <kuninori.morimoto.gx@renesas.com>
Signed-off-by: Simon Horman <horms+renesas@verge.net.au>
10 years agoARM: shmobile: r8a7779: add SDHI clock support for DT
Kuninori Morimoto [Fri, 11 Oct 2013 06:36:10 +0000 (23:36 -0700)]
ARM: shmobile: r8a7779: add SDHI clock support for DT

Signed-off-by: Kuninori Morimoto <kuninori.morimoto.gx@renesas.com>
Signed-off-by: Simon Horman <horms+renesas@verge.net.au>
10 years agoARM: shmobile: r8a7778: add SDHI clock support for DT
Kuninori Morimoto [Fri, 11 Oct 2013 06:35:17 +0000 (23:35 -0700)]
ARM: shmobile: r8a7778: add SDHI clock support for DT

Signed-off-by: Kuninori Morimoto <kuninori.morimoto.gx@renesas.com>
Signed-off-by: Simon Horman <horms+renesas@verge.net.au>
10 years agoARM: shmobile: r8a7778: add MMCIF clock support for DT
Kuninori Morimoto [Fri, 4 Oct 2013 01:32:10 +0000 (18:32 -0700)]
ARM: shmobile: r8a7778: add MMCIF clock support for DT

Signed-off-by: Kuninori Morimoto <kuninori.morimoto.gx@renesas.com>
Signed-off-by: Simon Horman <horms+renesas@verge.net.au>
10 years agoARM: shmobile: Select GPIO in case of the r7s72100 SoC
Magnus Damm [Wed, 9 Oct 2013 22:57:42 +0000 (07:57 +0900)]
ARM: shmobile: Select GPIO in case of the r7s72100 SoC

The r7s72100 contains GPIO controllers so make sure the GPIO
subsystem can be built by selecting ARCH_WANT_OPTIONAL_GPIOLIB.

Signed-off-by: Magnus Damm <damm@opensource.se>
Signed-off-by: Simon Horman <horms+renesas@verge.net.au>
10 years agoARM: shmobile: r8a73a4: don't use named irq for DMAEngine
Kuninori Morimoto [Tue, 8 Oct 2013 05:59:23 +0000 (22:59 -0700)]
ARM: shmobile: r8a73a4: don't use named irq for DMAEngine

sh-dma-engine driver doesn't care irq name.

Signed-off-by: Kuninori Morimoto <kuninori.morimoto.gx@renesas.com>
Acked-by: Magnus Damm <damm@opensource.se>
Signed-off-by: Simon Horman <horms+renesas@verge.net.au>
10 years agoARM: shmobile: r8a7791 GPIO platform device support
Magnus Damm [Tue, 8 Oct 2013 03:39:39 +0000 (12:39 +0900)]
ARM: shmobile: r8a7791 GPIO platform device support

Add GPIO controller platform devices for the r8a7791 SoC.

Signed-off-by: Magnus Damm <damm@opensource.se>
Signed-off-by: Simon Horman <horms+renesas@verge.net.au>
10 years agoARM: shmobile: Select GPIO in case of the r8a7791 SoC
Magnus Damm [Tue, 8 Oct 2013 03:39:20 +0000 (12:39 +0900)]
ARM: shmobile: Select GPIO in case of the r8a7791 SoC

Make it possible to build GPIO on r8a7791.

Signed-off-by: Magnus Damm <damm@opensource.se>
Signed-off-by: Simon Horman <horms+renesas@verge.net.au>
10 years agoARM: shmobile: r8a7791 PFC platform device support
Magnus Damm [Tue, 8 Oct 2013 03:39:10 +0000 (12:39 +0900)]
ARM: shmobile: r8a7791 PFC platform device support

Add a platform device for the r8a7791 PFC device.

Signed-off-by: Magnus Damm <damm@opensource.se>
Signed-off-by: Simon Horman <horms+renesas@verge.net.au>
10 years agoARM: shmobile: Select IRQC in case of the r8a7791 SoC
Magnus Damm [Tue, 8 Oct 2013 03:16:08 +0000 (12:16 +0900)]
ARM: shmobile: Select IRQC in case of the r8a7791 SoC

The r8a7791 contains IRQC hardware so make sure
the driver gets built by selecting RENESAS_IRQC.

Signed-off-by: Magnus Damm <damm@opensource.se>
Signed-off-by: Simon Horman <horms+renesas@verge.net.au>
10 years agoARM: shmobile: r8a7779: add I2C clock for DT
Kuninori Morimoto [Fri, 4 Oct 2013 06:45:03 +0000 (23:45 -0700)]
ARM: shmobile: r8a7779: add I2C clock for DT

10e8d4f6dddb0f9dc408c2f2bde8399b243a42ca
(ARM: mach-shmobile: r8a7779: Minimal setup using DT)
added I2C driver for DT, but it didn't add clock.
This patch adds missing clock for I2C

Signed-off-by: Kuninori Morimoto <kuninori.morimoto.gx@renesas.com>
Signed-off-by: Simon Horman <horms+renesas@verge.net.au>
10 years agoARM: shmobile: r8a7778: add I2C clock for DT
Kuninori Morimoto [Fri, 4 Oct 2013 06:43:54 +0000 (23:43 -0700)]
ARM: shmobile: r8a7778: add I2C clock for DT

Signed-off-by: Kuninori Morimoto <kuninori.morimoto.gx@renesas.com>
Signed-off-by: Simon Horman <horms+renesas@verge.net.au>
10 years agoARM: shmobile: sh7372: tidyup clock table order
Kuninori Morimoto [Tue, 19 Nov 2013 09:05:33 +0000 (01:05 -0800)]
ARM: shmobile: sh7372: tidyup clock table order

SuperH lookups clock is using CLKDEV_CON/DEV/ICK_ID() macro
for a long term.
But in these days, the ICK clock is defined in random place.
This patch arranges it.

Signed-off-by: Kuninori Morimoto <kuninori.morimoto.gx@renesas.com>
Signed-off-by: Simon Horman <horms+renesas@verge.net.au>
10 years agoARM: shmobile: r7s72100: tidyup clock table order
Kuninori Morimoto [Tue, 19 Nov 2013 09:05:10 +0000 (01:05 -0800)]
ARM: shmobile: r7s72100: tidyup clock table order

SuperH lookups clock is using CLKDEV_CON/DEV/ICK_ID() macro
for a long term.
But in these days, the ICK clock is defined in random place.
This patch arranges it.

Signed-off-by: Kuninori Morimoto <kuninori.morimoto.gx@renesas.com>
Signed-off-by: Simon Horman <horms+renesas@verge.net.au>
10 years agoARM: shmobile: sh73a0: tidyup clock table order
Kuninori Morimoto [Tue, 19 Nov 2013 09:04:57 +0000 (01:04 -0800)]
ARM: shmobile: sh73a0: tidyup clock table order

SuperH lookups clock is using CLKDEV_CON/DEV/ICK_ID() macro
for a long term.
But in these days, the ICK clock is defined in random place.
This patch arranges it.

Signed-off-by: Kuninori Morimoto <kuninori.morimoto.gx@renesas.com>
Signed-off-by: Simon Horman <horms+renesas@verge.net.au>
10 years agoARM: shmobile: r8a7779: cleanup registration of sh_eth
Kuninori Morimoto [Tue, 12 Nov 2013 04:23:36 +0000 (20:23 -0800)]
ARM: shmobile: r8a7779: cleanup registration of sh_eth

sh_eth driver which needs platform data at the time of
registration is not used.
Now, ARM/shmobile aims to support DT,
and the C code base board support will be removed
if DT support is completed.
Current driver registration method which needs platform data
and which is not shared complicates codes.
This means legacy C code cleanup after DT supporting
will be more complicated
This patch removes r8a7779_add_ether_device()

Signed-off-by: Kuninori Morimoto <kuninori.morimoto.gx@renesas.com>
Signed-off-by: Simon Horman <horms+renesas@verge.net.au>
10 years agoARM: shmobile: r8a7790: Correct typo in clocks
Simon Horman [Wed, 6 Nov 2013 04:58:55 +0000 (13:58 +0900)]
ARM: shmobile: r8a7790: Correct typo in clocks

This is the r8a7790 SoC not the r8a77a4 SoC and
clocks are updated in r8a7790_clock_init not r8a73a4_clock_init.

Acked-by: Kuninori Morimoto <kuninori.morimoto.gx@renesas.com>
Signed-off-by: Simon Horman <horms+renesas@verge.net.au>
10 years agoARM: shmobile: r8a7779: camera-rcar header cleanup
Kuninori Morimoto [Fri, 1 Nov 2013 01:00:03 +0000 (18:00 -0700)]
ARM: shmobile: r8a7779: camera-rcar header cleanup

<linux/platform_data/camera-rcar.h> is needed on Marzen,
not setup-r8a7779.c

Signed-off-by: Kuninori Morimoto <kuninori.morimoto.gx@renesas.com>
Signed-off-by: Simon Horman <horms+renesas@verge.net.au>
10 years agoARM: shmobile: Add r8a7790_register_pfc() function
Magnus Damm [Wed, 16 Oct 2013 21:51:46 +0000 (06:51 +0900)]
ARM: shmobile: Add r8a7790_register_pfc() function

Break out the r8a7790 PFC platform device creation code
to increase readability and follow same style as r8a7791.

Signed-off-by: Magnus Damm <damm@opensource.se>
Signed-off-by: Simon Horman <horms+renesas@verge.net.au>
10 years agoARM: shmobile: Cosmetic update of Lager DT Reference
Magnus Damm [Wed, 16 Oct 2013 21:58:19 +0000 (06:58 +0900)]
ARM: shmobile: Cosmetic update of Lager DT Reference

Clean up the Lager DT reference board code to make it match
Koelsch DT reference including using the rcar-gen2 header.

Signed-off-by: Magnus Damm <damm@opensource.se>
Signed-off-by: Simon Horman <horms+renesas@verge.net.au>
10 years agoARM: shmobile: r8a7779: cleanup registration of VIN
Kuninori Morimoto [Tue, 15 Oct 2013 02:45:36 +0000 (19:45 -0700)]
ARM: shmobile: r8a7779: cleanup registration of VIN

VIN driver which needs platform data at the time of
registration is used from Marzen only.
Now, ARM/shmobile aims to support DT,
and the C code base board support will be removed
if DT support is completed.
Current driver registration method which needs platform data
and which is not shared complicates codes.
This means legacy C code cleanup after DT supporting
will be more complicated
This patch registers it on board code as cleanup C code

Signed-off-by: Kuninori Morimoto <kuninori.morimoto.gx@renesas.com>
Acked-by: Magnus Damm <damm@opensource.se>
Signed-off-by: Simon Horman <horms+renesas@verge.net.au>
10 years agoARM: shmobile: bockw: header cleanup
Kuninori Morimoto [Fri, 11 Oct 2013 06:34:53 +0000 (23:34 -0700)]
ARM: shmobile: bockw: header cleanup

linux/pinctrl/machine.h is no longer needed

Signed-off-by: Kuninori Morimoto <kuninori.morimoto.gx@renesas.com>
Signed-off-by: Simon Horman <horms+renesas@verge.net.au>
10 years agoARM: shmobile: sh73a0: don't use named resource for IPMMU
Kuninori Morimoto [Tue, 8 Oct 2013 05:59:06 +0000 (22:59 -0700)]
ARM: shmobile: sh73a0: don't use named resource for IPMMU

shmobile-ipmmu driver doesn't care resource name.

Signed-off-by: Kuninori Morimoto <kuninori.morimoto.gx@renesas.com>
Acked-by: Magnus Damm <damm@opensource.se>
Signed-off-by: Simon Horman <horms+renesas@verge.net.au>
10 years agoARM: shmobile: sh73a0: don't use named resource for I2C
Kuninori Morimoto [Tue, 8 Oct 2013 05:58:55 +0000 (22:58 -0700)]
ARM: shmobile: sh73a0: don't use named resource for I2C

i2c-sh_mobile driver doesn't care resource name.

Signed-off-by: Kuninori Morimoto <kuninori.morimoto.gx@renesas.com>
Acked-by: Magnus Damm <damm@opensource.se>
Signed-off-by: Simon Horman <horms+renesas@verge.net.au>
10 years agoARM: shmobile: sh73a0: don't use named resource for TMU
Kuninori Morimoto [Tue, 8 Oct 2013 05:58:42 +0000 (22:58 -0700)]
ARM: shmobile: sh73a0: don't use named resource for TMU

sh_tmu driver doesn't care resource name.

Signed-off-by: Kuninori Morimoto <kuninori.morimoto.gx@renesas.com>
Acked-by: Magnus Damm <damm@opensource.se>
Signed-off-by: Simon Horman <horms+renesas@verge.net.au>
10 years agoARM: shmobile: lager: don't use named resource for MMCIF
Kuninori Morimoto [Tue, 8 Oct 2013 05:58:30 +0000 (22:58 -0700)]
ARM: shmobile: lager: don't use named resource for MMCIF

sh_mmcif driver doesn't care resource name.

Signed-off-by: Kuninori Morimoto <kuninori.morimoto.gx@renesas.com>
Acked-by: Magnus Damm <damm@opensource.se>
Signed-off-by: Simon Horman <horms+renesas@verge.net.au>
10 years agoARM: shmobile: ape6evm: don't use named resource for SDHI
Kuninori Morimoto [Tue, 8 Oct 2013 05:58:20 +0000 (22:58 -0700)]
ARM: shmobile: ape6evm: don't use named resource for SDHI

sh_mobile_sdhi driver doesn't care resource name.

Signed-off-by: Kuninori Morimoto <kuninori.morimoto.gx@renesas.com>
Acked-by: Magnus Damm <damm@opensource.se>
Signed-off-by: Simon Horman <horms+renesas@verge.net.au>
10 years agoARM: shmobile: ape6evm: don't use named resource for MMCIF
Kuninori Morimoto [Tue, 8 Oct 2013 05:58:08 +0000 (22:58 -0700)]
ARM: shmobile: ape6evm: don't use named resource for MMCIF

sh_mmcif driver doesn't care resource name.

Signed-off-by: Kuninori Morimoto <kuninori.morimoto.gx@renesas.com>
Acked-by: Magnus Damm <damm@opensource.se>
Signed-off-by: Simon Horman <horms+renesas@verge.net.au>
10 years agoLinux 3.13-rc1
Linus Torvalds [Fri, 22 Nov 2013 19:30:55 +0000 (11:30 -0800)]
Linux 3.13-rc1

10 years agoMerge tag 'ecryptfs-3.13-rc1-quiet-checkers' of git://git.kernel.org/pub/scm/linux...
Linus Torvalds [Fri, 22 Nov 2013 18:58:14 +0000 (10:58 -0800)]
Merge tag 'ecryptfs-3.13-rc1-quiet-checkers' of git://git.kernel.org/pub/scm/linux/kernel/git/tyhicks/ecryptfs

Pull minor eCryptfs fix from Tyler Hicks:
 "Quiet static checkers by removing unneeded conditionals"

* tag 'ecryptfs-3.13-rc1-quiet-checkers' of git://git.kernel.org/pub/scm/linux/kernel/git/tyhicks/ecryptfs:
  eCryptfs: file->private_data is always valid

10 years agoMerge tag 'sound-fix2-3.13-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git...
Linus Torvalds [Fri, 22 Nov 2013 18:57:31 +0000 (10:57 -0800)]
Merge tag 'sound-fix2-3.13-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/tiwai/sound

Pull second set of sound fixes from Takashi Iwai:
 "A collection of small fixes in HD-audio quirks and runtime PM, ASoC
  rcar, abs8500 and other codecs.  Most of commits are for stable
  kernels, too"

* tag 'sound-fix2-3.13-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/tiwai/sound:
  ALSA: hda - Set current_headset_type to ALC_HEADSET_TYPE_ENUM (janitorial)
  ALSA: hda - Provide missing pin configs for VAIO with ALC260
  ALSA: hda - Add headset quirk for Dell Inspiron 3135
  ALSA: hda - Fix the headphone jack detection on Sony VAIO TX
  ALSA: hda - Fix missing bass speaker on ASUS N550
  ALSA: hda - Fix unbalanced runtime PM notification at resume
  ASoC: arizona: Set FLL to free-run before disabling
  ALSA: hda - A casual Dell Headset quirk
  ASoC: rcar: fixup dma_async_issue_pending() timing
  ASoC: rcar: off by one in rsnd_scu_set_route()
  ASoC: wm5110: Add post SYSCLK register patch for rev D chip
  ASoC: ab8500: Revert to using custom I/O functions
  ALSA: hda - Also enable mute/micmute LED control for "Lenovo dock" fixup
  ALSA: firewire-lib: include sound/asound.h to refer to snd_pcm_format_t
  ALSA: hda - Select FW_LOADER from CONFIG_SND_HDA_CODEC_CA0132_DSP
  ALSA: hda - Enable mute/mic-mute LEDs for more Thinkpads with Realtek codec
  ASoC: rcar: fixup mod access before checking

10 years agoMerge branch 'drm-fixes' of git://people.freedesktop.org/~airlied/linux
Linus Torvalds [Fri, 22 Nov 2013 18:56:11 +0000 (10:56 -0800)]
Merge branch 'drm-fixes' of git://people.freedesktop.org/~airlied/linux

Pull DRM fixes from Dave Airlie:
 "I was going to leave this until post -rc1 but sysfs fixes broke
  hotplug in userspace, so I had to fix it harder, otherwise a set of
  pulls from intel, radeon and vmware,

  The vmware/ttm changes are bit larger but since its early and they are
  unlikely to break anything else I put them in, it lets vmware work
  with dri3"

* 'drm-fixes' of git://people.freedesktop.org/~airlied/linux: (36 commits)
  drm/sysfs: fix hotplug regression since lifetime changes
  drm/exynos: g2d: fix memory leak to userptr
  drm/i915: Fix gen3 self-refresh watermarks
  drm/ttm: Remove set_need_resched from the ttm fault handler
  drm/ttm: Don't move non-existing data
  drm/radeon: hook up backlight functions for CI and KV family.
  drm/i915: Replicate BIOS eDP bpp clamping hack for hsw
  drm/i915: Do not enable package C8 on unsupported hardware
  drm/i915: Hold pc8 lock around toggling pc8.gpu_idle
  drm/i915: encoder->get_config is no longer optional
  drm/i915/tv: add ->get_config callback
  drm/radeon/cik: Add macrotile mode array query
  drm/radeon/cik: Return backend map information to userspace
  drm/vmwgfx: Make vmwgfx dma buffers prime aware
  drm/vmwgfx: Make surfaces prime-aware
  drm/vmwgfx: Hook up the prime ioctls
  drm/ttm: Add a minimal prime implementation for ttm base objects
  drm/vmwgfx: Fix false lockdep warning
  drm/ttm: Allow execbuf util reserves without ticket
  drm/i915: restore the early forcewake cleanup
  ...

10 years agoMerge tag 'pci-v3.13-fixes-1' of git://git.kernel.org/pub/scm/linux/kernel/git/helgaa...
Linus Torvalds [Fri, 22 Nov 2013 18:53:47 +0000 (10:53 -0800)]
Merge tag 'pci-v3.13-fixes-1' of git://git.kernel.org/pub/scm/linux/kernel/git/helgaas/pci

Pull PCI updates from Bjorn Helgaas:
 "Miscellaneous
   - Remove duplicate disable from pcie_portdrv_remove() (Yinghai Lu)
   - Fix whitespace, capitalization, and spelling errors (Bjorn Helgaas)"

* tag 'pci-v3.13-fixes-1' of git://git.kernel.org/pub/scm/linux/kernel/git/helgaas/pci:
  PCI: Remove duplicate pci_disable_device() from pcie_portdrv_remove()
  PCI: Fix whitespace, capitalization, and spelling errors

10 years agoMerge branch 'for-next' of git://git.kernel.org/pub/scm/linux/kernel/git/nab/target...
Linus Torvalds [Fri, 22 Nov 2013 18:52:03 +0000 (10:52 -0800)]
Merge branch 'for-next' of git://git.kernel.org/pub/scm/linux/kernel/git/nab/target-pending

Pull SCSI target updates from Nicholas Bellinger:
 "Things have been quiet this round with mostly bugfixes, percpu
  conversions, and other minor iscsi-target conformance testing changes.

  The highlights include:

   - Add demo_mode_discovery attribute for iscsi-target (Thomas)
   - Convert tcm_fc(FCoE) to use percpu-ida pre-allocation
   - Add send completion interrupt coalescing for ib_isert
   - Convert target-core to use percpu-refcounting for se_lun
   - Fix mutex_trylock usage bug in iscsit_increment_maxcmdsn
   - tcm_loop updates (Hannes)
   - target-core ALUA cleanups + prep for v3.14 SCSI Referrals support (Hannes)

  v3.14 is currently shaping to be a busy development cycle in target
  land, with initial support for T10 Referrals and T10 DIF currently on
  the roadmap"

* 'for-next' of git://git.kernel.org/pub/scm/linux/kernel/git/nab/target-pending: (40 commits)
  iscsi-target: chap auth shouldn't match username with trailing garbage
  iscsi-target: fix extract_param to handle buffer length corner case
  iscsi-target: Expose default_erl as TPG attribute
  target_core_configfs: split up ALUA supported states
  target_core_alua: Make supported states configurable
  target_core_alua: Store supported ALUA states
  target_core_alua: Rename ALUA_ACCESS_STATE_OPTIMIZED
  target_core_alua: spellcheck
  target core: rename (ex,im)plict -> (ex,im)plicit
  percpu-refcount: Add percpu-refcount.o to obj-y
  iscsi-target: Do not reject non-immediate CmdSNs exceeding MaxCmdSN
  iscsi-target: Convert iscsi_session statistics to atomic_long_t
  target: Convert se_device statistics to atomic_long_t
  target: Fix delayed Task Aborted Status (TAS) handling bug
  iscsi-target: Reject unsupported multi PDU text command sequence
  ib_isert: Avoid duplicate iscsit_increment_maxcmdsn call
  iscsi-target: Fix mutex_trylock usage in iscsit_increment_maxcmdsn
  target: Core does not need blkdev.h
  target: Pass through I/O topology for block backstores
  iser-target: Avoid using FRMR for single dma entry requests
  ...

10 years agoMerge tag 'hwmon-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/groeck...
Linus Torvalds [Fri, 22 Nov 2013 18:49:14 +0000 (10:49 -0800)]
Merge tag 'hwmon-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/groeck/linux-staging

Pull hwmon fixes from Guenter Roeck:
 - acpi_power_meter: Fix return value check from call to
   acpi_bus_get_device
 - nct6775: Fix/improve NCT6791 support
 - lm75: Add support for GMT G751

* tag 'hwmon-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/groeck/linux-staging:
  hwmon: (acpi_power_meter) Fix acpi_bus_get_device() return value check
  hwmon: (nct6775) NCT6791 supports weight control only for CPUFAN
  hwmon: (nct6775) Monitor additional temperature registers
  hwmon: (lm75) Add support for GMT G751 chip

10 years agoMerge git://git.kernel.org/pub/scm/linux/kernel/git/davem/net
Linus Torvalds [Fri, 22 Nov 2013 17:57:35 +0000 (09:57 -0800)]
Merge git://git.kernel.org/pub/scm/linux/kernel/git/davem/net

Pull networking fixes from David Miller:

 1) Fix memory leaks and other issues in mwifiex driver, from Amitkumar
    Karwar.

 2) skb_segment() can choke on packets using frag lists, fix from
    Herbert Xu with help from Eric Dumazet and others.

 3) IPv4 output cached route instantiation properly handles races
    involving two threads trying to install the same route, but we
    forgot to propagate this logic to input routes as well.  Fix from
    Alexei Starovoitov.

 4) Put protections in place to make sure that recvmsg() paths never
    accidently copy uninitialized memory back into userspace and also
    make sure that we never try to use more that sockaddr_storage for
    building the on-kernel-stack copy of a sockaddr.  Fixes from Hannes
    Frederic Sowa.

 5) R8152 driver transmit flow bug fixes from Hayes Wang.

 6) Fix some minor fallouts from genetlink changes, from Johannes Berg
    and Michael Opdenacker.

 7) AF_PACKET sendmsg path can race with netdevice unregister notifier,
    fix by using RCU to make sure the network device doesn't go away
    from under us.  Fix from Daniel Borkmann.

* git://git.kernel.org/pub/scm/linux/kernel/git/davem/net: (43 commits)
  gso: handle new frag_list of frags GRO packets
  genetlink: fix genl_set_err() group ID
  genetlink: fix genlmsg_multicast() bug
  packet: fix use after free race in send path when dev is released
  xen-netback: stop the VIF thread before unbinding IRQs
  wimax: remove dead code
  net/phy: Add the autocross feature for forced links on VSC82x4
  net/phy: Add VSC8662 support
  net/phy: Add VSC8574 support
  net/phy: Add VSC8234 support
  net: add BUG_ON if kernel advertises msg_namelen > sizeof(struct sockaddr_storage)
  net: rework recvmsg handler msg_name and msg_namelen logic
  bridge: flush br's address entry in fdb when remove the
  net: core: Always propagate flag changes to interfaces
  ipv4: fix race in concurrent ip_route_input_slow()
  r8152: fix incorrect type in assignment
  r8152: support stopping/waking tx queue
  r8152: modify the tx flow
  r8152: fix tx/rx memory overflow
  netfilter: ebt_ip6: fix source and destination matching
  ...

10 years agoMerge branch 'fixes' of git://git.linaro.org/people/rmk/linux-arm
Linus Torvalds [Fri, 22 Nov 2013 17:56:51 +0000 (09:56 -0800)]
Merge branch 'fixes' of git://git.linaro.org/people/rmk/linux-arm

Pull ARM fixes from Russell King:
 "Some small fixes for this merge window, most of them quite self
  explanatory - the biggest thing here is a fix for the ARMv7 LPAE
  suspend/resume support"

* 'fixes' of git://git.linaro.org/people/rmk/linux-arm:
  ARM: 7894/1: kconfig: select GENERIC_CLOCKEVENTS if HAVE_ARM_ARCH_TIMER
  ARM: 7893/1: bitops: only emit .arch_extension mp if CONFIG_SMP
  ARM: 7892/1: Fix warning for V7M builds
  ARM: 7888/1: seccomp: not compatible with ARM OABI
  ARM: 7886/1: make OABI default to off
  ARM: 7885/1: Save/Restore 64-bit TTBR registers on LPAE suspend/resume
  ARM: 7884/1: mm: Fix ECC mem policy printk
  ARM: 7883/1: fix mov to mvn conversion in case of 64 bit phys_addr_t and BE
  ARM: 7882/1: mm: fix __phys_to_virt to work with 64 bit phys_addr_t in BE case
  ARM: 7881/1: __fixup_smp read of SCU config should do byteswap in BE case
  ARM: Fix nommu.c build warning

10 years agoMerge branch 'next' of git://git.kernel.org/pub/scm/virt/kvm/kvm
Linus Torvalds [Fri, 22 Nov 2013 17:56:07 +0000 (09:56 -0800)]
Merge branch 'next' of git://git.kernel.org/pub/scm/virt/kvm/kvm

Pull KVM fixes from Gleb Natapov.

* 'next' of git://git.kernel.org/pub/scm/virt/kvm/kvm:
  KVM: kvm_clear_guest_page(): fix empty_zero_page usage
  kvm: mmu: delay mmu audit activation
  arm/arm64: KVM: Fix hyp mappings of vmalloc regions

10 years agoMerge git://git.kvack.org/~bcrl/aio-next
Linus Torvalds [Fri, 22 Nov 2013 16:42:14 +0000 (08:42 -0800)]
Merge git://git.kvack.org/~bcrl/aio-next

Pull aio fixes from Benjamin LaHaise.

* git://git.kvack.org/~bcrl/aio-next:
  aio: nullify aio->ring_pages after freeing it
  aio: prevent double free in ioctx_alloc
  aio: Fix a trinity splat

10 years agoMerge branch 'for-3.13' of git://linux-nfs.org/~bfields/linux
Linus Torvalds [Fri, 22 Nov 2013 16:41:17 +0000 (08:41 -0800)]
Merge branch 'for-3.13' of git://linux-nfs.org/~bfields/linux

Pull nfsd bugfixes from Bruce Fields:
 "A couple nfsd bugfixes"

* 'for-3.13' of git://linux-nfs.org/~bfields/linux:
  nfsd4: fix xdr decoding of large non-write compounds
  nfsd: make sure to balance get/put_write_access
  nfsd: split up nfsd_setattr

10 years agoMerge tag 'gfs2-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/steve/gfs2...
Linus Torvalds [Fri, 22 Nov 2013 16:39:44 +0000 (08:39 -0800)]
Merge tag 'gfs2-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/steve/gfs2-3.0-fixes

Pull GFS2 fixes from Steven Whitehouse:
 "A couple of small, but important bug fixes for GFS2.  The first one
  fixes a possible NULL pointer dereference, and the second one resolves
  a reference counting issue in one of the lesser used paths through
  atomic_open"

* tag 'gfs2-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/steve/gfs2-3.0-fixes:
  GFS2: Fix ref count bug relating to atomic_open
  GFS2: fix potential NULL pointer dereference

10 years agoMerge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/mason/linux...
Linus Torvalds [Fri, 22 Nov 2013 16:38:55 +0000 (08:38 -0800)]
Merge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/mason/linux-btrfs

Pull btrfs fixes from Chris Mason:
 "Almost all of these are bug fixes.  Dave Sterba's documentation update
  is the big exception because he removed our promises to set any
  machine running Btrfs on fire"

* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/mason/linux-btrfs:
  Documentation: filesystems: update btrfs tools section
  Documentation: filesystems: add new btrfs mount options
  btrfs: update kconfig help text
  btrfs: fix bio_size_ok() for max_sectors > 0xffff
  btrfs: Use trace condition for get_extent tracepoint
  btrfs: fix typo in the log message
  Btrfs: fix list delete warning when removing ordered root from the list
  Btrfs: print bytenr instead of page pointer in check-int
  Btrfs: remove dead codes from ctree.h
  Btrfs: don't wait for ordered data outside desired range
  Btrfs: fix lockdep error in async commit
  Btrfs: avoid heavy operations in btrfs_commit_super
  Btrfs: fix __btrfs_start_workers retval
  Btrfs: disable online raid-repair on ro mounts
  Btrfs: do not inc uncorrectable_errors counter on ro scrubs
  Btrfs: only drop modified extents if we logged the whole inode
  Btrfs: make sure to copy everything if we rename
  Btrfs: don't BUG_ON() if we get an error walking backrefs

10 years agoMerge tag 'xfs-for-linus-v3.13-rc1-2' of git://oss.sgi.com/xfs/xfs
Linus Torvalds [Fri, 22 Nov 2013 16:37:47 +0000 (08:37 -0800)]
Merge tag 'xfs-for-linus-v3.13-rc1-2' of git://oss.sgi.com/xfs/xfs

Pull second xfs update from Ben Myers:
 "There are a couple of patches that I wasn't quite sure about in time
  for our initial 3.13 pull request, a bugfix, and an update to add Dave
  to MAINTAINERS:

  Here we have a performance fix for inode iversion, increased inode
  cluster size for v5 superblock filesystems, a fix for error handling
  in xfs_bmap_add_attrfork, and a MAINTAINERS update to add Dave"

* tag 'xfs-for-linus-v3.13-rc1-2' of git://oss.sgi.com/xfs/xfs:
  xfs: open code inc_inode_iversion when logging an inode
  xfs: increase inode cluster size for v5 filesystems
  xfs: fix unlock in xfs_bmap_add_attrfork
  xfs: update maintainers

10 years agoMerge branch 'slab/next' of git://git.kernel.org/pub/scm/linux/kernel/git/penberg...
Linus Torvalds [Fri, 22 Nov 2013 16:10:34 +0000 (08:10 -0800)]
Merge branch 'slab/next' of git://git.kernel.org/pub/scm/linux/kernel/git/penberg/linux

Pull SLAB changes from Pekka Enberg:
 "The patches from Joonsoo Kim switch mm/slab.c to use 'struct page' for
  slab internals similar to mm/slub.c.  This reduces memory usage and
  improves performance:

    https://lkml.org/lkml/2013/10/16/155

  Rest of the changes are bug fixes from various people"

* 'slab/next' of git://git.kernel.org/pub/scm/linux/kernel/git/penberg/linux: (21 commits)
  mm, slub: fix the typo in mm/slub.c
  mm, slub: fix the typo in include/linux/slub_def.h
  slub: Handle NULL parameter in kmem_cache_flags
  slab: replace non-existing 'struct freelist *' with 'void *'
  slab: fix to calm down kmemleak warning
  slub: proper kmemleak tracking if CONFIG_SLUB_DEBUG disabled
  slab: rename slab_bufctl to slab_freelist
  slab: remove useless statement for checking pfmemalloc
  slab: use struct page for slab management
  slab: replace free and inuse in struct slab with newly introduced active
  slab: remove SLAB_LIMIT
  slab: remove kmem_bufctl_t
  slab: change the management method of free objects of the slab
  slab: use __GFP_COMP flag for allocating slab pages
  slab: use well-defined macro, virt_to_slab()
  slab: overloading the RCU head over the LRU for RCU free
  slab: remove cachep in struct slab_rcu
  slab: remove nodeid in struct slab
  slab: remove colouroff in struct slab
  slab: change return type of kmem_getpages() to struct page
  ...

10 years agoMerge branch 'merge' of git://git.kernel.org/pub/scm/linux/kernel/git/benh/powerpc
Linus Torvalds [Fri, 22 Nov 2013 16:07:11 +0000 (08:07 -0800)]
Merge branch 'merge' of git://git.kernel.org/pub/scm/linux/kernel/git/benh/powerpc

Pull third set of powerpc updates from Benjamin Herrenschmidt:
 "This is a small collection of random bug fixes and a few improvements
  of Oops output which I deemed valuable enough to include as well.

  The fixes are essentially recent build breakage and regressions, and a
  couple of older bugs such as the DTL log duplication, the EEH issue
  with PCI_COMMAND_MASTER and the problem with small contexts passed to
  get/set_context with VSX enabled"

* 'merge' of git://git.kernel.org/pub/scm/linux/kernel/git/benh/powerpc:
  powerpc/signals: Mark VSX not saved with small contexts
  powerpc/pseries: Fix SMP=n build of rng.c
  powerpc: Make cpu_to_chip_id() available when SMP=n
  powerpc/vio: Fix a dma_mask issue of vio
  powerpc: booke: Fix build failures
  powerpc: ppc64 address space capped at 32TB, mmap randomisation disabled
  powerpc: Only print PACATMSCRATCH in oops when TM is active
  powerpc/pseries: Duplicate dtl entries sometimes sent to userspace
  powerpc: Remove a few lines of oops output
  powerpc: Print DAR and DSISR on machine check oopses
  powerpc: Fix __get_user_pages_fast() irq handling
  powerpc/eeh: More accurate log
  powerpc/eeh: Enable PCI_COMMAND_MASTER for PCI bridges

10 years agoALSA: hda - Set current_headset_type to ALC_HEADSET_TYPE_ENUM (janitorial)
David Henningsson [Fri, 22 Nov 2013 11:17:06 +0000 (12:17 +0100)]
ALSA: hda - Set current_headset_type to ALC_HEADSET_TYPE_ENUM (janitorial)

current_headset_type should be of the HEADSET_TYPE enum, not the
HEADSET_MODE enum. Since ALC_HEADSET_TYPE_UNKNOWN and ALC_HEADSET_MODE_UNKNOWN
are both 0, this patch is just janitorial.

Signed-off-by: David Henningsson <david.henningsson@canonical.com>
Signed-off-by: Takashi Iwai <tiwai@suse.de>
10 years agoALSA: hda - Provide missing pin configs for VAIO with ALC260
Takashi Iwai [Fri, 22 Nov 2013 07:06:36 +0000 (08:06 +0100)]
ALSA: hda - Provide missing pin configs for VAIO with ALC260

Some models (or maybe depending on BIOS version) of Sony VAIO with
ALC260 give no proper pin configurations as default, resulting in the
non-working speaker, etc.  Just provide the whole pin configurations
via a fixup.

Reported-by: Matthew Markus <mmarkus@hearit.co>
Cc: <stable@vger.kernel.org>
Signed-off-by: Takashi Iwai <tiwai@suse.de>
10 years agoMerge branch 'akpm' (fixes from Andrew)
Linus Torvalds [Fri, 22 Nov 2013 05:32:38 +0000 (21:32 -0800)]
Merge branch 'akpm' (fixes from Andrew)

Merge patches from Andrew Morton:
 "13 fixes"

* emailed patches from Andrew Morton <akpm@linux-foundation.org>:
  mm: place page->pmd_huge_pte to right union
  MAINTAINERS: add keyboard driver to Hyper-V file list
  x86, mm: do not leak page->ptl for pmd page tables
  ipc,shm: correct error return value in shmctl (SHM_UNLOCK)
  mm, mempolicy: silence gcc warning
  block/partitions/efi.c: fix bound check
  ARM: drivers/rtc/rtc-at91rm9200.c: disable interrupts at shutdown
  mm: hugetlbfs: fix hugetlbfs optimization
  kernel: remove CONFIG_USE_GENERIC_SMP_HELPERS cleanly
  ipc,shm: fix shm_file deletion races
  mm: thp: give transparent hugepage code a separate copy_page
  checkpatch: fix "Use of uninitialized value" warnings
  configfs: fix race between dentry put and lookup

10 years agoMerge branch 'for-linus2' of git://git.kernel.org/pub/scm/linux/kernel/git/jmorris...
Linus Torvalds [Fri, 22 Nov 2013 03:46:00 +0000 (19:46 -0800)]
Merge branch 'for-linus2' of git://git.kernel.org/pub/scm/linux/kernel/git/jmorris/linux-security

Pull security subsystem updates from James Morris:
 "In this patchset, we finally get an SELinux update, with Paul Moore
  taking over as maintainer of that code.

  Also a significant update for the Keys subsystem, as well as
  maintenance updates to Smack, IMA, TPM, and Apparmor"

and since I wanted to know more about the updates to key handling,
here's the explanation from David Howells on that:

 "Okay.  There are a number of separate bits.  I'll go over the big bits
  and the odd important other bit, most of the smaller bits are just
  fixes and cleanups.  If you want the small bits accounting for, I can
  do that too.

   (1) Keyring capacity expansion.

        KEYS: Consolidate the concept of an 'index key' for key access
        KEYS: Introduce a search context structure
        KEYS: Search for auth-key by name rather than target key ID
        Add a generic associative array implementation.
        KEYS: Expand the capacity of a keyring

     Several of the patches are providing an expansion of the capacity of a
     keyring.  Currently, the maximum size of a keyring payload is one page.
     Subtract a small header and then divide up into pointers, that only gives
     you ~500 pointers on an x86_64 box.  However, since the NFS idmapper uses
     a keyring to store ID mapping data, that has proven to be insufficient to
     the cause.

     Whatever data structure I use to handle the keyring payload, it can only
     store pointers to keys, not the keys themselves because several keyrings
     may point to a single key.  This precludes inserting, say, and rb_node
     struct into the key struct for this purpose.

     I could make an rbtree of records such that each record has an rb_node
     and a key pointer, but that would use four words of space per key stored
     in the keyring.  It would, however, be able to use much existing code.

     I selected instead a non-rebalancing radix-tree type approach as that
     could have a better space-used/key-pointer ratio.  I could have used the
     radix tree implementation that we already have and insert keys into it by
     their serial numbers, but that means any sort of search must iterate over
     the whole radix tree.  Further, its nodes are a bit on the capacious side
     for what I want - especially given that key serial numbers are randomly
     allocated, thus leaving a lot of empty space in the tree.

     So what I have is an associative array that internally is a radix-tree
     with 16 pointers per node where the index key is constructed from the key
     type pointer and the key description.  This means that an exact lookup by
     type+description is very fast as this tells us how to navigate directly to
     the target key.

     I made the data structure general in lib/assoc_array.c as far as it is
     concerned, its index key is just a sequence of bits that leads to a
     pointer.  It's possible that someone else will be able to make use of it
     also.  FS-Cache might, for example.

   (2) Mark keys as 'trusted' and keyrings as 'trusted only'.

        KEYS: verify a certificate is signed by a 'trusted' key
        KEYS: Make the system 'trusted' keyring viewable by userspace
        KEYS: Add a 'trusted' flag and a 'trusted only' flag
        KEYS: Separate the kernel signature checking keyring from module signing

     These patches allow keys carrying asymmetric public keys to be marked as
     being 'trusted' and allow keyrings to be marked as only permitting the
     addition or linkage of trusted keys.

     Keys loaded from hardware during kernel boot or compiled into the kernel
     during build are marked as being trusted automatically.  New keys can be
     loaded at runtime with add_key().  They are checked against the system
     keyring contents and if their signatures can be validated with keys that
     are already marked trusted, then they are marked trusted also and can
     thus be added into the master keyring.

     Patches from Mimi Zohar make this usable with the IMA keyrings also.

   (3) Remove the date checks on the key used to validate a module signature.

        X.509: Remove certificate date checks

     It's not reasonable to reject a signature just because the key that it was
     generated with is no longer valid datewise - especially if the kernel
     hasn't yet managed to set the system clock when the first module is
     loaded - so just remove those checks.

   (4) Make it simpler to deal with additional X.509 being loaded into the kernel.

        KEYS: Load *.x509 files into kernel keyring
        KEYS: Have make canonicalise the paths of the X.509 certs better to deduplicate

     The builder of the kernel now just places files with the extension ".x509"
     into the kernel source or build trees and they're concatenated by the
     kernel build and stuffed into the appropriate section.

   (5) Add support for userspace kerberos to use keyrings.

        KEYS: Add per-user_namespace registers for persistent per-UID kerberos caches
        KEYS: Implement a big key type that can save to tmpfs

     Fedora went to, by default, storing kerberos tickets and tokens in tmpfs.
     We looked at storing it in keyrings instead as that confers certain
     advantages such as tickets being automatically deleted after a certain
     amount of time and the ability for the kernel to get at these tokens more
     easily.

     To make this work, two things were needed:

     (a) A way for the tickets to persist beyond the lifetime of all a user's
         sessions so that cron-driven processes can still use them.

         The problem is that a user's session keyrings are deleted when the
         session that spawned them logs out and the user's user keyring is
         deleted when the UID is deleted (typically when the last log out
         happens), so neither of these places is suitable.

         I've added a system keyring into which a 'persistent' keyring is
         created for each UID on request.  Each time a user requests their
         persistent keyring, the expiry time on it is set anew.  If the user
         doesn't ask for it for, say, three days, the keyring is automatically
         expired and garbage collected using the existing gc.  All the kerberos
         tokens it held are then also gc'd.

     (b) A key type that can hold really big tickets (up to 1MB in size).

         The problem is that Active Directory can return huge tickets with lots
         of auxiliary data attached.  We don't, however, want to eat up huge
         tracts of unswappable kernel space for this, so if the ticket is
         greater than a certain size, we create a swappable shmem file and dump
         the contents in there and just live with the fact we then have an
         inode and a dentry overhead.  If the ticket is smaller than that, we
         slap it in a kmalloc()'d buffer"

* 'for-linus2' of git://git.kernel.org/pub/scm/linux/kernel/git/jmorris/linux-security: (121 commits)
  KEYS: Fix keyring content gc scanner
  KEYS: Fix error handling in big_key instantiation
  KEYS: Fix UID check in keyctl_get_persistent()
  KEYS: The RSA public key algorithm needs to select MPILIB
  ima: define '_ima' as a builtin 'trusted' keyring
  ima: extend the measurement list to include the file signature
  kernel/system_certificate.S: use real contents instead of macro GLOBAL()
  KEYS: fix error return code in big_key_instantiate()
  KEYS: Fix keyring quota misaccounting on key replacement and unlink
  KEYS: Fix a race between negating a key and reading the error set
  KEYS: Make BIG_KEYS boolean
  apparmor: remove the "task" arg from may_change_ptraced_domain()
  apparmor: remove parent task info from audit logging
  apparmor: remove tsk field from the apparmor_audit_struct
  apparmor: fix capability to not use the current task, during reporting
  Smack: Ptrace access check mode
  ima: provide hash algo info in the xattr
  ima: enable support for larger default filedata hash algorithms
  ima: define kernel parameter 'ima_template=' to change configured default
  ima: add Kconfig default measurement list template
  ...

10 years agoMerge git://git.infradead.org/users/eparis/audit
Linus Torvalds [Fri, 22 Nov 2013 03:18:14 +0000 (19:18 -0800)]
Merge git://git.infradead.org/users/eparis/audit

Pull audit updates from Eric Paris:
 "Nothing amazing.  Formatting, small bug fixes, couple of fixes where
  we didn't get records due to some old VFS changes, and a change to how
  we collect execve info..."

Fixed conflict in fs/exec.c as per Eric and linux-next.

* git://git.infradead.org/users/eparis/audit: (28 commits)
  audit: fix type of sessionid in audit_set_loginuid()
  audit: call audit_bprm() only once to add AUDIT_EXECVE information
  audit: move audit_aux_data_execve contents into audit_context union
  audit: remove unused envc member of audit_aux_data_execve
  audit: Kill the unused struct audit_aux_data_capset
  audit: do not reject all AUDIT_INODE filter types
  audit: suppress stock memalloc failure warnings since already managed
  audit: log the audit_names record type
  audit: add child record before the create to handle case where create fails
  audit: use given values in tty_audit enable api
  audit: use nlmsg_len() to get message payload length
  audit: use memset instead of trying to initialize field by field
  audit: fix info leak in AUDIT_GET requests
  audit: update AUDIT_INODE filter rule to comparator function
  audit: audit feature to set loginuid immutable
  audit: audit feature to only allow unsetting the loginuid
  audit: allow unsetting the loginuid (with priv)
  audit: remove CONFIG_AUDIT_LOGINUID_IMMUTABLE
  audit: loginuid functions coding style
  selinux: apply selinux checks on new audit message types
  ...

10 years agomm: place page->pmd_huge_pte to right union
Kirill A. Shutemov [Thu, 21 Nov 2013 22:32:11 +0000 (14:32 -0800)]
mm: place page->pmd_huge_pte to right union

I don't know what went wrong, mis-merge or something, but ->pmd_huge_pte
placed in wrong union within struct page.

In original patch[1] it's placed to union with ->lru and ->slab, but in
commit e009bb30c8df ("mm: implement split page table lock for PMD
level") it's in union with ->index and ->freelist.

That union seems also unused for pages with table tables and safe to
re-use, but it's not what I've tested.

Let's move it to original place.  It fixes indentation at least.  :)

[1] https://lkml.org/lkml/2013/10/7/288

Signed-off-by: Kirill A. Shutemov <kirill.shutemov@linux.intel.com>
Reviewed-by: Naoya Horiguchi <n-horiguchi@ah.jp.nec.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
10 years agoMAINTAINERS: add keyboard driver to Hyper-V file list
Haiyang Zhang [Thu, 21 Nov 2013 22:32:10 +0000 (14:32 -0800)]
MAINTAINERS: add keyboard driver to Hyper-V file list

Signed-off-by: Haiyang Zhang <haiyangz@microsoft.com>
Cc: Dmitry Torokhov <dmitry.torokhov@gmail.com>
Cc: "K. Y. Srinivasan" <kys@microsoft.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
10 years agox86, mm: do not leak page->ptl for pmd page tables
Kirill A. Shutemov [Thu, 21 Nov 2013 22:32:09 +0000 (14:32 -0800)]
x86, mm: do not leak page->ptl for pmd page tables

There are two code paths how page with pmd page table can be freed:
pmd_free() and pmd_free_tlb().

I've missed the second one and didn't add page table destructor call
there.  It leads to leak of page->ptl for pmd page tables, if
dynamically allocated page->ptl is in use.

The patch adds the missed destructor and modifies documentation
accordingly.

Signed-off-by: Kirill A. Shutemov <kirill.shutemov@linux.intel.com>
Reported-by: Andrey Vagin <avagin@openvz.org>
Tested-by: Andrey Vagin <avagin@openvz.org>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
10 years agoipc,shm: correct error return value in shmctl (SHM_UNLOCK)
Jesper Nilsson [Thu, 21 Nov 2013 22:32:08 +0000 (14:32 -0800)]
ipc,shm: correct error return value in shmctl (SHM_UNLOCK)

Commit 2caacaa82a51 ("ipc,shm: shorten critical region for shmctl")
restructured the ipc shm to shorten critical region, but introduced a
path where the return value could be -EPERM, even if the operation
actually was performed.

Before the commit, the err return value was reset by the return value
from security_shm_shmctl() after the if (!ns_capable(...)) statement.

Now, we still exit the if statement with err set to -EPERM, and in the
case of SHM_UNLOCK, it is not reset at all, and used as the return value
from shmctl.

To fix this, we only set err when errors occur, leaving the fallthrough
case alone.

Signed-off-by: Jesper Nilsson <jesper.nilsson@axis.com>
Cc: Davidlohr Bueso <davidlohr@hp.com>
Cc: Rik van Riel <riel@redhat.com>
Cc: Michel Lespinasse <walken@google.com>
Cc: Al Viro <viro@zeniv.linux.org.uk>
Cc: <stable@vger.kernel.org> [3.12.x]
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
10 years agomm, mempolicy: silence gcc warning
David Rientjes [Thu, 21 Nov 2013 22:32:06 +0000 (14:32 -0800)]
mm, mempolicy: silence gcc warning

Fengguang Wu reports that compiling mm/mempolicy.c results in a warning:

  mm/mempolicy.c: In function 'mpol_to_str':
  mm/mempolicy.c:2878:2: error: format not a string literal and no format arguments

Kees says this is because he is using -Wformat-security.

Silence the warning.

Signed-off-by: David Rientjes <rientjes@google.com>
Reported-by: Fengguang Wu <fengguang.wu@intel.com>
Suggested-by: Kees Cook <keescook@chromium.org>
Acked-by: Kees Cook <keescook@chromium.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
10 years agoblock/partitions/efi.c: fix bound check
Antti P Miettinen [Thu, 21 Nov 2013 22:32:05 +0000 (14:32 -0800)]
block/partitions/efi.c: fix bound check

Use ARRAY_SIZE instead of sizeof to get proper max for label length.

Since this is just a read out of bounds it's not that bad, but the
problem becomes user-visible eg if one tries to use DEBUG_PAGEALLOC and
DEBUG_RODATA, at least with some enhancements from Hiroshi.  Of course
the destination array can contain garbage when we read beyond the end of
source array so that would be another user-visible problem.

Signed-off-by: Antti P Miettinen <amiettinen@nvidia.com>
Reviewed-by: Hiroshi Doyu <hdoyu@nvidia.com>
Tested-by: Hiroshi Doyu <hdoyu@nvidia.com>
Cc: Will Drewry <wad@chromium.org>
Cc: Matt Fleming <matt.fleming@intel.com>
Acked-by: Davidlohr Bueso <davidlohr@hp.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
10 years agoARM: drivers/rtc/rtc-at91rm9200.c: disable interrupts at shutdown
Johan Hovold [Thu, 21 Nov 2013 22:32:04 +0000 (14:32 -0800)]
ARM: drivers/rtc/rtc-at91rm9200.c: disable interrupts at shutdown

Make sure RTC-interrupts are disabled at shutdown.

As the RTC is generally powered by backup power (VDDBU), its interrupts
are not disabled on wake-up, user, watchdog or software reset.  This
could cause troubles on other systems (e.g.  older kernels) if an
interrupt occurs before a handler has been installed at next boot.

Let us be well-behaved and disable them on clean shutdowns at least (as
do the RTT-based rtc-at91sam9 driver).

Signed-off-by: Johan Hovold <jhovold@gmail.com>
Acked-by: Nicolas Ferre <nicolas.ferre@atmel.com>
Cc: Jean-Christophe Plagniol-Villard <plagnioj@jcrosoft.com>
Cc: Andrew Victor <linux@maxim.org.za>
Cc: Alessandro Zummo <a.zummo@towertech.it>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
10 years agomm: hugetlbfs: fix hugetlbfs optimization
Andrea Arcangeli [Thu, 21 Nov 2013 22:32:02 +0000 (14:32 -0800)]
mm: hugetlbfs: fix hugetlbfs optimization

Commit 7cb2ef56e6a8 ("mm: fix aio performance regression for database
caused by THP") can cause dereference of a dangling pointer if
split_huge_page runs during PageHuge() if there are updates to the
tail_page->private field.

Also it is repeating compound_head twice for hugetlbfs and it is running
compound_head+compound_trans_head for THP when a single one is needed in
both cases.

The new code within the PageSlab() check doesn't need to verify that the
THP page size is never bigger than the smallest hugetlbfs page size, to
avoid memory corruption.

A longstanding theoretical race condition was found while fixing the
above (see the change right after the skip_unlock label, that is
relevant for the compound_lock path too).

By re-establishing the _mapcount tail refcounting for all compound
pages, this also fixes the below problem:

  echo 0 >/sys/kernel/mm/hugepages/hugepages-2048kB/nr_hugepages

  BUG: Bad page state in process bash  pfn:59a01
  page:ffffea000139b038 count:0 mapcount:10 mapping:          (null) index:0x0
  page flags: 0x1c00000000008000(tail)
  Modules linked in:
  CPU: 6 PID: 2018 Comm: bash Not tainted 3.12.0+ #25
  Hardware name: Bochs Bochs, BIOS Bochs 01/01/2011
  Call Trace:
    dump_stack+0x55/0x76
    bad_page+0xd5/0x130
    free_pages_prepare+0x213/0x280
    __free_pages+0x36/0x80
    update_and_free_page+0xc1/0xd0
    free_pool_huge_page+0xc2/0xe0
    set_max_huge_pages.part.58+0x14c/0x220
    nr_hugepages_store_common.isra.60+0xd0/0xf0
    nr_hugepages_store+0x13/0x20
    kobj_attr_store+0xf/0x20
    sysfs_write_file+0x189/0x1e0
    vfs_write+0xc5/0x1f0
    SyS_write+0x55/0xb0
    system_call_fastpath+0x16/0x1b

Signed-off-by: Khalid Aziz <khalid.aziz@oracle.com>
Signed-off-by: Andrea Arcangeli <aarcange@redhat.com>
Tested-by: Khalid Aziz <khalid.aziz@oracle.com>
Cc: Pravin Shelar <pshelar@nicira.com>
Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Cc: Ben Hutchings <bhutchings@solarflare.com>
Cc: Christoph Lameter <cl@linux.com>
Cc: Johannes Weiner <jweiner@redhat.com>
Cc: Mel Gorman <mgorman@suse.de>
Cc: Rik van Riel <riel@redhat.com>
Cc: Andi Kleen <andi@firstfloor.org>
Cc: Minchan Kim <minchan@kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
10 years agokernel: remove CONFIG_USE_GENERIC_SMP_HELPERS cleanly
Yuanhan Liu [Thu, 21 Nov 2013 22:32:01 +0000 (14:32 -0800)]
kernel: remove CONFIG_USE_GENERIC_SMP_HELPERS cleanly

Remove CONFIG_USE_GENERIC_SMP_HELPERS left by commit 0a06ff068f12
("kernel: remove CONFIG_USE_GENERIC_SMP_HELPERS").

Signed-off-by: Yuanhan Liu <yuanhan.liu@linux.intel.com>
Cc: Christoph Hellwig <hch@infradead.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
10 years agoipc,shm: fix shm_file deletion races
Greg Thelen [Thu, 21 Nov 2013 22:32:00 +0000 (14:32 -0800)]
ipc,shm: fix shm_file deletion races

When IPC_RMID races with other shm operations there's potential for
use-after-free of the shm object's associated file (shm_file).

Here's the race before this patch:

  TASK 1                     TASK 2
  ------                     ------
  shm_rmid()
    ipc_lock_object()
                             shmctl()
                             shp = shm_obtain_object_check()

    shm_destroy()
      shum_unlock()
      fput(shp->shm_file)
                             ipc_lock_object()
                             shmem_lock(shp->shm_file)
                             <OOPS>

The oops is caused because shm_destroy() calls fput() after dropping the
ipc_lock.  fput() clears the file's f_inode, f_path.dentry, and
f_path.mnt, which causes various NULL pointer references in task 2.  I
reliably see the oops in task 2 if with shmlock, shmu

This patch fixes the races by:
1) set shm_file=NULL in shm_destroy() while holding ipc_object_lock().
2) modify at risk operations to check shm_file while holding
   ipc_object_lock().

Example workloads, which each trigger oops...

Workload 1:
  while true; do
    id=$(shmget 1 4096)
    shm_rmid $id &
    shmlock $id &
    wait
  done

  The oops stack shows accessing NULL f_inode due to racing fput:
    _raw_spin_lock
    shmem_lock
    SyS_shmctl

Workload 2:
  while true; do
    id=$(shmget 1 4096)
    shmat $id 4096 &
    shm_rmid $id &
    wait
  done

  The oops stack is similar to workload 1 due to NULL f_inode:
    touch_atime
    shmem_mmap
    shm_mmap
    mmap_region
    do_mmap_pgoff
    do_shmat
    SyS_shmat

Workload 3:
  while true; do
    id=$(shmget 1 4096)
    shmlock $id
    shm_rmid $id &
    shmunlock $id &
    wait
  done

  The oops stack shows second fput tripping on an NULL f_inode.  The
  first fput() completed via from shm_destroy(), but a racing thread did
  a get_file() and queued this fput():
    locks_remove_flock
    __fput
    ____fput
    task_work_run
    do_notify_resume
    int_signal

Fixes: c2c737a0461e ("ipc,shm: shorten critical region for shmat")
Fixes: 2caacaa82a51 ("ipc,shm: shorten critical region for shmctl")
Signed-off-by: Greg Thelen <gthelen@google.com>
Cc: Davidlohr Bueso <davidlohr@hp.com>
Cc: Rik van Riel <riel@redhat.com>
Cc: Manfred Spraul <manfred@colorfullife.com>
Cc: <stable@vger.kernel.org> # 3.10.17+ 3.11.6+
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
10 years agomm: thp: give transparent hugepage code a separate copy_page
Dave Hansen [Thu, 21 Nov 2013 22:31:58 +0000 (14:31 -0800)]
mm: thp: give transparent hugepage code a separate copy_page

Right now, the migration code in migrate_page_copy() uses copy_huge_page()
for hugetlbfs and thp pages:

       if (PageHuge(page) || PageTransHuge(page))
                copy_huge_page(newpage, page);

So, yay for code reuse.  But:

  void copy_huge_page(struct page *dst, struct page *src)
  {
        struct hstate *h = page_hstate(src);

and a non-hugetlbfs page has no page_hstate().  This works 99% of the
time because page_hstate() determines the hstate from the page order
alone.  Since the page order of a THP page matches the default hugetlbfs
page order, it works.

But, if you change the default huge page size on the boot command-line
(say default_hugepagesz=1G), then we might not even *have* a 2MB hstate
so page_hstate() returns null and copy_huge_page() oopses pretty fast
since copy_huge_page() dereferences the hstate:

  void copy_huge_page(struct page *dst, struct page *src)
  {
        struct hstate *h = page_hstate(src);
        if (unlikely(pages_per_huge_page(h) > MAX_ORDER_NR_PAGES)) {
  ...

Mel noticed that the migration code is really the only user of these
functions.  This moves all the copy code over to migrate.c and makes
copy_huge_page() work for THP by checking for it explicitly.

I believe the bug was introduced in commit b32967ff101a ("mm: numa: Add
THP migration for the NUMA working set scanning fault case")

[akpm@linux-foundation.org: fix coding-style and comment text, per Naoya Horiguchi]
Signed-off-by: Dave Hansen <dave.hansen@linux.intel.com>
Acked-by: Mel Gorman <mgorman@suse.de>
Reviewed-by: Naoya Horiguchi <n-horiguchi@ah.jp.nec.com>
Cc: Hillf Danton <dhillf@gmail.com>
Cc: Andrea Arcangeli <aarcange@redhat.com>
Tested-by: Dave Jiang <dave.jiang@intel.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
10 years agocheckpatch: fix "Use of uninitialized value" warnings
Joe Perches [Thu, 21 Nov 2013 22:31:57 +0000 (14:31 -0800)]
checkpatch: fix "Use of uninitialized value" warnings

checkpatch is currently confused about some complex macros and references
undefined variables $stat and $cond.

Make sure these are defined before using them.

Signed-off-by: Joe Perches <joe@perches.com>
Reported-by: Gerhard Sittig <gsi@denx.de>
Acked-by: Andy Whitcroft <apw@canonical.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
10 years agoconfigfs: fix race between dentry put and lookup
Junxiao Bi [Thu, 21 Nov 2013 22:31:56 +0000 (14:31 -0800)]
configfs: fix race between dentry put and lookup

A race window in configfs, it starts from one dentry is UNHASHED and end
before configfs_d_iput is called.  In this window, if a lookup happen,
since the original dentry was UNHASHED, so a new dentry will be
allocated, and then in configfs_attach_attr(), sd->s_dentry will be
updated to the new dentry.  Then in configfs_d_iput(),
BUG_ON(sd->s_dentry != dentry) will be triggered and system panic.

sys_open:                     sys_close:
 ...                           fput
                                dput
                                 dentry_kill
                                  __d_drop <--- dentry unhashed here,
                                           but sd->dentry still point
                                           to this dentry.

 lookup_real
  configfs_lookup
   configfs_attach_attr---> update sd->s_dentry
                            to new allocated dentry here.

                                   d_kill
                                     configfs_d_iput <--- BUG_ON(sd->s_dentry != dentry)
                                                     triggered here.

To fix it, change configfs_d_iput to not update sd->s_dentry if
sd->s_count > 2, that means there are another dentry is using the sd
beside the one that is going to be put.  Use configfs_dirent_lock in
configfs_attach_attr to sync with configfs_d_iput.

With the following steps, you can reproduce the bug.

1. enable ocfs2, this will mount configfs at /sys/kernel/config and
   fill configure in it.

2. run the following script.
while [ 1 ]; do cat /sys/kernel/config/cluster/$your_cluster_name/idle_timeout_ms > /dev/null; done &
while [ 1 ]; do cat /sys/kernel/config/cluster/$your_cluster_name/idle_timeout_ms > /dev/null; done &

Signed-off-by: Junxiao Bi <junxiao.bi@oracle.com>
Cc: Joel Becker <jlbec@evilplan.org>
Cc: Al Viro <viro@zeniv.linux.org.uk>
Cc: <stable@vger.kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
10 years agogso: handle new frag_list of frags GRO packets
Herbert Xu [Thu, 21 Nov 2013 19:10:04 +0000 (11:10 -0800)]
gso: handle new frag_list of frags GRO packets

Recently GRO started generating packets with frag_lists of frags.
This was not handled by GSO, thus leading to a crash.

Thankfully these packets are of a regular form and are easy to
handle.  This patch handles them in two ways.  For completely
non-linear frag_list entries, we simply continue to iterate over
the frag_list frags once we exhaust the normal frags.  For frag_list
entries with linear parts, we call pskb_trim on the first part
of the frag_list skb, and then process the rest of the frags in
the usual way.

This patch also kills a chunk of dead frag_list code that has
obviously never ever been run since it ends up generating a bogus
GSO-segmented packet with a frag_list entry.

Future work is planned to split super big packets into TSO
ones.

Fixes: 8a29111c7ca6 ("net: gro: allow to build full sized skb")
Reported-by: Christoph Paasch <christoph.paasch@uclouvain.be>
Reported-by: Jerry Chu <hkchu@google.com>
Reported-by: Sander Eikelenboom <linux@eikelenboom.it>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Signed-off-by: Eric Dumazet <edumazet@google.com>
Tested-by: Sander Eikelenboom <linux@eikelenboom.it>
Tested-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
10 years agoGFS2: Fix ref count bug relating to atomic_open
Steven Whitehouse [Thu, 21 Nov 2013 18:47:57 +0000 (18:47 +0000)]
GFS2: Fix ref count bug relating to atomic_open

In the case that atomic_open calls finish_no_open() with
the dentry that was supplied to gfs2_atomic_open() an
extra reference count is required. This patch fixes that
issue preventing a bug trap triggering at umount time.

Signed-off-by: Steven Whitehouse <swhiteho@redhat.com>
10 years agogenetlink: fix genl_set_err() group ID
Johannes Berg [Thu, 21 Nov 2013 17:20:28 +0000 (18:20 +0100)]
genetlink: fix genl_set_err() group ID

Fix another really stupid bug - I introduced genl_set_err()
precisely to be able to adjust the group and reject invalid
ones, but then forgot to do so.

Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
10 years agogenetlink: fix genlmsg_multicast() bug
Johannes Berg [Thu, 21 Nov 2013 17:17:04 +0000 (18:17 +0100)]
genetlink: fix genlmsg_multicast() bug

Unfortunately, I introduced a tremendously stupid bug into
genlmsg_multicast() when doing all those multicast group
changes: it adjusts the group number, but then passes it
to genlmsg_multicast_netns() which does that again.

Somehow, my tests failed to catch this, so add a warning
into genlmsg_multicast_netns() and remove the offending
group ID adjustment.

Also add a warning to the similar code in other functions
so people who misuse them are more loudly warned.

Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
10 years agopacket: fix use after free race in send path when dev is released
Daniel Borkmann [Thu, 21 Nov 2013 15:50:58 +0000 (16:50 +0100)]
packet: fix use after free race in send path when dev is released

Salam reported a use after free bug in PF_PACKET that occurs when
we're sending out frames on a socket bound device and suddenly the
net device is being unregistered. It appears that commit 827d9780
introduced a possible race condition between {t,}packet_snd() and
packet_notifier(). In the case of a bound socket, packet_notifier()
can drop the last reference to the net_device and {t,}packet_snd()
might end up suddenly sending a packet over a freed net_device.

To avoid reverting 827d9780 and thus introducing a performance
regression compared to the current state of things, we decided to
hold a cached RCU protected pointer to the net device and maintain
it on write side via bind spin_lock protected register_prot_hook()
and __unregister_prot_hook() calls.

In {t,}packet_snd() path, we access this pointer under rcu_read_lock
through packet_cached_dev_get() that holds reference to the device
to prevent it from being freed through packet_notifier() while
we're in send path. This is okay to do as dev_put()/dev_hold() are
per-cpu counters, so this should not be a performance issue. Also,
the code simplifies a bit as we don't need need_rls_dev anymore.

Fixes: 827d978037d7 ("af-packet: Use existing netdev reference for bound sockets.")
Reported-by: Salam Noureddine <noureddine@aristanetworks.com>
Signed-off-by: Daniel Borkmann <dborkman@redhat.com>
Signed-off-by: Salam Noureddine <noureddine@aristanetworks.com>
Cc: Ben Greear <greearb@candelatech.com>
Cc: Eric Dumazet <eric.dumazet@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>