GitHub/LineageOS/android_kernel_motorola_exynos9610.git
9 years agoARM: dts: am33xx: Update DCAN nodes
Roger Quadros [Tue, 9 Sep 2014 13:15:35 +0000 (16:15 +0300)]
ARM: dts: am33xx: Update DCAN nodes

Add "raminit-syscon" property to specify the RAMINIT register.
Add clock information.
Rename can nodes from "d_can" to "can" to be compliant
with the ePAPR specs.

Signed-off-by: Roger Quadros <rogerq@ti.com>
Signed-off-by: Tony Lindgren <tony@atomide.com>
9 years agoARM: dts: am33xx: Add control module syscon node
Roger Quadros [Tue, 9 Sep 2014 13:28:01 +0000 (16:28 +0300)]
ARM: dts: am33xx: Add control module syscon node

Use syscon regmap to expose the Control module register space.
This register space is shared between many users e.g. DCAN, USB, display, etc.

Signed-off-by: Roger Quadros <rogerq@ti.com>
Signed-off-by: Tony Lindgren <tony@atomide.com>
9 years agoARM: dts: am437x-gp: Add dcan support
Mugunthan V N [Fri, 18 Jul 2014 05:59:11 +0000 (11:29 +0530)]
ARM: dts: am437x-gp: Add dcan support

Add DCAN support for AM437x GP EVM with both DCAN instances.

[Roger Q] Updated output pin to not use pull up.

Signed-off-by: Mugunthan V N <mugunthanvnm@ti.com>
Signed-off-by: George Cherian <george.cherian@ti.com>
Signed-off-by: Sekhar Nori <nsekhar@ti.com>
Signed-off-by: Roger Quadros <rogerq@ti.com>
Signed-off-by: Tony Lindgren <tony@atomide.com>
9 years agoARM: dts: am4372: Add DCAN nodes
Roger Quadros [Thu, 4 Sep 2014 12:36:03 +0000 (15:36 +0300)]
ARM: dts: am4372: Add DCAN nodes

The SoC contains 2 DCAN modules. Add them.

Signed-off-by: Roger Quadros <rogerq@ti.com>
Signed-off-by: Tony Lindgren <tony@atomide.com>
9 years agoARM: dts: am4372: Add control module syscon node
Roger Quadros [Mon, 8 Sep 2014 12:45:10 +0000 (15:45 +0300)]
ARM: dts: am4372: Add control module syscon node

Use syscon regmap to expose the Control module register space.
This register space is shared between many users e.g. DCAN, USB, display, etc.

Signed-off-by: Roger Quadros <rogerq@ti.com>
Signed-off-by: Tony Lindgren <tony@atomide.com>
9 years agoARM: dts: dra72-evm: Add CAN support
Roger Quadros [Mon, 3 Nov 2014 11:07:18 +0000 (13:07 +0200)]
ARM: dts: dra72-evm: Add CAN support

The board has 2 CAN ports but only the first one can be used.
Enable the first CAN port.

WAKEUP0 pin doesn't have INPUT enable bit so we just disable
weak PULLs.

The second CAN port cannot be used without hardware modification
so we don't enable the second port.

Signed-off-by: Roger Quadros <rogerq@ti.com>
Signed-off-by: Tony Lindgren <tony@atomide.com>
9 years agoARM: dts: dra7-evm: Add CAN support
Roger Quadros [Fri, 15 Aug 2014 13:09:19 +0000 (16:09 +0300)]
ARM: dts: dra7-evm: Add CAN support

The board has 2 CAN ports but only the first one can be used.
Enable the first CAN port.

WAKEUP0 pin doesn't have INPUT enable bit so we just disable
weak PULLs.

The second CAN port cannot be used without hardware modification
so we don't enable the second port.

Signed-off-by: Roger Quadros <rogerq@ti.com>
Signed-off-by: Tony Lindgren <tony@atomide.com>
9 years agoARM: dts: DRA7: Add DCAN nodes
Roger Quadros [Fri, 15 Aug 2014 13:08:36 +0000 (16:08 +0300)]
ARM: dts: DRA7: Add DCAN nodes

The SoC supports 2 DCAN nodes. Add them.

Signed-off-by: Roger Quadros <rogerq@ti.com>
Signed-off-by: Tony Lindgren <tony@atomide.com>
9 years agoARM: dts: dra7: Add syscon regmap for CORE CONTROL area
Roger Quadros [Wed, 3 Sep 2014 14:21:45 +0000 (17:21 +0300)]
ARM: dts: dra7: Add syscon regmap for CORE CONTROL area

Display and DCAN drivers use syscon regmap to access some registers
in the CORE control area. Add the syscon regmap node for this
area.

Cc: Tomi Valkeinen <tomi.valkeinen@ti.com>
Cc: Nishanth Menon <nm@ti.com>
Signed-off-by: Roger Quadros <rogerq@ti.com>
Signed-off-by: Tony Lindgren <tony@atomide.com>
9 years agoARM: dts: sbc-t3x30: add audio support
Dmitry Lifshitz [Sun, 23 Nov 2014 09:42:02 +0000 (11:42 +0200)]
ARM: dts: sbc-t3x30: add audio support

Add audio related DT nodes

Signed-off-by: Dmitry Lifshitz <lifshitz@compulab.co.il>
Signed-off-by: Tony Lindgren <tony@atomide.com>
9 years agoARM: dts: sbc-t3x: add TV out display alias
Dmitry Lifshitz [Sun, 23 Nov 2014 09:42:01 +0000 (11:42 +0200)]
ARM: dts: sbc-t3x: add TV out display alias

Add display alias for TV out.

Signed-off-by: Dmitry Lifshitz <lifshitz@compulab.co.il>
Signed-off-by: Tony Lindgren <tony@atomide.com>
9 years agoARM: dts: cm-t3x: add TV out support
Dmitry Lifshitz [Sun, 23 Nov 2014 09:42:00 +0000 (11:42 +0200)]
ARM: dts: cm-t3x: add TV out support

Add TV out support.

Signed-off-by: Dmitry Lifshitz <lifshitz@compulab.co.il>
Signed-off-by: Tony Lindgren <tony@atomide.com>
9 years agoARM: dts: cm-t3x: add I2C1 pinmux
Dmitry Lifshitz [Sun, 23 Nov 2014 09:41:59 +0000 (11:41 +0200)]
ARM: dts: cm-t3x: add I2C1 pinmux

Add missing I2C1 pinmux setup.

Signed-off-by: Dmitry Lifshitz <lifshitz@compulab.co.il>
Signed-off-by: Tony Lindgren <tony@atomide.com>
9 years agoARM: dts: AM43xx: add tscadc DT entries for am437x-evm and am43x-epos-evm
Vignesh R [Fri, 21 Nov 2014 10:14:22 +0000 (15:44 +0530)]
ARM: dts: AM43xx: add tscadc DT entries for am437x-evm and am43x-epos-evm

This patch adds tscadc DT entries for am437x-gp-evm
and am43x-epos-evm.

Signed-off-by: Vignesh R <vigneshr@ti.com>
Signed-off-by: Tony Lindgren <tony@atomide.com>
9 years agoARM: dts: cm-t3x30: add keypad support
Dmitry Lifshitz [Tue, 18 Nov 2014 09:13:24 +0000 (11:13 +0200)]
ARM: dts: cm-t3x30: add keypad support

Add twl4030 matrtix keypad support.

Signed-off-by: Dmitry Lifshitz <lifshitz@compulab.co.il>
Signed-off-by: Tony Lindgren <tony@atomide.com>
9 years agoARM: dts: sb-t35: add EEPROM support
Dmitry Lifshitz [Tue, 18 Nov 2014 09:13:20 +0000 (11:13 +0200)]
ARM: dts: sb-t35: add EEPROM support

Add at24 EEPROM chip support.

Signed-off-by: Dmitry Lifshitz <lifshitz@compulab.co.il>
Signed-off-by: Tony Lindgren <tony@atomide.com>
9 years agoARM: dts: cm-t3x: add EEPROM support
Dmitry Lifshitz [Tue, 18 Nov 2014 09:13:19 +0000 (11:13 +0200)]
ARM: dts: cm-t3x: add EEPROM support

Add at24 EEPROM chip support.

Signed-off-by: Dmitry Lifshitz <lifshitz@compulab.co.il>
[tony@atomide.com: updated to remove missing i2c1_pins]
Signed-off-by: Tony Lindgren <tony@atomide.com>
9 years agoARM: OMAP2+: remove cm-t3x touchscreen pdata quirk
Dmitry Lifshitz [Tue, 18 Nov 2014 09:13:18 +0000 (11:13 +0200)]
ARM: OMAP2+: remove cm-t3x touchscreen pdata quirk

Remove ADS7846 touchscreen pdata quirk for CM-T3x CoMs

Signed-off-by: Dmitry Lifshitz <lifshitz@compulab.co.il>
Signed-off-by: Tony Lindgren <tony@atomide.com>
9 years agoARM: dts: cm-t3x: add ADS7846 touchscreen support
Dmitry Lifshitz [Tue, 18 Nov 2014 09:13:17 +0000 (11:13 +0200)]
ARM: dts: cm-t3x: add ADS7846 touchscreen support

Add ADS7846 touchscreen support.

Signed-off-by: Dmitry Lifshitz <lifshitz@compulab.co.il>
Signed-off-by: Tony Lindgren <tony@atomide.com>
9 years agoARM: dts: cm-t3x: cleanup comments indentation
Dmitry Lifshitz [Tue, 18 Nov 2014 09:13:16 +0000 (11:13 +0200)]
ARM: dts: cm-t3x: cleanup comments indentation

Fix comment style

Signed-off-by: Dmitry Lifshitz <lifshitz@compulab.co.il>
Signed-off-by: Tony Lindgren <tony@atomide.com>
9 years agoARM: dts: DRA7: Add node for RTC
Lokesh Vutla [Wed, 19 Nov 2014 12:23:08 +0000 (17:53 +0530)]
ARM: dts: DRA7: Add node for RTC

Add node for RTC.

Note that on dra7xx are no separate interrupts for alram
and timer unlike for SoCs.

Signed-off-by: Lokesh Vutla <lokeshvutla@ti.com>
[nm@ti.com: update with rtc crossbar number]
Signed-off-by: Nishanth Menon <nm@ti.com>
[tony@atomide.com: updated comments]
Signed-off-by: Tony Lindgren <tony@atomide.com>
9 years agoARM: dst: OMAP3-N900: Add n900-battery support
Sebastian Reichel [Sat, 15 Nov 2014 02:08:24 +0000 (03:08 +0100)]
ARM: dst: OMAP3-N900: Add n900-battery support

This adds support for the N900's battery to the
Nokia N900 DTS file.

Signed-off-by: Sebastian Reichel <sre@kernel.org>
Signed-off-by: Tony Lindgren <tony@atomide.com>
9 years agoARM: dts: OMAP3-N900: add si4713 support
Sebastian Reichel [Sat, 15 Nov 2014 01:47:30 +0000 (02:47 +0100)]
ARM: dts: OMAP3-N900: add si4713 support

Add si4713 node to the N900 device tree file.

Signed-off-by: Sebastian Reichel <sre@kernel.org>
Signed-off-by: Tony Lindgren <tony@atomide.com>
9 years agoARM: dts: Gumstix DuoVero: Bind vdac regulator to hdmi node
Adam YH Lee [Thu, 13 Nov 2014 18:04:33 +0000 (10:04 -0800)]
ARM: dts: Gumstix DuoVero: Bind vdac regulator to hdmi node

The HDMI node does not have a power supply attached. As a result its
power regulator, VDAC, shuts off on boot and screen loses signal.
This attaches VDAC (vdda_hdmi_dac) to HDMI's vdda-supply.

Signed-off-by: Adam YH Lee <adam.yh.lee@gmail.com>
Signed-off-by: Tony Lindgren <tony@atomide.com>
9 years agoARM: dts: Add devicetree for NovaTech OrionLXm
George McCollister [Mon, 17 Nov 2014 19:02:35 +0000 (13:02 -0600)]
ARM: dts: Add devicetree for NovaTech OrionLXm

This adds the NovaTech OrionLXm which is based on the AM335x SoC
http://www.novatechweb.com/substation-automation/orionlxm/

RAM: 512MiB
Flash: 4GB eMMC
Ethernet PHYs: 2x Micrel KSZ8041FTLI
USB ports are used internally by the expansion cards.
Internal micro SD slot is available.

Signed-off-by: George McCollister <george.mccollister@gmail.com>
Reviewed-by: Felipe Balbi <balbi@ti.com>
Signed-off-by: Tony Lindgren <tony@atomide.com>
10 years agoARM: dts: DRA7: Add aliases for all serial ports
Nishanth Menon [Tue, 21 Oct 2014 16:18:15 +0000 (11:18 -0500)]
ARM: dts: DRA7: Add aliases for all serial ports

Add serial port aliases for consoles > 6.

Signed-off-by: Nishanth Menon <nm@ti.com>
Signed-off-by: Tony Lindgren <tony@atomide.com>
10 years agoARM: dts: Add am57xx-beagle-x15
Nishanth Menon [Fri, 7 Nov 2014 17:47:13 +0000 (11:47 -0600)]
ARM: dts: Add am57xx-beagle-x15

BeagleBoard-X15 is the next generation Open Source Hardware
BeagleBoard based on TI's AM5728 SoC featuring dual core 1.5GHz A15
processor. The platform features 2GB DDR3L (w/dual 32bit busses),
eSATA, 3 USB3.0 ports, integrated HDMI (1920x1080@60), separate LCD
port, video In port, 4GB eMMC, uSD, Analog audio in/out, dual 1G
Ethernet.

For more information, refer to:
BeagleBoard-X15 Wiki:
http://www.elinux.org/Beagleboard:BeagleBoard-X15

AM5728 is part of the Sitara product family whose additional details
will be available: http://www.ti.com/lsds/ti/arm/overview.page

Technical Reference Manual for AM5728 is public domain at:
http://www.ti.com/lit/spruhz6

Just add basic support for the moment, the following updates are needed:
  i) Ethernet - depends on SoC dts fixes
  ii) USB Client (USB2) - depends on GPIO extcon
  ii) HDMI - additional driver fixes pending
  iii) Audio - additional driver fixes pending

NOTE:
AM5728 Data Manual (SPRS915L - August 2014) section 4.1.1 states: "All
unused power supply balls must be supplied with the voltages specified
in the Section 5.2, Recommended Operating Conditions". This implies
that all unused voltage rails for AM5728 can never be switched off even
if the hardware blocks inside that voltage domain is unused. Switching
off these unused rails may result in stability issues on other domains
and increased leakage and power-on-hour impacts.

Signed-off-by: Felipe Balbi <balbi@ti.com>
Signed-off-by: Nishanth Menon <nm@ti.com>
Reviewed-by: Tom Rini <trini@ti.com>
Signed-off-by: Tony Lindgren <tony@atomide.com>
10 years agoARM: OMAP2+: igep00x0: Add pdata-quirks for the btwilink device.
Enric Balletbo i Serra [Thu, 6 Nov 2014 12:01:51 +0000 (13:01 +0100)]
ARM: OMAP2+: igep00x0: Add pdata-quirks for the btwilink device.

Add btwilink device for IGEPv2 Rev. F and IGEP COM MODULE Rev. G.

Signed-off-by: Enric Balletbo i Serra <eballetbo@iseebcn.com>
Acked-by: Javier Martinez Canillas <javier@dowhile0.org>
Signed-off-by: Tony Lindgren <tony@atomide.com>
10 years agoARM: dts: omap3-igep00x0: Remove i2c2 node.
Enric Balletbo i Serra [Thu, 6 Nov 2014 12:01:50 +0000 (13:01 +0100)]
ARM: dts: omap3-igep00x0: Remove i2c2 node.

We can't suppose that the i2c2 pins are configured as I2C bus, these pins are
connected to expansion connectors.

Signed-off-by: Enric Balletbo i Serra <eballetbo@iseebcn.com>
Acked-by: Javier Martinez Canillas <javier@dowhile0.org>
Signed-off-by: Tony Lindgren <tony@atomide.com>
10 years agoARM: dts: omap3-igep0020-rev-f: Support IGEPv2 Rev. F
Enric Balletbo i Serra [Thu, 6 Nov 2014 12:01:49 +0000 (13:01 +0100)]
ARM: dts: omap3-igep0020-rev-f: Support IGEPv2 Rev. F

Add support for the new hardware revision of the IGEPv2. Basically, the new
revision F replaces the old Wifi module for a Wilink8 based module.

Signed-off-by: Enric Balletbo i Serra <eballetbo@iseebcn.com>
Acked-by: Javier Martinez Canillas <javier@dowhile0.org>
Signed-off-by: Tony Lindgren <tony@atomide.com>
10 years agoARM: dts: omap3-igep0020-common: Introduce igep0020 common dtsi file.
Enric Balletbo i Serra [Thu, 6 Nov 2014 12:01:48 +0000 (13:01 +0100)]
ARM: dts: omap3-igep0020-common: Introduce igep0020 common dtsi file.

Use the omap3-igep0020-common.dtsi file and remove repeated parts leaving
the nodes that are not common between IGEPv2 hardware revisions.

Signed-off-by: Enric Balletbo i Serra <eballetbo@iseebcn.com>
Acked-by: Javier Martinez Canillas <javier@dowhile0.org>
Signed-off-by: Tony Lindgren <tony@atomide.com>
10 years agoARM: dts: omap3-igep0030-rev-g: Support IGEP COM MODULE Rev. G
Enric Balletbo i Serra [Thu, 6 Nov 2014 12:01:47 +0000 (13:01 +0100)]
ARM: dts: omap3-igep0030-rev-g: Support IGEP COM MODULE Rev. G

Add support for the new hardware revision of the IGEP COM MODULE. Basically,
the new revision G replaces the old Wifi module for a Wilink8 based module.

Signed-off-by: Enric Balletbo i Serra <eballetbo@iseebcn.com>
Acked-by: Javier Martinez Canillas <javier@dowhile0.org>
Signed-off-by: Tony Lindgren <tony@atomide.com>
10 years agoARM: dts: omap3-igep0030-common: Introduce igep0030 common dtsi file.
Enric Balletbo i Serra [Thu, 6 Nov 2014 12:01:46 +0000 (13:01 +0100)]
ARM: dts: omap3-igep0030-common: Introduce igep0030 common dtsi file.

Use the omap3-igep0030-common.dtsi file and remove repeated parts leaving
the nodes that are not common between IGEP COM MODULE hardware revisions.

Signed-off-by: Enric Balletbo i Serra <eballetbo@iseebcn.com>
Acked-by: Javier Martinez Canillas <javier@dowhile0.org>
Signed-off-by: Tony Lindgren <tony@atomide.com>
10 years agoARM: dts: omap3-igep00x0: Move outside common file the on board Wifi module.
Enric Balletbo i Serra [Thu, 6 Nov 2014 12:01:45 +0000 (13:01 +0100)]
ARM: dts: omap3-igep00x0: Move outside common file the on board Wifi module.

New IGEP boards revisions will use another Wifi module, so this patch moves
the DT nodes outside the common omap3-igep.dtsi file to specific DT for every
board.

Signed-off-by: Enric Balletbo i Serra <eballetbo@iseebcn.com>
Acked-by: Javier Martinez Canillas <javier@dowhile0.org>
Signed-off-by: Tony Lindgren <tony@atomide.com>
10 years agoARM: dts: omap3-igep0020: Specify IGEPv2 revision in device tree.
Enric Balletbo i Serra [Thu, 6 Nov 2014 12:01:44 +0000 (13:01 +0100)]
ARM: dts: omap3-igep0020: Specify IGEPv2 revision in device tree.

We'll introduce new hardware revisions soon. This patch is only to
indicate which board revision supports this device tree file in order
to avoid confusions.

Signed-off-by: Enric Balletbo i Serra <eballetbo@iseebcn.com>
Acked-by: Javier Martinez Canillas <javier@dowhile0.org>
Signed-off-by: Tony Lindgren <tony@atomide.com>
10 years agoARM: dts: omap3-igep0030: Specify IGEP COM revision in device tree.
Enric Balletbo i Serra [Thu, 6 Nov 2014 12:01:43 +0000 (13:01 +0100)]
ARM: dts: omap3-igep0030: Specify IGEP COM revision in device tree.

We'll introduce new hardware revisions soon. This patch is only to
indicate which board revision supports this device tree file in order
to avoid confusions.

Signed-off-by: Enric Balletbo i Serra <eballetbo@iseebcn.com>
Acked-by: Javier Martinez Canillas <javier@dowhile0.org>
Signed-off-by: Tony Lindgren <tony@atomide.com>
10 years agoARM: dts: omap3-igep00x0: Move NAND configuration to a common place.
Enric Balletbo i Serra [Thu, 6 Nov 2014 12:01:42 +0000 (13:01 +0100)]
ARM: dts: omap3-igep00x0: Move NAND configuration to a common place.

At this moment all supported boards use same NAND chip, so has more sense
move the GPMC and NAND configuration to the omap3-igep.dtsi common place.

Signed-off-by: Enric Balletbo i Serra <eballetbo@iseebcn.com>
Acked-by: Javier Martinez Canillas <javier@dowhile0.org>
Signed-off-by: Tony Lindgren <tony@atomide.com>
10 years agoARM: dts: omap3-igep00x0: Fix UART2 pins that aren't common.
Enric Balletbo i Serra [Thu, 6 Nov 2014 12:01:41 +0000 (13:01 +0100)]
ARM: dts: omap3-igep00x0: Fix UART2 pins that aren't common.

UART2 is used to connect the processor with the bluetooth chip, these pins
are not common between IGEPv2 boards and IGEP COM MODULE boards. This patch
muxes the correct pins for every board and removes UART2 configuration from
common omap3-igep.dtsi file.

Signed-off-by: Enric Balletbo i Serra <eballetbo@iseebcn.com>
Acked-by: Javier Martinez Canillas <javier@dowhile0.org>
Signed-off-by: Tony Lindgren <tony@atomide.com>
10 years agoARM: dts: dra7: add labels to DWC3 nodes
Felipe Balbi [Mon, 3 Nov 2014 16:28:42 +0000 (10:28 -0600)]
ARM: dts: dra7: add labels to DWC3 nodes

by adding labels to DWC3 nodes, it's far easier
for boards to reference them.

Signed-off-by: Felipe Balbi <balbi@ti.com>
[tony@atomide.com: updated for otg 4 move to dra74x.dtsi]
Signed-off-by: Tony Lindgren <tony@atomide.com>
10 years agoARM: dts: dra72x-evm: Enable CPSW and MDIO
Mugunthan V N [Mon, 3 Nov 2014 09:58:13 +0000 (15:28 +0530)]
ARM: dts: dra72x-evm: Enable CPSW and MDIO

Adding CPSW phy-id, CPSW and MDIO pinmux configuration for active and
sleep states and enable them in board evm dts file.

Signed-off-by: Mugunthan V N <mugunthanvnm@ti.com>
Signed-off-by: Tony Lindgren <tony@atomide.com>
10 years agoARM: dts: dra7-evm: Keep all VDD rails always-on
Nishanth Menon [Tue, 21 Oct 2014 14:38:10 +0000 (09:38 -0500)]
ARM: dts: dra7-evm: Keep all VDD rails always-on

DRA7 Data Manual (SPRS857L - August 2014) section 4.1.1 states: "All
unused power supply balls must be supplied with the voltages specified
in the Section 5.2, Recommended Operating Conditions".

This implies that all unused voltage rails for Vayu can never be
switched off even if the hardware blocks inside that voltage domain is
unused. Switching off these unused rails may result in stability issues
on other domains and increased leakage and power-on-hour impacts.

J6eco-evm dts file already considers this, however j6evm-dts file needs
to be fixed to consider this constraint of the SoC.

Signed-off-by: Nishanth Menon <nm@ti.com>
Acked-by: Roger Quadros <rogerq@ti.com>
Signed-off-by: Tony Lindgren <tony@atomide.com>
10 years agoARM: dts: dra72-evm: Add MMC nodes
Nishanth Menon [Tue, 21 Oct 2014 14:35:56 +0000 (09:35 -0500)]
ARM: dts: dra72-evm: Add MMC nodes

Add MMC1 and 2 nodes. MMC1 is SDcard and MMC2 is eMMC.

NOTE on MMC1 card detect: Ideally, we should be using in-built SDCD
support, but we dont have it yet. So, use the fact that control module
of DRA7 is setup such that no matter what mode one configures it, GPIO
option is always hardwired in - use GPIO mode for SDcard detection.

[peter.ujfalusi@ti.com]
The power line feeding the SD card is also used by other devices on the EVM.
Use generic name instead of mmc2_3v3 so when other devices want to use the
same regulator it will look a bit better.

Signed-off-by: Nishanth Menon <nm@ti.com>
Signed-off-by: Peter Ujfalusi <peter.ujfalusi@ti.com>
Signed-off-by: Tony Lindgren <tony@atomide.com>
10 years agoARM: dts: dra72-evm: Add power button node
Nishanth Menon [Tue, 21 Oct 2014 14:30:47 +0000 (09:30 -0500)]
ARM: dts: dra72-evm: Add power button node

With Commit adff5962fdd2 ("Input: introduce palmas-pwrbutton"), we can
now support tps power button as a event source - This is SW7 (PB/WAKE)
on the J6-evm.

Signed-off-by: Nishanth Menon <nm@ti.com>
Signed-off-by: Tony Lindgren <tony@atomide.com>
10 years agoARM: dts: dra72-evm: Provide explicit pinmux for TPS PMIC
Nishanth Menon [Tue, 21 Oct 2014 14:30:46 +0000 (09:30 -0500)]
ARM: dts: dra72-evm: Provide explicit pinmux for TPS PMIC

Even thought sys_nirq1 is hardwired on the SoC for the pin, it is
better to configure the pin to the required mux configuration.

Signed-off-by: Nishanth Menon <nm@ti.com>
Signed-off-by: Tony Lindgren <tony@atomide.com>
10 years agoARM: dts: dra72-evm: Add regulator information to USB2 PHYs
Roger Quadros [Tue, 21 Oct 2014 10:41:20 +0000 (13:41 +0300)]
ARM: dts: dra72-evm: Add regulator information to USB2 PHYs

The ldo4_reg regulator provides power to the USB1 and USB2
High Speed PHYs.

Signed-off-by: Roger Quadros <rogerq@ti.com>
Signed-off-by: Tony Lindgren <tony@atomide.com>
10 years agoARM: dts: dra72-evm: Enable USB support for dra72-evm.
George Cherian [Tue, 21 Oct 2014 10:41:19 +0000 (13:41 +0300)]
ARM: dts: dra72-evm: Enable USB support for dra72-evm.

Add USB data and pinctrl for USB.

Signed-off-by: George Cherian <george.cherian@ti.com>
Signed-off-by: Sekhar Nori <nsekhar@ti.com>
Signed-off-by: Roger Quadros <rogerq@ti.com>
Signed-off-by: Tony Lindgren <tony@atomide.com>
10 years agoARM: dts: DRA7: Move USB_OTG 4 to dra74x.dtsi
Roger Quadros [Tue, 21 Oct 2014 10:41:18 +0000 (13:41 +0300)]
ARM: dts: DRA7: Move USB_OTG 4 to dra74x.dtsi

The 4th USB controller instance present only on the DRA74x family of
devices so move it there.

Signed-off-by: Roger Quadros <rogerq@ti.com>
Acked-by: Nishanth Menon <nm@ti.com>
Signed-off-by: Sekhar Nori <nsekhar@ti.com>
Signed-off-by: Tony Lindgren <tony@atomide.com>
10 years agoARM: dts: dra72-evm: Add NAND support
Roger Quadros [Tue, 21 Oct 2014 10:41:17 +0000 (13:41 +0300)]
ARM: dts: dra72-evm: Add NAND support

DRA72-evm has a 256MB 16-bit wide NAND chip. Add
pinmux and NAND node.

The NAND chips 'Chip select' and 'Write protect' can be
controlled using DIP Switch SW5. To use NAND,
the switch must be configured like so:

SW5.1 (NAND_SELn) = ON (LOW)
SW5.9 (GPMC_WPN) = OFF (HIGH)

Signed-off-by: Roger Quadros <rogerq@ti.com>
Signed-off-by: Sekhar Nori <nsekhar@ti.com>
Tested-by: Nishanth Menon <nm@ti.com>
Acked-by: Nishanth Menon <nm@ti.com>
Signed-off-by: Tony Lindgren <tony@atomide.com>
10 years agoARM: dts: dra7-evm: Enable CPSW and MDIO for dra7xx EVM
Mugunthan V N [Tue, 21 Oct 2014 10:01:01 +0000 (15:31 +0530)]
ARM: dts: dra7-evm: Enable CPSW and MDIO for dra7xx EVM

Adding CPSW phy-id, CPSW and MDIO pinmux configuration for active and
sleep states and enable them in board evm dts file.

Signed-off-by: Mugunthan V N <mugunthanvnm@ti.com>
Signed-off-by: Tony Lindgren <tony@atomide.com>
10 years agoARM: dts: dra7: Add CPSW and MDIO module nodes for dra7
Mugunthan V N [Tue, 21 Oct 2014 10:01:00 +0000 (15:31 +0530)]
ARM: dts: dra7: Add CPSW and MDIO module nodes for dra7

Add CPSW and MDIO related device tree data for DRA7XX and made as status
disabled. Phy-id, pinmux for active and sleep state needs to be added in
board dts files and enable the CPSW device.

Signed-off-by: Mugunthan V N <mugunthanvnm@ti.com>
Signed-off-by: Tony Lindgren <tony@atomide.com>
10 years agoARM: dts: omap3-gta04: Use omap specific pinctrl defines
Marek Belisko [Fri, 17 Oct 2014 20:32:03 +0000 (22:32 +0200)]
ARM: dts: omap3-gta04: Use omap specific pinctrl defines

Use omap specific pinctrl defines (OMAP3_CORE1_IOPAD) to configure
the padconf register offset.

Signed-off-by: Marek Belisko <marek@goldelico.com>
Signed-off-by: Tony Lindgren <tony@atomide.com>
10 years agoARM: dts: sbc-t3x: add DVI display data
Dmitry Lifshitz [Sun, 2 Nov 2014 11:19:58 +0000 (13:19 +0200)]
ARM: dts: sbc-t3x: add DVI display data

Add DSS related pinmux and display data nodes required to support
DVI video out on SBC-T3530, SBC-T3730 and SBC-T3517.

Signed-off-by: Dmitry Lifshitz <lifshitz@compulab.co.il>
Acked-by: Igor Grinberg <grinberg@compulab.co.il>
Signed-off-by: Tony Lindgren <tony@atomide.com>
10 years agoARM: dts: dra7: add DMA properties for UART
Sebastian Andrzej Siewior [Mon, 29 Sep 2014 18:06:47 +0000 (20:06 +0200)]
ARM: dts: dra7: add DMA properties for UART

Cc: devicetree@vger.kernel.org
Reviewed-by: Tony Lindgren <tony@atomide.com>
Tested-by: Tony Lindgren <tony@atomide.com>
Signed-off-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de>
Signed-off-by: Tony Lindgren <tony@atomide.com>
10 years agoARM: dts: am33xx: add DMA properties for UART
Sebastian Andrzej Siewior [Mon, 29 Sep 2014 18:06:46 +0000 (20:06 +0200)]
ARM: dts: am33xx: add DMA properties for UART

Cc: devicetree@vger.kernel.org
Reviewed-by: Tony Lindgren <tony@atomide.com>
Tested-by: Tony Lindgren <tony@atomide.com>
Signed-off-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de>
Signed-off-by: Tony Lindgren <tony@atomide.com>
10 years agoARM: dts: OMAP2+: Add #mbox-cells property to all mailbox nodes
Suman Anna [Mon, 3 Nov 2014 23:07:35 +0000 (17:07 -0600)]
ARM: dts: OMAP2+: Add #mbox-cells property to all mailbox nodes

The '#mbox-cells' property is added to all the OMAP mailbox
nodes. This property is mandatory with the new mailbox framework.

Cc: "Benoît Cousson" <bcousson@baylibre.com>
Cc: Rob Herring <robh+dt@kernel.org>
Cc: Pawel Moll <pawel.moll@arm.com>
Cc: Mark Rutland <mark.rutland@arm.com>
Cc: Ian Campbell <ijc+devicetree@hellion.org.uk>
Cc: Kumar Gala <galak@codeaurora.org>
Signed-off-by: Suman Anna <s-anna@ti.com>
Signed-off-by: Tony Lindgren <tony@atomide.com>
10 years agoARM: dts: DRA7: Add interrupts property to mailbox nodes
Suman Anna [Mon, 3 Nov 2014 23:07:34 +0000 (17:07 -0600)]
ARM: dts: DRA7: Add interrupts property to mailbox nodes

Add the interrupts property to all the 13 mailbox nodes in
DRA7xx. The interrupts property information added is inline
with the expected values with the DRA7xx crossbar driver,
and is common to both DRA74x and DRA72x SoCs.

Do note that the mailbox 1 is only capable of generating out
3 interrupts, while all the remaining mailboxes have 4
interrupts each.

Signed-off-by: Suman Anna <s-anna@ti.com>
Signed-off-by: Tony Lindgren <tony@atomide.com>
10 years agoARM: dts: Fix missing GPMC NAND device width for omap3 boards
Tony Lindgren [Tue, 4 Nov 2014 01:42:16 +0000 (17:42 -0800)]
ARM: dts: Fix missing GPMC NAND device width for omap3 boards

Looks like we have some GPMC NAND timings missing device
width. This fixes "gpmc_cs_program_settings: invalid width 0!"
errors during boot.

Signed-off-by: Tony Lindgren <tony@atomide.com>
10 years agoARM: dts: Use better omap GPMC timings for LAN9220
Tony Lindgren [Tue, 4 Nov 2014 00:48:16 +0000 (16:48 -0800)]
ARM: dts: Use better omap GPMC timings for LAN9220

With the GPMC warnings now enabled, I noticed the LAN9220 timings
can overflow the GPMC registers with 200MHz L3 speed. Earlier we
were just skipping the bad timings and would continue with the
bootloader timings. Now we no longer allow to continue with bad
timings as we have the timings in the .dts files.

We could start using the GPMC clock divider, but let's instead
use the u-boot timings that are known to be working and a bit
faster. These are basically the u-boot NET_GPMC_CONFIG[1-6]
defines deciphered. Except that we don't set gpmc,burst-length
as that's only partially configured and does not seem to work
if fully enabled.

[tony@atomide.com: updated to remove gpmc,burst-length]
Signed-off-by: Tony Lindgren <tony@atomide.com>
10 years agoARM: dts: Add GPMC timings for omap zoom serial port
Tony Lindgren [Thu, 30 Oct 2014 00:16:48 +0000 (17:16 -0700)]
ARM: dts: Add GPMC timings for omap zoom serial port

The four port serial port on the zoom debug board uses a TL16CP754C
with a single interrupt and GPMC chip select. The serial ports each
use a 8 bytes for IO registers, and are 256 bytes apart on the GPMC
line.

Let's add timings for all four ports so we can remove the GPMC
workarounds for using bootloader timings.

Not caused by this patch, but looks like u-boot only properly
initializes the fifo on the first serial port. Currently the other
ports produce garbage at least with my version of u-boot. I suspect
that TL16CP754C needs non-standard initialization added to 8250
driver to properly fix this issue.

Cc: Roger Quadros <rogerq@ti.com>
Signed-off-by: Tony Lindgren <tony@atomide.com>
10 years agoARM: dts: Add smc91x GPMC configuration for 2430sdp
Tony Lindgren [Thu, 30 Oct 2014 00:16:47 +0000 (17:16 -0700)]
ARM: dts: Add smc91x GPMC configuration for 2430sdp

Let's use the bootloader values except for the partially configured
wait-pin that does not seem to work.

Cc: Roger Quadros <rogerq@ti.com>
Signed-off-by: Tony Lindgren <tony@atomide.com>
10 years agoARM: dts: Fix wrong GPMC size mappings for omaps
Tony Lindgren [Thu, 30 Oct 2014 00:16:47 +0000 (17:16 -0700)]
ARM: dts: Fix wrong GPMC size mappings for omaps

The GPMC binding is obviously very confusing as the values
are all over the place. People seem to confuse the GPMC partition
size for the chip select, and the device IO size within the GPMC
partition easily.

The ranges entry contains the GPMC partition size. And the
reg entry contains the size of the IO registers of the
device connected to the GPMC.

Let's fix the issue according to the following table:

Device          GPMC partition size     Device IO size
connected       in the ranges entry     in the reg entry

NAND            0x01000000 (16MB)       4
16550           0x01000000 (16MB)       8
smc91x          0x01000000 (16MB)       0xf
smc911x         0x01000000 (16MB)       0xff
OneNAND         0x01000000 (16MB)       0x20000 (128KB)
16MB NOR        0x01000000 (16MB)       0x01000000 (16MB)
32MB NOR        0x02000000 (32MB)       0x02000000 (32MB)
64MB NOR        0x04000000 (64MB)       0x04000000 (64MB)
128MB NOR       0x08000000 (128MB)      0x08000000 (128MB)
256MB NOR       0x10000000 (256MB)      0x10000000 (256MB)

Let's also add comments to the fixed entries while at it.

Acked-by: Roger Quadros <rogerq@ti.com>
Signed-off-by: Tony Lindgren <tony@atomide.com>
10 years agoARM: dts: Fix bootloader version dependencies by muxing n900 smc91x pins
Tony Lindgren [Thu, 30 Oct 2014 00:16:47 +0000 (17:16 -0700)]
ARM: dts: Fix bootloader version dependencies by muxing n900 smc91x pins

Apparently some versions of nolo don't mux the all the necessary GPMC
pins for the smc91x probe to work properly. Let's fix this issue
by adding mux support for GPMC to the kernel.

Note that GPMC clk needs input enabled for OnenNAND to work.

Cc: Kevin Hilman <khilman@kernel.org>
Cc: Roger Quadros <rogerq@ti.com>
Signed-off-by: Tony Lindgren <tony@atomide.com>
10 years agoLinux 3.18-rc2
Linus Torvalds [Sun, 26 Oct 2014 23:48:41 +0000 (16:48 -0700)]
Linux 3.18-rc2

10 years agoMerge tag 'armsoc-for-rc2' of git://git.kernel.org/pub/scm/linux/kernel/git/arm/arm-soc
Linus Torvalds [Sun, 26 Oct 2014 18:35:51 +0000 (11:35 -0700)]
Merge tag 'armsoc-for-rc2' of git://git./linux/kernel/git/arm/arm-soc

Pull ARM SoC fixes from Olof Johansson:
 "Another week, another small batch of fixes.

  Most of these make zynq, socfpga and sunxi platforms work a bit
  better:

   - due to new requirements for regulators, DWMMC on socfpga broke past
     v3.17
   - SMP spinup fix for socfpga
   - a few DT fixes for zynq
   - another option (FIXED_REGULATOR) for sunxi is needed that used to
     be selected by other options but no longer is.
   - a couple of small DT fixes for at91
   - ...and a couple for i.MX"

* tag 'armsoc-for-rc2' of git://git.kernel.org/pub/scm/linux/kernel/git/arm/arm-soc:
  ARM: dts: imx28-evk: Let i2c0 run at 100kHz
  ARM: i.MX6: Fix "emi" clock name typo
  ARM: multi_v7_defconfig: enable CONFIG_MMC_DW_ROCKCHIP
  ARM: sunxi_defconfig: enable CONFIG_REGULATOR_FIXED_VOLTAGE
  ARM: dts: socfpga: Add a 3.3V fixed regulator node
  ARM: dts: socfpga: Fix SD card detect
  ARM: dts: socfpga: rename gpio nodes
  ARM: at91/dt: sam9263: fix PLLB frequencies
  power: reset: at91-reset: fix power down register
  MAINTAINERS: add atmel ssc driver maintainer entry
  arm: socfpga: fix fetching cpu1start_addr for SMP
  ARM: zynq: DT: trivial: Fix mc node
  ARM: zynq: DT: Add cadence watchdog node
  ARM: zynq: DT: Add missing reference for memory-controller
  ARM: zynq: DT: Add missing reference for ADC
  ARM: zynq: DT: Add missing address for L2 pl310
  ARM: zynq: DT: Remove 222 MHz OPP
  ARM: zynq: DT: Fix GEM register area size

10 years agoMerge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/viro/vfs
Linus Torvalds [Sun, 26 Oct 2014 18:19:18 +0000 (11:19 -0700)]
Merge branch 'for-linus' of git://git./linux/kernel/git/viro/vfs

Pull vfs updates from Al Viro:
 "overlayfs merge + leak fix for d_splice_alias() failure exits"

* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/viro/vfs:
  overlayfs: embed middle into overlay_readdir_data
  overlayfs: embed root into overlay_readdir_data
  overlayfs: make ovl_cache_entry->name an array instead of pointer
  overlayfs: don't hold ->i_mutex over opening the real directory
  fix inode leaks on d_splice_alias() failure exits
  fs: limit filesystem stacking depth
  overlay: overlay filesystem documentation
  overlayfs: implement show_options
  overlayfs: add statfs support
  overlay filesystem
  shmem: support RENAME_WHITEOUT
  ext4: support RENAME_WHITEOUT
  vfs: add RENAME_WHITEOUT
  vfs: add whiteout support
  vfs: export check_sticky()
  vfs: introduce clone_private_mount()
  vfs: export __inode_permission() to modules
  vfs: export do_splice_direct() to modules
  vfs: add i_op->dentry_open()

10 years agoMerge tag 'imx-fixes-3.18' of git://git.kernel.org/pub/scm/linux/kernel/git/shawnguo...
Olof Johansson [Sun, 26 Oct 2014 03:44:05 +0000 (20:44 -0700)]
Merge tag 'imx-fixes-3.18' of git://git./linux/kernel/git/shawnguo/linux into fixes

Merge "ARM: imx: fixes for 3.18" from Shawn Guo:

The i.MX fixes for 3.18:
 - Revert one patch which increases I2C bus frequency on imx28-evk
 - Fix a typo on imx6q EIM clock name

* tag 'imx-fixes-3.18' of git://git.kernel.org/pub/scm/linux/kernel/git/shawnguo/linux:
  ARM: dts: imx28-evk: Let i2c0 run at 100kHz
  ARM: i.MX6: Fix "emi" clock name typo

Signed-off-by: Olof Johansson <olof@lixom.net>
10 years agoARM: dts: imx28-evk: Let i2c0 run at 100kHz
Fabio Estevam [Mon, 20 Oct 2014 13:08:01 +0000 (11:08 -0200)]
ARM: dts: imx28-evk: Let i2c0 run at 100kHz

Commit 78b81f4666fb ("ARM: dts: imx28-evk: Run I2C0 at 400kHz") caused issues
when doing the following sequence in loop:

- Boot the kernel
- Perform audio playback
- Reboot the system via 'reboot' command

In many times the audio card cannot be probed, which causes playback to fail.

After restoring to the original i2c0 frequency of 100kHz there is no such
problem anymore.

This reverts commit 78b81f4666fbb22a20b1e63e5baf197ad2e90e88.

Cc: <stable@vger.kernel.org> # 3.16+
Signed-off-by: Fabio Estevam <fabio.estevam@freescale.com>
Signed-off-by: Shawn Guo <shawn.guo@linaro.org>
10 years agoARM: i.MX6: Fix "emi" clock name typo
Steve Longerbeam [Tue, 14 Oct 2014 17:41:49 +0000 (20:41 +0300)]
ARM: i.MX6: Fix "emi" clock name typo

Fix a typo error, the "emi" names refer to the eim clocks.

The change fixes typo in EIM and EIM_SLOW pre-output dividers and
selectors clock names. Notably EIM_SLOW clock itself is named correctly.

Signed-off-by: Steve Longerbeam <steve_longerbeam@mentor.com>
[vladimir_zapolskiy@mentor.com: ported to v3.17]
Signed-off-by: Vladimir Zapolskiy <vladimir_zapolskiy@mentor.com>
Cc: Sascha Hauer <kernel@pengutronix.de>
Signed-off-by: Shawn Guo <shawn.guo@linaro.org>
10 years agooverlayfs: embed middle into overlay_readdir_data
Al Viro [Fri, 24 Oct 2014 03:03:03 +0000 (23:03 -0400)]
overlayfs: embed middle into overlay_readdir_data

same story...

Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
10 years agooverlayfs: embed root into overlay_readdir_data
Al Viro [Fri, 24 Oct 2014 03:00:53 +0000 (23:00 -0400)]
overlayfs: embed root into overlay_readdir_data

no sense having it a pointer - all instances have it pointing to
local variable in the same stack frame

Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
10 years agooverlayfs: make ovl_cache_entry->name an array instead of pointer
Al Viro [Fri, 24 Oct 2014 02:58:56 +0000 (22:58 -0400)]
overlayfs: make ovl_cache_entry->name an array instead of pointer

Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
10 years agooverlayfs: don't hold ->i_mutex over opening the real directory
Al Viro [Fri, 24 Oct 2014 02:56:05 +0000 (22:56 -0400)]
overlayfs: don't hold ->i_mutex over opening the real directory

just use it to serialize the assignment

Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
10 years agoMerge branch 'upstream' of git://git.linux-mips.org/pub/scm/ralf/upstream-linus
Linus Torvalds [Fri, 24 Oct 2014 19:48:47 +0000 (12:48 -0700)]
Merge branch 'upstream' of git://git.linux-mips.org/ralf/upstream-linus

Pull MIPS fixes from Ralf Baechle:
 "This is the first round of fixes and tying up loose ends for MIPS.

   - plenty of fixes for build errors in specific obscure configurations
   - remove redundant code on the Lantiq platform
   - removal of a useless SEAD I2C driver that was causing a build issue
   - fix an earlier TLB exeption handler fix to also work on Octeon.
   - fix ISA level dependencies in FPU emulator's instruction decoding.
   - don't hardcode kernel command line in Octeon software emulator.
   - fix an earlier fix for the Loondson 2 clock setting"

* 'upstream' of git://git.linux-mips.org/pub/scm/ralf/upstream-linus:
  MIPS: SEAD3: Fix I2C device registration.
  MIPS: SEAD3: Nuke PIC32 I2C driver.
  MIPS: ftrace: Fix a microMIPS build problem
  MIPS: MSP71xx: Fix build error
  MIPS: Malta: Do not build the malta-amon.c file if CMP is not enabled
  MIPS: Prevent compiler warning from cop2_{save,restore}
  MIPS: Kconfig: Add missing MIPS_CPS dependencies to PM and cpuidle
  MIPS: idle: Remove leftover __pastwait symbol and its references
  MIPS: Sibyte: Include the swarm subdir to the sb1250 LittleSur builds
  MIPS: ptrace.h: Add a missing include
  MIPS: ath79: Fix compilation error when CONFIG_PCI is disabled
  MIPS: MSP71xx: Remove compilation error when CONFIG_MIPS_MT is present
  MIPS: Octeon: Remove special case for simulator command line.
  MIPS: tlbex: Properly fix HUGE TLB Refill exception handler
  MIPS: loongson2_cpufreq: Fix CPU clock rate setting mismerge
  pci: pci-lantiq: remove duplicate check on resource
  MIPS: Lasat: Add missing CONFIG_PROC_FS dependency to PICVUE_PROC
  MIPS: cp1emu: Fix ISA restrictions for cop1x_op instructions

10 years agoMerge tag 'arm64-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/arm64/linux
Linus Torvalds [Fri, 24 Oct 2014 19:48:04 +0000 (12:48 -0700)]
Merge tag 'arm64-fixes' of git://git./linux/kernel/git/arm64/linux

Pull arm64 fixes from Catalin Marinas:

 - enable 48-bit VA space now that KVM has been fixed, together with a
   couple of fixes for pgd allocation alignment and initial memblock
   current_limit.  There is still a dependency on !ARM_SMMU which needs
   to be updated as it uses the page table manipulation macros of the
   host kernel
 - eBPF fixes following changes/conflicts during the merging window
 - Compat types affecting compat_elf_prpsinfo
 - Compilation error on UP builds
 - ASLR fix when /proc/sys/kernel/randomize_va_space == 0
 - DT definitions for CLCD support on ARMv8 model platform

* tag 'arm64-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/arm64/linux:
  arm64: Fix memblock current_limit with 64K pages and 48-bit VA
  arm64: ASLR: Don't randomise text when randomise_va_space == 0
  arm64: vexpress: Add CLCD support to the ARMv8 model platform
  arm64: Fix compilation error on UP builds
  Documentation/arm64/memory.txt: fix typo
  net: bpf: arm64: minor fix of type in jited
  arm64: bpf: add 'load 64-bit immediate' instruction
  arm64: bpf: add 'shift by register' instructions
  net: bpf: arm64: address randomize and write protect JIT code
  arm64: mm: Correct fixmap pagetable types
  arm64: compat: fix compat types affecting struct compat_elf_prpsinfo
  arm64: Align less than PAGE_SIZE pgds naturally
  arm64: Allow 48-bits VA space without ARM_SMMU

10 years agoMerge git://git.kernel.org/pub/scm/linux/kernel/git/davem/sparc
Linus Torvalds [Fri, 24 Oct 2014 19:45:47 +0000 (12:45 -0700)]
Merge git://git./linux/kernel/git/davem/sparc

Pull two sparc fixes from David Miller:

 1) Fix boots with gcc-4.9 compiled sparc64 kernels.

 2) Add missing __get_user_pages_fast() on sparc64 to fix hangs on
    futexes used in transparent hugepage areas.

    It's really idiotic to have a weak symbolled fallback that just
    returns zero, and causes this kind of bug.  There should be no
    backup implementation and the link should fail if the architecture
    fails to provide __get_user_pages_fast() and supports transparent
    hugepages.

* git://git.kernel.org/pub/scm/linux/kernel/git/davem/sparc:
  sparc64: Implement __get_user_pages_fast().
  sparc64: Fix register corruption in top-most kernel stack frame during boot.

10 years agoMerge tag 'for-linus' of git://git.kernel.org/pub/scm/virt/kvm/kvm
Linus Torvalds [Fri, 24 Oct 2014 19:42:55 +0000 (12:42 -0700)]
Merge tag 'for-linus' of git://git./virt/kvm/kvm

Pull kvm fixes from Paolo Bonzini:
 "This is a pretty large update.  I think it is roughly as big as what I
  usually had for the _whole_ rc period.

  There are a few bad bugs where the guest can OOPS or crash the host.
  We have also started looking at attack models for nested
  virtualization; bugs that usually result in the guest ring 0 crashing
  itself become more worrisome if you have nested virtualization,
  because the nested guest might bring down the non-nested guest as
  well.  For current uses of nested virtualization these do not really
  have a security impact, but you never know and bugs are bugs
  nevertheless.

  A lot of these bugs are in 3.17 too, resulting in a large number of
  stable@ Ccs.  I checked that all the patches apply there with no
  conflicts"

* tag 'for-linus' of git://git.kernel.org/pub/scm/virt/kvm/kvm:
  kvm: vfio: fix unregister kvm_device_ops of vfio
  KVM: x86: Wrong assertion on paging_tmpl.h
  kvm: fix excessive pages un-pinning in kvm_iommu_map error path.
  KVM: x86: PREFETCH and HINT_NOP should have SrcMem flag
  KVM: x86: Emulator does not decode clflush well
  KVM: emulate: avoid accessing NULL ctxt->memopp
  KVM: x86: Decoding guest instructions which cross page boundary may fail
  kvm: x86: don't kill guest on unknown exit reason
  kvm: vmx: handle invvpid vm exit gracefully
  KVM: x86: Handle errors when RIP is set during far jumps
  KVM: x86: Emulator fixes for eip canonical checks on near branches
  KVM: x86: Fix wrong masking on relative jump/call
  KVM: x86: Improve thread safety in pit
  KVM: x86: Prevent host from panicking on shared MSR writes.
  KVM: x86: Check non-canonical addresses upon WRMSR

10 years agoMerge tag 'stable/for-linus-3.18-b-rc1-tag' of git://git.kernel.org/pub/scm/linux...
Linus Torvalds [Fri, 24 Oct 2014 19:41:50 +0000 (12:41 -0700)]
Merge tag 'stable/for-linus-3.18-b-rc1-tag' of git://git./linux/kernel/git/xen/tip

Pull xen bug fixes from David Vrabel:

 - Fix regression in xen_clocksource_read() which caused all Xen guests
   to crash early in boot.
 - Several fixes for super rare race conditions in the p2m.
 - Assorted other minor fixes.

* tag 'stable/for-linus-3.18-b-rc1-tag' of git://git.kernel.org/pub/scm/linux/kernel/git/xen/tip:
  xen/pci: Allocate memory for physdev_pci_device_add's optarr
  x86/xen: panic on bad Xen-provided memory map
  x86/xen: Fix incorrect per_cpu accessor in xen_clocksource_read()
  x86/xen: avoid race in p2m handling
  x86/xen: delay construction of mfn_list_list
  x86/xen: avoid writing to freed memory after race in p2m handling
  xen/balloon: Don't continue ballooning when BP_ECANCELED is encountered

10 years agoMerge tag 'sound-3.18-rc2' of git://git.kernel.org/pub/scm/linux/kernel/git/tiwai...
Linus Torvalds [Fri, 24 Oct 2014 19:35:48 +0000 (12:35 -0700)]
Merge tag 'sound-3.18-rc2' of git://git./linux/kernel/git/tiwai/sound

Pull sound fixes from Takashi Iwai:
 "Here are a chunk of small fixes since rc1: two PCM core fixes, one is
  a long-standing annoyance about lockdep and another is an ARM64 mmap
  fix.

  The rest are a HD-audio HDMI hotplug notification fix, a fix for
  missing NULL termination in Realtek codec quirks and a few new
  device/codec-specific quirks as usual"

* tag 'sound-3.18-rc2' of git://git.kernel.org/pub/scm/linux/kernel/git/tiwai/sound:
  ALSA: hda - Add missing terminating entry to SND_HDA_PIN_QUIRK macro
  ALSA: pcm: Fix false lockdep warnings
  ALSA: hda - Fix inverted LED gpio setup for Lenovo Ideapad
  ALSA: hda - hdmi: Fix missing ELD change event on plug/unplug
  ALSA: usb-audio: Add support for Steinberg UR22 USB interface
  ALSA: ALC283 codec - Avoid pop noise on headphones during suspend/resume
  ALSA: pcm: use the same dma mmap codepath both for arm and arm64

10 years agoMerge tag 'random_for_linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tytso...
Linus Torvalds [Fri, 24 Oct 2014 19:33:32 +0000 (12:33 -0700)]
Merge tag 'random_for_linus' of git://git./linux/kernel/git/tytso/random

Pull /dev/random updates from Ted Ts'o:
 "This adds a memzero_explicit() call which is guaranteed not to be
  optimized away by GCC.  This is important when we are wiping
  cryptographically sensitive material"

* tag 'random_for_linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tytso/random:
  crypto: memzero_explicit - make sure to clear out sensitive data
  random: add and use memzero_explicit() for clearing data

10 years agoMerge tag 'pm+acpi-3.18-rc2' of git://git.kernel.org/pub/scm/linux/kernel/git/rafael...
Linus Torvalds [Fri, 24 Oct 2014 18:29:31 +0000 (11:29 -0700)]
Merge tag 'pm+acpi-3.18-rc2' of git://git./linux/kernel/git/rafael/linux-pm

Pull ACPI and power management updates from Rafael Wysocki:
 "This is material that didn't make it to my 3.18-rc1 pull request for
  various reasons, mostly related to timing and travel (LinuxCon EU /
  LPC) plus a couple of fixes for recent bugs.

  The only really new thing here is the PM QoS class for memory
  bandwidth, but it is simple enough and users of it will be added in
  the next cycle.  One major change in behavior is that platform devices
  enumerated by ACPI will use 32-bit DMA mask by default.  Also included
  is an ACPICA update to a new upstream release, but that's mostly
  cleanups, changes in tools and similar.  The rest is fixes and
  cleanups mostly.

  Specifics:

   - Fix for a recent PCI power management change that overlooked the
     fact that some IRQ chips might not be able to configure PCIe PME
     for system wakeup from Lucas Stach.

   - Fix for a bug introduced in 3.17 where acpi_device_wakeup() is
     called with a wrong ordering of arguments from Zhang Rui.

   - A bunch of intel_pstate driver fixes (all -stable candidates) from
     Dirk Brandewie, Gabriele Mazzotta and Pali Rohár.

   - Fixes for a rather long-standing problem with the OOM killer and
     the freezer that frozen processes killed by the OOM do not actually
     release any memory until they are thawed, so OOM-killing them is
     rather pointless, with a couple of cleanups on top (Michal Hocko,
     Cong Wang, Rafael J Wysocki).

   - ACPICA update to upstream release 20140926, inlcuding mostly
     cleanups reducing differences between the upstream ACPICA and the
     kernel code, tools changes (acpidump, acpiexec) and support for the
     _DDN object (Bob Moore, Lv Zheng).

   - New PM QoS class for memory bandwidth from Tomeu Vizoso.

   - Default 32-bit DMA mask for platform devices enumerated by ACPI
     (this change is mostly needed for some drivers development in
     progress targeted at 3.19) from Heikki Krogerus.

   - ACPI EC driver cleanups, mostly related to debugging, from Lv
     Zheng.

   - cpufreq-dt driver updates from Thomas Petazzoni.

   - powernv cpuidle driver update from Preeti U Murthy"

* tag 'pm+acpi-3.18-rc2' of git://git.kernel.org/pub/scm/linux/kernel/git/rafael/linux-pm: (34 commits)
  intel_pstate: Correct BYT VID values.
  intel_pstate: Fix BYT frequency reporting
  intel_pstate: Don't lose sysfs settings during cpu offline
  cpufreq: intel_pstate: Reflect current no_turbo state correctly
  cpufreq: expose scaling_cur_freq sysfs file for set_policy() drivers
  cpufreq: intel_pstate: Fix setting max_perf_pct in performance policy
  PCI / PM: handle failure to enable wakeup on PCIe PME
  ACPI: invoke acpi_device_wakeup() with correct parameters
  PM / freezer: Clean up code after recent fixes
  PM: convert do_each_thread to for_each_process_thread
  OOM, PM: OOM killed task shouldn't escape PM suspend
  freezer: remove obsolete comments in __thaw_task()
  freezer: Do not freeze tasks killed by OOM killer
  ACPI / platform: provide default DMA mask
  cpuidle: powernv: Populate cpuidle state details by querying the device-tree
  cpufreq: cpufreq-dt: adjust message related to regulators
  cpufreq: cpufreq-dt: extend with platform_data
  cpufreq: allow driver-specific data
  ACPI / EC: Cleanup coding style.
  ACPI / EC: Refine event/query debugging messages.
  ...

10 years agoMerge branch 'next' of git://git.kernel.org/pub/scm/linux/kernel/git/rzhang/linux
Linus Torvalds [Fri, 24 Oct 2014 18:21:43 +0000 (11:21 -0700)]
Merge branch 'next' of git://git./linux/kernel/git/rzhang/linux

Pull thermal management updates from Zhang Rui:
 "Sorry that I missed the merge window as there is a bug found in the
  last minute, and I have to fix it and wait for the code to be tested
  in linux-next tree for a few days.  Now the buggy patch has been
  dropped entirely from my next branch.  Thus I hope those changes can
  still be merged in 3.18-rc2 as most of them are platform thermal
  driver changes.

  Specifics:

   - introduce ACPI INT340X thermal drivers.

     Newer laptops and tablets may have thermal sensors and other
     devices with thermal control capabilities that are exposed for the
     OS to use via the ACPI INT340x device objects.  Several drivers are
     introduced to expose the temperature information and cooling
     ability from these objects to user-space via the normal thermal
     framework.

     From: Lu Aaron, Lan Tianyu, Jacob Pan and Zhang Rui.

   - introduce a new thermal governor, which just uses a hysteresis to
     switch abruptly on/off a cooling device.  This governor can be used
     to control certain fan devices that can not be throttled but just
     switched on or off.  From: Peter Feuerer.

   - introduce support for some new thermal interrupt functions on
     i.MX6SX, in IMX thermal driver.  From: Anson, Huang.

   - introduce tracing support on thermal framework.  From: Punit
     Agrawal.

   - small fixes in OF thermal and thermal step_wise governor"

* 'next' of git://git.kernel.org/pub/scm/linux/kernel/git/rzhang/linux: (25 commits)
  Thermal: int340x thermal: select ACPI fan driver
  Thermal: int3400_thermal: use acpi_thermal_rel parsing APIs
  Thermal: int340x_thermal: expose acpi thermal relationship tables
  Thermal: introduce int3403 thermal driver
  Thermal: introduce INT3402 thermal driver
  Thermal: move the KELVIN_TO_MILLICELSIUS macro to thermal.h
  ACPI / Fan: support INT3404 thermal device
  ACPI / Fan: add ACPI 4.0 style fan support
  ACPI / fan: convert to platform driver
  ACPI / fan: use acpi_device_xxx_power instead of acpi_bus equivelant
  ACPI / fan: remove no need check for device pointer
  ACPI / fan: remove unused macro
  Thermal: int3400 thermal: register to thermal framework
  Thermal: int3400 thermal: add capability to detect supporting UUIDs
  Thermal: introduce int3400 thermal driver
  ACPI: add ACPI_TYPE_LOCAL_REFERENCE support to acpi_extract_package()
  ACPI: make acpi_create_platform_device() an external API
  thermal: step_wise: fix: Prevent from binary overflow when trend is dropping
  ACPI: introduce ACPI int340x thermal scan handler
  thermal: Added Bang-bang thermal governor
  ...

10 years agoarm64: Fix memblock current_limit with 64K pages and 48-bit VA
Catalin Marinas [Fri, 24 Oct 2014 17:16:47 +0000 (18:16 +0100)]
arm64: Fix memblock current_limit with 64K pages and 48-bit VA

With 48-bit VA space, the 64K page configuration uses 3 levels instead
of 2 and PUD_SIZE != PMD_SIZE. Since with 64K pages we only cover
PMD_SIZE with the initial swapper_pg_dir populated in head.S, the
memblock current_limit needs to be set accordingly in map_mem() to avoid
allocating unmapped memory. The memblock current_limit is progressively
increased as more blocks are mapped.

Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
10 years agosparc64: Implement __get_user_pages_fast().
David S. Miller [Fri, 24 Oct 2014 16:59:02 +0000 (09:59 -0700)]
sparc64: Implement __get_user_pages_fast().

It is not sufficient to only implement get_user_pages_fast(), you
must also implement the atomic version __get_user_pages_fast()
otherwise you end up using the weak symbol fallback implementation
which simply returns zero.

This is dangerous, because it causes the futex code to loop forever
if transparent hugepages are supported (see get_futex_key()).

Signed-off-by: David S. Miller <davem@davemloft.net>
10 years agosparc64: Fix register corruption in top-most kernel stack frame during boot.
David S. Miller [Thu, 23 Oct 2014 19:58:13 +0000 (12:58 -0700)]
sparc64: Fix register corruption in top-most kernel stack frame during boot.

Meelis Roos reported that kernels built with gcc-4.9 do not boot, we
eventually narrowed this down to only impacting machines using
UltraSPARC-III and derivitive cpus.

The crash happens right when the first user process is spawned:

[   54.451346] Kernel panic - not syncing: Attempted to kill init! exitcode=0x00000004
[   54.451346]
[   54.571516] CPU: 1 PID: 1 Comm: init Not tainted 3.16.0-rc2-00211-gd7933ab #96
[   54.666431] Call Trace:
[   54.698453]  [0000000000762f8c] panic+0xb0/0x224
[   54.759071]  [000000000045cf68] do_exit+0x948/0x960
[   54.823123]  [000000000042cbc0] fault_in_user_windows+0xe0/0x100
[   54.902036]  [0000000000404ad0] __handle_user_windows+0x0/0x10
[   54.978662] Press Stop-A (L1-A) to return to the boot prom
[   55.050713] ---[ end Kernel panic - not syncing: Attempted to kill init! exitcode=0x00000004

Further investigation showed that compiling only per_cpu_patch() with
an older compiler fixes the boot.

Detailed analysis showed that the function is not being miscompiled by
gcc-4.9, but it is using a different register allocation ordering.

With the gcc-4.9 compiled function, something during the code patching
causes some of the %i* input registers to get corrupted.  Perhaps
we have a TLB miss path into the firmware that is deep enough to
cause a register window spill and subsequent restore when we get
back from the TLB miss trap.

Let's plug this up by doing two things:

1) Stop using the firmware stack for client interface calls into
   the firmware.  Just use the kernel's stack.

2) As soon as we can, call into a new function "start_early_boot()"
   to put a one-register-window buffer between the firmware's
   deepest stack frame and the top-most initial kernel one.

Reported-by: Meelis Roos <mroos@linux.ee>
Tested-by: Meelis Roos <mroos@linux.ee>
Signed-off-by: David S. Miller <davem@davemloft.net>
10 years agoarm64: ASLR: Don't randomise text when randomise_va_space == 0
Arun Chandran [Fri, 10 Oct 2014 11:31:24 +0000 (12:31 +0100)]
arm64: ASLR: Don't randomise text when randomise_va_space == 0

When user asks to turn off ASLR by writing "0" to
/proc/sys/kernel/randomize_va_space there should not be
any randomization to mmap base, stack, VDSO, libs, text and heap

Currently arm64 violates this behavior by randomising text.
Fix this by defining a constant ELF_ET_DYN_BASE. The randomisation of
mm->mmap_base is done by setup_new_exec -> arch_pick_mmap_layout ->
mmap_base -> mmap_rnd.

Signed-off-by: Arun Chandran <achandran@mvista.com>
Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
10 years agoMIPS: SEAD3: Fix I2C device registration.
Ralf Baechle [Fri, 24 Oct 2014 10:23:26 +0000 (12:23 +0200)]
MIPS: SEAD3: Fix I2C device registration.

This isn't a module and shouldn't be one.

Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
10 years agokvm: vfio: fix unregister kvm_device_ops of vfio
Wanpeng Li [Thu, 9 Oct 2014 10:30:08 +0000 (18:30 +0800)]
kvm: vfio: fix unregister kvm_device_ops of vfio

After commit 80ce163 (KVM: VFIO: register kvm_device_ops dynamically),
kvm_device_ops of vfio can be registered dynamically. Commit 3c3c29fd
(kvm-vfio: do not use module_init) move the dynamic register invoked by
kvm_init in order to fix broke unloading of the kvm module. However,
kvm_device_ops of vfio is unregistered after rmmod kvm-intel module
which lead to device type collision detection warning after kvm-intel
module reinsmod.

    WARNING: CPU: 1 PID: 10358 at /root/cathy/kvm/arch/x86/kvm/../../../virt/kvm/kvm_main.c:3289 kvm_init+0x234/0x282 [kvm]()
    Modules linked in: kvm_intel(O+) kvm(O) nfsv3 nfs_acl auth_rpcgss oid_registry nfsv4 dns_resolver nfs fscache lockd sunrpc pci_stub bridge stp llc autofs4 8021q cpufreq_ondemand ipv6 joydev microcode pcspkr igb i2c_algo_bit ehci_pci ehci_hcd e1000e i2c_i801 ixgbe ptp pps_core hwmon mdio tpm_tis tpm ipmi_si ipmi_msghandler acpi_cpufreq isci libsas scsi_transport_sas button dm_mirror dm_region_hash dm_log dm_mod [last unloaded: kvm_intel]
    CPU: 1 PID: 10358 Comm: insmod Tainted: G        W  O   3.17.0-rc1 #2
    Hardware name: Intel Corporation S2600CP/S2600CP, BIOS RMLSDP.86I.00.29.D696.1311111329 11/11/2013
     0000000000000cd9 ffff880ff08cfd18 ffffffff814a61d9 0000000000000cd9
     0000000000000000 ffff880ff08cfd58 ffffffff810417b7 ffff880ff08cfd48
     ffffffffa045bcac ffffffffa049c420 0000000000000040 00000000000000ff
    Call Trace:
     [<ffffffff814a61d9>] dump_stack+0x49/0x60
     [<ffffffff810417b7>] warn_slowpath_common+0x7c/0x96
     [<ffffffffa045bcac>] ? kvm_init+0x234/0x282 [kvm]
     [<ffffffff810417e6>] warn_slowpath_null+0x15/0x17
     [<ffffffffa045bcac>] kvm_init+0x234/0x282 [kvm]
     [<ffffffffa016e995>] vmx_init+0x1bf/0x42a [kvm_intel]
     [<ffffffffa016e7d6>] ? vmx_check_processor_compat+0x64/0x64 [kvm_intel]
     [<ffffffff810002ab>] do_one_initcall+0xe3/0x170
     [<ffffffff811168a9>] ? __vunmap+0xad/0xb8
     [<ffffffff8109c58f>] do_init_module+0x2b/0x174
     [<ffffffff8109d414>] load_module+0x43e/0x569
     [<ffffffff8109c6d8>] ? do_init_module+0x174/0x174
     [<ffffffff8109c75a>] ? copy_module_from_user+0x39/0x82
     [<ffffffff8109b7dd>] ? module_sect_show+0x20/0x20
     [<ffffffff8109d65f>] SyS_init_module+0x54/0x81
     [<ffffffff814a9a12>] system_call_fastpath+0x16/0x1b
    ---[ end trace 0626f4a3ddea56f3 ]---

The bug can be reproduced by:

    rmmod kvm_intel.ko
    insmod kvm_intel.ko

without rmmod/insmod kvm.ko
This patch fixes the bug by unregistering kvm_device_ops of vfio when the
kvm-intel module is removed.

Reported-by: Liu Rongrong <rongrongx.liu@intel.com>
Fixes: 3c3c29fd0d7cddc32862c350d0700ce69953e3bd
Signed-off-by: Wanpeng Li <wanpeng.li@linux.intel.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
10 years agoKVM: x86: Wrong assertion on paging_tmpl.h
Nadav Amit [Tue, 30 Sep 2014 17:49:18 +0000 (20:49 +0300)]
KVM: x86: Wrong assertion on paging_tmpl.h

Even after the recent fix, the assertion on paging_tmpl.h is triggered.
Apparently, the assertion wants to check that the PAE is always set on
long-mode, but does it in incorrect way.  Note that the assertion is not
enabled unless the code is debugged by defining MMU_DEBUG.

Signed-off-by: Nadav Amit <namit@cs.technion.ac.il>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
10 years agokvm: fix excessive pages un-pinning in kvm_iommu_map error path.
Quentin Casasnovas [Fri, 17 Oct 2014 20:55:59 +0000 (22:55 +0200)]
kvm: fix excessive pages un-pinning in kvm_iommu_map error path.

The third parameter of kvm_unpin_pages() when called from
kvm_iommu_map_pages() is wrong, it should be the number of pages to un-pin
and not the page size.

This error was facilitated with an inconsistent API: kvm_pin_pages() takes
a size, but kvn_unpin_pages() takes a number of pages, so fix the problem
by matching the two.

This was introduced by commit 350b8bd ("kvm: iommu: fix the third parameter
of kvm_iommu_put_pages (CVE-2014-3601)"), which fixes the lack of
un-pinning for pages intended to be un-pinned (i.e. memory leak) but
unfortunately potentially aggravated the number of pages we un-pin that
should have stayed pinned. As far as I understand though, the same
practical mitigations apply.

This issue was found during review of Red Hat 6.6 patches to prepare
Ksplice rebootless updates.

Thanks to Vegard for his time on a late Friday evening to help me in
understanding this code.

Fixes: 350b8bd ("kvm: iommu: fix the third parameter of... (CVE-2014-3601)")
Cc: stable@vger.kernel.org
Signed-off-by: Quentin Casasnovas <quentin.casasnovas@oracle.com>
Signed-off-by: Vegard Nossum <vegard.nossum@oracle.com>
Signed-off-by: Jamie Iles <jamie.iles@oracle.com>
Reviewed-by: Sasha Levin <sasha.levin@oracle.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
10 years agoKVM: x86: PREFETCH and HINT_NOP should have SrcMem flag
Nadav Amit [Mon, 13 Oct 2014 10:04:14 +0000 (13:04 +0300)]
KVM: x86: PREFETCH and HINT_NOP should have SrcMem flag

The decode phase of the x86 emulator assumes that every instruction with the
ModRM flag, and which can be used with RIP-relative addressing, has either
SrcMem or DstMem.  This is not the case for several instructions - prefetch,
hint-nop and clflush.

Adding SrcMem|NoAccess for prefetch and hint-nop and SrcMem for clflush.

This fixes CVE-2014-8480.

Fixes: 41061cdb98a0bec464278b4db8e894a3121671f5
Cc: stable@vger.kernel.org
Signed-off-by: Nadav Amit <namit@cs.technion.ac.il>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
10 years agoKVM: x86: Emulator does not decode clflush well
Nadav Amit [Mon, 13 Oct 2014 10:04:13 +0000 (13:04 +0300)]
KVM: x86: Emulator does not decode clflush well

Currently, all group15 instructions are decoded as clflush (e.g., mfence,
xsave).  In addition, the clflush instruction requires no prefix (66/f2/f3)
would exist. If prefix exists it may encode a different instruction (e.g.,
clflushopt).

Creating a group for clflush, and different group for each prefix.

This has been the case forever, but the next patch needs the cflush group
in order to fix a bug introduced in 3.17.

Fixes: 41061cdb98a0bec464278b4db8e894a3121671f5
Cc: stable@vger.kernel.org
Signed-off-by: Nadav Amit <namit@cs.technion.ac.il>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
10 years agoKVM: emulate: avoid accessing NULL ctxt->memopp
Paolo Bonzini [Thu, 23 Oct 2014 12:54:14 +0000 (14:54 +0200)]
KVM: emulate: avoid accessing NULL ctxt->memopp

A failure to decode the instruction can cause a NULL pointer access.
This is fixed simply by moving the "done" label as close as possible
to the return.

This fixes CVE-2014-8481.

Reported-by: Andy Lutomirski <luto@amacapital.net>
Cc: stable@vger.kernel.org
Fixes: 41061cdb98a0bec464278b4db8e894a3121671f5
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
10 years agoMIPS: SEAD3: Nuke PIC32 I2C driver.
Ralf Baechle [Fri, 24 Oct 2014 11:27:37 +0000 (13:27 +0200)]
MIPS: SEAD3: Nuke PIC32 I2C driver.

A platform driver for which nothing ever registers the corresponding
platform device.

Also it was driving the same hardware as sead3-i2c-drv.c so redundant
anyway and couldn't co-exist with that driver because each of them was
using a private spinlock to protect access to the same hardware
resources.

This also fixes a randconfig problem:

arch/mips/mti-sead3/sead3-pic32-i2c-drv.c: In function 'i2c_platform_probe':
arch/mips/mti-sead3/sead3-pic32-i2c-drv.c:345:2: error: implicit declaration of
function 'i2c_add_numbered_adapter' [-Werror=implicit-function-declaration]
  ret = i2c_add_numbered_adapter(&priv->adap);
    ^
arch/mips/mti-sead3/sead3-pic32-i2c-drv.c: In function
'i2c_platform_remove':
arch/mips/mti-sead3/sead3-pic32-i2c-drv.c:361:2: error: implicit declaration
of function 'i2c_del_adapter' [-Werror=implicit-function-declaration]
i2c_del_adapter(&priv->adap);

Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
10 years agoKVM: x86: Decoding guest instructions which cross page boundary may fail
Nadav Amit [Thu, 2 Oct 2014 22:10:04 +0000 (01:10 +0300)]
KVM: x86: Decoding guest instructions which cross page boundary may fail

Once an instruction crosses a page boundary, the size read from the second page
disregards the common case that part of the operand resides on the first page.
As a result, fetch of long insturctions may fail, and thereby cause the
decoding to fail as well.

Cc: stable@vger.kernel.org
Fixes: 5cfc7e0f5e5e1adf998df94f8e36edaf5d30d38e
Signed-off-by: Nadav Amit <namit@cs.technion.ac.il>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
10 years agokvm: x86: don't kill guest on unknown exit reason
Michael S. Tsirkin [Thu, 18 Sep 2014 13:21:16 +0000 (16:21 +0300)]
kvm: x86: don't kill guest on unknown exit reason

KVM_EXIT_UNKNOWN is a kvm bug, we don't really know whether it was
triggered by a priveledged application.  Let's not kill the guest: WARN
and inject #UD instead.

Cc: stable@vger.kernel.org
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
10 years agokvm: vmx: handle invvpid vm exit gracefully
Petr Matousek [Tue, 23 Sep 2014 18:22:30 +0000 (20:22 +0200)]
kvm: vmx: handle invvpid vm exit gracefully

On systems with invvpid instruction support (corresponding bit in
IA32_VMX_EPT_VPID_CAP MSR is set) guest invocation of invvpid
causes vm exit, which is currently not handled and results in
propagation of unknown exit to userspace.

Fix this by installing an invvpid vm exit handler.

This is CVE-2014-3646.

Cc: stable@vger.kernel.org
Signed-off-by: Petr Matousek <pmatouse@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
10 years agoKVM: x86: Handle errors when RIP is set during far jumps
Nadav Amit [Thu, 18 Sep 2014 19:39:39 +0000 (22:39 +0300)]
KVM: x86: Handle errors when RIP is set during far jumps

Far jmp/call/ret may fault while loading a new RIP.  Currently KVM does not
handle this case, and may result in failed vm-entry once the assignment is
done.  The tricky part of doing so is that loading the new CS affects the
VMCS/VMCB state, so if we fail during loading the new RIP, we are left in
unconsistent state.  Therefore, this patch saves on 64-bit the old CS
descriptor and restores it if loading RIP failed.

This fixes CVE-2014-3647.

Cc: stable@vger.kernel.org
Signed-off-by: Nadav Amit <namit@cs.technion.ac.il>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
10 years agoKVM: x86: Emulator fixes for eip canonical checks on near branches
Nadav Amit [Thu, 18 Sep 2014 19:39:38 +0000 (22:39 +0300)]
KVM: x86: Emulator fixes for eip canonical checks on near branches

Before changing rip (during jmp, call, ret, etc.) the target should be asserted
to be canonical one, as real CPUs do.  During sysret, both target rsp and rip
should be canonical. If any of these values is noncanonical, a #GP exception
should occur.  The exception to this rule are syscall and sysenter instructions
in which the assigned rip is checked during the assignment to the relevant
MSRs.

This patch fixes the emulator to behave as real CPUs do for near branches.
Far branches are handled by the next patch.

This fixes CVE-2014-3647.

Cc: stable@vger.kernel.org
Signed-off-by: Nadav Amit <namit@cs.technion.ac.il>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
10 years agoKVM: x86: Fix wrong masking on relative jump/call
Nadav Amit [Thu, 18 Sep 2014 19:39:37 +0000 (22:39 +0300)]
KVM: x86: Fix wrong masking on relative jump/call

Relative jumps and calls do the masking according to the operand size, and not
according to the address size as the KVM emulator does today.

This patch fixes KVM behavior.

Cc: stable@vger.kernel.org
Signed-off-by: Nadav Amit <namit@cs.technion.ac.il>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
10 years agoKVM: x86: Improve thread safety in pit
Andy Honig [Wed, 27 Aug 2014 21:42:54 +0000 (14:42 -0700)]
KVM: x86: Improve thread safety in pit

There's a race condition in the PIT emulation code in KVM.  In
__kvm_migrate_pit_timer the pit_timer object is accessed without
synchronization.  If the race condition occurs at the wrong time this
can crash the host kernel.

This fixes CVE-2014-3611.

Cc: stable@vger.kernel.org
Signed-off-by: Andrew Honig <ahonig@google.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>