Merge branch 'arm-aesbs' of git://git.linaro.org/people/ardbiesheuvel/linux-arm into devel-stable

144 files changed, 6844 insertions, 858 deletions

diff --git a/arch/Kconfig b/arch/Kconfig
index 1feb169274fe..af2cc6eabcc7 100644
--- a/arch/Kconfig
+++ b/arch/Kconfig
@@ -286,9 +286,6 @@ config HAVE_PERF_USER_STACK_DUMP
 config HAVE_ARCH_JUMP_LABEL
 	bool
 
-config HAVE_ARCH_MUTEX_CPU_RELAX
-	bool
-
 config HAVE_RCU_TABLE_FREE
 	bool
 
diff --git a/arch/arm/Kconfig b/arch/arm/Kconfig
index b3135378ac39..22efc5d9c952 100644
--- a/arch/arm/Kconfig
+++ b/arch/arm/Kconfig
@@ -2243,8 +2243,7 @@ config NEON
 
 config KERNEL_MODE_NEON
 	bool "Support for NEON in kernel mode"
-	default n
-	depends on NEON
+	depends on NEON && AEABI
 	help
 	  Say Y to include support for NEON in kernel mode.
 
diff --git a/arch/arm/boot/dts/Makefile b/arch/arm/boot/dts/Makefile
index cc0f1fb61753..e95af3f5433b 100644
--- a/arch/arm/boot/dts/Makefile
+++ b/arch/arm/boot/dts/Makefile
@@ -183,6 +183,7 @@ dtb-$(CONFIG_ARCH_OMAP2PLUS) += omap2420-h4.dtb \
 	am335x-evm.dtb \
 	am335x-evmsk.dtb \
 	am335x-bone.dtb \
+	am335x-boneblack.dtb \
 	am3517-evm.dtb \
 	am3517_mt_ventoux.dtb \
 	am43x-epos-evm.dtb
diff --git a/arch/arm/boot/dts/am335x-bone-common.dtsi b/arch/arm/boot/dts/am335x-bone-common.dtsi
new file mode 100644
index 000000000000..2f66deda9f5c
--- /dev/null
+++ b/arch/arm/boot/dts/am335x-bone-common.dtsi
@@ -0,0 +1,262 @@
+/*
+ * Copyright (C) 2012 Texas Instruments Incorporated - http://www.ti.com/
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2 as
+ * published by the Free Software Foundation.
+ */
+
+/ {
+	model = "TI AM335x BeagleBone";
+	compatible = "ti,am335x-bone", "ti,am33xx";
+
+	cpus {
+		cpu@0 {
+			cpu0-supply = <&dcdc2_reg>;
+		};
+	};
+
+	memory {
+		device_type = "memory";
+		reg = <0x80000000 0x10000000>; /* 256 MB */
+	};
+
+	am33xx_pinmux: pinmux@44e10800 {
+		pinctrl-names = "default";
+		pinctrl-0 = <&clkout2_pin>;
+
+		user_leds_s0: user_leds_s0 {
+			pinctrl-single,pins = <
+				0x54 (PIN_OUTPUT_PULLDOWN | MUX_MODE7)	/* gpmc_a5.gpio1_21 */
+				0x58 (PIN_OUTPUT_PULLUP | MUX_MODE7)	/* gpmc_a6.gpio1_22 */
+				0x5c (PIN_OUTPUT_PULLDOWN | MUX_MODE7)	/* gpmc_a7.gpio1_23 */
+				0x60 (PIN_OUTPUT_PULLUP | MUX_MODE7)	/* gpmc_a8.gpio1_24 */
+			>;
+		};
+
+		i2c0_pins: pinmux_i2c0_pins {
+			pinctrl-single,pins = <
+				0x188 (PIN_INPUT_PULLUP | MUX_MODE0)	/* i2c0_sda.i2c0_sda */
+				0x18c (PIN_INPUT_PULLUP | MUX_MODE0)	/* i2c0_scl.i2c0_scl */
+			>;
+		};
+
+		uart0_pins: pinmux_uart0_pins {
+			pinctrl-single,pins = <
+				0x170 (PIN_INPUT_PULLUP | MUX_MODE0)	/* uart0_rxd.uart0_rxd */
+				0x174 (PIN_OUTPUT_PULLDOWN | MUX_MODE0)	/* uart0_txd.uart0_txd */
+			>;
+		};
+
+		clkout2_pin: pinmux_clkout2_pin {
+			pinctrl-single,pins = <
+				0x1b4 (PIN_OUTPUT_PULLDOWN | MUX_MODE3)	/* xdma_event_intr1.clkout2 */
+			>;
+		};
+
+		cpsw_default: cpsw_default {
+			pinctrl-single,pins = <
+				/* Slave 1 */
+				0x110 (PIN_INPUT_PULLUP | MUX_MODE0)	/* mii1_rxerr.mii1_rxerr */
+				0x114 (PIN_OUTPUT_PULLDOWN | MUX_MODE0)	/* mii1_txen.mii1_txen */
+				0x118 (PIN_INPUT_PULLUP | MUX_MODE0)	/* mii1_rxdv.mii1_rxdv */
+				0x11c (PIN_OUTPUT_PULLDOWN | MUX_MODE0)	/* mii1_txd3.mii1_txd3 */
+				0x120 (PIN_OUTPUT_PULLDOWN | MUX_MODE0)	/* mii1_txd2.mii1_txd2 */
+				0x124 (PIN_OUTPUT_PULLDOWN | MUX_MODE0)	/* mii1_txd1.mii1_txd1 */
+				0x128 (PIN_OUTPUT_PULLDOWN | MUX_MODE0)	/* mii1_txd0.mii1_txd0 */
+				0x12c (PIN_INPUT_PULLUP | MUX_MODE0)	/* mii1_txclk.mii1_txclk */
+				0x130 (PIN_INPUT_PULLUP | MUX_MODE0)	/* mii1_rxclk.mii1_rxclk */
+				0x134 (PIN_INPUT_PULLUP | MUX_MODE0)	/* mii1_rxd3.mii1_rxd3 */
+				0x138 (PIN_INPUT_PULLUP | MUX_MODE0)	/* mii1_rxd2.mii1_rxd2 */
+				0x13c (PIN_INPUT_PULLUP | MUX_MODE0)	/* mii1_rxd1.mii1_rxd1 */
+				0x140 (PIN_INPUT_PULLUP | MUX_MODE0)	/* mii1_rxd0.mii1_rxd0 */
+			>;
+		};
+
+		cpsw_sleep: cpsw_sleep {
+			pinctrl-single,pins = <
+				/* Slave 1 reset value */
+				0x110 (PIN_INPUT_PULLDOWN | MUX_MODE7)
+				0x114 (PIN_INPUT_PULLDOWN | MUX_MODE7)
+				0x118 (PIN_INPUT_PULLDOWN | MUX_MODE7)
+				0x11c (PIN_INPUT_PULLDOWN | MUX_MODE7)
+				0x120 (PIN_INPUT_PULLDOWN | MUX_MODE7)
+				0x124 (PIN_INPUT_PULLDOWN | MUX_MODE7)
+				0x128 (PIN_INPUT_PULLDOWN | MUX_MODE7)
+				0x12c (PIN_INPUT_PULLDOWN | MUX_MODE7)
+				0x130 (PIN_INPUT_PULLDOWN | MUX_MODE7)
+				0x134 (PIN_INPUT_PULLDOWN | MUX_MODE7)
+				0x138 (PIN_INPUT_PULLDOWN | MUX_MODE7)
+				0x13c (PIN_INPUT_PULLDOWN | MUX_MODE7)
+				0x140 (PIN_INPUT_PULLDOWN | MUX_MODE7)
+			>;
+		};
+
+		davinci_mdio_default: davinci_mdio_default {
+			pinctrl-single,pins = <
+				/* MDIO */
+				0x148 (PIN_INPUT_PULLUP | SLEWCTRL_FAST | MUX_MODE0)	/* mdio_data.mdio_data */
+				0x14c (PIN_OUTPUT_PULLUP | MUX_MODE0)			/* mdio_clk.mdio_clk */
+			>;
+		};
+
+		davinci_mdio_sleep: davinci_mdio_sleep {
+			pinctrl-single,pins = <
+				/* MDIO reset value */
+				0x148 (PIN_INPUT_PULLDOWN | MUX_MODE7)
+				0x14c (PIN_INPUT_PULLDOWN | MUX_MODE7)
+			>;
+		};
+	};
+
+	ocp {
+		uart0: serial@44e09000 {
+			pinctrl-names = "default";
+			pinctrl-0 = <&uart0_pins>;
+
+			status = "okay";
+		};
+
+		musb: usb@47400000 {
+			status = "okay";
+
+			control@44e10000 {
+				status = "okay";
+			};
+
+			usb-phy@47401300 {
+				status = "okay";
+			};
+
+			usb-phy@47401b00 {
+				status = "okay";
+			};
+
+			usb@47401000 {
+				status = "okay";
+			};
+
+			usb@47401800 {
+				status = "okay";
+				dr_mode = "host";
+			};
+
+			dma-controller@07402000  {
+				status = "okay";
+			};
+		};
+
+		i2c0: i2c@44e0b000 {
+			pinctrl-names = "default";
+			pinctrl-0 = <&i2c0_pins>;
+
+			status = "okay";
+			clock-frequency = <400000>;
+
+			tps: tps@24 {
+				reg = <0x24>;
+			};
+
+		};
+	};
+
+	leds {
+		pinctrl-names = "default";
+		pinctrl-0 = <&user_leds_s0>;
+
+		compatible = "gpio-leds";
+
+		led@2 {
+			label = "beaglebone:green:heartbeat";
+			gpios = <&gpio1 21 GPIO_ACTIVE_HIGH>;
+			linux,default-trigger = "heartbeat";
+			default-state = "off";
+		};
+
+		led@3 {
+			label = "beaglebone:green:mmc0";
+			gpios = <&gpio1 22 GPIO_ACTIVE_HIGH>;
+			linux,default-trigger = "mmc0";
+			default-state = "off";
+		};
+
+		led@4 {
+			label = "beaglebone:green:usr2";
+			gpios = <&gpio1 23 GPIO_ACTIVE_HIGH>;
+			default-state = "off";
+		};
+
+		led@5 {
+			label = "beaglebone:green:usr3";
+			gpios = <&gpio1 24 GPIO_ACTIVE_HIGH>;
+			default-state = "off";
+		};
+	};
+};
+
+/include/ "tps65217.dtsi"
+
+&tps {
+	regulators {
+		dcdc1_reg: regulator@0 {
+			regulator-always-on;
+		};
+
+		dcdc2_reg: regulator@1 {
+			/* VDD_MPU voltage limits 0.95V - 1.26V with +/-4% tolerance */
+			regulator-name = "vdd_mpu";
+			regulator-min-microvolt = <925000>;
+			regulator-max-microvolt = <1325000>;
+			regulator-boot-on;
+			regulator-always-on;
+		};
+
+		dcdc3_reg: regulator@2 {
+			/* VDD_CORE voltage limits 0.95V - 1.1V with +/-4% tolerance */
+			regulator-name = "vdd_core";
+			regulator-min-microvolt = <925000>;
+			regulator-max-microvolt = <1150000>;
+			regulator-boot-on;
+			regulator-always-on;
+		};
+
+		ldo1_reg: regulator@3 {
+			regulator-always-on;
+		};
+
+		ldo2_reg: regulator@4 {
+			regulator-always-on;
+		};
+
+		ldo3_reg: regulator@5 {
+			regulator-always-on;
+		};
+
+		ldo4_reg: regulator@6 {
+			regulator-always-on;
+		};
+	};
+};
+
+&cpsw_emac0 {
+	phy_id = <&davinci_mdio>, <0>;
+	phy-mode = "mii";
+};
+
+&cpsw_emac1 {
+	phy_id = <&davinci_mdio>, <1>;
+	phy-mode = "mii";
+};
+
+&mac {
+	pinctrl-names = "default", "sleep";
+	pinctrl-0 = <&cpsw_default>;
+	pinctrl-1 = <&cpsw_sleep>;
+
+};
+
+&davinci_mdio {
+	pinctrl-names = "default", "sleep";
+	pinctrl-0 = <&davinci_mdio_default>;
+	pinctrl-1 = <&davinci_mdio_sleep>;
+};
diff --git a/arch/arm/boot/dts/am335x-bone.dts b/arch/arm/boot/dts/am335x-bone.dts
index d318987d44a1..7993c489982c 100644
--- a/arch/arm/boot/dts/am335x-bone.dts
+++ b/arch/arm/boot/dts/am335x-bone.dts
@@ -8,258 +8,4 @@
 /dts-v1/;
 
 #include "am33xx.dtsi"
-
-/ {
-	model = "TI AM335x BeagleBone";
-	compatible = "ti,am335x-bone", "ti,am33xx";
-
-	cpus {
-		cpu@0 {
-			cpu0-supply = <&dcdc2_reg>;
-		};
-	};
-
-	memory {
-		device_type = "memory";
-		reg = <0x80000000 0x10000000>; /* 256 MB */
-	};
-
-	am33xx_pinmux: pinmux@44e10800 {
-		pinctrl-names = "default";
-		pinctrl-0 = <&clkout2_pin>;
-
-		user_leds_s0: user_leds_s0 {
-			pinctrl-single,pins = <
-				0x54 (PIN_OUTPUT_PULLDOWN | MUX_MODE7)	/* gpmc_a5.gpio1_21 */
-				0x58 (PIN_OUTPUT_PULLUP | MUX_MODE7)	/* gpmc_a6.gpio1_22 */
-				0x5c (PIN_OUTPUT_PULLDOWN | MUX_MODE7)	/* gpmc_a7.gpio1_23 */
-				0x60 (PIN_OUTPUT_PULLUP | MUX_MODE7)	/* gpmc_a8.gpio1_24 */
-			>;
-		};
-
-		i2c0_pins: pinmux_i2c0_pins {
-			pinctrl-single,pins = <
-				0x188 (PIN_INPUT_PULLUP | MUX_MODE0)	/* i2c0_sda.i2c0_sda */
-				0x18c (PIN_INPUT_PULLUP | MUX_MODE0)	/* i2c0_scl.i2c0_scl */
-			>;
-		};
-
-		uart0_pins: pinmux_uart0_pins {
-			pinctrl-single,pins = <
-				0x170 (PIN_INPUT_PULLUP | MUX_MODE0)	/* uart0_rxd.uart0_rxd */
-				0x174 (PIN_OUTPUT_PULLDOWN | MUX_MODE0)	/* uart0_txd.uart0_txd */
-			>;
-		};
-
-		clkout2_pin: pinmux_clkout2_pin {
-			pinctrl-single,pins = <
-				0x1b4 (PIN_OUTPUT_PULLDOWN | MUX_MODE3)	/* xdma_event_intr1.clkout2 */
-			>;
-		};
-
-		cpsw_default: cpsw_default {
-			pinctrl-single,pins = <
-				/* Slave 1 */
-				0x110 (PIN_INPUT_PULLUP | MUX_MODE0)	/* mii1_rxerr.mii1_rxerr */
-				0x114 (PIN_OUTPUT_PULLDOWN | MUX_MODE0)	/* mii1_txen.mii1_txen */
-				0x118 (PIN_INPUT_PULLUP | MUX_MODE0)	/* mii1_rxdv.mii1_rxdv */
-				0x11c (PIN_OUTPUT_PULLDOWN | MUX_MODE0)	/* mii1_txd3.mii1_txd3 */
-				0x120 (PIN_OUTPUT_PULLDOWN | MUX_MODE0)	/* mii1_txd2.mii1_txd2 */
-				0x124 (PIN_OUTPUT_PULLDOWN | MUX_MODE0)	/* mii1_txd1.mii1_txd1 */
-				0x128 (PIN_OUTPUT_PULLDOWN | MUX_MODE0)	/* mii1_txd0.mii1_txd0 */
-				0x12c (PIN_INPUT_PULLUP | MUX_MODE0)	/* mii1_txclk.mii1_txclk */
-				0x130 (PIN_INPUT_PULLUP | MUX_MODE0)	/* mii1_rxclk.mii1_rxclk */
-				0x134 (PIN_INPUT_PULLUP | MUX_MODE0)	/* mii1_rxd3.mii1_rxd3 */
-				0x138 (PIN_INPUT_PULLUP | MUX_MODE0)	/* mii1_rxd2.mii1_rxd2 */
-				0x13c (PIN_INPUT_PULLUP | MUX_MODE0)	/* mii1_rxd1.mii1_rxd1 */
-				0x140 (PIN_INPUT_PULLUP | MUX_MODE0)	/* mii1_rxd0.mii1_rxd0 */
-			>;
-		};
-
-		cpsw_sleep: cpsw_sleep {
-			pinctrl-single,pins = <
-				/* Slave 1 reset value */
-				0x110 (PIN_INPUT_PULLDOWN | MUX_MODE7)
-				0x114 (PIN_INPUT_PULLDOWN | MUX_MODE7)
-				0x118 (PIN_INPUT_PULLDOWN | MUX_MODE7)
-				0x11c (PIN_INPUT_PULLDOWN | MUX_MODE7)
-				0x120 (PIN_INPUT_PULLDOWN | MUX_MODE7)
-				0x124 (PIN_INPUT_PULLDOWN | MUX_MODE7)
-				0x128 (PIN_INPUT_PULLDOWN | MUX_MODE7)
-				0x12c (PIN_INPUT_PULLDOWN | MUX_MODE7)
-				0x130 (PIN_INPUT_PULLDOWN | MUX_MODE7)
-				0x134 (PIN_INPUT_PULLDOWN | MUX_MODE7)
-				0x138 (PIN_INPUT_PULLDOWN | MUX_MODE7)
-				0x13c (PIN_INPUT_PULLDOWN | MUX_MODE7)
-				0x140 (PIN_INPUT_PULLDOWN | MUX_MODE7)
-			>;
-		};
-
-		davinci_mdio_default: davinci_mdio_default {
-			pinctrl-single,pins = <
-				/* MDIO */
-				0x148 (PIN_INPUT_PULLUP | SLEWCTRL_FAST | MUX_MODE0)	/* mdio_data.mdio_data */
-				0x14c (PIN_OUTPUT_PULLUP | MUX_MODE0)			/* mdio_clk.mdio_clk */
-			>;
-		};
-
-		davinci_mdio_sleep: davinci_mdio_sleep {
-			pinctrl-single,pins = <
-				/* MDIO reset value */
-				0x148 (PIN_INPUT_PULLDOWN | MUX_MODE7)
-				0x14c (PIN_INPUT_PULLDOWN | MUX_MODE7)
-			>;
-		};
-	};
-
-	ocp {
-		uart0: serial@44e09000 {
-			pinctrl-names = "default";
-			pinctrl-0 = <&uart0_pins>;
-
-			status = "okay";
-		};
-
-		musb: usb@47400000 {
-			status = "okay";
-
-			control@44e10000 {
-				status = "okay";
-			};
-
-			usb-phy@47401300 {
-				status = "okay";
-			};
-
-			usb-phy@47401b00 {
-				status = "okay";
-			};
-
-			usb@47401000 {
-				status = "okay";
-			};
-
-			usb@47401800 {
-				status = "okay";
-				dr_mode = "host";
-			};
-
-			dma-controller@07402000  {
-				status = "okay";
-			};
-		};
-
-		i2c0: i2c@44e0b000 {
-			pinctrl-names = "default";
-			pinctrl-0 = <&i2c0_pins>;
-
-			status = "okay";
-			clock-frequency = <400000>;
-
-			tps: tps@24 {
-				reg = <0x24>;
-			};
-
-		};
-	};
-
-	leds {
-		pinctrl-names = "default";
-		pinctrl-0 = <&user_leds_s0>;
-
-		compatible = "gpio-leds";
-
-		led@2 {
-			label = "beaglebone:green:heartbeat";
-			gpios = <&gpio1 21 GPIO_ACTIVE_HIGH>;
-			linux,default-trigger = "heartbeat";
-			default-state = "off";
-		};
-
-		led@3 {
-			label = "beaglebone:green:mmc0";
-			gpios = <&gpio1 22 GPIO_ACTIVE_HIGH>;
-			linux,default-trigger = "mmc0";
-			default-state = "off";
-		};
-
-		led@4 {
-			label = "beaglebone:green:usr2";
-			gpios = <&gpio1 23 GPIO_ACTIVE_HIGH>;
-			default-state = "off";
-		};
-
-		led@5 {
-			label = "beaglebone:green:usr3";
-			gpios = <&gpio1 24 GPIO_ACTIVE_HIGH>;
-			default-state = "off";
-		};
-	};
-};
-
-/include/ "tps65217.dtsi"
-
-&tps {
-	regulators {
-		dcdc1_reg: regulator@0 {
-			regulator-always-on;
-		};
-
-		dcdc2_reg: regulator@1 {
-			/* VDD_MPU voltage limits 0.95V - 1.26V with +/-4% tolerance */
-			regulator-name = "vdd_mpu";
-			regulator-min-microvolt = <925000>;
-			regulator-max-microvolt = <1325000>;
-			regulator-boot-on;
-			regulator-always-on;
-		};
-
-		dcdc3_reg: regulator@2 {
-			/* VDD_CORE voltage limits 0.95V - 1.1V with +/-4% tolerance */
-			regulator-name = "vdd_core";
-			regulator-min-microvolt = <925000>;
-			regulator-max-microvolt = <1150000>;
-			regulator-boot-on;
-			regulator-always-on;
-		};
-
-		ldo1_reg: regulator@3 {
-			regulator-always-on;
-		};
-
-		ldo2_reg: regulator@4 {
-			regulator-always-on;
-		};
-
-		ldo3_reg: regulator@5 {
-			regulator-always-on;
-		};
-
-		ldo4_reg: regulator@6 {
-			regulator-always-on;
-		};
-	};
-};
-
-&cpsw_emac0 {
-	phy_id = <&davinci_mdio>, <0>;
-	phy-mode = "mii";
-};
-
-&cpsw_emac1 {
-	phy_id = <&davinci_mdio>, <1>;
-	phy-mode = "mii";
-};
-
-&mac {
-	pinctrl-names = "default", "sleep";
-	pinctrl-0 = <&cpsw_default>;
-	pinctrl-1 = <&cpsw_sleep>;
-
-};
-
-&davinci_mdio {
-	pinctrl-names = "default", "sleep";
-	pinctrl-0 = <&davinci_mdio_default>;
-	pinctrl-1 = <&davinci_mdio_sleep>;
-};
+#include "am335x-bone-common.dtsi"
diff --git a/arch/arm/boot/dts/am335x-boneblack.dts b/arch/arm/boot/dts/am335x-boneblack.dts
new file mode 100644
index 000000000000..197cadf72d2c
--- /dev/null
+++ b/arch/arm/boot/dts/am335x-boneblack.dts
@@ -0,0 +1,17 @@
+/*
+ * Copyright (C) 2012 Texas Instruments Incorporated - http://www.ti.com/
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2 as
+ * published by the Free Software Foundation.
+ */
+/dts-v1/;
+
+#include "am33xx.dtsi"
+#include "am335x-bone-common.dtsi"
+
+&ldo3_reg {
+	regulator-min-microvolt = <1800000>;
+	regulator-max-microvolt = <1800000>;
+	regulator-always-on;
+};
diff --git a/arch/arm/boot/dts/imx27.dtsi b/arch/arm/boot/dts/imx27.dtsi
index c037c223619a..b7a1c6d950b9 100644
--- a/arch/arm/boot/dts/imx27.dtsi
+++ b/arch/arm/boot/dts/imx27.dtsi
@@ -187,7 +187,7 @@
 				compatible = "fsl,imx27-cspi";
 				reg = <0x1000e000 0x1000>;
 				interrupts = <16>;
-				clocks = <&clks 53>, <&clks 53>;
+				clocks = <&clks 53>, <&clks 60>;
 				clock-names = "ipg", "per";
 				status = "disabled";
 			};
@@ -198,7 +198,7 @@
 				compatible = "fsl,imx27-cspi";
 				reg = <0x1000f000 0x1000>;
 				interrupts = <15>;
-				clocks = <&clks 52>, <&clks 52>;
+				clocks = <&clks 52>, <&clks 60>;
 				clock-names = "ipg", "per";
 				status = "disabled";
 			};
@@ -309,7 +309,7 @@
 				compatible = "fsl,imx27-cspi";
 				reg = <0x10017000 0x1000>;
 				interrupts = <6>;
-				clocks = <&clks 51>, <&clks 51>;
+				clocks = <&clks 51>, <&clks 60>;
 				clock-names = "ipg", "per";
 				status = "disabled";
 			};
diff --git a/arch/arm/boot/dts/imx51.dtsi b/arch/arm/boot/dts/imx51.dtsi
index a85abb424c34..54cee6517902 100644
--- a/arch/arm/boot/dts/imx51.dtsi
+++ b/arch/arm/boot/dts/imx51.dtsi
@@ -474,7 +474,7 @@
 				compatible = "fsl,imx51-pata", "fsl,imx27-pata";
 				reg = <0x83fe0000 0x4000>;
 				interrupts = <70>;
-				clocks = <&clks 161>;
+				clocks = <&clks 172>;
 				status = "disabled";
 			};
 
diff --git a/arch/arm/boot/dts/imx6q-pinfunc.h b/arch/arm/boot/dts/imx6q-pinfunc.h
index c0e38a45e4bb..9bbe82bdee41 100644
--- a/arch/arm/boot/dts/imx6q-pinfunc.h
+++ b/arch/arm/boot/dts/imx6q-pinfunc.h
@@ -207,8 +207,8 @@
 #define MX6QDL_PAD_EIM_D29__ECSPI4_SS0              0x0c8 0x3dc 0x824 0x2 0x1
 #define MX6QDL_PAD_EIM_D29__UART2_RTS_B             0x0c8 0x3dc 0x924 0x4 0x1
 #define MX6QDL_PAD_EIM_D29__UART2_CTS_B             0x0c8 0x3dc 0x000 0x4 0x0
-#define MX6QDL_PAD_EIM_D29__UART2_DTE_RTS_B         0x0c4 0x3dc 0x000 0x4 0x0
-#define MX6QDL_PAD_EIM_D29__UART2_DTE_CTS_B         0x0c4 0x3dc 0x924 0x4 0x1
+#define MX6QDL_PAD_EIM_D29__UART2_DTE_RTS_B         0x0c8 0x3dc 0x000 0x4 0x0
+#define MX6QDL_PAD_EIM_D29__UART2_DTE_CTS_B         0x0c8 0x3dc 0x924 0x4 0x1
 #define MX6QDL_PAD_EIM_D29__GPIO3_IO29              0x0c8 0x3dc 0x000 0x5 0x0
 #define MX6QDL_PAD_EIM_D29__IPU2_CSI1_VSYNC         0x0c8 0x3dc 0x8e4 0x6 0x0
 #define MX6QDL_PAD_EIM_D29__IPU1_DI0_PIN14          0x0c8 0x3dc 0x000 0x7 0x0
diff --git a/arch/arm/boot/dts/omap3-beagle-xm.dts b/arch/arm/boot/dts/omap3-beagle-xm.dts
index afdb16417d4e..0c514dc8460c 100644
--- a/arch/arm/boot/dts/omap3-beagle-xm.dts
+++ b/arch/arm/boot/dts/omap3-beagle-xm.dts
@@ -11,7 +11,7 @@
 
 / {
 	model = "TI OMAP3 BeagleBoard xM";
-	compatible = "ti,omap3-beagle-xm, ti,omap3-beagle", "ti,omap3";
+	compatible = "ti,omap3-beagle-xm", "ti,omap3-beagle", "ti,omap3";
 
 	cpus {
 		cpu@0 {
diff --git a/arch/arm/boot/dts/omap3-igep.dtsi b/arch/arm/boot/dts/omap3-igep.dtsi
index bc48b114eae6..2326d11462a5 100644
--- a/arch/arm/boot/dts/omap3-igep.dtsi
+++ b/arch/arm/boot/dts/omap3-igep.dtsi
@@ -48,6 +48,15 @@
 		>;
 	};
 
+	mcbsp2_pins: pinmux_mcbsp2_pins {
+		pinctrl-single,pins = <
+			0x10c (PIN_INPUT | MUX_MODE0)		/* mcbsp2_fsx.mcbsp2_fsx */
+			0x10e (PIN_INPUT | MUX_MODE0)		/* mcbsp2_clkx.mcbsp2_clkx */
+			0x110 (PIN_INPUT | MUX_MODE0)		/* mcbsp2_dr.mcbsp2.dr */
+			0x112 (PIN_OUTPUT | MUX_MODE0)		/* mcbsp2_dx.mcbsp2_dx */
+		>;
+	};
+
 	mmc1_pins: pinmux_mmc1_pins {
 		pinctrl-single,pins = <
 			0x114 (PIN_INPUT_PULLUP | MUX_MODE0)	/* sdmmc1_clk.sdmmc1_clk */
@@ -93,6 +102,11 @@
 	clock-frequency = <400000>;
 };
 
+&mcbsp2 {
+	pinctrl-names = "default";
+	pinctrl-0 = <&mcbsp2_pins>;
+};
+
 &mmc1 {
       pinctrl-names = "default";
       pinctrl-0 = <&mmc1_pins>;
diff --git a/arch/arm/boot/dts/omap4-panda-common.dtsi b/arch/arm/boot/dts/omap4-panda-common.dtsi
index faa95b5b242e..814ab67c8c29 100644
--- a/arch/arm/boot/dts/omap4-panda-common.dtsi
+++ b/arch/arm/boot/dts/omap4-panda-common.dtsi
@@ -107,6 +107,19 @@
 	 */
 		clock-frequency = <19200000>;
 	};
+
+	/* regulator for wl12xx on sdio5 */
+	wl12xx_vmmc: wl12xx_vmmc {
+		pinctrl-names = "default";
+		pinctrl-0 = <&wl12xx_gpio>;
+		compatible = "regulator-fixed";
+		regulator-name = "vwl1271";
+		regulator-min-microvolt = <1800000>;
+		regulator-max-microvolt = <1800000>;
+		gpio = <&gpio2 11 0>;
+		startup-delay-us = <70000>;
+		enable-active-high;
+	};
 };
 
 &omap4_pmx_wkup {
@@ -235,6 +248,33 @@
 			0x1c (PIN_OUTPUT | MUX_MODE3)	/* gpio_wk8 */
 		>;
 	};
+
+	/*
+	 * wl12xx GPIO outputs for WLAN_EN, BT_EN, FM_EN, BT_WAKEUP
+	 * REVISIT: Are the pull-ups needed for GPIO 48 and 49?
+	 */
+	wl12xx_gpio: pinmux_wl12xx_gpio {
+		pinctrl-single,pins = <
+			0x26 (PIN_OUTPUT | MUX_MODE3)		/* gpmc_a19.gpio_43 */
+			0x2c (PIN_OUTPUT | MUX_MODE3)		/* gpmc_a22.gpio_46 */
+			0x30 (PIN_OUTPUT_PULLUP | MUX_MODE3)	/* gpmc_a24.gpio_48 */
+			0x32 (PIN_OUTPUT_PULLUP | MUX_MODE3)	/* gpmc_a25.gpio_49 */
+		>;
+	};
+
+	/* wl12xx GPIO inputs and SDIO pins */
+	wl12xx_pins: pinmux_wl12xx_pins {
+		pinctrl-single,pins = <
+			0x38 (PIN_INPUT | MUX_MODE3)		/* gpmc_ncs2.gpio_52 */
+			0x3a (PIN_INPUT | MUX_MODE3)		/* gpmc_ncs3.gpio_53 */
+			0x108 (PIN_OUTPUT | MUX_MODE0)		/* sdmmc5_clk.sdmmc5_clk */
+			0x10a (PIN_INPUT_PULLUP | MUX_MODE0)	/* sdmmc5_cmd.sdmmc5_cmd */
+			0x10c (PIN_INPUT_PULLUP | MUX_MODE0)	/* sdmmc5_dat0.sdmmc5_dat0 */
+			0x10e (PIN_INPUT_PULLUP | MUX_MODE0)	/* sdmmc5_dat1.sdmmc5_dat1 */
+			0x110 (PIN_INPUT_PULLUP | MUX_MODE0)	/* sdmmc5_dat2.sdmmc5_dat2 */
+			0x112 (PIN_INPUT_PULLUP | MUX_MODE0)	/* sdmmc5_dat3.sdmmc5_dat3 */
+		>;
+	};
 };
 
 &i2c1 {
@@ -314,8 +354,12 @@
 };
 
 &mmc5 {
-	ti,non-removable;
+	pinctrl-names = "default";
+	pinctrl-0 = <&wl12xx_pins>;
+	vmmc-supply = <&wl12xx_vmmc>;
+	non-removable;
 	bus-width = <4>;
+	cap-power-off-card;
 };
 
 &emif1 {
diff --git a/arch/arm/boot/dts/omap4-sdp.dts b/arch/arm/boot/dts/omap4-sdp.dts
index 7951b4ea500a..4f78380ecdb8 100644
--- a/arch/arm/boot/dts/omap4-sdp.dts
+++ b/arch/arm/boot/dts/omap4-sdp.dts
@@ -140,6 +140,19 @@
 			"DMic", "Digital Mic",
 			"Digital Mic", "Digital Mic1 Bias";
 	};
+
+	/* regulator for wl12xx on sdio5 */
+	wl12xx_vmmc: wl12xx_vmmc {
+		pinctrl-names = "default";
+		pinctrl-0 = <&wl12xx_gpio>;
+		compatible = "regulator-fixed";
+		regulator-name = "vwl1271";
+		regulator-min-microvolt = <1800000>;
+		regulator-max-microvolt = <1800000>;
+		gpio = <&gpio2 22 0>;
+		startup-delay-us = <70000>;
+		enable-active-high;
+	};
 };
 
 &omap4_pmx_wkup {
@@ -295,6 +308,26 @@
 			0xf0 (PIN_INPUT_PULLUP | MUX_MODE0)	/* i2c4_sda */
 		>;
 	};
+
+	/* wl12xx GPIO output for WLAN_EN */
+	wl12xx_gpio: pinmux_wl12xx_gpio {
+		pinctrl-single,pins = <
+			0x3c (PIN_OUTPUT | MUX_MODE3)		/* gpmc_nwp.gpio_54 */
+		>;
+	};
+
+	/* wl12xx GPIO inputs and SDIO pins */
+	wl12xx_pins: pinmux_wl12xx_pins {
+		pinctrl-single,pins = <
+			0x3a (PIN_INPUT | MUX_MODE3)		/* gpmc_ncs3.gpio_53 */
+			0x108 (PIN_OUTPUT | MUX_MODE3)		/* sdmmc5_clk.sdmmc5_clk */
+			0x10a (PIN_INPUT_PULLUP | MUX_MODE3)	/* sdmmc5_cmd.sdmmc5_cmd */
+			0x10c (PIN_INPUT_PULLUP | MUX_MODE3)	/* sdmmc5_dat0.sdmmc5_dat0 */
+			0x10e (PIN_INPUT_PULLUP | MUX_MODE3)	/* sdmmc5_dat1.sdmmc5_dat1 */
+			0x110 (PIN_INPUT_PULLUP | MUX_MODE3)	/* sdmmc5_dat2.sdmmc5_dat2 */
+			0x112 (PIN_INPUT_PULLUP | MUX_MODE3)	/* sdmmc5_dat3.sdmmc5_dat3 */
+		>;
+	};
 };
 
 &i2c1 {
@@ -420,8 +453,12 @@
 };
 
 &mmc5 {
+	pinctrl-names = "default";
+	pinctrl-0 = <&wl12xx_pins>;
+	vmmc-supply = <&wl12xx_vmmc>;
+	non-removable;
 	bus-width = <4>;
-	ti,non-removable;
+	cap-power-off-card;
 };
 
 &emif1 {
diff --git a/arch/arm/boot/dts/omap5.dtsi b/arch/arm/boot/dts/omap5.dtsi
index 07be2cd7b318..7cdea1bfea09 100644
--- a/arch/arm/boot/dts/omap5.dtsi
+++ b/arch/arm/boot/dts/omap5.dtsi
@@ -637,7 +637,7 @@
 		omap_dwc3@4a020000 {
 			compatible = "ti,dwc3";
 			ti,hwmods = "usb_otg_ss";
-			reg = <0x4a020000 0x1000>;
+			reg = <0x4a020000 0x10000>;
 			interrupts = <GIC_SPI 93 IRQ_TYPE_LEVEL_HIGH>;
 			#address-cells = <1>;
 			#size-cells = <1>;
@@ -645,17 +645,18 @@
 			ranges;
 			dwc3@4a030000 {
 				compatible = "snps,dwc3";
-				reg = <0x4a030000 0x1000>;
+				reg = <0x4a030000 0x10000>;
 				interrupts = <GIC_SPI 92 IRQ_TYPE_LEVEL_HIGH>;
 				usb-phy = <&usb2_phy>, <&usb3_phy>;
 				tx-fifo-resize;
 			};
 		};
 
-		ocp2scp {
+		ocp2scp@4a080000 {
 			compatible = "ti,omap-ocp2scp";
 			#address-cells = <1>;
 			#size-cells = <1>;
+			reg = <0x4a080000 0x20>;
 			ranges;
 			ti,hwmods = "ocp2scp1";
 			usb2_phy: usb2phy@4a084000 {
diff --git a/arch/arm/configs/multi_v7_defconfig b/arch/arm/configs/multi_v7_defconfig
index 6e572c64cf5a..f3935b46df29 100644
--- a/arch/arm/configs/multi_v7_defconfig
+++ b/arch/arm/configs/multi_v7_defconfig
@@ -36,6 +36,7 @@ CONFIG_ARCH_TEGRA_114_SOC=y
 CONFIG_TEGRA_PCI=y
 CONFIG_TEGRA_EMC_SCALING_ENABLE=y
 CONFIG_ARCH_U8500=y
+CONFIG_MACH_HREFV60=y
 CONFIG_MACH_SNOWBALL=y
 CONFIG_MACH_UX500_DT=y
 CONFIG_ARCH_VEXPRESS=y
@@ -46,6 +47,7 @@ CONFIG_ARCH_ZYNQ=y
 CONFIG_SMP=y
 CONFIG_HIGHPTE=y
 CONFIG_ARM_APPENDED_DTB=y
+CONFIG_ARM_ATAG_DTB_COMPAT=y
 CONFIG_NET=y
 CONFIG_UNIX=y
 CONFIG_INET=y
diff --git a/arch/arm/crypto/Makefile b/arch/arm/crypto/Makefile
index a2c83851bc90..81cda39860c5 100644
--- a/arch/arm/crypto/Makefile
+++ b/arch/arm/crypto/Makefile
@@ -3,7 +3,17 @@
 #
 
 obj-$(CONFIG_CRYPTO_AES_ARM) += aes-arm.o
+obj-$(CONFIG_CRYPTO_AES_ARM_BS) += aes-arm-bs.o
 obj-$(CONFIG_CRYPTO_SHA1_ARM) += sha1-arm.o
 
-aes-arm-y  := aes-armv4.o aes_glue.o
-sha1-arm-y := sha1-armv4-large.o sha1_glue.o
+aes-arm-y	:= aes-armv4.o aes_glue.o
+aes-arm-bs-y	:= aesbs-core.o aesbs-glue.o
+sha1-arm-y	:= sha1-armv4-large.o sha1_glue.o
+
+quiet_cmd_perl = PERL    $@
+      cmd_perl = $(PERL) $(<) > $(@)
+
+$(src)/aesbs-core.S_shipped: $(src)/bsaes-armv7.pl
+	$(call cmd,perl)
+
+.PRECIOUS: $(obj)/aesbs-core.S
diff --git a/arch/arm/crypto/aes-armv4.S b/arch/arm/crypto/aes-armv4.S
index 19d6cd6f29f9..3a14ea8fe97e 100644
--- a/arch/arm/crypto/aes-armv4.S
+++ b/arch/arm/crypto/aes-armv4.S
@@ -148,7 +148,7 @@ AES_Te:
 @ 		 const AES_KEY *key) {
 .align	5
 ENTRY(AES_encrypt)
-	sub	r3,pc,#8		@ AES_encrypt
+	adr	r3,AES_encrypt
 	stmdb   sp!,{r1,r4-r12,lr}
 	mov	r12,r0		@ inp
 	mov	r11,r2
@@ -381,7 +381,7 @@ _armv4_AES_encrypt:
 .align	5
 ENTRY(private_AES_set_encrypt_key)
 _armv4_AES_set_encrypt_key:
-	sub	r3,pc,#8		@ AES_set_encrypt_key
+	adr	r3,_armv4_AES_set_encrypt_key
 	teq	r0,#0
 	moveq	r0,#-1
 	beq	.Labrt
@@ -843,7 +843,7 @@ AES_Td:
 @ 		 const AES_KEY *key) {
 .align	5
 ENTRY(AES_decrypt)
-	sub	r3,pc,#8		@ AES_decrypt
+	adr	r3,AES_decrypt
 	stmdb   sp!,{r1,r4-r12,lr}
 	mov	r12,r0		@ inp
 	mov	r11,r2
diff --git a/arch/arm/crypto/aes_glue.c b/arch/arm/crypto/aes_glue.c
index 59f7877ead6a..3003fa1f6fb4 100644
--- a/arch/arm/crypto/aes_glue.c
+++ b/arch/arm/crypto/aes_glue.c
@@ -6,22 +6,12 @@
 #include <linux/crypto.h>
 #include <crypto/aes.h>
 
-#define AES_MAXNR 14
+#include "aes_glue.h"
 
-typedef struct {
-	unsigned int rd_key[4 *(AES_MAXNR + 1)];
-	int rounds;
-} AES_KEY;
-
-struct AES_CTX {
-	AES_KEY enc_key;
-	AES_KEY dec_key;
-};
-
-asmlinkage void AES_encrypt(const u8 *in, u8 *out, AES_KEY *ctx);
-asmlinkage void AES_decrypt(const u8 *in, u8 *out, AES_KEY *ctx);
-asmlinkage int private_AES_set_decrypt_key(const unsigned char *userKey, const int bits, AES_KEY *key);
-asmlinkage int private_AES_set_encrypt_key(const unsigned char *userKey, const int bits, AES_KEY *key);
+EXPORT_SYMBOL(AES_encrypt);
+EXPORT_SYMBOL(AES_decrypt);
+EXPORT_SYMBOL(private_AES_set_encrypt_key);
+EXPORT_SYMBOL(private_AES_set_decrypt_key);
 
 static void aes_encrypt(struct crypto_tfm *tfm, u8 *dst, const u8 *src)
 {
@@ -81,7 +71,7 @@ static struct crypto_alg aes_alg = {
 		.cipher	= {
 			.cia_min_keysize	= AES_MIN_KEY_SIZE,
 			.cia_max_keysize	= AES_MAX_KEY_SIZE,
-			.cia_setkey			= aes_set_key,
+			.cia_setkey		= aes_set_key,
 			.cia_encrypt		= aes_encrypt,
 			.cia_decrypt		= aes_decrypt
 		}
diff --git a/arch/arm/crypto/aes_glue.h b/arch/arm/crypto/aes_glue.h
new file mode 100644
index 000000000000..cca3e51eb606
--- /dev/null
+++ b/arch/arm/crypto/aes_glue.h
@@ -0,0 +1,19 @@
+
+#define AES_MAXNR 14
+
+struct AES_KEY {
+	unsigned int rd_key[4 * (AES_MAXNR + 1)];
+	int rounds;
+};
+
+struct AES_CTX {
+	struct AES_KEY enc_key;
+	struct AES_KEY dec_key;
+};
+
+asmlinkage void AES_encrypt(const u8 *in, u8 *out, struct AES_KEY *ctx);
+asmlinkage void AES_decrypt(const u8 *in, u8 *out, struct AES_KEY *ctx);
+asmlinkage int private_AES_set_decrypt_key(const unsigned char *userKey,
+					   const int bits, struct AES_KEY *key);
+asmlinkage int private_AES_set_encrypt_key(const unsigned char *userKey,
+					   const int bits, struct AES_KEY *key);
diff --git a/arch/arm/crypto/aesbs-core.S_shipped b/arch/arm/crypto/aesbs-core.S_shipped
new file mode 100644
index 000000000000..64205d453260
--- /dev/null
+++ b/arch/arm/crypto/aesbs-core.S_shipped
@@ -0,0 +1,2544 @@
+
+@ ====================================================================
+@ Written by Andy Polyakov <appro@openssl.org> for the OpenSSL
+@ project. The module is, however, dual licensed under OpenSSL and
+@ CRYPTOGAMS licenses depending on where you obtain it. For further
+@ details see http://www.openssl.org/~appro/cryptogams/.
+@
+@ Specific modes and adaptation for Linux kernel by Ard Biesheuvel
+@ <ard.biesheuvel@linaro.org>. Permission to use under GPL terms is
+@ granted.
+@ ====================================================================
+
+@ Bit-sliced AES for ARM NEON
+@
+@ February 2012.
+@
+@ This implementation is direct adaptation of bsaes-x86_64 module for
+@ ARM NEON. Except that this module is endian-neutral [in sense that
+@ it can be compiled for either endianness] by courtesy of vld1.8's
+@ neutrality. Initial version doesn't implement interface to OpenSSL,
+@ only low-level primitives and unsupported entry points, just enough
+@ to collect performance results, which for Cortex-A8 core are:
+@
+@ encrypt	19.5 cycles per byte processed with 128-bit key
+@ decrypt	22.1 cycles per byte processed with 128-bit key
+@ key conv.	440  cycles per 128-bit key/0.18 of 8x block
+@
+@ Snapdragon S4 encrypts byte in 17.6 cycles and decrypts in 19.7,
+@ which is [much] worse than anticipated (for further details see
+@ http://www.openssl.org/~appro/Snapdragon-S4.html).
+@
+@ Cortex-A15 manages in 14.2/16.1 cycles [when integer-only code
+@ manages in 20.0 cycles].
+@
+@ When comparing to x86_64 results keep in mind that NEON unit is
+@ [mostly] single-issue and thus can't [fully] benefit from
+@ instruction-level parallelism. And when comparing to aes-armv4
+@ results keep in mind key schedule conversion overhead (see
+@ bsaes-x86_64.pl for further details)...
+@
+@						<appro@openssl.org>
+
+@ April-August 2013
+@
+@ Add CBC, CTR and XTS subroutines, adapt for kernel use.
+@
+@					<ard.biesheuvel@linaro.org>
+
+#ifndef __KERNEL__
+# include "arm_arch.h"
+
+# define VFP_ABI_PUSH	vstmdb	sp!,{d8-d15}
+# define VFP_ABI_POP	vldmia	sp!,{d8-d15}
+# define VFP_ABI_FRAME	0x40
+#else
+# define VFP_ABI_PUSH
+# define VFP_ABI_POP
+# define VFP_ABI_FRAME	0
+# define BSAES_ASM_EXTENDED_KEY
+# define XTS_CHAIN_TWEAK
+# define __ARM_ARCH__ __LINUX_ARM_ARCH__
+#endif
+
+#ifdef __thumb__
+# define adrl adr
+#endif
+
+#if __ARM_ARCH__>=7
+.text
+.syntax	unified 	@ ARMv7-capable assembler is expected to handle this
+#ifdef __thumb2__
+.thumb
+#else
+.code   32
+#endif
+
+.fpu	neon
+
+.type	_bsaes_decrypt8,%function
+.align	4
+_bsaes_decrypt8:
+	adr	r6,_bsaes_decrypt8
+	vldmia	r4!, {q9}		@ round 0 key
+	add	r6,r6,#.LM0ISR-_bsaes_decrypt8
+
+	vldmia	r6!, {q8}		@ .LM0ISR
+	veor	q10, q0, q9	@ xor with round0 key
+	veor	q11, q1, q9
+	 vtbl.8	d0, {q10}, d16
+	 vtbl.8	d1, {q10}, d17
+	veor	q12, q2, q9
+	 vtbl.8	d2, {q11}, d16
+	 vtbl.8	d3, {q11}, d17
+	veor	q13, q3, q9
+	 vtbl.8	d4, {q12}, d16
+	 vtbl.8	d5, {q12}, d17
+	veor	q14, q4, q9
+	 vtbl.8	d6, {q13}, d16
+	 vtbl.8	d7, {q13}, d17
+	veor	q15, q5, q9
+	 vtbl.8	d8, {q14}, d16
+	 vtbl.8	d9, {q14}, d17
+	veor	q10, q6, q9
+	 vtbl.8	d10, {q15}, d16
+	 vtbl.8	d11, {q15}, d17
+	veor	q11, q7, q9
+	 vtbl.8	d12, {q10}, d16
+	 vtbl.8	d13, {q10}, d17
+	 vtbl.8	d14, {q11}, d16
+	 vtbl.8	d15, {q11}, d17
+	vmov.i8	q8,#0x55			@ compose .LBS0
+	vmov.i8	q9,#0x33			@ compose .LBS1
+	vshr.u64	q10, q6, #1
+	 vshr.u64	q11, q4, #1
+	veor		q10, q10, q7
+	 veor		q11, q11, q5
+	vand		q10, q10, q8
+	 vand		q11, q11, q8
+	veor		q7, q7, q10
+	vshl.u64	q10, q10, #1
+	 veor		q5, q5, q11
+	 vshl.u64	q11, q11, #1
+	veor		q6, q6, q10
+	 veor		q4, q4, q11
+	vshr.u64	q10, q2, #1
+	 vshr.u64	q11, q0, #1
+	veor		q10, q10, q3
+	 veor		q11, q11, q1
+	vand		q10, q10, q8
+	 vand		q11, q11, q8
+	veor		q3, q3, q10
+	vshl.u64	q10, q10, #1
+	 veor		q1, q1, q11
+	 vshl.u64	q11, q11, #1
+	veor		q2, q2, q10
+	 veor		q0, q0, q11
+	vmov.i8	q8,#0x0f			@ compose .LBS2
+	vshr.u64	q10, q5, #2
+	 vshr.u64	q11, q4, #2
+	veor		q10, q10, q7
+	 veor		q11, q11, q6
+	vand		q10, q10, q9
+	 vand		q11, q11, q9
+	veor		q7, q7, q10
+	vshl.u64	q10, q10, #2
+	 veor		q6, q6, q11
+	 vshl.u64	q11, q11, #2
+	veor		q5, q5, q10
+	 veor		q4, q4, q11
+	vshr.u64	q10, q1, #2
+	 vshr.u64	q11, q0, #2
+	veor		q10, q10, q3
+	 veor		q11, q11, q2
+	vand		q10, q10, q9
+	 vand		q11, q11, q9
+	veor		q3, q3, q10
+	vshl.u64	q10, q10, #2
+	 veor		q2, q2, q11
+	 vshl.u64	q11, q11, #2
+	veor		q1, q1, q10
+	 veor		q0, q0, q11
+	vshr.u64	q10, q3, #4
+	 vshr.u64	q11, q2, #4
+	veor		q10, q10, q7
+	 veor		q11, q11, q6
+	vand		q10, q10, q8
+	 vand		q11, q11, q8
+	veor		q7, q7, q10
+	vshl.u64	q10, q10, #4
+	 veor		q6, q6, q11
+	 vshl.u64	q11, q11, #4
+	veor		q3, q3, q10
+	 veor		q2, q2, q11
+	vshr.u64	q10, q1, #4
+	 vshr.u64	q11, q0, #4
+	veor		q10, q10, q5
+	 veor		q11, q11, q4
+	vand		q10, q10, q8
+	 vand		q11, q11, q8
+	veor		q5, q5, q10
+	vshl.u64	q10, q10, #4
+	 veor		q4, q4, q11
+	 vshl.u64	q11, q11, #4
+	veor		q1, q1, q10
+	 veor		q0, q0, q11
+	sub	r5,r5,#1
+	b	.Ldec_sbox
+.align	4
+.Ldec_loop:
+	vldmia	r4!, {q8-q11}
+	veor	q8, q8, q0
+	veor	q9, q9, q1
+	vtbl.8	d0, {q8}, d24
+	vtbl.8	d1, {q8}, d25
+	vldmia	r4!, {q8}
+	veor	q10, q10, q2
+	vtbl.8	d2, {q9}, d24
+	vtbl.8	d3, {q9}, d25
+	vldmia	r4!, {q9}
+	veor	q11, q11, q3
+	vtbl.8	d4, {q10}, d24
+	vtbl.8	d5, {q10}, d25
+	vldmia	r4!, {q10}
+	vtbl.8	d6, {q11}, d24
+	vtbl.8	d7, {q11}, d25
+	vldmia	r4!, {q11}
+	veor	q8, q8, q4
+	veor	q9, q9, q5
+	vtbl.8	d8, {q8}, d24
+	vtbl.8	d9, {q8}, d25
+	veor	q10, q10, q6
+	vtbl.8	d10, {q9}, d24
+	vtbl.8	d11, {q9}, d25
+	veor	q11, q11, q7
+	vtbl.8	d12, {q10}, d24
+	vtbl.8	d13, {q10}, d25
+	vtbl.8	d14, {q11}, d24
+	vtbl.8	d15, {q11}, d25
+.Ldec_sbox:
+	 veor	q1, q1, q4
+	veor	q3, q3, q4
+
+	veor	q4, q4, q7
+	 veor	q1, q1, q6
+	veor	q2, q2, q7
+	veor	q6, q6, q4
+
+	veor	q0, q0, q1
+	veor	q2, q2, q5
+	 veor	q7, q7, q6
+	veor	q3, q3, q0
+	veor	q5, q5, q0
+	veor	q1, q1, q3
+	veor	q11, q3, q0
+	veor	q10, q7, q4
+	veor	q9, q1, q6
+	veor	q13, q4, q0
+	 vmov	q8, q10
+	veor	q12, q5, q2
+
+	vorr	q10, q10, q9
+	veor	q15, q11, q8
+	vand	q14, q11, q12
+	vorr	q11, q11, q12
+	veor	q12, q12, q9
+	vand	q8, q8, q9
+	veor	q9, q6, q2
+	vand	q15, q15, q12
+	vand	q13, q13, q9
+	veor	q9, q3, q7
+	veor	q12, q1, q5
+	veor	q11, q11, q13
+	veor	q10, q10, q13
+	vand	q13, q9, q12
+	vorr	q9, q9, q12
+	veor	q11, q11, q15
+	veor	q8, q8, q13
+	veor	q10, q10, q14
+	veor	q9, q9, q15
+	veor	q8, q8, q14
+	vand	q12, q4, q6
+	veor	q9, q9, q14
+	vand	q13, q0, q2
+	vand	q14, q7, q1
+	vorr	q15, q3, q5
+	veor	q11, q11, q12
+	veor	q9, q9, q14
+	veor	q8, q8, q15
+	veor	q10, q10, q13
+
+	@ Inv_GF16 	0, 	1, 	2, 	3, s0, s1, s2, s3
+
+	@ new smaller inversion
+
+	vand	q14, q11, q9
+	vmov	q12, q8
+
+	veor	q13, q10, q14
+	veor	q15, q8, q14
+	veor	q14, q8, q14	@ q14=q15
+
+	vbsl	q13, q9, q8
+	vbsl	q15, q11, q10
+	veor	q11, q11, q10
+
+	vbsl	q12, q13, q14
+	vbsl	q8, q14, q13
+
+	vand	q14, q12, q15
+	veor	q9, q9, q8
+
+	veor	q14, q14, q11
+	veor	q12, q5, q2
+	veor	q8, q1, q6
+	veor 	q10, q15, q14
+	vand	q10, q10, q5
+	veor	q5, q5, q1
+	vand	q11, q1, q15
+	vand	q5, q5, q14
+	veor	q1, q11, q10
+	veor	q5, q5, q11
+	veor	q15, q15, q13
+	veor	q14, q14, q9
+	veor	q11, q15, q14
+	 veor 	q10, q13, q9
+	vand	q11, q11, q12
+	 vand	q10, q10, q2
+	veor	q12, q12, q8
+	 veor	q2, q2, q6
+	vand	q8, q8, q15
+	 vand	q6, q6, q13
+	vand	q12, q12, q14
+	 vand	q2, q2, q9
+	veor	q8, q8, q12
+	 veor	q2, q2, q6
+	veor	q12, q12, q11
+	 veor	q6, q6, q10
+	veor	q5, q5, q12
+	veor	q2, q2, q12
+	veor	q1, q1, q8
+	veor	q6, q6, q8
+
+	veor	q12, q3, q0
+	veor	q8, q7, q4
+	veor	q11, q15, q14
+	 veor 	q10, q13, q9
+	vand	q11, q11, q12
+	 vand	q10, q10, q0
+	veor	q12, q12, q8
+	 veor	q0, q0, q4
+	vand	q8, q8, q15
+	 vand	q4, q4, q13
+	vand	q12, q12, q14
+	 vand	q0, q0, q9
+	veor	q8, q8, q12
+	 veor	q0, q0, q4
+	veor	q12, q12, q11
+	 veor	q4, q4, q10
+	veor	q15, q15, q13
+	veor	q14, q14, q9
+	veor 	q10, q15, q14
+	vand	q10, q10, q3
+	veor	q3, q3, q7
+	vand	q11, q7, q15
+	vand	q3, q3, q14
+	veor	q7, q11, q10
+	veor	q3, q3, q11
+	veor	q3, q3, q12
+	veor	q0, q0, q12
+	veor	q7, q7, q8
+	veor	q4, q4, q8
+	veor	q1, q1, q7
+	veor	q6, q6, q5
+
+	veor	q4, q4, q1
+	veor	q2, q2, q7
+	veor	q5, q5, q7
+	veor	q4, q4, q2
+	 veor 	q7, q7, q0
+	veor	q4, q4, q5
+	 veor	q3, q3, q6
+	 veor	q6, q6, q1
+	veor	q3, q3, q4
+
+	veor	q4, q4, q0
+	veor	q7, q7, q3
+	subs	r5,r5,#1
+	bcc	.Ldec_done
+	@ multiplication by 0x05-0x00-0x04-0x00
+	vext.8	q8, q0, q0, #8
+	vext.8	q14, q3, q3, #8
+	vext.8	q15, q5, q5, #8
+	veor	q8, q8, q0
+	vext.8	q9, q1, q1, #8
+	veor	q14, q14, q3
+	vext.8	q10, q6, q6, #8
+	veor	q15, q15, q5
+	vext.8	q11, q4, q4, #8
+	veor	q9, q9, q1
+	vext.8	q12, q2, q2, #8
+	veor	q10, q10, q6
+	vext.8	q13, q7, q7, #8
+	veor	q11, q11, q4
+	veor	q12, q12, q2
+	veor	q13, q13, q7
+
+	 veor	q0, q0, q14
+	 veor	q1, q1, q14
+	 veor	q6, q6, q8
+	 veor	q2, q2, q10
+	 veor	q4, q4, q9
+	 veor	q1, q1, q15
+	 veor	q6, q6, q15
+	 veor	q2, q2, q14
+	 veor	q7, q7, q11
+	 veor	q4, q4, q14
+	 veor	q3, q3, q12
+	 veor	q2, q2, q15
+	 veor	q7, q7, q15
+	 veor	q5, q5, q13
+	vext.8	q8, q0, q0, #12	@ x0 <<< 32
+	vext.8	q9, q1, q1, #12
+	 veor	q0, q0, q8		@ x0 ^ (x0 <<< 32)
+	vext.8	q10, q6, q6, #12
+	 veor	q1, q1, q9
+	vext.8	q11, q4, q4, #12
+	 veor	q6, q6, q10
+	vext.8	q12, q2, q2, #12
+	 veor	q4, q4, q11
+	vext.8	q13, q7, q7, #12
+	 veor	q2, q2, q12
+	vext.8	q14, q3, q3, #12
+	 veor	q7, q7, q13
+	vext.8	q15, q5, q5, #12
+	 veor	q3, q3, q14
+
+	veor	q9, q9, q0
+	 veor	q5, q5, q15
+	 vext.8	q0, q0, q0, #8		@ (x0 ^ (x0 <<< 32)) <<< 64)
+	veor	q10, q10, q1
+	veor	q8, q8, q5
+	veor	q9, q9, q5
+	 vext.8	q1, q1, q1, #8
+	veor	q13, q13, q2
+	 veor	q0, q0, q8
+	veor	q14, q14, q7
+	 veor	q1, q1, q9
+	 vext.8	q8, q2, q2, #8
+	veor	q12, q12, q4
+	 vext.8	q9, q7, q7, #8
+	veor	q15, q15, q3
+	 vext.8	q2, q4, q4, #8
+	veor	q11, q11, q6
+	 vext.8	q7, q5, q5, #8
+	veor	q12, q12, q5
+	 vext.8	q4, q3, q3, #8
+	veor	q11, q11, q5
+	 vext.8	q3, q6, q6, #8
+	veor	q5, q9, q13
+	veor	q11, q11, q2
+	veor	q7, q7, q15
+	veor	q6, q4, q14
+	veor	q4, q8, q12
+	veor	q2, q3, q10
+	vmov	q3, q11
+	 @ vmov	q5, q9
+	vldmia	r6, {q12}		@ .LISR
+	ite	eq				@ Thumb2 thing, sanity check in ARM
+	addeq	r6,r6,#0x10
+	bne	.Ldec_loop
+	vldmia	r6, {q12}		@ .LISRM0
+	b	.Ldec_loop
+.align	4
+.Ldec_done:
+	vmov.i8	q8,#0x55			@ compose .LBS0
+	vmov.i8	q9,#0x33			@ compose .LBS1
+	vshr.u64	q10, q3, #1
+	 vshr.u64	q11, q2, #1
+	veor		q10, q10, q5
+	 veor		q11, q11, q7
+	vand		q10, q10, q8
+	 vand		q11, q11, q8
+	veor		q5, q5, q10
+	vshl.u64	q10, q10, #1
+	 veor		q7, q7, q11
+	 vshl.u64	q11, q11, #1
+	veor		q3, q3, q10
+	 veor		q2, q2, q11
+	vshr.u64	q10, q6, #1
+	 vshr.u64	q11, q0, #1
+	veor		q10, q10, q4
+	 veor		q11, q11, q1
+	vand		q10, q10, q8
+	 vand		q11, q11, q8
+	veor		q4, q4, q10
+	vshl.u64	q10, q10, #1
+	 veor		q1, q1, q11
+	 vshl.u64	q11, q11, #1
+	veor		q6, q6, q10
+	 veor		q0, q0, q11
+	vmov.i8	q8,#0x0f			@ compose .LBS2
+	vshr.u64	q10, q7, #2
+	 vshr.u64	q11, q2, #2
+	veor		q10, q10, q5
+	 veor		q11, q11, q3
+	vand		q10, q10, q9
+	 vand		q11, q11, q9
+	veor		q5, q5, q10
+	vshl.u64	q10, q10, #2
+	 veor		q3, q3, q11
+	 vshl.u64	q11, q11, #2
+	veor		q7, q7, q10
+	 veor		q2, q2, q11
+	vshr.u64	q10, q1, #2
+	 vshr.u64	q11, q0, #2
+	veor		q10, q10, q4
+	 veor		q11, q11, q6
+	vand		q10, q10, q9
+	 vand		q11, q11, q9
+	veor		q4, q4, q10
+	vshl.u64	q10, q10, #2
+	 veor		q6, q6, q11
+	 vshl.u64	q11, q11, #2
+	veor		q1, q1, q10
+	 veor		q0, q0, q11
+	vshr.u64	q10, q4, #4
+	 vshr.u64	q11, q6, #4
+	veor		q10, q10, q5
+	 veor		q11, q11, q3
+	vand		q10, q10, q8
+	 vand		q11, q11, q8
+	veor		q5, q5, q10
+	vshl.u64	q10, q10, #4
+	 veor		q3, q3, q11
+	 vshl.u64	q11, q11, #4
+	veor		q4, q4, q10
+	 veor		q6, q6, q11
+	vshr.u64	q10, q1, #4
+	 vshr.u64	q11, q0, #4
+	veor		q10, q10, q7
+	 veor		q11, q11, q2
+	vand		q10, q10, q8
+	 vand		q11, q11, q8
+	veor		q7, q7, q10
+	vshl.u64	q10, q10, #4
+	 veor		q2, q2, q11
+	 vshl.u64	q11, q11, #4
+	veor		q1, q1, q10
+	 veor		q0, q0, q11
+	vldmia	r4, {q8}			@ last round key
+	veor	q6, q6, q8
+	veor	q4, q4, q8
+	veor	q2, q2, q8
+	veor	q7, q7, q8
+	veor	q3, q3, q8
+	veor	q5, q5, q8
+	veor	q0, q0, q8
+	veor	q1, q1, q8
+	bx	lr
+.size	_bsaes_decrypt8,.-_bsaes_decrypt8
+
+.type	_bsaes_const,%object
+.align	6
+_bsaes_const:
+.LM0ISR:	@ InvShiftRows constants
+	.quad	0x0a0e0206070b0f03, 0x0004080c0d010509
+.LISR:
+	.quad	0x0504070602010003, 0x0f0e0d0c080b0a09
+.LISRM0:
+	.quad	0x01040b0e0205080f, 0x0306090c00070a0d
+.LM0SR:		@ ShiftRows constants
+	.quad	0x0a0e02060f03070b, 0x0004080c05090d01
+.LSR:
+	.quad	0x0504070600030201, 0x0f0e0d0c0a09080b
+.LSRM0:
+	.quad	0x0304090e00050a0f, 0x01060b0c0207080d
+.LM0:
+	.quad	0x02060a0e03070b0f, 0x0004080c0105090d
+.LREVM0SR:
+	.quad	0x090d01050c000408, 0x03070b0f060a0e02
+.asciz	"Bit-sliced AES for NEON, CRYPTOGAMS by <appro@openssl.org>"
+.align	6
+.size	_bsaes_const,.-_bsaes_const
+
+.type	_bsaes_encrypt8,%function
+.align	4
+_bsaes_encrypt8:
+	adr	r6,_bsaes_encrypt8
+	vldmia	r4!, {q9}		@ round 0 key
+	sub	r6,r6,#_bsaes_encrypt8-.LM0SR
+
+	vldmia	r6!, {q8}		@ .LM0SR
+_bsaes_encrypt8_alt:
+	veor	q10, q0, q9	@ xor with round0 key
+	veor	q11, q1, q9
+	 vtbl.8	d0, {q10}, d16
+	 vtbl.8	d1, {q10}, d17
+	veor	q12, q2, q9
+	 vtbl.8	d2, {q11}, d16
+	 vtbl.8	d3, {q11}, d17
+	veor	q13, q3, q9
+	 vtbl.8	d4, {q12}, d16
+	 vtbl.8	d5, {q12}, d17
+	veor	q14, q4, q9
+	 vtbl.8	d6, {q13}, d16
+	 vtbl.8	d7, {q13}, d17
+	veor	q15, q5, q9
+	 vtbl.8	d8, {q14}, d16
+	 vtbl.8	d9, {q14}, d17
+	veor	q10, q6, q9
+	 vtbl.8	d10, {q15}, d16
+	 vtbl.8	d11, {q15}, d17
+	veor	q11, q7, q9
+	 vtbl.8	d12, {q10}, d16
+	 vtbl.8	d13, {q10}, d17
+	 vtbl.8	d14, {q11}, d16
+	 vtbl.8	d15, {q11}, d17
+_bsaes_encrypt8_bitslice:
+	vmov.i8	q8,#0x55			@ compose .LBS0
+	vmov.i8	q9,#0x33			@ compose .LBS1
+	vshr.u64	q10, q6, #1
+	 vshr.u64	q11, q4, #1
+	veor		q10, q10, q7
+	 veor		q11, q11, q5
+	vand		q10, q10, q8
+	 vand		q11, q11, q8
+	veor		q7, q7, q10
+	vshl.u64	q10, q10, #1
+	 veor		q5, q5, q11
+	 vshl.u64	q11, q11, #1
+	veor		q6, q6, q10
+	 veor		q4, q4, q11
+	vshr.u64	q10, q2, #1
+	 vshr.u64	q11, q0, #1
+	veor		q10, q10, q3
+	 veor		q11, q11, q1
+	vand		q10, q10, q8
+	 vand		q11, q11, q8
+	veor		q3, q3, q10
+	vshl.u64	q10, q10, #1
+	 veor		q1, q1, q11
+	 vshl.u64	q11, q11, #1
+	veor		q2, q2, q10
+	 veor		q0, q0, q11
+	vmov.i8	q8,#0x0f			@ compose .LBS2
+	vshr.u64	q10, q5, #2
+	 vshr.u64	q11, q4, #2
+	veor		q10, q10, q7
+	 veor		q11, q11, q6
+	vand		q10, q10, q9
+	 vand		q11, q11, q9
+	veor		q7, q7, q10
+	vshl.u64	q10, q10, #2
+	 veor		q6, q6, q11
+	 vshl.u64	q11, q11, #2
+	veor		q5, q5, q10
+	 veor		q4, q4, q11
+	vshr.u64	q10, q1, #2
+	 vshr.u64	q11, q0, #2
+	veor		q10, q10, q3
+	 veor		q11, q11, q2
+	vand		q10, q10, q9
+	 vand		q11, q11, q9
+	veor		q3, q3, q10
+	vshl.u64	q10, q10, #2
+	 veor		q2, q2, q11
+	 vshl.u64	q11, q11, #2
+	veor		q1, q1, q10
+	 veor		q0, q0, q11
+	vshr.u64	q10, q3, #4
+	 vshr.u64	q11, q2, #4
+	veor		q10, q10, q7
+	 veor		q11, q11, q6
+	vand		q10, q10, q8
+	 vand		q11, q11, q8
+	veor		q7, q7, q10
+	vshl.u64	q10, q10, #4
+	 veor		q6, q6, q11
+	 vshl.u64	q11, q11, #4
+	veor		q3, q3, q10
+	 veor		q2, q2, q11
+	vshr.u64	q10, q1, #4
+	 vshr.u64	q11, q0, #4
+	veor		q10, q10, q5
+	 veor		q11, q11, q4
+	vand		q10, q10, q8
+	 vand		q11, q11, q8
+	veor		q5, q5, q10
+	vshl.u64	q10, q10, #4
+	 veor		q4, q4, q11
+	 vshl.u64	q11, q11, #4
+	veor		q1, q1, q10
+	 veor		q0, q0, q11
+	sub	r5,r5,#1
+	b	.Lenc_sbox
+.align	4
+.Lenc_loop:
+	vldmia	r4!, {q8-q11}
+	veor	q8, q8, q0
+	veor	q9, q9, q1
+	vtbl.8	d0, {q8}, d24
+	vtbl.8	d1, {q8}, d25
+	vldmia	r4!, {q8}
+	veor	q10, q10, q2
+	vtbl.8	d2, {q9}, d24
+	vtbl.8	d3, {q9}, d25
+	vldmia	r4!, {q9}
+	veor	q11, q11, q3
+	vtbl.8	d4, {q10}, d24
+	vtbl.8	d5, {q10}, d25
+	vldmia	r4!, {q10}
+	vtbl.8	d6, {q11}, d24
+	vtbl.8	d7, {q11}, d25
+	vldmia	r4!, {q11}
+	veor	q8, q8, q4
+	veor	q9, q9, q5
+	vtbl.8	d8, {q8}, d24
+	vtbl.8	d9, {q8}, d25
+	veor	q10, q10, q6
+	vtbl.8	d10, {q9}, d24
+	vtbl.8	d11, {q9}, d25
+	veor	q11, q11, q7
+	vtbl.8	d12, {q10}, d24
+	vtbl.8	d13, {q10}, d25
+	vtbl.8	d14, {q11}, d24
+	vtbl.8	d15, {q11}, d25
+.Lenc_sbox:
+	veor	q2, q2, q1
+	veor	q5, q5, q6
+	veor	q3, q3, q0
+	veor	q6, q6, q2
+	veor	q5, q5, q0
+
+	veor	q6, q6, q3
+	veor	q3, q3, q7
+	veor	q7, q7, q5
+	veor	q3, q3, q4
+	veor	q4, q4, q5
+
+	veor	q2, q2, q7
+	veor	q3, q3, q1
+	veor	q1, q1, q5
+	veor	q11, q7, q4
+	veor	q10, q1, q2
+	veor	q9, q5, q3
+	veor	q13, q2, q4
+	 vmov	q8, q10
+	veor	q12, q6, q0
+
+	vorr	q10, q10, q9
+	veor	q15, q11, q8
+	vand	q14, q11, q12
+	vorr	q11, q11, q12
+	veor	q12, q12, q9
+	vand	q8, q8, q9
+	veor	q9, q3, q0
+	vand	q15, q15, q12
+	vand	q13, q13, q9
+	veor	q9, q7, q1
+	veor	q12, q5, q6
+	veor	q11, q11, q13
+	veor	q10, q10, q13
+	vand	q13, q9, q12
+	vorr	q9, q9, q12
+	veor	q11, q11, q15
+	veor	q8, q8, q13
+	veor	q10, q10, q14
+	veor	q9, q9, q15
+	veor	q8, q8, q14
+	vand	q12, q2, q3
+	veor	q9, q9, q14
+	vand	q13, q4, q0
+	vand	q14, q1, q5
+	vorr	q15, q7, q6
+	veor	q11, q11, q12
+	veor	q9, q9, q14
+	veor	q8, q8, q15
+	veor	q10, q10, q13
+
+	@ Inv_GF16 	0, 	1, 	2, 	3, s0, s1, s2, s3
+
+	@ new smaller inversion
+
+	vand	q14, q11, q9
+	vmov	q12, q8
+
+	veor	q13, q10, q14
+	veor	q15, q8, q14
+	veor	q14, q8, q14	@ q14=q15
+
+	vbsl	q13, q9, q8
+	vbsl	q15, q11, q10
+	veor	q11, q11, q10
+
+	vbsl	q12, q13, q14
+	vbsl	q8, q14, q13
+
+	vand	q14, q12, q15
+	veor	q9, q9, q8
+
+	veor	q14, q14, q11
+	veor	q12, q6, q0
+	veor	q8, q5, q3
+	veor 	q10, q15, q14
+	vand	q10, q10, q6
+	veor	q6, q6, q5
+	vand	q11, q5, q15
+	vand	q6, q6, q14
+	veor	q5, q11, q10
+	veor	q6, q6, q11
+	veor	q15, q15, q13
+	veor	q14, q14, q9
+	veor	q11, q15, q14
+	 veor 	q10, q13, q9
+	vand	q11, q11, q12
+	 vand	q10, q10, q0
+	veor	q12, q12, q8
+	 veor	q0, q0, q3
+	vand	q8, q8, q15
+	 vand	q3, q3, q13
+	vand	q12, q12, q14
+	 vand	q0, q0, q9
+	veor	q8, q8, q12
+	 veor	q0, q0, q3
+	veor	q12, q12, q11
+	 veor	q3, q3, q10
+	veor	q6, q6, q12
+	veor	q0, q0, q12
+	veor	q5, q5, q8
+	veor	q3, q3, q8
+
+	veor	q12, q7, q4
+	veor	q8, q1, q2
+	veor	q11, q15, q14
+	 veor 	q10, q13, q9
+	vand	q11, q11, q12
+	 vand	q10, q10, q4
+	veor	q12, q12, q8
+	 veor	q4, q4, q2
+	vand	q8, q8, q15
+	 vand	q2, q2, q13
+	vand	q12, q12, q14
+	 vand	q4, q4, q9
+	veor	q8, q8, q12
+	 veor	q4, q4, q2
+	veor	q12, q12, q11
+	 veor	q2, q2, q10
+	veor	q15, q15, q13
+	veor	q14, q14, q9
+	veor 	q10, q15, q14
+	vand	q10, q10, q7
+	veor	q7, q7, q1
+	vand	q11, q1, q15
+	vand	q7, q7, q14
+	veor	q1, q11, q10
+	veor	q7, q7, q11
+	veor	q7, q7, q12
+	veor	q4, q4, q12
+	veor	q1, q1, q8
+	veor	q2, q2, q8
+	veor	q7, q7, q0
+	veor	q1, q1, q6
+	veor	q6, q6, q0
+	veor	q4, q4, q7
+	veor	q0, q0, q1
+
+	veor	q1, q1, q5
+	veor	q5, q5, q2
+	veor	q2, q2, q3
+	veor	q3, q3, q5
+	veor	q4, q4, q5
+
+	veor	q6, q6, q3
+	subs	r5,r5,#1
+	bcc	.Lenc_done
+	vext.8	q8, q0, q0, #12	@ x0 <<< 32
+	vext.8	q9, q1, q1, #12
+	 veor	q0, q0, q8		@ x0 ^ (x0 <<< 32)
+	vext.8	q10, q4, q4, #12
+	 veor	q1, q1, q9
+	vext.8	q11, q6, q6, #12
+	 veor	q4, q4, q10
+	vext.8	q12, q3, q3, #12
+	 veor	q6, q6, q11
+	vext.8	q13, q7, q7, #12
+	 veor	q3, q3, q12
+	vext.8	q14, q2, q2, #12
+	 veor	q7, q7, q13
+	vext.8	q15, q5, q5, #12
+	 veor	q2, q2, q14
+
+	veor	q9, q9, q0
+	 veor	q5, q5, q15
+	 vext.8	q0, q0, q0, #8		@ (x0 ^ (x0 <<< 32)) <<< 64)
+	veor	q10, q10, q1
+	veor	q8, q8, q5
+	veor	q9, q9, q5
+	 vext.8	q1, q1, q1, #8
+	veor	q13, q13, q3
+	 veor	q0, q0, q8
+	veor	q14, q14, q7
+	 veor	q1, q1, q9
+	 vext.8	q8, q3, q3, #8
+	veor	q12, q12, q6
+	 vext.8	q9, q7, q7, #8
+	veor	q15, q15, q2
+	 vext.8	q3, q6, q6, #8
+	veor	q11, q11, q4
+	 vext.8	q7, q5, q5, #8
+	veor	q12, q12, q5
+	 vext.8	q6, q2, q2, #8
+	veor	q11, q11, q5
+	 vext.8	q2, q4, q4, #8
+	veor	q5, q9, q13
+	veor	q4, q8, q12
+	veor	q3, q3, q11
+	veor	q7, q7, q15
+	veor	q6, q6, q14
+	 @ vmov	q4, q8
+	veor	q2, q2, q10
+	 @ vmov	q5, q9
+	vldmia	r6, {q12}		@ .LSR
+	ite	eq				@ Thumb2 thing, samity check in ARM
+	addeq	r6,r6,#0x10
+	bne	.Lenc_loop
+	vldmia	r6, {q12}		@ .LSRM0
+	b	.Lenc_loop
+.align	4
+.Lenc_done:
+	vmov.i8	q8,#0x55			@ compose .LBS0
+	vmov.i8	q9,#0x33			@ compose .LBS1
+	vshr.u64	q10, q2, #1
+	 vshr.u64	q11, q3, #1
+	veor		q10, q10, q5
+	 veor		q11, q11, q7
+	vand		q10, q10, q8
+	 vand		q11, q11, q8
+	veor		q5, q5, q10
+	vshl.u64	q10, q10, #1
+	 veor		q7, q7, q11
+	 vshl.u64	q11, q11, #1
+	veor		q2, q2, q10
+	 veor		q3, q3, q11
+	vshr.u64	q10, q4, #1
+	 vshr.u64	q11, q0, #1
+	veor		q10, q10, q6
+	 veor		q11, q11, q1
+	vand		q10, q10, q8
+	 vand		q11, q11, q8
+	veor		q6, q6, q10
+	vshl.u64	q10, q10, #1
+	 veor		q1, q1, q11
+	 vshl.u64	q11, q11, #1
+	veor		q4, q4, q10
+	 veor		q0, q0, q11
+	vmov.i8	q8,#0x0f			@ compose .LBS2
+	vshr.u64	q10, q7, #2
+	 vshr.u64	q11, q3, #2
+	veor		q10, q10, q5
+	 veor		q11, q11, q2
+	vand		q10, q10, q9
+	 vand		q11, q11, q9
+	veor		q5, q5, q10
+	vshl.u64	q10, q10, #2
+	 veor		q2, q2, q11
+	 vshl.u64	q11, q11, #2
+	veor		q7, q7, q10
+	 veor		q3, q3, q11
+	vshr.u64	q10, q1, #2
+	 vshr.u64	q11, q0, #2
+	veor		q10, q10, q6
+	 veor		q11, q11, q4
+	vand		q10, q10, q9
+	 vand		q11, q11, q9
+	veor		q6, q6, q10
+	vshl.u64	q10, q10, #2
+	 veor		q4, q4, q11
+	 vshl.u64	q11, q11, #2
+	veor		q1, q1, q10
+	 veor		q0, q0, q11
+	vshr.u64	q10, q6, #4
+	 vshr.u64	q11, q4, #4
+	veor		q10, q10, q5
+	 veor		q11, q11, q2
+	vand		q10, q10, q8
+	 vand		q11, q11, q8
+	veor		q5, q5, q10
+	vshl.u64	q10, q10, #4
+	 veor		q2, q2, q11
+	 vshl.u64	q11, q11, #4
+	veor		q6, q6, q10
+	 veor		q4, q4, q11
+	vshr.u64	q10, q1, #4
+	 vshr.u64	q11, q0, #4
+	veor		q10, q10, q7
+	 veor		q11, q11, q3
+	vand		q10, q10, q8
+	 vand		q11, q11, q8
+	veor		q7, q7, q10
+	vshl.u64	q10, q10, #4
+	 veor		q3, q3, q11
+	 vshl.u64	q11, q11, #4
+	veor		q1, q1, q10
+	 veor		q0, q0, q11
+	vldmia	r4, {q8}			@ last round key
+	veor	q4, q4, q8
+	veor	q6, q6, q8
+	veor	q3, q3, q8
+	veor	q7, q7, q8
+	veor	q2, q2, q8
+	veor	q5, q5, q8
+	veor	q0, q0, q8
+	veor	q1, q1, q8
+	bx	lr
+.size	_bsaes_encrypt8,.-_bsaes_encrypt8
+.type	_bsaes_key_convert,%function
+.align	4
+_bsaes_key_convert:
+	adr	r6,_bsaes_key_convert
+	vld1.8	{q7},  [r4]!		@ load round 0 key
+	sub	r6,r6,#_bsaes_key_convert-.LM0
+	vld1.8	{q15}, [r4]!		@ load round 1 key
+
+	vmov.i8	q8,  #0x01			@ bit masks
+	vmov.i8	q9,  #0x02
+	vmov.i8	q10, #0x04
+	vmov.i8	q11, #0x08
+	vmov.i8	q12, #0x10
+	vmov.i8	q13, #0x20
+	vldmia	r6, {q14}		@ .LM0
+
+#ifdef __ARMEL__
+	vrev32.8	q7,  q7
+	vrev32.8	q15, q15
+#endif
+	sub	r5,r5,#1
+	vstmia	r12!, {q7}		@ save round 0 key
+	b	.Lkey_loop
+
+.align	4
+.Lkey_loop:
+	vtbl.8	d14,{q15},d28
+	vtbl.8	d15,{q15},d29
+	vmov.i8	q6,  #0x40
+	vmov.i8	q15, #0x80
+
+	vtst.8	q0, q7, q8
+	vtst.8	q1, q7, q9
+	vtst.8	q2, q7, q10
+	vtst.8	q3, q7, q11
+	vtst.8	q4, q7, q12
+	vtst.8	q5, q7, q13
+	vtst.8	q6, q7, q6
+	vtst.8	q7, q7, q15
+	vld1.8	{q15}, [r4]!		@ load next round key
+	vmvn	q0, q0		@ "pnot"
+	vmvn	q1, q1
+	vmvn	q5, q5
+	vmvn	q6, q6
+#ifdef __ARMEL__
+	vrev32.8	q15, q15
+#endif
+	subs	r5,r5,#1
+	vstmia	r12!,{q0-q7}		@ write bit-sliced round key
+	bne	.Lkey_loop
+
+	vmov.i8	q7,#0x63			@ compose .L63
+	@ don't save last round key
+	bx	lr
+.size	_bsaes_key_convert,.-_bsaes_key_convert
+.extern AES_cbc_encrypt
+.extern AES_decrypt
+
+.global	bsaes_cbc_encrypt
+.type	bsaes_cbc_encrypt,%function
+.align	5
+bsaes_cbc_encrypt:
+#ifndef	__KERNEL__
+	cmp	r2, #128
+#ifndef	__thumb__
+	blo	AES_cbc_encrypt
+#else
+	bhs	1f
+	b	AES_cbc_encrypt
+1:
+#endif
+#endif
+
+	@ it is up to the caller to make sure we are called with enc == 0
+
+	mov	ip, sp
+	stmdb	sp!, {r4-r10, lr}
+	VFP_ABI_PUSH
+	ldr	r8, [ip]			@ IV is 1st arg on the stack
+	mov	r2, r2, lsr#4		@ len in 16 byte blocks
+	sub	sp, #0x10			@ scratch space to carry over the IV
+	mov	r9, sp				@ save sp
+
+	ldr	r10, [r3, #240]		@ get # of rounds
+#ifndef	BSAES_ASM_EXTENDED_KEY
+	@ allocate the key schedule on the stack
+	sub	r12, sp, r10, lsl#7		@ 128 bytes per inner round key
+	add	r12, #96			@ sifze of bit-slices key schedule
+
+	@ populate the key schedule
+	mov	r4, r3			@ pass key
+	mov	r5, r10			@ pass # of rounds
+	mov	sp, r12				@ sp is sp
+	bl	_bsaes_key_convert
+	vldmia	sp, {q6}
+	vstmia	r12,  {q15}		@ save last round key
+	veor	q7, q7, q6	@ fix up round 0 key
+	vstmia	sp, {q7}
+#else
+	ldr	r12, [r3, #244]
+	eors	r12, #1
+	beq	0f
+
+	@ populate the key schedule
+	str	r12, [r3, #244]
+	mov	r4, r3			@ pass key
+	mov	r5, r10			@ pass # of rounds
+	add	r12, r3, #248			@ pass key schedule
+	bl	_bsaes_key_convert
+	add	r4, r3, #248
+	vldmia	r4, {q6}
+	vstmia	r12, {q15}			@ save last round key
+	veor	q7, q7, q6	@ fix up round 0 key
+	vstmia	r4, {q7}
+
+.align	2
+0:
+#endif
+
+	vld1.8	{q15}, [r8]		@ load IV
+	b	.Lcbc_dec_loop
+
+.align	4
+.Lcbc_dec_loop:
+	subs	r2, r2, #0x8
+	bmi	.Lcbc_dec_loop_finish
+
+	vld1.8	{q0-q1}, [r0]!	@ load input
+	vld1.8	{q2-q3}, [r0]!
+#ifndef	BSAES_ASM_EXTENDED_KEY
+	mov	r4, sp			@ pass the key
+#else
+	add	r4, r3, #248
+#endif
+	vld1.8	{q4-q5}, [r0]!
+	mov	r5, r10
+	vld1.8	{q6-q7}, [r0]
+	sub	r0, r0, #0x60
+	vstmia	r9, {q15}			@ put aside IV
+
+	bl	_bsaes_decrypt8
+
+	vldmia	r9, {q14}			@ reload IV
+	vld1.8	{q8-q9}, [r0]!	@ reload input
+	veor	q0, q0, q14	@ ^= IV
+	vld1.8	{q10-q11}, [r0]!
+	veor	q1, q1, q8
+	veor	q6, q6, q9
+	vld1.8	{q12-q13}, [r0]!
+	veor	q4, q4, q10
+	veor	q2, q2, q11
+	vld1.8	{q14-q15}, [r0]!
+	veor	q7, q7, q12
+	vst1.8	{q0-q1}, [r1]!	@ write output
+	veor	q3, q3, q13
+	vst1.8	{q6}, [r1]!
+	veor	q5, q5, q14
+	vst1.8	{q4}, [r1]!
+	vst1.8	{q2}, [r1]!
+	vst1.8	{q7}, [r1]!
+	vst1.8	{q3}, [r1]!
+	vst1.8	{q5}, [r1]!
+
+	b	.Lcbc_dec_loop
+
+.Lcbc_dec_loop_finish:
+	adds	r2, r2, #8
+	beq	.Lcbc_dec_done
+
+	vld1.8	{q0}, [r0]!		@ load input
+	cmp	r2, #2
+	blo	.Lcbc_dec_one
+	vld1.8	{q1}, [r0]!
+#ifndef	BSAES_ASM_EXTENDED_KEY
+	mov	r4, sp			@ pass the key
+#else
+	add	r4, r3, #248
+#endif
+	mov	r5, r10
+	vstmia	r9, {q15}			@ put aside IV
+	beq	.Lcbc_dec_two
+	vld1.8	{q2}, [r0]!
+	cmp	r2, #4
+	blo	.Lcbc_dec_three
+	vld1.8	{q3}, [r0]!
+	beq	.Lcbc_dec_four
+	vld1.8	{q4}, [r0]!
+	cmp	r2, #6
+	blo	.Lcbc_dec_five
+	vld1.8	{q5}, [r0]!
+	beq	.Lcbc_dec_six
+	vld1.8	{q6}, [r0]!
+	sub	r0, r0, #0x70
+
+	bl	_bsaes_decrypt8
+
+	vldmia	r9, {q14}			@ reload IV
+	vld1.8	{q8-q9}, [r0]!	@ reload input
+	veor	q0, q0, q14	@ ^= IV
+	vld1.8	{q10-q11}, [r0]!
+	veor	q1, q1, q8
+	veor	q6, q6, q9
+	vld1.8	{q12-q13}, [r0]!
+	veor	q4, q4, q10
+	veor	q2, q2, q11
+	vld1.8	{q15}, [r0]!
+	veor	q7, q7, q12
+	vst1.8	{q0-q1}, [r1]!	@ write output
+	veor	q3, q3, q13
+	vst1.8	{q6}, [r1]!
+	vst1.8	{q4}, [r1]!
+	vst1.8	{q2}, [r1]!
+	vst1.8	{q7}, [r1]!
+	vst1.8	{q3}, [r1]!
+	b	.Lcbc_dec_done
+.align	4
+.Lcbc_dec_six:
+	sub	r0, r0, #0x60
+	bl	_bsaes_decrypt8
+	vldmia	r9,{q14}			@ reload IV
+	vld1.8	{q8-q9}, [r0]!	@ reload input
+	veor	q0, q0, q14	@ ^= IV
+	vld1.8	{q10-q11}, [r0]!
+	veor	q1, q1, q8
+	veor	q6, q6, q9
+	vld1.8	{q12}, [r0]!
+	veor	q4, q4, q10
+	veor	q2, q2, q11
+	vld1.8	{q15}, [r0]!
+	veor	q7, q7, q12
+	vst1.8	{q0-q1}, [r1]!	@ write output
+	vst1.8	{q6}, [r1]!
+	vst1.8	{q4}, [r1]!
+	vst1.8	{q2}, [r1]!
+	vst1.8	{q7}, [r1]!
+	b	.Lcbc_dec_done
+.align	4
+.Lcbc_dec_five:
+	sub	r0, r0, #0x50
+	bl	_bsaes_decrypt8
+	vldmia	r9, {q14}			@ reload IV
+	vld1.8	{q8-q9}, [r0]!	@ reload input
+	veor	q0, q0, q14	@ ^= IV
+	vld1.8	{q10-q11}, [r0]!
+	veor	q1, q1, q8
+	veor	q6, q6, q9
+	vld1.8	{q15}, [r0]!
+	veor	q4, q4, q10
+	vst1.8	{q0-q1}, [r1]!	@ write output
+	veor	q2, q2, q11
+	vst1.8	{q6}, [r1]!
+	vst1.8	{q4}, [r1]!
+	vst1.8	{q2}, [r1]!
+	b	.Lcbc_dec_done
+.align	4
+.Lcbc_dec_four:
+	sub	r0, r0, #0x40
+	bl	_bsaes_decrypt8
+	vldmia	r9, {q14}			@ reload IV
+	vld1.8	{q8-q9}, [r0]!	@ reload input
+	veor	q0, q0, q14	@ ^= IV
+	vld1.8	{q10}, [r0]!
+	veor	q1, q1, q8
+	veor	q6, q6, q9
+	vld1.8	{q15}, [r0]!
+	veor	q4, q4, q10
+	vst1.8	{q0-q1}, [r1]!	@ write output
+	vst1.8	{q6}, [r1]!
+	vst1.8	{q4}, [r1]!
+	b	.Lcbc_dec_done
+.align	4
+.Lcbc_dec_three:
+	sub	r0, r0, #0x30
+	bl	_bsaes_decrypt8
+	vldmia	r9, {q14}			@ reload IV
+	vld1.8	{q8-q9}, [r0]!	@ reload input
+	veor	q0, q0, q14	@ ^= IV
+	vld1.8	{q15}, [r0]!
+	veor	q1, q1, q8
+	veor	q6, q6, q9
+	vst1.8	{q0-q1}, [r1]!	@ write output
+	vst1.8	{q6}, [r1]!
+	b	.Lcbc_dec_done
+.align	4
+.Lcbc_dec_two:
+	sub	r0, r0, #0x20
+	bl	_bsaes_decrypt8
+	vldmia	r9, {q14}			@ reload IV
+	vld1.8	{q8}, [r0]!		@ reload input
+	veor	q0, q0, q14	@ ^= IV
+	vld1.8	{q15}, [r0]!		@ reload input
+	veor	q1, q1, q8
+	vst1.8	{q0-q1}, [r1]!	@ write output
+	b	.Lcbc_dec_done
+.align	4
+.Lcbc_dec_one:
+	sub	r0, r0, #0x10
+	mov	r10, r1			@ save original out pointer
+	mov	r1, r9			@ use the iv scratch space as out buffer
+	mov	r2, r3
+	vmov	q4,q15		@ just in case ensure that IV
+	vmov	q5,q0			@ and input are preserved
+	bl	AES_decrypt
+	vld1.8	{q0}, [r9,:64]		@ load result
+	veor	q0, q0, q4	@ ^= IV
+	vmov	q15, q5		@ q5 holds input
+	vst1.8	{q0}, [r10]		@ write output
+
+.Lcbc_dec_done:
+#ifndef	BSAES_ASM_EXTENDED_KEY
+	vmov.i32	q0, #0
+	vmov.i32	q1, #0
+.Lcbc_dec_bzero:				@ wipe key schedule [if any]
+	vstmia		sp!, {q0-q1}
+	cmp		sp, r9
+	bne		.Lcbc_dec_bzero
+#endif
+
+	mov	sp, r9
+	add	sp, #0x10			@ add sp,r9,#0x10 is no good for thumb
+	vst1.8	{q15}, [r8]		@ return IV
+	VFP_ABI_POP
+	ldmia	sp!, {r4-r10, pc}
+.size	bsaes_cbc_encrypt,.-bsaes_cbc_encrypt
+.extern	AES_encrypt
+.global	bsaes_ctr32_encrypt_blocks
+.type	bsaes_ctr32_encrypt_blocks,%function
+.align	5
+bsaes_ctr32_encrypt_blocks:
+	cmp	r2, #8			@ use plain AES for
+	blo	.Lctr_enc_short			@ small sizes
+
+	mov	ip, sp
+	stmdb	sp!, {r4-r10, lr}
+	VFP_ABI_PUSH
+	ldr	r8, [ip]			@ ctr is 1st arg on the stack
+	sub	sp, sp, #0x10			@ scratch space to carry over the ctr
+	mov	r9, sp				@ save sp
+
+	ldr	r10, [r3, #240]		@ get # of rounds
+#ifndef	BSAES_ASM_EXTENDED_KEY
+	@ allocate the key schedule on the stack
+	sub	r12, sp, r10, lsl#7		@ 128 bytes per inner round key
+	add	r12, #96			@ size of bit-sliced key schedule
+
+	@ populate the key schedule
+	mov	r4, r3			@ pass key
+	mov	r5, r10			@ pass # of rounds
+	mov	sp, r12				@ sp is sp
+	bl	_bsaes_key_convert
+	veor	q7,q7,q15	@ fix up last round key
+	vstmia	r12, {q7}			@ save last round key
+
+	vld1.8	{q0}, [r8]		@ load counter
+	add	r8, r6, #.LREVM0SR-.LM0	@ borrow r8
+	vldmia	sp, {q4}		@ load round0 key
+#else
+	ldr	r12, [r3, #244]
+	eors	r12, #1
+	beq	0f
+
+	@ populate the key schedule
+	str	r12, [r3, #244]
+	mov	r4, r3			@ pass key
+	mov	r5, r10			@ pass # of rounds
+	add	r12, r3, #248			@ pass key schedule
+	bl	_bsaes_key_convert
+	veor	q7,q7,q15	@ fix up last round key
+	vstmia	r12, {q7}			@ save last round key
+
+.align	2
+0:	add	r12, r3, #248
+	vld1.8	{q0}, [r8]		@ load counter
+	adrl	r8, .LREVM0SR			@ borrow r8
+	vldmia	r12, {q4}			@ load round0 key
+	sub	sp, #0x10			@ place for adjusted round0 key
+#endif
+
+	vmov.i32	q8,#1		@ compose 1<<96
+	veor		q9,q9,q9
+	vrev32.8	q0,q0
+	vext.8		q8,q9,q8,#4
+	vrev32.8	q4,q4
+	vadd.u32	q9,q8,q8	@ compose 2<<96
+	vstmia	sp, {q4}		@ save adjusted round0 key
+	b	.Lctr_enc_loop
+
+.align	4
+.Lctr_enc_loop:
+	vadd.u32	q10, q8, q9	@ compose 3<<96
+	vadd.u32	q1, q0, q8	@ +1
+	vadd.u32	q2, q0, q9	@ +2
+	vadd.u32	q3, q0, q10	@ +3
+	vadd.u32	q4, q1, q10
+	vadd.u32	q5, q2, q10
+	vadd.u32	q6, q3, q10
+	vadd.u32	q7, q4, q10
+	vadd.u32	q10, q5, q10	@ next counter
+
+	@ Borrow prologue from _bsaes_encrypt8 to use the opportunity
+	@ to flip byte order in 32-bit counter
+
+	vldmia		sp, {q9}		@ load round0 key
+#ifndef	BSAES_ASM_EXTENDED_KEY
+	add		r4, sp, #0x10		@ pass next round key
+#else
+	add		r4, r3, #264
+#endif
+	vldmia		r8, {q8}			@ .LREVM0SR
+	mov		r5, r10			@ pass rounds
+	vstmia		r9, {q10}			@ save next counter
+	sub		r6, r8, #.LREVM0SR-.LSR	@ pass constants
+
+	bl		_bsaes_encrypt8_alt
+
+	subs		r2, r2, #8
+	blo		.Lctr_enc_loop_done
+
+	vld1.8		{q8-q9}, [r0]!	@ load input
+	vld1.8		{q10-q11}, [r0]!
+	veor		q0, q8
+	veor		q1, q9
+	vld1.8		{q12-q13}, [r0]!
+	veor		q4, q10
+	veor		q6, q11
+	vld1.8		{q14-q15}, [r0]!
+	veor		q3, q12
+	vst1.8		{q0-q1}, [r1]!	@ write output
+	veor		q7, q13
+	veor		q2, q14
+	vst1.8		{q4}, [r1]!
+	veor		q5, q15
+	vst1.8		{q6}, [r1]!
+	vmov.i32	q8, #1			@ compose 1<<96
+	vst1.8		{q3}, [r1]!
+	veor		q9, q9, q9
+	vst1.8		{q7}, [r1]!
+	vext.8		q8, q9, q8, #4
+	vst1.8		{q2}, [r1]!
+	vadd.u32	q9,q8,q8		@ compose 2<<96
+	vst1.8		{q5}, [r1]!
+	vldmia		r9, {q0}			@ load counter
+
+	bne		.Lctr_enc_loop
+	b		.Lctr_enc_done
+
+.align	4
+.Lctr_enc_loop_done:
+	add		r2, r2, #8
+	vld1.8		{q8}, [r0]!	@ load input
+	veor		q0, q8
+	vst1.8		{q0}, [r1]!	@ write output
+	cmp		r2, #2
+	blo		.Lctr_enc_done
+	vld1.8		{q9}, [r0]!
+	veor		q1, q9
+	vst1.8		{q1}, [r1]!
+	beq		.Lctr_enc_done
+	vld1.8		{q10}, [r0]!
+	veor		q4, q10
+	vst1.8		{q4}, [r1]!
+	cmp		r2, #4
+	blo		.Lctr_enc_done
+	vld1.8		{q11}, [r0]!
+	veor		q6, q11
+	vst1.8		{q6}, [r1]!
+	beq		.Lctr_enc_done
+	vld1.8		{q12}, [r0]!
+	veor		q3, q12
+	vst1.8		{q3}, [r1]!
+	cmp		r2, #6
+	blo		.Lctr_enc_done
+	vld1.8		{q13}, [r0]!
+	veor		q7, q13
+	vst1.8		{q7}, [r1]!
+	beq		.Lctr_enc_done
+	vld1.8		{q14}, [r0]
+	veor		q2, q14
+	vst1.8		{q2}, [r1]!
+
+.Lctr_enc_done:
+	vmov.i32	q0, #0
+	vmov.i32	q1, #0
+#ifndef	BSAES_ASM_EXTENDED_KEY
+.Lctr_enc_bzero:			@ wipe key schedule [if any]
+	vstmia		sp!, {q0-q1}
+	cmp		sp, r9
+	bne		.Lctr_enc_bzero
+#else
+	vstmia		sp, {q0-q1}
+#endif
+
+	mov	sp, r9
+	add	sp, #0x10		@ add sp,r9,#0x10 is no good for thumb
+	VFP_ABI_POP
+	ldmia	sp!, {r4-r10, pc}	@ return
+
+.align	4
+.Lctr_enc_short:
+	ldr	ip, [sp]		@ ctr pointer is passed on stack
+	stmdb	sp!, {r4-r8, lr}
+
+	mov	r4, r0		@ copy arguments
+	mov	r5, r1
+	mov	r6, r2
+	mov	r7, r3
+	ldr	r8, [ip, #12]		@ load counter LSW
+	vld1.8	{q1}, [ip]		@ load whole counter value
+#ifdef __ARMEL__
+	rev	r8, r8
+#endif
+	sub	sp, sp, #0x10
+	vst1.8	{q1}, [sp,:64]	@ copy counter value
+	sub	sp, sp, #0x10
+
+.Lctr_enc_short_loop:
+	add	r0, sp, #0x10		@ input counter value
+	mov	r1, sp			@ output on the stack
+	mov	r2, r7			@ key
+
+	bl	AES_encrypt
+
+	vld1.8	{q0}, [r4]!	@ load input
+	vld1.8	{q1}, [sp,:64]	@ load encrypted counter
+	add	r8, r8, #1
+#ifdef __ARMEL__
+	rev	r0, r8
+	str	r0, [sp, #0x1c]		@ next counter value
+#else
+	str	r8, [sp, #0x1c]		@ next counter value
+#endif
+	veor	q0,q0,q1
+	vst1.8	{q0}, [r5]!	@ store output
+	subs	r6, r6, #1
+	bne	.Lctr_enc_short_loop
+
+	vmov.i32	q0, #0
+	vmov.i32	q1, #0
+	vstmia		sp!, {q0-q1}
+
+	ldmia	sp!, {r4-r8, pc}
+.size	bsaes_ctr32_encrypt_blocks,.-bsaes_ctr32_encrypt_blocks
+.globl	bsaes_xts_encrypt
+.type	bsaes_xts_encrypt,%function
+.align	4
+bsaes_xts_encrypt:
+	mov	ip, sp
+	stmdb	sp!, {r4-r10, lr}		@ 0x20
+	VFP_ABI_PUSH
+	mov	r6, sp				@ future r3
+
+	mov	r7, r0
+	mov	r8, r1
+	mov	r9, r2
+	mov	r10, r3
+
+	sub	r0, sp, #0x10			@ 0x10
+	bic	r0, #0xf			@ align at 16 bytes
+	mov	sp, r0
+
+#ifdef	XTS_CHAIN_TWEAK
+	ldr	r0, [ip]			@ pointer to input tweak
+#else
+	@ generate initial tweak
+	ldr	r0, [ip, #4]			@ iv[]
+	mov	r1, sp
+	ldr	r2, [ip, #0]			@ key2
+	bl	AES_encrypt
+	mov	r0,sp				@ pointer to initial tweak
+#endif
+
+	ldr	r1, [r10, #240]		@ get # of rounds
+	mov	r3, r6
+#ifndef	BSAES_ASM_EXTENDED_KEY
+	@ allocate the key schedule on the stack
+	sub	r12, sp, r1, lsl#7		@ 128 bytes per inner round key
+	@ add	r12, #96			@ size of bit-sliced key schedule
+	sub	r12, #48			@ place for tweak[9]
+
+	@ populate the key schedule
+	mov	r4, r10			@ pass key
+	mov	r5, r1			@ pass # of rounds
+	mov	sp, r12
+	add	r12, #0x90			@ pass key schedule
+	bl	_bsaes_key_convert
+	veor	q7, q7, q15	@ fix up last round key
+	vstmia	r12, {q7}			@ save last round key
+#else
+	ldr	r12, [r10, #244]
+	eors	r12, #1
+	beq	0f
+
+	str	r12, [r10, #244]
+	mov	r4, r10			@ pass key
+	mov	r5, r1			@ pass # of rounds
+	add	r12, r10, #248			@ pass key schedule
+	bl	_bsaes_key_convert
+	veor	q7, q7, q15	@ fix up last round key
+	vstmia	r12, {q7}
+
+.align	2
+0:	sub	sp, #0x90			@ place for tweak[9]
+#endif
+
+	vld1.8	{q8}, [r0]			@ initial tweak
+	adr	r2, .Lxts_magic
+
+	subs	r9, #0x80
+	blo	.Lxts_enc_short
+	b	.Lxts_enc_loop
+
+.align	4
+.Lxts_enc_loop:
+	vldmia		r2, {q5}	@ load XTS magic
+	vshr.s64	q6, q8, #63
+	mov		r0, sp
+	vand		q6, q6, q5
+	vadd.u64	q9, q8, q8
+	vst1.64		{q8}, [r0,:128]!
+	vswp		d13,d12
+	vshr.s64	q7, q9, #63
+	veor		q9, q9, q6
+	vand		q7, q7, q5
+	vadd.u64	q10, q9, q9
+	vst1.64		{q9}, [r0,:128]!
+	vswp		d15,d14
+	vshr.s64	q6, q10, #63
+	veor		q10, q10, q7
+	vand		q6, q6, q5
+	vld1.8		{q0}, [r7]!
+	vadd.u64	q11, q10, q10
+	vst1.64		{q10}, [r0,:128]!
+	vswp		d13,d12
+	vshr.s64	q7, q11, #63
+	veor		q11, q11, q6
+	vand		q7, q7, q5
+	vld1.8		{q1}, [r7]!
+	veor		q0, q0, q8
+	vadd.u64	q12, q11, q11
+	vst1.64		{q11}, [r0,:128]!
+	vswp		d15,d14
+	vshr.s64	q6, q12, #63
+	veor		q12, q12, q7
+	vand		q6, q6, q5
+	vld1.8		{q2}, [r7]!
+	veor		q1, q1, q9
+	vadd.u64	q13, q12, q12
+	vst1.64		{q12}, [r0,:128]!
+	vswp		d13,d12
+	vshr.s64	q7, q13, #63
+	veor		q13, q13, q6
+	vand		q7, q7, q5
+	vld1.8		{q3}, [r7]!
+	veor		q2, q2, q10
+	vadd.u64	q14, q13, q13
+	vst1.64		{q13}, [r0,:128]!
+	vswp		d15,d14
+	vshr.s64	q6, q14, #63
+	veor		q14, q14, q7
+	vand		q6, q6, q5
+	vld1.8		{q4}, [r7]!
+	veor		q3, q3, q11
+	vadd.u64	q15, q14, q14
+	vst1.64		{q14}, [r0,:128]!
+	vswp		d13,d12
+	vshr.s64	q7, q15, #63
+	veor		q15, q15, q6
+	vand		q7, q7, q5
+	vld1.8		{q5}, [r7]!
+	veor		q4, q4, q12
+	vadd.u64	q8, q15, q15
+	vst1.64		{q15}, [r0,:128]!
+	vswp		d15,d14
+	veor		q8, q8, q7
+	vst1.64		{q8}, [r0,:128]		@ next round tweak
+
+	vld1.8		{q6-q7}, [r7]!
+	veor		q5, q5, q13
+#ifndef	BSAES_ASM_EXTENDED_KEY
+	add		r4, sp, #0x90			@ pass key schedule
+#else
+	add		r4, r10, #248			@ pass key schedule
+#endif
+	veor		q6, q6, q14
+	mov		r5, r1			@ pass rounds
+	veor		q7, q7, q15
+	mov		r0, sp
+
+	bl		_bsaes_encrypt8
+
+	vld1.64		{q8-q9}, [r0,:128]!
+	vld1.64		{q10-q11}, [r0,:128]!
+	veor		q0, q0, q8
+	vld1.64		{q12-q13}, [r0,:128]!
+	veor		q1, q1, q9
+	veor		q8, q4, q10
+	vst1.8		{q0-q1}, [r8]!
+	veor		q9, q6, q11
+	vld1.64		{q14-q15}, [r0,:128]!
+	veor		q10, q3, q12
+	vst1.8		{q8-q9}, [r8]!
+	veor		q11, q7, q13
+	veor		q12, q2, q14
+	vst1.8		{q10-q11}, [r8]!
+	veor		q13, q5, q15
+	vst1.8		{q12-q13}, [r8]!
+
+	vld1.64		{q8}, [r0,:128]		@ next round tweak
+
+	subs		r9, #0x80
+	bpl		.Lxts_enc_loop
+
+.Lxts_enc_short:
+	adds		r9, #0x70
+	bmi		.Lxts_enc_done
+
+	vldmia		r2, {q5}	@ load XTS magic
+	vshr.s64	q7, q8, #63
+	mov		r0, sp
+	vand		q7, q7, q5
+	vadd.u64	q9, q8, q8
+	vst1.64		{q8}, [r0,:128]!
+	vswp		d15,d14
+	vshr.s64	q6, q9, #63
+	veor		q9, q9, q7
+	vand		q6, q6, q5
+	vadd.u64	q10, q9, q9
+	vst1.64		{q9}, [r0,:128]!
+	vswp		d13,d12
+	vshr.s64	q7, q10, #63
+	veor		q10, q10, q6
+	vand		q7, q7, q5
+	vld1.8		{q0}, [r7]!
+	subs		r9, #0x10
+	bmi		.Lxts_enc_1
+	vadd.u64	q11, q10, q10
+	vst1.64		{q10}, [r0,:128]!
+	vswp		d15,d14
+	vshr.s64	q6, q11, #63
+	veor		q11, q11, q7
+	vand		q6, q6, q5
+	vld1.8		{q1}, [r7]!
+	subs		r9, #0x10
+	bmi		.Lxts_enc_2
+	veor		q0, q0, q8
+	vadd.u64	q12, q11, q11
+	vst1.64		{q11}, [r0,:128]!
+	vswp		d13,d12
+	vshr.s64	q7, q12, #63
+	veor		q12, q12, q6
+	vand		q7, q7, q5
+	vld1.8		{q2}, [r7]!
+	subs		r9, #0x10
+	bmi		.Lxts_enc_3
+	veor		q1, q1, q9
+	vadd.u64	q13, q12, q12
+	vst1.64		{q12}, [r0,:128]!
+	vswp		d15,d14
+	vshr.s64	q6, q13, #63
+	veor		q13, q13, q7
+	vand		q6, q6, q5
+	vld1.8		{q3}, [r7]!
+	subs		r9, #0x10
+	bmi		.Lxts_enc_4
+	veor		q2, q2, q10
+	vadd.u64	q14, q13, q13
+	vst1.64		{q13}, [r0,:128]!
+	vswp		d13,d12
+	vshr.s64	q7, q14, #63
+	veor		q14, q14, q6
+	vand		q7, q7, q5
+	vld1.8		{q4}, [r7]!
+	subs		r9, #0x10
+	bmi		.Lxts_enc_5
+	veor		q3, q3, q11
+	vadd.u64	q15, q14, q14
+	vst1.64		{q14}, [r0,:128]!
+	vswp		d15,d14
+	vshr.s64	q6, q15, #63
+	veor		q15, q15, q7
+	vand		q6, q6, q5
+	vld1.8		{q5}, [r7]!
+	subs		r9, #0x10
+	bmi		.Lxts_enc_6
+	veor		q4, q4, q12
+	sub		r9, #0x10
+	vst1.64		{q15}, [r0,:128]		@ next round tweak
+
+	vld1.8		{q6}, [r7]!
+	veor		q5, q5, q13
+#ifndef	BSAES_ASM_EXTENDED_KEY
+	add		r4, sp, #0x90			@ pass key schedule
+#else
+	add		r4, r10, #248			@ pass key schedule
+#endif
+	veor		q6, q6, q14
+	mov		r5, r1			@ pass rounds
+	mov		r0, sp
+
+	bl		_bsaes_encrypt8
+
+	vld1.64		{q8-q9}, [r0,:128]!
+	vld1.64		{q10-q11}, [r0,:128]!
+	veor		q0, q0, q8
+	vld1.64		{q12-q13}, [r0,:128]!
+	veor		q1, q1, q9
+	veor		q8, q4, q10
+	vst1.8		{q0-q1}, [r8]!
+	veor		q9, q6, q11
+	vld1.64		{q14}, [r0,:128]!
+	veor		q10, q3, q12
+	vst1.8		{q8-q9}, [r8]!
+	veor		q11, q7, q13
+	veor		q12, q2, q14
+	vst1.8		{q10-q11}, [r8]!
+	vst1.8		{q12}, [r8]!
+
+	vld1.64		{q8}, [r0,:128]		@ next round tweak
+	b		.Lxts_enc_done
+.align	4
+.Lxts_enc_6:
+	vst1.64		{q14}, [r0,:128]		@ next round tweak
+
+	veor		q4, q4, q12
+#ifndef	BSAES_ASM_EXTENDED_KEY
+	add		r4, sp, #0x90			@ pass key schedule
+#else
+	add		r4, r10, #248			@ pass key schedule
+#endif
+	veor		q5, q5, q13
+	mov		r5, r1			@ pass rounds
+	mov		r0, sp
+
+	bl		_bsaes_encrypt8
+
+	vld1.64		{q8-q9}, [r0,:128]!
+	vld1.64		{q10-q11}, [r0,:128]!
+	veor		q0, q0, q8
+	vld1.64		{q12-q13}, [r0,:128]!
+	veor		q1, q1, q9
+	veor		q8, q4, q10
+	vst1.8		{q0-q1}, [r8]!
+	veor		q9, q6, q11
+	veor		q10, q3, q12
+	vst1.8		{q8-q9}, [r8]!
+	veor		q11, q7, q13
+	vst1.8		{q10-q11}, [r8]!
+
+	vld1.64		{q8}, [r0,:128]		@ next round tweak
+	b		.Lxts_enc_done
+
+@ put this in range for both ARM and Thumb mode adr instructions
+.align	5
+.Lxts_magic:
+	.quad	1, 0x87
+
+.align	5
+.Lxts_enc_5:
+	vst1.64		{q13}, [r0,:128]		@ next round tweak
+
+	veor		q3, q3, q11
+#ifndef	BSAES_ASM_EXTENDED_KEY
+	add		r4, sp, #0x90			@ pass key schedule
+#else
+	add		r4, r10, #248			@ pass key schedule
+#endif
+	veor		q4, q4, q12
+	mov		r5, r1			@ pass rounds
+	mov		r0, sp
+
+	bl		_bsaes_encrypt8
+
+	vld1.64		{q8-q9}, [r0,:128]!
+	vld1.64		{q10-q11}, [r0,:128]!
+	veor		q0, q0, q8
+	vld1.64		{q12}, [r0,:128]!
+	veor		q1, q1, q9
+	veor		q8, q4, q10
+	vst1.8		{q0-q1}, [r8]!
+	veor		q9, q6, q11
+	veor		q10, q3, q12
+	vst1.8		{q8-q9}, [r8]!
+	vst1.8		{q10}, [r8]!
+
+	vld1.64		{q8}, [r0,:128]		@ next round tweak
+	b		.Lxts_enc_done
+.align	4
+.Lxts_enc_4:
+	vst1.64		{q12}, [r0,:128]		@ next round tweak
+
+	veor		q2, q2, q10
+#ifndef	BSAES_ASM_EXTENDED_KEY
+	add		r4, sp, #0x90			@ pass key schedule
+#else
+	add		r4, r10, #248			@ pass key schedule
+#endif
+	veor		q3, q3, q11
+	mov		r5, r1			@ pass rounds
+	mov		r0, sp
+
+	bl		_bsaes_encrypt8
+
+	vld1.64		{q8-q9}, [r0,:128]!
+	vld1.64		{q10-q11}, [r0,:128]!
+	veor		q0, q0, q8
+	veor		q1, q1, q9
+	veor		q8, q4, q10
+	vst1.8		{q0-q1}, [r8]!
+	veor		q9, q6, q11
+	vst1.8		{q8-q9}, [r8]!
+
+	vld1.64		{q8}, [r0,:128]		@ next round tweak
+	b		.Lxts_enc_done
+.align	4
+.Lxts_enc_3:
+	vst1.64		{q11}, [r0,:128]		@ next round tweak
+
+	veor		q1, q1, q9
+#ifndef	BSAES_ASM_EXTENDED_KEY
+	add		r4, sp, #0x90			@ pass key schedule
+#else
+	add		r4, r10, #248			@ pass key schedule
+#endif
+	veor		q2, q2, q10
+	mov		r5, r1			@ pass rounds
+	mov		r0, sp
+
+	bl		_bsaes_encrypt8
+
+	vld1.64		{q8-q9}, [r0,:128]!
+	vld1.64		{q10}, [r0,:128]!
+	veor		q0, q0, q8
+	veor		q1, q1, q9
+	veor		q8, q4, q10
+	vst1.8		{q0-q1}, [r8]!
+	vst1.8		{q8}, [r8]!
+
+	vld1.64		{q8}, [r0,:128]		@ next round tweak
+	b		.Lxts_enc_done
+.align	4
+.Lxts_enc_2:
+	vst1.64		{q10}, [r0,:128]		@ next round tweak
+
+	veor		q0, q0, q8
+#ifndef	BSAES_ASM_EXTENDED_KEY
+	add		r4, sp, #0x90			@ pass key schedule
+#else
+	add		r4, r10, #248			@ pass key schedule
+#endif
+	veor		q1, q1, q9
+	mov		r5, r1			@ pass rounds
+	mov		r0, sp
+
+	bl		_bsaes_encrypt8
+
+	vld1.64		{q8-q9}, [r0,:128]!
+	veor		q0, q0, q8
+	veor		q1, q1, q9
+	vst1.8		{q0-q1}, [r8]!
+
+	vld1.64		{q8}, [r0,:128]		@ next round tweak
+	b		.Lxts_enc_done
+.align	4
+.Lxts_enc_1:
+	mov		r0, sp
+	veor		q0, q8
+	mov		r1, sp
+	vst1.8		{q0}, [sp,:128]
+	mov		r2, r10
+	mov		r4, r3				@ preserve fp
+
+	bl		AES_encrypt
+
+	vld1.8		{q0}, [sp,:128]
+	veor		q0, q0, q8
+	vst1.8		{q0}, [r8]!
+	mov		r3, r4
+
+	vmov		q8, q9		@ next round tweak
+
+.Lxts_enc_done:
+#ifndef	XTS_CHAIN_TWEAK
+	adds		r9, #0x10
+	beq		.Lxts_enc_ret
+	sub		r6, r8, #0x10
+
+.Lxts_enc_steal:
+	ldrb		r0, [r7], #1
+	ldrb		r1, [r8, #-0x10]
+	strb		r0, [r8, #-0x10]
+	strb		r1, [r8], #1
+
+	subs		r9, #1
+	bhi		.Lxts_enc_steal
+
+	vld1.8		{q0}, [r6]
+	mov		r0, sp
+	veor		q0, q0, q8
+	mov		r1, sp
+	vst1.8		{q0}, [sp,:128]
+	mov		r2, r10
+	mov		r4, r3			@ preserve fp
+
+	bl		AES_encrypt
+
+	vld1.8		{q0}, [sp,:128]
+	veor		q0, q0, q8
+	vst1.8		{q0}, [r6]
+	mov		r3, r4
+#endif
+
+.Lxts_enc_ret:
+	bic		r0, r3, #0xf
+	vmov.i32	q0, #0
+	vmov.i32	q1, #0
+#ifdef	XTS_CHAIN_TWEAK
+	ldr		r1, [r3, #0x20+VFP_ABI_FRAME]	@ chain tweak
+#endif
+.Lxts_enc_bzero:				@ wipe key schedule [if any]
+	vstmia		sp!, {q0-q1}
+	cmp		sp, r0
+	bne		.Lxts_enc_bzero
+
+	mov		sp, r3
+#ifdef	XTS_CHAIN_TWEAK
+	vst1.8		{q8}, [r1]
+#endif
+	VFP_ABI_POP
+	ldmia		sp!, {r4-r10, pc}	@ return
+
+.size	bsaes_xts_encrypt,.-bsaes_xts_encrypt
+
+.globl	bsaes_xts_decrypt
+.type	bsaes_xts_decrypt,%function
+.align	4
+bsaes_xts_decrypt:
+	mov	ip, sp
+	stmdb	sp!, {r4-r10, lr}		@ 0x20
+	VFP_ABI_PUSH
+	mov	r6, sp				@ future r3
+
+	mov	r7, r0
+	mov	r8, r1
+	mov	r9, r2
+	mov	r10, r3
+
+	sub	r0, sp, #0x10			@ 0x10
+	bic	r0, #0xf			@ align at 16 bytes
+	mov	sp, r0
+
+#ifdef	XTS_CHAIN_TWEAK
+	ldr	r0, [ip]			@ pointer to input tweak
+#else
+	@ generate initial tweak
+	ldr	r0, [ip, #4]			@ iv[]
+	mov	r1, sp
+	ldr	r2, [ip, #0]			@ key2
+	bl	AES_encrypt
+	mov	r0, sp				@ pointer to initial tweak
+#endif
+
+	ldr	r1, [r10, #240]		@ get # of rounds
+	mov	r3, r6
+#ifndef	BSAES_ASM_EXTENDED_KEY
+	@ allocate the key schedule on the stack
+	sub	r12, sp, r1, lsl#7		@ 128 bytes per inner round key
+	@ add	r12, #96			@ size of bit-sliced key schedule
+	sub	r12, #48			@ place for tweak[9]
+
+	@ populate the key schedule
+	mov	r4, r10			@ pass key
+	mov	r5, r1			@ pass # of rounds
+	mov	sp, r12
+	add	r12, #0x90			@ pass key schedule
+	bl	_bsaes_key_convert
+	add	r4, sp, #0x90
+	vldmia	r4, {q6}
+	vstmia	r12,  {q15}		@ save last round key
+	veor	q7, q7, q6	@ fix up round 0 key
+	vstmia	r4, {q7}
+#else
+	ldr	r12, [r10, #244]
+	eors	r12, #1
+	beq	0f
+
+	str	r12, [r10, #244]
+	mov	r4, r10			@ pass key
+	mov	r5, r1			@ pass # of rounds
+	add	r12, r10, #248			@ pass key schedule
+	bl	_bsaes_key_convert
+	add	r4, r10, #248
+	vldmia	r4, {q6}
+	vstmia	r12,  {q15}		@ save last round key
+	veor	q7, q7, q6	@ fix up round 0 key
+	vstmia	r4, {q7}
+
+.align	2
+0:	sub	sp, #0x90			@ place for tweak[9]
+#endif
+	vld1.8	{q8}, [r0]			@ initial tweak
+	adr	r2, .Lxts_magic
+
+	tst	r9, #0xf			@ if not multiple of 16
+	it	ne				@ Thumb2 thing, sanity check in ARM
+	subne	r9, #0x10			@ subtract another 16 bytes
+	subs	r9, #0x80
+
+	blo	.Lxts_dec_short
+	b	.Lxts_dec_loop
+
+.align	4
+.Lxts_dec_loop:
+	vldmia		r2, {q5}	@ load XTS magic
+	vshr.s64	q6, q8, #63
+	mov		r0, sp
+	vand		q6, q6, q5
+	vadd.u64	q9, q8, q8
+	vst1.64		{q8}, [r0,:128]!
+	vswp		d13,d12
+	vshr.s64	q7, q9, #63
+	veor		q9, q9, q6
+	vand		q7, q7, q5
+	vadd.u64	q10, q9, q9
+	vst1.64		{q9}, [r0,:128]!
+	vswp		d15,d14
+	vshr.s64	q6, q10, #63
+	veor		q10, q10, q7
+	vand		q6, q6, q5
+	vld1.8		{q0}, [r7]!
+	vadd.u64	q11, q10, q10
+	vst1.64		{q10}, [r0,:128]!
+	vswp		d13,d12
+	vshr.s64	q7, q11, #63
+	veor		q11, q11, q6
+	vand		q7, q7, q5
+	vld1.8		{q1}, [r7]!
+	veor		q0, q0, q8
+	vadd.u64	q12, q11, q11
+	vst1.64		{q11}, [r0,:128]!
+	vswp		d15,d14
+	vshr.s64	q6, q12, #63
+	veor		q12, q12, q7
+	vand		q6, q6, q5
+	vld1.8		{q2}, [r7]!
+	veor		q1, q1, q9
+	vadd.u64	q13, q12, q12
+	vst1.64		{q12}, [r0,:128]!
+	vswp		d13,d12
+	vshr.s64	q7, q13, #63
+	veor		q13, q13, q6
+	vand		q7, q7, q5
+	vld1.8		{q3}, [r7]!
+	veor		q2, q2, q10
+	vadd.u64	q14, q13, q13
+	vst1.64		{q13}, [r0,:128]!
+	vswp		d15,d14
+	vshr.s64	q6, q14, #63
+	veor		q14, q14, q7
+	vand		q6, q6, q5
+	vld1.8		{q4}, [r7]!
+	veor		q3, q3, q11
+	vadd.u64	q15, q14, q14
+	vst1.64		{q14}, [r0,:128]!
+	vswp		d13,d12
+	vshr.s64	q7, q15, #63
+	veor		q15, q15, q6
+	vand		q7, q7, q5
+	vld1.8		{q5}, [r7]!
+	veor		q4, q4, q12
+	vadd.u64	q8, q15, q15
+	vst1.64		{q15}, [r0,:128]!
+	vswp		d15,d14
+	veor		q8, q8, q7
+	vst1.64		{q8}, [r0,:128]		@ next round tweak
+
+	vld1.8		{q6-q7}, [r7]!
+	veor		q5, q5, q13
+#ifndef	BSAES_ASM_EXTENDED_KEY
+	add		r4, sp, #0x90			@ pass key schedule
+#else
+	add		r4, r10, #248			@ pass key schedule
+#endif
+	veor		q6, q6, q14
+	mov		r5, r1			@ pass rounds
+	veor		q7, q7, q15
+	mov		r0, sp
+
+	bl		_bsaes_decrypt8
+
+	vld1.64		{q8-q9}, [r0,:128]!
+	vld1.64		{q10-q11}, [r0,:128]!
+	veor		q0, q0, q8
+	vld1.64		{q12-q13}, [r0,:128]!
+	veor		q1, q1, q9
+	veor		q8, q6, q10
+	vst1.8		{q0-q1}, [r8]!
+	veor		q9, q4, q11
+	vld1.64		{q14-q15}, [r0,:128]!
+	veor		q10, q2, q12
+	vst1.8		{q8-q9}, [r8]!
+	veor		q11, q7, q13
+	veor		q12, q3, q14
+	vst1.8		{q10-q11}, [r8]!
+	veor		q13, q5, q15
+	vst1.8		{q12-q13}, [r8]!
+
+	vld1.64		{q8}, [r0,:128]		@ next round tweak
+
+	subs		r9, #0x80
+	bpl		.Lxts_dec_loop
+
+.Lxts_dec_short:
+	adds		r9, #0x70
+	bmi		.Lxts_dec_done
+
+	vldmia		r2, {q5}	@ load XTS magic
+	vshr.s64	q7, q8, #63
+	mov		r0, sp
+	vand		q7, q7, q5
+	vadd.u64	q9, q8, q8
+	vst1.64		{q8}, [r0,:128]!
+	vswp		d15,d14
+	vshr.s64	q6, q9, #63
+	veor		q9, q9, q7
+	vand		q6, q6, q5
+	vadd.u64	q10, q9, q9
+	vst1.64		{q9}, [r0,:128]!
+	vswp		d13,d12
+	vshr.s64	q7, q10, #63
+	veor		q10, q10, q6
+	vand		q7, q7, q5
+	vld1.8		{q0}, [r7]!
+	subs		r9, #0x10
+	bmi		.Lxts_dec_1
+	vadd.u64	q11, q10, q10
+	vst1.64		{q10}, [r0,:128]!
+	vswp		d15,d14
+	vshr.s64	q6, q11, #63
+	veor		q11, q11, q7
+	vand		q6, q6, q5
+	vld1.8		{q1}, [r7]!
+	subs		r9, #0x10
+	bmi		.Lxts_dec_2
+	veor		q0, q0, q8
+	vadd.u64	q12, q11, q11
+	vst1.64		{q11}, [r0,:128]!
+	vswp		d13,d12
+	vshr.s64	q7, q12, #63
+	veor		q12, q12, q6
+	vand		q7, q7, q5
+	vld1.8		{q2}, [r7]!
+	subs		r9, #0x10
+	bmi		.Lxts_dec_3
+	veor		q1, q1, q9
+	vadd.u64	q13, q12, q12
+	vst1.64		{q12}, [r0,:128]!
+	vswp		d15,d14
+	vshr.s64	q6, q13, #63
+	veor		q13, q13, q7
+	vand		q6, q6, q5
+	vld1.8		{q3}, [r7]!
+	subs		r9, #0x10
+	bmi		.Lxts_dec_4
+	veor		q2, q2, q10
+	vadd.u64	q14, q13, q13
+	vst1.64		{q13}, [r0,:128]!
+	vswp		d13,d12
+	vshr.s64	q7, q14, #63
+	veor		q14, q14, q6
+	vand		q7, q7, q5
+	vld1.8		{q4}, [r7]!
+	subs		r9, #0x10
+	bmi		.Lxts_dec_5
+	veor		q3, q3, q11
+	vadd.u64	q15, q14, q14
+	vst1.64		{q14}, [r0,:128]!
+	vswp		d15,d14
+	vshr.s64	q6, q15, #63
+	veor		q15, q15, q7
+	vand		q6, q6, q5
+	vld1.8		{q5}, [r7]!
+	subs		r9, #0x10
+	bmi		.Lxts_dec_6
+	veor		q4, q4, q12
+	sub		r9, #0x10
+	vst1.64		{q15}, [r0,:128]		@ next round tweak
+
+	vld1.8		{q6}, [r7]!
+	veor		q5, q5, q13
+#ifndef	BSAES_ASM_EXTENDED_KEY
+	add		r4, sp, #0x90			@ pass key schedule
+#else
+	add		r4, r10, #248			@ pass key schedule
+#endif
+	veor		q6, q6, q14
+	mov		r5, r1			@ pass rounds
+	mov		r0, sp
+
+	bl		_bsaes_decrypt8
+
+	vld1.64		{q8-q9}, [r0,:128]!
+	vld1.64		{q10-q11}, [r0,:128]!
+	veor		q0, q0, q8
+	vld1.64		{q12-q13}, [r0,:128]!
+	veor		q1, q1, q9
+	veor		q8, q6, q10
+	vst1.8		{q0-q1}, [r8]!
+	veor		q9, q4, q11
+	vld1.64		{q14}, [r0,:128]!
+	veor		q10, q2, q12
+	vst1.8		{q8-q9}, [r8]!
+	veor		q11, q7, q13
+	veor		q12, q3, q14
+	vst1.8		{q10-q11}, [r8]!
+	vst1.8		{q12}, [r8]!
+
+	vld1.64		{q8}, [r0,:128]		@ next round tweak
+	b		.Lxts_dec_done
+.align	4
+.Lxts_dec_6:
+	vst1.64		{q14}, [r0,:128]		@ next round tweak
+
+	veor		q4, q4, q12
+#ifndef	BSAES_ASM_EXTENDED_KEY
+	add		r4, sp, #0x90			@ pass key schedule
+#else
+	add		r4, r10, #248			@ pass key schedule
+#endif
+	veor		q5, q5, q13
+	mov		r5, r1			@ pass rounds
+	mov		r0, sp
+
+	bl		_bsaes_decrypt8
+
+	vld1.64		{q8-q9}, [r0,:128]!
+	vld1.64		{q10-q11}, [r0,:128]!
+	veor		q0, q0, q8
+	vld1.64		{q12-q13}, [r0,:128]!
+	veor		q1, q1, q9
+	veor		q8, q6, q10
+	vst1.8		{q0-q1}, [r8]!
+	veor		q9, q4, q11
+	veor		q10, q2, q12
+	vst1.8		{q8-q9}, [r8]!
+	veor		q11, q7, q13
+	vst1.8		{q10-q11}, [r8]!
+
+	vld1.64		{q8}, [r0,:128]		@ next round tweak
+	b		.Lxts_dec_done
+.align	4
+.Lxts_dec_5:
+	vst1.64		{q13}, [r0,:128]		@ next round tweak
+
+	veor		q3, q3, q11
+#ifndef	BSAES_ASM_EXTENDED_KEY
+	add		r4, sp, #0x90			@ pass key schedule
+#else
+	add		r4, r10, #248			@ pass key schedule
+#endif
+	veor		q4, q4, q12
+	mov		r5, r1			@ pass rounds
+	mov		r0, sp
+
+	bl		_bsaes_decrypt8
+
+	vld1.64		{q8-q9}, [r0,:128]!
+	vld1.64		{q10-q11}, [r0,:128]!
+	veor		q0, q0, q8
+	vld1.64		{q12}, [r0,:128]!
+	veor		q1, q1, q9
+	veor		q8, q6, q10
+	vst1.8		{q0-q1}, [r8]!
+	veor		q9, q4, q11
+	veor		q10, q2, q12
+	vst1.8		{q8-q9}, [r8]!
+	vst1.8		{q10}, [r8]!
+
+	vld1.64		{q8}, [r0,:128]		@ next round tweak
+	b		.Lxts_dec_done
+.align	4
+.Lxts_dec_4:
+	vst1.64		{q12}, [r0,:128]		@ next round tweak
+
+	veor		q2, q2, q10
+#ifndef	BSAES_ASM_EXTENDED_KEY
+	add		r4, sp, #0x90			@ pass key schedule
+#else
+	add		r4, r10, #248			@ pass key schedule
+#endif
+	veor		q3, q3, q11
+	mov		r5, r1			@ pass rounds
+	mov		r0, sp
+
+	bl		_bsaes_decrypt8
+
+	vld1.64		{q8-q9}, [r0,:128]!
+	vld1.64		{q10-q11}, [r0,:128]!
+	veor		q0, q0, q8
+	veor		q1, q1, q9
+	veor		q8, q6, q10
+	vst1.8		{q0-q1}, [r8]!
+	veor		q9, q4, q11
+	vst1.8		{q8-q9}, [r8]!
+
+	vld1.64		{q8}, [r0,:128]		@ next round tweak
+	b		.Lxts_dec_done
+.align	4
+.Lxts_dec_3:
+	vst1.64		{q11}, [r0,:128]		@ next round tweak
+
+	veor		q1, q1, q9
+#ifndef	BSAES_ASM_EXTENDED_KEY
+	add		r4, sp, #0x90			@ pass key schedule
+#else
+	add		r4, r10, #248			@ pass key schedule
+#endif
+	veor		q2, q2, q10
+	mov		r5, r1			@ pass rounds
+	mov		r0, sp
+
+	bl		_bsaes_decrypt8
+
+	vld1.64		{q8-q9}, [r0,:128]!
+	vld1.64		{q10}, [r0,:128]!
+	veor		q0, q0, q8
+	veor		q1, q1, q9
+	veor		q8, q6, q10
+	vst1.8		{q0-q1}, [r8]!
+	vst1.8		{q8}, [r8]!
+
+	vld1.64		{q8}, [r0,:128]		@ next round tweak
+	b		.Lxts_dec_done
+.align	4
+.Lxts_dec_2:
+	vst1.64		{q10}, [r0,:128]		@ next round tweak
+
+	veor		q0, q0, q8
+#ifndef	BSAES_ASM_EXTENDED_KEY
+	add		r4, sp, #0x90			@ pass key schedule
+#else
+	add		r4, r10, #248			@ pass key schedule
+#endif
+	veor		q1, q1, q9
+	mov		r5, r1			@ pass rounds
+	mov		r0, sp
+
+	bl		_bsaes_decrypt8
+
+	vld1.64		{q8-q9}, [r0,:128]!
+	veor		q0, q0, q8
+	veor		q1, q1, q9
+	vst1.8		{q0-q1}, [r8]!
+
+	vld1.64		{q8}, [r0,:128]		@ next round tweak
+	b		.Lxts_dec_done
+.align	4
+.Lxts_dec_1:
+	mov		r0, sp
+	veor		q0, q8
+	mov		r1, sp
+	vst1.8		{q0}, [sp,:128]
+	mov		r2, r10
+	mov		r4, r3				@ preserve fp
+	mov		r5, r2			@ preserve magic
+
+	bl		AES_decrypt
+
+	vld1.8		{q0}, [sp,:128]
+	veor		q0, q0, q8
+	vst1.8		{q0}, [r8]!
+	mov		r3, r4
+	mov		r2, r5
+
+	vmov		q8, q9		@ next round tweak
+
+.Lxts_dec_done:
+#ifndef	XTS_CHAIN_TWEAK
+	adds		r9, #0x10
+	beq		.Lxts_dec_ret
+
+	@ calculate one round of extra tweak for the stolen ciphertext
+	vldmia		r2, {q5}
+	vshr.s64	q6, q8, #63
+	vand		q6, q6, q5
+	vadd.u64	q9, q8, q8
+	vswp		d13,d12
+	veor		q9, q9, q6
+
+	@ perform the final decryption with the last tweak value
+	vld1.8		{q0}, [r7]!
+	mov		r0, sp
+	veor		q0, q0, q9
+	mov		r1, sp
+	vst1.8		{q0}, [sp,:128]
+	mov		r2, r10
+	mov		r4, r3			@ preserve fp
+
+	bl		AES_decrypt
+
+	vld1.8		{q0}, [sp,:128]
+	veor		q0, q0, q9
+	vst1.8		{q0}, [r8]
+
+	mov		r6, r8
+.Lxts_dec_steal:
+	ldrb		r1, [r8]
+	ldrb		r0, [r7], #1
+	strb		r1, [r8, #0x10]
+	strb		r0, [r8], #1
+
+	subs		r9, #1
+	bhi		.Lxts_dec_steal
+
+	vld1.8		{q0}, [r6]
+	mov		r0, sp
+	veor		q0, q8
+	mov		r1, sp
+	vst1.8		{q0}, [sp,:128]
+	mov		r2, r10
+
+	bl		AES_decrypt
+
+	vld1.8		{q0}, [sp,:128]
+	veor		q0, q0, q8
+	vst1.8		{q0}, [r6]
+	mov		r3, r4
+#endif
+
+.Lxts_dec_ret:
+	bic		r0, r3, #0xf
+	vmov.i32	q0, #0
+	vmov.i32	q1, #0
+#ifdef	XTS_CHAIN_TWEAK
+	ldr		r1, [r3, #0x20+VFP_ABI_FRAME]	@ chain tweak
+#endif
+.Lxts_dec_bzero:				@ wipe key schedule [if any]
+	vstmia		sp!, {q0-q1}
+	cmp		sp, r0
+	bne		.Lxts_dec_bzero
+
+	mov		sp, r3
+#ifdef	XTS_CHAIN_TWEAK
+	vst1.8		{q8}, [r1]
+#endif
+	VFP_ABI_POP
+	ldmia		sp!, {r4-r10, pc}	@ return
+
+.size	bsaes_xts_decrypt,.-bsaes_xts_decrypt
+#endif
diff --git a/arch/arm/crypto/aesbs-glue.c b/arch/arm/crypto/aesbs-glue.c
new file mode 100644
index 000000000000..4522366da759
--- /dev/null
+++ b/arch/arm/crypto/aesbs-glue.c
@@ -0,0 +1,434 @@
+/*
+ * linux/arch/arm/crypto/aesbs-glue.c - glue code for NEON bit sliced AES
+ *
+ * Copyright (C) 2013 Linaro Ltd <ard.biesheuvel@linaro.org>
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2 as
+ * published by the Free Software Foundation.
+ */
+
+#include <asm/neon.h>
+#include <crypto/aes.h>
+#include <crypto/ablk_helper.h>
+#include <crypto/algapi.h>
+#include <linux/module.h>
+
+#include "aes_glue.h"
+
+#define BIT_SLICED_KEY_MAXSIZE	(128 * (AES_MAXNR - 1) + 2 * AES_BLOCK_SIZE)
+
+struct BS_KEY {
+	struct AES_KEY	rk;
+	int		converted;
+	u8 __aligned(8)	bs[BIT_SLICED_KEY_MAXSIZE];
+} __aligned(8);
+
+asmlinkage void bsaes_enc_key_convert(u8 out[], struct AES_KEY const *in);
+asmlinkage void bsaes_dec_key_convert(u8 out[], struct AES_KEY const *in);
+
+asmlinkage void bsaes_cbc_encrypt(u8 const in[], u8 out[], u32 bytes,
+				  struct BS_KEY *key, u8 iv[]);
+
+asmlinkage void bsaes_ctr32_encrypt_blocks(u8 const in[], u8 out[], u32 blocks,
+					   struct BS_KEY *key, u8 const iv[]);
+
+asmlinkage void bsaes_xts_encrypt(u8 const in[], u8 out[], u32 bytes,
+				  struct BS_KEY *key, u8 tweak[]);
+
+asmlinkage void bsaes_xts_decrypt(u8 const in[], u8 out[], u32 bytes,
+				  struct BS_KEY *key, u8 tweak[]);
+
+struct aesbs_cbc_ctx {
+	struct AES_KEY	enc;
+	struct BS_KEY	dec;
+};
+
+struct aesbs_ctr_ctx {
+	struct BS_KEY	enc;
+};
+
+struct aesbs_xts_ctx {
+	struct BS_KEY	enc;
+	struct BS_KEY	dec;
+	struct AES_KEY	twkey;
+};
+
+static int aesbs_cbc_set_key(struct crypto_tfm *tfm, const u8 *in_key,
+			     unsigned int key_len)
+{
+	struct aesbs_cbc_ctx *ctx = crypto_tfm_ctx(tfm);
+	int bits = key_len * 8;
+
+	if (private_AES_set_encrypt_key(in_key, bits, &ctx->enc)) {
+		tfm->crt_flags |= CRYPTO_TFM_RES_BAD_KEY_LEN;
+		return -EINVAL;
+	}
+	ctx->dec.rk = ctx->enc;
+	private_AES_set_decrypt_key(in_key, bits, &ctx->dec.rk);
+	ctx->dec.converted = 0;
+	return 0;
+}
+
+static int aesbs_ctr_set_key(struct crypto_tfm *tfm, const u8 *in_key,
+			     unsigned int key_len)
+{
+	struct aesbs_ctr_ctx *ctx = crypto_tfm_ctx(tfm);
+	int bits = key_len * 8;
+
+	if (private_AES_set_encrypt_key(in_key, bits, &ctx->enc.rk)) {
+		tfm->crt_flags |= CRYPTO_TFM_RES_BAD_KEY_LEN;
+		return -EINVAL;
+	}
+	ctx->enc.converted = 0;
+	return 0;
+}
+
+static int aesbs_xts_set_key(struct crypto_tfm *tfm, const u8 *in_key,
+			     unsigned int key_len)
+{
+	struct aesbs_xts_ctx *ctx = crypto_tfm_ctx(tfm);
+	int bits = key_len * 4;
+
+	if (private_AES_set_encrypt_key(in_key, bits, &ctx->enc.rk)) {
+		tfm->crt_flags |= CRYPTO_TFM_RES_BAD_KEY_LEN;
+		return -EINVAL;
+	}
+	ctx->dec.rk = ctx->enc.rk;
+	private_AES_set_decrypt_key(in_key, bits, &ctx->dec.rk);
+	private_AES_set_encrypt_key(in_key + key_len / 2, bits, &ctx->twkey);
+	ctx->enc.converted = ctx->dec.converted = 0;
+	return 0;
+}
+
+static int aesbs_cbc_encrypt(struct blkcipher_desc *desc,
+			     struct scatterlist *dst,
+			     struct scatterlist *src, unsigned int nbytes)
+{
+	struct aesbs_cbc_ctx *ctx = crypto_blkcipher_ctx(desc->tfm);
+	struct blkcipher_walk walk;
+	int err;
+
+	blkcipher_walk_init(&walk, dst, src, nbytes);
+	err = blkcipher_walk_virt(desc, &walk);
+
+	while (walk.nbytes) {
+		u32 blocks = walk.nbytes / AES_BLOCK_SIZE;
+		u8 *src = walk.src.virt.addr;
+
+		if (walk.dst.virt.addr == walk.src.virt.addr) {
+			u8 *iv = walk.iv;
+
+			do {
+				crypto_xor(src, iv, AES_BLOCK_SIZE);
+				AES_encrypt(src, src, &ctx->enc);
+				iv = src;
+				src += AES_BLOCK_SIZE;
+			} while (--blocks);
+			memcpy(walk.iv, iv, AES_BLOCK_SIZE);
+		} else {
+			u8 *dst = walk.dst.virt.addr;
+
+			do {
+				crypto_xor(walk.iv, src, AES_BLOCK_SIZE);
+				AES_encrypt(walk.iv, dst, &ctx->enc);
+				memcpy(walk.iv, dst, AES_BLOCK_SIZE);
+				src += AES_BLOCK_SIZE;
+				dst += AES_BLOCK_SIZE;
+			} while (--blocks);
+		}
+		err = blkcipher_walk_done(desc, &walk, 0);
+	}
+	return err;
+}
+
+static int aesbs_cbc_decrypt(struct blkcipher_desc *desc,
+			     struct scatterlist *dst,
+			     struct scatterlist *src, unsigned int nbytes)
+{
+	struct aesbs_cbc_ctx *ctx = crypto_blkcipher_ctx(desc->tfm);
+	struct blkcipher_walk walk;
+	int err;
+
+	blkcipher_walk_init(&walk, dst, src, nbytes);
+	err = blkcipher_walk_virt_block(desc, &walk, 8 * AES_BLOCK_SIZE);
+
+	while ((walk.nbytes / AES_BLOCK_SIZE) >= 8) {
+		kernel_neon_begin();
+		bsaes_cbc_encrypt(walk.src.virt.addr, walk.dst.virt.addr,
+				  walk.nbytes, &ctx->dec, walk.iv);
+		kernel_neon_end();
+		err = blkcipher_walk_done(desc, &walk, 0);
+	}
+	while (walk.nbytes) {
+		u32 blocks = walk.nbytes / AES_BLOCK_SIZE;
+		u8 *dst = walk.dst.virt.addr;
+		u8 *src = walk.src.virt.addr;
+		u8 bk[2][AES_BLOCK_SIZE];
+		u8 *iv = walk.iv;
+
+		do {
+			if (walk.dst.virt.addr == walk.src.virt.addr)
+				memcpy(bk[blocks & 1], src, AES_BLOCK_SIZE);
+
+			AES_decrypt(src, dst, &ctx->dec.rk);
+			crypto_xor(dst, iv, AES_BLOCK_SIZE);
+
+			if (walk.dst.virt.addr == walk.src.virt.addr)
+				iv = bk[blocks & 1];
+			else
+				iv = src;
+
+			dst += AES_BLOCK_SIZE;
+			src += AES_BLOCK_SIZE;
+		} while (--blocks);
+		err = blkcipher_walk_done(desc, &walk, 0);
+	}
+	return err;
+}
+
+static void inc_be128_ctr(__be32 ctr[], u32 addend)
+{
+	int i;
+
+	for (i = 3; i >= 0; i--, addend = 1) {
+		u32 n = be32_to_cpu(ctr[i]) + addend;
+
+		ctr[i] = cpu_to_be32(n);
+		if (n >= addend)
+			break;
+	}
+}
+
+static int aesbs_ctr_encrypt(struct blkcipher_desc *desc,
+			     struct scatterlist *dst, struct scatterlist *src,
+			     unsigned int nbytes)
+{
+	struct aesbs_ctr_ctx *ctx = crypto_blkcipher_ctx(desc->tfm);
+	struct blkcipher_walk walk;
+	u32 blocks;
+	int err;
+
+	blkcipher_walk_init(&walk, dst, src, nbytes);
+	err = blkcipher_walk_virt_block(desc, &walk, 8 * AES_BLOCK_SIZE);
+
+	while ((blocks = walk.nbytes / AES_BLOCK_SIZE)) {
+		u32 tail = walk.nbytes % AES_BLOCK_SIZE;
+		__be32 *ctr = (__be32 *)walk.iv;
+		u32 headroom = UINT_MAX - be32_to_cpu(ctr[3]);
+
+		/* avoid 32 bit counter overflow in the NEON code */
+		if (unlikely(headroom < blocks)) {
+			blocks = headroom + 1;
+			tail = walk.nbytes - blocks * AES_BLOCK_SIZE;
+		}
+		kernel_neon_begin();
+		bsaes_ctr32_encrypt_blocks(walk.src.virt.addr,
+					   walk.dst.virt.addr, blocks,
+					   &ctx->enc, walk.iv);
+		kernel_neon_end();
+		inc_be128_ctr(ctr, blocks);
+
+		nbytes -= blocks * AES_BLOCK_SIZE;
+		if (nbytes && nbytes == tail && nbytes <= AES_BLOCK_SIZE)
+			break;
+
+		err = blkcipher_walk_done(desc, &walk, tail);
+	}
+	if (walk.nbytes) {
+		u8 *tdst = walk.dst.virt.addr + blocks * AES_BLOCK_SIZE;
+		u8 *tsrc = walk.src.virt.addr + blocks * AES_BLOCK_SIZE;
+		u8 ks[AES_BLOCK_SIZE];
+
+		AES_encrypt(walk.iv, ks, &ctx->enc.rk);
+		if (tdst != tsrc)
+			memcpy(tdst, tsrc, nbytes);
+		crypto_xor(tdst, ks, nbytes);
+		err = blkcipher_walk_done(desc, &walk, 0);
+	}
+	return err;
+}
+
+static int aesbs_xts_encrypt(struct blkcipher_desc *desc,
+			     struct scatterlist *dst,
+			     struct scatterlist *src, unsigned int nbytes)
+{
+	struct aesbs_xts_ctx *ctx = crypto_blkcipher_ctx(desc->tfm);
+	struct blkcipher_walk walk;
+	int err;
+
+	blkcipher_walk_init(&walk, dst, src, nbytes);
+	err = blkcipher_walk_virt_block(desc, &walk, 8 * AES_BLOCK_SIZE);
+
+	/* generate the initial tweak */
+	AES_encrypt(walk.iv, walk.iv, &ctx->twkey);
+
+	while (walk.nbytes) {
+		kernel_neon_begin();
+		bsaes_xts_encrypt(walk.src.virt.addr, walk.dst.virt.addr,
+				  walk.nbytes, &ctx->enc, walk.iv);
+		kernel_neon_end();
+		err = blkcipher_walk_done(desc, &walk, 0);
+	}
+	return err;
+}
+
+static int aesbs_xts_decrypt(struct blkcipher_desc *desc,
+			     struct scatterlist *dst,
+			     struct scatterlist *src, unsigned int nbytes)
+{
+	struct aesbs_xts_ctx *ctx = crypto_blkcipher_ctx(desc->tfm);
+	struct blkcipher_walk walk;
+	int err;
+
+	blkcipher_walk_init(&walk, dst, src, nbytes);
+	err = blkcipher_walk_virt_block(desc, &walk, 8 * AES_BLOCK_SIZE);
+
+	/* generate the initial tweak */
+	AES_encrypt(walk.iv, walk.iv, &ctx->twkey);
+
+	while (walk.nbytes) {
+		kernel_neon_begin();
+		bsaes_xts_decrypt(walk.src.virt.addr, walk.dst.virt.addr,
+				  walk.nbytes, &ctx->dec, walk.iv);
+		kernel_neon_end();
+		err = blkcipher_walk_done(desc, &walk, 0);
+	}
+	return err;
+}
+
+static struct crypto_alg aesbs_algs[] = { {
+	.cra_name		= "__cbc-aes-neonbs",
+	.cra_driver_name	= "__driver-cbc-aes-neonbs",
+	.cra_priority		= 0,
+	.cra_flags		= CRYPTO_ALG_TYPE_BLKCIPHER,
+	.cra_blocksize		= AES_BLOCK_SIZE,
+	.cra_ctxsize		= sizeof(struct aesbs_cbc_ctx),
+	.cra_alignmask		= 7,
+	.cra_type		= &crypto_blkcipher_type,
+	.cra_module		= THIS_MODULE,
+	.cra_blkcipher = {
+		.min_keysize	= AES_MIN_KEY_SIZE,
+		.max_keysize	= AES_MAX_KEY_SIZE,
+		.ivsize		= AES_BLOCK_SIZE,
+		.setkey		= aesbs_cbc_set_key,
+		.encrypt	= aesbs_cbc_encrypt,
+		.decrypt	= aesbs_cbc_decrypt,
+	},
+}, {
+	.cra_name		= "__ctr-aes-neonbs",
+	.cra_driver_name	= "__driver-ctr-aes-neonbs",
+	.cra_priority		= 0,
+	.cra_flags		= CRYPTO_ALG_TYPE_BLKCIPHER,
+	.cra_blocksize		= 1,
+	.cra_ctxsize		= sizeof(struct aesbs_ctr_ctx),
+	.cra_alignmask		= 7,
+	.cra_type		= &crypto_blkcipher_type,
+	.cra_module		= THIS_MODULE,
+	.cra_blkcipher = {
+		.min_keysize	= AES_MIN_KEY_SIZE,
+		.max_keysize	= AES_MAX_KEY_SIZE,
+		.ivsize		= AES_BLOCK_SIZE,
+		.setkey		= aesbs_ctr_set_key,
+		.encrypt	= aesbs_ctr_encrypt,
+		.decrypt	= aesbs_ctr_encrypt,
+	},
+}, {
+	.cra_name		= "__xts-aes-neonbs",
+	.cra_driver_name	= "__driver-xts-aes-neonbs",
+	.cra_priority		= 0,
+	.cra_flags		= CRYPTO_ALG_TYPE_BLKCIPHER,
+	.cra_blocksize		= AES_BLOCK_SIZE,
+	.cra_ctxsize		= sizeof(struct aesbs_xts_ctx),
+	.cra_alignmask		= 7,
+	.cra_type		= &crypto_blkcipher_type,
+	.cra_module		= THIS_MODULE,
+	.cra_blkcipher = {
+		.min_keysize	= 2 * AES_MIN_KEY_SIZE,
+		.max_keysize	= 2 * AES_MAX_KEY_SIZE,
+		.ivsize		= AES_BLOCK_SIZE,
+		.setkey		= aesbs_xts_set_key,
+		.encrypt	= aesbs_xts_encrypt,
+		.decrypt	= aesbs_xts_decrypt,
+	},
+}, {
+	.cra_name		= "cbc(aes)",
+	.cra_driver_name	= "cbc-aes-neonbs",
+	.cra_priority		= 300,
+	.cra_flags		= CRYPTO_ALG_TYPE_ABLKCIPHER|CRYPTO_ALG_ASYNC,
+	.cra_blocksize		= AES_BLOCK_SIZE,
+	.cra_ctxsize		= sizeof(struct async_helper_ctx),
+	.cra_alignmask		= 7,
+	.cra_type		= &crypto_ablkcipher_type,
+	.cra_module		= THIS_MODULE,
+	.cra_init		= ablk_init,
+	.cra_exit		= ablk_exit,
+	.cra_ablkcipher = {
+		.min_keysize	= AES_MIN_KEY_SIZE,
+		.max_keysize	= AES_MAX_KEY_SIZE,
+		.ivsize		= AES_BLOCK_SIZE,
+		.setkey		= ablk_set_key,
+		.encrypt	= __ablk_encrypt,
+		.decrypt	= ablk_decrypt,
+	}
+}, {
+	.cra_name		= "ctr(aes)",
+	.cra_driver_name	= "ctr-aes-neonbs",
+	.cra_priority		= 300,
+	.cra_flags		= CRYPTO_ALG_TYPE_ABLKCIPHER|CRYPTO_ALG_ASYNC,
+	.cra_blocksize		= 1,
+	.cra_ctxsize		= sizeof(struct async_helper_ctx),
+	.cra_alignmask		= 7,
+	.cra_type		= &crypto_ablkcipher_type,
+	.cra_module		= THIS_MODULE,
+	.cra_init		= ablk_init,
+	.cra_exit		= ablk_exit,
+	.cra_ablkcipher = {
+		.min_keysize	= AES_MIN_KEY_SIZE,
+		.max_keysize	= AES_MAX_KEY_SIZE,
+		.ivsize		= AES_BLOCK_SIZE,
+		.setkey		= ablk_set_key,
+		.encrypt	= ablk_encrypt,
+		.decrypt	= ablk_decrypt,
+	}
+}, {
+	.cra_name		= "xts(aes)",
+	.cra_driver_name	= "xts-aes-neonbs",
+	.cra_priority		= 300,
+	.cra_flags		= CRYPTO_ALG_TYPE_ABLKCIPHER|CRYPTO_ALG_ASYNC,
+	.cra_blocksize		= AES_BLOCK_SIZE,
+	.cra_ctxsize		= sizeof(struct async_helper_ctx),
+	.cra_alignmask		= 7,
+	.cra_type		= &crypto_ablkcipher_type,
+	.cra_module		= THIS_MODULE,
+	.cra_init		= ablk_init,
+	.cra_exit		= ablk_exit,
+	.cra_ablkcipher = {
+		.min_keysize	= 2 * AES_MIN_KEY_SIZE,
+		.max_keysize	= 2 * AES_MAX_KEY_SIZE,
+		.ivsize		= AES_BLOCK_SIZE,
+		.setkey		= ablk_set_key,
+		.encrypt	= ablk_encrypt,
+		.decrypt	= ablk_decrypt,
+	}
+} };
+
+static int __init aesbs_mod_init(void)
+{
+	if (!cpu_has_neon())
+		return -ENODEV;
+
+	return crypto_register_algs(aesbs_algs, ARRAY_SIZE(aesbs_algs));
+}
+
+static void __exit aesbs_mod_exit(void)
+{
+	crypto_unregister_algs(aesbs_algs, ARRAY_SIZE(aesbs_algs));
+}
+
+module_init(aesbs_mod_init);
+module_exit(aesbs_mod_exit);
+
+MODULE_DESCRIPTION("Bit sliced AES in CBC/CTR/XTS modes using NEON");
+MODULE_AUTHOR("Ard Biesheuvel <ard.biesheuvel@linaro.org>");
+MODULE_LICENSE("GPL");
diff --git a/arch/arm/crypto/bsaes-armv7.pl b/arch/arm/crypto/bsaes-armv7.pl
new file mode 100644
index 000000000000..f3d96d932573
--- /dev/null
+++ b/arch/arm/crypto/bsaes-armv7.pl
@@ -0,0 +1,2467 @@
+#!/usr/bin/env perl
+
+# ====================================================================
+# Written by Andy Polyakov <appro@openssl.org> for the OpenSSL
+# project. The module is, however, dual licensed under OpenSSL and
+# CRYPTOGAMS licenses depending on where you obtain it. For further
+# details see http://www.openssl.org/~appro/cryptogams/.
+#
+# Specific modes and adaptation for Linux kernel by Ard Biesheuvel
+# <ard.biesheuvel@linaro.org>. Permission to use under GPL terms is
+# granted.
+# ====================================================================
+
+# Bit-sliced AES for ARM NEON
+#
+# February 2012.
+#
+# This implementation is direct adaptation of bsaes-x86_64 module for
+# ARM NEON. Except that this module is endian-neutral [in sense that
+# it can be compiled for either endianness] by courtesy of vld1.8's
+# neutrality. Initial version doesn't implement interface to OpenSSL,
+# only low-level primitives and unsupported entry points, just enough
+# to collect performance results, which for Cortex-A8 core are:
+#
+# encrypt	19.5 cycles per byte processed with 128-bit key
+# decrypt	22.1 cycles per byte processed with 128-bit key
+# key conv.	440  cycles per 128-bit key/0.18 of 8x block
+#
+# Snapdragon S4 encrypts byte in 17.6 cycles and decrypts in 19.7,
+# which is [much] worse than anticipated (for further details see
+# http://www.openssl.org/~appro/Snapdragon-S4.html).
+#
+# Cortex-A15 manages in 14.2/16.1 cycles [when integer-only code
+# manages in 20.0 cycles].
+#
+# When comparing to x86_64 results keep in mind that NEON unit is
+# [mostly] single-issue and thus can't [fully] benefit from
+# instruction-level parallelism. And when comparing to aes-armv4
+# results keep in mind key schedule conversion overhead (see
+# bsaes-x86_64.pl for further details)...
+#
+#						<appro@openssl.org>
+
+# April-August 2013
+#
+# Add CBC, CTR and XTS subroutines, adapt for kernel use.
+#
+#					<ard.biesheuvel@linaro.org>
+
+while (($output=shift) && ($output!~/^\w[\w\-]*\.\w+$/)) {}
+open STDOUT,">$output";
+
+my ($inp,$out,$len,$key)=("r0","r1","r2","r3");
+my @XMM=map("q$_",(0..15));
+
+{
+my ($key,$rounds,$const)=("r4","r5","r6");
+
+sub Dlo()   { shift=~m|q([1]?[0-9])|?"d".($1*2):"";     }
+sub Dhi()   { shift=~m|q([1]?[0-9])|?"d".($1*2+1):"";   }
+
+sub Sbox {
+# input in  lsb > [b0, b1, b2, b3, b4, b5, b6, b7] < msb
+# output in lsb > [b0, b1, b4, b6, b3, b7, b2, b5] < msb
+my @b=@_[0..7];
+my @t=@_[8..11];
+my @s=@_[12..15];
+	&InBasisChange	(@b);
+	&Inv_GF256	(@b[6,5,0,3,7,1,4,2],@t,@s);
+	&OutBasisChange	(@b[7,1,4,2,6,5,0,3]);
+}
+
+sub InBasisChange {
+# input in  lsb > [b0, b1, b2, b3, b4, b5, b6, b7] < msb
+# output in lsb > [b6, b5, b0, b3, b7, b1, b4, b2] < msb 
+my @b=@_[0..7];
+$code.=<<___;
+	veor	@b[2], @b[2], @b[1]
+	veor	@b[5], @b[5], @b[6]
+	veor	@b[3], @b[3], @b[0]
+	veor	@b[6], @b[6], @b[2]
+	veor	@b[5], @b[5], @b[0]
+
+	veor	@b[6], @b[6], @b[3]
+	veor	@b[3], @b[3], @b[7]
+	veor	@b[7], @b[7], @b[5]
+	veor	@b[3], @b[3], @b[4]
+	veor	@b[4], @b[4], @b[5]
+
+	veor	@b[2], @b[2], @b[7]
+	veor	@b[3], @b[3], @b[1]
+	veor	@b[1], @b[1], @b[5]
+___
+}
+
+sub OutBasisChange {
+# input in  lsb > [b0, b1, b2, b3, b4, b5, b6, b7] < msb
+# output in lsb > [b6, b1, b2, b4, b7, b0, b3, b5] < msb
+my @b=@_[0..7];
+$code.=<<___;
+	veor	@b[0], @b[0], @b[6]
+	veor	@b[1], @b[1], @b[4]
+	veor	@b[4], @b[4], @b[6]
+	veor	@b[2], @b[2], @b[0]
+	veor	@b[6], @b[6], @b[1]
+
+	veor	@b[1], @b[1], @b[5]
+	veor	@b[5], @b[5], @b[3]
+	veor	@b[3], @b[3], @b[7]
+	veor	@b[7], @b[7], @b[5]
+	veor	@b[2], @b[2], @b[5]
+
+	veor	@b[4], @b[4], @b[7]
+___
+}
+
+sub InvSbox {
+# input in lsb 	> [b0, b1, b2, b3, b4, b5, b6, b7] < msb
+# output in lsb	> [b0, b1, b6, b4, b2, b7, b3, b5] < msb
+my @b=@_[0..7];
+my @t=@_[8..11];
+my @s=@_[12..15];
+	&InvInBasisChange	(@b);
+	&Inv_GF256		(@b[5,1,2,6,3,7,0,4],@t,@s);
+	&InvOutBasisChange	(@b[3,7,0,4,5,1,2,6]);
+}
+
+sub InvInBasisChange {		# OutBasisChange in reverse (with twist)
+my @b=@_[5,1,2,6,3,7,0,4];
+$code.=<<___
+	 veor	@b[1], @b[1], @b[7]
+	veor	@b[4], @b[4], @b[7]
+
+	veor	@b[7], @b[7], @b[5]
+	 veor	@b[1], @b[1], @b[3]
+	veor	@b[2], @b[2], @b[5]
+	veor	@b[3], @b[3], @b[7]
+
+	veor	@b[6], @b[6], @b[1]
+	veor	@b[2], @b[2], @b[0]
+	 veor	@b[5], @b[5], @b[3]
+	veor	@b[4], @b[4], @b[6]
+	veor	@b[0], @b[0], @b[6]
+	veor	@b[1], @b[1], @b[4]
+___
+}
+
+sub InvOutBasisChange {		# InBasisChange in reverse
+my @b=@_[2,5,7,3,6,1,0,4];
+$code.=<<___;
+	veor	@b[1], @b[1], @b[5]
+	veor	@b[2], @b[2], @b[7]
+
+	veor	@b[3], @b[3], @b[1]
+	veor	@b[4], @b[4], @b[5]
+	veor	@b[7], @b[7], @b[5]
+	veor	@b[3], @b[3], @b[4]
+	 veor 	@b[5], @b[5], @b[0]
+	veor	@b[3], @b[3], @b[7]
+	 veor	@b[6], @b[6], @b[2]
+	 veor	@b[2], @b[2], @b[1]
+	veor	@b[6], @b[6], @b[3]
+
+	veor	@b[3], @b[3], @b[0]
+	veor	@b[5], @b[5], @b[6]
+___
+}
+
+sub Mul_GF4 {
+#;*************************************************************
+#;* Mul_GF4: Input x0-x1,y0-y1 Output x0-x1 Temp t0 (8) *
+#;*************************************************************
+my ($x0,$x1,$y0,$y1,$t0,$t1)=@_;
+$code.=<<___;
+	veor 	$t0, $y0, $y1
+	vand	$t0, $t0, $x0
+	veor	$x0, $x0, $x1
+	vand	$t1, $x1, $y0
+	vand	$x0, $x0, $y1
+	veor	$x1, $t1, $t0
+	veor	$x0, $x0, $t1
+___
+}
+
+sub Mul_GF4_N {				# not used, see next subroutine
+# multiply and scale by N
+my ($x0,$x1,$y0,$y1,$t0)=@_;
+$code.=<<___;
+	veor	$t0, $y0, $y1
+	vand	$t0, $t0, $x0
+	veor	$x0, $x0, $x1
+	vand	$x1, $x1, $y0
+	vand	$x0, $x0, $y1
+	veor	$x1, $x1, $x0
+	veor	$x0, $x0, $t0
+___
+}
+
+sub Mul_GF4_N_GF4 {
+# interleaved Mul_GF4_N and Mul_GF4
+my ($x0,$x1,$y0,$y1,$t0,
+    $x2,$x3,$y2,$y3,$t1)=@_;
+$code.=<<___;
+	veor	$t0, $y0, $y1
+	 veor 	$t1, $y2, $y3
+	vand	$t0, $t0, $x0
+	 vand	$t1, $t1, $x2
+	veor	$x0, $x0, $x1
+	 veor	$x2, $x2, $x3
+	vand	$x1, $x1, $y0
+	 vand	$x3, $x3, $y2
+	vand	$x0, $x0, $y1
+	 vand	$x2, $x2, $y3
+	veor	$x1, $x1, $x0
+	 veor	$x2, $x2, $x3
+	veor	$x0, $x0, $t0
+	 veor	$x3, $x3, $t1
+___
+}
+sub Mul_GF16_2 {
+my @x=@_[0..7];
+my @y=@_[8..11];
+my @t=@_[12..15];
+$code.=<<___;
+	veor	@t[0], @x[0], @x[2]
+	veor	@t[1], @x[1], @x[3]
+___
+	&Mul_GF4  	(@x[0], @x[1], @y[0], @y[1], @t[2..3]);
+$code.=<<___;
+	veor	@y[0], @y[0], @y[2]
+	veor	@y[1], @y[1], @y[3]
+___
+	Mul_GF4_N_GF4	(@t[0], @t[1], @y[0], @y[1], @t[3],
+			 @x[2], @x[3], @y[2], @y[3], @t[2]);
+$code.=<<___;
+	veor	@x[0], @x[0], @t[0]
+	veor	@x[2], @x[2], @t[0]
+	veor	@x[1], @x[1], @t[1]
+	veor	@x[3], @x[3], @t[1]
+
+	veor	@t[0], @x[4], @x[6]
+	veor	@t[1], @x[5], @x[7]
+___
+	&Mul_GF4_N_GF4	(@t[0], @t[1], @y[0], @y[1], @t[3],
+			 @x[6], @x[7], @y[2], @y[3], @t[2]);
+$code.=<<___;
+	veor	@y[0], @y[0], @y[2]
+	veor	@y[1], @y[1], @y[3]
+___
+	&Mul_GF4  	(@x[4], @x[5], @y[0], @y[1], @t[2..3]);
+$code.=<<___;
+	veor	@x[4], @x[4], @t[0]
+	veor	@x[6], @x[6], @t[0]
+	veor	@x[5], @x[5], @t[1]
+	veor	@x[7], @x[7], @t[1]
+___
+}
+sub Inv_GF256 {
+#;********************************************************************
+#;* Inv_GF256: Input x0-x7 Output x0-x7 Temp t0-t3,s0-s3 (144)       *
+#;********************************************************************
+my @x=@_[0..7];
+my @t=@_[8..11];
+my @s=@_[12..15];
+# direct optimizations from hardware
+$code.=<<___;
+	veor	@t[3], @x[4], @x[6]
+	veor	@t[2], @x[5], @x[7]
+	veor	@t[1], @x[1], @x[3]
+	veor	@s[1], @x[7], @x[6]
+	 vmov	@t[0], @t[2]
+	veor	@s[0], @x[0], @x[2]
+
+	vorr	@t[2], @t[2], @t[1]
+	veor	@s[3], @t[3], @t[0]
+	vand	@s[2], @t[3], @s[0]
+	vorr	@t[3], @t[3], @s[0]
+	veor	@s[0], @s[0], @t[1]
+	vand	@t[0], @t[0], @t[1]
+	veor	@t[1], @x[3], @x[2]
+	vand	@s[3], @s[3], @s[0]
+	vand	@s[1], @s[1], @t[1]
+	veor	@t[1], @x[4], @x[5]
+	veor	@s[0], @x[1], @x[0]
+	veor	@t[3], @t[3], @s[1]
+	veor	@t[2], @t[2], @s[1]
+	vand	@s[1], @t[1], @s[0]
+	vorr	@t[1], @t[1], @s[0]
+	veor	@t[3], @t[3], @s[3]
+	veor	@t[0], @t[0], @s[1]
+	veor	@t[2], @t[2], @s[2]
+	veor	@t[1], @t[1], @s[3]
+	veor	@t[0], @t[0], @s[2]
+	vand	@s[0], @x[7], @x[3]
+	veor	@t[1], @t[1], @s[2]
+	vand	@s[1], @x[6], @x[2]
+	vand	@s[2], @x[5], @x[1]
+	vorr	@s[3], @x[4], @x[0]
+	veor	@t[3], @t[3], @s[0]
+	veor	@t[1], @t[1], @s[2]
+	veor	@t[0], @t[0], @s[3]
+	veor	@t[2], @t[2], @s[1]
+
+	@ Inv_GF16 \t0, \t1, \t2, \t3, \s0, \s1, \s2, \s3
+
+	@ new smaller inversion
+
+	vand	@s[2], @t[3], @t[1]
+	vmov	@s[0], @t[0]
+
+	veor	@s[1], @t[2], @s[2]
+	veor	@s[3], @t[0], @s[2]
+	veor	@s[2], @t[0], @s[2]	@ @s[2]=@s[3]
+
+	vbsl	@s[1], @t[1], @t[0]
+	vbsl	@s[3], @t[3], @t[2]
+	veor	@t[3], @t[3], @t[2]
+
+	vbsl	@s[0], @s[1], @s[2]
+	vbsl	@t[0], @s[2], @s[1]
+
+	vand	@s[2], @s[0], @s[3]
+	veor	@t[1], @t[1], @t[0]
+
+	veor	@s[2], @s[2], @t[3]
+___
+# output in s3, s2, s1, t1
+
+# Mul_GF16_2 \x0, \x1, \x2, \x3, \x4, \x5, \x6, \x7, \t2, \t3, \t0, \t1, \s0, \s1, \s2, \s3
+
+# Mul_GF16_2 \x0, \x1, \x2, \x3, \x4, \x5, \x6, \x7, \s3, \s2, \s1, \t1, \s0, \t0, \t2, \t3
+	&Mul_GF16_2(@x,@s[3,2,1],@t[1],@s[0],@t[0,2,3]);
+
+### output msb > [x3,x2,x1,x0,x7,x6,x5,x4] < lsb
+}
+
+# AES linear components
+
+sub ShiftRows {
+my @x=@_[0..7];
+my @t=@_[8..11];
+my $mask=pop;
+$code.=<<___;
+	vldmia	$key!, {@t[0]-@t[3]}
+	veor	@t[0], @t[0], @x[0]
+	veor	@t[1], @t[1], @x[1]
+	vtbl.8	`&Dlo(@x[0])`, {@t[0]}, `&Dlo($mask)`
+	vtbl.8	`&Dhi(@x[0])`, {@t[0]}, `&Dhi($mask)`
+	vldmia	$key!, {@t[0]}
+	veor	@t[2], @t[2], @x[2]
+	vtbl.8	`&Dlo(@x[1])`, {@t[1]}, `&Dlo($mask)`
+	vtbl.8	`&Dhi(@x[1])`, {@t[1]}, `&Dhi($mask)`
+	vldmia	$key!, {@t[1]}
+	veor	@t[3], @t[3], @x[3]
+	vtbl.8	`&Dlo(@x[2])`, {@t[2]}, `&Dlo($mask)`
+	vtbl.8	`&Dhi(@x[2])`, {@t[2]}, `&Dhi($mask)`
+	vldmia	$key!, {@t[2]}
+	vtbl.8	`&Dlo(@x[3])`, {@t[3]}, `&Dlo($mask)`
+	vtbl.8	`&Dhi(@x[3])`, {@t[3]}, `&Dhi($mask)`
+	vldmia	$key!, {@t[3]}
+	veor	@t[0], @t[0], @x[4]
+	veor	@t[1], @t[1], @x[5]
+	vtbl.8	`&Dlo(@x[4])`, {@t[0]}, `&Dlo($mask)`
+	vtbl.8	`&Dhi(@x[4])`, {@t[0]}, `&Dhi($mask)`
+	veor	@t[2], @t[2], @x[6]
+	vtbl.8	`&Dlo(@x[5])`, {@t[1]}, `&Dlo($mask)`
+	vtbl.8	`&Dhi(@x[5])`, {@t[1]}, `&Dhi($mask)`
+	veor	@t[3], @t[3], @x[7]
+	vtbl.8	`&Dlo(@x[6])`, {@t[2]}, `&Dlo($mask)`
+	vtbl.8	`&Dhi(@x[6])`, {@t[2]}, `&Dhi($mask)`
+	vtbl.8	`&Dlo(@x[7])`, {@t[3]}, `&Dlo($mask)`
+	vtbl.8	`&Dhi(@x[7])`, {@t[3]}, `&Dhi($mask)`
+___
+}
+
+sub MixColumns {
+# modified to emit output in order suitable for feeding back to aesenc[last]
+my @x=@_[0..7];
+my @t=@_[8..15];
+my $inv=@_[16];	# optional
+$code.=<<___;
+	vext.8	@t[0], @x[0], @x[0], #12	@ x0 <<< 32
+	vext.8	@t[1], @x[1], @x[1], #12
+	 veor	@x[0], @x[0], @t[0]		@ x0 ^ (x0 <<< 32)
+	vext.8	@t[2], @x[2], @x[2], #12
+	 veor	@x[1], @x[1], @t[1]
+	vext.8	@t[3], @x[3], @x[3], #12
+	 veor	@x[2], @x[2], @t[2]
+	vext.8	@t[4], @x[4], @x[4], #12
+	 veor	@x[3], @x[3], @t[3]
+	vext.8	@t[5], @x[5], @x[5], #12
+	 veor	@x[4], @x[4], @t[4]
+	vext.8	@t[6], @x[6], @x[6], #12
+	 veor	@x[5], @x[5], @t[5]
+	vext.8	@t[7], @x[7], @x[7], #12
+	 veor	@x[6], @x[6], @t[6]
+
+	veor	@t[1], @t[1], @x[0]
+	 veor	@x[7], @x[7], @t[7]
+	 vext.8	@x[0], @x[0], @x[0], #8		@ (x0 ^ (x0 <<< 32)) <<< 64)
+	veor	@t[2], @t[2], @x[1]
+	veor	@t[0], @t[0], @x[7]
+	veor	@t[1], @t[1], @x[7]
+	 vext.8	@x[1], @x[1], @x[1], #8
+	veor	@t[5], @t[5], @x[4]
+	 veor	@x[0], @x[0], @t[0]
+	veor	@t[6], @t[6], @x[5]
+	 veor	@x[1], @x[1], @t[1]
+	 vext.8	@t[0], @x[4], @x[4], #8
+	veor	@t[4], @t[4], @x[3]
+	 vext.8	@t[1], @x[5], @x[5], #8
+	veor	@t[7], @t[7], @x[6]
+	 vext.8	@x[4], @x[3], @x[3], #8
+	veor	@t[3], @t[3], @x[2]
+	 vext.8	@x[5], @x[7], @x[7], #8
+	veor	@t[4], @t[4], @x[7]
+	 vext.8	@x[3], @x[6], @x[6], #8
+	veor	@t[3], @t[3], @x[7]
+	 vext.8	@x[6], @x[2], @x[2], #8
+	veor	@x[7], @t[1], @t[5]
+___
+$code.=<<___ if (!$inv);
+	veor	@x[2], @t[0], @t[4]
+	veor	@x[4], @x[4], @t[3]
+	veor	@x[5], @x[5], @t[7]
+	veor	@x[3], @x[3], @t[6]
+	 @ vmov	@x[2], @t[0]
+	veor	@x[6], @x[6], @t[2]
+	 @ vmov	@x[7], @t[1]
+___
+$code.=<<___ if ($inv);
+	veor	@t[3], @t[3], @x[4]
+	veor	@x[5], @x[5], @t[7]
+	veor	@x[2], @x[3], @t[6]
+	veor	@x[3], @t[0], @t[4]
+	veor	@x[4], @x[6], @t[2]
+	vmov	@x[6], @t[3]
+	 @ vmov	@x[7], @t[1]
+___
+}
+
+sub InvMixColumns_orig {
+my @x=@_[0..7];
+my @t=@_[8..15];
+
+$code.=<<___;
+	@ multiplication by 0x0e
+	vext.8	@t[7], @x[7], @x[7], #12
+	vmov	@t[2], @x[2]
+	veor	@x[2], @x[2], @x[5]		@ 2 5
+	veor	@x[7], @x[7], @x[5]		@ 7 5
+	vext.8	@t[0], @x[0], @x[0], #12
+	vmov	@t[5], @x[5]
+	veor	@x[5], @x[5], @x[0]		@ 5 0		[1]
+	veor	@x[0], @x[0], @x[1]		@ 0 1
+	vext.8	@t[1], @x[1], @x[1], #12
+	veor	@x[1], @x[1], @x[2]		@ 1 25
+	veor	@x[0], @x[0], @x[6]		@ 01 6		[2]
+	vext.8	@t[3], @x[3], @x[3], #12
+	veor	@x[1], @x[1], @x[3]		@ 125 3		[4]
+	veor	@x[2], @x[2], @x[0]		@ 25 016	[3]
+	veor	@x[3], @x[3], @x[7]		@ 3 75
+	veor	@x[7], @x[7], @x[6]		@ 75 6		[0]
+	vext.8	@t[6], @x[6], @x[6], #12
+	vmov	@t[4], @x[4]
+	veor	@x[6], @x[6], @x[4]		@ 6 4
+	veor	@x[4], @x[4], @x[3]		@ 4 375		[6]
+	veor	@x[3], @x[3], @x[7]		@ 375 756=36
+	veor	@x[6], @x[6], @t[5]		@ 64 5		[7]
+	veor	@x[3], @x[3], @t[2]		@ 36 2
+	vext.8	@t[5], @t[5], @t[5], #12
+	veor	@x[3], @x[3], @t[4]		@ 362 4		[5]
+___
+					my @y = @x[7,5,0,2,1,3,4,6];
+$code.=<<___;
+	@ multiplication by 0x0b
+	veor	@y[1], @y[1], @y[0]
+	veor	@y[0], @y[0], @t[0]
+	vext.8	@t[2], @t[2], @t[2], #12
+	veor	@y[1], @y[1], @t[1]
+	veor	@y[0], @y[0], @t[5]
+	vext.8	@t[4], @t[4], @t[4], #12
+	veor	@y[1], @y[1], @t[6]
+	veor	@y[0], @y[0], @t[7]
+	veor	@t[7], @t[7], @t[6]		@ clobber t[7]
+
+	veor	@y[3], @y[3], @t[0]
+	 veor	@y[1], @y[1], @y[0]
+	vext.8	@t[0], @t[0], @t[0], #12
+	veor	@y[2], @y[2], @t[1]
+	veor	@y[4], @y[4], @t[1]
+	vext.8	@t[1], @t[1], @t[1], #12
+	veor	@y[2], @y[2], @t[2]
+	veor	@y[3], @y[3], @t[2]
+	veor	@y[5], @y[5], @t[2]
+	veor	@y[2], @y[2], @t[7]
+	vext.8	@t[2], @t[2], @t[2], #12
+	veor	@y[3], @y[3], @t[3]
+	veor	@y[6], @y[6], @t[3]
+	veor	@y[4], @y[4], @t[3]
+	veor	@y[7], @y[7], @t[4]
+	vext.8	@t[3], @t[3], @t[3], #12
+	veor	@y[5], @y[5], @t[4]
+	veor	@y[7], @y[7], @t[7]
+	veor	@t[7], @t[7], @t[5]		@ clobber t[7] even more
+	veor	@y[3], @y[3], @t[5]
+	veor	@y[4], @y[4], @t[4]
+
+	veor	@y[5], @y[5], @t[7]
+	vext.8	@t[4], @t[4], @t[4], #12
+	veor	@y[6], @y[6], @t[7]
+	veor	@y[4], @y[4], @t[7]
+
+	veor	@t[7], @t[7], @t[5]
+	vext.8	@t[5], @t[5], @t[5], #12
+
+	@ multiplication by 0x0d
+	veor	@y[4], @y[4], @y[7]
+	 veor	@t[7], @t[7], @t[6]		@ restore t[7]
+	veor	@y[7], @y[7], @t[4]
+	vext.8	@t[6], @t[6], @t[6], #12
+	veor	@y[2], @y[2], @t[0]
+	veor	@y[7], @y[7], @t[5]
+	vext.8	@t[7], @t[7], @t[7], #12
+	veor	@y[2], @y[2], @t[2]
+
+	veor	@y[3], @y[3], @y[1]
+	veor	@y[1], @y[1], @t[1]
+	veor	@y[0], @y[0], @t[0]
+	veor	@y[3], @y[3], @t[0]
+	veor	@y[1], @y[1], @t[5]
+	veor	@y[0], @y[0], @t[5]
+	vext.8	@t[0], @t[0], @t[0], #12
+	veor	@y[1], @y[1], @t[7]
+	veor	@y[0], @y[0], @t[6]
+	veor	@y[3], @y[3], @y[1]
+	veor	@y[4], @y[4], @t[1]
+	vext.8	@t[1], @t[1], @t[1], #12
+
+	veor	@y[7], @y[7], @t[7]
+	veor	@y[4], @y[4], @t[2]
+	veor	@y[5], @y[5], @t[2]
+	veor	@y[2], @y[2], @t[6]
+	veor	@t[6], @t[6], @t[3]		@ clobber t[6]
+	vext.8	@t[2], @t[2], @t[2], #12
+	veor	@y[4], @y[4], @y[7]
+	veor	@y[3], @y[3], @t[6]
+
+	veor	@y[6], @y[6], @t[6]
+	veor	@y[5], @y[5], @t[5]
+	vext.8	@t[5], @t[5], @t[5], #12
+	veor	@y[6], @y[6], @t[4]
+	vext.8	@t[4], @t[4], @t[4], #12
+	veor	@y[5], @y[5], @t[6]
+	veor	@y[6], @y[6], @t[7]
+	vext.8	@t[7], @t[7], @t[7], #12
+	veor	@t[6], @t[6], @t[3]		@ restore t[6]
+	vext.8	@t[3], @t[3], @t[3], #12
+
+	@ multiplication by 0x09
+	veor	@y[4], @y[4], @y[1]
+	veor	@t[1], @t[1], @y[1]		@ t[1]=y[1]
+	veor	@t[0], @t[0], @t[5]		@ clobber t[0]
+	vext.8	@t[6], @t[6], @t[6], #12
+	veor	@t[1], @t[1], @t[5]
+	veor	@y[3], @y[3], @t[0]
+	veor	@t[0], @t[0], @y[0]		@ t[0]=y[0]
+	veor	@t[1], @t[1], @t[6]
+	veor	@t[6], @t[6], @t[7]		@ clobber t[6]
+	veor	@y[4], @y[4], @t[1]
+	veor	@y[7], @y[7], @t[4]
+	veor	@y[6], @y[6], @t[3]
+	veor	@y[5], @y[5], @t[2]
+	veor	@t[4], @t[4], @y[4]		@ t[4]=y[4]
+	veor	@t[3], @t[3], @y[3]		@ t[3]=y[3]
+	veor	@t[5], @t[5], @y[5]		@ t[5]=y[5]
+	veor	@t[2], @t[2], @y[2]		@ t[2]=y[2]
+	veor	@t[3], @t[3], @t[7]
+	veor	@XMM[5], @t[5], @t[6]
+	veor	@XMM[6], @t[6], @y[6]		@ t[6]=y[6]
+	veor	@XMM[2], @t[2], @t[6]
+	veor	@XMM[7], @t[7], @y[7]		@ t[7]=y[7]
+
+	vmov	@XMM[0], @t[0]
+	vmov	@XMM[1], @t[1]
+	@ vmov	@XMM[2], @t[2]
+	vmov	@XMM[3], @t[3]
+	vmov	@XMM[4], @t[4]
+	@ vmov	@XMM[5], @t[5]
+	@ vmov	@XMM[6], @t[6]
+	@ vmov	@XMM[7], @t[7]
+___
+}
+
+sub InvMixColumns {
+my @x=@_[0..7];
+my @t=@_[8..15];
+
+# Thanks to Jussi Kivilinna for providing pointer to
+#
+# | 0e 0b 0d 09 |   | 02 03 01 01 |   | 05 00 04 00 |
+# | 09 0e 0b 0d | = | 01 02 03 01 | x | 00 05 00 04 |
+# | 0d 09 0e 0b |   | 01 01 02 03 |   | 04 00 05 00 |
+# | 0b 0d 09 0e |   | 03 01 01 02 |   | 00 04 00 05 |
+
+$code.=<<___;
+	@ multiplication by 0x05-0x00-0x04-0x00
+	vext.8	@t[0], @x[0], @x[0], #8
+	vext.8	@t[6], @x[6], @x[6], #8
+	vext.8	@t[7], @x[7], @x[7], #8
+	veor	@t[0], @t[0], @x[0]
+	vext.8	@t[1], @x[1], @x[1], #8
+	veor	@t[6], @t[6], @x[6]
+	vext.8	@t[2], @x[2], @x[2], #8
+	veor	@t[7], @t[7], @x[7]
+	vext.8	@t[3], @x[3], @x[3], #8
+	veor	@t[1], @t[1], @x[1]
+	vext.8	@t[4], @x[4], @x[4], #8
+	veor	@t[2], @t[2], @x[2]
+	vext.8	@t[5], @x[5], @x[5], #8
+	veor	@t[3], @t[3], @x[3]
+	veor	@t[4], @t[4], @x[4]
+	veor	@t[5], @t[5], @x[5]
+
+	 veor	@x[0], @x[0], @t[6]
+	 veor	@x[1], @x[1], @t[6]
+	 veor	@x[2], @x[2], @t[0]
+	 veor	@x[4], @x[4], @t[2]
+	 veor	@x[3], @x[3], @t[1]
+	 veor	@x[1], @x[1], @t[7]
+	 veor	@x[2], @x[2], @t[7]
+	 veor	@x[4], @x[4], @t[6]
+	 veor	@x[5], @x[5], @t[3]
+	 veor	@x[3], @x[3], @t[6]
+	 veor	@x[6], @x[6], @t[4]
+	 veor	@x[4], @x[4], @t[7]
+	 veor	@x[5], @x[5], @t[7]
+	 veor	@x[7], @x[7], @t[5]
+___
+	&MixColumns	(@x,@t,1);	# flipped 2<->3 and 4<->6
+}
+
+sub swapmove {
+my ($a,$b,$n,$mask,$t)=@_;
+$code.=<<___;
+	vshr.u64	$t, $b, #$n
+	veor		$t, $t, $a
+	vand		$t, $t, $mask
+	veor		$a, $a, $t
+	vshl.u64	$t, $t, #$n
+	veor		$b, $b, $t
+___
+}
+sub swapmove2x {
+my ($a0,$b0,$a1,$b1,$n,$mask,$t0,$t1)=@_;
+$code.=<<___;
+	vshr.u64	$t0, $b0, #$n
+	 vshr.u64	$t1, $b1, #$n
+	veor		$t0, $t0, $a0
+	 veor		$t1, $t1, $a1
+	vand		$t0, $t0, $mask
+	 vand		$t1, $t1, $mask
+	veor		$a0, $a0, $t0
+	vshl.u64	$t0, $t0, #$n
+	 veor		$a1, $a1, $t1
+	 vshl.u64	$t1, $t1, #$n
+	veor		$b0, $b0, $t0
+	 veor		$b1, $b1, $t1
+___
+}
+
+sub bitslice {
+my @x=reverse(@_[0..7]);
+my ($t0,$t1,$t2,$t3)=@_[8..11];
+$code.=<<___;
+	vmov.i8	$t0,#0x55			@ compose .LBS0
+	vmov.i8	$t1,#0x33			@ compose .LBS1
+___
+	&swapmove2x(@x[0,1,2,3],1,$t0,$t2,$t3);
+	&swapmove2x(@x[4,5,6,7],1,$t0,$t2,$t3);
+$code.=<<___;
+	vmov.i8	$t0,#0x0f			@ compose .LBS2
+___
+	&swapmove2x(@x[0,2,1,3],2,$t1,$t2,$t3);
+	&swapmove2x(@x[4,6,5,7],2,$t1,$t2,$t3);
+
+	&swapmove2x(@x[0,4,1,5],4,$t0,$t2,$t3);
+	&swapmove2x(@x[2,6,3,7],4,$t0,$t2,$t3);
+}
+
+$code.=<<___;
+#ifndef __KERNEL__
+# include "arm_arch.h"
+
+# define VFP_ABI_PUSH	vstmdb	sp!,{d8-d15}
+# define VFP_ABI_POP	vldmia	sp!,{d8-d15}
+# define VFP_ABI_FRAME	0x40
+#else
+# define VFP_ABI_PUSH
+# define VFP_ABI_POP
+# define VFP_ABI_FRAME	0
+# define BSAES_ASM_EXTENDED_KEY
+# define XTS_CHAIN_TWEAK
+# define __ARM_ARCH__ __LINUX_ARM_ARCH__
+#endif
+
+#ifdef __thumb__
+# define adrl adr
+#endif
+
+#if __ARM_ARCH__>=7
+.text
+.syntax	unified 	@ ARMv7-capable assembler is expected to handle this
+#ifdef __thumb2__
+.thumb
+#else
+.code   32
+#endif
+
+.fpu	neon
+
+.type	_bsaes_decrypt8,%function
+.align	4
+_bsaes_decrypt8:
+	adr	$const,_bsaes_decrypt8
+	vldmia	$key!, {@XMM[9]}		@ round 0 key
+	add	$const,$const,#.LM0ISR-_bsaes_decrypt8
+
+	vldmia	$const!, {@XMM[8]}		@ .LM0ISR
+	veor	@XMM[10], @XMM[0], @XMM[9]	@ xor with round0 key
+	veor	@XMM[11], @XMM[1], @XMM[9]
+	 vtbl.8	`&Dlo(@XMM[0])`, {@XMM[10]}, `&Dlo(@XMM[8])`
+	 vtbl.8	`&Dhi(@XMM[0])`, {@XMM[10]}, `&Dhi(@XMM[8])`
+	veor	@XMM[12], @XMM[2], @XMM[9]
+	 vtbl.8	`&Dlo(@XMM[1])`, {@XMM[11]}, `&Dlo(@XMM[8])`
+	 vtbl.8	`&Dhi(@XMM[1])`, {@XMM[11]}, `&Dhi(@XMM[8])`
+	veor	@XMM[13], @XMM[3], @XMM[9]
+	 vtbl.8	`&Dlo(@XMM[2])`, {@XMM[12]}, `&Dlo(@XMM[8])`
+	 vtbl.8	`&Dhi(@XMM[2])`, {@XMM[12]}, `&Dhi(@XMM[8])`
+	veor	@XMM[14], @XMM[4], @XMM[9]
+	 vtbl.8	`&Dlo(@XMM[3])`, {@XMM[13]}, `&Dlo(@XMM[8])`
+	 vtbl.8	`&Dhi(@XMM[3])`, {@XMM[13]}, `&Dhi(@XMM[8])`
+	veor	@XMM[15], @XMM[5], @XMM[9]
+	 vtbl.8	`&Dlo(@XMM[4])`, {@XMM[14]}, `&Dlo(@XMM[8])`
+	 vtbl.8	`&Dhi(@XMM[4])`, {@XMM[14]}, `&Dhi(@XMM[8])`
+	veor	@XMM[10], @XMM[6], @XMM[9]
+	 vtbl.8	`&Dlo(@XMM[5])`, {@XMM[15]}, `&Dlo(@XMM[8])`
+	 vtbl.8	`&Dhi(@XMM[5])`, {@XMM[15]}, `&Dhi(@XMM[8])`
+	veor	@XMM[11], @XMM[7], @XMM[9]
+	 vtbl.8	`&Dlo(@XMM[6])`, {@XMM[10]}, `&Dlo(@XMM[8])`
+	 vtbl.8	`&Dhi(@XMM[6])`, {@XMM[10]}, `&Dhi(@XMM[8])`
+	 vtbl.8	`&Dlo(@XMM[7])`, {@XMM[11]}, `&Dlo(@XMM[8])`
+	 vtbl.8	`&Dhi(@XMM[7])`, {@XMM[11]}, `&Dhi(@XMM[8])`
+___
+	&bitslice	(@XMM[0..7, 8..11]);
+$code.=<<___;
+	sub	$rounds,$rounds,#1
+	b	.Ldec_sbox
+.align	4
+.Ldec_loop:
+___
+	&ShiftRows	(@XMM[0..7, 8..12]);
+$code.=".Ldec_sbox:\n";
+	&InvSbox	(@XMM[0..7, 8..15]);
+$code.=<<___;
+	subs	$rounds,$rounds,#1
+	bcc	.Ldec_done
+___
+	&InvMixColumns	(@XMM[0,1,6,4,2,7,3,5, 8..15]);
+$code.=<<___;
+	vldmia	$const, {@XMM[12]}		@ .LISR
+	ite	eq				@ Thumb2 thing, sanity check in ARM
+	addeq	$const,$const,#0x10
+	bne	.Ldec_loop
+	vldmia	$const, {@XMM[12]}		@ .LISRM0
+	b	.Ldec_loop
+.align	4
+.Ldec_done:
+___
+	&bitslice	(@XMM[0,1,6,4,2,7,3,5, 8..11]);
+$code.=<<___;
+	vldmia	$key, {@XMM[8]}			@ last round key
+	veor	@XMM[6], @XMM[6], @XMM[8]
+	veor	@XMM[4], @XMM[4], @XMM[8]
+	veor	@XMM[2], @XMM[2], @XMM[8]
+	veor	@XMM[7], @XMM[7], @XMM[8]
+	veor	@XMM[3], @XMM[3], @XMM[8]
+	veor	@XMM[5], @XMM[5], @XMM[8]
+	veor	@XMM[0], @XMM[0], @XMM[8]
+	veor	@XMM[1], @XMM[1], @XMM[8]
+	bx	lr
+.size	_bsaes_decrypt8,.-_bsaes_decrypt8
+
+.type	_bsaes_const,%object
+.align	6
+_bsaes_const:
+.LM0ISR:	@ InvShiftRows constants
+	.quad	0x0a0e0206070b0f03, 0x0004080c0d010509
+.LISR:
+	.quad	0x0504070602010003, 0x0f0e0d0c080b0a09
+.LISRM0:
+	.quad	0x01040b0e0205080f, 0x0306090c00070a0d
+.LM0SR:		@ ShiftRows constants
+	.quad	0x0a0e02060f03070b, 0x0004080c05090d01
+.LSR:
+	.quad	0x0504070600030201, 0x0f0e0d0c0a09080b
+.LSRM0:
+	.quad	0x0304090e00050a0f, 0x01060b0c0207080d
+.LM0:
+	.quad	0x02060a0e03070b0f, 0x0004080c0105090d
+.LREVM0SR:
+	.quad	0x090d01050c000408, 0x03070b0f060a0e02
+.asciz	"Bit-sliced AES for NEON, CRYPTOGAMS by <appro\@openssl.org>"
+.align	6
+.size	_bsaes_const,.-_bsaes_const
+
+.type	_bsaes_encrypt8,%function
+.align	4
+_bsaes_encrypt8:
+	adr	$const,_bsaes_encrypt8
+	vldmia	$key!, {@XMM[9]}		@ round 0 key
+	sub	$const,$const,#_bsaes_encrypt8-.LM0SR
+
+	vldmia	$const!, {@XMM[8]}		@ .LM0SR
+_bsaes_encrypt8_alt:
+	veor	@XMM[10], @XMM[0], @XMM[9]	@ xor with round0 key
+	veor	@XMM[11], @XMM[1], @XMM[9]
+	 vtbl.8	`&Dlo(@XMM[0])`, {@XMM[10]}, `&Dlo(@XMM[8])`
+	 vtbl.8	`&Dhi(@XMM[0])`, {@XMM[10]}, `&Dhi(@XMM[8])`
+	veor	@XMM[12], @XMM[2], @XMM[9]
+	 vtbl.8	`&Dlo(@XMM[1])`, {@XMM[11]}, `&Dlo(@XMM[8])`
+	 vtbl.8	`&Dhi(@XMM[1])`, {@XMM[11]}, `&Dhi(@XMM[8])`
+	veor	@XMM[13], @XMM[3], @XMM[9]
+	 vtbl.8	`&Dlo(@XMM[2])`, {@XMM[12]}, `&Dlo(@XMM[8])`
+	 vtbl.8	`&Dhi(@XMM[2])`, {@XMM[12]}, `&Dhi(@XMM[8])`
+	veor	@XMM[14], @XMM[4], @XMM[9]
+	 vtbl.8	`&Dlo(@XMM[3])`, {@XMM[13]}, `&Dlo(@XMM[8])`
+	 vtbl.8	`&Dhi(@XMM[3])`, {@XMM[13]}, `&Dhi(@XMM[8])`
+	veor	@XMM[15], @XMM[5], @XMM[9]
+	 vtbl.8	`&Dlo(@XMM[4])`, {@XMM[14]}, `&Dlo(@XMM[8])`
+	 vtbl.8	`&Dhi(@XMM[4])`, {@XMM[14]}, `&Dhi(@XMM[8])`
+	veor	@XMM[10], @XMM[6], @XMM[9]
+	 vtbl.8	`&Dlo(@XMM[5])`, {@XMM[15]}, `&Dlo(@XMM[8])`
+	 vtbl.8	`&Dhi(@XMM[5])`, {@XMM[15]}, `&Dhi(@XMM[8])`
+	veor	@XMM[11], @XMM[7], @XMM[9]
+	 vtbl.8	`&Dlo(@XMM[6])`, {@XMM[10]}, `&Dlo(@XMM[8])`
+	 vtbl.8	`&Dhi(@XMM[6])`, {@XMM[10]}, `&Dhi(@XMM[8])`
+	 vtbl.8	`&Dlo(@XMM[7])`, {@XMM[11]}, `&Dlo(@XMM[8])`
+	 vtbl.8	`&Dhi(@XMM[7])`, {@XMM[11]}, `&Dhi(@XMM[8])`
+_bsaes_encrypt8_bitslice:
+___
+	&bitslice	(@XMM[0..7, 8..11]);
+$code.=<<___;
+	sub	$rounds,$rounds,#1
+	b	.Lenc_sbox
+.align	4
+.Lenc_loop:
+___
+	&ShiftRows	(@XMM[0..7, 8..12]);
+$code.=".Lenc_sbox:\n";
+	&Sbox		(@XMM[0..7, 8..15]);
+$code.=<<___;
+	subs	$rounds,$rounds,#1
+	bcc	.Lenc_done
+___
+	&MixColumns	(@XMM[0,1,4,6,3,7,2,5, 8..15]);
+$code.=<<___;
+	vldmia	$const, {@XMM[12]}		@ .LSR
+	ite	eq				@ Thumb2 thing, samity check in ARM
+	addeq	$const,$const,#0x10
+	bne	.Lenc_loop
+	vldmia	$const, {@XMM[12]}		@ .LSRM0
+	b	.Lenc_loop
+.align	4
+.Lenc_done:
+___
+	# output in lsb > [t0, t1, t4, t6, t3, t7, t2, t5] < msb
+	&bitslice	(@XMM[0,1,4,6,3,7,2,5, 8..11]);
+$code.=<<___;
+	vldmia	$key, {@XMM[8]}			@ last round key
+	veor	@XMM[4], @XMM[4], @XMM[8]
+	veor	@XMM[6], @XMM[6], @XMM[8]
+	veor	@XMM[3], @XMM[3], @XMM[8]
+	veor	@XMM[7], @XMM[7], @XMM[8]
+	veor	@XMM[2], @XMM[2], @XMM[8]
+	veor	@XMM[5], @XMM[5], @XMM[8]
+	veor	@XMM[0], @XMM[0], @XMM[8]
+	veor	@XMM[1], @XMM[1], @XMM[8]
+	bx	lr
+.size	_bsaes_encrypt8,.-_bsaes_encrypt8
+___
+}
+{
+my ($out,$inp,$rounds,$const)=("r12","r4","r5","r6");
+
+sub bitslice_key {
+my @x=reverse(@_[0..7]);
+my ($bs0,$bs1,$bs2,$t2,$t3)=@_[8..12];
+
+	&swapmove	(@x[0,1],1,$bs0,$t2,$t3);
+$code.=<<___;
+	@ &swapmove(@x[2,3],1,$t0,$t2,$t3);
+	vmov	@x[2], @x[0]
+	vmov	@x[3], @x[1]
+___
+	#&swapmove2x(@x[4,5,6,7],1,$t0,$t2,$t3);
+
+	&swapmove2x	(@x[0,2,1,3],2,$bs1,$t2,$t3);
+$code.=<<___;
+	@ &swapmove2x(@x[4,6,5,7],2,$t1,$t2,$t3);
+	vmov	@x[4], @x[0]
+	vmov	@x[6], @x[2]
+	vmov	@x[5], @x[1]
+	vmov	@x[7], @x[3]
+___
+	&swapmove2x	(@x[0,4,1,5],4,$bs2,$t2,$t3);
+	&swapmove2x	(@x[2,6,3,7],4,$bs2,$t2,$t3);
+}
+
+$code.=<<___;
+.type	_bsaes_key_convert,%function
+.align	4
+_bsaes_key_convert:
+	adr	$const,_bsaes_key_convert
+	vld1.8	{@XMM[7]},  [$inp]!		@ load round 0 key
+	sub	$const,$const,#_bsaes_key_convert-.LM0
+	vld1.8	{@XMM[15]}, [$inp]!		@ load round 1 key
+
+	vmov.i8	@XMM[8],  #0x01			@ bit masks
+	vmov.i8	@XMM[9],  #0x02
+	vmov.i8	@XMM[10], #0x04
+	vmov.i8	@XMM[11], #0x08
+	vmov.i8	@XMM[12], #0x10
+	vmov.i8	@XMM[13], #0x20
+	vldmia	$const, {@XMM[14]}		@ .LM0
+
+#ifdef __ARMEL__
+	vrev32.8	@XMM[7],  @XMM[7]
+	vrev32.8	@XMM[15], @XMM[15]
+#endif
+	sub	$rounds,$rounds,#1
+	vstmia	$out!, {@XMM[7]}		@ save round 0 key
+	b	.Lkey_loop
+
+.align	4
+.Lkey_loop:
+	vtbl.8	`&Dlo(@XMM[7])`,{@XMM[15]},`&Dlo(@XMM[14])`
+	vtbl.8	`&Dhi(@XMM[7])`,{@XMM[15]},`&Dhi(@XMM[14])`
+	vmov.i8	@XMM[6],  #0x40
+	vmov.i8	@XMM[15], #0x80
+
+	vtst.8	@XMM[0], @XMM[7], @XMM[8]
+	vtst.8	@XMM[1], @XMM[7], @XMM[9]
+	vtst.8	@XMM[2], @XMM[7], @XMM[10]
+	vtst.8	@XMM[3], @XMM[7], @XMM[11]
+	vtst.8	@XMM[4], @XMM[7], @XMM[12]
+	vtst.8	@XMM[5], @XMM[7], @XMM[13]
+	vtst.8	@XMM[6], @XMM[7], @XMM[6]
+	vtst.8	@XMM[7], @XMM[7], @XMM[15]
+	vld1.8	{@XMM[15]}, [$inp]!		@ load next round key
+	vmvn	@XMM[0], @XMM[0]		@ "pnot"
+	vmvn	@XMM[1], @XMM[1]
+	vmvn	@XMM[5], @XMM[5]
+	vmvn	@XMM[6], @XMM[6]
+#ifdef __ARMEL__
+	vrev32.8	@XMM[15], @XMM[15]
+#endif
+	subs	$rounds,$rounds,#1
+	vstmia	$out!,{@XMM[0]-@XMM[7]}		@ write bit-sliced round key
+	bne	.Lkey_loop
+
+	vmov.i8	@XMM[7],#0x63			@ compose .L63
+	@ don't save last round key
+	bx	lr
+.size	_bsaes_key_convert,.-_bsaes_key_convert
+___
+}
+
+if (0) {		# following four functions are unsupported interface
+			# used for benchmarking...
+$code.=<<___;
+.globl	bsaes_enc_key_convert
+.type	bsaes_enc_key_convert,%function
+.align	4
+bsaes_enc_key_convert:
+	stmdb	sp!,{r4-r6,lr}
+	vstmdb	sp!,{d8-d15}		@ ABI specification says so
+
+	ldr	r5,[$inp,#240]			@ pass rounds
+	mov	r4,$inp				@ pass key
+	mov	r12,$out			@ pass key schedule
+	bl	_bsaes_key_convert
+	veor	@XMM[7],@XMM[7],@XMM[15]	@ fix up last round key
+	vstmia	r12, {@XMM[7]}			@ save last round key
+
+	vldmia	sp!,{d8-d15}
+	ldmia	sp!,{r4-r6,pc}
+.size	bsaes_enc_key_convert,.-bsaes_enc_key_convert
+
+.globl	bsaes_encrypt_128
+.type	bsaes_encrypt_128,%function
+.align	4
+bsaes_encrypt_128:
+	stmdb	sp!,{r4-r6,lr}
+	vstmdb	sp!,{d8-d15}		@ ABI specification says so
+.Lenc128_loop:
+	vld1.8	{@XMM[0]-@XMM[1]}, [$inp]!	@ load input
+	vld1.8	{@XMM[2]-@XMM[3]}, [$inp]!
+	mov	r4,$key				@ pass the key
+	vld1.8	{@XMM[4]-@XMM[5]}, [$inp]!
+	mov	r5,#10				@ pass rounds
+	vld1.8	{@XMM[6]-@XMM[7]}, [$inp]!
+
+	bl	_bsaes_encrypt8
+
+	vst1.8	{@XMM[0]-@XMM[1]}, [$out]!	@ write output
+	vst1.8	{@XMM[4]}, [$out]!
+	vst1.8	{@XMM[6]}, [$out]!
+	vst1.8	{@XMM[3]}, [$out]!
+	vst1.8	{@XMM[7]}, [$out]!
+	vst1.8	{@XMM[2]}, [$out]!
+	subs	$len,$len,#0x80
+	vst1.8	{@XMM[5]}, [$out]!
+	bhi	.Lenc128_loop
+
+	vldmia	sp!,{d8-d15}
+	ldmia	sp!,{r4-r6,pc}
+.size	bsaes_encrypt_128,.-bsaes_encrypt_128
+
+.globl	bsaes_dec_key_convert
+.type	bsaes_dec_key_convert,%function
+.align	4
+bsaes_dec_key_convert:
+	stmdb	sp!,{r4-r6,lr}
+	vstmdb	sp!,{d8-d15}		@ ABI specification says so
+
+	ldr	r5,[$inp,#240]			@ pass rounds
+	mov	r4,$inp				@ pass key
+	mov	r12,$out			@ pass key schedule
+	bl	_bsaes_key_convert
+	vldmia	$out, {@XMM[6]}
+	vstmia	r12,  {@XMM[15]}		@ save last round key
+	veor	@XMM[7], @XMM[7], @XMM[6]	@ fix up round 0 key
+	vstmia	$out, {@XMM[7]}
+
+	vldmia	sp!,{d8-d15}
+	ldmia	sp!,{r4-r6,pc}
+.size	bsaes_dec_key_convert,.-bsaes_dec_key_convert
+
+.globl	bsaes_decrypt_128
+.type	bsaes_decrypt_128,%function
+.align	4
+bsaes_decrypt_128:
+	stmdb	sp!,{r4-r6,lr}
+	vstmdb	sp!,{d8-d15}		@ ABI specification says so
+.Ldec128_loop:
+	vld1.8	{@XMM[0]-@XMM[1]}, [$inp]!	@ load input
+	vld1.8	{@XMM[2]-@XMM[3]}, [$inp]!
+	mov	r4,$key				@ pass the key
+	vld1.8	{@XMM[4]-@XMM[5]}, [$inp]!
+	mov	r5,#10				@ pass rounds
+	vld1.8	{@XMM[6]-@XMM[7]}, [$inp]!
+
+	bl	_bsaes_decrypt8
+
+	vst1.8	{@XMM[0]-@XMM[1]}, [$out]!	@ write output
+	vst1.8	{@XMM[6]}, [$out]!
+	vst1.8	{@XMM[4]}, [$out]!
+	vst1.8	{@XMM[2]}, [$out]!
+	vst1.8	{@XMM[7]}, [$out]!
+	vst1.8	{@XMM[3]}, [$out]!
+	subs	$len,$len,#0x80
+	vst1.8	{@XMM[5]}, [$out]!
+	bhi	.Ldec128_loop
+
+	vldmia	sp!,{d8-d15}
+	ldmia	sp!,{r4-r6,pc}
+.size	bsaes_decrypt_128,.-bsaes_decrypt_128
+___
+}
+{
+my ($inp,$out,$len,$key, $ivp,$fp,$rounds)=map("r$_",(0..3,8..10));
+my ($keysched)=("sp");
+
+$code.=<<___;
+.extern AES_cbc_encrypt
+.extern AES_decrypt
+
+.global	bsaes_cbc_encrypt
+.type	bsaes_cbc_encrypt,%function
+.align	5
+bsaes_cbc_encrypt:
+#ifndef	__KERNEL__
+	cmp	$len, #128
+#ifndef	__thumb__
+	blo	AES_cbc_encrypt
+#else
+	bhs	1f
+	b	AES_cbc_encrypt
+1:
+#endif
+#endif
+
+	@ it is up to the caller to make sure we are called with enc == 0
+
+	mov	ip, sp
+	stmdb	sp!, {r4-r10, lr}
+	VFP_ABI_PUSH
+	ldr	$ivp, [ip]			@ IV is 1st arg on the stack
+	mov	$len, $len, lsr#4		@ len in 16 byte blocks
+	sub	sp, #0x10			@ scratch space to carry over the IV
+	mov	$fp, sp				@ save sp
+
+	ldr	$rounds, [$key, #240]		@ get # of rounds
+#ifndef	BSAES_ASM_EXTENDED_KEY
+	@ allocate the key schedule on the stack
+	sub	r12, sp, $rounds, lsl#7		@ 128 bytes per inner round key
+	add	r12, #`128-32`			@ sifze of bit-slices key schedule
+
+	@ populate the key schedule
+	mov	r4, $key			@ pass key
+	mov	r5, $rounds			@ pass # of rounds
+	mov	sp, r12				@ sp is $keysched
+	bl	_bsaes_key_convert
+	vldmia	$keysched, {@XMM[6]}
+	vstmia	r12,  {@XMM[15]}		@ save last round key
+	veor	@XMM[7], @XMM[7], @XMM[6]	@ fix up round 0 key
+	vstmia	$keysched, {@XMM[7]}
+#else
+	ldr	r12, [$key, #244]
+	eors	r12, #1
+	beq	0f
+
+	@ populate the key schedule
+	str	r12, [$key, #244]
+	mov	r4, $key			@ pass key
+	mov	r5, $rounds			@ pass # of rounds
+	add	r12, $key, #248			@ pass key schedule
+	bl	_bsaes_key_convert
+	add	r4, $key, #248
+	vldmia	r4, {@XMM[6]}
+	vstmia	r12, {@XMM[15]}			@ save last round key
+	veor	@XMM[7], @XMM[7], @XMM[6]	@ fix up round 0 key
+	vstmia	r4, {@XMM[7]}
+
+.align	2
+0:
+#endif
+
+	vld1.8	{@XMM[15]}, [$ivp]		@ load IV
+	b	.Lcbc_dec_loop
+
+.align	4
+.Lcbc_dec_loop:
+	subs	$len, $len, #0x8
+	bmi	.Lcbc_dec_loop_finish
+
+	vld1.8	{@XMM[0]-@XMM[1]}, [$inp]!	@ load input
+	vld1.8	{@XMM[2]-@XMM[3]}, [$inp]!
+#ifndef	BSAES_ASM_EXTENDED_KEY
+	mov	r4, $keysched			@ pass the key
+#else
+	add	r4, $key, #248
+#endif
+	vld1.8	{@XMM[4]-@XMM[5]}, [$inp]!
+	mov	r5, $rounds
+	vld1.8	{@XMM[6]-@XMM[7]}, [$inp]
+	sub	$inp, $inp, #0x60
+	vstmia	$fp, {@XMM[15]}			@ put aside IV
+
+	bl	_bsaes_decrypt8
+
+	vldmia	$fp, {@XMM[14]}			@ reload IV
+	vld1.8	{@XMM[8]-@XMM[9]}, [$inp]!	@ reload input
+	veor	@XMM[0], @XMM[0], @XMM[14]	@ ^= IV
+	vld1.8	{@XMM[10]-@XMM[11]}, [$inp]!
+	veor	@XMM[1], @XMM[1], @XMM[8]
+	veor	@XMM[6], @XMM[6], @XMM[9]
+	vld1.8	{@XMM[12]-@XMM[13]}, [$inp]!
+	veor	@XMM[4], @XMM[4], @XMM[10]
+	veor	@XMM[2], @XMM[2], @XMM[11]
+	vld1.8	{@XMM[14]-@XMM[15]}, [$inp]!
+	veor	@XMM[7], @XMM[7], @XMM[12]
+	vst1.8	{@XMM[0]-@XMM[1]}, [$out]!	@ write output
+	veor	@XMM[3], @XMM[3], @XMM[13]
+	vst1.8	{@XMM[6]}, [$out]!
+	veor	@XMM[5], @XMM[5], @XMM[14]
+	vst1.8	{@XMM[4]}, [$out]!
+	vst1.8	{@XMM[2]}, [$out]!
+	vst1.8	{@XMM[7]}, [$out]!
+	vst1.8	{@XMM[3]}, [$out]!
+	vst1.8	{@XMM[5]}, [$out]!
+
+	b	.Lcbc_dec_loop
+
+.Lcbc_dec_loop_finish:
+	adds	$len, $len, #8
+	beq	.Lcbc_dec_done
+
+	vld1.8	{@XMM[0]}, [$inp]!		@ load input
+	cmp	$len, #2
+	blo	.Lcbc_dec_one
+	vld1.8	{@XMM[1]}, [$inp]!
+#ifndef	BSAES_ASM_EXTENDED_KEY
+	mov	r4, $keysched			@ pass the key
+#else
+	add	r4, $key, #248
+#endif
+	mov	r5, $rounds
+	vstmia	$fp, {@XMM[15]}			@ put aside IV
+	beq	.Lcbc_dec_two
+	vld1.8	{@XMM[2]}, [$inp]!
+	cmp	$len, #4
+	blo	.Lcbc_dec_three
+	vld1.8	{@XMM[3]}, [$inp]!
+	beq	.Lcbc_dec_four
+	vld1.8	{@XMM[4]}, [$inp]!
+	cmp	$len, #6
+	blo	.Lcbc_dec_five
+	vld1.8	{@XMM[5]}, [$inp]!
+	beq	.Lcbc_dec_six
+	vld1.8	{@XMM[6]}, [$inp]!
+	sub	$inp, $inp, #0x70
+
+	bl	_bsaes_decrypt8
+
+	vldmia	$fp, {@XMM[14]}			@ reload IV
+	vld1.8	{@XMM[8]-@XMM[9]}, [$inp]!	@ reload input
+	veor	@XMM[0], @XMM[0], @XMM[14]	@ ^= IV
+	vld1.8	{@XMM[10]-@XMM[11]}, [$inp]!
+	veor	@XMM[1], @XMM[1], @XMM[8]
+	veor	@XMM[6], @XMM[6], @XMM[9]
+	vld1.8	{@XMM[12]-@XMM[13]}, [$inp]!
+	veor	@XMM[4], @XMM[4], @XMM[10]
+	veor	@XMM[2], @XMM[2], @XMM[11]
+	vld1.8	{@XMM[15]}, [$inp]!
+	veor	@XMM[7], @XMM[7], @XMM[12]
+	vst1.8	{@XMM[0]-@XMM[1]}, [$out]!	@ write output
+	veor	@XMM[3], @XMM[3], @XMM[13]
+	vst1.8	{@XMM[6]}, [$out]!
+	vst1.8	{@XMM[4]}, [$out]!
+	vst1.8	{@XMM[2]}, [$out]!
+	vst1.8	{@XMM[7]}, [$out]!
+	vst1.8	{@XMM[3]}, [$out]!
+	b	.Lcbc_dec_done
+.align	4
+.Lcbc_dec_six:
+	sub	$inp, $inp, #0x60
+	bl	_bsaes_decrypt8
+	vldmia	$fp,{@XMM[14]}			@ reload IV
+	vld1.8	{@XMM[8]-@XMM[9]}, [$inp]!	@ reload input
+	veor	@XMM[0], @XMM[0], @XMM[14]	@ ^= IV
+	vld1.8	{@XMM[10]-@XMM[11]}, [$inp]!
+	veor	@XMM[1], @XMM[1], @XMM[8]
+	veor	@XMM[6], @XMM[6], @XMM[9]
+	vld1.8	{@XMM[12]}, [$inp]!
+	veor	@XMM[4], @XMM[4], @XMM[10]
+	veor	@XMM[2], @XMM[2], @XMM[11]
+	vld1.8	{@XMM[15]}, [$inp]!
+	veor	@XMM[7], @XMM[7], @XMM[12]
+	vst1.8	{@XMM[0]-@XMM[1]}, [$out]!	@ write output
+	vst1.8	{@XMM[6]}, [$out]!
+	vst1.8	{@XMM[4]}, [$out]!
+	vst1.8	{@XMM[2]}, [$out]!
+	vst1.8	{@XMM[7]}, [$out]!
+	b	.Lcbc_dec_done
+.align	4
+.Lcbc_dec_five:
+	sub	$inp, $inp, #0x50
+	bl	_bsaes_decrypt8
+	vldmia	$fp, {@XMM[14]}			@ reload IV
+	vld1.8	{@XMM[8]-@XMM[9]}, [$inp]!	@ reload input
+	veor	@XMM[0], @XMM[0], @XMM[14]	@ ^= IV
+	vld1.8	{@XMM[10]-@XMM[11]}, [$inp]!
+	veor	@XMM[1], @XMM[1], @XMM[8]
+	veor	@XMM[6], @XMM[6], @XMM[9]
+	vld1.8	{@XMM[15]}, [$inp]!
+	veor	@XMM[4], @XMM[4], @XMM[10]
+	vst1.8	{@XMM[0]-@XMM[1]}, [$out]!	@ write output
+	veor	@XMM[2], @XMM[2], @XMM[11]
+	vst1.8	{@XMM[6]}, [$out]!
+	vst1.8	{@XMM[4]}, [$out]!
+	vst1.8	{@XMM[2]}, [$out]!
+	b	.Lcbc_dec_done
+.align	4
+.Lcbc_dec_four:
+	sub	$inp, $inp, #0x40
+	bl	_bsaes_decrypt8
+	vldmia	$fp, {@XMM[14]}			@ reload IV
+	vld1.8	{@XMM[8]-@XMM[9]}, [$inp]!	@ reload input
+	veor	@XMM[0], @XMM[0], @XMM[14]	@ ^= IV
+	vld1.8	{@XMM[10]}, [$inp]!
+	veor	@XMM[1], @XMM[1], @XMM[8]
+	veor	@XMM[6], @XMM[6], @XMM[9]
+	vld1.8	{@XMM[15]}, [$inp]!
+	veor	@XMM[4], @XMM[4], @XMM[10]
+	vst1.8	{@XMM[0]-@XMM[1]}, [$out]!	@ write output
+	vst1.8	{@XMM[6]}, [$out]!
+	vst1.8	{@XMM[4]}, [$out]!
+	b	.Lcbc_dec_done
+.align	4
+.Lcbc_dec_three:
+	sub	$inp, $inp, #0x30
+	bl	_bsaes_decrypt8
+	vldmia	$fp, {@XMM[14]}			@ reload IV
+	vld1.8	{@XMM[8]-@XMM[9]}, [$inp]!	@ reload input
+	veor	@XMM[0], @XMM[0], @XMM[14]	@ ^= IV
+	vld1.8	{@XMM[15]}, [$inp]!
+	veor	@XMM[1], @XMM[1], @XMM[8]
+	veor	@XMM[6], @XMM[6], @XMM[9]
+	vst1.8	{@XMM[0]-@XMM[1]}, [$out]!	@ write output
+	vst1.8	{@XMM[6]}, [$out]!
+	b	.Lcbc_dec_done
+.align	4
+.Lcbc_dec_two:
+	sub	$inp, $inp, #0x20
+	bl	_bsaes_decrypt8
+	vldmia	$fp, {@XMM[14]}			@ reload IV
+	vld1.8	{@XMM[8]}, [$inp]!		@ reload input
+	veor	@XMM[0], @XMM[0], @XMM[14]	@ ^= IV
+	vld1.8	{@XMM[15]}, [$inp]!		@ reload input
+	veor	@XMM[1], @XMM[1], @XMM[8]
+	vst1.8	{@XMM[0]-@XMM[1]}, [$out]!	@ write output
+	b	.Lcbc_dec_done
+.align	4
+.Lcbc_dec_one:
+	sub	$inp, $inp, #0x10
+	mov	$rounds, $out			@ save original out pointer
+	mov	$out, $fp			@ use the iv scratch space as out buffer
+	mov	r2, $key
+	vmov	@XMM[4],@XMM[15]		@ just in case ensure that IV
+	vmov	@XMM[5],@XMM[0]			@ and input are preserved
+	bl	AES_decrypt
+	vld1.8	{@XMM[0]}, [$fp,:64]		@ load result
+	veor	@XMM[0], @XMM[0], @XMM[4]	@ ^= IV
+	vmov	@XMM[15], @XMM[5]		@ @XMM[5] holds input
+	vst1.8	{@XMM[0]}, [$rounds]		@ write output
+
+.Lcbc_dec_done:
+#ifndef	BSAES_ASM_EXTENDED_KEY
+	vmov.i32	q0, #0
+	vmov.i32	q1, #0
+.Lcbc_dec_bzero:				@ wipe key schedule [if any]
+	vstmia		$keysched!, {q0-q1}
+	cmp		$keysched, $fp
+	bne		.Lcbc_dec_bzero
+#endif
+
+	mov	sp, $fp
+	add	sp, #0x10			@ add sp,$fp,#0x10 is no good for thumb
+	vst1.8	{@XMM[15]}, [$ivp]		@ return IV
+	VFP_ABI_POP
+	ldmia	sp!, {r4-r10, pc}
+.size	bsaes_cbc_encrypt,.-bsaes_cbc_encrypt
+___
+}
+{
+my ($inp,$out,$len,$key, $ctr,$fp,$rounds)=(map("r$_",(0..3,8..10)));
+my $const = "r6";	# shared with _bsaes_encrypt8_alt
+my $keysched = "sp";
+
+$code.=<<___;
+.extern	AES_encrypt
+.global	bsaes_ctr32_encrypt_blocks
+.type	bsaes_ctr32_encrypt_blocks,%function
+.align	5
+bsaes_ctr32_encrypt_blocks:
+	cmp	$len, #8			@ use plain AES for
+	blo	.Lctr_enc_short			@ small sizes
+
+	mov	ip, sp
+	stmdb	sp!, {r4-r10, lr}
+	VFP_ABI_PUSH
+	ldr	$ctr, [ip]			@ ctr is 1st arg on the stack
+	sub	sp, sp, #0x10			@ scratch space to carry over the ctr
+	mov	$fp, sp				@ save sp
+
+	ldr	$rounds, [$key, #240]		@ get # of rounds
+#ifndef	BSAES_ASM_EXTENDED_KEY
+	@ allocate the key schedule on the stack
+	sub	r12, sp, $rounds, lsl#7		@ 128 bytes per inner round key
+	add	r12, #`128-32`			@ size of bit-sliced key schedule
+
+	@ populate the key schedule
+	mov	r4, $key			@ pass key
+	mov	r5, $rounds			@ pass # of rounds
+	mov	sp, r12				@ sp is $keysched
+	bl	_bsaes_key_convert
+	veor	@XMM[7],@XMM[7],@XMM[15]	@ fix up last round key
+	vstmia	r12, {@XMM[7]}			@ save last round key
+
+	vld1.8	{@XMM[0]}, [$ctr]		@ load counter
+	add	$ctr, $const, #.LREVM0SR-.LM0	@ borrow $ctr
+	vldmia	$keysched, {@XMM[4]}		@ load round0 key
+#else
+	ldr	r12, [$key, #244]
+	eors	r12, #1
+	beq	0f
+
+	@ populate the key schedule
+	str	r12, [$key, #244]
+	mov	r4, $key			@ pass key
+	mov	r5, $rounds			@ pass # of rounds
+	add	r12, $key, #248			@ pass key schedule
+	bl	_bsaes_key_convert
+	veor	@XMM[7],@XMM[7],@XMM[15]	@ fix up last round key
+	vstmia	r12, {@XMM[7]}			@ save last round key
+
+.align	2
+0:	add	r12, $key, #248
+	vld1.8	{@XMM[0]}, [$ctr]		@ load counter
+	adrl	$ctr, .LREVM0SR			@ borrow $ctr
+	vldmia	r12, {@XMM[4]}			@ load round0 key
+	sub	sp, #0x10			@ place for adjusted round0 key
+#endif
+
+	vmov.i32	@XMM[8],#1		@ compose 1<<96
+	veor		@XMM[9],@XMM[9],@XMM[9]
+	vrev32.8	@XMM[0],@XMM[0]
+	vext.8		@XMM[8],@XMM[9],@XMM[8],#4
+	vrev32.8	@XMM[4],@XMM[4]
+	vadd.u32	@XMM[9],@XMM[8],@XMM[8]	@ compose 2<<96
+	vstmia	$keysched, {@XMM[4]}		@ save adjusted round0 key
+	b	.Lctr_enc_loop
+
+.align	4
+.Lctr_enc_loop:
+	vadd.u32	@XMM[10], @XMM[8], @XMM[9]	@ compose 3<<96
+	vadd.u32	@XMM[1], @XMM[0], @XMM[8]	@ +1
+	vadd.u32	@XMM[2], @XMM[0], @XMM[9]	@ +2
+	vadd.u32	@XMM[3], @XMM[0], @XMM[10]	@ +3
+	vadd.u32	@XMM[4], @XMM[1], @XMM[10]
+	vadd.u32	@XMM[5], @XMM[2], @XMM[10]
+	vadd.u32	@XMM[6], @XMM[3], @XMM[10]
+	vadd.u32	@XMM[7], @XMM[4], @XMM[10]
+	vadd.u32	@XMM[10], @XMM[5], @XMM[10]	@ next counter
+
+	@ Borrow prologue from _bsaes_encrypt8 to use the opportunity
+	@ to flip byte order in 32-bit counter
+
+	vldmia		$keysched, {@XMM[9]}		@ load round0 key
+#ifndef	BSAES_ASM_EXTENDED_KEY
+	add		r4, $keysched, #0x10		@ pass next round key
+#else
+	add		r4, $key, #`248+16`
+#endif
+	vldmia		$ctr, {@XMM[8]}			@ .LREVM0SR
+	mov		r5, $rounds			@ pass rounds
+	vstmia		$fp, {@XMM[10]}			@ save next counter
+	sub		$const, $ctr, #.LREVM0SR-.LSR	@ pass constants
+
+	bl		_bsaes_encrypt8_alt
+
+	subs		$len, $len, #8
+	blo		.Lctr_enc_loop_done
+
+	vld1.8		{@XMM[8]-@XMM[9]}, [$inp]!	@ load input
+	vld1.8		{@XMM[10]-@XMM[11]}, [$inp]!
+	veor		@XMM[0], @XMM[8]
+	veor		@XMM[1], @XMM[9]
+	vld1.8		{@XMM[12]-@XMM[13]}, [$inp]!
+	veor		@XMM[4], @XMM[10]
+	veor		@XMM[6], @XMM[11]
+	vld1.8		{@XMM[14]-@XMM[15]}, [$inp]!
+	veor		@XMM[3], @XMM[12]
+	vst1.8		{@XMM[0]-@XMM[1]}, [$out]!	@ write output
+	veor		@XMM[7], @XMM[13]
+	veor		@XMM[2], @XMM[14]
+	vst1.8		{@XMM[4]}, [$out]!
+	veor		@XMM[5], @XMM[15]
+	vst1.8		{@XMM[6]}, [$out]!
+	vmov.i32	@XMM[8], #1			@ compose 1<<96
+	vst1.8		{@XMM[3]}, [$out]!
+	veor		@XMM[9], @XMM[9], @XMM[9]
+	vst1.8		{@XMM[7]}, [$out]!
+	vext.8		@XMM[8], @XMM[9], @XMM[8], #4
+	vst1.8		{@XMM[2]}, [$out]!
+	vadd.u32	@XMM[9],@XMM[8],@XMM[8]		@ compose 2<<96
+	vst1.8		{@XMM[5]}, [$out]!
+	vldmia		$fp, {@XMM[0]}			@ load counter
+
+	bne		.Lctr_enc_loop
+	b		.Lctr_enc_done
+
+.align	4
+.Lctr_enc_loop_done:
+	add		$len, $len, #8
+	vld1.8		{@XMM[8]}, [$inp]!	@ load input
+	veor		@XMM[0], @XMM[8]
+	vst1.8		{@XMM[0]}, [$out]!	@ write output
+	cmp		$len, #2
+	blo		.Lctr_enc_done
+	vld1.8		{@XMM[9]}, [$inp]!
+	veor		@XMM[1], @XMM[9]
+	vst1.8		{@XMM[1]}, [$out]!
+	beq		.Lctr_enc_done
+	vld1.8		{@XMM[10]}, [$inp]!
+	veor		@XMM[4], @XMM[10]
+	vst1.8		{@XMM[4]}, [$out]!
+	cmp		$len, #4
+	blo		.Lctr_enc_done
+	vld1.8		{@XMM[11]}, [$inp]!
+	veor		@XMM[6], @XMM[11]
+	vst1.8		{@XMM[6]}, [$out]!
+	beq		.Lctr_enc_done
+	vld1.8		{@XMM[12]}, [$inp]!
+	veor		@XMM[3], @XMM[12]
+	vst1.8		{@XMM[3]}, [$out]!
+	cmp		$len, #6
+	blo		.Lctr_enc_done
+	vld1.8		{@XMM[13]}, [$inp]!
+	veor		@XMM[7], @XMM[13]
+	vst1.8		{@XMM[7]}, [$out]!
+	beq		.Lctr_enc_done
+	vld1.8		{@XMM[14]}, [$inp]
+	veor		@XMM[2], @XMM[14]
+	vst1.8		{@XMM[2]}, [$out]!
+
+.Lctr_enc_done:
+	vmov.i32	q0, #0
+	vmov.i32	q1, #0
+#ifndef	BSAES_ASM_EXTENDED_KEY
+.Lctr_enc_bzero:			@ wipe key schedule [if any]
+	vstmia		$keysched!, {q0-q1}
+	cmp		$keysched, $fp
+	bne		.Lctr_enc_bzero
+#else
+	vstmia		$keysched, {q0-q1}
+#endif
+
+	mov	sp, $fp
+	add	sp, #0x10		@ add sp,$fp,#0x10 is no good for thumb
+	VFP_ABI_POP
+	ldmia	sp!, {r4-r10, pc}	@ return
+
+.align	4
+.Lctr_enc_short:
+	ldr	ip, [sp]		@ ctr pointer is passed on stack
+	stmdb	sp!, {r4-r8, lr}
+
+	mov	r4, $inp		@ copy arguments
+	mov	r5, $out
+	mov	r6, $len
+	mov	r7, $key
+	ldr	r8, [ip, #12]		@ load counter LSW
+	vld1.8	{@XMM[1]}, [ip]		@ load whole counter value
+#ifdef __ARMEL__
+	rev	r8, r8
+#endif
+	sub	sp, sp, #0x10
+	vst1.8	{@XMM[1]}, [sp,:64]	@ copy counter value
+	sub	sp, sp, #0x10
+
+.Lctr_enc_short_loop:
+	add	r0, sp, #0x10		@ input counter value
+	mov	r1, sp			@ output on the stack
+	mov	r2, r7			@ key
+
+	bl	AES_encrypt
+
+	vld1.8	{@XMM[0]}, [r4]!	@ load input
+	vld1.8	{@XMM[1]}, [sp,:64]	@ load encrypted counter
+	add	r8, r8, #1
+#ifdef __ARMEL__
+	rev	r0, r8
+	str	r0, [sp, #0x1c]		@ next counter value
+#else
+	str	r8, [sp, #0x1c]		@ next counter value
+#endif
+	veor	@XMM[0],@XMM[0],@XMM[1]
+	vst1.8	{@XMM[0]}, [r5]!	@ store output
+	subs	r6, r6, #1
+	bne	.Lctr_enc_short_loop
+
+	vmov.i32	q0, #0
+	vmov.i32	q1, #0
+	vstmia		sp!, {q0-q1}
+
+	ldmia	sp!, {r4-r8, pc}
+.size	bsaes_ctr32_encrypt_blocks,.-bsaes_ctr32_encrypt_blocks
+___
+}
+{
+######################################################################
+# void bsaes_xts_[en|de]crypt(const char *inp,char *out,size_t len,
+#	const AES_KEY *key1, const AES_KEY *key2,
+#	const unsigned char iv[16]);
+#
+my ($inp,$out,$len,$key,$rounds,$magic,$fp)=(map("r$_",(7..10,1..3)));
+my $const="r6";		# returned by _bsaes_key_convert
+my $twmask=@XMM[5];
+my @T=@XMM[6..7];
+
+$code.=<<___;
+.globl	bsaes_xts_encrypt
+.type	bsaes_xts_encrypt,%function
+.align	4
+bsaes_xts_encrypt:
+	mov	ip, sp
+	stmdb	sp!, {r4-r10, lr}		@ 0x20
+	VFP_ABI_PUSH
+	mov	r6, sp				@ future $fp
+
+	mov	$inp, r0
+	mov	$out, r1
+	mov	$len, r2
+	mov	$key, r3
+
+	sub	r0, sp, #0x10			@ 0x10
+	bic	r0, #0xf			@ align at 16 bytes
+	mov	sp, r0
+
+#ifdef	XTS_CHAIN_TWEAK
+	ldr	r0, [ip]			@ pointer to input tweak
+#else
+	@ generate initial tweak
+	ldr	r0, [ip, #4]			@ iv[]
+	mov	r1, sp
+	ldr	r2, [ip, #0]			@ key2
+	bl	AES_encrypt
+	mov	r0,sp				@ pointer to initial tweak
+#endif
+
+	ldr	$rounds, [$key, #240]		@ get # of rounds
+	mov	$fp, r6
+#ifndef	BSAES_ASM_EXTENDED_KEY
+	@ allocate the key schedule on the stack
+	sub	r12, sp, $rounds, lsl#7		@ 128 bytes per inner round key
+	@ add	r12, #`128-32`			@ size of bit-sliced key schedule
+	sub	r12, #`32+16`			@ place for tweak[9]
+
+	@ populate the key schedule
+	mov	r4, $key			@ pass key
+	mov	r5, $rounds			@ pass # of rounds
+	mov	sp, r12
+	add	r12, #0x90			@ pass key schedule
+	bl	_bsaes_key_convert
+	veor	@XMM[7], @XMM[7], @XMM[15]	@ fix up last round key
+	vstmia	r12, {@XMM[7]}			@ save last round key
+#else
+	ldr	r12, [$key, #244]
+	eors	r12, #1
+	beq	0f
+
+	str	r12, [$key, #244]
+	mov	r4, $key			@ pass key
+	mov	r5, $rounds			@ pass # of rounds
+	add	r12, $key, #248			@ pass key schedule
+	bl	_bsaes_key_convert
+	veor	@XMM[7], @XMM[7], @XMM[15]	@ fix up last round key
+	vstmia	r12, {@XMM[7]}
+
+.align	2
+0:	sub	sp, #0x90			@ place for tweak[9]
+#endif
+
+	vld1.8	{@XMM[8]}, [r0]			@ initial tweak
+	adr	$magic, .Lxts_magic
+
+	subs	$len, #0x80
+	blo	.Lxts_enc_short
+	b	.Lxts_enc_loop
+
+.align	4
+.Lxts_enc_loop:
+	vldmia		$magic, {$twmask}	@ load XTS magic
+	vshr.s64	@T[0], @XMM[8], #63
+	mov		r0, sp
+	vand		@T[0], @T[0], $twmask
+___
+for($i=9;$i<16;$i++) {
+$code.=<<___;
+	vadd.u64	@XMM[$i], @XMM[$i-1], @XMM[$i-1]
+	vst1.64		{@XMM[$i-1]}, [r0,:128]!
+	vswp		`&Dhi("@T[0]")`,`&Dlo("@T[0]")`
+	vshr.s64	@T[1], @XMM[$i], #63
+	veor		@XMM[$i], @XMM[$i], @T[0]
+	vand		@T[1], @T[1], $twmask
+___
+	@T=reverse(@T);
+
+$code.=<<___ if ($i>=10);
+	vld1.8		{@XMM[$i-10]}, [$inp]!
+___
+$code.=<<___ if ($i>=11);
+	veor		@XMM[$i-11], @XMM[$i-11], @XMM[$i-3]
+___
+}
+$code.=<<___;
+	vadd.u64	@XMM[8], @XMM[15], @XMM[15]
+	vst1.64		{@XMM[15]}, [r0,:128]!
+	vswp		`&Dhi("@T[0]")`,`&Dlo("@T[0]")`
+	veor		@XMM[8], @XMM[8], @T[0]
+	vst1.64		{@XMM[8]}, [r0,:128]		@ next round tweak
+
+	vld1.8		{@XMM[6]-@XMM[7]}, [$inp]!
+	veor		@XMM[5], @XMM[5], @XMM[13]
+#ifndef	BSAES_ASM_EXTENDED_KEY
+	add		r4, sp, #0x90			@ pass key schedule
+#else
+	add		r4, $key, #248			@ pass key schedule
+#endif
+	veor		@XMM[6], @XMM[6], @XMM[14]
+	mov		r5, $rounds			@ pass rounds
+	veor		@XMM[7], @XMM[7], @XMM[15]
+	mov		r0, sp
+
+	bl		_bsaes_encrypt8
+
+	vld1.64		{@XMM[ 8]-@XMM[ 9]}, [r0,:128]!
+	vld1.64		{@XMM[10]-@XMM[11]}, [r0,:128]!
+	veor		@XMM[0], @XMM[0], @XMM[ 8]
+	vld1.64		{@XMM[12]-@XMM[13]}, [r0,:128]!
+	veor		@XMM[1], @XMM[1], @XMM[ 9]
+	veor		@XMM[8], @XMM[4], @XMM[10]
+	vst1.8		{@XMM[0]-@XMM[1]}, [$out]!
+	veor		@XMM[9], @XMM[6], @XMM[11]
+	vld1.64		{@XMM[14]-@XMM[15]}, [r0,:128]!
+	veor		@XMM[10], @XMM[3], @XMM[12]
+	vst1.8		{@XMM[8]-@XMM[9]}, [$out]!
+	veor		@XMM[11], @XMM[7], @XMM[13]
+	veor		@XMM[12], @XMM[2], @XMM[14]
+	vst1.8		{@XMM[10]-@XMM[11]}, [$out]!
+	veor		@XMM[13], @XMM[5], @XMM[15]
+	vst1.8		{@XMM[12]-@XMM[13]}, [$out]!
+
+	vld1.64		{@XMM[8]}, [r0,:128]		@ next round tweak
+
+	subs		$len, #0x80
+	bpl		.Lxts_enc_loop
+
+.Lxts_enc_short:
+	adds		$len, #0x70
+	bmi		.Lxts_enc_done
+
+	vldmia		$magic, {$twmask}	@ load XTS magic
+	vshr.s64	@T[0], @XMM[8], #63
+	mov		r0, sp
+	vand		@T[0], @T[0], $twmask
+___
+for($i=9;$i<16;$i++) {
+$code.=<<___;
+	vadd.u64	@XMM[$i], @XMM[$i-1], @XMM[$i-1]
+	vst1.64		{@XMM[$i-1]}, [r0,:128]!
+	vswp		`&Dhi("@T[0]")`,`&Dlo("@T[0]")`
+	vshr.s64	@T[1], @XMM[$i], #63
+	veor		@XMM[$i], @XMM[$i], @T[0]
+	vand		@T[1], @T[1], $twmask
+___
+	@T=reverse(@T);
+
+$code.=<<___ if ($i>=10);
+	vld1.8		{@XMM[$i-10]}, [$inp]!
+	subs		$len, #0x10
+	bmi		.Lxts_enc_`$i-9`
+___
+$code.=<<___ if ($i>=11);
+	veor		@XMM[$i-11], @XMM[$i-11], @XMM[$i-3]
+___
+}
+$code.=<<___;
+	sub		$len, #0x10
+	vst1.64		{@XMM[15]}, [r0,:128]		@ next round tweak
+
+	vld1.8		{@XMM[6]}, [$inp]!
+	veor		@XMM[5], @XMM[5], @XMM[13]
+#ifndef	BSAES_ASM_EXTENDED_KEY
+	add		r4, sp, #0x90			@ pass key schedule
+#else
+	add		r4, $key, #248			@ pass key schedule
+#endif
+	veor		@XMM[6], @XMM[6], @XMM[14]
+	mov		r5, $rounds			@ pass rounds
+	mov		r0, sp
+
+	bl		_bsaes_encrypt8
+
+	vld1.64		{@XMM[ 8]-@XMM[ 9]}, [r0,:128]!
+	vld1.64		{@XMM[10]-@XMM[11]}, [r0,:128]!
+	veor		@XMM[0], @XMM[0], @XMM[ 8]
+	vld1.64		{@XMM[12]-@XMM[13]}, [r0,:128]!
+	veor		@XMM[1], @XMM[1], @XMM[ 9]
+	veor		@XMM[8], @XMM[4], @XMM[10]
+	vst1.8		{@XMM[0]-@XMM[1]}, [$out]!
+	veor		@XMM[9], @XMM[6], @XMM[11]
+	vld1.64		{@XMM[14]}, [r0,:128]!
+	veor		@XMM[10], @XMM[3], @XMM[12]
+	vst1.8		{@XMM[8]-@XMM[9]}, [$out]!
+	veor		@XMM[11], @XMM[7], @XMM[13]
+	veor		@XMM[12], @XMM[2], @XMM[14]
+	vst1.8		{@XMM[10]-@XMM[11]}, [$out]!
+	vst1.8		{@XMM[12]}, [$out]!
+
+	vld1.64		{@XMM[8]}, [r0,:128]		@ next round tweak
+	b		.Lxts_enc_done
+.align	4
+.Lxts_enc_6:
+	vst1.64		{@XMM[14]}, [r0,:128]		@ next round tweak
+
+	veor		@XMM[4], @XMM[4], @XMM[12]
+#ifndef	BSAES_ASM_EXTENDED_KEY
+	add		r4, sp, #0x90			@ pass key schedule
+#else
+	add		r4, $key, #248			@ pass key schedule
+#endif
+	veor		@XMM[5], @XMM[5], @XMM[13]
+	mov		r5, $rounds			@ pass rounds
+	mov		r0, sp
+
+	bl		_bsaes_encrypt8
+
+	vld1.64		{@XMM[ 8]-@XMM[ 9]}, [r0,:128]!
+	vld1.64		{@XMM[10]-@XMM[11]}, [r0,:128]!
+	veor		@XMM[0], @XMM[0], @XMM[ 8]
+	vld1.64		{@XMM[12]-@XMM[13]}, [r0,:128]!
+	veor		@XMM[1], @XMM[1], @XMM[ 9]
+	veor		@XMM[8], @XMM[4], @XMM[10]
+	vst1.8		{@XMM[0]-@XMM[1]}, [$out]!
+	veor		@XMM[9], @XMM[6], @XMM[11]
+	veor		@XMM[10], @XMM[3], @XMM[12]
+	vst1.8		{@XMM[8]-@XMM[9]}, [$out]!
+	veor		@XMM[11], @XMM[7], @XMM[13]
+	vst1.8		{@XMM[10]-@XMM[11]}, [$out]!
+
+	vld1.64		{@XMM[8]}, [r0,:128]		@ next round tweak
+	b		.Lxts_enc_done
+
+@ put this in range for both ARM and Thumb mode adr instructions
+.align	5
+.Lxts_magic:
+	.quad	1, 0x87
+
+.align	5
+.Lxts_enc_5:
+	vst1.64		{@XMM[13]}, [r0,:128]		@ next round tweak
+
+	veor		@XMM[3], @XMM[3], @XMM[11]
+#ifndef	BSAES_ASM_EXTENDED_KEY
+	add		r4, sp, #0x90			@ pass key schedule
+#else
+	add		r4, $key, #248			@ pass key schedule
+#endif
+	veor		@XMM[4], @XMM[4], @XMM[12]
+	mov		r5, $rounds			@ pass rounds
+	mov		r0, sp
+
+	bl		_bsaes_encrypt8
+
+	vld1.64		{@XMM[ 8]-@XMM[ 9]}, [r0,:128]!
+	vld1.64		{@XMM[10]-@XMM[11]}, [r0,:128]!
+	veor		@XMM[0], @XMM[0], @XMM[ 8]
+	vld1.64		{@XMM[12]}, [r0,:128]!
+	veor		@XMM[1], @XMM[1], @XMM[ 9]
+	veor		@XMM[8], @XMM[4], @XMM[10]
+	vst1.8		{@XMM[0]-@XMM[1]}, [$out]!
+	veor		@XMM[9], @XMM[6], @XMM[11]
+	veor		@XMM[10], @XMM[3], @XMM[12]
+	vst1.8		{@XMM[8]-@XMM[9]}, [$out]!
+	vst1.8		{@XMM[10]}, [$out]!
+
+	vld1.64		{@XMM[8]}, [r0,:128]		@ next round tweak
+	b		.Lxts_enc_done
+.align	4
+.Lxts_enc_4:
+	vst1.64		{@XMM[12]}, [r0,:128]		@ next round tweak
+
+	veor		@XMM[2], @XMM[2], @XMM[10]
+#ifndef	BSAES_ASM_EXTENDED_KEY
+	add		r4, sp, #0x90			@ pass key schedule
+#else
+	add		r4, $key, #248			@ pass key schedule
+#endif
+	veor		@XMM[3], @XMM[3], @XMM[11]
+	mov		r5, $rounds			@ pass rounds
+	mov		r0, sp
+
+	bl		_bsaes_encrypt8
+
+	vld1.64		{@XMM[ 8]-@XMM[ 9]}, [r0,:128]!
+	vld1.64		{@XMM[10]-@XMM[11]}, [r0,:128]!
+	veor		@XMM[0], @XMM[0], @XMM[ 8]
+	veor		@XMM[1], @XMM[1], @XMM[ 9]
+	veor		@XMM[8], @XMM[4], @XMM[10]
+	vst1.8		{@XMM[0]-@XMM[1]}, [$out]!
+	veor		@XMM[9], @XMM[6], @XMM[11]
+	vst1.8		{@XMM[8]-@XMM[9]}, [$out]!
+
+	vld1.64		{@XMM[8]}, [r0,:128]		@ next round tweak
+	b		.Lxts_enc_done
+.align	4
+.Lxts_enc_3:
+	vst1.64		{@XMM[11]}, [r0,:128]		@ next round tweak
+
+	veor		@XMM[1], @XMM[1], @XMM[9]
+#ifndef	BSAES_ASM_EXTENDED_KEY
+	add		r4, sp, #0x90			@ pass key schedule
+#else
+	add		r4, $key, #248			@ pass key schedule
+#endif
+	veor		@XMM[2], @XMM[2], @XMM[10]
+	mov		r5, $rounds			@ pass rounds
+	mov		r0, sp
+
+	bl		_bsaes_encrypt8
+
+	vld1.64		{@XMM[8]-@XMM[9]}, [r0,:128]!
+	vld1.64		{@XMM[10]}, [r0,:128]!
+	veor		@XMM[0], @XMM[0], @XMM[ 8]
+	veor		@XMM[1], @XMM[1], @XMM[ 9]
+	veor		@XMM[8], @XMM[4], @XMM[10]
+	vst1.8		{@XMM[0]-@XMM[1]}, [$out]!
+	vst1.8		{@XMM[8]}, [$out]!
+
+	vld1.64		{@XMM[8]}, [r0,:128]		@ next round tweak
+	b		.Lxts_enc_done
+.align	4
+.Lxts_enc_2:
+	vst1.64		{@XMM[10]}, [r0,:128]		@ next round tweak
+
+	veor		@XMM[0], @XMM[0], @XMM[8]
+#ifndef	BSAES_ASM_EXTENDED_KEY
+	add		r4, sp, #0x90			@ pass key schedule
+#else
+	add		r4, $key, #248			@ pass key schedule
+#endif
+	veor		@XMM[1], @XMM[1], @XMM[9]
+	mov		r5, $rounds			@ pass rounds
+	mov		r0, sp
+
+	bl		_bsaes_encrypt8
+
+	vld1.64		{@XMM[8]-@XMM[9]}, [r0,:128]!
+	veor		@XMM[0], @XMM[0], @XMM[ 8]
+	veor		@XMM[1], @XMM[1], @XMM[ 9]
+	vst1.8		{@XMM[0]-@XMM[1]}, [$out]!
+
+	vld1.64		{@XMM[8]}, [r0,:128]		@ next round tweak
+	b		.Lxts_enc_done
+.align	4
+.Lxts_enc_1:
+	mov		r0, sp
+	veor		@XMM[0], @XMM[8]
+	mov		r1, sp
+	vst1.8		{@XMM[0]}, [sp,:128]
+	mov		r2, $key
+	mov		r4, $fp				@ preserve fp
+
+	bl		AES_encrypt
+
+	vld1.8		{@XMM[0]}, [sp,:128]
+	veor		@XMM[0], @XMM[0], @XMM[8]
+	vst1.8		{@XMM[0]}, [$out]!
+	mov		$fp, r4
+
+	vmov		@XMM[8], @XMM[9]		@ next round tweak
+
+.Lxts_enc_done:
+#ifndef	XTS_CHAIN_TWEAK
+	adds		$len, #0x10
+	beq		.Lxts_enc_ret
+	sub		r6, $out, #0x10
+
+.Lxts_enc_steal:
+	ldrb		r0, [$inp], #1
+	ldrb		r1, [$out, #-0x10]
+	strb		r0, [$out, #-0x10]
+	strb		r1, [$out], #1
+
+	subs		$len, #1
+	bhi		.Lxts_enc_steal
+
+	vld1.8		{@XMM[0]}, [r6]
+	mov		r0, sp
+	veor		@XMM[0], @XMM[0], @XMM[8]
+	mov		r1, sp
+	vst1.8		{@XMM[0]}, [sp,:128]
+	mov		r2, $key
+	mov		r4, $fp			@ preserve fp
+
+	bl		AES_encrypt
+
+	vld1.8		{@XMM[0]}, [sp,:128]
+	veor		@XMM[0], @XMM[0], @XMM[8]
+	vst1.8		{@XMM[0]}, [r6]
+	mov		$fp, r4
+#endif
+
+.Lxts_enc_ret:
+	bic		r0, $fp, #0xf
+	vmov.i32	q0, #0
+	vmov.i32	q1, #0
+#ifdef	XTS_CHAIN_TWEAK
+	ldr		r1, [$fp, #0x20+VFP_ABI_FRAME]	@ chain tweak
+#endif
+.Lxts_enc_bzero:				@ wipe key schedule [if any]
+	vstmia		sp!, {q0-q1}
+	cmp		sp, r0
+	bne		.Lxts_enc_bzero
+
+	mov		sp, $fp
+#ifdef	XTS_CHAIN_TWEAK
+	vst1.8		{@XMM[8]}, [r1]
+#endif
+	VFP_ABI_POP
+	ldmia		sp!, {r4-r10, pc}	@ return
+
+.size	bsaes_xts_encrypt,.-bsaes_xts_encrypt
+
+.globl	bsaes_xts_decrypt
+.type	bsaes_xts_decrypt,%function
+.align	4
+bsaes_xts_decrypt:
+	mov	ip, sp
+	stmdb	sp!, {r4-r10, lr}		@ 0x20
+	VFP_ABI_PUSH
+	mov	r6, sp				@ future $fp
+
+	mov	$inp, r0
+	mov	$out, r1
+	mov	$len, r2
+	mov	$key, r3
+
+	sub	r0, sp, #0x10			@ 0x10
+	bic	r0, #0xf			@ align at 16 bytes
+	mov	sp, r0
+
+#ifdef	XTS_CHAIN_TWEAK
+	ldr	r0, [ip]			@ pointer to input tweak
+#else
+	@ generate initial tweak
+	ldr	r0, [ip, #4]			@ iv[]
+	mov	r1, sp
+	ldr	r2, [ip, #0]			@ key2
+	bl	AES_encrypt
+	mov	r0, sp				@ pointer to initial tweak
+#endif
+
+	ldr	$rounds, [$key, #240]		@ get # of rounds
+	mov	$fp, r6
+#ifndef	BSAES_ASM_EXTENDED_KEY
+	@ allocate the key schedule on the stack
+	sub	r12, sp, $rounds, lsl#7		@ 128 bytes per inner round key
+	@ add	r12, #`128-32`			@ size of bit-sliced key schedule
+	sub	r12, #`32+16`			@ place for tweak[9]
+
+	@ populate the key schedule
+	mov	r4, $key			@ pass key
+	mov	r5, $rounds			@ pass # of rounds
+	mov	sp, r12
+	add	r12, #0x90			@ pass key schedule
+	bl	_bsaes_key_convert
+	add	r4, sp, #0x90
+	vldmia	r4, {@XMM[6]}
+	vstmia	r12,  {@XMM[15]}		@ save last round key
+	veor	@XMM[7], @XMM[7], @XMM[6]	@ fix up round 0 key
+	vstmia	r4, {@XMM[7]}
+#else
+	ldr	r12, [$key, #244]
+	eors	r12, #1
+	beq	0f
+
+	str	r12, [$key, #244]
+	mov	r4, $key			@ pass key
+	mov	r5, $rounds			@ pass # of rounds
+	add	r12, $key, #248			@ pass key schedule
+	bl	_bsaes_key_convert
+	add	r4, $key, #248
+	vldmia	r4, {@XMM[6]}
+	vstmia	r12,  {@XMM[15]}		@ save last round key
+	veor	@XMM[7], @XMM[7], @XMM[6]	@ fix up round 0 key
+	vstmia	r4, {@XMM[7]}
+
+.align	2
+0:	sub	sp, #0x90			@ place for tweak[9]
+#endif
+	vld1.8	{@XMM[8]}, [r0]			@ initial tweak
+	adr	$magic, .Lxts_magic
+
+	tst	$len, #0xf			@ if not multiple of 16
+	it	ne				@ Thumb2 thing, sanity check in ARM
+	subne	$len, #0x10			@ subtract another 16 bytes
+	subs	$len, #0x80
+
+	blo	.Lxts_dec_short
+	b	.Lxts_dec_loop
+
+.align	4
+.Lxts_dec_loop:
+	vldmia		$magic, {$twmask}	@ load XTS magic
+	vshr.s64	@T[0], @XMM[8], #63
+	mov		r0, sp
+	vand		@T[0], @T[0], $twmask
+___
+for($i=9;$i<16;$i++) {
+$code.=<<___;
+	vadd.u64	@XMM[$i], @XMM[$i-1], @XMM[$i-1]
+	vst1.64		{@XMM[$i-1]}, [r0,:128]!
+	vswp		`&Dhi("@T[0]")`,`&Dlo("@T[0]")`
+	vshr.s64	@T[1], @XMM[$i], #63
+	veor		@XMM[$i], @XMM[$i], @T[0]
+	vand		@T[1], @T[1], $twmask
+___
+	@T=reverse(@T);
+
+$code.=<<___ if ($i>=10);
+	vld1.8		{@XMM[$i-10]}, [$inp]!
+___
+$code.=<<___ if ($i>=11);
+	veor		@XMM[$i-11], @XMM[$i-11], @XMM[$i-3]
+___
+}
+$code.=<<___;
+	vadd.u64	@XMM[8], @XMM[15], @XMM[15]
+	vst1.64		{@XMM[15]}, [r0,:128]!
+	vswp		`&Dhi("@T[0]")`,`&Dlo("@T[0]")`
+	veor		@XMM[8], @XMM[8], @T[0]
+	vst1.64		{@XMM[8]}, [r0,:128]		@ next round tweak
+
+	vld1.8		{@XMM[6]-@XMM[7]}, [$inp]!
+	veor		@XMM[5], @XMM[5], @XMM[13]
+#ifndef	BSAES_ASM_EXTENDED_KEY
+	add		r4, sp, #0x90			@ pass key schedule
+#else
+	add		r4, $key, #248			@ pass key schedule
+#endif
+	veor		@XMM[6], @XMM[6], @XMM[14]
+	mov		r5, $rounds			@ pass rounds
+	veor		@XMM[7], @XMM[7], @XMM[15]
+	mov		r0, sp
+
+	bl		_bsaes_decrypt8
+
+	vld1.64		{@XMM[ 8]-@XMM[ 9]}, [r0,:128]!
+	vld1.64		{@XMM[10]-@XMM[11]}, [r0,:128]!
+	veor		@XMM[0], @XMM[0], @XMM[ 8]
+	vld1.64		{@XMM[12]-@XMM[13]}, [r0,:128]!
+	veor		@XMM[1], @XMM[1], @XMM[ 9]
+	veor		@XMM[8], @XMM[6], @XMM[10]
+	vst1.8		{@XMM[0]-@XMM[1]}, [$out]!
+	veor		@XMM[9], @XMM[4], @XMM[11]
+	vld1.64		{@XMM[14]-@XMM[15]}, [r0,:128]!
+	veor		@XMM[10], @XMM[2], @XMM[12]
+	vst1.8		{@XMM[8]-@XMM[9]}, [$out]!
+	veor		@XMM[11], @XMM[7], @XMM[13]
+	veor		@XMM[12], @XMM[3], @XMM[14]
+	vst1.8		{@XMM[10]-@XMM[11]}, [$out]!
+	veor		@XMM[13], @XMM[5], @XMM[15]
+	vst1.8		{@XMM[12]-@XMM[13]}, [$out]!
+
+	vld1.64		{@XMM[8]}, [r0,:128]		@ next round tweak
+
+	subs		$len, #0x80
+	bpl		.Lxts_dec_loop
+
+.Lxts_dec_short:
+	adds		$len, #0x70
+	bmi		.Lxts_dec_done
+
+	vldmia		$magic, {$twmask}	@ load XTS magic
+	vshr.s64	@T[0], @XMM[8], #63
+	mov		r0, sp
+	vand		@T[0], @T[0], $twmask
+___
+for($i=9;$i<16;$i++) {
+$code.=<<___;
+	vadd.u64	@XMM[$i], @XMM[$i-1], @XMM[$i-1]
+	vst1.64		{@XMM[$i-1]}, [r0,:128]!
+	vswp		`&Dhi("@T[0]")`,`&Dlo("@T[0]")`
+	vshr.s64	@T[1], @XMM[$i], #63
+	veor		@XMM[$i], @XMM[$i], @T[0]
+	vand		@T[1], @T[1], $twmask
+___
+	@T=reverse(@T);
+
+$code.=<<___ if ($i>=10);
+	vld1.8		{@XMM[$i-10]}, [$inp]!
+	subs		$len, #0x10
+	bmi		.Lxts_dec_`$i-9`
+___
+$code.=<<___ if ($i>=11);
+	veor		@XMM[$i-11], @XMM[$i-11], @XMM[$i-3]
+___
+}
+$code.=<<___;
+	sub		$len, #0x10
+	vst1.64		{@XMM[15]}, [r0,:128]		@ next round tweak
+
+	vld1.8		{@XMM[6]}, [$inp]!
+	veor		@XMM[5], @XMM[5], @XMM[13]
+#ifndef	BSAES_ASM_EXTENDED_KEY
+	add		r4, sp, #0x90			@ pass key schedule
+#else
+	add		r4, $key, #248			@ pass key schedule
+#endif
+	veor		@XMM[6], @XMM[6], @XMM[14]
+	mov		r5, $rounds			@ pass rounds
+	mov		r0, sp
+
+	bl		_bsaes_decrypt8
+
+	vld1.64		{@XMM[ 8]-@XMM[ 9]}, [r0,:128]!
+	vld1.64		{@XMM[10]-@XMM[11]}, [r0,:128]!
+	veor		@XMM[0], @XMM[0], @XMM[ 8]
+	vld1.64		{@XMM[12]-@XMM[13]}, [r0,:128]!
+	veor		@XMM[1], @XMM[1], @XMM[ 9]
+	veor		@XMM[8], @XMM[6], @XMM[10]
+	vst1.8		{@XMM[0]-@XMM[1]}, [$out]!
+	veor		@XMM[9], @XMM[4], @XMM[11]
+	vld1.64		{@XMM[14]}, [r0,:128]!
+	veor		@XMM[10], @XMM[2], @XMM[12]
+	vst1.8		{@XMM[8]-@XMM[9]}, [$out]!
+	veor		@XMM[11], @XMM[7], @XMM[13]
+	veor		@XMM[12], @XMM[3], @XMM[14]
+	vst1.8		{@XMM[10]-@XMM[11]}, [$out]!
+	vst1.8		{@XMM[12]}, [$out]!
+
+	vld1.64		{@XMM[8]}, [r0,:128]		@ next round tweak
+	b		.Lxts_dec_done
+.align	4
+.Lxts_dec_6:
+	vst1.64		{@XMM[14]}, [r0,:128]		@ next round tweak
+
+	veor		@XMM[4], @XMM[4], @XMM[12]
+#ifndef	BSAES_ASM_EXTENDED_KEY
+	add		r4, sp, #0x90			@ pass key schedule
+#else
+	add		r4, $key, #248			@ pass key schedule
+#endif
+	veor		@XMM[5], @XMM[5], @XMM[13]
+	mov		r5, $rounds			@ pass rounds
+	mov		r0, sp
+
+	bl		_bsaes_decrypt8
+
+	vld1.64		{@XMM[ 8]-@XMM[ 9]}, [r0,:128]!
+	vld1.64		{@XMM[10]-@XMM[11]}, [r0,:128]!
+	veor		@XMM[0], @XMM[0], @XMM[ 8]
+	vld1.64		{@XMM[12]-@XMM[13]}, [r0,:128]!
+	veor		@XMM[1], @XMM[1], @XMM[ 9]
+	veor		@XMM[8], @XMM[6], @XMM[10]
+	vst1.8		{@XMM[0]-@XMM[1]}, [$out]!
+	veor		@XMM[9], @XMM[4], @XMM[11]
+	veor		@XMM[10], @XMM[2], @XMM[12]
+	vst1.8		{@XMM[8]-@XMM[9]}, [$out]!
+	veor		@XMM[11], @XMM[7], @XMM[13]
+	vst1.8		{@XMM[10]-@XMM[11]}, [$out]!
+
+	vld1.64		{@XMM[8]}, [r0,:128]		@ next round tweak
+	b		.Lxts_dec_done
+.align	4
+.Lxts_dec_5:
+	vst1.64		{@XMM[13]}, [r0,:128]		@ next round tweak
+
+	veor		@XMM[3], @XMM[3], @XMM[11]
+#ifndef	BSAES_ASM_EXTENDED_KEY
+	add		r4, sp, #0x90			@ pass key schedule
+#else
+	add		r4, $key, #248			@ pass key schedule
+#endif
+	veor		@XMM[4], @XMM[4], @XMM[12]
+	mov		r5, $rounds			@ pass rounds
+	mov		r0, sp
+
+	bl		_bsaes_decrypt8
+
+	vld1.64		{@XMM[ 8]-@XMM[ 9]}, [r0,:128]!
+	vld1.64		{@XMM[10]-@XMM[11]}, [r0,:128]!
+	veor		@XMM[0], @XMM[0], @XMM[ 8]
+	vld1.64		{@XMM[12]}, [r0,:128]!
+	veor		@XMM[1], @XMM[1], @XMM[ 9]
+	veor		@XMM[8], @XMM[6], @XMM[10]
+	vst1.8		{@XMM[0]-@XMM[1]}, [$out]!
+	veor		@XMM[9], @XMM[4], @XMM[11]
+	veor		@XMM[10], @XMM[2], @XMM[12]
+	vst1.8		{@XMM[8]-@XMM[9]}, [$out]!
+	vst1.8		{@XMM[10]}, [$out]!
+
+	vld1.64		{@XMM[8]}, [r0,:128]		@ next round tweak
+	b		.Lxts_dec_done
+.align	4
+.Lxts_dec_4:
+	vst1.64		{@XMM[12]}, [r0,:128]		@ next round tweak
+
+	veor		@XMM[2], @XMM[2], @XMM[10]
+#ifndef	BSAES_ASM_EXTENDED_KEY
+	add		r4, sp, #0x90			@ pass key schedule
+#else
+	add		r4, $key, #248			@ pass key schedule
+#endif
+	veor		@XMM[3], @XMM[3], @XMM[11]
+	mov		r5, $rounds			@ pass rounds
+	mov		r0, sp
+
+	bl		_bsaes_decrypt8
+
+	vld1.64		{@XMM[ 8]-@XMM[ 9]}, [r0,:128]!
+	vld1.64		{@XMM[10]-@XMM[11]}, [r0,:128]!
+	veor		@XMM[0], @XMM[0], @XMM[ 8]
+	veor		@XMM[1], @XMM[1], @XMM[ 9]
+	veor		@XMM[8], @XMM[6], @XMM[10]
+	vst1.8		{@XMM[0]-@XMM[1]}, [$out]!
+	veor		@XMM[9], @XMM[4], @XMM[11]
+	vst1.8		{@XMM[8]-@XMM[9]}, [$out]!
+
+	vld1.64		{@XMM[8]}, [r0,:128]		@ next round tweak
+	b		.Lxts_dec_done
+.align	4
+.Lxts_dec_3:
+	vst1.64		{@XMM[11]}, [r0,:128]		@ next round tweak
+
+	veor		@XMM[1], @XMM[1], @XMM[9]
+#ifndef	BSAES_ASM_EXTENDED_KEY
+	add		r4, sp, #0x90			@ pass key schedule
+#else
+	add		r4, $key, #248			@ pass key schedule
+#endif
+	veor		@XMM[2], @XMM[2], @XMM[10]
+	mov		r5, $rounds			@ pass rounds
+	mov		r0, sp
+
+	bl		_bsaes_decrypt8
+
+	vld1.64		{@XMM[8]-@XMM[9]}, [r0,:128]!
+	vld1.64		{@XMM[10]}, [r0,:128]!
+	veor		@XMM[0], @XMM[0], @XMM[ 8]
+	veor		@XMM[1], @XMM[1], @XMM[ 9]
+	veor		@XMM[8], @XMM[6], @XMM[10]
+	vst1.8		{@XMM[0]-@XMM[1]}, [$out]!
+	vst1.8		{@XMM[8]}, [$out]!
+
+	vld1.64		{@XMM[8]}, [r0,:128]		@ next round tweak
+	b		.Lxts_dec_done
+.align	4
+.Lxts_dec_2:
+	vst1.64		{@XMM[10]}, [r0,:128]		@ next round tweak
+
+	veor		@XMM[0], @XMM[0], @XMM[8]
+#ifndef	BSAES_ASM_EXTENDED_KEY
+	add		r4, sp, #0x90			@ pass key schedule
+#else
+	add		r4, $key, #248			@ pass key schedule
+#endif
+	veor		@XMM[1], @XMM[1], @XMM[9]
+	mov		r5, $rounds			@ pass rounds
+	mov		r0, sp
+
+	bl		_bsaes_decrypt8
+
+	vld1.64		{@XMM[8]-@XMM[9]}, [r0,:128]!
+	veor		@XMM[0], @XMM[0], @XMM[ 8]
+	veor		@XMM[1], @XMM[1], @XMM[ 9]
+	vst1.8		{@XMM[0]-@XMM[1]}, [$out]!
+
+	vld1.64		{@XMM[8]}, [r0,:128]		@ next round tweak
+	b		.Lxts_dec_done
+.align	4
+.Lxts_dec_1:
+	mov		r0, sp
+	veor		@XMM[0], @XMM[8]
+	mov		r1, sp
+	vst1.8		{@XMM[0]}, [sp,:128]
+	mov		r2, $key
+	mov		r4, $fp				@ preserve fp
+	mov		r5, $magic			@ preserve magic
+
+	bl		AES_decrypt
+
+	vld1.8		{@XMM[0]}, [sp,:128]
+	veor		@XMM[0], @XMM[0], @XMM[8]
+	vst1.8		{@XMM[0]}, [$out]!
+	mov		$fp, r4
+	mov		$magic, r5
+
+	vmov		@XMM[8], @XMM[9]		@ next round tweak
+
+.Lxts_dec_done:
+#ifndef	XTS_CHAIN_TWEAK
+	adds		$len, #0x10
+	beq		.Lxts_dec_ret
+
+	@ calculate one round of extra tweak for the stolen ciphertext
+	vldmia		$magic, {$twmask}
+	vshr.s64	@XMM[6], @XMM[8], #63
+	vand		@XMM[6], @XMM[6], $twmask
+	vadd.u64	@XMM[9], @XMM[8], @XMM[8]
+	vswp		`&Dhi("@XMM[6]")`,`&Dlo("@XMM[6]")`
+	veor		@XMM[9], @XMM[9], @XMM[6]
+
+	@ perform the final decryption with the last tweak value
+	vld1.8		{@XMM[0]}, [$inp]!
+	mov		r0, sp
+	veor		@XMM[0], @XMM[0], @XMM[9]
+	mov		r1, sp
+	vst1.8		{@XMM[0]}, [sp,:128]
+	mov		r2, $key
+	mov		r4, $fp			@ preserve fp
+
+	bl		AES_decrypt
+
+	vld1.8		{@XMM[0]}, [sp,:128]
+	veor		@XMM[0], @XMM[0], @XMM[9]
+	vst1.8		{@XMM[0]}, [$out]
+
+	mov		r6, $out
+.Lxts_dec_steal:
+	ldrb		r1, [$out]
+	ldrb		r0, [$inp], #1
+	strb		r1, [$out, #0x10]
+	strb		r0, [$out], #1
+
+	subs		$len, #1
+	bhi		.Lxts_dec_steal
+
+	vld1.8		{@XMM[0]}, [r6]
+	mov		r0, sp
+	veor		@XMM[0], @XMM[8]
+	mov		r1, sp
+	vst1.8		{@XMM[0]}, [sp,:128]
+	mov		r2, $key
+
+	bl		AES_decrypt
+
+	vld1.8		{@XMM[0]}, [sp,:128]
+	veor		@XMM[0], @XMM[0], @XMM[8]
+	vst1.8		{@XMM[0]}, [r6]
+	mov		$fp, r4
+#endif
+
+.Lxts_dec_ret:
+	bic		r0, $fp, #0xf
+	vmov.i32	q0, #0
+	vmov.i32	q1, #0
+#ifdef	XTS_CHAIN_TWEAK
+	ldr		r1, [$fp, #0x20+VFP_ABI_FRAME]	@ chain tweak
+#endif
+.Lxts_dec_bzero:				@ wipe key schedule [if any]
+	vstmia		sp!, {q0-q1}
+	cmp		sp, r0
+	bne		.Lxts_dec_bzero
+
+	mov		sp, $fp
+#ifdef	XTS_CHAIN_TWEAK
+	vst1.8		{@XMM[8]}, [r1]
+#endif
+	VFP_ABI_POP
+	ldmia		sp!, {r4-r10, pc}	@ return
+
+.size	bsaes_xts_decrypt,.-bsaes_xts_decrypt
+___
+}
+$code.=<<___;
+#endif
+___
+
+$code =~ s/\`([^\`]*)\`/eval($1)/gem;
+
+open SELF,$0;
+while(<SELF>) {
+	next if (/^#!/);
+        last if (!s/^#/@/ and !/^$/);
+        print;
+}
+close SELF;
+
+print $code;
+
+close STDOUT;
diff --git a/arch/arm/include/asm/Kbuild b/arch/arm/include/asm/Kbuild
index d3db39860b9c..6577b8aeb711 100644
--- a/arch/arm/include/asm/Kbuild
+++ b/arch/arm/include/asm/Kbuild
@@ -24,6 +24,7 @@ generic-y += sembuf.h
 generic-y += serial.h
 generic-y += shmbuf.h
 generic-y += siginfo.h
+generic-y += simd.h
 generic-y += sizes.h
 generic-y += socket.h
 generic-y += sockios.h
diff --git a/arch/arm/include/asm/uaccess.h b/arch/arm/include/asm/uaccess.h
index 7e1f76027f66..72abdc541f38 100644
--- a/arch/arm/include/asm/uaccess.h
+++ b/arch/arm/include/asm/uaccess.h
@@ -19,6 +19,13 @@
 #include <asm/unified.h>
 #include <asm/compiler.h>
 
+#if __LINUX_ARM_ARCH__ < 6
+#include <asm-generic/uaccess-unaligned.h>
+#else
+#define __get_user_unaligned __get_user
+#define __put_user_unaligned __put_user
+#endif
+
 #define VERIFY_READ 0
 #define VERIFY_WRITE 1
 
diff --git a/arch/arm/kernel/entry-common.S b/arch/arm/kernel/entry-common.S
index 74ad15d1a065..bc6bd9683ba4 100644
--- a/arch/arm/kernel/entry-common.S
+++ b/arch/arm/kernel/entry-common.S
@@ -442,10 +442,10 @@ local_restart:
 	ldrcc	pc, [tbl, scno, lsl #2]		@ call sys_* routine
 
 	add	r1, sp, #S_OFF
-	cmp	scno, #(__ARM_NR_BASE - __NR_SYSCALL_BASE)
+2:	cmp	scno, #(__ARM_NR_BASE - __NR_SYSCALL_BASE)
 	eor	r0, scno, #__NR_SYSCALL_BASE	@ put OS number back
 	bcs	arm_syscall
-2:	mov	why, #0				@ no longer a real syscall
+	mov	why, #0				@ no longer a real syscall
 	b	sys_ni_syscall			@ not private func
 
 #if defined(CONFIG_OABI_COMPAT) || !defined(CONFIG_AEABI)
diff --git a/arch/arm/kernel/entry-header.S b/arch/arm/kernel/entry-header.S
index de23a9beed13..39f89fbd5111 100644
--- a/arch/arm/kernel/entry-header.S
+++ b/arch/arm/kernel/entry-header.S
@@ -329,10 +329,10 @@
 #ifdef CONFIG_CONTEXT_TRACKING
 	.if	\save
 	stmdb   sp!, {r0-r3, ip, lr}
-	bl	user_exit
+	bl	context_tracking_user_exit
 	ldmia	sp!, {r0-r3, ip, lr}
 	.else
-	bl	user_exit
+	bl	context_tracking_user_exit
 	.endif
 #endif
 	.endm
@@ -341,10 +341,10 @@
 #ifdef CONFIG_CONTEXT_TRACKING
 	.if	\save
 	stmdb   sp!, {r0-r3, ip, lr}
-	bl	user_enter
+	bl	context_tracking_user_enter
 	ldmia	sp!, {r0-r3, ip, lr}
 	.else
-	bl	user_enter
+	bl	context_tracking_user_enter
 	.endif
 #endif
 	.endm
diff --git a/arch/arm/mach-imx/clk-fixup-mux.c b/arch/arm/mach-imx/clk-fixup-mux.c
index deb4b8093b30..0d40b35c557c 100644
--- a/arch/arm/mach-imx/clk-fixup-mux.c
+++ b/arch/arm/mach-imx/clk-fixup-mux.c
@@ -90,6 +90,7 @@ struct clk *imx_clk_fixup_mux(const char *name, void __iomem *reg,
 	init.ops = &clk_fixup_mux_ops;
 	init.parent_names = parents;
 	init.num_parents = num_parents;
+	init.flags = 0;
 
 	fixup_mux->mux.reg = reg;
 	fixup_mux->mux.shift = shift;
diff --git a/arch/arm/mach-imx/clk-imx27.c b/arch/arm/mach-imx/clk-imx27.c
index c3cfa4116dc0..c6b40f386786 100644
--- a/arch/arm/mach-imx/clk-imx27.c
+++ b/arch/arm/mach-imx/clk-imx27.c
@@ -285,7 +285,7 @@ int __init mx27_clocks_init(unsigned long fref)
 	clk_register_clkdev(clk[ata_ahb_gate], "ata", NULL);
 	clk_register_clkdev(clk[rtc_ipg_gate], NULL, "imx21-rtc");
 	clk_register_clkdev(clk[scc_ipg_gate], "scc", NULL);
-	clk_register_clkdev(clk[cpu_div], NULL, "cpufreq-cpu0.0");
+	clk_register_clkdev(clk[cpu_div], NULL, "cpu0");
 	clk_register_clkdev(clk[emi_ahb_gate], "emi_ahb" , NULL);
 
 	mxc_timer_init(MX27_IO_ADDRESS(MX27_GPT1_BASE_ADDR), MX27_INT_GPT1);
diff --git a/arch/arm/mach-imx/clk-imx51-imx53.c b/arch/arm/mach-imx/clk-imx51-imx53.c
index 1a56a3319997..7c0dc4540aa4 100644
--- a/arch/arm/mach-imx/clk-imx51-imx53.c
+++ b/arch/arm/mach-imx/clk-imx51-imx53.c
@@ -328,7 +328,7 @@ static void __init mx5_clocks_common_init(unsigned long rate_ckil,
 	clk_register_clkdev(clk[ssi2_ipg_gate], NULL, "imx-ssi.1");
 	clk_register_clkdev(clk[ssi3_ipg_gate], NULL, "imx-ssi.2");
 	clk_register_clkdev(clk[sdma_gate], NULL, "imx35-sdma");
-	clk_register_clkdev(clk[cpu_podf], NULL, "cpufreq-cpu0.0");
+	clk_register_clkdev(clk[cpu_podf], NULL, "cpu0");
 	clk_register_clkdev(clk[iim_gate], "iim", NULL);
 	clk_register_clkdev(clk[dummy], NULL, "imx2-wdt.0");
 	clk_register_clkdev(clk[dummy], NULL, "imx2-wdt.1");
@@ -397,7 +397,7 @@ int __init mx51_clocks_init(unsigned long rate_ckil, unsigned long rate_osc,
 				mx51_spdif_xtal_sel, ARRAY_SIZE(mx51_spdif_xtal_sel));
 	clk[spdif1_sel] = imx_clk_mux("spdif1_sel", MXC_CCM_CSCMR2, 2, 2,
 				spdif_sel, ARRAY_SIZE(spdif_sel));
-	clk[spdif1_pred] = imx_clk_divider("spdif1_podf", "spdif1_sel", MXC_CCM_CDCDR, 16, 3);
+	clk[spdif1_pred] = imx_clk_divider("spdif1_pred", "spdif1_sel", MXC_CCM_CDCDR, 16, 3);
 	clk[spdif1_podf] = imx_clk_divider("spdif1_podf", "spdif1_pred", MXC_CCM_CDCDR, 9, 6);
 	clk[spdif1_com_sel] = imx_clk_mux("spdif1_com_sel", MXC_CCM_CSCMR2, 5, 1,
 				mx51_spdif1_com_sel, ARRAY_SIZE(mx51_spdif1_com_sel));
diff --git a/arch/arm/mach-imx/mach-imx6q.c b/arch/arm/mach-imx/mach-imx6q.c
index 85a1b51346c8..90372a21087f 100644
--- a/arch/arm/mach-imx/mach-imx6q.c
+++ b/arch/arm/mach-imx/mach-imx6q.c
@@ -233,10 +233,15 @@ put_node:
 	of_node_put(np);
 }
 
-static void __init imx6q_opp_init(struct device *cpu_dev)
+static void __init imx6q_opp_init(void)
 {
 	struct device_node *np;
+	struct device *cpu_dev = get_cpu_device(0);
 
+	if (!cpu_dev) {
+		pr_warn("failed to get cpu0 device\n");
+		return;
+	}
 	np = of_node_get(cpu_dev->of_node);
 	if (!np) {
 		pr_warn("failed to find cpu0 node\n");
@@ -268,7 +273,7 @@ static void __init imx6q_init_late(void)
 		imx6q_cpuidle_init();
 
 	if (IS_ENABLED(CONFIG_ARM_IMX6Q_CPUFREQ)) {
-		imx6q_opp_init(&imx6q_cpufreq_pdev.dev);
+		imx6q_opp_init();
 		platform_device_register(&imx6q_cpufreq_pdev);
 	}
 }
diff --git a/arch/arm/mach-imx/system.c b/arch/arm/mach-imx/system.c
index 64ff37ea72b1..80c177c36c5f 100644
--- a/arch/arm/mach-imx/system.c
+++ b/arch/arm/mach-imx/system.c
@@ -117,6 +117,17 @@ void __init imx_init_l2cache(void)
 	/* Configure the L2 PREFETCH and POWER registers */
 	val = readl_relaxed(l2x0_base + L2X0_PREFETCH_CTRL);
 	val |= 0x70800000;
+	/*
+	 * The L2 cache controller(PL310) version on the i.MX6D/Q is r3p1-50rel0
+	 * The L2 cache controller(PL310) version on the i.MX6DL/SOLO/SL is r3p2
+	 * But according to ARM PL310 errata: 752271
+	 * ID: 752271: Double linefill feature can cause data corruption
+	 * Fault Status: Present in: r3p0, r3p1, r3p1-50rel0. Fixed in r3p2
+	 * Workaround: The only workaround to this erratum is to disable the
+	 * double linefill feature. This is the default behavior.
+	 */
+	if (cpu_is_imx6q())
+		val &= ~(1 << 30 | 1 << 23);
 	writel_relaxed(val, l2x0_base + L2X0_PREFETCH_CTRL);
 	val = L2X0_DYNAMIC_CLK_GATING_EN | L2X0_STNDBY_MODE_EN;
 	writel_relaxed(val, l2x0_base + L2X0_POWER_CTRL);
diff --git a/arch/arm/mach-omap2/cclock44xx_data.c b/arch/arm/mach-omap2/cclock44xx_data.c
index 1d5b5290d2af..b237950eb8a3 100644
--- a/arch/arm/mach-omap2/cclock44xx_data.c
+++ b/arch/arm/mach-omap2/cclock44xx_data.c
@@ -1632,7 +1632,7 @@ static struct omap_clk omap44xx_clks[] = {
 	CLK(NULL,	"auxclk5_src_ck",		&auxclk5_src_ck),
 	CLK(NULL,	"auxclk5_ck",			&auxclk5_ck),
 	CLK(NULL,	"auxclkreq5_ck",		&auxclkreq5_ck),
-	CLK("omap-gpmc",	"fck",			&dummy_ck),
+	CLK("50000000.gpmc",	"fck",			&dummy_ck),
 	CLK("omap_i2c.1",	"ick",			&dummy_ck),
 	CLK("omap_i2c.2",	"ick",			&dummy_ck),
 	CLK("omap_i2c.3",	"ick",			&dummy_ck),
diff --git a/arch/arm/mach-omap2/cpuidle44xx.c b/arch/arm/mach-omap2/cpuidle44xx.c
index c443f2e97e10..4c8982ae9529 100644
--- a/arch/arm/mach-omap2/cpuidle44xx.c
+++ b/arch/arm/mach-omap2/cpuidle44xx.c
@@ -143,7 +143,7 @@ static int omap_enter_idle_coupled(struct cpuidle_device *dev,
 	 * Call idle CPU cluster PM exit notifier chain
 	 * to restore GIC and wakeupgen context.
 	 */
-	if ((cx->mpu_state == PWRDM_POWER_RET) &&
+	if (dev->cpu == 0 && (cx->mpu_state == PWRDM_POWER_RET) &&
 		(cx->mpu_logic_state == PWRDM_POWER_OFF))
 		cpu_cluster_pm_exit();
 
diff --git a/arch/arm/mach-omap2/gpmc.c b/arch/arm/mach-omap2/gpmc.c
index 9f4795aff48a..579697adaae7 100644
--- a/arch/arm/mach-omap2/gpmc.c
+++ b/arch/arm/mach-omap2/gpmc.c
@@ -1491,8 +1491,8 @@ static int gpmc_probe_generic_child(struct platform_device *pdev,
 	 */
 	ret = gpmc_cs_remap(cs, res.start);
 	if (ret < 0) {
-		dev_err(&pdev->dev, "cannot remap GPMC CS %d to 0x%x\n",
-			cs, res.start);
+		dev_err(&pdev->dev, "cannot remap GPMC CS %d to %pa\n",
+			cs, &res.start);
 		goto err;
 	}
 
diff --git a/arch/arm/mach-omap2/mux34xx.c b/arch/arm/mach-omap2/mux34xx.c
index c53609f46294..be271f1d585b 100644
--- a/arch/arm/mach-omap2/mux34xx.c
+++ b/arch/arm/mach-omap2/mux34xx.c
@@ -620,7 +620,7 @@ static struct omap_mux __initdata omap3_muxmodes[] = {
 		"uart1_rts", "ssi1_flag_tx", NULL, NULL,
 		"gpio_149", NULL, NULL, "safe_mode"),
 	_OMAP3_MUXENTRY(UART1_RX, 151,
-		"uart1_rx", "ss1_wake_tx", "mcbsp1_clkr", "mcspi4_clk",
+		"uart1_rx", "ssi1_wake_tx", "mcbsp1_clkr", "mcspi4_clk",
 		"gpio_151", NULL, NULL, "safe_mode"),
 	_OMAP3_MUXENTRY(UART1_TX, 148,
 		"uart1_tx", "ssi1_dat_tx", NULL, NULL,
diff --git a/arch/arm/mach-omap2/omap-smp.c b/arch/arm/mach-omap2/omap-smp.c
index 8708b2a9da45..891211093295 100644
--- a/arch/arm/mach-omap2/omap-smp.c
+++ b/arch/arm/mach-omap2/omap-smp.c
@@ -1,5 +1,5 @@
 /*
- * OMAP4 SMP source file. It contains platform specific fucntions
+ * OMAP4 SMP source file. It contains platform specific functions
  * needed for the linux smp kernel.
  *
  * Copyright (C) 2009 Texas Instruments, Inc.
diff --git a/arch/arm/mach-omap2/omap_device.c b/arch/arm/mach-omap2/omap_device.c
index f99f68e1e85b..b69dd9abb50a 100644
--- a/arch/arm/mach-omap2/omap_device.c
+++ b/arch/arm/mach-omap2/omap_device.c
@@ -158,7 +158,7 @@ static int omap_device_build_from_dt(struct platform_device *pdev)
 	}
 
 	od = omap_device_alloc(pdev, hwmods, oh_cnt);
-	if (!od) {
+	if (IS_ERR(od)) {
 		dev_err(&pdev->dev, "Cannot allocate omap_device for :%s\n",
 			oh_name);
 		ret = PTR_ERR(od);
diff --git a/arch/arm/mach-sa1100/collie.c b/arch/arm/mach-sa1100/collie.c
index 612a45689770..7fb96ebdc0fb 100644
--- a/arch/arm/mach-sa1100/collie.c
+++ b/arch/arm/mach-sa1100/collie.c
@@ -289,7 +289,7 @@ static void collie_flash_exit(void)
 }
 
 static struct flash_platform_data collie_flash_data = {
-	.map_name	= "cfi_probe",
+	.map_name	= "jedec_probe",
 	.init		= collie_flash_init,
 	.set_vpp	= collie_set_vpp,
 	.exit		= collie_flash_exit,
diff --git a/arch/arm/mach-shmobile/clock-r8a73a4.c b/arch/arm/mach-shmobile/clock-r8a73a4.c
index 8ea5ef6c79cc..5bd2e851e3c7 100644
--- a/arch/arm/mach-shmobile/clock-r8a73a4.c
+++ b/arch/arm/mach-shmobile/clock-r8a73a4.c
@@ -555,7 +555,7 @@ static struct clk_lookup lookups[] = {
 	CLKDEV_CON_ID("pll2h",			&pll2h_clk),
 
 	/* CPU clock */
-	CLKDEV_DEV_ID("cpufreq-cpu0",		&z_clk),
+	CLKDEV_DEV_ID("cpu0",			&z_clk),
 
 	/* DIV6 */
 	CLKDEV_CON_ID("zb",			&div6_clks[DIV6_ZB]),
diff --git a/arch/arm/mach-shmobile/clock-sh73a0.c b/arch/arm/mach-shmobile/clock-sh73a0.c
index 1942eaef5181..c92c023f0d27 100644
--- a/arch/arm/mach-shmobile/clock-sh73a0.c
+++ b/arch/arm/mach-shmobile/clock-sh73a0.c
@@ -616,7 +616,7 @@ static struct clk_lookup lookups[] = {
 	CLKDEV_DEV_ID("smp_twd", &twd_clk), /* smp_twd */
 
 	/* DIV4 clocks */
-	CLKDEV_DEV_ID("cpufreq-cpu0", &div4_clks[DIV4_Z]),
+	CLKDEV_DEV_ID("cpu0", &div4_clks[DIV4_Z]),
 
 	/* DIV6 clocks */
 	CLKDEV_CON_ID("vck1_clk", &div6_clks[DIV6_VCK1]),
diff --git a/arch/arm/mach-u300/Kconfig b/arch/arm/mach-u300/Kconfig
index a85adcd00882..a1659863bfd5 100644
--- a/arch/arm/mach-u300/Kconfig
+++ b/arch/arm/mach-u300/Kconfig
@@ -1,7 +1,3 @@
-menu "ST-Ericsson AB U300/U335 Platform"
-
-comment "ST-Ericsson Mobile Platform Products"
-
 config ARCH_U300
 	bool "ST-Ericsson U300 Series" if ARCH_MULTI_V5
 	depends on MMU
@@ -25,7 +21,9 @@ config ARCH_U300
 	help
 	  Support for ST-Ericsson U300 series mobile platforms.
 
-comment "ST-Ericsson U300/U335 Feature Selections"
+if ARCH_U300
+
+menu "ST-Ericsson AB U300/U335 Platform"
 
 config MACH_U300
 	depends on ARCH_U300
@@ -53,3 +51,5 @@ config MACH_U300_SPIDUMMY
 		SPI framework and ARM PL022 support.
 
 endmenu
+
+endif
diff --git a/arch/arm/mach-ux500/cache-l2x0.c b/arch/arm/mach-ux500/cache-l2x0.c
index 82ccf1d98735..264f894c0e3d 100644
--- a/arch/arm/mach-ux500/cache-l2x0.c
+++ b/arch/arm/mach-ux500/cache-l2x0.c
@@ -69,6 +69,7 @@ static int __init ux500_l2x0_init(void)
 	 * some SMI service available.
 	 */
 	outer_cache.disable = NULL;
+	outer_cache.set_debug = NULL;
 
 	return 0;
 }
diff --git a/arch/arm64/include/asm/hwcap.h b/arch/arm64/include/asm/hwcap.h
index 6d4482fa35bc..e2950b098e76 100644
--- a/arch/arm64/include/asm/hwcap.h
+++ b/arch/arm64/include/asm/hwcap.h
@@ -43,6 +43,6 @@
 				 COMPAT_HWCAP_VFPv3|COMPAT_HWCAP_VFPv4|\
 				 COMPAT_HWCAP_NEON|COMPAT_HWCAP_IDIV)
 
-extern unsigned int elf_hwcap;
+extern unsigned long elf_hwcap;
 #endif
 #endif
diff --git a/arch/arm64/kernel/process.c b/arch/arm64/kernel/process.c
index 57fb55c44c90..7ae8a1f00c3c 100644
--- a/arch/arm64/kernel/process.c
+++ b/arch/arm64/kernel/process.c
@@ -143,15 +143,26 @@ void machine_restart(char *cmd)
 
 void __show_regs(struct pt_regs *regs)
 {
-	int i;
+	int i, top_reg;
+	u64 lr, sp;
+
+	if (compat_user_mode(regs)) {
+		lr = regs->compat_lr;
+		sp = regs->compat_sp;
+		top_reg = 12;
+	} else {
+		lr = regs->regs[30];
+		sp = regs->sp;
+		top_reg = 29;
+	}
 
 	show_regs_print_info(KERN_DEFAULT);
 	print_symbol("PC is at %s\n", instruction_pointer(regs));
-	print_symbol("LR is at %s\n", regs->regs[30]);
+	print_symbol("LR is at %s\n", lr);
 	printk("pc : [<%016llx>] lr : [<%016llx>] pstate: %08llx\n",
-	       regs->pc, regs->regs[30], regs->pstate);
-	printk("sp : %016llx\n", regs->sp);
-	for (i = 29; i >= 0; i--) {
+	       regs->pc, lr, regs->pstate);
+	printk("sp : %016llx\n", sp);
+	for (i = top_reg; i >= 0; i--) {
 		printk("x%-2d: %016llx ", i, regs->regs[i]);
 		if (i % 2 == 0)
 			printk("\n");
diff --git a/arch/arm64/kernel/setup.c b/arch/arm64/kernel/setup.c
index 12ad8f3d0cfd..055cfb80e05c 100644
--- a/arch/arm64/kernel/setup.c
+++ b/arch/arm64/kernel/setup.c
@@ -57,7 +57,7 @@
 unsigned int processor_id;
 EXPORT_SYMBOL(processor_id);
 
-unsigned int elf_hwcap __read_mostly;
+unsigned long elf_hwcap __read_mostly;
 EXPORT_SYMBOL_GPL(elf_hwcap);
 
 static const char *cpu_name;
diff --git a/arch/arm64/mm/fault.c b/arch/arm64/mm/fault.c
index 6d6acf153bff..c23751b06120 100644
--- a/arch/arm64/mm/fault.c
+++ b/arch/arm64/mm/fault.c
@@ -130,7 +130,7 @@ static void __do_user_fault(struct task_struct *tsk, unsigned long addr,
 	force_sig_info(sig, &si, tsk);
 }
 
-void do_bad_area(unsigned long addr, unsigned int esr, struct pt_regs *regs)
+static void do_bad_area(unsigned long addr, unsigned int esr, struct pt_regs *regs)
 {
 	struct task_struct *tsk = current;
 	struct mm_struct *mm = tsk->active_mm;
diff --git a/arch/mips/Makefile b/arch/mips/Makefile
index 75a36ad11ff5..ca8f8340d75f 100644
--- a/arch/mips/Makefile
+++ b/arch/mips/Makefile
@@ -288,9 +288,6 @@ endif
 vmlinux.32: vmlinux
 	$(OBJCOPY) -O $(32bit-bfd) $(OBJCOPYFLAGS) $< $@
 
-
-#obj-$(CONFIG_KPROBES)		+= kprobes.o
-
 #
 # The 64-bit ELF tools are pretty broken so at this time we generate 64-bit
 # ELF files from 32-bit files by conversion.
diff --git a/arch/mips/alchemy/common/usb.c b/arch/mips/alchemy/common/usb.c
index fcc695626117..2adc7edda49c 100644
--- a/arch/mips/alchemy/common/usb.c
+++ b/arch/mips/alchemy/common/usb.c
@@ -14,6 +14,7 @@
 #include <linux/module.h>
 #include <linux/spinlock.h>
 #include <linux/syscore_ops.h>
+#include <asm/cpu.h>
 #include <asm/mach-au1x00/au1000.h>
 
 /* control register offsets */
@@ -358,7 +359,7 @@ static inline int au1200_coherency_bug(void)
 {
 #if defined(CONFIG_DMA_COHERENT)
 	/* Au1200 AB USB does not support coherent memory */
-	if (!(read_c0_prid() & 0xff)) {
+	if (!(read_c0_prid() & PRID_REV_MASK)) {
 		printk(KERN_INFO "Au1200 USB: this is chip revision AB !!\n");
 		printk(KERN_INFO "Au1200 USB: update your board or re-configure"
 				 " the kernel\n");
diff --git a/arch/mips/bcm63xx/cpu.c b/arch/mips/bcm63xx/cpu.c
index 7e17374a9ae8..b713cd64b087 100644
--- a/arch/mips/bcm63xx/cpu.c
+++ b/arch/mips/bcm63xx/cpu.c
@@ -306,14 +306,14 @@ void __init bcm63xx_cpu_init(void)
 
 	switch (c->cputype) {
 	case CPU_BMIPS3300:
-		if ((read_c0_prid() & 0xff00) != PRID_IMP_BMIPS3300_ALT)
+		if ((read_c0_prid() & PRID_IMP_MASK) != PRID_IMP_BMIPS3300_ALT)
 			__cpu_name[cpu] = "Broadcom BCM6338";
 		/* fall-through */
 	case CPU_BMIPS32:
 		chipid_reg = BCM_6345_PERF_BASE;
 		break;
 	case CPU_BMIPS4350:
-		switch ((read_c0_prid() & 0xff)) {
+		switch ((read_c0_prid() & PRID_REV_MASK)) {
 		case 0x04:
 			chipid_reg = BCM_3368_PERF_BASE;
 			break;
diff --git a/arch/mips/boot/dts/include/dt-bindings b/arch/mips/boot/dts/include/dt-bindings
index 68ae3887b3e5..08c00e4972fa 120000
--- a/arch/mips/boot/dts/include/dt-bindings
+++ b/arch/mips/boot/dts/include/dt-bindings
@@ -1 +1 @@
-../../../../../include/dt-bindings
+../../../../../include/dt-bindings
\ No newline at end of file
diff --git a/arch/mips/cavium-octeon/csrc-octeon.c b/arch/mips/cavium-octeon/csrc-octeon.c
index 02193953eb9e..b752c4ed0b79 100644
--- a/arch/mips/cavium-octeon/csrc-octeon.c
+++ b/arch/mips/cavium-octeon/csrc-octeon.c
@@ -12,6 +12,7 @@
 #include <linux/smp.h>
 
 #include <asm/cpu-info.h>
+#include <asm/cpu-type.h>
 #include <asm/time.h>
 
 #include <asm/octeon/octeon.h>
diff --git a/arch/mips/dec/prom/init.c b/arch/mips/dec/prom/init.c
index ab169046e442..468f665de7bb 100644
--- a/arch/mips/dec/prom/init.c
+++ b/arch/mips/dec/prom/init.c
@@ -13,6 +13,7 @@
 
 #include <asm/bootinfo.h>
 #include <asm/cpu.h>
+#include <asm/cpu-type.h>
 #include <asm/processor.h>
 
 #include <asm/dec/prom.h>
diff --git a/arch/mips/include/asm/cpu-features.h b/arch/mips/include/asm/cpu-features.h
index fa44f3ec5302..d445d060e346 100644
--- a/arch/mips/include/asm/cpu-features.h
+++ b/arch/mips/include/asm/cpu-features.h
@@ -13,12 +13,6 @@
 #include <asm/cpu-info.h>
 #include <cpu-feature-overrides.h>
 
-#ifndef current_cpu_type
-#define current_cpu_type()	current_cpu_data.cputype
-#endif
-
-#define boot_cpu_type()		cpu_data[0].cputype
-
 /*
  * SMP assumption: Options of CPU 0 are a superset of all processors.
  * This is true for all known MIPS systems.
@@ -193,7 +187,7 @@
 
 /*
  * MIPS32, MIPS64, VR5500, IDT32332, IDT32334 and maybe a few other
- * pre-MIPS32/MIPS53 processors have CLO, CLZ.	The IDT RC64574 is 64-bit and
+ * pre-MIPS32/MIPS64 processors have CLO, CLZ.	The IDT RC64574 is 64-bit and
  * has CLO and CLZ but not DCLO nor DCLZ.  For 64-bit kernels
  * cpu_has_clo_clz also indicates the availability of DCLO and DCLZ.
  */
diff --git a/arch/mips/include/asm/cpu-info.h b/arch/mips/include/asm/cpu-info.h
index 41401d8eb7d1..21c8e29c8f91 100644
--- a/arch/mips/include/asm/cpu-info.h
+++ b/arch/mips/include/asm/cpu-info.h
@@ -84,6 +84,7 @@ struct cpuinfo_mips {
 extern struct cpuinfo_mips cpu_data[];
 #define current_cpu_data cpu_data[smp_processor_id()]
 #define raw_current_cpu_data cpu_data[raw_smp_processor_id()]
+#define boot_cpu_data cpu_data[0]
 
 extern void cpu_probe(void);
 extern void cpu_report(void);
diff --git a/arch/mips/include/asm/cpu-type.h b/arch/mips/include/asm/cpu-type.h
new file mode 100644
index 000000000000..4a402cc60c03
--- /dev/null
+++ b/arch/mips/include/asm/cpu-type.h
@@ -0,0 +1,203 @@
+/*
+ * This file is subject to the terms and conditions of the GNU General Public
+ * License.  See the file "COPYING" in the main directory of this archive
+ * for more details.
+ *
+ * Copyright (C) 2003, 2004 Ralf Baechle
+ * Copyright (C) 2004  Maciej W. Rozycki
+ */
+#ifndef __ASM_CPU_TYPE_H
+#define __ASM_CPU_TYPE_H
+
+#include <linux/smp.h>
+#include <linux/compiler.h>
+
+static inline int __pure __get_cpu_type(const int cpu_type)
+{
+	switch (cpu_type) {
+#if defined(CONFIG_SYS_HAS_CPU_LOONGSON2E) || \
+    defined(CONFIG_SYS_HAS_CPU_LOONGSON2F)
+	case CPU_LOONGSON2:
+#endif
+
+#ifdef CONFIG_SYS_HAS_CPU_LOONGSON1B
+	case CPU_LOONGSON1:
+#endif
+
+#ifdef CONFIG_SYS_HAS_CPU_MIPS32_R1
+	case CPU_4KC:
+	case CPU_ALCHEMY:
+	case CPU_BMIPS3300:
+	case CPU_BMIPS4350:
+	case CPU_PR4450:
+	case CPU_BMIPS32:
+	case CPU_JZRISC:
+#endif
+
+#if defined(CONFIG_SYS_HAS_CPU_MIPS32_R1) || \
+    defined(CONFIG_SYS_HAS_CPU_MIPS32_R2)
+	case CPU_4KEC:
+#endif
+
+#ifdef CONFIG_SYS_HAS_CPU_MIPS32_R2
+	case CPU_4KSC:
+	case CPU_24K:
+	case CPU_34K:
+	case CPU_1004K:
+	case CPU_74K:
+	case CPU_M14KC:
+	case CPU_M14KEC:
+#endif
+
+#ifdef CONFIG_SYS_HAS_CPU_MIPS64_R1
+	case CPU_5KC:
+	case CPU_5KE:
+	case CPU_20KC:
+	case CPU_25KF:
+	case CPU_SB1:
+	case CPU_SB1A:
+#endif
+
+#ifdef CONFIG_SYS_HAS_CPU_MIPS64_R2
+	/*
+	 * All MIPS64 R2 processors have their own special symbols.  That is,
+	 * there currently is no pure R2 core
+	 */
+#endif
+
+#ifdef CONFIG_SYS_HAS_CPU_R3000
+	case CPU_R2000:
+	case CPU_R3000:
+	case CPU_R3000A:
+	case CPU_R3041:
+	case CPU_R3051:
+	case CPU_R3052:
+	case CPU_R3081:
+	case CPU_R3081E:
+#endif
+
+#ifdef CONFIG_SYS_HAS_CPU_TX39XX
+	case CPU_TX3912:
+	case CPU_TX3922:
+	case CPU_TX3927:
+#endif
+
+#ifdef CONFIG_SYS_HAS_CPU_VR41XX
+	case CPU_VR41XX:
+	case CPU_VR4111:
+	case CPU_VR4121:
+	case CPU_VR4122:
+	case CPU_VR4131:
+	case CPU_VR4133:
+	case CPU_VR4181:
+	case CPU_VR4181A:
+#endif
+
+#ifdef CONFIG_SYS_HAS_CPU_R4300
+	case CPU_R4300:
+	case CPU_R4310:
+#endif
+
+#ifdef CONFIG_SYS_HAS_CPU_R4X00
+	case CPU_R4000PC:
+	case CPU_R4000SC:
+	case CPU_R4000MC:
+	case CPU_R4200:
+	case CPU_R4400PC:
+	case CPU_R4400SC:
+	case CPU_R4400MC:
+	case CPU_R4600:
+	case CPU_R4700:
+	case CPU_R4640:
+	case CPU_R4650:
+#endif
+
+#ifdef CONFIG_SYS_HAS_CPU_TX49XX
+	case CPU_TX49XX:
+#endif
+
+#ifdef CONFIG_SYS_HAS_CPU_R5000
+	case CPU_R5000:
+#endif
+
+#ifdef CONFIG_SYS_HAS_CPU_R5432
+	case CPU_R5432:
+#endif
+
+#ifdef CONFIG_SYS_HAS_CPU_R5500
+	case CPU_R5500:
+#endif
+
+#ifdef CONFIG_SYS_HAS_CPU_R6000
+	case CPU_R6000:
+	case CPU_R6000A:
+#endif
+
+#ifdef CONFIG_SYS_HAS_CPU_NEVADA
+	case CPU_NEVADA:
+#endif
+
+#ifdef CONFIG_SYS_HAS_CPU_R8000
+	case CPU_R8000:
+#endif
+
+#ifdef CONFIG_SYS_HAS_CPU_R10000
+	case CPU_R10000:
+	case CPU_R12000:
+	case CPU_R14000:
+#endif
+#ifdef CONFIG_SYS_HAS_CPU_RM7000
+	case CPU_RM7000:
+	case CPU_SR71000:
+#endif
+#ifdef CONFIG_SYS_HAS_CPU_RM9000
+	case CPU_RM9000:
+#endif
+#ifdef CONFIG_SYS_HAS_CPU_SB1
+	case CPU_SB1:
+	case CPU_SB1A:
+#endif
+#ifdef CONFIG_SYS_HAS_CPU_CAVIUM_OCTEON
+	case CPU_CAVIUM_OCTEON:
+	case CPU_CAVIUM_OCTEON_PLUS:
+	case CPU_CAVIUM_OCTEON2:
+#endif
+
+#ifdef CONFIG_SYS_HAS_CPU_BMIPS4380
+	case CPU_BMIPS4380:
+#endif
+
+#ifdef CONFIG_SYS_HAS_CPU_BMIPS5000
+	case CPU_BMIPS5000:
+#endif
+
+#ifdef CONFIG_SYS_HAS_CPU_XLP
+	case CPU_XLP:
+#endif
+
+#ifdef CONFIG_SYS_HAS_CPU_XLR
+	case CPU_XLR:
+#endif
+		break;
+	default:
+		unreachable();
+	}
+
+	return cpu_type;
+}
+
+static inline int __pure current_cpu_type(void)
+{
+	const int cpu_type = current_cpu_data.cputype;
+
+	return __get_cpu_type(cpu_type);
+}
+
+static inline int __pure boot_cpu_type(void)
+{
+	const int cpu_type = cpu_data[0].cputype;
+
+	return __get_cpu_type(cpu_type);
+}
+
+#endif /* __ASM_CPU_TYPE_H */
diff --git a/arch/mips/include/asm/cpu.h b/arch/mips/include/asm/cpu.h
index 71b9f1998be7..d2035e16502a 100644
--- a/arch/mips/include/asm/cpu.h
+++ b/arch/mips/include/asm/cpu.h
@@ -3,15 +3,14 @@
  *	  various MIPS cpu types.
  *
  * Copyright (C) 1996 David S. Miller (davem@davemloft.net)
- * Copyright (C) 2004  Maciej W. Rozycki
+ * Copyright (C) 2004, 2013  Maciej W. Rozycki
  */
 #ifndef _ASM_CPU_H
 #define _ASM_CPU_H
 
-/* Assigned Company values for bits 23:16 of the PRId Register
-   (CP0 register 15, select 0).	 As of the MIPS32 and MIPS64 specs from
-   MTI, the PRId register is defined in this (backwards compatible)
-   way:
+/*
+   As of the MIPS32 and MIPS64 specs from MTI, the PRId register (CP0
+   register 15, select 0) is defined in this (backwards compatible) way:
 
   +----------------+----------------+----------------+----------------+
   | Company Options| Company ID	    | Processor ID   | Revision	      |
@@ -23,6 +22,14 @@
    spec.
 */
 
+#define PRID_OPT_MASK		0xff000000
+
+/*
+ * Assigned Company values for bits 23:16 of the PRId register.
+ */
+
+#define PRID_COMP_MASK		0xff0000
+
 #define PRID_COMP_LEGACY	0x000000
 #define PRID_COMP_MIPS		0x010000
 #define PRID_COMP_BROADCOM	0x020000
@@ -38,10 +45,17 @@
 #define PRID_COMP_INGENIC	0xd00000
 
 /*
- * Assigned values for the product ID register.	 In order to detect a
- * certain CPU type exactly eventually additional registers may need to
- * be examined.	 These are valid when 23:16 == PRID_COMP_LEGACY
+ * Assigned Processor ID (implementation) values for bits 15:8 of the PRId
+ * register.  In order to detect a certain CPU type exactly eventually
+ * additional registers may need to be examined.
  */
+
+#define PRID_IMP_MASK		0xff00
+
+/*
+ * These are valid when 23:16 == PRID_COMP_LEGACY
+ */
+
 #define PRID_IMP_R2000		0x0100
 #define PRID_IMP_AU1_REV1	0x0100
 #define PRID_IMP_AU1_REV2	0x0200
@@ -182,11 +196,15 @@
 #define PRID_IMP_NETLOGIC_XLP2XX	0x1200
 
 /*
- * Definitions for 7:0 on legacy processors
+ * Particular Revision values for bits 7:0 of the PRId register.
  */
 
 #define PRID_REV_MASK		0x00ff
 
+/*
+ * Definitions for 7:0 on legacy processors
+ */
+
 #define PRID_REV_TX4927		0x0022
 #define PRID_REV_TX4937		0x0030
 #define PRID_REV_R4400		0x0040
@@ -227,6 +245,8 @@
  *  31				   16 15	     8 7	      0
  */
 
+#define FPIR_IMP_MASK		0xff00
+
 #define FPIR_IMP_NONE		0x0000
 
 enum cpu_type_enum {
diff --git a/arch/mips/include/asm/mach-au1x00/au1000.h b/arch/mips/include/asm/mach-au1x00/au1000.h
index 3e11a468cdf8..54f9e84db8ac 100644
--- a/arch/mips/include/asm/mach-au1x00/au1000.h
+++ b/arch/mips/include/asm/mach-au1x00/au1000.h
@@ -43,6 +43,8 @@
 #include <linux/io.h>
 #include <linux/irq.h>
 
+#include <asm/cpu.h>
+
 /* cpu pipeline flush */
 void static inline au_sync(void)
 {
@@ -140,7 +142,7 @@ static inline int au1xxx_cpu_needs_config_od(void)
 
 static inline int alchemy_get_cputype(void)
 {
-	switch (read_c0_prid() & 0xffff0000) {
+	switch (read_c0_prid() & (PRID_OPT_MASK | PRID_COMP_MASK)) {
 	case 0x00030000:
 		return ALCHEMY_CPU_AU1000;
 		break;
diff --git a/arch/mips/include/asm/mach-ip22/cpu-feature-overrides.h b/arch/mips/include/asm/mach-ip22/cpu-feature-overrides.h
index f4caacd25552..1bcb6421205e 100644
--- a/arch/mips/include/asm/mach-ip22/cpu-feature-overrides.h
+++ b/arch/mips/include/asm/mach-ip22/cpu-feature-overrides.h
@@ -8,6 +8,8 @@
 #ifndef __ASM_MACH_IP22_CPU_FEATURE_OVERRIDES_H
 #define __ASM_MACH_IP22_CPU_FEATURE_OVERRIDES_H
 
+#include <asm/cpu.h>
+
 /*
  * IP22 with a variety of processors so we can't use defaults for everything.
  */
diff --git a/arch/mips/include/asm/mach-ip27/cpu-feature-overrides.h b/arch/mips/include/asm/mach-ip27/cpu-feature-overrides.h
index 1d2b6ff60d33..d6111aa2e886 100644
--- a/arch/mips/include/asm/mach-ip27/cpu-feature-overrides.h
+++ b/arch/mips/include/asm/mach-ip27/cpu-feature-overrides.h
@@ -8,6 +8,8 @@
 #ifndef __ASM_MACH_IP27_CPU_FEATURE_OVERRIDES_H
 #define __ASM_MACH_IP27_CPU_FEATURE_OVERRIDES_H
 
+#include <asm/cpu.h>
+
 /*
  * IP27 only comes with R10000 family processors all using the same config
  */
diff --git a/arch/mips/include/asm/mach-ip28/cpu-feature-overrides.h b/arch/mips/include/asm/mach-ip28/cpu-feature-overrides.h
index 65e9c856390d..4cec06d133db 100644
--- a/arch/mips/include/asm/mach-ip28/cpu-feature-overrides.h
+++ b/arch/mips/include/asm/mach-ip28/cpu-feature-overrides.h
@@ -9,6 +9,8 @@
 #ifndef __ASM_MACH_IP28_CPU_FEATURE_OVERRIDES_H
 #define __ASM_MACH_IP28_CPU_FEATURE_OVERRIDES_H
 
+#include <asm/cpu.h>
+
 /*
  * IP28 only comes with R10000 family processors all using the same config
  */
diff --git a/arch/mips/include/asm/mipsregs.h b/arch/mips/include/asm/mipsregs.h
index fed1c3e9b486..e0331414c7d6 100644
--- a/arch/mips/include/asm/mipsregs.h
+++ b/arch/mips/include/asm/mipsregs.h
@@ -603,6 +603,13 @@
 #define MIPS_CONF4_MMUEXTDEF	(_ULCAST_(3) << 14)
 #define MIPS_CONF4_MMUEXTDEF_MMUSIZEEXT (_ULCAST_(1) << 14)
 
+#define MIPS_CONF5_NF		(_ULCAST_(1) << 0)
+#define MIPS_CONF5_UFR		(_ULCAST_(1) << 2)
+#define MIPS_CONF5_MSAEN	(_ULCAST_(1) << 27)
+#define MIPS_CONF5_EVA		(_ULCAST_(1) << 28)
+#define MIPS_CONF5_CV		(_ULCAST_(1) << 29)
+#define MIPS_CONF5_K		(_ULCAST_(1) << 30)
+
 #define MIPS_CONF6_SYND		(_ULCAST_(1) << 13)
 
 #define MIPS_CONF7_WII		(_ULCAST_(1) << 31)
diff --git a/arch/mips/include/asm/pci.h b/arch/mips/include/asm/pci.h
index f194c08bd057..12d6842962be 100644
--- a/arch/mips/include/asm/pci.h
+++ b/arch/mips/include/asm/pci.h
@@ -83,6 +83,18 @@ static inline void pcibios_penalize_isa_irq(int irq, int active)
 extern int pci_mmap_page_range(struct pci_dev *dev, struct vm_area_struct *vma,
 	enum pci_mmap_state mmap_state, int write_combine);
 
+#define HAVE_ARCH_PCI_RESOURCE_TO_USER
+
+static inline void pci_resource_to_user(const struct pci_dev *dev, int bar,
+		const struct resource *rsrc, resource_size_t *start,
+		resource_size_t *end)
+{
+	phys_t size = resource_size(rsrc);
+
+	*start = fixup_bigphys_addr(rsrc->start, size);
+	*end = rsrc->start + size;
+}
+
 /*
  * Dynamic DMA mapping stuff.
  * MIPS has everything mapped statically.
diff --git a/arch/mips/include/asm/timex.h b/arch/mips/include/asm/timex.h
index 6529704aa73a..c5424757da65 100644
--- a/arch/mips/include/asm/timex.h
+++ b/arch/mips/include/asm/timex.h
@@ -10,7 +10,9 @@
 
 #ifdef __KERNEL__
 
+#include <asm/cpu-features.h>
 #include <asm/mipsregs.h>
+#include <asm/cpu-type.h>
 
 /*
  * This is the clock rate of the i8253 PIT.  A MIPS system may not have
@@ -33,9 +35,38 @@
 
 typedef unsigned int cycles_t;
 
+/*
+ * On R4000/R4400 before version 5.0 an erratum exists such that if the
+ * cycle counter is read in the exact moment that it is matching the
+ * compare register, no interrupt will be generated.
+ *
+ * There is a suggested workaround and also the erratum can't strike if
+ * the compare interrupt isn't being used as the clock source device.
+ * However for now the implementaton of this function doesn't get these
+ * fine details right.
+ */
 static inline cycles_t get_cycles(void)
 {
-	return 0;
+	switch (boot_cpu_type()) {
+	case CPU_R4400PC:
+	case CPU_R4400SC:
+	case CPU_R4400MC:
+		if ((read_c0_prid() & 0xff) >= 0x0050)
+			return read_c0_count();
+		break;
+
+        case CPU_R4000PC:
+        case CPU_R4000SC:
+        case CPU_R4000MC:
+		break;
+
+	default:
+		if (cpu_has_counter)
+			return read_c0_count();
+		break;
+	}
+
+	return 0;	/* no usable counter */
 }
 
 #endif /* __KERNEL__ */
diff --git a/arch/mips/include/asm/vga.h b/arch/mips/include/asm/vga.h
index f4cff7e4fa8a..f82c83749a08 100644
--- a/arch/mips/include/asm/vga.h
+++ b/arch/mips/include/asm/vga.h
@@ -6,6 +6,7 @@
 #ifndef _ASM_VGA_H
 #define _ASM_VGA_H
 
+#include <asm/addrspace.h>
 #include <asm/byteorder.h>
 
 /*
@@ -13,7 +14,7 @@
  *	access the videoram directly without any black magic.
  */
 
-#define VGA_MAP_MEM(x, s)	(0xb0000000L + (unsigned long)(x))
+#define VGA_MAP_MEM(x, s)	CKSEG1ADDR(0x10000000L + (unsigned long)(x))
 
 #define vga_readb(x)	(*(x))
 #define vga_writeb(x, y)	(*(y) = (x))
diff --git a/arch/mips/kernel/cpu-probe.c b/arch/mips/kernel/cpu-probe.c
index 37663c7862a5..5465dc183e5a 100644
--- a/arch/mips/kernel/cpu-probe.c
+++ b/arch/mips/kernel/cpu-probe.c
@@ -20,6 +20,7 @@
 
 #include <asm/bugs.h>
 #include <asm/cpu.h>
+#include <asm/cpu-type.h>
 #include <asm/fpu.h>
 #include <asm/mipsregs.h>
 #include <asm/watch.h>
@@ -55,7 +56,7 @@ static inline void check_errata(void)
 {
 	struct cpuinfo_mips *c = &current_cpu_data;
 
-	switch (c->cputype) {
+	switch (current_cpu_type()) {
 	case CPU_34K:
 		/*
 		 * Erratum "RPS May Cause Incorrect Instruction Execution"
@@ -122,7 +123,7 @@ static inline unsigned long cpu_get_fpu_id(void)
  */
 static inline int __cpu_has_fpu(void)
 {
-	return ((cpu_get_fpu_id() & 0xff00) != FPIR_IMP_NONE);
+	return ((cpu_get_fpu_id() & FPIR_IMP_MASK) != FPIR_IMP_NONE);
 }
 
 static inline void cpu_probe_vmbits(struct cpuinfo_mips *c)
@@ -290,6 +291,17 @@ static inline unsigned int decode_config4(struct cpuinfo_mips *c)
 	return config4 & MIPS_CONF_M;
 }
 
+static inline unsigned int decode_config5(struct cpuinfo_mips *c)
+{
+	unsigned int config5;
+
+	config5 = read_c0_config5();
+	config5 &= ~MIPS_CONF5_UFR;
+	write_c0_config5(config5);
+
+	return config5 & MIPS_CONF_M;
+}
+
 static void decode_configs(struct cpuinfo_mips *c)
 {
 	int ok;
@@ -310,6 +322,8 @@ static void decode_configs(struct cpuinfo_mips *c)
 		ok = decode_config3(c);
 	if (ok)
 		ok = decode_config4(c);
+	if (ok)
+		ok = decode_config5(c);
 
 	mips_probe_watch_registers(c);
 
@@ -322,7 +336,7 @@ static void decode_configs(struct cpuinfo_mips *c)
 
 static inline void cpu_probe_legacy(struct cpuinfo_mips *c, unsigned int cpu)
 {
-	switch (c->processor_id & 0xff00) {
+	switch (c->processor_id & PRID_IMP_MASK) {
 	case PRID_IMP_R2000:
 		c->cputype = CPU_R2000;
 		__cpu_name[cpu] = "R2000";
@@ -333,7 +347,7 @@ static inline void cpu_probe_legacy(struct cpuinfo_mips *c, unsigned int cpu)
 		c->tlbsize = 64;
 		break;
 	case PRID_IMP_R3000:
-		if ((c->processor_id & 0xff) == PRID_REV_R3000A) {
+		if ((c->processor_id & PRID_REV_MASK) == PRID_REV_R3000A) {
 			if (cpu_has_confreg()) {
 				c->cputype = CPU_R3081E;
 				__cpu_name[cpu] = "R3081";
@@ -353,7 +367,8 @@ static inline void cpu_probe_legacy(struct cpuinfo_mips *c, unsigned int cpu)
 		break;
 	case PRID_IMP_R4000:
 		if (read_c0_config() & CONF_SC) {
-			if ((c->processor_id & 0xff) >= PRID_REV_R4400) {
+			if ((c->processor_id & PRID_REV_MASK) >=
+			    PRID_REV_R4400) {
 				c->cputype = CPU_R4400PC;
 				__cpu_name[cpu] = "R4400PC";
 			} else {
@@ -361,7 +376,8 @@ static inline void cpu_probe_legacy(struct cpuinfo_mips *c, unsigned int cpu)
 				__cpu_name[cpu] = "R4000PC";
 			}
 		} else {
-			if ((c->processor_id & 0xff) >= PRID_REV_R4400) {
+			if ((c->processor_id & PRID_REV_MASK) >=
+			    PRID_REV_R4400) {
 				c->cputype = CPU_R4400SC;
 				__cpu_name[cpu] = "R4400SC";
 			} else {
@@ -454,7 +470,7 @@ static inline void cpu_probe_legacy(struct cpuinfo_mips *c, unsigned int cpu)
 			__cpu_name[cpu] = "TX3927";
 			c->tlbsize = 64;
 		} else {
-			switch (c->processor_id & 0xff) {
+			switch (c->processor_id & PRID_REV_MASK) {
 			case PRID_REV_TX3912:
 				c->cputype = CPU_TX3912;
 				__cpu_name[cpu] = "TX3912";
@@ -640,7 +656,7 @@ static inline void cpu_probe_legacy(struct cpuinfo_mips *c, unsigned int cpu)
 static inline void cpu_probe_mips(struct cpuinfo_mips *c, unsigned int cpu)
 {
 	decode_configs(c);
-	switch (c->processor_id & 0xff00) {
+	switch (c->processor_id & PRID_IMP_MASK) {
 	case PRID_IMP_4KC:
 		c->cputype = CPU_4KC;
 		__cpu_name[cpu] = "MIPS 4Kc";
@@ -711,7 +727,7 @@ static inline void cpu_probe_mips(struct cpuinfo_mips *c, unsigned int cpu)
 static inline void cpu_probe_alchemy(struct cpuinfo_mips *c, unsigned int cpu)
 {
 	decode_configs(c);
-	switch (c->processor_id & 0xff00) {
+	switch (c->processor_id & PRID_IMP_MASK) {
 	case PRID_IMP_AU1_REV1:
 	case PRID_IMP_AU1_REV2:
 		c->cputype = CPU_ALCHEMY;
@@ -730,7 +746,7 @@ static inline void cpu_probe_alchemy(struct cpuinfo_mips *c, unsigned int cpu)
 			break;
 		case 4:
 			__cpu_name[cpu] = "Au1200";
-			if ((c->processor_id & 0xff) == 2)
+			if ((c->processor_id & PRID_REV_MASK) == 2)
 				__cpu_name[cpu] = "Au1250";
 			break;
 		case 5:
@@ -748,12 +764,12 @@ static inline void cpu_probe_sibyte(struct cpuinfo_mips *c, unsigned int cpu)
 {
 	decode_configs(c);
 
-	switch (c->processor_id & 0xff00) {
+	switch (c->processor_id & PRID_IMP_MASK) {
 	case PRID_IMP_SB1:
 		c->cputype = CPU_SB1;
 		__cpu_name[cpu] = "SiByte SB1";
 		/* FPU in pass1 is known to have issues. */
-		if ((c->processor_id & 0xff) < 0x02)
+		if ((c->processor_id & PRID_REV_MASK) < 0x02)
 			c->options &= ~(MIPS_CPU_FPU | MIPS_CPU_32FPR);
 		break;
 	case PRID_IMP_SB1A:
@@ -766,7 +782,7 @@ static inline void cpu_probe_sibyte(struct cpuinfo_mips *c, unsigned int cpu)
 static inline void cpu_probe_sandcraft(struct cpuinfo_mips *c, unsigned int cpu)
 {
 	decode_configs(c);
-	switch (c->processor_id & 0xff00) {
+	switch (c->processor_id & PRID_IMP_MASK) {
 	case PRID_IMP_SR71000:
 		c->cputype = CPU_SR71000;
 		__cpu_name[cpu] = "Sandcraft SR71000";
@@ -779,7 +795,7 @@ static inline void cpu_probe_sandcraft(struct cpuinfo_mips *c, unsigned int cpu)
 static inline void cpu_probe_nxp(struct cpuinfo_mips *c, unsigned int cpu)
 {
 	decode_configs(c);
-	switch (c->processor_id & 0xff00) {
+	switch (c->processor_id & PRID_IMP_MASK) {
 	case PRID_IMP_PR4450:
 		c->cputype = CPU_PR4450;
 		__cpu_name[cpu] = "Philips PR4450";
@@ -791,7 +807,7 @@ static inline void cpu_probe_nxp(struct cpuinfo_mips *c, unsigned int cpu)
 static inline void cpu_probe_broadcom(struct cpuinfo_mips *c, unsigned int cpu)
 {
 	decode_configs(c);
-	switch (c->processor_id & 0xff00) {
+	switch (c->processor_id & PRID_IMP_MASK) {
 	case PRID_IMP_BMIPS32_REV4:
 	case PRID_IMP_BMIPS32_REV8:
 		c->cputype = CPU_BMIPS32;
@@ -806,7 +822,7 @@ static inline void cpu_probe_broadcom(struct cpuinfo_mips *c, unsigned int cpu)
 		set_elf_platform(cpu, "bmips3300");
 		break;
 	case PRID_IMP_BMIPS43XX: {
-		int rev = c->processor_id & 0xff;
+		int rev = c->processor_id & PRID_REV_MASK;
 
 		if (rev >= PRID_REV_BMIPS4380_LO &&
 				rev <= PRID_REV_BMIPS4380_HI) {
@@ -832,7 +848,7 @@ static inline void cpu_probe_broadcom(struct cpuinfo_mips *c, unsigned int cpu)
 static inline void cpu_probe_cavium(struct cpuinfo_mips *c, unsigned int cpu)
 {
 	decode_configs(c);
-	switch (c->processor_id & 0xff00) {
+	switch (c->processor_id & PRID_IMP_MASK) {
 	case PRID_IMP_CAVIUM_CN38XX:
 	case PRID_IMP_CAVIUM_CN31XX:
 	case PRID_IMP_CAVIUM_CN30XX:
@@ -875,7 +891,7 @@ static inline void cpu_probe_ingenic(struct cpuinfo_mips *c, unsigned int cpu)
 	decode_configs(c);
 	/* JZRISC does not implement the CP0 counter. */
 	c->options &= ~MIPS_CPU_COUNTER;
-	switch (c->processor_id & 0xff00) {
+	switch (c->processor_id & PRID_IMP_MASK) {
 	case PRID_IMP_JZRISC:
 		c->cputype = CPU_JZRISC;
 		__cpu_name[cpu] = "Ingenic JZRISC";
@@ -890,7 +906,7 @@ static inline void cpu_probe_netlogic(struct cpuinfo_mips *c, int cpu)
 {
 	decode_configs(c);
 
-	if ((c->processor_id & 0xff00) == PRID_IMP_NETLOGIC_AU13XX) {
+	if ((c->processor_id & PRID_IMP_MASK) == PRID_IMP_NETLOGIC_AU13XX) {
 		c->cputype = CPU_ALCHEMY;
 		__cpu_name[cpu] = "Au1300";
 		/* following stuff is not for Alchemy */
@@ -905,7 +921,7 @@ static inline void cpu_probe_netlogic(struct cpuinfo_mips *c, int cpu)
 			MIPS_CPU_EJTAG	 |
 			MIPS_CPU_LLSC);
 
-	switch (c->processor_id & 0xff00) {
+	switch (c->processor_id & PRID_IMP_MASK) {
 	case PRID_IMP_NETLOGIC_XLP2XX:
 		c->cputype = CPU_XLP;
 		__cpu_name[cpu] = "Broadcom XLPII";
@@ -984,7 +1000,7 @@ void cpu_probe(void)
 	c->cputype	= CPU_UNKNOWN;
 
 	c->processor_id = read_c0_prid();
-	switch (c->processor_id & 0xff0000) {
+	switch (c->processor_id & PRID_COMP_MASK) {
 	case PRID_COMP_LEGACY:
 		cpu_probe_legacy(c, cpu);
 		break;
diff --git a/arch/mips/kernel/idle.c b/arch/mips/kernel/idle.c
index 42f8875d2444..f7991d95bff9 100644
--- a/arch/mips/kernel/idle.c
+++ b/arch/mips/kernel/idle.c
@@ -18,6 +18,7 @@
 #include <linux/sched.h>
 #include <asm/cpu.h>
 #include <asm/cpu-info.h>
+#include <asm/cpu-type.h>
 #include <asm/idle.h>
 #include <asm/mipsregs.h>
 
@@ -136,7 +137,7 @@ void __init check_wait(void)
 		return;
 	}
 
-	switch (c->cputype) {
+	switch (current_cpu_type()) {
 	case CPU_R3081:
 	case CPU_R3081E:
 		cpu_wait = r3081_wait;
diff --git a/arch/mips/kernel/time.c b/arch/mips/kernel/time.c
index 364d26ae4215..dcb8e5d3bb8a 100644
--- a/arch/mips/kernel/time.c
+++ b/arch/mips/kernel/time.c
@@ -24,6 +24,7 @@
 #include <linux/export.h>
 
 #include <asm/cpu-features.h>
+#include <asm/cpu-type.h>
 #include <asm/div64.h>
 #include <asm/smtc_ipi.h>
 #include <asm/time.h>
diff --git a/arch/mips/kernel/traps.c b/arch/mips/kernel/traps.c
index aec3408edd4b..524841f02803 100644
--- a/arch/mips/kernel/traps.c
+++ b/arch/mips/kernel/traps.c
@@ -39,6 +39,7 @@
 #include <asm/break.h>
 #include <asm/cop2.h>
 #include <asm/cpu.h>
+#include <asm/cpu-type.h>
 #include <asm/dsp.h>
 #include <asm/fpu.h>
 #include <asm/fpu_emulator.h>
@@ -622,7 +623,7 @@ static int simulate_rdhwr(struct pt_regs *regs, int rd, int rt)
 		regs->regs[rt] = read_c0_count();
 		return 0;
 	case 3:		/* Count register resolution */
-		switch (current_cpu_data.cputype) {
+		switch (current_cpu_type()) {
 		case CPU_20KC:
 		case CPU_25KF:
 			regs->regs[rt] = 1;
diff --git a/arch/mips/mm/c-octeon.c b/arch/mips/mm/c-octeon.c
index 729e7702b1de..c8efdb5b6ee0 100644
--- a/arch/mips/mm/c-octeon.c
+++ b/arch/mips/mm/c-octeon.c
@@ -19,6 +19,7 @@
 #include <asm/bootinfo.h>
 #include <asm/cacheops.h>
 #include <asm/cpu-features.h>
+#include <asm/cpu-type.h>
 #include <asm/page.h>
 #include <asm/pgtable.h>
 #include <asm/r4kcache.h>
@@ -186,9 +187,10 @@ static void probe_octeon(void)
 	unsigned long dcache_size;
 	unsigned int config1;
 	struct cpuinfo_mips *c = &current_cpu_data;
+	int cputype = current_cpu_type();
 
 	config1 = read_c0_config1();
-	switch (c->cputype) {
+	switch (cputype) {
 	case CPU_CAVIUM_OCTEON:
 	case CPU_CAVIUM_OCTEON_PLUS:
 		c->icache.linesz = 2 << ((config1 >> 19) & 7);
@@ -199,7 +201,7 @@ static void probe_octeon(void)
 			c->icache.sets * c->icache.ways * c->icache.linesz;
 		c->icache.waybit = ffs(icache_size / c->icache.ways) - 1;
 		c->dcache.linesz = 128;
-		if (c->cputype == CPU_CAVIUM_OCTEON_PLUS)
+		if (cputype == CPU_CAVIUM_OCTEON_PLUS)
 			c->dcache.sets = 2; /* CN5XXX has two Dcache sets */
 		else
 			c->dcache.sets = 1; /* CN3XXX has one Dcache set */
diff --git a/arch/mips/mm/c-r4k.c b/arch/mips/mm/c-r4k.c
index f749f687ee87..627883bc6d5f 100644
--- a/arch/mips/mm/c-r4k.c
+++ b/arch/mips/mm/c-r4k.c
@@ -12,6 +12,7 @@
 #include <linux/highmem.h>
 #include <linux/kernel.h>
 #include <linux/linkage.h>
+#include <linux/preempt.h>
 #include <linux/sched.h>
 #include <linux/smp.h>
 #include <linux/mm.h>
@@ -24,6 +25,7 @@
 #include <asm/cacheops.h>
 #include <asm/cpu.h>
 #include <asm/cpu-features.h>
+#include <asm/cpu-type.h>
 #include <asm/io.h>
 #include <asm/page.h>
 #include <asm/pgtable.h>
@@ -601,6 +603,7 @@ static void r4k_dma_cache_wback_inv(unsigned long addr, unsigned long size)
 	/* Catch bad driver code */
 	BUG_ON(size == 0);
 
+	preempt_disable();
 	if (cpu_has_inclusive_pcaches) {
 		if (size >= scache_size)
 			r4k_blast_scache();
@@ -621,6 +624,7 @@ static void r4k_dma_cache_wback_inv(unsigned long addr, unsigned long size)
 		R4600_HIT_CACHEOP_WAR_IMPL;
 		blast_dcache_range(addr, addr + size);
 	}
+	preempt_enable();
 
 	bc_wback_inv(addr, size);
 	__sync();
@@ -631,6 +635,7 @@ static void r4k_dma_cache_inv(unsigned long addr, unsigned long size)
 	/* Catch bad driver code */
 	BUG_ON(size == 0);
 
+	preempt_disable();
 	if (cpu_has_inclusive_pcaches) {
 		if (size >= scache_size)
 			r4k_blast_scache();
@@ -655,6 +660,7 @@ static void r4k_dma_cache_inv(unsigned long addr, unsigned long size)
 		R4600_HIT_CACHEOP_WAR_IMPL;
 		blast_inv_dcache_range(addr, addr + size);
 	}
+	preempt_enable();
 
 	bc_inv(addr, size);
 	__sync();
@@ -780,20 +786,30 @@ static inline void rm7k_erratum31(void)
 
 static inline void alias_74k_erratum(struct cpuinfo_mips *c)
 {
+	unsigned int imp = c->processor_id & PRID_IMP_MASK;
+	unsigned int rev = c->processor_id & PRID_REV_MASK;
+
 	/*
 	 * Early versions of the 74K do not update the cache tags on a
 	 * vtag miss/ptag hit which can occur in the case of KSEG0/KUSEG
 	 * aliases. In this case it is better to treat the cache as always
 	 * having aliases.
 	 */
-	if ((c->processor_id & 0xff) <= PRID_REV_ENCODE_332(2, 4, 0))
-		c->dcache.flags |= MIPS_CACHE_VTAG;
-	if ((c->processor_id & 0xff) == PRID_REV_ENCODE_332(2, 4, 0))
-		write_c0_config6(read_c0_config6() | MIPS_CONF6_SYND);
-	if (((c->processor_id & 0xff00) == PRID_IMP_1074K) &&
-	    ((c->processor_id & 0xff) <= PRID_REV_ENCODE_332(1, 1, 0))) {
-		c->dcache.flags |= MIPS_CACHE_VTAG;
-		write_c0_config6(read_c0_config6() | MIPS_CONF6_SYND);
+	switch (imp) {
+	case PRID_IMP_74K:
+		if (rev <= PRID_REV_ENCODE_332(2, 4, 0))
+			c->dcache.flags |= MIPS_CACHE_VTAG;
+		if (rev == PRID_REV_ENCODE_332(2, 4, 0))
+			write_c0_config6(read_c0_config6() | MIPS_CONF6_SYND);
+		break;
+	case PRID_IMP_1074K:
+		if (rev <= PRID_REV_ENCODE_332(1, 1, 0)) {
+			c->dcache.flags |= MIPS_CACHE_VTAG;
+			write_c0_config6(read_c0_config6() | MIPS_CONF6_SYND);
+		}
+		break;
+	default:
+		BUG();
 	}
 }
 
@@ -809,7 +825,7 @@ static void probe_pcache(void)
 	unsigned long config1;
 	unsigned int lsize;
 
-	switch (c->cputype) {
+	switch (current_cpu_type()) {
 	case CPU_R4600:			/* QED style two way caches? */
 	case CPU_R4700:
 	case CPU_R5000:
@@ -1025,7 +1041,8 @@ static void probe_pcache(void)
 	 * presumably no vendor is shipping his hardware in the "bad"
 	 * configuration.
 	 */
-	if ((prid & 0xff00) == PRID_IMP_R4000 && (prid & 0xff) < 0x40 &&
+	if ((prid & PRID_IMP_MASK) == PRID_IMP_R4000 &&
+	    (prid & PRID_REV_MASK) < PRID_REV_R4400 &&
 	    !(config & CONF_SC) && c->icache.linesz != 16 &&
 	    PAGE_SIZE <= 0x8000)
 		panic("Improper R4000SC processor configuration detected");
@@ -1045,7 +1062,7 @@ static void probe_pcache(void)
 	 * normally they'd suffer from aliases but magic in the hardware deals
 	 * with that for us so we don't need to take care ourselves.
 	 */
-	switch (c->cputype) {
+	switch (current_cpu_type()) {
 	case CPU_20KC:
 	case CPU_25KF:
 	case CPU_SB1:
@@ -1065,7 +1082,7 @@ static void probe_pcache(void)
 	case CPU_34K:
 	case CPU_74K:
 	case CPU_1004K:
-		if (c->cputype == CPU_74K)
+		if (current_cpu_type() == CPU_74K)
 			alias_74k_erratum(c);
 		if ((read_c0_config7() & (1 << 16))) {
 			/* effectively physically indexed dcache,
@@ -1078,7 +1095,7 @@ static void probe_pcache(void)
 			c->dcache.flags |= MIPS_CACHE_ALIASES;
 	}
 
-	switch (c->cputype) {
+	switch (current_cpu_type()) {
 	case CPU_20KC:
 		/*
 		 * Some older 20Kc chips doesn't have the 'VI' bit in
@@ -1207,7 +1224,7 @@ static void setup_scache(void)
 	 * processors don't have a S-cache that would be relevant to the
 	 * Linux memory management.
 	 */
-	switch (c->cputype) {
+	switch (current_cpu_type()) {
 	case CPU_R4000SC:
 	case CPU_R4000MC:
 	case CPU_R4400SC:
@@ -1384,9 +1401,8 @@ static void r4k_cache_error_setup(void)
 {
 	extern char __weak except_vec2_generic;
 	extern char __weak except_vec2_sb1;
-	struct cpuinfo_mips *c = &current_cpu_data;
 
-	switch (c->cputype) {
+	switch (current_cpu_type()) {
 	case CPU_SB1:
 	case CPU_SB1A:
 		set_uncached_handler(0x100, &except_vec2_sb1, 0x80);
diff --git a/arch/mips/mm/dma-default.c b/arch/mips/mm/dma-default.c
index 664e523653d0..5f8b95512580 100644
--- a/arch/mips/mm/dma-default.c
+++ b/arch/mips/mm/dma-default.c
@@ -18,6 +18,7 @@
 #include <linux/highmem.h>
 
 #include <asm/cache.h>
+#include <asm/cpu-type.h>
 #include <asm/io.h>
 
 #include <dma-coherence.h>
@@ -307,12 +308,10 @@ static void mips_dma_sync_sg_for_cpu(struct device *dev,
 {
 	int i;
 
-	/* Make sure that gcc doesn't leave the empty loop body.  */
-	for (i = 0; i < nelems; i++, sg++) {
-		if (cpu_needs_post_dma_flush(dev))
+	if (cpu_needs_post_dma_flush(dev))
+		for (i = 0; i < nelems; i++, sg++)
 			__dma_sync(sg_page(sg), sg->offset, sg->length,
 				   direction);
-	}
 }
 
 static void mips_dma_sync_sg_for_device(struct device *dev,
@@ -320,12 +319,10 @@ static void mips_dma_sync_sg_for_device(struct device *dev,
 {
 	int i;
 
-	/* Make sure that gcc doesn't leave the empty loop body.  */
-	for (i = 0; i < nelems; i++, sg++) {
-		if (!plat_device_is_coherent(dev))
+	if (!plat_device_is_coherent(dev))
+		for (i = 0; i < nelems; i++, sg++)
 			__dma_sync(sg_page(sg), sg->offset, sg->length,
 				   direction);
-	}
 }
 
 int mips_dma_mapping_error(struct device *dev, dma_addr_t dma_addr)
diff --git a/arch/mips/mm/page.c b/arch/mips/mm/page.c
index 218c2109a55d..cbd81d17793a 100644
--- a/arch/mips/mm/page.c
+++ b/arch/mips/mm/page.c
@@ -18,6 +18,7 @@
 
 #include <asm/bugs.h>
 #include <asm/cacheops.h>
+#include <asm/cpu-type.h>
 #include <asm/inst.h>
 #include <asm/io.h>
 #include <asm/page.h>
diff --git a/arch/mips/mm/sc-mips.c b/arch/mips/mm/sc-mips.c
index 5d01392e3518..08d05aee8788 100644
--- a/arch/mips/mm/sc-mips.c
+++ b/arch/mips/mm/sc-mips.c
@@ -6,6 +6,7 @@
 #include <linux/sched.h>
 #include <linux/mm.h>
 
+#include <asm/cpu-type.h>
 #include <asm/mipsregs.h>
 #include <asm/bcache.h>
 #include <asm/cacheops.h>
@@ -71,7 +72,7 @@ static inline int mips_sc_is_activated(struct cpuinfo_mips *c)
 	unsigned int tmp;
 
 	/* Check the bypass bit (L2B) */
-	switch (c->cputype) {
+	switch (current_cpu_type()) {
 	case CPU_34K:
 	case CPU_74K:
 	case CPU_1004K:
diff --git a/arch/mips/mm/tlb-r4k.c b/arch/mips/mm/tlb-r4k.c
index 00b26a67a06d..bb3a5f643e97 100644
--- a/arch/mips/mm/tlb-r4k.c
+++ b/arch/mips/mm/tlb-r4k.c
@@ -16,6 +16,7 @@
 #include <linux/module.h>
 
 #include <asm/cpu.h>
+#include <asm/cpu-type.h>
 #include <asm/bootinfo.h>
 #include <asm/mmu_context.h>
 #include <asm/pgtable.h>
diff --git a/arch/mips/mm/tlbex.c b/arch/mips/mm/tlbex.c
index 821b45175dc1..9bb3a9363b06 100644
--- a/arch/mips/mm/tlbex.c
+++ b/arch/mips/mm/tlbex.c
@@ -30,6 +30,7 @@
 #include <linux/cache.h>
 
 #include <asm/cacheflush.h>
+#include <asm/cpu-type.h>
 #include <asm/pgtable.h>
 #include <asm/war.h>
 #include <asm/uasm.h>
diff --git a/arch/mips/mti-malta/malta-time.c b/arch/mips/mti-malta/malta-time.c
index 53aad4a35375..a18af5fce67e 100644
--- a/arch/mips/mti-malta/malta-time.c
+++ b/arch/mips/mti-malta/malta-time.c
@@ -27,6 +27,7 @@
 #include <linux/timex.h>
 #include <linux/mc146818rtc.h>
 
+#include <asm/cpu.h>
 #include <asm/mipsregs.h>
 #include <asm/mipsmtregs.h>
 #include <asm/hardirq.h>
@@ -76,7 +77,7 @@ static void __init estimate_frequencies(void)
 #endif
 
 #if defined (CONFIG_KVM_GUEST) && defined (CONFIG_KVM_HOST_FREQ)
-	unsigned int prid = read_c0_prid() & 0xffff00;
+	unsigned int prid = read_c0_prid() & (PRID_COMP_MASK | PRID_IMP_MASK);
 
 	/*
 	 * XXXKYMA: hardwire the CPU frequency to Host Freq/4
@@ -169,7 +170,7 @@ unsigned int get_c0_compare_int(void)
 
 void __init plat_time_init(void)
 {
-	unsigned int prid = read_c0_prid() & 0xffff00;
+	unsigned int prid = read_c0_prid() & (PRID_COMP_MASK | PRID_IMP_MASK);
 	unsigned int freq;
 
 	estimate_frequencies();
diff --git a/arch/mips/mti-sead3/sead3-time.c b/arch/mips/mti-sead3/sead3-time.c
index a43ea3cc0a3b..552d26c34386 100644
--- a/arch/mips/mti-sead3/sead3-time.c
+++ b/arch/mips/mti-sead3/sead3-time.c
@@ -7,6 +7,7 @@
  */
 #include <linux/init.h>
 
+#include <asm/cpu.h>
 #include <asm/setup.h>
 #include <asm/time.h>
 #include <asm/irq.h>
@@ -34,7 +35,7 @@ static void __iomem *status_reg = (void __iomem *)0xbf000410;
  */
 static unsigned int __init estimate_cpu_frequency(void)
 {
-	unsigned int prid = read_c0_prid() & 0xffff00;
+	unsigned int prid = read_c0_prid() & (PRID_COMP_MASK | PRID_IMP_MASK);
 	unsigned int tick = 0;
 	unsigned int freq;
 	unsigned int orig;
diff --git a/arch/mips/netlogic/xlr/fmn-config.c b/arch/mips/netlogic/xlr/fmn-config.c
index ed3bf0e3f309..c7622c6e5f67 100644
--- a/arch/mips/netlogic/xlr/fmn-config.c
+++ b/arch/mips/netlogic/xlr/fmn-config.c
@@ -36,6 +36,7 @@
 #include <linux/irq.h>
 #include <linux/interrupt.h>
 
+#include <asm/cpu.h>
 #include <asm/mipsregs.h>
 #include <asm/netlogic/xlr/fmn.h>
 #include <asm/netlogic/xlr/xlr.h>
@@ -187,7 +188,7 @@ void xlr_board_info_setup(void)
 	int processor_id, num_core;
 
 	num_core = hweight32(nlm_current_node()->coremask);
-	processor_id = read_c0_prid() & 0xff00;
+	processor_id = read_c0_prid() & PRID_IMP_MASK;
 
 	setup_cpu_fmninfo(cpu, num_core);
 	switch (processor_id) {
diff --git a/arch/mips/oprofile/common.c b/arch/mips/oprofile/common.c
index 5e5424753b56..4d1736fc1955 100644
--- a/arch/mips/oprofile/common.c
+++ b/arch/mips/oprofile/common.c
@@ -12,6 +12,7 @@
 #include <linux/oprofile.h>
 #include <linux/smp.h>
 #include <asm/cpu-info.h>
+#include <asm/cpu-type.h>
 
 #include "op_impl.h"
 
diff --git a/arch/mips/pci/pci-bcm1480.c b/arch/mips/pci/pci-bcm1480.c
index 44dd5aa2e36f..5ec2a7bae02c 100644
--- a/arch/mips/pci/pci-bcm1480.c
+++ b/arch/mips/pci/pci-bcm1480.c
@@ -39,6 +39,7 @@
 #include <linux/mm.h>
 #include <linux/console.h>
 #include <linux/tty.h>
+#include <linux/vt.h>
 
 #include <asm/sibyte/bcm1480_regs.h>
 #include <asm/sibyte/bcm1480_scd.h>
diff --git a/arch/mips/sibyte/bcm1480/setup.c b/arch/mips/sibyte/bcm1480/setup.c
index 05ed92c92b69..8e2e04f77870 100644
--- a/arch/mips/sibyte/bcm1480/setup.c
+++ b/arch/mips/sibyte/bcm1480/setup.c
@@ -22,6 +22,7 @@
 #include <linux/string.h>
 
 #include <asm/bootinfo.h>
+#include <asm/cpu.h>
 #include <asm/mipsregs.h>
 #include <asm/io.h>
 #include <asm/sibyte/sb1250.h>
@@ -119,7 +120,7 @@ void __init bcm1480_setup(void)
 	uint64_t sys_rev;
 	int plldiv;
 
-	sb1_pass = read_c0_prid() & 0xff;
+	sb1_pass = read_c0_prid() & PRID_REV_MASK;
 	sys_rev = __raw_readq(IOADDR(A_SCD_SYSTEM_REVISION));
 	soc_type = SYS_SOC_TYPE(sys_rev);
 	part_type = G_SYS_PART(sys_rev);
diff --git a/arch/mips/sibyte/sb1250/setup.c b/arch/mips/sibyte/sb1250/setup.c
index a14bd4cb0bc0..3c02b2a77ae9 100644
--- a/arch/mips/sibyte/sb1250/setup.c
+++ b/arch/mips/sibyte/sb1250/setup.c
@@ -22,6 +22,7 @@
 #include <linux/string.h>
 
 #include <asm/bootinfo.h>
+#include <asm/cpu.h>
 #include <asm/mipsregs.h>
 #include <asm/io.h>
 #include <asm/sibyte/sb1250.h>
@@ -182,7 +183,7 @@ void __init sb1250_setup(void)
 	int plldiv;
 	int bad_config = 0;
 
-	sb1_pass = read_c0_prid() & 0xff;
+	sb1_pass = read_c0_prid() & PRID_REV_MASK;
 	sys_rev = __raw_readq(IOADDR(A_SCD_SYSTEM_REVISION));
 	soc_type = SYS_SOC_TYPE(sys_rev);
 	soc_pass = G_SYS_REVISION(sys_rev);
diff --git a/arch/mips/sni/setup.c b/arch/mips/sni/setup.c
index 5b09b3544edd..efad85c8c823 100644
--- a/arch/mips/sni/setup.c
+++ b/arch/mips/sni/setup.c
@@ -25,6 +25,7 @@
 #endif
 
 #include <asm/bootinfo.h>
+#include <asm/cpu.h>
 #include <asm/io.h>
 #include <asm/reboot.h>
 #include <asm/sni.h>
@@ -173,7 +174,7 @@ void __init plat_mem_setup(void)
 		system_type = "RM300-Cxx";
 		break;
 	case SNI_BRD_PCI_DESKTOP:
-		switch (read_c0_prid() & 0xff00) {
+		switch (read_c0_prid() & PRID_IMP_MASK) {
 		case PRID_IMP_R4600:
 		case PRID_IMP_R4700:
 			system_type = "RM200-C20";
diff --git a/arch/openrisc/include/asm/prom.h b/arch/openrisc/include/asm/prom.h
index eb59bfe23e85..93c9980e1b6b 100644
--- a/arch/openrisc/include/asm/prom.h
+++ b/arch/openrisc/include/asm/prom.h
@@ -14,53 +14,9 @@
  * the Free Software Foundation; either version 2 of the License, or
  * (at your option) any later version.
  */
-
-#include <linux/of.h>	/* linux/of.h gets to determine #include ordering */
-
 #ifndef _ASM_OPENRISC_PROM_H
 #define _ASM_OPENRISC_PROM_H
-#ifdef __KERNEL__
-#ifndef __ASSEMBLY__
 
-#include <linux/types.h>
-#include <asm/irq.h>
-#include <linux/irqdomain.h>
-#include <linux/atomic.h>
-#include <linux/of_irq.h>
-#include <linux/of_fdt.h>
-#include <linux/of_address.h>
-#include <linux/proc_fs.h>
-#include <linux/platform_device.h>
 #define HAVE_ARCH_DEVTREE_FIXUPS
 
-/* Other Prototypes */
-extern int early_uartlite_console(void);
-
-/* Parse the ibm,dma-window property of an OF node into the busno, phys and
- * size parameters.
- */
-void of_parse_dma_window(struct device_node *dn, const void *dma_window_prop,
-		unsigned long *busno, unsigned long *phys, unsigned long *size);
-
-extern void kdump_move_device_tree(void);
-
-/* Get the MAC address */
-extern const void *of_get_mac_address(struct device_node *np);
-
-/**
- * of_irq_map_pci - Resolve the interrupt for a PCI device
- * @pdev:	the device whose interrupt is to be resolved
- * @out_irq:	structure of_irq filled by this function
- *
- * This function resolves the PCI interrupt for a given PCI device. If a
- * device-node exists for a given pci_dev, it will use normal OF tree
- * walking. If not, it will implement standard swizzling and walk up the
- * PCI tree until an device-node is found, at which point it will finish
- * resolving using the OF tree walking.
- */
-struct pci_dev;
-extern int of_irq_map_pci(struct pci_dev *pdev, struct of_irq *out_irq);
-
-#endif /* __ASSEMBLY__ */
-#endif /* __KERNEL__ */
 #endif /* _ASM_OPENRISC_PROM_H */
diff --git a/arch/powerpc/boot/Makefile b/arch/powerpc/boot/Makefile
index 6a15c968d214..15ca2255f438 100644
--- a/arch/powerpc/boot/Makefile
+++ b/arch/powerpc/boot/Makefile
@@ -74,7 +74,7 @@ src-wlib-$(CONFIG_8xx) += mpc8xx.c planetcore.c
 src-wlib-$(CONFIG_PPC_82xx) += pq2.c fsl-soc.c planetcore.c
 src-wlib-$(CONFIG_EMBEDDED6xx) += mv64x60.c mv64x60_i2c.c ugecon.c
 
-src-plat-y := of.c
+src-plat-y := of.c epapr.c
 src-plat-$(CONFIG_40x) += fixed-head.S ep405.c cuboot-hotfoot.c \
 				treeboot-walnut.c cuboot-acadia.c \
 				cuboot-kilauea.c simpleboot.c \
@@ -97,7 +97,7 @@ src-plat-$(CONFIG_EMBEDDED6xx) += cuboot-pq2.c cuboot-mpc7448hpc2.c \
 					prpmc2800.c
 src-plat-$(CONFIG_AMIGAONE) += cuboot-amigaone.c
 src-plat-$(CONFIG_PPC_PS3) += ps3-head.S ps3-hvcall.S ps3.c
-src-plat-$(CONFIG_EPAPR_BOOT) += epapr.c
+src-plat-$(CONFIG_EPAPR_BOOT) += epapr.c epapr-wrapper.c
 
 src-wlib := $(sort $(src-wlib-y))
 src-plat := $(sort $(src-plat-y))
diff --git a/arch/powerpc/boot/epapr-wrapper.c b/arch/powerpc/boot/epapr-wrapper.c
new file mode 100644
index 000000000000..c10191006673
--- /dev/null
+++ b/arch/powerpc/boot/epapr-wrapper.c
@@ -0,0 +1,9 @@
+extern void epapr_platform_init(unsigned long r3, unsigned long r4,
+				unsigned long r5, unsigned long r6,
+				unsigned long r7);
+
+void platform_init(unsigned long r3, unsigned long r4, unsigned long r5,
+		   unsigned long r6, unsigned long r7)
+{
+	epapr_platform_init(r3, r4, r5, r6, r7);
+}
diff --git a/arch/powerpc/boot/epapr.c b/arch/powerpc/boot/epapr.c
index 06c1961bd124..02e91aa2194a 100644
--- a/arch/powerpc/boot/epapr.c
+++ b/arch/powerpc/boot/epapr.c
@@ -48,8 +48,8 @@ static void platform_fixups(void)
 		       fdt_addr, fdt_totalsize((void *)fdt_addr), ima_size);
 }
 
-void platform_init(unsigned long r3, unsigned long r4, unsigned long r5,
-		   unsigned long r6, unsigned long r7)
+void epapr_platform_init(unsigned long r3, unsigned long r4, unsigned long r5,
+			 unsigned long r6, unsigned long r7)
 {
 	epapr_magic = r6;
 	ima_size = r7;
diff --git a/arch/powerpc/boot/of.c b/arch/powerpc/boot/of.c
index 61d9899aa0d0..62e2f43ec1df 100644
--- a/arch/powerpc/boot/of.c
+++ b/arch/powerpc/boot/of.c
@@ -26,6 +26,9 @@
 
 static unsigned long claim_base;
 
+void epapr_platform_init(unsigned long r3, unsigned long r4, unsigned long r5,
+			 unsigned long r6, unsigned long r7);
+
 static void *of_try_claim(unsigned long size)
 {
 	unsigned long addr = 0;
@@ -61,7 +64,7 @@ static void of_image_hdr(const void *hdr)
 	}
 }
 
-void platform_init(unsigned long a1, unsigned long a2, void *promptr)
+static void of_platform_init(unsigned long a1, unsigned long a2, void *promptr)
 {
 	platform_ops.image_hdr = of_image_hdr;
 	platform_ops.malloc = of_try_claim;
@@ -81,3 +84,14 @@ void platform_init(unsigned long a1, unsigned long a2, void *promptr)
 		loader_info.initrd_size = a2;
 	}
 }
+
+void platform_init(unsigned long r3, unsigned long r4, unsigned long r5,
+		   unsigned long r6, unsigned long r7)
+{
+	/* Detect OF vs. ePAPR boot */
+	if (r5)
+		of_platform_init(r3, r4, (void *)r5);
+	else
+		epapr_platform_init(r3, r4, r5, r6, r7);
+}
+
diff --git a/arch/powerpc/boot/wrapper b/arch/powerpc/boot/wrapper
index 6761c746048d..cd7af841ba05 100755
--- a/arch/powerpc/boot/wrapper
+++ b/arch/powerpc/boot/wrapper
@@ -148,18 +148,18 @@ make_space=y
 
 case "$platform" in
 pseries)
-    platformo=$object/of.o
+    platformo="$object/of.o $object/epapr.o"
     link_address='0x4000000'
     ;;
 maple)
-    platformo=$object/of.o
+    platformo="$object/of.o $object/epapr.o"
     link_address='0x400000'
     ;;
 pmac|chrp)
-    platformo=$object/of.o
+    platformo="$object/of.o $object/epapr.o"
     ;;
 coff)
-    platformo="$object/crt0.o $object/of.o"
+    platformo="$object/crt0.o $object/of.o $object/epapr.o"
     lds=$object/zImage.coff.lds
     link_address='0x500000'
     pie=
@@ -253,6 +253,7 @@ treeboot-iss4xx-mpic)
     platformo="$object/treeboot-iss4xx.o"
     ;;
 epapr)
+    platformo="$object/epapr.o $object/epapr-wrapper.o"
     link_address='0x20000000'
     pie=-pie
     ;;
diff --git a/arch/powerpc/include/asm/irq.h b/arch/powerpc/include/asm/irq.h
index 0e40843a1c6e..41f13cec8a8f 100644
--- a/arch/powerpc/include/asm/irq.h
+++ b/arch/powerpc/include/asm/irq.h
@@ -69,9 +69,9 @@ extern struct thread_info *softirq_ctx[NR_CPUS];
 
 extern void irq_ctx_init(void);
 extern void call_do_softirq(struct thread_info *tp);
-extern int call_handle_irq(int irq, void *p1,
-			   struct thread_info *tp, void *func);
+extern void call_do_irq(struct pt_regs *regs, struct thread_info *tp);
 extern void do_IRQ(struct pt_regs *regs);
+extern void __do_irq(struct pt_regs *regs);
 
 int irq_choose_cpu(const struct cpumask *mask);
 
diff --git a/arch/powerpc/include/asm/processor.h b/arch/powerpc/include/asm/processor.h
index e378cccfca55..ce4de5aed7b5 100644
--- a/arch/powerpc/include/asm/processor.h
+++ b/arch/powerpc/include/asm/processor.h
@@ -149,8 +149,6 @@ typedef struct {
 
 struct thread_struct {
 	unsigned long	ksp;		/* Kernel stack pointer */
-	unsigned long	ksp_limit;	/* if ksp <= ksp_limit stack overflow */
-
 #ifdef CONFIG_PPC64
 	unsigned long	ksp_vsid;
 #endif
@@ -162,6 +160,7 @@ struct thread_struct {
 #endif
 #ifdef CONFIG_PPC32
 	void		*pgdir;		/* root of page-table tree */
+	unsigned long	ksp_limit;	/* if ksp <= ksp_limit stack overflow */
 #endif
 #ifdef CONFIG_PPC_ADV_DEBUG_REGS
 	/*
@@ -321,7 +320,6 @@ struct thread_struct {
 #else
 #define INIT_THREAD  { \
 	.ksp = INIT_SP, \
-	.ksp_limit = INIT_SP_LIMIT, \
 	.regs = (struct pt_regs *)INIT_SP - 1, /* XXX bogus, I think */ \
 	.fs = KERNEL_DS, \
 	.fpr = {{0}}, \
diff --git a/arch/powerpc/kernel/asm-offsets.c b/arch/powerpc/kernel/asm-offsets.c
index d8958be5f31a..502c7a4e73f7 100644
--- a/arch/powerpc/kernel/asm-offsets.c
+++ b/arch/powerpc/kernel/asm-offsets.c
@@ -80,10 +80,11 @@ int main(void)
 	DEFINE(TASKTHREADPPR, offsetof(struct task_struct, thread.ppr));
 #else
 	DEFINE(THREAD_INFO, offsetof(struct task_struct, stack));
+	DEFINE(THREAD_INFO_GAP, _ALIGN_UP(sizeof(struct thread_info), 16));
+	DEFINE(KSP_LIMIT, offsetof(struct thread_struct, ksp_limit));
 #endif /* CONFIG_PPC64 */
 
 	DEFINE(KSP, offsetof(struct thread_struct, ksp));
-	DEFINE(KSP_LIMIT, offsetof(struct thread_struct, ksp_limit));
 	DEFINE(PT_REGS, offsetof(struct thread_struct, regs));
 #ifdef CONFIG_BOOKE
 	DEFINE(THREAD_NORMSAVES, offsetof(struct thread_struct, normsave[0]));
diff --git a/arch/powerpc/kernel/irq.c b/arch/powerpc/kernel/irq.c
index c69440cef7af..57d286a78f86 100644
--- a/arch/powerpc/kernel/irq.c
+++ b/arch/powerpc/kernel/irq.c
@@ -441,50 +441,6 @@ void migrate_irqs(void)
 }
 #endif
 
-static inline void handle_one_irq(unsigned int irq)
-{
-	struct thread_info *curtp, *irqtp;
-	unsigned long saved_sp_limit;
-	struct irq_desc *desc;
-
-	desc = irq_to_desc(irq);
-	if (!desc)
-		return;
-
-	/* Switch to the irq stack to handle this */
-	curtp = current_thread_info();
-	irqtp = hardirq_ctx[smp_processor_id()];
-
-	if (curtp == irqtp) {
-		/* We're already on the irq stack, just handle it */
-		desc->handle_irq(irq, desc);
-		return;
-	}
-
-	saved_sp_limit = current->thread.ksp_limit;
-
-	irqtp->task = curtp->task;
-	irqtp->flags = 0;
-
-	/* Copy the softirq bits in preempt_count so that the
-	 * softirq checks work in the hardirq context. */
-	irqtp->preempt_count = (irqtp->preempt_count & ~SOFTIRQ_MASK) |
-			       (curtp->preempt_count & SOFTIRQ_MASK);
-
-	current->thread.ksp_limit = (unsigned long)irqtp +
-		_ALIGN_UP(sizeof(struct thread_info), 16);
-
-	call_handle_irq(irq, desc, irqtp, desc->handle_irq);
-	current->thread.ksp_limit = saved_sp_limit;
-	irqtp->task = NULL;
-
-	/* Set any flag that may have been set on the
-	 * alternate stack
-	 */
-	if (irqtp->flags)
-		set_bits(irqtp->flags, &curtp->flags);
-}
-
 static inline void check_stack_overflow(void)
 {
 #ifdef CONFIG_DEBUG_STACKOVERFLOW
@@ -501,9 +457,9 @@ static inline void check_stack_overflow(void)
 #endif
 }
 
-void do_IRQ(struct pt_regs *regs)
+void __do_irq(struct pt_regs *regs)
 {
-	struct pt_regs *old_regs = set_irq_regs(regs);
+	struct irq_desc *desc;
 	unsigned int irq;
 
 	irq_enter();
@@ -519,18 +475,56 @@ void do_IRQ(struct pt_regs *regs)
 	 */
 	irq = ppc_md.get_irq();
 
-	/* We can hard enable interrupts now */
+	/* We can hard enable interrupts now to allow perf interrupts */
 	may_hard_irq_enable();
 
 	/* And finally process it */
-	if (irq != NO_IRQ)
-		handle_one_irq(irq);
-	else
+	if (unlikely(irq == NO_IRQ))
 		__get_cpu_var(irq_stat).spurious_irqs++;
+	else {
+		desc = irq_to_desc(irq);
+		if (likely(desc))
+			desc->handle_irq(irq, desc);
+	}
 
 	trace_irq_exit(regs);
 
 	irq_exit();
+}
+
+void do_IRQ(struct pt_regs *regs)
+{
+	struct pt_regs *old_regs = set_irq_regs(regs);
+	struct thread_info *curtp, *irqtp;
+
+	/* Switch to the irq stack to handle this */
+	curtp = current_thread_info();
+	irqtp = hardirq_ctx[raw_smp_processor_id()];
+
+	/* Already there ? */
+	if (unlikely(curtp == irqtp)) {
+		__do_irq(regs);
+		set_irq_regs(old_regs);
+		return;
+	}
+
+	/* Prepare the thread_info in the irq stack */
+	irqtp->task = curtp->task;
+	irqtp->flags = 0;
+
+	/* Copy the preempt_count so that the [soft]irq checks work. */
+	irqtp->preempt_count = curtp->preempt_count;
+
+	/* Switch stack and call */
+	call_do_irq(regs, irqtp);
+
+	/* Restore stack limit */
+	irqtp->task = NULL;
+
+	/* Copy back updates to the thread_info */
+	if (irqtp->flags)
+		set_bits(irqtp->flags, &curtp->flags);
+
 	set_irq_regs(old_regs);
 }
 
@@ -592,28 +586,22 @@ void irq_ctx_init(void)
 		memset((void *)softirq_ctx[i], 0, THREAD_SIZE);
 		tp = softirq_ctx[i];
 		tp->cpu = i;
-		tp->preempt_count = 0;
 
 		memset((void *)hardirq_ctx[i], 0, THREAD_SIZE);
 		tp = hardirq_ctx[i];
 		tp->cpu = i;
-		tp->preempt_count = HARDIRQ_OFFSET;
 	}
 }
 
 static inline void do_softirq_onstack(void)
 {
 	struct thread_info *curtp, *irqtp;
-	unsigned long saved_sp_limit = current->thread.ksp_limit;
 
 	curtp = current_thread_info();
 	irqtp = softirq_ctx[smp_processor_id()];
 	irqtp->task = curtp->task;
 	irqtp->flags = 0;
-	current->thread.ksp_limit = (unsigned long)irqtp +
-				    _ALIGN_UP(sizeof(struct thread_info), 16);
 	call_do_softirq(irqtp);
-	current->thread.ksp_limit = saved_sp_limit;
 	irqtp->task = NULL;
 
 	/* Set any flag that may have been set on the
diff --git a/arch/powerpc/kernel/misc_32.S b/arch/powerpc/kernel/misc_32.S
index 777d999f563b..2b0ad9845363 100644
--- a/arch/powerpc/kernel/misc_32.S
+++ b/arch/powerpc/kernel/misc_32.S
@@ -36,26 +36,41 @@
 
 	.text
 
+/*
+ * We store the saved ksp_limit in the unused part
+ * of the STACK_FRAME_OVERHEAD
+ */
 _GLOBAL(call_do_softirq)
 	mflr	r0
 	stw	r0,4(r1)
+	lwz	r10,THREAD+KSP_LIMIT(r2)
+	addi	r11,r3,THREAD_INFO_GAP
 	stwu	r1,THREAD_SIZE-STACK_FRAME_OVERHEAD(r3)
 	mr	r1,r3
+	stw	r10,8(r1)
+	stw	r11,THREAD+KSP_LIMIT(r2)
 	bl	__do_softirq
+	lwz	r10,8(r1)
 	lwz	r1,0(r1)
 	lwz	r0,4(r1)
+	stw	r10,THREAD+KSP_LIMIT(r2)
 	mtlr	r0
 	blr
 
-_GLOBAL(call_handle_irq)
+_GLOBAL(call_do_irq)
 	mflr	r0
 	stw	r0,4(r1)
-	mtctr	r6
-	stwu	r1,THREAD_SIZE-STACK_FRAME_OVERHEAD(r5)
-	mr	r1,r5
-	bctrl
+	lwz	r10,THREAD+KSP_LIMIT(r2)
+	addi	r11,r3,THREAD_INFO_GAP
+	stwu	r1,THREAD_SIZE-STACK_FRAME_OVERHEAD(r4)
+	mr	r1,r4
+	stw	r10,8(r1)
+	stw	r11,THREAD+KSP_LIMIT(r2)
+	bl	__do_irq
+	lwz	r10,8(r1)
 	lwz	r1,0(r1)
 	lwz	r0,4(r1)
+	stw	r10,THREAD+KSP_LIMIT(r2)
 	mtlr	r0
 	blr
 
diff --git a/arch/powerpc/kernel/misc_64.S b/arch/powerpc/kernel/misc_64.S
index 971d7e78aff2..e59caf874d05 100644
--- a/arch/powerpc/kernel/misc_64.S
+++ b/arch/powerpc/kernel/misc_64.S
@@ -40,14 +40,12 @@ _GLOBAL(call_do_softirq)
 	mtlr	r0
 	blr
 
-_GLOBAL(call_handle_irq)
-	ld	r8,0(r6)
+_GLOBAL(call_do_irq)
 	mflr	r0
 	std	r0,16(r1)
-	mtctr	r8
-	stdu	r1,THREAD_SIZE-STACK_FRAME_OVERHEAD(r5)
-	mr	r1,r5
-	bctrl
+	stdu	r1,THREAD_SIZE-STACK_FRAME_OVERHEAD(r4)
+	mr	r1,r4
+	bl	.__do_irq
 	ld	r1,0(r1)
 	ld	r0,16(r1)
 	mtlr	r0
diff --git a/arch/powerpc/kernel/process.c b/arch/powerpc/kernel/process.c
index 6f428da53e20..96d2fdf3aa9e 100644
--- a/arch/powerpc/kernel/process.c
+++ b/arch/powerpc/kernel/process.c
@@ -1000,9 +1000,10 @@ int copy_thread(unsigned long clone_flags, unsigned long usp,
 	kregs = (struct pt_regs *) sp;
 	sp -= STACK_FRAME_OVERHEAD;
 	p->thread.ksp = sp;
+#ifdef CONFIG_PPC32
 	p->thread.ksp_limit = (unsigned long)task_stack_page(p) +
 				_ALIGN_UP(sizeof(struct thread_info), 16);
-
+#endif
 #ifdef CONFIG_HAVE_HW_BREAKPOINT
 	p->thread.ptrace_bps[0] = NULL;
 #endif
diff --git a/arch/powerpc/kernel/prom_init.c b/arch/powerpc/kernel/prom_init.c
index 12e656ffe60e..5fe2842e8bab 100644
--- a/arch/powerpc/kernel/prom_init.c
+++ b/arch/powerpc/kernel/prom_init.c
@@ -196,6 +196,8 @@ static int __initdata mem_reserve_cnt;
 
 static cell_t __initdata regbuf[1024];
 
+static bool rtas_has_query_cpu_stopped;
+
 
 /*
  * Error results ... some OF calls will return "-1" on error, some
@@ -1574,6 +1576,11 @@ static void __init prom_instantiate_rtas(void)
 	prom_setprop(rtas_node, "/rtas", "linux,rtas-entry",
 		     &val, sizeof(val));
 
+	/* Check if it supports "query-cpu-stopped-state" */
+	if (prom_getprop(rtas_node, "query-cpu-stopped-state",
+			 &val, sizeof(val)) != PROM_ERROR)
+		rtas_has_query_cpu_stopped = true;
+
 #if defined(CONFIG_PPC_POWERNV) && defined(__BIG_ENDIAN__)
 	/* PowerVN takeover hack */
 	prom_rtas_data = base;
@@ -1815,6 +1822,18 @@ static void __init prom_hold_cpus(void)
 		= (void *) LOW_ADDR(__secondary_hold_acknowledge);
 	unsigned long secondary_hold = LOW_ADDR(__secondary_hold);
 
+	/*
+	 * On pseries, if RTAS supports "query-cpu-stopped-state",
+	 * we skip this stage, the CPUs will be started by the
+	 * kernel using RTAS.
+	 */
+	if ((of_platform == PLATFORM_PSERIES ||
+	     of_platform == PLATFORM_PSERIES_LPAR) &&
+	    rtas_has_query_cpu_stopped) {
+		prom_printf("prom_hold_cpus: skipped\n");
+		return;
+	}
+
 	prom_debug("prom_hold_cpus: start...\n");
 	prom_debug("    1) spinloop       = 0x%x\n", (unsigned long)spinloop);
 	prom_debug("    1) *spinloop      = 0x%x\n", *spinloop);
@@ -3011,6 +3030,8 @@ unsigned long __init prom_init(unsigned long r3, unsigned long r4,
 	 * On non-powermacs, put all CPUs in spin-loops.
 	 *
 	 * PowerMacs use a different mechanism to spin CPUs
+	 *
+	 * (This must be done after instanciating RTAS)
 	 */
 	if (of_platform != PLATFORM_POWERMAC &&
 	    of_platform != PLATFORM_OPAL)
diff --git a/arch/powerpc/lib/sstep.c b/arch/powerpc/lib/sstep.c
index a7ee978fb860..b1faa1593c90 100644
--- a/arch/powerpc/lib/sstep.c
+++ b/arch/powerpc/lib/sstep.c
@@ -1505,6 +1505,7 @@ int __kprobes emulate_step(struct pt_regs *regs, unsigned int instr)
 		 */
 		if ((ra == 1) && !(regs->msr & MSR_PR) \
 			&& (val3 >= (regs->gpr[1] - STACK_INT_FRAME_SIZE))) {
+#ifdef CONFIG_PPC32
 			/*
 			 * Check if we will touch kernel sack overflow
 			 */
@@ -1513,7 +1514,7 @@ int __kprobes emulate_step(struct pt_regs *regs, unsigned int instr)
 				err = -EINVAL;
 				break;
 			}
-
+#endif /* CONFIG_PPC32 */
 			/*
 			 * Check if we already set since that means we'll
 			 * lose the previous value.
diff --git a/arch/powerpc/platforms/pseries/smp.c b/arch/powerpc/platforms/pseries/smp.c
index 1c1771a40250..24f58cb0a543 100644
--- a/arch/powerpc/platforms/pseries/smp.c
+++ b/arch/powerpc/platforms/pseries/smp.c
@@ -233,18 +233,24 @@ static void __init smp_init_pseries(void)
 
 	alloc_bootmem_cpumask_var(&of_spin_mask);
 
-	/* Mark threads which are still spinning in hold loops. */
-	if (cpu_has_feature(CPU_FTR_SMT)) {
-		for_each_present_cpu(i) { 
-			if (cpu_thread_in_core(i) == 0)
-				cpumask_set_cpu(i, of_spin_mask);
-		}
-	} else {
-		cpumask_copy(of_spin_mask, cpu_present_mask);
+	/*
+	 * Mark threads which are still spinning in hold loops
+	 *
+	 * We know prom_init will not have started them if RTAS supports
+	 * query-cpu-stopped-state.
+	 */
+	if (rtas_token("query-cpu-stopped-state") == RTAS_UNKNOWN_SERVICE) {
+		if (cpu_has_feature(CPU_FTR_SMT)) {
+			for_each_present_cpu(i) {
+				if (cpu_thread_in_core(i) == 0)
+					cpumask_set_cpu(i, of_spin_mask);
+			}
+		} else
+			cpumask_copy(of_spin_mask, cpu_present_mask);
+
+		cpumask_clear_cpu(boot_cpuid, of_spin_mask);
 	}
 
-	cpumask_clear_cpu(boot_cpuid, of_spin_mask);
-
 	/* Non-lpar has additional take/give timebase */
 	if (rtas_token("freeze-time-base") != RTAS_UNKNOWN_SERVICE) {
 		smp_ops->give_timebase = rtas_give_timebase;
diff --git a/arch/s390/Kconfig b/arch/s390/Kconfig
index dcc6ac2d8026..7143793859fa 100644
--- a/arch/s390/Kconfig
+++ b/arch/s390/Kconfig
@@ -93,6 +93,7 @@ config S390
 	select ARCH_INLINE_WRITE_UNLOCK_IRQ
 	select ARCH_INLINE_WRITE_UNLOCK_IRQRESTORE
 	select ARCH_SAVE_PAGE_KEYS if HIBERNATION
+	select ARCH_USE_CMPXCHG_LOCKREF
 	select ARCH_WANT_IPC_PARSE_VERSION
 	select BUILDTIME_EXTABLE_SORT
 	select CLONE_BACKWARDS2
@@ -102,7 +103,6 @@ config S390
 	select GENERIC_TIME_VSYSCALL_OLD
 	select HAVE_ALIGNED_STRUCT_PAGE if SLUB
 	select HAVE_ARCH_JUMP_LABEL if !MARCH_G5
-	select HAVE_ARCH_MUTEX_CPU_RELAX
 	select HAVE_ARCH_SECCOMP_FILTER
 	select HAVE_ARCH_TRACEHOOK
 	select HAVE_ARCH_TRANSPARENT_HUGEPAGE if 64BIT
diff --git a/arch/s390/include/asm/mutex.h b/arch/s390/include/asm/mutex.h
index 688271f5f2e4..458c1f7fbc18 100644
--- a/arch/s390/include/asm/mutex.h
+++ b/arch/s390/include/asm/mutex.h
@@ -7,5 +7,3 @@
  */
 
 #include <asm-generic/mutex-dec.h>
-
-#define arch_mutex_cpu_relax()	barrier()
diff --git a/arch/s390/include/asm/processor.h b/arch/s390/include/asm/processor.h
index 0eb37505cab1..ca7821f07260 100644
--- a/arch/s390/include/asm/processor.h
+++ b/arch/s390/include/asm/processor.h
@@ -198,6 +198,8 @@ static inline void cpu_relax(void)
 	barrier();
 }
 
+#define arch_mutex_cpu_relax()  barrier()
+
 static inline void psw_set_key(unsigned int key)
 {
 	asm volatile("spka 0(%0)" : : "d" (key));
diff --git a/arch/s390/include/asm/spinlock.h b/arch/s390/include/asm/spinlock.h
index 701fe8c59e1f..83e5d216105e 100644
--- a/arch/s390/include/asm/spinlock.h
+++ b/arch/s390/include/asm/spinlock.h
@@ -44,6 +44,11 @@ extern void arch_spin_lock_wait_flags(arch_spinlock_t *, unsigned long flags);
 extern int arch_spin_trylock_retry(arch_spinlock_t *);
 extern void arch_spin_relax(arch_spinlock_t *lock);
 
+static inline int arch_spin_value_unlocked(arch_spinlock_t lock)
+{
+	return lock.owner_cpu == 0;
+}
+
 static inline void arch_spin_lock(arch_spinlock_t *lp)
 {
 	int old;
diff --git a/arch/tile/Kconfig b/arch/tile/Kconfig
index 8a7cc663b3f8..d45a2c48f185 100644
--- a/arch/tile/Kconfig
+++ b/arch/tile/Kconfig
@@ -361,7 +361,7 @@ config CMDLINE_OVERRIDE
 
 config VMALLOC_RESERVE
 	hex
-	default 0x1000000
+	default 0x2000000
 
 config HARDWALL
 	bool "Hardwall support to allow access to user dynamic network"
diff --git a/arch/tile/gxio/iorpc_mpipe.c b/arch/tile/gxio/iorpc_mpipe.c
index 4f8f3d619c4a..e19325c4c431 100644
--- a/arch/tile/gxio/iorpc_mpipe.c
+++ b/arch/tile/gxio/iorpc_mpipe.c
@@ -21,7 +21,7 @@ struct alloc_buffer_stacks_param {
 	unsigned int flags;
 };
 
-int gxio_mpipe_alloc_buffer_stacks(gxio_mpipe_context_t * context,
+int gxio_mpipe_alloc_buffer_stacks(gxio_mpipe_context_t *context,
 				   unsigned int count, unsigned int first,
 				   unsigned int flags)
 {
@@ -45,7 +45,7 @@ struct init_buffer_stack_aux_param {
 	unsigned int buffer_size_enum;
 };
 
-int gxio_mpipe_init_buffer_stack_aux(gxio_mpipe_context_t * context,
+int gxio_mpipe_init_buffer_stack_aux(gxio_mpipe_context_t *context,
 				     void *mem_va, size_t mem_size,
 				     unsigned int mem_flags, unsigned int stack,
 				     unsigned int buffer_size_enum)
@@ -80,7 +80,7 @@ struct alloc_notif_rings_param {
 	unsigned int flags;
 };
 
-int gxio_mpipe_alloc_notif_rings(gxio_mpipe_context_t * context,
+int gxio_mpipe_alloc_notif_rings(gxio_mpipe_context_t *context,
 				 unsigned int count, unsigned int first,
 				 unsigned int flags)
 {
@@ -102,7 +102,7 @@ struct init_notif_ring_aux_param {
 	unsigned int ring;
 };
 
-int gxio_mpipe_init_notif_ring_aux(gxio_mpipe_context_t * context, void *mem_va,
+int gxio_mpipe_init_notif_ring_aux(gxio_mpipe_context_t *context, void *mem_va,
 				   size_t mem_size, unsigned int mem_flags,
 				   unsigned int ring)
 {
@@ -133,7 +133,7 @@ struct request_notif_ring_interrupt_param {
 	unsigned int ring;
 };
 
-int gxio_mpipe_request_notif_ring_interrupt(gxio_mpipe_context_t * context,
+int gxio_mpipe_request_notif_ring_interrupt(gxio_mpipe_context_t *context,
 					    int inter_x, int inter_y,
 					    int inter_ipi, int inter_event,
 					    unsigned int ring)
@@ -158,7 +158,7 @@ struct enable_notif_ring_interrupt_param {
 	unsigned int ring;
 };
 
-int gxio_mpipe_enable_notif_ring_interrupt(gxio_mpipe_context_t * context,
+int gxio_mpipe_enable_notif_ring_interrupt(gxio_mpipe_context_t *context,
 					   unsigned int ring)
 {
 	struct enable_notif_ring_interrupt_param temp;
@@ -179,7 +179,7 @@ struct alloc_notif_groups_param {
 	unsigned int flags;
 };
 
-int gxio_mpipe_alloc_notif_groups(gxio_mpipe_context_t * context,
+int gxio_mpipe_alloc_notif_groups(gxio_mpipe_context_t *context,
 				  unsigned int count, unsigned int first,
 				  unsigned int flags)
 {
@@ -201,7 +201,7 @@ struct init_notif_group_param {
 	gxio_mpipe_notif_group_bits_t bits;
 };
 
-int gxio_mpipe_init_notif_group(gxio_mpipe_context_t * context,
+int gxio_mpipe_init_notif_group(gxio_mpipe_context_t *context,
 				unsigned int group,
 				gxio_mpipe_notif_group_bits_t bits)
 {
@@ -223,7 +223,7 @@ struct alloc_buckets_param {
 	unsigned int flags;
 };
 
-int gxio_mpipe_alloc_buckets(gxio_mpipe_context_t * context, unsigned int count,
+int gxio_mpipe_alloc_buckets(gxio_mpipe_context_t *context, unsigned int count,
 			     unsigned int first, unsigned int flags)
 {
 	struct alloc_buckets_param temp;
@@ -244,7 +244,7 @@ struct init_bucket_param {
 	MPIPE_LBL_INIT_DAT_BSTS_TBL_t bucket_info;
 };
 
-int gxio_mpipe_init_bucket(gxio_mpipe_context_t * context, unsigned int bucket,
+int gxio_mpipe_init_bucket(gxio_mpipe_context_t *context, unsigned int bucket,
 			   MPIPE_LBL_INIT_DAT_BSTS_TBL_t bucket_info)
 {
 	struct init_bucket_param temp;
@@ -265,7 +265,7 @@ struct alloc_edma_rings_param {
 	unsigned int flags;
 };
 
-int gxio_mpipe_alloc_edma_rings(gxio_mpipe_context_t * context,
+int gxio_mpipe_alloc_edma_rings(gxio_mpipe_context_t *context,
 				unsigned int count, unsigned int first,
 				unsigned int flags)
 {
@@ -288,7 +288,7 @@ struct init_edma_ring_aux_param {
 	unsigned int channel;
 };
 
-int gxio_mpipe_init_edma_ring_aux(gxio_mpipe_context_t * context, void *mem_va,
+int gxio_mpipe_init_edma_ring_aux(gxio_mpipe_context_t *context, void *mem_va,
 				  size_t mem_size, unsigned int mem_flags,
 				  unsigned int ring, unsigned int channel)
 {
@@ -315,7 +315,7 @@ int gxio_mpipe_init_edma_ring_aux(gxio_mpipe_context_t * context, void *mem_va,
 EXPORT_SYMBOL(gxio_mpipe_init_edma_ring_aux);
 
 
-int gxio_mpipe_commit_rules(gxio_mpipe_context_t * context, const void *blob,
+int gxio_mpipe_commit_rules(gxio_mpipe_context_t *context, const void *blob,
 			    size_t blob_size)
 {
 	const void *params = blob;
@@ -332,7 +332,7 @@ struct register_client_memory_param {
 	unsigned int flags;
 };
 
-int gxio_mpipe_register_client_memory(gxio_mpipe_context_t * context,
+int gxio_mpipe_register_client_memory(gxio_mpipe_context_t *context,
 				      unsigned int iotlb, HV_PTE pte,
 				      unsigned int flags)
 {
@@ -355,7 +355,7 @@ struct link_open_aux_param {
 	unsigned int flags;
 };
 
-int gxio_mpipe_link_open_aux(gxio_mpipe_context_t * context,
+int gxio_mpipe_link_open_aux(gxio_mpipe_context_t *context,
 			     _gxio_mpipe_link_name_t name, unsigned int flags)
 {
 	struct link_open_aux_param temp;
@@ -374,7 +374,7 @@ struct link_close_aux_param {
 	int mac;
 };
 
-int gxio_mpipe_link_close_aux(gxio_mpipe_context_t * context, int mac)
+int gxio_mpipe_link_close_aux(gxio_mpipe_context_t *context, int mac)
 {
 	struct link_close_aux_param temp;
 	struct link_close_aux_param *params = &temp;
@@ -393,7 +393,7 @@ struct link_set_attr_aux_param {
 	int64_t val;
 };
 
-int gxio_mpipe_link_set_attr_aux(gxio_mpipe_context_t * context, int mac,
+int gxio_mpipe_link_set_attr_aux(gxio_mpipe_context_t *context, int mac,
 				 uint32_t attr, int64_t val)
 {
 	struct link_set_attr_aux_param temp;
@@ -415,8 +415,8 @@ struct get_timestamp_aux_param {
 	uint64_t cycles;
 };
 
-int gxio_mpipe_get_timestamp_aux(gxio_mpipe_context_t * context, uint64_t * sec,
-				 uint64_t * nsec, uint64_t * cycles)
+int gxio_mpipe_get_timestamp_aux(gxio_mpipe_context_t *context, uint64_t *sec,
+				 uint64_t *nsec, uint64_t *cycles)
 {
 	int __result;
 	struct get_timestamp_aux_param temp;
@@ -440,7 +440,7 @@ struct set_timestamp_aux_param {
 	uint64_t cycles;
 };
 
-int gxio_mpipe_set_timestamp_aux(gxio_mpipe_context_t * context, uint64_t sec,
+int gxio_mpipe_set_timestamp_aux(gxio_mpipe_context_t *context, uint64_t sec,
 				 uint64_t nsec, uint64_t cycles)
 {
 	struct set_timestamp_aux_param temp;
@@ -460,8 +460,7 @@ struct adjust_timestamp_aux_param {
 	int64_t nsec;
 };
 
-int gxio_mpipe_adjust_timestamp_aux(gxio_mpipe_context_t * context,
-				    int64_t nsec)
+int gxio_mpipe_adjust_timestamp_aux(gxio_mpipe_context_t *context, int64_t nsec)
 {
 	struct adjust_timestamp_aux_param temp;
 	struct adjust_timestamp_aux_param *params = &temp;
@@ -475,25 +474,6 @@ int gxio_mpipe_adjust_timestamp_aux(gxio_mpipe_context_t * context,
 
 EXPORT_SYMBOL(gxio_mpipe_adjust_timestamp_aux);
 
-struct adjust_timestamp_freq_param {
-	int32_t ppb;
-};
-
-int gxio_mpipe_adjust_timestamp_freq(gxio_mpipe_context_t * context,
-				     int32_t ppb)
-{
-	struct adjust_timestamp_freq_param temp;
-	struct adjust_timestamp_freq_param *params = &temp;
-
-	params->ppb = ppb;
-
-	return hv_dev_pwrite(context->fd, 0, (HV_VirtAddr) params,
-			     sizeof(*params),
-			     GXIO_MPIPE_OP_ADJUST_TIMESTAMP_FREQ);
-}
-
-EXPORT_SYMBOL(gxio_mpipe_adjust_timestamp_freq);
-
 struct config_edma_ring_blks_param {
 	unsigned int ering;
 	unsigned int max_blks;
@@ -501,7 +481,7 @@ struct config_edma_ring_blks_param {
 	unsigned int db;
 };
 
-int gxio_mpipe_config_edma_ring_blks(gxio_mpipe_context_t * context,
+int gxio_mpipe_config_edma_ring_blks(gxio_mpipe_context_t *context,
 				     unsigned int ering, unsigned int max_blks,
 				     unsigned int min_snf_blks, unsigned int db)
 {
@@ -520,11 +500,29 @@ int gxio_mpipe_config_edma_ring_blks(gxio_mpipe_context_t * context,
 
 EXPORT_SYMBOL(gxio_mpipe_config_edma_ring_blks);
 
+struct adjust_timestamp_freq_param {
+	int32_t ppb;
+};
+
+int gxio_mpipe_adjust_timestamp_freq(gxio_mpipe_context_t *context, int32_t ppb)
+{
+	struct adjust_timestamp_freq_param temp;
+	struct adjust_timestamp_freq_param *params = &temp;
+
+	params->ppb = ppb;
+
+	return hv_dev_pwrite(context->fd, 0, (HV_VirtAddr) params,
+			     sizeof(*params),
+			     GXIO_MPIPE_OP_ADJUST_TIMESTAMP_FREQ);
+}
+
+EXPORT_SYMBOL(gxio_mpipe_adjust_timestamp_freq);
+
 struct arm_pollfd_param {
 	union iorpc_pollfd pollfd;
 };
 
-int gxio_mpipe_arm_pollfd(gxio_mpipe_context_t * context, int pollfd_cookie)
+int gxio_mpipe_arm_pollfd(gxio_mpipe_context_t *context, int pollfd_cookie)
 {
 	struct arm_pollfd_param temp;
 	struct arm_pollfd_param *params = &temp;
@@ -541,7 +539,7 @@ struct close_pollfd_param {
 	union iorpc_pollfd pollfd;
 };
 
-int gxio_mpipe_close_pollfd(gxio_mpipe_context_t * context, int pollfd_cookie)
+int gxio_mpipe_close_pollfd(gxio_mpipe_context_t *context, int pollfd_cookie)
 {
 	struct close_pollfd_param temp;
 	struct close_pollfd_param *params = &temp;
@@ -558,7 +556,7 @@ struct get_mmio_base_param {
 	HV_PTE base;
 };
 
-int gxio_mpipe_get_mmio_base(gxio_mpipe_context_t * context, HV_PTE *base)
+int gxio_mpipe_get_mmio_base(gxio_mpipe_context_t *context, HV_PTE *base)
 {
 	int __result;
 	struct get_mmio_base_param temp;
@@ -579,7 +577,7 @@ struct check_mmio_offset_param {
 	unsigned long size;
 };
 
-int gxio_mpipe_check_mmio_offset(gxio_mpipe_context_t * context,
+int gxio_mpipe_check_mmio_offset(gxio_mpipe_context_t *context,
 				 unsigned long offset, unsigned long size)
 {
 	struct check_mmio_offset_param temp;
diff --git a/arch/tile/gxio/iorpc_mpipe_info.c b/arch/tile/gxio/iorpc_mpipe_info.c
index 64883aabeb9c..77019c6e9b4a 100644
--- a/arch/tile/gxio/iorpc_mpipe_info.c
+++ b/arch/tile/gxio/iorpc_mpipe_info.c
@@ -15,12 +15,11 @@
 /* This file is machine-generated; DO NOT EDIT! */
 #include "gxio/iorpc_mpipe_info.h"
 
-
 struct instance_aux_param {
 	_gxio_mpipe_link_name_t name;
 };
 
-int gxio_mpipe_info_instance_aux(gxio_mpipe_info_context_t * context,
+int gxio_mpipe_info_instance_aux(gxio_mpipe_info_context_t *context,
 				 _gxio_mpipe_link_name_t name)
 {
 	struct instance_aux_param temp;
@@ -39,10 +38,10 @@ struct enumerate_aux_param {
 	_gxio_mpipe_link_mac_t mac;
 };
 
-int gxio_mpipe_info_enumerate_aux(gxio_mpipe_info_context_t * context,
+int gxio_mpipe_info_enumerate_aux(gxio_mpipe_info_context_t *context,
 				  unsigned int idx,
-				  _gxio_mpipe_link_name_t * name,
-				  _gxio_mpipe_link_mac_t * mac)
+				  _gxio_mpipe_link_name_t *name,
+				  _gxio_mpipe_link_mac_t *mac)
 {
 	int __result;
 	struct enumerate_aux_param temp;
@@ -50,7 +49,7 @@ int gxio_mpipe_info_enumerate_aux(gxio_mpipe_info_context_t * context,
 
 	__result =
 	    hv_dev_pread(context->fd, 0, (HV_VirtAddr) params, sizeof(*params),
-			 (((uint64_t) idx << 32) |
+			 (((uint64_t)idx << 32) |
 			  GXIO_MPIPE_INFO_OP_ENUMERATE_AUX));
 	*name = params->name;
 	*mac = params->mac;
@@ -64,7 +63,7 @@ struct get_mmio_base_param {
 	HV_PTE base;
 };
 
-int gxio_mpipe_info_get_mmio_base(gxio_mpipe_info_context_t * context,
+int gxio_mpipe_info_get_mmio_base(gxio_mpipe_info_context_t *context,
 				  HV_PTE *base)
 {
 	int __result;
@@ -86,7 +85,7 @@ struct check_mmio_offset_param {
 	unsigned long size;
 };
 
-int gxio_mpipe_info_check_mmio_offset(gxio_mpipe_info_context_t * context,
+int gxio_mpipe_info_check_mmio_offset(gxio_mpipe_info_context_t *context,
 				      unsigned long offset, unsigned long size)
 {
 	struct check_mmio_offset_param temp;
diff --git a/arch/tile/gxio/iorpc_trio.c b/arch/tile/gxio/iorpc_trio.c
index da6e18e049c3..1d3cedb9aeb4 100644
--- a/arch/tile/gxio/iorpc_trio.c
+++ b/arch/tile/gxio/iorpc_trio.c
@@ -21,7 +21,7 @@ struct alloc_asids_param {
 	unsigned int flags;
 };
 
-int gxio_trio_alloc_asids(gxio_trio_context_t * context, unsigned int count,
+int gxio_trio_alloc_asids(gxio_trio_context_t *context, unsigned int count,
 			  unsigned int first, unsigned int flags)
 {
 	struct alloc_asids_param temp;
@@ -44,7 +44,7 @@ struct alloc_memory_maps_param {
 	unsigned int flags;
 };
 
-int gxio_trio_alloc_memory_maps(gxio_trio_context_t * context,
+int gxio_trio_alloc_memory_maps(gxio_trio_context_t *context,
 				unsigned int count, unsigned int first,
 				unsigned int flags)
 {
@@ -67,7 +67,7 @@ struct alloc_scatter_queues_param {
 	unsigned int flags;
 };
 
-int gxio_trio_alloc_scatter_queues(gxio_trio_context_t * context,
+int gxio_trio_alloc_scatter_queues(gxio_trio_context_t *context,
 				   unsigned int count, unsigned int first,
 				   unsigned int flags)
 {
@@ -91,7 +91,7 @@ struct alloc_pio_regions_param {
 	unsigned int flags;
 };
 
-int gxio_trio_alloc_pio_regions(gxio_trio_context_t * context,
+int gxio_trio_alloc_pio_regions(gxio_trio_context_t *context,
 				unsigned int count, unsigned int first,
 				unsigned int flags)
 {
@@ -115,7 +115,7 @@ struct init_pio_region_aux_param {
 	unsigned int flags;
 };
 
-int gxio_trio_init_pio_region_aux(gxio_trio_context_t * context,
+int gxio_trio_init_pio_region_aux(gxio_trio_context_t *context,
 				  unsigned int pio_region, unsigned int mac,
 				  uint32_t bus_address_hi, unsigned int flags)
 {
@@ -145,7 +145,7 @@ struct init_memory_map_mmu_aux_param {
 	unsigned int order_mode;
 };
 
-int gxio_trio_init_memory_map_mmu_aux(gxio_trio_context_t * context,
+int gxio_trio_init_memory_map_mmu_aux(gxio_trio_context_t *context,
 				      unsigned int map, unsigned long va,
 				      uint64_t size, unsigned int asid,
 				      unsigned int mac, uint64_t bus_address,
@@ -175,7 +175,7 @@ struct get_port_property_param {
 	struct pcie_trio_ports_property trio_ports;
 };
 
-int gxio_trio_get_port_property(gxio_trio_context_t * context,
+int gxio_trio_get_port_property(gxio_trio_context_t *context,
 				struct pcie_trio_ports_property *trio_ports)
 {
 	int __result;
@@ -198,7 +198,7 @@ struct config_legacy_intr_param {
 	unsigned int intx;
 };
 
-int gxio_trio_config_legacy_intr(gxio_trio_context_t * context, int inter_x,
+int gxio_trio_config_legacy_intr(gxio_trio_context_t *context, int inter_x,
 				 int inter_y, int inter_ipi, int inter_event,
 				 unsigned int mac, unsigned int intx)
 {
@@ -227,7 +227,7 @@ struct config_msi_intr_param {
 	unsigned int asid;
 };
 
-int gxio_trio_config_msi_intr(gxio_trio_context_t * context, int inter_x,
+int gxio_trio_config_msi_intr(gxio_trio_context_t *context, int inter_x,
 			      int inter_y, int inter_ipi, int inter_event,
 			      unsigned int mac, unsigned int mem_map,
 			      uint64_t mem_map_base, uint64_t mem_map_limit,
@@ -259,7 +259,7 @@ struct set_mps_mrs_param {
 	unsigned int mac;
 };
 
-int gxio_trio_set_mps_mrs(gxio_trio_context_t * context, uint16_t mps,
+int gxio_trio_set_mps_mrs(gxio_trio_context_t *context, uint16_t mps,
 			  uint16_t mrs, unsigned int mac)
 {
 	struct set_mps_mrs_param temp;
@@ -279,7 +279,7 @@ struct force_rc_link_up_param {
 	unsigned int mac;
 };
 
-int gxio_trio_force_rc_link_up(gxio_trio_context_t * context, unsigned int mac)
+int gxio_trio_force_rc_link_up(gxio_trio_context_t *context, unsigned int mac)
 {
 	struct force_rc_link_up_param temp;
 	struct force_rc_link_up_param *params = &temp;
@@ -296,7 +296,7 @@ struct force_ep_link_up_param {
 	unsigned int mac;
 };
 
-int gxio_trio_force_ep_link_up(gxio_trio_context_t * context, unsigned int mac)
+int gxio_trio_force_ep_link_up(gxio_trio_context_t *context, unsigned int mac)
 {
 	struct force_ep_link_up_param temp;
 	struct force_ep_link_up_param *params = &temp;
@@ -313,7 +313,7 @@ struct get_mmio_base_param {
 	HV_PTE base;
 };
 
-int gxio_trio_get_mmio_base(gxio_trio_context_t * context, HV_PTE *base)
+int gxio_trio_get_mmio_base(gxio_trio_context_t *context, HV_PTE *base)
 {
 	int __result;
 	struct get_mmio_base_param temp;
@@ -334,7 +334,7 @@ struct check_mmio_offset_param {
 	unsigned long size;
 };
 
-int gxio_trio_check_mmio_offset(gxio_trio_context_t * context,
+int gxio_trio_check_mmio_offset(gxio_trio_context_t *context,
 				unsigned long offset, unsigned long size)
 {
 	struct check_mmio_offset_param temp;
diff --git a/arch/tile/gxio/iorpc_usb_host.c b/arch/tile/gxio/iorpc_usb_host.c
index cf3c3cc12204..9c820073bfc0 100644
--- a/arch/tile/gxio/iorpc_usb_host.c
+++ b/arch/tile/gxio/iorpc_usb_host.c
@@ -19,7 +19,7 @@ struct cfg_interrupt_param {
 	union iorpc_interrupt interrupt;
 };
 
-int gxio_usb_host_cfg_interrupt(gxio_usb_host_context_t * context, int inter_x,
+int gxio_usb_host_cfg_interrupt(gxio_usb_host_context_t *context, int inter_x,
 				int inter_y, int inter_ipi, int inter_event)
 {
 	struct cfg_interrupt_param temp;
@@ -41,7 +41,7 @@ struct register_client_memory_param {
 	unsigned int flags;
 };
 
-int gxio_usb_host_register_client_memory(gxio_usb_host_context_t * context,
+int gxio_usb_host_register_client_memory(gxio_usb_host_context_t *context,
 					 HV_PTE pte, unsigned int flags)
 {
 	struct register_client_memory_param temp;
@@ -61,7 +61,7 @@ struct get_mmio_base_param {
 	HV_PTE base;
 };
 
-int gxio_usb_host_get_mmio_base(gxio_usb_host_context_t * context, HV_PTE *base)
+int gxio_usb_host_get_mmio_base(gxio_usb_host_context_t *context, HV_PTE *base)
 {
 	int __result;
 	struct get_mmio_base_param temp;
@@ -82,7 +82,7 @@ struct check_mmio_offset_param {
 	unsigned long size;
 };
 
-int gxio_usb_host_check_mmio_offset(gxio_usb_host_context_t * context,
+int gxio_usb_host_check_mmio_offset(gxio_usb_host_context_t *context,
 				    unsigned long offset, unsigned long size)
 {
 	struct check_mmio_offset_param temp;
diff --git a/arch/tile/gxio/usb_host.c b/arch/tile/gxio/usb_host.c
index 66b002f54ecc..785afad7922e 100644
--- a/arch/tile/gxio/usb_host.c
+++ b/arch/tile/gxio/usb_host.c
@@ -26,7 +26,7 @@
 #include <gxio/kiorpc.h>
 #include <gxio/usb_host.h>
 
-int gxio_usb_host_init(gxio_usb_host_context_t * context, int usb_index,
+int gxio_usb_host_init(gxio_usb_host_context_t *context, int usb_index,
 		       int is_ehci)
 {
 	char file[32];
@@ -63,7 +63,7 @@ int gxio_usb_host_init(gxio_usb_host_context_t * context, int usb_index,
 
 EXPORT_SYMBOL_GPL(gxio_usb_host_init);
 
-int gxio_usb_host_destroy(gxio_usb_host_context_t * context)
+int gxio_usb_host_destroy(gxio_usb_host_context_t *context)
 {
 	iounmap((void __force __iomem *)(context->mmio_base));
 	hv_dev_close(context->fd);
@@ -76,14 +76,14 @@ int gxio_usb_host_destroy(gxio_usb_host_context_t * context)
 
 EXPORT_SYMBOL_GPL(gxio_usb_host_destroy);
 
-void *gxio_usb_host_get_reg_start(gxio_usb_host_context_t * context)
+void *gxio_usb_host_get_reg_start(gxio_usb_host_context_t *context)
 {
 	return context->mmio_base;
 }
 
 EXPORT_SYMBOL_GPL(gxio_usb_host_get_reg_start);
 
-size_t gxio_usb_host_get_reg_len(gxio_usb_host_context_t * context)
+size_t gxio_usb_host_get_reg_len(gxio_usb_host_context_t *context)
 {
 	return HV_USB_HOST_MMIO_SIZE;
 }
diff --git a/arch/tile/include/arch/mpipe.h b/arch/tile/include/arch/mpipe.h
index 8a33912fd6cc..904538e754d8 100644
--- a/arch/tile/include/arch/mpipe.h
+++ b/arch/tile/include/arch/mpipe.h
@@ -176,7 +176,18 @@ typedef union
      */
     uint_reg_t stack_idx    : 5;
     /* Reserved. */
-    uint_reg_t __reserved_2 : 5;
+    uint_reg_t __reserved_2 : 3;
+    /*
+     * Instance ID.  For devices that support automatic buffer return between
+     * mPIPE instances, this field indicates the buffer owner.  If the INST
+     * field does not match the mPIPE's instance number when a packet is
+     * egressed, buffers with HWB set will be returned to the other mPIPE
+     * instance.  Note that not all devices support multi-mPIPE buffer
+     * return.  The MPIPE_EDMA_INFO.REMOTE_BUFF_RTN_SUPPORT bit indicates
+     * whether the INST field in the buffer descriptor is populated by iDMA
+     * hardware. This field is ignored on writes.
+     */
+    uint_reg_t inst         : 2;
     /*
      * Reads as one to indicate that this is a hardware managed buffer.
      * Ignored on writes since all buffers on a given stack are the same size.
@@ -205,7 +216,8 @@ typedef union
     uint_reg_t c            : 2;
     uint_reg_t size         : 3;
     uint_reg_t hwb          : 1;
-    uint_reg_t __reserved_2 : 5;
+    uint_reg_t inst         : 2;
+    uint_reg_t __reserved_2 : 3;
     uint_reg_t stack_idx    : 5;
     uint_reg_t __reserved_1 : 6;
     int_reg_t va           : 35;
@@ -231,9 +243,9 @@ typedef union
     /* Reserved. */
     uint_reg_t __reserved_0 : 3;
     /* eDMA ring being accessed */
-    uint_reg_t ring         : 5;
+    uint_reg_t ring         : 6;
     /* Reserved. */
-    uint_reg_t __reserved_1 : 18;
+    uint_reg_t __reserved_1 : 17;
     /*
      * This field of the address selects the region (address space) to be
      * accessed.  For the egress DMA post region, this field must be 5.
@@ -250,8 +262,8 @@ typedef union
     uint_reg_t svc_dom      : 5;
     uint_reg_t __reserved_2 : 6;
     uint_reg_t region       : 3;
-    uint_reg_t __reserved_1 : 18;
-    uint_reg_t ring         : 5;
+    uint_reg_t __reserved_1 : 17;
+    uint_reg_t ring         : 6;
     uint_reg_t __reserved_0 : 3;
 #endif
   };
diff --git a/arch/tile/include/arch/mpipe_constants.h b/arch/tile/include/arch/mpipe_constants.h
index 410a0400e055..84022ac5fe82 100644
--- a/arch/tile/include/arch/mpipe_constants.h
+++ b/arch/tile/include/arch/mpipe_constants.h
@@ -16,13 +16,13 @@
 #ifndef __ARCH_MPIPE_CONSTANTS_H__
 #define __ARCH_MPIPE_CONSTANTS_H__
 
-#define MPIPE_NUM_CLASSIFIERS 10
+#define MPIPE_NUM_CLASSIFIERS 16
 #define MPIPE_CLS_MHZ 1200
 
-#define MPIPE_NUM_EDMA_RINGS 32
+#define MPIPE_NUM_EDMA_RINGS 64
 
 #define MPIPE_NUM_SGMII_MACS 16
-#define MPIPE_NUM_XAUI_MACS 4
+#define MPIPE_NUM_XAUI_MACS 16
 #define MPIPE_NUM_LOOPBACK_CHANNELS 4
 #define MPIPE_NUM_NON_LB_CHANNELS 28
 
diff --git a/arch/tile/include/arch/mpipe_shm.h b/arch/tile/include/arch/mpipe_shm.h
index f2e9e122818d..13b3c4300e50 100644
--- a/arch/tile/include/arch/mpipe_shm.h
+++ b/arch/tile/include/arch/mpipe_shm.h
@@ -44,8 +44,14 @@ typedef union
      * descriptors toggles each time the ring tail pointer wraps.
      */
     uint_reg_t gen        : 1;
+    /**
+     * For devices with EDMA reorder support, this field allows the
+     * descriptor to select the egress FIFO.  The associated DMA ring must
+     * have ALLOW_EFIFO_SEL enabled.
+     */
+    uint_reg_t efifo_sel  : 6;
     /** Reserved.  Must be zero. */
-    uint_reg_t r0         : 7;
+    uint_reg_t r0         : 1;
     /** Checksum generation enabled for this transfer. */
     uint_reg_t csum       : 1;
     /**
@@ -110,7 +116,8 @@ typedef union
     uint_reg_t notif      : 1;
     uint_reg_t ns         : 1;
     uint_reg_t csum       : 1;
-    uint_reg_t r0         : 7;
+    uint_reg_t r0         : 1;
+    uint_reg_t efifo_sel  : 6;
     uint_reg_t gen        : 1;
 #endif
 
@@ -126,14 +133,16 @@ typedef union
     /** Reserved. */
     uint_reg_t __reserved_1 : 3;
     /**
-     * Instance ID.  For devices that support more than one mPIPE instance,
-     * this field indicates the buffer owner.  If the INST field does not
-     * match the mPIPE's instance number when a packet is egressed, buffers
-     * with HWB set will be returned to the other mPIPE instance.
+     * Instance ID.  For devices that support automatic buffer return between
+     * mPIPE instances, this field indicates the buffer owner.  If the INST
+     * field does not match the mPIPE's instance number when a packet is
+     * egressed, buffers with HWB set will be returned to the other mPIPE
+     * instance.  Note that not all devices support multi-mPIPE buffer
+     * return.  The MPIPE_EDMA_INFO.REMOTE_BUFF_RTN_SUPPORT bit indicates
+     * whether the INST field in the buffer descriptor is populated by iDMA
+     * hardware.
      */
-    uint_reg_t inst         : 1;
-    /** Reserved. */
-    uint_reg_t __reserved_2 : 1;
+    uint_reg_t inst         : 2;
     /**
      * Always set to one by hardware in iDMA packet descriptors.  For eDMA,
      * indicates whether the buffer will be released to the buffer stack
@@ -166,8 +175,7 @@ typedef union
     uint_reg_t c            : 2;
     uint_reg_t size         : 3;
     uint_reg_t hwb          : 1;
-    uint_reg_t __reserved_2 : 1;
-    uint_reg_t inst         : 1;
+    uint_reg_t inst         : 2;
     uint_reg_t __reserved_1 : 3;
     uint_reg_t stack_idx    : 5;
     uint_reg_t __reserved_0 : 6;
@@ -408,7 +416,10 @@ typedef union
     /**
      * Sequence number applied when packet is distributed.   Classifier
      * selects which sequence number is to be applied by writing the 13-bit
-     * SQN-selector into this field.
+     * SQN-selector into this field.  For devices that support EXT_SQN (as
+     * indicated in IDMA_INFO.EXT_SQN_SUPPORT), the GP_SQN can be extended to
+     * 32-bits via the IDMA_CTL.EXT_SQN register.  In this case the
+     * PACKET_SQN will be reduced to 32 bits.
      */
     uint_reg_t gp_sqn     : 16;
     /**
@@ -451,14 +462,16 @@ typedef union
     /** Reserved. */
     uint_reg_t __reserved_5 : 3;
     /**
-     * Instance ID.  For devices that support more than one mPIPE instance,
-     * this field indicates the buffer owner.  If the INST field does not
-     * match the mPIPE's instance number when a packet is egressed, buffers
-     * with HWB set will be returned to the other mPIPE instance.
+     * Instance ID.  For devices that support automatic buffer return between
+     * mPIPE instances, this field indicates the buffer owner.  If the INST
+     * field does not match the mPIPE's instance number when a packet is
+     * egressed, buffers with HWB set will be returned to the other mPIPE
+     * instance.  Note that not all devices support multi-mPIPE buffer
+     * return.  The MPIPE_EDMA_INFO.REMOTE_BUFF_RTN_SUPPORT bit indicates
+     * whether the INST field in the buffer descriptor is populated by iDMA
+     * hardware.
      */
-    uint_reg_t inst         : 1;
-    /** Reserved. */
-    uint_reg_t __reserved_6 : 1;
+    uint_reg_t inst         : 2;
     /**
      * Always set to one by hardware in iDMA packet descriptors.  For eDMA,
      * indicates whether the buffer will be released to the buffer stack
@@ -491,8 +504,7 @@ typedef union
     uint_reg_t c            : 2;
     uint_reg_t size         : 3;
     uint_reg_t hwb          : 1;
-    uint_reg_t __reserved_6 : 1;
-    uint_reg_t inst         : 1;
+    uint_reg_t inst         : 2;
     uint_reg_t __reserved_5 : 3;
     uint_reg_t stack_idx    : 5;
     uint_reg_t __reserved_4 : 6;
diff --git a/arch/tile/include/arch/trio_constants.h b/arch/tile/include/arch/trio_constants.h
index 628b045436b8..85647e91a458 100644
--- a/arch/tile/include/arch/trio_constants.h
+++ b/arch/tile/include/arch/trio_constants.h
@@ -16,21 +16,21 @@
 #ifndef __ARCH_TRIO_CONSTANTS_H__
 #define __ARCH_TRIO_CONSTANTS_H__
 
-#define TRIO_NUM_ASIDS 16
+#define TRIO_NUM_ASIDS 32
 #define TRIO_NUM_TLBS_PER_ASID 16
 
 #define TRIO_NUM_TPIO_REGIONS 8
 #define TRIO_LOG2_NUM_TPIO_REGIONS 3
 
-#define TRIO_NUM_MAP_MEM_REGIONS 16
-#define TRIO_LOG2_NUM_MAP_MEM_REGIONS 4
+#define TRIO_NUM_MAP_MEM_REGIONS 32
+#define TRIO_LOG2_NUM_MAP_MEM_REGIONS 5
 #define TRIO_NUM_MAP_SQ_REGIONS 8
 #define TRIO_LOG2_NUM_MAP_SQ_REGIONS 3
 
 #define TRIO_LOG2_NUM_SQ_FIFO_ENTRIES 6
 
-#define TRIO_NUM_PUSH_DMA_RINGS 32
+#define TRIO_NUM_PUSH_DMA_RINGS 64
 
-#define TRIO_NUM_PULL_DMA_RINGS 32
+#define TRIO_NUM_PULL_DMA_RINGS 64
 
 #endif /* __ARCH_TRIO_CONSTANTS_H__ */
diff --git a/arch/tile/include/asm/page.h b/arch/tile/include/asm/page.h
index 6346888f7bdc..672768008618 100644
--- a/arch/tile/include/asm/page.h
+++ b/arch/tile/include/asm/page.h
@@ -182,10 +182,9 @@ static inline __attribute_const__ int get_order(unsigned long size)
 
 #define PAGE_OFFSET		(-(_AC(1, UL) << (MAX_VA_WIDTH - 1)))
 #define KERNEL_HIGH_VADDR	_AC(0xfffffff800000000, UL)  /* high 32GB */
-#define FIXADDR_BASE		(KERNEL_HIGH_VADDR - 0x400000000) /* 4 GB */
-#define FIXADDR_TOP		(KERNEL_HIGH_VADDR - 0x300000000) /* 4 GB */
+#define FIXADDR_BASE		(KERNEL_HIGH_VADDR - 0x300000000) /* 4 GB */
+#define FIXADDR_TOP		(KERNEL_HIGH_VADDR - 0x200000000) /* 4 GB */
 #define _VMALLOC_START		FIXADDR_TOP
-#define HUGE_VMAP_BASE		(KERNEL_HIGH_VADDR - 0x200000000) /* 4 GB */
 #define MEM_SV_START		(KERNEL_HIGH_VADDR - 0x100000000) /* 256 MB */
 #define MEM_MODULE_START	(MEM_SV_START + (256*1024*1024)) /* 256 MB */
 #define MEM_MODULE_END		(MEM_MODULE_START + (256*1024*1024))
diff --git a/arch/tile/include/asm/pgtable_32.h b/arch/tile/include/asm/pgtable_32.h
index 63142ab3b3dd..d26a42279036 100644
--- a/arch/tile/include/asm/pgtable_32.h
+++ b/arch/tile/include/asm/pgtable_32.h
@@ -55,17 +55,9 @@
 #define PKMAP_BASE   ((FIXADDR_BOOT_START - PAGE_SIZE*LAST_PKMAP) & PGDIR_MASK)
 
 #ifdef CONFIG_HIGHMEM
-# define __VMAPPING_END	(PKMAP_BASE & ~(HPAGE_SIZE-1))
+# define _VMALLOC_END	(PKMAP_BASE & ~(HPAGE_SIZE-1))
 #else
-# define __VMAPPING_END	(FIXADDR_START & ~(HPAGE_SIZE-1))
-#endif
-
-#ifdef CONFIG_HUGEVMAP
-#define HUGE_VMAP_END	__VMAPPING_END
-#define HUGE_VMAP_BASE	(HUGE_VMAP_END - CONFIG_NR_HUGE_VMAPS * HPAGE_SIZE)
-#define _VMALLOC_END	HUGE_VMAP_BASE
-#else
-#define _VMALLOC_END	__VMAPPING_END
+# define _VMALLOC_END	(FIXADDR_START & ~(HPAGE_SIZE-1))
 #endif
 
 /*
diff --git a/arch/tile/include/asm/pgtable_64.h b/arch/tile/include/asm/pgtable_64.h
index 3421177f7370..2c8a9cd102d3 100644
--- a/arch/tile/include/asm/pgtable_64.h
+++ b/arch/tile/include/asm/pgtable_64.h
@@ -52,12 +52,10 @@
  * memory allocation code).  The vmalloc code puts in an internal
  * guard page between each allocation.
  */
-#define _VMALLOC_END	HUGE_VMAP_BASE
+#define _VMALLOC_END	MEM_SV_START
 #define VMALLOC_END	_VMALLOC_END
 #define VMALLOC_START	_VMALLOC_START
 
-#define HUGE_VMAP_END	(HUGE_VMAP_BASE + PGDIR_SIZE)
-
 #ifndef __ASSEMBLY__
 
 /* We have no pud since we are a three-level page table. */
diff --git a/arch/tile/include/gxio/iorpc_mpipe.h b/arch/tile/include/gxio/iorpc_mpipe.h
index fdd07f88cfd7..4cda03de734f 100644
--- a/arch/tile/include/gxio/iorpc_mpipe.h
+++ b/arch/tile/include/gxio/iorpc_mpipe.h
@@ -56,89 +56,89 @@
 #define GXIO_MPIPE_OP_GET_MMIO_BASE    IORPC_OPCODE(IORPC_FORMAT_NONE_NOUSER, 0x8000)
 #define GXIO_MPIPE_OP_CHECK_MMIO_OFFSET IORPC_OPCODE(IORPC_FORMAT_NONE_NOUSER, 0x8001)
 
-int gxio_mpipe_alloc_buffer_stacks(gxio_mpipe_context_t * context,
+int gxio_mpipe_alloc_buffer_stacks(gxio_mpipe_context_t *context,
 				   unsigned int count, unsigned int first,
 				   unsigned int flags);
 
-int gxio_mpipe_init_buffer_stack_aux(gxio_mpipe_context_t * context,
+int gxio_mpipe_init_buffer_stack_aux(gxio_mpipe_context_t *context,
 				     void *mem_va, size_t mem_size,
 				     unsigned int mem_flags, unsigned int stack,
 				     unsigned int buffer_size_enum);
 
 
-int gxio_mpipe_alloc_notif_rings(gxio_mpipe_context_t * context,
+int gxio_mpipe_alloc_notif_rings(gxio_mpipe_context_t *context,
 				 unsigned int count, unsigned int first,
 				 unsigned int flags);
 
-int gxio_mpipe_init_notif_ring_aux(gxio_mpipe_context_t * context, void *mem_va,
+int gxio_mpipe_init_notif_ring_aux(gxio_mpipe_context_t *context, void *mem_va,
 				   size_t mem_size, unsigned int mem_flags,
 				   unsigned int ring);
 
-int gxio_mpipe_request_notif_ring_interrupt(gxio_mpipe_context_t * context,
+int gxio_mpipe_request_notif_ring_interrupt(gxio_mpipe_context_t *context,
 					    int inter_x, int inter_y,
 					    int inter_ipi, int inter_event,
 					    unsigned int ring);
 
-int gxio_mpipe_enable_notif_ring_interrupt(gxio_mpipe_context_t * context,
+int gxio_mpipe_enable_notif_ring_interrupt(gxio_mpipe_context_t *context,
 					   unsigned int ring);
 
-int gxio_mpipe_alloc_notif_groups(gxio_mpipe_context_t * context,
+int gxio_mpipe_alloc_notif_groups(gxio_mpipe_context_t *context,
 				  unsigned int count, unsigned int first,
 				  unsigned int flags);
 
-int gxio_mpipe_init_notif_group(gxio_mpipe_context_t * context,
+int gxio_mpipe_init_notif_group(gxio_mpipe_context_t *context,
 				unsigned int group,
 				gxio_mpipe_notif_group_bits_t bits);
 
-int gxio_mpipe_alloc_buckets(gxio_mpipe_context_t * context, unsigned int count,
+int gxio_mpipe_alloc_buckets(gxio_mpipe_context_t *context, unsigned int count,
 			     unsigned int first, unsigned int flags);
 
-int gxio_mpipe_init_bucket(gxio_mpipe_context_t * context, unsigned int bucket,
+int gxio_mpipe_init_bucket(gxio_mpipe_context_t *context, unsigned int bucket,
 			   MPIPE_LBL_INIT_DAT_BSTS_TBL_t bucket_info);
 
-int gxio_mpipe_alloc_edma_rings(gxio_mpipe_context_t * context,
+int gxio_mpipe_alloc_edma_rings(gxio_mpipe_context_t *context,
 				unsigned int count, unsigned int first,
 				unsigned int flags);
 
-int gxio_mpipe_init_edma_ring_aux(gxio_mpipe_context_t * context, void *mem_va,
+int gxio_mpipe_init_edma_ring_aux(gxio_mpipe_context_t *context, void *mem_va,
 				  size_t mem_size, unsigned int mem_flags,
 				  unsigned int ring, unsigned int channel);
 
 
-int gxio_mpipe_commit_rules(gxio_mpipe_context_t * context, const void *blob,
+int gxio_mpipe_commit_rules(gxio_mpipe_context_t *context, const void *blob,
 			    size_t blob_size);
 
-int gxio_mpipe_register_client_memory(gxio_mpipe_context_t * context,
+int gxio_mpipe_register_client_memory(gxio_mpipe_context_t *context,
 				      unsigned int iotlb, HV_PTE pte,
 				      unsigned int flags);
 
-int gxio_mpipe_link_open_aux(gxio_mpipe_context_t * context,
+int gxio_mpipe_link_open_aux(gxio_mpipe_context_t *context,
 			     _gxio_mpipe_link_name_t name, unsigned int flags);
 
-int gxio_mpipe_link_close_aux(gxio_mpipe_context_t * context, int mac);
+int gxio_mpipe_link_close_aux(gxio_mpipe_context_t *context, int mac);
 
-int gxio_mpipe_link_set_attr_aux(gxio_mpipe_context_t * context, int mac,
+int gxio_mpipe_link_set_attr_aux(gxio_mpipe_context_t *context, int mac,
 				 uint32_t attr, int64_t val);
 
-int gxio_mpipe_get_timestamp_aux(gxio_mpipe_context_t * context, uint64_t * sec,
-				 uint64_t * nsec, uint64_t * cycles);
+int gxio_mpipe_get_timestamp_aux(gxio_mpipe_context_t *context, uint64_t *sec,
+				 uint64_t *nsec, uint64_t *cycles);
 
-int gxio_mpipe_set_timestamp_aux(gxio_mpipe_context_t * context, uint64_t sec,
+int gxio_mpipe_set_timestamp_aux(gxio_mpipe_context_t *context, uint64_t sec,
 				 uint64_t nsec, uint64_t cycles);
 
-int gxio_mpipe_adjust_timestamp_aux(gxio_mpipe_context_t * context,
+int gxio_mpipe_adjust_timestamp_aux(gxio_mpipe_context_t *context,
 				    int64_t nsec);
 
-int gxio_mpipe_adjust_timestamp_freq(gxio_mpipe_context_t * context,
+int gxio_mpipe_adjust_timestamp_freq(gxio_mpipe_context_t *context,
 				     int32_t ppb);
 
-int gxio_mpipe_arm_pollfd(gxio_mpipe_context_t * context, int pollfd_cookie);
+int gxio_mpipe_arm_pollfd(gxio_mpipe_context_t *context, int pollfd_cookie);
 
-int gxio_mpipe_close_pollfd(gxio_mpipe_context_t * context, int pollfd_cookie);
+int gxio_mpipe_close_pollfd(gxio_mpipe_context_t *context, int pollfd_cookie);
 
-int gxio_mpipe_get_mmio_base(gxio_mpipe_context_t * context, HV_PTE *base);
+int gxio_mpipe_get_mmio_base(gxio_mpipe_context_t *context, HV_PTE *base);
 
-int gxio_mpipe_check_mmio_offset(gxio_mpipe_context_t * context,
+int gxio_mpipe_check_mmio_offset(gxio_mpipe_context_t *context,
 				 unsigned long offset, unsigned long size);
 
 #endif /* !__GXIO_MPIPE_LINUX_RPC_H__ */
diff --git a/arch/tile/include/gxio/iorpc_mpipe_info.h b/arch/tile/include/gxio/iorpc_mpipe_info.h
index 476c5e5ca22c..f0b04284468b 100644
--- a/arch/tile/include/gxio/iorpc_mpipe_info.h
+++ b/arch/tile/include/gxio/iorpc_mpipe_info.h
@@ -33,18 +33,18 @@
 #define GXIO_MPIPE_INFO_OP_CHECK_MMIO_OFFSET IORPC_OPCODE(IORPC_FORMAT_NONE_NOUSER, 0x8001)
 
 
-int gxio_mpipe_info_instance_aux(gxio_mpipe_info_context_t * context,
+int gxio_mpipe_info_instance_aux(gxio_mpipe_info_context_t *context,
 				 _gxio_mpipe_link_name_t name);
 
-int gxio_mpipe_info_enumerate_aux(gxio_mpipe_info_context_t * context,
+int gxio_mpipe_info_enumerate_aux(gxio_mpipe_info_context_t *context,
 				  unsigned int idx,
-				  _gxio_mpipe_link_name_t * name,
-				  _gxio_mpipe_link_mac_t * mac);
+				  _gxio_mpipe_link_name_t *name,
+				  _gxio_mpipe_link_mac_t *mac);
 
-int gxio_mpipe_info_get_mmio_base(gxio_mpipe_info_context_t * context,
+int gxio_mpipe_info_get_mmio_base(gxio_mpipe_info_context_t *context,
 				  HV_PTE *base);
 
-int gxio_mpipe_info_check_mmio_offset(gxio_mpipe_info_context_t * context,
+int gxio_mpipe_info_check_mmio_offset(gxio_mpipe_info_context_t *context,
 				      unsigned long offset, unsigned long size);
 
 #endif /* !__GXIO_MPIPE_INFO_LINUX_RPC_H__ */
diff --git a/arch/tile/include/gxio/iorpc_trio.h b/arch/tile/include/gxio/iorpc_trio.h
index d95b96fd6c93..376a4f771167 100644
--- a/arch/tile/include/gxio/iorpc_trio.h
+++ b/arch/tile/include/gxio/iorpc_trio.h
@@ -46,59 +46,59 @@
 #define GXIO_TRIO_OP_GET_MMIO_BASE     IORPC_OPCODE(IORPC_FORMAT_NONE_NOUSER, 0x8000)
 #define GXIO_TRIO_OP_CHECK_MMIO_OFFSET IORPC_OPCODE(IORPC_FORMAT_NONE_NOUSER, 0x8001)
 
-int gxio_trio_alloc_asids(gxio_trio_context_t * context, unsigned int count,
+int gxio_trio_alloc_asids(gxio_trio_context_t *context, unsigned int count,
 			  unsigned int first, unsigned int flags);
 
 
-int gxio_trio_alloc_memory_maps(gxio_trio_context_t * context,
+int gxio_trio_alloc_memory_maps(gxio_trio_context_t *context,
 				unsigned int count, unsigned int first,
 				unsigned int flags);
 
 
-int gxio_trio_alloc_scatter_queues(gxio_trio_context_t * context,
+int gxio_trio_alloc_scatter_queues(gxio_trio_context_t *context,
 				   unsigned int count, unsigned int first,
 				   unsigned int flags);
 
-int gxio_trio_alloc_pio_regions(gxio_trio_context_t * context,
+int gxio_trio_alloc_pio_regions(gxio_trio_context_t *context,
 				unsigned int count, unsigned int first,
 				unsigned int flags);
 
-int gxio_trio_init_pio_region_aux(gxio_trio_context_t * context,
+int gxio_trio_init_pio_region_aux(gxio_trio_context_t *context,
 				  unsigned int pio_region, unsigned int mac,
 				  uint32_t bus_address_hi, unsigned int flags);
 
 
-int gxio_trio_init_memory_map_mmu_aux(gxio_trio_context_t * context,
+int gxio_trio_init_memory_map_mmu_aux(gxio_trio_context_t *context,
 				      unsigned int map, unsigned long va,
 				      uint64_t size, unsigned int asid,
 				      unsigned int mac, uint64_t bus_address,
 				      unsigned int node,
 				      unsigned int order_mode);
 
-int gxio_trio_get_port_property(gxio_trio_context_t * context,
+int gxio_trio_get_port_property(gxio_trio_context_t *context,
 				struct pcie_trio_ports_property *trio_ports);
 
-int gxio_trio_config_legacy_intr(gxio_trio_context_t * context, int inter_x,
+int gxio_trio_config_legacy_intr(gxio_trio_context_t *context, int inter_x,
 				 int inter_y, int inter_ipi, int inter_event,
 				 unsigned int mac, unsigned int intx);
 
-int gxio_trio_config_msi_intr(gxio_trio_context_t * context, int inter_x,
+int gxio_trio_config_msi_intr(gxio_trio_context_t *context, int inter_x,
 			      int inter_y, int inter_ipi, int inter_event,
 			      unsigned int mac, unsigned int mem_map,
 			      uint64_t mem_map_base, uint64_t mem_map_limit,
 			      unsigned int asid);
 
 
-int gxio_trio_set_mps_mrs(gxio_trio_context_t * context, uint16_t mps,
+int gxio_trio_set_mps_mrs(gxio_trio_context_t *context, uint16_t mps,
 			  uint16_t mrs, unsigned int mac);
 
-int gxio_trio_force_rc_link_up(gxio_trio_context_t * context, unsigned int mac);
+int gxio_trio_force_rc_link_up(gxio_trio_context_t *context, unsigned int mac);
 
-int gxio_trio_force_ep_link_up(gxio_trio_context_t * context, unsigned int mac);
+int gxio_trio_force_ep_link_up(gxio_trio_context_t *context, unsigned int mac);
 
-int gxio_trio_get_mmio_base(gxio_trio_context_t * context, HV_PTE *base);
+int gxio_trio_get_mmio_base(gxio_trio_context_t *context, HV_PTE *base);
 
-int gxio_trio_check_mmio_offset(gxio_trio_context_t * context,
+int gxio_trio_check_mmio_offset(gxio_trio_context_t *context,
 				unsigned long offset, unsigned long size);
 
 #endif /* !__GXIO_TRIO_LINUX_RPC_H__ */
diff --git a/arch/tile/include/gxio/iorpc_usb_host.h b/arch/tile/include/gxio/iorpc_usb_host.h
index 8622e7d126ad..79962a97de8e 100644
--- a/arch/tile/include/gxio/iorpc_usb_host.h
+++ b/arch/tile/include/gxio/iorpc_usb_host.h
@@ -31,16 +31,16 @@
 #define GXIO_USB_HOST_OP_GET_MMIO_BASE IORPC_OPCODE(IORPC_FORMAT_NONE_NOUSER, 0x8000)
 #define GXIO_USB_HOST_OP_CHECK_MMIO_OFFSET IORPC_OPCODE(IORPC_FORMAT_NONE_NOUSER, 0x8001)
 
-int gxio_usb_host_cfg_interrupt(gxio_usb_host_context_t * context, int inter_x,
+int gxio_usb_host_cfg_interrupt(gxio_usb_host_context_t *context, int inter_x,
 				int inter_y, int inter_ipi, int inter_event);
 
-int gxio_usb_host_register_client_memory(gxio_usb_host_context_t * context,
+int gxio_usb_host_register_client_memory(gxio_usb_host_context_t *context,
 					 HV_PTE pte, unsigned int flags);
 
-int gxio_usb_host_get_mmio_base(gxio_usb_host_context_t * context,
+int gxio_usb_host_get_mmio_base(gxio_usb_host_context_t *context,
 				HV_PTE *base);
 
-int gxio_usb_host_check_mmio_offset(gxio_usb_host_context_t * context,
+int gxio_usb_host_check_mmio_offset(gxio_usb_host_context_t *context,
 				    unsigned long offset, unsigned long size);
 
 #endif /* !__GXIO_USB_HOST_LINUX_RPC_H__ */
diff --git a/arch/tile/include/gxio/usb_host.h b/arch/tile/include/gxio/usb_host.h
index 5eedec0e988e..93c9636d2dd7 100644
--- a/arch/tile/include/gxio/usb_host.h
+++ b/arch/tile/include/gxio/usb_host.h
@@ -53,7 +53,7 @@ typedef struct {
  * @return Zero if the context was successfully initialized, else a
  *  GXIO_ERR_xxx error code.
  */
-extern int gxio_usb_host_init(gxio_usb_host_context_t * context, int usb_index,
+extern int gxio_usb_host_init(gxio_usb_host_context_t *context, int usb_index,
 			      int is_ehci);
 
 /* Destroy a USB context.
@@ -68,20 +68,20 @@ extern int gxio_usb_host_init(gxio_usb_host_context_t * context, int usb_index,
  * @return Zero if the context was successfully destroyed, else a
  *  GXIO_ERR_xxx error code.
  */
-extern int gxio_usb_host_destroy(gxio_usb_host_context_t * context);
+extern int gxio_usb_host_destroy(gxio_usb_host_context_t *context);
 
 /* Retrieve the address of the shim's MMIO registers.
  *
  * @param context Pointer to a properly initialized gxio_usb_host_context_t.
  * @return The address of the shim's MMIO registers.
  */
-extern void *gxio_usb_host_get_reg_start(gxio_usb_host_context_t * context);
+extern void *gxio_usb_host_get_reg_start(gxio_usb_host_context_t *context);
 
 /* Retrieve the length of the shim's MMIO registers.
  *
  * @param context Pointer to a properly initialized gxio_usb_host_context_t.
  * @return The length of the shim's MMIO registers.
  */
-extern size_t gxio_usb_host_get_reg_len(gxio_usb_host_context_t * context);
+extern size_t gxio_usb_host_get_reg_len(gxio_usb_host_context_t *context);
 
 #endif /* _GXIO_USB_H_ */
diff --git a/arch/tile/kernel/compat.c b/arch/tile/kernel/compat.c
index ed378416b86a..49120843ff96 100644
--- a/arch/tile/kernel/compat.c
+++ b/arch/tile/kernel/compat.c
@@ -84,7 +84,7 @@ COMPAT_SYSCALL_DEFINE5(llseek, unsigned int, fd, unsigned int, offset_high,
 {
 	return sys_llseek(fd, offset_high, offset_low, result, origin);
 }
- 
+
 /* Provide the compat syscall number to call mapping. */
 #undef __SYSCALL
 #define __SYSCALL(nr, call) [nr] = (call),
diff --git a/arch/tile/kernel/futex_64.S b/arch/tile/kernel/futex_64.S
deleted file mode 100644
index f465d1eda20f..000000000000
--- a/arch/tile/kernel/futex_64.S
+++ /dev/null
@@ -1,55 +0,0 @@
-/*
- * Copyright 2011 Tilera Corporation. All Rights Reserved.
- *
- *   This program is free software; you can redistribute it and/or
- *   modify it under the terms of the GNU General Public License
- *   as published by the Free Software Foundation, version 2.
- *
- *   This program is distributed in the hope that it will be useful, but
- *   WITHOUT ANY WARRANTY; without even the implied warranty of
- *   MERCHANTABILITY OR FITNESS FOR A PARTICULAR PURPOSE, GOOD TITLE or
- *   NON INFRINGEMENT.  See the GNU General Public License for
- *   more details.
- *
- * Atomically access user memory, but use MMU to avoid propagating
- * kernel exceptions.
- */
-
-#include <linux/linkage.h>
-#include <asm/errno.h>
-#include <asm/futex.h>
-#include <asm/page.h>
-#include <asm/processor.h>
-
-/*
- * Provide a set of atomic memory operations supporting <asm/futex.h>.
- *
- * r0: user address to manipulate
- * r1: new value to write, or for cmpxchg, old value to compare against
- * r2: (cmpxchg only) new value to write
- *
- * Return __get_user struct, r0 with value, r1 with error.
- */
-#define FUTEX_OP(name, ...) \
-STD_ENTRY(futex_##name)			\
-	__VA_ARGS__;			\
-	{				\
-	 move   r1, zero;		\
-	 jrp    lr			\
-	};				\
-	STD_ENDPROC(futex_##name);	\
-	.pushsection __ex_table,"a";	\
-	.quad 1b, get_user_fault;	\
-	.popsection
-
-	.pushsection .fixup,"ax"
-get_user_fault:
-	{ movei r1, -EFAULT; jrp lr }
-	ENDPROC(get_user_fault)
-	.popsection
-
-FUTEX_OP(cmpxchg, mtspr CMPEXCH_VALUE, r1; 1: cmpexch4 r0, r0, r2)
-FUTEX_OP(set, 1: exch4 r0, r0, r1)
-FUTEX_OP(add, 1: fetchadd4 r0, r0, r1)
-FUTEX_OP(or, 1: fetchor4 r0, r0, r1)
-FUTEX_OP(andn, nor r1, r1, zero; 1: fetchand4 r0, r0, r1)
diff --git a/arch/tile/kernel/setup.c b/arch/tile/kernel/setup.c
index 4c34caea9dd3..74c91729a62a 100644
--- a/arch/tile/kernel/setup.c
+++ b/arch/tile/kernel/setup.c
@@ -1268,8 +1268,7 @@ static void __init validate_va(void)
 	if ((long)VMALLOC_START >= 0)
 		early_panic(
 			"Linux VMALLOC region below the 2GB line (%#lx)!\n"
-			"Reconfigure the kernel with fewer NR_HUGE_VMAPS\n"
-			"or smaller VMALLOC_RESERVE.\n",
+			"Reconfigure the kernel with smaller VMALLOC_RESERVE.\n",
 			VMALLOC_START);
 #endif
 }
diff --git a/arch/tile/kernel/unaligned.c b/arch/tile/kernel/unaligned.c
index b425fb6a480d..b030b4e78845 100644
--- a/arch/tile/kernel/unaligned.c
+++ b/arch/tile/kernel/unaligned.c
@@ -551,8 +551,8 @@ static tilegx_bundle_bits  jit_x1_bnezt(int ra, int broff)
 /*
  * This function generates unalign fixup JIT.
  *
- * We fist find unalign load/store instruction's destination, source
- * reguisters: ra, rb and rd. and 3 scratch registers by calling
+ * We first find unalign load/store instruction's destination, source
+ * registers: ra, rb and rd. and 3 scratch registers by calling
  * find_regs(...). 3 scratch clobbers should not alias with any register
  * used in the fault bundle. Then analyze the fault bundle to determine
  * if it's a load or store, operand width, branch or address increment etc.
diff --git a/arch/tile/mm/fault.c b/arch/tile/mm/fault.c
index 4c288f199453..6c0571216a9d 100644
--- a/arch/tile/mm/fault.c
+++ b/arch/tile/mm/fault.c
@@ -149,8 +149,6 @@ static inline int vmalloc_fault(pgd_t *pgd, unsigned long address)
 	pmd_k = vmalloc_sync_one(pgd, address);
 	if (!pmd_k)
 		return -1;
-	if (pmd_huge(*pmd_k))
-		return 0;   /* support TILE huge_vmap() API */
 	pte_k = pte_offset_kernel(pmd_k, address);
 	if (!pte_present(*pte_k))
 		return -1;
diff --git a/arch/tile/mm/init.c b/arch/tile/mm/init.c
index 4e316deb92fd..0fa1acfac79a 100644
--- a/arch/tile/mm/init.c
+++ b/arch/tile/mm/init.c
@@ -828,10 +828,6 @@ void __init mem_init(void)
 	printk(KERN_DEBUG "  PKMAP   %#lx - %#lx\n",
 	       PKMAP_BASE, PKMAP_ADDR(LAST_PKMAP) - 1);
 #endif
-#ifdef CONFIG_HUGEVMAP
-	printk(KERN_DEBUG "  HUGEMAP %#lx - %#lx\n",
-	       HUGE_VMAP_BASE, HUGE_VMAP_END - 1);
-#endif
 	printk(KERN_DEBUG "  VMALLOC %#lx - %#lx\n",
 	       _VMALLOC_START, _VMALLOC_END - 1);
 #ifdef __tilegx__
diff --git a/arch/tile/mm/pgtable.c b/arch/tile/mm/pgtable.c
index 2deaddf3e01f..4fd9ec0b58ed 100644
--- a/arch/tile/mm/pgtable.c
+++ b/arch/tile/mm/pgtable.c
@@ -127,8 +127,7 @@ void shatter_huge_page(unsigned long addr)
 	}
 
 	/* Shatter the huge page into the preallocated L2 page table. */
-	pmd_populate_kernel(&init_mm, pmd,
-			    get_prealloc_pte(pte_pfn(*(pte_t *)pmd)));
+	pmd_populate_kernel(&init_mm, pmd, get_prealloc_pte(pmd_pfn(*pmd)));
 
 #ifdef __PAGETABLE_PMD_FOLDED
 	/* Walk every pgd on the system and update the pmd there. */
diff --git a/arch/x86/Kconfig b/arch/x86/Kconfig
index e241a1930c98..ee2fb9d37745 100644
--- a/arch/x86/Kconfig
+++ b/arch/x86/Kconfig
@@ -481,11 +481,12 @@ config X86_INTEL_LPSS
 	bool "Intel Low Power Subsystem Support"
 	depends on ACPI
 	select COMMON_CLK
+	select PINCTRL
 	---help---
 	  Select to build support for Intel Low Power Subsystem such as
 	  found on Intel Lynxpoint PCH. Selecting this option enables
-	  things like clock tree (common clock framework) which are needed
-	  by the LPSS peripheral drivers.
+	  things like clock tree (common clock framework) and pincontrol
+	  which are needed by the LPSS peripheral drivers.
 
 config X86_RDC321X
 	bool "RDC R-321x SoC"
diff --git a/arch/x86/include/asm/xen/page.h b/arch/x86/include/asm/xen/page.h
index 6aef9fbc09b7..b913915e8e63 100644
--- a/arch/x86/include/asm/xen/page.h
+++ b/arch/x86/include/asm/xen/page.h
@@ -79,30 +79,38 @@ static inline int phys_to_machine_mapping_valid(unsigned long pfn)
 	return get_phys_to_machine(pfn) != INVALID_P2M_ENTRY;
 }
 
-static inline unsigned long mfn_to_pfn(unsigned long mfn)
+static inline unsigned long mfn_to_pfn_no_overrides(unsigned long mfn)
 {
 	unsigned long pfn;
-	int ret = 0;
+	int ret;
 
 	if (xen_feature(XENFEAT_auto_translated_physmap))
 		return mfn;
 
-	if (unlikely(mfn >= machine_to_phys_nr)) {
-		pfn = ~0;
-		goto try_override;
-	}
-	pfn = 0;
+	if (unlikely(mfn >= machine_to_phys_nr))
+		return ~0;
+
 	/*
 	 * The array access can fail (e.g., device space beyond end of RAM).
 	 * In such cases it doesn't matter what we return (we return garbage),
 	 * but we must handle the fault without crashing!
 	 */
 	ret = __get_user(pfn, &machine_to_phys_mapping[mfn]);
-try_override:
-	/* ret might be < 0 if there are no entries in the m2p for mfn */
 	if (ret < 0)
-		pfn = ~0;
-	else if (get_phys_to_machine(pfn) != mfn)
+		return ~0;
+
+	return pfn;
+}
+
+static inline unsigned long mfn_to_pfn(unsigned long mfn)
+{
+	unsigned long pfn;
+
+	if (xen_feature(XENFEAT_auto_translated_physmap))
+		return mfn;
+
+	pfn = mfn_to_pfn_no_overrides(mfn);
+	if (get_phys_to_machine(pfn) != mfn) {
 		/*
 		 * If this appears to be a foreign mfn (because the pfn
 		 * doesn't map back to the mfn), then check the local override
@@ -111,6 +119,7 @@ try_override:
 		 * m2p_find_override_pfn returns ~0 if it doesn't find anything.
 		 */
 		pfn = m2p_find_override_pfn(mfn, ~0);
+	}
 
 	/* 
 	 * pfn is ~0 if there are no entries in the m2p for mfn or if the
diff --git a/arch/x86/kernel/cpu/perf_event.c b/arch/x86/kernel/cpu/perf_event.c
index 8355c84b9729..897783b3302a 100644
--- a/arch/x86/kernel/cpu/perf_event.c
+++ b/arch/x86/kernel/cpu/perf_event.c
@@ -1506,7 +1506,7 @@ static int __init init_hw_perf_events(void)
 		err = amd_pmu_init();
 		break;
 	default:
-		return 0;
+		err = -ENOTSUPP;
 	}
 	if (err != 0) {
 		pr_cont("no PMU driver, software events only.\n");
@@ -1883,9 +1883,9 @@ static struct pmu pmu = {
 
 void arch_perf_update_userpage(struct perf_event_mmap_page *userpg, u64 now)
 {
-	userpg->cap_usr_time = 0;
-	userpg->cap_usr_time_zero = 0;
-	userpg->cap_usr_rdpmc = x86_pmu.attr_rdpmc;
+	userpg->cap_user_time = 0;
+	userpg->cap_user_time_zero = 0;
+	userpg->cap_user_rdpmc = x86_pmu.attr_rdpmc;
 	userpg->pmc_width = x86_pmu.cntval_bits;
 
 	if (!boot_cpu_has(X86_FEATURE_CONSTANT_TSC))
@@ -1894,13 +1894,13 @@ void arch_perf_update_userpage(struct perf_event_mmap_page *userpg, u64 now)
 	if (!boot_cpu_has(X86_FEATURE_NONSTOP_TSC))
 		return;
 
-	userpg->cap_usr_time = 1;
+	userpg->cap_user_time = 1;
 	userpg->time_mult = this_cpu_read(cyc2ns);
 	userpg->time_shift = CYC2NS_SCALE_FACTOR;
 	userpg->time_offset = this_cpu_read(cyc2ns_offset) - now;
 
 	if (sched_clock_stable && !check_tsc_disabled()) {
-		userpg->cap_usr_time_zero = 1;
+		userpg->cap_user_time_zero = 1;
 		userpg->time_zero = this_cpu_read(cyc2ns_offset);
 	}
 }
diff --git a/arch/x86/kernel/cpu/perf_event_intel.c b/arch/x86/kernel/cpu/perf_event_intel.c
index c62d88396ad5..f31a1655d1ff 100644
--- a/arch/x86/kernel/cpu/perf_event_intel.c
+++ b/arch/x86/kernel/cpu/perf_event_intel.c
@@ -899,8 +899,8 @@ static __initconst const u64 atom_hw_cache_event_ids
 static struct extra_reg intel_slm_extra_regs[] __read_mostly =
 {
 	/* must define OFFCORE_RSP_X first, see intel_fixup_er() */
-	INTEL_UEVENT_EXTRA_REG(0x01b7, MSR_OFFCORE_RSP_0, 0x768005ffff, RSP_0),
-	INTEL_UEVENT_EXTRA_REG(0x02b7, MSR_OFFCORE_RSP_1, 0x768005ffff, RSP_1),
+	INTEL_UEVENT_EXTRA_REG(0x01b7, MSR_OFFCORE_RSP_0, 0x768005ffffull, RSP_0),
+	INTEL_UEVENT_EXTRA_REG(0x02b7, MSR_OFFCORE_RSP_1, 0x768005ffffull, RSP_1),
 	EVENT_EXTRA_END
 };
 
@@ -2325,6 +2325,7 @@ __init int intel_pmu_init(void)
 		break;
 
 	case 55: /* Atom 22nm "Silvermont" */
+	case 77: /* Avoton "Silvermont" */
 		memcpy(hw_cache_event_ids, slm_hw_cache_event_ids,
 			sizeof(hw_cache_event_ids));
 		memcpy(hw_cache_extra_regs, slm_hw_cache_extra_regs,
diff --git a/arch/x86/kernel/cpu/perf_event_intel_ds.c b/arch/x86/kernel/cpu/perf_event_intel_ds.c
index 63438aad177f..ab3ba1c1b7dd 100644
--- a/arch/x86/kernel/cpu/perf_event_intel_ds.c
+++ b/arch/x86/kernel/cpu/perf_event_intel_ds.c
@@ -584,6 +584,7 @@ struct event_constraint intel_snb_pebs_event_constraints[] = {
 	INTEL_EVENT_CONSTRAINT(0xd0, 0xf),    /* MEM_UOP_RETIRED.* */
 	INTEL_EVENT_CONSTRAINT(0xd1, 0xf),    /* MEM_LOAD_UOPS_RETIRED.* */
 	INTEL_EVENT_CONSTRAINT(0xd2, 0xf),    /* MEM_LOAD_UOPS_LLC_HIT_RETIRED.* */
+	INTEL_EVENT_CONSTRAINT(0xd3, 0xf),    /* MEM_LOAD_UOPS_LLC_MISS_RETIRED.* */
 	INTEL_UEVENT_CONSTRAINT(0x02d4, 0xf), /* MEM_LOAD_UOPS_MISC_RETIRED.LLC_MISS */
 	EVENT_CONSTRAINT_END
 };
diff --git a/arch/x86/kernel/cpu/perf_event_intel_uncore.c b/arch/x86/kernel/cpu/perf_event_intel_uncore.c
index 8ed44589b0e4..4118f9f68315 100644
--- a/arch/x86/kernel/cpu/perf_event_intel_uncore.c
+++ b/arch/x86/kernel/cpu/perf_event_intel_uncore.c
@@ -2706,14 +2706,14 @@ static void uncore_pmu_init_hrtimer(struct intel_uncore_box *box)
 	box->hrtimer.function = uncore_pmu_hrtimer;
 }
 
-struct intel_uncore_box *uncore_alloc_box(struct intel_uncore_type *type, int cpu)
+static struct intel_uncore_box *uncore_alloc_box(struct intel_uncore_type *type, int node)
 {
 	struct intel_uncore_box *box;
 	int i, size;
 
 	size = sizeof(*box) + type->num_shared_regs * sizeof(struct intel_uncore_extra_reg);
 
-	box = kzalloc_node(size, GFP_KERNEL, cpu_to_node(cpu));
+	box = kzalloc_node(size, GFP_KERNEL, node);
 	if (!box)
 		return NULL;
 
@@ -3031,7 +3031,7 @@ static int uncore_validate_group(struct intel_uncore_pmu *pmu,
 	struct intel_uncore_box *fake_box;
 	int ret = -EINVAL, n;
 
-	fake_box = uncore_alloc_box(pmu->type, smp_processor_id());
+	fake_box = uncore_alloc_box(pmu->type, NUMA_NO_NODE);
 	if (!fake_box)
 		return -ENOMEM;
 
@@ -3294,7 +3294,7 @@ static int uncore_pci_probe(struct pci_dev *pdev, const struct pci_device_id *id
 	}
 
 	type = pci_uncores[UNCORE_PCI_DEV_TYPE(id->driver_data)];
-	box = uncore_alloc_box(type, 0);
+	box = uncore_alloc_box(type, NUMA_NO_NODE);
 	if (!box)
 		return -ENOMEM;
 
@@ -3499,7 +3499,7 @@ static int uncore_cpu_prepare(int cpu, int phys_id)
 			if (pmu->func_id < 0)
 				pmu->func_id = j;
 
-			box = uncore_alloc_box(type, cpu);
+			box = uncore_alloc_box(type, cpu_to_node(cpu));
 			if (!box)
 				return -ENOMEM;
 
diff --git a/arch/x86/kernel/entry_64.S b/arch/x86/kernel/entry_64.S
index 1b69951a81e2..b077f4cc225a 100644
--- a/arch/x86/kernel/entry_64.S
+++ b/arch/x86/kernel/entry_64.S
@@ -487,21 +487,6 @@ ENDPROC(native_usergs_sysret64)
 	TRACE_IRQS_OFF
 	.endm
 
-ENTRY(save_rest)
-	PARTIAL_FRAME 1 (REST_SKIP+8)
-	movq 5*8+16(%rsp), %r11	/* save return address */
-	movq_cfi rbx, RBX+16
-	movq_cfi rbp, RBP+16
-	movq_cfi r12, R12+16
-	movq_cfi r13, R13+16
-	movq_cfi r14, R14+16
-	movq_cfi r15, R15+16
-	movq %r11, 8(%rsp)	/* return address */
-	FIXUP_TOP_OF_STACK %r11, 16
-	ret
-	CFI_ENDPROC
-END(save_rest)
-
 /* save complete stack frame */
 	.pushsection .kprobes.text, "ax"
 ENTRY(save_paranoid)
diff --git a/arch/x86/kernel/microcode_amd.c b/arch/x86/kernel/microcode_amd.c
index 7123b5df479d..af99f71aeb7f 100644
--- a/arch/x86/kernel/microcode_amd.c
+++ b/arch/x86/kernel/microcode_amd.c
@@ -216,6 +216,7 @@ int apply_microcode_amd(int cpu)
 	/* need to apply patch? */
 	if (rev >= mc_amd->hdr.patch_id) {
 		c->microcode = rev;
+		uci->cpu_sig.rev = rev;
 		return 0;
 	}
 
diff --git a/arch/x86/kernel/reboot.c b/arch/x86/kernel/reboot.c
index 563ed91e6faa..e643e744e4d8 100644
--- a/arch/x86/kernel/reboot.c
+++ b/arch/x86/kernel/reboot.c
@@ -352,12 +352,28 @@ static struct dmi_system_id __initdata reboot_dmi_table[] = {
 	},
 	{	/* Handle problems with rebooting on the Precision M6600. */
 		.callback = set_pci_reboot,
-		.ident = "Dell OptiPlex 990",
+		.ident = "Dell Precision M6600",
 		.matches = {
 			DMI_MATCH(DMI_SYS_VENDOR, "Dell Inc."),
 			DMI_MATCH(DMI_PRODUCT_NAME, "Precision M6600"),
 		},
 	},
+	{	/* Handle problems with rebooting on the Dell PowerEdge C6100. */
+		.callback = set_pci_reboot,
+		.ident = "Dell PowerEdge C6100",
+		.matches = {
+			DMI_MATCH(DMI_SYS_VENDOR, "Dell Inc."),
+			DMI_MATCH(DMI_PRODUCT_NAME, "C6100"),
+		},
+	},
+	{	/* Some C6100 machines were shipped with vendor being 'Dell'. */
+		.callback = set_pci_reboot,
+		.ident = "Dell PowerEdge C6100",
+		.matches = {
+			DMI_MATCH(DMI_SYS_VENDOR, "Dell"),
+			DMI_MATCH(DMI_PRODUCT_NAME, "C6100"),
+		},
+	},
 	{ }
 };
 
diff --git a/arch/x86/kernel/smpboot.c b/arch/x86/kernel/smpboot.c
index aecc98a93d1b..6cacab671f9b 100644
--- a/arch/x86/kernel/smpboot.c
+++ b/arch/x86/kernel/smpboot.c
@@ -653,6 +653,7 @@ static void announce_cpu(int cpu, int apicid)
 {
 	static int current_node = -1;
 	int node = early_cpu_to_node(cpu);
+	int max_cpu_present = find_last_bit(cpumask_bits(cpu_present_mask), NR_CPUS);
 
 	if (system_state == SYSTEM_BOOTING) {
 		if (node != current_node) {
@@ -661,7 +662,7 @@ static void announce_cpu(int cpu, int apicid)
 			current_node = node;
 			pr_info("Booting Node %3d, Processors ", node);
 		}
-		pr_cont(" #%d%s", cpu, cpu == (nr_cpu_ids - 1) ? " OK\n" : "");
+		pr_cont(" #%4d%s", cpu, cpu == max_cpu_present ? " OK\n" : "");
 		return;
 	} else
 		pr_info("Booting Node %d Processor %d APIC 0x%x\n",
diff --git a/arch/x86/kvm/emulate.c b/arch/x86/kvm/emulate.c
index 2bc1e81045b0..ddc3f3d2afdb 100644
--- a/arch/x86/kvm/emulate.c
+++ b/arch/x86/kvm/emulate.c
@@ -2025,6 +2025,17 @@ static int em_ret_far(struct x86_emulate_ctxt *ctxt)
 	return rc;
 }
 
+static int em_ret_far_imm(struct x86_emulate_ctxt *ctxt)
+{
+        int rc;
+
+        rc = em_ret_far(ctxt);
+        if (rc != X86EMUL_CONTINUE)
+                return rc;
+        rsp_increment(ctxt, ctxt->src.val);
+        return X86EMUL_CONTINUE;
+}
+
 static int em_cmpxchg(struct x86_emulate_ctxt *ctxt)
 {
 	/* Save real source value, then compare EAX against destination. */
@@ -3763,7 +3774,8 @@ static const struct opcode opcode_table[256] = {
 	G(ByteOp, group11), G(0, group11),
 	/* 0xC8 - 0xCF */
 	I(Stack | SrcImmU16 | Src2ImmByte, em_enter), I(Stack, em_leave),
-	N, I(ImplicitOps | Stack, em_ret_far),
+	I(ImplicitOps | Stack | SrcImmU16, em_ret_far_imm),
+	I(ImplicitOps | Stack, em_ret_far),
 	D(ImplicitOps), DI(SrcImmByte, intn),
 	D(ImplicitOps | No64), II(ImplicitOps, em_iret, iret),
 	/* 0xD0 - 0xD7 */
diff --git a/arch/x86/kvm/paging_tmpl.h b/arch/x86/kvm/paging_tmpl.h
index 043330159179..ad75d77999d0 100644
--- a/arch/x86/kvm/paging_tmpl.h
+++ b/arch/x86/kvm/paging_tmpl.h
@@ -99,6 +99,7 @@ struct guest_walker {
 	pt_element_t prefetch_ptes[PTE_PREFETCH_NUM];
 	gpa_t pte_gpa[PT_MAX_FULL_LEVELS];
 	pt_element_t __user *ptep_user[PT_MAX_FULL_LEVELS];
+	bool pte_writable[PT_MAX_FULL_LEVELS];
 	unsigned pt_access;
 	unsigned pte_access;
 	gfn_t gfn;
@@ -235,6 +236,22 @@ static int FNAME(update_accessed_dirty_bits)(struct kvm_vcpu *vcpu,
 		if (pte == orig_pte)
 			continue;
 
+		/*
+		 * If the slot is read-only, simply do not process the accessed
+		 * and dirty bits.  This is the correct thing to do if the slot
+		 * is ROM, and page tables in read-as-ROM/write-as-MMIO slots
+		 * are only supported if the accessed and dirty bits are already
+		 * set in the ROM (so that MMIO writes are never needed).
+		 *
+		 * Note that NPT does not allow this at all and faults, since
+		 * it always wants nested page table entries for the guest
+		 * page tables to be writable.  And EPT works but will simply
+		 * overwrite the read-only memory to set the accessed and dirty
+		 * bits.
+		 */
+		if (unlikely(!walker->pte_writable[level - 1]))
+			continue;
+
 		ret = FNAME(cmpxchg_gpte)(vcpu, mmu, ptep_user, index, orig_pte, pte);
 		if (ret)
 			return ret;
@@ -309,7 +326,8 @@ retry_walk:
 			goto error;
 		real_gfn = gpa_to_gfn(real_gfn);
 
-		host_addr = gfn_to_hva(vcpu->kvm, real_gfn);
+		host_addr = gfn_to_hva_prot(vcpu->kvm, real_gfn,
+					    &walker->pte_writable[walker->level - 1]);
 		if (unlikely(kvm_is_error_hva(host_addr)))
 			goto error;
 
diff --git a/arch/x86/kvm/vmx.c b/arch/x86/kvm/vmx.c
index 1f1da43ff2a2..a1216de9ffda 100644
--- a/arch/x86/kvm/vmx.c
+++ b/arch/x86/kvm/vmx.c
@@ -5339,6 +5339,15 @@ static int handle_ept_violation(struct kvm_vcpu *vcpu)
 		return 0;
 	}
 
+	/*
+	 * EPT violation happened while executing iret from NMI,
+	 * "blocked by NMI" bit has to be set before next VM entry.
+	 * There are errata that may cause this bit to not be set:
+	 * AAK134, BY25.
+	 */
+	if (exit_qualification & INTR_INFO_UNBLOCK_NMI)
+		vmcs_set_bits(GUEST_INTERRUPTIBILITY_INFO, GUEST_INTR_STATE_NMI);
+
 	gpa = vmcs_read64(GUEST_PHYSICAL_ADDRESS);
 	trace_kvm_page_fault(gpa, exit_qualification);
 
@@ -7766,6 +7775,10 @@ static void prepare_vmcs02(struct kvm_vcpu *vcpu, struct vmcs12 *vmcs12)
 		vmcs_write64(GUEST_PDPTR1, vmcs12->guest_pdptr1);
 		vmcs_write64(GUEST_PDPTR2, vmcs12->guest_pdptr2);
 		vmcs_write64(GUEST_PDPTR3, vmcs12->guest_pdptr3);
+		__clear_bit(VCPU_EXREG_PDPTR,
+				(unsigned long *)&vcpu->arch.regs_avail);
+		__clear_bit(VCPU_EXREG_PDPTR,
+				(unsigned long *)&vcpu->arch.regs_dirty);
 	}
 
 	kvm_register_write(vcpu, VCPU_REGS_RSP, vmcs12->guest_rsp);
diff --git a/arch/x86/platform/efi/efi.c b/arch/x86/platform/efi/efi.c
index 90f6ed127096..c7e22ab29a5a 100644
--- a/arch/x86/platform/efi/efi.c
+++ b/arch/x86/platform/efi/efi.c
@@ -912,10 +912,13 @@ void __init efi_enter_virtual_mode(void)
 
 	for (p = memmap.map; p < memmap.map_end; p += memmap.desc_size) {
 		md = p;
-		if (!(md->attribute & EFI_MEMORY_RUNTIME) &&
-		    md->type != EFI_BOOT_SERVICES_CODE &&
-		    md->type != EFI_BOOT_SERVICES_DATA)
-			continue;
+		if (!(md->attribute & EFI_MEMORY_RUNTIME)) {
+#ifdef CONFIG_X86_64
+			if (md->type != EFI_BOOT_SERVICES_CODE &&
+			    md->type != EFI_BOOT_SERVICES_DATA)
+#endif
+				continue;
+		}
 
 		size = md->num_pages << EFI_PAGE_SHIFT;
 		end = md->phys_addr + size;
diff --git a/arch/x86/xen/p2m.c b/arch/x86/xen/p2m.c
index 8b901e8d782d..a61c7d5811be 100644
--- a/arch/x86/xen/p2m.c
+++ b/arch/x86/xen/p2m.c
@@ -879,7 +879,6 @@ int m2p_add_override(unsigned long mfn, struct page *page,
 	unsigned long uninitialized_var(address);
 	unsigned level;
 	pte_t *ptep = NULL;
-	int ret = 0;
 
 	pfn = page_to_pfn(page);
 	if (!PageHighMem(page)) {
@@ -926,8 +925,8 @@ int m2p_add_override(unsigned long mfn, struct page *page,
 	 * frontend pages while they are being shared with the backend,
 	 * because mfn_to_pfn (that ends up being called by GUPF) will
 	 * return the backend pfn rather than the frontend pfn. */
-	ret = __get_user(pfn, &machine_to_phys_mapping[mfn]);
-	if (ret == 0 && get_phys_to_machine(pfn) == mfn)
+	pfn = mfn_to_pfn_no_overrides(mfn);
+	if (get_phys_to_machine(pfn) == mfn)
 		set_phys_to_machine(pfn, FOREIGN_FRAME(mfn));
 
 	return 0;
@@ -942,7 +941,6 @@ int m2p_remove_override(struct page *page,
 	unsigned long uninitialized_var(address);
 	unsigned level;
 	pte_t *ptep = NULL;
-	int ret = 0;
 
 	pfn = page_to_pfn(page);
 	mfn = get_phys_to_machine(pfn);
@@ -1029,8 +1027,8 @@ int m2p_remove_override(struct page *page,
 	 * the original pfn causes mfn_to_pfn(mfn) to return the frontend
 	 * pfn again. */
 	mfn &= ~FOREIGN_FRAME_BIT;
-	ret = __get_user(pfn, &machine_to_phys_mapping[mfn]);
-	if (ret == 0 && get_phys_to_machine(pfn) == FOREIGN_FRAME(mfn) &&
+	pfn = mfn_to_pfn_no_overrides(mfn);
+	if (get_phys_to_machine(pfn) == FOREIGN_FRAME(mfn) &&
 			m2p_find_override(mfn) == NULL)
 		set_phys_to_machine(pfn, mfn);
 
diff --git a/arch/x86/xen/spinlock.c b/arch/x86/xen/spinlock.c
index 253f63fceea1..be6b86078957 100644
--- a/arch/x86/xen/spinlock.c
+++ b/arch/x86/xen/spinlock.c
@@ -259,6 +259,14 @@ void xen_uninit_lock_cpu(int cpu)
 }
 
 
+/*
+ * Our init of PV spinlocks is split in two init functions due to us
+ * using paravirt patching and jump labels patching and having to do
+ * all of this before SMP code is invoked.
+ *
+ * The paravirt patching needs to be done _before_ the alternative asm code
+ * is started, otherwise we would not patch the core kernel code.
+ */
 void __init xen_init_spinlocks(void)
 {
 
@@ -267,12 +275,26 @@ void __init xen_init_spinlocks(void)
 		return;
 	}
 
-	static_key_slow_inc(&paravirt_ticketlocks_enabled);
-
 	pv_lock_ops.lock_spinning = PV_CALLEE_SAVE(xen_lock_spinning);
 	pv_lock_ops.unlock_kick = xen_unlock_kick;
 }
 
+/*
+ * While the jump_label init code needs to happend _after_ the jump labels are
+ * enabled and before SMP is started. Hence we use pre-SMP initcall level
+ * init. We cannot do it in xen_init_spinlocks as that is done before
+ * jump labels are activated.
+ */
+static __init int xen_init_spinlocks_jump(void)
+{
+	if (!xen_pvspin)
+		return 0;
+
+	static_key_slow_inc(&paravirt_ticketlocks_enabled);
+	return 0;
+}
+early_initcall(xen_init_spinlocks_jump);
+
 static __init int xen_parse_nopvspin(char *arg)
 {
 	xen_pvspin = false;