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

CREDITS

@@ -2808,8 +2808,7 @@ S: Ottawa, Ontario
 S: Canada K2P 0X8
 
 N: Mikael Pettersson
-E: mikpe@it.uu.se
-W: http://user.it.uu.se/~mikpe/linux/
+E: mikpelinux@gmail.com
 D: Miscellaneous fixes
 
 N: Reed H. Petty

Documentation/arm64/tagged-pointers.txt

@@ -18,17 +18,17 @@ this byte for application use, with the following caveats:
 	    parameters containing user virtual addresses *must* have
 	    their top byte cleared before trapping to the kernel.
 
-	(2) Tags are not guaranteed to be preserved when delivering
-	    signals. This means that signal handlers in applications
-	    making use of tags cannot rely on the tag information for
-	    user virtual addresses being maintained for fields inside
-	    siginfo_t. One exception to this rule is for signals raised
-	    in response to debug exceptions, where the tag information
+	(2) Non-zero tags are not preserved when delivering signals.
+	    This means that signal handlers in applications making use
+	    of tags cannot rely on the tag information for user virtual
+	    addresses being maintained for fields inside siginfo_t.
+	    One exception to this rule is for signals raised in response
+	    to watchpoint debug exceptions, where the tag information
 	    will be preserved.
 
 	(3) Special care should be taken when using tagged pointers,
 	    since it is likely that C compilers will not hazard two
-	    addresses differing only in the upper bits.
+	    virtual addresses differing only in the upper byte.
 
 The architecture prevents the use of a tagged PC, so the upper byte will
 be set to a sign-extension of bit 55 on exception return.

Documentation/devicetree/bindings/mmc/exynos-dw-mshc.txt

@@ -1,11 +1,11 @@
-* Samsung Exynos specific extensions to the Synopsis Designware Mobile
+* Samsung Exynos specific extensions to the Synopsys Designware Mobile
   Storage Host Controller
 
-The Synopsis designware mobile storage host controller is used to interface
+The Synopsys designware mobile storage host controller is used to interface
 a SoC with storage medium such as eMMC or SD/MMC cards. This file documents
-differences between the core Synopsis dw mshc controller properties described
-by synopsis-dw-mshc.txt and the properties used by the Samsung Exynos specific
-extensions to the Synopsis Designware Mobile Storage Host Controller.
+differences between the core Synopsys dw mshc controller properties described
+by synopsys-dw-mshc.txt and the properties used by the Samsung Exynos specific
+extensions to the Synopsys Designware Mobile Storage Host Controller.
 
 Required Properties:
 

Documentation/devicetree/bindings/mmc/rockchip-dw-mshc.txt

@@ -1,11 +1,11 @@
-* Rockchip specific extensions to the Synopsis Designware Mobile
+* Rockchip specific extensions to the Synopsys Designware Mobile
   Storage Host Controller
 
-The Synopsis designware mobile storage host controller is used to interface
+The Synopsys designware mobile storage host controller is used to interface
 a SoC with storage medium such as eMMC or SD/MMC cards. This file documents
-differences between the core Synopsis dw mshc controller properties described
-by synopsis-dw-mshc.txt and the properties used by the Rockchip specific
-extensions to the Synopsis Designware Mobile Storage Host Controller.
+differences between the core Synopsys dw mshc controller properties described
+by synopsys-dw-mshc.txt and the properties used by the Rockchip specific
+extensions to the Synopsys Designware Mobile Storage Host Controller.
 
 Required Properties:
 

Documentation/devicetree/bindings/mmc/synopsis-dw-mshc.txt → Documentation/devicetree/bindings/mmc/synopsys-dw-mshc.txt

@@ -1,14 +1,14 @@
-* Synopsis Designware Mobile Storage Host Controller
+* Synopsys Designware Mobile Storage Host Controller
 
-The Synopsis designware mobile storage host controller is used to interface
+The Synopsys designware mobile storage host controller is used to interface
 a SoC with storage medium such as eMMC or SD/MMC cards. This file documents
 differences between the core mmc properties described by mmc.txt and the
-properties used by the Synopsis Designware Mobile Storage Host Controller.
+properties used by the Synopsys Designware Mobile Storage Host Controller.
 
 Required Properties:
 
 * compatible: should be
-	- snps,dw-mshc: for controllers compliant with synopsis dw-mshc.
+	- snps,dw-mshc: for controllers compliant with synopsys dw-mshc.
 * #address-cells: should be 1.
 * #size-cells: should be 0.
 

Documentation/devicetree/bindings/pci/designware-pcie.txt

@@ -1,4 +1,4 @@
-* Synopsis Designware PCIe interface
+* Synopsys Designware PCIe interface
 
 Required properties:
 - compatible: should contain "snps,dw-pcie" to identify the

Documentation/filesystems/vfs.txt

@@ -359,11 +359,9 @@ struct inode_operations {
 	ssize_t (*listxattr) (struct dentry *, char *, size_t);
 	int (*removexattr) (struct dentry *, const char *);
 	void (*update_time)(struct inode *, struct timespec *, int);
-	int (*atomic_open)(struct inode *, struct dentry *,
+	int (*atomic_open)(struct inode *, struct dentry *, struct file *,
+			unsigned open_flag, umode_t create_mode, int *opened);
 	int (*tmpfile) (struct inode *, struct dentry *, umode_t);
-} ____cacheline_aligned;
-				struct file *, unsigned open_flag,
-				umode_t create_mode, int *opened);
 };
 
 Again, all methods are called without any locks being held, unless
@@ -470,9 +468,11 @@ otherwise noted.
   	method the filesystem can look up, possibly create and open the file in
   	one atomic operation.  If it cannot perform this (e.g. the file type
   	turned out to be wrong) it may signal this by returning 1 instead of
-  	usual 0 or -ve .  This method is only called if the last
-  	component is negative or needs lookup.  Cached positive dentries are
-  	still handled by f_op->open().
+	usual 0 or -ve .  This method is only called if the last component is
+	negative or needs lookup.  Cached positive dentries are still handled by
+	f_op->open().  If the file was created, the FILE_CREATED flag should be
+	set in "opened".  In case of O_EXCL the method must only succeed if the
+	file didn't exist and hence FILE_CREATED shall always be set on success.
 
   tmpfile: called in the end of O_TMPFILE open().  Optional, equivalent to
 	atomically creating, opening and unlinking a file in given directory.

Documentation/kernel-parameters.txt

@@ -3485,6 +3485,10 @@ bytes respectively. Such letter suffixes can also be entirely omitted.
 				the unplug protocol
 			never -- do not unplug even if version check succeeds
 
+	xen_nopvspin	[X86,XEN]
+			Disables the ticketlock slowpath using Xen PV
+			optimizations.
+
 	xirc2ps_cs=	[NET,PCMCIA]
 			Format:
 			<irq>,<irq_mask>,<io>,<full_duplex>,<do_sound>,<lockup_hack>[,<irq2>[,<irq3>[,<irq4>]]]

Documentation/networking/bonding.txt

@@ -1362,6 +1362,12 @@ To add ARP targets:
 To remove an ARP target:
 # echo -192.168.0.100 > /sys/class/net/bond0/bonding/arp_ip_target
 
+To configure the interval between learning packet transmits:
+# echo 12 > /sys/class/net/bond0/bonding/lp_interval
+	NOTE: the lp_inteval is the number of seconds between instances where
+the bonding driver sends learning packets to each slaves peer switch.  The
+default interval is 1 second.
+
 Example Configuration
 ---------------------
 	We begin with the same example that is shown in section 3.3,

Documentation/scheduler/sched-design-CFS.txt

@@ -66,9 +66,7 @@ rq->cfs.load value, which is the sum of the weights of the tasks queued on the
 runqueue.
 
 CFS maintains a time-ordered rbtree, where all runnable tasks are sorted by the
-p->se.vruntime key (there is a subtraction using rq->cfs.min_vruntime to
-account for possible wraparounds).  CFS picks the "leftmost" task from this
-tree and sticks to it.
+p->se.vruntime key. CFS picks the "leftmost" task from this tree and sticks to it.
 As the system progresses forwards, the executed tasks are put into the tree
 more and more to the right --- slowly but surely giving a chance for every task
 to become the "leftmost task" and thus get on the CPU within a deterministic

Documentation/sound/alsa/HD-Audio-Models.txt

@@ -296,6 +296,12 @@ Cirrus Logic CS4206/4207
   imac27	IMac 27 Inch
   auto		BIOS setup (default)
 
+Cirrus Logic CS4208
+===================
+  mba6		MacBook Air 6,1 and 6,2
+  gpio0		Enable GPIO 0 amp
+  auto		BIOS setup (default)
+
 VIA VT17xx/VT18xx/VT20xx
 ========================
   auto		BIOS setup (default)

MAINTAINERS

@@ -1812,7 +1812,8 @@ S:	Supported
 F:	drivers/net/ethernet/broadcom/bnx2x/
 
 BROADCOM BCM281XX/BCM11XXX ARM ARCHITECTURE
-M:	Christian Daudt <csd@broadcom.com>
+M:	Christian Daudt <bcm@fixthebug.org>
+L:	bcm-kernel-feedback-list@broadcom.com
 T:	git git://git.github.com/broadcom/bcm11351
 S:	Maintained
 F:	arch/arm/mach-bcm/
@@ -2639,6 +2640,18 @@ F:	include/linux/device-mapper.h
 F:	include/linux/dm-*.h
 F:	include/uapi/linux/dm-*.h
 
+DIGI NEO AND CLASSIC PCI PRODUCTS
+M:	Lidza Louina <lidza.louina@gmail.com>
+L:	driverdev-devel@linuxdriverproject.org
+S:	Maintained
+F:	drivers/staging/dgnc/
+
+DIGI EPCA PCI PRODUCTS
+M:	Lidza Louina <lidza.louina@gmail.com>
+L:	driverdev-devel@linuxdriverproject.org
+S:	Maintained
+F:	drivers/staging/dgap/
+
 DIOLAN U2C-12 I2C DRIVER
 M:	Guenter Roeck <linux@roeck-us.net>
 L:	linux-i2c@vger.kernel.org
@@ -6595,7 +6608,7 @@ S:	Obsolete
 F:	drivers/net/wireless/prism54/
 
 PROMISE SATA TX2/TX4 CONTROLLER LIBATA DRIVER
-M:	Mikael Pettersson <mikpe@it.uu.se>
+M:	Mikael Pettersson <mikpelinux@gmail.com>
 L:	linux-ide@vger.kernel.org
 S:	Maintained
 F:	drivers/ata/sata_promise.*
@@ -8724,9 +8737,8 @@ F:	Documentation/hid/hiddev.txt
 F:	drivers/hid/usbhid/
 
 USB/IP DRIVERS
-M:	Matt Mooney <mfm@muteddisk.com>
 L:	linux-usb@vger.kernel.org
-S:	Maintained
+S:	Orphan
 F:	drivers/staging/usbip/
 
 USB ISP116X DRIVER

Makefile

@@ -1,7 +1,7 @@
 VERSION = 3
 PATCHLEVEL = 12
 SUBLEVEL = 0
-EXTRAVERSION = -rc1
+EXTRAVERSION = -rc3
 NAME = One Giant Leap for Frogkind
 
 # *DOCUMENTATION*

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
 

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.
 

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

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>;
+};

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"

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;
+};

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";
 			};

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";
 			};
 

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

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 {

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>;

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 {

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 {

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 {

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

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

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

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
 		}

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);

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

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");

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;

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

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
 

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)

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

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;

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);

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));

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);
 	}
 }

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);

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),

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();
 

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;
 	}
 

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,

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.

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);

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,

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]),

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]),

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

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;
 }

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

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");

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;

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;

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.

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");

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;

arch/mips/boot/dts/include/dt-bindings

@@ -1 +1 @@
-../../../../../include/dt-bindings
+../../../../../include/dt-bindings

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>

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>

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.
  */

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);

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 */

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 {

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;

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.
  */

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
  */

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
  */

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)

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.

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__ */

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))

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;

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;

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>

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;

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 */

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);

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)

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>

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:

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>

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>

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();

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;

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) {

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"
 

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>

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);

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);

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";

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 */

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))

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);
+}