Merge tag 'sh-for-4.8' of git://git.libc.org/linux-sh

Pull arch/sh updates from Rich Felker:
 "These changes improve device tree support (including builtin DTB), add
  support for the J-Core J2 processor, an open source synthesizable
  reimplementation of the SH-2 ISA, resolve a longstanding sigcontext
  ABI mismatch issue, and fix various bugs including nommu-specific
  issues and minor regressions introduced in 4.6.

  The J-Core arch support is included here but to be usable it needs
  drivers that are waiting on approval/inclusion from their subsystem
  maintainers"

* tag 'sh-for-4.8' of git://git.libc.org/linux-sh: (23 commits)
  sh: add device tree source for J2 FPGA on Mimas v2 board
  sh: add defconfig for J-Core J2
  sh: use common clock framework with device tree boards
  sh: system call wire up
  sh: Delete unnecessary checks before the function call "mempool_destroy"
  sh: do not perform IPI-based cache flush except on boards that need it
  sh: add SMP support for J2
  sh: SMP support for SH2 entry.S
  sh: add working futex atomic ops on userspace addresses for smp
  sh: add J2 atomics using the cas.l instruction
  sh: add AT_HWCAP flag for J-Core cas.l instruction
  sh: add support for J-Core J2 processor
  sh: fix build regression with CONFIG_OF && !CONFIG_OF_FLATTREE
  sh: allow clocksource drivers to register sched_clock backends
  sh: make heartbeat driver explicitly non-modular
  sh: make board-secureedge5410 explicitly non-modular
  sh: make mm/asids-debugfs explicitly non-modular
  sh: make time.c explicitly non-modular
  sh: fix futex/robust_list on nommu models
  sh: disable aliased page logic on NOMMU models
  ...

arch/sh/Kconfig

@@ -38,6 +38,7 @@ config SUPERH
 	select GENERIC_IDLE_POLL_SETUP
 	select GENERIC_CLOCKEVENTS
 	select GENERIC_CMOS_UPDATE if SH_SH03 || SH_DREAMCAST
+	select GENERIC_SCHED_CLOCK
 	select GENERIC_STRNCPY_FROM_USER
 	select GENERIC_STRNLEN_USER
 	select HAVE_MOD_ARCH_SPECIFIC if DWARF_UNWINDER
@@ -45,6 +46,7 @@ config SUPERH
 	select OLD_SIGSUSPEND
 	select OLD_SIGACTION
 	select HAVE_ARCH_AUDITSYSCALL
+	select HAVE_FUTEX_CMPXCHG if FUTEX
 	select HAVE_NMI
 	help
 	  The SuperH is a RISC processor targeted for use in embedded systems
@@ -184,6 +186,12 @@ config CPU_SH2A
 	select CPU_SH2
 	select UNCACHED_MAPPING
 
+config CPU_J2
+	bool
+	select CPU_SH2
+	select OF
+	select OF_EARLY_FLATTREE
+
 config CPU_SH3
 	bool
 	select CPU_HAS_INTEVT
@@ -250,6 +258,12 @@ config CPU_SUBTYPE_SH7619
 	select CPU_SH2
 	select SYS_SUPPORTS_SH_CMT
 
+config CPU_SUBTYPE_J2
+	bool "Support J2 processor"
+	select CPU_J2
+	select SYS_SUPPORTS_SMP
+	select GENERIC_CLOCKEVENTS_BROADCAST if SMP
+
 # SH-2A Processor Support
 
 config CPU_SUBTYPE_SH7201
@@ -739,6 +753,26 @@ endmenu
 
 menu "Boot options"
 
+config USE_BUILTIN_DTB
+	bool "Use builtin DTB"
+	default n
+	depends on SH_DEVICE_TREE
+	help
+	  Link a device tree blob for particular hardware into the kernel,
+	  suppressing use of the DTB pointer provided by the bootloader.
+	  This option should only be used with legacy bootloaders that are
+	  not capable of providing a DTB to the kernel, or for experimental
+	  hardware without stable device tree bindings.
+
+config BUILTIN_DTB_SOURCE
+	string "Source file for builtin DTB"
+	default ""
+	depends on USE_BUILTIN_DTB
+	help
+	  Base name (without suffix, relative to arch/sh/boot/dts) for the
+	  a DTS file that will be used to produce the DTB linked into the
+	  kernel.
+
 config ZERO_PAGE_OFFSET
 	hex
 	default "0x00010000" if PAGE_SIZE_64KB || SH_RTS7751R2D || \

arch/sh/Makefile

@@ -31,6 +31,7 @@ isa-y					:= $(isa-y)-up
 endif
 
 cflags-$(CONFIG_CPU_SH2)		:= $(call cc-option,-m2,)
+cflags-$(CONFIG_CPU_J2)			:= $(call cc-option,-mj2,)
 cflags-$(CONFIG_CPU_SH2A)		+= $(call cc-option,-m2a,) \
 					   $(call cc-option,-m2a-nofpu,) \
 					   $(call cc-option,-m4-nofpu,)
@@ -130,6 +131,8 @@ head-y	:= arch/sh/kernel/head_$(BITS).o
 core-y				+= arch/sh/kernel/ arch/sh/mm/ arch/sh/boards/
 core-$(CONFIG_SH_FPU_EMU)	+= arch/sh/math-emu/
 
+core-$(CONFIG_USE_BUILTIN_DTB)	+= arch/sh/boot/dts/
+
 # Mach groups
 machdir-$(CONFIG_SOLUTION_ENGINE)		+= mach-se
 machdir-$(CONFIG_SH_HP6XX)			+= mach-hp6xx

arch/sh/boards/Kconfig

@@ -11,6 +11,7 @@ config SH_DEVICE_TREE
 	select OF
 	select OF_EARLY_FLATTREE
 	select CLKSRC_OF
+	select COMMON_CLK
 	select GENERIC_CALIBRATE_DELAY
 	help
 	  Select Board Described by Device Tree to build a kernel that

arch/sh/boards/board-secureedge5410.c

@@ -14,7 +14,6 @@
 #include <linux/interrupt.h>
 #include <linux/timer.h>
 #include <linux/delay.h>
-#include <linux/module.h>
 #include <linux/sched.h>
 #include <asm/machvec.h>
 #include <mach/secureedge5410.h>
@@ -49,7 +48,7 @@ static int __init eraseconfig_init(void)
 				irq);
 	return 0;
 }
-module_init(eraseconfig_init);
+device_initcall(eraseconfig_init);
 
 /*
  * Initialize IRQ setting

arch/sh/boards/of-generic.c

@@ -124,13 +124,22 @@ static void __init sh_of_time_init(void)
 
 static void __init sh_of_setup(char **cmdline_p)
 {
+	struct device_node *root;
+
+#ifdef CONFIG_USE_BUILTIN_DTB
+	unflatten_and_copy_device_tree();
+#else
 	unflatten_device_tree();
+#endif
 
 	board_time_init = sh_of_time_init;
 
-	sh_mv.mv_name = of_flat_dt_get_machine_name();
-	if (!sh_mv.mv_name)
-		sh_mv.mv_name = "Unknown SH model";
+	sh_mv.mv_name = "Unknown SH model";
+	root = of_find_node_by_path("/");
+	if (root) {
+		of_property_read_string(root, "model", &sh_mv.mv_name);
+		of_node_put(root);
+	}
 
 	sh_of_smp_probe();
 }

arch/sh/boot/dts/Makefile

+obj-$(CONFIG_USE_BUILTIN_DTB) += $(patsubst "%",%,$(CONFIG_BUILTIN_DTB_SOURCE)).dtb.o
+
+clean-files := *.dtb.S

arch/sh/boot/dts/j2_mimas_v2.dts

+/dts-v1/;
+
+/ {
+	compatible = "jcore,j2-soc";
+	model = "J2 FPGA SoC on Mimas v2 board";
+
+	#address-cells = <1>;
+	#size-cells = <1>;
+
+	interrupt-parent = <&aic>;
+
+	cpus {
+		#address-cells = <1>;
+		#size-cells = <0>;
+
+		cpu@0 {
+			device_type = "cpu";
+			compatible = "jcore,j2";
+			reg = <0>;
+			clock-frequency = <50000000>;
+			d-cache-size = <8192>;
+			i-cache-size = <8192>;
+			d-cache-block-size = <16>;
+			i-cache-block-size = <16>;
+		};
+	};
+
+	memory@10000000 {
+		device_type = "memory";
+		reg = <0x10000000 0x4000000>;
+	};
+
+	aliases {
+		serial0 = &uart0;
+		spi0 = &spi0;
+	};
+
+	chosen {
+		stdout-path = "serial0";
+	};
+
+	soc@abcd0000 {
+		compatible = "simple-bus";
+		ranges = <0 0xabcd0000 0x100000>;
+
+		#address-cells = <1>;
+		#size-cells = <1>;
+
+		aic: interrupt-controller@200 {
+			compatible = "jcore,aic1";
+			reg = <0x200 0x10>;
+			interrupt-controller;
+			#interrupt-cells = <1>;
+		};
+
+		cache-controller@c0 {
+			compatible = "jcore,cache";
+			reg = <0xc0 4>;
+		};
+
+		timer@200 {
+			compatible = "jcore,pit";
+			reg = <0x200 0x30>;
+			interrupts = <0x48>;
+		};
+
+		spi0: spi@40 {
+			compatible = "jcore,spi2";
+
+			#address-cells = <1>;
+			#size-cells = <0>;
+
+			spi-max-frequency = <25000000>;
+
+			reg = <0x40 0x8>;
+
+			sdcard@0 {
+				compatible = "mmc-spi-slot";
+				reg = <0>;
+				spi-max-frequency = <25000000>;
+				voltage-ranges = <3200 3400>;
+				mode = <0>;
+			};
+		};
+
+		uart0: serial@100 {
+			clock-frequency = <125000000>;
+			compatible = "xlnx,xps-uartlite-1.00.a";
+			current-speed = <19200>;
+			device_type = "serial";
+			interrupts = <0x12>;
+			port-number = <0>;
+			reg = <0x100 0x10>;
+		};
+	};
+};

arch/sh/configs/j2_defconfig

+CONFIG_SMP=y
+CONFIG_SYSVIPC=y
+CONFIG_POSIX_MQUEUE=y
+CONFIG_NO_HZ=y
+CONFIG_HIGH_RES_TIMERS=y
+CONFIG_CPU_SUBTYPE_J2=y
+CONFIG_MEMORY_START=0x10000000
+CONFIG_MEMORY_SIZE=0x04000000
+CONFIG_CPU_BIG_ENDIAN=y
+CONFIG_SH_DEVICE_TREE=y
+CONFIG_HZ_100=y
+CONFIG_CMDLINE_OVERWRITE=y
+CONFIG_CMDLINE="console=ttyUL0 earlycon"
+CONFIG_BINFMT_ELF_FDPIC=y
+CONFIG_BINFMT_FLAT=y
+CONFIG_NET=y
+CONFIG_PACKET=y
+CONFIG_UNIX=y
+CONFIG_INET=y
+CONFIG_DEVTMPFS=y
+CONFIG_DEVTMPFS_MOUNT=y
+CONFIG_NETDEVICES=y
+CONFIG_SERIAL_UARTLITE=y
+CONFIG_SERIAL_UARTLITE_CONSOLE=y
+CONFIG_I2C=y
+CONFIG_SPI=y
+CONFIG_SPI_JCORE=y
+CONFIG_WATCHDOG=y
+CONFIG_MMC=y
+CONFIG_MMC_SPI=y
+CONFIG_CLKSRC_JCORE_PIT=y
+CONFIG_JCORE_AIC=y
+CONFIG_EXT4_FS=y
+CONFIG_VFAT_FS=y
+CONFIG_FAT_DEFAULT_IOCHARSET="ascii"
+CONFIG_FAT_DEFAULT_UTF8=y
+CONFIG_NLS_DEFAULT="utf8"
+CONFIG_NLS_CODEPAGE_437=y
+CONFIG_NLS_ASCII=y
+CONFIG_NLS_UTF8=y

arch/sh/drivers/heartbeat.c

@@ -19,7 +19,6 @@
  * for more details.
  */
 #include <linux/init.h>
-#include <linux/module.h>
 #include <linux/platform_device.h>
 #include <linux/sched.h>
 #include <linux/timer.h>
@@ -139,26 +138,11 @@ static int heartbeat_drv_probe(struct platform_device *pdev)
 	return mod_timer(&hd->timer, jiffies + 1);
 }
 
-static int heartbeat_drv_remove(struct platform_device *pdev)
-{
-	struct heartbeat_data *hd = platform_get_drvdata(pdev);
-
-	del_timer_sync(&hd->timer);
-	iounmap(hd->base);
-
-	platform_set_drvdata(pdev, NULL);
-
-	if (!pdev->dev.platform_data)
-		kfree(hd);
-
-	return 0;
-}
-
 static struct platform_driver heartbeat_driver = {
 	.probe		= heartbeat_drv_probe,
-	.remove		= heartbeat_drv_remove,
 	.driver		= {
-		.name	= DRV_NAME,
+		.name			= DRV_NAME,
+		.suppress_bind_attrs	= true,
 	},
 };
 
@@ -167,14 +151,4 @@ static int __init heartbeat_init(void)
 	printk(KERN_NOTICE DRV_NAME ": version %s loaded\n", DRV_VERSION);
 	return platform_driver_register(&heartbeat_driver);
 }
-
-static void __exit heartbeat_exit(void)
-{
-	platform_driver_unregister(&heartbeat_driver);
-}
-module_init(heartbeat_init);
-module_exit(heartbeat_exit);
-
-MODULE_VERSION(DRV_VERSION);
-MODULE_AUTHOR("Paul Mundt");
-MODULE_LICENSE("GPL v2");
+device_initcall(heartbeat_init);

arch/sh/include/asm/atomic.h

 #ifndef __ASM_SH_ATOMIC_H
 #define __ASM_SH_ATOMIC_H
 
+#if defined(CONFIG_CPU_J2)
+
+#include <asm-generic/atomic.h>
+
+#else
+
 /*
  * Atomic operations that C can't guarantee us.  Useful for
  * resource counting etc..
@@ -63,4 +69,6 @@ static inline int __atomic_add_unless(atomic_t *v, int a, int u)
 	return c;
 }
 
+#endif /* CONFIG_CPU_J2 */
+
 #endif /* __ASM_SH_ATOMIC_H */

arch/sh/include/asm/barrier.h

@@ -29,6 +29,11 @@
 #define wmb()		mb()
 #define ctrl_barrier()	__icbi(PAGE_OFFSET)
 #else
+#if defined(CONFIG_CPU_J2) && defined(CONFIG_SMP)
+#define __smp_mb()	do { int tmp = 0; __asm__ __volatile__ ("cas.l %0,%0,@%1" : "+r"(tmp) : "z"(&tmp) : "memory", "t"); } while(0)
+#define __smp_rmb()	__smp_mb()
+#define __smp_wmb()	__smp_mb()
+#endif
 #define ctrl_barrier()	__asm__ __volatile__ ("nop;nop;nop;nop;nop;nop;nop;nop")
 #endif
 

arch/sh/include/asm/bitops-cas.h

+#ifndef __ASM_SH_BITOPS_CAS_H
+#define __ASM_SH_BITOPS_CAS_H
+
+static inline unsigned __bo_cas(volatile unsigned *p, unsigned old, unsigned new)
+{
+	__asm__ __volatile__("cas.l %1,%0,@r0"
+		: "+r"(new)
+		: "r"(old), "z"(p)
+		: "t", "memory" );
+	return new;
+}
+
+static inline void set_bit(int nr, volatile void *addr)
+{
+	unsigned mask, old;
+	volatile unsigned *a = addr;
+
+	a += nr >> 5;
+	mask = 1U << (nr & 0x1f);
+
+	do old = *a;
+	while (__bo_cas(a, old, old|mask) != old);
+}
+
+static inline void clear_bit(int nr, volatile void *addr)
+{
+	unsigned mask, old;
+	volatile unsigned *a = addr;
+
+	a += nr >> 5;
+	mask = 1U << (nr & 0x1f);
+
+	do old = *a;
+	while (__bo_cas(a, old, old&~mask) != old);
+}
+
+static inline void change_bit(int nr, volatile void *addr)
+{
+	unsigned mask, old;
+	volatile unsigned *a = addr;
+
+	a += nr >> 5;
+	mask = 1U << (nr & 0x1f);
+
+	do old = *a;
+	while (__bo_cas(a, old, old^mask) != old);
+}
+
+static inline int test_and_set_bit(int nr, volatile void *addr)
+{
+	unsigned mask, old;
+	volatile unsigned *a = addr;
+
+	a += nr >> 5;
+	mask = 1U << (nr & 0x1f);
+
+	do old = *a;
+	while (__bo_cas(a, old, old|mask) != old);
+
+	return !!(old & mask);
+}
+
+static inline int test_and_clear_bit(int nr, volatile void *addr)
+{
+	unsigned mask, old;
+	volatile unsigned *a = addr;
+
+	a += nr >> 5;
+	mask = 1U << (nr & 0x1f);
+
+	do old = *a;
+	while (__bo_cas(a, old, old&~mask) != old);
+
+	return !!(old & mask);
+}
+
+static inline int test_and_change_bit(int nr, volatile void *addr)
+{
+	unsigned mask, old;
+	volatile unsigned *a = addr;
+
+	a += nr >> 5;
+	mask = 1U << (nr & 0x1f);
+
+	do old = *a;
+	while (__bo_cas(a, old, old^mask) != old);
+
+	return !!(old & mask);
+}
+
+#include <asm-generic/bitops/non-atomic.h>
+
+#endif /* __ASM_SH_BITOPS_CAS_H */

arch/sh/include/asm/bitops.h

@@ -18,6 +18,8 @@
 #include <asm/bitops-op32.h>
 #elif defined(CONFIG_CPU_SH4A)
 #include <asm/bitops-llsc.h>
+#elif defined(CONFIG_CPU_J2) && defined(CONFIG_SMP)
+#include <asm/bitops-cas.h>
 #else
 #include <asm-generic/bitops/atomic.h>
 #include <asm-generic/bitops/non-atomic.h>

arch/sh/include/asm/cmpxchg-cas.h

+#ifndef __ASM_SH_CMPXCHG_CAS_H
+#define __ASM_SH_CMPXCHG_CAS_H
+
+static inline unsigned long
+__cmpxchg_u32(volatile u32 *m, unsigned long old, unsigned long new)
+{
+	__asm__ __volatile__("cas.l %1,%0,@r0"
+		: "+r"(new)
+		: "r"(old), "z"(m)
+		: "t", "memory" );
+	return new;
+}
+
+static inline unsigned long xchg_u32(volatile u32 *m, unsigned long val)
+{
+	unsigned long old;
+	do old = *m;
+	while (__cmpxchg_u32(m, old, val) != old);
+	return old;
+}
+
+#include <asm/cmpxchg-xchg.h>
+
+#endif /* __ASM_SH_CMPXCHG_CAS_H */

arch/sh/include/asm/cmpxchg-xchg.h

@@ -21,7 +21,7 @@ static inline u32 __xchg_cmpxchg(volatile void *ptr, u32 x, int size)
 	int off = (unsigned long)ptr % sizeof(u32);
 	volatile u32 *p = ptr - off;
 #ifdef __BIG_ENDIAN
-	int bitoff = (sizeof(u32) - 1 - off) * BITS_PER_BYTE;
+	int bitoff = (sizeof(u32) - size - off) * BITS_PER_BYTE;
 #else
 	int bitoff = off * BITS_PER_BYTE;
 #endif

arch/sh/include/asm/cmpxchg.h

@@ -13,6 +13,8 @@
 #include <asm/cmpxchg-grb.h>
 #elif defined(CONFIG_CPU_SH4A)
 #include <asm/cmpxchg-llsc.h>
+#elif defined(CONFIG_CPU_J2) && defined(CONFIG_SMP)
+#include <asm/cmpxchg-cas.h>
 #else
 #include <asm/cmpxchg-irq.h>
 #endif

arch/sh/include/asm/futex-cas.h

+#ifndef __ASM_SH_FUTEX_CAS_H
+#define __ASM_SH_FUTEX_CAS_H
+
+static inline int atomic_futex_op_cmpxchg_inatomic(u32 *uval,
+						   u32 __user *uaddr,
+						   u32 oldval, u32 newval)
+{
+	int err = 0;
+	__asm__ __volatile__(
+		"1:\n\t"
+		"cas.l	%2, %1, @r0\n"
+		"2:\n\t"
+#ifdef CONFIG_MMU
+		".section	.fixup,\"ax\"\n"
+		"3:\n\t"
+		"mov.l	4f, %0\n\t"
+		"jmp	@%0\n\t"
+		" mov	%3, %0\n\t"
+		".balign	4\n"
+		"4:	.long	2b\n\t"
+		".previous\n"
+		".section	__ex_table,\"a\"\n\t"
+		".long	1b, 3b\n\t"
+		".previous"
+#endif
+		:"+r" (err), "+r" (newval)
+		:"r" (oldval), "i" (-EFAULT), "z" (uaddr)
+		:"t", "memory");
+	if (err) return err;
+	*uval = newval;
+	return 0;
+}
+
+#endif /* __ASM_SH_FUTEX_CAS_H */

arch/sh/include/asm/futex-irq.h

 #ifndef __ASM_SH_FUTEX_IRQ_H
 #define __ASM_SH_FUTEX_IRQ_H
 
-
-static inline int atomic_futex_op_xchg_set(int oparg, u32 __user *uaddr,
-					   int *oldval)
-{
-	unsigned long flags;
-	int ret;
-
-	local_irq_save(flags);
-
-	ret = get_user(*oldval, uaddr);
-	if (!ret)
-		ret = put_user(oparg, uaddr);
-
-	local_irq_restore(flags);
-
-	return ret;
-}
-
-static inline int atomic_futex_op_xchg_add(int oparg, u32 __user *uaddr,
-					   int *oldval)
-{
-	unsigned long flags;
-	int ret;
-
-	local_irq_save(flags);
-
-	ret = get_user(*oldval, uaddr);
-	if (!ret)
-		ret = put_user(*oldval + oparg, uaddr);
-
-	local_irq_restore(flags);
-
-	return ret;
-}
-
-static inline int atomic_futex_op_xchg_or(int oparg, u32 __user *uaddr,
-					  int *oldval)
-{
-	unsigned long flags;
-	int ret;
-
-	local_irq_save(flags);
-
-	ret = get_user(*oldval, uaddr);
-	if (!ret)
-		ret = put_user(*oldval | oparg, uaddr);
-
-	local_irq_restore(flags);
-
-	return ret;
-}
-
-static inline int atomic_futex_op_xchg_and(int oparg, u32 __user *uaddr,
-					   int *oldval)
-{
-	unsigned long flags;
-	int ret;
-
-	local_irq_save(flags);
-
-	ret = get_user(*oldval, uaddr);
-	if (!ret)
-		ret = put_user(*oldval & oparg, uaddr);
-
-	local_irq_restore(flags);
-
-	return ret;
-}
-
-static inline int atomic_futex_op_xchg_xor(int oparg, u32 __user *uaddr,
-					   int *oldval)
-{
-	unsigned long flags;
-	int ret;
-
-	local_irq_save(flags);
-
-	ret = get_user(*oldval, uaddr);
-	if (!ret)
-		ret = put_user(*oldval ^ oparg, uaddr);
-
-	local_irq_restore(flags);
-
-	return ret;
-}
-
 static inline int atomic_futex_op_cmpxchg_inatomic(u32 *uval,
 						   u32 __user *uaddr,
 						   u32 oldval, u32 newval)

arch/sh/include/asm/futex-llsc.h

+#ifndef __ASM_SH_FUTEX_LLSC_H
+#define __ASM_SH_FUTEX_LLSC_H
+
+static inline int atomic_futex_op_cmpxchg_inatomic(u32 *uval,
+						   u32 __user *uaddr,
+						   u32 oldval, u32 newval)
+{
+	int err = 0;
+	__asm__ __volatile__(
+		"synco\n"
+		"1:\n\t"
+		"movli.l	@%2, r0\n\t"
+		"mov	r0, %1\n\t"
+		"cmp/eq	%1, %4\n\t"
+		"bf	2f\n\t"
+		"mov	%5, r0\n\t"
+		"movco.l	r0, @%2\n\t"
+		"bf	1b\n"
+		"2:\n\t"
+		"synco\n\t"
+#ifdef CONFIG_MMU
+		".section	.fixup,\"ax\"\n"
+		"3:\n\t"
+		"mov.l	4f, %0\n\t"
+		"jmp	@%0\n\t"
+		" mov	%3, %0\n\t"
+		".balign	4\n"
+		"4:	.long	2b\n\t"
+		".previous\n"
+		".section	__ex_table,\"a\"\n\t"
+		".long	1b, 3b\n\t"
+		".previous"
+#endif
+		:"+r" (err), "=&r" (*uval)
+		:"r" (uaddr), "i" (-EFAULT), "r" (oldval), "r" (newval)
+		:"t", "memory", "r0");
+	if (err) return err;
+	return 0;
+}
+
+#endif /* __ASM_SH_FUTEX_LLSC_H */

arch/sh/include/asm/futex.h

@@ -7,16 +7,34 @@
 #include <linux/uaccess.h>
 #include <asm/errno.h>
 
-/* XXX: UP variants, fix for SH-4A and SMP.. */
+#if !defined(CONFIG_SMP)
 #include <asm/futex-irq.h>
+#elif defined(CONFIG_CPU_J2)
+#include <asm/futex-cas.h>
+#elif defined(CONFIG_CPU_SH4A)
+#include <asm/futex-llsc.h>
+#else
+#error SMP not supported on this configuration.
+#endif
+
+static inline int
+futex_atomic_cmpxchg_inatomic(u32 *uval, u32 __user *uaddr,
+			      u32 oldval, u32 newval)
+{
+	if (!access_ok(VERIFY_WRITE, uaddr, sizeof(u32)))
+		return -EFAULT;
+
+	return atomic_futex_op_cmpxchg_inatomic(uval, uaddr, oldval, newval);
+}
 
 static inline int futex_atomic_op_inuser(int encoded_op, u32 __user *uaddr)
 {
 	int op = (encoded_op >> 28) & 7;
 	int cmp = (encoded_op >> 24) & 15;
-	int oparg = (encoded_op << 8) >> 20;
-	int cmparg = (encoded_op << 20) >> 20;
-	int oldval = 0, ret;
+	u32 oparg = (encoded_op << 8) >> 20;
+	u32 cmparg = (encoded_op << 20) >> 20;
+	u32 oldval, newval, prev;
+	int ret;
 
 	if (encoded_op & (FUTEX_OP_OPARG_SHIFT << 28))
 		oparg = 1 << oparg;
@@ -26,26 +44,39 @@ static inline int futex_atomic_op_inuser(int encoded_op, u32 __user *uaddr)
 
 	pagefault_disable();
 
-	switch (op) {
-	case FUTEX_OP_SET:
-		ret = atomic_futex_op_xchg_set(oparg, uaddr, &oldval);
-		break;
-	case FUTEX_OP_ADD:
-		ret = atomic_futex_op_xchg_add(oparg, uaddr, &oldval);
-		break;
-	case FUTEX_OP_OR:
-		ret = atomic_futex_op_xchg_or(oparg, uaddr, &oldval);
-		break;
-	case FUTEX_OP_ANDN:
-		ret = atomic_futex_op_xchg_and(~oparg, uaddr, &oldval);
-		break;
-	case FUTEX_OP_XOR:
-		ret = atomic_futex_op_xchg_xor(oparg, uaddr, &oldval);
-		break;
-	default:
-		ret = -ENOSYS;
-		break;
-	}
+	do {
+		if (op == FUTEX_OP_SET)
+			ret = oldval = 0;
+		else
+			ret = get_user(oldval, uaddr);
+
+		if (ret) break;
+
+		switch (op) {
+		case FUTEX_OP_SET:
+			newval = oparg;
+			break;
+		case FUTEX_OP_ADD:
+			newval = oldval + oparg;
+			break;
+		case FUTEX_OP_OR:
+			newval = oldval | oparg;
+			break;
+		case FUTEX_OP_ANDN:
+			newval = oldval & ~oparg;
+			break;
+		case FUTEX_OP_XOR:
+			newval = oldval ^ oparg;
+			break;
+		default:
+			ret = -ENOSYS;
+			break;
+		}
+
+		if (ret) break;
+
+		ret = futex_atomic_cmpxchg_inatomic(&prev, uaddr, oldval, newval);
+	} while (!ret && prev != oldval);
 
 	pagefault_enable();
 
@@ -53,10 +84,10 @@ static inline int futex_atomic_op_inuser(int encoded_op, u32 __user *uaddr)
 		switch (cmp) {
 		case FUTEX_OP_CMP_EQ: ret = (oldval == cmparg); break;
 		case FUTEX_OP_CMP_NE: ret = (oldval != cmparg); break;
-		case FUTEX_OP_CMP_LT: ret = (oldval < cmparg); break;
-		case FUTEX_OP_CMP_GE: ret = (oldval >= cmparg); break;
-		case FUTEX_OP_CMP_LE: ret = (oldval <= cmparg); break;
-		case FUTEX_OP_CMP_GT: ret = (oldval > cmparg); break;
+		case FUTEX_OP_CMP_LT: ret = ((int)oldval < (int)cmparg); break;
+		case FUTEX_OP_CMP_GE: ret = ((int)oldval >= (int)cmparg); break;
+		case FUTEX_OP_CMP_LE: ret = ((int)oldval <= (int)cmparg); break;
+		case FUTEX_OP_CMP_GT: ret = ((int)oldval > (int)cmparg); break;
 		default: ret = -ENOSYS;
 		}
 	}
@@ -64,15 +95,5 @@ static inline int futex_atomic_op_inuser(int encoded_op, u32 __user *uaddr)
 	return ret;
 }
 
-static inline int
-futex_atomic_cmpxchg_inatomic(u32 *uval, u32 __user *uaddr,
-			      u32 oldval, u32 newval)
-{
-	if (!access_ok(VERIFY_WRITE, uaddr, sizeof(u32)))
-		return -EFAULT;
-
-	return atomic_futex_op_cmpxchg_inatomic(uval, uaddr, oldval, newval);
-}
-
 #endif /* __KERNEL__ */
 #endif /* __ASM_SH_FUTEX_H */