Initial import

11 files changed, 3405 insertions, 0 deletions

diff --git a/arch/sh/kernel/cpu/sh4/Makefile b/arch/sh/kernel/cpu/sh4/Makefile
new file mode 100644
index 000000000..3a1dbc709
--- /dev/null
+++ b/arch/sh/kernel/cpu/sh4/Makefile
@@ -0,0 +1,36 @@
+#
+# Makefile for the Linux/SuperH SH-4 backends.
+#
+
+obj-y	:= probe.o common.o
+common-y	+= $(addprefix ../sh3/, entry.o ex.o)
+
+obj-$(CONFIG_HIBERNATION)		+= $(addprefix ../sh3/, swsusp.o)
+obj-$(CONFIG_SH_FPU)			+= fpu.o softfloat.o
+obj-$(CONFIG_SH_STORE_QUEUES)		+= sq.o
+
+# Perf events
+perf-$(CONFIG_CPU_SUBTYPE_SH7750)	:= perf_event.o
+perf-$(CONFIG_CPU_SUBTYPE_SH7750S)	:= perf_event.o
+perf-$(CONFIG_CPU_SUBTYPE_SH7091)	:= perf_event.o
+
+# CPU subtype setup
+obj-$(CONFIG_CPU_SUBTYPE_SH7750)	+= setup-sh7750.o
+obj-$(CONFIG_CPU_SUBTYPE_SH7750R)	+= setup-sh7750.o
+obj-$(CONFIG_CPU_SUBTYPE_SH7750S)	+= setup-sh7750.o
+obj-$(CONFIG_CPU_SUBTYPE_SH7091)	+= setup-sh7750.o
+obj-$(CONFIG_CPU_SUBTYPE_SH7751)	+= setup-sh7750.o
+obj-$(CONFIG_CPU_SUBTYPE_SH7751R)	+= setup-sh7750.o
+obj-$(CONFIG_CPU_SUBTYPE_SH7760)	+= setup-sh7760.o
+obj-$(CONFIG_CPU_SUBTYPE_SH4_202)	+= setup-sh4-202.o
+
+# Primary on-chip clocks (common)
+ifndef CONFIG_CPU_SH4A
+clock-$(CONFIG_CPU_SH4)			:= clock-sh4.o
+endif
+
+# Additional clocks by subtype
+clock-$(CONFIG_CPU_SUBTYPE_SH4_202)	+= clock-sh4-202.o
+
+obj-y					+= $(clock-y)
+obj-$(CONFIG_PERF_EVENTS)		+= $(perf-y)
diff --git a/arch/sh/kernel/cpu/sh4/clock-sh4-202.c b/arch/sh/kernel/cpu/sh4/clock-sh4-202.c
new file mode 100644
index 000000000..4b5bab5f8
--- /dev/null
+++ b/arch/sh/kernel/cpu/sh4/clock-sh4-202.c
@@ -0,0 +1,177 @@
+/*
+ * arch/sh/kernel/cpu/sh4/clock-sh4-202.c
+ *
+ * Additional SH4-202 support for the clock framework
+ *
+ *  Copyright (C) 2005  Paul Mundt
+ *
+ * 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.
+ */
+#include <linux/init.h>
+#include <linux/kernel.h>
+#include <linux/err.h>
+#include <linux/io.h>
+#include <linux/clkdev.h>
+#include <asm/clock.h>
+#include <asm/freq.h>
+
+#define CPG2_FRQCR3	0xfe0a0018
+
+static int frqcr3_divisors[] = { 1, 2, 3, 4, 6, 8, 16 };
+static int frqcr3_values[]   = { 0, 1, 2, 3, 4, 5, 6  };
+
+static unsigned long emi_clk_recalc(struct clk *clk)
+{
+	int idx = __raw_readl(CPG2_FRQCR3) & 0x0007;
+	return clk->parent->rate / frqcr3_divisors[idx];
+}
+
+static inline int frqcr3_lookup(struct clk *clk, unsigned long rate)
+{
+	int divisor = clk->parent->rate / rate;
+	int i;
+
+	for (i = 0; i < ARRAY_SIZE(frqcr3_divisors); i++)
+		if (frqcr3_divisors[i] == divisor)
+			return frqcr3_values[i];
+
+	/* Safe fallback */
+	return 5;
+}
+
+static struct sh_clk_ops sh4202_emi_clk_ops = {
+	.recalc		= emi_clk_recalc,
+};
+
+static struct clk sh4202_emi_clk = {
+	.flags		= CLK_ENABLE_ON_INIT,
+	.ops		= &sh4202_emi_clk_ops,
+};
+
+static unsigned long femi_clk_recalc(struct clk *clk)
+{
+	int idx = (__raw_readl(CPG2_FRQCR3) >> 3) & 0x0007;
+	return clk->parent->rate / frqcr3_divisors[idx];
+}
+
+static struct sh_clk_ops sh4202_femi_clk_ops = {
+	.recalc		= femi_clk_recalc,
+};
+
+static struct clk sh4202_femi_clk = {
+	.flags		= CLK_ENABLE_ON_INIT,
+	.ops		= &sh4202_femi_clk_ops,
+};
+
+static void shoc_clk_init(struct clk *clk)
+{
+	int i;
+
+	/*
+	 * For some reason, the shoc_clk seems to be set to some really
+	 * insane value at boot (values outside of the allowable frequency
+	 * range for instance). We deal with this by scaling it back down
+	 * to something sensible just in case.
+	 *
+	 * Start scaling from the high end down until we find something
+	 * that passes rate verification..
+	 */
+	for (i = 0; i < ARRAY_SIZE(frqcr3_divisors); i++) {
+		int divisor = frqcr3_divisors[i];
+
+		if (clk->ops->set_rate(clk, clk->parent->rate / divisor) == 0)
+			break;
+	}
+
+	WARN_ON(i == ARRAY_SIZE(frqcr3_divisors));	/* Undefined clock */
+}
+
+static unsigned long shoc_clk_recalc(struct clk *clk)
+{
+	int idx = (__raw_readl(CPG2_FRQCR3) >> 6) & 0x0007;
+	return clk->parent->rate / frqcr3_divisors[idx];
+}
+
+static int shoc_clk_verify_rate(struct clk *clk, unsigned long rate)
+{
+	struct clk *bclk = clk_get(NULL, "bus_clk");
+	unsigned long bclk_rate = clk_get_rate(bclk);
+
+	clk_put(bclk);
+
+	if (rate > bclk_rate)
+		return 1;
+	if (rate > 66000000)
+		return 1;
+
+	return 0;
+}
+
+static int shoc_clk_set_rate(struct clk *clk, unsigned long rate)
+{
+	unsigned long frqcr3;
+	unsigned int tmp;
+
+	/* Make sure we have something sensible to switch to */
+	if (shoc_clk_verify_rate(clk, rate) != 0)
+		return -EINVAL;
+
+	tmp = frqcr3_lookup(clk, rate);
+
+	frqcr3 = __raw_readl(CPG2_FRQCR3);
+	frqcr3 &= ~(0x0007 << 6);
+	frqcr3 |= tmp << 6;
+	__raw_writel(frqcr3, CPG2_FRQCR3);
+
+	clk->rate = clk->parent->rate / frqcr3_divisors[tmp];
+
+	return 0;
+}
+
+static struct sh_clk_ops sh4202_shoc_clk_ops = {
+	.init		= shoc_clk_init,
+	.recalc		= shoc_clk_recalc,
+	.set_rate	= shoc_clk_set_rate,
+};
+
+static struct clk sh4202_shoc_clk = {
+	.flags		= CLK_ENABLE_ON_INIT,
+	.ops		= &sh4202_shoc_clk_ops,
+};
+
+static struct clk *sh4202_onchip_clocks[] = {
+	&sh4202_emi_clk,
+	&sh4202_femi_clk,
+	&sh4202_shoc_clk,
+};
+
+static struct clk_lookup lookups[] = {
+	/* main clocks */
+	CLKDEV_CON_ID("emi_clk", &sh4202_emi_clk),
+	CLKDEV_CON_ID("femi_clk", &sh4202_femi_clk),
+	CLKDEV_CON_ID("shoc_clk", &sh4202_shoc_clk),
+};
+
+int __init arch_clk_init(void)
+{
+	struct clk *clk;
+	int i, ret = 0;
+
+	cpg_clk_init();
+
+	clk = clk_get(NULL, "master_clk");
+	for (i = 0; i < ARRAY_SIZE(sh4202_onchip_clocks); i++) {
+		struct clk *clkp = sh4202_onchip_clocks[i];
+
+		clkp->parent = clk;
+		ret |= clk_register(clkp);
+	}
+
+	clk_put(clk);
+
+	clkdev_add_table(lookups, ARRAY_SIZE(lookups));
+
+	return ret;
+}
diff --git a/arch/sh/kernel/cpu/sh4/clock-sh4.c b/arch/sh/kernel/cpu/sh4/clock-sh4.c
new file mode 100644
index 000000000..99e5ec8b4
--- /dev/null
+++ b/arch/sh/kernel/cpu/sh4/clock-sh4.c
@@ -0,0 +1,80 @@
+/*
+ * arch/sh/kernel/cpu/sh4/clock-sh4.c
+ *
+ * Generic SH-4 support for the clock framework
+ *
+ *  Copyright (C) 2005  Paul Mundt
+ *
+ * FRQCR parsing hacked out of arch/sh/kernel/time.c
+ *
+ *  Copyright (C) 1999  Tetsuya Okada & Niibe Yutaka
+ *  Copyright (C) 2000  Philipp Rumpf <prumpf@tux.org>
+ *  Copyright (C) 2002, 2003, 2004  Paul Mundt
+ *  Copyright (C) 2002  M. R. Brown  <mrbrown@linux-sh.org>
+ *
+ * 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.
+ */
+#include <linux/init.h>
+#include <linux/kernel.h>
+#include <asm/clock.h>
+#include <asm/freq.h>
+#include <asm/io.h>
+
+static int ifc_divisors[] = { 1, 2, 3, 4, 6, 8, 1, 1 };
+#define bfc_divisors ifc_divisors	/* Same */
+static int pfc_divisors[] = { 2, 3, 4, 6, 8, 2, 2, 2 };
+
+static void master_clk_init(struct clk *clk)
+{
+	clk->rate *= pfc_divisors[__raw_readw(FRQCR) & 0x0007];
+}
+
+static struct sh_clk_ops sh4_master_clk_ops = {
+	.init		= master_clk_init,
+};
+
+static unsigned long module_clk_recalc(struct clk *clk)
+{
+	int idx = (__raw_readw(FRQCR) & 0x0007);
+	return clk->parent->rate / pfc_divisors[idx];
+}
+
+static struct sh_clk_ops sh4_module_clk_ops = {
+	.recalc		= module_clk_recalc,
+};
+
+static unsigned long bus_clk_recalc(struct clk *clk)
+{
+	int idx = (__raw_readw(FRQCR) >> 3) & 0x0007;
+	return clk->parent->rate / bfc_divisors[idx];
+}
+
+static struct sh_clk_ops sh4_bus_clk_ops = {
+	.recalc		= bus_clk_recalc,
+};
+
+static unsigned long cpu_clk_recalc(struct clk *clk)
+{
+	int idx = (__raw_readw(FRQCR) >> 6) & 0x0007;
+	return clk->parent->rate / ifc_divisors[idx];
+}
+
+static struct sh_clk_ops sh4_cpu_clk_ops = {
+	.recalc		= cpu_clk_recalc,
+};
+
+static struct sh_clk_ops *sh4_clk_ops[] = {
+	&sh4_master_clk_ops,
+	&sh4_module_clk_ops,
+	&sh4_bus_clk_ops,
+	&sh4_cpu_clk_ops,
+};
+
+void __init arch_init_clk_ops(struct sh_clk_ops **ops, int idx)
+{
+	if (idx < ARRAY_SIZE(sh4_clk_ops))
+		*ops = sh4_clk_ops[idx];
+}
+
diff --git a/arch/sh/kernel/cpu/sh4/fpu.c b/arch/sh/kernel/cpu/sh4/fpu.c
new file mode 100644
index 000000000..69ab4d3c8
--- /dev/null
+++ b/arch/sh/kernel/cpu/sh4/fpu.c
@@ -0,0 +1,429 @@
+/*
+ * Save/restore floating point context for signal handlers.
+ *
+ * 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) 1999, 2000  Kaz Kojima & Niibe Yutaka
+ * Copyright (C) 2006  ST Microelectronics Ltd. (denorm support)
+ *
+ * FIXME! These routines have not been tested for big endian case.
+ */
+#include <linux/sched.h>
+#include <linux/signal.h>
+#include <linux/io.h>
+#include <cpu/fpu.h>
+#include <asm/processor.h>
+#include <asm/fpu.h>
+#include <asm/traps.h>
+
+/* The PR (precision) bit in the FP Status Register must be clear when
+ * an frchg instruction is executed, otherwise the instruction is undefined.
+ * Executing frchg with PR set causes a trap on some SH4 implementations.
+ */
+
+#define FPSCR_RCHG 0x00000000
+extern unsigned long long float64_div(unsigned long long a,
+				      unsigned long long b);
+extern unsigned long int float32_div(unsigned long int a, unsigned long int b);
+extern unsigned long long float64_mul(unsigned long long a,
+				      unsigned long long b);
+extern unsigned long int float32_mul(unsigned long int a, unsigned long int b);
+extern unsigned long long float64_add(unsigned long long a,
+				      unsigned long long b);
+extern unsigned long int float32_add(unsigned long int a, unsigned long int b);
+extern unsigned long long float64_sub(unsigned long long a,
+				      unsigned long long b);
+extern unsigned long int float32_sub(unsigned long int a, unsigned long int b);
+extern unsigned long int float64_to_float32(unsigned long long a);
+static unsigned int fpu_exception_flags;
+
+/*
+ * Save FPU registers onto task structure.
+ */
+void save_fpu(struct task_struct *tsk)
+{
+	unsigned long dummy;
+
+	enable_fpu();
+	asm volatile ("sts.l	fpul, @-%0\n\t"
+		      "sts.l	fpscr, @-%0\n\t"
+		      "lds	%2, fpscr\n\t"
+		      "frchg\n\t"
+		      "fmov.s	fr15, @-%0\n\t"
+		      "fmov.s	fr14, @-%0\n\t"
+		      "fmov.s	fr13, @-%0\n\t"
+		      "fmov.s	fr12, @-%0\n\t"
+		      "fmov.s	fr11, @-%0\n\t"
+		      "fmov.s	fr10, @-%0\n\t"
+		      "fmov.s	fr9, @-%0\n\t"
+		      "fmov.s	fr8, @-%0\n\t"
+		      "fmov.s	fr7, @-%0\n\t"
+		      "fmov.s	fr6, @-%0\n\t"
+		      "fmov.s	fr5, @-%0\n\t"
+		      "fmov.s	fr4, @-%0\n\t"
+		      "fmov.s	fr3, @-%0\n\t"
+		      "fmov.s	fr2, @-%0\n\t"
+		      "fmov.s	fr1, @-%0\n\t"
+		      "fmov.s	fr0, @-%0\n\t"
+		      "frchg\n\t"
+		      "fmov.s	fr15, @-%0\n\t"
+		      "fmov.s	fr14, @-%0\n\t"
+		      "fmov.s	fr13, @-%0\n\t"
+		      "fmov.s	fr12, @-%0\n\t"
+		      "fmov.s	fr11, @-%0\n\t"
+		      "fmov.s	fr10, @-%0\n\t"
+		      "fmov.s	fr9, @-%0\n\t"
+		      "fmov.s	fr8, @-%0\n\t"
+		      "fmov.s	fr7, @-%0\n\t"
+		      "fmov.s	fr6, @-%0\n\t"
+		      "fmov.s	fr5, @-%0\n\t"
+		      "fmov.s	fr4, @-%0\n\t"
+		      "fmov.s	fr3, @-%0\n\t"
+		      "fmov.s	fr2, @-%0\n\t"
+		      "fmov.s	fr1, @-%0\n\t"
+		      "fmov.s	fr0, @-%0\n\t"
+		      "lds	%3, fpscr\n\t":"=r" (dummy)
+		      :"0"((char *)(&tsk->thread.xstate->hardfpu.status)),
+		      "r"(FPSCR_RCHG), "r"(FPSCR_INIT)
+		      :"memory");
+
+	disable_fpu();
+}
+
+void restore_fpu(struct task_struct *tsk)
+{
+	unsigned long dummy;
+
+	enable_fpu();
+	asm volatile ("lds	%2, fpscr\n\t"
+		      "fmov.s	@%0+, fr0\n\t"
+		      "fmov.s	@%0+, fr1\n\t"
+		      "fmov.s	@%0+, fr2\n\t"
+		      "fmov.s	@%0+, fr3\n\t"
+		      "fmov.s	@%0+, fr4\n\t"
+		      "fmov.s	@%0+, fr5\n\t"
+		      "fmov.s	@%0+, fr6\n\t"
+		      "fmov.s	@%0+, fr7\n\t"
+		      "fmov.s	@%0+, fr8\n\t"
+		      "fmov.s	@%0+, fr9\n\t"
+		      "fmov.s	@%0+, fr10\n\t"
+		      "fmov.s	@%0+, fr11\n\t"
+		      "fmov.s	@%0+, fr12\n\t"
+		      "fmov.s	@%0+, fr13\n\t"
+		      "fmov.s	@%0+, fr14\n\t"
+		      "fmov.s	@%0+, fr15\n\t"
+		      "frchg\n\t"
+		      "fmov.s	@%0+, fr0\n\t"
+		      "fmov.s	@%0+, fr1\n\t"
+		      "fmov.s	@%0+, fr2\n\t"
+		      "fmov.s	@%0+, fr3\n\t"
+		      "fmov.s	@%0+, fr4\n\t"
+		      "fmov.s	@%0+, fr5\n\t"
+		      "fmov.s	@%0+, fr6\n\t"
+		      "fmov.s	@%0+, fr7\n\t"
+		      "fmov.s	@%0+, fr8\n\t"
+		      "fmov.s	@%0+, fr9\n\t"
+		      "fmov.s	@%0+, fr10\n\t"
+		      "fmov.s	@%0+, fr11\n\t"
+		      "fmov.s	@%0+, fr12\n\t"
+		      "fmov.s	@%0+, fr13\n\t"
+		      "fmov.s	@%0+, fr14\n\t"
+		      "fmov.s	@%0+, fr15\n\t"
+		      "frchg\n\t"
+		      "lds.l	@%0+, fpscr\n\t"
+		      "lds.l	@%0+, fpul\n\t"
+		      :"=r" (dummy)
+		      :"0" (tsk->thread.xstate), "r" (FPSCR_RCHG)
+		      :"memory");
+	disable_fpu();
+}
+
+/**
+ *      denormal_to_double - Given denormalized float number,
+ *                           store double float
+ *
+ *      @fpu: Pointer to sh_fpu_hard structure
+ *      @n: Index to FP register
+ */
+static void denormal_to_double(struct sh_fpu_hard_struct *fpu, int n)
+{
+	unsigned long du, dl;
+	unsigned long x = fpu->fpul;
+	int exp = 1023 - 126;
+
+	if (x != 0 && (x & 0x7f800000) == 0) {
+		du = (x & 0x80000000);
+		while ((x & 0x00800000) == 0) {
+			x <<= 1;
+			exp--;
+		}
+		x &= 0x007fffff;
+		du |= (exp << 20) | (x >> 3);
+		dl = x << 29;
+
+		fpu->fp_regs[n] = du;
+		fpu->fp_regs[n + 1] = dl;
+	}
+}
+
+/**
+ *	ieee_fpe_handler - Handle denormalized number exception
+ *
+ *	@regs: Pointer to register structure
+ *
+ *	Returns 1 when it's handled (should not cause exception).
+ */
+static int ieee_fpe_handler(struct pt_regs *regs)
+{
+	unsigned short insn = *(unsigned short *)regs->pc;
+	unsigned short finsn;
+	unsigned long nextpc;
+	int nib[4] = {
+		(insn >> 12) & 0xf,
+		(insn >> 8) & 0xf,
+		(insn >> 4) & 0xf,
+		insn & 0xf
+	};
+
+	if (nib[0] == 0xb || (nib[0] == 0x4 && nib[2] == 0x0 && nib[3] == 0xb))
+		regs->pr = regs->pc + 4;  /* bsr & jsr */
+
+	if (nib[0] == 0xa || nib[0] == 0xb) {
+		/* bra & bsr */
+		nextpc = regs->pc + 4 + ((short)((insn & 0xfff) << 4) >> 3);
+		finsn = *(unsigned short *)(regs->pc + 2);
+	} else if (nib[0] == 0x8 && nib[1] == 0xd) {
+		/* bt/s */
+		if (regs->sr & 1)
+			nextpc = regs->pc + 4 + ((char)(insn & 0xff) << 1);
+		else
+			nextpc = regs->pc + 4;
+		finsn = *(unsigned short *)(regs->pc + 2);
+	} else if (nib[0] == 0x8 && nib[1] == 0xf) {
+		/* bf/s */
+		if (regs->sr & 1)
+			nextpc = regs->pc + 4;
+		else
+			nextpc = regs->pc + 4 + ((char)(insn & 0xff) << 1);
+		finsn = *(unsigned short *)(regs->pc + 2);
+	} else if (nib[0] == 0x4 && nib[3] == 0xb &&
+		   (nib[2] == 0x0 || nib[2] == 0x2)) {
+		/* jmp & jsr */
+		nextpc = regs->regs[nib[1]];
+		finsn = *(unsigned short *)(regs->pc + 2);
+	} else if (nib[0] == 0x0 && nib[3] == 0x3 &&
+		   (nib[2] == 0x0 || nib[2] == 0x2)) {
+		/* braf & bsrf */
+		nextpc = regs->pc + 4 + regs->regs[nib[1]];
+		finsn = *(unsigned short *)(regs->pc + 2);
+	} else if (insn == 0x000b) {
+		/* rts */
+		nextpc = regs->pr;
+		finsn = *(unsigned short *)(regs->pc + 2);
+	} else {
+		nextpc = regs->pc + instruction_size(insn);
+		finsn = insn;
+	}
+
+	if ((finsn & 0xf1ff) == 0xf0ad) {
+		/* fcnvsd */
+		struct task_struct *tsk = current;
+
+		if ((tsk->thread.xstate->hardfpu.fpscr & FPSCR_CAUSE_ERROR))
+			/* FPU error */
+			denormal_to_double(&tsk->thread.xstate->hardfpu,
+					   (finsn >> 8) & 0xf);
+		else
+			return 0;
+
+		regs->pc = nextpc;
+		return 1;
+	} else if ((finsn & 0xf00f) == 0xf002) {
+		/* fmul */
+		struct task_struct *tsk = current;
+		int fpscr;
+		int n, m, prec;
+		unsigned int hx, hy;
+
+		n = (finsn >> 8) & 0xf;
+		m = (finsn >> 4) & 0xf;
+		hx = tsk->thread.xstate->hardfpu.fp_regs[n];
+		hy = tsk->thread.xstate->hardfpu.fp_regs[m];
+		fpscr = tsk->thread.xstate->hardfpu.fpscr;
+		prec = fpscr & FPSCR_DBL_PRECISION;
+
+		if ((fpscr & FPSCR_CAUSE_ERROR)
+		    && (prec && ((hx & 0x7fffffff) < 0x00100000
+				 || (hy & 0x7fffffff) < 0x00100000))) {
+			long long llx, lly;
+
+			/* FPU error because of denormal (doubles) */
+			llx = ((long long)hx << 32)
+			    | tsk->thread.xstate->hardfpu.fp_regs[n + 1];
+			lly = ((long long)hy << 32)
+			    | tsk->thread.xstate->hardfpu.fp_regs[m + 1];
+			llx = float64_mul(llx, lly);
+			tsk->thread.xstate->hardfpu.fp_regs[n] = llx >> 32;
+			tsk->thread.xstate->hardfpu.fp_regs[n + 1] = llx & 0xffffffff;
+		} else if ((fpscr & FPSCR_CAUSE_ERROR)
+			   && (!prec && ((hx & 0x7fffffff) < 0x00800000
+					 || (hy & 0x7fffffff) < 0x00800000))) {
+			/* FPU error because of denormal (floats) */
+			hx = float32_mul(hx, hy);
+			tsk->thread.xstate->hardfpu.fp_regs[n] = hx;
+		} else
+			return 0;
+
+		regs->pc = nextpc;
+		return 1;
+	} else if ((finsn & 0xf00e) == 0xf000) {
+		/* fadd, fsub */
+		struct task_struct *tsk = current;
+		int fpscr;
+		int n, m, prec;
+		unsigned int hx, hy;
+
+		n = (finsn >> 8) & 0xf;
+		m = (finsn >> 4) & 0xf;
+		hx = tsk->thread.xstate->hardfpu.fp_regs[n];
+		hy = tsk->thread.xstate->hardfpu.fp_regs[m];
+		fpscr = tsk->thread.xstate->hardfpu.fpscr;
+		prec = fpscr & FPSCR_DBL_PRECISION;
+
+		if ((fpscr & FPSCR_CAUSE_ERROR)
+		    && (prec && ((hx & 0x7fffffff) < 0x00100000
+				 || (hy & 0x7fffffff) < 0x00100000))) {
+			long long llx, lly;
+
+			/* FPU error because of denormal (doubles) */
+			llx = ((long long)hx << 32)
+			    | tsk->thread.xstate->hardfpu.fp_regs[n + 1];
+			lly = ((long long)hy << 32)
+			    | tsk->thread.xstate->hardfpu.fp_regs[m + 1];
+			if ((finsn & 0xf00f) == 0xf000)
+				llx = float64_add(llx, lly);
+			else
+				llx = float64_sub(llx, lly);
+			tsk->thread.xstate->hardfpu.fp_regs[n] = llx >> 32;
+			tsk->thread.xstate->hardfpu.fp_regs[n + 1] = llx & 0xffffffff;
+		} else if ((fpscr & FPSCR_CAUSE_ERROR)
+			   && (!prec && ((hx & 0x7fffffff) < 0x00800000
+					 || (hy & 0x7fffffff) < 0x00800000))) {
+			/* FPU error because of denormal (floats) */
+			if ((finsn & 0xf00f) == 0xf000)
+				hx = float32_add(hx, hy);
+			else
+				hx = float32_sub(hx, hy);
+			tsk->thread.xstate->hardfpu.fp_regs[n] = hx;
+		} else
+			return 0;
+
+		regs->pc = nextpc;
+		return 1;
+	} else if ((finsn & 0xf003) == 0xf003) {
+		/* fdiv */
+		struct task_struct *tsk = current;
+		int fpscr;
+		int n, m, prec;
+		unsigned int hx, hy;
+
+		n = (finsn >> 8) & 0xf;
+		m = (finsn >> 4) & 0xf;
+		hx = tsk->thread.xstate->hardfpu.fp_regs[n];
+		hy = tsk->thread.xstate->hardfpu.fp_regs[m];
+		fpscr = tsk->thread.xstate->hardfpu.fpscr;
+		prec = fpscr & FPSCR_DBL_PRECISION;
+
+		if ((fpscr & FPSCR_CAUSE_ERROR)
+		    && (prec && ((hx & 0x7fffffff) < 0x00100000
+				 || (hy & 0x7fffffff) < 0x00100000))) {
+			long long llx, lly;
+
+			/* FPU error because of denormal (doubles) */
+			llx = ((long long)hx << 32)
+			    | tsk->thread.xstate->hardfpu.fp_regs[n + 1];
+			lly = ((long long)hy << 32)
+			    | tsk->thread.xstate->hardfpu.fp_regs[m + 1];
+
+			llx = float64_div(llx, lly);
+
+			tsk->thread.xstate->hardfpu.fp_regs[n] = llx >> 32;
+			tsk->thread.xstate->hardfpu.fp_regs[n + 1] = llx & 0xffffffff;
+		} else if ((fpscr & FPSCR_CAUSE_ERROR)
+			   && (!prec && ((hx & 0x7fffffff) < 0x00800000
+					 || (hy & 0x7fffffff) < 0x00800000))) {
+			/* FPU error because of denormal (floats) */
+			hx = float32_div(hx, hy);
+			tsk->thread.xstate->hardfpu.fp_regs[n] = hx;
+		} else
+			return 0;
+
+		regs->pc = nextpc;
+		return 1;
+	} else if ((finsn & 0xf0bd) == 0xf0bd) {
+		/* fcnvds - double to single precision convert */
+		struct task_struct *tsk = current;
+		int m;
+		unsigned int hx;
+
+		m = (finsn >> 8) & 0x7;
+		hx = tsk->thread.xstate->hardfpu.fp_regs[m];
+
+		if ((tsk->thread.xstate->hardfpu.fpscr & FPSCR_CAUSE_ERROR)
+			&& ((hx & 0x7fffffff) < 0x00100000)) {
+			/* subnormal double to float conversion */
+			long long llx;
+
+			llx = ((long long)tsk->thread.xstate->hardfpu.fp_regs[m] << 32)
+			    | tsk->thread.xstate->hardfpu.fp_regs[m + 1];
+
+			tsk->thread.xstate->hardfpu.fpul = float64_to_float32(llx);
+		} else
+			return 0;
+
+		regs->pc = nextpc;
+		return 1;
+	}
+
+	return 0;
+}
+
+void float_raise(unsigned int flags)
+{
+	fpu_exception_flags |= flags;
+}
+
+int float_rounding_mode(void)
+{
+	struct task_struct *tsk = current;
+	int roundingMode = FPSCR_ROUNDING_MODE(tsk->thread.xstate->hardfpu.fpscr);
+	return roundingMode;
+}
+
+BUILD_TRAP_HANDLER(fpu_error)
+{
+	struct task_struct *tsk = current;
+	TRAP_HANDLER_DECL;
+
+	__unlazy_fpu(tsk, regs);
+	fpu_exception_flags = 0;
+	if (ieee_fpe_handler(regs)) {
+		tsk->thread.xstate->hardfpu.fpscr &=
+		    ~(FPSCR_CAUSE_MASK | FPSCR_FLAG_MASK);
+		tsk->thread.xstate->hardfpu.fpscr |= fpu_exception_flags;
+		/* Set the FPSCR flag as well as cause bits - simply
+		 * replicate the cause */
+		tsk->thread.xstate->hardfpu.fpscr |= (fpu_exception_flags >> 10);
+		grab_fpu(regs);
+		restore_fpu(tsk);
+		task_thread_info(tsk)->status |= TS_USEDFPU;
+		if ((((tsk->thread.xstate->hardfpu.fpscr & FPSCR_ENABLE_MASK) >> 7) &
+		     (fpu_exception_flags >> 2)) == 0) {
+			return;
+		}
+	}
+
+	force_sig(SIGFPE, tsk);
+}
diff --git a/arch/sh/kernel/cpu/sh4/perf_event.c b/arch/sh/kernel/cpu/sh4/perf_event.c
new file mode 100644
index 000000000..fa4f724b2
--- /dev/null
+++ b/arch/sh/kernel/cpu/sh4/perf_event.c
@@ -0,0 +1,268 @@
+/*
+ * Performance events support for SH7750-style performance counters
+ *
+ *  Copyright (C) 2009  Paul Mundt
+ *
+ * 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.
+ */
+#include <linux/kernel.h>
+#include <linux/init.h>
+#include <linux/io.h>
+#include <linux/irq.h>
+#include <linux/perf_event.h>
+#include <asm/processor.h>
+
+#define PM_CR_BASE	0xff000084	/* 16-bit */
+#define PM_CTR_BASE	0xff100004	/* 32-bit */
+
+#define PMCR(n)		(PM_CR_BASE + ((n) * 0x04))
+#define PMCTRH(n)	(PM_CTR_BASE + 0x00 + ((n) * 0x08))
+#define PMCTRL(n)	(PM_CTR_BASE + 0x04 + ((n) * 0x08))
+
+#define PMCR_PMM_MASK	0x0000003f
+
+#define PMCR_CLKF	0x00000100
+#define PMCR_PMCLR	0x00002000
+#define PMCR_PMST	0x00004000
+#define PMCR_PMEN	0x00008000
+
+static struct sh_pmu sh7750_pmu;
+
+/*
+ * There are a number of events supported by each counter (33 in total).
+ * Since we have 2 counters, each counter will take the event code as it
+ * corresponds to the PMCR PMM setting. Each counter can be configured
+ * independently.
+ *
+ *	Event Code	Description
+ *	----------	-----------
+ *
+ *	0x01		Operand read access
+ *	0x02		Operand write access
+ *	0x03		UTLB miss
+ *	0x04		Operand cache read miss
+ *	0x05		Operand cache write miss
+ *	0x06		Instruction fetch (w/ cache)
+ *	0x07		Instruction TLB miss
+ *	0x08		Instruction cache miss
+ *	0x09		All operand accesses
+ *	0x0a		All instruction accesses
+ *	0x0b		OC RAM operand access
+ *	0x0d		On-chip I/O space access
+ *	0x0e		Operand access (r/w)
+ *	0x0f		Operand cache miss (r/w)
+ *	0x10		Branch instruction
+ *	0x11		Branch taken
+ *	0x12		BSR/BSRF/JSR
+ *	0x13		Instruction execution
+ *	0x14		Instruction execution in parallel
+ *	0x15		FPU Instruction execution
+ *	0x16		Interrupt
+ *	0x17		NMI
+ *	0x18		trapa instruction execution
+ *	0x19		UBCA match
+ *	0x1a		UBCB match
+ *	0x21		Instruction cache fill
+ *	0x22		Operand cache fill
+ *	0x23		Elapsed time
+ *	0x24		Pipeline freeze by I-cache miss
+ *	0x25		Pipeline freeze by D-cache miss
+ *	0x27		Pipeline freeze by branch instruction
+ *	0x28		Pipeline freeze by CPU register
+ *	0x29		Pipeline freeze by FPU
+ */
+
+static const int sh7750_general_events[] = {
+	[PERF_COUNT_HW_CPU_CYCLES]		= 0x0023,
+	[PERF_COUNT_HW_INSTRUCTIONS]		= 0x000a,
+	[PERF_COUNT_HW_CACHE_REFERENCES]	= 0x0006,	/* I-cache */
+	[PERF_COUNT_HW_CACHE_MISSES]		= 0x0008,	/* I-cache */
+	[PERF_COUNT_HW_BRANCH_INSTRUCTIONS]	= 0x0010,
+	[PERF_COUNT_HW_BRANCH_MISSES]		= -1,
+	[PERF_COUNT_HW_BUS_CYCLES]		= -1,
+};
+
+#define C(x)	PERF_COUNT_HW_CACHE_##x
+
+static const int sh7750_cache_events
+			[PERF_COUNT_HW_CACHE_MAX]
+			[PERF_COUNT_HW_CACHE_OP_MAX]
+			[PERF_COUNT_HW_CACHE_RESULT_MAX] =
+{
+	[ C(L1D) ] = {
+		[ C(OP_READ) ] = {
+			[ C(RESULT_ACCESS) ] = 0x0001,
+			[ C(RESULT_MISS)   ] = 0x0004,
+		},
+		[ C(OP_WRITE) ] = {
+			[ C(RESULT_ACCESS) ] = 0x0002,
+			[ C(RESULT_MISS)   ] = 0x0005,
+		},
+		[ C(OP_PREFETCH) ] = {
+			[ C(RESULT_ACCESS) ] = 0,
+			[ C(RESULT_MISS)   ] = 0,
+		},
+	},
+
+	[ C(L1I) ] = {
+		[ C(OP_READ) ] = {
+			[ C(RESULT_ACCESS) ] = 0x0006,
+			[ C(RESULT_MISS)   ] = 0x0008,
+		},
+		[ C(OP_WRITE) ] = {
+			[ C(RESULT_ACCESS) ] = -1,
+			[ C(RESULT_MISS)   ] = -1,
+		},
+		[ C(OP_PREFETCH) ] = {
+			[ C(RESULT_ACCESS) ] = 0,
+			[ C(RESULT_MISS)   ] = 0,
+		},
+	},
+
+	[ C(LL) ] = {
+		[ C(OP_READ) ] = {
+			[ C(RESULT_ACCESS) ] = 0,
+			[ C(RESULT_MISS)   ] = 0,
+		},
+		[ C(OP_WRITE) ] = {
+			[ C(RESULT_ACCESS) ] = 0,
+			[ C(RESULT_MISS)   ] = 0,
+		},
+		[ C(OP_PREFETCH) ] = {
+			[ C(RESULT_ACCESS) ] = 0,
+			[ C(RESULT_MISS)   ] = 0,
+		},
+	},
+
+	[ C(DTLB) ] = {
+		[ C(OP_READ) ] = {
+			[ C(RESULT_ACCESS) ] = 0,
+			[ C(RESULT_MISS)   ] = 0x0003,
+		},
+		[ C(OP_WRITE) ] = {
+			[ C(RESULT_ACCESS) ] = 0,
+			[ C(RESULT_MISS)   ] = 0,
+		},
+		[ C(OP_PREFETCH) ] = {
+			[ C(RESULT_ACCESS) ] = 0,
+			[ C(RESULT_MISS)   ] = 0,
+		},
+	},
+
+	[ C(ITLB) ] = {
+		[ C(OP_READ) ] = {
+			[ C(RESULT_ACCESS) ] = 0,
+			[ C(RESULT_MISS)   ] = 0x0007,
+		},
+		[ C(OP_WRITE) ] = {
+			[ C(RESULT_ACCESS) ] = -1,
+			[ C(RESULT_MISS)   ] = -1,
+		},
+		[ C(OP_PREFETCH) ] = {
+			[ C(RESULT_ACCESS) ] = -1,
+			[ C(RESULT_MISS)   ] = -1,
+		},
+	},
+
+	[ C(BPU) ] = {
+		[ C(OP_READ) ] = {
+			[ C(RESULT_ACCESS) ] = -1,
+			[ C(RESULT_MISS)   ] = -1,
+		},
+		[ C(OP_WRITE) ] = {
+			[ C(RESULT_ACCESS) ] = -1,
+			[ C(RESULT_MISS)   ] = -1,
+		},
+		[ C(OP_PREFETCH) ] = {
+			[ C(RESULT_ACCESS) ] = -1,
+			[ C(RESULT_MISS)   ] = -1,
+		},
+	},
+
+	[ C(NODE) ] = {
+		[ C(OP_READ) ] = {
+			[ C(RESULT_ACCESS) ] = -1,
+			[ C(RESULT_MISS)   ] = -1,
+		},
+		[ C(OP_WRITE) ] = {
+			[ C(RESULT_ACCESS) ] = -1,
+			[ C(RESULT_MISS)   ] = -1,
+		},
+		[ C(OP_PREFETCH) ] = {
+			[ C(RESULT_ACCESS) ] = -1,
+			[ C(RESULT_MISS)   ] = -1,
+		},
+	},
+};
+
+static int sh7750_event_map(int event)
+{
+	return sh7750_general_events[event];
+}
+
+static u64 sh7750_pmu_read(int idx)
+{
+	return (u64)((u64)(__raw_readl(PMCTRH(idx)) & 0xffff) << 32) |
+			   __raw_readl(PMCTRL(idx));
+}
+
+static void sh7750_pmu_disable(struct hw_perf_event *hwc, int idx)
+{
+	unsigned int tmp;
+
+	tmp = __raw_readw(PMCR(idx));
+	tmp &= ~(PMCR_PMM_MASK | PMCR_PMEN);
+	__raw_writew(tmp, PMCR(idx));
+}
+
+static void sh7750_pmu_enable(struct hw_perf_event *hwc, int idx)
+{
+	__raw_writew(__raw_readw(PMCR(idx)) | PMCR_PMCLR, PMCR(idx));
+	__raw_writew(hwc->config | PMCR_PMEN | PMCR_PMST, PMCR(idx));
+}
+
+static void sh7750_pmu_disable_all(void)
+{
+	int i;
+
+	for (i = 0; i < sh7750_pmu.num_events; i++)
+		__raw_writew(__raw_readw(PMCR(i)) & ~PMCR_PMEN, PMCR(i));
+}
+
+static void sh7750_pmu_enable_all(void)
+{
+	int i;
+
+	for (i = 0; i < sh7750_pmu.num_events; i++)
+		__raw_writew(__raw_readw(PMCR(i)) | PMCR_PMEN, PMCR(i));
+}
+
+static struct sh_pmu sh7750_pmu = {
+	.name		= "sh7750",
+	.num_events	= 2,
+	.event_map	= sh7750_event_map,
+	.max_events	= ARRAY_SIZE(sh7750_general_events),
+	.raw_event_mask	= PMCR_PMM_MASK,
+	.cache_events	= &sh7750_cache_events,
+	.read		= sh7750_pmu_read,
+	.disable	= sh7750_pmu_disable,
+	.enable		= sh7750_pmu_enable,
+	.disable_all	= sh7750_pmu_disable_all,
+	.enable_all	= sh7750_pmu_enable_all,
+};
+
+static int __init sh7750_pmu_init(void)
+{
+	/*
+	 * Make sure this CPU actually has perf counters.
+	 */
+	if (!(boot_cpu_data.flags & CPU_HAS_PERF_COUNTER)) {
+		pr_notice("HW perf events unsupported, software events only.\n");
+		return -ENODEV;
+	}
+
+	return register_sh_pmu(&sh7750_pmu);
+}
+early_initcall(sh7750_pmu_init);
diff --git a/arch/sh/kernel/cpu/sh4/probe.c b/arch/sh/kernel/cpu/sh4/probe.c
new file mode 100644
index 000000000..a521bcf50
--- /dev/null
+++ b/arch/sh/kernel/cpu/sh4/probe.c
@@ -0,0 +1,263 @@
+/*
+ * arch/sh/kernel/cpu/sh4/probe.c
+ *
+ * CPU Subtype Probing for SH-4.
+ *
+ * Copyright (C) 2001 - 2007  Paul Mundt
+ * Copyright (C) 2003  Richard Curnow
+ *
+ * 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.
+ */
+#include <linux/init.h>
+#include <linux/io.h>
+#include <asm/processor.h>
+#include <asm/cache.h>
+
+void cpu_probe(void)
+{
+	unsigned long pvr, prr, cvr;
+	unsigned long size;
+
+	static unsigned long sizes[16] = {
+		[1] = (1 << 12),
+		[2] = (1 << 13),
+		[4] = (1 << 14),
+		[8] = (1 << 15),
+		[9] = (1 << 16)
+	};
+
+	pvr = (__raw_readl(CCN_PVR) >> 8) & 0xffffff;
+	prr = (__raw_readl(CCN_PRR) >> 4) & 0xff;
+	cvr = (__raw_readl(CCN_CVR));
+
+	/*
+	 * Setup some sane SH-4 defaults for the icache
+	 */
+	boot_cpu_data.icache.way_incr		= (1 << 13);
+	boot_cpu_data.icache.entry_shift	= 5;
+	boot_cpu_data.icache.sets		= 256;
+	boot_cpu_data.icache.ways		= 1;
+	boot_cpu_data.icache.linesz		= L1_CACHE_BYTES;
+
+	/*
+	 * And again for the dcache ..
+	 */
+	boot_cpu_data.dcache.way_incr		= (1 << 14);
+	boot_cpu_data.dcache.entry_shift	= 5;
+	boot_cpu_data.dcache.sets		= 512;
+	boot_cpu_data.dcache.ways		= 1;
+	boot_cpu_data.dcache.linesz		= L1_CACHE_BYTES;
+
+	/* We don't know the chip cut */
+	boot_cpu_data.cut_major = boot_cpu_data.cut_minor = -1;
+
+	/*
+	 * Setup some generic flags we can probe on SH-4A parts
+	 */
+	if (((pvr >> 16) & 0xff) == 0x10) {
+		boot_cpu_data.family = CPU_FAMILY_SH4A;
+
+		if ((cvr & 0x10000000) == 0) {
+			boot_cpu_data.flags |= CPU_HAS_DSP;
+			boot_cpu_data.family = CPU_FAMILY_SH4AL_DSP;
+		}
+
+		boot_cpu_data.flags |= CPU_HAS_LLSC | CPU_HAS_PERF_COUNTER;
+		boot_cpu_data.cut_major = pvr & 0x7f;
+
+		boot_cpu_data.icache.ways = 4;
+		boot_cpu_data.dcache.ways = 4;
+	} else {
+		/* And some SH-4 defaults.. */
+		boot_cpu_data.flags |= CPU_HAS_PTEA | CPU_HAS_FPU;
+		boot_cpu_data.family = CPU_FAMILY_SH4;
+	}
+
+	/* FPU detection works for almost everyone */
+	if ((cvr & 0x20000000))
+		boot_cpu_data.flags |= CPU_HAS_FPU;
+
+	/* Mask off the upper chip ID */
+	pvr &= 0xffff;
+
+	/*
+	 * Probe the underlying processor version/revision and
+	 * adjust cpu_data setup accordingly.
+	 */
+	switch (pvr) {
+	case 0x205:
+		boot_cpu_data.type = CPU_SH7750;
+		boot_cpu_data.flags |= CPU_HAS_P2_FLUSH_BUG |
+				       CPU_HAS_PERF_COUNTER;
+		break;
+	case 0x206:
+		boot_cpu_data.type = CPU_SH7750S;
+		boot_cpu_data.flags |= CPU_HAS_P2_FLUSH_BUG |
+				       CPU_HAS_PERF_COUNTER;
+		break;
+	case 0x1100:
+		boot_cpu_data.type = CPU_SH7751;
+		break;
+	case 0x2001:
+	case 0x2004:
+		boot_cpu_data.type = CPU_SH7770;
+		break;
+	case 0x2006:
+	case 0x200A:
+		if (prr == 0x61)
+			boot_cpu_data.type = CPU_SH7781;
+		else if (prr == 0xa1)
+			boot_cpu_data.type = CPU_SH7763;
+		else
+			boot_cpu_data.type = CPU_SH7780;
+
+		break;
+	case 0x3000:
+	case 0x3003:
+	case 0x3009:
+		boot_cpu_data.type = CPU_SH7343;
+		break;
+	case 0x3004:
+	case 0x3007:
+		boot_cpu_data.type = CPU_SH7785;
+		break;
+	case 0x4004:
+	case 0x4005:
+		boot_cpu_data.type = CPU_SH7786;
+		boot_cpu_data.flags |= CPU_HAS_PTEAEX | CPU_HAS_L2_CACHE;
+		break;
+	case 0x3008:
+		switch (prr) {
+		case 0x50:
+		case 0x51:
+			boot_cpu_data.type = CPU_SH7723;
+			boot_cpu_data.flags |= CPU_HAS_L2_CACHE;
+			break;
+		case 0x70:
+			boot_cpu_data.type = CPU_SH7366;
+			break;
+		case 0xa0:
+		case 0xa1:
+			boot_cpu_data.type = CPU_SH7722;
+			break;
+		}
+		break;
+	case 0x300b:
+		switch (prr) {
+		case 0x20:
+			boot_cpu_data.type = CPU_SH7724;
+			boot_cpu_data.flags |= CPU_HAS_L2_CACHE;
+			break;
+		case 0x10:
+		case 0x11:
+			boot_cpu_data.type = CPU_SH7757;
+			break;
+		case 0xd0:
+		case 0x40: /* yon-ten-go */
+			boot_cpu_data.type = CPU_SH7372;
+			break;
+		case 0xE0: /* 0x4E0 */
+			boot_cpu_data.type = CPU_SH7734; /* SH7733/SH7734 */
+			break;
+
+		}
+		break;
+	case 0x4000:	/* 1st cut */
+	case 0x4001:	/* 2nd cut */
+		boot_cpu_data.type = CPU_SHX3;
+		break;
+	case 0x700:
+		boot_cpu_data.type = CPU_SH4_501;
+		boot_cpu_data.flags &= ~CPU_HAS_FPU;
+		boot_cpu_data.icache.ways = 2;
+		boot_cpu_data.dcache.ways = 2;
+		break;
+	case 0x600:
+		boot_cpu_data.type = CPU_SH4_202;
+		boot_cpu_data.icache.ways = 2;
+		boot_cpu_data.dcache.ways = 2;
+		break;
+	case 0x500 ... 0x501:
+		switch (prr) {
+		case 0x10:
+			boot_cpu_data.type = CPU_SH7750R;
+			break;
+		case 0x11:
+			boot_cpu_data.type = CPU_SH7751R;
+			break;
+		case 0x50 ... 0x5f:
+			boot_cpu_data.type = CPU_SH7760;
+			break;
+		}
+
+		boot_cpu_data.icache.ways = 2;
+		boot_cpu_data.dcache.ways = 2;
+
+		break;
+	}
+
+	/*
+	 * On anything that's not a direct-mapped cache, look to the CVR
+	 * for I/D-cache specifics.
+	 */
+	if (boot_cpu_data.icache.ways > 1) {
+		size = sizes[(cvr >> 20) & 0xf];
+		boot_cpu_data.icache.way_incr	= (size >> 1);
+		boot_cpu_data.icache.sets	= (size >> 6);
+
+	}
+
+	/* And the rest of the D-cache */
+	if (boot_cpu_data.dcache.ways > 1) {
+		size = sizes[(cvr >> 16) & 0xf];
+		boot_cpu_data.dcache.way_incr	= (size >> 1);
+		boot_cpu_data.dcache.sets	= (size >> 6);
+	}
+
+	/*
+	 * SH-4A's have an optional PIPT L2.
+	 */
+	if (boot_cpu_data.flags & CPU_HAS_L2_CACHE) {
+		/*
+		 * Verify that it really has something hooked up, this
+		 * is the safety net for CPUs that have optional L2
+		 * support yet do not implement it.
+		 */
+		if ((cvr & 0xf) == 0)
+			boot_cpu_data.flags &= ~CPU_HAS_L2_CACHE;
+		else {
+			/*
+			 * Silicon and specifications have clearly never
+			 * met..
+			 */
+			cvr ^= 0xf;
+
+			/*
+			 * Size calculation is much more sensible
+			 * than it is for the L1.
+			 *
+			 * Sizes are 128KB, 256KB, 512KB, and 1MB.
+			 */
+			size = (cvr & 0xf) << 17;
+
+			boot_cpu_data.scache.way_incr		= (1 << 16);
+			boot_cpu_data.scache.entry_shift	= 5;
+			boot_cpu_data.scache.ways		= 4;
+			boot_cpu_data.scache.linesz		= L1_CACHE_BYTES;
+
+			boot_cpu_data.scache.entry_mask	=
+				(boot_cpu_data.scache.way_incr -
+				 boot_cpu_data.scache.linesz);
+
+			boot_cpu_data.scache.sets	= size /
+				(boot_cpu_data.scache.linesz *
+				 boot_cpu_data.scache.ways);
+
+			boot_cpu_data.scache.way_size	=
+				(boot_cpu_data.scache.sets *
+				 boot_cpu_data.scache.linesz);
+		}
+	}
+}
diff --git a/arch/sh/kernel/cpu/sh4/setup-sh4-202.c b/arch/sh/kernel/cpu/sh4/setup-sh4-202.c
new file mode 100644
index 000000000..e7a7b3cdf
--- /dev/null
+++ b/arch/sh/kernel/cpu/sh4/setup-sh4-202.c
@@ -0,0 +1,142 @@
+/*
+ * SH4-202 Setup
+ *
+ *  Copyright (C) 2006  Paul Mundt
+ *  Copyright (C) 2009  Magnus Damm
+ *
+ * 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.
+ */
+#include <linux/platform_device.h>
+#include <linux/init.h>
+#include <linux/serial.h>
+#include <linux/serial_sci.h>
+#include <linux/sh_timer.h>
+#include <linux/sh_intc.h>
+#include <linux/io.h>
+
+static struct plat_sci_port scif0_platform_data = {
+	.flags		= UPF_BOOT_AUTOCONF,
+	.scscr		= SCSCR_RE | SCSCR_TE | SCSCR_REIE,
+	.type		= PORT_SCIF,
+};
+
+static struct resource scif0_resources[] = {
+	DEFINE_RES_MEM(0xffe80000, 0x100),
+	DEFINE_RES_IRQ(evt2irq(0x700)),
+	DEFINE_RES_IRQ(evt2irq(0x720)),
+	DEFINE_RES_IRQ(evt2irq(0x760)),
+	DEFINE_RES_IRQ(evt2irq(0x740)),
+};
+
+static struct platform_device scif0_device = {
+	.name		= "sh-sci",
+	.id		= 0,
+	.resource	= scif0_resources,
+	.num_resources	= ARRAY_SIZE(scif0_resources),
+	.dev		= {
+		.platform_data	= &scif0_platform_data,
+	},
+};
+
+static struct sh_timer_config tmu0_platform_data = {
+	.channels_mask = 7,
+};
+
+static struct resource tmu0_resources[] = {
+	DEFINE_RES_MEM(0xffd80000, 0x30),
+	DEFINE_RES_IRQ(evt2irq(0x400)),
+	DEFINE_RES_IRQ(evt2irq(0x420)),
+	DEFINE_RES_IRQ(evt2irq(0x440)),
+};
+
+static struct platform_device tmu0_device = {
+	.name		= "sh-tmu",
+	.id		= 0,
+	.dev = {
+		.platform_data	= &tmu0_platform_data,
+	},
+	.resource	= tmu0_resources,
+	.num_resources	= ARRAY_SIZE(tmu0_resources),
+};
+
+static struct platform_device *sh4202_devices[] __initdata = {
+	&scif0_device,
+	&tmu0_device,
+};
+
+static int __init sh4202_devices_setup(void)
+{
+	return platform_add_devices(sh4202_devices,
+				    ARRAY_SIZE(sh4202_devices));
+}
+arch_initcall(sh4202_devices_setup);
+
+static struct platform_device *sh4202_early_devices[] __initdata = {
+	&scif0_device,
+	&tmu0_device,
+};
+
+void __init plat_early_device_setup(void)
+{
+	early_platform_add_devices(sh4202_early_devices,
+				   ARRAY_SIZE(sh4202_early_devices));
+}
+
+enum {
+	UNUSED = 0,
+
+	/* interrupt sources */
+	IRL0, IRL1, IRL2, IRL3, /* only IRLM mode supported */
+	HUDI, TMU0, TMU1, TMU2, RTC, SCIF, WDT,
+};
+
+static struct intc_vect vectors[] __initdata = {
+	INTC_VECT(HUDI, 0x600),
+	INTC_VECT(TMU0, 0x400), INTC_VECT(TMU1, 0x420),
+	INTC_VECT(TMU2, 0x440), INTC_VECT(TMU2, 0x460),
+	INTC_VECT(RTC, 0x480), INTC_VECT(RTC, 0x4a0),
+	INTC_VECT(RTC, 0x4c0),
+	INTC_VECT(SCIF, 0x700), INTC_VECT(SCIF, 0x720),
+	INTC_VECT(SCIF, 0x740), INTC_VECT(SCIF, 0x760),
+	INTC_VECT(WDT, 0x560),
+};
+
+static struct intc_prio_reg prio_registers[] __initdata = {
+	{ 0xffd00004, 0, 16, 4, /* IPRA */ { TMU0, TMU1, TMU2, RTC } },
+	{ 0xffd00008, 0, 16, 4, /* IPRB */ { WDT, 0, 0, 0 } },
+	{ 0xffd0000c, 0, 16, 4, /* IPRC */ { 0, 0, SCIF, HUDI } },
+	{ 0xffd00010, 0, 16, 4, /* IPRD */ { IRL0, IRL1, IRL2, IRL3 } },
+};
+
+static DECLARE_INTC_DESC(intc_desc, "sh4-202", vectors, NULL,
+			 NULL, prio_registers, NULL);
+
+static struct intc_vect vectors_irlm[] __initdata = {
+	INTC_VECT(IRL0, 0x240), INTC_VECT(IRL1, 0x2a0),
+	INTC_VECT(IRL2, 0x300), INTC_VECT(IRL3, 0x360),
+};
+
+static DECLARE_INTC_DESC(intc_desc_irlm, "sh4-202_irlm", vectors_irlm, NULL,
+			 NULL, prio_registers, NULL);
+
+void __init plat_irq_setup(void)
+{
+	register_intc_controller(&intc_desc);
+}
+
+#define INTC_ICR	0xffd00000UL
+#define INTC_ICR_IRLM   (1<<7)
+
+void __init plat_irq_setup_pins(int mode)
+{
+	switch (mode) {
+	case IRQ_MODE_IRQ: /* individual interrupt mode for IRL3-0 */
+		__raw_writew(__raw_readw(INTC_ICR) | INTC_ICR_IRLM, INTC_ICR);
+		register_intc_controller(&intc_desc_irlm);
+		break;
+	default:
+		BUG();
+	}
+}
diff --git a/arch/sh/kernel/cpu/sh4/setup-sh7750.c b/arch/sh/kernel/cpu/sh4/setup-sh7750.c
new file mode 100644
index 000000000..5f08c59b9
--- /dev/null
+++ b/arch/sh/kernel/cpu/sh4/setup-sh7750.c
@@ -0,0 +1,366 @@
+/*
+ * SH7091/SH7750/SH7750S/SH7750R/SH7751/SH7751R Setup
+ *
+ *  Copyright (C) 2006  Paul Mundt
+ *  Copyright (C) 2006  Jamie Lenehan
+ *
+ * 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.
+ */
+#include <linux/platform_device.h>
+#include <linux/init.h>
+#include <linux/serial.h>
+#include <linux/io.h>
+#include <linux/sh_timer.h>
+#include <linux/sh_intc.h>
+#include <linux/serial_sci.h>
+#include <generated/machtypes.h>
+
+static struct resource rtc_resources[] = {
+	[0] = {
+		.start	= 0xffc80000,
+		.end	= 0xffc80000 + 0x58 - 1,
+		.flags	= IORESOURCE_IO,
+	},
+	[1] = {
+		/* Shared Period/Carry/Alarm IRQ */
+		.start	= evt2irq(0x480),
+		.flags	= IORESOURCE_IRQ,
+	},
+};
+
+static struct platform_device rtc_device = {
+	.name		= "sh-rtc",
+	.id		= -1,
+	.num_resources	= ARRAY_SIZE(rtc_resources),
+	.resource	= rtc_resources,
+};
+
+static struct plat_sci_port sci_platform_data = {
+	.port_reg	= 0xffe0001C,
+	.flags		= UPF_BOOT_AUTOCONF,
+	.scscr		= SCSCR_TE | SCSCR_RE,
+	.type		= PORT_SCI,
+	.regshift	= 2,
+};
+
+static struct resource sci_resources[] = {
+	DEFINE_RES_MEM(0xffe00000, 0x100),
+	DEFINE_RES_IRQ(evt2irq(0x4e0)),
+};
+
+static struct platform_device sci_device = {
+	.name		= "sh-sci",
+	.id		= 0,
+	.resource	= sci_resources,
+	.num_resources	= ARRAY_SIZE(sci_resources),
+	.dev		= {
+		.platform_data	= &sci_platform_data,
+	},
+};
+
+static struct plat_sci_port scif_platform_data = {
+	.flags		= UPF_BOOT_AUTOCONF,
+	.scscr		= SCSCR_TE | SCSCR_RE | SCSCR_REIE,
+	.type		= PORT_SCIF,
+};
+
+static struct resource scif_resources[] = {
+	DEFINE_RES_MEM(0xffe80000, 0x100),
+	DEFINE_RES_IRQ(evt2irq(0x700)),
+};
+
+static struct platform_device scif_device = {
+	.name		= "sh-sci",
+	.id		= 1,
+	.resource	= scif_resources,
+	.num_resources	= ARRAY_SIZE(scif_resources),
+	.dev		= {
+		.platform_data	= &scif_platform_data,
+	},
+};
+
+static struct sh_timer_config tmu0_platform_data = {
+	.channels_mask = 7,
+};
+
+static struct resource tmu0_resources[] = {
+	DEFINE_RES_MEM(0xffd80000, 0x30),
+	DEFINE_RES_IRQ(evt2irq(0x400)),
+	DEFINE_RES_IRQ(evt2irq(0x420)),
+	DEFINE_RES_IRQ(evt2irq(0x440)),
+};
+
+static struct platform_device tmu0_device = {
+	.name		= "sh-tmu",
+	.id		= 0,
+	.dev = {
+		.platform_data	= &tmu0_platform_data,
+	},
+	.resource	= tmu0_resources,
+	.num_resources	= ARRAY_SIZE(tmu0_resources),
+};
+
+/* SH7750R, SH7751 and SH7751R all have two extra timer channels */
+#if defined(CONFIG_CPU_SUBTYPE_SH7750R) || \
+	defined(CONFIG_CPU_SUBTYPE_SH7751) || \
+	defined(CONFIG_CPU_SUBTYPE_SH7751R)
+
+static struct sh_timer_config tmu1_platform_data = {
+	.channels_mask = 3,
+};
+
+static struct resource tmu1_resources[] = {
+	DEFINE_RES_MEM(0xfe100000, 0x20),
+	DEFINE_RES_IRQ(evt2irq(0xb00)),
+	DEFINE_RES_IRQ(evt2irq(0xb80)),
+};
+
+static struct platform_device tmu1_device = {
+	.name		= "sh-tmu",
+	.id		= 1,
+	.dev = {
+		.platform_data	= &tmu1_platform_data,
+	},
+	.resource	= tmu1_resources,
+	.num_resources	= ARRAY_SIZE(tmu1_resources),
+};
+
+#endif
+
+static struct platform_device *sh7750_devices[] __initdata = {
+	&rtc_device,
+	&tmu0_device,
+#if defined(CONFIG_CPU_SUBTYPE_SH7750R) || \
+	defined(CONFIG_CPU_SUBTYPE_SH7751) || \
+	defined(CONFIG_CPU_SUBTYPE_SH7751R)
+	&tmu1_device,
+#endif
+};
+
+static int __init sh7750_devices_setup(void)
+{
+	if (mach_is_rts7751r2d()) {
+		platform_device_register(&scif_device);
+	} else {
+		platform_device_register(&sci_device);
+		platform_device_register(&scif_device);
+	}
+
+	return platform_add_devices(sh7750_devices,
+				    ARRAY_SIZE(sh7750_devices));
+}
+arch_initcall(sh7750_devices_setup);
+
+static struct platform_device *sh7750_early_devices[] __initdata = {
+	&tmu0_device,
+#if defined(CONFIG_CPU_SUBTYPE_SH7750R) || \
+	defined(CONFIG_CPU_SUBTYPE_SH7751) || \
+	defined(CONFIG_CPU_SUBTYPE_SH7751R)
+	&tmu1_device,
+#endif
+};
+
+void __init plat_early_device_setup(void)
+{
+	struct platform_device *dev[1];
+
+	if (mach_is_rts7751r2d()) {
+		scif_platform_data.scscr |= SCSCR_CKE1;
+		dev[0] = &scif_device;
+		early_platform_add_devices(dev, 1);
+	} else {
+		dev[0] = &sci_device;
+		early_platform_add_devices(dev, 1);
+		dev[0] = &scif_device;
+		early_platform_add_devices(dev, 1);
+	}
+
+	early_platform_add_devices(sh7750_early_devices,
+				   ARRAY_SIZE(sh7750_early_devices));
+}
+
+enum {
+	UNUSED = 0,
+
+	/* interrupt sources */
+	IRL0, IRL1, IRL2, IRL3, /* only IRLM mode supported */
+	HUDI, GPIOI, DMAC,
+	PCIC0_PCISERR, PCIC1_PCIERR, PCIC1_PCIPWDWN, PCIC1_PCIPWON,
+	PCIC1_PCIDMA0, PCIC1_PCIDMA1, PCIC1_PCIDMA2, PCIC1_PCIDMA3,
+	TMU3, TMU4, TMU0, TMU1, TMU2, RTC, SCI1, SCIF, WDT, REF,
+
+	/* interrupt groups */
+	PCIC1,
+};
+
+static struct intc_vect vectors[] __initdata = {
+	INTC_VECT(HUDI, 0x600), INTC_VECT(GPIOI, 0x620),
+	INTC_VECT(TMU0, 0x400), INTC_VECT(TMU1, 0x420),
+	INTC_VECT(TMU2, 0x440), INTC_VECT(TMU2, 0x460),
+	INTC_VECT(RTC, 0x480), INTC_VECT(RTC, 0x4a0),
+	INTC_VECT(RTC, 0x4c0),
+	INTC_VECT(SCI1, 0x4e0), INTC_VECT(SCI1, 0x500),
+	INTC_VECT(SCI1, 0x520), INTC_VECT(SCI1, 0x540),
+	INTC_VECT(SCIF, 0x700), INTC_VECT(SCIF, 0x720),
+	INTC_VECT(SCIF, 0x740), INTC_VECT(SCIF, 0x760),
+	INTC_VECT(WDT, 0x560),
+	INTC_VECT(REF, 0x580), INTC_VECT(REF, 0x5a0),
+};
+
+static struct intc_prio_reg prio_registers[] __initdata = {
+	{ 0xffd00004, 0, 16, 4, /* IPRA */ { TMU0, TMU1, TMU2, RTC } },
+	{ 0xffd00008, 0, 16, 4, /* IPRB */ { WDT, REF, SCI1, 0 } },
+	{ 0xffd0000c, 0, 16, 4, /* IPRC */ { GPIOI, DMAC, SCIF, HUDI } },
+	{ 0xffd00010, 0, 16, 4, /* IPRD */ { IRL0, IRL1, IRL2, IRL3 } },
+	{ 0xfe080000, 0, 32, 4, /* INTPRI00 */ { 0, 0, 0, 0,
+						 TMU4, TMU3,
+						 PCIC1, PCIC0_PCISERR } },
+};
+
+static DECLARE_INTC_DESC(intc_desc, "sh7750", vectors, NULL,
+			 NULL, prio_registers, NULL);
+
+/* SH7750, SH7750S, SH7751 and SH7091 all have 4-channel DMA controllers */
+#if defined(CONFIG_CPU_SUBTYPE_SH7750) || \
+	defined(CONFIG_CPU_SUBTYPE_SH7750S) || \
+	defined(CONFIG_CPU_SUBTYPE_SH7751) || \
+	defined(CONFIG_CPU_SUBTYPE_SH7091)
+static struct intc_vect vectors_dma4[] __initdata = {
+	INTC_VECT(DMAC, 0x640), INTC_VECT(DMAC, 0x660),
+	INTC_VECT(DMAC, 0x680), INTC_VECT(DMAC, 0x6a0),
+	INTC_VECT(DMAC, 0x6c0),
+};
+
+static DECLARE_INTC_DESC(intc_desc_dma4, "sh7750_dma4",
+			 vectors_dma4, NULL,
+			 NULL, prio_registers, NULL);
+#endif
+
+/* SH7750R and SH7751R both have 8-channel DMA controllers */
+#if defined(CONFIG_CPU_SUBTYPE_SH7750R) || defined(CONFIG_CPU_SUBTYPE_SH7751R)
+static struct intc_vect vectors_dma8[] __initdata = {
+	INTC_VECT(DMAC, 0x640), INTC_VECT(DMAC, 0x660),
+	INTC_VECT(DMAC, 0x680), INTC_VECT(DMAC, 0x6a0),
+	INTC_VECT(DMAC, 0x780), INTC_VECT(DMAC, 0x7a0),
+	INTC_VECT(DMAC, 0x7c0), INTC_VECT(DMAC, 0x7e0),
+	INTC_VECT(DMAC, 0x6c0),
+};
+
+static DECLARE_INTC_DESC(intc_desc_dma8, "sh7750_dma8",
+			 vectors_dma8, NULL,
+			 NULL, prio_registers, NULL);
+#endif
+
+/* SH7750R, SH7751 and SH7751R all have two extra timer channels */
+#if defined(CONFIG_CPU_SUBTYPE_SH7750R) || \
+	defined(CONFIG_CPU_SUBTYPE_SH7751) || \
+	defined(CONFIG_CPU_SUBTYPE_SH7751R)
+static struct intc_vect vectors_tmu34[] __initdata = {
+	INTC_VECT(TMU3, 0xb00), INTC_VECT(TMU4, 0xb80),
+};
+
+static struct intc_mask_reg mask_registers[] __initdata = {
+	{ 0xfe080040, 0xfe080060, 32, /* INTMSK00 / INTMSKCLR00 */
+	  { 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
+	    0, 0, 0, 0, 0, 0, TMU4, TMU3,
+	    PCIC1_PCIERR, PCIC1_PCIPWDWN, PCIC1_PCIPWON,
+	    PCIC1_PCIDMA0, PCIC1_PCIDMA1, PCIC1_PCIDMA2,
+	    PCIC1_PCIDMA3, PCIC0_PCISERR } },
+};
+
+static DECLARE_INTC_DESC(intc_desc_tmu34, "sh7750_tmu34",
+			 vectors_tmu34, NULL,
+			 mask_registers, prio_registers, NULL);
+#endif
+
+/* SH7750S, SH7750R, SH7751 and SH7751R all have IRLM priority registers */
+static struct intc_vect vectors_irlm[] __initdata = {
+	INTC_VECT(IRL0, 0x240), INTC_VECT(IRL1, 0x2a0),
+	INTC_VECT(IRL2, 0x300), INTC_VECT(IRL3, 0x360),
+};
+
+static DECLARE_INTC_DESC(intc_desc_irlm, "sh7750_irlm", vectors_irlm, NULL,
+			 NULL, prio_registers, NULL);
+
+/* SH7751 and SH7751R both have PCI */
+#if defined(CONFIG_CPU_SUBTYPE_SH7751) || defined(CONFIG_CPU_SUBTYPE_SH7751R)
+static struct intc_vect vectors_pci[] __initdata = {
+	INTC_VECT(PCIC0_PCISERR, 0xa00), INTC_VECT(PCIC1_PCIERR, 0xae0),
+	INTC_VECT(PCIC1_PCIPWDWN, 0xac0), INTC_VECT(PCIC1_PCIPWON, 0xaa0),
+	INTC_VECT(PCIC1_PCIDMA0, 0xa80), INTC_VECT(PCIC1_PCIDMA1, 0xa60),
+	INTC_VECT(PCIC1_PCIDMA2, 0xa40), INTC_VECT(PCIC1_PCIDMA3, 0xa20),
+};
+
+static struct intc_group groups_pci[] __initdata = {
+	INTC_GROUP(PCIC1, PCIC1_PCIERR, PCIC1_PCIPWDWN, PCIC1_PCIPWON,
+		   PCIC1_PCIDMA0, PCIC1_PCIDMA1, PCIC1_PCIDMA2, PCIC1_PCIDMA3),
+};
+
+static DECLARE_INTC_DESC(intc_desc_pci, "sh7750_pci", vectors_pci, groups_pci,
+			 mask_registers, prio_registers, NULL);
+#endif
+
+#if defined(CONFIG_CPU_SUBTYPE_SH7750) || \
+	defined(CONFIG_CPU_SUBTYPE_SH7750S) || \
+	defined(CONFIG_CPU_SUBTYPE_SH7091)
+void __init plat_irq_setup(void)
+{
+	/*
+	 * same vectors for SH7750, SH7750S and SH7091 except for IRLM,
+	 * see below..
+	 */
+	register_intc_controller(&intc_desc);
+	register_intc_controller(&intc_desc_dma4);
+}
+#endif
+
+#if defined(CONFIG_CPU_SUBTYPE_SH7750R)
+void __init plat_irq_setup(void)
+{
+	register_intc_controller(&intc_desc);
+	register_intc_controller(&intc_desc_dma8);
+	register_intc_controller(&intc_desc_tmu34);
+}
+#endif
+
+#if defined(CONFIG_CPU_SUBTYPE_SH7751)
+void __init plat_irq_setup(void)
+{
+	register_intc_controller(&intc_desc);
+	register_intc_controller(&intc_desc_dma4);
+	register_intc_controller(&intc_desc_tmu34);
+	register_intc_controller(&intc_desc_pci);
+}
+#endif
+
+#if defined(CONFIG_CPU_SUBTYPE_SH7751R)
+void __init plat_irq_setup(void)
+{
+	register_intc_controller(&intc_desc);
+	register_intc_controller(&intc_desc_dma8);
+	register_intc_controller(&intc_desc_tmu34);
+	register_intc_controller(&intc_desc_pci);
+}
+#endif
+
+#define INTC_ICR	0xffd00000UL
+#define INTC_ICR_IRLM   (1<<7)
+
+void __init plat_irq_setup_pins(int mode)
+{
+#if defined(CONFIG_CPU_SUBTYPE_SH7750) || defined(CONFIG_CPU_SUBTYPE_SH7091)
+	BUG(); /* impossible to mask interrupts on SH7750 and SH7091 */
+	return;
+#endif
+
+	switch (mode) {
+	case IRQ_MODE_IRQ: /* individual interrupt mode for IRL3-0 */
+		__raw_writew(__raw_readw(INTC_ICR) | INTC_ICR_IRLM, INTC_ICR);
+		register_intc_controller(&intc_desc_irlm);
+		break;
+	default:
+		BUG();
+	}
+}
diff --git a/arch/sh/kernel/cpu/sh4/setup-sh7760.c b/arch/sh/kernel/cpu/sh4/setup-sh7760.c
new file mode 100644
index 000000000..973b736b3
--- /dev/null
+++ b/arch/sh/kernel/cpu/sh4/setup-sh7760.c
@@ -0,0 +1,298 @@
+/*
+ * SH7760 Setup
+ *
+ *  Copyright (C) 2006  Paul Mundt
+ *
+ * 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.
+ */
+#include <linux/platform_device.h>
+#include <linux/init.h>
+#include <linux/serial.h>
+#include <linux/sh_timer.h>
+#include <linux/sh_intc.h>
+#include <linux/serial_sci.h>
+#include <linux/io.h>
+
+enum {
+	UNUSED = 0,
+
+	/* interrupt sources */
+	IRL0, IRL1, IRL2, IRL3,
+	HUDI, GPIOI, DMAC,
+	IRQ4, IRQ5, IRQ6, IRQ7,
+	HCAN20, HCAN21,
+	SSI0, SSI1,
+	HAC0, HAC1,
+	I2C0, I2C1,
+	USB, LCDC,
+	DMABRG0, DMABRG1, DMABRG2,
+	SCIF0_ERI, SCIF0_RXI, SCIF0_BRI, SCIF0_TXI,
+	SCIF1_ERI, SCIF1_RXI, SCIF1_BRI, SCIF1_TXI,
+	SCIF2_ERI, SCIF2_RXI, SCIF2_BRI, SCIF2_TXI,
+	SIM_ERI, SIM_RXI, SIM_TXI, SIM_TEI,
+	HSPI,
+	MMCIF0, MMCIF1, MMCIF2, MMCIF3,
+	MFI, ADC, CMT,
+	TMU0, TMU1, TMU2,
+	WDT, REF,
+
+	/* interrupt groups */
+	DMABRG, SCIF0, SCIF1, SCIF2, SIM, MMCIF,
+};
+
+static struct intc_vect vectors[] __initdata = {
+	INTC_VECT(HUDI, 0x600), INTC_VECT(GPIOI, 0x620),
+	INTC_VECT(DMAC, 0x640), INTC_VECT(DMAC, 0x660),
+	INTC_VECT(DMAC, 0x680), INTC_VECT(DMAC, 0x6a0),
+	INTC_VECT(DMAC, 0x780), INTC_VECT(DMAC, 0x7a0),
+	INTC_VECT(DMAC, 0x7c0), INTC_VECT(DMAC, 0x7e0),
+	INTC_VECT(DMAC, 0x6c0),
+	INTC_VECT(IRQ4, 0x800), INTC_VECT(IRQ5, 0x820),
+	INTC_VECT(IRQ6, 0x840), INTC_VECT(IRQ6, 0x860),
+	INTC_VECT(HCAN20, 0x900), INTC_VECT(HCAN21, 0x920),
+	INTC_VECT(SSI0, 0x940), INTC_VECT(SSI1, 0x960),
+	INTC_VECT(HAC0, 0x980), INTC_VECT(HAC1, 0x9a0),
+	INTC_VECT(I2C0, 0x9c0), INTC_VECT(I2C1, 0x9e0),
+	INTC_VECT(USB, 0xa00), INTC_VECT(LCDC, 0xa20),
+	INTC_VECT(DMABRG0, 0xa80), INTC_VECT(DMABRG1, 0xaa0),
+	INTC_VECT(DMABRG2, 0xac0),
+	INTC_VECT(SCIF0_ERI, 0x880), INTC_VECT(SCIF0_RXI, 0x8a0),
+	INTC_VECT(SCIF0_BRI, 0x8c0), INTC_VECT(SCIF0_TXI, 0x8e0),
+	INTC_VECT(SCIF1_ERI, 0xb00), INTC_VECT(SCIF1_RXI, 0xb20),
+	INTC_VECT(SCIF1_BRI, 0xb40), INTC_VECT(SCIF1_TXI, 0xb60),
+	INTC_VECT(SCIF2_ERI, 0xb80), INTC_VECT(SCIF2_RXI, 0xba0),
+	INTC_VECT(SCIF2_BRI, 0xbc0), INTC_VECT(SCIF2_TXI, 0xbe0),
+	INTC_VECT(SIM_ERI, 0xc00), INTC_VECT(SIM_RXI, 0xc20),
+	INTC_VECT(SIM_TXI, 0xc40), INTC_VECT(SIM_TEI, 0xc60),
+	INTC_VECT(HSPI, 0xc80),
+	INTC_VECT(MMCIF0, 0xd00), INTC_VECT(MMCIF1, 0xd20),
+	INTC_VECT(MMCIF2, 0xd40), INTC_VECT(MMCIF3, 0xd60),
+	INTC_VECT(MFI, 0xe80), /* 0xf80 according to data sheet */
+	INTC_VECT(ADC, 0xf80), INTC_VECT(CMT, 0xfa0),
+	INTC_VECT(TMU0, 0x400), INTC_VECT(TMU1, 0x420),
+	INTC_VECT(TMU2, 0x440), INTC_VECT(TMU2, 0x460),
+	INTC_VECT(WDT, 0x560),
+	INTC_VECT(REF, 0x580), INTC_VECT(REF, 0x5a0),
+};
+
+static struct intc_group groups[] __initdata = {
+	INTC_GROUP(DMABRG, DMABRG0, DMABRG1, DMABRG2),
+	INTC_GROUP(SCIF0, SCIF0_ERI, SCIF0_RXI, SCIF0_BRI, SCIF0_TXI),
+	INTC_GROUP(SCIF1, SCIF1_ERI, SCIF1_RXI, SCIF1_BRI, SCIF1_TXI),
+	INTC_GROUP(SCIF2, SCIF2_ERI, SCIF2_RXI, SCIF2_BRI, SCIF2_TXI),
+	INTC_GROUP(SIM, SIM_ERI, SIM_RXI, SIM_TXI, SIM_TEI),
+	INTC_GROUP(MMCIF, MMCIF0, MMCIF1, MMCIF2, MMCIF3),
+};
+
+static struct intc_mask_reg mask_registers[] __initdata = {
+	{ 0xfe080040, 0xfe080060, 32, /* INTMSK00 / INTMSKCLR00 */
+	  { IRQ4, IRQ5, IRQ6, IRQ7, 0, 0, HCAN20, HCAN21,
+	    SSI0, SSI1, HAC0, HAC1, I2C0, I2C1, USB, LCDC,
+	    0, DMABRG0, DMABRG1, DMABRG2,
+	    SCIF0_ERI, SCIF0_RXI, SCIF0_BRI, SCIF0_TXI,
+	    SCIF1_ERI, SCIF1_RXI, SCIF1_BRI, SCIF1_TXI,
+	    SCIF2_ERI, SCIF2_RXI, SCIF2_BRI, SCIF2_TXI, } },
+	{ 0xfe080044, 0xfe080064, 32, /* INTMSK04 / INTMSKCLR04 */
+	  { 0, 0, 0, 0, 0, 0, 0, 0,
+	    SIM_ERI, SIM_RXI, SIM_TXI, SIM_TEI,
+	    HSPI, MMCIF0, MMCIF1, MMCIF2,
+	    MMCIF3, 0, 0, 0, 0, 0, 0, 0,
+	    0, MFI, 0, 0, 0, 0, ADC, CMT, } },
+};
+
+static struct intc_prio_reg prio_registers[] __initdata = {
+	{ 0xffd00004, 0, 16, 4, /* IPRA */ { TMU0, TMU1, TMU2 } },
+	{ 0xffd00008, 0, 16, 4, /* IPRB */ { WDT, REF, 0, 0 } },
+	{ 0xffd0000c, 0, 16, 4, /* IPRC */ { GPIOI, DMAC, 0, HUDI } },
+	{ 0xffd00010, 0, 16, 4, /* IPRD */ { IRL0, IRL1, IRL2, IRL3 } },
+	{ 0xfe080000, 0, 32, 4, /* INTPRI00 */ { IRQ4, IRQ5, IRQ6, IRQ7 } },
+	{ 0xfe080004, 0, 32, 4, /* INTPRI04 */ { HCAN20, HCAN21, SSI0, SSI1,
+						 HAC0, HAC1, I2C0, I2C1 } },
+	{ 0xfe080008, 0, 32, 4, /* INTPRI08 */ { USB, LCDC, DMABRG, SCIF0,
+						 SCIF1, SCIF2, SIM, HSPI } },
+	{ 0xfe08000c, 0, 32, 4, /* INTPRI0C */ { 0, 0, MMCIF, 0,
+						 MFI, 0, ADC, CMT } },
+};
+
+static DECLARE_INTC_DESC(intc_desc, "sh7760", vectors, groups,
+			 mask_registers, prio_registers, NULL);
+
+static struct intc_vect vectors_irq[] __initdata = {
+	INTC_VECT(IRL0, 0x240), INTC_VECT(IRL1, 0x2a0),
+	INTC_VECT(IRL2, 0x300), INTC_VECT(IRL3, 0x360),
+};
+
+static DECLARE_INTC_DESC(intc_desc_irq, "sh7760-irq", vectors_irq, groups,
+			 mask_registers, prio_registers, NULL);
+
+static struct plat_sci_port scif0_platform_data = {
+	.flags		= UPF_BOOT_AUTOCONF,
+	.scscr		= SCSCR_RE | SCSCR_TE | SCSCR_REIE,
+	.type		= PORT_SCIF,
+	.regtype	= SCIx_SH4_SCIF_FIFODATA_REGTYPE,
+};
+
+static struct resource scif0_resources[] = {
+	DEFINE_RES_MEM(0xfe600000, 0x100),
+	DEFINE_RES_IRQ(evt2irq(0x880)),
+	DEFINE_RES_IRQ(evt2irq(0x8a0)),
+	DEFINE_RES_IRQ(evt2irq(0x8e0)),
+	DEFINE_RES_IRQ(evt2irq(0x8c0)),
+};
+
+static struct platform_device scif0_device = {
+	.name		= "sh-sci",
+	.id		= 0,
+	.resource	= scif0_resources,
+	.num_resources	= ARRAY_SIZE(scif0_resources),
+	.dev		= {
+		.platform_data	= &scif0_platform_data,
+	},
+};
+
+static struct plat_sci_port scif1_platform_data = {
+	.flags		= UPF_BOOT_AUTOCONF,
+	.type		= PORT_SCIF,
+	.scscr		= SCSCR_RE | SCSCR_TE | SCSCR_REIE,
+	.regtype	= SCIx_SH4_SCIF_FIFODATA_REGTYPE,
+};
+
+static struct resource scif1_resources[] = {
+	DEFINE_RES_MEM(0xfe610000, 0x100),
+	DEFINE_RES_IRQ(evt2irq(0xb00)),
+	DEFINE_RES_IRQ(evt2irq(0xb20)),
+	DEFINE_RES_IRQ(evt2irq(0xb60)),
+	DEFINE_RES_IRQ(evt2irq(0xb40)),
+};
+
+static struct platform_device scif1_device = {
+	.name		= "sh-sci",
+	.id		= 1,
+	.resource	= scif1_resources,
+	.num_resources	= ARRAY_SIZE(scif1_resources),
+	.dev		= {
+		.platform_data	= &scif1_platform_data,
+	},
+};
+
+static struct plat_sci_port scif2_platform_data = {
+	.flags		= UPF_BOOT_AUTOCONF,
+	.scscr		= SCSCR_RE | SCSCR_TE | SCSCR_REIE,
+	.type		= PORT_SCIF,
+	.regtype	= SCIx_SH4_SCIF_FIFODATA_REGTYPE,
+};
+
+static struct resource scif2_resources[] = {
+	DEFINE_RES_MEM(0xfe620000, 0x100),
+	DEFINE_RES_IRQ(evt2irq(0xb80)),
+	DEFINE_RES_IRQ(evt2irq(0xba0)),
+	DEFINE_RES_IRQ(evt2irq(0xbe0)),
+	DEFINE_RES_IRQ(evt2irq(0xbc0)),
+};
+
+static struct platform_device scif2_device = {
+	.name		= "sh-sci",
+	.id		= 2,
+	.resource	= scif2_resources,
+	.num_resources	= ARRAY_SIZE(scif2_resources),
+	.dev		= {
+		.platform_data	= &scif2_platform_data,
+	},
+};
+
+static struct plat_sci_port scif3_platform_data = {
+	.flags		= UPF_BOOT_AUTOCONF,
+	.scscr		= SCSCR_RE | SCSCR_TE | SCSCR_REIE,
+	.type		= PORT_SCI,
+	.regshift	= 2,
+};
+
+static struct resource scif3_resources[] = {
+	DEFINE_RES_MEM(0xfe480000, 0x100),
+	DEFINE_RES_IRQ(evt2irq(0xc00)),
+	DEFINE_RES_IRQ(evt2irq(0xc20)),
+	DEFINE_RES_IRQ(evt2irq(0xc40)),
+};
+
+static struct platform_device scif3_device = {
+	.name		= "sh-sci",
+	.id		= 3,
+	.resource	= scif3_resources,
+	.num_resources	= ARRAY_SIZE(scif3_resources),
+	.dev		= {
+		.platform_data	= &scif3_platform_data,
+	},
+};
+
+static struct sh_timer_config tmu0_platform_data = {
+	.channels_mask = 7,
+};
+
+static struct resource tmu0_resources[] = {
+	DEFINE_RES_MEM(0xffd80000, 0x30),
+	DEFINE_RES_IRQ(evt2irq(0x400)),
+	DEFINE_RES_IRQ(evt2irq(0x420)),
+	DEFINE_RES_IRQ(evt2irq(0x440)),
+};
+
+static struct platform_device tmu0_device = {
+	.name		= "sh-tmu",
+	.id		= 0,
+	.dev = {
+		.platform_data	= &tmu0_platform_data,
+	},
+	.resource	= tmu0_resources,
+	.num_resources	= ARRAY_SIZE(tmu0_resources),
+};
+
+
+static struct platform_device *sh7760_devices[] __initdata = {
+	&scif0_device,
+	&scif1_device,
+	&scif2_device,
+	&scif3_device,
+	&tmu0_device,
+};
+
+static int __init sh7760_devices_setup(void)
+{
+	return platform_add_devices(sh7760_devices,
+				    ARRAY_SIZE(sh7760_devices));
+}
+arch_initcall(sh7760_devices_setup);
+
+static struct platform_device *sh7760_early_devices[] __initdata = {
+	&scif0_device,
+	&scif1_device,
+	&scif2_device,
+	&scif3_device,
+	&tmu0_device,
+};
+
+void __init plat_early_device_setup(void)
+{
+	early_platform_add_devices(sh7760_early_devices,
+				   ARRAY_SIZE(sh7760_early_devices));
+}
+
+#define INTC_ICR	0xffd00000UL
+#define INTC_ICR_IRLM	(1 << 7)
+
+void __init plat_irq_setup_pins(int mode)
+{
+	switch (mode) {
+	case IRQ_MODE_IRQ:
+		__raw_writew(__raw_readw(INTC_ICR) | INTC_ICR_IRLM, INTC_ICR);
+		register_intc_controller(&intc_desc_irq);
+		break;
+	default:
+		BUG();
+	}
+}
+
+void __init plat_irq_setup(void)
+{
+	register_intc_controller(&intc_desc);
+}
diff --git a/arch/sh/kernel/cpu/sh4/softfloat.c b/arch/sh/kernel/cpu/sh4/softfloat.c
new file mode 100644
index 000000000..42edf2e54
--- /dev/null
+++ b/arch/sh/kernel/cpu/sh4/softfloat.c
@@ -0,0 +1,930 @@
+/*
+ * Floating point emulation support for subnormalised numbers on SH4
+ * architecture This file is derived from the SoftFloat IEC/IEEE
+ * Floating-point Arithmetic Package, Release 2 the original license of
+ * which is reproduced below.
+ *
+ * ========================================================================
+ *
+ * This C source file is part of the SoftFloat IEC/IEEE Floating-point
+ * Arithmetic Package, Release 2.
+ *
+ * Written by John R. Hauser.  This work was made possible in part by the
+ * International Computer Science Institute, located at Suite 600, 1947 Center
+ * Street, Berkeley, California 94704.  Funding was partially provided by the
+ * National Science Foundation under grant MIP-9311980.  The original version
+ * of this code was written as part of a project to build a fixed-point vector
+ * processor in collaboration with the University of California at Berkeley,
+ * overseen by Profs. Nelson Morgan and John Wawrzynek.  More information
+ * is available through the web page `http://HTTP.CS.Berkeley.EDU/~jhauser/
+ * arithmetic/softfloat.html'.
+ *
+ * THIS SOFTWARE IS DISTRIBUTED AS IS, FOR FREE.  Although reasonable effort
+ * has been made to avoid it, THIS SOFTWARE MAY CONTAIN FAULTS THAT WILL AT
+ * TIMES RESULT IN INCORRECT BEHAVIOR.  USE OF THIS SOFTWARE IS RESTRICTED TO
+ * PERSONS AND ORGANIZATIONS WHO CAN AND WILL TAKE FULL RESPONSIBILITY FOR ANY
+ * AND ALL LOSSES, COSTS, OR OTHER PROBLEMS ARISING FROM ITS USE.
+ *
+ * Derivative works are acceptable, even for commercial purposes, so long as
+ * (1) they include prominent notice that the work is derivative, and (2) they
+ * include prominent notice akin to these three paragraphs for those parts of
+ * this code that are retained.
+ *
+ * ========================================================================
+ *
+ * SH4 modifications by Ismail Dhaoui <ismail.dhaoui@st.com>
+ * and Kamel Khelifi <kamel.khelifi@st.com>
+ */
+#include <linux/kernel.h>
+#include <cpu/fpu.h>
+#include <asm/div64.h>
+
+#define LIT64( a ) a##LL
+
+typedef char flag;
+typedef unsigned char uint8;
+typedef signed char int8;
+typedef int uint16;
+typedef int int16;
+typedef unsigned int uint32;
+typedef signed int int32;
+
+typedef unsigned long long int bits64;
+typedef signed long long int sbits64;
+
+typedef unsigned char bits8;
+typedef signed char sbits8;
+typedef unsigned short int bits16;
+typedef signed short int sbits16;
+typedef unsigned int bits32;
+typedef signed int sbits32;
+
+typedef unsigned long long int uint64;
+typedef signed long long int int64;
+
+typedef unsigned long int float32;
+typedef unsigned long long float64;
+
+extern void float_raise(unsigned int flags);	/* in fpu.c */
+extern int float_rounding_mode(void);	/* in fpu.c */
+
+bits64 extractFloat64Frac(float64 a);
+flag extractFloat64Sign(float64 a);
+int16 extractFloat64Exp(float64 a);
+int16 extractFloat32Exp(float32 a);
+flag extractFloat32Sign(float32 a);
+bits32 extractFloat32Frac(float32 a);
+float64 packFloat64(flag zSign, int16 zExp, bits64 zSig);
+void shift64RightJamming(bits64 a, int16 count, bits64 * zPtr);
+float32 packFloat32(flag zSign, int16 zExp, bits32 zSig);
+void shift32RightJamming(bits32 a, int16 count, bits32 * zPtr);
+float64 float64_sub(float64 a, float64 b);
+float32 float32_sub(float32 a, float32 b);
+float32 float32_add(float32 a, float32 b);
+float64 float64_add(float64 a, float64 b);
+float64 float64_div(float64 a, float64 b);
+float32 float32_div(float32 a, float32 b);
+float32 float32_mul(float32 a, float32 b);
+float64 float64_mul(float64 a, float64 b);
+float32 float64_to_float32(float64 a);
+void add128(bits64 a0, bits64 a1, bits64 b0, bits64 b1, bits64 * z0Ptr,
+		   bits64 * z1Ptr);
+void sub128(bits64 a0, bits64 a1, bits64 b0, bits64 b1, bits64 * z0Ptr,
+		   bits64 * z1Ptr);
+void mul64To128(bits64 a, bits64 b, bits64 * z0Ptr, bits64 * z1Ptr);
+
+static int8 countLeadingZeros32(bits32 a);
+static int8 countLeadingZeros64(bits64 a);
+static float64 normalizeRoundAndPackFloat64(flag zSign, int16 zExp,
+					    bits64 zSig);
+static float64 subFloat64Sigs(float64 a, float64 b, flag zSign);
+static float64 addFloat64Sigs(float64 a, float64 b, flag zSign);
+static float32 roundAndPackFloat32(flag zSign, int16 zExp, bits32 zSig);
+static float32 normalizeRoundAndPackFloat32(flag zSign, int16 zExp,
+					    bits32 zSig);
+static float64 roundAndPackFloat64(flag zSign, int16 zExp, bits64 zSig);
+static float32 subFloat32Sigs(float32 a, float32 b, flag zSign);
+static float32 addFloat32Sigs(float32 a, float32 b, flag zSign);
+static void normalizeFloat64Subnormal(bits64 aSig, int16 * zExpPtr,
+				      bits64 * zSigPtr);
+static bits64 estimateDiv128To64(bits64 a0, bits64 a1, bits64 b);
+static void normalizeFloat32Subnormal(bits32 aSig, int16 * zExpPtr,
+				      bits32 * zSigPtr);
+
+bits64 extractFloat64Frac(float64 a)
+{
+	return a & LIT64(0x000FFFFFFFFFFFFF);
+}
+
+flag extractFloat64Sign(float64 a)
+{
+	return a >> 63;
+}
+
+int16 extractFloat64Exp(float64 a)
+{
+	return (a >> 52) & 0x7FF;
+}
+
+int16 extractFloat32Exp(float32 a)
+{
+	return (a >> 23) & 0xFF;
+}
+
+flag extractFloat32Sign(float32 a)
+{
+	return a >> 31;
+}
+
+bits32 extractFloat32Frac(float32 a)
+{
+	return a & 0x007FFFFF;
+}
+
+float64 packFloat64(flag zSign, int16 zExp, bits64 zSig)
+{
+	return (((bits64) zSign) << 63) + (((bits64) zExp) << 52) + zSig;
+}
+
+void shift64RightJamming(bits64 a, int16 count, bits64 * zPtr)
+{
+	bits64 z;
+
+	if (count == 0) {
+		z = a;
+	} else if (count < 64) {
+		z = (a >> count) | ((a << ((-count) & 63)) != 0);
+	} else {
+		z = (a != 0);
+	}
+	*zPtr = z;
+}
+
+static int8 countLeadingZeros32(bits32 a)
+{
+	static const int8 countLeadingZerosHigh[] = {
+		8, 7, 6, 6, 5, 5, 5, 5, 4, 4, 4, 4, 4, 4, 4, 4,
+		3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3,
+		2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2,
+		2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2,
+		1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,
+		1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,
+		1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,
+		1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,
+		0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
+		0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
+		0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
+		0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
+		0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
+		0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
+		0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
+		0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0
+	};
+	int8 shiftCount;
+
+	shiftCount = 0;
+	if (a < 0x10000) {
+		shiftCount += 16;
+		a <<= 16;
+	}
+	if (a < 0x1000000) {
+		shiftCount += 8;
+		a <<= 8;
+	}
+	shiftCount += countLeadingZerosHigh[a >> 24];
+	return shiftCount;
+
+}
+
+static int8 countLeadingZeros64(bits64 a)
+{
+	int8 shiftCount;
+
+	shiftCount = 0;
+	if (a < ((bits64) 1) << 32) {
+		shiftCount += 32;
+	} else {
+		a >>= 32;
+	}
+	shiftCount += countLeadingZeros32(a);
+	return shiftCount;
+
+}
+
+static float64 normalizeRoundAndPackFloat64(flag zSign, int16 zExp, bits64 zSig)
+{
+	int8 shiftCount;
+
+	shiftCount = countLeadingZeros64(zSig) - 1;
+	return roundAndPackFloat64(zSign, zExp - shiftCount,
+				   zSig << shiftCount);
+
+}
+
+static float64 subFloat64Sigs(float64 a, float64 b, flag zSign)
+{
+	int16 aExp, bExp, zExp;
+	bits64 aSig, bSig, zSig;
+	int16 expDiff;
+
+	aSig = extractFloat64Frac(a);
+	aExp = extractFloat64Exp(a);
+	bSig = extractFloat64Frac(b);
+	bExp = extractFloat64Exp(b);
+	expDiff = aExp - bExp;
+	aSig <<= 10;
+	bSig <<= 10;
+	if (0 < expDiff)
+		goto aExpBigger;
+	if (expDiff < 0)
+		goto bExpBigger;
+	if (aExp == 0) {
+		aExp = 1;
+		bExp = 1;
+	}
+	if (bSig < aSig)
+		goto aBigger;
+	if (aSig < bSig)
+		goto bBigger;
+	return packFloat64(float_rounding_mode() == FPSCR_RM_ZERO, 0, 0);
+      bExpBigger:
+	if (bExp == 0x7FF) {
+		return packFloat64(zSign ^ 1, 0x7FF, 0);
+	}
+	if (aExp == 0) {
+		++expDiff;
+	} else {
+		aSig |= LIT64(0x4000000000000000);
+	}
+	shift64RightJamming(aSig, -expDiff, &aSig);
+	bSig |= LIT64(0x4000000000000000);
+      bBigger:
+	zSig = bSig - aSig;
+	zExp = bExp;
+	zSign ^= 1;
+	goto normalizeRoundAndPack;
+      aExpBigger:
+	if (aExp == 0x7FF) {
+		return a;
+	}
+	if (bExp == 0) {
+		--expDiff;
+	} else {
+		bSig |= LIT64(0x4000000000000000);
+	}
+	shift64RightJamming(bSig, expDiff, &bSig);
+	aSig |= LIT64(0x4000000000000000);
+      aBigger:
+	zSig = aSig - bSig;
+	zExp = aExp;
+      normalizeRoundAndPack:
+	--zExp;
+	return normalizeRoundAndPackFloat64(zSign, zExp, zSig);
+
+}
+static float64 addFloat64Sigs(float64 a, float64 b, flag zSign)
+{
+	int16 aExp, bExp, zExp;
+	bits64 aSig, bSig, zSig;
+	int16 expDiff;
+
+	aSig = extractFloat64Frac(a);
+	aExp = extractFloat64Exp(a);
+	bSig = extractFloat64Frac(b);
+	bExp = extractFloat64Exp(b);
+	expDiff = aExp - bExp;
+	aSig <<= 9;
+	bSig <<= 9;
+	if (0 < expDiff) {
+		if (aExp == 0x7FF) {
+			return a;
+		}
+		if (bExp == 0) {
+			--expDiff;
+		} else {
+			bSig |= LIT64(0x2000000000000000);
+		}
+		shift64RightJamming(bSig, expDiff, &bSig);
+		zExp = aExp;
+	} else if (expDiff < 0) {
+		if (bExp == 0x7FF) {
+			return packFloat64(zSign, 0x7FF, 0);
+		}
+		if (aExp == 0) {
+			++expDiff;
+		} else {
+			aSig |= LIT64(0x2000000000000000);
+		}
+		shift64RightJamming(aSig, -expDiff, &aSig);
+		zExp = bExp;
+	} else {
+		if (aExp == 0x7FF) {
+			return a;
+		}
+		if (aExp == 0)
+			return packFloat64(zSign, 0, (aSig + bSig) >> 9);
+		zSig = LIT64(0x4000000000000000) + aSig + bSig;
+		zExp = aExp;
+		goto roundAndPack;
+	}
+	aSig |= LIT64(0x2000000000000000);
+	zSig = (aSig + bSig) << 1;
+	--zExp;
+	if ((sbits64) zSig < 0) {
+		zSig = aSig + bSig;
+		++zExp;
+	}
+      roundAndPack:
+	return roundAndPackFloat64(zSign, zExp, zSig);
+
+}
+
+float32 packFloat32(flag zSign, int16 zExp, bits32 zSig)
+{
+	return (((bits32) zSign) << 31) + (((bits32) zExp) << 23) + zSig;
+}
+
+void shift32RightJamming(bits32 a, int16 count, bits32 * zPtr)
+{
+	bits32 z;
+	if (count == 0) {
+		z = a;
+	} else if (count < 32) {
+		z = (a >> count) | ((a << ((-count) & 31)) != 0);
+	} else {
+		z = (a != 0);
+	}
+	*zPtr = z;
+}
+
+static float32 roundAndPackFloat32(flag zSign, int16 zExp, bits32 zSig)
+{
+	flag roundNearestEven;
+	int8 roundIncrement, roundBits;
+	flag isTiny;
+
+	/* SH4 has only 2 rounding modes - round to nearest and round to zero */
+	roundNearestEven = (float_rounding_mode() == FPSCR_RM_NEAREST);
+	roundIncrement = 0x40;
+	if (!roundNearestEven) {
+		roundIncrement = 0;
+	}
+	roundBits = zSig & 0x7F;
+	if (0xFD <= (bits16) zExp) {
+		if ((0xFD < zExp)
+		    || ((zExp == 0xFD)
+			&& ((sbits32) (zSig + roundIncrement) < 0))
+		    ) {
+			float_raise(FPSCR_CAUSE_OVERFLOW | FPSCR_CAUSE_INEXACT);
+			return packFloat32(zSign, 0xFF,
+					   0) - (roundIncrement == 0);
+		}
+		if (zExp < 0) {
+			isTiny = (zExp < -1)
+			    || (zSig + roundIncrement < 0x80000000);
+			shift32RightJamming(zSig, -zExp, &zSig);
+			zExp = 0;
+			roundBits = zSig & 0x7F;
+			if (isTiny && roundBits)
+				float_raise(FPSCR_CAUSE_UNDERFLOW);
+		}
+	}
+	if (roundBits)
+		float_raise(FPSCR_CAUSE_INEXACT);
+	zSig = (zSig + roundIncrement) >> 7;
+	zSig &= ~(((roundBits ^ 0x40) == 0) & roundNearestEven);
+	if (zSig == 0)
+		zExp = 0;
+	return packFloat32(zSign, zExp, zSig);
+
+}
+
+static float32 normalizeRoundAndPackFloat32(flag zSign, int16 zExp, bits32 zSig)
+{
+	int8 shiftCount;
+
+	shiftCount = countLeadingZeros32(zSig) - 1;
+	return roundAndPackFloat32(zSign, zExp - shiftCount,
+				   zSig << shiftCount);
+}
+
+static float64 roundAndPackFloat64(flag zSign, int16 zExp, bits64 zSig)
+{
+	flag roundNearestEven;
+	int16 roundIncrement, roundBits;
+	flag isTiny;
+
+	/* SH4 has only 2 rounding modes - round to nearest and round to zero */
+	roundNearestEven = (float_rounding_mode() == FPSCR_RM_NEAREST);
+	roundIncrement = 0x200;
+	if (!roundNearestEven) {
+		roundIncrement = 0;
+	}
+	roundBits = zSig & 0x3FF;
+	if (0x7FD <= (bits16) zExp) {
+		if ((0x7FD < zExp)
+		    || ((zExp == 0x7FD)
+			&& ((sbits64) (zSig + roundIncrement) < 0))
+		    ) {
+			float_raise(FPSCR_CAUSE_OVERFLOW | FPSCR_CAUSE_INEXACT);
+			return packFloat64(zSign, 0x7FF,
+					   0) - (roundIncrement == 0);
+		}
+		if (zExp < 0) {
+			isTiny = (zExp < -1)
+			    || (zSig + roundIncrement <
+				LIT64(0x8000000000000000));
+			shift64RightJamming(zSig, -zExp, &zSig);
+			zExp = 0;
+			roundBits = zSig & 0x3FF;
+			if (isTiny && roundBits)
+				float_raise(FPSCR_CAUSE_UNDERFLOW);
+		}
+	}
+	if (roundBits)
+		float_raise(FPSCR_CAUSE_INEXACT);
+	zSig = (zSig + roundIncrement) >> 10;
+	zSig &= ~(((roundBits ^ 0x200) == 0) & roundNearestEven);
+	if (zSig == 0)
+		zExp = 0;
+	return packFloat64(zSign, zExp, zSig);
+
+}
+
+static float32 subFloat32Sigs(float32 a, float32 b, flag zSign)
+{
+	int16 aExp, bExp, zExp;
+	bits32 aSig, bSig, zSig;
+	int16 expDiff;
+
+	aSig = extractFloat32Frac(a);
+	aExp = extractFloat32Exp(a);
+	bSig = extractFloat32Frac(b);
+	bExp = extractFloat32Exp(b);
+	expDiff = aExp - bExp;
+	aSig <<= 7;
+	bSig <<= 7;
+	if (0 < expDiff)
+		goto aExpBigger;
+	if (expDiff < 0)
+		goto bExpBigger;
+	if (aExp == 0) {
+		aExp = 1;
+		bExp = 1;
+	}
+	if (bSig < aSig)
+		goto aBigger;
+	if (aSig < bSig)
+		goto bBigger;
+	return packFloat32(float_rounding_mode() == FPSCR_RM_ZERO, 0, 0);
+      bExpBigger:
+	if (bExp == 0xFF) {
+		return packFloat32(zSign ^ 1, 0xFF, 0);
+	}
+	if (aExp == 0) {
+		++expDiff;
+	} else {
+		aSig |= 0x40000000;
+	}
+	shift32RightJamming(aSig, -expDiff, &aSig);
+	bSig |= 0x40000000;
+      bBigger:
+	zSig = bSig - aSig;
+	zExp = bExp;
+	zSign ^= 1;
+	goto normalizeRoundAndPack;
+      aExpBigger:
+	if (aExp == 0xFF) {
+		return a;
+	}
+	if (bExp == 0) {
+		--expDiff;
+	} else {
+		bSig |= 0x40000000;
+	}
+	shift32RightJamming(bSig, expDiff, &bSig);
+	aSig |= 0x40000000;
+      aBigger:
+	zSig = aSig - bSig;
+	zExp = aExp;
+      normalizeRoundAndPack:
+	--zExp;
+	return normalizeRoundAndPackFloat32(zSign, zExp, zSig);
+
+}
+
+static float32 addFloat32Sigs(float32 a, float32 b, flag zSign)
+{
+	int16 aExp, bExp, zExp;
+	bits32 aSig, bSig, zSig;
+	int16 expDiff;
+
+	aSig = extractFloat32Frac(a);
+	aExp = extractFloat32Exp(a);
+	bSig = extractFloat32Frac(b);
+	bExp = extractFloat32Exp(b);
+	expDiff = aExp - bExp;
+	aSig <<= 6;
+	bSig <<= 6;
+	if (0 < expDiff) {
+		if (aExp == 0xFF) {
+			return a;
+		}
+		if (bExp == 0) {
+			--expDiff;
+		} else {
+			bSig |= 0x20000000;
+		}
+		shift32RightJamming(bSig, expDiff, &bSig);
+		zExp = aExp;
+	} else if (expDiff < 0) {
+		if (bExp == 0xFF) {
+			return packFloat32(zSign, 0xFF, 0);
+		}
+		if (aExp == 0) {
+			++expDiff;
+		} else {
+			aSig |= 0x20000000;
+		}
+		shift32RightJamming(aSig, -expDiff, &aSig);
+		zExp = bExp;
+	} else {
+		if (aExp == 0xFF) {
+			return a;
+		}
+		if (aExp == 0)
+			return packFloat32(zSign, 0, (aSig + bSig) >> 6);
+		zSig = 0x40000000 + aSig + bSig;
+		zExp = aExp;
+		goto roundAndPack;
+	}
+	aSig |= 0x20000000;
+	zSig = (aSig + bSig) << 1;
+	--zExp;
+	if ((sbits32) zSig < 0) {
+		zSig = aSig + bSig;
+		++zExp;
+	}
+      roundAndPack:
+	return roundAndPackFloat32(zSign, zExp, zSig);
+
+}
+
+float64 float64_sub(float64 a, float64 b)
+{
+	flag aSign, bSign;
+
+	aSign = extractFloat64Sign(a);
+	bSign = extractFloat64Sign(b);
+	if (aSign == bSign) {
+		return subFloat64Sigs(a, b, aSign);
+	} else {
+		return addFloat64Sigs(a, b, aSign);
+	}
+
+}
+
+float32 float32_sub(float32 a, float32 b)
+{
+	flag aSign, bSign;
+
+	aSign = extractFloat32Sign(a);
+	bSign = extractFloat32Sign(b);
+	if (aSign == bSign) {
+		return subFloat32Sigs(a, b, aSign);
+	} else {
+		return addFloat32Sigs(a, b, aSign);
+	}
+
+}
+
+float32 float32_add(float32 a, float32 b)
+{
+	flag aSign, bSign;
+
+	aSign = extractFloat32Sign(a);
+	bSign = extractFloat32Sign(b);
+	if (aSign == bSign) {
+		return addFloat32Sigs(a, b, aSign);
+	} else {
+		return subFloat32Sigs(a, b, aSign);
+	}
+
+}
+
+float64 float64_add(float64 a, float64 b)
+{
+	flag aSign, bSign;
+
+	aSign = extractFloat64Sign(a);
+	bSign = extractFloat64Sign(b);
+	if (aSign == bSign) {
+		return addFloat64Sigs(a, b, aSign);
+	} else {
+		return subFloat64Sigs(a, b, aSign);
+	}
+}
+
+static void
+normalizeFloat64Subnormal(bits64 aSig, int16 * zExpPtr, bits64 * zSigPtr)
+{
+	int8 shiftCount;
+
+	shiftCount = countLeadingZeros64(aSig) - 11;
+	*zSigPtr = aSig << shiftCount;
+	*zExpPtr = 1 - shiftCount;
+}
+
+void add128(bits64 a0, bits64 a1, bits64 b0, bits64 b1, bits64 * z0Ptr,
+		   bits64 * z1Ptr)
+{
+	bits64 z1;
+
+	z1 = a1 + b1;
+	*z1Ptr = z1;
+	*z0Ptr = a0 + b0 + (z1 < a1);
+}
+
+void
+sub128(bits64 a0, bits64 a1, bits64 b0, bits64 b1, bits64 * z0Ptr,
+       bits64 * z1Ptr)
+{
+	*z1Ptr = a1 - b1;
+	*z0Ptr = a0 - b0 - (a1 < b1);
+}
+
+static bits64 estimateDiv128To64(bits64 a0, bits64 a1, bits64 b)
+{
+	bits64 b0, b1;
+	bits64 rem0, rem1, term0, term1;
+	bits64 z, tmp;
+	if (b <= a0)
+		return LIT64(0xFFFFFFFFFFFFFFFF);
+	b0 = b >> 32;
+	tmp = a0;
+	do_div(tmp, b0);
+
+	z = (b0 << 32 <= a0) ? LIT64(0xFFFFFFFF00000000) : tmp << 32;
+	mul64To128(b, z, &term0, &term1);
+	sub128(a0, a1, term0, term1, &rem0, &rem1);
+	while (((sbits64) rem0) < 0) {
+		z -= LIT64(0x100000000);
+		b1 = b << 32;
+		add128(rem0, rem1, b0, b1, &rem0, &rem1);
+	}
+	rem0 = (rem0 << 32) | (rem1 >> 32);
+	tmp = rem0;
+	do_div(tmp, b0);
+	z |= (b0 << 32 <= rem0) ? 0xFFFFFFFF : tmp;
+	return z;
+}
+
+void mul64To128(bits64 a, bits64 b, bits64 * z0Ptr, bits64 * z1Ptr)
+{
+	bits32 aHigh, aLow, bHigh, bLow;
+	bits64 z0, zMiddleA, zMiddleB, z1;
+
+	aLow = a;
+	aHigh = a >> 32;
+	bLow = b;
+	bHigh = b >> 32;
+	z1 = ((bits64) aLow) * bLow;
+	zMiddleA = ((bits64) aLow) * bHigh;
+	zMiddleB = ((bits64) aHigh) * bLow;
+	z0 = ((bits64) aHigh) * bHigh;
+	zMiddleA += zMiddleB;
+	z0 += (((bits64) (zMiddleA < zMiddleB)) << 32) + (zMiddleA >> 32);
+	zMiddleA <<= 32;
+	z1 += zMiddleA;
+	z0 += (z1 < zMiddleA);
+	*z1Ptr = z1;
+	*z0Ptr = z0;
+
+}
+
+static void normalizeFloat32Subnormal(bits32 aSig, int16 * zExpPtr,
+				      bits32 * zSigPtr)
+{
+	int8 shiftCount;
+
+	shiftCount = countLeadingZeros32(aSig) - 8;
+	*zSigPtr = aSig << shiftCount;
+	*zExpPtr = 1 - shiftCount;
+
+}
+
+float64 float64_div(float64 a, float64 b)
+{
+	flag aSign, bSign, zSign;
+	int16 aExp, bExp, zExp;
+	bits64 aSig, bSig, zSig;
+	bits64 rem0, rem1;
+	bits64 term0, term1;
+
+	aSig = extractFloat64Frac(a);
+	aExp = extractFloat64Exp(a);
+	aSign = extractFloat64Sign(a);
+	bSig = extractFloat64Frac(b);
+	bExp = extractFloat64Exp(b);
+	bSign = extractFloat64Sign(b);
+	zSign = aSign ^ bSign;
+	if (aExp == 0x7FF) {
+		if (bExp == 0x7FF) {
+		}
+		return packFloat64(zSign, 0x7FF, 0);
+	}
+	if (bExp == 0x7FF) {
+		return packFloat64(zSign, 0, 0);
+	}
+	if (bExp == 0) {
+		if (bSig == 0) {
+			if ((aExp | aSig) == 0) {
+				float_raise(FPSCR_CAUSE_INVALID);
+			}
+			return packFloat64(zSign, 0x7FF, 0);
+		}
+		normalizeFloat64Subnormal(bSig, &bExp, &bSig);
+	}
+	if (aExp == 0) {
+		if (aSig == 0)
+			return packFloat64(zSign, 0, 0);
+		normalizeFloat64Subnormal(aSig, &aExp, &aSig);
+	}
+	zExp = aExp - bExp + 0x3FD;
+	aSig = (aSig | LIT64(0x0010000000000000)) << 10;
+	bSig = (bSig | LIT64(0x0010000000000000)) << 11;
+	if (bSig <= (aSig + aSig)) {
+		aSig >>= 1;
+		++zExp;
+	}
+	zSig = estimateDiv128To64(aSig, 0, bSig);
+	if ((zSig & 0x1FF) <= 2) {
+		mul64To128(bSig, zSig, &term0, &term1);
+		sub128(aSig, 0, term0, term1, &rem0, &rem1);
+		while ((sbits64) rem0 < 0) {
+			--zSig;
+			add128(rem0, rem1, 0, bSig, &rem0, &rem1);
+		}
+		zSig |= (rem1 != 0);
+	}
+	return roundAndPackFloat64(zSign, zExp, zSig);
+
+}
+
+float32 float32_div(float32 a, float32 b)
+{
+	flag aSign, bSign, zSign;
+	int16 aExp, bExp, zExp;
+	bits32 aSig, bSig;
+	uint64_t zSig;
+
+	aSig = extractFloat32Frac(a);
+	aExp = extractFloat32Exp(a);
+	aSign = extractFloat32Sign(a);
+	bSig = extractFloat32Frac(b);
+	bExp = extractFloat32Exp(b);
+	bSign = extractFloat32Sign(b);
+	zSign = aSign ^ bSign;
+	if (aExp == 0xFF) {
+		if (bExp == 0xFF) {
+		}
+		return packFloat32(zSign, 0xFF, 0);
+	}
+	if (bExp == 0xFF) {
+		return packFloat32(zSign, 0, 0);
+	}
+	if (bExp == 0) {
+		if (bSig == 0) {
+			return packFloat32(zSign, 0xFF, 0);
+		}
+		normalizeFloat32Subnormal(bSig, &bExp, &bSig);
+	}
+	if (aExp == 0) {
+		if (aSig == 0)
+			return packFloat32(zSign, 0, 0);
+		normalizeFloat32Subnormal(aSig, &aExp, &aSig);
+	}
+	zExp = aExp - bExp + 0x7D;
+	aSig = (aSig | 0x00800000) << 7;
+	bSig = (bSig | 0x00800000) << 8;
+	if (bSig <= (aSig + aSig)) {
+		aSig >>= 1;
+		++zExp;
+	}
+	zSig = (((bits64) aSig) << 32);
+	do_div(zSig, bSig);
+
+	if ((zSig & 0x3F) == 0) {
+		zSig |= (((bits64) bSig) * zSig != ((bits64) aSig) << 32);
+	}
+	return roundAndPackFloat32(zSign, zExp, (bits32)zSig);
+
+}
+
+float32 float32_mul(float32 a, float32 b)
+{
+	char aSign, bSign, zSign;
+	int aExp, bExp, zExp;
+	unsigned int aSig, bSig;
+	unsigned long long zSig64;
+	unsigned int zSig;
+
+	aSig = extractFloat32Frac(a);
+	aExp = extractFloat32Exp(a);
+	aSign = extractFloat32Sign(a);
+	bSig = extractFloat32Frac(b);
+	bExp = extractFloat32Exp(b);
+	bSign = extractFloat32Sign(b);
+	zSign = aSign ^ bSign;
+	if (aExp == 0) {
+		if (aSig == 0)
+			return packFloat32(zSign, 0, 0);
+		normalizeFloat32Subnormal(aSig, &aExp, &aSig);
+	}
+	if (bExp == 0) {
+		if (bSig == 0)
+			return packFloat32(zSign, 0, 0);
+		normalizeFloat32Subnormal(bSig, &bExp, &bSig);
+	}
+	if ((bExp == 0xff && bSig == 0) || (aExp == 0xff && aSig == 0))
+		return roundAndPackFloat32(zSign, 0xff, 0);
+
+	zExp = aExp + bExp - 0x7F;
+	aSig = (aSig | 0x00800000) << 7;
+	bSig = (bSig | 0x00800000) << 8;
+	shift64RightJamming(((unsigned long long)aSig) * bSig, 32, &zSig64);
+	zSig = zSig64;
+	if (0 <= (signed int)(zSig << 1)) {
+		zSig <<= 1;
+		--zExp;
+	}
+	return roundAndPackFloat32(zSign, zExp, zSig);
+
+}
+
+float64 float64_mul(float64 a, float64 b)
+{
+	char aSign, bSign, zSign;
+	int aExp, bExp, zExp;
+	unsigned long long int aSig, bSig, zSig0, zSig1;
+
+	aSig = extractFloat64Frac(a);
+	aExp = extractFloat64Exp(a);
+	aSign = extractFloat64Sign(a);
+	bSig = extractFloat64Frac(b);
+	bExp = extractFloat64Exp(b);
+	bSign = extractFloat64Sign(b);
+	zSign = aSign ^ bSign;
+
+	if (aExp == 0) {
+		if (aSig == 0)
+			return packFloat64(zSign, 0, 0);
+		normalizeFloat64Subnormal(aSig, &aExp, &aSig);
+	}
+	if (bExp == 0) {
+		if (bSig == 0)
+			return packFloat64(zSign, 0, 0);
+		normalizeFloat64Subnormal(bSig, &bExp, &bSig);
+	}
+	if ((aExp == 0x7ff && aSig == 0) || (bExp == 0x7ff && bSig == 0))
+		return roundAndPackFloat64(zSign, 0x7ff, 0);
+
+	zExp = aExp + bExp - 0x3FF;
+	aSig = (aSig | 0x0010000000000000LL) << 10;
+	bSig = (bSig | 0x0010000000000000LL) << 11;
+	mul64To128(aSig, bSig, &zSig0, &zSig1);
+	zSig0 |= (zSig1 != 0);
+	if (0 <= (signed long long int)(zSig0 << 1)) {
+		zSig0 <<= 1;
+		--zExp;
+	}
+	return roundAndPackFloat64(zSign, zExp, zSig0);
+}
+
+/*
+ * -------------------------------------------------------------------------------
+ *  Returns the result of converting the double-precision floating-point value
+ *  `a' to the single-precision floating-point format.  The conversion is
+ *  performed according to the IEC/IEEE Standard for Binary Floating-point
+ *  Arithmetic.
+ *  -------------------------------------------------------------------------------
+ *  */
+float32 float64_to_float32(float64 a)
+{
+    flag aSign;
+    int16 aExp;
+    bits64 aSig;
+    bits32 zSig;
+
+    aSig = extractFloat64Frac( a );
+    aExp = extractFloat64Exp( a );
+    aSign = extractFloat64Sign( a );
+
+    shift64RightJamming( aSig, 22, &aSig );
+    zSig = aSig;
+    if ( aExp || zSig ) {
+        zSig |= 0x40000000;
+        aExp -= 0x381;
+    }
+    return roundAndPackFloat32(aSign, aExp, zSig);
+}
diff --git a/arch/sh/kernel/cpu/sh4/sq.c b/arch/sh/kernel/cpu/sh4/sq.c
new file mode 100644
index 000000000..0a47bd3e7
--- /dev/null
+++ b/arch/sh/kernel/cpu/sh4/sq.c
@@ -0,0 +1,416 @@
+/*
+ * arch/sh/kernel/cpu/sh4/sq.c
+ *
+ * General management API for SH-4 integrated Store Queues
+ *
+ * Copyright (C) 2001 - 2006  Paul Mundt
+ * Copyright (C) 2001, 2002  M. R. Brown
+ *
+ * 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.
+ */
+#include <linux/init.h>
+#include <linux/cpu.h>
+#include <linux/bitmap.h>
+#include <linux/device.h>
+#include <linux/kernel.h>
+#include <linux/module.h>
+#include <linux/slab.h>
+#include <linux/vmalloc.h>
+#include <linux/mm.h>
+#include <linux/io.h>
+#include <linux/prefetch.h>
+#include <asm/page.h>
+#include <asm/cacheflush.h>
+#include <cpu/sq.h>
+
+struct sq_mapping;
+
+struct sq_mapping {
+	const char *name;
+
+	unsigned long sq_addr;
+	unsigned long addr;
+	unsigned int size;
+
+	struct sq_mapping *next;
+};
+
+static struct sq_mapping *sq_mapping_list;
+static DEFINE_SPINLOCK(sq_mapping_lock);
+static struct kmem_cache *sq_cache;
+static unsigned long *sq_bitmap;
+
+#define store_queue_barrier()			\
+do {						\
+	(void)__raw_readl(P4SEG_STORE_QUE);	\
+	__raw_writel(0, P4SEG_STORE_QUE + 0);	\
+	__raw_writel(0, P4SEG_STORE_QUE + 8);	\
+} while (0);
+
+/**
+ * sq_flush_range - Flush (prefetch) a specific SQ range
+ * @start: the store queue address to start flushing from
+ * @len: the length to flush
+ *
+ * Flushes the store queue cache from @start to @start + @len in a
+ * linear fashion.
+ */
+void sq_flush_range(unsigned long start, unsigned int len)
+{
+	unsigned long *sq = (unsigned long *)start;
+
+	/* Flush the queues */
+	for (len >>= 5; len--; sq += 8)
+		prefetchw(sq);
+
+	/* Wait for completion */
+	store_queue_barrier();
+}
+EXPORT_SYMBOL(sq_flush_range);
+
+static inline void sq_mapping_list_add(struct sq_mapping *map)
+{
+	struct sq_mapping **p, *tmp;
+
+	spin_lock_irq(&sq_mapping_lock);
+
+	p = &sq_mapping_list;
+	while ((tmp = *p) != NULL)
+		p = &tmp->next;
+
+	map->next = tmp;
+	*p = map;
+
+	spin_unlock_irq(&sq_mapping_lock);
+}
+
+static inline void sq_mapping_list_del(struct sq_mapping *map)
+{
+	struct sq_mapping **p, *tmp;
+
+	spin_lock_irq(&sq_mapping_lock);
+
+	for (p = &sq_mapping_list; (tmp = *p); p = &tmp->next)
+		if (tmp == map) {
+			*p = tmp->next;
+			break;
+		}
+
+	spin_unlock_irq(&sq_mapping_lock);
+}
+
+static int __sq_remap(struct sq_mapping *map, pgprot_t prot)
+{
+#if defined(CONFIG_MMU)
+	struct vm_struct *vma;
+
+	vma = __get_vm_area(map->size, VM_ALLOC, map->sq_addr, SQ_ADDRMAX);
+	if (!vma)
+		return -ENOMEM;
+
+	vma->phys_addr = map->addr;
+
+	if (ioremap_page_range((unsigned long)vma->addr,
+			       (unsigned long)vma->addr + map->size,
+			       vma->phys_addr, prot)) {
+		vunmap(vma->addr);
+		return -EAGAIN;
+	}
+#else
+	/*
+	 * Without an MMU (or with it turned off), this is much more
+	 * straightforward, as we can just load up each queue's QACR with
+	 * the physical address appropriately masked.
+	 */
+	__raw_writel(((map->addr >> 26) << 2) & 0x1c, SQ_QACR0);
+	__raw_writel(((map->addr >> 26) << 2) & 0x1c, SQ_QACR1);
+#endif
+
+	return 0;
+}
+
+/**
+ * sq_remap - Map a physical address through the Store Queues
+ * @phys: Physical address of mapping.
+ * @size: Length of mapping.
+ * @name: User invoking mapping.
+ * @prot: Protection bits.
+ *
+ * Remaps the physical address @phys through the next available store queue
+ * address of @size length. @name is logged at boot time as well as through
+ * the sysfs interface.
+ */
+unsigned long sq_remap(unsigned long phys, unsigned int size,
+		       const char *name, pgprot_t prot)
+{
+	struct sq_mapping *map;
+	unsigned long end;
+	unsigned int psz;
+	int ret, page;
+
+	/* Don't allow wraparound or zero size */
+	end = phys + size - 1;
+	if (unlikely(!size || end < phys))
+		return -EINVAL;
+	/* Don't allow anyone to remap normal memory.. */
+	if (unlikely(phys < virt_to_phys(high_memory)))
+		return -EINVAL;
+
+	phys &= PAGE_MASK;
+	size = PAGE_ALIGN(end + 1) - phys;
+
+	map = kmem_cache_alloc(sq_cache, GFP_KERNEL);
+	if (unlikely(!map))
+		return -ENOMEM;
+
+	map->addr = phys;
+	map->size = size;
+	map->name = name;
+
+	page = bitmap_find_free_region(sq_bitmap, 0x04000000 >> PAGE_SHIFT,
+				       get_order(map->size));
+	if (unlikely(page < 0)) {
+		ret = -ENOSPC;
+		goto out;
+	}
+
+	map->sq_addr = P4SEG_STORE_QUE + (page << PAGE_SHIFT);
+
+	ret = __sq_remap(map, prot);
+	if (unlikely(ret != 0))
+		goto out;
+
+	psz = (size + (PAGE_SIZE - 1)) >> PAGE_SHIFT;
+	pr_info("sqremap: %15s  [%4d page%s]  va 0x%08lx   pa 0x%08lx\n",
+		likely(map->name) ? map->name : "???",
+		psz, psz == 1 ? " " : "s",
+		map->sq_addr, map->addr);
+
+	sq_mapping_list_add(map);
+
+	return map->sq_addr;
+
+out:
+	kmem_cache_free(sq_cache, map);
+	return ret;
+}
+EXPORT_SYMBOL(sq_remap);
+
+/**
+ * sq_unmap - Unmap a Store Queue allocation
+ * @vaddr: Pre-allocated Store Queue mapping.
+ *
+ * Unmaps the store queue allocation @map that was previously created by
+ * sq_remap(). Also frees up the pte that was previously inserted into
+ * the kernel page table and discards the UTLB translation.
+ */
+void sq_unmap(unsigned long vaddr)
+{
+	struct sq_mapping **p, *map;
+	int page;
+
+	for (p = &sq_mapping_list; (map = *p); p = &map->next)
+		if (map->sq_addr == vaddr)
+			break;
+
+	if (unlikely(!map)) {
+		printk("%s: bad store queue address 0x%08lx\n",
+		       __func__, vaddr);
+		return;
+	}
+
+	page = (map->sq_addr - P4SEG_STORE_QUE) >> PAGE_SHIFT;
+	bitmap_release_region(sq_bitmap, page, get_order(map->size));
+
+#ifdef CONFIG_MMU
+	{
+		/*
+		 * Tear down the VMA in the MMU case.
+		 */
+		struct vm_struct *vma;
+
+		vma = remove_vm_area((void *)(map->sq_addr & PAGE_MASK));
+		if (!vma) {
+			printk(KERN_ERR "%s: bad address 0x%08lx\n",
+			       __func__, map->sq_addr);
+			return;
+		}
+	}
+#endif
+
+	sq_mapping_list_del(map);
+
+	kmem_cache_free(sq_cache, map);
+}
+EXPORT_SYMBOL(sq_unmap);
+
+/*
+ * Needlessly complex sysfs interface. Unfortunately it doesn't seem like
+ * there is any other easy way to add things on a per-cpu basis without
+ * putting the directory entries somewhere stupid and having to create
+ * links in sysfs by hand back in to the per-cpu directories.
+ *
+ * Some day we may want to have an additional abstraction per store
+ * queue, but considering the kobject hell we already have to deal with,
+ * it's simply not worth the trouble.
+ */
+static struct kobject *sq_kobject[NR_CPUS];
+
+struct sq_sysfs_attr {
+	struct attribute attr;
+	ssize_t (*show)(char *buf);
+	ssize_t (*store)(const char *buf, size_t count);
+};
+
+#define to_sq_sysfs_attr(a)	container_of(a, struct sq_sysfs_attr, attr)
+
+static ssize_t sq_sysfs_show(struct kobject *kobj, struct attribute *attr,
+			     char *buf)
+{
+	struct sq_sysfs_attr *sattr = to_sq_sysfs_attr(attr);
+
+	if (likely(sattr->show))
+		return sattr->show(buf);
+
+	return -EIO;
+}
+
+static ssize_t sq_sysfs_store(struct kobject *kobj, struct attribute *attr,
+			      const char *buf, size_t count)
+{
+	struct sq_sysfs_attr *sattr = to_sq_sysfs_attr(attr);
+
+	if (likely(sattr->store))
+		return sattr->store(buf, count);
+
+	return -EIO;
+}
+
+static ssize_t mapping_show(char *buf)
+{
+	struct sq_mapping **list, *entry;
+	char *p = buf;
+
+	for (list = &sq_mapping_list; (entry = *list); list = &entry->next)
+		p += sprintf(p, "%08lx-%08lx [%08lx]: %s\n",
+			     entry->sq_addr, entry->sq_addr + entry->size,
+			     entry->addr, entry->name);
+
+	return p - buf;
+}
+
+static ssize_t mapping_store(const char *buf, size_t count)
+{
+	unsigned long base = 0, len = 0;
+
+	sscanf(buf, "%lx %lx", &base, &len);
+	if (!base)
+		return -EIO;
+
+	if (likely(len)) {
+		int ret = sq_remap(base, len, "Userspace", PAGE_SHARED);
+		if (ret < 0)
+			return ret;
+	} else
+		sq_unmap(base);
+
+	return count;
+}
+
+static struct sq_sysfs_attr mapping_attr =
+	__ATTR(mapping, 0644, mapping_show, mapping_store);
+
+static struct attribute *sq_sysfs_attrs[] = {
+	&mapping_attr.attr,
+	NULL,
+};
+
+static const struct sysfs_ops sq_sysfs_ops = {
+	.show	= sq_sysfs_show,
+	.store	= sq_sysfs_store,
+};
+
+static struct kobj_type ktype_percpu_entry = {
+	.sysfs_ops	= &sq_sysfs_ops,
+	.default_attrs	= sq_sysfs_attrs,
+};
+
+static int sq_dev_add(struct device *dev, struct subsys_interface *sif)
+{
+	unsigned int cpu = dev->id;
+	struct kobject *kobj;
+	int error;
+
+	sq_kobject[cpu] = kzalloc(sizeof(struct kobject), GFP_KERNEL);
+	if (unlikely(!sq_kobject[cpu]))
+		return -ENOMEM;
+
+	kobj = sq_kobject[cpu];
+	error = kobject_init_and_add(kobj, &ktype_percpu_entry, &dev->kobj,
+				     "%s", "sq");
+	if (!error)
+		kobject_uevent(kobj, KOBJ_ADD);
+	return error;
+}
+
+static int sq_dev_remove(struct device *dev, struct subsys_interface *sif)
+{
+	unsigned int cpu = dev->id;
+	struct kobject *kobj = sq_kobject[cpu];
+
+	kobject_put(kobj);
+	return 0;
+}
+
+static struct subsys_interface sq_interface = {
+	.name		= "sq",
+	.subsys		= &cpu_subsys,
+	.add_dev	= sq_dev_add,
+	.remove_dev	= sq_dev_remove,
+};
+
+static int __init sq_api_init(void)
+{
+	unsigned int nr_pages = 0x04000000 >> PAGE_SHIFT;
+	unsigned int size = (nr_pages + (BITS_PER_LONG - 1)) / BITS_PER_LONG;
+	int ret = -ENOMEM;
+
+	printk(KERN_NOTICE "sq: Registering store queue API.\n");
+
+	sq_cache = kmem_cache_create("store_queue_cache",
+				sizeof(struct sq_mapping), 0, 0, NULL);
+	if (unlikely(!sq_cache))
+		return ret;
+
+	sq_bitmap = kzalloc(size, GFP_KERNEL);
+	if (unlikely(!sq_bitmap))
+		goto out;
+
+	ret = subsys_interface_register(&sq_interface);
+	if (unlikely(ret != 0))
+		goto out;
+
+	return 0;
+
+out:
+	kfree(sq_bitmap);
+	kmem_cache_destroy(sq_cache);
+
+	return ret;
+}
+
+static void __exit sq_api_exit(void)
+{
+	subsys_interface_unregister(&sq_interface);
+	kfree(sq_bitmap);
+	kmem_cache_destroy(sq_cache);
+}
+
+module_init(sq_api_init);
+module_exit(sq_api_exit);
+
+MODULE_AUTHOR("Paul Mundt <lethal@linux-sh.org>, M. R. Brown <mrbrown@0xd6.org>");
+MODULE_DESCRIPTION("Simple API for SH-4 integrated Store Queues");
+MODULE_LICENSE("GPL");