Merge intel.com:/home/lenb/bk/linux-2.6.4

into intel.com:/home/lenb/src/linux-acpi-test-2.6.4

arch/arm/Makefile

@@ -31,6 +31,7 @@ comma = ,
 # Note that GCC does not numerically define an architecture version
 # macro, but instead defines a whole series of macros which makes
 # testing for a specific architecture or later rather impossible.
+arch-$(CONFIG_CPU_32v6)		:=-D__LINUX_ARM_ARCH__=6 -march=armv5t -Wa,-march=armv6
 arch-$(CONFIG_CPU_32v5)		:=-D__LINUX_ARM_ARCH__=5 $(call check_gcc,-march=armv5te,-march=armv4)
 arch-$(CONFIG_CPU_32v4)		:=-D__LINUX_ARM_ARCH__=4 -march=armv4
 arch-$(CONFIG_CPU_32v3)		:=-D__LINUX_ARM_ARCH__=3 -march=armv3
@@ -45,6 +46,7 @@ tune-$(CONFIG_CPU_ARM926T)	:=-mtune=arm9tdmi
 tune-$(CONFIG_CPU_SA110)	:=-mtune=strongarm110
 tune-$(CONFIG_CPU_SA1100)	:=-mtune=strongarm1100
 tune-$(CONFIG_CPU_XSCALE)	:=$(call check_gcc,-mtune=xscale,-mtune=strongarm110) -Wa,-mcpu=xscale
+tune-$(CONFIG_CPU_V6)		:=-mtune=strongarm
 
 # Need -Uarm for gcc < 3.x
 CFLAGS_BOOT	:=-mapcs-32 $(arch-y) $(tune-y) -mshort-load-bytes -msoft-float -Wa,-mno-fpu -Uarm

arch/arm/boot/compressed/head.S

@@ -585,7 +585,7 @@ __armv3_cache_off:
  * On entry,
  *  r6 = processor ID
  * On exit,
- *  r1, r2, r3, r12 corrupted
+ *  r1, r2, r3, r11, r12 corrupted
  * This routine must preserve:
  *  r0, r4, r5, r6, r7
  */
@@ -595,9 +595,25 @@ cache_clean_flush:
 		b	call_cache_fn
 
 __armv4_cache_flush:
-		bic	r1, pc, #31
-		add	r2, r1, #65536		@ 2x the largest dcache size
-1:		ldr	r3, [r1], #32		@ s/w flush D cache
+		mov	r2, #64*1024		@ default: 32K dcache size (*2)
+		mov	r11, #32		@ default: 32 byte line size
+		mrc	p15, 0, r3, c0, c0, 1	@ read cache type
+		teq	r3, r6			@ cache ID register present?
+		beq	no_cache_id
+		mov	r1, r3, lsr #18
+		and	r1, r1, #7
+		mov	r2, #1024
+		mov	r2, r2, lsl r1		@ base dcache size *2
+		tst	r3, #1 << 14		@ test M bit
+		addne	r2, r2, r2, lsr #1	@ +1/2 size if M == 1
+		mov	r3, r3, lsr #12
+		and	r3, r3, #3
+		mov	r11, #8
+		mov	r11, r11, lsl r3	@ cache line size in bytes
+no_cache_id:
+		bic	r1, pc, #63		@ align to longest cache line
+		add	r2, r1, r2
+1:		ldr	r3, [r1], r11		@ s/w flush D cache
 		teq	r1, r2
 		bne	1b
 

arch/arm/common/Makefile

@@ -5,6 +5,6 @@
 obj-y				+= platform.o
 obj-$(CONFIG_ARM_AMBA)		+= amba.o
 obj-$(CONFIG_ICST525)		+= icst525.o
-obj-$(CONFIG_SA1111)		+= sa1111.o sa1111-pcibuf.o sa1111-pcipool.o
+obj-$(CONFIG_SA1111)		+= sa1111.o sa1111-pcibuf.o
 obj-$(CONFIG_PCI_HOST_PLX90X0)	+= plx90x0.o
 obj-$(CONFIG_PCI_HOST_VIA82C505) += via82c505.o

arch/arm/common/amba.c

@@ -315,16 +315,30 @@ amba_find_device(const char *busid, struct device *parent, unsigned int id,
 	return data.dev;
 }
 
+/**
+ *	amba_request_regions - request all mem regions associated with device
+ *	@dev: amba_device structure for device
+ *	@name: name, or NULL to use driver name
+ */
 int amba_request_regions(struct amba_device *dev, const char *name)
 {
 	int ret = 0;
 
+	if (!name)
+		name = dev->dev.driver->name;
+
 	if (!request_mem_region(dev->res.start, SZ_4K, name))
 		ret = -EBUSY;
 
 	return ret;
 }
 
+/**
+ *	amba_release_regions - release mem regions assoicated with device
+ *	@dev: amba_device structure for device
+ *
+ *	Release regions claimed by a successful call to amba_request_regions.
+ */
 void amba_release_regions(struct amba_device *dev)
 {
 	release_mem_region(dev->res.start, SZ_4K);

arch/arm/kernel/Makefile

@@ -11,7 +11,6 @@ obj-y		:= arch.o compat.o dma.o entry-armv.o entry-common.o irq.o   \
 		   time.o traps.o
 
 obj-$(CONFIG_APM)		+= apm.o
-obj-$(CONFIG_PM)		+= pm.o
 obj-$(CONFIG_ARCH_ACORN)	+= ecard.o time-acorn.o
 obj-$(CONFIG_ARCH_CLPS7500)	+= time-acorn.o
 obj-$(CONFIG_FOOTBRIDGE)	+= isa.o

arch/arm/kernel/asm-offsets.c

@@ -47,6 +47,10 @@ int main(void)
 {
   DEFINE(TSK_ACTIVE_MM,		offsetof(struct task_struct, active_mm));
   BLANK();
+#if __LINUX_ARM_ARCH__ >= 6
+  DEFINE(MM_CONTEXT_ID,		offsetof(struct mm_struct, context.id));
+#endif
+  BLANK();
   DEFINE(VMA_VM_MM,		offsetof(struct vm_area_struct, vm_mm));
   DEFINE(VMA_VM_FLAGS,		offsetof(struct vm_area_struct, vm_flags));
   BLANK();

arch/arm/kernel/entry-header.S

@@ -66,6 +66,15 @@
 	msr	cpsr_c, \mode
 	.endm
 
+#if __LINUX_ARM_ARCH__ >= 6
+	.macro	disable_irq, temp
+	cpsid	i
+	.endm
+
+	.macro	enable_irq, temp
+	cpsie	i
+	.endm
+#else
 	.macro	disable_irq, temp
 	set_cpsr_c \temp, #PSR_I_BIT | MODE_SVC
 	.endm
@@ -73,6 +82,7 @@
 	.macro	enable_irq, temp
 	set_cpsr_c \temp, #MODE_SVC
 	.endm
+#endif
 
 	.macro	save_user_regs
 	sub	sp, sp, #S_FRAME_SIZE

arch/arm/kernel/irq.c

@@ -31,6 +31,7 @@
 #include <linux/seq_file.h>
 #include <linux/errno.h>
 #include <linux/list.h>
+#include <linux/kallsyms.h>
 
 #include <asm/irq.h>
 #include <asm/system.h>
@@ -225,6 +226,34 @@ static int check_irq_lock(struct irqdesc *desc, int irq, struct pt_regs *regs)
 }
 
 static void
+report_bad_irq(unsigned int irq, struct pt_regs *regs, struct irqdesc *desc, int ret)
+{
+	static int count = 100;
+	struct irqaction *action;
+
+	if (!count)
+		return;
+
+	count--;
+
+	if (ret != IRQ_HANDLED && ret != IRQ_NONE) {
+		printk("irq%u: bogus retval mask %x\n", irq, ret);
+	} else {
+		printk("irq%u: nobody cared\n", irq);
+	}
+	show_regs(regs);
+	dump_stack();
+	printk(KERN_ERR "handlers:");
+	action = desc->action;
+	do {
+		printk("\n" KERN_ERR "[<%p>]", action->handler);
+		print_symbol(" (%s)", (unsigned long)action->handler);
+		action = action->next;
+	} while (action);
+	printk("\n");
+}
+
+static int
 __do_irq(unsigned int irq, struct irqaction *action, struct pt_regs *regs)
 {
 	unsigned int status;
@@ -247,18 +276,7 @@ __do_irq(unsigned int irq, struct irqaction *action, struct pt_regs *regs)
 
 	spin_lock_irq(&irq_controller_lock);
 
-	if (retval != 1) {
-		static int count = 100;
-		if (count) {
-			count--;
-			if (retval) {
-				printk("irq event %d: bogus retval mask %x\n",
-					irq, retval);
-			} else {
-				printk("irq %d: nobody cared\n", irq);
-			}
-		}
-	}
+	return retval;
 }
 
 /*
@@ -276,8 +294,11 @@ do_simple_IRQ(unsigned int irq, struct irqdesc *desc, struct pt_regs *regs)
 	kstat_cpu(cpu).irqs[irq]++;
 
 	action = desc->action;
-	if (action)
-		__do_irq(irq, desc->action, regs);
+	if (action) {
+		int ret = __do_irq(irq, action, regs);
+		if (ret != IRQ_HANDLED)
+			report_bad_irq(irq, regs, desc, ret);
+	}
 }
 
 /*
@@ -313,6 +334,7 @@ do_edge_IRQ(unsigned int irq, struct irqdesc *desc, struct pt_regs *regs)
 
 	do {
 		struct irqaction *action;
+		int ret;
 
 		action = desc->action;
 		if (!action)
@@ -323,7 +345,9 @@ do_edge_IRQ(unsigned int irq, struct irqdesc *desc, struct pt_regs *regs)
 			desc->chip->unmask(irq);
 		}
 
-		__do_irq(irq, action, regs);
+		ret = __do_irq(irq, action, regs);
+		if (ret != IRQ_HANDLED)
+			report_bad_irq(irq, regs, desc, ret);
 	} while (desc->pending && !desc->disable_depth);
 
 	desc->running = 0;
@@ -368,7 +392,10 @@ do_level_IRQ(unsigned int irq, struct irqdesc *desc, struct pt_regs *regs)
 		 */
 		action = desc->action;
 		if (action) {
-			__do_irq(irq, desc->action, regs);
+			int ret = __do_irq(irq, desc->action, regs);
+
+			if (ret != IRQ_HANDLED)
+				report_bad_irq(irq, regs, desc, ret);
 
 			if (likely(!desc->disable_depth &&
 				   !check_irq_lock(desc, irq, regs)))

arch/arm/kernel/pm.c

-/*
- *  linux/arch/arm/kernel/suspend.c
- *
- * This program is free software; you can redistribute it and/or
- * modify it under the terms of the GNU General Public License.
- *
- *  This is the common support code for suspending an ARM machine.
- *  pm_do_suspend() is responsible for actually putting the CPU to
- *  sleep.
- */
-#include <linux/config.h>
-#include <linux/init.h>
-#include <linux/sysctl.h>
-#include <linux/pm.h>
-#include <linux/errno.h>
-#include <linux/sched.h>
-
-#ifdef CONFIG_SYSCTL
-/*
- * We really want this to die.  It's a disgusting hack using unallocated
- * sysctl numbers.  We should be using a real interface.
- */
-
-static int
-pm_sysctl_proc_handler(ctl_table *ctl, int write, struct file *filp,
-		       void *buffer, size_t *lenp)
-{
-	int ret = -EIO;
-	printk("PM: task %s (pid %d) uses deprecated sysctl PM interface\n",
-		current->comm, current->pid);
-	if (write)
-		ret = pm_suspend(PM_SUSPEND_MEM);
-	return ret;
-}
-
-/*
- * This came from arch/arm/mach-sa1100/pm.c:
- * Copyright (c) 2001 Cliff Brake <cbrake@accelent.com>
- *  with modifications by Nicolas Pitre and Russell King.
- *
- * ARGH!  ACPI people defined CTL_ACPI in linux/acpi.h rather than
- * linux/sysctl.h.
- *
- * This means our interface here won't survive long - it needs a new
- * interface.  Quick hack to get this working - use sysctl id 9999.
- */
-#warning ACPI broke the kernel, this interface needs to be fixed up.
-#define CTL_ACPI 9999
-#define ACPI_S1_SLP_TYP 19
-
-static struct ctl_table pm_table[] =
-{
-	{
-		.ctl_name	= ACPI_S1_SLP_TYP,
-		.procname	= "suspend",
-		.mode		= 0200,
-		.proc_handler	= pm_sysctl_proc_handler,
-	},
-	{0}
-};
-
-static struct ctl_table pm_dir_table[] =
-{
-	{
-		.ctl_name	= CTL_ACPI,
-		.procname	= "pm",
-		.mode		= 0555,
-		.child		= pm_table,
-	},
-	{0}
-};
-
-/*
- * Initialize power interface
- */
-static int __init pm_init(void)
-{
-	register_sysctl_table(pm_dir_table, 1);
-	return 0;
-}
-
-fs_initcall(pm_init);
-
-#endif

arch/arm/kernel/setup.c

@@ -118,21 +118,21 @@ static struct resource io_res[] = {
 #define lp2 io_res[2]
 
 static const char *cache_types[16] = {
-	"write-through",
-	"write-back",
-	"write-back",
+	"VIVT write-through",
+	"VIVT write-back",
+	"VIVT write-back",
 	"undefined 3",
 	"undefined 4",
 	"undefined 5",
-	"write-back",
-	"write-back",
+	"VIVT write-back",
+	"VIVT write-back",
 	"undefined 8",
 	"undefined 9",
 	"undefined 10",
 	"undefined 11",
 	"undefined 12",
 	"undefined 13",
-	"undefined 14",
+	"VIPT write-back",
 	"undefined 15",
 };
 
@@ -151,7 +151,7 @@ static const char *cache_clean[16] = {
 	"undefined 11",
 	"undefined 12",
 	"undefined 13",
-	"undefined 14",
+	"cp15 c7 ops",
 	"undefined 15",
 };
 
@@ -170,7 +170,7 @@ static const char *cache_lockdown[16] = {
 	"undefined 11",
 	"undefined 12",
 	"undefined 13",
-	"undefined 14",
+	"format C",
 	"undefined 15",
 };
 
@@ -183,7 +183,7 @@ static const char *proc_arch[] = {
 	"5T",
 	"5TE",
 	"5TEJ",
-	"?(9)",
+	"6TEJ",
 	"?(10)",
 	"?(11)",
 	"?(12)",

arch/arm/lib/io-readsl-armv4.S

@@ -19,10 +19,24 @@ ENTRY(__raw_readsl)
 		ands	ip, r1, #3
 		bne	2f
 
-1:		ldr	r3, [r0]
-		str	r3, [r1], #4
-		subs	r2, r2, #1
-		bne	1b
+		subs	r2, r2, #4
+		bmi	1001f
+		stmfd	sp!, {r4, lr}
+1000:		ldr	r3, [r0, #0]
+		ldr	r4, [r0, #0]
+		ldr	ip, [r0, #0]
+		ldr	lr, [r0, #0]
+		subs	r2, r2, #4
+		stmia	r1!, {r3, r4, ip, lr}
+		bpl	1000b
+		ldmfd	sp!, {r4, lr}
+1001:		tst	r2, #2
+		ldrne	r3, [r0, #0]
+		ldrne	ip, [r0, #0]
+		stmneia	r1!, {r3, ip}
+		tst	r2, #1
+		ldrne	r3, [r0, #0]
+		strne	r3, [r1, #0]
 		mov	pc, lr
 
 2:		cmp	ip, #2

arch/arm/mach-sa1100/generic.c

@@ -18,6 +18,7 @@
 #include <linux/cpufreq.h>
 #include <linux/ioport.h>
 
+#include <asm/div64.h>
 #include <asm/hardware.h>
 #include <asm/system.h>
 #include <asm/pgtable.h>
@@ -111,6 +112,21 @@ unsigned int cpufreq_get(unsigned int cpu)
 EXPORT_SYMBOL(cpufreq_get);
 #endif
 
+/*
+ * This is the SA11x0 sched_clock implementation.  This has
+ * a resolution of 271ns, and a maximum value of 1165s.
+ *  ( * 1E9 / 3686400 => * 78125 / 288)
+ */
+unsigned long long sched_clock(void)
+{
+	unsigned long long v;
+
+	v = (unsigned long long)OSCR * 78125;
+	do_div(v, 288);
+
+	return v;
+}
+
 /*
  * Default power-off for SA1100
  */
@@ -151,6 +167,36 @@ static struct platform_device sa11x0udc_device = {
 	.resource	= sa11x0udc_resources,
 };
 
+static struct resource sa11x0uart1_resources[] = {
+	[0] = {
+		.start	= 0x80010000,
+		.end	= 0x8001ffff,
+		.flags	= IORESOURCE_MEM,
+	},
+};
+
+static struct platform_device sa11x0uart1_device = {
+	.name		= "sa11x0-uart",
+	.id		= 1,
+	.num_resources	= ARRAY_SIZE(sa11x0uart1_resources),
+	.resource	= sa11x0uart1_resources,
+};
+
+static struct resource sa11x0uart3_resources[] = {
+	[0] = {
+		.start	= 0x80050000,
+		.end	= 0x8005ffff,
+		.flags	= IORESOURCE_MEM,
+	},
+};
+
+static struct platform_device sa11x0uart3_device = {
+	.name		= "sa11x0-uart",
+	.id		= 3,
+	.num_resources	= ARRAY_SIZE(sa11x0uart3_resources),
+	.resource	= sa11x0uart3_resources,
+};
+
 static struct resource sa11x0mcp_resources[] = {
 	[0] = {
 		.start	= 0x80060000,
@@ -218,6 +264,8 @@ static struct platform_device sa11x0pcmcia_device = {
 
 static struct platform_device *sa11x0_devices[] __initdata = {
 	&sa11x0udc_device,
+	&sa11x0uart1_device,
+	&sa11x0uart3_device,
 	&sa11x0mcp_device,
 	&sa11x0ssp_device,
 	&sa11x0pcmcia_device,

arch/arm/mm/Kconfig

@@ -223,6 +223,17 @@ config CPU_XSCALE
 	select CPU_TLB_V4WBI
 	select CPU_MINICACHE
 
+# ARMv6
+config CPU_V6
+	bool "Support ARM V6 processor"
+	depends on ARCH_INTEGRATOR
+	select CPU_32v6
+	select CPU_ABRT_EV6
+	select CPU_CACHE_V6
+	select CPU_COPY_V6
+	select CPU_TLB_V6
+
+# Figure out what processor architecture version we should be using.
 # This defines the compiler instruction set which depends on the machine type.
 config CPU_32v3
 	bool
@@ -233,6 +244,9 @@ config CPU_32v4
 config CPU_32v5
 	bool
 
+config CPU_32v6
+	bool
+
 # The abort model
 config CPU_ABRT_EV4
 	bool
@@ -249,6 +263,9 @@ config CPU_ABRT_EV5T
 config CPU_ABRT_EV5TJ
 	bool
 
+config CPU_ABRT_EV6
+	bool
+
 # The cache model
 config CPU_CACHE_V3
 	bool
@@ -262,6 +279,9 @@ config CPU_CACHE_V4WT
 config CPU_CACHE_V4WB
 	bool
 
+config CPU_CACHE_V6
+	bool
+
 # The copy-page model
 config CPU_COPY_V3
 	bool
@@ -272,6 +292,9 @@ config CPU_COPY_V4WT
 config CPU_COPY_V4WB
 	bool
 
+config CPU_COPY_V6
+	bool
+
 # This selects the TLB model
 config CPU_TLB_V3
 	bool
@@ -306,7 +329,7 @@ comment "Processor Features"
 
 config ARM_THUMB
 	bool "Support Thumb user binaries"
-	depends on CPU_ARM720T || CPU_ARM920T || CPU_ARM922T || CPU_ARM925T || CPU_ARM926T || CPU_ARM1020 || CPU_ARM1020E || CPU_ARM1022 || CPU_ARM1026 || CPU_XSCALE
+	depends on CPU_ARM720T || CPU_ARM920T || CPU_ARM922T || CPU_ARM925T || CPU_ARM926T || CPU_ARM1020 || CPU_ARM1020E || CPU_ARM1022 || CPU_ARM1026 || CPU_XSCALE || CPU_V6
 	default y
 	help
 	  Say Y if you want to include kernel support for running user space

arch/arm/mm/Makefile

@@ -15,15 +15,18 @@ obj-$(CONFIG_CPU_ABRT_EV4T)	+= abort-ev4t.o
 obj-$(CONFIG_CPU_ABRT_LV4T)	+= abort-lv4t.o
 obj-$(CONFIG_CPU_ABRT_EV5T)	+= abort-ev5t.o
 obj-$(CONFIG_CPU_ABRT_EV5TJ)	+= abort-ev5tj.o
+obj-$(CONFIG_CPU_ABRT_EV6)	+= abort-ev6.o
 
 obj-$(CONFIG_CPU_CACHE_V3)	+= cache-v3.o
 obj-$(CONFIG_CPU_CACHE_V4)	+= cache-v4.o
 obj-$(CONFIG_CPU_CACHE_V4WT)	+= cache-v4wt.o
 obj-$(CONFIG_CPU_CACHE_V4WB)	+= cache-v4wb.o
+obj-$(CONFIG_CPU_CACHE_V6)	+= cache-v6.o
 
 obj-$(CONFIG_CPU_COPY_V3)	+= copypage-v3.o
 obj-$(CONFIG_CPU_COPY_V4WT)	+= copypage-v4wt.o
 obj-$(CONFIG_CPU_COPY_V4WB)	+= copypage-v4wb.o
+obj-$(CONFIG_CPU_COPY_V6)	+= copypage-v6.o mmu.o
 obj-$(CONFIG_CPU_SA1100)	+= copypage-v4mc.o
 obj-$(CONFIG_CPU_XSCALE)	+= copypage-xscale.o
 
@@ -33,6 +36,7 @@ obj-$(CONFIG_CPU_TLB_V3)	+= tlb-v3.o
 obj-$(CONFIG_CPU_TLB_V4WT)	+= tlb-v4.o
 obj-$(CONFIG_CPU_TLB_V4WB)	+= tlb-v4wb.o
 obj-$(CONFIG_CPU_TLB_V4WBI)	+= tlb-v4wbi.o
+obj-$(CONFIG_CPU_TLB_V6)	+= tlb-v6.o
 
 obj-$(CONFIG_CPU_ARM610)	+= proc-arm6_7.o
 obj-$(CONFIG_CPU_ARM710)	+= proc-arm6_7.o
@@ -48,3 +52,4 @@ obj-$(CONFIG_CPU_ARM1026)	+= proc-arm1026.o
 obj-$(CONFIG_CPU_SA110)		+= proc-sa110.o
 obj-$(CONFIG_CPU_SA1100)	+= proc-sa1100.o
 obj-$(CONFIG_CPU_XSCALE)	+= proc-xscale.o
+obj-$(CONFIG_CPU_V6)		+= proc-v6.o blockops.o

arch/arm/mm/abort-ev6.S

+#include <linux/linkage.h>
+#include <asm/assembler.h>
+/*
+ * Function: v6_early_abort
+ *
+ * Params  : r2 = address of aborted instruction
+ *         : r3 = saved SPSR
+ *
+ * Returns : r0 = address of abort
+ *	   : r1 = FSR, bit 11 = write
+ *	   : r2-r8 = corrupted
+ *	   : r9 = preserved
+ *	   : sp = pointer to registers
+ *
+ * Purpose : obtain information about current aborted instruction.
+ */
+	.align	5
+ENTRY(v6_early_abort)
+	mrc	p15, 0, r1, c5, c0, 0		@ get FSR
+	mrc	p15, 0, r0, c6, c0, 0		@ get FAR
+	mov	pc, lr
+
+

arch/arm/mm/abort-lv4t.S

@@ -188,38 +188,33 @@ ENTRY(v4t_late_abort)
 .data_thumb_pushpop:
 	tst	r8, #1 << 10
 	beq	.data_unknown
-	mov	r7, #0x11
-	and	r6, r8, r7
-	and	r2, r8, r7, lsl #1
+	and	r6, r8, #0x55			@ hweight8(r8) + R bit
+	and	r2, r8, #0xaa
 	add	r6, r6, r2, lsr #1
-	and	r2, r8, r7, lsl #2
+	and	r2, r6, #0xcc
+	and	r6, r6, #0x33
 	add	r6, r6, r2, lsr #2
-	and	r2, r8, r7, lsl #3
-	add	r6, r6, r2, lsr #3
-	add	r6, r6, r6, lsr #4
-	and	r2, r8, #0x0100			@ catch 'R' bit for push/pop
-	add	r6, r6, r2, lsr #8
+	movs	r7, r8, lsr #9			@ C = r8 bit 8 (R bit)
+	adc	r6, r6, r6, lsr #4		@ high + low nibble + R bit
 	and	r6, r6, #15			@ number of regs to transfer
 	ldr	r7, [sp, #13 << 2]
 	tst	r8, #1 << 11
-	addne	r7, r7, r6, lsl #2		@ increment SP if PUSH
-	subeq	r7, r7, r6, lsl #2		@ decrement SP if POP
+	addeq	r7, r7, r6, lsl #2		@ increment SP if PUSH
+	subne	r7, r7, r6, lsl #2		@ decrement SP if POP
 	str	r7, [sp, #13 << 2]
 	mov	pc, lr
 
 .data_thumb_ldmstm:
-	mov	r7, #0x11
-	and	r6, r8, r7
-	and	r2, r8, r7, lsl #1
+	and	r6, r8, #0x55			@ hweight8(r8)
+	and	r2, r8, #0xaa
 	add	r6, r6, r2, lsr #1
-	and	r2, r8, r7, lsl #2
+	and	r2, r6, #0xcc
+	and	r6, r6, #0x33
 	add	r6, r6, r2, lsr #2
-	and	r2, r8, r7, lsl #3
-	add	r6, r6, r2, lsr #3
 	add	r6, r6, r6, lsr #4
-	and	r6, r6, #15			@ number of regs to transfer
 	and	r5, r8, #7 << 8
 	ldr	r7, [sp, r5, lsr #6]
+	and	r6, r6, #15			@ number of regs to transfer
 	sub	r7, r7, r6, lsl #2		@ always decrement
 	str	r7, [sp, r5, lsr #6]
 	mov	pc, lr

arch/arm/mm/blockops.c

+#include <linux/kernel.h>
+#include <linux/init.h>
+#include <linux/errno.h>
+#include <linux/mm.h>
+
+#include <asm/memory.h>
+#include <asm/ptrace.h>
+#include <asm/cacheflush.h>
+#include <asm/traps.h>
+
+extern struct cpu_cache_fns blk_cache_fns;
+
+#define HARVARD_CACHE
+
+/*
+ *	blk_flush_kern_dcache_page(kaddr)
+ *
+ *	Ensure that the data held in the page kaddr is written back
+ *	to the page in question.
+ *
+ *	- kaddr   - kernel address (guaranteed to be page aligned)
+ */
+static void __attribute__((naked))
+blk_flush_kern_dcache_page(void *kaddr)
+{
+	asm(
+	"add	r1, r0, %0							\n\
+1:	.word	0xec401f0e	@ mcrr	p15, 0, r0, r1, c14, 0	@ blocking	\n\
+	mov	r0, #0								\n\
+	mcr	p15, 0, r0, c7, c5, 0						\n\
+	mcr	p15, 0, r0, c7, c10, 4						\n\
+	mov	pc, lr"
+	:
+	: "I" (PAGE_SIZE));
+}
+
+/*
+ *	blk_dma_inv_range(start,end)
+ *
+ *	Invalidate the data cache within the specified region; we will
+ *	be performing a DMA operation in this region and we want to
+ *	purge old data in the cache.
+ *
+ *	- start   - virtual start address of region
+ *	- end     - virtual end address of region
+ */
+static void __attribute__((naked))
+blk_dma_inv_range_unified(unsigned long start, unsigned long end)
+{
+	asm(
+	"tst	r0, %0								\n\
+	mcrne	p15, 0, r0, c7, c11, 1		@ clean unified line		\n\
+	tst	r1, %0								\n\
+	mcrne	p15, 0, r1, c7, c15, 1		@ clean & invalidate unified line\n\
+	.word	0xec401f06	@ mcrr	p15, 0, r1, r0, c6, 0	@ blocking	\n\
+	mov	r0, #0								\n\
+	mcr	p15, 0, r0, c7, c10, 4		@ drain write buffer		\n\
+	mov	pc, lr"
+	:
+	: "I" (L1_CACHE_BYTES - 1));
+}
+
+static void __attribute__((naked))
+blk_dma_inv_range_harvard(unsigned long start, unsigned long end)
+{
+	asm(
+	"tst	r0, %0								\n\
+	mcrne	p15, 0, r0, c7, c10, 1		@ clean D line			\n\
+	tst	r1, %0								\n\
+	mcrne	p15, 0, r1, c7, c14, 1		@ clean & invalidate D line	\n\
+	.word	0xec401f06	@ mcrr	p15, 0, r1, r0, c6, 0	@ blocking	\n\
+	mov	r0, #0								\n\
+	mcr	p15, 0, r0, c7, c10, 4		@ drain write buffer		\n\
+	mov	pc, lr"
+	:
+	: "I" (L1_CACHE_BYTES - 1));
+}
+
+/*
+ *	blk_dma_clean_range(start,end)
+ *	- start   - virtual start address of region
+ *	- end     - virtual end address of region
+ */
+static void __attribute__((naked))
+blk_dma_clean_range(unsigned long start, unsigned long end)
+{
+	asm(
+	".word	0xec401f0c	@ mcrr	p15, 0, r1, r0, c12, 0	@ blocking	\n\
+	mov	r0, #0								\n\
+	mcr	p15, 0, r0, c7, c10, 4		@ drain write buffer		\n\
+	mov	pc, lr");
+}
+
+/*
+ *	blk_dma_flush_range(start,end)
+ *	- start   - virtual start address of region
+ *	- end     - virtual end address of region
+ */
+static void __attribute__((naked))
+blk_dma_flush_range(unsigned long start, unsigned long end)
+{
+	asm(
+	".word	0xec401f0e	@ mcrr	p15, 0, r1, r0, c14, 0	@ blocking	\n\
+	mov	pc, lr");
+}
+
+static int blockops_trap(struct pt_regs *regs, unsigned int instr)
+{
+	regs->ARM_r4 |= regs->ARM_r2;
+	regs->ARM_pc += 4;
+	return 0;
+}
+
+static char *func[] = {
+	"Prefetch data range",
+	"Clean+Invalidate data range",
+	"Clean data range",
+	"Invalidate data range",
+	"Invalidate instr range"
+};
+
+static struct undef_hook blockops_hook __initdata = {
+	.instr_mask	= 0x0fffffd0,
+	.instr_val	= 0x0c401f00,
+	.cpsr_mask	= PSR_T_BIT,
+	.cpsr_val	= 0,
+	.fn		= blockops_trap,
+};
+
+static int __init blockops_check(void)
+{
+	register unsigned int err asm("r4") = 0;
+	unsigned int cache_type;
+	int i;
+
+	asm("mcr p15, 0, %0, c0, c0, 1" : "=r" (cache_type));
+
+	printk("Checking V6 block cache operations:\n");
+	register_undef_hook(&blockops_hook);
+
+	__asm__ ("mov	r0, %0\n\t"
+		"mov	r1, %1\n\t"
+		"mov	r2, #1\n\t"
+		".word	0xec401f2c @ mcrr p15, 0, r1, r0, c12, 2\n\t"
+		"mov	r2, #2\n\t"
+		".word	0xec401f0e @ mcrr p15, 0, r1, r0, c14, 0\n\t"
+		"mov	r2, #4\n\t"
+		".word	0xec401f0c @ mcrr p15, 0, r1, r0, c12, 0\n\t"
+		"mov	r2, #8\n\t"
+		".word	0xec401f06 @ mcrr p15, 0, r1, r0, c6, 0\n\t"
+		"mov	r2, #16\n\t"
+		".word	0xec401f05 @ mcrr p15, 0, r1, r0, c5, 0\n\t"
+		:
+		: "r" (PAGE_OFFSET), "r" (PAGE_OFFSET + 128)
+		: "r0", "r1", "r2");
+
+	unregister_undef_hook(&blockops_hook);
+
+	for (i = 0; i < ARRAY_SIZE(func); i++, err >>= 1)
+		printk("%30s: %ssupported\n", func[i], err & 1 ? "not " : "");
+
+	if ((err & 8) == 0) {
+		printk(" --> Using %s block cache invalidate\n",
+			cache_type & (1 << 24) ? "harvard" : "unified");
+		if (cache_type & (1 << 24))
+			cpu_cache.dma_inv_range = blk_dma_inv_range_harvard;
+		else
+			cpu_cache.dma_inv_range = blk_dma_inv_range_unified;
+	}
+	if ((err & 4) == 0) {
+		printk(" --> Using block cache clean\n");
+		cpu_cache.dma_clean_range        = blk_dma_clean_range;
+	}
+	if ((err & 2) == 0) {
+		printk(" --> Using block cache clean+invalidate\n");
+		cpu_cache.dma_flush_range        = blk_dma_flush_range;
+		cpu_cache.flush_kern_dcache_page = blk_flush_kern_dcache_page;
+	}
+
+	return 0;
+}
+
+__initcall(blockops_check);

arch/arm/mm/cache-v6.S

+/*
+ *  linux/arch/arm/mm/cache-v6.S
+ *
+ *  Copyright (C) 2001 Deep Blue Solutions Ltd.
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2 as
+ * published by the Free Software Foundation.
+ *
+ *  This is the "shell" of the ARMv6 processor support.
+ */
+#include <linux/linkage.h>
+#include <linux/init.h>
+#include <asm/assembler.h>
+
+#include "proc-macros.S"
+
+#define HARVARD_CACHE
+#define CACHE_LINE_SIZE		32
+#define D_CACHE_LINE_SIZE	32
+
+/*
+ *	v6_flush_cache_all()
+ *
+ *	Flush the entire cache.
+ *
+ *	It is assumed that:
+ */
+ENTRY(v6_flush_kern_cache_all)
+	mov	r0, #0
+#ifdef HARVARD_CACHE
+	mcr	p15, 0, r0, c7, c14, 0		@ D cache clean+invalidate
+	mcr	p15, 0, r0, c7, c5, 0		@ I+BTB cache invalidate
+#else
+	mcr	p15, 0, r0, c7, c15, 0		@ Cache clean+invalidate
+#endif
+	mov	pc, lr
+
+/*
+ *	v6_flush_cache_all()
+ *
+ *	Flush all TLB entries in a particular address space
+ *
+ *	- mm    - mm_struct describing address space
+ */
+ENTRY(v6_flush_user_cache_all)
+	/*FALLTHROUGH*/
+
+/*
+ *	v6_flush_cache_range(start, end, flags)
+ *
+ *	Flush a range of TLB entries in the specified address space.
+ *
+ *	- start - start address (may not be aligned)
+ *	- end   - end address (exclusive, may not be aligned)
+ *	- flags	- vm_area_struct flags describing address space
+ *
+ *	It is assumed that:
+ *	- we have a VIPT cache.
+ */
+ENTRY(v6_flush_user_cache_range)
+	mov	pc, lr
+
+/*
+ *	v6_coherent_kern_range(start,end)
+ *
+ *	Ensure that the I and D caches are coherent within specified
+ *	region.  This is typically used when code has been written to
+ *	a memory region, and will be executed.
+ *
+ *	- start   - virtual start address of region
+ *	- end     - virtual end address of region
+ *
+ *	It is assumed that:
+ *	- the Icache does not read data from the write buffer
+ */
+ENTRY(v6_coherent_kern_range)
+	bic	r0, r0, #CACHE_LINE_SIZE - 1
+1:
+#ifdef HARVARD_CACHE
+	mcr	p15, 0, r0, c7, c10, 1		@ clean D line
+	mcr	p15, 0, r0, c7, c5, 1		@ invalidate I line
+#endif
+	mcr	p15, 0, r0, c7, c5, 7		@ invalidate BTB entry
+	add	r0, r0, #CACHE_LINE_SIZE
+	cmp	r0, r1
+	blo	1b
+#ifdef HARVARD_CACHE
+	mov	r0, #0
+	mcr	p15, 0, r0, c7, c10, 4		@ drain write buffer
+#endif
+	mov	pc, lr
+
+/*
+ *	v6_flush_kern_dcache_page(kaddr)
+ *
+ *	Ensure that the data held in the page kaddr is written back
+ *	to the page in question.
+ *
+ *	- kaddr   - kernel address (guaranteed to be page aligned)
+ */
+ENTRY(v6_flush_kern_dcache_page)
+	add	r1, r0, #PAGE_SZ
+1:
+#ifdef HARVARD_CACHE
+	mcr	p15, 0, r0, c7, c14, 1		@ clean & invalidate D line
+#else
+	mcr	p15, 0, r0, c7, c15, 1		@ clean & invalidate unified line
+#endif	
+	add	r0, r0, #D_CACHE_LINE_SIZE
+	cmp	r0, r1
+	blo	1b
+#ifdef HARVARD_CACHE
+	mov	r0, #0
+	mcr	p15, 0, r0, c7, c10, 4
+#endif
+	mov	pc, lr
+
+
+/*
+ *	v6_dma_inv_range(start,end)
+ *
+ *	Invalidate the data cache within the specified region; we will
+ *	be performing a DMA operation in this region and we want to
+ *	purge old data in the cache.
+ *
+ *	- start   - virtual start address of region
+ *	- end     - virtual end address of region
+ */
+ENTRY(v6_dma_inv_range)
+	tst	r0, #D_CACHE_LINE_SIZE - 1
+	bic	r0, r0, #D_CACHE_LINE_SIZE - 1
+#ifdef HARVARD_CACHE
+	mcrne	p15, 0, r0, c7, c10, 1		@ clean D line
+#else
+	mcrne	p15, 0, r0, c7, c11, 1		@ clean unified line
+#endif
+	tst	r1, #D_CACHE_LINE_SIZE - 1
+	bic	r1, r1, #D_CACHE_LINE_SIZE - 1
+#ifdef HARVARD_CACHE
+	mcrne	p15, 0, r1, c7, c14, 1		@ clean & invalidate D line
+#else
+	mcrne	p15, 0, r1, c7, c15, 1		@ clean & invalidate unified line
+#endif
+1:
+#ifdef HARVARD_CACHE
+	mcr	p15, 0, r0, c7, c6, 1		@ invalidate D line
+#else
+	mcr	p15, 0, r0, c7, c7, 1		@ invalidate unified line
+#endif
+	add	r0, r0, #D_CACHE_LINE_SIZE
+	cmp	r0, r1
+	blo	1b
+	mcr	p15, 0, r0, c7, c10, 4		@ drain write buffer
+	mov	pc, lr
+
+/*
+ *	v6_dma_clean_range(start,end)
+ *	- start   - virtual start address of region
+ *	- end     - virtual end address of region
+ */
+ENTRY(v6_dma_clean_range)
+	bic	r0, r0, #D_CACHE_LINE_SIZE - 1
+1:
+#ifdef HARVARD_CACHE
+	mcr	p15, 0, r0, c7, c10, 1		@ clean D line
+#else
+	mcr	p15, 0, r0, c7, c11, 1		@ clean unified line
+#endif
+	add	r0, r0, #D_CACHE_LINE_SIZE
+	cmp	r0, r1
+	blo	1b
+	mov	r0, #0
+	mcr	p15, 0, r0, c7, c10, 4		@ drain write buffer
+	mov	pc, lr
+
+/*
+ *	v6_dma_flush_range(start,end)
+ *	- start   - virtual start address of region
+ *	- end     - virtual end address of region
+ */
+ENTRY(v6_dma_flush_range)
+	bic	r0, r0, #D_CACHE_LINE_SIZE - 1
+1:
+#ifdef HARVARD_CACHE
+	mcr	p15, 0, r0, c7, c14, 1		@ clean & invalidate D line
+#else
+	mcr	p15, 0, r0, c7, c15, 1		@ clean & invalidate line
+#endif
+	add	r0, r0, #D_CACHE_LINE_SIZE
+	blo	1b
+	mov	r0, #0
+	mcr	p15, 0, r0, c7, c10, 4		@ drain write buffer
+	mov	pc, lr
+
+	__INITDATA
+
+	.type	v6_cache_fns, #object
+ENTRY(v6_cache_fns)
+	.long	v6_flush_kern_cache_all
+	.long	v6_flush_user_cache_all
+	.long	v6_flush_user_cache_range
+	.long	v6_coherent_kern_range
+	.long	v6_flush_kern_dcache_page
+	.long	v6_dma_inv_range
+	.long	v6_dma_clean_range
+	.long	v6_dma_flush_range
+	.size	v6_cache_fns, . - v6_cache_fns

arch/arm/mm/copypage-v6.c

+/*
+ *  linux/arch/arm/mm/copypage-v6.c
+ *
+ *  Copyright (C) 2002 Deep Blue Solutions Ltd, All Rights Reserved.
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2 as
+ * published by the Free Software Foundation.
+ */
+#include <linux/init.h>
+#include <linux/spinlock.h>
+#include <linux/mm.h>
+
+#include <asm/page.h>
+#include <asm/pgalloc.h>
+#include <asm/pgtable.h>
+#include <asm/shmparam.h>
+#include <asm/tlbflush.h>
+
+#if SHMLBA > 16384
+#error FIX ME
+#endif
+
+#define from_address	(0xffff8000)
+#define from_pgprot	PAGE_KERNEL
+#define to_address	(0xffffc000)
+#define to_pgprot	PAGE_KERNEL
+
+static pte_t *from_pte;
+static pte_t *to_pte;
+static spinlock_t v6_lock = SPIN_LOCK_UNLOCKED;
+
+#define DCACHE_COLOUR(vaddr) ((vaddr & (SHMLBA - 1)) >> PAGE_SHIFT)
+
+/*
+ * Copy the page, taking account of the cache colour.
+ */
+void v6_copy_user_page(void *kto, const void *kfrom, unsigned long vaddr)
+{
+	unsigned int offset = DCACHE_COLOUR(vaddr);
+	unsigned long from, to;
+
+	spin_lock(&v6_lock);
+
+	set_pte(from_pte + offset, pfn_pte(__pa(kfrom) >> PAGE_SHIFT, from_pgprot));
+	set_pte(to_pte + offset, pfn_pte(__pa(kto) >> PAGE_SHIFT, to_pgprot));
+
+	from = from_address + (offset << PAGE_SHIFT);
+	to   = to_address + (offset << PAGE_SHIFT);
+
+	flush_tlb_kernel_page(from);
+	flush_tlb_kernel_page(to);
+
+	copy_page((void *)to, (void *)from);
+
+	spin_unlock(&v6_lock);
+}
+
+void v6_clear_user_page(void *kaddr, unsigned long vaddr)
+{
+	unsigned int offset = DCACHE_COLOUR(vaddr);
+	unsigned long to = to_address + (offset << PAGE_SHIFT);
+
+	spin_lock(&v6_lock);
+
+	set_pte(to_pte + offset, pfn_pte(__pa(kaddr) >> PAGE_SHIFT, to_pgprot));
+	flush_tlb_kernel_page(to);
+	clear_page((void *)to);
+
+	spin_unlock(&v6_lock);
+}
+
+struct cpu_user_fns v6_user_fns __initdata = {
+	.cpu_clear_user_page	= v6_clear_user_page,
+	.cpu_copy_user_page	= v6_copy_user_page,
+};
+
+static int __init v6_userpage_init(void)
+{
+	pgd_t *pgd;
+	pmd_t *pmd;
+
+	pgd = pgd_offset_k(from_address);
+	pmd = pmd_alloc(&init_mm, pgd, from_address);
+	if (!pmd)
+		BUG();
+	from_pte = pte_alloc_kernel(&init_mm, pmd, from_address);
+	if (!from_pte)
+		BUG();
+
+	to_pte = pte_alloc_kernel(&init_mm, pmd, to_address);
+	if (!to_pte)
+		BUG();
+
+	return 0;
+}
+
+__initcall(v6_userpage_init);
+

arch/arm/mm/mm-armv.c

@@ -585,20 +585,31 @@ void __init iotable_init(struct map_desc *io_desc, int nr)
 		create_mapping(io_desc + i);
 }
 
-static inline void free_memmap(int node, unsigned long start, unsigned long end)
+static inline void
+free_memmap(int node, unsigned long start_pfn, unsigned long end_pfn)
 {
+	struct page *start_pg, *end_pg;
 	unsigned long pg, pgend;
 
-	start = __phys_to_virt(start);
-	end   = __phys_to_virt(end);
-
-	pg    = PAGE_ALIGN((unsigned long)(virt_to_page(start)));
-	pgend = ((unsigned long)(virt_to_page(end))) & PAGE_MASK;
+	/*
+	 * Convert start_pfn/end_pfn to a struct page pointer.
+	 */
+	start_pg = pfn_to_page(start_pfn);
+	end_pg = pfn_to_page(end_pfn);
 
-	start = __virt_to_phys(pg);
-	end   = __virt_to_phys(pgend);
+	/*
+	 * Convert to physical addresses, and
+	 * round start upwards and end downwards.
+	 */
+	pg = PAGE_ALIGN(__pa(start_pg));
+	pgend = __pa(end_pg) & PAGE_MASK;
 
-	free_bootmem_node(NODE_DATA(node), start, end - start);
+	/*
+	 * If there are free pages between these,
+	 * free the section of the memmap array.
+	 */
+	if (pg < pgend)
+		free_bootmem_node(NODE_DATA(node), pg, pgend - pg);
 }
 
 static inline void free_unused_memmap_node(int node, struct meminfo *mi)
@@ -615,7 +626,12 @@ static inline void free_unused_memmap_node(int node, struct meminfo *mi)
 		if (mi->bank[i].size == 0 || mi->bank[i].node != node)
 			continue;
 
-		bank_start = mi->bank[i].start & PAGE_MASK;
+		bank_start = mi->bank[i].start >> PAGE_SHIFT;
+		if (bank_start < prev_bank_end) {
+			printk(KERN_ERR "MEM: unordered memory banks.  "
+				"Not freeing memmap.\n");
+			break;
+		}
 
 		/*
 		 * If we had a previous bank, and there is a space
@@ -625,7 +641,7 @@ static inline void free_unused_memmap_node(int node, struct meminfo *mi)
 			free_memmap(node, prev_bank_end, bank_start);
 
 		prev_bank_end = PAGE_ALIGN(mi->bank[i].start +
-					   mi->bank[i].size);
+					   mi->bank[i].size) >> PAGE_SHIFT;
 	}
 }
 

arch/arm/mm/proc-macros.S

@@ -35,3 +35,17 @@
 	ldr	\rd, [\rd, #TI_TASK]
 	ldr	\rd, [\rd, #TSK_ACTIVE_MM]
 	.endm
+
+/*
+ * mmid - get context id from mm pointer (mm->context.id)
+ */
+	.macro	mmid, rd, rn
+	ldr	\rd, [\rn, #MM_CONTEXT_ID]
+	.endm
+
+/*
+ * mask_asid - mask the ASID from the context ID
+ */
+	.macro	asid, rd, rn
+	and	\rd, \rn, #255
+	.endm

arch/arm/mm/proc-v6.S

+/*
+ *  linux/arch/arm/mm/proc-v6.S
+ *
+ *  Copyright (C) 2001 Deep Blue Solutions Ltd.
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2 as
+ * published by the Free Software Foundation.
+ *
+ *  This is the "shell" of the ARMv6 processor support.
+ */
+#include <linux/linkage.h>
+#include <asm/assembler.h>
+#include <asm/constants.h>
+#include <asm/procinfo.h>
+#include <asm/pgtable.h>
+
+#include "proc-macros.S"
+
+#define D_CACHE_LINE_SIZE	32
+
+	.macro	cpsie, flags
+	.ifc \flags, f
+	.long	0xf1080040
+	.exitm
+	.endif
+	.ifc \flags, i
+	.long	0xf1080080
+	.exitm
+	.endif
+	.ifc \flags, if
+	.long	0xf10800c0
+	.exitm
+	.endif
+	.err
+	.endm
+
+	.macro	cpsid, flags
+	.ifc \flags, f
+	.long	0xf10c0040
+	.exitm
+	.endif
+	.ifc \flags, i
+	.long	0xf10c0080
+	.exitm
+	.endif
+	.ifc \flags, if
+	.long	0xf10c00c0
+	.exitm
+	.endif
+	.err
+	.endm
+
+ENTRY(cpu_v6_proc_init)
+	mov	pc, lr
+
+ENTRY(cpu_v6_proc_fin)
+	mov	pc, lr
+
+/*
+ *	cpu_v6_reset(loc)
+ *
+ *	Perform a soft reset of the system.  Put the CPU into the
+ *	same state as it would be if it had been reset, and branch
+ *	to what would be the reset vector.
+ *
+ *	- loc   - location to jump to for soft reset
+ *
+ *	It is assumed that:
+ */
+	.align	5
+ENTRY(cpu_v6_reset)
+	mov	pc, r0
+
+/*
+ *	cpu_v6_do_idle()
+ *
+ *	Idle the processor (eg, wait for interrupt).
+ *
+ *	IRQs are already disabled.
+ */
+ENTRY(cpu_v6_do_idle)
+	mcr	p15, 0, r1, c7, c0, 4		@ wait for interrupt
+	mov	pc, lr
+
+ENTRY(cpu_v6_dcache_clean_area)
+#ifndef TLB_CAN_READ_FROM_L1_CACHE
+1:	mcr	p15, 0, r0, c7, c10, 1		@ clean D entry
+	add	r0, r0, #D_CACHE_LINE_SIZE
+	subs	r1, r1, #D_CACHE_LINE_SIZE
+	bhi	1b
+#endif
+	mov	pc, lr
+
+/*
+ *	cpu_arm926_switch_mm(pgd_phys, tsk)
+ *
+ *	Set the translation table base pointer to be pgd_phys
+ *
+ *	- pgd_phys - physical address of new TTB
+ *
+ *	It is assumed that:
+ *	- we are not using split page tables
+ */
+ENTRY(cpu_v6_switch_mm)
+	mov	r2, #0
+	ldr	r1, [r1, #MM_CONTEXT_ID]	@ get mm->context.id
+	mcr	p15, 0, r2, c7, c10, 4		@ drain write buffer
+	mcr	p15, 0, r0, c2, c0, 0		@ set TTB 0
+	mcr	p15, 0, r1, c13, c0, 1		@ set context ID
+	mov	pc, lr
+
+#define nG	(1 << 11)
+#define APX	(1 << 9)
+#define AP1	(1 << 5)
+#define AP0	(1 << 4)
+#define XN	(1 << 0)
+
+/*
+ *	cpu_v6_set_pte(ptep, pte)
+ *
+ *	Set a level 2 translation table entry.
+ *
+ *	- ptep  - pointer to level 2 translation table entry
+ *		  (hardware version is stored at -1024 bytes)
+ *	- pte   - PTE value to store
+ *
+ *	Permissions:
+ *	  YUWD  APX AP1 AP0	SVC	User
+ *	  0xxx   0   0   0	no acc	no acc
+ *	  100x   1   0   1	r/o	no acc
+ *	  10x0   1   0   1	r/o	no acc
+ *	  1011   0   0   1	r/w	no acc
+ *	  110x   1   1   0	r/o	r/o
+ *	  11x0   1   1   0	r/o	r/o
+ *	  1111   0   1   1	r/w	r/w
+ */
+ENTRY(cpu_v6_set_pte)
+	str	r1, [r0], #-2048		@ linux version
+
+	bic	r2, r1, #0x00000ff0
+	bic	r2, r2, #0x00000003
+	orr	r2, r2, #AP0 | 2
+
+	tst	r1, #L_PTE_WRITE
+	tstne	r1, #L_PTE_DIRTY
+	orreq	r2, r2, #APX
+
+	tst	r1, #L_PTE_USER
+	orrne	r2, r2, #AP1 | nG
+	tstne	r2, #APX
+	eorne	r2, r2, #AP0
+
+	tst	r1, #L_PTE_YOUNG
+	biceq	r2, r2, #APX | AP1 | AP0
+
+@	tst	r1, #L_PTE_EXEC
+@	orreq	r2, r2, #XN
+
+	tst	r1, #L_PTE_PRESENT
+	moveq	r2, #0
+
+	str	r2, [r0]
+	mcr	p15, 0, r0, c7, c10, 1 @ flush_pte
+	mov	pc, lr
+
+
+
+
+cpu_v6_name:
+	.asciz	"Some Random V6 Processor"
+	.align
+
+	.section ".text.init", #alloc, #execinstr
+
+/*
+ *	__v6_setup
+ *
+ *	Initialise TLB, Caches, and MMU state ready to switch the MMU
+ *	on.  Return in r0 the new CP15 C1 control register setting.
+ *
+ *	We automatically detect if we have a Harvard cache, and use the
+ *	Harvard cache control instructions insead of the unified cache
+ *	control instructions.
+ *
+ *	This should be able to cover all ARMv6 cores.
+ *
+ *	It is assumed that:
+ *	- cache type register is implemented
+ */
+__v6_setup:
+	mrc	p15, 0, r10, c0, c0, 1		@ read cache type register
+	tst	r10, #1 << 24			@ Harvard cache?
+	mov	r10, #0
+	mcrne	p15, 0, r10, c7, c14, 0		@ clean+invalidate D cache
+	mcrne	p15, 0, r10, c7, c5, 0		@ invalidate I cache
+	mcreq	p15, 0, r10, c7, c15, 0		@ clean+invalidate cache
+	mcr	p15, 0, r10, c7, c10, 4		@ drain write buffer
+	mcr	p15, 0, r10, c8, c7, 0		@ invalidate I + D TLBs
+	mcr	p15, 0, r10, c2, c0, 2		@ TTB control register
+	mcr	p15, 0, r4, c2, c0, 0		@ load TTB0
+	mcr	p15, 0, r4, c2, c0, 1		@ load TTB1
+	mov	r10, #0x1f			@ domains 0, 1 = manager
+	mcr	p15, 0, r10, c3, c0, 0		@ load domain access register
+	mrc	p15, 0, r0, c1, c0, 0		@ read control register
+	ldr	r10, cr1_clear			@ get mask for bits to clear
+	bic	r0, r0, r10			@ clear bits them
+	ldr	r10, cr1_set			@ get mask for bits to set
+	orr	r0, r0, r10			@ set them
+	mov	pc, lr				@ return to head.S:__ret
+
+	/*
+	 *         V X F   I D LR
+	 * .... ...E PUI. .T.T 4RVI ZFRS BLDP WCAM
+	 * rrrr rrrx xxx0 0101 xxxx xxxx x111 xxxx < forced
+	 *         0 110       0011 1.00 .111 1101 < we want
+	 */
+	.type	cr1_clear, #object
+	.type	cr1_set, #object
+cr1_clear:
+	.word	0x0120c302
+cr1_set:
+	.word	0x00c0387d
+
+	.type	v6_processor_functions, #object
+ENTRY(v6_processor_functions)
+	.word	v6_early_abort
+	.word	cpu_v6_proc_init
+	.word	cpu_v6_proc_fin
+	.word	cpu_v6_reset
+	.word	cpu_v6_do_idle
+	.word	cpu_v6_dcache_clean_area
+	.word	cpu_v6_switch_mm
+	.word	cpu_v6_set_pte
+	.size	v6_processor_functions, . - v6_processor_functions
+
+	.type	cpu_arch_name, #object
+cpu_arch_name:
+	.asciz	"armv6"
+	.size	cpu_arch_name, . - cpu_arch_name
+
+	.type	cpu_elf_name, #object
+cpu_elf_name:
+	.asciz	"v6"
+	.size	cpu_elf_name, . - cpu_elf_name
+	.align
+
+	.section ".proc.info", #alloc, #execinstr
+
+	/*
+	 * Match any ARMv6 processor core.
+	 */
+	.type	__v6_proc_info, #object
+__v6_proc_info:
+	.long	0x00070000
+	.long	0x00ff0000
+	.long	0x00000c0e
+	b	__v6_setup
+	.long	cpu_arch_name
+	.long	cpu_elf_name
+	.long	HWCAP_SWP | HWCAP_HALF | HWCAP_FAST_MULT | HWCAP_VFP
+	.long	cpu_v6_name
+	.long	v6_processor_functions
+	.long	v6wbi_tlb_fns
+	.long	v6_user_fns
+	.long	v6_cache_fns
+	.size	__v6_proc_info, . - __v6_proc_info

arch/arm/mm/tlb-v6.S

+/*
+ *  linux/arch/arm/mm/tlb-v6.S
+ *
+ *  Copyright (C) 1997-2002 Russell King
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2 as
+ * published by the Free Software Foundation.
+ *
+ *  ARM architecture version 6 TLB handling functions.
+ *  These assume a split I/D TLB.
+ */
+#include <linux/linkage.h>
+#include <asm/constants.h>
+#include <asm/page.h>
+#include <asm/tlbflush.h>
+#include "proc-macros.S"
+
+#define HARVARD_TLB
+
+/*
+ *	v6wbi_flush_user_tlb_range(start, end, vma)
+ *
+ *	Invalidate a range of TLB entries in the specified address space.
+ *
+ *	- start - start address (may not be aligned)
+ *	- end   - end address (exclusive, may not be aligned)
+ *	- vma   - vma_struct describing address range
+ *
+ *	It is assumed that:
+ *	- the "Invalidate single entry" instruction will invalidate
+ *	  both the I and the D TLBs on Harvard-style TLBs
+ */
+ENTRY(v6wbi_flush_user_tlb_range)
+	vma_vm_mm r3, r2			@ get vma->vm_mm
+	mov	ip, #0
+	mmid	r3, r3				@ get vm_mm->context.id
+	mcr	p15, 0, ip, c7, c10, 4		@ drain write buffer
+	mov	r0, r0, lsr #PAGE_SHIFT		@ align address
+	mov	r1, r1, lsr #PAGE_SHIFT
+	asid	r3, r3				@ mask ASID
+	orr	r0, r3, r0, lsl #PAGE_SHIFT	@ Create initial MVA
+	mov	r1, r1, lsl #PAGE_SHIFT
+	vma_vm_flags r2, r2			@ get vma->vm_flags
+1:
+#ifdef HARVARD_TLB
+	mcr	p15, 0, r0, c8, c6, 1		@ TLB invalidate D MVA (was 1)
+	tst	r2, #VM_EXEC			@ Executable area ?
+	mcrne	p15, 0, r0, c8, c5, 1		@ TLB invalidate I MVA (was 1)
+#else
+	mcr	p15, 0, r0, c8, c7, 1		@ TLB invalidate MVA (was 1)
+#endif
+	add	r0, r0, #PAGE_SZ
+	cmp	r0, r1
+	blo	1b
+	mov	pc, lr
+
+/*
+ *	v6wbi_flush_kern_tlb_range(start,end)
+ *
+ *	Invalidate a range of kernel TLB entries
+ *
+ *	- start - start address (may not be aligned)
+ *	- end   - end address (exclusive, may not be aligned)
+ */
+ENTRY(v6wbi_flush_kern_tlb_range)
+	mov	r2, #0
+	mcr	p15, 0, r2, c7, c10, 4		@ drain write buffer
+	mov	r0, r0, lsr #PAGE_SHIFT		@ align address
+	mov	r1, r1, lsr #PAGE_SHIFT
+	mov	r0, r0, lsl #PAGE_SHIFT
+	mov	r1, r1, lsl #PAGE_SHIFT
+1:
+#ifdef HARVARD_TLB
+	mcr	p15, 0, r0, c8, c6, 1		@ TLB invalidate D MVA
+	mcr	p15, 0, r0, c8, c5, 1		@ TLB invalidate I MVA
+#else
+	mcr	p15, 0, r0, c8, c7, 1		@ TLB invalidate MVA
+#endif
+	add	r0, r0, #PAGE_SZ
+	cmp	r0, r1
+	blo	1b
+	mov	pc, lr
+
+	.section ".text.init", #alloc, #execinstr
+
+	.type	v6wbi_tlb_fns, #object
+ENTRY(v6wbi_tlb_fns)
+	.long	v6wbi_flush_user_tlb_range
+	.long	v6wbi_flush_kern_tlb_range
+	.long	v6wbi_tlb_flags
+	.size	v6wbi_tlb_fns, . - v6wbi_tlb_fns

arch/arm/nwfpe/fpmodule.c

@@ -50,11 +50,6 @@
 
 #ifdef MODULE
 void fp_send_sig(unsigned long sig, struct task_struct *p, int priv);
-#if LINUX_VERSION_CODE > 0x20115
-MODULE_AUTHOR("Scott Bambrough <scottb@rebel.com>");
-MODULE_DESCRIPTION("NWFPE floating point emulator (" NWFPE_BITS " precision)");
-#endif
-
 #else
 #define fp_send_sig	send_sig
 #define kern_fp_enter	fp_enter
@@ -172,3 +167,7 @@ void float_raise(signed char flags)
 
 module_init(fpe_init);
 module_exit(fpe_exit);
+
+MODULE_AUTHOR("Scott Bambrough <scottb@rebel.com>");
+MODULE_DESCRIPTION("NWFPE floating point emulator (" NWFPE_BITS " precision)");
+MODULE_LICENSE("GPL");

arch/arm/tools/mach-types

@@ -6,7 +6,7 @@
 # To add an entry into this database, please see Documentation/arm/README,
 # or contact rmk@arm.linux.org.uk
 #
-# Last update: Thu Sep 18 17:15:55 2003
+# Last update: Tue Feb 24 17:17:50 2004
 #
 # machine_is_xxx	CONFIG_xxxx		MACH_TYPE_xxx		number
 #
@@ -202,7 +202,7 @@ karo			ARCH_KARO		KARO			190
 fester			SA1100_FESTER		FESTER			191
 gpi			ARCH_GPI		GPI			192
 smdk2410		ARCH_SMDK2410		SMDK2410		193
-premium			ARCH_PREMIUM		PREMIUM			194
+i519			ARCH_I519		I519			194
 nexio			SA1100_NEXIO		NEXIO			195
 bitbox			SA1100_BITBOX		BITBOX			196
 g200			SA1100_G200		G200			197
@@ -259,7 +259,7 @@ stork_nest		ARCH_STORK_NEST		STORK_NEST		247
 stork_egg		ARCH_STORK_EGG		STORK_EGG		248
 wismo			SA1100_WISMO		WISMO			249
 ezlinx			ARCH_EZLINX		EZLINX			250
-at91rm9200		ARCH_AT91		AT91			251
+at91rm9200		ARCH_AT91RM9200		AT91RM9200		251
 orion			ARCH_ORION		ORION			252
 neptune			ARCH_NEPTUNE		NEPTUNE			253
 hackkit			SA1100_HACKKIT		HACKKIT			254
@@ -295,7 +295,7 @@ viper			ARCH_VIPER		VIPER			283
 adsbitsyplus		SA1100_ADSBITSYPLUS	ADSBITSYPLUS		284
 adsagc			SA1100_ADSAGC		ADSAGC			285
 stp7312			ARCH_STP7312		STP7312			286
-nx_phnx			ARCH_PXA255		PXA255			287
+nx_phnx			MACH_NX_PHNX		NX_PHNX			287
 wep_ep250		ARCH_WEP_EP250		WEP_EP250		288
 inhandelf3		ARCH_INHANDELF3		INHANDELF3		289
 adi_coyote		ARCH_ADI_COYOTE		ADI_COYOTE		290
@@ -364,7 +364,7 @@ ixrd425			ARCH_IXRD425		IXRD425			352
 iq80315			ARCH_IQ80315		IQ80315			353
 nmp7312			ARCH_NMP7312		NMP7312			354
 cx861xx			ARCH_CX861XX		CX861XX			355
-ixp2000			ARCH_IXP2000		IXP2000			356
+enp2611			ARCH_ENP2611		ENP2611			356
 xda			SA1100_XDA		XDA			357
 csir_ims		ARCH_CSIR_IMS		CSIR_IMS		358
 ixp421_dnaeeth		ARCH_IXP421_DNAEETH	IXP421_DNAEETH		359
@@ -385,3 +385,92 @@ gumstik			ARCH_GUMSTIK		GUMSTIK			373
 rcube			ARCH_RCUBE		RCUBE			374
 rea_olv			ARCH_REA_OLV		REA_OLV			375
 pxa_iphone		ARCH_PXA_IPHONE		PXA_IPHONE		376
+s3c3410			ARCH_S3C3410		S3C3410			377
+espd_4510b		ARCH_ESPD_4510B		ESPD_4510B		378
+mp1x			ARCH_MP1X		MP1X			379
+at91rm9200tb		ARCH_AT91RM9200TB	AT91RM9200TB		380
+adsvgx			ARCH_ADSVGX		ADSVGX			381
+omap1610		ARCH_OMAP1610		OMAP1610		382
+pelee			ARCH_PELEE		PELEE			383
+e7xx			ARCH_E7XX		E7XX			384
+iq80331			ARCH_IQ80331		IQ80331			385
+versatile_pb		ARCH_VERSATILE_PB	VERSATILE_PB		387
+kev7a400		MACH_KEV7A400		KEV7A400		388
+lpd7a400		MACH_LPD7A400		LPD7A400		389
+lpd7a404		MACH_LPD7A404		LPD7A404		390
+fujitsu_camelot		ARCH_FUJITSU_CAMELOT	FUJITSU_CAMELOT		391
+janus2m			ARCH_JANUS2M		JANUS2M			392
+embtf			MACH_EMBTF		EMBTF			393
+hpm			MACH_HPM		HPM			394
+smdk2410tk		MACH_SMDK2410TK		SMDK2410TK		395
+smdk2410aj		MACH_SMDK2410AJ		SMDK2410AJ		396
+streetracer		MACH_STREETRACER	STREETRACER		397
+eframe			MACH_EFRAME		EFRAME			398
+csb337			MACH_CSB337		CSB337			399
+pxa_lark		MACH_PXA_LARK		PXA_LARK		400
+pxa_pnp2110		MACH_PNP2110		PNP2110			401
+tcc72x			MACH_TCC72X		TCC72X			402
+altair			MACH_ALTAIR		ALTAIR			403
+kc3			MACH_KC3		KC3			404
+sinteftd		MACH_SINTEFTD		SINTEFTD		405
+mainstone		MACH_MAINSTONE		MAINSTONE		406
+aday4x			MACH_ADAY4X		ADAY4X			407
+lite300			MACH_LITE300		LITE300			408
+s5c7376			MACH_S5C7376		S5C7376			409
+mt02			MACH_MT02		MT02			410
+mport3s			MACH_MPORT3S		MPORT3S			411
+ra_alpha		MACH_RA_ALPHA		RA_ALPHA		412
+xcep			MACH_XCEP		XCEP			413
+arcom_mercury		MACH_ARCOM_MERCURY	ARCOM_MERCURY		414
+stargate		MACH_STARGATE		STARGATE		415
+armadilloj		MACH_ARMADILLOJ		ARMADILLOJ		416
+elroy_jack		MACH_ELROY_JACK		ELROY_JACK		417
+backend			MACH_BACKEND		BACKEND			418
+s5linbox		MACH_S5LINBOX		S5LINBOX		419
+nomadik			MACH_NOMADIK		NOMADIK			420
+ia_cpu_9200		MACH_IA_CPU_9200	IA_CPU_9200		421
+at91_bja1		MACH_AT91_BJA1		AT91_BJA1		422
+corgi			MACH_CORGI		CORGI			423
+poodle			MACH_POODLE		POODLE			424
+ten			MACH_TEN		TEN			425
+roverp5p		MACH_ROVERP5P		ROVERP5P		426
+sc2700			MACH_SC2700		SC2700			427
+ex_eagle		MACH_EX_EAGLE		EX_EAGLE		428
+nx_pxa12		MACH_NX_PXA12		NX_PXA12		429
+nx_pxa5			MACH_NX_PXA5		NX_PXA5			430
+blackboard2		MACH_BLACKBOARD2	BLACKBOARD2		431
+i819			MACH_I819		I819			432
+ixmb995e		MACH_IXMB995E		IXMB995E		433
+skyrider		MACH_SKYRIDER		SKYRIDER		434
+skyhawk			MACH_SKYHAWK		SKYHAWK			435
+enterprise		MACH_ENTERPRISE		ENTERPRISE		436
+dep2410			MACH_DEP2410		DEP2410			437
+armcore			MACH_ARMCORE		ARMCORE			438
+hobbit			MACH_HOBBIT		HOBBIT			439
+h7210			MACH_H7210		H7210			440
+pxa_netdcu5		MACH_PXA_NETDCU5	PXA_NETDCU5		441
+acc			MACH_ACC		ACC			442
+esl_sarva		MACH_ESL_SARVA		ESL_SARVA		443
+xm250			MACH_XM250		XM250			444
+t6tc1xb			MACH_T6TC1XB		T6TC1XB			445
+ess710			MACH_ESS710		ESS710			446
+mx3ads			MACH_MX3ADS		MX3ADS			447
+himalaya		MACH_HIMALAYA		HIMALAYA		448
+bolfenk			MACH_BOLFENK		BOLFENK			449
+at91rm9200kr		MACH_AT91RM9200KR	AT91RM9200KR		450
+edb9312			MACH_EDB9312		EDB9312			451
+omap_generic		MACH_OMAP_GENERIC	OMAP_GENERIC		452
+aximx3			MACH_AXIMX3		AXIMX3			453
+eb67xdip		MACH_EB67XDIP		EB67XDIP		454
+webtxs			MACH_WEBTXS		WEBTXS			455
+hawk			MACH_HAWK		HAWK			456
+ccat91sbc001		MACH_CCAT91SBC001	CCAT91SBC001		457
+expresso		MACH_EXPRESSO		EXPRESSO		458
+h4000			MACH_H4000		H4000			459
+dino			MACH_DINO		DINO			460
+ml675k			MACH_ML675K		ML675K			461
+edb9301			MACH_EDB9301		EDB9301			462
+edb9315			MACH_EDB9315		EDB9315			463
+reciva_tt		MACH_RECIVA_TT		RECIVA_TT		464
+cstcb01			MACH_CSTCB01		CSTCB01			465
+cstcb1			MACH_CSTCB1		CSTCB1			466

arch/ppc64/Kconfig

@@ -89,6 +89,17 @@ config PPC_PMAC
 	bool "Apple PowerMac G5 support"
 	select ADB_PMU
 
+config PMAC_DART
+	bool "Enable DART/IOMMU on PowerMac (allow >2G of RAM)"
+	depends on PPC_PMAC
+	depends on EXPERIMENTAL
+	default n
+	help
+	  Enabling DART makes it possible to boot a PowerMac G5 with more
+	  than 2GB of memory. Note that the code is very new and untested
+	  at this time, so it has to be considered experimental. Enabling
+	  this might result in data loss.
+
 config PPC_PMAC64
 	bool
 	depends on PPC_PMAC
@@ -109,6 +120,18 @@ config POWER4_ONLY
 	  binary will not work on POWER3 or RS64 processors when compiled with
 	  binutils 2.15 or later.
 
+config IOMMU_VMERGE
+	bool "Enable IOMMU virtual merging (EXPERIMENTAL)"
+	depends on EXPERIMENTAL
+	default n
+	help
+	  Cause IO segments sent to a device for DMA to be merged virtually
+	  by the IOMMU when they happen to have been allocated contiguously.
+	  This doesn't add pressure to the IOMMU allocator. However, some
+	  drivers don't support getting large merged segments coming back
+	  from *_map_sg(). Say Y if you know the drivers you are using are
+	  properly handling this case.
+
 config SMP
 	bool "Symmetric multi-processing support"
 	---help---
@@ -287,9 +310,6 @@ config VIOTAPE
 	  If you are running Linux on an iSeries system and you want Linux
 	  to read and/or write a tape drive owned by OS/400, say Y here.
 
-config VETH
-	tristate "iSeries Virtual Ethernet driver support"
-
 endmenu
 
 config VIOPATH

arch/ppc64/kernel/Makefile

@@ -10,11 +10,12 @@ obj-y               :=	setup.o entry.o traps.o irq.o idle.o \
 			align.o semaphore.o bitops.o stab.o pacaData.o \
 			udbg.o binfmt_elf32.o sys_ppc32.o ioctl32.o \
 			ptrace32.o signal32.o pmc.o rtc.o init_task.o \
-			lmb.o cputable.o cpu_setup_power4.o idle_power4.o
+			lmb.o cputable.o cpu_setup_power4.o idle_power4.o \
+			iommu.o
 
 obj-$(CONFIG_PPC_OF) +=	of_device.o
 
-obj-$(CONFIG_PCI)	+= pci.o pci_dn.o pci_dma.o
+obj-$(CONFIG_PCI)	+= pci.o pci_dn.o pci_iommu.o
 
 ifdef CONFIG_PPC_ISERIES
 obj-$(CONFIG_PCI)	+= iSeries_pci.o iSeries_pci_reset.o \
@@ -28,12 +29,12 @@ obj-$(CONFIG_PPC_ISERIES) += iSeries_irq.o \
 			     HvCall.o HvLpConfig.o LparData.o mf_proc.o \
 			     iSeries_setup.o ItLpQueue.o hvCall.o \
 			     mf.o HvLpEvent.o iSeries_proc.o iSeries_htab.o \
-			     proc_pmc.o
+			     proc_pmc.o iSeries_iommu.o
 
 obj-$(CONFIG_PPC_PSERIES) += pSeries_pci.o pSeries_lpar.o pSeries_hvCall.o \
 			     eeh.o nvram.o pSeries_nvram.o rtasd.o ras.o \
 			     open_pic.o xics.o pSeries_htab.o rtas.o \
-			     chrp_setup.o i8259.o prom.o vio.o
+			     chrp_setup.o i8259.o prom.o vio.o pSeries_iommu.o
 
 obj-$(CONFIG_PROC_FS)		+= proc_ppc64.o
 obj-$(CONFIG_RTAS_FLASH)	+= rtas_flash.o
@@ -49,6 +50,8 @@ obj-$(CONFIG_BOOTX_TEXT)	+= btext.o
 obj-$(CONFIG_PPC_PMAC)		+= pmac_setup.o pmac_feature.o pmac_pci.o \
 				   pmac_time.o pmac_nvram.o pmac_low_i2c.o \
 				   open_pic_u3.o
+obj-$(CONFIG_PMAC_DART)		+= pmac_iommu.o
+
 ifdef CONFIG_SMP
 obj-$(CONFIG_PPC_PMAC)		+= pmac_smp.o smp-tbsync.o
 endif

arch/ppc64/kernel/chrp_setup.c

@@ -51,7 +51,7 @@
 #include <asm/prom.h>
 #include <asm/rtas.h>
 #include <asm/pci-bridge.h>
-#include <asm/pci_dma.h>
+#include <asm/iommu.h>
 #include <asm/dma.h>
 #include <asm/machdep.h>
 #include <asm/irq.h>

arch/ppc64/kernel/head.S

@@ -826,7 +826,14 @@ SystemCall_common:
 _GLOBAL(do_hash_page_ISI)
 	li	r4,0
 _GLOBAL(do_hash_page_DSI)
-	rlwimi	r4,r23,32-13,30,30	/* Insert MSR_PR as _PAGE_USER */
+	/*
+	 * We need to set the _PAGE_USER bit if MSR_PR is set or if we are
+	 * accessing a userspace segment (even from the kernel). We assume
+	 * kernel addresses always have the high bit set.
+	 */
+	rotldi	r0,r3,15		/* Move high bit into MSR_PR position */
+	orc	r0,r23,r0
+	rlwimi	r4,r0,32-13,30,30	/* Insert into _PAGE_USER */
 	ori	r4,r4,1			/* add _PAGE_PRESENT */
 
 	mflr	r21			/* Save LR in r21 */

arch/ppc64/kernel/iSeries_iommu.c

+/*
+ * arch/ppc64/kernel/iSeries_iommu.c
+ *
+ * Copyright (C) 2001 Mike Corrigan & Dave Engebretsen, IBM Corporation
+ * 
+ * Rewrite, cleanup: 
+ *
+ * Copyright (C) 2004 Olof Johansson <olof@austin.ibm.com>, IBM Corporation
+ *
+ * Dynamic DMA mapping support, iSeries-specific parts.
+ *
+ * 
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation; either version 2 of the License, or
+ * (at your option) any later version.
+ * 
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ * 
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA  02111-1307 USA
+ */
+
+#include <linux/config.h>
+#include <linux/init.h>
+#include <linux/types.h>
+#include <linux/slab.h>
+#include <linux/mm.h>
+#include <linux/spinlock.h>
+#include <linux/string.h>
+#include <linux/pci.h>
+#include <asm/io.h>
+#include <asm/prom.h>
+#include <asm/rtas.h>
+#include <asm/ppcdebug.h>
+
+#include <asm/iSeries/HvCallXm.h>
+#include <asm/iSeries/LparData.h>
+#include <asm/iommu.h>
+#include <asm/pci-bridge.h>
+#include <asm/iSeries/iSeries_pci.h>
+
+#include <asm/machdep.h>
+
+#include "pci.h"
+
+
+static struct iommu_table veth_iommu_table;	/* Tce table for virtual ethernet */
+static struct iommu_table vio_iommu_table;	/* Tce table for virtual I/O */
+
+static struct iSeries_Device_Node veth_dev_node = { .LogicalSlot = 0xFF, .iommu_table = &veth_iommu_table };
+static struct iSeries_Device_Node vio_dev_node  = { .LogicalSlot = 0xFF, .iommu_table = &vio_iommu_table };
+
+static struct pci_dev _veth_dev = { .sysdata = &veth_dev_node };
+static struct pci_dev _vio_dev  = { .sysdata = &vio_dev_node, .dev.bus = &pci_bus_type  };
+
+/*
+ * I wonder what the deal is with these.  Nobody uses them.  Why do they
+ * exist? Why do we export them to modules? Why is this comment here, and
+ * why didn't I just delete them?
+ */
+struct pci_dev *iSeries_veth_dev = &_veth_dev;
+struct device *iSeries_vio_dev = &_vio_dev.dev;
+
+EXPORT_SYMBOL(iSeries_veth_dev);
+EXPORT_SYMBOL(iSeries_vio_dev);
+
+extern struct list_head iSeries_Global_Device_List;
+
+
+static void tce_build_iSeries(struct iommu_table *tbl, long index, long npages,
+			      unsigned long uaddr, int direction)
+{
+	u64 rc;
+	union tce_entry tce;
+	
+	while (npages--) {
+		tce.te_word = 0;
+		tce.te_bits.tb_rpn = (virt_to_absolute(uaddr)) >> PAGE_SHIFT;
+
+		if (tbl->it_type == TCE_VB) {
+			/* Virtual Bus */
+			tce.te_bits.tb_valid = 1;
+			tce.te_bits.tb_allio = 1;
+			if (direction != PCI_DMA_TODEVICE)
+				tce.te_bits.tb_rdwr = 1;
+		} else {
+			/* PCI Bus */
+			tce.te_bits.tb_rdwr = 1; /* Read allowed */
+			if (direction != PCI_DMA_TODEVICE)
+				tce.te_bits.tb_pciwr = 1;
+		}
+		
+		rc = HvCallXm_setTce((u64)tbl->it_index, 
+				     (u64)index, 
+				     tce.te_word);
+		if (rc)
+			panic("PCI_DMA: HvCallXm_setTce failed, Rc: 0x%lx\n", rc);
+
+		index++;
+		uaddr += PAGE_SIZE;
+	}
+}
+
+static void tce_free_iSeries(struct iommu_table *tbl, long index, long npages)
+{
+	u64 rc;
+	union tce_entry tce;
+
+	while (npages--) {
+		tce.te_word = 0;
+		rc = HvCallXm_setTce((u64)tbl->it_index,
+				     (u64)index,
+				     tce.te_word);
+
+		if (rc) 
+			panic("PCI_DMA: HvCallXm_setTce failed, Rc: 0x%lx\n", rc);
+
+		index++;
+	}
+
+}
+
+void __init iommu_vio_init(void)
+{
+	struct iommu_table *t;
+	struct iommu_table_cb cb;
+	unsigned long cbp;
+
+	cb.itc_busno = 255;    /* Bus 255 is the virtual bus */
+	cb.itc_virtbus = 0xff; /* Ask for virtual bus */
+	
+	cbp = virt_to_absolute((unsigned long)&cb);
+	HvCallXm_getTceTableParms(cbp);
+	
+	veth_iommu_table.it_size        = cb.itc_size / 2;
+	veth_iommu_table.it_busno       = cb.itc_busno;
+	veth_iommu_table.it_offset      = cb.itc_offset;
+	veth_iommu_table.it_index       = cb.itc_index;
+	veth_iommu_table.it_type        = TCE_VB;
+	veth_iommu_table.it_entrysize	= sizeof(union tce_entry);
+	veth_iommu_table.it_blocksize	= 1;
+
+	t = iommu_init_table(&veth_iommu_table);
+
+	if (!t)
+		printk("Virtual Bus VETH TCE table failed.\n");
+
+	vio_iommu_table.it_size         = cb.itc_size - veth_iommu_table.it_size;
+	vio_iommu_table.it_busno        = cb.itc_busno;
+	vio_iommu_table.it_offset       = cb.itc_offset +
+		veth_iommu_table.it_size * (PAGE_SIZE/sizeof(union tce_entry));
+	vio_iommu_table.it_index        = cb.itc_index;
+	vio_iommu_table.it_type         = TCE_VB; 
+	vio_iommu_table.it_entrysize	= sizeof(union tce_entry);
+	vio_iommu_table.it_blocksize	= 1;
+
+	t = iommu_init_table(&vio_iommu_table);
+
+	if (!t) 
+		printk("Virtual Bus VIO TCE table failed.\n");
+}
+
+
+/*
+ * This function compares the known tables to find an iommu_table
+ * that has already been built for hardware TCEs.                          
+ */
+static struct iommu_table *iommu_table_find(struct iommu_table * tbl)
+{
+	struct iSeries_Device_Node *dp;
+
+	for (dp =  (struct iSeries_Device_Node *)iSeries_Global_Device_List.next;
+	     dp != (struct iSeries_Device_Node *)&iSeries_Global_Device_List;
+	     dp =  (struct iSeries_Device_Node *)dp->Device_List.next) 
+		if (dp->iommu_table                 != NULL &&
+		    dp->iommu_table->it_type        == TCE_PCI &&
+		    dp->iommu_table->it_offset      == tbl->it_offset &&
+		    dp->iommu_table->it_index       == tbl->it_index &&
+		    dp->iommu_table->it_size        == tbl->it_size) 
+			return dp->iommu_table;
+			
+
+	return NULL;
+}
+
+/*
+ * Call Hv with the architected data structure to get TCE table info.
+ * info. Put the returned data into the Linux representation of the   
+ * TCE table data.                                                     
+ * The Hardware Tce table comes in three flavors.                     
+ * 1. TCE table shared between Buses.                                  
+ * 2. TCE table per Bus.                                               
+ * 3. TCE Table per IOA.                                               
+ */
+static void iommu_table_getparms(struct iSeries_Device_Node* dn,
+				 struct iommu_table* tbl)
+{
+	struct iommu_table_cb *parms;
+
+	parms = (struct iommu_table_cb*)kmalloc(sizeof(*parms), GFP_KERNEL);
+
+	if (parms == NULL) 
+		panic("PCI_DMA: TCE Table Allocation failed.");
+
+	memset(parms, 0, sizeof(*parms));
+
+	parms->itc_busno   = ISERIES_BUS(dn);
+	parms->itc_slotno  = dn->LogicalSlot;
+	parms->itc_virtbus = 0;
+
+	HvCallXm_getTceTableParms(REALADDR(parms));
+
+	if (parms->itc_size == 0)
+		panic("PCI_DMA: parms->size is zero, parms is 0x%p", parms);
+
+	tbl->it_size        = parms->itc_size;
+	tbl->it_busno       = parms->itc_busno;
+	tbl->it_offset      = parms->itc_offset;
+	tbl->it_index       = parms->itc_index;
+	tbl->it_entrysize   = sizeof(union tce_entry);
+	tbl->it_blocksize   = 1;
+	tbl->it_type        = TCE_PCI;
+
+	kfree(parms);
+}
+
+
+void iommu_devnode_init(struct iSeries_Device_Node *dn) {
+	struct iommu_table *tbl;
+
+	tbl = (struct iommu_table *)kmalloc(sizeof(struct iommu_table), GFP_KERNEL);
+
+	iommu_table_getparms(dn, tbl);
+
+	/* Look for existing tce table */
+	dn->iommu_table = iommu_table_find(tbl);
+
+	if (dn->iommu_table == NULL)
+		dn->iommu_table = iommu_init_table(tbl);
+	else
+		kfree(tbl);
+
+	return;
+}
+
+
+void tce_init_iSeries(void)
+{
+	ppc_md.tce_build = tce_build_iSeries;
+	ppc_md.tce_free  = tce_free_iSeries;
+
+	pci_iommu_init();
+}

arch/ppc64/kernel/iSeries_pci.c

@@ -36,7 +36,7 @@
 #include <asm/pci-bridge.h>
 #include <asm/ppcdebug.h>
 #include <asm/naca.h>
-#include <asm/pci_dma.h>
+#include <asm/iommu.h>
 
 #include <asm/iSeries/HvCallPci.h>
 #include <asm/iSeries/HvCallSm.h>
@@ -53,7 +53,7 @@ extern int panic_timeout;
 
 extern unsigned long iSeries_Base_Io_Memory;    
 
-extern struct TceTable *tceTables[256];
+extern struct iommu_table *tceTables[256];
 
 extern void iSeries_MmIoTest(void);
 
@@ -273,7 +273,7 @@ void __init iSeries_pci_final_fixup(void)
 			iSeries_Device_Information(pdev, Buffer,
 					sizeof(Buffer));
 			printk("%d. %s\n", DeviceCount, Buffer);
-			create_pci_bus_tce_table((unsigned long)node);
+			iommu_devnode_init(node);
 		} else
 			printk("PCI: Device Tree not found for 0x%016lX\n",
 					(unsigned long)pdev);

arch/ppc64/kernel/lmb.c

@@ -269,11 +269,13 @@ lmb_phys_mem_size(void)
 	return _lmb->memory.size;
 #else
 	struct lmb_region *_mem = &(_lmb->memory);
-	unsigned long idx = _mem->cnt-1;
-	unsigned long lastbase = _mem->region[idx].physbase;
-	unsigned long lastsize = _mem->region[idx].size;
-	
-	return (lastbase + lastsize);
+	unsigned long total = 0;
+	int i;
+
+	/* add all physical memory to the bootmem map */
+	for (i=0; i < _mem->cnt; i++)
+		total += _mem->region[i].size;
+	return total;
 #endif /* CONFIG_MSCHUNKS */
 }
 
@@ -283,15 +285,13 @@ lmb_end_of_DRAM(void)
 	unsigned long offset = reloc_offset();
 	struct lmb *_lmb = PTRRELOC(&lmb);
 	struct lmb_region *_mem = &(_lmb->memory);
-	unsigned long idx;
+	int idx = _mem->cnt - 1;
 
-	for(idx=_mem->cnt-1; idx >= 0; idx--) {
 #ifdef CONFIG_MSCHUNKS
-		return (_mem->region[idx].physbase + _mem->region[idx].size);
+	return (_mem->region[idx].physbase + _mem->region[idx].size);
 #else
-		return (_mem->region[idx].base + _mem->region[idx].size);
+	return (_mem->region[idx].base + _mem->region[idx].size);
 #endif /* CONFIG_MSCHUNKS */
-	}
 
 	return 0;
 }

arch/ppc64/kernel/misc.S

@@ -209,6 +209,48 @@ _GLOBAL(flush_dcache_range)
 	sync
 	blr
 
+/*
+ * Like above, but works on non-mapped physical addresses.
+ * Use only for non-LPAR setups ! It also assumes real mode
+ * is cacheable. Used for flushing out the DART before using
+ * it as uncacheable memory 
+ *
+ * flush_dcache_phys_range(unsigned long start, unsigned long stop)
+ *
+ *    flush all bytes from start to stop-1 inclusive
+ */
+_GLOBAL(flush_dcache_phys_range)
+	LOADADDR(r10,naca)		/* Get Naca address */
+	ld	r10,0(r10)
+	LOADADDR(r11,systemcfg)		/* Get systemcfg address */
+	ld	r11,0(r11)
+	lwz	r7,DCACHEL1LINESIZE(r11)	/* Get dcache line size */
+	addi	r5,r7,-1
+	andc	r6,r3,r5		/* round low to line bdy */
+	subf	r8,r6,r4		/* compute length */
+	add	r8,r8,r5		/* ensure we get enough */
+	lwz	r9,DCACHEL1LOGLINESIZE(r10)	/* Get log-2 of dcache line size */
+	srw.	r8,r8,r9		/* compute line count */
+	beqlr				/* nothing to do? */
+	mfmsr	r5			/* Disable MMU Data Relocation */
+	ori	r0,r5,MSR_DR
+	xori	r0,r0,MSR_DR
+	sync
+	mtmsr	r0
+	sync
+	isync
+	mtctr	r8
+0:	dcbst	0,r6
+	add	r6,r6,r7
+	bdnz	0b
+	sync
+	isync
+	mtmsr	r5			/* Re-enable MMU Data Relocation */
+	sync
+	isync
+	blr
+
+
 /*
  * Flush a particular page from the data cache to RAM.
  * Note: this is necessary because the instruction cache does *not*

arch/ppc64/kernel/pSeries_htab.c

@@ -300,7 +300,7 @@ static void pSeries_flush_hash_range(unsigned long context,
 	int i, j;
 	HPTE *hptep;
 	Hpte_dword0 dw0;
-	struct ppc64_tlb_batch *batch = &ppc64_tlb_batch[smp_processor_id()];
+	struct ppc64_tlb_batch *batch = &__get_cpu_var(ppc64_tlb_batch);
 
 	/* XXX fix for large ptes */
 	unsigned long large = 0;