Merge branch 'devel-stable' into for-next

Conflicts:
	arch/arm/kernel/entry-armv.S

Documentation/arm/kernel_user_helpers.txt

@@ -0,0 +1,267 @@
+Kernel-provided User Helpers
+============================
+
+These are segment of kernel provided user code reachable from user space
+at a fixed address in kernel memory.  This is used to provide user space
+with some operations which require kernel help because of unimplemented
+native feature and/or instructions in many ARM CPUs. The idea is for this
+code to be executed directly in user mode for best efficiency but which is
+too intimate with the kernel counter part to be left to user libraries.
+In fact this code might even differ from one CPU to another depending on
+the available instruction set, or whether it is a SMP systems. In other
+words, the kernel reserves the right to change this code as needed without
+warning. Only the entry points and their results as documented here are
+guaranteed to be stable.
+
+This is different from (but doesn't preclude) a full blown VDSO
+implementation, however a VDSO would prevent some assembly tricks with
+constants that allows for efficient branching to those code segments. And
+since those code segments only use a few cycles before returning to user
+code, the overhead of a VDSO indirect far call would add a measurable
+overhead to such minimalistic operations.
+
+User space is expected to bypass those helpers and implement those things
+inline (either in the code emitted directly by the compiler, or part of
+the implementation of a library call) when optimizing for a recent enough
+processor that has the necessary native support, but only if resulting
+binaries are already to be incompatible with earlier ARM processors due to
+useage of similar native instructions for other things.  In other words
+don't make binaries unable to run on earlier processors just for the sake
+of not using these kernel helpers if your compiled code is not going to
+use new instructions for other purpose.
+
+New helpers may be added over time, so an older kernel may be missing some
+helpers present in a newer kernel.  For this reason, programs must check
+the value of __kuser_helper_version (see below) before assuming that it is
+safe to call any particular helper.  This check should ideally be
+performed only once at process startup time, and execution aborted early
+if the required helpers are not provided by the kernel version that
+process is running on.
+
+kuser_helper_version
+--------------------
+
+Location:	0xffff0ffc
+
+Reference declaration:
+
+  extern int32_t __kuser_helper_version;
+
+Definition:
+
+  This field contains the number of helpers being implemented by the
+  running kernel.  User space may read this to determine the availability
+  of a particular helper.
+
+Usage example:
+
+#define __kuser_helper_version (*(int32_t *)0xffff0ffc)
+
+void check_kuser_version(void)
+{
+	if (__kuser_helper_version < 2) {
+		fprintf(stderr, "can't do atomic operations, kernel too old\n");
+		abort();
+	}
+}
+
+Notes:
+
+  User space may assume that the value of this field never changes
+  during the lifetime of any single process.  This means that this
+  field can be read once during the initialisation of a library or
+  startup phase of a program.
+
+kuser_get_tls
+-------------
+
+Location:	0xffff0fe0
+
+Reference prototype:
+
+  void * __kuser_get_tls(void);
+
+Input:
+
+  lr = return address
+
+Output:
+
+  r0 = TLS value
+
+Clobbered registers:
+
+  none
+
+Definition:
+
+  Get the TLS value as previously set via the __ARM_NR_set_tls syscall.
+
+Usage example:
+
+typedef void * (__kuser_get_tls_t)(void);
+#define __kuser_get_tls (*(__kuser_get_tls_t *)0xffff0fe0)
+
+void foo()
+{
+	void *tls = __kuser_get_tls();
+	printf("TLS = %p\n", tls);
+}
+
+Notes:
+
+  - Valid only if __kuser_helper_version >= 1 (from kernel version 2.6.12).
+
+kuser_cmpxchg
+-------------
+
+Location:	0xffff0fc0
+
+Reference prototype:
+
+  int __kuser_cmpxchg(int32_t oldval, int32_t newval, volatile int32_t *ptr);
+
+Input:
+
+  r0 = oldval
+  r1 = newval
+  r2 = ptr
+  lr = return address
+
+Output:
+
+  r0 = success code (zero or non-zero)
+  C flag = set if r0 == 0, clear if r0 != 0
+
+Clobbered registers:
+
+  r3, ip, flags
+
+Definition:
+
+  Atomically store newval in *ptr only if *ptr is equal to oldval.
+  Return zero if *ptr was changed or non-zero if no exchange happened.
+  The C flag is also set if *ptr was changed to allow for assembly
+  optimization in the calling code.
+
+Usage example:
+
+typedef int (__kuser_cmpxchg_t)(int oldval, int newval, volatile int *ptr);
+#define __kuser_cmpxchg (*(__kuser_cmpxchg_t *)0xffff0fc0)
+
+int atomic_add(volatile int *ptr, int val)
+{
+	int old, new;
+
+	do {
+		old = *ptr;
+		new = old + val;
+	} while(__kuser_cmpxchg(old, new, ptr));
+
+	return new;
+}
+
+Notes:
+
+  - This routine already includes memory barriers as needed.
+
+  - Valid only if __kuser_helper_version >= 2 (from kernel version 2.6.12).
+
+kuser_memory_barrier
+--------------------
+
+Location:	0xffff0fa0
+
+Reference prototype:
+
+  void __kuser_memory_barrier(void);
+
+Input:
+
+  lr = return address
+
+Output:
+
+  none
+
+Clobbered registers:
+
+  none
+
+Definition:
+
+  Apply any needed memory barrier to preserve consistency with data modified
+  manually and __kuser_cmpxchg usage.
+
+Usage example:
+
+typedef void (__kuser_dmb_t)(void);
+#define __kuser_dmb (*(__kuser_dmb_t *)0xffff0fa0)
+
+Notes:
+
+  - Valid only if __kuser_helper_version >= 3 (from kernel version 2.6.15).
+
+kuser_cmpxchg64
+---------------
+
+Location:	0xffff0f60
+
+Reference prototype:
+
+  int __kuser_cmpxchg64(const int64_t *oldval,
+                        const int64_t *newval,
+                        volatile int64_t *ptr);
+
+Input:
+
+  r0 = pointer to oldval
+  r1 = pointer to newval
+  r2 = pointer to target value
+  lr = return address
+
+Output:
+
+  r0 = success code (zero or non-zero)
+  C flag = set if r0 == 0, clear if r0 != 0
+
+Clobbered registers:
+
+  r3, lr, flags
+
+Definition:
+
+  Atomically store the 64-bit value pointed by *newval in *ptr only if *ptr
+  is equal to the 64-bit value pointed by *oldval.  Return zero if *ptr was
+  changed or non-zero if no exchange happened.
+
+  The C flag is also set if *ptr was changed to allow for assembly
+  optimization in the calling code.
+
+Usage example:
+
+typedef int (__kuser_cmpxchg64_t)(const int64_t *oldval,
+                                  const int64_t *newval,
+                                  volatile int64_t *ptr);
+#define __kuser_cmpxchg64 (*(__kuser_cmpxchg64_t *)0xffff0f60)
+
+int64_t atomic_add64(volatile int64_t *ptr, int64_t val)
+{
+	int64_t old, new;
+
+	do {
+		old = *ptr;
+		new = old + val;
+	} while(__kuser_cmpxchg64(&old, &new, ptr));
+
+	return new;
+}
+
+Notes:
+
+  - This routine already includes memory barriers as needed.
+
+  - Due to the length of this sequence, this spans 2 conventional kuser
+    "slots", therefore 0xffff0f80 is not used as a valid entry point.
+
+  - Valid only if __kuser_helper_version >= 5 (from kernel version 3.1).

arch/arm/Kconfig

@@ -10,7 +10,7 @@ config ARM
 	select GENERIC_ATOMIC64 if (CPU_V6 || !CPU_32v6K || !AEABI)
 	select HAVE_OPROFILE if (HAVE_PERF_EVENTS)
 	select HAVE_ARCH_KGDB
-	select HAVE_KPROBES if (!XIP_KERNEL && !THUMB2_KERNEL)
+	select HAVE_KPROBES if !XIP_KERNEL
 	select HAVE_KRETPROBES if (HAVE_KPROBES)
 	select HAVE_FUNCTION_TRACER if (!XIP_KERNEL)
 	select HAVE_FTRACE_MCOUNT_RECORD if (!XIP_KERNEL)

arch/arm/include/asm/assembler.h

@@ -293,4 +293,13 @@
 	.macro	ldrusr, reg, ptr, inc, cond=al, rept=1, abort=9001f
 	usracc	ldr, \reg, \ptr, \inc, \cond, \rept, \abort
 	.endm
+
+/* Utility macro for declaring string literals */
+	.macro	string name:req, string
+	.type \name , #object
+\name:
+	.asciz "\string"
+	.size \name , . - \name
+	.endm
+
 #endif /* __ASM_ASSEMBLER_H__ */

arch/arm/include/asm/dma.h

@@ -1,15 +1,16 @@
 #ifndef __ASM_ARM_DMA_H
 #define __ASM_ARM_DMA_H
 
-#include <asm/memory.h>
-
 /*
  * This is the maximum virtual address which can be DMA'd from.
  */
-#ifndef ARM_DMA_ZONE_SIZE
+#ifndef CONFIG_ZONE_DMA
-#define MAX_DMA_ADDRESS	0xffffffff
+#define MAX_DMA_ADDRESS	0xffffffffUL
 #else
-#define MAX_DMA_ADDRESS	(PAGE_OFFSET + ARM_DMA_ZONE_SIZE)
+#define MAX_DMA_ADDRESS	({ \
+	extern unsigned long arm_dma_zone_size; \
+	arm_dma_zone_size ? \
+		(PAGE_OFFSET + arm_dma_zone_size) : 0xffffffffUL; })
 #endif
 
 #ifdef CONFIG_ISA_DMA_API

arch/arm/include/asm/hwcap.h

@@ -4,22 +4,26 @@
 /*
  * HWCAP flags - for elf_hwcap (in kernel) and AT_HWCAP
  */
-#define HWCAP_SWP	1
+#define HWCAP_SWP	(1 << 0)
-#define HWCAP_HALF	2
+#define HWCAP_HALF	(1 << 1)
-#define HWCAP_THUMB	4
+#define HWCAP_THUMB	(1 << 2)
-#define HWCAP_26BIT	8	/* Play it safe */
+#define HWCAP_26BIT	(1 << 3)	/* Play it safe */
-#define HWCAP_FAST_MULT	16
+#define HWCAP_FAST_MULT	(1 << 4)
-#define HWCAP_FPA	32
+#define HWCAP_FPA	(1 << 5)
-#define HWCAP_VFP	64
+#define HWCAP_VFP	(1 << 6)
-#define HWCAP_EDSP	128
+#define HWCAP_EDSP	(1 << 7)
-#define HWCAP_JAVA	256
+#define HWCAP_JAVA	(1 << 8)
-#define HWCAP_IWMMXT	512
+#define HWCAP_IWMMXT	(1 << 9)
-#define HWCAP_CRUNCH	1024
+#define HWCAP_CRUNCH	(1 << 10)
-#define HWCAP_THUMBEE	2048
+#define HWCAP_THUMBEE	(1 << 11)
-#define HWCAP_NEON	4096
+#define HWCAP_NEON	(1 << 12)
-#define HWCAP_VFPv3	8192
+#define HWCAP_VFPv3	(1 << 13)
-#define HWCAP_VFPv3D16	16384
+#define HWCAP_VFPv3D16	(1 << 14)
-#define HWCAP_TLS	32768
+#define HWCAP_TLS	(1 << 15)
+#define HWCAP_VFPv4	(1 << 16)
+#define HWCAP_IDIVA	(1 << 17)
+#define HWCAP_IDIVT	(1 << 18)
+#define HWCAP_IDIV	(HWCAP_IDIVA | HWCAP_IDIVT)
 
 #if defined(__KERNEL__) && !defined(__ASSEMBLY__)
 /*

arch/arm/include/asm/kprobes.h

@@ -24,12 +24,6 @@
 #define MAX_INSN_SIZE			2
 #define MAX_STACK_SIZE			64	/* 32 would probably be OK */
 
-/*
- * This undefined instruction must be unique and
- * reserved solely for kprobes' use.
- */
-#define KPROBE_BREAKPOINT_INSTRUCTION	0xe7f001f8
-
 #define regs_return_value(regs)		((regs)->ARM_r0)
 #define flush_insn_slot(p)		do { } while (0)
 #define kretprobe_blacklist_size	0
@@ -38,14 +32,17 @@ typedef u32 kprobe_opcode_t;
 
 struct kprobe;
 typedef void (kprobe_insn_handler_t)(struct kprobe *, struct pt_regs *);
-
 typedef unsigned long (kprobe_check_cc)(unsigned long);
+typedef void (kprobe_insn_singlestep_t)(struct kprobe *, struct pt_regs *);
+typedef void (kprobe_insn_fn_t)(void);
 
 /* Architecture specific copy of original instruction. */
 struct arch_specific_insn {
-	kprobe_opcode_t		*insn;
+	kprobe_opcode_t			*insn;
-	kprobe_insn_handler_t	*insn_handler;
+	kprobe_insn_handler_t		*insn_handler;
-	kprobe_check_cc		*insn_check_cc;
+	kprobe_check_cc			*insn_check_cc;
+	kprobe_insn_singlestep_t	*insn_singlestep;
+	kprobe_insn_fn_t		*insn_fn;
 };
 
 struct prev_kprobe {
@@ -62,20 +59,9 @@ struct kprobe_ctlblk {
 };
 
 void arch_remove_kprobe(struct kprobe *);
-void kretprobe_trampoline(void);
-
 int kprobe_fault_handler(struct pt_regs *regs, unsigned int fsr);
 int kprobe_exceptions_notify(struct notifier_block *self,
 			     unsigned long val, void *data);
 
-enum kprobe_insn {
-	INSN_REJECTED,
-	INSN_GOOD,
-	INSN_GOOD_NO_SLOT
-};
-
-enum kprobe_insn arm_kprobe_decode_insn(kprobe_opcode_t,
-					struct arch_specific_insn *);
-void __init arm_kprobe_decode_init(void);
 
 #endif /* _ARM_KPROBES_H */

arch/arm/include/asm/mach/arch.h

@@ -23,6 +23,10 @@ struct machine_desc {
 
 	unsigned int		nr_irqs;	/* number of IRQs */
 
+#ifdef CONFIG_ZONE_DMA
+	unsigned long		dma_zone_size;	/* size of DMA-able area */
+#endif
+
 	unsigned int		video_start;	/* start of video RAM	*/
 	unsigned int		video_end;	/* end of video RAM	*/
 

arch/arm/include/asm/perf_event.h

@@ -24,6 +24,8 @@ enum arm_perf_pmu_ids {
 	ARM_PERF_PMU_ID_V6MP,
 	ARM_PERF_PMU_ID_CA8,
 	ARM_PERF_PMU_ID_CA9,
+	ARM_PERF_PMU_ID_CA5,
+	ARM_PERF_PMU_ID_CA15,
 	ARM_NUM_PMU_IDS,
 };
 

arch/arm/include/asm/ptrace.h

@@ -69,8 +69,9 @@
 #define PSR_c		0x000000ff	/* Control		*/
 
 /*
- * ARMv7 groups of APSR bits
+ * ARMv7 groups of PSR bits
  */
+#define APSR_MASK	0xf80f0000	/* N, Z, C, V, Q and GE flags */
 #define PSR_ISET_MASK	0x01000010	/* ISA state (J, T) mask */
 #define PSR_IT_MASK	0x0600fc00	/* If-Then execution state mask */
 #define PSR_ENDIAN_MASK	0x00000200	/* Endianness state mask */
@@ -199,6 +200,14 @@ extern unsigned long profile_pc(struct pt_regs *regs);
 #define predicate(x)		((x) & 0xf0000000)
 #define PREDICATE_ALWAYS	0xe0000000
 
+/*
+ * True if instr is a 32-bit thumb instruction. This works if instr
+ * is the first or only half-word of a thumb instruction. It also works
+ * when instr holds all 32-bits of a wide thumb instruction if stored
+ * in the form (first_half<<16)|(second_half)
+ */
+#define is_wide_instruction(instr)	((unsigned)(instr) >= 0xe800)
+
 /*
  * kprobe-based event tracer support
  */

arch/arm/kernel/Makefile

@@ -37,7 +37,12 @@ obj-$(CONFIG_HAVE_ARM_TWD)	+= smp_twd.o
 obj-$(CONFIG_DYNAMIC_FTRACE)	+= ftrace.o
 obj-$(CONFIG_FUNCTION_GRAPH_TRACER)	+= ftrace.o
 obj-$(CONFIG_KEXEC)		+= machine_kexec.o relocate_kernel.o
-obj-$(CONFIG_KPROBES)		+= kprobes.o kprobes-decode.o
+obj-$(CONFIG_KPROBES)		+= kprobes.o kprobes-common.o
+ifdef CONFIG_THUMB2_KERNEL
+obj-$(CONFIG_KPROBES)		+= kprobes-thumb.o
+else
+obj-$(CONFIG_KPROBES)		+= kprobes-arm.o
+endif
 obj-$(CONFIG_ATAGS_PROC)	+= atags.o
 obj-$(CONFIG_OABI_COMPAT)	+= sys_oabi-compat.o
 obj-$(CONFIG_ARM_THUMBEE)	+= thumbee.o

arch/arm/kernel/entry-armv.S

@@ -377,7 +377,7 @@ ENDPROC(__pabt_svc)
 	.endm
 
 	.macro	kuser_cmpxchg_check
-#if __LINUX_ARM_ARCH__ < 6 && !defined(CONFIG_NEEDS_SYSCALL_FOR_CMPXCHG)
+#if !defined(CONFIG_CPU_32v6K) && !defined(CONFIG_NEEDS_SYSCALL_FOR_CMPXCHG)
 #ifndef CONFIG_MMU
 #warning "NPTL on non MMU needs fixing"
 #else
@@ -386,7 +386,7 @@ ENDPROC(__pabt_svc)
 	@ perform a quick test inline since it should be false
 	@ 99.9999% of the time.  The rest is done out of line.
 	cmp	r4, #TASK_SIZE
-	blhs	kuser_cmpxchg_fixup
+	blhs	kuser_cmpxchg64_fixup
 #endif
 #endif
 	.endm
@@ -701,31 +701,12 @@ ENDPROC(__switch_to)
 /*
  * User helpers.
  *
- * These are segment of kernel provided user code reachable from user space
- * at a fixed address in kernel memory.  This is used to provide user space
- * with some operations which require kernel help because of unimplemented
- * native feature and/or instructions in many ARM CPUs. The idea is for
- * this code to be executed directly in user mode for best efficiency but
- * which is too intimate with the kernel counter part to be left to user
- * libraries.  In fact this code might even differ from one CPU to another
- * depending on the available  instruction set and restrictions like on
- * SMP systems.  In other words, the kernel reserves the right to change
- * this code as needed without warning. Only the entry points and their
- * results are guaranteed to be stable.
- *
  * Each segment is 32-byte aligned and will be moved to the top of the high
  * vector page.  New segments (if ever needed) must be added in front of
  * existing ones.  This mechanism should be used only for things that are
  * really small and justified, and not be abused freely.
  *
- * User space is expected to implement those things inline when optimizing
+ * See Documentation/arm/kernel_user_helpers.txt for formal definitions.
- * for a processor that has the necessary native support, but only if such
- * resulting binaries are already to be incompatible with earlier ARM
- * processors due to the use of unsupported instructions other than what
- * is provided here.  In other words don't make binaries unable to run on
- * earlier processors just for the sake of not using these kernel helpers
- * if your compiled code is not going to use the new instructions for other
- * purpose.
  */
  THUMB(	.arm	)
 
@@ -742,97 +723,104 @@ ENDPROC(__switch_to)
 __kuser_helper_start:
 
 /*
- * Reference prototype:
+ * Due to the length of some sequences, __kuser_cmpxchg64 spans 2 regular
- *
+ * kuser "slots", therefore 0xffff0f80 is not used as a valid entry point.
- *	void __kernel_memory_barrier(void)
- *
- * Input:
- *
- *	lr = return address
- *
- * Output:
- *
- *	none
- *
- * Clobbered:
- *
- *	none
- *
- * Definition and user space usage example:
- *
- *	typedef void (__kernel_dmb_t)(void);
- *	#define __kernel_dmb (*(__kernel_dmb_t *)0xffff0fa0)
- *
- * Apply any needed memory barrier to preserve consistency with data modified
- * manually and __kuser_cmpxchg usage.
- *
- * This could be used as follows:
- *
- * #define __kernel_dmb() \
- *         asm volatile ( "mov r0, #0xffff0fff; mov lr, pc; sub pc, r0, #95" \
- *	        : : : "r0", "lr","cc" )
  */
 
+__kuser_cmpxchg64:				@ 0xffff0f60
+
+#if defined(CONFIG_NEEDS_SYSCALL_FOR_CMPXCHG)
+
+	/*
+	 * Poor you.  No fast solution possible...
+	 * The kernel itself must perform the operation.
+	 * A special ghost syscall is used for that (see traps.c).
+	 */
+	stmfd	sp!, {r7, lr}
+	ldr	r7, 1f			@ it's 20 bits
+	swi	__ARM_NR_cmpxchg64
+	ldmfd	sp!, {r7, pc}
+1:	.word	__ARM_NR_cmpxchg64
+
+#elif defined(CONFIG_CPU_32v6K)
+
+	stmfd	sp!, {r4, r5, r6, r7}
+	ldrd	r4, r5, [r0]			@ load old val
+	ldrd	r6, r7, [r1]			@ load new val
+	smp_dmb	arm
+1:	ldrexd	r0, r1, [r2]			@ load current val
+	eors	r3, r0, r4			@ compare with oldval (1)
+	eoreqs	r3, r1, r5			@ compare with oldval (2)
+	strexdeq r3, r6, r7, [r2]		@ store newval if eq
+	teqeq	r3, #1				@ success?
+	beq	1b				@ if no then retry
+	smp_dmb	arm
+	rsbs	r0, r3, #0			@ set returned val and C flag
+	ldmfd	sp!, {r4, r5, r6, r7}
+	bx	lr
+
+#elif !defined(CONFIG_SMP)
+
+#ifdef CONFIG_MMU
+
+	/*
+	 * The only thing that can break atomicity in this cmpxchg64
+	 * implementation is either an IRQ or a data abort exception
+	 * causing another process/thread to be scheduled in the middle of
+	 * the critical sequence.  The same strategy as for cmpxchg is used.
+	 */
+	stmfd	sp!, {r4, r5, r6, lr}
+	ldmia	r0, {r4, r5}			@ load old val
+	ldmia	r1, {r6, lr}			@ load new val
+1:	ldmia	r2, {r0, r1}			@ load current val
+	eors	r3, r0, r4			@ compare with oldval (1)
+	eoreqs	r3, r1, r5			@ compare with oldval (2)
+2:	stmeqia	r2, {r6, lr}			@ store newval if eq
+	rsbs	r0, r3, #0			@ set return val and C flag
+	ldmfd	sp!, {r4, r5, r6, pc}
+
+	.text
+kuser_cmpxchg64_fixup:
+	@ Called from kuser_cmpxchg_fixup.
+	@ r4 = address of interrupted insn (must be preserved).
+	@ sp = saved regs. r7 and r8 are clobbered.
+	@ 1b = first critical insn, 2b = last critical insn.
+	@ If r4 >= 1b and r4 <= 2b then saved pc_usr is set to 1b.
+	mov	r7, #0xffff0fff
+	sub	r7, r7, #(0xffff0fff - (0xffff0f60 + (1b - __kuser_cmpxchg64)))
+	subs	r8, r4, r7
+	rsbcss	r8, r8, #(2b - 1b)
+	strcs	r7, [sp, #S_PC]
+#if __LINUX_ARM_ARCH__ < 6
+	bcc	kuser_cmpxchg32_fixup
+#endif
+	mov	pc, lr
+	.previous
+
+#else
+#warning "NPTL on non MMU needs fixing"
+	mov	r0, #-1
+	adds	r0, r0, #0
+	usr_ret	lr
+#endif
+
+#else
+#error "incoherent kernel configuration"
+#endif
+
+	/* pad to next slot */
+	.rept	(16 - (. - __kuser_cmpxchg64)/4)
+	.word	0
+	.endr
+
+	.align	5
+
 __kuser_memory_barrier:				@ 0xffff0fa0
 	smp_dmb	arm
 	usr_ret	lr
 
 	.align	5
 
-/*
- * Reference prototype:
- *
- *	int __kernel_cmpxchg(int oldval, int newval, int *ptr)
- *
- * Input:
- *
- *	r0 = oldval
- *	r1 = newval
- *	r2 = ptr
- *	lr = return address
- *
- * Output:
- *
- *	r0 = returned value (zero or non-zero)
- *	C flag = set if r0 == 0, clear if r0 != 0
- *
- * Clobbered:
- *
- *	r3, ip, flags
- *
- * Definition and user space usage example:
- *
- *	typedef int (__kernel_cmpxchg_t)(int oldval, int newval, int *ptr);
- *	#define __kernel_cmpxchg (*(__kernel_cmpxchg_t *)0xffff0fc0)
- *
- * Atomically store newval in *ptr if *ptr is equal to oldval for user space.
- * Return zero if *ptr was changed or non-zero if no exchange happened.
- * The C flag is also set if *ptr was changed to allow for assembly
- * optimization in the calling code.
- *
- * Notes:
- *
- *    - This routine already includes memory barriers as needed.
- *
- * For example, a user space atomic_add implementation could look like this:
- *
- * #define atomic_add(ptr, val) \
- *	({ register unsigned int *__ptr asm("r2") = (ptr); \
- *	   register unsigned int __result asm("r1"); \
- *	   asm volatile ( \
- *	       "1: @ atomic_add\n\t" \
- *	       "ldr	r0, [r2]\n\t" \
- *	       "mov	r3, #0xffff0fff\n\t" \
- *	       "add	lr, pc, #4\n\t" \
- *	       "add	r1, r0, %2\n\t" \
- *	       "add	pc, r3, #(0xffff0fc0 - 0xffff0fff)\n\t" \
- *	       "bcc	1b" \
- *	       : "=&r" (__result) \
- *	       : "r" (__ptr), "rIL" (val) \
- *	       : "r0","r3","ip","lr","cc","memory" ); \
- *	   __result; })
- */
-
 __kuser_cmpxchg:				@ 0xffff0fc0
 
 #if defined(CONFIG_NEEDS_SYSCALL_FOR_CMPXCHG)
@@ -868,7 +856,7 @@ __kuser_cmpxchg:				@ 0xffff0fc0
 	usr_ret	lr
 
 	.text
-kuser_cmpxchg_fixup:
+kuser_cmpxchg32_fixup:
 	@ Called from kuser_cmpxchg_check macro.
 	@ r4 = address of interrupted insn (must be preserved).
 	@ sp = saved regs. r7 and r8 are clobbered.
@@ -906,39 +894,6 @@ kuser_cmpxchg_fixup:
 
 	.align	5
 
-/*
- * Reference prototype:
- *
- *	int __kernel_get_tls(void)
- *
- * Input:
- *
- *	lr = return address
- *
- * Output:
- *
- *	r0 = TLS value
- *
- * Clobbered:
- *
- *	none
- *
- * Definition and user space usage example:
- *
- *	typedef int (__kernel_get_tls_t)(void);
- *	#define __kernel_get_tls (*(__kernel_get_tls_t *)0xffff0fe0)
- *
- * Get the TLS value as previously set via the __ARM_NR_set_tls syscall.
- *
- * This could be used as follows:
- *
- * #define __kernel_get_tls() \
- *	({ register unsigned int __val asm("r0"); \
- *         asm( "mov r0, #0xffff0fff; mov lr, pc; sub pc, r0, #31" \
- *	        : "=r" (__val) : : "lr","cc" ); \
- *	   __val; })
- */
-
 __kuser_get_tls:				@ 0xffff0fe0
 	ldr	r0, [pc, #(16 - 8)]	@ read TLS, set in kuser_get_tls_init
 	usr_ret	lr
@@ -947,19 +902,6 @@ __kuser_get_tls:				@ 0xffff0fe0
 	.word	0			@ 0xffff0ff0 software TLS value, then
 	.endr				@ pad up to __kuser_helper_version
 
-/*
- * Reference declaration:
- *
- *	extern unsigned int __kernel_helper_version;
- *
- * Definition and user space usage example:
- *
- *	#define __kernel_helper_version (*(unsigned int *)0xffff0ffc)
- *
- * User space may read this to determine the curent number of helpers
- * available.
- */
-
 __kuser_helper_version:				@ 0xffff0ffc
 	.word	((__kuser_helper_end - __kuser_helper_start) >> 5)
 

arch/arm/kernel/entry-header.S

@@ -121,15 +121,13 @@
 	.endm
 #else	/* CONFIG_THUMB2_KERNEL */
 	.macro	svc_exit, rpsr
+	ldr	lr, [sp, #S_SP]			@ top of the stack
+	ldrd	r0, r1, [sp, #S_LR]		@ calling lr and pc
 	clrex					@ clear the exclusive monitor
-	ldr	r0, [sp, #S_SP]			@ top of the stack
+	stmdb	lr!, {r0, r1, \rpsr}		@ calling lr and rfe context
-	ldr	r1, [sp, #S_PC]			@ return address
-	tst	r0, #4				@ orig stack 8-byte aligned?
-	stmdb	r0, {r1, \rpsr}			@ rfe context
 	ldmia	sp, {r0 - r12}
-	ldr	lr, [sp, #S_LR]
+	mov	sp, lr
-	addeq	sp, sp, #S_FRAME_SIZE - 8	@ aligned
+	ldr	lr, [sp], #4
-	addne	sp, sp, #S_FRAME_SIZE - 4	@ not aligned
 	rfeia	sp!
 	.endm
 

arch/arm/kernel/kprobes-arm.c

@@ -0,0 +1,999 @@
+/*
+ * arch/arm/kernel/kprobes-decode.c
+ *
+ * Copyright (C) 2006, 2007 Motorola Inc.
+ *
+ * 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 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.
+ */
+
+/*
+ * We do not have hardware single-stepping on ARM, This
+ * effort is further complicated by the ARM not having a
+ * "next PC" register.  Instructions that change the PC
+ * can't be safely single-stepped in a MP environment, so
+ * we have a lot of work to do:
+ *
+ * In the prepare phase:
+ *   *) If it is an instruction that does anything
+ *      with the CPU mode, we reject it for a kprobe.
+ *      (This is out of laziness rather than need.  The
+ *      instructions could be simulated.)
+ *
+ *   *) Otherwise, decode the instruction rewriting its
+ *      registers to take fixed, ordered registers and
+ *      setting a handler for it to run the instruction.
+ *
+ * In the execution phase by an instruction's handler:
+ *
+ *   *) If the PC is written to by the instruction, the
+ *      instruction must be fully simulated in software.
+ *
+ *   *) Otherwise, a modified form of the instruction is
+ *      directly executed.  Its handler calls the
+ *      instruction in insn[0].  In insn[1] is a
+ *      "mov pc, lr" to return.
+ *
+ *      Before calling, load up the reordered registers
+ *      from the original instruction's registers.  If one
+ *      of the original input registers is the PC, compute
+ *      and adjust the appropriate input register.
+ *
+ *	After call completes, copy the output registers to
+ *      the original instruction's original registers.
+ *
+ * We don't use a real breakpoint instruction since that
+ * would have us in the kernel go from SVC mode to SVC
+ * mode losing the link register.  Instead we use an
+ * undefined instruction.  To simplify processing, the
+ * undefined instruction used for kprobes must be reserved
+ * exclusively for kprobes use.
+ *
+ * TODO: ifdef out some instruction decoding based on architecture.
+ */
+
+#include <linux/kernel.h>
+#include <linux/kprobes.h>
+
+#include "kprobes.h"
+
+#define sign_extend(x, signbit) ((x) | (0 - ((x) & (1 << (signbit)))))
+
+#define branch_displacement(insn) sign_extend(((insn) & 0xffffff) << 2, 25)
+
+#if  __LINUX_ARM_ARCH__ >= 6
+#define BLX(reg)	"blx	"reg"		\n\t"
+#else
+#define BLX(reg)	"mov	lr, pc		\n\t"	\
+			"mov	pc, "reg"	\n\t"
+#endif
+
+/*
+ * To avoid the complications of mimicing single-stepping on a
+ * processor without a Next-PC or a single-step mode, and to
+ * avoid having to deal with the side-effects of boosting, we
+ * simulate or emulate (almost) all ARM instructions.
+ *
+ * "Simulation" is where the instruction's behavior is duplicated in
+ * C code.  "Emulation" is where the original instruction is rewritten
+ * and executed, often by altering its registers.
+ *
+ * By having all behavior of the kprobe'd instruction completed before
+ * returning from the kprobe_handler(), all locks (scheduler and
+ * interrupt) can safely be released.  There is no need for secondary
+ * breakpoints, no race with MP or preemptable kernels, nor having to
+ * clean up resources counts at a later time impacting overall system
+ * performance.  By rewriting the instruction, only the minimum registers
+ * need to be loaded and saved back optimizing performance.
+ *
+ * Calling the insnslot_*_rwflags version of a function doesn't hurt
+ * anything even when the CPSR flags aren't updated by the
+ * instruction.  It's just a little slower in return for saving
+ * a little space by not having a duplicate function that doesn't
+ * update the flags.  (The same optimization can be said for
+ * instructions that do or don't perform register writeback)
+ * Also, instructions can either read the flags, only write the
+ * flags, or read and write the flags.  To save combinations
+ * rather than for sheer performance, flag functions just assume
+ * read and write of flags.
+ */
+
+static void __kprobes simulate_bbl(struct kprobe *p, struct pt_regs *regs)
+{
+	kprobe_opcode_t insn = p->opcode;
+	long iaddr = (long)p->addr;
+	int disp  = branch_displacement(insn);
+
+	if (insn & (1 << 24))
+		regs->ARM_lr = iaddr + 4;
+
+	regs->ARM_pc = iaddr + 8 + disp;
+}
+
+static void __kprobes simulate_blx1(struct kprobe *p, struct pt_regs *regs)
+{
+	kprobe_opcode_t insn = p->opcode;
+	long iaddr = (long)p->addr;
+	int disp = branch_displacement(insn);
+
+	regs->ARM_lr = iaddr + 4;
+	regs->ARM_pc = iaddr + 8 + disp + ((insn >> 23) & 0x2);
+	regs->ARM_cpsr |= PSR_T_BIT;
+}
+
+static void __kprobes simulate_blx2bx(struct kprobe *p, struct pt_regs *regs)
+{
+	kprobe_opcode_t insn = p->opcode;
+	int rm = insn & 0xf;
+	long rmv = regs->uregs[rm];
+
+	if (insn & (1 << 5))
+		regs->ARM_lr = (long)p->addr + 4;
+
+	regs->ARM_pc = rmv & ~0x1;
+	regs->ARM_cpsr &= ~PSR_T_BIT;
+	if (rmv & 0x1)
+		regs->ARM_cpsr |= PSR_T_BIT;
+}
+
+static void __kprobes simulate_mrs(struct kprobe *p, struct pt_regs *regs)
+{
+	kprobe_opcode_t insn = p->opcode;
+	int rd = (insn >> 12) & 0xf;
+	unsigned long mask = 0xf8ff03df; /* Mask out execution state */
+	regs->uregs[rd] = regs->ARM_cpsr & mask;
+}
+
+static void __kprobes simulate_mov_ipsp(struct kprobe *p, struct pt_regs *regs)
+{
+	regs->uregs[12] = regs->uregs[13];
+}
+
+static void __kprobes
+emulate_ldrdstrd(struct kprobe *p, struct pt_regs *regs)
+{
+	kprobe_opcode_t insn = p->opcode;
+	unsigned long pc = (unsigned long)p->addr + 8;
+	int rt = (insn >> 12) & 0xf;
+	int rn = (insn >> 16) & 0xf;
+	int rm = insn & 0xf;
+
+	register unsigned long rtv asm("r0") = regs->uregs[rt];
+	register unsigned long rt2v asm("r1") = regs->uregs[rt+1];
+	register unsigned long rnv asm("r2") = (rn == 15) ? pc
+							  : regs->uregs[rn];
+	register unsigned long rmv asm("r3") = regs->uregs[rm];
+
+	__asm__ __volatile__ (
+		BLX("%[fn]")
+		: "=r" (rtv), "=r" (rt2v), "=r" (rnv)
+		: "0" (rtv), "1" (rt2v), "2" (rnv), "r" (rmv),
+		  [fn] "r" (p->ainsn.insn_fn)
+		: "lr", "memory", "cc"
+	);
+
+	regs->uregs[rt] = rtv;
+	regs->uregs[rt+1] = rt2v;
+	if (is_writeback(insn))
+		regs->uregs[rn] = rnv;
+}
+
+static void __kprobes
+emulate_ldr(struct kprobe *p, struct pt_regs *regs)
+{
+	kprobe_opcode_t insn = p->opcode;
+	unsigned long pc = (unsigned long)p->addr + 8;
+	int rt = (insn >> 12) & 0xf;
+	int rn = (insn >> 16) & 0xf;
+	int rm = insn & 0xf;
+
+	register unsigned long rtv asm("r0");
+	register unsigned long rnv asm("r2") = (rn == 15) ? pc
+							  : regs->uregs[rn];
+	register unsigned long rmv asm("r3") = regs->uregs[rm];
+
+	__asm__ __volatile__ (
+		BLX("%[fn]")
+		: "=r" (rtv), "=r" (rnv)
+		: "1" (rnv), "r" (rmv), [fn] "r" (p->ainsn.insn_fn)
+		: "lr", "memory", "cc"
+	);
+
+	if (rt == 15)
+		load_write_pc(rtv, regs);
+	else
+		regs->uregs[rt] = rtv;
+
+	if (is_writeback(insn))
+		regs->uregs[rn] = rnv;
+}
+
+static void __kprobes
+emulate_str(struct kprobe *p, struct pt_regs *regs)
+{
+	kprobe_opcode_t insn = p->opcode;
+	unsigned long rtpc = (unsigned long)p->addr + str_pc_offset;
+	unsigned long rnpc = (unsigned long)p->addr + 8;
+	int rt = (insn >> 12) & 0xf;
+	int rn = (insn >> 16) & 0xf;
+	int rm = insn & 0xf;
+
+	register unsigned long rtv asm("r0") = (rt == 15) ? rtpc
+							  : regs->uregs[rt];
+	register unsigned long rnv asm("r2") = (rn == 15) ? rnpc
+							  : regs->uregs[rn];
+	register unsigned long rmv asm("r3") = regs->uregs[rm];
+
+	__asm__ __volatile__ (
+		BLX("%[fn]")
+		: "=r" (rnv)
+		: "r" (rtv), "0" (rnv), "r" (rmv), [fn] "r" (p->ainsn.insn_fn)
+		: "lr", "memory", "cc"
+	);
+
+	if (is_writeback(insn))
+		regs->uregs[rn] = rnv;
+}
+
+static void __kprobes
+emulate_rd12rn16rm0rs8_rwflags(struct kprobe *p, struct pt_regs *regs)
+{
+	kprobe_opcode_t insn = p->opcode;
+	unsigned long pc = (unsigned long)p->addr + 8;
+	int rd = (insn >> 12) & 0xf;
+	int rn = (insn >> 16) & 0xf;
+	int rm = insn & 0xf;
+	int rs = (insn >> 8) & 0xf;
+
+	register unsigned long rdv asm("r0") = regs->uregs[rd];
+	register unsigned long rnv asm("r2") = (rn == 15) ? pc
+							  : regs->uregs[rn];
+	register unsigned long rmv asm("r3") = (rm == 15) ? pc
+							  : regs->uregs[rm];
+	register unsigned long rsv asm("r1") = regs->uregs[rs];
+	unsigned long cpsr = regs->ARM_cpsr;
+
+	__asm__ __volatile__ (
+		"msr	cpsr_fs, %[cpsr]	\n\t"
+		BLX("%[fn]")
+		"mrs	%[cpsr], cpsr		\n\t"
+		: "=r" (rdv), [cpsr] "=r" (cpsr)
+		: "0" (rdv), "r" (rnv), "r" (rmv), "r" (rsv),
+		  "1" (cpsr), [fn] "r" (p->ainsn.insn_fn)
+		: "lr", "memory", "cc"
+	);
+
+	if (rd == 15)
+		alu_write_pc(rdv, regs);
+	else
+		regs->uregs[rd] = rdv;
+	regs->ARM_cpsr = (regs->ARM_cpsr & ~APSR_MASK) | (cpsr & APSR_MASK);
+}
+
+static void __kprobes
+emulate_rd12rn16rm0_rwflags_nopc(struct kprobe *p, struct pt_regs *regs)
+{
+	kprobe_opcode_t insn = p->opcode;
+	int rd = (insn >> 12) & 0xf;
+	int rn = (insn >> 16) & 0xf;
+	int rm = insn & 0xf;
+
+	register unsigned long rdv asm("r0") = regs->uregs[rd];
+	register unsigned long rnv asm("r2") = regs->uregs[rn];
+	register unsigned long rmv asm("r3") = regs->uregs[rm];
+	unsigned long cpsr = regs->ARM_cpsr;
+
+	__asm__ __volatile__ (
+		"msr	cpsr_fs, %[cpsr]	\n\t"
+		BLX("%[fn]")
+		"mrs	%[cpsr], cpsr		\n\t"
+		: "=r" (rdv), [cpsr] "=r" (cpsr)
+		: "0" (rdv), "r" (rnv), "r" (rmv),
+		  "1" (cpsr), [fn] "r" (p->ainsn.insn_fn)
+		: "lr", "memory", "cc"
+	);
+
+	regs->uregs[rd] = rdv;
+	regs->ARM_cpsr = (regs->ARM_cpsr & ~APSR_MASK) | (cpsr & APSR_MASK);
+}
+
+static void __kprobes
+emulate_rd16rn12rm0rs8_rwflags_nopc(struct kprobe *p, struct pt_regs *regs)
+{
+	kprobe_opcode_t insn = p->opcode;
+	int rd = (insn >> 16) & 0xf;
+	int rn = (insn >> 12) & 0xf;
+	int rm = insn & 0xf;
+	int rs = (insn >> 8) & 0xf;
+
+	register unsigned long rdv asm("r2") = regs->uregs[rd];
+	register unsigned long rnv asm("r0") = regs->uregs[rn];
+	register unsigned long rmv asm("r3") = regs->uregs[rm];
+	register unsigned long rsv asm("r1") = regs->uregs[rs];
+	unsigned long cpsr = regs->ARM_cpsr;
+
+	__asm__ __volatile__ (
+		"msr	cpsr_fs, %[cpsr]	\n\t"
+		BLX("%[fn]")
+		"mrs	%[cpsr], cpsr		\n\t"
+		: "=r" (rdv), [cpsr] "=r" (cpsr)
+		: "0" (rdv), "r" (rnv), "r" (rmv), "r" (rsv),
+		  "1" (cpsr), [fn] "r" (p->ainsn.insn_fn)
+		: "lr", "memory", "cc"
+	);
+
+	regs->uregs[rd] = rdv;
+	regs->ARM_cpsr = (regs->ARM_cpsr & ~APSR_MASK) | (cpsr & APSR_MASK);
+}
+
+static void __kprobes
+emulate_rd12rm0_noflags_nopc(struct kprobe *p, struct pt_regs *regs)
+{
+	kprobe_opcode_t insn = p->opcode;
+	int rd = (insn >> 12) & 0xf;
+	int rm = insn & 0xf;
+
+	register unsigned long rdv asm("r0") = regs->uregs[rd];
+	register unsigned long rmv asm("r3") = regs->uregs[rm];
+
+	__asm__ __volatile__ (
+		BLX("%[fn]")
+		: "=r" (rdv)
+		: "0" (rdv), "r" (rmv), [fn] "r" (p->ainsn.insn_fn)
+		: "lr", "memory", "cc"
+	);
+
+	regs->uregs[rd] = rdv;
+}
+
+static void __kprobes
+emulate_rdlo12rdhi16rn0rm8_rwflags_nopc(struct kprobe *p, struct pt_regs *regs)
+{
+	kprobe_opcode_t insn = p->opcode;
+	int rdlo = (insn >> 12) & 0xf;
+	int rdhi = (insn >> 16) & 0xf;
+	int rn = insn & 0xf;
+	int rm = (insn >> 8) & 0xf;
+
+	register unsigned long rdlov asm("r0") = regs->uregs[rdlo];
+	register unsigned long rdhiv asm("r2") = regs->uregs[rdhi];
+	register unsigned long rnv asm("r3") = regs->uregs[rn];
+	register unsigned long rmv asm("r1") = regs->uregs[rm];
+	unsigned long cpsr = regs->ARM_cpsr;
+
+	__asm__ __volatile__ (
+		"msr	cpsr_fs, %[cpsr]	\n\t"
+		BLX("%[fn]")
+		"mrs	%[cpsr], cpsr		\n\t"
+		: "=r" (rdlov), "=r" (rdhiv), [cpsr] "=r" (cpsr)
+		: "0" (rdlov), "1" (rdhiv), "r" (rnv), "r" (rmv),
+		  "2" (cpsr), [fn] "r" (p->ainsn.insn_fn)
+		: "lr", "memory", "cc"
+	);
+
+	regs->uregs[rdlo] = rdlov;
+	regs->uregs[rdhi] = rdhiv;
+	regs->ARM_cpsr = (regs->ARM_cpsr & ~APSR_MASK) | (cpsr & APSR_MASK);
+}
+
+/*
+ * For the instruction masking and comparisons in all the "space_*"
+ * functions below, Do _not_ rearrange the order of tests unless
+ * you're very, very sure of what you are doing.  For the sake of
+ * efficiency, the masks for some tests sometimes assume other test
+ * have been done prior to them so the number of patterns to test
+ * for an instruction set can be as broad as possible to reduce the
+ * number of tests needed.
+ */
+
+static const union decode_item arm_1111_table[] = {
+	/* Unconditional instructions					*/
+
+	/* memory hint		1111 0100 x001 xxxx xxxx xxxx xxxx xxxx */
+	/* PLDI (immediate)	1111 0100 x101 xxxx xxxx xxxx xxxx xxxx */
+	/* PLDW (immediate)	1111 0101 x001 xxxx xxxx xxxx xxxx xxxx */
+	/* PLD (immediate)	1111 0101 x101 xxxx xxxx xxxx xxxx xxxx */
+	DECODE_SIMULATE	(0xfe300000, 0xf4100000, kprobe_simulate_nop),
+
+	/* memory hint		1111 0110 x001 xxxx xxxx xxxx xxx0 xxxx */
+	/* PLDI (register)	1111 0110 x101 xxxx xxxx xxxx xxx0 xxxx */
+	/* PLDW (register)	1111 0111 x001 xxxx xxxx xxxx xxx0 xxxx */
+	/* PLD (register)	1111 0111 x101 xxxx xxxx xxxx xxx0 xxxx */
+	DECODE_SIMULATE	(0xfe300010, 0xf6100000, kprobe_simulate_nop),
+
+	/* BLX (immediate)	1111 101x xxxx xxxx xxxx xxxx xxxx xxxx */
+	DECODE_SIMULATE	(0xfe000000, 0xfa000000, simulate_blx1),
+
+	/* CPS			1111 0001 0000 xxx0 xxxx xxxx xx0x xxxx */
+	/* SETEND		1111 0001 0000 0001 xxxx xxxx 0000 xxxx */
+	/* SRS			1111 100x x1x0 xxxx xxxx xxxx xxxx xxxx */
+	/* RFE			1111 100x x0x1 xxxx xxxx xxxx xxxx xxxx */
+
+	/* Coprocessor instructions... */
+	/* MCRR2		1111 1100 0100 xxxx xxxx xxxx xxxx xxxx */
+	/* MRRC2		1111 1100 0101 xxxx xxxx xxxx xxxx xxxx */
+	/* LDC2			1111 110x xxx1 xxxx xxxx xxxx xxxx xxxx */
+	/* STC2			1111 110x xxx0 xxxx xxxx xxxx xxxx xxxx */
+	/* CDP2			1111 1110 xxxx xxxx xxxx xxxx xxx0 xxxx */
+	/* MCR2			1111 1110 xxx0 xxxx xxxx xxxx xxx1 xxxx */
+	/* MRC2			1111 1110 xxx1 xxxx xxxx xxxx xxx1 xxxx */
+
+	/* Other unallocated instructions...				*/
+	DECODE_END
+};
+
+static const union decode_item arm_cccc_0001_0xx0____0xxx_table[] = {
+	/* Miscellaneous instructions					*/
+
+	/* MRS cpsr		cccc 0001 0000 xxxx xxxx xxxx 0000 xxxx */
+	DECODE_SIMULATEX(0x0ff000f0, 0x01000000, simulate_mrs,
+						 REGS(0, NOPC, 0, 0, 0)),
+
+	/* BX			cccc 0001 0010 xxxx xxxx xxxx 0001 xxxx */
+	DECODE_SIMULATE	(0x0ff000f0, 0x01200010, simulate_blx2bx),
+
+	/* BLX (register)	cccc 0001 0010 xxxx xxxx xxxx 0011 xxxx */
+	DECODE_SIMULATEX(0x0ff000f0, 0x01200030, simulate_blx2bx,
+						 REGS(0, 0, 0, 0, NOPC)),
+
+	/* CLZ			cccc 0001 0110 xxxx xxxx xxxx 0001 xxxx */
+	DECODE_EMULATEX	(0x0ff000f0, 0x01600010, emulate_rd12rm0_noflags_nopc,
+						 REGS(0, NOPC, 0, 0, NOPC)),
+
+	/* QADD			cccc 0001 0000 xxxx xxxx xxxx 0101 xxxx */
+	/* QSUB			cccc 0001 0010 xxxx xxxx xxxx 0101 xxxx */
+	/* QDADD		cccc 0001 0100 xxxx xxxx xxxx 0101 xxxx */
+	/* QDSUB		cccc 0001 0110 xxxx xxxx xxxx 0101 xxxx */
+	DECODE_EMULATEX	(0x0f9000f0, 0x01000050, emulate_rd12rn16rm0_rwflags_nopc,
+						 REGS(NOPC, NOPC, 0, 0, NOPC)),
+
+	/* BXJ			cccc 0001 0010 xxxx xxxx xxxx 0010 xxxx */
+	/* MSR			cccc 0001 0x10 xxxx xxxx xxxx 0000 xxxx */
+	/* MRS spsr		cccc 0001 0100 xxxx xxxx xxxx 0000 xxxx */
+	/* BKPT			1110 0001 0010 xxxx xxxx xxxx 0111 xxxx */
+	/* SMC			cccc 0001 0110 xxxx xxxx xxxx 0111 xxxx */
+	/* And unallocated instructions...				*/
+	DECODE_END
+};
+
+static const union decode_item arm_cccc_0001_0xx0____1xx0_table[] = {
+	/* Halfword multiply and multiply-accumulate			*/
+
+	/* SMLALxy		cccc 0001 0100 xxxx xxxx xxxx 1xx0 xxxx */
+	DECODE_EMULATEX	(0x0ff00090, 0x01400080, emulate_rdlo12rdhi16rn0rm8_rwflags_nopc,
+						 REGS(NOPC, NOPC, NOPC, 0, NOPC)),
+
+	/* SMULWy		cccc 0001 0010 xxxx xxxx xxxx 1x10 xxxx */
+	DECODE_OR	(0x0ff000b0, 0x012000a0),
+	/* SMULxy		cccc 0001 0110 xxxx xxxx xxxx 1xx0 xxxx */
+	DECODE_EMULATEX	(0x0ff00090, 0x01600080, emulate_rd16rn12rm0rs8_rwflags_nopc,
+						 REGS(NOPC, 0, NOPC, 0, NOPC)),
+
+	/* SMLAxy		cccc 0001 0000 xxxx xxxx xxxx 1xx0 xxxx */
+	DECODE_OR	(0x0ff00090, 0x01000080),
+	/* SMLAWy		cccc 0001 0010 xxxx xxxx xxxx 1x00 xxxx */
+	DECODE_EMULATEX	(0x0ff000b0, 0x01200080, emulate_rd16rn12rm0rs8_rwflags_nopc,
+						 REGS(NOPC, NOPC, NOPC, 0, NOPC)),
+
+	DECODE_END
+};
+
+static const union decode_item arm_cccc_0000_____1001_table[] = {
+	/* Multiply and multiply-accumulate				*/
+
+	/* MUL			cccc 0000 0000 xxxx xxxx xxxx 1001 xxxx */
+	/* MULS			cccc 0000 0001 xxxx xxxx xxxx 1001 xxxx */
+	DECODE_EMULATEX	(0x0fe000f0, 0x00000090, emulate_rd16rn12rm0rs8_rwflags_nopc,
+						 REGS(NOPC, 0, NOPC, 0, NOPC)),
+
+	/* MLA			cccc 0000 0010 xxxx xxxx xxxx 1001 xxxx */
+	/* MLAS			cccc 0000 0011 xxxx xxxx xxxx 1001 xxxx */
+	DECODE_OR	(0x0fe000f0, 0x00200090),
+	/* MLS			cccc 0000 0110 xxxx xxxx xxxx 1001 xxxx */
+	DECODE_EMULATEX	(0x0ff000f0, 0x00600090, emulate_rd16rn12rm0rs8_rwflags_nopc,
+						 REGS(NOPC, NOPC, NOPC, 0, NOPC)),
+
+	/* UMAAL		cccc 0000 0100 xxxx xxxx xxxx 1001 xxxx */
+	DECODE_OR	(0x0ff000f0, 0x00400090),
+	/* UMULL		cccc 0000 1000 xxxx xxxx xxxx 1001 xxxx */
+	/* UMULLS		cccc 0000 1001 xxxx xxxx xxxx 1001 xxxx */
+	/* UMLAL		cccc 0000 1010 xxxx xxxx xxxx 1001 xxxx */
+	/* UMLALS		cccc 0000 1011 xxxx xxxx xxxx 1001 xxxx */
+	/* SMULL		cccc 0000 1100 xxxx xxxx xxxx 1001 xxxx */
+	/* SMULLS		cccc 0000 1101 xxxx xxxx xxxx 1001 xxxx */
+	/* SMLAL		cccc 0000 1110 xxxx xxxx xxxx 1001 xxxx */
+	/* SMLALS		cccc 0000 1111 xxxx xxxx xxxx 1001 xxxx */
+	DECODE_EMULATEX	(0x0f8000f0, 0x00800090, emulate_rdlo12rdhi16rn0rm8_rwflags_nopc,
+						 REGS(NOPC, NOPC, NOPC, 0, NOPC)),
+
+	DECODE_END
+};
+
+static const union decode_item arm_cccc_0001_____1001_table[] = {
+	/* Synchronization primitives					*/
+
+	/* SMP/SWPB		cccc 0001 0x00 xxxx xxxx xxxx 1001 xxxx */
+	DECODE_EMULATEX	(0x0fb000f0, 0x01000090, emulate_rd12rn16rm0_rwflags_nopc,
+						 REGS(NOPC, NOPC, 0, 0, NOPC)),
+
+	/* LDREX/STREX{,D,B,H}	cccc 0001 1xxx xxxx xxxx xxxx 1001 xxxx */
+	/* And unallocated instructions...				*/
+	DECODE_END
+};
+
+static const union decode_item arm_cccc_000x_____1xx1_table[] = {
+	/* Extra load/store instructions				*/
+
+	/* STRHT		cccc 0000 xx10 xxxx xxxx xxxx 1011 xxxx */
+	/* ???			cccc 0000 xx10 xxxx xxxx xxxx 11x1 xxxx */
+	/* LDRHT		cccc 0000 xx11 xxxx xxxx xxxx 1011 xxxx */
+	/* LDRSBT		cccc 0000 xx11 xxxx xxxx xxxx 1101 xxxx */
+	/* LDRSHT		cccc 0000 xx11 xxxx xxxx xxxx 1111 xxxx */
+	DECODE_REJECT	(0x0f200090, 0x00200090),
+
+	/* LDRD/STRD lr,pc,{...	cccc 000x x0x0 xxxx 111x xxxx 1101 xxxx */
+	DECODE_REJECT	(0x0e10e0d0, 0x0000e0d0),
+
+	/* LDRD (register)	cccc 000x x0x0 xxxx xxxx xxxx 1101 xxxx */
+	/* STRD (register)	cccc 000x x0x0 xxxx xxxx xxxx 1111 xxxx */
+	DECODE_EMULATEX	(0x0e5000d0, 0x000000d0, emulate_ldrdstrd,
+						 REGS(NOPCWB, NOPCX, 0, 0, NOPC)),
+
+	/* LDRD (immediate)	cccc 000x x1x0 xxxx xxxx xxxx 1101 xxxx */
+	/* STRD (immediate)	cccc 000x x1x0 xxxx xxxx xxxx 1111 xxxx */
+	DECODE_EMULATEX	(0x0e5000d0, 0x004000d0, emulate_ldrdstrd,
+						 REGS(NOPCWB, NOPCX, 0, 0, 0)),
+
+	/* STRH (register)	cccc 000x x0x0 xxxx xxxx xxxx 1011 xxxx */
+	DECODE_EMULATEX	(0x0e5000f0, 0x000000b0, emulate_str,
+						 REGS(NOPCWB, NOPC, 0, 0, NOPC)),
+
+	/* LDRH (register)	cccc 000x x0x1 xxxx xxxx xxxx 1011 xxxx */
+	/* LDRSB (register)	cccc 000x x0x1 xxxx xxxx xxxx 1101 xxxx */
+	/* LDRSH (register)	cccc 000x x0x1 xxxx xxxx xxxx 1111 xxxx */
+	DECODE_EMULATEX	(0x0e500090, 0x00100090, emulate_ldr,
+						 REGS(NOPCWB, NOPC, 0, 0, NOPC)),
+
+	/* STRH (immediate)	cccc 000x x1x0 xxxx xxxx xxxx 1011 xxxx */
+	DECODE_EMULATEX	(0x0e5000f0, 0x004000b0, emulate_str,
+						 REGS(NOPCWB, NOPC, 0, 0, 0)),
+
+	/* LDRH (immediate)	cccc 000x x1x1 xxxx xxxx xxxx 1011 xxxx */
+	/* LDRSB (immediate)	cccc 000x x1x1 xxxx xxxx xxxx 1101 xxxx */
+	/* LDRSH (immediate)	cccc 000x x1x1 xxxx xxxx xxxx 1111 xxxx */
+	DECODE_EMULATEX	(0x0e500090, 0x00500090, emulate_ldr,
+						 REGS(NOPCWB, NOPC, 0, 0, 0)),
+
+	DECODE_END
+};
+
+static const union decode_item arm_cccc_000x_table[] = {
+	/* Data-processing (register)					*/
+
+	/* <op>S PC, ...	cccc 000x xxx1 xxxx 1111 xxxx xxxx xxxx */
+	DECODE_REJECT	(0x0e10f000, 0x0010f000),
+
+	/* MOV IP, SP		1110 0001 1010 0000 1100 0000 0000 1101 */
+	DECODE_SIMULATE	(0xffffffff, 0xe1a0c00d, simulate_mov_ipsp),
+
+	/* TST (register)	cccc 0001 0001 xxxx xxxx xxxx xxx0 xxxx */
+	/* TEQ (register)	cccc 0001 0011 xxxx xxxx xxxx xxx0 xxxx */
+	/* CMP (register)	cccc 0001 0101 xxxx xxxx xxxx xxx0 xxxx */
+	/* CMN (register)	cccc 0001 0111 xxxx xxxx xxxx xxx0 xxxx */
+	DECODE_EMULATEX	(0x0f900010, 0x01100000, emulate_rd12rn16rm0rs8_rwflags,
+						 REGS(ANY, 0, 0, 0, ANY)),
+
+	/* MOV (register)	cccc 0001 101x xxxx xxxx xxxx xxx0 xxxx */
+	/* MVN (register)	cccc 0001 111x xxxx xxxx xxxx xxx0 xxxx */
+	DECODE_EMULATEX	(0x0fa00010, 0x01a00000, emulate_rd12rn16rm0rs8_rwflags,
+						 REGS(0, ANY, 0, 0, ANY)),
+
+	/* AND (register)	cccc 0000 000x xxxx xxxx xxxx xxx0 xxxx */
+	/* EOR (register)	cccc 0000 001x xxxx xxxx xxxx xxx0 xxxx */
+	/* SUB (register)	cccc 0000 010x xxxx xxxx xxxx xxx0 xxxx */
+	/* RSB (register)	cccc 0000 011x xxxx xxxx xxxx xxx0 xxxx */
+	/* ADD (register)	cccc 0000 100x xxxx xxxx xxxx xxx0 xxxx */
+	/* ADC (register)	cccc 0000 101x xxxx xxxx xxxx xxx0 xxxx */
+	/* SBC (register)	cccc 0000 110x xxxx xxxx xxxx xxx0 xxxx */
+	/* RSC (register)	cccc 0000 111x xxxx xxxx xxxx xxx0 xxxx */
+	/* ORR (register)	cccc 0001 100x xxxx xxxx xxxx xxx0 xxxx */
+	/* BIC (register)	cccc 0001 110x xxxx xxxx xxxx xxx0 xxxx */
+	DECODE_EMULATEX	(0x0e000010, 0x00000000, emulate_rd12rn16rm0rs8_rwflags,
+						 REGS(ANY, ANY, 0, 0, ANY)),
+
+	/* TST (reg-shift reg)	cccc 0001 0001 xxxx xxxx xxxx 0xx1 xxxx */
+	/* TEQ (reg-shift reg)	cccc 0001 0011 xxxx xxxx xxxx 0xx1 xxxx */
+	/* CMP (reg-shift reg)	cccc 0001 0101 xxxx xxxx xxxx 0xx1 xxxx */
+	/* CMN (reg-shift reg)	cccc 0001 0111 xxxx xxxx xxxx 0xx1 xxxx */
+	DECODE_EMULATEX	(0x0f900090, 0x01100010, emulate_rd12rn16rm0rs8_rwflags,
+						 REGS(ANY, 0, NOPC, 0, ANY)),
+
+	/* MOV (reg-shift reg)	cccc 0001 101x xxxx xxxx xxxx 0xx1 xxxx */
+	/* MVN (reg-shift reg)	cccc 0001 111x xxxx xxxx xxxx 0xx1 xxxx */
+	DECODE_EMULATEX	(0x0fa00090, 0x01a00010, emulate_rd12rn16rm0rs8_rwflags,
+						 REGS(0, ANY, NOPC, 0, ANY)),
+
+	/* AND (reg-shift reg)	cccc 0000 000x xxxx xxxx xxxx 0xx1 xxxx */
+	/* EOR (reg-shift reg)	cccc 0000 001x xxxx xxxx xxxx 0xx1 xxxx */
+	/* SUB (reg-shift reg)	cccc 0000 010x xxxx xxxx xxxx 0xx1 xxxx */
+	/* RSB (reg-shift reg)	cccc 0000 011x xxxx xxxx xxxx 0xx1 xxxx */
+	/* ADD (reg-shift reg)	cccc 0000 100x xxxx xxxx xxxx 0xx1 xxxx */
+	/* ADC (reg-shift reg)	cccc 0000 101x xxxx xxxx xxxx 0xx1 xxxx */
+	/* SBC (reg-shift reg)	cccc 0000 110x xxxx xxxx xxxx 0xx1 xxxx */
+	/* RSC (reg-shift reg)	cccc 0000 111x xxxx xxxx xxxx 0xx1 xxxx */
+	/* ORR (reg-shift reg)	cccc 0001 100x xxxx xxxx xxxx 0xx1 xxxx */
+	/* BIC (reg-shift reg)	cccc 0001 110x xxxx xxxx xxxx 0xx1 xxxx */
+	DECODE_EMULATEX	(0x0e000090, 0x00000010, emulate_rd12rn16rm0rs8_rwflags,
+						 REGS(ANY, ANY, NOPC, 0, ANY)),
+
+	DECODE_END
+};
+
+static const union decode_item arm_cccc_001x_table[] = {
+	/* Data-processing (immediate)					*/
+
+	/* MOVW			cccc 0011 0000 xxxx xxxx xxxx xxxx xxxx */
+	/* MOVT			cccc 0011 0100 xxxx xxxx xxxx xxxx xxxx */
+	DECODE_EMULATEX	(0x0fb00000, 0x03000000, emulate_rd12rm0_noflags_nopc,
+						 REGS(0, NOPC, 0, 0, 0)),
+
+	/* YIELD		cccc 0011 0010 0000 xxxx xxxx 0000 0001 */
+	DECODE_OR	(0x0fff00ff, 0x03200001),
+	/* SEV			cccc 0011 0010 0000 xxxx xxxx 0000 0100 */
+	DECODE_EMULATE	(0x0fff00ff, 0x03200004, kprobe_emulate_none),
+	/* NOP			cccc 0011 0010 0000 xxxx xxxx 0000 0000 */
+	/* WFE			cccc 0011 0010 0000 xxxx xxxx 0000 0010 */
+	/* WFI			cccc 0011 0010 0000 xxxx xxxx 0000 0011 */
+	DECODE_SIMULATE	(0x0fff00fc, 0x03200000, kprobe_simulate_nop),
+	/* DBG			cccc 0011 0010 0000 xxxx xxxx ffff xxxx */
+	/* unallocated hints	cccc 0011 0010 0000 xxxx xxxx xxxx xxxx */
+	/* MSR (immediate)	cccc 0011 0x10 xxxx xxxx xxxx xxxx xxxx */
+	DECODE_REJECT	(0x0fb00000, 0x03200000),
+
+	/* <op>S PC, ...	cccc 001x xxx1 xxxx 1111 xxxx xxxx xxxx */
+	DECODE_REJECT	(0x0e10f000, 0x0210f000),
+
+	/* TST (immediate)	cccc 0011 0001 xxxx xxxx xxxx xxxx xxxx */
+	/* TEQ (immediate)	cccc 0011 0011 xxxx xxxx xxxx xxxx xxxx */
+	/* CMP (immediate)	cccc 0011 0101 xxxx xxxx xxxx xxxx xxxx */
+	/* CMN (immediate)	cccc 0011 0111 xxxx xxxx xxxx xxxx xxxx */
+	DECODE_EMULATEX	(0x0f900000, 0x03100000, emulate_rd12rn16rm0rs8_rwflags,
+						 REGS(ANY, 0, 0, 0, 0)),
+
+	/* MOV (immediate)	cccc 0011 101x xxxx xxxx xxxx xxxx xxxx */
+	/* MVN (immediate)	cccc 0011 111x xxxx xxxx xxxx xxxx xxxx */
+	DECODE_EMULATEX	(0x0fa00000, 0x03a00000, emulate_rd12rn16rm0rs8_rwflags,
+						 REGS(0, ANY, 0, 0, 0)),
+
+	/* AND (immediate)	cccc 0010 000x xxxx xxxx xxxx xxxx xxxx */
+	/* EOR (immediate)	cccc 0010 001x xxxx xxxx xxxx xxxx xxxx */
+	/* SUB (immediate)	cccc 0010 010x xxxx xxxx xxxx xxxx xxxx */
+	/* RSB (immediate)	cccc 0010 011x xxxx xxxx xxxx xxxx xxxx */
+	/* ADD (immediate)	cccc 0010 100x xxxx xxxx xxxx xxxx xxxx */
+	/* ADC (immediate)	cccc 0010 101x xxxx xxxx xxxx xxxx xxxx */
+	/* SBC (immediate)	cccc 0010 110x xxxx xxxx xxxx xxxx xxxx */
+	/* RSC (immediate)	cccc 0010 111x xxxx xxxx xxxx xxxx xxxx */
+	/* ORR (immediate)	cccc 0011 100x xxxx xxxx xxxx xxxx xxxx */
+	/* BIC (immediate)	cccc 0011 110x xxxx xxxx xxxx xxxx xxxx */
+	DECODE_EMULATEX	(0x0e000000, 0x02000000, emulate_rd12rn16rm0rs8_rwflags,
+						 REGS(ANY, ANY, 0, 0, 0)),
+
+	DECODE_END
+};
+
+static const union decode_item arm_cccc_0110_____xxx1_table[] = {
+	/* Media instructions						*/
+
+	/* SEL			cccc 0110 1000 xxxx xxxx xxxx 1011 xxxx */
+	DECODE_EMULATEX	(0x0ff000f0, 0x068000b0, emulate_rd12rn16rm0_rwflags_nopc,
+						 REGS(NOPC, NOPC, 0, 0, NOPC)),
+
+	/* SSAT			cccc 0110 101x xxxx xxxx xxxx xx01 xxxx */
+	/* USAT			cccc 0110 111x xxxx xxxx xxxx xx01 xxxx */
+	DECODE_OR(0x0fa00030, 0x06a00010),
+	/* SSAT16		cccc 0110 1010 xxxx xxxx xxxx 0011 xxxx */
+	/* USAT16		cccc 0110 1110 xxxx xxxx xxxx 0011 xxxx */
+	DECODE_EMULATEX	(0x0fb000f0, 0x06a00030, emulate_rd12rn16rm0_rwflags_nopc,
+						 REGS(0, NOPC, 0, 0, NOPC)),
+
+	/* REV			cccc 0110 1011 xxxx xxxx xxxx 0011 xxxx */
+	/* REV16		cccc 0110 1011 xxxx xxxx xxxx 1011 xxxx */
+	/* RBIT			cccc 0110 1111 xxxx xxxx xxxx 0011 xxxx */
+	/* REVSH		cccc 0110 1111 xxxx xxxx xxxx 1011 xxxx */
+	DECODE_EMULATEX	(0x0fb00070, 0x06b00030, emulate_rd12rm0_noflags_nopc,
+						 REGS(0, NOPC, 0, 0, NOPC)),
+
+	/* ???			cccc 0110 0x00 xxxx xxxx xxxx xxx1 xxxx */
+	DECODE_REJECT	(0x0fb00010, 0x06000010),
+	/* ???			cccc 0110 0xxx xxxx xxxx xxxx 1011 xxxx */
+	DECODE_REJECT	(0x0f8000f0, 0x060000b0),
+	/* ???			cccc 0110 0xxx xxxx xxxx xxxx 1101 xxxx */
+	DECODE_REJECT	(0x0f8000f0, 0x060000d0),
+	/* SADD16		cccc 0110 0001 xxxx xxxx xxxx 0001 xxxx */
+	/* SADDSUBX		cccc 0110 0001 xxxx xxxx xxxx 0011 xxxx */
+	/* SSUBADDX		cccc 0110 0001 xxxx xxxx xxxx 0101 xxxx */
+	/* SSUB16		cccc 0110 0001 xxxx xxxx xxxx 0111 xxxx */
+	/* SADD8		cccc 0110 0001 xxxx xxxx xxxx 1001 xxxx */
+	/* SSUB8		cccc 0110 0001 xxxx xxxx xxxx 1111 xxxx */
+	/* QADD16		cccc 0110 0010 xxxx xxxx xxxx 0001 xxxx */
+	/* QADDSUBX		cccc 0110 0010 xxxx xxxx xxxx 0011 xxxx */
+	/* QSUBADDX		cccc 0110 0010 xxxx xxxx xxxx 0101 xxxx */
+	/* QSUB16		cccc 0110 0010 xxxx xxxx xxxx 0111 xxxx */
+	/* QADD8		cccc 0110 0010 xxxx xxxx xxxx 1001 xxxx */
+	/* QSUB8		cccc 0110 0010 xxxx xxxx xxxx 1111 xxxx */
+	/* SHADD16		cccc 0110 0011 xxxx xxxx xxxx 0001 xxxx */
+	/* SHADDSUBX		cccc 0110 0011 xxxx xxxx xxxx 0011 xxxx */
+	/* SHSUBADDX		cccc 0110 0011 xxxx xxxx xxxx 0101 xxxx */
+	/* SHSUB16		cccc 0110 0011 xxxx xxxx xxxx 0111 xxxx */
+	/* SHADD8		cccc 0110 0011 xxxx xxxx xxxx 1001 xxxx */
+	/* SHSUB8		cccc 0110 0011 xxxx xxxx xxxx 1111 xxxx */
+	/* UADD16		cccc 0110 0101 xxxx xxxx xxxx 0001 xxxx */
+	/* UADDSUBX		cccc 0110 0101 xxxx xxxx xxxx 0011 xxxx */
+	/* USUBADDX		cccc 0110 0101 xxxx xxxx xxxx 0101 xxxx */
+	/* USUB16		cccc 0110 0101 xxxx xxxx xxxx 0111 xxxx */
+	/* UADD8		cccc 0110 0101 xxxx xxxx xxxx 1001 xxxx */
+	/* USUB8		cccc 0110 0101 xxxx xxxx xxxx 1111 xxxx */
+	/* UQADD16		cccc 0110 0110 xxxx xxxx xxxx 0001 xxxx */
+	/* UQADDSUBX		cccc 0110 0110 xxxx xxxx xxxx 0011 xxxx */
+	/* UQSUBADDX		cccc 0110 0110 xxxx xxxx xxxx 0101 xxxx */
+	/* UQSUB16		cccc 0110 0110 xxxx xxxx xxxx 0111 xxxx */
+	/* UQADD8		cccc 0110 0110 xxxx xxxx xxxx 1001 xxxx */
+	/* UQSUB8		cccc 0110 0110 xxxx xxxx xxxx 1111 xxxx */
+	/* UHADD16		cccc 0110 0111 xxxx xxxx xxxx 0001 xxxx */
+	/* UHADDSUBX		cccc 0110 0111 xxxx xxxx xxxx 0011 xxxx */
+	/* UHSUBADDX		cccc 0110 0111 xxxx xxxx xxxx 0101 xxxx */
+	/* UHSUB16		cccc 0110 0111 xxxx xxxx xxxx 0111 xxxx */
+	/* UHADD8		cccc 0110 0111 xxxx xxxx xxxx 1001 xxxx */
+	/* UHSUB8		cccc 0110 0111 xxxx xxxx xxxx 1111 xxxx */
+	DECODE_EMULATEX	(0x0f800010, 0x06000010, emulate_rd12rn16rm0_rwflags_nopc,
+						 REGS(NOPC, NOPC, 0, 0, NOPC)),
+
+	/* PKHBT		cccc 0110 1000 xxxx xxxx xxxx x001 xxxx */
+	/* PKHTB		cccc 0110 1000 xxxx xxxx xxxx x101 xxxx */
+	DECODE_EMULATEX	(0x0ff00030, 0x06800010, emulate_rd12rn16rm0_rwflags_nopc,
+						 REGS(NOPC, NOPC, 0, 0, NOPC)),
+
+	/* ???			cccc 0110 1001 xxxx xxxx xxxx 0111 xxxx */
+	/* ???			cccc 0110 1101 xxxx xxxx xxxx 0111 xxxx */
+	DECODE_REJECT	(0x0fb000f0, 0x06900070),
+
+	/* SXTB16		cccc 0110 1000 1111 xxxx xxxx 0111 xxxx */
+	/* SXTB			cccc 0110 1010 1111 xxxx xxxx 0111 xxxx */
+	/* SXTH			cccc 0110 1011 1111 xxxx xxxx 0111 xxxx */
+	/* UXTB16		cccc 0110 1100 1111 xxxx xxxx 0111 xxxx */
+	/* UXTB			cccc 0110 1110 1111 xxxx xxxx 0111 xxxx */
+	/* UXTH			cccc 0110 1111 1111 xxxx xxxx 0111 xxxx */
+	DECODE_EMULATEX	(0x0f8f00f0, 0x068f0070, emulate_rd12rm0_noflags_nopc,
+						 REGS(0, NOPC, 0, 0, NOPC)),
+
+	/* SXTAB16		cccc 0110 1000 xxxx xxxx xxxx 0111 xxxx */
+	/* SXTAB		cccc 0110 1010 xxxx xxxx xxxx 0111 xxxx */
+	/* SXTAH		cccc 0110 1011 xxxx xxxx xxxx 0111 xxxx */
+	/* UXTAB16		cccc 0110 1100 xxxx xxxx xxxx 0111 xxxx */
+	/* UXTAB		cccc 0110 1110 xxxx xxxx xxxx 0111 xxxx */
+	/* UXTAH		cccc 0110 1111 xxxx xxxx xxxx 0111 xxxx */
+	DECODE_EMULATEX	(0x0f8000f0, 0x06800070, emulate_rd12rn16rm0_rwflags_nopc,
+						 REGS(NOPCX, NOPC, 0, 0, NOPC)),
+
+	DECODE_END
+};
+
+static const union decode_item arm_cccc_0111_____xxx1_table[] = {
+	/* Media instructions						*/
+
+	/* UNDEFINED		cccc 0111 1111 xxxx xxxx xxxx 1111 xxxx */
+	DECODE_REJECT	(0x0ff000f0, 0x07f000f0),
+
+	/* SMLALD		cccc 0111 0100 xxxx xxxx xxxx 00x1 xxxx */
+	/* SMLSLD		cccc 0111 0100 xxxx xxxx xxxx 01x1 xxxx */
+	DECODE_EMULATEX	(0x0ff00090, 0x07400010, emulate_rdlo12rdhi16rn0rm8_rwflags_nopc,
+						 REGS(NOPC, NOPC, NOPC, 0, NOPC)),
+
+	/* SMUAD		cccc 0111 0000 xxxx 1111 xxxx 00x1 xxxx */
+	/* SMUSD		cccc 0111 0000 xxxx 1111 xxxx 01x1 xxxx */
+	DECODE_OR	(0x0ff0f090, 0x0700f010),
+	/* SMMUL		cccc 0111 0101 xxxx 1111 xxxx 00x1 xxxx */
+	DECODE_OR	(0x0ff0f0d0, 0x0750f010),
+	/* USAD8		cccc 0111 1000 xxxx 1111 xxxx 0001 xxxx */
+	DECODE_EMULATEX	(0x0ff0f0f0, 0x0780f010, emulate_rd16rn12rm0rs8_rwflags_nopc,
+						 REGS(NOPC, 0, NOPC, 0, NOPC)),
+
+	/* SMLAD		cccc 0111 0000 xxxx xxxx xxxx 00x1 xxxx */
+	/* SMLSD		cccc 0111 0000 xxxx xxxx xxxx 01x1 xxxx */
+	DECODE_OR	(0x0ff00090, 0x07000010),
+	/* SMMLA		cccc 0111 0101 xxxx xxxx xxxx 00x1 xxxx */
+	DECODE_OR	(0x0ff000d0, 0x07500010),
+	/* USADA8		cccc 0111 1000 xxxx xxxx xxxx 0001 xxxx */
+	DECODE_EMULATEX	(0x0ff000f0, 0x07800010, emulate_rd16rn12rm0rs8_rwflags_nopc,
+						 REGS(NOPC, NOPCX, NOPC, 0, NOPC)),
+
+	/* SMMLS		cccc 0111 0101 xxxx xxxx xxxx 11x1 xxxx */
+	DECODE_EMULATEX	(0x0ff000d0, 0x075000d0, emulate_rd16rn12rm0rs8_rwflags_nopc,
+						 REGS(NOPC, NOPC, NOPC, 0, NOPC)),
+
+	/* SBFX			cccc 0111 101x xxxx xxxx xxxx x101 xxxx */
+	/* UBFX			cccc 0111 111x xxxx xxxx xxxx x101 xxxx */
+	DECODE_EMULATEX	(0x0fa00070, 0x07a00050, emulate_rd12rm0_noflags_nopc,
+						 REGS(0, NOPC, 0, 0, NOPC)),
+
+	/* BFC			cccc 0111 110x xxxx xxxx xxxx x001 1111 */
+	DECODE_EMULATEX	(0x0fe0007f, 0x07c0001f, emulate_rd12rm0_noflags_nopc,
+						 REGS(0, NOPC, 0, 0, 0)),
+
+	/* BFI			cccc 0111 110x xxxx xxxx xxxx x001 xxxx */
+	DECODE_EMULATEX	(0x0fe00070, 0x07c00010, emulate_rd12rm0_noflags_nopc,
+						 REGS(0, NOPC, 0, 0, NOPCX)),
+
+	DECODE_END
+};
+
+static const union decode_item arm_cccc_01xx_table[] = {
+	/* Load/store word and unsigned byte				*/
+
+	/* LDRB/STRB pc,[...]	cccc 01xx x0xx xxxx xxxx xxxx xxxx xxxx */
+	DECODE_REJECT	(0x0c40f000, 0x0440f000),
+
+	/* STRT			cccc 01x0 x010 xxxx xxxx xxxx xxxx xxxx */
+	/* LDRT			cccc 01x0 x011 xxxx xxxx xxxx xxxx xxxx */
+	/* STRBT		cccc 01x0 x110 xxxx xxxx xxxx xxxx xxxx */
+	/* LDRBT		cccc 01x0 x111 xxxx xxxx xxxx xxxx xxxx */
+	DECODE_REJECT	(0x0d200000, 0x04200000),
+
+	/* STR (immediate)	cccc 010x x0x0 xxxx xxxx xxxx xxxx xxxx */
+	/* STRB (immediate)	cccc 010x x1x0 xxxx xxxx xxxx xxxx xxxx */
+	DECODE_EMULATEX	(0x0e100000, 0x04000000, emulate_str,
+						 REGS(NOPCWB, ANY, 0, 0, 0)),
+
+	/* LDR (immediate)	cccc 010x x0x1 xxxx xxxx xxxx xxxx xxxx */
+	/* LDRB (immediate)	cccc 010x x1x1 xxxx xxxx xxxx xxxx xxxx */
+	DECODE_EMULATEX	(0x0e100000, 0x04100000, emulate_ldr,
+						 REGS(NOPCWB, ANY, 0, 0, 0)),
+
+	/* STR (register)	cccc 011x x0x0 xxxx xxxx xxxx xxxx xxxx */
+	/* STRB (register)	cccc 011x x1x0 xxxx xxxx xxxx xxxx xxxx */
+	DECODE_EMULATEX	(0x0e100000, 0x06000000, emulate_str,
+						 REGS(NOPCWB, ANY, 0, 0, NOPC)),
+
+	/* LDR (register)	cccc 011x x0x1 xxxx xxxx xxxx xxxx xxxx */
+	/* LDRB (register)	cccc 011x x1x1 xxxx xxxx xxxx xxxx xxxx */
+	DECODE_EMULATEX	(0x0e100000, 0x06100000, emulate_ldr,
+						 REGS(NOPCWB, ANY, 0, 0, NOPC)),
+
+	DECODE_END
+};
+
+static const union decode_item arm_cccc_100x_table[] = {
+	/* Block data transfer instructions				*/
+
+	/* LDM			cccc 100x x0x1 xxxx xxxx xxxx xxxx xxxx */
+	/* STM			cccc 100x x0x0 xxxx xxxx xxxx xxxx xxxx */
+	DECODE_CUSTOM	(0x0e400000, 0x08000000, kprobe_decode_ldmstm),
+
+	/* STM (user registers)	cccc 100x x1x0 xxxx xxxx xxxx xxxx xxxx */
+	/* LDM (user registers)	cccc 100x x1x1 xxxx 0xxx xxxx xxxx xxxx */
+	/* LDM (exception ret)	cccc 100x x1x1 xxxx 1xxx xxxx xxxx xxxx */
+	DECODE_END
+};
+
+const union decode_item kprobe_decode_arm_table[] = {
+	/*
+	 * Unconditional instructions
+	 *			1111 xxxx xxxx xxxx xxxx xxxx xxxx xxxx
+	 */
+	DECODE_TABLE	(0xf0000000, 0xf0000000, arm_1111_table),
+
+	/*
+	 * Miscellaneous instructions
+	 *			cccc 0001 0xx0 xxxx xxxx xxxx 0xxx xxxx
+	 */
+	DECODE_TABLE	(0x0f900080, 0x01000000, arm_cccc_0001_0xx0____0xxx_table),
+
+	/*
+	 * Halfword multiply and multiply-accumulate
+	 *			cccc 0001 0xx0 xxxx xxxx xxxx 1xx0 xxxx
+	 */
+	DECODE_TABLE	(0x0f900090, 0x01000080, arm_cccc_0001_0xx0____1xx0_table),
+
+	/*
+	 * Multiply and multiply-accumulate
+	 *			cccc 0000 xxxx xxxx xxxx xxxx 1001 xxxx
+	 */
+	DECODE_TABLE	(0x0f0000f0, 0x00000090, arm_cccc_0000_____1001_table),
+
+	/*
+	 * Synchronization primitives
+	 *			cccc 0001 xxxx xxxx xxxx xxxx 1001 xxxx
+	 */
+	DECODE_TABLE	(0x0f0000f0, 0x01000090, arm_cccc_0001_____1001_table),
+
+	/*
+	 * Extra load/store instructions
+	 *			cccc 000x xxxx xxxx xxxx xxxx 1xx1 xxxx
+	 */
+	DECODE_TABLE	(0x0e000090, 0x00000090, arm_cccc_000x_____1xx1_table),
+
+	/*
+	 * Data-processing (register)
+	 *			cccc 000x xxxx xxxx xxxx xxxx xxx0 xxxx
+	 * Data-processing (register-shifted register)
+	 *			cccc 000x xxxx xxxx xxxx xxxx 0xx1 xxxx
+	 */
+	DECODE_TABLE	(0x0e000000, 0x00000000, arm_cccc_000x_table),
+
+	/*
+	 * Data-processing (immediate)
+	 *			cccc 001x xxxx xxxx xxxx xxxx xxxx xxxx
+	 */
+	DECODE_TABLE	(0x0e000000, 0x02000000, arm_cccc_001x_table),
+
+	/*
+	 * Media instructions
+	 *			cccc 011x xxxx xxxx xxxx xxxx xxx1 xxxx
+	 */
+	DECODE_TABLE	(0x0f000010, 0x06000010, arm_cccc_0110_____xxx1_table),
+	DECODE_TABLE	(0x0f000010, 0x07000010, arm_cccc_0111_____xxx1_table),
+
+	/*
+	 * Load/store word and unsigned byte
+	 *			cccc 01xx xxxx xxxx xxxx xxxx xxxx xxxx
+	 */
+	DECODE_TABLE	(0x0c000000, 0x04000000, arm_cccc_01xx_table),
+
+	/*
+	 * Block data transfer instructions
+	 *			cccc 100x xxxx xxxx xxxx xxxx xxxx xxxx
+	 */
+	DECODE_TABLE	(0x0e000000, 0x08000000, arm_cccc_100x_table),
+
+	/* B			cccc 1010 xxxx xxxx xxxx xxxx xxxx xxxx */
+	/* BL			cccc 1011 xxxx xxxx xxxx xxxx xxxx xxxx */
+	DECODE_SIMULATE	(0x0e000000, 0x0a000000, simulate_bbl),
+
+	/*
+	 * Supervisor Call, and coprocessor instructions
+	 */
+
+	/* MCRR			cccc 1100 0100 xxxx xxxx xxxx xxxx xxxx */
+	/* MRRC			cccc 1100 0101 xxxx xxxx xxxx xxxx xxxx */
+	/* LDC			cccc 110x xxx1 xxxx xxxx xxxx xxxx xxxx */
+	/* STC			cccc 110x xxx0 xxxx xxxx xxxx xxxx xxxx */
+	/* CDP			cccc 1110 xxxx xxxx xxxx xxxx xxx0 xxxx */
+	/* MCR			cccc 1110 xxx0 xxxx xxxx xxxx xxx1 xxxx */
+	/* MRC			cccc 1110 xxx1 xxxx xxxx xxxx xxx1 xxxx */
+	/* SVC			cccc 1111 xxxx xxxx xxxx xxxx xxxx xxxx */
+	DECODE_REJECT	(0x0c000000, 0x0c000000),
+
+	DECODE_END
+};
+
+static void __kprobes arm_singlestep(struct kprobe *p, struct pt_regs *regs)
+{
+	regs->ARM_pc += 4;
+	p->ainsn.insn_handler(p, regs);
+}
+
+/* Return:
+ *   INSN_REJECTED     If instruction is one not allowed to kprobe,
+ *   INSN_GOOD         If instruction is supported and uses instruction slot,
+ *   INSN_GOOD_NO_SLOT If instruction is supported but doesn't use its slot.
+ *
+ * For instructions we don't want to kprobe (INSN_REJECTED return result):
+ *   These are generally ones that modify the processor state making
+ *   them "hard" to simulate such as switches processor modes or
+ *   make accesses in alternate modes.  Any of these could be simulated
+ *   if the work was put into it, but low return considering they
+ *   should also be very rare.
+ */
+enum kprobe_insn __kprobes
+arm_kprobe_decode_insn(kprobe_opcode_t insn, struct arch_specific_insn *asi)
+{
+	asi->insn_singlestep = arm_singlestep;
+	asi->insn_check_cc = kprobe_condition_checks[insn>>28];
+	return kprobe_decode_insn(insn, asi, kprobe_decode_arm_table, false);
+}

arch/arm/kernel/kprobes-common.c

@@ -0,0 +1,577 @@
+/*
+ * arch/arm/kernel/kprobes-common.c
+ *
+ * Copyright (C) 2011 Jon Medhurst <tixy@yxit.co.uk>.
+ *
+ * Some contents moved here from arch/arm/include/asm/kprobes-arm.c which is
+ * Copyright (C) 2006, 2007 Motorola Inc.
+ *
+ * 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/kernel.h>
+#include <linux/kprobes.h>
+
+#include "kprobes.h"
+
+
+#ifndef find_str_pc_offset
+
+/*
+ * For STR and STM instructions, an ARM core may choose to use either
+ * a +8 or a +12 displacement from the current instruction's address.
+ * Whichever value is chosen for a given core, it must be the same for
+ * both instructions and may not change.  This function measures it.
+ */
+
+int str_pc_offset;
+
+void __init find_str_pc_offset(void)
+{
+	int addr, scratch, ret;
+
+	__asm__ (
+		"sub	%[ret], pc, #4		\n\t"
+		"str	pc, %[addr]		\n\t"
+		"ldr	%[scr], %[addr]		\n\t"
+		"sub	%[ret], %[scr], %[ret]	\n\t"
+		: [ret] "=r" (ret), [scr] "=r" (scratch), [addr] "+m" (addr));
+
+	str_pc_offset = ret;
+}
+
+#endif /* !find_str_pc_offset */
+
+
+#ifndef test_load_write_pc_interworking
+
+bool load_write_pc_interworks;
+
+void __init test_load_write_pc_interworking(void)
+{
+	int arch = cpu_architecture();
+	BUG_ON(arch == CPU_ARCH_UNKNOWN);
+	load_write_pc_interworks = arch >= CPU_ARCH_ARMv5T;
+}
+
+#endif /* !test_load_write_pc_interworking */
+
+
+#ifndef test_alu_write_pc_interworking
+
+bool alu_write_pc_interworks;
+
+void __init test_alu_write_pc_interworking(void)
+{
+	int arch = cpu_architecture();
+	BUG_ON(arch == CPU_ARCH_UNKNOWN);
+	alu_write_pc_interworks = arch >= CPU_ARCH_ARMv7;
+}
+
+#endif /* !test_alu_write_pc_interworking */
+
+
+void __init arm_kprobe_decode_init(void)
+{
+	find_str_pc_offset();
+	test_load_write_pc_interworking();
+	test_alu_write_pc_interworking();
+}
+
+
+static unsigned long __kprobes __check_eq(unsigned long cpsr)
+{
+	return cpsr & PSR_Z_BIT;
+}
+
+static unsigned long __kprobes __check_ne(unsigned long cpsr)
+{
+	return (~cpsr) & PSR_Z_BIT;
+}
+
+static unsigned long __kprobes __check_cs(unsigned long cpsr)
+{
+	return cpsr & PSR_C_BIT;
+}
+
+static unsigned long __kprobes __check_cc(unsigned long cpsr)
+{
+	return (~cpsr) & PSR_C_BIT;
+}
+
+static unsigned long __kprobes __check_mi(unsigned long cpsr)
+{
+	return cpsr & PSR_N_BIT;
+}
+
+static unsigned long __kprobes __check_pl(unsigned long cpsr)
+{
+	return (~cpsr) & PSR_N_BIT;
+}
+
+static unsigned long __kprobes __check_vs(unsigned long cpsr)
+{
+	return cpsr & PSR_V_BIT;
+}
+
+static unsigned long __kprobes __check_vc(unsigned long cpsr)
+{
+	return (~cpsr) & PSR_V_BIT;
+}
+
+static unsigned long __kprobes __check_hi(unsigned long cpsr)
+{
+	cpsr &= ~(cpsr >> 1); /* PSR_C_BIT &= ~PSR_Z_BIT */
+	return cpsr & PSR_C_BIT;
+}
+
+static unsigned long __kprobes __check_ls(unsigned long cpsr)
+{
+	cpsr &= ~(cpsr >> 1); /* PSR_C_BIT &= ~PSR_Z_BIT */
+	return (~cpsr) & PSR_C_BIT;
+}
+
+static unsigned long __kprobes __check_ge(unsigned long cpsr)
+{
+	cpsr ^= (cpsr << 3); /* PSR_N_BIT ^= PSR_V_BIT */
+	return (~cpsr) & PSR_N_BIT;
+}
+
+static unsigned long __kprobes __check_lt(unsigned long cpsr)
+{
+	cpsr ^= (cpsr << 3); /* PSR_N_BIT ^= PSR_V_BIT */
+	return cpsr & PSR_N_BIT;
+}
+
+static unsigned long __kprobes __check_gt(unsigned long cpsr)
+{
+	unsigned long temp = cpsr ^ (cpsr << 3); /* PSR_N_BIT ^= PSR_V_BIT */
+	temp |= (cpsr << 1);			 /* PSR_N_BIT |= PSR_Z_BIT */
+	return (~temp) & PSR_N_BIT;
+}
+
+static unsigned long __kprobes __check_le(unsigned long cpsr)
+{
+	unsigned long temp = cpsr ^ (cpsr << 3); /* PSR_N_BIT ^= PSR_V_BIT */
+	temp |= (cpsr << 1);			 /* PSR_N_BIT |= PSR_Z_BIT */
+	return temp & PSR_N_BIT;
+}
+
+static unsigned long __kprobes __check_al(unsigned long cpsr)
+{
+	return true;
+}
+
+kprobe_check_cc * const kprobe_condition_checks[16] = {
+	&__check_eq, &__check_ne, &__check_cs, &__check_cc,
+	&__check_mi, &__check_pl, &__check_vs, &__check_vc,
+	&__check_hi, &__check_ls, &__check_ge, &__check_lt,
+	&__check_gt, &__check_le, &__check_al, &__check_al
+};
+
+
+void __kprobes kprobe_simulate_nop(struct kprobe *p, struct pt_regs *regs)
+{
+}
+
+void __kprobes kprobe_emulate_none(struct kprobe *p, struct pt_regs *regs)
+{
+	p->ainsn.insn_fn();
+}
+
+static void __kprobes simulate_ldm1stm1(struct kprobe *p, struct pt_regs *regs)
+{
+	kprobe_opcode_t insn = p->opcode;
+	int rn = (insn >> 16) & 0xf;
+	int lbit = insn & (1 << 20);
+	int wbit = insn & (1 << 21);
+	int ubit = insn & (1 << 23);
+	int pbit = insn & (1 << 24);
+	long *addr = (long *)regs->uregs[rn];
+	int reg_bit_vector;
+	int reg_count;
+
+	reg_count = 0;
+	reg_bit_vector = insn & 0xffff;
+	while (reg_bit_vector) {
+		reg_bit_vector &= (reg_bit_vector - 1);
+		++reg_count;
+	}
+
+	if (!ubit)
+		addr -= reg_count;
+	addr += (!pbit == !ubit);
+
+	reg_bit_vector = insn & 0xffff;
+	while (reg_bit_vector) {
+		int reg = __ffs(reg_bit_vector);
+		reg_bit_vector &= (reg_bit_vector - 1);
+		if (lbit)
+			regs->uregs[reg] = *addr++;
+		else
+			*addr++ = regs->uregs[reg];
+	}
+
+	if (wbit) {
+		if (!ubit)
+			addr -= reg_count;
+		addr -= (!pbit == !ubit);
+		regs->uregs[rn] = (long)addr;
+	}
+}
+
+static void __kprobes simulate_stm1_pc(struct kprobe *p, struct pt_regs *regs)
+{
+	regs->ARM_pc = (long)p->addr + str_pc_offset;
+	simulate_ldm1stm1(p, regs);
+	regs->ARM_pc = (long)p->addr + 4;
+}
+
+static void __kprobes simulate_ldm1_pc(struct kprobe *p, struct pt_regs *regs)
+{
+	simulate_ldm1stm1(p, regs);
+	load_write_pc(regs->ARM_pc, regs);
+}
+
+static void __kprobes
+emulate_generic_r0_12_noflags(struct kprobe *p, struct pt_regs *regs)
+{
+	register void *rregs asm("r1") = regs;
+	register void *rfn asm("lr") = p->ainsn.insn_fn;
+
+	__asm__ __volatile__ (
+		"stmdb	sp!, {%[regs], r11}	\n\t"
+		"ldmia	%[regs], {r0-r12}	\n\t"
+#if __LINUX_ARM_ARCH__ >= 6
+		"blx	%[fn]			\n\t"
+#else
+		"str	%[fn], [sp, #-4]!	\n\t"
+		"adr	lr, 1f			\n\t"
+		"ldr	pc, [sp], #4		\n\t"
+		"1:				\n\t"
+#endif
+		"ldr	lr, [sp], #4		\n\t" /* lr = regs */
+		"stmia	lr, {r0-r12}		\n\t"
+		"ldr	r11, [sp], #4		\n\t"
+		: [regs] "=r" (rregs), [fn] "=r" (rfn)
+		: "0" (rregs), "1" (rfn)
+		: "r0", "r2", "r3", "r4", "r5", "r6", "r7",
+		  "r8", "r9", "r10", "r12", "memory", "cc"
+		);
+}
+
+static void __kprobes
+emulate_generic_r2_14_noflags(struct kprobe *p, struct pt_regs *regs)
+{
+	emulate_generic_r0_12_noflags(p, (struct pt_regs *)(regs->uregs+2));
+}
+
+static void __kprobes
+emulate_ldm_r3_15(struct kprobe *p, struct pt_regs *regs)
+{
+	emulate_generic_r0_12_noflags(p, (struct pt_regs *)(regs->uregs+3));
+	load_write_pc(regs->ARM_pc, regs);
+}
+
+enum kprobe_insn __kprobes
+kprobe_decode_ldmstm(kprobe_opcode_t insn, struct arch_specific_insn *asi)
+{
+	kprobe_insn_handler_t *handler = 0;
+	unsigned reglist = insn & 0xffff;
+	int is_ldm = insn & 0x100000;
+	int rn = (insn >> 16) & 0xf;
+
+	if (rn <= 12 && (reglist & 0xe000) == 0) {
+		/* Instruction only uses registers in the range R0..R12 */
+		handler = emulate_generic_r0_12_noflags;
+
+	} else if (rn >= 2 && (reglist & 0x8003) == 0) {
+		/* Instruction only uses registers in the range R2..R14 */
+		rn -= 2;
+		reglist >>= 2;
+		handler = emulate_generic_r2_14_noflags;
+
+	} else if (rn >= 3 && (reglist & 0x0007) == 0) {
+		/* Instruction only uses registers in the range R3..R15 */
+		if (is_ldm && (reglist & 0x8000)) {
+			rn -= 3;
+			reglist >>= 3;
+			handler = emulate_ldm_r3_15;
+		}
+	}
+
+	if (handler) {
+		/* We can emulate the instruction in (possibly) modified form */
+		asi->insn[0] = (insn & 0xfff00000) | (rn << 16) | reglist;
+		asi->insn_handler = handler;
+		return INSN_GOOD;
+	}
+
+	/* Fallback to slower simulation... */
+	if (reglist & 0x8000)
+		handler = is_ldm ? simulate_ldm1_pc : simulate_stm1_pc;
+	else
+		handler = simulate_ldm1stm1;
+	asi->insn_handler = handler;
+	return INSN_GOOD_NO_SLOT;
+}
+
+
+/*
+ * Prepare an instruction slot to receive an instruction for emulating.
+ * This is done by placing a subroutine return after the location where the
+ * instruction will be placed. We also modify ARM instructions to be
+ * unconditional as the condition code will already be checked before any
+ * emulation handler is called.
+ */
+static kprobe_opcode_t __kprobes
+prepare_emulated_insn(kprobe_opcode_t insn, struct arch_specific_insn *asi,
+								bool thumb)
+{
+#ifdef CONFIG_THUMB2_KERNEL
+	if (thumb) {
+		u16 *thumb_insn = (u16 *)asi->insn;
+		thumb_insn[1] = 0x4770; /* Thumb bx lr */
+		thumb_insn[2] = 0x4770; /* Thumb bx lr */
+		return insn;
+	}
+	asi->insn[1] = 0xe12fff1e; /* ARM bx lr */
+#else
+	asi->insn[1] = 0xe1a0f00e; /* mov pc, lr */
+#endif
+	/* Make an ARM instruction unconditional */
+	if (insn < 0xe0000000)
+		insn = (insn | 0xe0000000) & ~0x10000000;
+	return insn;
+}
+
+/*
+ * Write a (probably modified) instruction into the slot previously prepared by
+ * prepare_emulated_insn
+ */
+static void  __kprobes
+set_emulated_insn(kprobe_opcode_t insn, struct arch_specific_insn *asi,
+								bool thumb)
+{
+#ifdef CONFIG_THUMB2_KERNEL
+	if (thumb) {
+		u16 *ip = (u16 *)asi->insn;
+		if (is_wide_instruction(insn))
+			*ip++ = insn >> 16;
+		*ip++ = insn;
+		return;
+	}
+#endif
+	asi->insn[0] = insn;
+}
+
+/*
+ * When we modify the register numbers encoded in an instruction to be emulated,
+ * the new values come from this define. For ARM and 32-bit Thumb instructions
+ * this gives...
+ *
+ *	bit position	  16  12   8   4   0
+ *	---------------+---+---+---+---+---+
+ *	register	 r2  r0  r1  --  r3
+ */
+#define INSN_NEW_BITS		0x00020103
+
+/* Each nibble has same value as that at INSN_NEW_BITS bit 16 */
+#define INSN_SAMEAS16_BITS	0x22222222
+
+/*
+ * Validate and modify each of the registers encoded in an instruction.
+ *
+ * Each nibble in regs contains a value from enum decode_reg_type. For each
+ * non-zero value, the corresponding nibble in pinsn is validated and modified
+ * according to the type.
+ */
+static bool __kprobes decode_regs(kprobe_opcode_t* pinsn, u32 regs)
+{
+	kprobe_opcode_t insn = *pinsn;
+	kprobe_opcode_t mask = 0xf; /* Start at least significant nibble */
+
+	for (; regs != 0; regs >>= 4, mask <<= 4) {
+
+		kprobe_opcode_t new_bits = INSN_NEW_BITS;
+
+		switch (regs & 0xf) {
+
+		case REG_TYPE_NONE:
+			/* Nibble not a register, skip to next */
+			continue;
+
+		case REG_TYPE_ANY:
+			/* Any register is allowed */
+			break;
+
+		case REG_TYPE_SAMEAS16:
+			/* Replace register with same as at bit position 16 */
+			new_bits = INSN_SAMEAS16_BITS;
+			break;
+
+		case REG_TYPE_SP:
+			/* Only allow SP (R13) */
+			if ((insn ^ 0xdddddddd) & mask)
+				goto reject;
+			break;
+
+		case REG_TYPE_PC:
+			/* Only allow PC (R15) */
+			if ((insn ^ 0xffffffff) & mask)
+				goto reject;
+			break;
+
+		case REG_TYPE_NOSP:
+			/* Reject SP (R13) */
+			if (((insn ^ 0xdddddddd) & mask) == 0)
+				goto reject;
+			break;
+
+		case REG_TYPE_NOSPPC:
+		case REG_TYPE_NOSPPCX:
+			/* Reject SP and PC (R13 and R15) */
+			if (((insn ^ 0xdddddddd) & 0xdddddddd & mask) == 0)
+				goto reject;
+			break;
+
+		case REG_TYPE_NOPCWB:
+			if (!is_writeback(insn))
+				break; /* No writeback, so any register is OK */
+			/* fall through... */
+		case REG_TYPE_NOPC:
+		case REG_TYPE_NOPCX:
+			/* Reject PC (R15) */
+			if (((insn ^ 0xffffffff) & mask) == 0)
+				goto reject;
+			break;
+		}
+
+		/* Replace value of nibble with new register number... */
+		insn &= ~mask;
+		insn |= new_bits & mask;
+	}
+
+	*pinsn = insn;
+	return true;
+
+reject:
+	return false;
+}
+
+static const int decode_struct_sizes[NUM_DECODE_TYPES] = {
+	[DECODE_TYPE_TABLE]	= sizeof(struct decode_table),
+	[DECODE_TYPE_CUSTOM]	= sizeof(struct decode_custom),
+	[DECODE_TYPE_SIMULATE]	= sizeof(struct decode_simulate),
+	[DECODE_TYPE_EMULATE]	= sizeof(struct decode_emulate),
+	[DECODE_TYPE_OR]	= sizeof(struct decode_or),
+	[DECODE_TYPE_REJECT]	= sizeof(struct decode_reject)
+};
+
+/*
+ * kprobe_decode_insn operates on data tables in order to decode an ARM
+ * architecture instruction onto which a kprobe has been placed.
+ *
+ * These instruction decoding tables are a concatenation of entries each
+ * of which consist of one of the following structs:
+ *
+ *	decode_table
+ *	decode_custom
+ *	decode_simulate
+ *	decode_emulate
+ *	decode_or
+ *	decode_reject
+ *
+ * Each of these starts with a struct decode_header which has the following
+ * fields:
+ *
+ *	type_regs
+ *	mask
+ *	value
+ *
+ * The least significant DECODE_TYPE_BITS of type_regs contains a value
+ * from enum decode_type, this indicates which of the decode_* structs
+ * the entry contains. The value DECODE_TYPE_END indicates the end of the
+ * table.
+ *
+ * When the table is parsed, each entry is checked in turn to see if it
+ * matches the instruction to be decoded using the test:
+ *
+ *	(insn & mask) == value
+ *
+ * If no match is found before the end of the table is reached then decoding
+ * fails with INSN_REJECTED.
+ *
+ * When a match is found, decode_regs() is called to validate and modify each
+ * of the registers encoded in the instruction; the data it uses to do this
+ * is (type_regs >> DECODE_TYPE_BITS). A validation failure will cause decoding
+ * to fail with INSN_REJECTED.
+ *
+ * Once the instruction has passed the above tests, further processing
+ * depends on the type of the table entry's decode struct.
+ *
+ */
+int __kprobes
+kprobe_decode_insn(kprobe_opcode_t insn, struct arch_specific_insn *asi,
+				const union decode_item *table, bool thumb)
+{
+	const struct decode_header *h = (struct decode_header *)table;
+	const struct decode_header *next;
+	bool matched = false;
+
+	insn = prepare_emulated_insn(insn, asi, thumb);
+
+	for (;; h = next) {
+		enum decode_type type = h->type_regs.bits & DECODE_TYPE_MASK;
+		u32 regs = h->type_regs.bits >> DECODE_TYPE_BITS;
+
+		if (type == DECODE_TYPE_END)
+			return INSN_REJECTED;
+
+		next = (struct decode_header *)
+				((uintptr_t)h + decode_struct_sizes[type]);
+
+		if (!matched && (insn & h->mask.bits) != h->value.bits)
+			continue;
+
+		if (!decode_regs(&insn, regs))
+			return INSN_REJECTED;
+
+		switch (type) {
+
+		case DECODE_TYPE_TABLE: {
+			struct decode_table *d = (struct decode_table *)h;
+			next = (struct decode_header *)d->table.table;
+			break;
+		}
+
+		case DECODE_TYPE_CUSTOM: {
+			struct decode_custom *d = (struct decode_custom *)h;
+			return (*d->decoder.decoder)(insn, asi);
+		}
+
+		case DECODE_TYPE_SIMULATE: {
+			struct decode_simulate *d = (struct decode_simulate *)h;
+			asi->insn_handler = d->handler.handler;
+			return INSN_GOOD_NO_SLOT;
+		}
+
+		case DECODE_TYPE_EMULATE: {
+			struct decode_emulate *d = (struct decode_emulate *)h;
+			asi->insn_handler = d->handler.handler;
+			set_emulated_insn(insn, asi, thumb);
+			return INSN_GOOD;
+		}
+
+		case DECODE_TYPE_OR:
+			matched = true;
+			break;
+
+		case DECODE_TYPE_REJECT:
+		default:
+			return INSN_REJECTED;
+		}
+		}
+	}

arch/arm/kernel/kprobes.c

@@ -28,14 +28,16 @@
 #include <asm/traps.h>
 #include <asm/cacheflush.h>
 
+#include "kprobes.h"
+
 #define MIN_STACK_SIZE(addr) 				\
 	min((unsigned long)MAX_STACK_SIZE,		\
 	    (unsigned long)current_thread_info() + THREAD_START_SP - (addr))
 
-#define flush_insns(addr, cnt) 				\
+#define flush_insns(addr, size)				\
 	flush_icache_range((unsigned long)(addr),	\
 			   (unsigned long)(addr) +	\
-			   sizeof(kprobe_opcode_t) * (cnt))
+			   (size))
 
 /* Used as a marker in ARM_pc to note when we're in a jprobe. */
 #define JPROBE_MAGIC_ADDR		0xffffffff
@@ -49,16 +51,35 @@ int __kprobes arch_prepare_kprobe(struct kprobe *p)
 	kprobe_opcode_t insn;
 	kprobe_opcode_t tmp_insn[MAX_INSN_SIZE];
 	unsigned long addr = (unsigned long)p->addr;
+	bool thumb;
+	kprobe_decode_insn_t *decode_insn;
 	int is;
 
-	if (addr & 0x3 || in_exception_text(addr))
+	if (in_exception_text(addr))
 		return -EINVAL;
 
+#ifdef CONFIG_THUMB2_KERNEL
+	thumb = true;
+	addr &= ~1; /* Bit 0 would normally be set to indicate Thumb code */
+	insn = ((u16 *)addr)[0];
+	if (is_wide_instruction(insn)) {
+		insn <<= 16;
+		insn |= ((u16 *)addr)[1];
+		decode_insn = thumb32_kprobe_decode_insn;
+	} else
+		decode_insn = thumb16_kprobe_decode_insn;
+#else /* !CONFIG_THUMB2_KERNEL */
+	thumb = false;
+	if (addr & 0x3)
+		return -EINVAL;
 	insn = *p->addr;
+	decode_insn = arm_kprobe_decode_insn;
+#endif
+
 	p->opcode = insn;
 	p->ainsn.insn = tmp_insn;
 
-	switch (arm_kprobe_decode_insn(insn, &p->ainsn)) {
+	switch ((*decode_insn)(insn, &p->ainsn)) {
 	case INSN_REJECTED:	/* not supported */
 		return -EINVAL;
 
@@ -68,7 +89,10 @@ int __kprobes arch_prepare_kprobe(struct kprobe *p)
 			return -ENOMEM;
 		for (is = 0; is < MAX_INSN_SIZE; ++is)
 			p->ainsn.insn[is] = tmp_insn[is];
-		flush_insns(p->ainsn.insn, MAX_INSN_SIZE);
+		flush_insns(p->ainsn.insn,
+				sizeof(p->ainsn.insn[0]) * MAX_INSN_SIZE);
+		p->ainsn.insn_fn = (kprobe_insn_fn_t *)
+					((uintptr_t)p->ainsn.insn | thumb);
 		break;
 
 	case INSN_GOOD_NO_SLOT:	/* instruction doesn't need insn slot */
@@ -79,24 +103,88 @@ int __kprobes arch_prepare_kprobe(struct kprobe *p)
 	return 0;
 }
 
+#ifdef CONFIG_THUMB2_KERNEL
+
+/*
+ * For a 32-bit Thumb breakpoint spanning two memory words we need to take
+ * special precautions to insert the breakpoint atomically, especially on SMP
+ * systems. This is achieved by calling this arming function using stop_machine.
+ */
+static int __kprobes set_t32_breakpoint(void *addr)
+{
+	((u16 *)addr)[0] = KPROBE_THUMB32_BREAKPOINT_INSTRUCTION >> 16;
+	((u16 *)addr)[1] = KPROBE_THUMB32_BREAKPOINT_INSTRUCTION & 0xffff;
+	flush_insns(addr, 2*sizeof(u16));
+	return 0;
+}
+
 void __kprobes arch_arm_kprobe(struct kprobe *p)
 {
-	*p->addr = KPROBE_BREAKPOINT_INSTRUCTION;
+	uintptr_t addr = (uintptr_t)p->addr & ~1; /* Remove any Thumb flag */
-	flush_insns(p->addr, 1);
+
+	if (!is_wide_instruction(p->opcode)) {
+		*(u16 *)addr = KPROBE_THUMB16_BREAKPOINT_INSTRUCTION;
+		flush_insns(addr, sizeof(u16));
+	} else if (addr & 2) {
+		/* A 32-bit instruction spanning two words needs special care */
+		stop_machine(set_t32_breakpoint, (void *)addr, &cpu_online_map);
+	} else {
+		/* Word aligned 32-bit instruction can be written atomically */
+		u32 bkp = KPROBE_THUMB32_BREAKPOINT_INSTRUCTION;
+#ifndef __ARMEB__ /* Swap halfwords for little-endian */
+		bkp = (bkp >> 16) | (bkp << 16);
+#endif
+		*(u32 *)addr = bkp;
+		flush_insns(addr, sizeof(u32));
+	}
 }
 
+#else /* !CONFIG_THUMB2_KERNEL */
+
+void __kprobes arch_arm_kprobe(struct kprobe *p)
+{
+	kprobe_opcode_t insn = p->opcode;
+	kprobe_opcode_t brkp = KPROBE_ARM_BREAKPOINT_INSTRUCTION;
+	if (insn >= 0xe0000000)
+		brkp |= 0xe0000000;  /* Unconditional instruction */
+	else
+		brkp |= insn & 0xf0000000;  /* Copy condition from insn */
+	*p->addr = brkp;
+	flush_insns(p->addr, sizeof(p->addr[0]));
+}
+
+#endif /* !CONFIG_THUMB2_KERNEL */
+
 /*
  * The actual disarming is done here on each CPU and synchronized using
  * stop_machine. This synchronization is necessary on SMP to avoid removing
  * a probe between the moment the 'Undefined Instruction' exception is raised
  * and the moment the exception handler reads the faulting instruction from
- * memory.
+ * memory. It is also needed to atomically set the two half-words of a 32-bit
+ * Thumb breakpoint.
  */
 int __kprobes __arch_disarm_kprobe(void *p)
 {
 	struct kprobe *kp = p;
+#ifdef CONFIG_THUMB2_KERNEL
+	u16 *addr = (u16 *)((uintptr_t)kp->addr & ~1);
+	kprobe_opcode_t insn = kp->opcode;
+	unsigned int len;
+
+	if (is_wide_instruction(insn)) {
+		((u16 *)addr)[0] = insn>>16;
+		((u16 *)addr)[1] = insn;
+		len = 2*sizeof(u16);
+	} else {
+		((u16 *)addr)[0] = insn;
+		len = sizeof(u16);
+	}
+	flush_insns(addr, len);
+
+#else /* !CONFIG_THUMB2_KERNEL */
 	*kp->addr = kp->opcode;
-	flush_insns(kp->addr, 1);
+	flush_insns(kp->addr, sizeof(kp->addr[0]));
+#endif
 	return 0;
 }
 
@@ -130,12 +218,24 @@ static void __kprobes set_current_kprobe(struct kprobe *p)
 	__get_cpu_var(current_kprobe) = p;
 }
 
-static void __kprobes singlestep(struct kprobe *p, struct pt_regs *regs,
+static void __kprobes
-				 struct kprobe_ctlblk *kcb)
+singlestep_skip(struct kprobe *p, struct pt_regs *regs)
 {
+#ifdef CONFIG_THUMB2_KERNEL
+	regs->ARM_cpsr = it_advance(regs->ARM_cpsr);
+	if (is_wide_instruction(p->opcode))
+		regs->ARM_pc += 4;
+	else
+		regs->ARM_pc += 2;
+#else
 	regs->ARM_pc += 4;
-	if (p->ainsn.insn_check_cc(regs->ARM_cpsr))
+#endif
-		p->ainsn.insn_handler(p, regs);
+}
+
+static inline void __kprobes
+singlestep(struct kprobe *p, struct pt_regs *regs, struct kprobe_ctlblk *kcb)
+{
+	p->ainsn.insn_singlestep(p, regs);
 }
 
 /*
@@ -149,11 +249,23 @@ void __kprobes kprobe_handler(struct pt_regs *regs)
 {
 	struct kprobe *p, *cur;
 	struct kprobe_ctlblk *kcb;
-	kprobe_opcode_t	*addr = (kprobe_opcode_t *)regs->ARM_pc;
 
 	kcb = get_kprobe_ctlblk();
 	cur = kprobe_running();
-	p = get_kprobe(addr);
+
+#ifdef CONFIG_THUMB2_KERNEL
+	/*
+	 * First look for a probe which was registered using an address with
+	 * bit 0 set, this is the usual situation for pointers to Thumb code.
+	 * If not found, fallback to looking for one with bit 0 clear.
+	 */
+	p = get_kprobe((kprobe_opcode_t *)(regs->ARM_pc | 1));
+	if (!p)
+		p = get_kprobe((kprobe_opcode_t *)regs->ARM_pc);
+
+#else /* ! CONFIG_THUMB2_KERNEL */
+	p = get_kprobe((kprobe_opcode_t *)regs->ARM_pc);
+#endif
 
 	if (p) {
 		if (cur) {
@@ -173,7 +285,8 @@ void __kprobes kprobe_handler(struct pt_regs *regs)
 				/* impossible cases */
 				BUG();
 			}
-		} else {
+		} else if (p->ainsn.insn_check_cc(regs->ARM_cpsr)) {
+			/* Probe hit and conditional execution check ok. */
 			set_current_kprobe(p);
 			kcb->kprobe_status = KPROBE_HIT_ACTIVE;
 
@@ -193,6 +306,13 @@ void __kprobes kprobe_handler(struct pt_regs *regs)
 				}
 				reset_current_kprobe();
 			}
+		} else {
+			/*
+			 * Probe hit but conditional execution check failed,
+			 * so just skip the instruction and continue as if
+			 * nothing had happened.
+			 */
+			singlestep_skip(p, regs);
 		}
 	} else if (cur) {
 		/* We probably hit a jprobe.  Call its break handler. */
@@ -300,7 +420,11 @@ void __naked __kprobes kretprobe_trampoline(void)
 		"bl	trampoline_handler	\n\t"
 		"mov	lr, r0			\n\t"
 		"ldmia	sp!, {r0 - r11}		\n\t"
+#ifdef CONFIG_THUMB2_KERNEL
+		"bx	lr			\n\t"
+#else
 		"mov	pc, lr			\n\t"
+#endif
 		: : : "memory");
 }
 
@@ -378,11 +502,22 @@ int __kprobes setjmp_pre_handler(struct kprobe *p, struct pt_regs *regs)
 	struct jprobe *jp = container_of(p, struct jprobe, kp);
 	struct kprobe_ctlblk *kcb = get_kprobe_ctlblk();
 	long sp_addr = regs->ARM_sp;
+	long cpsr;
 
 	kcb->jprobe_saved_regs = *regs;
 	memcpy(kcb->jprobes_stack, (void *)sp_addr, MIN_STACK_SIZE(sp_addr));
 	regs->ARM_pc = (long)jp->entry;
-	regs->ARM_cpsr |= PSR_I_BIT;
+
+	cpsr = regs->ARM_cpsr | PSR_I_BIT;
+#ifdef CONFIG_THUMB2_KERNEL
+	/* Set correct Thumb state in cpsr */
+	if (regs->ARM_pc & 1)
+		cpsr |= PSR_T_BIT;
+	else
+		cpsr &= ~PSR_T_BIT;
+#endif
+	regs->ARM_cpsr = cpsr;
+
 	preempt_disable();
 	return 1;
 }
@@ -404,7 +539,12 @@ void __kprobes jprobe_return(void)
 		 * This is to prevent any simulated instruction from writing
 		 * over the regs when they are accessing the stack.
 		 */
+#ifdef CONFIG_THUMB2_KERNEL
+		"sub    r0, %0, %1		\n\t"
+		"mov    sp, r0			\n\t"
+#else
 		"sub    sp, %0, %1		\n\t"
+#endif
 		"ldr    r0, ="__stringify(JPROBE_MAGIC_ADDR)"\n\t"
 		"str    %0, [sp, %2]		\n\t"
 		"str    r0, [sp, %3]		\n\t"
@@ -415,15 +555,28 @@ void __kprobes jprobe_return(void)
 		 * Return to the context saved by setjmp_pre_handler
 		 * and restored by longjmp_break_handler.
 		 */
+#ifdef CONFIG_THUMB2_KERNEL
+		"ldr	lr, [sp, %2]		\n\t" /* lr = saved sp */
+		"ldrd	r0, r1, [sp, %5]	\n\t" /* r0,r1 = saved lr,pc */
+		"ldr	r2, [sp, %4]		\n\t" /* r2 = saved psr */
+		"stmdb	lr!, {r0, r1, r2}	\n\t" /* push saved lr and */
+						      /* rfe context */
+		"ldmia	sp, {r0 - r12}		\n\t"
+		"mov	sp, lr			\n\t"
+		"ldr	lr, [sp], #4		\n\t"
+		"rfeia	sp!			\n\t"
+#else
 		"ldr	r0, [sp, %4]		\n\t"
 		"msr	cpsr_cxsf, r0		\n\t"
 		"ldmia	sp, {r0 - pc}		\n\t"
+#endif
 		:
 		: "r" (kcb->jprobe_saved_regs.ARM_sp),
 		  "I" (sizeof(struct pt_regs) * 2),
 		  "J" (offsetof(struct pt_regs, ARM_sp)),
 		  "J" (offsetof(struct pt_regs, ARM_pc)),
-		  "J" (offsetof(struct pt_regs, ARM_cpsr))
+		  "J" (offsetof(struct pt_regs, ARM_cpsr)),
+		  "J" (offsetof(struct pt_regs, ARM_lr))
 		: "memory", "cc");
 }
 
@@ -460,17 +613,44 @@ int __kprobes arch_trampoline_kprobe(struct kprobe *p)
 	return 0;
 }
 
-static struct undef_hook kprobes_break_hook = {
+#ifdef CONFIG_THUMB2_KERNEL
-	.instr_mask	= 0xffffffff,
+
-	.instr_val	= KPROBE_BREAKPOINT_INSTRUCTION,
+static struct undef_hook kprobes_thumb16_break_hook = {
+	.instr_mask	= 0xffff,
+	.instr_val	= KPROBE_THUMB16_BREAKPOINT_INSTRUCTION,
 	.cpsr_mask	= MODE_MASK,
 	.cpsr_val	= SVC_MODE,
 	.fn		= kprobe_trap_handler,
 };
 
+static struct undef_hook kprobes_thumb32_break_hook = {
+	.instr_mask	= 0xffffffff,
+	.instr_val	= KPROBE_THUMB32_BREAKPOINT_INSTRUCTION,
+	.cpsr_mask	= MODE_MASK,
+	.cpsr_val	= SVC_MODE,
+	.fn		= kprobe_trap_handler,
+};
+
+#else  /* !CONFIG_THUMB2_KERNEL */
+
+static struct undef_hook kprobes_arm_break_hook = {
+	.instr_mask	= 0x0fffffff,
+	.instr_val	= KPROBE_ARM_BREAKPOINT_INSTRUCTION,
+	.cpsr_mask	= MODE_MASK,
+	.cpsr_val	= SVC_MODE,
+	.fn		= kprobe_trap_handler,
+};
+
+#endif /* !CONFIG_THUMB2_KERNEL */
+
 int __init arch_init_kprobes()
 {
 	arm_kprobe_decode_init();
-	register_undef_hook(&kprobes_break_hook);
+#ifdef CONFIG_THUMB2_KERNEL
+	register_undef_hook(&kprobes_thumb16_break_hook);
+	register_undef_hook(&kprobes_thumb32_break_hook);
+#else
+	register_undef_hook(&kprobes_arm_break_hook);
+#endif
 	return 0;
 }

arch/arm/kernel/kprobes.h

@@ -0,0 +1,420 @@
+/*
+ * arch/arm/kernel/kprobes.h
+ *
+ * Copyright (C) 2011 Jon Medhurst <tixy@yxit.co.uk>.
+ *
+ * Some contents moved here from arch/arm/include/asm/kprobes.h which is
+ * Copyright (C) 2006, 2007 Motorola Inc.
+ *
+ * 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 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.
+ */
+
+#ifndef _ARM_KERNEL_KPROBES_H
+#define _ARM_KERNEL_KPROBES_H
+
+/*
+ * These undefined instructions must be unique and
+ * reserved solely for kprobes' use.
+ */
+#define KPROBE_ARM_BREAKPOINT_INSTRUCTION	0x07f001f8
+#define KPROBE_THUMB16_BREAKPOINT_INSTRUCTION	0xde18
+#define KPROBE_THUMB32_BREAKPOINT_INSTRUCTION	0xf7f0a018
+
+
+enum kprobe_insn {
+	INSN_REJECTED,
+	INSN_GOOD,
+	INSN_GOOD_NO_SLOT
+};
+
+typedef enum kprobe_insn (kprobe_decode_insn_t)(kprobe_opcode_t,
+						struct arch_specific_insn *);
+
+#ifdef CONFIG_THUMB2_KERNEL
+
+enum kprobe_insn thumb16_kprobe_decode_insn(kprobe_opcode_t,
+						struct arch_specific_insn *);
+enum kprobe_insn thumb32_kprobe_decode_insn(kprobe_opcode_t,
+						struct arch_specific_insn *);
+
+#else /* !CONFIG_THUMB2_KERNEL */
+
+enum kprobe_insn arm_kprobe_decode_insn(kprobe_opcode_t,
+					struct arch_specific_insn *);
+#endif
+
+void __init arm_kprobe_decode_init(void);
+
+extern kprobe_check_cc * const kprobe_condition_checks[16];
+
+
+#if __LINUX_ARM_ARCH__ >= 7
+
+/* str_pc_offset is architecturally defined from ARMv7 onwards */
+#define str_pc_offset 8
+#define find_str_pc_offset()
+
+#else /* __LINUX_ARM_ARCH__ < 7 */
+
+/* We need a run-time check to determine str_pc_offset */
+extern int str_pc_offset;
+void __init find_str_pc_offset(void);
+
+#endif
+
+
+/*
+ * Update ITSTATE after normal execution of an IT block instruction.
+ *
+ * The 8 IT state bits are split into two parts in CPSR:
+ *	ITSTATE<1:0> are in CPSR<26:25>
+ *	ITSTATE<7:2> are in CPSR<15:10>
+ */
+static inline unsigned long it_advance(unsigned long cpsr)
+	{
+	if ((cpsr & 0x06000400) == 0) {
+		/* ITSTATE<2:0> == 0 means end of IT block, so clear IT state */
+		cpsr &= ~PSR_IT_MASK;
+	} else {
+		/* We need to shift left ITSTATE<4:0> */
+		const unsigned long mask = 0x06001c00;  /* Mask ITSTATE<4:0> */
+		unsigned long it = cpsr & mask;
+		it <<= 1;
+		it |= it >> (27 - 10);  /* Carry ITSTATE<2> to correct place */
+		it &= mask;
+		cpsr &= ~mask;
+		cpsr |= it;
+	}
+	return cpsr;
+}
+
+static inline void __kprobes bx_write_pc(long pcv, struct pt_regs *regs)
+{
+	long cpsr = regs->ARM_cpsr;
+	if (pcv & 0x1) {
+		cpsr |= PSR_T_BIT;
+		pcv &= ~0x1;
+	} else {
+		cpsr &= ~PSR_T_BIT;
+		pcv &= ~0x2;	/* Avoid UNPREDICTABLE address allignment */
+	}
+	regs->ARM_cpsr = cpsr;
+	regs->ARM_pc = pcv;
+}
+
+
+#if __LINUX_ARM_ARCH__ >= 6
+
+/* Kernels built for >= ARMv6 should never run on <= ARMv5 hardware, so... */
+#define load_write_pc_interworks true
+#define test_load_write_pc_interworking()
+
+#else /* __LINUX_ARM_ARCH__ < 6 */
+
+/* We need run-time testing to determine if load_write_pc() should interwork. */
+extern bool load_write_pc_interworks;
+void __init test_load_write_pc_interworking(void);
+
+#endif
+
+static inline void __kprobes load_write_pc(long pcv, struct pt_regs *regs)
+{
+	if (load_write_pc_interworks)
+		bx_write_pc(pcv, regs);
+	else
+		regs->ARM_pc = pcv;
+}
+
+
+#if __LINUX_ARM_ARCH__ >= 7
+
+#define alu_write_pc_interworks true
+#define test_alu_write_pc_interworking()
+
+#elif __LINUX_ARM_ARCH__ <= 5
+
+/* Kernels built for <= ARMv5 should never run on >= ARMv6 hardware, so... */
+#define alu_write_pc_interworks false
+#define test_alu_write_pc_interworking()
+
+#else /* __LINUX_ARM_ARCH__ == 6 */
+
+/* We could be an ARMv6 binary on ARMv7 hardware so we need a run-time check. */
+extern bool alu_write_pc_interworks;
+void __init test_alu_write_pc_interworking(void);
+
+#endif /* __LINUX_ARM_ARCH__ == 6 */
+
+static inline void __kprobes alu_write_pc(long pcv, struct pt_regs *regs)
+{
+	if (alu_write_pc_interworks)
+		bx_write_pc(pcv, regs);
+	else
+		regs->ARM_pc = pcv;
+}
+
+
+void __kprobes kprobe_simulate_nop(struct kprobe *p, struct pt_regs *regs);
+void __kprobes kprobe_emulate_none(struct kprobe *p, struct pt_regs *regs);
+
+enum kprobe_insn __kprobes
+kprobe_decode_ldmstm(kprobe_opcode_t insn, struct arch_specific_insn *asi);
+
+/*
+ * Test if load/store instructions writeback the address register.
+ * if P (bit 24) == 0 or W (bit 21) == 1
+ */
+#define is_writeback(insn) ((insn ^ 0x01000000) & 0x01200000)
+
+/*
+ * The following definitions and macros are used to build instruction
+ * decoding tables for use by kprobe_decode_insn.
+ *
+ * These tables are a concatenation of entries each of which consist of one of
+ * the decode_* structs. All of the fields in every type of decode structure
+ * are of the union type decode_item, therefore the entire decode table can be
+ * viewed as an array of these and declared like:
+ *
+ *	static const union decode_item table_name[] = {};
+ *
+ * In order to construct each entry in the table, macros are used to
+ * initialise a number of sequential decode_item values in a layout which
+ * matches the relevant struct. E.g. DECODE_SIMULATE initialise a struct
+ * decode_simulate by initialising four decode_item objects like this...
+ *
+ *	{.bits = _type},
+ *	{.bits = _mask},
+ *	{.bits = _value},
+ *	{.handler = _handler},
+ *
+ * Initialising a specified member of the union means that the compiler
+ * will produce a warning if the argument is of an incorrect type.
+ *
+ * Below is a list of each of the macros used to initialise entries and a
+ * description of the action performed when that entry is matched to an
+ * instruction. A match is found when (instruction & mask) == value.
+ *
+ * DECODE_TABLE(mask, value, table)
+ *	Instruction decoding jumps to parsing the new sub-table 'table'.
+ *
+ * DECODE_CUSTOM(mask, value, decoder)
+ *	The custom function 'decoder' is called to the complete decoding
+ *	of an instruction.
+ *
+ * DECODE_SIMULATE(mask, value, handler)
+ *	Set the probes instruction handler to 'handler', this will be used
+ *	to simulate the instruction when the probe is hit. Decoding returns
+ *	with INSN_GOOD_NO_SLOT.
+ *
+ * DECODE_EMULATE(mask, value, handler)
+ *	Set the probes instruction handler to 'handler', this will be used
+ *	to emulate the instruction when the probe is hit. The modified
+ *	instruction (see below) is placed in the probes instruction slot so it
+ *	may be called by the emulation code. Decoding returns with INSN_GOOD.
+ *
+ * DECODE_REJECT(mask, value)
+ *	Instruction decoding fails with INSN_REJECTED
+ *
+ * DECODE_OR(mask, value)
+ *	This allows the mask/value test of multiple table entries to be
+ *	logically ORed. Once an 'or' entry is matched the decoding action to
+ *	be performed is that of the next entry which isn't an 'or'. E.g.
+ *
+ *		DECODE_OR	(mask1, value1)
+ *		DECODE_OR	(mask2, value2)
+ *		DECODE_SIMULATE	(mask3, value3, simulation_handler)
+ *
+ *	This means that if any of the three mask/value pairs match the
+ *	instruction being decoded, then 'simulation_handler' will be used
+ *	for it.
+ *
+ * Both the SIMULATE and EMULATE macros have a second form which take an
+ * additional 'regs' argument.
+ *
+ *	DECODE_SIMULATEX(mask, value, handler, regs)
+ *	DECODE_EMULATEX	(mask, value, handler, regs)
+ *
+ * These are used to specify what kind of CPU register is encoded in each of the
+ * least significant 5 nibbles of the instruction being decoded. The regs value
+ * is specified using the REGS macro, this takes any of the REG_TYPE_* values
+ * from enum decode_reg_type as arguments; only the '*' part of the name is
+ * given. E.g.
+ *
+ *	REGS(0, ANY, NOPC, 0, ANY)
+ *
+ * This indicates an instruction is encoded like:
+ *
+ *	bits 19..16	ignore
+ *	bits 15..12	any register allowed here
+ *	bits 11.. 8	any register except PC allowed here
+ *	bits  7.. 4	ignore
+ *	bits  3.. 0	any register allowed here
+ *
+ * This register specification is checked after a decode table entry is found to
+ * match an instruction (through the mask/value test). Any invalid register then
+ * found in the instruction will cause decoding to fail with INSN_REJECTED. In
+ * the above example this would happen if bits 11..8 of the instruction were
+ * 1111, indicating R15 or PC.
+ *
+ * As well as checking for legal combinations of registers, this data is also
+ * used to modify the registers encoded in the instructions so that an
+ * emulation routines can use it. (See decode_regs() and INSN_NEW_BITS.)
+ *
+ * Here is a real example which matches ARM instructions of the form
+ * "AND <Rd>,<Rn>,<Rm>,<shift> <Rs>"
+ *
+ *	DECODE_EMULATEX	(0x0e000090, 0x00000010, emulate_rd12rn16rm0rs8_rwflags,
+ *						 REGS(ANY, ANY, NOPC, 0, ANY)),
+ *						      ^    ^    ^        ^
+ *						      Rn   Rd   Rs       Rm
+ *
+ * Decoding the instruction "AND R4, R5, R6, ASL R15" will be rejected because
+ * Rs == R15
+ *
+ * Decoding the instruction "AND R4, R5, R6, ASL R7" will be accepted and the
+ * instruction will be modified to "AND R0, R2, R3, ASL R1" and then placed into
+ * the kprobes instruction slot. This can then be called later by the handler
+ * function emulate_rd12rn16rm0rs8_rwflags in order to simulate the instruction.
+ */
+
+enum decode_type {
+	DECODE_TYPE_END,
+	DECODE_TYPE_TABLE,
+	DECODE_TYPE_CUSTOM,
+	DECODE_TYPE_SIMULATE,
+	DECODE_TYPE_EMULATE,
+	DECODE_TYPE_OR,
+	DECODE_TYPE_REJECT,
+	NUM_DECODE_TYPES /* Must be last enum */
+};
+
+#define DECODE_TYPE_BITS	4
+#define DECODE_TYPE_MASK	((1 << DECODE_TYPE_BITS) - 1)
+
+enum decode_reg_type {
+	REG_TYPE_NONE = 0, /* Not a register, ignore */
+	REG_TYPE_ANY,	   /* Any register allowed */
+	REG_TYPE_SAMEAS16, /* Register should be same as that at bits 19..16 */
+	REG_TYPE_SP,	   /* Register must be SP */
+	REG_TYPE_PC,	   /* Register must be PC */
+	REG_TYPE_NOSP,	   /* Register must not be SP */
+	REG_TYPE_NOSPPC,   /* Register must not be SP or PC */
+	REG_TYPE_NOPC,	   /* Register must not be PC */
+	REG_TYPE_NOPCWB,   /* No PC if load/store write-back flag also set */
+
+	/* The following types are used when the encoding for PC indicates
+	 * another instruction form. This distiction only matters for test
+	 * case coverage checks.
+	 */
+	REG_TYPE_NOPCX,	   /* Register must not be PC */
+	REG_TYPE_NOSPPCX,  /* Register must not be SP or PC */
+
+	/* Alias to allow '0' arg to be used in REGS macro. */
+	REG_TYPE_0 = REG_TYPE_NONE
+};
+
+#define REGS(r16, r12, r8, r4, r0)	\
+	((REG_TYPE_##r16) << 16) +	\
+	((REG_TYPE_##r12) << 12) +	\
+	((REG_TYPE_##r8) << 8) +	\
+	((REG_TYPE_##r4) << 4) +	\
+	(REG_TYPE_##r0)
+
+union decode_item {
+	u32			bits;
+	const union decode_item	*table;
+	kprobe_insn_handler_t	*handler;
+	kprobe_decode_insn_t	*decoder;
+};
+
+
+#define DECODE_END			\
+	{.bits = DECODE_TYPE_END}
+
+
+struct decode_header {
+	union decode_item	type_regs;
+	union decode_item	mask;
+	union decode_item	value;
+};
+
+#define DECODE_HEADER(_type, _mask, _value, _regs)		\
+	{.bits = (_type) | ((_regs) << DECODE_TYPE_BITS)},	\
+	{.bits = (_mask)},					\
+	{.bits = (_value)}
+
+
+struct decode_table {
+	struct decode_header	header;
+	union decode_item	table;
+};
+
+#define DECODE_TABLE(_mask, _value, _table)			\
+	DECODE_HEADER(DECODE_TYPE_TABLE, _mask, _value, 0),	\
+	{.table = (_table)}
+
+
+struct decode_custom {
+	struct decode_header	header;
+	union decode_item	decoder;
+};
+
+#define DECODE_CUSTOM(_mask, _value, _decoder)			\
+	DECODE_HEADER(DECODE_TYPE_CUSTOM, _mask, _value, 0),	\
+	{.decoder = (_decoder)}
+
+
+struct decode_simulate {
+	struct decode_header	header;
+	union decode_item	handler;
+};
+
+#define DECODE_SIMULATEX(_mask, _value, _handler, _regs)		\
+	DECODE_HEADER(DECODE_TYPE_SIMULATE, _mask, _value, _regs),	\
+	{.handler = (_handler)}
+
+#define DECODE_SIMULATE(_mask, _value, _handler)	\
+	DECODE_SIMULATEX(_mask, _value, _handler, 0)
+
+
+struct decode_emulate {
+	struct decode_header	header;
+	union decode_item	handler;
+};
+
+#define DECODE_EMULATEX(_mask, _value, _handler, _regs)			\
+	DECODE_HEADER(DECODE_TYPE_EMULATE, _mask, _value, _regs),	\
+	{.handler = (_handler)}
+
+#define DECODE_EMULATE(_mask, _value, _handler)		\
+	DECODE_EMULATEX(_mask, _value, _handler, 0)
+
+
+struct decode_or {
+	struct decode_header	header;
+};
+
+#define DECODE_OR(_mask, _value)				\
+	DECODE_HEADER(DECODE_TYPE_OR, _mask, _value, 0)
+
+
+struct decode_reject {
+	struct decode_header	header;
+};
+
+#define DECODE_REJECT(_mask, _value)				\
+	DECODE_HEADER(DECODE_TYPE_REJECT, _mask, _value, 0)
+
+
+int kprobe_decode_insn(kprobe_opcode_t insn, struct arch_specific_insn *asi,
+			const union decode_item *table, bool thumb16);
+
+
+#endif /* _ARM_KERNEL_KPROBES_H */

arch/arm/kernel/perf_event.c

@@ -662,6 +662,12 @@ init_hw_perf_events(void)
 		case 0xC090:	/* Cortex-A9 */
 			armpmu = armv7_a9_pmu_init();
 			break;
+		case 0xC050:	/* Cortex-A5 */
+			armpmu = armv7_a5_pmu_init();
+			break;
+		case 0xC0F0:	/* Cortex-A15 */
+			armpmu = armv7_a15_pmu_init();
+			break;
 		}
 	/* Intel CPUs [xscale]. */
 	} else if (0x69 == implementor) {

arch/arm/kernel/perf_event_v7.c

@@ -17,17 +17,23 @@
  */
 
 #ifdef CONFIG_CPU_V7
-/* Common ARMv7 event types */
+/*
+ * Common ARMv7 event types
+ *
+ * Note: An implementation may not be able to count all of these events
+ * but the encodings are considered to be `reserved' in the case that
+ * they are not available.
+ */
 enum armv7_perf_types {
 	ARMV7_PERFCTR_PMNC_SW_INCR		= 0x00,
 	ARMV7_PERFCTR_IFETCH_MISS		= 0x01,
 	ARMV7_PERFCTR_ITLB_MISS			= 0x02,
-	ARMV7_PERFCTR_DCACHE_REFILL		= 0x03,
+	ARMV7_PERFCTR_DCACHE_REFILL		= 0x03,	/* L1 */
-	ARMV7_PERFCTR_DCACHE_ACCESS		= 0x04,
+	ARMV7_PERFCTR_DCACHE_ACCESS		= 0x04,	/* L1 */
 	ARMV7_PERFCTR_DTLB_REFILL		= 0x05,
 	ARMV7_PERFCTR_DREAD			= 0x06,
 	ARMV7_PERFCTR_DWRITE			= 0x07,
-
+	ARMV7_PERFCTR_INSTR_EXECUTED		= 0x08,
 	ARMV7_PERFCTR_EXC_TAKEN			= 0x09,
 	ARMV7_PERFCTR_EXC_EXECUTED		= 0x0A,
 	ARMV7_PERFCTR_CID_WRITE			= 0x0B,
@@ -39,21 +45,30 @@ enum armv7_perf_types {
 	 */
 	ARMV7_PERFCTR_PC_WRITE			= 0x0C,
 	ARMV7_PERFCTR_PC_IMM_BRANCH		= 0x0D,
+	ARMV7_PERFCTR_PC_PROC_RETURN		= 0x0E,
 	ARMV7_PERFCTR_UNALIGNED_ACCESS		= 0x0F,
+
+	/* These events are defined by the PMUv2 supplement (ARM DDI 0457A). */
 	ARMV7_PERFCTR_PC_BRANCH_MIS_PRED	= 0x10,
 	ARMV7_PERFCTR_CLOCK_CYCLES		= 0x11,
-
+	ARMV7_PERFCTR_PC_BRANCH_PRED		= 0x12,
-	ARMV7_PERFCTR_PC_BRANCH_MIS_USED	= 0x12,
+	ARMV7_PERFCTR_MEM_ACCESS		= 0x13,
+	ARMV7_PERFCTR_L1_ICACHE_ACCESS		= 0x14,
+	ARMV7_PERFCTR_L1_DCACHE_WB		= 0x15,
+	ARMV7_PERFCTR_L2_DCACHE_ACCESS		= 0x16,
+	ARMV7_PERFCTR_L2_DCACHE_REFILL		= 0x17,
+	ARMV7_PERFCTR_L2_DCACHE_WB		= 0x18,
+	ARMV7_PERFCTR_BUS_ACCESS		= 0x19,
+	ARMV7_PERFCTR_MEMORY_ERROR		= 0x1A,
+	ARMV7_PERFCTR_INSTR_SPEC		= 0x1B,
+	ARMV7_PERFCTR_TTBR_WRITE		= 0x1C,
+	ARMV7_PERFCTR_BUS_CYCLES		= 0x1D,
 
 	ARMV7_PERFCTR_CPU_CYCLES		= 0xFF
 };
 
 /* ARMv7 Cortex-A8 specific event types */
 enum armv7_a8_perf_types {
-	ARMV7_PERFCTR_INSTR_EXECUTED		= 0x08,
-
-	ARMV7_PERFCTR_PC_PROC_RETURN		= 0x0E,
-
 	ARMV7_PERFCTR_WRITE_BUFFER_FULL		= 0x40,
 	ARMV7_PERFCTR_L2_STORE_MERGED		= 0x41,
 	ARMV7_PERFCTR_L2_STORE_BUFF		= 0x42,
@@ -138,6 +153,39 @@ enum armv7_a9_perf_types {
 	ARMV7_PERFCTR_PLE_RQST_PROG		= 0xA5
 };
 
+/* ARMv7 Cortex-A5 specific event types */
+enum armv7_a5_perf_types {
+	ARMV7_PERFCTR_IRQ_TAKEN			= 0x86,
+	ARMV7_PERFCTR_FIQ_TAKEN			= 0x87,
+
+	ARMV7_PERFCTR_EXT_MEM_RQST		= 0xc0,
+	ARMV7_PERFCTR_NC_EXT_MEM_RQST		= 0xc1,
+	ARMV7_PERFCTR_PREFETCH_LINEFILL		= 0xc2,
+	ARMV7_PERFCTR_PREFETCH_LINEFILL_DROP	= 0xc3,
+	ARMV7_PERFCTR_ENTER_READ_ALLOC		= 0xc4,
+	ARMV7_PERFCTR_READ_ALLOC		= 0xc5,
+
+	ARMV7_PERFCTR_STALL_SB_FULL		= 0xc9,
+};
+
+/* ARMv7 Cortex-A15 specific event types */
+enum armv7_a15_perf_types {
+	ARMV7_PERFCTR_L1_DCACHE_READ_ACCESS	= 0x40,
+	ARMV7_PERFCTR_L1_DCACHE_WRITE_ACCESS	= 0x41,
+	ARMV7_PERFCTR_L1_DCACHE_READ_REFILL	= 0x42,
+	ARMV7_PERFCTR_L1_DCACHE_WRITE_REFILL	= 0x43,
+
+	ARMV7_PERFCTR_L1_DTLB_READ_REFILL	= 0x4C,
+	ARMV7_PERFCTR_L1_DTLB_WRITE_REFILL	= 0x4D,
+
+	ARMV7_PERFCTR_L2_DCACHE_READ_ACCESS	= 0x50,
+	ARMV7_PERFCTR_L2_DCACHE_WRITE_ACCESS	= 0x51,
+	ARMV7_PERFCTR_L2_DCACHE_READ_REFILL	= 0x52,
+	ARMV7_PERFCTR_L2_DCACHE_WRITE_REFILL	= 0x53,
+
+	ARMV7_PERFCTR_SPEC_PC_WRITE		= 0x76,
+};
+
 /*
  * Cortex-A8 HW events mapping
  *
@@ -207,11 +255,6 @@ static const unsigned armv7_a8_perf_cache_map[PERF_COUNT_HW_CACHE_MAX]
 		},
 	},
 	[C(DTLB)] = {
-		/*
-		 * Only ITLB misses and DTLB refills are supported.
-		 * If users want the DTLB refills misses a raw counter
-		 * must be used.
-		 */
 		[C(OP_READ)] = {
 			[C(RESULT_ACCESS)]	= CACHE_OP_UNSUPPORTED,
 			[C(RESULT_MISS)]	= ARMV7_PERFCTR_DTLB_REFILL,
@@ -323,11 +366,6 @@ static const unsigned armv7_a9_perf_cache_map[PERF_COUNT_HW_CACHE_MAX]
 		},
 	},
 	[C(DTLB)] = {
-		/*
-		 * Only ITLB misses and DTLB refills are supported.
-		 * If users want the DTLB refills misses a raw counter
-		 * must be used.
-		 */
 		[C(OP_READ)] = {
 			[C(RESULT_ACCESS)]	= CACHE_OP_UNSUPPORTED,
 			[C(RESULT_MISS)]	= ARMV7_PERFCTR_DTLB_REFILL,
@@ -373,6 +411,242 @@ static const unsigned armv7_a9_perf_cache_map[PERF_COUNT_HW_CACHE_MAX]
 	},
 };
 
+/*
+ * Cortex-A5 HW events mapping
+ */
+static const unsigned armv7_a5_perf_map[PERF_COUNT_HW_MAX] = {
+	[PERF_COUNT_HW_CPU_CYCLES]	    = ARMV7_PERFCTR_CPU_CYCLES,
+	[PERF_COUNT_HW_INSTRUCTIONS]	    = ARMV7_PERFCTR_INSTR_EXECUTED,
+	[PERF_COUNT_HW_CACHE_REFERENCES]    = HW_OP_UNSUPPORTED,
+	[PERF_COUNT_HW_CACHE_MISSES]	    = HW_OP_UNSUPPORTED,
+	[PERF_COUNT_HW_BRANCH_INSTRUCTIONS] = ARMV7_PERFCTR_PC_WRITE,
+	[PERF_COUNT_HW_BRANCH_MISSES]	    = ARMV7_PERFCTR_PC_BRANCH_MIS_PRED,
+	[PERF_COUNT_HW_BUS_CYCLES]	    = HW_OP_UNSUPPORTED,
+};
+
+static const unsigned armv7_a5_perf_cache_map[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)]
+					= ARMV7_PERFCTR_DCACHE_ACCESS,
+			[C(RESULT_MISS)]
+					= ARMV7_PERFCTR_DCACHE_REFILL,
+		},
+		[C(OP_WRITE)] = {
+			[C(RESULT_ACCESS)]
+					= ARMV7_PERFCTR_DCACHE_ACCESS,
+			[C(RESULT_MISS)]
+					= ARMV7_PERFCTR_DCACHE_REFILL,
+		},
+		[C(OP_PREFETCH)] = {
+			[C(RESULT_ACCESS)]
+					= ARMV7_PERFCTR_PREFETCH_LINEFILL,
+			[C(RESULT_MISS)]
+					= ARMV7_PERFCTR_PREFETCH_LINEFILL_DROP,
+		},
+	},
+	[C(L1I)] = {
+		[C(OP_READ)] = {
+			[C(RESULT_ACCESS)]	= ARMV7_PERFCTR_L1_ICACHE_ACCESS,
+			[C(RESULT_MISS)]	= ARMV7_PERFCTR_IFETCH_MISS,
+		},
+		[C(OP_WRITE)] = {
+			[C(RESULT_ACCESS)]	= ARMV7_PERFCTR_L1_ICACHE_ACCESS,
+			[C(RESULT_MISS)]	= ARMV7_PERFCTR_IFETCH_MISS,
+		},
+		/*
+		 * The prefetch counters don't differentiate between the I
+		 * side and the D side.
+		 */
+		[C(OP_PREFETCH)] = {
+			[C(RESULT_ACCESS)]
+					= ARMV7_PERFCTR_PREFETCH_LINEFILL,
+			[C(RESULT_MISS)]
+					= ARMV7_PERFCTR_PREFETCH_LINEFILL_DROP,
+		},
+	},
+	[C(LL)] = {
+		[C(OP_READ)] = {
+			[C(RESULT_ACCESS)]	= CACHE_OP_UNSUPPORTED,
+			[C(RESULT_MISS)]	= CACHE_OP_UNSUPPORTED,
+		},
+		[C(OP_WRITE)] = {
+			[C(RESULT_ACCESS)]	= CACHE_OP_UNSUPPORTED,
+			[C(RESULT_MISS)]	= CACHE_OP_UNSUPPORTED,
+		},
+		[C(OP_PREFETCH)] = {
+			[C(RESULT_ACCESS)]	= CACHE_OP_UNSUPPORTED,
+			[C(RESULT_MISS)]	= CACHE_OP_UNSUPPORTED,
+		},
+	},
+	[C(DTLB)] = {
+		[C(OP_READ)] = {
+			[C(RESULT_ACCESS)]	= CACHE_OP_UNSUPPORTED,
+			[C(RESULT_MISS)]	= ARMV7_PERFCTR_DTLB_REFILL,
+		},
+		[C(OP_WRITE)] = {
+			[C(RESULT_ACCESS)]	= CACHE_OP_UNSUPPORTED,
+			[C(RESULT_MISS)]	= ARMV7_PERFCTR_DTLB_REFILL,
+		},
+		[C(OP_PREFETCH)] = {
+			[C(RESULT_ACCESS)]	= CACHE_OP_UNSUPPORTED,
+			[C(RESULT_MISS)]	= CACHE_OP_UNSUPPORTED,
+		},
+	},
+	[C(ITLB)] = {
+		[C(OP_READ)] = {
+			[C(RESULT_ACCESS)]	= CACHE_OP_UNSUPPORTED,
+			[C(RESULT_MISS)]	= ARMV7_PERFCTR_ITLB_MISS,
+		},
+		[C(OP_WRITE)] = {
+			[C(RESULT_ACCESS)]	= CACHE_OP_UNSUPPORTED,
+			[C(RESULT_MISS)]	= ARMV7_PERFCTR_ITLB_MISS,
+		},
+		[C(OP_PREFETCH)] = {
+			[C(RESULT_ACCESS)]	= CACHE_OP_UNSUPPORTED,
+			[C(RESULT_MISS)]	= CACHE_OP_UNSUPPORTED,
+		},
+	},
+	[C(BPU)] = {
+		[C(OP_READ)] = {
+			[C(RESULT_ACCESS)]	= ARMV7_PERFCTR_PC_BRANCH_PRED,
+			[C(RESULT_MISS)]
+					= ARMV7_PERFCTR_PC_BRANCH_MIS_PRED,
+		},
+		[C(OP_WRITE)] = {
+			[C(RESULT_ACCESS)]	= ARMV7_PERFCTR_PC_BRANCH_PRED,
+			[C(RESULT_MISS)]
+					= ARMV7_PERFCTR_PC_BRANCH_MIS_PRED,
+		},
+		[C(OP_PREFETCH)] = {
+			[C(RESULT_ACCESS)]	= CACHE_OP_UNSUPPORTED,
+			[C(RESULT_MISS)]	= CACHE_OP_UNSUPPORTED,
+		},
+	},
+};
+
+/*
+ * Cortex-A15 HW events mapping
+ */
+static const unsigned armv7_a15_perf_map[PERF_COUNT_HW_MAX] = {
+	[PERF_COUNT_HW_CPU_CYCLES]	    = ARMV7_PERFCTR_CPU_CYCLES,
+	[PERF_COUNT_HW_INSTRUCTIONS]	    = ARMV7_PERFCTR_INSTR_EXECUTED,
+	[PERF_COUNT_HW_CACHE_REFERENCES]    = HW_OP_UNSUPPORTED,
+	[PERF_COUNT_HW_CACHE_MISSES]	    = HW_OP_UNSUPPORTED,
+	[PERF_COUNT_HW_BRANCH_INSTRUCTIONS] = ARMV7_PERFCTR_SPEC_PC_WRITE,
+	[PERF_COUNT_HW_BRANCH_MISSES]	    = ARMV7_PERFCTR_PC_BRANCH_MIS_PRED,
+	[PERF_COUNT_HW_BUS_CYCLES]	    = ARMV7_PERFCTR_BUS_CYCLES,
+};
+
+static const unsigned armv7_a15_perf_cache_map[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)]
+					= ARMV7_PERFCTR_L1_DCACHE_READ_ACCESS,
+			[C(RESULT_MISS)]
+					= ARMV7_PERFCTR_L1_DCACHE_READ_REFILL,
+		},
+		[C(OP_WRITE)] = {
+			[C(RESULT_ACCESS)]
+					= ARMV7_PERFCTR_L1_DCACHE_WRITE_ACCESS,
+			[C(RESULT_MISS)]
+					= ARMV7_PERFCTR_L1_DCACHE_WRITE_REFILL,
+		},
+		[C(OP_PREFETCH)] = {
+			[C(RESULT_ACCESS)]	= CACHE_OP_UNSUPPORTED,
+			[C(RESULT_MISS)]	= CACHE_OP_UNSUPPORTED,
+		},
+	},
+	[C(L1I)] = {
+		/*
+		 * Not all performance counters differentiate between read
+		 * and write accesses/misses so we're not always strictly
+		 * correct, but it's the best we can do. Writes and reads get
+		 * combined in these cases.
+		 */
+		[C(OP_READ)] = {
+			[C(RESULT_ACCESS)]	= ARMV7_PERFCTR_L1_ICACHE_ACCESS,
+			[C(RESULT_MISS)]	= ARMV7_PERFCTR_IFETCH_MISS,
+		},
+		[C(OP_WRITE)] = {
+			[C(RESULT_ACCESS)]	= ARMV7_PERFCTR_L1_ICACHE_ACCESS,
+			[C(RESULT_MISS)]	= ARMV7_PERFCTR_IFETCH_MISS,
+		},
+		[C(OP_PREFETCH)] = {
+			[C(RESULT_ACCESS)]	= CACHE_OP_UNSUPPORTED,
+			[C(RESULT_MISS)]	= CACHE_OP_UNSUPPORTED,
+		},
+	},
+	[C(LL)] = {
+		[C(OP_READ)] = {
+			[C(RESULT_ACCESS)]
+					= ARMV7_PERFCTR_L2_DCACHE_READ_ACCESS,
+			[C(RESULT_MISS)]
+					= ARMV7_PERFCTR_L2_DCACHE_READ_REFILL,
+		},
+		[C(OP_WRITE)] = {
+			[C(RESULT_ACCESS)]
+					= ARMV7_PERFCTR_L2_DCACHE_WRITE_ACCESS,
+			[C(RESULT_MISS)]
+					= ARMV7_PERFCTR_L2_DCACHE_WRITE_REFILL,
+		},
+		[C(OP_PREFETCH)] = {
+			[C(RESULT_ACCESS)]	= CACHE_OP_UNSUPPORTED,
+			[C(RESULT_MISS)]	= CACHE_OP_UNSUPPORTED,
+		},
+	},
+	[C(DTLB)] = {
+		[C(OP_READ)] = {
+			[C(RESULT_ACCESS)]	= CACHE_OP_UNSUPPORTED,
+			[C(RESULT_MISS)]
+					= ARMV7_PERFCTR_L1_DTLB_READ_REFILL,
+		},
+		[C(OP_WRITE)] = {
+			[C(RESULT_ACCESS)]	= CACHE_OP_UNSUPPORTED,
+			[C(RESULT_MISS)]
+					= ARMV7_PERFCTR_L1_DTLB_WRITE_REFILL,
+		},
+		[C(OP_PREFETCH)] = {
+			[C(RESULT_ACCESS)]	= CACHE_OP_UNSUPPORTED,
+			[C(RESULT_MISS)]	= CACHE_OP_UNSUPPORTED,
+		},
+	},
+	[C(ITLB)] = {
+		[C(OP_READ)] = {
+			[C(RESULT_ACCESS)]	= CACHE_OP_UNSUPPORTED,
+			[C(RESULT_MISS)]	= ARMV7_PERFCTR_ITLB_MISS,
+		},
+		[C(OP_WRITE)] = {
+			[C(RESULT_ACCESS)]	= CACHE_OP_UNSUPPORTED,
+			[C(RESULT_MISS)]	= ARMV7_PERFCTR_ITLB_MISS,
+		},
+		[C(OP_PREFETCH)] = {
+			[C(RESULT_ACCESS)]	= CACHE_OP_UNSUPPORTED,
+			[C(RESULT_MISS)]	= CACHE_OP_UNSUPPORTED,
+		},
+	},
+	[C(BPU)] = {
+		[C(OP_READ)] = {
+			[C(RESULT_ACCESS)]	= ARMV7_PERFCTR_PC_BRANCH_PRED,
+			[C(RESULT_MISS)]
+					= ARMV7_PERFCTR_PC_BRANCH_MIS_PRED,
+		},
+		[C(OP_WRITE)] = {
+			[C(RESULT_ACCESS)]	= ARMV7_PERFCTR_PC_BRANCH_PRED,
+			[C(RESULT_MISS)]
+					= ARMV7_PERFCTR_PC_BRANCH_MIS_PRED,
+		},
+		[C(OP_PREFETCH)] = {
+			[C(RESULT_ACCESS)]	= CACHE_OP_UNSUPPORTED,
+			[C(RESULT_MISS)]	= CACHE_OP_UNSUPPORTED,
+		},
+	},
+};
+
 /*
  * Perf Events counters
  */
@@ -905,6 +1179,26 @@ static const struct arm_pmu *__init armv7_a9_pmu_init(void)
 	armv7pmu.num_events	= armv7_read_num_pmnc_events();
 	return &armv7pmu;
 }
+
+static const struct arm_pmu *__init armv7_a5_pmu_init(void)
+{
+	armv7pmu.id		= ARM_PERF_PMU_ID_CA5;
+	armv7pmu.name		= "ARMv7 Cortex-A5";
+	armv7pmu.cache_map	= &armv7_a5_perf_cache_map;
+	armv7pmu.event_map	= &armv7_a5_perf_map;
+	armv7pmu.num_events	= armv7_read_num_pmnc_events();
+	return &armv7pmu;
+}
+
+static const struct arm_pmu *__init armv7_a15_pmu_init(void)
+{
+	armv7pmu.id		= ARM_PERF_PMU_ID_CA15;
+	armv7pmu.name		= "ARMv7 Cortex-A15";
+	armv7pmu.cache_map	= &armv7_a15_perf_cache_map;
+	armv7pmu.event_map	= &armv7_a15_perf_map;
+	armv7pmu.num_events	= armv7_read_num_pmnc_events();
+	return &armv7pmu;
+}
 #else
 static const struct arm_pmu *__init armv7_a8_pmu_init(void)
 {
@@ -915,4 +1209,14 @@ static const struct arm_pmu *__init armv7_a9_pmu_init(void)
 {
 	return NULL;
 }
+
+static const struct arm_pmu *__init armv7_a5_pmu_init(void)
+{
+	return NULL;
+}
+
+static const struct arm_pmu *__init armv7_a15_pmu_init(void)
+{
+	return NULL;
+}
 #endif	/* CONFIG_CPU_V7 */

arch/arm/kernel/ptrace.c

@@ -228,34 +228,12 @@ static struct undef_hook thumb_break_hook = {
 	.fn		= break_trap,
 };
 
-static int thumb2_break_trap(struct pt_regs *regs, unsigned int instr)
-{
-	unsigned int instr2;
-	void __user *pc;
-
-	/* Check the second half of the instruction.  */
-	pc = (void __user *)(instruction_pointer(regs) + 2);
-
-	if (processor_mode(regs) == SVC_MODE) {
-		instr2 = *(u16 *) pc;
-	} else {
-		get_user(instr2, (u16 __user *)pc);
-	}
-
-	if (instr2 == 0xa000) {
-		ptrace_break(current, regs);
-		return 0;
-	} else {
-		return 1;
-	}
-}
-
 static struct undef_hook thumb2_break_hook = {
-	.instr_mask	= 0xffff,
+	.instr_mask	= 0xffffffff,
-	.instr_val	= 0xf7f0,
+	.instr_val	= 0xf7f0a000,
 	.cpsr_mask	= PSR_T_BIT,
 	.cpsr_val	= PSR_T_BIT,
-	.fn		= thumb2_break_trap,
+	.fn		= break_trap,
 };
 
 static int __init ptrace_break_init(void)

arch/arm/kernel/setup.c

@@ -919,6 +919,12 @@ void __init setup_arch(char **cmdline_p)
 
 	tcm_init();
 
+#ifdef CONFIG_ZONE_DMA
+	if (mdesc->dma_zone_size) {
+		extern unsigned long arm_dma_zone_size;
+		arm_dma_zone_size = mdesc->dma_zone_size;
+	}
+#endif
 #ifdef CONFIG_MULTI_IRQ_HANDLER
 	handle_arch_irq = mdesc->handle_irq;
 #endif
@@ -980,6 +986,10 @@ static const char *hwcap_str[] = {
 	"neon",
 	"vfpv3",
 	"vfpv3d16",
+	"tls",
+	"vfpv4",
+	"idiva",
+	"idivt",
 	NULL
 };
 

arch/arm/kernel/traps.c

@@ -355,9 +355,24 @@ asmlinkage void __exception do_undefinstr(struct pt_regs *regs)
 	pc = (void __user *)instruction_pointer(regs);
 
 	if (processor_mode(regs) == SVC_MODE) {
-		instr = *(u32 *) pc;
+#ifdef CONFIG_THUMB2_KERNEL
+		if (thumb_mode(regs)) {
+			instr = ((u16 *)pc)[0];
+			if (is_wide_instruction(instr)) {
+				instr <<= 16;
+				instr |= ((u16 *)pc)[1];
+			}
+		} else
+#endif
+			instr = *(u32 *) pc;
 	} else if (thumb_mode(regs)) {
 		get_user(instr, (u16 __user *)pc);
+		if (is_wide_instruction(instr)) {
+			unsigned int instr2;
+			get_user(instr2, (u16 __user *)pc+1);
+			instr <<= 16;
+			instr |= instr2;
+		}
 	} else {
 		get_user(instr, (u32 __user *)pc);
 	}

arch/arm/mach-davinci/board-da830-evm.c

@@ -681,4 +681,5 @@ MACHINE_START(DAVINCI_DA830_EVM, "DaVinci DA830/OMAP-L137/AM17x EVM")
 	.init_irq	= cp_intc_init,
 	.timer		= &davinci_timer,
 	.init_machine	= da830_evm_init,
+	.dma_zone_size	= SZ_128M,
 MACHINE_END

arch/arm/mach-davinci/board-da850-evm.c

@@ -1261,4 +1261,5 @@ MACHINE_START(DAVINCI_DA850_EVM, "DaVinci DA850/OMAP-L138/AM18x EVM")
 	.init_irq	= cp_intc_init,
 	.timer		= &davinci_timer,
 	.init_machine	= da850_evm_init,
+	.dma_zone_size	= SZ_128M,
 MACHINE_END

arch/arm/mach-davinci/board-dm355-evm.c

@@ -356,4 +356,5 @@ MACHINE_START(DAVINCI_DM355_EVM, "DaVinci DM355 EVM")
 	.init_irq     = davinci_irq_init,
 	.timer	      = &davinci_timer,
 	.init_machine = dm355_evm_init,
+	.dma_zone_size	= SZ_128M,
 MACHINE_END

arch/arm/mach-davinci/board-dm355-leopard.c

@@ -275,4 +275,5 @@ MACHINE_START(DM355_LEOPARD, "DaVinci DM355 leopard")
 	.init_irq     = davinci_irq_init,
 	.timer	      = &davinci_timer,
 	.init_machine = dm355_leopard_init,
+	.dma_zone_size	= SZ_128M,
 MACHINE_END

arch/arm/mach-davinci/board-dm365-evm.c

@@ -617,5 +617,6 @@ MACHINE_START(DAVINCI_DM365_EVM, "DaVinci DM365 EVM")
 	.init_irq	= davinci_irq_init,
 	.timer		= &davinci_timer,
 	.init_machine	= dm365_evm_init,
+	.dma_zone_size	= SZ_128M,
 MACHINE_END
 

arch/arm/mach-davinci/board-dm644x-evm.c

@@ -717,4 +717,5 @@ MACHINE_START(DAVINCI_EVM, "DaVinci DM644x EVM")
 	.init_irq     = davinci_irq_init,
 	.timer	      = &davinci_timer,
 	.init_machine = davinci_evm_init,
+	.dma_zone_size	= SZ_128M,
 MACHINE_END

arch/arm/mach-davinci/board-dm646x-evm.c

@@ -802,6 +802,7 @@ MACHINE_START(DAVINCI_DM6467_EVM, "DaVinci DM646x EVM")
 	.init_irq     = davinci_irq_init,
 	.timer        = &davinci_timer,
 	.init_machine = evm_init,
+	.dma_zone_size	= SZ_128M,
 MACHINE_END
 
 MACHINE_START(DAVINCI_DM6467TEVM, "DaVinci DM6467T EVM")
@@ -810,5 +811,6 @@ MACHINE_START(DAVINCI_DM6467TEVM, "DaVinci DM6467T EVM")
 	.init_irq     = davinci_irq_init,
 	.timer        = &davinci_timer,
 	.init_machine = evm_init,
+	.dma_zone_size	= SZ_128M,
 MACHINE_END
 

arch/arm/mach-davinci/board-mityomapl138.c

@@ -570,4 +570,5 @@ MACHINE_START(MITYOMAPL138, "MityDSP-L138/MityARM-1808")
 	.init_irq	= cp_intc_init,
 	.timer		= &davinci_timer,
 	.init_machine	= mityomapl138_init,
+	.dma_zone_size	= SZ_128M,
 MACHINE_END

arch/arm/mach-davinci/board-neuros-osd2.c

@@ -277,4 +277,5 @@ MACHINE_START(NEUROS_OSD2, "Neuros OSD2")
 	.init_irq	= davinci_irq_init,
 	.timer		= &davinci_timer,
 	.init_machine = davinci_ntosd2_init,
+	.dma_zone_size	= SZ_128M,
 MACHINE_END

arch/arm/mach-davinci/board-omapl138-hawk.c

@@ -343,4 +343,5 @@ MACHINE_START(OMAPL138_HAWKBOARD, "AM18x/OMAP-L138 Hawkboard")
 	.init_irq	= cp_intc_init,
 	.timer		= &davinci_timer,
 	.init_machine	= omapl138_hawk_init,
+	.dma_zone_size	= SZ_128M,
 MACHINE_END

arch/arm/mach-davinci/board-sffsdr.c

@@ -156,4 +156,5 @@ MACHINE_START(SFFSDR, "Lyrtech SFFSDR")
 	.init_irq     = davinci_irq_init,
 	.timer	      = &davinci_timer,
 	.init_machine = davinci_sffsdr_init,
+	.dma_zone_size	= SZ_128M,
 MACHINE_END

arch/arm/mach-davinci/board-tnetv107x-evm.c

@@ -282,4 +282,5 @@ MACHINE_START(TNETV107X, "TNETV107X EVM")
 	.init_irq	= cp_intc_init,
 	.timer		= &davinci_timer,
 	.init_machine	= tnetv107x_evm_board_init,
+	.dma_zone_size	= SZ_128M,
 MACHINE_END

arch/arm/mach-davinci/include/mach/memory.h

@@ -41,11 +41,4 @@
  */
 #define CONSISTENT_DMA_SIZE (14<<20)
 
-/*
- * Restrict DMA-able region to workaround silicon bug.  The bug
- * restricts buffers available for DMA to video hardware to be
- * below 128M
- */
-#define ARM_DMA_ZONE_SIZE	SZ_128M
-
 #endif /* __ASM_ARCH_MEMORY_H */

arch/arm/mach-h720x/h7201-eval.c

@@ -33,4 +33,5 @@ MACHINE_START(H7201, "Hynix GMS30C7201")
 	.map_io		= h720x_map_io,
 	.init_irq	= h720x_init_irq,
 	.timer		= &h7201_timer,
+	.dma_zone_size	= SZ_256M,
 MACHINE_END

arch/arm/mach-h720x/h7202-eval.c

@@ -76,4 +76,5 @@ MACHINE_START(H7202, "Hynix HMS30C7202")
 	.init_irq	= h7202_init_irq,
 	.timer		= &h7202_timer,
 	.init_machine	= init_eval_h7202,
+	.dma_zone_size	= SZ_256M,
 MACHINE_END

arch/arm/mach-h720x/include/mach/memory.h

@@ -8,11 +8,4 @@
 #define __ASM_ARCH_MEMORY_H
 
 #define PLAT_PHYS_OFFSET	UL(0x40000000)
-/*
- * This is the maximum DMA address that can be DMAd to.
- * There should not be more than (0xd0000000 - 0xc0000000)
- * bytes of RAM.
- */
-#define ARM_DMA_ZONE_SIZE	SZ_256M
-
 #endif

arch/arm/mach-ixp4xx/avila-setup.c

@@ -169,6 +169,9 @@ MACHINE_START(AVILA, "Gateworks Avila Network Platform")
 	.timer		= &ixp4xx_timer,
 	.boot_params	= 0x0100,
 	.init_machine	= avila_init,
+#if defined(CONFIG_PCI)
+	.dma_zone_size	= SZ_64M,
+#endif
 MACHINE_END
 
  /*
@@ -184,6 +187,9 @@ MACHINE_START(LOFT, "Giant Shoulder Inc Loft board")
 	.timer		= &ixp4xx_timer,
 	.boot_params	= 0x0100,
 	.init_machine	= avila_init,
+#if defined(CONFIG_PCI)
+	.dma_zone_size	= SZ_64M,
+#endif
 MACHINE_END
 #endif
 

arch/arm/mach-ixp4xx/coyote-setup.c

@@ -114,6 +114,9 @@ MACHINE_START(ADI_COYOTE, "ADI Engineering Coyote")
 	.timer		= &ixp4xx_timer,
 	.boot_params	= 0x0100,
 	.init_machine	= coyote_init,
+#if defined(CONFIG_PCI)
+	.dma_zone_size	= SZ_64M,
+#endif
 MACHINE_END
 #endif
 

arch/arm/mach-ixp4xx/dsmg600-setup.c

@@ -284,4 +284,7 @@ MACHINE_START(DSMG600, "D-Link DSM-G600 RevA")
 	.init_irq	= ixp4xx_init_irq,
 	.timer          = &dsmg600_timer,
 	.init_machine	= dsmg600_init,
+#if defined(CONFIG_PCI)
+	.dma_zone_size	= SZ_64M,
+#endif
 MACHINE_END

arch/arm/mach-ixp4xx/fsg-setup.c

@@ -275,5 +275,8 @@ MACHINE_START(FSG, "Freecom FSG-3")
 	.timer		= &ixp4xx_timer,
 	.boot_params	= 0x0100,
 	.init_machine	= fsg_init,
+#if defined(CONFIG_PCI)
+	.dma_zone_size	= SZ_64M,
+#endif
 MACHINE_END
 

arch/arm/mach-ixp4xx/gateway7001-setup.c

@@ -101,5 +101,8 @@ MACHINE_START(GATEWAY7001, "Gateway 7001 AP")
 	.timer		= &ixp4xx_timer,
 	.boot_params	= 0x0100,
 	.init_machine	= gateway7001_init,
+#if defined(CONFIG_PCI)
+	.dma_zone_size	= SZ_64M,
+#endif
 MACHINE_END
 #endif

arch/arm/mach-ixp4xx/goramo_mlr.c

@@ -501,4 +501,7 @@ MACHINE_START(GORAMO_MLR, "MultiLink")
 	.timer		= &ixp4xx_timer,
 	.boot_params	= 0x0100,
 	.init_machine	= gmlr_init,
+#if defined(CONFIG_PCI)
+	.dma_zone_size	= SZ_64M,
+#endif
 MACHINE_END

arch/arm/mach-ixp4xx/gtwx5715-setup.c

@@ -169,6 +169,9 @@ MACHINE_START(GTWX5715, "Gemtek GTWX5715 (Linksys WRV54G)")
 	.timer		= &ixp4xx_timer,
 	.boot_params	= 0x0100,
 	.init_machine	= gtwx5715_init,
+#if defined(CONFIG_PCI)
+	.dma_zone_size	= SZ_64M,
+#endif
 MACHINE_END
 
 

arch/arm/mach-ixp4xx/include/mach/memory.h

@@ -14,8 +14,4 @@
  */
 #define PLAT_PHYS_OFFSET	UL(0x00000000)
 
-#ifdef CONFIG_PCI
-#define ARM_DMA_ZONE_SIZE	SZ_64M
-#endif
-
 #endif

arch/arm/mach-ixp4xx/ixdp425-setup.c

@@ -258,6 +258,9 @@ MACHINE_START(IXDP425, "Intel IXDP425 Development Platform")
 	.timer		= &ixp4xx_timer,
 	.boot_params	= 0x0100,
 	.init_machine	= ixdp425_init,
+#if defined(CONFIG_PCI)
+	.dma_zone_size	= SZ_64M,
+#endif
 MACHINE_END
 #endif
 
@@ -269,6 +272,9 @@ MACHINE_START(IXDP465, "Intel IXDP465 Development Platform")
 	.timer		= &ixp4xx_timer,
 	.boot_params	= 0x0100,
 	.init_machine	= ixdp425_init,
+#if defined(CONFIG_PCI)
+	.dma_zone_size	= SZ_64M,
+#endif
 MACHINE_END
 #endif
 
@@ -280,6 +286,9 @@ MACHINE_START(IXCDP1100, "Intel IXCDP1100 Development Platform")
 	.timer		= &ixp4xx_timer,
 	.boot_params	= 0x0100,
 	.init_machine	= ixdp425_init,
+#if defined(CONFIG_PCI)
+	.dma_zone_size	= SZ_64M,
+#endif
 MACHINE_END
 #endif
 
@@ -291,5 +300,8 @@ MACHINE_START(KIXRP435, "Intel KIXRP435 Reference Platform")
 	.timer		= &ixp4xx_timer,
 	.boot_params	= 0x0100,
 	.init_machine	= ixdp425_init,
+#if defined(CONFIG_PCI)
+	.dma_zone_size	= SZ_64M,
+#endif
 MACHINE_END
 #endif

arch/arm/mach-ixp4xx/nas100d-setup.c

@@ -319,4 +319,7 @@ MACHINE_START(NAS100D, "Iomega NAS 100d")
 	.init_irq	= ixp4xx_init_irq,
 	.timer          = &ixp4xx_timer,
 	.init_machine	= nas100d_init,
+#if defined(CONFIG_PCI)
+	.dma_zone_size	= SZ_64M,
+#endif
 MACHINE_END

arch/arm/mach-ixp4xx/nslu2-setup.c

@@ -305,4 +305,7 @@ MACHINE_START(NSLU2, "Linksys NSLU2")
 	.init_irq	= ixp4xx_init_irq,
 	.timer          = &nslu2_timer,
 	.init_machine	= nslu2_init,
+#if defined(CONFIG_PCI)
+	.dma_zone_size	= SZ_64M,
+#endif
 MACHINE_END

arch/arm/mach-ixp4xx/vulcan-setup.c

@@ -241,4 +241,7 @@ MACHINE_START(ARCOM_VULCAN, "Arcom/Eurotech Vulcan")
 	.timer		= &ixp4xx_timer,
 	.boot_params	= 0x0100,
 	.init_machine	= vulcan_init,
+#if defined(CONFIG_PCI)
+	.dma_zone_size	= SZ_64M,
+#endif
 MACHINE_END

arch/arm/mach-ixp4xx/wg302v2-setup.c

@@ -102,5 +102,8 @@ MACHINE_START(WG302V2, "Netgear WG302 v2 / WAG302 v2")
 	.timer		= &ixp4xx_timer,
 	.boot_params	= 0x0100,
 	.init_machine	= wg302v2_init,
+#if defined(CONFIG_PCI)
+	.dma_zone_size	= SZ_64M,
+#endif
 MACHINE_END
 #endif

arch/arm/mach-pxa/cm-x2xx.c

@@ -518,4 +518,7 @@ MACHINE_START(ARMCORE, "Compulab CM-X2XX")
 	.init_irq	= cmx2xx_init_irq,
 	.timer		= &pxa_timer,
 	.init_machine	= cmx2xx_init,
+#ifdef CONFIG_PCI
+	.dma_zone_size	= SZ_64M,
+#endif
 MACHINE_END

arch/arm/mach-pxa/include/mach/memory.h

@@ -17,8 +17,4 @@
  */
 #define PLAT_PHYS_OFFSET	UL(0xa0000000)
 
-#if defined(CONFIG_MACH_ARMCORE) && defined(CONFIG_PCI)
-#define ARM_DMA_ZONE_SIZE	SZ_64M
-#endif
-
 #endif

arch/arm/mach-realview/include/mach/memory.h

@@ -29,10 +29,6 @@
 #define PLAT_PHYS_OFFSET		UL(0x00000000)
 #endif
 
-#ifdef CONFIG_ZONE_DMA
-#define ARM_DMA_ZONE_SIZE	SZ_256M
-#endif
-
 #ifdef CONFIG_SPARSEMEM
 
 /*

arch/arm/mach-realview/realview_eb.c

@@ -470,4 +470,7 @@ MACHINE_START(REALVIEW_EB, "ARM-RealView EB")
 	.init_irq	= gic_init_irq,
 	.timer		= &realview_eb_timer,
 	.init_machine	= realview_eb_init,
+#ifdef CONFIG_ZONE_DMA
+	.dma_zone_size	= SZ_256M,
+#endif
 MACHINE_END

arch/arm/mach-realview/realview_pb1176.c

@@ -365,4 +365,7 @@ MACHINE_START(REALVIEW_PB1176, "ARM-RealView PB1176")
 	.init_irq	= gic_init_irq,
 	.timer		= &realview_pb1176_timer,
 	.init_machine	= realview_pb1176_init,
+#ifdef CONFIG_ZONE_DMA
+	.dma_zone_size	= SZ_256M,
+#endif
 MACHINE_END

arch/arm/mach-realview/realview_pb11mp.c

@@ -367,4 +367,7 @@ MACHINE_START(REALVIEW_PB11MP, "ARM-RealView PB11MPCore")
 	.init_irq	= gic_init_irq,
 	.timer		= &realview_pb11mp_timer,
 	.init_machine	= realview_pb11mp_init,
+#ifdef CONFIG_ZONE_DMA
+	.dma_zone_size	= SZ_256M,
+#endif
 MACHINE_END

arch/arm/mach-realview/realview_pba8.c

@@ -317,4 +317,7 @@ MACHINE_START(REALVIEW_PBA8, "ARM-RealView PB-A8")
 	.init_irq	= gic_init_irq,
 	.timer		= &realview_pba8_timer,
 	.init_machine	= realview_pba8_init,
+#ifdef CONFIG_ZONE_DMA
+	.dma_zone_size	= SZ_256M,
+#endif
 MACHINE_END

arch/arm/mach-realview/realview_pbx.c

@@ -400,4 +400,7 @@ MACHINE_START(REALVIEW_PBX, "ARM-RealView PBX")
 	.init_irq	= gic_init_irq,
 	.timer		= &realview_pbx_timer,
 	.init_machine	= realview_pbx_init,
+#ifdef CONFIG_ZONE_DMA
+	.dma_zone_size	= SZ_256M,
+#endif
 MACHINE_END

arch/arm/mach-sa1100/assabet.c

@@ -453,4 +453,7 @@ MACHINE_START(ASSABET, "Intel-Assabet")
 	.init_irq	= sa1100_init_irq,
 	.timer		= &sa1100_timer,
 	.init_machine	= assabet_init,
+#ifdef CONFIG_SA1111
+	.dma_zone_size	= SZ_1M,
+#endif
 MACHINE_END

arch/arm/mach-sa1100/badge4.c

@@ -306,4 +306,7 @@ MACHINE_START(BADGE4, "Hewlett-Packard Laboratories BadgePAD 4")
 	.map_io		= badge4_map_io,
 	.init_irq	= sa1100_init_irq,
 	.timer		= &sa1100_timer,
+#ifdef CONFIG_SA1111
+	.dma_zone_size	= SZ_1M,
+#endif
 MACHINE_END

arch/arm/mach-sa1100/include/mach/memory.h

@@ -14,10 +14,6 @@
  */
 #define PLAT_PHYS_OFFSET	UL(0xc0000000)
 
-#ifdef CONFIG_SA1111
-#define ARM_DMA_ZONE_SIZE	SZ_1M
-#endif
-
 /*
  * Because of the wide memory address space between physical RAM banks on the
  * SA1100, it's much convenient to use Linux's SparseMEM support to implement

arch/arm/mach-sa1100/jornada720.c

@@ -369,4 +369,7 @@ MACHINE_START(JORNADA720, "HP Jornada 720")
 	.init_irq	= sa1100_init_irq,
 	.timer		= &sa1100_timer,
 	.init_machine	= jornada720_mach_init,
+#ifdef CONFIG_SA1111
+	.dma_zone_size	= SZ_1M,
+#endif
 MACHINE_END

arch/arm/mach-shark/core.c

@@ -156,4 +156,5 @@ MACHINE_START(SHARK, "Shark")
 	.map_io		= shark_map_io,
 	.init_irq	= shark_init_irq,
 	.timer		= &shark_timer,
+	.dma_zone_size	= SZ_4M,
 MACHINE_END

arch/arm/mach-shark/include/mach/memory.h

@@ -17,8 +17,6 @@
  */
 #define PLAT_PHYS_OFFSET     UL(0x08000000)
 
-#define ARM_DMA_ZONE_SIZE	SZ_4M
-
 /*
  * Cache flushing area
  */

arch/arm/mm/cache-fa.S

@@ -242,16 +242,5 @@ ENDPROC(fa_dma_unmap_area)
 
 	__INITDATA
 
-	.type	fa_cache_fns, #object
+	@ define struct cpu_cache_fns (see <asm/cacheflush.h> and proc-macros.S)
-ENTRY(fa_cache_fns)
+	define_cache_functions fa
-	.long	fa_flush_icache_all
-	.long	fa_flush_kern_cache_all
-	.long	fa_flush_user_cache_all
-	.long	fa_flush_user_cache_range
-	.long	fa_coherent_kern_range
-	.long	fa_coherent_user_range
-	.long	fa_flush_kern_dcache_area
-	.long	fa_dma_map_area
-	.long	fa_dma_unmap_area
-	.long	fa_dma_flush_range
-	.size	fa_cache_fns, . - fa_cache_fns

arch/arm/mm/cache-v3.S

@@ -129,16 +129,5 @@ ENDPROC(v3_dma_map_area)
 
 	__INITDATA
 
-	.type	v3_cache_fns, #object
+	@ define struct cpu_cache_fns (see <asm/cacheflush.h> and proc-macros.S)
-ENTRY(v3_cache_fns)
+	define_cache_functions v3
-	.long	v3_flush_icache_all
-	.long	v3_flush_kern_cache_all
-	.long	v3_flush_user_cache_all
-	.long	v3_flush_user_cache_range
-	.long	v3_coherent_kern_range
-	.long	v3_coherent_user_range
-	.long	v3_flush_kern_dcache_area
-	.long	v3_dma_map_area
-	.long	v3_dma_unmap_area
-	.long	v3_dma_flush_range
-	.size	v3_cache_fns, . - v3_cache_fns

arch/arm/mm/cache-v4.S

@@ -141,16 +141,5 @@ ENDPROC(v4_dma_map_area)
 
 	__INITDATA
 
-	.type	v4_cache_fns, #object
+	@ define struct cpu_cache_fns (see <asm/cacheflush.h> and proc-macros.S)
-ENTRY(v4_cache_fns)
+	define_cache_functions v4
-	.long	v4_flush_icache_all
-	.long	v4_flush_kern_cache_all
-	.long	v4_flush_user_cache_all
-	.long	v4_flush_user_cache_range
-	.long	v4_coherent_kern_range
-	.long	v4_coherent_user_range
-	.long	v4_flush_kern_dcache_area
-	.long	v4_dma_map_area
-	.long	v4_dma_unmap_area
-	.long	v4_dma_flush_range
-	.size	v4_cache_fns, . - v4_cache_fns

arch/arm/mm/cache-v4wb.S

@@ -253,16 +253,5 @@ ENDPROC(v4wb_dma_unmap_area)
 
 	__INITDATA
 
-	.type	v4wb_cache_fns, #object
+	@ define struct cpu_cache_fns (see <asm/cacheflush.h> and proc-macros.S)
-ENTRY(v4wb_cache_fns)
+	define_cache_functions v4wb
-	.long	v4wb_flush_icache_all
-	.long	v4wb_flush_kern_cache_all
-	.long	v4wb_flush_user_cache_all
-	.long	v4wb_flush_user_cache_range
-	.long	v4wb_coherent_kern_range
-	.long	v4wb_coherent_user_range
-	.long	v4wb_flush_kern_dcache_area
-	.long	v4wb_dma_map_area
-	.long	v4wb_dma_unmap_area
-	.long	v4wb_dma_flush_range
-	.size	v4wb_cache_fns, . - v4wb_cache_fns

arch/arm/mm/cache-v4wt.S

@@ -197,16 +197,5 @@ ENDPROC(v4wt_dma_map_area)
 
 	__INITDATA
 
-	.type	v4wt_cache_fns, #object
+	@ define struct cpu_cache_fns (see <asm/cacheflush.h> and proc-macros.S)
-ENTRY(v4wt_cache_fns)
+	define_cache_functions v4wt
-	.long	v4wt_flush_icache_all
-	.long	v4wt_flush_kern_cache_all
-	.long	v4wt_flush_user_cache_all
-	.long	v4wt_flush_user_cache_range
-	.long	v4wt_coherent_kern_range
-	.long	v4wt_coherent_user_range
-	.long	v4wt_flush_kern_dcache_area
-	.long	v4wt_dma_map_area
-	.long	v4wt_dma_unmap_area
-	.long	v4wt_dma_flush_range
-	.size	v4wt_cache_fns, . - v4wt_cache_fns

arch/arm/mm/cache-v6.S

@@ -330,16 +330,5 @@ ENDPROC(v6_dma_unmap_area)
 
 	__INITDATA
 
-	.type	v6_cache_fns, #object
+	@ define struct cpu_cache_fns (see <asm/cacheflush.h> and proc-macros.S)
-ENTRY(v6_cache_fns)
+	define_cache_functions v6
-	.long	v6_flush_icache_all
-	.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_coherent_user_range
-	.long	v6_flush_kern_dcache_area
-	.long	v6_dma_map_area
-	.long	v6_dma_unmap_area
-	.long	v6_dma_flush_range
-	.size	v6_cache_fns, . - v6_cache_fns

arch/arm/mm/cache-v7.S

@@ -325,16 +325,5 @@ ENDPROC(v7_dma_unmap_area)
 
 	__INITDATA
 
-	.type	v7_cache_fns, #object
+	@ define struct cpu_cache_fns (see <asm/cacheflush.h> and proc-macros.S)
-ENTRY(v7_cache_fns)
+	define_cache_functions v7
-	.long	v7_flush_icache_all
-	.long	v7_flush_kern_cache_all
-	.long	v7_flush_user_cache_all
-	.long	v7_flush_user_cache_range
-	.long	v7_coherent_kern_range
-	.long	v7_coherent_user_range
-	.long	v7_flush_kern_dcache_area
-	.long	v7_dma_map_area
-	.long	v7_dma_unmap_area
-	.long	v7_dma_flush_range
-	.size	v7_cache_fns, . - v7_cache_fns

arch/arm/mm/init.c

@@ -212,6 +212,10 @@ static void __init arm_bootmem_init(unsigned long start_pfn,
 }
 
 #ifdef CONFIG_ZONE_DMA
+
+unsigned long arm_dma_zone_size __read_mostly;
+EXPORT_SYMBOL(arm_dma_zone_size);
+
 /*
  * The DMA mask corresponding to the maximum bus address allocatable
  * using GFP_DMA.  The default here places no restriction on DMA
@@ -275,19 +279,17 @@ static void __init arm_bootmem_free(unsigned long min, unsigned long max_low,
 #endif
 	}
 
-#ifdef ARM_DMA_ZONE_SIZE
+#ifdef CONFIG_ZONE_DMA
-#ifndef CONFIG_ZONE_DMA
-#error ARM_DMA_ZONE_SIZE set but no DMA zone to limit allocations
-#endif
-
 	/*
 	 * Adjust the sizes according to any special requirements for
 	 * this machine type.
 	 */
-	arm_adjust_dma_zone(zone_size, zhole_size,
+	if (arm_dma_zone_size) {
-		ARM_DMA_ZONE_SIZE >> PAGE_SHIFT);
+		arm_adjust_dma_zone(zone_size, zhole_size,
-
+			arm_dma_zone_size >> PAGE_SHIFT);
-	arm_dma_limit = PHYS_OFFSET + ARM_DMA_ZONE_SIZE - 1;
+		arm_dma_limit = PHYS_OFFSET + arm_dma_zone_size - 1;
+	} else
+		arm_dma_limit = 0xffffffff;
 #endif
 
 	free_area_init_node(0, zone_size, min, zhole_size);

arch/arm/mm/proc-arm1020.S

@@ -364,17 +364,8 @@ ENTRY(arm1020_dma_unmap_area)
 	mov	pc, lr
 ENDPROC(arm1020_dma_unmap_area)
 
-ENTRY(arm1020_cache_fns)
+	@ define struct cpu_cache_fns (see <asm/cacheflush.h> and proc-macros.S)
-	.long	arm1020_flush_icache_all
+	define_cache_functions arm1020
-	.long	arm1020_flush_kern_cache_all
-	.long	arm1020_flush_user_cache_all
-	.long	arm1020_flush_user_cache_range
-	.long	arm1020_coherent_kern_range
-	.long	arm1020_coherent_user_range
-	.long	arm1020_flush_kern_dcache_area
-	.long	arm1020_dma_map_area
-	.long	arm1020_dma_unmap_area
-	.long	arm1020_dma_flush_range
 
 	.align	5
 ENTRY(cpu_arm1020_dcache_clean_area)
@@ -477,38 +468,14 @@ arm1020_crval:
 	crval	clear=0x0000593f, mmuset=0x00003935, ucset=0x00001930
 
 	__INITDATA
+	@ define struct processor (see <asm/proc-fns.h> and proc-macros.S)
+	define_processor_functions arm1020, dabort=v4t_early_abort, pabort=legacy_pabort
 
-/*
- * Purpose : Function pointers used to access above functions - all calls
- *	     come through these
- */
-	.type	arm1020_processor_functions, #object
-arm1020_processor_functions:
-	.word	v4t_early_abort
-	.word	legacy_pabort
-	.word	cpu_arm1020_proc_init
-	.word	cpu_arm1020_proc_fin
-	.word	cpu_arm1020_reset
-	.word	cpu_arm1020_do_idle
-	.word	cpu_arm1020_dcache_clean_area
-	.word	cpu_arm1020_switch_mm
-	.word	cpu_arm1020_set_pte_ext
-	.word	0
-	.word	0
-	.word	0
-	.size	arm1020_processor_functions, . - arm1020_processor_functions
 
 	.section ".rodata"
 
-	.type	cpu_arch_name, #object
+	string	cpu_arch_name, "armv5t"
-cpu_arch_name:
+	string	cpu_elf_name, "v5"
-	.asciz	"armv5t"
-	.size	cpu_arch_name, . - cpu_arch_name
-
-	.type	cpu_elf_name, #object
-cpu_elf_name:
-	.asciz	"v5"
-	.size	cpu_elf_name, . - cpu_elf_name
 
 	.type	cpu_arm1020_name, #object
 cpu_arm1020_name:

arch/arm/mm/proc-arm1020e.S

@@ -350,17 +350,8 @@ ENTRY(arm1020e_dma_unmap_area)
 	mov	pc, lr
 ENDPROC(arm1020e_dma_unmap_area)
 
-ENTRY(arm1020e_cache_fns)
+	@ define struct cpu_cache_fns (see <asm/cacheflush.h> and proc-macros.S)
-	.long	arm1020e_flush_icache_all
+	define_cache_functions arm1020e
-	.long	arm1020e_flush_kern_cache_all
-	.long	arm1020e_flush_user_cache_all
-	.long	arm1020e_flush_user_cache_range
-	.long	arm1020e_coherent_kern_range
-	.long	arm1020e_coherent_user_range
-	.long	arm1020e_flush_kern_dcache_area
-	.long	arm1020e_dma_map_area
-	.long	arm1020e_dma_unmap_area
-	.long	arm1020e_dma_flush_range
 
 	.align	5
 ENTRY(cpu_arm1020e_dcache_clean_area)
@@ -458,43 +449,14 @@ arm1020e_crval:
 	crval	clear=0x00007f3f, mmuset=0x00003935, ucset=0x00001930
 
 	__INITDATA
-
+	@ define struct processor (see <asm/proc-fns.h> and proc-macros.S)
-/*
+	define_processor_functions arm1020e, dabort=v4t_early_abort, pabort=legacy_pabort
- * Purpose : Function pointers used to access above functions - all calls
- *	     come through these
- */
-	.type	arm1020e_processor_functions, #object
-arm1020e_processor_functions:
-	.word	v4t_early_abort
-	.word	legacy_pabort
-	.word	cpu_arm1020e_proc_init
-	.word	cpu_arm1020e_proc_fin
-	.word	cpu_arm1020e_reset
-	.word	cpu_arm1020e_do_idle
-	.word	cpu_arm1020e_dcache_clean_area
-	.word	cpu_arm1020e_switch_mm
-	.word	cpu_arm1020e_set_pte_ext
-	.word	0
-	.word	0
-	.word	0
-	.size	arm1020e_processor_functions, . - arm1020e_processor_functions
 
 	.section ".rodata"
 
-	.type	cpu_arch_name, #object
+	string	cpu_arch_name, "armv5te"
-cpu_arch_name:
+	string	cpu_elf_name, "v5"
-	.asciz	"armv5te"
+	string	cpu_arm1020e_name, "ARM1020E"
-	.size	cpu_arch_name, . - cpu_arch_name
-
-	.type	cpu_elf_name, #object
-cpu_elf_name:
-	.asciz	"v5"
-	.size	cpu_elf_name, . - cpu_elf_name
-
-	.type	cpu_arm1020e_name, #object
-cpu_arm1020e_name:
-	.asciz	"ARM1020E"
-	.size	cpu_arm1020e_name, . - cpu_arm1020e_name
 
 	.align
 

arch/arm/mm/proc-arm1022.S

@@ -339,17 +339,8 @@ ENTRY(arm1022_dma_unmap_area)
 	mov	pc, lr
 ENDPROC(arm1022_dma_unmap_area)
 
-ENTRY(arm1022_cache_fns)
+	@ define struct cpu_cache_fns (see <asm/cacheflush.h> and proc-macros.S)
-	.long	arm1022_flush_icache_all
+	define_cache_functions arm1022
-	.long	arm1022_flush_kern_cache_all
-	.long	arm1022_flush_user_cache_all
-	.long	arm1022_flush_user_cache_range
-	.long	arm1022_coherent_kern_range
-	.long	arm1022_coherent_user_range
-	.long	arm1022_flush_kern_dcache_area
-	.long	arm1022_dma_map_area
-	.long	arm1022_dma_unmap_area
-	.long	arm1022_dma_flush_range
 
 	.align	5
 ENTRY(cpu_arm1022_dcache_clean_area)
@@ -441,43 +432,14 @@ arm1022_crval:
 	crval	clear=0x00007f3f, mmuset=0x00003935, ucset=0x00001930
 
 	__INITDATA
-
+	@ define struct processor (see <asm/proc-fns.h> and proc-macros.S)
-/*
+	define_processor_functions arm1022, dabort=v4t_early_abort, pabort=legacy_pabort
- * Purpose : Function pointers used to access above functions - all calls
- *	     come through these
- */
-	.type	arm1022_processor_functions, #object
-arm1022_processor_functions:
-	.word	v4t_early_abort
-	.word	legacy_pabort
-	.word	cpu_arm1022_proc_init
-	.word	cpu_arm1022_proc_fin
-	.word	cpu_arm1022_reset
-	.word	cpu_arm1022_do_idle
-	.word	cpu_arm1022_dcache_clean_area
-	.word	cpu_arm1022_switch_mm
-	.word	cpu_arm1022_set_pte_ext
-	.word	0
-	.word	0
-	.word	0
-	.size	arm1022_processor_functions, . - arm1022_processor_functions
 
 	.section ".rodata"
 
-	.type	cpu_arch_name, #object
+	string	cpu_arch_name, "armv5te"
-cpu_arch_name:
+	string	cpu_elf_name, "v5"
-	.asciz	"armv5te"
+	string	cpu_arm1022_name, "ARM1022"
-	.size	cpu_arch_name, . - cpu_arch_name
-
-	.type	cpu_elf_name, #object
-cpu_elf_name:
-	.asciz	"v5"
-	.size	cpu_elf_name, . - cpu_elf_name
-
-	.type	cpu_arm1022_name, #object
-cpu_arm1022_name:
-	.asciz	"ARM1022"
-	.size	cpu_arm1022_name, . - cpu_arm1022_name
 
 	.align
 

arch/arm/mm/proc-arm1026.S

@@ -333,17 +333,8 @@ ENTRY(arm1026_dma_unmap_area)
 	mov	pc, lr
 ENDPROC(arm1026_dma_unmap_area)
 
-ENTRY(arm1026_cache_fns)
+	@ define struct cpu_cache_fns (see <asm/cacheflush.h> and proc-macros.S)
-	.long	arm1026_flush_icache_all
+	define_cache_functions arm1026
-	.long	arm1026_flush_kern_cache_all
-	.long	arm1026_flush_user_cache_all
-	.long	arm1026_flush_user_cache_range
-	.long	arm1026_coherent_kern_range
-	.long	arm1026_coherent_user_range
-	.long	arm1026_flush_kern_dcache_area
-	.long	arm1026_dma_map_area
-	.long	arm1026_dma_unmap_area
-	.long	arm1026_dma_flush_range
 
 	.align	5
 ENTRY(cpu_arm1026_dcache_clean_area)
@@ -436,45 +427,15 @@ arm1026_crval:
 	crval	clear=0x00007f3f, mmuset=0x00003935, ucset=0x00001934
 
 	__INITDATA
-
+	@ define struct processor (see <asm/proc-fns.h> and proc-macros.S)
-/*
+	define_processor_functions arm1026, dabort=v5t_early_abort, pabort=legacy_pabort
- * Purpose : Function pointers used to access above functions - all calls
- *	     come through these
- */
-	.type	arm1026_processor_functions, #object
-arm1026_processor_functions:
-	.word	v5t_early_abort
-	.word	legacy_pabort
-	.word	cpu_arm1026_proc_init
-	.word	cpu_arm1026_proc_fin
-	.word	cpu_arm1026_reset
-	.word	cpu_arm1026_do_idle
-	.word	cpu_arm1026_dcache_clean_area
-	.word	cpu_arm1026_switch_mm
-	.word	cpu_arm1026_set_pte_ext
-	.word	0
-	.word	0
-	.word	0
-	.size	arm1026_processor_functions, . - arm1026_processor_functions
 
 	.section .rodata
 
-	.type	cpu_arch_name, #object
+	string	cpu_arch_name, "armv5tej"
-cpu_arch_name:
+	string	cpu_elf_name, "v5"
-	.asciz	"armv5tej"
-	.size	cpu_arch_name, . - cpu_arch_name
-
-	.type	cpu_elf_name, #object
-cpu_elf_name:
-	.asciz	"v5"
-	.size	cpu_elf_name, . - cpu_elf_name
 	.align
-
+	string	cpu_arm1026_name, "ARM1026EJ-S"
-	.type	cpu_arm1026_name, #object
-cpu_arm1026_name:
-	.asciz	"ARM1026EJ-S"
-	.size	cpu_arm1026_name, . - cpu_arm1026_name
-
 	.align
 
 	.section ".proc.info.init", #alloc, #execinstr

arch/arm/mm/proc-arm6_7.S

@@ -267,159 +267,57 @@ __arm7_setup:	mov	r0, #0
 
 		__INITDATA
 
-/*
+		@ define struct processor (see <asm/proc-fns.h> and proc-macros.S)
- * Purpose : Function pointers used to access above functions - all calls
+		define_processor_functions arm6, dabort=cpu_arm6_data_abort, pabort=legacy_pabort
- *	     come through these
+		define_processor_functions arm7, dabort=cpu_arm7_data_abort, pabort=legacy_pabort
- */
-		.type	arm6_processor_functions, #object
-ENTRY(arm6_processor_functions)
-		.word	cpu_arm6_data_abort
-		.word	legacy_pabort
-		.word	cpu_arm6_proc_init
-		.word	cpu_arm6_proc_fin
-		.word	cpu_arm6_reset
-		.word	cpu_arm6_do_idle
-		.word	cpu_arm6_dcache_clean_area
-		.word	cpu_arm6_switch_mm
-		.word	cpu_arm6_set_pte_ext
-		.word	0
-		.word	0
-		.word	0
-		.size	arm6_processor_functions, . - arm6_processor_functions
-
-/*
- * Purpose : Function pointers used to access above functions - all calls
- *	     come through these
- */
-		.type	arm7_processor_functions, #object
-ENTRY(arm7_processor_functions)
-		.word	cpu_arm7_data_abort
-		.word	legacy_pabort
-		.word	cpu_arm7_proc_init
-		.word	cpu_arm7_proc_fin
-		.word	cpu_arm7_reset
-		.word	cpu_arm7_do_idle
-		.word	cpu_arm7_dcache_clean_area
-		.word	cpu_arm7_switch_mm
-		.word	cpu_arm7_set_pte_ext
-		.word	0
-		.word	0
-		.word	0
-		.size	arm7_processor_functions, . - arm7_processor_functions
 
 		.section ".rodata"
 
-		.type	cpu_arch_name, #object
+		string	cpu_arch_name, "armv3"
-cpu_arch_name:	.asciz	"armv3"
+		string	cpu_elf_name, "v3"
-		.size	cpu_arch_name, . - cpu_arch_name
+		string	cpu_arm6_name, "ARM6"
-
+		string	cpu_arm610_name, "ARM610"
-		.type	cpu_elf_name, #object
+		string	cpu_arm7_name, "ARM7"
-cpu_elf_name:	.asciz	"v3"
+		string	cpu_arm710_name, "ARM710"
-		.size	cpu_elf_name, . - cpu_elf_name
-
-		.type	cpu_arm6_name, #object
-cpu_arm6_name:	.asciz	"ARM6"
-		.size	cpu_arm6_name, . - cpu_arm6_name
-
-		.type	cpu_arm610_name, #object
-cpu_arm610_name:
-		.asciz	"ARM610"
-		.size	cpu_arm610_name, . - cpu_arm610_name
-
-		.type	cpu_arm7_name, #object
-cpu_arm7_name:	.asciz	"ARM7"
-		.size	cpu_arm7_name, . - cpu_arm7_name
-
-		.type	cpu_arm710_name, #object
-cpu_arm710_name:
-		.asciz	"ARM710"
-		.size	cpu_arm710_name, . - cpu_arm710_name
 
 		.align
 
 		.section ".proc.info.init", #alloc, #execinstr
 
-		.type	__arm6_proc_info, #object
+.macro arm67_proc_info name:req, cpu_val:req, cpu_mask:req, cpu_name:req, \
-__arm6_proc_info:
+	cpu_mm_mmu_flags:req, cpu_flush:req, cpu_proc_funcs:req
-		.long	0x41560600
+		.type	__\name\()_proc_info, #object
-		.long	0xfffffff0
+__\name\()_proc_info:
-		.long	0x00000c1e
+		.long	\cpu_val
+		.long	\cpu_mask
+		.long	\cpu_mm_mmu_flags
 		.long   PMD_TYPE_SECT | \
 			PMD_BIT4 | \
 			PMD_SECT_AP_WRITE | \
 			PMD_SECT_AP_READ
-		b	__arm6_setup
+		b	\cpu_flush
 		.long	cpu_arch_name
 		.long	cpu_elf_name
 		.long	HWCAP_SWP | HWCAP_26BIT
-		.long	cpu_arm6_name
+		.long	\cpu_name
-		.long	arm6_processor_functions
+		.long	\cpu_proc_funcs
 		.long	v3_tlb_fns
 		.long	v3_user_fns
 		.long	v3_cache_fns
-		.size	__arm6_proc_info, . - __arm6_proc_info
+		.size	__\name\()_proc_info, . - __\name\()_proc_info
+.endm
 
-		.type	__arm610_proc_info, #object
+	arm67_proc_info	arm6,	0x41560600, 0xfffffff0, cpu_arm6_name, \
-__arm610_proc_info:
+		0x00000c1e, __arm6_setup, arm6_processor_functions
-		.long	0x41560610
+	arm67_proc_info	arm610,	0x41560610, 0xfffffff0, cpu_arm610_name, \
-		.long	0xfffffff0
+		0x00000c1e, __arm6_setup, arm6_processor_functions
-		.long	0x00000c1e
+	arm67_proc_info	arm7,	0x41007000, 0xffffff00, cpu_arm7_name, \
-		.long   PMD_TYPE_SECT | \
+		0x00000c1e, __arm7_setup, arm7_processor_functions
-			PMD_BIT4 | \
+	arm67_proc_info	arm710,	0x41007100, 0xfff8ff00, cpu_arm710_name, \
-			PMD_SECT_AP_WRITE | \
+			PMD_TYPE_SECT | \
-			PMD_SECT_AP_READ
-		b	__arm6_setup
-		.long	cpu_arch_name
-		.long	cpu_elf_name
-		.long	HWCAP_SWP | HWCAP_26BIT
-		.long	cpu_arm610_name
-		.long	arm6_processor_functions
-		.long	v3_tlb_fns
-		.long	v3_user_fns
-		.long	v3_cache_fns
-		.size	__arm610_proc_info, . - __arm610_proc_info
-
-		.type	__arm7_proc_info, #object
-__arm7_proc_info:
-		.long	0x41007000
-		.long	0xffffff00
-		.long	0x00000c1e
-		.long   PMD_TYPE_SECT | \
-			PMD_BIT4 | \
-			PMD_SECT_AP_WRITE | \
-			PMD_SECT_AP_READ
-		b	__arm7_setup
-		.long	cpu_arch_name
-		.long	cpu_elf_name
-		.long	HWCAP_SWP | HWCAP_26BIT
-		.long	cpu_arm7_name
-		.long	arm7_processor_functions
-		.long	v3_tlb_fns
-		.long	v3_user_fns
-		.long	v3_cache_fns
-		.size	__arm7_proc_info, . - __arm7_proc_info
-
-		.type	__arm710_proc_info, #object
-__arm710_proc_info:
-		.long	0x41007100
-		.long	0xfff8ff00
-		.long   PMD_TYPE_SECT | \
 			PMD_SECT_BUFFERABLE | \
 			PMD_SECT_CACHEABLE | \
 			PMD_BIT4 | \
 			PMD_SECT_AP_WRITE | \
-			PMD_SECT_AP_READ
+			PMD_SECT_AP_READ, \
-		.long   PMD_TYPE_SECT | \
+		__arm7_setup, arm7_processor_functions
-			PMD_BIT4 | \
-			PMD_SECT_AP_WRITE | \
-			PMD_SECT_AP_READ
-		b	__arm7_setup
-		.long	cpu_arch_name
-		.long	cpu_elf_name
-		.long	HWCAP_SWP | HWCAP_26BIT
-		.long	cpu_arm710_name
-		.long	arm7_processor_functions
-		.long	v3_tlb_fns
-		.long	v3_user_fns
-		.long	v3_cache_fns
-		.size	__arm710_proc_info, . - __arm710_proc_info

arch/arm/mm/proc-arm720.S

@@ -169,46 +169,15 @@ arm720_crval:
 	crval	clear=0x00002f3f, mmuset=0x0000213d, ucset=0x00000130
 
 		__INITDATA
-
+	@ define struct processor (see <asm/proc-fns.h> and proc-macros.S)
-/*
+	define_processor_functions arm720, dabort=v4t_late_abort, pabort=legacy_pabort
- * Purpose : Function pointers used to access above functions - all calls
- *	     come through these
- */
-		.type	arm720_processor_functions, #object
-ENTRY(arm720_processor_functions)
-		.word	v4t_late_abort
-		.word	legacy_pabort
-		.word	cpu_arm720_proc_init
-		.word	cpu_arm720_proc_fin
-		.word	cpu_arm720_reset
-		.word	cpu_arm720_do_idle
-		.word	cpu_arm720_dcache_clean_area
-		.word	cpu_arm720_switch_mm
-		.word	cpu_arm720_set_pte_ext
-		.word	0
-		.word	0
-		.word	0
-		.size	arm720_processor_functions, . - arm720_processor_functions
 
 		.section ".rodata"
 
-		.type	cpu_arch_name, #object
+	string	cpu_arch_name, "armv4t"
-cpu_arch_name:	.asciz	"armv4t"
+	string	cpu_elf_name, "v4"
-		.size	cpu_arch_name, . - cpu_arch_name
+	string	cpu_arm710_name, "ARM710T"
-
+	string	cpu_arm720_name, "ARM720T"
-		.type	cpu_elf_name, #object
-cpu_elf_name:	.asciz	"v4"
-		.size	cpu_elf_name, . - cpu_elf_name
-
-		.type	cpu_arm710_name, #object
-cpu_arm710_name:
-		.asciz	"ARM710T"
-		.size	cpu_arm710_name, . - cpu_arm710_name
-
-		.type	cpu_arm720_name, #object
-cpu_arm720_name:
-		.asciz	"ARM720T"
-		.size	cpu_arm720_name, . - cpu_arm720_name
 
 		.align
 
@@ -218,10 +187,11 @@ cpu_arm720_name:
 	
 		.section ".proc.info.init", #alloc, #execinstr
 
-		.type	__arm710_proc_info, #object
+.macro arm720_proc_info name:req, cpu_val:req, cpu_mask:req, cpu_name:req, cpu_flush:req
-__arm710_proc_info:
+		.type	__\name\()_proc_info,#object
-		.long	0x41807100				@ cpu_val
+__\name\()_proc_info:
-		.long	0xffffff00				@ cpu_mask
+		.long	\cpu_val
+		.long	\cpu_mask
 		.long   PMD_TYPE_SECT | \
 			PMD_SECT_BUFFERABLE | \
 			PMD_SECT_CACHEABLE | \
@@ -232,38 +202,17 @@ __arm710_proc_info:
 			PMD_BIT4 | \
 			PMD_SECT_AP_WRITE | \
 			PMD_SECT_AP_READ
-		b	__arm710_setup				@ cpu_flush
+		b	\cpu_flush				@ cpu_flush
 		.long	cpu_arch_name				@ arch_name
 		.long	cpu_elf_name				@ elf_name
 		.long	HWCAP_SWP | HWCAP_HALF | HWCAP_THUMB	@ elf_hwcap
-		.long	cpu_arm710_name				@ name
+		.long	\cpu_name
 		.long	arm720_processor_functions
 		.long	v4_tlb_fns
 		.long	v4wt_user_fns
 		.long	v4_cache_fns
-		.size	__arm710_proc_info, . - __arm710_proc_info
+		.size	__\name\()_proc_info, . - __\name\()_proc_info
+.endm
 
-		.type	__arm720_proc_info, #object
+	arm720_proc_info arm710, 0x41807100, 0xffffff00, cpu_arm710_name, __arm710_setup
-__arm720_proc_info:
+	arm720_proc_info arm720, 0x41807200, 0xffffff00, cpu_arm720_name, __arm720_setup
-		.long	0x41807200				@ cpu_val
-		.long	0xffffff00				@ cpu_mask
-		.long   PMD_TYPE_SECT | \
-			PMD_SECT_BUFFERABLE | \
-			PMD_SECT_CACHEABLE | \
-			PMD_BIT4 | \
-			PMD_SECT_AP_WRITE | \
-			PMD_SECT_AP_READ
-		.long   PMD_TYPE_SECT | \
-			PMD_BIT4 | \
-			PMD_SECT_AP_WRITE | \
-			PMD_SECT_AP_READ
-		b	__arm720_setup				@ cpu_flush
-		.long	cpu_arch_name				@ arch_name
-		.long	cpu_elf_name				@ elf_name
-		.long	HWCAP_SWP | HWCAP_HALF | HWCAP_THUMB	@ elf_hwcap
-		.long	cpu_arm720_name				@ name
-		.long	arm720_processor_functions
-		.long	v4_tlb_fns
-		.long	v4wt_user_fns
-		.long	v4_cache_fns
-		.size	__arm720_proc_info, . - __arm720_proc_info

arch/arm/mm/proc-arm740.S

@@ -17,6 +17,8 @@
 #include <asm/pgtable.h>
 #include <asm/ptrace.h>
 
+#include "proc-macros.S"
+
 	.text
 /*
  * cpu_arm740_proc_init()
@@ -115,42 +117,14 @@ __arm740_setup:
 
 	__INITDATA
 
-/*
+	@ define struct processor (see <asm/proc-fns.h> and proc-macros.S)
- * Purpose : Function pointers used to access above functions - all calls
+	define_processor_functions arm740, dabort=v4t_late_abort, pabort=legacy_pabort, nommu=1
- *	     come through these
- */
-	.type	arm740_processor_functions, #object
-ENTRY(arm740_processor_functions)
-	.word	v4t_late_abort
-	.word	legacy_pabort
-	.word	cpu_arm740_proc_init
-	.word	cpu_arm740_proc_fin
-	.word	cpu_arm740_reset
-	.word   cpu_arm740_do_idle
-	.word	cpu_arm740_dcache_clean_area
-	.word	cpu_arm740_switch_mm
-	.word	0			@ cpu_*_set_pte
-	.word	0
-	.word	0
-	.word	0
-	.size	arm740_processor_functions, . - arm740_processor_functions
 
 	.section ".rodata"
 
-	.type	cpu_arch_name, #object
+	string	cpu_arch_name, "armv4"
-cpu_arch_name:
+	string	cpu_elf_name, "v4"
-	.asciz	"armv4"
+	string	cpu_arm740_name, "ARM740T"
-	.size	cpu_arch_name, . - cpu_arch_name
-
-	.type	cpu_elf_name, #object
-cpu_elf_name:
-	.asciz	"v4"
-	.size	cpu_elf_name, . - cpu_elf_name
-
-	.type	cpu_arm740_name, #object
-cpu_arm740_name:
-	.ascii	"ARM740T"
-	.size	cpu_arm740_name, . - cpu_arm740_name
 
 	.align
 
@@ -170,5 +144,3 @@ __arm740_proc_info:
 	.long	0
 	.long	v3_cache_fns			@ cache model
 	.size	__arm740_proc_info, . - __arm740_proc_info
-
-

arch/arm/mm/proc-arm7tdmi.S

@@ -17,6 +17,8 @@
 #include <asm/pgtable.h>
 #include <asm/ptrace.h>
 
+#include "proc-macros.S"
+
 	.text
 /*
  * cpu_arm7tdmi_proc_init()
@@ -55,197 +57,57 @@ __arm7tdmi_setup:
 
 		__INITDATA
 
-/*
+		@ define struct processor (see <asm/proc-fns.h> and proc-macros.S)
- * Purpose : Function pointers used to access above functions - all calls
+		define_processor_functions arm7tdmi, dabort=v4t_late_abort, pabort=legacy_pabort, nommu=1
- *	     come through these
- */
-		.type	arm7tdmi_processor_functions, #object
-ENTRY(arm7tdmi_processor_functions)
-		.word	v4t_late_abort
-		.word	legacy_pabort
-		.word	cpu_arm7tdmi_proc_init
-		.word	cpu_arm7tdmi_proc_fin
-		.word	cpu_arm7tdmi_reset
-		.word	cpu_arm7tdmi_do_idle
-		.word	cpu_arm7tdmi_dcache_clean_area
-		.word	cpu_arm7tdmi_switch_mm
-		.word	0		@ cpu_*_set_pte
-		.word	0
-		.word	0
-		.word	0
-		.size	arm7tdmi_processor_functions, . - arm7tdmi_processor_functions
 
 		.section ".rodata"
 
-		.type	cpu_arch_name, #object
+		string	cpu_arch_name, "armv4t"
-cpu_arch_name:
+		string	cpu_elf_name, "v4"
-		.asciz	"armv4t"
+		string	cpu_arm7tdmi_name, "ARM7TDMI"
-		.size	cpu_arch_name, . - cpu_arch_name
+		string	cpu_triscenda7_name, "Triscend-A7x"
-
+		string	cpu_at91_name, "Atmel-AT91M40xxx"
-		.type	cpu_elf_name, #object
+		string	cpu_s3c3410_name, "Samsung-S3C3410"
-cpu_elf_name:
+		string	cpu_s3c44b0x_name, "Samsung-S3C44B0x"
-		.asciz	"v4"
+		string	cpu_s3c4510b_name, "Samsung-S3C4510B"
-		.size	cpu_elf_name, . - cpu_elf_name
+		string	cpu_s3c4530_name, "Samsung-S3C4530"
-
+		string	cpu_netarm_name, "NETARM"
-		.type	cpu_arm7tdmi_name, #object
-cpu_arm7tdmi_name:
-		.asciz	"ARM7TDMI"
-		.size	cpu_arm7tdmi_name, . - cpu_arm7tdmi_name
-
-		.type	cpu_triscenda7_name, #object
-cpu_triscenda7_name:
-		.asciz	"Triscend-A7x"
-		.size	cpu_triscenda7_name, . - cpu_triscenda7_name
-
-		.type	cpu_at91_name, #object
-cpu_at91_name:
-		.asciz	"Atmel-AT91M40xxx"
-		.size	cpu_at91_name, . - cpu_at91_name
-
-		.type	cpu_s3c3410_name, #object
-cpu_s3c3410_name:
-		.asciz	"Samsung-S3C3410"
-		.size	cpu_s3c3410_name, . - cpu_s3c3410_name
-
-		.type	cpu_s3c44b0x_name, #object
-cpu_s3c44b0x_name:
-		.asciz	"Samsung-S3C44B0x"
-		.size	cpu_s3c44b0x_name, . - cpu_s3c44b0x_name
-
-		.type	cpu_s3c4510b, #object
-cpu_s3c4510b_name:
-		.asciz	"Samsung-S3C4510B"
-		.size	cpu_s3c4510b_name, . - cpu_s3c4510b_name
-
-		.type	cpu_s3c4530_name, #object
-cpu_s3c4530_name:
-		.asciz	"Samsung-S3C4530"
-		.size	cpu_s3c4530_name, . - cpu_s3c4530_name
-
-		.type	cpu_netarm_name, #object
-cpu_netarm_name:
-		.asciz	"NETARM"
-		.size	cpu_netarm_name, . - cpu_netarm_name
 
 		.align
 
 		.section ".proc.info.init", #alloc, #execinstr
 
-		.type	__arm7tdmi_proc_info, #object
+.macro arm7tdmi_proc_info name:req, cpu_val:req, cpu_mask:req, cpu_name:req, \
-__arm7tdmi_proc_info:
+	extra_hwcaps=0
-		.long	0x41007700
+		.type	__\name\()_proc_info, #object
-		.long	0xfff8ff00
+__\name\()_proc_info:
+		.long	\cpu_val
+		.long	\cpu_mask
 		.long	0
 		.long	0
 		b	__arm7tdmi_setup
 		.long	cpu_arch_name
 		.long	cpu_elf_name
-		.long	HWCAP_SWP | HWCAP_26BIT
+		.long	HWCAP_SWP | HWCAP_26BIT | ( \extra_hwcaps )
-		.long	cpu_arm7tdmi_name
+		.long	\cpu_name
 		.long	arm7tdmi_processor_functions
 		.long	0
 		.long	0
 		.long	v4_cache_fns
-		.size	__arm7tdmi_proc_info, . - __arm7tdmi_proc_info
+		.size	__\name\()_proc_info, . - __\name\()_proc_info
+.endm
 
-		.type	__triscenda7_proc_info, #object
+		arm7tdmi_proc_info arm7tdmi, 0x41007700, 0xfff8ff00, \
-__triscenda7_proc_info:
+			cpu_arm7tdmi_name
-		.long	0x0001d2ff
+		arm7tdmi_proc_info triscenda7, 0x0001d2ff, 0x0001ffff, \
-		.long	0x0001ffff
+			cpu_triscenda7_name, extra_hwcaps=HWCAP_THUMB
-		.long	0
+		arm7tdmi_proc_info at91, 0x14000040, 0xfff000e0, \
-		.long	0
+			cpu_at91_name, extra_hwcaps=HWCAP_THUMB
-		b	__arm7tdmi_setup
+		arm7tdmi_proc_info s3c4510b, 0x36365000, 0xfffff000, \
-		.long	cpu_arch_name
+			cpu_s3c4510b_name, extra_hwcaps=HWCAP_THUMB
-		.long	cpu_elf_name
+		arm7tdmi_proc_info s3c4530, 0x4c000000, 0xfff000e0, \
-		.long	HWCAP_SWP | HWCAP_THUMB | HWCAP_26BIT
+			cpu_s3c4530_name, extra_hwcaps=HWCAP_THUMB
-		.long	cpu_triscenda7_name
+		arm7tdmi_proc_info s3c3410, 0x34100000, 0xffff0000, \
-		.long	arm7tdmi_processor_functions
+			cpu_s3c3410_name, extra_hwcaps=HWCAP_THUMB
-		.long	0
+		arm7tdmi_proc_info s3c44b0x, 0x44b00000, 0xffff0000, \
-		.long	0
+			cpu_s3c44b0x_name, extra_hwcaps=HWCAP_THUMB
-		.long	v4_cache_fns
-		.size	__triscenda7_proc_info, . - __triscenda7_proc_info
-
-		.type	__at91_proc_info, #object
-__at91_proc_info:
-		.long	0x14000040
-		.long	0xfff000e0
-		.long	0
-		.long	0
-		b	__arm7tdmi_setup
-		.long	cpu_arch_name
-		.long	cpu_elf_name
-		.long	HWCAP_SWP | HWCAP_THUMB | HWCAP_26BIT
-		.long	cpu_at91_name
-		.long	arm7tdmi_processor_functions
-		.long	0
-		.long	0
-		.long	v4_cache_fns
-		.size	__at91_proc_info, . - __at91_proc_info
-
-		.type	__s3c4510b_proc_info, #object
-__s3c4510b_proc_info:
-		.long	0x36365000
-		.long	0xfffff000
-		.long	0
-		.long	0
-		b	__arm7tdmi_setup
-		.long	cpu_arch_name
-		.long	cpu_elf_name
-		.long	HWCAP_SWP | HWCAP_THUMB | HWCAP_26BIT
-		.long	cpu_s3c4510b_name
-		.long	arm7tdmi_processor_functions
-		.long	0
-		.long	0
-		.long	v4_cache_fns
-		.size	__s3c4510b_proc_info, . - __s3c4510b_proc_info
-
-		.type	__s3c4530_proc_info, #object
-__s3c4530_proc_info:
-		.long	0x4c000000
-		.long	0xfff000e0
-		.long	0
-		.long	0
-		b	__arm7tdmi_setup
-		.long	cpu_arch_name
-		.long	cpu_elf_name
-		.long	HWCAP_SWP | HWCAP_THUMB | HWCAP_26BIT
-		.long	cpu_s3c4530_name
-		.long	arm7tdmi_processor_functions
-		.long	0
-		.long	0
-		.long	v4_cache_fns
-		.size	__s3c4530_proc_info, . - __s3c4530_proc_info
-
-		.type	__s3c3410_proc_info, #object
-__s3c3410_proc_info:
-		.long	0x34100000
-		.long	0xffff0000
-		.long	0
-		.long	0
-		b	__arm7tdmi_setup
-		.long	cpu_arch_name
-		.long	cpu_elf_name
-		.long	HWCAP_SWP | HWCAP_THUMB | HWCAP_26BIT
-		.long	cpu_s3c3410_name
-		.long	arm7tdmi_processor_functions
-		.long	0
-		.long	0
-		.long	v4_cache_fns
-		.size	__s3c3410_proc_info, . - __s3c3410_proc_info
-
-		.type	__s3c44b0x_proc_info, #object
-__s3c44b0x_proc_info:
-		.long	0x44b00000
-		.long	0xffff0000
-		.long	0
-		.long	0
-		b	__arm7tdmi_setup
-		.long	cpu_arch_name
-		.long	cpu_elf_name
-		.long	HWCAP_SWP | HWCAP_THUMB | HWCAP_26BIT
-		.long	cpu_s3c44b0x_name
-		.long	arm7tdmi_processor_functions
-		.long	0
-		.long	0
-		.long	v4_cache_fns
-		.size	__s3c44b0x_proc_info, . - __s3c44b0x_proc_info

arch/arm/mm/proc-arm920.S

@@ -315,18 +315,8 @@ ENTRY(arm920_dma_unmap_area)
 	mov	pc, lr
 ENDPROC(arm920_dma_unmap_area)
 
-ENTRY(arm920_cache_fns)
+	@ define struct cpu_cache_fns (see <asm/cacheflush.h> and proc-macros.S)
-	.long	arm920_flush_icache_all
+	define_cache_functions arm920
-	.long	arm920_flush_kern_cache_all
-	.long	arm920_flush_user_cache_all
-	.long	arm920_flush_user_cache_range
-	.long	arm920_coherent_kern_range
-	.long	arm920_coherent_user_range
-	.long	arm920_flush_kern_dcache_area
-	.long	arm920_dma_map_area
-	.long	arm920_dma_unmap_area
-	.long	arm920_dma_flush_range
-
 #endif
 
 
@@ -416,9 +406,6 @@ ENTRY(cpu_arm920_do_resume)
 		     PMD_SECT_CACHEABLE | PMD_BIT4 | PMD_SECT_AP_WRITE
 	b	cpu_resume_mmu
 ENDPROC(cpu_arm920_do_resume)
-#else
-#define cpu_arm920_do_suspend	0
-#define cpu_arm920_do_resume	0
 #endif
 
 	__CPUINIT
@@ -450,43 +437,14 @@ arm920_crval:
 	crval	clear=0x00003f3f, mmuset=0x00003135, ucset=0x00001130
 
 	__INITDATA
-
+	@ define struct processor (see <asm/proc-fns.h> and proc-macros.S)
-/*
+	define_processor_functions arm920, dabort=v4t_early_abort, pabort=legacy_pabort, suspend=1
- * Purpose : Function pointers used to access above functions - all calls
- *	     come through these
- */
-	.type	arm920_processor_functions, #object
-arm920_processor_functions:
-	.word	v4t_early_abort
-	.word	legacy_pabort
-	.word	cpu_arm920_proc_init
-	.word	cpu_arm920_proc_fin
-	.word	cpu_arm920_reset
-	.word   cpu_arm920_do_idle
-	.word	cpu_arm920_dcache_clean_area
-	.word	cpu_arm920_switch_mm
-	.word	cpu_arm920_set_pte_ext
-	.word	cpu_arm920_suspend_size
-	.word	cpu_arm920_do_suspend
-	.word	cpu_arm920_do_resume
-	.size	arm920_processor_functions, . - arm920_processor_functions
 
 	.section ".rodata"
 
-	.type	cpu_arch_name, #object
+	string	cpu_arch_name, "armv4t"
-cpu_arch_name:
+	string	cpu_elf_name, "v4"
-	.asciz	"armv4t"
+	string	cpu_arm920_name, "ARM920T"
-	.size	cpu_arch_name, . - cpu_arch_name
-
-	.type	cpu_elf_name, #object
-cpu_elf_name:
-	.asciz	"v4"
-	.size	cpu_elf_name, . - cpu_elf_name
-
-	.type	cpu_arm920_name, #object
-cpu_arm920_name:
-	.asciz	"ARM920T"
-	.size	cpu_arm920_name, . - cpu_arm920_name
 
 	.align
 

arch/arm/mm/proc-arm922.S

@@ -317,18 +317,8 @@ ENTRY(arm922_dma_unmap_area)
 	mov	pc, lr
 ENDPROC(arm922_dma_unmap_area)
 
-ENTRY(arm922_cache_fns)
+	@ define struct cpu_cache_fns (see <asm/cacheflush.h> and proc-macros.S)
-	.long	arm922_flush_icache_all
+	define_cache_functions arm922
-	.long	arm922_flush_kern_cache_all
-	.long	arm922_flush_user_cache_all
-	.long	arm922_flush_user_cache_range
-	.long	arm922_coherent_kern_range
-	.long	arm922_coherent_user_range
-	.long	arm922_flush_kern_dcache_area
-	.long	arm922_dma_map_area
-	.long	arm922_dma_unmap_area
-	.long	arm922_dma_flush_range
-
 #endif
 
 
@@ -420,43 +410,14 @@ arm922_crval:
 	crval	clear=0x00003f3f, mmuset=0x00003135, ucset=0x00001130
 
 	__INITDATA
-
+	@ define struct processor (see <asm/proc-fns.h> and proc-macros.S)
-/*
+	define_processor_functions arm922, dabort=v4t_early_abort, pabort=legacy_pabort
- * Purpose : Function pointers used to access above functions - all calls
- *	     come through these
- */
-	.type	arm922_processor_functions, #object
-arm922_processor_functions:
-	.word	v4t_early_abort
-	.word	legacy_pabort
-	.word	cpu_arm922_proc_init
-	.word	cpu_arm922_proc_fin
-	.word	cpu_arm922_reset
-	.word   cpu_arm922_do_idle
-	.word	cpu_arm922_dcache_clean_area
-	.word	cpu_arm922_switch_mm
-	.word	cpu_arm922_set_pte_ext
-	.word	0
-	.word	0
-	.word	0
-	.size	arm922_processor_functions, . - arm922_processor_functions
 
 	.section ".rodata"
 
-	.type	cpu_arch_name, #object
+	string	cpu_arch_name, "armv4t"
-cpu_arch_name:
+	string	cpu_elf_name, "v4"
-	.asciz	"armv4t"
+	string	cpu_arm922_name, "ARM922T"
-	.size	cpu_arch_name, . - cpu_arch_name
-
-	.type	cpu_elf_name, #object
-cpu_elf_name:
-	.asciz	"v4"
-	.size	cpu_elf_name, . - cpu_elf_name
-
-	.type	cpu_arm922_name, #object
-cpu_arm922_name:
-	.asciz	"ARM922T"
-	.size	cpu_arm922_name, . - cpu_arm922_name
 
 	.align
 

arch/arm/mm/proc-arm925.S

@@ -372,17 +372,8 @@ ENTRY(arm925_dma_unmap_area)
 	mov	pc, lr
 ENDPROC(arm925_dma_unmap_area)
 
-ENTRY(arm925_cache_fns)
+	@ define struct cpu_cache_fns (see <asm/cacheflush.h> and proc-macros.S)
-	.long	arm925_flush_icache_all
+	define_cache_functions arm925
-	.long	arm925_flush_kern_cache_all
-	.long	arm925_flush_user_cache_all
-	.long	arm925_flush_user_cache_range
-	.long	arm925_coherent_kern_range
-	.long	arm925_coherent_user_range
-	.long	arm925_flush_kern_dcache_area
-	.long	arm925_dma_map_area
-	.long	arm925_dma_unmap_area
-	.long	arm925_dma_flush_range
 
 ENTRY(cpu_arm925_dcache_clean_area)
 #ifndef CONFIG_CPU_DCACHE_WRITETHROUGH
@@ -487,52 +478,24 @@ arm925_crval:
 	crval	clear=0x00007f3f, mmuset=0x0000313d, ucset=0x00001130
 
 	__INITDATA
-
+	@ define struct processor (see <asm/proc-fns.h> and proc-macros.S)
-/*
+	define_processor_functions arm925, dabort=v4t_early_abort, pabort=legacy_pabort
- * Purpose : Function pointers used to access above functions - all calls
- *	     come through these
- */
-	.type	arm925_processor_functions, #object
-arm925_processor_functions:
-	.word	v4t_early_abort
-	.word	legacy_pabort
-	.word	cpu_arm925_proc_init
-	.word	cpu_arm925_proc_fin
-	.word	cpu_arm925_reset
-	.word   cpu_arm925_do_idle
-	.word	cpu_arm925_dcache_clean_area
-	.word	cpu_arm925_switch_mm
-	.word	cpu_arm925_set_pte_ext
-	.word	0
-	.word	0
-	.word	0
-	.size	arm925_processor_functions, . - arm925_processor_functions
 
 	.section ".rodata"
 
-	.type	cpu_arch_name, #object
+	string	cpu_arch_name, "armv4t"
-cpu_arch_name:
+	string	cpu_elf_name, "v4"
-	.asciz	"armv4t"
+	string	cpu_arm925_name, "ARM925T"
-	.size	cpu_arch_name, . - cpu_arch_name
-
-	.type	cpu_elf_name, #object
-cpu_elf_name:
-	.asciz	"v4"
-	.size	cpu_elf_name, . - cpu_elf_name
-
-	.type	cpu_arm925_name, #object
-cpu_arm925_name:
-	.asciz	"ARM925T"
-	.size	cpu_arm925_name, . - cpu_arm925_name
 
 	.align
 
 	.section ".proc.info.init", #alloc, #execinstr
 
-	.type	__arm925_proc_info,#object
+.macro arm925_proc_info name:req, cpu_val:req, cpu_mask:req, cpu_name:req, cache
-__arm925_proc_info:
+	.type	__\name\()_proc_info,#object
-	.long	0x54029250
+__\name\()_proc_info:
-	.long	0xfffffff0
+	.long	\cpu_val
+	.long	\cpu_mask
 	.long   PMD_TYPE_SECT | \
 		PMD_BIT4 | \
 		PMD_SECT_AP_WRITE | \
@@ -550,27 +513,8 @@ __arm925_proc_info:
 	.long	v4wbi_tlb_fns
 	.long	v4wb_user_fns
 	.long	arm925_cache_fns
-	.size	__arm925_proc_info, . - __arm925_proc_info
+	.size	__\name\()_proc_info, . - __\name\()_proc_info
+.endm
 
-	.type	__arm915_proc_info,#object
+	arm925_proc_info arm925, 0x54029250, 0xfffffff0, cpu_arm925_name
-__arm915_proc_info:
+	arm925_proc_info arm915, 0x54029150, 0xfffffff0, cpu_arm925_name
-	.long	0x54029150
-	.long	0xfffffff0
-	.long   PMD_TYPE_SECT | \
-		PMD_BIT4 | \
-		PMD_SECT_AP_WRITE | \
-		PMD_SECT_AP_READ
-	.long   PMD_TYPE_SECT | \
-		PMD_BIT4 | \
-		PMD_SECT_AP_WRITE | \
-		PMD_SECT_AP_READ
-	b	__arm925_setup
-	.long	cpu_arch_name
-	.long	cpu_elf_name
-	.long	HWCAP_SWP | HWCAP_HALF | HWCAP_THUMB
-	.long	cpu_arm925_name
-	.long	arm925_processor_functions
-	.long	v4wbi_tlb_fns
-	.long	v4wb_user_fns
-	.long	arm925_cache_fns
-	.size	__arm925_proc_info, . - __arm925_proc_info

arch/arm/mm/proc-arm926.S

@@ -335,17 +335,8 @@ ENTRY(arm926_dma_unmap_area)
 	mov	pc, lr
 ENDPROC(arm926_dma_unmap_area)
 
-ENTRY(arm926_cache_fns)
+	@ define struct cpu_cache_fns (see <asm/cacheflush.h> and proc-macros.S)
-	.long	arm926_flush_icache_all
+	define_cache_functions arm926
-	.long	arm926_flush_kern_cache_all
-	.long	arm926_flush_user_cache_all
-	.long	arm926_flush_user_cache_range
-	.long	arm926_coherent_kern_range
-	.long	arm926_coherent_user_range
-	.long	arm926_flush_kern_dcache_area
-	.long	arm926_dma_map_area
-	.long	arm926_dma_unmap_area
-	.long	arm926_dma_flush_range
 
 ENTRY(cpu_arm926_dcache_clean_area)
 #ifndef CONFIG_CPU_DCACHE_WRITETHROUGH
@@ -430,9 +421,6 @@ ENTRY(cpu_arm926_do_resume)
 		     PMD_SECT_CACHEABLE | PMD_BIT4 | PMD_SECT_AP_WRITE
 	b	cpu_resume_mmu
 ENDPROC(cpu_arm926_do_resume)
-#else
-#define cpu_arm926_do_suspend	0
-#define cpu_arm926_do_resume	0
 #endif
 
 	__CPUINIT
@@ -475,42 +463,14 @@ arm926_crval:
 
 	__INITDATA
 
-/*
+	@ define struct processor (see <asm/proc-fns.h> and proc-macros.S)
- * Purpose : Function pointers used to access above functions - all calls
+	define_processor_functions arm926, dabort=v5tj_early_abort, pabort=legacy_pabort, suspend=1
- *	     come through these
- */
-	.type	arm926_processor_functions, #object
-arm926_processor_functions:
-	.word	v5tj_early_abort
-	.word	legacy_pabort
-	.word	cpu_arm926_proc_init
-	.word	cpu_arm926_proc_fin
-	.word	cpu_arm926_reset
-	.word	cpu_arm926_do_idle
-	.word	cpu_arm926_dcache_clean_area
-	.word	cpu_arm926_switch_mm
-	.word	cpu_arm926_set_pte_ext
-	.word	cpu_arm926_suspend_size
-	.word	cpu_arm926_do_suspend
-	.word	cpu_arm926_do_resume
-	.size	arm926_processor_functions, . - arm926_processor_functions
 
 	.section ".rodata"
 
-	.type	cpu_arch_name, #object
+	string	cpu_arch_name, "armv5tej"
-cpu_arch_name:
+	string	cpu_elf_name, "v5"
-	.asciz	"armv5tej"
+	string	cpu_arm926_name, "ARM926EJ-S"
-	.size	cpu_arch_name, . - cpu_arch_name
-
-	.type	cpu_elf_name, #object
-cpu_elf_name:
-	.asciz	"v5"
-	.size	cpu_elf_name, . - cpu_elf_name
-
-	.type	cpu_arm926_name, #object
-cpu_arm926_name:
-	.asciz	"ARM926EJ-S"
-	.size	cpu_arm926_name, . - cpu_arm926_name
 
 	.align
 

arch/arm/mm/proc-arm940.S

@@ -264,17 +264,8 @@ ENTRY(arm940_dma_unmap_area)
 	mov	pc, lr
 ENDPROC(arm940_dma_unmap_area)
 
-ENTRY(arm940_cache_fns)
+	@ define struct cpu_cache_fns (see <asm/cacheflush.h> and proc-macros.S)
-	.long	arm940_flush_icache_all
+	define_cache_functions arm940
-	.long	arm940_flush_kern_cache_all
-	.long	arm940_flush_user_cache_all
-	.long	arm940_flush_user_cache_range
-	.long	arm940_coherent_kern_range
-	.long	arm940_coherent_user_range
-	.long	arm940_flush_kern_dcache_area
-	.long	arm940_dma_map_area
-	.long	arm940_dma_unmap_area
-	.long	arm940_dma_flush_range
 
 	__CPUINIT
 
@@ -348,42 +339,14 @@ __arm940_setup:
 
 	__INITDATA
 
-/*
+	@ define struct processor (see <asm/proc-fns.h> and proc-macros.S)
- * Purpose : Function pointers used to access above functions - all calls
+	define_processor_functions arm940, dabort=nommu_early_abort, pabort=legacy_pabort, nommu=1
- *	     come through these
- */
-	.type	arm940_processor_functions, #object
-ENTRY(arm940_processor_functions)
-	.word	nommu_early_abort
-	.word	legacy_pabort
-	.word	cpu_arm940_proc_init
-	.word	cpu_arm940_proc_fin
-	.word	cpu_arm940_reset
-	.word   cpu_arm940_do_idle
-	.word	cpu_arm940_dcache_clean_area
-	.word	cpu_arm940_switch_mm
-	.word	0		@ cpu_*_set_pte
-	.word	0
-	.word	0
-	.word	0
-	.size	arm940_processor_functions, . - arm940_processor_functions
 
 	.section ".rodata"
 
-.type	cpu_arch_name, #object
+	string	cpu_arch_name, "armv4t"
-cpu_arch_name:
+	string	cpu_elf_name, "v4"
-	.asciz	"armv4t"
+	string	cpu_arm940_name, "ARM940T"
-	.size	cpu_arch_name, . - cpu_arch_name
-
-	.type	cpu_elf_name, #object
-cpu_elf_name:
-	.asciz	"v4"
-	.size	cpu_elf_name, . - cpu_elf_name
-
-	.type	cpu_arm940_name, #object
-cpu_arm940_name:
-	.ascii	"ARM940T"
-	.size	cpu_arm940_name, . - cpu_arm940_name
 
 	.align
 

arch/arm/mm/proc-arm946.S

@@ -306,18 +306,8 @@ ENTRY(arm946_dma_unmap_area)
 	mov	pc, lr
 ENDPROC(arm946_dma_unmap_area)
 
-ENTRY(arm946_cache_fns)
+	@ define struct cpu_cache_fns (see <asm/cacheflush.h> and proc-macros.S)
-	.long	arm946_flush_icache_all
+	define_cache_functions arm946
-	.long	arm946_flush_kern_cache_all
-	.long	arm946_flush_user_cache_all
-	.long	arm946_flush_user_cache_range
-	.long	arm946_coherent_kern_range
-	.long	arm946_coherent_user_range
-	.long	arm946_flush_kern_dcache_area
-	.long	arm946_dma_map_area
-	.long	arm946_dma_unmap_area
-	.long	arm946_dma_flush_range
-
 
 ENTRY(cpu_arm946_dcache_clean_area)
 #ifndef CONFIG_CPU_DCACHE_WRITETHROUGH
@@ -403,43 +393,14 @@ __arm946_setup:
 
 	__INITDATA
 
-/*
+	@ define struct processor (see <asm/proc-fns.h> and proc-macros.S)
- * Purpose : Function pointers used to access above functions - all calls
+	define_processor_functions arm946, dabort=nommu_early_abort, pabort=legacy_pabort, nommu=1
- *	     come through these
- */
-	.type	arm946_processor_functions, #object
-ENTRY(arm946_processor_functions)
-	.word	nommu_early_abort
-	.word	legacy_pabort
-	.word	cpu_arm946_proc_init
-	.word	cpu_arm946_proc_fin
-	.word	cpu_arm946_reset
-	.word   cpu_arm946_do_idle
-
-	.word	cpu_arm946_dcache_clean_area
-	.word	cpu_arm946_switch_mm
-	.word	0		@ cpu_*_set_pte
-	.word	0
-	.word	0
-	.word	0
-	.size	arm946_processor_functions, . - arm946_processor_functions
 
 	.section ".rodata"
 
-	.type	cpu_arch_name, #object
+	string	cpu_arch_name, "armv5te"
-cpu_arch_name:
+	string	cpu_elf_name, "v5t"
-	.asciz	"armv5te"
+	string	cpu_arm946_name, "ARM946E-S"
-	.size	cpu_arch_name, . - cpu_arch_name
-
-	.type	cpu_elf_name, #object
-cpu_elf_name:
-	.asciz	"v5t"
-	.size	cpu_elf_name, . - cpu_elf_name
-
-	.type	cpu_arm946_name, #object
-cpu_arm946_name:
-	.ascii	"ARM946E-S"
-	.size	cpu_arm946_name, . - cpu_arm946_name
 
 	.align
 

arch/arm/mm/proc-arm9tdmi.S

@@ -17,6 +17,8 @@
 #include <asm/pgtable.h>
 #include <asm/ptrace.h>
 
+#include "proc-macros.S"
+
 	.text
 /*
  * cpu_arm9tdmi_proc_init()
@@ -55,82 +57,38 @@ __arm9tdmi_setup:
 
 		__INITDATA
 
-/*
+		@ define struct processor (see <asm/proc-fns.h> and proc-macros.S)
- * Purpose : Function pointers used to access above functions - all calls
+		define_processor_functions arm9tdmi, dabort=nommu_early_abort, pabort=legacy_pabort, nommu=1
- *	     come through these
- */
-		.type	arm9tdmi_processor_functions, #object
-ENTRY(arm9tdmi_processor_functions)
-		.word	nommu_early_abort
-		.word	legacy_pabort
-		.word	cpu_arm9tdmi_proc_init
-		.word	cpu_arm9tdmi_proc_fin
-		.word	cpu_arm9tdmi_reset
-		.word	cpu_arm9tdmi_do_idle
-		.word	cpu_arm9tdmi_dcache_clean_area
-		.word	cpu_arm9tdmi_switch_mm
-		.word	0		@ cpu_*_set_pte
-		.word	0
-		.word	0
-		.word	0
-		.size	arm9tdmi_processor_functions, . - arm9tdmi_processor_functions
 
 		.section ".rodata"
 
-		.type	cpu_arch_name, #object
+		string	cpu_arch_name, "armv4t"
-cpu_arch_name:
+		string	cpu_elf_name, "v4"
-		.asciz	"armv4t"
+		string	cpu_arm9tdmi_name, "ARM9TDMI"
-		.size	cpu_arch_name, . - cpu_arch_name
+		string	cpu_p2001_name, "P2001"
-
-		.type	cpu_elf_name, #object
-cpu_elf_name:
-		.asciz	"v4"
-		.size	cpu_elf_name, . - cpu_elf_name
-
-		.type	cpu_arm9tdmi_name, #object
-cpu_arm9tdmi_name:
-		.asciz	"ARM9TDMI"
-		.size	cpu_arm9tdmi_name, . - cpu_arm9tdmi_name
-
-		.type	cpu_p2001_name, #object
-cpu_p2001_name:
-		.asciz	"P2001"
-		.size	cpu_p2001_name, . - cpu_p2001_name
 
 		.align
 
 		.section ".proc.info.init", #alloc, #execinstr
 
-		.type	__arm9tdmi_proc_info, #object
+.macro arm9tdmi_proc_info name:req, cpu_val:req, cpu_mask:req, cpu_name:req
-__arm9tdmi_proc_info:
+		.type	__\name\()_proc_info, #object
-		.long	0x41009900
+__\name\()_proc_info:
-		.long	0xfff8ff00
+		.long	\cpu_val
+		.long	\cpu_mask
 		.long	0
 		.long	0
 		b	__arm9tdmi_setup
 		.long	cpu_arch_name
 		.long	cpu_elf_name
 		.long	HWCAP_SWP | HWCAP_THUMB | HWCAP_26BIT
-		.long	cpu_arm9tdmi_name
+		.long	\cpu_name
 		.long	arm9tdmi_processor_functions
 		.long	0
 		.long	0
 		.long	v4_cache_fns
-		.size	__arm9tdmi_proc_info, . - __arm9tdmi_proc_info
+		.size	__\name\()_proc_info, . - __\name\()_proc_info
+.endm
 
-		.type	__p2001_proc_info, #object
+	arm9tdmi_proc_info arm9tdmi, 0x41009900, 0xfff8ff00, cpu_arm9tdmi_name
-__p2001_proc_info:
+	arm9tdmi_proc_info p2001, 0x41029000, 0xffffffff, cpu_p2001_name
-		.long	0x41029000
-		.long	0xffffffff
-		.long	0
-		.long	0
-		b	__arm9tdmi_setup
-		.long	cpu_arch_name
-		.long	cpu_elf_name
-		.long	HWCAP_SWP | HWCAP_THUMB | HWCAP_26BIT
-		.long	cpu_p2001_name
-		.long	arm9tdmi_processor_functions
-		.long	0
-		.long	0
-		.long	v4_cache_fns
-		.size	__p2001_proc_info, . - __p2001_proc_info

arch/arm/mm/proc-fa526.S

@@ -180,42 +180,14 @@ fa526_cr1_set:
 
 	__INITDATA
 
-/*
+	@ define struct processor (see <asm/proc-fns.h> and proc-macros.S)
- * Purpose : Function pointers used to access above functions - all calls
+	define_processor_functions fa526, dabort=v4_early_abort, pabort=legacy_pabort
- *	     come through these
- */
-	.type	fa526_processor_functions, #object
-fa526_processor_functions:
-	.word	v4_early_abort
-	.word	legacy_pabort
-	.word	cpu_fa526_proc_init
-	.word	cpu_fa526_proc_fin
-	.word	cpu_fa526_reset
-	.word   cpu_fa526_do_idle
-	.word	cpu_fa526_dcache_clean_area
-	.word	cpu_fa526_switch_mm
-	.word	cpu_fa526_set_pte_ext
-	.word	0
-	.word	0
-	.word	0
-	.size	fa526_processor_functions, . - fa526_processor_functions
 
 	.section ".rodata"
 
-	.type	cpu_arch_name, #object
+	string	cpu_arch_name, "armv4"
-cpu_arch_name:
+	string	cpu_elf_name, "v4"
-	.asciz	"armv4"
+	string	cpu_fa526_name, "FA526"
-	.size	cpu_arch_name, . - cpu_arch_name
-
-	.type	cpu_elf_name, #object
-cpu_elf_name:
-	.asciz	"v4"
-	.size	cpu_elf_name, . - cpu_elf_name
-
-	.type	cpu_fa526_name, #object
-cpu_fa526_name:
-	.asciz	"FA526"
-	.size	cpu_fa526_name, . - cpu_fa526_name
 
 	.align
 

arch/arm/mm/proc-feroceon.S

@@ -411,29 +411,28 @@ ENTRY(feroceon_dma_unmap_area)
 	mov	pc, lr
 ENDPROC(feroceon_dma_unmap_area)
 
-ENTRY(feroceon_cache_fns)
+	@ define struct cpu_cache_fns (see <asm/cacheflush.h> and proc-macros.S)
-	.long	feroceon_flush_icache_all
+	define_cache_functions feroceon
-	.long	feroceon_flush_kern_cache_all
-	.long	feroceon_flush_user_cache_all
-	.long	feroceon_flush_user_cache_range
-	.long	feroceon_coherent_kern_range
-	.long	feroceon_coherent_user_range
-	.long	feroceon_flush_kern_dcache_area
-	.long	feroceon_dma_map_area
-	.long	feroceon_dma_unmap_area
-	.long	feroceon_dma_flush_range
 
-ENTRY(feroceon_range_cache_fns)
+.macro range_alias basename
-	.long	feroceon_flush_icache_all
+	.globl feroceon_range_\basename
-	.long	feroceon_flush_kern_cache_all
+	.type feroceon_range_\basename , %function
-	.long	feroceon_flush_user_cache_all
+	.equ feroceon_range_\basename , feroceon_\basename
-	.long	feroceon_flush_user_cache_range
+.endm
-	.long	feroceon_coherent_kern_range
+
-	.long	feroceon_coherent_user_range
+/*
-	.long	feroceon_range_flush_kern_dcache_area
+ * Most of the cache functions are unchanged for this case.
-	.long	feroceon_range_dma_map_area
+ * Export suitable alias symbols for the unchanged functions:
-	.long	feroceon_dma_unmap_area
+ */
-	.long	feroceon_range_dma_flush_range
+	range_alias flush_icache_all
+	range_alias flush_user_cache_all
+	range_alias flush_kern_cache_all
+	range_alias flush_user_cache_range
+	range_alias coherent_kern_range
+	range_alias coherent_user_range
+	range_alias dma_unmap_area
+
+	define_cache_functions feroceon_range
 
 	.align	5
 ENTRY(cpu_feroceon_dcache_clean_area)
@@ -539,67 +538,27 @@ feroceon_crval:
 
 	__INITDATA
 
-/*
+	@ define struct processor (see <asm/proc-fns.h> and proc-macros.S)
- * Purpose : Function pointers used to access above functions - all calls
+	define_processor_functions feroceon, dabort=v5t_early_abort, pabort=legacy_pabort
- *	     come through these
- */
-	.type	feroceon_processor_functions, #object
-feroceon_processor_functions:
-	.word	v5t_early_abort
-	.word	legacy_pabort
-	.word	cpu_feroceon_proc_init
-	.word	cpu_feroceon_proc_fin
-	.word	cpu_feroceon_reset
-	.word	cpu_feroceon_do_idle
-	.word	cpu_feroceon_dcache_clean_area
-	.word	cpu_feroceon_switch_mm
-	.word	cpu_feroceon_set_pte_ext
-	.word	0
-	.word	0
-	.word	0
-	.size	feroceon_processor_functions, . - feroceon_processor_functions
 
 	.section ".rodata"
 
-	.type	cpu_arch_name, #object
+	string	cpu_arch_name, "armv5te"
-cpu_arch_name:
+	string	cpu_elf_name, "v5"
-	.asciz	"armv5te"
+	string	cpu_feroceon_name, "Feroceon"
-	.size	cpu_arch_name, . - cpu_arch_name
+	string	cpu_88fr531_name, "Feroceon 88FR531-vd"
-
+	string	cpu_88fr571_name, "Feroceon 88FR571-vd"
-	.type	cpu_elf_name, #object
+	string	cpu_88fr131_name, "Feroceon 88FR131"
-cpu_elf_name:
-	.asciz	"v5"
-	.size	cpu_elf_name, . - cpu_elf_name
-
-	.type	cpu_feroceon_name, #object
-cpu_feroceon_name:
-	.asciz	"Feroceon"
-	.size	cpu_feroceon_name, . - cpu_feroceon_name
-
-	.type	cpu_88fr531_name, #object
-cpu_88fr531_name:
-	.asciz	"Feroceon 88FR531-vd"
-	.size	cpu_88fr531_name, . - cpu_88fr531_name
-
-	.type	cpu_88fr571_name, #object
-cpu_88fr571_name:
-	.asciz	"Feroceon 88FR571-vd"
-	.size	cpu_88fr571_name, . - cpu_88fr571_name
-
-	.type	cpu_88fr131_name, #object
-cpu_88fr131_name:
-	.asciz	"Feroceon 88FR131"
-	.size	cpu_88fr131_name, . - cpu_88fr131_name
 
 	.align
 
 	.section ".proc.info.init", #alloc, #execinstr
 
-#ifdef CONFIG_CPU_FEROCEON_OLD_ID
+.macro feroceon_proc_info name:req, cpu_val:req, cpu_mask:req, cpu_name:req, cache:req
-	.type	__feroceon_old_id_proc_info,#object
+	.type	__\name\()_proc_info,#object
-__feroceon_old_id_proc_info:
+__\name\()_proc_info:
-	.long	0x41009260
+	.long	\cpu_val
-	.long	0xff00fff0
+	.long	\cpu_mask
 	.long	PMD_TYPE_SECT | \
 		PMD_SECT_BUFFERABLE | \
 		PMD_SECT_CACHEABLE | \
@@ -614,85 +573,22 @@ __feroceon_old_id_proc_info:
 	.long	cpu_arch_name
 	.long	cpu_elf_name
 	.long	HWCAP_SWP|HWCAP_HALF|HWCAP_THUMB|HWCAP_FAST_MULT|HWCAP_EDSP
-	.long	cpu_feroceon_name
+	.long	\cpu_name
 	.long	feroceon_processor_functions
 	.long	v4wbi_tlb_fns
 	.long	feroceon_user_fns
-	.long	feroceon_cache_fns
+	.long	\cache
-	.size	__feroceon_old_id_proc_info, . - __feroceon_old_id_proc_info
+	 .size	__\name\()_proc_info, . - __\name\()_proc_info
+.endm
+
+#ifdef CONFIG_CPU_FEROCEON_OLD_ID
+	feroceon_proc_info feroceon_old_id, 0x41009260, 0xff00fff0, \
+		cpu_name=cpu_feroceon_name, cache=feroceon_cache_fns
 #endif
 
-	.type	__88fr531_proc_info,#object
+	feroceon_proc_info 88fr531, 0x56055310, 0xfffffff0, cpu_88fr531_name, \
-__88fr531_proc_info:
+		cache=feroceon_cache_fns
-	.long	0x56055310
+	feroceon_proc_info 88fr571, 0x56155710, 0xfffffff0, cpu_88fr571_name, \
-	.long	0xfffffff0
+		cache=feroceon_range_cache_fns
-	.long	PMD_TYPE_SECT | \
+	feroceon_proc_info 88fr131, 0x56251310, 0xfffffff0, cpu_88fr131_name, \
-		PMD_SECT_BUFFERABLE | \
+		cache=feroceon_range_cache_fns
-		PMD_SECT_CACHEABLE | \
-		PMD_BIT4 | \
-		PMD_SECT_AP_WRITE | \
-		PMD_SECT_AP_READ
-	.long	PMD_TYPE_SECT | \
-		PMD_BIT4 | \
-		PMD_SECT_AP_WRITE | \
-		PMD_SECT_AP_READ
-	b	__feroceon_setup
-	.long	cpu_arch_name
-	.long	cpu_elf_name
-	.long	HWCAP_SWP|HWCAP_HALF|HWCAP_THUMB|HWCAP_FAST_MULT|HWCAP_EDSP
-	.long	cpu_88fr531_name
-	.long	feroceon_processor_functions
-	.long	v4wbi_tlb_fns
-	.long	feroceon_user_fns
-	.long	feroceon_cache_fns
-	.size	__88fr531_proc_info, . - __88fr531_proc_info
-
-	.type	__88fr571_proc_info,#object
-__88fr571_proc_info:
-	.long	0x56155710
-	.long	0xfffffff0
-	.long	PMD_TYPE_SECT | \
-		PMD_SECT_BUFFERABLE | \
-		PMD_SECT_CACHEABLE | \
-		PMD_BIT4 | \
-		PMD_SECT_AP_WRITE | \
-		PMD_SECT_AP_READ
-	.long	PMD_TYPE_SECT | \
-		PMD_BIT4 | \
-		PMD_SECT_AP_WRITE | \
-		PMD_SECT_AP_READ
-	b	__feroceon_setup
-	.long	cpu_arch_name
-	.long	cpu_elf_name
-	.long	HWCAP_SWP|HWCAP_HALF|HWCAP_THUMB|HWCAP_FAST_MULT|HWCAP_EDSP
-	.long	cpu_88fr571_name
-	.long	feroceon_processor_functions
-	.long	v4wbi_tlb_fns
-	.long	feroceon_user_fns
-	.long	feroceon_range_cache_fns
-	.size	__88fr571_proc_info, . - __88fr571_proc_info
-
-	.type	__88fr131_proc_info,#object
-__88fr131_proc_info:
-	.long	0x56251310
-	.long	0xfffffff0
-	.long	PMD_TYPE_SECT | \
-		PMD_SECT_BUFFERABLE | \
-		PMD_SECT_CACHEABLE | \
-		PMD_BIT4 | \
-		PMD_SECT_AP_WRITE | \
-		PMD_SECT_AP_READ
-	.long	PMD_TYPE_SECT | \
-		PMD_BIT4 | \
-		PMD_SECT_AP_WRITE | \
-		PMD_SECT_AP_READ
-	b	__feroceon_setup
-	.long	cpu_arch_name
-	.long	cpu_elf_name
-	.long	HWCAP_SWP|HWCAP_HALF|HWCAP_THUMB|HWCAP_FAST_MULT|HWCAP_EDSP
-	.long	cpu_88fr131_name
-	.long	feroceon_processor_functions
-	.long	v4wbi_tlb_fns
-	.long	feroceon_user_fns
-	.long	feroceon_range_cache_fns
-	.size	__88fr131_proc_info, . - __88fr131_proc_info

arch/arm/mm/proc-macros.S

@@ -254,3 +254,71 @@
 	mcr	p15, 0, r0, c7, c10, 1		@ clean L1 D line
 	mcr	p15, 0, ip, c7, c10, 4		@ data write barrier
 	.endm
+
+.macro define_processor_functions name:req, dabort:req, pabort:req, nommu=0, suspend=0
+	.type	\name\()_processor_functions, #object
+	.align 2
+ENTRY(\name\()_processor_functions)
+	.word	\dabort
+	.word	\pabort
+	.word	cpu_\name\()_proc_init
+	.word	cpu_\name\()_proc_fin
+	.word	cpu_\name\()_reset
+	.word	cpu_\name\()_do_idle
+	.word	cpu_\name\()_dcache_clean_area
+	.word	cpu_\name\()_switch_mm
+
+	.if \nommu
+	.word	0
+	.else
+	.word	cpu_\name\()_set_pte_ext
+	.endif
+
+	.if \suspend
+	.word	cpu_\name\()_suspend_size
+#ifdef CONFIG_PM_SLEEP
+	.word	cpu_\name\()_do_suspend
+	.word	cpu_\name\()_do_resume
+#else
+	.word	0
+	.word	0
+#endif
+	.else
+	.word	0
+	.word	0
+	.word	0
+	.endif
+
+	.size	\name\()_processor_functions, . - \name\()_processor_functions
+.endm
+
+.macro define_cache_functions name:req
+	.align 2
+	.type	\name\()_cache_fns, #object
+ENTRY(\name\()_cache_fns)
+	.long	\name\()_flush_icache_all
+	.long	\name\()_flush_kern_cache_all
+	.long	\name\()_flush_user_cache_all
+	.long	\name\()_flush_user_cache_range
+	.long	\name\()_coherent_kern_range
+	.long	\name\()_coherent_user_range
+	.long	\name\()_flush_kern_dcache_area
+	.long	\name\()_dma_map_area
+	.long	\name\()_dma_unmap_area
+	.long	\name\()_dma_flush_range
+	.size	\name\()_cache_fns, . - \name\()_cache_fns
+.endm
+
+.macro define_tlb_functions name:req, flags_up:req, flags_smp
+	.type	\name\()_tlb_fns, #object
+ENTRY(\name\()_tlb_fns)
+	.long	\name\()_flush_user_tlb_range
+	.long	\name\()_flush_kern_tlb_range
+	.ifnb \flags_smp
+		ALT_SMP(.long	\flags_smp )
+		ALT_UP(.long	\flags_up )
+	.else
+		.long	\flags_up
+	.endif
+	.size	\name\()_tlb_fns, . - \name\()_tlb_fns
+.endm

arch/arm/mm/proc-mohawk.S

@@ -92,6 +92,17 @@ ENTRY(cpu_mohawk_do_idle)
 	mcr	p15, 0, r0, c7, c0, 4		@ wait for interrupt
 	mov	pc, lr
 
+/*
+ *	flush_icache_all()
+ *
+ *	Unconditionally clean and invalidate the entire icache.
+ */
+ENTRY(mohawk_flush_icache_all)
+	mov	r0, #0
+	mcr	p15, 0, r0, c7, c5, 0		@ invalidate I cache
+	mov	pc, lr
+ENDPROC(mohawk_flush_icache_all)
+
 /*
  *	flush_user_cache_all()
  *
@@ -288,16 +299,8 @@ ENTRY(mohawk_dma_unmap_area)
 	mov	pc, lr
 ENDPROC(mohawk_dma_unmap_area)
 
-ENTRY(mohawk_cache_fns)
+	@ define struct cpu_cache_fns (see <asm/cacheflush.h> and proc-macros.S)
-	.long	mohawk_flush_kern_cache_all
+	define_cache_functions mohawk
-	.long	mohawk_flush_user_cache_all
-	.long	mohawk_flush_user_cache_range
-	.long	mohawk_coherent_kern_range
-	.long	mohawk_coherent_user_range
-	.long	mohawk_flush_kern_dcache_area
-	.long	mohawk_dma_map_area
-	.long	mohawk_dma_unmap_area
-	.long	mohawk_dma_flush_range
 
 ENTRY(cpu_mohawk_dcache_clean_area)
 1:	mcr	p15, 0, r0, c7, c10, 1		@ clean D entry
@@ -373,42 +376,14 @@ mohawk_crval:
 
 	__INITDATA
 
-/*
+	@ define struct processor (see <asm/proc-fns.h> and proc-macros.S)
- * Purpose : Function pointers used to access above functions - all calls
+	define_processor_functions mohawk, dabort=v5t_early_abort, pabort=legacy_pabort
- *	     come through these
- */
-	.type	mohawk_processor_functions, #object
-mohawk_processor_functions:
-	.word	v5t_early_abort
-	.word	legacy_pabort
-	.word	cpu_mohawk_proc_init
-	.word	cpu_mohawk_proc_fin
-	.word	cpu_mohawk_reset
-	.word	cpu_mohawk_do_idle
-	.word	cpu_mohawk_dcache_clean_area
-	.word	cpu_mohawk_switch_mm
-	.word	cpu_mohawk_set_pte_ext
-	.word	0
-	.word	0
-	.word	0
-	.size	mohawk_processor_functions, . - mohawk_processor_functions
 
 	.section ".rodata"
 
-	.type	cpu_arch_name, #object
+	string	cpu_arch_name, "armv5te"
-cpu_arch_name:
+	string	cpu_elf_name, "v5"
-	.asciz	"armv5te"
+	string	cpu_mohawk_name, "Marvell 88SV331x"
-	.size	cpu_arch_name, . - cpu_arch_name
-
-	.type	cpu_elf_name, #object
-cpu_elf_name:
-	.asciz	"v5"
-	.size	cpu_elf_name, . - cpu_elf_name
-
-	.type	cpu_mohawk_name, #object
-cpu_mohawk_name:
-	.asciz	"Marvell 88SV331x"
-	.size	cpu_mohawk_name, . - cpu_mohawk_name
 
 	.align
 

arch/arm/mm/proc-sa110.S

@@ -187,43 +187,14 @@ sa110_crval:
 
 	__INITDATA
 
-/*
+	@ define struct processor (see <asm/proc-fns.h> and proc-macros.S)
- * Purpose : Function pointers used to access above functions - all calls
+	define_processor_functions sa110, dabort=v4_early_abort, pabort=legacy_pabort
- *	     come through these
- */
-
-	.type	sa110_processor_functions, #object
-ENTRY(sa110_processor_functions)
-	.word	v4_early_abort
-	.word	legacy_pabort
-	.word	cpu_sa110_proc_init
-	.word	cpu_sa110_proc_fin
-	.word	cpu_sa110_reset
-	.word	cpu_sa110_do_idle
-	.word	cpu_sa110_dcache_clean_area
-	.word	cpu_sa110_switch_mm
-	.word	cpu_sa110_set_pte_ext
-	.word	0
-	.word	0
-	.word	0
-	.size	sa110_processor_functions, . - sa110_processor_functions
 
 	.section ".rodata"
 
-	.type	cpu_arch_name, #object
+	string	cpu_arch_name, "armv4"
-cpu_arch_name:
+	string	cpu_elf_name, "v4"
-	.asciz	"armv4"
+	string	cpu_sa110_name, "StrongARM-110"
-	.size	cpu_arch_name, . - cpu_arch_name
-
-	.type	cpu_elf_name, #object
-cpu_elf_name:
-	.asciz	"v4"
-	.size	cpu_elf_name, . - cpu_elf_name
-
-	.type	cpu_sa110_name, #object
-cpu_sa110_name:
-	.asciz	"StrongARM-110"
-	.size	cpu_sa110_name, . - cpu_sa110_name
 
 	.align
 

arch/arm/mm/proc-sa1100.S

@@ -198,9 +198,6 @@ ENTRY(cpu_sa1100_do_resume)
 		     PMD_SECT_CACHEABLE | PMD_SECT_AP_WRITE
 	b	cpu_resume_mmu
 ENDPROC(cpu_sa1100_do_resume)
-#else
-#define cpu_sa1100_do_suspend	0
-#define cpu_sa1100_do_resume	0
 #endif
 
 	__CPUINIT
@@ -233,60 +230,29 @@ sa1100_crval:
 
 	__INITDATA
 
-/*
- * Purpose : Function pointers used to access above functions - all calls
- *	     come through these
- */
-
 /*
  * SA1100 and SA1110 share the same function calls
  */
-	.type	sa1100_processor_functions, #object
+
-ENTRY(sa1100_processor_functions)
+	@ define struct processor (see <asm/proc-fns.h> and proc-macros.S)
-	.word	v4_early_abort
+	define_processor_functions sa1100, dabort=v4_early_abort, pabort=legacy_pabort, suspend=1
-	.word	legacy_pabort
-	.word	cpu_sa1100_proc_init
-	.word	cpu_sa1100_proc_fin
-	.word	cpu_sa1100_reset
-	.word	cpu_sa1100_do_idle
-	.word	cpu_sa1100_dcache_clean_area
-	.word	cpu_sa1100_switch_mm
-	.word	cpu_sa1100_set_pte_ext
-	.word	cpu_sa1100_suspend_size
-	.word	cpu_sa1100_do_suspend
-	.word	cpu_sa1100_do_resume
-	.size	sa1100_processor_functions, . - sa1100_processor_functions
 
 	.section ".rodata"
 
-	.type	cpu_arch_name, #object
+	string	cpu_arch_name, "armv4"
-cpu_arch_name:
+	string	cpu_elf_name, "v4"
-	.asciz	"armv4"
+	string	cpu_sa1100_name, "StrongARM-1100"
-	.size	cpu_arch_name, . - cpu_arch_name
+	string	cpu_sa1110_name, "StrongARM-1110"
-
-	.type	cpu_elf_name, #object
-cpu_elf_name:
-	.asciz	"v4"
-	.size	cpu_elf_name, . - cpu_elf_name
-
-	.type	cpu_sa1100_name, #object
-cpu_sa1100_name:
-	.asciz	"StrongARM-1100"
-	.size	cpu_sa1100_name, . - cpu_sa1100_name
-
-	.type	cpu_sa1110_name, #object
-cpu_sa1110_name:
-	.asciz	"StrongARM-1110"
-	.size	cpu_sa1110_name, . - cpu_sa1110_name
 
 	.align
 
 	.section ".proc.info.init", #alloc, #execinstr
 
-	.type	__sa1100_proc_info,#object
+.macro sa1100_proc_info name:req, cpu_val:req, cpu_mask:req, cpu_name:req
-__sa1100_proc_info:
+	.type	__\name\()_proc_info,#object
-	.long	0x4401a110
+__\name\()_proc_info:
-	.long	0xfffffff0
+	.long	\cpu_val
+	.long	\cpu_mask
 	.long   PMD_TYPE_SECT | \
 		PMD_SECT_BUFFERABLE | \
 		PMD_SECT_CACHEABLE | \
@@ -299,32 +265,13 @@ __sa1100_proc_info:
 	.long	cpu_arch_name
 	.long	cpu_elf_name
 	.long	HWCAP_SWP | HWCAP_HALF | HWCAP_26BIT | HWCAP_FAST_MULT
-	.long	cpu_sa1100_name
+	.long	\cpu_name
 	.long	sa1100_processor_functions
 	.long	v4wb_tlb_fns
 	.long	v4_mc_user_fns
 	.long	v4wb_cache_fns
-	.size	__sa1100_proc_info, . - __sa1100_proc_info
+	.size	__\name\()_proc_info, . - __\name\()_proc_info
+.endm
 
-	.type	__sa1110_proc_info,#object
+	sa1100_proc_info sa1100, 0x4401a110, 0xfffffff0, cpu_sa1100_name
-__sa1110_proc_info:
+	sa1100_proc_info sa1110, 0x6901b110, 0xfffffff0, cpu_sa1110_name
-	.long	0x6901b110
-	.long	0xfffffff0
-	.long   PMD_TYPE_SECT | \
-		PMD_SECT_BUFFERABLE | \
-		PMD_SECT_CACHEABLE | \
-		PMD_SECT_AP_WRITE | \
-		PMD_SECT_AP_READ
-	.long   PMD_TYPE_SECT | \
-		PMD_SECT_AP_WRITE | \
-		PMD_SECT_AP_READ
-	b	__sa1100_setup
-	.long	cpu_arch_name
-	.long	cpu_elf_name
-	.long	HWCAP_SWP | HWCAP_HALF | HWCAP_26BIT | HWCAP_FAST_MULT
-	.long	cpu_sa1110_name
-	.long	sa1100_processor_functions
-	.long	v4wb_tlb_fns
-	.long	v4_mc_user_fns
-	.long	v4wb_cache_fns
-	.size	__sa1110_proc_info, . - __sa1110_proc_info

arch/arm/mm/proc-v6.S

@@ -56,6 +56,11 @@ ENTRY(cpu_v6_proc_fin)
  */
 	.align	5
 ENTRY(cpu_v6_reset)
+	mrc	p15, 0, r1, c1, c0, 0		@ ctrl register
+	bic	r1, r1, #0x1			@ ...............m
+	mcr	p15, 0, r1, c1, c0, 0		@ disable MMU
+	mov	r1, #0
+	mcr	p15, 0, r1, c7, c5, 4		@ ISB
 	mov	pc, r0
 
 /*
@@ -164,16 +169,9 @@ ENDPROC(cpu_v6_do_resume)
 cpu_resume_l1_flags:
 	ALT_SMP(.long PMD_TYPE_SECT | PMD_SECT_AP_WRITE | PMD_FLAGS_SMP)
 	ALT_UP(.long  PMD_TYPE_SECT | PMD_SECT_AP_WRITE | PMD_FLAGS_UP)
-#else
-#define cpu_v6_do_suspend 0
-#define cpu_v6_do_resume 0
 #endif
 
-
+	string	cpu_v6_name, "ARMv6-compatible processor"
-	.type	cpu_v6_name, #object
-cpu_v6_name:
-	.asciz	"ARMv6-compatible processor"
-	.size	cpu_v6_name, . - cpu_v6_name
 
 	.align
 
@@ -239,33 +237,13 @@ v6_crval:
 
 	__INITDATA
 
-	.type	v6_processor_functions, #object
+	@ define struct processor (see <asm/proc-fns.h> and proc-macros.S)
-ENTRY(v6_processor_functions)
+	define_processor_functions v6, dabort=v6_early_abort, pabort=v6_pabort, suspend=1
-	.word	v6_early_abort
-	.word	v6_pabort
-	.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_ext
-	.word	cpu_v6_suspend_size
-	.word	cpu_v6_do_suspend
-	.word	cpu_v6_do_resume
-	.size	v6_processor_functions, . - v6_processor_functions
 
 	.section ".rodata"
 
-	.type	cpu_arch_name, #object
+	string	cpu_arch_name, "armv6"
-cpu_arch_name:
+	string	cpu_elf_name, "v6"
-	.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.init", #alloc, #execinstr

arch/arm/mm/proc-v7.S

@@ -58,9 +58,16 @@ ENDPROC(cpu_v7_proc_fin)
  *	to what would be the reset vector.
  *
  *	- loc   - location to jump to for soft reset
+ *
+ *	This code must be executed using a flat identity mapping with
+ *      caches disabled.
  */
 	.align	5
 ENTRY(cpu_v7_reset)
+	mrc	p15, 0, r1, c1, c0, 0		@ ctrl register
+	bic	r1, r1, #0x1			@ ...............m
+	mcr	p15, 0, r1, c1, c0, 0		@ disable MMU
+	isb
 	mov	pc, r0
 ENDPROC(cpu_v7_reset)
 
@@ -173,8 +180,7 @@ ENTRY(cpu_v7_set_pte_ext)
 	mov	pc, lr
 ENDPROC(cpu_v7_set_pte_ext)
 
-cpu_v7_name:
+	string	cpu_v7_name, "ARMv7 Processor"
-	.ascii	"ARMv7 Processor"
 	.align
 
 	/*
@@ -257,9 +263,6 @@ ENDPROC(cpu_v7_do_resume)
 cpu_resume_l1_flags:
 	ALT_SMP(.long PMD_TYPE_SECT | PMD_SECT_AP_WRITE | PMD_FLAGS_SMP)
 	ALT_UP(.long  PMD_TYPE_SECT | PMD_SECT_AP_WRITE | PMD_FLAGS_UP)
-#else
-#define cpu_v7_do_suspend	0
-#define cpu_v7_do_resume	0
 #endif
 
 	__CPUINIT
@@ -279,13 +282,20 @@ cpu_resume_l1_flags:
  *	It is assumed that:
  *	- cache type register is implemented
  */
+__v7_ca5mp_setup:
 __v7_ca9mp_setup:
+	mov	r10, #(1 << 0)			@ TLB ops broadcasting
+	b	1f
+__v7_ca15mp_setup:
+	mov	r10, #0
+1:
 #ifdef CONFIG_SMP
 	ALT_SMP(mrc	p15, 0, r0, c1, c0, 1)
 	ALT_UP(mov	r0, #(1 << 6))		@ fake it for UP
 	tst	r0, #(1 << 6)			@ SMP/nAMP mode enabled?
-	orreq	r0, r0, #(1 << 6) | (1 << 0)	@ Enable SMP/nAMP mode and
+	orreq	r0, r0, #(1 << 6)		@ Enable SMP/nAMP mode
-	mcreq	p15, 0, r0, c1, c0, 1		@ TLB ops broadcasting
+	orreq	r0, r0, r10			@ Enable CPU-specific SMP bits
+	mcreq	p15, 0, r0, c1, c0, 1
 #endif
 __v7_setup:
 	adr	r12, __v7_setup_stack		@ the local stack
@@ -411,66 +421,69 @@ __v7_setup_stack:
 
 	__INITDATA
 
-	.type	v7_processor_functions, #object
+	@ define struct processor (see <asm/proc-fns.h> and proc-macros.S)
-ENTRY(v7_processor_functions)
+	define_processor_functions v7, dabort=v7_early_abort, pabort=v7_pabort, suspend=1
-	.word	v7_early_abort
-	.word	v7_pabort
-	.word	cpu_v7_proc_init
-	.word	cpu_v7_proc_fin
-	.word	cpu_v7_reset
-	.word	cpu_v7_do_idle
-	.word	cpu_v7_dcache_clean_area
-	.word	cpu_v7_switch_mm
-	.word	cpu_v7_set_pte_ext
-	.word	cpu_v7_suspend_size
-	.word	cpu_v7_do_suspend
-	.word	cpu_v7_do_resume
-	.size	v7_processor_functions, . - v7_processor_functions
 
 	.section ".rodata"
 
-	.type	cpu_arch_name, #object
+	string	cpu_arch_name, "armv7"
-cpu_arch_name:
+	string	cpu_elf_name, "v7"
-	.asciz	"armv7"
-	.size	cpu_arch_name, . - cpu_arch_name
-
-	.type	cpu_elf_name, #object
-cpu_elf_name:
-	.asciz	"v7"
-	.size	cpu_elf_name, . - cpu_elf_name
 	.align
 
 	.section ".proc.info.init", #alloc, #execinstr
 
-	.type   __v7_ca9mp_proc_info, #object
+	/*
-__v7_ca9mp_proc_info:
+	 * Standard v7 proc info content
-	.long	0x410fc090		@ Required ID value
+	 */
-	.long	0xff0ffff0		@ Mask for ID
+.macro __v7_proc initfunc, mm_mmuflags = 0, io_mmuflags = 0, hwcaps = 0
-	ALT_SMP(.long \
+	ALT_SMP(.long	PMD_TYPE_SECT | PMD_SECT_AP_WRITE | PMD_SECT_AP_READ | \
-		PMD_TYPE_SECT | \
+			PMD_FLAGS_SMP | \mm_mmuflags)
-		PMD_SECT_AP_WRITE | \
+	ALT_UP(.long	PMD_TYPE_SECT | PMD_SECT_AP_WRITE | PMD_SECT_AP_READ | \
-		PMD_SECT_AP_READ | \
+			PMD_FLAGS_UP | \mm_mmuflags)
-		PMD_FLAGS_SMP)
+	.long	PMD_TYPE_SECT | PMD_SECT_XN | PMD_SECT_AP_WRITE | \
-	ALT_UP(.long \
+		PMD_SECT_AP_READ | \io_mmuflags
-		PMD_TYPE_SECT | \
+	W(b)	\initfunc
-		PMD_SECT_AP_WRITE | \
-		PMD_SECT_AP_READ | \
-		PMD_FLAGS_UP)
-	.long   PMD_TYPE_SECT | \
-		PMD_SECT_XN | \
-		PMD_SECT_AP_WRITE | \
-		PMD_SECT_AP_READ
-	W(b)	__v7_ca9mp_setup
 	.long	cpu_arch_name
 	.long	cpu_elf_name
-	.long	HWCAP_SWP|HWCAP_HALF|HWCAP_THUMB|HWCAP_FAST_MULT|HWCAP_EDSP|HWCAP_TLS
+	.long	HWCAP_SWP | HWCAP_HALF | HWCAP_THUMB | HWCAP_FAST_MULT | \
+		HWCAP_EDSP | HWCAP_TLS | \hwcaps
 	.long	cpu_v7_name
 	.long	v7_processor_functions
 	.long	v7wbi_tlb_fns
 	.long	v6_user_fns
 	.long	v7_cache_fns
+.endm
+
+	/*
+	 * ARM Ltd. Cortex A5 processor.
+	 */
+	.type   __v7_ca5mp_proc_info, #object
+__v7_ca5mp_proc_info:
+	.long	0x410fc050
+	.long	0xff0ffff0
+	__v7_proc __v7_ca5mp_setup
+	.size	__v7_ca5mp_proc_info, . - __v7_ca5mp_proc_info
+
+	/*
+	 * ARM Ltd. Cortex A9 processor.
+	 */
+	.type   __v7_ca9mp_proc_info, #object
+__v7_ca9mp_proc_info:
+	.long	0x410fc090
+	.long	0xff0ffff0
+	__v7_proc __v7_ca9mp_setup
 	.size	__v7_ca9mp_proc_info, . - __v7_ca9mp_proc_info
 
+	/*
+	 * ARM Ltd. Cortex A15 processor.
+	 */
+	.type	__v7_ca15mp_proc_info, #object
+__v7_ca15mp_proc_info:
+	.long	0x410fc0f0
+	.long	0xff0ffff0
+	__v7_proc __v7_ca15mp_setup, hwcaps = HWCAP_IDIV
+	.size	__v7_ca15mp_proc_info, . - __v7_ca15mp_proc_info
+
 	/*
 	 * Match any ARMv7 processor core.
 	 */
@@ -478,27 +491,5 @@ __v7_ca9mp_proc_info:
 __v7_proc_info:
 	.long	0x000f0000		@ Required ID value
 	.long	0x000f0000		@ Mask for ID
-	ALT_SMP(.long \
+	__v7_proc __v7_setup
-		PMD_TYPE_SECT | \
-		PMD_SECT_AP_WRITE | \
-		PMD_SECT_AP_READ | \
-		PMD_FLAGS_SMP)
-	ALT_UP(.long \
-		PMD_TYPE_SECT | \
-		PMD_SECT_AP_WRITE | \
-		PMD_SECT_AP_READ | \
-		PMD_FLAGS_UP)
-	.long   PMD_TYPE_SECT | \
-		PMD_SECT_XN | \
-		PMD_SECT_AP_WRITE | \
-		PMD_SECT_AP_READ
-	W(b)	__v7_setup
-	.long	cpu_arch_name
-	.long	cpu_elf_name
-	.long	HWCAP_SWP|HWCAP_HALF|HWCAP_THUMB|HWCAP_FAST_MULT|HWCAP_EDSP|HWCAP_TLS
-	.long	cpu_v7_name
-	.long	v7_processor_functions
-	.long	v7wbi_tlb_fns
-	.long	v6_user_fns
-	.long	v7_cache_fns
 	.size	__v7_proc_info, . - __v7_proc_info

arch/arm/mm/proc-xsc3.S

@@ -335,17 +335,8 @@ ENTRY(xsc3_dma_unmap_area)
 	mov	pc, lr
 ENDPROC(xsc3_dma_unmap_area)
 
-ENTRY(xsc3_cache_fns)
+	@ define struct cpu_cache_fns (see <asm/cacheflush.h> and proc-macros.S)
-	.long	xsc3_flush_icache_all
+	define_cache_functions xsc3
-	.long	xsc3_flush_kern_cache_all
-	.long	xsc3_flush_user_cache_all
-	.long	xsc3_flush_user_cache_range
-	.long	xsc3_coherent_kern_range
-	.long	xsc3_coherent_user_range
-	.long	xsc3_flush_kern_dcache_area
-	.long	xsc3_dma_map_area
-	.long	xsc3_dma_unmap_area
-	.long	xsc3_dma_flush_range
 
 ENTRY(cpu_xsc3_dcache_clean_area)
 1:	mcr	p15, 0, r0, c7, c10, 1		@ clean L1 D line
@@ -454,9 +445,6 @@ ENTRY(cpu_xsc3_do_resume)
 	ldr	r3, =0x542e		@ section flags
 	b	cpu_resume_mmu
 ENDPROC(cpu_xsc3_do_resume)
-#else
-#define cpu_xsc3_do_suspend	0
-#define cpu_xsc3_do_resume	0
 #endif
 
 	__CPUINIT
@@ -503,52 +491,24 @@ xsc3_crval:
 
 	__INITDATA
 
-/*
+	@ define struct processor (see <asm/proc-fns.h> and proc-macros.S)
- * Purpose : Function pointers used to access above functions - all calls
+	define_processor_functions xsc3, dabort=v5t_early_abort, pabort=legacy_pabort, suspend=1
- *	     come through these
- */
-
-	.type	xsc3_processor_functions, #object
-ENTRY(xsc3_processor_functions)
-	.word	v5t_early_abort
-	.word	legacy_pabort
-	.word	cpu_xsc3_proc_init
-	.word	cpu_xsc3_proc_fin
-	.word	cpu_xsc3_reset
-	.word	cpu_xsc3_do_idle
-	.word	cpu_xsc3_dcache_clean_area
-	.word	cpu_xsc3_switch_mm
-	.word	cpu_xsc3_set_pte_ext
-	.word	cpu_xsc3_suspend_size
-	.word	cpu_xsc3_do_suspend
-	.word	cpu_xsc3_do_resume
-	.size	xsc3_processor_functions, . - xsc3_processor_functions
 
 	.section ".rodata"
 
-	.type	cpu_arch_name, #object
+	string	cpu_arch_name, "armv5te"
-cpu_arch_name:
+	string	cpu_elf_name, "v5"
-	.asciz	"armv5te"
+	string	cpu_xsc3_name, "XScale-V3 based processor"
-	.size	cpu_arch_name, . - cpu_arch_name
-
-	.type	cpu_elf_name, #object
-cpu_elf_name:
-	.asciz	"v5"
-	.size	cpu_elf_name, . - cpu_elf_name
-
-	.type	cpu_xsc3_name, #object
-cpu_xsc3_name:
-	.asciz	"XScale-V3 based processor"
-	.size	cpu_xsc3_name, . - cpu_xsc3_name
 
 	.align
 
 	.section ".proc.info.init", #alloc, #execinstr
 
-	.type	__xsc3_proc_info,#object
+.macro xsc3_proc_info name:req, cpu_val:req, cpu_mask:req
-__xsc3_proc_info:
+	.type	__\name\()_proc_info,#object
-	.long	0x69056000
+__\name\()_proc_info:
-	.long	0xffffe000
+	.long	\cpu_val
+	.long	\cpu_mask
 	.long	PMD_TYPE_SECT | \
 		PMD_SECT_BUFFERABLE | \
 		PMD_SECT_CACHEABLE | \
@@ -566,29 +526,10 @@ __xsc3_proc_info:
 	.long	v4wbi_tlb_fns
 	.long	xsc3_mc_user_fns
 	.long	xsc3_cache_fns
-	.size	__xsc3_proc_info, . - __xsc3_proc_info
+	.size	__\name\()_proc_info, . - __\name\()_proc_info
+.endm
+
+	xsc3_proc_info xsc3, 0x69056000, 0xffffe000
 
 /* Note: PXA935 changed its implementor ID from Intel to Marvell */
-
+	xsc3_proc_info xsc3_pxa935, 0x56056000, 0xffffe000
-	.type	__xsc3_pxa935_proc_info,#object
-__xsc3_pxa935_proc_info:
-	.long	0x56056000
-	.long	0xffffe000
-	.long	PMD_TYPE_SECT | \
-		PMD_SECT_BUFFERABLE | \
-		PMD_SECT_CACHEABLE | \
-		PMD_SECT_AP_WRITE | \
-		PMD_SECT_AP_READ
-	.long	PMD_TYPE_SECT | \
-		PMD_SECT_AP_WRITE | \
-		PMD_SECT_AP_READ
-	b	__xsc3_setup
-	.long	cpu_arch_name
-	.long	cpu_elf_name
-	.long	HWCAP_SWP|HWCAP_HALF|HWCAP_THUMB|HWCAP_FAST_MULT|HWCAP_EDSP
-	.long	cpu_xsc3_name
-	.long	xsc3_processor_functions
-	.long	v4wbi_tlb_fns
-	.long	xsc3_mc_user_fns
-	.long	xsc3_cache_fns
-	.size	__xsc3_pxa935_proc_info, . - __xsc3_pxa935_proc_info

arch/arm/mm/proc-xscale.S

@@ -390,12 +390,12 @@ ENDPROC(xscale_dma_map_area)
  *	- size	- size of region
  *	- dir	- DMA direction
  */
-ENTRY(xscale_dma_a0_map_area)
+ENTRY(xscale_80200_A0_A1_dma_map_area)
 	add	r1, r1, r0
 	teq	r2, #DMA_TO_DEVICE
 	beq	xscale_dma_clean_range
 	b	xscale_dma_flush_range
-ENDPROC(xscale_dma_a0_map_area)
+ENDPROC(xscale_80200_A0_A1_dma_map_area)
 
 /*
  *	dma_unmap_area(start, size, dir)
@@ -407,17 +407,8 @@ ENTRY(xscale_dma_unmap_area)
 	mov	pc, lr
 ENDPROC(xscale_dma_unmap_area)
 
-ENTRY(xscale_cache_fns)
+	@ define struct cpu_cache_fns (see <asm/cacheflush.h> and proc-macros.S)
-	.long	xscale_flush_icache_all
+	define_cache_functions xscale
-	.long	xscale_flush_kern_cache_all
-	.long	xscale_flush_user_cache_all
-	.long	xscale_flush_user_cache_range
-	.long	xscale_coherent_kern_range
-	.long	xscale_coherent_user_range
-	.long	xscale_flush_kern_dcache_area
-	.long	xscale_dma_map_area
-	.long	xscale_dma_unmap_area
-	.long	xscale_dma_flush_range
 
 /*
  * On stepping A0/A1 of the 80200, invalidating D-cache by line doesn't
@@ -432,16 +423,28 @@ ENTRY(xscale_cache_fns)
  * revision January 22, 2003, available at:
  *     http://www.intel.com/design/iio/specupdt/273415.htm
  */
-ENTRY(xscale_80200_A0_A1_cache_fns)
+.macro a0_alias basename
-	.long	xscale_flush_kern_cache_all
+	.globl xscale_80200_A0_A1_\basename
-	.long	xscale_flush_user_cache_all
+	.type xscale_80200_A0_A1_\basename , %function
-	.long	xscale_flush_user_cache_range
+	.equ xscale_80200_A0_A1_\basename , xscale_\basename
-	.long	xscale_coherent_kern_range
+.endm
-	.long	xscale_coherent_user_range
+
-	.long	xscale_flush_kern_dcache_area
+/*
-	.long	xscale_dma_a0_map_area
+ * Most of the cache functions are unchanged for these processor revisions.
-	.long	xscale_dma_unmap_area
+ * Export suitable alias symbols for the unchanged functions:
-	.long	xscale_dma_flush_range
+ */
+	a0_alias flush_icache_all
+	a0_alias flush_user_cache_all
+	a0_alias flush_kern_cache_all
+	a0_alias flush_user_cache_range
+	a0_alias coherent_kern_range
+	a0_alias coherent_user_range
+	a0_alias flush_kern_dcache_area
+	a0_alias dma_flush_range
+	a0_alias dma_unmap_area
+
+	@ define struct cpu_cache_fns (see <asm/cacheflush.h> and proc-macros.S)
+	define_cache_functions xscale_80200_A0_A1
 
 ENTRY(cpu_xscale_dcache_clean_area)
 1:	mcr	p15, 0, r0, c7, c10, 1		@ clean D entry
@@ -551,9 +554,6 @@ ENTRY(cpu_xscale_do_resume)
 		     PMD_SECT_CACHEABLE | PMD_SECT_AP_WRITE
 	b	cpu_resume_mmu
 ENDPROC(cpu_xscale_do_resume)
-#else
-#define cpu_xscale_do_suspend	0
-#define cpu_xscale_do_resume	0
 #endif
 
 	__CPUINIT
@@ -587,432 +587,74 @@ xscale_crval:
 
 	__INITDATA
 
-/*
+	@ define struct processor (see <asm/proc-fns.h> and proc-macros.S)
- * Purpose : Function pointers used to access above functions - all calls
+	define_processor_functions xscale, dabort=v5t_early_abort, pabort=legacy_pabort, suspend=1
- *	     come through these
- */
-
-	.type	xscale_processor_functions, #object
-ENTRY(xscale_processor_functions)
-	.word	v5t_early_abort
-	.word	legacy_pabort
-	.word	cpu_xscale_proc_init
-	.word	cpu_xscale_proc_fin
-	.word	cpu_xscale_reset
-	.word	cpu_xscale_do_idle
-	.word	cpu_xscale_dcache_clean_area
-	.word	cpu_xscale_switch_mm
-	.word	cpu_xscale_set_pte_ext
-	.word	cpu_xscale_suspend_size
-	.word	cpu_xscale_do_suspend
-	.word	cpu_xscale_do_resume
-	.size	xscale_processor_functions, . - xscale_processor_functions
 
 	.section ".rodata"
 
-	.type	cpu_arch_name, #object
+	string	cpu_arch_name, "armv5te"
-cpu_arch_name:
+	string	cpu_elf_name, "v5"
-	.asciz	"armv5te"
-	.size	cpu_arch_name, . - cpu_arch_name
 
-	.type	cpu_elf_name, #object
+	string	cpu_80200_A0_A1_name, "XScale-80200 A0/A1"
-cpu_elf_name:
+	string	cpu_80200_name, "XScale-80200"
-	.asciz	"v5"
+	string	cpu_80219_name, "XScale-80219"
-	.size	cpu_elf_name, . - cpu_elf_name
+	string	cpu_8032x_name, "XScale-IOP8032x Family"
-
+	string	cpu_8033x_name, "XScale-IOP8033x Family"
-	.type	cpu_80200_A0_A1_name, #object
+	string	cpu_pxa250_name, "XScale-PXA250"
-cpu_80200_A0_A1_name:
+	string	cpu_pxa210_name, "XScale-PXA210"
-	.asciz	"XScale-80200 A0/A1"
+	string	cpu_ixp42x_name, "XScale-IXP42x Family"
-	.size	cpu_80200_A0_A1_name, . - cpu_80200_A0_A1_name
+	string	cpu_ixp43x_name, "XScale-IXP43x Family"
-
+	string	cpu_ixp46x_name, "XScale-IXP46x Family"
-	.type	cpu_80200_name, #object
+	string	cpu_ixp2400_name, "XScale-IXP2400"
-cpu_80200_name:
+	string	cpu_ixp2800_name, "XScale-IXP2800"
-	.asciz	"XScale-80200"
+	string	cpu_pxa255_name, "XScale-PXA255"
-	.size	cpu_80200_name, . - cpu_80200_name
+	string	cpu_pxa270_name, "XScale-PXA270"
-
-	.type	cpu_80219_name, #object
-cpu_80219_name:
-	.asciz	"XScale-80219"
-	.size	cpu_80219_name, . - cpu_80219_name
-
-	.type	cpu_8032x_name, #object
-cpu_8032x_name:
-	.asciz	"XScale-IOP8032x Family"
-	.size	cpu_8032x_name, . - cpu_8032x_name
-
-	.type	cpu_8033x_name, #object
-cpu_8033x_name:
-	.asciz	"XScale-IOP8033x Family"
-	.size	cpu_8033x_name, . - cpu_8033x_name
-
-	.type	cpu_pxa250_name, #object
-cpu_pxa250_name:
-	.asciz	"XScale-PXA250"
-	.size	cpu_pxa250_name, . - cpu_pxa250_name
-
-	.type	cpu_pxa210_name, #object
-cpu_pxa210_name:
-	.asciz	"XScale-PXA210"
-	.size	cpu_pxa210_name, . - cpu_pxa210_name
-
-	.type	cpu_ixp42x_name, #object
-cpu_ixp42x_name:
-	.asciz	"XScale-IXP42x Family"
-	.size	cpu_ixp42x_name, . - cpu_ixp42x_name
-
-	.type	cpu_ixp43x_name, #object
-cpu_ixp43x_name:
-	.asciz	"XScale-IXP43x Family"
-	.size	cpu_ixp43x_name, . - cpu_ixp43x_name
-
-	.type	cpu_ixp46x_name, #object
-cpu_ixp46x_name:
-	.asciz	"XScale-IXP46x Family"
-	.size	cpu_ixp46x_name, . - cpu_ixp46x_name
-
-	.type	cpu_ixp2400_name, #object
-cpu_ixp2400_name:
-	.asciz	"XScale-IXP2400"
-	.size	cpu_ixp2400_name, . - cpu_ixp2400_name
-
-	.type	cpu_ixp2800_name, #object
-cpu_ixp2800_name:
-	.asciz	"XScale-IXP2800"
-	.size	cpu_ixp2800_name, . - cpu_ixp2800_name
-
-	.type	cpu_pxa255_name, #object
-cpu_pxa255_name:
-	.asciz	"XScale-PXA255"
-	.size	cpu_pxa255_name, . - cpu_pxa255_name
-
-	.type	cpu_pxa270_name, #object
-cpu_pxa270_name:
-	.asciz	"XScale-PXA270"
-	.size	cpu_pxa270_name, . - cpu_pxa270_name
 
 	.align
 
 	.section ".proc.info.init", #alloc, #execinstr
 
-	.type	__80200_A0_A1_proc_info,#object
+.macro xscale_proc_info name:req, cpu_val:req, cpu_mask:req, cpu_name:req, cache
-__80200_A0_A1_proc_info:
+	.type	__\name\()_proc_info,#object
-	.long	0x69052000
+__\name\()_proc_info:
-	.long	0xfffffffe
+	.long	\cpu_val
-	.long   PMD_TYPE_SECT | \
+	.long	\cpu_mask
+	.long	PMD_TYPE_SECT | \
 		PMD_SECT_BUFFERABLE | \
 		PMD_SECT_CACHEABLE | \
 		PMD_SECT_AP_WRITE | \
 		PMD_SECT_AP_READ
-	.long   PMD_TYPE_SECT | \
+	.long	PMD_TYPE_SECT | \
 		PMD_SECT_AP_WRITE | \
 		PMD_SECT_AP_READ
 	b	__xscale_setup
 	.long	cpu_arch_name
 	.long	cpu_elf_name
 	.long	HWCAP_SWP|HWCAP_HALF|HWCAP_THUMB|HWCAP_FAST_MULT|HWCAP_EDSP
-	.long	cpu_80200_name
+	.long	\cpu_name
 	.long	xscale_processor_functions
 	.long	v4wbi_tlb_fns
 	.long	xscale_mc_user_fns
-	.long	xscale_80200_A0_A1_cache_fns
+	.ifb \cache
-	.size	__80200_A0_A1_proc_info, . - __80200_A0_A1_proc_info
+		.long	xscale_cache_fns
-
+	.else
-	.type	__80200_proc_info,#object
+		.long	\cache
-__80200_proc_info:
+	.endif
-	.long	0x69052000
+	.size	__\name\()_proc_info, . - __\name\()_proc_info
-	.long	0xfffffff0
+.endm
-	.long   PMD_TYPE_SECT | \
-		PMD_SECT_BUFFERABLE | \
-		PMD_SECT_CACHEABLE | \
-		PMD_SECT_AP_WRITE | \
-		PMD_SECT_AP_READ
-	.long   PMD_TYPE_SECT | \
-		PMD_SECT_AP_WRITE | \
-		PMD_SECT_AP_READ
-	b	__xscale_setup
-	.long	cpu_arch_name
-	.long	cpu_elf_name
-	.long	HWCAP_SWP|HWCAP_HALF|HWCAP_THUMB|HWCAP_FAST_MULT|HWCAP_EDSP
-	.long	cpu_80200_name
-	.long	xscale_processor_functions
-	.long	v4wbi_tlb_fns
-	.long	xscale_mc_user_fns
-	.long	xscale_cache_fns
-	.size	__80200_proc_info, . - __80200_proc_info
-
-	.type	__80219_proc_info,#object
-__80219_proc_info:
-	.long	0x69052e20
-	.long	0xffffffe0
-	.long   PMD_TYPE_SECT | \
-		PMD_SECT_BUFFERABLE | \
-		PMD_SECT_CACHEABLE | \
-		PMD_SECT_AP_WRITE | \
-		PMD_SECT_AP_READ
-	.long   PMD_TYPE_SECT | \
-		PMD_SECT_AP_WRITE | \
-		PMD_SECT_AP_READ
-	b	__xscale_setup
-	.long	cpu_arch_name
-	.long	cpu_elf_name
-	.long	HWCAP_SWP|HWCAP_HALF|HWCAP_THUMB|HWCAP_FAST_MULT|HWCAP_EDSP
-	.long	cpu_80219_name
-	.long	xscale_processor_functions
-	.long	v4wbi_tlb_fns
-	.long	xscale_mc_user_fns
-	.long	xscale_cache_fns
-	.size	__80219_proc_info, . - __80219_proc_info
-
-	.type	__8032x_proc_info,#object
-__8032x_proc_info:
-	.long	0x69052420
-	.long	0xfffff7e0
-	.long   PMD_TYPE_SECT | \
-		PMD_SECT_BUFFERABLE | \
-		PMD_SECT_CACHEABLE | \
-		PMD_SECT_AP_WRITE | \
-		PMD_SECT_AP_READ
-	.long   PMD_TYPE_SECT | \
-		PMD_SECT_AP_WRITE | \
-		PMD_SECT_AP_READ
-	b	__xscale_setup
-	.long	cpu_arch_name
-	.long	cpu_elf_name
-	.long	HWCAP_SWP|HWCAP_HALF|HWCAP_THUMB|HWCAP_FAST_MULT|HWCAP_EDSP
-	.long	cpu_8032x_name
-	.long	xscale_processor_functions
-	.long	v4wbi_tlb_fns
-	.long	xscale_mc_user_fns
-	.long	xscale_cache_fns
-	.size	__8032x_proc_info, . - __8032x_proc_info
-
-	.type	__8033x_proc_info,#object
-__8033x_proc_info:
-	.long	0x69054010
-	.long	0xfffffd30
-	.long   PMD_TYPE_SECT | \
-		PMD_SECT_BUFFERABLE | \
-		PMD_SECT_CACHEABLE | \
-		PMD_SECT_AP_WRITE | \
-		PMD_SECT_AP_READ
-	.long   PMD_TYPE_SECT | \
-		PMD_SECT_AP_WRITE | \
-		PMD_SECT_AP_READ
-	b	__xscale_setup
-	.long	cpu_arch_name
-	.long	cpu_elf_name
-	.long	HWCAP_SWP|HWCAP_HALF|HWCAP_THUMB|HWCAP_FAST_MULT|HWCAP_EDSP
-	.long	cpu_8033x_name
-	.long	xscale_processor_functions
-	.long	v4wbi_tlb_fns
-	.long	xscale_mc_user_fns
-	.long	xscale_cache_fns
-	.size	__8033x_proc_info, . - __8033x_proc_info
-
-	.type	__pxa250_proc_info,#object
-__pxa250_proc_info:
-	.long	0x69052100
-	.long	0xfffff7f0
-	.long   PMD_TYPE_SECT | \
-		PMD_SECT_BUFFERABLE | \
-		PMD_SECT_CACHEABLE | \
-		PMD_SECT_AP_WRITE | \
-		PMD_SECT_AP_READ
-	.long   PMD_TYPE_SECT | \
-		PMD_SECT_AP_WRITE | \
-		PMD_SECT_AP_READ
-	b	__xscale_setup
-	.long	cpu_arch_name
-	.long	cpu_elf_name
-	.long	HWCAP_SWP|HWCAP_HALF|HWCAP_THUMB|HWCAP_FAST_MULT|HWCAP_EDSP
-	.long	cpu_pxa250_name
-	.long	xscale_processor_functions
-	.long	v4wbi_tlb_fns
-	.long	xscale_mc_user_fns
-	.long	xscale_cache_fns
-	.size	__pxa250_proc_info, . - __pxa250_proc_info
-
-	.type	__pxa210_proc_info,#object
-__pxa210_proc_info:
-	.long	0x69052120
-	.long	0xfffff3f0
-	.long   PMD_TYPE_SECT | \
-		PMD_SECT_BUFFERABLE | \
-		PMD_SECT_CACHEABLE | \
-		PMD_SECT_AP_WRITE | \
-		PMD_SECT_AP_READ
-	.long   PMD_TYPE_SECT | \
-		PMD_SECT_AP_WRITE | \
-		PMD_SECT_AP_READ
-	b	__xscale_setup
-	.long	cpu_arch_name
-	.long	cpu_elf_name
-	.long	HWCAP_SWP|HWCAP_HALF|HWCAP_THUMB|HWCAP_FAST_MULT|HWCAP_EDSP
-	.long	cpu_pxa210_name
-	.long	xscale_processor_functions
-	.long	v4wbi_tlb_fns
-	.long	xscale_mc_user_fns
-	.long	xscale_cache_fns
-	.size	__pxa210_proc_info, . - __pxa210_proc_info
-
-	.type	__ixp2400_proc_info, #object
-__ixp2400_proc_info:
-	.long   0x69054190
-	.long   0xfffffff0
-	.long   PMD_TYPE_SECT | \
-		PMD_SECT_BUFFERABLE | \
-		PMD_SECT_CACHEABLE | \
-		PMD_SECT_AP_WRITE | \
-		PMD_SECT_AP_READ
-	.long   PMD_TYPE_SECT | \
-		PMD_SECT_AP_WRITE | \
-		PMD_SECT_AP_READ
-	b       __xscale_setup
-	.long   cpu_arch_name
-	.long   cpu_elf_name
-	.long   HWCAP_SWP|HWCAP_HALF|HWCAP_THUMB|HWCAP_FAST_MULT|HWCAP_EDSP
-	.long   cpu_ixp2400_name
-	.long   xscale_processor_functions
-	.long	v4wbi_tlb_fns
-	.long	xscale_mc_user_fns
-	.long	xscale_cache_fns
-	.size   __ixp2400_proc_info, . - __ixp2400_proc_info                
-
-	.type	__ixp2800_proc_info, #object
-__ixp2800_proc_info:
-	.long   0x690541a0
-	.long   0xfffffff0
-	.long   PMD_TYPE_SECT | \
-		PMD_SECT_BUFFERABLE | \
-		PMD_SECT_CACHEABLE | \
-		PMD_SECT_AP_WRITE | \
-		PMD_SECT_AP_READ
-	.long   PMD_TYPE_SECT | \
-		PMD_SECT_AP_WRITE | \
-		PMD_SECT_AP_READ
-	b       __xscale_setup
-	.long   cpu_arch_name
-	.long   cpu_elf_name
-	.long   HWCAP_SWP|HWCAP_HALF|HWCAP_THUMB|HWCAP_FAST_MULT|HWCAP_EDSP
-	.long   cpu_ixp2800_name
-	.long   xscale_processor_functions
-	.long	v4wbi_tlb_fns
-	.long	xscale_mc_user_fns
-	.long	xscale_cache_fns
-	.size   __ixp2800_proc_info, . - __ixp2800_proc_info                
-
-	.type	__ixp42x_proc_info, #object
-__ixp42x_proc_info:
-	.long   0x690541c0
-	.long   0xffffffc0
-	.long   PMD_TYPE_SECT | \
-		PMD_SECT_BUFFERABLE | \
-		PMD_SECT_CACHEABLE | \
-		PMD_SECT_AP_WRITE | \
-		PMD_SECT_AP_READ
-	.long   PMD_TYPE_SECT | \
-		PMD_SECT_AP_WRITE | \
-		PMD_SECT_AP_READ
-	b       __xscale_setup
-	.long   cpu_arch_name
-	.long   cpu_elf_name
-	.long   HWCAP_SWP|HWCAP_HALF|HWCAP_THUMB|HWCAP_FAST_MULT|HWCAP_EDSP
-	.long   cpu_ixp42x_name
-	.long   xscale_processor_functions
-	.long	v4wbi_tlb_fns
-	.long	xscale_mc_user_fns
-	.long	xscale_cache_fns
-	.size   __ixp42x_proc_info, . - __ixp42x_proc_info                
-
-	.type   __ixp43x_proc_info, #object
-__ixp43x_proc_info:
-	.long   0x69054040
-	.long   0xfffffff0
-	.long   PMD_TYPE_SECT | \
-		PMD_SECT_BUFFERABLE | \
-		PMD_SECT_CACHEABLE | \
-		PMD_SECT_AP_WRITE | \
-		PMD_SECT_AP_READ
-	.long   PMD_TYPE_SECT | \
-		PMD_SECT_AP_WRITE | \
-		PMD_SECT_AP_READ
-	b       __xscale_setup
-	.long   cpu_arch_name
-	.long   cpu_elf_name
-	.long   HWCAP_SWP|HWCAP_HALF|HWCAP_THUMB|HWCAP_FAST_MULT|HWCAP_EDSP
-	.long   cpu_ixp43x_name
-	.long   xscale_processor_functions
-	.long   v4wbi_tlb_fns
-	.long   xscale_mc_user_fns
-	.long   xscale_cache_fns
-	.size   __ixp43x_proc_info, . - __ixp43x_proc_info
-
-	.type	__ixp46x_proc_info, #object
-__ixp46x_proc_info:
-	.long   0x69054200
-	.long   0xffffff00
-	.long   PMD_TYPE_SECT | \
-		PMD_SECT_BUFFERABLE | \
-		PMD_SECT_CACHEABLE | \
-		PMD_SECT_AP_WRITE | \
-		PMD_SECT_AP_READ
-	.long   PMD_TYPE_SECT | \
-		PMD_SECT_AP_WRITE | \
-		PMD_SECT_AP_READ
-	b       __xscale_setup
-	.long   cpu_arch_name
-	.long   cpu_elf_name
-	.long   HWCAP_SWP|HWCAP_HALF|HWCAP_THUMB|HWCAP_FAST_MULT|HWCAP_EDSP
-	.long   cpu_ixp46x_name
-	.long   xscale_processor_functions
-	.long	v4wbi_tlb_fns
-	.long	xscale_mc_user_fns
-	.long	xscale_cache_fns
-	.size   __ixp46x_proc_info, . - __ixp46x_proc_info
-
-	.type	__pxa255_proc_info,#object
-__pxa255_proc_info:
-	.long	0x69052d00
-	.long	0xfffffff0
-	.long   PMD_TYPE_SECT | \
-		PMD_SECT_BUFFERABLE | \
-		PMD_SECT_CACHEABLE | \
-		PMD_SECT_AP_WRITE | \
-		PMD_SECT_AP_READ
-	.long   PMD_TYPE_SECT | \
-		PMD_SECT_AP_WRITE | \
-		PMD_SECT_AP_READ
-	b	__xscale_setup
-	.long	cpu_arch_name
-	.long	cpu_elf_name
-	.long	HWCAP_SWP|HWCAP_HALF|HWCAP_THUMB|HWCAP_FAST_MULT|HWCAP_EDSP
-	.long	cpu_pxa255_name
-	.long	xscale_processor_functions
-	.long	v4wbi_tlb_fns
-	.long	xscale_mc_user_fns
-	.long	xscale_cache_fns
-	.size	__pxa255_proc_info, . - __pxa255_proc_info
-
-	.type	__pxa270_proc_info,#object
-__pxa270_proc_info:
-	.long	0x69054110
-	.long	0xfffffff0
-	.long   PMD_TYPE_SECT | \
-		PMD_SECT_BUFFERABLE | \
-		PMD_SECT_CACHEABLE | \
-		PMD_SECT_AP_WRITE | \
-		PMD_SECT_AP_READ
-	.long   PMD_TYPE_SECT | \
-		PMD_SECT_AP_WRITE | \
-		PMD_SECT_AP_READ
-	b	__xscale_setup
-	.long	cpu_arch_name
-	.long	cpu_elf_name
-	.long	HWCAP_SWP|HWCAP_HALF|HWCAP_THUMB|HWCAP_FAST_MULT|HWCAP_EDSP
-	.long	cpu_pxa270_name
-	.long	xscale_processor_functions
-	.long	v4wbi_tlb_fns
-	.long	xscale_mc_user_fns
-	.long	xscale_cache_fns
-	.size	__pxa270_proc_info, . - __pxa270_proc_info
 
+	xscale_proc_info 80200_A0_A1, 0x69052000, 0xfffffffe, cpu_80200_name, \
+		cache=xscale_80200_A0_A1_cache_fns
+	xscale_proc_info 80200, 0x69052000, 0xfffffff0, cpu_80200_name
+	xscale_proc_info 80219, 0x69052e20, 0xffffffe0, cpu_80219_name
+	xscale_proc_info 8032x, 0x69052420, 0xfffff7e0, cpu_8032x_name
+	xscale_proc_info 8033x, 0x69054010, 0xfffffd30, cpu_8033x_name
+	xscale_proc_info pxa250, 0x69052100, 0xfffff7f0, cpu_pxa250_name
+	xscale_proc_info pxa210, 0x69052120, 0xfffff3f0, cpu_pxa210_name
+	xscale_proc_info ixp2400, 0x69054190, 0xfffffff0, cpu_ixp2400_name
+	xscale_proc_info ixp2800, 0x690541a0, 0xfffffff0, cpu_ixp2800_name
+	xscale_proc_info ixp42x, 0x690541c0, 0xffffffc0, cpu_ixp42x_name
+	xscale_proc_info ixp43x, 0x69054040, 0xfffffff0, cpu_ixp43x_name
+	xscale_proc_info ixp46x, 0x69054200, 0xffffff00, cpu_ixp46x_name
+	xscale_proc_info pxa255, 0x69052d00, 0xfffffff0, cpu_pxa255_name
+	xscale_proc_info pxa270, 0x69054110, 0xfffffff0, cpu_pxa270_name

arch/arm/mm/tlb-fa.S

@@ -65,9 +65,5 @@ ENTRY(fa_flush_kern_tlb_range)
 
 	__INITDATA
 
-	.type	fa_tlb_fns, #object
+	/* define struct cpu_tlb_fns (see <asm/tlbflush.h> and proc-macros.S) */
-ENTRY(fa_tlb_fns)
+	define_tlb_functions fa, fa_tlb_flags
-	.long	fa_flush_user_tlb_range
-	.long	fa_flush_kern_tlb_range
-	.long	fa_tlb_flags
-	.size	fa_tlb_fns, . - fa_tlb_fns