Merge branch 'tracing/for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/linux-2.6-tip

* 'tracing/for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/linux-2.6-tip: (228 commits)
  ftrace: build fix for ftraced_suspend
  ftrace: separate out the function enabled variable
  ftrace: add ftrace_kill_atomic
  ftrace: use current CPU for function startup
  ftrace: start wakeup tracing after setting function tracer
  ftrace: check proper config for preempt type
  ftrace: trace schedule
  ftrace: define function trace nop
  ftrace: move sched_switch enable after markers
  ftrace: prevent ftrace modifications while being kprobe'd, v2
  fix "ftrace: store mcount address in rec->ip"
  mmiotrace broken in linux-next (8-bit writes only)
  ftrace: avoid modifying kprobe'd records
  ftrace: freeze kprobe'd records
  kprobes: enable clean usage of get_kprobe
  ftrace: store mcount address in rec->ip
  ftrace: build fix with gcc 4.3
  namespacecheck: fixes
  ftrace: fix "notrace" filtering priority
  ftrace: fix printout
  ...

Documentation/tracers/mmiotrace.txt

@@ -0,0 +1,164 @@
+		In-kernel memory-mapped I/O tracing
+
+
+Home page and links to optional user space tools:
+
+	http://nouveau.freedesktop.org/wiki/MmioTrace
+
+MMIO tracing was originally developed by Intel around 2003 for their Fault
+Injection Test Harness. In Dec 2006 - Jan 2007, using the code from Intel,
+Jeff Muizelaar created a tool for tracing MMIO accesses with the Nouveau
+project in mind. Since then many people have contributed.
+
+Mmiotrace was built for reverse engineering any memory-mapped IO device with
+the Nouveau project as the first real user. Only x86 and x86_64 architectures
+are supported.
+
+Out-of-tree mmiotrace was originally modified for mainline inclusion and
+ftrace framework by Pekka Paalanen <pq@iki.fi>.
+
+
+Preparation
+-----------
+
+Mmiotrace feature is compiled in by the CONFIG_MMIOTRACE option. Tracing is
+disabled by default, so it is safe to have this set to yes. SMP systems are
+supported, but tracing is unreliable and may miss events if more than one CPU
+is on-line, therefore mmiotrace takes all but one CPU off-line during run-time
+activation. You can re-enable CPUs by hand, but you have been warned, there
+is no way to automatically detect if you are losing events due to CPUs racing.
+
+
+Usage Quick Reference
+---------------------
+
+$ mount -t debugfs debugfs /debug
+$ echo mmiotrace > /debug/tracing/current_tracer
+$ cat /debug/tracing/trace_pipe > mydump.txt &
+Start X or whatever.
+$ echo "X is up" > /debug/tracing/marker
+$ echo none > /debug/tracing/current_tracer
+Check for lost events.
+
+
+Usage
+-----
+
+Make sure debugfs is mounted to /debug. If not, (requires root privileges)
+$ mount -t debugfs debugfs /debug
+
+Check that the driver you are about to trace is not loaded.
+
+Activate mmiotrace (requires root privileges):
+$ echo mmiotrace > /debug/tracing/current_tracer
+
+Start storing the trace:
+$ cat /debug/tracing/trace_pipe > mydump.txt &
+The 'cat' process should stay running (sleeping) in the background.
+
+Load the driver you want to trace and use it. Mmiotrace will only catch MMIO
+accesses to areas that are ioremapped while mmiotrace is active.
+
+[Unimplemented feature:]
+During tracing you can place comments (markers) into the trace by
+$ echo "X is up" > /debug/tracing/marker
+This makes it easier to see which part of the (huge) trace corresponds to
+which action. It is recommended to place descriptive markers about what you
+do.
+
+Shut down mmiotrace (requires root privileges):
+$ echo none > /debug/tracing/current_tracer
+The 'cat' process exits. If it does not, kill it by issuing 'fg' command and
+pressing ctrl+c.
+
+Check that mmiotrace did not lose events due to a buffer filling up. Either
+$ grep -i lost mydump.txt
+which tells you exactly how many events were lost, or use
+$ dmesg
+to view your kernel log and look for "mmiotrace has lost events" warning. If
+events were lost, the trace is incomplete. You should enlarge the buffers and
+try again. Buffers are enlarged by first seeing how large the current buffers
+are:
+$ cat /debug/tracing/trace_entries
+gives you a number. Approximately double this number and write it back, for
+instance:
+$ echo 128000 > /debug/tracing/trace_entries
+Then start again from the top.
+
+If you are doing a trace for a driver project, e.g. Nouveau, you should also
+do the following before sending your results:
+$ lspci -vvv > lspci.txt
+$ dmesg > dmesg.txt
+$ tar zcf pciid-nick-mmiotrace.tar.gz mydump.txt lspci.txt dmesg.txt
+and then send the .tar.gz file. The trace compresses considerably. Replace
+"pciid" and "nick" with the PCI ID or model name of your piece of hardware
+under investigation and your nick name.
+
+
+How Mmiotrace Works
+-------------------
+
+Access to hardware IO-memory is gained by mapping addresses from PCI bus by
+calling one of the ioremap_*() functions. Mmiotrace is hooked into the
+__ioremap() function and gets called whenever a mapping is created. Mapping is
+an event that is recorded into the trace log. Note, that ISA range mappings
+are not caught, since the mapping always exists and is returned directly.
+
+MMIO accesses are recorded via page faults. Just before __ioremap() returns,
+the mapped pages are marked as not present. Any access to the pages causes a
+fault. The page fault handler calls mmiotrace to handle the fault. Mmiotrace
+marks the page present, sets TF flag to achieve single stepping and exits the
+fault handler. The instruction that faulted is executed and debug trap is
+entered. Here mmiotrace again marks the page as not present. The instruction
+is decoded to get the type of operation (read/write), data width and the value
+read or written. These are stored to the trace log.
+
+Setting the page present in the page fault handler has a race condition on SMP
+machines. During the single stepping other CPUs may run freely on that page
+and events can be missed without a notice. Re-enabling other CPUs during
+tracing is discouraged.
+
+
+Trace Log Format
+----------------
+
+The raw log is text and easily filtered with e.g. grep and awk. One record is
+one line in the log. A record starts with a keyword, followed by keyword
+dependant arguments. Arguments are separated by a space, or continue until the
+end of line. The format for version 20070824 is as follows:
+
+Explanation	Keyword	Space separated arguments
+---------------------------------------------------------------------------
+
+read event	R	width, timestamp, map id, physical, value, PC, PID
+write event	W	width, timestamp, map id, physical, value, PC, PID
+ioremap event	MAP	timestamp, map id, physical, virtual, length, PC, PID
+iounmap event	UNMAP	timestamp, map id, PC, PID
+marker		MARK	timestamp, text
+version		VERSION	the string "20070824"
+info for reader	LSPCI	one line from lspci -v
+PCI address map	PCIDEV	space separated /proc/bus/pci/devices data
+unk. opcode	UNKNOWN	timestamp, map id, physical, data, PC, PID
+
+Timestamp is in seconds with decimals. Physical is a PCI bus address, virtual
+is a kernel virtual address. Width is the data width in bytes and value is the
+data value. Map id is an arbitrary id number identifying the mapping that was
+used in an operation. PC is the program counter and PID is process id. PC is
+zero if it is not recorded. PID is always zero as tracing MMIO accesses
+originating in user space memory is not yet supported.
+
+For instance, the following awk filter will pass all 32-bit writes that target
+physical addresses in the range [0xfb73ce40, 0xfb800000[
+
+$ awk '/W 4 / { adr=strtonum($5); if (adr >= 0xfb73ce40 &&
+adr < 0xfb800000) print; }'
+
+
+Tools for Developers
+--------------------
+
+The user space tools include utilities for:
+- replacing numeric addresses and values with hardware register names
+- replaying MMIO logs, i.e., re-executing the recorded writes
+
+

Makefile

@@ -528,6 +528,10 @@ KBUILD_CFLAGS	+= -g
 KBUILD_AFLAGS	+= -gdwarf-2
 endif
 
+ifdef CONFIG_FTRACE
+KBUILD_CFLAGS	+= -pg
+endif
+
 # We trigger additional mismatches with less inlining
 ifdef CONFIG_DEBUG_SECTION_MISMATCH
 KBUILD_CFLAGS += $(call cc-option, -fno-inline-functions-called-once)

arch/arm/Kconfig

@@ -14,6 +14,8 @@ config ARM
 	select HAVE_OPROFILE
 	select HAVE_KPROBES if (!XIP_KERNEL)
 	select HAVE_KRETPROBES if (HAVE_KPROBES)
+	select HAVE_FTRACE if (!XIP_KERNEL)
+	select HAVE_DYNAMIC_FTRACE if (HAVE_FTRACE)
 	help
 	  The ARM series is a line of low-power-consumption RISC chip designs
 	  licensed by ARM Ltd and targeted at embedded applications and

arch/arm/boot/compressed/Makefile

@@ -69,6 +69,12 @@ SEDFLAGS	= s/TEXT_START/$(ZTEXTADDR)/;s/BSS_START/$(ZBSSADDR)/
 
 targets       := vmlinux vmlinux.lds piggy.gz piggy.o font.o font.c \
 		 head.o misc.o $(OBJS)
+
+ifeq ($(CONFIG_FTRACE),y)
+ORIG_CFLAGS := $(KBUILD_CFLAGS)
+KBUILD_CFLAGS = $(subst -pg, , $(ORIG_CFLAGS))
+endif
+
 EXTRA_CFLAGS  := -fpic -fno-builtin
 EXTRA_AFLAGS  :=
 

arch/arm/kernel/Makefile

@@ -4,6 +4,10 @@
 
 AFLAGS_head.o := -DTEXT_OFFSET=$(TEXT_OFFSET)
 
+ifdef CONFIG_DYNAMIC_FTRACE
+CFLAGS_REMOVE_ftrace.o = -pg
+endif
+
 # Object file lists.
 
 obj-y		:= compat.o entry-armv.o entry-common.o irq.o \
@@ -18,6 +22,7 @@ obj-$(CONFIG_ARTHUR)		+= arthur.o
 obj-$(CONFIG_ISA_DMA)		+= dma-isa.o
 obj-$(CONFIG_PCI)		+= bios32.o isa.o
 obj-$(CONFIG_SMP)		+= smp.o
+obj-$(CONFIG_DYNAMIC_FTRACE)	+= ftrace.o
 obj-$(CONFIG_KEXEC)		+= machine_kexec.o relocate_kernel.o
 obj-$(CONFIG_KPROBES)		+= kprobes.o kprobes-decode.o
 obj-$(CONFIG_ATAGS_PROC)	+= atags.o

arch/arm/kernel/armksyms.c

@@ -18,6 +18,7 @@
 #include <asm/io.h>
 #include <asm/system.h>
 #include <asm/uaccess.h>
+#include <asm/ftrace.h>
 
 /*
  * libgcc functions - functions that are used internally by the
@@ -181,3 +182,7 @@ EXPORT_SYMBOL(_find_next_bit_be);
 #endif
 
 EXPORT_SYMBOL(copy_page);
+
+#ifdef CONFIG_FTRACE
+EXPORT_SYMBOL(mcount);
+#endif

arch/arm/kernel/entry-common.S

@@ -9,6 +9,7 @@
  */
 
 #include <asm/unistd.h>
+#include <asm/ftrace.h>
 #include <asm/arch/entry-macro.S>
 
 #include "entry-header.S"
@@ -99,6 +100,56 @@ ENTRY(ret_from_fork)
 #undef CALL
 #define CALL(x) .long x
 
+#ifdef CONFIG_FTRACE
+#ifdef CONFIG_DYNAMIC_FTRACE
+ENTRY(mcount)
+	stmdb sp!, {r0-r3, lr}
+	mov r0, lr
+	sub r0, r0, #MCOUNT_INSN_SIZE
+
+	.globl mcount_call
+mcount_call:
+	bl ftrace_stub
+	ldmia sp!, {r0-r3, pc}
+
+ENTRY(ftrace_caller)
+	stmdb sp!, {r0-r3, lr}
+	ldr r1, [fp, #-4]
+	mov r0, lr
+	sub r0, r0, #MCOUNT_INSN_SIZE
+
+	.globl ftrace_call
+ftrace_call:
+	bl ftrace_stub
+	ldmia sp!, {r0-r3, pc}
+
+#else
+
+ENTRY(mcount)
+	stmdb sp!, {r0-r3, lr}
+	ldr r0, =ftrace_trace_function
+	ldr r2, [r0]
+	adr r0, ftrace_stub
+	cmp r0, r2
+	bne trace
+	ldmia sp!, {r0-r3, pc}
+
+trace:
+	ldr r1, [fp, #-4]
+	mov r0, lr
+	sub r0, r0, #MCOUNT_INSN_SIZE
+	mov lr, pc
+	mov pc, r2
+	ldmia sp!, {r0-r3, pc}
+
+#endif /* CONFIG_DYNAMIC_FTRACE */
+
+	.globl ftrace_stub
+ftrace_stub:
+	mov pc, lr
+
+#endif /* CONFIG_FTRACE */
+
 /*=============================================================================
  * SWI handler
  *-----------------------------------------------------------------------------

arch/arm/kernel/ftrace.c

@@ -0,0 +1,116 @@
+/*
+ * Dynamic function tracing support.
+ *
+ * Copyright (C) 2008 Abhishek Sagar <sagar.abhishek@gmail.com>
+ *
+ * For licencing details, see COPYING.
+ *
+ * Defines low-level handling of mcount calls when the kernel
+ * is compiled with the -pg flag. When using dynamic ftrace, the
+ * mcount call-sites get patched lazily with NOP till they are
+ * enabled. All code mutation routines here take effect atomically.
+ */
+
+#include <linux/ftrace.h>
+
+#include <asm/cacheflush.h>
+#include <asm/ftrace.h>
+
+#define PC_OFFSET      8
+#define BL_OPCODE      0xeb000000
+#define BL_OFFSET_MASK 0x00ffffff
+
+static unsigned long bl_insn;
+static const unsigned long NOP = 0xe1a00000; /* mov r0, r0 */
+
+unsigned char *ftrace_nop_replace(void)
+{
+	return (char *)&NOP;
+}
+
+/* construct a branch (BL) instruction to addr */
+unsigned char *ftrace_call_replace(unsigned long pc, unsigned long addr)
+{
+	long offset;
+
+	offset = (long)addr - (long)(pc + PC_OFFSET);
+	if (unlikely(offset < -33554432 || offset > 33554428)) {
+		/* Can't generate branches that far (from ARM ARM). Ftrace
+		 * doesn't generate branches outside of kernel text.
+		 */
+		WARN_ON_ONCE(1);
+		return NULL;
+	}
+	offset = (offset >> 2) & BL_OFFSET_MASK;
+	bl_insn = BL_OPCODE | offset;
+	return (unsigned char *)&bl_insn;
+}
+
+int ftrace_modify_code(unsigned long pc, unsigned char *old_code,
+		       unsigned char *new_code)
+{
+	unsigned long err = 0, replaced = 0, old, new;
+
+	old = *(unsigned long *)old_code;
+	new = *(unsigned long *)new_code;
+
+	__asm__ __volatile__ (
+		"1:  ldr    %1, [%2]  \n"
+		"    cmp    %1, %4    \n"
+		"2:  streq  %3, [%2]  \n"
+		"    cmpne  %1, %3    \n"
+		"    movne  %0, #2    \n"
+		"3:\n"
+
+		".section .fixup, \"ax\"\n"
+		"4:  mov  %0, #1  \n"
+		"    b    3b      \n"
+		".previous\n"
+
+		".section __ex_table, \"a\"\n"
+		"    .long 1b, 4b \n"
+		"    .long 2b, 4b \n"
+		".previous\n"
+
+		: "=r"(err), "=r"(replaced)
+		: "r"(pc), "r"(new), "r"(old), "0"(err), "1"(replaced)
+		: "memory");
+
+	if (!err && (replaced == old))
+		flush_icache_range(pc, pc + MCOUNT_INSN_SIZE);
+
+	return err;
+}
+
+int ftrace_update_ftrace_func(ftrace_func_t func)
+{
+	int ret;
+	unsigned long pc, old;
+	unsigned char *new;
+
+	pc = (unsigned long)&ftrace_call;
+	memcpy(&old, &ftrace_call, MCOUNT_INSN_SIZE);
+	new = ftrace_call_replace(pc, (unsigned long)func);
+	ret = ftrace_modify_code(pc, (unsigned char *)&old, new);
+	return ret;
+}
+
+int ftrace_mcount_set(unsigned long *data)
+{
+	unsigned long pc, old;
+	unsigned long *addr = data;
+	unsigned char *new;
+
+	pc = (unsigned long)&mcount_call;
+	memcpy(&old, &mcount_call, MCOUNT_INSN_SIZE);
+	new = ftrace_call_replace(pc, *addr);
+	*addr = ftrace_modify_code(pc, (unsigned char *)&old, new);
+	return 0;
+}
+
+/* run from kstop_machine */
+int __init ftrace_dyn_arch_init(void *data)
+{
+	ftrace_mcount_set(data);
+	return 0;
+}

arch/arm/kernel/kprobes.c

@@ -274,7 +274,7 @@ int __kprobes kprobe_exceptions_notify(struct notifier_block *self,
  * for kretprobe handlers which should normally be interested in r0 only
  * anyway.
  */
-static void __attribute__((naked)) __kprobes kretprobe_trampoline(void)
+void __naked __kprobes kretprobe_trampoline(void)
 {
 	__asm__ __volatile__ (
 		"stmdb	sp!, {r0 - r11}		\n\t"

arch/powerpc/Kconfig

@@ -105,11 +105,13 @@ config ARCH_NO_VIRT_TO_BUS
 config PPC
 	bool
 	default y
+	select HAVE_DYNAMIC_FTRACE
+	select HAVE_FTRACE
 	select HAVE_IDE
-	select HAVE_OPROFILE
 	select HAVE_KPROBES
 	select HAVE_KRETPROBES
 	select HAVE_LMB
+	select HAVE_OPROFILE
 
 config EARLY_PRINTK
 	bool

arch/powerpc/kernel/Makefile

@@ -12,6 +12,18 @@ CFLAGS_prom_init.o      += -fPIC
 CFLAGS_btext.o		+= -fPIC
 endif
 
+ifdef CONFIG_FTRACE
+# Do not trace early boot code
+CFLAGS_REMOVE_cputable.o = -pg
+CFLAGS_REMOVE_prom_init.o = -pg
+
+ifdef CONFIG_DYNAMIC_FTRACE
+# dynamic ftrace setup.
+CFLAGS_REMOVE_ftrace.o = -pg
+endif
+
+endif
+
 obj-y				:= cputable.o ptrace.o syscalls.o \
 				   irq.o align.o signal_32.o pmc.o vdso.o \
 				   init_task.o process.o systbl.o idle.o \
@@ -78,6 +90,8 @@ obj-$(CONFIG_KEXEC)		+= machine_kexec.o crash.o \
 obj-$(CONFIG_AUDIT)		+= audit.o
 obj64-$(CONFIG_AUDIT)		+= compat_audit.o
 
+obj-$(CONFIG_DYNAMIC_FTRACE)	+= ftrace.o
+
 obj-$(CONFIG_8XX_MINIMAL_FPEMU) += softemu8xx.o
 
 ifneq ($(CONFIG_PPC_INDIRECT_IO),y)

arch/powerpc/kernel/entry_32.S

@@ -30,6 +30,7 @@
 #include <asm/ppc_asm.h>
 #include <asm/asm-offsets.h>
 #include <asm/unistd.h>
+#include <asm/ftrace.h>
 
 #undef SHOW_SYSCALLS
 #undef SHOW_SYSCALLS_TASK
@@ -1035,3 +1036,129 @@ machine_check_in_rtas:
 	/* XXX load up BATs and panic */
 
 #endif /* CONFIG_PPC_RTAS */
+
+#ifdef CONFIG_FTRACE
+#ifdef CONFIG_DYNAMIC_FTRACE
+_GLOBAL(mcount)
+_GLOBAL(_mcount)
+	stwu	r1,-48(r1)
+	stw	r3, 12(r1)
+	stw	r4, 16(r1)
+	stw	r5, 20(r1)
+	stw	r6, 24(r1)
+	mflr	r3
+	stw	r7, 28(r1)
+	mfcr	r5
+	stw	r8, 32(r1)
+	stw	r9, 36(r1)
+	stw	r10,40(r1)
+	stw	r3, 44(r1)
+	stw	r5, 8(r1)
+	subi	r3, r3, MCOUNT_INSN_SIZE
+	.globl mcount_call
+mcount_call:
+	bl	ftrace_stub
+	nop
+	lwz	r6, 8(r1)
+	lwz	r0, 44(r1)
+	lwz	r3, 12(r1)
+	mtctr	r0
+	lwz	r4, 16(r1)
+	mtcr	r6
+	lwz	r5, 20(r1)
+	lwz	r6, 24(r1)
+	lwz	r0, 52(r1)
+	lwz	r7, 28(r1)
+	lwz	r8, 32(r1)
+	mtlr	r0
+	lwz	r9, 36(r1)
+	lwz	r10,40(r1)
+	addi	r1, r1, 48
+	bctr
+
+_GLOBAL(ftrace_caller)
+	/* Based off of objdump optput from glibc */
+	stwu	r1,-48(r1)
+	stw	r3, 12(r1)
+	stw	r4, 16(r1)
+	stw	r5, 20(r1)
+	stw	r6, 24(r1)
+	mflr	r3
+	lwz	r4, 52(r1)
+	mfcr	r5
+	stw	r7, 28(r1)
+	stw	r8, 32(r1)
+	stw	r9, 36(r1)
+	stw	r10,40(r1)
+	stw	r3, 44(r1)
+	stw	r5, 8(r1)
+	subi	r3, r3, MCOUNT_INSN_SIZE
+.globl ftrace_call
+ftrace_call:
+	bl	ftrace_stub
+	nop
+	lwz	r6, 8(r1)
+	lwz	r0, 44(r1)
+	lwz	r3, 12(r1)
+	mtctr	r0
+	lwz	r4, 16(r1)
+	mtcr	r6
+	lwz	r5, 20(r1)
+	lwz	r6, 24(r1)
+	lwz	r0, 52(r1)
+	lwz	r7, 28(r1)
+	lwz	r8, 32(r1)
+	mtlr	r0
+	lwz	r9, 36(r1)
+	lwz	r10,40(r1)
+	addi	r1, r1, 48
+	bctr
+#else
+_GLOBAL(mcount)
+_GLOBAL(_mcount)
+	stwu	r1,-48(r1)
+	stw	r3, 12(r1)
+	stw	r4, 16(r1)
+	stw	r5, 20(r1)
+	stw	r6, 24(r1)
+	mflr	r3
+	lwz	r4, 52(r1)
+	mfcr	r5
+	stw	r7, 28(r1)
+	stw	r8, 32(r1)
+	stw	r9, 36(r1)
+	stw	r10,40(r1)
+	stw	r3, 44(r1)
+	stw	r5, 8(r1)
+
+	subi	r3, r3, MCOUNT_INSN_SIZE
+	LOAD_REG_ADDR(r5, ftrace_trace_function)
+	lwz	r5,0(r5)
+
+	mtctr	r5
+	bctrl
+
+	nop
+
+	lwz	r6, 8(r1)
+	lwz	r0, 44(r1)
+	lwz	r3, 12(r1)
+	mtctr	r0
+	lwz	r4, 16(r1)
+	mtcr	r6
+	lwz	r5, 20(r1)
+	lwz	r6, 24(r1)
+	lwz	r0, 52(r1)
+	lwz	r7, 28(r1)
+	lwz	r8, 32(r1)
+	mtlr	r0
+	lwz	r9, 36(r1)
+	lwz	r10,40(r1)
+	addi	r1, r1, 48
+	bctr
+#endif
+
+_GLOBAL(ftrace_stub)
+	blr
+
+#endif /* CONFIG_MCOUNT */

arch/powerpc/kernel/entry_64.S

@@ -31,6 +31,7 @@
 #include <asm/bug.h>
 #include <asm/ptrace.h>
 #include <asm/irqflags.h>
+#include <asm/ftrace.h>
 
 /*
  * System calls.
@@ -870,3 +871,67 @@ _GLOBAL(enter_prom)
 	ld	r0,16(r1)
 	mtlr    r0
         blr
+
+#ifdef CONFIG_FTRACE
+#ifdef CONFIG_DYNAMIC_FTRACE
+_GLOBAL(mcount)
+_GLOBAL(_mcount)
+	/* Taken from output of objdump from lib64/glibc */
+	mflr	r3
+	stdu	r1, -112(r1)
+	std	r3, 128(r1)
+	subi	r3, r3, MCOUNT_INSN_SIZE
+	.globl mcount_call
+mcount_call:
+	bl	ftrace_stub
+	nop
+	ld	r0, 128(r1)
+	mtlr	r0
+	addi	r1, r1, 112
+	blr
+
+_GLOBAL(ftrace_caller)
+	/* Taken from output of objdump from lib64/glibc */
+	mflr	r3
+	ld	r11, 0(r1)
+	stdu	r1, -112(r1)
+	std	r3, 128(r1)
+	ld	r4, 16(r11)
+	subi	r3, r3, MCOUNT_INSN_SIZE
+.globl ftrace_call
+ftrace_call:
+	bl	ftrace_stub
+	nop
+	ld	r0, 128(r1)
+	mtlr	r0
+	addi	r1, r1, 112
+_GLOBAL(ftrace_stub)
+	blr
+#else
+_GLOBAL(mcount)
+	blr
+
+_GLOBAL(_mcount)
+	/* Taken from output of objdump from lib64/glibc */
+	mflr	r3
+	ld	r11, 0(r1)
+	stdu	r1, -112(r1)
+	std	r3, 128(r1)
+	ld	r4, 16(r11)
+
+	subi	r3, r3, MCOUNT_INSN_SIZE
+	LOAD_REG_ADDR(r5,ftrace_trace_function)
+	ld	r5,0(r5)
+	ld	r5,0(r5)
+	mtctr	r5
+	bctrl
+
+	nop
+	ld	r0, 128(r1)
+	mtlr	r0
+	addi	r1, r1, 112
+_GLOBAL(ftrace_stub)
+	blr
+
+#endif
+#endif

arch/powerpc/kernel/ftrace.c

@@ -0,0 +1,154 @@
+/*
+ * Code for replacing ftrace calls with jumps.
+ *
+ * Copyright (C) 2007-2008 Steven Rostedt <srostedt@redhat.com>
+ *
+ * Thanks goes out to P.A. Semi, Inc for supplying me with a PPC64 box.
+ *
+ */
+
+#include <linux/spinlock.h>
+#include <linux/hardirq.h>
+#include <linux/ftrace.h>
+#include <linux/percpu.h>
+#include <linux/init.h>
+#include <linux/list.h>
+
+#include <asm/cacheflush.h>
+#include <asm/ftrace.h>
+
+
+static unsigned int ftrace_nop = 0x60000000;
+
+#ifdef CONFIG_PPC32
+# define GET_ADDR(addr) addr
+#else
+/* PowerPC64's functions are data that points to the functions */
+# define GET_ADDR(addr) *(unsigned long *)addr
+#endif
+
+
+static unsigned int notrace ftrace_calc_offset(long ip, long addr)
+{
+	return (int)(addr - ip);
+}
+
+notrace unsigned char *ftrace_nop_replace(void)
+{
+	return (char *)&ftrace_nop;
+}
+
+notrace unsigned char *ftrace_call_replace(unsigned long ip, unsigned long addr)
+{
+	static unsigned int op;
+
+	/*
+	 * It would be nice to just use create_function_call, but that will
+	 * update the code itself. Here we need to just return the
+	 * instruction that is going to be modified, without modifying the
+	 * code.
+	 */
+	addr = GET_ADDR(addr);
+
+	/* Set to "bl addr" */
+	op = 0x48000001 | (ftrace_calc_offset(ip, addr) & 0x03fffffc);
+
+	/*
+	 * No locking needed, this must be called via kstop_machine
+	 * which in essence is like running on a uniprocessor machine.
+	 */
+	return (unsigned char *)&op;
+}
+
+#ifdef CONFIG_PPC64
+# define _ASM_ALIGN	" .align 3 "
+# define _ASM_PTR	" .llong "
+#else
+# define _ASM_ALIGN	" .align 2 "
+# define _ASM_PTR	" .long "
+#endif
+
+notrace int
+ftrace_modify_code(unsigned long ip, unsigned char *old_code,
+		   unsigned char *new_code)
+{
+	unsigned replaced;
+	unsigned old = *(unsigned *)old_code;
+	unsigned new = *(unsigned *)new_code;
+	int faulted = 0;
+
+	/*
+	 * Note: Due to modules and __init, code can
+	 *  disappear and change, we need to protect against faulting
+	 *  as well as code changing.
+	 *
+	 * No real locking needed, this code is run through
+	 * kstop_machine.
+	 */
+	asm volatile (
+		"1: lwz		%1, 0(%2)\n"
+		"   cmpw	%1, %5\n"
+		"   bne		2f\n"
+		"   stwu	%3, 0(%2)\n"
+		"2:\n"
+		".section .fixup, \"ax\"\n"
+		"3:	li %0, 1\n"
+		"	b 2b\n"
+		".previous\n"
+		".section __ex_table,\"a\"\n"
+		_ASM_ALIGN "\n"
+		_ASM_PTR "1b, 3b\n"
+		".previous"
+		: "=r"(faulted), "=r"(replaced)
+		: "r"(ip), "r"(new),
+		  "0"(faulted), "r"(old)
+		: "memory");
+
+	if (replaced != old && replaced != new)
+		faulted = 2;
+
+	if (!faulted)
+		flush_icache_range(ip, ip + 8);
+
+	return faulted;
+}
+
+notrace int ftrace_update_ftrace_func(ftrace_func_t func)
+{
+	unsigned long ip = (unsigned long)(&ftrace_call);
+	unsigned char old[MCOUNT_INSN_SIZE], *new;
+	int ret;
+
+	memcpy(old, &ftrace_call, MCOUNT_INSN_SIZE);
+	new = ftrace_call_replace(ip, (unsigned long)func);
+	ret = ftrace_modify_code(ip, old, new);
+
+	return ret;
+}
+
+notrace int ftrace_mcount_set(unsigned long *data)
+{
+	unsigned long ip = (long)(&mcount_call);
+	unsigned long *addr = data;
+	unsigned char old[MCOUNT_INSN_SIZE], *new;
+
+	/*
+	 * Replace the mcount stub with a pointer to the
+	 * ip recorder function.
+	 */
+	memcpy(old, &mcount_call, MCOUNT_INSN_SIZE);
+	new = ftrace_call_replace(ip, *addr);
+	*addr = ftrace_modify_code(ip, old, new);
+
+	return 0;
+}
+
+int __init ftrace_dyn_arch_init(void *data)
+{
+	/* This is running in kstop_machine */
+
+	ftrace_mcount_set(data);
+
+	return 0;
+}
+

arch/powerpc/kernel/io.c

@@ -120,7 +120,8 @@ EXPORT_SYMBOL(_outsl_ns);
 
 #define IO_CHECK_ALIGN(v,a) ((((unsigned long)(v)) & ((a) - 1)) == 0)
 
-void _memset_io(volatile void __iomem *addr, int c, unsigned long n)
+notrace void
+_memset_io(volatile void __iomem *addr, int c, unsigned long n)
 {
 	void *p = (void __force *)addr;
 	u32 lc = c;

arch/powerpc/kernel/irq.c

@@ -98,7 +98,7 @@ EXPORT_SYMBOL(irq_desc);
 
 int distribute_irqs = 1;
 
-static inline unsigned long get_hard_enabled(void)
+static inline notrace unsigned long get_hard_enabled(void)
 {
 	unsigned long enabled;
 
@@ -108,13 +108,13 @@ static inline unsigned long get_hard_enabled(void)
 	return enabled;
 }
 
-static inline void set_soft_enabled(unsigned long enable)
+static inline notrace void set_soft_enabled(unsigned long enable)
 {
 	__asm__ __volatile__("stb %0,%1(13)"
 	: : "r" (enable), "i" (offsetof(struct paca_struct, soft_enabled)));
 }
 
-void raw_local_irq_restore(unsigned long en)
+notrace void raw_local_irq_restore(unsigned long en)
 {
 	/*
 	 * get_paca()->soft_enabled = en;

arch/powerpc/kernel/ppc_ksyms.c

@@ -42,6 +42,7 @@
 #include <asm/div64.h>
 #include <asm/signal.h>
 #include <asm/dcr.h>
+#include <asm/ftrace.h>
 
 #ifdef CONFIG_PPC32
 extern void transfer_to_handler(void);
@@ -67,6 +68,10 @@ EXPORT_SYMBOL(single_step_exception);
 EXPORT_SYMBOL(sys_sigreturn);
 #endif
 
+#ifdef CONFIG_FTRACE
+EXPORT_SYMBOL(_mcount);
+#endif
+
 EXPORT_SYMBOL(strcpy);
 EXPORT_SYMBOL(strncpy);
 EXPORT_SYMBOL(strcat);

arch/powerpc/kernel/setup_32.c

@@ -81,7 +81,7 @@ int ucache_bsize;
  * from the address that it was linked at, so we must use RELOC/PTRRELOC
  * to access static data (including strings).  -- paulus
  */
-unsigned long __init early_init(unsigned long dt_ptr)
+notrace unsigned long __init early_init(unsigned long dt_ptr)
 {
 	unsigned long offset = reloc_offset();
 	struct cpu_spec *spec;
@@ -111,7 +111,7 @@ unsigned long __init early_init(unsigned long dt_ptr)
  * This is called very early on the boot process, after a minimal
  * MMU environment has been set up but before MMU_init is called.
  */
-void __init machine_init(unsigned long dt_ptr, unsigned long phys)
+notrace void __init machine_init(unsigned long dt_ptr, unsigned long phys)
 {
 	/* Enable early debugging if any specified (see udbg.h) */
 	udbg_early_init();
@@ -133,7 +133,7 @@ void __init machine_init(unsigned long dt_ptr, unsigned long phys)
 
 #ifdef CONFIG_BOOKE_WDT
 /* Checks wdt=x and wdt_period=xx command-line option */
-int __init early_parse_wdt(char *p)
+notrace int __init early_parse_wdt(char *p)
 {
 	if (p && strncmp(p, "0", 1) != 0)
 	       booke_wdt_enabled = 1;

arch/powerpc/platforms/powermac/Makefile

@@ -1,5 +1,10 @@
 CFLAGS_bootx_init.o  		+= -fPIC
 
+ifdef CONFIG_FTRACE
+# Do not trace early boot code
+CFLAGS_REMOVE_bootx_init.o = -pg
+endif
+
 obj-y				+= pic.o setup.o time.o feature.o pci.o \
 				   sleep.o low_i2c.o cache.o pfunc_core.o \
 				   pfunc_base.o

arch/sparc64/Kconfig

@@ -11,6 +11,8 @@ config SPARC
 config SPARC64
 	bool
 	default y
+	select HAVE_DYNAMIC_FTRACE
+	select HAVE_FTRACE
 	select HAVE_IDE
 	select HAVE_LMB
 	select HAVE_ARCH_KGDB

arch/sparc64/Kconfig.debug

@@ -33,7 +33,7 @@ config DEBUG_PAGEALLOC
 
 config MCOUNT
 	bool
-	depends on STACK_DEBUG
+	depends on STACK_DEBUG || FTRACE
 	default y
 
 config FRAME_POINTER

arch/sparc64/kernel/Makefile

@@ -14,6 +14,7 @@ obj-y		:= process.o setup.o cpu.o idprom.o \
 		   power.o sbus.o sparc64_ksyms.o chmc.o \
 		   visemul.o prom.o of_device.o hvapi.o sstate.o mdesc.o
 
+obj-$(CONFIG_DYNAMIC_FTRACE) += ftrace.o
 obj-$(CONFIG_STACKTRACE) += stacktrace.o
 obj-$(CONFIG_PCI)	 += ebus.o pci_common.o \
 			    pci_psycho.o pci_sabre.o pci_schizo.o \

arch/sparc64/kernel/ftrace.c

@@ -0,0 +1,94 @@
+#include <linux/spinlock.h>
+#include <linux/hardirq.h>
+#include <linux/ftrace.h>
+#include <linux/percpu.h>
+#include <linux/init.h>
+#include <linux/list.h>
+
+#include <asm/ftrace.h>
+
+static const u32 ftrace_nop = 0x01000000;
+
+notrace unsigned char *ftrace_nop_replace(void)
+{
+	return (char *)&ftrace_nop;
+}
+
+notrace unsigned char *ftrace_call_replace(unsigned long ip, unsigned long addr)
+{
+	static u32 call;
+	s32 off;
+
+	off = ((s32)addr - (s32)ip);
+	call = 0x40000000 | ((u32)off >> 2);
+
+	return (unsigned char *) &call;
+}
+
+notrace int
+ftrace_modify_code(unsigned long ip, unsigned char *old_code,
+		   unsigned char *new_code)
+{
+	u32 old = *(u32 *)old_code;
+	u32 new = *(u32 *)new_code;
+	u32 replaced;
+	int faulted;
+
+	__asm__ __volatile__(
+	"1:	cas	[%[ip]], %[old], %[new]\n"
+	"	flush	%[ip]\n"
+	"	mov	0, %[faulted]\n"
+	"2:\n"
+	"	.section .fixup,#alloc,#execinstr\n"
+	"	.align	4\n"
+	"3:	sethi	%%hi(2b), %[faulted]\n"
+	"	jmpl	%[faulted] + %%lo(2b), %%g0\n"
+	"	 mov	1, %[faulted]\n"
+	"	.previous\n"
+	"	.section __ex_table,\"a\"\n"
+	"	.align	4\n"
+	"	.word	1b, 3b\n"
+	"	.previous\n"
+	: "=r" (replaced), [faulted] "=r" (faulted)
+	: [new] "0" (new), [old] "r" (old), [ip] "r" (ip)
+	: "memory");
+
+	if (replaced != old && replaced != new)
+		faulted = 2;
+
+	return faulted;
+}
+
+notrace int ftrace_update_ftrace_func(ftrace_func_t func)
+{
+	unsigned long ip = (unsigned long)(&ftrace_call);
+	unsigned char old[MCOUNT_INSN_SIZE], *new;
+
+	memcpy(old, &ftrace_call, MCOUNT_INSN_SIZE);
+	new = ftrace_call_replace(ip, (unsigned long)func);
+	return ftrace_modify_code(ip, old, new);
+}
+
+notrace int ftrace_mcount_set(unsigned long *data)
+{
+	unsigned long ip = (long)(&mcount_call);
+	unsigned long *addr = data;
+	unsigned char old[MCOUNT_INSN_SIZE], *new;
+
+	/*
+	 * Replace the mcount stub with a pointer to the
+	 * ip recorder function.
+	 */
+	memcpy(old, &mcount_call, MCOUNT_INSN_SIZE);
+	new = ftrace_call_replace(ip, *addr);
+	*addr = ftrace_modify_code(ip, old, new);
+
+	return 0;
+}
+
+
+int __init ftrace_dyn_arch_init(void *data)
+{
+	ftrace_mcount_set(data);
+	return 0;
+}

arch/sparc64/kernel/sparc64_ksyms.c

@@ -53,6 +53,7 @@
 #include <asm/ns87303.h>
 #include <asm/timer.h>
 #include <asm/cpudata.h>
+#include <asm/ftrace.h>
 
 struct poll {
 	int fd;
@@ -111,8 +112,7 @@ EXPORT_SYMBOL(__write_trylock);
 EXPORT_SYMBOL(smp_call_function);
 #endif /* CONFIG_SMP */
 
-#if defined(CONFIG_MCOUNT)
-extern void _mcount(void);
+#ifdef CONFIG_MCOUNT
 EXPORT_SYMBOL(_mcount);
 #endif
 

arch/sparc64/lib/mcount.S

@@ -28,10 +28,13 @@ ovstack:
 	.skip		OVSTACKSIZE
 #endif
 	.text
-	.align 32
-	.globl mcount, _mcount
-mcount:
+	.align		32
+	.globl		_mcount
+	.type		_mcount,#function
+	.globl		mcount
+	.type		mcount,#function
 _mcount:
+mcount:
 #ifdef CONFIG_STACK_DEBUG
 	/*
 	 * Check whether %sp is dangerously low.
@@ -55,6 +58,53 @@ _mcount:
 	 or		%g3, %lo(panicstring), %o0
 	call		prom_halt
 	 nop
+1:
 #endif
-1:	retl
+#ifdef CONFIG_FTRACE
+#ifdef CONFIG_DYNAMIC_FTRACE
+	mov		%o7, %o0
+	.globl		mcount_call
+mcount_call:
+	call		ftrace_stub
+	 mov		%o0, %o7
+#else
+	sethi		%hi(ftrace_trace_function), %g1
+	sethi		%hi(ftrace_stub), %g2
+	ldx		[%g1 + %lo(ftrace_trace_function)], %g1
+	or		%g2, %lo(ftrace_stub), %g2
+	cmp		%g1, %g2
+	be,pn		%icc, 1f
+	 mov		%i7, %o1
+	jmpl		%g1, %g0
+	 mov		%o7, %o0
+	/* not reached */
+1:
+#endif
+#endif
+	retl
 	 nop
+	.size		_mcount,.-_mcount
+	.size		mcount,.-mcount
+
+#ifdef CONFIG_FTRACE
+	.globl		ftrace_stub
+	.type		ftrace_stub,#function
+ftrace_stub:
+	retl
+	 nop
+	.size		ftrace_stub,.-ftrace_stub
+#ifdef CONFIG_DYNAMIC_FTRACE
+	.globl		ftrace_caller
+	.type		ftrace_caller,#function
+ftrace_caller:
+	mov		%i7, %o1
+	mov		%o7, %o0
+	.globl		ftrace_call
+ftrace_call:
+	call		ftrace_stub
+	 mov		%o0, %o7
+	retl
+	 nop
+	.size		ftrace_caller,.-ftrace_caller
+#endif
+#endif

arch/x86/Kconfig

@@ -23,6 +23,8 @@ config X86
 	select HAVE_OPROFILE
 	select HAVE_KPROBES
 	select HAVE_KRETPROBES
+	select HAVE_DYNAMIC_FTRACE
+	select HAVE_FTRACE
 	select HAVE_KVM if ((X86_32 && !X86_VOYAGER && !X86_VISWS && !X86_NUMAQ) || X86_64)
 	select HAVE_ARCH_KGDB if !X86_VOYAGER
 

arch/x86/Kconfig.debug

@@ -171,6 +171,34 @@ config IOMMU_LEAK
 	  Add a simple leak tracer to the IOMMU code. This is useful when you
 	  are debugging a buggy device driver that leaks IOMMU mappings.
 
+config MMIOTRACE_HOOKS
+	bool
+
+config MMIOTRACE
+	bool "Memory mapped IO tracing"
+	depends on DEBUG_KERNEL && PCI
+	select TRACING
+	select MMIOTRACE_HOOKS
+	default y
+	help
+	  Mmiotrace traces Memory Mapped I/O access and is meant for
+	  debugging and reverse engineering. It is called from the ioremap
+	  implementation and works via page faults. Tracing is disabled by
+	  default and can be enabled at run-time.
+
+	  See Documentation/tracers/mmiotrace.txt.
+	  If you are not helping to develop drivers, say N.
+
+config MMIOTRACE_TEST
+	tristate "Test module for mmiotrace"
+	depends on MMIOTRACE && m
+	help
+	  This is a dumb module for testing mmiotrace. It is very dangerous
+	  as it will write garbage to IO memory starting at a given address.
+	  However, it should be safe to use on e.g. unused portion of VRAM.
+
+	  Say N, unless you absolutely know what you are doing.
+
 #
 # IO delay types:
 #

arch/x86/kernel/Makefile

@@ -6,6 +6,13 @@ extra-y                := head_$(BITS).o head$(BITS).o head.o init_task.o vmlinu
 
 CPPFLAGS_vmlinux.lds += -U$(UTS_MACHINE)
 
+ifdef CONFIG_FTRACE
+# Do not profile debug utilities
+CFLAGS_REMOVE_tsc_64.o = -pg
+CFLAGS_REMOVE_tsc_32.o = -pg
+CFLAGS_REMOVE_rtc.o = -pg
+endif
+
 #
 # vsyscalls (which work on the user stack) should have
 # no stack-protector checks:
@@ -57,6 +64,7 @@ obj-$(CONFIG_X86_MPPARSE)	+= mpparse.o
 obj-$(CONFIG_X86_LOCAL_APIC)	+= apic_$(BITS).o nmi.o
 obj-$(CONFIG_X86_IO_APIC)	+= io_apic_$(BITS).o
 obj-$(CONFIG_X86_REBOOTFIXUPS)	+= reboot_fixups_32.o
+obj-$(CONFIG_DYNAMIC_FTRACE)	+= ftrace.o
 obj-$(CONFIG_KEXEC)		+= machine_kexec_$(BITS).o
 obj-$(CONFIG_KEXEC)		+= relocate_kernel_$(BITS).o crash.o
 obj-$(CONFIG_CRASH_DUMP)	+= crash_dump_$(BITS).o

arch/x86/kernel/alternative.c

@@ -1,6 +1,6 @@
 #include <linux/module.h>
 #include <linux/sched.h>
-#include <linux/spinlock.h>
+#include <linux/mutex.h>
 #include <linux/list.h>
 #include <linux/kprobes.h>
 #include <linux/mm.h>
@@ -143,7 +143,7 @@ static const unsigned char *const p6_nops[ASM_NOP_MAX+1] = {
 #ifdef CONFIG_X86_64
 
 extern char __vsyscall_0;
-static inline const unsigned char*const * find_nop_table(void)
+const unsigned char *const *find_nop_table(void)
 {
 	return boot_cpu_data.x86_vendor != X86_VENDOR_INTEL ||
 	       boot_cpu_data.x86 < 6 ? k8_nops : p6_nops;
@@ -162,7 +162,7 @@ static const struct nop {
 	{ -1, NULL }
 };
 
-static const unsigned char*const * find_nop_table(void)
+const unsigned char *const *find_nop_table(void)
 {
 	const unsigned char *const *noptable = intel_nops;
 	int i;
@@ -279,7 +279,7 @@ struct smp_alt_module {
 	struct list_head next;
 };
 static LIST_HEAD(smp_alt_modules);
-static DEFINE_SPINLOCK(smp_alt);
+static DEFINE_MUTEX(smp_alt);
 static int smp_mode = 1;	/* protected by smp_alt */
 
 void alternatives_smp_module_add(struct module *mod, char *name,
@@ -312,12 +312,12 @@ void alternatives_smp_module_add(struct module *mod, char *name,
 		__func__, smp->locks, smp->locks_end,
 		smp->text, smp->text_end, smp->name);
 
-	spin_lock(&smp_alt);
+	mutex_lock(&smp_alt);
 	list_add_tail(&smp->next, &smp_alt_modules);
 	if (boot_cpu_has(X86_FEATURE_UP))
 		alternatives_smp_unlock(smp->locks, smp->locks_end,
 					smp->text, smp->text_end);
-	spin_unlock(&smp_alt);
+	mutex_unlock(&smp_alt);
 }
 
 void alternatives_smp_module_del(struct module *mod)
@@ -327,17 +327,17 @@ void alternatives_smp_module_del(struct module *mod)
 	if (smp_alt_once || noreplace_smp)
 		return;
 
-	spin_lock(&smp_alt);
+	mutex_lock(&smp_alt);
 	list_for_each_entry(item, &smp_alt_modules, next) {
 		if (mod != item->mod)
 			continue;
 		list_del(&item->next);
-		spin_unlock(&smp_alt);
+		mutex_unlock(&smp_alt);
 		DPRINTK("%s: %s\n", __func__, item->name);
 		kfree(item);
 		return;
 	}
-	spin_unlock(&smp_alt);
+	mutex_unlock(&smp_alt);
 }
 
 void alternatives_smp_switch(int smp)
@@ -359,7 +359,7 @@ void alternatives_smp_switch(int smp)
 		return;
 	BUG_ON(!smp && (num_online_cpus() > 1));
 
-	spin_lock(&smp_alt);
+	mutex_lock(&smp_alt);
 
 	/*
 	 * Avoid unnecessary switches because it forces JIT based VMs to
@@ -383,7 +383,7 @@ void alternatives_smp_switch(int smp)
 						mod->text, mod->text_end);
 	}
 	smp_mode = smp;
-	spin_unlock(&smp_alt);
+	mutex_unlock(&smp_alt);
 }
 
 #endif

arch/x86/kernel/entry_32.S

@@ -51,6 +51,7 @@
 #include <asm/percpu.h>
 #include <asm/dwarf2.h>
 #include <asm/processor-flags.h>
+#include <asm/ftrace.h>
 #include <asm/irq_vectors.h>
 
 /*
@@ -1111,6 +1112,77 @@ ENDPROC(xen_failsafe_callback)
 
 #endif	/* CONFIG_XEN */
 
+#ifdef CONFIG_FTRACE
+#ifdef CONFIG_DYNAMIC_FTRACE
+
+ENTRY(mcount)
+	pushl %eax
+	pushl %ecx
+	pushl %edx
+	movl 0xc(%esp), %eax
+	subl $MCOUNT_INSN_SIZE, %eax
+
+.globl mcount_call
+mcount_call:
+	call ftrace_stub
+
+	popl %edx
+	popl %ecx
+	popl %eax
+
+	ret
+END(mcount)
+
+ENTRY(ftrace_caller)
+	pushl %eax
+	pushl %ecx
+	pushl %edx
+	movl 0xc(%esp), %eax
+	movl 0x4(%ebp), %edx
+	subl $MCOUNT_INSN_SIZE, %eax
+
+.globl ftrace_call
+ftrace_call:
+	call ftrace_stub
+
+	popl %edx
+	popl %ecx
+	popl %eax
+
+.globl ftrace_stub
+ftrace_stub:
+	ret
+END(ftrace_caller)
+
+#else /* ! CONFIG_DYNAMIC_FTRACE */
+
+ENTRY(mcount)
+	cmpl $ftrace_stub, ftrace_trace_function
+	jnz trace
+.globl ftrace_stub
+ftrace_stub:
+	ret
+
+	/* taken from glibc */
+trace:
+	pushl %eax
+	pushl %ecx
+	pushl %edx
+	movl 0xc(%esp), %eax
+	movl 0x4(%ebp), %edx
+	subl $MCOUNT_INSN_SIZE, %eax
+
+	call *ftrace_trace_function
+
+	popl %edx
+	popl %ecx
+	popl %eax
+
+	jmp ftrace_stub
+END(mcount)
+#endif /* CONFIG_DYNAMIC_FTRACE */
+#endif /* CONFIG_FTRACE */
+
 .section .rodata,"a"
 #include "syscall_table_32.S"
 

arch/x86/kernel/entry_64.S

@@ -51,9 +51,115 @@
 #include <asm/page.h>
 #include <asm/irqflags.h>
 #include <asm/paravirt.h>
+#include <asm/ftrace.h>
 
 	.code64
 
+#ifdef CONFIG_FTRACE
+#ifdef CONFIG_DYNAMIC_FTRACE
+ENTRY(mcount)
+
+	subq $0x38, %rsp
+	movq %rax, (%rsp)
+	movq %rcx, 8(%rsp)
+	movq %rdx, 16(%rsp)
+	movq %rsi, 24(%rsp)
+	movq %rdi, 32(%rsp)
+	movq %r8, 40(%rsp)
+	movq %r9, 48(%rsp)
+
+	movq 0x38(%rsp), %rdi
+	subq $MCOUNT_INSN_SIZE, %rdi
+
+.globl mcount_call
+mcount_call:
+	call ftrace_stub
+
+	movq 48(%rsp), %r9
+	movq 40(%rsp), %r8
+	movq 32(%rsp), %rdi
+	movq 24(%rsp), %rsi
+	movq 16(%rsp), %rdx
+	movq 8(%rsp), %rcx
+	movq (%rsp), %rax
+	addq $0x38, %rsp
+
+	retq
+END(mcount)
+
+ENTRY(ftrace_caller)
+
+	/* taken from glibc */
+	subq $0x38, %rsp
+	movq %rax, (%rsp)
+	movq %rcx, 8(%rsp)
+	movq %rdx, 16(%rsp)
+	movq %rsi, 24(%rsp)
+	movq %rdi, 32(%rsp)
+	movq %r8, 40(%rsp)
+	movq %r9, 48(%rsp)
+
+	movq 0x38(%rsp), %rdi
+	movq 8(%rbp), %rsi
+	subq $MCOUNT_INSN_SIZE, %rdi
+
+.globl ftrace_call
+ftrace_call:
+	call ftrace_stub
+
+	movq 48(%rsp), %r9
+	movq 40(%rsp), %r8
+	movq 32(%rsp), %rdi
+	movq 24(%rsp), %rsi
+	movq 16(%rsp), %rdx
+	movq 8(%rsp), %rcx
+	movq (%rsp), %rax
+	addq $0x38, %rsp
+
+.globl ftrace_stub
+ftrace_stub:
+	retq
+END(ftrace_caller)
+
+#else /* ! CONFIG_DYNAMIC_FTRACE */
+ENTRY(mcount)
+	cmpq $ftrace_stub, ftrace_trace_function
+	jnz trace
+.globl ftrace_stub
+ftrace_stub:
+	retq
+
+trace:
+	/* taken from glibc */
+	subq $0x38, %rsp
+	movq %rax, (%rsp)
+	movq %rcx, 8(%rsp)
+	movq %rdx, 16(%rsp)
+	movq %rsi, 24(%rsp)
+	movq %rdi, 32(%rsp)
+	movq %r8, 40(%rsp)
+	movq %r9, 48(%rsp)
+
+	movq 0x38(%rsp), %rdi
+	movq 8(%rbp), %rsi
+	subq $MCOUNT_INSN_SIZE, %rdi
+
+	call   *ftrace_trace_function
+
+	movq 48(%rsp), %r9
+	movq 40(%rsp), %r8
+	movq 32(%rsp), %rdi
+	movq 24(%rsp), %rsi
+	movq 16(%rsp), %rdx
+	movq 8(%rsp), %rcx
+	movq (%rsp), %rax
+	addq $0x38, %rsp
+
+	jmp ftrace_stub
+END(mcount)
+#endif /* CONFIG_DYNAMIC_FTRACE */
+#endif /* CONFIG_FTRACE */
+
 #ifndef CONFIG_PREEMPT
 #define retint_kernel retint_restore_args
 #endif	

arch/x86/kernel/ftrace.c

@@ -0,0 +1,141 @@
+/*
+ * Code for replacing ftrace calls with jumps.
+ *
+ * Copyright (C) 2007-2008 Steven Rostedt <srostedt@redhat.com>
+ *
+ * Thanks goes to Ingo Molnar, for suggesting the idea.
+ * Mathieu Desnoyers, for suggesting postponing the modifications.
+ * Arjan van de Ven, for keeping me straight, and explaining to me
+ * the dangers of modifying code on the run.
+ */
+
+#include <linux/spinlock.h>
+#include <linux/hardirq.h>
+#include <linux/ftrace.h>
+#include <linux/percpu.h>
+#include <linux/init.h>
+#include <linux/list.h>
+
+#include <asm/alternative.h>
+#include <asm/ftrace.h>
+
+
+/* Long is fine, even if it is only 4 bytes ;-) */
+static long *ftrace_nop;
+
+union ftrace_code_union {
+	char code[MCOUNT_INSN_SIZE];
+	struct {
+		char e8;
+		int offset;
+	} __attribute__((packed));
+};
+
+
+static int notrace ftrace_calc_offset(long ip, long addr)
+{
+	return (int)(addr - ip);
+}
+
+notrace unsigned char *ftrace_nop_replace(void)
+{
+	return (char *)ftrace_nop;
+}
+
+notrace unsigned char *ftrace_call_replace(unsigned long ip, unsigned long addr)
+{
+	static union ftrace_code_union calc;
+
+	calc.e8		= 0xe8;
+	calc.offset	= ftrace_calc_offset(ip + MCOUNT_INSN_SIZE, addr);
+
+	/*
+	 * No locking needed, this must be called via kstop_machine
+	 * which in essence is like running on a uniprocessor machine.
+	 */
+	return calc.code;
+}
+
+notrace int
+ftrace_modify_code(unsigned long ip, unsigned char *old_code,
+		   unsigned char *new_code)
+{
+	unsigned replaced;
+	unsigned old = *(unsigned *)old_code; /* 4 bytes */
+	unsigned new = *(unsigned *)new_code; /* 4 bytes */
+	unsigned char newch = new_code[4];
+	int faulted = 0;
+
+	/*
+	 * Note: Due to modules and __init, code can
+	 *  disappear and change, we need to protect against faulting
+	 *  as well as code changing.
+	 *
+	 * No real locking needed, this code is run through
+	 * kstop_machine.
+	 */
+	asm volatile (
+		"1: lock\n"
+		"   cmpxchg %3, (%2)\n"
+		"   jnz 2f\n"
+		"   movb %b4, 4(%2)\n"
+		"2:\n"
+		".section .fixup, \"ax\"\n"
+		"3:	movl $1, %0\n"
+		"	jmp 2b\n"
+		".previous\n"
+		_ASM_EXTABLE(1b, 3b)
+		: "=r"(faulted), "=a"(replaced)
+		: "r"(ip), "r"(new), "c"(newch),
+		  "0"(faulted), "a"(old)
+		: "memory");
+	sync_core();
+
+	if (replaced != old && replaced != new)
+		faulted = 2;
+
+	return faulted;
+}
+
+notrace int ftrace_update_ftrace_func(ftrace_func_t func)
+{
+	unsigned long ip = (unsigned long)(&ftrace_call);
+	unsigned char old[MCOUNT_INSN_SIZE], *new;
+	int ret;
+
+	memcpy(old, &ftrace_call, MCOUNT_INSN_SIZE);
+	new = ftrace_call_replace(ip, (unsigned long)func);
+	ret = ftrace_modify_code(ip, old, new);
+
+	return ret;
+}
+
+notrace int ftrace_mcount_set(unsigned long *data)
+{
+	unsigned long ip = (long)(&mcount_call);
+	unsigned long *addr = data;
+	unsigned char old[MCOUNT_INSN_SIZE], *new;
+
+	/*
+	 * Replace the mcount stub with a pointer to the
+	 * ip recorder function.
+	 */
+	memcpy(old, &mcount_call, MCOUNT_INSN_SIZE);
+	new = ftrace_call_replace(ip, *addr);
+	*addr = ftrace_modify_code(ip, old, new);
+
+	return 0;
+}
+
+int __init ftrace_dyn_arch_init(void *data)
+{
+	const unsigned char *const *noptable = find_nop_table();
+
+	/* This is running in kstop_machine */
+
+	ftrace_mcount_set(data);
+
+	ftrace_nop = (unsigned long *)noptable[MCOUNT_INSN_SIZE];
+
+	return 0;
+}

arch/x86/kernel/i386_ksyms_32.c

@@ -1,7 +1,14 @@
 #include <linux/module.h>
+
 #include <asm/checksum.h>
-#include <asm/desc.h>
 #include <asm/pgtable.h>
+#include <asm/desc.h>
+#include <asm/ftrace.h>
+
+#ifdef CONFIG_FTRACE
+/* mcount is defined in assembly */
+EXPORT_SYMBOL(mcount);
+#endif
 
 /* Networking helper routines. */
 EXPORT_SYMBOL(csum_partial_copy_generic);

arch/x86/kernel/machine_kexec_32.c

@@ -11,6 +11,8 @@
 #include <linux/delay.h>
 #include <linux/init.h>
 #include <linux/numa.h>
+#include <linux/ftrace.h>
+
 #include <asm/pgtable.h>
 #include <asm/pgalloc.h>
 #include <asm/tlbflush.h>
@@ -107,6 +109,8 @@ NORET_TYPE void machine_kexec(struct kimage *image)
 	unsigned long page_list[PAGES_NR];
 	void *control_page;
 
+	tracer_disable();
+
 	/* Interrupts aren't acceptable while we reboot */
 	local_irq_disable();
 

arch/x86/kernel/machine_kexec_64.c

@@ -11,6 +11,8 @@
 #include <linux/string.h>
 #include <linux/reboot.h>
 #include <linux/numa.h>
+#include <linux/ftrace.h>
+
 #include <asm/pgtable.h>
 #include <asm/tlbflush.h>
 #include <asm/mmu_context.h>
@@ -184,6 +186,8 @@ NORET_TYPE void machine_kexec(struct kimage *image)
 	unsigned long page_list[PAGES_NR];
 	void *control_page;
 
+	tracer_disable();
+
 	/* Interrupts aren't acceptable while we reboot */
 	local_irq_disable();
 

arch/x86/kernel/process_32.c

@@ -142,7 +142,10 @@ void cpu_idle(void)
 
 			local_irq_disable();
 			__get_cpu_var(irq_stat).idle_timestamp = jiffies;
+			/* Don't trace irqs off for idle */
+			stop_critical_timings();
 			pm_idle();
+			start_critical_timings();
 		}
 		tick_nohz_restart_sched_tick();
 		preempt_enable_no_resched();

arch/x86/kernel/process_64.c

@@ -134,7 +134,10 @@ void cpu_idle(void)
 			 */
 			local_irq_disable();
 			enter_idle();
+			/* Don't trace irqs off for idle */
+			stop_critical_timings();
 			pm_idle();
+			start_critical_timings();
 			/* In many cases the interrupt that ended idle
 			   has already called exit_idle. But some idle
 			   loops can be woken up without interrupt. */

arch/x86/kernel/vsyscall_64.c

@@ -42,7 +42,8 @@
 #include <asm/topology.h>
 #include <asm/vgtod.h>
 
-#define __vsyscall(nr) __attribute__ ((unused,__section__(".vsyscall_" #nr)))
+#define __vsyscall(nr) \
+		__attribute__ ((unused, __section__(".vsyscall_" #nr))) notrace
 #define __syscall_clobber "r11","cx","memory"
 
 /*

arch/x86/kernel/x8664_ksyms_64.c

@@ -2,13 +2,20 @@
    All C exports should go in the respective C files. */
 
 #include <linux/module.h>
-#include <net/checksum.h>
 #include <linux/smp.h>
 
+#include <net/checksum.h>
+
 #include <asm/processor.h>
-#include <asm/uaccess.h>
 #include <asm/pgtable.h>
+#include <asm/uaccess.h>
 #include <asm/desc.h>
+#include <asm/ftrace.h>
+
+#ifdef CONFIG_FTRACE
+/* mcount is defined in assembly */
+EXPORT_SYMBOL(mcount);
+#endif
 
 EXPORT_SYMBOL(kernel_thread);
 

arch/x86/lib/Makefile

@@ -5,6 +5,7 @@
 obj-$(CONFIG_SMP) := msr-on-cpu.o
 
 lib-y := delay.o
+lib-y += thunk_$(BITS).o
 lib-y += usercopy_$(BITS).o getuser.o putuser.o
 lib-y += memcpy_$(BITS).o
 

arch/x86/lib/thunk_32.S

@@ -0,0 +1,47 @@
+/*
+ * Trampoline to trace irqs off. (otherwise CALLER_ADDR1 might crash)
+ * Copyright 2008 by Steven Rostedt, Red Hat, Inc
+ *  (inspired by Andi Kleen's thunk_64.S)
+ * Subject to the GNU public license, v.2. No warranty of any kind.
+ */
+
+	#include <linux/linkage.h>
+
+#define ARCH_TRACE_IRQS_ON			\
+	pushl %eax;				\
+	pushl %ecx;				\
+	pushl %edx;				\
+	call trace_hardirqs_on;			\
+	popl %edx;				\
+	popl %ecx;				\
+	popl %eax;
+
+#define ARCH_TRACE_IRQS_OFF			\
+	pushl %eax;				\
+	pushl %ecx;				\
+	pushl %edx;				\
+	call trace_hardirqs_off;		\
+	popl %edx;				\
+	popl %ecx;				\
+	popl %eax;
+
+#ifdef CONFIG_TRACE_IRQFLAGS
+	/* put return address in eax (arg1) */
+	.macro thunk_ra name,func
+	.globl \name
+\name:
+	pushl %eax
+	pushl %ecx
+	pushl %edx
+	/* Place EIP in the arg1 */
+	movl 3*4(%esp), %eax
+	call \func
+	popl %edx
+	popl %ecx
+	popl %eax
+	ret
+	.endm
+
+	thunk_ra trace_hardirqs_on_thunk,trace_hardirqs_on_caller
+	thunk_ra trace_hardirqs_off_thunk,trace_hardirqs_off_caller
+#endif

arch/x86/lib/thunk_64.S

@@ -2,6 +2,7 @@
  * Save registers before calling assembly functions. This avoids
  * disturbance of register allocation in some inline assembly constructs.
  * Copyright 2001,2002 by Andi Kleen, SuSE Labs.
+ * Added trace_hardirqs callers - Copyright 2007 Steven Rostedt, Red Hat, Inc.
  * Subject to the GNU public license, v.2. No warranty of any kind.
  */
 
@@ -42,8 +43,22 @@
 #endif	
 	
 #ifdef CONFIG_TRACE_IRQFLAGS
-	thunk trace_hardirqs_on_thunk,trace_hardirqs_on
-	thunk trace_hardirqs_off_thunk,trace_hardirqs_off
+	/* put return address in rdi (arg1) */
+	.macro thunk_ra name,func
+	.globl \name
+\name:
+	CFI_STARTPROC
+	SAVE_ARGS
+	/* SAVE_ARGS pushs 9 elements */
+	/* the next element would be the rip */
+	movq 9*8(%rsp), %rdi
+	call \func
+	jmp  restore
+	CFI_ENDPROC
+	.endm
+
+	thunk_ra trace_hardirqs_on_thunk,trace_hardirqs_on_caller
+	thunk_ra trace_hardirqs_off_thunk,trace_hardirqs_off_caller
 #endif
 
 #ifdef CONFIG_DEBUG_LOCK_ALLOC

arch/x86/mm/Makefile

@@ -8,6 +8,11 @@ obj-$(CONFIG_X86_PTDUMP)	+= dump_pagetables.o
 
 obj-$(CONFIG_HIGHMEM)		+= highmem_32.o
 
+obj-$(CONFIG_MMIOTRACE_HOOKS)	+= kmmio.o
+obj-$(CONFIG_MMIOTRACE)		+= mmiotrace.o
+mmiotrace-y			:= pf_in.o mmio-mod.o
+obj-$(CONFIG_MMIOTRACE_TEST)	+= testmmiotrace.o
+
 ifeq ($(CONFIG_X86_32),y)
 obj-$(CONFIG_NUMA)		+= discontig_32.o
 else

arch/x86/mm/fault.c

@@ -10,6 +10,7 @@
 #include <linux/string.h>
 #include <linux/types.h>
 #include <linux/ptrace.h>
+#include <linux/mmiotrace.h>
 #include <linux/mman.h>
 #include <linux/mm.h>
 #include <linux/smp.h>
@@ -49,6 +50,16 @@
 #define PF_RSVD		(1<<3)
 #define PF_INSTR	(1<<4)
 
+static inline int kmmio_fault(struct pt_regs *regs, unsigned long addr)
+{
+#ifdef CONFIG_MMIOTRACE_HOOKS
+	if (unlikely(is_kmmio_active()))
+		if (kmmio_handler(regs, addr) == 1)
+			return -1;
+#endif
+	return 0;
+}
+
 static inline int notify_page_fault(struct pt_regs *regs)
 {
 #ifdef CONFIG_KPROBES
@@ -598,6 +609,8 @@ void __kprobes do_page_fault(struct pt_regs *regs, unsigned long error_code)
 
 	if (notify_page_fault(regs))
 		return;
+	if (unlikely(kmmio_fault(regs, address)))
+		return;
 
 	/*
 	 * We fault-in kernel-space virtual memory on-demand. The

arch/x86/mm/init_32.c

@@ -1035,6 +1035,8 @@ void mark_rodata_ro(void)
 	unsigned long start = PFN_ALIGN(_text);
 	unsigned long size = PFN_ALIGN(_etext) - start;
 
+#ifndef CONFIG_DYNAMIC_FTRACE
+	/* Dynamic tracing modifies the kernel text section */
 	set_pages_ro(virt_to_page(start), size >> PAGE_SHIFT);
 	printk(KERN_INFO "Write protecting the kernel text: %luk\n",
 		size >> 10);
@@ -1047,6 +1049,8 @@ void mark_rodata_ro(void)
 	printk(KERN_INFO "Testing CPA: write protecting again\n");
 	set_pages_ro(virt_to_page(start), size>>PAGE_SHIFT);
 #endif
+#endif /* CONFIG_DYNAMIC_FTRACE */
+
 	start += size;
 	size = (unsigned long)__end_rodata - start;
 	set_pages_ro(virt_to_page(start), size >> PAGE_SHIFT);

arch/x86/mm/init_64.c

@@ -991,6 +991,13 @@ EXPORT_SYMBOL_GPL(rodata_test_data);
 void mark_rodata_ro(void)
 {
 	unsigned long start = PFN_ALIGN(_stext), end = PFN_ALIGN(__end_rodata);
+	unsigned long rodata_start =
+		((unsigned long)__start_rodata + PAGE_SIZE - 1) & PAGE_MASK;
+
+#ifdef CONFIG_DYNAMIC_FTRACE
+	/* Dynamic tracing modifies the kernel text section */
+	start = rodata_start;
+#endif
 
 	printk(KERN_INFO "Write protecting the kernel read-only data: %luk\n",
 	       (end - start) >> 10);
@@ -1000,8 +1007,7 @@ void mark_rodata_ro(void)
 	 * The rodata section (but not the kernel text!) should also be
 	 * not-executable.
 	 */
-	start = ((unsigned long)__start_rodata + PAGE_SIZE - 1) & PAGE_MASK;
-	set_memory_nx(start, (end - start) >> PAGE_SHIFT);
+	set_memory_nx(rodata_start, (end - rodata_start) >> PAGE_SHIFT);
 
 	rodata_test();
 

arch/x86/mm/ioremap.c

@@ -12,6 +12,7 @@
 #include <linux/module.h>
 #include <linux/slab.h>
 #include <linux/vmalloc.h>
+#include <linux/mmiotrace.h>
 
 #include <asm/cacheflush.h>
 #include <asm/e820.h>
@@ -122,10 +123,13 @@ static void __iomem *__ioremap_caller(resource_size_t phys_addr,
 {
 	unsigned long pfn, offset, vaddr;
 	resource_size_t last_addr;
+	const resource_size_t unaligned_phys_addr = phys_addr;
+	const unsigned long unaligned_size = size;
 	struct vm_struct *area;
 	unsigned long new_prot_val;
 	pgprot_t prot;
 	int retval;
+	void __iomem *ret_addr;
 
 	/* Don't allow wraparound or zero size */
 	last_addr = phys_addr + size - 1;
@@ -233,7 +237,10 @@ static void __iomem *__ioremap_caller(resource_size_t phys_addr,
 		return NULL;
 	}
 
-	return (void __iomem *) (vaddr + offset);
+	ret_addr = (void __iomem *) (vaddr + offset);
+	mmiotrace_ioremap(unaligned_phys_addr, unaligned_size, ret_addr);
+
+	return ret_addr;
 }
 
 /**
@@ -348,6 +355,8 @@ void iounmap(volatile void __iomem *addr)
 	addr = (volatile void __iomem *)
 		(PAGE_MASK & (unsigned long __force)addr);
 
+	mmiotrace_iounmap(addr);
+
 	/* Use the vm area unlocked, assuming the caller
 	   ensures there isn't another iounmap for the same address
 	   in parallel. Reuse of the virtual address is prevented by

arch/x86/mm/kmmio.c

@@ -0,0 +1,510 @@
+/* Support for MMIO probes.
+ * Benfit many code from kprobes
+ * (C) 2002 Louis Zhuang <louis.zhuang@intel.com>.
+ *     2007 Alexander Eichner
+ *     2008 Pekka Paalanen <pq@iki.fi>
+ */
+
+#include <linux/list.h>
+#include <linux/rculist.h>
+#include <linux/spinlock.h>
+#include <linux/hash.h>
+#include <linux/init.h>
+#include <linux/module.h>
+#include <linux/kernel.h>
+#include <linux/uaccess.h>
+#include <linux/ptrace.h>
+#include <linux/preempt.h>
+#include <linux/percpu.h>
+#include <linux/kdebug.h>
+#include <linux/mutex.h>
+#include <linux/io.h>
+#include <asm/cacheflush.h>
+#include <asm/tlbflush.h>
+#include <linux/errno.h>
+#include <asm/debugreg.h>
+#include <linux/mmiotrace.h>
+
+#define KMMIO_PAGE_HASH_BITS 4
+#define KMMIO_PAGE_TABLE_SIZE (1 << KMMIO_PAGE_HASH_BITS)
+
+struct kmmio_fault_page {
+	struct list_head list;
+	struct kmmio_fault_page *release_next;
+	unsigned long page; /* location of the fault page */
+
+	/*
+	 * Number of times this page has been registered as a part
+	 * of a probe. If zero, page is disarmed and this may be freed.
+	 * Used only by writers (RCU).
+	 */
+	int count;
+};
+
+struct kmmio_delayed_release {
+	struct rcu_head rcu;
+	struct kmmio_fault_page *release_list;
+};
+
+struct kmmio_context {
+	struct kmmio_fault_page *fpage;
+	struct kmmio_probe *probe;
+	unsigned long saved_flags;
+	unsigned long addr;
+	int active;
+};
+
+static DEFINE_SPINLOCK(kmmio_lock);
+
+/* Protected by kmmio_lock */
+unsigned int kmmio_count;
+
+/* Read-protected by RCU, write-protected by kmmio_lock. */
+static struct list_head kmmio_page_table[KMMIO_PAGE_TABLE_SIZE];
+static LIST_HEAD(kmmio_probes);
+
+static struct list_head *kmmio_page_list(unsigned long page)
+{
+	return &kmmio_page_table[hash_long(page, KMMIO_PAGE_HASH_BITS)];
+}
+
+/* Accessed per-cpu */
+static DEFINE_PER_CPU(struct kmmio_context, kmmio_ctx);
+
+/*
+ * this is basically a dynamic stabbing problem:
+ * Could use the existing prio tree code or
+ * Possible better implementations:
+ * The Interval Skip List: A Data Structure for Finding All Intervals That
+ * Overlap a Point (might be simple)
+ * Space Efficient Dynamic Stabbing with Fast Queries - Mikkel Thorup
+ */
+/* Get the kmmio at this addr (if any). You must be holding RCU read lock. */
+static struct kmmio_probe *get_kmmio_probe(unsigned long addr)
+{
+	struct kmmio_probe *p;
+	list_for_each_entry_rcu(p, &kmmio_probes, list) {
+		if (addr >= p->addr && addr <= (p->addr + p->len))
+			return p;
+	}
+	return NULL;
+}
+
+/* You must be holding RCU read lock. */
+static struct kmmio_fault_page *get_kmmio_fault_page(unsigned long page)
+{
+	struct list_head *head;
+	struct kmmio_fault_page *p;
+
+	page &= PAGE_MASK;
+	head = kmmio_page_list(page);
+	list_for_each_entry_rcu(p, head, list) {
+		if (p->page == page)
+			return p;
+	}
+	return NULL;
+}
+
+static void set_page_present(unsigned long addr, bool present,
+							unsigned int *pglevel)
+{
+	pteval_t pteval;
+	pmdval_t pmdval;
+	unsigned int level;
+	pmd_t *pmd;
+	pte_t *pte = lookup_address(addr, &level);
+
+	if (!pte) {
+		pr_err("kmmio: no pte for page 0x%08lx\n", addr);
+		return;
+	}
+
+	if (pglevel)
+		*pglevel = level;
+
+	switch (level) {
+	case PG_LEVEL_2M:
+		pmd = (pmd_t *)pte;
+		pmdval = pmd_val(*pmd) & ~_PAGE_PRESENT;
+		if (present)
+			pmdval |= _PAGE_PRESENT;
+		set_pmd(pmd, __pmd(pmdval));
+		break;
+
+	case PG_LEVEL_4K:
+		pteval = pte_val(*pte) & ~_PAGE_PRESENT;
+		if (present)
+			pteval |= _PAGE_PRESENT;
+		set_pte_atomic(pte, __pte(pteval));
+		break;
+
+	default:
+		pr_err("kmmio: unexpected page level 0x%x.\n", level);
+		return;
+	}
+
+	__flush_tlb_one(addr);
+}
+
+/** Mark the given page as not present. Access to it will trigger a fault. */
+static void arm_kmmio_fault_page(unsigned long page, unsigned int *pglevel)
+{
+	set_page_present(page & PAGE_MASK, false, pglevel);
+}
+
+/** Mark the given page as present. */
+static void disarm_kmmio_fault_page(unsigned long page, unsigned int *pglevel)
+{
+	set_page_present(page & PAGE_MASK, true, pglevel);
+}
+
+/*
+ * This is being called from do_page_fault().
+ *
+ * We may be in an interrupt or a critical section. Also prefecthing may
+ * trigger a page fault. We may be in the middle of process switch.
+ * We cannot take any locks, because we could be executing especially
+ * within a kmmio critical section.
+ *
+ * Local interrupts are disabled, so preemption cannot happen.
+ * Do not enable interrupts, do not sleep, and watch out for other CPUs.
+ */
+/*
+ * Interrupts are disabled on entry as trap3 is an interrupt gate
+ * and they remain disabled thorough out this function.
+ */
+int kmmio_handler(struct pt_regs *regs, unsigned long addr)
+{
+	struct kmmio_context *ctx;
+	struct kmmio_fault_page *faultpage;
+	int ret = 0; /* default to fault not handled */
+
+	/*
+	 * Preemption is now disabled to prevent process switch during
+	 * single stepping. We can only handle one active kmmio trace
+	 * per cpu, so ensure that we finish it before something else
+	 * gets to run. We also hold the RCU read lock over single
+	 * stepping to avoid looking up the probe and kmmio_fault_page
+	 * again.
+	 */
+	preempt_disable();
+	rcu_read_lock();
+
+	faultpage = get_kmmio_fault_page(addr);
+	if (!faultpage) {
+		/*
+		 * Either this page fault is not caused by kmmio, or
+		 * another CPU just pulled the kmmio probe from under
+		 * our feet. The latter case should not be possible.
+		 */
+		goto no_kmmio;
+	}
+
+	ctx = &get_cpu_var(kmmio_ctx);
+	if (ctx->active) {
+		disarm_kmmio_fault_page(faultpage->page, NULL);
+		if (addr == ctx->addr) {
+			/*
+			 * On SMP we sometimes get recursive probe hits on the
+			 * same address. Context is already saved, fall out.
+			 */
+			pr_debug("kmmio: duplicate probe hit on CPU %d, for "
+						"address 0x%08lx.\n",
+						smp_processor_id(), addr);
+			ret = 1;
+			goto no_kmmio_ctx;
+		}
+		/*
+		 * Prevent overwriting already in-flight context.
+		 * This should not happen, let's hope disarming at least
+		 * prevents a panic.
+		 */
+		pr_emerg("kmmio: recursive probe hit on CPU %d, "
+					"for address 0x%08lx. Ignoring.\n",
+					smp_processor_id(), addr);
+		pr_emerg("kmmio: previous hit was at 0x%08lx.\n",
+					ctx->addr);
+		goto no_kmmio_ctx;
+	}
+	ctx->active++;
+
+	ctx->fpage = faultpage;
+	ctx->probe = get_kmmio_probe(addr);
+	ctx->saved_flags = (regs->flags & (X86_EFLAGS_TF | X86_EFLAGS_IF));
+	ctx->addr = addr;
+
+	if (ctx->probe && ctx->probe->pre_handler)
+		ctx->probe->pre_handler(ctx->probe, regs, addr);
+
+	/*
+	 * Enable single-stepping and disable interrupts for the faulting
+	 * context. Local interrupts must not get enabled during stepping.
+	 */
+	regs->flags |= X86_EFLAGS_TF;
+	regs->flags &= ~X86_EFLAGS_IF;
+
+	/* Now we set present bit in PTE and single step. */
+	disarm_kmmio_fault_page(ctx->fpage->page, NULL);
+
+	/*
+	 * If another cpu accesses the same page while we are stepping,
+	 * the access will not be caught. It will simply succeed and the
+	 * only downside is we lose the event. If this becomes a problem,
+	 * the user should drop to single cpu before tracing.
+	 */
+
+	put_cpu_var(kmmio_ctx);
+	return 1; /* fault handled */
+
+no_kmmio_ctx:
+	put_cpu_var(kmmio_ctx);
+no_kmmio:
+	rcu_read_unlock();
+	preempt_enable_no_resched();
+	return ret;
+}
+
+/*
+ * Interrupts are disabled on entry as trap1 is an interrupt gate
+ * and they remain disabled thorough out this function.
+ * This must always get called as the pair to kmmio_handler().
+ */
+static int post_kmmio_handler(unsigned long condition, struct pt_regs *regs)
+{
+	int ret = 0;
+	struct kmmio_context *ctx = &get_cpu_var(kmmio_ctx);
+
+	if (!ctx->active) {
+		pr_debug("kmmio: spurious debug trap on CPU %d.\n",
+							smp_processor_id());
+		goto out;
+	}
+
+	if (ctx->probe && ctx->probe->post_handler)
+		ctx->probe->post_handler(ctx->probe, condition, regs);
+
+	arm_kmmio_fault_page(ctx->fpage->page, NULL);
+
+	regs->flags &= ~X86_EFLAGS_TF;
+	regs->flags |= ctx->saved_flags;
+
+	/* These were acquired in kmmio_handler(). */
+	ctx->active--;
+	BUG_ON(ctx->active);
+	rcu_read_unlock();
+	preempt_enable_no_resched();
+
+	/*
+	 * if somebody else is singlestepping across a probe point, flags
+	 * will have TF set, in which case, continue the remaining processing
+	 * of do_debug, as if this is not a probe hit.
+	 */
+	if (!(regs->flags & X86_EFLAGS_TF))
+		ret = 1;
+out:
+	put_cpu_var(kmmio_ctx);
+	return ret;
+}
+
+/* You must be holding kmmio_lock. */
+static int add_kmmio_fault_page(unsigned long page)
+{
+	struct kmmio_fault_page *f;
+
+	page &= PAGE_MASK;
+	f = get_kmmio_fault_page(page);
+	if (f) {
+		if (!f->count)
+			arm_kmmio_fault_page(f->page, NULL);
+		f->count++;
+		return 0;
+	}
+
+	f = kmalloc(sizeof(*f), GFP_ATOMIC);
+	if (!f)
+		return -1;
+
+	f->count = 1;
+	f->page = page;
+	list_add_rcu(&f->list, kmmio_page_list(f->page));
+
+	arm_kmmio_fault_page(f->page, NULL);
+
+	return 0;
+}
+
+/* You must be holding kmmio_lock. */
+static void release_kmmio_fault_page(unsigned long page,
+				struct kmmio_fault_page **release_list)
+{
+	struct kmmio_fault_page *f;
+
+	page &= PAGE_MASK;
+	f = get_kmmio_fault_page(page);
+	if (!f)
+		return;
+
+	f->count--;
+	BUG_ON(f->count < 0);
+	if (!f->count) {
+		disarm_kmmio_fault_page(f->page, NULL);
+		f->release_next = *release_list;
+		*release_list = f;
+	}
+}
+
+/*
+ * With page-unaligned ioremaps, one or two armed pages may contain
+ * addresses from outside the intended mapping. Events for these addresses
+ * are currently silently dropped. The events may result only from programming
+ * mistakes by accessing addresses before the beginning or past the end of a
+ * mapping.
+ */
+int register_kmmio_probe(struct kmmio_probe *p)
+{
+	unsigned long flags;
+	int ret = 0;
+	unsigned long size = 0;
+	const unsigned long size_lim = p->len + (p->addr & ~PAGE_MASK);
+
+	spin_lock_irqsave(&kmmio_lock, flags);
+	if (get_kmmio_probe(p->addr)) {
+		ret = -EEXIST;
+		goto out;
+	}
+	kmmio_count++;
+	list_add_rcu(&p->list, &kmmio_probes);
+	while (size < size_lim) {
+		if (add_kmmio_fault_page(p->addr + size))
+			pr_err("kmmio: Unable to set page fault.\n");
+		size += PAGE_SIZE;
+	}
+out:
+	spin_unlock_irqrestore(&kmmio_lock, flags);
+	/*
+	 * XXX: What should I do here?
+	 * Here was a call to global_flush_tlb(), but it does not exist
+	 * anymore. It seems it's not needed after all.
+	 */
+	return ret;
+}
+EXPORT_SYMBOL(register_kmmio_probe);
+
+static void rcu_free_kmmio_fault_pages(struct rcu_head *head)
+{
+	struct kmmio_delayed_release *dr = container_of(
+						head,
+						struct kmmio_delayed_release,
+						rcu);
+	struct kmmio_fault_page *p = dr->release_list;
+	while (p) {
+		struct kmmio_fault_page *next = p->release_next;
+		BUG_ON(p->count);
+		kfree(p);
+		p = next;
+	}
+	kfree(dr);
+}
+
+static void remove_kmmio_fault_pages(struct rcu_head *head)
+{
+	struct kmmio_delayed_release *dr = container_of(
+						head,
+						struct kmmio_delayed_release,
+						rcu);
+	struct kmmio_fault_page *p = dr->release_list;
+	struct kmmio_fault_page **prevp = &dr->release_list;
+	unsigned long flags;
+	spin_lock_irqsave(&kmmio_lock, flags);
+	while (p) {
+		if (!p->count)
+			list_del_rcu(&p->list);
+		else
+			*prevp = p->release_next;
+		prevp = &p->release_next;
+		p = p->release_next;
+	}
+	spin_unlock_irqrestore(&kmmio_lock, flags);
+	/* This is the real RCU destroy call. */
+	call_rcu(&dr->rcu, rcu_free_kmmio_fault_pages);
+}
+
+/*
+ * Remove a kmmio probe. You have to synchronize_rcu() before you can be
+ * sure that the callbacks will not be called anymore. Only after that
+ * you may actually release your struct kmmio_probe.
+ *
+ * Unregistering a kmmio fault page has three steps:
+ * 1. release_kmmio_fault_page()
+ *    Disarm the page, wait a grace period to let all faults finish.
+ * 2. remove_kmmio_fault_pages()
+ *    Remove the pages from kmmio_page_table.
+ * 3. rcu_free_kmmio_fault_pages()
+ *    Actally free the kmmio_fault_page structs as with RCU.
+ */
+void unregister_kmmio_probe(struct kmmio_probe *p)
+{
+	unsigned long flags;
+	unsigned long size = 0;
+	const unsigned long size_lim = p->len + (p->addr & ~PAGE_MASK);
+	struct kmmio_fault_page *release_list = NULL;
+	struct kmmio_delayed_release *drelease;
+
+	spin_lock_irqsave(&kmmio_lock, flags);
+	while (size < size_lim) {
+		release_kmmio_fault_page(p->addr + size, &release_list);
+		size += PAGE_SIZE;
+	}
+	list_del_rcu(&p->list);
+	kmmio_count--;
+	spin_unlock_irqrestore(&kmmio_lock, flags);
+
+	drelease = kmalloc(sizeof(*drelease), GFP_ATOMIC);
+	if (!drelease) {
+		pr_crit("kmmio: leaking kmmio_fault_page objects.\n");
+		return;
+	}
+	drelease->release_list = release_list;
+
+	/*
+	 * This is not really RCU here. We have just disarmed a set of
+	 * pages so that they cannot trigger page faults anymore. However,
+	 * we cannot remove the pages from kmmio_page_table,
+	 * because a probe hit might be in flight on another CPU. The
+	 * pages are collected into a list, and they will be removed from
+	 * kmmio_page_table when it is certain that no probe hit related to
+	 * these pages can be in flight. RCU grace period sounds like a
+	 * good choice.
+	 *
+	 * If we removed the pages too early, kmmio page fault handler might
+	 * not find the respective kmmio_fault_page and determine it's not
+	 * a kmmio fault, when it actually is. This would lead to madness.
+	 */
+	call_rcu(&drelease->rcu, remove_kmmio_fault_pages);
+}
+EXPORT_SYMBOL(unregister_kmmio_probe);
+
+static int kmmio_die_notifier(struct notifier_block *nb, unsigned long val,
+								void *args)
+{
+	struct die_args *arg = args;
+
+	if (val == DIE_DEBUG && (arg->err & DR_STEP))
+		if (post_kmmio_handler(arg->err, arg->regs) == 1)
+			return NOTIFY_STOP;
+
+	return NOTIFY_DONE;
+}
+
+static struct notifier_block nb_die = {
+	.notifier_call = kmmio_die_notifier
+};
+
+static int __init init_kmmio(void)
+{
+	int i;
+	for (i = 0; i < KMMIO_PAGE_TABLE_SIZE; i++)
+		INIT_LIST_HEAD(&kmmio_page_table[i]);
+	return register_die_notifier(&nb_die);
+}
+fs_initcall(init_kmmio); /* should be before device_initcall() */

arch/x86/mm/mmio-mod.c

@@ -0,0 +1,515 @@
+/*
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation; either version 2 of the License, or
+ * (at your option) any later version.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA.
+ *
+ * Copyright (C) IBM Corporation, 2005
+ *               Jeff Muizelaar, 2006, 2007
+ *               Pekka Paalanen, 2008 <pq@iki.fi>
+ *
+ * Derived from the read-mod example from relay-examples by Tom Zanussi.
+ */
+#define DEBUG 1
+
+#include <linux/module.h>
+#include <linux/debugfs.h>
+#include <linux/uaccess.h>
+#include <linux/io.h>
+#include <linux/version.h>
+#include <linux/kallsyms.h>
+#include <asm/pgtable.h>
+#include <linux/mmiotrace.h>
+#include <asm/e820.h> /* for ISA_START_ADDRESS */
+#include <asm/atomic.h>
+#include <linux/percpu.h>
+#include <linux/cpu.h>
+
+#include "pf_in.h"
+
+#define NAME "mmiotrace: "
+
+struct trap_reason {
+	unsigned long addr;
+	unsigned long ip;
+	enum reason_type type;
+	int active_traces;
+};
+
+struct remap_trace {
+	struct list_head list;
+	struct kmmio_probe probe;
+	resource_size_t phys;
+	unsigned long id;
+};
+
+/* Accessed per-cpu. */
+static DEFINE_PER_CPU(struct trap_reason, pf_reason);
+static DEFINE_PER_CPU(struct mmiotrace_rw, cpu_trace);
+
+#if 0 /* XXX: no way gather this info anymore */
+/* Access to this is not per-cpu. */
+static DEFINE_PER_CPU(atomic_t, dropped);
+#endif
+
+static struct dentry *marker_file;
+
+static DEFINE_MUTEX(mmiotrace_mutex);
+static DEFINE_SPINLOCK(trace_lock);
+static atomic_t mmiotrace_enabled;
+static LIST_HEAD(trace_list);		/* struct remap_trace */
+
+/*
+ * Locking in this file:
+ * - mmiotrace_mutex enforces enable/disable_mmiotrace() critical sections.
+ * - mmiotrace_enabled may be modified only when holding mmiotrace_mutex
+ *   and trace_lock.
+ * - Routines depending on is_enabled() must take trace_lock.
+ * - trace_list users must hold trace_lock.
+ * - is_enabled() guarantees that mmio_trace_record is allowed.
+ * - pre/post callbacks assume the effect of is_enabled() being true.
+ */
+
+/* module parameters */
+static unsigned long	filter_offset;
+static int		nommiotrace;
+static int		trace_pc;
+
+module_param(filter_offset, ulong, 0);
+module_param(nommiotrace, bool, 0);
+module_param(trace_pc, bool, 0);
+
+MODULE_PARM_DESC(filter_offset, "Start address of traced mappings.");
+MODULE_PARM_DESC(nommiotrace, "Disable actual MMIO tracing.");
+MODULE_PARM_DESC(trace_pc, "Record address of faulting instructions.");
+
+static bool is_enabled(void)
+{
+	return atomic_read(&mmiotrace_enabled);
+}
+
+#if 0 /* XXX: needs rewrite */
+/*
+ * Write callback for the debugfs entry:
+ * Read a marker and write it to the mmio trace log
+ */
+static ssize_t write_marker(struct file *file, const char __user *buffer,
+						size_t count, loff_t *ppos)
+{
+	char *event = NULL;
+	struct mm_io_header *headp;
+	ssize_t len = (count > 65535) ? 65535 : count;
+
+	event = kzalloc(sizeof(*headp) + len, GFP_KERNEL);
+	if (!event)
+		return -ENOMEM;
+
+	headp = (struct mm_io_header *)event;
+	headp->type = MMIO_MAGIC | (MMIO_MARKER << MMIO_OPCODE_SHIFT);
+	headp->data_len = len;
+
+	if (copy_from_user(event + sizeof(*headp), buffer, len)) {
+		kfree(event);
+		return -EFAULT;
+	}
+
+	spin_lock_irq(&trace_lock);
+#if 0 /* XXX: convert this to use tracing */
+	if (is_enabled())
+		relay_write(chan, event, sizeof(*headp) + len);
+	else
+#endif
+		len = -EINVAL;
+	spin_unlock_irq(&trace_lock);
+	kfree(event);
+	return len;
+}
+#endif
+
+static void print_pte(unsigned long address)
+{
+	unsigned int level;
+	pte_t *pte = lookup_address(address, &level);
+
+	if (!pte) {
+		pr_err(NAME "Error in %s: no pte for page 0x%08lx\n",
+							__func__, address);
+		return;
+	}
+
+	if (level == PG_LEVEL_2M) {
+		pr_emerg(NAME "4MB pages are not currently supported: "
+							"0x%08lx\n", address);
+		BUG();
+	}
+	pr_info(NAME "pte for 0x%lx: 0x%llx 0x%llx\n", address,
+		(unsigned long long)pte_val(*pte),
+		(unsigned long long)pte_val(*pte) & _PAGE_PRESENT);
+}
+
+/*
+ * For some reason the pre/post pairs have been called in an
+ * unmatched order. Report and die.
+ */
+static void die_kmmio_nesting_error(struct pt_regs *regs, unsigned long addr)
+{
+	const struct trap_reason *my_reason = &get_cpu_var(pf_reason);
+	pr_emerg(NAME "unexpected fault for address: 0x%08lx, "
+					"last fault for address: 0x%08lx\n",
+					addr, my_reason->addr);
+	print_pte(addr);
+	print_symbol(KERN_EMERG "faulting IP is at %s\n", regs->ip);
+	print_symbol(KERN_EMERG "last faulting IP was at %s\n", my_reason->ip);
+#ifdef __i386__
+	pr_emerg("eax: %08lx   ebx: %08lx   ecx: %08lx   edx: %08lx\n",
+			regs->ax, regs->bx, regs->cx, regs->dx);
+	pr_emerg("esi: %08lx   edi: %08lx   ebp: %08lx   esp: %08lx\n",
+			regs->si, regs->di, regs->bp, regs->sp);
+#else
+	pr_emerg("rax: %016lx   rcx: %016lx   rdx: %016lx\n",
+					regs->ax, regs->cx, regs->dx);
+	pr_emerg("rsi: %016lx   rdi: %016lx   rbp: %016lx   rsp: %016lx\n",
+				regs->si, regs->di, regs->bp, regs->sp);
+#endif
+	put_cpu_var(pf_reason);
+	BUG();
+}
+
+static void pre(struct kmmio_probe *p, struct pt_regs *regs,
+						unsigned long addr)
+{
+	struct trap_reason *my_reason = &get_cpu_var(pf_reason);
+	struct mmiotrace_rw *my_trace = &get_cpu_var(cpu_trace);
+	const unsigned long instptr = instruction_pointer(regs);
+	const enum reason_type type = get_ins_type(instptr);
+	struct remap_trace *trace = p->private;
+
+	/* it doesn't make sense to have more than one active trace per cpu */
+	if (my_reason->active_traces)
+		die_kmmio_nesting_error(regs, addr);
+	else
+		my_reason->active_traces++;
+
+	my_reason->type = type;
+	my_reason->addr = addr;
+	my_reason->ip = instptr;
+
+	my_trace->phys = addr - trace->probe.addr + trace->phys;
+	my_trace->map_id = trace->id;
+
+	/*
+	 * Only record the program counter when requested.
+	 * It may taint clean-room reverse engineering.
+	 */
+	if (trace_pc)
+		my_trace->pc = instptr;
+	else
+		my_trace->pc = 0;
+
+	/*
+	 * XXX: the timestamp recorded will be *after* the tracing has been
+	 * done, not at the time we hit the instruction. SMP implications
+	 * on event ordering?
+	 */
+
+	switch (type) {
+	case REG_READ:
+		my_trace->opcode = MMIO_READ;
+		my_trace->width = get_ins_mem_width(instptr);
+		break;
+	case REG_WRITE:
+		my_trace->opcode = MMIO_WRITE;
+		my_trace->width = get_ins_mem_width(instptr);
+		my_trace->value = get_ins_reg_val(instptr, regs);
+		break;
+	case IMM_WRITE:
+		my_trace->opcode = MMIO_WRITE;
+		my_trace->width = get_ins_mem_width(instptr);
+		my_trace->value = get_ins_imm_val(instptr);
+		break;
+	default:
+		{
+			unsigned char *ip = (unsigned char *)instptr;
+			my_trace->opcode = MMIO_UNKNOWN_OP;
+			my_trace->width = 0;
+			my_trace->value = (*ip) << 16 | *(ip + 1) << 8 |
+								*(ip + 2);
+		}
+	}
+	put_cpu_var(cpu_trace);
+	put_cpu_var(pf_reason);
+}
+
+static void post(struct kmmio_probe *p, unsigned long condition,
+							struct pt_regs *regs)
+{
+	struct trap_reason *my_reason = &get_cpu_var(pf_reason);
+	struct mmiotrace_rw *my_trace = &get_cpu_var(cpu_trace);
+
+	/* this should always return the active_trace count to 0 */
+	my_reason->active_traces--;
+	if (my_reason->active_traces) {
+		pr_emerg(NAME "unexpected post handler");
+		BUG();
+	}
+
+	switch (my_reason->type) {
+	case REG_READ:
+		my_trace->value = get_ins_reg_val(my_reason->ip, regs);
+		break;
+	default:
+		break;
+	}
+
+	mmio_trace_rw(my_trace);
+	put_cpu_var(cpu_trace);
+	put_cpu_var(pf_reason);
+}
+
+static void ioremap_trace_core(resource_size_t offset, unsigned long size,
+							void __iomem *addr)
+{
+	static atomic_t next_id;
+	struct remap_trace *trace = kmalloc(sizeof(*trace), GFP_KERNEL);
+	/* These are page-unaligned. */
+	struct mmiotrace_map map = {
+		.phys = offset,
+		.virt = (unsigned long)addr,
+		.len = size,
+		.opcode = MMIO_PROBE
+	};
+
+	if (!trace) {
+		pr_err(NAME "kmalloc failed in ioremap\n");
+		return;
+	}
+
+	*trace = (struct remap_trace) {
+		.probe = {
+			.addr = (unsigned long)addr,
+			.len = size,
+			.pre_handler = pre,
+			.post_handler = post,
+			.private = trace
+		},
+		.phys = offset,
+		.id = atomic_inc_return(&next_id)
+	};
+	map.map_id = trace->id;
+
+	spin_lock_irq(&trace_lock);
+	if (!is_enabled())
+		goto not_enabled;
+
+	mmio_trace_mapping(&map);
+	list_add_tail(&trace->list, &trace_list);
+	if (!nommiotrace)
+		register_kmmio_probe(&trace->probe);
+
+not_enabled:
+	spin_unlock_irq(&trace_lock);
+}
+
+void mmiotrace_ioremap(resource_size_t offset, unsigned long size,
+						void __iomem *addr)
+{
+	if (!is_enabled()) /* recheck and proper locking in *_core() */
+		return;
+
+	pr_debug(NAME "ioremap_*(0x%llx, 0x%lx) = %p\n",
+				(unsigned long long)offset, size, addr);
+	if ((filter_offset) && (offset != filter_offset))
+		return;
+	ioremap_trace_core(offset, size, addr);
+}
+
+static void iounmap_trace_core(volatile void __iomem *addr)
+{
+	struct mmiotrace_map map = {
+		.phys = 0,
+		.virt = (unsigned long)addr,
+		.len = 0,
+		.opcode = MMIO_UNPROBE
+	};
+	struct remap_trace *trace;
+	struct remap_trace *tmp;
+	struct remap_trace *found_trace = NULL;
+
+	pr_debug(NAME "Unmapping %p.\n", addr);
+
+	spin_lock_irq(&trace_lock);
+	if (!is_enabled())
+		goto not_enabled;
+
+	list_for_each_entry_safe(trace, tmp, &trace_list, list) {
+		if ((unsigned long)addr == trace->probe.addr) {
+			if (!nommiotrace)
+				unregister_kmmio_probe(&trace->probe);
+			list_del(&trace->list);
+			found_trace = trace;
+			break;
+		}
+	}
+	map.map_id = (found_trace) ? found_trace->id : -1;
+	mmio_trace_mapping(&map);
+
+not_enabled:
+	spin_unlock_irq(&trace_lock);
+	if (found_trace) {
+		synchronize_rcu(); /* unregister_kmmio_probe() requirement */
+		kfree(found_trace);
+	}
+}
+
+void mmiotrace_iounmap(volatile void __iomem *addr)
+{
+	might_sleep();
+	if (is_enabled()) /* recheck and proper locking in *_core() */
+		iounmap_trace_core(addr);
+}
+
+static void clear_trace_list(void)
+{
+	struct remap_trace *trace;
+	struct remap_trace *tmp;
+
+	/*
+	 * No locking required, because the caller ensures we are in a
+	 * critical section via mutex, and is_enabled() is false,
+	 * i.e. nothing can traverse or modify this list.
+	 * Caller also ensures is_enabled() cannot change.
+	 */
+	list_for_each_entry(trace, &trace_list, list) {
+		pr_notice(NAME "purging non-iounmapped "
+					"trace @0x%08lx, size 0x%lx.\n",
+					trace->probe.addr, trace->probe.len);
+		if (!nommiotrace)
+			unregister_kmmio_probe(&trace->probe);
+	}
+	synchronize_rcu(); /* unregister_kmmio_probe() requirement */
+
+	list_for_each_entry_safe(trace, tmp, &trace_list, list) {
+		list_del(&trace->list);
+		kfree(trace);
+	}
+}
+
+#ifdef CONFIG_HOTPLUG_CPU
+static cpumask_t downed_cpus;
+
+static void enter_uniprocessor(void)
+{
+	int cpu;
+	int err;
+
+	get_online_cpus();
+	downed_cpus = cpu_online_map;
+	cpu_clear(first_cpu(cpu_online_map), downed_cpus);
+	if (num_online_cpus() > 1)
+		pr_notice(NAME "Disabling non-boot CPUs...\n");
+	put_online_cpus();
+
+	for_each_cpu_mask(cpu, downed_cpus) {
+		err = cpu_down(cpu);
+		if (!err)
+			pr_info(NAME "CPU%d is down.\n", cpu);
+		else
+			pr_err(NAME "Error taking CPU%d down: %d\n", cpu, err);
+	}
+	if (num_online_cpus() > 1)
+		pr_warning(NAME "multiple CPUs still online, "
+						"may miss events.\n");
+}
+
+static void leave_uniprocessor(void)
+{
+	int cpu;
+	int err;
+
+	if (cpus_weight(downed_cpus) == 0)
+		return;
+	pr_notice(NAME "Re-enabling CPUs...\n");
+	for_each_cpu_mask(cpu, downed_cpus) {
+		err = cpu_up(cpu);
+		if (!err)
+			pr_info(NAME "enabled CPU%d.\n", cpu);
+		else
+			pr_err(NAME "cannot re-enable CPU%d: %d\n", cpu, err);
+	}
+}
+
+#else /* !CONFIG_HOTPLUG_CPU */
+static void enter_uniprocessor(void)
+{
+	if (num_online_cpus() > 1)
+		pr_warning(NAME "multiple CPUs are online, may miss events. "
+			"Suggest booting with maxcpus=1 kernel argument.\n");
+}
+
+static void leave_uniprocessor(void)
+{
+}
+#endif
+
+#if 0 /* XXX: out of order */
+static struct file_operations fops_marker = {
+	.owner =	THIS_MODULE,
+	.write =	write_marker
+};
+#endif
+
+void enable_mmiotrace(void)
+{
+	mutex_lock(&mmiotrace_mutex);
+	if (is_enabled())
+		goto out;
+
+#if 0 /* XXX: tracing does not support text entries */
+	marker_file = debugfs_create_file("marker", 0660, dir, NULL,
+								&fops_marker);
+	if (!marker_file)
+		pr_err(NAME "marker file creation failed.\n");
+#endif
+
+	if (nommiotrace)
+		pr_info(NAME "MMIO tracing disabled.\n");
+	enter_uniprocessor();
+	spin_lock_irq(&trace_lock);
+	atomic_inc(&mmiotrace_enabled);
+	spin_unlock_irq(&trace_lock);
+	pr_info(NAME "enabled.\n");
+out:
+	mutex_unlock(&mmiotrace_mutex);
+}
+
+void disable_mmiotrace(void)
+{
+	mutex_lock(&mmiotrace_mutex);
+	if (!is_enabled())
+		goto out;
+
+	spin_lock_irq(&trace_lock);
+	atomic_dec(&mmiotrace_enabled);
+	BUG_ON(is_enabled());
+	spin_unlock_irq(&trace_lock);
+
+	clear_trace_list(); /* guarantees: no more kmmio callbacks */
+	leave_uniprocessor();
+	if (marker_file) {
+		debugfs_remove(marker_file);
+		marker_file = NULL;
+	}
+
+	pr_info(NAME "disabled.\n");
+out:
+	mutex_unlock(&mmiotrace_mutex);
+}

arch/x86/mm/pageattr.c

@@ -262,6 +262,7 @@ pte_t *lookup_address(unsigned long address, unsigned int *level)
 
 	return pte_offset_kernel(pmd, address);
 }
+EXPORT_SYMBOL_GPL(lookup_address);
 
 /*
  * Set the new pmd in all the pgds we know about:

arch/x86/mm/pf_in.c

@@ -0,0 +1,489 @@
+/*
+ *  Fault Injection Test harness (FI)
+ *  Copyright (C) Intel Crop.
+ *
+ *  This program is free software; you can redistribute it and/or
+ *  modify it under the terms of the GNU General Public License
+ *  as published by the Free Software Foundation; either version 2
+ *  of the License, or (at your option) any later version.
+ *
+ *  This program is distributed in the hope that it will be useful,
+ *  but WITHOUT ANY WARRANTY; without even the implied warranty of
+ *  MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ *  GNU General Public License for more details.
+ *
+ *  You should have received a copy of the GNU General Public License
+ *  along with this program; if not, write to the Free Software
+ *  Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA  02111-1307,
+ *  USA.
+ *
+ */
+
+/*  Id: pf_in.c,v 1.1.1.1 2002/11/12 05:56:32 brlock Exp
+ *  Copyright by Intel Crop., 2002
+ *  Louis Zhuang (louis.zhuang@intel.com)
+ *
+ *  Bjorn Steinbrink (B.Steinbrink@gmx.de), 2007
+ */
+
+#include <linux/module.h>
+#include <linux/ptrace.h> /* struct pt_regs */
+#include "pf_in.h"
+
+#ifdef __i386__
+/* IA32 Manual 3, 2-1 */
+static unsigned char prefix_codes[] = {
+	0xF0, 0xF2, 0xF3, 0x2E, 0x36, 0x3E, 0x26, 0x64,
+	0x65, 0x2E, 0x3E, 0x66, 0x67
+};
+/* IA32 Manual 3, 3-432*/
+static unsigned int reg_rop[] = {
+	0x8A, 0x8B, 0xB60F, 0xB70F, 0xBE0F, 0xBF0F
+};
+static unsigned int reg_wop[] = { 0x88, 0x89 };
+static unsigned int imm_wop[] = { 0xC6, 0xC7 };
+/* IA32 Manual 3, 3-432*/
+static unsigned int rw8[] = { 0x88, 0x8A, 0xC6 };
+static unsigned int rw32[] = {
+	0x89, 0x8B, 0xC7, 0xB60F, 0xB70F, 0xBE0F, 0xBF0F
+};
+static unsigned int mw8[] = { 0x88, 0x8A, 0xC6, 0xB60F, 0xBE0F };
+static unsigned int mw16[] = { 0xB70F, 0xBF0F };
+static unsigned int mw32[] = { 0x89, 0x8B, 0xC7 };
+static unsigned int mw64[] = {};
+#else /* not __i386__ */
+static unsigned char prefix_codes[] = {
+	0x66, 0x67, 0x2E, 0x3E, 0x26, 0x64, 0x65, 0x36,
+	0xF0, 0xF3, 0xF2,
+	/* REX Prefixes */
+	0x40, 0x41, 0x42, 0x43, 0x44, 0x45, 0x46, 0x47,
+	0x48, 0x49, 0x4a, 0x4b, 0x4c, 0x4d, 0x4e, 0x4f
+};
+/* AMD64 Manual 3, Appendix A*/
+static unsigned int reg_rop[] = {
+	0x8A, 0x8B, 0xB60F, 0xB70F, 0xBE0F, 0xBF0F
+};
+static unsigned int reg_wop[] = { 0x88, 0x89 };
+static unsigned int imm_wop[] = { 0xC6, 0xC7 };
+static unsigned int rw8[] = { 0xC6, 0x88, 0x8A };
+static unsigned int rw32[] = {
+	0xC7, 0x89, 0x8B, 0xB60F, 0xB70F, 0xBE0F, 0xBF0F
+};
+/* 8 bit only */
+static unsigned int mw8[] = { 0xC6, 0x88, 0x8A, 0xB60F, 0xBE0F };
+/* 16 bit only */
+static unsigned int mw16[] = { 0xB70F, 0xBF0F };
+/* 16 or 32 bit */
+static unsigned int mw32[] = { 0xC7 };
+/* 16, 32 or 64 bit */
+static unsigned int mw64[] = { 0x89, 0x8B };
+#endif /* not __i386__ */
+
+static int skip_prefix(unsigned char *addr, int *shorted, int *enlarged,
+								int *rexr)
+{
+	int i;
+	unsigned char *p = addr;
+	*shorted = 0;
+	*enlarged = 0;
+	*rexr = 0;
+
+restart:
+	for (i = 0; i < ARRAY_SIZE(prefix_codes); i++) {
+		if (*p == prefix_codes[i]) {
+			if (*p == 0x66)
+				*shorted = 1;
+#ifdef __amd64__
+			if ((*p & 0xf8) == 0x48)
+				*enlarged = 1;
+			if ((*p & 0xf4) == 0x44)
+				*rexr = 1;
+#endif
+			p++;
+			goto restart;
+		}
+	}
+
+	return (p - addr);
+}
+
+static int get_opcode(unsigned char *addr, unsigned int *opcode)
+{
+	int len;
+
+	if (*addr == 0x0F) {
+		/* 0x0F is extension instruction */
+		*opcode = *(unsigned short *)addr;
+		len = 2;
+	} else {
+		*opcode = *addr;
+		len = 1;
+	}
+
+	return len;
+}
+
+#define CHECK_OP_TYPE(opcode, array, type) \
+	for (i = 0; i < ARRAY_SIZE(array); i++) { \
+		if (array[i] == opcode) { \
+			rv = type; \
+			goto exit; \
+		} \
+	}
+
+enum reason_type get_ins_type(unsigned long ins_addr)
+{
+	unsigned int opcode;
+	unsigned char *p;
+	int shorted, enlarged, rexr;
+	int i;
+	enum reason_type rv = OTHERS;
+
+	p = (unsigned char *)ins_addr;
+	p += skip_prefix(p, &shorted, &enlarged, &rexr);
+	p += get_opcode(p, &opcode);
+
+	CHECK_OP_TYPE(opcode, reg_rop, REG_READ);
+	CHECK_OP_TYPE(opcode, reg_wop, REG_WRITE);
+	CHECK_OP_TYPE(opcode, imm_wop, IMM_WRITE);
+
+exit:
+	return rv;
+}
+#undef CHECK_OP_TYPE
+
+static unsigned int get_ins_reg_width(unsigned long ins_addr)
+{
+	unsigned int opcode;
+	unsigned char *p;
+	int i, shorted, enlarged, rexr;
+
+	p = (unsigned char *)ins_addr;
+	p += skip_prefix(p, &shorted, &enlarged, &rexr);
+	p += get_opcode(p, &opcode);
+
+	for (i = 0; i < ARRAY_SIZE(rw8); i++)
+		if (rw8[i] == opcode)
+			return 1;
+
+	for (i = 0; i < ARRAY_SIZE(rw32); i++)
+		if (rw32[i] == opcode)
+			return (shorted ? 2 : (enlarged ? 8 : 4));
+
+	printk(KERN_ERR "mmiotrace: Unknown opcode 0x%02x\n", opcode);
+	return 0;
+}
+
+unsigned int get_ins_mem_width(unsigned long ins_addr)
+{
+	unsigned int opcode;
+	unsigned char *p;
+	int i, shorted, enlarged, rexr;
+
+	p = (unsigned char *)ins_addr;
+	p += skip_prefix(p, &shorted, &enlarged, &rexr);
+	p += get_opcode(p, &opcode);
+
+	for (i = 0; i < ARRAY_SIZE(mw8); i++)
+		if (mw8[i] == opcode)
+			return 1;
+
+	for (i = 0; i < ARRAY_SIZE(mw16); i++)
+		if (mw16[i] == opcode)
+			return 2;
+
+	for (i = 0; i < ARRAY_SIZE(mw32); i++)
+		if (mw32[i] == opcode)
+			return shorted ? 2 : 4;
+
+	for (i = 0; i < ARRAY_SIZE(mw64); i++)
+		if (mw64[i] == opcode)
+			return shorted ? 2 : (enlarged ? 8 : 4);
+
+	printk(KERN_ERR "mmiotrace: Unknown opcode 0x%02x\n", opcode);
+	return 0;
+}
+
+/*
+ * Define register ident in mod/rm byte.
+ * Note: these are NOT the same as in ptrace-abi.h.
+ */
+enum {
+	arg_AL = 0,
+	arg_CL = 1,
+	arg_DL = 2,
+	arg_BL = 3,
+	arg_AH = 4,
+	arg_CH = 5,
+	arg_DH = 6,
+	arg_BH = 7,
+
+	arg_AX = 0,
+	arg_CX = 1,
+	arg_DX = 2,
+	arg_BX = 3,
+	arg_SP = 4,
+	arg_BP = 5,
+	arg_SI = 6,
+	arg_DI = 7,
+#ifdef __amd64__
+	arg_R8  = 8,
+	arg_R9  = 9,
+	arg_R10 = 10,
+	arg_R11 = 11,
+	arg_R12 = 12,
+	arg_R13 = 13,
+	arg_R14 = 14,
+	arg_R15 = 15
+#endif
+};
+
+static unsigned char *get_reg_w8(int no, struct pt_regs *regs)
+{
+	unsigned char *rv = NULL;
+
+	switch (no) {
+	case arg_AL:
+		rv = (unsigned char *)&regs->ax;
+		break;
+	case arg_BL:
+		rv = (unsigned char *)&regs->bx;
+		break;
+	case arg_CL:
+		rv = (unsigned char *)&regs->cx;
+		break;
+	case arg_DL:
+		rv = (unsigned char *)&regs->dx;
+		break;
+	case arg_AH:
+		rv = 1 + (unsigned char *)&regs->ax;
+		break;
+	case arg_BH:
+		rv = 1 + (unsigned char *)&regs->bx;
+		break;
+	case arg_CH:
+		rv = 1 + (unsigned char *)&regs->cx;
+		break;
+	case arg_DH:
+		rv = 1 + (unsigned char *)&regs->dx;
+		break;
+#ifdef __amd64__
+	case arg_R8:
+		rv = (unsigned char *)&regs->r8;
+		break;
+	case arg_R9:
+		rv = (unsigned char *)&regs->r9;
+		break;
+	case arg_R10:
+		rv = (unsigned char *)&regs->r10;
+		break;
+	case arg_R11:
+		rv = (unsigned char *)&regs->r11;
+		break;
+	case arg_R12:
+		rv = (unsigned char *)&regs->r12;
+		break;
+	case arg_R13:
+		rv = (unsigned char *)&regs->r13;
+		break;
+	case arg_R14:
+		rv = (unsigned char *)&regs->r14;
+		break;
+	case arg_R15:
+		rv = (unsigned char *)&regs->r15;
+		break;
+#endif
+	default:
+		printk(KERN_ERR "mmiotrace: Error reg no# %d\n", no);
+		break;
+	}
+	return rv;
+}
+
+static unsigned long *get_reg_w32(int no, struct pt_regs *regs)
+{
+	unsigned long *rv = NULL;
+
+	switch (no) {
+	case arg_AX:
+		rv = &regs->ax;
+		break;
+	case arg_BX:
+		rv = &regs->bx;
+		break;
+	case arg_CX:
+		rv = &regs->cx;
+		break;
+	case arg_DX:
+		rv = &regs->dx;
+		break;
+	case arg_SP:
+		rv = &regs->sp;
+		break;
+	case arg_BP:
+		rv = &regs->bp;
+		break;
+	case arg_SI:
+		rv = &regs->si;
+		break;
+	case arg_DI:
+		rv = &regs->di;
+		break;
+#ifdef __amd64__
+	case arg_R8:
+		rv = &regs->r8;
+		break;
+	case arg_R9:
+		rv = &regs->r9;
+		break;
+	case arg_R10:
+		rv = &regs->r10;
+		break;
+	case arg_R11:
+		rv = &regs->r11;
+		break;
+	case arg_R12:
+		rv = &regs->r12;
+		break;
+	case arg_R13:
+		rv = &regs->r13;
+		break;
+	case arg_R14:
+		rv = &regs->r14;
+		break;
+	case arg_R15:
+		rv = &regs->r15;
+		break;
+#endif
+	default:
+		printk(KERN_ERR "mmiotrace: Error reg no# %d\n", no);
+	}
+
+	return rv;
+}
+
+unsigned long get_ins_reg_val(unsigned long ins_addr, struct pt_regs *regs)
+{
+	unsigned int opcode;
+	unsigned char mod_rm;
+	int reg;
+	unsigned char *p;
+	int i, shorted, enlarged, rexr;
+	unsigned long rv;
+
+	p = (unsigned char *)ins_addr;
+	p += skip_prefix(p, &shorted, &enlarged, &rexr);
+	p += get_opcode(p, &opcode);
+	for (i = 0; i < ARRAY_SIZE(reg_rop); i++)
+		if (reg_rop[i] == opcode) {
+			rv = REG_READ;
+			goto do_work;
+		}
+
+	for (i = 0; i < ARRAY_SIZE(reg_wop); i++)
+		if (reg_wop[i] == opcode) {
+			rv = REG_WRITE;
+			goto do_work;
+		}
+
+	printk(KERN_ERR "mmiotrace: Not a register instruction, opcode "
+							"0x%02x\n", opcode);
+	goto err;
+
+do_work:
+	mod_rm = *p;
+	reg = ((mod_rm >> 3) & 0x7) | (rexr << 3);
+	switch (get_ins_reg_width(ins_addr)) {
+	case 1:
+		return *get_reg_w8(reg, regs);
+
+	case 2:
+		return *(unsigned short *)get_reg_w32(reg, regs);
+
+	case 4:
+		return *(unsigned int *)get_reg_w32(reg, regs);
+
+#ifdef __amd64__
+	case 8:
+		return *(unsigned long *)get_reg_w32(reg, regs);
+#endif
+
+	default:
+		printk(KERN_ERR "mmiotrace: Error width# %d\n", reg);
+	}
+
+err:
+	return 0;
+}
+
+unsigned long get_ins_imm_val(unsigned long ins_addr)
+{
+	unsigned int opcode;
+	unsigned char mod_rm;
+	unsigned char mod;
+	unsigned char *p;
+	int i, shorted, enlarged, rexr;
+	unsigned long rv;
+
+	p = (unsigned char *)ins_addr;
+	p += skip_prefix(p, &shorted, &enlarged, &rexr);
+	p += get_opcode(p, &opcode);
+	for (i = 0; i < ARRAY_SIZE(imm_wop); i++)
+		if (imm_wop[i] == opcode) {
+			rv = IMM_WRITE;
+			goto do_work;
+		}
+
+	printk(KERN_ERR "mmiotrace: Not an immediate instruction, opcode "
+							"0x%02x\n", opcode);
+	goto err;
+
+do_work:
+	mod_rm = *p;
+	mod = mod_rm >> 6;
+	p++;
+	switch (mod) {
+	case 0:
+		/* if r/m is 5 we have a 32 disp (IA32 Manual 3, Table 2-2)  */
+		/* AMD64: XXX Check for address size prefix? */
+		if ((mod_rm & 0x7) == 0x5)
+			p += 4;
+		break;
+
+	case 1:
+		p += 1;
+		break;
+
+	case 2:
+		p += 4;
+		break;
+
+	case 3:
+	default:
+		printk(KERN_ERR "mmiotrace: not a memory access instruction "
+						"at 0x%lx, rm_mod=0x%02x\n",
+						ins_addr, mod_rm);
+	}
+
+	switch (get_ins_reg_width(ins_addr)) {
+	case 1:
+		return *(unsigned char *)p;
+
+	case 2:
+		return *(unsigned short *)p;
+
+	case 4:
+		return *(unsigned int *)p;
+
+#ifdef __amd64__
+	case 8:
+		return *(unsigned long *)p;
+#endif
+
+	default:
+		printk(KERN_ERR "mmiotrace: Error: width.\n");
+	}
+
+err:
+	return 0;
+}

arch/x86/mm/pf_in.h

@@ -0,0 +1,39 @@
+/*
+ *  Fault Injection Test harness (FI)
+ *  Copyright (C) Intel Crop.
+ *
+ *  This program is free software; you can redistribute it and/or
+ *  modify it under the terms of the GNU General Public License
+ *  as published by the Free Software Foundation; either version 2
+ *  of the License, or (at your option) any later version.
+ *
+ *  This program is distributed in the hope that it will be useful,
+ *  but WITHOUT ANY WARRANTY; without even the implied warranty of
+ *  MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ *  GNU General Public License for more details.
+ *
+ *  You should have received a copy of the GNU General Public License
+ *  along with this program; if not, write to the Free Software
+ *  Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA  02111-1307,
+ *  USA.
+ *
+ */
+
+#ifndef __PF_H_
+#define __PF_H_
+
+enum reason_type {
+	NOT_ME,	/* page fault is not in regions */
+	NOTHING,	/* access others point in regions */
+	REG_READ,	/* read from addr to reg */
+	REG_WRITE,	/* write from reg to addr */
+	IMM_WRITE,	/* write from imm to addr */
+	OTHERS	/* Other instructions can not intercept */
+};
+
+enum reason_type get_ins_type(unsigned long ins_addr);
+unsigned int get_ins_mem_width(unsigned long ins_addr);
+unsigned long get_ins_reg_val(unsigned long ins_addr, struct pt_regs *regs);
+unsigned long get_ins_imm_val(unsigned long ins_addr);
+
+#endif /* __PF_H_ */

arch/x86/mm/testmmiotrace.c

@@ -0,0 +1,71 @@
+/*
+ * Written by Pekka Paalanen, 2008 <pq@iki.fi>
+ */
+#include <linux/module.h>
+#include <linux/io.h>
+
+#define MODULE_NAME "testmmiotrace"
+
+static unsigned long mmio_address;
+module_param(mmio_address, ulong, 0);
+MODULE_PARM_DESC(mmio_address, "Start address of the mapping of 16 kB.");
+
+static void do_write_test(void __iomem *p)
+{
+	unsigned int i;
+	for (i = 0; i < 256; i++)
+		iowrite8(i, p + i);
+	for (i = 1024; i < (5 * 1024); i += 2)
+		iowrite16(i * 12 + 7, p + i);
+	for (i = (5 * 1024); i < (16 * 1024); i += 4)
+		iowrite32(i * 212371 + 13, p + i);
+}
+
+static void do_read_test(void __iomem *p)
+{
+	unsigned int i;
+	for (i = 0; i < 256; i++)
+		ioread8(p + i);
+	for (i = 1024; i < (5 * 1024); i += 2)
+		ioread16(p + i);
+	for (i = (5 * 1024); i < (16 * 1024); i += 4)
+		ioread32(p + i);
+}
+
+static void do_test(void)
+{
+	void __iomem *p = ioremap_nocache(mmio_address, 0x4000);
+	if (!p) {
+		pr_err(MODULE_NAME ": could not ioremap, aborting.\n");
+		return;
+	}
+	do_write_test(p);
+	do_read_test(p);
+	iounmap(p);
+}
+
+static int __init init(void)
+{
+	if (mmio_address == 0) {
+		pr_err(MODULE_NAME ": you have to use the module argument "
+							"mmio_address.\n");
+		pr_err(MODULE_NAME ": DO NOT LOAD THIS MODULE UNLESS"
+				" YOU REALLY KNOW WHAT YOU ARE DOING!\n");
+		return -ENXIO;
+	}
+
+	pr_warning(MODULE_NAME ": WARNING: mapping 16 kB @ 0x%08lx "
+					"in PCI address space, and writing "
+					"rubbish in there.\n", mmio_address);
+	do_test();
+	return 0;
+}
+
+static void __exit cleanup(void)
+{
+	pr_debug(MODULE_NAME ": unloaded.\n");
+}
+
+module_init(init);
+module_exit(cleanup);
+MODULE_LICENSE("GPL");

arch/x86/vdso/vclock_gettime.c

@@ -23,7 +23,7 @@
 
 #define gtod vdso_vsyscall_gtod_data
 
-static long vdso_fallback_gettime(long clock, struct timespec *ts)
+notrace static long vdso_fallback_gettime(long clock, struct timespec *ts)
 {
 	long ret;
 	asm("syscall" : "=a" (ret) :
@@ -31,7 +31,7 @@ static long vdso_fallback_gettime(long clock, struct timespec *ts)
 	return ret;
 }
 
-static inline long vgetns(void)
+notrace static inline long vgetns(void)
 {
 	long v;
 	cycles_t (*vread)(void);
@@ -40,7 +40,7 @@ static inline long vgetns(void)
 	return (v * gtod->clock.mult) >> gtod->clock.shift;
 }
 
-static noinline int do_realtime(struct timespec *ts)
+notrace static noinline int do_realtime(struct timespec *ts)
 {
 	unsigned long seq, ns;
 	do {
@@ -54,7 +54,8 @@ static noinline int do_realtime(struct timespec *ts)
 }
 
 /* Copy of the version in kernel/time.c which we cannot directly access */
-static void vset_normalized_timespec(struct timespec *ts, long sec, long nsec)
+notrace static void
+vset_normalized_timespec(struct timespec *ts, long sec, long nsec)
 {
 	while (nsec >= NSEC_PER_SEC) {
 		nsec -= NSEC_PER_SEC;
@@ -68,7 +69,7 @@ static void vset_normalized_timespec(struct timespec *ts, long sec, long nsec)
 	ts->tv_nsec = nsec;
 }
 
-static noinline int do_monotonic(struct timespec *ts)
+notrace static noinline int do_monotonic(struct timespec *ts)
 {
 	unsigned long seq, ns, secs;
 	do {
@@ -82,7 +83,7 @@ static noinline int do_monotonic(struct timespec *ts)
 	return 0;
 }
 
-int __vdso_clock_gettime(clockid_t clock, struct timespec *ts)
+notrace int __vdso_clock_gettime(clockid_t clock, struct timespec *ts)
 {
 	if (likely(gtod->sysctl_enabled && gtod->clock.vread))
 		switch (clock) {
@@ -96,7 +97,7 @@ int __vdso_clock_gettime(clockid_t clock, struct timespec *ts)
 int clock_gettime(clockid_t, struct timespec *)
 	__attribute__((weak, alias("__vdso_clock_gettime")));
 
-int __vdso_gettimeofday(struct timeval *tv, struct timezone *tz)
+notrace int __vdso_gettimeofday(struct timeval *tv, struct timezone *tz)
 {
 	long ret;
 	if (likely(gtod->sysctl_enabled && gtod->clock.vread)) {

arch/x86/vdso/vgetcpu.c

@@ -13,7 +13,8 @@
 #include <asm/vgtod.h>
 #include "vextern.h"
 
-long __vdso_getcpu(unsigned *cpu, unsigned *node, struct getcpu_cache *unused)
+notrace long
+__vdso_getcpu(unsigned *cpu, unsigned *node, struct getcpu_cache *unused)
 {
 	unsigned int p;
 

include/asm-arm/ftrace.h

@@ -0,0 +1,14 @@
+#ifndef _ASM_ARM_FTRACE
+#define _ASM_ARM_FTRACE
+
+#ifdef CONFIG_FTRACE
+#define MCOUNT_ADDR		((long)(mcount))
+#define MCOUNT_INSN_SIZE	4 /* sizeof mcount call */
+
+#ifndef __ASSEMBLY__
+extern void mcount(void);
+#endif
+
+#endif
+
+#endif /* _ASM_ARM_FTRACE */

include/asm-arm/kprobes.h

@@ -59,6 +59,7 @@ struct kprobe_ctlblk {
 };
 
 void arch_remove_kprobe(struct kprobe *);
+void kretprobe_trampoline(void);
 
 int kprobe_trap_handler(struct pt_regs *regs, unsigned int instr);
 int kprobe_fault_handler(struct pt_regs *regs, unsigned int fsr);

include/asm-powerpc/ftrace.h

@@ -0,0 +1,14 @@
+#ifndef _ASM_POWERPC_FTRACE
+#define _ASM_POWERPC_FTRACE
+
+#ifdef CONFIG_FTRACE
+#define MCOUNT_ADDR		((long)(_mcount))
+#define MCOUNT_INSN_SIZE	4 /* sizeof mcount call */
+
+#ifndef __ASSEMBLY__
+extern void _mcount(void);
+#endif
+
+#endif
+
+#endif /* _ASM_POWERPC_FTRACE */

include/asm-powerpc/hw_irq.h

@@ -59,6 +59,11 @@ extern void iseries_handle_interrupts(void);
 		get_paca()->hard_enabled = 0;	\
 	} while(0)
 
+static inline int irqs_disabled_flags(unsigned long flags)
+{
+	return flags == 0;
+}
+
 #else
 
 #if defined(CONFIG_BOOKE)
@@ -113,6 +118,11 @@ static inline void local_irq_save_ptr(unsigned long *flags)
 #define hard_irq_enable()	local_irq_enable()
 #define hard_irq_disable()	local_irq_disable()
 
+static inline int irqs_disabled_flags(unsigned long flags)
+{
+	return (flags & MSR_EE) == 0;
+}
+
 #endif /* CONFIG_PPC64 */
 
 /*

include/asm-sparc64/ftrace.h

@@ -0,0 +1,14 @@
+#ifndef _ASM_SPARC64_FTRACE
+#define _ASM_SPARC64_FTRACE
+
+#ifdef CONFIG_MCOUNT
+#define MCOUNT_ADDR		((long)(_mcount))
+#define MCOUNT_INSN_SIZE	4 /* sizeof mcount call */
+
+#ifndef __ASSEMBLY__
+extern void _mcount(void);
+#endif
+
+#endif
+
+#endif /* _ASM_SPARC64_FTRACE */

include/asm-x86/alternative.h

@@ -72,6 +72,8 @@ static inline void alternatives_smp_module_del(struct module *mod) {}
 static inline void alternatives_smp_switch(int smp) {}
 #endif	/* CONFIG_SMP */
 
+const unsigned char *const *find_nop_table(void);
+
 /*
  * Alternative instructions for different CPU types or capabilities.
  *

include/asm-x86/ftrace.h

@@ -0,0 +1,14 @@
+#ifndef _ASM_X86_FTRACE
+#define _ASM_SPARC64_FTRACE
+
+#ifdef CONFIG_FTRACE
+#define MCOUNT_ADDR		((long)(mcount))
+#define MCOUNT_INSN_SIZE	5 /* sizeof mcount call */
+
+#ifndef __ASSEMBLY__
+extern void mcount(void);
+#endif
+
+#endif /* CONFIG_FTRACE */
+
+#endif /* _ASM_X86_FTRACE */

include/asm-x86/irqflags.h

@@ -190,8 +190,6 @@ static inline void trace_hardirqs_fixup(void)
 #else
 
 #ifdef CONFIG_X86_64
-#define ARCH_TRACE_IRQS_ON		call trace_hardirqs_on_thunk
-#define ARCH_TRACE_IRQS_OFF		call trace_hardirqs_off_thunk
 #define ARCH_LOCKDEP_SYS_EXIT		call lockdep_sys_exit_thunk
 #define ARCH_LOCKDEP_SYS_EXIT_IRQ	\
 	TRACE_IRQS_ON; \
@@ -203,24 +201,6 @@ static inline void trace_hardirqs_fixup(void)
 	TRACE_IRQS_OFF;
 
 #else
-#define ARCH_TRACE_IRQS_ON			\
-	pushl %eax;				\
-	pushl %ecx;				\
-	pushl %edx;				\
-	call trace_hardirqs_on;			\
-	popl %edx;				\
-	popl %ecx;				\
-	popl %eax;
-
-#define ARCH_TRACE_IRQS_OFF			\
-	pushl %eax;				\
-	pushl %ecx;				\
-	pushl %edx;				\
-	call trace_hardirqs_off;		\
-	popl %edx;				\
-	popl %ecx;				\
-	popl %eax;
-
 #define ARCH_LOCKDEP_SYS_EXIT			\
 	pushl %eax;				\
 	pushl %ecx;				\
@@ -234,8 +214,8 @@ static inline void trace_hardirqs_fixup(void)
 #endif
 
 #ifdef CONFIG_TRACE_IRQFLAGS
-#  define TRACE_IRQS_ON		ARCH_TRACE_IRQS_ON
-#  define TRACE_IRQS_OFF	ARCH_TRACE_IRQS_OFF
+#  define TRACE_IRQS_ON		call trace_hardirqs_on_thunk;
+#  define TRACE_IRQS_OFF	call trace_hardirqs_off_thunk;
 #else
 #  define TRACE_IRQS_ON
 #  define TRACE_IRQS_OFF

include/asm-x86/vsyscall.h

@@ -24,7 +24,8 @@ enum vsyscall_num {
 	((unused, __section__ (".vsyscall_gtod_data"),aligned(16)))
 #define __section_vsyscall_clock __attribute__ \
 	((unused, __section__ (".vsyscall_clock"),aligned(16)))
-#define __vsyscall_fn __attribute__ ((unused,__section__(".vsyscall_fn")))
+#define __vsyscall_fn \
+	__attribute__ ((unused, __section__(".vsyscall_fn"))) notrace
 
 #define VGETCPU_RDTSCP	1
 #define VGETCPU_LSL	2

include/linux/ftrace.h

@@ -0,0 +1,144 @@
+#ifndef _LINUX_FTRACE_H
+#define _LINUX_FTRACE_H
+
+#ifdef CONFIG_FTRACE
+
+#include <linux/linkage.h>
+#include <linux/fs.h>
+
+extern int ftrace_enabled;
+extern int
+ftrace_enable_sysctl(struct ctl_table *table, int write,
+		     struct file *filp, void __user *buffer, size_t *lenp,
+		     loff_t *ppos);
+
+typedef void (*ftrace_func_t)(unsigned long ip, unsigned long parent_ip);
+
+struct ftrace_ops {
+	ftrace_func_t	  func;
+	struct ftrace_ops *next;
+};
+
+/*
+ * The ftrace_ops must be a static and should also
+ * be read_mostly.  These functions do modify read_mostly variables
+ * so use them sparely. Never free an ftrace_op or modify the
+ * next pointer after it has been registered. Even after unregistering
+ * it, the next pointer may still be used internally.
+ */
+int register_ftrace_function(struct ftrace_ops *ops);
+int unregister_ftrace_function(struct ftrace_ops *ops);
+void clear_ftrace_function(void);
+
+extern void ftrace_stub(unsigned long a0, unsigned long a1);
+
+#else /* !CONFIG_FTRACE */
+# define register_ftrace_function(ops) do { } while (0)
+# define unregister_ftrace_function(ops) do { } while (0)
+# define clear_ftrace_function(ops) do { } while (0)
+#endif /* CONFIG_FTRACE */
+
+#ifdef CONFIG_DYNAMIC_FTRACE
+# define FTRACE_HASHBITS	10
+# define FTRACE_HASHSIZE	(1<<FTRACE_HASHBITS)
+
+enum {
+	FTRACE_FL_FREE		= (1 << 0),
+	FTRACE_FL_FAILED	= (1 << 1),
+	FTRACE_FL_FILTER	= (1 << 2),
+	FTRACE_FL_ENABLED	= (1 << 3),
+	FTRACE_FL_NOTRACE	= (1 << 4),
+	FTRACE_FL_CONVERTED	= (1 << 5),
+	FTRACE_FL_FROZEN	= (1 << 6),
+};
+
+struct dyn_ftrace {
+	struct hlist_node node;
+	unsigned long	  ip; /* address of mcount call-site */
+	unsigned long	  flags;
+};
+
+int ftrace_force_update(void);
+void ftrace_set_filter(unsigned char *buf, int len, int reset);
+
+/* defined in arch */
+extern int ftrace_ip_converted(unsigned long ip);
+extern unsigned char *ftrace_nop_replace(void);
+extern unsigned char *ftrace_call_replace(unsigned long ip, unsigned long addr);
+extern int ftrace_dyn_arch_init(void *data);
+extern int ftrace_mcount_set(unsigned long *data);
+extern int ftrace_modify_code(unsigned long ip, unsigned char *old_code,
+			      unsigned char *new_code);
+extern int ftrace_update_ftrace_func(ftrace_func_t func);
+extern void ftrace_caller(void);
+extern void ftrace_call(void);
+extern void mcount_call(void);
+
+extern int skip_trace(unsigned long ip);
+
+void ftrace_disable_daemon(void);
+void ftrace_enable_daemon(void);
+
+#else
+# define skip_trace(ip)				({ 0; })
+# define ftrace_force_update()			({ 0; })
+# define ftrace_set_filter(buf, len, reset)	do { } while (0)
+# define ftrace_disable_daemon()		do { } while (0)
+# define ftrace_enable_daemon()			do { } while (0)
+#endif /* CONFIG_DYNAMIC_FTRACE */
+
+/* totally disable ftrace - can not re-enable after this */
+void ftrace_kill(void);
+void ftrace_kill_atomic(void);
+
+static inline void tracer_disable(void)
+{
+#ifdef CONFIG_FTRACE
+	ftrace_enabled = 0;
+#endif
+}
+
+#ifdef CONFIG_FRAME_POINTER
+/* TODO: need to fix this for ARM */
+# define CALLER_ADDR0 ((unsigned long)__builtin_return_address(0))
+# define CALLER_ADDR1 ((unsigned long)__builtin_return_address(1))
+# define CALLER_ADDR2 ((unsigned long)__builtin_return_address(2))
+# define CALLER_ADDR3 ((unsigned long)__builtin_return_address(3))
+# define CALLER_ADDR4 ((unsigned long)__builtin_return_address(4))
+# define CALLER_ADDR5 ((unsigned long)__builtin_return_address(5))
+# define CALLER_ADDR6 ((unsigned long)__builtin_return_address(6))
+#else
+# define CALLER_ADDR0 ((unsigned long)__builtin_return_address(0))
+# define CALLER_ADDR1 0UL
+# define CALLER_ADDR2 0UL
+# define CALLER_ADDR3 0UL
+# define CALLER_ADDR4 0UL
+# define CALLER_ADDR5 0UL
+# define CALLER_ADDR6 0UL
+#endif
+
+#ifdef CONFIG_IRQSOFF_TRACER
+  extern void time_hardirqs_on(unsigned long a0, unsigned long a1);
+  extern void time_hardirqs_off(unsigned long a0, unsigned long a1);
+#else
+# define time_hardirqs_on(a0, a1)		do { } while (0)
+# define time_hardirqs_off(a0, a1)		do { } while (0)
+#endif
+
+#ifdef CONFIG_PREEMPT_TRACER
+  extern void trace_preempt_on(unsigned long a0, unsigned long a1);
+  extern void trace_preempt_off(unsigned long a0, unsigned long a1);
+#else
+# define trace_preempt_on(a0, a1)		do { } while (0)
+# define trace_preempt_off(a0, a1)		do { } while (0)
+#endif
+
+#ifdef CONFIG_TRACING
+extern void
+ftrace_special(unsigned long arg1, unsigned long arg2, unsigned long arg3);
+#else
+static inline void
+ftrace_special(unsigned long arg1, unsigned long arg2, unsigned long arg3) { }
+#endif
+
+#endif /* _LINUX_FTRACE_H */

include/linux/irqflags.h

@@ -12,10 +12,10 @@
 #define _LINUX_TRACE_IRQFLAGS_H
 
 #ifdef CONFIG_TRACE_IRQFLAGS
-  extern void trace_hardirqs_on(void);
-  extern void trace_hardirqs_off(void);
   extern void trace_softirqs_on(unsigned long ip);
   extern void trace_softirqs_off(unsigned long ip);
+  extern void trace_hardirqs_on(void);
+  extern void trace_hardirqs_off(void);
 # define trace_hardirq_context(p)	((p)->hardirq_context)
 # define trace_softirq_context(p)	((p)->softirq_context)
 # define trace_hardirqs_enabled(p)	((p)->hardirqs_enabled)
@@ -41,6 +41,15 @@
 # define INIT_TRACE_IRQFLAGS
 #endif
 
+#if defined(CONFIG_IRQSOFF_TRACER) || \
+	defined(CONFIG_PREEMPT_TRACER)
+ extern void stop_critical_timings(void);
+ extern void start_critical_timings(void);
+#else
+# define stop_critical_timings() do { } while (0)
+# define start_critical_timings() do { } while (0)
+#endif
+
 #ifdef CONFIG_TRACE_IRQFLAGS_SUPPORT
 
 #include <asm/irqflags.h>

include/linux/kprobes.h

@@ -259,6 +259,10 @@ void recycle_rp_inst(struct kretprobe_instance *ri, struct hlist_head *head);
 struct jprobe;
 struct kretprobe;
 
+static inline struct kprobe *get_kprobe(void *addr)
+{
+	return NULL;
+}
 static inline struct kprobe *kprobe_running(void)
 {
 	return NULL;

include/linux/linkage.h

@@ -4,6 +4,8 @@
 #include <linux/compiler.h>
 #include <asm/linkage.h>
 
+#define notrace __attribute__((no_instrument_function))
+
 #ifdef __cplusplus
 #define CPP_ASMLINKAGE extern "C"
 #else

include/linux/marker.h

@@ -44,8 +44,8 @@ struct marker {
 				 */
 	char state;		/* Marker state. */
 	char ptype;		/* probe type : 0 : single, 1 : multi */
-	void (*call)(const struct marker *mdata,	/* Probe wrapper */
-		void *call_private, const char *fmt, ...);
+				/* Probe wrapper */
+	void (*call)(const struct marker *mdata, void *call_private, ...);
 	struct marker_probe_closure single;
 	struct marker_probe_closure *multi;
 } __attribute__((aligned(8)));
@@ -58,8 +58,12 @@ struct marker {
  * Make sure the alignment of the structure in the __markers section will
  * not add unwanted padding between the beginning of the section and the
  * structure. Force alignment to the same alignment as the section start.
+ *
+ * The "generic" argument controls which marker enabling mechanism must be used.
+ * If generic is true, a variable read is used.
+ * If generic is false, immediate values are used.
  */
-#define __trace_mark(name, call_private, format, args...)		\
+#define __trace_mark(generic, name, call_private, format, args...)	\
 	do {								\
 		static const char __mstrtab_##name[]			\
 		__attribute__((section("__markers_strings")))		\
@@ -72,15 +76,14 @@ struct marker {
 		__mark_check_format(format, ## args);			\
 		if (unlikely(__mark_##name.state)) {			\
 			(*__mark_##name.call)				\
-				(&__mark_##name, call_private,		\
-				format, ## args);			\
+				(&__mark_##name, call_private, ## args);\
 		}							\
 	} while (0)
 
 extern void marker_update_probe_range(struct marker *begin,
 	struct marker *end);
 #else /* !CONFIG_MARKERS */
-#define __trace_mark(name, call_private, format, args...) \
+#define __trace_mark(generic, name, call_private, format, args...) \
 		__mark_check_format(format, ## args)
 static inline void marker_update_probe_range(struct marker *begin,
 	struct marker *end)
@@ -88,15 +91,30 @@ static inline void marker_update_probe_range(struct marker *begin,
 #endif /* CONFIG_MARKERS */
 
 /**
- * trace_mark - Marker
+ * trace_mark - Marker using code patching
  * @name: marker name, not quoted.
  * @format: format string
  * @args...: variable argument list
  *
- * Places a marker.
+ * Places a marker using optimized code patching technique (imv_read())
+ * to be enabled when immediate values are present.
  */
 #define trace_mark(name, format, args...) \
-	__trace_mark(name, NULL, format, ## args)
+	__trace_mark(0, name, NULL, format, ## args)
+
+/**
+ * _trace_mark - Marker using variable read
+ * @name: marker name, not quoted.
+ * @format: format string
+ * @args...: variable argument list
+ *
+ * Places a marker using a standard memory read (_imv_read()) to be
+ * enabled. Should be used for markers in code paths where instruction
+ * modification based enabling is not welcome. (__init and __exit functions,
+ * lockdep, some traps, printk).
+ */
+#define _trace_mark(name, format, args...) \
+	__trace_mark(1, name, NULL, format, ## args)
 
 /**
  * MARK_NOARGS - Format string for a marker with no argument.
@@ -117,9 +135,9 @@ static inline void __printf(1, 2) ___mark_check_format(const char *fmt, ...)
 extern marker_probe_func __mark_empty_function;
 
 extern void marker_probe_cb(const struct marker *mdata,
-	void *call_private, const char *fmt, ...);
+	void *call_private, ...);
 extern void marker_probe_cb_noarg(const struct marker *mdata,
-	void *call_private, const char *fmt, ...);
+	void *call_private, ...);
 
 /*
  * Connect a probe to a marker.

include/linux/mmiotrace.h

@@ -0,0 +1,85 @@
+#ifndef MMIOTRACE_H
+#define MMIOTRACE_H
+
+#include <linux/types.h>
+#include <linux/list.h>
+
+struct kmmio_probe;
+struct pt_regs;
+
+typedef void (*kmmio_pre_handler_t)(struct kmmio_probe *,
+				struct pt_regs *, unsigned long addr);
+typedef void (*kmmio_post_handler_t)(struct kmmio_probe *,
+				unsigned long condition, struct pt_regs *);
+
+struct kmmio_probe {
+	struct list_head list; /* kmmio internal list */
+	unsigned long addr; /* start location of the probe point */
+	unsigned long len; /* length of the probe region */
+	kmmio_pre_handler_t pre_handler; /* Called before addr is executed. */
+	kmmio_post_handler_t post_handler; /* Called after addr is executed */
+	void *private;
+};
+
+/* kmmio is active by some kmmio_probes? */
+static inline int is_kmmio_active(void)
+{
+	extern unsigned int kmmio_count;
+	return kmmio_count;
+}
+
+extern int register_kmmio_probe(struct kmmio_probe *p);
+extern void unregister_kmmio_probe(struct kmmio_probe *p);
+
+/* Called from page fault handler. */
+extern int kmmio_handler(struct pt_regs *regs, unsigned long addr);
+
+/* Called from ioremap.c */
+#ifdef CONFIG_MMIOTRACE
+extern void mmiotrace_ioremap(resource_size_t offset, unsigned long size,
+							void __iomem *addr);
+extern void mmiotrace_iounmap(volatile void __iomem *addr);
+#else
+static inline void mmiotrace_ioremap(resource_size_t offset,
+					unsigned long size, void __iomem *addr)
+{
+}
+
+static inline void mmiotrace_iounmap(volatile void __iomem *addr)
+{
+}
+#endif /* CONFIG_MMIOTRACE_HOOKS */
+
+enum mm_io_opcode {
+	MMIO_READ = 0x1,     /* struct mmiotrace_rw */
+	MMIO_WRITE = 0x2,    /* struct mmiotrace_rw */
+	MMIO_PROBE = 0x3,    /* struct mmiotrace_map */
+	MMIO_UNPROBE = 0x4,  /* struct mmiotrace_map */
+	MMIO_MARKER = 0x5,   /* raw char data */
+	MMIO_UNKNOWN_OP = 0x6, /* struct mmiotrace_rw */
+};
+
+struct mmiotrace_rw {
+	resource_size_t phys;	/* PCI address of register */
+	unsigned long value;
+	unsigned long pc;	/* optional program counter */
+	int map_id;
+	unsigned char opcode;	/* one of MMIO_{READ,WRITE,UNKNOWN_OP} */
+	unsigned char width;	/* size of register access in bytes */
+};
+
+struct mmiotrace_map {
+	resource_size_t phys;	/* base address in PCI space */
+	unsigned long virt;	/* base virtual address */
+	unsigned long len;	/* mapping size */
+	int map_id;
+	unsigned char opcode;	/* MMIO_PROBE or MMIO_UNPROBE */
+};
+
+/* in kernel/trace/trace_mmiotrace.c */
+extern void enable_mmiotrace(void);
+extern void disable_mmiotrace(void);
+extern void mmio_trace_rw(struct mmiotrace_rw *rw);
+extern void mmio_trace_mapping(struct mmiotrace_map *map);
+
+#endif /* MMIOTRACE_H */

include/linux/preempt.h

@@ -10,7 +10,7 @@
 #include <linux/linkage.h>
 #include <linux/list.h>
 
-#ifdef CONFIG_DEBUG_PREEMPT
+#if defined(CONFIG_DEBUG_PREEMPT) || defined(CONFIG_PREEMPT_TRACER)
   extern void add_preempt_count(int val);
   extern void sub_preempt_count(int val);
 #else
@@ -52,6 +52,34 @@ do { \
 	preempt_check_resched(); \
 } while (0)
 
+/* For debugging and tracer internals only! */
+#define add_preempt_count_notrace(val)			\
+	do { preempt_count() += (val); } while (0)
+#define sub_preempt_count_notrace(val)			\
+	do { preempt_count() -= (val); } while (0)
+#define inc_preempt_count_notrace() add_preempt_count_notrace(1)
+#define dec_preempt_count_notrace() sub_preempt_count_notrace(1)
+
+#define preempt_disable_notrace() \
+do { \
+	inc_preempt_count_notrace(); \
+	barrier(); \
+} while (0)
+
+#define preempt_enable_no_resched_notrace() \
+do { \
+	barrier(); \
+	dec_preempt_count_notrace(); \
+} while (0)
+
+/* preempt_check_resched is OK to trace */
+#define preempt_enable_notrace() \
+do { \
+	preempt_enable_no_resched_notrace(); \
+	barrier(); \
+	preempt_check_resched(); \
+} while (0)
+
 #else
 
 #define preempt_disable()		do { } while (0)
@@ -59,6 +87,10 @@ do { \
 #define preempt_enable()		do { } while (0)
 #define preempt_check_resched()		do { } while (0)
 
+#define preempt_disable_notrace()		do { } while (0)
+#define preempt_enable_no_resched_notrace()	do { } while (0)
+#define preempt_enable_notrace()		do { } while (0)
+
 #endif
 
 #ifdef CONFIG_PREEMPT_NOTIFIERS

include/linux/sched.h

@@ -245,6 +245,8 @@ extern asmlinkage void schedule_tail(struct task_struct *prev);
 extern void init_idle(struct task_struct *idle, int cpu);
 extern void init_idle_bootup_task(struct task_struct *idle);
 
+extern int runqueue_is_locked(void);
+
 extern cpumask_t nohz_cpu_mask;
 #if defined(CONFIG_SMP) && defined(CONFIG_NO_HZ)
 extern int select_nohz_load_balancer(int cpu);
@@ -2132,6 +2134,18 @@ static inline void arch_pick_mmap_layout(struct mm_struct *mm)
 }
 #endif
 
+#ifdef CONFIG_TRACING
+extern void
+__trace_special(void *__tr, void *__data,
+		unsigned long arg1, unsigned long arg2, unsigned long arg3);
+#else
+static inline void
+__trace_special(void *__tr, void *__data,
+		unsigned long arg1, unsigned long arg2, unsigned long arg3)
+{
+}
+#endif
+
 extern long sched_setaffinity(pid_t pid, const cpumask_t *new_mask);
 extern long sched_getaffinity(pid_t pid, cpumask_t *mask);
 
@@ -2226,6 +2240,8 @@ static inline void mm_init_owner(struct mm_struct *mm, struct task_struct *p)
 }
 #endif /* CONFIG_MM_OWNER */
 
+#define TASK_STATE_TO_CHAR_STR "RSDTtZX"
+
 #endif /* __KERNEL__ */
 
 #endif

include/linux/writeback.h

@@ -105,6 +105,8 @@ extern int vm_highmem_is_dirtyable;
 extern int block_dump;
 extern int laptop_mode;
 
+extern unsigned long determine_dirtyable_memory(void);
+
 extern int dirty_ratio_handler(struct ctl_table *table, int write,
 		struct file *filp, void __user *buffer, size_t *lenp,
 		loff_t *ppos);

kernel/Makefile

@@ -11,6 +11,18 @@ obj-y     = sched.o fork.o exec_domain.o panic.o printk.o profile.o \
 	    hrtimer.o rwsem.o nsproxy.o srcu.o semaphore.o \
 	    notifier.o ksysfs.o pm_qos_params.o sched_clock.o
 
+CFLAGS_REMOVE_sched.o = -mno-spe
+
+ifdef CONFIG_FTRACE
+# Do not trace debug files and internal ftrace files
+CFLAGS_REMOVE_lockdep.o = -pg
+CFLAGS_REMOVE_lockdep_proc.o = -pg
+CFLAGS_REMOVE_mutex-debug.o = -pg
+CFLAGS_REMOVE_rtmutex-debug.o = -pg
+CFLAGS_REMOVE_cgroup-debug.o = -pg
+CFLAGS_REMOVE_sched_clock.o = -pg
+endif
+
 obj-$(CONFIG_SYSCTL_SYSCALL_CHECK) += sysctl_check.o
 obj-$(CONFIG_STACKTRACE) += stacktrace.o
 obj-y += time/
@@ -69,6 +81,8 @@ obj-$(CONFIG_TASK_DELAY_ACCT) += delayacct.o
 obj-$(CONFIG_TASKSTATS) += taskstats.o tsacct.o
 obj-$(CONFIG_MARKERS) += marker.o
 obj-$(CONFIG_LATENCYTOP) += latencytop.o
+obj-$(CONFIG_FTRACE) += trace/
+obj-$(CONFIG_TRACING) += trace/
 obj-$(CONFIG_SMP) += sched_cpupri.o
 
 ifneq ($(CONFIG_SCHED_NO_NO_OMIT_FRAME_POINTER),y)

kernel/fork.c

@@ -910,7 +910,7 @@ static struct task_struct *copy_process(unsigned long clone_flags,
 
 	rt_mutex_init_task(p);
 
-#ifdef CONFIG_TRACE_IRQFLAGS
+#if defined(CONFIG_TRACE_IRQFLAGS) && defined(CONFIG_LOCKDEP)
 	DEBUG_LOCKS_WARN_ON(!p->hardirqs_enabled);
 	DEBUG_LOCKS_WARN_ON(!p->softirqs_enabled);
 #endif

kernel/lockdep.c

@@ -39,6 +39,7 @@
 #include <linux/irqflags.h>
 #include <linux/utsname.h>
 #include <linux/hash.h>
+#include <linux/ftrace.h>
 
 #include <asm/sections.h>
 
@@ -81,6 +82,8 @@ static int graph_lock(void)
 		__raw_spin_unlock(&lockdep_lock);
 		return 0;
 	}
+	/* prevent any recursions within lockdep from causing deadlocks */
+	current->lockdep_recursion++;
 	return 1;
 }
 
@@ -89,6 +92,7 @@ static inline int graph_unlock(void)
 	if (debug_locks && !__raw_spin_is_locked(&lockdep_lock))
 		return DEBUG_LOCKS_WARN_ON(1);
 
+	current->lockdep_recursion--;
 	__raw_spin_unlock(&lockdep_lock);
 	return 0;
 }
@@ -982,7 +986,7 @@ check_noncircular(struct lock_class *source, unsigned int depth)
 	return 1;
 }
 
-#ifdef CONFIG_TRACE_IRQFLAGS
+#if defined(CONFIG_TRACE_IRQFLAGS) && defined(CONFIG_PROVE_LOCKING)
 /*
  * Forwards and backwards subgraph searching, for the purposes of
  * proving that two subgraphs can be connected by a new dependency
@@ -1680,7 +1684,7 @@ valid_state(struct task_struct *curr, struct held_lock *this,
 static int mark_lock(struct task_struct *curr, struct held_lock *this,
 		     enum lock_usage_bit new_bit);
 
-#ifdef CONFIG_TRACE_IRQFLAGS
+#if defined(CONFIG_TRACE_IRQFLAGS) && defined(CONFIG_PROVE_LOCKING)
 
 /*
  * print irq inversion bug:
@@ -2013,11 +2017,13 @@ void early_boot_irqs_on(void)
 /*
  * Hardirqs will be enabled:
  */
-void trace_hardirqs_on(void)
+void trace_hardirqs_on_caller(unsigned long a0)
 {
 	struct task_struct *curr = current;
 	unsigned long ip;
 
+	time_hardirqs_on(CALLER_ADDR0, a0);
+
 	if (unlikely(!debug_locks || current->lockdep_recursion))
 		return;
 
@@ -2055,16 +2061,23 @@ void trace_hardirqs_on(void)
 	curr->hardirq_enable_event = ++curr->irq_events;
 	debug_atomic_inc(&hardirqs_on_events);
 }
+EXPORT_SYMBOL(trace_hardirqs_on_caller);
 
+void trace_hardirqs_on(void)
+{
+	trace_hardirqs_on_caller(CALLER_ADDR0);
+}
 EXPORT_SYMBOL(trace_hardirqs_on);
 
 /*
  * Hardirqs were disabled:
  */
-void trace_hardirqs_off(void)
+void trace_hardirqs_off_caller(unsigned long a0)
 {
 	struct task_struct *curr = current;
 
+	time_hardirqs_off(CALLER_ADDR0, a0);
+
 	if (unlikely(!debug_locks || current->lockdep_recursion))
 		return;
 
@@ -2082,7 +2095,12 @@ void trace_hardirqs_off(void)
 	} else
 		debug_atomic_inc(&redundant_hardirqs_off);
 }
+EXPORT_SYMBOL(trace_hardirqs_off_caller);
 
+void trace_hardirqs_off(void)
+{
+	trace_hardirqs_off_caller(CALLER_ADDR0);
+}
 EXPORT_SYMBOL(trace_hardirqs_off);
 
 /*
@@ -2246,7 +2264,7 @@ static inline int separate_irq_context(struct task_struct *curr,
  * Mark a lock with a usage bit, and validate the state transition:
  */
 static int mark_lock(struct task_struct *curr, struct held_lock *this,
-		     enum lock_usage_bit new_bit)
+			     enum lock_usage_bit new_bit)
 {
 	unsigned int new_mask = 1 << new_bit, ret = 1;
 
@@ -2686,7 +2704,7 @@ static void check_flags(unsigned long flags)
  * and also avoid lockdep recursion:
  */
 void lock_acquire(struct lockdep_map *lock, unsigned int subclass,
-		  int trylock, int read, int check, unsigned long ip)
+			  int trylock, int read, int check, unsigned long ip)
 {
 	unsigned long flags;
 
@@ -2708,7 +2726,8 @@ void lock_acquire(struct lockdep_map *lock, unsigned int subclass,
 
 EXPORT_SYMBOL_GPL(lock_acquire);
 
-void lock_release(struct lockdep_map *lock, int nested, unsigned long ip)
+void lock_release(struct lockdep_map *lock, int nested,
+			  unsigned long ip)
 {
 	unsigned long flags;
 

kernel/marker.c

@@ -55,8 +55,8 @@ static DEFINE_MUTEX(markers_mutex);
 struct marker_entry {
 	struct hlist_node hlist;
 	char *format;
-	void (*call)(const struct marker *mdata,	/* Probe wrapper */
-		void *call_private, const char *fmt, ...);
+			/* Probe wrapper */
+	void (*call)(const struct marker *mdata, void *call_private, ...);
 	struct marker_probe_closure single;
 	struct marker_probe_closure *multi;
 	int refcount;	/* Number of times armed. 0 if disarmed. */
@@ -91,15 +91,13 @@ EXPORT_SYMBOL_GPL(__mark_empty_function);
  * marker_probe_cb Callback that prepares the variable argument list for probes.
  * @mdata: pointer of type struct marker
  * @call_private: caller site private data
- * @fmt: format string
  * @...:  Variable argument list.
  *
  * Since we do not use "typical" pointer based RCU in the 1 argument case, we
  * need to put a full smp_rmb() in this branch. This is why we do not use
  * rcu_dereference() for the pointer read.
  */
-void marker_probe_cb(const struct marker *mdata, void *call_private,
-	const char *fmt, ...)
+void marker_probe_cb(const struct marker *mdata, void *call_private, ...)
 {
 	va_list args;
 	char ptype;
@@ -120,8 +118,9 @@ void marker_probe_cb(const struct marker *mdata, void *call_private,
 		/* Must read the ptr before private data. They are not data
 		 * dependant, so we put an explicit smp_rmb() here. */
 		smp_rmb();
-		va_start(args, fmt);
-		func(mdata->single.probe_private, call_private, fmt, &args);
+		va_start(args, call_private);
+		func(mdata->single.probe_private, call_private, mdata->format,
+			&args);
 		va_end(args);
 	} else {
 		struct marker_probe_closure *multi;
@@ -136,9 +135,9 @@ void marker_probe_cb(const struct marker *mdata, void *call_private,
 		smp_read_barrier_depends();
 		multi = mdata->multi;
 		for (i = 0; multi[i].func; i++) {
-			va_start(args, fmt);
-			multi[i].func(multi[i].probe_private, call_private, fmt,
-				&args);
+			va_start(args, call_private);
+			multi[i].func(multi[i].probe_private, call_private,
+				mdata->format, &args);
 			va_end(args);
 		}
 	}
@@ -150,13 +149,11 @@ EXPORT_SYMBOL_GPL(marker_probe_cb);
  * marker_probe_cb Callback that does not prepare the variable argument list.
  * @mdata: pointer of type struct marker
  * @call_private: caller site private data
- * @fmt: format string
  * @...:  Variable argument list.
  *
  * Should be connected to markers "MARK_NOARGS".
  */
-void marker_probe_cb_noarg(const struct marker *mdata,
-	void *call_private, const char *fmt, ...)
+void marker_probe_cb_noarg(const struct marker *mdata, void *call_private, ...)
 {
 	va_list args;	/* not initialized */
 	char ptype;
@@ -172,7 +169,8 @@ void marker_probe_cb_noarg(const struct marker *mdata,
 		/* Must read the ptr before private data. They are not data
 		 * dependant, so we put an explicit smp_rmb() here. */
 		smp_rmb();
-		func(mdata->single.probe_private, call_private, fmt, &args);
+		func(mdata->single.probe_private, call_private, mdata->format,
+			&args);
 	} else {
 		struct marker_probe_closure *multi;
 		int i;
@@ -186,8 +184,8 @@ void marker_probe_cb_noarg(const struct marker *mdata,
 		smp_read_barrier_depends();
 		multi = mdata->multi;
 		for (i = 0; multi[i].func; i++)
-			multi[i].func(multi[i].probe_private, call_private, fmt,
-				&args);
+			multi[i].func(multi[i].probe_private, call_private,
+				mdata->format, &args);
 	}
 	preempt_enable();
 }

kernel/printk.c

@@ -1046,7 +1046,9 @@ void release_console_sem(void)
 		_log_end = log_end;
 		con_start = log_end;		/* Flush */
 		spin_unlock(&logbuf_lock);
+		stop_critical_timings();	/* don't trace print latency */
 		call_console_drivers(_con_start, _log_end);
+		start_critical_timings();
 		local_irq_restore(flags);
 	}
 	console_locked = 0;

kernel/sched.c

@@ -70,6 +70,7 @@
 #include <linux/bootmem.h>
 #include <linux/debugfs.h>
 #include <linux/ctype.h>
+#include <linux/ftrace.h>
 
 #include <asm/tlb.h>
 #include <asm/irq_regs.h>
@@ -645,6 +646,24 @@ static inline void update_rq_clock(struct rq *rq)
 # define const_debug static const
 #endif
 
+/**
+ * runqueue_is_locked
+ *
+ * Returns true if the current cpu runqueue is locked.
+ * This interface allows printk to be called with the runqueue lock
+ * held and know whether or not it is OK to wake up the klogd.
+ */
+int runqueue_is_locked(void)
+{
+	int cpu = get_cpu();
+	struct rq *rq = cpu_rq(cpu);
+	int ret;
+
+	ret = spin_is_locked(&rq->lock);
+	put_cpu();
+	return ret;
+}
+
 /*
  * Debugging: various feature bits
  */
@@ -2318,6 +2337,9 @@ out_activate:
 	success = 1;
 
 out_running:
+	trace_mark(kernel_sched_wakeup,
+		"pid %d state %ld ## rq %p task %p rq->curr %p",
+		p->pid, p->state, rq, p, rq->curr);
 	check_preempt_curr(rq, p);
 
 	p->state = TASK_RUNNING;
@@ -2450,6 +2472,9 @@ void wake_up_new_task(struct task_struct *p, unsigned long clone_flags)
 		p->sched_class->task_new(rq, p);
 		inc_nr_running(rq);
 	}
+	trace_mark(kernel_sched_wakeup_new,
+		"pid %d state %ld ## rq %p task %p rq->curr %p",
+		p->pid, p->state, rq, p, rq->curr);
 	check_preempt_curr(rq, p);
 #ifdef CONFIG_SMP
 	if (p->sched_class->task_wake_up)
@@ -2622,6 +2647,11 @@ context_switch(struct rq *rq, struct task_struct *prev,
 	struct mm_struct *mm, *oldmm;
 
 	prepare_task_switch(rq, prev, next);
+	trace_mark(kernel_sched_schedule,
+		"prev_pid %d next_pid %d prev_state %ld "
+		"## rq %p prev %p next %p",
+		prev->pid, next->pid, prev->state,
+		rq, prev, next);
 	mm = next->mm;
 	oldmm = prev->active_mm;
 	/*
@@ -4221,26 +4251,44 @@ void scheduler_tick(void)
 #endif
 }
 
-#if defined(CONFIG_PREEMPT) && defined(CONFIG_DEBUG_PREEMPT)
+#if defined(CONFIG_PREEMPT) && (defined(CONFIG_DEBUG_PREEMPT) || \
+				defined(CONFIG_PREEMPT_TRACER))
+
+static inline unsigned long get_parent_ip(unsigned long addr)
+{
+	if (in_lock_functions(addr)) {
+		addr = CALLER_ADDR2;
+		if (in_lock_functions(addr))
+			addr = CALLER_ADDR3;
+	}
+	return addr;
+}
 
 void __kprobes add_preempt_count(int val)
 {
+#ifdef CONFIG_DEBUG_PREEMPT
 	/*
 	 * Underflow?
 	 */
 	if (DEBUG_LOCKS_WARN_ON((preempt_count() < 0)))
 		return;
+#endif
 	preempt_count() += val;
+#ifdef CONFIG_DEBUG_PREEMPT
 	/*
 	 * Spinlock count overflowing soon?
 	 */
 	DEBUG_LOCKS_WARN_ON((preempt_count() & PREEMPT_MASK) >=
 				PREEMPT_MASK - 10);
+#endif
+	if (preempt_count() == val)
+		trace_preempt_off(CALLER_ADDR0, get_parent_ip(CALLER_ADDR1));
 }
 EXPORT_SYMBOL(add_preempt_count);
 
 void __kprobes sub_preempt_count(int val)
 {
+#ifdef CONFIG_DEBUG_PREEMPT
 	/*
 	 * Underflow?
 	 */
@@ -4252,7 +4300,10 @@ void __kprobes sub_preempt_count(int val)
 	if (DEBUG_LOCKS_WARN_ON((val < PREEMPT_MASK) &&
 			!(preempt_count() & PREEMPT_MASK)))
 		return;
+#endif
 
+	if (preempt_count() == val)
+		trace_preempt_on(CALLER_ADDR0, get_parent_ip(CALLER_ADDR1));
 	preempt_count() -= val;
 }
 EXPORT_SYMBOL(sub_preempt_count);
@@ -5566,7 +5617,7 @@ out_unlock:
 	return retval;
 }
 
-static const char stat_nam[] = "RSDTtZX";
+static const char stat_nam[] = TASK_STATE_TO_CHAR_STR;
 
 void sched_show_task(struct task_struct *p)
 {

kernel/semaphore.c

@@ -31,6 +31,7 @@
 #include <linux/sched.h>
 #include <linux/semaphore.h>
 #include <linux/spinlock.h>
+#include <linux/ftrace.h>
 
 static noinline void __down(struct semaphore *sem);
 static noinline int __down_interruptible(struct semaphore *sem);

kernel/spinlock.c

@@ -436,7 +436,7 @@ int __lockfunc _spin_trylock_bh(spinlock_t *lock)
 }
 EXPORT_SYMBOL(_spin_trylock_bh);
 
-int in_lock_functions(unsigned long addr)
+notrace int in_lock_functions(unsigned long addr)
 {
 	/* Linker adds these: start and end of __lockfunc functions */
 	extern char __lock_text_start[], __lock_text_end[];

kernel/sysctl.c

@@ -46,6 +46,7 @@
 #include <linux/nfs_fs.h>
 #include <linux/acpi.h>
 #include <linux/reboot.h>
+#include <linux/ftrace.h>
 
 #include <asm/uaccess.h>
 #include <asm/processor.h>
@@ -463,6 +464,16 @@ static struct ctl_table kern_table[] = {
 		.mode		= 0644,
 		.proc_handler	= &proc_dointvec,
 	},
+#ifdef CONFIG_FTRACE
+	{
+		.ctl_name	= CTL_UNNUMBERED,
+		.procname	= "ftrace_enabled",
+		.data		= &ftrace_enabled,
+		.maxlen		= sizeof(int),
+		.mode		= 0644,
+		.proc_handler	= &ftrace_enable_sysctl,
+	},
+#endif
 #ifdef CONFIG_KMOD
 	{
 		.ctl_name	= KERN_MODPROBE,

kernel/trace/Kconfig

@@ -0,0 +1,135 @@
+#
+# Architectures that offer an FTRACE implementation should select HAVE_FTRACE:
+#
+config HAVE_FTRACE
+	bool
+
+config HAVE_DYNAMIC_FTRACE
+	bool
+
+config TRACER_MAX_TRACE
+	bool
+
+config TRACING
+	bool
+	select DEBUG_FS
+	select STACKTRACE
+
+config FTRACE
+	bool "Kernel Function Tracer"
+	depends on HAVE_FTRACE
+	select FRAME_POINTER
+	select TRACING
+	select CONTEXT_SWITCH_TRACER
+	help
+	  Enable the kernel to trace every kernel function. This is done
+	  by using a compiler feature to insert a small, 5-byte No-Operation
+	  instruction to the beginning of every kernel function, which NOP
+	  sequence is then dynamically patched into a tracer call when
+	  tracing is enabled by the administrator. If it's runtime disabled
+	  (the bootup default), then the overhead of the instructions is very
+	  small and not measurable even in micro-benchmarks.
+
+config IRQSOFF_TRACER
+	bool "Interrupts-off Latency Tracer"
+	default n
+	depends on TRACE_IRQFLAGS_SUPPORT
+	depends on GENERIC_TIME
+	depends on HAVE_FTRACE
+	select TRACE_IRQFLAGS
+	select TRACING
+	select TRACER_MAX_TRACE
+	help
+	  This option measures the time spent in irqs-off critical
+	  sections, with microsecond accuracy.
+
+	  The default measurement method is a maximum search, which is
+	  disabled by default and can be runtime (re-)started
+	  via:
+
+	      echo 0 > /debugfs/tracing/tracing_max_latency
+
+	  (Note that kernel size and overhead increases with this option
+	  enabled. This option and the preempt-off timing option can be
+	  used together or separately.)
+
+config PREEMPT_TRACER
+	bool "Preemption-off Latency Tracer"
+	default n
+	depends on GENERIC_TIME
+	depends on PREEMPT
+	depends on HAVE_FTRACE
+	select TRACING
+	select TRACER_MAX_TRACE
+	help
+	  This option measures the time spent in preemption off critical
+	  sections, with microsecond accuracy.
+
+	  The default measurement method is a maximum search, which is
+	  disabled by default and can be runtime (re-)started
+	  via:
+
+	      echo 0 > /debugfs/tracing/tracing_max_latency
+
+	  (Note that kernel size and overhead increases with this option
+	  enabled. This option and the irqs-off timing option can be
+	  used together or separately.)
+
+config SYSPROF_TRACER
+	bool "Sysprof Tracer"
+	depends on X86
+	select TRACING
+	help
+	  This tracer provides the trace needed by the 'Sysprof' userspace
+	  tool.
+
+config SCHED_TRACER
+	bool "Scheduling Latency Tracer"
+	depends on HAVE_FTRACE
+	select TRACING
+	select CONTEXT_SWITCH_TRACER
+	select TRACER_MAX_TRACE
+	help
+	  This tracer tracks the latency of the highest priority task
+	  to be scheduled in, starting from the point it has woken up.
+
+config CONTEXT_SWITCH_TRACER
+	bool "Trace process context switches"
+	depends on HAVE_FTRACE
+	select TRACING
+	select MARKERS
+	help
+	  This tracer gets called from the context switch and records
+	  all switching of tasks.
+
+config DYNAMIC_FTRACE
+	bool "enable/disable ftrace tracepoints dynamically"
+	depends on FTRACE
+	depends on HAVE_DYNAMIC_FTRACE
+	default y
+	help
+         This option will modify all the calls to ftrace dynamically
+	 (will patch them out of the binary image and replaces them
+	 with a No-Op instruction) as they are called. A table is
+	 created to dynamically enable them again.
+
+	 This way a CONFIG_FTRACE kernel is slightly larger, but otherwise
+	 has native performance as long as no tracing is active.
+
+	 The changes to the code are done by a kernel thread that
+	 wakes up once a second and checks to see if any ftrace calls
+	 were made. If so, it runs stop_machine (stops all CPUS)
+	 and modifies the code to jump over the call to ftrace.
+
+config FTRACE_SELFTEST
+	bool
+
+config FTRACE_STARTUP_TEST
+	bool "Perform a startup test on ftrace"
+	depends on TRACING
+	select FTRACE_SELFTEST
+	help
+	  This option performs a series of startup tests on ftrace. On bootup
+	  a series of tests are made to verify that the tracer is
+	  functioning properly. It will do tests on all the configured
+	  tracers of ftrace.

kernel/trace/Makefile

@@ -0,0 +1,24 @@
+
+# Do not instrument the tracer itself:
+
+ifdef CONFIG_FTRACE
+ORIG_CFLAGS := $(KBUILD_CFLAGS)
+KBUILD_CFLAGS = $(subst -pg,,$(ORIG_CFLAGS))
+
+# selftest needs instrumentation
+CFLAGS_trace_selftest_dynamic.o = -pg
+obj-y += trace_selftest_dynamic.o
+endif
+
+obj-$(CONFIG_FTRACE) += libftrace.o
+
+obj-$(CONFIG_TRACING) += trace.o
+obj-$(CONFIG_CONTEXT_SWITCH_TRACER) += trace_sched_switch.o
+obj-$(CONFIG_SYSPROF_TRACER) += trace_sysprof.o
+obj-$(CONFIG_FTRACE) += trace_functions.o
+obj-$(CONFIG_IRQSOFF_TRACER) += trace_irqsoff.o
+obj-$(CONFIG_PREEMPT_TRACER) += trace_irqsoff.o
+obj-$(CONFIG_SCHED_TRACER) += trace_sched_wakeup.o
+obj-$(CONFIG_MMIOTRACE) += trace_mmiotrace.o
+
+libftrace-y := ftrace.o

kernel/trace/trace.h

@@ -0,0 +1,339 @@
+#ifndef _LINUX_KERNEL_TRACE_H
+#define _LINUX_KERNEL_TRACE_H
+
+#include <linux/fs.h>
+#include <asm/atomic.h>
+#include <linux/sched.h>
+#include <linux/clocksource.h>
+#include <linux/mmiotrace.h>
+
+enum trace_type {
+	__TRACE_FIRST_TYPE = 0,
+
+	TRACE_FN,
+	TRACE_CTX,
+	TRACE_WAKE,
+	TRACE_STACK,
+	TRACE_SPECIAL,
+	TRACE_MMIO_RW,
+	TRACE_MMIO_MAP,
+
+	__TRACE_LAST_TYPE
+};
+
+/*
+ * Function trace entry - function address and parent function addres:
+ */
+struct ftrace_entry {
+	unsigned long		ip;
+	unsigned long		parent_ip;
+};
+
+/*
+ * Context switch trace entry - which task (and prio) we switched from/to:
+ */
+struct ctx_switch_entry {
+	unsigned int		prev_pid;
+	unsigned char		prev_prio;
+	unsigned char		prev_state;
+	unsigned int		next_pid;
+	unsigned char		next_prio;
+	unsigned char		next_state;
+};
+
+/*
+ * Special (free-form) trace entry:
+ */
+struct special_entry {
+	unsigned long		arg1;
+	unsigned long		arg2;
+	unsigned long		arg3;
+};
+
+/*
+ * Stack-trace entry:
+ */
+
+#define FTRACE_STACK_ENTRIES	8
+
+struct stack_entry {
+	unsigned long		caller[FTRACE_STACK_ENTRIES];
+};
+
+/*
+ * The trace entry - the most basic unit of tracing. This is what
+ * is printed in the end as a single line in the trace output, such as:
+ *
+ *     bash-15816 [01]   235.197585: idle_cpu <- irq_enter
+ */
+struct trace_entry {
+	char			type;
+	char			cpu;
+	char			flags;
+	char			preempt_count;
+	int			pid;
+	cycle_t			t;
+	union {
+		struct ftrace_entry		fn;
+		struct ctx_switch_entry		ctx;
+		struct special_entry		special;
+		struct stack_entry		stack;
+		struct mmiotrace_rw		mmiorw;
+		struct mmiotrace_map		mmiomap;
+	};
+};
+
+#define TRACE_ENTRY_SIZE	sizeof(struct trace_entry)
+
+/*
+ * The CPU trace array - it consists of thousands of trace entries
+ * plus some other descriptor data: (for example which task started
+ * the trace, etc.)
+ */
+struct trace_array_cpu {
+	struct list_head	trace_pages;
+	atomic_t		disabled;
+	raw_spinlock_t		lock;
+	struct lock_class_key	lock_key;
+
+	/* these fields get copied into max-trace: */
+	unsigned		trace_head_idx;
+	unsigned		trace_tail_idx;
+	void			*trace_head; /* producer */
+	void			*trace_tail; /* consumer */
+	unsigned long		trace_idx;
+	unsigned long		overrun;
+	unsigned long		saved_latency;
+	unsigned long		critical_start;
+	unsigned long		critical_end;
+	unsigned long		critical_sequence;
+	unsigned long		nice;
+	unsigned long		policy;
+	unsigned long		rt_priority;
+	cycle_t			preempt_timestamp;
+	pid_t			pid;
+	uid_t			uid;
+	char			comm[TASK_COMM_LEN];
+};
+
+struct trace_iterator;
+
+/*
+ * The trace array - an array of per-CPU trace arrays. This is the
+ * highest level data structure that individual tracers deal with.
+ * They have on/off state as well:
+ */
+struct trace_array {
+	unsigned long		entries;
+	long			ctrl;
+	int			cpu;
+	cycle_t			time_start;
+	struct task_struct	*waiter;
+	struct trace_array_cpu	*data[NR_CPUS];
+};
+
+/*
+ * A specific tracer, represented by methods that operate on a trace array:
+ */
+struct tracer {
+	const char		*name;
+	void			(*init)(struct trace_array *tr);
+	void			(*reset)(struct trace_array *tr);
+	void			(*open)(struct trace_iterator *iter);
+	void			(*pipe_open)(struct trace_iterator *iter);
+	void			(*close)(struct trace_iterator *iter);
+	void			(*start)(struct trace_iterator *iter);
+	void			(*stop)(struct trace_iterator *iter);
+	ssize_t			(*read)(struct trace_iterator *iter,
+					struct file *filp, char __user *ubuf,
+					size_t cnt, loff_t *ppos);
+	void			(*ctrl_update)(struct trace_array *tr);
+#ifdef CONFIG_FTRACE_STARTUP_TEST
+	int			(*selftest)(struct tracer *trace,
+					    struct trace_array *tr);
+#endif
+	int			(*print_line)(struct trace_iterator *iter);
+	struct tracer		*next;
+	int			print_max;
+};
+
+struct trace_seq {
+	unsigned char		buffer[PAGE_SIZE];
+	unsigned int		len;
+	unsigned int		readpos;
+};
+
+/*
+ * Trace iterator - used by printout routines who present trace
+ * results to users and which routines might sleep, etc:
+ */
+struct trace_iterator {
+	struct trace_array	*tr;
+	struct tracer		*trace;
+	void			*private;
+	long			last_overrun[NR_CPUS];
+	long			overrun[NR_CPUS];
+
+	/* The below is zeroed out in pipe_read */
+	struct trace_seq	seq;
+	struct trace_entry	*ent;
+	int			cpu;
+
+	struct trace_entry	*prev_ent;
+	int			prev_cpu;
+
+	unsigned long		iter_flags;
+	loff_t			pos;
+	unsigned long		next_idx[NR_CPUS];
+	struct list_head	*next_page[NR_CPUS];
+	unsigned		next_page_idx[NR_CPUS];
+	long			idx;
+};
+
+void tracing_reset(struct trace_array_cpu *data);
+int tracing_open_generic(struct inode *inode, struct file *filp);
+struct dentry *tracing_init_dentry(void);
+void init_tracer_sysprof_debugfs(struct dentry *d_tracer);
+
+void ftrace(struct trace_array *tr,
+			    struct trace_array_cpu *data,
+			    unsigned long ip,
+			    unsigned long parent_ip,
+			    unsigned long flags);
+void tracing_sched_switch_trace(struct trace_array *tr,
+				struct trace_array_cpu *data,
+				struct task_struct *prev,
+				struct task_struct *next,
+				unsigned long flags);
+void tracing_record_cmdline(struct task_struct *tsk);
+
+void tracing_sched_wakeup_trace(struct trace_array *tr,
+				struct trace_array_cpu *data,
+				struct task_struct *wakee,
+				struct task_struct *cur,
+				unsigned long flags);
+void trace_special(struct trace_array *tr,
+		   struct trace_array_cpu *data,
+		   unsigned long arg1,
+		   unsigned long arg2,
+		   unsigned long arg3);
+void trace_function(struct trace_array *tr,
+		    struct trace_array_cpu *data,
+		    unsigned long ip,
+		    unsigned long parent_ip,
+		    unsigned long flags);
+
+void tracing_start_cmdline_record(void);
+void tracing_stop_cmdline_record(void);
+int register_tracer(struct tracer *type);
+void unregister_tracer(struct tracer *type);
+
+extern unsigned long nsecs_to_usecs(unsigned long nsecs);
+
+extern unsigned long tracing_max_latency;
+extern unsigned long tracing_thresh;
+
+void update_max_tr(struct trace_array *tr, struct task_struct *tsk, int cpu);
+void update_max_tr_single(struct trace_array *tr,
+			  struct task_struct *tsk, int cpu);
+
+extern cycle_t ftrace_now(int cpu);
+
+#ifdef CONFIG_FTRACE
+void tracing_start_function_trace(void);
+void tracing_stop_function_trace(void);
+#else
+# define tracing_start_function_trace()		do { } while (0)
+# define tracing_stop_function_trace()		do { } while (0)
+#endif
+
+#ifdef CONFIG_CONTEXT_SWITCH_TRACER
+typedef void
+(*tracer_switch_func_t)(void *private,
+			void *__rq,
+			struct task_struct *prev,
+			struct task_struct *next);
+
+struct tracer_switch_ops {
+	tracer_switch_func_t		func;
+	void				*private;
+	struct tracer_switch_ops	*next;
+};
+
+#endif /* CONFIG_CONTEXT_SWITCH_TRACER */
+
+#ifdef CONFIG_DYNAMIC_FTRACE
+extern unsigned long ftrace_update_tot_cnt;
+#define DYN_FTRACE_TEST_NAME trace_selftest_dynamic_test_func
+extern int DYN_FTRACE_TEST_NAME(void);
+#endif
+
+#ifdef CONFIG_MMIOTRACE
+extern void __trace_mmiotrace_rw(struct trace_array *tr,
+				struct trace_array_cpu *data,
+				struct mmiotrace_rw *rw);
+extern void __trace_mmiotrace_map(struct trace_array *tr,
+				struct trace_array_cpu *data,
+				struct mmiotrace_map *map);
+#endif
+
+#ifdef CONFIG_FTRACE_STARTUP_TEST
+#ifdef CONFIG_FTRACE
+extern int trace_selftest_startup_function(struct tracer *trace,
+					   struct trace_array *tr);
+#endif
+#ifdef CONFIG_IRQSOFF_TRACER
+extern int trace_selftest_startup_irqsoff(struct tracer *trace,
+					  struct trace_array *tr);
+#endif
+#ifdef CONFIG_PREEMPT_TRACER
+extern int trace_selftest_startup_preemptoff(struct tracer *trace,
+					     struct trace_array *tr);
+#endif
+#if defined(CONFIG_IRQSOFF_TRACER) && defined(CONFIG_PREEMPT_TRACER)
+extern int trace_selftest_startup_preemptirqsoff(struct tracer *trace,
+						 struct trace_array *tr);
+#endif
+#ifdef CONFIG_SCHED_TRACER
+extern int trace_selftest_startup_wakeup(struct tracer *trace,
+					 struct trace_array *tr);
+#endif
+#ifdef CONFIG_CONTEXT_SWITCH_TRACER
+extern int trace_selftest_startup_sched_switch(struct tracer *trace,
+					       struct trace_array *tr);
+#endif
+#ifdef CONFIG_SYSPROF_TRACER
+extern int trace_selftest_startup_sysprof(struct tracer *trace,
+					       struct trace_array *tr);
+#endif
+#endif /* CONFIG_FTRACE_STARTUP_TEST */
+
+extern void *head_page(struct trace_array_cpu *data);
+extern int trace_seq_printf(struct trace_seq *s, const char *fmt, ...);
+extern ssize_t trace_seq_to_user(struct trace_seq *s, char __user *ubuf,
+				 size_t cnt);
+extern long ns2usecs(cycle_t nsec);
+
+extern unsigned long trace_flags;
+
+/*
+ * trace_iterator_flags is an enumeration that defines bit
+ * positions into trace_flags that controls the output.
+ *
+ * NOTE: These bits must match the trace_options array in
+ *       trace.c.
+ */
+enum trace_iterator_flags {
+	TRACE_ITER_PRINT_PARENT		= 0x01,
+	TRACE_ITER_SYM_OFFSET		= 0x02,
+	TRACE_ITER_SYM_ADDR		= 0x04,
+	TRACE_ITER_VERBOSE		= 0x08,
+	TRACE_ITER_RAW			= 0x10,
+	TRACE_ITER_HEX			= 0x20,
+	TRACE_ITER_BIN			= 0x40,
+	TRACE_ITER_BLOCK		= 0x80,
+	TRACE_ITER_STACKTRACE		= 0x100,
+	TRACE_ITER_SCHED_TREE		= 0x200,
+};
+
+#endif /* _LINUX_KERNEL_TRACE_H */

kernel/trace/trace_functions.c

@@ -0,0 +1,81 @@
+/*
+ * ring buffer based function tracer
+ *
+ * Copyright (C) 2007-2008 Steven Rostedt <srostedt@redhat.com>
+ * Copyright (C) 2008 Ingo Molnar <mingo@redhat.com>
+ *
+ * Based on code from the latency_tracer, that is:
+ *
+ *  Copyright (C) 2004-2006 Ingo Molnar
+ *  Copyright (C) 2004 William Lee Irwin III
+ */
+#include <linux/debugfs.h>
+#include <linux/uaccess.h>
+#include <linux/ftrace.h>
+#include <linux/fs.h>
+
+#include "trace.h"
+
+static void function_reset(struct trace_array *tr)
+{
+	int cpu;
+
+	tr->time_start = ftrace_now(tr->cpu);
+
+	for_each_online_cpu(cpu)
+		tracing_reset(tr->data[cpu]);
+}
+
+static void start_function_trace(struct trace_array *tr)
+{
+	tr->cpu = get_cpu();
+	function_reset(tr);
+	put_cpu();
+
+	tracing_start_cmdline_record();
+	tracing_start_function_trace();
+}
+
+static void stop_function_trace(struct trace_array *tr)
+{
+	tracing_stop_function_trace();
+	tracing_stop_cmdline_record();
+}
+
+static void function_trace_init(struct trace_array *tr)
+{
+	if (tr->ctrl)
+		start_function_trace(tr);
+}
+
+static void function_trace_reset(struct trace_array *tr)
+{
+	if (tr->ctrl)
+		stop_function_trace(tr);
+}
+
+static void function_trace_ctrl_update(struct trace_array *tr)
+{
+	if (tr->ctrl)
+		start_function_trace(tr);
+	else
+		stop_function_trace(tr);
+}
+
+static struct tracer function_trace __read_mostly =
+{
+	.name	     = "ftrace",
+	.init	     = function_trace_init,
+	.reset	     = function_trace_reset,
+	.ctrl_update = function_trace_ctrl_update,
+#ifdef CONFIG_FTRACE_SELFTEST
+	.selftest    = trace_selftest_startup_function,
+#endif
+};
+
+static __init int init_function_trace(void)
+{
+	return register_tracer(&function_trace);
+}
+
+device_initcall(init_function_trace);

kernel/trace/trace_irqsoff.c

@@ -0,0 +1,486 @@
+/*
+ * trace irqs off criticall timings
+ *
+ * Copyright (C) 2007-2008 Steven Rostedt <srostedt@redhat.com>
+ * Copyright (C) 2008 Ingo Molnar <mingo@redhat.com>
+ *
+ * From code in the latency_tracer, that is:
+ *
+ *  Copyright (C) 2004-2006 Ingo Molnar
+ *  Copyright (C) 2004 William Lee Irwin III
+ */
+#include <linux/kallsyms.h>
+#include <linux/debugfs.h>
+#include <linux/uaccess.h>
+#include <linux/module.h>
+#include <linux/ftrace.h>
+#include <linux/fs.h>
+
+#include "trace.h"
+
+static struct trace_array		*irqsoff_trace __read_mostly;
+static int				tracer_enabled __read_mostly;
+
+static DEFINE_PER_CPU(int, tracing_cpu);
+
+static DEFINE_SPINLOCK(max_trace_lock);
+
+enum {
+	TRACER_IRQS_OFF		= (1 << 1),
+	TRACER_PREEMPT_OFF	= (1 << 2),
+};
+
+static int trace_type __read_mostly;
+
+#ifdef CONFIG_PREEMPT_TRACER
+static inline int
+preempt_trace(void)
+{
+	return ((trace_type & TRACER_PREEMPT_OFF) && preempt_count());
+}
+#else
+# define preempt_trace() (0)
+#endif
+
+#ifdef CONFIG_IRQSOFF_TRACER
+static inline int
+irq_trace(void)
+{
+	return ((trace_type & TRACER_IRQS_OFF) &&
+		irqs_disabled());
+}
+#else
+# define irq_trace() (0)
+#endif
+
+/*
+ * Sequence count - we record it when starting a measurement and
+ * skip the latency if the sequence has changed - some other section
+ * did a maximum and could disturb our measurement with serial console
+ * printouts, etc. Truly coinciding maximum latencies should be rare
+ * and what happens together happens separately as well, so this doesnt
+ * decrease the validity of the maximum found:
+ */
+static __cacheline_aligned_in_smp	unsigned long max_sequence;
+
+#ifdef CONFIG_FTRACE
+/*
+ * irqsoff uses its own tracer function to keep the overhead down:
+ */
+static void
+irqsoff_tracer_call(unsigned long ip, unsigned long parent_ip)
+{
+	struct trace_array *tr = irqsoff_trace;
+	struct trace_array_cpu *data;
+	unsigned long flags;
+	long disabled;
+	int cpu;
+
+	/*
+	 * Does not matter if we preempt. We test the flags
+	 * afterward, to see if irqs are disabled or not.
+	 * If we preempt and get a false positive, the flags
+	 * test will fail.
+	 */
+	cpu = raw_smp_processor_id();
+	if (likely(!per_cpu(tracing_cpu, cpu)))
+		return;
+
+	local_save_flags(flags);
+	/* slight chance to get a false positive on tracing_cpu */
+	if (!irqs_disabled_flags(flags))
+		return;
+
+	data = tr->data[cpu];
+	disabled = atomic_inc_return(&data->disabled);
+
+	if (likely(disabled == 1))
+		trace_function(tr, data, ip, parent_ip, flags);
+
+	atomic_dec(&data->disabled);
+}
+
+static struct ftrace_ops trace_ops __read_mostly =
+{
+	.func = irqsoff_tracer_call,
+};
+#endif /* CONFIG_FTRACE */
+
+/*
+ * Should this new latency be reported/recorded?
+ */
+static int report_latency(cycle_t delta)
+{
+	if (tracing_thresh) {
+		if (delta < tracing_thresh)
+			return 0;
+	} else {
+		if (delta <= tracing_max_latency)
+			return 0;
+	}
+	return 1;
+}
+
+static void
+check_critical_timing(struct trace_array *tr,
+		      struct trace_array_cpu *data,
+		      unsigned long parent_ip,
+		      int cpu)
+{
+	unsigned long latency, t0, t1;
+	cycle_t T0, T1, delta;
+	unsigned long flags;
+
+	/*
+	 * usecs conversion is slow so we try to delay the conversion
+	 * as long as possible:
+	 */
+	T0 = data->preempt_timestamp;
+	T1 = ftrace_now(cpu);
+	delta = T1-T0;
+
+	local_save_flags(flags);
+
+	if (!report_latency(delta))
+		goto out;
+
+	spin_lock_irqsave(&max_trace_lock, flags);
+
+	/* check if we are still the max latency */
+	if (!report_latency(delta))
+		goto out_unlock;
+
+	trace_function(tr, data, CALLER_ADDR0, parent_ip, flags);
+
+	latency = nsecs_to_usecs(delta);
+
+	if (data->critical_sequence != max_sequence)
+		goto out_unlock;
+
+	tracing_max_latency = delta;
+	t0 = nsecs_to_usecs(T0);
+	t1 = nsecs_to_usecs(T1);
+
+	data->critical_end = parent_ip;
+
+	update_max_tr_single(tr, current, cpu);
+
+	max_sequence++;
+
+out_unlock:
+	spin_unlock_irqrestore(&max_trace_lock, flags);
+
+out:
+	data->critical_sequence = max_sequence;
+	data->preempt_timestamp = ftrace_now(cpu);
+	tracing_reset(data);
+	trace_function(tr, data, CALLER_ADDR0, parent_ip, flags);
+}
+
+static inline void
+start_critical_timing(unsigned long ip, unsigned long parent_ip)
+{
+	int cpu;
+	struct trace_array *tr = irqsoff_trace;
+	struct trace_array_cpu *data;
+	unsigned long flags;
+
+	if (likely(!tracer_enabled))
+		return;
+
+	cpu = raw_smp_processor_id();
+
+	if (per_cpu(tracing_cpu, cpu))
+		return;
+
+	data = tr->data[cpu];
+
+	if (unlikely(!data) || atomic_read(&data->disabled))
+		return;
+
+	atomic_inc(&data->disabled);
+
+	data->critical_sequence = max_sequence;
+	data->preempt_timestamp = ftrace_now(cpu);
+	data->critical_start = parent_ip ? : ip;
+	tracing_reset(data);
+
+	local_save_flags(flags);
+
+	trace_function(tr, data, ip, parent_ip, flags);
+
+	per_cpu(tracing_cpu, cpu) = 1;
+
+	atomic_dec(&data->disabled);
+}
+
+static inline void
+stop_critical_timing(unsigned long ip, unsigned long parent_ip)
+{
+	int cpu;
+	struct trace_array *tr = irqsoff_trace;
+	struct trace_array_cpu *data;
+	unsigned long flags;
+
+	cpu = raw_smp_processor_id();
+	/* Always clear the tracing cpu on stopping the trace */
+	if (unlikely(per_cpu(tracing_cpu, cpu)))
+		per_cpu(tracing_cpu, cpu) = 0;
+	else
+		return;
+
+	if (!tracer_enabled)
+		return;
+
+	data = tr->data[cpu];
+
+	if (unlikely(!data) || unlikely(!head_page(data)) ||
+	    !data->critical_start || atomic_read(&data->disabled))
+		return;
+
+	atomic_inc(&data->disabled);
+
+	local_save_flags(flags);
+	trace_function(tr, data, ip, parent_ip, flags);
+	check_critical_timing(tr, data, parent_ip ? : ip, cpu);
+	data->critical_start = 0;
+	atomic_dec(&data->disabled);
+}
+
+/* start and stop critical timings used to for stoppage (in idle) */
+void start_critical_timings(void)
+{
+	if (preempt_trace() || irq_trace())
+		start_critical_timing(CALLER_ADDR0, CALLER_ADDR1);
+}
+
+void stop_critical_timings(void)
+{
+	if (preempt_trace() || irq_trace())
+		stop_critical_timing(CALLER_ADDR0, CALLER_ADDR1);
+}
+
+#ifdef CONFIG_IRQSOFF_TRACER
+#ifdef CONFIG_PROVE_LOCKING
+void time_hardirqs_on(unsigned long a0, unsigned long a1)
+{
+	if (!preempt_trace() && irq_trace())
+		stop_critical_timing(a0, a1);
+}
+
+void time_hardirqs_off(unsigned long a0, unsigned long a1)
+{
+	if (!preempt_trace() && irq_trace())
+		start_critical_timing(a0, a1);
+}
+
+#else /* !CONFIG_PROVE_LOCKING */
+
+/*
+ * Stubs:
+ */
+
+void early_boot_irqs_off(void)
+{
+}
+
+void early_boot_irqs_on(void)
+{
+}
+
+void trace_softirqs_on(unsigned long ip)
+{
+}
+
+void trace_softirqs_off(unsigned long ip)
+{
+}
+
+inline void print_irqtrace_events(struct task_struct *curr)
+{
+}
+
+/*
+ * We are only interested in hardirq on/off events:
+ */
+void trace_hardirqs_on(void)
+{
+	if (!preempt_trace() && irq_trace())
+		stop_critical_timing(CALLER_ADDR0, CALLER_ADDR1);
+}
+EXPORT_SYMBOL(trace_hardirqs_on);
+
+void trace_hardirqs_off(void)
+{
+	if (!preempt_trace() && irq_trace())
+		start_critical_timing(CALLER_ADDR0, CALLER_ADDR1);
+}
+EXPORT_SYMBOL(trace_hardirqs_off);
+
+void trace_hardirqs_on_caller(unsigned long caller_addr)
+{
+	if (!preempt_trace() && irq_trace())
+		stop_critical_timing(CALLER_ADDR0, caller_addr);
+}
+EXPORT_SYMBOL(trace_hardirqs_on_caller);
+
+void trace_hardirqs_off_caller(unsigned long caller_addr)
+{
+	if (!preempt_trace() && irq_trace())
+		start_critical_timing(CALLER_ADDR0, caller_addr);
+}
+EXPORT_SYMBOL(trace_hardirqs_off_caller);
+
+#endif /* CONFIG_PROVE_LOCKING */
+#endif /*  CONFIG_IRQSOFF_TRACER */
+
+#ifdef CONFIG_PREEMPT_TRACER
+void trace_preempt_on(unsigned long a0, unsigned long a1)
+{
+	stop_critical_timing(a0, a1);
+}
+
+void trace_preempt_off(unsigned long a0, unsigned long a1)
+{
+	start_critical_timing(a0, a1);
+}
+#endif /* CONFIG_PREEMPT_TRACER */
+
+static void start_irqsoff_tracer(struct trace_array *tr)
+{
+	register_ftrace_function(&trace_ops);
+	tracer_enabled = 1;
+}
+
+static void stop_irqsoff_tracer(struct trace_array *tr)
+{
+	tracer_enabled = 0;
+	unregister_ftrace_function(&trace_ops);
+}
+
+static void __irqsoff_tracer_init(struct trace_array *tr)
+{
+	irqsoff_trace = tr;
+	/* make sure that the tracer is visible */
+	smp_wmb();
+
+	if (tr->ctrl)
+		start_irqsoff_tracer(tr);
+}
+
+static void irqsoff_tracer_reset(struct trace_array *tr)
+{
+	if (tr->ctrl)
+		stop_irqsoff_tracer(tr);
+}
+
+static void irqsoff_tracer_ctrl_update(struct trace_array *tr)
+{
+	if (tr->ctrl)
+		start_irqsoff_tracer(tr);
+	else
+		stop_irqsoff_tracer(tr);
+}
+
+static void irqsoff_tracer_open(struct trace_iterator *iter)
+{
+	/* stop the trace while dumping */
+	if (iter->tr->ctrl)
+		stop_irqsoff_tracer(iter->tr);
+}
+
+static void irqsoff_tracer_close(struct trace_iterator *iter)
+{
+	if (iter->tr->ctrl)
+		start_irqsoff_tracer(iter->tr);
+}
+
+#ifdef CONFIG_IRQSOFF_TRACER
+static void irqsoff_tracer_init(struct trace_array *tr)
+{
+	trace_type = TRACER_IRQS_OFF;
+
+	__irqsoff_tracer_init(tr);
+}
+static struct tracer irqsoff_tracer __read_mostly =
+{
+	.name		= "irqsoff",
+	.init		= irqsoff_tracer_init,
+	.reset		= irqsoff_tracer_reset,
+	.open		= irqsoff_tracer_open,
+	.close		= irqsoff_tracer_close,
+	.ctrl_update	= irqsoff_tracer_ctrl_update,
+	.print_max	= 1,
+#ifdef CONFIG_FTRACE_SELFTEST
+	.selftest    = trace_selftest_startup_irqsoff,
+#endif
+};
+# define register_irqsoff(trace) register_tracer(&trace)
+#else
+# define register_irqsoff(trace) do { } while (0)
+#endif
+
+#ifdef CONFIG_PREEMPT_TRACER
+static void preemptoff_tracer_init(struct trace_array *tr)
+{
+	trace_type = TRACER_PREEMPT_OFF;
+
+	__irqsoff_tracer_init(tr);
+}
+
+static struct tracer preemptoff_tracer __read_mostly =
+{
+	.name		= "preemptoff",
+	.init		= preemptoff_tracer_init,
+	.reset		= irqsoff_tracer_reset,
+	.open		= irqsoff_tracer_open,
+	.close		= irqsoff_tracer_close,
+	.ctrl_update	= irqsoff_tracer_ctrl_update,
+	.print_max	= 1,
+#ifdef CONFIG_FTRACE_SELFTEST
+	.selftest    = trace_selftest_startup_preemptoff,
+#endif
+};
+# define register_preemptoff(trace) register_tracer(&trace)
+#else
+# define register_preemptoff(trace) do { } while (0)
+#endif
+
+#if defined(CONFIG_IRQSOFF_TRACER) && \
+	defined(CONFIG_PREEMPT_TRACER)
+
+static void preemptirqsoff_tracer_init(struct trace_array *tr)
+{
+	trace_type = TRACER_IRQS_OFF | TRACER_PREEMPT_OFF;
+
+	__irqsoff_tracer_init(tr);
+}
+
+static struct tracer preemptirqsoff_tracer __read_mostly =
+{
+	.name		= "preemptirqsoff",
+	.init		= preemptirqsoff_tracer_init,
+	.reset		= irqsoff_tracer_reset,
+	.open		= irqsoff_tracer_open,
+	.close		= irqsoff_tracer_close,
+	.ctrl_update	= irqsoff_tracer_ctrl_update,
+	.print_max	= 1,
+#ifdef CONFIG_FTRACE_SELFTEST
+	.selftest    = trace_selftest_startup_preemptirqsoff,
+#endif
+};
+
+# define register_preemptirqsoff(trace) register_tracer(&trace)
+#else
+# define register_preemptirqsoff(trace) do { } while (0)
+#endif
+
+__init static int init_irqsoff_tracer(void)
+{
+	register_irqsoff(irqsoff_tracer);
+	register_preemptoff(preemptoff_tracer);
+	register_preemptirqsoff(preemptirqsoff_tracer);
+
+	return 0;
+}
+device_initcall(init_irqsoff_tracer);

kernel/trace/trace_mmiotrace.c

@@ -0,0 +1,295 @@
+/*
+ * Memory mapped I/O tracing
+ *
+ * Copyright (C) 2008 Pekka Paalanen <pq@iki.fi>
+ */
+
+#define DEBUG 1
+
+#include <linux/kernel.h>
+#include <linux/mmiotrace.h>
+#include <linux/pci.h>
+
+#include "trace.h"
+
+struct header_iter {
+	struct pci_dev *dev;
+};
+
+static struct trace_array *mmio_trace_array;
+static bool overrun_detected;
+
+static void mmio_reset_data(struct trace_array *tr)
+{
+	int cpu;
+
+	overrun_detected = false;
+	tr->time_start = ftrace_now(tr->cpu);
+
+	for_each_online_cpu(cpu)
+		tracing_reset(tr->data[cpu]);
+}
+
+static void mmio_trace_init(struct trace_array *tr)
+{
+	pr_debug("in %s\n", __func__);
+	mmio_trace_array = tr;
+	if (tr->ctrl) {
+		mmio_reset_data(tr);
+		enable_mmiotrace();
+	}
+}
+
+static void mmio_trace_reset(struct trace_array *tr)
+{
+	pr_debug("in %s\n", __func__);
+	if (tr->ctrl)
+		disable_mmiotrace();
+	mmio_reset_data(tr);
+	mmio_trace_array = NULL;
+}
+
+static void mmio_trace_ctrl_update(struct trace_array *tr)
+{
+	pr_debug("in %s\n", __func__);
+	if (tr->ctrl) {
+		mmio_reset_data(tr);
+		enable_mmiotrace();
+	} else {
+		disable_mmiotrace();
+	}
+}
+
+static int mmio_print_pcidev(struct trace_seq *s, const struct pci_dev *dev)
+{
+	int ret = 0;
+	int i;
+	resource_size_t start, end;
+	const struct pci_driver *drv = pci_dev_driver(dev);
+
+	/* XXX: incomplete checks for trace_seq_printf() return value */
+	ret += trace_seq_printf(s, "PCIDEV %02x%02x %04x%04x %x",
+				dev->bus->number, dev->devfn,
+				dev->vendor, dev->device, dev->irq);
+	/*
+	 * XXX: is pci_resource_to_user() appropriate, since we are
+	 * supposed to interpret the __ioremap() phys_addr argument based on
+	 * these printed values?
+	 */
+	for (i = 0; i < 7; i++) {
+		pci_resource_to_user(dev, i, &dev->resource[i], &start, &end);
+		ret += trace_seq_printf(s, " %llx",
+			(unsigned long long)(start |
+			(dev->resource[i].flags & PCI_REGION_FLAG_MASK)));
+	}
+	for (i = 0; i < 7; i++) {
+		pci_resource_to_user(dev, i, &dev->resource[i], &start, &end);
+		ret += trace_seq_printf(s, " %llx",
+			dev->resource[i].start < dev->resource[i].end ?
+			(unsigned long long)(end - start) + 1 : 0);
+	}
+	if (drv)
+		ret += trace_seq_printf(s, " %s\n", drv->name);
+	else
+		ret += trace_seq_printf(s, " \n");
+	return ret;
+}
+
+static void destroy_header_iter(struct header_iter *hiter)
+{
+	if (!hiter)
+		return;
+	pci_dev_put(hiter->dev);
+	kfree(hiter);
+}
+
+static void mmio_pipe_open(struct trace_iterator *iter)
+{
+	struct header_iter *hiter;
+	struct trace_seq *s = &iter->seq;
+
+	trace_seq_printf(s, "VERSION 20070824\n");
+
+	hiter = kzalloc(sizeof(*hiter), GFP_KERNEL);
+	if (!hiter)
+		return;
+
+	hiter->dev = pci_get_device(PCI_ANY_ID, PCI_ANY_ID, NULL);
+	iter->private = hiter;
+}
+
+/* XXX: This is not called when the pipe is closed! */
+static void mmio_close(struct trace_iterator *iter)
+{
+	struct header_iter *hiter = iter->private;
+	destroy_header_iter(hiter);
+	iter->private = NULL;
+}
+
+static unsigned long count_overruns(struct trace_iterator *iter)
+{
+	int cpu;
+	unsigned long cnt = 0;
+	for_each_online_cpu(cpu) {
+		cnt += iter->overrun[cpu];
+		iter->overrun[cpu] = 0;
+	}
+	return cnt;
+}
+
+static ssize_t mmio_read(struct trace_iterator *iter, struct file *filp,
+				char __user *ubuf, size_t cnt, loff_t *ppos)
+{
+	ssize_t ret;
+	struct header_iter *hiter = iter->private;
+	struct trace_seq *s = &iter->seq;
+	unsigned long n;
+
+	n = count_overruns(iter);
+	if (n) {
+		/* XXX: This is later than where events were lost. */
+		trace_seq_printf(s, "MARK 0.000000 Lost %lu events.\n", n);
+		if (!overrun_detected)
+			pr_warning("mmiotrace has lost events.\n");
+		overrun_detected = true;
+		goto print_out;
+	}
+
+	if (!hiter)
+		return 0;
+
+	mmio_print_pcidev(s, hiter->dev);
+	hiter->dev = pci_get_device(PCI_ANY_ID, PCI_ANY_ID, hiter->dev);
+
+	if (!hiter->dev) {
+		destroy_header_iter(hiter);
+		iter->private = NULL;
+	}
+
+print_out:
+	ret = trace_seq_to_user(s, ubuf, cnt);
+	return (ret == -EBUSY) ? 0 : ret;
+}
+
+static int mmio_print_rw(struct trace_iterator *iter)
+{
+	struct trace_entry *entry = iter->ent;
+	struct mmiotrace_rw *rw	= &entry->mmiorw;
+	struct trace_seq *s	= &iter->seq;
+	unsigned long long t	= ns2usecs(entry->t);
+	unsigned long usec_rem	= do_div(t, 1000000ULL);
+	unsigned secs		= (unsigned long)t;
+	int ret = 1;
+
+	switch (entry->mmiorw.opcode) {
+	case MMIO_READ:
+		ret = trace_seq_printf(s,
+			"R %d %lu.%06lu %d 0x%llx 0x%lx 0x%lx %d\n",
+			rw->width, secs, usec_rem, rw->map_id,
+			(unsigned long long)rw->phys,
+			rw->value, rw->pc, 0);
+		break;
+	case MMIO_WRITE:
+		ret = trace_seq_printf(s,
+			"W %d %lu.%06lu %d 0x%llx 0x%lx 0x%lx %d\n",
+			rw->width, secs, usec_rem, rw->map_id,
+			(unsigned long long)rw->phys,
+			rw->value, rw->pc, 0);
+		break;
+	case MMIO_UNKNOWN_OP:
+		ret = trace_seq_printf(s,
+			"UNKNOWN %lu.%06lu %d 0x%llx %02x,%02x,%02x 0x%lx %d\n",
+			secs, usec_rem, rw->map_id,
+			(unsigned long long)rw->phys,
+			(rw->value >> 16) & 0xff, (rw->value >> 8) & 0xff,
+			(rw->value >> 0) & 0xff, rw->pc, 0);
+		break;
+	default:
+		ret = trace_seq_printf(s, "rw what?\n");
+		break;
+	}
+	if (ret)
+		return 1;
+	return 0;
+}
+
+static int mmio_print_map(struct trace_iterator *iter)
+{
+	struct trace_entry *entry = iter->ent;
+	struct mmiotrace_map *m	= &entry->mmiomap;
+	struct trace_seq *s	= &iter->seq;
+	unsigned long long t	= ns2usecs(entry->t);
+	unsigned long usec_rem	= do_div(t, 1000000ULL);
+	unsigned secs		= (unsigned long)t;
+	int ret = 1;
+
+	switch (entry->mmiorw.opcode) {
+	case MMIO_PROBE:
+		ret = trace_seq_printf(s,
+			"MAP %lu.%06lu %d 0x%llx 0x%lx 0x%lx 0x%lx %d\n",
+			secs, usec_rem, m->map_id,
+			(unsigned long long)m->phys, m->virt, m->len,
+			0UL, 0);
+		break;
+	case MMIO_UNPROBE:
+		ret = trace_seq_printf(s,
+			"UNMAP %lu.%06lu %d 0x%lx %d\n",
+			secs, usec_rem, m->map_id, 0UL, 0);
+		break;
+	default:
+		ret = trace_seq_printf(s, "map what?\n");
+		break;
+	}
+	if (ret)
+		return 1;
+	return 0;
+}
+
+/* return 0 to abort printing without consuming current entry in pipe mode */
+static int mmio_print_line(struct trace_iterator *iter)
+{
+	switch (iter->ent->type) {
+	case TRACE_MMIO_RW:
+		return mmio_print_rw(iter);
+	case TRACE_MMIO_MAP:
+		return mmio_print_map(iter);
+	default:
+		return 1; /* ignore unknown entries */
+	}
+}
+
+static struct tracer mmio_tracer __read_mostly =
+{
+	.name		= "mmiotrace",
+	.init		= mmio_trace_init,
+	.reset		= mmio_trace_reset,
+	.pipe_open	= mmio_pipe_open,
+	.close		= mmio_close,
+	.read		= mmio_read,
+	.ctrl_update	= mmio_trace_ctrl_update,
+	.print_line	= mmio_print_line,
+};
+
+__init static int init_mmio_trace(void)
+{
+	return register_tracer(&mmio_tracer);
+}
+device_initcall(init_mmio_trace);
+
+void mmio_trace_rw(struct mmiotrace_rw *rw)
+{
+	struct trace_array *tr = mmio_trace_array;
+	struct trace_array_cpu *data = tr->data[smp_processor_id()];
+	__trace_mmiotrace_rw(tr, data, rw);
+}
+
+void mmio_trace_mapping(struct mmiotrace_map *map)
+{
+	struct trace_array *tr = mmio_trace_array;
+	struct trace_array_cpu *data;
+
+	preempt_disable();
+	data = tr->data[smp_processor_id()];
+	__trace_mmiotrace_map(tr, data, map);
+	preempt_enable();
+}

kernel/trace/trace_sched_switch.c

@@ -0,0 +1,286 @@
+/*
+ * trace context switch
+ *
+ * Copyright (C) 2007 Steven Rostedt <srostedt@redhat.com>
+ *
+ */
+#include <linux/module.h>
+#include <linux/fs.h>
+#include <linux/debugfs.h>
+#include <linux/kallsyms.h>
+#include <linux/uaccess.h>
+#include <linux/marker.h>
+#include <linux/ftrace.h>
+
+#include "trace.h"
+
+static struct trace_array	*ctx_trace;
+static int __read_mostly	tracer_enabled;
+static atomic_t			sched_ref;
+
+static void
+sched_switch_func(void *private, void *__rq, struct task_struct *prev,
+			struct task_struct *next)
+{
+	struct trace_array **ptr = private;
+	struct trace_array *tr = *ptr;
+	struct trace_array_cpu *data;
+	unsigned long flags;
+	long disabled;
+	int cpu;
+
+	tracing_record_cmdline(prev);
+	tracing_record_cmdline(next);
+
+	if (!tracer_enabled)
+		return;
+
+	local_irq_save(flags);
+	cpu = raw_smp_processor_id();
+	data = tr->data[cpu];
+	disabled = atomic_inc_return(&data->disabled);
+
+	if (likely(disabled == 1))
+		tracing_sched_switch_trace(tr, data, prev, next, flags);
+
+	atomic_dec(&data->disabled);
+	local_irq_restore(flags);
+}
+
+static notrace void
+sched_switch_callback(void *probe_data, void *call_data,
+		      const char *format, va_list *args)
+{
+	struct task_struct *prev;
+	struct task_struct *next;
+	struct rq *__rq;
+
+	if (!atomic_read(&sched_ref))
+		return;
+
+	/* skip prev_pid %d next_pid %d prev_state %ld */
+	(void)va_arg(*args, int);
+	(void)va_arg(*args, int);
+	(void)va_arg(*args, long);
+	__rq = va_arg(*args, typeof(__rq));
+	prev = va_arg(*args, typeof(prev));
+	next = va_arg(*args, typeof(next));
+
+	/*
+	 * If tracer_switch_func only points to the local
+	 * switch func, it still needs the ptr passed to it.
+	 */
+	sched_switch_func(probe_data, __rq, prev, next);
+}
+
+static void
+wakeup_func(void *private, void *__rq, struct task_struct *wakee, struct
+			task_struct *curr)
+{
+	struct trace_array **ptr = private;
+	struct trace_array *tr = *ptr;
+	struct trace_array_cpu *data;
+	unsigned long flags;
+	long disabled;
+	int cpu;
+
+	if (!tracer_enabled)
+		return;
+
+	tracing_record_cmdline(curr);
+
+	local_irq_save(flags);
+	cpu = raw_smp_processor_id();
+	data = tr->data[cpu];
+	disabled = atomic_inc_return(&data->disabled);
+
+	if (likely(disabled == 1))
+		tracing_sched_wakeup_trace(tr, data, wakee, curr, flags);
+
+	atomic_dec(&data->disabled);
+	local_irq_restore(flags);
+}
+
+static notrace void
+wake_up_callback(void *probe_data, void *call_data,
+		 const char *format, va_list *args)
+{
+	struct task_struct *curr;
+	struct task_struct *task;
+	struct rq *__rq;
+
+	if (likely(!tracer_enabled))
+		return;
+
+	/* Skip pid %d state %ld */
+	(void)va_arg(*args, int);
+	(void)va_arg(*args, long);
+	/* now get the meat: "rq %p task %p rq->curr %p" */
+	__rq = va_arg(*args, typeof(__rq));
+	task = va_arg(*args, typeof(task));
+	curr = va_arg(*args, typeof(curr));
+
+	tracing_record_cmdline(task);
+	tracing_record_cmdline(curr);
+
+	wakeup_func(probe_data, __rq, task, curr);
+}
+
+static void sched_switch_reset(struct trace_array *tr)
+{
+	int cpu;
+
+	tr->time_start = ftrace_now(tr->cpu);
+
+	for_each_online_cpu(cpu)
+		tracing_reset(tr->data[cpu]);
+}
+
+static int tracing_sched_register(void)
+{
+	int ret;
+
+	ret = marker_probe_register("kernel_sched_wakeup",
+			"pid %d state %ld ## rq %p task %p rq->curr %p",
+			wake_up_callback,
+			&ctx_trace);
+	if (ret) {
+		pr_info("wakeup trace: Couldn't add marker"
+			" probe to kernel_sched_wakeup\n");
+		return ret;
+	}
+
+	ret = marker_probe_register("kernel_sched_wakeup_new",
+			"pid %d state %ld ## rq %p task %p rq->curr %p",
+			wake_up_callback,
+			&ctx_trace);
+	if (ret) {
+		pr_info("wakeup trace: Couldn't add marker"
+			" probe to kernel_sched_wakeup_new\n");
+		goto fail_deprobe;
+	}
+
+	ret = marker_probe_register("kernel_sched_schedule",
+		"prev_pid %d next_pid %d prev_state %ld "
+		"## rq %p prev %p next %p",
+		sched_switch_callback,
+		&ctx_trace);
+	if (ret) {
+		pr_info("sched trace: Couldn't add marker"
+			" probe to kernel_sched_schedule\n");
+		goto fail_deprobe_wake_new;
+	}
+
+	return ret;
+fail_deprobe_wake_new:
+	marker_probe_unregister("kernel_sched_wakeup_new",
+				wake_up_callback,
+				&ctx_trace);
+fail_deprobe:
+	marker_probe_unregister("kernel_sched_wakeup",
+				wake_up_callback,
+				&ctx_trace);
+	return ret;
+}
+
+static void tracing_sched_unregister(void)
+{
+	marker_probe_unregister("kernel_sched_schedule",
+				sched_switch_callback,
+				&ctx_trace);
+	marker_probe_unregister("kernel_sched_wakeup_new",
+				wake_up_callback,
+				&ctx_trace);
+	marker_probe_unregister("kernel_sched_wakeup",
+				wake_up_callback,
+				&ctx_trace);
+}
+
+static void tracing_start_sched_switch(void)
+{
+	long ref;
+
+	ref = atomic_inc_return(&sched_ref);
+	if (ref == 1)
+		tracing_sched_register();
+}
+
+static void tracing_stop_sched_switch(void)
+{
+	long ref;
+
+	ref = atomic_dec_and_test(&sched_ref);
+	if (ref)
+		tracing_sched_unregister();
+}
+
+void tracing_start_cmdline_record(void)
+{
+	tracing_start_sched_switch();
+}
+
+void tracing_stop_cmdline_record(void)
+{
+	tracing_stop_sched_switch();
+}
+
+static void start_sched_trace(struct trace_array *tr)
+{
+	sched_switch_reset(tr);
+	tracing_start_cmdline_record();
+	tracer_enabled = 1;
+}
+
+static void stop_sched_trace(struct trace_array *tr)
+{
+	tracer_enabled = 0;
+	tracing_stop_cmdline_record();
+}
+
+static void sched_switch_trace_init(struct trace_array *tr)
+{
+	ctx_trace = tr;
+
+	if (tr->ctrl)
+		start_sched_trace(tr);
+}
+
+static void sched_switch_trace_reset(struct trace_array *tr)
+{
+	if (tr->ctrl)
+		stop_sched_trace(tr);
+}
+
+static void sched_switch_trace_ctrl_update(struct trace_array *tr)
+{
+	/* When starting a new trace, reset the buffers */
+	if (tr->ctrl)
+		start_sched_trace(tr);
+	else
+		stop_sched_trace(tr);
+}
+
+static struct tracer sched_switch_trace __read_mostly =
+{
+	.name		= "sched_switch",
+	.init		= sched_switch_trace_init,
+	.reset		= sched_switch_trace_reset,
+	.ctrl_update	= sched_switch_trace_ctrl_update,
+#ifdef CONFIG_FTRACE_SELFTEST
+	.selftest    = trace_selftest_startup_sched_switch,
+#endif
+};
+
+__init static int init_sched_switch_trace(void)
+{
+	int ret = 0;
+
+	if (atomic_read(&sched_ref))
+		ret = tracing_sched_register();
+	if (ret) {
+		pr_info("error registering scheduler trace\n");
+		return ret;
+	}
+	return register_tracer(&sched_switch_trace);
+}
+device_initcall(init_sched_switch_trace);

kernel/trace/trace_sched_wakeup.c

@@ -0,0 +1,448 @@
+/*
+ * trace task wakeup timings
+ *
+ * Copyright (C) 2007-2008 Steven Rostedt <srostedt@redhat.com>
+ * Copyright (C) 2008 Ingo Molnar <mingo@redhat.com>
+ *
+ * Based on code from the latency_tracer, that is:
+ *
+ *  Copyright (C) 2004-2006 Ingo Molnar
+ *  Copyright (C) 2004 William Lee Irwin III
+ */
+#include <linux/module.h>
+#include <linux/fs.h>
+#include <linux/debugfs.h>
+#include <linux/kallsyms.h>
+#include <linux/uaccess.h>
+#include <linux/ftrace.h>
+#include <linux/marker.h>
+
+#include "trace.h"
+
+static struct trace_array	*wakeup_trace;
+static int __read_mostly	tracer_enabled;
+
+static struct task_struct	*wakeup_task;
+static int			wakeup_cpu;
+static unsigned			wakeup_prio = -1;
+
+static DEFINE_SPINLOCK(wakeup_lock);
+
+static void __wakeup_reset(struct trace_array *tr);
+
+#ifdef CONFIG_FTRACE
+/*
+ * irqsoff uses its own tracer function to keep the overhead down:
+ */
+static void
+wakeup_tracer_call(unsigned long ip, unsigned long parent_ip)
+{
+	struct trace_array *tr = wakeup_trace;
+	struct trace_array_cpu *data;
+	unsigned long flags;
+	long disabled;
+	int resched;
+	int cpu;
+
+	if (likely(!wakeup_task))
+		return;
+
+	resched = need_resched();
+	preempt_disable_notrace();
+
+	cpu = raw_smp_processor_id();
+	data = tr->data[cpu];
+	disabled = atomic_inc_return(&data->disabled);
+	if (unlikely(disabled != 1))
+		goto out;
+
+	spin_lock_irqsave(&wakeup_lock, flags);
+
+	if (unlikely(!wakeup_task))
+		goto unlock;
+
+	/*
+	 * The task can't disappear because it needs to
+	 * wake up first, and we have the wakeup_lock.
+	 */
+	if (task_cpu(wakeup_task) != cpu)
+		goto unlock;
+
+	trace_function(tr, data, ip, parent_ip, flags);
+
+ unlock:
+	spin_unlock_irqrestore(&wakeup_lock, flags);
+
+ out:
+	atomic_dec(&data->disabled);
+
+	/*
+	 * To prevent recursion from the scheduler, if the
+	 * resched flag was set before we entered, then
+	 * don't reschedule.
+	 */
+	if (resched)
+		preempt_enable_no_resched_notrace();
+	else
+		preempt_enable_notrace();
+}
+
+static struct ftrace_ops trace_ops __read_mostly =
+{
+	.func = wakeup_tracer_call,
+};
+#endif /* CONFIG_FTRACE */
+
+/*
+ * Should this new latency be reported/recorded?
+ */
+static int report_latency(cycle_t delta)
+{
+	if (tracing_thresh) {
+		if (delta < tracing_thresh)
+			return 0;
+	} else {
+		if (delta <= tracing_max_latency)
+			return 0;
+	}
+	return 1;
+}
+
+static void notrace
+wakeup_sched_switch(void *private, void *rq, struct task_struct *prev,
+	struct task_struct *next)
+{
+	unsigned long latency = 0, t0 = 0, t1 = 0;
+	struct trace_array **ptr = private;
+	struct trace_array *tr = *ptr;
+	struct trace_array_cpu *data;
+	cycle_t T0, T1, delta;
+	unsigned long flags;
+	long disabled;
+	int cpu;
+
+	if (unlikely(!tracer_enabled))
+		return;
+
+	/*
+	 * When we start a new trace, we set wakeup_task to NULL
+	 * and then set tracer_enabled = 1. We want to make sure
+	 * that another CPU does not see the tracer_enabled = 1
+	 * and the wakeup_task with an older task, that might
+	 * actually be the same as next.
+	 */
+	smp_rmb();
+
+	if (next != wakeup_task)
+		return;
+
+	/* The task we are waiting for is waking up */
+	data = tr->data[wakeup_cpu];
+
+	/* disable local data, not wakeup_cpu data */
+	cpu = raw_smp_processor_id();
+	disabled = atomic_inc_return(&tr->data[cpu]->disabled);
+	if (likely(disabled != 1))
+		goto out;
+
+	spin_lock_irqsave(&wakeup_lock, flags);
+
+	/* We could race with grabbing wakeup_lock */
+	if (unlikely(!tracer_enabled || next != wakeup_task))
+		goto out_unlock;
+
+	trace_function(tr, data, CALLER_ADDR1, CALLER_ADDR2, flags);
+
+	/*
+	 * usecs conversion is slow so we try to delay the conversion
+	 * as long as possible:
+	 */
+	T0 = data->preempt_timestamp;
+	T1 = ftrace_now(cpu);
+	delta = T1-T0;
+
+	if (!report_latency(delta))
+		goto out_unlock;
+
+	latency = nsecs_to_usecs(delta);
+
+	tracing_max_latency = delta;
+	t0 = nsecs_to_usecs(T0);
+	t1 = nsecs_to_usecs(T1);
+
+	update_max_tr(tr, wakeup_task, wakeup_cpu);
+
+out_unlock:
+	__wakeup_reset(tr);
+	spin_unlock_irqrestore(&wakeup_lock, flags);
+out:
+	atomic_dec(&tr->data[cpu]->disabled);
+}
+
+static notrace void
+sched_switch_callback(void *probe_data, void *call_data,
+		      const char *format, va_list *args)
+{
+	struct task_struct *prev;
+	struct task_struct *next;
+	struct rq *__rq;
+
+	/* skip prev_pid %d next_pid %d prev_state %ld */
+	(void)va_arg(*args, int);
+	(void)va_arg(*args, int);
+	(void)va_arg(*args, long);
+	__rq = va_arg(*args, typeof(__rq));
+	prev = va_arg(*args, typeof(prev));
+	next = va_arg(*args, typeof(next));
+
+	tracing_record_cmdline(prev);
+
+	/*
+	 * If tracer_switch_func only points to the local
+	 * switch func, it still needs the ptr passed to it.
+	 */
+	wakeup_sched_switch(probe_data, __rq, prev, next);
+}
+
+static void __wakeup_reset(struct trace_array *tr)
+{
+	struct trace_array_cpu *data;
+	int cpu;
+
+	assert_spin_locked(&wakeup_lock);
+
+	for_each_possible_cpu(cpu) {
+		data = tr->data[cpu];
+		tracing_reset(data);
+	}
+
+	wakeup_cpu = -1;
+	wakeup_prio = -1;
+
+	if (wakeup_task)
+		put_task_struct(wakeup_task);
+
+	wakeup_task = NULL;
+}
+
+static void wakeup_reset(struct trace_array *tr)
+{
+	unsigned long flags;
+
+	spin_lock_irqsave(&wakeup_lock, flags);
+	__wakeup_reset(tr);
+	spin_unlock_irqrestore(&wakeup_lock, flags);
+}
+
+static void
+wakeup_check_start(struct trace_array *tr, struct task_struct *p,
+		   struct task_struct *curr)
+{
+	int cpu = smp_processor_id();
+	unsigned long flags;
+	long disabled;
+
+	if (likely(!rt_task(p)) ||
+			p->prio >= wakeup_prio ||
+			p->prio >= curr->prio)
+		return;
+
+	disabled = atomic_inc_return(&tr->data[cpu]->disabled);
+	if (unlikely(disabled != 1))
+		goto out;
+
+	/* interrupts should be off from try_to_wake_up */
+	spin_lock(&wakeup_lock);
+
+	/* check for races. */
+	if (!tracer_enabled || p->prio >= wakeup_prio)
+		goto out_locked;
+
+	/* reset the trace */
+	__wakeup_reset(tr);
+
+	wakeup_cpu = task_cpu(p);
+	wakeup_prio = p->prio;
+
+	wakeup_task = p;
+	get_task_struct(wakeup_task);
+
+	local_save_flags(flags);
+
+	tr->data[wakeup_cpu]->preempt_timestamp = ftrace_now(cpu);
+	trace_function(tr, tr->data[wakeup_cpu],
+		       CALLER_ADDR1, CALLER_ADDR2, flags);
+
+out_locked:
+	spin_unlock(&wakeup_lock);
+out:
+	atomic_dec(&tr->data[cpu]->disabled);
+}
+
+static notrace void
+wake_up_callback(void *probe_data, void *call_data,
+		 const char *format, va_list *args)
+{
+	struct trace_array **ptr = probe_data;
+	struct trace_array *tr = *ptr;
+	struct task_struct *curr;
+	struct task_struct *task;
+	struct rq *__rq;
+
+	if (likely(!tracer_enabled))
+		return;
+
+	/* Skip pid %d state %ld */
+	(void)va_arg(*args, int);
+	(void)va_arg(*args, long);
+	/* now get the meat: "rq %p task %p rq->curr %p" */
+	__rq = va_arg(*args, typeof(__rq));
+	task = va_arg(*args, typeof(task));
+	curr = va_arg(*args, typeof(curr));
+
+	tracing_record_cmdline(task);
+	tracing_record_cmdline(curr);
+
+	wakeup_check_start(tr, task, curr);
+}
+
+static void start_wakeup_tracer(struct trace_array *tr)
+{
+	int ret;
+
+	ret = marker_probe_register("kernel_sched_wakeup",
+			"pid %d state %ld ## rq %p task %p rq->curr %p",
+			wake_up_callback,
+			&wakeup_trace);
+	if (ret) {
+		pr_info("wakeup trace: Couldn't add marker"
+			" probe to kernel_sched_wakeup\n");
+		return;
+	}
+
+	ret = marker_probe_register("kernel_sched_wakeup_new",
+			"pid %d state %ld ## rq %p task %p rq->curr %p",
+			wake_up_callback,
+			&wakeup_trace);
+	if (ret) {
+		pr_info("wakeup trace: Couldn't add marker"
+			" probe to kernel_sched_wakeup_new\n");
+		goto fail_deprobe;
+	}
+
+	ret = marker_probe_register("kernel_sched_schedule",
+		"prev_pid %d next_pid %d prev_state %ld "
+		"## rq %p prev %p next %p",
+		sched_switch_callback,
+		&wakeup_trace);
+	if (ret) {
+		pr_info("sched trace: Couldn't add marker"
+			" probe to kernel_sched_schedule\n");
+		goto fail_deprobe_wake_new;
+	}
+
+	wakeup_reset(tr);
+
+	/*
+	 * Don't let the tracer_enabled = 1 show up before
+	 * the wakeup_task is reset. This may be overkill since
+	 * wakeup_reset does a spin_unlock after setting the
+	 * wakeup_task to NULL, but I want to be safe.
+	 * This is a slow path anyway.
+	 */
+	smp_wmb();
+
+	register_ftrace_function(&trace_ops);
+
+	tracer_enabled = 1;
+
+	return;
+fail_deprobe_wake_new:
+	marker_probe_unregister("kernel_sched_wakeup_new",
+				wake_up_callback,
+				&wakeup_trace);
+fail_deprobe:
+	marker_probe_unregister("kernel_sched_wakeup",
+				wake_up_callback,
+				&wakeup_trace);
+}
+
+static void stop_wakeup_tracer(struct trace_array *tr)
+{
+	tracer_enabled = 0;
+	unregister_ftrace_function(&trace_ops);
+	marker_probe_unregister("kernel_sched_schedule",
+				sched_switch_callback,
+				&wakeup_trace);
+	marker_probe_unregister("kernel_sched_wakeup_new",
+				wake_up_callback,
+				&wakeup_trace);
+	marker_probe_unregister("kernel_sched_wakeup",
+				wake_up_callback,
+				&wakeup_trace);
+}
+
+static void wakeup_tracer_init(struct trace_array *tr)
+{
+	wakeup_trace = tr;
+
+	if (tr->ctrl)
+		start_wakeup_tracer(tr);
+}
+
+static void wakeup_tracer_reset(struct trace_array *tr)
+{
+	if (tr->ctrl) {
+		stop_wakeup_tracer(tr);
+		/* make sure we put back any tasks we are tracing */
+		wakeup_reset(tr);
+	}
+}
+
+static void wakeup_tracer_ctrl_update(struct trace_array *tr)
+{
+	if (tr->ctrl)
+		start_wakeup_tracer(tr);
+	else
+		stop_wakeup_tracer(tr);
+}
+
+static void wakeup_tracer_open(struct trace_iterator *iter)
+{
+	/* stop the trace while dumping */
+	if (iter->tr->ctrl)
+		stop_wakeup_tracer(iter->tr);
+}
+
+static void wakeup_tracer_close(struct trace_iterator *iter)
+{
+	/* forget about any processes we were recording */
+	if (iter->tr->ctrl)
+		start_wakeup_tracer(iter->tr);
+}
+
+static struct tracer wakeup_tracer __read_mostly =
+{
+	.name		= "wakeup",
+	.init		= wakeup_tracer_init,
+	.reset		= wakeup_tracer_reset,
+	.open		= wakeup_tracer_open,
+	.close		= wakeup_tracer_close,
+	.ctrl_update	= wakeup_tracer_ctrl_update,
+	.print_max	= 1,
+#ifdef CONFIG_FTRACE_SELFTEST
+	.selftest    = trace_selftest_startup_wakeup,
+#endif
+};
+
+__init static int init_wakeup_tracer(void)
+{
+	int ret;
+
+	ret = register_tracer(&wakeup_tracer);
+	if (ret)
+		return ret;
+
+	return 0;
+}
+device_initcall(init_wakeup_tracer);

kernel/trace/trace_selftest.c

@@ -0,0 +1,563 @@
+/* Include in trace.c */
+
+#include <linux/kthread.h>
+#include <linux/delay.h>
+
+static inline int trace_valid_entry(struct trace_entry *entry)
+{
+	switch (entry->type) {
+	case TRACE_FN:
+	case TRACE_CTX:
+	case TRACE_WAKE:
+	case TRACE_STACK:
+	case TRACE_SPECIAL:
+		return 1;
+	}
+	return 0;
+}
+
+static int
+trace_test_buffer_cpu(struct trace_array *tr, struct trace_array_cpu *data)
+{
+	struct trace_entry *entries;
+	struct page *page;
+	int idx = 0;
+	int i;
+
+	BUG_ON(list_empty(&data->trace_pages));
+	page = list_entry(data->trace_pages.next, struct page, lru);
+	entries = page_address(page);
+
+	check_pages(data);
+	if (head_page(data) != entries)
+		goto failed;
+
+	/*
+	 * The starting trace buffer always has valid elements,
+	 * if any element exists.
+	 */
+	entries = head_page(data);
+
+	for (i = 0; i < tr->entries; i++) {
+
+		if (i < data->trace_idx && !trace_valid_entry(&entries[idx])) {
+			printk(KERN_CONT ".. invalid entry %d ",
+				entries[idx].type);
+			goto failed;
+		}
+
+		idx++;
+		if (idx >= ENTRIES_PER_PAGE) {
+			page = virt_to_page(entries);
+			if (page->lru.next == &data->trace_pages) {
+				if (i != tr->entries - 1) {
+					printk(KERN_CONT ".. entries buffer mismatch");
+					goto failed;
+				}
+			} else {
+				page = list_entry(page->lru.next, struct page, lru);
+				entries = page_address(page);
+			}
+			idx = 0;
+		}
+	}
+
+	page = virt_to_page(entries);
+	if (page->lru.next != &data->trace_pages) {
+		printk(KERN_CONT ".. too many entries");
+		goto failed;
+	}
+
+	return 0;
+
+ failed:
+	/* disable tracing */
+	tracing_disabled = 1;
+	printk(KERN_CONT ".. corrupted trace buffer .. ");
+	return -1;
+}
+
+/*
+ * Test the trace buffer to see if all the elements
+ * are still sane.
+ */
+static int trace_test_buffer(struct trace_array *tr, unsigned long *count)
+{
+	unsigned long flags, cnt = 0;
+	int cpu, ret = 0;
+
+	/* Don't allow flipping of max traces now */
+	raw_local_irq_save(flags);
+	__raw_spin_lock(&ftrace_max_lock);
+	for_each_possible_cpu(cpu) {
+		if (!head_page(tr->data[cpu]))
+			continue;
+
+		cnt += tr->data[cpu]->trace_idx;
+
+		ret = trace_test_buffer_cpu(tr, tr->data[cpu]);
+		if (ret)
+			break;
+	}
+	__raw_spin_unlock(&ftrace_max_lock);
+	raw_local_irq_restore(flags);
+
+	if (count)
+		*count = cnt;
+
+	return ret;
+}
+
+#ifdef CONFIG_FTRACE
+
+#ifdef CONFIG_DYNAMIC_FTRACE
+
+#define __STR(x) #x
+#define STR(x) __STR(x)
+
+/* Test dynamic code modification and ftrace filters */
+int trace_selftest_startup_dynamic_tracing(struct tracer *trace,
+					   struct trace_array *tr,
+					   int (*func)(void))
+{
+	unsigned long count;
+	int ret;
+	int save_ftrace_enabled = ftrace_enabled;
+	int save_tracer_enabled = tracer_enabled;
+	char *func_name;
+
+	/* The ftrace test PASSED */
+	printk(KERN_CONT "PASSED\n");
+	pr_info("Testing dynamic ftrace: ");
+
+	/* enable tracing, and record the filter function */
+	ftrace_enabled = 1;
+	tracer_enabled = 1;
+
+	/* passed in by parameter to fool gcc from optimizing */
+	func();
+
+	/* update the records */
+	ret = ftrace_force_update();
+	if (ret) {
+		printk(KERN_CONT ".. ftraced failed .. ");
+		return ret;
+	}
+
+	/*
+	 * Some archs *cough*PowerPC*cough* add charachters to the
+	 * start of the function names. We simply put a '*' to
+	 * accomodate them.
+	 */
+	func_name = "*" STR(DYN_FTRACE_TEST_NAME);
+
+	/* filter only on our function */
+	ftrace_set_filter(func_name, strlen(func_name), 1);
+
+	/* enable tracing */
+	tr->ctrl = 1;
+	trace->init(tr);
+	/* Sleep for a 1/10 of a second */
+	msleep(100);
+
+	/* we should have nothing in the buffer */
+	ret = trace_test_buffer(tr, &count);
+	if (ret)
+		goto out;
+
+	if (count) {
+		ret = -1;
+		printk(KERN_CONT ".. filter did not filter .. ");
+		goto out;
+	}
+
+	/* call our function again */
+	func();
+
+	/* sleep again */
+	msleep(100);
+
+	/* stop the tracing. */
+	tr->ctrl = 0;
+	trace->ctrl_update(tr);
+	ftrace_enabled = 0;
+
+	/* check the trace buffer */
+	ret = trace_test_buffer(tr, &count);
+	trace->reset(tr);
+
+	/* we should only have one item */
+	if (!ret && count != 1) {
+		printk(KERN_CONT ".. filter failed count=%ld ..", count);
+		ret = -1;
+		goto out;
+	}
+ out:
+	ftrace_enabled = save_ftrace_enabled;
+	tracer_enabled = save_tracer_enabled;
+
+	/* Enable tracing on all functions again */
+	ftrace_set_filter(NULL, 0, 1);
+
+	return ret;
+}
+#else
+# define trace_selftest_startup_dynamic_tracing(trace, tr, func) ({ 0; })
+#endif /* CONFIG_DYNAMIC_FTRACE */
+/*
+ * Simple verification test of ftrace function tracer.
+ * Enable ftrace, sleep 1/10 second, and then read the trace
+ * buffer to see if all is in order.
+ */
+int
+trace_selftest_startup_function(struct tracer *trace, struct trace_array *tr)
+{
+	unsigned long count;
+	int ret;
+	int save_ftrace_enabled = ftrace_enabled;
+	int save_tracer_enabled = tracer_enabled;
+
+	/* make sure msleep has been recorded */
+	msleep(1);
+
+	/* force the recorded functions to be traced */
+	ret = ftrace_force_update();
+	if (ret) {
+		printk(KERN_CONT ".. ftraced failed .. ");
+		return ret;
+	}
+
+	/* start the tracing */
+	ftrace_enabled = 1;
+	tracer_enabled = 1;
+
+	tr->ctrl = 1;
+	trace->init(tr);
+	/* Sleep for a 1/10 of a second */
+	msleep(100);
+	/* stop the tracing. */
+	tr->ctrl = 0;
+	trace->ctrl_update(tr);
+	ftrace_enabled = 0;
+
+	/* check the trace buffer */
+	ret = trace_test_buffer(tr, &count);
+	trace->reset(tr);
+
+	if (!ret && !count) {
+		printk(KERN_CONT ".. no entries found ..");
+		ret = -1;
+		goto out;
+	}
+
+	ret = trace_selftest_startup_dynamic_tracing(trace, tr,
+						     DYN_FTRACE_TEST_NAME);
+
+ out:
+	ftrace_enabled = save_ftrace_enabled;
+	tracer_enabled = save_tracer_enabled;
+
+	/* kill ftrace totally if we failed */
+	if (ret)
+		ftrace_kill();
+
+	return ret;
+}
+#endif /* CONFIG_FTRACE */
+
+#ifdef CONFIG_IRQSOFF_TRACER
+int
+trace_selftest_startup_irqsoff(struct tracer *trace, struct trace_array *tr)
+{
+	unsigned long save_max = tracing_max_latency;
+	unsigned long count;
+	int ret;
+
+	/* start the tracing */
+	tr->ctrl = 1;
+	trace->init(tr);
+	/* reset the max latency */
+	tracing_max_latency = 0;
+	/* disable interrupts for a bit */
+	local_irq_disable();
+	udelay(100);
+	local_irq_enable();
+	/* stop the tracing. */
+	tr->ctrl = 0;
+	trace->ctrl_update(tr);
+	/* check both trace buffers */
+	ret = trace_test_buffer(tr, NULL);
+	if (!ret)
+		ret = trace_test_buffer(&max_tr, &count);
+	trace->reset(tr);
+
+	if (!ret && !count) {
+		printk(KERN_CONT ".. no entries found ..");
+		ret = -1;
+	}
+
+	tracing_max_latency = save_max;
+
+	return ret;
+}
+#endif /* CONFIG_IRQSOFF_TRACER */
+
+#ifdef CONFIG_PREEMPT_TRACER
+int
+trace_selftest_startup_preemptoff(struct tracer *trace, struct trace_array *tr)
+{
+	unsigned long save_max = tracing_max_latency;
+	unsigned long count;
+	int ret;
+
+	/* start the tracing */
+	tr->ctrl = 1;
+	trace->init(tr);
+	/* reset the max latency */
+	tracing_max_latency = 0;
+	/* disable preemption for a bit */
+	preempt_disable();
+	udelay(100);
+	preempt_enable();
+	/* stop the tracing. */
+	tr->ctrl = 0;
+	trace->ctrl_update(tr);
+	/* check both trace buffers */
+	ret = trace_test_buffer(tr, NULL);
+	if (!ret)
+		ret = trace_test_buffer(&max_tr, &count);
+	trace->reset(tr);
+
+	if (!ret && !count) {
+		printk(KERN_CONT ".. no entries found ..");
+		ret = -1;
+	}
+
+	tracing_max_latency = save_max;
+
+	return ret;
+}
+#endif /* CONFIG_PREEMPT_TRACER */
+
+#if defined(CONFIG_IRQSOFF_TRACER) && defined(CONFIG_PREEMPT_TRACER)
+int
+trace_selftest_startup_preemptirqsoff(struct tracer *trace, struct trace_array *tr)
+{
+	unsigned long save_max = tracing_max_latency;
+	unsigned long count;
+	int ret;
+
+	/* start the tracing */
+	tr->ctrl = 1;
+	trace->init(tr);
+
+	/* reset the max latency */
+	tracing_max_latency = 0;
+
+	/* disable preemption and interrupts for a bit */
+	preempt_disable();
+	local_irq_disable();
+	udelay(100);
+	preempt_enable();
+	/* reverse the order of preempt vs irqs */
+	local_irq_enable();
+
+	/* stop the tracing. */
+	tr->ctrl = 0;
+	trace->ctrl_update(tr);
+	/* check both trace buffers */
+	ret = trace_test_buffer(tr, NULL);
+	if (ret)
+		goto out;
+
+	ret = trace_test_buffer(&max_tr, &count);
+	if (ret)
+		goto out;
+
+	if (!ret && !count) {
+		printk(KERN_CONT ".. no entries found ..");
+		ret = -1;
+		goto out;
+	}
+
+	/* do the test by disabling interrupts first this time */
+	tracing_max_latency = 0;
+	tr->ctrl = 1;
+	trace->ctrl_update(tr);
+	preempt_disable();
+	local_irq_disable();
+	udelay(100);
+	preempt_enable();
+	/* reverse the order of preempt vs irqs */
+	local_irq_enable();
+
+	/* stop the tracing. */
+	tr->ctrl = 0;
+	trace->ctrl_update(tr);
+	/* check both trace buffers */
+	ret = trace_test_buffer(tr, NULL);
+	if (ret)
+		goto out;
+
+	ret = trace_test_buffer(&max_tr, &count);
+
+	if (!ret && !count) {
+		printk(KERN_CONT ".. no entries found ..");
+		ret = -1;
+		goto out;
+	}
+
+ out:
+	trace->reset(tr);
+	tracing_max_latency = save_max;
+
+	return ret;
+}
+#endif /* CONFIG_IRQSOFF_TRACER && CONFIG_PREEMPT_TRACER */
+
+#ifdef CONFIG_SCHED_TRACER
+static int trace_wakeup_test_thread(void *data)
+{
+	/* Make this a RT thread, doesn't need to be too high */
+	struct sched_param param = { .sched_priority = 5 };
+	struct completion *x = data;
+
+	sched_setscheduler(current, SCHED_FIFO, &param);
+
+	/* Make it know we have a new prio */
+	complete(x);
+
+	/* now go to sleep and let the test wake us up */
+	set_current_state(TASK_INTERRUPTIBLE);
+	schedule();
+
+	/* we are awake, now wait to disappear */
+	while (!kthread_should_stop()) {
+		/*
+		 * This is an RT task, do short sleeps to let
+		 * others run.
+		 */
+		msleep(100);
+	}
+
+	return 0;
+}
+
+int
+trace_selftest_startup_wakeup(struct tracer *trace, struct trace_array *tr)
+{
+	unsigned long save_max = tracing_max_latency;
+	struct task_struct *p;
+	struct completion isrt;
+	unsigned long count;
+	int ret;
+
+	init_completion(&isrt);
+
+	/* create a high prio thread */
+	p = kthread_run(trace_wakeup_test_thread, &isrt, "ftrace-test");
+	if (IS_ERR(p)) {
+		printk(KERN_CONT "Failed to create ftrace wakeup test thread ");
+		return -1;
+	}
+
+	/* make sure the thread is running at an RT prio */
+	wait_for_completion(&isrt);
+
+	/* start the tracing */
+	tr->ctrl = 1;
+	trace->init(tr);
+	/* reset the max latency */
+	tracing_max_latency = 0;
+
+	/* sleep to let the RT thread sleep too */
+	msleep(100);
+
+	/*
+	 * Yes this is slightly racy. It is possible that for some
+	 * strange reason that the RT thread we created, did not
+	 * call schedule for 100ms after doing the completion,
+	 * and we do a wakeup on a task that already is awake.
+	 * But that is extremely unlikely, and the worst thing that
+	 * happens in such a case, is that we disable tracing.
+	 * Honestly, if this race does happen something is horrible
+	 * wrong with the system.
+	 */
+
+	wake_up_process(p);
+
+	/* stop the tracing. */
+	tr->ctrl = 0;
+	trace->ctrl_update(tr);
+	/* check both trace buffers */
+	ret = trace_test_buffer(tr, NULL);
+	if (!ret)
+		ret = trace_test_buffer(&max_tr, &count);
+
+
+	trace->reset(tr);
+
+	tracing_max_latency = save_max;
+
+	/* kill the thread */
+	kthread_stop(p);
+
+	if (!ret && !count) {
+		printk(KERN_CONT ".. no entries found ..");
+		ret = -1;
+	}
+
+	return ret;
+}
+#endif /* CONFIG_SCHED_TRACER */
+
+#ifdef CONFIG_CONTEXT_SWITCH_TRACER
+int
+trace_selftest_startup_sched_switch(struct tracer *trace, struct trace_array *tr)
+{
+	unsigned long count;
+	int ret;
+
+	/* start the tracing */
+	tr->ctrl = 1;
+	trace->init(tr);
+	/* Sleep for a 1/10 of a second */
+	msleep(100);
+	/* stop the tracing. */
+	tr->ctrl = 0;
+	trace->ctrl_update(tr);
+	/* check the trace buffer */
+	ret = trace_test_buffer(tr, &count);
+	trace->reset(tr);
+
+	if (!ret && !count) {
+		printk(KERN_CONT ".. no entries found ..");
+		ret = -1;
+	}
+
+	return ret;
+}
+#endif /* CONFIG_CONTEXT_SWITCH_TRACER */
+
+#ifdef CONFIG_SYSPROF_TRACER
+int
+trace_selftest_startup_sysprof(struct tracer *trace, struct trace_array *tr)
+{
+	unsigned long count;
+	int ret;
+
+	/* start the tracing */
+	tr->ctrl = 1;
+	trace->init(tr);
+	/* Sleep for a 1/10 of a second */
+	msleep(100);
+	/* stop the tracing. */
+	tr->ctrl = 0;
+	trace->ctrl_update(tr);
+	/* check the trace buffer */
+	ret = trace_test_buffer(tr, &count);
+	trace->reset(tr);
+
+	return ret;
+}
+#endif /* CONFIG_SYSPROF_TRACER */

kernel/trace/trace_selftest_dynamic.c

@@ -0,0 +1,7 @@
+#include "trace.h"
+
+int DYN_FTRACE_TEST_NAME(void)
+{
+	/* used to call mcount */
+	return 0;
+}

kernel/trace/trace_sysprof.c

@@ -0,0 +1,363 @@
+/*
+ * trace stack traces
+ *
+ * Copyright (C) 2004-2008, Soeren Sandmann
+ * Copyright (C) 2007 Steven Rostedt <srostedt@redhat.com>
+ * Copyright (C) 2008 Ingo Molnar <mingo@redhat.com>
+ */
+#include <linux/kallsyms.h>
+#include <linux/debugfs.h>
+#include <linux/hrtimer.h>
+#include <linux/uaccess.h>
+#include <linux/ftrace.h>
+#include <linux/module.h>
+#include <linux/irq.h>
+#include <linux/fs.h>
+
+#include <asm/stacktrace.h>
+
+#include "trace.h"
+
+static struct trace_array	*sysprof_trace;
+static int __read_mostly	tracer_enabled;
+
+/*
+ * 1 msec sample interval by default:
+ */
+static unsigned long sample_period = 1000000;
+static const unsigned int sample_max_depth = 512;
+
+static DEFINE_MUTEX(sample_timer_lock);
+/*
+ * Per CPU hrtimers that do the profiling:
+ */
+static DEFINE_PER_CPU(struct hrtimer, stack_trace_hrtimer);
+
+struct stack_frame {
+	const void __user	*next_fp;
+	unsigned long		return_address;
+};
+
+static int copy_stack_frame(const void __user *fp, struct stack_frame *frame)
+{
+	int ret;
+
+	if (!access_ok(VERIFY_READ, fp, sizeof(*frame)))
+		return 0;
+
+	ret = 1;
+	pagefault_disable();
+	if (__copy_from_user_inatomic(frame, fp, sizeof(*frame)))
+		ret = 0;
+	pagefault_enable();
+
+	return ret;
+}
+
+struct backtrace_info {
+	struct trace_array_cpu	*data;
+	struct trace_array	*tr;
+	int			pos;
+};
+
+static void
+backtrace_warning_symbol(void *data, char *msg, unsigned long symbol)
+{
+	/* Ignore warnings */
+}
+
+static void backtrace_warning(void *data, char *msg)
+{
+	/* Ignore warnings */
+}
+
+static int backtrace_stack(void *data, char *name)
+{
+	/* Don't bother with IRQ stacks for now */
+	return -1;
+}
+
+static void backtrace_address(void *data, unsigned long addr, int reliable)
+{
+	struct backtrace_info *info = data;
+
+	if (info->pos < sample_max_depth && reliable) {
+		__trace_special(info->tr, info->data, 1, addr, 0);
+
+		info->pos++;
+	}
+}
+
+const static struct stacktrace_ops backtrace_ops = {
+	.warning		= backtrace_warning,
+	.warning_symbol		= backtrace_warning_symbol,
+	.stack			= backtrace_stack,
+	.address		= backtrace_address,
+};
+
+static int
+trace_kernel(struct pt_regs *regs, struct trace_array *tr,
+	     struct trace_array_cpu *data)
+{
+	struct backtrace_info info;
+	unsigned long bp;
+	char *stack;
+
+	info.tr = tr;
+	info.data = data;
+	info.pos = 1;
+
+	__trace_special(info.tr, info.data, 1, regs->ip, 0);
+
+	stack = ((char *)regs + sizeof(struct pt_regs));
+#ifdef CONFIG_FRAME_POINTER
+	bp = regs->bp;
+#else
+	bp = 0;
+#endif
+
+	dump_trace(NULL, regs, (void *)stack, bp, &backtrace_ops, &info);
+
+	return info.pos;
+}
+
+static void timer_notify(struct pt_regs *regs, int cpu)
+{
+	struct trace_array_cpu *data;
+	struct stack_frame frame;
+	struct trace_array *tr;
+	const void __user *fp;
+	int is_user;
+	int i;
+
+	if (!regs)
+		return;
+
+	tr = sysprof_trace;
+	data = tr->data[cpu];
+	is_user = user_mode(regs);
+
+	if (!current || current->pid == 0)
+		return;
+
+	if (is_user && current->state != TASK_RUNNING)
+		return;
+
+	__trace_special(tr, data, 0, 0, current->pid);
+
+	if (!is_user)
+		i = trace_kernel(regs, tr, data);
+	else
+		i = 0;
+
+	/*
+	 * Trace user stack if we are not a kernel thread
+	 */
+	if (current->mm && i < sample_max_depth) {
+		regs = (struct pt_regs *)current->thread.sp0 - 1;
+
+		fp = (void __user *)regs->bp;
+
+		__trace_special(tr, data, 2, regs->ip, 0);
+
+		while (i < sample_max_depth) {
+			frame.next_fp = 0;
+			frame.return_address = 0;
+			if (!copy_stack_frame(fp, &frame))
+				break;
+			if ((unsigned long)fp < regs->sp)
+				break;
+
+			__trace_special(tr, data, 2, frame.return_address,
+					(unsigned long)fp);
+			fp = frame.next_fp;
+
+			i++;
+		}
+
+	}
+
+	/*
+	 * Special trace entry if we overflow the max depth:
+	 */
+	if (i == sample_max_depth)
+		__trace_special(tr, data, -1, -1, -1);
+
+	__trace_special(tr, data, 3, current->pid, i);
+}
+
+static enum hrtimer_restart stack_trace_timer_fn(struct hrtimer *hrtimer)
+{
+	/* trace here */
+	timer_notify(get_irq_regs(), smp_processor_id());
+
+	hrtimer_forward_now(hrtimer, ns_to_ktime(sample_period));
+
+	return HRTIMER_RESTART;
+}
+
+static void start_stack_timer(int cpu)
+{
+	struct hrtimer *hrtimer = &per_cpu(stack_trace_hrtimer, cpu);
+
+	hrtimer_init(hrtimer, CLOCK_MONOTONIC, HRTIMER_MODE_REL);
+	hrtimer->function = stack_trace_timer_fn;
+	hrtimer->cb_mode = HRTIMER_CB_IRQSAFE_NO_SOFTIRQ;
+
+	hrtimer_start(hrtimer, ns_to_ktime(sample_period), HRTIMER_MODE_REL);
+}
+
+static void start_stack_timers(void)
+{
+	cpumask_t saved_mask = current->cpus_allowed;
+	int cpu;
+
+	for_each_online_cpu(cpu) {
+		set_cpus_allowed_ptr(current, &cpumask_of_cpu(cpu));
+		start_stack_timer(cpu);
+	}
+	set_cpus_allowed_ptr(current, &saved_mask);
+}
+
+static void stop_stack_timer(int cpu)
+{
+	struct hrtimer *hrtimer = &per_cpu(stack_trace_hrtimer, cpu);
+
+	hrtimer_cancel(hrtimer);
+}
+
+static void stop_stack_timers(void)
+{
+	int cpu;
+
+	for_each_online_cpu(cpu)
+		stop_stack_timer(cpu);
+}
+
+static void stack_reset(struct trace_array *tr)
+{
+	int cpu;
+
+	tr->time_start = ftrace_now(tr->cpu);
+
+	for_each_online_cpu(cpu)
+		tracing_reset(tr->data[cpu]);
+}
+
+static void start_stack_trace(struct trace_array *tr)
+{
+	mutex_lock(&sample_timer_lock);
+	stack_reset(tr);
+	start_stack_timers();
+	tracer_enabled = 1;
+	mutex_unlock(&sample_timer_lock);
+}
+
+static void stop_stack_trace(struct trace_array *tr)
+{
+	mutex_lock(&sample_timer_lock);
+	stop_stack_timers();
+	tracer_enabled = 0;
+	mutex_unlock(&sample_timer_lock);
+}
+
+static void stack_trace_init(struct trace_array *tr)
+{
+	sysprof_trace = tr;
+
+	if (tr->ctrl)
+		start_stack_trace(tr);
+}
+
+static void stack_trace_reset(struct trace_array *tr)
+{
+	if (tr->ctrl)
+		stop_stack_trace(tr);
+}
+
+static void stack_trace_ctrl_update(struct trace_array *tr)
+{
+	/* When starting a new trace, reset the buffers */
+	if (tr->ctrl)
+		start_stack_trace(tr);
+	else
+		stop_stack_trace(tr);
+}
+
+static struct tracer stack_trace __read_mostly =
+{
+	.name		= "sysprof",
+	.init		= stack_trace_init,
+	.reset		= stack_trace_reset,
+	.ctrl_update	= stack_trace_ctrl_update,
+#ifdef CONFIG_FTRACE_SELFTEST
+	.selftest    = trace_selftest_startup_sysprof,
+#endif
+};
+
+__init static int init_stack_trace(void)
+{
+	return register_tracer(&stack_trace);
+}
+device_initcall(init_stack_trace);
+
+#define MAX_LONG_DIGITS 22
+
+static ssize_t
+sysprof_sample_read(struct file *filp, char __user *ubuf,
+		    size_t cnt, loff_t *ppos)
+{
+	char buf[MAX_LONG_DIGITS];
+	int r;
+
+	r = sprintf(buf, "%ld\n", nsecs_to_usecs(sample_period));
+
+	return simple_read_from_buffer(ubuf, cnt, ppos, buf, r);
+}
+
+static ssize_t
+sysprof_sample_write(struct file *filp, const char __user *ubuf,
+		     size_t cnt, loff_t *ppos)
+{
+	char buf[MAX_LONG_DIGITS];
+	unsigned long val;
+
+	if (cnt > MAX_LONG_DIGITS-1)
+		cnt = MAX_LONG_DIGITS-1;
+
+	if (copy_from_user(&buf, ubuf, cnt))
+		return -EFAULT;
+
+	buf[cnt] = 0;
+
+	val = simple_strtoul(buf, NULL, 10);
+	/*
+	 * Enforce a minimum sample period of 100 usecs:
+	 */
+	if (val < 100)
+		val = 100;
+
+	mutex_lock(&sample_timer_lock);
+	stop_stack_timers();
+	sample_period = val * 1000;
+	start_stack_timers();
+	mutex_unlock(&sample_timer_lock);
+
+	return cnt;
+}
+
+static struct file_operations sysprof_sample_fops = {
+	.read		= sysprof_sample_read,
+	.write		= sysprof_sample_write,
+};
+
+void init_tracer_sysprof_debugfs(struct dentry *d_tracer)
+{
+	struct dentry *entry;
+
+	entry = debugfs_create_file("sysprof_sample_period", 0644,
+			d_tracer, NULL, &sysprof_sample_fops);
+	if (entry)
+		return;
+	pr_warning("Could not create debugfs 'dyn_ftrace_total_info' entry\n");
+}

lib/Kconfig.debug

@@ -634,6 +634,8 @@ config LATENCYTOP
 	  Enable this option if you want to use the LatencyTOP tool
 	  to find out which userspace is blocking on what kernel operations.
 
+source kernel/trace/Kconfig
+
 config PROVIDE_OHCI1394_DMA_INIT
 	bool "Remote debugging over FireWire early on boot"
 	depends on PCI && X86

lib/Makefile

@@ -8,6 +8,15 @@ lib-y := ctype.o string.o vsprintf.o cmdline.o \
 	 sha1.o irq_regs.o reciprocal_div.o argv_split.o \
 	 proportions.o prio_heap.o ratelimit.o
 
+ifdef CONFIG_FTRACE
+# Do not profile string.o, since it may be used in early boot or vdso
+CFLAGS_REMOVE_string.o = -pg
+# Also do not profile any debug utilities
+CFLAGS_REMOVE_spinlock_debug.o = -pg
+CFLAGS_REMOVE_list_debug.o = -pg
+CFLAGS_REMOVE_debugobjects.o = -pg
+endif
+
 lib-$(CONFIG_MMU) += ioremap.o
 lib-$(CONFIG_SMP) += cpumask.o
 

lib/smp_processor_id.c

@@ -7,7 +7,7 @@
 #include <linux/kallsyms.h>
 #include <linux/sched.h>
 
-unsigned int debug_smp_processor_id(void)
+notrace unsigned int debug_smp_processor_id(void)
 {
 	unsigned long preempt_count = preempt_count();
 	int this_cpu = raw_smp_processor_id();
@@ -37,7 +37,7 @@ unsigned int debug_smp_processor_id(void)
 	/*
 	 * Avoid recursion:
 	 */
-	preempt_disable();
+	preempt_disable_notrace();
 
 	if (!printk_ratelimit())
 		goto out_enable;
@@ -49,7 +49,7 @@ unsigned int debug_smp_processor_id(void)
 	dump_stack();
 
 out_enable:
-	preempt_enable_no_resched();
+	preempt_enable_no_resched_notrace();
 out:
 	return this_cpu;
 }

mm/page-writeback.c

@@ -126,8 +126,6 @@ static void background_writeout(unsigned long _min_pages);
 static struct prop_descriptor vm_completions;
 static struct prop_descriptor vm_dirties;
 
-static unsigned long determine_dirtyable_memory(void);
-
 /*
  * couple the period to the dirty_ratio:
  *
@@ -347,7 +345,13 @@ static unsigned long highmem_dirtyable_memory(unsigned long total)
 #endif
 }
 
-static unsigned long determine_dirtyable_memory(void)
+/**
+ * determine_dirtyable_memory - amount of memory that may be used
+ *
+ * Returns the numebr of pages that can currently be freed and used
+ * by the kernel for direct mappings.
+ */
+unsigned long determine_dirtyable_memory(void)
 {
 	unsigned long x;
 

scripts/Makefile.lib

@@ -96,7 +96,8 @@ basename_flags = -D"KBUILD_BASENAME=KBUILD_STR($(call name-fix,$(basetarget)))"
 modname_flags  = $(if $(filter 1,$(words $(modname))),\
                  -D"KBUILD_MODNAME=KBUILD_STR($(call name-fix,$(modname)))")
 
-_c_flags       = $(KBUILD_CFLAGS) $(ccflags-y) $(CFLAGS_$(basetarget).o)
+orig_c_flags   = $(KBUILD_CFLAGS) $(ccflags-y) $(CFLAGS_$(basetarget).o)
+_c_flags       = $(filter-out $(CFLAGS_REMOVE_$(basetarget).o), $(orig_c_flags))
 _a_flags       = $(KBUILD_AFLAGS) $(asflags-y) $(AFLAGS_$(basetarget).o)
 _cpp_flags     = $(KBUILD_CPPFLAGS) $(cppflags-y) $(CPPFLAGS_$(@F))