Merge branches 'tracing/branch-tracer', 'tracing/fastboot', 'tracing/function-return-tracer' and 'tracing/urgent' into tracing/core
diff --git a/arch/x86/kernel/vsyscall_64.c b/arch/x86/kernel/vsyscall_64.c
index 0b8b6690..6f3d3d4 100644
--- a/arch/x86/kernel/vsyscall_64.c
+++ b/arch/x86/kernel/vsyscall_64.c
@@ -17,6 +17,9 @@
* want per guest time just set the kernel.vsyscall64 sysctl to 0.
*/
+/* Disable profiling for userspace code: */
+#define DISABLE_BRANCH_PROFILING
+
#include <linux/time.h>
#include <linux/init.h>
#include <linux/kernel.h>
diff --git a/arch/x86/vdso/vclock_gettime.c b/arch/x86/vdso/vclock_gettime.c
index 1ef0f90..d9d3582 100644
--- a/arch/x86/vdso/vclock_gettime.c
+++ b/arch/x86/vdso/vclock_gettime.c
@@ -9,6 +9,9 @@
* Also alternative() doesn't work.
*/
+/* Disable profiling for userspace code: */
+#define DISABLE_BRANCH_PROFILING
+
#include <linux/kernel.h>
#include <linux/posix-timers.h>
#include <linux/time.h>
diff --git a/include/asm-generic/vmlinux.lds.h b/include/asm-generic/vmlinux.lds.h
index 80744606..a5e4ed9 100644
--- a/include/asm-generic/vmlinux.lds.h
+++ b/include/asm-generic/vmlinux.lds.h
@@ -45,6 +45,17 @@
#define MCOUNT_REC()
#endif
+#ifdef CONFIG_TRACE_BRANCH_PROFILING
+#define LIKELY_PROFILE() VMLINUX_SYMBOL(__start_likely_profile) = .; \
+ *(_ftrace_likely) \
+ VMLINUX_SYMBOL(__stop_likely_profile) = .; \
+ VMLINUX_SYMBOL(__start_unlikely_profile) = .; \
+ *(_ftrace_unlikely) \
+ VMLINUX_SYMBOL(__stop_unlikely_profile) = .;
+#else
+#define LIKELY_PROFILE()
+#endif
+
/* .data section */
#define DATA_DATA \
*(.data) \
@@ -62,7 +73,8 @@
VMLINUX_SYMBOL(__stop___markers) = .; \
VMLINUX_SYMBOL(__start___tracepoints) = .; \
*(__tracepoints) \
- VMLINUX_SYMBOL(__stop___tracepoints) = .;
+ VMLINUX_SYMBOL(__stop___tracepoints) = .; \
+ LIKELY_PROFILE()
#define RO_DATA(align) \
. = ALIGN((align)); \
diff --git a/include/linux/compiler.h b/include/linux/compiler.h
index 98115d9..c7d804a 100644
--- a/include/linux/compiler.h
+++ b/include/linux/compiler.h
@@ -59,8 +59,70 @@
* specific implementations come from the above header files
*/
-#define likely(x) __builtin_expect(!!(x), 1)
-#define unlikely(x) __builtin_expect(!!(x), 0)
+struct ftrace_branch_data {
+ const char *func;
+ const char *file;
+ unsigned line;
+ unsigned long correct;
+ unsigned long incorrect;
+};
+
+/*
+ * Note: DISABLE_BRANCH_PROFILING can be used by special lowlevel code
+ * to disable branch tracing on a per file basis.
+ */
+#if defined(CONFIG_TRACE_BRANCH_PROFILING) && !defined(DISABLE_BRANCH_PROFILING)
+void ftrace_likely_update(struct ftrace_branch_data *f, int val, int expect);
+
+#define likely_notrace(x) __builtin_expect(!!(x), 1)
+#define unlikely_notrace(x) __builtin_expect(!!(x), 0)
+
+#define likely_check(x) ({ \
+ int ______r; \
+ static struct ftrace_branch_data \
+ __attribute__((__aligned__(4))) \
+ __attribute__((section("_ftrace_likely"))) \
+ ______f = { \
+ .func = __func__, \
+ .file = __FILE__, \
+ .line = __LINE__, \
+ }; \
+ ______f.line = __LINE__; \
+ ______r = likely_notrace(x); \
+ ftrace_likely_update(&______f, ______r, 1); \
+ ______r; \
+ })
+#define unlikely_check(x) ({ \
+ int ______r; \
+ static struct ftrace_branch_data \
+ __attribute__((__aligned__(4))) \
+ __attribute__((section("_ftrace_unlikely"))) \
+ ______f = { \
+ .func = __func__, \
+ .file = __FILE__, \
+ .line = __LINE__, \
+ }; \
+ ______f.line = __LINE__; \
+ ______r = unlikely_notrace(x); \
+ ftrace_likely_update(&______f, ______r, 0); \
+ ______r; \
+ })
+
+/*
+ * Using __builtin_constant_p(x) to ignore cases where the return
+ * value is always the same. This idea is taken from a similar patch
+ * written by Daniel Walker.
+ */
+# ifndef likely
+# define likely(x) (__builtin_constant_p(x) ? !!(x) : likely_check(x))
+# endif
+# ifndef unlikely
+# define unlikely(x) (__builtin_constant_p(x) ? !!(x) : unlikely_check(x))
+# endif
+#else
+# define likely(x) __builtin_expect(!!(x), 1)
+# define unlikely(x) __builtin_expect(!!(x), 0)
+#endif
/* Optimization barrier */
#ifndef barrier
diff --git a/kernel/trace/Kconfig b/kernel/trace/Kconfig
index d986216..9c89526 100644
--- a/kernel/trace/Kconfig
+++ b/kernel/trace/Kconfig
@@ -159,6 +159,44 @@
selected, because the self-tests are an initcall as well and that
would invalidate the boot trace. )
+config TRACE_BRANCH_PROFILING
+ bool "Trace likely/unlikely profiler"
+ depends on DEBUG_KERNEL
+ select TRACING
+ help
+ This tracer profiles all the the likely and unlikely macros
+ in the kernel. It will display the results in:
+
+ /debugfs/tracing/profile_likely
+ /debugfs/tracing/profile_unlikely
+
+ Note: this will add a significant overhead, only turn this
+ on if you need to profile the system's use of these macros.
+
+ Say N if unsure.
+
+config TRACING_BRANCHES
+ bool
+ help
+ Selected by tracers that will trace the likely and unlikely
+ conditions. This prevents the tracers themselves from being
+ profiled. Profiling the tracing infrastructure can only happen
+ when the likelys and unlikelys are not being traced.
+
+config BRANCH_TRACER
+ bool "Trace likely/unlikely instances"
+ depends on TRACE_BRANCH_PROFILING
+ select TRACING_BRANCHES
+ help
+ This traces the events of likely and unlikely condition
+ calls in the kernel. The difference between this and the
+ "Trace likely/unlikely profiler" is that this is not a
+ histogram of the callers, but actually places the calling
+ events into a running trace buffer to see when and where the
+ events happened, as well as their results.
+
+ Say N if unsure.
+
config STACK_TRACER
bool "Trace max stack"
depends on HAVE_FUNCTION_TRACER
diff --git a/kernel/trace/Makefile b/kernel/trace/Makefile
index 3e1f361..1a8c925 100644
--- a/kernel/trace/Makefile
+++ b/kernel/trace/Makefile
@@ -10,6 +10,11 @@
obj-y += trace_selftest_dynamic.o
endif
+# If unlikely tracing is enabled, do not trace these files
+ifdef CONFIG_TRACING_BRANCHES
+KBUILD_CFLAGS += -DDISABLE_BRANCH_PROFILING
+endif
+
obj-$(CONFIG_FUNCTION_TRACER) += libftrace.o
obj-$(CONFIG_RING_BUFFER) += ring_buffer.o
@@ -25,5 +30,6 @@
obj-$(CONFIG_MMIOTRACE) += trace_mmiotrace.o
obj-$(CONFIG_BOOT_TRACER) += trace_boot.o
obj-$(CONFIG_FUNCTION_RET_TRACER) += trace_functions_return.o
+obj-$(CONFIG_TRACE_BRANCH_PROFILING) += trace_branch.o
libftrace-y := ftrace.o
diff --git a/kernel/trace/ring_buffer.c b/kernel/trace/ring_buffer.c
index 86dc353..2d6c2cf 100644
--- a/kernel/trace/ring_buffer.c
+++ b/kernel/trace/ring_buffer.c
@@ -53,8 +53,14 @@
/* FIXME!!! */
u64 ring_buffer_time_stamp(int cpu)
{
+ u64 time;
+
+ preempt_disable_notrace();
/* shift to debug/test normalization and TIME_EXTENTS */
- return sched_clock() << DEBUG_SHIFT;
+ time = sched_clock() << DEBUG_SHIFT;
+ preempt_enable_notrace();
+
+ return time;
}
void ring_buffer_normalize_time_stamp(int cpu, u64 *ts)
diff --git a/kernel/trace/trace.c b/kernel/trace/trace.c
index a3f7ae9..4bf070b 100644
--- a/kernel/trace/trace.c
+++ b/kernel/trace/trace.c
@@ -258,6 +258,9 @@
"sched-tree",
"ftrace_printk",
"ftrace_preempt",
+#ifdef CONFIG_BRANCH_TRACER
+ "branch",
+#endif
NULL
};
@@ -1648,6 +1651,18 @@
trace_seq_print_cont(s, iter);
break;
}
+ case TRACE_BRANCH: {
+ struct trace_branch *field;
+
+ trace_assign_type(field, entry);
+
+ trace_seq_printf(s, "[%s] %s:%s:%d\n",
+ field->correct ? " ok " : " MISS ",
+ field->func,
+ field->file,
+ field->line);
+ break;
+ }
default:
trace_seq_printf(s, "Unknown type %d\n", entry->type);
}
@@ -1787,6 +1802,18 @@
return print_return_function(iter);
break;
}
+ case TRACE_BRANCH: {
+ struct trace_branch *field;
+
+ trace_assign_type(field, entry);
+
+ trace_seq_printf(s, "[%s] %s:%s:%d\n",
+ field->correct ? " ok " : " MISS ",
+ field->func,
+ field->file,
+ field->line);
+ break;
+ }
}
return TRACE_TYPE_HANDLED;
}
@@ -2592,6 +2619,7 @@
if (t == current_trace)
goto out;
+ trace_branch_disable();
if (current_trace && current_trace->reset)
current_trace->reset(tr);
@@ -2599,6 +2627,7 @@
if (t->init)
t->init(tr);
+ trace_branch_enable(tr);
out:
mutex_unlock(&trace_types_lock);
diff --git a/kernel/trace/trace.h b/kernel/trace/trace.h
index b5f91f1..9e015f5 100644
--- a/kernel/trace/trace.h
+++ b/kernel/trace/trace.h
@@ -22,6 +22,7 @@
TRACE_SPECIAL,
TRACE_MMIO_RW,
TRACE_MMIO_MAP,
+ TRACE_BRANCH,
TRACE_BOOT_CALL,
TRACE_BOOT_RET,
TRACE_FN_RET,
@@ -134,6 +135,16 @@
struct boot_trace_ret boot_ret;
};
+#define TRACE_FUNC_SIZE 30
+#define TRACE_FILE_SIZE 20
+struct trace_branch {
+ struct trace_entry ent;
+ unsigned line;
+ char func[TRACE_FUNC_SIZE+1];
+ char file[TRACE_FILE_SIZE+1];
+ char correct;
+};
+
/*
* trace_flag_type is an enumeration that holds different
* states when a trace occurs. These are:
@@ -236,6 +247,7 @@
TRACE_MMIO_MAP); \
IF_ASSIGN(var, ent, struct trace_boot_call, TRACE_BOOT_CALL);\
IF_ASSIGN(var, ent, struct trace_boot_ret, TRACE_BOOT_RET);\
+ IF_ASSIGN(var, ent, struct trace_branch, TRACE_BRANCH); \
IF_ASSIGN(var, ent, struct ftrace_ret_entry, TRACE_FN_RET);\
__ftrace_bad_type(); \
} while (0)
@@ -408,6 +420,8 @@
struct trace_array *tr);
extern int trace_selftest_startup_sysprof(struct tracer *trace,
struct trace_array *tr);
+extern int trace_selftest_startup_branch(struct tracer *trace,
+ struct trace_array *tr);
#endif /* CONFIG_FTRACE_STARTUP_TEST */
extern void *head_page(struct trace_array_cpu *data);
@@ -456,6 +470,9 @@
TRACE_ITER_SCHED_TREE = 0x200,
TRACE_ITER_PRINTK = 0x400,
TRACE_ITER_PREEMPTONLY = 0x800,
+#ifdef CONFIG_BRANCH_TRACER
+ TRACE_ITER_BRANCH = 0x1000,
+#endif
};
/*
@@ -515,4 +532,28 @@
preempt_enable_notrace();
}
+#ifdef CONFIG_BRANCH_TRACER
+extern int enable_branch_tracing(struct trace_array *tr);
+extern void disable_branch_tracing(void);
+static inline int trace_branch_enable(struct trace_array *tr)
+{
+ if (trace_flags & TRACE_ITER_BRANCH)
+ return enable_branch_tracing(tr);
+ return 0;
+}
+static inline void trace_branch_disable(void)
+{
+ /* due to races, always disable */
+ disable_branch_tracing();
+}
+#else
+static inline int trace_branch_enable(struct trace_array *tr)
+{
+ return 0;
+}
+static inline void trace_branch_disable(void)
+{
+}
+#endif /* CONFIG_BRANCH_TRACER */
+
#endif /* _LINUX_KERNEL_TRACE_H */
diff --git a/kernel/trace/trace_branch.c b/kernel/trace/trace_branch.c
new file mode 100644
index 0000000..8526555
--- /dev/null
+++ b/kernel/trace/trace_branch.c
@@ -0,0 +1,320 @@
+/*
+ * unlikely profiler
+ *
+ * Copyright (C) 2008 Steven Rostedt <srostedt@redhat.com>
+ */
+#include <linux/kallsyms.h>
+#include <linux/seq_file.h>
+#include <linux/spinlock.h>
+#include <linux/debugfs.h>
+#include <linux/uaccess.h>
+#include <linux/module.h>
+#include <linux/ftrace.h>
+#include <linux/hash.h>
+#include <linux/fs.h>
+#include <asm/local.h>
+#include "trace.h"
+
+#ifdef CONFIG_BRANCH_TRACER
+
+static int branch_tracing_enabled __read_mostly;
+static DEFINE_MUTEX(branch_tracing_mutex);
+static struct trace_array *branch_tracer;
+
+static void
+probe_likely_condition(struct ftrace_branch_data *f, int val, int expect)
+{
+ struct trace_array *tr = branch_tracer;
+ struct ring_buffer_event *event;
+ struct trace_branch *entry;
+ unsigned long flags, irq_flags;
+ int cpu, pc;
+ const char *p;
+
+ /*
+ * I would love to save just the ftrace_likely_data pointer, but
+ * this code can also be used by modules. Ugly things can happen
+ * if the module is unloaded, and then we go and read the
+ * pointer. This is slower, but much safer.
+ */
+
+ if (unlikely(!tr))
+ return;
+
+ local_irq_save(flags);
+ cpu = raw_smp_processor_id();
+ if (atomic_inc_return(&tr->data[cpu]->disabled) != 1)
+ goto out;
+
+ event = ring_buffer_lock_reserve(tr->buffer, sizeof(*entry),
+ &irq_flags);
+ if (!event)
+ goto out;
+
+ pc = preempt_count();
+ entry = ring_buffer_event_data(event);
+ tracing_generic_entry_update(&entry->ent, flags, pc);
+ entry->ent.type = TRACE_BRANCH;
+
+ /* Strip off the path, only save the file */
+ p = f->file + strlen(f->file);
+ while (p >= f->file && *p != '/')
+ p--;
+ p++;
+
+ strncpy(entry->func, f->func, TRACE_FUNC_SIZE);
+ strncpy(entry->file, p, TRACE_FILE_SIZE);
+ entry->func[TRACE_FUNC_SIZE] = 0;
+ entry->file[TRACE_FILE_SIZE] = 0;
+ entry->line = f->line;
+ entry->correct = val == expect;
+
+ ring_buffer_unlock_commit(tr->buffer, event, irq_flags);
+
+ out:
+ atomic_dec(&tr->data[cpu]->disabled);
+ local_irq_restore(flags);
+}
+
+static inline
+void trace_likely_condition(struct ftrace_branch_data *f, int val, int expect)
+{
+ if (!branch_tracing_enabled)
+ return;
+
+ probe_likely_condition(f, val, expect);
+}
+
+int enable_branch_tracing(struct trace_array *tr)
+{
+ int ret = 0;
+
+ mutex_lock(&branch_tracing_mutex);
+ branch_tracer = tr;
+ /*
+ * Must be seen before enabling. The reader is a condition
+ * where we do not need a matching rmb()
+ */
+ smp_wmb();
+ branch_tracing_enabled++;
+ mutex_unlock(&branch_tracing_mutex);
+
+ return ret;
+}
+
+void disable_branch_tracing(void)
+{
+ mutex_lock(&branch_tracing_mutex);
+
+ if (!branch_tracing_enabled)
+ goto out_unlock;
+
+ branch_tracing_enabled--;
+
+ out_unlock:
+ mutex_unlock(&branch_tracing_mutex);
+}
+
+static void start_branch_trace(struct trace_array *tr)
+{
+ enable_branch_tracing(tr);
+}
+
+static void stop_branch_trace(struct trace_array *tr)
+{
+ disable_branch_tracing();
+}
+
+static void branch_trace_init(struct trace_array *tr)
+{
+ int cpu;
+
+ for_each_online_cpu(cpu)
+ tracing_reset(tr, cpu);
+
+ start_branch_trace(tr);
+}
+
+static void branch_trace_reset(struct trace_array *tr)
+{
+ stop_branch_trace(tr);
+}
+
+struct tracer branch_trace __read_mostly =
+{
+ .name = "branch",
+ .init = branch_trace_init,
+ .reset = branch_trace_reset,
+#ifdef CONFIG_FTRACE_SELFTEST
+ .selftest = trace_selftest_startup_branch,
+#endif
+};
+
+__init static int init_branch_trace(void)
+{
+ return register_tracer(&branch_trace);
+}
+
+device_initcall(init_branch_trace);
+#else
+static inline
+void trace_likely_condition(struct ftrace_branch_data *f, int val, int expect)
+{
+}
+#endif /* CONFIG_BRANCH_TRACER */
+
+void ftrace_likely_update(struct ftrace_branch_data *f, int val, int expect)
+{
+ /*
+ * I would love to have a trace point here instead, but the
+ * trace point code is so inundated with unlikely and likely
+ * conditions that the recursive nightmare that exists is too
+ * much to try to get working. At least for now.
+ */
+ trace_likely_condition(f, val, expect);
+
+ /* FIXME: Make this atomic! */
+ if (val == expect)
+ f->correct++;
+ else
+ f->incorrect++;
+}
+EXPORT_SYMBOL(ftrace_likely_update);
+
+struct ftrace_pointer {
+ void *start;
+ void *stop;
+};
+
+static void *
+t_next(struct seq_file *m, void *v, loff_t *pos)
+{
+ struct ftrace_pointer *f = m->private;
+ struct ftrace_branch_data *p = v;
+
+ (*pos)++;
+
+ if (v == (void *)1)
+ return f->start;
+
+ ++p;
+
+ if ((void *)p >= (void *)f->stop)
+ return NULL;
+
+ return p;
+}
+
+static void *t_start(struct seq_file *m, loff_t *pos)
+{
+ void *t = (void *)1;
+ loff_t l = 0;
+
+ for (; t && l < *pos; t = t_next(m, t, &l))
+ ;
+
+ return t;
+}
+
+static void t_stop(struct seq_file *m, void *p)
+{
+}
+
+static int t_show(struct seq_file *m, void *v)
+{
+ struct ftrace_branch_data *p = v;
+ const char *f;
+ unsigned long percent;
+
+ if (v == (void *)1) {
+ seq_printf(m, " correct incorrect %% "
+ " Function "
+ " File Line\n"
+ " ------- --------- - "
+ " -------- "
+ " ---- ----\n");
+ return 0;
+ }
+
+ /* Only print the file, not the path */
+ f = p->file + strlen(p->file);
+ while (f >= p->file && *f != '/')
+ f--;
+ f++;
+
+ if (p->correct) {
+ percent = p->incorrect * 100;
+ percent /= p->correct + p->incorrect;
+ } else
+ percent = p->incorrect ? 100 : 0;
+
+ seq_printf(m, "%8lu %8lu %3lu ", p->correct, p->incorrect, percent);
+ seq_printf(m, "%-30.30s %-20.20s %d\n", p->func, f, p->line);
+ return 0;
+}
+
+static struct seq_operations tracing_likely_seq_ops = {
+ .start = t_start,
+ .next = t_next,
+ .stop = t_stop,
+ .show = t_show,
+};
+
+static int tracing_likely_open(struct inode *inode, struct file *file)
+{
+ int ret;
+
+ ret = seq_open(file, &tracing_likely_seq_ops);
+ if (!ret) {
+ struct seq_file *m = file->private_data;
+ m->private = (void *)inode->i_private;
+ }
+
+ return ret;
+}
+
+static struct file_operations tracing_likely_fops = {
+ .open = tracing_likely_open,
+ .read = seq_read,
+ .llseek = seq_lseek,
+};
+
+extern unsigned long __start_likely_profile[];
+extern unsigned long __stop_likely_profile[];
+extern unsigned long __start_unlikely_profile[];
+extern unsigned long __stop_unlikely_profile[];
+
+static struct ftrace_pointer ftrace_likely_pos = {
+ .start = __start_likely_profile,
+ .stop = __stop_likely_profile,
+};
+
+static struct ftrace_pointer ftrace_unlikely_pos = {
+ .start = __start_unlikely_profile,
+ .stop = __stop_unlikely_profile,
+};
+
+static __init int ftrace_branch_init(void)
+{
+ struct dentry *d_tracer;
+ struct dentry *entry;
+
+ d_tracer = tracing_init_dentry();
+
+ entry = debugfs_create_file("profile_likely", 0444, d_tracer,
+ &ftrace_likely_pos,
+ &tracing_likely_fops);
+ if (!entry)
+ pr_warning("Could not create debugfs 'profile_likely' entry\n");
+
+ entry = debugfs_create_file("profile_unlikely", 0444, d_tracer,
+ &ftrace_unlikely_pos,
+ &tracing_likely_fops);
+ if (!entry)
+ pr_warning("Could not create debugfs"
+ " 'profile_unlikely' entry\n");
+
+ return 0;
+}
+
+device_initcall(ftrace_branch_init);
diff --git a/kernel/trace/trace_selftest.c b/kernel/trace/trace_selftest.c
index 0728a10..24e6e07 100644
--- a/kernel/trace/trace_selftest.c
+++ b/kernel/trace/trace_selftest.c
@@ -13,6 +13,7 @@
case TRACE_STACK:
case TRACE_PRINT:
case TRACE_SPECIAL:
+ case TRACE_BRANCH:
return 1;
}
return 0;
@@ -544,3 +545,25 @@
return ret;
}
#endif /* CONFIG_SYSPROF_TRACER */
+
+#ifdef CONFIG_BRANCH_TRACER
+int
+trace_selftest_startup_branch(struct tracer *trace, struct trace_array *tr)
+{
+ unsigned long count;
+ int ret;
+
+ /* start the tracing */
+ trace->init(tr);
+ /* Sleep for a 1/10 of a second */
+ msleep(100);
+ /* stop the tracing. */
+ tracing_stop();
+ /* check the trace buffer */
+ ret = trace_test_buffer(tr, &count);
+ trace->reset(tr);
+ tracing_start();
+
+ return ret;
+}
+#endif /* CONFIG_BRANCH_TRACER */
diff --git a/scripts/bootgraph.pl b/scripts/bootgraph.pl
index d2c61ef..a8635a1 100644
--- a/scripts/bootgraph.pl
+++ b/scripts/bootgraph.pl
@@ -109,8 +109,8 @@
my %rows;
my $rowscount = 1;
my @initcalls = sort { $start{$a} <=> $start{$b} } keys(%start);
-my $key;
-foreach $key (@initcalls) {
+
+foreach my $key (@initcalls) {
my $duration = $end{$key} - $start{$key};
if ($duration >= $threshold) {