From a32c7765e2796395aec49f699bd25c407155e9c5 Mon Sep 17 00:00:00 2001 From: Li Zefan Date: Fri, 26 Jun 2009 16:55:51 +0800 Subject: tracing: Fix stack tracer sysctl handling This made my machine completely frozen: # echo 1 > /proc/sys/kernel/stack_tracer_enabled # echo 2 > /proc/sys/kernel/stack_tracer_enabled The cause is register_ftrace_function() was called twice. Also fix ftrace_enabled sysctl, though seems nothing bad happened as I tested it. Signed-off-by: Li Zefan Cc: Steven Rostedt Cc: Frederic Weisbecker LKML-Reference: <4A448D17.9010305@cn.fujitsu.com> Signed-off-by: Ingo Molnar --- kernel/trace/ftrace.c | 4 ++-- kernel/trace/trace_stack.c | 4 ++-- 2 files changed, 4 insertions(+), 4 deletions(-) (limited to 'kernel/trace') diff --git a/kernel/trace/ftrace.c b/kernel/trace/ftrace.c index f3716bf04df..bce9e01a29c 100644 --- a/kernel/trace/ftrace.c +++ b/kernel/trace/ftrace.c @@ -3160,10 +3160,10 @@ ftrace_enable_sysctl(struct ctl_table *table, int write, ret = proc_dointvec(table, write, file, buffer, lenp, ppos); - if (ret || !write || (last_ftrace_enabled == ftrace_enabled)) + if (ret || !write || (last_ftrace_enabled == !!ftrace_enabled)) goto out; - last_ftrace_enabled = ftrace_enabled; + last_ftrace_enabled = !!ftrace_enabled; if (ftrace_enabled) { diff --git a/kernel/trace/trace_stack.c b/kernel/trace/trace_stack.c index 2d7aebd71db..e644af91012 100644 --- a/kernel/trace/trace_stack.c +++ b/kernel/trace/trace_stack.c @@ -326,10 +326,10 @@ stack_trace_sysctl(struct ctl_table *table, int write, ret = proc_dointvec(table, write, file, buffer, lenp, ppos); if (ret || !write || - (last_stack_tracer_enabled == stack_tracer_enabled)) + (last_stack_tracer_enabled == !!stack_tracer_enabled)) goto out; - last_stack_tracer_enabled = stack_tracer_enabled; + last_stack_tracer_enabled = !!stack_tracer_enabled; if (stack_tracer_enabled) register_ftrace_function(&trace_ops); -- cgit v1.2.3 From 82d5308127c3e3404ffbf41e503853c68660b18b Mon Sep 17 00:00:00 2001 From: Lai Jiangshan Date: Fri, 26 Jun 2009 17:07:02 +0800 Subject: trace_export: Repair missed fields Some fields for struct ftrace_graph_ret are missed when they are exported to user. Signed-off-by: Lai Jiangshan Cc: Frederic Weisbecker Cc: Steven Rostedt LKML-Reference: <4A448FB6.5000302@cn.fujitsu.com> Signed-off-by: Ingo Molnar --- kernel/trace/trace_event_types.h | 3 +++ 1 file changed, 3 insertions(+) (limited to 'kernel/trace') diff --git a/kernel/trace/trace_event_types.h b/kernel/trace/trace_event_types.h index 5e32e375134..6db005e1248 100644 --- a/kernel/trace/trace_event_types.h +++ b/kernel/trace/trace_event_types.h @@ -26,6 +26,9 @@ TRACE_EVENT_FORMAT(funcgraph_exit, TRACE_GRAPH_RET, ftrace_graph_ret_entry, ignore, TRACE_STRUCT( TRACE_FIELD(unsigned long, ret.func, func) + TRACE_FIELD(unsigned long long, ret.calltime, calltime) + TRACE_FIELD(unsigned long long, ret.rettime, rettime) + TRACE_FIELD(unsigned long, ret.overrun, overrun) TRACE_FIELD(int, ret.depth, depth) ), TP_RAW_FMT("<-- %lx (%d)") -- cgit v1.2.3 From 238a24f626628cb16a3015f332d649f08246ca89 Mon Sep 17 00:00:00 2001 From: Li Zefan Date: Mon, 29 Jun 2009 15:55:10 +0800 Subject: tracing/fastboot: Document the need of initcall_debug To use boot tracer, one should pass initcall_debug as well as ftrace=initcall to the command line. Signed-off-by: Li Zefan Cc: Frederic Weisbecker Cc: Steven Rostedt LKML-Reference: <4A48735E.9050002@cn.fujitsu.com> Signed-off-by: Ingo Molnar --- kernel/trace/Kconfig | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-) (limited to 'kernel/trace') diff --git a/kernel/trace/Kconfig b/kernel/trace/Kconfig index 1551f47e766..019f380fd76 100644 --- a/kernel/trace/Kconfig +++ b/kernel/trace/Kconfig @@ -226,13 +226,13 @@ config BOOT_TRACER the timings of the initcalls and traces key events and the identity of tasks that can cause boot delays, such as context-switches. - Its aim is to be parsed by the /scripts/bootgraph.pl tool to + Its aim is to be parsed by the scripts/bootgraph.pl tool to produce pretty graphics about boot inefficiencies, giving a visual representation of the delays during initcalls - but the raw /debug/tracing/trace text output is readable too. - You must pass in ftrace=initcall to the kernel command line - to enable this on bootup. + You must pass in initcall_debug and ftrace=initcall to the kernel + command line to enable this on bootup. config TRACE_BRANCH_PROFILING bool -- cgit v1.2.3 From e1af3aec3e2e7d21d4b3054323779d478c19d907 Mon Sep 17 00:00:00 2001 From: Xiao Guangrong Date: Wed, 1 Jul 2009 16:50:25 +0800 Subject: tracing: Fix trace_print_seq() We will lose something if trace_seq->buffer[0] is 0, because the copy length is calculated by strlen() in seq_puts(), so using seq_write() instead of seq_puts(). There have a example: after reboot: # echo kmemtrace > current_tracer # echo 0 > options/kmem_minimalistic # cat trace # tracer: kmemtrace # # Nothing is exported, because the first byte of trace_seq->buffer[ ] is KMEMTRACE_USER_ALLOC. ( the value of KMEMTRACE_USER_ALLOC is zero, seeing kmemtrace_print_alloc_user() in kernel/trace/kmemtrace.c) Signed-off-by: Xiao Guangrong Acked-by: Frederic Weisbecker Acked-by: Pekka Enberg Acked-by: Eduard - Gabriel Munteanu Cc: Steven Rostedt LKML-Reference: <4A4B2351.5010300@cn.fujitsu.com> Signed-off-by: Ingo Molnar --- kernel/trace/trace_output.c | 3 +-- 1 file changed, 1 insertion(+), 2 deletions(-) (limited to 'kernel/trace') diff --git a/kernel/trace/trace_output.c b/kernel/trace/trace_output.c index 7938f3ae93e..e0c2545622e 100644 --- a/kernel/trace/trace_output.c +++ b/kernel/trace/trace_output.c @@ -27,8 +27,7 @@ void trace_print_seq(struct seq_file *m, struct trace_seq *s) { int len = s->len >= PAGE_SIZE ? PAGE_SIZE - 1 : s->len; - s->buffer[len] = 0; - seq_puts(m, s->buffer); + seq_write(m, s->buffer, len); trace_seq_init(s); } -- cgit v1.2.3 From 405f55712dfe464b3240d7816cc4fe4174831be2 Mon Sep 17 00:00:00 2001 From: Alexey Dobriyan Date: Sat, 11 Jul 2009 22:08:37 +0400 Subject: headers: smp_lock.h redux * Remove smp_lock.h from files which don't need it (including some headers!) * Add smp_lock.h to files which do need it * Make smp_lock.h include conditional in hardirq.h It's needed only for one kernel_locked() usage which is under CONFIG_PREEMPT This will make hardirq.h inclusion cheaper for every PREEMPT=n config (which includes allmodconfig/allyesconfig, BTW) Signed-off-by: Alexey Dobriyan Signed-off-by: Linus Torvalds --- kernel/trace/blktrace.c | 1 + kernel/trace/trace.c | 1 + 2 files changed, 2 insertions(+) (limited to 'kernel/trace') diff --git a/kernel/trace/blktrace.c b/kernel/trace/blktrace.c index 39af8af6fc3..1090b0aed9b 100644 --- a/kernel/trace/blktrace.c +++ b/kernel/trace/blktrace.c @@ -22,6 +22,7 @@ #include #include #include +#include #include #include diff --git a/kernel/trace/trace.c b/kernel/trace/trace.c index 3aa0a0dfdfa..8bc8d8afea6 100644 --- a/kernel/trace/trace.c +++ b/kernel/trace/trace.c @@ -17,6 +17,7 @@ #include #include #include +#include #include #include #include -- cgit v1.2.3 From 6ab5d668b131d3c5416f6df1d3ca95b82d4fe8a2 Mon Sep 17 00:00:00 2001 From: Steven Rostedt Date: Thu, 4 Jun 2009 00:55:45 -0400 Subject: tracing/function-profiler: do not free per cpu variable stat The per cpu variable stat is freeded if we fail to allocate a name on start up. This was due to stat at first being allocated in the initial design. But since then, it has become a static per cpu variable but the free on error was not removed. Also added __init annotation to the function that this is in. [ Impact: prevent possible memory corruption on low mem at boot up ] Signed-off-by: Steven Rostedt Signed-off-by: Ingo Molnar --- kernel/trace/ftrace.c | 5 ++--- 1 file changed, 2 insertions(+), 3 deletions(-) (limited to 'kernel/trace') diff --git a/kernel/trace/ftrace.c b/kernel/trace/ftrace.c index bce9e01a29c..4521c77d1a1 100644 --- a/kernel/trace/ftrace.c +++ b/kernel/trace/ftrace.c @@ -768,7 +768,7 @@ static struct tracer_stat function_stats __initdata = { .stat_show = function_stat_show }; -static void ftrace_profile_debugfs(struct dentry *d_tracer) +static __init void ftrace_profile_debugfs(struct dentry *d_tracer) { struct ftrace_profile_stat *stat; struct dentry *entry; @@ -786,7 +786,6 @@ static void ftrace_profile_debugfs(struct dentry *d_tracer) * The files created are permanent, if something happens * we still do not free memory. */ - kfree(stat); WARN(1, "Could not allocate stat file for cpu %d\n", cpu); @@ -813,7 +812,7 @@ static void ftrace_profile_debugfs(struct dentry *d_tracer) } #else /* CONFIG_FUNCTION_PROFILER */ -static void ftrace_profile_debugfs(struct dentry *d_tracer) +static __init void ftrace_profile_debugfs(struct dentry *d_tracer) { } #endif /* CONFIG_FUNCTION_PROFILER */ -- cgit v1.2.3 From 04aef32d39cc4ef80087c0ce8ed113c6d64f1a6b Mon Sep 17 00:00:00 2001 From: Xiao Guangrong Date: Wed, 15 Jul 2009 12:29:06 +0800 Subject: tracing/function: Fix the return value of ftrace_trace_onoff_callback() ftrace_trace_onoff_callback() will return an error even if we do the right operation, for example: # echo _spin_*:traceon:10 > set_ftrace_filter -bash: echo: write error: Invalid argument # cat set_ftrace_filter #### all functions enabled #### _spin_trylock_bh:traceon:count=10 _spin_unlock_irq:traceon:count=10 _spin_unlock_bh:traceon:count=10 _spin_lock_irq:traceon:count=10 _spin_unlock:traceon:count=10 _spin_trylock:traceon:count=10 _spin_unlock_irqrestore:traceon:count=10 _spin_lock_irqsave:traceon:count=10 _spin_lock_bh:traceon:count=10 _spin_lock:traceon:count=10 We want to set _spin_*:traceon:10 to set_ftrace_filter, it complains with "Invalid argument", but the operation is successful. This is because ftrace_process_regex() returns the number of functions that matched the pattern. If the number is not 0, this value is returned by ftrace_regex_write() whereas we want to return the number of bytes virtually written. Also the file offset pointer is not updated in this case. If the number of matched functions is lower than the number of bytes written by the user, this results to a reprocessing of the string given by the user with a lower size, leading to a malformed ftrace regex and then a -EINVAL returned. So, this patch fixes it by returning 0 if no error occured. The fix also applies on 2.6.30 Signed-off-by: Xiao Guangrong Reviewed-by: Li Zefan Cc: stable@kernel.org Signed-off-by: Frederic Weisbecker --- kernel/trace/trace_functions.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'kernel/trace') diff --git a/kernel/trace/trace_functions.c b/kernel/trace/trace_functions.c index 7402144bff2..75ef000613c 100644 --- a/kernel/trace/trace_functions.c +++ b/kernel/trace/trace_functions.c @@ -363,7 +363,7 @@ ftrace_trace_onoff_callback(char *glob, char *cmd, char *param, int enable) out_reg: ret = register_ftrace_function_probe(glob, ops, count); - return ret; + return ret < 0 ? ret : 0; } static struct ftrace_func_command ftrace_traceon_cmd = { -- cgit v1.2.3 From d8cc1ab793993c886c62abf77c93287df33ffd8b Mon Sep 17 00:00:00 2001 From: Li Zefan Date: Thu, 23 Jul 2009 11:28:40 +0800 Subject: trace_stack: Fix seqfile memory leak Every time we cat stack_trace, we leak memory allocated by seq_open(). Signed-off-by: Li Zefan LKML-Reference: <4A67D8E8.3020500@cn.fujitsu.com> Signed-off-by: Steven Rostedt --- kernel/trace/trace_stack.c | 7 ++----- 1 file changed, 2 insertions(+), 5 deletions(-) (limited to 'kernel/trace') diff --git a/kernel/trace/trace_stack.c b/kernel/trace/trace_stack.c index e644af91012..6a2a9d484cd 100644 --- a/kernel/trace/trace_stack.c +++ b/kernel/trace/trace_stack.c @@ -301,17 +301,14 @@ static const struct seq_operations stack_trace_seq_ops = { static int stack_trace_open(struct inode *inode, struct file *file) { - int ret; - - ret = seq_open(file, &stack_trace_seq_ops); - - return ret; + return seq_open(file, &stack_trace_seq_ops); } static const struct file_operations stack_trace_fops = { .open = stack_trace_open, .read = seq_read, .llseek = seq_lseek, + .release = seq_release, }; int -- cgit v1.2.3 From 87827111a5538633b18e5c641ced673c4c2bb6ce Mon Sep 17 00:00:00 2001 From: Li Zefan Date: Thu, 23 Jul 2009 11:29:11 +0800 Subject: function-graph: Fix seqfile memory leak Every time we cat set_graph_function, we leak memory allocated by seq_open(). Signed-off-by: Li Zefan LKML-Reference: <4A67D907.2010500@cn.fujitsu.com> Signed-off-by: Steven Rostedt --- kernel/trace/ftrace.c | 15 ++++++++++++--- 1 file changed, 12 insertions(+), 3 deletions(-) (limited to 'kernel/trace') diff --git a/kernel/trace/ftrace.c b/kernel/trace/ftrace.c index 4521c77d1a1..1f3ec2afa51 100644 --- a/kernel/trace/ftrace.c +++ b/kernel/trace/ftrace.c @@ -2595,6 +2595,14 @@ ftrace_graph_open(struct inode *inode, struct file *file) return ret; } +static int +ftrace_graph_release(struct inode *inode, struct file *file) +{ + if (file->f_mode & FMODE_READ) + seq_release(inode, file); + return 0; +} + static int ftrace_set_func(unsigned long *array, int *idx, char *buffer) { @@ -2724,9 +2732,10 @@ ftrace_graph_write(struct file *file, const char __user *ubuf, } static const struct file_operations ftrace_graph_fops = { - .open = ftrace_graph_open, - .read = seq_read, - .write = ftrace_graph_write, + .open = ftrace_graph_open, + .read = seq_read, + .write = ftrace_graph_write, + .release = ftrace_graph_release, }; #endif /* CONFIG_FUNCTION_GRAPH_TRACER */ -- cgit v1.2.3 From 636eacee3b0c76915151db37203cc624becb6d7b Mon Sep 17 00:00:00 2001 From: Li Zefan Date: Thu, 23 Jul 2009 11:29:47 +0800 Subject: tracing/stat: Fix seqfile memory leak Every time we cat a trace_stat file, we leak memory allocated by seq_open(). Also fix memory leak in a failure path in tracing_stat_open(). Signed-off-by: Li Zefan LKML-Reference: <4A67D92B.4060704@cn.fujitsu.com> Signed-off-by: Steven Rostedt --- kernel/trace/trace_stat.c | 34 ++++++++++++++++++++++------------ 1 file changed, 22 insertions(+), 12 deletions(-) (limited to 'kernel/trace') diff --git a/kernel/trace/trace_stat.c b/kernel/trace/trace_stat.c index e66f5e49334..aea321c82fa 100644 --- a/kernel/trace/trace_stat.c +++ b/kernel/trace/trace_stat.c @@ -73,7 +73,7 @@ static struct rb_node *release_next(struct rb_node *node) } } -static void reset_stat_session(struct stat_session *session) +static void __reset_stat_session(struct stat_session *session) { struct rb_node *node = session->stat_root.rb_node; @@ -83,10 +83,17 @@ static void reset_stat_session(struct stat_session *session) session->stat_root = RB_ROOT; } +static void reset_stat_session(struct stat_session *session) +{ + mutex_lock(&session->stat_mutex); + __reset_stat_session(session); + mutex_unlock(&session->stat_mutex); +} + static void destroy_session(struct stat_session *session) { debugfs_remove(session->file); - reset_stat_session(session); + __reset_stat_session(session); mutex_destroy(&session->stat_mutex); kfree(session); } @@ -150,7 +157,7 @@ static int stat_seq_init(struct stat_session *session) int i; mutex_lock(&session->stat_mutex); - reset_stat_session(session); + __reset_stat_session(session); if (!ts->stat_cmp) ts->stat_cmp = dummy_cmp; @@ -183,7 +190,7 @@ exit: return ret; exit_free_rbtree: - reset_stat_session(session); + __reset_stat_session(session); mutex_unlock(&session->stat_mutex); return ret; } @@ -250,16 +257,21 @@ static const struct seq_operations trace_stat_seq_ops = { static int tracing_stat_open(struct inode *inode, struct file *file) { int ret; - + struct seq_file *m; struct stat_session *session = inode->i_private; + ret = stat_seq_init(session); + if (ret) + return ret; + ret = seq_open(file, &trace_stat_seq_ops); - if (!ret) { - struct seq_file *m = file->private_data; - m->private = session; - ret = stat_seq_init(session); + if (ret) { + reset_stat_session(session); + return ret; } + m = file->private_data; + m->private = session; return ret; } @@ -270,11 +282,9 @@ static int tracing_stat_release(struct inode *i, struct file *f) { struct stat_session *session = i->i_private; - mutex_lock(&session->stat_mutex); reset_stat_session(session); - mutex_unlock(&session->stat_mutex); - return 0; + return seq_release(i, f); } static const struct file_operations tracing_stat_fops = { -- cgit v1.2.3 From 4c739ff043e5787d97c9691d62cabf7a29e75a9d Mon Sep 17 00:00:00 2001 From: Steven Rostedt Date: Wed, 22 Jul 2009 23:11:03 -0400 Subject: tracing: show proper address for trace-printk format Since the trace_printk may use pointers to the format fields in the buffer, they are exported via debugfs/tracing/printk_formats. This is used by utilities that read the ring buffer in binary format. It helps the utilities map the address of the format in the binary buffer to what the printf format looks like. Unfortunately, the way the output code works, it exports the address of the pointer to the format address, and not the format address itself. This makes the file totally useless in trying to figure out what format string a binary address belongs to. Signed-off-by: Steven Rostedt --- kernel/trace/trace_printk.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'kernel/trace') diff --git a/kernel/trace/trace_printk.c b/kernel/trace/trace_printk.c index 7b627811082..687699d365a 100644 --- a/kernel/trace/trace_printk.c +++ b/kernel/trace/trace_printk.c @@ -176,7 +176,7 @@ static int t_show(struct seq_file *m, void *v) const char *str = *fmt; int i; - seq_printf(m, "0x%lx : \"", (unsigned long)fmt); + seq_printf(m, "0x%lx : \"", *(unsigned long *)fmt); /* * Tabs and new lines need to be converted. -- cgit v1.2.3 From 8650ae32ef7045e763825dee6256dde7f331bb85 Mon Sep 17 00:00:00 2001 From: Steven Rostedt Date: Wed, 22 Jul 2009 23:29:30 -0400 Subject: tracing: only truncate ftrace files when O_TRUNC is set The current code will truncate the ftrace files contents if O_APPEND is not set and the file is opened in write mode. This is incorrect. It should only truncate the file if O_TRUNC is set. Otherwise if one of these files is opened by a C program with fopen "r+", it will incorrectly truncate the file. Reported-by: Jiri Olsa Signed-off-by: Steven Rostedt --- kernel/trace/ftrace.c | 4 ++-- kernel/trace/trace.c | 2 +- kernel/trace/trace_events.c | 2 +- 3 files changed, 4 insertions(+), 4 deletions(-) (limited to 'kernel/trace') diff --git a/kernel/trace/ftrace.c b/kernel/trace/ftrace.c index 1f3ec2afa51..1e1d23c2630 100644 --- a/kernel/trace/ftrace.c +++ b/kernel/trace/ftrace.c @@ -1662,7 +1662,7 @@ ftrace_regex_open(struct inode *inode, struct file *file, int enable) mutex_lock(&ftrace_regex_lock); if ((file->f_mode & FMODE_WRITE) && - !(file->f_flags & O_APPEND)) + (file->f_flags & O_TRUNC)) ftrace_filter_reset(enable); if (file->f_mode & FMODE_READ) { @@ -2577,7 +2577,7 @@ ftrace_graph_open(struct inode *inode, struct file *file) mutex_lock(&graph_lock); if ((file->f_mode & FMODE_WRITE) && - !(file->f_flags & O_APPEND)) { + (file->f_flags & O_TRUNC)) { ftrace_graph_count = 0; memset(ftrace_graph_funcs, 0, sizeof(ftrace_graph_funcs)); } diff --git a/kernel/trace/trace.c b/kernel/trace/trace.c index 8bc8d8afea6..d8ef28574aa 100644 --- a/kernel/trace/trace.c +++ b/kernel/trace/trace.c @@ -2031,7 +2031,7 @@ static int tracing_open(struct inode *inode, struct file *file) /* If this file was open for write, then erase contents */ if ((file->f_mode & FMODE_WRITE) && - !(file->f_flags & O_APPEND)) { + (file->f_flags & O_TRUNC)) { long cpu = (long) inode->i_private; if (cpu == TRACE_PIPE_ALL_CPU) diff --git a/kernel/trace/trace_events.c b/kernel/trace/trace_events.c index 53c8fd376a8..23d2972b22d 100644 --- a/kernel/trace/trace_events.c +++ b/kernel/trace/trace_events.c @@ -376,7 +376,7 @@ ftrace_event_seq_open(struct inode *inode, struct file *file) const struct seq_operations *seq_ops; if ((file->f_mode & FMODE_WRITE) && - !(file->f_flags & O_APPEND)) + (file->f_flags & O_TRUNC)) ftrace_clear_events(); seq_ops = inode->i_private; -- cgit v1.2.3 From 38ceb592fcac9110c6b3c87ea0a27bff68c43486 Mon Sep 17 00:00:00 2001 From: Lai Jiangshan Date: Tue, 28 Jul 2009 20:11:24 +0800 Subject: tracing: Fix invalid function_graph entry When print_graph_entry() computes a function call entry event, it needs to also check the next entry to guess if it matches the return event of the current function entry. In order to look at this next event, it needs to consume the current entry before going ahead in the ring buffer. However, if the current event that gets consumed is the last one in the ring buffer head page, the ring_buffer may reuse the page for writers. The consumed entry will then become invalid because of possible racy overwriting. Me must then handle this entry by making a copy of it. The fix also applies on 2.6.30 Signed-off-by: Lai Jiangshan Cc: Steven Rostedt Cc: stable@kernel.org LKML-Reference: <4A6EEAEC.3050508@cn.fujitsu.com> Signed-off-by: Frederic Weisbecker --- kernel/trace/trace_functions_graph.c | 11 +++++++++-- 1 file changed, 9 insertions(+), 2 deletions(-) (limited to 'kernel/trace') diff --git a/kernel/trace/trace_functions_graph.c b/kernel/trace/trace_functions_graph.c index d2249abafb5..420ec348757 100644 --- a/kernel/trace/trace_functions_graph.c +++ b/kernel/trace/trace_functions_graph.c @@ -843,9 +843,16 @@ print_graph_function(struct trace_iterator *iter) switch (entry->type) { case TRACE_GRAPH_ENT: { - struct ftrace_graph_ent_entry *field; + /* + * print_graph_entry() may consume the current event, + * thus @field may become invalid, so we need to save it. + * sizeof(struct ftrace_graph_ent_entry) is very small, + * it can be safely saved at the stack. + */ + struct ftrace_graph_ent_entry *field, saved; trace_assign_type(field, entry); - return print_graph_entry(field, s, iter); + saved = *field; + return print_graph_entry(&saved, s, iter); } case TRACE_GRAPH_RET: { struct ftrace_graph_ret_entry *field; -- cgit v1.2.3 From 74e7ff8c50b6b022e6ffaa736b16a4dc161d3eaf Mon Sep 17 00:00:00 2001 From: Lai Jiangshan Date: Tue, 28 Jul 2009 20:17:22 +0800 Subject: tracing: Fix missing function_graph events when we splice_read from trace_pipe About a half events are missing when we splice_read from trace_pipe. They are unexpectedly consumed because we ignore the TRACE_TYPE_NO_CONSUME return value used by the function graph tracer when it needs to consume the events by itself to walk on the ring buffer. The same problem appears with ftrace_dump() Example of an output before this patch: 1) | ktime_get_real() { 1) 2.846 us | read_hpet(); 1) 4.558 us | } 1) 6.195 us | } After this patch: 0) | ktime_get_real() { 0) | getnstimeofday() { 0) 1.960 us | read_hpet(); 0) 3.597 us | } 0) 5.196 us | } The fix also applies on 2.6.30 Signed-off-by: Lai Jiangshan Cc: Steven Rostedt Cc: stable@kernel.org LKML-Reference: <4A6EEC52.90704@cn.fujitsu.com> Signed-off-by: Frederic Weisbecker --- kernel/trace/trace.c | 10 +++++++--- 1 file changed, 7 insertions(+), 3 deletions(-) (limited to 'kernel/trace') diff --git a/kernel/trace/trace.c b/kernel/trace/trace.c index 8bc8d8afea6..da984ad065a 100644 --- a/kernel/trace/trace.c +++ b/kernel/trace/trace.c @@ -3085,7 +3085,8 @@ tracing_fill_pipe_page(size_t rem, struct trace_iterator *iter) break; } - trace_consume(iter); + if (ret != TRACE_TYPE_NO_CONSUME) + trace_consume(iter); rem -= count; if (!find_next_entry_inc(iter)) { rem = 0; @@ -4233,8 +4234,11 @@ static void __ftrace_dump(bool disable_tracing) iter.pos = -1; if (find_next_entry_inc(&iter) != NULL) { - print_trace_line(&iter); - trace_consume(&iter); + int ret; + + ret = print_trace_line(&iter); + if (ret != TRACE_TYPE_NO_CONSUME) + trace_consume(&iter); } trace_printk_seq(&iter.seq); -- cgit v1.2.3 From 0f2541d299d233eddddee4345795e0c46264fd56 Mon Sep 17 00:00:00 2001 From: Steven Rostedt Date: Wed, 5 Aug 2009 12:02:48 -0400 Subject: ring-buffer: fix check of try_to_discard result The function ring_buffer_discard_commit inversed the code path of the result of try_to_discard. It should skip incrementing the entry counter if try_to_discard succeeded. But instead, it increments the entry conder if it succeeded to discard, and does not increment it if it fails. The result of this bug is that filtering will make the stat counters incorrect. Signed-off-by: Steven Rostedt --- kernel/trace/ring_buffer.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'kernel/trace') diff --git a/kernel/trace/ring_buffer.c b/kernel/trace/ring_buffer.c index bf27bb7a63e..2fd1752f0c8 100644 --- a/kernel/trace/ring_buffer.c +++ b/kernel/trace/ring_buffer.c @@ -1785,7 +1785,7 @@ void ring_buffer_discard_commit(struct ring_buffer *buffer, */ RB_WARN_ON(buffer, !local_read(&cpu_buffer->committing)); - if (!rb_try_to_discard(cpu_buffer, event)) + if (rb_try_to_discard(cpu_buffer, event)) goto out; /* -- cgit v1.2.3 From 464e85eb0e63096bd52e4c3e2a6fb8357fb95828 Mon Sep 17 00:00:00 2001 From: Steven Rostedt Date: Wed, 5 Aug 2009 15:26:37 -0400 Subject: ring-buffer: do not disable ring buffer on oops_in_progress The commit: commit e0fdace10e75dac67d906213b780ff1b1a4cc360 Author: David Miller Date: Fri Aug 1 01:11:22 2008 -0700 debug_locks: set oops_in_progress if we will log messages. Otherwise lock debugging messages on runqueue locks can deadlock the system due to the wakeups performed by printk(). Signed-off-by: David S. Miller Signed-off-by: Ingo Molnar Will permanently set oops_in_progress on any lockdep failure. When this triggers it will cause any read from the ring buffer to permanently disable the ring buffer (not to mention no locking of printk). This patch removes the check. It keeps the print in NMI which makes sense. This is probably OK, since the ring buffer should not cause something to set oops_in_progress anyway. Signed-off-by: Steven Rostedt --- kernel/trace/ring_buffer.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'kernel/trace') diff --git a/kernel/trace/ring_buffer.c b/kernel/trace/ring_buffer.c index 2fd1752f0c8..2606cee433d 100644 --- a/kernel/trace/ring_buffer.c +++ b/kernel/trace/ring_buffer.c @@ -2486,7 +2486,7 @@ static inline int rb_ok_to_lock(void) * buffer too. A one time deal is all you get from reading * the ring buffer from an NMI. */ - if (likely(!in_nmi() && !oops_in_progress)) + if (likely(!in_nmi())) return 1; tracing_off_permanent(); -- cgit v1.2.3 From af6af30c0fcd77e621638e53ef8b176bca8bd3b4 Mon Sep 17 00:00:00 2001 From: Peter Zijlstra Date: Wed, 5 Aug 2009 20:41:04 +0200 Subject: ftrace: Fix perf-tracepoint OOPS Not all tracepoints are created equal, in specific the ftrace tracepoints are created with TRACE_EVENT_FORMAT() which does not generate the needed bits to tie them into perf counters. For those events, don't create the 'id' file and fail ->profile_enable when their ID is specified through other means. Reported-by: Chris Mason Signed-off-by: Peter Zijlstra Cc: Steven Rostedt LKML-Reference: <1249497664.5890.4.camel@laptop> [ v2: fix build error in the !CONFIG_EVENT_PROFILE case ] Signed-off-by: Ingo Molnar --- kernel/trace/trace_event_profile.c | 2 +- kernel/trace/trace_events.c | 2 +- 2 files changed, 2 insertions(+), 2 deletions(-) (limited to 'kernel/trace') diff --git a/kernel/trace/trace_event_profile.c b/kernel/trace/trace_event_profile.c index 5b5895afecf..11ba5bb4ed0 100644 --- a/kernel/trace/trace_event_profile.c +++ b/kernel/trace/trace_event_profile.c @@ -14,7 +14,7 @@ int ftrace_profile_enable(int event_id) mutex_lock(&event_mutex); list_for_each_entry(event, &ftrace_events, list) { - if (event->id == event_id) { + if (event->id == event_id && event->profile_enable) { ret = event->profile_enable(event); break; } diff --git a/kernel/trace/trace_events.c b/kernel/trace/trace_events.c index 23d2972b22d..e75276a49cf 100644 --- a/kernel/trace/trace_events.c +++ b/kernel/trace/trace_events.c @@ -940,7 +940,7 @@ event_create_dir(struct ftrace_event_call *call, struct dentry *d_events, entry = trace_create_file("enable", 0644, call->dir, call, enable); - if (call->id) + if (call->id && call->profile_enable) entry = trace_create_file("id", 0444, call->dir, call, id); -- cgit v1.2.3 From 469535a598f28c13a2a42037e1b778f671af1d16 Mon Sep 17 00:00:00 2001 From: Robert Richter Date: Thu, 30 Jul 2009 19:19:18 +0200 Subject: ring-buffer: Fix advance of reader in rb_buffer_peek() When calling rb_buffer_peek() from ring_buffer_consume() and a padding event is returned, the function rb_advance_reader() is called twice. This may lead to missing samples or under high workloads to the warning below. This patch fixes this. If a padding event is returned by rb_buffer_peek() it will be consumed by the calling function now. Also, I simplified some code in ring_buffer_consume(). ------------[ cut here ]------------ WARNING: at /dev/shm/.source/linux/kernel/trace/ring_buffer.c:2289 rb_advance_reader+0x2e/0xc5() Hardware name: Anaheim Modules linked in: Pid: 29, comm: events/2 Tainted: G W 2.6.31-rc3-oprofile-x86_64-standard-00059-g5050dc2 #1 Call Trace: [] ? rb_advance_reader+0x2e/0xc5 [] warn_slowpath_common+0x77/0x8f [] warn_slowpath_null+0xf/0x11 [] rb_advance_reader+0x2e/0xc5 [] ring_buffer_consume+0xa0/0xd2 [] op_cpu_buffer_read_entry+0x21/0x9e [] ? __find_get_block+0x4b/0x165 [] sync_buffer+0xa5/0x401 [] ? __find_get_block+0x4b/0x165 [] ? wq_sync_buffer+0x0/0x78 [] wq_sync_buffer+0x5b/0x78 [] worker_thread+0x113/0x1ac [] ? autoremove_wake_function+0x0/0x38 [] ? worker_thread+0x0/0x1ac [] kthread+0x88/0x92 [] child_rip+0xa/0x20 [] ? kthread+0x0/0x92 [] ? child_rip+0x0/0x20 ---[ end trace f561c0a58fcc89bd ]--- Cc: Steven Rostedt Cc: Signed-off-by: Robert Richter Signed-off-by: Ingo Molnar --- kernel/trace/ring_buffer.c | 10 ++++------ 1 file changed, 4 insertions(+), 6 deletions(-) (limited to 'kernel/trace') diff --git a/kernel/trace/ring_buffer.c b/kernel/trace/ring_buffer.c index 2606cee433d..d4d3580a894 100644 --- a/kernel/trace/ring_buffer.c +++ b/kernel/trace/ring_buffer.c @@ -2383,7 +2383,6 @@ rb_buffer_peek(struct ring_buffer *buffer, int cpu, u64 *ts) * the box. Return the padding, and we will release * the current locks, and try again. */ - rb_advance_reader(cpu_buffer); return event; case RINGBUF_TYPE_TIME_EXTEND: @@ -2519,6 +2518,8 @@ ring_buffer_peek(struct ring_buffer *buffer, int cpu, u64 *ts) if (dolock) spin_lock(&cpu_buffer->reader_lock); event = rb_buffer_peek(buffer, cpu, ts); + if (event && event->type_len == RINGBUF_TYPE_PADDING) + rb_advance_reader(cpu_buffer); if (dolock) spin_unlock(&cpu_buffer->reader_lock); local_irq_restore(flags); @@ -2590,12 +2591,9 @@ ring_buffer_consume(struct ring_buffer *buffer, int cpu, u64 *ts) spin_lock(&cpu_buffer->reader_lock); event = rb_buffer_peek(buffer, cpu, ts); - if (!event) - goto out_unlock; - - rb_advance_reader(cpu_buffer); + if (event) + rb_advance_reader(cpu_buffer); - out_unlock: if (dolock) spin_unlock(&cpu_buffer->reader_lock); local_irq_restore(flags); -- cgit v1.2.3 From bd3f02212d6a457267e0c9c02c426151c436d9d4 Mon Sep 17 00:00:00 2001 From: Eric Dumazet Date: Fri, 7 Aug 2009 12:49:29 +0200 Subject: ring-buffer: Fix memleak in ring_buffer_free() I noticed oprofile memleaked in linux-2.6 current tree, and tracked this ring-buffer leak. Signed-off-by: Eric Dumazet LKML-Reference: <4A7C06B9.2090302@gmail.com> Cc: stable@kernel.org Signed-off-by: Steven Rostedt --- kernel/trace/ring_buffer.c | 1 + 1 file changed, 1 insertion(+) (limited to 'kernel/trace') diff --git a/kernel/trace/ring_buffer.c b/kernel/trace/ring_buffer.c index d4d3580a894..a330513d96c 100644 --- a/kernel/trace/ring_buffer.c +++ b/kernel/trace/ring_buffer.c @@ -735,6 +735,7 @@ ring_buffer_free(struct ring_buffer *buffer) put_online_cpus(); + kfree(buffer->buffers); free_cpumask_var(buffer->cpumask); kfree(buffer); -- cgit v1.2.3 From 96b2de313b1e0e02aea80ee47df6a2b5cbdf8e13 Mon Sep 17 00:00:00 2001 From: Tom Zanussi Date: Sat, 8 Aug 2009 10:49:09 -0500 Subject: tracing/filters: Don't use pred on alloc failure Dan Carpenter sent me a fix to prevent pred from being used if it couldn't be allocated. I noticed the same problem also existed for the create_pred() case and added a fix for that. Reported-by: Dan Carpenter Signed-off-by: Tom Zanussi Cc: Steven Rostedt Cc: Frederic Weisbecker Cc: Li Zefan LKML-Reference: <1249746549.6453.29.camel@tropicana> Signed-off-by: Ingo Molnar --- kernel/trace/trace_events_filter.c | 4 ++++ 1 file changed, 4 insertions(+) (limited to 'kernel/trace') diff --git a/kernel/trace/trace_events_filter.c b/kernel/trace/trace_events_filter.c index 936c621bbf4..1557148be34 100644 --- a/kernel/trace/trace_events_filter.c +++ b/kernel/trace/trace_events_filter.c @@ -1029,6 +1029,8 @@ static int replace_preds(struct event_subsystem *system, if (elt->op == OP_AND || elt->op == OP_OR) { pred = create_logical_pred(elt->op); + if (!pred) + return -ENOMEM; if (call) { err = filter_add_pred(ps, call, pred); filter_free_pred(pred); @@ -1048,6 +1050,8 @@ static int replace_preds(struct event_subsystem *system, } pred = create_pred(elt->op, operand1, operand2); + if (!pred) + return -ENOMEM; if (call) { err = filter_add_pred(ps, call, pred); filter_free_pred(pred); -- cgit v1.2.3 From 26528e773ecc74fb1b61b7275f86f761cbb340ec Mon Sep 17 00:00:00 2001 From: Tom Zanussi Date: Sat, 8 Aug 2009 10:49:53 -0500 Subject: tracing/filters: Always free pred on filter_add_subsystem_pred() failure If filter_add_subsystem_pred() fails due to ENOSPC or ENOMEM, the pred doesn't get freed, while as a side effect it does for other errors. Make it so the caller always frees the pred for any error. Signed-off-by: Tom Zanussi Cc: Steven Rostedt Cc: Frederic Weisbecker Cc: Li Zefan LKML-Reference: <1249746593.6453.32.camel@tropicana> Signed-off-by: Ingo Molnar --- kernel/trace/trace_events_filter.c | 16 +++++++++++----- 1 file changed, 11 insertions(+), 5 deletions(-) (limited to 'kernel/trace') diff --git a/kernel/trace/trace_events_filter.c b/kernel/trace/trace_events_filter.c index 1557148be34..f32dc9d1ea7 100644 --- a/kernel/trace/trace_events_filter.c +++ b/kernel/trace/trace_events_filter.c @@ -624,9 +624,6 @@ static int filter_add_subsystem_pred(struct filter_parse_state *ps, return -ENOSPC; } - filter->preds[filter->n_preds] = pred; - filter->n_preds++; - list_for_each_entry(call, &ftrace_events, list) { if (!call->define_fields) @@ -643,6 +640,9 @@ static int filter_add_subsystem_pred(struct filter_parse_state *ps, } replace_filter_string(call->filter, filter_string); } + + filter->preds[filter->n_preds] = pred; + filter->n_preds++; out: return err; } @@ -1034,9 +1034,12 @@ static int replace_preds(struct event_subsystem *system, if (call) { err = filter_add_pred(ps, call, pred); filter_free_pred(pred); - } else + } else { err = filter_add_subsystem_pred(ps, system, pred, filter_string); + if (err) + filter_free_pred(pred); + } if (err) return err; @@ -1055,9 +1058,12 @@ static int replace_preds(struct event_subsystem *system, if (call) { err = filter_add_pred(ps, call, pred); filter_free_pred(pred); - } else + } else { err = filter_add_subsystem_pred(ps, system, pred, filter_string); + if (err) + filter_free_pred(pred); + } if (err) return err; -- cgit v1.2.3 From f413cdb80ce00ec1a4d0ab949b5d96c81cae7f75 Mon Sep 17 00:00:00 2001 From: Frederic Weisbecker Date: Fri, 7 Aug 2009 01:25:54 +0200 Subject: perf_counter: Fix/complete ftrace event records sampling This patch implements the kernel side support for ftrace event record sampling. A new counter sampling attribute is added: PERF_SAMPLE_TP_RECORD which requests ftrace events record sampling. In this case if a PERF_TYPE_TRACEPOINT counter is active and a tracepoint fires, we emit the tracepoint binary record to the perfcounter event buffer, as a sample. Result, after setting PERF_SAMPLE_TP_RECORD attribute from perf record: perf record -f -F 1 -a -e workqueue:workqueue_execution perf report -D 0x21e18 [0x48]: event: 9 . . ... raw event: size 72 bytes . 0000: 09 00 00 00 01 00 48 00 d0 c7 00 81 ff ff ff ff ......H........ . 0010: 0a 00 00 00 0a 00 00 00 21 00 00 00 00 00 00 00 ........!...... . 0020: 2b 00 01 02 0a 00 00 00 0a 00 00 00 65 76 65 6e +...........eve . 0030: 74 73 2f 31 00 00 00 00 00 00 00 00 0a 00 00 00 ts/1........... . 0040: e0 b1 31 81 ff ff ff ff ....... . 0x21e18 [0x48]: PERF_EVENT_SAMPLE (IP, 1): 10: 0xffffffff8100c7d0 period: 33 The raw ftrace binary record starts at offset 0020. Translation: struct trace_entry { type = 0x2b = 43; flags = 1; preempt_count = 2; pid = 0xa = 10; tgid = 0xa = 10; } thread_comm = "events/1" thread_pid = 0xa = 10; func = 0xffffffff8131b1e0 = flush_to_ldisc() What will come next? - Userspace support ('perf trace'), 'flight data recorder' mode for perf trace, etc. - The unconditional copy from the profiling callback brings some costs however if someone wants no such sampling to occur, and needs to be fixed in the future. For that we need to have an instant access to the perf counter attribute. This is a matter of a flag to add in the struct ftrace_event. - Take care of the events recursivity! Don't ever try to record a lock event for example, it seems some locking is used in the profiling fast path and lead to a tracing recursivity. That will be fixed using raw spinlock or recursivity protection. - [...] - Profit! :-) Signed-off-by: Frederic Weisbecker Cc: Li Zefan Cc: Tom Zanussi Cc: Arnaldo Carvalho de Melo Cc: Peter Zijlstra Cc: Mike Galbraith Cc: Steven Rostedt Cc: Paul Mackerras Cc: Pekka Enberg Cc: Gabriel Munteanu Cc: Lai Jiangshan Signed-off-by: Ingo Molnar --- kernel/trace/trace.c | 1 + kernel/trace/trace.h | 4 ---- 2 files changed, 1 insertion(+), 4 deletions(-) (limited to 'kernel/trace') diff --git a/kernel/trace/trace.c b/kernel/trace/trace.c index 8930e39b9d8..c22b40f8f57 100644 --- a/kernel/trace/trace.c +++ b/kernel/trace/trace.c @@ -848,6 +848,7 @@ tracing_generic_entry_update(struct trace_entry *entry, unsigned long flags, ((pc & SOFTIRQ_MASK) ? TRACE_FLAG_SOFTIRQ : 0) | (need_resched() ? TRACE_FLAG_NEED_RESCHED : 0); } +EXPORT_SYMBOL_GPL(tracing_generic_entry_update); struct ring_buffer_event *trace_buffer_lock_reserve(struct trace_array *tr, int type, diff --git a/kernel/trace/trace.h b/kernel/trace/trace.h index 3548ae5cc78..8b9f4f6e955 100644 --- a/kernel/trace/trace.h +++ b/kernel/trace/trace.h @@ -438,10 +438,6 @@ struct trace_entry *tracing_get_trace_entry(struct trace_array *tr, struct trace_entry *trace_find_next_entry(struct trace_iterator *iter, int *ent_cpu, u64 *ent_ts); -void tracing_generic_entry_update(struct trace_entry *entry, - unsigned long flags, - int pc); - void default_wait_pipe(struct trace_iterator *iter); void poll_wait_pipe(struct trace_iterator *iter); -- cgit v1.2.3 From 39cbb602b543e477df71dca84b5b2e36f8bd29fc Mon Sep 17 00:00:00 2001 From: "Alan D. Brunelle" Date: Fri, 7 Aug 2009 12:01:08 -0400 Subject: Remove double removal of blktrace directory commit fd51d251e4cdb21f68e9dbc4336514d64a105a79 Author: Stefan Raspl Date: Tue May 19 09:59:08 2009 +0200 blktrace: remove debugfs entries on bad path added in an explicit invocation of debugfs_remove for bt->dir, in blk_remove_buf_file_callback we are also getting the directory removed. On occasion I am seeing memory corruption that I have bisected down to this commit. [The testing involves a (long) series of I/O benchmarks with blktrace invoked around the actual runs.] I believe that this committed patch is correct, but the problem actually lies in the code in blk_remove_buf_file_callback. With this patch I am able to consistently get complete runs whereas previously I could not get a single run to complete. The first part of the patch simply moves the debugfs_remove below the relay_close: the relay_close call will remove files under bt->dir, and so we should not remove the directory until all the files we created have been removed. (Note: This is not sufficient to fix the problem - the file system code has ref counts on the directoy, so our invocation does not cause the directory to actually be removed. Nonetheless, we should not rely upon that feature.) Signed-off-by: Alan D. Brunelle Signed-off-by: Jens Axboe --- kernel/trace/blktrace.c | 12 +----------- 1 file changed, 1 insertion(+), 11 deletions(-) (limited to 'kernel/trace') diff --git a/kernel/trace/blktrace.c b/kernel/trace/blktrace.c index 1090b0aed9b..7a34cb563fe 100644 --- a/kernel/trace/blktrace.c +++ b/kernel/trace/blktrace.c @@ -267,8 +267,8 @@ static void blk_trace_free(struct blk_trace *bt) { debugfs_remove(bt->msg_file); debugfs_remove(bt->dropped_file); - debugfs_remove(bt->dir); relay_close(bt->rchan); + debugfs_remove(bt->dir); free_percpu(bt->sequence); free_percpu(bt->msg_data); kfree(bt); @@ -378,18 +378,8 @@ static int blk_subbuf_start_callback(struct rchan_buf *buf, void *subbuf, static int blk_remove_buf_file_callback(struct dentry *dentry) { - struct dentry *parent = dentry->d_parent; debugfs_remove(dentry); - /* - * this will fail for all but the last file, but that is ok. what we - * care about is the top level buts->name directory going away, when - * the last trace file is gone. Then we don't have to rmdir() that - * manually on trace stop, so it nicely solves the issue with - * force killing of running traces. - */ - - debugfs_remove(parent); return 0; } -- cgit v1.2.3