From e1c7e2a6e67fe9db19dd15e71614526a31b5fdb1 Mon Sep 17 00:00:00 2001 From: Li Zefan Date: Wed, 24 Jun 2009 09:52:29 +0800 Subject: tracing/events: Don't increment @pos in s_start() While testing syscall tracepoints posted by Jason, I found 3 entries were missing when reading available_events. The output size of available_events is < 4 pages, which means we lost 1 entry per page. The cause is, it's wrong to increment @pos in s_start(). Actually there's another bug here -- reading avaiable_events/set_events can race with module unload: # cat available_events | s_start() | s_stop() | | # rmmod foo.ko s_start() | call = list_entry(m->private) | @call might be freed and accessing it will lead to crash. Reviewed-by: Liming Wang Signed-off-by: Li Zefan Cc: Steven Rostedt Cc: Frederic Weisbecker LKML-Reference: <4A4186DD.6090405@cn.fujitsu.com> Signed-off-by: Ingo Molnar --- kernel/trace/trace_events.c | 28 ++++++++++++++++++++++------ 1 file changed, 22 insertions(+), 6 deletions(-) diff --git a/kernel/trace/trace_events.c b/kernel/trace/trace_events.c index aa08be69a1b..53c8fd376a8 100644 --- a/kernel/trace/trace_events.c +++ b/kernel/trace/trace_events.c @@ -300,10 +300,18 @@ t_next(struct seq_file *m, void *v, loff_t *pos) static void *t_start(struct seq_file *m, loff_t *pos) { + struct ftrace_event_call *call = NULL; + loff_t l; + mutex_lock(&event_mutex); - if (*pos == 0) - m->private = ftrace_events.next; - return t_next(m, NULL, pos); + + m->private = ftrace_events.next; + for (l = 0; l <= *pos; ) { + call = t_next(m, NULL, &l); + if (!call) + break; + } + return call; } static void * @@ -332,10 +340,18 @@ s_next(struct seq_file *m, void *v, loff_t *pos) static void *s_start(struct seq_file *m, loff_t *pos) { + struct ftrace_event_call *call = NULL; + loff_t l; + mutex_lock(&event_mutex); - if (*pos == 0) - m->private = ftrace_events.next; - return s_next(m, NULL, pos); + + m->private = ftrace_events.next; + for (l = 0; l <= *pos; ) { + call = s_next(m, NULL, &l); + if (!call) + break; + } + return call; } static int t_show(struct seq_file *m, void *v) -- cgit v1.2.3 From c8961ec6da22ea010bf4470a8e0fb3fdad0f11c4 Mon Sep 17 00:00:00 2001 From: Li Zefan Date: Wed, 24 Jun 2009 09:52:58 +0800 Subject: tracing_bprintk: Don't increment @pos in t_start() It's wrong to increment @pos in t_start(), otherwise we'll lose some entries when reading printk_formats, if the output is larger than PAGE_SIZE. Reported-by: Lai Jiangshan Reviewed-by: Liming Wang Signed-off-by: Li Zefan Cc: Steven Rostedt Cc: Frederic Weisbecker LKML-Reference: <4A4186FA.1020106@cn.fujitsu.com> Signed-off-by: Ingo Molnar --- kernel/trace/trace_printk.c | 26 ++++++-------------------- 1 file changed, 6 insertions(+), 20 deletions(-) diff --git a/kernel/trace/trace_printk.c b/kernel/trace/trace_printk.c index 9bece9687b6..7b627811082 100644 --- a/kernel/trace/trace_printk.c +++ b/kernel/trace/trace_printk.c @@ -155,25 +155,19 @@ int __ftrace_vprintk(unsigned long ip, const char *fmt, va_list ap) EXPORT_SYMBOL_GPL(__ftrace_vprintk); static void * -t_next(struct seq_file *m, void *v, loff_t *pos) +t_start(struct seq_file *m, loff_t *pos) { - const char **fmt = m->private; - const char **next = fmt; - - (*pos)++; + const char **fmt = __start___trace_bprintk_fmt + *pos; if ((unsigned long)fmt >= (unsigned long)__stop___trace_bprintk_fmt) return NULL; - - next = fmt; - m->private = ++next; - return fmt; } -static void *t_start(struct seq_file *m, loff_t *pos) +static void *t_next(struct seq_file *m, void * v, loff_t *pos) { - return t_next(m, NULL, pos); + (*pos)++; + return t_start(m, pos); } static int t_show(struct seq_file *m, void *v) @@ -224,15 +218,7 @@ static const struct seq_operations show_format_seq_ops = { static int ftrace_formats_open(struct inode *inode, struct file *file) { - int ret; - - ret = seq_open(file, &show_format_seq_ops); - if (!ret) { - struct seq_file *m = file->private_data; - - m->private = __start___trace_bprintk_fmt; - } - return ret; + return seq_open(file, &show_format_seq_ops); } static const struct file_operations ftrace_formats_fops = { -- cgit v1.2.3 From 2961bf345fd1b736c3db46cad0f69855f67fbe9c Mon Sep 17 00:00:00 2001 From: Li Zefan Date: Wed, 24 Jun 2009 09:53:26 +0800 Subject: trace_stat: Don't increment @pos in seq start() It's wrong to increment @pos in stat_seq_start(). It causes some stat entries lost when reading stat file, if the output of the file is larger than PAGE_SIZE. Reviewed-by: Liming Wang Signed-off-by: Li Zefan Cc: Steven Rostedt Cc: Frederic Weisbecker LKML-Reference: <4A418716.90209@cn.fujitsu.com> Signed-off-by: Ingo Molnar --- kernel/trace/trace_stat.c | 6 +----- 1 file changed, 1 insertion(+), 5 deletions(-) diff --git a/kernel/trace/trace_stat.c b/kernel/trace/trace_stat.c index c00643733f4..e66f5e49334 100644 --- a/kernel/trace/trace_stat.c +++ b/kernel/trace/trace_stat.c @@ -199,17 +199,13 @@ static void *stat_seq_start(struct seq_file *s, loff_t *pos) mutex_lock(&session->stat_mutex); /* If we are in the beginning of the file, print the headers */ - if (!*pos && session->ts->stat_headers) { - (*pos)++; + if (!*pos && session->ts->stat_headers) return SEQ_START_TOKEN; - } node = rb_first(&session->stat_root); for (i = 0; node && i < *pos; i++) node = rb_next(node); - (*pos)++; - return node; } -- cgit v1.2.3 From f129e965bef40c6153e4fe505f1e408286213424 Mon Sep 17 00:00:00 2001 From: Li Zefan Date: Wed, 24 Jun 2009 09:53:44 +0800 Subject: tracing: Reset iterator in t_start() The iterator is m->private, but it's not reset to trace_types in t_start(). If the output is larger than PAGE_SIZE and t_start() is called the 2nd time, things will go wrong. Reviewed-by: Liming Wang Signed-off-by: Li Zefan Cc: Steven Rostedt Cc: Frederic Weisbecker LKML-Reference: <4A418728.5020506@cn.fujitsu.com> Signed-off-by: Ingo Molnar --- kernel/trace/trace.c | 18 ++++-------------- 1 file changed, 4 insertions(+), 14 deletions(-) diff --git a/kernel/trace/trace.c b/kernel/trace/trace.c index 076fa6f0ee4..3bb31006b5c 100644 --- a/kernel/trace/trace.c +++ b/kernel/trace/trace.c @@ -2053,25 +2053,23 @@ static int tracing_open(struct inode *inode, struct file *file) static void * t_next(struct seq_file *m, void *v, loff_t *pos) { - struct tracer *t = m->private; + struct tracer *t = v; (*pos)++; if (t) t = t->next; - m->private = t; - return t; } static void *t_start(struct seq_file *m, loff_t *pos) { - struct tracer *t = m->private; + struct tracer *t; loff_t l = 0; mutex_lock(&trace_types_lock); - for (; t && l < *pos; t = t_next(m, t, &l)) + for (t = trace_types; t && l < *pos; t = t_next(m, t, &l)) ; return t; @@ -2107,18 +2105,10 @@ static struct seq_operations show_traces_seq_ops = { static int show_traces_open(struct inode *inode, struct file *file) { - int ret; - if (tracing_disabled) return -ENODEV; - ret = seq_open(file, &show_traces_seq_ops); - if (!ret) { - struct seq_file *m = file->private_data; - m->private = trace_types; - } - - return ret; + return seq_open(file, &show_traces_seq_ops); } static ssize_t -- cgit v1.2.3 From 85951842a1020669f0a9eb0f0d1853b41341f097 Mon Sep 17 00:00:00 2001 From: Li Zefan Date: Wed, 24 Jun 2009 09:54:00 +0800 Subject: ftrace: Don't increment @pos in g_start() It's wrong to increment @pos in g_start(). It causes some entries lost when reading set_graph_function, if the output of the file is larger than PAGE_SIZE. Reviewed-by: Liming Wang Signed-off-by: Li Zefan Cc: Steven Rostedt Cc: Frederic Weisbecker LKML-Reference: <4A418738.7090401@cn.fujitsu.com> Signed-off-by: Ingo Molnar --- kernel/trace/ftrace.c | 21 ++++++++++----------- 1 file changed, 10 insertions(+), 11 deletions(-) diff --git a/kernel/trace/ftrace.c b/kernel/trace/ftrace.c index 3718d55fb4c..cde74b9973b 100644 --- a/kernel/trace/ftrace.c +++ b/kernel/trace/ftrace.c @@ -2500,32 +2500,31 @@ int ftrace_graph_count; unsigned long ftrace_graph_funcs[FTRACE_GRAPH_MAX_FUNCS] __read_mostly; static void * -g_next(struct seq_file *m, void *v, loff_t *pos) +__g_next(struct seq_file *m, loff_t *pos) { unsigned long *array = m->private; - int index = *pos; - - (*pos)++; - if (index >= ftrace_graph_count) + if (*pos >= ftrace_graph_count) return NULL; + return &array[*pos]; +} - return &array[index]; +static void * +g_next(struct seq_file *m, void *v, loff_t *pos) +{ + (*pos)++; + return __g_next(m, pos); } static void *g_start(struct seq_file *m, loff_t *pos) { - void *p = NULL; - mutex_lock(&graph_lock); /* Nothing, tell g_show to print all functions are enabled */ if (!ftrace_graph_count && !*pos) return (void *)1; - p = g_next(m, p, pos); - - return p; + return __g_next(m, pos); } static void g_stop(struct seq_file *m, void *p) -- cgit v1.2.3 From 694ce0a544fba37a60025a6803ee6265be8a2a22 Mon Sep 17 00:00:00 2001 From: Li Zefan Date: Wed, 24 Jun 2009 09:54:19 +0800 Subject: ftrace: Don't manipulate @pos in t_start() It's rather confusing that in t_start(), in some cases @pos is incremented, and in some cases it's decremented and then incremented. This patch rewrites t_start() in a much more general way. Thus we fix a bug that if ftrace_filtered == 1, functions have tracer hooks won't be printed, because the branch is always unreachable: static void *t_start(...) { ... if (!p) return t_hash_start(m, pos); return p; } Before: # echo 'sys_open' > /mnt/tracing/set_ftrace_filter # echo 'sys_write:traceon:4' >> /mnt/tracing/set_ftrace_filter sys_open After: # echo 'sys_open' > /mnt/tracing/set_ftrace_filter # echo 'sys_write:traceon:4' >> /mnt/tracing/set_ftrace_filter sys_open sys_write:traceon:count=4 Reviewed-by: Liming Wang Signed-off-by: Li Zefan Cc: Steven Rostedt Cc: Frederic Weisbecker LKML-Reference: <4A41874B.4090507@cn.fujitsu.com> Signed-off-by: Ingo Molnar --- kernel/trace/ftrace.c | 19 ++++++++----------- 1 file changed, 8 insertions(+), 11 deletions(-) diff --git a/kernel/trace/ftrace.c b/kernel/trace/ftrace.c index cde74b9973b..dc810208edd 100644 --- a/kernel/trace/ftrace.c +++ b/kernel/trace/ftrace.c @@ -1467,8 +1467,6 @@ t_next(struct seq_file *m, void *v, loff_t *pos) iter->pg = iter->pg->next; iter->idx = 0; goto retry; - } else { - iter->idx = -1; } } else { rec = &iter->pg->records[iter->idx++]; @@ -1497,6 +1495,7 @@ static void *t_start(struct seq_file *m, loff_t *pos) { struct ftrace_iterator *iter = m->private; void *p = NULL; + loff_t l; mutex_lock(&ftrace_lock); /* @@ -1508,23 +1507,21 @@ static void *t_start(struct seq_file *m, loff_t *pos) if (*pos > 0) return t_hash_start(m, pos); iter->flags |= FTRACE_ITER_PRINTALL; - (*pos)++; return iter; } if (iter->flags & FTRACE_ITER_HASH) return t_hash_start(m, pos); - if (*pos > 0) { - if (iter->idx < 0) - return p; - (*pos)--; - iter->idx--; + iter->pg = ftrace_pages_start; + iter->idx = 0; + for (l = 0; l <= *pos; ) { + p = t_next(m, p, &l); + if (!p) + break; } - p = t_next(m, p, pos); - - if (!p) + if (!p && iter->flags & FTRACE_ITER_FILTER) return t_hash_start(m, pos); return p; -- cgit v1.2.3 From d82d62444f87e5993af2fa82ed636b2206e052ea Mon Sep 17 00:00:00 2001 From: Li Zefan Date: Wed, 24 Jun 2009 09:54:54 +0800 Subject: ftrace: Fix t_hash_start() When the output of set_ftrace_filter is larger than PAGE_SIZE, t_hash_start() will be called the 2nd time, and then we start from the head of a hlist, which is wrong and causes some entries to be outputed twice. The worse is, if the hlist is large enough, reading set_ftrace_filter won't stop but in a dead loop. Reviewed-by: Liming Wang Signed-off-by: Li Zefan Cc: Steven Rostedt Cc: Frederic Weisbecker LKML-Reference: <4A41876E.2060407@cn.fujitsu.com> Signed-off-by: Ingo Molnar --- kernel/trace/ftrace.c | 12 +++++++++++- 1 file changed, 11 insertions(+), 1 deletion(-) diff --git a/kernel/trace/ftrace.c b/kernel/trace/ftrace.c index dc810208edd..71a52c17214 100644 --- a/kernel/trace/ftrace.c +++ b/kernel/trace/ftrace.c @@ -1417,10 +1417,20 @@ static void *t_hash_start(struct seq_file *m, loff_t *pos) { struct ftrace_iterator *iter = m->private; void *p = NULL; + loff_t l; + + if (!(iter->flags & FTRACE_ITER_HASH)) + *pos = 0; iter->flags |= FTRACE_ITER_HASH; - return t_hash_next(m, p, pos); + iter->hidx = 0; + for (l = 0; l <= *pos; ) { + p = t_hash_next(m, p, &l); + if (!p) + break; + } + return p; } static int t_hash_show(struct seq_file *m, void *v) -- cgit v1.2.3 From 9d612beff5089b89a295a2331883a8ce3fff08c1 Mon Sep 17 00:00:00 2001 From: Li Zefan Date: Wed, 24 Jun 2009 17:33:15 +0800 Subject: tracing: Fix trace_buf_size boot option We should be able to specify [KMG] when setting trace_buf_size boot option, as documented in kernel-parameters.txt Signed-off-by: Li Zefan Cc: Steven Rostedt Cc: Frederic Weisbecker LKML-Reference: <4A41F2DB.4020102@cn.fujitsu.com> Signed-off-by: Ingo Molnar --- Documentation/kernel-parameters.txt | 3 ++- kernel/trace/trace.c | 5 ++--- 2 files changed, 4 insertions(+), 4 deletions(-) diff --git a/Documentation/kernel-parameters.txt b/Documentation/kernel-parameters.txt index 92e1ab8178a..d3f41db3ed4 100644 --- a/Documentation/kernel-parameters.txt +++ b/Documentation/kernel-parameters.txt @@ -2475,7 +2475,8 @@ and is between 256 and 4096 characters. It is defined in the file tp720= [HW,PS2] - trace_buf_size=nn[KMG] [ftrace] will set tracing buffer size. + trace_buf_size=nn[KMG] + [FTRACE] will set tracing buffer size. trix= [HW,OSS] MediaTrix AudioTrix Pro Format: diff --git a/kernel/trace/trace.c b/kernel/trace/trace.c index 3bb31006b5c..3aa0a0dfdfa 100644 --- a/kernel/trace/trace.c +++ b/kernel/trace/trace.c @@ -284,13 +284,12 @@ void trace_wake_up(void) static int __init set_buf_size(char *str) { unsigned long buf_size; - int ret; if (!str) return 0; - ret = strict_strtoul(str, 0, &buf_size); + buf_size = memparse(str, &str); /* nr_entries can not be zero */ - if (ret < 0 || buf_size == 0) + if (buf_size == 0) return 0; trace_buf_size = buf_size; return 1; -- cgit v1.2.3 From 00e54d087afb3867b0b461aef6c1ff433d0df564 Mon Sep 17 00:00:00 2001 From: Li Zefan Date: Thu, 25 Jun 2009 14:05:27 +0800 Subject: ftrace: Remove duplicate newline Before: # echo 'sys_open:traceon:' > set_ftrace_filter # echo 'sys_close:traceoff:5' > set_ftrace_filter # cat set_ftrace_filter #### all functions enabled #### sys_open:traceon:unlimited sys_close:traceoff:count=0 After: # cat set_ftrace_filter #### all functions enabled #### sys_open:traceon:unlimited sys_close:traceoff:count=0 Signed-off-by: Li Zefan Cc: Steven Rostedt Cc: Frederic Weisbecker LKML-Reference: <4A4313A7.7030105@cn.fujitsu.com> Signed-off-by: Ingo Molnar --- kernel/trace/trace_functions.c | 3 +-- 1 file changed, 1 insertion(+), 2 deletions(-) diff --git a/kernel/trace/trace_functions.c b/kernel/trace/trace_functions.c index 90f13476483..7402144bff2 100644 --- a/kernel/trace/trace_functions.c +++ b/kernel/trace/trace_functions.c @@ -302,8 +302,7 @@ ftrace_trace_onoff_print(struct seq_file *m, unsigned long ip, if (count == -1) seq_printf(m, ":unlimited\n"); else - seq_printf(m, ":count=%ld", count); - seq_putc(m, '\n'); + seq_printf(m, ":count=%ld\n", count); return 0; } -- cgit v1.2.3 From 1155de47cd66d0c496d5a6fb2223e980ef1285b2 Mon Sep 17 00:00:00 2001 From: Paul Mundt Date: Thu, 25 Jun 2009 14:30:12 +0900 Subject: ring-buffer: Make it generally available In hunting down the cause for the hwlat_detector ring buffer spew in my failed -next builds it became obvious that folks are now treating ring_buffer as something that is generic independent of tracing and thus, suitable for public driver consumption. Given that there are only a few minor areas in ring_buffer that have any reliance on CONFIG_TRACING or CONFIG_FUNCTION_TRACER, provide stubs for those and make it generally available. Signed-off-by: Paul Mundt Cc: Jon Masters Cc: Steven Rostedt LKML-Reference: <20090625053012.GB19944@linux-sh.org> Signed-off-by: Ingo Molnar --- kernel/Makefile | 1 + kernel/trace/ring_buffer.c | 11 +++++++++++ kernel/trace/trace.h | 7 +++++++ 3 files changed, 19 insertions(+) diff --git a/kernel/Makefile b/kernel/Makefile index 0a32cb21ec9..0630e293cd4 100644 --- a/kernel/Makefile +++ b/kernel/Makefile @@ -96,6 +96,7 @@ obj-$(CONFIG_HAVE_GENERIC_DMA_COHERENT) += dma-coherent.o obj-$(CONFIG_FUNCTION_TRACER) += trace/ obj-$(CONFIG_TRACING) += trace/ obj-$(CONFIG_X86_DS) += trace/ +obj-$(CONFIG_RING_BUFFER) += trace/ obj-$(CONFIG_SMP) += sched_cpupri.o obj-$(CONFIG_SLOW_WORK) += slow-work.o obj-$(CONFIG_PERF_COUNTERS) += perf_counter.o diff --git a/kernel/trace/ring_buffer.c b/kernel/trace/ring_buffer.c index 04dac263825..bf27bb7a63e 100644 --- a/kernel/trace/ring_buffer.c +++ b/kernel/trace/ring_buffer.c @@ -1563,6 +1563,8 @@ rb_reserve_next_event(struct ring_buffer_per_cpu *cpu_buffer, return NULL; } +#ifdef CONFIG_TRACING + #define TRACE_RECURSIVE_DEPTH 16 static int trace_recursive_lock(void) @@ -1593,6 +1595,13 @@ static void trace_recursive_unlock(void) current->trace_recursion--; } +#else + +#define trace_recursive_lock() (0) +#define trace_recursive_unlock() do { } while (0) + +#endif + static DEFINE_PER_CPU(int, rb_need_resched); /** @@ -3104,6 +3113,7 @@ int ring_buffer_read_page(struct ring_buffer *buffer, } EXPORT_SYMBOL_GPL(ring_buffer_read_page); +#ifdef CONFIG_TRACING static ssize_t rb_simple_read(struct file *filp, char __user *ubuf, size_t cnt, loff_t *ppos) @@ -3171,6 +3181,7 @@ static __init int rb_init_debugfs(void) } fs_initcall(rb_init_debugfs); +#endif #ifdef CONFIG_HOTPLUG_CPU static int rb_cpu_notify(struct notifier_block *self, diff --git a/kernel/trace/trace.h b/kernel/trace/trace.h index 6e735d4771f..3548ae5cc78 100644 --- a/kernel/trace/trace.h +++ b/kernel/trace/trace.h @@ -597,6 +597,7 @@ print_graph_function(struct trace_iterator *iter) extern struct pid *ftrace_pid_trace; +#ifdef CONFIG_FUNCTION_TRACER static inline int ftrace_trace_task(struct task_struct *task) { if (!ftrace_pid_trace) @@ -604,6 +605,12 @@ static inline int ftrace_trace_task(struct task_struct *task) return test_tsk_trace_trace(task); } +#else +static inline int ftrace_trace_task(struct task_struct *task) +{ + return 1; +} +#endif /* * trace_iterator_flags is an enumeration that defines bit -- cgit v1.2.3 From 0296e4254f3318e0dcad9706fa1daf8e5addc1e9 Mon Sep 17 00:00:00 2001 From: Li Zefan Date: Fri, 26 Jun 2009 11:15:37 +0800 Subject: ftrace: Fix the output of profile The first entry of the ftrace profile was always skipped when reading trace_stat/functionX. Signed-off-by: Li Zefan Cc: Steven Rostedt Cc: Frederic Weisbecker LKML-Reference: <4A443D59.4080307@cn.fujitsu.com> Signed-off-by: Ingo Molnar --- kernel/trace/ftrace.c | 4 +++- 1 file changed, 3 insertions(+), 1 deletion(-) diff --git a/kernel/trace/ftrace.c b/kernel/trace/ftrace.c index 71a52c17214..f3716bf04df 100644 --- a/kernel/trace/ftrace.c +++ b/kernel/trace/ftrace.c @@ -291,7 +291,9 @@ function_stat_next(void *v, int idx) pg = (struct ftrace_profile_page *)((unsigned long)rec & PAGE_MASK); again: - rec++; + if (idx != 0) + rec++; + if ((void *)rec >= (void *)&pg->records[pg->index]) { pg = pg->next; if (!pg) -- cgit v1.2.3