Skip to contentSkip to conversation

update for tip/tracing/ftrace

GIT PULL New Steven Rostedt 2 participants 5 messages Mar 17, 2009

Subsystems (3)
lore
Share & tools
lore

Steven Rostedt

author maintainer
#0 lore raw

Ingo,

Please pull the latest tip/tracing/ftrace tree, which can be found at:

git://git.kernel.org/pub/scm/linux/kernel/git/rostedt/linux-2.6-trace.git
tip/tracing/ftrace

Frederic Weisbecker (1):
tracing/ftrace: fix the check on nopped sites

Steven Rostedt (2):
tracing: protect reader of cmdline output
tracing: stop comm recording on tracing off

----

 kernel/trace/blktrace.c              |   23 ++++++++++++++++++-----
 kernel/trace/ftrace.c                |    2 +-
 kernel/trace/trace.c                 |   22 +++++++++++++---------
 kernel/trace/trace.h                 |    2 +-
 kernel/trace/trace_functions_graph.c |   12 ++++++------
 kernel/trace/trace_output.c          |   18 ++++++++++++------
 6 files changed, 51 insertions(+), 28 deletions(-)
author maintainer
#1 lore raw

PATCH 1/3 tracing/ftrace: fix the check on nopped sites

From: Frederic Weisbecker <[email protected]>

Impact: fix a dynamic tracing failure

Recently, the function and function graph tracers failed to use dynamic tracing after the following commit:

fa9d13cf135efbd454453a53b6299976bea245a9
(ftrace: don't try to __ftrace_replace_code on !FTRACE_FL_CONVERTED rec)

The patch is right except a mistake on the check for the FTRACE_FL_CONVERTED flag. The code patching is aborted in case of successfully nopped sites. What we want is the opposite: ignore the callsites that haven't been nopped.

---
 kernel/trace/ftrace.c |    2 +-
 1 files changed, 1 insertions(+), 1 deletions(-)
kernel/trace/ftrace.c +1 -1
index 90d5729..7847806 100644
@@ -537,7 +537,7 @@ static void ftrace_replace_code(int enable)
 		 */
 		if (rec->flags & FTRACE_FL_FREE ||
 		    rec->flags & FTRACE_FL_FAILED ||
-		    rec->flags & FTRACE_FL_CONVERTED)
+		    !(rec->flags & FTRACE_FL_CONVERTED))
 			continue;
 
 		/* ignore updates to this record's mcount site */
Signature · 3 lines
1.6.2

-- 
author maintainer
#3 lore raw

PATCH 3/3 tracing: stop comm recording on tracing off

In reply to Steven Rostedt

From: Steven Rostedt <[email protected]>

Impact: fix for losing comms in trace

The command lines of tasks are cached at sched switch to not need to record them at every trace point. Disabling the tracing on stops the recording of traces, but does not stop the caching of command lines. When the tracing is off the cache may overflow and cause the tracing to show incorrect tasks matching the PIDs.

This patch disables prevents updates to the comm cache when the ring buffer is off.

---
 kernel/trace/trace.c |    2 +-
 1 files changed, 1 insertions(+), 1 deletions(-)
kernel/trace/trace.c +1 -1
index 2796bd2..8f89690 100644
@@ -797,7 +797,7 @@ void trace_find_cmdline(int pid, char comm[])
 
 void tracing_record_cmdline(struct task_struct *tsk)
 {
-	if (atomic_read(&trace_record_cmdline_disabled))
+	if (atomic_read(&trace_record_cmdline_disabled) || !tracing_is_on())
 		return;
 
 	trace_save_cmdline(tsk);
Signature · 3 lines
1.6.2

-- 
author maintainer
#2 lore raw

PATCH 2/3 tracing: protect reader of cmdline output

In reply to Steven Rostedt

From: Steven Rostedt <[email protected]>

Impact: fix to one cause of incorrect comm outputs in trace

The spinlock only protected the creation of a comm <=> pid pair. But it was possible that a reader could look up a pid, and get the wrong comm because it had no locking.

This also required changing trace_find_cmdline to copy the comm cache and not just send back a pointer to it.

---
 kernel/trace/blktrace.c              |   23 ++++++++++++++++++-----
 kernel/trace/trace.c                 |   20 ++++++++++++--------
 kernel/trace/trace.h                 |    2 +-
 kernel/trace/trace_functions_graph.c |   12 ++++++------
 kernel/trace/trace_output.c          |   18 ++++++++++++------
 5 files changed, 49 insertions(+), 26 deletions(-)
kernel/trace/blktrace.c +18 -5
index 1f32e4e..b171778 100644
@@ -1027,7 +1027,9 @@ static int blk_log_action_seq(struct trace_seq *s, const struct blk_io_trace *t,
 
 static int blk_log_generic(struct trace_seq *s, const struct trace_entry *ent)
 {
-	const char *cmd = trace_find_cmdline(ent->pid);
+	char cmd[TASK_COMM_LEN];
+
+	trace_find_cmdline(ent->pid, cmd);
 
 	if (t_sec(ent))
 		return trace_seq_printf(s, "%llu + %u [%s]\n",
@@ -1057,19 +1059,30 @@ static int blk_log_remap(struct trace_seq *s, const struct trace_entry *ent)
 
 static int blk_log_plug(struct trace_seq *s, const struct trace_entry *ent)
 {
-	return trace_seq_printf(s, "[%s]\n", trace_find_cmdline(ent->pid));
+	char cmd[TASK_COMM_LEN];
+
+	trace_find_cmdline(ent->pid, cmd);
+
+	return trace_seq_printf(s, "[%s]\n", cmd);
 }
 
 static int blk_log_unplug(struct trace_seq *s, const struct trace_entry *ent)
 {
-	return trace_seq_printf(s, "[%s] %llu\n", trace_find_cmdline(ent->pid),
-				get_pdu_int(ent));
+	char cmd[TASK_COMM_LEN];
+
+	trace_find_cmdline(ent->pid, cmd);
+
+	return trace_seq_printf(s, "[%s] %llu\n", cmd, get_pdu_int(ent));
 }
 
 static int blk_log_split(struct trace_seq *s, const struct trace_entry *ent)
 {
+	char cmd[TASK_COMM_LEN];
+
+	trace_find_cmdline(ent->pid, cmd);
+
 	return trace_seq_printf(s, "%llu / %llu [%s]\n", t_sector(ent),
-				get_pdu_int(ent), trace_find_cmdline(ent->pid));
+				get_pdu_int(ent), cmd);
 }
 
 /*
kernel/trace/trace.c +12 -8
index efe3202..2796bd2 100644
@@ -770,25 +770,29 @@ static void trace_save_cmdline(struct task_struct *tsk)
 	__raw_spin_unlock(&trace_cmdline_lock);
 }
 
-char *trace_find_cmdline(int pid)
+void trace_find_cmdline(int pid, char comm[])
 {
-	char *cmdline = "<...>";
 	unsigned map;
 
-	if (!pid)
-		return "<idle>";
+	if (!pid) {
+		strcpy(comm, "<idle>");
+		return;
+	}
 
-	if (pid > PID_MAX_DEFAULT)
-		goto out;
+	if (pid > PID_MAX_DEFAULT) {
+		strcpy(comm, "<...>");
+		return;
+	}
 
+	__raw_spin_lock(&trace_cmdline_lock);
 	map = map_pid_to_cmdline[pid];
 	if (map >= SAVED_CMDLINES)
 		goto out;
 
-	cmdline = saved_cmdlines[map];
+	strcpy(comm, saved_cmdlines[map]);
 
  out:
-	return cmdline;
+	__raw_spin_unlock(&trace_cmdline_lock);
 }
 
 void tracing_record_cmdline(struct task_struct *tsk)
kernel/trace/trace.h +1 -1
index 56ce34d..b0ecad8 100644
@@ -547,7 +547,7 @@ struct tracer_switch_ops {
 };
 #endif /* CONFIG_CONTEXT_SWITCH_TRACER */
 
-extern char *trace_find_cmdline(int pid);
+extern void trace_find_cmdline(int pid, char comm[]);
 
 #ifdef CONFIG_DYNAMIC_FTRACE
 extern unsigned long ftrace_update_tot_cnt;
kernel/trace/trace_functions_graph.c +6 -6
index 4c38860..6004cca 100644
@@ -190,15 +190,15 @@ print_graph_cpu(struct trace_seq *s, int cpu)
 static enum print_line_t
 print_graph_proc(struct trace_seq *s, pid_t pid)
 {
-	int i;
-	int ret;
-	int len;
-	char comm[8];
-	int spaces = 0;
+	char comm[TASK_COMM_LEN];
 	/* sign + log10(MAX_INT) + '\0' */
 	char pid_str[11];
+	int spaces = 0;
+	int ret;
+	int len;
+	int i;
 
-	strncpy(comm, trace_find_cmdline(pid), 7);
+	trace_find_cmdline(pid, comm);
 	comm[7] = '\0';
 	sprintf(pid_str, "%d", pid);
 
kernel/trace/trace_output.c +12 -6
index ea9d3b4..6a4c9de 100644
@@ -309,9 +309,9 @@ static int
 lat_print_generic(struct trace_seq *s, struct trace_entry *entry, int cpu)
 {
 	int hardirq, softirq;
-	char *comm;
+	char comm[TASK_COMM_LEN];
 
-	comm = trace_find_cmdline(entry->pid);
+	trace_find_cmdline(entry->pid, comm);
 	hardirq = entry->flags & TRACE_FLAG_HARDIRQ;
 	softirq = entry->flags & TRACE_FLAG_SOFTIRQ;
 
@@ -346,10 +346,12 @@ int trace_print_context(struct trace_iterator *iter)
 {
 	struct trace_seq *s = &iter->seq;
 	struct trace_entry *entry = iter->ent;
-	char *comm = trace_find_cmdline(entry->pid);
 	unsigned long long t = ns2usecs(iter->ts);
 	unsigned long usec_rem = do_div(t, USEC_PER_SEC);
 	unsigned long secs = (unsigned long)t;
+	char comm[TASK_COMM_LEN];
+
+	trace_find_cmdline(entry->pid, comm);
 
 	return trace_seq_printf(s, "%16s-%-5d [%03d] %5lu.%06lu: ",
 				comm, entry->pid, iter->cpu, secs, usec_rem);
@@ -372,7 +374,10 @@ int trace_print_lat_context(struct trace_iterator *iter)
 	rel_usecs = ns2usecs(next_ts - iter->ts);
 
 	if (verbose) {
-		char *comm = trace_find_cmdline(entry->pid);
+		char comm[TASK_COMM_LEN];
+
+		trace_find_cmdline(entry->pid, comm);
+
 		ret = trace_seq_printf(s, "%16s %5d %3d %d %08x %08lx [%08lx]"
 				       " %ld.%03ldms (+%ld.%03ldms): ", comm,
 				       entry->pid, iter->cpu, entry->flags,
@@ -577,14 +582,15 @@ static enum print_line_t trace_ctxwake_print(struct trace_iterator *iter,
 					     char *delim)
 {
 	struct ctx_switch_entry *field;
-	char *comm;
+	char comm[TASK_COMM_LEN];
 	int S, T;
 
+
 	trace_assign_type(field, iter->ent);
 
 	T = task_state_char(field->next_state);
 	S = task_state_char(field->prev_state);
-	comm = trace_find_cmdline(field->next_pid);
+	trace_find_cmdline(field->next_pid, comm);
 	if (!trace_seq_printf(&iter->seq,
 			      " %5d:%3d:%c %s [%03d] %5d:%3d:%c %s\n",
 			      field->prev_pid,
Signature · 3 lines
1.6.2

-- 
#4 lore raw
In reply to Steven Rostedt
Quoted 15 lines · Steven Rostedt

Ingo,

Please pull the latest tip/tracing/ftrace tree, which can be found at:

git://git.kernel.org/pub/scm/linux/kernel/git/rostedt/linux-2.6-trace.git
tip/tracing/ftrace

Frederic Weisbecker (1):
tracing/ftrace: fix the check on nopped sites

Steven Rostedt (2):
tracing: protect reader of cmdline output
tracing: stop comm recording on tracing off

Quoted 8 lines

----

 kernel/trace/blktrace.c              |   23 ++++++++++++++++++-----
 kernel/trace/ftrace.c                |    2 +-
 kernel/trace/trace.c                 |   22 +++++++++++++---------
 kernel/trace/trace.h                 |    2 +-
 kernel/trace/trace_functions_graph.c |   12 ++++++------
 kernel/trace/trace_output.c          |   18 ++++++++++++------
 6 files changed, 51 insertions(+), 28 deletions(-)

Pulled, thanks Steve!

Ingo