Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]

Groups > linux.kernel > #1682803

[PATCH 3/3] tracing: Attempt to record other information even if some fail

Show all headers | View raw

In recent patches where we record comm and tgid at the same time, we skip
continuing to record if any fail. Fix that by trying to record as many things
as we can even if some couldn't be recorded. If any information isn't recorded,
then we don't set trace_taskinfo_save as before.

Cc: kernel-team@android.com
Cc: Steven Rostedt <rostedt@goodmis.org>
Cc: Ingo Molnar <mingo@redhat.com>
Signed-off-by: Joel Fernandes <joelaf@google.com>
---
 kernel/trace/trace.c | 32 ++++++++++++++++++++++++--------
 1 file changed, 24 insertions(+), 8 deletions(-)

diff --git a/kernel/trace/trace.c b/kernel/trace/trace.c
index 363924f90e2f..8a44e9cb2594 100644
--- a/kernel/trace/trace.c
+++ b/kernel/trace/trace.c
@@ -2037,11 +2037,20 @@ static bool tracing_record_taskinfo_skip(int flags)
  */
 void tracing_record_taskinfo(struct task_struct *task, int flags)
 {
+	bool done;
+
 	if (tracing_record_taskinfo_skip(flags))
 		return;
-	if ((flags & TRACE_RECORD_CMDLINE) && !trace_save_cmdline(task))
-		return;
-	if ((flags & TRACE_RECORD_TGID) && !trace_save_tgid(task))
+
+	/*
+	 * Record as much task information as possible. If some fail, continue
+	 * to try to record the others.
+	 */
+	done = !(flags & TRACE_RECORD_CMDLINE) || trace_save_cmdline(task);
+	done &= !(flags & TRACE_RECORD_TGID) || trace_save_tgid(task);
+
+	/* If recording any information failed, retry again soon. */
+	if (!done)
 		return;
 
 	__this_cpu_write(trace_taskinfo_save, false);
@@ -2058,15 +2067,22 @@ void tracing_record_taskinfo(struct task_struct *task, int flags)
 void tracing_record_taskinfo_sched_switch(struct task_struct *prev,
 					  struct task_struct *next, int flags)
 {
+	bool done;
+
 	if (tracing_record_taskinfo_skip(flags))
 		return;
 
-	if ((flags & TRACE_RECORD_CMDLINE) &&
-	    (!trace_save_cmdline(prev) || !trace_save_cmdline(next)))
-		return;
+	/*
+	 * Record as much task information as possible. If some fail, continue
+	 * to try to record the others.
+	 */
+	done  = !(flags & TRACE_RECORD_CMDLINE) || trace_save_cmdline(prev);
+	done &= !(flags & TRACE_RECORD_CMDLINE) || trace_save_cmdline(next);
+	done &= !(flags & TRACE_RECORD_TGID) || trace_save_tgid(prev);
+	done &= !(flags & TRACE_RECORD_TGID) || trace_save_tgid(next);
 
-	if ((flags & TRACE_RECORD_TGID) &&
-	    (!trace_save_tgid(prev) || !trace_save_tgid(next)))
+	/* If recording any information failed, retry again soon. */
+	if (!done)
 		return;
 
 	__this_cpu_write(trace_taskinfo_save, false);
-- 
2.13.2.725.g09c95d1e9-goog

Back to linux.kernel | Previous | Next — Previous in thread | Find similar | Unroll thread

Thread

[PATCH 1/3] tracing: Treat recording comm for idle task as a success Joel Fernandes <joelaf@google.com> - 2017-07-07 01:10 +0200
  [PATCH 3/3] tracing: Attempt to record other information even if some fail Joel Fernandes <joelaf@google.com> - 2017-07-07 01:10 +0200

csiph-web