2015-12-21 07:55:14

by Namhyung Kim

Subject: [PATCH] trace-cmd: Print relate stacktrace at once

From: Namhyung Kim <[email protected]>

Currently trace-cmd prints ring buffer events in a strict time order.
But it sometimes annoying that stacktrace from the same cpu can be
intermixed with events from other cpu. This patch looks next event when
print last record and prints it also if it's a stacktrace (from the same
cpu).

Requested-by: Joonsoo Kim <[email protected]>
Signed-off-by: Namhyung Kim <[email protected]>
---
trace-read.c | 11 +++++++++++
1 file changed, 11 insertions(+)

diff --git a/trace-read.c b/trace-read.c
index aec8532..10edf5d 100644
--- a/trace-read.c
+++ b/trace-read.c
@@ -1155,9 +1155,20 @@ static void read_data_info(struct list_head *handle_list, enum output_type otype
}
}
if (last_record) {
+ int last_cpu = last_record->cpu;
+
print_handle_file(last_handle);
trace_show_data(last_handle->handle, last_record, profile);
free_handle_record(last_handle);
+
+ /* print related stacktrace at once */
+ record = tracecmd_peek_data(last_handle->handle, last_cpu);
+ if (record && pevent_data_type(pevent, record) == stacktrace_id &&
+ test_stacktrace(last_handle, record, 1)) {
+ print_handle_file(last_handle);
+ trace_show_data(last_handle->handle, record, profile);
+ tracecmd_read_data(last_handle->handle, last_cpu);
+ }
}
} while (last_record);

--
2.6.4

2015-12-21 12:43:46

by Joonsoo Kim

Subject: Re: [PATCH] trace-cmd: Print relate stacktrace at once

2015-12-21 16:54 GMT+09:00 Namhyung Kim <[email protected]>:
> From: Namhyung Kim <[email protected]>
>
> Currently trace-cmd prints ring buffer events in a strict time order.
> But it sometimes annoying that stacktrace from the same cpu can be
> intermixed with events from other cpu. This patch looks next event when
> print last record and prints it also if it's a stacktrace (from the same
> cpu).
>
> Requested-by: Joonsoo Kim <[email protected]>
> Signed-off-by: Namhyung Kim <[email protected]>

Good Job!
I tested it and it seems that it works well for me.

Thanks!!!