Hi Jiri,
On Sat, 8 Sep 2012 05:50:59 -0700, Arnaldo Carvalho de Melo wrote:
> Em Sat, Sep 08, 2012 at 02:35:14PM +0200, Jiri Olsa escreveu:
>> On Fri, Sep 07, 2012 at 11:55:02AM +0900, Namhyung Kim wrote:
>> > Hi, Jiri
>> >
>> > On Thu, 6 Sep 2012 17:46:59 +0200, Jiri Olsa wrote:
>> > > Currently for any of the data columns (like Overhead/Period..) in
>> > > stdio ui, there's separate code to print header/dots/value scattered
>> > > along the display code path.
>> > >
>> > > Adding hists_stdio_column struct to centralize all info needed
>> > > to print column header/dots/value.
>> > >
>> > > This change eases up addition for new columns, which is now mostly
>> > > matter only of adding new hists_stdio_column struct.
>> >
>> > As you may know, I submitted a similar patchset few days ago for the
>> > same reason and it handles TUI/GTK cases as well. I'm waiting for
>> > reviews.
>>
>> ok, I'll rebase this to acme/tmp.perf/hpp
>
> well, you can, but that one is buggy and you will not be able to test
> 'perf diff' at all...
I posted the fix right before, and you will also need patch below for
fixing broken "baseline" output.
>From 8f2d9010979e244c6565f3a9134e7bfa61936e38 Mon Sep 17 00:00:00 2001
From: Namhyung Kim <namhy...@kernel.org>
Date: Sat, 8 Sep 2012 23:28:59 +0900
Subject: [PATCH] perf hist: Fix perf diff baseline output
When perf diff is running, the overhead field should print old hist
entry's period as a baseline.
Signed-off-by: Namhyung Kim <namhy...@kernel.org>
---
tools/perf/ui/hist.c | 23 +++++++++++++++++++++++
1 file changed, 23 insertions(+)
diff --git a/tools/perf/ui/hist.c b/tools/perf/ui/hist.c
index 16dc486d02be..326f4e9e6911 100644
--- a/tools/perf/ui/hist.c
+++ b/tools/perf/ui/hist.c
@@ -21,6 +21,18 @@ static int hpp__width_overhead(struct perf_hpp *hpp __used)
static int hpp__color_overhead(struct perf_hpp *hpp, struct hist_entry *he)
{
double percent = 100.0 * he->period / hpp->total_period;
+
+ if (hpp->ptr) {
+ struct hists *old_hists = hpp->ptr;
+ u64 total_period = old_hists->stats.total_period;
+ u64 base_period = he->pair ? he->pair->period : 0;
+
+ if (total_period)
+ percent = 100.0 * base_period / total_period;
+ else
+ percent = 0.0;
+ }
+
return percent_color_snprintf(hpp->buf, hpp->size, " %5.2f%%",
percent);
}
@@ -29,6 +41,17 @@ static int hpp__entry_overhead(struct perf_hpp *hpp, struct
hist_entry *he)
double percent = 100.0 * he->period / hpp->total_period;
const char *fmt = symbol_conf.field_sep ? "%.2f" : " %5.2f%%";
+ if (hpp->ptr) {
+ struct hists *old_hists = hpp->ptr;
+ u64 total_period = old_hists->stats.total_period;
+ u64 base_period = he->pair ? he->pair->period : 0;
+
+ if (total_period)
+ percent = 100.0 * base_period / total_period;
+ else
+ percent = 0.0;
+ }
+
return scnprintf(hpp->buf, hpp->size, fmt, percent);
}
--
1.7.9.2
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majord...@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/
