perf report: Add support for event group view (v8)
Namhyung Kim
maintainercover perf report: Add support for event group view (v8)
Hi,
This is my 8th attempt to enable the event group view for perf report. For basic idea and usage example, please see my initial post [1].
This version is rebased on the current acme/perf/core and it's mainly changed due to a refactoring of gtk browser.
You can get this via 'perf/group-v8' branch on my tree:
git://git.kernel.org/pub/scm/linux/kernel/git/namhyung/linux-perf.git
Any comments are welcome, thanks,
Namhyung
v7 -> v8:
* rebase onto acme/perf/core
* patch 3 in the previous version is merged
* add Jiri's Acked-by's
v6 -> v7:
* hists__{match,link} changes are merged separately
* factor out inc_group_count() from parsing group_def (Jiri)
* add checking for group fields in evlist/evsel (Jiri)
* check return value of during header processing (Arnaldo)
* get rid of a temporal array in hpp functions (Arnaldo)
* convert hpp macros to static inline functions (Jiri)
v5 -> v6:
* set ->leader alse for leader evsel (Arnaldo)
* use hists__{match,link} (Arnaldo)
v4 -> v5:
* rebase onto acme/perf/core
v3 -> v4:
* patch 1-9 in previous post are merged.
* add Jiri's Acked-by
* add report.group config option
v2 -> v3:
* drop patch 1 since it's merged into acme/perf/core
* cherry-pick Jiri's hpp changes
* add missing bswap_32 on reading nr_groups (Jiri)
* remove perf_evlist__recalc_nr_groups() in favor of list_is_last (Jiri)
v1 -> v2:
* save group relation to header (Jiri)
[1] https://lkml.org/lkml/2012/7/24/81
Namhyung Kim (13):
perf tools: Keep group information
perf tests: Add group test conditions
perf header: Add HEADER_GROUP_DESC feature
perf report: Make another loop for linking group hists perf hists: Resort hist entries using group members for output perf ui/hist: Add support for event group view
perf hist browser: Add support for event group view
perf gtk/browser: Add support for event group view
perf gtk/browser: Trim column header string when event group enabled perf report: Bypass non-leader events when event group is enabled perf report: Show group description when event group is enabled perf report: Add --group option
perf report: Add report.group config option
tools/perf/Documentation/perf-report.txt | 3 + tools/perf/builtin-record.c | 3 + tools/perf/builtin-report.c | 47 ++++- tools/perf/builtin-script.c | 12 -- tools/perf/tests/parse-events.c | 28 +++ tools/perf/ui/browsers/hists.c | 219 +++++++++++++++++---- tools/perf/ui/gtk/hists.c | 128 ++++++++++--- tools/perf/ui/hist.c | 316 ++++++++++++++++--------------- tools/perf/ui/stdio/hist.c | 2 + tools/perf/util/evlist.c | 7 +- tools/perf/util/evlist.h | 1 + tools/perf/util/evsel.c | 25 +++ tools/perf/util/evsel.h | 16 ++ tools/perf/util/header.c | 164 ++++++++++++++++ tools/perf/util/header.h | 2 + tools/perf/util/hist.c | 59 +++++- tools/perf/util/parse-events.c | 1 + tools/perf/util/parse-events.h | 1 + tools/perf/util/parse-events.y | 10 + tools/perf/util/string.c | 18 ++ tools/perf/util/symbol.h | 3 +- tools/perf/util/util.h | 1 + 22 files changed, 840 insertions(+), 226 deletions(-)
Signature · 1 line
1.7.11.7
Namhyung Kim
maintainerPATCH 1/13 perf tools: Keep group information
From: Namhyung Kim <[email protected]>
Add a few of group-related field in struct perf_{evlist,evsel} so that the group information in a evlist can be known easily. It only counts groups which have more than 1 members since leader-only groups are treated as non-group events.
--- tools/perf/util/evlist.c | 7 ++++++- tools/perf/util/evlist.h | 1 + tools/perf/util/evsel.h | 6 ++++++ tools/perf/util/parse-events.c | 1 + tools/perf/util/parse-events.h | 1 + tools/perf/util/parse-events.y | 10 ++++++++++ 6 files changed, 25 insertions(+), 1 deletion(-)
tools/perf/util/evlist.c +6 -1
@@ -117,6 +117,9 @@ void __perf_evlist__set_leader(struct list_head *list)struct perf_evsel *evsel, *leader; leader = list_entry(list->next, struct perf_evsel, node); + evsel = list_entry(list->prev, struct perf_evsel, node); + + leader->nr_members = evsel->idx - leader->idx +1 ;list_for_each_entry (evsel, list, node) {if (evsel != leader)
@@ -126,8 +129,10 @@ void __perf_evlist__set_leader(struct list_head *list)void perf_evlist__set_leader (struct perf_evlist *evlist) { -if (evlist->nr_entries) +if (evlist->nr_entries) { + evlist->nr_groups = evlist->nr_entries >1 ?1 :0 ;__perf_evlist__set_leader (&evlist->entries); + } }int perf_evlist__add_default (struct perf_evlist *evlist)
tools/perf/util/evlist.h +1 -0
@@ -21,6 +21,7 @@ struct perf_evlist {struct list_head entries;struct hlist_head heads[PERF_EVLIST__HLIST_SIZE ];int nr_entries; +int nr_groups;int nr_fds;int nr_mmaps;int mmap_len;
tools/perf/util/evsel.h +6 -0
@@ -73,6 +73,7 @@ struct perf_evsel {bool needs_swap;/* parse modifier helper */ int exclude_GH; +int nr_members;struct perf_evsel *leader;char *group_name; };
@@ -257,4 +258,9 @@ bool perf_evsel__fallback(struct perf_evsel *evsel, int err,int perf_evsel__open_strerror (struct perf_evsel *evsel,struct perf_target *target,int err,char *msg,size_t size); + +static inline int perf_evsel__group_idx (struct perf_evsel *evsel) +{ +return evsel->idx - evsel->leader->idx; +}#endif /* __PERF_EVSEL_H */
tools/perf/util/parse-events.c +1 -0
@@ -885,6 +885,7 @@ int parse_events(struct perf_evlist *evlist, const char *str,if (!ret) {int entries = data.idx - evlist->nr_entries;perf_evlist__splice_list_tail (evlist, &data.list, entries); + evlist->nr_groups += data.nr_groups;return 0 ; }
tools/perf/util/parse-events.h +1 -0
@@ -65,6 +65,7 @@ struct parse_events__term {struct parse_events_data__events {struct list_head list;int idx; +int nr_groups; };struct parse_events_data__terms {
tools/perf/util/parse-events.y +10 -0
@@ -23,6 +23,14 @@ do { \ YYABORT; \ } while (0) +static inc_group_count(struct list_head *list, + struct parse_events_data__events *data) +{ + /* Count groups only have more than 1 members */ + if (!list_is_last(list->next, list)) + data->nr_groups++; +} + %} %token PE_START_EVENTS PE_START_TERMS
@@ -123,6 +131,7 @@ PE_NAME '{' events '}' { struct list_head *list = $3; + inc_group_count(list, _data); parse_events__set_leader($1, list); $$ = list; }
@@ -131,6 +140,7 @@ PE_NAME '{' events '}' { struct list_head *list = $2; + inc_group_count(list, _data); parse_events__set_leader(NULL, list); $$ = list; }
Signature · 1 line
1.7.11.7
Namhyung Kim
maintainerPATCH 2/13 perf tests: Add group test conditions
In reply to Namhyung KimFrom: Namhyung Kim <[email protected]>
As some new fields for handling groups added, check them to be sure to have valid values in test__group* cases.
--- tools/perf/tests/parse-events.c | 28 ++++++++++++++++++++++++++++ 1 file changed, 28 insertions(+)
tools/perf/tests/parse-events.c +28 -0
@@ -23,6 +23,7 @@ static int test__checkevent_tracepoint(struct perf_evlist *evlist)struct perf_evsel *evsel =perf_evlist__first (evlist);TEST_ASSERT_VAL ("wrong number of entries" ,1 == evlist->nr_entries); +TEST_ASSERT_VAL ("wrong number of groups" ,0 == evlist->nr_groups);TEST_ASSERT_VAL ("wrong type" ,PERF_TYPE_TRACEPOINT == evsel->attr.type);TEST_ASSERT_VAL ("wrong sample_type" ,PERF_TP_SAMPLE_TYPE == evsel->attr.sample_type);
@@ -35,6 +36,7 @@ static int test__checkevent_tracepoint_multi(struct perf_evlist *evlist)struct perf_evsel *evsel;TEST_ASSERT_VAL ("wrong number of entries" , evlist->nr_entries >1 ); +TEST_ASSERT_VAL ("wrong number of groups" ,0 == evlist->nr_groups);list_for_each_entry (evsel, &evlist->entries, node) { TEST_ASSERT_VAL("wrong type",
@@ -510,6 +512,7 @@ static int test__group1(struct perf_evlist *evlist)struct perf_evsel *evsel, *leader;TEST_ASSERT_VAL ("wrong number of entries" ,2 == evlist->nr_entries); +TEST_ASSERT_VAL ("wrong number of groups" ,1 == evlist->nr_groups);/* instructions:k */ evsel = leader =perf_evlist__first (evlist);
@@ -523,6 +526,8 @@ static int test__group1(struct perf_evlist *evlist)TEST_ASSERT_VAL ("wrong exclude host" , !evsel->attr.exclude_host);TEST_ASSERT_VAL ("wrong precise_ip" , !evsel->attr.precise_ip);TEST_ASSERT_VAL ("wrong leader" ,perf_evsel__is_group_leader (evsel)); +TEST_ASSERT_VAL ("wrong nr_members" , evsel->nr_members ==2 ); +TEST_ASSERT_VAL ("wrong group_idx" ,perf_evsel__group_idx (evsel) ==0 );/* cycles:upp */ evsel =perf_evsel__next (evsel);
@@ -537,6 +542,7 @@ static int test__group1(struct perf_evlist *evlist)TEST_ASSERT_VAL ("wrong exclude host" , !evsel->attr.exclude_host);TEST_ASSERT_VAL ("wrong precise_ip" , evsel->attr.precise_ip ==2 );TEST_ASSERT_VAL ("wrong leader" , evsel->leader == leader); +TEST_ASSERT_VAL ("wrong group_idx" ,perf_evsel__group_idx (evsel) ==1 );return 0 ; }
@@ -546,6 +552,7 @@ static int test__group2(struct perf_evlist *evlist)struct perf_evsel *evsel, *leader;TEST_ASSERT_VAL ("wrong number of entries" ,3 == evlist->nr_entries); +TEST_ASSERT_VAL ("wrong number of groups" ,1 == evlist->nr_groups);/* faults + :ku modifier */ evsel = leader =perf_evlist__first (evlist);
@@ -559,6 +566,8 @@ static int test__group2(struct perf_evlist *evlist)TEST_ASSERT_VAL ("wrong exclude host" , !evsel->attr.exclude_host);TEST_ASSERT_VAL ("wrong precise_ip" , !evsel->attr.precise_ip);TEST_ASSERT_VAL ("wrong leader" ,perf_evsel__is_group_leader (evsel)); +TEST_ASSERT_VAL ("wrong nr_members" , evsel->nr_members ==2 ); +TEST_ASSERT_VAL ("wrong group_idx" ,perf_evsel__group_idx (evsel) ==0 );/* cache-references + :u modifier */ evsel =perf_evsel__next (evsel);
@@ -572,6 +581,7 @@ static int test__group2(struct perf_evlist *evlist)TEST_ASSERT_VAL ("wrong exclude host" , !evsel->attr.exclude_host);TEST_ASSERT_VAL ("wrong precise_ip" , !evsel->attr.precise_ip);TEST_ASSERT_VAL ("wrong leader" , evsel->leader == leader); +TEST_ASSERT_VAL ("wrong group_idx" ,perf_evsel__group_idx (evsel) ==1 );/* cycles:k */ evsel =perf_evsel__next (evsel);
@@ -594,6 +604,7 @@ static int test__group3(struct perf_evlist *evlist __maybe_unused)struct perf_evsel *evsel, *leader;TEST_ASSERT_VAL ("wrong number of entries" ,5 == evlist->nr_entries); +TEST_ASSERT_VAL ("wrong number of groups" ,2 == evlist->nr_groups);/* group1 syscalls:sys_enter_open:H */ evsel = leader =perf_evlist__first (evlist);
@@ -610,6 +621,8 @@ static int test__group3(struct perf_evlist *evlist __maybe_unused)TEST_ASSERT_VAL ("wrong leader" ,perf_evsel__is_group_leader (evsel));TEST_ASSERT_VAL ("wrong group name" , !strcmp (leader->group_name,"group1" )); +TEST_ASSERT_VAL ("wrong nr_members" , evsel->nr_members ==2 ); +TEST_ASSERT_VAL ("wrong group_idx" ,perf_evsel__group_idx (evsel) ==0 );/* group1 cycles:kppp */ evsel =perf_evsel__next (evsel);
@@ -625,6 +638,7 @@ static int test__group3(struct perf_evlist *evlist __maybe_unused)TEST_ASSERT_VAL ("wrong precise_ip" , evsel->attr.precise_ip ==3 );TEST_ASSERT_VAL ("wrong leader" , evsel->leader == leader);TEST_ASSERT_VAL ("wrong group name" , !evsel->group_name); +TEST_ASSERT_VAL ("wrong group_idx" ,perf_evsel__group_idx (evsel) ==1 );/* group2 cycles + G modifier */ evsel = leader =perf_evsel__next (evsel);
@@ -640,6 +654,8 @@ static int test__group3(struct perf_evlist *evlist __maybe_unused)TEST_ASSERT_VAL ("wrong leader" ,perf_evsel__is_group_leader (evsel));TEST_ASSERT_VAL ("wrong group name" , !strcmp (leader->group_name,"group2" )); +TEST_ASSERT_VAL ("wrong nr_members" , evsel->nr_members ==2 ); +TEST_ASSERT_VAL ("wrong group_idx" ,perf_evsel__group_idx (evsel) ==0 );/* group2 1:3 + G modifier */ evsel =perf_evsel__next (evsel);
@@ -652,6 +668,7 @@ static int test__group3(struct perf_evlist *evlist __maybe_unused)TEST_ASSERT_VAL ("wrong exclude host" , evsel->attr.exclude_host);TEST_ASSERT_VAL ("wrong precise_ip" , !evsel->attr.precise_ip);TEST_ASSERT_VAL ("wrong leader" , evsel->leader == leader); +TEST_ASSERT_VAL ("wrong group_idx" ,perf_evsel__group_idx (evsel) ==1 );/* instructions:u */ evsel =perf_evsel__next (evsel);
@@ -674,6 +691,7 @@ static int test__group4(struct perf_evlist *evlist __maybe_unused)struct perf_evsel *evsel, *leader;TEST_ASSERT_VAL ("wrong number of entries" ,2 == evlist->nr_entries); +TEST_ASSERT_VAL ("wrong number of groups" ,1 == evlist->nr_groups);/* cycles:u + p */ evsel = leader =perf_evlist__first (evlist);
@@ -689,6 +707,8 @@ static int test__group4(struct perf_evlist *evlist __maybe_unused)TEST_ASSERT_VAL ("wrong precise_ip" , evsel->attr.precise_ip ==1 );TEST_ASSERT_VAL ("wrong group name" , !evsel->group_name);TEST_ASSERT_VAL ("wrong leader" ,perf_evsel__is_group_leader (evsel)); +TEST_ASSERT_VAL ("wrong nr_members" , evsel->nr_members ==2 ); +TEST_ASSERT_VAL ("wrong group_idx" ,perf_evsel__group_idx (evsel) ==0 );/* instructions:kp + p */ evsel =perf_evsel__next (evsel);
@@ -703,6 +723,7 @@ static int test__group4(struct perf_evlist *evlist __maybe_unused)TEST_ASSERT_VAL ("wrong exclude host" , !evsel->attr.exclude_host);TEST_ASSERT_VAL ("wrong precise_ip" , evsel->attr.precise_ip ==2 );TEST_ASSERT_VAL ("wrong leader" , evsel->leader == leader); +TEST_ASSERT_VAL ("wrong group_idx" ,perf_evsel__group_idx (evsel) ==1 );return 0 ; }
@@ -712,6 +733,7 @@ static int test__group5(struct perf_evlist *evlist __maybe_unused)struct perf_evsel *evsel, *leader;TEST_ASSERT_VAL ("wrong number of entries" ,5 == evlist->nr_entries); +TEST_ASSERT_VAL ("wrong number of groups" ,2 == evlist->nr_groups);/* cycles + G */ evsel = leader =perf_evlist__first (evlist);
@@ -726,6 +748,8 @@ static int test__group5(struct perf_evlist *evlist __maybe_unused)TEST_ASSERT_VAL ("wrong precise_ip" , !evsel->attr.precise_ip);TEST_ASSERT_VAL ("wrong group name" , !evsel->group_name);TEST_ASSERT_VAL ("wrong leader" ,perf_evsel__is_group_leader (evsel)); +TEST_ASSERT_VAL ("wrong nr_members" , evsel->nr_members ==2 ); +TEST_ASSERT_VAL ("wrong group_idx" ,perf_evsel__group_idx (evsel) ==0 );/* instructions + G */ evsel =perf_evsel__next (evsel);
@@ -739,6 +763,7 @@ static int test__group5(struct perf_evlist *evlist __maybe_unused)TEST_ASSERT_VAL ("wrong exclude host" , evsel->attr.exclude_host);TEST_ASSERT_VAL ("wrong precise_ip" , !evsel->attr.precise_ip);TEST_ASSERT_VAL ("wrong leader" , evsel->leader == leader); +TEST_ASSERT_VAL ("wrong group_idx" ,perf_evsel__group_idx (evsel) ==1 );/* cycles:G */ evsel = leader =perf_evsel__next (evsel);
@@ -753,6 +778,8 @@ static int test__group5(struct perf_evlist *evlist __maybe_unused)TEST_ASSERT_VAL ("wrong precise_ip" , !evsel->attr.precise_ip);TEST_ASSERT_VAL ("wrong group name" , !evsel->group_name);TEST_ASSERT_VAL ("wrong leader" ,perf_evsel__is_group_leader (evsel)); +TEST_ASSERT_VAL ("wrong nr_members" , evsel->nr_members ==2 ); +TEST_ASSERT_VAL ("wrong group_idx" ,perf_evsel__group_idx (evsel) ==0 );/* instructions:G */ evsel =perf_evsel__next (evsel);
@@ -766,6 +793,7 @@ static int test__group5(struct perf_evlist *evlist __maybe_unused)TEST_ASSERT_VAL ("wrong exclude host" , evsel->attr.exclude_host);TEST_ASSERT_VAL ("wrong precise_ip" , !evsel->attr.precise_ip);TEST_ASSERT_VAL ("wrong leader" , evsel->leader == leader); +TEST_ASSERT_VAL ("wrong group_idx" ,perf_evsel__group_idx (evsel) ==1 );/* cycles */ evsel =perf_evsel__next (evsel);
Signature · 1 line
1.7.11.7
Namhyung Kim
maintainerPATCH 3/13 perf header: Add HEADER_GROUP_DESC feature
In reply to Namhyung KimFrom: Namhyung Kim <[email protected]>
Save group relationship information so that it can be restored when perf report is running.
--- tools/perf/builtin-record.c | 3 + tools/perf/util/header.c | 164 ++++++++++++++++++++++++++++++++++++++++++++ tools/perf/util/header.h | 2 + 3 files changed, 169 insertions(+)
tools/perf/builtin-record.c +3 -0
@@ -486,6 +486,9 @@ static int __cmd_record(struct perf_record *rec, int argc, const char **argv) goto out_delete_session; } +if (!evsel_list->nr_groups) +perf_header__clear_feat (&session->header,HEADER_GROUP_DESC ); +/* * perf_session__delete(session) will be called at perf_record__exit() */
tools/perf/util/header.c +164 -0
188 lines not on this page · raw message
tools/perf/util/header.h +2 -0
@@ -29,6 +29,7 @@ enum {HEADER_NUMA_TOPOLOGY ,HEADER_BRANCH_STACK ,HEADER_PMU_MAPPINGS , +HEADER_GROUP_DESC ,HEADER_LAST_FEATURE , HEADER_FEAT_BITS = 256, };
@@ -79,6 +80,7 @@ struct perf_session_env {char *numa_nodes;int nr_pmu_mappings;char *pmu_mappings; +int nr_groups; };struct perf_header {
Signature · 1 line
1.7.11.7
Namhyung Kim
maintainerPATCH 4/13 perf report: Make another loop for linking group hists
In reply to Namhyung KimFrom: Namhyung Kim <[email protected]>
Now the event grouping viewing requires linking all member hists in a group to the leader's. Thus hists__output_resort should be called after linking all events in evlist.
Introduce symbol_conf.event_group flag to determine whether the feature is enabled or not.
--- tools/perf/builtin-report.c | 13 ++++++++++++- tools/perf/util/symbol.h | 3 ++- 2 files changed, 14 insertions(+), 2 deletions(-)
tools/perf/builtin-report.c +12 -1
@@ -416,8 +416,16 @@ static int __cmd_report(struct perf_report *rep) hists->symbol_filter_str = rep->symbol_filter_str;hists__collapse_resort (hists); -hists__output_resort (hists); nr_samples += hists->stats.nr_events[PERF_RECORD_SAMPLE ]; + +/* Non-group events are considered as leader */ +if (symbol_conf.event_group && + !perf_evsel__is_group_leader (pos)) { +struct hists *leader_hists = &pos->leader->hists; + +hists__match (leader_hists, hists); +hists__link (leader_hists, hists); + } }if (nr_samples ==0 ) {
@@ -425,6 +433,9 @@ static int __cmd_report(struct perf_report *rep) goto out_delete; } +list_for_each_entry (pos, &session->evlist->entries, node) +hists__output_resort (&pos->hists); +if (use_browser >0 ) {if (use_browser ==1 ) { perf_evlist__tui_browse_hists(session->evlist, help,
tools/perf/util/symbol.h +2 -1
@@ -96,7 +96,8 @@ struct symbol_conf { initialized, kptr_restrict, annotate_asm_raw, - annotate_src; + annotate_src, + event_group;const char *vmlinux_name, *kallsyms_name, *source_prefix,
Signature · 1 line
1.7.11.7
Namhyung Kim
maintainerPATCH 5/13 perf hists: Resort hist entries using group members for output
In reply to Namhyung KimFrom: Namhyung Kim <[email protected]>
When event group is enabled, sorting hist entries on periods for output should consider groups members' period also. To do that, build period table using link/pair information and compare the table.
--- tools/perf/util/evsel.h | 2 ++ tools/perf/util/hist.c | 59 ++++++++++++++++++++++++++++++++++++++++++++++++- 2 files changed, 60 insertions(+), 1 deletion(-)
tools/perf/util/evsel.h +2 -0
@@ -78,6 +78,8 @@ struct perf_evsel {char *group_name; }; +#define hists_to_evsel (h) container_of(h, struct perf_evsel, hists) +struct cpu_map ;struct thread_map ;struct perf_evlist ;
tools/perf/util/hist.c +58 -1
77 lines not on this page · raw message
Signature · 1 line
1.7.11.7
Re: [PATCH 05/13] perf hists: Resort hist entries using group members for output
Em Tue, Jan 15, 2013 at 04:31:49PM +0900, Namhyung Kim escreveu:
Quoted 59 lines
From: Namhyung Kim <[email protected]>
When event group is enabled, sorting hist entries on periods for output should consider groups members' period also. To do that, build period table using link/pair information and compare the table.
Signed-off-by Namhyung Kim <[email protected]>Acked-by Jiri Olsa <[email protected]>Cc Ingo Molnar <[email protected]>Cc Jiri Olsa <[email protected]>Cc Paul Mackerras <[email protected]>Cc Peter Zijlstra <[email protected]>Cc Stephane Eranian <[email protected]>Signed-off-by Arnaldo Carvalho de Melo <[email protected]>--- tools/perf/util/evsel.h | 2 ++ tools/perf/util/hist.c | 59 ++++++++++++++++++++++++++++++++++++++++++++++++- 2 files changed, 60 insertions(+), 1 deletion(-)tools/perf/util/evsel.h +2
-0@@ -78,6 +78,8 @@ struct perf_evsel {char *group_name; }; +#define hists_to_evsel (h) container_of(h, struct perf_evsel, hists) +struct cpu_map ;struct thread_map ;struct perf_evlist ;tools/perf/util/hist.c +11
-0@@ -4,6 +4,7 @@#include "hist.h" #include "session.h" #include "sort.h" +#include "evsel.h" #include <math.h> static bool hists__filter_entry_by_dso(struct hists *hists,@@ -540,6 +541,62 @@ void hists__collapse_resort_threaded(struct hists *hists)* reverse the map, sort on period. */ +static int period_cmp (u64 period_a,u64 period_b) +{ +if (period_a > period_b) +return 1 ; +if (period_a < period_b) +return -1 ; +return 0 ; +} + +static int hist_entry__sort_on_period(struct hist_entry *a,
tools/perf/util/hist.c void hists__collapse_resort_threaded(struct hists *hists) · line 563
Quoted 10 lines
tools/perf/util/hist.c +10
-0+ struct hist_entry *b) +{ +int ret; +int i, nr_members; +struct perf_evsel *evsel; +struct hist_entry *pair; +u64 *periods_a, *periods_b; + + ret =period_cmp (a->stat.period, b->stat.period); +if (ret || !symbol_conf.event_group)
Why do we need to call period_cmp() when symbol_conf.event_group is false?
I see, confusing neverthless :-\
tools/perf/util/hist.c void hists__collapse_resort_threaded(struct hists *hists) · line 564
Quoted 1 line
tools/perf/util/hist.c +1
-0+return ret;
I would expect two functions, one for sorting on period and another for not sorting on period, i.e. don't call a function called 'sort_on_period' when _not_ sorting on period.
- Arnaldo
Quoted 49 lines
tools/perf/util/hist.c +36
-1+ + evsel =hists_to_evsel (a->hists); + nr_members = evsel->nr_members; +if (nr_members <=1 ) +return ret; + + periods_a =zalloc (sizeof (periods_a) * nr_members); + periods_b =zalloc (sizeof (periods_b) * nr_members); + +if (!periods_a || !periods_b) + goto out; + +list_for_each_entry (pair, &a->pairs.head, pairs.node) { + evsel =hists_to_evsel (pair->hists); + periods_a[perf_evsel__group_idx (evsel)] = pair->stat.period; + } + +list_for_each_entry (pair, &b->pairs.head, pairs.node) { + evsel =hists_to_evsel (pair->hists); + periods_b[perf_evsel__group_idx (evsel)] = pair->stat.period; + } + +for (i =1 ; i < nr_members; i++) { + ret =period_cmp (periods_a[i], periods_b[i]); +if (ret) +break ; + } + +out: +free (periods_a); +free (periods_b); + +return ret; +} +static void __hists__insert_output_entry (struct rb_root *entries,struct hist_entry *he,u64 min_callchain_hits)@@ -556,7 +613,7 @@ static void __hists__insert_output_entry(struct rb_root *entries, parent = *p; iter = rb_entry(parent, struct hist_entry, rb_node); -if (he->stat.period > iter->stat.period) +if (hist_entry__sort_on_period (he, iter) >0 ) p = &(*p)->rb_left;else p = &(*p)->rb_right;Signature · 1 line
1.7.11.7
Namhyung Kim
maintainerRe: [PATCH 05/13] perf hists: Resort hist entries using group members for output
Hi Arnaldo,
Quoted 6 lines · Arnaldo Carvalho de Melo
Em Tue, Jan 15, 2013 at 04:31:49PM +0900, Namhyung Kim escreveu:
Quoted 5 lines
From: Namhyung Kim <[email protected]>
When event group is enabled, sorting hist entries on periods for output should consider groups members' period also. To do that, build period table using link/pair information and compare the table.
[snip]
Quoted 15 lines
tools/perf/util/hist.c void hists__collapse_resort_threaded(struct hists *hists) · line 563
Quoted 11 lines
tools/perf/util/hist.c +11
-0+static int hist_entry__sort_on_period (struct hist_entry *a, +struct hist_entry *b) +{ +int ret; +int i, nr_members; +struct perf_evsel *evsel; +struct hist_entry *pair; +u64 *periods_a, *periods_b; + + ret =period_cmp (a->stat.period, b->stat.period); +if (ret || !symbol_conf.event_group)Why do we need to call
period_cmp()when symbol_conf.event_group is false?
Quoted 1 line
I see, confusing neverthless :-\
In this function @a and @b are entries of group leaders and should be compared first - other group members are compared only if periods of leader entries are same.
For non-group events or when event group view is disabled, this function will be called for every events as if they're leaders.
Quoted 6 lines
Review
Quoted 2 lines
(quoted diff) +1
-0+ return ret;I would expect two functions, one for sorting on period and another for not sorting on period, i.e. don't call a function called '
sort_on_period' when _not_ sorting on period.
Sorry, I can't understand.
This 'sort_on_period' function is called during the (final) output resorting and it's unconditional. IOW hist entries should be sorted by their period to be shown accordingly in the output.
Am I missing something?
Thanks,
Namhyung
Namhyung Kim
maintainerPATCH 6/13 perf ui/hist: Add support for event group view
In reply to Namhyung KimFrom: Namhyung Kim <[email protected]>
Show group members' overhead also when showing the leader's if event group is enabled. Use macro for defining hpp functions which looks almost identical.
Thanks to Arnaldo for suggesting a better way to print group members without allocating temporary array.
--- tools/perf/ui/hist.c | 311 +++++++++++++++++++++++---------------------- tools/perf/ui/stdio/hist.c | 2 + 2 files changed, 164 insertions(+), 149 deletions(-)
tools/perf/ui/hist.c +162 -149
@@ -3,151 +3,190 @@#include "../util/hist.h" #include "../util/util.h" #include "../util/sort.h" - +#include "../util/evsel.h" /* hist period print (hpp) functions */ -static int hpp__header_overhead (struct perf_hpp *hpp) -{ -return scnprintf (hpp->buf, hpp->size,"Overhead" ); -} -static int hpp__width_overhead (struct perf_hpp *hpp__maybe_unused ) -{ -return 8 ; -} +typedef int (*hpp_snprint_fn )(char *buf,size_t size,const char *fmt, ...); -static int hpp__color_overhead (struct perf_hpp *hpp,struct hist_entry *he) +static int __hpp__percent_fmt (struct perf_hpp *hpp,struct hist_entry *he, +u64 (*get_field)(struct hist_entry *), +const char *fmt,hpp_snprint_fn print_fn) { +int ret; +double percent =0.0 ;struct hists *hists = he->hists; -double percent =100.0 * he->stat.period / hists->stats.total_period; -return percent_color_snprintf (hpp->buf, hpp->size," %6.2f%%" , percent); -} +if (hists->stats.total_period) + percent =100.0 *get_field (he) / hists->stats.total_period; -static int hpp__entry_overhead (struct perf_hpp *hpp,struct hist_entry *he) -{ -struct hists *hists = he->hists; -double percent =100.0 * he->stat.period / hists->stats.total_period; -const char *fmt = symbol_conf.field_sep ?"%.2f" :" %6.2f%%" ; + ret =print_fn (hpp->buf, hpp->size, fmt, percent); -return scnprintf (hpp->buf, hpp->size, fmt, percent); -} +if (symbol_conf.event_group) { +int prev_idx, idx_delta, i; +struct perf_evsel *evsel =hists_to_evsel (hists); +struct hist_entry *pair; +int nr_members = evsel->nr_members; -static int hpp__header_overhead_sys (struct perf_hpp *hpp) -{ -const char *fmt = symbol_conf.field_sep ?"%s" :"%7s" ; +if (nr_members <=1 ) +return ret; -return scnprintf (hpp->buf, hpp->size, fmt,"sys" ); -} + prev_idx =perf_evsel__group_idx (evsel); -static int hpp__width_overhead_sys (struct perf_hpp *hpp__maybe_unused ) -{ -return 7 ; -} +list_for_each_entry (pair, &he->pairs.head, pairs.node) { +u64 period =get_field (pair); +u64 total = pair->hists->stats.total_period; -static int hpp__color_overhead_sys (struct perf_hpp *hpp,struct hist_entry *he) -{ -struct hists *hists = he->hists; -double percent =100.0 * he->stat.period_sys / hists->stats.total_period; +if (!total) +continue ; -return percent_color_snprintf (hpp->buf, hpp->size,"%6.2f%%" , percent); -} + evsel =hists_to_evsel (pair->hists); + idx_delta =perf_evsel__group_idx (evsel) - prev_idx -1 ; -static int hpp__entry_overhead_sys (struct perf_hpp *hpp,struct hist_entry *he) -{ -struct hists *hists = he->hists; -double percent =100.0 * he->stat.period_sys / hists->stats.total_period; -const char *fmt = symbol_conf.field_sep ?"%.2f" :"%6.2f%%" ; +for (i =0 ; i < idx_delta; i++) { + ret +=print_fn (hpp->buf + ret, hpp->size - ret, + fmt,0.0 ); + } -return scnprintf (hpp->buf, hpp->size, fmt, percent); -} + ret +=print_fn (hpp->buf + ret, hpp->size - ret, + fmt,100.0 * period / total); -static int hpp__header_overhead_us (struct perf_hpp *hpp) -{ -const char *fmt = symbol_conf.field_sep ?"%s" :"%7s" ; + prev_idx +=1 + idx_delta; + } -return scnprintf (hpp->buf, hpp->size, fmt,"user" ); -} + idx_delta = nr_members - prev_idx -1 ; -static int hpp__width_overhead_us (struct perf_hpp *hpp__maybe_unused ) -{ -return 7 ; +for (i =0 ; i < idx_delta; i++) { + ret +=print_fn (hpp->buf + ret, hpp->size - ret, + fmt,0.0 ); + } + } +return ret; } -static int hpp__color_overhead_us (struct perf_hpp *hpp,struct hist_entry *he) +static int __hpp__raw_fmt (struct perf_hpp *hpp,struct hist_entry *he, +u64 (*get_field)(struct hist_entry *), +const char *fmt,hpp_snprint_fn print_fn) { -struct hists *hists = he->hists; -double percent =100.0 * he->stat.period_us / hists->stats.total_period; - -return percent_color_snprintf (hpp->buf, hpp->size,"%6.2f%%" , percent); -} +int ret; + ret =print_fn (hpp->buf, hpp->size, fmt,get_field (he)); -static int hpp__entry_overhead_us (struct perf_hpp *hpp,struct hist_entry *he) -{ -struct hists *hists = he->hists; -double percent =100.0 * he->stat.period_us / hists->stats.total_period; -const char *fmt = symbol_conf.field_sep ?"%.2f" :"%6.2f%%" ; +if (symbol_conf.event_group) { +int prev_idx, idx_delta, i; +struct perf_evsel *evsel =hists_to_evsel (he->hists); +struct hist_entry *pair; +int nr_members = evsel->nr_members; -return scnprintf (hpp->buf, hpp->size, fmt, percent); -} +if (nr_members <=1 ) +return ret; -static int hpp__header_overhead_guest_sys (struct perf_hpp *hpp) -{ -return scnprintf (hpp->buf, hpp->size,"guest sys" ); -} + prev_idx =perf_evsel__group_idx (evsel); -static int hpp__width_overhead_guest_sys (struct perf_hpp *hpp__maybe_unused ) -{ -return 9 ; -} +list_for_each_entry (pair, &he->pairs.head, pairs.node) { + evsel =hists_to_evsel (pair->hists); + idx_delta =perf_evsel__group_idx (evsel) - prev_idx -1 ; -static int hpp__color_overhead_guest_sys (struct perf_hpp *hpp, -struct hist_entry *he) -{ -struct hists *hists = he->hists; -double percent =100.0 * he->stat.period_guest_sys / hists->stats.total_period; +for (i =0 ; i < idx_delta; i++) { + ret +=print_fn (hpp->buf + ret, hpp->size - ret, + fmt,0.0 ); + } -return percent_color_snprintf (hpp->buf, hpp->size," %6.2f%% " , percent); -} + ret +=print_fn (hpp->buf + ret, hpp->size - ret, + fmt,get_field (pair)); -static int hpp__entry_overhead_guest_sys (struct perf_hpp *hpp, -struct hist_entry *he) -{ -struct hists *hists = he->hists; -double percent =100.0 * he->stat.period_guest_sys / hists->stats.total_period; -const char *fmt = symbol_conf.field_sep ?"%.2f" :" %6.2f%% " ; - -return scnprintf (hpp->buf, hpp->size, fmt, percent); -} + prev_idx +=1 + idx_delta; + } -static int hpp__header_overhead_guest_us (struct perf_hpp *hpp) -{ -return scnprintf (hpp->buf, hpp->size,"guest usr" ); -} + idx_delta = nr_members - prev_idx -1 ; -static int hpp__width_overhead_guest_us (struct perf_hpp *hpp__maybe_unused ) -{ -return 9 ; +for (i =0 ; i < idx_delta; i++) { + ret +=print_fn (hpp->buf + ret, hpp->size - ret, + fmt,0.0 ); + } + } +return ret; } -static int hpp__color_overhead_guest_us (struct perf_hpp *hpp, -struct hist_entry *he) -{ -struct hists *hists = he->hists; -double percent =100.0 * he->stat.period_guest_us / hists->stats.total_period; - -return percent_color_snprintf (hpp->buf, hpp->size," %6.2f%% " , percent); -} -static int hpp__entry_overhead_guest_us (struct perf_hpp *hpp, -struct hist_entry *he) -{ -struct hists *hists = he->hists; -double percent =100.0 * he->stat.period_guest_us / hists->stats.total_period; -const char *fmt = symbol_conf.field_sep ?"%.2f" :" %6.2f%% " ; +#define __HPP_HEADER_FN (_type, _str, _min_width, _unit_width) \ +static int hpp__header_##_type(struct perf_hpp *hpp) \ +{ \ + int len = _min_width; \ + \ + if (symbol_conf.event_group) { \ + struct perf_evsel *evsel = hpp->ptr; \ + \ + len = max(len, evsel->nr_members * _unit_width); \ + } \ + return scnprintf(hpp->buf, hpp->size, "%*s", len, _str); \ +} + +#define __HPP_WIDTH_FN (_type, _min_width, _unit_width) \ +static int hpp__width_##_type(struct perf_hpp *hpp) \ +{ \ + int len = _min_width; \ + \ + if (symbol_conf.event_group) { \ + struct perf_evsel *evsel = hpp->ptr; \ + \ + len = max(len, evsel->nr_members * _unit_width); \ + } \ + return len; \ +} + +#define __HPP_COLOR_PERCENT_FN (_type, _field) \ +static u64 he_get_##_field(struct hist_entry *he) \ +{ \ + return he->stat._field; \ +} \ + \ +static int hpp__color_##_type(struct perf_hpp *hpp, struct hist_entry *he) \ +{ \ + return __hpp__percent_fmt(hpp, he, he_get_##_field, " %6.2f%%", \ + (hpp_snprint_fn)percent_color_snprintf); \ +} + +#define __HPP_ENTRY_PERCENT_FN (_type, _field) \ +static int hpp__entry_##_type(struct perf_hpp *hpp, struct hist_entry *he) \ +{ \ + const char *fmt = symbol_conf.field_sep ? " %.2f" : " %6.2f%%"; \ + return __hpp__percent_fmt(hpp, he, he_get_##_field, fmt, \ + scnprintf); \ +} + +#define __HPP_ENTRY_RAW_FN (_type, _field) \ +static u64 he_get_raw_##_field(struct hist_entry *he) \ +{ \ + return he->stat._field; \ +} \ + \ +static int hpp__entry_##_type(struct perf_hpp *hpp, struct hist_entry *he) \ +{ \ + const char *fmt = symbol_conf.field_sep ? " %"PRIu64 : " %11"PRIu64; \ + return __hpp__raw_fmt(hpp, he, he_get_raw_##_field, fmt, scnprintf); \ +} + +#define HPP_PERCENT_FNS (_type, _str, _field, _min_width, _unit_width) \ +__HPP_HEADER_FN(_type, _str, _min_width, _unit_width) \ +__HPP_WIDTH_FN(_type, _min_width, _unit_width) \ +__HPP_COLOR_PERCENT_FN(_type, _field) \ +__HPP_ENTRY_PERCENT_FN(_type, _field) + +#define HPP_RAW_FNS (_type, _str, _field, _min_width, _unit_width) \ +__HPP_HEADER_FN(_type, _str, _min_width, _unit_width) \ +__HPP_WIDTH_FN(_type, _min_width, _unit_width) \ +__HPP_ENTRY_RAW_FN(_type, _field) + + +HPP_PERCENT_FNS(overhead, "Overhead", period, 8, 8) +HPP_PERCENT_FNS(overhead_sys, "sys", period_sys, 8, 8) +HPP_PERCENT_FNS(overhead_us, "usr", period_us, 8, 8) +HPP_PERCENT_FNS(overhead_guest_sys, "guest sys", period_guest_sys, 9, 8) +HPP_PERCENT_FNS(overhead_guest_us, "guest usr", period_guest_us, 9, 8) + +HPP_RAW_FNS(samples, "Samples", nr_events, 12, 12) +HPP_RAW_FNS(period, "Period", period, 12, 12) -return scnprintf (hpp->buf, hpp->size, fmt, percent); -}static int hpp__header_baseline (struct perf_hpp *hpp) {
@@ -179,7 +218,7 @@ static int hpp__color_baseline(struct perf_hpp *hpp, struct hist_entry *he) {double percent =baseline_percent (he); -if (hist_entry__has_pairs (he)) +if (hist_entry__has_pairs (he) || symbol_conf.field_sep)return percent_color_snprintf (hpp->buf, hpp->size," %6.2f%%" , percent);else return scnprintf (hpp->buf, hpp->size," " );
@@ -196,44 +235,6 @@ static int hpp__entry_baseline(struct perf_hpp *hpp, struct hist_entry *he)return scnprintf (hpp->buf, hpp->size," " ); } -static int hpp__header_samples (struct perf_hpp *hpp) -{ -const char *fmt = symbol_conf.field_sep ?"%s" :"%11s" ; - -return scnprintf (hpp->buf, hpp->size, fmt,"Samples" ); -} - -static int hpp__width_samples (struct perf_hpp *hpp__maybe_unused ) -{ -return 11 ; -} - -static int hpp__entry_samples (struct perf_hpp *hpp,struct hist_entry *he) -{ -const char *fmt = symbol_conf.field_sep ?"%" PRIu64 :"%11" PRIu64; - -return scnprintf (hpp->buf, hpp->size, fmt, he->stat.nr_events); -} - -static int hpp__header_period (struct perf_hpp *hpp) -{ -const char *fmt = symbol_conf.field_sep ?"%s" :"%12s" ; - -return scnprintf (hpp->buf, hpp->size, fmt,"Period" ); -} - -static int hpp__width_period (struct perf_hpp *hpp__maybe_unused ) -{ -return 12 ; -} - -static int hpp__entry_period (struct perf_hpp *hpp,struct hist_entry *he) -{ -const char *fmt = symbol_conf.field_sep ?"%" PRIu64 :"%12" PRIu64; - -return scnprintf (hpp->buf, hpp->size, fmt, he->stat.period); -} -static int hpp__header_period_baseline (struct perf_hpp *hpp) {const char *fmt = symbol_conf.field_sep ?"%s" :"%12s" ;
@@ -254,6 +255,7 @@ static int hpp__entry_period_baseline(struct perf_hpp *hpp, struct hist_entry *hreturn scnprintf (hpp->buf, hpp->size, fmt, period); } +static int hpp__header_delta (struct perf_hpp *hpp) {const char *fmt = symbol_conf.field_sep ?"%s" :"%7s" ;
@@ -408,9 +410,20 @@ struct perf_hpp_fmt perf_hpp__format[] = {LIST_HEAD (perf_hpp__list); +#undef HPP__COLOR_PRINT_FNS#undef HPP__PRINT_FNS +#undef HPP_PERCENT_FNS +#undef HPP_RAW_FNS + +#undef __HPP_HEADER_FN +#undef __HPP_WIDTH_FN +#undef __HPP_COLOR_PERCENT_FN +#undef __HPP_ENTRY_PERCENT_FN +#undef __HPP_ENTRY_RAW_FN + +void perf_hpp__init (void ) {if (symbol_conf.show_cpu_utilization) {
tools/perf/ui/stdio/hist.c +2 -0
@@ -3,6 +3,7 @@#include "../../util/util.h" #include "../../util/hist.h" #include "../../util/sort.h" +#include "../../util/evsel.h" static size_t callchain__fprintf_left_margin (FILE *fp,int left_margin)
@@ -347,6 +348,7 @@ size_t hists__fprintf(struct hists *hists, bool show_header, int max_rows,struct perf_hpp dummy_hpp = { .buf = bf, .size =sizeof (bf), + .ptr =hists_to_evsel (hists), };bool first = true;
Signature · 1 line
1.7.11.7
Re: [PATCH 06/13] perf ui/hist: Add support for event group view
Em Tue, Jan 15, 2013 at 04:31:50PM +0900, Namhyung Kim escreveu:
Quoted 8 lines
From: Namhyung Kim <[email protected]>
Show group members' overhead also when showing the leader's if event group is enabled. Use macro for defining hpp functions which looks almost identical.
Thanks to Arnaldo for suggesting a better way to print group members without allocating temporary array.
Thanks! Its a pity the patch ended up mixing up things, I'll have to apply it and look at the end result.
Quoted 418 lines
Cc Jiri Olsa <[email protected]>Cc Stephane Eranian <[email protected]>Signed-off-by Namhyung Kim <[email protected]>--- tools/perf/ui/hist.c | 311 +++++++++++++++++++++++---------------------- tools/perf/ui/stdio/hist.c | 2 + 2 files changed, 164 insertions(+), 149 deletions(-)tools/perf/ui/hist.c +162
-149@@ -3,151 +3,190 @@#include "../util/hist.h" #include "../util/util.h" #include "../util/sort.h" - +#include "../util/evsel.h" /* hist period print (hpp) functions */ -static int hpp__header_overhead (struct perf_hpp *hpp) -{ -return scnprintf (hpp->buf, hpp->size,"Overhead" ); -} -static int hpp__width_overhead (struct perf_hpp *hpp__maybe_unused ) -{ -return 8 ; -} +typedef int (*hpp_snprint_fn )(char *buf,size_t size,const char *fmt, ...); -static int hpp__color_overhead (struct perf_hpp *hpp,struct hist_entry *he) +static int __hpp__percent_fmt (struct perf_hpp *hpp,struct hist_entry *he, +u64 (*get_field)(struct hist_entry *), +const char *fmt,hpp_snprint_fn print_fn) { +int ret; +double percent =0.0 ;struct hists *hists = he->hists; -double percent =100.0 * he->stat.period / hists->stats.total_period; -return percent_color_snprintf (hpp->buf, hpp->size," %6.2f%%" , percent); -} +if (hists->stats.total_period) + percent =100.0 *get_field (he) / hists->stats.total_period; -static int hpp__entry_overhead (struct perf_hpp *hpp,struct hist_entry *he) -{ -struct hists *hists = he->hists; -double percent =100.0 * he->stat.period / hists->stats.total_period; -const char *fmt = symbol_conf.field_sep ?"%.2f" :" %6.2f%%" ; + ret =print_fn (hpp->buf, hpp->size, fmt, percent); -return scnprintf (hpp->buf, hpp->size, fmt, percent); -} +if (symbol_conf.event_group) { +int prev_idx, idx_delta, i; +struct perf_evsel *evsel =hists_to_evsel (hists); +struct hist_entry *pair; +int nr_members = evsel->nr_members; -static int hpp__header_overhead_sys (struct perf_hpp *hpp) -{ -const char *fmt = symbol_conf.field_sep ?"%s" :"%7s" ; +if (nr_members <=1 ) +return ret; -return scnprintf (hpp->buf, hpp->size, fmt,"sys" ); -} + prev_idx =perf_evsel__group_idx (evsel); -static int hpp__width_overhead_sys (struct perf_hpp *hpp__maybe_unused ) -{ -return 7 ; -} +list_for_each_entry (pair, &he->pairs.head, pairs.node) { +u64 period =get_field (pair); +u64 total = pair->hists->stats.total_period; -static int hpp__color_overhead_sys (struct perf_hpp *hpp,struct hist_entry *he) -{ -struct hists *hists = he->hists; -double percent =100.0 * he->stat.period_sys / hists->stats.total_period; +if (!total) +continue ; -return percent_color_snprintf (hpp->buf, hpp->size,"%6.2f%%" , percent); -} + evsel =hists_to_evsel (pair->hists); + idx_delta =perf_evsel__group_idx (evsel) - prev_idx -1 ; -static int hpp__entry_overhead_sys (struct perf_hpp *hpp,struct hist_entry *he) -{ -struct hists *hists = he->hists; -double percent =100.0 * he->stat.period_sys / hists->stats.total_period; -const char *fmt = symbol_conf.field_sep ?"%.2f" :"%6.2f%%" ; +for (i =0 ; i < idx_delta; i++) { + ret +=print_fn (hpp->buf + ret, hpp->size - ret, + fmt,0.0 ); + } -return scnprintf (hpp->buf, hpp->size, fmt, percent); -} + ret +=print_fn (hpp->buf + ret, hpp->size - ret, + fmt,100.0 * period / total); -static int hpp__header_overhead_us (struct perf_hpp *hpp) -{ -const char *fmt = symbol_conf.field_sep ?"%s" :"%7s" ; + prev_idx +=1 + idx_delta; + } -return scnprintf (hpp->buf, hpp->size, fmt,"user" ); -} + idx_delta = nr_members - prev_idx -1 ; -static int hpp__width_overhead_us (struct perf_hpp *hpp__maybe_unused ) -{ -return 7 ; +for (i =0 ; i < idx_delta; i++) { + ret +=print_fn (hpp->buf + ret, hpp->size - ret, + fmt,0.0 ); + } + } +return ret; } -static int hpp__color_overhead_us (struct perf_hpp *hpp,struct hist_entry *he) +static int __hpp__raw_fmt (struct perf_hpp *hpp,struct hist_entry *he, +u64 (*get_field)(struct hist_entry *), +const char *fmt,hpp_snprint_fn print_fn) { -struct hists *hists = he->hists; -double percent =100.0 * he->stat.period_us / hists->stats.total_period; - -return percent_color_snprintf (hpp->buf, hpp->size,"%6.2f%%" , percent); -} +int ret; + ret =print_fn (hpp->buf, hpp->size, fmt,get_field (he)); -static int hpp__entry_overhead_us (struct perf_hpp *hpp,struct hist_entry *he) -{ -struct hists *hists = he->hists; -double percent =100.0 * he->stat.period_us / hists->stats.total_period; -const char *fmt = symbol_conf.field_sep ?"%.2f" :"%6.2f%%" ; +if (symbol_conf.event_group) { +int prev_idx, idx_delta, i; +struct perf_evsel *evsel =hists_to_evsel (he->hists); +struct hist_entry *pair; +int nr_members = evsel->nr_members; -return scnprintf (hpp->buf, hpp->size, fmt, percent); -} +if (nr_members <=1 ) +return ret; -static int hpp__header_overhead_guest_sys (struct perf_hpp *hpp) -{ -return scnprintf (hpp->buf, hpp->size,"guest sys" ); -} + prev_idx =perf_evsel__group_idx (evsel); -static int hpp__width_overhead_guest_sys (struct perf_hpp *hpp__maybe_unused ) -{ -return 9 ; -} +list_for_each_entry (pair, &he->pairs.head, pairs.node) { + evsel =hists_to_evsel (pair->hists); + idx_delta =perf_evsel__group_idx (evsel) - prev_idx -1 ; -static int hpp__color_overhead_guest_sys (struct perf_hpp *hpp, -struct hist_entry *he) -{ -struct hists *hists = he->hists; -double percent =100.0 * he->stat.period_guest_sys / hists->stats.total_period; +for (i =0 ; i < idx_delta; i++) { + ret +=print_fn (hpp->buf + ret, hpp->size - ret, + fmt,0.0 ); + } -return percent_color_snprintf (hpp->buf, hpp->size," %6.2f%% " , percent); -} + ret +=print_fn (hpp->buf + ret, hpp->size - ret, + fmt,get_field (pair)); -static int hpp__entry_overhead_guest_sys (struct perf_hpp *hpp, -struct hist_entry *he) -{ -struct hists *hists = he->hists; -double percent =100.0 * he->stat.period_guest_sys / hists->stats.total_period; -const char *fmt = symbol_conf.field_sep ?"%.2f" :" %6.2f%% " ; - -return scnprintf (hpp->buf, hpp->size, fmt, percent); -} + prev_idx +=1 + idx_delta; + } -static int hpp__header_overhead_guest_us (struct perf_hpp *hpp) -{ -return scnprintf (hpp->buf, hpp->size,"guest usr" ); -} + idx_delta = nr_members - prev_idx -1 ; -static int hpp__width_overhead_guest_us (struct perf_hpp *hpp__maybe_unused ) -{ -return 9 ; +for (i =0 ; i < idx_delta; i++) { + ret +=print_fn (hpp->buf + ret, hpp->size - ret, + fmt,0.0 ); + } + } +return ret; } -static int hpp__color_overhead_guest_us (struct perf_hpp *hpp, -struct hist_entry *he) -{ -struct hists *hists = he->hists; -double percent =100.0 * he->stat.period_guest_us / hists->stats.total_period; - -return percent_color_snprintf (hpp->buf, hpp->size," %6.2f%% " , percent); -} -static int hpp__entry_overhead_guest_us (struct perf_hpp *hpp, -struct hist_entry *he) -{ -struct hists *hists = he->hists; -double percent =100.0 * he->stat.period_guest_us / hists->stats.total_period; -const char *fmt = symbol_conf.field_sep ?"%.2f" :" %6.2f%% " ; +#define __HPP_HEADER_FN (_type, _str, _min_width, _unit_width) \ +static int hpp__header_##_type(struct perf_hpp *hpp) \ +{ \ + int len = _min_width; \ + \ + if (symbol_conf.event_group) { \ + struct perf_evsel *evsel = hpp->ptr; \ + \ + len = max(len, evsel->nr_members * _unit_width); \ + } \ + return scnprintf(hpp->buf, hpp->size, "%*s", len, _str); \ +} + +#define __HPP_WIDTH_FN (_type, _min_width, _unit_width) \ +static int hpp__width_##_type(struct perf_hpp *hpp) \ +{ \ + int len = _min_width; \ + \ + if (symbol_conf.event_group) { \ + struct perf_evsel *evsel = hpp->ptr; \ + \ + len = max(len, evsel->nr_members * _unit_width); \ + } \ + return len; \ +} + +#define __HPP_COLOR_PERCENT_FN (_type, _field) \ +static u64 he_get_##_field(struct hist_entry *he) \ +{ \ + return he->stat._field; \ +} \ + \ +static int hpp__color_##_type(struct perf_hpp *hpp, struct hist_entry *he) \ +{ \ + return __hpp__percent_fmt(hpp, he, he_get_##_field, " %6.2f%%", \ + (hpp_snprint_fn)percent_color_snprintf); \ +} + +#define __HPP_ENTRY_PERCENT_FN (_type, _field) \ +static int hpp__entry_##_type(struct perf_hpp *hpp, struct hist_entry *he) \ +{ \ + const char *fmt = symbol_conf.field_sep ? " %.2f" : " %6.2f%%"; \ + return __hpp__percent_fmt(hpp, he, he_get_##_field, fmt, \ + scnprintf); \ +} + +#define __HPP_ENTRY_RAW_FN (_type, _field) \ +static u64 he_get_raw_##_field(struct hist_entry *he) \ +{ \ + return he->stat._field; \ +} \ + \ +static int hpp__entry_##_type(struct perf_hpp *hpp, struct hist_entry *he) \ +{ \ + const char *fmt = symbol_conf.field_sep ? " %"PRIu64 : " %11"PRIu64; \ + return __hpp__raw_fmt(hpp, he, he_get_raw_##_field, fmt, scnprintf); \ +} + +#define HPP_PERCENT_FNS (_type, _str, _field, _min_width, _unit_width) \ +__HPP_HEADER_FN(_type, _str, _min_width, _unit_width) \ +__HPP_WIDTH_FN(_type, _min_width, _unit_width) \ +__HPP_COLOR_PERCENT_FN(_type, _field) \ +__HPP_ENTRY_PERCENT_FN(_type, _field) + +#define HPP_RAW_FNS (_type, _str, _field, _min_width, _unit_width) \ +__HPP_HEADER_FN(_type, _str, _min_width, _unit_width) \ +__HPP_WIDTH_FN(_type, _min_width, _unit_width) \ +__HPP_ENTRY_RAW_FN(_type, _field) + + +HPP_PERCENT_FNS(overhead, "Overhead", period, 8, 8) +HPP_PERCENT_FNS(overhead_sys, "sys", period_sys, 8, 8) +HPP_PERCENT_FNS(overhead_us, "usr", period_us, 8, 8) +HPP_PERCENT_FNS(overhead_guest_sys, "guest sys", period_guest_sys, 9, 8) +HPP_PERCENT_FNS(overhead_guest_us, "guest usr", period_guest_us, 9, 8) + +HPP_RAW_FNS(samples, "Samples", nr_events, 12, 12) +HPP_RAW_FNS(period, "Period", period, 12, 12) -return scnprintf (hpp->buf, hpp->size, fmt, percent); -}static int hpp__header_baseline (struct perf_hpp *hpp) {@@ -179,7 +218,7 @@ static int hpp__color_baseline(struct perf_hpp *hpp, struct hist_entry *he) {double percent =baseline_percent (he); -if (hist_entry__has_pairs (he)) +if (hist_entry__has_pairs (he) || symbol_conf.field_sep)return percent_color_snprintf (hpp->buf, hpp->size," %6.2f%%" , percent);else return scnprintf (hpp->buf, hpp->size," " );@@ -196,44 +235,6 @@ static int hpp__entry_baseline(struct perf_hpp *hpp, struct hist_entry *he)return scnprintf (hpp->buf, hpp->size," " ); } -static int hpp__header_samples (struct perf_hpp *hpp) -{ -const char *fmt = symbol_conf.field_sep ?"%s" :"%11s" ; - -return scnprintf (hpp->buf, hpp->size, fmt,"Samples" ); -} - -static int hpp__width_samples (struct perf_hpp *hpp__maybe_unused ) -{ -return 11 ; -} - -static int hpp__entry_samples (struct perf_hpp *hpp,struct hist_entry *he) -{ -const char *fmt = symbol_conf.field_sep ?"%" PRIu64 :"%11" PRIu64; - -return scnprintf (hpp->buf, hpp->size, fmt, he->stat.nr_events); -} - -static int hpp__header_period (struct perf_hpp *hpp) -{ -const char *fmt = symbol_conf.field_sep ?"%s" :"%12s" ; - -return scnprintf (hpp->buf, hpp->size, fmt,"Period" ); -} - -static int hpp__width_period (struct perf_hpp *hpp__maybe_unused ) -{ -return 12 ; -} - -static int hpp__entry_period (struct perf_hpp *hpp,struct hist_entry *he) -{ -const char *fmt = symbol_conf.field_sep ?"%" PRIu64 :"%12" PRIu64; - -return scnprintf (hpp->buf, hpp->size, fmt, he->stat.period); -} -static int hpp__header_period_baseline (struct perf_hpp *hpp) {const char *fmt = symbol_conf.field_sep ?"%s" :"%12s" ;@@ -254,6 +255,7 @@ static int hpp__entry_period_baseline(struct perf_hpp *hpp, struct hist_entry *hreturn scnprintf (hpp->buf, hpp->size, fmt, period); } +static int hpp__header_delta (struct perf_hpp *hpp) {const char *fmt = symbol_conf.field_sep ?"%s" :"%7s" ;@@ -408,9 +410,20 @@ struct perf_hpp_fmt perf_hpp__format[] = {LIST_HEAD (perf_hpp__list); +#undef HPP__COLOR_PRINT_FNS#undef HPP__PRINT_FNS +#undef HPP_PERCENT_FNS +#undef HPP_RAW_FNS + +#undef __HPP_HEADER_FN +#undef __HPP_WIDTH_FN +#undef __HPP_COLOR_PERCENT_FN +#undef __HPP_ENTRY_PERCENT_FN +#undef __HPP_ENTRY_RAW_FN + +void perf_hpp__init (void ) {if (symbol_conf.show_cpu_utilization) {tools/perf/ui/stdio/hist.c +2
-0@@ -3,6 +3,7 @@#include "../../util/util.h" #include "../../util/hist.h" #include "../../util/sort.h" +#include "../../util/evsel.h" static size_t callchain__fprintf_left_margin (FILE *fp,int left_margin)@@ -347,6 +348,7 @@ size_t hists__fprintf(struct hists *hists, bool show_header, int max_rows,struct perf_hpp dummy_hpp = { .buf = bf, .size =sizeof (bf), + .ptr =hists_to_evsel (hists), };bool first = true;Signature · 1 line
1.7.11.7
Namhyung Kim
maintainerRe: [PATCH 06/13] perf ui/hist: Add support for event group view
Quoted 7 lines · Arnaldo Carvalho de Melo
Em Tue, Jan 15, 2013 at 04:31:50PM +0900, Namhyung Kim escreveu:
Quoted 6 lines
From: Namhyung Kim <[email protected]>
Show group members' overhead also when showing the leader's if event group is enabled. Use macro for defining hpp functions which looks almost identical.
Quoted 5 lines
Quoted 2 lines
Thanks to Arnaldo for suggesting a better way to print group members without allocating temporary array.
Thanks! Its a pity the patch ended up mixing up things, I'll have to apply it and look at the end result.
Do you mean I have to separate this into two - one for converting to macro and another for event group view itself?
Thanks,
Namhyung
Re: [PATCH 06/13] perf ui/hist: Add support for event group view
Em Wed, Jan 16, 2013 at 06:22:49PM +0900, Namhyung Kim escreveu:
Quoted 8 lines · Arnaldo Carvalho de Melo
Quoted 7 lines
Em Tue, Jan 15, 2013 at 04:31:50PM +0900, Namhyung Kim escreveu:
Quoted 6 lines
From: Namhyung Kim <[email protected]>
Show group members' overhead also when showing the leader's if event group is enabled. Use macro for defining hpp functions which looks almost identical.
Quoted 8 lines
Quoted 5 lines
Quoted 2 lines
Thanks to Arnaldo for suggesting a better way to print group members without allocating temporary array.
Thanks! Its a pity the patch ended up mixing up things, I'll have to apply it and look at the end result.
Do you mean I have to separate this into two - one for converting to macro and another for event group view itself?
If you can, please, that way we may see more clearly each of these changes in the separated patches.
Thanks,
- Arnaldo
Namhyung Kim
maintainerPATCH 7/13 perf hist browser: Add support for event group view
In reply to Namhyung KimFrom: Namhyung Kim <[email protected]>
Show group members' overhead also when showing the leader's if event group is enabled. Use macro for defining hpp functions which looks almost identical.
Also move coloring logic into the hpp function so that each value can be colored independently.
--- tools/perf/ui/browsers/hists.c | 153 ++++++++++++++++++++++++++++++++--------- tools/perf/ui/hist.c | 5 +- 2 files changed, 126 insertions(+), 32 deletions(-)
tools/perf/ui/browsers/hists.c +122 -31
@@ -567,23 +567,126 @@ static int hist_browser__show_callchain(struct hist_browser *browser,return row - first_row; } -#define HPP__COLOR_FN (_name, _field) \ -static int hist_browser__hpp_color_ ## _name(struct perf_hpp *hpp, \ - struct hist_entry *he) \ +struct hpp_arg { +struct ui_browser *b; +char folded_sign; +bool current_entry; +}; + +static int hist_browser__hpp_color_callchain (struct hpp_arg *arg) +{ +if (!symbol_conf.use_callchain) +return 0 ; + +slsmg_printf ("%c " , arg->folded_sign); +return 2 ; +} + +static int __hpp__color_fmt (struct perf_hpp *hpp,struct hist_entry *he, +u64 (*get_field)(struct hist_entry *), +int (*callchain_cb)(struct hpp_arg *)) +{ +int ret =0 ; +double percent =0.0 ; +struct hists *hists = he->hists; +struct hpp_arg *arg = hpp->ptr; + +if (hists->stats.total_period) + percent =100.0 *get_field (he) / hists->stats.total_period; + +ui_browser__set_percent_color (arg->b, percent, arg->current_entry); + +if (callchain_cb) + ret +=callchain_cb (arg); + + ret +=scnprintf (hpp->buf, hpp->size,"%6.2f%%" , percent); +slsmg_printf ("%s" , hpp->buf); + +if (symbol_conf.event_group) { +int prev_idx, idx_delta, i; +struct perf_evsel *evsel =hists_to_evsel (hists); +struct hist_entry *pair; +int nr_members = evsel->nr_members; + +if (nr_members <=1 ) + goto out; + + prev_idx =perf_evsel__group_idx (evsel); + +list_for_each_entry (pair, &he->pairs.head, pairs.node) { +u64 period =get_field (pair); +u64 total = pair->hists->stats.total_period; + +if (!total) +continue ; + + evsel =hists_to_evsel (pair->hists); + idx_delta =perf_evsel__group_idx (evsel) - prev_idx -1 ; + +for (i =0 ; i < idx_delta; i++) { +ui_browser__set_percent_color (arg->b,0.0 , + arg->current_entry); + ret +=scnprintf (hpp->buf, hpp->size, +" %6.2f%%" ,0.0 ); +slsmg_printf ("%s" , hpp->buf); + } + + percent =100.0 * period / total; +ui_browser__set_percent_color (arg->b, percent, + arg->current_entry); + ret +=scnprintf (hpp->buf, hpp->size, +" %6.2f%%" , percent); +slsmg_printf ("%s" , hpp->buf); + + prev_idx +=1 + idx_delta; + } + + idx_delta = nr_members - prev_idx -1 ; + +for (i =0 ; i < idx_delta; i++) { +ui_browser__set_percent_color (arg->b,0.0 , + arg->current_entry); + ret +=scnprintf (hpp->buf, hpp->size," %6.2f%%" ,0.0 ); +slsmg_printf ("%s" , hpp->buf); + } + } +out: +if (!arg->current_entry || !arg->b->navkeypressed) +ui_browser__set_color (arg->b,HE_COLORSET_NORMAL ); + +return ret; +} + +static u64 __hpp_get_period (struct hist_entry *he) +{ +return he->stat.period; +} + +static int hist_browser__hpp_color_overhead (struct perf_hpp *hpp, +struct hist_entry *he) +{ +return __hpp__color_fmt (hpp, he, __hpp_get_period, + hist_browser__hpp_color_callchain); +} + +#define __HPP_COLOR_PERCENT_FN (_type, _field) \ +static u64 __hpp_get_##_field(struct hist_entry *he) \ +{ \ + return he->stat._field; \ +} \ + \ +static int hist_browser__hpp_color_##_type(struct perf_hpp *hpp, \ + struct hist_entry *he) \ { \ - struct hists *hists = he->hists; \ - double percent = 100.0 * he->stat._field / hists->stats.total_period; \ - *(double *)hpp->ptr = percent; \ - return scnprintf(hpp->buf, hpp->size, "%6.2f%%", percent); \ + return __hpp__color_fmt(hpp, he, __hpp_get_##_field, NULL); \ } -HPP__COLOR_FN(overhead, period) -HPP__COLOR_FN(overhead_sys, period_sys) -HPP__COLOR_FN(overhead_us, period_us) -HPP__COLOR_FN(overhead_guest_sys, period_guest_sys) -HPP__COLOR_FN(overhead_guest_us, period_guest_us) +__HPP_COLOR_PERCENT_FN(overhead_sys, period_sys) +__HPP_COLOR_PERCENT_FN(overhead_us, period_us) +__HPP_COLOR_PERCENT_FN(overhead_guest_sys, period_guest_sys) +__HPP_COLOR_PERCENT_FN(overhead_guest_us, period_guest_us) -#undef HPP__COLOR_FN +#undef __HPP_COLOR_PERCENT_FNvoid hist_browser__init_hpp (void ) {
@@ -608,7 +711,6 @@ static int hist_browser__show_entry(struct hist_browser *browser, unsigned short row) {char s[256 ]; -double percent;int printed =0 ;int width = browser->b.width;char folded_sign =' ' ;
@@ -628,11 +730,16 @@ static int hist_browser__show_entry(struct hist_browser *browser, }if (row_offset ==0 ) { +struct hpp_arg arg = { + .b = &browser->b, + .folded_sign = folded_sign, + .current_entry = current_entry, + };struct perf_hpp hpp = { .buf = s, .size =sizeof (s), + .ptr = &arg, }; -int i =0 ;ui_browser__gotorc (&browser->b, row,0 );
@@ -645,27 +752,11 @@ static int hist_browser__show_entry(struct hist_browser *browser, first = false;if (fmt->color) { - hpp.ptr = &percent; -/* It will set percent for us. See HPP__COLOR_FN above. */ width -= fmt->color (&hpp, entry); - -ui_browser__set_percent_color (&browser->b, percent, current_entry); - -if (!i && symbol_conf.use_callchain) { -slsmg_printf ("%c " , folded_sign); - width -=2 ; - } - -slsmg_printf ("%s" , s); - -if (!current_entry || !browser->b.navkeypressed) -ui_browser__set_color (&browser->b,HE_COLORSET_NORMAL ); }else { width -= fmt->entry (&hpp, entry);slsmg_printf ("%s" , s); } - - i++; }/* The scroll bar isn't being used */
tools/perf/ui/hist.c +4 -1
@@ -521,12 +521,15 @@ unsigned int hists__sort_list_width(struct hists *hists)struct perf_hpp_fmt *fmt;struct sort_entry *se;int i =0 , ret =0 ; +struct perf_hpp dummy_hpp = { + .ptr =hists_to_evsel (hists), + };perf_hpp__for_each_format (fmt) {if (i) ret +=2 ; - ret += fmt->width (NULL ); + ret += fmt->width (&dummy_hpp); }list_for_each_entry (se, &hist_entry__sort_list, list)
Signature · 1 line
1.7.11.7
Re: [PATCH 07/13] perf hist browser: Add support for event group view
Em Tue, Jan 15, 2013 at 04:31:51PM +0900, Namhyung Kim escreveu:
Quoted 8 lines
From: Namhyung Kim <[email protected]>
Show group members' overhead also when showing the leader's if event group is enabled. Use macro for defining hpp functions which looks almost identical.
Also move coloring logic into the hpp function so that each value can be colored independently.
Just from reading the description: when we have this "Also do this extra thing" it smells like we need to break the patch in two... Can this be done?
- Arnaldo
Quoted 229 lines
Cc Jiri Olsa <[email protected]>Cc Stephane Eranian <[email protected]>Signed-off-by Namhyung Kim <[email protected]>--- tools/perf/ui/browsers/hists.c | 153 ++++++++++++++++++++++++++++++++--------- tools/perf/ui/hist.c | 5 +- 2 files changed, 126 insertions(+), 32 deletions(-)tools/perf/ui/browsers/hists.c +122
-31@@ -567,23 +567,126 @@ static int hist_browser__show_callchain(struct hist_browser *browser,return row - first_row; } -#define HPP__COLOR_FN (_name, _field) \ -static int hist_browser__hpp_color_ ## _name(struct perf_hpp *hpp, \ - struct hist_entry *he) \ +struct hpp_arg { +struct ui_browser *b; +char folded_sign; +bool current_entry; +}; + +static int hist_browser__hpp_color_callchain (struct hpp_arg *arg) +{ +if (!symbol_conf.use_callchain) +return 0 ; + +slsmg_printf ("%c " , arg->folded_sign); +return 2 ; +} + +static int __hpp__color_fmt (struct perf_hpp *hpp,struct hist_entry *he, +u64 (*get_field)(struct hist_entry *), +int (*callchain_cb)(struct hpp_arg *)) +{ +int ret =0 ; +double percent =0.0 ; +struct hists *hists = he->hists; +struct hpp_arg *arg = hpp->ptr; + +if (hists->stats.total_period) + percent =100.0 *get_field (he) / hists->stats.total_period; + +ui_browser__set_percent_color (arg->b, percent, arg->current_entry); + +if (callchain_cb) + ret +=callchain_cb (arg); + + ret +=scnprintf (hpp->buf, hpp->size,"%6.2f%%" , percent); +slsmg_printf ("%s" , hpp->buf); + +if (symbol_conf.event_group) { +int prev_idx, idx_delta, i; +struct perf_evsel *evsel =hists_to_evsel (hists); +struct hist_entry *pair; +int nr_members = evsel->nr_members; + +if (nr_members <=1 ) + goto out; + + prev_idx =perf_evsel__group_idx (evsel); + +list_for_each_entry (pair, &he->pairs.head, pairs.node) { +u64 period =get_field (pair); +u64 total = pair->hists->stats.total_period; + +if (!total) +continue ; + + evsel =hists_to_evsel (pair->hists); + idx_delta =perf_evsel__group_idx (evsel) - prev_idx -1 ; + +for (i =0 ; i < idx_delta; i++) { +ui_browser__set_percent_color (arg->b,0.0 , + arg->current_entry); + ret +=scnprintf (hpp->buf, hpp->size, +" %6.2f%%" ,0.0 ); +slsmg_printf ("%s" , hpp->buf); + } + + percent =100.0 * period / total; +ui_browser__set_percent_color (arg->b, percent, + arg->current_entry); + ret +=scnprintf (hpp->buf, hpp->size, +" %6.2f%%" , percent); +slsmg_printf ("%s" , hpp->buf); + + prev_idx +=1 + idx_delta; + } + + idx_delta = nr_members - prev_idx -1 ; + +for (i =0 ; i < idx_delta; i++) { +ui_browser__set_percent_color (arg->b,0.0 , + arg->current_entry); + ret +=scnprintf (hpp->buf, hpp->size," %6.2f%%" ,0.0 ); +slsmg_printf ("%s" , hpp->buf); + } + } +out: +if (!arg->current_entry || !arg->b->navkeypressed) +ui_browser__set_color (arg->b,HE_COLORSET_NORMAL ); + +return ret; +} + +static u64 __hpp_get_period (struct hist_entry *he) +{ +return he->stat.period; +} + +static int hist_browser__hpp_color_overhead (struct perf_hpp *hpp, +struct hist_entry *he) +{ +return __hpp__color_fmt (hpp, he, __hpp_get_period, + hist_browser__hpp_color_callchain); +} + +#define __HPP_COLOR_PERCENT_FN (_type, _field) \ +static u64 __hpp_get_##_field(struct hist_entry *he) \ +{ \ + return he->stat._field; \ +} \ + \ +static int hist_browser__hpp_color_##_type(struct perf_hpp *hpp, \ + struct hist_entry *he) \ { \ - struct hists *hists = he->hists; \ - double percent = 100.0 * he->stat._field / hists->stats.total_period; \ - *(double *)hpp->ptr = percent; \ - return scnprintf(hpp->buf, hpp->size, "%6.2f%%", percent); \ + return __hpp__color_fmt(hpp, he, __hpp_get_##_field, NULL); \ } -HPP__COLOR_FN(overhead, period) -HPP__COLOR_FN(overhead_sys, period_sys) -HPP__COLOR_FN(overhead_us, period_us) -HPP__COLOR_FN(overhead_guest_sys, period_guest_sys) -HPP__COLOR_FN(overhead_guest_us, period_guest_us) +__HPP_COLOR_PERCENT_FN(overhead_sys, period_sys) +__HPP_COLOR_PERCENT_FN(overhead_us, period_us) +__HPP_COLOR_PERCENT_FN(overhead_guest_sys, period_guest_sys) +__HPP_COLOR_PERCENT_FN(overhead_guest_us, period_guest_us) -#undef HPP__COLOR_FN +#undef __HPP_COLOR_PERCENT_FNvoid hist_browser__init_hpp (void ) {@@ -608,7 +711,6 @@ static int hist_browser__show_entry(struct hist_browser *browser, unsigned short row) {char s[256 ]; -double percent;int printed =0 ;int width = browser->b.width;char folded_sign =' ' ;@@ -628,11 +730,16 @@ static int hist_browser__show_entry(struct hist_browser *browser, }if (row_offset ==0 ) { +struct hpp_arg arg = { + .b = &browser->b, + .folded_sign = folded_sign, + .current_entry = current_entry, + };struct perf_hpp hpp = { .buf = s, .size =sizeof (s), + .ptr = &arg, }; -int i =0 ;ui_browser__gotorc (&browser->b, row,0 );@@ -645,27 +752,11 @@ static int hist_browser__show_entry(struct hist_browser *browser, first = false;if (fmt->color) { - hpp.ptr = &percent; -/* It will set percent for us. See HPP__COLOR_FN above. */ width -= fmt->color (&hpp, entry); - -ui_browser__set_percent_color (&browser->b, percent, current_entry); - -if (!i && symbol_conf.use_callchain) { -slsmg_printf ("%c " , folded_sign); - width -=2 ; - } - -slsmg_printf ("%s" , s); - -if (!current_entry || !browser->b.navkeypressed) -ui_browser__set_color (&browser->b,HE_COLORSET_NORMAL ); }else { width -= fmt->entry (&hpp, entry);slsmg_printf ("%s" , s); } - - i++; }/* The scroll bar isn't being used */ tools/perf/ui/hist.c +4
-1@@ -521,12 +521,15 @@ unsigned int hists__sort_list_width(struct hists *hists)struct perf_hpp_fmt *fmt;struct sort_entry *se;int i =0 , ret =0 ; +struct perf_hpp dummy_hpp = { + .ptr =hists_to_evsel (hists), + };perf_hpp__for_each_format (fmt) {if (i) ret +=2 ; - ret += fmt->width (NULL ); + ret += fmt->width (&dummy_hpp); }list_for_each_entry (se, &hist_entry__sort_list, list)Signature · 1 line
1.7.11.7
Namhyung Kim
maintainerRe: [PATCH 07/13] perf hist browser: Add support for event group view
Quoted 7 lines · Arnaldo Carvalho de Melo
Em Tue, Jan 15, 2013 at 04:31:51PM +0900, Namhyung Kim escreveu:
Quoted 6 lines
From: Namhyung Kim <[email protected]>
Show group members' overhead also when showing the leader's if event group is enabled. Use macro for defining hpp functions which looks almost identical.
Quoted 6 lines
Quoted 2 lines
Also move coloring logic into the hpp function so that each value can be colored independently.
Just from reading the description: when we have this "Also do this extra thing" it smells like we need to break the patch in two... Can this be done?
Sure. I'll break it.
Thanks,
Namhyung
Namhyung Kim
maintainerPATCH 8/13 perf gtk/browser: Add support for event group view
In reply to Namhyung KimFrom: Namhyung Kim <[email protected]>
Show group members' overhead also when showing the leader's if event group is enabled. Use macro for defining hpp functions which looks almost identical.
Unlike other hpp backend, GTK+ needs to print dummy 0.00% output since it's displayed with variable width fonts. So that simply skipping with %*s trick won't work well here.
--- tools/perf/ui/gtk/hists.c | 114 +++++++++++++++++++++++++++++++++++++--------- 1 file changed, 92 insertions(+), 22 deletions(-)
tools/perf/ui/gtk/hists.c +92 -22
@@ -8,32 +8,102 @@#define MAX_COLUMNS 32 -#define HPP__COLOR_FN (_name, _field) \ -static int perf_gtk__hpp_color_ ## _name(struct perf_hpp *hpp, \ - struct hist_entry *he) \ +static int perf_gtk__percent_color_snprintf (char *buf,size_t size, +double percent) +{ +int ret =0 ; +const char *markup; + + markup =perf_gtk__get_percent_color (percent); +if (markup) + ret +=scnprintf (buf, size, markup); + + ret +=scnprintf (buf + ret, size - ret,"%6.2f%%" , percent); + +if (markup) + ret +=scnprintf (buf + ret, size - ret,"</span>" ); + +return ret; +} + + +static int __hpp__color_fmt (struct perf_hpp *hpp,struct hist_entry *he, +u64 (*get_field)(struct hist_entry *)) +{ +int ret; +double percent =0.0 ; +struct hists *hists = he->hists; + +if (hists->stats.total_period) + percent =100.0 *get_field (he) / hists->stats.total_period; + + ret =perf_gtk__percent_color_snprintf (hpp->buf, hpp->size, percent); + +if (symbol_conf.event_group) { +int prev_idx, idx_delta; +struct perf_evsel *evsel =hists_to_evsel (hists); +struct hist_entry *pair; +int nr_members = evsel->nr_members; + +if (nr_members <=1 ) +return ret; + + prev_idx =perf_evsel__group_idx (evsel); + +list_for_each_entry (pair, &he->pairs.head, pairs.node) { +u64 period =get_field (pair); +u64 total = pair->hists->stats.total_period; + + evsel =hists_to_evsel (pair->hists); + idx_delta =perf_evsel__group_idx (evsel) - prev_idx -1 ; + +while (idx_delta--) { + ret +=scnprintf (hpp->buf + ret, hpp->size -ret," " ); + ret +=perf_gtk__percent_color_snprintf (hpp->buf + ret, + hpp->size - ret, +0.0 ); + } + + percent =100.0 * period / total; + ret +=scnprintf (hpp->buf + ret, hpp->size -ret," " ); + ret +=perf_gtk__percent_color_snprintf (hpp->buf + ret, + hpp->size - ret, + percent); + + prev_idx =perf_evsel__group_idx (evsel); + } + + idx_delta = nr_members - prev_idx -1 ; + +while (idx_delta--) { + ret +=scnprintf (hpp->buf + ret, hpp->size -ret," " ); + ret +=perf_gtk__percent_color_snprintf (hpp->buf + ret, + hpp->size - ret, +0.0 ); + } + } +return ret; +} + +#define __HPP_COLOR_PERCENT_FN (_type, _field) \ +static u64 he_get_##_field(struct hist_entry *he) \ { \ - struct hists *hists = he->hists; \ - double percent = 100.0 * he->stat._field / hists->stats.total_period; \ - const char *markup; \ - int ret = 0; \ + return he->stat._field; \ +} \ \ - markup = perf_gtk__get_percent_color(percent); \ - if (markup) \ - ret += scnprintf(hpp->buf, hpp->size, "%s", markup); \ - ret += scnprintf(hpp->buf + ret, hpp->size - ret, "%6.2f%%", percent); \ - if (markup) \ - ret += scnprintf(hpp->buf + ret, hpp->size - ret, "</span>"); \ - \ - return ret; \ +static int perf_gtk__hpp_color_##_type(struct perf_hpp *hpp, \ + struct hist_entry *he) \ +{ \ + return __hpp__color_fmt(hpp, he, he_get_##_field); \ } -HPP__COLOR_FN(overhead, period) -HPP__COLOR_FN(overhead_sys, period_sys) -HPP__COLOR_FN(overhead_us, period_us) -HPP__COLOR_FN(overhead_guest_sys, period_guest_sys) -HPP__COLOR_FN(overhead_guest_us, period_guest_us) +__HPP_COLOR_PERCENT_FN(overhead, period) +__HPP_COLOR_PERCENT_FN(overhead_sys, period_sys) +__HPP_COLOR_PERCENT_FN(overhead_us, period_us) +__HPP_COLOR_PERCENT_FN(overhead_guest_sys, period_guest_sys) +__HPP_COLOR_PERCENT_FN(overhead_guest_us, period_guest_us) -#undef HPP__COLOR_FN +#undef __HPP_COLOR_PERCENT_FNvoid perf_gtk__init_hpp (void )
@@ -70,6 +140,7 @@ static void perf_gtk__show_hists(GtkWidget *window, struct hists *hists)struct perf_hpp hpp = { .buf = s, .size =sizeof (s), + .ptr =hists_to_evsel (hists), }; nr_cols =0 ;
@@ -94,7 +165,6 @@ static void perf_gtk__show_hists(GtkWidget *window, struct hists *hists)perf_hpp__for_each_format (fmt) { fmt->header (&hpp); - gtk_tree_view_insert_column_with_attributes(GTK_TREE_VIEW(view), -1, s, renderer, "markup",
Signature · 1 line
1.7.11.7
Re: [PATCH 08/13] perf gtk/browser: Add support for event group view
Em Tue, Jan 15, 2013 at 04:31:52PM +0900, Namhyung Kim escreveu:
Quoted 9 lines
From: Namhyung Kim <[email protected]>
Show group members' overhead also when showing the leader's if event group is enabled. Use macro for defining hpp functions which looks almost identical.
Unlike other hpp backend, GTK+ needs to print dummy 0.00% output since it's displayed with variable width fonts. So that simply skipping with %*s trick won't work well here.
I'm illiterate on GTK+ internals, but I kinda expected that it could provide a way to avoid such things, i.e. a way to position the cursor in such circumstance without having to actually print the '0.00%' :-\
- Arnaldo
Quoted 154 lines
Cc Jiri Olsa <[email protected]>Cc Stephane Eranian <[email protected]>Cc Pekka Enberg <[email protected]>Signed-off-by Namhyung Kim <[email protected]>--- tools/perf/ui/gtk/hists.c | 114 +++++++++++++++++++++++++++++++++++++--------- 1 file changed, 92 insertions(+), 22 deletions(-)tools/perf/ui/gtk/hists.c +92
-22@@ -8,32 +8,102 @@#define MAX_COLUMNS 32 -#define HPP__COLOR_FN (_name, _field) \ -static int perf_gtk__hpp_color_ ## _name(struct perf_hpp *hpp, \ - struct hist_entry *he) \ +static int perf_gtk__percent_color_snprintf (char *buf,size_t size, +double percent) +{ +int ret =0 ; +const char *markup; + + markup =perf_gtk__get_percent_color (percent); +if (markup) + ret +=scnprintf (buf, size, markup); + + ret +=scnprintf (buf + ret, size - ret,"%6.2f%%" , percent); + +if (markup) + ret +=scnprintf (buf + ret, size - ret,"</span>" ); + +return ret; +} + + +static int __hpp__color_fmt (struct perf_hpp *hpp,struct hist_entry *he, +u64 (*get_field)(struct hist_entry *)) +{ +int ret; +double percent =0.0 ; +struct hists *hists = he->hists; + +if (hists->stats.total_period) + percent =100.0 *get_field (he) / hists->stats.total_period; + + ret =perf_gtk__percent_color_snprintf (hpp->buf, hpp->size, percent); + +if (symbol_conf.event_group) { +int prev_idx, idx_delta; +struct perf_evsel *evsel =hists_to_evsel (hists); +struct hist_entry *pair; +int nr_members = evsel->nr_members; + +if (nr_members <=1 ) +return ret; + + prev_idx =perf_evsel__group_idx (evsel); + +list_for_each_entry (pair, &he->pairs.head, pairs.node) { +u64 period =get_field (pair); +u64 total = pair->hists->stats.total_period; + + evsel =hists_to_evsel (pair->hists); + idx_delta =perf_evsel__group_idx (evsel) - prev_idx -1 ; + +while (idx_delta--) { + ret +=scnprintf (hpp->buf + ret, hpp->size -ret," " ); + ret +=perf_gtk__percent_color_snprintf (hpp->buf + ret, + hpp->size - ret, +0.0 ); + } + + percent =100.0 * period / total; + ret +=scnprintf (hpp->buf + ret, hpp->size -ret," " ); + ret +=perf_gtk__percent_color_snprintf (hpp->buf + ret, + hpp->size - ret, + percent); + + prev_idx =perf_evsel__group_idx (evsel); + } + + idx_delta = nr_members - prev_idx -1 ; + +while (idx_delta--) { + ret +=scnprintf (hpp->buf + ret, hpp->size -ret," " ); + ret +=perf_gtk__percent_color_snprintf (hpp->buf + ret, + hpp->size - ret, +0.0 ); + } + } +return ret; +} + +#define __HPP_COLOR_PERCENT_FN (_type, _field) \ +static u64 he_get_##_field(struct hist_entry *he) \ { \ - struct hists *hists = he->hists; \ - double percent = 100.0 * he->stat._field / hists->stats.total_period; \ - const char *markup; \ - int ret = 0; \ + return he->stat._field; \ +} \ \ - markup = perf_gtk__get_percent_color(percent); \ - if (markup) \ - ret += scnprintf(hpp->buf, hpp->size, "%s", markup); \ - ret += scnprintf(hpp->buf + ret, hpp->size - ret, "%6.2f%%", percent); \ - if (markup) \ - ret += scnprintf(hpp->buf + ret, hpp->size - ret, "</span>"); \ - \ - return ret; \ +static int perf_gtk__hpp_color_##_type(struct perf_hpp *hpp, \ + struct hist_entry *he) \ +{ \ + return __hpp__color_fmt(hpp, he, he_get_##_field); \ } -HPP__COLOR_FN(overhead, period) -HPP__COLOR_FN(overhead_sys, period_sys) -HPP__COLOR_FN(overhead_us, period_us) -HPP__COLOR_FN(overhead_guest_sys, period_guest_sys) -HPP__COLOR_FN(overhead_guest_us, period_guest_us) +__HPP_COLOR_PERCENT_FN(overhead, period) +__HPP_COLOR_PERCENT_FN(overhead_sys, period_sys) +__HPP_COLOR_PERCENT_FN(overhead_us, period_us) +__HPP_COLOR_PERCENT_FN(overhead_guest_sys, period_guest_sys) +__HPP_COLOR_PERCENT_FN(overhead_guest_us, period_guest_us) -#undef HPP__COLOR_FN +#undef __HPP_COLOR_PERCENT_FNvoid perf_gtk__init_hpp (void )@@ -70,6 +140,7 @@ static void perf_gtk__show_hists(GtkWidget *window, struct hists *hists)struct perf_hpp hpp = { .buf = s, .size =sizeof (s), + .ptr =hists_to_evsel (hists), }; nr_cols =0 ;@@ -94,7 +165,6 @@ static void perf_gtk__show_hists(GtkWidget *window, struct hists *hists)perf_hpp__for_each_format (fmt) { fmt->header (&hpp); - gtk_tree_view_insert_column_with_attributes(GTK_TREE_VIEW(view), -1, s, renderer, "markup",Signature · 1 line
1.7.11.7
Namhyung Kim
maintainerRe: [PATCH 08/13] perf gtk/browser: Add support for event group view
Quoted 7 lines · Arnaldo Carvalho de Melo
Em Tue, Jan 15, 2013 at 04:31:52PM +0900, Namhyung Kim escreveu:
Quoted 6 lines
From: Namhyung Kim <[email protected]>
Show group members' overhead also when showing the leader's if event group is enabled. Use macro for defining hpp functions which looks almost identical.
Quoted 7 lines
Quoted 3 lines
Unlike other hpp backend, GTK+ needs to print dummy 0.00% output since it's displayed with variable width fonts. So that simply skipping with %*s trick won't work well here.
I'm illiterate on GTK+ internals, but I kinda expected that it could provide a way to avoid such things, i.e. a way to position the cursor in such circumstance without having to actually print the '0.00%' :-\
I also don't understand the internals so I just guess that it might (only?) be acheived by adding new columns for each group members.
Of course I'm fine with this way too, but one thing I'm not sure is how the column header look like. Currently "Overhead" is displayed on the header for period of every group members - say, there's a group consists of two events: {branches,branch-misses}. It'd look like:
Overhead Command Shared Object Symbol .............. ....... ............. .......
98.32% 31.16% a.out a.out [.] foo
Maybe we can use simple event name for this, but it seems it makes to hard to recognize which one is which if other output options like --show-total-period, --showcpuutilization and/or --show-nr-samples.
Adding current header name to event name will fix the problem but it probably occupies too much screen width especially for long named tracepoint or PMU-specific events like "compaction:mm_compaction_isolate_migratepages".
Overhead/branches Overhead/branch-misses sys/branches sys/branch-misses usr/branches usr/branch-misses Command Shared Object Symbol
................. ...................... ............ ................. ............ ................. ....... ............. .......
98.32% 31.16% 0.00% 0.00% 98.32% 31.16% a.out a.out [.] fooIf you have a better idea or other way to place the cursor without printing bogus 0.00% on GTK, please let me know.
Thanks,
Namhyung
Re: [PATCH 08/13] perf gtk/browser: Add support for event group view
Em Wed, Jan 16, 2013 at 07:25:02PM +0900, Namhyung Kim escreveu:
Quoted 5 lines
Adding current header name to event name will fix the problem but it probably occupies too much screen width especially for long named tracepoint or PMU-specific events like "compaction:mm_compaction_isolate_migratepages".
Quoted 7 lines
Overhead/branches Overhead/branch-misses sys/branches sys/branch-misses usr/branches usr/branch-misses Command Shared Object Symbol ................. ...................... ............ ................. ............ ................. ....... ............. ....... 98.32% 31.16% 0.00% 0.00% 98.32% 31.16% a.out a.out [.] fooIf you have a better idea or other way to place the cursor without printing bogus 0.00% on GTK, please let me know.
Compacting it using an extra line:
Overhead............... sys.................... usr.................... branches branch-misses branches branch-misses branches branch-misses Command Shared Object Symbol ........ ............. ........ ............. ........ ............. ....... ............. .......
98.32% 31.16% 0.00% 0.00% 98.32% 31.16% a.out a.out [.] foo
It could even use some reference:
Overhead..................... sys......... usr...........
branches(1) branch-misses(2) (1) (2) (1) (2) Command Shared Object Symbol
........... ................ ..... ..... ...... ...... ....... ............. .......
98.32% 31.16% 0.00% 0.00% 98.32% 31.16% a.out a.out [.] fooThe (1) could be done with a superscript number or even just using a different fore/background color, to use fewer columns.
One other way, that would scale for really long event names, would be to have the event list in the first few lines and then:
Events:
1. branches
2. branch-misses
Overhead...... sys......... usr........... (1) (2) (1) (2) (1) (2) Command Shared Object Symbol ...... ...... ..... ..... ...... ...... ....... ............. ....... 98.32% 31.16% 0.00% 0.00% 98.32% 31.16% a.out a.out [.] foo
I think you could switch to/from each of these forms using a hotkey, that would influence how the hist_entry__snprintf() routine would work, either using perf_evsel__name() or evsel->idx :-)
This way if at some point the user wants to expand/compress the lines, it will be possible to do so quickly, just pressing the hotkey.
- Arnaldo
Namhyung Kim
maintainerRe: [PATCH 08/13] perf gtk/browser: Add support for event group view
Hi Arnaldo,
Quoted 43 lines · Arnaldo Carvalho de Melo
Em Wed, Jan 16, 2013 at 07:25:02PM +0900, Namhyung Kim escreveu:
Quoted 5 lines
Adding current header name to event name will fix the problem but it probably occupies too much screen width especially for long named tracepoint or PMU-specific events like "compaction:mm_compaction_isolate_migratepages".
Quoted 7 lines
Overhead/branches Overhead/branch-misses sys/branches sys/branch-misses usr/branches usr/branch-misses Command Shared Object Symbol ................. ...................... ............ ................. ............ ................. ....... ............. ....... 98.32% 31.16% 0.00% 0.00% 98.32% 31.16% a.out a.out [.] fooIf you have a better idea or other way to place the cursor without printing bogus 0.00% on GTK, please let me know.
Compacting it using an extra line:
Overhead............... sys.................... usr.................... branches branch-misses branches branch-misses branches branch-misses Command Shared Object Symbol ........ ............. ........ ............. ........ ............. ....... ............. .......98.32% 31.16% 0.00% 0.00% 98.32% 31.16% a.out a.out [.] foo
It could even use some reference:
Overhead..................... sys......... usr........... branches(1) branch-misses(2) (1) (2) (1) (2) Command Shared Object Symbol ........... ................ ..... ..... ...... ...... ....... ............. ....... 98.32% 31.16% 0.00% 0.00% 98.32% 31.16% a.out a.out [.] fooThe (1) could be done with a superscript number or even just using a different fore/background color, to use fewer columns.
One other way, that would scale for really long event names, would be to have the event list in the first few lines and then:
Events:
1. branches
2. branch-missesOverhead...... sys......... usr........... (1) (2) (1) (2) (1) (2) Command Shared Object Symbol ...... ...... ..... ..... ...... ...... ....... ............. ....... 98.32% 31.16% 0.00% 0.00% 98.32% 31.16% a.out a.out [.] foo
Quoted 6 lines
I think you could switch to/from each of these forms using a hotkey, that would influence how the
hist_entry__snprintf()routine would work, either usingperf_evsel__name()orevsel->idx:-)This way if at some point the user wants to expand/compress the lines, it will be possible to do so quickly, just pressing the hotkey.
By saying hotkey, I guess you meant to use it for TUI. However TUI doesn't provide those header lines. ;-)
As this extra line (and hotkey) thing might add complexity to the patchset, I'd like to separate it to a different work and to focus on the basic feature with current behavior. Is it acceptable for you?
Thanks,
Namhyung
Re: [PATCH 08/13] perf gtk/browser: Add support for event group view
Em Tue, Jan 22, 2013 at 02:58:36PM +0900, Namhyung Kim escreveu:
Quoted 50 lines
Hi Arnaldo,
Quoted 47 lines · Arnaldo Carvalho de Melo
Em Wed, Jan 16, 2013 at 07:25:02PM +0900, Namhyung Kim escreveu:
Quoted 5 lines
Adding current header name to event name will fix the problem but it probably occupies too much screen width especially for long named tracepoint or PMU-specific events like "compaction:mm_compaction_isolate_migratepages".
Quoted 7 lines
Overhead/branches Overhead/branch-misses sys/branches sys/branch-misses usr/branches usr/branch-misses Command Shared Object Symbol ................. ...................... ............ ................. ............ ................. ....... ............. ....... 98.32% 31.16% 0.00% 0.00% 98.32% 31.16% a.out a.out [.] fooIf you have a better idea or other way to place the cursor without printing bogus 0.00% on GTK, please let me know.
Compacting it using an extra line:
Overhead............... sys.................... usr.................... branches branch-misses branches branch-misses branches branch-misses Command Shared Object Symbol ........ ............. ........ ............. ........ ............. ....... ............. .......98.32% 31.16% 0.00% 0.00% 98.32% 31.16% a.out a.out [.] foo
It could even use some reference:
Overhead..................... sys......... usr........... branches(1) branch-misses(2) (1) (2) (1) (2) Command Shared Object Symbol ........... ................ ..... ..... ...... ...... ....... ............. ....... 98.32% 31.16% 0.00% 0.00% 98.32% 31.16% a.out a.out [.] fooThe (1) could be done with a superscript number or even just using a different fore/background color, to use fewer columns.
One other way, that would scale for really long event names, would be to have the event list in the first few lines and then:
Events:
1. branches
2. branch-missesOverhead...... sys......... usr........... (1) (2) (1) (2) (1) (2) Command Shared Object Symbol ...... ...... ..... ..... ...... ...... ....... ............. ....... 98.32% 31.16% 0.00% 0.00% 98.32% 31.16% a.out a.out [.] fooI think you could switch to/from each of these forms using a hotkey, that would influence how the
hist_entry__snprintf()routine would work, either usingperf_evsel__name()orevsel->idx:-)
Quoted 5 lines
Quoted 2 lines
This way if at some point the user wants to expand/compress the lines, it will be possible to do so quickly, just pressing the hotkey.
By saying hotkey, I guess you meant to use it for TUI. However TUI doesn't provide those header lines. ;-)
See? I was trying to hit two rabbits with one stone, but you realised that...
Quoted 3 lines
As this extra line (and hotkey) thing might add complexity to the patchset, I'd like to separate it to a different work and to focus on the basic feature with current behavior. Is it acceptable for you?
Sure thing.
- Arnaldo
Namhyung Kim
maintainerPATCH 9/13 perf gtk/browser: Trim column header string when event group enabled
In reply to Namhyung KimFrom: Namhyung Kim <[email protected]>
When event group feature is enabled, each column header is expanded to match with the whole group column width. But this is not needed for GTK+ browser since it usually use variable-width fonts. So trim it.
As we have ltrim() implementation in builtin-script.c move it to the more generic location of util/string.c.
--- tools/perf/builtin-script.c | 12 ------------ tools/perf/ui/gtk/hists.c | 2 +- tools/perf/util/string.c | 18 ++++++++++++++++++ tools/perf/util/util.h | 1 + 4 files changed, 20 insertions(+), 13 deletions(-)
tools/perf/builtin-script.c +0 -12
@@ -909,18 +909,6 @@ static const char *ends_with(const char *str, const char *suffix)return NULL ; } -static char *ltrim (char *str) -{ -int len =strlen (str); - -while (len &&isspace (*str)) { - len--; - str++; - } - -return str; -} -static int read_script_info (struct script_desc *desc,const char *filename) {char line[BUFSIZ ], *p;
tools/perf/ui/gtk/hists.c +1 -1
@@ -166,7 +166,7 @@ static void perf_gtk__show_hists(GtkWidget *window, struct hists *hists)perf_hpp__for_each_format (fmt) { fmt->header (&hpp);gtk_tree_view_insert_column_with_attributes (GTK_TREE_VIEW (view), --1 , s, +-1 ,ltrim (s), renderer,"markup" , col_idx++,NULL ); }
tools/perf/util/string.c +18 -0
@@ -332,6 +332,24 @@ char *strxfrchar(char *s, char from, char to) }/** +* ltrim - Removes leading whitespace from @s. +* @s: The string to be stripped. +* +* Return pointer to the first non-whitespace character in @s. +*/ +char *ltrim (char *s) +{ +int len =strlen (s); + +while (len &&isspace (*s)) { + len--; + s++; + } + +return s; +} + +/** * rtrim - Removes trailing whitespace from @s. * @s: The string to be stripped. *
tools/perf/util/util.h +1 -0
@@ -265,6 +265,7 @@ bool is_power_of_2(unsigned long n)size_t hex_width (u64 v);int hex2u64 (const char *ptr,u64 *val); +char *ltrim (char *s);char *rtrim (char *s);void dump_stack (void );
Signature · 1 line
1.7.11.7
Namhyung Kim
maintainerPATCH 10/13 perf report: Bypass non-leader events when event group is enabled
In reply to Namhyung KimFrom: Namhyung Kim <[email protected]>
Since we have all necessary information in the leader events and other members don't, bypass members. Member events will be shown along with the leaders if event group is enabled.
--- tools/perf/builtin-report.c | 4 ++++ tools/perf/ui/browsers/hists.c | 41 +++++++++++++++++++++++++++++++++++------ tools/perf/ui/gtk/hists.c | 4 ++++ 3 files changed, 43 insertions(+), 6 deletions(-)
tools/perf/builtin-report.c +4 -0
@@ -319,6 +319,10 @@ static int perf_evlist__tty_browse_hists(struct perf_evlist *evlist,struct hists *hists = &pos->hists;const char *evname =perf_evsel__name (pos); +if (symbol_conf.event_group && + !perf_evsel__is_group_leader (pos)) +continue ; +hists__fprintf_nr_sample_events (hists, evname, stdout);hists__fprintf (hists, true,0 ,0 , stdout);fprintf (stdout,"\n\n" );
tools/perf/ui/browsers/hists.c +35 -6
70 lines not on this page · raw message
tools/perf/ui/gtk/hists.c +4 -0
@@ -267,6 +267,10 @@ int perf_evlist__gtk_browse_hists(struct perf_evlist *evlist,GtkWidget *scrolled_window;GtkWidget *tab_label; +if (symbol_conf.event_group && + !perf_evsel__is_group_leader (pos)) +continue ; + scrolled_window =gtk_scrolled_window_new (NULL ,NULL ); gtk_scrolled_window_set_policy(GTK_SCROLLED_WINDOW(scrolled_window),
Signature · 1 line
1.7.11.7
Namhyung Kim
maintainerPATCH 11/13 perf report: Show group description when event group is enabled
In reply to Namhyung KimFrom: Namhyung Kim <[email protected]>
When using event group viewer, it's better to show the group description rather than the leader information alone.
If a leader did not contain any member, it's a non-group event.
--- tools/perf/builtin-report.c | 15 +++++++++++++++ tools/perf/ui/browsers/hists.c | 25 +++++++++++++++++++++++++ tools/perf/ui/gtk/hists.c | 14 +++++++++++--- tools/perf/util/evsel.c | 25 +++++++++++++++++++++++++ tools/perf/util/evsel.h | 8 ++++++++ 5 files changed, 84 insertions(+), 3 deletions(-)
tools/perf/builtin-report.c +15 -0
@@ -299,6 +299,21 @@ static size_t hists__fprintf_nr_sample_events(struct hists *self,char unit;unsigned long nr_samples = self->stats.nr_events[PERF_RECORD_SAMPLE ];u64 nr_events = self->stats.total_period; +struct perf_evsel *evsel =hists_to_evsel (self); +char buf[512 ]; +size_t size =sizeof (buf); + +if (symbol_conf.event_group && evsel->nr_members >1 ) { +struct perf_evsel *pos; + +perf_evsel__group_desc (evsel, buf, size); + evname = buf; + +for_each_group_member (pos, evsel) { + nr_samples += pos->hists.stats.nr_events[PERF_RECORD_SAMPLE ]; + nr_events += pos->hists.stats.total_period; + } + } nr_samples =convert_unit (nr_samples, &unit); ret =fprintf (fp,"# Samples: %lu%c" , nr_samples, unit);
tools/perf/ui/browsers/hists.c +25 -0
@@ -1193,6 +1193,21 @@ static int hists__browser_title(struct hists *hists, char *bf, size_t size,const struct thread *thread = hists->thread_filter;unsigned long nr_samples = hists->stats.nr_events[PERF_RECORD_SAMPLE ];u64 nr_events = hists->stats.total_period; +struct perf_evsel *evsel =hists_to_evsel (hists); +char buf[512 ]; +size_t buflen =sizeof (buf); + +if (symbol_conf.event_group && evsel->nr_members >1 ) { +struct perf_evsel *pos; + +perf_evsel__group_desc (evsel, buf, buflen); + ev_name = buf; + +for_each_group_member (pos, evsel) { + nr_samples += pos->hists.stats.nr_events[PERF_RECORD_SAMPLE ]; + nr_events += pos->hists.stats.total_period; + } + } nr_samples =convert_unit (nr_samples, &unit); printed = scnprintf(bf, size,
@@ -1589,6 +1604,16 @@ static void perf_evsel_menu__write(struct ui_browser *browser,ui_browser__set_color (browser, current_entry ?HE_COLORSET_SELECTED :HE_COLORSET_NORMAL ); +if (symbol_conf.event_group && evsel->nr_members >1 ) { +struct perf_evsel *pos; + + ev_name =perf_evsel__group_name (evsel); + +for_each_group_member (pos, evsel) { + nr_events += pos->hists.stats.nr_events[PERF_RECORD_SAMPLE ]; + } + } + nr_events =convert_unit (nr_events, &unit); printed =scnprintf (bf,sizeof (bf),"%lu%c%s%s" , nr_events, unit, unit ==' ' ?"" :" " , ev_name);
tools/perf/ui/gtk/hists.c +11 -3
@@ -266,10 +266,18 @@ int perf_evlist__gtk_browse_hists(struct perf_evlist *evlist,const char *evname =perf_evsel__name (pos);GtkWidget *scrolled_window;GtkWidget *tab_label; +char buf[512 ]; +size_t size =sizeof (buf); -if (symbol_conf.event_group && - !perf_evsel__is_group_leader (pos)) -continue ; +if (symbol_conf.event_group) { +if (!perf_evsel__is_group_leader (pos)) +continue ; + +if (pos->nr_members >1 ) { +perf_evsel__group_desc (pos, buf, size); + evname = buf; + } + } scrolled_window =gtk_scrolled_window_new (NULL ,NULL );
tools/perf/util/evsel.c +25 -0
@@ -434,6 +434,31 @@ const char *perf_evsel__name(struct perf_evsel *evsel)return evsel->name ?:"unknown" ; } +const char *perf_evsel__group_name (struct perf_evsel *evsel) +{ +return evsel->group_name ?:"anon group" ; +} + +int perf_evsel__group_desc (struct perf_evsel *evsel,char *buf,size_t size) +{ +int ret; +struct perf_evsel *pos; +const char *group_name =perf_evsel__group_name (evsel); + + ret =scnprintf (buf, size,"%s" , group_name); + + ret +=scnprintf (buf + ret, size - ret," { %s" , +perf_evsel__name (evsel)); + +for_each_group_member (pos, evsel) + ret +=scnprintf (buf + ret, size - ret,", %s" , +perf_evsel__name (pos)); + + ret +=scnprintf (buf + ret, size - ret," }" ); + +return ret; +} +/* * The enable_on_exec/disabled value strategy: *
tools/perf/util/evsel.h +8 -0
@@ -113,6 +113,8 @@ extern const char *perf_evsel__sw_names[PERF_COUNT_SW_MAX];int __perf_evsel__hw_cache_type_op_res_name (u8 type,u8 op,u8 result,char *bf,size_t size);const char *perf_evsel__name (struct perf_evsel *evsel); +const char *perf_evsel__group_name (struct perf_evsel *evsel); +int perf_evsel__group_desc (struct perf_evsel *evsel,char *buf,size_t size);int perf_evsel__alloc_fd (struct perf_evsel *evsel,int ncpus,int nthreads);int perf_evsel__alloc_id (struct perf_evsel *evsel,int ncpus,int nthreads);
@@ -265,4 +267,10 @@ static inline int perf_evsel__group_idx(struct perf_evsel *evsel) {return evsel->idx - evsel->leader->idx; } + +#define for_each_group_member (_evsel, _leader) \ +for ((_evsel) = list_entry((_leader)->node.next, struct perf_evsel, node); \ + (_evsel) && (_evsel)->leader == (_leader); \ + (_evsel) = list_entry((_evsel)->node.next, struct perf_evsel, node)) +#endif /* __PERF_EVSEL_H */
Signature · 1 line
1.7.11.7
Namhyung Kim
maintainerPATCH 12/13 perf report: Add --group option
In reply to Namhyung KimFrom: Namhyung Kim <[email protected]>
Add --group option to enable event grouping. When enabled, all the group members information will be shown together with the leader.
--- tools/perf/Documentation/perf-report.txt | 3 +++ tools/perf/builtin-report.c | 2 ++ 2 files changed, 5 insertions(+)
tools/perf/Documentation/perf-report.txt +3 -0
@@ -203,6 +203,9 @@ OPTIONS --objdump=<path>:: Path to objdump binary. +--group:: + Show event group information together. + SEE ALSO -------- linkperf:perf-stat[1], linkperf:perf-annotate[1]
tools/perf/builtin-report.c +2 -0
@@ -668,6 +668,8 @@ int cmd_report(int argc, const char **argv, const char *prefix __maybe_unused) "Specify disassembler style (e.g. -M intel for intel syntax)"),OPT_BOOLEAN (0 ,"show-total-period" , &symbol_conf.show_total_period,"Show a column with the sum of periods" ), +OPT_BOOLEAN (0 ,"group" , &symbol_conf.event_group, +"Show event group information together" ),OPT_CALLBACK_NOOPT ('b' ,"branch-stack" , &sort__branch_mode,"" ,"use branch records for histogram filling" , parse_branch_mode), OPT_STRING(0, "objdump", &objdump_path, "path",
Signature · 1 line
1.7.11.7
Namhyung Kim
maintainerPATCH 13/13 perf report: Add report.group config option
In reply to Namhyung KimFrom: Namhyung Kim <[email protected]>
Add report.group config option for setting default value of event group view. It affects the report output only if perf.data contains event group info.
A user can write .perfconfig file like below to enable group view by default:
$ cat ~/.perfconfig [report] group = true
And it can be disabled through command line:
$ perf report --no-group
--- tools/perf/builtin-report.c | 13 +++++++++++++ 1 file changed, 13 insertions(+)
tools/perf/builtin-report.c +13 -0
@@ -8,6 +8,7 @@#include "builtin.h" #include "util/util.h" +#include "util/cache.h" #include "util/annotate.h" #include "util/color.h"
@@ -54,6 +55,16 @@ struct perf_report {DECLARE_BITMAP (cpu_bitmap,MAX_NR_CPUS ); }; +static int perf_report_config (const char *var,const char *value,void *cb) +{ +if (!strcmp (var,"report.group" )) { + symbol_conf.event_group =perf_config_bool (var, value); +return 0 ; + } + +return perf_default_config (var, value, cb); +} + static int perf_report__add_branch_hist_entry(struct perf_tool *tool, struct addr_location *al, struct perf_sample *sample,
@@ -677,6 +688,8 @@ int cmd_report(int argc, const char **argv, const char *prefix __maybe_unused) OPT_END() }; +perf_config (perf_report_config,NULL ); + argc =parse_options (argc, argv, options, report_usage,0 );if (report.use_stdio)
Signature · 1 line
1.7.11.7
Em Tue, Jan 15, 2013 at 04:31:44PM +0900, Namhyung Kim escreveu:
Quoted 11 lines
This is my 8th attempt to enable the event group view for perf report. For basic idea and usage example, please see my initial post [1].
This version is rebased on the current acme/perf/core and it's mainly changed due to a refactoring of gtk browser.
You can get this via 'perf/group-v8' branch on my tree:
git://git.kernel.org/pub/scm/linux/kernel/git/namhyung/linux-perf.gitAny comments are welcome, thanks,
It would be great if examples of this in use were provided.
Detailed ones at that, showing how things were in the past, how the info can be retrieved when using 'perf report --stdio', perhaps 'perf evlist' can show the stored group information too, how the 'perf report' output looks like when showing group info, things like that.
- Arnaldo
Namhyung Kim
maintainerQuoted 11 lines · Arnaldo Carvalho de Melo
Em Tue, Jan 15, 2013 at 04:31:44PM +0900, Namhyung Kim escreveu:
Quoted 10 lines
This is my 8th attempt to enable the event group view for perf report. For basic idea and usage example, please see my initial post [1].
This version is rebased on the current acme/perf/core and it's mainly changed due to a refactoring of gtk browser.
You can get this via 'perf/group-v8' branch on my tree:
git://git.kernel.org/pub/scm/linux/kernel/git/namhyung/linux-perf.git
Quoted 8 lines
Quoted 1 line
Any comments are welcome, thanks,
It would be great if examples of this in use were provided.
Detailed ones at that, showing how things were in the past, how the info can be retrieved when using 'perf report --stdio', perhaps 'perf evlist' can show the stored group information too, how the 'perf report' output looks like when showing group info, things like that.
I think most of aforementioned things were on the initial post. I thought it's too verbose to be contained to the cover letter every time so I just put the link.
But I'll include it next time and work on the perf evlist too.
Thanks,
Namhyung
Em Wed, Jan 16, 2013 at 07:33:52PM +0900, Namhyung Kim escreveu:
Quoted 8 lines · Arnaldo Carvalho de Melo
Quoted 6 lines
It would be great if examples of this in use were provided.
Detailed ones at that, showing how things were in the past, how the info can be retrieved when using 'perf report --stdio', perhaps 'perf evlist' can show the stored group information too, how the 'perf report' output looks like when showing group info, things like that.
Quoted 5 lines
I think most of aforementioned things were on the initial post. I thought it's too verbose to be contained to the cover letter every time so I just put the link.
But I'll include it next time and work on the perf evlist too.
Thank you!
- Arnaldo