jevents/pmu-events improvements
Subsystems (2)
Ian Rogers
reviewerv2 cover jevents/pmu-events improvements
Add an optimization to jevents using the metric code, rewrite metrics
in terms of each other in order to minimize size and improve
readability. For example, on Power8other_stall_cpi is rewritten from:
"PM_CMPLU_STALL / PM_RUN_INST_CMPL - PM_CMPLU_STALL_BRU_CRU / PM_RUN_INST_CMPL - PM_CMPLU_STALL_FXU / PM_RUN_INST_CMPL - PM_CMPLU_STALL_VSU / PM_RUN_INST_CMPL - PM_CMPLU_STALL_LSU / PM_RUN_INST_CMPL - PM_CMPLU_STALL_NTCG_FLUSH / PM_RUN_INST_CMPL - PM_CMPLU_STALL_NO_NTF / PM_RUN_INST_CMPL" to:
"stall_cpi - bru_cru_stall_cpi - fxu_stall_cpi - vsu_stall_cpi - lsu_stall_cpi - ntcg_flush_cpi - no_ntf_stall_cpi"
Which more closely matches the definition on Power9.
A limitation of the substitutions are that they depend on strict equality and the shape of the tree. This means that for "a + b + c" then a substitution of "a + b" will succeed while "b + c" will fail (the LHS for "+ c" is "a + b" not just "b").
Separate out the events and metrics in the pmu-events tables saving 14.8% in the table size while making it that metrics no longer need to iterate over all events and vice versa. These changes remove evsel's direct metric support as the pmu_event no longer has a metric to populate it. This is a minor issue as the code wasn't working properly, metrics for this are rare and can still be properly ran using '-M'.
Add an ability to just build certain models into the jevents generated pmu-metrics.c code. This functionality is appropriate for operating systems like ChromeOS, that aim to minimize binary size and know all the target CPU models.
v2. Rebase. Modify the code that skips rewriting a metric with the
same name with itself, to make the name check case insensitive.
Ian Rogers (9):
perf jevents metric: Correct Function equality
perf jevents metric: Add ability to rewrite metrics in terms of others perf jevents: Rewrite metrics in the same file with each other perf pmu-events: Separate metric out of pmu_event
perf stat: Remove evsel metric_name/expr
perf jevents: Combine table prefix and suffix writing
perf pmu-events: Introduce pmu_metrics_table
perf jevents: Generate metrics and events as separate tables perf jevents: Add model list option
tools/perf/arch/arm64/util/pmu.c | 23 +- tools/perf/arch/powerpc/util/header.c | 4 +- tools/perf/builtin-list.c | 20 +- tools/perf/builtin-stat.c | 1 - tools/perf/pmu-events/Build | 3 +- tools/perf/pmu-events/empty-pmu-events.c | 111 ++++++- tools/perf/pmu-events/jevents.py | 353 ++++++++++++++++++----- tools/perf/pmu-events/metric.py | 79 ++++- tools/perf/pmu-events/metric_test.py | 10 + tools/perf/pmu-events/pmu-events.h | 26 +- tools/perf/tests/expand-cgroup.c | 4 +- tools/perf/tests/parse-metric.c | 4 +- tools/perf/tests/pmu-events.c | 68 ++--- tools/perf/util/cgroup.c | 1 - tools/perf/util/evsel.c | 2 - tools/perf/util/evsel.h | 2 - tools/perf/util/metricgroup.c | 203 +++++++------ tools/perf/util/metricgroup.h | 4 +- tools/perf/util/parse-events.c | 2 - tools/perf/util/pmu.c | 44 +-- tools/perf/util/pmu.h | 10 +- tools/perf/util/print-events.c | 32 +- tools/perf/util/print-events.h | 3 +- tools/perf/util/python.c | 7 - tools/perf/util/stat-shadow.c | 112 ------- tools/perf/util/stat.h | 1 - 26 files changed, 666 insertions(+), 463 deletions(-)
Signature · 1 line
2.39.0.314.g84b9a713c41-goog
Ian Rogers
reviewerPATCH v2 1/9 perf jevents metric: Correct Function equality
rhs may not be defined, say for source_count, so add a guard.
--- tools/perf/pmu-events/metric.py | 6 ++++-- 1 file changed, 4 insertions(+), 2 deletions(-)
tools/perf/pmu-events/metric.py +4 -2
@@ -261,8 +261,10 @@ class Function(Expression): def Equals(self, other: Expression) -> bool: if isinstance(other, Function): - return self.fn == other.fn and self.lhs.Equals( - other.lhs) and self.rhs.Equals(other.rhs) + result = self.fn == other.fn and self.lhs.Equals(other.lhs) + if self.rhs: + result = result and self.rhs.Equals(other.rhs) + return result return False
Signature · 1 line
2.39.0.314.g84b9a713c41-goog
Ian Rogers
reviewerPATCH v2 2/9 perf jevents metric: Add ability to rewrite metrics in terms of others
In reply to Ian RogersAdd RewriteMetricsInTermsOfOthers that iterates over pairs of names and expressions trying to replace an expression, within the current expression, with its name.
--- tools/perf/pmu-events/metric.py | 73 +++++++++++++++++++++++++++- tools/perf/pmu-events/metric_test.py | 10 ++++ 2 files changed, 81 insertions(+), 2 deletions(-)
tools/perf/pmu-events/metric.py +71 -2
@@ -4,7 +4,7 @@ import ast import decimal import json import re -from typing import Dict, List, Optional, Set, Union +from typing import Dict, List, Optional, Set, Tuple, Union class Expression:
@@ -26,6 +26,9 @@ class Expression: """Returns true when two expressions are the same.""" raise NotImplementedError() + def Substitute(self, name: str, expression: 'Expression') -> 'Expression': + raise NotImplementedError() + def __str__(self) -> str: return self.ToPerfJson()
@@ -186,6 +189,15 @@ class Operator(Expression): other.lhs) and self.rhs.Equals(other.rhs) return False + def Substitute(self, name: str, expression: Expression) -> Expression: + if self.Equals(expression): + return Event(name) + lhs = self.lhs.Substitute(name, expression) + rhs = None + if self.rhs: + rhs = self.rhs.Substitute(name, expression) + return Operator(self.operator, lhs, rhs) + class Select(Expression): """Represents a select ternary in the parse tree."""
@@ -225,6 +237,14 @@ class Select(Expression): other.false_val) and self.true_val.Equals(other.true_val) return False + def Substitute(self, name: str, expression: Expression) -> Expression: + if self.Equals(expression): + return Event(name) + true_val = self.true_val.Substitute(name, expression) + cond = self.cond.Substitute(name, expression) + false_val = self.false_val.Substitute(name, expression) + return Select(true_val, cond, false_val) + class Function(Expression): """A function in an expression like min, max, d_ratio."""
@@ -267,6 +287,15 @@ class Function(Expression): return result return False + def Substitute(self, name: str, expression: Expression) -> Expression: + if self.Equals(expression): + return Event(name) + lhs = self.lhs.Substitute(name, expression) + rhs = None + if self.rhs: + rhs = self.rhs.Substitute(name, expression) + return Function(self.fn, lhs, rhs) + def _FixEscapes(s: str) -> str: s = re.sub(r'([^\\]),', r'\1\\,', s)
@@ -293,6 +322,9 @@ class Event(Expression): def Equals(self, other: Expression) -> bool: return isinstance(other, Event) and self.name == other.name + def Substitute(self, name: str, expression: Expression) -> Expression: + return self + class Constant(Expression): """A constant within the expression tree."""
@@ -317,6 +349,9 @@ class Constant(Expression): def Equals(self, other: Expression) -> bool: return isinstance(other, Constant) and self.value == other.value + def Substitute(self, name: str, expression: Expression) -> Expression: + return self + class Literal(Expression): """A runtime literal within the expression tree."""
@@ -336,6 +371,9 @@ class Literal(Expression): def Equals(self, other: Expression) -> bool: return isinstance(other, Literal) and self.value == other.value + def Substitute(self, name: str, expression: Expression) -> Expression: + return self + def min(lhs: Union[int, float, Expression], rhs: Union[int, float, Expression]) -> Function:
@@ -461,6 +499,7 @@ class MetricGroup: class _RewriteIfExpToSelect(ast.NodeTransformer): + """Transformer to convert if-else nodes to Select expressions.""" def visit_IfExp(self, node): # pylint: disable=invalid-name
@@ -498,7 +537,37 @@ def ParsePerfJson(orig: str) -> Expression: for kw in keywords: py = re.sub(rf'Event\(r"{kw}"\)', kw, py) - parsed = ast.parse(py, mode='eval') + try: + parsed = ast.parse(py, mode='eval') + except SyntaxError as e: + raise SyntaxError(f'Parsing expression:\n{orig}') from e _RewriteIfExpToSelect().visit(parsed) parsed = ast.fix_missing_locations(parsed) return _Constify(eval(compile(parsed, orig, 'eval'))) + + +def RewriteMetricsInTermsOfOthers(metrics: list[Tuple[str, Expression]] + )-> Dict[str, Expression]: + """Shorten metrics by rewriting in terms of others. + + Args: + metrics (list): pairs of metric names and their expressions. + Returns: + Dict: mapping from a metric name to a shortened expression. + """ + updates: Dict[str, Expression] = dict() + for outer_name, outer_expression in metrics: + updated = outer_expression + while True: + for inner_name, inner_expression in metrics: + if inner_name.lower() == outer_name.lower(): + continue + if inner_name in updates: + inner_expression = updates[inner_name] + updated = updated.Substitute(inner_name, inner_expression) + if updated.Equals(outer_expression): + break + if outer_name in updates and updated.Equals(updates[outer_name]): + break + updates[outer_name] = updated + return updates
tools/perf/pmu-events/metric_test.py +10 -0
@@ -2,7 +2,9 @@ import unittest from metric import Constant from metric import Event +from metric import Expression from metric import ParsePerfJson +from metric import RewriteMetricsInTermsOfOthers class TestMetricExpressions(unittest.TestCase):
@@ -153,5 +155,13 @@ class TestMetricExpressions(unittest.TestCase): after = '0 * SLOTS' self.assertEqual(ParsePerfJson(before).Simplify().ToPerfJson(), after) + def test_RewriteMetricsInTermsOfOthers(self): + Expression.__eq__ = lambda e1, e2: e1.Equals(e2) + before = [('m1', ParsePerfJson('a + b + c + d')), + ('m2', ParsePerfJson('a + b + c'))] + after = {'m1': ParsePerfJson('m2 + d')} + self.assertEqual(RewriteMetricsInTermsOfOthers(before), after) + Expression.__eq__ = None + if __name__ == '__main__': unittest.main()
Signature · 1 line
2.39.0.314.g84b9a713c41-goog
Ian Rogers
reviewerPATCH v2 3/9 perf jevents: Rewrite metrics in the same file with each other
In reply to Ian RogersRewrite metrics within the same file in terms of each other. For example, on Power8other_stall_cpi is rewritten from:
"PM_CMPLU_STALL / PM_RUN_INST_CMPL - PM_CMPLU_STALL_BRU_CRU / PM_RUN_INST_CMPL - PM_CMPLU_STALL_FXU / PM_RUN_INST_CMPL - PM_CMPLU_STALL_VSU / PM_RUN_INST_CMPL - PM_CMPLU_STALL_LSU / PM_RUN_INST_CMPL - PM_CMPLU_STALL_NTCG_FLUSH / PM_RUN_INST_CMPL - PM_CMPLU_STALL_NO_NTF / PM_RUN_INST_CMPL" to:
"stall_cpi - bru_cru_stall_cpi - fxu_stall_cpi - vsu_stall_cpi - lsu_stall_cpi - ntcg_flush_cpi - no_ntf_stall_cpi"
Which more closely matches the definition on Power9.
To avoid recomputation decorate the function with a cache.
--- tools/perf/pmu-events/jevents.py | 21 ++++++++++++++++----- 1 file changed, 16 insertions(+), 5 deletions(-)
tools/perf/pmu-events/jevents.py +16 -5
@@ -3,6 +3,7 @@ """Convert directories of JSON events to C code.""" import argparse import csv +from functools import lru_cache import json import metric import os
@@ -337,18 +338,28 @@ class JsonEvent: s = self.build_c_string() return f'{{ { _bcs.offsets[s] } }}, /* {s} */\n' - +@lru_cache(maxsize=None) def read_json_events(path: str, topic: str) -> Sequence[JsonEvent]: """Read json events from the specified file.""" - try: - result = json.load(open(path), object_hook=JsonEvent) + events = json.load(open(path), object_hook=JsonEvent) except BaseException as err: print(f"Exception processing {path}") raise - for event in result: + metrics: list[Tuple[str, metric.Expression]] = [] + for event in events: event.topic = topic - return result + if event.metric_name and '-' not in event.metric_name: + metrics.append((event.metric_name, event.metric_expr)) + updates = metric.RewriteMetricsInTermsOfOthers(metrics) + if updates: + for event in events: + if event.metric_name in updates: + # print(f'Updated {event.metric_name} from\n"{event.metric_expr}"\n' + # f'to\n"{updates[event.metric_name]}"') + event.metric_expr = updates[event.metric_name] + + return events def preprocess_arch_std_files(archpath: str) -> None: """Read in all architecture standard events."""
Signature · 1 line
2.39.0.314.g84b9a713c41-goog
Ian Rogers
reviewerPATCH v2 4/9 perf pmu-events: Separate metric out of pmu_event
In reply to Ian RogersPreviously both events and metrics were encoded in struct pmu_event. Create a new pmu_metric that has the metric related variables and remove these from pmu_event. Add iterators for pmu_metric and use in places that metrics are desired rather than events.
Note, this change removes the setting of evsel's metric_name/expr as these fields are no longer part of struct pmu_event. The metric remains but is no longer implicitly requested when the event is. This impacts a few Intel uncore events, however, as the ScaleUnit is shared by the event and the metric this utility is questionable. Also the MetricNames look broken (contain spaces) in some cases and when trying to use the functionality with '-e' the metrics fail but regular metrics with '-M' work. For example, on SkylakeX '-M' works:
```
$ perf stat -M LLC_MISSES.PCIE_WRITE -a sleep 1
Performance counter stats for 'system wide':
0 UNC_IIO_DATA_REQ_OF_CPU.MEM_WRITE.PART2 # 57896.0 Bytes LLC_MISSES.PCIE_WRITE (49.84%)
7,174 UNC_IIO_DATA_REQ_OF_CPU.MEM_WRITE.PART1 (49.85%)
0 UNC_IIO_DATA_REQ_OF_CPU.MEM_WRITE.PART3 (50.16%)
63 UNC_IIO_DATA_REQ_OF_CPU.MEM_WRITE.PART0 (50.15%)
1.004576381 seconds time elapsed
```whilst the event '-e' version is broken even with --group/-g (fwiw, we should also remove -g [1]):
```
$ perf stat -g -e LLC_MISSES.PCIE_WRITE -g -a sleep 1
Add UNC_IIO_DATA_REQ_OF_CPU.MEM_WRITE.PART2 event to groups to get metric expression for LLC_MISSES.PCIE_WRITE
Add UNC_IIO_DATA_REQ_OF_CPU.MEM_WRITE.PART1 event to groups to get metric expression for LLC_MISSES.PCIE_WRITE
Add UNC_IIO_DATA_REQ_OF_CPU.MEM_WRITE.PART3 event to groups to get metric expression for LLC_MISSES.PCIE_WRITE
Add UNC_IIO_DATA_REQ_OF_CPU.MEM_WRITE.PART0 event to groups to get metric expression for LLC_MISSES.PCIE_WRITE
Add UNC_IIO_DATA_REQ_OF_CPU.MEM_WRITE.PART2 event to groups to get metric expression for LLC_MISSES.PCIE_WRITE
Add UNC_IIO_DATA_REQ_OF_CPU.MEM_WRITE.PART1 event to groups to get metric expression for LLC_MISSES.PCIE_WRITE
Add UNC_IIO_DATA_REQ_OF_CPU.MEM_WRITE.PART3 event to groups to get metric expression for LLC_MISSES.PCIE_WRITE
Add UNC_IIO_DATA_REQ_OF_CPU.MEM_WRITE.PART0 event to groups to get metric expression for LLC_MISSES.PCIE_WRITE
Add UNC_IIO_DATA_REQ_OF_CPU.MEM_WRITE.PART2 event to groups to get metric expression for LLC_MISSES.PCIE_WRITE
Add UNC_IIO_DATA_REQ_OF_CPU.MEM_WRITE.PART1 event to groups to get metric expression for LLC_MISSES.PCIE_WRITE
Add UNC_IIO_DATA_REQ_OF_CPU.MEM_WRITE.PART3 event to groups to get metric expression for LLC_MISSES.PCIE_WRITE
Add UNC_IIO_DATA_REQ_OF_CPU.MEM_WRITE.PART0 event to groups to get metric expression for LLC_MISSES.PCIE_WRITE
Add UNC_IIO_DATA_REQ_OF_CPU.MEM_WRITE.PART2 event to groups to get metric expression for LLC_MISSES.PCIE_WRITE
Add UNC_IIO_DATA_REQ_OF_CPU.MEM_WRITE.PART1 event to groups to get metric expression for LLC_MISSES.PCIE_WRITE
Add UNC_IIO_DATA_REQ_OF_CPU.MEM_WRITE.PART3 event to groups to get metric expression for LLC_MISSES.PCIE_WRITE
Add UNC_IIO_DATA_REQ_OF_CPU.MEM_WRITE.PART0 event to groups to get metric expression for LLC_MISSES.PCIE_WRITE
Add UNC_IIO_DATA_REQ_OF_CPU.MEM_WRITE.PART2 event to groups to get metric expression for LLC_MISSES.PCIE_WRITE
Add UNC_IIO_DATA_REQ_OF_CPU.MEM_WRITE.PART1 event to groups to get metric expression for LLC_MISSES.PCIE_WRITE
Add UNC_IIO_DATA_REQ_OF_CPU.MEM_WRITE.PART3 event to groups to get metric expression for LLC_MISSES.PCIE_WRITE
Add UNC_IIO_DATA_REQ_OF_CPU.MEM_WRITE.PART0 event to groups to get metric expression for LLC_MISSES.PCIE_WRITE
Add UNC_IIO_DATA_REQ_OF_CPU.MEM_WRITE.PART2 event to groups to get metric expression for LLC_MISSES.PCIE_WRITE
Add UNC_IIO_DATA_REQ_OF_CPU.MEM_WRITE.PART1 event to groups to get metric expression for LLC_MISSES.PCIE_WRITE
Add UNC_IIO_DATA_REQ_OF_CPU.MEM_WRITE.PART3 event to groups to get metric expression for LLC_MISSES.PCIE_WRITE
Add UNC_IIO_DATA_REQ_OF_CPU.MEM_WRITE.PART0 event to groups to get metric expression for LLC_MISSES.PCIE_WRITE
Performance counter stats for 'system wide':
27,316 Bytes LLC_MISSES.PCIE_WRITE
1.004505469 seconds time elapsed
```The code also carries warnings where the user is supposed to select events for metrics [2] but given the lack of use of such a feature, let's clean the code and just remove.
With NO_JEVENTS=1 the empty-pmu-events.c is used and the separation to use metrics causes "Parse and process metrics" and "Event expansion for cgroups" to fail as they fail to find necessary metrics. This is fixed in a later change.
[1] https://lore.kernel.org/lkml/[email protected]/
[2] https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/tree/tools/perf/util/stat-shadow.c?id=01b8957b738f42f96a130079bc951b3cc78c5b8a#n425
--- tools/perf/arch/powerpc/util/header.c | 4 +- tools/perf/builtin-list.c | 20 +-- tools/perf/pmu-events/empty-pmu-events.c | 73 ++++++++-- tools/perf/pmu-events/jevents.py | 82 +++++++++++- tools/perf/pmu-events/pmu-events.h | 20 ++- tools/perf/tests/pmu-events.c | 62 +++------ tools/perf/util/metricgroup.c | 161 +++++++++++------------ tools/perf/util/metricgroup.h | 2 +- tools/perf/util/parse-events.c | 2 - tools/perf/util/pmu.c | 35 +---- tools/perf/util/pmu.h | 9 -- tools/perf/util/print-events.c | 32 ++--- tools/perf/util/print-events.h | 3 +- 13 files changed, 266 insertions(+), 239 deletions(-)
tools/perf/arch/powerpc/util/header.c +2 -2
@@ -40,11 +40,11 @@ get_cpuid_str(struct perf_pmu *pmu __maybe_unused)return bufp; } -int arch_get_runtimeparam (const struct pmu_event *pe) +int arch_get_runtimeparam (const struct pmu_metric *pm) {int count;char path[PATH_MAX ] ="/devices/hv_24x7/interface/" ; -atoi (pe->aggr_mode) == PerChip ?strcat (path,"sockets" ) :strcat (path,"coresperchip" ); +atoi (pm->aggr_mode) == PerChip ?strcat (path,"sockets" ) :strcat (path,"coresperchip" );return sysfs__read_int (path, &count) <0 ?1 : count; }
tools/perf/builtin-list.c +2 -18
@@ -99,8 +99,7 @@ static void default_print_event(void *ps, const char *pmu_name, const char *topiconst char *scale_unit__maybe_unused , bool deprecated, const char *event_type_desc, const char *desc, const char *long_desc, - const char *encoding_desc, - const char *metric_name, const char *metric_expr) + const char *encoding_desc) { struct print_state *print_state = ps;int pos;
@@ -159,10 +158,6 @@ static void default_print_event(void *ps, const char *pmu_name, const char *topiif (print_state->detailed && encoding_desc) {printf ("%*s" ,8 ,"" );wordwrap (encoding_desc,8 ,pager_get_columns (),0 ); -if (metric_name) -printf (" MetricName: %s" , metric_name); -if (metric_expr) -printf (" MetricExpr: %s" , metric_expr);putchar ('\n' ); } }
@@ -308,8 +303,7 @@ static void json_print_event(void *ps, const char *pmu_name, const char *topic,const char *scale_unit, bool deprecated, const char *event_type_desc, const char *desc, const char *long_desc, - const char *encoding_desc, - const char *metric_name, const char *metric_expr) + const char *encoding_desc) { struct json_print_state *print_state = ps;bool need_sep = false;
@@ -366,16 +360,6 @@ static void json_print_event(void *ps, const char *pmu_name, const char *topic, encoding_desc); need_sep = true; } -if (metric_name) { -fix_escape_printf (&buf,"%s\t\"MetricName\": \"%S\"" , need_sep ?",\n" :"" , - metric_name); - need_sep = true; - } -if (metric_expr) { -fix_escape_printf (&buf,"%s\t\"MetricExpr\": \"%S\"" , need_sep ?",\n" :"" , - metric_expr); - need_sep = true; - }printf ("%s}" , need_sep ?"\n" :"" );strbuf_release (&buf); }
tools/perf/pmu-events/empty-pmu-events.c +60 -13
146 lines not on this page · raw message
tools/perf/pmu-events/jevents.py +76 -6
122 lines not on this page · raw message
tools/perf/pmu-events/pmu-events.h +18 -2
@@ -23,11 +23,19 @@ struct pmu_event {const char *unit;const char *perpkg;const char *aggr_mode; -const char *metric_expr; +const char *deprecated; +}; + +struct pmu_metric {const char *metric_name;const char *metric_group; -const char *deprecated; +const char *metric_expr; +const char *unit; +const char *compat; +const char *aggr_mode;const char *metric_constraint; +const char *desc; +const char *long_desc; };struct pmu_events_table ;
@@ -36,14 +44,22 @@ typedef int (*pmu_event_iter_fn)(const struct pmu_event *pe,const struct pmu_events_table *table, void *data); +typedef int (*pmu_metric_iter_fn )(const struct pmu_metric *pm, +const struct pmu_events_table *table, +void *data); +int pmu_events_table_for_each_event (const struct pmu_events_table *table,pmu_event_iter_fn fn,void *data); +int pmu_events_table_for_each_metric (const struct pmu_events_table *table,pmu_metric_iter_fn fn, +void *data);const struct pmu_events_table *perf_pmu__find_table (struct perf_pmu *pmu);const struct pmu_events_table *find_core_events_table (const char *arch,const char *cpuid);int pmu_for_each_core_event (pmu_event_iter_fn fn,void *data); +int pmu_for_each_core_metric (pmu_metric_iter_fn fn,void *data);const struct pmu_events_table *find_sys_events_table (const char *name);int pmu_for_each_sys_event (pmu_event_iter_fn fn,void *data); +int pmu_for_each_sys_metric (pmu_metric_iter_fn fn,void *data);#endif
tools/perf/tests/pmu-events.c +16 -46
143 lines not on this page · raw message
tools/perf/util/metricgroup.c +79 -82
456 lines not on this page · raw message
tools/perf/util/metricgroup.h +1 -1
@@ -81,7 +81,7 @@ int metricgroup__parse_groups_test(struct evlist *evlist,void metricgroup__print (const struct print_callbacks *print_cb,void *print_state);bool metricgroup__has_metric (const char *metric); -int arch_get_runtimeparam (const struct pmu_event *pe__maybe_unused ); +int arch_get_runtimeparam (const struct pmu_metric *pm);void metricgroup__rblist_exit (struct rblist *metric_events); int metricgroup__copy_metric_events(struct evlist *evlist, struct cgroup *cgrp,
tools/perf/util/parse-events.c +0 -2
@@ -1570,8 +1570,6 @@ int parse_events_add_pmu(struct parse_events_state *parse_state, evsel->scale = info.scale; evsel->per_pkg = info.per_pkg; evsel->snapshot = info.snapshot; - evsel->metric_expr = info.metric_expr; - evsel->metric_name = info.metric_name;return 0 ; }
tools/perf/util/pmu.c +3 -32
@@ -283,10 +283,6 @@ static void perf_pmu_update_alias(struct perf_pmu_alias *old,perf_pmu_assign_str (old->name,"long_desc" , &old->long_desc, &newalias->long_desc);perf_pmu_assign_str (old->name,"topic" , &old->topic, &newalias->topic); -perf_pmu_assign_str (old->name,"metric_expr" , &old->metric_expr, - &newalias->metric_expr); -perf_pmu_assign_str (old->name,"metric_name" , &old->metric_name, - &newalias->metric_name);perf_pmu_assign_str (old->name,"value" , &old->str, &newalias->str); old->scale = newalias->scale; old->per_pkg = newalias->per_pkg;
@@ -302,8 +298,6 @@ void perf_pmu_free_alias(struct perf_pmu_alias *newalias)zfree (&newalias->long_desc);zfree (&newalias->topic);zfree (&newalias->str); -zfree (&newalias->metric_expr); -zfree (&newalias->metric_name);zfree (&newalias->pmu_name);parse_events_terms__purge (&newalias->terms);free (newalias);
@@ -340,16 +334,13 @@ static int __perf_pmu__new_alias(struct list_head *list, char *dir, char *name,int num;char newval[256 ];char *long_desc =NULL , *topic =NULL , *unit =NULL , *perpkg =NULL , - *metric_expr =NULL , *metric_name =NULL , *deprecated =NULL , - *pmu_name =NULL ; + *deprecated =NULL , *pmu_name =NULL ;if (pe) { long_desc = (char *)pe->long_desc; topic = (char *)pe->topic; unit = (char *)pe->unit; perpkg = (char *)pe->perpkg; - metric_expr = (char *)pe->metric_expr; - metric_name = (char *)pe->metric_name; deprecated = (char *)pe->deprecated; pmu_name = (char *)pe->pmu; }
@@ -404,8 +395,6 @@ static int __perf_pmu__new_alias(struct list_head *list, char *dir, char *name,perf_pmu__parse_snapshot (alias, dir, name); } - alias->metric_expr = metric_expr ?strdup (metric_expr) :NULL ; - alias->metric_name = metric_name ?strdup (metric_name):NULL ; alias->desc = desc ?strdup (desc) :NULL ; alias->long_desc = long_desc ?strdup (long_desc) : desc ?strdup (desc) :NULL ;
@@ -822,9 +811,6 @@ static int pmu_add_cpu_aliases_map_callback(const struct pmu_event *pe,struct pmu_add_cpu_aliases_map_data *data = vdata;const char *pname = pe->pmu ? pe->pmu : data->cpu_name; -if (!pe->name) -return 0 ; -if (data->pmu->is_uncore &&pmu_uncore_alias_match (pname, data->name)) goto new_alias;
@@ -879,12 +865,6 @@ static int pmu_add_sys_aliases_iter_fn(const struct pmu_event *pe,struct pmu_sys_event_iter_data *idata = data;struct perf_pmu *pmu = idata->pmu; -if (!pe->name) { -if (pe->metric_group || pe->metric_name) -return 0 ; -return -EINVAL ; - } -if (!pe->compat || !pe->pmu)return 0 ;
@@ -1469,8 +1449,6 @@ int perf_pmu__check_alias(struct perf_pmu *pmu, struct list_head *head_terms, info->unit =NULL ; info->scale =0.0 ; info->snapshot = false; - info->metric_expr =NULL ; - info->metric_name =NULL ;list_for_each_entry_safe (term, h, head_terms, list) { alias =pmu_find_alias (pmu, term);
@@ -1486,8 +1464,6 @@ int perf_pmu__check_alias(struct perf_pmu *pmu, struct list_head *head_terms,if (alias->per_pkg) info->per_pkg = true; - info->metric_expr = alias->metric_expr; - info->metric_name = alias->metric_name;list_del_init (&term->list);parse_events_term__delete (term);
@@ -1703,8 +1679,7 @@ void print_pmu_events(const struct print_callbacks *print_cb, void *print_state)for (j =0 ; j < len; j++) {const char *name, *alias =NULL , *scale_unit =NULL , *desc =NULL , *long_desc =NULL , - *encoding_desc =NULL , *topic =NULL , - *metric_name =NULL , *metric_expr =NULL ; + *encoding_desc =NULL , *topic =NULL ;bool deprecated = false;size_t buf_used;
@@ -1742,8 +1717,6 @@ void print_pmu_events(const struct print_callbacks *print_cb, void *print_state) buf_used +=snprintf (buf + buf_used,sizeof (buf) - buf_used,"%s/%s/" , aliases[j].pmu->name, aliases[j].event->str) +1 ; - metric_name = aliases[j].event->metric_name; - metric_expr = aliases[j].event->metric_expr; deprecated = aliases[j].event->deprecated; } print_cb->print_event(print_state,
@@ -1756,9 +1729,7 @@ void print_pmu_events(const struct print_callbacks *print_cb, void *print_state)"Kernel PMU event" , desc, long_desc, - encoding_desc, - metric_name, - metric_expr); + encoding_desc); }if (printed &&pager_in_use ())printf ("\n" );
tools/perf/util/pmu.h +0 -9
@@ -133,8 +133,6 @@ extern struct perf_pmu perf_pmu__fake;struct perf_pmu_info {const char *unit; -const char *metric_expr; -const char *metric_name;double scale;bool per_pkg;bool snapshot;
@@ -188,13 +186,6 @@ struct perf_pmu_alias {* default. */ bool deprecated; -/** -* @metric_expr: A metric expression associated with an event. Doing -* this makes little sense due to scale and unit applying to both. -*/ -char *metric_expr; -/** @metric_name: A name for the metric. unit applying to both. */ -char *metric_name;/** @pmu_name: The name copied from struct perf_pmu. */ char *pmu_name; };
tools/perf/util/print-events.c +8 -24
@@ -101,9 +101,7 @@ void print_tracepoint_events(const struct print_callbacks *print_cb, void *print"Tracepoint event" ,/*desc=*/ NULL ,/*long_desc=*/ NULL , -/*encoding_desc=*/ NULL , -/*metric_name=*/ NULL , - /*metric_expr=*/NULL); + /*encoding_desc=*/NULL); }free (dir_path);free (evt_namelist);
@@ -195,9 +193,7 @@ void print_sdt_events(const struct print_callbacks *print_cb, void *print_state)"SDT event" ,/*desc=*/ NULL ,/*long_desc=*/ NULL , -/*encoding_desc=*/ NULL , -/*metric_name=*/ NULL , - /*metric_expr=*/NULL); + /*encoding_desc=*/NULL);free (evt_name); }
@@ -255,9 +251,7 @@ int print_hwcache_events(const struct print_callbacks *print_cb, void *print_sta event_type_descriptors[PERF_TYPE_HW_CACHE ],/*desc=*/ NULL ,/*long_desc=*/ NULL , -/*encoding_desc=*/ NULL , -/*metric_name=*/ NULL , - /*metric_expr=*/NULL); + /*encoding_desc=*/NULL); }strlist__delete (evt_name_list);return 0 ;
@@ -277,9 +271,7 @@ void print_tool_events(const struct print_callbacks *print_cb, void *print_state"Tool event" ,/*desc=*/ NULL ,/*long_desc=*/ NULL , -/*encoding_desc=*/ NULL , -/*metric_name=*/ NULL , - /*metric_expr=*/NULL); + /*encoding_desc=*/NULL); } }
@@ -331,9 +323,7 @@ void print_symbol_events(const struct print_callbacks *print_cb, void *print_sta event_type_descriptors[type],/*desc=*/ NULL ,/*long_desc=*/ NULL , -/*encoding_desc=*/ NULL , -/*metric_name=*/ NULL , - /*metric_expr=*/NULL); + /*encoding_desc=*/NULL); }strlist__delete (evt_name_list); }
@@ -364,9 +354,7 @@ void print_events(const struct print_callbacks *print_cb, void *print_state) event_type_descriptors[PERF_TYPE_RAW ],/*desc=*/ NULL ,/*long_desc=*/ NULL , -/*encoding_desc=*/ NULL , -/*metric_name=*/ NULL , - /*metric_expr=*/NULL); + /*encoding_desc=*/NULL); print_cb->print_event(print_state, /*topic=*/NULL,
@@ -378,9 +366,7 @@ void print_events(const struct print_callbacks *print_cb, void *print_state) event_type_descriptors[PERF_TYPE_RAW ],"(see 'man perf-list' on how to encode it)" ,/*long_desc=*/ NULL , -/*encoding_desc=*/ NULL , -/*metric_name=*/ NULL , - /*metric_expr=*/NULL); + /*encoding_desc=*/NULL); print_cb->print_event(print_state, /*topic=*/NULL,
@@ -392,9 +378,7 @@ void print_events(const struct print_callbacks *print_cb, void *print_state) event_type_descriptors[PERF_TYPE_BREAKPOINT ],/*desc=*/ NULL ,/*long_desc=*/ NULL , -/*encoding_desc=*/ NULL , -/*metric_name=*/ NULL , - /*metric_expr=*/NULL); + /*encoding_desc=*/NULL);print_tracepoint_events (print_cb, print_state);
tools/perf/util/print-events.h +1 -2
@@ -16,8 +16,7 @@ struct print_callbacks {const char *scale_unit, bool deprecated, const char *event_type_desc, const char *desc, const char *long_desc, - const char *encoding_desc, - const char *metric_name, const char *metric_expr); + const char *encoding_desc); void (*print_metric)(void *print_state, const char *group, const char *name,
Signature · 1 line
2.39.0.314.g84b9a713c41-goog
John Garry
reviewerRe: [PATCH v2 4/9] perf pmu-events: Separate metric out of pmu_event
Quoted 6 lines · Ian Rogers
Previously both events and metrics were encoded in struct
pmu_event. Create a newpmu_metricthat has the metric related variables and remove these frompmu_event. Add iterators forpmu_metricand use in places that metrics are desired rather than events.
Quoted 9 lines
Note, this change removes the setting of evsel's metric_name/expr as these fields are no longer part of struct
pmu_event. The metric remains but is no longer implicitly requested when the event is. This impacts a few Intel uncore events, however, as the ScaleUnit is shared by the event and the metric this utility is questionable. Also the MetricNames look broken (contain spaces) in some cases and when trying to use the functionality with '-e' the metrics fail but regular metrics with '-M' work. For example, on SkylakeX '-M' works:
I like this change. It's quite large for a single patch. Just some sparse comments below.
BTW, it there a better name for metric struct variable than "pm"? To me and many other people, pm is power management.
Quoted 239 lines
``` $ perf stat -M LLC_MISSES.PCIE_WRITE -a sleep 1 Performance counter stats for 'system wide': 0 UNC_IIO_DATA_REQ_OF_CPU.MEM_WRITE.PART2 # 57896.0 Bytes LLC_MISSES.PCIE_WRITE (49.84%) 7,174 UNC_IIO_DATA_REQ_OF_CPU.MEM_WRITE.PART1 (49.85%) 0 UNC_IIO_DATA_REQ_OF_CPU.MEM_WRITE.PART3 (50.16%) 63 UNC_IIO_DATA_REQ_OF_CPU.MEM_WRITE.PART0 (50.15%) 1.004576381 seconds time elapsed ```whilst the event '-e' version is broken even with --group/-g (fwiw, we should also remove -g [1]):
``` $ perf stat -g -e LLC_MISSES.PCIE_WRITE -g -a sleep 1 Add UNC_IIO_DATA_REQ_OF_CPU.MEM_WRITE.PART2 event to groups to get metric expression for LLC_MISSES.PCIE_WRITE Add UNC_IIO_DATA_REQ_OF_CPU.MEM_WRITE.PART1 event to groups to get metric expression for LLC_MISSES.PCIE_WRITE Add UNC_IIO_DATA_REQ_OF_CPU.MEM_WRITE.PART3 event to groups to get metric expression for LLC_MISSES.PCIE_WRITE Add UNC_IIO_DATA_REQ_OF_CPU.MEM_WRITE.PART0 event to groups to get metric expression for LLC_MISSES.PCIE_WRITE Add UNC_IIO_DATA_REQ_OF_CPU.MEM_WRITE.PART2 event to groups to get metric expression for LLC_MISSES.PCIE_WRITE Add UNC_IIO_DATA_REQ_OF_CPU.MEM_WRITE.PART1 event to groups to get metric expression for LLC_MISSES.PCIE_WRITE Add UNC_IIO_DATA_REQ_OF_CPU.MEM_WRITE.PART3 event to groups to get metric expression for LLC_MISSES.PCIE_WRITE Add UNC_IIO_DATA_REQ_OF_CPU.MEM_WRITE.PART0 event to groups to get metric expression for LLC_MISSES.PCIE_WRITE Add UNC_IIO_DATA_REQ_OF_CPU.MEM_WRITE.PART2 event to groups to get metric expression for LLC_MISSES.PCIE_WRITE Add UNC_IIO_DATA_REQ_OF_CPU.MEM_WRITE.PART1 event to groups to get metric expression for LLC_MISSES.PCIE_WRITE Add UNC_IIO_DATA_REQ_OF_CPU.MEM_WRITE.PART3 event to groups to get metric expression for LLC_MISSES.PCIE_WRITE Add UNC_IIO_DATA_REQ_OF_CPU.MEM_WRITE.PART0 event to groups to get metric expression for LLC_MISSES.PCIE_WRITE Add UNC_IIO_DATA_REQ_OF_CPU.MEM_WRITE.PART2 event to groups to get metric expression for LLC_MISSES.PCIE_WRITE Add UNC_IIO_DATA_REQ_OF_CPU.MEM_WRITE.PART1 event to groups to get metric expression for LLC_MISSES.PCIE_WRITE Add UNC_IIO_DATA_REQ_OF_CPU.MEM_WRITE.PART3 event to groups to get metric expression for LLC_MISSES.PCIE_WRITE Add UNC_IIO_DATA_REQ_OF_CPU.MEM_WRITE.PART0 event to groups to get metric expression for LLC_MISSES.PCIE_WRITE Add UNC_IIO_DATA_REQ_OF_CPU.MEM_WRITE.PART2 event to groups to get metric expression for LLC_MISSES.PCIE_WRITE Add UNC_IIO_DATA_REQ_OF_CPU.MEM_WRITE.PART1 event to groups to get metric expression for LLC_MISSES.PCIE_WRITE Add UNC_IIO_DATA_REQ_OF_CPU.MEM_WRITE.PART3 event to groups to get metric expression for LLC_MISSES.PCIE_WRITE Add UNC_IIO_DATA_REQ_OF_CPU.MEM_WRITE.PART0 event to groups to get metric expression for LLC_MISSES.PCIE_WRITE Add UNC_IIO_DATA_REQ_OF_CPU.MEM_WRITE.PART2 event to groups to get metric expression for LLC_MISSES.PCIE_WRITE Add UNC_IIO_DATA_REQ_OF_CPU.MEM_WRITE.PART1 event to groups to get metric expression for LLC_MISSES.PCIE_WRITE Add UNC_IIO_DATA_REQ_OF_CPU.MEM_WRITE.PART3 event to groups to get metric expression for LLC_MISSES.PCIE_WRITE Add UNC_IIO_DATA_REQ_OF_CPU.MEM_WRITE.PART0 event to groups to get metric expression for LLC_MISSES.PCIE_WRITE Performance counter stats for 'system wide': 27,316 Bytes LLC_MISSES.PCIE_WRITE 1.004505469 seconds time elapsed ```The code also carries warnings where the user is supposed to select events for metrics [2] but given the lack of use of such a feature, let's clean the code and just remove.
With NO_JEVENTS=1 the
empty-pmu-events.cis used and the separation to use metrics causes "Parse and process metrics" and "Event expansion for cgroups" to fail as they fail to find necessary metrics. This is fixed in a later change.[1] https://urldefense.com/v3/__https://lore.kernel.org/lkml/[email protected]/__;!!ACWV5N9M2RV99hQ!MTNgf7sIIl2VCxBzo-biSeul-ZQkuU73pT6gBbsb6y0AqsVlgGCcmuigDyevtrgmpHDN1QkEJwbJYvREu3Q$
[2] https://urldefense.com/v3/__https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/tree/tools/perf/util/stat-shadow.c?id=01b8957b738f42f96a130079bc951b3cc78c5b8a*n425__;Iw!!ACWV5N9M2RV99hQ!MTNgf7sIIl2VCxBzo-biSeul-ZQkuU73pT6gBbsb6y0AqsVlgGCcmuigDyevtrgmpHDN1QkEJwbJ8H0tppQ$Signed-off-by Ian Rogers <[email protected]>--- tools/perf/arch/powerpc/util/header.c | 4 +- tools/perf/builtin-list.c | 20 +-- tools/perf/pmu-events/empty-pmu-events.c | 73 ++++++++-- tools/perf/pmu-events/jevents.py | 82 +++++++++++- tools/perf/pmu-events/pmu-events.h | 20 ++- tools/perf/tests/pmu-events.c | 62 +++------ tools/perf/util/metricgroup.c | 161 +++++++++++------------ tools/perf/util/metricgroup.h | 2 +- tools/perf/util/parse-events.c | 2 - tools/perf/util/pmu.c | 35 +---- tools/perf/util/pmu.h | 9 -- tools/perf/util/print-events.c | 32 ++--- tools/perf/util/print-events.h | 3 +- 13 files changed, 266 insertions(+), 239 deletions(-)tools/perf/arch/powerpc/util/header.c +2
-2@@ -40,11 +40,11 @@ get_cpuid_str(struct perf_pmu *pmu __maybe_unused)return bufp; } -int arch_get_runtimeparam (const struct pmu_event *pe) +int arch_get_runtimeparam (const struct pmu_metric *pm) {int count;char path[PATH_MAX ] ="/devices/hv_24x7/interface/" ; -atoi (pe->aggr_mode) == PerChip ?strcat (path,"sockets" ) :strcat (path,"coresperchip" ); +atoi (pm->aggr_mode) == PerChip ?strcat (path,"sockets" ) :strcat (path,"coresperchip" );return sysfs__read_int (path, &count) <0 ?1 : count; }tools/perf/builtin-list.c +2
-18@@ -99,8 +99,7 @@ static void default_print_event(void *ps, const char *pmu_name, const char *topiconst char *scale_unit__maybe_unused , bool deprecated, const char *event_type_desc, const char *desc, const char *long_desc, - const char *encoding_desc, - const char *metric_name, const char *metric_expr) + const char *encoding_desc) { struct print_state *print_state = ps;int pos;@@ -159,10 +158,6 @@ static void default_print_event(void *ps, const char *pmu_name, const char *topiif (print_state->detailed && encoding_desc) {printf ("%*s" ,8 ,"" );wordwrap (encoding_desc,8 ,pager_get_columns (),0 ); -if (metric_name) -printf (" MetricName: %s" , metric_name); -if (metric_expr) -printf (" MetricExpr: %s" , metric_expr);putchar ('\n' ); } }@@ -308,8 +303,7 @@ static void json_print_event(void *ps, const char *pmu_name, const char *topic,const char *scale_unit, bool deprecated, const char *event_type_desc, const char *desc, const char *long_desc, - const char *encoding_desc, - const char *metric_name, const char *metric_expr) + const char *encoding_desc) { struct json_print_state *print_state = ps;bool need_sep = false;@@ -366,16 +360,6 @@ static void json_print_event(void *ps, const char *pmu_name, const char *topic, encoding_desc); need_sep = true; } -if (metric_name) { -fix_escape_printf (&buf,"%s\t\"MetricName\": \"%S\"" , need_sep ?",\n" :"" , - metric_name); - need_sep = true; - } -if (metric_expr) { -fix_escape_printf (&buf,"%s\t\"MetricExpr\": \"%S\"" , need_sep ?",\n" :"" , - metric_expr); - need_sep = true; - }printf ("%s}" , need_sep ?"\n" :"" );strbuf_release (&buf); }tools/perf/pmu-events/empty-pmu-events.c +23
-10@@ -11,7 +11,7 @@#include <string.h> #include <stddef.h> -static const struct pmu_event pme_test_soc_cpu[] = { +static const struct pmu_event pmu_events__test_soc_cpu[] = { { .name ="l3_cache_rd" , .event ="event=0x40" ,@@ -105,6 +105,14 @@ static const struct pmu_event pme_test_soc_cpu[] = { .desc = "L2 BTB Correction", .topic = "branch", }, + { + .name = 0, + .event = 0, + .desc = 0, + }, +}; + +static const struct pmu_metric pmu_metrics__test_soc_cpu[] = { { .metric_expr ="1 / IPC" , .metric_name ="CPI" ,@@ -170,9 +178,8 @@ static const struct pmu_event pme_test_soc_cpu[] = { .metric_name = "L1D_Cache_Fill_BW", }, { - .name = 0, - .event = 0, - .desc = 0, + .metric_expr = 0, + .metric_name = 0, }, };@@ -181,6 +188,11 @@ struct pmu_events_table {const struct pmu_event *entries; }; +/* Struct used to make the PMU metric table implementation opaque to callers. */ +struct pmu_metrics_table { +const struct pmu_metric *entries; +}; +/* * Map a CPU to its table of PMU events. The CPU is identified by the * cpuid field, which is an arch-specific identifier for the CPU. @@ -192,7 +204,8 @@ struct pmu_events_table {struct pmu_events_map {const char *arch;const char *cpuid; -const struct pmu_events_table table; +const struct pmu_events_table event_table; +const struct pmu_metrics_table metric_table; };/* @@ -203,12 +216,14 @@ static const struct pmu_events_map pmu_events_map[] = { { .arch = "testarch", .cpuid = "testcpu", - .table = { pme_test_soc_cpu }, + .event_table = { pmu_events__test_soc_cpu }, + .metric_table = { pmu_metrics__test_soc_cpu }, }, { .arch = 0, .cpuid = 0, - .table = { 0 }, + .event_table = { 0 }, + .metric_table = { 0 }, }, };@@ -254,9 +269,7 @@ static const struct pmu_sys_events pmu_sys_event_tables[] = {int pmu_events_table_for_each_event (const struct pmu_events_table *table,pmu_event_iter_fn fn,void *data) { -for (const struct pmu_event *pe = &table->entries[0 ]; - pe->name || pe->metric_group || pe->metric_name; - pe++) { +for (const struct pmu_event *pe = &table->entries[0 ]; pe->name; pe++) {int ret =fn (pe, table, data);if (ret)
tools/perf/pmu-events/empty-pmu-events.c int pmu_events_table_for_each_event(const struct pmu_events_table *table, pmu_ev · line 284
Quoted 7 lines
tools/perf/pmu-events/empty-pmu-events.c +4
-0@@ -265,6 +278,21 @@ int pmu_events_table_for_each_event(const struct pmu_events_table *table, pmu_evreturn 0 ; } +int pmu_events_table_for_each_metric (const struct pmu_events_table *etable,pmu_metric_iter_fn fn, +void *data) +{ +struct pmu_metrics_table *table = (struct pmu_metrics_table *)etable;
This is odd. Would the change be too big if we combined patches to avoid casting like this? I assume that this is going away later...
tools/perf/pmu-events/empty-pmu-events.c int pmu_events_table_for_each_event(const struct pmu_events_table *table, pmu_ev · line 286
Quoted 2 lines
tools/perf/pmu-events/empty-pmu-events.c +2
-0+ + for (const struct pmu_metric *pm = &table->entries[0]; pm->metric_group || pm->metric_name;
Is it valid to only have metric_group or metric_name set?
Quoted 38 lines
tools/perf/pmu-events/empty-pmu-events.c +14
-3+ pm++) { +int ret =fn (pm, etable, data); + +if (ret) +return ret; + } +return 0 ; +} +const struct pmu_events_table *perf_pmu__find_table (struct perf_pmu *pmu) {const struct pmu_events_table *table =NULL ;@@ -285,7 +313,7 @@ const struct pmu_events_table *perf_pmu__find_table(struct perf_pmu *pmu)break ;if (!strcmp_cpuid_str (map->cpuid, cpuid)) { - table = &map->table; + table = &map->event_table;break ; } }@@ -299,7 +327,7 @@ const struct pmu_events_table *find_core_events_table(const char *arch, const ch tables->arch; tables++) {if (!strcmp (tables->arch, arch) && !strcmp_cpuid_str (tables->cpuid, cpuid)) -return &tables->table; +return &tables->event_table; }return NULL ; }@@ -309,7 +337,21 @@ int pmu_for_each_core_event(pmu_event_iter_fn fn, void *data)for (const struct pmu_events_map *tables = &pmu_events_map[0 ]; tables->arch; tables++) { -int ret =pmu_events_table_for_each_event (&tables->table, fn, data); +int ret =pmu_events_table_for_each_event (&tables->event_table, fn, data); + +if (ret)
tools/perf/pmu-events/empty-pmu-events.c int pmu_for_each_core_event(pmu_event_iter_fn fn, void *data) · line 352
Quoted 10 lines
tools/perf/pmu-events/empty-pmu-events.c +10
-0+return ret; + } +return 0 ; +} + +int pmu_for_each_core_metric (pmu_metric_iter_fn fn,void *data) +{ +for (const struct pmu_events_map *tables = &pmu_events_map[0 ]; + tables->arch; + tables++) {
again, you can combine here
Quoted 151 lines
tools/perf/pmu-events/empty-pmu-events.c +7
-0+int ret =pmu_events_table_for_each_metric ( + (const struct pmu_events_table *)&tables->metric_table, fn, data);if (ret)return ret;@@ -340,3 +382,8 @@ int pmu_for_each_sys_event(pmu_event_iter_fn fn, void *data) }return 0 ; } + +int pmu_for_each_sys_metric (pmu_metric_iter_fn fn__maybe_unused ,void *data__maybe_unused ) +{ +return 0 ; +}tools/perf/pmu-events/jevents.py +76
-6@@ -37,6 +37,11 @@ _json_event_attributes = [ 'metric_constraint', 'metric_expr', 'long_desc' ] +# Attributes that are in pmu_metric rather than pmu_event. +_json_metric_attributes = [ + 'metric_name', 'metric_group', 'metric_constraint', 'metric_expr', 'desc', + 'long_desc', 'unit', 'compat', 'aggr_mode' +] def removesuffix(s: str, suffix: str) -> str: """Remove the suffix from a string@@ -562,12 +567,15 @@ static const struct pmu_sys_events pmu_sys_event_tables[] = { \t}, }; -static void decompress(int offset, struct pmu_event *pe) +static void decompress_event(int offset, struct pmu_event *pe) { \tconst char *p = &big_c_string[offset]; """) for attr in _json_event_attributes: - _args.output_file.write(f""" + if attr in _json_metric_attributes and 'metric_' in attr: + _args.output_file.write(f'\n\t/* Skip {attr} */\n') + else: + _args.output_file.write(f""" \tpe->{attr} = (*p == '\\0' ? NULL : p); """) if attr == _json_event_attributes[-1]:@@ -575,6 +583,22 @@ static void decompress(int offset, struct pmu_event *pe) _args.output_file.write('\twhile (*p++);') _args.output_file.write("""} +static void decompress_metric(int offset, struct pmu_metric *pm) +{ +\tconst char *p = &big_c_string[offset]; +""") + for attr in _json_event_attributes: + if attr in _json_metric_attributes: + _args.output_file.write(f""" +\tpm->{attr} = (*p == '\\0' ? NULL : p); +""") + else: + _args.output_file.write(f'\n\t/* Skip {attr} */\n') + if attr == _json_event_attributes[-1]: + continue + _args.output_file.write('\twhile (*p++);') + _args.output_file.write("""} + int pmu_events_table_for_each_event(const struct pmu_events_table *table, pmu_event_iter_fn fn, void *data)@@ -583,10 +607,30 @@ int pmu_events_table_for_each_event(const struct pmu_events_table *table, struct pmu_event pe; int ret; - decompress(table->entries[i].offset, &pe); - ret = fn(&pe, table, data); - if (ret) - return ret; + decompress_event(table->entries[i].offset, &pe); + if (pe.name) { + ret = fn(&pe, table, data); + if (ret) + return ret; + } + } + return 0; +} + +int pmu_events_table_for_each_metric(const struct pmu_events_table *table, + pmu_metric_iter_fn fn, + void *data) +{ + for (size_t i = 0; i < table->length; i++) { + struct pmu_metric pm; + int ret; + + decompress_metric(table->entries[i].offset, &pm); + if (pm.metric_name) { + ret = fn(&pm, table, data); + if (ret) + return ret; + } } return 0; }@@ -642,6 +686,19 @@ int pmu_for_each_core_event(pmu_event_iter_fn fn, void *data) return 0; } +int pmu_for_each_core_metric(pmu_metric_iter_fn fn, void *data) +{ + for (const struct pmu_events_map *tables = &pmu_events_map[0]; + tables->arch; + tables++) { + int ret = pmu_events_table_for_each_metric(&tables->table, fn, data); + + if (ret) + return ret; + } + return 0; +} + const struct pmu_events_table *find_sys_events_table(const char *name) { for (const struct pmu_sys_events *tables = &pmu_sys_event_tables[0];@@ -665,6 +722,19 @@ int pmu_for_each_sys_event(pmu_event_iter_fn fn, void *data) } return 0; } + +int pmu_for_each_sys_metric(pmu_metric_iter_fn fn, void *data) +{ + for (const struct pmu_sys_events *tables = &pmu_sys_event_tables[0]; + tables->name; + tables++) { + int ret = pmu_events_table_for_each_metric(&tables->table, fn, data); + + if (ret) + return ret; + } + return 0; +} """)tools/perf/pmu-events/pmu-events.h +0
-0
tools/perf/pmu-events/pmu-events.h struct pmu_event { · line 29
Quoted 8 lines
tools/perf/pmu-events/pmu-events.h +4
-1@@ -23,11 +23,19 @@ struct pmu_event {const char *unit;const char *perpkg;const char *aggr_mode; -const char *metric_expr; +const char *deprecated; +}; + +struct pmu_metric {
In terms of formatting patches, it could have been better to first introdue struct pmu_metric being same as struct pmu_event, and make the changes of /s/struct pmu_event pe;/struct pmu_event pm;/, and then secondly the meaningful change to get rid of the inappropiate members in struct pmu_metric and struct pmu_event. It's just harder to review it all together.
tools/perf/pmu-events/pmu-events.h struct pmu_event { · line 32
Quoted 4 lines
tools/perf/pmu-events/pmu-events.h +1
-1const char *metric_name;const char *metric_group; -const char *deprecated; +const char *metric_expr;
I would prefer the shorter name "expr" (as opposed to "metric_expr") as we know it's a metric struct already. But that does not seem to be coding style, so ...
Quoted 37 lines
tools/perf/pmu-events/pmu-events.h +13
-0+const char *unit; +const char *compat; +const char *aggr_mode;const char *metric_constraint; +const char *desc; +const char *long_desc; };struct pmu_events_table ;@@ -36,14 +44,22 @@ typedef int (*pmu_event_iter_fn)(const struct pmu_event *pe,const struct pmu_events_table *table, void *data); +typedef int (*pmu_metric_iter_fn )(const struct pmu_metric *pm, +const struct pmu_events_table *table, +void *data); +int pmu_events_table_for_each_event (const struct pmu_events_table *table,pmu_event_iter_fn fn,void *data); +int pmu_events_table_for_each_metric (const struct pmu_events_table *table,pmu_metric_iter_fn fn, +void *data);const struct pmu_events_table *perf_pmu__find_table (struct perf_pmu *pmu);const struct pmu_events_table *find_core_events_table (const char *arch,const char *cpuid);int pmu_for_each_core_event (pmu_event_iter_fn fn,void *data); +int pmu_for_each_core_metric (pmu_metric_iter_fn fn,void *data);const struct pmu_events_table *find_sys_events_table (const char *name);int pmu_for_each_sys_event (pmu_event_iter_fn fn,void *data); +int pmu_for_each_sys_metric (pmu_metric_iter_fn fn,void *data);#endif tools/perf/tests/pmu-events.c +0
-0
tools/perf/tests/pmu-events.c static int compare_pmu_events(const struct pmu_event *e1, const struct pmu_event · line 343
Quoted 7 lines
tools/perf/tests/pmu-events.c +0
-4@@ -337,36 +337,12 @@ static int compare_pmu_events(const struct pmu_event *e1, const struct pmu_eventreturn -1 ; } -if (!is_same (e1->metric_expr, e2->metric_expr)) { -pr_debug2 ("testing event e1 %s: mismatched metric_expr, %s vs %s\n" , - e1->name, e1->metric_expr, e2->metric_expr); -return -1 ;
is there something new to replace this (I didn't check)?
Quoted 243 lines
tools/perf/tests/pmu-events.c +16
-42- } - -if (!is_same (e1->metric_name, e2->metric_name)) { -pr_debug2 ("testing event e1 %s: mismatched metric_name, %s vs %s\n" , - e1->name, e1->metric_name, e2->metric_name); -return -1 ; - } - -if (!is_same (e1->metric_group, e2->metric_group)) { -pr_debug2 ("testing event e1 %s: mismatched metric_group, %s vs %s\n" , - e1->name, e1->metric_group, e2->metric_group); -return -1 ; - } -if (!is_same (e1->deprecated, e2->deprecated)) {pr_debug2 ("testing event e1 %s: mismatched deprecated, %s vs %s\n" , e1->name, e1->deprecated, e2->deprecated);return -1 ; } -if (!is_same (e1->metric_constraint, e2->metric_constraint)) { -pr_debug2 ("testing event e1 %s: mismatched metric_constant, %s vs %s\n" , - e1->name, e1->metric_constraint, e2->metric_constraint); -return -1 ; - } -return 0 ; }@@ -432,9 +408,6 @@ static int test__pmu_event_table_core_callback(const struct pmu_event *pe,struct perf_pmu_test_event const **test_event_table;bool found = false; -if (!pe->name) -return 0 ; -if (pe->pmu) test_event_table = &uncore_events[0 ];else @@ -840,7 +813,7 @@ struct metric {struct metric_ref metric_ref; }; -static int test__parsing_callback (const struct pmu_event *pe,const struct pmu_events_table *table, +static int test__parsing_callback (const struct pmu_metric *pm,const struct pmu_events_table *table,void *data) {int *failures = data;@@ -854,10 +827,10 @@ static int test__parsing_callback(const struct pmu_event *pe, const struct pmu_e };int err =0 ; -if (!pe->metric_expr) +if (!pm->metric_expr)return 0 ; -pr_debug ("Found metric '%s'\n" , pe->metric_name); +pr_debug ("Found metric '%s'\n" , pm->metric_name); (*failures)++;/* @@ -877,14 +850,14 @@ static int test__parsing_callback(const struct pmu_event *pe, const struct pmu_eperf_evlist__set_maps (&evlist->core, cpus,NULL );runtime_stat__init (&st); - err =metricgroup__parse_groups_test (evlist, table, pe->metric_name, + err =metricgroup__parse_groups_test (evlist, table, pm->metric_name, false, false, &metric_events);if (err) { -if (!strcmp (pe->metric_name,"M1" ) || !strcmp (pe->metric_name,"M2" ) || - !strcmp (pe->metric_name,"M3" )) { +if (!strcmp (pm->metric_name,"M1" ) || !strcmp (pm->metric_name,"M2" ) || + !strcmp (pm->metric_name,"M3" )) { (*failures)--; -pr_debug ("Expected broken metric %s skipping\n" , pe->metric_name); +pr_debug ("Expected broken metric %s skipping\n" , pm->metric_name); err =0 ; } goto out_err;@@ -912,7 +885,7 @@ static int test__parsing_callback(const struct pmu_event *pe, const struct pmu_estruct metric_expr *mexp; list_for_each_entry (mexp, &me->head, nd) { -if (strcmp (mexp->metric_name, pe->metric_name)) +if (strcmp (mexp->metric_name, pm->metric_name))continue ;pr_debug ("Result %f\n" ,test_generic_metric (mexp,0 , &st)); err =0 ;@@ -921,11 +894,11 @@ static int test__parsing_callback(const struct pmu_event *pe, const struct pmu_e } } } -pr_debug ("Didn't find parsed metric %s" , pe->metric_name); +pr_debug ("Didn't find parsed metric %s" , pm->metric_name); err =1 ; out_err:if (err) -pr_debug ("Broken metric %s\n" , pe->metric_name); +pr_debug ("Broken metric %s\n" , pm->metric_name);/* ... cleanup. */ metricgroup__rblist_exit (&metric_events);@@ -941,8 +914,8 @@ static int test__parsing(struct test_suite *test __maybe_unused, {int failures =0 ; -pmu_for_each_core_event (test__parsing_callback, &failures); -pmu_for_each_sys_event (test__parsing_callback, &failures); +pmu_for_each_core_metric (test__parsing_callback, &failures); +pmu_for_each_sys_metric (test__parsing_callback, &failures);return failures ==0 ?TEST_OK :TEST_FAIL ; }@@ -1021,14 +994,11 @@ static int metric_parse_fake(const char *metric_name, const char *str)return ret; } -static int test__parsing_fake_callback (const struct pmu_event *pe, +static int test__parsing_fake_callback (const struct pmu_metric *pm,const struct pmu_events_table *table__maybe_unused ,void *data__maybe_unused ) { -if (!pe->metric_expr) -return 0 ; - -return metric_parse_fake (pe->metric_name, pe->metric_expr); +return metric_parse_fake (pm->metric_name, pm->metric_expr); }/* @@ -1047,11 +1017,11 @@ static int test__parsing_fake(struct test_suite *test __maybe_unused,return err; } - err =pmu_for_each_core_event (test__parsing_fake_callback,NULL ); + err =pmu_for_each_core_metric (test__parsing_fake_callback,NULL );if (err)return err; -return pmu_for_each_sys_event (test__parsing_fake_callback,NULL ); +return pmu_for_each_sys_metric (test__parsing_fake_callback,NULL ); }static struct test_case pmu_events_tests[] = {tools/perf/util/metricgroup.c +15
-16@@ -167,14 +167,14 @@ static void metricgroup___watchdog_constraint_hint(const char *name, bool foot) " echo 1 > /proc/sys/kernel/nmi_watchdog\n"); } -static bool metricgroup__has_constraint (const struct pmu_event *pe) +static bool metricgroup__has_constraint (const struct pmu_metric *pm) { -if (!pe->metric_constraint) +if (!pm->metric_constraint)return false; -if (!strcmp (pe->metric_constraint,"NO_NMI_WATCHDOG" ) && +if (!strcmp (pm->metric_constraint,"NO_NMI_WATCHDOG" ) &&sysctl__nmi_watchdog_enabled ()) { -metricgroup___watchdog_constraint_hint (pe->metric_name, false); +metricgroup___watchdog_constraint_hint (pm->metric_name, false);return true; }@@ -193,7 +193,7 @@ static void metric__free(struct metric *m)free (m); } -static struct metric *metric__new(const struct pmu_event *pe, +static struct metric *metric__new(const struct pmu_metric *pm, const char *modifier, bool metric_no_group, int runtime,@@ -210,15 +210,15 @@ static struct metric *metric__new(const struct pmu_event *pe,if (!m->pctx) goto out_err; - m->metric_name = pe->metric_name; + m->metric_name = pm->metric_name; m->modifier =NULL ;if (modifier) { m->modifier =strdup (modifier);if (!m->modifier) goto out_err; } - m->metric_expr = pe->metric_expr; - m->metric_unit = pe->unit; + m->metric_expr = pm->metric_expr; + m->metric_unit = pm->unit; m->pctx->sctx.user_requested_cpu_list =NULL ;if (user_requested_cpu_list) { m->pctx->sctx.user_requested_cpu_list =strdup (user_requested_cpu_list);@@ -227,7 +227,7 @@ static struct metric *metric__new(const struct pmu_event *pe, } m->pctx->sctx.runtime = runtime; m->pctx->sctx.system_wide = system_wide; - m->has_constraint = metric_no_group ||metricgroup__has_constraint (pe); + m->has_constraint = metric_no_group ||metricgroup__has_constraint (pm); m->metric_refs =NULL ; m->evlist =NULL ;@@ -348,10 +348,10 @@ static bool match_metric(const char *n, const char *list)return false; } -static bool match_pe_metric (const struct pmu_event *pe,const char *metric) +static bool match_pm_metric (const struct pmu_metric *pm,const char *metric) { -return match_metric (pe->metric_group, metric) || -match_metric (pe->metric_name, metric); +return match_metric (pm->metric_group, metric) || +match_metric (pm->metric_name, metric); }/** struct mep - RB-tree node for building printing information. */ @@ -420,13 +420,13 @@ static struct mep *mep_lookup(struct rblist *groups, const char *metric_group,return NULL ; } -static int metricgroup__add_to_mep_groups (const struct pmu_event *pe, +static int metricgroup__add_to_mep_groups (const struct pmu_metric *pm,struct rblist *groups) {const char *g;char *omg, *mg; - mg =strdup (pe->metric_group ?:"No_group" ); + mg =strdup (pm->metric_group ?:"No_group" );if (!mg)return -ENOMEM ; omg = mg;@@ -435,15 +435,15 @@ static int metricgroup__add_to_mep_groups(const struct pmu_event *pe, g =skip_spaces (g);if (strlen (g)) - me =mep_lookup (groups, g, pe->metric_name); + me =mep_lookup (groups, g, pm->metric_name);else - me =mep_lookup (groups,"No_group" , pe->metric_name);
tools/perf/util/metricgroup.c static int metricgroup__add_to_mep_groups(const struct pmu_event *pe, · line 445
Quoted 10 lines
tools/perf/util/metricgroup.c +4
-4+ me =mep_lookup (groups,"No_group" , pm->metric_name);if (me) { - me->metric_desc = pe->desc; - me->metric_long_desc = pe->long_desc; - me->metric_expr = pe->metric_expr; - me->metric_unit = pe->unit; + me->metric_desc = pm->desc; + me->metric_long_desc = pm->long_desc; + me->metric_expr = pm->metric_expr;
could struct pmu_metric be a sub-struct in struct mep?
Quoted 237 lines
tools/perf/util/metricgroup.c +36
-38+ me->metric_unit = pm->unit; } }free (omg);@@ -452,40 +452,37 @@ static int metricgroup__add_to_mep_groups(const struct pmu_event *pe, }struct metricgroup_iter_data { -pmu_event_iter_fn fn; +pmu_metric_iter_fn fn;void *data; }; -static int metricgroup__sys_event_iter (const struct pmu_event *pe, +static int metricgroup__sys_event_iter (const struct pmu_metric *pm,const struct pmu_events_table *table,void *data) {struct metricgroup_iter_data *d = data;struct perf_pmu *pmu =NULL ; -if (!pe->metric_expr || !pe->compat) +if (!pm->metric_expr || !pm->compat)return 0 ;while ((pmu =perf_pmu__scan (pmu))) { -if (!pmu->id ||strcmp (pmu->id, pe->compat)) +if (!pmu->id ||strcmp (pmu->id, pm->compat))continue ; -return d->fn (pe, table, d->data); +return d->fn (pm, table, d->data); }return 0 ; } -static int metricgroup__add_to_mep_groups_callback (const struct pmu_event *pe, +static int metricgroup__add_to_mep_groups_callback (const struct pmu_metric *pm,const struct pmu_events_table *table__maybe_unused ,void *vdata) {struct rblist *groups = vdata; -if (!pe->metric_name) -return 0 ; - -return metricgroup__add_to_mep_groups (pe, groups); +return metricgroup__add_to_mep_groups (pm, groups); }void metricgroup__print (const struct print_callbacks *print_cb,void *print_state)@@ -500,16 +497,16 @@ void metricgroup__print(const struct print_callbacks *print_cb, void *print_stat groups.node_delete = mep_delete; table =pmu_events_table__find ();if (table) { -pmu_events_table_for_each_event (table, - metricgroup__add_to_mep_groups_callback, - &groups); +pmu_events_table_for_each_metric (table, + metricgroup__add_to_mep_groups_callback, + &groups); } {struct metricgroup_iter_data data = { .fn = metricgroup__add_to_mep_groups_callback, .data = &groups, }; -pmu_for_each_sys_event (metricgroup__sys_event_iter, &data); +pmu_for_each_sys_metric (metricgroup__sys_event_iter, &data); }for (node =rb_first_cached (&groups.entries); node; node = next) {@@ -743,7 +740,7 @@ static int metricgroup__build_event_string(struct strbuf *events,#undef RETURN_IF_NON_ZERO } -int __weak arch_get_runtimeparam (const struct pmu_event *pe__maybe_unused ) +int __weak arch_get_runtimeparam (const struct pmu_metric *pm__maybe_unused ) {return 1 ; }@@ -773,10 +770,10 @@ struct metricgroup_add_iter_data {static bool metricgroup__find_metric (const char *metric,const struct pmu_events_table *table, -struct pmu_event *pe); +struct pmu_metric *pm); static int add_metric(struct list_head *metric_list, - const struct pmu_event *pe, + const struct pmu_metric *pm, const char *modifier, bool metric_no_group, const char *user_requested_cpu_list,@@ -816,10 +813,10 @@ static int resolve_metric(struct list_head *metric_list,size_t bkt;struct to_resolve {/* The metric to resolve. */ -struct pmu_event pe; +struct pmu_metric pm;/* * The key in the IDs map, this may differ from in case, -* etc. from pe->metric_name. +* etc. from pm->metric_name. */ const char *key; } *pending =NULL ;@@ -830,15 +827,15 @@ static int resolve_metric(struct list_head *metric_list,* the pending array. */ hashmap__for_each_entry (root_metric->pctx->ids, cur, bkt) { -struct pmu_event pe; +struct pmu_metric pm; -if (metricgroup__find_metric (cur->pkey, table, &pe)) { +if (metricgroup__find_metric (cur->pkey, table, &pm)) { pending =realloc (pending, (pending_cnt +1 ) *sizeof (struct to_resolve ));if (!pending)return -ENOMEM ; -memcpy (&pending[pending_cnt].pe, &pe,sizeof (pe)); +memcpy (&pending[pending_cnt].pm, &pm,sizeof (pm)); pending[pending_cnt].key = cur->pkey; pending_cnt++; }@@ -853,7 +850,7 @@ static int resolve_metric(struct list_head *metric_list,* context. */ for (i =0 ; i < pending_cnt; i++) { - ret =add_metric (metric_list, &pending[i].pe, modifier, metric_no_group, + ret =add_metric (metric_list, &pending[i].pm, modifier, metric_no_group, user_requested_cpu_list, system_wide, root_metric, visited, table);if (ret)@@ -867,7 +864,7 @@ static int resolve_metric(struct list_head *metric_list,/** * __add_metric - Add a metric to metric_list. * @metric_list: The list the metric is added to. -* @pe: The pmu_event containing the metric to be added. +* @pm: The pmu_metric containing the metric to be added. * @modifier: if non-null event modifiers like "u". * @metric_no_group: Should events written to events be grouped "{}" or * global. Grouping is the default but due to multiplexing the @@ -884,7 +881,7 @@ static int resolve_metric(struct list_head *metric_list,* architecture perf is running upon. */ static int __add_metric(struct list_head *metric_list, - const struct pmu_event *pe, + const struct pmu_metric *pm, const char *modifier, bool metric_no_group, int runtime,@@ -898,13 +895,13 @@ static int __add_metric(struct list_head *metric_list,int ret;bool is_root = !root_metric;struct visited_metric visited_node = { - .name = pe->metric_name, + .name = pm->metric_name, .parent = visited, };for (vm = visited; vm; vm = vm->parent) { -if (!strcmp (pe->metric_name, vm->name)) { -pr_err ("failed: recursion detected for %s\n" , pe->metric_name); +if (!strcmp (pm->metric_name, vm->name)) { +pr_err ("failed: recursion detected for %s\n" , pm->metric_name);return -1 ; } }@@ -914,7 +911,7 @@ static int __add_metric(struct list_head *metric_list,* This metric is the root of a tree and may reference other * metrics that are added recursively. */ - root_metric =metric__new (pe, modifier, metric_no_group, runtime, + root_metric =metric__new (pm, modifier, metric_no_group, runtime, user_requested_cpu_list, system_wide);if (!root_metric)return -ENOMEM ;@@ -929,7 +926,7 @@ static int __add_metric(struct list_head *metric_list,*/ if (root_metric->metric_refs) {for (; root_metric->metric_refs[cnt].metric_name; cnt++) { -if (!strcmp (pe->metric_name, +if (!strcmp (pm->metric_name, root_metric->metric_refs[cnt].metric_name))return 0 ; }@@ -947,8 +944,8 @@ static int __add_metric(struct list_head *metric_list,* need to change them, so there's no need to create * our own copy. */ - root_metric->metric_refs[cnt].metric_name = pe->metric_name; - root_metric->metric_refs[cnt].metric_expr = pe->metric_expr; + root_metric->metric_refs[cnt].metric_name = pm->metric_name; + root_metric->metric_refs[cnt].metric_expr = pm->metric_expr;/* Null terminate array. */ root_metric->metric_refs[cnt+1 ].metric_name =NULL ;@@ -959,7 +956,7 @@ static int __add_metric(struct list_head *metric_list,* For both the parent and referenced metrics, we parse * all the metric's IDs and add it to the root context. */ -if (expr__find_ids (pe->metric_expr,NULL , root_metric->pctx) <0 ) { +if (expr__find_ids (pm->metric_expr,NULL , root_metric->pctx) <0 ) {/* Broken metric. */ ret = -EINVAL ; }else {@@ -981,37 +978,37 @@ static int __add_metric(struct list_head *metric_list,struct metricgroup__find_metric_data {const char *metric; -struct pmu_event *pe; +struct pmu_metric *pm; }; -static int metricgroup__find_metric_callback (const struct pmu_event *pe, +static int metricgroup__find_metric_callback (const struct pmu_metric *pm,const struct pmu_events_table *table__maybe_unused ,void *vdata) {struct metricgroup__find_metric_data *data = vdata; -if (!match_metric (pe->metric_name, data->metric)) +if (!match_metric (pm->metric_name, data->metric))return 0 ; -memcpy (data->pe, pe,sizeof (*pe)); +memcpy (data->pm, pm,sizeof (*pm));return 1 ; }static bool metricgroup__find_metric (const char *metric,const struct pmu_events_table *table, -struct pmu_event *pe) +struct pmu_metric *pm)
tools/perf/util/metricgroup.c static int __add_metric(struct list_head *metric_list, · line 1007
Quoted 10 lines
tools/perf/util/metricgroup.c +2
-2{struct metricgroup__find_metric_data data = { .metric = metric, - .pe = pe, + .pm = pm, }; -return pmu_events_table_for_each_event (table, metricgroup__find_metric_callback, &data) +return pmu_events_table_for_each_metric (table, metricgroup__find_metric_callback, &data) ? true : false;
comment on pre-existing code: since it returns bool, I don't think the ternary operator is required at all. There are many instances of this.
Quoted 407 lines
tools/perf/util/metricgroup.c +22
-22} static int add_metric(struct list_head *metric_list, - const struct pmu_event *pe, + const struct pmu_metric *pm, const char *modifier, bool metric_no_group, const char *user_requested_cpu_list,@@ -1022,16 +1019,16 @@ static int add_metric(struct list_head *metric_list, {int ret =0 ; -pr_debug ("metric expr %s for %s\n" , pe->metric_expr, pe->metric_name); +pr_debug ("metric expr %s for %s\n" , pm->metric_expr, pm->metric_name); -if (!strstr (pe->metric_expr,"?" )) { - ret =__add_metric (metric_list, pe, modifier, metric_no_group,0 , +if (!strstr (pm->metric_expr,"?" )) { + ret =__add_metric (metric_list, pm, modifier, metric_no_group,0 , user_requested_cpu_list, system_wide, root_metric, visited, table); }else {int j, count; - count =arch_get_runtimeparam (pe); + count =arch_get_runtimeparam (pm);/* This loop is added to create multiple * events depend on count value and add @@ -1039,7 +1036,7 @@ static int add_metric(struct list_head *metric_list,*/ for (j =0 ; j < count && !ret; j++) - ret =__add_metric (metric_list, pe, modifier, metric_no_group, j, + ret =__add_metric (metric_list, pm, modifier, metric_no_group, j, user_requested_cpu_list, system_wide, root_metric, visited, table); }@@ -1047,17 +1044,17 @@ static int add_metric(struct list_head *metric_list,return ret; } -static int metricgroup__add_metric_sys_event_iter (const struct pmu_event *pe, +static int metricgroup__add_metric_sys_event_iter (const struct pmu_metric *pm,const struct pmu_events_table *table__maybe_unused ,void *data) {struct metricgroup_add_iter_data *d = data;int ret; -if (!match_pe_metric (pe, d->metric_name)) +if (!match_pm_metric (pm, d->metric_name))return 0 ; - ret =add_metric (d->metric_list, pe, d->modifier, d->metric_no_group, + ret =add_metric (d->metric_list, pm, d->modifier, d->metric_no_group, d->user_requested_cpu_list, d->system_wide, d->root_metric, d->visited, d->table);if (ret)@@ -1107,19 +1104,19 @@ struct metricgroup__add_metric_data {bool has_match; }; -static int metricgroup__add_metric_callback (const struct pmu_event *pe, +static int metricgroup__add_metric_callback (const struct pmu_metric *pm,const struct pmu_events_table *table,void *vdata) {struct metricgroup__add_metric_data *data = vdata;int ret =0 ; -if (pe->metric_expr && - (match_metric (pe->metric_group, data->metric_name) || -match_metric (pe->metric_name, data->metric_name))) { +if (pm->metric_expr && + (match_metric (pm->metric_group, data->metric_name) || +match_metric (pm->metric_name, data->metric_name))) { data->has_match = true; - ret =add_metric (data->list, pe, data->modifier, data->metric_no_group, + ret =add_metric (data->list, pm, data->modifier, data->metric_no_group, data->user_requested_cpu_list, data->system_wide,/*root_metric=*/ NULL ,/*visited_metrics=*/ NULL , table); }@@ -1166,8 +1163,8 @@ static int metricgroup__add_metric(const char *metric_name, const char *modifier* Iterate over all metrics seeing if metric matches either the * name or group. When it does add the metric to the list. */ - ret =pmu_events_table_for_each_event (table, metricgroup__add_metric_callback, - &data); + ret =pmu_events_table_for_each_metric (table, metricgroup__add_metric_callback, + &data);if (ret) goto out;@@ -1189,7 +1186,7 @@ static int metricgroup__add_metric(const char *metric_name, const char *modifier }, }; -pmu_for_each_sys_event (metricgroup__sys_event_iter, &data); +pmu_for_each_sys_metric (metricgroup__sys_event_iter, &data); }/* End of pmu events. */ if (!has_match)@@ -1603,16 +1600,16 @@ int metricgroup__parse_groups_test(struct evlist *evlist, &perf_pmu__fake, metric_events, table); } -static int metricgroup__has_metric_callback (const struct pmu_event *pe, +static int metricgroup__has_metric_callback (const struct pmu_metric *pm,const struct pmu_events_table *table__maybe_unused ,void *vdata) {const char *metric = vdata; -if (!pe->metric_expr) +if (!pm->metric_expr)return 0 ; -if (match_metric (pe->metric_name, metric)) +if (match_metric (pm->metric_name, metric))return 1 ;return 0 ;@@ -1625,8 +1622,8 @@ bool metricgroup__has_metric(const char *metric)if (!table)return false; -return pmu_events_table_for_each_event (table, metricgroup__has_metric_callback, - (void *)metric) ? true : false; +return pmu_events_table_for_each_metric (table, metricgroup__has_metric_callback, + (void *)metric) ? true : false; } int metricgroup__copy_metric_events(struct evlist *evlist, struct cgroup *cgrp,tools/perf/util/metricgroup.h +1
-1@@ -81,7 +81,7 @@ int metricgroup__parse_groups_test(struct evlist *evlist,void metricgroup__print (const struct print_callbacks *print_cb,void *print_state);bool metricgroup__has_metric (const char *metric); -int arch_get_runtimeparam (const struct pmu_event *pe__maybe_unused ); +int arch_get_runtimeparam (const struct pmu_metric *pm);void metricgroup__rblist_exit (struct rblist *metric_events); int metricgroup__copy_metric_events(struct evlist *evlist, struct cgroup *cgrp,tools/perf/util/parse-events.c +0
-2@@ -1570,8 +1570,6 @@ int parse_events_add_pmu(struct parse_events_state *parse_state, evsel->scale = info.scale; evsel->per_pkg = info.per_pkg; evsel->snapshot = info.snapshot; - evsel->metric_expr = info.metric_expr; - evsel->metric_name = info.metric_name;return 0 ; }tools/perf/util/pmu.c +3
-32@@ -283,10 +283,6 @@ static void perf_pmu_update_alias(struct perf_pmu_alias *old,perf_pmu_assign_str (old->name,"long_desc" , &old->long_desc, &newalias->long_desc);perf_pmu_assign_str (old->name,"topic" , &old->topic, &newalias->topic); -perf_pmu_assign_str (old->name,"metric_expr" , &old->metric_expr, - &newalias->metric_expr); -perf_pmu_assign_str (old->name,"metric_name" , &old->metric_name, - &newalias->metric_name);perf_pmu_assign_str (old->name,"value" , &old->str, &newalias->str); old->scale = newalias->scale; old->per_pkg = newalias->per_pkg;@@ -302,8 +298,6 @@ void perf_pmu_free_alias(struct perf_pmu_alias *newalias)zfree (&newalias->long_desc);zfree (&newalias->topic);zfree (&newalias->str); -zfree (&newalias->metric_expr); -zfree (&newalias->metric_name);zfree (&newalias->pmu_name);parse_events_terms__purge (&newalias->terms);free (newalias);@@ -340,16 +334,13 @@ static int __perf_pmu__new_alias(struct list_head *list, char *dir, char *name,int num;char newval[256 ];char *long_desc =NULL , *topic =NULL , *unit =NULL , *perpkg =NULL , - *metric_expr =NULL , *metric_name =NULL , *deprecated =NULL , - *pmu_name =NULL ; + *deprecated =NULL , *pmu_name =NULL ;if (pe) { long_desc = (char *)pe->long_desc; topic = (char *)pe->topic; unit = (char *)pe->unit; perpkg = (char *)pe->perpkg; - metric_expr = (char *)pe->metric_expr; - metric_name = (char *)pe->metric_name; deprecated = (char *)pe->deprecated; pmu_name = (char *)pe->pmu; }@@ -404,8 +395,6 @@ static int __perf_pmu__new_alias(struct list_head *list, char *dir, char *name,perf_pmu__parse_snapshot (alias, dir, name); } - alias->metric_expr = metric_expr ?strdup (metric_expr) :NULL ; - alias->metric_name = metric_name ?strdup (metric_name):NULL ; alias->desc = desc ?strdup (desc) :NULL ; alias->long_desc = long_desc ?strdup (long_desc) : desc ?strdup (desc) :NULL ;@@ -822,9 +811,6 @@ static int pmu_add_cpu_aliases_map_callback(const struct pmu_event *pe,struct pmu_add_cpu_aliases_map_data *data = vdata;const char *pname = pe->pmu ? pe->pmu : data->cpu_name; -if (!pe->name) -return 0 ; -if (data->pmu->is_uncore &&pmu_uncore_alias_match (pname, data->name)) goto new_alias;@@ -879,12 +865,6 @@ static int pmu_add_sys_aliases_iter_fn(const struct pmu_event *pe,struct pmu_sys_event_iter_data *idata = data;struct perf_pmu *pmu = idata->pmu; -if (!pe->name) { -if (pe->metric_group || pe->metric_name) -return 0 ; -return -EINVAL ; - } -if (!pe->compat || !pe->pmu)return 0 ;@@ -1469,8 +1449,6 @@ int perf_pmu__check_alias(struct perf_pmu *pmu, struct list_head *head_terms, info->unit =NULL ; info->scale =0.0 ; info->snapshot = false; - info->metric_expr =NULL ; - info->metric_name =NULL ;list_for_each_entry_safe (term, h, head_terms, list) { alias =pmu_find_alias (pmu, term);@@ -1486,8 +1464,6 @@ int perf_pmu__check_alias(struct perf_pmu *pmu, struct list_head *head_terms,if (alias->per_pkg) info->per_pkg = true; - info->metric_expr = alias->metric_expr; - info->metric_name = alias->metric_name;list_del_init (&term->list);parse_events_term__delete (term);@@ -1703,8 +1679,7 @@ void print_pmu_events(const struct print_callbacks *print_cb, void *print_state)for (j =0 ; j < len; j++) {const char *name, *alias =NULL , *scale_unit =NULL , *desc =NULL , *long_desc =NULL , - *encoding_desc =NULL , *topic =NULL , - *metric_name =NULL , *metric_expr =NULL ; + *encoding_desc =NULL , *topic =NULL ;bool deprecated = false;size_t buf_used;@@ -1742,8 +1717,6 @@ void print_pmu_events(const struct print_callbacks *print_cb, void *print_state) buf_used +=snprintf (buf + buf_used,sizeof (buf) - buf_used,"%s/%s/" , aliases[j].pmu->name, aliases[j].event->str) +1 ; - metric_name = aliases[j].event->metric_name; - metric_expr = aliases[j].event->metric_expr; deprecated = aliases[j].event->deprecated; } print_cb->print_event(print_state,@@ -1756,9 +1729,7 @@ void print_pmu_events(const struct print_callbacks *print_cb, void *print_state)"Kernel PMU event" , desc, long_desc, - encoding_desc, - metric_name, - metric_expr); + encoding_desc); }if (printed &&pager_in_use ())printf ("\n" );tools/perf/util/pmu.h +0
-9@@ -133,8 +133,6 @@ extern struct perf_pmu perf_pmu__fake;struct perf_pmu_info {const char *unit; -const char *metric_expr; -const char *metric_name;double scale;bool per_pkg;bool snapshot;@@ -188,13 +186,6 @@ struct perf_pmu_alias {* default. */ bool deprecated; -/** -* @metric_expr: A metric expression associated with an event. Doing -* this makes little sense due to scale and unit applying to both. -*/ -char *metric_expr; -/** @metric_name: A name for the metric. unit applying to both. */ -char *metric_name;/** @pmu_name: The name copied from struct perf_pmu. */ char *pmu_name; };tools/perf/util/print-events.c +8
-24@@ -101,9 +101,7 @@ void print_tracepoint_events(const struct print_callbacks *print_cb, void *print"Tracepoint event" ,/*desc=*/ NULL ,/*long_desc=*/ NULL , -/*encoding_desc=*/ NULL , -/*metric_name=*/ NULL , - /*metric_expr=*/NULL); + /*encoding_desc=*/NULL); }free (dir_path);free (evt_namelist);@@ -195,9 +193,7 @@ void print_sdt_events(const struct print_callbacks *print_cb, void *print_state)"SDT event" ,/*desc=*/ NULL ,/*long_desc=*/ NULL , -/*encoding_desc=*/ NULL , -/*metric_name=*/ NULL , - /*metric_expr=*/NULL); + /*encoding_desc=*/NULL);free (evt_name); }@@ -255,9 +251,7 @@ int print_hwcache_events(const struct print_callbacks *print_cb, void *print_sta event_type_descriptors[PERF_TYPE_HW_CACHE ],/*desc=*/ NULL ,/*long_desc=*/ NULL , -/*encoding_desc=*/ NULL , -/*metric_name=*/ NULL , - /*metric_expr=*/NULL); + /*encoding_desc=*/NULL); }strlist__delete (evt_name_list);return 0 ;@@ -277,9 +271,7 @@ void print_tool_events(const struct print_callbacks *print_cb, void *print_state"Tool event" ,/*desc=*/ NULL ,/*long_desc=*/ NULL , -/*encoding_desc=*/ NULL , -/*metric_name=*/ NULL , - /*metric_expr=*/NULL); + /*encoding_desc=*/NULL); } }@@ -331,9 +323,7 @@ void print_symbol_events(const struct print_callbacks *print_cb, void *print_sta event_type_descriptors[type],/*desc=*/ NULL ,/*long_desc=*/ NULL , -/*encoding_desc=*/ NULL , -/*metric_name=*/ NULL , - /*metric_expr=*/NULL); + /*encoding_desc=*/NULL); }strlist__delete (evt_name_list); }@@ -364,9 +354,7 @@ void print_events(const struct print_callbacks *print_cb, void *print_state) event_type_descriptors[PERF_TYPE_RAW ],/*desc=*/ NULL ,/*long_desc=*/ NULL , -/*encoding_desc=*/ NULL , -/*metric_name=*/ NULL , - /*metric_expr=*/NULL); + /*encoding_desc=*/NULL); print_cb->print_event(print_state, /*topic=*/NULL,@@ -378,9 +366,7 @@ void print_events(const struct print_callbacks *print_cb, void *print_state) event_type_descriptors[PERF_TYPE_RAW ],"(see 'man perf-list' on how to encode it)" ,/*long_desc=*/ NULL , -/*encoding_desc=*/ NULL , -/*metric_name=*/ NULL , - /*metric_expr=*/NULL); + /*encoding_desc=*/NULL); print_cb->print_event(print_state, /*topic=*/NULL,@@ -392,9 +378,7 @@ void print_events(const struct print_callbacks *print_cb, void *print_state) event_type_descriptors[PERF_TYPE_BREAKPOINT ],/*desc=*/ NULL ,/*long_desc=*/ NULL , -/*encoding_desc=*/ NULL , -/*metric_name=*/ NULL , - /*metric_expr=*/NULL); + /*encoding_desc=*/NULL);print_tracepoint_events (print_cb, print_state);tools/perf/util/print-events.h +0
-0
tools/perf/util/print-events.h struct print_callbacks { · line 22
Quoted 9 lines
tools/perf/util/print-events.h +1
-2@@ -16,8 +16,7 @@ struct print_callbacks {const char *scale_unit, bool deprecated, const char *event_type_desc, const char *desc, const char *long_desc, - const char *encoding_desc, - const char *metric_name, const char *metric_expr); + const char *encoding_desc); void (*print_metric)(void *print_state, const char *group, const char *name,
_______________________________________________
linux-arm-kernel mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel
Ian Rogers
reviewerRe: [PATCH v2 4/9] perf pmu-events: Separate metric out of pmu_event
Quoted 18 lines · John Garry
Quoted 6 lines · Ian Rogers
Previously both events and metrics were encoded in struct
pmu_event. Create a newpmu_metricthat has the metric related variables and remove these frompmu_event. Add iterators forpmu_metricand use in places that metrics are desired rather than events.Quoted 9 lines
Note, this change removes the setting of evsel's metric_name/expr as these fields are no longer part of struct
pmu_event. The metric remains but is no longer implicitly requested when the event is. This impacts a few Intel uncore events, however, as the ScaleUnit is shared by the event and the metric this utility is questionable. Also the MetricNames look broken (contain spaces) in some cases and when trying to use the functionality with '-e' the metrics fail but regular metrics with '-M' work. For example, on SkylakeX '-M' works:
Quoted 5 lines
I like this change. It's quite large for a single patch. Just some sparse comments below.
BTW, it there a better name for metric struct variable than "pm"? To me and many other people, pm is power management.
Agreed. There a few things like that in the code, I dislike the overload on "core". I've left it as pm as pmu_event became pe, so it is consistent for pmu_metric to become pm. We can do a global rename as a follow up.
Quoted 242 lines
Quoted 242 lines
``` $ perf stat -M LLC_MISSES.PCIE_WRITE -a sleep 1 Performance counter stats for 'system wide': 0 UNC_IIO_DATA_REQ_OF_CPU.MEM_WRITE.PART2 # 57896.0 Bytes LLC_MISSES.PCIE_WRITE (49.84%) 7,174 UNC_IIO_DATA_REQ_OF_CPU.MEM_WRITE.PART1 (49.85%) 0 UNC_IIO_DATA_REQ_OF_CPU.MEM_WRITE.PART3 (50.16%) 63 UNC_IIO_DATA_REQ_OF_CPU.MEM_WRITE.PART0 (50.15%) 1.004576381 seconds time elapsed ```whilst the event '-e' version is broken even with --group/-g (fwiw, we should also remove -g [1]):
``` $ perf stat -g -e LLC_MISSES.PCIE_WRITE -g -a sleep 1 Add UNC_IIO_DATA_REQ_OF_CPU.MEM_WRITE.PART2 event to groups to get metric expression for LLC_MISSES.PCIE_WRITE Add UNC_IIO_DATA_REQ_OF_CPU.MEM_WRITE.PART1 event to groups to get metric expression for LLC_MISSES.PCIE_WRITE Add UNC_IIO_DATA_REQ_OF_CPU.MEM_WRITE.PART3 event to groups to get metric expression for LLC_MISSES.PCIE_WRITE Add UNC_IIO_DATA_REQ_OF_CPU.MEM_WRITE.PART0 event to groups to get metric expression for LLC_MISSES.PCIE_WRITE Add UNC_IIO_DATA_REQ_OF_CPU.MEM_WRITE.PART2 event to groups to get metric expression for LLC_MISSES.PCIE_WRITE Add UNC_IIO_DATA_REQ_OF_CPU.MEM_WRITE.PART1 event to groups to get metric expression for LLC_MISSES.PCIE_WRITE Add UNC_IIO_DATA_REQ_OF_CPU.MEM_WRITE.PART3 event to groups to get metric expression for LLC_MISSES.PCIE_WRITE Add UNC_IIO_DATA_REQ_OF_CPU.MEM_WRITE.PART0 event to groups to get metric expression for LLC_MISSES.PCIE_WRITE Add UNC_IIO_DATA_REQ_OF_CPU.MEM_WRITE.PART2 event to groups to get metric expression for LLC_MISSES.PCIE_WRITE Add UNC_IIO_DATA_REQ_OF_CPU.MEM_WRITE.PART1 event to groups to get metric expression for LLC_MISSES.PCIE_WRITE Add UNC_IIO_DATA_REQ_OF_CPU.MEM_WRITE.PART3 event to groups to get metric expression for LLC_MISSES.PCIE_WRITE Add UNC_IIO_DATA_REQ_OF_CPU.MEM_WRITE.PART0 event to groups to get metric expression for LLC_MISSES.PCIE_WRITE Add UNC_IIO_DATA_REQ_OF_CPU.MEM_WRITE.PART2 event to groups to get metric expression for LLC_MISSES.PCIE_WRITE Add UNC_IIO_DATA_REQ_OF_CPU.MEM_WRITE.PART1 event to groups to get metric expression for LLC_MISSES.PCIE_WRITE Add UNC_IIO_DATA_REQ_OF_CPU.MEM_WRITE.PART3 event to groups to get metric expression for LLC_MISSES.PCIE_WRITE Add UNC_IIO_DATA_REQ_OF_CPU.MEM_WRITE.PART0 event to groups to get metric expression for LLC_MISSES.PCIE_WRITE Add UNC_IIO_DATA_REQ_OF_CPU.MEM_WRITE.PART2 event to groups to get metric expression for LLC_MISSES.PCIE_WRITE Add UNC_IIO_DATA_REQ_OF_CPU.MEM_WRITE.PART1 event to groups to get metric expression for LLC_MISSES.PCIE_WRITE Add UNC_IIO_DATA_REQ_OF_CPU.MEM_WRITE.PART3 event to groups to get metric expression for LLC_MISSES.PCIE_WRITE Add UNC_IIO_DATA_REQ_OF_CPU.MEM_WRITE.PART0 event to groups to get metric expression for LLC_MISSES.PCIE_WRITE Add UNC_IIO_DATA_REQ_OF_CPU.MEM_WRITE.PART2 event to groups to get metric expression for LLC_MISSES.PCIE_WRITE Add UNC_IIO_DATA_REQ_OF_CPU.MEM_WRITE.PART1 event to groups to get metric expression for LLC_MISSES.PCIE_WRITE Add UNC_IIO_DATA_REQ_OF_CPU.MEM_WRITE.PART3 event to groups to get metric expression for LLC_MISSES.PCIE_WRITE Add UNC_IIO_DATA_REQ_OF_CPU.MEM_WRITE.PART0 event to groups to get metric expression for LLC_MISSES.PCIE_WRITE Performance counter stats for 'system wide': 27,316 Bytes LLC_MISSES.PCIE_WRITE 1.004505469 seconds time elapsed ```The code also carries warnings where the user is supposed to select events for metrics [2] but given the lack of use of such a feature, let's clean the code and just remove.
With NO_JEVENTS=1 the
empty-pmu-events.cis used and the separation to use metrics causes "Parse and process metrics" and "Event expansion for cgroups" to fail as they fail to find necessary metrics. This is fixed in a later change.[1] https://urldefense.com/v3/__https://lore.kernel.org/lkml/[email protected]/__;!!ACWV5N9M2RV99hQ!MTNgf7sIIl2VCxBzo-biSeul-ZQkuU73pT6gBbsb6y0AqsVlgGCcmuigDyevtrgmpHDN1QkEJwbJYvREu3Q$
[2] https://urldefense.com/v3/__https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/tree/tools/perf/util/stat-shadow.c?id=01b8957b738f42f96a130079bc951b3cc78c5b8a*n425__;Iw!!ACWV5N9M2RV99hQ!MTNgf7sIIl2VCxBzo-biSeul-ZQkuU73pT6gBbsb6y0AqsVlgGCcmuigDyevtrgmpHDN1QkEJwbJ8H0tppQ$Signed-off-by Ian Rogers <[email protected]>--- tools/perf/arch/powerpc/util/header.c | 4 +- tools/perf/builtin-list.c | 20 +-- tools/perf/pmu-events/empty-pmu-events.c | 73 ++++++++-- tools/perf/pmu-events/jevents.py | 82 +++++++++++- tools/perf/pmu-events/pmu-events.h | 20 ++- tools/perf/tests/pmu-events.c | 62 +++------ tools/perf/util/metricgroup.c | 161 +++++++++++------------ tools/perf/util/metricgroup.h | 2 +- tools/perf/util/parse-events.c | 2 - tools/perf/util/pmu.c | 35 +---- tools/perf/util/pmu.h | 9 -- tools/perf/util/print-events.c | 32 ++--- tools/perf/util/print-events.h | 3 +- 13 files changed, 266 insertions(+), 239 deletions(-)tools/perf/arch/powerpc/util/header.c +2
-2@@ -40,11 +40,11 @@ get_cpuid_str(struct perf_pmu *pmu __maybe_unused)return bufp; } -int arch_get_runtimeparam (const struct pmu_event *pe) +int arch_get_runtimeparam (const struct pmu_metric *pm) {int count;char path[PATH_MAX ] ="/devices/hv_24x7/interface/" ; -atoi (pe->aggr_mode) == PerChip ?strcat (path,"sockets" ) :strcat (path,"coresperchip" ); +atoi (pm->aggr_mode) == PerChip ?strcat (path,"sockets" ) :strcat (path,"coresperchip" );return sysfs__read_int (path, &count) <0 ?1 : count; }tools/perf/builtin-list.c +2
-18@@ -99,8 +99,7 @@ static void default_print_event(void *ps, const char *pmu_name, const char *topiconst char *scale_unit__maybe_unused , bool deprecated, const char *event_type_desc, const char *desc, const char *long_desc, - const char *encoding_desc, - const char *metric_name, const char *metric_expr) + const char *encoding_desc) { struct print_state *print_state = ps;int pos;@@ -159,10 +158,6 @@ static void default_print_event(void *ps, const char *pmu_name, const char *topiif (print_state->detailed && encoding_desc) {printf ("%*s" ,8 ,"" );wordwrap (encoding_desc,8 ,pager_get_columns (),0 ); -if (metric_name) -printf (" MetricName: %s" , metric_name); -if (metric_expr) -printf (" MetricExpr: %s" , metric_expr);putchar ('\n' ); } }@@ -308,8 +303,7 @@ static void json_print_event(void *ps, const char *pmu_name, const char *topic,const char *scale_unit, bool deprecated, const char *event_type_desc, const char *desc, const char *long_desc, - const char *encoding_desc, - const char *metric_name, const char *metric_expr) + const char *encoding_desc) { struct json_print_state *print_state = ps;bool need_sep = false;@@ -366,16 +360,6 @@ static void json_print_event(void *ps, const char *pmu_name, const char *topic, encoding_desc); need_sep = true; } -if (metric_name) { -fix_escape_printf (&buf,"%s\t\"MetricName\": \"%S\"" , need_sep ?",\n" :"" , - metric_name); - need_sep = true; - } -if (metric_expr) { -fix_escape_printf (&buf,"%s\t\"MetricExpr\": \"%S\"" , need_sep ?",\n" :"" , - metric_expr); - need_sep = true; - }printf ("%s}" , need_sep ?"\n" :"" );strbuf_release (&buf); }tools/perf/pmu-events/empty-pmu-events.c +23
-10@@ -11,7 +11,7 @@#include <string.h> #include <stddef.h> -static const struct pmu_event pme_test_soc_cpu[] = { +static const struct pmu_event pmu_events__test_soc_cpu[] = { { .name ="l3_cache_rd" , .event ="event=0x40" ,@@ -105,6 +105,14 @@ static const struct pmu_event pme_test_soc_cpu[] = { .desc = "L2 BTB Correction", .topic = "branch", }, + { + .name = 0, + .event = 0, + .desc = 0, + }, +}; + +static const struct pmu_metric pmu_metrics__test_soc_cpu[] = { { .metric_expr ="1 / IPC" , .metric_name ="CPI" ,@@ -170,9 +178,8 @@ static const struct pmu_event pme_test_soc_cpu[] = { .metric_name = "L1D_Cache_Fill_BW", }, { - .name = 0, - .event = 0, - .desc = 0, + .metric_expr = 0, + .metric_name = 0, }, };@@ -181,6 +188,11 @@ struct pmu_events_table {const struct pmu_event *entries; }; +/* Struct used to make the PMU metric table implementation opaque to callers. */ +struct pmu_metrics_table { +const struct pmu_metric *entries; +}; +/* * Map a CPU to its table of PMU events. The CPU is identified by the * cpuid field, which is an arch-specific identifier for the CPU. @@ -192,7 +204,8 @@ struct pmu_events_table {struct pmu_events_map {const char *arch;const char *cpuid; -const struct pmu_events_table table; +const struct pmu_events_table event_table; +const struct pmu_metrics_table metric_table; };/* @@ -203,12 +216,14 @@ static const struct pmu_events_map pmu_events_map[] = { { .arch = "testarch", .cpuid = "testcpu", - .table = { pme_test_soc_cpu }, + .event_table = { pmu_events__test_soc_cpu }, + .metric_table = { pmu_metrics__test_soc_cpu }, }, { .arch = 0, .cpuid = 0, - .table = { 0 }, + .event_table = { 0 }, + .metric_table = { 0 }, }, };@@ -254,9 +269,7 @@ static const struct pmu_sys_events pmu_sys_event_tables[] = {int pmu_events_table_for_each_event (const struct pmu_events_table *table,pmu_event_iter_fn fn,void *data) { -for (const struct pmu_event *pe = &table->entries[0 ]; - pe->name || pe->metric_group || pe->metric_name; - pe++) { +for (const struct pmu_event *pe = &table->entries[0 ]; pe->name; pe++) {int ret =fn (pe, table, data);if (ret)@@ -265,6 +278,21 @@ int pmu_events_table_for_each_event(const struct pmu_events_table *table, pmu_evreturn 0 ; }
Quoted 7 lines
tools/perf/pmu-events/empty-pmu-events.c int pmu_events_table_for_each_event(const struct pmu_events_table *table, pmu_ev · line 284
Quoted 4 lines
tools/perf/pmu-events/empty-pmu-events.c +4
-0+int pmu_events_table_for_each_metric (const struct pmu_events_table *etable,pmu_metric_iter_fn fn, +void *data) +{ +struct pmu_metrics_table *table = (struct pmu_metrics_table *)etable;This is odd. Would the change be too big if we combined patches to avoid casting like this? I assume that this is going away later...
Yep. Done this way to minimize the patch sizes.
Quoted 4 lines
tools/perf/pmu-events/empty-pmu-events.c int pmu_events_table_for_each_event(const struct pmu_events_table *table, pmu_ev · line 286
Quoted 2 lines
tools/perf/pmu-events/empty-pmu-events.c +2
-0+ + for (const struct pmu_metric *pm = &table->entries[0]; pm->metric_group || pm->metric_name;Is it valid to only have
metric_groupormetric_nameset?
Agreed. I think using the expr is the best solution, albeit temporary in the patch series. Fwiw, it was done this way as the logic was moved from pmu_events_table_for_each_event.
Quoted 42 lines
Quoted 42 lines
tools/perf/pmu-events/empty-pmu-events.c +18
-3+ pm++) { +int ret =fn (pm, etable, data); + +if (ret) +return ret; + } +return 0 ; +} +const struct pmu_events_table *perf_pmu__find_table (struct perf_pmu *pmu) {const struct pmu_events_table *table =NULL ;@@ -285,7 +313,7 @@ const struct pmu_events_table *perf_pmu__find_table(struct perf_pmu *pmu)break ;if (!strcmp_cpuid_str (map->cpuid, cpuid)) { - table = &map->table; + table = &map->event_table;break ; } }@@ -299,7 +327,7 @@ const struct pmu_events_table *find_core_events_table(const char *arch, const ch tables->arch; tables++) {if (!strcmp (tables->arch, arch) && !strcmp_cpuid_str (tables->cpuid, cpuid)) -return &tables->table; +return &tables->event_table; }return NULL ; }@@ -309,7 +337,21 @@ int pmu_for_each_core_event(pmu_event_iter_fn fn, void *data)for (const struct pmu_events_map *tables = &pmu_events_map[0 ]; tables->arch; tables++) { -int ret =pmu_events_table_for_each_event (&tables->table, fn, data); +int ret =pmu_events_table_for_each_event (&tables->event_table, fn, data); + +if (ret) +return ret; + } +return 0 ; +}
Quoted 8 lines
tools/perf/pmu-events/empty-pmu-events.c int pmu_for_each_core_event(pmu_event_iter_fn fn, void *data) · line 352
Quoted 6 lines
tools/perf/pmu-events/empty-pmu-events.c +6
-0+ +int pmu_for_each_core_metric (pmu_metric_iter_fn fn,void *data) +{ +for (const struct pmu_events_map *tables = &pmu_events_map[0 ]; + tables->arch; + tables++) {again, you can combine here
Right, done this way to reduce patch size. Later patch splits the tables.
Quoted 158 lines
Quoted 158 lines
tools/perf/pmu-events/empty-pmu-events.c +7
-0+int ret =pmu_events_table_for_each_metric ( + (const struct pmu_events_table *)&tables->metric_table, fn, data);if (ret)return ret;@@ -340,3 +382,8 @@ int pmu_for_each_sys_event(pmu_event_iter_fn fn, void *data) }return 0 ; } + +int pmu_for_each_sys_metric (pmu_metric_iter_fn fn__maybe_unused ,void *data__maybe_unused ) +{ +return 0 ; +}tools/perf/pmu-events/jevents.py +76
-6@@ -37,6 +37,11 @@ _json_event_attributes = [ 'metric_constraint', 'metric_expr', 'long_desc' ] +# Attributes that are in pmu_metric rather than pmu_event. +_json_metric_attributes = [ + 'metric_name', 'metric_group', 'metric_constraint', 'metric_expr', 'desc', + 'long_desc', 'unit', 'compat', 'aggr_mode' +] def removesuffix(s: str, suffix: str) -> str: """Remove the suffix from a string@@ -562,12 +567,15 @@ static const struct pmu_sys_events pmu_sys_event_tables[] = { \t}, }; -static void decompress(int offset, struct pmu_event *pe) +static void decompress_event(int offset, struct pmu_event *pe) { \tconst char *p = &big_c_string[offset]; """) for attr in _json_event_attributes: - _args.output_file.write(f""" + if attr in _json_metric_attributes and 'metric_' in attr: + _args.output_file.write(f'\n\t/* Skip {attr} */\n') + else: + _args.output_file.write(f""" \tpe->{attr} = (*p == '\\0' ? NULL : p); """) if attr == _json_event_attributes[-1]:@@ -575,6 +583,22 @@ static void decompress(int offset, struct pmu_event *pe) _args.output_file.write('\twhile (*p++);') _args.output_file.write("""} +static void decompress_metric(int offset, struct pmu_metric *pm) +{ +\tconst char *p = &big_c_string[offset]; +""") + for attr in _json_event_attributes: + if attr in _json_metric_attributes: + _args.output_file.write(f""" +\tpm->{attr} = (*p == '\\0' ? NULL : p); +""") + else: + _args.output_file.write(f'\n\t/* Skip {attr} */\n') + if attr == _json_event_attributes[-1]: + continue + _args.output_file.write('\twhile (*p++);') + _args.output_file.write("""} + int pmu_events_table_for_each_event(const struct pmu_events_table *table, pmu_event_iter_fn fn, void *data)@@ -583,10 +607,30 @@ int pmu_events_table_for_each_event(const struct pmu_events_table *table, struct pmu_event pe; int ret; - decompress(table->entries[i].offset, &pe); - ret = fn(&pe, table, data); - if (ret) - return ret; + decompress_event(table->entries[i].offset, &pe); + if (pe.name) { + ret = fn(&pe, table, data); + if (ret) + return ret; + } + } + return 0; +} + +int pmu_events_table_for_each_metric(const struct pmu_events_table *table, + pmu_metric_iter_fn fn, + void *data) +{ + for (size_t i = 0; i < table->length; i++) { + struct pmu_metric pm; + int ret; + + decompress_metric(table->entries[i].offset, &pm); + if (pm.metric_name) { + ret = fn(&pm, table, data); + if (ret) + return ret; + } } return 0; }@@ -642,6 +686,19 @@ int pmu_for_each_core_event(pmu_event_iter_fn fn, void *data) return 0; } +int pmu_for_each_core_metric(pmu_metric_iter_fn fn, void *data) +{ + for (const struct pmu_events_map *tables = &pmu_events_map[0]; + tables->arch; + tables++) { + int ret = pmu_events_table_for_each_metric(&tables->table, fn, data); + + if (ret) + return ret; + } + return 0; +} + const struct pmu_events_table *find_sys_events_table(const char *name) { for (const struct pmu_sys_events *tables = &pmu_sys_event_tables[0];@@ -665,6 +722,19 @@ int pmu_for_each_sys_event(pmu_event_iter_fn fn, void *data) } return 0; } + +int pmu_for_each_sys_metric(pmu_metric_iter_fn fn, void *data) +{ + for (const struct pmu_sys_events *tables = &pmu_sys_event_tables[0]; + tables->name; + tables++) { + int ret = pmu_events_table_for_each_metric(&tables->table, fn, data); + + if (ret) + return ret; + } + return 0; +} """)tools/perf/pmu-events/pmu-events.h +3
-1@@ -23,11 +23,19 @@ struct pmu_event {const char *unit;const char *perpkg;const char *aggr_mode; -const char *metric_expr; +const char *deprecated; +}; +
Quoted 8 lines
tools/perf/pmu-events/pmu-events.h struct pmu_event { · line 29
Quoted 1 line
tools/perf/pmu-events/pmu-events.h +1
-0+struct pmu_metric {In terms of formatting patches, it could have been better to first introdue struct
pmu_metricbeing same as structpmu_event, and make the changes of /s/structpmu_eventpe;/structpmu_eventpm;/, and then secondly the meaningful change to get rid of the inappropiate members in structpmu_metricand structpmu_event. It's just harder to review it all together.
Ok, I'll switch it this way in v3. Having done v3 I've now made this 3 changes following this suggestion, I've also folded in the empty pmu metrics fixes. Fwiw, I'd done the non empty case and then tried to retro fit. I think we should probably change the empty-pmu-events.c and have a way to generate from jevents.py, then check in the generated file to avoid the python dependency.
Quoted 8 lines
tools/perf/pmu-events/pmu-events.h struct pmu_event { · line 32
Quoted 4 lines
tools/perf/pmu-events/pmu-events.h +1
-1const char *metric_name;const char *metric_group; -const char *deprecated; +const char *metric_expr;I would prefer the shorter name "expr" (as opposed to "
metric_expr") as we know it's a metric struct already. But that does not seem to be coding style, so ...
Agreed. Let's do it as a later cleanup.
Quoted 40 lines
Quoted 40 lines
tools/perf/pmu-events/pmu-events.h +13
-0+const char *unit; +const char *compat; +const char *aggr_mode;const char *metric_constraint; +const char *desc; +const char *long_desc; };struct pmu_events_table ;@@ -36,14 +44,22 @@ typedef int (*pmu_event_iter_fn)(const struct pmu_event *pe,const struct pmu_events_table *table, void *data); +typedef int (*pmu_metric_iter_fn )(const struct pmu_metric *pm, +const struct pmu_events_table *table, +void *data); +int pmu_events_table_for_each_event (const struct pmu_events_table *table,pmu_event_iter_fn fn,void *data); +int pmu_events_table_for_each_metric (const struct pmu_events_table *table,pmu_metric_iter_fn fn, +void *data);const struct pmu_events_table *perf_pmu__find_table (struct perf_pmu *pmu);const struct pmu_events_table *find_core_events_table (const char *arch,const char *cpuid);int pmu_for_each_core_event (pmu_event_iter_fn fn,void *data); +int pmu_for_each_core_metric (pmu_metric_iter_fn fn,void *data);const struct pmu_events_table *find_sys_events_table (const char *name);int pmu_for_each_sys_event (pmu_event_iter_fn fn,void *data); +int pmu_for_each_sys_metric (pmu_metric_iter_fn fn,void *data);#endif tools/perf/tests/pmu-events.c +0
-0@@ -337,36 +337,12 @@ static int compare_pmu_events(const struct pmu_event *e1, const struct pmu_eventreturn -1 ; }
Quoted 6 lines
tools/perf/tests/pmu-events.c static int compare_pmu_events(const struct pmu_event *e1, const struct pmu_event · line 343
Quoted 4 lines
tools/perf/tests/pmu-events.c +0
-4-if (!is_same (e1->metric_expr, e2->metric_expr)) { -pr_debug2 ("testing event e1 %s: mismatched metric_expr, %s vs %s\n" , - e1->name, e1->metric_expr, e2->metric_expr); -return -1 ;is there something new to replace this (I didn't check)?
The variables no longer exist and in the original test they weren't ever defined.
Quoted 245 lines
Quoted 245 lines
tools/perf/tests/pmu-events.c +16
-42- } - -if (!is_same (e1->metric_name, e2->metric_name)) { -pr_debug2 ("testing event e1 %s: mismatched metric_name, %s vs %s\n" , - e1->name, e1->metric_name, e2->metric_name); -return -1 ; - } - -if (!is_same (e1->metric_group, e2->metric_group)) { -pr_debug2 ("testing event e1 %s: mismatched metric_group, %s vs %s\n" , - e1->name, e1->metric_group, e2->metric_group); -return -1 ; - } -if (!is_same (e1->deprecated, e2->deprecated)) {pr_debug2 ("testing event e1 %s: mismatched deprecated, %s vs %s\n" , e1->name, e1->deprecated, e2->deprecated);return -1 ; } -if (!is_same (e1->metric_constraint, e2->metric_constraint)) { -pr_debug2 ("testing event e1 %s: mismatched metric_constant, %s vs %s\n" , - e1->name, e1->metric_constraint, e2->metric_constraint); -return -1 ; - } -return 0 ; }@@ -432,9 +408,6 @@ static int test__pmu_event_table_core_callback(const struct pmu_event *pe,struct perf_pmu_test_event const **test_event_table;bool found = false; -if (!pe->name) -return 0 ; -if (pe->pmu) test_event_table = &uncore_events[0 ];else @@ -840,7 +813,7 @@ struct metric {struct metric_ref metric_ref; }; -static int test__parsing_callback (const struct pmu_event *pe,const struct pmu_events_table *table, +static int test__parsing_callback (const struct pmu_metric *pm,const struct pmu_events_table *table,void *data) {int *failures = data;@@ -854,10 +827,10 @@ static int test__parsing_callback(const struct pmu_event *pe, const struct pmu_e };int err =0 ; -if (!pe->metric_expr) +if (!pm->metric_expr)return 0 ; -pr_debug ("Found metric '%s'\n" , pe->metric_name); +pr_debug ("Found metric '%s'\n" , pm->metric_name); (*failures)++;/* @@ -877,14 +850,14 @@ static int test__parsing_callback(const struct pmu_event *pe, const struct pmu_eperf_evlist__set_maps (&evlist->core, cpus,NULL );runtime_stat__init (&st); - err =metricgroup__parse_groups_test (evlist, table, pe->metric_name, + err =metricgroup__parse_groups_test (evlist, table, pm->metric_name, false, false, &metric_events);if (err) { -if (!strcmp (pe->metric_name,"M1" ) || !strcmp (pe->metric_name,"M2" ) || - !strcmp (pe->metric_name,"M3" )) { +if (!strcmp (pm->metric_name,"M1" ) || !strcmp (pm->metric_name,"M2" ) || + !strcmp (pm->metric_name,"M3" )) { (*failures)--; -pr_debug ("Expected broken metric %s skipping\n" , pe->metric_name); +pr_debug ("Expected broken metric %s skipping\n" , pm->metric_name); err =0 ; } goto out_err;@@ -912,7 +885,7 @@ static int test__parsing_callback(const struct pmu_event *pe, const struct pmu_estruct metric_expr *mexp; list_for_each_entry (mexp, &me->head, nd) { -if (strcmp (mexp->metric_name, pe->metric_name)) +if (strcmp (mexp->metric_name, pm->metric_name))continue ;pr_debug ("Result %f\n" ,test_generic_metric (mexp,0 , &st)); err =0 ;@@ -921,11 +894,11 @@ static int test__parsing_callback(const struct pmu_event *pe, const struct pmu_e } } } -pr_debug ("Didn't find parsed metric %s" , pe->metric_name); +pr_debug ("Didn't find parsed metric %s" , pm->metric_name); err =1 ; out_err:if (err) -pr_debug ("Broken metric %s\n" , pe->metric_name); +pr_debug ("Broken metric %s\n" , pm->metric_name);/* ... cleanup. */ metricgroup__rblist_exit (&metric_events);@@ -941,8 +914,8 @@ static int test__parsing(struct test_suite *test __maybe_unused, {int failures =0 ; -pmu_for_each_core_event (test__parsing_callback, &failures); -pmu_for_each_sys_event (test__parsing_callback, &failures); +pmu_for_each_core_metric (test__parsing_callback, &failures); +pmu_for_each_sys_metric (test__parsing_callback, &failures);return failures ==0 ?TEST_OK :TEST_FAIL ; }@@ -1021,14 +994,11 @@ static int metric_parse_fake(const char *metric_name, const char *str)return ret; } -static int test__parsing_fake_callback (const struct pmu_event *pe, +static int test__parsing_fake_callback (const struct pmu_metric *pm,const struct pmu_events_table *table__maybe_unused ,void *data__maybe_unused ) { -if (!pe->metric_expr) -return 0 ; - -return metric_parse_fake (pe->metric_name, pe->metric_expr); +return metric_parse_fake (pm->metric_name, pm->metric_expr); }/* @@ -1047,11 +1017,11 @@ static int test__parsing_fake(struct test_suite *test __maybe_unused,return err; } - err =pmu_for_each_core_event (test__parsing_fake_callback,NULL ); + err =pmu_for_each_core_metric (test__parsing_fake_callback,NULL );if (err)return err; -return pmu_for_each_sys_event (test__parsing_fake_callback,NULL ); +return pmu_for_each_sys_metric (test__parsing_fake_callback,NULL ); }static struct test_case pmu_events_tests[] = {tools/perf/util/metricgroup.c +16
-16@@ -167,14 +167,14 @@ static void metricgroup___watchdog_constraint_hint(const char *name, bool foot) " echo 1 > /proc/sys/kernel/nmi_watchdog\n"); } -static bool metricgroup__has_constraint (const struct pmu_event *pe) +static bool metricgroup__has_constraint (const struct pmu_metric *pm) { -if (!pe->metric_constraint) +if (!pm->metric_constraint)return false; -if (!strcmp (pe->metric_constraint,"NO_NMI_WATCHDOG" ) && +if (!strcmp (pm->metric_constraint,"NO_NMI_WATCHDOG" ) &&sysctl__nmi_watchdog_enabled ()) { -metricgroup___watchdog_constraint_hint (pe->metric_name, false); +metricgroup___watchdog_constraint_hint (pm->metric_name, false);return true; }@@ -193,7 +193,7 @@ static void metric__free(struct metric *m)free (m); } -static struct metric *metric__new(const struct pmu_event *pe, +static struct metric *metric__new(const struct pmu_metric *pm, const char *modifier, bool metric_no_group, int runtime,@@ -210,15 +210,15 @@ static struct metric *metric__new(const struct pmu_event *pe,if (!m->pctx) goto out_err; - m->metric_name = pe->metric_name; + m->metric_name = pm->metric_name; m->modifier =NULL ;if (modifier) { m->modifier =strdup (modifier);if (!m->modifier) goto out_err; } - m->metric_expr = pe->metric_expr; - m->metric_unit = pe->unit; + m->metric_expr = pm->metric_expr; + m->metric_unit = pm->unit; m->pctx->sctx.user_requested_cpu_list =NULL ;if (user_requested_cpu_list) { m->pctx->sctx.user_requested_cpu_list =strdup (user_requested_cpu_list);@@ -227,7 +227,7 @@ static struct metric *metric__new(const struct pmu_event *pe, } m->pctx->sctx.runtime = runtime; m->pctx->sctx.system_wide = system_wide; - m->has_constraint = metric_no_group ||metricgroup__has_constraint (pe); + m->has_constraint = metric_no_group ||metricgroup__has_constraint (pm); m->metric_refs =NULL ; m->evlist =NULL ;@@ -348,10 +348,10 @@ static bool match_metric(const char *n, const char *list)return false; } -static bool match_pe_metric (const struct pmu_event *pe,const char *metric) +static bool match_pm_metric (const struct pmu_metric *pm,const char *metric) { -return match_metric (pe->metric_group, metric) || -match_metric (pe->metric_name, metric); +return match_metric (pm->metric_group, metric) || +match_metric (pm->metric_name, metric); }/** struct mep - RB-tree node for building printing information. */ @@ -420,13 +420,13 @@ static struct mep *mep_lookup(struct rblist *groups, const char *metric_group,return NULL ; } -static int metricgroup__add_to_mep_groups (const struct pmu_event *pe, +static int metricgroup__add_to_mep_groups (const struct pmu_metric *pm,struct rblist *groups) {const char *g;char *omg, *mg; - mg =strdup (pe->metric_group ?:"No_group" ); + mg =strdup (pm->metric_group ?:"No_group" );if (!mg)return -ENOMEM ; omg = mg;@@ -435,15 +435,15 @@ static int metricgroup__add_to_mep_groups(const struct pmu_event *pe, g =skip_spaces (g);if (strlen (g)) - me =mep_lookup (groups, g, pe->metric_name); + me =mep_lookup (groups, g, pm->metric_name);else - me =mep_lookup (groups,"No_group" , pe->metric_name); + me =mep_lookup (groups,"No_group" , pm->metric_name);
Quoted 10 lines
tools/perf/util/metricgroup.c static int metricgroup__add_to_mep_groups(const struct pmu_event *pe, · line 445
Quoted 8 lines
tools/perf/util/metricgroup.c +3
-4if (me) { - me->metric_desc = pe->desc; - me->metric_long_desc = pe->long_desc; - me->metric_expr = pe->metric_expr; - me->metric_unit = pe->unit; + me->metric_desc = pm->desc; + me->metric_long_desc = pm->long_desc; + me->metric_expr = pm->metric_expr;could struct
pmu_metricbe a sub-struct in struct mep?
struct mep is only used for 'perf list' and has a subset of the information in struct pmu_metric. The metric_group is also mutated. I like the idea of merging them and having a single assignment here, but it would add some overhead for limited gain.
Quoted 244 lines
Quoted 244 lines
tools/perf/util/metricgroup.c +37
-39+ me->metric_unit = pm->unit; } }free (omg);@@ -452,40 +452,37 @@ static int metricgroup__add_to_mep_groups(const struct pmu_event *pe, }struct metricgroup_iter_data { -pmu_event_iter_fn fn; +pmu_metric_iter_fn fn;void *data; }; -static int metricgroup__sys_event_iter (const struct pmu_event *pe, +static int metricgroup__sys_event_iter (const struct pmu_metric *pm,const struct pmu_events_table *table,void *data) {struct metricgroup_iter_data *d = data;struct perf_pmu *pmu =NULL ; -if (!pe->metric_expr || !pe->compat) +if (!pm->metric_expr || !pm->compat)return 0 ;while ((pmu =perf_pmu__scan (pmu))) { -if (!pmu->id ||strcmp (pmu->id, pe->compat)) +if (!pmu->id ||strcmp (pmu->id, pm->compat))continue ; -return d->fn (pe, table, d->data); +return d->fn (pm, table, d->data); }return 0 ; } -static int metricgroup__add_to_mep_groups_callback (const struct pmu_event *pe, +static int metricgroup__add_to_mep_groups_callback (const struct pmu_metric *pm,const struct pmu_events_table *table__maybe_unused ,void *vdata) {struct rblist *groups = vdata; -if (!pe->metric_name) -return 0 ; - -return metricgroup__add_to_mep_groups (pe, groups); +return metricgroup__add_to_mep_groups (pm, groups); }void metricgroup__print (const struct print_callbacks *print_cb,void *print_state)@@ -500,16 +497,16 @@ void metricgroup__print(const struct print_callbacks *print_cb, void *print_stat groups.node_delete = mep_delete; table =pmu_events_table__find ();if (table) { -pmu_events_table_for_each_event (table, - metricgroup__add_to_mep_groups_callback, - &groups); +pmu_events_table_for_each_metric (table, + metricgroup__add_to_mep_groups_callback, + &groups); } {struct metricgroup_iter_data data = { .fn = metricgroup__add_to_mep_groups_callback, .data = &groups, }; -pmu_for_each_sys_event (metricgroup__sys_event_iter, &data); +pmu_for_each_sys_metric (metricgroup__sys_event_iter, &data); }for (node =rb_first_cached (&groups.entries); node; node = next) {@@ -743,7 +740,7 @@ static int metricgroup__build_event_string(struct strbuf *events,#undef RETURN_IF_NON_ZERO } -int __weak arch_get_runtimeparam (const struct pmu_event *pe__maybe_unused ) +int __weak arch_get_runtimeparam (const struct pmu_metric *pm__maybe_unused ) {return 1 ; }@@ -773,10 +770,10 @@ struct metricgroup_add_iter_data {static bool metricgroup__find_metric (const char *metric,const struct pmu_events_table *table, -struct pmu_event *pe); +struct pmu_metric *pm); static int add_metric(struct list_head *metric_list, - const struct pmu_event *pe, + const struct pmu_metric *pm, const char *modifier, bool metric_no_group, const char *user_requested_cpu_list,@@ -816,10 +813,10 @@ static int resolve_metric(struct list_head *metric_list,size_t bkt;struct to_resolve {/* The metric to resolve. */ -struct pmu_event pe; +struct pmu_metric pm;/* * The key in the IDs map, this may differ from in case, -* etc. from pe->metric_name. +* etc. from pm->metric_name. */ const char *key; } *pending =NULL ;@@ -830,15 +827,15 @@ static int resolve_metric(struct list_head *metric_list,* the pending array. */ hashmap__for_each_entry (root_metric->pctx->ids, cur, bkt) { -struct pmu_event pe; +struct pmu_metric pm; -if (metricgroup__find_metric (cur->pkey, table, &pe)) { +if (metricgroup__find_metric (cur->pkey, table, &pm)) { pending =realloc (pending, (pending_cnt +1 ) *sizeof (struct to_resolve ));if (!pending)return -ENOMEM ; -memcpy (&pending[pending_cnt].pe, &pe,sizeof (pe)); +memcpy (&pending[pending_cnt].pm, &pm,sizeof (pm)); pending[pending_cnt].key = cur->pkey; pending_cnt++; }@@ -853,7 +850,7 @@ static int resolve_metric(struct list_head *metric_list,* context. */ for (i =0 ; i < pending_cnt; i++) { - ret =add_metric (metric_list, &pending[i].pe, modifier, metric_no_group, + ret =add_metric (metric_list, &pending[i].pm, modifier, metric_no_group, user_requested_cpu_list, system_wide, root_metric, visited, table);if (ret)@@ -867,7 +864,7 @@ static int resolve_metric(struct list_head *metric_list,/** * __add_metric - Add a metric to metric_list. * @metric_list: The list the metric is added to. -* @pe: The pmu_event containing the metric to be added. +* @pm: The pmu_metric containing the metric to be added. * @modifier: if non-null event modifiers like "u". * @metric_no_group: Should events written to events be grouped "{}" or * global. Grouping is the default but due to multiplexing the @@ -884,7 +881,7 @@ static int resolve_metric(struct list_head *metric_list,* architecture perf is running upon. */ static int __add_metric(struct list_head *metric_list, - const struct pmu_event *pe, + const struct pmu_metric *pm, const char *modifier, bool metric_no_group, int runtime,@@ -898,13 +895,13 @@ static int __add_metric(struct list_head *metric_list,int ret;bool is_root = !root_metric;struct visited_metric visited_node = { - .name = pe->metric_name, + .name = pm->metric_name, .parent = visited, };for (vm = visited; vm; vm = vm->parent) { -if (!strcmp (pe->metric_name, vm->name)) { -pr_err ("failed: recursion detected for %s\n" , pe->metric_name); +if (!strcmp (pm->metric_name, vm->name)) { +pr_err ("failed: recursion detected for %s\n" , pm->metric_name);return -1 ; } }@@ -914,7 +911,7 @@ static int __add_metric(struct list_head *metric_list,* This metric is the root of a tree and may reference other * metrics that are added recursively. */ - root_metric =metric__new (pe, modifier, metric_no_group, runtime, + root_metric =metric__new (pm, modifier, metric_no_group, runtime, user_requested_cpu_list, system_wide);if (!root_metric)return -ENOMEM ;@@ -929,7 +926,7 @@ static int __add_metric(struct list_head *metric_list,*/ if (root_metric->metric_refs) {for (; root_metric->metric_refs[cnt].metric_name; cnt++) { -if (!strcmp (pe->metric_name, +if (!strcmp (pm->metric_name, root_metric->metric_refs[cnt].metric_name))return 0 ; }@@ -947,8 +944,8 @@ static int __add_metric(struct list_head *metric_list,* need to change them, so there's no need to create * our own copy. */ - root_metric->metric_refs[cnt].metric_name = pe->metric_name; - root_metric->metric_refs[cnt].metric_expr = pe->metric_expr; + root_metric->metric_refs[cnt].metric_name = pm->metric_name; + root_metric->metric_refs[cnt].metric_expr = pm->metric_expr;/* Null terminate array. */ root_metric->metric_refs[cnt+1 ].metric_name =NULL ;@@ -959,7 +956,7 @@ static int __add_metric(struct list_head *metric_list,* For both the parent and referenced metrics, we parse * all the metric's IDs and add it to the root context. */ -if (expr__find_ids (pe->metric_expr,NULL , root_metric->pctx) <0 ) { +if (expr__find_ids (pm->metric_expr,NULL , root_metric->pctx) <0 ) {/* Broken metric. */ ret = -EINVAL ; }else {@@ -981,37 +978,37 @@ static int __add_metric(struct list_head *metric_list,struct metricgroup__find_metric_data {const char *metric; -struct pmu_event *pe; +struct pmu_metric *pm; }; -static int metricgroup__find_metric_callback (const struct pmu_event *pe, +static int metricgroup__find_metric_callback (const struct pmu_metric *pm,const struct pmu_events_table *table__maybe_unused ,void *vdata) {struct metricgroup__find_metric_data *data = vdata; -if (!match_metric (pe->metric_name, data->metric)) +if (!match_metric (pm->metric_name, data->metric))return 0 ; -memcpy (data->pe, pe,sizeof (*pe)); +memcpy (data->pm, pm,sizeof (*pm));return 1 ; }static bool metricgroup__find_metric (const char *metric,const struct pmu_events_table *table, -struct pmu_event *pe) +struct pmu_metric *pm) {struct metricgroup__find_metric_data data = { .metric = metric, - .pe = pe, + .pm = pm, };
Quoted 6 lines
tools/perf/util/metricgroup.c static int __add_metric(struct list_head *metric_list, · line 1007
Quoted 3 lines
tools/perf/util/metricgroup.c +1
-1-return pmu_events_table_for_each_event (table, metricgroup__find_metric_callback, &data) +return pmu_events_table_for_each_metric (table, metricgroup__find_metric_callback, &data) ? true : false;comment on pre-existing code: since it returns bool, I don't think the ternary operator is required at all. There are many instances of this.
There'd be an implicit int -> bool conversion, which is why this exists. Perhaps changing the return type to int as the int pattern is ubiquitous.
Thanks,
Ian
Quoted 417 lines
Quoted 416 lines
tools/perf/util/metricgroup.c +22
-22} static int add_metric(struct list_head *metric_list, - const struct pmu_event *pe, + const struct pmu_metric *pm, const char *modifier, bool metric_no_group, const char *user_requested_cpu_list,@@ -1022,16 +1019,16 @@ static int add_metric(struct list_head *metric_list, {int ret =0 ; -pr_debug ("metric expr %s for %s\n" , pe->metric_expr, pe->metric_name); +pr_debug ("metric expr %s for %s\n" , pm->metric_expr, pm->metric_name); -if (!strstr (pe->metric_expr,"?" )) { - ret =__add_metric (metric_list, pe, modifier, metric_no_group,0 , +if (!strstr (pm->metric_expr,"?" )) { + ret =__add_metric (metric_list, pm, modifier, metric_no_group,0 , user_requested_cpu_list, system_wide, root_metric, visited, table); }else {int j, count; - count =arch_get_runtimeparam (pe); + count =arch_get_runtimeparam (pm);/* This loop is added to create multiple * events depend on count value and add @@ -1039,7 +1036,7 @@ static int add_metric(struct list_head *metric_list,*/ for (j =0 ; j < count && !ret; j++) - ret =__add_metric (metric_list, pe, modifier, metric_no_group, j, + ret =__add_metric (metric_list, pm, modifier, metric_no_group, j, user_requested_cpu_list, system_wide, root_metric, visited, table); }@@ -1047,17 +1044,17 @@ static int add_metric(struct list_head *metric_list,return ret; } -static int metricgroup__add_metric_sys_event_iter (const struct pmu_event *pe, +static int metricgroup__add_metric_sys_event_iter (const struct pmu_metric *pm,const struct pmu_events_table *table__maybe_unused ,void *data) {struct metricgroup_add_iter_data *d = data;int ret; -if (!match_pe_metric (pe, d->metric_name)) +if (!match_pm_metric (pm, d->metric_name))return 0 ; - ret =add_metric (d->metric_list, pe, d->modifier, d->metric_no_group, + ret =add_metric (d->metric_list, pm, d->modifier, d->metric_no_group, d->user_requested_cpu_list, d->system_wide, d->root_metric, d->visited, d->table);if (ret)@@ -1107,19 +1104,19 @@ struct metricgroup__add_metric_data {bool has_match; }; -static int metricgroup__add_metric_callback (const struct pmu_event *pe, +static int metricgroup__add_metric_callback (const struct pmu_metric *pm,const struct pmu_events_table *table,void *vdata) {struct metricgroup__add_metric_data *data = vdata;int ret =0 ; -if (pe->metric_expr && - (match_metric (pe->metric_group, data->metric_name) || -match_metric (pe->metric_name, data->metric_name))) { +if (pm->metric_expr && + (match_metric (pm->metric_group, data->metric_name) || +match_metric (pm->metric_name, data->metric_name))) { data->has_match = true; - ret =add_metric (data->list, pe, data->modifier, data->metric_no_group, + ret =add_metric (data->list, pm, data->modifier, data->metric_no_group, data->user_requested_cpu_list, data->system_wide,/*root_metric=*/ NULL ,/*visited_metrics=*/ NULL , table); }@@ -1166,8 +1163,8 @@ static int metricgroup__add_metric(const char *metric_name, const char *modifier* Iterate over all metrics seeing if metric matches either the * name or group. When it does add the metric to the list. */ - ret =pmu_events_table_for_each_event (table, metricgroup__add_metric_callback, - &data); + ret =pmu_events_table_for_each_metric (table, metricgroup__add_metric_callback, + &data);if (ret) goto out;@@ -1189,7 +1186,7 @@ static int metricgroup__add_metric(const char *metric_name, const char *modifier }, }; -pmu_for_each_sys_event (metricgroup__sys_event_iter, &data); +pmu_for_each_sys_metric (metricgroup__sys_event_iter, &data); }/* End of pmu events. */ if (!has_match)@@ -1603,16 +1600,16 @@ int metricgroup__parse_groups_test(struct evlist *evlist, &perf_pmu__fake, metric_events, table); } -static int metricgroup__has_metric_callback (const struct pmu_event *pe, +static int metricgroup__has_metric_callback (const struct pmu_metric *pm,const struct pmu_events_table *table__maybe_unused ,void *vdata) {const char *metric = vdata; -if (!pe->metric_expr) +if (!pm->metric_expr)return 0 ; -if (match_metric (pe->metric_name, metric)) +if (match_metric (pm->metric_name, metric))return 1 ;return 0 ;@@ -1625,8 +1622,8 @@ bool metricgroup__has_metric(const char *metric)if (!table)return false; -return pmu_events_table_for_each_event (table, metricgroup__has_metric_callback, - (void *)metric) ? true : false; +return pmu_events_table_for_each_metric (table, metricgroup__has_metric_callback, + (void *)metric) ? true : false; } int metricgroup__copy_metric_events(struct evlist *evlist, struct cgroup *cgrp,tools/perf/util/metricgroup.h +1
-1@@ -81,7 +81,7 @@ int metricgroup__parse_groups_test(struct evlist *evlist,void metricgroup__print (const struct print_callbacks *print_cb,void *print_state);bool metricgroup__has_metric (const char *metric); -int arch_get_runtimeparam (const struct pmu_event *pe__maybe_unused ); +int arch_get_runtimeparam (const struct pmu_metric *pm);void metricgroup__rblist_exit (struct rblist *metric_events); int metricgroup__copy_metric_events(struct evlist *evlist, struct cgroup *cgrp,tools/perf/util/parse-events.c +0
-2@@ -1570,8 +1570,6 @@ int parse_events_add_pmu(struct parse_events_state *parse_state, evsel->scale = info.scale; evsel->per_pkg = info.per_pkg; evsel->snapshot = info.snapshot; - evsel->metric_expr = info.metric_expr; - evsel->metric_name = info.metric_name;return 0 ; }tools/perf/util/pmu.c +3
-32@@ -283,10 +283,6 @@ static void perf_pmu_update_alias(struct perf_pmu_alias *old,perf_pmu_assign_str (old->name,"long_desc" , &old->long_desc, &newalias->long_desc);perf_pmu_assign_str (old->name,"topic" , &old->topic, &newalias->topic); -perf_pmu_assign_str (old->name,"metric_expr" , &old->metric_expr, - &newalias->metric_expr); -perf_pmu_assign_str (old->name,"metric_name" , &old->metric_name, - &newalias->metric_name);perf_pmu_assign_str (old->name,"value" , &old->str, &newalias->str); old->scale = newalias->scale; old->per_pkg = newalias->per_pkg;@@ -302,8 +298,6 @@ void perf_pmu_free_alias(struct perf_pmu_alias *newalias)zfree (&newalias->long_desc);zfree (&newalias->topic);zfree (&newalias->str); -zfree (&newalias->metric_expr); -zfree (&newalias->metric_name);zfree (&newalias->pmu_name);parse_events_terms__purge (&newalias->terms);free (newalias);@@ -340,16 +334,13 @@ static int __perf_pmu__new_alias(struct list_head *list, char *dir, char *name,int num;char newval[256 ];char *long_desc =NULL , *topic =NULL , *unit =NULL , *perpkg =NULL , - *metric_expr =NULL , *metric_name =NULL , *deprecated =NULL , - *pmu_name =NULL ; + *deprecated =NULL , *pmu_name =NULL ;if (pe) { long_desc = (char *)pe->long_desc; topic = (char *)pe->topic; unit = (char *)pe->unit; perpkg = (char *)pe->perpkg; - metric_expr = (char *)pe->metric_expr; - metric_name = (char *)pe->metric_name; deprecated = (char *)pe->deprecated; pmu_name = (char *)pe->pmu; }@@ -404,8 +395,6 @@ static int __perf_pmu__new_alias(struct list_head *list, char *dir, char *name,perf_pmu__parse_snapshot (alias, dir, name); } - alias->metric_expr = metric_expr ?strdup (metric_expr) :NULL ; - alias->metric_name = metric_name ?strdup (metric_name):NULL ; alias->desc = desc ?strdup (desc) :NULL ; alias->long_desc = long_desc ?strdup (long_desc) : desc ?strdup (desc) :NULL ;@@ -822,9 +811,6 @@ static int pmu_add_cpu_aliases_map_callback(const struct pmu_event *pe,struct pmu_add_cpu_aliases_map_data *data = vdata;const char *pname = pe->pmu ? pe->pmu : data->cpu_name; -if (!pe->name) -return 0 ; -if (data->pmu->is_uncore &&pmu_uncore_alias_match (pname, data->name)) goto new_alias;@@ -879,12 +865,6 @@ static int pmu_add_sys_aliases_iter_fn(const struct pmu_event *pe,struct pmu_sys_event_iter_data *idata = data;struct perf_pmu *pmu = idata->pmu; -if (!pe->name) { -if (pe->metric_group || pe->metric_name) -return 0 ; -return -EINVAL ; - } -if (!pe->compat || !pe->pmu)return 0 ;@@ -1469,8 +1449,6 @@ int perf_pmu__check_alias(struct perf_pmu *pmu, struct list_head *head_terms, info->unit =NULL ; info->scale =0.0 ; info->snapshot = false; - info->metric_expr =NULL ; - info->metric_name =NULL ;list_for_each_entry_safe (term, h, head_terms, list) { alias =pmu_find_alias (pmu, term);@@ -1486,8 +1464,6 @@ int perf_pmu__check_alias(struct perf_pmu *pmu, struct list_head *head_terms,if (alias->per_pkg) info->per_pkg = true; - info->metric_expr = alias->metric_expr; - info->metric_name = alias->metric_name;list_del_init (&term->list);parse_events_term__delete (term);@@ -1703,8 +1679,7 @@ void print_pmu_events(const struct print_callbacks *print_cb, void *print_state)for (j =0 ; j < len; j++) {const char *name, *alias =NULL , *scale_unit =NULL , *desc =NULL , *long_desc =NULL , - *encoding_desc =NULL , *topic =NULL , - *metric_name =NULL , *metric_expr =NULL ; + *encoding_desc =NULL , *topic =NULL ;bool deprecated = false;size_t buf_used;@@ -1742,8 +1717,6 @@ void print_pmu_events(const struct print_callbacks *print_cb, void *print_state) buf_used +=snprintf (buf + buf_used,sizeof (buf) - buf_used,"%s/%s/" , aliases[j].pmu->name, aliases[j].event->str) +1 ; - metric_name = aliases[j].event->metric_name; - metric_expr = aliases[j].event->metric_expr; deprecated = aliases[j].event->deprecated; } print_cb->print_event(print_state,@@ -1756,9 +1729,7 @@ void print_pmu_events(const struct print_callbacks *print_cb, void *print_state)"Kernel PMU event" , desc, long_desc, - encoding_desc, - metric_name, - metric_expr); + encoding_desc); }if (printed &&pager_in_use ())printf ("\n" );tools/perf/util/pmu.h +0
-9@@ -133,8 +133,6 @@ extern struct perf_pmu perf_pmu__fake;struct perf_pmu_info {const char *unit; -const char *metric_expr; -const char *metric_name;double scale;bool per_pkg;bool snapshot;@@ -188,13 +186,6 @@ struct perf_pmu_alias {* default. */ bool deprecated; -/** -* @metric_expr: A metric expression associated with an event. Doing -* this makes little sense due to scale and unit applying to both. -*/ -char *metric_expr; -/** @metric_name: A name for the metric. unit applying to both. */ -char *metric_name;/** @pmu_name: The name copied from struct perf_pmu. */ char *pmu_name; };tools/perf/util/print-events.c +8
-24@@ -101,9 +101,7 @@ void print_tracepoint_events(const struct print_callbacks *print_cb, void *print"Tracepoint event" ,/*desc=*/ NULL ,/*long_desc=*/ NULL , -/*encoding_desc=*/ NULL , -/*metric_name=*/ NULL , - /*metric_expr=*/NULL); + /*encoding_desc=*/NULL); }free (dir_path);free (evt_namelist);@@ -195,9 +193,7 @@ void print_sdt_events(const struct print_callbacks *print_cb, void *print_state)"SDT event" ,/*desc=*/ NULL ,/*long_desc=*/ NULL , -/*encoding_desc=*/ NULL , -/*metric_name=*/ NULL , - /*metric_expr=*/NULL); + /*encoding_desc=*/NULL);free (evt_name); }@@ -255,9 +251,7 @@ int print_hwcache_events(const struct print_callbacks *print_cb, void *print_sta event_type_descriptors[PERF_TYPE_HW_CACHE ],/*desc=*/ NULL ,/*long_desc=*/ NULL , -/*encoding_desc=*/ NULL , -/*metric_name=*/ NULL , - /*metric_expr=*/NULL); + /*encoding_desc=*/NULL); }strlist__delete (evt_name_list);return 0 ;@@ -277,9 +271,7 @@ void print_tool_events(const struct print_callbacks *print_cb, void *print_state"Tool event" ,/*desc=*/ NULL ,/*long_desc=*/ NULL , -/*encoding_desc=*/ NULL , -/*metric_name=*/ NULL , - /*metric_expr=*/NULL); + /*encoding_desc=*/NULL); } }@@ -331,9 +323,7 @@ void print_symbol_events(const struct print_callbacks *print_cb, void *print_sta event_type_descriptors[type],/*desc=*/ NULL ,/*long_desc=*/ NULL , -/*encoding_desc=*/ NULL , -/*metric_name=*/ NULL , - /*metric_expr=*/NULL); + /*encoding_desc=*/NULL); }strlist__delete (evt_name_list); }@@ -364,9 +354,7 @@ void print_events(const struct print_callbacks *print_cb, void *print_state) event_type_descriptors[PERF_TYPE_RAW ],/*desc=*/ NULL ,/*long_desc=*/ NULL , -/*encoding_desc=*/ NULL , -/*metric_name=*/ NULL , - /*metric_expr=*/NULL); + /*encoding_desc=*/NULL); print_cb->print_event(print_state, /*topic=*/NULL,@@ -378,9 +366,7 @@ void print_events(const struct print_callbacks *print_cb, void *print_state) event_type_descriptors[PERF_TYPE_RAW ],"(see 'man perf-list' on how to encode it)" ,/*long_desc=*/ NULL , -/*encoding_desc=*/ NULL , -/*metric_name=*/ NULL , - /*metric_expr=*/NULL); + /*encoding_desc=*/NULL); print_cb->print_event(print_state, /*topic=*/NULL,@@ -392,9 +378,7 @@ void print_events(const struct print_callbacks *print_cb, void *print_state) event_type_descriptors[PERF_TYPE_BREAKPOINT ],/*desc=*/ NULL ,/*long_desc=*/ NULL , -/*encoding_desc=*/ NULL , -/*metric_name=*/ NULL , - /*metric_expr=*/NULL); + /*encoding_desc=*/NULL);print_tracepoint_events (print_cb, print_state);tools/perf/util/print-events.h +1
-2@@ -16,8 +16,7 @@ struct print_callbacks {const char *scale_unit, bool deprecated, const char *event_type_desc, const char *desc, const char *long_desc, - const char *encoding_desc, - const char *metric_name, const char *metric_expr); + const char *encoding_desc); void (*print_metric)(void *print_state, const char *group, const char *name,
Ian Rogers
reviewerPATCH v2 5/9 perf stat: Remove evsel metric_name/expr
In reply to Ian RogersMetrics are their own unit and these variables held broken metrics previously and now just hold the value NULL. Remove code that used these variables.
--- tools/perf/builtin-stat.c | 1 - tools/perf/util/cgroup.c | 1 - tools/perf/util/evsel.c | 2 - tools/perf/util/evsel.h | 2 - tools/perf/util/python.c | 7 --- tools/perf/util/stat-shadow.c | 112 ---------------------------------- tools/perf/util/stat.h | 1 - 7 files changed, 126 deletions(-)
tools/perf/builtin-stat.c +0 -1
@@ -2524,7 +2524,6 @@ int cmd_stat(int argc, const char **argv) &stat_config.metric_events);zfree (&metrics); } -perf_stat__collect_metric_expr (evsel_list);perf_stat__init_shadow_stats ();if (add_default_attributes ())
tools/perf/util/cgroup.c +0 -1
@@ -468,7 +468,6 @@ int evlist__expand_cgroup(struct evlist *evlist, const char *str, nr_cgroups++;if (metric_events) { -perf_stat__collect_metric_expr (tmp_list);if (metricgroup__copy_metric_events (tmp_list, cgrp, metric_events, &orig_metric_events) <0 )
tools/perf/util/evsel.c +0 -2
@@ -285,8 +285,6 @@ void evsel__init(struct evsel *evsel, evsel->sample_size =__evsel__sample_size (attr->sample_type);evsel__calc_id_pos (evsel); evsel->cmdline_group_boundary = false; - evsel->metric_expr =NULL ; - evsel->metric_name =NULL ; evsel->metric_events =NULL ; evsel->per_pkg_mask =NULL ; evsel->collect_stat = false;
tools/perf/util/evsel.h +0 -2
@@ -105,8 +105,6 @@ struct evsel {* metric fields are similar, but needs more care as they can have * references to other metric (evsel). */ -const char * metric_expr; -const char * metric_name;struct evsel **metric_events;struct evsel *metric_leader;
tools/perf/util/python.c +0 -7
@@ -75,13 +75,6 @@ const char *perf_env__arch(struct perf_env *env __maybe_unused)return NULL ; } -/* -* Add this one here not to drag util/stat-shadow.c -*/ -void perf_stat__collect_metric_expr (struct evlist *evsel_list) -{ -} -/* * This one is needed not to drag the PMU bandwagon, jevents generated * pmu_sys_event_tables, etc and evsel__find_pmu() is used so far just for
tools/perf/util/stat-shadow.c +0 -112
124 lines not on this page · raw message
tools/perf/util/stat.h +0 -1
@@ -257,7 +257,6 @@ void perf_stat__print_shadow_stats(struct perf_stat_config *config,struct perf_stat_output_ctx *out, struct rblist *metric_events, struct runtime_stat *st); -void perf_stat__collect_metric_expr (struct evlist *);int evlist__alloc_stats (struct perf_stat_config *config,struct evlist *evlist,bool alloc_raw);
Signature · 1 line
2.39.0.314.g84b9a713c41-goog
Re: [PATCH v2 5/9] perf stat: Remove evsel metric_name/expr
Quoted 5 lines · Ian Rogers
Metrics are their own unit and these variables held broken metrics previously and now just hold the value NULL. Remove code that used these variables.
Signed-off-by Ian Rogers<[email protected]>
Ian Rogers
reviewerPATCH v2 6/9 perf jevents: Combine table prefix and suffix writing
In reply to Ian RogersCombine into a single function to simplify, in a later change, writing metrics separately.
--- tools/perf/pmu-events/jevents.py | 36 +++++++++++++------------------- 1 file changed, 14 insertions(+), 22 deletions(-)
tools/perf/pmu-events/jevents.py +14 -22
@@ -19,10 +19,10 @@ _sys_event_tables = [] # JsonEvent. Architecture standard events are in json files in the top # f'{_args.starting_dir}/{_args.arch}' directory. _arch_std_events = {} -# Track whether an events table is currently being defined and needs closing. -_close_table = False # Events to write out when the table is closed _pending_events = [] +# Name of table to be written out +_pending_events_tblname = None # Global BigCString shared by all structures. _bcs = None # Order specific JsonEvent attributes will be visited.
@@ -376,24 +376,13 @@ def preprocess_arch_std_files(archpath: str) -> None: _arch_std_events[event.name.lower()] = event -def print_events_table_prefix(tblname: str) -> None: - """Called when a new events table is started.""" - global _close_table - if _close_table: - raise IOError('Printing table prefix but last table has no suffix') - _args.output_file.write(f'static const struct compact_pmu_event {tblname}[] = {{\n') - _close_table = True - - def add_events_table_entries(item: os.DirEntry, topic: str) -> None: """Add contents of file to _pending_events table.""" - if not _close_table: - raise IOError('Table entries missing prefix') for e in read_json_events(item.path, topic): _pending_events.append(e) -def print_events_table_suffix() -> None: +def print_pending_events() -> None: """Optionally close events table.""" def event_cmp_key(j: JsonEvent) -> Tuple[bool, str, str, str, str]:
@@ -405,17 +394,19 @@ def print_events_table_suffix() -> None: return (j.desc is not None, fix_none(j.topic), fix_none(j.name), fix_none(j.pmu), fix_none(j.metric_name)) - global _close_table - if not _close_table: + global _pending_events + if not _pending_events: return - global _pending_events + global _pending_events_tblname + _args.output_file.write( + f'static const struct compact_pmu_event {_pending_events_tblname}[] = {{\n') + for event in sorted(_pending_events, key=event_cmp_key): _args.output_file.write(event.to_c_string()) - _pending_events = [] + _pending_events = [] _args.output_file.write('};\n\n') - _close_table = False def get_topic(topic: str) -> str: if topic.endswith('metrics.json'):
@@ -453,12 +444,13 @@ def process_one_file(parents: Sequence[str], item: os.DirEntry) -> None: # model directory, reset topic if item.is_dir() and is_leaf_dir(item.path): - print_events_table_suffix() + print_pending_events() tblname = file_name_to_table_name(parents, item.name) if item.name == 'sys': _sys_event_tables.append(tblname) - print_events_table_prefix(tblname) + global _pending_events_tblname + _pending_events_tblname = tblname return # base dir or too deep
@@ -802,7 +794,7 @@ struct compact_pmu_event { for arch in archs: arch_path = f'{_args.starting_dir}/{arch}' ftw(arch_path, [], process_one_file) - print_events_table_suffix() + print_pending_events() print_mapping_table(archs) print_system_mapping_table()
Signature · 1 line
2.39.0.314.g84b9a713c41-goog
Ian Rogers
reviewerPATCH v2 7/9 perf pmu-events: Introduce pmu_metrics_table
In reply to Ian RogersAdd a metrics table that is just a cast from pmu_events_table. This changes the APIs so that event and metric usage of the underlying table is different. Later changes will separate the tables.
This introduction fixes a NO_JEVENTS=1 regression on:
68: Parse and process metrics : Ok
70: Event expansion for cgroups : Ok
caused by the necessary test metrics not being found.
--- tools/perf/arch/arm64/util/pmu.c | 23 ++++++++++- tools/perf/pmu-events/empty-pmu-events.c | 52 ++++++++++++++++++++---- tools/perf/pmu-events/jevents.py | 24 ++++++++--- tools/perf/pmu-events/pmu-events.h | 10 +++-- tools/perf/tests/expand-cgroup.c | 4 +- tools/perf/tests/parse-metric.c | 4 +- tools/perf/tests/pmu-events.c | 5 ++- tools/perf/util/metricgroup.c | 50 +++++++++++------------ tools/perf/util/metricgroup.h | 2 +- tools/perf/util/pmu.c | 9 +++- tools/perf/util/pmu.h | 1 + 11 files changed, 133 insertions(+), 51 deletions(-)
tools/perf/arch/arm64/util/pmu.c +22 -1
@@ -19,7 +19,28 @@ const struct pmu_events_table *pmu_events_table__find(void)if (pmu->cpus->nr !=cpu__max_cpu ().cpu)return NULL ; -return perf_pmu__find_table (pmu); +return perf_pmu__find_events_table (pmu); + } + +return NULL ; +} + +const struct pmu_metrics_table *pmu_metrics_table__find (void ) +{ +struct perf_pmu *pmu =NULL ; + +while ((pmu =perf_pmu__scan (pmu))) { +if (!is_pmu_core (pmu->name)) +continue ; + +/* +* The cpumap should cover all CPUs. Otherwise, some CPUs may +* not support some events or have different event IDs. +*/ +if (pmu->cpus->nr !=cpu__max_cpu ().cpu) +return NULL ; + +return perf_pmu__find_metrics_table (pmu); }return NULL ;
tools/perf/pmu-events/empty-pmu-events.c +44 -8
85 lines not on this page · raw message
tools/perf/pmu-events/jevents.py +19 -5
@@ -609,17 +609,19 @@ int pmu_events_table_for_each_event(const struct pmu_events_table *table, return 0; } -int pmu_events_table_for_each_metric(const struct pmu_events_table *table, +int pmu_metrics_table_for_each_metric(const struct pmu_metrics_table *mtable, pmu_metric_iter_fn fn, void *data) { + struct pmu_events_table *table = (struct pmu_events_table *)mtable; + for (size_t i = 0; i < table->length; i++) { struct pmu_metric pm; int ret; decompress_metric(table->entries[i].offset, &pm); if (pm.metric_name) { - ret = fn(&pm, table, data); + ret = fn(&pm, mtable, data); if (ret) return ret; }
@@ -627,7 +629,7 @@ int pmu_events_table_for_each_metric(const struct pmu_events_table *table, return 0; } -const struct pmu_events_table *perf_pmu__find_table(struct perf_pmu *pmu) +const struct pmu_events_table *perf_pmu__find_events_table(struct perf_pmu *pmu) { const struct pmu_events_table *table = NULL; char *cpuid = perf_pmu__getcpuid(pmu);
@@ -654,6 +656,11 @@ const struct pmu_events_table *perf_pmu__find_table(struct perf_pmu *pmu) return table; } +const struct pmu_metrics_table *perf_pmu__find_metrics_table(struct perf_pmu *pmu) +{ + return (struct pmu_metrics_table *)perf_pmu__find_events_table(pmu); +} + const struct pmu_events_table *find_core_events_table(const char *arch, const char *cpuid) { for (const struct pmu_events_map *tables = &pmu_events_map[0];
@@ -665,6 +672,11 @@ const struct pmu_events_table *find_core_events_table(const char *arch, const ch return NULL; } +const struct pmu_metrics_table *find_core_metrics_table(const char *arch, const char *cpuid) +{ + return (struct pmu_metrics_table *)find_core_events_table(arch, cpuid); +} + int pmu_for_each_core_event(pmu_event_iter_fn fn, void *data) { for (const struct pmu_events_map *tables = &pmu_events_map[0];
@@ -683,7 +695,8 @@ int pmu_for_each_core_metric(pmu_metric_iter_fn fn, void *data) for (const struct pmu_events_map *tables = &pmu_events_map[0]; tables->arch; tables++) { - int ret = pmu_events_table_for_each_metric(&tables->table, fn, data); + int ret = pmu_metrics_table_for_each_metric( + (struct pmu_metrics_table *)&tables->table, fn, data); if (ret) return ret;
@@ -720,7 +733,8 @@ int pmu_for_each_sys_metric(pmu_metric_iter_fn fn, void *data) for (const struct pmu_sys_events *tables = &pmu_sys_event_tables[0]; tables->name; tables++) { - int ret = pmu_events_table_for_each_metric(&tables->table, fn, data); + int ret = pmu_metrics_table_for_each_metric( + (struct pmu_metrics_table *)&tables->table, fn, data); if (ret) return ret;
tools/perf/pmu-events/pmu-events.h +7 -3
@@ -39,26 +39,30 @@ struct pmu_metric { };struct pmu_events_table ; +struct pmu_metrics_table ;typedef int (*pmu_event_iter_fn )(const struct pmu_event *pe,const struct pmu_events_table *table,void *data);typedef int (*pmu_metric_iter_fn )(const struct pmu_metric *pm, -const struct pmu_events_table *table, +const struct pmu_metrics_table *table,void *data);int pmu_events_table_for_each_event (const struct pmu_events_table *table,pmu_event_iter_fn fn,void *data); -int pmu_events_table_for_each_metric (const struct pmu_events_table *table,pmu_metric_iter_fn fn, +int pmu_metrics_table_for_each_metric (const struct pmu_metrics_table *table,pmu_metric_iter_fn fn,void *data); -const struct pmu_events_table *perf_pmu__find_table (struct perf_pmu *pmu); +const struct pmu_events_table *perf_pmu__find_events_table (struct perf_pmu *pmu); +const struct pmu_metrics_table *perf_pmu__find_metrics_table (struct perf_pmu *pmu);const struct pmu_events_table *find_core_events_table (const char *arch,const char *cpuid); +const struct pmu_metrics_table *find_core_metrics_table (const char *arch,const char *cpuid);int pmu_for_each_core_event (pmu_event_iter_fn fn,void *data);int pmu_for_each_core_metric (pmu_metric_iter_fn fn,void *data);const struct pmu_events_table *find_sys_events_table (const char *name); +const struct pmu_metrics_table *find_sys_metrics_table (const char *name);int pmu_for_each_sys_event (pmu_event_iter_fn fn,void *data);int pmu_for_each_sys_metric (pmu_metric_iter_fn fn,void *data);
tools/perf/tests/expand-cgroup.c +2 -2
@@ -180,13 +180,13 @@ static int expand_metric_events(void)struct evlist *evlist;struct rblist metric_events;const char metric_str[] ="CPI" ; -const struct pmu_events_table *pme_test; +const struct pmu_metrics_table *pme_test; evlist =evlist__new ();TEST_ASSERT_VAL ("failed to get evlist" , evlist);rblist__init (&metric_events); - pme_test =find_core_events_table ("testarch" ,"testcpu" ); + pme_test =find_core_metrics_table ("testarch" ,"testcpu" ); ret =metricgroup__parse_groups_test (evlist, pme_test, metric_str, false, false, &metric_events);if (ret <0 ) {
tools/perf/tests/parse-metric.c +2 -2
@@ -72,7 +72,7 @@ static int __compute_metric(const char *name, struct value *vals,struct rblist metric_events = { .nr_entries =0 , }; -const struct pmu_events_table *pme_test; +const struct pmu_metrics_table *pme_test;struct perf_cpu_map *cpus;struct runtime_stat st;struct evlist *evlist;
@@ -96,7 +96,7 @@ static int __compute_metric(const char *name, struct value *vals,runtime_stat__init (&st);/* Parse the metric into metric_events list. */ - pme_test =find_core_events_table ("testarch" ,"testcpu" ); + pme_test =find_core_metrics_table ("testarch" ,"testcpu" ); err =metricgroup__parse_groups_test (evlist, pme_test, name, false, false, &metric_events);
tools/perf/tests/pmu-events.c +3 -2
@@ -813,7 +813,8 @@ struct metric {struct metric_ref metric_ref; }; -static int test__parsing_callback (const struct pmu_metric *pm,const struct pmu_events_table *table, +static int test__parsing_callback (const struct pmu_metric *pm, +const struct pmu_metrics_table *table,void *data) {int *failures = data;
@@ -995,7 +996,7 @@ static int metric_parse_fake(const char *metric_name, const char *str) }static int test__parsing_fake_callback (const struct pmu_metric *pm, -const struct pmu_events_table *table__maybe_unused , +const struct pmu_metrics_table *table__maybe_unused ,void *data__maybe_unused ) {return metric_parse_fake (pm->metric_name, pm->metric_expr);
tools/perf/util/metricgroup.c +25 -25
190 lines not on this page · raw message
tools/perf/util/metricgroup.h +1 -1
@@ -73,7 +73,7 @@ int metricgroup__parse_groups(struct evlist *perf_evlist,bool system_wide, struct rblist *metric_events); int metricgroup__parse_groups_test(struct evlist *evlist, - const struct pmu_events_table *table, + const struct pmu_metrics_table *table, const char *str, bool metric_no_group, bool metric_no_merge,
tools/perf/util/pmu.c +7 -2
@@ -718,7 +718,12 @@ char *perf_pmu__getcpuid(struct perf_pmu *pmu)__weak const struct pmu_events_table *pmu_events_table__find (void ) { -return perf_pmu__find_table (NULL ); +return perf_pmu__find_events_table (NULL ); +} + +__weak const struct pmu_metrics_table *pmu_metrics_table__find (void ) +{ +return perf_pmu__find_metrics_table (NULL ); }/*
@@ -846,7 +851,7 @@ static void pmu_add_cpu_aliases(struct list_head *head, struct perf_pmu *pmu) {const struct pmu_events_table *table; - table =perf_pmu__find_table (pmu); + table =perf_pmu__find_events_table (pmu);if (!table)return ;
tools/perf/util/pmu.h +1 -0
@@ -230,6 +230,7 @@ void pmu_add_cpu_aliases_table(struct list_head *head, struct perf_pmu *pmu,char *perf_pmu__getcpuid (struct perf_pmu *pmu);const struct pmu_events_table *pmu_events_table__find (void ); +const struct pmu_metrics_table *pmu_metrics_table__find (void );bool pmu_uncore_alias_match (const char *pmu_name,const char *name);void perf_pmu_free_alias (struct perf_pmu_alias *alias);
Signature · 1 line
2.39.0.314.g84b9a713c41-goog
John Garry
reviewerRe: [PATCH v2 7/9] perf pmu-events: Introduce pmu_metrics_table
Quoted 9 lines · Ian Rogers
Add a metrics table that is just a cast from
pmu_events_table. This changes the APIs so that event and metric usage of the underlying table is different. Later changes will separate the tables.This introduction fixes a NO_JEVENTS=1 regression on:
68: Parse and process metrics : Ok
70: Event expansion for cgroups : Ok
caused by the necessary test metrics not being found.
I have just checked some of this code so far...
Quoted 36 lines
Signed-off-by Ian Rogers <[email protected]>--- tools/perf/arch/arm64/util/pmu.c | 23 ++++++++++- tools/perf/pmu-events/empty-pmu-events.c | 52 ++++++++++++++++++++---- tools/perf/pmu-events/jevents.py | 24 ++++++++--- tools/perf/pmu-events/pmu-events.h | 10 +++-- tools/perf/tests/expand-cgroup.c | 4 +- tools/perf/tests/parse-metric.c | 4 +- tools/perf/tests/pmu-events.c | 5 ++- tools/perf/util/metricgroup.c | 50 +++++++++++------------ tools/perf/util/metricgroup.h | 2 +- tools/perf/util/pmu.c | 9 +++- tools/perf/util/pmu.h | 1 + 11 files changed, 133 insertions(+), 51 deletions(-)tools/perf/arch/arm64/util/pmu.c +12
-1@@ -19,7 +19,28 @@ const struct pmu_events_table *pmu_events_table__find(void)if (pmu->cpus->nr !=cpu__max_cpu ().cpu)return NULL ; -return perf_pmu__find_table (pmu); +return perf_pmu__find_events_table (pmu); + } + +return NULL ; +} + +const struct pmu_metrics_table *pmu_metrics_table__find (void ) +{ +struct perf_pmu *pmu =NULL ; + +while ((pmu =perf_pmu__scan (pmu))) { +if (!is_pmu_core (pmu->name))
tools/perf/arch/arm64/util/pmu.c const struct pmu_events_table *pmu_events_table__find(void) · line 43
Quoted 10 lines
tools/perf/arch/arm64/util/pmu.c +10
-0+continue ; + +/* +* The cpumap should cover all CPUs. Otherwise, some CPUs may +* not support some events or have different event IDs. +*/ +if (pmu->cpus->nr !=cpu__max_cpu ().cpu) +return NULL ; + +return perf_pmu__find_metrics_table (pmu);
I think that this code will be conflicting with the recent arm64 metric support. And now it seems even more scope for factoring out code.
Quoted 9 lines
tools/perf/arch/arm64/util/pmu.c +0
-0}return NULL ;tools/perf/pmu-events/empty-pmu-events.c +0
-0@@ -278,14 +278,12 @@ int pmu_events_table_for_each_event(const struct pmu_events_table *table, pmu_evreturn 0 ;
tools/perf/pmu-events/empty-pmu-events.c int pmu_events_table_for_each_event(const struct pmu_events_table *table, pmu_ev · line 284
Quoted 10 lines
tools/perf/pmu-events/empty-pmu-events.c +2
-4} -int pmu_events_table_for_each_metric (const struct pmu_events_table *etable,pmu_metric_iter_fn fn, -void *data) +int pmu_metrics_table_for_each_metric (const struct pmu_metrics_table *table,pmu_metric_iter_fn fn, +void *data) { -struct pmu_metrics_table *table = (struct pmu_metrics_table *)etable; - for (const struct pmu_metric *pm = &table->entries[0]
nit on coding style: do we normally declare local variables like this? It condenses the code but makes a bit less readable, IMHO
Quoted 25 lines
tools/perf/pmu-events/empty-pmu-events.c +1
-1; pm->metric_group || pm->metric_name; pm++) { -int ret =fn (pm, etable, data); +int ret =fn (pm, table, data);if (ret)return ret;
tools/perf/pmu-events/empty-pmu-events.c +6
-1@@ -293,7 +291,7 @@ int pmu_events_table_for_each_metric(const struct pmu_events_table *etable, pmu_return 0 ; } -const struct pmu_events_table *perf_pmu__find_table (struct perf_pmu *pmu) +const struct pmu_events_table *perf_pmu__find_events_table (struct perf_pmu *pmu) {const struct pmu_events_table *table =NULL ;char *cpuid =perf_pmu__getcpuid (pmu);@@ -321,6 +319,34 @@ const struct pmu_events_table *perf_pmu__find_table(struct perf_pmu *pmu)return table; } +const struct pmu_metrics_table *perf_pmu__find_metrics_table (struct perf_pmu *pmu) +{ +const struct pmu_metrics_table *table =NULL ; +char *cpuid =perf_pmu__getcpuid (pmu); +int i;
tools/perf/pmu-events/empty-pmu-events.c const struct pmu_events_table *perf_pmu__find_table(struct perf_pmu *pmu) · line 336
Quoted 10 lines
tools/perf/pmu-events/empty-pmu-events.c +10
-0+ +/* on some platforms which uses cpus map, cpuid can be NULL for +* PMUs other than CORE PMUs. +*/ +if (!cpuid) +return NULL ; + + i =0 ; +for (;;) { +const struct pmu_events_map *map = &pmu_events_map[i++];
To me, this is all strange code. Again this is a comment on the current code: Consider pmu_for_each_sys_event() as an example, we have a while loop for each member of pmu_sys_event_tables[]. But pmu_sys_event_tables is hardcoded for a single member, so why loop? It seems the same for all these "for each" helper in the "empty" events c file.
Quoted 17 lines
tools/perf/pmu-events/empty-pmu-events.c +13
-0+ +if (!map->cpuid) +break ; + +if (!strcmp_cpuid_str (map->cpuid, cpuid)) { + table = &map->metric_table; +break ; + } + } +free (cpuid); +return table; +} +const struct pmu_events_table *find_core_events_table (const char *arch,const char *cpuid) { for (const struct pmu_events_map *tables = &pmu_events_map[0];
tools/perf/pmu-events/empty-pmu-events.c const struct pmu_events_table *find_core_events_table(const char *arch, const ch · line 365
Quoted 8 lines
tools/perf/pmu-events/empty-pmu-events.c +5
-0@@ -332,6 +358,17 @@ const struct pmu_events_table *find_core_events_table(const char *arch, const chreturn NULL ; } +const struct pmu_metrics_table *find_core_metrics_table (const char *arch,const char *cpuid) +{ +for (const struct pmu_events_map *tables = &pmu_events_map[0 ]; + tables->arch; + tables++) {
combine with previous line?
Quoted 24 lines
tools/perf/pmu-events/empty-pmu-events.c +7
-2+if (!strcmp (tables->arch, arch) && !strcmp_cpuid_str (tables->cpuid, cpuid)) +return &tables->metric_table; + } +return NULL ; +} +int pmu_for_each_core_event (pmu_event_iter_fn fn,void *data) { for (const struct pmu_events_map *tables = &pmu_events_map[0];@@ -350,8 +387,7 @@ int pmu_for_each_core_metric(pmu_metric_iter_fn fn, void *data)for (const struct pmu_events_map *tables = &pmu_events_map[0 ]; tables->arch; tables++) { -int ret =pmu_events_table_for_each_metric ( - (const struct pmu_events_table *)&tables->metric_table, fn, data); +int ret =pmu_metrics_table_for_each_metric (&tables->metric_table, fn, data);if (ret)return ret;tools/perf/pmu-events/jevents.py +0
-0
tools/perf/pmu-events/jevents.py int pmu_events_table_for_each_event(const struct pmu_events_table *table, · line 616
Quoted 9 lines
tools/perf/pmu-events/jevents.py +2
-1@@ -609,17 +609,19 @@ int pmu_events_table_for_each_event(const struct pmu_events_table *table, return 0; } -int pmu_events_table_for_each_metric(const struct pmu_events_table *table, +int pmu_metrics_table_for_each_metric(const struct pmu_metrics_table *mtable, pmu_metric_iter_fn fn, void *data) { + struct pmu_events_table *table = (struct pmu_events_table *)mtable;
As I may have hinted before, can we avoid casts like this, even if transient?
Quoted 12 lines
tools/perf/pmu-events/jevents.py +2
-1+ for (size_t i = 0; i < table->length; i++) { struct pmu_metric pm; int ret; decompress_metric(table->entries[i].offset, &pm); if (pm.metric_name) { - ret = fn(&pm, table, data); + ret = fn(&pm, mtable, data); if (ret) return ret; }
Ian Rogers
reviewerRe: [PATCH v2 7/9] perf pmu-events: Introduce pmu_metrics_table
Quoted 55 lines · John Garry
Quoted 9 lines · Ian Rogers
Add a metrics table that is just a cast from
pmu_events_table. This changes the APIs so that event and metric usage of the underlying table is different. Later changes will separate the tables.This introduction fixes a NO_JEVENTS=1 regression on:
68: Parse and process metrics : Ok
70: Event expansion for cgroups : Ok
caused by the necessary test metrics not being found.I have just checked some of this code so far...
Quoted 41 lines
Signed-off-by Ian Rogers <[email protected]>--- tools/perf/arch/arm64/util/pmu.c | 23 ++++++++++- tools/perf/pmu-events/empty-pmu-events.c | 52 ++++++++++++++++++++---- tools/perf/pmu-events/jevents.py | 24 ++++++++--- tools/perf/pmu-events/pmu-events.h | 10 +++-- tools/perf/tests/expand-cgroup.c | 4 +- tools/perf/tests/parse-metric.c | 4 +- tools/perf/tests/pmu-events.c | 5 ++- tools/perf/util/metricgroup.c | 50 +++++++++++------------ tools/perf/util/metricgroup.h | 2 +- tools/perf/util/pmu.c | 9 +++- tools/perf/util/pmu.h | 1 + 11 files changed, 133 insertions(+), 51 deletions(-)tools/perf/arch/arm64/util/pmu.c +17
-1@@ -19,7 +19,28 @@ const struct pmu_events_table *pmu_events_table__find(void)if (pmu->cpus->nr !=cpu__max_cpu ().cpu)return NULL ; -return perf_pmu__find_table (pmu); +return perf_pmu__find_events_table (pmu); + } + +return NULL ; +} + +const struct pmu_metrics_table *pmu_metrics_table__find (void ) +{ +struct perf_pmu *pmu =NULL ; + +while ((pmu =perf_pmu__scan (pmu))) { +if (!is_pmu_core (pmu->name)) +continue ; + +/* +* The cpumap should cover all CPUs. Otherwise, some CPUs may +* not support some events or have different event IDs.
Quoted 8 lines
tools/perf/arch/arm64/util/pmu.c const struct pmu_events_table *pmu_events_table__find(void) · line 43
Quoted 5 lines
tools/perf/arch/arm64/util/pmu.c +5
-0+*/ +if (pmu->cpus->nr !=cpu__max_cpu ().cpu) +return NULL ; + +return perf_pmu__find_metrics_table (pmu);I think that this code will be conflicting with the recent arm64 metric support. And now it seems even more scope for factoring out code.
v3 will rebase and fix.
Quoted 11 lines
Quoted 11 lines
tools/perf/arch/arm64/util/pmu.c +0
-0}return NULL ;tools/perf/pmu-events/empty-pmu-events.c +0
-0@@ -278,14 +278,12 @@ int pmu_events_table_for_each_event(const struct pmu_events_table *table, pmu_evreturn 0 ; }
Quoted 11 lines
tools/perf/pmu-events/empty-pmu-events.c int pmu_events_table_for_each_event(const struct pmu_events_table *table, pmu_ev
Quoted 8 lines
tools/perf/pmu-events/empty-pmu-events.c +2
-4-int pmu_events_table_for_each_metric (const struct pmu_events_table *etable,pmu_metric_iter_fn fn, -void *data) +int pmu_metrics_table_for_each_metric (const struct pmu_metrics_table *table,pmu_metric_iter_fn fn, +void *data) { -struct pmu_metrics_table *table = (struct pmu_metrics_table *)etable; - for (const struct pmu_metric *pm = &table->entries[0]nit on coding style: do we normally declare local variables like this? It condenses the code but makes a bit less readable, IMHO
The main reason to do it is to reduce the scope of pm to just be the loop body. There's some discussion relating to this to do with the move to C11:
https://lwn.net/Articles/885941/
Quoted 33 lines
Quoted 33 lines
tools/perf/pmu-events/empty-pmu-events.c +1
-1; pm->metric_group || pm->metric_name; pm++) { -int ret =fn (pm, etable, data); +int ret =fn (pm, table, data);if (ret)return ret;tools/perf/pmu-events/empty-pmu-events.c +14
-1@@ -293,7 +291,7 @@ int pmu_events_table_for_each_metric(const struct pmu_events_table *etable, pmu_return 0 ; } -const struct pmu_events_table *perf_pmu__find_table (struct perf_pmu *pmu) +const struct pmu_events_table *perf_pmu__find_events_table (struct perf_pmu *pmu) {const struct pmu_events_table *table =NULL ;char *cpuid =perf_pmu__getcpuid (pmu);@@ -321,6 +319,34 @@ const struct pmu_events_table *perf_pmu__find_table(struct perf_pmu *pmu)return table; } +const struct pmu_metrics_table *perf_pmu__find_metrics_table (struct perf_pmu *pmu) +{ +const struct pmu_metrics_table *table =NULL ; +char *cpuid =perf_pmu__getcpuid (pmu); +int i; + +/* on some platforms which uses cpus map, cpuid can be NULL for +* PMUs other than CORE PMUs. +*/ +if (!cpuid) +return NULL ; + + i =0 ;
Quoted 8 lines
tools/perf/pmu-events/empty-pmu-events.c const struct pmu_events_table *perf_pmu__find_table(struct perf_pmu *pmu) · line 336
Quoted 2 lines
tools/perf/pmu-events/empty-pmu-events.c +2
-0+for (;;) { +const struct pmu_events_map *map = &pmu_events_map[i++];To me, this is all strange code. Again this is a comment on the current
code: Considerpmu_for_each_sys_event()as an example, we have a while loop for each member of pmu_sys_event_tables[]. Butpmu_sys_event_tablesis hardcoded for a single member, so why loop? It seems the same for all these "for each" helper in the "empty" events c file.
Agreed. I think we should generate the empty case and then event if there is just 1 iteration, we know that the code is that way because of the auto-generation.
Quoted 20 lines
Quoted 20 lines
tools/perf/pmu-events/empty-pmu-events.c +13
-0+ +if (!map->cpuid) +break ; + +if (!strcmp_cpuid_str (map->cpuid, cpuid)) { + table = &map->metric_table; +break ; + } + } +free (cpuid); +return table; +} +const struct pmu_events_table *find_core_events_table (const char *arch,const char *cpuid) { for (const struct pmu_events_map *tables = &pmu_events_map[0];@@ -332,6 +358,17 @@ const struct pmu_events_table *find_core_events_table(const char *arch, const chreturn NULL ; }
Quoted 7 lines
tools/perf/pmu-events/empty-pmu-events.c const struct pmu_events_table *find_core_events_table(const char *arch, const ch · line 365
Quoted 5 lines
tools/perf/pmu-events/empty-pmu-events.c +5
-0+const struct pmu_metrics_table *find_core_metrics_table (const char *arch,const char *cpuid) +{ +for (const struct pmu_events_map *tables = &pmu_events_map[0 ]; + tables->arch; + tables++) {combine with previous line?
Done.
Quoted 27 lines
Quoted 27 lines
tools/perf/pmu-events/empty-pmu-events.c +7
-2+if (!strcmp (tables->arch, arch) && !strcmp_cpuid_str (tables->cpuid, cpuid)) +return &tables->metric_table; + } +return NULL ; +} +int pmu_for_each_core_event (pmu_event_iter_fn fn,void *data) { for (const struct pmu_events_map *tables = &pmu_events_map[0];@@ -350,8 +387,7 @@ int pmu_for_each_core_metric(pmu_metric_iter_fn fn, void *data)for (const struct pmu_events_map *tables = &pmu_events_map[0 ]; tables->arch; tables++) { -int ret =pmu_events_table_for_each_metric ( - (const struct pmu_events_table *)&tables->metric_table, fn, data); +int ret =pmu_metrics_table_for_each_metric (&tables->metric_table, fn, data);if (ret)return ret;tools/perf/pmu-events/jevents.py +0
-0@@ -609,17 +609,19 @@ int pmu_events_table_for_each_event(const struct pmu_events_table *table, return 0; }
Quoted 9 lines
tools/perf/pmu-events/jevents.py int pmu_events_table_for_each_event(const struct pmu_events_table *table, · line 616
Quoted 6 lines
tools/perf/pmu-events/jevents.py +2
-1-int pmu_events_table_for_each_metric(const struct pmu_events_table *table, +int pmu_metrics_table_for_each_metric(const struct pmu_metrics_table *mtable, pmu_metric_iter_fn fn, void *data) { + struct pmu_events_table *table = (struct pmu_events_table *)mtable;As I may have hinted before, can we avoid casts like this, even if transient?
It was a trade-off with patch size. Will try to improve in v3.
Thanks,
Ian
Quoted 15 lines
Quoted 12 lines
tools/perf/pmu-events/jevents.py +2
-1+ for (size_t i = 0; i < table->length; i++) { struct pmu_metric pm; int ret; decompress_metric(table->entries[i].offset, &pm); if (pm.metric_name) { - ret = fn(&pm, table, data); + ret = fn(&pm, mtable, data); if (ret) return ret; }
Ian Rogers
reviewerPATCH v2 8/9 perf jevents: Generate metrics and events as separate tables
In reply to Ian RogersTurn a perf json event into an event, metric or both. This reduces the number of events needed to scan to find an event or metric. As events no longer need the relatively seldom used metric fields, 4 bytes is saved per event. This reduces the big C string's size by 335kb (14.8%) on x86.
--- tools/perf/pmu-events/jevents.py | 244 +++++++++++++++++++++++-------- tools/perf/tests/pmu-events.c | 3 +- 2 files changed, 189 insertions(+), 58 deletions(-)
tools/perf/pmu-events/jevents.py +187 -57
@@ -13,28 +13,40 @@ import collections # Global command line arguments. _args = None +# List of regular event tables. +_event_tables = [] # List of event tables generated from "/sys" directories. _sys_event_tables = [] +# List of regular metric tables. +_metric_tables = [] +# List of metric tables generated from "/sys" directories. +_sys_metric_tables = [] +# Mapping between sys event table names and sys metric table names. +_sys_event_table_to_metric_table_mapping = {} # Map from an event name to an architecture standard # JsonEvent. Architecture standard events are in json files in the top # f'{_args.starting_dir}/{_args.arch}' directory. _arch_std_events = {} # Events to write out when the table is closed _pending_events = [] -# Name of table to be written out +# Name of events table to be written out _pending_events_tblname = None +# Metrics to write out when the table is closed +_pending_metrics = [] +# Name of metrics table to be written out +_pending_metrics_tblname = None # Global BigCString shared by all structures. _bcs = None # Order specific JsonEvent attributes will be visited. _json_event_attributes = [ # cmp_sevent related attributes. - 'name', 'pmu', 'topic', 'desc', 'metric_name', 'metric_group', + 'name', 'pmu', 'topic', 'desc', # Seems useful, put it early. 'event', # Short things in alphabetical order. 'aggr_mode', 'compat', 'deprecated', 'perpkg', 'unit', # Longer things (the last won't be iterated over during decompress). - 'metric_constraint', 'metric_expr', 'long_desc' + 'long_desc' ] # Attributes that are in pmu_metric rather than pmu_event.
@@ -52,14 +64,16 @@ def removesuffix(s: str, suffix: str) -> str: return s[0:-len(suffix)] if s.endswith(suffix) else s -def file_name_to_table_name(parents: Sequence[str], dirname: str) -> str: +def file_name_to_table_name(prefix: str, parents: Sequence[str], + dirname: str) -> str: """Generate a C table name from directory names.""" - tblname = 'pme' + tblname = prefix for p in parents: tblname += '_' + p tblname += '_' + dirname return tblname.replace('-', '_') + def c_len(s: str) -> int: """Return the length of s a C string
@@ -277,7 +291,7 @@ class JsonEvent: self.metric_constraint = jd.get('MetricConstraint') self.metric_expr = None if 'MetricExpr' in jd: - self.metric_expr = metric.ParsePerfJson(jd['MetricExpr']).Simplify() + self.metric_expr = metric.ParsePerfJson(jd['MetricExpr']).Simplify() arch_std = jd.get('ArchStdEvent') if precise and self.desc and '(Precise Event)' not in self.desc:
@@ -326,23 +340,24 @@ class JsonEvent: s += f'\t{attr} = {value},\n' return s + '}' - def build_c_string(self) -> str: + def build_c_string(self, metric: bool) -> str: s = '' - for attr in _json_event_attributes: + for attr in _json_metric_attributes if metric else _json_event_attributes: x = getattr(self, attr) - if x and attr == 'metric_expr': + if metric and x and attr == 'metric_expr': # Convert parsed metric expressions into a string. Slashes # must be doubled in the file. x = x.ToPerfJson().replace('\\', '\\\\') s += f'{x}\\000' if x else '\\000' return s - def to_c_string(self) -> str: + def to_c_string(self, metric: bool) -> str: """Representation of the event as a C struct initializer.""" - s = self.build_c_string() + s = self.build_c_string(metric) return f'{{ { _bcs.offsets[s] } }}, /* {s} */\n' + @lru_cache(maxsize=None) def read_json_events(path: str, topic: str) -> Sequence[JsonEvent]: """Read json events from the specified file."""
@@ -379,7 +394,10 @@ def preprocess_arch_std_files(archpath: str) -> None: def add_events_table_entries(item: os.DirEntry, topic: str) -> None: """Add contents of file to _pending_events table.""" for e in read_json_events(item.path, topic): - _pending_events.append(e) + if e.name: + _pending_events.append(e) + if e.metric_name: + _pending_metrics.append(e) def print_pending_events() -> None:
@@ -399,15 +417,54 @@ def print_pending_events() -> None: return global _pending_events_tblname + if _pending_events_tblname.endswith('_sys'): + global _sys_event_tables + _sys_event_tables.append(_pending_events_tblname) + else: + global event_tables + _event_tables.append(_pending_events_tblname) + _args.output_file.write( f'static const struct compact_pmu_event {_pending_events_tblname}[] = {{\n') for event in sorted(_pending_events, key=event_cmp_key): - _args.output_file.write(event.to_c_string()) + _args.output_file.write(event.to_c_string(metric=False)) _pending_events = [] _args.output_file.write('};\n\n') +def print_pending_metrics() -> None: + """Optionally close metrics table.""" + + def metric_cmp_key(j: JsonEvent) -> Tuple[bool, str, str]: + def fix_none(s: Optional[str]) -> str: + if s is None: + return '' + return s + + return (j.desc is not None, fix_none(j.pmu), fix_none(j.metric_name)) + + global _pending_metrics + if not _pending_metrics: + return + + global _pending_metrics_tblname + if _pending_metrics_tblname.endswith('_sys'): + global _sys_metric_tables + _sys_metric_tables.append(_pending_metrics_tblname) + else: + global metric_tables + _metric_tables.append(_pending_metrics_tblname) + + _args.output_file.write( + f'static const struct compact_pmu_event {_pending_metrics_tblname}[] = {{\n') + + for metric in sorted(_pending_metrics, key=metric_cmp_key): + _args.output_file.write(metric.to_c_string(metric=True)) + _pending_metrics = [] + + _args.output_file.write('};\n\n') + def get_topic(topic: str) -> str: if topic.endswith('metrics.json'): return 'metrics'
@@ -430,12 +487,13 @@ def preprocess_one_file(parents: Sequence[str], item: os.DirEntry) -> None: topic = get_topic(item.name) for event in read_json_events(item.path, topic): - _bcs.add(event.build_c_string()) + if event.name: + _bcs.add(event.build_c_string(metric=False)) + if event.metric_name: + _bcs.add(event.build_c_string(metric=True)) def process_one_file(parents: Sequence[str], item: os.DirEntry) -> None: """Process a JSON file during the main walk.""" - global _sys_event_tables - def is_leaf_dir(path: str) -> bool: for item in os.scandir(path): if item.is_dir():
@@ -445,12 +503,15 @@ def process_one_file(parents: Sequence[str], item: os.DirEntry) -> None: # model directory, reset topic if item.is_dir() and is_leaf_dir(item.path): print_pending_events() + print_pending_metrics() - tblname = file_name_to_table_name(parents, item.name) - if item.name == 'sys': - _sys_event_tables.append(tblname) global _pending_events_tblname - _pending_events_tblname = tblname + _pending_events_tblname = file_name_to_table_name('pmu_events_', parents, item.name) + global _pending_metrics_tblname + _pending_metrics_tblname = file_name_to_table_name('pmu_metrics_', parents, item.name) + + if item.name == 'sys': + _sys_event_table_to_metric_table_mapping[_pending_events_tblname] = _pending_metrics_tblname return # base dir or too deep
@@ -475,6 +536,12 @@ struct pmu_events_table { size_t length; }; +/* Struct used to make the PMU metric table implementation opaque to callers. */ +struct pmu_metrics_table { + const struct compact_pmu_event *entries; + size_t length; +}; + /* * Map a CPU to its table of PMU events. The CPU is identified by the * cpuid field, which is an arch-specific identifier for the CPU.
@@ -486,7 +553,8 @@ struct pmu_events_table { struct pmu_events_map { const char *arch; const char *cpuid; - struct pmu_events_table table; + struct pmu_events_table event_table; + struct pmu_metrics_table metric_table; }; /*
@@ -500,9 +568,13 @@ const struct pmu_events_map pmu_events_map[] = { _args.output_file.write("""{ \t.arch = "testarch", \t.cpuid = "testcpu", -\t.table = { -\t.entries = pme_test_soc_cpu, -\t.length = ARRAY_SIZE(pme_test_soc_cpu), +\t.event_table = { +\t\t.entries = pmu_events__test_soc_cpu, +\t\t.length = ARRAY_SIZE(pmu_events__test_soc_cpu), +\t}, +\t.metric_table = { +\t\t.entries = pmu_metrics__test_soc_cpu, +\t\t.length = ARRAY_SIZE(pmu_metrics__test_soc_cpu), \t} }, """)
@@ -513,14 +585,29 @@ const struct pmu_events_map pmu_events_map[] = { for row in table: # Skip the first row or any row beginning with #. if not first and len(row) > 0 and not row[0].startswith('#'): - tblname = file_name_to_table_name([], row[2].replace('/', '_')) + event_tblname = file_name_to_table_name('pmu_events_', [], row[2].replace('/', '_')) + if event_tblname in _event_tables: + event_size = f'ARRAY_SIZE({event_tblname})' + else: + event_tblname = 'NULL' + event_size = '0' + metric_tblname = file_name_to_table_name('pmu_metrics_', [], row[2].replace('/', '_')) + if metric_tblname in _metric_tables: + metric_size = f'ARRAY_SIZE({metric_tblname})' + else: + metric_tblname = 'NULL' + metric_size = '0' cpuid = row[0].replace('\\', '\\\\') _args.output_file.write(f"""{{ \t.arch = "{arch}", \t.cpuid = "{cpuid}", -\t.table = {{ -\t\t.entries = {tblname}, -\t\t.length = ARRAY_SIZE({tblname}) +\t.event_table = {{ +\t\t.entries = {event_tblname}, +\t\t.length = {event_size} +\t}}, +\t.metric_table = {{ +\t\t.entries = {metric_tblname}, +\t\t.length = {metric_size} \t}} }}, """)
@@ -529,7 +616,8 @@ const struct pmu_events_map pmu_events_map[] = { _args.output_file.write("""{ \t.arch = 0, \t.cpuid = 0, -\t.table = { 0, 0 }, +\t.event_table = { 0, 0 }, +\t.metric_table = { 0, 0 }, } }; """)
@@ -540,14 +628,36 @@ def print_system_mapping_table() -> None: _args.output_file.write(""" struct pmu_sys_events { \tconst char *name; -\tstruct pmu_events_table table; +\tstruct pmu_events_table event_table; +\tstruct pmu_metrics_table metric_table; }; static const struct pmu_sys_events pmu_sys_event_tables[] = { """) + printed_metric_tables = [] for tblname in _sys_event_tables: _args.output_file.write(f"""\t{{ -\t\t.table = {{ +\t\t.event_table = {{ +\t\t\t.entries = {tblname}, +\t\t\t.length = ARRAY_SIZE({tblname}) +\t\t}},""") + metric_tblname = _sys_event_table_to_metric_table_mapping[tblname] + if metric_tblname in _sys_metric_tables: + _args.output_file.write(f""" +\t\t.metric_table = {{ +\t\t\t.entries = {metric_tblname}, +\t\t\t.length = ARRAY_SIZE({metric_tblname}) +\t\t}},""") + printed_metric_tables.append(metric_tblname) + _args.output_file.write(f""" +\t\t.name = \"{tblname}\", +\t}}, +""") + for tblname in _sys_metric_tables: + if tblname in printed_metric_tables: + continue + _args.output_file.write(f"""\t{{ +\t\t.metric_table = {{ \t\t\t.entries = {tblname}, \t\t\t.length = ARRAY_SIZE({tblname}) \t\t}},
@@ -555,7 +665,8 @@ static const struct pmu_sys_events pmu_sys_event_tables[] = { \t}}, """) _args.output_file.write("""\t{ -\t\t.table = { 0, 0 } +\t\t.event_table = { 0, 0 }, +\t\t.metric_table = { 0, 0 }, \t}, };
@@ -564,10 +675,7 @@ static void decompress_event(int offset, struct pmu_event *pe) \tconst char *p = &big_c_string[offset]; """) for attr in _json_event_attributes: - if attr in _json_metric_attributes and 'metric_' in attr: - _args.output_file.write(f'\n\t/* Skip {attr} */\n') - else: - _args.output_file.write(f""" + _args.output_file.write(f""" \tpe->{attr} = (*p == '\\0' ? NULL : p); """) if attr == _json_event_attributes[-1]:
@@ -579,14 +687,11 @@ static void decompress_metric(int offset, struct pmu_metric *pm) { \tconst char *p = &big_c_string[offset]; """) - for attr in _json_event_attributes: - if attr in _json_metric_attributes: - _args.output_file.write(f""" + for attr in _json_metric_attributes: + _args.output_file.write(f""" \tpm->{attr} = (*p == '\\0' ? NULL : p); """) - else: - _args.output_file.write(f'\n\t/* Skip {attr} */\n') - if attr == _json_event_attributes[-1]: + if attr == _json_metric_attributes[-1]: continue _args.output_file.write('\twhile (*p++);') _args.output_file.write("""}
@@ -609,19 +714,17 @@ int pmu_events_table_for_each_event(const struct pmu_events_table *table, return 0; } -int pmu_metrics_table_for_each_metric(const struct pmu_metrics_table *mtable, +int pmu_metrics_table_for_each_metric(const struct pmu_metrics_table *table, pmu_metric_iter_fn fn, void *data) { - struct pmu_events_table *table = (struct pmu_events_table *)mtable; - for (size_t i = 0; i < table->length; i++) { struct pmu_metric pm; int ret; decompress_metric(table->entries[i].offset, &pm); if (pm.metric_name) { - ret = fn(&pm, mtable, data); + ret = fn(&pm, table, data); if (ret) return ret; }
@@ -648,7 +751,7 @@ const struct pmu_events_table *perf_pmu__find_events_table(struct perf_pmu *pmu) break; if (!strcmp_cpuid_str(map->cpuid, cpuid)) { - table = &map->table; + table = &map->event_table; break; } }
@@ -658,7 +761,29 @@ const struct pmu_events_table *perf_pmu__find_events_table(struct perf_pmu *pmu) const struct pmu_metrics_table *perf_pmu__find_metrics_table(struct perf_pmu *pmu) { - return (struct pmu_metrics_table *)perf_pmu__find_events_table(pmu); + const struct pmu_metrics_table *table = NULL; + char *cpuid = perf_pmu__getcpuid(pmu); + int i; + + /* on some platforms which uses cpus map, cpuid can be NULL for + * PMUs other than CORE PMUs. + */ + if (!cpuid) + return NULL; + + i = 0; + for (;;) { + const struct pmu_events_map *map = &pmu_events_map[i++]; + if (!map->arch) + break; + + if (!strcmp_cpuid_str(map->cpuid, cpuid)) { + table = &map->metric_table; + break; + } + } + free(cpuid); + return table; } const struct pmu_events_table *find_core_events_table(const char *arch, const char *cpuid)
@@ -667,14 +792,20 @@ const struct pmu_events_table *find_core_events_table(const char *arch, const ch tables->arch; tables++) { if (!strcmp(tables->arch, arch) && !strcmp_cpuid_str(tables->cpuid, cpuid)) - return &tables->table; + return &tables->event_table; } return NULL; } const struct pmu_metrics_table *find_core_metrics_table(const char *arch, const char *cpuid) { - return (struct pmu_metrics_table *)find_core_events_table(arch, cpuid); + for (const struct pmu_events_map *tables = &pmu_events_map[0]; + tables->arch; + tables++) { + if (!strcmp(tables->arch, arch) && !strcmp_cpuid_str(tables->cpuid, cpuid)) + return &tables->metric_table; + } + return NULL; } int pmu_for_each_core_event(pmu_event_iter_fn fn, void *data)
@@ -682,7 +813,7 @@ int pmu_for_each_core_event(pmu_event_iter_fn fn, void *data) for (const struct pmu_events_map *tables = &pmu_events_map[0]; tables->arch; tables++) { - int ret = pmu_events_table_for_each_event(&tables->table, fn, data); + int ret = pmu_events_table_for_each_event(&tables->event_table, fn, data); if (ret) return ret;
@@ -695,8 +826,7 @@ int pmu_for_each_core_metric(pmu_metric_iter_fn fn, void *data) for (const struct pmu_events_map *tables = &pmu_events_map[0]; tables->arch; tables++) { - int ret = pmu_metrics_table_for_each_metric( - (struct pmu_metrics_table *)&tables->table, fn, data); + int ret = pmu_metrics_table_for_each_metric(&tables->metric_table, fn, data); if (ret) return ret;
@@ -710,7 +840,7 @@ const struct pmu_events_table *find_sys_events_table(const char *name) tables->name; tables++) { if (!strcmp(tables->name, name)) - return &tables->table; + return &tables->event_table; } return NULL; }
@@ -720,7 +850,7 @@ int pmu_for_each_sys_event(pmu_event_iter_fn fn, void *data) for (const struct pmu_sys_events *tables = &pmu_sys_event_tables[0]; tables->name; tables++) { - int ret = pmu_events_table_for_each_event(&tables->table, fn, data); + int ret = pmu_events_table_for_each_event(&tables->event_table, fn, data); if (ret) return ret;
@@ -733,8 +863,7 @@ int pmu_for_each_sys_metric(pmu_metric_iter_fn fn, void *data) for (const struct pmu_sys_events *tables = &pmu_sys_event_tables[0]; tables->name; tables++) { - int ret = pmu_metrics_table_for_each_metric( - (struct pmu_metrics_table *)&tables->table, fn, data); + int ret = pmu_metrics_table_for_each_metric(&tables->metric_table, fn, data); if (ret) return ret;
@@ -809,6 +938,7 @@ struct compact_pmu_event { arch_path = f'{_args.starting_dir}/{arch}' ftw(arch_path, [], process_one_file) print_pending_events() + print_pending_metrics() print_mapping_table(archs) print_system_mapping_table()
tools/perf/tests/pmu-events.c +2 -1
@@ -469,7 +469,8 @@ static int test__pmu_event_table_sys_callback(const struct pmu_event *pe,static int test__pmu_event_table (struct test_suite *test__maybe_unused ,int subtest__maybe_unused ) { -const struct pmu_events_table *sys_event_table =find_sys_events_table ("pme_test_soc_sys" ); +const struct pmu_events_table *sys_event_table = +find_sys_events_table ("pmu_events__test_soc_sys" );const struct pmu_events_table *table =find_core_events_table ("testarch" ,"testcpu" );int map_events =0 , expected_events, err;
Signature · 1 line
2.39.0.314.g84b9a713c41-goog
John Garry
reviewerRe: [PATCH v2 8/9] perf jevents: Generate metrics and events as separate tables
Quoted 7 lines · Ian Rogers
Turn a perf json event into an event, metric or both. This reduces the number of events needed to scan to find an event or metric. As events no longer need the relatively seldom used metric fields, 4 bytes is saved per event. This reduces the big C string's size by 335kb (14.8%) on x86.
Signed-off-by Ian Rogers<[email protected]>
It would have been good to show an example of how the output changes. I could not apply the series (see cover), and knowing what to expect makes reviewing the code easier...
Thanks,
John
Ian Rogers
reviewerRe: [PATCH v2 8/9] perf jevents: Generate metrics and events as separate tables
Quoted 8 lines · John Garry
Quoted 6 lines · Ian Rogers
Turn a perf json event into an event, metric or both. This reduces the number of events needed to scan to find an event or metric. As events no longer need the relatively seldom used metric fields, 4 bytes is saved per event. This reduces the big C string's size by 335kb (14.8%) on x86.
Quoted 8 lines
Quoted 1 line
Signed-off-by Ian Rogers<[email protected]>It would have been good to show an example of how the output changes. I could not apply the series (see cover), and knowing what to expect makes reviewing the code easier...
Thanks,
John
Thanks, will add in v3.
Ian
Ian Rogers
reviewerPATCH v2 9/9 perf jevents: Add model list option
In reply to Ian RogersThis allows the set of generated jevents events and metrics be limited to a subset of the model names. Appropriate if trying to minimize the binary size where only a set of models are possible.
--- tools/perf/pmu-events/Build | 3 ++- tools/perf/pmu-events/jevents.py | 4 ++++ 2 files changed, 6 insertions(+), 1 deletion(-)
tools/perf/pmu-events/Build +2 -1
@@ -10,6 +10,7 @@ JEVENTS_PY = pmu-events/jevents.py ifeq ($(JEVENTS_ARCH),) JEVENTS_ARCH=$(SRCARCH) endif +JEVENTS_MODEL ?= all # # Locate/process JSON files in pmu-events/arch/
@@ -23,5 +24,5 @@ $(OUTPUT)pmu-events/pmu-events.c: pmu-events/empty-pmu-events.c else $(OUTPUT)pmu-events/pmu-events.c: $(JSON) $(JSON_TEST) $(JEVENTS_PY) pmu-events/metric.py $(call rule_mkdir) - $(Q)$(call echo-cmd,gen)$(PYTHON) $(JEVENTS_PY) $(JEVENTS_ARCH) pmu-events/arch $@ + $(Q)$(call echo-cmd,gen)$(PYTHON) $(JEVENTS_PY) $(JEVENTS_ARCH) $(JEVENTS_MODEL) pmu-events/arch $@ endif
tools/perf/pmu-events/jevents.py +4 -0
@@ -886,12 +886,16 @@ def main() -> None: action: Callable[[Sequence[str], os.DirEntry], None]) -> None: """Replicate the directory/file walking behavior of C's file tree walk.""" for item in os.scandir(path): + if (len(parents) == 0 and item.is_dir() and _args.model != 'all' and + 'test' not in item.name and item.name not in _args.model.split(',')): + continue action(parents, item) if item.is_dir(): ftw(item.path, parents + [item.name], action) ap = argparse.ArgumentParser() ap.add_argument('arch', help='Architecture name like x86') + ap.add_argument('model', help='Model such as skylake, normally "all"', default='all') ap.add_argument( 'starting_dir', type=dir_path,
Signature · 1 line
2.39.0.314.g84b9a713c41-goog
Ian Rogers
reviewerQuoted 31 lines · Ian Rogers
Add an optimization to jevents using the metric code, rewrite metrics
in terms of each other in order to minimize size and improve
readability. For example, on Power8other_stall_cpiis rewritten from:
"PM_CMPLU_STALL/PM_RUN_INST_CMPL-PM_CMPLU_STALL_BRU_CRU/PM_RUN_INST_CMPL-PM_CMPLU_STALL_FXU/PM_RUN_INST_CMPL-PM_CMPLU_STALL_VSU/PM_RUN_INST_CMPL-PM_CMPLU_STALL_LSU/PM_RUN_INST_CMPL-PM_CMPLU_STALL_NTCG_FLUSH/PM_RUN_INST_CMPL-PM_CMPLU_STALL_NO_NTF/PM_RUN_INST_CMPL" to:
"stall_cpi-bru_cru_stall_cpi-fxu_stall_cpi-vsu_stall_cpi-lsu_stall_cpi-ntcg_flush_cpi-no_ntf_stall_cpi"
Which more closely matches the definition on Power9.A limitation of the substitutions are that they depend on strict equality and the shape of the tree. This means that for "a + b + c" then a substitution of "a + b" will succeed while "b + c" will fail (the LHS for "+ c" is "a + b" not just "b").
Separate out the events and metrics in the pmu-events tables saving 14.8% in the table size while making it that metrics no longer need to iterate over all events and vice versa. These changes remove evsel's direct metric support as the
pmu_eventno longer has a metric to populate it. This is a minor issue as the code wasn't working properly, metrics for this are rare and can still be properly ran using '-M'.Add an ability to just build certain models into the jevents generated
pmu-metrics.ccode. This functionality is appropriate for operating systems like ChromeOS, that aim to minimize binary size and know all the target CPU models.v2. Rebase. Modify the code that skips rewriting a metric with the
same name with itself, to make the name check case insensitive.
Quoted 10 lines
Ian Rogers (9):
perf jevents metric: Correct Function equality
perf jevents metric: Add ability to rewrite metrics in terms of others perf jevents: Rewrite metrics in the same file with each other perf pmu-events: Separate metric out ofpmu_event
perf stat: Remove evsel metric_name/expr
perf jevents: Combine table prefix and suffix writing
perf pmu-events: Introducepmu_metrics_table
perf jevents: Generate metrics and events as separate tables perf jevents: Add model list option
Ping. Looking for reviews.
Thanks,
Ian
Quoted 31 lines
tools/perf/arch/arm64/util/pmu.c | 23 +- tools/perf/arch/powerpc/util/header.c | 4 +- tools/perf/builtin-list.c | 20 +- tools/perf/builtin-stat.c | 1 - tools/perf/pmu-events/Build | 3 +- tools/perf/pmu-events/empty-pmu-events.c | 111 ++++++- tools/perf/pmu-events/jevents.py | 353 ++++++++++++++++++----- tools/perf/pmu-events/metric.py | 79 ++++- tools/perf/pmu-events/metric_test.py | 10 + tools/perf/pmu-events/pmu-events.h | 26 +- tools/perf/tests/expand-cgroup.c | 4 +- tools/perf/tests/parse-metric.c | 4 +- tools/perf/tests/pmu-events.c | 68 ++--- tools/perf/util/cgroup.c | 1 - tools/perf/util/evsel.c | 2 - tools/perf/util/evsel.h | 2 - tools/perf/util/metricgroup.c | 203 +++++++------ tools/perf/util/metricgroup.h | 4 +- tools/perf/util/parse-events.c | 2 - tools/perf/util/pmu.c | 44 +-- tools/perf/util/pmu.h | 10 +- tools/perf/util/print-events.c | 32 +- tools/perf/util/print-events.h | 3 +- tools/perf/util/python.c | 7 - tools/perf/util/stat-shadow.c | 112 ------- tools/perf/util/stat.h | 1 - 26 files changed, 666 insertions(+), 463 deletions(-)--
2.39.0.314.g84b9a713c41-goog
John Garry
reviewerQuoted 22 lines · Ian Rogers
Add an optimization to jevents using the metric code, rewrite metrics
in terms of each other in order to minimize size and improve
readability. For example, on Power8other_stall_cpiis rewritten from:
"PM_CMPLU_STALL/PM_RUN_INST_CMPL-PM_CMPLU_STALL_BRU_CRU/PM_RUN_INST_CMPL-PM_CMPLU_STALL_FXU/PM_RUN_INST_CMPL-PM_CMPLU_STALL_VSU/PM_RUN_INST_CMPL-PM_CMPLU_STALL_LSU/PM_RUN_INST_CMPL-PM_CMPLU_STALL_NTCG_FLUSH/PM_RUN_INST_CMPL-PM_CMPLU_STALL_NO_NTF/PM_RUN_INST_CMPL" to:
"stall_cpi-bru_cru_stall_cpi-fxu_stall_cpi-vsu_stall_cpi-lsu_stall_cpi-ntcg_flush_cpi-no_ntf_stall_cpi"
Which more closely matches the definition on Power9.A limitation of the substitutions are that they depend on strict equality and the shape of the tree. This means that for "a + b + c" then a substitution of "a + b" will succeed while "b + c" will fail (the LHS for "+ c" is "a + b" not just "b").
Separate out the events and metrics in the pmu-events tables saving 14.8% in the table size while making it that metrics no longer need to iterate over all events and vice versa. These changes remove evsel's direct metric support as the
pmu_eventno longer has a metric to populate it. This is a minor issue as the code wasn't working properly, metrics for this are rare and can still be properly ran using '-M'.
Quoted 4 lines
Add an ability to just build certain models into the jevents generated
pmu-metrics.ccode. This functionality is appropriate for operating systems like ChromeOS, that aim to minimize binary size and know all the target CPU models.
From a glance, this does not look like it would work for arm64. As I see in the code, we check the model in the arch folder for the test to see if built. For arm64, as it uses arch/implementator/model folder org, and not just arch/model (like x86)
So on the assumption that it does not work for arm64 (or just any arch which uses arch/implementator/model folder org), it would be nice to have that feature also. Or maybe also support not just specifying model but also implementator.
Quoted 4 lines
v2. Rebase. Modify the code that skips rewriting a metric with the
same name with itself, to make the name check case insensitive.
Unfortunately you might need another rebase as this does not apply to acme perf/core (if that is what you want), now for me at:
5670ebf54bd2 (HEAD, origin/tmp.perf/core, origin/perf/core, perf/core) perf cs-etm: Ensure that Coresight timestamps don't go backwards
Quoted 10 lines
Ian Rogers (9):
perf jevents metric: Correct Function equality
perf jevents metric: Add ability to rewrite metrics in terms of others perf jevents: Rewrite metrics in the same file with each other perf pmu-events: Separate metric out ofpmu_event
perf stat: Remove evsel metric_name/expr
perf jevents: Combine table prefix and suffix writing
perf pmu-events: Introducepmu_metrics_table
perf jevents: Generate metrics and events as separate tables perf jevents: Add model list option
Thanks,
John
Ian Rogers
reviewerQuoted 34 lines · John Garry
Quoted 22 lines · Ian Rogers
Add an optimization to jevents using the metric code, rewrite metrics
in terms of each other in order to minimize size and improve
readability. For example, on Power8other_stall_cpiis rewritten from:
"PM_CMPLU_STALL/PM_RUN_INST_CMPL-PM_CMPLU_STALL_BRU_CRU/PM_RUN_INST_CMPL-PM_CMPLU_STALL_FXU/PM_RUN_INST_CMPL-PM_CMPLU_STALL_VSU/PM_RUN_INST_CMPL-PM_CMPLU_STALL_LSU/PM_RUN_INST_CMPL-PM_CMPLU_STALL_NTCG_FLUSH/PM_RUN_INST_CMPL-PM_CMPLU_STALL_NO_NTF/PM_RUN_INST_CMPL" to:
"stall_cpi-bru_cru_stall_cpi-fxu_stall_cpi-vsu_stall_cpi-lsu_stall_cpi-ntcg_flush_cpi-no_ntf_stall_cpi"
Which more closely matches the definition on Power9.A limitation of the substitutions are that they depend on strict equality and the shape of the tree. This means that for "a + b + c" then a substitution of "a + b" will succeed while "b + c" will fail (the LHS for "+ c" is "a + b" not just "b").
Separate out the events and metrics in the pmu-events tables saving 14.8% in the table size while making it that metrics no longer need to iterate over all events and vice versa. These changes remove evsel's direct metric support as the
pmu_eventno longer has a metric to populate it. This is a minor issue as the code wasn't working properly, metrics for this are rare and can still be properly ran using '-M'.Quoted 4 lines
Add an ability to just build certain models into the jevents generated
pmu-metrics.ccode. This functionality is appropriate for operating systems like ChromeOS, that aim to minimize binary size and know all the target CPU models.From a glance, this does not look like it would work for arm64. As I
see in the code, we check the model in the arch folder for the test to see if built. For arm64, as it usesarch/implementator/modelfolder org, and not justarch/model(like x86)
Quoted 4 lines
So on the assumption that it does not work for arm64 (or just any arch which uses
arch/implementator/modelfolder org), it would be nice to have that feature also. Or maybe also support not just specifying model but also implementator.
Hmm.. this is tricky as x86 isn't following the implementor pattern. I will tweak the comment for the ARM64 case where --model will select an implementor.
Quoted 11 lines
Quoted 4 lines
v2. Rebase. Modify the code that skips rewriting a metric with the
same name with itself, to make the name check case insensitive.Unfortunately you might need another rebase as this does not apply to acme perf/core (if that is what you want), now for me at:
5670ebf54bd2(HEAD, origin/tmp.perf/core, origin/perf/core, perf/core) perf cs-etm: Ensure that Coresight timestamps don't go backwards
Will do, thanks!
Ian
Quoted 13 lines
Quoted 10 lines
Ian Rogers (9):
perf jevents metric: Correct Function equality
perf jevents metric: Add ability to rewrite metrics in terms of others perf jevents: Rewrite metrics in the same file with each other perf pmu-events: Separate metric out ofpmu_event
perf stat: Remove evsel metric_name/expr
perf jevents: Combine table prefix and suffix writing
perf pmu-events: Introducepmu_metrics_table
perf jevents: Generate metrics and events as separate tables perf jevents: Add model list optionThanks,
John