Commit cd657187 authored by Ingo Molnar's avatar Ingo Molnar

Merge tag 'perf-urgent-for-mingo' of...

Merge tag 'perf-urgent-for-mingo' of git://git.kernel.org/pub/scm/linux/kernel/git/acme/linux into perf/urgent

Pull perf/urgent fixes from Arnaldo Carvalho de Melo:

 * Add color overhead for stdio output buffer, which fixes
   --stdio output being chopped up on the hot (red) entries,
   fix from Jiri Olsa.

 * Get 'perf record -g -a sleep 1' working again, removing the
   need for -- separating perf options from the workload, restoring
   ages old behaviour, fix from Jiri Olsa.
   More patches allowing ~/.perfconfig setting up of default
   callchain collecting method ("fp" or "dwarf") left for next
   merge window.

 * Fixup mmap event consumption, where we were acking the
   consumption by writing the tail before actually accessing
   the event, which could lead to using overwritten records
   in things like 'perf record --call-graph'. From Zhouyi Zhou.
Signed-off-by: default avatarArnaldo Carvalho de Melo <acme@redhat.com>
parents d17cccbe 8e50d384
......@@ -90,8 +90,20 @@ OPTIONS
Number of mmap data pages. Must be a power of two.
-g::
Enables call-graph (stack chain/backtrace) recording.
--call-graph::
Do call-graph (stack chain/backtrace) recording.
Setup and enable call-graph (stack chain/backtrace) recording,
implies -g.
Allows specifying "fp" (frame pointer) or "dwarf"
(DWARF's CFI - Call Frame Information) as the method to collect
the information used to show the call graphs.
In some systems, where binaries are build with gcc
--fomit-frame-pointer, using the "fp" method will produce bogus
call graphs, using "dwarf", if available (perf tools linked to
the libunwind library) should be used instead.
-q::
--quiet::
......
......@@ -140,20 +140,12 @@ Default is to monitor all CPUS.
--asm-raw::
Show raw instruction encoding of assembly instructions.
-G [type,min,order]::
-G::
Enables call-graph (stack chain/backtrace) recording.
--call-graph::
Display call chains using type, min percent threshold and order.
type can be either:
- flat: single column, linear exposure of call chains.
- graph: use a graph tree, displaying absolute overhead rates.
- fractal: like graph, but displays relative rates. Each branch of
the tree is considered as a new profiled object.
order can be either:
- callee: callee based call graph.
- caller: inverted caller based call graph.
Default: fractal,0.5,callee.
Setup and enable call-graph (stack chain/backtrace) recording,
implies -G.
--ignore-callees=<regex>::
Ignore callees of the function(s) matching the given regex.
......
......@@ -888,11 +888,18 @@ static s64 perf_kvm__mmap_read_idx(struct perf_kvm_stat *kvm, int idx,
while ((event = perf_evlist__mmap_read(kvm->evlist, idx)) != NULL) {
err = perf_evlist__parse_sample(kvm->evlist, event, &sample);
if (err) {
perf_evlist__mmap_consume(kvm->evlist, idx);
pr_err("Failed to parse sample\n");
return -1;
}
err = perf_session_queue_event(kvm->session, event, &sample, 0);
/*
* FIXME: Here we can't consume the event, as perf_session_queue_event will
* point to it, and it'll get possibly overwritten by the kernel.
*/
perf_evlist__mmap_consume(kvm->evlist, idx);
if (err) {
pr_err("Failed to enqueue sample: %d\n", err);
return -1;
......
......@@ -712,21 +712,12 @@ static int get_stack_size(char *str, unsigned long *_size)
}
#endif /* LIBUNWIND_SUPPORT */
int record_parse_callchain_opt(const struct option *opt,
const char *arg, int unset)
int record_parse_callchain(const char *arg, struct perf_record_opts *opts)
{
struct perf_record_opts *opts = opt->value;
char *tok, *name, *saveptr = NULL;
char *buf;
int ret = -1;
/* --no-call-graph */
if (unset)
return 0;
/* We specified default option if none is provided. */
BUG_ON(!arg);
/* We need buffer that we know we can write to. */
buf = malloc(strlen(arg) + 1);
if (!buf)
......@@ -764,13 +755,9 @@ int record_parse_callchain_opt(const struct option *opt,
ret = get_stack_size(tok, &size);
opts->stack_dump_size = size;
}
if (!ret)
pr_debug("callchain: stack dump size %d\n",
opts->stack_dump_size);
#endif /* LIBUNWIND_SUPPORT */
} else {
pr_err("callchain: Unknown -g option "
pr_err("callchain: Unknown --call-graph option "
"value: %s\n", arg);
break;
}
......@@ -778,13 +765,52 @@ int record_parse_callchain_opt(const struct option *opt,
} while (0);
free(buf);
return ret;
}
if (!ret)
static void callchain_debug(struct perf_record_opts *opts)
{
pr_debug("callchain: type %d\n", opts->call_graph);
if (opts->call_graph == CALLCHAIN_DWARF)
pr_debug("callchain: stack dump size %d\n",
opts->stack_dump_size);
}
int record_parse_callchain_opt(const struct option *opt,
const char *arg,
int unset)
{
struct perf_record_opts *opts = opt->value;
int ret;
/* --no-call-graph */
if (unset) {
opts->call_graph = CALLCHAIN_NONE;
pr_debug("callchain: disabled\n");
return 0;
}
ret = record_parse_callchain(arg, opts);
if (!ret)
callchain_debug(opts);
return ret;
}
int record_callchain_opt(const struct option *opt,
const char *arg __maybe_unused,
int unset __maybe_unused)
{
struct perf_record_opts *opts = opt->value;
if (opts->call_graph == CALLCHAIN_NONE)
opts->call_graph = CALLCHAIN_FP;
callchain_debug(opts);
return 0;
}
static const char * const record_usage[] = {
"perf record [<options>] [<command>]",
"perf record [<options>] -- <command> [<options>]",
......@@ -813,12 +839,12 @@ static struct perf_record record = {
},
};
#define CALLCHAIN_HELP "do call-graph (stack chain/backtrace) recording: "
#define CALLCHAIN_HELP "setup and enables call-graph (stack chain/backtrace) recording: "
#ifdef LIBUNWIND_SUPPORT
const char record_callchain_help[] = CALLCHAIN_HELP "[fp] dwarf";
const char record_callchain_help[] = CALLCHAIN_HELP "fp dwarf";
#else
const char record_callchain_help[] = CALLCHAIN_HELP "[fp]";
const char record_callchain_help[] = CALLCHAIN_HELP "fp";
#endif
/*
......@@ -858,9 +884,12 @@ const struct option record_options[] = {
"number of mmap data pages"),
OPT_BOOLEAN(0, "group", &record.opts.group,
"put the counters into a counter group"),
OPT_CALLBACK_DEFAULT('g', "call-graph", &record.opts,
OPT_CALLBACK_NOOPT('g', NULL, &record.opts,
NULL, "enables call-graph recording" ,
&record_callchain_opt),
OPT_CALLBACK(0, "call-graph", &record.opts,
"mode[,dump_size]", record_callchain_help,
&record_parse_callchain_opt, "fp"),
&record_parse_callchain_opt),
OPT_INCR('v', "verbose", &verbose,
"be more verbose (show counter open errors, etc)"),
OPT_BOOLEAN('q', "quiet", &quiet, "don't print any message"),
......
......@@ -810,7 +810,7 @@ static void perf_top__mmap_read_idx(struct perf_top *top, int idx)
ret = perf_evlist__parse_sample(top->evlist, event, &sample);
if (ret) {
pr_err("Can't parse sample, err = %d\n", ret);
continue;
goto next_event;
}
evsel = perf_evlist__id2evsel(session->evlist, sample.id);
......@@ -825,13 +825,13 @@ static void perf_top__mmap_read_idx(struct perf_top *top, int idx)
case PERF_RECORD_MISC_USER:
++top->us_samples;
if (top->hide_user_symbols)
continue;
goto next_event;
machine = &session->machines.host;
break;
case PERF_RECORD_MISC_KERNEL:
++top->kernel_samples;
if (top->hide_kernel_symbols)
continue;
goto next_event;
machine = &session->machines.host;
break;
case PERF_RECORD_MISC_GUEST_KERNEL:
......@@ -847,7 +847,7 @@ static void perf_top__mmap_read_idx(struct perf_top *top, int idx)
*/
/* Fall thru */
default:
continue;
goto next_event;
}
......@@ -859,6 +859,8 @@ static void perf_top__mmap_read_idx(struct perf_top *top, int idx)
machine__process_event(machine, event);
} else
++session->stats.nr_unknown_events;
next_event:
perf_evlist__mmap_consume(top->evlist, idx);
}
}
......@@ -1016,16 +1018,16 @@ static int __cmd_top(struct perf_top *top)
}
static int
parse_callchain_opt(const struct option *opt, const char *arg, int unset)
callchain_opt(const struct option *opt, const char *arg, int unset)
{
/*
* --no-call-graph
*/
if (unset)
return 0;
symbol_conf.use_callchain = true;
return record_callchain_opt(opt, arg, unset);
}
static int
parse_callchain_opt(const struct option *opt, const char *arg, int unset)
{
symbol_conf.use_callchain = true;
return record_parse_callchain_opt(opt, arg, unset);
}
......@@ -1106,9 +1108,12 @@ int cmd_top(int argc, const char **argv, const char *prefix __maybe_unused)
"sort by key(s): pid, comm, dso, symbol, parent, weight, local_weight"),
OPT_BOOLEAN('n', "show-nr-samples", &symbol_conf.show_nr_samples,
"Show a column with the number of samples"),
OPT_CALLBACK_DEFAULT('G', "call-graph", &top.record_opts,
OPT_CALLBACK_NOOPT('G', NULL, &top.record_opts,
NULL, "enables call-graph recording",
&callchain_opt),
OPT_CALLBACK(0, "call-graph", &top.record_opts,
"mode[,dump_size]", record_callchain_help,
&parse_callchain_opt, "fp"),
&parse_callchain_opt),
OPT_CALLBACK(0, "ignore-callees", NULL, "regex",
"ignore callees of these functions in call graphs",
report_parse_ignore_callees_opt),
......
......@@ -987,7 +987,7 @@ static int trace__run(struct trace *trace, int argc, const char **argv)
err = perf_evlist__parse_sample(evlist, event, &sample);
if (err) {
fprintf(trace->output, "Can't parse sample, err = %d, skipping...\n", err);
continue;
goto next_event;
}
if (trace->base_time == 0)
......@@ -1001,18 +1001,20 @@ static int trace__run(struct trace *trace, int argc, const char **argv)
evsel = perf_evlist__id2evsel(evlist, sample.id);
if (evsel == NULL) {
fprintf(trace->output, "Unknown tp ID %" PRIu64 ", skipping...\n", sample.id);
continue;
goto next_event;
}
if (sample.raw_data == NULL) {
fprintf(trace->output, "%s sample with no payload for tid: %d, cpu %d, raw_size=%d, skipping...\n",
perf_evsel__name(evsel), sample.tid,
sample.cpu, sample.raw_size);
continue;
goto next_event;
}
handler = evsel->handler.func;
handler(trace, evsel, &sample);
next_event:
perf_evlist__mmap_consume(evlist, i);
if (done)
goto out_unmap_evlist;
......
......@@ -290,6 +290,7 @@ static int process_events(struct machine *machine, struct perf_evlist *evlist,
for (i = 0; i < evlist->nr_mmaps; i++) {
while ((event = perf_evlist__mmap_read(evlist, i)) != NULL) {
ret = process_event(machine, evlist, event, state);
perf_evlist__mmap_consume(evlist, i);
if (ret < 0)
return ret;
}
......
......@@ -36,6 +36,7 @@ static int find_comm(struct perf_evlist *evlist, const char *comm)
(pid_t)event->comm.tid == getpid() &&
strcmp(event->comm.comm, comm) == 0)
found += 1;
perf_evlist__mmap_consume(evlist, i);
}
}
return found;
......
......@@ -122,6 +122,7 @@ int test__basic_mmap(void)
goto out_munmap;
}
nr_events[evsel->idx]++;
perf_evlist__mmap_consume(evlist, 0);
}
err = 0;
......
......@@ -77,8 +77,10 @@ int test__syscall_open_tp_fields(void)
++nr_events;
if (type != PERF_RECORD_SAMPLE)
if (type != PERF_RECORD_SAMPLE) {
perf_evlist__mmap_consume(evlist, i);
continue;
}
err = perf_evsel__parse_sample(evsel, event, &sample);
if (err) {
......
......@@ -263,6 +263,8 @@ int test__PERF_RECORD(void)
type);
++errs;
}
perf_evlist__mmap_consume(evlist, i);
}
}
......
......@@ -122,7 +122,7 @@ int test__perf_time_to_tsc(void)
if (event->header.type != PERF_RECORD_COMM ||
(pid_t)event->comm.pid != getpid() ||
(pid_t)event->comm.tid != getpid())
continue;
goto next_event;
if (strcmp(event->comm.comm, comm1) == 0) {
CHECK__(perf_evsel__parse_sample(evsel, event,
......@@ -134,6 +134,8 @@ int test__perf_time_to_tsc(void)
&sample));
comm2_time = sample.time;
}
next_event:
perf_evlist__mmap_consume(evlist, i);
}
}
......
......@@ -78,7 +78,7 @@ static int __test__sw_clock_freq(enum perf_sw_ids clock_id)
struct perf_sample sample;
if (event->header.type != PERF_RECORD_SAMPLE)
continue;
goto next_event;
err = perf_evlist__parse_sample(evlist, event, &sample);
if (err < 0) {
......@@ -88,6 +88,8 @@ static int __test__sw_clock_freq(enum perf_sw_ids clock_id)
total_periods += sample.period;
nr_samples++;
next_event:
perf_evlist__mmap_consume(evlist, 0);
}
if ((u64) nr_samples == total_periods) {
......
......@@ -96,10 +96,10 @@ int test__task_exit(void)
retry:
while ((event = perf_evlist__mmap_read(evlist, 0)) != NULL) {
if (event->header.type != PERF_RECORD_EXIT)
continue;
if (event->header.type == PERF_RECORD_EXIT)
nr_exit++;
perf_evlist__mmap_consume(evlist, 0);
}
if (!exited || !nr_exit) {
......
......@@ -315,8 +315,7 @@ static inline void advance_hpp(struct perf_hpp *hpp, int inc)
}
static int hist_entry__period_snprintf(struct perf_hpp *hpp,
struct hist_entry *he,
bool color)
struct hist_entry *he)
{
const char *sep = symbol_conf.field_sep;
struct perf_hpp_fmt *fmt;
......@@ -338,7 +337,7 @@ static int hist_entry__period_snprintf(struct perf_hpp *hpp,
} else
first = false;
if (color && fmt->color)
if (perf_hpp__use_color() && fmt->color)
ret = fmt->color(fmt, hpp, he);
else
ret = fmt->entry(fmt, hpp, he);
......@@ -358,12 +357,11 @@ static int hist_entry__fprintf(struct hist_entry *he, size_t size,
.buf = bf,
.size = size,
};
bool color = !symbol_conf.field_sep;
if (size == 0 || size > bfsz)
size = hpp.size = bfsz;
ret = hist_entry__period_snprintf(&hpp, he, color);
ret = hist_entry__period_snprintf(&hpp, he);
hist_entry__sort_snprintf(he, bf + ret, size - ret, hists);
ret = fprintf(fp, "%s\n", bf);
......@@ -482,6 +480,7 @@ size_t hists__fprintf(struct hists *hists, bool show_header, int max_rows,
print_entries:
linesz = hists__sort_list_width(hists) + 3 + 1;
linesz += perf_hpp__color_overhead();
line = malloc(linesz);
if (line == NULL) {
ret = -1;
......
......@@ -147,6 +147,9 @@ static inline void callchain_cursor_advance(struct callchain_cursor *cursor)
struct option;
int record_parse_callchain(const char *arg, struct perf_record_opts *opts);
int record_parse_callchain_opt(const struct option *opt, const char *arg, int unset);
int record_callchain_opt(const struct option *opt, const char *arg, int unset);
extern const char record_callchain_help[];
#endif /* __PERF_CALLCHAIN_H */
......@@ -545,12 +545,19 @@ union perf_event *perf_evlist__mmap_read(struct perf_evlist *evlist, int idx)
md->prev = old;
if (!evlist->overwrite)
perf_mmap__write_tail(md, old);
return event;
}
void perf_evlist__mmap_consume(struct perf_evlist *evlist, int idx)
{
if (!evlist->overwrite) {
struct perf_mmap *md = &evlist->mmap[idx];
unsigned int old = md->prev;
perf_mmap__write_tail(md, old);
}
}
static void __perf_evlist__munmap(struct perf_evlist *evlist, int idx)
{
if (evlist->mmap[idx].base != NULL) {
......
......@@ -89,6 +89,8 @@ struct perf_sample_id *perf_evlist__id2sid(struct perf_evlist *evlist, u64 id);
union perf_event *perf_evlist__mmap_read(struct perf_evlist *self, int idx);
void perf_evlist__mmap_consume(struct perf_evlist *evlist, int idx);
int perf_evlist__open(struct perf_evlist *evlist);
void perf_evlist__close(struct perf_evlist *evlist);
......
......@@ -5,6 +5,7 @@
#include <pthread.h>
#include "callchain.h"
#include "header.h"
#include "color.h"
extern struct callchain_param callchain_param;
......@@ -175,6 +176,18 @@ void perf_hpp__init(void);
void perf_hpp__column_register(struct perf_hpp_fmt *format);
void perf_hpp__column_enable(unsigned col);
static inline size_t perf_hpp__use_color(void)
{
return !symbol_conf.field_sep;
}
static inline size_t perf_hpp__color_overhead(void)
{
return perf_hpp__use_color() ?
(COLOR_MAXLEN + sizeof(PERF_COLOR_RESET)) * PERF_HPP__MAX_INDEX
: 0;
}
struct perf_evlist;
struct hist_browser_timer {
......
......@@ -822,6 +822,8 @@ static PyObject *pyrf_evlist__read_on_cpu(struct pyrf_evlist *pevlist,
PyObject *pyevent = pyrf_event__new(event);
struct pyrf_event *pevent = (struct pyrf_event *)pyevent;
perf_evlist__mmap_consume(evlist, cpu);
if (pyevent == NULL)
return PyErr_NoMemory();
......
Markdown is supported
0%
or
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment