diff options
author | Ingo Molnar <mingo@elte.hu> | 2011-03-11 09:28:31 +0100 |
---|---|---|
committer | Ingo Molnar <mingo@elte.hu> | 2011-03-11 09:28:31 +0100 |
commit | 137ee20ddd10fdc20600c389fe63edab0c39cb1a (patch) | |
tree | f2f9db93cbbe1929d2c3380658a546a748181dc9 /tools/perf/util | |
parent | 4a0b1665db09cf2da9ad7d0f12da386373c10bfa (diff) | |
parent | 1c0b04d10bbe35279c50e3b36cf5b8ec2a0050d8 (diff) | |
download | lwn-137ee20ddd10fdc20600c389fe63edab0c39cb1a.tar.gz lwn-137ee20ddd10fdc20600c389fe63edab0c39cb1a.zip |
Merge branch 'perf/core' of git://git.kernel.org/pub/scm/linux/kernel/git/acme/linux-2.6 into perf/core
Diffstat (limited to 'tools/perf/util')
-rw-r--r-- | tools/perf/util/evlist.c | 26 | ||||
-rw-r--r-- | tools/perf/util/evlist.h | 4 | ||||
-rw-r--r-- | tools/perf/util/evsel.c | 21 | ||||
-rw-r--r-- | tools/perf/util/evsel.h | 9 | ||||
-rw-r--r-- | tools/perf/util/header.c | 422 | ||||
-rw-r--r-- | tools/perf/util/header.h | 48 | ||||
-rw-r--r-- | tools/perf/util/session.c | 74 | ||||
-rw-r--r-- | tools/perf/util/session.h | 2 | ||||
-rw-r--r-- | tools/perf/util/top.c | 48 |
9 files changed, 278 insertions, 376 deletions
diff --git a/tools/perf/util/evlist.c b/tools/perf/util/evlist.c index 190c64c6e266..d852cefa20de 100644 --- a/tools/perf/util/evlist.c +++ b/tools/perf/util/evlist.c @@ -19,7 +19,7 @@ #include <linux/hash.h> #define FD(e, x, y) (*(int *)xyarray__entry(e->fd, x, y)) -#define SID(e, x, y) xyarray__entry(e->id, x, y) +#define SID(e, x, y) xyarray__entry(e->sample_id, x, y) void perf_evlist__init(struct perf_evlist *evlist, struct cpu_map *cpus, struct thread_map *threads) @@ -106,8 +106,9 @@ void perf_evlist__add_pollfd(struct perf_evlist *evlist, int fd) evlist->nr_fds++; } -void perf_evlist__id_hash(struct perf_evlist *evlist, struct perf_evsel *evsel, - int cpu, int thread, u64 id) +static void perf_evlist__id_hash(struct perf_evlist *evlist, + struct perf_evsel *evsel, + int cpu, int thread, u64 id) { int hash; struct perf_sample_id *sid = SID(evsel, cpu, thread); @@ -118,9 +119,16 @@ void perf_evlist__id_hash(struct perf_evlist *evlist, struct perf_evsel *evsel, hlist_add_head(&sid->node, &evlist->heads[hash]); } -static int perf_evlist__id_hash_fd(struct perf_evlist *evlist, - struct perf_evsel *evsel, - int cpu, int thread, int fd) +void perf_evlist__id_add(struct perf_evlist *evlist, struct perf_evsel *evsel, + int cpu, int thread, u64 id) +{ + perf_evlist__id_hash(evlist, evsel, cpu, thread, id); + evsel->id[evsel->ids++] = id; +} + +static int perf_evlist__id_add_fd(struct perf_evlist *evlist, + struct perf_evsel *evsel, + int cpu, int thread, int fd) { u64 read_data[4] = { 0, }; int id_idx = 1; /* The first entry is the counter value */ @@ -134,7 +142,7 @@ static int perf_evlist__id_hash_fd(struct perf_evlist *evlist, if (evsel->attr.read_format & PERF_FORMAT_TOTAL_TIME_RUNNING) ++id_idx; - perf_evlist__id_hash(evlist, evsel, cpu, thread, read_data[id_idx]); + perf_evlist__id_add(evlist, evsel, cpu, thread, read_data[id_idx]); return 0; } @@ -292,7 +300,7 @@ int perf_evlist__mmap(struct perf_evlist *evlist, int pages, bool overwrite) list_for_each_entry(evsel, &evlist->entries, node) { if ((evsel->attr.read_format & PERF_FORMAT_ID) && - evsel->id == NULL && + evsel->sample_id == NULL && perf_evsel__alloc_id(evsel, cpus->nr, threads->nr) < 0) return -ENOMEM; @@ -308,7 +316,7 @@ int perf_evlist__mmap(struct perf_evlist *evlist, int pages, bool overwrite) goto out_unmap; if ((evsel->attr.read_format & PERF_FORMAT_ID) && - perf_evlist__id_hash_fd(evlist, evsel, cpu, thread, fd) < 0) + perf_evlist__id_add_fd(evlist, evsel, cpu, thread, fd) < 0) goto out_unmap; } } diff --git a/tools/perf/util/evlist.h b/tools/perf/util/evlist.h index 078d51256085..8b1cb7a4c5f1 100644 --- a/tools/perf/util/evlist.h +++ b/tools/perf/util/evlist.h @@ -38,8 +38,8 @@ void perf_evlist__delete(struct perf_evlist *evlist); void perf_evlist__add(struct perf_evlist *evlist, struct perf_evsel *entry); int perf_evlist__add_default(struct perf_evlist *evlist); -void perf_evlist__id_hash(struct perf_evlist *evlist, struct perf_evsel *evsel, - int cpu, int thread, u64 id); +void perf_evlist__id_add(struct perf_evlist *evlist, struct perf_evsel *evsel, + int cpu, int thread, u64 id); int perf_evlist__alloc_pollfd(struct perf_evlist *evlist); void perf_evlist__add_pollfd(struct perf_evlist *evlist, int fd); diff --git a/tools/perf/util/evsel.c b/tools/perf/util/evsel.c index 8083d5126fca..662596afd7f1 100644 --- a/tools/perf/util/evsel.c +++ b/tools/perf/util/evsel.c @@ -41,8 +41,18 @@ int perf_evsel__alloc_fd(struct perf_evsel *evsel, int ncpus, int nthreads) int perf_evsel__alloc_id(struct perf_evsel *evsel, int ncpus, int nthreads) { - evsel->id = xyarray__new(ncpus, nthreads, sizeof(struct perf_sample_id)); - return evsel->id != NULL ? 0 : -ENOMEM; + evsel->sample_id = xyarray__new(ncpus, nthreads, sizeof(struct perf_sample_id)); + if (evsel->sample_id == NULL) + return -ENOMEM; + + evsel->id = zalloc(ncpus * nthreads * sizeof(u64)); + if (evsel->id == NULL) { + xyarray__delete(evsel->sample_id); + evsel->sample_id = NULL; + return -ENOMEM; + } + + return 0; } int perf_evsel__alloc_counts(struct perf_evsel *evsel, int ncpus) @@ -60,7 +70,9 @@ void perf_evsel__free_fd(struct perf_evsel *evsel) void perf_evsel__free_id(struct perf_evsel *evsel) { - xyarray__delete(evsel->id); + xyarray__delete(evsel->sample_id); + evsel->sample_id = NULL; + free(evsel->id); evsel->id = NULL; } @@ -79,7 +91,8 @@ void perf_evsel__exit(struct perf_evsel *evsel) { assert(list_empty(&evsel->node)); xyarray__delete(evsel->fd); - xyarray__delete(evsel->id); + xyarray__delete(evsel->sample_id); + free(evsel->id); } void perf_evsel__delete(struct perf_evsel *evsel) diff --git a/tools/perf/util/evsel.h b/tools/perf/util/evsel.h index 281b60e5fc7b..6710ab538342 100644 --- a/tools/perf/util/evsel.h +++ b/tools/perf/util/evsel.h @@ -49,12 +49,17 @@ struct perf_evsel { struct perf_event_attr attr; char *filter; struct xyarray *fd; - struct xyarray *id; + struct xyarray *sample_id; + u64 *id; struct perf_counts *counts; int idx; + int ids; struct hists hists; char *name; - void *priv; + union { + void *priv; + off_t id_offset; + }; struct cgroup_sel *cgrp; }; diff --git a/tools/perf/util/header.c b/tools/perf/util/header.c index 108b0db7bbef..5a72d421e211 100644 --- a/tools/perf/util/header.c +++ b/tools/perf/util/header.c @@ -9,6 +9,7 @@ #include <linux/kernel.h> #include "evlist.h" +#include "evsel.h" #include "util.h" #include "header.h" #include "../perf.h" @@ -19,89 +20,6 @@ static bool no_buildid_cache = false; -/* - * Create new perf.data header attribute: - */ -struct perf_header_attr *perf_header_attr__new(struct perf_event_attr *attr) -{ - struct perf_header_attr *self = malloc(sizeof(*self)); - - if (self != NULL) { - self->attr = *attr; - self->ids = 0; - self->size = 1; - self->id = malloc(sizeof(u64)); - if (self->id == NULL) { - free(self); - self = NULL; - } - } - - return self; -} - -void perf_header_attr__delete(struct perf_header_attr *self) -{ - free(self->id); - free(self); -} - -int perf_header_attr__add_id(struct perf_header_attr *self, u64 id) -{ - int pos = self->ids; - - self->ids++; - if (self->ids > self->size) { - int nsize = self->size * 2; - u64 *nid = realloc(self->id, nsize * sizeof(u64)); - - if (nid == NULL) - return -1; - - self->size = nsize; - self->id = nid; - } - self->id[pos] = id; - return 0; -} - -int perf_header__init(struct perf_header *self) -{ - self->size = 1; - self->attr = malloc(sizeof(void *)); - return self->attr == NULL ? -ENOMEM : 0; -} - -void perf_header__exit(struct perf_header *self) -{ - int i; - for (i = 0; i < self->attrs; ++i) - perf_header_attr__delete(self->attr[i]); - free(self->attr); -} - -int perf_header__add_attr(struct perf_header *self, - struct perf_header_attr *attr) -{ - if (self->frozen) - return -1; - - if (self->attrs == self->size) { - int nsize = self->size * 2; - struct perf_header_attr **nattr; - - nattr = realloc(self->attr, nsize * sizeof(void *)); - if (nattr == NULL) - return -1; - - self->size = nsize; - self->attr = nattr; - } - - self->attr[self->attrs++] = attr; - return 0; -} - static int event_count; static struct perf_trace_event_type *events; @@ -148,19 +66,19 @@ struct perf_file_attr { struct perf_file_section ids; }; -void perf_header__set_feat(struct perf_header *self, int feat) +void perf_header__set_feat(struct perf_header *header, int feat) { - set_bit(feat, self->adds_features); + set_bit(feat, header->adds_features); } -void perf_header__clear_feat(struct perf_header *self, int feat) +void perf_header__clear_feat(struct perf_header *header, int feat) { - clear_bit(feat, self->adds_features); + clear_bit(feat, header->adds_features); } -bool perf_header__has_feat(const struct perf_header *self, int feat) +bool perf_header__has_feat(const struct perf_header *header, int feat) { - return test_bit(feat, self->adds_features); + return test_bit(feat, header->adds_features); } static int do_write(int fd, const void *buf, size_t size) @@ -229,22 +147,22 @@ static int __dsos__write_buildid_table(struct list_head *head, pid_t pid, return 0; } -static int machine__write_buildid_table(struct machine *self, int fd) +static int machine__write_buildid_table(struct machine *machine, int fd) { int err; u16 kmisc = PERF_RECORD_MISC_KERNEL, umisc = PERF_RECORD_MISC_USER; - if (!machine__is_host(self)) { + if (!machine__is_host(machine)) { kmisc = PERF_RECORD_MISC_GUEST_KERNEL; umisc = PERF_RECORD_MISC_GUEST_USER; } - err = __dsos__write_buildid_table(&self->kernel_dsos, self->pid, + err = __dsos__write_buildid_table(&machine->kernel_dsos, machine->pid, kmisc, fd); if (err == 0) - err = __dsos__write_buildid_table(&self->user_dsos, - self->pid, umisc, fd); + err = __dsos__write_buildid_table(&machine->user_dsos, + machine->pid, umisc, fd); return err; } @@ -362,12 +280,12 @@ out_free: return err; } -static int dso__cache_build_id(struct dso *self, const char *debugdir) +static int dso__cache_build_id(struct dso *dso, const char *debugdir) { - bool is_kallsyms = self->kernel && self->long_name[0] != '/'; + bool is_kallsyms = dso->kernel && dso->long_name[0] != '/'; - return build_id_cache__add_b(self->build_id, sizeof(self->build_id), - self->long_name, debugdir, is_kallsyms); + return build_id_cache__add_b(dso->build_id, sizeof(dso->build_id), + dso->long_name, debugdir, is_kallsyms); } static int __dsos__cache_build_ids(struct list_head *head, const char *debugdir) @@ -382,14 +300,14 @@ static int __dsos__cache_build_ids(struct list_head *head, const char *debugdir) return err; } -static int machine__cache_build_ids(struct machine *self, const char *debugdir) +static int machine__cache_build_ids(struct machine *machine, const char *debugdir) { - int ret = __dsos__cache_build_ids(&self->kernel_dsos, debugdir); - ret |= __dsos__cache_build_ids(&self->user_dsos, debugdir); + int ret = __dsos__cache_build_ids(&machine->kernel_dsos, debugdir); + ret |= __dsos__cache_build_ids(&machine->user_dsos, debugdir); return ret; } -static int perf_session__cache_build_ids(struct perf_session *self) +static int perf_session__cache_build_ids(struct perf_session *session) { struct rb_node *nd; int ret; @@ -400,28 +318,28 @@ static int perf_session__cache_build_ids(struct perf_session *self) if (mkdir(debugdir, 0755) != 0 && errno != EEXIST) return -1; - ret = machine__cache_build_ids(&self->host_machine, debugdir); + ret = machine__cache_build_ids(&session->host_machine, debugdir); - for (nd = rb_first(&self->machines); nd; nd = rb_next(nd)) { + for (nd = rb_first(&session->machines); nd; nd = rb_next(nd)) { struct machine *pos = rb_entry(nd, struct machine, rb_node); ret |= machine__cache_build_ids(pos, debugdir); } return ret ? -1 : 0; } -static bool machine__read_build_ids(struct machine *self, bool with_hits) +static bool machine__read_build_ids(struct machine *machine, bool with_hits) { - bool ret = __dsos__read_build_ids(&self->kernel_dsos, with_hits); - ret |= __dsos__read_build_ids(&self->user_dsos, with_hits); + bool ret = __dsos__read_build_ids(&machine->kernel_dsos, with_hits); + ret |= __dsos__read_build_ids(&machine->user_dsos, with_hits); return ret; } -static bool perf_session__read_build_ids(struct perf_session *self, bool with_hits) +static bool perf_session__read_build_ids(struct perf_session *session, bool with_hits) { struct rb_node *nd; - bool ret = machine__read_build_ids(&self->host_machine, with_hits); + bool ret = machine__read_build_ids(&session->host_machine, with_hits); - for (nd = rb_first(&self->machines); nd; nd = rb_next(nd)) { + for (nd = rb_first(&session->machines); nd; nd = rb_next(nd)) { struct machine *pos = rb_entry(nd, struct machine, rb_node); ret |= machine__read_build_ids(pos, with_hits); } @@ -429,7 +347,7 @@ static bool perf_session__read_build_ids(struct perf_session *self, bool with_hi return ret; } -static int perf_header__adds_write(struct perf_header *self, +static int perf_header__adds_write(struct perf_header *header, struct perf_evlist *evlist, int fd) { int nr_sections; @@ -439,13 +357,13 @@ static int perf_header__adds_write(struct perf_header *self, u64 sec_start; int idx = 0, err; - session = container_of(self, struct perf_session, header); + session = container_of(header, struct perf_session, header); - if (perf_header__has_feat(self, HEADER_BUILD_ID && + if (perf_header__has_feat(header, HEADER_BUILD_ID && !perf_session__read_build_ids(session, true))) - perf_header__clear_feat(self, HEADER_BUILD_ID); + perf_header__clear_feat(header, HEADER_BUILD_ID); - nr_sections = bitmap_weight(self->adds_features, HEADER_FEAT_BITS); + nr_sections = bitmap_weight(header->adds_features, HEADER_FEAT_BITS); if (!nr_sections) return 0; @@ -455,10 +373,10 @@ static int perf_header__adds_write(struct perf_header *self, sec_size = sizeof(*feat_sec) * nr_sections; - sec_start = self->data_offset + self->data_size; + sec_start = header->data_offset + header->data_size; lseek(fd, sec_start + sec_size, SEEK_SET); - if (perf_header__has_feat(self, HEADER_TRACE_INFO)) { + if (perf_header__has_feat(header, HEADER_TRACE_INFO)) { struct perf_file_section *trace_sec; trace_sec = &feat_sec[idx++]; @@ -469,14 +387,14 @@ static int perf_header__adds_write(struct perf_header *self, trace_sec->size = lseek(fd, 0, SEEK_CUR) - trace_sec->offset; } - if (perf_header__has_feat(self, HEADER_BUILD_ID)) { + if (perf_header__has_feat(header, HEADER_BUILD_ID)) { struct perf_file_section *buildid_sec; buildid_sec = &feat_sec[idx++]; /* Write build-ids */ buildid_sec->offset = lseek(fd, 0, SEEK_CUR); - err = dsos__write_buildid_table(self, fd); + err = dsos__write_buildid_table(header, fd); if (err < 0) { pr_debug("failed to write buildid table\n"); goto out_free; @@ -515,33 +433,41 @@ int perf_header__write_pipe(int fd) return 0; } -int perf_header__write(struct perf_header *self, struct perf_evlist *evlist, - int fd, bool at_exit) +int perf_session__write_header(struct perf_session *session, + struct perf_evlist *evlist, + int fd, bool at_exit) { struct perf_file_header f_header; struct perf_file_attr f_attr; - struct perf_header_attr *attr; - int i, err; + struct perf_header *header = &session->header; + struct perf_evsel *attr, *pair = NULL; + int err; lseek(fd, sizeof(f_header), SEEK_SET); - for (i = 0; i < self->attrs; i++) { - attr = self->attr[i]; + if (session->evlist != evlist) + pair = list_entry(session->evlist->entries.next, struct perf_evsel, node); + list_for_each_entry(attr, &evlist->entries, node) { attr->id_offset = lseek(fd, 0, SEEK_CUR); err = do_write(fd, attr->id, attr->ids * sizeof(u64)); if (err < 0) { +out_err_write: pr_debug("failed to write perf header\n"); return err; } + if (session->evlist != evlist) { + err = do_write(fd, pair->id, pair->ids * sizeof(u64)); + if (err < 0) + goto out_err_write; + attr->ids += pair->ids; + pair = list_entry(pair->node.next, struct perf_evsel, node); + } } + header->attr_offset = lseek(fd, 0, SEEK_CUR); - self->attr_offset = lseek(fd, 0, SEEK_CUR); - - for (i = 0; i < self->attrs; i++) { - attr = self->attr[i]; - + list_for_each_entry(attr, &evlist->entries, node) { f_attr = (struct perf_file_attr){ .attr = attr->attr, .ids = { @@ -556,20 +482,20 @@ int perf_header__write(struct perf_header *self, struct perf_evlist *evlist, } } - self->event_offset = lseek(fd, 0, SEEK_CUR); - self->event_size = event_count * sizeof(struct perf_trace_event_type); + header->event_offset = lseek(fd, 0, SEEK_CUR); + header->event_size = event_count * sizeof(struct perf_trace_event_type); if (events) { - err = do_write(fd, events, self->event_size); + err = do_write(fd, events, header->event_size); if (err < 0) { pr_debug("failed to write perf header events\n"); return err; } } - self->data_offset = lseek(fd, 0, SEEK_CUR); + header->data_offset = lseek(fd, 0, SEEK_CUR); if (at_exit) { - err = perf_header__adds_write(self, evlist, fd); + err = perf_header__adds_write(header, evlist, fd); if (err < 0) return err; } @@ -579,20 +505,20 @@ int perf_header__write(struct perf_header *self, struct perf_evlist *evlist, .size = sizeof(f_header), .attr_size = sizeof(f_attr), .attrs = { - .offset = self->attr_offset, - .size = self->attrs * sizeof(f_attr), + .offset = header->attr_offset, + .size = evlist->nr_entries * sizeof(f_attr), }, .data = { - .offset = self->data_offset, - .size = self->data_size, + .offset = header->data_offset, + .size = header->data_size, }, .event_types = { - .offset = self->event_offset, - .size = self->event_size, + .offset = header->event_offset, + .size = header->event_size, }, }; - memcpy(&f_header.adds_features, &self->adds_features, sizeof(self->adds_features)); + memcpy(&f_header.adds_features, &header->adds_features, sizeof(header->adds_features)); lseek(fd, 0, SEEK_SET); err = do_write(fd, &f_header, sizeof(f_header)); @@ -600,26 +526,26 @@ int perf_header__write(struct perf_header *self, struct perf_evlist *evlist, pr_debug("failed to write perf header\n"); return err; } - lseek(fd, self->data_offset + self->data_size, SEEK_SET); + lseek(fd, header->data_offset + header->data_size, SEEK_SET); - self->frozen = 1; + header->frozen = 1; return 0; } -static int perf_header__getbuffer64(struct perf_header *self, +static int perf_header__getbuffer64(struct perf_header *header, int fd, void *buf, size_t size) { if (readn(fd, buf, size) <= 0) return -1; - if (self->needs_swap) + if (header->needs_swap) mem_bswap_64(buf, size); return 0; } -int perf_header__process_sections(struct perf_header *self, int fd, - int (*process)(struct perf_file_section *self, +int perf_header__process_sections(struct perf_header *header, int fd, + int (*process)(struct perf_file_section *section, struct perf_header *ph, int feat, int fd)) { @@ -629,7 +555,7 @@ int perf_header__process_sections(struct perf_header *self, int fd, int idx = 0; int err = -1, feat = 1; - nr_sections = bitmap_weight(self->adds_features, HEADER_FEAT_BITS); + nr_sections = bitmap_weight(header->adds_features, HEADER_FEAT_BITS); if (!nr_sections) return 0; @@ -639,17 +565,17 @@ int perf_header__process_sections(struct perf_header *self, int fd, sec_size = sizeof(*feat_sec) * nr_sections; - lseek(fd, self->data_offset + self->data_size, SEEK_SET); + lseek(fd, header->data_offset + header->data_size, SEEK_SET); - if (perf_header__getbuffer64(self, fd, feat_sec, sec_size)) + if (perf_header__getbuffer64(header, fd, feat_sec, sec_size)) goto out_free; err = 0; while (idx < nr_sections && feat < HEADER_LAST_FEATURE) { - if (perf_header__has_feat(self, feat)) { + if (perf_header__has_feat(header, feat)) { struct perf_file_section *sec = &feat_sec[idx++]; - err = process(sec, self, feat, fd); + err = process(sec, header, feat, fd); if (err < 0) break; } @@ -660,35 +586,35 @@ out_free: return err; } -int perf_file_header__read(struct perf_file_header *self, +int perf_file_header__read(struct perf_file_header *header, struct perf_header *ph, int fd) { lseek(fd, 0, SEEK_SET); - if (readn(fd, self, sizeof(*self)) <= 0 || - memcmp(&self->magic, __perf_magic, sizeof(self->magic))) + if (readn(fd, header, sizeof(*header)) <= 0 || + memcmp(&header->magic, __perf_magic, sizeof(header->magic))) return -1; - if (self->attr_size != sizeof(struct perf_file_attr)) { - u64 attr_size = bswap_64(self->attr_size); + if (header->attr_size != sizeof(struct perf_file_attr)) { + u64 attr_size = bswap_64(header->attr_size); if (attr_size != sizeof(struct perf_file_attr)) return -1; - mem_bswap_64(self, offsetof(struct perf_file_header, + mem_bswap_64(header, offsetof(struct perf_file_header, adds_features)); ph->needs_swap = true; } - if (self->size != sizeof(*self)) { + if (header->size != sizeof(*header)) { /* Support the previous format */ - if (self->size == offsetof(typeof(*self), adds_features)) - bitmap_zero(self->adds_features, HEADER_FEAT_BITS); + if (header->size == offsetof(typeof(*header), adds_features)) + bitmap_zero(header->adds_features, HEADER_FEAT_BITS); else return -1; } - memcpy(&ph->adds_features, &self->adds_features, + memcpy(&ph->adds_features, &header->adds_features, sizeof(ph->adds_features)); /* * FIXME: hack that assumes that if we need swap the perf.data file @@ -702,10 +628,10 @@ int perf_file_header__read(struct perf_file_header *self, perf_header__set_feat(ph, HEADER_BUILD_ID); } - ph->event_offset = self->event_types.offset; - ph->event_size = self->event_types.size; - ph->data_offset = self->data.offset; - ph->data_size = self->data.size; + ph->event_offset = header->event_types.offset; + ph->event_size = header->event_types.size; + ph->data_offset = header->data.offset; + ph->data_size = header->data.size; return 0; } @@ -764,11 +690,10 @@ out: return err; } -static int perf_header__read_build_ids(struct perf_header *self, - int input, u64 offset, u64 size) +static int perf_header__read_build_ids(struct perf_header *header, + int input, u64 offset, u64 size) { - struct perf_session *session = container_of(self, - struct perf_session, header); + struct perf_session *session = container_of(header, struct perf_session, header); struct build_id_event bev; char filename[PATH_MAX]; u64 limit = offset + size; @@ -780,7 +705,7 @@ static int perf_header__read_build_ids(struct perf_header *self, if (read(input, &bev, sizeof(bev)) != sizeof(bev)) goto out; - if (self->needs_swap) + if (header->needs_swap) perf_event_header__bswap(&bev.header); len = bev.header.size - sizeof(bev); @@ -796,13 +721,13 @@ out: return err; } -static int perf_file_section__process(struct perf_file_section *self, +static int perf_file_section__process(struct perf_file_section *section, struct perf_header *ph, int feat, int fd) { - if (lseek(fd, self->offset, SEEK_SET) == (off_t)-1) { + if (lseek(fd, section->offset, SEEK_SET) == (off_t)-1) { pr_debug("Failed to lseek to %" PRIu64 " offset for feature " - "%d, continuing...\n", self->offset, feat); + "%d, continuing...\n", section->offset, feat); return 0; } @@ -812,7 +737,7 @@ static int perf_file_section__process(struct perf_file_section *self, break; case HEADER_BUILD_ID: - if (perf_header__read_build_ids(ph, fd, self->offset, self->size)) + if (perf_header__read_build_ids(ph, fd, section->offset, section->size)) pr_debug("Failed to read buildids, continuing...\n"); break; default: @@ -822,21 +747,21 @@ static int perf_file_section__process(struct perf_file_section *self, return 0; } -static int perf_file_header__read_pipe(struct perf_pipe_file_header *self, +static int perf_file_header__read_pipe(struct perf_pipe_file_header *header, struct perf_header *ph, int fd, bool repipe) { - if (readn(fd, self, sizeof(*self)) <= 0 || - memcmp(&self->magic, __perf_magic, sizeof(self->magic))) + if (readn(fd, header, sizeof(*header)) <= 0 || + memcmp(&header->magic, __perf_magic, sizeof(header->magic))) return -1; - if (repipe && do_write(STDOUT_FILENO, self, sizeof(*self)) < 0) + if (repipe && do_write(STDOUT_FILENO, header, sizeof(*header)) < 0) return -1; - if (self->size != sizeof(*self)) { - u64 size = bswap_64(self->size); + if (header->size != sizeof(*header)) { + u64 size = bswap_64(header->size); - if (size != sizeof(*self)) + if (size != sizeof(*header)) return -1; ph->needs_swap = true; @@ -847,10 +772,10 @@ static int perf_file_header__read_pipe(struct perf_pipe_file_header *self, static int perf_header__read_pipe(struct perf_session *session, int fd) { - struct perf_header *self = &session->header; + struct perf_header *header = &session->header; struct perf_pipe_file_header f_header; - if (perf_file_header__read_pipe(&f_header, self, fd, + if (perf_file_header__read_pipe(&f_header, header, fd, session->repipe) < 0) { pr_debug("incompatible file format\n"); return -EINVAL; @@ -861,18 +786,22 @@ static int perf_header__read_pipe(struct perf_session *session, int fd) return 0; } -int perf_header__read(struct perf_session *session, int fd) +int perf_session__read_header(struct perf_session *session, int fd) { - struct perf_header *self = &session->header; + struct perf_header *header = &session->header; struct perf_file_header f_header; struct perf_file_attr f_attr; u64 f_id; int nr_attrs, nr_ids, i, j; + session->evlist = perf_evlist__new(NULL, NULL); + if (session->evlist == NULL) + return -ENOMEM; + if (session->fd_pipe) return perf_header__read_pipe(session, fd); - if (perf_file_header__read(&f_header, self, fd) < 0) { + if (perf_file_header__read(&f_header, header, fd) < 0) { pr_debug("incompatible file format\n"); return -EINVAL; } @@ -881,33 +810,39 @@ int perf_header__read(struct perf_session *session, int fd) lseek(fd, f_header.attrs.offset, SEEK_SET); for (i = 0; i < nr_attrs; i++) { - struct perf_header_attr *attr; + struct perf_evsel *evsel; off_t tmp; - if (perf_header__getbuffer64(self, fd, &f_attr, sizeof(f_attr))) + if (perf_header__getbuffer64(header, fd, &f_attr, sizeof(f_attr))) goto out_errno; tmp = lseek(fd, 0, SEEK_CUR); + evsel = perf_evsel__new(&f_attr.attr, i); - attr = perf_header_attr__new(&f_attr.attr); - if (attr == NULL) - return -ENOMEM; + if (evsel == NULL) + goto out_delete_evlist; + /* + * Do it before so that if perf_evsel__alloc_id fails, this + * entry gets purged too at perf_evlist__delete(). + */ + perf_evlist__add(session->evlist, evsel); nr_ids = f_attr.ids.size / sizeof(u64); + /* + * We don't have the cpu and thread maps on the header, so + * for allocating the perf_sample_id table we fake 1 cpu and + * hattr->ids threads. + */ + if (perf_evsel__alloc_id(evsel, 1, nr_ids)) + goto out_delete_evlist; + lseek(fd, f_attr.ids.offset, SEEK_SET); for (j = 0; j < nr_ids; j++) { - if (perf_header__getbuffer64(self, fd, &f_id, sizeof(f_id))) + if (perf_header__getbuffer64(header, fd, &f_id, sizeof(f_id))) goto out_errno; - if (perf_header_attr__add_id(attr, f_id) < 0) { - perf_header_attr__delete(attr); - return -ENOMEM; - } - } - if (perf_header__add_attr(self, attr) < 0) { - perf_header_attr__delete(attr); - return -ENOMEM; + perf_evlist__id_add(session->evlist, evsel, 0, j, f_id); } lseek(fd, tmp, SEEK_SET); @@ -918,51 +853,52 @@ int perf_header__read(struct perf_session *session, int fd) events = malloc(f_header.event_types.size); if (events == NULL) return -ENOMEM; - if (perf_header__getbuffer64(self, fd, events, + if (perf_header__getbuffer64(header, fd, events, f_header.event_types.size)) goto out_errno; event_count = f_header.event_types.size / sizeof(struct perf_trace_event_type); } - perf_header__process_sections(self, fd, perf_file_section__process); + perf_header__process_sections(header, fd, perf_file_section__process); - lseek(fd, self->data_offset, SEEK_SET); + lseek(fd, header->data_offset, SEEK_SET); - self->frozen = 1; + header->frozen = 1; return 0; out_errno: return -errno; + +out_delete_evlist: + perf_evlist__delete(session->evlist); + session->evlist = NULL; + return -ENOMEM; } -u64 perf_header__sample_type(struct perf_header *header) +u64 perf_evlist__sample_type(struct perf_evlist *evlist) { + struct perf_evsel *pos; u64 type = 0; - int i; - - for (i = 0; i < header->attrs; i++) { - struct perf_header_attr *attr = header->attr[i]; + list_for_each_entry(pos, &evlist->entries, node) { if (!type) - type = attr->attr.sample_type; - else if (type != attr->attr.sample_type) + type = pos->attr.sample_type; + else if (type != pos->attr.sample_type) die("non matching sample_type"); } return type; } -bool perf_header__sample_id_all(const struct perf_header *header) +bool perf_evlist__sample_id_all(const struct perf_evlist *evlist) { bool value = false, first = true; - int i; - - for (i = 0; i < header->attrs; i++) { - struct perf_header_attr *attr = header->attr[i]; + struct perf_evsel *pos; + list_for_each_entry(pos, &evlist->entries, node) { if (first) { - value = attr->attr.sample_id_all; + value = pos->attr.sample_id_all; first = false; - } else if (value != attr->attr.sample_id_all) + } else if (value != pos->attr.sample_id_all) die("non matching sample_id_all"); } @@ -1000,16 +936,13 @@ int perf_event__synthesize_attr(struct perf_event_attr *attr, u16 ids, u64 *id, return err; } -int perf_event__synthesize_attrs(struct perf_header *self, - perf_event__handler_t process, - struct perf_session *session) +int perf_session__synthesize_attrs(struct perf_session *session, + perf_event__handler_t process) { - struct perf_header_attr *attr; - int i, err = 0; - - for (i = 0; i < self->attrs; i++) { - attr = self->attr[i]; + struct perf_evsel *attr; + int err = 0; + list_for_each_entry(attr, &session->evlist->entries, node) { err = perf_event__synthesize_attr(&attr->attr, attr->ids, attr->id, process, session); if (err) { @@ -1024,27 +957,36 @@ int perf_event__synthesize_attrs(struct perf_header *self, int perf_event__process_attr(union perf_event *event, struct perf_session *session) { - struct perf_header_attr *attr; unsigned int i, ids, n_ids; + struct perf_evsel *evsel; - attr = perf_header_attr__new(&event->attr.attr); - if (attr == NULL) + if (session->evlist == NULL) { + session->evlist = perf_evlist__new(NULL, NULL); + if (session->evlist == NULL) + return -ENOMEM; + } + + evsel = perf_evsel__new(&event->attr.attr, + session->evlist->nr_entries); + if (evsel == NULL) return -ENOMEM; + perf_evlist__add(session->evlist, evsel); + ids = event->header.size; ids -= (void *)&event->attr.id - (void *)event; n_ids = ids / sizeof(u64); + /* + * We don't have the cpu and thread maps on the header, so + * for allocating the perf_sample_id table we fake 1 cpu and + * hattr->ids threads. + */ + if (perf_evsel__alloc_id(evsel, 1, n_ids)) + return -ENOMEM; for (i = 0; i < n_ids; i++) { - if (perf_header_attr__add_id(attr, event->attr.id[i]) < 0) { - perf_header_attr__delete(attr); - return -ENOMEM; - } - } - - if (perf_header__add_attr(&session->header, attr) < 0) { - perf_header_attr__delete(attr); - return -ENOMEM; + perf_evlist__id_add(session->evlist, evsel, 0, i, + event->attr.id[i]); } perf_session__update_sample_type(session); diff --git a/tools/perf/util/header.h b/tools/perf/util/header.h index 2fab13348aab..456661d7f10e 100644 --- a/tools/perf/util/header.h +++ b/tools/perf/util/header.h @@ -9,13 +9,6 @@ #include <linux/bitmap.h> -struct perf_header_attr { - struct perf_event_attr attr; - int ids, size; - u64 *id; - off_t id_offset; -}; - enum { HEADER_TRACE_INFO = 1, HEADER_BUILD_ID, @@ -46,14 +39,12 @@ struct perf_pipe_file_header { struct perf_header; -int perf_file_header__read(struct perf_file_header *self, +int perf_file_header__read(struct perf_file_header *header, struct perf_header *ph, int fd); struct perf_header { int frozen; - int attrs, size; bool needs_swap; - struct perf_header_attr **attr; s64 attr_offset; u64 data_offset; u64 data_size; @@ -62,35 +53,25 @@ struct perf_header { DECLARE_BITMAP(adds_features, HEADER_FEAT_BITS); }; -int perf_header__init(struct perf_header *self); -void perf_header__exit(struct perf_header *self); - struct perf_evlist; -int perf_header__read(struct perf_session *session, int fd); -int perf_header__write(struct perf_header *self, struct perf_evlist *evlist, - int fd, bool at_exit); +int perf_session__read_header(struct perf_session *session, int fd); +int perf_session__write_header(struct perf_session *session, + struct perf_evlist *evlist, + int fd, bool at_exit); int perf_header__write_pipe(int fd); -int perf_header__add_attr(struct perf_header *self, - struct perf_header_attr *attr); - int perf_header__push_event(u64 id, const char *name); char *perf_header__find_event(u64 id); -struct perf_header_attr *perf_header_attr__new(struct perf_event_attr *attr); -void perf_header_attr__delete(struct perf_header_attr *self); +u64 perf_evlist__sample_type(struct perf_evlist *evlist); +bool perf_evlist__sample_id_all(const struct perf_evlist *evlist); +void perf_header__set_feat(struct perf_header *header, int feat); +void perf_header__clear_feat(struct perf_header *header, int feat); +bool perf_header__has_feat(const struct perf_header *header, int feat); -int perf_header_attr__add_id(struct perf_header_attr *self, u64 id); - -u64 perf_header__sample_type(struct perf_header *header); -bool perf_header__sample_id_all(const struct perf_header *header); -void perf_header__set_feat(struct perf_header *self, int feat); -void perf_header__clear_feat(struct perf_header *self, int feat); -bool perf_header__has_feat(const struct perf_header *self, int feat); - -int perf_header__process_sections(struct perf_header *self, int fd, - int (*process)(struct perf_file_section *self, +int perf_header__process_sections(struct perf_header *header, int fd, + int (*process)(struct perf_file_section *section, struct perf_header *ph, int feat, int fd)); @@ -101,9 +82,8 @@ int build_id_cache__remove_s(const char *sbuild_id, const char *debugdir); int perf_event__synthesize_attr(struct perf_event_attr *attr, u16 ids, u64 *id, perf_event__handler_t process, struct perf_session *session); -int perf_event__synthesize_attrs(struct perf_header *self, - perf_event__handler_t process, - struct perf_session *session); +int perf_session__synthesize_attrs(struct perf_session *session, + perf_event__handler_t process); int perf_event__process_attr(union perf_event *event, struct perf_session *session); int perf_event__synthesize_event_type(u64 event_id, char *name, diff --git a/tools/perf/util/session.c b/tools/perf/util/session.c index 0d414199889d..f26639fa0fb3 100644 --- a/tools/perf/util/session.c +++ b/tools/perf/util/session.c @@ -13,46 +13,6 @@ #include "sort.h" #include "util.h" -static int perf_session__read_evlist(struct perf_session *session) -{ - int i, j; - - session->evlist = perf_evlist__new(NULL, NULL); - if (session->evlist == NULL) - return -ENOMEM; - - for (i = 0; i < session->header.attrs; ++i) { - struct perf_header_attr *hattr = session->header.attr[i]; - struct perf_evsel *evsel = perf_evsel__new(&hattr->attr, i); - - if (evsel == NULL) - goto out_delete_evlist; - /* - * Do it before so that if perf_evsel__alloc_id fails, this - * entry gets purged too at perf_evlist__delete(). - */ - perf_evlist__add(session->evlist, evsel); - /* - * We don't have the cpu and thread maps on the header, so - * for allocating the perf_sample_id table we fake 1 cpu and - * hattr->ids threads. - */ - if (perf_evsel__alloc_id(evsel, 1, hattr->ids)) - goto out_delete_evlist; - - for (j = 0; j < hattr->ids; ++j) - perf_evlist__id_hash(session->evlist, evsel, 0, j, - hattr->id[j]); - } - - return 0; - -out_delete_evlist: - perf_evlist__delete(session->evlist); - session->evlist = NULL; - return -ENOMEM; -} - static int perf_session__open(struct perf_session *self, bool force) { struct stat input_stat; @@ -61,7 +21,7 @@ static int perf_session__open(struct perf_session *self, bool force) self->fd_pipe = true; self->fd = STDIN_FILENO; - if (perf_header__read(self, self->fd) < 0) + if (perf_session__read_header(self, self->fd) < 0) pr_err("incompatible file format"); return 0; @@ -93,16 +53,11 @@ static int perf_session__open(struct perf_session *self, bool force) goto out_close; } - if (perf_header__read(self, self->fd) < 0) { + if (perf_session__read_header(self, self->fd) < 0) { pr_err("incompatible file format"); goto out_close; } - if (perf_session__read_evlist(self) < 0) { - pr_err("Not enough memory to read the event selector list\n"); - goto out_close; - } - self->size = input_stat.st_size; return 0; @@ -139,21 +94,10 @@ out: session->id_hdr_size = size; } -void perf_session__set_sample_id_all(struct perf_session *session, bool value) -{ - session->sample_id_all = value; - perf_session__id_header_size(session); -} - -void perf_session__set_sample_type(struct perf_session *session, u64 type) -{ - session->sample_type = type; -} - void perf_session__update_sample_type(struct perf_session *self) { - self->sample_type = perf_header__sample_type(&self->header); - self->sample_id_all = perf_header__sample_id_all(&self->header); + self->sample_type = perf_evlist__sample_type(self->evlist); + self->sample_id_all = perf_evlist__sample_id_all(self->evlist); perf_session__id_header_size(self); } @@ -182,9 +126,6 @@ struct perf_session *perf_session__new(const char *filename, int mode, if (self == NULL) goto out; - if (perf_header__init(&self->header) < 0) - goto out_free; - memcpy(self->filename, filename, len); self->threads = RB_ROOT; INIT_LIST_HEAD(&self->dead_threads); @@ -208,6 +149,7 @@ struct perf_session *perf_session__new(const char *filename, int mode, if (mode == O_RDONLY) { if (perf_session__open(self, force) < 0) goto out_delete; + perf_session__update_sample_type(self); } else if (mode == O_WRONLY) { /* * In O_RDONLY mode this will be performed when reading the @@ -217,8 +159,6 @@ struct perf_session *perf_session__new(const char *filename, int mode, goto out_delete; } - perf_session__update_sample_type(self); - if (ops && ops->ordering_requires_timestamps && ops->ordered_samples && !self->sample_id_all) { dump_printf("WARNING: No sample_id_all support, falling back to unordered processing\n"); @@ -227,9 +167,6 @@ struct perf_session *perf_session__new(const char *filename, int mode, out: return self; -out_free: - free(self); - return NULL; out_delete: perf_session__delete(self); return NULL; @@ -260,7 +197,6 @@ static void perf_session__delete_threads(struct perf_session *self) void perf_session__delete(struct perf_session *self) { - perf_header__exit(&self->header); perf_session__destroy_kernel_maps(self); perf_session__delete_dead_threads(self); perf_session__delete_threads(self); diff --git a/tools/perf/util/session.h b/tools/perf/util/session.h index 05dd7bcb9453..b5b148b0aaca 100644 --- a/tools/perf/util/session.h +++ b/tools/perf/util/session.h @@ -112,8 +112,6 @@ void mem_bswap_64(void *src, int byte_size); int perf_session__create_kernel_maps(struct perf_session *self); void perf_session__update_sample_type(struct perf_session *self); -void perf_session__set_sample_id_all(struct perf_session *session, bool value); -void perf_session__set_sample_type(struct perf_session *session, u64 type); void perf_session__remove_thread(struct perf_session *self, struct thread *th); static inline diff --git a/tools/perf/util/top.c b/tools/perf/util/top.c index 70a9c13f4ad5..75cfe4d45119 100644 --- a/tools/perf/util/top.c +++ b/tools/perf/util/top.c @@ -61,6 +61,12 @@ static void rb_insert_active_sym(struct rb_root *tree, struct sym_entry *se) rb_insert_color(&se->rb_node, tree); } +#define SNPRINTF(buf, size, fmt, args...) \ +({ \ + size_t r = snprintf(buf, size, fmt, ## args); \ + r > size ? size : r; \ +}) + size_t perf_top__header_snprintf(struct perf_top *top, char *bf, size_t size) { struct perf_evsel *counter; @@ -70,7 +76,7 @@ size_t perf_top__header_snprintf(struct perf_top *top, char *bf, size_t size) size_t ret = 0; if (!perf_guest) { - ret = snprintf(bf, size, + ret = SNPRINTF(bf, size, " PerfTop:%8.0f irqs/sec kernel:%4.1f%%" " exact: %4.1f%% [", samples_per_sec, 100.0 - (100.0 * ((samples_per_sec - ksamples_per_sec) / @@ -81,7 +87,7 @@ size_t perf_top__header_snprintf(struct perf_top *top, char *bf, size_t size) float guest_kernel_samples_per_sec = top->guest_kernel_samples / top->delay_secs; float guest_us_samples_per_sec = top->guest_us_samples / top->delay_secs; - ret = snprintf(bf, size, + ret = SNPRINTF(bf, size, " PerfTop:%8.0f irqs/sec kernel:%4.1f%% us:%4.1f%%" " guest kernel:%4.1f%% guest us:%4.1f%%" " exact: %4.1f%% [", samples_per_sec, @@ -101,38 +107,52 @@ size_t perf_top__header_snprintf(struct perf_top *top, char *bf, size_t size) if (top->evlist->nr_entries == 1 || !top->display_weighted) { struct perf_evsel *first; first = list_entry(top->evlist->entries.next, struct perf_evsel, node); - ret += snprintf(bf + ret, size - ret, "%" PRIu64 "%s ", + ret += SNPRINTF(bf + ret, size - ret, "%" PRIu64 "%s ", (uint64_t)first->attr.sample_period, top->freq ? "Hz" : ""); } if (!top->display_weighted) { - ret += snprintf(bf + ret, size - ret, "%s", + ret += SNPRINTF(bf + ret, size - ret, "%s", event_name(top->sym_evsel)); - } else list_for_each_entry(counter, &top->evlist->entries, node) { - ret += snprintf(bf + ret, size - ret, "%s%s", - counter->idx ? "/" : "", event_name(counter)); + } else { + /* + * Don't let events eat all the space. Leaving 30 bytes + * for the rest should be enough. + */ + size_t last_pos = size - 30; + + list_for_each_entry(counter, &top->evlist->entries, node) { + ret += SNPRINTF(bf + ret, size - ret, "%s%s", + counter->idx ? "/" : "", + event_name(counter)); + if (ret > last_pos) { + sprintf(bf + last_pos - 3, ".."); + ret = last_pos - 1; + break; + } + } } - ret += snprintf(bf + ret, size - ret, "], "); + ret += SNPRINTF(bf + ret, size - ret, "], "); if (top->target_pid != -1) - ret += snprintf(bf + ret, size - ret, " (target_pid: %d", + ret += SNPRINTF(bf + ret, size - ret, " (target_pid: %d", top->target_pid); else if (top->target_tid != -1) - ret += snprintf(bf + ret, size - ret, " (target_tid: %d", + ret += SNPRINTF(bf + ret, size - ret, " (target_tid: %d", top->target_tid); else - ret += snprintf(bf + ret, size - ret, " (all"); + ret += SNPRINTF(bf + ret, size - ret, " (all"); if (top->cpu_list) - ret += snprintf(bf + ret, size - ret, ", CPU%s: %s)", + ret += SNPRINTF(bf + ret, size - ret, ", CPU%s: %s)", top->evlist->cpus->nr > 1 ? "s" : "", top->cpu_list); else { if (top->target_tid != -1) - ret += snprintf(bf + ret, size - ret, ")"); + ret += SNPRINTF(bf + ret, size - ret, ")"); else - ret += snprintf(bf + ret, size - ret, ", %d CPU%s)", + ret += SNPRINTF(bf + ret, size - ret, ", %d CPU%s)", top->evlist->cpus->nr, top->evlist->cpus->nr > 1 ? "s" : ""); } |