From: Ian Rogers <irogers@google.com>
To: Namhyung Kim <namhyung@kernel.org>,
Arnaldo Carvalho de Melo <acme@kernel.org>,
Alireza Haghdoost <haghdoost@uber.com>,
Ian Rogers <irogers@google.com>
Cc: Peter Zijlstra <peterz@infradead.org>,
Ingo Molnar <mingo@redhat.com>, Jiri Olsa <jolsa@kernel.org>,
Adrian Hunter <adrian.hunter@intel.com>,
James Clark <james.clark@linaro.org>,
linux-perf-users@vger.kernel.org, linux-kernel@vger.kernel.org
Subject: [PATCH v7 03/10] perf symbol: Convert map__for_each_symbol() to callback-based iteration
Date: Fri, 20 Nov 2026 23:16:44 -0800 [thread overview]
Message-ID: <20261121071651.67118-4-irogers@google.com> (raw)
In-Reply-To: <20261121071651.67118-1-irogers@google.com>
Currently symbols__for_each_entry(), dso__for_each_symbol(), and
map__for_each_symbol() are macros that expose struct rb_node * cursors to
callers and perform unlocked rb_first_cached() / rb_next() walks.
To prepare for replacing the unprotected symbol rbtree with a locked
symbol array, replace symbols__for_each_entry() with a callback-based
symbols__visit() helper and convert dso__for_each_symbol() and
map__for_each_symbol() to callback-based iteration matching
dsos__for_each_dso() and maps__for_each_map(), updating the existing
map__for_each_symbol() callers.
Signed-off-by: Ian Rogers <irogers@google.com>
---
tools/perf/arch/powerpc/util/sym-handling.c | 34 ++--
tools/perf/builtin-kmem.c | 52 +++---
tools/perf/tests/vmlinux-kallsyms.c | 181 ++++++++++----------
tools/perf/util/dso.h | 8 +-
tools/perf/util/map.h | 8 +-
tools/perf/util/probe-event.c | 82 +++++----
tools/perf/util/symbol.c | 19 ++
tools/perf/util/symbol.h | 16 +-
8 files changed, 228 insertions(+), 172 deletions(-)
diff --git a/tools/perf/arch/powerpc/util/sym-handling.c b/tools/perf/arch/powerpc/util/sym-handling.c
index 947bfad7aa59..f87d77262c00 100644
--- a/tools/perf/arch/powerpc/util/sym-handling.c
+++ b/tools/perf/arch/powerpc/util/sym-handling.c
@@ -115,13 +115,27 @@ void arch__fix_tev_from_maps(struct perf_probe_event *pev,
}
#ifdef HAVE_LIBELF_SUPPORT
+struct post_process_probe_trace_events_args {
+ struct perf_probe_event *pev;
+ struct probe_trace_event *tev;
+ struct map *map;
+};
+
+static int post_process_probe_trace_events_cb(struct symbol *sym, void *data)
+{
+ struct post_process_probe_trace_events_args *args = data;
+
+ if (map__unmap_ip(args->map, sym->start) == args->tev->point.address) {
+ arch__fix_tev_from_maps(args->pev, args->tev, args->map, sym);
+ return 1;
+ }
+ return 0;
+}
+
void arch__post_process_probe_trace_events(struct perf_probe_event *pev,
int ntevs)
{
- struct probe_trace_event *tev;
struct map *map;
- struct symbol *sym = NULL;
- struct rb_node *tmp;
int i = 0;
map = get_target_map(pev->target, pev->nsi, pev->uprobes);
@@ -129,13 +143,13 @@ void arch__post_process_probe_trace_events(struct perf_probe_event *pev,
return;
for (i = 0; i < ntevs; i++) {
- tev = &pev->tevs[i];
- map__for_each_symbol(map, sym, tmp) {
- if (map__unmap_ip(map, sym->start) == tev->point.address) {
- arch__fix_tev_from_maps(pev, tev, map, sym);
- break;
- }
- }
+ struct post_process_probe_trace_events_args args = {
+ .pev = pev,
+ .tev = &pev->tevs[i],
+ .map = map,
+ };
+
+ map__for_each_symbol(map, post_process_probe_trace_events_cb, &args);
}
}
#endif /* HAVE_LIBELF_SUPPORT */
diff --git a/tools/perf/builtin-kmem.c b/tools/perf/builtin-kmem.c
index 1404b523a785..87869c717266 100644
--- a/tools/perf/builtin-kmem.c
+++ b/tools/perf/builtin-kmem.c
@@ -341,13 +341,33 @@ static int callcmp(const void *a, const void *b)
return -1;
}
+static int build_alloc_func_list_cb(struct symbol *sym, void *data)
+{
+ regex_t *alloc_func_regex = data;
+ struct alloc_func *func;
+
+ if (regexec(alloc_func_regex, sym->name, 0, NULL, 0))
+ return 0;
+
+ func = realloc(alloc_func_list,
+ (nr_alloc_funcs + 1) * sizeof(*func));
+ if (func == NULL)
+ return -ENOMEM;
+
+ pr_debug("alloc func: %s\n", sym->name);
+ func[nr_alloc_funcs].start = sym->start;
+ func[nr_alloc_funcs].end = sym->end;
+ func[nr_alloc_funcs].name = sym->name;
+
+ alloc_func_list = func;
+ nr_alloc_funcs++;
+ return 0;
+}
+
static int build_alloc_func_list(void)
{
int ret;
struct map *kernel_map;
- struct symbol *sym;
- struct rb_node *node;
- struct alloc_func *func;
struct machine *machine = &kmem_session->machines.host;
regex_t alloc_func_regex;
static const char pattern[] = "^_?_?(alloc|get_free|get_zeroed)_pages?";
@@ -368,29 +388,13 @@ static int build_alloc_func_list(void)
return -ENOENT;
}
- map__for_each_symbol(kernel_map, sym, node) {
- if (regexec(&alloc_func_regex, sym->name, 0, NULL, 0))
- continue;
-
- func = realloc(alloc_func_list,
- (nr_alloc_funcs + 1) * sizeof(*func));
- if (func == NULL) {
- regfree(&alloc_func_regex);
- return -ENOMEM;
- }
-
- pr_debug("alloc func: %s\n", sym->name);
- func[nr_alloc_funcs].start = sym->start;
- func[nr_alloc_funcs].end = sym->end;
- func[nr_alloc_funcs].name = sym->name;
-
- alloc_func_list = func;
- nr_alloc_funcs++;
- }
+ ret = map__for_each_symbol(kernel_map, build_alloc_func_list_cb, &alloc_func_regex);
+ regfree(&alloc_func_regex);
+ if (ret)
+ return ret;
- qsort(alloc_func_list, nr_alloc_funcs, sizeof(*func), funcmp);
+ qsort(alloc_func_list, nr_alloc_funcs, sizeof(*alloc_func_list), funcmp);
- regfree(&alloc_func_regex);
return 0;
}
diff --git a/tools/perf/tests/vmlinux-kallsyms.c b/tools/perf/tests/vmlinux-kallsyms.c
index 9396c8a77c86..38f5c9584e15 100644
--- a/tools/perf/tests/vmlinux-kallsyms.c
+++ b/tools/perf/tests/vmlinux-kallsyms.c
@@ -113,6 +113,7 @@ static bool is_ignored_symbol(const char *name, char type)
struct test__vmlinux_matches_kallsyms_cb_args {
struct machine kallsyms;
struct map *vmlinux_map;
+ int err;
bool header_printed;
};
@@ -185,16 +186,101 @@ static int test__vmlinux_matches_kallsyms_cb3(struct map *map, void *data)
return 0;
}
+static int test__vmlinux_matches_kallsyms_sym_cb(struct symbol *sym, void *data)
+{
+ struct test__vmlinux_matches_kallsyms_cb_args *args = data;
+ /*
+ * Step 4:
+ *
+ * kallsyms will be internally on demand sorted by name so that we can
+ * find the reference relocation * symbol, i.e. the symbol we will use
+ * to see if the running kernel was relocated by checking if it has the
+ * same value in the vmlinux file we load.
+ */
+ struct map *kallsyms_map = machine__kernel_map(&args->kallsyms);
+ struct symbol *pair, *first_pair;
+ u64 mem_start, mem_end;
+
+ if (sym->start == sym->end)
+ return 0;
+
+ mem_start = map__unmap_ip(args->vmlinux_map, sym->start);
+ mem_end = map__unmap_ip(args->vmlinux_map, sym->end);
+
+ first_pair = machine__find_kernel_symbol(&args->kallsyms, mem_start, NULL);
+ pair = first_pair;
+
+ if (pair && UM(pair->start) == mem_start) {
+next_pair:
+ if (arch__compare_symbol_names(sym->name, pair->name) == 0) {
+ /*
+ * kallsyms don't have the symbol end, so we
+ * set that by using the next symbol start - 1,
+ * in some cases we get this up to a page
+ * wrong, trace_kmalloc when I was developing
+ * this code was one such example, 2106 bytes
+ * off the real size. More than that and we
+ * _really_ have a problem.
+ */
+ s64 skew = mem_end - UM(pair->end);
+
+ if (llabs(skew) >= page_size) {
+ pr_debug("WARN: %#" PRIx64 ": diff end addr for %s v: %#"
+ PRIx64 " k: %#" PRIx64 "\n",
+ mem_start, sym->name, mem_end,
+ UM(pair->end));
+ }
+
+ /*
+ * Do not count this as a failure, because we
+ * could really find a case where it's not
+ * possible to get proper function end from
+ * kallsyms.
+ */
+ return 0;
+ }
+ pair = machine__find_kernel_symbol_by_name(&args->kallsyms,
+ sym->name, NULL);
+ if (pair) {
+ if (UM(pair->start) == mem_start)
+ goto next_pair;
+
+ pr_debug("WARN: %#" PRIx64 ": diff name v: %s k: %s\n",
+ mem_start, sym->name, pair->name);
+ } else {
+ pr_debug("WARN: %#" PRIx64 ": diff name v: %s k: %s\n",
+ mem_start, sym->name, first_pair->name);
+ }
+
+ return 0;
+ }
+
+ if (mem_start == map__end(args->kallsyms.vmlinux_map)) {
+ /*
+ * Ignore aliases to _etext, i.e. to the end of the kernel text area,
+ * such as __indirect_thunk_end.
+ */
+ return 0;
+ }
+ if (is_ignored_symbol(sym->name, symbol__type(sym))) {
+ /*
+ * Ignore hidden symbols, see scripts/kallsyms.c for the details
+ */
+ return 0;
+ }
+ pr_debug("ERR : %#" PRIx64 ": %s not on kallsyms\n",
+ mem_start, sym->name);
+
+ args->err = -1;
+ return 0;
+}
+
static int test__vmlinux_matches_kallsyms(struct test_suite *test __maybe_unused,
int subtest __maybe_unused)
{
int err = TEST_FAIL;
- struct rb_node *nd;
- struct symbol *sym;
- struct map *kallsyms_map;
struct machine vmlinux = { 0 };
struct maps *maps;
- u64 mem_start, mem_end;
struct test__vmlinux_matches_kallsyms_cb_args args;
/*
@@ -240,16 +326,6 @@ static int test__vmlinux_matches_kallsyms(struct test_suite *test __maybe_unused
goto out;
}
- /*
- * Step 4:
- *
- * kallsyms will be internally on demand sorted by name so that we can
- * find the reference relocation * symbol, i.e. the symbol we will use
- * to see if the running kernel was relocated by checking if it has the
- * same value in the vmlinux file we load.
- */
- kallsyms_map = machine__kernel_map(&args.kallsyms);
-
/*
* Step 5:
*
@@ -279,7 +355,7 @@ static int test__vmlinux_matches_kallsyms(struct test_suite *test __maybe_unused
goto out;
}
- err = 0;
+ args.err = 0;
/*
* Step 7:
*
@@ -287,79 +363,8 @@ static int test__vmlinux_matches_kallsyms(struct test_suite *test __maybe_unused
* in the kallsyms dso. For the ones that are in both, check its names and
* end addresses too.
*/
- map__for_each_symbol(args.vmlinux_map, sym, nd) {
- struct symbol *pair, *first_pair;
-
- sym = rb_entry(nd, struct symbol, rb_node);
-
- if (sym->start == sym->end)
- continue;
-
- mem_start = map__unmap_ip(args.vmlinux_map, sym->start);
- mem_end = map__unmap_ip(args.vmlinux_map, sym->end);
-
- first_pair = machine__find_kernel_symbol(&args.kallsyms, mem_start, NULL);
- pair = first_pair;
-
- if (pair && UM(pair->start) == mem_start) {
-next_pair:
- if (arch__compare_symbol_names(sym->name, pair->name) == 0) {
- /*
- * kallsyms don't have the symbol end, so we
- * set that by using the next symbol start - 1,
- * in some cases we get this up to a page
- * wrong, trace_kmalloc when I was developing
- * this code was one such example, 2106 bytes
- * off the real size. More than that and we
- * _really_ have a problem.
- */
- s64 skew = mem_end - UM(pair->end);
- if (llabs(skew) >= page_size)
- pr_debug("WARN: %#" PRIx64 ": diff end addr for %s v: %#" PRIx64 " k: %#" PRIx64 "\n",
- mem_start, sym->name, mem_end,
- UM(pair->end));
-
- /*
- * Do not count this as a failure, because we
- * could really find a case where it's not
- * possible to get proper function end from
- * kallsyms.
- */
- continue;
- } else {
- pair = machine__find_kernel_symbol_by_name(&args.kallsyms,
- sym->name, NULL);
- if (pair) {
- if (UM(pair->start) == mem_start)
- goto next_pair;
-
- pr_debug("WARN: %#" PRIx64 ": diff name v: %s k: %s\n",
- mem_start, sym->name, pair->name);
- } else {
- pr_debug("WARN: %#" PRIx64 ": diff name v: %s k: %s\n",
- mem_start, sym->name, first_pair->name);
- }
-
- continue;
- }
- } else if (mem_start == map__end(args.kallsyms.vmlinux_map)) {
- /*
- * Ignore aliases to _etext, i.e. to the end of the kernel text area,
- * such as __indirect_thunk_end.
- */
- continue;
- } else if (is_ignored_symbol(sym->name, symbol__type(sym))) {
- /*
- * Ignore hidden symbols, see scripts/kallsyms.c for the details
- */
- continue;
- } else {
- pr_debug("ERR : %#" PRIx64 ": %s not on kallsyms\n",
- mem_start, sym->name);
- }
-
- err = -1;
- }
+ map__for_each_symbol(args.vmlinux_map, test__vmlinux_matches_kallsyms_sym_cb, &args);
+ err = args.err;
if (verbose <= 0)
goto out;
diff --git a/tools/perf/util/dso.h b/tools/perf/util/dso.h
index d5f7382d4e13..532ad3ec156f 100644
--- a/tools/perf/util/dso.h
+++ b/tools/perf/util/dso.h
@@ -365,11 +365,11 @@ int dso_id__cmp(const struct dso_id *a, const struct dso_id *b);
/* dso__for_each_symbol - iterate over the symbols of given type
*
* @dso: the 'struct dso *' in which symbols are iterated
- * @pos: the 'struct symbol *' to use as a loop cursor
- * @n: the 'struct rb_node *' to use as a temporary storage
+ * @cb: callback function invoked for each symbol
+ * @data: opaque data pointer passed to @cb
*/
-#define dso__for_each_symbol(dso, pos, n) \
- symbols__for_each_entry(dso__symbols(dso), pos, n)
+#define dso__for_each_symbol(dso, cb, data) \
+ symbols__visit(dso__symbols(dso), cb, data)
static inline void *dso__a2l(const struct dso *dso)
{
diff --git a/tools/perf/util/map.h b/tools/perf/util/map.h
index c2e31be232c8..32f1fd9a7404 100644
--- a/tools/perf/util/map.h
+++ b/tools/perf/util/map.h
@@ -146,12 +146,12 @@ struct thread;
/* map__for_each_symbol - iterate over the symbols in the given map
*
* @map: the 'struct map *' in which symbols are iterated
- * @pos: the 'struct symbol *' to use as a loop cursor
- * @n: the 'struct rb_node *' to use as a temporary storage
+ * @cb: callback function invoked for each symbol
+ * @data: opaque data pointer passed to @cb
* Note: caller must ensure map->dso is not NULL (map is loaded).
*/
-#define map__for_each_symbol(map, pos, n) \
- dso__for_each_symbol(map__dso(map), pos, n)
+#define map__for_each_symbol(map, cb, data) \
+ dso__for_each_symbol(map__dso(map), cb, data)
/* map__for_each_symbol_with_name - iterate over the symbols in the given map
* that have the given name
diff --git a/tools/perf/util/probe-event.c b/tools/perf/util/probe-event.c
index 1534280c0c1d..3d506c09b749 100644
--- a/tools/perf/util/probe-event.c
+++ b/tools/perf/util/probe-event.c
@@ -3047,49 +3047,63 @@ static int __add_probe_trace_events(struct perf_probe_event *pev,
return ret;
}
-static int find_probe_functions(struct map *map, char *name,
- struct symbol **syms)
+struct find_probe_functions_args {
+ const char *name;
+ struct symbol **syms;
+ int found;
+ bool cut_version;
+};
+
+static int find_probe_functions_cb(struct symbol *sym, void *data)
{
- int found = 0;
- struct symbol *sym;
- struct rb_node *tmp;
+ struct find_probe_functions_args *args = data;
const char *norm, *ver;
char *buf = NULL;
- bool cut_version = true;
-
- if (map__load(map) < 0)
- return -EACCES; /* Possible permission error to load symbols */
- /* If user gives a version, don't cut off the version from symbols */
- if (strchr(name, '@'))
- cut_version = false;
-
- map__for_each_symbol(map, sym, tmp) {
- norm = arch__normalize_symbol_name(sym->name);
- if (!norm)
- continue;
+ norm = arch__normalize_symbol_name(sym->name);
+ if (!norm)
+ return 0;
- if (cut_version) {
- /* We don't care about default symbol or not */
- ver = strchr(norm, '@');
- if (ver) {
- buf = strndup(norm, ver - norm);
- if (!buf)
- return -ENOMEM;
- norm = buf;
- }
+ if (args->cut_version) {
+ /* We don't care about default symbol or not */
+ ver = strchr(norm, '@');
+ if (ver) {
+ buf = strndup(norm, ver - norm);
+ if (!buf)
+ return -ENOMEM;
+ norm = buf;
}
+ }
- if (strglobmatch(norm, name)) {
- found++;
- if (syms && found < probe_conf.max_probes)
- syms[found - 1] = sym;
- }
- if (buf)
- zfree(&buf);
+ if (strglobmatch(norm, args->name)) {
+ args->found++;
+ if (args->syms && args->found < probe_conf.max_probes)
+ args->syms[args->found - 1] = sym;
}
+ free(buf);
+ return 0;
+}
+
+static int find_probe_functions(struct map *map, char *name,
+ struct symbol **syms)
+{
+ struct find_probe_functions_args args = {
+ .name = name,
+ .syms = syms,
+ /* If user gives a version, don't cut off the version from symbols */
+ .cut_version = strchr(name, '@') == NULL,
+ .found = 0,
+ };
+ int ret;
+
+ if (map__load(map) < 0)
+ return -EACCES; /* Possible permission error to load symbols */
+
+ ret = map__for_each_symbol(map, find_probe_functions_cb, &args);
+ if (ret < 0)
+ return ret;
- return found;
+ return args.found;
}
void __weak arch__fix_tev_from_maps(struct perf_probe_event *pev __maybe_unused,
diff --git a/tools/perf/util/symbol.c b/tools/perf/util/symbol.c
index a1ac1cd1d2e7..1f7faae5a357 100644
--- a/tools/perf/util/symbol.c
+++ b/tools/perf/util/symbol.c
@@ -613,6 +613,25 @@ struct symbol *dso__find_symbol(struct dso *dso, u64 addr)
return dso__last_find_result_symbol(dso);
}
+int symbols__visit(struct rb_root_cached *symbols,
+ int (*cb)(struct symbol *sym, void *data),
+ void *data)
+{
+ struct rb_node *nd;
+
+ if (!symbols)
+ return 0;
+
+ for (nd = rb_first_cached(symbols); nd; nd = rb_next(nd)) {
+ struct symbol *sym = rb_entry(nd, struct symbol, rb_node);
+ int ret = cb(sym, data);
+
+ if (ret)
+ return ret;
+ }
+ return 0;
+}
+
struct symbol *dso__find_symbol_nocache(struct dso *dso, u64 addr)
{
return symbols__find(dso__symbols(dso), addr);
diff --git a/tools/perf/util/symbol.h b/tools/perf/util/symbol.h
index a6a2ab200f98..5a105b10a089 100644
--- a/tools/perf/util/symbol.h
+++ b/tools/perf/util/symbol.h
@@ -141,16 +141,16 @@ void symbol__set_annotate2(struct symbol *sym, bool annotate2);
void symbol__set_inlined(struct symbol *sym, bool inlined);
void symbol__set_ifunc_alias(struct symbol *sym, bool ifunc_alias);
-/* symbols__for_each_entry - iterate over symbols (rb_root)
+/*
+ * symbols__visit - iterate over symbols calling @cb for each entry
*
- * @symbols: the rb_root of symbols
- * @pos: the 'struct symbol *' to use as a loop cursor
- * @nd: the 'struct rb_node *' to use as a temporary storage
+ * @symbols: the rb_root_cached of symbols
+ * @cb: callback function invoked for each symbol
+ * @data: opaque data pointer passed to @cb
*/
-#define symbols__for_each_entry(symbols, pos, nd) \
- for (nd = rb_first_cached(symbols); \
- nd && (pos = rb_entry(nd, struct symbol, rb_node)); \
- nd = rb_next(nd))
+int symbols__visit(struct rb_root_cached *symbols,
+ int (*cb)(struct symbol *sym, void *data),
+ void *data);
static inline size_t symbol__size(const struct symbol *sym)
{
--
2.56.1.750.gf70dc682cf-goog
next prev parent reply other threads:[~2026-11-21 7:17 UTC|newest]
Thread overview: 11+ messages / expand[flat|nested] mbox.gz Atom feed top
2026-11-21 7:16 [PATCH v7 00/10] perf symbol: Sorted arrays, DSO-owned lifetimes, and memory shrinking Ian Rogers
2026-11-21 7:16 ` [PATCH v7 01/10] perf kmem: Fix alloc_func_regex leak on error in build_alloc_func_list() Ian Rogers
2026-11-21 7:16 ` [PATCH v7 02/10] perf maps: Add tail fast-path check in first_ending_after() Ian Rogers
2026-11-21 7:16 ` Ian Rogers [this message]
2026-11-21 7:16 ` [PATCH v7 04/10] perf symbol: Use callback iteration instead of open-coded symbol loops Ian Rogers
2026-11-21 7:16 ` [PATCH v7 05/10] perf symbol: Switch backing storage from rbtree to struct symbols array Ian Rogers
2026-11-21 7:16 ` [PATCH v7 06/10] perf symbol: Replace qsort() with natural-run merge sort Ian Rogers
2026-11-21 7:16 ` [PATCH v7 07/10] perf symbol: Optimize incremental symbol insertion with sorted tail Ian Rogers
2026-11-21 7:16 ` [PATCH v7 08/10] perf symbol: Add memory shrinking for symbols, DSOs, and machines Ian Rogers
2026-11-21 7:16 ` [PATCH v7 09/10] perf session: Add dso__session() and session-scoped allocation wrappers Ian Rogers
2026-11-21 7:16 ` [PATCH v7 10/10] perf test symbols: Add tests for symbol shrinking and session alloc wrappers Ian Rogers
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20261121071651.67118-4-irogers@google.com \
--to=irogers@google.com \
--cc=acme@kernel.org \
--cc=adrian.hunter@intel.com \
--cc=haghdoost@uber.com \
--cc=james.clark@linaro.org \
--cc=jolsa@kernel.org \
--cc=linux-kernel@vger.kernel.org \
--cc=linux-perf-users@vger.kernel.org \
--cc=mingo@redhat.com \
--cc=namhyung@kernel.org \
--cc=peterz@infradead.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.