perf lock: Move common lock contention code to new file

Avoid references from util code to builtin-lock that require python
stubs. Move the functions and related variables to
util/lock-contention.c. Add max_stack_depth parameter to
match_callstack_filter to avoid sharing a global variable.

Signed-off-by: Ian Rogers <[e-mail]>
Tested-by: Arnaldo Carvalho de Melo <[e-mail]>
Acked-by: Arnaldo Carvalho de Melo <[e-mail]>
Acked-by: Namhyung Kim <[e-mail]>
Cc: Adrian Hunter <[e-mail]>
Cc: Alexander Shishkin <[e-mail]>
Cc: Andi Kleen <[e-mail]>
Cc: Athira Rajeev <[e-mail]>
Cc: Colin Ian King <[e-mail]>
Cc: Dapeng Mi <[e-mail]>
Cc: Howard Chu <[e-mail]>
Cc: Ilya Leoshkevich <[e-mail]>
Cc: Ingo Molnar <[e-mail]>
Cc: James Clark <[e-mail]>
Cc: Jiri Olsa <[e-mail]>
Cc: Josh Poimboeuf <[e-mail]>
Cc: Kan Liang <[e-mail]>
Cc: Mark Rutland <[e-mail]>
Cc: Michael Petlan <[e-mail]>
Cc: Peter Zijlstra <[e-mail]>
Cc: Thomas Richter <[e-mail]>
Cc: Veronika Molnarova <[e-mail]>
Cc: Weilin Wang <[e-mail]>
Link: https://lore.kernel.org/r/[e-mail]
Signed-off-by: Arnaldo Carvalho de Melo <[e-mail]>

tools/perf/builtin-lock.c

···
 static struct perf_session *session;
 static struct target target;
 
-/* based on kernel/lockdep.c */
-#define LOCKHASH_BITS		12
-#define LOCKHASH_SIZE		(1UL << LOCKHASH_BITS)
-
-static struct hlist_head *lockhash_table;
-
-#define __lockhashfn(key)	hash_long((unsigned long)key, LOCKHASH_BITS)
-#define lockhashentry(key)	(lockhash_table + __lockhashfn((key)))
-
 static struct rb_root		thread_stats;
 
 static bool combine_locks;
···
 static int max_stack_depth = CONTENTION_STACK_DEPTH;
 static int stack_skip = CONTENTION_STACK_SKIP;
 static int print_nr_entries = INT_MAX / 2;
-static LIST_HEAD(callstack_filters);
 static const char *output_name = NULL;
 static FILE *lock_output;
-
-struct callstack_filter {
-	struct list_head list;
-	char name[];
-};
 
 static struct lock_filter filters;
 
 static enum lock_aggr_mode aggr_mode = LOCK_AGGR_ADDR;
-
-static bool needs_callstack(void)
-{
-	return !list_empty(&callstack_filters);
-}
 
 static struct thread_stat *thread_stat_find(u32 tid)
 {
···
 
 	rb_erase(node, &result);
 	return container_of(node, struct lock_stat, rb);
-}
-
-struct lock_stat *lock_stat_find(u64 addr)
-{
-	struct hlist_head *entry = lockhashentry(addr);
-	struct lock_stat *ret;
-
-	hlist_for_each_entry(ret, entry, hash_entry) {
-		if (ret->addr == addr)
-			return ret;
-	}
-	return NULL;
-}
-
-struct lock_stat *lock_stat_findnew(u64 addr, const char *name, int flags)
-{
-	struct hlist_head *entry = lockhashentry(addr);
-	struct lock_stat *ret, *new;
-
-	hlist_for_each_entry(ret, entry, hash_entry) {
-		if (ret->addr == addr)
-			return ret;
-	}
-
-	new = zalloc(sizeof(struct lock_stat));
-	if (!new)
-		goto alloc_failed;
-
-	new->addr = addr;
-	new->name = strdup(name);
-	if (!new->name) {
-		free(new);
-		goto alloc_failed;
-	}
-
-	new->flags = flags;
-	new->wait_time_min = ULLONG_MAX;
-
-	hlist_add_head(&new->hash_entry, entry);
-	return new;
-
-alloc_failed:
-	pr_err("memory allocation failed\n");
-	return NULL;
-}
-
-bool match_callstack_filter(struct machine *machine, u64 *callstack)
-{
-	struct map *kmap;
-	struct symbol *sym;
-	u64 ip;
-	const char *arch = perf_env__arch(machine->env);
-
-	if (list_empty(&callstack_filters))
-		return true;
-
-	for (int i = 0; i < max_stack_depth; i++) {
-		struct callstack_filter *filter;
-
-		/*
-		 * In powerpc, the callchain saved by kernel always includes
-		 * first three entries as the NIP (next instruction pointer),
-		 * LR (link register), and the contents of LR save area in the
-		 * second stack frame. In certain scenarios its possible to have
-		 * invalid kernel instruction addresses in either LR or the second
-		 * stack frame's LR. In that case, kernel will store that address as
-		 * zero.
-		 *
-		 * The below check will continue to look into callstack,
-		 * incase first or second callstack index entry has 0
-		 * address for powerpc.
-		 */
-		if (!callstack || (!callstack[i] && (strcmp(arch, "powerpc") ||
-						(i != 1 && i != 2))))
-			break;
-
-		ip = callstack[i];
-		sym = machine__find_kernel_symbol(machine, ip, &kmap);
-		if (sym == NULL)
-			continue;
-
-		list_for_each_entry(filter, &callstack_filters, list) {
-			if (strstr(sym->name, filter->name))
-				return true;
-		}
-	}
-	return false;
 }
 
 struct trace_lock_handler {
···
 		if (callstack == NULL)
 			return -ENOMEM;
 
-		if (!match_callstack_filter(machine, callstack)) {
+		if (!match_callstack_filter(machine, callstack, max_stack_depth)) {
 			free(callstack);
 			return 0;
 		}
···
 			ret = -1;
 			break;
 		}
-	}
-
-	free(s);
-	return ret;
-}
-
-static int parse_call_stack(const struct option *opt __maybe_unused, const char *str,
-			   int unset __maybe_unused)
-{
-	char *s, *tmp, *tok;
-	int ret = 0;
-
-	s = strdup(str);
-	if (s == NULL)
-		return -1;
-
-	for (tok = strtok_r(s, ", ", &tmp); tok; tok = strtok_r(NULL, ", ", &tmp)) {
-		struct callstack_filter *entry;
-
-		entry = malloc(sizeof(*entry) + strlen(tok) + 1);
-		if (entry == NULL) {
-			pr_err("Memory allocation failure\n");
-			free(s);
-			return -1;
-		}
-
-		strcpy(entry->name, tok);
-		list_add_tail(&entry->list, &callstack_filters);
 	}
 
 	free(s);

tools/perf/util/Build

···
 perf-util-y += iostat.o
 perf-util-y += stream.o
 perf-util-y += kvm-stat.o
+perf-util-y += lock-contention.o
 perf-util-$(CONFIG_AUXTRACE) += auxtrace.o
 perf-util-y += intel-pt-decoder/
 perf-util-$(CONFIG_AUXTRACE) += intel-pt.o

tools/perf/util/bpf_lock_contention.c

···
 		if (con->save_callstack) {
 			bpf_map_lookup_elem(stack, &key.stack_id, stack_trace);
 
-			if (!match_callstack_filter(machine, stack_trace)) {
+			if (!match_callstack_filter(machine, stack_trace, con->max_stack)) {
 				con->nr_filtered += data.count;
 				goto next;
 			}

tools/perf/util/lock-contention.c

···
+// SPDX-License-Identifier: GPL-2.0
+#include "debug.h"
+#include "env.h"
+#include "lock-contention.h"
+#include "machine.h"
+#include "symbol.h"
+
+#include <limits.h>
+#include <string.h>
+
+#include <linux/hash.h>
+#include <linux/zalloc.h>
+
+#define __lockhashfn(key)	hash_long((unsigned long)key, LOCKHASH_BITS)
+#define lockhashentry(key)	(lockhash_table + __lockhashfn((key)))
+
+struct callstack_filter {
+	struct list_head list;
+	char name[];
+};
+
+static LIST_HEAD(callstack_filters);
+struct hlist_head *lockhash_table;
+
+int parse_call_stack(const struct option *opt __maybe_unused, const char *str,
+		     int unset __maybe_unused)
+{
+	char *s, *tmp, *tok;
+	int ret = 0;
+
+	s = strdup(str);
+	if (s == NULL)
+		return -1;
+
+	for (tok = strtok_r(s, ", ", &tmp); tok; tok = strtok_r(NULL, ", ", &tmp)) {
+		struct callstack_filter *entry;
+
+		entry = malloc(sizeof(*entry) + strlen(tok) + 1);
+		if (entry == NULL) {
+			pr_err("Memory allocation failure\n");
+			free(s);
+			return -1;
+		}
+
+		strcpy(entry->name, tok);
+		list_add_tail(&entry->list, &callstack_filters);
+	}
+
+	free(s);
+	return ret;
+}
+
+bool needs_callstack(void)
+{
+	return !list_empty(&callstack_filters);
+}
+
+struct lock_stat *lock_stat_find(u64 addr)
+{
+	struct hlist_head *entry = lockhashentry(addr);
+	struct lock_stat *ret;
+
+	hlist_for_each_entry(ret, entry, hash_entry) {
+		if (ret->addr == addr)
+			return ret;
+	}
+	return NULL;
+}
+
+struct lock_stat *lock_stat_findnew(u64 addr, const char *name, int flags)
+{
+	struct hlist_head *entry = lockhashentry(addr);
+	struct lock_stat *ret, *new;
+
+	hlist_for_each_entry(ret, entry, hash_entry) {
+		if (ret->addr == addr)
+			return ret;
+	}
+
+	new = zalloc(sizeof(struct lock_stat));
+	if (!new)
+		goto alloc_failed;
+
+	new->addr = addr;
+	new->name = strdup(name);
+	if (!new->name) {
+		free(new);
+		goto alloc_failed;
+	}
+
+	new->flags = flags;
+	new->wait_time_min = ULLONG_MAX;
+
+	hlist_add_head(&new->hash_entry, entry);
+	return new;
+
+alloc_failed:
+	pr_err("memory allocation failed\n");
+	return NULL;
+}
+
+bool match_callstack_filter(struct machine *machine, u64 *callstack, int max_stack_depth)
+{
+	struct map *kmap;
+	struct symbol *sym;
+	u64 ip;
+	const char *arch = perf_env__arch(machine->env);
+
+	if (list_empty(&callstack_filters))
+		return true;
+
+	for (int i = 0; i < max_stack_depth; i++) {
+		struct callstack_filter *filter;
+
+		/*
+		 * In powerpc, the callchain saved by kernel always includes
+		 * first three entries as the NIP (next instruction pointer),
+		 * LR (link register), and the contents of LR save area in the
+		 * second stack frame. In certain scenarios its possible to have
+		 * invalid kernel instruction addresses in either LR or the second
+		 * stack frame's LR. In that case, kernel will store that address as
+		 * zero.
+		 *
+		 * The below check will continue to look into callstack,
+		 * incase first or second callstack index entry has 0
+		 * address for powerpc.
+		 */
+		if (!callstack || (!callstack[i] && (strcmp(arch, "powerpc") ||
+						(i != 1 && i != 2))))
+			break;
+
+		ip = callstack[i];
+		sym = machine__find_kernel_symbol(machine, ip, &kmap);
+		if (sym == NULL)
+			continue;
+
+		list_for_each_entry(filter, &callstack_filters, list) {
+			if (strstr(sym->name, filter->name))
+				return true;
+		}
+	}
+	return false;
+}

tools/perf/util/lock-contention.h

···
  */
 #define MAX_LOCK_DEPTH 48
 
-struct lock_stat *lock_stat_find(u64 addr);
-struct lock_stat *lock_stat_findnew(u64 addr, const char *name, int flags);
+/* based on kernel/lockdep.c */
+#define LOCKHASH_BITS		12
+#define LOCKHASH_SIZE		(1UL << LOCKHASH_BITS)
 
-bool match_callstack_filter(struct machine *machine, u64 *callstack);
+extern struct hlist_head *lockhash_table;
 
 /*
  * struct lock_seq_stat:
···
 	bool save_callstack;
 };
 
-#ifdef HAVE_BPF_SKEL
+struct option;
+int parse_call_stack(const struct option *opt, const char *str, int unset);
+bool needs_callstack(void);
 
+struct lock_stat *lock_stat_find(u64 addr);
+struct lock_stat *lock_stat_findnew(u64 addr, const char *name, int flags);
+
+bool match_callstack_filter(struct machine *machine, u64 *callstack, int max_stack_depth);
+
+
+#ifdef HAVE_BPF_SKEL
 int lock_contention_prepare(struct lock_contention *con);
 int lock_contention_start(void);
 int lock_contention_stop(void);

tools/perf/util/python.c

···
 #include "mmap.h"
 #include "util/kwork.h"
 #include "util/sample.h"
-#include "util/lock-contention.h"
 #include <internal/lib.h>
 #include "../builtin.h"
 
···
 struct kwork_work *perf_kwork_add_work(struct perf_kwork *kwork __maybe_unused,
 				       struct kwork_class *class __maybe_unused,
 				       struct kwork_work *key  __maybe_unused)
-{
-	return NULL;
-}
-
-bool match_callstack_filter(struct machine *machine __maybe_unused, u64 *callstack __maybe_unused)
-{
-	return false;
-}
-
-struct lock_stat *lock_stat_find(u64 addr __maybe_unused)
-{
-	return NULL;
-}
-
-struct lock_stat *lock_stat_findnew(u64 addr __maybe_unused, const char *name __maybe_unused,
-				int flags __maybe_unused)
 {
 	return NULL;
 }