[PATCH bpf-next 0/4] bpf: Introduce global percpu data

[PATCH bpf-next 0/4] bpf: Introduce global percpu data

[Leon Hwang]

This patch set introduces global percpu data, similar to commit
6316f78306c1 ("Merge branch 'support-global-data'"), to reduce restrictions
in C for BPF programs.

With this enhancement, it becomes possible to define and use global percpu
variables, much like the DEFINE_PER_CPU() macro in the kernel[0].

The idea stems from the bpflbr project[1], which itself was inspired by
retsnoop[2]. During testing of bpflbr on the v6.6 kernel, two LBR
(Last Branch Record) entries were observed related to the
bpf_get_smp_processor_id() helper.

Since commit 1ae6921009e5 ("bpf: inline bpf_get_smp_processor_id() helper"),
the bpf_get_smp_processor_id() helper has been inlined on x86_64, reducing
the overhead and consequently minimizing these two LBR records.

However, the introduction of global percpu data offers a more robust
solution. By leveraging the percpu_array map and percpu instructions,
global percpu data can be implemented intrinsically.

This feature also facilitates sharing percpu information between tail
callers and callees or between freplace callers and callees through a
shared global percpu variable. Previously, this was achieved using a
1-entry percpu_array map, which this patch set aims to improve upon.

Links:
[0] https://github.com/torvalds/linux/blob/fbfd64d25c7af3b8695201ebc85efe90be28c5a3/include/linux/percpu-defs.h#L114
[1] https://github.com/Asphaltt/bpflbr
[2] https://github.com/anakryiko/retsnoop

Changes:
rfc -> v1:
  * Address comments from Andrii:
    * Keep one image of global percpu variable for all CPUs.
    * Reject non-ARRAY map in bpf_map_direct_read(), check_reg_const_str(),
      and check_bpf_snprintf_call() in verifier.
    * Split out libbpf changes from kernel-side changes.
    * Use ".percpu" as PERCPU_DATA_SEC.
    * Use enum libbpf_map_type to distinguish BSS, DATA, RODATA and
      PERCPU_DATA.
    * Avoid using errno for checking err from libbpf_num_possible_cpus().
    * Use "map '%s': " prefix for error message.

Leon Hwang (4):
  bpf: Introduce global percpu data
  bpf, libbpf: Support global percpu data
  bpf, bpftool: Generate skeleton for global percpu data
  selftests/bpf: Add a case to test global percpu data

 kernel/bpf/arraymap.c                         |  39 +++-
 kernel/bpf/verifier.c                         |  45 +++++
 tools/bpf/bpftool/gen.c                       |  13 +-
 tools/lib/bpf/libbpf.c                        | 175 ++++++++++++++----
 tools/lib/bpf/libbpf.h                        |   1 +
 .../bpf/prog_tests/global_data_init.c         |  89 ++++++++-
 .../bpf/progs/test_global_percpu_data.c       |  21 +++
 7 files changed, 340 insertions(+), 43 deletions(-)
 create mode 100644 tools/testing/selftests/bpf/progs/test_global_percpu_data.c

-- 
2.47.1

[PATCH bpf-next 1/4] bpf: Introduce global percpu data

[Leon Hwang]

This patch introduces global percpu data, inspired by commit
6316f78306c1 ("Merge branch 'support-global-data'"). It enables the
definition of global percpu variables in BPF, similar to the
DEFINE_PER_CPU() macro in the kernel[0].

For example, in BPF, it is able to define a global percpu variable like
this:

int percpu_data SEC(".percpu");

With this patch, tools like retsnoop[1] and bpflbr[2] can simplify their
BPF code for handling LBRs. The code can be updated from

static struct perf_branch_entry lbrs[1][MAX_LBR_ENTRIES] SEC(".data.lbrs");

to

static struct perf_branch_entry lbrs[MAX_LBR_ENTRIES] SEC(".percpu.lbrs");

This eliminates the need to retrieve the CPU ID using the
bpf_get_smp_processor_id() helper.

Additionally, by reusing global percpu data map, sharing information
between tail callers and callees or freplace callers and callees becomes
simpler compared to reusing percpu_array maps.

Links:
[0] https://github.com/torvalds/linux/blob/fbfd64d25c7af3b8695201ebc85efe90be28c5a3/include/linux/percpu-defs.h#L114
[1] https://github.com/anakryiko/retsnoop
[2] https://github.com/Asphaltt/bpflbr

Signed-off-by: Leon Hwang <leon.hwang@linux.dev>
---
 kernel/bpf/arraymap.c | 39 ++++++++++++++++++++++++++++++++++++-
 kernel/bpf/verifier.c | 45 +++++++++++++++++++++++++++++++++++++++++++
 2 files changed, 83 insertions(+), 1 deletion(-)

diff --git a/kernel/bpf/arraymap.c b/kernel/bpf/arraymap.c
index eb28c0f219ee4..f8c60d8331975 100644
--- a/kernel/bpf/arraymap.c
+++ b/kernel/bpf/arraymap.c
@@ -249,6 +249,40 @@ static void *percpu_array_map_lookup_elem(struct bpf_map *map, void *key)
 	return this_cpu_ptr(array->pptrs[index & array->index_mask]);
 }
 
+static int percpu_array_map_direct_value_addr(const struct bpf_map *map,
+					      u64 *imm, u32 off)
+{
+	struct bpf_array *array = container_of(map, struct bpf_array, map);
+
+	if (map->max_entries != 1)
+		return -EOPNOTSUPP;
+	if (off >= map->value_size)
+		return -EINVAL;
+	if (!bpf_jit_supports_percpu_insn())
+		return -EOPNOTSUPP;
+
+	*imm = (u64) array->pptrs[0];
+	return 0;
+}
+
+static int percpu_array_map_direct_value_meta(const struct bpf_map *map,
+					      u64 imm, u32 *off)
+{
+	struct bpf_array *array = container_of(map, struct bpf_array, map);
+	u64 base = (u64) array->pptrs[0];
+	u64 range = array->elem_size;
+
+	if (map->max_entries != 1)
+		return -EOPNOTSUPP;
+	if (imm < base || imm >= base + range)
+		return -ENOENT;
+	if (!bpf_jit_supports_percpu_insn())
+		return -EOPNOTSUPP;
+
+	*off = imm - base;
+	return 0;
+}
+
 /* emit BPF instructions equivalent to C code of percpu_array_map_lookup_elem() */
 static int percpu_array_map_gen_lookup(struct bpf_map *map, struct bpf_insn *insn_buf)
 {
@@ -534,7 +568,8 @@ static int array_map_check_btf(const struct bpf_map *map,
 
 	/* One exception for keyless BTF: .bss/.data/.rodata map */
 	if (btf_type_is_void(key_type)) {
-		if (map->map_type != BPF_MAP_TYPE_ARRAY ||
+		if ((map->map_type != BPF_MAP_TYPE_ARRAY &&
+		     map->map_type != BPF_MAP_TYPE_PERCPU_ARRAY) ||
 		    map->max_entries != 1)
 			return -EINVAL;
 
@@ -815,6 +850,8 @@ const struct bpf_map_ops percpu_array_map_ops = {
 	.map_get_next_key = array_map_get_next_key,
 	.map_lookup_elem = percpu_array_map_lookup_elem,
 	.map_gen_lookup = percpu_array_map_gen_lookup,
+	.map_direct_value_addr = percpu_array_map_direct_value_addr,
+	.map_direct_value_meta = percpu_array_map_direct_value_meta,
 	.map_update_elem = array_map_update_elem,
 	.map_delete_elem = array_map_delete_elem,
 	.map_lookup_percpu_elem = percpu_array_map_lookup_percpu_elem,
diff --git a/kernel/bpf/verifier.c b/kernel/bpf/verifier.c
index 9971c03adfd5d..9d99497c2b94c 100644
--- a/kernel/bpf/verifier.c
+++ b/kernel/bpf/verifier.c
@@ -6810,6 +6810,8 @@ static int bpf_map_direct_read(struct bpf_map *map, int off, int size, u64 *val,
 	u64 addr;
 	int err;
 
+	if (map->map_type != BPF_MAP_TYPE_ARRAY)
+		return -EINVAL;
 	err = map->ops->map_direct_value_addr(map, &addr, off);
 	if (err)
 		return err;
@@ -7322,6 +7324,7 @@ static int check_mem_access(struct bpf_verifier_env *env, int insn_idx, u32 regn
 			/* if map is read-only, track its contents as scalars */
 			if (tnum_is_const(reg->var_off) &&
 			    bpf_map_is_rdonly(map) &&
+			    map->map_type != BPF_MAP_TYPE_PERCPU_ARRAY &&
 			    map->ops->map_direct_value_addr) {
 				int map_off = off + reg->var_off.value;
 				u64 val = 0;
@@ -9128,6 +9131,11 @@ static int check_reg_const_str(struct bpf_verifier_env *env,
 		return -EACCES;
 	}
 
+	if (map->map_type != BPF_MAP_TYPE_ARRAY) {
+		verbose(env, "only array map supports direct string value access\n");
+		return -EINVAL;
+	}
+
 	err = check_map_access(env, regno, reg->off,
 			       map->value_size - reg->off, false,
 			       ACCESS_HELPER);
@@ -10802,6 +10810,11 @@ static int check_bpf_snprintf_call(struct bpf_verifier_env *env,
 		return -EINVAL;
 	num_args = data_len_reg->var_off.value / 8;
 
+	if (fmt_map->map_type != BPF_MAP_TYPE_ARRAY) {
+		verbose(env, "only array map supports snprintf\n");
+		return -EINVAL;
+	}
+
 	/* fmt being ARG_PTR_TO_CONST_STR guarantees that var_off is const
 	 * and map_direct_value_addr is set.
 	 */
@@ -21381,6 +21394,38 @@ static int do_misc_fixups(struct bpf_verifier_env *env)
 			goto next_insn;
 		}
 
+#ifdef CONFIG_SMP
+		if (insn->code == (BPF_LD | BPF_IMM | BPF_DW) &&
+		    (insn->src_reg == BPF_PSEUDO_MAP_VALUE ||
+		     insn->src_reg == BPF_PSEUDO_MAP_IDX_VALUE)) {
+			struct bpf_map *map;
+
+			aux = &env->insn_aux_data[i + delta];
+			map = env->used_maps[aux->map_index];
+			if (map->map_type != BPF_MAP_TYPE_PERCPU_ARRAY)
+				goto next_insn;
+
+			/* Reuse the original ld_imm64 insn. And add one
+			 * mov64_percpu_reg insn.
+			 */
+
+			insn_buf[0] = insn[1];
+			insn_buf[1] = BPF_MOV64_PERCPU_REG(insn->dst_reg, insn->dst_reg);
+			cnt = 2;
+
+			i++;
+			new_prog = bpf_patch_insn_data(env, i + delta, insn_buf, cnt);
+			if (!new_prog)
+				return -ENOMEM;
+
+			delta    += cnt - 1;
+			env->prog = prog = new_prog;
+			insn      = new_prog->insnsi + i + delta;
+
+			goto next_insn;
+		}
+#endif
+
 		if (insn->code != (BPF_JMP | BPF_CALL))
 			goto next_insn;
 		if (insn->src_reg == BPF_PSEUDO_CALL)
-- 
2.47.1

[PATCH bpf-next 2/4] bpf, libbpf: Support global percpu data

[Leon Hwang]

This patch introduces support for global percpu data in libbpf. A new
section named ".percpu" is added, similar to the existing ".data" section.
Internal maps are created for ".percpu" sections, which are then
initialized and populated accordingly.

The changes include:

* Introduction of the ".percpu" section in libbpf.
* Creation of internal maps for percpu data.
* Initialization and population of these maps.

This enhancement allows BPF programs to efficiently manage and access
percpu global data, improving performance for use cases that require
percpu buffer.

Signed-off-by: Leon Hwang <leon.hwang@linux.dev>
---
 tools/lib/bpf/libbpf.c | 172 ++++++++++++++++++++++++++++++++---------
 1 file changed, 135 insertions(+), 37 deletions(-)

diff --git a/tools/lib/bpf/libbpf.c b/tools/lib/bpf/libbpf.c
index 194809da51725..6da6004c5c84d 100644
--- a/tools/lib/bpf/libbpf.c
+++ b/tools/lib/bpf/libbpf.c
@@ -516,6 +516,7 @@ struct bpf_struct_ops {
 };
 
 #define DATA_SEC ".data"
+#define PERCPU_DATA_SEC ".percpu"
 #define BSS_SEC ".bss"
 #define RODATA_SEC ".rodata"
 #define KCONFIG_SEC ".kconfig"
@@ -530,6 +531,7 @@ enum libbpf_map_type {
 	LIBBPF_MAP_BSS,
 	LIBBPF_MAP_RODATA,
 	LIBBPF_MAP_KCONFIG,
+	LIBBPF_MAP_PERCPU_DATA,
 };
 
 struct bpf_map_def {
@@ -562,6 +564,7 @@ struct bpf_map {
 	__u32 btf_value_type_id;
 	__u32 btf_vmlinux_value_type_id;
 	enum libbpf_map_type libbpf_type;
+	void *data;
 	void *mmaped;
 	struct bpf_struct_ops *st_ops;
 	struct bpf_map *inner_map;
@@ -640,6 +643,7 @@ enum sec_type {
 	SEC_DATA,
 	SEC_RODATA,
 	SEC_ST_OPS,
+	SEC_PERCPU_DATA,
 };
 
 struct elf_sec_desc {
@@ -1923,13 +1927,24 @@ static bool map_is_mmapable(struct bpf_object *obj, struct bpf_map *map)
 	return false;
 }
 
+static void map_copy_data(struct bpf_map *map, const void *data)
+{
+	bool is_percpu_data = map->libbpf_type == LIBBPF_MAP_PERCPU_DATA;
+	size_t data_sz = map->def.value_size;
+
+	if (data)
+		memcpy(is_percpu_data ? map->data : map->mmaped, data, data_sz);
+}
+
 static int
 bpf_object__init_internal_map(struct bpf_object *obj, enum libbpf_map_type type,
 			      const char *real_name, int sec_idx, void *data, size_t data_sz)
 {
+	bool is_percpu_data = type == LIBBPF_MAP_PERCPU_DATA;
 	struct bpf_map_def *def;
 	struct bpf_map *map;
 	size_t mmap_sz;
+	size_t elem_sz;
 	int err;
 
 	map = bpf_object__add_map(obj);
@@ -1948,7 +1963,8 @@ bpf_object__init_internal_map(struct bpf_object *obj, enum libbpf_map_type type,
 	}
 
 	def = &map->def;
-	def->type = BPF_MAP_TYPE_ARRAY;
+	def->type = is_percpu_data ? BPF_MAP_TYPE_PERCPU_ARRAY
+				   : BPF_MAP_TYPE_ARRAY;
 	def->key_size = sizeof(int);
 	def->value_size = data_sz;
 	def->max_entries = 1;
@@ -1958,29 +1974,53 @@ bpf_object__init_internal_map(struct bpf_object *obj, enum libbpf_map_type type,
 	/* failures are fine because of maps like .rodata.str1.1 */
 	(void) map_fill_btf_type_info(obj, map);
 
-	if (map_is_mmapable(obj, map))
-		def->map_flags |= BPF_F_MMAPABLE;
+	pr_debug("map '%s' (global %sdata): at sec_idx %d, offset %zu, flags %x.\n",
+		 map->name, is_percpu_data ? "percpu " : "", map->sec_idx,
+		 map->sec_offset, def->map_flags);
 
-	pr_debug("map '%s' (global data): at sec_idx %d, offset %zu, flags %x.\n",
-		 map->name, map->sec_idx, map->sec_offset, def->map_flags);
+	if (is_percpu_data) {
+		elem_sz = roundup(data_sz, 8);
 
-	mmap_sz = bpf_map_mmap_sz(map);
-	map->mmaped = mmap(NULL, mmap_sz, PROT_READ | PROT_WRITE,
-			   MAP_SHARED | MAP_ANONYMOUS, -1, 0);
-	if (map->mmaped == MAP_FAILED) {
-		err = -errno;
-		map->mmaped = NULL;
-		pr_warn("failed to alloc map '%s' content buffer: %s\n", map->name, errstr(err));
-		zfree(&map->real_name);
-		zfree(&map->name);
-		return err;
-	}
+		map->data = malloc(elem_sz);
+		if (!map->data) {
+			err = -ENOMEM;
+			pr_warn("map '%s': failed to alloc content buffer: %s\n",
+				map->name, errstr(err));
+			goto free_name;
+		}
 
-	if (data)
-		memcpy(map->mmaped, data, data_sz);
+		if (data) {
+			memcpy(map->data, data, data_sz);
+			if (data_sz != elem_sz)
+				memset(map->data + data_sz, 0, elem_sz - data_sz);
+		} else {
+			memset(map->data, 0, elem_sz);
+		}
+	} else {
+		if (map_is_mmapable(obj, map))
+			def->map_flags |= BPF_F_MMAPABLE;
+
+		mmap_sz = bpf_map_mmap_sz(map);
+		map->mmaped = mmap(NULL, mmap_sz, PROT_READ | PROT_WRITE,
+				   MAP_SHARED | MAP_ANONYMOUS, -1, 0);
+		if (map->mmaped == MAP_FAILED) {
+			err = -errno;
+			map->mmaped = NULL;
+			pr_warn("map '%s': failed to alloc content buffer: %s\n",
+				map->name, errstr(err));
+			goto free_name;
+		}
+
+		map_copy_data(map, data);
+	}
 
 	pr_debug("map %td is \"%s\"\n", map - obj->maps, map->name);
 	return 0;
+
+free_name:
+	zfree(&map->real_name);
+	zfree(&map->name);
+	return err;
 }
 
 static int bpf_object__init_global_data_maps(struct bpf_object *obj)
@@ -2015,6 +2055,13 @@ static int bpf_object__init_global_data_maps(struct bpf_object *obj)
 							    sec_desc->data->d_buf,
 							    sec_desc->data->d_size);
 			break;
+		case SEC_PERCPU_DATA:
+			sec_name = elf_sec_name(obj, elf_sec_by_idx(obj, sec_idx));
+			err = bpf_object__init_internal_map(obj, LIBBPF_MAP_PERCPU_DATA,
+							    sec_name, sec_idx,
+							    sec_desc->data->d_buf,
+							    sec_desc->data->d_size);
+			break;
 		case SEC_BSS:
 			sec_name = elf_sec_name(obj, elf_sec_by_idx(obj, sec_idx));
 			err = bpf_object__init_internal_map(obj, LIBBPF_MAP_BSS,
@@ -3934,6 +3981,11 @@ static int bpf_object__elf_collect(struct bpf_object *obj)
 				sec_desc->sec_type = SEC_RODATA;
 				sec_desc->shdr = sh;
 				sec_desc->data = data;
+			} else if (strcmp(name, PERCPU_DATA_SEC) == 0 ||
+				   str_has_pfx(name, PERCPU_DATA_SEC)) {
+				sec_desc->sec_type = SEC_PERCPU_DATA;
+				sec_desc->shdr = sh;
+				sec_desc->data = data;
 			} else if (strcmp(name, STRUCT_OPS_SEC) == 0 ||
 				   strcmp(name, STRUCT_OPS_LINK_SEC) == 0 ||
 				   strcmp(name, "?" STRUCT_OPS_SEC) == 0 ||
@@ -4453,6 +4505,7 @@ static bool bpf_object__shndx_is_data(const struct bpf_object *obj,
 	case SEC_BSS:
 	case SEC_DATA:
 	case SEC_RODATA:
+	case SEC_PERCPU_DATA:
 		return true;
 	default:
 		return false;
@@ -4478,6 +4531,8 @@ bpf_object__section_to_libbpf_map_type(const struct bpf_object *obj, int shndx)
 		return LIBBPF_MAP_DATA;
 	case SEC_RODATA:
 		return LIBBPF_MAP_RODATA;
+	case SEC_PERCPU_DATA:
+		return LIBBPF_MAP_PERCPU_DATA;
 	default:
 		return LIBBPF_MAP_UNSPEC;
 	}
@@ -4795,7 +4850,7 @@ static int map_fill_btf_type_info(struct bpf_object *obj, struct bpf_map *map)
 
 	/*
 	 * LLVM annotates global data differently in BTF, that is,
-	 * only as '.data', '.bss' or '.rodata'.
+	 * only as '.data', '.bss', '.percpu' or '.rodata'.
 	 */
 	if (!bpf_map__is_internal(map))
 		return -ENOENT;
@@ -5125,23 +5180,54 @@ static int
 bpf_object__populate_internal_map(struct bpf_object *obj, struct bpf_map *map)
 {
 	enum libbpf_map_type map_type = map->libbpf_type;
-	int err, zero = 0;
+	bool is_percpu_data = map_type == LIBBPF_MAP_PERCPU_DATA;
+	int err = 0, zero = 0;
+	void *data = NULL;
+	int num_cpus, i;
+	size_t data_sz;
+	size_t elem_sz;
 	size_t mmap_sz;
 
+	data_sz = map->def.value_size;
+	if (is_percpu_data) {
+		num_cpus = libbpf_num_possible_cpus();
+		if (num_cpus < 0) {
+			err = libbpf_err_errno(num_cpus);
+			pr_warn("map '%s': failed to get num_cpus: %s\n",
+				bpf_map__name(map), errstr(err));
+			return err;
+		}
+
+		elem_sz = roundup(data_sz, 8);
+		data_sz = elem_sz * num_cpus;
+		data = malloc(data_sz);
+		if (!data) {
+			err = -ENOMEM;
+			pr_warn("map '%s': failed to malloc memory: %s\n",
+				bpf_map__name(map), errstr(err));
+			return err;
+		}
+
+		for (i = 0; i < num_cpus; i++)
+			memcpy(data + i * elem_sz, map->data, elem_sz);
+	} else {
+		data = map->mmaped;
+	}
+
 	if (obj->gen_loader) {
 		bpf_gen__map_update_elem(obj->gen_loader, map - obj->maps,
-					 map->mmaped, map->def.value_size);
+					 data, data_sz);
 		if (map_type == LIBBPF_MAP_RODATA || map_type == LIBBPF_MAP_KCONFIG)
 			bpf_gen__map_freeze(obj->gen_loader, map - obj->maps);
-		return 0;
+		goto free_data;
 	}
 
-	err = bpf_map_update_elem(map->fd, &zero, map->mmaped, 0);
+	err = bpf_map_update_elem(map->fd, &zero, data, 0);
 	if (err) {
 		err = -errno;
 		pr_warn("map '%s': failed to set initial contents: %s\n",
 			bpf_map__name(map), errstr(err));
-		return err;
+		goto free_data;
 	}
 
 	/* Freeze .rodata and .kconfig map as read-only from syscall side. */
@@ -5151,7 +5237,7 @@ bpf_object__populate_internal_map(struct bpf_object *obj, struct bpf_map *map)
 			err = -errno;
 			pr_warn("map '%s': failed to freeze as read-only: %s\n",
 				bpf_map__name(map), errstr(err));
-			return err;
+			goto free_data;
 		}
 	}
 
@@ -5178,7 +5264,7 @@ bpf_object__populate_internal_map(struct bpf_object *obj, struct bpf_map *map)
 			err = -errno;
 			pr_warn("map '%s': failed to re-mmap() contents: %s\n",
 				bpf_map__name(map), errstr(err));
-			return err;
+			goto free_data;
 		}
 		map->mmaped = mmaped;
 	} else if (map->mmaped) {
@@ -5186,7 +5272,10 @@ bpf_object__populate_internal_map(struct bpf_object *obj, struct bpf_map *map)
 		map->mmaped = NULL;
 	}
 
-	return 0;
+free_data:
+	if (is_percpu_data)
+		free(data);
+	return err;
 }
 
 static void bpf_map__destroy(struct bpf_map *map);
@@ -8120,7 +8209,9 @@ static int bpf_object__sanitize_maps(struct bpf_object *obj)
 	struct bpf_map *m;
 
 	bpf_object__for_each_map(m, obj) {
-		if (!bpf_map__is_internal(m))
+		if (!bpf_map__is_internal(m) ||
+		    /* percpu data map is internal and not-mmapable. */
+		    m->libbpf_type == LIBBPF_MAP_PERCPU_DATA)
 			continue;
 		if (!kernel_supports(obj, FEAT_ARRAY_MMAP))
 			m->def.map_flags &= ~BPF_F_MMAPABLE;
@@ -9041,6 +9132,8 @@ static void bpf_map__destroy(struct bpf_map *map)
 	if (map->mmaped && map->mmaped != map->obj->arena_data)
 		munmap(map->mmaped, bpf_map_mmap_sz(map));
 	map->mmaped = NULL;
+	if (map->data)
+		zfree(&map->data);
 
 	if (map->st_ops) {
 		zfree(&map->st_ops->data);
@@ -10132,14 +10225,18 @@ int bpf_map__fd(const struct bpf_map *map)
 
 static bool map_uses_real_name(const struct bpf_map *map)
 {
-	/* Since libbpf started to support custom .data.* and .rodata.* maps,
-	 * their user-visible name differs from kernel-visible name. Users see
-	 * such map's corresponding ELF section name as a map name.
-	 * This check distinguishes .data/.rodata from .data.* and .rodata.*
-	 * maps to know which name has to be returned to the user.
+	/* Since libbpf started to support custom .data.*, .percpu.* and
+	 * .rodata.* maps, their user-visible name differs from kernel-visible
+	 * name. Users see such map's corresponding ELF section name as a map
+	 * name. This check distinguishes .data/.percpu/.rodata from .data.*,
+	 * .percpu.* and .rodata.* maps to know which name has to be returned to
+	 * the user.
 	 */
 	if (map->libbpf_type == LIBBPF_MAP_DATA && strcmp(map->real_name, DATA_SEC) != 0)
 		return true;
+	if (map->libbpf_type == LIBBPF_MAP_PERCPU_DATA &&
+	    strcmp(map->real_name, PERCPU_DATA_SEC) != 0)
+		return true;
 	if (map->libbpf_type == LIBBPF_MAP_RODATA && strcmp(map->real_name, RODATA_SEC) != 0)
 		return true;
 	return false;
@@ -10348,7 +10445,8 @@ int bpf_map__set_initial_value(struct bpf_map *map,
 	if (map->obj->loaded || map->reused)
 		return libbpf_err(-EBUSY);
 
-	if (!map->mmaped || map->libbpf_type == LIBBPF_MAP_KCONFIG)
+	if ((!map->mmaped && !map->data) ||
+	    map->libbpf_type == LIBBPF_MAP_KCONFIG)
 		return libbpf_err(-EINVAL);
 
 	if (map->def.type == BPF_MAP_TYPE_ARENA)
@@ -10358,7 +10456,7 @@ int bpf_map__set_initial_value(struct bpf_map *map,
 	if (size != actual_sz)
 		return libbpf_err(-EINVAL);
 
-	memcpy(map->mmaped, data, size);
+	map_copy_data(map, data);
 	return 0;
 }
 
@@ -10370,7 +10468,7 @@ void *bpf_map__initial_value(const struct bpf_map *map, size_t *psize)
 		return map->st_ops->data;
 	}
 
-	if (!map->mmaped)
+	if (!map->mmaped && !map->data)
 		return NULL;
 
 	if (map->def.type == BPF_MAP_TYPE_ARENA)
@@ -10378,7 +10476,7 @@ void *bpf_map__initial_value(const struct bpf_map *map, size_t *psize)
 	else
 		*psize = map->def.value_size;
 
-	return map->mmaped;
+	return map->libbpf_type == LIBBPF_MAP_PERCPU_DATA ? map->data : map->mmaped;
 }
 
 bool bpf_map__is_internal(const struct bpf_map *map)
-- 
2.47.1

[PATCH bpf-next 3/4] bpf, bpftool: Generate skeleton for global percpu data

[Leon Hwang]

This patch enhances bpftool to generate skeletons for global percpu
variables. The generated skeleton includes a dedicated structure for
percpu data, allowing users to initialize and access percpu variables
efficiently.

Changes:

1. skeleton structure:

For global percpu variables, the skeleton now includes a nested
structure, e.g.:

struct test_global_percpu_data {
	struct bpf_object_skeleton *skeleton;
	struct bpf_object *obj;
	struct {
		struct bpf_map *bss;
		struct bpf_map *percpu;
	} maps;
	// ...
	struct test_global_percpu_data__percpu {
		int percpu_data;
	} *percpu;

	// ...
};

  * The "struct test_global_percpu_data__percpu" points to initialized
    data, which is actually "maps.percpu->data".
  * Before loading the skeleton, updating the
    "struct test_global_percpu_data__percpu" modifies the initial value
    of the corresponding global percpu variables.
  * After loading the skeleton, accessing or updating this struct is no
    longer meaningful. Instead, users must interact with the global
    percpu variables via the "maps.percpu" map.

2. code changes:

  * Added support for ".percpu" sections in bpftool's map identification
    logic.
  * Modified skeleton generation to handle percpu data maps
    appropriately.
  * Updated libbpf to make "percpu" pointing to "maps.percpu->data".

Signed-off-by: Leon Hwang <leon.hwang@linux.dev>
---
 tools/bpf/bpftool/gen.c | 13 +++++++++----
 tools/lib/bpf/libbpf.c  |  3 +++
 tools/lib/bpf/libbpf.h  |  1 +
 3 files changed, 13 insertions(+), 4 deletions(-)

diff --git a/tools/bpf/bpftool/gen.c b/tools/bpf/bpftool/gen.c
index 5a4d3240689ed..975775683ca12 100644
--- a/tools/bpf/bpftool/gen.c
+++ b/tools/bpf/bpftool/gen.c
@@ -92,7 +92,7 @@ static void get_header_guard(char *guard, const char *obj_name, const char *suff
 
 static bool get_map_ident(const struct bpf_map *map, char *buf, size_t buf_sz)
 {
-	static const char *sfxs[] = { ".data", ".rodata", ".bss", ".kconfig" };
+	static const char *sfxs[] = { ".data", ".rodata", ".percpu", ".bss", ".kconfig" };
 	const char *name = bpf_map__name(map);
 	int i, n;
 
@@ -117,7 +117,7 @@ static bool get_map_ident(const struct bpf_map *map, char *buf, size_t buf_sz)
 
 static bool get_datasec_ident(const char *sec_name, char *buf, size_t buf_sz)
 {
-	static const char *pfxs[] = { ".data", ".rodata", ".bss", ".kconfig" };
+	static const char *pfxs[] = { ".data", ".rodata", ".percpu", ".bss", ".kconfig" };
 	int i, n;
 
 	/* recognize hard coded LLVM section name */
@@ -263,7 +263,9 @@ static bool is_mmapable_map(const struct bpf_map *map, char *buf, size_t sz)
 		return true;
 	}
 
-	if (!bpf_map__is_internal(map) || !(bpf_map__map_flags(map) & BPF_F_MMAPABLE))
+	if (!bpf_map__is_internal(map) ||
+	    (!(bpf_map__map_flags(map) & BPF_F_MMAPABLE) &&
+	     bpf_map__type(map) != BPF_MAP_TYPE_PERCPU_ARRAY))
 		return false;
 
 	if (!get_map_ident(map, buf, sz))
@@ -903,7 +905,10 @@ codegen_maps_skeleton(struct bpf_object *obj, size_t map_cnt, bool mmaped, bool
 			i, bpf_map__name(map), ident);
 		/* memory-mapped internal maps */
 		if (mmaped && is_mmapable_map(map, ident, sizeof(ident))) {
-			printf("\tmap->mmaped = (void **)&obj->%s;\n", ident);
+			if (bpf_map__type(map) == BPF_MAP_TYPE_PERCPU_ARRAY)
+				printf("\tmap->data = (void **)&obj->%s;\n", ident);
+			else
+				printf("\tmap->mmaped = (void **)&obj->%s;\n", ident);
 		}
 
 		if (populate_links && bpf_map__type(map) == BPF_MAP_TYPE_STRUCT_OPS) {
diff --git a/tools/lib/bpf/libbpf.c b/tools/lib/bpf/libbpf.c
index 6da6004c5c84d..dafb419bc5b86 100644
--- a/tools/lib/bpf/libbpf.c
+++ b/tools/lib/bpf/libbpf.c
@@ -13915,6 +13915,7 @@ static int populate_skeleton_maps(const struct bpf_object *obj,
 		struct bpf_map **map = map_skel->map;
 		const char *name = map_skel->name;
 		void **mmaped = map_skel->mmaped;
+		void **data = map_skel->data;
 
 		*map = bpf_object__find_map_by_name(obj, name);
 		if (!*map) {
@@ -13925,6 +13926,8 @@ static int populate_skeleton_maps(const struct bpf_object *obj,
 		/* externs shouldn't be pre-setup from user code */
 		if (mmaped && (*map)->libbpf_type != LIBBPF_MAP_KCONFIG)
 			*mmaped = (*map)->mmaped;
+		if (data && (*map)->libbpf_type == LIBBPF_MAP_PERCPU_DATA)
+			*data = (*map)->data;
 	}
 	return 0;
 }
diff --git a/tools/lib/bpf/libbpf.h b/tools/lib/bpf/libbpf.h
index 3020ee45303a0..c49d6e44b5630 100644
--- a/tools/lib/bpf/libbpf.h
+++ b/tools/lib/bpf/libbpf.h
@@ -1701,6 +1701,7 @@ struct bpf_map_skeleton {
 	const char *name;
 	struct bpf_map **map;
 	void **mmaped;
+	void **data;
 	struct bpf_link **link;
 };
 
-- 
2.47.1

[PATCH bpf-next 4/4] selftests/bpf: Add a case to test global percpu data

[Leon Hwang]

If the arch, like s390x, does not support percpu insn, this case won't
test global percpu data by checking -EOPNOTSUPP when load prog.

The following APIs have been tested for global percpu data:
1. bpf_map__set_initial_value()
2. bpf_map__initial_value()
3. generated percpu struct pointer that points to internal map's data
4. bpf_map__lookup_elem() for global percpu data map

cd tools/testing/selftests/bpf; ./test_progs -t global_percpu_data
124     global_percpu_data_init:OK
Summary: 1/0 PASSED, 0 SKIPPED, 0 FAILED

Signed-off-by: Leon Hwang <leon.hwang@linux.dev>
---
 .../bpf/prog_tests/global_data_init.c         | 89 ++++++++++++++++++-
 .../bpf/progs/test_global_percpu_data.c       | 21 +++++
 2 files changed, 109 insertions(+), 1 deletion(-)
 create mode 100644 tools/testing/selftests/bpf/progs/test_global_percpu_data.c

diff --git a/tools/testing/selftests/bpf/prog_tests/global_data_init.c b/tools/testing/selftests/bpf/prog_tests/global_data_init.c
index 8466332d7406f..a5d0890444f67 100644
--- a/tools/testing/selftests/bpf/prog_tests/global_data_init.c
+++ b/tools/testing/selftests/bpf/prog_tests/global_data_init.c
@@ -1,5 +1,6 @@
 // SPDX-License-Identifier: GPL-2.0
 #include <test_progs.h>
+#include "test_global_percpu_data.skel.h"
 
 void test_global_data_init(void)
 {
@@ -8,7 +9,7 @@ void test_global_data_init(void)
 	__u8 *buff = NULL, *newval = NULL;
 	struct bpf_object *obj;
 	struct bpf_map *map;
-        __u32 duration = 0;
+	__u32 duration = 0;
 	size_t sz;
 
 	obj = bpf_object__open_file(file, NULL);
@@ -60,3 +61,89 @@ void test_global_data_init(void)
 	free(newval);
 	bpf_object__close(obj);
 }
+
+void test_global_percpu_data_init(void)
+{
+	struct test_global_percpu_data *skel = NULL;
+	u64 *percpu_data = NULL;
+	struct bpf_map *map;
+	size_t init_data_sz;
+	char buff[128] = {};
+	int init_value = 2;
+	int key, value_sz;
+	int prog_fd, err;
+	int *init_data;
+	int num_cpus;
+
+	LIBBPF_OPTS(bpf_test_run_opts, topts,
+		    .data_in = buff,
+		    .data_size_in = sizeof(buff),
+		    .repeat = 1,
+	);
+
+	num_cpus = libbpf_num_possible_cpus();
+	if (!ASSERT_GT(num_cpus, 0, "libbpf_num_possible_cpus"))
+		return;
+
+	percpu_data = calloc(num_cpus, sizeof(*percpu_data));
+	if (!ASSERT_FALSE(percpu_data == NULL, "calloc percpu_data"))
+		return;
+
+	value_sz = sizeof(*percpu_data) * num_cpus;
+	memset(percpu_data, 0, value_sz);
+
+	skel = test_global_percpu_data__open();
+	if (!ASSERT_OK_PTR(skel, "test_global_percpu_data__open"))
+		goto out;
+
+	ASSERT_EQ(skel->percpu->percpu_data, -1, "skel->percpu->percpu_data");
+
+	map = skel->maps.percpu;
+	err = bpf_map__set_initial_value(map, &init_value,
+					 sizeof(init_value));
+	if (!ASSERT_OK(err, "bpf_map__set_initial_value"))
+		goto out;
+
+	init_data = bpf_map__initial_value(map, &init_data_sz);
+	if (!ASSERT_OK_PTR(init_data, "bpf_map__initial_value"))
+		goto out;
+
+	ASSERT_EQ(*init_data, init_value, "initial_value");
+	ASSERT_EQ(init_data_sz, sizeof(init_value), "initial_value size");
+
+	if (!ASSERT_EQ((void *) init_data, (void *) skel->percpu, "skel->percpu"))
+		goto out;
+	ASSERT_EQ(skel->percpu->percpu_data, init_value, "skel->percpu->percpu_data");
+
+	err = test_global_percpu_data__load(skel);
+	if (err == -EOPNOTSUPP)
+		goto out;
+	if (!ASSERT_OK(err, "test_global_percpu_data__load"))
+		goto out;
+
+	ASSERT_EQ(bpf_map__type(map), BPF_MAP_TYPE_PERCPU_ARRAY,
+		  "bpf_map__type");
+
+	prog_fd = bpf_program__fd(skel->progs.update_percpu_data);
+	err = bpf_prog_test_run_opts(prog_fd, &topts);
+	ASSERT_OK(err, "update_percpu_data");
+	ASSERT_EQ(topts.retval, 0, "update_percpu_data retval");
+
+	key = 0;
+	err = bpf_map__lookup_elem(map, &key, sizeof(key), percpu_data,
+				   value_sz, 0);
+	if (!ASSERT_OK(err, "bpf_map__lookup_elem"))
+		goto out;
+
+	if (!ASSERT_LT(skel->bss->curr_cpu, num_cpus, "curr_cpu"))
+		goto out;
+	ASSERT_EQ((int) percpu_data[skel->bss->curr_cpu], 1, "percpu_data");
+	if (num_cpus > 1)
+		ASSERT_EQ((int) percpu_data[(skel->bss->curr_cpu+1)%num_cpus],
+			  init_value, "init_value");
+
+out:
+	test_global_percpu_data__destroy(skel);
+	if (percpu_data)
+		free(percpu_data);
+}
diff --git a/tools/testing/selftests/bpf/progs/test_global_percpu_data.c b/tools/testing/selftests/bpf/progs/test_global_percpu_data.c
new file mode 100644
index 0000000000000..731c3214b0bb4
--- /dev/null
+++ b/tools/testing/selftests/bpf/progs/test_global_percpu_data.c
@@ -0,0 +1,21 @@
+// SPDX-License-Identifier: GPL-2.0
+/* Copyright Leon Hwang */
+
+#include <linux/bpf.h>
+#include <linux/pkt_cls.h>
+
+#include <bpf/bpf_helpers.h>
+
+int percpu_data SEC(".percpu") = -1;
+int curr_cpu;
+
+SEC("tc")
+int update_percpu_data(struct __sk_buff *skb)
+{
+	curr_cpu = bpf_get_smp_processor_id();
+	percpu_data = 1;
+
+	return TC_ACT_OK;
+}
+
+char _license[] SEC("license") = "GPL";
-- 
2.47.1

Re: [PATCH bpf-next 1/4] bpf: Introduce global percpu data

[Andrii Nakryiko]

On Mon, Jan 27, 2025 at 8:22 AM Leon Hwang <leon.hwang@linux.dev> wrote:
>
> This patch introduces global percpu data, inspired by commit
> 6316f78306c1 ("Merge branch 'support-global-data'"). It enables the
> definition of global percpu variables in BPF, similar to the
> DEFINE_PER_CPU() macro in the kernel[0].
>
> For example, in BPF, it is able to define a global percpu variable like
> this:
>
> int percpu_data SEC(".percpu");
>
> With this patch, tools like retsnoop[1] and bpflbr[2] can simplify their
> BPF code for handling LBRs. The code can be updated from
>
> static struct perf_branch_entry lbrs[1][MAX_LBR_ENTRIES] SEC(".data.lbrs");
>
> to
>
> static struct perf_branch_entry lbrs[MAX_LBR_ENTRIES] SEC(".percpu.lbrs");
>
> This eliminates the need to retrieve the CPU ID using the
> bpf_get_smp_processor_id() helper.
>
> Additionally, by reusing global percpu data map, sharing information
> between tail callers and callees or freplace callers and callees becomes
> simpler compared to reusing percpu_array maps.
>
> Links:
> [0] https://github.com/torvalds/linux/blob/fbfd64d25c7af3b8695201ebc85efe90be28c5a3/include/linux/percpu-defs.h#L114
> [1] https://github.com/anakryiko/retsnoop
> [2] https://github.com/Asphaltt/bpflbr
>
> Signed-off-by: Leon Hwang <leon.hwang@linux.dev>
> ---
>  kernel/bpf/arraymap.c | 39 ++++++++++++++++++++++++++++++++++++-
>  kernel/bpf/verifier.c | 45 +++++++++++++++++++++++++++++++++++++++++++
>  2 files changed, 83 insertions(+), 1 deletion(-)
>

[...]

> diff --git a/kernel/bpf/verifier.c b/kernel/bpf/verifier.c
> index 9971c03adfd5d..9d99497c2b94c 100644
> --- a/kernel/bpf/verifier.c
> +++ b/kernel/bpf/verifier.c
> @@ -6810,6 +6810,8 @@ static int bpf_map_direct_read(struct bpf_map *map, int off, int size, u64 *val,
>         u64 addr;
>         int err;
>
> +       if (map->map_type != BPF_MAP_TYPE_ARRAY)
> +               return -EINVAL;
>         err = map->ops->map_direct_value_addr(map, &addr, off);
>         if (err)
>                 return err;
> @@ -7322,6 +7324,7 @@ static int check_mem_access(struct bpf_verifier_env *env, int insn_idx, u32 regn
>                         /* if map is read-only, track its contents as scalars */
>                         if (tnum_is_const(reg->var_off) &&
>                             bpf_map_is_rdonly(map) &&
> +                           map->map_type != BPF_MAP_TYPE_PERCPU_ARRAY &&

shouldn't this rather be a safer `map->map_type == BPF_MAP_TYPE_ARRAY` check?

>                             map->ops->map_direct_value_addr) {
>                                 int map_off = off + reg->var_off.value;
>                                 u64 val = 0;

[...]

Re: [PATCH bpf-next 2/4] bpf, libbpf: Support global percpu data

[Andrii Nakryiko]

On Mon, Jan 27, 2025 at 8:22 AM Leon Hwang <leon.hwang@linux.dev> wrote:
>
> This patch introduces support for global percpu data in libbpf. A new
> section named ".percpu" is added, similar to the existing ".data" section.
> Internal maps are created for ".percpu" sections, which are then
> initialized and populated accordingly.
>
> The changes include:
>
> * Introduction of the ".percpu" section in libbpf.
> * Creation of internal maps for percpu data.
> * Initialization and population of these maps.
>
> This enhancement allows BPF programs to efficiently manage and access
> percpu global data, improving performance for use cases that require
> percpu buffer.
>
> Signed-off-by: Leon Hwang <leon.hwang@linux.dev>
> ---
>  tools/lib/bpf/libbpf.c | 172 ++++++++++++++++++++++++++++++++---------
>  1 file changed, 135 insertions(+), 37 deletions(-)
>
> diff --git a/tools/lib/bpf/libbpf.c b/tools/lib/bpf/libbpf.c
> index 194809da51725..6da6004c5c84d 100644
> --- a/tools/lib/bpf/libbpf.c
> +++ b/tools/lib/bpf/libbpf.c
> @@ -516,6 +516,7 @@ struct bpf_struct_ops {
>  };
>
>  #define DATA_SEC ".data"
> +#define PERCPU_DATA_SEC ".percpu"
>  #define BSS_SEC ".bss"
>  #define RODATA_SEC ".rodata"
>  #define KCONFIG_SEC ".kconfig"
> @@ -530,6 +531,7 @@ enum libbpf_map_type {
>         LIBBPF_MAP_BSS,
>         LIBBPF_MAP_RODATA,
>         LIBBPF_MAP_KCONFIG,
> +       LIBBPF_MAP_PERCPU_DATA,

nit: let's keep it shorter: LIBBPF_MAP_PERCPU

>  };
>
>  struct bpf_map_def {
> @@ -562,6 +564,7 @@ struct bpf_map {
>         __u32 btf_value_type_id;
>         __u32 btf_vmlinux_value_type_id;
>         enum libbpf_map_type libbpf_type;
> +       void *data;
>         void *mmaped;
>         struct bpf_struct_ops *st_ops;
>         struct bpf_map *inner_map;
> @@ -640,6 +643,7 @@ enum sec_type {
>         SEC_DATA,
>         SEC_RODATA,
>         SEC_ST_OPS,
> +       SEC_PERCPU_DATA,

ditto, just SEC_PERCPU?

>  };
>
>  struct elf_sec_desc {
> @@ -1923,13 +1927,24 @@ static bool map_is_mmapable(struct bpf_object *obj, struct bpf_map *map)
>         return false;
>  }
>
> +static void map_copy_data(struct bpf_map *map, const void *data)
> +{
> +       bool is_percpu_data = map->libbpf_type == LIBBPF_MAP_PERCPU_DATA;
> +       size_t data_sz = map->def.value_size;
> +
> +       if (data)
> +               memcpy(is_percpu_data ? map->data : map->mmaped, data, data_sz);
> +}
> +
>  static int
>  bpf_object__init_internal_map(struct bpf_object *obj, enum libbpf_map_type type,
>                               const char *real_name, int sec_idx, void *data, size_t data_sz)
>  {
> +       bool is_percpu_data = type == LIBBPF_MAP_PERCPU_DATA;
>         struct bpf_map_def *def;
>         struct bpf_map *map;
>         size_t mmap_sz;
> +       size_t elem_sz;

nit: just:

size_t mmap_sz, elem_sz;

>         int err;
>
>         map = bpf_object__add_map(obj);
> @@ -1948,7 +1963,8 @@ bpf_object__init_internal_map(struct bpf_object *obj, enum libbpf_map_type type,
>         }
>
>         def = &map->def;
> -       def->type = BPF_MAP_TYPE_ARRAY;
> +       def->type = is_percpu_data ? BPF_MAP_TYPE_PERCPU_ARRAY
> +                                  : BPF_MAP_TYPE_ARRAY;

nit: single line

>         def->key_size = sizeof(int);
>         def->value_size = data_sz;
>         def->max_entries = 1;

[...]

> +               if (map_is_mmapable(obj, map))
> +                       def->map_flags |= BPF_F_MMAPABLE;
> +
> +               mmap_sz = bpf_map_mmap_sz(map);
> +               map->mmaped = mmap(NULL, mmap_sz, PROT_READ | PROT_WRITE,
> +                                  MAP_SHARED | MAP_ANONYMOUS, -1, 0);
> +               if (map->mmaped == MAP_FAILED) {
> +                       err = -errno;
> +                       map->mmaped = NULL;
> +                       pr_warn("map '%s': failed to alloc content buffer: %s\n",
> +                               map->name, errstr(err));
> +                       goto free_name;
> +               }
> +
> +               map_copy_data(map, data);

why not memcpy() here? you know it's not percpu map, so why obscuring
that memcpy?


> +       }
>
>         pr_debug("map %td is \"%s\"\n", map - obj->maps, map->name);
>         return 0;

[...]

> @@ -5125,23 +5180,54 @@ static int
>  bpf_object__populate_internal_map(struct bpf_object *obj, struct bpf_map *map)
>  {
>         enum libbpf_map_type map_type = map->libbpf_type;
> -       int err, zero = 0;
> +       bool is_percpu_data = map_type == LIBBPF_MAP_PERCPU_DATA;
> +       int err = 0, zero = 0;
> +       void *data = NULL;
> +       int num_cpus, i;
> +       size_t data_sz;
> +       size_t elem_sz;
>         size_t mmap_sz;

nit: keep those size_t variables grouped: `size_t mmap_sz, data_sz, elem_sz;`

>
> +       data_sz = map->def.value_size;
> +       if (is_percpu_data) {
> +               num_cpus = libbpf_num_possible_cpus();
> +               if (num_cpus < 0) {
> +                       err = libbpf_err_errno(num_cpus);

why? num_cpus *IS* error code if num_cpus < 0

> +                       pr_warn("map '%s': failed to get num_cpus: %s\n",
> +                               bpf_map__name(map), errstr(err));

this is unlikely to happen, I'd drop pr_warn()

> +                       return err;
> +               }
> +
> +               elem_sz = roundup(data_sz, 8);
> +               data_sz = elem_sz * num_cpus;
> +               data = malloc(data_sz);
> +               if (!data) {
> +                       err = -ENOMEM;
> +                       pr_warn("map '%s': failed to malloc memory: %s\n",
> +                               bpf_map__name(map), errstr(err));

-ENOMEM is rather self-descriptive (and generally not expected), so
don't add pr_warn() for such cases

> +                       return err;
> +               }
> +
> +               for (i = 0; i < num_cpus; i++)
> +                       memcpy(data + i * elem_sz, map->data, elem_sz);
> +       } else {
> +               data = map->mmaped;
> +       }
> +

[...]

>  static void bpf_map__destroy(struct bpf_map *map);
> @@ -8120,7 +8209,9 @@ static int bpf_object__sanitize_maps(struct bpf_object *obj)
>         struct bpf_map *m;
>
>         bpf_object__for_each_map(m, obj) {
> -               if (!bpf_map__is_internal(m))
> +               if (!bpf_map__is_internal(m) ||
> +                   /* percpu data map is internal and not-mmapable. */
> +                   m->libbpf_type == LIBBPF_MAP_PERCPU_DATA)

original logic would work anyways, no? let's not add unnecessary
special casing here

>                         continue;
>                 if (!kernel_supports(obj, FEAT_ARRAY_MMAP))
>                         m->def.map_flags &= ~BPF_F_MMAPABLE;
> @@ -9041,6 +9132,8 @@ static void bpf_map__destroy(struct bpf_map *map)
>         if (map->mmaped && map->mmaped != map->obj->arena_data)
>                 munmap(map->mmaped, bpf_map_mmap_sz(map));
>         map->mmaped = NULL;
> +       if (map->data)
> +               zfree(&map->data);
>

this whole map->mmaped and map->data duality and duplication seems not
worth it, tbh. Maybe we should keep using map->mmaped (we probably
could name it more generically at some point, but I don't want to
start bike shedding now) even for malloc'ed memory? After all, we
already have ARENA as another special case? WDYT, can your changes be
implemented by reusing map->mmaped, taking into account a type of map?

pw-bot: cr

>         if (map->st_ops) {
>                 zfree(&map->st_ops->data);
> @@ -10132,14 +10225,18 @@ int bpf_map__fd(const struct bpf_map *map)
>
>  static bool map_uses_real_name(const struct bpf_map *map)
>  {
> -       /* Since libbpf started to support custom .data.* and .rodata.* maps,
> -        * their user-visible name differs from kernel-visible name. Users see
> -        * such map's corresponding ELF section name as a map name.
> -        * This check distinguishes .data/.rodata from .data.* and .rodata.*
> -        * maps to know which name has to be returned to the user.
> +       /* Since libbpf started to support custom .data.*, .percpu.* and
> +        * .rodata.* maps, their user-visible name differs from kernel-visible
> +        * name. Users see such map's corresponding ELF section name as a map
> +        * name. This check distinguishes .data/.percpu/.rodata from .data.*,
> +        * .percpu.* and .rodata.* maps to know which name has to be returned to
> +        * the user.
>          */
>         if (map->libbpf_type == LIBBPF_MAP_DATA && strcmp(map->real_name, DATA_SEC) != 0)
>                 return true;
> +       if (map->libbpf_type == LIBBPF_MAP_PERCPU_DATA &&
> +           strcmp(map->real_name, PERCPU_DATA_SEC) != 0)
> +               return true;

nit: shorten LIBBPF_MAP_PERCPU_DATA and keep single line, please

>         if (map->libbpf_type == LIBBPF_MAP_RODATA && strcmp(map->real_name, RODATA_SEC) != 0)
>                 return true;
>         return false;
> @@ -10348,7 +10445,8 @@ int bpf_map__set_initial_value(struct bpf_map *map,
>         if (map->obj->loaded || map->reused)
>                 return libbpf_err(-EBUSY);
>
> -       if (!map->mmaped || map->libbpf_type == LIBBPF_MAP_KCONFIG)
> +       if ((!map->mmaped && !map->data) ||
> +           map->libbpf_type == LIBBPF_MAP_KCONFIG)
>                 return libbpf_err(-EINVAL);
>
>         if (map->def.type == BPF_MAP_TYPE_ARENA)
> @@ -10358,7 +10456,7 @@ int bpf_map__set_initial_value(struct bpf_map *map,
>         if (size != actual_sz)
>                 return libbpf_err(-EINVAL);
>
> -       memcpy(map->mmaped, data, size);
> +       map_copy_data(map, data);
>         return 0;
>  }
>
> @@ -10370,7 +10468,7 @@ void *bpf_map__initial_value(const struct bpf_map *map, size_t *psize)
>                 return map->st_ops->data;
>         }
>
> -       if (!map->mmaped)
> +       if (!map->mmaped && !map->data)
>                 return NULL;
>
>         if (map->def.type == BPF_MAP_TYPE_ARENA)
> @@ -10378,7 +10476,7 @@ void *bpf_map__initial_value(const struct bpf_map *map, size_t *psize)
>         else
>                 *psize = map->def.value_size;
>
> -       return map->mmaped;
> +       return map->libbpf_type == LIBBPF_MAP_PERCPU_DATA ? map->data : map->mmaped;

Good chunk of changes like this wouldn't be necessary if we just reuse
map->mmaped.


>  }
>
>  bool bpf_map__is_internal(const struct bpf_map *map)
> --
> 2.47.1
>

Re: [PATCH bpf-next 3/4] bpf, bpftool: Generate skeleton for global percpu data

[Andrii Nakryiko]

On Mon, Jan 27, 2025 at 8:22 AM Leon Hwang <leon.hwang@linux.dev> wrote:
>
> This patch enhances bpftool to generate skeletons for global percpu
> variables. The generated skeleton includes a dedicated structure for
> percpu data, allowing users to initialize and access percpu variables
> efficiently.
>
> Changes:
>
> 1. skeleton structure:
>
> For global percpu variables, the skeleton now includes a nested
> structure, e.g.:
>
> struct test_global_percpu_data {
>         struct bpf_object_skeleton *skeleton;
>         struct bpf_object *obj;
>         struct {
>                 struct bpf_map *bss;
>                 struct bpf_map *percpu;
>         } maps;
>         // ...
>         struct test_global_percpu_data__percpu {
>                 int percpu_data;
>         } *percpu;
>
>         // ...
> };
>
>   * The "struct test_global_percpu_data__percpu" points to initialized
>     data, which is actually "maps.percpu->data".
>   * Before loading the skeleton, updating the
>     "struct test_global_percpu_data__percpu" modifies the initial value
>     of the corresponding global percpu variables.
>   * After loading the skeleton, accessing or updating this struct is no
>     longer meaningful. Instead, users must interact with the global
>     percpu variables via the "maps.percpu" map.

can we set this pointer to NULL after load to avoid accidental
(successful) reads/writes to it?

>
> 2. code changes:
>
>   * Added support for ".percpu" sections in bpftool's map identification
>     logic.
>   * Modified skeleton generation to handle percpu data maps
>     appropriately.
>   * Updated libbpf to make "percpu" pointing to "maps.percpu->data".
>
> Signed-off-by: Leon Hwang <leon.hwang@linux.dev>
> ---
>  tools/bpf/bpftool/gen.c | 13 +++++++++----
>  tools/lib/bpf/libbpf.c  |  3 +++
>  tools/lib/bpf/libbpf.h  |  1 +
>  3 files changed, 13 insertions(+), 4 deletions(-)
>
> diff --git a/tools/bpf/bpftool/gen.c b/tools/bpf/bpftool/gen.c
> index 5a4d3240689ed..975775683ca12 100644
> --- a/tools/bpf/bpftool/gen.c
> +++ b/tools/bpf/bpftool/gen.c
> @@ -92,7 +92,7 @@ static void get_header_guard(char *guard, const char *obj_name, const char *suff
>
>  static bool get_map_ident(const struct bpf_map *map, char *buf, size_t buf_sz)
>  {
> -       static const char *sfxs[] = { ".data", ".rodata", ".bss", ".kconfig" };
> +       static const char *sfxs[] = { ".data", ".rodata", ".percpu", ".bss", ".kconfig" };
>         const char *name = bpf_map__name(map);
>         int i, n;
>
> @@ -117,7 +117,7 @@ static bool get_map_ident(const struct bpf_map *map, char *buf, size_t buf_sz)
>
>  static bool get_datasec_ident(const char *sec_name, char *buf, size_t buf_sz)
>  {
> -       static const char *pfxs[] = { ".data", ".rodata", ".bss", ".kconfig" };
> +       static const char *pfxs[] = { ".data", ".rodata", ".percpu", ".bss", ".kconfig" };
>         int i, n;
>
>         /* recognize hard coded LLVM section name */
> @@ -263,7 +263,9 @@ static bool is_mmapable_map(const struct bpf_map *map, char *buf, size_t sz)
>                 return true;
>         }
>
> -       if (!bpf_map__is_internal(map) || !(bpf_map__map_flags(map) & BPF_F_MMAPABLE))
> +       if (!bpf_map__is_internal(map) ||
> +           (!(bpf_map__map_flags(map) & BPF_F_MMAPABLE) &&
> +            bpf_map__type(map) != BPF_MAP_TYPE_PERCPU_ARRAY))

there will be no BPF_F_MMAPABLE set for PERCPU_ARRAY, why adding these
unnecessary extra conditionals?

>                 return false;
>
>         if (!get_map_ident(map, buf, sz))
> @@ -903,7 +905,10 @@ codegen_maps_skeleton(struct bpf_object *obj, size_t map_cnt, bool mmaped, bool
>                         i, bpf_map__name(map), ident);
>                 /* memory-mapped internal maps */
>                 if (mmaped && is_mmapable_map(map, ident, sizeof(ident))) {
> -                       printf("\tmap->mmaped = (void **)&obj->%s;\n", ident);
> +                       if (bpf_map__type(map) == BPF_MAP_TYPE_PERCPU_ARRAY)
> +                               printf("\tmap->data = (void **)&obj->%s;\n", ident);
> +                       else
> +                               printf("\tmap->mmaped = (void **)&obj->%s;\n", ident);
>                 }
>
>                 if (populate_links && bpf_map__type(map) == BPF_MAP_TYPE_STRUCT_OPS) {
> diff --git a/tools/lib/bpf/libbpf.c b/tools/lib/bpf/libbpf.c
> index 6da6004c5c84d..dafb419bc5b86 100644
> --- a/tools/lib/bpf/libbpf.c
> +++ b/tools/lib/bpf/libbpf.c
> @@ -13915,6 +13915,7 @@ static int populate_skeleton_maps(const struct bpf_object *obj,
>                 struct bpf_map **map = map_skel->map;
>                 const char *name = map_skel->name;
>                 void **mmaped = map_skel->mmaped;
> +               void **data = map_skel->data;
>
>                 *map = bpf_object__find_map_by_name(obj, name);
>                 if (!*map) {
> @@ -13925,6 +13926,8 @@ static int populate_skeleton_maps(const struct bpf_object *obj,
>                 /* externs shouldn't be pre-setup from user code */
>                 if (mmaped && (*map)->libbpf_type != LIBBPF_MAP_KCONFIG)
>                         *mmaped = (*map)->mmaped;
> +               if (data && (*map)->libbpf_type == LIBBPF_MAP_PERCPU_DATA)
> +                       *data = (*map)->data;
>         }
>         return 0;
>  }
> diff --git a/tools/lib/bpf/libbpf.h b/tools/lib/bpf/libbpf.h
> index 3020ee45303a0..c49d6e44b5630 100644
> --- a/tools/lib/bpf/libbpf.h
> +++ b/tools/lib/bpf/libbpf.h
> @@ -1701,6 +1701,7 @@ struct bpf_map_skeleton {
>         const char *name;
>         struct bpf_map **map;
>         void **mmaped;
> +       void **data;

this is breaking backwards/forward compatibility. let's try to reuse mmaped

>         struct bpf_link **link;
>  };
>
> --
> 2.47.1
>

Re: [PATCH bpf-next 4/4] selftests/bpf: Add a case to test global percpu data

[Andrii Nakryiko]

On Mon, Jan 27, 2025 at 8:22 AM Leon Hwang <leon.hwang@linux.dev> wrote:
>
> If the arch, like s390x, does not support percpu insn, this case won't
> test global percpu data by checking -EOPNOTSUPP when load prog.
>
> The following APIs have been tested for global percpu data:
> 1. bpf_map__set_initial_value()
> 2. bpf_map__initial_value()
> 3. generated percpu struct pointer that points to internal map's data
> 4. bpf_map__lookup_elem() for global percpu data map
>
> cd tools/testing/selftests/bpf; ./test_progs -t global_percpu_data
> 124     global_percpu_data_init:OK
> Summary: 1/0 PASSED, 0 SKIPPED, 0 FAILED
>
> Signed-off-by: Leon Hwang <leon.hwang@linux.dev>
> ---
>  .../bpf/prog_tests/global_data_init.c         | 89 ++++++++++++++++++-
>  .../bpf/progs/test_global_percpu_data.c       | 21 +++++
>  2 files changed, 109 insertions(+), 1 deletion(-)
>  create mode 100644 tools/testing/selftests/bpf/progs/test_global_percpu_data.c
>
> diff --git a/tools/testing/selftests/bpf/prog_tests/global_data_init.c b/tools/testing/selftests/bpf/prog_tests/global_data_init.c
> index 8466332d7406f..a5d0890444f67 100644
> --- a/tools/testing/selftests/bpf/prog_tests/global_data_init.c
> +++ b/tools/testing/selftests/bpf/prog_tests/global_data_init.c
> @@ -1,5 +1,6 @@
>  // SPDX-License-Identifier: GPL-2.0
>  #include <test_progs.h>
> +#include "test_global_percpu_data.skel.h"
>
>  void test_global_data_init(void)
>  {
> @@ -8,7 +9,7 @@ void test_global_data_init(void)
>         __u8 *buff = NULL, *newval = NULL;
>         struct bpf_object *obj;
>         struct bpf_map *map;
> -        __u32 duration = 0;
> +       __u32 duration = 0;
>         size_t sz;
>
>         obj = bpf_object__open_file(file, NULL);
> @@ -60,3 +61,89 @@ void test_global_data_init(void)
>         free(newval);
>         bpf_object__close(obj);
>  }
> +
> +void test_global_percpu_data_init(void)
> +{
> +       struct test_global_percpu_data *skel = NULL;
> +       u64 *percpu_data = NULL;

there is that test_global_percpu_data__percpu type you are declaring
in the BPF skeleton, right? We should try using it here.

And for that array access, we should make sure that it's __aligned(8),
so indexing by CPU index works correctly.

Also, you define per-CPU variable as int, but here it is u64, what's
up with that?

> +       struct bpf_map *map;
> +       size_t init_data_sz;
> +       char buff[128] = {};
> +       int init_value = 2;
> +       int key, value_sz;
> +       int prog_fd, err;
> +       int *init_data;
> +       int num_cpus;
> +

nit: LIBBPF_OPTS below is variable declaration, so there shouldn't be
an empty line here (and maybe group those int variables a bit more
tightly?)

> +       LIBBPF_OPTS(bpf_test_run_opts, topts,
> +                   .data_in = buff,
> +                   .data_size_in = sizeof(buff),
> +                   .repeat = 1,
> +       );
> +
> +       num_cpus = libbpf_num_possible_cpus();
> +       if (!ASSERT_GT(num_cpus, 0, "libbpf_num_possible_cpus"))
> +               return;
> +
> +       percpu_data = calloc(num_cpus, sizeof(*percpu_data));
> +       if (!ASSERT_FALSE(percpu_data == NULL, "calloc percpu_data"))

ASSERT_OK_PTR()

> +               return;
> +
> +       value_sz = sizeof(*percpu_data) * num_cpus;
> +       memset(percpu_data, 0, value_sz);

you calloc()'ed it, it's already zero-initialized


> +
> +       skel = test_global_percpu_data__open();
> +       if (!ASSERT_OK_PTR(skel, "test_global_percpu_data__open"))
> +               goto out;
> +
> +       ASSERT_EQ(skel->percpu->percpu_data, -1, "skel->percpu->percpu_data");
> +
> +       map = skel->maps.percpu;
> +       err = bpf_map__set_initial_value(map, &init_value,
> +                                        sizeof(init_value));
> +       if (!ASSERT_OK(err, "bpf_map__set_initial_value"))
> +               goto out;
> +
> +       init_data = bpf_map__initial_value(map, &init_data_sz);
> +       if (!ASSERT_OK_PTR(init_data, "bpf_map__initial_value"))
> +               goto out;
> +
> +       ASSERT_EQ(*init_data, init_value, "initial_value");
> +       ASSERT_EQ(init_data_sz, sizeof(init_value), "initial_value size");
> +
> +       if (!ASSERT_EQ((void *) init_data, (void *) skel->percpu, "skel->percpu"))
> +               goto out;
> +       ASSERT_EQ(skel->percpu->percpu_data, init_value, "skel->percpu->percpu_data");
> +
> +       err = test_global_percpu_data__load(skel);
> +       if (err == -EOPNOTSUPP)
> +               goto out;
> +       if (!ASSERT_OK(err, "test_global_percpu_data__load"))
> +               goto out;
> +
> +       ASSERT_EQ(bpf_map__type(map), BPF_MAP_TYPE_PERCPU_ARRAY,
> +                 "bpf_map__type");
> +
> +       prog_fd = bpf_program__fd(skel->progs.update_percpu_data);
> +       err = bpf_prog_test_run_opts(prog_fd, &topts);

at least one of BPF programs (don't remember which one, could be
raw_tp) supports specifying CPU index to run on, it would be nice to
loop over CPUs, triggering BPF program on each one and filling per-CPU
variable with current CPU index. Then we can check that all per-CPU
values have expected values.


> +       ASSERT_OK(err, "update_percpu_data");
> +       ASSERT_EQ(topts.retval, 0, "update_percpu_data retval");
> +
> +       key = 0;
> +       err = bpf_map__lookup_elem(map, &key, sizeof(key), percpu_data,
> +                                  value_sz, 0);
> +       if (!ASSERT_OK(err, "bpf_map__lookup_elem"))
> +               goto out;
> +
> +       if (!ASSERT_LT(skel->bss->curr_cpu, num_cpus, "curr_cpu"))
> +               goto out;
> +       ASSERT_EQ((int) percpu_data[skel->bss->curr_cpu], 1, "percpu_data");
> +       if (num_cpus > 1)
> +               ASSERT_EQ((int) percpu_data[(skel->bss->curr_cpu+1)%num_cpus],
> +                         init_value, "init_value");
> +
> +out:
> +       test_global_percpu_data__destroy(skel);
> +       if (percpu_data)
> +               free(percpu_data);
> +}
> diff --git a/tools/testing/selftests/bpf/progs/test_global_percpu_data.c b/tools/testing/selftests/bpf/progs/test_global_percpu_data.c
> new file mode 100644
> index 0000000000000..731c3214b0bb4
> --- /dev/null
> +++ b/tools/testing/selftests/bpf/progs/test_global_percpu_data.c
> @@ -0,0 +1,21 @@
> +// SPDX-License-Identifier: GPL-2.0
> +/* Copyright Leon Hwang */
> +
> +#include <linux/bpf.h>
> +#include <linux/pkt_cls.h>
> +
> +#include <bpf/bpf_helpers.h>
> +
> +int percpu_data SEC(".percpu") = -1;
> +int curr_cpu;
> +
> +SEC("tc")
> +int update_percpu_data(struct __sk_buff *skb)
> +{
> +       curr_cpu = bpf_get_smp_processor_id();
> +       percpu_data = 1;
> +
> +       return TC_ACT_OK;
> +}
> +
> +char _license[] SEC("license") = "GPL";
> --
> 2.47.1
>

Re: [PATCH bpf-next 1/4] bpf: Introduce global percpu data

[Leon Hwang]

On 6/2/25 08:09, Andrii Nakryiko wrote:
> On Mon, Jan 27, 2025 at 8:22 AM Leon Hwang <leon.hwang@linux.dev> wrote:
>>
>> This patch introduces global percpu data, inspired by commit
>> 6316f78306c1 ("Merge branch 'support-global-data'"). It enables the
>> definition of global percpu variables in BPF, similar to the
>> DEFINE_PER_CPU() macro in the kernel[0].
>>
>> For example, in BPF, it is able to define a global percpu variable like
>> this:
>>
>> int percpu_data SEC(".percpu");
>>
>> With this patch, tools like retsnoop[1] and bpflbr[2] can simplify their
>> BPF code for handling LBRs. The code can be updated from
>>
>> static struct perf_branch_entry lbrs[1][MAX_LBR_ENTRIES] SEC(".data.lbrs");
>>
>> to
>>
>> static struct perf_branch_entry lbrs[MAX_LBR_ENTRIES] SEC(".percpu.lbrs");
>>
>> This eliminates the need to retrieve the CPU ID using the
>> bpf_get_smp_processor_id() helper.
>>
>> Additionally, by reusing global percpu data map, sharing information
>> between tail callers and callees or freplace callers and callees becomes
>> simpler compared to reusing percpu_array maps.
>>
>> Links:
>> [0] https://github.com/torvalds/linux/blob/fbfd64d25c7af3b8695201ebc85efe90be28c5a3/include/linux/percpu-defs.h#L114
>> [1] https://github.com/anakryiko/retsnoop
>> [2] https://github.com/Asphaltt/bpflbr
>>
>> Signed-off-by: Leon Hwang <leon.hwang@linux.dev>
>> ---
>>  kernel/bpf/arraymap.c | 39 ++++++++++++++++++++++++++++++++++++-
>>  kernel/bpf/verifier.c | 45 +++++++++++++++++++++++++++++++++++++++++++
>>  2 files changed, 83 insertions(+), 1 deletion(-)
>>
> 
> [...]
> 
>> diff --git a/kernel/bpf/verifier.c b/kernel/bpf/verifier.c
>> index 9971c03adfd5d..9d99497c2b94c 100644
>> --- a/kernel/bpf/verifier.c
>> +++ b/kernel/bpf/verifier.c
>> @@ -6810,6 +6810,8 @@ static int bpf_map_direct_read(struct bpf_map *map, int off, int size, u64 *val,
>>         u64 addr;
>>         int err;
>>
>> +       if (map->map_type != BPF_MAP_TYPE_ARRAY)
>> +               return -EINVAL;
>>         err = map->ops->map_direct_value_addr(map, &addr, off);
>>         if (err)
>>                 return err;
>> @@ -7322,6 +7324,7 @@ static int check_mem_access(struct bpf_verifier_env *env, int insn_idx, u32 regn
>>                         /* if map is read-only, track its contents as scalars */
>>                         if (tnum_is_const(reg->var_off) &&
>>                             bpf_map_is_rdonly(map) &&
>> +                           map->map_type != BPF_MAP_TYPE_PERCPU_ARRAY &&
> 
> shouldn't this rather be a safer `map->map_type == BPF_MAP_TYPE_ARRAY` check?
> 

Ack. I will update it to `map->map_type == BPF_MAP_TYPE_ARRAY`.

Thanks,
Leon

>>                             map->ops->map_direct_value_addr) {
>>                                 int map_off = off + reg->var_off.value;
>>                                 u64 val = 0;
> 
> [...]

Re: [PATCH bpf-next 2/4] bpf, libbpf: Support global percpu data

[Leon Hwang]

On 6/2/25 08:09, Andrii Nakryiko wrote:
> On Mon, Jan 27, 2025 at 8:22 AM Leon Hwang <leon.hwang@linux.dev> wrote:
>>
>> This patch introduces support for global percpu data in libbpf. A new
>> section named ".percpu" is added, similar to the existing ".data" section.
>> Internal maps are created for ".percpu" sections, which are then
>> initialized and populated accordingly.
>>
>> The changes include:
>>
>> * Introduction of the ".percpu" section in libbpf.
>> * Creation of internal maps for percpu data.
>> * Initialization and population of these maps.
>>
>> This enhancement allows BPF programs to efficiently manage and access
>> percpu global data, improving performance for use cases that require
>> percpu buffer.
>>
>> Signed-off-by: Leon Hwang <leon.hwang@linux.dev>
>> ---
>>  tools/lib/bpf/libbpf.c | 172 ++++++++++++++++++++++++++++++++---------
>>  1 file changed, 135 insertions(+), 37 deletions(-)
>>
>> diff --git a/tools/lib/bpf/libbpf.c b/tools/lib/bpf/libbpf.c
>> index 194809da51725..6da6004c5c84d 100644
>> --- a/tools/lib/bpf/libbpf.c
>> +++ b/tools/lib/bpf/libbpf.c
>> @@ -516,6 +516,7 @@ struct bpf_struct_ops {
>>  };
>>
>>  #define DATA_SEC ".data"
>> +#define PERCPU_DATA_SEC ".percpu"
>>  #define BSS_SEC ".bss"
>>  #define RODATA_SEC ".rodata"
>>  #define KCONFIG_SEC ".kconfig"
>> @@ -530,6 +531,7 @@ enum libbpf_map_type {
>>         LIBBPF_MAP_BSS,
>>         LIBBPF_MAP_RODATA,
>>         LIBBPF_MAP_KCONFIG,
>> +       LIBBPF_MAP_PERCPU_DATA,
> 
> nit: let's keep it shorter: LIBBPF_MAP_PERCPU
> 

Ack.

>>  };
>>
>>  struct bpf_map_def {
>> @@ -562,6 +564,7 @@ struct bpf_map {
>>         __u32 btf_value_type_id;
>>         __u32 btf_vmlinux_value_type_id;
>>         enum libbpf_map_type libbpf_type;
>> +       void *data;
>>         void *mmaped;
>>         struct bpf_struct_ops *st_ops;
>>         struct bpf_map *inner_map;
>> @@ -640,6 +643,7 @@ enum sec_type {
>>         SEC_DATA,
>>         SEC_RODATA,
>>         SEC_ST_OPS,
>> +       SEC_PERCPU_DATA,
> 
> ditto, just SEC_PERCPU?
> 

Ack.

>>  };
>>
>>  struct elf_sec_desc {
>> @@ -1923,13 +1927,24 @@ static bool map_is_mmapable(struct bpf_object *obj, struct bpf_map *map)
>>         return false;
>>  }
>>
>> +static void map_copy_data(struct bpf_map *map, const void *data)
>> +{
>> +       bool is_percpu_data = map->libbpf_type == LIBBPF_MAP_PERCPU_DATA;
>> +       size_t data_sz = map->def.value_size;
>> +
>> +       if (data)
>> +               memcpy(is_percpu_data ? map->data : map->mmaped, data, data_sz);
>> +}
>> +
>>  static int
>>  bpf_object__init_internal_map(struct bpf_object *obj, enum libbpf_map_type type,
>>                               const char *real_name, int sec_idx, void *data, size_t data_sz)
>>  {
>> +       bool is_percpu_data = type == LIBBPF_MAP_PERCPU_DATA;
>>         struct bpf_map_def *def;
>>         struct bpf_map *map;
>>         size_t mmap_sz;
>> +       size_t elem_sz;
> 
> nit: just:
> 
> size_t mmap_sz, elem_sz;
> 
>>         int err;
>>
>>         map = bpf_object__add_map(obj);
>> @@ -1948,7 +1963,8 @@ bpf_object__init_internal_map(struct bpf_object *obj, enum libbpf_map_type type,
>>         }
>>
>>         def = &map->def;
>> -       def->type = BPF_MAP_TYPE_ARRAY;
>> +       def->type = is_percpu_data ? BPF_MAP_TYPE_PERCPU_ARRAY
>> +                                  : BPF_MAP_TYPE_ARRAY;
> 
> nit: single line
> 
>>         def->key_size = sizeof(int);
>>         def->value_size = data_sz;
>>         def->max_entries = 1;
> 
> [...]
> 
>> +               if (map_is_mmapable(obj, map))
>> +                       def->map_flags |= BPF_F_MMAPABLE;
>> +
>> +               mmap_sz = bpf_map_mmap_sz(map);
>> +               map->mmaped = mmap(NULL, mmap_sz, PROT_READ | PROT_WRITE,
>> +                                  MAP_SHARED | MAP_ANONYMOUS, -1, 0);
>> +               if (map->mmaped == MAP_FAILED) {
>> +                       err = -errno;
>> +                       map->mmaped = NULL;
>> +                       pr_warn("map '%s': failed to alloc content buffer: %s\n",
>> +                               map->name, errstr(err));
>> +                       goto free_name;
>> +               }
>> +
>> +               map_copy_data(map, data);
> 
> why not memcpy() here? you know it's not percpu map, so why obscuring
> that memcpy?
> 
> 
>> +       }
>>
>>         pr_debug("map %td is \"%s\"\n", map - obj->maps, map->name);
>>         return 0;
> 
> [...]
> 
>> @@ -5125,23 +5180,54 @@ static int
>>  bpf_object__populate_internal_map(struct bpf_object *obj, struct bpf_map *map)
>>  {
>>         enum libbpf_map_type map_type = map->libbpf_type;
>> -       int err, zero = 0;
>> +       bool is_percpu_data = map_type == LIBBPF_MAP_PERCPU_DATA;
>> +       int err = 0, zero = 0;
>> +       void *data = NULL;
>> +       int num_cpus, i;
>> +       size_t data_sz;
>> +       size_t elem_sz;
>>         size_t mmap_sz;
> 
> nit: keep those size_t variables grouped: `size_t mmap_sz, data_sz, elem_sz;`
> 

Ack.

>>
>> +       data_sz = map->def.value_size;
>> +       if (is_percpu_data) {
>> +               num_cpus = libbpf_num_possible_cpus();
>> +               if (num_cpus < 0) {
>> +                       err = libbpf_err_errno(num_cpus);
> 
> why? num_cpus *IS* error code if num_cpus < 0
> 

Ack. Thanks for pointing this out.

>> +                       pr_warn("map '%s': failed to get num_cpus: %s\n",
>> +                               bpf_map__name(map), errstr(err));
> 
> this is unlikely to happen, I'd drop pr_warn()
> 

Ack.

>> +                       return err;
>> +               }
>> +
>> +               elem_sz = roundup(data_sz, 8);
>> +               data_sz = elem_sz * num_cpus;
>> +               data = malloc(data_sz);
>> +               if (!data) {
>> +                       err = -ENOMEM;
>> +                       pr_warn("map '%s': failed to malloc memory: %s\n",
>> +                               bpf_map__name(map), errstr(err));
> 
> -ENOMEM is rather self-descriptive (and generally not expected), so
> don't add pr_warn() for such cases
> 

Ack.

>> +                       return err;
>> +               }
>> +
>> +               for (i = 0; i < num_cpus; i++)
>> +                       memcpy(data + i * elem_sz, map->data, elem_sz);
>> +       } else {
>> +               data = map->mmaped;
>> +       }
>> +
> 
> [...]
> 
>>  static void bpf_map__destroy(struct bpf_map *map);
>> @@ -8120,7 +8209,9 @@ static int bpf_object__sanitize_maps(struct bpf_object *obj)
>>         struct bpf_map *m;
>>
>>         bpf_object__for_each_map(m, obj) {
>> -               if (!bpf_map__is_internal(m))
>> +               if (!bpf_map__is_internal(m) ||
>> +                   /* percpu data map is internal and not-mmapable. */
>> +                   m->libbpf_type == LIBBPF_MAP_PERCPU_DATA)
> 
> original logic would work anyways, no? let's not add unnecessary
> special casing here
> 

The original logic works for it. I'll remove it.

>>                         continue;
>>                 if (!kernel_supports(obj, FEAT_ARRAY_MMAP))
>>                         m->def.map_flags &= ~BPF_F_MMAPABLE;
>> @@ -9041,6 +9132,8 @@ static void bpf_map__destroy(struct bpf_map *map)
>>         if (map->mmaped && map->mmaped != map->obj->arena_data)
>>                 munmap(map->mmaped, bpf_map_mmap_sz(map));
>>         map->mmaped = NULL;
>> +       if (map->data)
>> +               zfree(&map->data);
>>
> 
> this whole map->mmaped and map->data duality and duplication seems not
> worth it, tbh. Maybe we should keep using map->mmaped (we probably
> could name it more generically at some point, but I don't want to
> start bike shedding now) even for malloc'ed memory? After all, we
> already have ARENA as another special case? WDYT, can your changes be
> implemented by reusing map->mmaped, taking into account a type of map?
> 

It is better to reuse map->mmaped. I'll take a try.

> pw-bot: cr
> 
>>         if (map->st_ops) {
>>                 zfree(&map->st_ops->data);
>> @@ -10132,14 +10225,18 @@ int bpf_map__fd(const struct bpf_map *map)
>>
>>  static bool map_uses_real_name(const struct bpf_map *map)
>>  {
>> -       /* Since libbpf started to support custom .data.* and .rodata.* maps,
>> -        * their user-visible name differs from kernel-visible name. Users see
>> -        * such map's corresponding ELF section name as a map name.
>> -        * This check distinguishes .data/.rodata from .data.* and .rodata.*
>> -        * maps to know which name has to be returned to the user.
>> +       /* Since libbpf started to support custom .data.*, .percpu.* and
>> +        * .rodata.* maps, their user-visible name differs from kernel-visible
>> +        * name. Users see such map's corresponding ELF section name as a map
>> +        * name. This check distinguishes .data/.percpu/.rodata from .data.*,
>> +        * .percpu.* and .rodata.* maps to know which name has to be returned to
>> +        * the user.
>>          */
>>         if (map->libbpf_type == LIBBPF_MAP_DATA && strcmp(map->real_name, DATA_SEC) != 0)
>>                 return true;
>> +       if (map->libbpf_type == LIBBPF_MAP_PERCPU_DATA &&
>> +           strcmp(map->real_name, PERCPU_DATA_SEC) != 0)
>> +               return true;
> 
> nit: shorten LIBBPF_MAP_PERCPU_DATA and keep single line, please
> 

Ack.

>>         if (map->libbpf_type == LIBBPF_MAP_RODATA && strcmp(map->real_name, RODATA_SEC) != 0)
>>                 return true;
>>         return false;
>> @@ -10348,7 +10445,8 @@ int bpf_map__set_initial_value(struct bpf_map *map,
>>         if (map->obj->loaded || map->reused)
>>                 return libbpf_err(-EBUSY);
>>
>> -       if (!map->mmaped || map->libbpf_type == LIBBPF_MAP_KCONFIG)
>> +       if ((!map->mmaped && !map->data) ||
>> +           map->libbpf_type == LIBBPF_MAP_KCONFIG)
>>                 return libbpf_err(-EINVAL);
>>
>>         if (map->def.type == BPF_MAP_TYPE_ARENA)
>> @@ -10358,7 +10456,7 @@ int bpf_map__set_initial_value(struct bpf_map *map,
>>         if (size != actual_sz)
>>                 return libbpf_err(-EINVAL);
>>
>> -       memcpy(map->mmaped, data, size);
>> +       map_copy_data(map, data);
>>         return 0;
>>  }
>>
>> @@ -10370,7 +10468,7 @@ void *bpf_map__initial_value(const struct bpf_map *map, size_t *psize)
>>                 return map->st_ops->data;
>>         }
>>
>> -       if (!map->mmaped)
>> +       if (!map->mmaped && !map->data)
>>                 return NULL;
>>
>>         if (map->def.type == BPF_MAP_TYPE_ARENA)
>> @@ -10378,7 +10476,7 @@ void *bpf_map__initial_value(const struct bpf_map *map, size_t *psize)
>>         else
>>                 *psize = map->def.value_size;
>>
>> -       return map->mmaped;
>> +       return map->libbpf_type == LIBBPF_MAP_PERCPU_DATA ? map->data : map->mmaped;
> 
> Good chunk of changes like this wouldn't be necessary if we just reuse
> map->mmaped.
> 

Yes. you're right. It's unnecessary to change it if reuse map->mmaped.

> 
>>  }
>>
>>  bool bpf_map__is_internal(const struct bpf_map *map)
>> --
>> 2.47.1
>>


Thanks,
Leon

Re: [PATCH bpf-next 3/4] bpf, bpftool: Generate skeleton for global percpu data

[Leon Hwang]

On 6/2/25 08:09, Andrii Nakryiko wrote:
> On Mon, Jan 27, 2025 at 8:22 AM Leon Hwang <leon.hwang@linux.dev> wrote:
>>
>> This patch enhances bpftool to generate skeletons for global percpu
>> variables. The generated skeleton includes a dedicated structure for
>> percpu data, allowing users to initialize and access percpu variables
>> efficiently.
>>
>> Changes:
>>
>> 1. skeleton structure:
>>
>> For global percpu variables, the skeleton now includes a nested
>> structure, e.g.:
>>
>> struct test_global_percpu_data {
>>         struct bpf_object_skeleton *skeleton;
>>         struct bpf_object *obj;
>>         struct {
>>                 struct bpf_map *bss;
>>                 struct bpf_map *percpu;
>>         } maps;
>>         // ...
>>         struct test_global_percpu_data__percpu {
>>                 int percpu_data;
>>         } *percpu;
>>
>>         // ...
>> };
>>
>>   * The "struct test_global_percpu_data__percpu" points to initialized
>>     data, which is actually "maps.percpu->data".
>>   * Before loading the skeleton, updating the
>>     "struct test_global_percpu_data__percpu" modifies the initial value
>>     of the corresponding global percpu variables.
>>   * After loading the skeleton, accessing or updating this struct is no
>>     longer meaningful. Instead, users must interact with the global
>>     percpu variables via the "maps.percpu" map.
> 
> can we set this pointer to NULL after load to avoid accidental
> (successful) reads/writes to it?
> 

Good idea. I'll try to achieve it.

>>
>> 2. code changes:
>>
>>   * Added support for ".percpu" sections in bpftool's map identification
>>     logic.
>>   * Modified skeleton generation to handle percpu data maps
>>     appropriately.
>>   * Updated libbpf to make "percpu" pointing to "maps.percpu->data".
>>
>> Signed-off-by: Leon Hwang <leon.hwang@linux.dev>
>> ---
>>  tools/bpf/bpftool/gen.c | 13 +++++++++----
>>  tools/lib/bpf/libbpf.c  |  3 +++
>>  tools/lib/bpf/libbpf.h  |  1 +
>>  3 files changed, 13 insertions(+), 4 deletions(-)
>>
>> diff --git a/tools/bpf/bpftool/gen.c b/tools/bpf/bpftool/gen.c
>> index 5a4d3240689ed..975775683ca12 100644
>> --- a/tools/bpf/bpftool/gen.c
>> +++ b/tools/bpf/bpftool/gen.c
>> @@ -92,7 +92,7 @@ static void get_header_guard(char *guard, const char *obj_name, const char *suff
>>
>>  static bool get_map_ident(const struct bpf_map *map, char *buf, size_t buf_sz)
>>  {
>> -       static const char *sfxs[] = { ".data", ".rodata", ".bss", ".kconfig" };
>> +       static const char *sfxs[] = { ".data", ".rodata", ".percpu", ".bss", ".kconfig" };
>>         const char *name = bpf_map__name(map);
>>         int i, n;
>>
>> @@ -117,7 +117,7 @@ static bool get_map_ident(const struct bpf_map *map, char *buf, size_t buf_sz)
>>
>>  static bool get_datasec_ident(const char *sec_name, char *buf, size_t buf_sz)
>>  {
>> -       static const char *pfxs[] = { ".data", ".rodata", ".bss", ".kconfig" };
>> +       static const char *pfxs[] = { ".data", ".rodata", ".percpu", ".bss", ".kconfig" };
>>         int i, n;
>>
>>         /* recognize hard coded LLVM section name */
>> @@ -263,7 +263,9 @@ static bool is_mmapable_map(const struct bpf_map *map, char *buf, size_t sz)
>>                 return true;
>>         }
>>
>> -       if (!bpf_map__is_internal(map) || !(bpf_map__map_flags(map) & BPF_F_MMAPABLE))
>> +       if (!bpf_map__is_internal(map) ||
>> +           (!(bpf_map__map_flags(map) & BPF_F_MMAPABLE) &&
>> +            bpf_map__type(map) != BPF_MAP_TYPE_PERCPU_ARRAY))
> 
> there will be no BPF_F_MMAPABLE set for PERCPU_ARRAY, why adding these
> unnecessary extra conditionals?
> 

Ack. It's unnecessary.

>>                 return false;
>>
>>         if (!get_map_ident(map, buf, sz))
>> @@ -903,7 +905,10 @@ codegen_maps_skeleton(struct bpf_object *obj, size_t map_cnt, bool mmaped, bool
>>                         i, bpf_map__name(map), ident);
>>                 /* memory-mapped internal maps */
>>                 if (mmaped && is_mmapable_map(map, ident, sizeof(ident))) {
>> -                       printf("\tmap->mmaped = (void **)&obj->%s;\n", ident);
>> +                       if (bpf_map__type(map) == BPF_MAP_TYPE_PERCPU_ARRAY)
>> +                               printf("\tmap->data = (void **)&obj->%s;\n", ident);
>> +                       else
>> +                               printf("\tmap->mmaped = (void **)&obj->%s;\n", ident);
>>                 }
>>
>>                 if (populate_links && bpf_map__type(map) == BPF_MAP_TYPE_STRUCT_OPS) {
>> diff --git a/tools/lib/bpf/libbpf.c b/tools/lib/bpf/libbpf.c
>> index 6da6004c5c84d..dafb419bc5b86 100644
>> --- a/tools/lib/bpf/libbpf.c
>> +++ b/tools/lib/bpf/libbpf.c
>> @@ -13915,6 +13915,7 @@ static int populate_skeleton_maps(const struct bpf_object *obj,
>>                 struct bpf_map **map = map_skel->map;
>>                 const char *name = map_skel->name;
>>                 void **mmaped = map_skel->mmaped;
>> +               void **data = map_skel->data;
>>
>>                 *map = bpf_object__find_map_by_name(obj, name);
>>                 if (!*map) {
>> @@ -13925,6 +13926,8 @@ static int populate_skeleton_maps(const struct bpf_object *obj,
>>                 /* externs shouldn't be pre-setup from user code */
>>                 if (mmaped && (*map)->libbpf_type != LIBBPF_MAP_KCONFIG)
>>                         *mmaped = (*map)->mmaped;
>> +               if (data && (*map)->libbpf_type == LIBBPF_MAP_PERCPU_DATA)
>> +                       *data = (*map)->data;
>>         }
>>         return 0;
>>  }
>> diff --git a/tools/lib/bpf/libbpf.h b/tools/lib/bpf/libbpf.h
>> index 3020ee45303a0..c49d6e44b5630 100644
>> --- a/tools/lib/bpf/libbpf.h
>> +++ b/tools/lib/bpf/libbpf.h
>> @@ -1701,6 +1701,7 @@ struct bpf_map_skeleton {
>>         const char *name;
>>         struct bpf_map **map;
>>         void **mmaped;
>> +       void **data;
> 
> this is breaking backwards/forward compatibility. let's try to reuse mmaped
> 

Indeed. If reuse map->mmaped for global percpu variables, it's
unnecessary to add this "data".

>>         struct bpf_link **link;
>>  };
>>
>> --
>> 2.47.1
>>

Thanks,
Leon

Re: [PATCH bpf-next 4/4] selftests/bpf: Add a case to test global percpu data

[Leon Hwang]

On 6/2/25 08:09, Andrii Nakryiko wrote:
> On Mon, Jan 27, 2025 at 8:22 AM Leon Hwang <leon.hwang@linux.dev> wrote:
>>
>> If the arch, like s390x, does not support percpu insn, this case won't
>> test global percpu data by checking -EOPNOTSUPP when load prog.
>>
>> The following APIs have been tested for global percpu data:
>> 1. bpf_map__set_initial_value()
>> 2. bpf_map__initial_value()
>> 3. generated percpu struct pointer that points to internal map's data
>> 4. bpf_map__lookup_elem() for global percpu data map
>>
>> cd tools/testing/selftests/bpf; ./test_progs -t global_percpu_data
>> 124     global_percpu_data_init:OK
>> Summary: 1/0 PASSED, 0 SKIPPED, 0 FAILED
>>
>> Signed-off-by: Leon Hwang <leon.hwang@linux.dev>
>> ---
>>  .../bpf/prog_tests/global_data_init.c         | 89 ++++++++++++++++++-
>>  .../bpf/progs/test_global_percpu_data.c       | 21 +++++
>>  2 files changed, 109 insertions(+), 1 deletion(-)
>>  create mode 100644 tools/testing/selftests/bpf/progs/test_global_percpu_data.c
>>
>> diff --git a/tools/testing/selftests/bpf/prog_tests/global_data_init.c b/tools/testing/selftests/bpf/prog_tests/global_data_init.c
>> index 8466332d7406f..a5d0890444f67 100644
>> --- a/tools/testing/selftests/bpf/prog_tests/global_data_init.c
>> +++ b/tools/testing/selftests/bpf/prog_tests/global_data_init.c
>> @@ -1,5 +1,6 @@
>>  // SPDX-License-Identifier: GPL-2.0
>>  #include <test_progs.h>
>> +#include "test_global_percpu_data.skel.h"
>>
>>  void test_global_data_init(void)
>>  {
>> @@ -8,7 +9,7 @@ void test_global_data_init(void)
>>         __u8 *buff = NULL, *newval = NULL;
>>         struct bpf_object *obj;
>>         struct bpf_map *map;
>> -        __u32 duration = 0;
>> +       __u32 duration = 0;
>>         size_t sz;
>>
>>         obj = bpf_object__open_file(file, NULL);
>> @@ -60,3 +61,89 @@ void test_global_data_init(void)
>>         free(newval);
>>         bpf_object__close(obj);
>>  }
>> +
>> +void test_global_percpu_data_init(void)
>> +{
>> +       struct test_global_percpu_data *skel = NULL;
>> +       u64 *percpu_data = NULL;
> 
> there is that test_global_percpu_data__percpu type you are declaring
> in the BPF skeleton, right? We should try using it here.
> 

No. bpftool does not generate test_global_percpu_data__percpu. The
struct for global variables is embedded into skeleton struct.

Should we generate type for global variables?

> And for that array access, we should make sure that it's __aligned(8),
> so indexing by CPU index works correctly.
> 

Ack.

> Also, you define per-CPU variable as int, but here it is u64, what's
> up with that?
> 

Like __aligned(8), it's to make sure 8-bytes aligned. It's better to use
__aligned(8).

>> +       struct bpf_map *map;
>> +       size_t init_data_sz;
>> +       char buff[128] = {};
>> +       int init_value = 2;
>> +       int key, value_sz;
>> +       int prog_fd, err;
>> +       int *init_data;
>> +       int num_cpus;
>> +
> 
> nit: LIBBPF_OPTS below is variable declaration, so there shouldn't be
> an empty line here (and maybe group those int variables a bit more
> tightly?)
> 

Ack.

>> +       LIBBPF_OPTS(bpf_test_run_opts, topts,
>> +                   .data_in = buff,
>> +                   .data_size_in = sizeof(buff),
>> +                   .repeat = 1,
>> +       );
>> +
>> +       num_cpus = libbpf_num_possible_cpus();
>> +       if (!ASSERT_GT(num_cpus, 0, "libbpf_num_possible_cpus"))
>> +               return;
>> +
>> +       percpu_data = calloc(num_cpus, sizeof(*percpu_data));
>> +       if (!ASSERT_FALSE(percpu_data == NULL, "calloc percpu_data"))
> 
> ASSERT_OK_PTR()
> 

Ack.

>> +               return;
>> +
>> +       value_sz = sizeof(*percpu_data) * num_cpus;
>> +       memset(percpu_data, 0, value_sz);
> 
> you calloc()'ed it, it's already zero-initialized
> 

Ack. Thanks. I should check "man calloc" to use it.

> 
>> +
>> +       skel = test_global_percpu_data__open();
>> +       if (!ASSERT_OK_PTR(skel, "test_global_percpu_data__open"))
>> +               goto out;
>> +
>> +       ASSERT_EQ(skel->percpu->percpu_data, -1, "skel->percpu->percpu_data");
>> +
>> +       map = skel->maps.percpu;
>> +       err = bpf_map__set_initial_value(map, &init_value,
>> +                                        sizeof(init_value));
>> +       if (!ASSERT_OK(err, "bpf_map__set_initial_value"))
>> +               goto out;
>> +
>> +       init_data = bpf_map__initial_value(map, &init_data_sz);
>> +       if (!ASSERT_OK_PTR(init_data, "bpf_map__initial_value"))
>> +               goto out;
>> +
>> +       ASSERT_EQ(*init_data, init_value, "initial_value");
>> +       ASSERT_EQ(init_data_sz, sizeof(init_value), "initial_value size");
>> +
>> +       if (!ASSERT_EQ((void *) init_data, (void *) skel->percpu, "skel->percpu"))
>> +               goto out;
>> +       ASSERT_EQ(skel->percpu->percpu_data, init_value, "skel->percpu->percpu_data");
>> +
>> +       err = test_global_percpu_data__load(skel);
>> +       if (err == -EOPNOTSUPP)
>> +               goto out;
>> +       if (!ASSERT_OK(err, "test_global_percpu_data__load"))
>> +               goto out;
>> +
>> +       ASSERT_EQ(bpf_map__type(map), BPF_MAP_TYPE_PERCPU_ARRAY,
>> +                 "bpf_map__type");
>> +
>> +       prog_fd = bpf_program__fd(skel->progs.update_percpu_data);
>> +       err = bpf_prog_test_run_opts(prog_fd, &topts);
> 
> at least one of BPF programs (don't remember which one, could be
> raw_tp) supports specifying CPU index to run on, it would be nice to
> loop over CPUs, triggering BPF program on each one and filling per-CPU
> variable with current CPU index. Then we can check that all per-CPU
> values have expected values.
> 

Do you mean prog_tests/perf_buffer.c::trigger_on_cpu()?

Your suggestion looks good to me. I'll do it.

> 
>> +       ASSERT_OK(err, "update_percpu_data");
>> +       ASSERT_EQ(topts.retval, 0, "update_percpu_data retval");
>> +
>> +       key = 0;
>> +       err = bpf_map__lookup_elem(map, &key, sizeof(key), percpu_data,
>> +                                  value_sz, 0);
>> +       if (!ASSERT_OK(err, "bpf_map__lookup_elem"))
>> +               goto out;
>> +

[...]

Thanks,
Leon

Re: [PATCH bpf-next 4/4] selftests/bpf: Add a case to test global percpu data

[Andrii Nakryiko]

On Fri, Feb 7, 2025 at 2:00 AM Leon Hwang <leon.hwang@linux.dev> wrote:
>
>
>
> On 6/2/25 08:09, Andrii Nakryiko wrote:
> > On Mon, Jan 27, 2025 at 8:22 AM Leon Hwang <leon.hwang@linux.dev> wrote:
> >>
> >> If the arch, like s390x, does not support percpu insn, this case won't
> >> test global percpu data by checking -EOPNOTSUPP when load prog.
> >>
> >> The following APIs have been tested for global percpu data:
> >> 1. bpf_map__set_initial_value()
> >> 2. bpf_map__initial_value()
> >> 3. generated percpu struct pointer that points to internal map's data
> >> 4. bpf_map__lookup_elem() for global percpu data map
> >>
> >> cd tools/testing/selftests/bpf; ./test_progs -t global_percpu_data
> >> 124     global_percpu_data_init:OK
> >> Summary: 1/0 PASSED, 0 SKIPPED, 0 FAILED
> >>
> >> Signed-off-by: Leon Hwang <leon.hwang@linux.dev>
> >> ---
> >>  .../bpf/prog_tests/global_data_init.c         | 89 ++++++++++++++++++-
> >>  .../bpf/progs/test_global_percpu_data.c       | 21 +++++
> >>  2 files changed, 109 insertions(+), 1 deletion(-)
> >>  create mode 100644 tools/testing/selftests/bpf/progs/test_global_percpu_data.c
> >>

[...]

> >> +void test_global_percpu_data_init(void)
> >> +{
> >> +       struct test_global_percpu_data *skel = NULL;
> >> +       u64 *percpu_data = NULL;
> >
> > there is that test_global_percpu_data__percpu type you are declaring
> > in the BPF skeleton, right? We should try using it here.
> >
>
> No. bpftool does not generate test_global_percpu_data__percpu. The
> struct for global variables is embedded into skeleton struct.
>
> Should we generate type for global variables?

we already have custom skeleton-specific type for .data, .rodata,
.bss, so we should provide one for .percpu as well, yes

>
> > And for that array access, we should make sure that it's __aligned(8),
> > so indexing by CPU index works correctly.
> >
>
> Ack.
>
> > Also, you define per-CPU variable as int, but here it is u64, what's
> > up with that?
> >
>
> Like __aligned(8), it's to make sure 8-bytes aligned. It's better to use
> __aligned(8).

It's hacky, and it won't work correctly on big-endian architectures.
But you shouldn't need that if we have a struct representing this
.percpu memory image. Just make sure that struct has 8 byte alignment
(from bpftool side during skeleton generation, probably).

[...]

> > at least one of BPF programs (don't remember which one, could be
> > raw_tp) supports specifying CPU index to run on, it would be nice to
> > loop over CPUs, triggering BPF program on each one and filling per-CPU
> > variable with current CPU index. Then we can check that all per-CPU
> > values have expected values.
> >
>
> Do you mean prog_tests/perf_buffer.c::trigger_on_cpu()?
>

No, look at `cpu` field of `struct bpf_test_run_opts`. We should have
a selftest using it, so you can work backwards from that.

> Your suggestion looks good to me. I'll do it.
>
> >
> >> +       ASSERT_OK(err, "update_percpu_data");
> >> +       ASSERT_EQ(topts.retval, 0, "update_percpu_data retval");
> >> +
> >> +       key = 0;
> >> +       err = bpf_map__lookup_elem(map, &key, sizeof(key), percpu_data,
> >> +                                  value_sz, 0);
> >> +       if (!ASSERT_OK(err, "bpf_map__lookup_elem"))
> >> +               goto out;
> >> +
>
> [...]
>
> Thanks,
> Leon
>
>

Re: [PATCH bpf-next 1/4] bpf: Introduce global percpu data

[Alexei Starovoitov]

On Mon, Jan 27, 2025 at 8:22 AM Leon Hwang <leon.hwang@linux.dev> wrote:
>
> +
> +static int percpu_array_map_direct_value_meta(const struct bpf_map *map,
> +                                             u64 imm, u32 *off)
> +{
> +       struct bpf_array *array = container_of(map, struct bpf_array, map);
> +       u64 base = (u64) array->pptrs[0];
> +       u64 range = array->elem_size;
> +
> +       if (map->max_entries != 1)
> +               return -EOPNOTSUPP;
> +       if (imm < base || imm >= base + range)
> +               return -ENOENT;
> +       if (!bpf_jit_supports_percpu_insn())
> +               return -EOPNOTSUPP;
> +
> +       *off = imm - base;
> +       return 0;
> +}

Pls add a selftest for off != 0.
I think the above should work, but this is not obvious.

>
> +#ifdef CONFIG_SMP
> +               if (insn->code == (BPF_LD | BPF_IMM | BPF_DW) &&
> +                   (insn->src_reg == BPF_PSEUDO_MAP_VALUE ||
> +                    insn->src_reg == BPF_PSEUDO_MAP_IDX_VALUE)) {

Is there a selftest for BPF_PSEUDO_MAP_IDX_VALUE part ?
I couldn't find it.

Re: [PATCH bpf-next 1/4] bpf: Introduce global percpu data

[Leon Hwang]

On 8/2/25 08:23, Alexei Starovoitov wrote:
> On Mon, Jan 27, 2025 at 8:22 AM Leon Hwang <leon.hwang@linux.dev> wrote:
>>
>> +
>> +static int percpu_array_map_direct_value_meta(const struct bpf_map *map,
>> +                                             u64 imm, u32 *off)
>> +{
>> +       struct bpf_array *array = container_of(map, struct bpf_array, map);
>> +       u64 base = (u64) array->pptrs[0];
>> +       u64 range = array->elem_size;
>> +
>> +       if (map->max_entries != 1)
>> +               return -EOPNOTSUPP;
>> +       if (imm < base || imm >= base + range)
>> +               return -ENOENT;
>> +       if (!bpf_jit_supports_percpu_insn())
>> +               return -EOPNOTSUPP;
>> +
>> +       *off = imm - base;
>> +       return 0;
>> +}
> 
> Pls add a selftest for off != 0.
> I think the above should work, but this is not obvious.
> 

Ack.

>>
>> +#ifdef CONFIG_SMP
>> +               if (insn->code == (BPF_LD | BPF_IMM | BPF_DW) &&
>> +                   (insn->src_reg == BPF_PSEUDO_MAP_VALUE ||
>> +                    insn->src_reg == BPF_PSEUDO_MAP_IDX_VALUE)) {
> 
> Is there a selftest for BPF_PSEUDO_MAP_IDX_VALUE part ?
> I couldn't find it.

Do you mean add a selftest with 'bpftool gen skeleton -L'?

Indeed, it's better to test it.

Thanks,
Leon

Re: [PATCH bpf-next 4/4] selftests/bpf: Add a case to test global percpu data

[Leon Hwang]

On 8/2/25 03:48, Andrii Nakryiko wrote:
> On Fri, Feb 7, 2025 at 2:00 AM Leon Hwang <leon.hwang@linux.dev> wrote:
>>
>>
>>
>> On 6/2/25 08:09, Andrii Nakryiko wrote:
>>> On Mon, Jan 27, 2025 at 8:22 AM Leon Hwang <leon.hwang@linux.dev> wrote:
>>>>

[...]

>>>> +void test_global_percpu_data_init(void)
>>>> +{
>>>> +       struct test_global_percpu_data *skel = NULL;
>>>> +       u64 *percpu_data = NULL;
>>>
>>> there is that test_global_percpu_data__percpu type you are declaring
>>> in the BPF skeleton, right? We should try using it here.
>>>
>>
>> No. bpftool does not generate test_global_percpu_data__percpu. The
>> struct for global variables is embedded into skeleton struct.
>>
>> Should we generate type for global variables?
> 
> we already have custom skeleton-specific type for .data, .rodata,
> .bss, so we should provide one for .percpu as well, yes
> 

Yes, I've generated it. But it should not add '__aligned(8)' to it. Or
bpf_map__set_initial_value() will fails because the aligned size is
different from the actual spec's value size.

If the actual value size is not __aligned(8), how should we lookup
element from percpu_array map?

The doc[0] does not provide a good practice for this case.

[0] https://docs.kernel.org/bpf/map_array.html#bpf-map-type-percpu-array

>>
>>> And for that array access, we should make sure that it's __aligned(8),
>>> so indexing by CPU index works correctly.
>>>
>>
>> Ack.
>>
>>> Also, you define per-CPU variable as int, but here it is u64, what's
>>> up with that?
>>>
>>
>> Like __aligned(8), it's to make sure 8-bytes aligned. It's better to use
>> __aligned(8).
> 
> It's hacky, and it won't work correctly on big-endian architectures.
> But you shouldn't need that if we have a struct representing this
> .percpu memory image. Just make sure that struct has 8 byte alignment
> (from bpftool side during skeleton generation, probably).
> 
> [...]
> 
>>> at least one of BPF programs (don't remember which one, could be
>>> raw_tp) supports specifying CPU index to run on, it would be nice to
>>> loop over CPUs, triggering BPF program on each one and filling per-CPU
>>> variable with current CPU index. Then we can check that all per-CPU
>>> values have expected values.
>>>
>>
>> Do you mean prog_tests/perf_buffer.c::trigger_on_cpu()?
>>
> 
> No, look at `cpu` field of `struct bpf_test_run_opts`. We should have
> a selftest using it, so you can work backwards from that.
> 

By referencing raw_tp, which uses `opts.cpu`, if use it to test global
percpu data, it will fail to test on non-zero CPU, because
bpf_prog_test_run_skb() disallows setting `opts.cpu`.

Then, when `setaffinity` like perf_buffer.c::trigger_on_cpu(), it's OK
to run the adding selftests on all CPUs.

So, should I use `setaffinity` or change the bpf prog type from tc to
raw_tp to use `opts.cpu`?

Thanks,
Leon

Re: [PATCH bpf-next 4/4] selftests/bpf: Add a case to test global percpu data

[Andrii Nakryiko]

On Mon, Feb 10, 2025 at 1:52 AM Leon Hwang <leon.hwang@linux.dev> wrote:
>
>
>
> On 8/2/25 03:48, Andrii Nakryiko wrote:
> > On Fri, Feb 7, 2025 at 2:00 AM Leon Hwang <leon.hwang@linux.dev> wrote:
> >>
> >>
> >>
> >> On 6/2/25 08:09, Andrii Nakryiko wrote:
> >>> On Mon, Jan 27, 2025 at 8:22 AM Leon Hwang <leon.hwang@linux.dev> wrote:
> >>>>
>
> [...]
>
> >>>> +void test_global_percpu_data_init(void)
> >>>> +{
> >>>> +       struct test_global_percpu_data *skel = NULL;
> >>>> +       u64 *percpu_data = NULL;
> >>>
> >>> there is that test_global_percpu_data__percpu type you are declaring
> >>> in the BPF skeleton, right? We should try using it here.
> >>>
> >>
> >> No. bpftool does not generate test_global_percpu_data__percpu. The
> >> struct for global variables is embedded into skeleton struct.
> >>
> >> Should we generate type for global variables?
> >
> > we already have custom skeleton-specific type for .data, .rodata,
> > .bss, so we should provide one for .percpu as well, yes
> >
>
> Yes, I've generated it. But it should not add '__aligned(8)' to it. Or
> bpf_map__set_initial_value() will fails because the aligned size is
> different from the actual spec's value size.
>
> If the actual value size is not __aligned(8), how should we lookup
> element from percpu_array map?

for .percpu libbpf can ensure that map is created with correct value
size that matches __aligned(8) size? It's an error-prone corner case
to non-multiple-of-8 size anyways (for per-CPU data), so just prevent
the issue altogether?...

>
> The doc[0] does not provide a good practice for this case.
>
> [0] https://docs.kernel.org/bpf/map_array.html#bpf-map-type-percpu-array
>
> >>
> >>> And for that array access, we should make sure that it's __aligned(8),
> >>> so indexing by CPU index works correctly.
> >>>
> >>
> >> Ack.
> >>
> >>> Also, you define per-CPU variable as int, but here it is u64, what's
> >>> up with that?
> >>>
> >>
> >> Like __aligned(8), it's to make sure 8-bytes aligned. It's better to use
> >> __aligned(8).
> >
> > It's hacky, and it won't work correctly on big-endian architectures.
> > But you shouldn't need that if we have a struct representing this
> > .percpu memory image. Just make sure that struct has 8 byte alignment
> > (from bpftool side during skeleton generation, probably).
> >
> > [...]
> >
> >>> at least one of BPF programs (don't remember which one, could be
> >>> raw_tp) supports specifying CPU index to run on, it would be nice to
> >>> loop over CPUs, triggering BPF program on each one and filling per-CPU
> >>> variable with current CPU index. Then we can check that all per-CPU
> >>> values have expected values.
> >>>
> >>
> >> Do you mean prog_tests/perf_buffer.c::trigger_on_cpu()?
> >>
> >
> > No, look at `cpu` field of `struct bpf_test_run_opts`. We should have
> > a selftest using it, so you can work backwards from that.
> >
>
> By referencing raw_tp, which uses `opts.cpu`, if use it to test global
> percpu data, it will fail to test on non-zero CPU, because
> bpf_prog_test_run_skb() disallows setting `opts.cpu`.
>
> Then, when `setaffinity` like perf_buffer.c::trigger_on_cpu(), it's OK
> to run the adding selftests on all CPUs.
>
> So, should I use `setaffinity` or change the bpf prog type from tc to
> raw_tp to use `opts.cpu`?

Is it a problem to use raw_tp (we do it a lot)? If not, I'd go with
raw_tp and opts.cpu.

>
> Thanks,
> Leon
>

Re: [PATCH bpf-next 4/4] selftests/bpf: Add a case to test global percpu data

[Leon Hwang]

On 11/2/25 08:15, Andrii Nakryiko wrote:
> On Mon, Feb 10, 2025 at 1:52 AM Leon Hwang <leon.hwang@linux.dev> wrote:
>>
>>
>>
>> On 8/2/25 03:48, Andrii Nakryiko wrote:
>>> On Fri, Feb 7, 2025 at 2:00 AM Leon Hwang <leon.hwang@linux.dev> wrote:

[...]

>>
>> Yes, I've generated it. But it should not add '__aligned(8)' to it. Or
>> bpf_map__set_initial_value() will fails because the aligned size is
>> different from the actual spec's value size.
>>
>> If the actual value size is not __aligned(8), how should we lookup
>> element from percpu_array map?
> 
> for .percpu libbpf can ensure that map is created with correct value
> size that matches __aligned(8) size? It's an error-prone corner case
> to non-multiple-of-8 size anyways (for per-CPU data), so just prevent
> the issue altogether?...
> 

Ack.

I'll update value size of .percpu maps to match __aligned(8) size, and
add '__aligned(8)' to the generated .percpu types.

>>
>> The doc[0] does not provide a good practice for this case.
>>
>> [0] https://docs.kernel.org/bpf/map_array.html#bpf-map-type-percpu-array
>>
>>>>
>>>>> And for that array access, we should make sure that it's __aligned(8),
>>>>> so indexing by CPU index works correctly.
>>>>>
>>>>
>>>> Ack.
>>>>
>>>>> Also, you define per-CPU variable as int, but here it is u64, what's
>>>>> up with that?
>>>>>
>>>>
>>>> Like __aligned(8), it's to make sure 8-bytes aligned. It's better to use
>>>> __aligned(8).
>>>
>>> It's hacky, and it won't work correctly on big-endian architectures.
>>> But you shouldn't need that if we have a struct representing this
>>> .percpu memory image. Just make sure that struct has 8 byte alignment
>>> (from bpftool side during skeleton generation, probably).
>>>
>>> [...]
>>>
>>>>> at least one of BPF programs (don't remember which one, could be
>>>>> raw_tp) supports specifying CPU index to run on, it would be nice to
>>>>> loop over CPUs, triggering BPF program on each one and filling per-CPU
>>>>> variable with current CPU index. Then we can check that all per-CPU
>>>>> values have expected values.
>>>>>
>>>>
>>>> Do you mean prog_tests/perf_buffer.c::trigger_on_cpu()?
>>>>
>>>
>>> No, look at `cpu` field of `struct bpf_test_run_opts`. We should have
>>> a selftest using it, so you can work backwards from that.
>>>
>>
>> By referencing raw_tp, which uses `opts.cpu`, if use it to test global
>> percpu data, it will fail to test on non-zero CPU, because
>> bpf_prog_test_run_skb() disallows setting `opts.cpu`.
>>
>> Then, when `setaffinity` like perf_buffer.c::trigger_on_cpu(), it's OK
>> to run the adding selftests on all CPUs.
>>
>> So, should I use `setaffinity` or change the bpf prog type from tc to
>> raw_tp to use `opts.cpu`?
> 
> Is it a problem to use raw_tp (we do it a lot)? If not, I'd go with
> raw_tp and opts.cpu.
> 

There's no problem to use raw_tp. Let us use raw_tp and opts.cpu.

Thanks,
Leon

[PATCH bpf-next 2/4] bpf, libbpf: Support global percpu data

[Leon Hwang]

This patch introduces support for global percpu data in libbpf by adding a
new ".percpu" section, similar to ".data". It enables efficient handling of
percpu global variables in bpf programs.

Key changes:

* Introduces the ".percpu" section in libbpf.
* Correct value size to __aligned(8) of ".percpu" map definition and btf.
* Creates internal maps for percpu data.
* Initializes and populates these maps accordingly.

This enhancement improves performance for workloads that benefit from
percpu storage.

Meanwhile, add bpf_map__is_internal_percpu() API to check whether the map
is an internal map used for global percpu variables.

Signed-off-by: Leon Hwang <leon.hwang@linux.dev>
---
 tools/lib/bpf/libbpf.c   | 101 +++++++++++++++++++++++++++++++--------
 tools/lib/bpf/libbpf.h   |   9 ++++
 tools/lib/bpf/libbpf.map |   1 +
 3 files changed, 90 insertions(+), 21 deletions(-)

diff --git a/tools/lib/bpf/libbpf.c b/tools/lib/bpf/libbpf.c
index 194809da51725..736a902a667e9 100644
--- a/tools/lib/bpf/libbpf.c
+++ b/tools/lib/bpf/libbpf.c
@@ -516,6 +516,7 @@ struct bpf_struct_ops {
 };
 
 #define DATA_SEC ".data"
+#define PERCPU_SEC ".percpu"
 #define BSS_SEC ".bss"
 #define RODATA_SEC ".rodata"
 #define KCONFIG_SEC ".kconfig"
@@ -530,6 +531,7 @@ enum libbpf_map_type {
 	LIBBPF_MAP_BSS,
 	LIBBPF_MAP_RODATA,
 	LIBBPF_MAP_KCONFIG,
+	LIBBPF_MAP_PERCPU,
 };
 
 struct bpf_map_def {
@@ -640,6 +642,7 @@ enum sec_type {
 	SEC_DATA,
 	SEC_RODATA,
 	SEC_ST_OPS,
+	SEC_PERCPU,
 };
 
 struct elf_sec_desc {
@@ -1903,7 +1906,7 @@ static bool map_is_mmapable(struct bpf_object *obj, struct bpf_map *map)
 	struct btf_var_secinfo *vsi;
 	int i, n;
 
-	if (!map->btf_value_type_id)
+	if (!map->btf_value_type_id || map->libbpf_type == LIBBPF_MAP_PERCPU)
 		return false;
 
 	t = btf__type_by_id(obj->btf, map->btf_value_type_id);
@@ -1927,6 +1930,7 @@ static int
 bpf_object__init_internal_map(struct bpf_object *obj, enum libbpf_map_type type,
 			      const char *real_name, int sec_idx, void *data, size_t data_sz)
 {
+	bool is_percpu = type == LIBBPF_MAP_PERCPU;
 	struct bpf_map_def *def;
 	struct bpf_map *map;
 	size_t mmap_sz;
@@ -1948,9 +1952,9 @@ bpf_object__init_internal_map(struct bpf_object *obj, enum libbpf_map_type type,
 	}
 
 	def = &map->def;
-	def->type = BPF_MAP_TYPE_ARRAY;
+	def->type = is_percpu ? BPF_MAP_TYPE_PERCPU_ARRAY : BPF_MAP_TYPE_ARRAY;
 	def->key_size = sizeof(int);
-	def->value_size = data_sz;
+	def->value_size = is_percpu ? roundup(data_sz, 8) : data_sz;
 	def->max_entries = 1;
 	def->map_flags = type == LIBBPF_MAP_RODATA || type == LIBBPF_MAP_KCONFIG
 		? BPF_F_RDONLY_PROG : 0;
@@ -1961,10 +1965,11 @@ bpf_object__init_internal_map(struct bpf_object *obj, enum libbpf_map_type type,
 	if (map_is_mmapable(obj, map))
 		def->map_flags |= BPF_F_MMAPABLE;
 
-	pr_debug("map '%s' (global data): at sec_idx %d, offset %zu, flags %x.\n",
-		 map->name, map->sec_idx, map->sec_offset, def->map_flags);
+	pr_debug("map '%s' (global %sdata): at sec_idx %d, offset %zu, flags %x.\n",
+		 map->name, is_percpu ? "percpu " : "", map->sec_idx,
+		 map->sec_offset, def->map_flags);
 
-	mmap_sz = bpf_map_mmap_sz(map);
+	mmap_sz = is_percpu ? def->value_size : bpf_map_mmap_sz(map);
 	map->mmaped = mmap(NULL, mmap_sz, PROT_READ | PROT_WRITE,
 			   MAP_SHARED | MAP_ANONYMOUS, -1, 0);
 	if (map->mmaped == MAP_FAILED) {
@@ -2015,6 +2020,13 @@ static int bpf_object__init_global_data_maps(struct bpf_object *obj)
 							    sec_desc->data->d_buf,
 							    sec_desc->data->d_size);
 			break;
+		case SEC_PERCPU:
+			sec_name = elf_sec_name(obj, elf_sec_by_idx(obj, sec_idx));
+			err = bpf_object__init_internal_map(obj, LIBBPF_MAP_PERCPU,
+							    sec_name, sec_idx,
+							    sec_desc->data->d_buf,
+							    sec_desc->data->d_size);
+			break;
 		case SEC_BSS:
 			sec_name = elf_sec_name(obj, elf_sec_by_idx(obj, sec_idx));
 			err = bpf_object__init_internal_map(obj, LIBBPF_MAP_BSS,
@@ -3364,6 +3376,10 @@ static int btf_fixup_datasec(struct bpf_object *obj, struct btf *btf,
 		fixup_offsets = true;
 	}
 
+	/* .percpu DATASEC must has __aligned(8) size. */
+	if (strcmp(sec_name, PERCPU_SEC) == 0 || str_has_pfx(sec_name, PERCPU_SEC))
+		t->size = roundup(t->size, 8);
+
 	for (i = 0, vsi = btf_var_secinfos(t); i < vars; i++, vsi++) {
 		const struct btf_type *t_var;
 		struct btf_var *var;
@@ -3934,6 +3950,11 @@ static int bpf_object__elf_collect(struct bpf_object *obj)
 				sec_desc->sec_type = SEC_RODATA;
 				sec_desc->shdr = sh;
 				sec_desc->data = data;
+			} else if (strcmp(name, PERCPU_SEC) == 0 ||
+				   str_has_pfx(name, PERCPU_SEC)) {
+				sec_desc->sec_type = SEC_PERCPU;
+				sec_desc->shdr = sh;
+				sec_desc->data = data;
 			} else if (strcmp(name, STRUCT_OPS_SEC) == 0 ||
 				   strcmp(name, STRUCT_OPS_LINK_SEC) == 0 ||
 				   strcmp(name, "?" STRUCT_OPS_SEC) == 0 ||
@@ -4453,6 +4474,7 @@ static bool bpf_object__shndx_is_data(const struct bpf_object *obj,
 	case SEC_BSS:
 	case SEC_DATA:
 	case SEC_RODATA:
+	case SEC_PERCPU:
 		return true;
 	default:
 		return false;
@@ -4478,6 +4500,8 @@ bpf_object__section_to_libbpf_map_type(const struct bpf_object *obj, int shndx)
 		return LIBBPF_MAP_DATA;
 	case SEC_RODATA:
 		return LIBBPF_MAP_RODATA;
+	case SEC_PERCPU:
+		return LIBBPF_MAP_PERCPU;
 	default:
 		return LIBBPF_MAP_UNSPEC;
 	}
@@ -4795,7 +4819,7 @@ static int map_fill_btf_type_info(struct bpf_object *obj, struct bpf_map *map)
 
 	/*
 	 * LLVM annotates global data differently in BTF, that is,
-	 * only as '.data', '.bss' or '.rodata'.
+	 * only as '.data', '.bss', '.percpu' or '.rodata'.
 	 */
 	if (!bpf_map__is_internal(map))
 		return -ENOENT;
@@ -5125,23 +5149,47 @@ static int
 bpf_object__populate_internal_map(struct bpf_object *obj, struct bpf_map *map)
 {
 	enum libbpf_map_type map_type = map->libbpf_type;
-	int err, zero = 0;
-	size_t mmap_sz;
+	bool is_percpu = map_type == LIBBPF_MAP_PERCPU;
+	int err = 0, zero = 0, num_cpus, i;
+	size_t data_sz, elem_sz, mmap_sz;
+	void *data = NULL;
+
+	data_sz = map->def.value_size;
+	if (is_percpu) {
+		num_cpus = libbpf_num_possible_cpus();
+		if (num_cpus < 0) {
+			err = num_cpus;
+			return err;
+		}
+
+		data_sz = data_sz * num_cpus;
+		data = malloc(data_sz);
+		if (!data) {
+			err = -ENOMEM;
+			return err;
+		}
+
+		elem_sz = map->def.value_size;
+		for (i = 0; i < num_cpus; i++)
+			memcpy(data + i * elem_sz, map->mmaped, elem_sz);
+	} else {
+		data = map->mmaped;
+	}
 
 	if (obj->gen_loader) {
 		bpf_gen__map_update_elem(obj->gen_loader, map - obj->maps,
-					 map->mmaped, map->def.value_size);
+					 data, data_sz);
 		if (map_type == LIBBPF_MAP_RODATA || map_type == LIBBPF_MAP_KCONFIG)
 			bpf_gen__map_freeze(obj->gen_loader, map - obj->maps);
-		return 0;
+		goto free_data;
 	}
 
-	err = bpf_map_update_elem(map->fd, &zero, map->mmaped, 0);
+	err = bpf_map_update_elem(map->fd, &zero, data, 0);
 	if (err) {
 		err = -errno;
 		pr_warn("map '%s': failed to set initial contents: %s\n",
 			bpf_map__name(map), errstr(err));
-		return err;
+		goto free_data;
 	}
 
 	/* Freeze .rodata and .kconfig map as read-only from syscall side. */
@@ -5151,7 +5199,7 @@ bpf_object__populate_internal_map(struct bpf_object *obj, struct bpf_map *map)
 			err = -errno;
 			pr_warn("map '%s': failed to freeze as read-only: %s\n",
 				bpf_map__name(map), errstr(err));
-			return err;
+			goto free_data;
 		}
 	}
 
@@ -5178,7 +5226,7 @@ bpf_object__populate_internal_map(struct bpf_object *obj, struct bpf_map *map)
 			err = -errno;
 			pr_warn("map '%s': failed to re-mmap() contents: %s\n",
 				bpf_map__name(map), errstr(err));
-			return err;
+			goto free_data;
 		}
 		map->mmaped = mmaped;
 	} else if (map->mmaped) {
@@ -5186,7 +5234,10 @@ bpf_object__populate_internal_map(struct bpf_object *obj, struct bpf_map *map)
 		map->mmaped = NULL;
 	}
 
-	return 0;
+free_data:
+	if (is_percpu)
+		free(data);
+	return err;
 }
 
 static void bpf_map__destroy(struct bpf_map *map);
@@ -10132,14 +10183,17 @@ int bpf_map__fd(const struct bpf_map *map)
 
 static bool map_uses_real_name(const struct bpf_map *map)
 {
-	/* Since libbpf started to support custom .data.* and .rodata.* maps,
-	 * their user-visible name differs from kernel-visible name. Users see
-	 * such map's corresponding ELF section name as a map name.
-	 * This check distinguishes .data/.rodata from .data.* and .rodata.*
-	 * maps to know which name has to be returned to the user.
+	/* Since libbpf started to support custom .data.*, .percpu.* and
+	 * .rodata.* maps, their user-visible name differs from kernel-visible
+	 * name. Users see such map's corresponding ELF section name as a map
+	 * name. This check distinguishes .data/.percpu/.rodata from .data.*,
+	 * .percpu.* and .rodata.* maps to know which name has to be returned to
+	 * the user.
 	 */
 	if (map->libbpf_type == LIBBPF_MAP_DATA && strcmp(map->real_name, DATA_SEC) != 0)
 		return true;
+	if (map->libbpf_type == LIBBPF_MAP_PERCPU && strcmp(map->real_name, PERCPU_SEC) != 0)
+		return true;
 	if (map->libbpf_type == LIBBPF_MAP_RODATA && strcmp(map->real_name, RODATA_SEC) != 0)
 		return true;
 	return false;
@@ -10386,6 +10440,11 @@ bool bpf_map__is_internal(const struct bpf_map *map)
 	return map->libbpf_type != LIBBPF_MAP_UNSPEC;
 }
 
+bool bpf_map__is_internal_percpu(const struct bpf_map *map)
+{
+	return map->libbpf_type == LIBBPF_MAP_PERCPU;
+}
+
 __u32 bpf_map__ifindex(const struct bpf_map *map)
 {
 	return map->map_ifindex;
diff --git a/tools/lib/bpf/libbpf.h b/tools/lib/bpf/libbpf.h
index 3020ee45303a0..1d8ca33d370d1 100644
--- a/tools/lib/bpf/libbpf.h
+++ b/tools/lib/bpf/libbpf.h
@@ -1072,6 +1072,15 @@ LIBBPF_API void *bpf_map__initial_value(const struct bpf_map *map, size_t *psize
  */
 LIBBPF_API bool bpf_map__is_internal(const struct bpf_map *map);
 
+/**
+ * @brief **bpf_map__is_internal_percpu()** tells the caller whether or not
+ * the passed map is an internal map used for global percpu variables.
+ * @param map the bpf_map
+ * @return true, if the map is an internal map used for global percpu
+ * variables; false, otherwise
+ */
+LIBBPF_API bool bpf_map__is_internal_percpu(const struct bpf_map *map);
+
 /**
  * @brief **bpf_map__set_pin_path()** sets the path attribute that tells where the
  * BPF map should be pinned. This does not actually create the 'pin'.
diff --git a/tools/lib/bpf/libbpf.map b/tools/lib/bpf/libbpf.map
index b5a838de6f47c..09cdbd6e32218 100644
--- a/tools/lib/bpf/libbpf.map
+++ b/tools/lib/bpf/libbpf.map
@@ -438,4 +438,5 @@ LIBBPF_1.6.0 {
 		bpf_linker__new_fd;
 		btf__add_decl_attr;
 		btf__add_type_attr;
+		bpf_map__is_internal_percpu;
 } LIBBPF_1.5.0;
-- 
2.47.1