AuxXxilium/linux_dsm_epyc7002
1
0
Fork 0
mirror of https://github.com/AuxXxilium/linux_dsm_epyc7002.git synced 2024-11-23 23:40:53 +07:00
Code Issues Actions Packages Projects Releases Wiki Activity

libbpf: Add BTF-defined map-in-map support

Browse Source 
As discussed at LPC 2019 ([0]), this patch brings (a quite belated) support
for declarative BTF-defined map-in-map support in libbpf. It allows to define
ARRAY_OF_MAPS and HASH_OF_MAPS BPF maps without any user-space initialization
code involved.

Additionally, it allows to initialize outer map's slots with references to
respective inner maps at load time, also completely declaratively.

Despite a weak type system of C, the way BTF-defined map-in-map definition
works, it's actually quite hard to accidentally initialize outer map with
incompatible inner maps. This being C, of course, it's still possible, but
even that would be caught at load time and error returned with helpful debug
log pointing exactly to the slot that failed to be initialized.

As an example, here's a rather advanced HASH_OF_MAPS declaration and
initialization example, filling slots #0 and #4 with two inner maps:

  #include <bpf/bpf_helpers.h>

  struct inner_map {
          __uint(type, BPF_MAP_TYPE_ARRAY);
          __uint(max_entries, 1);
          __type(key, int);
          __type(value, int);
  } inner_map1 SEC(".maps"),
    inner_map2 SEC(".maps");

  struct outer_hash {
          __uint(type, BPF_MAP_TYPE_HASH_OF_MAPS);
          __uint(max_entries, 5);
          __uint(key_size, sizeof(int));
          __array(values, struct inner_map);
  } outer_hash SEC(".maps") = {
          .values = {
                  [0] = &inner_map2,
                  [4] = &inner_map1,
          },
  };

Here's the relevant part of libbpf debug log showing pretty clearly of what's
going on with map-in-map initialization:

  libbpf: .maps relo #0: for 6 value 0 rel.r_offset 96 name 260 ('inner_map1')
  libbpf: .maps relo #0: map 'outer_arr' slot [0] points to map 'inner_map1'
  libbpf: .maps relo #1: for 7 value 32 rel.r_offset 112 name 249 ('inner_map2')
  libbpf: .maps relo #1: map 'outer_arr' slot [2] points to map 'inner_map2'
  libbpf: .maps relo #2: for 7 value 32 rel.r_offset 144 name 249 ('inner_map2')
  libbpf: .maps relo #2: map 'outer_hash' slot [0] points to map 'inner_map2'
  libbpf: .maps relo #3: for 6 value 0 rel.r_offset 176 name 260 ('inner_map1')
  libbpf: .maps relo #3: map 'outer_hash' slot [4] points to map 'inner_map1'
  libbpf: map 'inner_map1': created successfully, fd=4
  libbpf: map 'inner_map2': created successfully, fd=5
  libbpf: map 'outer_hash': created successfully, fd=7
  libbpf: map 'outer_hash': slot [0] set to map 'inner_map2' fd=5
  libbpf: map 'outer_hash': slot [4] set to map 'inner_map1' fd=4

Notice from the log above that fd=6 (not logged explicitly) is used for inner
"prototype" map, necessary for creation of outer map. It is destroyed
immediately after outer map is created.

See also included selftest with some extra comments explaining extra details
of usage. Additionally, similar initialization syntax and libbpf functionality
can be used to do initialization of BPF_PROG_ARRAY with references to BPF
sub-programs. This can be done in follow up patches, if there will be a demand
for this.

  [0] https://linuxplumbersconf.org/event/4/contributions/448/

Signed-off-by: Andrii Nakryiko <andriin@fb.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Acked-by: Toke Høiland-Jørgensen <toke@redhat.com>
Link: https://lore.kernel.org/bpf/20200429002739.48006-4-andriin@fb.com
This commit is contained in:
Andrii Nakryiko 2020-04-28 17:27:39 -07:00 committed by Alexei Starovoitov
parent 2d39d7c56f
commit 646f02ffdd
4 changed files with 384 additions and 23 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

1
tools/lib/bpf/bpf_helpers.h
View File

@ -12,6 +12,7 @@
#define __uint(name, val) int (*name)[val]
#define __type(name, val) typeof(val) *name
#define __array(name, val) typeof(val) *name[]
/* Helper macro to print out debug messages */
#define bpf_printk(fmt, ...) \

281
tools/lib/bpf/libbpf.c
View File

@ -310,6 +310,7 @@ struct bpf_map {
int map_ifindex;
int inner_map_fd;
struct bpf_map_def def;
__u32 btf_var_idx;
__u32 btf_key_type_id;
__u32 btf_value_type_id;
__u32 btf_vmlinux_value_type_id;
@ -318,6 +319,9 @@ struct bpf_map {
enum libbpf_map_type libbpf_type;
void *mmaped;
struct bpf_struct_ops *st_ops;
struct bpf_map *inner_map;
void **init_slots;
int init_slots_sz;
char *pin_path;
bool pinned;
bool reused;
@ -389,6 +393,7 @@ struct bpf_object {
int nr_reloc_sects;
int maps_shndx;
int btf_maps_shndx;
__u32 btf_maps_sec_btf_id;
int text_shndx;
int symbols_shndx;
int data_shndx;
@ -1918,7 +1923,7 @@ static int build_map_pin_path(struct bpf_map *map, const char *path)
static int parse_btf_map_def(struct bpf_object *obj,
struct bpf_map *map,
const struct btf_type *def,
bool strict,
bool strict, bool is_inner,
const char *pin_root_path)
{
const struct btf_type *t;
@ -2036,10 +2041,79 @@ static int parse_btf_map_def(struct bpf_object *obj,
}
map->def.value_size = sz;
map->btf_value_type_id = t->type;
}
else if (strcmp(name, "values") == 0) {
int err;
if (is_inner) {
pr_warn("map '%s': multi-level inner maps not supported.\n",
map->name);
return -ENOTSUP;
}
if (i != vlen - 1) {
pr_warn("map '%s': '%s' member should be last.\n",
map->name, name);
return -EINVAL;
}
if (!bpf_map_type__is_map_in_map(map->def.type)) {
pr_warn("map '%s': should be map-in-map.\n",
map->name);
return -ENOTSUP;
}
if (map->def.value_size && map->def.value_size != 4) {
pr_warn("map '%s': conflicting value size %u != 4.\n",
map->name, map->def.value_size);
return -EINVAL;
}
map->def.value_size = 4;
t = btf__type_by_id(obj->btf, m->type);
if (!t) {
pr_warn("map '%s': map-in-map inner type [%d] not found.\n",
map->name, m->type);
return -EINVAL;
}
if (!btf_is_array(t) || btf_array(t)->nelems) {
pr_warn("map '%s': map-in-map inner spec is not a zero-sized array.\n",
map->name);
return -EINVAL;
}
t = skip_mods_and_typedefs(obj->btf, btf_array(t)->type,
NULL);
if (!btf_is_ptr(t)) {
pr_warn("map '%s': map-in-map inner def is of unexpected kind %u.\n",
map->name, btf_kind(t));
return -EINVAL;
}
t = skip_mods_and_typedefs(obj->btf, t->type, NULL);
if (!btf_is_struct(t)) {
pr_warn("map '%s': map-in-map inner def is of unexpected kind %u.\n",
map->name, btf_kind(t));
return -EINVAL;
}
map->inner_map = calloc(1, sizeof(*map->inner_map));
if (!map->inner_map)
return -ENOMEM;
map->inner_map->sec_idx = obj->efile.btf_maps_shndx;
map->inner_map->name = malloc(strlen(map->name) +
sizeof(".inner") + 1);
if (!map->inner_map->name)
return -ENOMEM;
sprintf(map->inner_map->name, "%s.inner", map->name);
err = parse_btf_map_def(obj, map->inner_map, t, strict,
true /* is_inner */, NULL);
if (err)
return err;
} else if (strcmp(name, "pinning") == 0) {
__u32 val;
int err;
if (is_inner) {
pr_debug("map '%s': inner def can't be pinned.\n",
map->name);
return -EINVAL;
}
if (!get_map_field_int(map->name, obj->btf, m, &val))
return -EINVAL;
pr_debug("map '%s': found pinning = %u.\n",
@ -2138,10 +2212,11 @@ static int bpf_object__init_user_btf_map(struct bpf_object *obj,
map->def.type = BPF_MAP_TYPE_UNSPEC;
map->sec_idx = sec_idx;
map->sec_offset = vi->offset;
map->btf_var_idx = var_idx;
pr_debug("map '%s': at sec_idx %d, offset %zu.\n",
map_name, map->sec_idx, map->sec_offset);
return parse_btf_map_def(obj, map, def, strict, pin_root_path);
return parse_btf_map_def(obj, map, def, strict, false, pin_root_path);
}
static int bpf_object__init_user_btf_maps(struct bpf_object *obj, bool strict,
@ -2174,6 +2249,7 @@ static int bpf_object__init_user_btf_maps(struct bpf_object *obj, bool strict,
name = btf__name_by_offset(obj->btf, t->name_off);
if (strcmp(name, MAPS_ELF_SEC) == 0) {
sec = t;
obj->efile.btf_maps_sec_btf_id = i;
break;
}
}
@ -2560,7 +2636,8 @@ static int bpf_object__elf_collect(struct bpf_object *obj)
/* Only do relo for section with exec instructions */
if (!section_have_execinstr(obj, sec) &&
strcmp(name, ".rel" STRUCT_OPS_SEC)) {
strcmp(name, ".rel" STRUCT_OPS_SEC) &&
strcmp(name, ".rel" MAPS_ELF_SEC)) {
pr_debug("skip relo %s(%d) for section(%d)\n",
name, idx, sec);
continue;
@ -3538,6 +3615,22 @@ static int bpf_object__create_map(struct bpf_object *obj, struct bpf_map *map)
create_attr.btf_value_type_id = map->btf_value_type_id;
}
if (bpf_map_type__is_map_in_map(def->type)) {
if (map->inner_map) {
int err;
err = bpf_object__create_map(obj, map->inner_map);
if (err) {
pr_warn("map '%s': failed to create inner map: %d\n",
map->name, err);
return err;
}
map->inner_map_fd = bpf_map__fd(map->inner_map);
}
if (map->inner_map_fd >= 0)
create_attr.inner_map_fd = map->inner_map_fd;
}
map->fd = bpf_create_map_xattr(&create_attr);
if (map->fd < 0 && (create_attr.btf_key_type_id ||
create_attr.btf_value_type_id)) {
@ -3558,6 +3651,11 @@ static int bpf_object__create_map(struct bpf_object *obj, struct bpf_map *map)
if (map->fd < 0)
return -errno;
if (bpf_map_type__is_map_in_map(def->type) && map->inner_map) {
bpf_map__destroy(map->inner_map);
zfree(&map->inner_map);
}
return 0;
}
@ -3602,6 +3700,31 @@ bpf_object__create_maps(struct bpf_object *obj)
}
}
if (map->init_slots_sz) {
for (j = 0; j < map->init_slots_sz; j++) {
const struct bpf_map *targ_map;
int fd;
if (!map->init_slots[j])
continue;
targ_map = map->init_slots[j];
fd = bpf_map__fd(targ_map);
err = bpf_map_update_elem(map->fd, &j, &fd, 0);
if (err) {
err = -errno;
pr_warn("map '%s': failed to initialize slot [%d] to map '%s' fd=%d: %d\n",
map->name, j, targ_map->name,
fd, err);
goto err_out;
}
pr_debug("map '%s': slot [%d] set to map '%s' fd=%d\n",
map->name, j, targ_map->name, fd);
}
zfree(&map->init_slots);
map->init_slots_sz = 0;
}
if (map->pin_path && !map->pinned) {
err = bpf_map__pin(map, NULL);
if (err) {
@ -4873,9 +4996,118 @@ bpf_object__relocate(struct bpf_object *obj, const char *targ_btf_path)
return 0;
}
static int bpf_object__collect_struct_ops_map_reloc(struct bpf_object *obj,
GElf_Shdr *shdr,
Elf_Data *data);
static int bpf_object__collect_st_ops_relos(struct bpf_object *obj,
GElf_Shdr *shdr, Elf_Data *data);
static int bpf_object__collect_map_relos(struct bpf_object *obj,
GElf_Shdr *shdr, Elf_Data *data)
{
int i, j, nrels, new_sz, ptr_sz = sizeof(void *);
const struct btf_type *sec, *var, *def;
const struct btf_var_secinfo *vi;
const struct btf_member *member;
struct bpf_map *map, *targ_map;
const char *name, *mname;
Elf_Data *symbols;
unsigned int moff;
GElf_Sym sym;
GElf_Rel rel;
void *tmp;
if (!obj->efile.btf_maps_sec_btf_id || !obj->btf)
return -EINVAL;
sec = btf__type_by_id(obj->btf, obj->efile.btf_maps_sec_btf_id);
if (!sec)
return -EINVAL;
symbols = obj->efile.symbols;
nrels = shdr->sh_size / shdr->sh_entsize;
for (i = 0; i < nrels; i++) {
if (!gelf_getrel(data, i, &rel)) {
pr_warn(".maps relo #%d: failed to get ELF relo\n", i);
return -LIBBPF_ERRNO__FORMAT;
}
if (!gelf_getsym(symbols, GELF_R_SYM(rel.r_info), &sym)) {
pr_warn(".maps relo #%d: symbol %zx not found\n",
i, (size_t)GELF_R_SYM(rel.r_info));
return -LIBBPF_ERRNO__FORMAT;
}
name = elf_strptr(obj->efile.elf, obj->efile.strtabidx,
sym.st_name) ? : "<?>";
if (sym.st_shndx != obj->efile.btf_maps_shndx) {
pr_warn(".maps relo #%d: '%s' isn't a BTF-defined map\n",
i, name);
return -LIBBPF_ERRNO__RELOC;
}
pr_debug(".maps relo #%d: for %zd value %zd rel.r_offset %zu name %d ('%s')\n",
i, (ssize_t)(rel.r_info >> 32), (size_t)sym.st_value,
(size_t)rel.r_offset, sym.st_name, name);
for (j = 0; j < obj->nr_maps; j++) {
map = &obj->maps[j];
if (map->sec_idx != obj->efile.btf_maps_shndx)
continue;
vi = btf_var_secinfos(sec) + map->btf_var_idx;
if (vi->offset <= rel.r_offset &&
rel.r_offset + sizeof(void *) <= vi->offset + vi->size)
break;
}
if (j == obj->nr_maps) {
pr_warn(".maps relo #%d: cannot find map '%s' at rel.r_offset %zu\n",
i, name, (size_t)rel.r_offset);
return -EINVAL;
}
if (!bpf_map_type__is_map_in_map(map->def.type))
return -EINVAL;
if (map->def.type == BPF_MAP_TYPE_HASH_OF_MAPS &&
map->def.key_size != sizeof(int)) {
pr_warn(".maps relo #%d: hash-of-maps '%s' should have key size %zu.\n",
i, map->name, sizeof(int));
return -EINVAL;
}
targ_map = bpf_object__find_map_by_name(obj, name);
if (!targ_map)
return -ESRCH;
var = btf__type_by_id(obj->btf, vi->type);
def = skip_mods_and_typedefs(obj->btf, var->type, NULL);
if (btf_vlen(def) == 0)
return -EINVAL;
member = btf_members(def) + btf_vlen(def) - 1;
mname = btf__name_by_offset(obj->btf, member->name_off);
if (strcmp(mname, "values"))
return -EINVAL;
moff = btf_member_bit_offset(def, btf_vlen(def) - 1) / 8;
if (rel.r_offset - vi->offset < moff)
return -EINVAL;
moff = rel.r_offset - vi->offset - moff;
if (moff % ptr_sz)
return -EINVAL;
moff /= ptr_sz;
if (moff >= map->init_slots_sz) {
new_sz = moff + 1;
tmp = realloc(map->init_slots, new_sz * ptr_sz);
if (!tmp)
return -ENOMEM;
map->init_slots = tmp;
memset(map->init_slots + map->init_slots_sz, 0,
(new_sz - map->init_slots_sz) * ptr_sz);
map->init_slots_sz = new_sz;
}
map->init_slots[moff] = targ_map;
pr_debug(".maps relo #%d: map '%s' slot [%d] points to map '%s'\n",
i, map->name, moff, name);
}
return 0;
}
static int bpf_object__collect_reloc(struct bpf_object *obj)
{
@ -4898,21 +5130,17 @@ static int bpf_object__collect_reloc(struct bpf_object *obj)
}
if (idx == obj->efile.st_ops_shndx) {
err = bpf_object__collect_struct_ops_map_reloc(obj,
shdr,
data);
if (err)
return err;
continue;
err = bpf_object__collect_st_ops_relos(obj, shdr, data);
} else if (idx == obj->efile.btf_maps_shndx) {
err = bpf_object__collect_map_relos(obj, shdr, data);
} else {
prog = bpf_object__find_prog_by_idx(obj, idx);
if (!prog) {
pr_warn("relocation failed: no prog in section(%d)\n", idx);
return -LIBBPF_ERRNO__RELOC;
}
err = bpf_program__collect_reloc(prog, shdr, data, obj);
}
prog = bpf_object__find_prog_by_idx(obj, idx);
if (!prog) {
pr_warn("relocation failed: no section(%d)\n", idx);
return -LIBBPF_ERRNO__RELOC;
}
err = bpf_program__collect_reloc(prog, shdr, data, obj);
if (err)
return err;
}
@ -5984,6 +6212,14 @@ static void bpf_map__destroy(struct bpf_map *map)
map->priv = NULL;
map->clear_priv = NULL;
if (map->inner_map) {
bpf_map__destroy(map->inner_map);
zfree(&map->inner_map);
}
zfree(&map->init_slots);
map->init_slots_sz = 0;
if (map->mmaped) {
munmap(map->mmaped, bpf_map_mmap_sz(map));
map->mmaped = NULL;
@ -6543,9 +6779,8 @@ static struct bpf_map *find_struct_ops_map_by_offset(struct bpf_object *obj,
}
/* Collect the reloc from ELF and populate the st_ops->progs[] */
static int bpf_object__collect_struct_ops_map_reloc(struct bpf_object *obj,
GElf_Shdr *shdr,
Elf_Data *data)
static int bpf_object__collect_st_ops_relos(struct bpf_object *obj,
GElf_Shdr *shdr, Elf_Data *data)
{
const struct btf_member *member;
struct bpf_struct_ops *st_ops;

49
tools/testing/selftests/bpf/prog_tests/btf_map_in_map.c Normal file
View File

@ -0,0 +1,49 @@
// SPDX-License-Identifier: GPL-2.0
/* Copyright (c) 2020 Facebook */
#include <test_progs.h>
#include "test_btf_map_in_map.skel.h"
void test_btf_map_in_map(void)
{
int duration = 0, err, key = 0, val;
struct test_btf_map_in_map* skel;
skel = test_btf_map_in_map__open_and_load();
if (CHECK(!skel, "skel_open", "failed to open&load skeleton\n"))
return;
err = test_btf_map_in_map__attach(skel);
if (CHECK(err, "skel_attach", "skeleton attach failed: %d\n", err))
goto cleanup;
/* inner1 = input, inner2 = input + 1 */
val = bpf_map__fd(skel->maps.inner_map1);
bpf_map_update_elem(bpf_map__fd(skel->maps.outer_arr), &key, &val, 0);
val = bpf_map__fd(skel->maps.inner_map2);
bpf_map_update_elem(bpf_map__fd(skel->maps.outer_hash), &key, &val, 0);
skel->bss->input = 1;
usleep(1);
bpf_map_lookup_elem(bpf_map__fd(skel->maps.inner_map1), &key, &val);
CHECK(val != 1, "inner1", "got %d != exp %d\n", val, 1);
bpf_map_lookup_elem(bpf_map__fd(skel->maps.inner_map2), &key, &val);
CHECK(val != 2, "inner2", "got %d != exp %d\n", val, 2);
/* inner1 = input + 1, inner2 = input */
val = bpf_map__fd(skel->maps.inner_map2);
bpf_map_update_elem(bpf_map__fd(skel->maps.outer_arr), &key, &val, 0);
val = bpf_map__fd(skel->maps.inner_map1);
bpf_map_update_elem(bpf_map__fd(skel->maps.outer_hash), &key, &val, 0);
skel->bss->input = 3;
usleep(1);
bpf_map_lookup_elem(bpf_map__fd(skel->maps.inner_map1), &key, &val);
CHECK(val != 4, "inner1", "got %d != exp %d\n", val, 4);
bpf_map_lookup_elem(bpf_map__fd(skel->maps.inner_map2), &key, &val);
CHECK(val != 3, "inner2", "got %d != exp %d\n", val, 3);
cleanup:
test_btf_map_in_map__destroy(skel);
}

76
tools/testing/selftests/bpf/progs/test_btf_map_in_map.c Normal file
View File

@ -0,0 +1,76 @@
/* SPDX-License-Identifier: GPL-2.0 */
/* Copyright (c) 2020 Facebook */
#include <linux/bpf.h>
#include <bpf/bpf_helpers.h>
struct inner_map {
__uint(type, BPF_MAP_TYPE_ARRAY);
__uint(max_entries, 1);
__type(key, int);
__type(value, int);
} inner_map1 SEC(".maps"),
inner_map2 SEC(".maps");
struct outer_arr {
__uint(type, BPF_MAP_TYPE_ARRAY_OF_MAPS);
__uint(max_entries, 3);
__uint(key_size, sizeof(int));
__uint(value_size, sizeof(int));
/* it's possible to use anonymous struct as inner map definition here */
__array(values, struct {
__uint(type, BPF_MAP_TYPE_ARRAY);
/* changing max_entries to 2 will fail during load
* due to incompatibility with inner_map definition */
__uint(max_entries, 1);
__type(key, int);
__type(value, int);
});
} outer_arr SEC(".maps") = {
/* (void *) cast is necessary because we didn't use `struct inner_map`
* in __inner(values, ...)
* Actually, a conscious effort is required to screw up initialization
* of inner map slots, which is a great thing!
*/
.values = { (void *)&inner_map1, 0, (void *)&inner_map2 },
};
struct outer_hash {
__uint(type, BPF_MAP_TYPE_HASH_OF_MAPS);
__uint(max_entries, 5);
__uint(key_size, sizeof(int));
/* Here everything works flawlessly due to reuse of struct inner_map
* and compiler will complain at the attempt to use non-inner_map
* references below. This is great experience.
*/
__array(values, struct inner_map);
} outer_hash SEC(".maps") = {
.values = {
[0] = &inner_map2,
[4] = &inner_map1,
},
};
int input = 0;
SEC("raw_tp/sys_enter")
int handle__sys_enter(void *ctx)
{
struct inner_map *inner_map;
int key = 0, val;
inner_map = bpf_map_lookup_elem(&outer_arr, &key);
if (!inner_map)
return 1;
val = input;
bpf_map_update_elem(inner_map, &key, &val, 0);
inner_map = bpf_map_lookup_elem(&outer_hash, &key);
if (!inner_map)
return 1;
val = input + 1;
bpf_map_update_elem(inner_map, &key, &val, 0);
return 0;
}
char _license[] SEC("license") = "GPL";