bpftool: Ensure bytes_memlock json output is correct
authorChris J Arges <carges@cloudflare.com>
Wed, 9 Mar 2022 21:41:58 +0000 (15:41 -0600)
committerDaniel Borkmann <daniel@iogearbox.net>
Thu, 10 Mar 2022 23:06:11 +0000 (00:06 +0100)
If a BPF map is created over 2^32 the memlock value as displayed in JSON
format will be incorrect. Use atoll instead of atoi so that the correct
number is displayed.

  ```
  $ bpftool map create /sys/fs/bpf/test_bpfmap type hash key 4 \
    value 1024 entries 4194304 name test_bpfmap
  $ bpftool map list
  1: hash  name test_bpfmap  flags 0x0
          key 4B  value 1024B  max_entries 4194304  memlock 4328521728B
  $ sudo bpftool map list -j | jq .[].bytes_memlock
  33554432
  ```

Signed-off-by: Chris J Arges <carges@cloudflare.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Link: https://lore.kernel.org/bpf/b6601087-0b11-33cc-904a-1133d1500a10@cloudflare.com
tools/bpf/bpftool/map.c
tools/bpf/bpftool/prog.c

index e746642..f91d9bf 100644 (file)
@@ -504,7 +504,7 @@ static int show_map_close_json(int fd, struct bpf_map_info *info)
        jsonw_uint_field(json_wtr, "max_entries", info->max_entries);
 
        if (memlock)
-               jsonw_int_field(json_wtr, "bytes_memlock", atoi(memlock));
+               jsonw_int_field(json_wtr, "bytes_memlock", atoll(memlock));
        free(memlock);
 
        if (info->type == BPF_MAP_TYPE_PROG_ARRAY) {
index 8a52eed..bc4e055 100644 (file)
@@ -485,7 +485,7 @@ static void print_prog_json(struct bpf_prog_info *info, int fd)
 
        memlock = get_fdinfo(fd, "memlock");
        if (memlock)
-               jsonw_int_field(json_wtr, "bytes_memlock", atoi(memlock));
+               jsonw_int_field(json_wtr, "bytes_memlock", atoll(memlock));
        free(memlock);
 
        if (info->nr_map_ids)