[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <cc32d3db-60ef-4046-8988-289cd0cc8c26@oracle.com>
Date: Wed, 29 Oct 2025 21:35:37 +0530
From: Harshit Mogalapalli <harshit.m.mogalapalli@...cle.com>
To: Yonghong Song <yonghong.song@...ux.dev>, bpf@...r.kernel.org
Cc: alan.maguire@...cle.com, Quentin Monnet <qmo@...nel.org>,
Alexei Starovoitov <ast@...nel.org>,
Daniel Borkmann <daniel@...earbox.net>,
Andrii Nakryiko <andrii@...nel.org>,
Martin KaFai Lau
<martin.lau@...ux.dev>,
Eduard Zingerman <eddyz87@...il.com>, Song Liu <song@...nel.org>,
John Fastabend <john.fastabend@...il.com>,
KP Singh <kpsingh@...nel.org>, Stanislav Fomichev <sdf@...ichev.me>,
Hao Luo <haoluo@...gle.com>, Jiri Olsa <jolsa@...nel.org>,
Shuah Khan <shuah@...nel.org>, linux-kernel@...r.kernel.org,
linux-kselftest@...r.kernel.org
Subject: Re: [RFC bpf-next 1/2] bpftool: Print map ID upon creation and
support JSON output
Hi Yonghong,
On 29/10/25 07:44, Yonghong Song wrote:
>
>
> On 10/28/25 5:57 AM, Harshit Mogalapalli wrote:
>> It is useful to print map ID on successful creation.
>>
>> JSON case:
>> $ ./bpftool -j map create /sys/fs/bpf/test_map4 type hash key 4 value
>> 8 entries 128 name map4
>> {"id":12}
>>
>> Generic case:
>> $ ./bpftool map create /sys/fs/bpf/test_map5 type hash key 4 value 8
>> entries 128 name map5
>> Map successfully created with ID: 15
>>
>> Bpftool Issue: https://github.com/libbpf/bpftool/issues/121
>> Signed-off-by: Harshit Mogalapalli <harshit.m.mogalapalli@...cle.com>
>> ---
>> tools/bpf/bpftool/map.c | 24 ++++++++++++++++++++----
>> 1 file changed, 20 insertions(+), 4 deletions(-)
>>
>> diff --git a/tools/bpf/bpftool/map.c b/tools/bpf/bpftool/map.c
>> index c9de44a45778..b6580f25361d 100644
>> --- a/tools/bpf/bpftool/map.c
>> +++ b/tools/bpf/bpftool/map.c
>> @@ -1251,6 +1251,8 @@ static int do_create(int argc, char **argv)
>> LIBBPF_OPTS(bpf_map_create_opts, attr);
>> enum bpf_map_type map_type = BPF_MAP_TYPE_UNSPEC;
>> __u32 key_size = 0, value_size = 0, max_entries = 0;
>> + struct bpf_map_info map_info = {};
>> + __u32 map_info_len = sizeof(map_info);
>> const char *map_name = NULL;
>> const char *pinfile;
>> int err = -1, fd;
>> @@ -1353,13 +1355,27 @@ static int do_create(int argc, char **argv)
>> }
>> err = do_pin_fd(fd, pinfile);
>> - close(fd);
>> - if (err)
>> + if (err) {
>> + close(fd);
>
> I think you can remove close(fd) here,
>
>> goto exit;
>> + }
>> - if (json_output)
>> - jsonw_null(json_wtr);
>> + err = bpf_obj_get_info_by_fd(fd, &map_info, &map_info_len);
>> + if (err) {
>> + p_err("Failed to fetch map info: %s\n", strerror(errno));
>> + close(fd);
>
> and here
>
>> + goto exit;
>> + }
>> + close(fd);
>
> and here,
>
>> +
>> + if (json_output) {
>> + jsonw_start_object(json_wtr);
>> + jsonw_int_field(json_wtr, "id", map_info.id);
>> + jsonw_end_object(json_wtr);
>> + } else {
>> + printf("Map successfully created with ID: %u\n", map_info.id);
>> + }
>> exit:
>
> and put close(fd) here.
I think we need one more close_fd: label and then put a close(fd); here.
As there are other gotos to exit earlier in this function when fd is
uninitialized, which can the error like:
map.c: In function ‘do_create’:
map.c:1375:9: warning: ‘fd’ may be used uninitialized
[-Wmaybe-uninitialized]
1375 | close(fd);
| ^~~~~~~~~
map.c:1258:23: note: ‘fd’ was declared here
1258 | int err = -1, fd;
| ^~
So, maybe we could do something like this:
err = do_pin_fd(fd, pinfile);
- close(fd);
if (err)
- goto exit;
+ goto close_fd;
- if (json_output)
- jsonw_null(json_wtr);
+ err = bpf_obj_get_info_by_fd(fd, &map_info, &map_info_len);
+ if (err) {
+ p_err("Failed to fetch map info: %s\n", strerror(errno));
+ goto close_fd;
+ }
+ if (json_output) {
+ jsonw_start_object(json_wtr);
+ jsonw_int_field(json_wtr, "id", map_info.id);
+ jsonw_end_object(json_wtr);
+ } else {
+ printf("Map successfully created with ID: %u\n",
map_info.id);
+ }
+close_fd:
+ close(fd);
exit:
if (attr.inner_map_fd > 0)
close(attr.inner_map_fd);
I can prepare a v2 with this change, but wouldn't it be simpler to add a
direct close(fd); on the few error paths instead of introducing an
additional label for close(fd);?
Thoughts/Suggestions ?
Thanks,
Harshit
>
>> if (attr.inner_map_fd > 0)
>> close(attr.inner_map_fd);
>
>
Powered by blists - more mailing lists