Re: [PATCH bpf-next v6 3/3] bpf: Add show_fdinfo for kprobe_multi

[Tao Chen <chen.dylane@linux.dev>]

在 2025/7/2 04:37, Andrii Nakryiko 写道:
> On Fri, Jun 27, 2025 at 1:23 AM Tao Chen <chen.dylane@linux.dev> wrote:
>>
>> Show kprobe_multi link info with fdinfo, the info as follows:
>>
>> link_type:      kprobe_multi
>> link_id:        1
>> prog_tag:       33be53a4fd673e1d
>> prog_id:        21
>> kprobe_cnt:     8
>> missed: 0
>> cookie           func
>> 1                bpf_fentry_test1+0x0/0x20
>> 7                bpf_fentry_test2+0x0/0x20
>> 2                bpf_fentry_test3+0x0/0x20
>> 3                bpf_fentry_test4+0x0/0x20
>> 4                bpf_fentry_test5+0x0/0x20
>> 5                bpf_fentry_test6+0x0/0x20
>> 6                bpf_fentry_test7+0x0/0x20
>> 8                bpf_fentry_test8+0x0/0x10
> 
> two nits:
> 
> 1) order of cookie. For uprobes you have cookie at the end, here in
> the front. Given variable-sized func name, I'd move cookie to the
> front for uprobes for consistency.
> 

Ok, will change it in v7.

> 2) field sizing for cookie (16) is a) not sufficient for maximum
> possible u64 (20 digits) and b) very wasteful in common case of small
> numbers. So use tab instead of fixed-sized column? And why 16
> character sizing for the func column? Just to have more spaces
> emitted?
> 

I just referred to the implementation in bpftool/link.c, i will use tab 
format in v7. "16 character sizing for the func" just keeps consistency 
with cookie, " %s" for func name maybe better.

> 
> Other than that the series looks good to me.
> 
> pw-bot: cr
> 
> 
>>
>> Signed-off-by: Tao Chen <chen.dylane@linux.dev>
>> ---
>>   kernel/trace/bpf_trace.c | 27 +++++++++++++++++++++++++++
>>   1 file changed, 27 insertions(+)
>>
>> diff --git a/kernel/trace/bpf_trace.c b/kernel/trace/bpf_trace.c
>> index 1c75f9c6c66..e8f070504c4 100644
>> --- a/kernel/trace/bpf_trace.c
>> +++ b/kernel/trace/bpf_trace.c
>> @@ -2622,10 +2622,37 @@ static int bpf_kprobe_multi_link_fill_link_info(const struct bpf_link *link,
>>          return err;
>>   }
>>
>> +#ifdef CONFIG_PROC_FS
>> +static void bpf_kprobe_multi_show_fdinfo(const struct bpf_link *link,
>> +                                        struct seq_file *seq)
>> +{
>> +       struct bpf_kprobe_multi_link *kmulti_link;
>> +
>> +       kmulti_link = container_of(link, struct bpf_kprobe_multi_link, link);
>> +
>> +       seq_printf(seq,
>> +                  "kprobe_cnt:\t%u\n"
>> +                  "missed:\t%lu\n",
>> +                  kmulti_link->cnt,
>> +                  kmulti_link->fp.nmissed);
>> +
>> +       seq_printf(seq, "%-16s %-16s\n", "cookie", "func");
>> +       for (int i = 0; i < kmulti_link->cnt; i++) {
>> +               seq_printf(seq,
>> +                          "%-16llu %-16pS\n",
>> +                          kmulti_link->cookies[i],
>> +                          (void *)kmulti_link->addrs[i]);
>> +       }
>> +}
>> +#endif
>> +
>>   static const struct bpf_link_ops bpf_kprobe_multi_link_lops = {
>>          .release = bpf_kprobe_multi_link_release,
>>          .dealloc_deferred = bpf_kprobe_multi_link_dealloc,
>>          .fill_link_info = bpf_kprobe_multi_link_fill_link_info,
>> +#ifdef CONFIG_PROC_FS
>> +       .show_fdinfo = bpf_kprobe_multi_show_fdinfo,
>> +#endif
>>   };
>>
>>   static void bpf_kprobe_multi_cookie_swap(void *a, void *b, int size, const void *priv)
>> --
>> 2.48.1
>>


-- 
Best Regards
Tao Chen