Re: [PATCH bpf-next v2 03/20] bpf: support bpf tracing/iter programs for BPF_LINK_CREATE

[Alexei Starovoitov <ast@fb.com>]

On 5/5/20 5:14 PM, Yonghong Song wrote:
> 
> 
> On 5/5/20 2:30 PM, Andrii Nakryiko wrote:
>> On Sun, May 3, 2020 at 11:26 PM Yonghong Song <yhs@fb.com> wrote:
>>>
>>> Given a bpf program, the step to create an anonymous bpf iterator is:
>>>    - create a bpf_iter_link, which combines bpf program and the target.
>>>      In the future, there could be more information recorded in the 
>>> link.
>>>      A link_fd will be returned to the user space.
>>>    - create an anonymous bpf iterator with the given link_fd.
>>>
>>> The bpf_iter_link can be pinned to bpffs mount file system to
>>> create a file based bpf iterator as well.
>>>
>>> The benefit to use of bpf_iter_link:
>>>    - using bpf link simplifies design and implementation as bpf link
>>>      is used for other tracing bpf programs.
>>>    - for file based bpf iterator, bpf_iter_link provides a standard
>>>      way to replace underlying bpf programs.
>>>    - for both anonymous and free based iterators, bpf link query
>>>      capability can be leveraged.
>>>
>>> The patch added support of tracing/iter programs for BPF_LINK_CREATE.
>>> A new link type BPF_LINK_TYPE_ITER is added to facilitate link
>>> querying. Currently, only prog_id is needed, so there is no
>>> additional in-kernel show_fdinfo() and fill_link_info() hook
>>> is needed for BPF_LINK_TYPE_ITER link.
>>>
>>> Signed-off-by: Yonghong Song <yhs@fb.com>
>>> ---
>>
>> LGTM. See small nit about __GFP_NOWARN.
>>
>> Acked-by: Andrii Nakryiko <andriin@fb.com>
>>
>>
>>>   include/linux/bpf.h            |  1 +
>>>   include/linux/bpf_types.h      |  1 +
>>>   include/uapi/linux/bpf.h       |  1 +
>>>   kernel/bpf/bpf_iter.c          | 62 ++++++++++++++++++++++++++++++++++
>>>   kernel/bpf/syscall.c           | 14 ++++++++
>>>   tools/include/uapi/linux/bpf.h |  1 +
>>>   6 files changed, 80 insertions(+)
>>>
>>
>> [...]
>>
>>> +int bpf_iter_link_attach(const union bpf_attr *attr, struct bpf_prog 
>>> *prog)
>>> +{
>>> +       struct bpf_link_primer link_primer;
>>> +       struct bpf_iter_target_info *tinfo;
>>> +       struct bpf_iter_link *link;
>>> +       bool existed = false;
>>> +       u32 prog_btf_id;
>>> +       int err;
>>> +
>>> +       if (attr->link_create.target_fd || attr->link_create.flags)
>>> +               return -EINVAL;
>>> +
>>> +       prog_btf_id = prog->aux->attach_btf_id;
>>> +       mutex_lock(&targets_mutex);
>>> +       list_for_each_entry(tinfo, &targets, list) {
>>> +               if (tinfo->btf_id == prog_btf_id) {
>>> +                       existed = true;
>>> +                       break;
>>> +               }
>>> +       }
>>> +       mutex_unlock(&targets_mutex);
>>> +       if (!existed)
>>> +               return -ENOENT;
>>> +
>>> +       link = kzalloc(sizeof(*link), GFP_USER | __GFP_NOWARN);
>>
>> nit: all existing link implementation don't specify __GFP_NOWARN,
>> wonder if bpf_iter_link should be special?
> 
> Nothing special. Just feel __GFP_NOWARN is the right thing to do to 
> avoid pollute dmesg since -ENOMEM is returned to user space. But in
> reality, unlike some key/value allocation where the size could be huge
> and __GFP_NOWARN might be more useful, here, sizeof(*link) is fixed
> and small, __GFP_NOWARN probably not that useful.
> 
> Will drop it.

actually all existing user space driven allocation have nowarn.
If we missed it in other link allocs we should probably add it.