BPF List
 help / color / mirror / Atom feed
From: Martin KaFai Lau <martin.lau@linux.dev>
To: Kui-Feng Lee <sinquersw@gmail.com>, Kui-Feng Lee <kuifeng@meta.com>
Cc: bpf@vger.kernel.org, ast@kernel.org, song@kernel.org,
	kernel-team@meta.com, andrii@kernel.org, sdf@google.com
Subject: Re: [PATCH bpf-next v7 5/8] bpf: Update the struct_ops of a bpf_link.
Date: Fri, 17 Mar 2023 22:38:49 -0700	[thread overview]
Message-ID: <b977a7d4-4c82-eb46-3457-4a92ec94de46@linux.dev> (raw)
In-Reply-To: <6fed9361-ba07-c387-14d4-2fee2d161b5f@gmail.com>

On 3/17/23 6:11 PM, Kui-Feng Lee wrote:
>>> --- a/net/bpf/bpf_dummy_struct_ops.c
>>> +++ b/net/bpf/bpf_dummy_struct_ops.c
>>> @@ -222,12 +222,18 @@ static void bpf_dummy_unreg(void *kdata)
>>>   {
>>>   }
>>> +static int bpf_dummy_update(void *kdata, void *old_kdata)
>>> +{
>>> +    return -EOPNOTSUPP;
>>> +}
>>> +
>>>   struct bpf_struct_ops bpf_bpf_dummy_ops = {
>>>       .verifier_ops = &bpf_dummy_verifier_ops,
>>>       .init = bpf_dummy_init,
>>>       .check_member = bpf_dummy_ops_check_member,
>>>       .init_member = bpf_dummy_init_member,
>>>       .reg = bpf_dummy_reg,
>>> +    .update = bpf_dummy_update,
>>
>> When looking at this together in patch 5, the changes in 
>> bpf_dummy_struct_ops.c should not be needed.
> 
> I don't follow you.
> If we don't assign a function to .update, it will fail in
> bpf_struct_ops_map_link_update(). Of course, I can add a check
> in bpf_struct_ops_map_link_update() to return an error if .update
> is NULL.

Yes, test ->update in bpf_struct_ops.c is better but not in 
bpf_struct_ops_map_link_update (more on this later). It does not need the dummy 
struct_ops to test the link. The dummy struct_ops was created to catch the 
trampoline img issue with ops/func having different args and return value.

It is better to enforce the BPF_F_LINK struct_ops must support both ->validate 
and ->update at the beginning and it can be revisited later. The current 
'->validate' testing in bpf_struct_ops_map_update_elem() in patch 3 is too late. 
Being able to '->validate' is particularly important for BPF_F_LINK struct_ops. 
Testing '->validate' and '->update' in bpf_struct_ops_init() will be too strict 
for now though when considering other on-going efforts to support struct_ops in 
other subsystems. A better place to test both '->validate' and '->update' should 
be in bpf_struct_ops_map_alloc(). It should return -ENOTSUPP when trying to 
creating a struct_ops BPF_F_LINK map without st_ops->validate and st_ops->update.

  reply	other threads:[~2023-03-18  5:38 UTC|newest]

Thread overview: 33+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2023-03-16  2:36 [PATCH bpf-next v7 0/8] Transit between BPF TCP congestion controls Kui-Feng Lee
2023-03-16  2:36 ` [PATCH bpf-next v7 1/8] bpf: Retire the struct_ops map kvalue->refcnt Kui-Feng Lee
2023-03-17 16:47   ` Martin KaFai Lau
2023-03-17 20:41     ` Kui-Feng Lee
2023-03-16  2:36 ` [PATCH bpf-next v7 2/8] net: Update an existing TCP congestion control algorithm Kui-Feng Lee
2023-03-17 15:23   ` Daniel Borkmann
2023-03-17 17:18     ` Martin KaFai Lau
2023-03-17 17:23       ` Daniel Borkmann
2023-03-17 21:46         ` Kui-Feng Lee
2023-03-17 23:07     ` Kui-Feng Lee
2023-03-16  2:36 ` [PATCH bpf-next v7 3/8] bpf: Create links for BPF struct_ops maps Kui-Feng Lee
2023-03-17 18:10   ` Martin KaFai Lau
2023-03-17 20:52     ` Kui-Feng Lee
2023-03-16  2:36 ` [PATCH bpf-next v7 4/8] libbpf: Create a bpf_link in bpf_map__attach_struct_ops() Kui-Feng Lee
2023-03-17 18:44   ` Martin KaFai Lau
2023-03-17 21:00     ` Kui-Feng Lee
2023-03-17 22:23   ` Andrii Nakryiko
2023-03-17 23:48     ` Kui-Feng Lee
2023-03-16  2:36 ` [PATCH bpf-next v7 5/8] bpf: Update the struct_ops of a bpf_link Kui-Feng Lee
2023-03-17 19:23   ` Martin KaFai Lau
2023-03-17 21:39     ` Kui-Feng Lee
2023-03-18  1:11     ` Kui-Feng Lee
2023-03-18  5:38       ` Martin KaFai Lau [this message]
2023-03-17 22:27   ` Andrii Nakryiko
2023-03-18  0:41     ` Kui-Feng Lee
2023-03-16  2:36 ` [PATCH bpf-next v7 6/8] libbpf: Update a bpf_link with another struct_ops Kui-Feng Lee
2023-03-17 19:42   ` Martin KaFai Lau
2023-03-17 21:40     ` Kui-Feng Lee
2023-03-17 22:33   ` Andrii Nakryiko
2023-03-18  1:17     ` Kui-Feng Lee
2023-03-16  2:36 ` [PATCH bpf-next v7 7/8] libbpf: Use .struct_ops.link section to indicate a struct_ops with a link Kui-Feng Lee
2023-03-17 22:35   ` Andrii Nakryiko
2023-03-16  2:36 ` [PATCH bpf-next v7 8/8] selftests/bpf: Test switching TCP Congestion Control algorithms Kui-Feng Lee

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=b977a7d4-4c82-eb46-3457-4a92ec94de46@linux.dev \
    --to=martin.lau@linux.dev \
    --cc=andrii@kernel.org \
    --cc=ast@kernel.org \
    --cc=bpf@vger.kernel.org \
    --cc=kernel-team@meta.com \
    --cc=kuifeng@meta.com \
    --cc=sdf@google.com \
    --cc=sinquersw@gmail.com \
    --cc=song@kernel.org \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox