public inbox for bpf@vger.kernel.org
 help / color / mirror / Atom feed
From: Kui-Feng Lee <sinquersw@gmail.com>
To: Martin KaFai Lau <martin.lau@linux.dev>, thinker.li@gmail.com
Cc: kuifeng@meta.com, bpf@vger.kernel.org, ast@kernel.org,
	song@kernel.org, kernel-team@meta.com, andrii@kernel.org
Subject: Re: [RFC bpf-next v3 05/11] bpf: hold module for bpf_struct_ops_map.
Date: Mon, 25 Sep 2023 16:42:57 -0700	[thread overview]
Message-ID: <437cd950-90ad-9cf4-8cb4-caa9538fef36@gmail.com> (raw)
In-Reply-To: <8393a1f3-b4cf-9e4c-ce76-4b09a3f1622b@linux.dev>



On 9/25/23 16:23, Martin KaFai Lau wrote:
> On 9/20/23 8:59 AM, thinker.li@gmail.com wrote:
>> From: Kui-Feng Lee <thinker.li@gmail.com>
>>
>> Ensure a module doesn't go away when a struct_ops object is still alive,
>> being a struct_ops type that is registered by the module.
>>
>> Signed-off-by: Kui-Feng Lee <thinker.li@gmail.com>
>> ---
>>   include/linux/bpf.h         | 1 +
>>   kernel/bpf/bpf_struct_ops.c | 6 ++++++
>>   2 files changed, 7 insertions(+)
>>
>> diff --git a/include/linux/bpf.h b/include/linux/bpf.h
>> index 0776cb584b3f..faaec20156f1 100644
>> --- a/include/linux/bpf.h
>> +++ b/include/linux/bpf.h
>> @@ -1627,6 +1627,7 @@ struct bpf_struct_ops {
>>       int (*update)(void *kdata, void *old_kdata);
>>       int (*validate)(void *kdata);
>>       const struct btf *btf;
>> +    struct module *owner;
>>       const struct btf_type *type;
>>       const struct btf_type *value_type;
>>       const char *name;
>> diff --git a/kernel/bpf/bpf_struct_ops.c b/kernel/bpf/bpf_struct_ops.c
>> index 7c2ef53687ef..ef8a1edec891 100644
>> --- a/kernel/bpf/bpf_struct_ops.c
>> +++ b/kernel/bpf/bpf_struct_ops.c
>> @@ -632,6 +632,8 @@ static void __bpf_struct_ops_map_free(struct 
>> bpf_map *map)
>>   static void bpf_struct_ops_map_free(struct bpf_map *map)
>>   {
>> +    struct bpf_struct_ops_map *st_map = (struct bpf_struct_ops_map 
>> *)map;
>> +
>>       /* The struct_ops's function may switch to another struct_ops.
>>        *
>>        * For example, bpf_tcp_cc_x->init() may switch to
>> @@ -649,6 +651,7 @@ static void bpf_struct_ops_map_free(struct bpf_map 
>> *map)
>>        */
>>       synchronize_rcu_mult(call_rcu, call_rcu_tasks);
>> +    module_put(st_map->st_ops->owner);
>>       __bpf_struct_ops_map_free(map);
>>   }
>> @@ -673,6 +676,9 @@ static struct bpf_map 
>> *bpf_struct_ops_map_alloc(union bpf_attr *attr)
>>       if (!st_ops)
>>           return ERR_PTR(-ENOTSUPP);
>> +    if (!try_module_get(st_ops->owner))
>> +        return ERR_PTR(-EINVAL);
> 
> The module can be gone at this point?
> I don't think try_module_get is safe. btf_try_get_module should be used 
> instead.

At this point, it holds btf, but not module. Module can go away while
some one still holding a refcount to the btf.

And, you are right, I should use btf_try_get_module().

> 
>> +
>>       vt = st_ops->value_type;
>>       if (attr->value_size != vt->size)
>>           return ERR_PTR(-EINVAL);
> 

  reply	other threads:[~2023-09-25 23:43 UTC|newest]

Thread overview: 34+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2023-09-20 15:59 [RFC bpf-next v3 00/11] Registrating struct_ops types from modules thinker.li
2023-09-20 15:59 ` [RFC bpf-next v3 01/11] bpf: refactory struct_ops type initialization to a function thinker.li
2023-09-20 15:59 ` [RFC bpf-next v3 02/11] bpf: add struct_ops_tab to btf thinker.li
2023-09-25 21:10   ` Martin KaFai Lau
2023-09-25 21:45     ` Kui-Feng Lee
2023-09-20 15:59 ` [RFC bpf-next v3 03/11] bpf: add register and unregister functions for struct_ops thinker.li
2023-09-25 23:07   ` Martin KaFai Lau
2023-09-25 23:13     ` Kui-Feng Lee
2023-09-25 23:31   ` Martin KaFai Lau
2023-09-26  0:19     ` Kui-Feng Lee
2023-09-20 15:59 ` [RFC bpf-next v3 04/11] bpf: attach a module BTF to a bpf_struct_ops thinker.li
2023-09-25 22:57   ` Martin KaFai Lau
2023-09-25 23:25     ` Kui-Feng Lee
2023-09-20 15:59 ` [RFC bpf-next v3 05/11] bpf: hold module for bpf_struct_ops_map thinker.li
2023-09-25 23:23   ` Martin KaFai Lau
2023-09-25 23:42     ` Kui-Feng Lee [this message]
2023-09-20 15:59 ` [RFC bpf-next v3 06/11] bpf: validate value_type thinker.li
2023-09-26  1:03   ` Martin KaFai Lau
2023-09-27 20:27     ` Kui-Feng Lee
2023-09-20 15:59 ` [RFC bpf-next v3 07/11] bpf, net: switch to storing struct_ops in btf thinker.li
2023-09-26  0:02   ` Martin KaFai Lau
2023-09-26  0:18     ` Kui-Feng Lee
2023-09-20 15:59 ` [RFC bpf-next v3 08/11] bpf: pass attached BTF to find correct type info of struct_ops progs thinker.li
2023-09-25 22:58   ` Andrii Nakryiko
2023-09-25 23:50     ` Kui-Feng Lee
2023-09-26  0:24   ` Martin KaFai Lau
2023-09-26  0:58     ` Kui-Feng Lee
2023-09-20 15:59 ` [RFC bpf-next v3 09/11] libbpf: Find correct module BTFs for struct_ops maps and progs thinker.li
2023-09-25 23:09   ` Andrii Nakryiko
2023-09-26  0:12     ` Kui-Feng Lee
2023-09-20 15:59 ` [RFC bpf-next v3 10/11] bpf: export btf_ctx_access to modules thinker.li
2023-09-20 15:59 ` [RFC bpf-next v3 11/11] selftests/bpf: test case for register_bpf_struct_ops() thinker.li
2023-09-26  1:19   ` Martin KaFai Lau
2023-09-26  1:33 ` [RFC bpf-next v3 00/11] Registrating struct_ops types from modules Martin KaFai Lau

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=437cd950-90ad-9cf4-8cb4-caa9538fef36@gmail.com \
    --to=sinquersw@gmail.com \
    --cc=andrii@kernel.org \
    --cc=ast@kernel.org \
    --cc=bpf@vger.kernel.org \
    --cc=kernel-team@meta.com \
    --cc=kuifeng@meta.com \
    --cc=martin.lau@linux.dev \
    --cc=song@kernel.org \
    --cc=thinker.li@gmail.com \
    /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