[PATCH bpf-next] bpf: Define bpf_token_show_fdinfo with CONFIG_PROC_FS

[PATCH bpf-next] bpf: Define bpf_token_show_fdinfo with CONFIG_PROC_FS

[Tao Chen]

Protect bpf_token_show_fdinfo with CONFIG_PROC_FS check, follow the
pattern used with other *_show_fdinfo functions.

Fixes: 35f96de04127 ("bpf: Introduce BPF token object")
Signed-off-by: Tao Chen <chen.dylane@linux.dev>
---
 kernel/bpf/token.c | 4 ++++
 1 file changed, 4 insertions(+)

diff --git a/kernel/bpf/token.c b/kernel/bpf/token.c
index 26057aa13..104ca37e9 100644
--- a/kernel/bpf/token.c
+++ b/kernel/bpf/token.c
@@ -65,6 +65,7 @@ static int bpf_token_release(struct inode *inode, struct file *filp)
 	return 0;
 }
 
+#ifdef CONFIG_PROC_FS
 static void bpf_token_show_fdinfo(struct seq_file *m, struct file *filp)
 {
 	struct bpf_token *token = filp->private_data;
@@ -98,6 +99,7 @@ static void bpf_token_show_fdinfo(struct seq_file *m, struct file *filp)
 	else
 		seq_printf(m, "allowed_attachs:\t0x%llx\n", token->allowed_attachs);
 }
+#endif
 
 #define BPF_TOKEN_INODE_NAME "bpf-token"
 
@@ -105,7 +107,9 @@ static const struct inode_operations bpf_token_iops = { };
 
 static const struct file_operations bpf_token_fops = {
 	.release	= bpf_token_release,
+#ifdef CONFIG_PROC_FS
 	.show_fdinfo	= bpf_token_show_fdinfo,
+#endif
 };
 
 int bpf_token_create(union bpf_attr *attr)
-- 
2.43.0

Re: [PATCH bpf-next] bpf: Define bpf_token_show_fdinfo with CONFIG_PROC_FS

[Jiri Olsa]

On Tue, Mar 18, 2025 at 02:25:57PM +0800, Tao Chen wrote:
> Protect bpf_token_show_fdinfo with CONFIG_PROC_FS check, follow the
> pattern used with other *_show_fdinfo functions.
> 
> Fixes: 35f96de04127 ("bpf: Introduce BPF token object")
> Signed-off-by: Tao Chen <chen.dylane@linux.dev>
> ---
>  kernel/bpf/token.c | 4 ++++
>  1 file changed, 4 insertions(+)
> 
> diff --git a/kernel/bpf/token.c b/kernel/bpf/token.c
> index 26057aa13..104ca37e9 100644
> --- a/kernel/bpf/token.c
> +++ b/kernel/bpf/token.c
> @@ -65,6 +65,7 @@ static int bpf_token_release(struct inode *inode, struct file *filp)
>  	return 0;
>  }
>  
> +#ifdef CONFIG_PROC_FS
>  static void bpf_token_show_fdinfo(struct seq_file *m, struct file *filp)
>  {
>  	struct bpf_token *token = filp->private_data;
> @@ -98,6 +99,7 @@ static void bpf_token_show_fdinfo(struct seq_file *m, struct file *filp)
>  	else
>  		seq_printf(m, "allowed_attachs:\t0x%llx\n", token->allowed_attachs);
>  }
> +#endif
>  
>  #define BPF_TOKEN_INODE_NAME "bpf-token"
>  
> @@ -105,7 +107,9 @@ static const struct inode_operations bpf_token_iops = { };
>  
>  static const struct file_operations bpf_token_fops = {
>  	.release	= bpf_token_release,
> +#ifdef CONFIG_PROC_FS
>  	.show_fdinfo	= bpf_token_show_fdinfo,
> +#endif

there's many more of such cases.. I'm not sure if it makes sense to fix that,
because it does not break the build and only save space for !CONFIG_PROC_FS
kernels

jirka


>  };
>  
>  int bpf_token_create(union bpf_attr *attr)
> -- 
> 2.43.0
> 

Re: [PATCH bpf-next] bpf: Define bpf_token_show_fdinfo with CONFIG_PROC_FS

[Alexei Starovoitov]

On Tue, Mar 18, 2025 at 2:29 AM Jiri Olsa <olsajiri@gmail.com> wrote:
>
> On Tue, Mar 18, 2025 at 02:25:57PM +0800, Tao Chen wrote:
> > Protect bpf_token_show_fdinfo with CONFIG_PROC_FS check, follow the
> > pattern used with other *_show_fdinfo functions.
> >
> > Fixes: 35f96de04127 ("bpf: Introduce BPF token object")
> > Signed-off-by: Tao Chen <chen.dylane@linux.dev>
> > ---
> >  kernel/bpf/token.c | 4 ++++
> >  1 file changed, 4 insertions(+)
> >
> > diff --git a/kernel/bpf/token.c b/kernel/bpf/token.c
> > index 26057aa13..104ca37e9 100644
> > --- a/kernel/bpf/token.c
> > +++ b/kernel/bpf/token.c
> > @@ -65,6 +65,7 @@ static int bpf_token_release(struct inode *inode, struct file *filp)
> >       return 0;
> >  }
> >
> > +#ifdef CONFIG_PROC_FS
> >  static void bpf_token_show_fdinfo(struct seq_file *m, struct file *filp)
> >  {
> >       struct bpf_token *token = filp->private_data;
> > @@ -98,6 +99,7 @@ static void bpf_token_show_fdinfo(struct seq_file *m, struct file *filp)
> >       else
> >               seq_printf(m, "allowed_attachs:\t0x%llx\n", token->allowed_attachs);
> >  }
> > +#endif
> >
> >  #define BPF_TOKEN_INODE_NAME "bpf-token"
> >
> > @@ -105,7 +107,9 @@ static const struct inode_operations bpf_token_iops = { };
> >
> >  static const struct file_operations bpf_token_fops = {
> >       .release        = bpf_token_release,
> > +#ifdef CONFIG_PROC_FS
> >       .show_fdinfo    = bpf_token_show_fdinfo,
> > +#endif
>
> there's many more of such cases.. I'm not sure if it makes sense to fix that,
> because it does not break the build and only save space for !CONFIG_PROC_FS
> kernels

+1.
let's keep the code as-is.

pw-bot: cr

Re: [PATCH bpf-next] bpf: Define bpf_token_show_fdinfo with CONFIG_PROC_FS

[Tao Chen]

在 2025/3/18 22:59, Alexei Starovoitov 写道:
> On Tue, Mar 18, 2025 at 2:29 AM Jiri Olsa <olsajiri@gmail.com> wrote:
>>
>> On Tue, Mar 18, 2025 at 02:25:57PM +0800, Tao Chen wrote:
>>> Protect bpf_token_show_fdinfo with CONFIG_PROC_FS check, follow the
>>> pattern used with other *_show_fdinfo functions.
>>>
>>> Fixes: 35f96de04127 ("bpf: Introduce BPF token object")
>>> Signed-off-by: Tao Chen <chen.dylane@linux.dev>
>>> ---
>>>   kernel/bpf/token.c | 4 ++++
>>>   1 file changed, 4 insertions(+)
>>>
>>> diff --git a/kernel/bpf/token.c b/kernel/bpf/token.c
>>> index 26057aa13..104ca37e9 100644
>>> --- a/kernel/bpf/token.c
>>> +++ b/kernel/bpf/token.c
>>> @@ -65,6 +65,7 @@ static int bpf_token_release(struct inode *inode, struct file *filp)
>>>        return 0;
>>>   }
>>>
>>> +#ifdef CONFIG_PROC_FS
>>>   static void bpf_token_show_fdinfo(struct seq_file *m, struct file *filp)
>>>   {
>>>        struct bpf_token *token = filp->private_data;
>>> @@ -98,6 +99,7 @@ static void bpf_token_show_fdinfo(struct seq_file *m, struct file *filp)
>>>        else
>>>                seq_printf(m, "allowed_attachs:\t0x%llx\n", token->allowed_attachs);
>>>   }
>>> +#endif
>>>
>>>   #define BPF_TOKEN_INODE_NAME "bpf-token"
>>>
>>> @@ -105,7 +107,9 @@ static const struct inode_operations bpf_token_iops = { };
>>>
>>>   static const struct file_operations bpf_token_fops = {
>>>        .release        = bpf_token_release,
>>> +#ifdef CONFIG_PROC_FS
>>>        .show_fdinfo    = bpf_token_show_fdinfo,
>>> +#endif
>>
>> there's many more of such cases.. I'm not sure if it makes sense to fix that,
>> because it does not break the build and only save space for !CONFIG_PROC_FS
>> kernels
> 
> +1.
> let's keep the code as-is.
> 

Got it, thanks anyway.

> pw-bot: cr


-- 
Best Regards
Tao Chen

[PATCH bpf-next] bpf: Define bpf_token_show_fdinfo with CONFIG_PROC_FS

[Tao Chen]

Protect bpf_token_show_fdinfo with CONFIG_PROC_FS check, otherwise
it will compile error if CONFIG_PROC_FS is not set.

Fixes: 35f96de04127 ("bpf: Introduce BPF token object")
Signed-off-by: Tao Chen <chen.dylane@linux.dev>
---
 kernel/bpf/token.c | 2 ++
 1 file changed, 2 insertions(+)

diff --git a/kernel/bpf/token.c b/kernel/bpf/token.c
index 26057aa13..4396eefde 100644
--- a/kernel/bpf/token.c
+++ b/kernel/bpf/token.c
@@ -105,7 +105,9 @@ static const struct inode_operations bpf_token_iops = { };
 
 static const struct file_operations bpf_token_fops = {
 	.release	= bpf_token_release,
+#ifdef CONFIG_PROC_FS
 	.show_fdinfo	= bpf_token_show_fdinfo,
+#endif
 };
 
 int bpf_token_create(union bpf_attr *attr)
-- 
2.43.0

Re: [PATCH bpf-next] bpf: Define bpf_token_show_fdinfo with CONFIG_PROC_FS

[Song Liu]

On Mon, Mar 17, 2025 at 8:52 AM Tao Chen <chen.dylane@linux.dev> wrote:
>
> Protect bpf_token_show_fdinfo with CONFIG_PROC_FS check, otherwise
> it will compile error if CONFIG_PROC_FS is not set.

On bpt-next/master, I don't see compile error with

CONFIG_BPF_SYSCALL=y
...
# CONFIG_PROC_FS is not set

Are you testing with a different tree/branch?

Thanks,
Song

>
> Fixes: 35f96de04127 ("bpf: Introduce BPF token object")
> Signed-off-by: Tao Chen <chen.dylane@linux.dev>
> ---
>  kernel/bpf/token.c | 2 ++
>  1 file changed, 2 insertions(+)
>
> diff --git a/kernel/bpf/token.c b/kernel/bpf/token.c
> index 26057aa13..4396eefde 100644
> --- a/kernel/bpf/token.c
> +++ b/kernel/bpf/token.c
> @@ -105,7 +105,9 @@ static const struct inode_operations bpf_token_iops = { };
>
>  static const struct file_operations bpf_token_fops = {
>         .release        = bpf_token_release,
> +#ifdef CONFIG_PROC_FS
>         .show_fdinfo    = bpf_token_show_fdinfo,
> +#endif
>  };
>
>  int bpf_token_create(union bpf_attr *attr)
> --
> 2.43.0
>
>

Re: [PATCH bpf-next] bpf: Define bpf_token_show_fdinfo with CONFIG_PROC_FS

[Tao Chen]

在 2025/3/18 02:01, Song Liu 写道:
> On Mon, Mar 17, 2025 at 8:52 AM Tao Chen <chen.dylane@linux.dev> wrote:
>>
>> Protect bpf_token_show_fdinfo with CONFIG_PROC_FS check, otherwise
>> it will compile error if CONFIG_PROC_FS is not set.
> 
> On bpt-next/master, I don't see compile error with
> 
> CONFIG_BPF_SYSCALL=y
> ...
> # CONFIG_PROC_FS is not set
> 
> Are you testing with a different tree/branch?
> 
> Thanks,
> Song
> 
>>
>> Fixes: 35f96de04127 ("bpf: Introduce BPF token object")
>> Signed-off-by: Tao Chen <chen.dylane@linux.dev>
>> ---
>>   kernel/bpf/token.c | 2 ++
>>   1 file changed, 2 insertions(+)
>>
>> diff --git a/kernel/bpf/token.c b/kernel/bpf/token.c
>> index 26057aa13..4396eefde 100644
>> --- a/kernel/bpf/token.c
>> +++ b/kernel/bpf/token.c
>> @@ -105,7 +105,9 @@ static const struct inode_operations bpf_token_iops = { };
>>
>>   static const struct file_operations bpf_token_fops = {
>>          .release        = bpf_token_release,
>> +#ifdef CONFIG_PROC_FS
>>          .show_fdinfo    = bpf_token_show_fdinfo,
>> +#endif
>>   };
>>
>>   int bpf_token_create(union bpf_attr *attr)
>> --
>> 2.43.0
>>
>>

Hi Song,

I try it, it seems that i had some misunderstandings before about 
CONFIG_PROC_FS, and the description in the changelog is not quite 
correct. Please ignore this patch, and i will resend it. Just to follow 
the pattern used with other *_show_fdinfo functions and only define 
bpf_token_show_fdinfo when CONFIG_PROC_FS is set. Thank you for your 
reminder.

-- 
Best Regards
Tao Chen