bpf: Remove unused field initialization

Message ID 20221221-bpf-syscall-v1-0-9550f5f2c3fc@chromium.org
State New
Headers
Series bpf: Remove unused field initialization |

Commit Message

Ricardo Ribalda Dec. 21, 2022, 7:55 p.m. UTC
  Maxlen is used by standard proc_handlers such as proc_dointvec(), but in this
case we have our own proc_handler. Remove the initialization.

Signed-off-by: Ricardo Ribalda <ribalda@chromium.org>
---
bpf: Trivial remove of unitialised field.

I have inspired myself in your code and heritaded this bug :). Fixing this
here so none else makes the same mistake.

To: Alexei Starovoitov <ast@kernel.org>
To: Daniel Borkmann <daniel@iogearbox.net>
To: John Fastabend <john.fastabend@gmail.com>
To: Andrii Nakryiko <andrii@kernel.org>
To: Martin KaFai Lau <martin.lau@linux.dev>
To: Song Liu <song@kernel.org>
To: Yonghong Song <yhs@fb.com>
To: KP Singh <kpsingh@kernel.org>
To: Stanislav Fomichev <sdf@google.com>
To: Hao Luo <haoluo@google.com>
To: Jiri Olsa <jolsa@kernel.org>
Cc: bpf@vger.kernel.org
Cc: linux-kernel@vger.kernel.org
---
 kernel/bpf/syscall.c | 1 -
 1 file changed, 1 deletion(-)


---
base-commit: b6bb9676f2165d518b35ba3bea5f1fcfc0d969bf
change-id: 20221221-bpf-syscall-58d1ac3f817a

Best regards,
  

Comments

Stanislav Fomichev Dec. 21, 2022, 8:10 p.m. UTC | #1
On Wed, Dec 21, 2022 at 11:55 AM Ricardo Ribalda <ribalda@chromium.org> wrote:
>
> Maxlen is used by standard proc_handlers such as proc_dointvec(), but in this
> case we have our own proc_handler. Remove the initialization.

Are you sure?

bpf_stats_handler
  proc_dointvec_minmax
    do_proc_dointvec
      __do_proc_dointvec
        vleft = table->maxlen / sizeof(*i);

Maybe we should really do the following instead?

.maxlen: sizeof(int)

?

> Signed-off-by: Ricardo Ribalda <ribalda@chromium.org>
> ---
> bpf: Trivial remove of unitialised field.
>
> I have inspired myself in your code and heritaded this bug :). Fixing this
> here so none else makes the same mistake.
>
> To: Alexei Starovoitov <ast@kernel.org>
> To: Daniel Borkmann <daniel@iogearbox.net>
> To: John Fastabend <john.fastabend@gmail.com>
> To: Andrii Nakryiko <andrii@kernel.org>
> To: Martin KaFai Lau <martin.lau@linux.dev>
> To: Song Liu <song@kernel.org>
> To: Yonghong Song <yhs@fb.com>
> To: KP Singh <kpsingh@kernel.org>
> To: Stanislav Fomichev <sdf@google.com>
> To: Hao Luo <haoluo@google.com>
> To: Jiri Olsa <jolsa@kernel.org>
> Cc: bpf@vger.kernel.org
> Cc: linux-kernel@vger.kernel.org
> ---
>  kernel/bpf/syscall.c | 1 -
>  1 file changed, 1 deletion(-)
>
> diff --git a/kernel/bpf/syscall.c b/kernel/bpf/syscall.c
> index 35972afb6850..8e55456bd648 100644
> --- a/kernel/bpf/syscall.c
> +++ b/kernel/bpf/syscall.c
> @@ -5319,7 +5319,6 @@ static struct ctl_table bpf_syscall_table[] = {
>         {
>                 .procname       = "bpf_stats_enabled",
>                 .data           = &bpf_stats_enabled_key.key,
> -               .maxlen         = sizeof(bpf_stats_enabled_key),
>                 .mode           = 0644,
>                 .proc_handler   = bpf_stats_handler,
>         },
>
> ---
> base-commit: b6bb9676f2165d518b35ba3bea5f1fcfc0d969bf
> change-id: 20221221-bpf-syscall-58d1ac3f817a
>
> Best regards,
> --
> Ricardo Ribalda <ribalda@chromium.org>
  
Ricardo Ribalda Dec. 21, 2022, 8:13 p.m. UTC | #2
Hi Stanislav

On Wed, 21 Dec 2022 at 21:10, Stanislav Fomichev <sdf@google.com> wrote:
>
> On Wed, Dec 21, 2022 at 11:55 AM Ricardo Ribalda <ribalda@chromium.org> wrote:
> >
> > Maxlen is used by standard proc_handlers such as proc_dointvec(), but in this
> > case we have our own proc_handler. Remove the initialization.
>
> Are you sure?
>
> bpf_stats_handler
>   proc_dointvec_minmax
>     do_proc_dointvec
>       __do_proc_dointvec
>         vleft = table->maxlen / sizeof(*i);

I believe do_proc_dointvec is using the value from:

struct ctl_table tmp = {
  .maxlen=sixeof(val);
}

>
> Maybe we should really do the following instead?
>
> .maxlen: sizeof(int)
>
> ?
>
> > Signed-off-by: Ricardo Ribalda <ribalda@chromium.org>
> > ---
> > bpf: Trivial remove of unitialised field.
> >
> > I have inspired myself in your code and heritaded this bug :). Fixing this
> > here so none else makes the same mistake.
> >
> > To: Alexei Starovoitov <ast@kernel.org>
> > To: Daniel Borkmann <daniel@iogearbox.net>
> > To: John Fastabend <john.fastabend@gmail.com>
> > To: Andrii Nakryiko <andrii@kernel.org>
> > To: Martin KaFai Lau <martin.lau@linux.dev>
> > To: Song Liu <song@kernel.org>
> > To: Yonghong Song <yhs@fb.com>
> > To: KP Singh <kpsingh@kernel.org>
> > To: Stanislav Fomichev <sdf@google.com>
> > To: Hao Luo <haoluo@google.com>
> > To: Jiri Olsa <jolsa@kernel.org>
> > Cc: bpf@vger.kernel.org
> > Cc: linux-kernel@vger.kernel.org
> > ---
> >  kernel/bpf/syscall.c | 1 -
> >  1 file changed, 1 deletion(-)
> >
> > diff --git a/kernel/bpf/syscall.c b/kernel/bpf/syscall.c
> > index 35972afb6850..8e55456bd648 100644
> > --- a/kernel/bpf/syscall.c
> > +++ b/kernel/bpf/syscall.c
> > @@ -5319,7 +5319,6 @@ static struct ctl_table bpf_syscall_table[] = {
> >         {
> >                 .procname       = "bpf_stats_enabled",
> >                 .data           = &bpf_stats_enabled_key.key,
> > -               .maxlen         = sizeof(bpf_stats_enabled_key),
> >                 .mode           = 0644,
> >                 .proc_handler   = bpf_stats_handler,
> >         },
> >
> > ---
> > base-commit: b6bb9676f2165d518b35ba3bea5f1fcfc0d969bf
> > change-id: 20221221-bpf-syscall-58d1ac3f817a
> >
> > Best regards,
> > --
> > Ricardo Ribalda <ribalda@chromium.org>
  
Stanislav Fomichev Dec. 21, 2022, 8:18 p.m. UTC | #3
On Wed, Dec 21, 2022 at 12:13 PM Ricardo Ribalda <ribalda@chromium.org> wrote:
>
> Hi Stanislav
>
> On Wed, 21 Dec 2022 at 21:10, Stanislav Fomichev <sdf@google.com> wrote:
> >
> > On Wed, Dec 21, 2022 at 11:55 AM Ricardo Ribalda <ribalda@chromium.org> wrote:
> > >
> > > Maxlen is used by standard proc_handlers such as proc_dointvec(), but in this
> > > case we have our own proc_handler. Remove the initialization.
> >
> > Are you sure?
> >
> > bpf_stats_handler
> >   proc_dointvec_minmax
> >     do_proc_dointvec
> >       __do_proc_dointvec
> >         vleft = table->maxlen / sizeof(*i);
>
> I believe do_proc_dointvec is using the value from:
>
> struct ctl_table tmp = {
>   .maxlen=sixeof(val);
> }

Oh, you're right, I've missed that!

Acked-by: Stanislav Fomichev <sdf@google.com>

> >
> > Maybe we should really do the following instead?
> >
> > .maxlen: sizeof(int)
> >
> > ?
> >
> > > Signed-off-by: Ricardo Ribalda <ribalda@chromium.org>
> > > ---
> > > bpf: Trivial remove of unitialised field.
> > >
> > > I have inspired myself in your code and heritaded this bug :). Fixing this
> > > here so none else makes the same mistake.
> > >
> > > To: Alexei Starovoitov <ast@kernel.org>
> > > To: Daniel Borkmann <daniel@iogearbox.net>
> > > To: John Fastabend <john.fastabend@gmail.com>
> > > To: Andrii Nakryiko <andrii@kernel.org>
> > > To: Martin KaFai Lau <martin.lau@linux.dev>
> > > To: Song Liu <song@kernel.org>
> > > To: Yonghong Song <yhs@fb.com>
> > > To: KP Singh <kpsingh@kernel.org>
> > > To: Stanislav Fomichev <sdf@google.com>
> > > To: Hao Luo <haoluo@google.com>
> > > To: Jiri Olsa <jolsa@kernel.org>
> > > Cc: bpf@vger.kernel.org
> > > Cc: linux-kernel@vger.kernel.org
> > > ---
> > >  kernel/bpf/syscall.c | 1 -
> > >  1 file changed, 1 deletion(-)
> > >
> > > diff --git a/kernel/bpf/syscall.c b/kernel/bpf/syscall.c
> > > index 35972afb6850..8e55456bd648 100644
> > > --- a/kernel/bpf/syscall.c
> > > +++ b/kernel/bpf/syscall.c
> > > @@ -5319,7 +5319,6 @@ static struct ctl_table bpf_syscall_table[] = {
> > >         {
> > >                 .procname       = "bpf_stats_enabled",
> > >                 .data           = &bpf_stats_enabled_key.key,
> > > -               .maxlen         = sizeof(bpf_stats_enabled_key),
> > >                 .mode           = 0644,
> > >                 .proc_handler   = bpf_stats_handler,
> > >         },
> > >
> > > ---
> > > base-commit: b6bb9676f2165d518b35ba3bea5f1fcfc0d969bf
> > > change-id: 20221221-bpf-syscall-58d1ac3f817a
> > >
> > > Best regards,
> > > --
> > > Ricardo Ribalda <ribalda@chromium.org>
>
>
>
> --
> Ricardo Ribalda
  
patchwork-bot+netdevbpf@kernel.org Dec. 22, 2022, 3 p.m. UTC | #4
Hello:

This patch was applied to bpf/bpf-next.git (master)
by Daniel Borkmann <daniel@iogearbox.net>:

On Wed, 21 Dec 2022 20:55:29 +0100 you wrote:
> Maxlen is used by standard proc_handlers such as proc_dointvec(), but in this
> case we have our own proc_handler. Remove the initialization.
> 
> Signed-off-by: Ricardo Ribalda <ribalda@chromium.org>
> ---
> bpf: Trivial remove of unitialised field.
> 
> [...]

Here is the summary with links:
  - bpf: Remove unused field initialization
    https://git.kernel.org/bpf/bpf-next/c/cfca00767feb

You are awesome, thank you!
  

Patch

diff --git a/kernel/bpf/syscall.c b/kernel/bpf/syscall.c
index 35972afb6850..8e55456bd648 100644
--- a/kernel/bpf/syscall.c
+++ b/kernel/bpf/syscall.c
@@ -5319,7 +5319,6 @@  static struct ctl_table bpf_syscall_table[] = {
 	{
 		.procname	= "bpf_stats_enabled",
 		.data		= &bpf_stats_enabled_key.key,
-		.maxlen		= sizeof(bpf_stats_enabled_key),
 		.mode		= 0644,
 		.proc_handler	= bpf_stats_handler,
 	},