[PATCH v3 bpf-next] net: remove check in __cgroup_bpf_run_filter_skb

[PATCH v3 bpf-next] net: remove check in __cgroup_bpf_run_filter_skb

[Oliver Crumrine]

Originally, this patch removed a redundant check in
BPF_CGROUP_RUN_PROG_INET_EGRESS, as the check was already being done in
the function it called, __cgroup_bpf_run_filter_skb. For v2, it was
reccomended that I remove the check from __cgroup_bpf_run_filter_skb,
and add the checks to the other macro that calls that function,
BPF_CGROUP_RUN_PROG_INET_INGRESS.

To sum it up, checking that the socket exists and that it is a full
socket is now part of both macros BPF_CGROUP_RUN_PROG_INET_EGRESS and
BPF_CGROUP_RUN_PROG_INET_INGRESS, and it is no longer part of the
function they call, __cgroup_bpf_run_filter_skb.

Signed-off-by: Oliver Crumrine <ozlinuxc@gmail.com>

v2->v3: Sent to bpf-next instead of generic patch
v1->v2: Addressed feedback about where check should be removed.
---
 include/linux/bpf-cgroup.h | 7 ++++---
 kernel/bpf/cgroup.c        | 3 ---
 2 files changed, 4 insertions(+), 6 deletions(-)

diff --git a/include/linux/bpf-cgroup.h b/include/linux/bpf-cgroup.h
index a789266feac3..b28dc0ff4218 100644
--- a/include/linux/bpf-cgroup.h
+++ b/include/linux/bpf-cgroup.h
@@ -195,10 +195,11 @@ static inline bool cgroup_bpf_sock_enabled(struct sock *sk,
 #define BPF_CGROUP_RUN_PROG_INET_INGRESS(sk, skb)			      \
 ({									      \
 	int __ret = 0;							      \
-	if (cgroup_bpf_enabled(CGROUP_INET_INGRESS) &&			      \
-	    cgroup_bpf_sock_enabled(sk, CGROUP_INET_INGRESS))		      \
+	if (cgroup_bpf_enabled(CGROUP_INET_INGRESS) &&			      \
+	    cgroup_bpf_sock_enabled(sk, CGROUP_INET_INGRESS) && sk &&	      \
+	    sk_fullsock(sk))						      \
 		__ret = __cgroup_bpf_run_filter_skb(sk, skb,		      \
-						    CGROUP_INET_INGRESS);     \
+						    CGROUP_INET_INGRESS);     \
 									      \
 	__ret;								      \
 })
diff --git a/kernel/bpf/cgroup.c b/kernel/bpf/cgroup.c
index 491d20038cbe..644bfb39cf9d 100644
--- a/kernel/bpf/cgroup.c
+++ b/kernel/bpf/cgroup.c
@@ -1364,9 +1364,6 @@ int __cgroup_bpf_run_filter_skb(struct sock *sk,
 	struct cgroup *cgrp;
 	int ret;
 
-	if (!sk || !sk_fullsock(sk))
-		return 0;
-
 	if (sk->sk_family != AF_INET && sk->sk_family != AF_INET6)
 		return 0;
 
-- 
2.43.0

Re: [PATCH v3 bpf-next] net: remove check in __cgroup_bpf_run_filter_skb

[Stanislav Fomichev]

On 02/09, Oliver Crumrine wrote:
> Originally, this patch removed a redundant check in
> BPF_CGROUP_RUN_PROG_INET_EGRESS, as the check was already being done in
> the function it called, __cgroup_bpf_run_filter_skb. For v2, it was
> reccomended that I remove the check from __cgroup_bpf_run_filter_skb,
> and add the checks to the other macro that calls that function,
> BPF_CGROUP_RUN_PROG_INET_INGRESS.
> 
> To sum it up, checking that the socket exists and that it is a full
> socket is now part of both macros BPF_CGROUP_RUN_PROG_INET_EGRESS and
> BPF_CGROUP_RUN_PROG_INET_INGRESS, and it is no longer part of the
> function they call, __cgroup_bpf_run_filter_skb.
> 
> Signed-off-by: Oliver Crumrine <ozlinuxc@gmail.com>
> 
> v2->v3: Sent to bpf-next instead of generic patch
> v1->v2: Addressed feedback about where check should be removed.
> ---
>  include/linux/bpf-cgroup.h | 7 ++++---
>  kernel/bpf/cgroup.c        | 3 ---
>  2 files changed, 4 insertions(+), 6 deletions(-)
> 
> diff --git a/include/linux/bpf-cgroup.h b/include/linux/bpf-cgroup.h
> index a789266feac3..b28dc0ff4218 100644
> --- a/include/linux/bpf-cgroup.h
> +++ b/include/linux/bpf-cgroup.h
> @@ -195,10 +195,11 @@ static inline bool cgroup_bpf_sock_enabled(struct sock *sk,
>  #define BPF_CGROUP_RUN_PROG_INET_INGRESS(sk, skb)			      \
>  ({									      \
>  	int __ret = 0;							      \
> -	if (cgroup_bpf_enabled(CGROUP_INET_INGRESS) &&			      \
> -	    cgroup_bpf_sock_enabled(sk, CGROUP_INET_INGRESS))		      \
> +	if (cgroup_bpf_enabled(CGROUP_INET_INGRESS) &&			      \
> +	    cgroup_bpf_sock_enabled(sk, CGROUP_INET_INGRESS) && sk &&	      \
> +	    sk_fullsock(sk))						      \
>  		__ret = __cgroup_bpf_run_filter_skb(sk, skb,		      \

[..]

> -						    CGROUP_INET_INGRESS);     \
> +						    CGROUP_INET_INGRESS);     \

The bot still can't git-am it. And I can't either. Did you somehow
manually mangle that part above? The original line has less trailing spaces
than what your diff source has, look at:

https://git.kernel.org/pub/scm/linux/kernel/git/bpf/bpf-next.git/tree/include/linux/bpf-cgroup.h#n201

Can you drop this part? Let the idents stay broken :-)

Re: [PATCH v3 bpf-next] net: remove check in __cgroup_bpf_run_filter_skb

[Oliver Crumrine]

On Fri, Feb 09, 2024 at 04:01:21PM -0800, Stanislav Fomichev wrote:
> The bot still can't git-am it. And I can't either. Did you somehow
> manually mangle that part above? The original line has less trailing spaces
> than what your diff source has, look at:
I literally saw the bot issue and the spaces just before you sent this.
> Can you drop this part? Let the idents stay broken :-)
Just did. Didn't manually edit the patch so I don't exactly know how
that happened, but thanks for pointing it out.