Re: [PATCH bpf-next v3 03/10] bpf: Allow read access to addr_len from cgroup sockaddr programs

[Martin KaFai Lau <martin.lau@linux.dev>]

On 4/21/23 9:27 AM, Daan De Meyer wrote:
> As prep for adding unix socket support to the cgroup sockaddr hooks,
> let's expose the sockaddr addrlen in bpf_sock_addr_kern. While not
> important for AF_INET or AF_INET6, the sockaddr length is important
> when working with AF_UNIX sockaddrs as the size of the sockaddr cannot
> be determined just from the address family or the sockaddr's contents.
> 
> __cgroup_bpf_run_filter_sock_addr() is modified to return the addr_len
> in preparation for adding unix socket support for which we'll need to
> return the modified address length.
> 
> Signed-off-by: Daan De Meyer <daan.j.demeyer@gmail.com>
> ---
>   include/linux/bpf-cgroup.h | 73 +++++++++++++++++++-------------------
>   include/linux/filter.h     |  1 +
>   kernel/bpf/cgroup.c        | 16 +++++++--
>   net/ipv4/af_inet.c         |  8 ++---
>   net/ipv4/ping.c            |  8 ++++-
>   net/ipv4/tcp_ipv4.c        |  8 ++++-
>   net/ipv4/udp.c             | 17 ++++++---
>   net/ipv6/af_inet6.c        |  8 ++---
>   net/ipv6/ping.c            |  8 ++++-
>   net/ipv6/tcp_ipv6.c        |  8 ++++-
>   net/ipv6/udp.c             | 14 ++++++--
>   11 files changed, 111 insertions(+), 58 deletions(-)
> 
> diff --git a/include/linux/bpf-cgroup.h b/include/linux/bpf-cgroup.h
> index 57e9e109257e..f3f5adf3881f 100644
> --- a/include/linux/bpf-cgroup.h
> +++ b/include/linux/bpf-cgroup.h
> @@ -120,6 +120,7 @@ int __cgroup_bpf_run_filter_sk(struct sock *sk,
>   
>   int __cgroup_bpf_run_filter_sock_addr(struct sock *sk,
>   				      struct sockaddr *uaddr,
> +				      u32 uaddrlen,

If the bpf_sock_addr_set() kfunc can only change the sin[6]_addr and unix_path 
(the comment in patch 5), the "u32 uaddrlen" can be changed to "u32 *uaddrlen" 
here. The new unix_path length can be passed back to af_unix.c in "*uaddrlen". 
The inet[6] code path can just pass NULL and most of the code churn in this 
patch will no longer be needed?