Re: [PATCH v4 net-next 11/15] selftests: xsk: send UDP over IPv4 or IPv6 between the interface MACs

[sashiko-bot@kernel.org]

Thank you for your contribution! Sashiko AI review found 1 potential issue(s) to consider:
- [Low] bpf: multi-line comment style violation
--

commit c93ec22f2ead9dd358b23391b62cb4c9037e31b7
Author: Maciej Fijalkowski <maciej.fijalkowski@intel.com>

    selftests: xsk: send UDP over IPv4 or IPv6 between the interface MACs
    
    xskxceiver sends frames of the loopback ethertype between made-up MAC
    addresses, and its XDP programs redirect every frame to the XSKMAP.
    That suits a veth pair that carries nothing else, but not a physical
    link, where the peer NIC filters out the made-up addresses and the
    programs take neighbour discovery, LLDP or SSH away from the stack.
    
    Send the test packets as UDP between the MAC addresses of the two
    interfaces on any link. Each endpoint takes the IPv4 or IPv6 addresses
    of the flow (--src-ip, --dst-ip), the MAC address of its peer
    (--dst-mac) and the UDP port (--udp-port), and reads its own MAC
    address from the interface. Over IPv4 the UDP checksum is left at zero,
    as IPv4 allows; over IPv6 it is computed from the known payload.
    PKT_HDR_SIZE and MIN_PKT_SIZE now stand for variables that
    xsk_set_flow() sets, so the smallest packet grows with the headers to
    64 bytes over IPv4 and 88 over IPv6, and a verbatim stream adds up the
    descriptors of a packet for its IP and UDP lengths.

> diff --git a/tools/testing/selftests/net/lib/xsk/xsk_xdp_progs.bpf.c b/tools/testing/selftests/net/lib/xsk/xsk_xdp_progs.bpf.c
> index 023d8befd4cab..2b55dbaf13b24 100644
> --- a/tools/testing/selftests/net/lib/xsk/xsk_xdp_progs.bpf.c
> +++ b/tools/testing/selftests/net/lib/xsk/xsk_xdp_progs.bpf.c
> @@ -15,12 +23,64 @@ struct {
>  	__uint(value_size, sizeof(int));
>  } xsk SEC(".maps");
>  
> -static unsigned int idx;
>  int adjust_value = 0;
>  int count = 0;
> +/* UDP destination port of the test flow, other traffic goes to the stack */
> +__u16 test_port;
> +
> +/* Return the UDP header of a test packet, and in hdr_len the size of the
> + * headers in front of its payload, or NULL for any other packet.
> + */

[Severity: Low]
This is a minor coding style observation. The BPF subsystem prefers the
opening of multi-line comments to be on its own line. Should this be
formatted as follows?

/*
 * Return the UDP header of a test packet, and in hdr_len the size of the
 * headers in front of its payload, or NULL for any other packet.
 */

> +static __always_inline struct udphdr *test_udp_hdr(struct xdp_md *xdp,
> +						   __u32 *hdr_len)
> +{
> +	void *data_end = (void *)(long)xdp->data_end;
> +	void *data = (void *)(long)xdp->data;
> +	struct ethhdr *eth = data;
> +	struct ipv6hdr *ip6;
> +	struct udphdr *udp;
> +	struct iphdr *ip;

-- 
Sashiko AI review · https://sashiko.dev/#/patchset/20261016174338.787135-1-maciej.fijalkowski@intel.com?part=11