Re: [PATCH net-next v5] net: ipv6: seg6: report lwtunnel setup errors via extack

[Andrea Mayer <andrea.mayer@uniroma2.it>]

On Fri,  9 Oct 2026 17:20:13 +0200
Gabriel Goller <g.goller@proxmox.com> wrote:

Hi Gabriel,

thanks for the v5.

> seg6_build_state() rejected invalid configurations with a generic -EINVAL,
> so "ip route add ... encap seg6 ..." sometimes reported "Invalid argument".
> 
> Add extack messages to the checks that lacked one, and use
> NL_REQ_ATTR_CHECK() for the missing SRH attribute so that userspace can
> tell which attribute is missing. The checks themselves and their return
> values are unchanged.
> 
> Signed-off-by: Gabriel Goller <g.goller@proxmox.com>
> ---
> 
> v5 (https://lore.kernel.org/all/20261005140637.644840-1-g.goller@proxmox.com/):
>  * use NL_REQ_ATTR_CHECK() for the missing SRH attribute (Jakub)
>  * reword the inline mode message to "inline mode requires the IPv6
>    address family", since it also applies to IPv4 nexthop objects
>    (Andrea)
>  * consistently call the header "IPv6 Segment Routing Header" (RFC 8754)
>    in all messages (Andrea)
>  * include the offending mode value in the invalid mode message
>  * reword the commit message to match
> 
> v4 (https://lore.kernel.org/all/20260923120228.172643-1-g.goller@proxmox.com/):
>  * use NL_SET_ERR_MSG when the SRH attr is missing
>  * use NL_SET_ERR_MSG_ATTR when we have an invalid attr
> 
> v3 (https://lore.kernel.org/all/20260922090851.38978-1-g.goller@proxmox.com/):
>  * format fix
>  * use NL_SET_ERR_MSG_ATTR
> 
> v2 (https://lore.kernel.org/all/20260918153544.1178884-1-g.goller@proxmox.com/):
>  * dropped tests
> 
>  net/ipv6/seg6_iptunnel.c | 26 +++++++++++++++++++++-----
>  1 file changed, 21 insertions(+), 5 deletions(-)
> 
> [snip]
>  	default:
> +		NL_SET_ERR_MSG_ATTR_FMT(extack, tb[SEG6_IPTUNNEL_SRH],
> +					"invalid SRv6 encapsulation mode %d",
> +					tuninfo->mode);
>  		return -EINVAL;

A small nit on this one, as it is new in v5: I would have kept the
v4 text here, "invalid SRv6 encapsulation mode". No need for a new
version just for this.

Reviewed-by: Andrea Mayer <andrea.mayer@uniroma2.it>

Thanks,
Andrea