Re: [PATCH 1/2] test: Don't require 64-bit prefixes in perf tests

[Stefano Brivio <sbrivio@redhat.com>]

On Wed,  6 Nov 2024 12:44:14 +1100
David Gibson <david@gibson.dropbear.id.au> wrote:

> When determining the namespace's IPv6 address in the perf test setup, we
> explicitly filter for addresses with a 64-bit prefix length.  There's no
> real reason we need that - as long as it's a global address we can use it.
> I suspect this was copied without thinking from a similar example in the
> NDP tests, where the 64-bit prefix length _is_ meaningful (though it's not
> entirely clear if the handling is correct there either).

So, I'm not sure why we had those filters, I also suspect a copy and
past, but:

> Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
> ---
>  test/perf/pasta_tcp | 2 +-
>  test/perf/pasta_udp | 2 +-
>  2 files changed, 2 insertions(+), 2 deletions(-)
> 
> diff --git a/test/perf/pasta_tcp b/test/perf/pasta_tcp
> index d1ccf7d..88284b2 100644
> --- a/test/perf/pasta_tcp
> +++ b/test/perf/pasta_tcp
> @@ -211,7 +211,7 @@ tr	TCP throughput over IPv6: host to ns
>  iperf3s	ns 10002
>  
>  nsout	IFNAME ip -j link show | jq -rM '.[] | select(.link_type == "ether").ifname'
> -nsout	ADDR6 ip -j -6 addr show|jq -rM '.[] | select(.ifname == "__IFNAME__").addr_info[] | select(.scope == "global" and .prefixlen == 64).local'
> +nsout	ADDR6 ip -j -6 addr show|jq -rM '.[] | select(.ifname == "__IFNAME__").addr_info[] | select(.scope == "global").local'
>  bw	-
>  bw	-
>  bw	-
> diff --git a/test/perf/pasta_udp b/test/perf/pasta_udp
> index 544bf17..3d07091 100644
> --- a/test/perf/pasta_udp
> +++ b/test/perf/pasta_udp
> @@ -196,7 +196,7 @@ tr	UDP throughput over IPv6: host to ns
>  iperf3s	ns 10002
>  
>  nsout	IFNAME ip -j link show | jq -rM '.[] | select(.link_type == "ether").ifname'
> -nsout	ADDR6 ip -j -6 addr show|jq -rM '.[] | select(.ifname == "__IFNAME__").addr_info[] | select(.scope == "global" and .prefixlen == 64).local'
> +nsout	ADDR6 ip -j -6 addr show|jq -rM '.[] | select(.ifname == "__IFNAME__").addr_info[] | select(.scope == "global").local'

...these cause failures on a "typical" pasta-in-passt-in-passt setup,
because in the guest we typically have two addresses, one assigned via
SLAAC, and one assigned via DHCPv6, say:

    inet6 2a01:4f8:222:904::2/128 scope global noprefixroute 
       valid_lft forever preferred_lft forever
    inet6 2a01:4f8:222:904:5054:ff:fe7c:cea1/64 scope global mngtmpaddr noprefixroute 
       valid_lft forever preferred_lft forever

and, whoops:

ns$ ip -j -6 addr show|jq -rM '.[] | select(.ifname == "ens3").addr_info[] | select(.scope == "global").local'
2a01:4f8:222:904:5054:ff:fe7c:cea1
2a01:4f8:222:904::2

I guess we should come up with a jq expression selecting the longest
prefix. Any address will do, but it's probably better to keep this
consistent. On the subject, see also:

  https://github.com/siderolabs/talos/issues/9725#issuecomment-2478509622

...or should we just pick the first one as we do elsewhere?

-- 
Stefano