Re: [PATCH v4 net-next 8/8] selftests: forwarding: ethtool_rmon: Add histogram counter test

[Vladimir Oltean <olteanv@gmail.com>]

On Thu, Dec 14, 2023 at 02:50:29PM +0100, Tobias Waldekranz wrote:
> Validate the operation of rx and tx histogram counters, if supported
> by the interface, by sending batches of packets targeted for each
> bucket.
> 
> Signed-off-by: Tobias Waldekranz <tobias@waldekranz.com>
> ---

Tested-by: Vladimir Oltean <vladimir.oltean@nxp.com>
Reviewed-by: Vladimir Oltean <vladimir.oltean@nxp.com>

# enetc + ocelot
./ethtool_rmon.sh eno0 swp0
TEST: rx-pkts64to64 on eno0                                         [ OK ]
TEST: rx-pkts65to127 on eno0                                        [ OK ]
TEST: rx-pkts128to255 on eno0                                       [ OK ]
TEST: rx-pkts256to511 on eno0                                       [ OK ]
TEST: rx-pkts512to1023 on eno0                                      [ OK ]
TEST: rx-pkts1024to1522 on eno0                                     [ OK ]
TEST: rx-pkts1523to9600 on eno0                                     [ OK ]
TEST: rx-pkts64to64 on swp0                                         [ OK ]
TEST: rx-pkts65to127 on swp0                                        [ OK ]
TEST: rx-pkts128to255 on swp0                                       [ OK ]
TEST: rx-pkts256to511 on swp0                                       [ OK ]
TEST: rx-pkts512to1023 on swp0                                      [ OK ]
TEST: rx-pkts1024to1526 on swp0                                     [ OK ]
TEST: rx-pkts1527to65535 on swp0                                    [ OK ]
TEST: tx-pkts64to64 on eno0                                         [ OK ]
TEST: tx-pkts65to127 on eno0                                        [ OK ]
TEST: tx-pkts128to255 on eno0                                       [ OK ]
TEST: tx-pkts256to511 on eno0                                       [ OK ]
TEST: tx-pkts512to1023 on eno0                                      [ OK ]
TEST: tx-pkts1024to1522 on eno0                                     [ OK ]
TEST: tx-pkts1523to9600 on eno0                                     [ OK ]
TEST: tx-pkts64to64 on swp0                                         [ OK ]
TEST: tx-pkts65to127 on swp0                                        [ OK ]
TEST: tx-pkts128to255 on swp0                                       [ OK ]
TEST: tx-pkts256to511 on swp0                                       [ OK ]
TEST: tx-pkts512to1023 on swp0                                      [ OK ]
TEST: tx-pkts1024to1526 on swp0                                     [ OK ]
TEST: tx-pkts1527to65535 on swp0                                    [ OK ]

# mv88e6xxx
./ethtool_rmon.sh lan1 lan2
TEST: rx-pkts64to64 on lan1                                         [ OK ]
TEST: rx-pkts65to127 on lan1                                        [ OK ]
TEST: rx-pkts128to255 on lan1                                       [ OK ]
TEST: rx-pkts256to511 on lan1                                       [ OK ]
TEST: rx-pkts512to1023 on lan1                                      [ OK ]
TEST: rx-pkts1024to65535 on lan1                                    [ OK ]
TEST: rx-pkts64to64 on lan2                                         [ OK ]
TEST: rx-pkts65to127 on lan2                                        [ OK ]
TEST: rx-pkts128to255 on lan2                                       [ OK ]
TEST: rx-pkts256to511 on lan2                                       [ OK ]
TEST: rx-pkts512to1023 on lan2                                      [ OK ]
TEST: rx-pkts1024to65535 on lan2                                    [ OK ]
TEST: lan1 does not support tx histogram counters                   [SKIP]
TEST: lan2 does not support tx histogram counters                   [SKIP]

This is just lovely, thanks for the work, Tobias.

Just one nitpick below.

> +rmon_histogram()
> +{
> +	local iface=$1; shift
> +	local neigh=$1; shift
> +	local set=$1; shift
> +	local nbuckets=0
> +	local step=
> +
> +	RET=0
> +
> +	while read -r -a bucket; do
> +		step="$set-pkts${bucket[0]}to${bucket[1]} on $iface"
> +
> +		for if in $iface $neigh; do

My syntax highlighting in vim gets confused by the fact that you name a
variable "if". I guess something like "netif" would do the trick. But
bash doesn't seem to be confused. I can send a patch renaming this after
it gets merged. There's no reason to resend - it's not a functional
change.

> +			if ! ensure_mtu $if ${bucket[0]}; then
> +				log_test_skip "$if does not support the required MTU for $step"
> +				return
> +			fi
> +		done
> +
> +		if ! bucket_test $iface $neigh $set $nbuckets ${bucket[0]}; then
> +			check_err 1 "$step failed"
> +			return 1
> +		fi
> +		log_test "$step"
> +		nbuckets=$((nbuckets + 1))
> +	done < <(ethtool --json -S $iface --groups rmon | \
> +		jq -r ".[0].rmon[\"${set}-pktsNtoM\"][]|[.low, .high]|@tsv" 2>/dev/null)