Re: [PATCH v7 4/4] test-pmd: add more packet verbose decode options

[Alex Chapman <alex.chapman@arm.com>]

Hi Stephen,

I have gone through your patch series and the hexdump option would be 
quite valuable for use in DTS.

However I am currently facing the issue of distinguishing noise packets 
from intentional packets within the verbose output. Prior to your patch, 
the intention was to use the Layer 4 port to distinguish between them, 
however with the hexdump option, the plan is to now use a custom payload.

The one issue is that with verbose level 5 does not print the required 
RSS hash and RSS queue information.

Would it be possible for you to add this to your patch series? 
Otherwise, do you have any ideas and/or solutions to tackle this 
specific problem?

On 8/2/24 20:56, Stephen Hemminger wrote:
<snip>
> +static uint16_t
> +dump_pkt_burst(uint16_t port_id, uint16_t queue, struct rte_mbuf *pkts[],
> +	      uint16_t nb_pkts, int is_rx)
> +{
> +	if (unlikely(nb_pkts == 0))
> +		return 0;
> +
> +	switch (verbose_level) {
> +	case VERBOSE_RX ... VERBOSE_BOTH:
> +		dump_pkt_verbose(port_id, queue, pkts, nb_pkts, is_rx);
> +		break;
> +	case VERBOSE_DISSECT:
> +		dump_pkt_brief(port_id, queue, pkts, nb_pkts, is_rx);
> +		break;
> +	case VERBOSE_HEX:
> +		dump_pkt_hex(pkts, nb_pkts);
> +	}
> +	fflush(stdout);
> +	return nb_pkts;
> +}

I have identified the function above that would require altering in 
order to accommodate the additional RSS information.

Thanks,
Alex