Re: [PATCH v3 21/25] pesto: Read current ruleset from passt/pasta and display it

[Stefano Brivio <sbrivio@redhat.com>]

On Mon, 23 Mar 2026 18:37:28 +1100
David Gibson <david@gibson.dropbear.id.au> wrote:

> Implement serialisation of our current forwarding rules in conf.c,
> deserialising it to display in the pesto client.
> 
> Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
> ---
>  conf.c     | 44 +++++++++++++++++++++++++++++++
>  fwd_rule.c | 40 ++++++++++++++++++++++++++++
>  fwd_rule.h |  3 +++
>  pesto.c    | 77 +++++++++++++++++++++++++++++++++++++++++++++++++++---
>  4 files changed, 160 insertions(+), 4 deletions(-)
> 
> diff --git a/conf.c b/conf.c
> index b878db94..7f311914 100644
> --- a/conf.c
> +++ b/conf.c
> @@ -2338,6 +2338,47 @@ static int conf_send_pifs(const struct ctx *c, int fd)
>  	return 0;
>  }
>  
> +/**
> + * conf_send_rules() - Send current forwarding rules to dynamic update client (pesto)

What about:

 * conf_send_rules() - Send forwarding rules to configuration client (pesto)

?

> + * @c:		Execution context
> + * @fd:		Socket to the client
> + *
> + * Return: 0 on success, -1 on failure
> + */
> +static int conf_send_rules(const struct ctx *c, int fd)
> +{
> +	unsigned pif;
> +
> +	for (pif = 0; pif < PIF_NUM_TYPES; pif++) {
> +		const struct fwd_table *fwd = c->fwd[pif];
> +		unsigned i;
> +
> +		if (!fwd)
> +			continue;
> +
> +		assert(pif);
> +
> +		/* PIF id */
> +		if (write_u8(fd, pif))
> +			return -1;
> +
> +		/* Number of rules */
> +		if (write_u32(fd, fwd->count))
> +			return -1;
> +
> +		for (i = 0; i < fwd->count; i++) {
> +			if (fwd_rule_write(fd, &fwd->rules[i].rule))
> +				return -1;
> +		}
> +	}
> +
> +	/* Write 0 PIF id to finish */
> +	if (write_u8(fd, 0))

We have PIF_NONE, shouldn't we use that to make it obvious that it's an
invalid value?

> +		return -1;
> +
> +	return 0;
> +}
>
> [...]

I stopped reviewing here, the next patches in my series are anyway mine
and after those I guess it doesn't make sense, yet, that I review in
fine detail, as we'll probably need to change a bunch of things.

-- 
Stefano