Re: [PATCH v8 17/19] conf, fwd: Allow switching to new rules received from pesto

[Laurent Vivier <lvivier@redhat.com>]

On 5/6/26 10:52, David Gibson wrote:
> On Wed, May 06, 2026 at 10:49:55AM +0200, Stefano Brivio wrote:
>> On Wed, 6 May 2026 10:39:30 +0200
>> Stefano Brivio <sbrivio@redhat.com> wrote:
>>
>>> On Wed, 6 May 2026 10:12:21 +0200
>>> Laurent Vivier <lvivier@redhat.com> wrote:
>>>
>>>> On 5/6/26 01:47, Stefano Brivio wrote:
>>>>> From: David Gibson <david@gibson.dropbear.id.au>
>>>>>
>>>>> We can now receive updates to the forwarding rules from the pesto client
>>>>> and store them in a "pending" copy of the forwarding tables.  Implement
>>>>> switching to using the new rules.
>>>>>
>>>>> The logic is in a new fwd_listen_switch().  For now this closes all
>>>>> listening sockets related to the old tables, swaps the active and pending
>>>>> tables, then listens based on the new tables.  In future we look to improve
>>>>> this so that we don't temporarily stop listening on ports that both the
>>>>> old and new tables specify.
>>>>>
>>>>> Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
>>>>> [sbrivio: In fwd_listen_switch(), use the destination size as argument
>>>>>    to memcpy(), instead of sizeof(tmp), as suggested by Laurent]
>>>>> Signed-off-by: Stefano Brivio <sbrivio@redhat.com>
>>>>> ---
>>>>>    conf.c |  5 ++---
>>>>>    fwd.c  | 34 ++++++++++++++++++++++++++++++++++
>>>>>    fwd.h  |  1 +
>>>>>    3 files changed, 37 insertions(+), 3 deletions(-)
>>>>>
>>>>> diff --git a/conf.c b/conf.c
>>>>> index 76344da..3f48793 100644
>>>>> --- a/conf.c
>>>>> +++ b/conf.c
>>>>> @@ -2160,15 +2160,14 @@ void conf_handler(struct ctx *c, uint32_t events)
>>>>>    			fwd_rules_dump(info, fwd->rules, fwd->count,
>>>>>    				       "    ", "");
>>>>>    		}
>>>>> +
>>>>> +		fwd_listen_switch(c);
>>>>>    	}
>>>>>    
>>>>>    	if (events & EPOLLHUP) {
>>>>>    		debug("Configuration client hangup");
>>>>> -		goto close;
>>>>>    	}
>>>>>    
>>>>> -	return;
>>>>> -
>>>>>    close:
>>>>>    	conf_close(c);
>>>>>    
>>>>> diff --git a/fwd.c b/fwd.c
>>>>> index d93d2e5..0697435 100644
>>>>> --- a/fwd.c
>>>>> +++ b/fwd.c
>>>>> @@ -534,6 +534,40 @@ int fwd_listen_init(const struct ctx *c)
>>>>>    	return 0;
>>>>>    }
>>>>>    
>>>>> +/**
>>>>> + * fwd_listen_switch() - Switch from current to pending rules table
>>>>> + * @c:		Execution context
>>>>> + */
>>>>> +void fwd_listen_switch(struct ctx *c)
>>>>> +{
>>>>> +	struct fwd_table *tmp[PIF_NUM_TYPES];
>>>>> +	unsigned i;
>>>>> +
>>>>> +	/* Stop listening on the old tables */
>>>>> +	for (i = 0; i < PIF_NUM_TYPES; i++) {
>>>>> +		struct fwd_table *fwd = c->fwd[i];
>>>>> +
>>>>> +		if (!fwd)
>>>>> +			continue;
>>>>> +
>>>>> +		debug("Flushing %u old %s rules", fwd->count, pif_name(i));
>>>>> +		fwd_listen_close(fwd);
>>>>> +		fwd->count = fwd->sock_count = 0;
>>>>> +	}
>>>>> +
>>>>> +	/* Swap active and pending tables */
>>>>> +	static_assert(sizeof(tmp) == sizeof(c->fwd) &&
>>>>> +		      sizeof(tmp) == sizeof(c->fwd_pending),
>>>>> +		      "Temporary has wrong size");
>>>>
>>>> At this point:
>>>>
>>>> c->fwd[PIF_HOST] = &fwd_in;
>>>> c->fwd[PIF_SPLICE] = &fwd_out;
>>>>
>>>> c->fwd_pending[PIF_HOST] = &fwd_in_pending;
>>>> c->fwd_pending[PIF_SPLICE] = &fwd_out_pending;
>>>>    
>>>>> +	memcpy(&tmp, (void *)c->fwd, sizeof(tmp));
>>>>> +	memcpy((void *)c->fwd, (void *)c->fwd_pending, sizeof(c->fwd));
>>>>> +	memcpy((void *)c->fwd_pending, &tmp, sizeof(c->fwd_pending));
>>>>
>>>> At this point:
>>>>
>>>> c->fwd[PIF_HOST] = &fwd_in_pending;
>>>> c->fwd[PIF_SPLICE] = &fwd_out_pending;
>>>>
>>>> c->fwd_pending[PIF_HOST] = &fwd_in;
>>>> c->fwd_pending[PIF_SPLICE] = &fwd_out;
>>>
>>> Yeah, makes sense, I can change that in v9.
>>>
>>>> Perhaps it should be noted somewhere to avoid confusion in the future?
>>>
>>> What do you think should be noted exactly, and where? Can you show a
>>> practical example of the change you're proposing?
>>
>> ...I'm leaving like it is in v9 to make sure I'm not misinterpreting
>> you, and also because the current (v8) version is obviously correct and
>> I also tested it fairly heavily by now.
>>
>> I'd suggest optimising this (and commenting as needed) in a separate
>> patch later.
> 
> As noted in another branch of this thread, I think all it really needs
> is renaming the globals that c->fwd and c->fwd_pending point to.  They
> should just be fwd_in_[12] (or an array of 2 tables), instead of
> implying a semantic difference between the plain and "pending" copies.
> 

I agree with that.

Thanks,
Laurent