Re: [PATCH v7 20/27] udp: Create flows for datagrams from originating sockets

[Stefano Brivio <sbrivio@redhat.com>]

On Fri, 12 Jul 2024 08:58:57 +1000
David Gibson <david@gibson.dropbear.id.au> wrote:

> On Thu, Jul 11, 2024 at 10:20:01AM +0200, Stefano Brivio wrote:
> > On Thu, 11 Jul 2024 14:26:38 +1000
> > David Gibson <david@gibson.dropbear.id.au> wrote:
> >   
> > > On Wed, Jul 10, 2024 at 11:35:23PM +0200, Stefano Brivio wrote:  
> > > > On Wed, 10 Jul 2024 09:59:08 +1000
> > > > David Gibson <david@gibson.dropbear.id.au> wrote:
> > > >     
> > > > > On Wed, Jul 10, 2024 at 12:32:02AM +0200, Stefano Brivio wrote:    
> > > > > > Nits only, here:
> > > > > > 
> > > > > > On Fri,  5 Jul 2024 12:07:17 +1000
> > > > > > David Gibson <david@gibson.dropbear.id.au> wrote:
> > > > > >    
> > > > > > > [...]
> > > > > > >
> > > > > > > + * @c:		Execution context
> > > > > > > + * @proto:	Protocol of the flow (IP L4 protocol number)
> > > > > > > + * @pif:	Interface of the flow
> > > > > > > + * @esa:	Socket address of the endpoint
> > > > > > > + * @fport:	Forwarding port number
> > > > > > > + *
> > > > > > > + * Return: sidx of the matching flow & side, FLOW_SIDX_NONE if not found
> > > > > > > + */
> > > > > > > +flow_sidx_t flow_lookup_sa(const struct ctx *c, uint8_t proto, uint8_t pif,
> > > > > > > +			   const void *esa, in_port_t fport)
> > > > > > > +{
> > > > > > > +	struct flowside fside = {      
> > > > > > 
> > > > > > And the "f" in "fside" stands for "forwarding"... I don't have any
> > > > > > quick fix in mind, and it's _kind of_ clear anyway, but this makes me
> > > > > > doubt a bit about the "forwarding" / "endpoint" choice of words.      
> > > > > 
> > > > > Heh, no, here "fside" is simply short for "flowside".  Every flowside
> > > > > has both forwarding and endpoint elements.    
> > > > 
> > > > Oh, I thought you called it fside here because you're setting the
> > > > forwarding part of it directly, or something like that.
> > > >     
> > > > > So it is confusing, but
> > > > > for a different reason.  I need to find a different convention for
> > > > > naming struct flowside variables.  I'd say 'side', but sometimes
> > > > > that's used for the 1-bit integer indicating which side in a flow.
> > > > > 
> > > > > Hrm.. now that pif has been removed from here, maybe I could rename
> > > > > struct flowside back to 'flowaddrs' or 'sideaddrs' perhaps?    
> > > > 
> > > > That's also confusing because it contains ports too (even though sure,
> > > > in some sense they're part of the address).    
> > > 
> > > Right :/.
> > >   
> > > > I would suggest keeping it
> > > > like it is in for this series, but after that, if it's not too long,
> > > > what about flow_addrs_ports?    
> > > 
> > > Still need a conventional name for the variables.  "fap" probably
> > > isn't the best look, and still has the potentially confusing "f" in
> > > it.
> > >   
> > > > Actually, I don't think flowside is that bad. What I'm struggling with
> > > > is rather 'forwarding' and 'endpoint'. I don't have any good suggestion
> > > > at the moment, anyway. Using 'local' and 'remote' (laddr/lport,
> > > > raddr/rport) would be clearer to me and avoid the conflict with 'f' of
> > > > flowside, but you had good reasons to avoid that, if I recall correctly.    
> > > 
> > > Kind of, yeah.  Local and remote are great when it's clear we're
> > > talking specifically from the point of view of the passt process.
> > > There are a bunch of cases where it's not necessarily obvious if we're
> > > talking from that point of view, the point of view of the guest, or
> > > the point of view of the host (the last usually when the endpoint is
> > > somewhere on an entirely different system).  I wanted something that
> > > wherever we were talking about it is specifically relative to the
> > > passt process itself.
> > > 
> > > I get the impression that "forwarding" is causing more trouble here
> > > than "endpoint".  "midpoint address"?  "intercepted address"?
> > > "redirected address" (probably not, that's 'r' like remote but this
> > > would be the local address)?  
> > 
> > I think "forwarding" is still better than any of those. Perhaps "passt
> > address" (and paddr/pport)... but I'm not sure it's much better than
> > "forwarding".  
> 
> Hm.  "passthrough address"?  Kind of means the same thing as
> "forwarding" in context, and maybe evokes the idea that this is the
> address that passt itself owns?

I'd still prefer "forwarding" to that... at least I almost got used to
it, "passthrough" is even more confusing because not everything really
passes... through?

The part I'm missing (with any of those) is "here" vs. "there", that
is, a reference to the "where" and to the topology instead of what that
side does.

Again, I would just leave that as "forwarding" in this series, and then
change it everywhere if we come up with something nicer.

-- 
Stefano