Re: [PATCH 3/4] tcp, tcp_splice: CONN_IDX subtraction of pointers isn't always long

[David Gibson <david@gibson.dropbear.id.au>]

[-- Attachment #1: Type: text/plain, Size: 2135 bytes --]

On Wed, Nov 29, 2023 at 02:58:42PM +0100, Stefano Brivio wrote:
> On Wed, 29 Nov 2023 14:46:09 +0100
> Stefano Brivio <sbrivio@redhat.com> wrote:
> 
> > On 32-bit architectures, it's a regular int. C99 introduced ptrdiff_t
> > for this case, with a matching length modifier, 't'.
> > 
> > Signed-off-by: Stefano Brivio <sbrivio@redhat.com>
> > ---
> >  tcp.c        | 39 +++++++++++++++++++++------------------
> >  tcp_splice.c | 14 +++++++-------
> >  2 files changed, 28 insertions(+), 25 deletions(-)
> > 
> > diff --git a/tcp.c b/tcp.c
> > index 44468ca..c32c9cb 100644
> > --- a/tcp.c
> > +++ b/tcp.c
> > @@ -727,7 +727,7 @@ static void tcp_timer_ctl(const struct ctx *c, struct tcp_tap_conn *conn)
> >  		it.it_value.tv_sec = ACT_TIMEOUT;
> >  	}
> >  
> > -	debug("TCP: index %li, timer expires in %lu.%03lus", CONN_IDX(conn),
> > +	debug("TCP: index %ti, timer expires in %lu.%03lus", CONN_IDX(conn),
> >
> > [...]
> 
> Oops, I just realised this clashes with your "[PATCH v2 03/11] flow,
> tcp: Consolidate flow pointer<->index helpers".

And then a bunch will be obsoleted by "flow, tcp: Add logging helpers
for connection related messages".

> There, however, I guess that the new flow_idx() should return ptrdiff_t,
> which is signed.

Actually, no, I don't think so.  Yes the expression that generates it
is naturally of type ptrdiff_t.  But it's a bug to call flow_idx() on
something not in the flow table, and places where we want to pass *in*
a flow table index it makes more sense for it to be unsigned.  So I
think flow indices should be unsigned throughout.

> I can drop this patch if you re-spin it (assuming it makes sense to
> you), or I can adapt it on top of your patch -- whatever is most
> convenient for you.

I have a couple of reasons to re-spin anyway.  So how about you drop
this, and I'll double check that I get the format specifiers sane
after my series?

-- 
David Gibson			| I'll have my music baroque, and my code
david AT gibson.dropbear.id.au	| minimalist, thank you.  NOT _the_ _other_
				| _way_ _around_!
http://www.ozlabs.org/~dgibson

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 833 bytes --]