Do not apply this version of the patch: I just discovered it's
entirely breaking ping.  Investigating...


On Wed, Jul 05, 2023 at 11:04:27AM +1000, David Gibson wrote:
> On Tue, Jul 04, 2023 at 01:21:04PM +0200, Stefano Brivio wrote:
> > On Tue,  4 Jul 2023 14:36:23 +1000
> > David Gibson <david@gibson.dropbear.id.au> wrote:
> > 
> > > We don't handle defragmentation of IP packets coming from the tap side,
> > > and we're unlikely to any time soon (with our large MTU, it's not useful
> > > for practical use cases).  Currently, however, we simply ignore the
> > > fragmentation flags and treat fragments as though they were whole IP
> > > packets.  This isn't ideal and can lead to rather cryptic behaviour if we
> > > do receive IP fragments.
> > > 
> > > Change the code to explicitly drop fragmented packets, and print a rate
> > > limited warning if we do encounter them.
> > > 
> > > Bugzilla: https://bugs.passt.top/show_bug.cgi?id=62
> > 
> > By the way, I silently replaced those with "Link:" in the past, just in
> > case we want to automate something around it one day, to avoid
> > differences between references to different bug trackers.
> 
> Oh, ok.  I'll keep that in mind.
> 
> > Once upon a time, I wrote some scripting to automatically link HTML
> > reports with (Linux kernel) commits to bug trackers, and it was quite
> > painful to discover all possible spellings of "Bugzilla" plus a few
> > others, hence my thought. But let me know if something speaks against
> > this.
> 
> No, that makes sense.
> 
> > > Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
> > > ---
> > >  tap.c | 31 +++++++++++++++++++++++++++++++
> > >  1 file changed, 31 insertions(+)
> > > 
> > > diff --git a/tap.c b/tap.c
> > > index e3235299..2e6939fa 100644
> > > --- a/tap.c
> > > +++ b/tap.c
> > > @@ -62,6 +62,7 @@ static PACKET_POOL_NOINIT(pool_tap4, TAP_MSGS, pkt_buf);
> > >  static PACKET_POOL_NOINIT(pool_tap6, TAP_MSGS, pkt_buf);
> > >  
> > >  #define TAP_SEQS		128 /* Different L4 tuples in one batch */
> > > +#define FRAGMENT_MSG_RATE	10  /* # seconds between fragment warnings */
> > >  
> > >  /**
> > >   * tap_send() - Send frame, with qemu socket header if needed
> > > @@ -543,6 +544,32 @@ static void tap_packet_debug(const struct iphdr *iph,
> > >  	}
> > >  }
> > >  
> > > +/**
> > > + * tap4_is_fragment() - Determine if a packet is an IP fragment
> > 
> > This is actually independent from the "tap" "side", it could also be
> > e.g. ipv4_is_fragment(), in util.c. Not a strong preference though, I
> > guess we can also merge it as it is.
> 
> Well.. the detection of fragments is independent, but the warning
> message is specific to tap.  I'm inclined to leave it as is, at least
> until we have a need for this logic somewhere else, at which point we
> can refactor.
> 
> > 
> > > + * @iph:	IPv4 header (length already validated)
> > > + * @now:	Current timestamp
> > > + *
> > > + * Return: true if iph is an IP fragment, false otherwise
> > > + */
> > > +static bool tap4_is_fragment(const struct iphdr *iph,
> > > +			     const struct timespec *now)
> > > +{
> > > +	if (iph->frag_off & ~IP_DF) {
> > > +		/* Ratelimit messages */
> > > +		static time_t last_message;
> > > +		static unsigned num_dropped;
> > > +
> > > +		num_dropped++;
> > > +		if (now->tv_sec - last_message > FRAGMENT_MSG_RATE) {
> > > +			warn("Can't process IPv4 fragments (%lu dropped)", num_dropped);
> > > +			last_message = now->tv_sec;
> > > +			num_dropped = 0;
> > > +		}
> > > +		return true;
> > > +	}
> > > +	return false;
> > > +}
> > > +
> > >  /**
> > >   * tap4_handler() - IPv4 and ARP packet handler for tap file descriptor
> > >   * @c:		Execution context
> > > @@ -591,6 +618,10 @@ resume:
> > >  		    hlen > l3_len)
> > >  			continue;
> > >  
> > > +		/* We don't handle IP fragments, drop them */
> > > +		if (tap4_is_fragment(iph, now))
> > > +			continue;
> > > +
> > >  		l4_len = l3_len - hlen;
> > >  
> > >  		if (iph->saddr && c->ip4.addr_seen.s_addr != iph->saddr) {
> > 
> 



-- 
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