Re: [PATCHv10 bpf-next 2/4] xdp: extend xdp_redirect_map with broadcast support

On Mon, Apr 26, 2021 at 06:47:17PM +0800, Hangbin Liu wrote:

On Mon, Apr 26, 2021 at 11:53:50AM +0200, Jesper Dangaard Brouer wrote:

quoted

On Fri, 23 Apr 2021 10:00:17 +0800
Hangbin Liu [off-list ref] wrote:

quoted

This patch adds two flags BPF_F_BROADCAST and BPF_F_EXCLUDE_INGRESS to
extend xdp_redirect_map for broadcast support.

With BPF_F_BROADCAST the packet will be broadcasted to all the interfaces
in the map. with BPF_F_EXCLUDE_INGRESS the ingress interface will be
excluded when do broadcasting.

When getting the devices in dev hash map via dev_map_hash_get_next_key(),
there is a possibility that we fall back to the first key when a device
was removed. This will duplicate packets on some interfaces. So just walk
the whole buckets to avoid this issue. For dev array map, we also walk the
whole map to find valid interfaces.

Function bpf_clear_redirect_map() was removed in
commit ee75aef23afe ("bpf, xdp: Restructure redirect actions").
Add it back as we need to use ri->map again.

Here is the performance result by using 10Gb i40e NIC, do XDP_DROP on
veth peer, run xdp_redirect_{map, map_multi} in sample/bpf and send pkts
via pktgen cmd:
./pktgen_sample03_burst_single_flow.sh -i eno1 -d $dst_ip -m $dst_mac -t 10 -s 64

While running:
 $ sudo ./xdp_redirect_map_multi -F i40e2 i40e2
 Get interfaces 7 7
 libbpf: elf: skipping unrecognized data section(23) .eh_frame
 libbpf: elf: skipping relo section(24) .rel.eh_frame for section(23) .eh_frame
 Forwarding   10140845 pkt/s
 Forwarding   11767042 pkt/s
 Forwarding   11783437 pkt/s
 Forwarding   11767331 pkt/s

When starting:  sudo ./xdp_monitor --stats

That seems the same issue I reported previously in our meeting.
https://bugzilla.redhat.com/show_bug.cgi?id=1906820#c4

I only saw it 3 times and can't reproduce it easily.

Do you have any idea where is the root cause?

OK, I just re-did the test and could reproduce it now.
Maybe because the code changed and it's easy to reproduce now.

I will check this issue.

Thanks
Hangbin