Bug #6250
open
libbpf: elf: legacy map definitions in 'maps' section are not supported by libbpf v1.0+
Added by Vincent Li over 1 year ago.
Updated over 1 year ago.
Description
Hi
I followed https://docs.suricata.io/en/latest/capture-hardware/ebpf-xdp.html to compile most recent github libbpf and suricata. when I run:
suricata -c ./suricata-ebpf.yaml --pidfile /var/run/suricata.pid --af-packet -vvv
I got:
libbpf: elf: legacy map definitions in 'maps' section are not supported by libbpf v1.0+
Error: ebpf: Unable to load eBPF objects in '/usr/libexec/suricata/ebpf/xdp_filter.bpf': Operation not supported [EBPFLoadFile:util-ebpf.c:342]
Warning: af-packet: enp4s0f0: failed to load XDP filter file [ParseAFPConfig:runmode-af-packet.c:579]
libbpf: elf: legacy map definitions in 'maps' section are not supported by libbpf v1.0+
Error: ebpf: Unable to load eBPF objects in '/usr/libexec/suricata/ebpf/xdp_filter.bpf': Operation not supported [EBPFLoadFile:util-ebpf.c:342]
Warning: af-packet: enp4s0f1: failed to load XDP filter file [ParseAFPConfig:runmode-af-packet.c:579]
I think the legacy eBPF map definition needs to be updated to be BTF defined eBPF map, I could send up PR to update the map definition. what do you think?
Here is my suricata-ebpf.yaml config:
af-packet:
- interface: enp4s0f0
# Number of receive threads. "auto" uses the number of cores
#threads: auto
threads: 1
# Default clusterid. AF_PACKET will load balance packets based on flow.
cluster-id: 99 cluster-type: cluster_flow
# In some fragmentation cases, the hash can not be computed. If "defrag" is set
# to yes, the kernel will do the needed defragmentation before sending the packets.
defrag: yes
copy-mode: tap
copy-iface: enp4s0f1
xdp-mode: driver
pinned-maps: true
pinned-maps-name: flow_table_v4
xdp-filter-file: /usr/libexec/suricata/ebpf/xdp_filter.bpf
#bypass: yes
use-mmap: yes
ring-size: 200000
# For eBPF and XDP setup including bypass, filter and load balancing, please
# see doc/userguide/capture-hardware/ebpf-xdp.rst for more info.
# Put default values here. These will be used for an interface that is not
# in the list above.
- interface: enp4s0f1
threads: 1
cluster-id: 100
cluster-type: cluster_flow
defrag: yes
copy-mode: ips
copy-iface: enp4s0f0
xdp-mode: driver
pinned-maps: true pinned-maps-name: flow_table_v4
xdp-filter-file: /usr/libexec/suricata/ebpf/xdp_filter.bpf
#bypass: yes
use-mmap: yes
ring-size: 200000
Also available in: Atom
PDF