ee3fd1b982
This patch implements a generic approach to extract fields from the packet's header and copying them to packet metadata. The fields are selected at the desired offset on the basis of the mask specified in application configuration file. The extracted fields, for instance, can be used to compute hash for the lookup table. This feature exposes more flexibility to the users as they will be able to employ new protocol headers and specify the required fields to be extracted. The above feature has been implemented as port_in action handler of the passthrough pipeline. The example of the configuration file for passthrough pipeline is as below; [PIPELINE1] type = PASS-THROUGH core = 1 pktq_in = RXQ0.0 RXQ1.0 RXQ2.0 RXQ3.0 pktq_out = TXQ0.0 TXQ1.0 TXQ2.0 TXQ3.0 dma_size = 16 dma_dst_offset = 64 dma_src_offset = 150 dma_src_mask = 00FF0000FFFFFFFFFFFFFFFFFFFFFFFF dma_hash_offset = 80 Signed-off-by: Jasvinder Singh <jasvinder.singh@intel.com> Acked-by: Cristian Dumitrescu <cristian.dumitrescu@intel.com> |
||
---|---|---|
.. | ||
bond | ||
cmdline | ||
distributor | ||
dpdk_qat | ||
exception_path | ||
helloworld | ||
ip_fragmentation | ||
ip_pipeline | ||
ip_reassembly | ||
ipv4_multicast | ||
kni | ||
l2fwd | ||
l2fwd-crypto | ||
l2fwd-ivshmem | ||
l2fwd-jobstats | ||
l2fwd-keepalive | ||
l3fwd | ||
l3fwd-acl | ||
l3fwd-power | ||
l3fwd-vf | ||
link_status_interrupt | ||
load_balancer | ||
multi_process | ||
netmap_compat | ||
packet_ordering | ||
ptpclient | ||
qos_meter | ||
qos_sched | ||
quota_watermark | ||
rxtx_callbacks | ||
skeleton | ||
tep_termination | ||
timer | ||
vhost | ||
vhost_xen | ||
vm_power_manager | ||
vmdq | ||
vmdq_dcb | ||
Makefile |