Bug ID 1602
Summary igc: hardware timestamp is always zero with RTE_ETH_RX_OFFLOAD_TIMESTAMP
Product DPDK
Version 23.11
Hardware x86
OS Linux
Status UNCONFIRMED
Severity normal
Priority Normal
Component ethdev
Assignee dev@dpdk.org
Reporter rpf@mail.ustc.edu.cn
Target Milestone ---

Created attachment 298 [details]
Patch for rxtx_callbacks example to show hardware timestamp

I tried to use RTE_ETH_RX_OFFLOAD_TIMESTAMP with timestamp dynfield to get a
hardware timestamp on every packet. But it is always zero despite 
RTE_ETH_RX_OFFLOAD_TIMESTAMP is supported for igc and enabled.

I encountered this problem in my own project. But this problem could be
reproduced with PktGen-DPDK and the rxtx_callbacks example.

My setup contains two ports (02:00.0 and 03:00.0) connected together on the
same machine:

I am generating traffic with pktgen-dpdk on port 02:00.0:

sudo env LD_LIBRARY_PATH=/opt/dpdk24/lib ./usr/local/bin/pktgen  -l 0,1 -n 4 -b
03:00.0 --file-prefix pktgen -- -P -m "[1].0" -s 0:traces/small_flows.pcap 

And receiving them on port 03:00.0:

sudo ./examples/dpdk-rxtx_callbacks -c 2 -n 3 -b 02:00.0  -- -t

The rxtx_callbacks examples is modified with the attached patch to show hw ts
in rx callbacks.

But for all the timestamps I get are zeros.

I am using uio_pci_generic with Intel Corporation Ethernet Controller I226-V
(rev 04).

I am aware that there's some problem with hardware timestamp for igc and has
applied this patch to my dpdk installation:
https://lore.kernel.org/dpdk-dev/00f78eba-8eff-4c67-8ecf-3bef5f25b9f6@allegro-packets.com/#t

Please kindly ask if more information is needed to debug it. Thanks!
          


You are receiving this mail because: