Re: [dts] TestSuite_l2fwd.py questions

["Liu, Yong" <yong.liu@intel.com>]

Hi Angela,
L2fwd case was designed few years ago. At that time we forced l2fwd case using different ports on different cards for maximum throughput.
Currently all performance reports are based on l3fwd sample, so you can look at L3fwd test case for replacement.  

For the mismatched expected output, that maybe one bug. We haven't run l2fwd performance case for a long time.

Thanks,
Marvin

> -----Original Message-----
> From: dts [mailto:dts-bounces@dpdk.org] On Behalf Of Angela Czubak
> Sent: Tuesday, April 04, 2017 12:11 AM
> To: dts@dpdk.org
> Subject: [dts] TestSuite_l2fwd.py questions
> 
> Hi,
> 
> I was looking into L2fwd test case and was wondering if getting
> available ports (self.dut_ports = self.dut.get_ports_performance() in
> set_up_all() method) and leaving the default for force_different_nic
> argument was on purpose. I mean, was this test designed to run for two
> different NICs only?
> 
> One more question: why does test_l2fwd integrity run l2fwd and expect
> "memory mapped" in the output, whereas others seem to expect "L2FWD:
> entering main loop"? Is that a bug?
> 
> Regards,
> 
> Angela Czubak