[dts] [PATCH V1] tests/pmdrss_hash:change hard code
Rami Rosen
ramirose at gmail.com
Fri Mar 15 08:30:19 CET 2019
> - self.dut.send_expect("./%s/app/testpmd -c f -n 4 -- -i" %
self.target,
> "testpmd> ", 120)
> + self.dut.send_expect("./%s/app/testpmd -c %s -n 4 -- -i" %
> + (self.target, self.coremask), "testpmd> ", 120)
[Lijuan] Do you forget to change hardcode of memory channel?
> out = self.dut.send_expect("create bonded device 3 0", "testpmd>
", 30)
> bond_device_id = int(re.search("port \d+", out).group().split("
")[-1].strip())
Hi all,
Good point.
Actually, the number of channels is a non mandatory EAL parameter, and in
DTS it defaults to 1.
http://git.dpdk.org/tools/dts/tree/framework/dut.py#n565
In this module, the other invocations of testpmd were indeed with
get_memory_channels().
In this occurrence it is hardcoded to 4.
The number of memory channels is merely an optimization, as you can see in
the following thread:
https://dev.dpdk.narkive.com/78gtSXha/dpdk-dev-patch-eal-default-to-one-memory-channel-if-not-specified
According to it,
...
Obtaining the correct value of memory channels, especially from a
running system, can be anything from difficult to plain impossible.
...
Not sure what was the reason for hard coding it to 4 in this test case, but
anyhow on this occasion maybe it is better to replace it with
get_memory_channels() as with the rest of the module, for the sake of
consistency.
Regards,
Rami Rosen
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mails.dpdk.org/archives/dts/attachments/20190315/7067705f/attachment.html>
More information about the dts
mailing list