[dts] [PATCH V1] tests/ipfrag:Modify the way to get the core.

Zhu, ShuaiX shuaix.zhu at intel.com
Thu Jan 24 03:32:35 CET 2019


I think that using the all parameter should be possible in the current code. Some platforms have fewer core values than fixed values. For example, using 2S/4C/2T, the following error occurs: Cannot get requested core configuration requested 2S/4C/2T have.

From: Rami Rosen [mailto:ramirose at gmail.com]
Sent: Thursday, January 24, 2019 9:54 AM
To: Zhu, ShuaiX <shuaix.zhu at intel.com>
Cc: dts at dpdk.org
Subject: Re: [dts] [PATCH V1] tests/ipfrag:Modify the way to get the core.

Hi, Zhu,

> Some platforms have a core number less than a fixed value.

Thanks for this patch. You are right, obviously there are platforms with other combinations than
"2S/2C/2T", and actually I myself run DTS on diifferent platforms than this and had some issues related to it with various DTS test modules.

 I have a question, though,if I may:
On which platform ( i mean which combination of socket/core/thread in that platform) did the solution of using cores = self.dut.get_core_listst("all") instead of the  hard coded "2S/2C/2T" worked  for you for this ipfrag Test Module?

And a general question to DTS developers/maintainer:  all the DTS test suites use hard coded values for the S/C/T combination. And obviously this will fail on certain platforms. Wouldn't it be reasonable to replace all these occurrences with self.dut.get_core_listst("all") ? Will self.dut.get_core_listst("all") always work, for any combination of socket/core/thread ? Will this work for 2S/4C/2T ?

Regards,
Rami Rosen
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mails.dpdk.org/archives/dts/attachments/20190124/00bb3557/attachment.html>


More information about the dts mailing list