Skip to content

Add non-unicast noise filter to LossyQueueTest ingress drop checks#23340

Open
XuChen-MSFT wants to merge 1 commit intosonic-net:masterfrom
XuChen-MSFT:xuchen3/master/fix-lossy-queue-noise
Open

Add non-unicast noise filter to LossyQueueTest ingress drop checks#23340
XuChen-MSFT wants to merge 1 commit intosonic-net:masterfrom
XuChen-MSFT:xuchen3/master/fix-lossy-queue-noise

Conversation

@XuChen-MSFT
Copy link
Contributor

Description of PR

Extend the ignore_ingress_drop_caused_by_nonunicast_noise() function (introduced in PR #22871 for PFCtest/PFCXonTest) to also cover LossyQueueTest.

PR #22871 added non-unicast noise filtering to PFCtest and PFCXonTest, but missed LossyQueueTest. Environmental broadcast/multicast traffic causes false positive ingress drop failures in LossyQueueTest with the same root cause.

Type of change

  • Bug fix

Back port request

  • 202205
  • 202305
  • 202311
  • 202405
  • 202411
  • 202505
  • 202511

Approach

What is the motivation for this PR?

LossyQueueTest has the same false positive ingress drop issue as PFCtest/PFCXonTest. Environmental broadcast/multicast traffic (ARP, IPv6 NS/RA) arrives during testing and causes InDiscard counter increases unrelated to test traffic. PR #22871 fixed this for PFCtest/PFCXonTest but did not cover LossyQueueTest.

Kusto analysis shows LossyQueueTest still fails at 33% rate on SONiC.20251110.x across multiple platforms (Cisco-8101, Arista-7060X6, etc.) even after PR #22871.

How did you do it?

Applied the same ignore_ingress_drop_caused_by_nonunicast_noise() wrapper to both ingress drop check points in LossyQueueTest:

  1. Short-of-drop phase (no ingress drop expected)
  2. Post-excess-traffic phase (no ingress drop expected on non-DNX platforms)

How did you verify/test it?

Code review. Same pattern as PR #22871 which was verified on Cisco-8000 chassis topology.

Any platform specific information?

Counter margin (COUNTER_MARGIN) applied for broadcom-dnx, cisco-8000, and marvell-teralynx platforms.

ADO PBIs: #37142761

Extend the ignore_ingress_drop_caused_by_nonunicast_noise() function
(introduced in PR sonic-net#22871 for PFCtest/PFCXonTest) to also cover
LossyQueueTest. Environmental broadcast/multicast traffic causes false
positive ingress drop failures in LossyQueueTest, same root cause as
the PFC tests.

Applied to both ingress drop check points in LossyQueueTest:
1. Short-of-drop phase: no ingress drop expected
2. Post-excess-traffic phase: no ingress drop expected (non-DNX)

Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
@mssonicbld
Copy link
Collaborator

/azp run

@azure-pipelines
Copy link

Azure Pipelines successfully started running 1 pipeline(s).

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants