ignore ingress drop caused by non-unicast noise#22871
Merged
StormLiangMS merged 1 commit intosonic-net:masterfrom Mar 13, 2026
Merged
ignore ingress drop caused by non-unicast noise#22871StormLiangMS merged 1 commit intosonic-net:masterfrom
StormLiangMS merged 1 commit intosonic-net:masterfrom
Conversation
Collaborator
|
/azp run |
|
Azure Pipelines successfully started running 1 pipeline(s). |
more and more qos test break by environment noise: unexpected non-unicast traffic caused ingress drop. This PR adds a helper function to detect and ignore ingress drops caused by broadcast/multicast noise, preventing unnecessary test failures. Signed-off-by: Xu Chen <[email protected]>
XuChen-MSFT
force-pushed
the
xuchen3/master/ignore-ingress-drop
branch
from
5aad431 to
bb52178
Compare
Collaborator
|
/azp run |
|
Azure Pipelines successfully started running 1 pipeline(s). |
XuChen-MSFT
added
Included in 202505 branch
Included in 202511 branch
Request for 202505 branch
Request for 202511 branch
mssonicbld
pushed a commit
to mssonicbld/sonic-mgmt
that referenced
this pull request
Mar 13, 2026
more and more qos test break by environment noise: unexpected non-unicast traffic caused ingress drop. This PR adds a helper function to detect and ignore ingress drops caused by broadcast/multicast noise, preventing unnecessary test failures. Signed-off-by: Xu Chen <[email protected]> Signed-off-by: mssonicbld <[email protected]>
Collaborator
|
Cherry-pick PR to 202511: #22934 |
12 tasks
mssonicbld
pushed a commit
that referenced
this pull request
Mar 13, 2026
more and more qos test break by environment noise: unexpected non-unicast traffic caused ingress drop. This PR adds a helper function to detect and ignore ingress drops caused by broadcast/multicast noise, preventing unnecessary test failures. Signed-off-by: Xu Chen <[email protected]> Signed-off-by: mssonicbld <[email protected]>
mssonicbld
added
Included in 202511 branch
and removed
Created PR to 202511 branch
labels
selldinesh
pushed a commit
to selldinesh/sonic-mgmt
that referenced
this pull request
Mar 16, 2026
more and more qos test break by environment noise: unexpected non-unicast traffic caused ingress drop. This PR adds a helper function to detect and ignore ingress drops caused by broadcast/multicast noise, preventing unnecessary test failures. Signed-off-by: Xu Chen <[email protected]> Signed-off-by: selldinesh <[email protected]>
abhishek-nexthop
pushed a commit
to nexthop-ai/sonic-mgmt
that referenced
this pull request
Mar 17, 2026
more and more qos test break by environment noise: unexpected non-unicast traffic caused ingress drop. This PR adds a helper function to detect and ignore ingress drops caused by broadcast/multicast noise, preventing unnecessary test failures. Signed-off-by: Xu Chen <[email protected]> Signed-off-by: Abhishek <[email protected]>
vrajeshe
pushed a commit
to vrajeshe/sonic-mgmt
that referenced
this pull request
Mar 23, 2026
more and more qos test break by environment noise: unexpected non-unicast traffic caused ingress drop. This PR adds a helper function to detect and ignore ingress drops caused by broadcast/multicast noise, preventing unnecessary test failures. Signed-off-by: Xu Chen <[email protected]> Signed-off-by: Venkata Gouri Rajesh Etla <[email protected]>
XuChen-MSFT
added a commit
to XuChen-MSFT/sonic-mgmt
that referenced
this pull request
Mar 26, 2026
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 <[email protected]>
8 tasks
ravaliyel
pushed a commit
to ravaliyel/sonic-mgmt
that referenced
this pull request
Mar 27, 2026
more and more qos test break by environment noise: unexpected non-unicast traffic caused ingress drop. This PR adds a helper function to detect and ignore ingress drops caused by broadcast/multicast noise, preventing unnecessary test failures. Signed-off-by: Xu Chen <[email protected]>
XuChen-MSFT
added a commit
to XuChen-MSFT/sonic-mgmt
that referenced
this pull request
Mar 27, 2026
PR sonic-net#22871 added noise filtering to PFCtest for "short of PFC trigger" and "after trigger PFC" phases, but missed the "short of ingress drop" phase (line 2474-2483). This causes false positive test failures when environmental broadcast/multicast traffic arrives during this phase. Observed on Arista-7260CX3-C64 (broadcom, T1) with SONiC.20251110.17: - Phase 1 "short of PFC trigger": InDiscard Delta=1, noise filter correctly detected InNonUcPkt increase and IGNORED ✅ - Phase 3 "short of ingress drop": InDiscard Delta=1, but NO noise filter → assertion failed ❌ Fix: Apply the same ignore_ingress_drop_caused_by_nonunicast_noise() wrapper to the "short of ingress drop" phase ingress counter check. Co-authored-by: Copilot <[email protected]>
2 tasks
XuChen-MSFT
added a commit
to XuChen-MSFT/sonic-mgmt
that referenced
this pull request
Mar 27, 2026
…ssyQueueTest PR sonic-net#22871 added noise filtering but missed some ingress drop check points: 1. PFCtest "short of ingress drop" phase (was line 2474-2483): Observed on Arista-7260CX3-C64 (broadcom, T1) with SONiC.20251110.17. Phase 1 noise was correctly IGNORED, but Phase 3 had no filter. InDiscard Delta=1, InNonUcPkt Delta=1 — single broadcast packet. 2. LossyQueueTest 2nd ingress check (after "trigger egress drop"): Same testbed, same test plan. InDiscard Delta=1 at line 5106. Non-DNX platforms go through else branch with no noise filter. Fix: Apply ignore_ingress_drop_caused_by_nonunicast_noise() to both missed check points. Co-authored-by: Copilot <[email protected]>
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
3 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Description of PR
Summary:
More and more QoS tests break by environment noise: unexpected non-unicast traffic caused ingress drop counters to increase, leading to false positive test failures.
This PR adds a helper function
ignore_ingress_drop_caused_by_nonunicast_noiseto detect and ignore ingress drops caused by broadcast/multicast noise, preventing unnecessary test failures.Type of change
Back port request
Approach
What is the motivation for this PR?
Non-unicast environmental noise (broadcast/multicast traffic) causes unexpected ingress drop counter increases on DUT receive ports. This leads to false positive failures in PFC/QoS SAI tests (PFCXonTest, PFCtest, etc.), even when the DUT behavior is correct.
How did you do it?
ignore_ingress_drop_caused_by_nonunicast_noise()function that checks if ingress drop increase correlates with non-unicast packet counter increase on the same portHow did you verify/test it?
Verified on internal testbed via ADO PR #18719 (approved by Storm Liang and Austin Pham). The change was tested on chassis topology with Cisco-8000 platform.
Any platform specific information?
Counter margin (COUNTER_MARGIN) is applied for broadcom-dnx, cisco-8000, and marvell-teralynx platforms due to known extra IPv6 NS/RA packets from VM.
Supported testbed topology if it's a new test case?
N/A - this is a bug fix to existing test cases.
Documentation
N/A