From: Muna Sinada muna.sinada@oss.qualcomm.com
[ Upstream commit 1a4a6a22552ca9d723f28a1fe35eab1b9b3d8b33 ]
Currently for MLO, sending out multicast frames on each link is handled by mac80211 only when IEEE80211_HW_MLO_MCAST_MULTI_LINK_TX flag is not set.
Dynamic VLAN multicast traffic utilizes software encryption. Due to this, mac80211 should handle transmitting multicast frames on all links for multicast VLAN traffic.
Signed-off-by: Muna Sinada muna.sinada@oss.qualcomm.com Link: https://patch.msgid.link/20250325213125.1509362-4-muna.sinada@oss.qualcomm.c... [remove unnecessary parentheses] Signed-off-by: Johannes Berg johannes.berg@intel.com Signed-off-by: Sasha Levin sashal@kernel.org ---
**YES** This commit should be backported to stable kernel trees. Here's my detailed analysis: ## Backporting Analysis: YES ### Nature of the Change This is a **targeted bug fix** that addresses a functional gap in MLO (Multi-Link Operation) multicast handling for AP_VLAN interfaces. The commit extends existing logic rather than introducing new features. ### Technical Assessment **The Problem Fixed:** The original code in `ieee80211_subif_start_xmit()` only handled `NL80211_IFTYPE_AP` interfaces for MLO multicast transmission, but completely missed `NL80211_IFTYPE_AP_VLAN` interfaces. This created a behavioral inconsistency where: - AP interfaces: Properly handled MLO multicast transmission via `ieee80211_mlo_multicast_tx()` - AP_VLAN interfaces: Fell through to normal transmission path, bypassing MLO multicast coordination **The Fix Applied:** ```c } else if (ieee80211_vif_is_mld(&sdata->vif) && - sdata->vif.type == NL80211_IFTYPE_AP && - !ieee80211_hw_check(&sdata->local->hw, MLO_MCAST_MULTI_LINK_TX)) { + ((sdata->vif.type == NL80211_IFTYPE_AP && + !ieee80211_hw_check(&sdata->local->hw, MLO_MCAST_MULTI_LINK_TX)) || + (sdata->vif.type == NL80211_IFTYPE_AP_VLAN && + !sdata->wdev.use_4addr))) { ``` This change: 1. **Extends existing conditional logic** to include AP_VLAN interfaces 2. **Uses existing infrastructure** (`ieee80211_mlo_multicast_tx()`) 3. **Adds proper validation** (excludes 4-address/WDS mode via `!use_4addr`) ### Risk Analysis: LOW RISK **Minimal Change Scope:** - Single conditional modification in well-established code path - No new functions, data structures, or APIs introduced - Uses existing MLO multicast transmission infrastructure **Well-Contained Impact:** - Only affects MLO multicast transmission for AP_VLAN interfaces - Preserved existing behavior for all other interface types - Added validation ensures appropriate application scope **Backward Compatibility:** - No changes to external interfaces or user-visible behavior - No impact on non-MLO or non-VLAN configurations - Maintains existing AP interface behavior unchanged ### Comparison with Similar Commits This commit aligns with **Similar Commit #4** (MLD address translation fix - Status: YES): - Both are targeted bug fixes for MLO functionality - Both address specific functional gaps rather than adding new features - Both use existing infrastructure to correct behavioral issues - Both have minimal risk profiles suitable for stable backporting **Contrasts with rejected commits:** - Unlike Similar Commits #1, #2: This is not implementing new MLO features - Unlike Similar Commit #3: This is not changing encapsulation behavior - Unlike Similar Commit #5: This is not removing functionality ### Production Impact **Functional Benefits:** - **Fixes MLO+VLAN deployments**: Enables proper multicast operation in combined MLO/VLAN environments - **Consistency**: Ensures uniform multicast handling across AP and AP_VLAN interface types - **Reliability**: Prevents potential multicast delivery issues in enterprise VLAN configurations **Real-World Relevance:** - AP_VLAN interfaces are commonly used in enterprise environments for traffic segmentation - MLO is increasingly deployed in Wi-Fi 7 scenarios - The combination creates practical deployment scenarios where this fix is needed ### Stable Tree Criteria Compliance ✅ **Important bug fix**: Addresses functional gap affecting MLO+VLAN deployments ✅ **Minimal risk**: Single conditional extension with clear validation ✅ **No architectural changes**: Uses existing MLO multicast infrastructure ✅ **Well-contained**: Affects only specific interface type in specific scenario ✅ **Production relevant**: Impacts real-world enterprise VLAN configurations This commit represents a clear bug fix that resolves a functional inconsistency in MLO multicast handling, making it well-suited for stable kernel backporting according to established criteria.
net/mac80211/tx.c | 6 ++++-- 1 file changed, 4 insertions(+), 2 deletions(-)
diff --git a/net/mac80211/tx.c b/net/mac80211/tx.c index 45a093d3f1fa7..ec5469add68a2 100644 --- a/net/mac80211/tx.c +++ b/net/mac80211/tx.c @@ -4507,8 +4507,10 @@ netdev_tx_t ieee80211_subif_start_xmit(struct sk_buff *skb, IEEE80211_TX_CTRL_MLO_LINK_UNSPEC, NULL); } else if (ieee80211_vif_is_mld(&sdata->vif) && - sdata->vif.type == NL80211_IFTYPE_AP && - !ieee80211_hw_check(&sdata->local->hw, MLO_MCAST_MULTI_LINK_TX)) { + ((sdata->vif.type == NL80211_IFTYPE_AP && + !ieee80211_hw_check(&sdata->local->hw, MLO_MCAST_MULTI_LINK_TX)) || + (sdata->vif.type == NL80211_IFTYPE_AP_VLAN && + !sdata->wdev.use_4addr))) { ieee80211_mlo_multicast_tx(dev, skb); } else { normal: