* [BUG] ath12k: CSA with blocktx disrupts TX on other bands
@ 2026-02-04 16:31 Remi Pommarel
0 siblings, 0 replies; only message in thread
From: Remi Pommarel @ 2026-02-04 16:31 UTC (permalink / raw)
To: ath12k; +Cc: Remi Pommarel
Hi,
With the single phy changes introduced for MLO support, both 2.4 GHz
and 5 GHz interfaces on the same dual-band chip now share the same
hardware queues, even when operating on separate bands.
As a result, when a Channel Switch Announcement (CSA) with blocktx
is triggered on one interface (e.g., the 5 GHz vif),
ieee80211_vif_block_queues_csa() pauses the shared HW queues. This
unintentionally blocks TX traffic on the other interface (e.g., the
2.4 GHz vif), despite it being on a different band.
I haven't tested it, but the same issue is expected to affect a single
MLO AP setup, where a CSA could inadvertently pause transmission across
all links.
To address this class of problem, the IEEE80211_HW_HANDLES_QUIET_CSA
flag was introduced. When set, it delegates queue pausing to the
firmware, avoiding unnecessary disruptions in multi-vif or MLO
scenarios. This approach is already used in iwlwifi/mvm, as seen in
commit bbd6d0f8bc51 ("wifi: iwlwifi: mvm: advertise
IEEE80211_HW_HANDLES_QUIET_CSA").
Thus, does the current ath12k firmware properly handle TX quiescing
allowing IEEE80211_HW_HANDLES_QUIET_CSA to be set ? Or does it need
to be handle at the driver level (likewise what has been done in
iwlwifi/mvm in the commit above) ?
Thanks.
--
Remi
^ permalink raw reply [flat|nested] only message in thread
only message in thread, other threads:[~2026-02-04 16:49 UTC | newest]
Thread overview: (only message) (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2026-02-04 16:31 [BUG] ath12k: CSA with blocktx disrupts TX on other bands Remi Pommarel
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox