Linux wireless drivers development
 help / color / mirror / Atom feed
From: Priyansha Tiwari <priyansha.tiwari@oss.qualcomm.com>
To: johannes@sipsolutions.net
Cc: linux-wireless@vger.kernel.org, quic_drohan@quicinc.com
Subject: [PATCH wireless-next v3 0/4] wifi: nl80211: introduce PROBE_PEER for AP and STA
Date: Thu,  7 May 2026 22:45:59 +0530	[thread overview]
Message-ID: <20260507171603.1007812-1-pritiwa@qti.qualcomm.com> (raw)

From: Priyansha Tiwari <priyansha.tiwari@oss.qualcomm.com>

This series introduces a unified mechanism to probe connected peers.
It generalizes the legacy AP-only PROBE_CLIENT functionality by adding
NL80211_CMD_PROBE_PEER and enabling (feature-gated) STA-side probing.
With this, AP/GO continues to probe associated stations as before, and
STA/P2P-client can probe the connected AP for faster link health checks.
For MLO connections, mac80211 supports per-link STA probing to obtain
link-specific ACK information.

Patch 1 renames NL80211_CMD_PROBE_CLIENT to NL80211_CMD_PROBE_PEER in
the UAPI enum (keeping PROBE_CLIENT as a compatibility alias) and renames
the .probe_client cfg80211_ops callback to .probe_peer. All in-tree users
(wil6210, mwifiex, mac80211) are updated so the tree builds after this
patch. This is a pure rename with no behaviour change; documentation is
intentionally left unchanged.

Patch 2 updates the @probe_peer documentation in cfg80211_ops to describe
the STA-mode semantics, adds NL80211_EXT_FEATURE_PROBE_AP to advertise
STA-side support, extends cfg80211_probe_status() to carry an optional
peer address and a link_id (-1 for non-MLO), and extends the nl80211
handler to accept STA/P2P-client interfaces when the driver advertises
the feature (MAC attribute must be omitted; the AP is implied by the
association). All callers of cfg80211_probe_status() are updated.

Patch 3 adds per-link PROBE_PEER support in mac80211 for STA/P2P-client
mode. For MLO STA/P2P-client, it uses IEEE80211_LINK_UNSPECIFIED with
MLD addresses and lets the driver select the link. For non-MLO
STA/P2P-client, it uses link 0 with the per-link BSSID and link address.
The link_valid/link_id bitfields in ieee80211_tx_info.status are set
before transmitting and read back in ieee80211_report_ack_skb() to
report the actual link_id to userspace. AP/GO behaviour is unchanged.

Patch 4 makes mac80211_hwsim populate link_valid/link_id in TX status
for both the direct TX status path and the wmediumd/netlink TX status
path.

---
Changes in v3:
  - Restructured patch split:
    * Patch 1: pure rename (probe_client -> probe_peer), no doc changes
    * Patch 2: documentation update for STA-mode semantics +
               nl80211 API logic change + cfg80211_probe_status update
    * Patch 3: mac80211 implementation
  - Removed unnecessary bitfield padding (no pad2:3)
  - Moved MAC-omission check for STA mode into cfg80211/nl80211
    (not mac80211).
  - Used switch statement in both nl80211_probe_peer() and
    ieee80211_probe_peer().
  - Used guard(rcu)() instead of manual rcu_read_lock/unlock
  - Return -ENOLINK (not -ENOTCONN) for unconnected STA, consistent
    with cfg80211 conventions

Priyansha Tiwari (4):
  wifi: nl80211/cfg80211: rename probe_client to probe_peer
  wifi: cfg80211/nl80211: add STA-mode peer probing
  wifi: mac80211: implement STA-mode peer probing
  wifi: mac80211_hwsim: report TX status link_id

 drivers/net/wireless/ath/wil6210/cfg80211.c   |  10 +-
 .../net/wireless/marvell/mwifiex/cfg80211.c   |   8 +-
 .../wireless/virtual/mac80211_hwsim_main.c    |  43 ++++-
 include/net/cfg80211.h                        |  18 +-
 include/net/mac80211.h                        |   2 +-
 include/uapi/linux/nl80211.h                  |  23 ++-
 net/mac80211/cfg.c                            | 157 ++++++++++--------
 net/mac80211/status.c                         |   5 +-
 net/wireless/nl80211.c                        |  68 +++++---
 net/wireless/rdev-ops.h                       |  10 +-
 net/wireless/trace.h                          |   2 +-
 11 files changed, 216 insertions(+), 130 deletions(-)

-- 
2.34.1


             reply	other threads:[~2026-05-07 17:16 UTC|newest]

Thread overview: 5+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2026-05-07 17:15 Priyansha Tiwari [this message]
2026-05-07 17:16 ` [PATCH wireless-next v3 1/4] wifi: nl80211/cfg80211: rename probe_client to probe_peer Priyansha Tiwari
2026-05-07 17:16 ` [PATCH wireless-next v3 2/4] wifi: cfg80211/nl80211: add STA-mode peer probing Priyansha Tiwari
2026-05-07 17:16 ` [PATCH wireless-next v3 3/4] wifi: mac80211: implement " Priyansha Tiwari
2026-05-07 17:16 ` [PATCH wireless-next v3 4/4] wifi: mac80211_hwsim: report TX status link_id Priyansha Tiwari

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=20260507171603.1007812-1-pritiwa@qti.qualcomm.com \
    --to=priyansha.tiwari@oss.qualcomm.com \
    --cc=johannes@sipsolutions.net \
    --cc=linux-wireless@vger.kernel.org \
    --cc=quic_drohan@quicinc.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox