linux-stable.git/drivers/net/wireless, branch v7.1.13

wifi: mt76: mt7925: ensure tx headroom in usb_sdio_tx_prepare_skb

2026-09-02T12:32:39+00:00

commit ef3e34874d2332d0f63e72c2c35ce5c93568c125 upstream.

mt7925_usb_sdio_tx_prepare_skb() pushes a TX descriptor and a USB
header onto every skb and assumes the headroom for them is already
there. That holds for locally generated traffic, where mac80211
reserves hw->extra_tx_headroom, but forwarded frames are sent through
ieee80211_8023_xmit(), which does not reserve it. Bridge a wired
interface to an mt7925u AP and the first forwarded frame that arrives
short panics the kernel:

 skbuff: skb_under_panic: len:415 put:4 tail:0x19b end:0x640 dev:wlan1
 kernel BUG at net/core/skbuff.c:212!
 Call trace:
  skb_panic+0x58/0x60 (P)
  skb_push+0x58/0x60
  mt7925_usb_sdio_tx_prepare_skb+0xf8/0x1b8 [mt7925_common]
  mt76u_tx_queue_skb+0xa0/0x1f8 [mt76_usb]
  __mt76_tx_queue_skb+0x54/0xe8 [mt76]
  mt76_txq_schedule.part.0+0x204/0x478 [mt76]
  mt76_txq_schedule_all+0x50/0x80 [mt76]
  mt792x_tx_worker+0x68/0x100 [mt792x_lib]
  __mt76_worker_fn+0x84/0x150 [mt76]

Whether a given setup hits it depends on how much headroom the ingress
netdev leaves in its rx skbs. Reproduced on a Raspberry Pi 5 bridging
onboard ethernet to a Netgear A9000; originally reported on an MT7986
router running OpenWrt. Nick Morrow's testing on a Pi 4 (bcmgenet),
which leaves more headroom, helped narrow the trigger to the ingress
path.

The same bug was fixed on mt7921 by commit 98c4d0abf5c4 ("mt76:
mt7921: don't assume adequate headroom for SDIO headers"), but mt7925
was copied from mt7921 without the fix. Add the same guard here.

Fixes: c948b5da6bbe ("wifi: mt76: mt7925: add Mediatek Wi-Fi7 driver for mt7925 chips")
Cc: stable@vger.kernel.org
Link: https://github.com/morrownr/mt76/issues/52
Signed-off-by: Devin Wittmayer 
Link: https://patch.msgid.link/20260715023348.59506-1-lucid_duck@justthetip.ca
Signed-off-by: Felix Fietkau 
Signed-off-by: Greg Kroah-Hartman

wifi: mwifiex: use the subframe length when parsing A-MSDU TDLS frames

2026-08-09T18:26:45+00:00

commit 99a948382af8a225e2d5e54a7052158cd6281cc6 upstream.

mwifiex_11n_dispatch_amsdu_pkt() splits an A-MSDU with
ieee80211_amsdu_to_8023s() and walks the resulting subframes. For each
subframe it passes the subframe data pointer to
mwifiex_process_tdls_action_frame(), but pairs it with skb->len, the
length of the A-MSDU parent, instead of rx_skb->len:

	rx_skb = __skb_dequeue(&list);
	rx_hdr = (struct rx_packet_hdr *)rx_skb->data;
	if (ISSUPP_TDLS_ENABLED(priv->adapter->fw_cap_info) &&
	    ntohs(rx_hdr->eth803_hdr.h_proto) == ETH_P_TDLS) {
		mwifiex_process_tdls_action_frame(priv, (u8 *)rx_hdr,
						  skb->len);
	}

The parent is not a valid description of that buffer, and may not be
valid memory at all. ieee80211_amsdu_to_8023s() ends with

	if (!reuse_skb)
		dev_kfree_skb(skb);

and it only sets reuse_skb when the parent is linear, is not a
head_frag, and is being consumed as the *last* subframe. So when the
parent does not qualify for reuse it has already been freed, and the
read of skb->len is a use-after-free. When it is reused, skb->len is
the length of the last subframe, applied to every earlier subframe,
which over-states the buffer whenever an earlier subframe is shorter.

The callee cannot absorb a wrong length, because it derives its own
ceiling from the value it is given. Each frame type computes

	ies_len = len - sizeof(struct ethhdr) - TDLS_*_FIX_LEN;

and the element walk is then bounded entirely against that ceiling,

	for (end = pos + ies_len; pos + 1 < end; pos += 2 + pos[1]) {
		u8 ie_len = pos[1];

		if (pos + 2 + ie_len > end)
			break;

so a too-large len moves end past the end of the subframe and the walk
reads and copies beyond it. The A-MSDU layout is chosen by the sender,
which makes the difference between the last subframe and a shorter
earlier one remotely selectable. Reaching this requires TDLS support in
firmware and the TDLS ethertype on the subframe.

The other caller, mwifiex_process_rx_packet(), is correct: it passes a
pointer and a length that describe the same region of the RX buffer.

Pass rx_skb->len, the length of the subframe actually being parsed.

Fixes: 776f742040ca ("mwifiex: fix AMPDU not setup on TDLS link problem")
Assisted-by: Codex:gpt-5.6-sol
Assisted-by: Kimi:K3
Cc: stable@vger.kernel.org
Signed-off-by: Zhao Li 
Link: https://patch.msgid.link/20260728115325.19128-1-enderaoelyther@gmail.com
Signed-off-by: Johannes Berg 
Signed-off-by: Greg Kroah-Hartman

wifi: ath12k: fix out-of-bounds clear_bit in ath12k_mac_dp_peer_cleanup()

2026-08-09T18:26:29+00:00

[ Upstream commit 47abd2ca281531deee38a3b3770d885e270e9fc9 ]

ath12k_mac_dp_peer_cleanup() clears the ML peer ID slot on the
free_ml_peer_id_map bitmap by indexing it with dp_peer->peer_id. That is
wrong: dp_peer->peer_id for an MLO peer always carries the
ATH12K_PEER_ML_ID_VALID bit (BIT(13)), so clear_bit() is invoked with
index >= 0x2000, which is far outside the bitmap of ATH12K_MAX_MLO_PEERS
(256) bits and corrupts memory adjacent to ah->free_ml_peer_id_map. The
intended bitmap entry also never gets cleared, so subsequent
ath12k_peer_ml_alloc() calls eventually run out of IDs.

The ID without the VALID bit is what ath12k_peer_ml_alloc() returned and
is stored in ahsta->ml_peer_id. Use that instead.

While there, also reset ahsta->ml_peer_id to ATH12K_MLO_PEER_ID_INVALID so
the bitmap and ahsta->ml_peer_id stay in sync.

Tested-on: WCN7850 hw2.0 PCI WLAN.HMT.1.1.c5-00302-QCAHMTSWPL_V1.0_V2.0_SILICONZ-1.115823.3

Fixes: ee16dcf573d5 ("wifi: ath12k: Define ath12k_dp_peer structure & APIs for create & delete")
Signed-off-by: Baochen Qiang 
Reviewed-by: Rameshkumar Sundaram 
Link: https://patch.msgid.link/20260720-ath12k-fw-allocated-ml-peer-id-v2-1-630632758a80@oss.qualcomm.com
Signed-off-by: Jeff Johnson 
Signed-off-by: Sasha Levin

wifi: ath12k: fix NULL pointer dereference in rhash table destroy

2026-08-03T09:25:51+00:00

commit 70231dcd782201579990ded73e0435d18bb524ca upstream.

When unbinding the ath12k driver, kernel NULL pointer dereferences
occur in irq_work_sync() called from rhashtable_destroy().

Two hash tables are affected:
1. ath12k_link_sta hash table in ath12k_base
2. ath12k_dp_link_peer hash table in ath12k_dp

The issue happens because the destroy functions are called unconditionally
in cleanup paths, but the hash tables are only initialized late in their
respective init functions. If the device was never fully started or if the
init functions failed before initializing the hash tables, the pointers
will be NULL. The issues are always reproducible from a VM because the MSI
addressing initialization is failing.

Call trace for ath12k_link_sta_rhash_tbl_destroy:
 RIP: irq_work_sync+0x1e/0x70
 rhashtable_destroy+0x12/0x60
 ath12k_link_sta_rhash_tbl_destroy+0x19/0x40 [ath12k]
 ath12k_core_stop+0xe/0x80 [ath12k]
 ath12k_core_hw_group_cleanup+0x6b/0xb0 [ath12k]
 ath12k_pci_remove+0x60/0x110 [ath12k]

Call trace for ath12k_dp_link_peer_rhash_tbl_destroy:
 RIP: irq_work_sync+0x1e/0x70
 rhashtable_destroy+0x12/0x60
 ath12k_dp_link_peer_rhash_tbl_destroy+0x29/0x50 [ath12k]
 ath12k_dp_cmn_device_deinit+0x21/0x140 [ath12k]
 ath12k_core_hw_group_cleanup+0x6b/0xb0 [ath12k]
 ath12k_pci_remove+0x60/0x110 [ath12k]

Fix this by adding NULL checks before calling rhashtable_destroy() in
both destroy functions.

The NULL check approach was chosen because the rhashtable pointer
serves as the initialization state indicator. The init can fail at
various points, leaving some components uninitialized. Checking the
pointer directly is simpler than adding separate state flags that
would need synchronization.

Fixes: 57ccca410237 ("wifi: ath12k: Add hash table for ath12k_link_sta in ath12k_base")
Fixes: a88cf5f71adf ("wifi: ath12k: Add hash table for ath12k_dp_link_peer")
Cc: stable@vger.kernel.org
Signed-off-by: Jose Ignacio Tornos Martinez 
Reviewed-by: Vasanthakumar Thiagarajan 
Link: https://patch.msgid.link/20260615112103.601982-1-jtornosm@redhat.com
Signed-off-by: Jeff Johnson 
Signed-off-by: Greg Kroah-Hartman

wifi: ath11k: fix refcount leak in ath11k_ahb_fw_resources_init()

2026-08-03T09:25:51+00:00

commit 0e120ee0822b7cc650bd7b29682a34e137cec10d upstream.

of_get_child_by_name() returns a node pointer with refcount
incremented, but the error path when ath11k_ahb_setup_msa_resources()
fails does not release it.  Add the missing of_node_put() to avoid
leaking the reference.

Cc: stable@vger.kernel.org
Fixes: 095cb947490c ("wifi: ath11k: allow missing memory-regions")
Signed-off-by: Wentao Liang 
Reviewed-by: Baochen Qiang 
Reviewed-by: Rameshkumar Sundaram 
Link: https://patch.msgid.link/20260609092528.220547-1-vulab@iscas.ac.cn
Signed-off-by: Jeff Johnson 
Signed-off-by: Greg Kroah-Hartman

wifi: brcmfmac: set F2 blocksize to 256 for BCM43752

2026-08-03T09:25:51+00:00

commit 29ab31f3f27157648f2f7e6d5e1fd9792fdf0614 upstream.

The BCM43752 is not reliable with the default 512-byte SDIO function 2
block size: on an i.MX8MP board with an AMPAK AP6275S module at
SDR104 / 200 MHz, an iperf TX stress test kills WLAN within seconds:

  mmc_submit_one: CMD53 sg block write failed -84
  brcmf_sdio_dpc: failed backplane access over SDIO, halting operation

Commit d2587c57ffd8 ("brcmfmac: add 43752 SDIO ids and initialization")
set up the 43752 like the 4373 for the F2 watermark but missed the F2
block size, which the 4373 limits to 256 bytes. The vendor driver
(bcmdhd) also programs a 256-byte F2 block size for this chip and runs
the same hardware without errors.

Group the 43752 with the 4373, matching the F2 watermark handling.
With this change a 10-minute bidirectional iperf3 soak completes with
zero SDIO errors at ~270 Mbit/s in each direction.

Backporting note: kernels before v6.18 name this id
SDIO_DEVICE_ID_BROADCOM_CYPRESS_43752, so on those trees the case
label added by this patch must be adjusted to that name. Cherry-picking
the rename commit 74e2ef72bd4b ("wifi: brcmfmac: fix 43752 SDIO FWVID
incorrectly labelled as Cypress (CYW)") first is not a clean
alternative: on trees before v6.17 its context collides with the 43751
additions, and trees before v6.2 lack the FWVID framework it touches.

Fixes: d2587c57ffd8 ("brcmfmac: add 43752 SDIO ids and initialization")
Cc: stable@vger.kernel.org # see patch description, needs adjustments for <= 6.17
Signed-off-by: LiangCheng Wang 
Acked-by: Arend van Spriel 
Link: https://patch.msgid.link/20260715-b43752-f2-blksz-v2-1-f9be49856050@gmail.com
Signed-off-by: Johannes Berg 
Signed-off-by: Greg Kroah-Hartman

wifi: brcmfmac: make release_scratchbuffers idempotent

2026-08-03T09:25:51+00:00

commit 538c51e9d124cf656f2dd0c0394a8545efc7102d upstream.

brcmf_pcie_release_scratchbuffers() frees the shared.scratch and
shared.ringupd DMA buffers with dma_free_coherent() but does not clear
the pointers afterwards, unlike the sibling release_ringbuffers() which
NULLs commonrings/flowrings/idxbuf on release.

Both the bus_reset .reset callback (brcmf_pcie_reset) and
brcmf_pcie_remove() call release_scratchbuffers.  When reset teardown
has run before removal, remove's own teardown would call
dma_free_coherent() a second time on the already-freed DMA allocation.

NULL the pointers after free, matching release_ringbuffers(), so a later
release observes that the allocation has already been released.  This
patch makes repeated sequential release safe; the reset-work lifetime is
handled separately by the following patch.

This issue was found by an in-house static analysis tool.

Fixes: 4684997d9eea ("brcmfmac: reset PCIe bus on a firmware crash")
Cc: stable@vger.kernel.org
Signed-off-by: Fan Wu 
Assisted-by: Codex:gpt-5.6
Acked-by: Arend van Spriel 
Link: https://patch.msgid.link/20260718024353.3147201-2-fanwu01@zju.edu.cn
Signed-off-by: Johannes Berg 
Signed-off-by: Greg Kroah-Hartman

wifi: brcmfmac: drain bus_reset work on device removal

2026-08-03T09:25:51+00:00

commit 43b25879f004c98defa2776bedc6ca4763c51945 upstream.

brcmf_fw_crashed() and the debugfs "reset" entry both schedule
drvr->bus_reset, whose callback recovers drvr through container_of()
and dereferences it.  The removal path frees drvr (brcmf_free ->
wiphy_free) without draining the work, so a bus_reset callback pending
or running during removal can outlive drvr.

Cancellation cannot live in brcmf_detach() or brcmf_free(): the work
callback reaches teardown through the bus .reset op (PCIe
brcmf_pcie_reset -> brcmf_detach; SDIO brcmf_sdio_bus_reset ->
brcmf_sdiod_remove -> brcmf_free), so cancelling there would wait for
the running work and deadlock.

Add a per-bus mutex (bus_reset_lock) and route all arming through
brcmf_bus_schedule_reset(), which under the lock skips when the bus is
marked removing.  Each bus remove entry calls
brcmf_bus_cancel_reset_work(), which under the same lock sets removing
and cancels the work.  Holding the mutex across cancel_work_sync() makes
the set-removing + drain step atomic.  Every producer reaches the arming
path from process context -- the PCIe firmware-halt notification runs in
the threaded IRQ handler (brcmf_pcie_isr_thread) and the SDIO hostmail
path runs from the data workqueue -- so the mutex is taken only in
sleepable contexts.  Where applicable the remove entry first stops the
firmware-crash producer: on PCIe mask the mailbox and synchronize_irq;
on SDIO unregister the bus interrupt and cancel the data worker, which
also reports firmware halts through brcmf_fw_crashed().  The mutex is
initialized at bus allocation.  The SDIO suspend power-off path frees
drvr through the same brcmf_sdiod_remove() and takes the same lock;
resume re-allows the work only on a successful re-probe.

Also guard brcmf_fw_crashed() against a NULL bus_if/drvr: it can fire
before brcmf_attach() wires up drvr, and it dereferences drvr
(bphy_err/brcmf_dev_coredump) before reaching the arming gate.

The bus_reset work is shared across buses, so the drain is applied to
every remove path: PCIe (the .reset op introduced by the Fixes commit),
SDIO (arms the same work through brcmf_fw_crashed()), and USB (via the
debugfs "reset" entry).  cancel_work_sync() drains a running or pending
bus_reset work item before removal frees drvr, and patch 1/2 makes the
scratch-buffer release safe when reset teardown has already released
those DMA buffers.

This patch fixes the lifetime of the bus_reset work item itself.  It does
not attempt to address the separate, pre-existing lifetime of the
asynchronous firmware completion started by the PCIe reset path.  That
callback needs its own lifetime/ownership protocol and is being tracked
separately.

This issue was found by an in-house static analysis tool.

Fixes: 4684997d9eea ("brcmfmac: reset PCIe bus on a firmware crash")
Cc: stable@vger.kernel.org
Signed-off-by: Fan Wu 
Assisted-by: Codex:gpt-5.6
Acked-by: Arend van Spriel 
Link: https://patch.msgid.link/20260718024353.3147201-3-fanwu01@zju.edu.cn
Signed-off-by: Johannes Berg 
Signed-off-by: Greg Kroah-Hartman

wifi: mt76: restrict NPU/PPE active checks to MMIO devices

2026-08-03T09:25:51+00:00

commit 7981aca2bd28a1f7ad7eeab89715442a95b1f72e upstream.

mt76_npu_device_active() and mt76_ppe_device_active() read dev->mmio.npu
and dev->mmio.ppe_dev. The mmio, usb and sdio bus structs share a union in
struct mt76_dev, so on USB and SDIO these read unrelated data from the
usb/sdio struct, which is non-NULL in practice.

mt76_npu_device_active() then returns true on USB, and
mt76_rx_poll_complete() takes the offload path and skips
mt76_rx_aggr_reorder(). RX A-MPDU subframes are delivered out of order and
the peer's TCP stack treats that as loss: heavy retransmissions and reduced
throughput in AP mode. Seen on mt7921u, mt7925u, mt76x2u and mt76x0u.

Gate both helpers on mt76_is_mmio() so they only run for the bus type that
owns the mmio union member.

Fixes: 7fb554b1b623 ("wifi: mt76: Introduce the NPU generic layer")
Cc: stable@vger.kernel.org
Tested-by: Nick Morrow 
Signed-off-by: Devin Wittmayer 
Link: https://patch.msgid.link/20260720232640.41293-1-lucid_duck@justthetip.ca
Signed-off-by: Felix Fietkau 
Signed-off-by: Greg Kroah-Hartman

wifi: mt76: mt7925: drop TXRX_NOTIFY on non-mmio buses

2026-08-03T09:25:50+00:00

commit feeff151c83e7f0ffcdedcad5343852d23d1f6e1 upstream.

PKT_TYPE_TXRX_NOTIFY is an mmio-only event, but mt7925_rx_check() and
mt7925_queue_rx_skb() dispatch it to mt7925_mac_tx_free() on every bus.
mt7925_mac_tx_free() cleans the DMA tx queues with
mt76_queue_tx_cleanup(), which calls queue_ops->tx_cleanup(). Only the
mmio queue ops implement that callback; on USB it is NULL, so a
TXRX_NOTIFY there calls a NULL pointer in the RX worker:

  BUG: kernel NULL pointer dereference, address: 0000000000000000
  RIP: 0010:0x0
  Call Trace:
   mt7925_mac_tx_free+0x58/0x350 [mt7925_common]
   mt7925_rx_check+0xe2/0x130 [mt7925_common]
   mt76u_rx_worker+0x1b9/0x620 [mt76_usb]

Drop the event on non-mmio buses via mt76_is_mmio(), as in
commit 5683e1488aa9 ("wifi: mt76: connac: do not check WED status for
non-mmio devices").

Fixes: c948b5da6bbe ("wifi: mt76: mt7925: add Mediatek Wi-Fi7 driver for mt7925 chips")
Cc: stable@vger.kernel.org
Signed-off-by: Devin Wittmayer 
Link: https://patch.msgid.link/20260627191336.20223-3-lucid_duck@justthetip.ca
Signed-off-by: Felix Fietkau 
Signed-off-by: Greg Kroah-Hartman