When multiple links switch channel contexts around the same time, mac80211
may complete CSA for several links together and invoke
ath12k_mac_op_switch_vif_chanctx() with an array of vifs spanning more than
one underlying radio in a single-wiphy configuration.
The driver currently assumes that all entries in the vifs array belong to the
same radio and derives the radio context from the first element. On multi-radio
hardware, this can lead to incorrect vdev selection/updates and may corrupt
driver state when the number of vifs exceeds what a single radio supports.
Fix this by validating each vif's switch request and then processing vifs
grouped by their associated radio. For each vif, ensure the band does not
change across the switch and that both old/new channel contexts resolve to a
valid ath12k device. Reject attempts to move a vif between radios (not
supported for now) and return -EOPNOTSUPP to upper layers.
Then, iterate through the input vifs, collect all unprocessed entries that map
to the same radio, and invoke ath12k_mac_update_vif_chan() separately for each
radio group. This removes any reliance on mac80211 providing the array grouped
by radio or sharing old_ctx pointers across vifs.
Tested-on: QCN9274 hw2.0 PCI WLAN.WBE.1.5-01651-QCAHKSWPL_SILICONZ-1
Signed-off-by: Aditya Kumar Singh <aditya.kumar.singh@oss.qualcomm.com>
Co-developed-by: Maharaja Kennadyrajan <maharaja.kennadyrajan@oss.qualcomm.com>
Signed-off-by: Maharaja Kennadyrajan <maharaja.kennadyrajan@oss.qualcomm.com>
Reviewed-by: Baochen Qiang <baochen.qiang@oss.qualcomm.com>
Reviewed-by: Rameshkumar Sundaram <rameshkumar.sundaram@oss.qualcomm.com>
Link: https://patch.msgid.link/20260522091828.3199584-1-maharaja.kennadyrajan@oss.qualcomm.com
Signed-off-by: Jeff Johnson <jeff.johnson@oss.qualcomm.com>
Currently, the driver uses memory profile parameters to determine the
maximum number of supported clients, with a default limit of 512 for
single-radio and 128 for DBS and DBS+SBS configurations. However,
some devices have lower hardware limits depending on the radio
configuration. Exceeding these hardware-specific limits can lead to
firmware crashes.
Add hardware parameters in ath12k_hw_params to define the maximum supported
clients for each radio configuration. The driver uses the minimum of the
memory profile limit and the hardware capability limit to prevent exceeding
hardware constraints.
Tested-on: IPQ5332 hw1.0 AHB WLAN.WBE.1.6-01275-QCAHKSWPL_SILICONZ-1
Tested-on: QCN9274 hw2.0 PCI WLAN.WBE.1.6-01243-QCAHKSWPL_SILICONZ-1
Signed-off-by: Aaradhana Sahu <aaradhana.sahu@oss.qualcomm.com>
Reviewed-by: Rameshkumar Sundaram <rameshkumar.sundaram@oss.qualcomm.com>
Reviewed-by: Baochen Qiang <baochen.qiang@oss.qualcomm.com>
Link: https://patch.msgid.link/20260515030909.3312511-1-aaradhana.sahu@oss.qualcomm.com
Signed-off-by: Jeff Johnson <jeff.johnson@oss.qualcomm.com>
_ieee80211_set_active_links() calls _ieee80211_link_use_channel() for
each newly-added link and WARN_ON_ONCE()s if it fails. The call uses
assign_on_failure=true, which allows mac80211 to continue despite
driver failures, but when a mac80211-level channel validation fails
(e.g., combinations check, DFS, or no available radio),
drv_assign_vif_chanctx() is never reached. Since ath12k_mac_vdev_create()
is only called from that path, arvif->is_created remains false and
arvif->ar remains NULL for the failed link.
The subsequent drv_change_sta_links() call reaches
ath12k_mac_op_change_sta_links(), which allocates an arsta and sets
ahsta->links_map |= BIT(link_id) for the broken link before checking
whether the link is ready. When the vdev was never created, only
station_add() is skipped, but the link remains in links_map.
Any subsequent operation iterating links_map and dereferencing arvif->ar
without a NULL check will crash. Two observed examples are NULL deref in
ath12k_mac_ml_station_remove() on disconnect and in ath12k_mac_op_set_key()
when wpa_supplicant installs PTK keys.
BUG: Unable to handle kernel NULL pointer dereference at 0x00000000
pc : ath12k_mac_station_post_remove+0x40/0xe8 [ath12k]
Call trace:
ath12k_mac_station_post_remove+0x40/0xe8 [ath12k]
ath12k_mac_op_sta_state+0xb60/0x1720 [ath12k]
drv_sta_state+0x100/0xbd8 [mac80211]
__sta_info_destroy_part2+0x148/0x178 [mac80211]
ieee80211_set_disassoc+0x500/0x678 [mac80211]
BUG: Unable to handle kernel NULL pointer dereference at 0x00000000
pc : ath12k_mac_op_set_key+0x1f8/0x2c0 [ath12k]
Call trace:
ath12k_mac_op_set_key+0x1f8/0x2c0 [ath12k]
drv_set_key+0x70/0x100 [mac80211]
ieee80211_key_enable_hw_accel+0x78/0x260 [mac80211]
ieee80211_add_key+0x16c/0x2ac [mac80211]
nl80211_new_key+0x138/0x280 [cfg80211]
Fix this by checking arvif->is_created before calling
ath12k_mac_alloc_assign_link_sta(). This prevents the broken link from
entering links_map, so all subsequent operations iterating the bitmap
are protected. The reliability of arvif->is_created across all error
paths is ensured by the preceding patch.
Tested-on: WCN7850 hw2.0 PCI WLAN.HMT.1.1.c5-00302-QCAHMTSWPL_V1.0_V2.0_SILICONZ-1.115823.3
Fixes: a27fa6148d ("wifi: ath12k: support change_sta_links() mac80211 op")
Signed-off-by: Wei Zhang <wei.zhang@oss.qualcomm.com>
Reviewed-by: Baochen Qiang <baochen.qiang@oss.qualcomm.com>
Reviewed-by: Rameshkumar Sundaram <rameshkumar.sundaram@oss.qualcomm.com>
Link: https://patch.msgid.link/20260512044906.1735821-3-wei.zhang@oss.qualcomm.com
Signed-off-by: Jeff Johnson <jeff.johnson@oss.qualcomm.com>
ath12k_mac_vdev_create() has three error path issues that leave arvif
in an inconsistent state:
1. When ath12k_wmi_vdev_create() fails, the function returns directly
without clearing arvif->ar, which was already set before the WMI
call. Subsequent code checking arvif->ar to determine vdev readiness
will see a non-NULL value despite no vdev existing in firmware.
2. When ath12k_wmi_send_peer_delete_cmd() fails in err_peer_del, the
code jumped to err: skipping the DP peer cleanup and vdev rollback,
leaving num_created_vdevs, vdev maps and arvif list membership live.
3. When ath12k_wait_for_peer_delete_done() fails, the code jumped to
err_vdev_del: skipping the DP peer cleanup.
Fix by changing the ath12k_wmi_vdev_create() failure to goto err instead
of returning directly, routing both err_peer_del failure paths through
err_dp_peer_del: for proper DP peer and vdev rollback, and consolidating
the arvif state cleanup at err:.
Tested-on: WCN7850 hw2.0 PCI WLAN.HMT.1.1.c5-00302-QCAHMTSWPL_V1.0_V2.0_SILICONZ-1.115823.3
Fixes: 477cabfdb7 ("wifi: ath12k: modify link arvif creation and removal for MLO")
Signed-off-by: Wei Zhang <wei.zhang@oss.qualcomm.com>
Reviewed-by: Baochen Qiang <baochen.qiang@oss.qualcomm.com>
Reviewed-by: Rameshkumar Sundaram <rameshkumar.sundaram@oss.qualcomm.com>
Link: https://patch.msgid.link/20260512044906.1735821-2-wei.zhang@oss.qualcomm.com
Signed-off-by: Jeff Johnson <jeff.johnson@oss.qualcomm.com>
For some chipsets, firmware can report HTT_T2H_MSG_TYPE_PEER_MAP2 with
peer_id 0 as a valid value for mapping ath12k_dp_link_peer to
ath12k_dp_peer.
ath12k_dp_peer_find_by_peerid() currently treats peer_id 0 as invalid.
When firmware assigns peer_id 0, peer lookup fails. As a result,
DHCP OFFER packets are dropped in __ieee80211_rx_handle_packet()
because pubsta is NULL.
ath12k_dp_rx_deliver_msdu() <- rx_info->peer_id 0
ath12k_dp_peer_find_by_peerid -> peer NULL
ieee80211_rx_napi <- pubsta NULL
ieee80211_rx_list
__ieee80211_rx_handle_packet <- pubsta NULL, skb undelivered
The following error in the TX completion path is caused by the same issue:
ath12k_wifi7_pci 0000:04:00.0: dp_tx: failed to find the peer with peer_id 0
The error message is triggered by:
ath12k_wifi7_dp_tx_complete_msdu
ath12k_dp_link_peer_find_by_peerid <- ts->peer_id 0
ath12k_dp_peer_find_by_peerid -> peer NULL
ath12k_dp_tx_htt_tx_complete_buf
ath12k_dp_link_peer_find_by_peerid <- peer_id 0
ath12k_dp_peer_find_by_peerid -> peer NULL
Fix this by allowing peer_id 0 in ath12k_dp_peer_find_by_peerid() and
rejecting only values >= ATH12K_DP_PEER_ID_INVALID.
Also update peer_id 0 handling in monitor path:
Always call ath12k_dp_link_peer_find_by_peerid() in
ath12k_dp_rx_h_find_link_peer() to fetch the peer, including when
peer_id is 0.
Always store peer_id in ppdu_info->peer_id in
ath12k_wifi7_dp_mon_rx_parse_status_tlv(), including peer_id 0.
Tested-on: QCC2072 hw1.0 PCI WLAN.COL.1.0.c2-00074-QCACOLSWPL_V1_TO_SILICONZ-1
Tested-on: WCN7850 hw2.0 PCI WLAN.HMT.1.1.c7-00108-QCAHMTSWPL_V1.0_V2.0_SILICONZ_UPSTREAM-3
Tested-on: QCN9274 hw2.0 PCI WLAN.WBE.1.6-01243-QCAHKSWPL_SILICONZ-1
Signed-off-by: Hangtian Zhu <hangtian.zhu@oss.qualcomm.com>
Reviewed-by: Rameshkumar Sundaram <rameshkumar.sundaram@oss.qualcomm.com>
Reviewed-by: Baochen Qiang <baochen.qiang@oss.qualcomm.com>
Link: https://patch.msgid.link/20260512025732.1297849-1-hangtian.zhu@oss.qualcomm.com
Signed-off-by: Jeff Johnson <jeff.johnson@oss.qualcomm.com>
In ath12k_wifi7_dp_rx_h_verify_tkip_mic(), the call to
ath12k_dp_rx_check_nwifi_hdr_len_valid() may return false when the
NWIFI header length is invalid, causing the function to abort early with
-EINVAL.
When this happens, the error propagates to
ath12k_wifi7_dp_rx_h_defrag(), which clears first_frag by setting it
to NULL. As a result, the corresponding MSDU is no longer referenced
by the defragmentation path and is never freed.
This leads to a memory leak for the affected MSDU on this error path.
Proper cleanup is required to ensure the MSDU is released when header
validation fails during TKIP MIC verification.
Tested-on: WCN7850 hw2.0 PCI WLAN.HMT.1.1.c5-00302-QCAHMTSWPL_V1.0_V2.0_SILICONZ-1.115823.3
Fixes: 9a0dddfb30 ("wifi: ath12k: Fix invalid data access in ath12k_dp_rx_h_undecap_nwifi")
Signed-off-by: Miaoqing Pan <miaoqing.pan@oss.qualcomm.com>
Reviewed-by: Tamizh Chelvam Raja <tamizh.raja@oss.qualcomm.com>
Reviewed-by: Baochen Qiang <baochen.qiang@oss.qualcomm.com>
Link: https://patch.msgid.link/20260512021108.2031651-1-miaoqing.pan@oss.qualcomm.com
Signed-off-by: Jeff Johnson <jeff.johnson@oss.qualcomm.com>
In monitor mode, the driver incorrectly assigns the legacy rate
to the rate_idx field of the radiotap header for HT/VHT/HE/EHT
frames, ignoring the actual MCS value parsed from the hardware.
This causes packet analyzers (like Wireshark) to display incorrect
MCS values (e.g., legacy base rates instead of the true MCS).
Fix this by assigning ppdu_info->mcs as the default rate_mcs
in ath12k_dp_mon_fill_rx_rate(), and remove rate_idx assignments in
ath12k_dp_mon_update_radiotap() to preserve
the previously calculated MCS values (including the HT NSS offset).
Tested-on: QCN9274 hw2.0 PCI WLAN.WBE.1.4.1-00199-QCAHKSWPL_SILICONZ
Fixes: 5393dcb452 ("wifi: ath12k: change the status update in the monitor Rx")
Closes: https://bugzilla.kernel.org/show_bug.cgi?id=220864
Signed-off-by: Kwan Lai Chee Hou <laicheehou9@gmail.com>
Reviewed-by: Rameshkumar Sundaram <rameshkumar.sundaram@oss.qualcomm.com>
Reviewed-by: Baochen Qiang <baochen.qiang@oss.qualcomm.com>
Link: https://patch.msgid.link/20260507015336.14636-1-laicheehou9@gmail.com
Signed-off-by: Jeff Johnson <jeff.johnson@oss.qualcomm.com>
When P2P support is enabled, wpa_supplicant creates a p2p-device
interface by default, which implicitly consumes one vdev. On systems
managed by NetworkManager, this interface cannot be reliably disabled,
leaving only two usable interfaces for user configurations.
Increase num_vdevs to four for QCA6390 hw2.0, WCN6855 hw2.0/hw2.1,
QCA2066 hw2.1, and QCA6698AQ hw2.1 to account for the implicit
p2p-device and enable common concurrency scenarios such as AP + AP + STA.
This change increases interface concurrency in the two-channel scenario
by raising the maximum vdev limit, while keeping other combination rules
unchanged.
Tested-on: QCA6390 hw2.0 PCI WLAN.HST.1.0.1-05266-QCAHSTSWPLZ_V2_TO_X86-1
Tested-on: WCN6855 hw2.0 PCI WLAN.HSP.1.1-03125-QCAHSPSWPL_V1_V2_SILICONZ_LITE-3.6510.41
Tested-on: WCN6855 hw2.1 PCI WLAN.HSP.1.1-04685-QCAHSPSWPL_V1_V2_SILICONZ_IOE-1
Tested-on: QCA2066 hw2.1 PCI WLAN.HSP.1.1-03926.13-QCAHSPSWPL_V2_SILICONZ_CE-2.52297.9
Tested-on: QCA6698AQ hw2.1 PCI WLAN.HSP.1.1-04685-QCAHSPSWPL_V1_V2_SILICONZ_IOE-1
Signed-off-by: Wei Zhang <wei.zhang@oss.qualcomm.com>
Reviewed-by: Rameshkumar Sundaram <rameshkumar.sundaram@oss.qualcomm.com>
Reviewed-by: Baochen Qiang <baochen.qiang@oss.qualcomm.com>
Link: https://patch.msgid.link/20260525020711.2590815-1-wei.zhang@oss.qualcomm.com
Signed-off-by: Jeff Johnson <jeff.johnson@oss.qualcomm.com>
In the WBM error path, while processing TKIP MIC errors, MSDU length
is fetched from the hal_rx_desc's msdu_end. This MSDU length is
directly passed to skb_put() without validation. In stress test
scenarios, the WBM error ring may receive invalid descriptors, which
could lead to an invalid MSDU length.
To fix this, add a check to drop the skb when the calculated MSDU
length is greater than the skb size.
This is adapted from the discussion/patch of the ath12k driver [1].
Tested-on: WCN6855 hw2.1 PCI WLAN.HSP.1.1-04685-QCAHSPSWPL_V1_V2_SILICONZ_IOE-1
Link: https://lore.kernel.org/linux-wireless/20250416021903.3178962-1-nithyanantham.paramasivam@oss.qualcomm.com/ # [1]
Signed-off-by: Miaoqing Pan <miaoqing.pan@oss.qualcomm.com>
Reviewed-by: Rameshkumar Sundaram <rameshkumar.sundaram@oss.qualcomm.com>
Reviewed-by: Baochen Qiang <baochen.qiang@oss.qualcomm.com>
Link: https://patch.msgid.link/20260512022351.2033155-3-miaoqing.pan@oss.qualcomm.com
Signed-off-by: Jeff Johnson <jeff.johnson@oss.qualcomm.com>
In certain cases, hardware might provide packets with a
length greater than the maximum native Wi-Fi header length.
This can lead to accessing and modifying fields in the header
within the ath11k_dp_rx_h_undecap_nwifi() function for the
DP_RX_DECAP_TYPE_NATIVE_WIFI decap type and
potentially result in invalid data access and memory corruption.
Kernel stack is corrupted in: ath11k_dp_rx_h_undecap+0x6b0/0x6b0 [ath11k]
Call trace:
ath11k_dp_rx_h_mpdu+0x0/0x2e8 [ath11k]
ath11k_dp_rx_h_mpdu+0x1e0/0x2e8 [ath11k]
ath11k_dp_rx_wbm_err+0x1e0/0x450 [ath11k]
ath11k_dp_rx_process_wbm_err+0x2fc/0x460 [ath11k]
ath11k_dp_service_srng+0x2e0/0x348 [ath11k]
Add a sanity check before processing the SKB to prevent invalid
data access in the undecap native Wi-Fi function for the
DP_RX_DECAP_TYPE_NATIVE_WIFI decap type.
This adapted from the discussion/patch of the ath12k driver [1].
Tested-on: WCN6855 hw2.1 PCI WLAN.HSP.1.1-04685-QCAHSPSWPL_V1_V2_SILICONZ_IOE-1
Link: https://lore.kernel.org/linux-wireless/20250211090302.4105141-1-tamizh.raja@oss.qualcomm.com/ # [1]
Signed-off-by: Miaoqing Pan <miaoqing.pan@oss.qualcomm.com>
Reviewed-by: Rameshkumar Sundaram <rameshkumar.sundaram@oss.qualcomm.com>
Reviewed-by: Baochen Qiang <baochen.qiang@oss.qualcomm.com>
Link: https://patch.msgid.link/20260512022351.2033155-2-miaoqing.pan@oss.qualcomm.com
Signed-off-by: Jeff Johnson <jeff.johnson@oss.qualcomm.com>
Convert kzalloc_obj + kcalloc to kzalloc_flex to save an allocation.
Add __counted_by to get extra runtime analysis. Move counting variable
assignment immediately after allocation before any potential accesses.
kzalloc_flex does this anyway for GCC >= 15.
Signed-off-by: Rosen Penev <rosenp@gmail.com>
Reviewed-by: Baochen Qiang <baochen.qiang@oss.qualcomm.com>
Reviewed-by: Rameshkumar Sundaram <rameshkumar.sundaram@oss.qualcomm.com>
Link: https://patch.msgid.link/20260428205017.26288-1-rosenp@gmail.com
Signed-off-by: Jeff Johnson <jeff.johnson@oss.qualcomm.com>
Matthieu Baerts says:
====================
selftests: mptcp: reduce bufferbloat and cleanup
Bufferbloat is baaaad, even in our selftests: let's kill it (or at least
reduce it). By doing that, the tests (seem to) have a more stable
transfer, and are then less unstable. That's what patches 1-2 are doing,
and they can be backported up to 5.10.
Patch 3 is not related: a small fix in the selftests to remove temp
files that were not deleted in some conditions, since v5.13.
====================
Link: https://patch.msgid.link/20260527-net-mptcp-sft-bufferbloat-exit-v1-0-9afc4e742090@kernel.org
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Set the EXIT trap for cleanup immediately after creating temporary file
variables, before init and make_file, to ensure cleanup runs on any
failure or interruption during the early setup phase.
Signed-off-by: Geliang Tang <tanggeliang@kylinos.cn>
Reviewed-by: Matthieu Baerts (NGI0) <matttbe@kernel.org>
Signed-off-by: Matthieu Baerts (NGI0) <matttbe@kernel.org>
Link: https://patch.msgid.link/20260527-net-mptcp-sft-bufferbloat-exit-v1-3-9afc4e742090@kernel.org
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Avoid using a fixed limit, no matter the setup. This was causing too
high bufferbloat in some situations, e.g. with a low bandwidth and very
low delay because the default limit was too high for this case.
Instead, use more appropriated limits. Note that unbalanced bandwidth
modes seem to require slightly higher limits to cope with the different
bursts.
Reviewed-by: Mat Martineau <martineau@kernel.org>
Signed-off-by: Matthieu Baerts (NGI0) <matttbe@kernel.org>
Link: https://patch.msgid.link/20260527-net-mptcp-sft-bufferbloat-exit-v1-2-9afc4e742090@kernel.org
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Netem is used to apply a rate limit, and its 'limit' option is per
packet.
Disable GSO on both sides to work with packets of a specific size. That
increases the number of packets, but stabilise the throughput. As a
consequence, limits are more adapted, and the bufferbloat is reduced.
Reviewed-by: Mat Martineau <martineau@kernel.org>
Signed-off-by: Matthieu Baerts (NGI0) <matttbe@kernel.org>
Link: https://patch.msgid.link/20260527-net-mptcp-sft-bufferbloat-exit-v1-1-9afc4e742090@kernel.org
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Since commit 4ee58e1e56 ("net: promote SIOCSHWTSTAMP and SIOCGHWTSTAMP
ioctls to dedicated handlers"), SIOCSHWTSTAMP and SIOCGHWTSTAMP are no
longer dispatched through dev_eth_ioctl() / ndo_eth_ioctl(). They are
now handled by their own dedicated functions dev_set_hwtstamp() and
dev_get_hwtstamp() in the ioctl path.
However, the comment describing ndo_eth_ioctl in netdevice.h still
lists these two ioctls, which is misleading for driver developers who
may incorrectly assume they need to handle hardware timestamping
commands in their ndo_eth_ioctl implementation.
Remove the stale references from the comment to accurately reflect that
ndo_eth_ioctl only handles SIOCGMIIPHY, SIOCGMIIREG and SIOCSMIIREG.
Signed-off-by: Xuan Zhuo <xuanzhuo@linux.alibaba.com>
Reviewed-by: Vadim Fedorenko <vadim.fedorenko@linux.dev>
Link: https://patch.msgid.link/20260527120936.24169-1-xuanzhuo@linux.alibaba.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
ETHTOOL_MSG_STRSET_GET is the only op which sets allow_nodev_do.
When no device is provided it dumps static tables, there's no
need to hold rtnl_lock for this.
Not taking rtnl_lock is a minor win in itself so I think this
patch stands on its own merits. Later on it will be useful
to do locking only in paths which have access to a netdev,
so that we can decide which locks to take per-netdev.
Reviewed-by: Vadim Fedorenko <vadim.fedorenko@linux.dev>
Link: https://patch.msgid.link/20260527162522.3344231-1-kuba@kernel.org
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
This reverts commit 4157501b9a ("vsock/virtio: fix skb overhead
overflow on 32-bit builds"). The fix was semantically correct (although
it would have been better to use mul_u32_u32(), as David pointed out),
but in practice we are estimating the memory used to allocate the SKBs,
and this will never cause a 32-bit variable to overflow on a 32-bit
system, since the memory would have run out long before that. On 64-bit,
SKB_TRUESIZE() already evaluates to size_t, so the multiplication is
already in 64-bit arithmetic without the cast.
Let's revert this to avoid unnecessary 64-bit multiplies on the
per-packet receive path on 32-bit systems.
Reported-by: David Laight <david.laight.linux@gmail.com>
Closes: https://lore.kernel.org/netdev/20260523173557.5cc4f4f6@pumpkin
Suggested-by: "Michael S. Tsirkin" <mst@redhat.com>
Signed-off-by: Stefano Garzarella <sgarzare@redhat.com>
Acked-by: Michael S. Tsirkin <mst@redhat.com>
Reviewed-by: David Laight <david.laight.linux@gmail.com>
Link: https://patch.msgid.link/20260527171046.130211-1-sgarzare@redhat.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Jakub Kicinski says:
====================
docs: net: updates for old and cobwebbed docs
I'm hoping to start feeding our docs into the AI review tools, instead
of maintaining a separate repo with review prompts. To experiment with
that we have to refresh the docs a little bit.
A read thru our current docs makes one slightly question the value
of including them in reviews. But directionally, I feel, it's probably
still right. I'm hoping the Rx Checksum section about not dropping packets
for example to be impactful. I don't think the current AI agents or
review docs include this guidance.
====================
Link: https://patch.msgid.link/20260526160151.2793354-1-kuba@kernel.org
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Update the segmentation offload documentation to match current GSO types:
- clarify csum_start for encapsulated TSO
- document TCP AccECN GSO and NETIF_F_GSO_ACCECN
- distinguish legacy UFO from UDP L4 GSO
- add ESP and fraglist GSO entries
Link: https://patch.msgid.link/20260526160151.2793354-11-kuba@kernel.org
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
checksum-offloads.rst seems like a better place to render
the checksum comment than skbuff.rst.
Remove the stale references to sections in that comment
(it no longer has A, B, C, D, E sections).
Reviewed-by: Willem de Bruijn <willemb@google.com>
Link: https://patch.msgid.link/20260526160151.2793354-10-kuba@kernel.org
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
The Rx checksum processing gives people pause. The two main questions
in my experience are:
- what to do with bad IPv4 checksum; and
- what to do with packets with bad checksum.
Folks often feel the urge to drop the latter, to "avoid overloading
the host".
Reviewed-by: Willem de Bruijn <willemb@google.com>
Link: https://patch.msgid.link/20260526160151.2793354-9-kuba@kernel.org
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Update the checksum offload documentation to match current code:
- SCTP CRC32c offload requires NETIF_F_SCTP_CRC, not ordinary IP
checksum offload
- NETIF_F_IP_CSUM and NETIF_F_IPV6_CSUM are restricted legacy
features; new devices should use NETIF_F_HW_CSUM
- GRE LCO is handled by the shared gre_build_header() helper used by
both IPv4 and IPv6 GRE
- VXLAN_F_REMCSUM_TX is a VXLAN configuration flag, not a field of
struct vxlan_rdst
Link: https://patch.msgid.link/20260526160151.2793354-8-kuba@kernel.org
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Update netdev feature documentation for current locking rules and
feature semantics. Clarify hw_features updates and netdev_update_features()
locking, keep the NETIF_F_NEVER_CHANGE rule with the VLAN challenged
exception, fix the HSR duplication wording, and document netdev->netmem_tx
as a device flag rather than a feature bit.
Split the list of basic feature sets from the "extra" ones like
vlan_features. A bunch of the newer fields weren't documented and
having them all together would be confusing.
Link: https://patch.msgid.link/20260526160151.2793354-7-kuba@kernel.org
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Update the NAPI documentation to match current API behavior:
- repeated napi_disable() calls hang waiting for ownership, rather
than deadlock
- NAPI IDs are exposed through SO_INCOMING_NAPI_ID and netdev Netlink
- epoll uses the maxevents parameter spelling
- add that drivers holding the netdev instance lock may need _locked()
variants
Link: https://patch.msgid.link/20260526160151.2793354-6-kuba@kernel.org
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Update the code examples
- update the YNL sample with the latest(?) APIs
- struct dmabuf_tx_cmsg does not exist, use __u32 directly
Acked-by: Stanislav Fomichev <sdf@fomichev.me>
Link: https://patch.msgid.link/20260526160151.2793354-5-kuba@kernel.org
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Update the kernel-internal ethtool stats list to match current code:
- spell the entries as "struct ethtool_*_stats", not as functions
- list the full set of structures, not only pause and fec
- mention that fields are pre-initialized to ETHTOOL_STAT_NOT_SET by
ethtool_stats_init() and drivers should leave unsupported fields at
that value rather than zeroing them
Link: https://patch.msgid.link/20260526160151.2793354-4-kuba@kernel.org
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Update the driver documentation TX queue example to match current APIs:
- use the ring-local tx_ring_mask field in drv_tx_avail()
- stop the selected netdev_queue with netif_tx_stop_queue() instead of
stopping queue 0 with netif_stop_queue()
Link: https://patch.msgid.link/20260526160151.2793354-3-kuba@kernel.org
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
A handful of unrelated nits:
- free_netdevice() does not exist; replace two stray references
with free_netdev().
- The simple-driver probe example fell through into err_undo after
register_netdev() success; add return 0 for clarity.
- Clarify the netdev_priv() paragraph: "(netdev_priv())" was easy
to misread as the thing that needs explicit freeing; spell out
that it refers to extra pointers stored in the device private
struct.
- ndo_setup_tc synchronization note: TC_SETUP_BLOCK / TC_SETUP_FT
actually run under block->cb_lock, not "NFT locks", and rtnl_lock
may or may not be held depending on path.
- ->lltx guidance reads as very outdated, it's not really deprecated.
I suspect people may have been trying to use it for HW drivers
in the past but I can't think of such a case in the last decade.
Acked-by: Stanislav Fomichev <sdf@fomichev.me>
Link: https://patch.msgid.link/20260526160151.2793354-2-kuba@kernel.org
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Current release - regressions:
- netfilter: walk fib6_siblings under RCU
Previous releases - regressions:
- netlink: fix sending unassigned nsid after assigned one
- bridge: fix sleep in atomic context in netlink path
- sched: fix ethx:ingress -> ethy:egress -> ethx:ingress mirred loop
- ipv4: fix net->ipv4.sysctl_local_reserved_ports UaF
- eth: tun: free page on short-frame rejection in tun_xdp_one()
Previous releases - always broken:
- skbuff: fix missing zerocopy reference in pskb_carve helpers
- handshake: drain pending requests at net namespace exit
- ethtool:
- rss: avoid modifying the RSS context response
- module: avoid leaking a netdev ref on module flash errors
- coalesce: cap profile updates at NET_DIM_PARAMS_NUM_PROFILES
- netfilter: fix dst corruption in same register operation
- nfc: hci: fix out-of-bounds read in HCP header parsing
- ipv6: exthdrs: refresh nh pointer after ipv6_hop_jumbo()
- eth: vti: use ip6_tnl.net in vti6_changelink().
- eth: vxlan: do not reuse cached ip_hdr() value after skb_tunnel_check_pmtu()
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
-----BEGIN PGP SIGNATURE-----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=IBwj
-----END PGP SIGNATURE-----
Merge tag 'net-7.1-rc6' of git://git.kernel.org/pub/scm/linux/kernel/git/netdev/net
Pull networking fixes from Paolo Abeni:
"This is again significantly bigger than the same point into the
previous cycle, but at least smaller than last week.
I'm not aware of any pending regression for the current cycle.
Including fixes from netfilter.
Current release - regressions:
- netfilter: walk fib6_siblings under RCU
Previous releases - regressions:
- netlink: fix sending unassigned nsid after assigned one
- bridge: fix sleep in atomic context in netlink path
- sched: fix ethx:ingress -> ethy:egress -> ethx:ingress mirred loop
- ipv4: fix net->ipv4.sysctl_local_reserved_ports UaF
- eth: tun: free page on short-frame rejection in tun_xdp_one()
Previous releases - always broken:
- skbuff: fix missing zerocopy reference in pskb_carve helpers
- handshake: drain pending requests at net namespace exit
- ethtool:
- rss: avoid modifying the RSS context response
- module: avoid leaking a netdev ref on module flash errors
- coalesce: cap profile updates at NET_DIM_PARAMS_NUM_PROFILES
- netfilter: fix dst corruption in same register operation
- nfc: hci: fix out-of-bounds read in HCP header parsing
- ipv6: exthdrs: refresh nh pointer after ipv6_hop_jumbo()
- eth:
- vti: use ip6_tnl.net in vti6_changelink().
- vxlan: do not reuse cached ip_hdr() value after
skb_tunnel_check_pmtu()"
* tag 'net-7.1-rc6' of git://git.kernel.org/pub/scm/linux/kernel/git/netdev/net: (94 commits)
dpll: zl3073x: make frequency monitor a per-device attribute
dpll: zl3073x: use __dpll_device_change_ntf() and remove change_work
dpll: export __dpll_device_change_ntf() for use under dpll_lock
net/handshake: Drain pending requests at net namespace exit
net/handshake: Verify file-reference balance in submit paths
net/handshake: Close the submit-side sock_hold race
net/handshake: hand off the pinned file reference to accept_doit
net/handshake: Take a long-lived file reference at submit
net/handshake: Pass negative errno through handshake_complete()
nvme-tcp: store negative errno in queue->tls_err
net/handshake: Use spin_lock_bh for hn_lock
net: skbuff: fix missing zerocopy reference in pskb_carve helpers
net: hibmcge: move dma_rmb() after dma_sync_single_for_cpu() in RX path
net: hibmcge: disable Relaxed Ordering to fix RX packet corruption
selftests/tc-testing: Add netem test case exercising loops
selftests/tc-testing: Add mirred test cases exercising loops
net/sched: act_mirred: Fix return code in early mirred redirect error paths
net/sched: act_mirred: Fix blockcast recursion bypass leading to stack overflow
net/sched: Fix ethx:ingress -> ethy:egress -> ethx:ingress mirred loop
net/sched: fix packet loop on netem when duplicate is on
...
Nicholas Carlini reports that the keyring code calls assoc_array_find()
in find_key_to_update() without holding the RCU read lock, while the
assoc_array_gc() code really is designed around removing the node from
the tree and then freeing it after an RCU grace-period.
The regular key handling doesn't see this because holding the keyring
semaphore hides any lifetime issues, but the persistent key handling
uses a different model.
Instead of extending the keyring locking, just do the simple RCU locking
that the assoc_array was designed for.
Reported-by: Nicholas Carlini <npc@anthropic.com>
Cc: David Howells <dhowells@redhat.com>
Cc: Jarkko Sakkinen <jarkko@kernel.org>
Cc: Paul Moore <paul@paul-moore.com>
Cc: James Morris James Morris <jmorris@namei.org>
Cc: Serge E. Hallyn <serge@hallyn.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Address several teardown issues and resource leaks in the driver's remove
path and error handling:
1. Debounce clock reference leak: The debounce clock (bank->db_clk) is
obtained using of_clk_get() which increments the clock's reference
count, but clk_put() is never called. Register a devm action to
cleanly release it on unbind. Note that of_clk_get(..., 1) remains
necessary over devm_clk_get() because the DT binding does not define
clock-names, precluding name-based lookup.
2. Unregistered chained IRQ handler: The chained IRQ handler is not
disconnected in remove(). If a stray interrupt fires after the driver
is removed, the kernel attempts to execute a stale handler, leading
to a panic. Fix this by clearing the handler in remove().
3. IRQ domain leak: The linear IRQ domain and its generic chips are
allocated manually during probe but never removed. Remove the IRQ
domain during driver teardown to free the associated generic chips
and mappings.
Fixes: 936ee2675e ("gpio/rockchip: add driver for rockchip gpio")
Assisted-by: Antigravity:gemini-3.5-flash
Signed-off-by: Marco Scardovi <scardracs@disroot.org>
Link: https://patch.msgid.link/20260526171050.12785-3-scardracs@disroot.org
[Bartosz: don't emit an error message on devres allocation failure]
Signed-off-by: Bartosz Golaszewski <bartosz.golaszewski@oss.qualcomm.com>
The bank->clk was previously obtained via of_clk_get() and manually
prepared/enabled. However, it was missing a corresponding clk_put() in
both the error paths and the remove function, leading to a reference leak.
Convert the allocation to devm_clk_get_enabled(), which also properly
propagates failures from clk_prepare_enable() that were previously ignored.
The GPIO bank device uses the same OF node as the previous of_clk_get()
call, so devm_clk_get_enabled(dev, NULL) correctly resolves the same
clock provider entry.
Fix the reference leak and simplify the code by removing the manual
clk_disable_unprepare() calls in the probe error paths and in the
remove function.
Fixes: 936ee2675e ("gpio/rockchip: add driver for rockchip gpio")
Assisted-by: Antigravity:gemini-3.5-flash
Signed-off-by: Marco Scardovi <scardracs@disroot.org>
Link: https://patch.msgid.link/20260526171050.12785-2-scardracs@disroot.org
Signed-off-by: Bartosz Golaszewski <bartosz.golaszewski@oss.qualcomm.com>
If *ppos is non-zero (user-space write split over multiple calls to
write()) then simple_write_to_buffer() won't initialize the start of the
buffer. Really, non-zero values for *ppos aren't going to work at all.
Check for that and return -EINVAL at the start of the function.
Fixes: 91581c4b3f ("gpio: virtuser: new virtual testing driver for the GPIO API")
Signed-off-by: Dan Carpenter <error27@gmail.com>
Link: https://patch.msgid.link/ahP3BJWWy-m_qI0X@stanley.mountain
Signed-off-by: Bartosz Golaszewski <bartosz.golaszewski@oss.qualcomm.com>
By the time gpio_device_teardown_shared() is called, the parent device
is gone from the global list of GPIO devices and all outstanding SRCU
read-side critical sections have completed. That means that no
concurrent gpio_find_and_request() can call
gpio_shared_add_proxy_lookup() for this device at this time. There's
also no risk of the parent device being re-bound to the driver before
the unbinding completes (including the child devices).
Lockdep produces a false-positive report about a possible circular
dependency as it doesn't know the ordering guarantee. Not taking the
ref->lock in gpio_device_teardown_shared() silences it and is safe to do.
Cc: stable@vger.kernel.org
Fixes: ea513dd3c0 ("gpio: shared: make locking more fine-grained")
Reviewed-by: Linus Walleij <linusw@kernel.org>
Link: https://patch.msgid.link/20260522-gpio-shared-deadlock-v1-2-76bca088f8c0@oss.qualcomm.com
Signed-off-by: Bartosz Golaszewski <bartosz.golaszewski@oss.qualcomm.com>
Commit 710abda580 ("gpio: shared: call gpio_chip::of_xlate() if set")
used the mutex embedded in struct gpio_shared_entry to protect the
offset field which now can be modified after assignment. The critical
section however is too wide and introduced a potential deadlock on the
removal of the shared GPIO proxy's parent.
Make the critical section shorter - only protect the offset when it's
being read.
While at it: mention the fact that the entry lock is now also used to
protect against concurrent access to the offset field in the structure's
documentation.
Cc: stable@vger.kernel.org
Fixes: 710abda580 ("gpio: shared: call gpio_chip::of_xlate() if set")
Reviewed-by: Linus Walleij <linusw@kernel.org>
Link: https://patch.msgid.link/20260522-gpio-shared-deadlock-v1-1-76bca088f8c0@oss.qualcomm.com
Signed-off-by: Bartosz Golaszewski <bartosz.golaszewski@oss.qualcomm.com>
Ivan Vecera says:
====================
dpll: zl3073x: various fixes
Three fixes for the zl3073x DPLL driver.
Patch 1 exports __dpll_device_change_ntf() for use by drivers that
need to send device change notifications from within callbacks
already running under dpll_lock.
Patch 2 replaces the change_work workqueue mechanism with direct
calls to __dpll_device_change_ntf(), eliminating a race condition
where the work handler could dereference a freed dpll_dev pointer
during device teardown.
Patch 3 moves the freq_monitor flag from per-DPLL to per-device
scope to match the hardware behavior where frequency measurement
registers are shared across all DPLL channels.
====================
Link: https://patch.msgid.link/20260526074525.1451008-1-ivecera@redhat.com
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
The frequency monitoring feature uses shared hardware registers
that measure input reference frequencies independently of
individual DPLL channels. However, the freq_monitor flag was
incorrectly placed in the per-DPLL structure, causing each
channel to track its own enable/disable state independently.
Since the DPLL core calls measured_freq_get() only for the first
pin registration, the measured_freq_check() in the periodic worker
was gated by the per-DPLL freq_monitor flag of whichever channel
happens to be checked. If the first DPLL channel had frequency
monitoring disabled while another had it enabled, measurements
were never reported.
Move freq_monitor from struct zl3073x_dpll to struct zl3073x_dev
so all DPLL channels share a single flag, matching the hardware
behavior. Update freq_monitor_set() to notify other DPLL devices
about the change (like phase_offset_avg_factor_set() already does)
and remove the mode-dependent guard in zl3073x_dpll_changes_check()
since all input pin monitoring (pin state, phase offset, FFO, and
measured frequency) works correctly in all DPLL modes.
Fixes: bfc923b642 ("dpll: zl3073x: implement frequency monitoring")
Signed-off-by: Ivan Vecera <ivecera@redhat.com>
Link: https://patch.msgid.link/20260526074525.1451008-4-ivecera@redhat.com
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
The change_work was introduced to send device change notifications
from DPLL device callbacks without deadlocking on dpll_lock, since
the callbacks are already invoked under that lock. Now that
__dpll_device_change_ntf() is exported for callers that already
hold dpll_lock, use it directly and remove the change_work
infrastructure entirely.
This eliminates a race condition where change_work could be
re-scheduled after cancel_work_sync() during device teardown,
potentially causing the handler to dereference a freed or NULL
dpll_dev pointer.
Fixes: 9363b48376 ("dpll: zl3073x: Allow to configure phase offset averaging factor")
Signed-off-by: Ivan Vecera <ivecera@redhat.com>
Link: https://patch.msgid.link/20260526074525.1451008-3-ivecera@redhat.com
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
Export __dpll_device_change_ntf() so that drivers can send device
change notifications from within device callbacks, which are already
called under dpll_lock. Using dpll_device_change_ntf() in that
context would deadlock.
Add lockdep_assert_held() to catch misuse without the lock held.
Signed-off-by: Ivan Vecera <ivecera@redhat.com>
Reviewed-by: Jiri Pirko <jiri@nvidia.com>
Link: https://patch.msgid.link/20260526074525.1451008-2-ivecera@redhat.com
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
Chuck Lever says:
====================
net/handshake: anchor request lifetime to a pinned file reference
handshake_nl_accept_doit() has accumulated four follow-on fixes
since 3b3009ea8a ("net/handshake: Create a NETLINK service for
handling handshake requests"): 7ea9c1ec66, 7798b59409,
fe67b063f6, and dabac51b81. Each was a local refcount or
NULL-check correction; none moved where the file reference is
owned, and the same code keeps producing the same class of bug.
Reworking the ownership is what breaks the pattern.
For the duration of a request, sock->file has no single owner.
Submit publishes the request without taking a file reference;
accept_doit acquires one inside the handler, after the request
has already left the pending list. The consumer can drop its
own reference at any time, including the moment between
handshake_req_next() popping the request and accept_doit
reaching get_file(). The submit-side sock_hold() pins only
struct sock; struct socket and sock->file remain under the
consumer's control via the file descriptor.
This series places the file reference under unambiguous
ownership. handshake_req_submit() pins it on the request and
completion or cancel drops it (patches 4-5); the submit-side
sock_hold() then becomes redundant, and dropping it also closes
a publish-before-pin race the late sock_hold itself opened
(patch 6). The handshake_complete() API and its consumers move
to a uniform negative-errno sign convention (patch 3), with the
matching sign correction in nvme-tcp (patch 2). Patch 1
hardens hn_lock for BH context, the netns-exit drain fix
builds on the new file-pin infrastructure (patch 8), and new
KUnit file-count assertions verify the refcount contract
(patch 7).
Three things in this restructuring want a careful look. In
handshake_complete(), the fput() of the request's file
reference has to come after hp_done() -- fput() can transitively
run handshake_sk_destruct() and free the request, so the patch
stashes hr_file in a local first. handshake_sk_destruct()
itself is kept on purpose: it owns rhashtable removal and
kfree, and remains the backstop if a consumer path bypasses
handshake_complete() entirely. Third, handshake_req_next() now
returns its request with an extra get_file() held under
hn_lock; accept_doit must consume that reference (FD_PREPARE on
success, explicit fput on the fdf.err path), and any future
caller has to honor the same contract.
v2: https://patch.msgid.link/20260521-handshake-file-pin-v2-0-b9dadc472840@oracle.com
v1: https://patch.msgid.link/20260518-handshake-file-pin-v1-0-4bbcb7e62fda@oracle.com
====================
Link: https://patch.msgid.link/20260525-handshake-file-pin-v3-0-66c616906ead@oracle.com
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
The arguments to list_splice_init() in handshake_net_exit() are
reversed. The call moves the local empty "requests" list onto
hn->hn_requests, leaving the local list empty, so the subsequent
drain loop runs zero iterations. Pending handshake requests that
had not yet been accepted are not torn down when the net namespace
is destroyed; each one keeps a reference on a socket file and on
the handshake_req allocation.
Pass the source and destination in the documented order
(list_splice_init(list, head) moves list onto head) so the pending
list is transferred to the local scratch list and drained through
handshake_complete().
Fixing the splice direction exposes a list-corruption race. After
the splice each req->hr_list still has non-empty link pointers,
threading the stack-local scratch list rather than hn_requests.
A concurrent handshake_req_cancel() -- for example, from sunrpc's
TLS timeout on a kernel socket whose netns reference was not
taken -- finds the request through the rhashtable, calls
remove_pending(), and sees !list_empty(&req->hr_list).
__remove_pending_locked() then list_del_init()s an entry off the
scratch list while the drain iterates, corrupting it. The same
call arriving after the drain loop has run list_del() on an
entry hits LIST_POISON instead.
Have remove_pending() check HANDSHAKE_F_NET_DRAINING under
hn_lock and report not-found when drain is in progress. The
drain has already taken ownership; handshake_complete()'s existing
test_and_set on HANDSHAKE_F_REQ_COMPLETED still arbitrates
between drain and cancel for who calls the consumer's hp_done. Use
list_del_init() rather than list_del() in the drain so req->hr_list
does not carry LIST_POISON after drain releases the entry.
The DRAINING guard in remove_pending() makes cancel return false,
but cancel still falls through to test_and_set_bit on
HANDSHAKE_F_REQ_COMPLETED and drops the request's hr_file reference.
Without another pin, if that is the last reference, sk_destruct frees
the request while it is still linked on the drain loop's local list.
Pin each request's hr_file under hn_lock before releasing the list,
and drop that drain pin after the loop finishes with the request.
Fixes: 3b3009ea8a ("net/handshake: Create a NETLINK service for handling handshake requests")
Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Reviewed-by: Hannes Reinecke <hare@kernel.org>
Link: https://patch.msgid.link/20260525-handshake-file-pin-v3-8-66c616906ead@oracle.com
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
The new file-reference contract on struct handshake_req is silently
breakable: a missing get_file() at submit or a missing fput() on an
error path leaves the file leaked but does not crash the test, so
the existing absence-of-crash checks pass either way.
Snapshot file_count(filp) before each handshake_req_submit() in
the submit-success, EAGAIN, EBUSY, and cancel tests, and assert
the expected balance after submit and again after cancel. The
already-completed cancel test also asserts the post-complete
balance, which pins down that handshake_complete() drops the
reference and that the subsequent cancel does not double-fput.
The destroy test gets the same treatment before __fput_sync(),
which double-checks that cancel's fput() ran and the only
remaining reference is the one sock_alloc_file() established.
Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Reviewed-by: Hannes Reinecke <hare@kernel.org>
Link: https://patch.msgid.link/20260525-handshake-file-pin-v3-7-66c616906ead@oracle.com
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
handshake_req_submit() publishes the request via
handshake_req_hash_add() and __add_pending_locked(), drops
hn_lock, and calls handshake_genl_notify() (which can sleep)
before taking sock_hold() on req->hr_sk. A fast tlshd ACCEPT
followed by DONE can drive handshake_complete()'s sock_put()
into the window between the spin_unlock and the late
sock_hold(); on a system where the consumer's fd held the
only sk reference, the late sock_hold() then operates on an
sk whose refcount has reached zero.
The preceding two patches install an explicit file reference
on struct handshake_req. That file pins sock->file, which
pins the embedded struct socket, which defers inet_release()'s
sock_put(). As long as hr_file is held, sk cannot reach refcount
zero from the consumer side, and the submit-side sock_hold()
with its matching sock_put() calls in handshake_complete() and
handshake_req_cancel() is now redundant.
Drop all three. The file reference already keeps each request's
socket alive, and the lifetime story is contained in a single
get_file()/fput() pair.
Fixes: 3b3009ea8a ("net/handshake: Create a NETLINK service for handling handshake requests")
Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Reviewed-by: Hannes Reinecke <hare@kernel.org>
Link: https://patch.msgid.link/20260525-handshake-file-pin-v3-6-66c616906ead@oracle.com
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
handshake_req_next() removes the request from the per-net
pending list and drops hn_lock before handshake_nl_accept_doit()
reads req->hr_sk->sk_socket and dereferences sock->file (once in
FD_PREPARE() and again in get_file()). In that window a
consumer running tls_handshake_cancel() followed by sockfd_put()
(svc_sock_free) or __fput_sync() (xs_reset_transport) releases
sock->file. sock_release() then runs sock_orphan(), zeroing
sk_socket, and frees the struct socket. The accept-side code
either reads NULL through sk_socket or chases freed memory.
The submit-side sock_hold() does not prevent this. sk_refcnt
protects struct sock, but struct socket and sock->file are
independently refcounted via the file descriptor the consumer
owns. Pinning sk leaves sock and sock->file unprotected.
Retarget the accept-side dereferences at req->hr_file, which was
pinned at submit time, instead of req->hr_sk->sk_socket->file.
Pinning on its own is not sufficient: a consumer that cancels
between handshake_req_next() returning and accept_doit reaching
FD_PREPARE() takes the !remove_pending() branch in
handshake_req_cancel() and drops hr_file before the accept side
takes its own reference. Hand off an additional file reference
inside handshake_req_next(), under hn_lock, so the accept side
operates on a reference that no concurrent handshake_req_cancel()
can revoke. FD_PREPARE() consumes that handed-off reference,
either by transferring it to the new fd in fd_publish() or by
dropping it in the cleanup destructor on error; the explicit
get_file() that previously balanced FD_PREPARE() is therefore
redundant and goes away.
Update handshake_req_cancel_test2 and _test3 to simulate the
FD_PREPARE() consumption with an fput() so the kunit file-count
assertions stay balanced.
Reported-by: Chris Mason <clm@meta.com>
Fixes: 3b3009ea8a ("net/handshake: Create a NETLINK service for handling handshake requests")
Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Reviewed-by: Hannes Reinecke <hare@kernel.org>
Link: https://patch.msgid.link/20260525-handshake-file-pin-v3-5-66c616906ead@oracle.com
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
