[ Upstream commit b47adaab8b ]
In veth_get_ethtool_stats(), some statistics protected by
u64_stats_sync, are read and accumulated in ignorance of possible
u64_stats_fetch_retry() events. These statistics, peer_tq_xdp_xmit and
peer_tq_xdp_xmit_err, are already accumulated by veth_xdp_xmit(). Fix
this by reading them into a temporary buffer first.
Fixes: 5fe6e56776 ("veth: rely on peer veth_rq for ndo_xdp_xmit accounting")
Signed-off-by: David Yang <mmyangfl@gmail.com>
Link: https://patch.msgid.link/20260114122450.227982-1-mmyangfl@gmail.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
[ Upstream commit 7a9bc9e3f4 ]
fou_udp_recv() has the same problem mentioned in the previous
patch.
If FOU_ATTR_IPPROTO is set to 0, skb is not freed by
fou_udp_recv() nor "resubmit"-ted in ip_protocol_deliver_rcu().
Let's forbid 0 for FOU_ATTR_IPPROTO.
Fixes: 23461551c0 ("fou: Support for foo-over-udp RX path")
Signed-off-by: Kuniyuki Iwashima <kuniyu@google.com>
Reviewed-by: Eric Dumazet <edumazet@google.com>
Link: https://patch.msgid.link/20260115172533.693652-4-kuniyu@google.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
[ Upstream commit c158f985cf ]
On the receive path, packet can be damaged because of buffer
overflow in Rx FIFO. Avoid misleading per-packet error log when
packet->errors is set, this can flood the log. Instead, rely on the
standard rtnl_link_stats64 stats.
Fixes: c5aa9e3b81 ("amd-xgbe: Initial AMD 10GbE platform driver")
Signed-off-by: Raju Rangoju <Raju.Rangoju@amd.com>
Link: https://patch.msgid.link/20260114163037.2062606-1-Raju.Rangoju@amd.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
[ Upstream commit a80c9d945a ]
A null-ptr-deref was reported in the SCTP transmit path when SCTP-AUTH key
initialization fails:
==================================================================
KASAN: null-ptr-deref in range [0x0000000000000018-0x000000000000001f]
CPU: 0 PID: 16 Comm: ksoftirqd/0 Tainted: G W 6.6.0 #2
RIP: 0010:sctp_packet_bundle_auth net/sctp/output.c:264 [inline]
RIP: 0010:sctp_packet_append_chunk+0xb36/0x1260 net/sctp/output.c:401
Call Trace:
sctp_packet_transmit_chunk+0x31/0x250 net/sctp/output.c:189
sctp_outq_flush_data+0xa29/0x26d0 net/sctp/outqueue.c:1111
sctp_outq_flush+0xc80/0x1240 net/sctp/outqueue.c:1217
sctp_cmd_interpreter.isra.0+0x19a5/0x62c0 net/sctp/sm_sideeffect.c:1787
sctp_side_effects net/sctp/sm_sideeffect.c:1198 [inline]
sctp_do_sm+0x1a3/0x670 net/sctp/sm_sideeffect.c:1169
sctp_assoc_bh_rcv+0x33e/0x640 net/sctp/associola.c:1052
sctp_inq_push+0x1dd/0x280 net/sctp/inqueue.c:88
sctp_rcv+0x11ae/0x3100 net/sctp/input.c:243
sctp6_rcv+0x3d/0x60 net/sctp/ipv6.c:1127
The issue is triggered when sctp_auth_asoc_init_active_key() fails in
sctp_sf_do_5_1C_ack() while processing an INIT_ACK. In this case, the
command sequence is currently:
- SCTP_CMD_PEER_INIT
- SCTP_CMD_TIMER_STOP (T1_INIT)
- SCTP_CMD_TIMER_START (T1_COOKIE)
- SCTP_CMD_NEW_STATE (COOKIE_ECHOED)
- SCTP_CMD_ASSOC_SHKEY
- SCTP_CMD_GEN_COOKIE_ECHO
If SCTP_CMD_ASSOC_SHKEY fails, asoc->shkey remains NULL, while
asoc->peer.auth_capable and asoc->peer.peer_chunks have already been set by
SCTP_CMD_PEER_INIT. This allows a DATA chunk with auth = 1 and shkey = NULL
to be queued by sctp_datamsg_from_user().
Since command interpretation stops on failure, no COOKIE_ECHO should been
sent via SCTP_CMD_GEN_COOKIE_ECHO. However, the T1_COOKIE timer has already
been started, and it may enqueue a COOKIE_ECHO into the outqueue later. As
a result, the DATA chunk can be transmitted together with the COOKIE_ECHO
in sctp_outq_flush_data(), leading to the observed issue.
Similar to the other places where it calls sctp_auth_asoc_init_active_key()
right after sctp_process_init(), this patch moves the SCTP_CMD_ASSOC_SHKEY
immediately after SCTP_CMD_PEER_INIT, before stopping T1_INIT and starting
T1_COOKIE. This ensures that if shared key generation fails, authenticated
DATA cannot be sent. It also allows the T1_INIT timer to retransmit INIT,
giving the client another chance to process INIT_ACK and retry key setup.
Fixes: 730fc3d05c ("[SCTP]: Implete SCTP-AUTH parameter processing")
Reported-by: Zhen Chen <chenzhen126@huawei.com>
Tested-by: Zhen Chen <chenzhen126@huawei.com>
Signed-off-by: Xin Long <lucien.xin@gmail.com>
Link: https://patch.msgid.link/44881224b375aa8853f5e19b4055a1a56d895813.1768324226.git.lucien.xin@gmail.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
[ Upstream commit 79a6d1bfe1 ]
In commit 7352e1d593 ("can: gs_usb: gs_usb_receive_bulk_callback(): fix
URB memory leak"), the URB was re-anchored before usb_submit_urb() in
gs_usb_receive_bulk_callback() to prevent a leak of this URB during
cleanup.
However, this patch did not take into account that usb_submit_urb() could
fail. The URB remains anchored and
usb_kill_anchored_urbs(&parent->rx_submitted) in gs_can_close() loops
infinitely since the anchor list never becomes empty.
To fix the bug, unanchor the URB when an usb_submit_urb() error occurs,
also print an info message.
Fixes: 7352e1d593 ("can: gs_usb: gs_usb_receive_bulk_callback(): fix URB memory leak")
Reported-by: Jakub Kicinski <kuba@kernel.org>
Closes: https://lore.kernel.org/all/20260110223836.3890248-1-kuba@kernel.org/
Link: https://patch.msgid.link/20260116-can_usb-fix-reanchor-v1-1-9d74e7289225@pengutronix.de
Signed-off-by: Marc Kleine-Budde <mkl@pengutronix.de>
Signed-off-by: Sasha Levin <sashal@kernel.org>
[ Upstream commit a74c7a58ca ]
In ucc_geth's .mac_config(), we configure the TBI Serdes block represented by a
struct phy_device that we get from firmware.
While porting to phylink, a check was missed to make sure we don't try
to access the TBI PHY if we can't get it. Let's add it and return early
in case of error
Reported-by: kernel test robot <lkp@intel.com>
Reported-by: Dan Carpenter <dan.carpenter@linaro.org>
Closes: https://lore.kernel.org/r/202601130843.rFGNXA5a-lkp@intel.com/
Fixes: 53036aa8d0 ("net: freescale: ucc_geth: phylink conversion")
Signed-off-by: Maxime Chevallier <maxime.chevallier@bootlin.com>
Link: https://patch.msgid.link/20260114080247.366252-1-maxime.chevallier@bootlin.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
[ Upstream commit 4f5f148dd7 ]
Currently, the test breaks if the SUT already has a default route
configured for IPv6. Fix by avoiding the use of the default namespace.
Fixes: 4ed591c8ab ("net/ipv6: Allow onlink routes to have a device mismatch if it is the default route")
Suggested-by: Fernando Fernandez Mancera <fmancera@suse.de>
Signed-off-by: Ricardo B. Marlière <rbm@suse.com>
Reviewed-by: Ido Schimmel <idosch@nvidia.com>
Reviewed-by: Fernando Fernandez Mancera <fmancera@suse.de>
Link: https://patch.msgid.link/20260113-selftests-net-fib-onlink-v2-1-89de2b931389@suse.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
[ Upstream commit 7d7dbafefb ]
The SR9700 chip sends more than one packet in a USB transaction,
like the DM962x chips can optionally do, but the dm9601 driver does not
support this mode, and the hardware does not have the DM962x
MODE_CTL register to disable it, so this driver drops packets on SR9700
devices. The sr9700 driver correctly handles receiving more than one
packet per transaction.
While the dm9601 driver could be improved to handle this, the easiest
way to fix this issue in the short term is to remove the SR9700 device
ID from the dm9601 driver so the sr9700 driver is always used. This
device ID should not have been in more than one driver to begin with.
The "Fixes" commit was chosen so that the patch is automatically
included in all kernels that have the sr9700 driver, even though the
issue affects dm9601.
Fixes: c9b37458e9 ("USB2NET : SR9700 : One chip USB 1.1 USB2NET SR9700Device Driver Support")
Signed-off-by: Ethan Nelson-Moore <enelsonmoore@gmail.com>
Acked-by: Peter Korsgaard <peter@korsgaard.com>
Link: https://patch.msgid.link/20260113063924.74464-1-enelsonmoore@gmail.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
[ Upstream commit 0386bd321d ]
vsock/virtio common tries to coalesce buffers in rx queue: if a linear skb
(with a spare tail room) is followed by a small skb (length limited by
GOOD_COPY_LEN = 128), an attempt is made to join them.
Since the introduction of MSG_ZEROCOPY support, assumption that a small skb
will always be linear is incorrect. In the zerocopy case, data is lost and
the linear skb is appended with uninitialized kernel memory.
Of all 3 supported virtio-based transports, only loopback-transport is
affected. G2H virtio-transport rx queue operates on explicitly linear skbs;
see virtio_vsock_alloc_linear_skb() in virtio_vsock_rx_fill(). H2G
vhost-transport may allocate non-linear skbs, but only for sizes that are
not considered for coalescence; see PAGE_ALLOC_COSTLY_ORDER in
virtio_vsock_alloc_skb().
Ensure only linear skbs are coalesced. Note that skb_tailroom(last_skb) > 0
guarantees last_skb is linear.
Fixes: 581512a6dc ("vsock/virtio: MSG_ZEROCOPY flag support")
Signed-off-by: Michal Luczaj <mhal@rbox.co>
Reviewed-by: Stefano Garzarella <sgarzare@redhat.com>
Link: https://patch.msgid.link/20260113-vsock-recv-coalescence-v2-1-552b17837cf4@rbox.co
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
[ Upstream commit 8ad1b6c1e6 ]
The previous 7 KB per queue caused TX unit hangs under heavy
timestamping load. Reducing to 5 KB avoids these hangs and matches
the TSN recommendation in I225/I226 SW User Manual Section 7.5.4.
The 8 KB "freed" by this change is currently unused. This reduction
is not expected to impact throughput, as the i226 is PCIe-limited
for small TSN packets rather than TX-buffer-limited.
Fixes: 0d58cdc902 ("igc: optimize TX packet buffer utilization for TSN mode")
Reported-by: Zdenek Bouska <zdenek.bouska@siemens.com>
Closes: https://lore.kernel.org/netdev/AS1PR10MB5675DBFE7CE5F2A9336ABFA4EBEAA@AS1PR10MB5675.EURPRD10.PROD.OUTLOOK.COM/
Reviewed-by: Paul Menzel <pmenzel@molgen.mpg.de>
Reviewed-by: Simon Horman <horms@kernel.org>
Reviewed-by: Aleksandr Loktionov <aleksandr.loktionov@intel.com>
Signed-off-by: Chwee-Lin Choong <chwee.lin.choong@intel.com>
Tested-by: Avigail Dahan <avigailx.dahan@intel.com>
Signed-off-by: Tony Nguyen <anthony.l.nguyen@intel.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
[ Upstream commit 6990dc392a ]
The current HW bug workaround checks the TXTT_0 ready bit first,
then reads TXSTMPL_0 twice (before and after reading TXSTMPH_0)
to detect whether a new timestamp was captured by timestamp
register 0 during the workaround.
This sequence has a race: if a new timestamp is captured after
checking the TXTT_0 bit but before the first TXSTMPL_0 read, the
detection fails because both the "old" and "new" values come from
the same timestamp.
Fix by reading TXSTMPL_0 first to establish a baseline, then
checking the TXTT_0 bit. This ensures any timestamp captured
during the race window will be detected.
Old sequence:
1. Check TXTT_0 ready bit
2. Read TXSTMPL_0 (baseline)
3. Read TXSTMPH_0 (interrupt workaround)
4. Read TXSTMPL_0 (detect changes vs baseline)
New sequence:
1. Read TXSTMPL_0 (baseline)
2. Check TXTT_0 ready bit
3. Read TXSTMPH_0 (interrupt workaround)
4. Read TXSTMPL_0 (detect changes vs baseline)
Fixes: c789ad7cbe ("igc: Work around HW bug causing missing timestamps")
Suggested-by: Avi Shalev <avi.shalev@intel.com>
Reviewed-by: Aleksandr Loktionov <aleksandr.loktionov@intel.com>
Co-developed-by: Song Yoong Siang <yoong.siang.song@intel.com>
Signed-off-by: Song Yoong Siang <yoong.siang.song@intel.com>
Signed-off-by: Chwee-Lin Choong <chwee.lin.choong@intel.com>
Tested-by: Avigail Dahan <avigailx.dahan@intel.com>
Signed-off-by: Tony Nguyen <anthony.l.nguyen@intel.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
[ Upstream commit 41a9a6826f ]
The Multi-queue Priority (MQPRIO) and Earliest TxTime First (ETF) offloads
utilize the Time Sensitive Networking (TSN) Tx mode. This mode is always
coupled to IEEE 802.1Qbv time aware shaper (Qbv). Therefore, the driver
sets a default Qbv schedule of all gates opened and a cycle time of
1s. This schedule is set during probe.
However, the following sequence of events lead to Tx issues:
- Boot a dual core system
igc_probe():
igc_tsn_clear_schedule():
-> Default Schedule is set
Note: At this point the driver has allocated two Tx/Rx queues, because
there are only two CPUs.
- ethtool -L enp3s0 combined 4
igc_ethtool_set_channels():
igc_reinit_queues()
-> Default schedule is gone, per Tx ring start and end time are zero
- tc qdisc replace dev enp3s0 handle 100 parent root mqprio \
num_tc 4 map 3 3 2 2 0 1 1 1 3 3 3 3 3 3 3 3 \
queues 1@0 1@1 1@2 1@3 hw 1
igc_tsn_offload_apply():
igc_tsn_enable_offload():
-> Writes zeros to IGC_STQT(i) and IGC_ENDQT(i), causing Tx to stall/fail
Therefore, restore the default Qbv schedule after changing the number of
channels.
Furthermore, add a restriction to not allow queue reconfiguration when
TSN/Qbv is enabled, because it may lead to inconsistent states.
Fixes: c814a2d2d4 ("igc: Use default cycle 'start' and 'end' values for queues")
Signed-off-by: Kurt Kanzenbach <kurt@linutronix.de>
Reviewed-by: Aleksandr Loktionov <aleksandr.loktionov@intel.com>
Tested-by: Avigail Dahan <avigailx.dahan@intel.com>
Acked-by: Vinicius Costa Gomes <vinicius.gomes@intel.com>
Signed-off-by: Tony Nguyen <anthony.l.nguyen@intel.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
[ Upstream commit 01139a2ce5 ]
The commit 5f6df173f9 ("ice: implement and use rd32_poll_timeout for
ice_sq_done timeout") converted ICE_CTL_Q_SQ_CMD_TIMEOUT from jiffies
to microseconds.
But the ice_release_res() function was missed, and its logic still
treats ICE_CTL_Q_SQ_CMD_TIMEOUT as a jiffies value.
So correct the issue by usecs_to_jiffies().
Found by inspection of the DDP downloading process.
Compile and modprobe tested only.
Fixes: 5f6df173f9 ("ice: implement and use rd32_poll_timeout for ice_sq_done timeout")
Signed-off-by: Ding Hui <dinghui@sangfor.com.cn>
Reviewed-by: Simon Horman <horms@kernel.org>
Reviewed-by: Aleksandr Loktionov <aleksandr.loktionov@intel.com>
Reviewed-by: Jacob Keller <jacob.e.keller@intel.com>
Reviewed-by: Paul Menzel <pmenzel@molgen.mpg.de>
Tested-by: Rinitha S <sx.rinitha@intel.com> (A Contingent worker at Intel)
Signed-off-by: Tony Nguyen <anthony.l.nguyen@intel.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
[ Upstream commit a9d45c22ed ]
When the user issues an administrative down to an interface that is the
primary for an aggregate bond, the prune lists are being purged. This
breaks communication to the secondary interface, which shares a prune
list on the main switch block while bonded together.
For the primary interface of an aggregate, avoid deleting these prune
lists during stop, and since they are hardcoded to specific values for
the default vlan and QinQ vlans, the attempt to re-add them during the
up phase will quietly fail without any additional problem.
Fixes: 1e0f9881ef ("ice: Flesh out implementation of support for SRIOV on bonded interface")
Reviewed-by: Jacob Keller <jacob.e.keller@intel.com>
Reviewed-by: Marcin Szycik <marcin.szycik@linux.intel.com>
Signed-off-by: Dave Ertman <david.m.ertman@intel.com>
Tested-by: Rinitha S <sx.rinitha@intel.com> (A Contingent worker at Intel)
Signed-off-by: Tony Nguyen <anthony.l.nguyen@intel.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
[ Upstream commit 8439016c3b ]
The u64_stats_sync structure is empty on 64-bit systems. However, on 32-bit
systems it contains a seqcount_t which needs to be initialized. While the
memory is zero-initialized, a lack of u64_stats_init means that lockdep
won't get initialized properly. Fix this by adding u64_stats_init() calls
to the rings just after allocation.
Fixes: 2b245cb294 ("ice: Implement transmit and NAPI support")
Reviewed-by: Aleksandr Loktionov <aleksandr.loktionov@intel.com>
Signed-off-by: Jacob Keller <jacob.e.keller@intel.com>
Reviewed-by: Simon Horman <horms@kernel.org>
Tested-by: Rinitha S <sx.rinitha@intel.com> (A Contingent worker at Intel)
Signed-off-by: Tony Nguyen <anthony.l.nguyen@intel.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
[ Upstream commit 3170757210 ]
Wrong P2P device link id value of 0 was introduced in ath12k_mac_op_tx() by [1].
During the P2P negotiation process, there is only one scan vdev with link ID 15.
Currently, the device link ID is incorrectly set to 0 in ath12k_mac_op_tx()
during the P2P negotiation process, which leads to TX failures.
Set the correct P2P device link ID to 15 to fix the TX failure issue.
Tested-on: WCN7850 hw2.0 PCI WLAN.HMT.1.1.c5-00302-QCAHMTSWPL_V1.0_V2.0_SILICONZ-1.115823.3
Fixes: 648a121baf ("wifi: ath12k: ath12k_mac_op_tx(): MLO support") # [1]
Signed-off-by: Yingying Tang <yingying.tang@oss.qualcomm.com>
Reviewed-by: Baochen Qiang <baochen.qiang@oss.qualcomm.com>
Reviewed-by: Vasanthakumar Thiagarajan <vasanthakumar.thiagarajan@oss.qualcomm.com>
Cc: linux-next@vger.kernel.org
Cc: netdev@vger.kernel.org
Link: https://patch.msgid.link/20260113054636.2620035-1-yingying.tang@oss.qualcomm.com
Signed-off-by: Jeff Johnson <jeff.johnson@oss.qualcomm.com>
[ Upstream commit f88e9fc30a ]
Commit [1] converted the management transmission work item into a
wiphy work. Since a wiphy work can only run under wiphy lock
protection, a race condition happens in below scenario:
1. a management frame is queued for transmission.
2. ath12k_mac_op_flush() gets called to flush pending frames associated
with the hardware (i.e, vif being NULL). Then in ath12k_mac_flush()
the process waits for the transmission done.
3. Since wiphy lock has been taken by the flush process, the transmission
work item has no chance to run, hence the dead lock.
>From user view, this dead lock results in below issue:
wlp8s0: authenticate with xxxxxx (local address=xxxxxx)
wlp8s0: send auth to xxxxxx (try 1/3)
wlp8s0: authenticate with xxxxxx (local address=xxxxxx)
wlp8s0: send auth to xxxxxx (try 1/3)
wlp8s0: authenticated
wlp8s0: associate with xxxxxx (try 1/3)
wlp8s0: aborting association with xxxxxx by local choice (Reason: 3=DEAUTH_LEAVING)
ath12k_pci 0000:08:00.0: failed to flush mgmt transmit queue, mgmt pkts pending 1
The dead lock can be avoided by invoking wiphy_work_flush() to proactively
run the queued work item. Note actually it is already present in
ath12k_mac_op_flush(), however it does not protect the case where vif
being NULL. Hence move it ahead to cover this case as well.
Tested-on: WCN7850 hw2.0 PCI WLAN.HMT.1.1.c5-00302-QCAHMTSWPL_V1.0_V2.0_SILICONZ-1.115823.3
Fixes: 56dcbf0b52 ("wifi: ath12k: convert struct ath12k::wmi_mgmt_tx_work to struct wiphy_work") # [1]
Reported-by: Stuart Hayhurst <stuart.a.hayhurst@gmail.com>
Closes: https://bugzilla.kernel.org/show_bug.cgi?id=220959
Signed-off-by: Baochen Qiang <baochen.qiang@oss.qualcomm.com>
Reviewed-by: Vasanthakumar Thiagarajan <vasanthakumar.thiagarajan@oss.qualcomm.com>
Link: https://patch.msgid.link/20260113-ath12k-fix-dead-lock-while-flushing-v1-1-9713621f3a0f@oss.qualcomm.com
Signed-off-by: Jeff Johnson <jeff.johnson@oss.qualcomm.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
[ Upstream commit 8b8d6ee53d ]
Scan finish workqueue was introduced in __ath12k_mac_scan_finish() by [1].
During ath12k_mac_op_cancel_remain_on_channel(), scan state is set to
ABORTING and should be reset to IDLE in the queued work. However,
wiphy_work_cancel() is called before exiting
ath12k_mac_op_cancel_remain_on_channel(), which prevents the work
from running and leaves the state in ABORTING. This blocks all
subsequent scan requests.
Replace wiphy_work_cancel() with wiphy_work_flush() to ensure the
queued work runs and scan state is reset to IDLE.
Tested-on: WCN7850 hw2.0 PCI WLAN.HMT.1.1.c5-00302-QCAHMTSWPL_V1.0_V2.0_SILICONZ-1.115823.3
Fixes: 3863f014ad ("wifi: ath12k: symmetrize scan vdev creation and deletion during HW scan") # [1]
Signed-off-by: Yingying Tang <yingying.tang@oss.qualcomm.com>
Reviewed-by: Vasanthakumar Thiagarajan <vasanthakumar.thiagarajan@oss.qualcomm.com>
Reviewed-by: Baochen Qiang <baochen.qiang@oss.qualcomm.com>
Link: https://patch.msgid.link/20260112115516.2144219-1-yingying.tang@oss.qualcomm.com
Signed-off-by: Jeff Johnson <jeff.johnson@oss.qualcomm.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
[ Upstream commit 39c90b1a1d ]
Cancel the scheduled scan request only on the vdev that has an active
scan running. Currently, ahvif->links_map is used to obtain the links,
but this includes links for which no scan is scheduled. In failure cases
where the scan fails due to an invalid channel definition, other links
which are not yet brought up (vdev not created) may also be accessed,
leading to the following trace:
Unable to handle kernel paging request at virtual address 0000000000004c8c
pc : _raw_spin_lock_bh+0x1c/0x54
lr : ath12k_scan_abort+0x20/0xc8 [ath12k]
Call trace:
_raw_spin_lock_bh+0x1c/0x54 (P)
ath12k_mac_op_cancel_hw_scan+0xac/0xc4 [ath12k]
ieee80211_scan_cancel+0xcc/0x12c [mac80211]
ieee80211_do_stop+0x6c4/0x7a8 [mac80211]
ieee80211_stop+0x60/0xd8 [mac80211]
Skip links that are not created or are not the current scan vdev. This
ensures only the scan for the matching links is aborted and avoids
aborting unrelated links during cancellation, thus aligning with how
start/cleanup manage ar->scan.arvif.
Also, remove the redundant arvif->is_started check from
ath12k_mac_op_cancel_hw_scan() that was introduced in commit 3863f014ad
("wifi: ath12k: symmetrize scan vdev creation and deletion during HW
scan") to avoid deleting the scan interface if the scan is triggered on
the existing AP vdev as this use case is already handled in
ath12k_scan_vdev_clean_work().
Tested-on: QCN9274 hw2.0 PCI WLAN.WBE.1.4.1-00199-QCAHKSWPL_SILICONZ-1
Fixes: feed05f152 ("wifi: ath12k: Split scan request for split band device")
Signed-off-by: Manish Dharanenthiran <manish.dharanenthiran@oss.qualcomm.com>
Reviewed-by: Baochen Qiang <baochen.qiang@oss.qualcomm.com>
Reviewed-by: Vasanthakumar Thiagarajan <vasanthakumar.thiagarajan@oss.qualcomm.com>
Link: https://patch.msgid.link/20260107-scan_vdev-v1-1-b600aedc645a@qti.qualcomm.com
Signed-off-by: Jeff Johnson <jeff.johnson@oss.qualcomm.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
[ Upstream commit c8c6fb886f ]
Commit d633b8a702 ("libata: print feature list on device scan")
added a print of the features supported by the device for ATA_DEV_ATA and
ATA_DEV_ZAC devices, but not for ATA_DEV_ATAPI devices.
Fix this by printing the features also for ATAPI devices.
Before changes:
ata1.00: ATAPI: Slimtype DVD A DU8AESH, 6C2M, max UDMA/133
After changes:
ata1.00: ATAPI: Slimtype DVD A DU8AESH, 6C2M, max UDMA/133
ata1.00: Features: Dev-Attention HIPM DIPM
Fixes: d633b8a702 ("libata: print feature list on device scan")
Signed-off-by: Niklas Cassel <cassel@kernel.org>
Tested-by: Wolf <wolf@yoxt.cc>
Signed-off-by: Damien Le Moal <dlemoal@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
[ Upstream commit 89531b68fc ]
ata_dev_print_features() is supposed to return early and not print anything
if there are no features supported.
However, commit b1f5af54f1 ("ata: libata-core: Advertize device support
for DIPM and HIPM features") added additional features to
ata_dev_print_features() without updating the early return conditional.
Add the missing features to the early return conditional.
Fixes: b1f5af54f1 ("ata: libata-core: Advertize device support for DIPM and HIPM features")
Signed-off-by: Niklas Cassel <cassel@kernel.org>
Tested-by: Wolf <wolf@yoxt.cc>
Signed-off-by: Damien Le Moal <dlemoal@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
[ Upstream commit a6bee5e524 ]
ata_dev_print_features() is supposed to return early and not print anything
if there are no features supported.
However, commit fe22e1c2f7 ("libata: support concurrent positioning
ranges log") added another feature to ata_dev_print_features() without
updating the early return conditional.
Add the missing feature to the early return conditional.
Fixes: fe22e1c2f7 ("libata: support concurrent positioning ranges log")
Signed-off-by: Niklas Cassel <cassel@kernel.org>
Tested-by: Wolf <wolf@yoxt.cc>
Signed-off-by: Damien Le Moal <dlemoal@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
[ Upstream commit ce83767ea3 ]
The link_power_management_supported sysfs attribute is currently set as
true even for ata ports that lack a .set_lpm() callback, e.g. dummy ports.
This is a bit silly, because while writing to the
link_power_management_policy sysfs attribute will make ata_scsi_lpm_store()
update ap->target_lpm_policy (thus sysfs will reflect the new value) and
call ata_port_schedule_eh() for the port, it is essentially a no-op.
This is because for a port without a .set_lpm() callback, once EH gets to
run, the ata_eh_link_set_lpm() will simply return, since the port does not
provide a .set_lpm() callback.
Thus, make sure that the link_power_management_supported sysfs attribute
is set to false for ports that lack a .set_lpm() callback. This way the
link_power_management_policy sysfs attribute will no longer be writable,
so we will no longer be misleading users to think that their sysfs write
actually does something.
Fixes: 0060beec0b ("ata: libata-sata: Add link_power_management_supported sysfs attribute")
Signed-off-by: Niklas Cassel <cassel@kernel.org>
Tested-by: Wolf <wolf@yoxt.cc>
Signed-off-by: Damien Le Moal <dlemoal@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
[ Upstream commit 8f3fb33f8f ]
Commit d360121832 ("ata: libata-core: Introduce ata_dev_config_lpm()")
introduced ata_dev_config_lpm(). However, it only called this function for
ATA_DEV_ATA and ATA_DEV_ZAC devices, not for ATA_DEV_ATAPI devices.
Additionally, commit d99a9142e7 ("ata: libata-core: Move device LPM quirk
settings to ata_dev_config_lpm()") moved the LPM quirk application from
ata_dev_configure() to ata_dev_config_lpm(), causing LPM quirks for ATAPI
devices to no longer be applied.
Call ata_dev_config_lpm() also for ATAPI devices, such that LPM quirks are
applied for ATAPI devices with an entry in __ata_dev_quirks once again.
Fixes: d360121832 ("ata: libata-core: Introduce ata_dev_config_lpm()")
Fixes: d99a9142e7 ("ata: libata-core: Move device LPM quirk settings to ata_dev_config_lpm()")
Signed-off-by: Niklas Cassel <cassel@kernel.org>
Tested-by: Wolf <wolf@yoxt.cc>
Signed-off-by: Damien Le Moal <dlemoal@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
[ Upstream commit ea4d4ea6d1 ]
An AHCI HBA specifies the number of ports it supports using CAP.NP.
The HBA is free to only make a subset of the number of ports available
using the PI (Ports Implemented) register.
libata currently creates dummy ports for HBA ports that are provided by
the HBA, but which are marked as "unavailable" using the PI register.
Each port will have a per port area of registers in the HBA, regardless
if the port is marked as "unavailable" or not.
ahci_mark_external_port() currently reads this per port area of registers
using readl() to see if the port is marked as external/hotplug-capable.
However, AHCI 1.3.1, section "3.1.4 Offset 0Ch: PI – Ports Implemented"
states: "Software must not read or write to registers within unavailable
ports."
Thus, make sure that we only call ahci_mark_external_port() and
ahci_update_initial_lpm_policy() for ports that are implemented.
From a libata perspective, this should not change anything related to LPM,
as dummy ports do not provide any ap->ops (they do not have a .set_lpm()
callback), so even if EH were to call .set_lpm() on a dummy port, it was
already a no-op.
Fixes: f713193523 ("ata: ahci: move marking of external port earlier")
Signed-off-by: Niklas Cassel <cassel@kernel.org>
Tested-by: Wolf <wolf@yoxt.cc>
Signed-off-by: Damien Le Moal <dlemoal@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
[ Upstream commit 1fed08c551 ]
freq_to_idx() is used to map a channel to a survey index. Commit
acc152f9be ("wifi: ath12k: combine channel list for split-phy devices in
single-wiphy") adds radio specific frequency range check in this helper to
make sure an invalid index is returned if the channel falls outside that
range. However, this check introduces a race, resulting in below warnings
as reported in [1].
ath12k_pci 0000:08:00.0: chan info: invalid frequency 6455 (idx 101 out of bounds)
ath12k_pci 0000:08:00.0: chan info: invalid frequency 6535 (idx 101 out of bounds)
ath12k_pci 0000:08:00.0: chan info: invalid frequency 6615 (idx 101 out of bounds)
ath12k_pci 0000:08:00.0: chan info: invalid frequency 6695 (idx 101 out of bounds)
ath12k_pci 0000:08:00.0: chan info: invalid frequency 6775 (idx 101 out of bounds)
ath12k_pci 0000:08:00.0: chan info: invalid frequency 6855 (idx 101 out of bounds)
ath12k_pci 0000:08:00.0: chan info: invalid frequency 6935 (idx 101 out of bounds)
ath12k_pci 0000:08:00.0: chan info: invalid frequency 7015 (idx 101 out of bounds)
ath12k_pci 0000:08:00.0: chan info: invalid frequency 7095 (idx 101 out of bounds)
ath12k_pci 0000:08:00.0: chan info: invalid frequency 6435 (idx 101 out of bounds)
Race scenario:
1) A regdomain covering below frequency range is uploaded to host via
WMI_REG_CHAN_LIST_CC_EXT_EVENTID event:
Country 00, CFG Regdomain UNSET FW Regdomain 0, num_reg_rules 6
1. (2402 - 2472 @ 40) (0, 20) (0 ms) (FLAGS 360448) (0, 0)
2. (2457 - 2477 @ 20) (0, 20) (0 ms) (FLAGS 360576) (0, 0)
3. (5170 - 5330 @ 160) (0, 20) (0 ms) (FLAGS 264320) (0, 0)
4. (5490 - 5730 @ 160) (0, 20) (0 ms) (FLAGS 264320) (0, 0)
5. (5735 - 5895 @ 160) (0, 20) (0 ms) (FLAGS 264320) (0, 0)
6. (5925 - 7125 @ 320) (0, 24) (0 ms) (FLAGS 2056) (0, 255)
As a result, radio frequency range is updated as [2402, 7125]
ath12k_pci 0000:08:00.0: mac pdev 0 freq limit updated. New range 2402->7125 MHz
If no scan in progress or after scan finished, command
WMI_SCAN_CHAN_LIST_CMDID is sent to firmware notifying that firmware
is allowed to do scan on all channels within that range.
The running path is:
/* redomain uploaded */
1. WMI_REG_CHAN_LIST_CC_EXT_EVENTID
2. ath12k_reg_chan_list_event()
3. ath12k_reg_handle_chan_list()
4. queue_work(..., &ar->regd_update_work)
5. ath12k_regd_update_work()
6. ath12k_regd_update()
/* update radio frequency range */
7. ath12k_mac_update_freq_range()
8. regulatory_set_wiphy_regd()
9. ath12k_reg_notifier()
10. ath12k_reg_update_chan_list()
11. queue_work(..., &ar->regd_channel_update_work)
12. ath12k_regd_update_chan_list_work()
/* wait scan finishes */
13. wait_for_completion_timeout(&ar->scan.completed, ...)
/* command notifying list of valid channels */
14. ath12k_wmi_send_scan_chan_list_cmd()
2) Hardware scan is triggered on all allowed channels.
3) Before scan completed, 11D mechanism detects a new country code
ath12k_pci 0000:08:00.0: wmi 11d new cc GB
With this code sent to firmware, firmware uploads a new regdomain
Country GB, CFG Regdomain ETSI FW Regdomain 2, num_reg_rules 9
1. (2402 - 2482 @ 40) (0, 20) (0 ms) (FLAGS 360448) (0, 0)
2. (5170 - 5250 @ 80) (0, 23) (0 ms) (FLAGS 264192) (0, 0)
3. (5250 - 5330 @ 80) (0, 23) (0 ms) (FLAGS 264216) (0, 0)
4. (5490 - 5590 @ 80) (0, 30) (0 ms) (FLAGS 264208)
5. (5590 - 5650 @ 40) (0, 30) (600000 ms) (FLAGS 264208)
6. (5650 - 5730 @ 80) (0, 30) (0 ms) (FLAGS 264208)
7. (5735 - 5875 @ 80) (0, 14) (0 ms) (FLAGS 264192) (0, 0)
8. (5855 - 5875 @ 20) (0, 14) (0 ms) (FLAGS 264192) (0, 0)
9. (5945 - 6425 @ 320) (0, 24) (0 ms) (FLAGS 2056) (0, 11)
Then radio frequency range is updated as [2402, 6425]
ath12k_pci 0000:08:00.0: mac pdev 0 freq limit updated. New range 2402->6425 MHz
Please note this is a smaller range than the previous one. Later host
runs the same path for the purpose of notifying the new channel list.
However since scan not completed, host just waits there. Meanwhile,
firmware is possibly scanning channels outside the new range. As a
result, WMI_CHAN_INFO_EVENTID events for those channels fail
freq_to_idx() check and triggers warnings above.
Fix this issue by removing radio frequency check in freq_to_idx(). This is
valid because channels being scanned do not synchronize with frequency
range update. Besides, this won't cause any problem, since freq_to_idx()
is only used for survey data. Even out-of-range channels filled in the
survey, they won't get delivered to userspace due to the range check
already there in ath12k_mac_op_get_survey().
Tested-on: WCN7850 hw2.0 PCI WLAN.HMT.1.1.c5-00302-QCAHMTSWPL_V1.0_V2.0_SILICONZ-1.115823.3
Fixes: acc152f9be ("wifi: ath12k: combine channel list for split-phy devices in single-wiphy")
Closes: https://bugzilla.kernel.org/show_bug.cgi?id=220871 # 1
Signed-off-by: Baochen Qiang <baochen.qiang@oss.qualcomm.com>
Link: https://patch.msgid.link/20260108-ath12k-fix-freq-to-idx-v1-1-b2458cf7aa0d@oss.qualcomm.com
Signed-off-by: Jeff Johnson <jeff.johnson@oss.qualcomm.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
[ Upstream commit 5bc3e720e7 ]
This was apparently accounted for in dt-bindings, but never made its
way into the driver.
Fix it for SC8280XP and its VDD_GFX-less cousin, SA8540P.
Fixes: f68f1cb343 ("soc: qcom: rpmhpd: add sc8280xp & sa8540p rpmh power-domains")
Reviewed-by: Dmitry Baryshkov <dmitry.baryshkov@oss.qualcomm.com>
Signed-off-by: Konrad Dybcio <konrad.dybcio@oss.qualcomm.com>
Reviewed-by: Ulf Hansson <ulf.hansson@linaro.org>
Link: https://lore.kernel.org/r/20251202-topic-8280_mxc-v2-2-46cdf47a829e@oss.qualcomm.com
Signed-off-by: Bjorn Andersson <andersson@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
[ Upstream commit 45e1be5dde ]
Not sure how useful it's gonna be in practice, but the definition is
missing (unlike the previously-unused SC8280XP_MXC-non-_AO), so add it
to allow the driver to create the corresponding pmdomain.
Fixes: dbfb5f94e0 ("dt-bindings: power: rpmpd: Add sc8280xp RPMh power-domains")
Acked-by: Rob Herring (Arm) <robh@kernel.org>
Signed-off-by: Konrad Dybcio <konrad.dybcio@oss.qualcomm.com>
Reviewed-by: Ulf Hansson <ulf.hansson@linaro.org>
Link: https://lore.kernel.org/r/20251202-topic-8280_mxc-v2-1-46cdf47a829e@oss.qualcomm.com
Signed-off-by: Bjorn Andersson <andersson@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
[ Upstream commit 1d8f69f453 ]
When the BLOCK_GROUP_TREE compat_ro flag is set, the extent root and
csum root fields are getting missed.
This is because EXTENT_TREE_V2 treated these differently, and when
they were split off this special-casing was mistakenly assigned to
BGT rather than the rump EXTENT_TREE_V2. There's no reason why the
existence of the block group tree should mean that we don't record the
details of the last commit's extent root and csum root.
Fix the code in backup_super_roots() so that the correct check gets
made.
Fixes: 1c56ab9919 ("btrfs: separate BLOCK_GROUP_TREE compat RO flag from EXTENT_TREE_V2")
Reviewed-by: Qu Wenruo <wqu@suse.com>
Signed-off-by: Mark Harmstone <mark@harmstone.com>
Reviewed-by: David Sterba <dsterba@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
[ Upstream commit 49f49d47af ]
hv_kmsg_dump() currently skips the panic notification entirely if it
doesn't get any message bytes to pass to Hyper-V due to an error from
kmsg_dump_get_buffer(). Skipping the notification is undesirable because
it leaves the Hyper-V host uncertain about the state of a panic'ed guest.
Fix this by always doing the panic notification, even if bytes_written
is zero. Also ensure that bytes_written is initialized, which fixes a
kernel test robot warning. The warning is actually bogus because
kmsg_dump_get_buffer() happens to set bytes_written even if it fails, and
in the kernel test robot's CONFIG_PRINTK not set case, hv_kmsg_dump() is
never called. But do the initialization for robustness and to quiet the
static checker.
Fixes: 9c318a1d9b ("Drivers: hv: move panic report code from vmbus to hv early init code")
Reported-by: kernel test robot <lkp@intel.com>
Reported-by: Dan Carpenter <dan.carpenter@linaro.org>
Closes: https://lore.kernel.org/all/202512172103.OcUspn1Z-lkp@intel.com/
Signed-off-by: Michael Kelley <mhklinux@outlook.com>
Reviewed-by: Roman Kisel <vdso@mailbox.org>
Signed-off-by: Wei Liu <wei.liu@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
[ Upstream commit 1eb217ab2e ]
If evsel__new_idx() returns NULL, the function currently jumps to label
'out_err'. Here, references to `cpus` and `pmu_cpus` are dropped.
Also, resources held by evsel->name and evsel->metric_id are freed.
But if evsel__new_idx() returns NULL, it can lead to NULL pointer
dereference.
Fixes: cd63c22168 ("perf parse-events: Minor __add_event refactoring")
Signed-off-by: Faisal Bukhari <faisalbukhari523@gmail.com>
Reviewed-by: Arnaldo Carvalho de Melo <acme@redhat.com>
Signed-off-by: Namhyung Kim <namhyung@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
[ Upstream commit 955b263c42 ]
According to RK3576 TRM part1 Table 1-1 Address Mapping, the size of
the GPU registers is 128 KB.
The current mapping incorrectly includes the addresses of multiple
following IP like the eInk interface at 0x27900000. This has not
been detected by the DT tooling as none of the extra mapped IP is
described in the upstream RK3576 DT so far.
Fixes: 57b1ce9039 ("arm64: dts: rockchip: Add rk3576 SoC base DT")
Signed-off-by: Chaoyi Chen <chaoyi.chen@rock-chips.com>
Reviewed-by: Nicolas Frattaroli <nicolas.frattaroli@collabora.com>
Reviewed-by: Sebastian Reichel <sebastian.reichel@collabora.com>
Link: https://patch.msgid.link/20260106071513.209-1-kernel@airkyi.com
Signed-off-by: Heiko Stuebner <heiko@sntech.de>
Signed-off-by: Sasha Levin <sashal@kernel.org>
[ Upstream commit 1f6ca55708 ]
With W=1, the following error comes up:
Warning (avoid_unnecessary_addr_size): /soc@0/usb@a600000: unnecessary #address-cells/#size-cells without "ranges", "dma-ranges" or child "reg" or "ranges" property
This is because the child node being removed during flattening and moving
to latest bindings.
Fixes: 77e1f16b93 ("arm64: dts: qcom: sm8650: Flatten the USB nodes")
Signed-off-by: Krishna Kurapati <krishna.kurapati@oss.qualcomm.com>
Reviewed-by: Krzysztof Kozlowski <krzysztof.kozlowski@oss.qualcomm.com>
Link: https://lore.kernel.org/r/20251203144856.2711440-3-krishna.kurapati@oss.qualcomm.com
Signed-off-by: Bjorn Andersson <andersson@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
[ Upstream commit 9dbc9bed01 ]
With W=1, the following error comes up:
Warning (avoid_unnecessary_addr_size): /soc@0/usb@a600000: unnecessary #address-cells/#size-cells without "ranges", "dma-ranges" or child "reg" or "ranges" property
This is because the child node being removed during flattening and moving
to latest bindings.
Fixes: 33450878ad ("arm64: dts: qcom: sm8550: Flatten the USB nodes")
Signed-off-by: Krishna Kurapati <krishna.kurapati@oss.qualcomm.com>
Reviewed-by: Krzysztof Kozlowski <krzysztof.kozlowski@oss.qualcomm.com>
Link: https://lore.kernel.org/r/20251203144856.2711440-2-krishna.kurapati@oss.qualcomm.com
Signed-off-by: Bjorn Andersson <andersson@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
[ Upstream commit 868b979c53 ]
To make sure that power rail is voted for, wire it up to its consumers.
Fixes: 152d1faf1e ("arm64: dts: qcom: add SC8280XP platform")
Signed-off-by: Konrad Dybcio <konrad.dybcio@oss.qualcomm.com>
Reviewed-by: Ulf Hansson <ulf.hansson@linaro.org>
Link: https://lore.kernel.org/r/20251202-topic-8280_mxc-v2-3-46cdf47a829e@oss.qualcomm.com
Signed-off-by: Bjorn Andersson <andersson@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
commit 4b5c493ff7 upstream.
A call to mmu_notifier_arch_invalidate_secondary_tlbs() was introduced in
commit e37d5a2d60 ("iommu/sva: invalidate stale IOTLB entries for kernel
address space") but without explicitly adding its corresponding header
file <linux/mmu_notifier.h>. This was evidenced while trying to enable
compile testing support for IOMMU_SVA:
config IOMMU_SVA
select IOMMU_MM_DATA
- bool
+ bool "Shared Virtual Addressing" if COMPILE_TEST
The thing is for certain architectures this header file is indirectly
included via <asm/tlbflush.h>. However, for others such as 32-bit arm the
header is missing and it results in a build failure:
$ make ARCH=arm allmodconfig
[...]
drivers/iommu/iommu-sva.c:340:3: error: call to undeclared function 'mmu_notifier_arch_invalidate_secondary_tlbs' [...]
340 | mmu_notifier_arch_invalidate_secondary_tlbs(iommu_mm->mm, start, end);
| ^
Fix this by including the appropriate header file.
Link: https://lkml.kernel.org/r/20260105190747.625082-1-cmllamas@google.com
Fixes: e37d5a2d60 ("iommu/sva: invalidate stale IOTLB entries for kernel address space")
Signed-off-by: Carlos Llamas <cmllamas@google.com>
Cc: Baolu Lu <baolu.lu@linux.intel.com>
Cc: Jason Gunthorpe <jgg@ziepe.ca>
Cc: Joerg Roedel <joro@8bytes.org>
Cc: Kevin Tian <kevin.tian@intel.com>
Cc: Robin Murphy <robin.murphy@arm.com>
Cc: Vasant Hegde <vasant.hegde@amd.com>
Cc: Will Deacon <will@kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
This reverts commit b49c766856 which is
commit e5bf5ee266 upstream.
It has been reported to cause test problems in Android devices. As the
other functionfs changes were not also backported at the same time,
something is out of sync. So just revert this one for now and it can
come back in the future as a patch series if it is tested.
Cc: Al Viro <viro@zeniv.linux.org.uk>
Cc: Sasha Levin <sashal@kernel.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit 038a102535 upstream.
The kernel test robot has reported:
BUG: spinlock trylock failure on UP on CPU#0, kcompactd0/28
lock: 0xffff888807e35ef0, .magic: dead4ead, .owner: kcompactd0/28, .owner_cpu: 0
CPU: 0 UID: 0 PID: 28 Comm: kcompactd0 Not tainted 6.18.0-rc5-00127-ga06157804399 #1 PREEMPT 8cc09ef94dcec767faa911515ce9e609c45db470
Call Trace:
<IRQ>
__dump_stack (lib/dump_stack.c:95)
dump_stack_lvl (lib/dump_stack.c:123)
dump_stack (lib/dump_stack.c:130)
spin_dump (kernel/locking/spinlock_debug.c:71)
do_raw_spin_trylock (kernel/locking/spinlock_debug.c:?)
_raw_spin_trylock (include/linux/spinlock_api_smp.h:89 kernel/locking/spinlock.c:138)
__free_frozen_pages (mm/page_alloc.c:2973)
___free_pages (mm/page_alloc.c:5295)
__free_pages (mm/page_alloc.c:5334)
tlb_remove_table_rcu (include/linux/mm.h:? include/linux/mm.h:3122 include/asm-generic/tlb.h:220 mm/mmu_gather.c:227 mm/mmu_gather.c:290)
? __cfi_tlb_remove_table_rcu (mm/mmu_gather.c:289)
? rcu_core (kernel/rcu/tree.c:?)
rcu_core (include/linux/rcupdate.h:341 kernel/rcu/tree.c:2607 kernel/rcu/tree.c:2861)
rcu_core_si (kernel/rcu/tree.c:2879)
handle_softirqs (arch/x86/include/asm/jump_label.h:36 include/trace/events/irq.h:142 kernel/softirq.c:623)
__irq_exit_rcu (arch/x86/include/asm/jump_label.h:36 kernel/softirq.c:725)
irq_exit_rcu (kernel/softirq.c:741)
sysvec_apic_timer_interrupt (arch/x86/kernel/apic/apic.c:1052)
</IRQ>
<TASK>
RIP: 0010:_raw_spin_unlock_irqrestore (arch/x86/include/asm/preempt.h:95 include/linux/spinlock_api_smp.h:152 kernel/locking/spinlock.c:194)
free_pcppages_bulk (mm/page_alloc.c:1494)
drain_pages_zone (include/linux/spinlock.h:391 mm/page_alloc.c:2632)
__drain_all_pages (mm/page_alloc.c:2731)
drain_all_pages (mm/page_alloc.c:2747)
kcompactd (mm/compaction.c:3115)
kthread (kernel/kthread.c:465)
? __cfi_kcompactd (mm/compaction.c:3166)
? __cfi_kthread (kernel/kthread.c:412)
ret_from_fork (arch/x86/kernel/process.c:164)
? __cfi_kthread (kernel/kthread.c:412)
ret_from_fork_asm (arch/x86/entry/entry_64.S:255)
</TASK>
Matthew has analyzed the report and identified that in drain_page_zone()
we are in a section protected by spin_lock(&pcp->lock) and then get an
interrupt that attempts spin_trylock() on the same lock. The code is
designed to work this way without disabling IRQs and occasionally fail the
trylock with a fallback. However, the SMP=n spinlock implementation
assumes spin_trylock() will always succeed, and thus it's normally a
no-op. Here the enabled lock debugging catches the problem, but otherwise
it could cause a corruption of the pcp structure.
The problem has been introduced by commit 5749077415 ("mm/page_alloc:
leave IRQs enabled for per-cpu page allocations"). The pcp locking scheme
recognizes the need for disabling IRQs to prevent nesting spin_trylock()
sections on SMP=n, but the need to prevent the nesting in spin_lock() has
not been recognized. Fix it by introducing local wrappers that change the
spin_lock() to spin_lock_iqsave() with SMP=n and use them in all places
that do spin_lock(&pcp->lock).
[vbabka@suse.cz: add pcp_ prefix to the spin_lock_irqsave wrappers, per Steven]
Link: https://lkml.kernel.org/r/20260105-fix-pcp-up-v1-1-5579662d2071@suse.cz
Fixes: 5749077415 ("mm/page_alloc: leave IRQs enabled for per-cpu page allocations")
Signed-off-by: Vlastimil Babka <vbabka@suse.cz>
Reported-by: kernel test robot <oliver.sang@intel.com>
Closes: https://lore.kernel.org/oe-lkp/202512101320.e2f2dd6f-lkp@intel.com
Analyzed-by: Matthew Wilcox <willy@infradead.org>
Link: https://lore.kernel.org/all/aUW05pyc9nZkvY-1@casper.infradead.org/
Acked-by: Mel Gorman <mgorman@techsingularity.net>
Cc: Brendan Jackman <jackmanb@google.com>
Cc: Johannes Weiner <hannes@cmpxchg.org>
Cc: Michal Hocko <mhocko@suse.com>
Cc: Sebastian Andrzej Siewior <bigeasy@linutronix.de>
Cc: Steven Rostedt <rostedt@goodmis.org>
Cc: Suren Baghdasaryan <surenb@google.com>
Cc: Zi Yan <ziy@nvidia.com>
Cc: <stable@vger.kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit fc4b909c36 upstream.
It is possible for pcp->count - pcp->high to exceed pcp->batch by a lot.
When this happens, we should perform batching to ensure that
free_pcppages_bulk isn't called with too many pages to free at once and
starve out other threads that need the pcp or zone lock.
Since we are still only freeing the difference between the initial
pcp->count and pcp->high values, there should be no change to how many
pages are freed.
Link: https://lkml.kernel.org/r/20251014145011.3427205-3-joshua.hahnjy@gmail.com
Signed-off-by: Joshua Hahn <joshua.hahnjy@gmail.com>
Suggested-by: Chris Mason <clm@fb.com>
Suggested-by: Andrew Morton <akpm@linux-foundation.org>
Co-developed-by: Johannes Weiner <hannes@cmpxchg.org>
Reviewed-by: Vlastimil Babka <vbabka@suse.cz>
Cc: Brendan Jackman <jackmanb@google.com>
Cc: "Kirill A. Shutemov" <kirill@shutemov.name>
Cc: Michal Hocko <mhocko@suse.com>
Cc: SeongJae Park <sj@kernel.org>
Cc: Suren Baghdasaryan <surenb@google.com>
Cc: Zi Yan <ziy@nvidia.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Stable-dep-of: 038a102535 ("mm/page_alloc: prevent pcp corruption with SMP=n")
Signed-off-by: Sasha Levin <sashal@kernel.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit 0acc67c403 upstream.
Patch series "mm/page_alloc: Batch callers of free_pcppages_bulk", v5.
Motivation & Approach
=====================
While testing workloads with high sustained memory pressure on large
machines in the Meta fleet (1Tb memory, 316 CPUs), we saw an unexpectedly
high number of softlockups. Further investigation showed that the zone
lock in free_pcppages_bulk was being held for a long time, and was called
to free 2k+ pages over 100 times just during boot.
This causes starvation in other processes for the zone lock, which can
lead to the system stalling as multiple threads cannot make progress
without the locks. We can see these issues manifesting as warnings:
[ 4512.591979] rcu: INFO: rcu_sched self-detected stall on CPU
[ 4512.604370] rcu: 20-....: (9312 ticks this GP) idle=a654/1/0x4000000000000000 softirq=309340/309344 fqs=5426
[ 4512.626401] rcu: hardirqs softirqs csw/system
[ 4512.638793] rcu: number: 0 145 0
[ 4512.651177] rcu: cputime: 30 10410 174 ==> 10558(ms)
[ 4512.666657] rcu: (t=21077 jiffies g=783665 q=1242213 ncpus=316)
While these warnings don't indicate a crash or a kernel panic, they do
point to the underlying issue of lock contention. To prevent starvation
in both locks, batch the freeing of pages using pcp->batch.
Because free_pcppages_bulk is called with the pcp lock and acquires the
zone lock, relinquishing and reacquiring the locks are only effective when
both of them are broken together (unless the system was built with queued
spinlocks). Thus, instead of modifying free_pcppages_bulk to break both
locks, batch the freeing from its callers instead.
A similar fix has been implemented in the Meta fleet, and we have seen
significantly less softlockups.
Testing
=======
The following are a few synthetic benchmarks, made on three machines. The
first is a large machine with 754GiB memory and 316 processors.
The second is a relatively smaller machine with 251GiB memory and 176
processors. The third and final is the smallest of the three, which has 62GiB
memory and 36 processors.
On all machines, I kick off a kernel build with -j$(nproc).
Negative delta is better (faster compilation).
Large machine (754GiB memory, 316 processors)
make -j$(nproc)
+------------+---------------+-----------+
| Metric (s) | Variation (%) | Delta(%) |
+------------+---------------+-----------+
| real | 0.8070 | - 1.4865 |
| user | 0.2823 | + 0.4081 |
| sys | 5.0267 | -11.8737 |
+------------+---------------+-----------+
Medium machine (251GiB memory, 176 processors)
make -j$(nproc)
+------------+---------------+----------+
| Metric (s) | Variation (%) | Delta(%) |
+------------+---------------+----------+
| real | 0.2806 | +0.0351 |
| user | 0.0994 | +0.3170 |
| sys | 0.6229 | -0.6277 |
+------------+---------------+----------+
Small machine (62GiB memory, 36 processors)
make -j$(nproc)
+------------+---------------+----------+
| Metric (s) | Variation (%) | Delta(%) |
+------------+---------------+----------+
| real | 0.1503 | -2.6585 |
| user | 0.0431 | -2.2984 |
| sys | 0.1870 | -3.2013 |
+------------+---------------+----------+
Here, variation is the coefficient of variation, i.e. standard deviation
/ mean.
Based on these results, it seems like there are varying degrees to how
much lock contention this reduces. For the largest and smallest machines
that I ran the tests on, it seems like there is quite some significant
reduction. There is also some performance increases visible from
userspace.
Interestingly, the performance gains don't scale with the size of the
machine, but rather there seems to be a dip in the gain there is for the
medium-sized machine. One possible theory is that because the high
watermark depends on both memory and the number of local CPUs, what
impacts zone contention the most is not these individual values, but
rather the ratio of mem:processors.
This patch (of 5):
Currently, refresh_cpu_vm_stats returns an int, indicating how many
changes were made during its updates. Using this information, callers
like vmstat_update can heuristically determine if more work will be done
in the future.
However, all of refresh_cpu_vm_stats's callers either (a) ignore the
result, only caring about performing the updates, or (b) only care about
whether changes were made, but not *how many* changes were made.
Simplify the code by returning a bool instead to indicate if updates
were made.
In addition, simplify fold_diff and decay_pcp_high to return a bool
for the same reason.
Link: https://lkml.kernel.org/r/20251014145011.3427205-1-joshua.hahnjy@gmail.com
Link: https://lkml.kernel.org/r/20251014145011.3427205-2-joshua.hahnjy@gmail.com
Signed-off-by: Joshua Hahn <joshua.hahnjy@gmail.com>
Reviewed-by: Vlastimil Babka <vbabka@suse.cz>
Reviewed-by: SeongJae Park <sj@kernel.org>
Cc: Brendan Jackman <jackmanb@google.com>
Cc: Chris Mason <clm@fb.com>
Cc: Johannes Weiner <hannes@cmpxchg.org>
Cc: "Kirill A. Shutemov" <kirill@shutemov.name>
Cc: Michal Hocko <mhocko@suse.com>
Cc: Suren Baghdasaryan <surenb@google.com>
Cc: Zi Yan <ziy@nvidia.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Stable-dep-of: 038a102535 ("mm/page_alloc: prevent pcp corruption with SMP=n")
Signed-off-by: Sasha Levin <sashal@kernel.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit 5037b34282 upstream.
When wait_current_trans() is called during start_transaction(), it
currently waits for a blocked transaction without considering whether
the given transaction type actually needs to wait for that particular
transaction state. The btrfs_blocked_trans_types[] array already defines
which transaction types should wait for which transaction states, but
this check was missing in wait_current_trans().
This can lead to a deadlock scenario involving two transactions and
pending ordered extents:
1. Transaction A is in TRANS_STATE_COMMIT_DOING state
2. A worker processing an ordered extent calls start_transaction()
with TRANS_JOIN
3. join_transaction() returns -EBUSY because Transaction A is in
TRANS_STATE_COMMIT_DOING
4. Transaction A moves to TRANS_STATE_UNBLOCKED and completes
5. A new Transaction B is created (TRANS_STATE_RUNNING)
6. The ordered extent from step 2 is added to Transaction B's
pending ordered extents
7. Transaction B immediately starts commit by another task and
enters TRANS_STATE_COMMIT_START
8. The worker finally reaches wait_current_trans(), sees Transaction B
in TRANS_STATE_COMMIT_START (a blocked state), and waits
unconditionally
9. However, TRANS_JOIN should NOT wait for TRANS_STATE_COMMIT_START
according to btrfs_blocked_trans_types[]
10. Transaction B is waiting for pending ordered extents to complete
11. Deadlock: Transaction B waits for ordered extent, ordered extent
waits for Transaction B
This can be illustrated by the following call stacks:
CPU0 CPU1
btrfs_finish_ordered_io()
start_transaction(TRANS_JOIN)
join_transaction()
# -EBUSY (Transaction A is
# TRANS_STATE_COMMIT_DOING)
# Transaction A completes
# Transaction B created
# ordered extent added to
# Transaction B's pending list
btrfs_commit_transaction()
# Transaction B enters
# TRANS_STATE_COMMIT_START
# waiting for pending ordered
# extents
wait_current_trans()
# waits for Transaction B
# (should not wait!)
Task bstore_kv_sync in btrfs_commit_transaction waiting for ordered
extents:
__schedule+0x2e7/0x8a0
schedule+0x64/0xe0
btrfs_commit_transaction+0xbf7/0xda0 [btrfs]
btrfs_sync_file+0x342/0x4d0 [btrfs]
__x64_sys_fdatasync+0x4b/0x80
do_syscall_64+0x33/0x40
entry_SYSCALL_64_after_hwframe+0x44/0xa9
Task kworker in wait_current_trans waiting for transaction commit:
Workqueue: btrfs-syno_nocow btrfs_work_helper [btrfs]
__schedule+0x2e7/0x8a0
schedule+0x64/0xe0
wait_current_trans+0xb0/0x110 [btrfs]
start_transaction+0x346/0x5b0 [btrfs]
btrfs_finish_ordered_io.isra.0+0x49b/0x9c0 [btrfs]
btrfs_work_helper+0xe8/0x350 [btrfs]
process_one_work+0x1d3/0x3c0
worker_thread+0x4d/0x3e0
kthread+0x12d/0x150
ret_from_fork+0x1f/0x30
Fix this by passing the transaction type to wait_current_trans() and
checking btrfs_blocked_trans_types[cur_trans->state] against the given
type before deciding to wait. This ensures that transaction types which
are allowed to join during certain blocked states will not unnecessarily
wait and cause deadlocks.
Reviewed-by: Filipe Manana <fdmanana@suse.com>
Signed-off-by: Robbie Ko <robbieko@synology.com>
Signed-off-by: Filipe Manana <fdmanana@suse.com>
Reviewed-by: David Sterba <dsterba@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
Cc: Motiejus Jakštys <motiejus@jakstys.lt>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit 3644f44117 upstream.
Clang warns (or errors with CONFIG_WERROR=y / W=e):
drivers/hid/intel-ish-hid/ipc/ipc.c:935:36: error: cast from 'void (*)(struct workqueue_struct *)' to 'void (*)(void *)' converts to incompatible function type [-Werror,-Wcast-function-type-strict]
935 | if (devm_add_action_or_reset(dev, (void (*)(void *))destroy_workqueue,
| ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
include/linux/device/devres.h:168:34: note: expanded from macro 'devm_add_action_or_reset'
168 | __devm_add_action_or_ireset(dev, action, data, #action)
| ^~~~~~
This warning is pointing out a kernel control flow integrity (kCFI /
CONFIG_CFI=y) violation will occur due to this function cast when the
destroy_workqueue() is indirectly called via devm_action_release()
because the prototype of destroy_workqueue() does not match the
prototype of (*action)().
Use a local function with the correct prototype to wrap
destroy_workqueue() to resolve the warning and CFI violation.
Reported-by: kernel test robot <lkp@intel.com>
Closes: https://lore.kernel.org/oe-kbuild-all/202510190103.qTZvfdjj-lkp@intel.com/
Closes: https://github.com/ClangBuiltLinux/linux/issues/2139
Fixes: 0d30dae38f ("HID: intel-ish-hid: Use dedicated unbound workqueues to prevent resume blocking")
Signed-off-by: Nathan Chancellor <nathan@kernel.org>
Acked-by: Srinivas Pandruvada <srinivas.pandruvada@linux.intel.com>
Reviewed-by: Zhang Lixu <lixu.zhang@intel.com>
Signed-off-by: Jiri Kosina <jkosina@suse.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit 0d30dae38f upstream.
During suspend/resume tests with S2IDLE, some ISH functional failures were
observed because of delay in executing ISH resume handler. Here
schedule_work() is used from resume handler to do actual work.
schedule_work() uses system_wq, which is a per CPU work queue. Although
the queuing is not bound to a CPU, but it prefers local CPU of the caller,
unless prohibited.
Users of this work queue are not supposed to queue long running work.
But in practice, there are scenarios where long running work items are
queued on other unbound workqueues, occupying the CPU. As a result, the
ISH resume handler may not get a chance to execute in a timely manner.
In one scenario, one of the ish_resume_handler() executions was delayed
nearly 1 second because another work item on an unbound workqueue occupied
the same CPU. This delay causes ISH functionality failures.
A similar issue was previously observed where the ISH HID driver timed out
while getting the HID descriptor during S4 resume in the recovery kernel,
likely caused by the same workqueue contention problem.
Create dedicated unbound workqueues for all ISH operations to allow work
items to execute on any available CPU, eliminating CPU-specific bottlenecks
and improving resume reliability under varying system loads. Also ISH has
three different components, a bus driver which implements ISH protocols, a
PCI interface layer and HID interface. Use one dedicated work queue for all
of them.
Signed-off-by: Zhang Lixu <lixu.zhang@intel.com>
Signed-off-by: Jiri Kosina <jkosina@suse.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit e37d5a2d60 upstream.
Introduce a new IOMMU interface to flush IOTLB paging cache entries for
the CPU kernel address space. This interface is invoked from the x86
architecture code that manages combined user and kernel page tables,
specifically before any kernel page table page is freed and reused.
This addresses the main issue with vfree() which is a common occurrence
and can be triggered by unprivileged users. While this resolves the
primary problem, it doesn't address some extremely rare case related to
memory unplug of memory that was present as reserved memory at boot, which
cannot be triggered by unprivileged users. The discussion can be found at
the link below.
Enable SVA on x86 architecture since the IOMMU can now receive
notification to flush the paging cache before freeing the CPU kernel page
table pages.
Link: https://lkml.kernel.org/r/20251022082635.2462433-9-baolu.lu@linux.intel.com
Link: https://lore.kernel.org/linux-iommu/04983c62-3b1d-40d4-93ae-34ca04b827e5@intel.com/
Co-developed-by: Jason Gunthorpe <jgg@nvidia.com>
Signed-off-by: Jason Gunthorpe <jgg@nvidia.com>
Signed-off-by: Lu Baolu <baolu.lu@linux.intel.com>
Suggested-by: Jann Horn <jannh@google.com>
Reviewed-by: Jason Gunthorpe <jgg@nvidia.com>
Reviewed-by: Vasant Hegde <vasant.hegde@amd.com>
Reviewed-by: Kevin Tian <kevin.tian@intel.com>
Cc: Alistair Popple <apopple@nvidia.com>
Cc: Andy Lutomirski <luto@kernel.org>
Cc: Borislav Betkov <bp@alien8.de>
Cc: Dave Hansen <dave.hansen@linux.intel.com>
Cc: David Hildenbrand <david@redhat.com>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: Jean-Philippe Brucker <jean-philippe@linaro.org>
Cc: Joerg Roedel <joro@8bytes.org>
Cc: Liam Howlett <liam.howlett@oracle.com>
Cc: Lorenzo Stoakes <lorenzo.stoakes@oracle.com>
Cc: Matthew Wilcox (Oracle) <willy@infradead.org>
Cc: Michal Hocko <mhocko@kernel.org>
Cc: Mike Rapoport (Microsoft) <rppt@kernel.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Robin Murohy <robin.murphy@arm.com>
Cc: Thomas Gleinxer <tglx@linutronix.de>
Cc: "Uladzislau Rezki (Sony)" <urezki@gmail.com>
Cc: Vinicius Costa Gomes <vinicius.gomes@intel.com>
Cc: Vlastimil Babka <vbabka@suse.cz>
Cc: Will Deacon <will@kernel.org>
Cc: Yi Lai <yi1.lai@intel.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit 5ba2f0a155 upstream.
This introduces a conditional asynchronous mechanism, enabled by
CONFIG_ASYNC_KERNEL_PGTABLE_FREE. When enabled, this mechanism defers the
freeing of pages that are used as page tables for kernel address mappings.
These pages are now queued to a work struct instead of being freed
immediately.
This deferred freeing allows for batch-freeing of page tables, providing a
safe context for performing a single expensive operation (TLB flush) for a
batch of kernel page tables instead of performing that expensive operation
for each page table.
Link: https://lkml.kernel.org/r/20251022082635.2462433-8-baolu.lu@linux.intel.com
Signed-off-by: Dave Hansen <dave.hansen@linux.intel.com>
Signed-off-by: Lu Baolu <baolu.lu@linux.intel.com>
Reviewed-by: Jason Gunthorpe <jgg@nvidia.com>
Reviewed-by: Kevin Tian <kevin.tian@intel.com>
Acked-by: David Hildenbrand <david@redhat.com>
Acked-by: Mike Rapoport (Microsoft) <rppt@kernel.org>
Cc: Alistair Popple <apopple@nvidia.com>
Cc: Andy Lutomirski <luto@kernel.org>
Cc: Borislav Betkov <bp@alien8.de>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: Jann Horn <jannh@google.com>
Cc: Jean-Philippe Brucker <jean-philippe@linaro.org>
Cc: Joerg Roedel <joro@8bytes.org>
Cc: Liam Howlett <liam.howlett@oracle.com>
Cc: Lorenzo Stoakes <lorenzo.stoakes@oracle.com>
Cc: Matthew Wilcox (Oracle) <willy@infradead.org>
Cc: Michal Hocko <mhocko@kernel.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Robin Murohy <robin.murphy@arm.com>
Cc: Thomas Gleinxer <tglx@linutronix.de>
Cc: "Uladzislau Rezki (Sony)" <urezki@gmail.com>
Cc: Vasant Hegde <vasant.hegde@amd.com>
Cc: Vinicius Costa Gomes <vinicius.gomes@intel.com>
Cc: Vlastimil Babka <vbabka@suse.cz>
Cc: Will Deacon <will@kernel.org>
Cc: Yi Lai <yi1.lai@intel.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
