When CONFIG_IPV6 is disabled, the udp_sock_create6() function returns 0
(success) without actually creating a socket. Callers such as
fou_create() then proceed to dereference the uninitialized socket
pointer, resulting in a NULL pointer dereference.
This patch makes udp_sock_create6 return -EPFNOSUPPORT instead, so
callers correctly take their error paths. There is only one caller of
the vulnerable function and only privileged users can trigger it.
Fixes: fd384412e199b ("udp_tunnel: Seperate ipv6 functions into its own file.") Reported-by: Weiming Shi <bestswngs@gmail.com> Signed-off-by: Xiang Mei <xmei5@asu.edu> Link: https://patch.msgid.link/20260317010241.1893893-1-xmei5@asu.edu Signed-off-by: Jakub Kicinski <kuba@kernel.org> Signed-off-by: Sasha Levin <sashal@kernel.org>
In IPSec full offload mode, the device reports an ESN (Extended
Sequence Number) wrap event to the driver. The driver validates this
event by querying the IPSec ASO and checking that the esn_event_arm
field is 0x0, which indicates an event has occurred. After handling
the event, the driver must re-arm the context by setting esn_event_arm
back to 0x1.
A race condition exists in this handling path. After validating the
event, the driver calls mlx5_accel_esp_modify_xfrm() to update the
kernel's xfrm state. This function temporarily releases and
re-acquires the xfrm state lock.
So, need to acknowledge the event first by setting esn_event_arm to
0x1. This prevents the driver from reprocessing the same ESN update if
the hardware sends events for other reason. Since the next ESN update
only occurs after nearly 2^31 packets are received, there's no risk of
missing an update, as it will happen long after this handling has
finished.
Processing the event twice causes the ESN high-order bits (esn_msb) to
be incremented incorrectly. The driver then programs the hardware with
this invalid ESN state, which leads to anti-replay failures and a
complete halt of IPSec traffic.
Fix this by re-arming the ESN event immediately after it is validated,
before calling mlx5_accel_esp_modify_xfrm(). This ensures that any
spurious, duplicate events are correctly ignored, closing the race
window.
Fixes: fef06678931f ("net/mlx5e: Fix ESN update kernel panic") Signed-off-by: Jianbo Liu <jianbol@nvidia.com> Reviewed-by: Leon Romanovsky <leonro@nvidia.com> Signed-off-by: Tariq Toukan <tariqt@nvidia.com> Link: https://patch.msgid.link/20260316094603.6999-4-tariqt@nvidia.com Signed-off-by: Jakub Kicinski <kuba@kernel.org> Signed-off-by: Sasha Levin <sashal@kernel.org>
The query or updating IPSec offload object is through Access ASO WQE.
The driver uses a single mlx5e_ipsec_aso struct for each PF, which
contains a shared DMA-mapped context for all ASO operations.
A race condition exists because the ASO spinlock is released before
the hardware has finished processing WQE. If a second operation is
initiated immediately after, it overwrites the shared context in the
DMA area.
When the first operation's completion is processed later, it reads
this corrupted context, leading to unexpected behavior and incorrect
results.
This commit fixes the race by introducing a private context within
each IPSec offload object. The shared ASO context is now copied to
this private context while the ASO spinlock is held. Subsequent
processing uses this saved, per-object context, ensuring its integrity
is maintained.
Fixes: 1ed78fc03307 ("net/mlx5e: Update IPsec soft and hard limits") Signed-off-by: Jianbo Liu <jianbol@nvidia.com> Reviewed-by: Leon Romanovsky <leonro@nvidia.com> Signed-off-by: Tariq Toukan <tariqt@nvidia.com> Link: https://patch.msgid.link/20260316094603.6999-3-tariqt@nvidia.com Signed-off-by: Jakub Kicinski <kuba@kernel.org> Signed-off-by: Sasha Levin <sashal@kernel.org>
Fix that by restricting the rtnl-protected section to just the necessary
part, the call to netdev_master_upper_dev_get and speed querying, so
that the last lock dependency is avoided and the cycle doesn't close.
This is safe because mlx5_uplink_netdev_get uses netdev_hold to keep the
uplink netdev alive while its master device is queried.
Use this opportunity to rename the ambiguously-named "hold_rtnl_lock"
argument to "take_rtnl" and remove the "_locked" suffix from
mlx5_esw_qos_lag_link_speed_get_locked.
If hardware doesn't support RX Flow Filters, rx_fs_lock spinlock is not
initialized leading to the following assertion splat triggerable via
set_rxnfc callback.
INFO: trying to register non-static key.
The code is fine but needs lockdep annotation, or maybe
you didn't initialize this object before use?
turning off the locking correctness validator.
CPU: 1 PID: 949 Comm: syz.0.6 Not tainted 6.1.164+ #113
Hardware name: QEMU Standard PC (Q35 + ICH9, 2009), BIOS rel-1.16.1-0-g3208b098f51a-prebuilt.qemu.org 04/01/2014
Call Trace:
<TASK>
__dump_stack lib/dump_stack.c:88 [inline]
dump_stack_lvl+0x8d/0xba lib/dump_stack.c:106
assign_lock_key kernel/locking/lockdep.c:974 [inline]
register_lock_class+0x141b/0x17f0 kernel/locking/lockdep.c:1287
__lock_acquire+0x74f/0x6c40 kernel/locking/lockdep.c:4928
lock_acquire kernel/locking/lockdep.c:5662 [inline]
lock_acquire+0x190/0x4b0 kernel/locking/lockdep.c:5627
__raw_spin_lock_irqsave include/linux/spinlock_api_smp.h:110 [inline]
_raw_spin_lock_irqsave+0x33/0x50 kernel/locking/spinlock.c:162
gem_del_flow_filter drivers/net/ethernet/cadence/macb_main.c:3562 [inline]
gem_set_rxnfc+0x533/0xac0 drivers/net/ethernet/cadence/macb_main.c:3667
ethtool_set_rxnfc+0x18c/0x280 net/ethtool/ioctl.c:961
__dev_ethtool net/ethtool/ioctl.c:2956 [inline]
dev_ethtool+0x229c/0x6290 net/ethtool/ioctl.c:3095
dev_ioctl+0x637/0x1070 net/core/dev_ioctl.c:510
sock_do_ioctl+0x20d/0x2c0 net/socket.c:1215
sock_ioctl+0x577/0x6d0 net/socket.c:1320
vfs_ioctl fs/ioctl.c:51 [inline]
__do_sys_ioctl fs/ioctl.c:870 [inline]
__se_sys_ioctl fs/ioctl.c:856 [inline]
__x64_sys_ioctl+0x18c/0x210 fs/ioctl.c:856
do_syscall_x64 arch/x86/entry/common.c:46 [inline]
do_syscall_64+0x35/0x80 arch/x86/entry/common.c:76
entry_SYSCALL_64_after_hwframe+0x6e/0xd8
A more straightforward solution would be to always initialize rx_fs_lock,
just like rx_fs_list. However, in this case the driver set_rxnfc callback
would return with a rather confusing error code, e.g. -EINVAL. So deny
set_rxnfc attempts directly if the RX filtering feature is not supported
by hardware.
Fixes: ae8223de3df5 ("net: macb: Added support for RX filtering") Signed-off-by: Fedor Pchelkin <pchelkin@ispras.ru> Link: https://patch.msgid.link/20260316103826.74506-2-pchelkin@ispras.ru Signed-off-by: Jakub Kicinski <kuba@kernel.org> Signed-off-by: Sasha Levin <sashal@kernel.org>
After commi f132e089fe89 ("ACPI: processor: Fix NULL-pointer dereference
in acpi_processor_errata_piix4()"), device pointers may be dereferenced
after dropping references to the device objects pointed to by them,
which may cause a use-after-free to occur.
Moreover, debug messages about enabling the errata may be printed
if the errata flags corresponding to them are unset.
Address all of these issues by moving message printing to the points
in the code where the errata flags are set.
Fixes: f132e089fe89 ("ACPI: processor: Fix NULL-pointer dereference in acpi_processor_errata_piix4()") Reported-by: Guenter Roeck <linux@roeck-us.net> Closes: https://lore.kernel.org/linux-acpi/938e2206-def5-4b7a-9b2c-d1fd37681d8a@roeck-us.net/ Reviewed-by: Guenter Roeck <linux@roeck-us.net> Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com> Link: https://patch.msgid.link/5975693.DvuYhMxLoT@rafael.j.wysocki Signed-off-by: Sasha Levin <sashal@kernel.org>
Since upstream commit e75665dd0968 ("wifi: wlcore: ensure skb headroom
before skb_push"), wl1271_tx_allocate() and with it
wl1271_prepare_tx_frame() returns -EAGAIN if pskb_expand_head() fails.
However, in wlcore_tx_work_locked(), a return value of -EAGAIN from
wl1271_prepare_tx_frame() is interpreted as the aggregation buffer being
full. This causes the code to flush the buffer, put the skb back at the
head of the queue, and immediately retry the same skb in a tight while
loop.
Because wlcore_tx_work_locked() holds wl->mutex, and the retry happens
immediately with GFP_ATOMIC, this will result in an infinite loop and a
CPU soft lockup. Return -ENOMEM instead so the packet is dropped and
the loop terminates.
The problem was found by an experimental code review agent based on
gemini-3.1-pro while reviewing backports into v6.18.y.
Assisted-by: Gemini:gemini-3.1-pro Fixes: e75665dd0968 ("wifi: wlcore: ensure skb headroom before skb_push") Cc: Peter Astrand <astrand@lysator.liu.se> Signed-off-by: Guenter Roeck <linux@roeck-us.net> Link: https://patch.msgid.link/20260318064636.3065925-1-linux@roeck-us.net Signed-off-by: Johannes Berg <johannes.berg@intel.com> Signed-off-by: Sasha Levin <sashal@kernel.org>
mesh_matches_local() unconditionally dereferences ie->mesh_config to
compare mesh configuration parameters. When called from
mesh_rx_csa_frame(), the parsed action-frame elements may not contain a
Mesh Configuration IE, leaving ie->mesh_config NULL and triggering a
kernel NULL pointer dereference.
The other two callers are already safe:
- ieee80211_mesh_rx_bcn_presp() checks !elems->mesh_config before
calling mesh_matches_local()
- mesh_plink_get_event() is only reached through
mesh_process_plink_frame(), which checks !elems->mesh_config, too
mesh_rx_csa_frame() is the only caller that passes raw parsed elements
to mesh_matches_local() without guarding mesh_config. An adjacent
attacker can exploit this by sending a crafted CSA action frame that
includes a valid Mesh ID IE but omits the Mesh Configuration IE,
crashing the kernel.
This patch adds a NULL check for ie->mesh_config at the top of
mesh_matches_local() to return false early when the Mesh Configuration
IE is absent.
Fixes: 2e3c8736820b ("mac80211: support functions for mesh") Reported-by: Weiming Shi <bestswngs@gmail.com> Signed-off-by: Xiang Mei <xmei5@asu.edu> Link: https://patch.msgid.link/20260318034244.2595020-1-xmei5@asu.edu Signed-off-by: Johannes Berg <johannes.berg@intel.com> Signed-off-by: Sasha Levin <sashal@kernel.org>
When iavf_add_vlan() finds an existing filter in IAVF_VLAN_REMOVE
state, it transitions the filter to IAVF_VLAN_ACTIVE assuming the
pending delete can simply be cancelled. However, there is no guarantee
that iavf_del_vlans() has not already processed the delete AQ request
and removed the filter from the PF. In that case the filter remains in
the driver's list as IAVF_VLAN_ACTIVE but is no longer programmed on
the NIC. Since iavf_add_vlans() only picks up filters in
IAVF_VLAN_ADD state, the filter is never re-added, and spoof checking
drops all traffic for that VLAN.
CPU0 CPU1 Workqueue
---- ---- ---------
iavf_del_vlan(vlan 100)
f->state = REMOVE
schedule AQ_DEL_VLAN
iavf_add_vlan(vlan 100)
f->state = ACTIVE
iavf_del_vlans()
f is ACTIVE, skip
iavf_add_vlans()
f is ACTIVE, skip
Filter is ACTIVE in driver but absent from NIC.
Transition to IAVF_VLAN_ADD instead and schedule
IAVF_FLAG_AQ_ADD_VLAN_FILTER so iavf_add_vlans() re-programs the
filter. A duplicate add is idempotent on the PF.
If an XDP application that requested TX timestamping is shutting down
while the link of the interface in use is still up the following kernel
splat is reported:
igc_xmit_frame() misses updating skb->tail when the packet size is
shorter than the minimum one.
Use skb_put_padto() in alignment with other Intel Ethernet drivers.
Fixes: 0507ef8a0372 ("igc: Add transmit and receive fastpath and interrupt handlers") Signed-off-by: Kohei Enju <kohei@enjuk.jp> Reviewed-by: Simon Horman <horms@kernel.org> Reviewed-by: Paul Menzel <pmenzel@molgen.mpg.de> 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>
This is caused by aqc111_suspend calling
the PM variant of its write_cmd routine.
The simplified call trace looks like this:
rpm_suspend()
usb_suspend_both() - here udev->dev.power.runtime_status == RPM_SUSPENDING
aqc111_suspend() - called for the usb device interface
aqc111_write32_cmd()
usb_autopm_get_interface()
pm_runtime_resume_and_get()
rpm_resume() - here we call rpm_resume() on our parent
rpm_resume() - Here we wait for a status change that will never happen.
At this point we block another task which holds
rtnl_lock and locks up the whole networking stack.
Fix this by replacing the write_cmd calls with their _nopm variants
Reported-by: syzbot+48dc1e8dfc92faf1124c@syzkaller.appspotmail.com Closes: https://syzkaller.appspot.com/bug?extid=48dc1e8dfc92faf1124c Fixes: e58ba4544c77 ("net: usb: aqc111: Add support for wake on LAN by MAGIC packet") Signed-off-by: Nikola Z. Ivanov <zlatistiv@gmail.com> Link: https://patch.msgid.link/20260313141643.1181386-1-zlatistiv@gmail.com Signed-off-by: Paolo Abeni <pabeni@redhat.com> Signed-off-by: Sasha Levin <sashal@kernel.org>
Fix a use-after-free in the clsact qdisc upon init/destroy rollback asymmetry.
The latter is achieved by first fully initializing a clsact instance, and
then in a second step having a replacement failure for the new clsact qdisc
instance. clsact_init() initializes ingress first and then takes care of the
egress part. This can fail midway, for example, via tcf_block_get_ext(). Upon
failure, the kernel will trigger the clsact_destroy() callback.
Commit 1cb6f0bae504 ("bpf: Fix too early release of tcx_entry") details the
way how the transition is happening. If tcf_block_get_ext on the q->ingress_block
ends up failing, we took the tcx_miniq_inc reference count on the ingress
side, but not yet on the egress side. clsact_destroy() tests whether the
{ingress,egress}_entry was non-NULL. However, even in midway failure on the
replacement, both are in fact non-NULL with a valid egress_entry from the
previous clsact instance.
What we really need to test for is whether the qdisc instance-specific ingress
or egress side previously got initialized. This adds a small helper for checking
the miniq initialization called mini_qdisc_pair_inited, and utilizes that upon
clsact_destroy() in order to fix the use-after-free scenario. Convert the
ingress_destroy() side as well so both are consistent to each other.
Fixes: 1cb6f0bae504 ("bpf: Fix too early release of tcx_entry") Reported-by: Keenan Dong <keenanat2000@gmail.com> Signed-off-by: Daniel Borkmann <daniel@iogearbox.net> Cc: Martin KaFai Lau <martin.lau@kernel.org> Acked-by: Martin KaFai Lau <martin.lau@kernel.org> Link: https://patch.msgid.link/20260313065531.98639-1-daniel@iogearbox.net Signed-off-by: Paolo Abeni <pabeni@redhat.com> Signed-off-by: Sasha Levin <sashal@kernel.org>
The same bounds-check bug fixed for NDP16 in the previous patch also
exists in cdc_ncm_rx_verify_ndp32(). The DPE array size is validated
against the total skb length without accounting for ndpoffset, allowing
out-of-bounds reads when the NDP32 is placed near the end of the NTB.
Add ndpoffset to the nframes bounds check and use struct_size_t() to
express the NDP-plus-DPE-array size more clearly.
Compile-tested only.
Fixes: 0fa81b304a79 ("cdc_ncm: Implement the 32-bit version of NCM Transfer Block") Signed-off-by: Tobi Gaertner <tob.gaertner@me.com> Link: https://patch.msgid.link/20260314054640.2895026-3-tob.gaertner@me.com Signed-off-by: Jakub Kicinski <kuba@kernel.org> Signed-off-by: Sasha Levin <sashal@kernel.org>
cdc_ncm_rx_verify_ndp16() validates that the NDP header and its DPE
entries fit within the skb. The first check correctly accounts for
ndpoffset:
if ((ndpoffset + sizeof(struct usb_cdc_ncm_ndp16)) > skb_in->len)
but the second check omits it:
if ((sizeof(struct usb_cdc_ncm_ndp16) +
ret * (sizeof(struct usb_cdc_ncm_dpe16))) > skb_in->len)
This validates the DPE array size against the total skb length as if
the NDP were at offset 0, rather than at ndpoffset. When the NDP is
placed near the end of the NTB (large wNdpIndex), the DPE entries can
extend past the skb data buffer even though the check passes.
cdc_ncm_rx_fixup() then reads out-of-bounds memory when iterating
the DPE array.
Add ndpoffset to the nframes bounds check and use struct_size_t() to
express the NDP-plus-DPE-array size more clearly.
Do not run airoha_dev_stop routine explicitly in airoha_remove()
since ndo_stop() callback is already executed by unregister_netdev() in
__dev_close_many routine if necessary and, doing so, we will end up causing
an underflow in the qdma users atomic counters. Rely on networking subsystem
to stop the device removing the airoha_eth module.
In order to avoid any possible race, read completion queue head and
pending entry in airoha_qdma_tx_napi_poll routine instead of doing it in
airoha_irq_handler. Remove unused airoha_tx_irq_queue unused fields.
This is a preliminary patch to add Qdisc offload for airoha_eth driver.
Align PSE memory configuration to vendor SDK. In particular, increase
initial value of PSE reserved memory in airoha_fe_pse_ports_init()
routine by the value used for the second Packet Processor Engine (PPE2)
and do not overwrite the default value.
Introduced by commit 23020f049327 ("net: airoha: Introduce ethernet support
for EN7581 SoC")
Store the default value for the number of PSE reserved pages in orig_val
at the beginning of airoha_fe_set_pse_oq_rsv routine, before updating it
with airoha_fe_set_pse_queue_rsv_pages().
Introduce airoha_fe_get_pse_all_rsv utility routine.
Introduced by commit 23020f049327 ("net: airoha: Introduce ethernet support
for EN7581 SoC")
Whenever a TEQL devices has a lockless Qdisc as root, qdisc_reset should
be called using the seq_lock to avoid racing with the datapath. Failure
to do so may cause crashes like the following:
Workflow to reproduce:
1. Initialize a TEQL topology (dummy0 and ifb0 as slaves, teql0 up).
2. Start multiple sender workers continuously transmitting packets
through teql0 to drive teql_master_xmit().
3. In parallel, repeatedly delete and re-add the root qdisc on
dummy0 and ifb0 via RTNETLINK, forcing frequent teardown and reset activity
(teql_destroy() / qdisc_reset()).
4. After running both workloads concurrently for several iterations,
KASAN reports slab-use-after-free or double-free in the skb free path.
Fix this by moving dev_reset_queue to sch_generic.h and calling it, instead
of qdisc_reset, in teql_destroy since it handles both the lock and lockless
cases correctly for root qdiscs.
Fixes: 96009c7d500e ("sched: replace __QDISC_STATE_RUNNING bit with a spin lock") Reported-by: Xianrui Dong <keenanat2000@gmail.com> Tested-by: Xianrui Dong <keenanat2000@gmail.com> Co-developed-by: Victor Nogueira <victor@mojatatu.com> Signed-off-by: Victor Nogueira <victor@mojatatu.com> Signed-off-by: Jamal Hadi Salim <jhs@mojatatu.com> Link: https://patch.msgid.link/20260315155422.147256-1-jhs@mojatatu.com Signed-off-by: Jakub Kicinski <kuba@kernel.org> Signed-off-by: Sasha Levin <sashal@kernel.org>
Syzkaller reported a panic in smc_tcp_syn_recv_sock() [1].
smc_tcp_syn_recv_sock() is called in the TCP receive path
(softirq) via icsk_af_ops->syn_recv_sock on the clcsock (TCP
listening socket). It reads sk_user_data to get the smc_sock
pointer. However, when the SMC listen socket is being closed
concurrently, smc_close_active() sets clcsock->sk_user_data
to NULL under sk_callback_lock, and then the smc_sock itself
can be freed via sock_put() in smc_release().
This leads to two issues:
1) NULL pointer dereference: sk_user_data is NULL when
accessed.
2) Use-after-free: sk_user_data is read as non-NULL, but the
smc_sock is freed before its fields (e.g., queued_smc_hs,
ori_af_ops) are accessed.
The race window looks like this (the syzkaller crash [1]
triggers via the SYN cookie path: tcp_get_cookie_sock() ->
smc_tcp_syn_recv_sock(), but the normal tcp_check_req() path
has the same race):
CPU A (softirq) CPU B (process ctx)
tcp_v4_rcv()
TCP_NEW_SYN_RECV:
sk = req->rsk_listener
sock_hold(sk)
/* No lock on listener */
smc_close_active():
write_lock_bh(cb_lock)
sk_user_data = NULL
write_unlock_bh(cb_lock)
...
smc_clcsock_release()
sock_put(smc->sk) x2
-> smc_sock freed!
tcp_check_req()
smc_tcp_syn_recv_sock():
smc = user_data(sk)
-> NULL or dangling
smc->queued_smc_hs
-> crash!
Note that the clcsock and smc_sock are two independent objects
with separate refcounts. TCP stack holds a reference on the
clcsock, which keeps it alive, but this does NOT prevent the
smc_sock from being freed.
Fix this by using RCU and refcount_inc_not_zero() to safely
access smc_sock. Since smc_tcp_syn_recv_sock() is called in
the TCP three-way handshake path, taking read_lock_bh on
sk_callback_lock is too heavy and would not survive a SYN
flood attack. Using rcu_read_lock() is much more lightweight.
- Set SOCK_RCU_FREE on the SMC listen socket so that
smc_sock freeing is deferred until after the RCU grace
period. This guarantees the memory is still valid when
accessed inside rcu_read_lock().
- Use rcu_read_lock() to protect reading sk_user_data.
- Use refcount_inc_not_zero(&smc->sk.sk_refcnt) to pin the
smc_sock. If the refcount has already reached zero (close
path completed), it returns false and we bail out safely.
Note: smc_hs_congested() has a similar lockless read of
sk_user_data without rcu_read_lock(), but it only checks for
NULL and accesses the global smc_hs_wq, never dereferencing
any smc_sock field, so it is not affected.
Reproducer was verified with mdelay injection and smc_run,
the issue no longer occurs with this patch applied.
bond_header_parse() can loop if a stack of two bonding devices is setup,
because skb->dev always points to the hierarchy top.
Add new "const struct net_device *dev" parameter to
(struct header_ops)->parse() method to make sure the recursion
is bounded, and that the final leaf parse method is called.
Fixes: 950803f72547 ("bonding: fix type confusion in bond_setup_by_slave()") Signed-off-by: Eric Dumazet <edumazet@google.com> Reviewed-by: Jiayuan Chen <jiayuan.chen@shopee.com> Tested-by: Jiayuan Chen <jiayuan.chen@shopee.com> Cc: Jay Vosburgh <jv@jvosburgh.net> Cc: Andrew Lunn <andrew+netdev@lunn.ch> Link: https://patch.msgid.link/20260315104152.1436867-1-edumazet@google.com Signed-off-by: Jakub Kicinski <kuba@kernel.org> Signed-off-by: Sasha Levin <sashal@kernel.org>
There are two special cases in the idle loop that are handled
inconsistently even though they are analogous.
The first one is when a cpuidle driver is absent and the default CPU
idle time power management implemented by the architecture code is used.
In that case, the scheduler tick is stopped every time before invoking
default_idle_call().
The second one is when a cpuidle driver is present, but there is only
one idle state in its table. In that case, the scheduler tick is never
stopped at all.
Since each of these approaches has its drawbacks, reconcile them with
the help of one simple heuristic. Namely, stop the tick if the CPU has
been woken up by it in the previous iteration of the idle loop, or let
it tick otherwise.
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com> Reviewed-by: Christian Loehle <christian.loehle@arm.com> Reviewed-by: Frederic Weisbecker <frederic@kernel.org> Reviewed-by: Qais Yousef <qyousef@layalina.io> Reviewed-by: Aboorva Devarajan <aboorvad@linux.ibm.com> Fixes: ed98c3491998 ("sched: idle: Do not stop the tick before cpuidle_idle_call()")
[ rjw: Added Fixes tag, changelog edits ] Link: https://patch.msgid.link/4741364.LvFx2qVVIh@rafael.j.wysocki Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com> Signed-off-by: Sasha Levin <sashal@kernel.org>
A potential race condition exists in mana_hwc_destroy_channel() where
hwc->caller_ctx is freed before the HWC's Completion Queue (CQ) and
Event Queue (EQ) are destroyed. This allows an in-flight CQ interrupt
handler to dereference freed memory, leading to a use-after-free or
NULL pointer dereference in mana_hwc_handle_resp().
mana_smc_teardown_hwc() signals the hardware to stop but does not
synchronize against IRQ handlers already executing on other CPUs. The
IRQ synchronization only happens in mana_hwc_destroy_cq() via
mana_gd_destroy_eq() -> mana_gd_deregister_irq(). Since this runs
after kfree(hwc->caller_ctx), a concurrent mana_hwc_rx_event_handler()
can dereference freed caller_ctx (and rxq->msg_buf) in
mana_hwc_handle_resp().
Fix this by reordering teardown to reverse-of-creation order: destroy
the TX/RX work queues and CQ/EQ before freeing hwc->caller_ctx. This
ensures all in-flight interrupt handlers complete before the memory they
access is freed.
Fixes: ca9c54d2d6a5 ("net: mana: Add a driver for Microsoft Azure Network Adapter (MANA)") Reviewed-by: Haiyang Zhang <haiyangz@microsoft.com> Signed-off-by: Dipayaan Roy <dipayanroy@linux.microsoft.com> Reviewed-by: Simon Horman <horms@kernel.org> Link: https://patch.msgid.link/abHA3AjNtqa1nx9k@linuxonhyperv3.guj3yctzbm1etfxqx2vob5hsef.xx.internal.cloudapp.net Signed-off-by: Jakub Kicinski <kuba@kernel.org> Signed-off-by: Sasha Levin <sashal@kernel.org>
In DecodeQ931(), the UserUserIE code path reads a 16-bit length from
the packet, then decrements it by 1 to skip the protocol discriminator
byte before passing it to DecodeH323_UserInformation(). If the encoded
length is 0, the decrement wraps to -1, which is then passed as a
large value to the decoder, leading to an out-of-bounds read.
Add a check to ensure len is positive after the decrement.
The monthday field can be up to 31, and shifting a signed integer 1
by 31 positions (1 << 31) is undefined behavior in C, as the result
overflows a 32-bit signed int. Use 1U to ensure well-defined behavior
for all valid monthday values.
Change the weekday shift to 1U as well for consistency.
Packets sitting in nfqueue might hold a reference to:
- templates that specify the conntrack zone, because a percpu area is
used and module removal is possible.
- conntrack timeout policies and helper, where object removal leave
a stale reference.
Since these objects can just go away, drop enqueued packets to avoid
stale reference to them.
If there is a need for finer grain removal, this logic can be revisited
to make selective packet drop upon dependencies.
If cloning the second stateful expression in the element via GFP_ATOMIC
fails, then the first stateful expression remains in place without being
released.
Fixes: 563125a73ac3 ("netfilter: nftables: generalize set extension to support for several expressions") Reported-by: Gurpreet Shergill <giki.shergill@proton.me> Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org> Signed-off-by: Florian Westphal <fw@strlen.de> Signed-off-by: Sasha Levin <sashal@kernel.org>
In decode_int(), the CONS case calls get_bits(bs, 2) to read a length
value, then calls get_uint(bs, len) without checking that len bytes
remain in the buffer. The existing boundary check only validates the
2 bits for get_bits(), not the subsequent 1-4 bytes that get_uint()
reads. This allows a malformed H.323/RAS packet to cause a 1-4 byte
slab-out-of-bounds read.
Add a boundary check for len bytes after get_bits() and before
get_uint().
sip_help_tcp() parses the SIP Content-Length header with
simple_strtoul(), which returns unsigned long, but stores the result in
unsigned int clen. On 64-bit systems, values exceeding UINT_MAX are
silently truncated before computing the SIP message boundary.
For example, Content-Length 4294967328 (2^32 + 32) is truncated to 32,
causing the parser to miscalculate where the current message ends. The
loop then treats trailing data in the TCP segment as a second SIP
message and processes it through the SDP parser.
Fix this by changing clen to unsigned long to match the return type of
simple_strtoul(), and reject Content-Length values that exceed the
remaining TCP payload length.
ctnetlink_dump_exp_ct() stores a conntrack pointer in cb->data for the
netlink dump callback ctnetlink_exp_ct_dump_table(), but drops the
conntrack reference immediately after netlink_dump_start(). When the
dump spans multiple rounds, the second recvmsg() triggers the dump
callback which dereferences the now-freed conntrack via nfct_help(ct),
leading to a use-after-free on ct->ext.
The bug is that the netlink_dump_control has no .start or .done
callbacks to manage the conntrack reference across dump rounds. Other
dump functions in the same file (e.g. ctnetlink_get_conntrack) properly
use .start/.done callbacks for this purpose.
Fix this by adding .start and .done callbacks that hold and release the
conntrack reference for the duration of the dump, and move the
nfct_help() call after the cb->args[0] early-return check in the dump
callback to avoid dereferencing ct->ext unnecessarily.
BUG: KASAN: slab-use-after-free in ctnetlink_exp_ct_dump_table+0x4f/0x2e0
Read of size 8 at addr ffff88810597ebf0 by task ctnetlink_poc/133
Same pattern as previous patch: do not keep the expectation object
alive via refcount, only store a cookie value and then use that
as the skip hint for dump resumption.
AFAICS this has the same issue as the one resolved in the conntrack
dumper, when we do
if (!refcount_inc_not_zero(&exp->use))
to increment the refcount, there is a chance that exp == last, which
causes a double-increment of the refcount and subsequent memory leak.
Fixes: cf6994c2b981 ("[NETFILTER]: nf_conntrack_netlink: sync expectation dumping with conntrack table dumping") Fixes: e844a928431f ("netfilter: ctnetlink: allow to dump expectation per master conntrack") Signed-off-by: Florian Westphal <fw@strlen.de> Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
Stable-dep-of: 5cb81eeda909 ("netfilter: ctnetlink: fix use-after-free in ctnetlink_dump_exp_ct()") Signed-off-by: Sasha Levin <sashal@kernel.org>
If mpls_init() fails after registering mpls_dev_notifier, it never
gets removed. Add the missing unregister_netdevice_notifier() call to
the error handling path.
syzkaller reported a bug [1], and the reproducer is available at [2].
ROSE sockets use four sk->sk_state values: TCP_CLOSE, TCP_LISTEN,
TCP_SYN_SENT, and TCP_ESTABLISHED. rose_connect() already rejects
calls for TCP_ESTABLISHED (-EISCONN) and TCP_CLOSE with SS_CONNECTING
(-ECONNREFUSED), but lacks a check for TCP_SYN_SENT.
When rose_connect() is called a second time while the first connection
attempt is still in progress (TCP_SYN_SENT), it overwrites
rose->neighbour via rose_get_neigh(). If that returns NULL, the socket
is left with rose->state == ROSE_STATE_1 but rose->neighbour == NULL.
When the socket is subsequently closed, rose_release() sees
ROSE_STATE_1 and calls rose_write_internal() ->
rose_transmit_link(skb, NULL), causing a NULL pointer dereference.
Per connect(2), a second connect() while a connection is already in
progress should return -EALREADY. Add this missing check for
TCP_SYN_SENT to complete the state validation in rose_connect().
When a peer MEP is being deleted, cancel_delayed_work_sync() is called
on ccm_rx_dwork before freeing. However, br_cfm_frame_rx() runs in
softirq context under rcu_read_lock (without RTNL) and can re-schedule
ccm_rx_dwork via ccm_rx_timer_start() between cancel_delayed_work_sync()
returning and kfree_rcu() being called.
The following is a simple race scenario:
cpu0 cpu1
mep_delete_implementation()
cancel_delayed_work_sync(ccm_rx_dwork);
br_cfm_frame_rx()
// peer_mep still in hlist
if (peer_mep->ccm_defect)
ccm_rx_timer_start()
queue_delayed_work(ccm_rx_dwork)
hlist_del_rcu(&peer_mep->head);
kfree_rcu(peer_mep, rcu);
ccm_rx_work_expired()
// on freed peer_mep
To prevent this, cancel_delayed_work_sync() is replaced with
disable_delayed_work_sync() in both peer MEP deletion paths, so
that subsequent queue_delayed_work() calls from br_cfm_frame_rx()
are silently rejected.
The cc_peer_disable() helper retains cancel_delayed_work_sync()
because it is also used for the CC enable/disable toggle path where
the work must remain re-schedulable.
Fixes: dc32cbb3dbd7 ("bridge: cfm: Kernel space implementation of CFM. CCM frame RX added.") Signed-off-by: Hyunwoo Kim <imv4bel@gmail.com> Reviewed-by: Ido Schimmel <idosch@nvidia.com> Link: https://patch.msgid.link/abBgYT5K_FI9rD1a@v4bel Signed-off-by: Jakub Kicinski <kuba@kernel.org> Signed-off-by: Sasha Levin <sashal@kernel.org>
After commit ab4eedb790ca ("Bluetooth: L2CAP: Fix corrupted list in
hci_chan_del"), l2cap_conn_del() uses conn->lock to protect access to
conn->users. However, l2cap_register_user() and l2cap_unregister_user()
don't use conn->lock, creating a race condition where these functions can
access conn->users and conn->hchan concurrently with l2cap_conn_del().
This can lead to use-after-free and list corruption bugs, as reported
by syzbot.
Fix this by changing l2cap_register_user() and l2cap_unregister_user()
to use conn->lock instead of hci_dev_lock(), ensuring consistent locking
for the l2cap_conn structure.
Reported-by: syzbot+14b6d57fb728e27ce23c@syzkaller.appspotmail.com Closes: https://syzkaller.appspot.com/bug?extid=14b6d57fb728e27ce23c Fixes: ab4eedb790ca ("Bluetooth: L2CAP: Fix corrupted list in hci_chan_del") Signed-off-by: Shaurya Rane <ssrane_b23@ee.vjti.ac.in> Signed-off-by: Luiz Augusto von Dentz <luiz.von.dentz@intel.com> Signed-off-by: Sasha Levin <sashal@kernel.org>
Commit 302a1f674c00 ("Bluetooth: MGMT: Fix possible UAFs") introduced
mgmt_pending_valid(), which not only validates the pending command but
also unlinks it from the pending list if it is valid. This change in
semantics requires updates to several completion handlers to avoid list
corruption and memory safety issues.
This patch addresses two left-over issues from the aforementioned rework:
1. In mgmt_add_adv_patterns_monitor_complete(), mgmt_pending_remove()
is replaced with mgmt_pending_free() in the success path. Since
mgmt_pending_valid() already unlinks the command at the beginning of
the function, calling mgmt_pending_remove() leads to a double list_del()
and subsequent list corruption/kernel panic.
2. In set_mesh_complete(), the use of mgmt_pending_foreach() in the error
path is removed. Since the current command is already unlinked by
mgmt_pending_valid(), this foreach loop would incorrectly target other
pending mesh commands, potentially freeing them while they are still being
processed concurrently (leading to UAFs). The redundant mgmt_cmd_status()
is also simplified to use cmd->opcode directly.
Fixes: 302a1f674c00 ("Bluetooth: MGMT: Fix possible UAFs") Signed-off-by: Wang Tao <wangtao554@huawei.com> Signed-off-by: Luiz Augusto von Dentz <luiz.von.dentz@intel.com> Signed-off-by: Sasha Levin <sashal@kernel.org>
While introducing hci_le_create_conn_sync the functionality
of hci_connect_le was ported to hci_le_create_conn_sync including
the disable of the scan before starting the connection.
When this code was run non synchronously the immediate call that was
setting the flag HCI_LE_SCAN_INTERRUPTED had an impact. Since the
completion handler for the LE_SCAN_DISABLE was not immediately called.
In the completion handler of the LE_SCAN_DISABLE event, this flag is
checked to set the state of the hdev to DISCOVERY_STOPPED.
With the synchronised approach the later setting of the
HCI_LE_SCAN_INTERRUPTED flag has not the same effect. The completion
handler would immediately fire in the LE_SCAN_DISABLE call, check for
the flag, which is then not yet set and do nothing.
To fix this issue and make the function call work as before, we move the
setting of the flag HCI_LE_SCAN_INTERRUPTED before disabling the scan.
Fixes: 8e8b92ee60de ("Bluetooth: hci_sync: Add hci_le_create_conn_sync") Signed-off-by: Michael Grzeschik <m.grzeschik@pengutronix.de> Signed-off-by: Luiz Augusto von Dentz <luiz.von.dentz@intel.com> Signed-off-by: Sasha Levin <sashal@kernel.org>
iso-tester defer tests seem to fail with hci_conn_hash_lookup_cig
being unable to resolve a cig in set_cig_params_sync due a race
where it is run immediatelly before hci_bind_cis is able to set
the QoS settings into the hci_conn object.
So this moves the assigning of the QoS settings to be done directly
by hci_le_set_cig_params to prevent that from happening again.
Fixes: 26afbd826ee3 ("Bluetooth: Add initial implementation of CIS connections") Signed-off-by: Luiz Augusto von Dentz <luiz.von.dentz@intel.com> Signed-off-by: Sasha Levin <sashal@kernel.org>
The last test step ("Test with Invalid public key X and Y, all set to
0") expects to get an "DHKEY check failed" instead of "unspecified".
Fixes: 6d19628f539f ("Bluetooth: SMP: Fail if remote and local public keys are identical") Signed-off-by: Christian Eggers <ceggers@arri.de> Signed-off-by: Luiz Augusto von Dentz <luiz.von.dentz@intel.com> Signed-off-by: Sasha Levin <sashal@kernel.org>
Core 6.0, Vol 3, Part A, 3.4.3:
"... If the sum of the payload sizes for the K-frames exceeds the
specified SDU length, the receiver shall disconnect the channel."
This fixes L2CAP/LE/CFC/BV-27-C (running together with 'l2test -r -P
0x0027 -V le_public').
Fixes: aac23bf63659 ("Bluetooth: Implement LE L2CAP reassembly") Signed-off-by: Christian Eggers <ceggers@arri.de> Signed-off-by: Luiz Augusto von Dentz <luiz.von.dentz@intel.com> Signed-off-by: Sasha Levin <sashal@kernel.org>
A device_node reference obtained from the device tree is not released
on all error paths in the arm_scpi probe path. Specifically, a node
returned by of_parse_phandle() could be leaked when the probe failed
after the node was acquired. The probe function returns early and
the shmem reference is not released.
Use __free(device_node) scope-based cleanup to automatically release
the reference when the variable goes out of scope.
According to the FF-A specification (DEN0077, v1.1, §13.7), when
FFA_RXTX_UNMAP is invoked from any instance other than non-secure
physical, the w1 register must be zero (MBZ). If a non-zero value is
supplied in this context, the SPMC must return FFA_INVALID_PARAMETER.
The Arm FF-A driver operates exclusively as a guest or non-secure
physical instance where the partition ID is always zero and is not
invoked from a hypervisor context where w1 carries a VM ID. In this
execution model, the partition ID observed by the driver is always zero,
and passing a VM ID is unnecessary and potentially invalid.
Remove the vm_id parameter from ffa_rxtx_unmap() and ensure that the
SMC call is issued with w1 implicitly zeroed, as required by the
specification. This prevents invalid parameter errors and aligns the
implementation with the defined FF-A ABI behavior.
The HW user manual for the Renesas RZ/V2H(P) SoC (a.k.a r9a09g057)
states that only WDT1 is supposed to be accessed by the CA55 cores.
WDT0 is supposed to be used by the CM33 core, WDT2 is supposed
to be used by the CR8 core 0, and WDT3 is supposed to be used
by the CR8 core 1.
Remove wdt{0,2,3} from the SoC specific device tree to make it
compliant with the specification from the HW manual.
This change is harmless as there are currently no users of the
wdt{0,2,3} device tree nodes, only the wdt1 node is actually used.
When the nl80211 socket that originated a PMSR request is
closed, cfg80211_release_pmsr() sets the request's nl_portid
to zero and schedules pmsr_free_wk to process the abort
asynchronously. If the interface is concurrently torn down
before that work runs, cfg80211_pmsr_wdev_down() calls
cfg80211_pmsr_process_abort() directly. However, the already-
scheduled pmsr_free_wk work item remains pending and may run
after the interface has been removed from the driver. This
could cause the driver's abort_pmsr callback to operate on a
torn-down interface, leading to undefined behavior and
potential crashes.
Cancel pmsr_free_wk synchronously in cfg80211_pmsr_wdev_down()
before calling cfg80211_pmsr_process_abort(). This ensures any
pending or in-progress work is drained before interface teardown
proceeds, preventing the work from invoking the driver abort
callback after the interface is gone.
syzbot reported static_branch_dec() underflow in aql_enable_write(). [0]
The problem is that aql_enable_write() does not serialise concurrent
write()s to the debugfs.
aql_enable_write() checks static_key_false(&aql_disable.key) and
later calls static_branch_inc() or static_branch_dec(), but the
state may change between the two calls.
aql_disable does not need to track inc/dec.
Let's use static_branch_enable() and static_branch_disable().
Fix wrong variable used for error checking after devm_ioremap_resource()
call. The function checks qmc->scc_pram instead of qmc->dpram, which
could lead to incorrect error handling.
When QMAN_FQ_FLAG_DYNAMIC_FQID is set, there's a race condition between
fq_table[fq->idx] state and freeing/allocating from the pool and
WARN_ON(fq_table[fq->idx]) in qman_create_fq() gets triggered.
Indeed, we can have:
Thread A Thread B
qman_destroy_fq() qman_create_fq()
qman_release_fqid()
qman_shutdown_fq()
gen_pool_free()
-- At this point, the fqid is available again --
qman_alloc_fqid()
-- so, we can get the just-freed fqid in thread B --
fq->fqid = fqid;
fq->idx = fqid * 2;
WARN_ON(fq_table[fq->idx]);
fq_table[fq->idx] = fq;
fq_table[fq->idx] = NULL;
And adding some logs between qman_release_fqid() and
fq_table[fq->idx] = NULL makes the WARN_ON() trigger a lot more.
To prevent that, ensure that fq_table[fq->idx] is set to NULL before
gen_pool_free() is called by using smp_wmb().
In ax45mp_cache_init(), of_find_matching_node() returns a device node
with an incremented reference count that must be released with
of_node_put(). The current code fails to call of_node_put() which
causes a reference leak.
Use the __free(device_node) attribute to ensure automatic cleanup when
the variable goes out of scope.
Fixes: d34599bcd2e4 ("cache: Add L2 cache management for Andes AX45MP RISC-V core") Signed-off-by: Felix Gu <ustc.gu@gmail.com> Signed-off-by: Conor Dooley <conor.dooley@microchip.com> Signed-off-by: Sasha Levin <sashal@kernel.org>
In mpfs_sys_controller_probe(), if of_get_mtd_device_by_node() fails,
the function returns immediately without freeing the allocated memory
for sys_controller, leading to a memory leak.
Fix this by jumping to the out_free label to ensure the memory is
properly freed.
Also, consolidate the error handling for the mbox_request_channel()
failure case to use the same label.
If we log the parent directory of a conflicting inode, we are not logging
the new dentries of the directory, so when we finish we have the parent
directory's inode marked as logged but we did not log its new dentries.
As a consequence if the parent directory is explicitly fsynced later and
it does not have any new changes since we logged it, the fsync is a no-op
and after a power failure the new dentries are missing.
Example scenario:
$ mkdir foo
$ sync
$rmdir foo
$ mkdir dir1
$ mkdir dir2
# A file with the same name and parent as the directory we just deleted
# and was persisted in a past transaction. So the deleted directory's
# inode is a conflicting inode of this new file's inode.
$ touch foo
$ ln foo dir2/link
# The fsync on dir2 will log the parent directory (".") because the
# conflicting inode (deleted directory) does not exists anymore, but it
# it does not log its new dentries (dir1).
$ xfs_io -c "fsync" dir2
# This fsync on the parent directory is no-op, since the previous fsync
# logged it (but without logging its new dentries).
$ xfs_io -c "fsync" .
<power failure>
# After log replay dir1 is missing.
Fix this by ensuring we log new dir dentries whenever we log the parent
directory of a no longer existing conflicting inode.
A test case for fstests will follow soon.
Reported-by: Vyacheslav Kovalevsky <slava.kovalevskiy.2014@gmail.com> Link: https://lore.kernel.org/linux-btrfs/182055fa-e9ce-4089-9f5f-4b8a23e8dd91@gmail.com/ Fixes: a3baaf0d786e ("Btrfs: fix fsync after succession of renames and unlink/rmdir") Reviewed-by: Boris Burkov <boris@bur.io> Signed-off-by: Filipe Manana <fdmanana@suse.com> Signed-off-by: David Sterba <dsterba@suse.com> Signed-off-by: Sasha Levin <sashal@kernel.org>
Commit 4ab7bb976343 ("ata: libata-scsi: Refactor ata_scsiop_maint_in()")
modified ata_scsiop_maint_in() to directly call
ata_scsi_set_invalid_field() to set the field pointer of the sense data
of a failed MAINTENANCE IN command. However, in the case of an invalid
command format, the sense data field incorrectly indicates byte 1 of
the CDB. Fix this to indicate byte 2 of the command.
The function ata_scsi_rbuf_fill() used to fill the reply buffer of
emulated SCSI commands always copies the ATA reply buffer
(ata_scsi_rbuf) up to the size of the SCSI command buffer (the transfer
length for the command), even if the reply is shorter than the SCSI
command buffer. This leads to issuers of the SCSI command to always get
a result without any residual (resid is always 0) despite the
potentially shorter reply for the command.
Modify all fill actors used by ata_scsi_rbuf_fill() to return the number
of bytes filled for the reply and 0 in case of error. Using this value,
add a call to scsi_set_resid() in ata_scsi_rbuf_fill() to set the
correct residual for the SCSI command when the reply length is shorter
than the command buffer.
Signed-off-by: Damien Le Moal <dlemoal@kernel.org> Link: https://lore.kernel.org/r/20241022024537.251905-7-dlemoal@kernel.org Signed-off-by: Niklas Cassel <cassel@kernel.org>
Stable-dep-of: e6d7eba23b66 ("ata: libata-scsi: report correct sense field pointer in ata_scsiop_maint_in()") Signed-off-by: Sasha Levin <sashal@kernel.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Currently the code attempts to accept requests regardless of the
command identifier which may cause multiple requests to be marked
as pending (FLAG_DEFER_SETUP) which can cause more than
L2CAP_ECRED_MAX_CID(5) to be allocated in l2cap_ecred_rsp_defer
causing an overflow.
The spec is quite clear that the same identifier shall not be used on
subsequent requests:
'Within each signaling channel a different Identifier shall be used
for each successive request or indication.'
https://www.bluetooth.com/wp-content/uploads/Files/Specification/HTML/Core-62/out/en/host/logical-link-control-and-adaptation-protocol-specification.html#UUID-32a25a06-4aa4-c6c7-77c5-dcfe3682355d
So this attempts to check if there are any channels pending with the
same identifier and rejects if any are found.
Fixes: 15f02b910562 ("Bluetooth: L2CAP: Add initial code for Enhanced Credit Based Mode") Reported-by: Yiming Qian <yimingqian591@gmail.com> Signed-off-by: Luiz Augusto von Dentz <luiz.von.dentz@intel.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
GGTT MMIO access is currently protected by hotplug (drm_dev_enter),
which works correctly when the driver loads successfully and is later
unbound or unloaded. However, if driver load fails, this protection is
insufficient because drm_dev_unplug() is never called.
Additionally, devm release functions cannot guarantee that all BOs with
GGTT mappings are destroyed before the GGTT MMIO region is removed, as
some BOs may be freed asynchronously by worker threads.
To address this, introduce an open-coded flag, protected by the GGTT
lock, that guards GGTT MMIO access. The flag is cleared during the
dev_fini_ggtt devm release function to ensure MMIO access is disabled
once teardown begins.
Cc: stable@vger.kernel.org Fixes: 919bb54e989c ("drm/xe: Fix missing runtime outer protection for ggtt_remove_node") Reviewed-by: Zhanjun Dong <zhanjun.dong@intel.com> Signed-off-by: Matthew Brost <matthew.brost@intel.com> Link: https://patch.msgid.link/20260310225039.1320161-8-zhanjun.dong@intel.com
(cherry picked from commit 4f3a998a173b4325c2efd90bdadc6ccd3ad9a431) Signed-off-by: Thomas Hellström <thomas.hellstrom@linux.intel.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Some OA data might be present in the OA buffer when OA stream is
disabled. Allow UMD's to retrieve this data, so that all data till the
point when OA stream is disabled can be retrieved.
v2: Update tail pointer after disable (Umesh)
Fixes: efb315d0a013 ("drm/xe/oa/uapi: Read file_operation") Cc: stable@vger.kernel.org Signed-off-by: Ashutosh Dixit <ashutosh.dixit@intel.com> Reviewed-by: Umesh Nerlige Ramappa<umesh.nerlige.ramappa@intel.com> Link: https://patch.msgid.link/20260313053630.3176100-1-ashutosh.dixit@intel.com
(cherry picked from commit 4ff57c5e8dbba23b5457be12f9709d5c016da16e) Signed-off-by: Thomas Hellström <thomas.hellstrom@linux.intel.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
The soft reset sequence is currently executed from the threaded IRQ
handler, hence it cannot call disable_irq() which internally waits
for IRQ handlers, i.e. itself, to complete.
Use disable_irq_nosync() during a soft reset instead.
Fixes: cc1aeedb98ad ("drm/imagination: Implement firmware infrastructure and META FW support") Cc: stable@vger.kernel.org Signed-off-by: Alessio Belle <alessio.belle@imgtec.com> Reviewed-by: Matt Coster <matt.coster@imgtec.com> Link: https://patch.msgid.link/20260309-fix-soft-reset-v1-1-121113be554f@imgtec.com Signed-off-by: Matt Coster <matt.coster@imgtec.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
[Why]
The dcn32_override_min_req_memclk function is in dcn32_fpu.c, which is
compiled with CC_FLAGS_FPU into FP instructions. So when we call it we
must use DC_FP_{START,END} to save and restore the FP context, and
prepare the FP unit on architectures like LoongArch where the FP unit
isn't always on.
Reported-by: LiarOnce <liaronce@hotmail.com> Fixes: ee7be8f3de1c ("drm/amd/display: Limit DCN32 8 channel or less parts to DPM1 for FPO") Signed-off-by: Xi Ruoyao <xry111@xry111.site> Reviewed-by: Alex Hung <alex.hung@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
(cherry picked from commit 25bb1d54ba3983c064361033a8ec15474fece37e) Cc: stable@vger.kernel.org Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
When trying to do a rather aggressive test of igt's "xe_module_load
--r reload" with a full desktop environment and game running I noticed
a few OOPSes when dereferencing freed pointers, related to
framebuffers and property blobs after the compositor exits.
Solve this by guarding the freeing in drm_file with drm_dev_enter/exit,
and immediately put the references from struct drm_file objects during
drm_dev_unplug().
Add a guard around file close, and ensure the warnings from drm_mode_config
do not trigger. Fix those by allowing an open reference to the file descriptor
and cleaning up the file linked list entry in drm_mode_config_cleanup().
Cc: <stable@vger.kernel.org> # v4.18+ Fixes: bee330f3d672 ("drm: Use srcu to protect drm_device.unplugged") Cc: Thomas Hellström <thomas.hellstrom@linux.intel.com> Reviewed-by: Thomas Hellström <thomas.hellstrom@linux.intel.com> Link: https://patch.msgid.link/20260313151728.14990-4-dev@lankhorst.se Signed-off-by: Maarten Lankhorst <dev@lankhorst.se> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
When io_should_commit() returns true (eg for non-pollable files), buffer
commit happens at buffer selection time and sel->buf_list is set to
NULL. When __io_put_kbufs() generates CQE flags at completion time, it
calls __io_put_kbuf_ring() which finds a NULL buffer_list and hence
cannot determine whether the buffer was consumed or not. This means that
IORING_CQE_F_BUF_MORE is never set for non-pollable input with
incrementally consumed buffers.
Likewise for io_buffers_select(), which always commits upfront and
discards the return value of io_kbuf_commit().
Add REQ_F_BUF_MORE to store the result of io_kbuf_commit() during early
commit. Then __io_put_kbuf_ring() can check this flag and set
IORING_F_BUF_MORE accordingy.
Reported-by: Martin Michaelis <code@mgjm.de> Cc: stable@vger.kernel.org Fixes: ae98dbf43d75 ("io_uring/kbuf: add support for incremental buffer consumption") Link: https://github.com/axboe/liburing/issues/1553 Signed-off-by: Jens Axboe <axboe@kernel.dk> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
ulite_probe() calls pm_runtime_put_autosuspend() at the end of probe
without holding a corresponding PM runtime reference for non-console
ports.
During ulite_assign(), uart_add_one_port() triggers uart_configure_port()
which calls ulite_pm() via uart_change_pm(). For non-console ports, the
UART core performs a balanced get/put cycle:
This leaves no spare reference for the pm_runtime_put_autosuspend() at
the end of probe. The PM runtime core prevents the count from actually
going below zero, and instead triggers a
"Runtime PM usage count underflow!" warning.
For console ports the bug is masked: the UART core skips the
uart_change_pm(OFF) call, so the UART core's unbalanced get happens to
pair with probe's trailing put.
Add pm_runtime_get_noresume() before pm_runtime_enable() to take an
explicit probe-owned reference that the trailing
pm_runtime_put_autosuspend() can release. This ensures a correct usage
count regardless of whether the port is a console.
When DW UART is !uart_16550_compatible, it can indicate BUSY at any
point (when under constant Rx pressure) unless a complex sequence of
steps is performed. Any LCR write can run a foul with the condition
that prevents writing LCR while the UART is BUSY, which triggers
BUSY_DETECT interrupt that seems unmaskable using IER bits.
Normal flow is that dw8250_handle_irq() handles BUSY_DETECT condition
by reading USR register. This BUSY feature, however, breaks the
assumptions made in serial8250_do_shutdown(), which runs
synchronize_irq() after clearing IER and assumes no interrupts can
occur after that point but then proceeds to update LCR, which on DW
UART can trigger an interrupt.
If serial8250_do_shutdown() releases the interrupt handler before the
handler has run and processed the BUSY_DETECT condition by read the USR
register, the IRQ is not deasserted resulting in interrupt storm that
triggers "irq x: nobody cared" warning leading to disabling the IRQ.
Add late synchronize_irq() into serial8250_do_shutdown() to ensure
BUSY_DETECT from DW UART is handled before port's interrupt handler is
released. Alternative would be to add DW UART specific shutdown
function but it would mostly duplicate the generic code and the extra
synchronize_irq() seems pretty harmless in serial8250_do_shutdown().
`dmaengine_terminate_async` does not guarantee that the
`__dma_tx_complete` callback will run. The callback is currently the
only place where `dma->tx_running` gets cleared. If the transaction is
canceled and the callback never runs, then `dma->tx_running` will never
get cleared and we will never schedule new TX DMA transactions again.
This change makes it so we clear `dma->tx_running` after we terminate
the DMA transaction. This is "safe" because `serial8250_tx_dma_flush`
is holding the UART port lock. The first thing the callback does is also
grab the UART port lock, so access to `dma->tx_running` is serialized.
This is found in popular brands such as StarTech.com or Delock, and has
been a source of frustration to quite a few people, if I can trust
Amazon comments complaining about Linux support via the official
out-of-the-tree driver.
During the qi_check_fault process after an IOMMU ITE event, requests at
odd-numbered positions in the queue are set to QI_ABORT, only satisfying
single-request submissions. However, qi_submit_sync now supports multiple
simultaneous submissions, and can't guarantee that the wait_desc will be
at an odd-numbered position. Therefore, if an item times out, IOMMU can't
re-initiate the request, resulting in an infinite polling wait.
This modifies the process by setting the status of all requests already
fetched by IOMMU and recorded as QI_IN_USE status (including wait_desc
requests) to QI_ABORT, thus enabling multiple requests to be resubmitted.
Given CONFIG_FORTIFY_SOURCE=y and a recent compiler,
commit 439a1bcac648 ("fortify: Use __builtin_dynamic_object_size() when
available") produces the warning below and an oops.
Searching for RedBoot partition table in 50000000.flash at offset 0x7e0000
------------[ cut here ]------------
WARNING: lib/string_helpers.c:1035 at 0xc029e04c, CPU#0: swapper/0/1
memcmp: detected buffer overflow: 15 byte read of buffer size 14
Modules linked in:
CPU: 0 UID: 0 PID: 1 Comm: swapper/0 Not tainted 6.19.0 #1 NONE
As Kees said, "'names' is pointing to the final 'namelen' many bytes
of the allocation ... 'namelen' could be basically any length at all.
This fortify warning looks legit to me -- this code used to be reading
beyond the end of the allocation."
Since the size of the dynamic allocation is calculated with strlen()
we can use strcmp() instead of memcmp() and remain within bounds.
Fix wrong variable used for error checking after dma_alloc_coherent()
call. The function checks cdns_ctrl->dma_cdma_desc instead of
cdns_ctrl->cdma_desc, which could lead to incorrect error handling.
Fixes: ec4ba01e894d ("mtd: rawnand: Add new Cadence NAND driver to MTD subsystem") Cc: stable@vger.kernel.org Signed-off-by: Chen Ni <nichen@iscas.ac.cn> Reviewed-by: Alok Tiwari <alok.a.tiwari@oracle.com> Signed-off-by: Miquel Raynal <miquel.raynal@bootlin.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Timings of the nand are adjusted by pl35x_nfc_setup_interface() but
actually applied by the pl35x_nand_select_target() function.
If there is only one nand chip, the pl35x_nand_select_target() will only
apply the timings once since the test at its beginning will always be true
after the first call to this function. As a result, the hardware will
keep using the default timings set at boot to detect the nand chip, not
the optimal ones.
With this patch, we program directly the new timings when
pl35x_nfc_setup_interface() is called.
Fixes: 08d8c62164a3 ("mtd: rawnand: pl353: Add support for the ARM PL353 SMC NAND controller") Signed-off-by: Olivier Sobrie <olivier@sobrie.be> Cc: stable@vger.kernel.org Signed-off-by: Miquel Raynal <miquel.raynal@bootlin.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
The controller per-cpu statistics is not allocated until after the
controller has been registered with driver core, which leaves a window
where accessing the sysfs attributes can trigger a NULL-pointer
dereference.
Fix this by moving the statistics allocation to controller allocation
while tying its lifetime to that of the controller (rather than using
implicit devres).
Fixes: 6598b91b5ac3 ("spi: spi.c: Convert statistics to per-cpu u64_stats_t") Cc: stable@vger.kernel.org # 6.0 Cc: David Jander <david@protonic.nl> Signed-off-by: Johan Hovold <johan@kernel.org> Link: https://patch.msgid.link/20260312151817.32100-3-johan@kernel.org Signed-off-by: Mark Brown <broonie@kernel.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Make sure to deregister from driver core also in the unlikely event that
per-cpu statistics allocation fails during controller registration to
avoid use-after-free (of driver resources) and unclocked register
accesses.
Fixes: 6598b91b5ac3 ("spi: spi.c: Convert statistics to per-cpu u64_stats_t") Cc: stable@vger.kernel.org # 6.0 Cc: David Jander <david@protonic.nl> Signed-off-by: Johan Hovold <johan@kernel.org> Link: https://patch.msgid.link/20260312151817.32100-2-johan@kernel.org Signed-off-by: Mark Brown <broonie@kernel.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
The bcm2835_asb_control() function uses a tight polling loop to wait
for the ASB bridge to acknowledge a request. During intensive workloads,
this handshake intermittently fails for V3D's master ASB on BCM2711,
resulting in "Failed to disable ASB master for v3d" errors during
runtime PM suspend. As a consequence, the failed power-off leaves V3D in
a broken state, leading to bus faults or system hangs on later accesses.
As the timeout is insufficient in some scenarios, increase the polling
timeout from 1us to 5us, which is still negligible in the context of a
power domain transition. Also, replace the open-coded ktime_get_ns()/
cpu_relax() polling loop with readl_poll_timeout_atomic().
Cc: stable@vger.kernel.org Fixes: 670c672608a1 ("soc: bcm: bcm2835-pm: Add support for power domains under a new binding.") Signed-off-by: Maíra Canal <mcanal@igalia.com> Reviewed-by: Stefan Wahren <wahrenst@gmx.net> Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
When 1-bit bus width is used with HS200/HS400 capabilities set,
mmc_select_hs200() returns 0 without actually switching. This
causes mmc_select_timing() to skip mmc_select_hs(), leaving eMMC
in legacy mode (26MHz) instead of High Speed SDR (52MHz).
Per JEDEC eMMC spec section 5.3.2, 1-bit mode supports High Speed
SDR. Drop incompatible HS200/HS400/UHS/DDR caps early so timing
selection falls through to mmc_select_hs() correctly.
Fixes: f2119df6b764 ("mmc: sd: add support for signal voltage switch procedure") Signed-off-by: Luke Wang <ziniu.wang_1@nxp.com> Acked-by: Adrian Hunter <adrian.hunter@intel.com> Cc: stable@vger.kernel.org Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
The GL9750 SD host controller has intermittent data corruption during
DMA write operations. The GM_BURST register's R_OSRC_Lmt field
(bits 17:16), which limits outstanding DMA read requests from system
memory, is not being cleared during initialization. The Windows driver
sets R_OSRC_Lmt to zero, limiting requests to the smallest unit.
Clear R_OSRC_Lmt to match the Windows driver behavior. This eliminates
write corruption verified with f3write/f3read tests while maintaining
DMA performance.
ADATA SU680 SSDs suffer from NCQ read and write commands timeouts or bus
errors when link power management (LPM) is enabled. Flag these devices
with the ATA_QUIRK_NOLPM quirk to prevent the use of LPM and avoid these
command failures.
Reported-by: Mohammad Khaled Bayan <mhd.khaled.bayan@gmail.com> Closes: https://bugs.launchpad.net/ubuntu/+source/linux-hwe-6.17/+bug/2144060 Cc: stable@vger.kernel.org Tested-by: Mohammad-Khaled Bayan <mhd.khaled.bayan@gmail.com> Signed-off-by: Damien Le Moal <dlemoal@kernel.org> Reviewed-by: Martin K. Petersen <martin.petersen@oracle.com> Signed-off-by: Niklas Cassel <cassel@kernel.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
On certain platforms, such as AMD Versal boards, the tx/rx queue pointer
registers are cleared after suspend, and the rx queue pointer register
is also disabled during suspend if WOL is enabled. Previously, we assumed
that these registers would be restored by macb_mac_link_up(). However,
in commit bf9cf80cab81, macb_init_buffers() was moved from
macb_mac_link_up() to macb_open(). Therefore, we should call
macb_init_buffers() to reinitialize the tx/rx queue pointer registers
during resume.
Due to the reset of these two registers, we also need to adjust the
tx/rx rings accordingly. The tx ring will be handled by
gem_shuffle_tx_rings() in macb_mac_link_up(), so we only need to
initialize the rx ring here.
Fixes: bf9cf80cab81 ("net: macb: Fix tx/rx malfunction after phy link down and up") Reported-by: Quanyang Wang <quanyang.wang@windriver.com> Signed-off-by: Kevin Hao <haokexin@gmail.com> Tested-by: Quanyang Wang <quanyang.wang@windriver.com> Cc: stable@vger.kernel.org Reviewed-by: Simon Horman <horms@kernel.org> Link: https://patch.msgid.link/20260312-macb-versal-v1-2-467647173fa4@gmail.com Signed-off-by: Jakub Kicinski <kuba@kernel.org> Signed-off-by: Sasha Levin <sashal@kernel.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Extract the initialization code for the GEM RX ring into a new function.
This change will be utilized in a subsequent patch. No functional changes
are introduced.
Signed-off-by: Kevin Hao <haokexin@gmail.com> Reviewed-by: Simon Horman <horms@kernel.org> Link: https://patch.msgid.link/20260312-macb-versal-v1-1-467647173fa4@gmail.com Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Stable-dep-of: 718d0766ce4c ("net: macb: Reinitialize tx/rx queue pointer registers and rx ring during resume") Signed-off-by: Sasha Levin <sashal@kernel.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
When OGM aggregation state is toggled at runtime, an existing forwarded
packet may have been allocated with only packet_len bytes, while a later
packet can still be selected for aggregation. Appending in this case can
hit skb_put overflow conditions.
Reject aggregation when the target skb tailroom cannot accommodate the new
packet. The caller then falls back to creating a new forward packet
instead of appending.
Fixes: c6c8fea29769 ("net: Add batman-adv meshing protocol") Cc: stable@vger.kernel.org Reported-by: Yifan Wu <yifanwucs@gmail.com> Reported-by: Juefei Pu <tomapufckgml@gmail.com> Signed-off-by: Yuan Tan <tanyuan98@outlook.com> Signed-off-by: Xin Liu <bird@lzu.edu.cn> Signed-off-by: Ao Zhou <n05ec@lzu.edu.cn> Signed-off-by: Yang Yang <n05ec@lzu.edu.cn> Signed-off-by: Sven Eckelmann <sven@narfation.org> Signed-off-by: Simon Wunderlich <sw@simonwunderlich.de>
[ Adjust context ] Signed-off-by: Sven Eckelmann <sven@narfation.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Currently a user can trigger a transaction abort by snapshotting a
previously received snapshot a bunch of times until we reach a
BTRFS_UUID_KEY_RECEIVED_SUBVOL item overflow (the maximum item size we
can store in a leaf). This is very likely not common in practice, but
if it happens, it turns the filesystem into RO mode. The snapshot, send
and set_received_subvol and subvol_setflags (used by receive) don't
require CAP_SYS_ADMIN, just inode_owner_or_capable(). A malicious user
could use this to turn a filesystem into RO mode and disrupt a system.
Reproducer script:
$ cat test.sh
#!/bin/bash
DEV=/dev/sdi
MNT=/mnt/sdi
# Use smallest node size to make the test faster.
mkfs.btrfs -f --nodesize 4K $DEV
mount $DEV $MNT
# Create a subvolume and set it to RO so that it can be used for send.
btrfs subvolume create $MNT/sv
touch $MNT/sv/foo
btrfs property set $MNT/sv ro true
# Send and receive the subvolume into snaps/sv.
mkdir $MNT/snaps
btrfs send $MNT/sv | btrfs receive $MNT/snaps
# Now snapshot the received subvolume, which has a received_uuid, a
# lot of times to trigger the leaf overflow.
total=500
for ((i = 1; i <= $total; i++)); do
echo -ne "\rCreating snapshot $i/$total"
btrfs subvolume snapshot -r $MNT/snaps/sv $MNT/snaps/sv_$i > /dev/null
done
echo
umount $MNT
When running the test:
$ ./test.sh
(...)
Create subvolume '/mnt/sdi/sv'
At subvol /mnt/sdi/sv
At subvol sv
Creating snapshot 496/500ERROR: Could not create subvolume: Value too large for defined data type
Creating snapshot 497/500ERROR: Could not create subvolume: Read-only file system
Creating snapshot 498/500ERROR: Could not create subvolume: Read-only file system
Creating snapshot 499/500ERROR: Could not create subvolume: Read-only file system
Creating snapshot 500/500ERROR: Could not create subvolume: Read-only file system
Fix this by ignoring -EOVERFLOW errors from btrfs_uuid_tree_add() in the
snapshot creation code when attempting to add the
BTRFS_UUID_KEY_RECEIVED_SUBVOL item. This is OK because it's not critical
and we are still able to delete the snapshot, as snapshot/subvolume
deletion ignores if a BTRFS_UUID_KEY_RECEIVED_SUBVOL is missing (see
inode.c:btrfs_delete_subvolume()). As for send/receive, we can still do
send/receive operations since it always peeks the first root ID in the
existing BTRFS_UUID_KEY_RECEIVED_SUBVOL (it could peek any since all
snapshots have the same content), and even if the key is missing, it
falls back to searching by BTRFS_UUID_KEY_SUBVOL key.
A test case for fstests will be sent soon.
Fixes: dd5f9615fc5c ("Btrfs: maintain subvolume items in the UUID tree") CC: stable@vger.kernel.org # 3.12+ Reviewed-by: Boris Burkov <boris@bur.io> Reviewed-by: Qu Wenruo <wqu@suse.com> Signed-off-by: Filipe Manana <fdmanana@suse.com> Reviewed-by: David Sterba <dsterba@suse.com> Signed-off-by: David Sterba <dsterba@suse.com>
[ adapted error check condition to omit unlikely() wrapper ] Signed-off-by: Sasha Levin <sashal@kernel.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Commit 61688a82e047 ("powerpc/bpf: enable kfunc call") inadvertently
enabled kfunc call support for 32-bit powerpc but that support will
not be possible until ABI mismatch between 32-bit powerpc and eBPF is
handled in 32-bit powerpc JIT code. Till then, advertise support only
for 64-bit powerpc. Also, in powerpc ABI, caller needs to extend the
arguments properly based on signedness. The JIT code is responsible
for handling this explicitly for kfunc calls as verifier can't handle
this for each architecture-specific ABI needs. But this was not taken
care of while kfunc call support was enabled for powerpc. Fix it by
handling this with bpf_jit_find_kfunc_model() and using zero_extend()
& sign_extend() helper functions.
projects / thirdparty / kernel / stable.git / log
