Latest CVE Announcements

When TCP socket migration fails at inet_ehash_insert() in reqsk_timer_handler(), we jump to the no_ownership: label and free the new reqsk immediately with __reqsk_free(). Thus, we must stop the new reqsk's timer before jumping to the label, but the timer might be missed since the cited commit, resulting in UAF. As we are in the original reqsk's timer context, we can safely call timer_delete_sync() for the new reqsk. Let's pass false to __inet_csk_reqsk_queue_drop() to stop the new reqsk's timer.

A task can be found on a non-local DSQ while the CPU which dispatched it is still running ops.dequeue() and hasn't cleared SCX_OPSS_DISPATCHING. The DSQ consume and move paths don't wait for the dispatch to finish and can operate on the task concurrently with the dispatcher. Call scx_wait_dispatching() in scx_move_local_task_to_local_dsq(), unlink_dsq_and_switch_rq_lock() and move_task_between_dsqs() before they operate on the task. In scx_move_local_task_to_local_dsq(), wait before scx_resolve_local_dsq() as it may update p->scx.flags.

With IORING_SETUP_DEFER_TASKRUN completions are queued as local work, and only become CQEs when the submitter enters the kernel to wait for events. For an application that finds out via poll() or epoll on the ring fd when it is time to do that, io_uring_poll() reports EPOLLIN if local work is pending. That only helps if the work is there at the time of the call. When DEFER_TASKRUN was added, io_req_local_work_add() did io_cqring_wake(), which took care of tasks sleeping in poll() as well. Then polling got a wait queue of its own, ->poll_wq, and the wake up in io_req_local_work_add() was replaced by waking the submitter directly. Since then ->poll_wq is only woken when CQEs are committed, which for such a ring is done by the very task that is waiting to be told to: submitter completion poll(ring_fd) io_uring_poll() // nothing yet schedule() io_req_local_work_add() // ->cq_wait_nr: nobody waits // sleeps on Wake ->poll_wq when the first entry is put on the work list. Like for the eventfd, that must not be done right away if the work is queued from a wake up handler, as the chain of wake ups could get back to the wait queue whose lock is held. Use an RCU callback in that case, with a reference to the ring.

With dynamic tracking a NAPI id that hasn't been seen for a minute is meant to be dropped from the list the ring busy polls. The timeout is added to jiffies, but it is defined as 60 * SEC_CONVERSION, and SEC_CONVERSION isn't the number of jiffies in a second. It is the internal factor of __timespec64_to_jiffies(), HZ scaled up by SEC_JIFFIE_SC bits. For HZ=1000 that is 1000 << 22, which puts the timeout at about eight years. On 32-bit the multiplication overflows. So entries are never removed, and every id that a socket of the ring ever had keeps being polled on each wait.

io_zcrx_queue_cqe() computes the offset of the buffer in its area from net_iov_idx(), which is an unsigned int, so the shift is done in 32 bits and only the result is widened to the u64 that goes into the CQE. An area can be up to 1TB in size. For every niov beyond the first 4GB the CQE points the application at a buffer 4GB, or a multiple of that, below the one which has the data. It reads the wrong data, and then returns a buffer it doesn't own via the refill queue while the real one is never given back.

io_pipe_fixed() installs both ends of the pipe in the fixed file table, drops ->uring_lock, and copies the slot numbers to the application. If that faults, it takes the lock again and removes whatever is in these slots. If the request is run by io-wq the lock really is dropped in between, and the slots are visible to everybody else. By the time the error path gets to them, they may have been closed and reused, by IORING_OP_CLOSE and a direct open or accept, or by IORING_REGISTER_FILES_UPDATE. Then a file is removed from the table that has nothing to do with the failed request. Do the copy with the lock held. Faulting in user memory under ->uring_lock is fine, the registration paths do it all the time.

Commit 41b70df5b38b ("io_uring/net: commit partial buffers on retry") commits ring provided buffers when a receive is going to be retried after a partial transfer, so that another request using the same buffer group can't be handed the memory the first one keeps receiving into. It commits the number of bytes received so far. That does the job for a single buffer which is consumed as a whole, but the request hangs on to all it has mapped: - For a buffer ring with IOU_PBUF_RING_INC only these bytes are consumed. The rest of the buffer stays at the head of the ring. - For a bundle only the buffers which hold these bytes are consumed, the others that were peeked stay in the ring. In both cases the next receive with that buffer group gets memory that the retried request is still going to write to, and whoever comes last overwrites the data of the other. What the retry fills isn't committed at all, as the buffer list isn't known any more at that point. So the kernel hands it out again later on, while the CQE tells the application that it was used, and for an incremental buffer that it is done with. The remainder can't be committed in advance: if the request ends up receiving less, there is no way to tell the application that more was consumed than the CQE accounts for. Complete the request with what it has got instead of retrying it. MSG_WAITALL is no longer honoured for receives into incremental buffers or multi buffer bundles, they can return a short count, with the request marked as failed like for other short MSG_WAITALL transfers.

If a wakeup sets IO_POLL_RETRY_FLAG between the owning task consuming the existing instance of IO_POLL_RETRY_FLAG and the loop check of io_poll_check_events(), then we stall on processing the request until an unrelated wakeup comes along. Using atomic_cmpxchg instead of atomic_sub_return to guarantee both the count v and poll_refs are 0 and no wake-ups are missed.

An IORING_OP_TIMEOUT run by io-wq, because of IOSQE_ASYNC, has the same problem as poll, futex and waitid. io_wq_submit_work() looks at IO_WQ_WORK_CANCEL once, before it issues the request. A cancel in between flags the work, gets -EALREADY, and searches ->timeout_list in vain, while the worker then adds the request and starts the timer: io-wq worker io_try_cancel() io_wq_submit_work() IO_WQ_WORK_CANCEL isn't set io_async_cancel_one() sets IO_WQ_WORK_CANCEL = -EALREADY io_timeout_cancel() = -ENOENT io_timeout() adds the request to ->timeout_list hrtimer_start() = IOU_ISSUE_SKIP_COMPLETE Nobody looks at the flag again, and the timeout runs its course. At ring exit, io_uring_try_cancel_requests() cancels io-wq work and then kills the timeouts the same way, so an exiting task can get stuck for the full duration of such a timeout. Look at the flag under ->timeout_lock, which the search takes as well, and fail the request if it is set. If the search came first, the flag is visible. If it comes later, it finds the request on the list.

It fixes both the move overflow and the eviction fence add for evicting these per-vm BOs.

When setting sparse/compressed flags on a resident attribute, the function skipped the resident-to-non-resident conversion and terminated.

For legacy 32-bit Tegra devices, the warning in the tegra_pmc_enter_suspend_mode() function is being triggered because the early_pmc->rate variable is never configured and hence, always 0. This issue started occurring after updating the PMC driver to create the PMC context dynamically which causes the 'rate' to be set in the new dynamic context, but missed updating the rate in the early PMC context. Fix this by initialising the early_pmc->rate variable for 32-bit Tegra devices when the rate for the PMC context is initialised. [treding@nvidia.com: move assignment to existing #ifdef block]

hv_apicid_to_vp_index() currently uses the per-cpu hypercall input and output pages. This function is called when running in VTL2 and when running in an SEV-SNP CoCo VM with no paravisor. In the former case, the output page is allocated, but in the latter case it is not, so the hypervisor stores the output VP index in memory that has not been allocated by the guest. Fix this by using the input page for both input and output. The hypercall has very small input and output, so sharing the same page for both is straightforward. An alternative fix is to allocate the per-cpu output page when running in an SEV-SNP CoCo guest, but this uses significantly more memory, particularly with larger vCPUs counts.

We use platform lookup tables for shared GPIOs. This mechanism needs a struct device to work so fwnode-based lookup with no device associated currently causes a NULL-pointer deference. Adding this functionality will require a larger rework and it has never worked anyway so add a check that allows the kernel to gracefully handle this situation with an error ahead of implementing that functionality as a new feature.

dm cache used a rw_semaphore for background_work_lock. Write locks on rw_semaphores have strict owner semantics, but there was no guarantee that the process that locked background_work_lock was the same process that unlocked it. This can be easily seen using a kernel compiled with CONFIG_DEBUG_RWSEMS. Given a dm cache device , run: 'dmsetup suspend && dmsetup resume '. This will trigger a kernel warning: DEBUG_RWSEMS_WARN_ON((rwsem_owner(sem) != current) && !rwsem_test_oflags(sem, RWSEM_NONSPINNABLE)) triggered by cache_resume(). To fix this, switch from a rw_semaphore to a spinlock and a wait queue. dm cache already has a wait queue and associated counter, migration_wait and nr_allocated_migrations, that was getting woken up when background work was getting completed, but wasn't actually used by anything. This is replaced by the background_work queue and counter.

A task that exits cancels its tracked requests and waits until they are gone. io_uring_cancel_generic() looks everywhere a request can be, and goes to sleep if it found nothing to cancel. It is woken up each time one of the requests of the task is completed by task work or freed. When io-wq is done with a request that has a link, there is a time in which the next request of the link can't be found: io_worker_handle_work() io_assign_current_work(worker, NULL) io_wq_free_work(work) io_req_find_next() // takes the link off the request io_free_req() // task work for the task io_assign_current_work(worker, linked) io_run_task_work() cond_resched() worker->cur_work = linked It is not pending, it is not the current work of a worker, and it is not yet on the list that it is about to be parked on. If the pass that the exiting task makes because of io_free_req() falls into that time, it finds nothing and the task goes to sleep. The worker then runs the link, which IO_WQ_BIT_EXIT doesn't prevent for bounded work. If that parks the request nothing is completed or freed, and nobody wakes up the task. If it is the process that exits it can't be killed any more either. NOP, IOSQE_ASYNC | IOSQE_IO_LINK FUTEX_WAIT, FUTEX2_PRIVATE exit of the thread, or of the process, up to 63us later With four of these per round 5 to 8 of 10,000 rounds hang (4 CPUs, KVM): io_uring_cancel_generic do_exit Commit f8b632e89a10 ("io_uring: tighten task exit cancellations") is about the same gap, and makes the task look again for whatever happens before io_free_req(). Free the request once the link is the current work of the worker, or pending. Whoever is woken up by that can find the link, until it is parked where cancellations look as well. To do so split io_wq_free_work() in two.

io_ring_exit_work() asks each task which has used the ring to remove its tctx node, by task_work, and repeats that until the list is empty. io_tctx_exit_cb() doesn't touch the node while the task is in io_uring_cancel_generic(), which removes all the nodes at its end. So the work item queues the callback again as soon as it has completed, for nothing. That is more than a waste of time. task_work_run() only returns when it finds the list empty. If the rings are released from inside the cancellation, because requests of other rings held the last references to their files, up to 64 work items keep sending callbacks to the one task which is running them. It never finds the list empty, never gets back to the cancellation, the nodes stay, and so it goes on: a task with 60 pairs of rings which poll each other ran 280000 callbacks in 5 seconds, 20 seconds after exit_group(). It ends when all the senders happen to be late at the same time. Until then the task keeps a CPU and the work items busy. An exiting task can't be killed meanwhile. execve() cancels the requests in the same way, so a live task can get stuck there as well, for minutes. If it takes more than five, the work items run into the WARN_ON_ONCE(time_after(jiffies, timeout)) of io_ring_exit_work(). No privilege is needed for any of this. If the callback has met a task in cancellation, wait a little before it is sent again. Nobody has to wake the work item up for that: either the node is gone by then, or it is the turn of the next round.

io_uring_try_cancel_requests() sets ->cq_wait_nr to 1 so that the next io_req_local_work_add() wakes the cancelling task. The add that takes the count to zero is the only one that wakes, later ones see a count cq_wait_nr, 1) atomic_read(&ctx->cq_wait_nr) == 1 count taken, wake_up_state() // task is running, no-op prepare_to_wait(TASK_INTERRUPTIBLE) io_local_work_pending() == false schedule() mpscq_push(req B) cq_wait_nr <= 0, return Only the sleeping task can run req B, so it is stuck in execve() until a signal happens to wake it. An exiting task doesn't even get that. Arm the count again once the task state is set and before the work list is checked, for the rings whose local work this task is the one to run.

A push makes the node the new tail, and then links it to its predecessor. In between the queue isn't empty, but mpscq_pop() can't get at the node. The consumers deal with that by trying again, with a cond_resched() in the loop to give a preempted producer the chance to finish. cond_resched() doesn't do that if the consumer has a higher priority than the producer and both have to use the same CPU: producer (SCHED_OTHER) consumer (SCHED_FIFO), same CPU io_req_local_work_add() mpscq_push() prev = xchg(&q->tail, node) io_cqring_wait() __io_run_local_work() again: cond_resched() // nothing to do mpscq_pop() // NULL // work pending, no events yet goto again The producer can be any task that is able to queue task_work for the ring: an io-wq worker, or somebody who sends a message with IORING_OP_MSG_RING. The consumer spins, holding ->uring_lock, until the scheduler takes pity on the fair class, which takes about a second with the default settings of the fair server, and forever without it. The loops in tctx_task_work_run() and io_cancel_local_task_work() have the same problem. Don't allow preemption between the two steps. That bounds the time the window is open on the CPU of the producer by what interrupts take, and a consumer that waits for that is on another CPU. The cost is a preempt_disable()/preempt_enable() pair per push.

There is a retry loop in init_vp_index() where the CPUs from a certain node are stripped out if they have already been in the allocated cpumask or not in HK_TYPE_MANAGED_IRQ housekeeping cpumask. If there is no CPU left, the allocated cpumask is ignored and the process is retried again. However, if the HK_TYPE_MANAGED_IRQ housekeeping cpumask turns out not to contain any CPU in that particular node, that will become an infinite retry loop. This particular problem was reported by sashiko [1]. This should rarely happen, but we still need to guard against this. Fix this infinite loop problem by also skipping NUMA node that has no housekeeping CPU in the inner while loop of init_vp_index(). Also update the early abort check to check for the absence of online housekeeping CPUs instead of just the emptiness of the cpumask. As the outer for loop will only be reached if the housekeeping cpumask has at least one online CPU, a NUMA node with housekeeping CPUs will eventually be found.

There have been various bug reports related to SMP systems hanging on boot. The reports vary but seem to have some overlap with regard to missed IPI's during SMP setup. The reports also show that reverting NBCON console usage allows the bugs to disappear. The NBCON kthreads for console printing are started very early during boot using early_initcall(). Since these kthreads are asynchronously enabling/disabling interrupts during early boot, they may be interfering with early SMP setup on some hardware platforms. Starting the console printing kthreads later has shown to avoid the hanging on systems that could previously reproduce the issue. Postpone kthread starting until after all arch initcalls have completed.

When the init of a pid namespace exits, zap_pid_ns_processes() waits until all other pids in the namespace are freed. A tracer outside the namespace that does not reap its traced zombies keeps those pids allocated, so this wait has no upper bound, and while it lasts every suspend attempt fails: Freezing user space processes failed after 20.006 seconds (2 tasks refusing to freeze, wq_busy=0): Neither wait is freezable. The first loop sleeps in do_wait(), which returns -ERESTARTSYS on the freezer's fake signal; the loop then clears TIF_SIGPENDING and waits again without ever freezing. The final loop sleeps in TASK_INTERRUPTIBLE, and since nothing clears TIF_SIGPENDING there, schedule() returns immediately after the first freeze attempt and the task spins until the wait ends. Sleep in TASK_INTERRUPTIBLE | TASK_FREEZABLE in do_wait(), so that the freezer can freeze tasks waiting there directly, and in TASK_IDLE | TASK_FREEZABLE in the final loop, like coredump_task_exit() does. This was hit with Chromium-based browsers: crashpad ptrace-attaches the threads of a crashed renderer, which is a pid namespace init, and does not reap those killed during the dump.

The devres conversion replaced manual pm_runtime_enable()/disable() with devm_pm_runtime_set_active_enabled() and dropped the remove-time runtime PM teardown. The managed release tears runtime PM down in the wrong order: it calls pm_runtime_dont_use_autosuspend() before pm_runtime_disable(), i.e. while runtime PM is still enabled, and devres is LIFO so the devm_clk_get_enabled() release runs afterwards. At remove(), pm_runtime_put_sync() leaves the device active with the autosuspend timer armed. Clearing use_autosuspend then makes rpm_idle() suspend immediately, and xiic_i2c_runtime_suspend() clk_disable()s the clock. The later devm_clk_get_enabled() release clk_disable_unprepare()s the already-disabled clock, so clk_core_disable() WARNs ("clkN already disabled") on every teardown. Drop the managed helper and restore the non-managed runtime PM setup and teardown, so runtime PM is enabled once in probe and disabled once in remove and the clock enable count stays balanced. Order the remove teardown so runtime PM is disabled before the usage count is dropped: pm_runtime_disable() followed by pm_runtime_put_noidle() instead of pm_runtime_put_sync(). Dropping the count first with autosuspend still enabled arms the autosuspend timer, and a preemption before pm_runtime_disable() lets it expire, suspend the device and clk_disable() the clock, which the devm release then disables again - the same WARN. Restore the non-managed IRQ request as well. The probe error paths now unwind runtime PM by hand via goto, and a devm-registered handler is only freed later during devres unwind, so it would stay live across the manual teardown (and across any future failing step added after it). Request the IRQ with request_threaded_irq() and free it explicitly in the probe error path and in remove. The probe error path also clears use_autosuspend, to match the teardown in remove().

If trace buffer allocation fails during early boot (e.g. in early_trace_init() -> allocate_trace_buffers()), tracing_disabled is left set to 1 and the global trace array events list remains uninitialized. Later during boot, tracer_init_tracefs() (invoked via fs_initcall) queues tracerfs_init_work asynchronously without checking if tracing was disabled. When tracer_init_tracefs_work_func() runs, it calls event_trace_init() and init_tracer_tracefs(&global_trace, NULL) unconditionally. This leads to __find_event_file() dereferencing an uninitialized tr->events list, causing an early boot NULL pointer dereference Oops: BUG: Kernel NULL pointer dereference on read at 0x00000010 Faulting instruction address: 0xc000000000485be0 Oops: Kernel access of bad area, sig: 7 [#1] ... NIP [c000000000485be0] __find_event_file+0x70/0x3c0 LR [c000000000448874] init_tracer_tracefs+0x274/0xc80 Call Trace: init_tracer_tracefs+0x274/0xc80 tracer_init_tracefs_work_func+0x50/0x320 process_one_work+0x1e8/0x5c0 worker_thread+0x1dc/0x3d0 kthread+0x194/0x1b0 start_kernel_thread+0x14/0x18 Check tracing_disabled at the beginning of tracer_init_tracefs_work_func() and return immediately if tracing is disabled or failed to initialize, preventing the crash.

The tipc_node_reset_links trace event asks tipc_node_dump() to walk the node's link entries. Unlike the other node events that request link data, this event runs without the node lock. This permits bearer teardown to free a link while the trace callback is dumping it: CPU 0 CPU 1 trace_tipc_node_reset_links() tipc_node_dump() l = n->links[0].link tipc_node_write_lock() kfree(l) n->links[0].link = NULL tipc_node_write_unlock() tipc_link_dump(l) tipc_link_dump() then dereferences the stale pointer. KASAN reported: BUG: KASAN: slab-use-after-free in tipc_link_dump Read of size 4 by task poc/115 Call Trace: tipc_link_dump+0x10cb/0x16b0 tipc_node_dump+0x4bb/0x740 trace_event_raw_event_tipc_node_class+0x258/0x360 tipc_node_reset_links+0x14d/0x1a0 tipc_rcv+0x13f5/0x3030 tipc_udp_recv+0x4e3/0x670 Allocated by task 0: tipc_link_create+0x1e1/0x1020 tipc_node_check_dest+0x7d2/0x11a0 tipc_disc_rcv+0xdbf/0x1430 Freed by task 89: kfree+0x131/0x3c0 tipc_node_link_down+0x267/0x4b0 tipc_node_delete_links+0xec/0x160 bearer_disable+0x107/0x260 Take the node write lock around the trace event. This serializes the dump against tipc_node_link_down(delete=true), which frees the link under the same write lock.

The restrict_link check in key_reject_and_link() is done before taking the keyring semaphore, racing with a concurrent keyring_restrict() that installs the restriction while holding the semaphore in write mode. Move the check to after __key_link_lock(), so it is done under the semaphore, matching the pattern used in __key_create_or_update().

Since commit 7fea700e04bd ("zap_pid_ns_processes: clear TIF_NOTIFY_SIGNAL along with TIF_SIGPENDING") zap_pid_ns_processes() clears TIF_NOTIFY_SIGNAL before every kernel_wait4() call. Otherwise kernel_wait4() doesn't sleep and returns -ERESTARTSYS if signal_pending() is true and the reaper busy-waits. But clearing TIF_NOTIFY_SIGNAL doesn't fix it. It can get raised again. For example, klp_send_signals() sets it raw against any task that hasn't transitioned yet. And it keeps sending it until the livepatch finishes. The sleep loop waiting for pid_allocated to drop has the same problem and it doesn't even clear anything. So a schedule() in TASK_INTERRUPTIBLE returns immediately. Use a no_notify_signal guard. signal_pending() will ignore TIF_NOTIFY_SIGNAL and the pid namespace reaper sleeps until a child exits or free_pid() wakes it.

io_wq_exit_workers() cancels task work for queued worker creation with task_work_cancel_match(). That has a plain load of task->task_works. io_queue_worker_create() adds the new entry and tests IO_WQ_BIT_EXIT to cancel it if the workqueue is already on its way out. That must be ordered. task_work_add() has a full barrier via cmpxchg(). But io_wq_exit_start() sets the bit with set_bit() which doesn't have any memory ordering. So afaict, on weakly ordered architectures the exiting task may load task_works before the store of the bit is visible. The other side tests the bit before the store is visible as well. So the entry remains queued with worker_refs and the exiting task waits on worker_done indefinitely. If the task still has rings then io_uring_del_tctx_node() provides the barrier via test_and_set_bit() in io_wq_set_exit_on_idle(). When it has closed all rings though that barrier is gone. Add the barrier after the set_bit().

The fallocate implementation only takes invalidate_lock for punch hole, collapse range, and insert range operations. For standard allocation modes (mode == 0, FALLOC_FL_KEEP_SIZE), the lock is not held. During ntfs_attr_fallocate(), new clusters are mapped to the runlist via ntfs_attr_map_cluster() before being zeroed by ntfs_dio_zero_range(). This creates a window where concurrent mmap page faults can read uninitialized disk data. Since mmap uses filemap_fault() which takes invalidate_lock in shared mode, it can fault in pages during this window and expose old disk contents to userspace. This is an information leak and data integrity issue. Fix by taking invalidate_lock for all fallocate operations, not just for punch/collapse/insert modes. This prevents concurrent page faults from accessing unzeroed clusters during the allocation window.

ntfs_filemap_page_mkwrite() calls iomap_page_mkwrite() without holding mapping->invalidate_lock, so a concurrent truncate or fallocate can be in the middle of invalidating pagecache and rewriting the runlist while the write fault maps blocks and dirties the folio. This races with ntfs_attr_fallocate(), which merges clusters into the in-memory runlist, drops the runlist lock, and only afterwards zeroes the newly allocated clusters on disk; and with the punch-hole/insert/collapse paths that free clusters after truncating the cache. Per Documentation/filesystems/locking.rst, ->page_mkwrite() must ensure there are no truncate/invalidate races, "usually mapping->invalidate_lock is suitable for proper serialization". xfs takes its mmaplock (= the invalidate_lock rwsem) shared in exactly this path. Take invalidate_lock shared around iomap_page_mkwrite(). The read-only fault path is already covered because filemap_fault() itself grabs invalidate_lock shared on instantiation/read paths; only page_mkwrite was bypassing it in this driver.

Patch series "mm/damon: move zero damos quota target_value handling to the core layer". Having zero DAMOS quota target value can cause division by zero. DAMON API callers are checking the target value parameters to avoid that. It is easy to make mistakes in some of the multiple API callers. Move that to the core layer. Patch 1 adds the corner case handling into the core layer DAMON parameters validation logic. Patches 2 and 3 remove no more needed DAMON API callers side handling of the corner case in DAMON_LRU_SORT and DMON_SAMPLIE_MTIER, respectively. This patch (of 3): If a DAMOS scheme has a damos_quota_goal of zero target_value, damos_quota_goal() could trigger division-by-zero error. Hence each DAMON API callers should do the zero target_value validation. It is easy to make mistakes. Actually such bugs in DAMON_LRU_SORT and DAMON_SAMPLE_MTIER were found and fixed [1]. It is better to handle the corner case only once in the core layer, instead of multiple places in all DAMON API callers. One straightforward option is using an alternative denominator for the corner case in the damos_quota_goal(). However, the zero target_value is meaningless. In this case, the quota goal is always evaluated as achieved or over-achieved. The quota will only keep being reduced. Simply avoid using zero target_value by adding a check in the core layer DAMOS quota goal parameters validation/commit path, damos_commit_quota_goal(). Update it to return an error in the case. Also update its caller to propagate the error.

l2cap_chan may be linked to l2cap_conn at most once. This is assumed in several places, eg l2cap_chan_del cleanup. There is a TOCTOU race where the invariant is violated: [Task 1] [Task 2] l2cap_chan_connect l2cap_sock_bind l2cap_chan_lock lock_sock l2cap_state_change if (sk->sk_state != BT_OPEN) chan->state = BT_CONNECT l2cap_sock_state_change_cb chan->state = BT_BOUND sk->sk_state = BT_BOUND lock_sock sk_state = BT_CONNECT l2cap_sock_connect() does not check sk->sk_state, so since chan->state is now BT_BOUND, subsequent connect() ends up with second __l2cap_chan_add. Explicitly document and check the invariant in __l2cap_chan_add with WARN_ON_ONCE. The only callsite where it could be hit is l2cap_chan_connect, so add pre-check there to avoid relying on chan->state. chan->state read/write is not properly guarded currently so there can be other TOCTOUC problems. Add l2cap_lock_chan in l2cap_sock_bind() to guard chan->state write.

trace_module_record() walks ftrace_trace_arrays from the module notifier without trace_types_lock. A concurrent rmdir of an instance removes it from the list and frees it under that lock, so a module load or unload can walk into a freed trace_array.

update_last_data_if_empty() is called from tracing_open() with O_TRUNC, tracing_read_pipe() and tracing_buffers_read() without trace_types_lock. Two readers can both pass the LAST_BOOT test and free module_delta twice, and a module notifier can update module_delta while it is being freed. Take trace_types_lock in those paths and keep the callers that already hold it on an unlocked variant.

A non-empty expression normally finishes in predicate_parse()'s inner loop, which resolves the pending logical operators before jumping to finalization. If the input ends after an operator, possibly followed by unary '!' tokens, the outer loop instead falls through with unresolved operator state. For example, "common_pid == 1 || common_pid == 2 &&" leaves prog[0].target equal to -1. The second pass then reads prog[-1] before the forward-jump check can reject it. Depending on adjacent slab contents, it can also copy an out-of-bounds word into a branch target that passes the forward-only check. A self-targeted perf tracepoint counting event can reach this path through PERF_EVENT_IOC_SET_FILTER. Reject the outer-loop fallthrough when a predicate has already been parsed. This ensures that non-empty expressions only reach finalization after the inner loop has resolved their operators. Keep the existing no-filter error for input that has no predicates. Add a regression test for trailing logical operators and unary negations.

create_field_var_hist() creates a histogram on the matched event to supply a field variable, and records it on the target so it can be torn down later: hist_data = find_compatible_hist(target_hist_data, file); ... /* Save the compatible histogram information */ var_hist->hist_data = hist_data; hist_data is not the histogram it creates. It is one that was already there, whose keys the new one copies. The saved pointer has a single user, unregister_field_var_hists(), which runs when the target is removed: file = hist_data->field_var_hists[i]->hist_data->event_file; find_compatible_hist() matches on keys alone, so it can pick one with no variables of its own. check_var_refs() then returns false and the user can remove it while the target still points at it: BUG: KASAN: slab-use-after-free in event_hist_trigger_free+0x2b2/0x320 Read of size 8 at addr ffff8880093b90e0 by task init/1 Freed by task 1: kfree+0x154/0x420 event_hist_trigger_free+0x1d5/0x320 event_hist_trigger_parse+0x35f3/0x69e0 trigger_process_regex+0x1a6/0x250 Save the event file instead. It is all the dereference ever wanted, and it does not go away when the user removes a trigger.

An action that names a field on another event needs that field as a variable, so create_field_var_hist() registers a second hist trigger on that event to provide it, and records it in target_hist_data->field_var_hists[]. If a later part of the same command fails, create_actions() returns an error and event_hist_trigger_parse() jumps to out_free. The write fails with -EINVAL and sched_switch gets no trigger, but: # echo 'hist:keys=pid:ts0=common_timestamp.usecs' >> \ events/sched/sched_waking/trigger # echo 'my_synth u64 lat; int a; int b' >> synthetic_events # echo 'hist:keys=next_pid:wakeup_lat=common_timestamp.usecs-$ts0:\ onmatch(sched.sched_waking).my_synth($wakeup_lat,prio,nosuchfld)'\ >> events/sched/sched_switch/trigger # cat events/sched/sched_waking/trigger hist:keys=pid:vals=hitcount:ts0=common_timestamp.usecs:sort=hitcount:size=2048:clock=global [active] hist:keys=pid:vals=hitcount:synthetic_prio=prio:sort=hitcount:size=2048 [active] The second histogram is the one created for 'prio'. It is still active and still counting. out_free reaches destroy_field_var_hists(), which frees the field_var_hists[] entries, but never unregister_field_var_hists(), which is what removes the triggers and runs only from event_hist_trigger_free(). The entry is freed and the trigger is left with nothing tracking it. Unregister the field variable histograms in out_free as well, and record each one as soon as its trigger is registered rather than after the lookup that follows it, so no exit can leave a registered trigger untracked. The entry then belongs to field_var_hists[] and is released by destroy_field_var_hists(), so freeing it on that exit would be a double free.

In event_hist_trigger_parse() and hist_unreg_all(), the histogram registration and unregistration paths look up find_synth_event() using trace_event_name(file->event_call) and adjust se->ref if a match is found. Because find_synth_event() only compares the event basename, registering or removing a histogram on a non-synthetic trace event whose basename matches a synthetic event erroneously modifies the synthetic event's reference count. If a histogram is registered on the non-synthetic event before the synthetic event is created and removed afterwards, se->ref is decremented without a matching increment, allowing the synthetic event to be freed while an active histogram action still holds a pointer to it. Introduce find_synth_event_call() to verify that &se->call matches file->event_call before adjusting se->ref.

syscall_copy_user_array() copies the buffer argument of faultable syscalls (write, pwrite64, ...) with __copy_from_user(), which skips access_ok(). Use copy_from_user() instead so a non-user address is rejected and the entry is recorded as faulted. It could be a security concern for machines that trust perf a bit too much.

io_waitid_remove_wq() only takes the wait queue lock if ->head is set. io_waitid_wait() clears ->head first thing, before it tries to become the owner of the request by bumping ->refs. If somebody else owns the request at that point, nothing keeps it from completing the request while the wake up handler is still going to touch it: CPU0 (io_waitid_wait) CPU1 (io_waitid, holds a ref) list_del_init(&wait->entry) smp_store_release(&iw->head, NULL) __do_wait() finds another child io_waitid_remove_wq() // ->head == NULL, no locking io_waitid_finish(), IOU_COMPLETE // request freed and reused atomic_fetch_inc(&iw->refs) The same goes for __io_waitid_cancel() if it claims the request in that window. ->refs is part of the per opcode data of the request, so this increments whatever the next user of the request keeps there. If that happens to be zero, the handler concludes that it owns the request, and queues io_waitid_cb() for it. With the window held open by a delay, and the request reused for a receive that is waiting for data: BUG: KASAN: slab-use-after-free in __wake_up_common+0x1ca/0x1f0 __wake_up_sync_key+0x36/0x50 sock_def_readable+0xf1/0x420 unix_stream_sendmsg+0x60b/0x1110 Allocated by task 87: io_arm_apoll+0x1ed/0x940 Freed by task 87: __io_submit_flush_completions+0x118d/0x1a10 tctx_task_work_run+0x327/0x950 Only clear ->head once the handler owns the request. If it doesn't, the owner finds ->head set and has to wait for the lock, which is held by the caller of the handler. Deleting the already deleted entry again is harmless.

io_uring_cmd_del_cancelable() uses hlist_del() to take a command off ->cancelable_uring_cmd, which leaves LIST_POISON1/2 behind in ->hash_node. Requests are recycled through the request cache without being reinitialised, and not every user of ->hash_node sets it up before it is used: If a vectored futex wait is woken while futex_wait_multiple_setup() is still queueing the futexes, io_futexv_wait() doesn't add the request to ->futex_list, and leaves the completion to the task_work queued by the wake up. io_futexv_complete() ends in __io_futex_complete(), whose hlist_del_init() is fine for a request that is unhashed, which is how all other users of the node leave it. If the request was a cancelable uring_cmd in its previous life, e.g. a ublk FETCH_REQ, it writes through LIST_POISON2: Oops: general protection fault, probably for non-canonical address KASAN: maybe wild-memory-access in range [0xdead000000000120-...] RIP: 0010:io_futexv_complete+0x219/0x420 Call Trace: tctx_task_work_run+0x804/0x950 tctx_task_work+0x71/0xa0 task_work_run+0x118/0x200 Leave the node unhashed.

As long as IO_CHECK_CQ_OVERFLOW_BIT is set, io_cqring_wait_schedule() returns right away, before it gets to look at signals or arms the timeout, for io_cqring_wait() to flush the overflow list. If that doesn't empty the list, and the CQ ring doesn't hold the number of events the caller wants, the loop starts over, with the same result. The task spins in the kernel, and can't be killed. The assumption is that this can't happen: if CQEs can't be flushed, the ring is full, and min_events is capped to the size of the ring. But: - With IORING_SETUP_CQE_MIXED a 32b CQE needs two contiguous entries. It is stuck on the list if a single one is free, and then the ring holds less than cq_entries events. A single threaded application that asks for as many events as the ring has entries never gets out of io_uring_enter(). - Whether the ring is full is up to the CQ head, which the application can write to. While one thread waits for events on an empty ring, another one can move the head back so that the ring looks full and have a CQE overflow. The waiter can't be stopped any more, not even by SIGKILL. Only the application can make room in the CQ ring. Return to it if the list could not be flushed.

When a thread exits, only its tracked requests are cancelled. Then its io-wq is torn down. io_wq_exit_workers() signals every worker, to get it out of an interruptible sleep, and waits for all of them to exit. Before they do, they go through the work that is left. Unbounded work is cancelled. Bounded work is run, as it is expected to end by itself. Not all of it does. An open has no file yet, so it counts as bounded, and the open of a FIFO waits for the other end. That is what the signal is for. But the workers are signalled once, and a worker that hasn't started the work by then loses the signal: io_wq_exit_workers() io_worker_handle_work() io_wq_worker_wake() set TIF_NOTIFY_SIGNAL io_assign_current_work() io_run_task_work() clear TIF_NOTIFY_SIGNAL io_wq_submit_work() fifo_open() wait_for_partner() wait_for_completion_timeout() The thread never gets out of do_exit(): IORING_OP_OPENAT of a FIFO without a writer, IOSQE_ASYNC exit of the thread that has submitted it hangs in more than 9 of 10 attempts (4 CPUs, KVM), until the process is killed. Set TIF_NOTIFY_SIGNAL for bounded work that is started with IO_WQ_BIT_EXIT set. The bit is set before the workers are signalled, and it is tested after the flag has been cleared, with clear_notify_signal() providing the barrier. So either the flag stays, or the bit is seen. Work that is started while the io-wq exits is treated like work that is being run at that time: it isn't interrupted unless it waits interruptibly.

io_waitid packs the cancellation state into bit 31 of ->refs and uses bits 30:0 as a 31-bit reference counter: #define IO_WAITID_CANCEL_FLAG BIT(31) #define IO_WAITID_REF_MASK GENMASK(30, 0) __io_waitid_cancel() marks a request canceled, then increments the low bits of ->refs to claim completion ownership. When another path owns the request, the increment remains even though cancellation returns false. The request remains discoverable by cancellation, so repeated failed attempts can wrap the 31-bit reference field to zero. A later cancellation can claim ownership and complete the request a second time. Use atomic_fetch_or() when setting the cancel flag and inspect the old flag value. The first cancel attempt continues to the existing ownership increment, preserving its wake and issuer handoff semantics. Later attempts see the cancel flag and return without incrementing the reference field. This makes repeated failed cancellation attempts idempotent while leaving the first-cancel state transitions unchanged.

Named keyring lookup must keep the storage containing the owner UID alive through the namespace mapping check. find_keyring_by_name() reads keyring->user->uid under keyring_name_lock, but that lock protects the keyring's name entry and allocation, not its separate key_user. A named session-keyring join can overlap a privileged KEYCTL_CHOWN on another CPU. If the keyring holds the last reference to its old dynamic key_user and the ownership transfer succeeds, the following ordering is possible: Named join Chown find_keyring_by_name() keyctl_chown_key() read_lock(keyring_name_lock) down_write(key->sem) load old keyring->user replace key->user and key->uid up_write(key->sem) key_put(key) key_user_put(old user): free read old user->uid read_unlock(keyring_name_lock) Chown neither takes keyring_name_lock nor key_session_mutex, so it can free the old account between the pointer load and the UID read. The lookup then reads freed memory even though the keyring itself is alive. Use keyring->uid for the mapping check. key_alloc() initializes this inline owner UID and keyctl_chown_key() updates it on chown. The existing name lock keeps its containing keyring allocated throughout the check, so lookup no longer depends on the account's lifetime. This also matches the owner UID used by the permission check. KASAN report as below: BUG: KASAN: slab-use-after-free in find_keyring_by_name+0x577/0x5d0 Read of size 4 at addr ffff8881128ae2ec by task keycase/500

An authorization payload and its saved credentials must remain valid while search_process_keyrings_rcu() searches the requestor's keyrings. The function validates the authorization key, then separately loads payload.data[0] and unconditionally reads rka->cred. Validation does not prevent request_key_auth_revoke() from clearing the payload. A request-key helper can assume authority and fork a descendant that inherits the authorization key. If the descendant misses in its own keyrings, it falls back to the saved requestor credentials. Successful target instantiation invalidates the authorization key without clearing the payload. After the helper exits, UMH_WAIT_PROC returns to the requester, which calls complete_request_key() and revokes that key. With the descendant paused after validation, the paths can interleave as follows: 1. The descendant enters search_process_keyrings_rcu() under RCU and key_validate() succeeds. 2. The helper instantiates the target, invalidates the authorization key, and exits. 3. The requester resumes from UMH_WAIT_PROC, calls complete_request_key(), and revokes the authorization key. request_key_auth_revoke() sets its payload pointer to NULL. 4. The descendant loads NULL and dereferences rka->cred. This NULL dereference can oops the kernel. RCU delays payload disposal but does not prevent the pointer from being cleared. Load the payload with dereference_key_rcu() and require a non-NULL result before validating the authorization key and searching its saved credentials. A non-NULL payload and its credentials remain alive for the caller's RCU read-side section; a cleared payload skips the fallback search and follows the existing error selection. KASAN report as below: Oops: general protection fault, probably for non-canonical address 0xdffffc0000000005: 0000 [#1] SMP KASAN NOPTI KASAN: null-ptr-deref in range [0x0000000000000028-0x000000000000002f] RIP: 0010:search_process_keyrings_rcu+0x32b/0x790

keyctl_describe_key() must keep the key type's name alive until kasprintf() has consumed it. The retained key reference keeps the key and its description alive, but does not pin the module providing its type. The name is passed to kasprintf() without holding key->sem. kvasprintf() uses the saved name pointer in two formatting passes, with a GFP_KERNEL allocation between them. With AF_RXRPC=m, a task that passes View permission before type retirement can overlap a privileged module unload in this order: 1. KEYCTL_DESCRIBE looks up the key, saves key->type->name in the kasprintf() arguments and completes the first formatting pass. 2. While the formatter is delayed before its second pass, af_rxrpc_exit() calls unregister_key_type(), which removes the type from the registry and waits for key_gc_keytype(). 3. The collector takes key->sem for writing, retypes the retained key to key_type_dead and completes retirement. Unregistration and module exit return, allowing free_module() to release the type and name. 4. The formatter resumes its second pass and reads the saved name from released module storage, which can fault in string(). Hold key->sem for reading across kasprintf(). The collector then cannot retype the key or complete retirement until both formatting passes have finished. If retirement wins the lock first, formatting uses the core key_type_dead name instead. Release the semaphore immediately after kasprintf(), including on allocation failure, before copying to userspace. Oops report as below: BUG: unable to handle page fault for address: fffffbfff8054794 #PF: supervisor read access in kernel mode #PF: error_code(0x0000) - not-present page RIP: 0010:string+0x23f/0x470 Call Trace: vsnprintf+0x330/0x1110

KEYCTL_READ returns the serial numbers in a keyring in associative-array traversal order. The first word of the index is a public, deterministic hash of the type pointer, domain pointer and description. An unprivileged user can therefore insert keys with known descriptions, predict their order for each possible KASLR slide, and recover the kernel image address. On an x86-64 build with 16 MiB alignment, 45 user keys suffice to distinguish the candidate slides. Remove the addresses from the whole index, not just its first hash word. Assign each key type and domain a nonzero 64-bit ID and hash those IDs and the description with a once-initialized SipHash key. Use the full IDs in the later index chunks as well. Truncating or hashing the IDs there would lose the guarantee that distinct type/domain/description tuples have a differing index bit. Allocate IDs lazily since private key types need not be registered. Cache them in the index key so tree walks and copies of an existing index do not need to dereference a type descriptor that may have been unregistered. Keep the 32-bit chunk layout and diff_objects() bit offsets consistent, and retain the separate root slot for keyrings used by recursive search. The key comparison, permissions and userspace data format are unchanged; only the address-dependent in-memory ordering is replaced.

keyctl_read_key() checks key->type->read without holding key->sem. A reference keeps the key alive, but does not pin the module containing its type. A surviving RxRPC key can therefore be read while AF_RXRPC is being unloaded: KEYCTL_READ task Module unload and key GC ---------------- ------------------------ lookup_user_key() permission / possession checks load key->type af_rxrpc_exit() unregister_key_type() GC: down_write(key->sem) GC: retype to key_type_dead GC: up_write(key->sem) unregister_key_type() returns free_module() read saved type's read member The final access can read freed module storage. The read invocation in __keyctl_read_key() is already protected by key->sem, but the earlier method check is outside that critical section. Move the method check into __keyctl_read_key(), after key_validate() and under its existing read hold on key->sem. The check and invocation then use the same protected type for length queries and every read retry. The GC cannot retype the key until the semaphore is released, and type unregistration waits for that GC pass before the module can be freed. Return -EOPNOTSUPP when validation succeeds and the type has no read method. A temporary buffer may now be allocated before this result; allocation and copying to userspace remain outside the semaphore.

trusted_tpm_unseal() proceeds to pcrlock() when the TPM unseal operation fails. If pcrlock() succeeds, its return value overwrites the unseal error, causing key instantiation to succeed. Return immediately when unseal fails to preserve the original error.

The PKCS#8 parser deliberately leaves the asym_key_ids payload slot empty. key_or_keyring_common() unconditionally dereferences that slot when a PKCS#8 key is supplied to a key_or_keyring restriction. Both the plain and :chain forms reach this branch, allowing an unprivileged caller to trigger a NULL pointer dereference through KEYCTL_RESTRICT_KEYRING followed by add_key(). Oops: general protection fault, probably for non-canonical address 0xdffffc0000000000 KASAN: null-ptr-deref in range [0x0000000000000000-0x0000000000000007] RIP: 0010:key_or_keyring_common (crypto/asymmetric_keys/restrict.c:205 crypto/asymmetric_keys/restrict.c:279) Call Trace: __key_create_or_update (security/keys/key.c:884) key_create_or_update (security/keys/key.c:1021) __do_sys_add_key (security/keys/keyctl.c:134) do_syscall_64 (arch/x86/entry/common.c:52 arch/x86/entry/common.c:83) entry_SYSCALL_64_after_hwframe (arch/x86/entry/entry_64.S:130) Kernel panic - not syncing: Fatal exception Reject an asymmetric restriction key without key IDs. A PKCS#8 private key cannot identify the signer of a candidate certificate and is not a valid trust anchor.

print_rec() leaves 'modname' and 'offset' uninitialized and kallsyms_lookup() only sets them on success. When the lookup fails like when a module is being loaded or unloaded, 'modname' holds a garbage value, passes the 'if (modname)' check, and is dereferenced by seq_printf(). Oops: general protection fault RIP: 0010:string+0x48/0xe0 Call Trace: vsnprintf+0x1f5/0x650 seq_printf+0x8e/0xb0 print_rec+0x6c/0xc0 t_show+0x4c/0x440 seq_read_iter+0x119/0x480 This can be triggered by reading available_filter_functions while a module is being loaded or unloaded. Initialize both variables so a failed lookup skips the module name.

coredump_finish() releases the parked threads by clearing ->task in their core_thread entry and calling wake_up_process() on each of them. It does that without holding a reference on the task and without being inside rcu. But calling wake_up_process(task) without rcu here isn't safe. A parked thread doesn't need that wakeup to leave. A spurious wakeup or a preemption after the store is enough for the task to go away. So if the coredump client is preempted between the store and wake_up_process() the thread can exit and be freed in the meantime and try_to_wake_up() takes pi_lock in freed memory. Hold rcu across the loop.

Reorder copy_process() cleanup so sched_fork() taking scx_fork_rwsem scx_pre_fork() is safe and isn't held around exiting files. Right now, a fork that fails while another thread closed the bpf link fd will deadlock against its own read side. It also blocks every fork on the system: copy_process() holds scx_fork_rwsem for read exit_files() close_files() filp_close_sync() bpf_scx_unreg() kthread_flush_work() waits for the disable work scx_root_disable() percpu_down_write(&scx_fork_rwsem) Simply release the child's files after sched_cancel_fork() dropped the lock. The task starts out as a copy of its parent so p->files points to the parent's table until copy_files() replaces it. exit_files() must not run for a fork that failed before that. Clear p->files up front so exit_files() is a no-op for those and make copy_files() set it explicitly for CLONE_FILES.

In two places, current code forces vmbus_connection.conn_state to DISCONNECTED prior to calling vmbus_disconnect(). But vmbus_disconnect() calls vmbus_initiate_unload(), which unloads the VMBus connection only if the conn_state is CONNECTED. Consequently, the connection remains and the Hyper-V host might continue to send messages or post VMBus interrupts. The problem was introduced in commit 74347a99e73ae ("x86/Hyper-V: Unload vmbus channel in hv panic callback"), in that vmbus_initiate_unload() started checking the current connection state before doing the unload. This commit removed some occurrences where conn_state was forced to DISCONNECTED, but these two remained, apparently due to an oversight at the time. Fix this by not forcing conn_state to DISCONNECTED during VMBus exit, or if a failure occurs when establishing the connection in the first place. Whatever value conn_state has at the time is then read by vmbus_initiate_unload() to determine if there is a connection to be unloaded.

Once a power supply is registered, its callbacks can immediately start firing from other contexts, such as external_power_changed() triggered by the TCPM stack. If a power supply is unregistered while the callback is still running, the driver data can already be freed when the callback tries to access it, leading to a use-after-free. This happens e.g. when the hardware bus carrying the power supply device malfunctions (e.g. I2C is hogged down by another malfunctioning device) immediately after the power supply is registered, and thus the core is still processing the callbacks which were queued up when the driver starts the removal, leading in some cases to a kernel crash, e.g.: [ 11.645942] Unable to handle kernel NULL pointer dereference at virtual address 0000000000000005 [ 11.646751] Mem abort info: [ 11.647006] ESR = 0x0000000096000004 [ 11.647338] EC = 0x25: DABT (current EL), IL = 32 bits [ 11.647806] SET = 0, FnV = 0 [ 11.648077] EA = 0, S1PTW = 0 [ 11.648356] FSC = 0x04: level 0 translation fault [ 11.648785] Data abort info: [ 11.649041] ISV = 0, ISS = 0x00000004, ISS2 = 0x00000000 [ 11.649524] CM = 0, WnR = 0, TnD = 0, TagAccess = 0 [ 11.649981] GCS = 0, Overlay = 0, DirtyBit = 0 [ 11.650390] [0000000000000005] user address but active_mm is swapper [ 11.650955] Internal error: Oops: 0000000096000004 [#1] SMP [ 11.651460] Modules linked in: [ 11.651742] CPU: 1 UID: 0 PID: 144 Comm: kworker/1:2 Not tainted 7.2.0-rc6-g62a9297af2cd #1 PREEMPT [ 11.652553] Hardware name: Flipper One rev. F0B1C2 (DT) [ 11.653024] Workqueue: events power_supply_changed_work [ 11.653511] pstate: 60000005 (nZCv daif -PAN -UAO -TCO -DIT -SSBS BTYPE=--) [ 11.654135] pc : __power_supply_is_supplied_by+0x18/0x100 [ 11.654624] lr : __power_supply_am_i_supplied+0x40/0xb8 [ 11.655098] sp : ffff80008192bb30 [ 11.655399] x29: ffff80008192bb30 x28: 0000000000000000 x27: 0000000000000000 [ 11.656049] x26: 0000000000000000 x25: 0000000000000000 x24: 0000000000000000 [ 11.656695] x23: ffff0000c19f4200 x22: ffffdb2232ba4ea8 x21: ffff80008192bc28 [ 11.657344] x20: ffff0000c1eef000 x19: ffff80008192bc18 x18: 00000000a0886e62 [ 11.657991] x17: 000000040044ffff x16: 04500072b5503510 x15: 0000000000000000 [ 11.658639] x14: 0000000000000000 x13: 0000000000000220 x12: 0000000000000000 [ 11.659286] x11: 0000000000000000 x10: ffff0000c1fdb2b0 x9 : ffffdb2232ba5590 [ 11.659934] x8 : 00000000e5b906e6 x7 : ffff0000c2502778 x6 : ffffdb22339793d0 [ 11.660581] x5 : ffff80008192bc18 x4 : ffff0000c19cbca0 x3 : 0000000000000000 [ 11.661228] x2 : ffff0000c1fdaf40 x1 : ffffffffffffffed x0 : ffff0000c1eef000 [ 11.661878] Call trace: [ 11.662103] __power_supply_is_supplied_by+0x18/0x100 (P) [ 11.662596] __power_supply_am_i_supplied+0x40/0xb8 [ 11.663040] psy_for_each_psy_cb+0x20/0x40 [ 11.663416] class_for_each_device+0x110/0x150 [ 11.663825] power_supply_am_i_supplied+0x68/0x100 [ 11.664262] bq257xx_external_power_changed+0x58/0x140 [ 11.664733] __power_supply_changed_work+0x60/0x80 [ 11.665170] psy_for_each_psy_cb+0x20/0x40 [ 11.665545] class_for_each_device+0x110/0x150 [ 11.665953] power_supply_changed_work+0x98/0x1b8 [ 11.666382] process_one_work+0x164/0x4c0 [ 11.666758] worker_thread+0x19c/0x320 [ 11.667104] kthread+0x138/0x150 [ 11.667408] ret_from_fork+0x10/0x20 [ 11.667744] Code: d503233f a9bd7bfd 910003fd a90153f3 (f9400c34) [ 11.668294] ---[ end trace 0000000000000000 ]--- Add a read-write semaphore between external_power_changed() and power_supply_unregister() to prevent the latter from returning (and thus the driver from freeing its data) while the callback is still running.

Audio can be triggered by the HPD handler and start up before video has been restored, calling vc4_hdmi_write_infoframe() for the audio infoframe. vc4_hdmi_write_infoframe() needs Packet RAM to be on to store the infoframe, however, if the video hasn't been enabled, vc4_hdmi_encoder_post_crtc_enable() hasn't been called and therefore, the Packet RAM can be disabled by the time vc4_hdmi_write_infoframe() executes, causing the following splat: [ 201.570266] ------------[ cut here ]------------ [ 201.570293] Packet RAM has to be on to store the packet. [ 201.570367] WARNING: drivers/gpu/drm/vc4/vc4_hdmi.c:683 at vc4_hdmi_write_infoframe+0x404/0x494 [vc4], CPU#0: wireplumber/1606 [...] [ 201.658650] Hardware name: Raspberry Pi 4 Model B Rev 1.5 (DT) [...] [ 201.758499] Call trace: [ 201.760967] vc4_hdmi_write_infoframe+0x404/0x494 [vc4] (P) [ 201.766607] vc4_hdmi_write_audio_infoframe+0x1c/0x28 [vc4] [ 201.772247] write_infoframe+0xbc/0x140 [drm_display_helper] [ 201.777975] drm_atomic_helper_connector_hdmi_update_audio_infoframe+0x80/0xdc [drm_display_helper] [ 201.787140] vc4_hdmi_audio_prepare+0x4ac/0x570 [vc4] [ 201.792252] drm_connector_hdmi_audio_prepare+0x24/0x30 [drm_display_helper] [ 201.799390] hdmi_codec_prepare+0x100/0x160 [snd_soc_hdmi_codec] [ 201.805471] snd_soc_pcm_dai_prepare+0xd4/0x100 [snd_soc_core] [ 201.811375] __soc_pcm_prepare+0x60/0x12c [snd_soc_core] [ 201.816752] soc_pcm_prepare+0x34/0x58 [snd_soc_core] [ 201.821863] snd_pcm_do_prepare+0x70/0x8c [snd_pcm] [ 201.826798] snd_pcm_action_nonatomic+0xe4/0x174 [snd_pcm] [ 201.832350] snd_pcm_prepare+0xc8/0xd8 [snd_pcm] [ 201.837021] snd_pcm_common_ioctl+0x458/0x614 [snd_pcm] [ 201.842308] snd_pcm_ioctl+0x30/0x48 [snd_pcm] [ 201.846802] __arm64_sys_ioctl+0x88/0xd8 [ 201.850769] invoke_syscall+0x48/0xf4 [ 201.854470] el0_svc_common+0x80/0xd8 [ 201.858171] do_el0_svc+0x1c/0x28 [ 201.861520] el0_svc+0x3c/0xb0 [ 201.864605] el0t_64_sync_handler+0x68/0xdc [ 201.868834] el0t_64_sync+0x198/0x19c [ 201.872536] ---[ end trace 0000000000000000 ]--- Considering that Packet RAM enable is essentially a power management configuration, move its enabling/disabling to the Runtime PM hooks. Note that post_crtc_disable() still stops all infoframes, but keeps Packet RAM enabled, leaving it to be disabled by runtime suspend.

Handle the TCP_ESTABLISHED -> TCP_CLOSING transition on OP_RST, which can race with the connect loop. Immediately break and return 0 on ESTABLISHED/CLOSING. The return value of connect() should reflect what happened up to the point the connection was established or failed. Events that occur afterwards must not affect it. E.g. even if OP_RW has already set sk_err, connect() should still return 0, not ENOBUFS because of it. Adapt the inaccurate comment above signal_pending(). Resetting a socket that is still present in connected_table can lead to memory corruption. The reporter noted lost transports for in-flight skbs, and I have reproduced crashes caused by re-insertion into connected_table. list_add double add: new=, prev=, next=. kernel BUG at lib/list_debug.c:35! Oops: invalid opcode: 0000 [#1] SMP KASAN NOPTI Workqueue: vsock-loopback vsock_loopback_work RIP: 0010:__list_add_valid_or_report+0x11f/0x130 Call Trace: vsock_insert_connected.cold+0xe/0x13 virtio_transport_recv_pkt+0x10e9/0x1460 vsock_loopback_work+0x305/0x480 process_one_work+0xe4c/0x1560 worker_thread+0x4f1/0xd60 kthread+0x36e/0x470 ret_from_fork+0x47b/0x6b0 ret_from_fork_asm+0x1a/0x30 This fix is supplementary to commit 002541ef650b ("vsock: Ignore signal/timeout on connect() if already established"). Details under Link.

rds_rdma_cm_event_handler_cmn() hands the return value of rdma_resolve_route() straight back to the rdma_cm on RDMA_CM_EVENT_ADDR_RESOLVED. A synchronous failure there - -ENOMEM from the route work allocation, an SA query that cannot be set up, a RoCE route with no usable device - makes addr_handler() destroy the id the event was delivered on. That id is ic->i_cm_id, and nothing clears the pointer: the connection sits in RDS_CONN_CONNECTING with a freed id until its shutdown calls rdma_disconnect() and rdma_destroy_id() on it. The ROUTE_RESOLVED case already avoids this: rds_ib_cm_initiate_connect() forces a zero return while ic->i_cm_id == cm_id, as the comment there explains. Do the same here - drop the connection and return 0 - so the id stays RDS's to destroy from the shutdown, and the reconnect gets a fresh one.

Socket filters can use SKF_NET_OFF and SKF_LL_OFF to read relative to an skb's network and MAC headers. Netlink producers may reserve headroom without resetting those headers, allowing a receiving socket's filter to read data before the initialized message. For example, IWPM HELLO replies are allocated with dev_alloc_skb(), which reserves NET_SKB_PAD bytes. An unprivileged NETLINK_RDMA user can attach a classic socket filter and disclose this headroom through the length of the delivered reply. The resets in netlink_dump() do not help because IWPM sends a separately allocated skb through netlink_unicast(). Reset both headers before running the receiver's filter in unicast and broadcast delivery, as is already done for dump skbs. This also covers other netlink producers that reserve headroom without changing their allocation or message-building code.

rocket_probe() stores the return value from rocket_device_init() in the file-scoped rdev pointer before checking whether it is an error pointer. If rocket_device_init() fails, rdev is left as an ERR_PTR. The rdev pointer is also used as the marker for whether the shared rocket device has already been initialized. A later core probe can therefore see a non-NULL rdev, skip rocket_device_init(), and dereference the error pointer when reading rdev->num_cores. Save the error code and clear rdev before returning from the rocket_device_init() failure path.

xe_evict_flags() switches on tbo->resource->mem_type to choose an eviction placement. An xe BO can reach eviction with no ttm.resource, e.g. an imported dma-buf BO whose resource was freed during teardown while its attachment is still live: the exporter then invalidates it via xe_dma_buf_move_notify() -> xe_bo_evict(), and the switch dereferences a NULL resource and faults. A BO with no resource has nothing placed or mapped, so there is nothing to evict. Return an empty placement (num_placement = 0), mirroring the handling already used for non-xe scatter-gather BOs just above; ttm_bo_validate() then takes the pipeline-gutting path and returns 0. Without the fix the xe_dma_buf@xe_test_dmabuf_move_notify_no_resource dynamic live selftest will fail with below stack. BUG: kernel NULL pointer dereference, address: 0000000000000010 RIP: 0010:xe_evict_flags+0xf4/0x140 [xe] Call Trace: xe_bo_evict+0x40/0xa0 [xe] ? ttm_bo_pipeline_gutting+0x71/0x180 [ttm] xe_test_dmabuf_move_notify_no_resource+0x182/0x440 [xe] xe_dma_buf_kunit+0x387/0x580 [xe] v2: Add Fixes tag and failure info in commit text (Matt Brost) (cherry picked from commit dbcfaac6331f1fcb0c2fe3ad3e9adb1e5d43ec0c)

ttm_bo_type_kernel BOs are not guaranteed to be zeroed on allocation and may contain data from previously freed pages. Since the OA buffer is mmap'able by userspace, this can expose stale kernel memory contents. Streams opened with DRM_XE_OA_PROPERTY_OA_DISABLED allocate the OA buffer without calling xe_oa_init_oa_buffer(). Zero the buffer at allocation to prevent information disclosure through mmap. The existing zeroing in xe_oa_init_oa_buffer() remains necessary for zero report-id/timestamp detection on enable/re-enable. (cherry picked from commit 126c4c5388f9311a390b9b254a835b0cf0f55331)

target_do_delayed_work() takes dev->delayed_cmd_lock with a plain spin_lock(). It runs from system_percpu_wq, so interrupts stay enabled while the lock is held. Every other acquisition of that lock uses spin_lock_irqsave(): target_handle_task_attr() and transport_complete_ordered_sync() in target_core_transport.c, and target_non_ordered_release() in target_core_device.c. That was safe while every caller of transport_complete_task_attr() ran in process context: target_complete_ok_work() from target_completion_wq, transport_complete_qf() from dev->qf_work_queue, and transport_generic_request_failure() from the submit and execute paths. Since commit 06933066d88a ("scsi: target: Add support for completing commands from backend context") that is no longer true. With complete_type=1 on a fabric that sets direct_compl_supp, target_complete() runs target_complete_ok_work() inline in the backend's completion context, which target_complete_cmd_with_sense() documents as "May be called from interrupt context". For iblock that is the bio end_io. target_complete_ok_work() starts with transport_complete_task_attr(), which takes delayed_cmd_lock in two ways: - Every ORDERED command is put on dev->delayed_cmd_list by target_handle_task_attr() and dispatched from there by target_do_delayed_work(), which marks it SCF_TASK_ORDERED_SYNC. Its completion therefore goes through transport_complete_ordered_sync(). - A SIMPLE command that completes while an ORDERED command is pending drops the last reference to the killed dev->non_ordered, and percpu_ref_put() then runs target_non_ordered_release() in the completing context. Both take the lock and, if commands are waiting, call schedule_work(&dev->delayed_cmd_work), which queues on system_percpu_wq, i.e. on the CPU that took the completion. Further completions for the same queue are typically delivered there too, so one arriving while the work holds the lock spins on it forever. Reaching this needs complete_type=1, which is not the default (vhost-scsi, the only fabric with direct_compl_supp, defaults to TARGET_QUEUE_COMPL), a backend whose I/O completes from interrupt context, and ORDERED commands from the initiator. Linux virtio_scsi guests send only VIRTIO_SCSI_S_SIMPLE and never take the ordered path; any guest that uses the ORDERED task attribute, which vhost-scsi accepts, does. target_do_delayed_work() is a work item and always runs in process context, so spin_lock_irq() is sufficient. Found by inspection; no runtime report. A single ORDERED command goes through both acquisitions, so with CONFIG_PROVE_LOCKING the first one completed through direct completion should produce an inconsistent lock state report on delayed_cmd_lock (hardirq or softirq, depending on where the backend completes).

Separate configfs opens can concurrently update the discovery AuthMethod parameter. Both callers may free the same value in iscsi_update_param_value(), causing a double-free. Protect the lookup and replacement with the discovery TPG access lock, as used by the other TPG configfs parameter updates.

Paul reported recent RCU changes exposed a lost resched where CPU idles with a rcutorture thread waiting while being completely runnable. Dumping the rq state at the time of idling with runnable tasks showed the following: rq->next_class = &fair_sched_class rq->nr_running = 1 rq->cfs.h_nr_queued = 0 rq->rt.rt_nr_running = 1 Although the waiting task was an RT task, "rq->next_class" did not point to rt_sched_class and was stuck at the lower priority fair_sched_class. Inspecting sched_change_end() revealed the following race with __sched_setscheduler() for queued, but non-running tasks: CPU0 CPU1 ==== ==== __schedule() /* * rq->donor is a fair task that blocks. * rq->next_class = &fair_sched_class */ rq_lock(rq0) pick_next_task() /* nr_running = 0 */ pick_task_fair() sched_balance_newidle() raw_spin_rq_unlock(rq0) rq_lock(rq0) activate_task(rq0, p) /* p is a fair task */ rq0->nr_running++; rq0->cfs.h_nr_queued++; wakeup_preempt(rq, p) rq->next_class = &fair_sched_class resched_curr() rq_unlock(rq0) __sched_setscheduler(p) /* p is moved to SCHED_RR */ task_rq_lock(p) sched_change_begin() rq0->cfs.h_nr_queued--; rq0->rt.rt_nr_running++; sched_change_end() /* * p->on_rq = 1 (ctx->queued) * but p is not rq->donor * (!ctx->running) hence * resched_curr() is and * rq->next_class upgrade is * skipped. */ task_rq_unlock() raw_spin_rq_lock(rq0) /* * rq->nr_running = 1 * rq->cfs.h_nr_queued = 0 * rq->rt.rt_nr_running = 1 * rq->next_class = &fair_sched_class */ if (rq->cfs.h_nr_queued && ...) /* Above is false */ if (rq_modified_above(this_rq, &fair_sched_class)) /* Also false since rq->next_class is still fair */ /* * sched_balance_newidle() is considered to * have failed to pull any fair tasks and no * other wakeups were detected. */ return NULL; /* pick_task_fair() */ pick_task_idle() return rq->idle; next = rq->idle; /* Preemption indicators are lost */ clear_tsk_need_resched(prev); clear_preempt_need_resched(); rq->curr = rq->idle; !!! rq->idle switches in with rq->running == 1 but !need_resched() !!! In simple words, enqueue, followed by a class change of the newly queued task goes unnoticed in rq->next_class indicator and sched_balance_newidle() sees no fair task, nor any higher priority task enqueued. The above sequence will lead to CPU idling with a runnable task on the CPU. Call wakeup_preempt() for ctx->queued cases in sched_change_end() to re-evaluate rq->next_class after class change (ENQUEUE_CLASS) and set the NEED_RESCHED indicator accordingly. Newidle balance will correctly see an enqueue of a higher priority task if a queued task was upgraded when the rq_lock() was dropped and will retry the pick.

Recovery participant records must remain alive while debug_state_print() walks reco.node_data and formats their fields. The formatter holds dlm->spinlock, but the recovery participant list is protected by dlm_reco_state_lock. In particular, dlm_destroy_recovery_area() detaches the list under dlm_reco_state_lock and frees the detached records after releasing it. A concurrent debugfs open can therefore read freed memory. Protect the participant traversal and field reads in debug_state_print() with dlm_reco_state_lock, matching the existing list writers. This also serializes the traversal against list_add_tail() during recovery-area initialization. Make the existing lock available to the debug code through dlmcommon.h; the lock remains internal to the OCFS2 DLM module. Release dlm->spinlock after formatting the other domain state and before acquiring dlm_reco_state_lock, so the debug reader does not nest the two locks. Leave recovery list insertion, detachment and freeing unchanged. KASAN report as below: BUG: KASAN: slab-use-after-free in debug_state_open+0x1169/0x12d0 Read of size 4 at addr ffff88810662ef80 by task dlm-state-stres/896 CPU: 1 UID: 0 PID: 896 Comm: dlm-state-stres Not tainted 7.3.0-rc4-next-20260921-pmb-ocfs2-functional-v1+ #1 PREEMPT(lazy) Hardware name: QEMU Ubuntu 24.04 PC v2 (i440FX + PIIX, arch_caps fix, 1996), BIOS 1.16.3-debian-1.16.3-2 04/01/2014 Call Trace: dump_stack_lvl+0x93/0xd0 print_report+0xce/0x630 ? debug_state_open+0x1169/0x12d0 ? srso_alias_return_thunk+0x5/0xfbef5 ? __virt_addr_valid+0x20e/0x420 ? debug_state_open+0x1169/0x12d0 kasan_report+0xe0/0x110 ? debug_state_open+0x1169/0x12d0 debug_state_open+0x1169/0x12d0 ? __pfx_debug_state_open+0x10/0x10 full_proxy_open_regular+0x193/0x310 do_dentry_open+0x595/0x12d0 ? __pfx_full_proxy_open_regular+0x10/0x10 vfs_open_consume+0xd1/0x400 ? srso_alias_return_thunk+0x5/0xfbef5 path_openat+0x18d0/0x2020 ? __pfx_path_openat+0x10/0x10 do_file_open+0x21e/0x470 ? __pfx_do_file_open+0x10/0x10 ? _raw_spin_unlock+0x23/0x40 ? srso_alias_return_thunk+0x5/0xfbef5 ? alloc_fd+0x3a6/0x6b0 do_sys_openat2+0xf4/0x1b0 ? __pfx_do_sys_openat2+0x10/0x10 ? srso_alias_return_thunk+0x5/0xfbef5 ? __fput+0x5b5/0xa60 __x64_sys_openat+0x136/0x1e0 ? __pfx___x64_sys_openat+0x10/0x10 do_syscall_64+0x114/0x620 entry_SYSCALL_64_after_hwframe+0x77/0x7f [Register dump omitted.] Allocated by task 880: kasan_save_stack+0x33/0x60 kasan_save_track+0x14/0x30 __kasan_kmalloc+0xaa/0xb0 __kmalloc_cache_noprof+0x28d/0x610 dlm_remaster_locks+0x147/0x1d90 dlm_do_recovery+0xde1/0x1580 dlm_recovery_thread+0x109/0x300 kthread+0x351/0x460 ret_from_fork+0x659/0x940 ret_from_fork_asm+0x1a/0x30 Freed by task 880: kasan_save_stack+0x33/0x60 kasan_save_track+0x14/0x30 kasan_save_free_info+0x3b/0x60 __kasan_slab_free+0x5f/0x80 kfree+0x308/0x580 dlm_destroy_recovery_area+0x285/0x460 dlm_remaster_locks+0x17c4/0x1d90 dlm_do_recovery+0xde1/0x1580 dlm_recovery_thread+0x109/0x300 kthread+0x351/0x460 ret_from_fork+0x659/0x940 ret_from_fork_asm+0x1a/0x30 The buggy address belongs to the object at ffff88810662ef80 which belongs to the cache kmalloc-32 of size 32 The buggy address is located 0 bytes inside of freed 32-byte region [ffff88810662ef80, ffff88810662efa0) [Page and memory-state dumps omitted.]

F_SETFL and FIOASYNC decide whether to invoke ->fasync() before pipe_fasync() takes pipe->mutex. Two enables of one O_RDWR FIFO can therefore both enter the callback. fasync_helper() allocates before checking for an existing entry. If the first enable installs both entries, a second enable can update the reader entry and then fail the allocation for the writer entry. The error path removes the pre-existing reader entry, clearing FASYNC while the writer entry remains registered. __fput() then skips ->fasync(), so a later pipe notification can dereference the freed file or its f_owner. An unprivileged memory-pressure reproducer triggers a KASAN slab-use-after-free in send_sigio() through this path. Keep the reader helper's result and undo the registration only if this call added it. The pipe mutex serializes changes to both lists, so a failed new registration still removes its reader, while a failed duplicate preserves the existing registrations and FASYNC. Successful duplicate enables continue to update fa_fd as before.

smp_execute_task_sg() calls pm_runtime_get_sync() on the host before issuing an SMP command. When that command is itself issued from the HA resume path, the get_sync() deadlocks: it waits for the ongoing resume (the device is RPM_RESUMING), while the resume is blocked in sas_drain_work() waiting for that same SMP IO to complete. The deadlock needs an expander-attached SATA disk. During sas_resume_ha() -> sas_drain_work(), DISCE_RESUME -> sas_resume_sata() -> ata_sas_port_resume() requests ATA_EH_RESET, and the hard reset for such a disk is done via SMP PHY CONTROL (sas_ata_hard_reset() -> sas_phy_reset() -> sas_smp_phy_control() -> smp_execute_task_sg()). Direct-attached SATA resets through lldd_control_phy() and SSP devices use TMFs, so neither hits this. Replace the get_sync()/put_sync() pair with pm_runtime_get_noresume()/pm_runtime_put(). smp_execute_task_sg() only needs to hold off autosuspend while the SMP is in flight, and it must not try to resume the host: a sync resume issued from the HA resume path itself is what deadlocks, and by the time sas_resume_ha() runs, hw_init has already reinitialized the hardware, so the device is accessible without one. The usage reference is still required. Discovery work normally runs inside an event worker's PM reference, taken at sas_notify_port_event() notify time and held until the handler has flushed the disco queue. sas_rediscover_ex_phy() however requeues DISCE_REVALIDATE_DOMAIN from within the revalidation worker itself, and flush_workqueue() does not wait for work items queued during execution, so that chained revalidation runs with no outer PM reference - without the get_noresume(), its SMP could race autosuspend. For the BSG path, sas_smp_handler() is the only caller which may find the host autosuspended: expander SMP requests do not go through any SCSI device request queue, so nothing else in that path holds the host awake. Resume it there with pm_runtime_resume_and_get() and check the result.

The netlink rhashtable key includes a raw struct net pointer and a user-controlled port ID. Both /proc/net/netlink and socket diagnostics expose the table's bucket order. By binding and rebinding chosen NETLINK_USERSOCK port IDs, an unprivileged reader can distinguish equal buckets and recover the low bits of the Jenkins hash. Its 32-bit seed and the limited set of kernel-image slides can then be searched offline to recover the address of init_net. Use the namespace's unique, non-address net_cookie in the comparison key instead. It is assigned before the per-net initializers run and remains unchanged for the namespace's lifetime. The lookup key and object hash are still built by netlink_compare_arg_init(), keeping lookup, insertion, removal and rehashing consistent while preserving namespace separation. Neither public table walker needs to change. Reading the socket's namespace cookie in netlink_compare() is safe under RCU, including during namespace teardown. netlink_release() removes the socket from the hash table and defers its final put with call_rcu(). cleanup_net() runs the per-net exit methods and waits for outstanding RCU callbacks with rcu_barrier() before freeing namespace storage. Unlike ns.ns_id, net_cookie is also available and initialized during setup_net() in older stable kernels, making the change straightforward to backport to v5.15 and later.

hisi_femac_drv_probe() connects the PHY before dev->netdev_ops is assigned. phy_attach_direct() -> phy_link_topo_add_phy() reads dev->netdev_ops through netdev_need_ops_lock() since the commit in question, so probing crashes with a NULL pointer dereference when CONFIG_NET_SHAPER is enabled. phy_link_topo_add_phy from phy_attach_direct+0xec/0x37c of_phy_get_and_connect from hisi_femac_drv_probe+0x1f4/0x558 Kernel panic - not syncing: Attempted to kill init! exitcode=0x0000000b Assign netdev_ops before the PHY attach, as was done for emac in commit 7c9f391ec89c ("net: emac: move setting of netops to fix crash").

hip04_mac_probe() calls of_phy_connect() before ndev->netdev_ops is assigned. phy_attach_direct() -> phy_link_topo_add_phy() reads dev->netdev_ops through netdev_need_ops_lock() since the commit in question, so probing crashes with a NULL pointer dereference when CONFIG_NET_SHAPER is enabled. Assign netdev_ops before the PHY attach, as was done for emac in commit 7c9f391ec89c ("net: emac: move setting of netops to fix crash").

owl_emac_probe() calls owl_emac_phy_init(), which connects the PHY, before netdev->netdev_ops is assigned. phy_attach_direct() -> phy_link_topo_add_phy() reads dev->netdev_ops through netdev_need_ops_lock() since the commit in question, so probing crashes with a NULL pointer dereference when CONFIG_NET_SHAPER is enabled. Assign netdev_ops before the PHY attach, as was done for emac in commit 7c9f391ec89c ("net: emac: move setting of netops to fix crash").

mv643xx_eth_probe() connects the PHY before dev->netdev_ops is assigned. phy_attach_direct() -> phy_link_topo_add_phy() reads dev->netdev_ops through netdev_need_ops_lock() since the commit in question, so probing crashes with a NULL pointer dereference when CONFIG_NET_SHAPER is enabled. Assign netdev_ops before the PHY attach, as was done for emac in commit 7c9f391ec89c ("net: emac: move setting of netops to fix crash").

au1000_probe() calls au1000_mii_probe(), which connects the PHY, before dev->netdev_ops is assigned. phy_attach_direct() -> phy_link_topo_add_phy() reads dev->netdev_ops through netdev_need_ops_lock() since the commit in question, so probing crashes with a NULL pointer dereference when CONFIG_NET_SHAPER is enabled. Assign netdev_ops before the PHY attach, as was done for emac in commit 7c9f391ec89c ("net: emac: move setting of netops to fix crash").

ethoc_probe() calls ethoc_mdio_probe(), which connects the PHY, before netdev->netdev_ops is assigned. phy_attach_direct() -> phy_link_topo_add_phy() reads dev->netdev_ops through netdev_need_ops_lock() since the commit in question, so probing crashes with a NULL pointer dereference when CONFIG_NET_SHAPER is enabled. Assign netdev_ops before the PHY attach, as was done for emac in commit 7c9f391ec89c ("net: emac: move setting of netops to fix crash").

greth_of_probe() calls greth_mdio_init(), which connects the PHY, before dev->netdev_ops is assigned. phy_attach_direct() -> phy_link_topo_add_phy() reads dev->netdev_ops through netdev_need_ops_lock() since the commit in question, so probing crashes with a NULL pointer dereference when CONFIG_NET_SHAPER is enabled. Assign netdev_ops before the PHY attach, as was done for emac in commit 7c9f391ec89c ("net: emac: move setting of netops to fix crash").

icssm_prueth_netdev_init() calls of_phy_get_and_connect() before ndev->netdev_ops is assigned. phy_attach_direct() -> phy_link_topo_add_phy() reads dev->netdev_ops through netdev_need_ops_lock() since the commit in question, so probing crashes with a NULL pointer dereference when CONFIG_NET_SHAPER is enabled. Assign netdev_ops before the PHY attach, as was done for emac in commit 7c9f391ec89c ("net: emac: move setting of netops to fix crash").

bcmgenet stores first_cb, last_cb and bytes_sent in the skb control block and reads them back on completion to free the skb and credit BQL. ether_setup() leaves IFF_TX_SKB_SHARING set, so the stack can resubmit a shared skb that is still in the ring. The next transmit overwrites last_cb, so the earlier slots complete without crediting their bytes and BQL stalls the queue until the watchdog resets it. Unshare the skb before writing the control block. Reproduce with pktgen in clone_skb mode, which hands the driver the same skb while earlier copies are still in the ring: modprobe pktgen pg=/proc/net/pktgen echo "add_device eth0" > $pg/kpktgend_0 echo "dst_mac AA:BB:CC:DD:EE:FF" > $pg/eth0 echo "dst 192.0.2.1" > $pg/eth0 echo "clone_skb 8" > $pg/eth0 echo "count 0" > $pg/eth0 echo start > $pg/pgctrl & sleep 5 echo stop > $pg/pgctrl wait cat /sys/class/net/eth0/queues/tx-0/byte_queue_limits/inflight inflight stays nonzero on the now idle queue. clone_skb 0 leaves it at zero.

calipso_skbuff_setattr() adjusts payload_len by the length of the option it inserts or resizes: payload = ntohs(ip6_hdr->payload_len); ip6_hdr->payload_len = htons(payload + len_delta); GRO can build BIG TCP packets larger than IPV6_MAXPLEN, and stores payload_len == 0 for them; ipv6_payload_len() then falls back to skb->len. For such a packet forwarded to a destination mapped to CALIPSO, the code above writes len_delta (or 65536 + len_delta when shrinking) into payload_len, and ipv6_payload_len() no longer falls back to skb->len. nf_tables, conntrack and sch_cake rely on it. The same arithmetic also wraps when a packet grows past IPV6_MAXPLEN, and a BIG TCP packet that shrinks below IPV6_MAXPLEN should get its real length back. Set payload_len from the skb length with ipv6_set_payload_len(), which stores 0 above IPV6_MAXPLEN and the real length otherwise, as ipv6_gro_complete() does. The network header was just reset to skb->data, so skb->len - sizeof(*ip6_hdr) is the payload length. Tested on one kernel (arm64, QEMU) with SELinux permissive, a minimal mdp-generated policy (network_peer_controls=1), "netlabelctl calipso add pass doi:16" and a NetLabel mapping of fd00:2::/64 to CALIPSO DOI 16. calipso_skbuff_setattr() was called through netlbl_skbuff_setattr() from a test-only debugfs hook (not part of this patch) adding a level-only label; payload_len after the call: case before fix after fix real length 1000, unlabeled 1016 1016 1016 70000 GSO, payload_len 0 16 0 70016 65530 GSO, grows past MAXPLEN 10 0 65546 65540 GSO with a 24 byte hop header, shrinks below MAXPLEN 65528 65532 65532 1000 with a 24 byte hop header 992 992 992 (0 means ipv6_payload_len() returns the real length.) On the forward path, with a client injecting unlabeled UDP through a router to a server, the router adds the CALIPSO option and the server delivers all datagrams (Udp6InDatagrams 5, Ip6InTruncatedPkts 0) with this patch.

A drive-by review by Sashiko reports that, after a failed call to gpiochip_set_multiple(), gpiod_set_array_value() returns abruptly, bypassing the freeing of possibly alloced bitmaps, potentially leaking memory. It is an unlikely error path, but is possible. Defer the return code check until after the bitmap_free() block.

mxc_update_irq_chained_handler() is called before the IRQ domain, the generic IRQ chip, and the port list entry are set up. If an interrupt arrives in that window: - mx3_gpio_irq_handler() calls generic_handle_domain_irq() with port->domain still NULL. - mx2_gpio_irq_handler() walks mxc_gpio_ports, but the port has not been added to the list yet. Additionally, if any of the subsequent probe steps (gpio_generic_chip_init(), devm_gpiochip_add_data(), irq_domain_create_legacy(), or mxc_gpio_init_gc()) fail, the error paths never unregister the chained handler, leaving a dangling handler that points at freed memory. Move the handler installation after all its dependencies are ready and after list_add_tail(), so the handler is never live while the data structures it touches are incomplete, and is never installed if probe fails.

Since commit 2e2c878a3141 ("net: stmmac: remove SGMII/RGMII/SMII interrupt handling") nothing on the RGMII path reads GMAC_RGSMIIIS, so RGSMIIIS stays pending behind GMAC_INT_MASK from the first link change on. stmmac_resume() resets the MAC in stmmac_hw_setup(); the reset sets GMAC_INT_MASK to 0 until dwmac1000_core_init() writes the mask again, and the pending status storms the line: dwmac_dma_interrupt: unexpected status 04000000 A register dump at the first message on ODROID-HC4 showed it: core int_status 00000001 int_mask 00000000 pmt 00000000 rgsmiiis 0000000d When the resume thread runs on the CPU that takes the MAC interrupt, the storm starves it: a soft lockup, then a watchdog reset or a hang. Read GMAC_RGSMIIIS in dwmac1000_irq_status() when the status is not masked, as the driver did before 2e2c878a3141; the read clears it. Wake-on-LAN resume from s2idle with the resume thread and the MAC interrupt on one CPU: Helios64 (dwmac-rk), 7.3-rc6, soft-locked in 3 of 3 resumes without a fix and in 0 of 10 with this patch; it ran with an out-of-tree dwmac-rk change that keeps the GMAC powered across suspend. ODROID-HC4 (dwmac-meson8b): 0 of 10 with this patch, 2 of 2 without a fix on a 7.3-rc5 build. irq_rgmii_n went up by one per resume.

__perf_pmu_remove() walks the group leaders of a CPU and PMU with perf_event_groups_for_cpu_pmu(), removing each one as it goes. perf_remove_from_context() takes the leader out of the groups rbtree, after which rb_next() on that node returns NULL, prematurely ending the walk. Other groups stay behind in the source context. On CPU hotplug, they are then torn down with the dying CPU. They keep running until teardown, while the PMU's interrupt and migrated events are on a new CPU, which drivers rely on never happening. Walk ctx->event_list instead and pick out the PMU's events by their pmu_ctx.

Commit 45e1dccc0653 ("x86/insn: Simplify for_each_insn_prefix()") made for_each_insn_prefix() declare its own loop counter named idx. get_seg_reg_override_idx() keeps its result in a local of the same name, so the assignments in the loop body now write the loop counter: - the function always returns INAT_SEG_REG_DEFAULT, so FS and GS overrides are ignored; - a CS override sets the counter to INAT_SEG_REG_CS (2), the increment makes it 3, and prefixes.bytes[3] holds the last prefix, 0x2e again, so the loop never terminates. The UMIP emulation reaches this from the #GP handler for user space: an SGDT, SIDT or SMSW with a memory operand and a CS prefix leaves the task spinning in the kernel, never returning to user space, with a SIGKILL that is never delivered. Any user can thus tie up CPUs until reboot. With an FS or GS prefix, the emulated result is written to the wrong address. Rename the internal macro iterator variable. [ Lukas' original patch renamed the use site, I changed it to rename the iterator in the macro instead, but left Lukas' commit message and authorship - Linus ]

For a split transaction, either fw_core_handle_response() or split_transaction_timeout_callback() may manipulate the same transaction entry in the pending transaction list. The return value of timer_delete() is used to determine which function handles the entry. If the transaction is a split transaction and timer_delete() returns zero, the timeout callback may already be running. In this case, the response handler must leave the entry in the list for the timeout callback to remove. A previous code refactoring broke this behavior, introducing a race condition in which both functions could process the same transaction. This could lead to a use-after-free on the transaction initiator side. Fix the regression by ensuring that the response handler does not process the transaction when timer_delete() returns zero. The timeout callback is responsible for removing the entry from the pending transaction list.

rocket_remove() decrements rdev->num_cores for each core it removes, while find_core_for_dev() searches slots 0 to num_cores - 1. Unbinding the cores in the order they were bound therefore loses the last one: by the time it is removed the search range has already shrunk past its slot, so find_core_for_dev() returns -1 and rocket_remove() gives up without doing anything. num_cores never reaches zero, rocket_device_fini() never runs, and the file-scoped rdev keeps pointing at a device that is going away. Binding the cores again starts from that stale count, because rocket_probe() takes rdev->num_cores as the slot to fill. On an RK3588, which describes three cores, the second round lands on slots 1, 2 and 3 while rdev->cores was allocated with room for three: rocket fdab0000.npu: Rockchip NPU core 1 version: 1179210309 rocket fdac0000.npu: Rockchip NPU core 2 version: 1179210309 rocket fdad0000.npu: Rockchip NPU core 3 version: 1179210309 The write to rdev->cores[3] is past the end of the array. Nothing in tree reads the core array often enough to notice, so the overrun is silent today. It turned up while testing a devfreq series on top of this, where a worker walks every core a few times a second, and UBSAN caught the first bool it read out of the overrun entry: UBSAN: invalid-load in drivers/accel/rocket/rocket_devfreq.c:47:10 load of value 5 is not a valid value for type '_Bool' Workqueue: devfreq_wq devfreq_monitor Record how many slots were allocated and search all of them. Every core is then found on removal, num_cores reaches zero, the device is torn down and a later bind starts from a clean rdev. This does not make unbinding a single core out of several work. probe still takes num_cores as the slot to fill, so rebinding one core while its siblings stay bound would write over a slot that is already in use, and rocket_open() still reaches for cores[0] whether or not anything is there. Both of those want more thought than a fix should carry. Found by unbinding and rebinding all three cores on an Orange Pi 5 Plus. With this applied, 25 unbind/rebind rounds and 5 module unload/reload rounds run clean there: the cores land in slots 0, 1 and 2 every time, whichever order they are bound in, and the shared supply goes back to a single user after each round.

sched_to_core() walks rdev->cores[] up to rdev->num_cores, and rocket_remove() decrements num_cores for every core it removes. Unbind a core that is not the last one and the cores behind it fall outside the search, so sched_to_core() returns NULL for a core that is still bound and still running jobs. Neither caller checks the result: rocket_job_run(): rocket_fence_create(core), core->dev rocket_job_timedout(): dev_err(core->dev, "NPU job timed out") Unbinding the middle core of the three on an RK3588 while three clients are submitting to all of them faults twice, once from the surviving core's job queue and once from its reset work: KASAN: null-ptr-deref in range [0x0000000000000220-0x0000000000000227] Workqueue: fdad0000.npu drm_sched_run_job_work [gpu_sched] pc : rocket_job_run+0x234/0x838 [rocket] Call trace: rocket_job_run+0x234/0x838 [rocket] drm_sched_run_job_work+0x2cc/0xad8 [gpu_sched] process_one_work+0x640/0x14f0 KASAN: null-ptr-deref in range [0x0000000000000000-0x0000000000000007] Workqueue: rocket-reset-2 drm_sched_job_timedout [gpu_sched] pc : rocket_job_timedout+0xf0/0x1e0 [rocket] Call trace: rocket_job_timedout+0xf0/0x1e0 [rocket] drm_sched_job_timedout+0x188/0x6a0 [gpu_sched] Both are the third core: the workqueue names are its device and its core->index, and it was left at slot 2 while num_cores had dropped to 2. Search all the slots that were allocated, the way find_core_for_dev() now does. A core that is still bound is then found, and the two callers get the pointer they already assume they have. This does not make unbinding one core out of several safe. An open client keeps an entity pointing at the scheduler of the core that went away: drm_sched reports it as not ready for every job that lands on it, and the client waits in dma_fence_default_wait for a fence that will never signal. Stopping the NULL dereference is what belongs in a fix; the rest wants more thought.

When ch_count is smaller than blob->alh_cfg.device_count (e.g. a mono capture stream, such as a SmartMic, aggregated across multiple SDW links), step = ch_count / blob->alh_cfg.device_count truncates to 0. The subsequent GENMASK(step - 1, 0) then underflows step - 1 (u32) to UINT_MAX, causing GENMASK()'s internal shift to use an out-of-bounds exponent: UBSAN: shift-out-of-bounds in sound/soc/sof/ipc4-topology.c:2363:13 shift exponent 18446744069414584384 is too large for 64-bit type 'long unsigned int' Treat ch_count < device_count the same way as the existing "channels equal to output reference channels" case: apply the same ch_mask to every aggregated device instead of trying to split channels that can't be evenly divided.

autofs_notify_daemon() takes a reference to the current notification pipe under wq_mutex, then drops the mutex before writing the request. The write can block while the daemon enters catatonic mode and installs a replacement pipe. If the old pipe's reader is then closed, the delayed write returns -EPIPE and the generic error path enters catatonic mode again. That transition acts on the current superblock state, closes the replacement pipe, and releases the new daemon generation's wait queues. Only enter catatonic mode when the pipe which failed is still the current pipe. Factor the already-locked transition so the identity check and state change are atomic with respect to another replacement. This was reproduced on v7.3-rc5-337-gff47652a4b66c in a local QEMU guest. An unprivileged UID 65534 lookup filled the old packet pipe before a normal CATATONIC/SETPIPEFD restart. On the unmodified kernel, the new pipe received HUP and both a pending victim lookup and a later lookup failed with ENOENT in two out of two fresh-mount runs. With this change, the new pipe remained active, received an intact 304-byte request, and both victim lookups completed in two out of two runs. Normal lookup and restart-without-a-blocked-writer controls passed. No KASAN, oops, refcount, or lock diagnostic was observed. A source reproducer and complete local logs are available privately on request. The demonstrated impact is limited to denial of a shared autofs service during a legitimate daemon restart. No memory corruption, information leak, privilege escalation, or code execution primitive was observed.

generic_ci_d_hash() and generic_ci_d_compare() read sb->s_encoding and dereference the unicode map while a pathwalk in rcu mode holds nothing but rcu_read_lock(). The filesystems that use them, ext4, f2fs and tmpfs, drop the map with utf8_unload() from ->put_super(), which runs when the last reference to the mount goes away. A lookup that is still inside such a filesystem at that point reads a freed map. With a pathwalk held up in utf8_casefold_hash() while the last reference to a casefolded tmpfs goes away: BUG: KASAN: slab-use-after-free in utf8nlookup+0x74c/0x950 Read of size 8 at addr ff110001030fe658 by task walker/98 Call Trace: utf8nlookup+0x74c/0x950 utf8byte+0x160/0x1560 utf8_casefold_hash+0x167/0x230 generic_ci_d_hash+0xdd/0x1c0 link_path_walk+0x11ad/0x1d30 path_lookupat+0x63/0xb00 filename_lookup+0x1f1/0x5b0 vfs_statx+0xd7/0x3b0 [...] Allocated by task 94: __kmalloc_cache_noprof+0x249/0x560 utf8_load+0x4e/0x570 shmem_parse_opt_casefold.isra.0+0x92/0x270 [...] Freed by task 97: kfree+0x1db/0x5b0 shmem_put_super+0x72/0x170 generic_shutdown_super+0x14c/0x530 kill_anon_super+0x3a/0x60 deactivate_locked_super+0x9c/0xf0 cleanup_mnt+0x9f/0x130 So far that has been hard to hit. umount(2) waits for a grace period in namespace_unlock() before it drops the mount, so only a mount whose last reference goes away elsewhere, like the close of a file or a chdir out of a lazily unmounted directory, can race a walker. So it is reachable. Delay the free until a grace period has passed. The map and the data table it holds a reference to then outlive any pathwalk that found the superblock before the mount went away. The reference on the data table module is dropped from the rcu callback as well. utf8-core is built in whenever CONFIG_UNICODE is set, so there is no module exit that would have to wait for the callbacks. Overlayfs points the s_encoding of an overlay at the map of its first layer without taking a reference of its own. That stays correct. The layer is held by the overlay, so its map is dropped only after the overlay is gone.

There is a race between watch_queue_set_size() and post_one_notification(). watch_queue_set_size() publishes wqueue->notes, notes_bitmap, nr_pages and nr_notes with plain stores, and post_one_notification() reads them with plain loads under a different lock, so nothing orders the four stores against the poster's loads: depending on the build, notes can be stored last (gcc 14 on x86-64 does this). Since keyctl_watch_key() lets a watch attach to a queue before it is sized, post_one_notification() can already be running on another CPU at that moment: it reads nr_notes > 0, passes its bounds check, and then reads notes while it is still NULL, dereferencing notes[0]. An unprivileged user triggers this reliably by watching one of its own keys through a notification pipe and racing ioctl(IOC_WATCH_QUEUE_SET_SIZE) on that pipe against keyctl(KEYCTL_SETPERM) on the key from another CPU, triggering: BUG: kernel NULL pointer dereference, address: 0000000000000000 RIP: 0010:post_one_notification.isra.0+0xa2/0x1e0 Call Trace: __post_watch_notification+0x148/0x180 keyctl_setperm_key+0x119/0x130 do_syscall_64+0x108/0x4c0 entry_SYSCALL_64_after_hwframe+0x77/0x7f Kernel panic - not syncing: Fatal exception in interrupt Fix this by publishing nr_notes last with smp_store_release(), and by reading it once with smp_load_acquire() in post_one_notification() before the bitmap and the page array are touched. A poster that observes a non-zero nr_notes then also observes the pool behind it; one that observes zero drops the notification as before.

An async #PF "page not present" event arrives through the #PF vector, and its reason sits in the per-CPU apf_reason.flags until __kvm_handle_async_pf() reads and clears it. If an NMI arrives after the event is injected and before the #PF handler reads the flags, and the NMI handler itself takes a #PF, that fault reads and clears the flags belonging to the #PF it interrupted. It then finds interrupts disabled in the NMI context it interrupted, and panics: Kernel panic - not syncing: Host injected async #PF in interrupt disabled region Call Trace: panic+0x52/0x60 __kvm_handle_async_pf+0xb0/0xb0 exc_page_fault+0x71/0xb0 asm_exc_page_fault+0x27/0x30 RIP: 0010:__get_user_nocheck_8+0x6/0x20 perf_callchain_user+0x147/0x1a0 get_perf_callchain+0x12d/0x1c0 ... perf_event_nmi_handler+0x26/0x50 exc_nmi+0xe8/0x110 RIP: 0010:asm_exc_page_fault+0x0/0x30 The last RIP line shows the NMI landing on the first instruction of the #PF entry. perf with user call chains takes such a fault: the sampling NMI walks the user stack with __get_user(), which faults when a saved frame pointer is not a valid user address or the walk runs off the top of the stack. The same window exists when KVM, running as an L1 hypervisor with DELIVERY_AS_PF_VMEXIT, reads the flags itself after a #PF VM-exit. The NMI entry code saves and restores CR2 for this nesting, but not the async #PF reason. Andy Lutomirski described the race, with perf as the example, when async #PF delivery was reworked in 2020 [1]. Saving and clearing the flags across the NMI, next to CR2, was discussed there but not merged. Since that rework, Linux has stopped setting KVM_ASYNC_PF_SEND_ALWAYS, so the host delivers "page not present" only for user mode faults and a kernel-mode fault in NMI context (including kernel-mode #MC and #DB) cannot be an async page fault. In that case, return without touching the flags so that the interrupted handler finds them intact. Tested on v7.3-rc5 in a 4-vCPU, 2 GiB guest restored from a snapshot whose memory is served on demand through userfaultfd, so that reading it back takes async #PFs, while every thread samples hardware cycles with user call chains and points its frame pointer at an unmapped page. Without this patch the guest panicked as above in 3 of 3 runs. With it, the guest survived 3 of 3 runs of 60 seconds, and the host counted 148,441 to 153,334 kvm_async_pf_not_present events in the first 20 seconds of each. On 6.18.50 the results were the same: the guest panicked without the patch and survived with it.

set_page_dirty() requires that the caller holds some kind of lock to ensure that the page's mapping does not concurrently go away. That is not the case for a random page we got from get_user_pages_fast(), so use set_page_dirty_lock(). [sean: tag for stable]

Reset the vCPU's MMU pointer back to the root MMU if loading L2 state fails in the "not from VM-Entry" case, i.e. when emulating RSM or handling KVM_SET_NESTED_STATE. Failure to unwind the MMU pointer results in KVM trying to run L1 with a nested MMU, and is especially problematic if the vCPU manages to free its nested state before the VM crashes. Constructing the root for a nested EPT MMU goes through vmcs12, which is dynamically allocated/freed when nVMX is enabled/disabled. And so if the nested state is freed with the MMU still pointing at the nested MMU, KVM will hit a NULL pointer deref when attempting to get L1=>L2 EPTP. BUG: kernel NULL pointer dereference, address: 0000000000000078 #PF: supervisor read access in kernel mode #PF: error_code(0x0000) - not-present page PGD 114a2d067 P4D 111743067 PUD 11141f067 PMD 0 Oops: Oops: 0000 [#1] SMP CPU: 20 UID: 1000 PID: 1272 Comm: nested_tdp_faul Not tainted 7.3.0-rc4-8c89f11d1a99b-sink-vm #1914 PREEMPTLAZY Hardware name: QEMU Standard PC (Q35 + ICH9, 2009), BIOS 0.0.0 02/06/2015 RIP: 0010:nested_ept_get_eptp+0xb/0x10 [kvm_intel] Call Trace: kvm_mmu_load+0xcd/0x870 [kvm] kvm_arch_vcpu_ioctl_run+0x18f5/0x1f80 [kvm] kvm_vcpu_ioctl+0x2c7/0x970 [kvm] __x64_sys_ioctl+0x90/0xd0 do_syscall_64+0x9c/0x4d0 entry_SYSCALL_64_after_hwframe+0x4b/0x53 RIP: 0033:0x47491b Modules linked in: kvm_intel kvm irqbypass CR2: 0000000000000078 ---[ end trace 0000000000000000 ]--- RIP: 0010:nested_ept_get_eptp+0xb/0x10 [kvm_intel] [sean: write changelog]

fsnotify_destroy_marks() keeps a reference to its current mark while it drops conn->lock to destroy the mark. This stabilizes the forward-walk cursor, but does not prevent a new mark from being inserted before it. For example, an inotify watch re-added after IN_IGNORED sorts before the old, detached mark for the same group. The walk can miss the replacement and then detach the entire connector from the object. The replacement is still on its group's marks_list and, for inotify, in the watch IDR. The group mutex no longer protects the connector's object from disappearing. An unprivileged fdinfo read racing the final detach can consequently call igrab(NULL): BUG: KASAN: null-ptr-deref in igrab+0x24/0x100 Write of size 4 at addr 0000000000000148 Call Trace: igrab+0x24/0x100 inotify_show_fdinfo+0x1dc/0x5e0 seq_show+0x397/0x690 Set a DESTROYING flag under conn->lock before starting the walk, and reject new marks on that connector with -ENOENT. The existing add-mark error paths undo the temporary group-list entry and the backend's watch allocation. Do not wait for teardown while holding the group mutex, which the destroyer needs. Once the old connector is detached, an add can install a fresh connector as before. This closes the insertion race for all object types using the common connector teardown, without changing the mark reference or SRCU lifetime rules.

ext4_get_link() follows a symlink that keeps its target in a data block in rcu pathwalk mode with ext4_getblk(EXT4_GET_BLOCKS_CACHED_NOWAIT). That reads s_mount_state in the ASSERT() of ext4_getblk() and again in ext4_es_lookup_extent(), and the lookup bumps the extent status hit or miss counter, all through sb->s_fs_info. ext4_put_super() destroys those counters, clears sb->s_fs_info and frees the sbi without waiting for a grace period. A pathwalk that is still in ext4_get_link() when the last reference to the mount goes away then dereferences a NULL s_fs_info, the freed sbi or the destroyed percpu counters. With a pathwalk held up in ext4_get_link() while the last reference goes away: KASAN: null-ptr-deref in range [0x00000000000000a8-0x00000000000000af] RIP: 0010:ext4_getblk+0xf0/0x640 Call Trace: ext4_get_link+0x26c/0x410 pick_link+0xf96/0x1780 step_into_slowpath+0x9b8/0x1250 path_lookupat+0x484/0xb00 filename_lookup+0x1f1/0x5b0 vfs_statx+0xd7/0x3b0 vfs_fstatat+0x4d/0x90 __do_sys_newfstatat+0xa0/0x100 do_syscall_64+0x10e/0x5e0 entry_SYSCALL_64_after_hwframe+0x77/0x7f So far that has been hard to hit. umount(2) waits for a grace period in namespace_unlock() before it drops the mount, so only a mount whose last reference goes away elsewhere, like the close of a file or a chdir out of a lazily unmounted directory, can race a walker. Keep sb->s_fs_info and free the sbi together with the extent status counters after a grace period. Split the counters out of ext4_es_unregister_shrinker() for that. The shrinker still gets wasted in ext4_put_super(), the counters go with the sbi.

tracefs_d_revalidate() reads ->is_freed of the eventfs_inode that dentry->d_fsdata points to, and ->d_revalidate() is called in rcu pathwalk mode with nothing but rcu_read_lock() held. The dentry itself stays around for such a walker, but tracefs_d_release() drops its reference to the eventfs_inode when the dentry is killed, and release_ei() then frees it after a grace period of eventfs_srcu. An SRCU grace period doesn't wait for rcu_read_lock() sections, so the walker can read an eventfs_inode that is already gone. With a pathwalk held up in tracefs_d_revalidate() while the dentry goes away: BUG: KASAN: slab-use-after-free in tracefs_d_revalidate+0x76/0x90 Read of size 1 at addr ff1100010152d8d0 by task walker/97 Call Trace: tracefs_d_revalidate+0x76/0x90 lookup_fast+0x320/0x640 path_lookupat+0x160/0xb00 filename_lookup+0x1f1/0x5b0 vfs_statx+0xd7/0x3b0 vfs_fstatat+0x4d/0x90 __do_sys_newfstatat+0xa0/0x100 do_syscall_64+0x10e/0x5e0 entry_SYSCALL_64_after_hwframe+0x77/0x7f [...] Allocated by task 1: __kmalloc_cache_noprof+0x249/0x560 eventfs_create_dir+0x7a/0x590 event_create_dir+0x2fb/0xcf0 trace_add_event_call+0xbf/0x150 [...] Freed by task 11: kfree+0x1db/0x5b0 srcu_invoke_callbacks+0x3a3/0x620 process_one_work+0x937/0x1810 worker_thread+0x4f1/0xd60 Before commit 8e556432477e ("eventfs: Use SRCU for freeing eventfs_inodes") release_ei() used call_rcu(), which covered the walker, while eventfs_iterate() needs the SRCU grace period. Wait for both and chain call_rcu() from the SRCU callback.

mq_init_ns() creates the mqueue mount with fc_mount_longterm(), so its ->mnt_ns is MNT_NS_INTERNAL. When setting up the mqueue or ipc sysctls or msg_init_ns() fails afterwards, create_ipc_ns() drops it with a plain mntput(). With ->mnt_ns set that takes the fast path of mntput_no_expire(). The count drops to zero and nothing ever kills the mount, so the mount and the mqueue superblock are leaked. Commit dc55e35f9e81 ("ipc: Store mqueue sysctls in the ipc namespace") added the first failure path after mq_init_ns() and never dropped the mount. Commit 7229d74e5e8c ("ipc: create_ipc_ns: drop mqueue mount on sysctl setup failure") tried to fix that with a mntput() but that doesn't help for a long-term mount. Use kern_unmount(). free_ipc() open-codes the same thing for a namespace that made it.

If the "nocase" mount option ntfs3 uses custom hash and dentry compare functions. Both access ntfs3 superblock private information in rcu pathwalk. ntfs_put_super() and ntfs3_kill_sb() drop the info without any rcu delay. A lookup that's still inside the filesystem when the last reference to the mount is dropped causes a UAF. With a pathwalk held up in ntfs_d_hash() while the last reference to the mount goes away: BUG: KASAN: slab-use-after-free in ntfs_nls_to_utf16+0x752/0x940 Read of size 8 at addr ff1100010453c850 by task walker/99 Call Trace: ntfs_nls_to_utf16+0x752/0x940 ntfs_d_hash+0x22b/0x500 link_path_walk+0x11ad/0x1d30 path_lookupat+0x63/0xb00 filename_lookup+0x1f1/0x5b0 vfs_statx+0xd7/0x3b0 [...] Allocated by task 95: __kmalloc_cache_noprof+0x249/0x560 ntfs_init_fs_context+0x2db/0x5a0 alloc_fs_context+0x52e/0xc70 [...] Freed by task 98: kfree+0x1db/0x5b0 deactivate_locked_super+0x9c/0xf0 cleanup_mnt+0x9f/0x130 task_work_run+0x131/0x250 The bug is currently masked by the grace period in namespace_unlock() for umount(2). Keep the necessary information until the filesystem private info is freed and free it with an rcu delay. ntfs3_free_sbi() now runs from an rcu callback and drops the upcase table via ntfs_put_shared(). So switch s_shared_lock over to spin_lock_bh(). Otherwise the callback can deadlock against a task holding the lock on the same cpu.