In the Linux kernel, the following vulnerability has been resolved: serial: qcom_geni: fix kfifo underflow when flush precedes DMA completion IRQ When uart_flush_buffer() runs before the DMA completion IRQ is delivered, the following race can occur (all steps serialized by uart_port_lock): 1. DMA starts: tx_remaining = N, kfifo contains N bytes 2. DMA completes in hardware; IRQ is pending but not yet delivered 3. uart_flush_buffer() acquires the port lock and calls kfifo_reset(), making kfifo_len() = 0 while tx_remaining remains N 4. uart_flush_buffer() releases the port lock 5. DMA IRQ fires; handle_tx_dma() acquires the port lock and calls uart_xmit_advance(uport, tx_remaining) on an empty kfifo uart_xmit_advance() increments kfifo->out by tx_remaining. Since kfifo_reset() already set both in and out to 0, out wraps past in, causing kfifo_len() to return UART_XMIT_SIZE - tx_remaining. The next start_tx_dma() call then submits a DMA transfer of stale buffer data. Fix this by snapshotting kfifo_len() at the start of handle_tx_dma() and skipping uart_xmit_advance() when fifo_len < tx_remaining, which indicates the kfifo was reset by a preceding flush.

In the Linux kernel, the following vulnerability has been resolved: arm64: mm: call pagetable dtor when freeing hot-removed page tables Since 5e8eb9aeeda3 ("arm64: mm: always call PTE/PMD ctor in __create_pgd_mapping()") page-table allocation on ARM64 always calls pagetable_{pte,pmd,pud,p4d}_ctor(). This sets the page_type to PGTY_table, increments NR_PAGETABLE and possible allocates a PTL. However the matching pagetable_dtor() calls were never added. With DEBUG_VM enabled on kernel versions prior to v6.17 without 2dfcd1608f3a9 ("mm/page_alloc: let page freeing clear any set page type") this leads to the following warning when freeing these pages due to page->page_type sharing page->_mapcount: BUG: Bad page state in process ... pfn:284fbb page: refcount:0 mapcount:0 mapping:0000000000000000 index:0x0 pfn:0x284fbb flags: 0x17fffc000000000(node=0|zone=2|lastcpupid=0x1ffff) page_type: f2(table) page dumped because: nonzero mapcount Call trace: bad_page+0x13c/0x160 __free_frozen_pages+0x6cc/0x860 ___free_pages+0xf4/0x180 free_pages+0x54/0x80 free_hotplug_page_range.part.0+0x58/0x90 free_empty_tables+0x438/0x500 __remove_pgd_mapping.constprop.0+0x60/0xa8 arch_remove_memory+0x48/0x80 try_remove_memory+0x158/0x1d8 offline_and_remove_memory+0x138/0x180 It can also lead to leaking the ptl allocation if ALLOC_SPLIT_PTLOCKS is defined and incorrect NR_PAGETABLE stats. Fix this by calling pagetable_dtor() in free_hotplug_pgtable_page() prior to freeing the page to undo the effects of calling pagetable_*_ctor().

linux

In the Linux kernel, the following vulnerability has been resolved: RDMA/rxe: Fix iova-to-va conversion for MR page sizes != PAGE_SIZE The current implementation incorrectly handles memory regions (MRs) with page sizes different from the system PAGE_SIZE. The core issue is that rxe_set_page() is called with mr->page_size step increments, but the page_list stores individual struct page pointers, each representing PAGE_SIZE of memory. ib_sg_to_page() has ensured that when i>=1 either a) SG[i-1].dma_end and SG[i].dma_addr are contiguous or b) SG[i-1].dma_end and SG[i].dma_addr are mr->page_size aligned. This leads to incorrect iova-to-va conversion in scenarios: 1) page_size < PAGE_SIZE (e.g., MR: 4K, system: 64K): ibmr->iova = 0x181800 sg[0]: dma_addr=0x181800, len=0x800 sg[1]: dma_addr=0x173000, len=0x1000 Access iova = 0x181800 + 0x810 = 0x182010 Expected VA: 0x173010 (second SG, offset 0x10) Before fix: - index = (0x182010 >> 12) - (0x181800 >> 12) = 1 - page_offset = 0x182010 & 0xFFF = 0x10 - xarray[1] stores system page base 0x170000 - Resulting VA: 0x170000 + 0x10 = 0x170010 (wrong) 2) page_size > PAGE_SIZE (e.g., MR: 64K, system: 4K): ibmr->iova = 0x18f800 sg[0]: dma_addr=0x18f800, len=0x800 sg[1]: dma_addr=0x170000, len=0x1000 Access iova = 0x18f800 + 0x810 = 0x190010 Expected VA: 0x170010 (second SG, offset 0x10) Before fix: - index = (0x190010 >> 16) - (0x18f800 >> 16) = 1 - page_offset = 0x190010 & 0xFFFF = 0x10 - xarray[1] stores system page for dma_addr 0x170000 - Resulting VA: system page of 0x170000 + 0x10 = 0x170010 (wrong) Yi Zhang reported a kernel panic[1] years ago related to this defect. Solution: 1. Replace xarray with pre-allocated rxe_mr_page array for sequential indexing (all MR page indices are contiguous) 2. Each rxe_mr_page stores both struct page* and offset within the system page 3. Handle MR page_size != PAGE_SIZE relationships: - page_size > PAGE_SIZE: Split MR pages into multiple system pages - page_size <= PAGE_SIZE: Store offset within system page 4. Add boundary checks and compatibility validation This ensures correct iova-to-va conversion regardless of MR page size and system PAGE_SIZE relationship, while improving performance through array-based sequential access. Tests on 4K and 64K PAGE_SIZE hosts: - rdma-core/pytests $ ./build/bin/run_tests.py --dev eth0_rxe - blktest: $ TIMEOUT=30 QUICK_RUN=1 USE_RXE=1 NVMET_TRTYPES=rdma ./check nvme srp rnbd [1] https://lore.kernel.org/all/CAHj4cs9XRqE25jyVw9rj9YugffLn5+f=1znaBEnu1usLOciD+g@mail.gmail.com/T/

In the Linux kernel, the following vulnerability has been resolved: net: stmmac: Prevent NULL deref when RX memory exhausted The CPU receives frames from the MAC through conventional DMA: the CPU allocates buffers for the MAC, then the MAC fills them and returns ownership to the CPU. For each hardware RX queue, the CPU and MAC coordinate through a shared ring array of DMA descriptors: one descriptor per DMA buffer. Each descriptor includes the buffer's physical address and a status flag ("OWN") indicating which side owns the buffer: OWN=0 for CPU, OWN=1 for MAC. The CPU is only allowed to set the flag and the MAC is only allowed to clear it, and both must move through the ring in sequence: thus the ring is used for both "submissions" and "completions." In the stmmac driver, stmmac_rx() bookmarks its position in the ring with the `cur_rx` index. The main receive loop in that function checks for rx_descs[cur_rx].own=0, gives the corresponding buffer to the network stack (NULLing the pointer), and increments `cur_rx` modulo the ring size. After the loop exits, stmmac_rx_refill(), which bookmarks its position with `dirty_rx`, allocates fresh buffers and rearms the descriptors (setting OWN=1). If it fails any allocation, it simply stops early (leaving OWN=0) and will retry where it left off when next called. This means descriptors have a three-stage lifecycle (terms my own): - `empty` (OWN=1, buffer valid) - `full` (OWN=0, buffer valid and populated) - `dirty` (OWN=0, buffer NULL) But because stmmac_rx() only checks OWN, it confuses `full`/`dirty`. In the past (see 'Fixes:'), there was a bug where the loop could cycle `cur_rx` all the way back to the first descriptor it dirtied, resulting in a NULL dereference when mistaken for `full`. The aforementioned commit resolved that *specific* failure by capping the loop's iteration limit at `dma_rx_size - 1`, but this is only a partial fix: if the previous stmmac_rx_refill() didn't complete, then there are leftover `dirty` descriptors that the loop might encounter without needing to cycle fully around. The current code therefore panics (see 'Closes:') when stmmac_rx_refill() is memory-starved long enough for `cur_rx` to catch up to `dirty_rx`. Fix this by explicitly checking, before advancing `cur_rx`, if the next entry is dirty; exit the loop if so. This prevents processing of the final, used descriptor until stmmac_rx_refill() succeeds, but fully prevents the `cur_rx == dirty_rx` ambiguity as the previous bugfix intended: so remove the clamp as well. Since stmmac_rx_zc() is a copy-paste-and-tweak of stmmac_rx() and the code structure is identical, any fix to stmmac_rx() will also need a corresponding fix for stmmac_rx_zc(). Therefore, apply the same check there. In stmmac_rx() (not stmmac_rx_zc()), a related bug remains: after the MAC sets OWN=0 on the final descriptor, it will be unable to send any further DMA-complete IRQs until it's given more `empty` descriptors. Currently, the driver simply *hopes* that the next stmmac_rx_refill() succeeds, risking an indefinite stall of the receive process if not. But this is not a regression, so it can be addressed in a future change.

In the Linux kernel, the following vulnerability has been resolved: ext4: fix e4b bitmap inconsistency reports A bitmap inconsistency issue was observed during stress tests under mixed huge-page workloads. Ext4 reported multiple e4b bitmap check failures like: ext4_mb_complex_scan_group:2508: group 350, 8179 free clusters as per group info. But got 8192 blocks Analysis and experimentation confirmed that the issue is caused by a race condition between page migration and bitmap modification. Although this timing window is extremely narrow, it is still hit in practice: folio_lock ext4_mb_load_buddy __migrate_folio check ref count folio_mc_copy __filemap_get_folio folio_try_get(folio) ...... mb_mark_used ext4_mb_unload_buddy __folio_migrate_mapping folio_ref_freeze folio_unlock The root cause of this issue is that the fast path of load_buddy only increments the folio's reference count, which is insufficient to prevent concurrent folio migration. We observed that the folio migration process acquires the folio lock. Therefore, we can determine whether to take the fast path in load_buddy by checking the lock status. If the folio is locked, we opt for the slow path (which acquires the lock) to close this concurrency window. Additionally, this change addresses the following issues: When the DOUBLE_CHECK macro is enabled to inspect bitmap-related issues, the following error may be triggered: corruption in group 324 at byte 784(6272): f in copy != ff on disk/prealloc Analysis reveals that this is a false positive. There is a specific race window where the bitmap and the group descriptor become momentarily inconsistent, leading to this error report: ext4_mb_load_buddy ext4_mb_load_buddy __filemap_get_folio(create|lock) folio_lock ext4_mb_init_cache folio_mark_uptodate __filemap_get_folio(no lock) ...... mb_mark_used mb_mark_used_double mb_cmp_bitmaps mb_set_bits(e4b->bd_bitmap) folio_unlock The original logic assumed that since mb_cmp_bitmaps is called when the bitmap is newly loaded from disk, the folio lock would be sufficient to prevent concurrent access. However, this overlooks a specific race condition: if another process attempts to load buddy and finds the folio is already in an uptodate state, it will immediately begin using it without holding folio lock.

In the Linux kernel, the following vulnerability has been resolved: sched/rt: Skip currently executing CPU in rto_next_cpu() CPU0 becomes overloaded when hosting a CPU-bound RT task, a non-CPU-bound RT task, and a CFS task stuck in kernel space. When other CPUs switch from RT to non-RT tasks, RT load balancing (LB) is triggered; with HAVE_RT_PUSH_IPI enabled, they send IPIs to CPU0 to drive the execution of rto_push_irq_work_func. During push_rt_task on CPU0, if next_task->prio < rq->donor->prio, resched_curr() sets NEED_RESCHED and after the push operation completes, CPU0 calls rto_next_cpu(). Since only CPU0 is overloaded in this scenario, rto_next_cpu() should ideally return -1 (no further IPI needed). However, multiple CPUs invoking tell_cpu_to_push() during LB increments rd->rto_loop_next. Even when rd->rto_cpu is set to -1, the mismatch between rd->rto_loop and rd->rto_loop_next forces rto_next_cpu() to restart its search from -1. With CPU0 remaining overloaded (satisfying rt_nr_migratory && rt_nr_total > 1), it gets reselected, causing CPU0 to queue irq_work to itself and send self-IPIs repeatedly. As long as CPU0 stays overloaded and other CPUs run pull_rt_tasks(), it falls into an infinite self-IPI loop, which triggers a CPU hardlockup due to continuous self-interrupts. The trigging scenario is as follows: cpu0 cpu1 cpu2 pull_rt_task tell_cpu_to_push <------------irq_work_queue_on rto_push_irq_work_func push_rt_task resched_curr(rq) pull_rt_task rto_next_cpu tell_cpu_to_push <-------------------------- atomic_inc(rto_loop_next) rd->rto_loop != next rto_next_cpu irq_work_queue_on rto_push_irq_work_func Fix redundant self-IPI by filtering the initiating CPU in rto_next_cpu(). This solution has been verified to effectively eliminate spurious self-IPIs and prevent CPU hardlockup scenarios.

In the Linux kernel, the following vulnerability has been resolved: net/sched: taprio: fix NULL pointer dereference in class dump When a TAPRIO child qdisc is deleted via RTM_DELQDISC, taprio_graft() is called with new == NULL and stores NULL into q->qdiscs[cl - 1]. Subsequent RTM_GETTCLASS dump operations walk all classes via taprio_walk() and call taprio_dump_class(), which calls taprio_leaf() returning the NULL pointer, then dereferences it to read child->handle, causing a kernel NULL pointer dereference. The bug is reachable with namespace-scoped CAP_NET_ADMIN on any kernel with CONFIG_NET_SCH_TAPRIO enabled. On systems with unprivileged user namespaces enabled, an unprivileged local user can trigger a kernel panic by creating a taprio qdisc inside a new network namespace, grafting an explicit child qdisc, deleting it, and requesting a class dump. The RTM_GETTCLASS dump itself requires no capability. Oops: general protection fault, probably for non-canonical address 0xdffffc0000000007: 0000 [#1] SMP KASAN NOPTI KASAN: null-ptr-deref in range [0x0000000000000038-0x000000000000003f] RIP: 0010:taprio_dump_class (net/sched/sch_taprio.c:2478) Call Trace: <TASK> tc_fill_tclass (net/sched/sch_api.c:1966) qdisc_class_dump (net/sched/sch_api.c:2326) taprio_walk (net/sched/sch_taprio.c:2514) tc_dump_tclass_qdisc (net/sched/sch_api.c:2352) tc_dump_tclass_root (net/sched/sch_api.c:2370) tc_dump_tclass (net/sched/sch_api.c:2431) rtnl_dumpit (net/core/rtnetlink.c:6864) netlink_dump (net/netlink/af_netlink.c:2325) rtnetlink_rcv_msg (net/core/rtnetlink.c:6959) netlink_rcv_skb (net/netlink/af_netlink.c:2550) </TASK> Fix this by substituting &noop_qdisc when new is NULL in taprio_graft(), a common pattern used by other qdiscs (e.g., multiq_graft()) to ensure the q->qdiscs[] slots are never NULL. This makes control-plane dump paths safe without requiring individual NULL checks. Since the data-plane paths (taprio_enqueue and taprio_dequeue_from_txq) previously had explicit NULL guards that would drop/skip the packet cleanly, update those checks to test for &noop_qdisc instead. Without this, packets would reach taprio_enqueue_one() which increments the root qdisc's qlen and backlog before calling the child's enqueue; noop_qdisc drops the packet but those counters are never rolled back, permanently inflating the root qdisc's statistics. After this change *old can be a valid qdisc, NULL, or &noop_qdisc. Only call qdisc_put(*old) in the first case to avoid decreasing noop_qdisc's refcount, which was never increased.

libheif is a HEIF and AVIF file format decoder and encoder. In versions 1.21.2 and prior, a crafted HEIF sequence file where the saiz box declares more samples than actually exist in the track's chunk table causes a heap-buffer-overflow (out-of-bounds read) in the SampleAuxInfoReader constructor. The SampleAuxInfoReader constructor iterates over saiz->get_num_samples() samples but doesn't validate that this count is consistent with the number of chunks in the chunks vector. When saiz declares more samples than the chunks cover, the loop increments current_chunk past chunks.size(), causing an out-of-bounds read on the chunks vector. The vulnerability is triggered during file parsing (heif_context_read_from_file) without any additional user interaction. Any application using libheif to open untrusted HEIF files is affected. This issue has been fixed in version 1.22.0.

struktur

NLnet Labs Unbound up to and including version 1.25.0 has a vulnerability when handling replies with very large RRsets that Unbound needs to perform name compression for. Malicious upstream responses with very large RRsets with records that don't share a suffix above the root can cause Unbound to spend a considerable time applying name compression to downstream replies. This can lead to degraded performance and eventually denial of service in well orchestrated attacks. An adversary can exploit the vulnerability by querying Unbound for the specially crafted contents of a malicious zone with very large RRsets. Before Unbound replies to the query it will try to apply name compression which was an unbounded operation that could lock the CPU until the whole packet was complete. A compression limit was introduced in 1.21.1 for this but it didn't account for the case where records would not share any suffix above the root. That causes Unbound to go in a different code path because of the compression tree lookup failure and eventually not increment the compression counter for those operations. Unbound 1.25.1 contains a patch with a fix that increments the compression counter regardless of the compression tree lookup. This is a complement fix to CVE-2024-8508.

nlnetlabs

In the Linux kernel, the following vulnerability has been resolved: usb: gadget: f_subset: Fix unbalanced refcnt in geth_free geth_alloc() increments the reference count, but geth_free() fails to decrement it. This prevents the configuration of attributes via configfs after unlinking the function. Decrement the reference count in geth_free() to ensure proper cleanup.

linux

In the Linux kernel, the following vulnerability has been resolved: mm/hugetlb: restore failed global reservations to subpool Commit a833a693a490 ("mm: hugetlb: fix incorrect fallback for subpool") fixed an underflow error for hstate->resv_huge_pages caused by incorrectly attributing globally requested pages to the subpool's reservation. Unfortunately, this fix also introduced the opposite problem, which would leave spool->used_hpages elevated if the globally requested pages could not be acquired. This is because while a subpool's reserve pages only accounts for what is requested and allocated from the subpool, its "used" counter keeps track of what is consumed in total, both from the subpool and globally. Thus, we need to adjust spool->used_hpages in the other direction, and make sure that globally requested pages are uncharged from the subpool's used counter. Each failed allocation attempt increments the used_hpages counter by how many pages were requested from the global pool. Ultimately, this renders the subpool unusable, as used_hpages approaches the max limit. The issue can be reproduced as follows: 1. Allocate 4 hugetlb pages 2. Create a hugetlb mount with max=4, min=2 3. Consume 2 pages globally 4. Request 3 pages from the subpool (2 from subpool + 1 from global) 4.1 hugepage_subpool_get_pages(spool, 3) succeeds. used_hpages += 3 4.2 hugetlb_acct_memory(h, 1) fails: no global pages left used_hpages -= 2 5. Subpool now has used_hpages = 1, despite not being able to successfully allocate any hugepages. It believes it can now only allocate 3 more hugepages, not 4. With each failed allocation attempt incrementing the used counter, the subpool eventually reaches a point where its used counter equals its max counter. At that point, any future allocations that try to allocate hugeTLB pages from the subpool will fail, despite the subpool not having any of its hugeTLB pages consumed by any user. Once this happens, there is no way to make the subpool usable again, since there is no way to decrement the used counter as no process is really consuming the hugeTLB pages. The underflow issue that the original commit fixes still remains fixed as well. Without this fix, used_hpages would keep on leaking if hugetlb_acct_memory() fails.

linux

In the Linux kernel, the following vulnerability has been resolved: io_uring/net: fix slab-out-of-bounds read in io_bundle_nbufs() sqe->len is __u32 but gets stored into sr->len which is int. When userspace passes sqe->len values exceeding INT_MAX (e.g. 0xFFFFFFFF), sr->len overflows to a negative value. This negative value propagates through the bundle recv/send path: 1. io_recv(): sel.val = sr->len (ssize_t gets -1) 2. io_recv_buf_select(): arg.max_len = sel->val (size_t gets 0xFFFFFFFFFFFFFFFF) 3. io_ring_buffers_peek(): buf->len is not clamped because max_len is astronomically large 4. iov[].iov_len = 0xFFFFFFFF flows into io_bundle_nbufs() 5. io_bundle_nbufs(): min_t(int, 0xFFFFFFFF, ret) yields -1, causing ret to increase instead of decrease, creating an infinite loop that reads past the allocated iov[] array This results in a slab-out-of-bounds read in io_bundle_nbufs() from the kmalloc-64 slab, as nbufs increments past the allocated iovec entries. BUG: KASAN: slab-out-of-bounds in io_bundle_nbufs+0x128/0x160 Read of size 8 at addr ffff888100ae05c8 by task exp/145 Call Trace: io_bundle_nbufs+0x128/0x160 io_recv_finish+0x117/0xe20 io_recv+0x2db/0x1160 Fix this by rejecting negative sr->len values early in both io_sendmsg_prep() and io_recvmsg_prep(). Since sqe->len is __u32, any value > INT_MAX indicates overflow and is not a valid length.

linux

In the Linux kernel, the following vulnerability has been resolved: gpib: fix use-after-free in IO ioctl handlers The IBRD, IBWRT, IBCMD, and IBWAIT ioctl handlers use a gpib_descriptor pointer after board->big_gpib_mutex has been released. A concurrent IBCLOSEDEV ioctl can free the descriptor via close_dev_ioctl() during this window, causing a use-after-free. The IO handlers (read_ioctl, write_ioctl, command_ioctl) explicitly release big_gpib_mutex before calling their handler. wait_ioctl() is called with big_gpib_mutex held, but ibwait() releases it internally when wait_mask is non-zero. In all four cases, the descriptor pointer obtained from handle_to_descriptor() becomes unprotected. Fix this by introducing a kernel-only descriptor_busy reference count in struct gpib_descriptor. Each handler atomically increments descriptor_busy under file_priv->descriptors_mutex before releasing the lock, and decrements it when done. close_dev_ioctl() checks descriptor_busy under the same lock and rejects the close with -EBUSY if the count is non-zero. A reference count rather than a simple flag is necessary because multiple handlers can operate on the same descriptor concurrently (e.g. IBRD and IBWAIT on the same handle from different threads). A separate counter is needed because io_in_progress can be cleared from unprivileged userspace via the IBWAIT ioctl (through general_ibstatus() with set_mask containing CMPL), which would allow an attacker to bypass a check based solely on io_in_progress. The new descriptor_busy counter is only modified by the kernel IO paths. The lock ordering is consistent (big_gpib_mutex -> descriptors_mutex) and the handlers only hold descriptors_mutex briefly during the lookup, so there is no deadlock risk and no impact on IO throughput.

linux

In the Linux kernel, the following vulnerability has been resolved: smb: server: fix active_num_conn leak on transport allocation failure Commit 77ffbcac4e56 ("smb: server: fix leak of active_num_conn in ksmbd_tcp_new_connection()") addressed the kthread_run() failure path. The earlier alloc_transport() == NULL path in the same function has the same leak, is reachable pre-authentication via any TCP connect to port 445, and was empirically reproduced on UML (ARCH=um, v7.0-rc7): a small number of forced allocation failures were sufficient to put ksmbd into a state where every subsequent connection attempt was rejected for the remainder of the boot. ksmbd_kthread_fn() increments active_num_conn before calling ksmbd_tcp_new_connection() and discards the return value, so when alloc_transport() returns NULL the socket is released and -ENOMEM returned without decrementing the counter. Each such failure permanently consumes one slot from the max_connections pool; once cumulative failures reach the cap, atomic_inc_return() hits the threshold on every subsequent accept and every new connection is rejected. The counter is only reset by module reload. An unauthenticated remote attacker can drive the server toward the memory pressure that makes alloc_transport() fail by holding open connections with large RFC1002 lengths up to MAX_STREAM_PROT_LEN (0x00FFFFFF); natural transient allocation failures on a loaded host produce the same drift more slowly. Mirror the existing rollback pattern in ksmbd_kthread_fn(): on the alloc_transport() failure path, decrement active_num_conn gated on server_conf.max_connections. Repro details: with the patch reverted, forced alloc_transport() NULL returns leaked counter slots and subsequent connection attempts -- including legitimate connects issued after the forced-fail window had closed -- were all rejected with "Limit the maximum number of connections". With this patch applied, the same connect sequence produces no rejections and the counter cycles cleanly between zero and one on every accept.

linux

In the Linux kernel, the following vulnerability has been resolved: net/smc: fix double-free of smc_spd_priv when tee() duplicates splice pipe buffer smc_rx_splice() allocates one smc_spd_priv per pipe_buffer and stores the pointer in pipe_buffer.private. The pipe_buf_operations for these buffers used .get = generic_pipe_buf_get, which only increments the page reference count when tee(2) duplicates a pipe buffer. The smc_spd_priv pointer itself was not handled, so after tee() both the original and the cloned pipe_buffer share the same smc_spd_priv *. When both pipes are subsequently released, smc_rx_pipe_buf_release() is called twice against the same object: 1st call: kfree(priv) sock_put(sk) smc_rx_update_cons() [correct] 2nd call: kfree(priv) sock_put(sk) smc_rx_update_cons() [UAF] KASAN reports a slab-use-after-free in smc_rx_pipe_buf_release(), which then escalates to a NULL-pointer dereference and kernel panic via smc_rx_update_consumer() when it chases the freed priv->smc pointer: BUG: KASAN: slab-use-after-free in smc_rx_pipe_buf_release+0x78/0x2a0 Read of size 8 at addr ffff888004a45740 by task smc_splice_tee_/74 Call Trace: <TASK> dump_stack_lvl+0x53/0x70 print_report+0xce/0x650 kasan_report+0xc6/0x100 smc_rx_pipe_buf_release+0x78/0x2a0 free_pipe_info+0xd4/0x130 pipe_release+0x142/0x160 __fput+0x1c6/0x490 __x64_sys_close+0x4f/0x90 do_syscall_64+0xa6/0x1a0 entry_SYSCALL_64_after_hwframe+0x77/0x7f </TASK> BUG: kernel NULL pointer dereference, address: 0000000000000020 RIP: 0010:smc_rx_update_consumer+0x8d/0x350 Call Trace: <TASK> smc_rx_pipe_buf_release+0x121/0x2a0 free_pipe_info+0xd4/0x130 pipe_release+0x142/0x160 __fput+0x1c6/0x490 __x64_sys_close+0x4f/0x90 do_syscall_64+0xa6/0x1a0 entry_SYSCALL_64_after_hwframe+0x77/0x7f </TASK> Kernel panic - not syncing: Fatal exception Beyond the memory-safety problem, duplicating an SMC splice buffer is semantically questionable: smc_rx_update_cons() would advance the consumer cursor twice for the same data, corrupting receive-window accounting. A refcount on smc_spd_priv could fix the double-free, but the cursor-accounting issue would still need to be addressed separately. The .get callback is invoked by both tee(2) and splice_pipe_to_pipe() for partial transfers; both will now return -EFAULT. Users who need to duplicate SMC socket data must use a copy-based read path.

linux

In the Linux kernel, the following vulnerability has been resolved: net: fix fanout UAF in packet_release() via NETDEV_UP race `packet_release()` has a race window where `NETDEV_UP` can re-register a socket into a fanout group's `arr[]` array. The re-registration is not cleaned up by `fanout_release()`, leaving a dangling pointer in the fanout array. `packet_release()` does NOT zero `po->num` in its `bind_lock` section. After releasing `bind_lock`, `po->num` is still non-zero and `po->ifindex` still matches the bound device. A concurrent `packet_notifier(NETDEV_UP)` that already found the socket in `sklist` can re-register the hook. For fanout sockets, this re-registration calls `__fanout_link(sk, po)` which adds the socket back into `f->arr[]` and increments `f->num_members`, but does NOT increment `f->sk_ref`. The fix sets `po->num` to zero in `packet_release` while `bind_lock` is held to prevent NETDEV_UP from linking, preventing the race window. This bug was found following an additional audit with Claude Code based on CVE-2025-38617.

linux

In the Linux kernel, the following vulnerability has been resolved: ipv6: avoid overflows in ip6_datagram_send_ctl() Yiming Qian reported : <quote> I believe I found a locally triggerable kernel bug in the IPv6 sendmsg ancillary-data path that can panic the kernel via `skb_under_panic()` (local DoS). The core issue is a mismatch between: - a 16-bit length accumulator (`struct ipv6_txoptions::opt_flen`, type `__u16`) and - a pointer to the *last* provided destination-options header (`opt->dst1opt`) when multiple `IPV6_DSTOPTS` control messages (cmsgs) are provided. - `include/net/ipv6.h`: - `struct ipv6_txoptions::opt_flen` is `__u16` (wrap possible). (lines 291-307, especially 298) - `net/ipv6/datagram.c:ip6_datagram_send_ctl()`: - Accepts repeated `IPV6_DSTOPTS` and accumulates into `opt_flen` without rejecting duplicates. (lines 909-933) - `net/ipv6/ip6_output.c:__ip6_append_data()`: - Uses `opt->opt_flen + opt->opt_nflen` to compute header sizes/headroom decisions. (lines 1448-1466, especially 1463-1465) - `net/ipv6/ip6_output.c:__ip6_make_skb()`: - Calls `ipv6_push_frag_opts()` if `opt->opt_flen` is non-zero. (lines 1930-1934) - `net/ipv6/exthdrs.c:ipv6_push_frag_opts()` / `ipv6_push_exthdr()`: - Push size comes from `ipv6_optlen(opt->dst1opt)` (based on the pointed-to header). (lines 1179-1185 and 1206-1211) 1. `opt_flen` is a 16-bit accumulator: - `include/net/ipv6.h:298` defines `__u16 opt_flen; /* after fragment hdr */`. 2. `ip6_datagram_send_ctl()` accepts *repeated* `IPV6_DSTOPTS` cmsgs and increments `opt_flen` each time: - In `net/ipv6/datagram.c:909-933`, for `IPV6_DSTOPTS`: - It computes `len = ((hdr->hdrlen + 1) << 3);` - It checks `CAP_NET_RAW` using `ns_capable(net->user_ns, CAP_NET_RAW)`. (line 922) - Then it does: - `opt->opt_flen += len;` (line 927) - `opt->dst1opt = hdr;` (line 928) There is no duplicate rejection here (unlike the legacy `IPV6_2292DSTOPTS` path which rejects duplicates at `net/ipv6/datagram.c:901-904`). If enough large `IPV6_DSTOPTS` cmsgs are provided, `opt_flen` wraps while `dst1opt` still points to a large (2048-byte) destination-options header. In the attached PoC (`poc.c`): - 32 cmsgs with `hdrlen=255` => `len = (255+1)*8 = 2048` - 1 cmsg with `hdrlen=0` => `len = 8` - Total increment: `32*2048 + 8 = 65544`, so `(__u16)opt_flen == 8` - The last cmsg is 2048 bytes, so `dst1opt` points to a 2048-byte header. 3. The transmit path sizes headers using the wrapped `opt_flen`: - In `net/ipv6/ip6_output.c:1463-1465`: - `headersize = sizeof(struct ipv6hdr) + (opt ? opt->opt_flen + opt->opt_nflen : 0) + ...;` With wrapped `opt_flen`, `headersize`/headroom decisions underestimate what will be pushed later. 4. When building the final skb, the actual push length comes from `dst1opt` and is not limited by wrapped `opt_flen`: - In `net/ipv6/ip6_output.c:1930-1934`: - `if (opt->opt_flen) proto = ipv6_push_frag_opts(skb, opt, proto);` - In `net/ipv6/exthdrs.c:1206-1211`, `ipv6_push_frag_opts()` pushes `dst1opt` via `ipv6_push_exthdr()`. - In `net/ipv6/exthdrs.c:1179-1184`, `ipv6_push_exthdr()` does: - `skb_push(skb, ipv6_optlen(opt));` - `memcpy(h, opt, ipv6_optlen(opt));` With insufficient headroom, `skb_push()` underflows and triggers `skb_under_panic()` -> `BUG()`: - `net/core/skbuff.c:2669-2675` (`skb_push()` calls `skb_under_panic()`) - `net/core/skbuff.c:207-214` (`skb_panic()` ends in `BUG()`) - The `IPV6_DSTOPTS` cmsg path requires `CAP_NET_RAW` in the target netns user namespace (`ns_capable(net->user_ns, CAP_NET_RAW)`). - Root (or any task with `CAP_NET_RAW`) can trigger this without user namespaces. - An unprivileged `uid=1000` user can trigger this if unprivileged user namespaces are enabled and it can create a userns+netns to obtain namespaced `CAP_NET_RAW` (the attached PoC does this). - Local denial of service: kernel BUG/panic (system crash). - ---truncated---

linux

Vikunja is an open-source self-hosted task management platform. Prior to 2.3.0, the TOTP failed-attempt lockout mechanism is non-functional due to a database transaction handling bug. When a TOTP validation fails, the login handler in pkg/routes/api/v1/login.go calls HandleFailedTOTPAuth and then unconditionally rolls back. HandleFailedTOTPAuth in pkg/user/totp.go uses an in-memory counter (key-value store) to track failed attempts. When the counter reaches 10, it calls user.SetStatus(s, StatusAccountLocked) on the same database session s. Because the login handler always rolls back after a TOTP failure, the StatusAccountLocked write is undone. The in-memory counter correctly increments past 10, so the lockout code executes on every subsequent attempt, but the database write is rolled back every time. This allows unlimited brute-force attempts against TOTP codes. This vulnerability is fixed in 2.3.0.

vikunja

In the Linux kernel, the following vulnerability has been resolved: fs: ntfs3: fix infinite loop in attr_load_runs_range on inconsistent metadata We found an infinite loop bug in the ntfs3 file system that can lead to a Denial-of-Service (DoS) condition. A malformed NTFS image can cause an infinite loop when an attribute header indicates an empty run list, while directory entries reference it as containing actual data. In NTFS, setting evcn=-1 with svcn=0 is a valid way to represent an empty run list, and run_unpack() correctly handles this by checking if evcn + 1 equals svcn and returning early without parsing any run data. However, this creates a problem when there is metadata inconsistency, where the attribute header claims to be empty (evcn=-1) but the caller expects to read actual data. When run_unpack() immediately returns success upon seeing this condition, it leaves the runs_tree uninitialized with run->runs as a NULL. The calling function attr_load_runs_range() assumes that a successful return means that the runs were loaded and sets clen to 0, expecting the next run_lookup_entry() call to succeed. Because runs_tree remains uninitialized, run_lookup_entry() continues to fail, and the loop increments vcn by zero (vcn += 0), leading to an infinite loop. This patch adds a retry counter to detect when run_lookup_entry() fails consecutively after attr_load_runs_vcn(). If the run is still not found on the second attempt, it indicates corrupted metadata and returns -EINVAL, preventing the Denial-of-Service (DoS) vulnerability.

linux

In the Linux kernel, the following vulnerability has been resolved: dpll: Prevent duplicate registrations Modify the internal registration helpers dpll_xa_ref_{dpll,pin}_add() to reject duplicate registration attempts. Previously, if a caller attempted to register the same pin multiple times (with the same ops, priv, and cookie) on the same device, the core silently increments the reference count and return success. This behavior is incorrect because if the caller makes these duplicate registrations then for the first one dpll_pin_registration is allocated and for others the associated dpll_pin_ref.refcount is incremented. During the first unregistration the associated dpll_pin_registration is freed and for others WARN is fired. Fix this by updating the logic to return `-EEXIST` if a matching registration is found to enforce a strict "register once" policy.

linux

NanoMQ MQTT Broker (NanoMQ) is an all-around Edge Messaging Platform. In version 0.24.6, NanoMQ has a protocol parsing / forwarding inconsistency when handling shared subscriptions ($share/). A malformed SUBSCRIBE topic such as $share/ab (missing the second /) is not strictly validated during the subscription stage, so the invalid Topic Filter is stored into the subscription table. Later, when any PUBLISH matches this subscription, the broker send path (nmq_pipe_send_start_v4/v5) performs a second $share/ parsing using strchr() and increments the returned pointer without NULL checks. If the second strchr() returns NULL, sub_topic++ turns the pointer into an invalid address (e.g. 0x1). This invalid pointer is then passed into topic_filtern(), which triggers strlen() and crashes with SIGSEGV. The crash is stable and remotely triggerable. This issue has been patched in version 0.24.7.

emqx

In the Linux kernel, the following vulnerability has been resolved: soundwire: bus: Fix unbalanced pm_runtime_put() causing usage count underflow This reverts commit 443a98e649b4 ("soundwire: bus: use pm_runtime_resume_and_get()") Change calls to pm_runtime_resume_and_get() back to pm_runtime_get_sync(). This fixes a usage count underrun caused by doing a pm_runtime_put() even though pm_runtime_resume_and_get() returned an error. The three affected functions ignore -EACCES error from trying to get pm_runtime, and carry on, including a put at the end of the function. But pm_runtime_resume_and_get() does not increment the usage count if it returns an error. So in the -EACCES case you must not call pm_runtime_put(). The documentation for pm_runtime_get_sync() says: "Consider using pm_runtime_resume_and_get() ... as this is likely to result in cleaner code." In this case I don't think it results in cleaner code because the pm_runtime_put() at the end of the function would have to be conditional on the return value from pm_runtime_resume_and_get() at the top of the function. pm_runtime_get_sync() doesn't have this problem because it always increments the count, so always needs a put. The code can just flow through and do the pm_runtime_put() unconditionally.