Mercredi 26 août 2026

Linux

In the Linux kernel, the following vulnerability has been resolved: mptcp: reclaim forward-allocated memory on RX path errors After commit 9db5b3cec4ec ("mptcp: borrow forward memory from subflow"), errors in the receive path prior to queueing skbs into the receive queue do not trigger forward-allocated memory reclaiming. Prevent forward memory from growing unboundedly in pathological drop scenarios by explicitly reclaiming memory when skbs are dropped.

Linux

In the Linux kernel, the following vulnerability has been resolved: s390/qeth: validate user buffer length in SNMP and ARP query ioctls qeth_snmp_command() and qeth_l3_arp_query() allocate a buffer sized by a user-supplied length (udata_len) without checking a lower bound, then set udata_offset to a fixed non-zero value and pass both to a reply callback. The callback bounds-checks the copy with if ((udata_len - udata_offset) < len) Both fields are u32, so a udata_len smaller than udata_offset makes the subtraction wrap and the check pass, and the following memcpy() writes past the allocation. A udata_len of 0 also yields ZERO_SIZE_PTR from kzalloc(), which the existing NULL check does not catch. Reject buffers smaller than udata_offset before allocating, so the callback subtraction can no longer underflow.

Linux

In the Linux kernel, the following vulnerability has been resolved: ASoC: codecs: lpass-tx-macro: Fix enum kcontrol accesses The "DEC0 MODE" to "DEC7 MODE" controls are enumerated, but tx_macro_dec_mode_get() and tx_macro_dec_mode_put() access their value through ucontrol->value.integer.value[0] (a long) instead of ucontrol->value.enumerated.item[0] (an unsigned int). This same pattern was fixed in the sibling drivers by commit bcfe5f76cc40 ("ASoC: codecs: rx-macro: fix accessing array out of bounds for enum type") and commit 0ea5eff7c606 ("ASoC: codecs: va-macro: fix accessing array out of bounds for enum type"), but tx-macro was missed. On 64-bit kernels built with CONFIG_SND_CTL_DEBUG, the elem value sanity check catches the 4 bytes written past the enumerated item and every read of these controls fails with -EINVAL: snd-sm8250 sound: control 2:0:0:DEC0 MODE:0: access overflow

Linux

In the Linux kernel, the following vulnerability has been resolved: drm/shmem_helper: Check VMA boundaries for PMD mappings In the ->huge_fault handler do not install a PMD huge page mapping if the huge page exceeds the boundaries of the VMA. All other ->huge_fault handlers have similar checks and the resulting mapping will trigger a VM_BUG_ON_VMA() if it ever reaches copy_pmd_range().

Linux

In the Linux kernel, the following vulnerability has been resolved: fbdev: bound mode sysfs output to the sysfs buffer mode_string() uses snprintf() which can return a value larger than the remaining buffer space. show_modes() accumulates the return value into i without checking whether i has reached PAGE_SIZE, causing the offset to advance past the sysfs buffer if the modelist is long enough. Add a size parameter to mode_string() and use scnprintf() to return only the bytes actually written. Add an early return when offset already exceeds the buffer. In show_modes(), stop accumulating once the buffer is full.

Linux

In the Linux kernel, the following vulnerability has been resolved: fbdev: clear fb_info->mode before deleting a videomode fb_set_var() can delete a mode from info->modelist when userspace passes FB_ACTIVATE_INV_MODE through FBIOPUT_VSCREENINFO. The code checks that the mode being deleted is not the current info->var and that fbcon is not using it, but it does not check fb_info->mode. fb_info->mode may still point into the modelist entry being deleted. If the entry is freed, later mode sysfs reads through show_mode() can dereference a stale pointer. Clear fb_info->mode before calling fb_delete_videomode() when it matches the mode being removed.

Linux

In the Linux kernel, the following vulnerability has been resolved: fbdev: core: Fix pointer desynchronization in fb_io_read() In fb_io_read(), if copy_to_user() performs a partial copy (e.g., due to a faulty user buffer), the loop adjusts the chunk size 'c' and updates the remaining 'count'. However, the hardware 'src' pointer has already been eagerly advanced by the original chunk size. If the loop is allowed to continue, the read will resume from an incorrect, over-advanced offset. Since the remaining 'count' was only decremented by the successful bytes, this desynchronization causes the next iterations to execute more hardware reads than originally bounded, eventually leading to out-of-bounds I/O reads. Fix this by breaking out of the loop immediately upon a partial copy_to_user(). A partial copy indicates a faulty user buffer, making subsequent read attempts futile. Breaking out ensures we return the number of successfully read bytes without risking out-of-bounds hardware accesses in subsequent mismatched iterations.

Linux

In the Linux kernel, the following vulnerability has been resolved: drm/amdgpu: reject oversized IBs with per-ring packet limits On GFX rings, amdgpu_cs_p2_ib() passed user-supplied ib_bytes through to ib->length_dw without a limit, while ring_emit_ib() encodes length into packet fields. Oversized values can corrupt adjacent control bits and destabilize command submission. Add a per-ring IB packet size limit helper and reject command submissions exceeding the corresponding dword limit before IB allocation. Use the documented 20-bit limit for GFX/compute/SDMA/VPE, and apply the MM fallback limit for other ring types. (cherry picked from commit 7f48fa2cf62e3fa6c9c3870aa74988f773247e52)

Linux

In the Linux kernel, the following vulnerability has been resolved: Input: cs40l50-vibra - validate custom data from user space cs40l50_add() copies the custom data of an FF_PERIODIC/FF_CUSTOM effect straight from the ff_effect the user passed to EVIOCSFF, without requiring it to hold anything: work_data.custom_data = memdup_array_user(periodic->custom_data, periodic->custom_len, sizeof(s16)); work_data.custom_len = periodic->custom_len; The driver then reads two words out of that buffer: custom_data[0] as the waveform bank in cs40l50_effect_bank_set(), and custom_data[1] as the index within the bank in cs40l50_effect_index_set(). Neither read is covered by a length check, and custom_len is fully user controlled: - custom_len == 0 makes memdup_array_user() call memdup_user() with a length of zero, which returns ZERO_SIZE_PTR rather than an error, so custom_data[0] dereferences it. - custom_len == 1 allocates two bytes. A bank of ROM or RAM keeps effect->type out of the OWT case, and custom_data[1] is then read one word past the allocation. The bank value itself is also mishandled. It is masked with CS40L50_CUSTOM_DATA_MASK (0xffff) but stored in an s16, so a custom_data[0] of 0x8000 or above wraps to a negative value that passes the "bank_type >= CS40L50_WVFRM_BANK_NUM" test. cs40l50_effect_index_set() indexes vib->dsp.banks[] with it before the switch statement's default case gets a chance to reject it: base_index = vib->dsp.banks[effect->type].base_index; max_index = vib->dsp.banks[effect->type].max_index; Require the two words the driver reads to be present, and hold the masked bank in a u32 so the existing upper-bound test covers the whole range. The da7280 haptic driver already range checks custom_len this way.

Linux

In the Linux kernel, the following vulnerability has been resolved: Input: focaltech - fix array out-of-bounds in focaltech_process_rel_packet Make finger2 (and also finger1) unsigned, so that if the finger index in the packet is 0 then subtracting 1 creates an array index which overflows above the existing check for FOC_MAX_FINGERS, as the existing comment says it should, instead of writing to state->fingers[-1].

Linux

In the Linux kernel, the following vulnerability has been resolved: Input: byd - synchronize timer deletion before freeing private data byd_disconnect() uses timer_delete() before freeing the driver's private data. This does not wait for a running byd_clear_touch() callback, which dereferences the private data and its psmouse pointer. A callback racing with disconnect can therefore access the private data after it has been freed. The timer can also still be re-armed by byd_process_byte() while the disconnect is in progress. Use timer_shutdown_sync() before freeing the private data: it waits for a running callback and turns any later re-arm attempt into a no-op.

Linux

In the Linux kernel, the following vulnerability has been resolved: Input: synaptics-rmi4 - zero report size on F54 work error In rmi_f54_work(), if an error occurs during report request or command verification, the code jumped directly to the 'error' label, bypassing the 'abort' label where f54->report_size was normally zeroed out. This left f54->report_size containing its previous successful payload size. If a user then altered the V4L2 format to a smaller size, and a subsequent run failed, rmi_f54_buffer_queue() would copy the stale, larger payload size into the shrunken V4L2 buffer, causing a heap buffer overflow. Fix this by merging the 'abort' and 'error' labels into a single 'out' exit path, and ensuring that f54->report_size is always set to 0 on failure by checking for error and zeroing the local report_size first.

Linux

In the Linux kernel, the following vulnerability has been resolved: Input: synaptics-rmi4 - bound the F54 report size to the allocated buffer rmi_f54_work() reads a diagnostics report from the device into f54->report_data, sizing the transfer with rmi_f54_get_report_size(): report_size = rmi_f54_get_report_size(f54); ... for (i = 0; i < report_size; i += F54_REPORT_DATA_SIZE) { int size = min(F54_REPORT_DATA_SIZE, report_size - i); ... rmi_read_block(.., f54->report_data + i, size); } report_data is allocated once at probe from F54's own electrode counts (array3_size(f54->num_tx_electrodes, f54->num_rx_electrodes, sizeof(u16))), but rmi_f54_get_report_size() computes the size from drv_data->num_*_electrodes when those are set, i.e. from the F55 function's electrode counts. Both counts come straight from device queries (F54 and F55 each report up to 255 electrodes) and nothing constrains the F55 counts to the F54 ones. A malicious or malfunctioning RMI4 device that reports larger F55 electrode counts than its F54 counts makes report_size exceed the allocation, so the read loop writes past report_data (and the V4L2 dequeue memcpy() then reads past it). On conforming hardware the F55 configured electrodes are a subset of the F54 physical electrodes, so report_size never exceeds the buffer and well-behaved devices are unaffected. Record the allocation size and reject a report that does not fit, mirroring the existing zero-size check.

Linux

In the Linux kernel, the following vulnerability has been resolved: Input: synaptics-rmi4 - block s_input when F54 queue is busy Changing the input (diagnostic report type) mid-stream changes the report size. Since V4L2 buffers are allocated based on the size at stream start, changing the input while streaming could lead to a heap buffer overflow if the new size is larger than the allocated buffers. Prevent this by blocking VIDIOC_S_INPUT with -EBUSY if the V4L2 queue is busy (streaming).

Linux

In the Linux kernel, the following vulnerability has been resolved: crypto: qce - fix error path in devm_qce_register_algs If ops->register_algs() fails, the error path repeatedly calls the same ops->unregister_algs() from the failed registration. Use the loop index to unregister the previously registered algorithms instead.

Linux

In the Linux kernel, the following vulnerability has been resolved: openrisc: signal: do not restore privileged SR bits on sigreturn restore_sigcontext() copies the whole supervision register (SR) from the signal frame and only clears SPR_SR_SM before the value is reloaded into the hardware SR (through ESR and l.rfe) on the return to user space. All other SR bits are left under user control. An unprivileged task can thus return from a signal handler through a crafted sigframe that clears SPR_SR_DME. With the data MMU disabled the CPU performs no translation or protection on data accesses, so the task gains read and write access to arbitrary physical memory, a local privilege escalation. SPR_SR_IME, SPR_SR_SUMRA, SPR_SR_LEE, SPR_SR_EPH and the cache-enable bits are exposed the same way. The ptrace GPR regset already refuses any change to SR for exactly this reason. Restore only the arithmetic flag bits (F, CY, OV) from the signal frame and take every privileged control bit from the SR the kernel saved on signal entry. Verified with qemu-system-or1k -M or1k-sim: before this change an unprivileged PoC clears SPR_SR_DME in rt_sigreturn and writes a marker to physical address 0x03000000 (beyond the kernel's mem=32M); afterwards the same PoC receives SIGSEGV and physical memory is unchanged.

Linux

In the Linux kernel, the following vulnerability has been resolved: Input: sur40 - fix input device registration ordering In sur40_probe(), input_register_device() was previously called early before the V4L2 video device and vb2_queue components were fully initialized. If userspace opened the input device immediately upon registration, sur40_open() would trigger and start the sur40_poll() worker thread. This worker thread invokes sur40_process_video() and accesses the uninitialized vb2_queue structure, leading to a data race and potential system crash. Furthermore, if V4L2 or video registration failed after input_register_device() succeeded, the error path fell through to calling input_free_device() on a successfully registered device instead of input_unregister_device(), corrupting input core state. Move input_register_device() to the very end of sur40_probe(). This ensures the V4L2 and video queue structures are fully initialized before polling can start, and naturally resolves the error path bug since input_free_device() is now only called when input registration has not yet occurred. To maintain strict LIFO (Last-In, First-Out) teardown ordering, also move input_unregister_device() to the very beginning of sur40_disconnect(). This guarantees that the input polling worker thread is stopped before V4L2 video components or control handlers are unregistered.

Linux

In the Linux kernel, the following vulnerability has been resolved: mmc: atmel-mci: Fix use-after-free in atmci_remove due to race condition In atmci_probe, &host->bh_work is bound with atmci_work_func, and atmci_interrupt, atmci_timeout_timer and atmci_dma_complete can all queue this work on system_bh_wq. If we remove the module, atmci_remove makes cleanup and the memory allocated for host with devm_kzalloc() is released after the remove callback returns, while the work mentioned above may still be pending or running. The sequence of operations that may lead to a UAF bug is as follows: CPU0 CPU1 | atmci_interrupt | queue_work(system_bh_wq, | &host->bh_work) atmci_remove | atmci_cleanup_slot(...) | atmci_writel(host, ATMCI_IDR, ~0UL) | timer_delete_sync(&host->timer) | dma_release_channel(host->dma.chan) | free_irq(platform_get_irq(pdev, 0), host) | | atmci_work_func | // use host // devm resources released after | // remove returns, host is freed | | // use host (use-after-free) Fix it by canceling the work after all the sources that can schedule it (IRQ handler, timeout timer and DMA completion callback) have been stopped, and before proceeding with the remaining cleanup in atmci_remove.

Linux

In the Linux kernel, the following vulnerability has been resolved: s390/vfio_ccw: Free all memory if cp_init() fails The routine cp_free() is called to unpin/free any memory once an I/O is completed successfully, or if cp_prefetch() fails. But if cp_init() fails, and cp->initialized is not enabled, the same routine cannot be used to free all the memory. An attempt to address this exists in ccwchain_handle_ccw(), where a single call to ccwchain_free() is made for the currently-processed CCW segment. But this will leak other segments (created as a result of a Transfer in Channel) that had been allocated as part of the same channel program. Address this by performing the cleanup outside of the recursive ccwchain_handle_ccw()/ccwchain_loop_tic() logic.

Linux

In the Linux kernel, the following vulnerability has been resolved: s390/vfio_ccw: Cancel existing workqueues The initialization of the io_work and crw_work workqueues begs the question of whether they should be un-initialized. Add the corresponding cleanup tags in _release_dev to ensure work isn't dispatched after the private struct is free'd.

Linux

In the Linux kernel, the following vulnerability has been resolved: s390/vfio_ccw: Ensure index for read/write regions are within range The introduction of the capability chain rightly clamped the region indexes to the range of the capabilities itself, but neglected to do so for the existing read/write regions which should also be enforced.

Linux

In the Linux kernel, the following vulnerability has been resolved: s390/vfio_ccw: Fix out of bounds check on CCW array The routine ccwchain_calc_length() counts the number of channel command words (CCWs) that are chained together in a single channel program, and rejects anything larger than CCWCHAIN_LEN_MAX (256) CCWs. The loop itself is "do..while (count < 257)", and while the logic in is_cpa_within_range() correctly adjusts between the 0-index array of CCWs and the count of CCWs starting at 1, this means it would look at a possible 257th CCW before ending the loop and (correctly) returning an error. Fix this by restructuring the loop to break as soon as 256 CCWs (thus indexes 0-255) are examined, without looking at memory outside the range.

Linux

In the Linux kernel, the following vulnerability has been resolved: s390/vfio_ccw: Move cp cleanup out of not operational The fsm_notoper() routine is called when the device has been lost, and is (by definition) no longer operational. Since this can happen asynchronously from the normal behavior of the driver, the cleanup may happen when holding other locks in the calling sequence (notably, the cio subchannel lock). Push the cleanup of the private->cp resources to a workqueue, where it can be done out from under that lock sequence and a future patch can safely manage the locking requirements.

Linux

In the Linux kernel, the following vulnerability has been resolved: s390/vfio_ccw: Selectively expand io_mutex The io_mutex was defined to serialize the io_regions, but then has also sort of been associated with the I/O themselves because of the close relationship they share. With the handful of races that are possible, the choices are either to: A) expand the scope of io_mutex to close these remaining windows, or B) reduce the scope of io_mutex to just io_region, and introduce a new lock mechanism for the remaining I/O resources This patch implements A, since B brings with it a lot more interactions that would need to be tracked and kept in a correct hierarchy. It also takes advantage of the workqueue element for cp_free() that now gets called out of fsm_notoper(), which could be invoked out of an interrupt context and thus cannot acquire a mutex itself.

Linux

In the Linux kernel, the following vulnerability has been resolved: s390/vfio_ccw: Implement a crw lock Unlike the channel_program struct, which covers synchronous I/O submissions and asynchronous interrupts, the CRW region relies exclusively on asynchronous events coming from hardware. Implement a lock to manage the list of those payloads, to ensure they are read cohesively.

Linux

In the Linux kernel, the following vulnerability has been resolved: s390/zcrypt: Improve CCA CPRB length and overflow checks The xcrb_msg_to_type6cprb_msgx() function lacks proper input validation, creating security vulnerabilities: 1. Integer overflow after CEIL4 alignment: Signed int variables could overflow during 4-byte boundary alignment, causing undersized buffer allocations or incorrect bounds checking. 2. Missing minimum size validation: The CPRBX structure is copied from userspace without verifying sufficient buffer length. Undersized buffers cause uninitialized memory access when reading structure fields like cprbx.cprb_len and cprbx.domain. 3. Arithmetic overflow in sum calculations: Adding control block and data block sizes could overflow, bypassing size checks and enabling buffer overflows. Fix by using size_t for length calculations, adding U32_MAX boundary checks after alignment, validating minimum control block size before copying from userspace, and detecting sum calculation overflows.

Linux

In the Linux kernel, the following vulnerability has been resolved: s390/zcrypt: Improve EP11 CPRB length and overflow checks The xcrb_msg_to_type6_ep11cprb_msgx() function lacks proper input validation, creating security vulnerabilities: 1. Missing minimum size validation: The ep11_cprb structure and subsequent payload fields (pld_tag, pld_lenfmt) are copied from userspace without verifying sufficient buffer length. 2. Arithmetic overflow in length calculations: CEIL4 alignment could overflow, bypassing size checks and enabling buffer overflows. 3. The payload is asn1 encoded but the function just uses a simple c struct overlay to access some fields of the payload. Fix by using size_t for length calculations, adding U32_MAX boundary checks after alignment, and validating minimum request size and minimum reply size before copying from userspace. Do a very simple asn1 parsing of the payload up to the function value field.

Linux

In the Linux kernel, the following vulnerability has been resolved: s390/zcrypt: Improve EP11 CPRB domain handling with ASN.1 parsing The zcrypt_msgtype6_send_ep11_cprb() function uses fragile struct overlays to access and modify the domain field in the EP11 CPRB payload, creating maintainability and security concerns: 1. Struct overlay approach (pld_hdr) assumes fixed payload structure and doesn't validate the actual ASN.1 encoding. 2. Complex length format detection logic is error-prone and doesn't properly validate bounds at each parsing step. 3. Direct struct member access bypasses proper ASN.1 validation. Fix by replacing struct overlays with explicit ASN.1 parsing that validates each field (payload tag/length, function tag/length/value, optional domain tag/length/value) with proper bounds checking at every step. Add asn1_int_encode() helper function to safely write integer values with correct endianness conversion. This makes the code consistent with the validation pattern introduced with the rework of the xcrb_msg_to_type6_ep11cprb_msgx() function.

Linux

In the Linux kernel, the following vulnerability has been resolved: drm/amdgpu: validate GEM_CREATE domain combinations AMDGPU_GEM_CREATE checked domain bits against AMDGPU_GEM_DOMAIN_MASK, but did not validate domain combinations. Userspace could combine CPU|GTT|VRAM with DOORBELL, GDS, GWS, or OA, making amdgpu_bo_placement_from_domain() exceed AMDGPU_BO_MAX_PLACEMENTS and hit BUG_ON(). Allow combinations only within CPU/GTT/VRAM, and require non-CPU/GTT/ VRAM domains to be specified one at a time. Return -EINVAL for invalid combinations in amdgpu_gem_create_ioctl(). v2: Rename helper from amdgpu_gem_domain_valid() to amdgpu_gem_are_domains_valid() (Christian) (cherry picked from commit db39852d0c39843cb02048dfb47e4b8c703e9080)

Linux

In the Linux kernel, the following vulnerability has been resolved: drm/amdgpu: Fix UVD decode image min size calculation This needs to use pitch instead of width. Also reject pitch over 4096 to avoid overflow. (cherry picked from commit b41c8cb12e202b220353332ab87dc01a11f69304)

Linux

In the Linux kernel, the following vulnerability has been resolved: xfs: propagate errors from xfs_rtginode_load xfs_rtginode_ensure() treats every xfs_rtginode_load() error other than -ENOENT as success. This can leave the realtime group inode unset after an I/O, allocation, or corruption error. Growfs then continues as though the inode had been loaded. Only -ENOENT means that the inode needs to be created. Return all other errors to the growfs caller.

Linux

In the Linux kernel, the following vulnerability has been resolved: xfs: fix off-by-one in rtrefcount btree root level validation xfs_rtrefcountbt_compute_maxlevels() sets mp->m_rtrefc_maxlevels = min(d_maxlevels, r_maxlevels) + 1; where the trailing "+ 1" already accounts for the inode-root level, so the deepest valid on-disk root level is m_rtrefc_maxlevels - 1 and a cursor must satisfy bc_nlevels <= bc_maxlevels (= m_rtrefc_maxlevels). The two on-disk validation paths, xfs_rtrefcountbt_verify() and xfs_iformat_rtrefcount(), check the root level with ">" instead of ">=", so a crafted rtreflink (metadir + realtime + reflink) image whose /rtgroups/N.refcount inode has bb_level == m_rtrefc_maxlevels is accepted on mount. xfs_rtrefcountbt_init_cursor() then sets bc_nlevels = bb_level + 1, exceeding bc_maxlevels by one. Since the xfs_rtrefcountbt_cur slab object is sized for exactly bc_maxlevels entries, the first btree op on such a cursor indexes bc_levels[m_rtrefc_maxlevels] past the end of the object. This is reached by the first rtrefcount cursor built after mount, via log/CoW recovery (xfs_reflink_recover_cow() during xfs_mountfs()) or an FS_IOC_GETFSMAP over the realtime device. Reject a root level equal to m_rtrefc_maxlevels, matching the ">=" form already used by the sibling data-device refcount/rmap verifiers and the in-memory rtrmap verifier. BUG: KASAN: slab-out-of-bounds in xfs_btree_lookup (fs/xfs/libxfs/xfs_btree.c:2101) Write of size 2 at addr ffff888018391658 by task exploit/144 xfs_btree_lookup (fs/xfs/libxfs/xfs_btree.c:2101) xfs_btree_query_range (fs/xfs/libxfs/xfs_btree.c:5308) xfs_refcount_recover_cow_leftovers (fs/xfs/libxfs/xfs_refcount.c:2113) xfs_reflink_recover_cow (fs/xfs/xfs_reflink.c:1085) xlog_recover_finish (fs/xfs/xfs_log_recover.c:3551) xfs_mountfs (fs/xfs/xfs_mount.c:1158) xfs_fs_fill_super (fs/xfs/xfs_super.c:1940) get_tree_bdev_flags (fs/super.c:1634) vfs_get_tree (fs/super.c:1694) path_mount (fs/namespace.c:4161) __x64_sys_mount (fs/namespace.c:4367) entry_SYSCALL_64_after_hwframe (arch/x86/entry/entry_64.S:121) The buggy address belongs to the cache xfs_rtrefcountbt_cur of size 216 The buggy address is located 8 bytes to the right of allocated 216-byte region [ffff888018391578, ffff888018391650) Kernel panic - not syncing: Fatal exception

Linux

In the Linux kernel, the following vulnerability has been resolved: xfs: bounds-check buffer log item's dirty bitmap xlog_recover_do_reg_buffer() replays each dirty region described by a buffer log item's bitmap into the buffer read for that item: memcpy(xfs_buf_offset(bp, (uint)bit << XFS_BLF_SHIFT), item->ri_buf[i].iov_base, nbits << XFS_BLF_SHIFT); The destination offset (bit/nbits, from the logged dirty bitmap) and the buffer size (from the logged blf_len) are both attacker-controlled and otherwise unrelated, yet the only thing bounding the copy is an ASSERT(), which compiles away on production kernels. A crafted image logging a small blf_len together with a bitmap bit past the end of that buffer drives the memcpy() past the buffer's allocation, corrupting adjacent kernel heap during mount-time log recovery. This is reachable by anyone who can get a crafted image mounted -- the malicious-filesystem threat model XFS already guards against elsewhere. Turn the ASSERT() into a real XFS_IS_CORRUPT() check that aborts recovery of the buffer with -EFSCORRUPTED, consistent with the validate-and-fail idiom already used in xlog_recover_do_inode_buffer() and xfs_dquot_item_recover.c. xlog_recover_do_reg_buffer() therefore becomes STATIC int and its three callers propagate the error. Found and confirmed with KASAN on a CONFIG_XFS_DEBUG=n build: the crafted image trips a slab-out-of-bounds write before this change and fails recovery cleanly with -EFSCORRUPTED after it.

Linux

In the Linux kernel, the following vulnerability has been resolved: xfs: avoid UAF on sc->tempip in xrep_tempfile_create LOLLM noticed a potential UAF if the tempfile creation code fails after it set sc->tempip. Fix that.

Linux

In the Linux kernel, the following vulnerability has been resolved: xfs: fix exchange-range reflink flag clearing issue with INO1_WRITTEN When exchanging two full-file ranges, xmi_can_exchange_reflink_flags() can move the reflink inode flag from the file that currently has it to the other file, as long as exactly one side is marked. This assumes that the file contents, and therefore all shared extents, are exchanged. That assumption is not true when XFS_EXCHMAPS_INO1_WRITTEN is set. xfs_exchmaps_can_skip_mapping() can skip hole and unwritten mappings from file1, so an exchange can complete without moving every mapping that the earlier flag-swap decision accounted for. In that case the post-operation cleanup can clear the reflink flag from an inode that still owns shared written extents. Later writes then take the non-reflink write path and may update blocks that should still have been protected by CoW, which shows up as data corruption between reflink-related files. Fix this by disabling the reflink flag exchange whenever XFS_EXCHMAPS_INO1_WRITTEN is requested. The contents exchange can still proceed; the conservative outcome is that both inodes keep the reflink flag. The regular reflink flag cleanup path can drop the extra flag later once the inode no longer has shared extents.

Linux

In the Linux kernel, the following vulnerability has been resolved: ceph: fix hanging __ceph_get_caps() with stale mds_wanted A reader can hang forever in __ceph_get_caps() when the client no longer holds `FILE_RD`, but local cap state still says that the capability is already wanted (via `mds_wanted`). One way to trigger this is through MDS cap revocation. If another client performs a conflicting operation, the MDS can revoke `FILE_RD` from the reader; the next read then has to reacquire `FILE_RD`. If the cap update that should request `FILE_RD` never reaches the MDS after `cap->mds_wanted` was raised, the reader is left holding only non-file caps while local `mds_wanted` still includes the file read caps. In that state, try_get_cap_refs() sees `need <= mds_wanted` and returns 0, so __ceph_get_caps() just waits on `i_cap_wq`. If the cap update that was supposed to request `FILE_RD never reaches the MDS after `cap->mds_wanted was` raised, no further request is sent and the waiter can sleep indefinitely until unrelated cap traffic happens to wake it up. The ordering issue is that `cap->mds_wanted` is updated in __prep_cap() before the `CEPH_MSG_CLIENT_CAPS message` is actually queued for send. That makes one field serve two different meanings at once: what this client wants, and what the client believes the MDS already knows it wants. A proper fix would be to split those states and track whether a cap update is actually in flight or has been observed by the MDS. However, simply moving the `cap->mds_wanted assignment` later would not be sufficient: queueing the message in the messenger does not guarantee that the MDS processed that specific wanted set, and reconnect or message loss can still invalidate that assumption. Fixing that properly would require a larger rework of the cap state machine. To allow simpler backports to stable kernels, this patch implements a simpler workaround: - stop waiting forever in __ceph_get_caps(); after a bounded wait, fall back to the renew path - make ceph_renew_caps() issue a synchronous `OPEN` request whenever the inode still does not actually hold the wanted caps, instead of only calling ceph_check_caps() The extra issued-vs-wanted check in ceph_renew_caps() is necessary because the previous test only checked whether the inode still had any real caps at all. That is not enough after revocation: the client can still hold something like `pLs` and yet be missing `FILE_RD` completely. In that case, falling back to ceph_check_caps() is not sufficient, because it still trusts `cap->mds_wanted` and may resend nothing. By requiring `(issued & wanted) == wanted` before taking the asynchronous path, the code only uses ceph_check_caps() when the `wanted caps` are already actually issued. Otherwise, it sends the synchronous `OPEN` renew. This preserves the existing asynchronous fast path when the wanted caps are already issued, avoids changing cap-state semantics, and fixes the hang by guaranteeing that a stalled waiter eventually retries through a path that does not rely on the stale `mds_wanted` state. [ idryomov: move CEPH_GET_CAPS_WAIT_TIMEOUT from libceph.h to mds_client.h, formatting ]

Linux

In the Linux kernel, the following vulnerability has been resolved: ASoC: tas2562: Validate values for volume writes tas2562_volume_control_put() does not do any validation of the control value written by userspace, it uses it to look up a value in a fixed size array which can easily be overflowed and then writes whatever value it gets back to the device. Add validation that we are loading a value we have in the array.

Linux

In the Linux kernel, the following vulnerability has been resolved: clk: spacemit: k3: set hdma clock as critical HDMA clock is responsible for the internal TCM access path of X100 RISC-V core, so set the clock flag as critical to prevent it from being shut off, otherwise the Linux system will hang, for example in the case of a vector instruction access generates a page fault.

Linux

In the Linux kernel, the following vulnerability has been resolved: crypto: tegra - fix rctx->cryptlen calculation in tegra_gcm_do_one_req() Perform rctx->cryptlen calculation in tegra_gcm_do_one_req() the same way it is done in tegra_ccm_crypt_init(). The current formulae may lead to a crash if a caller does not call tegra_gcm_setauthsize() and so ctx->authsize remains zero. Then a decrypt operation with incorrect rctx->cryptlen will lead to a write beyound rctx->dst_sg buffer. As a follow-up cleanup delete struct tegra_aead_ctx->authsize field since it appears to be completely unused. Also simplify tegra_ccm_setauthsize() and tegra_gcm_setauthsize() functions respectively.

Linux

In the Linux kernel, the following vulnerability has been resolved: af_unix: Unlink scc_entry in unix_del_edge(). Kyle Zeng reported that GC could free a dead SCC partially. The scenario is as follows: 1) Create two SCCs: X -. A <-> B ^--' 2) Run the following concurrently: 2-1) send() sk-B to sk-B from sk-X 2-2) close() both A and B At 2-1), there is a small window where unix_add_edges() publishes a new edge (B <-> B) to GC but its skb is not queued by skb_queue_tail(). If 2-2) completes before skb_queue_tail() and GC is triggered, it judges A <-> B as dead, but B is not freed because GC cannot collect the not-yet-queued skb holding the B <-> B edge. X -. A <-> B -. This edge is visible ^--' ^..' but skb is not This itself is not a problem since the next GC run will judge B as dead as well and free it finally. X -. A <.> B -. ^--' ^--' However, X's SCC forces the next GC to call unix_walk_scc_fast(), and it iterates over A through B's scc_entry. Let's unlink scc_entry before freeing the vertex in unix_del_edge().

Linux

In the Linux kernel, the following vulnerability has been resolved: ovpn: fix NULL dereference when killing missing key ovpn_crypto_kill_key assumes both crypto slots are populated and dereferences each slot before checking it. That is not guaranteed: a peer can have only one installed key, and the kill path may be asked to remove a key that is not present. Read each slot once while holding the crypto state lock, check for NULL before looking at key_id, and only replace the slot that actually matches.

Linux

In the Linux kernel, the following vulnerability has been resolved: perf: Reject exited events as group leaders perf_event_remove_on_exec() sets remove-on-exec events to the EXIT state and detaches their group relationships. The event's file descriptor can remain open, however, and perf_event_open() currently accepts that event as a group leader because its early validation rejects only REVOKED and DEAD events. A new sibling can consequently be linked to the detached leader. When the leader is closed, perf_group_detach() observes that its PERF_ATTACH_GROUP bit is already clear and skips the new sibling. The sibling then retains a group_leader pointer to the freed event. Reject group leaders in the EXIT state. Perform the check while holding the shared context mutex so that an exec in the target task cannot detach the leader between validation and group attachment. [peterz: make the earlier test fully consistent]

Linux

In the Linux kernel, the following vulnerability has been resolved: ovpn: defer key slot crypto freeing to workqueue Key slots are released through a kref and the existing release path frees the AEAD transforms from an RCU callback. That is not safe for all crypto implementations: crypto_free_aead can sleep, for example when an async or hardware implementation has teardown work to complete. Use queue_rcu_work for key-slot release. This keeps the RCU grace period needed by lockless key-slot readers, but runs the actual crypto teardown from workqueue context where sleeping is allowed. Once the rcu_work callback runs, pre-existing RCU readers are gone, and the final kref put already proves that no transform user remains, so the worker can release the AEAD transforms and free the slot directly. The previous patch drains ovpn_wq during module exit, so queued key-slot teardown work cannot outlive module text.

Linux

In the Linux kernel, the following vulnerability has been resolved: netfilter: ipset: fix refcount race between list:set GC and swap __ip_set_put_byindex() resolved the index to a set pointer under RCU, then took ip_set_ref_lock in __ip_set_put() to decrement set->ref. ip_set_swap() holds that same lock while swapping both the ip_set_list slots and the two sets' ref counters, so it can interleave between the dereference and the lock acquisition, leaving the caller to decrement a set whose reference already moved to the other index and hit BUG_ON(set->ref == 0). list_set_gc() reaches this from timer softirq, which the nfnl mutex does not serialize against swap: an expiring list:set member calls list_set_del() -> ip_set_put_byindex() while IPSET_CMD_SWAP runs on the referenced sets. Resolve the index and decrement under ip_set_ref_lock, as ip_set_swap() already does, keeping the refcount tied to the index rather than to a stale set pointer. kernel BUG at net/netfilter/ipset/ip_set_core.c:685! Oops: invalid opcode: 0000 [#1] SMP KASAN NOPTI RIP: 0010:ip_set_put_byindex (net/netfilter/ipset/ip_set_core.c:870) Call Trace: <IRQ> list_set_del (net/netfilter/ipset/ip_set_list_set.c:159) set_cleanup_entries (net/netfilter/ipset/ip_set_list_set.c:181) list_set_gc (net/netfilter/ipset/ip_set_list_set.c:578) call_timer_fn (kernel/time/timer.c:1748) __run_timers (kernel/time/timer.c:1799 kernel/time/timer.c:2374) run_timer_softirq (kernel/time/timer.c:2405) </IRQ> Kernel panic - not syncing: Fatal exception in interrupt

Linux

In the Linux kernel, the following vulnerability has been resolved: ipvs: revalidate ihl to prevent out-of-bounds access While the outer IP header is already pulled into the skb head, we must be careful and revalidate the embedded headers after reading them from the skb frags to prevent out-of-bounds access. One such place reported by Sashiko is ip_vs_nat_icmp() where local process can change the ihl field and after skb_ensure_writable() we can see larger value which is a problem for the ip_send_check(cih) calls. Add check to drop the packet if the ihl field is changed.

Linux

In the Linux kernel, the following vulnerability has been resolved: eth: bnxt: avoid deadlock when canceling IRQ affinity notifier Unregistering IRQ affinity notifiers waits for the callback synchronously. bnxt takes the netdev instance lock in the notifier (to restart the queue) and cancels the work under the same lock. This may obviously deadlock. Move the restart to the async service task. The queue restart isn't super time sensitive. Store the new TPH tag, schedule the task. Safely canceling the service task is already ironed out. In bnxt_request_irq() the order of registering notifier, affinity and initial TPH programming has to be inverted. I think it was racy previously since user may trigger an update as soon as notifier is installed. There's a small known gap - if pcie_tph_get_cpu_st() fails at init and the target tag is 0 we may miss programming the entry. This does not seem worth fixing, the code has skip-on-failure all over the place, anyway.

Linux

In the Linux kernel, the following vulnerability has been resolved: veth: fix queue index used to wake the peer txq in veth_poll veth_poll() derives the index of the peer TX queue to wake from rq->xdp_rxq.queue_index. That field is only initialized by xdp_rxq_info_reg() in veth_enable_xdp_range(), which runs only when an XDP program is attached. On the plain GRO/NAPI path (veth_napi_enable_range()) xdp_rxq_info_reg() is never called, so queue_index stays 0 for every queue, as priv->rq is zero-allocated. So in a multi-queue setup with GRO enabled and no XDP program attached, every NAPI instance looks at the peer's TX queue 0. If veth_xmit() stops peer TX queue 1 because the ptr_ring is full (NETDEV_TX_BUSY), nothing ever wakes it again: the poller draining queue 1 wakes queue 0 instead. veth implements no ndo_tx_timeout, so the netdev watchdog does not kick in either, and the queue stays stopped indefinitely. Derive the index from the position of the rq within priv->rq instead, which is correct regardless of whether XDP was ever enabled. Scripts to reproduce the stall are available at https://github.com/netoptimizer/veth-backpressure-performance-testing

Linux

In the Linux kernel, the following vulnerability has been resolved: net: ngbe: fix NULL pointer dereference in non-MSI-X interrupt enabling In non-MSI-X mode (such as legacy INTx or single MSI), wx->msix_entry is not allocated or initialized. Calling NGBE_INTR_MISC(wx) dereferences wx->msix_entry->entry, leading to a NULL pointer dereference crash. This issue was introduced by fixing the IRQ vector when the number of VFs is 7. Fix the issue by explicitly checking `pdev->msix_enabled` to determine the correct vector index. Additionally, as a side fix, set the interrupt mask to BIT(0) for the non-MSI-X fallback. In MSI/INTx mode, the MISC and queue interrupts share vector 0, and the WX_PX_MISC_IVAR register is only valid in the MSI-X case. Thus, BIT(0) is the correct mask for the miscellaneous cause when MSI-X is disabled.

Linux

In the Linux kernel, the following vulnerability has been resolved: net/sched: cls_u32: skip hash tables in u32_bind_class() u32_walk() enumerates both struct tc_u_hnode and struct tc_u_knode through the walker callback. u32_bind_class() unconditionally casts the passed fh to tc_u_knode and accesses &n->res, so when fh is actually a tc_u_hnode, which has no tcf_result member, this results in a slab-out-of-bounds read of res->classid in tc_cls_bind_class(). The issue can be reproduced with the following commands: tc qdisc add dev lo root handle 1: hfsc tc class add dev lo parent 1: classid 1:1 hfsc sc rate 1000kbit tc filter add dev lo parent 1:1 protocol ip prio 1 u32 match u32 0 0 flowid 1:1 tc class add dev lo parent 1: classid 1:2 hfsc sc rate 2000kbit Fix this by skipping hash tables via the TC_U32_KEY(handle) check.

Linux

In the Linux kernel, the following vulnerability has been resolved: net/sched: cls_bpf: reject dev-bound programs bound to a different device cls_bpf_prog_from_efd() obtained a SCHED_CLS program via bpf_prog_get_type_dev() but never verified that a device-bound (offloaded) program's bound netdev matches the TC netdev the classifier is being attached to. This let a program loaded with prog_ifindex for device A be attached via cls_bpf + skip_sw to device B; deleting device A then destroyed the program's offload state while it was still attached to device B, triggering a netdevsim WARN (panic with panic_on_warn=1). Mirror the XDP attach path (net/core/dev.c) and reject the attach with -EINVAL when a dev-bound program's bound device does not match the target device.

OpenZFS

On Linux, several OpenZFS ioctl authorization checks accept a capability held only within a user-created, unprivileged namespace as equivalent to real host privilege, allowing an unprivileged local user to perform operations that should require root. Affected operations include pool-administrative operations (eg create, import, destroy), pool event log access (zpool events) and fault injection (zinject). Exploiting the problem requires only that the local user is permitted to open /dev/zfs (governed by local device permissions) and that the kernel permits unprivileged user namespace creation. No prior access to the target pool or its underlying devices is needed.