CVE Feed

In the Linux kernel, the following vulnerability has been resolved: RDMA/mlx5: Fix stack out-of-bounds read in cc_params debugfs get_param() reads a congestion parameter as a u32 but formats it with the signed "%d" into an 11-byte stack buffer. A value with bit 31 set, such as 0x80000000, renders as "-2147483648\n" whose full length is 12. snprintf() stores only 11 bytes yet returns 12, so simple_read_from_buffer() treats 12 bytes as valid and reads one byte past lbuf[]. Size the buffer for the widest unsigned decimal, format with "%u" to match the u32, and use scnprintf() so the length passed to simple_read_from_buffer() reflects the bytes actually stored.

In the Linux kernel, the following vulnerability has been resolved: RDMA/cxgb4: free STAG index when TPT entry write fails write_tpt_entry() allocates a new STAG index with c4iw_get_resource() and bumps stats.stag.cur before programming the entry. When write_adapter_mem() fails, it returns the error without releasing the index or reversing the statistic. No MR is inserted into rhp->mrs, so deregistration never reclaims it, leaking the index until device teardown. Record whether this call allocated the index and, on a failed write, return it to tpt_table and decrement stats.stag.cur. Key the rollback on both the write error and that flag, not the error alone: a non-reset update carries a caller-owned STAG that this call did not allocate and must not free.

In the Linux kernel, the following vulnerability has been resolved: IB/isert: reject PDUs declaring more data than was received isert_recv_done() hands each received PDU to the opcode handlers without ever looking at wc->byte_len, the number of bytes the HCA actually placed in the receive descriptor. The handlers then copy that many bytes - the data-segment length the initiator declared in the BHS (ntoh24(hdr->dlength), via the derived unsol_data_len / imm_data_len) - out of the fixed-size descriptor: isert_handle_iscsi_dataout(): sg_copy_from_buffer(sg_start, sg_nents, isert_get_data(rx_desc), unsol_data_len); isert_handle_scsi_cmd(): sg_copy_from_buffer(cmd->se_cmd.t_data_sg, sg_nents, isert_get_data(rx_desc), imm_data_len); Because the declared length is never checked against wc->byte_len, an initiator can declare a data segment larger than the bytes it actually sent (and larger than the descriptor) and cause an out-of-bounds read of the receive buffer. Nothing upstream of isert closes this door: - __iscsit_check_dataout_hdr() bounds the inbound payload against conn_ops->MaxXmitDataSegmentLength (MXDSL) - a transmit parameter, used here for the inbound check. - iscsi_set_connection_parameters() sets ops->MaxXmitDataSegmentLength = ops->TargetRecvDataSegmentLength; and TARGETRECVDATASEGMENTLENGTH is absent from the min()-clamp list in iscsi_check_acceptor_state(), so the value the initiator declares is adopted verbatim (type range 512..16777215). The initiator effectively raises its own ceiling. - isert never clamps the negotiated value to its own fixed receive descriptor (ISER_RX_SIZE, 9216 bytes), so the target core's bound and the descriptor size are unrelated. The imm_data_len == data_len path is more than an over-read: it aliases the receive descriptor via sg_set_buf() and passes it to the backend as the data source for the SCSI WRITE, so an over-declared length causes heap contents past the descriptor to be written through the backend to the backing store. The backend is the victim of the oversized scatterlist isert hands it, not the cause; no read-back of the written bytes was demonstrated. Trigger: after login completes (full feature phase), an initiator that has declared a large TargetRecvDataSegmentLength and a FirstBurstLength that permits unsolicited/immediate data sends a PDU whose declared data-segment length exceeds what was received. With KASAN: BUG: KASAN: slab-out-of-bounds in sg_copy_buffer+0x150/0x1c0 Read of size 4096 at addr ffff888109720800 by task kworker/1:0H/25 Workqueue: ib-comp-wq ib_cq_poll_work Call Trace: sg_copy_buffer+0x150/0x1c0 isert_recv_done+0xba6/0x2390 __ib_process_cq+0xe1/0x390 ib_cq_poll_work+0x46/0x150 isert_recv_done+0xba6 resolves to isert_handle_iscsi_dataout() (ib_isert.c:1160), inlined through isert_rx_opcode(). Validate wc->byte_len against the framing in isert_recv_done() before the PDU reaches any handler, and reinstate the connection if it is short. Because the test compares without subtracting the header length, it also rejects PDUs shorter than the iSER and iSCSI headers, which would otherwise be parsed out of stale descriptor contents. The login handler rejects PDUs shorter than ISER_HEADERS_LEN (commit 29e7b925ae6d ("IB/isert: Reject login PDUs shorter than ISER_HEADERS_LEN")) but does not bound the declared length either; that is fixed in the next patch. The data handlers had no length check at all. isert reads the data segment from a fixed offset: isert_get_data() returns the iSER header plus ISER_HEADERS_LEN and makes no adjustment for an AHS. The bytes the handlers touch are therefore exactly [ISER_HEADERS_LEN, ISER_HEADERS_LEN + dlength), and comparing that sum against wc->byte_len bounds precisely the region that is read. An AHS term would only make the test stricter without bounding anything furth ---truncated---

In the Linux kernel, the following vulnerability has been resolved: IB/isert: reject login PDUs declaring more data than was received isert_login_recv_done() records how many bytes the HCA actually placed in the login buffer, but nothing compares that against the length the login PDU's BHS declares. isert_rx_login_req() copies min(login_req_len, MAX_KEY_VALUE_PAIRS) bytes into login->req_buf, and the login code then reads the declared length back out of that buffer - for the first PDU in iscsi_target_locate_portal(), payload_length = ntoh24(login_req->dlength); tmpbuf = kmemdup_nul(login->req_buf, payload_length, GFP_KERNEL); and for the ones after it in iscsi_decode_text_input(), reached from iscsi_target_do_login(). login->req_buf is a fixed MAX_KEY_VALUE_PAIRS (8192) byte allocation, so an initiator that declares more than it sends reads off the end of it, before authentication and with the length under its control: BUG: KASAN: slab-out-of-bounds in kmemdup_nul+0x43/0x80 Read of size 8193 at addr ffff8881056a8000 by task iscsi_np/167 __asan_memcpy+0x23/0x60 kmemdup_nul+0x43/0x80 iscsi_target_locate_portal+0x48d/0x1180 iscsi_target_login_thread+0x19a9/0x3350 Allocated by task 167: __kmalloc_cache_noprof+0x158/0x370 iscsi_target_login_thread+0x971/0x3350 which belongs to the cache kmalloc-8k of size 8192 allocated 8192-byte region Falsifying the second login PDU instead reaches the other reader, on the same buffer: BUG: KASAN: slab-out-of-bounds in kmemdup_nul+0x43/0x80 Read of size 8193 at addr ffff888104d10000 by task kworker/1:1/50 Workqueue: isert_login_wq iscsi_target_do_login_rx __asan_memcpy+0x23/0x60 kmemdup_nul+0x43/0x80 iscsi_decode_text_input+0xc6/0x11c0 iscsi_target_do_login+0x261/0x1470 iscsi_target_do_login_rx+0x51d/0x7d0 iscsit over TCP is not exposed: iscsit_get_login_rx() validates the declared length with iscsi_target_check_login_request() and then reads exactly that many bytes off the socket, so the declared length governs how much arrives rather than how much is copied out of an already-filled buffer. isert does not call iscsi_target_check_login_request() at all. Reject a login PDU whose declared DataSegmentLength exceeds what was received, in both paths that reach isert_rx_login_req(): isert_get_login_rx() for the first login PDU and isert_login_recv_done() for the ones after it. dlength <= login_req_len is allowed because the received count can include up to three bytes of iSCSI padding. Once the check is in place the copy out can no longer exceed the copy in: the posted login SGE is ISER_RX_PAYLOAD_SIZE, so login_req_len cannot exceed MAX_KEY_VALUE_PAIRS and the min() in isert_rx_login_req() is login_req_len. Like the existing short-PDU check added by 29e7b925ae6d, the reject in isert_login_recv_done() returns without completing login_req_comp, so a malformed subsequent PDU leaves the login to be torn down by the login timer rather than failing immediately. The first-PDU path returns an error and fails straight away. Reproduced on 7.2.0-rc4 with soft-RoCE (rdma_rxe) under KASAN, using an initiator that sends the real key=value payload while declaring 8193 in the BHS, on the first login PDU and on the second in separate runs. The reported read size tracks the declared value exactly; 16384 and 61440 behave the same. Unpatched 3 of 3 runs report on each of the two paths, patched 0 of 3 on both, run alternately in a single session, and a normal login still completes on the patched build.

In the Linux kernel, the following vulnerability has been resolved: nvmet: fix return status of RMI log page on allocation failure nvmet_execute_get_log_page_rmi() leaves 'status' holding NVME_SC_SUCCESS (set by the successful nvmet_req_find_ns() call) when the kzalloc() for the log buffer fails. It then jumps to the out label and completes the request with a success status, so the host is told the command succeeded while no data was transferred. Initialize 'status' to NVME_SC_INTERNAL, matching the smart log handler, so an allocation failure is reported as an internal error.

In the Linux kernel, the following vulnerability has been resolved: nvme-fc: unmap cmd_iu DMA on rsp_iu mapping failure in init_request __nvme_fc_init_request() maps cmd_iu and then rsp_iu for DMA. If the rsp_iu mapping fails, the original code only recorded the error and fell through: it left the already-mapped cmd_iu unmapped and still marked the op as FCPOP_STATE_IDLE before returning. Since blk-mq does not call .exit_request() when .init_request() fails, the cmd_iu mapping is leaked for every op whose rsp_iu mapping fails. Jump to an error path on rsp_iu mapping failure that unmaps cmd_iu and returns the error without marking the op idle, so it stays in the FCPOP_STATE_UNINIT state set by the initial memset().

In the Linux kernel, the following vulnerability has been resolved: spi: davinci: switch to managed controller allocation The controller is allocated with the non-managed spi_alloc_host() while the interrupt is registered with devm_request_threaded_irq(). During removal, spi_bitbang_stop() only unregisters the controller; the subsequent spi_controller_put() then frees the controller together with its embedded davinci_spi devdata, which is the IRQ handler's dev_id. The devm_request_threaded_irq() release action (free_irq()), which drains the handler, does not run until after .remove() returns. A late or latched interrupt can therefore reach davinci_spi_irq() and dereference already-freed memory. Switch to devm_spi_alloc_host() so that the devres LIFO order releases the controller only after free_irq() has drained the handler, and drop the now-redundant spi_controller_put() from .remove(). The probe error path is simplified to direct returns. The clock is acquired with devm_clk_get_enabled(), which is registered after the IRQ and thus released before it by the devres LIFO order. Drain the interrupt explicitly with devm_free_irq() before disabling the controller so that a late interrupt cannot access the registers of a clock-gated controller. This issue was found by an in-house static analysis tool.

In the Linux kernel, the following vulnerability has been resolved: drm/panthor: Add vm_bind region with kbo range overlap check When a VM is created, caller has to specify the range of the address space carve-out set aside for mapping kernel BO's. That means vm_bind mappings of UM-exposed BO's should not intersect with that region, but at the moment we're not checking this. At first, I thought of giving these values to drm_gpuvm_init() through its reserve_{offset, range} arguments, but it turns out that is meant for VM address spans that are not managed through the usual drm_gpuvm split/merge circuit, so storing the end of the user VA range at VM creation time and doing a quick check in the vm_bind ioctl path was the simplest workaround. The new check also makes sure vm_bind range doesn't overflow the size of a 64-bit unsigned integer. That was already being done further down the call stack inside drm_gpuvm_sm_map -> drm_gpuvm_range_valid, but it's best to fail early in the driver before GPUVM functions are invoked so that we won't waste time allocating vm_bind context resources.

In the Linux kernel, the following vulnerability has been resolved: wifi: ath12k: fix overreads in ath12k_wmi_process_csa_switch_count_event() There is no policy entry for WMI_TAG_PDEV_CSA_SWITCH_COUNT_STATUS_EVENT, so the parse infrastructure does not enforce a minimum length for the event struct. Additionally, the num_vdevs field is taken directly from firmware and used as a loop bound over the vdev_ids array without checking that it fits within the TLV payload. Either condition can cause an out-of-bounds read. Add a TLV policy entry for WMI_TAG_PDEV_CSA_SWITCH_COUNT_STATUS_EVENT so the parse infrastructure enforces a minimum length for the fixed-size event struct. Add a helper ath12k_wmi_tlv_data_len() to recover the payload length of a parsed TLV from the header preceding its data pointer. Use it in ath12k_wmi_process_csa_switch_count_event() to bound num_vdevs before the loop. Compile tested only.

In the Linux kernel, the following vulnerability has been resolved: media: qcom: iris: handle runtime PM resume failure in core deinit Check the return value of pm_runtime_resume_and_get() in iris_core_deinit(). If runtime PM resume fails, skip hardware power-off operations but still perform software teardown and state transition. Also skip the corresponding pm_runtime_put_sync() call to avoid unbalanced runtime PM references.

In the Linux kernel, the following vulnerability has been resolved: media: stm32: dcmi: fix some error handling bugs in probe() There are a few issues here: 1) After we assign: chan = dma_request_chan(&pdev->dev, "tx"); Then the error paths need to clean up before returning. The first error path does a direct return. 2) The error paths check "dcmi->mdma_chan" but that is not assigned until later so it results in memory leaks. Test "mdma_chan" instead. 3) The error handling calls dma_release_channel(dcmi->dma_chan) before "dcmi->dma_chan" has been assigned which leads to a NULL pointer dereference. Use the "chan" variable instead. I also moved the call to dma_release_channel() after the call to dma_release_channel() so it mirrors the allocation code better.

In the Linux kernel, the following vulnerability has been resolved: platform/chrome: cros_ec_debugfs: Unregister panic notifier cros_ec_debugfs_probe() registers notifier_panic with the EC panic notifier chain. The remove path tears down debugfs and the console log, but leaves the notifier registered. A later panic notification can call back into the removed instance and queue work that accesses released data. Unregister the panic notifier before tearing down the debugfs and console log state. This issue was found by a static analysis tool.

In the Linux kernel, the following vulnerability has been resolved: wifi: rtlwifi: pci: fix error path in rtl_pci_probe() In the last error path in rtl_pci_probe(), the cleanup functions are skipped due to a wrong goto label. Moreover, the successful call to rtl_init_rfkill(), ieee80211_register_hw(), rtl_debug_add_one() have to be reverted. Fix this issue by updating the labels and adding the relevant cleanup functions to the last error path.

In the Linux kernel, the following vulnerability has been resolved: bus: mhi: host: Fix controller cleanup on EDL sysfs failure mhi_register_controller() adds the controller device before creating the optional trigger_edl sysfs file. If sysfs_create_file() fails, the error path only drops the device reference and leaves the device registered. Hence, call device_del() in the error path before put_device().

In the Linux kernel, the following vulnerability has been resolved: md: remove REQ_NOWAIT support from raid1/10/456 REQ_NOWAIT support in md personalities that can block internally is fundamentally incomplete. While reads can avoid some blocking paths, write requests can still encounter cases where one mirror succeeds while another returns -EAGAIN. At that point md cannot distinguish queue pressure from a real device failure, so it can neither record a bad block nor safely retry the write without REQ_NOWAIT, leaving mirrors with divergent data. Rather than continue advertising REQ_NOWAIT support for personalities that cannot implement it correctly, remove it from raid1, raid10 and raid456. Keep REQ_NOWAIT for linear and raid0, which only remap bios to their underlying devices; stacked limits will still clear the feature if any component device lacks REQ_NOWAIT support.

In the Linux kernel, the following vulnerability has been resolved: md: recheck spare changes before starting sync remove_spares() and remove_and_add_spares() modify the array's rdev configuration. These operations are only safe after the array has been suspended. md_start_sync() checks whether spare configuration changes are needed before taking reconfig_mutex. However, the rdev state can change before the mutex is acquired, so the initial check can become stale. In that case, md_choose_sync_action() may remove or replace rdevs while normal I/O is still accessing them. The race can occur as follows: raid10d Worker Normal IO ____________ _______________________ ______________________ raid10_write_request() wait_blocked_dev() set Blocked set Faulty Skip Faulty rdev rrdev->nr_pending++ .repl_bio = bio removeable_rdev = false . array not suspended . lock mddev goto err_handle lock mddev (wait) . update sb . clear Blocked . . unlock mddev . lock mddev (acquires) remove_spares() removeable_rdev = true raid10_remove_disk() rdev = replacement replacement = NULL rdev_dec_pending(NULL) unlock mddev (NULL)->nr_pending-- In this case, rdev_dec_pending() is called with a NULL pointer, resulting in a NULL pointer dereference when attempting to decrement nr_pending. Fix this by suspending the array when spare configuration changes are needed, including for non-read-write arrays, and checking again after taking reconfig_mutex. If the array was not already suspended and a change is now needed, release the mutex, suspend the array, and reacquire the mutex before continuing.

In the Linux kernel, the following vulnerability has been resolved: wifi: ath12k: fix stride mismatch in mac_phy_caps_parse() Currently, in ath12k_wmi_mac_phy_caps_parse(), kzalloc() sizes the mac_phy_caps buffer as tot_phy_id * len, where len is clamped to min(firmware_len, sizeof(struct ath12k_wmi_mac_phy_caps_params)). The subsequent memcpy() destination advances by sizeof(full struct) per slot via C pointer arithmetic, not by the clamped len. When firmware sends short TLVs, the second and later slots are written past the end of the allocation. The reader in ath12k_pull_mac_phy_cap_svc_ready_ext() also indexes the buffer with full-struct pointer arithmetic, so the allocation must match that stride. Fix by using kzalloc_objs(), which derives the element size from the pointer type, making allocation size and pointer stride provably consistent regardless of what len the firmware provides. Tested-on: WCN7850 hw2.0 PCI WLAN.HMT.1.1.c7-00108-QCAHMTSWPL_V1.0_V2.0_SILICONZ_UPSTREAM-3

In the Linux kernel, the following vulnerability has been resolved: wifi: ath11k: fix stride mismatch in mac_phy_caps_parse() Currently, in ath11k_wmi_tlv_mac_phy_caps_parse(), kcalloc() sizes the mac_phy_caps buffer as tot_phy_id * len, where len is clamped to min(firmware_len, sizeof(struct wmi_mac_phy_capabilities)). The subsequent memcpy() destination advances by sizeof(full struct) per slot via C pointer arithmetic, not by the clamped len. When firmware sends short TLVs, the second and later slots are written past the end of the allocation. The reader in ath11k_pull_mac_phy_cap_svc_ready_ext() also indexes the buffer with full-struct pointer arithmetic, so the allocation must match that stride. Fix by using kzalloc_objs(), which derives the element size from the pointer type, making allocation size and pointer stride provably consistent regardless of what len the firmware provides. Compile tested only.

In the Linux kernel, the following vulnerability has been resolved: firmware: qcom: scm: Fix NULL dereference in IRQ handler before __scm is published In qcom_scm_probe(), devm_request_threaded_irq() is called before smp_store_release(&__scm, scm). Two paths can dereference __scm before it is published, both causing a NULL pointer dereference. The IRQ handler receives scm via its data argument but passes only wq_ctx to qcom_scm_waitq_wakeup() and qcom_scm_get_completion(), which then dereference __scm directly. Thread scm through both functions so the IRQ handler path never touches __scm. Non-atomic SMC calls made during probe (e.g. from qcom_tzmem_init via qcom_scm_shm_bridge_enable) can return WAITQ_SLEEP, causing qcom_scm_wait_for_wq_completion() to run before __scm is published and dereference it. Add platform_set_drvdata(pdev, scm) early in probe and change qcom_scm_wait_for_wq_completion() to take the device pointer and use dev_get_drvdata() to reach scm, removing any dependency on __scm.

In the Linux kernel, the following vulnerability has been resolved: block: fix dio leak on metadata mapping error A failed integrity mapping holds a dio reference, so we need to go through the full bio ending in case there were previously submitted bio's in the sequence.

In the Linux kernel, the following vulnerability has been resolved: power: supply: isp1704_charger: cancel work on remove The USB notifier and initial VBUS detection can schedule isp->work. The remove path unregisters the notifier and power supply, but does not wait for queued or running work before tearing down the power supply state. Cancel the work after unregistering the notifier. Do this before unregistering the power supply. This issue was found by a static analysis tool.

In the Linux kernel, the following vulnerability has been resolved: power: supply: sc2731_charger: cancel work on remove The USB notifier and initial charger detection can schedule info->work. The remove path unregisters the notifier, but does not cancel queued or running work before the devm-allocated driver data is released. Set the platform drvdata used by remove, then cancel the work after unregistering the notifier. This issue was found by a static analysis tool.

In the Linux kernel, the following vulnerability has been resolved: bpf: Fix potential UAF in bpf_netns_link_update_prog In bpf_netns_link_update_prog, the checks for old_prog and prog type are currently performed locklessly before acquiring netns_bpf_mutex. This creates a race condition that can lead to a UAF issue. If two threads concurrently execute BPF_LINK_UPDATE on the same netns link, the following execution path can trigger a UAF: CPU0 CPU1 bpf_netns_link_update_prog if (old_prog && old_prog != link->prog) return -EPERM; bpf_netns_link_update_prog if (old_prog && old_prog != link->prog) ... old_prog = xchg(&link->prog, new_prog); bpf_prog_put(old_prog); if (new_prog->type != link->prog->type) <-- trigger UAF Fix this by moving the old_prog and prog->type checks inside the netns_bpf_mutex critical section. Meanwhile, use guard() to simplify lock management and avoid all the goto jumping.

In the Linux kernel, the following vulnerability has been resolved: bpf: Fix potential UAF when reading bpf link info In bpf_link_show_fdinfo and bpf_link_get_info_by_fd, link->prog is accessed without holding any locks. If the prog is concurrently replaced via bpf_link_update, the old prog can be freed, leading to a potential UAF issue. Fix this by accessing link->prog under RCU protection to safely fetch the pointer and guarantee its lifetime while reading its fields.

In the Linux kernel, the following vulnerability has been resolved: lib/test_hmm: fail dmirror_fault() when the mirrored mm is gone dmirror_fault() is called from the dmirror_read() and dmirror_write() retry loops after dmirror_do_read() or dmirror_do_write() finds a missing device page table entry. If the mirrored mm has already exited, mmget_not_zero() fails. The current code returns 0 in that case, which tells the caller that faulting succeeded even though no page was faulted and no device page table entry was installed. The caller then retries the same address, hits -ENOENT again, and can loop forever without making progress. Return -EFAULT instead, so the ioctl fails when the mirrored mm is no longer faultable.

In the Linux kernel, the following vulnerability has been resolved: md/bitmap: resume array on backlog_store() error path backlog_store() suspends the array before checking whether a write-mostly device exists. If no such device exists, the error path only unlocks reconfig_mutex and leaves the array suspended, blocking subsequent I/O. Use mddev_unlock_and_resume() to release both states.

In the Linux kernel, the following vulnerability has been resolved: md: scope memalloc_noio to allocation critical sections Storing a memalloc_noio_save() token in mddev->noio_flags lets one task save the token and another task restore it. With concurrent suspend sysfs writes, task A can enter PF_MEMALLOC_NOIO, return to userspace still in that scope, and later task B can restore A's saved token. Avoid tying the token lifetime to mddev. Keep mddev_suspend() and mddev_resume() only responsible for array suspension, and enter PF_MEMALLOC_NOIO only in the MD paths that allocate memory after the array has been suspended. Restore the token before resuming the array. A reproducer repeatedly writes suspend_lo and suspend_hi from concurrent workers and checks each worker's /proc/self/stat flags before and after the sysfs write.

In the Linux kernel, the following vulnerability has been resolved: iommu/dma: Check atomic pool allocation result directly The non-blocking, non-coherent allocation path uses dma_alloc_from_pool(), which returns the allocated page and fills cpu_addr only on success. Do not rely on cpu_addr to detect allocation failure in this path. Check the returned page directly before using it for the IOMMU mapping.

In the Linux kernel, the following vulnerability has been resolved: swiotlb: Preserve allocation virtual address for dynamic pools swiotlb_alloc_tlb() can allocate from the DMA atomic pool when a decrypted pool is needed from atomic context. With CONFIG_DMA_DIRECT_REMAP, the atomic pool is backed by remapped virtual addresses, which are not the same as the direct-map addresses returned by phys_to_virt(). swiotlb_init_io_tlb_pool() currently reconstructs the pool virtual address from the physical start address. For atomic-pool backed allocations this stores the wrong address in pool->vaddr. Later, swiotlb_free_tlb() passes that address to dma_free_from_pool(), which will fail to recognize the chunk Pass the virtual address returned by the allocation path into swiotlb_init_io_tlb_pool(), and store that address in pool->vaddr. This keeps the pool free path using the same virtual address as the allocator.

In the Linux kernel, the following vulnerability has been resolved: i3c: dw: avoid shift-out-of-bounds when DAA assigns no devices On an empty bus ENTDAA assigns nothing, so cmd->rx_len (the count of addresses left unassigned) equals master->maxdevs. The GENMASK() index master->maxdevs - cmd->rx_len - 1 then becomes -1, which trips up UBSAN. This happens every time on boot on a Gigabyte/AMD server: UBSAN: shift-out-of-bounds in drivers/i3c/master/dw-i3c-master.c:905:12 shift exponent 64 is too large for 64-bit type 'long unsigned int' CPU: 7 UID: 0 PID: 963 Comm: (udev-worker) Not tainted 7.0.11-200.fc44.x86_64 #1 PREEMPT(lazy) Hardware name: Giga Computing E163-Z34-AAH1-000/MZ33-DC1-000, BIOS R32_F45 04/01/2026 Call Trace: <TASK> dump_stack_lvl+0x5d/0x80 ubsan_epilogue+0x5/0x2b __ubsan_handle_shift_out_of_bounds.cold+0xd7/0x1ab dw_i3c_master_daa.cold+0x1b/0x96 [dw_i3c_master] i3c_master_do_daa_ext.part.0+0x3e/0xf0 [i3c] Skip the mask when no new device was assigned.

In the Linux kernel, the following vulnerability has been resolved: md/raid1: create serial pool adding rdev to array with serialize_policy=1 The following bug has been observed with kernel 7.1.3 after adding a new rdev to an existing RAID1 array with serialize_policy enabled: Oops: 0002 [#1] CPU: 0 UID: 0 PID: 19639 Comm: ext4lazyinit Not tainted 7.1.3-1-default RIP: _raw_spin_lock_irqsave+0x27/0x50 CR2: 0000000000004960 Call Trace: wait_for_serialization+0xb9/0x260 [raid1] raid1_make_request+0x762/0xaff [raid1] md_handle_request+0x1c9/0x2e0 [md_mod] The raid1.c code calls wait_for_serialization() if the MD_SERIALIZE_POLICY is set, and wait_for_serialization assumes that rdev->serial is initialized. Normally this will be the case for arrays that have the serialize_policy sysfs attribute set to 1. But when a new rdev is added to an existing array in bind_rdev_to_array(), the condition at mddev_create_serial_pool() causes creation of rdev->serial to be skipped. Fix it.

In the Linux kernel, the following vulnerability has been resolved: iomap: release the folio batch on iomap callback failures A sashiko review of an unrelated patch points out that the folio batch mechanism used for iomap zero range fails to release the batch in a couple error scenarios. If either calls to ->iomap_end() or ->iomap_begin() fail, the direct return paths bypass the batch cleanup. The ->iomap_end() case is not a practical issue at the moment because there is no user of the mechanism that returns an error from this path. The ->iomap_begin() case is theoretically possible because XFS can invoke the fill helper and error out at various points thereafter. This subtly complicates things because XFS does not transfer iomap_flags to the iomap data structure in the error path. To deal with both of these issues, first make sure to invoke the cleanup helper in the error path for either fs callback. Second, update the helper to clear the flag unconditionally and release the batch so long as it is populated. This more clearly delineates the purpose of the flag to control the I/O path and not necessarily the status of the fbatch, so add a comment around this as well.

In the Linux kernel, the following vulnerability has been resolved: misc: sgi-gru: remove interrupt-context page-table walks The GRU TLB miss handler walks a process's page tables without holding page-table locks or a reference to the mapped page. It also uses a kernel page-table accessor on user page tables and supports only PMD-level large mappings on x86-64. Remove the direct walker. Send interrupt faults directly to user polling mode so the existing call-OS fallback retries them in process context. Remove the mmap-lock failure statistic that can no longer be incremented.

In the Linux kernel, the following vulnerability has been resolved: wifi: mt76: mt76x02: do not WARN on invalid rx descriptor length The MPDU length in the rx descriptor comes from the hardware. In monitor mode with the fcsfail filter enabled, the hardware passes up corrupted frames, and a corrupted frame can report a length larger than the received buffer. The bounds check correctly discards such frames, but its WARN_ON_ONCE wrapper means any over-the-air garbage frame taints the kernel, and panics it on the first such frame when panic_on_warn is set. Drop the WARN and discard the frame silently, matching what commit c2d4c8723dbf ("mt76x2: remove some harmless WARN_ONs in tx status and rx path") did for the neighboring rx and tx status paths. Observed immediately on rx with an MT7612U in fcsfail monitor mode on a busy channel.

In the Linux kernel, the following vulnerability has been resolved: wifi: mt76: fix handling channel context with different bands in mt76_switch_vif_chanctx() When performing channel switches on different radios within a short timeframe, channel contexts with different bands can be carried for each struct ieee80211_vif_chanctx_switch. Rework mt76_switch_vif_chanctx() to properly handle this scenario.

In the Linux kernel, the following vulnerability has been resolved: wifi: mt76: mt792x: fix use-after-free in mt76_rx_poll_complete A use-after-free issue occurs in mt76_rx_poll_complete due to a race condition. The STA has already been removed, but the rx_status still had a pointer to the wcid in the STA. Set the links' wcid pointers to be NULL for a MLD in mt7925_sta_pre_rcu_remove() BUG: KASAN: invalid-access in mt76_rx_poll_complete+0x280/0x470 Call trace: dump_backtrace+0xec/0x128 show_stack+0x18/0x28 dump_stack_lvl+0x40/0xc8 print_report+0x1b8/0x710 kasan_report+0xe0/0x144 do_bad_area+0x120/0x260 do_tag_check_fault+0x20/0x34 do_mem_abort+0x54/0xa8 el1_abort+0x3c/0x5c el1h_64_sync_handler+0x40/0xcc el1h_64_sync+0x7c/0x80 mt76_rx_poll_complete+0x280/0x470 mt76_dma_rx_poll+0x114/0x51c mt792x_poll_rx+0x60/0xf8 napi_threaded_poll_loop+0xe0/0x450 napi_threaded_poll+0x80/0x9c kthread+0x11c/0x158 ret_from_fork+0x10/0x20

In the Linux kernel, the following vulnerability has been resolved: wifi: mt76: mt7921: Add PCIe AER handler support to prevent system crash When an AER error occurs and the bus is hung, the register reads return 0xFFFFFFFF, causing the DMA queue state to be corrupted and resulting in an invalid memory access when accessing q->desc[] or q->entry[]. Unable to handle kernel paging request at virtual address ffffffc01099eac0 pc : mt76_dma_add_buf+0x124/0x188 [mt76] lr : mt76_dma_rx_fill+0x11c/0x1d8 [mt76] sp : ffffffc016d9bbf0 x29: ffffffc016d9bc10 x28: 0000000000000000 x27: 0000000000000000 x26: ffffffb7855e50b8 x25: ffffffb80d04f000 x24: 0000000000000000 x23: 0000000000000ec0 x22: ffffffb796803648 x21: ffffffb796801f80 x20: ffffffb7968035f8 x19: 0000000000000ec0 x18: 0000000000000000 x17: 000000004ec00000 x16: 000000000ec00000 x15: ffffffc01099eac0 x14: 000000004ec00000 x13: 00000000ffc5a000 x12: ffffffc016d9bc32 x11: 00000000ffffffff x10: 0000000000000002 x9 : 0000000000000000 x8 : 000000000000b4ac x7 : 0000000000000a20 x6 : ffffffb6c1806400 x5 : 0000000000000000 x4 : ffffffb80d04f000 x3 : 0000000000000000 x2 : 0000000000000001 x1 : 000000000ec04000 x0 : ffffffb7968035f8 Call trace: mt76_dma_add_buf+0x124/0x188 [mt76 (HASH:1029 4)] mt76_dma_rx_reset+0xe8/0xfc [mt76 (HASH:1029 4)] mt7921_wpdma_reset+0x188/0x1b0 [mt7921e (HASH:ee48 5)] mt7921e_mac_reset+0x128/0x418 [mt7921e (HASH:ee48 5)] mt7921_mac_reset_work+0xac/0x1a8 [mt7921_common (HASH:f721 6)] process_one_work+0x188/0x514 worker_thread+0x12c/0x300 kthread+0x140/0x1fc ret_from_fork+0x10/0x30 Fix the invalid memory access by validating the DMA index read from the hardware before it is used as a queue index. An out-of-range value, such as the 0xFFFFFFFF returned while the bus is hung, is now clamped so it can no longer corrupt q->head or q->tail. In addition, check the bus_hung flag in mt7921_mac_reset_work() before attempting the reset sequence, reject MCU messages while the bus is hung, and install no-op bus operations when an unrecoverable AER error is detected, preventing further invalid hardware accesses. Due to hardware limitations - such as the lack of a connected hardware reset pin or the absence of host re-probe functionality - affected Wi-Fi devices may not fully recover to a normal operational state after certain errors, even with AER enabled.

In the Linux kernel, the following vulnerability has been resolved: wifi: mt76: mt792x: Fix memory leak in SDIO TX path When tx_prepare_skb() returns an error in the SDIO TX path, the skb is not freed, leading to a memory leak. This can occur when zero-length frames (such as WNM NULL frames) are dropped to prevent potential hardware TX hangs. Fix this by properly releasing the skb with ieee80211_tx_status_ext() when tx_prepare_skb() fails.

In the Linux kernel, the following vulnerability has been resolved: wifi: mt76: fix RX data queuing of RRO 3.0 For RRO 3.0, RX data released from a RRO data queue should be put to the indicator queue. The frames are processed and completed in the context of the indicator queue NAPI, which only polls skbs queued on the MT_RXQ_RRO_IND list; frames queued under the data queue id are left sitting on that list until the data queue NAPI happens to run, stalling and reordering RX data.

In the Linux kernel, the following vulnerability has been resolved: wifi: mt76: mt7996: fix MLD ID in MAC TXD and HIF TXP Problem: MCU command timeout while the firmware state is normal, and the firmware keeps showing the error log "ERROR!! NO PAUSE...". Root cause: If the MLD_ID field in the TXD is neither the primary link id nor the secondary link id, it may lead to a firmware busy loop when the third link is in power saving mode. Remap frames directed to a third link to the primary link wcid. Since TX status events and txfree completions carry the wcid the firmware saw, use the remapped wcid for packet id tracking and non-AQL packet accounting as well, while the frame keeps its original link context for addressing, band and OMAC selection.

In the Linux kernel, the following vulnerability has been resolved: wifi: mt76: fix non-AQL packet accounting for MLO stations __mt76_tx_queue_skb() overrides the wcid passed by the driver with sta->drv_priv, so the wcid might incorrectly be changed after TX, causing wcid->non_aql_packets to be counted on the wrong wcid. For example, on the AP side, if a station's setup link is the 5G link and the station uses 2G to transmit a frame, the value of non_aql_packets is increased on the 5G wcid but decreased on the 2G wcid. Once the inflated counter exceeds MT_MAX_NON_AQL_PKT, the TX scheduler permanently refuses to service the station. Drop the reassignment and account on the wcid used for transmission. This also records the actual wcid in the queue entry.

In the Linux kernel, the following vulnerability has been resolved: wifi: mt76: mt7996: validate RX band_idx before dereferencing phys[] band_idx comes from a 2-bit descriptor field (0-3) and was used directly to index dev->mt76.phys[] (size __MT_MAX_BAND == 3) and dereference the result. A corrupt or reserved descriptor value could index out of bounds or hit a NULL phy on parts with fewer bands. Reject invalid band indices, mirroring mt7996_rx_get_wcid().

In the Linux kernel, the following vulnerability has been resolved: wifi: mt76: mt7915: clear wcid mask under mutex after RCU pointer clear mt7915_remove_interface() cleared the wcid mask bit with no lock held and before clearing the RCU wcid pointer. The mask is a non-atomic RMW shared with the allocators, which all run under dev->mt76.mutex; on DBDC the two wiphys share one mt76_dev, so this raced add_interface/sta_add on the other band and could leak or double-hand-out a wcid. Clearing the bit before the RCU pointer also let a concurrent allocation reuse the index and publish its wcid, which the subsequent NULL assignment then wiped. Move the clear into the existing mutex section, after the RCU pointer is cleared.

In the Linux kernel, the following vulnerability has been resolved: wifi: mt76: mt7915: avoid nss underflow in mt7915_mcu_get_sta_nss If a peer's VHT/HE MCS map has no supported spatial stream (all fields 0x3), the loop exits with nss == 0 and the function returned (u8)-1 (255), which was then written into the firmware sta_rec_bf beamforming fields. Clamp the result to 0.

In the Linux kernel, the following vulnerability has been resolved: wifi: mt76: fix RXDMAD_C buffer recycling race The RXDMAD_C buffers come from the RRO data queues' page pools, which are bound to a different NAPI, so the direct page-pool recycle used here could race the owning NAPI; take the non-direct path as is already done for WED RX queues.

In the Linux kernel, the following vulnerability has been resolved: wifi: mt76: mt7996: bound TLV walk in mt7996_mcu_get_chip_config The response TLV loop advanced by tlv->len without a minimum, so a theoretical firmware response containing a zero-length TLV could spin forever, hanging the CPU during device probe. The u32 payload was also read without bounds checking. Reject a short fixed field, stop on a TLV whose length underruns the header or overruns the skb.

In the Linux kernel, the following vulnerability has been resolved: wifi: mt76: fix out-of-bounds access in mmio copy helpers mt76_mmio_write_copy() and mt76_mmio_read_copy() iterate up to ALIGN(len, 4), so a length that is not a multiple of four reads past the source buffer (write_copy) or writes past the destination (read_copy). Copy the aligned body in the loop and handle the remaining tail through a 4-byte bounce buffer, keeping the register access width unchanged.

In the Linux kernel, the following vulnerability has been resolved: wifi: mt76: mt7915: unwind state on add_interface failure When mt76_wcid_alloc() fails, mt7915_add_interface() returned without clearing the vif_mask/omac_mask bits it had already set, without removing the firmware dev info added earlier, and without clearing a monitor_vif pointer to the vif mac80211 is about to free. mac80211 does not call remove_interface() for a failed add, so the indices and firmware dev entry leaked permanently and testmode could dereference the stale monitor_vif. Add a proper error unwind.

In the Linux kernel, the following vulnerability has been resolved: wifi: mt76: mt7996: hold dev->mt76.mutex while disabling tx worker in SER mt7996_mac_reset_work() parked the tx worker and disabled the RX/TX NAPIs before taking dev->mt76.mutex. mt76_worker_disable()/_enable() are plain kthread park/unpark, not refcounted, and __mt76_set_channel() toggles the same worker and the MT76_RESET bit under the mutex. An L1 SER racing a channel switch could therefore have the worker unparked and MT76_RESET cleared while the reset path resets the DMA rings, corrupting descriptors or tokens. Take the mutex before disabling the worker, as mt7915 does.

In the Linux kernel, the following vulnerability has been resolved: wifi: mt76: mt7996: reserve space for the CSA-abort countdown TLV When a CSA countdown is active, mt7996_mcu_beacon_cntdwn() emits two bss_bcn_cntdwn_tlv entries (the CSA countdown and the CCA-abort BCC), but MT7996_BEACON_UPDATE_SIZE only reserved one. With MBSSID enabled and a near-maximum beacon template the extra 8 bytes could push the offload command past MT7996_MAX_BSS_OFFLOAD_SIZE and trigger skb_over_panic(). Reserve room for both countdown TLVs.