Vulnerabilities

In the Linux kernel, the following vulnerability has been resolved:<br /> <br /> wifi: rtw89: pci: validate sequence number of TX release report<br /> <br /> Hardware rarely reports abnormal sequence number in TX release report,<br /> which will access out-of-bounds of wd_ring-&gt;pages array, causing NULL<br /> pointer dereference.<br /> <br /> BUG: kernel NULL pointer dereference, address: 0000000000000000<br /> #PF: supervisor read access in kernel mode<br /> #PF: error_code(0x0000) - not-present page<br /> PGD 0 P4D 0<br /> Oops: 0000 [#1] PREEMPT SMP NOPTI<br /> CPU: 1 PID: 1085 Comm: irq/129-rtw89_p Tainted: G S U<br /> 6.1.145-17510-g2f3369c91536 #1 (HASH:69e8 1)<br /> Call Trace:<br /> <br /> rtw89_pci_release_tx+0x18f/0x300 [rtw89_pci (HASH:4c83 2)]<br /> rtw89_pci_napi_poll+0xc2/0x190 [rtw89_pci (HASH:4c83 2)]<br /> net_rx_action+0xfc/0x460 net/core/dev.c:6578 net/core/dev.c:6645 net/core/dev.c:6759<br /> handle_softirqs+0xbe/0x290 kernel/softirq.c:601<br /> ? rtw89_pci_interrupt_threadfn+0xc5/0x350 [rtw89_pci (HASH:4c83 2)]<br /> __local_bh_enable_ip+0xeb/0x120 kernel/softirq.c:499 kernel/softirq.c:423<br /> <br /> <br /> rtw89_pci_interrupt_threadfn+0xf8/0x350 [rtw89_pci (HASH:4c83 2)]<br /> ? irq_thread+0xa7/0x340 kernel/irq/manage.c:0<br /> irq_thread+0x177/0x340 kernel/irq/manage.c:1205 kernel/irq/manage.c:1314<br /> ? thaw_kernel_threads+0xb0/0xb0 kernel/irq/manage.c:1202<br /> ? irq_forced_thread_fn+0x80/0x80 kernel/irq/manage.c:1220<br /> kthread+0xea/0x110 kernel/kthread.c:376<br /> ? synchronize_irq+0x1a0/0x1a0 kernel/irq/manage.c:1287<br /> ? kthread_associate_blkcg+0x80/0x80 kernel/kthread.c:331<br /> ret_from_fork+0x1f/0x30 arch/x86/entry/entry_64.S:295<br /> <br /> <br /> To prevent crash, validate rpp_info.seq before using.

In the Linux kernel, the following vulnerability has been resolved:<br /> <br /> LoongArch: Make cpumask_of_node() robust against NUMA_NO_NODE<br /> <br /> The arch definition of cpumask_of_node() cannot handle NUMA_NO_NODE -<br /> which is a valid index - so add a check for this.

In the Linux kernel, the following vulnerability has been resolved:<br /> <br /> PCI: Fix pci_slot_trylock() error handling<br /> <br /> Commit a4e772898f8b ("PCI: Add missing bridge lock to pci_bus_lock()")<br /> delegates the bridge device&amp;#39;s pci_dev_trylock() to pci_bus_trylock() in<br /> pci_slot_trylock(), but it forgets to remove the corresponding<br /> pci_dev_unlock() when pci_bus_trylock() fails.<br /> <br /> Before a4e772898f8b, the code did:<br /> <br /> if (!pci_dev_trylock(dev)) /* subordinate) {<br /> if (!pci_bus_trylock(dev-&gt;subordinate)) {<br /> pci_dev_unlock(dev); /*

In the Linux kernel, the following vulnerability has been resolved:<br /> <br /> tracing: ring-buffer: Fix to check event length before using<br /> <br /> Check the event length before adding it for accessing next index in<br /> rb_read_data_buffer(). Since this function is used for validating<br /> possibly broken ring buffers, the length of the event could be broken.<br /> In that case, the new event (e + len) can point a wrong address.<br /> To avoid invalid memory access at boot, check whether the length of<br /> each event is in the possible range before using it.

In the Linux kernel, the following vulnerability has been resolved:<br /> <br /> net: do not pass flow_id to set_rps_cpu()<br /> <br /> Blamed commit made the assumption that the RPS table for each receive<br /> queue would have the same size, and that it would not change.<br /> <br /> Compute flow_id in set_rps_cpu(), do not assume we can use the value<br /> computed by get_rps_cpu(). Otherwise we risk out-of-bound access<br /> and/or crashes.

In the Linux kernel, the following vulnerability has been resolved:<br /> <br /> media: mtk-mdp: Fix error handling in probe function<br /> <br /> Add mtk_mdp_unregister_m2m_device() on the error handling path to prevent<br /> resource leak.<br /> <br /> Add check for the return value of vpu_get_plat_device() to prevent null<br /> pointer dereference. And vpu_get_plat_device() increases the reference<br /> count of the returned platform device. Add platform_device_put() to<br /> prevent reference leak.

In the Linux kernel, the following vulnerability has been resolved:<br /> <br /> minix: Add required sanity checking to minix_check_superblock()<br /> <br /> The fs/minix implementation of the minix filesystem does not currently<br /> support any other value for s_log_zone_size than 0. This is also the<br /> only value supported in util-linux; see mkfs.minix.c line 511. In<br /> addition, this patch adds some sanity checking for the other minix<br /> superblock fields, and moves the minix_blocks_needed() checks for the<br /> zmap and imap also to minix_check_super_block().<br /> <br /> This also closes a related syzbot bug report.

In the Linux kernel, the following vulnerability has been resolved:<br /> <br /> dpaa2-switch: validate num_ifs to prevent out-of-bounds write<br /> <br /> The driver obtains sw_attr.num_ifs from firmware via dpsw_get_attributes()<br /> but never validates it against DPSW_MAX_IF (64). This value controls<br /> iteration in dpaa2_switch_fdb_get_flood_cfg(), which writes port indices<br /> into the fixed-size cfg-&gt;if_id[DPSW_MAX_IF] array. When firmware reports<br /> num_ifs &gt;= 64, the loop can write past the array bounds.<br /> <br /> Add a bound check for num_ifs in dpaa2_switch_init().<br /> <br /> dpaa2_switch_fdb_get_flood_cfg() appends the control interface (port<br /> num_ifs) after all matched ports. When num_ifs == DPSW_MAX_IF and all<br /> ports match the flood filter, the loop fills all 64 slots and the control<br /> interface write overflows by one entry.<br /> <br /> The check uses &gt;= because num_ifs == DPSW_MAX_IF is also functionally<br /> broken.<br /> <br /> build_if_id_bitmap() silently drops any ID &gt;= 64:<br /> if (id[i]

In the Linux kernel, the following vulnerability has been resolved:<br /> <br /> drm/amdkfd: Fix out-of-bounds write in kfd_event_page_set()<br /> <br /> The kfd_event_page_set() function writes KFD_SIGNAL_EVENT_LIMIT * 8<br /> bytes via memset without checking the buffer size parameter. This allows<br /> unprivileged userspace to trigger an out-of bounds kernel memory write<br /> by passing a small buffer, leading to potential privilege<br /> escalation.

In the Linux kernel, the following vulnerability has been resolved:<br /> <br /> ASoC: qcom: q6asm: drop DSP responses for closed data streams<br /> <br /> &amp;#39;Commit a354f030dbce ("ASoC: qcom: q6asm: handle the responses<br /> after closing")&amp;#39; attempted to ignore DSP responses arriving<br /> after a stream had been closed.<br /> <br /> However, those responses were still handled, causing lockups.<br /> <br /> Fix this by unconditionally dropping all DSP responses associated with<br /> closed data streams.

In the Linux kernel, the following vulnerability has been resolved:<br /> <br /> atm: fore200e: fix use-after-free in tasklets during device removal<br /> <br /> When the PCA-200E or SBA-200E adapter is being detached, the fore200e<br /> is deallocated. However, the tx_tasklet or rx_tasklet may still be running<br /> or pending, leading to use-after-free bug when the already freed fore200e<br /> is accessed again in fore200e_tx_tasklet() or fore200e_rx_tasklet().<br /> <br /> One of the race conditions can occur as follows:<br /> <br /> CPU 0 (cleanup) | CPU 1 (tasklet)<br /> fore200e_pca_remove_one() | fore200e_interrupt()<br /> fore200e_shutdown() | tasklet_schedule()<br /> kfree(fore200e) | fore200e_tx_tasklet()<br /> | fore200e-&gt; // UAF<br /> <br /> Fix this by ensuring tx_tasklet or rx_tasklet is properly canceled before<br /> the fore200e is released. Add tasklet_kill() in fore200e_shutdown() to<br /> synchronize with any pending or running tasklets. Moreover, since<br /> fore200e_reset() could prevent further interrupts or data transfers,<br /> the tasklet_kill() should be placed after fore200e_reset() to prevent<br /> the tasklet from being rescheduled in fore200e_interrupt(). Finally,<br /> it only needs to do tasklet_kill() when the fore200e state is greater<br /> than or equal to FORE200E_STATE_IRQ, since tasklets are uninitialized<br /> in earlier states. In a word, the tasklet_kill() should be placed in<br /> the FORE200E_STATE_IRQ branch within the switch...case structure.<br /> <br /> This bug was identified through static analysis.

In the Linux kernel, the following vulnerability has been resolved:<br /> <br /> fbdev: vt8500lcdfb: fix missing dma_free_coherent()<br /> <br /> fbi-&gt;fb.screen_buffer is allocated with dma_alloc_coherent() but is not<br /> freed if the error path is reached.