Vulnerabilities

In the Linux kernel, the following vulnerability has been resolved:<br /> <br /> arm64: entry: avoid kprobe recursion<br /> <br /> The cortex_a76_erratum_1463225_debug_handler() function is called when<br /> handling debug exceptions (and synchronous exceptions from BRK<br /> instructions), and so is called when a probed function executes. If the<br /> compiler does not inline cortex_a76_erratum_1463225_debug_handler(), it<br /> can be probed.<br /> <br /> If cortex_a76_erratum_1463225_debug_handler() is probed, any debug<br /> exception or software breakpoint exception will result in recursive<br /> exceptions leading to a stack overflow. This can be triggered with the<br /> ftrace multiple_probes selftest, and as per the example splat below.<br /> <br /> This is a regression caused by commit:<br /> <br /> 6459b8469753e9fe ("arm64: entry: consolidate Cortex-A76 erratum 1463225 workaround")<br /> <br /> ... which removed the NOKPROBE_SYMBOL() annotation associated with the<br /> function.<br /> <br /> My intent was that cortex_a76_erratum_1463225_debug_handler() would be<br /> inlined into its caller, el1_dbg(), which is marked noinstr and cannot<br /> be probed. Mark cortex_a76_erratum_1463225_debug_handler() as<br /> __always_inline to ensure this.<br /> <br /> Example splat prior to this patch (with recursive entries elided):<br /> <br /> | # echo p cortex_a76_erratum_1463225_debug_handler &gt; /sys/kernel/debug/tracing/kprobe_events<br /> | # echo p do_el0_svc &gt;&gt; /sys/kernel/debug/tracing/kprobe_events<br /> | # echo 1 &gt; /sys/kernel/debug/tracing/events/kprobes/enable<br /> | Insufficient stack space to handle exception!<br /> | ESR: 0x0000000096000047 -- DABT (current EL)<br /> | FAR: 0xffff800009cefff0<br /> | Task stack: [0xffff800009cf0000..0xffff800009cf4000]<br /> | IRQ stack: [0xffff800008000000..0xffff800008004000]<br /> | Overflow stack: [0xffff00007fbc00f0..0xffff00007fbc10f0]<br /> | CPU: 0 PID: 145 Comm: sh Not tainted 6.0.0 #2<br /> | Hardware name: linux,dummy-virt (DT)<br /> | pstate: 604003c5 (nZCv DAIF +PAN -UAO -TCO -DIT -SSBS BTYPE=--)<br /> | pc : arm64_enter_el1_dbg+0x4/0x20<br /> | lr : el1_dbg+0x24/0x5c<br /> | sp : ffff800009cf0000<br /> | x29: ffff800009cf0000 x28: ffff000002c74740 x27: 0000000000000000<br /> | x26: 0000000000000000 x25: 0000000000000000 x24: 0000000000000000<br /> | x23: 00000000604003c5 x22: ffff80000801745c x21: 0000aaaac95ac068<br /> | x20: 00000000f2000004 x19: ffff800009cf0040 x18: 0000000000000000<br /> | x17: 0000000000000000 x16: 0000000000000000 x15: 0000000000000000<br /> | x14: 0000000000000000 x13: 0000000000000000 x12: 0000000000000000<br /> | x11: 0000000000000010 x10: ffff800008c87190 x9 : ffff800008ca00d0<br /> | x8 : 000000000000003c x7 : 0000000000000000 x6 : 0000000000000000<br /> | x5 : 0000000000000000 x4 : 0000000000000000 x3 : 00000000000043a4<br /> | x2 : 00000000f2000004 x1 : 00000000f2000004 x0 : ffff800009cf0040<br /> | Kernel panic - not syncing: kernel stack overflow<br /> | CPU: 0 PID: 145 Comm: sh Not tainted 6.0.0 #2<br /> | Hardware name: linux,dummy-virt (DT)<br /> | Call trace:<br /> | dump_backtrace+0xe4/0x104<br /> | show_stack+0x18/0x4c<br /> | dump_stack_lvl+0x64/0x7c<br /> | dump_stack+0x18/0x38<br /> | panic+0x14c/0x338<br /> | test_taint+0x0/0x2c<br /> | panic_bad_stack+0x104/0x118<br /> | handle_bad_stack+0x34/0x48<br /> | __bad_stack+0x78/0x7c<br /> | arm64_enter_el1_dbg+0x4/0x20<br /> | el1h_64_sync_handler+0x40/0x98<br /> | el1h_64_sync+0x64/0x68<br /> | cortex_a76_erratum_1463225_debug_handler+0x0/0x34<br /> ...<br /> | el1h_64_sync_handler+0x40/0x98<br /> | el1h_64_sync+0x64/0x68<br /> | cortex_a76_erratum_1463225_debug_handler+0x0/0x34<br /> ...<br /> | el1h_64_sync_handler+0x40/0x98<br /> | el1h_64_sync+0x64/0x68<br /> | cortex_a76_erratum_1463225_debug_handler+0x0/0x34<br /> | el1h_64_sync_handler+0x40/0x98<br /> | el1h_64_sync+0x64/0x68<br /> | do_el0_svc+0x0/0x28<br /> | el0t_64_sync_handler+0x84/0xf0<br /> | el0t_64_sync+0x18c/0x190<br /> | Kernel Offset: disabled<br /> | CPU features: 0x0080,00005021,19001080<br /> | Memory Limit: none<br /> | ---[ end Kernel panic - not syncing: kernel stack overflow ]---<br /> <br /> With this patch, cortex_a76_erratum_1463225_debug_handler() is inlined<br /> into el1_dbg(), and el1_dbg() cannot be probed:<br /> <br /> | # echo p cortex_a76_erratum_1463225_debug_handler &gt; /sys/kernel/debug/tracing/kprobe_events<br /> | sh: write error: No such file or directory<br /> | # grep -w cortex_a76_errat<br /> ---truncated---

In the Linux kernel, the following vulnerability has been resolved:<br /> <br /> ring-buffer: Check for NULL cpu_buffer in ring_buffer_wake_waiters()<br /> <br /> On some machines the number of listed CPUs may be bigger than the actual<br /> CPUs that exist. The tracing subsystem allocates a per_cpu directory with<br /> access to the per CPU ring buffer via a cpuX file. But to save space, the<br /> ring buffer will only allocate buffers for online CPUs, even though the<br /> CPU array will be as big as the nr_cpu_ids.<br /> <br /> With the addition of waking waiters on the ring buffer when closing the<br /> file, the ring_buffer_wake_waiters() now needs to make sure that the<br /> buffer is allocated (with the irq_work allocated with it) before trying to<br /> wake waiters, as it will cause a NULL pointer dereference.<br /> <br /> While debugging this, I added a NULL check for the buffer itself (which is<br /> OK to do), and also NULL pointer checks against buffer-&gt;buffers (which is<br /> not fine, and will WARN) as well as making sure the CPU number passed in<br /> is within the nr_cpu_ids (which is also not fine if it isn&amp;#39;t).<br /> <br /> <br /> Bugzilla: https://bugzilla.opensuse.org/show_bug.cgi?id=1204705

In the Linux kernel, the following vulnerability has been resolved:<br /> <br /> x86/tdx: Panic on bad configs that #VE on "private" memory access<br /> <br /> All normal kernel memory is "TDX private memory". This includes<br /> everything from kernel stacks to kernel text. Handling<br /> exceptions on arbitrary accesses to kernel memory is essentially<br /> impossible because they can happen in horribly nasty places like<br /> kernel entry/exit. But, TDX hardware can theoretically _deliver_<br /> a virtualization exception (#VE) on any access to private memory.<br /> <br /> But, it&amp;#39;s not as bad as it sounds. TDX can be configured to never<br /> deliver these exceptions on private memory with a "TD attribute"<br /> called ATTR_SEPT_VE_DISABLE. The guest has no way to *set* this<br /> attribute, but it can check it.<br /> <br /> Ensure ATTR_SEPT_VE_DISABLE is set in early boot. panic() if it<br /> is unset. There is no sane way for Linux to run with this<br /> attribute clear so a panic() is appropriate.<br /> <br /> There&amp;#39;s small window during boot before the check where kernel<br /> has an early #VE handler. But the handler is only for port I/O<br /> and will also panic() as soon as it sees any other #VE, such as<br /> a one generated by a private memory access.<br /> <br /> [ dhansen: Rewrite changelog and rebase on new tdx_parse_tdinfo().<br /> Add Kirill&amp;#39;s tested-by because I made changes since<br /> he wrote this. ]

In the Linux kernel, the following vulnerability has been resolved:<br /> <br /> KVM: Initialize gfn_to_pfn_cache locks in dedicated helper<br /> <br /> Move the gfn_to_pfn_cache lock initialization to another helper and<br /> call the new helper during VM/vCPU creation. There are race<br /> conditions possible due to kvm_gfn_to_pfn_cache_init()&amp;#39;s<br /> ability to re-initialize the cache&amp;#39;s locks.<br /> <br /> For example: a race between ioctl(KVM_XEN_HVM_EVTCHN_SEND) and<br /> kvm_gfn_to_pfn_cache_init() leads to a corrupted shinfo gpc lock.<br /> <br /> (thread 1) | (thread 2)<br /> |<br /> kvm_xen_set_evtchn_fast |<br /> read_lock_irqsave(&amp;gpc-&gt;lock, ...) |<br /> | kvm_gfn_to_pfn_cache_init<br /> | rwlock_init(&amp;gpc-&gt;lock)<br /> read_unlock_irqrestore(&amp;gpc-&gt;lock, ...) |<br /> <br /> Rename "cache_init" and "cache_destroy" to activate+deactivate to<br /> avoid implying that the cache really is destroyed/freed.<br /> <br /> Note, there more races in the newly named kvm_gpc_activate() that will<br /> be addressed separately.<br /> <br /> [sean: call out that this is a bug fix]

In the Linux kernel, the following vulnerability has been resolved:<br /> <br /> KVM: x86: smm: number of GPRs in the SMRAM image depends on the image format<br /> <br /> On 64 bit host, if the guest doesn&amp;#39;t have X86_FEATURE_LM, KVM will<br /> access 16 gprs to 32-bit smram image, causing out-ouf-bound ram<br /> access.<br /> <br /> On 32 bit host, the rsm_load_state_64/enter_smm_save_state_64<br /> is compiled out, thus access overflow can&amp;#39;t happen.

In the Linux kernel, the following vulnerability has been resolved:<br /> <br /> KVM: Reject attempts to consume or refresh inactive gfn_to_pfn_cache<br /> <br /> Reject kvm_gpc_check() and kvm_gpc_refresh() if the cache is inactive.<br /> Not checking the active flag during refresh is particularly egregious, as<br /> KVM can end up with a valid, inactive cache, which can lead to a variety<br /> of use-after-free bugs, e.g. consuming a NULL kernel pointer or missing<br /> an mmu_notifier invalidation due to the cache not being on the list of<br /> gfns to invalidate.<br /> <br /> Note, "active" needs to be set if and only if the cache is on the list<br /> of caches, i.e. is reachable via mmu_notifier events. If a relevant<br /> mmu_notifier event occurs while the cache is "active" but not on the<br /> list, KVM will not acquire the cache&amp;#39;s lock and so will not serailize<br /> the mmu_notifier event with active users and/or kvm_gpc_refresh().<br /> <br /> A race between KVM_XEN_ATTR_TYPE_SHARED_INFO and KVM_XEN_HVM_EVTCHN_SEND<br /> can be exploited to trigger the bug.<br /> <br /> 1. Deactivate shinfo cache:<br /> <br /> kvm_xen_hvm_set_attr<br /> case KVM_XEN_ATTR_TYPE_SHARED_INFO<br /> kvm_gpc_deactivate<br /> kvm_gpc_unmap<br /> gpc-&gt;valid = false<br /> gpc-&gt;khva = NULL<br /> gpc-&gt;active = false<br /> <br /> Result: active = false, valid = false<br /> <br /> 2. Cause cache refresh:<br /> <br /> kvm_arch_vm_ioctl<br /> case KVM_XEN_HVM_EVTCHN_SEND<br /> kvm_xen_hvm_evtchn_send<br /> kvm_xen_set_evtchn<br /> kvm_xen_set_evtchn_fast<br /> kvm_gpc_check<br /> return -EWOULDBLOCK because !gpc-&gt;valid<br /> kvm_xen_set_evtchn_fast<br /> return -EWOULDBLOCK<br /> kvm_gpc_refresh<br /> hva_to_pfn_retry<br /> gpc-&gt;valid = true<br /> gpc-&gt;khva = not NULL<br /> <br /> Result: active = false, valid = true<br /> <br /> 3. Race ioctl KVM_XEN_HVM_EVTCHN_SEND against ioctl<br /> KVM_XEN_ATTR_TYPE_SHARED_INFO:<br /> <br /> kvm_arch_vm_ioctl<br /> case KVM_XEN_HVM_EVTCHN_SEND<br /> kvm_xen_hvm_evtchn_send<br /> kvm_xen_set_evtchn<br /> kvm_xen_set_evtchn_fast<br /> read_lock gpc-&gt;lock<br /> kvm_xen_hvm_set_attr case<br /> KVM_XEN_ATTR_TYPE_SHARED_INFO<br /> mutex_lock kvm-&gt;lock<br /> kvm_xen_shared_info_init<br /> kvm_gpc_activate<br /> gpc-&gt;khva = NULL<br /> kvm_gpc_check<br /> [ Check passes because gpc-&gt;valid is<br /> still true, even though gpc-&gt;khva<br /> is already NULL. ]<br /> shinfo = gpc-&gt;khva<br /> pending_bits = shinfo-&gt;evtchn_pending<br /> CRASH: test_and_set_bit(..., pending_bits)

In the Linux kernel, the following vulnerability has been resolved:<br /> <br /> net: tun: Fix memory leaks of napi_get_frags<br /> <br /> kmemleak reports after running test_progs:<br /> <br /> unreferenced object 0xffff8881b1672dc0 (size 232):<br /> comm "test_progs", pid 394388, jiffies 4354712116 (age 841.975s)<br /> hex dump (first 32 bytes):<br /> e0 84 d7 a8 81 88 ff ff 80 2c 67 b1 81 88 ff ff .........,g.....<br /> 00 40 c5 9b 81 88 ff ff 00 00 00 00 00 00 00 00 .@..............<br /> backtrace:<br /> [] napi_skb_cache_get+0xd4/0x150<br /> [] __napi_build_skb+0x15/0x50<br /> [] __napi_alloc_skb+0x26e/0x540<br /> [] napi_get_frags+0x59/0x140<br /> [] tun_get_user+0x183d/0x3bb0 [tun]<br /> [] tun_chr_write_iter+0xc0/0x1b1 [tun]<br /> [] do_iter_readv_writev+0x19f/0x320<br /> [] do_iter_write+0x135/0x630<br /> [] vfs_writev+0x12e/0x440<br /> [] do_writev+0x104/0x280<br /> [] do_syscall_64+0x3b/0x90<br /> [] entry_SYSCALL_64_after_hwframe+0x63/0xcd<br /> <br /> The issue occurs in the following scenarios:<br /> tun_get_user()<br /> napi_gro_frags()<br /> napi_frags_finish()<br /> case GRO_NORMAL:<br /> gro_normal_one()<br /> list_add_tail(&amp;skb-&gt;list, &amp;napi-&gt;rx_list);<br /> rx_count

In the Linux kernel, the following vulnerability has been resolved:<br /> <br /> bpf: Fix wrong reg type conversion in release_reference()<br /> <br /> Some helper functions will allocate memory. To avoid memory leaks, the<br /> verifier requires the eBPF program to release these memories by calling<br /> the corresponding helper functions.<br /> <br /> When a resource is released, all pointer registers corresponding to the<br /> resource should be invalidated. The verifier use release_references() to<br /> do this job, by apply __mark_reg_unknown() to each relevant register.<br /> <br /> It will give these registers the type of SCALAR_VALUE. A register that<br /> will contain a pointer value at runtime, but of type SCALAR_VALUE, which<br /> may allow the unprivileged user to get a kernel pointer by storing this<br /> register into a map.<br /> <br /> Using __mark_reg_not_init() while NOT allow_ptr_leaks can mitigate this<br /> problem.

In the Linux kernel, the following vulnerability has been resolved:<br /> <br /> HID: hyperv: fix possible memory leak in mousevsc_probe()<br /> <br /> If hid_add_device() returns error, it should call hid_destroy_device()<br /> to free hid_dev which is allocated in hid_allocate_device().

In the Linux kernel, the following vulnerability has been resolved:<br /> <br /> bpftool: Fix NULL pointer dereference when pin {PROG, MAP, LINK} without FILE<br /> <br /> When using bpftool to pin {PROG, MAP, LINK} without FILE,<br /> segmentation fault will occur. The reson is that the lack<br /> of FILE will cause strlen to trigger NULL pointer dereference.<br /> The corresponding stacktrace is shown below:<br /> <br /> do_pin<br /> do_pin_any<br /> do_pin_fd<br /> mount_bpffs_for_pin<br /> strlen(name)

In the Linux kernel, the following vulnerability has been resolved:<br /> <br /> wifi: mac80211: fix general-protection-fault in ieee80211_subif_start_xmit()<br /> <br /> When device is running and the interface status is changed, the gpf issue<br /> is triggered. The problem triggering process is as follows:<br /> Thread A: Thread B<br /> ieee80211_runtime_change_iftype() process_one_work()<br /> ... ...<br /> ieee80211_do_stop() ...<br /> ... ...<br /> sdata-&gt;bss = NULL ...<br /> ... ieee80211_subif_start_xmit()<br /> ieee80211_multicast_to_unicast<br /> //!sdata-&gt;bss-&gt;multicast_to_unicast<br /> cause gpf issue<br /> <br /> When the interface status is changed, the sending queue continues to send<br /> packets. After the bss is set to NULL, the bss is accessed. As a result,<br /> this causes a general-protection-fault issue.<br /> <br /> The following is the stack information:<br /> general protection fault, probably for non-canonical address<br /> 0xdffffc000000002f: 0000 [#1] PREEMPT SMP KASAN<br /> KASAN: null-ptr-deref in range [0x0000000000000178-0x000000000000017f]<br /> Workqueue: mld mld_ifc_work<br /> RIP: 0010:ieee80211_subif_start_xmit+0x25b/0x1310<br /> Call Trace:<br /> <br /> dev_hard_start_xmit+0x1be/0x990<br /> __dev_queue_xmit+0x2c9a/0x3b60<br /> ip6_finish_output2+0xf92/0x1520<br /> ip6_finish_output+0x6af/0x11e0<br /> ip6_output+0x1ed/0x540<br /> mld_sendpack+0xa09/0xe70<br /> mld_ifc_work+0x71c/0xdb0<br /> process_one_work+0x9bf/0x1710<br /> worker_thread+0x665/0x1080<br /> kthread+0x2e4/0x3a0<br /> ret_from_fork+0x1f/0x30<br />

In the Linux kernel, the following vulnerability has been resolved:<br /> <br /> bpf, verifier: Fix memory leak in array reallocation for stack state<br /> <br /> If an error (NULL) is returned by krealloc(), callers of realloc_array()<br /> were setting their allocation pointers to NULL, but on error krealloc()<br /> does not touch the original allocation. This would result in a memory<br /> resource leak. Instead, free the old allocation on the error handling<br /> path.<br /> <br /> The memory leak information is as follows as also reported by Zhengchao:<br /> <br /> unreferenced object 0xffff888019801800 (size 256):<br /> comm "bpf_repo", pid 6490, jiffies 4294959200 (age 17.170s)<br /> hex dump (first 32 bytes):<br /> 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 ................<br /> 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 ................<br /> backtrace:<br /> [] __kmalloc_node_track_caller+0x45/0xc0<br /> [] krealloc+0x83/0xd0<br /> [] realloc_array+0x82/0xe2<br /> [] grow_stack_state+0xfb/0x186<br /> [] check_mem_access.cold+0x141/0x1341<br /> [] do_check_common+0x5358/0xb350<br /> [] bpf_check.cold+0xc3/0x29d<br /> [] bpf_prog_load+0x13db/0x2240<br /> [] __sys_bpf+0x1605/0x4ce0<br /> [] __x64_sys_bpf+0x75/0xb0<br /> [] do_syscall_64+0x35/0x80<br /> [] entry_SYSCALL_64_after_hwframe+0x63/0xcd