Wind River Support Network

Description

In the Linux kernel, the following vulnerability has been resolved:EOL][EOL]KVM: SVM: Reject SEV{-ES} intra host migration if vCPU creation is in-flight[EOL][EOL]Reject migration of SEV{-ES} state if either the source or destination VM[EOL]is actively creating a vCPU, i.e. if kvm_vm_ioctl_create_vcpu() is in the[EOL]section between incrementing created_vcpus and online_vcpus.  The bulk of[EOL]vCPU creation runs _outside_ of kvm->lock to allow creating multiple vCPUs[EOL]in parallel, and so sev_info.es_active can get toggled from false=>true in[EOL]the destination VM after (or during) svm_vcpu_create(), resulting in an[EOL]SEV{-ES} VM effectively having a non-SEV{-ES} vCPU.[EOL][EOL]The issue manifests most visibly as a crash when trying to free a vCPU's[EOL]NULL VMSA page in an SEV-ES VM, but any number of things can go wrong.[EOL][EOL]  BUG: unable to handle page fault for address: ffffebde00000000[EOL]  #PF: supervisor read access in kernel mode[EOL]  #PF: error_code(0x0000) - not-present page[EOL]  PGD 0 P4D 0[EOL]  Oops: Oops: 0000 [#1] SMP KASAN NOPTI[EOL]  CPU: 227 UID: 0 PID: 64063 Comm: syz.5.60023 Tainted: G     U     O        6.15.0-smp-DEV #2 NONE[EOL]  Tainted: [U]=USER, [O]=OOT_MODULE[EOL]  Hardware name: Google, Inc. Arcadia_IT_80/Arcadia_IT_80, BIOS 12.52.0-0 10/28/2024[EOL]  RIP: 0010:constant_test_bit arch/x86/include/asm/bitops.h:206 [inline][EOL]  RIP: 0010:arch_test_bit arch/x86/include/asm/bitops.h:238 [inline][EOL]  RIP: 0010:_test_bit include/asm-generic/bitops/instrumented-non-atomic.h:142 [inline][EOL]  RIP: 0010:PageHead include/linux/page-flags.h:866 [inline][EOL]  RIP: 0010:___free_pages+0x3e/0x120 mm/page_alloc.c:5067[EOL]  Code: <49> f7 06 40 00 00 00 75 05 45 31 ff eb 0c 66 90 4c 89 f0 4c 39 f0[EOL]  RSP: 0018:ffff8984551978d0 EFLAGS: 00010246[EOL]  RAX: 0000777f80000001 RBX: 0000000000000000 RCX: ffffffff918aeb98[EOL]  RDX: 0000000000000000 RSI: 0000000000000008 RDI: ffffebde00000000[EOL]  RBP: 0000000000000000 R08: ffffebde00000007 R09: 1ffffd7bc0000000[EOL]  R10: dffffc0000000000 R11: fffff97bc0000001 R12: dffffc0000000000[EOL]  R13: ffff8983e19751a8 R14: ffffebde00000000 R15: 1ffffd7bc0000000[EOL]  FS:  0000000000000000(0000) GS:ffff89ee661d3000(0000) knlGS:0000000000000000[EOL]  CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033[EOL]  CR2: ffffebde00000000 CR3: 000000793ceaa000 CR4: 0000000000350ef0[EOL]  DR0: 0000000000000000 DR1: 0000000000000b5f DR2: 0000000000000000[EOL]  DR3: 0000000000000000 DR6: 00000000ffff0ff0 DR7: 0000000000000400[EOL]  Call Trace:[EOL]   <TASK>[EOL]   sev_free_vcpu+0x413/0x630 arch/x86/kvm/svm/sev.c:3169[EOL]   svm_vcpu_free+0x13a/0x2a0 arch/x86/kvm/svm/svm.c:1515[EOL]   kvm_arch_vcpu_destroy+0x6a/0x1d0 arch/x86/kvm/x86.c:12396[EOL]   kvm_vcpu_destroy virt/kvm/kvm_main.c:470 [inline][EOL]   kvm_destroy_vcpus+0xd1/0x300 virt/kvm/kvm_main.c:490[EOL]   kvm_arch_destroy_vm+0x636/0x820 arch/x86/kvm/x86.c:12895[EOL]   kvm_put_kvm+0xb8e/0xfb0 virt/kvm/kvm_main.c:1310[EOL]   kvm_vm_release+0x48/0x60 virt/kvm/kvm_main.c:1369[EOL]   __fput+0x3e4/0x9e0 fs/file_table.c:465[EOL]   task_work_run+0x1a9/0x220 kernel/task_work.c:227[EOL]   exit_task_work include/linux/task_work.h:40 [inline][EOL]   do_exit+0x7f0/0x25b0 kernel/exit.c:953[EOL]   do_group_exit+0x203/0x2d0 kernel/exit.c:1102[EOL]   get_signal+0x1357/0x1480 kernel/signal.c:3034[EOL]   arch_do_signal_or_restart+0x40/0x690 arch/x86/kernel/signal.c:337[EOL]   exit_to_user_mode_loop kernel/entry/common.c:111 [inline][EOL]   exit_to_user_mode_prepare include/linux/entry-common.h:329 [inline][EOL]   __syscall_exit_to_user_mode_work kernel/entry/common.c:207 [inline][EOL]   syscall_exit_to_user_mode+0x67/0xb0 kernel/entry/common.c:218[EOL]   do_syscall_64+0x7c/0x150 arch/x86/entry/syscall_64.c:100[EOL]   entry_SYSCALL_64_after_hwframe+0x76/0x7e[EOL]  RIP: 0033:0x7f87a898e969[EOL]   </TASK>[EOL]  Modules linked in: gq(O)[EOL]  gsmi: Log Shutdown Reason 0x03[EOL]  CR2: ffffebde00000000[EOL]  ---[ end trace 0000000000000000 ]---[EOL][EOL]Deliberately don't check for a NULL VMSA when freeing the vCPU, as crashing[EOL]the host is likely desirable due to the VMSA being consumed by hardware.[EOL]E.g. if KVM manages to allow VMRUN on the vCPU, hardware may read/write a[EOL]bogus VMSA page.  Accessing P[EOL]---truncated---

CREATE(Triage):(User=admin) [CVE-2025-38455 (https://nvd.nist.gov/vuln/detail/CVE-2025-38455)