{"dataType":"CVE_RECORD","dataVersion":"5.2","cveMetadata":{"cveId":"CVE-2026-90347","assignerOrgId":"416baaa9-dc9f-4396-8d5f-8c081fb06d67","state":"PUBLISHED","assignerShortName":"Linux","dateReserved":"2026-09-11T19:38:34.805Z","datePublished":"2026-09-17T16:08:56.232Z","dateUpdated":"2026-09-18T17:54:46.679Z"},"containers":{"cna":{"providerMetadata":{"orgId":"416baaa9-dc9f-4396-8d5f-8c081fb06d67","shortName":"Linux","dateUpdated":"2026-09-18T17:54:46.679Z"},"descriptions":[{"lang":"en","value":"In the Linux kernel, the following vulnerability has been resolved:\n\narm64: ptrace: Keep 'orig_x0' in-sync with x0 on syscall entry\n\nCommit e057b9477232 (\"arm64: syscall: Ensure saved x0 is kept in-sync\nwith tracer updates\") attempted to resolve a long-standing issue with\nsyscall entry tracing, where a tracer is able to manipulate the first\nsyscall argument without being subjected to seccomp or audit checking.\n\nUnfortunately, that fix was incomplete [1], as it failed to update\n'orig_x0' between a tracer updating x0 during a seccomp ptrace exit\n(SECCOMP_RET_TRACE) and the seccomp filter being re-evaluated.\n\nRather than add hooks to the core seccomp code, instead move the\nsynchronisation code into the ptrace GPR and syscall setting code so\nthat 'orig_x0' is kept up to date with x0 whenever we're stopped on the\nsyscall entry path."}],"metrics":[{"cvssV3_1":{"version":"3.1","vectorString":"CVSS:3.1/AV:L/AC:L/PR:L/UI:N/S:C/C:H/I:H/A:N","baseScore":8.4,"baseSeverity":"HIGH"},"scenarios":[{"lang":"en","value":"AV:L - The attacker-controlled value is x0 written by ptrace(PTRACE_SETREGSET, NT_PRSTATUS) into gpr_set() while the tracee is stopped in syscall_trace_enter() after ptrace_report_syscall_entry() or ptrace_event(PTRACE_EVENT_SECCOMP) from SECCOMP_RET_TRACE. That write is a local ptrace syscall, not a received protocol message.\nAC:L - The tracer owns the stop: waitpid() for the PTRACE_SYSCALL enter-stop or PTRACE_EVENT_SECCOMP trap, PTRACE_SETREGSET to replace regs[0], then PTRACE_CONT. el0_svc_common() then reaches secure_computing() with orig_x0 still equal to the pre-ptrace x0. No race or uninfluenced layout is required.\nPR:L - __ptrace_may_access() allows a same-uid tracer, and default yama ptrace_scope (YAMA_SCOPE_RELATIONAL) permits PTRACE_TRACEME or PTRACE_ATTACH to a descendant without CAP_SYS_PTRACE. An ordinary user can fork, trace the child, and issue PTRACE_SETREGSET; init-namespace root is not required.\nUI:N - The attacker issues PTRACE_TRACEME/PTRACE_SYSCALL, PTRACE_SETREGSET of NT_PRSTATUS, and the in-flight syscall on their own descendant. No separate victim action such as mounting a filesystem or opening a file is required.\nS:C - populate_seccomp_data() exposes orig_x0 as sd.args[0] to a parent-installed seccomp filter (container runtime, systemd RestrictNamespaces, Android minijail). gpr_set() updates regs[0] without orig_x0, so invoke_syscall() via SC_ARM64_REGS_TO_ARGS runs a denied first argument while the filter still allows the old orig_x0, escaping that sandbox.\nC:H - syscall_get_arguments() returns stale orig_x0 to seccomp while __arm64_sys_* wrappers take regs[0]. A tracer starts an allowed openat(dfd) or socket(domain) and replaces x0 with a blocked directory fd or AF_NETLINK/AF_PACKET, so the syscall discloses files or netlink state the filter was written to deny.\nI:H - arm64 clone() takes clone_flags in x0 (CONFIG_CLONE_BACKWARDS). After gpr_set() writes regs[0] without orig_x0, __seccomp_filter() (including the SECCOMP_RET_TRACE recheck) still sees the allowed flags while invoke_syscall() honors CLONE_NEWUSER/CLONE_NEWNS the policy forbade; substituting write/ioctl's fd in x0 likewise modifies data the filter blocked.\nA:N - The bug is orig_x0 remaining stale after gpr_set() so populate_seccomp_data() and audit_syscall_entry() inspect the wrong first argument. That is a seccomp/audit policy bypass, not memory corruption, and it does not oops, panic, deadlock, or hang the kernel."}]}],"affected":[{"product":"Linux","vendor":"Linux","defaultStatus":"unaffected","repo":"https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git","programFiles":["arch/arm64/kernel/ptrace.c"],"versions":[{"version":"a5cd110cb8369d6b37ef5ccfe56b3fa1338c9615","lessThan":"f433869f23841a50455c4087d85540d5b4d37cde","status":"affected","versionType":"git"},{"version":"a5cd110cb8369d6b37ef5ccfe56b3fa1338c9615","lessThan":"88b839ce497ccb1ff92f7ae742c78dd2937ba572","status":"affected","versionType":"git"}]},{"product":"Linux","vendor":"Linux","defaultStatus":"affected","repo":"https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git","programFiles":["arch/arm64/kernel/ptrace.c"],"versions":[{"version":"4.8","status":"affected"},{"version":"0","lessThan":"4.8","status":"unaffected","versionType":"semver"},{"version":"7.2.6","lessThanOrEqual":"7.2.*","status":"unaffected","versionType":"semver"},{"version":"7.3-rc1","lessThanOrEqual":"*","status":"unaffected","versionType":"original_commit_for_fix"}]}],"cpeApplicability":[{"nodes":[{"operator":"OR","negate":false,"cpeMatch":[{"vulnerable":true,"criteria":"cpe:2.3:o:linux:linux_kernel:*:*:*:*:*:*:*:*","versionStartIncluding":"4.8","versionEndExcluding":"7.2.6"},{"vulnerable":true,"criteria":"cpe:2.3:o:linux:linux_kernel:*:*:*:*:*:*:*:*","versionStartIncluding":"4.8","versionEndExcluding":"7.3-rc1"}]}]}],"references":[{"url":"https://git.kernel.org/stable/c/f433869f23841a50455c4087d85540d5b4d37cde"},{"url":"https://git.kernel.org/stable/c/88b839ce497ccb1ff92f7ae742c78dd2937ba572"}],"title":"arm64: ptrace: Keep 'orig_x0' in-sync with x0 on syscall entry","x_generator":{"engine":"bippy-1.2.0"}}}}