{"id":"CVE-2026-9728","aliases":["EUVD-2026-64897","GHSA-47q2-w832-7w67"],"status":"published","title":"TOCTOU race in mbox_send syscall verifier allows userspace to leak kernel memory","description":"The userspace syscall verifier z_vrfy_mbox_send() in drivers/mbox/mbox_handlers.c validated the nested msg-\u003edata/msg-\u003esize fields by reading them directly out of live userspace memory, and then forwarded the original, still-mutable userspace struct mbox_msg * pointer to z_impl_mbox_send() and the underlying driver. Between the access check and the driver's use of msg-\u003edata, the validated pointer could be replaced, leaving a time-of-check/time-of-use window.\n\nOn a system built with CONFIG_USERSPACE, any unprivileged userspace thread may invoke the mbox_send() system call. A second thread sharing the caller's address space can race to overwrite msg-\u003edata with a supervisor (kernel) address after the verifier's bounds check has passed but before the driver dereferences it. The driver then reads from the attacker-chosen address in supervisor context (for example memcpy(\u0026data32, msg-\u003edata, msg-\u003esize) in the NXP mailbox driver, whose bytes are subsequently emitted to the peer mailbox endpoint).\n\nThe impact is a userspace-to-supervisor access-control bypass: disclosure of kernel memory contents (high confidentiality impact), or, for an invalid/unmapped target address, a faulting kernel read causing denial of service. The fix snapshots the entire struct mbox_msg into a kernel-stack copy with k_usermode_from_copy() and validates and forwards that immutable copy, closing the race.","severity":{"version":"3.1","score":6.4,"vector":"CVSS:3.1/AV:L/AC:H/PR:L/UI:N/S:C/C:H/I:N/A:L","level":"medium","assessor":"cna","from":"cve"},"epss":{"day":"2026-11-10","score":0.0011,"percentile":0.01273,"model":"v2026.06.15"},"facts":{"in_kev":false,"known_exploited":false,"public_exploit":false,"fix_available":false},"cwes":["CWE-367"],"published":"2026-08-24T14:28:46Z","modified":"2026-11-04T04:59:49Z","products":[{"vendor":"zephyrproject","product":"zephyr"},{"vendor":"zephyrproject-rtos","product":"zephyr"}],"references":[{"url":"https://github.com/zephyrproject-rtos/zephyr/commit/ab35eaccec5976f05c196f176d0c32885754496f","tags":["patch","fix"]},{"url":"https://github.com/zephyrproject-rtos/zephyr/security/advisories/GHSA-47q2-w832-7w67","tags":["advisory"]},{"url":"https://github.com/CVEProject/cvelistV5/tree/main/cves/2026/9xxx/CVE-2026-9728.json","tags":["advisory"]},{"url":"https://nvd.nist.gov/vuln/detail/CVE-2026-9728","tags":["advisory"]},{"url":"https://github.com/zephyrproject-rtos/zephyr","tags":["package"]}],"sources":{"cve":[{"id":"CVE-2026-9728","title":"TOCTOU race in mbox_send syscall verifier allows userspace to leak kernel memory","description":"The userspace syscall verifier z_vrfy_mbox_send() in drivers/mbox/mbox_handlers.c validated the nested msg-\u003edata/msg-\u003esize fields by reading them directly out of live userspace memory, and then forwarded the original, still-mutable userspace struct mbox_msg * pointer to z_impl_mbox_send() and the underlying driver. Between the access check and the driver's use of msg-\u003edata, the validated pointer could be replaced, leaving a time-of-check/time-of-use window.\n\nOn a system built with CONFIG_USERSPACE, any unprivileged userspace thread may invoke the mbox_send() system call. A second thread sharing the caller's address space can race to overwrite msg-\u003edata with a supervisor (kernel) address after the verifier's bounds check has passed but before the driver dereferences it. The driver then reads from the attacker-chosen address in supervisor context (for example memcpy(\u0026data32, msg-\u003edata, msg-\u003esize) in the NXP mailbox driver, whose bytes are subsequently emitted to the peer mailbox endpoint).\n\nThe impact is a userspace-to-supervisor access-control bypass: disclosure of kernel memory contents (high confidentiality impact), or, for an invalid/unmapped target address, a faulting kernel read causing denial of service. The fix snapshots the entire struct mbox_msg into a kernel-stack copy with k_usermode_from_copy() and validates and forwards that immutable copy, closing the race.","status":"published","published":"2026-08-24T14:28:46.429Z","modified":"2026-11-04T04:59:49.294Z","cvss":[{"version":"3.1","score":6.4,"vector":"CVSS:3.1/AV:L/AC:H/PR:L/UI:N/S:C/C:H/I:N/A:L","level":"medium","assessor":"cna"}],"cwes":["CWE-367"],"updated_at":"2026-11-04T05:21:48.223566Z"}],"euvd":[{"id":"EUVD-2026-64897","description":"The userspace syscall verifier z_vrfy_mbox_send() in drivers/mbox/mbox_handlers.c validated the nested msg-\u003edata/msg-\u003esize fields by reading them directly out of live userspace memory, and then forwarded the original, still-mutable userspace struct mbox_msg * pointer to z_impl_mbox_send() and the underlying driver. Between the access check and the driver's use of msg-\u003edata, the validated pointer could be replaced, leaving a time-of-check/time-of-use window.\n\nOn a system built with CONFIG_USERSPACE, any unprivileged userspace thread may invoke the mbox_send() system call. A second thread sharing the caller's address space can race to overwrite msg-\u003edata with a supervisor (kernel) address after the verifier's bounds check has passed but before the driver dereferences it. The driver then reads from the attacker-chosen address in supervisor context (for example memcpy(\u0026data32, msg-\u003edata, msg-\u003esize) in the NXP mailbox driver, whose bytes are subsequently emitted to the peer mailbox endpoint).\n\nThe impact is a userspace-to-supervisor access-control bypass: disclosure of kernel memory contents (high confidentiality impact), or, for an invalid/unmapped target address, a faulting kernel read causing denial of service. The fix snapshots the entire struct mbox_msg into a kernel-stack copy with k_usermode_from_copy() and validates and forwards that immutable copy, closing the race.","status":"published","published":"2026-08-24T14:28:46Z","modified":"2026-11-04T04:59:49Z","cvss":[{"version":"3.1","score":6.4,"vector":"CVSS:3.1/AV:L/AC:H/PR:L/UI:N/S:C/C:H/I:N/A:L","level":"medium"}],"updated_at":"2026-11-04T07:00:56.720215Z"}],"nvd":[{"id":"CVE-2026-9728","description":"The userspace syscall verifier z_vrfy_mbox_send() in drivers/mbox/mbox_handlers.c validated the nested msg-\u003edata/msg-\u003esize fields by reading them directly out of live userspace memory, and then forwarded the original, still-mutable userspace struct mbox_msg * pointer to z_impl_mbox_send() and the underlying driver. Between the access check and the driver's use of msg-\u003edata, the validated pointer could be replaced, leaving a time-of-check/time-of-use window.\n\nOn a system built with CONFIG_USERSPACE, any unprivileged userspace thread may invoke the mbox_send() system call. A second thread sharing the caller's address space can race to overwrite msg-\u003edata with a supervisor (kernel) address after the verifier's bounds check has passed but before the driver dereferences it. The driver then reads from the attacker-chosen address in supervisor context (for example memcpy(\u0026data32, msg-\u003edata, msg-\u003esize) in the NXP mailbox driver, whose bytes are subsequently emitted to the peer mailbox endpoint).\n\nThe impact is a userspace-to-supervisor access-control bypass: disclosure of kernel memory contents (high confidentiality impact), or, for an invalid/unmapped target address, a faulting kernel read causing denial of service. The fix snapshots the entire struct mbox_msg into a kernel-stack copy with k_usermode_from_copy() and validates and forwards that immutable copy, closing the race.","status":"published","published":"2026-08-24T15:16:49.01Z","modified":"2026-11-04T05:19:07.053Z","cvss":[{"version":"3.1","score":6.4,"vector":"CVSS:3.1/AV:L/AC:H/PR:L/UI:N/S:C/C:H/I:N/A:L","level":"medium","assessor":"secondary"}],"cwes":["CWE-367"],"updated_at":"2026-11-04T06:53:20.502063Z"}],"osv":[{"id":"CVE-2026-9728","title":"TOCTOU race in mbox_send syscall verifier allows userspace to leak kernel memory","description":"The userspace syscall verifier z_vrfy_mbox_send() in drivers/mbox/mbox_handlers.c validated the nested msg-\u003edata/msg-\u003esize fields by reading them directly out of live userspace memory, and then forwarded the original, still-mutable userspace struct mbox_msg * pointer to z_impl_mbox_send() and the underlying driver. Between the access check and the driver's use of msg-\u003edata, the validated pointer could be replaced, leaving a time-of-check/time-of-use window.\n\nOn a system built with CONFIG_USERSPACE, any unprivileged userspace thread may invoke the mbox_send() system call. A second thread sharing the caller's address space can race to overwrite msg-\u003edata with a supervisor (kernel) address after the verifier's bounds check has passed but before the driver dereferences it. The driver then reads from the attacker-chosen address in supervisor context (for example memcpy(\u0026data32, msg-\u003edata, msg-\u003esize) in the NXP mailbox driver, whose bytes are subsequently emitted to the peer mailbox endpoint).\n\nThe impact is a userspace-to-supervisor access-control bypass: disclosure of kernel memory contents (high confidentiality impact), or, for an invalid/unmapped target address, a faulting kernel read causing denial of service. The fix snapshots the entire struct mbox_msg into a kernel-stack copy with k_usermode_from_copy() and validates and forwards that immutable copy, closing the race.","status":"published","published":"2026-08-24T14:28:46.429Z","modified":"2026-11-05T08:03:58.316905181Z","cvss":[{"version":"3.1","score":6.4,"vector":"CVSS:3.1/AV:L/AC:H/PR:L/UI:N/S:C/C:H/I:N/A:L","level":"medium"}],"cwes":["CWE-367"],"updated_at":"2026-11-05T09:08:14.330889Z"}]}}