{ "schema_version": "1.6.1", "id": "RPI-CVE-2026-80671", "modified": "2026-10-09T11:39:44Z", "upstream": [ "CVE-2026-80671" ], "published": "2026-08-28T06:50:33Z", "summary": "perf sched: Fix register_pid() overflow, strcpy, and BUG_ON", "details": "In the Linux kernel, the following vulnerability has been resolved:\n\nperf sched: Fix register_pid() overflow, strcpy, and BUG_ON\n\nregister_pid() has several issues when processing untrusted perf.data:\n\n1. Integer overflow: (pid + 1) * sizeof(struct task_desc *) can wrap\n to a small value on 32-bit systems when pid is large (e.g.\n 0x40000000), causing realloc to return a tiny buffer followed by\n out-of-bounds writes in the initialization loop.\n\n2. Heap buffer overflow: strcpy(task->comm, comm) copies the\n untrusted comm string into a fixed 20-byte COMM_LEN buffer with\n no length check.\n\n3. BUG_ON on allocation failure: perf.data is untrusted input, so\n allocation failures should be handled gracefully rather than\n killing the process.\n\n4. Realloc of sched->tasks assigned directly back, leaking the old\n pointer on failure; nr_tasks incremented before the realloc,\n leaving corrupted state on failure.\n\nCap pid at PID_MAX_LIMIT (4194304, matching the kernel's maximum\non 64-bit), replace strcpy with strlcpy, guard against NULL comm,\nreplace BUG_ON with NULL returns using safe realloc patterns, and\nadd NULL checks in callers that dereference the result.", "severity": [ { "type": "CVSS_V3", "score": "CVSS:3.1/AV:L/AC:L/PR:N/UI:N/S:C/C:H/I:H/A:H" } ], "affected": [ { "package": { "ecosystem": "Raspberry Pi OS:12", "name": "linux" }, "versions": [ "1:6.1.47-1+rpt4" ], "ecosystem_specific": { "binaries": [ { "binary_name": "linux-image-6.1.0-rpi3-rpi-2712", "binary_version": "1:6.1.47-1+rpt4" }, { "binary_name": "linux-image-6.1.0-rpi3-rpi-v6", "binary_version": "1:6.1.47-1+rpt4" }, { "binary_name": "linux-image-6.1.0-rpi3-rpi-v7", "binary_version": "1:6.1.47-1+rpt4" }, { "binary_name": "linux-image-6.1.0-rpi3-rpi-v7l", "binary_version": "1:6.1.47-1+rpt4" }, { "binary_name": "linux-image-6.1.0-rpi3-rpi-v8", "binary_version": "1:6.1.47-1+rpt4" } ] }, "database_specific": { "per_version": { "1:6.1.47-1+rpt4": { "linux_commit": "655fc658a15ae7a6f37103754adb39ba52a9a14e", "upstream_version": "6.1.47", "assessment": "commit ancestry (strak rule)", "upstream_fixed": "6.1.188", "compiled_filter": false } } } } ], "references": [ { "type": "ADVISORY", "url": "https://www.cve.org/CVERecord?id=CVE-2026-80671" }, { "type": "FIX", "url": "https://git.kernel.org/stable/c/344e59fc4f2cba6dc04ce5e2a77388645d4ca834" }, { "type": "FIX", "url": "https://git.kernel.org/stable/c/22263d241bad753217d545f8085b7ead52e2873b" }, { "type": "FIX", "url": "https://git.kernel.org/stable/c/2003918e13f43cf6c5698be3cc9aa3657dd676a8" }, { "type": "FIX", "url": "https://git.kernel.org/stable/c/29639f00bfa49da7904ba0a499e0da7a0e5d735e" }, { "type": "FIX", "url": "https://git.kernel.org/stable/c/261d2b5f4afa0570ca906e553c93811c5abe4590" }, { "type": "FIX", "url": "https://git.kernel.org/stable/c/652cea73b7b7b7c622a2be670e44e3c499c6d49f" }, { "type": "FIX", "url": "https://git.kernel.org/stable/c/5ea1dcc9418c4e06ce29ed5170596f497ba86872" }, { "type": "FIX", "url": "https://git.kernel.org/stable/c/5949d339f5ec98752d56dcd4e36f619a59d513a5" } ], "database_specific": { "source": "https://git.kernel.org/pub/scm/linux/security/vulns.git", "upstream_modified": "2026-09-14T12:06:03Z", "severity": { "source": "kernel CNA", "cvss_v3_score": 9.3, "cvss_v3_severity": "Critical", "scenario": "AV:L - The flaw is in userspace perf's builtin-sched.c reached only when a local user runs perf sched replay or perf sched latency against a crafted perf.data file via perf_session__process_events; it is not triggered by remote network packets, kernel syscalls, or in-kernel sched code.\nAC:L - The attacker fully controls untrusted perf.data sched_switch/sched_wakeup fields (prev_pid, next_pid, prev_comm, next_comm); large PIDs and overlong comm strings deterministically hit the integer-overflow realloc path or strcpy heap overflow without races or victim-specific memory layout.\nPR:N - In the malicious-trace threat model an external party needs no account or privileges on the victim host\u2014they only supply a crafted perf.data file for a developer, support engineer, or CI job to analyze with perf sched replay/latency.\nUI:N - Once a trace is submitted to triage or CI, perf sched replay/latency parses perf.data non-interactively through perf_sched__read_events without further per-event operator actions beyond automated invocation of the perf command on the supplied file.\nS:C - Malicious sched metadata embedded in an externally supplied perf.data crosses from the untrusted reporter/artifact domain into the analyst or CI host security authority when perf parses that file, matching the kernel CNA untrusted-artifact model used for crafted filesystem images.\nC:H - Integer-overflow-driven out-of-bounds writes into the pid_to_task table and strcpy heap overflow past task_desc.comm[20] corrupt adjacent heap objects and can disclose process memory beyond a few bounded bytes.\nI:H - Attacker-controlled overflows corrupt task_desc function pointers and heap metadata (including the atoms pointer), providing controllable userspace memory corruption suitable for arbitrary write and code execution in the perf process.\nA:H - Pre-fix BUG_ON paths abort the perf process on allocation failure, and the heap out-of-bounds writes reliably crash or hang perf during register_pid() while parsing malicious perf.data." } } }