{ "dataType": "CVE_RECORD", "dataVersion": "5.2", "cveMetadata": { "cveId": "CVE-2025-22069", "assignerOrgId": "416baaa9-dc9f-4396-8d5f-8c081fb06d67", "state": "PUBLISHED", "assignerShortName": "Linux", "dateReserved": "2024-12-29T08:45:45.814Z", "datePublished": "2025-04-16T14:12:22.357Z", "dateUpdated": "2026-08-05T11:56:30.349Z" }, "containers": { "cna": { "providerMetadata": { "orgId": "416baaa9-dc9f-4396-8d5f-8c081fb06d67", "shortName": "Linux", "dateUpdated": "2026-08-05T11:56:30.349Z" }, "descriptions": [ { "lang": "en", "value": "In the Linux kernel, the following vulnerability has been resolved:\n\nriscv: fgraph: Fix stack layout to match __arch_ftrace_regs argument of ftrace_return_to_handler\n\nNaresh Kamboju reported a \"Bad frame pointer\" kernel warning while\nrunning LTP trace ftrace_stress_test.sh in riscv. We can reproduce the\nsame issue with the following command:\n\n```\n$ cd /sys/kernel/debug/tracing\n$ echo 'f:myprobe do_nanosleep%return args1=$retval' > dynamic_events\n$ echo 1 > events/fprobes/enable\n$ echo 1 > tracing_on\n$ sleep 1\n```\n\nAnd we can get the following kernel warning:\n\n[ 127.692888] ------------[ cut here ]------------\n[ 127.693755] Bad frame pointer: expected ff2000000065be50, received ba34c141e9594000\n[ 127.693755] from func do_nanosleep return to ffffffff800ccb16\n[ 127.698699] WARNING: CPU: 1 PID: 129 at kernel/trace/fgraph.c:755 ftrace_return_to_handler+0x1b2/0x1be\n[ 127.699894] Modules linked in:\n[ 127.700908] CPU: 1 UID: 0 PID: 129 Comm: sleep Not tainted 6.14.0-rc3-g0ab191c74642 #32\n[ 127.701453] Hardware name: riscv-virtio,qemu (DT)\n[ 127.701859] epc : ftrace_return_to_handler+0x1b2/0x1be\n[ 127.702032] ra : ftrace_return_to_handler+0x1b2/0x1be\n[ 127.702151] epc : ffffffff8013b5e0 ra : ffffffff8013b5e0 sp : ff2000000065bd10\n[ 127.702221] gp : ffffffff819c12f8 tp : ff60000080853100 t0 : 6e00000000000000\n[ 127.702284] t1 : 0000000000000020 t2 : 6e7566206d6f7266 s0 : ff2000000065bd80\n[ 127.702346] s1 : ff60000081262000 a0 : 000000000000007b a1 : ffffffff81894f20\n[ 127.702408] a2 : 0000000000000010 a3 : fffffffffffffffe a4 : 0000000000000000\n[ 127.702470] a5 : 0000000000000000 a6 : 0000000000000008 a7 : 0000000000000038\n[ 127.702530] s2 : ba34c141e9594000 s3 : 0000000000000000 s4 : ff2000000065bdd0\n[ 127.702591] s5 : 00007fff8adcf400 s6 : 000055556dc1d8c0 s7 : 0000000000000068\n[ 127.702651] s8 : 00007fff8adf5d10 s9 : 000000000000006d s10: 0000000000000001\n[ 127.702710] s11: 00005555737377c8 t3 : ffffffff819d899e t4 : ffffffff819d899e\n[ 127.702769] t5 : ffffffff819d89a0 t6 : ff2000000065bb18\n[ 127.702826] status: 0000000200000120 badaddr: 0000000000000000 cause: 0000000000000003\n[ 127.703292] [] ftrace_return_to_handler+0x1b2/0x1be\n[ 127.703760] [] return_to_handler+0x16/0x26\n[ 127.704009] [] return_to_handler+0x0/0x26\n[ 127.704057] [] common_nsleep+0x42/0x54\n[ 127.704117] [] __riscv_sys_clock_nanosleep+0xba/0x10a\n[ 127.704176] [] do_trap_ecall_u+0x188/0x218\n[ 127.704295] [] handle_exception+0x14a/0x156\n[ 127.705436] ---[ end trace 0000000000000000 ]---\n\nThe reason is that the stack layout for constructing argument for the\nftrace_return_to_handler in the return_to_handler does not match the\n__arch_ftrace_regs structure of riscv, leading to unexpected results." } ], "metrics": [ { "cvssV3_1": { "version": "3.1", "vectorString": "CVSS:3.1/AV:L/AC:L/PR:L/UI:N/S:U/C:H/I:H/A:H", "baseScore": 7.8, "baseSeverity": "HIGH" }, "scenarios": [ { "lang": "en", "value": "AV:L - The vulnerable trampoline is armed only by writing to tracefs (`current_tracer`, `dynamic_events`, `events/*/enable`) and fires on ordinary syscall execution, both of which require local access to the machine. There is no network-facing consumer of `arch/riscv/kernel/mcount.S`.\nAC:L - The mismatch is a fixed compile-time stack layout error, not a race or a memory-layout gamble: the very first return of any traced function reads `s0` from `sp+24` instead of `sp+64` and writes `a0`/`a1` past the 80-byte frame, so the commit's four-command reproducer fails deterministically on the first `sleep 1`.\nPR:L - Reaching the trampoline requires only write access to tracefs, which is enforced by file permissions rather than any `capable()` check and is routinely delegated to an unprivileged tracing group via tracefs's `gid=` mount option (as Android and ChromeOS do for `atrace`). A non-root member of that group can arm `function_graph` or an fprobe and trigger the corruption.\nUI:N - The same local user both arms the tracer and triggers it by making any traced function return in their own process; no action by an administrator or other victim is needed.\nS:U - The out-of-bounds writes and reads all land in the kernel stack of the running task, within the kernel's own security authority — no VM, IOMMU, or sandbox boundary is crossed.\nC:H - The C side treats `sp..sp+103` as `__arch_ftrace_regs` while only 80 bytes were allocated, so `a5`/`a6`/`a7` and the 64-byte `ftrace_partial_regs()` `memcpy` read 24 bytes past the frame into the caller's live stack, and `s0`/`t1`/`a0`–`a4` are read from uninitialized stack — these stale kernel stack values (pointers, saved registers) are handed to fprobe/BPF callbacks as `$retval`/`$argN` and dumped into the trace buffer and the WARN register dump.\nI:H - `REG_S a0, 80(sp)` / `REG_S a1, 88(sp)` after `addi sp, sp, -80` is an unconditional 16-byte out-of-bounds stack write into the live frame of the traced function's caller, with `a0` holding an often user-influenceable return value; the same overhang lets a retfunc callback's writes to `args[5..7]` corrupt the caller's frame and even substitute the value `RESTORE_RET_ABI_STATE` reloads into `a0`, giving control over saved state and potentially return addresses further up the stack.\nA:H - With `HAVE_FUNCTION_GRAPH_FP_TEST` — effectively always enabled on riscv given `ARCH_WANT_FRAME_POINTERS` and `select FRAME_POINTER if PERF_EVENTS` — the garbage frame pointer makes `ftrace_pop_return_trace()` set `*ret = (unsigned long)panic`, so `return_to_handler`'s `jalr a2` jumps directly into `panic()`, in addition to the WARN and `ftrace_graph_stop()` seen in the reported LTP failure." } ] } ], "affected": [ { "product": "Linux", "vendor": "Linux", "defaultStatus": "unaffected", "repo": "https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git", "programFiles": [ "arch/riscv/kernel/mcount.S" ], "versions": [ { "version": "33d4e904e24d14ff0fbc528b657ddc7c7b636e6a", "lessThan": "7ed384db061a264bd806898f7ccab9b98b591488", "status": "affected", "versionType": "git" }, { "version": "a3ed4157b7d89800a0008de0c9e46a438a5c3745", "lessThan": "78b39c587b8f6c69140177108f9c08a75b1c7c37", "status": "affected", "versionType": "git" }, { "version": "a3ed4157b7d89800a0008de0c9e46a438a5c3745", "lessThan": "67a5ba8f742f247bc83e46dd2313c142b1383276", "status": "affected", "versionType": "git" }, { "version": "6.12.75", "lessThan": "6.12.92", "status": "affected", "versionType": "semver" } ] }, { "product": "Linux", "vendor": "Linux", "defaultStatus": "affected", "repo": "https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git", "programFiles": [ "arch/riscv/kernel/mcount.S" ], "versions": [ { "version": "6.14", "status": "affected" }, { "version": "0", "lessThan": "6.14", "status": "unaffected", "versionType": "semver" }, { "version": "6.12.92", "lessThanOrEqual": "6.12.*", "status": "unaffected", "versionType": "semver" }, { "version": "6.14.2", "lessThanOrEqual": "6.14.*", "status": "unaffected", "versionType": "semver" }, { "version": "6.15", "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": "6.12.75", "versionEndExcluding": "6.12.92" }, { "vulnerable": true, "criteria": "cpe:2.3:o:linux:linux_kernel:*:*:*:*:*:*:*:*", "versionStartIncluding": "6.14", "versionEndExcluding": "6.14.2" }, { "vulnerable": true, "criteria": "cpe:2.3:o:linux:linux_kernel:*:*:*:*:*:*:*:*", "versionStartIncluding": "6.14", "versionEndExcluding": "6.15" } ] } ] } ], "references": [ { "url": "https://git.kernel.org/stable/c/7ed384db061a264bd806898f7ccab9b98b591488" }, { "url": "https://git.kernel.org/stable/c/78b39c587b8f6c69140177108f9c08a75b1c7c37" }, { "url": "https://git.kernel.org/stable/c/67a5ba8f742f247bc83e46dd2313c142b1383276" } ], "title": "riscv: fgraph: Fix stack layout to match __arch_ftrace_regs argument of ftrace_return_to_handler", "x_generator": { "engine": "bippy-1.2.0" } } } }