{ "schema_version": "1.6.1", "id": "RPI-CVE-2023-52761", "modified": "2026-10-09T11:39:44Z", "upstream": [ "CVE-2023-52761" ], "published": "2024-05-21T15:34:20Z", "summary": "riscv: VMAP_STACK overflow detection thread-safe", "details": "In the Linux kernel, the following vulnerability has been resolved:\n\nriscv: VMAP_STACK overflow detection thread-safe\n\ncommit 31da94c25aea (\"riscv: add VMAP_STACK overflow detection\") added\nsupport for CONFIG_VMAP_STACK. If overflow is detected, CPU switches to\n`shadow_stack` temporarily before switching finally to per-cpu\n`overflow_stack`.\n\nIf two CPUs/harts are racing and end up in over flowing kernel stack, one\nor both will end up corrupting each other state because `shadow_stack` is\nnot per-cpu. This patch optimizes per-cpu overflow stack switch by\ndirectly picking per-cpu `overflow_stack` and gets rid of `shadow_stack`.\n\nFollowing are the changes in this patch\n\n - Defines an asm macro to obtain per-cpu symbols in destination\n register.\n - In entry.S, when overflow is detected, per-cpu overflow stack is\n located using per-cpu asm macro. Computing per-cpu symbol requires\n a temporary register. x31 is saved away into CSR_SCRATCH\n (CSR_SCRATCH is anyways zero since we're in kernel).\n\nPlease see Links for additional relevant disccussion and alternative\nsolution.\n\nTested by `echo EXHAUST_STACK > /sys/kernel/debug/provoke-crash/DIRECT`\nKernel crash log below\n\n Insufficient stack space to handle exception!/debug/provoke-crash/DIRECT\n Task stack: [0xff20000010a98000..0xff20000010a9c000]\n Overflow stack: [0xff600001f7d98370..0xff600001f7d99370]\n CPU: 1 PID: 205 Comm: bash Not tainted 6.1.0-rc2-00001-g328a1f96f7b9 #34\n Hardware name: riscv-virtio,qemu (DT)\n epc : __memset+0x60/0xfc\n ra : recursive_loop+0x48/0xc6 [lkdtm]\n epc : ffffffff808de0e4 ra : ffffffff0163a752 sp : ff20000010a97e80\n gp : ffffffff815c0330 tp : ff600000820ea280 t0 : ff20000010a97e88\n t1 : 000000000000002e t2 : 3233206874706564 s0 : ff20000010a982b0\n s1 : 0000000000000012 a0 : ff20000010a97e88 a1 : 0000000000000000\n a2 : 0000000000000400 a3 : ff20000010a98288 a4 : 0000000000000000\n a5 : 0000000000000000 a6 : fffffffffffe43f0 a7 : 00007fffffffffff\n s2 : ff20000010a97e88 s3 : ffffffff01644680 s4 : ff20000010a9be90\n s5 : ff600000842ba6c0 s6 : 00aaaaaac29e42b0 s7 : 00fffffff0aa3684\n s8 : 00aaaaaac2978040 s9 : 0000000000000065 s10: 00ffffff8a7cad10\n s11: 00ffffff8a76a4e0 t3 : ffffffff815dbaf4 t4 : ffffffff815dbaf4\n t5 : ffffffff815dbab8 t6 : ff20000010a9bb48\n status: 0000000200000120 badaddr: ff20000010a97e88 cause: 000000000000000f\n Kernel panic - not syncing: Kernel stack overflow\n CPU: 1 PID: 205 Comm: bash Not tainted 6.1.0-rc2-00001-g328a1f96f7b9 #34\n Hardware name: riscv-virtio,qemu (DT)\n Call Trace:\n [] dump_backtrace+0x30/0x38\n [] show_stack+0x40/0x4c\n [] dump_stack_lvl+0x44/0x5c\n [] dump_stack+0x18/0x20\n [] panic+0x126/0x2fe\n [] walk_stackframe+0x0/0xf0\n [] recursive_loop+0x48/0xc6 [lkdtm]\n SMP: stopping secondary CPUs\n ---[ end Kernel panic - not syncing: Kernel stack overflow ]---", "severity": [ { "type": "CVSS_V3", "score": "CVSS:3.1/AV:L/AC:L/PR:L/UI:N/S:U/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": null, "compiled_filter": false } } } } ], "references": [ { "type": "ADVISORY", "url": "https://www.cve.org/CVERecord?id=CVE-2023-52761" }, { "type": "FIX", "url": "https://git.kernel.org/stable/c/1493baaf09e3c1899959c8a107cd1207e16d1788" }, { "type": "FIX", "url": "https://git.kernel.org/stable/c/eff53aea3855f71992c043cebb1c00988c17ee20" }, { "type": "FIX", "url": "https://git.kernel.org/stable/c/be97d0db5f44c0674480cb79ac6f5b0529b84c76" } ], "database_specific": { "source": "https://git.kernel.org/pub/scm/linux/security/vulns.git", "upstream_modified": "2026-08-05T11:24:39Z", "severity": { "source": "kernel CNA", "cvss_v3_score": 7.8, "cvss_v3_severity": "High", "scenario": "AV:L - The vulnerable path is arch/riscv kernel exception entry (handle_kernel_stack_overflow in entry.S), reached only after a local task overflows its kernel stack\u2014not via network protocols or adjacent-link frames.\nAC:L - An attacker who can drive deep kernel recursion can pin threads to multiple harts and trigger both sides of the shared shadow_stack race, so the race is attacker-controlled rather than dependent on external timing.\nPR:L - Kernel stack overflow can be induced through unprivileged syscall/kernel paths that recurse deeply; the overflow handler itself has no capability check, so a normal local user is sufficient.\nUI:N - No victim action is required; the attacker triggers the overflows directly from their own processes.\nS:U - Impact stays within the same kernel privilege domain; this is not a VM escape, IOMMU bypass, or other cross-boundary break.\nC:H - Concurrent corruption of the shared shadow_stack while another hart executes C code on it can overwrite saved registers and stack frames, yielding a kernel read primitive consistent with memory-corruption impact.\nI:H - The same stack-frame/register corruption can overwrite return addresses and control state in the overflow path, enabling control-flow hijacking in kernel context.\nA:H - The overflow path panics with \"Kernel stack overflow\"; the race also causes unpredictable corruption that can oops or hang the system." } } }