mlock_fixup() adjusts mm->locked_vm, an unsigned long, by the number of pages whose VM_LOCKED state it changes, but computes that number in an int: int nr_pages; ... nr_pages = (end - start) >> PAGE_SHIFT; ... mm->locked_vm += nr_pages; When a single VMA range of 2^31 pages (8 TiB with 4 KiB pages) or more is locked or unlocked, the count is truncated to 32 bits and then sign-extended, so locked_vm is changed by the wrong amount, while munmap(), mremap() and the other paths that maintain locked_vm use the full unsigned long page count. Such a range is reachable on 64-bit architectures with a 47-bit or larger user address space (x86_64, arm64 with 48-bit VA, riscv64 Sv48) by a task with CAP_IPC_LOCK or a raised RLIMIT_MEMLOCK, which do_mlock() checks against the full untruncated length. It costs nothing with MAP_NORESERVE and MLOCK_ONFAULT/MCL_ONFAULT, since then no page is populated. On x86_64, after p = mmap(NULL, 8UL << 40, PROT_READ | PROT_WRITE, MAP_PRIVATE | MAP_ANONYMOUS | MAP_NORESERVE, -1, 0); mlock2(p, 8UL << 40, MLOCK_ONFAULT); VmLck in /proc/self/status reads 18446744065119617024 kB instead of 8589934592 kB; munmap() of the region then subtracts the true size and leaves VmLck at 18446744056529682432 kB instead of 0. With 16 TiB, locked_vm is not changed at all, VmLck stays 0 kB, and after the task drops CAP_IPC_LOCK or lowers RLIMIT_MEMLOCK, mlock(), mmap(MAP_LOCKED) and the other mlock_future_ok() and account_locked_vm() users enforce RLIMIT_MEMLOCK against that count, so they let the task lock memory beyond its limit (or, in the 8 TiB case, refuse every further lock). The error persists for the life of the mm; unlocking the range in pieces skews it further. The wrong count is per-mm accounting only: it does not change which pages are locked. The int has been there since before git history (the variable was named "pages" until it was renamed in 2.6.28); every other updater of locked_vm uses an unsigned long or long count. Make nr_pages a long, which holds the page count of any user range and its negation, so the existing sign handling is unchanged. Found by formal verification of mm/mlock.c against the kernel binary. Fixes: 1da177e4c3f4 ("Linux-2.6.12-rc2") Assisted-by: LLM Signed-off-by: Nickolai Zeldovich --- Notes: Reproducer: the program below, run as root from an initramfs on v7.3-rc7-28-g4056b3c9d3c9 defconfig (LLVM=1) under QEMU, x86_64 (TCG, -cpu max) and riscv64 (virt, -cpu max); a longer version also checks munmap(), a 16 TiB lock and RLIMIT_MEMLOCK enforcement. before after VmLck after locking 8 TiB 18446744065119617024 kB 8589934592 kB VmLck after munmap() of it 18446744056529682432 kB 0 kB VmLck after locking 16 TiB 0 kB 17179869184 kB mlock() of 16 pages over a 16-page succeeds ENOMEM RLIMIT_MEMLOCK, CAP_IPC_LOCK dropped // Needs root (or CAP_IPC_LOCK), >= 16 TiB user VA, vm.overcommit_memory != 2. #define _GNU_SOURCE #include #include #include #include #include int main(void) { size_t len = 8UL << 40; /* 2^31 pages of 4 KiB */ char line[128]; FILE *f; void *p; p = mmap(NULL, len, PROT_READ | PROT_WRITE, MAP_PRIVATE | MAP_ANONYMOUS | MAP_NORESERVE, -1, 0); if (p == MAP_FAILED || syscall(SYS_mlock2, p, len, 1 /* MLOCK_ONFAULT */)) { perror("mmap/mlock2"); return 2; } f = fopen("/proc/self/status", "r"); while (fgets(line, sizeof(line), f)) if (!strncmp(line, "VmLck:", 6)) printf("%s", line); /* expected 8589934592 kB */ return 0; } mm/mlock.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/mm/mlock.c b/mm/mlock.c index 39215a3eab1f..4fb43cddafa8 100644 --- a/mm/mlock.c +++ b/mm/mlock.c @@ -470,7 +470,7 @@ static int mlock_fixup(struct vma_iterator *vmi, struct vm_area_struct *vma, { const vma_flags_t old_vma_flags = vma->flags; struct mm_struct *mm = vma->vm_mm; - int nr_pages; + long nr_pages; int ret = 0; if (vma_flags_same_pair(&old_vma_flags, new_vma_flags) || base-commit: 4056b3c9d3c92ea45955641dbafeb9b960e146ef -- 2.53.0