arena_vm_fault() allocated the page while holding arena->spinlock, so it could only use the non-blocking allocator. Once the memcg is at memory.max that allocation just fails, the fault turns into VM_FAULT_SIGSEGV, and the process gets a SIGSEGV on a perfectly valid arena address. Hitting memory.max is routine (e.g. page cache from reading a big file), so this kills innocent processes.
Rework the fault handler: - Preallocate the page before taking the lock, like do_anonymous_page() does, so it can sleep, reclaim and go through the OOM path, and return VM_FAULT_OOM on failure so the memcg OOM handler runs instead of a fake segfault. - A lockless probe skips that preallocation when a page is already mapped (e.g. allocated by the bpf program), so the common case wastes no allocation. The rare race where such a page is freed before we take the lock falls back to the non-blocking allocator under the lock. - Return VM_FAULT_SIGBUS for the non-recoverable errors (lock failure, range-tree and page-table failures) instead of VM_FAULT_SIGSEGV; only BPF_F_SEGV_ON_FAULT, and a scratch-page hole under that flag, is a real user addressing error and keeps VM_FAULT_SIGSEGV. - Tidy up the error labels. Signed-off-by: Jiayuan Chen <[email protected]> --- kernel/bpf/arena.c | 92 +++++++++++++++++++++++++++++++++++----------- 1 file changed, 71 insertions(+), 21 deletions(-) diff --git a/kernel/bpf/arena.c b/kernel/bpf/arena.c index 555ee2531ef9..09a718ca4c8b 100644 --- a/kernel/bpf/arena.c +++ b/kernel/bpf/arena.c @@ -481,7 +481,8 @@ static vm_fault_t arena_vm_fault(struct vm_fault *vmf) struct bpf_map *map = vmf->vma->vm_file->private_data; struct bpf_arena *arena = container_of(map, struct bpf_arena, map); struct mem_cgroup *new_memcg, *old_memcg; - struct page *page; + struct page *page, *new_page = NULL; + vm_fault_t fault_ret; long kbase, kaddr; unsigned long flags; int ret; @@ -489,59 +490,108 @@ static vm_fault_t arena_vm_fault(struct vm_fault *vmf) kbase = bpf_arena_get_kern_vm_start(arena); kaddr = kbase + (u32)(vmf->address); - if (raw_res_spin_lock_irqsave(&arena->spinlock, flags)) + page = vmalloc_to_page((void *)kaddr); + if (!page && !(arena->map.map_flags & BPF_F_SEGV_ON_FAULT)) { + /* + * We run in process context here, so preallocate the page + * outside the lock with an explicitly sleepable allocator. It + * can then go through reclaim (both memcg and global) and the + * OOM path, the way do_anonymous_page() does; under + * arena->spinlock only the non-blocking allocator is available, + * which never reclaims. That also decides the return value: + * VM_FAULT_OOM below is only meaningful if the OOM machinery was + * actually engaged, which the non-blocking allocator never does. + */ + bpf_map_memcg_enter(&arena->map, &old_memcg, &new_memcg); + new_page = bpf_map_alloc_page_sleepable(map); + bpf_map_memcg_exit(old_memcg, new_memcg); + if (!new_page) + return VM_FAULT_OOM; + } + + if (raw_res_spin_lock_irqsave(&arena->spinlock, flags)) { /* * A failed lock means a possible deadlock was detected. Don't * return VM_FAULT_RETRY: this handler never took mmap_lock, but * the fault path would re-take it on retry and deadlock. Fail. */ + if (new_page) + free_pages_nolock(new_page, 0); return VM_FAULT_SIGBUS; + } page = vmalloc_to_page((void *)kaddr); if (page) { - if (page == arena->scratch_page) - /* BPF triggered scratch here; don't lazy-alloc over it */ - goto out_sigsegv; + if (page == arena->scratch_page) { + /* + * A scratch page marks a hole. Segfault only if the user + * asked for it; otherwise we could lazy-allocate but + * choose not to over a hole, so report a bus error. + */ + fault_ret = (arena->map.map_flags & BPF_F_SEGV_ON_FAULT) ? + VM_FAULT_SIGSEGV : VM_FAULT_SIGBUS; + goto out_err_locked; + } /* already have a page vmap-ed */ goto out; } + if (arena->map.map_flags & BPF_F_SEGV_ON_FAULT) { + /* User space requested to segfault when page is not allocated by bpf prog */ + fault_ret = VM_FAULT_SIGSEGV; + goto out_err_locked; + } + bpf_map_memcg_enter(&arena->map, &old_memcg, &new_memcg); - if (arena->map.map_flags & BPF_F_SEGV_ON_FAULT) - /* User space requested to segfault when page is not allocated by bpf prog */ - goto out_sigsegv_memcg; + if (!new_page) { + /* + * Very rare race: the bpf program had allocated a page here, so + * the lockless probe saw it and we skipped preallocation, but it + * freed the page before we took the lock. Now we do need one; + * sleeping is not allowed here, so fall back to the non-blocking + * allocator and give up if it fails. + */ + ret = bpf_map_alloc_pages(map, map->numa_node, 1, &new_page); + if (ret) { + fault_ret = VM_FAULT_SIGBUS; + goto out_err_locked_memcg; + } + } ret = range_tree_clear(&arena->rt, vmf->pgoff, 1); - if (ret) - goto out_sigsegv_memcg; - - struct apply_range_data data = { .arena = arena, .pages = &page, .i = 0 }; - /* Account into memcg of the process that created bpf_arena */ - ret = bpf_map_alloc_pages(map, NUMA_NO_NODE, 1, &page); if (ret) { - range_tree_set(&arena->rt, vmf->pgoff, 1); - goto out_sigsegv_memcg; + fault_ret = VM_FAULT_SIGBUS; + goto out_err_locked_memcg; } + struct apply_range_data data = { .arena = arena, .pages = &new_page, .i = 0 }; ret = apply_to_page_range(&init_mm, kaddr, PAGE_SIZE, apply_range_set_cb, &data); if (ret) { range_tree_set(&arena->rt, vmf->pgoff, 1); - free_pages_nolock(page, 0); - goto out_sigsegv_memcg; + fault_ret = VM_FAULT_SIGBUS; + goto out_err_locked_memcg; } flush_vmap_cache(kaddr, PAGE_SIZE); bpf_map_memcg_exit(old_memcg, new_memcg); + /* new_page was consumed */ + page = new_page; + new_page = NULL; out: page_ref_add(page, 1); raw_res_spin_unlock_irqrestore(&arena->spinlock, flags); + if (new_page) + free_pages_nolock(new_page, 0); vmf->page = page; return 0; -out_sigsegv_memcg: + +out_err_locked_memcg: bpf_map_memcg_exit(old_memcg, new_memcg); -out_sigsegv: +out_err_locked: raw_res_spin_unlock_irqrestore(&arena->spinlock, flags); - return VM_FAULT_SIGSEGV; + if (new_page) + free_pages_nolock(new_page, 0); + return fault_ret; } static const struct vm_operations_struct arena_vm_ops = { -- 2.43.0
