diff options
Diffstat (limited to 'fs/userfaultfd.c')
| -rw-r--r-- | fs/userfaultfd.c | 46 |
1 files changed, 27 insertions, 19 deletions
diff --git a/fs/userfaultfd.c b/fs/userfaultfd.c index 3b30301c90ec..ccbdbd62f0d8 100644 --- a/fs/userfaultfd.c +++ b/fs/userfaultfd.c @@ -1,3 +1,4 @@ +// SPDX-License-Identifier: GPL-2.0-only /* * fs/userfaultfd.c * @@ -5,9 +6,6 @@ * Copyright (C) 2008-2009 Red Hat, Inc. * Copyright (C) 2015 Red Hat, Inc. * - * This work is licensed under the terms of the GNU GPL, version 2. See - * the COPYING file in the top-level directory. - * * Some part derived from fs/eventfd.c (anon inode setup) and * mm/ksm.c (mm hashing). */ @@ -42,6 +40,16 @@ enum userfaultfd_state { /* * Start with fault_pending_wqh and fault_wqh so they're more likely * to be in the same cacheline. + * + * Locking order: + * fd_wqh.lock + * fault_pending_wqh.lock + * fault_wqh.lock + * event_wqh.lock + * + * To avoid deadlocks, IRQs must be disabled when taking any of the above locks, + * since fd_wqh.lock is taken by aio_poll() while it's holding a lock that's + * also taken in IRQ context. */ struct userfaultfd_ctx { /* waitqueue head for the pending (i.e. not read) userfaults */ @@ -460,7 +468,7 @@ vm_fault_t handle_userfault(struct vm_fault *vmf, unsigned long reason) blocking_state = return_to_userland ? TASK_INTERRUPTIBLE : TASK_KILLABLE; - spin_lock(&ctx->fault_pending_wqh.lock); + spin_lock_irq(&ctx->fault_pending_wqh.lock); /* * After the __add_wait_queue the uwq is visible to userland * through poll/read(). @@ -472,7 +480,7 @@ vm_fault_t handle_userfault(struct vm_fault *vmf, unsigned long reason) * __add_wait_queue. */ set_current_state(blocking_state); - spin_unlock(&ctx->fault_pending_wqh.lock); + spin_unlock_irq(&ctx->fault_pending_wqh.lock); if (!is_vm_hugetlb_page(vmf->vma)) must_wait = userfaultfd_must_wait(ctx, vmf->address, vmf->flags, @@ -554,13 +562,13 @@ vm_fault_t handle_userfault(struct vm_fault *vmf, unsigned long reason) * kernel stack can be released after the list_del_init. */ if (!list_empty_careful(&uwq.wq.entry)) { - spin_lock(&ctx->fault_pending_wqh.lock); + spin_lock_irq(&ctx->fault_pending_wqh.lock); /* * No need of list_del_init(), the uwq on the stack * will be freed shortly anyway. */ list_del(&uwq.wq.entry); - spin_unlock(&ctx->fault_pending_wqh.lock); + spin_unlock_irq(&ctx->fault_pending_wqh.lock); } /* @@ -585,7 +593,7 @@ static void userfaultfd_event_wait_completion(struct userfaultfd_ctx *ctx, init_waitqueue_entry(&ewq->wq, current); release_new_ctx = NULL; - spin_lock(&ctx->event_wqh.lock); + spin_lock_irq(&ctx->event_wqh.lock); /* * After the __add_wait_queue the uwq is visible to userland * through poll/read(). @@ -615,15 +623,15 @@ static void userfaultfd_event_wait_completion(struct userfaultfd_ctx *ctx, break; } - spin_unlock(&ctx->event_wqh.lock); + spin_unlock_irq(&ctx->event_wqh.lock); wake_up_poll(&ctx->fd_wqh, EPOLLIN); schedule(); - spin_lock(&ctx->event_wqh.lock); + spin_lock_irq(&ctx->event_wqh.lock); } __set_current_state(TASK_RUNNING); - spin_unlock(&ctx->event_wqh.lock); + spin_unlock_irq(&ctx->event_wqh.lock); if (release_new_ctx) { struct vm_area_struct *vma; @@ -920,10 +928,10 @@ wakeup: * the last page faults that may have been already waiting on * the fault_*wqh. */ - spin_lock(&ctx->fault_pending_wqh.lock); + spin_lock_irq(&ctx->fault_pending_wqh.lock); __wake_up_locked_key(&ctx->fault_pending_wqh, TASK_NORMAL, &range); __wake_up(&ctx->fault_wqh, TASK_NORMAL, 1, &range); - spin_unlock(&ctx->fault_pending_wqh.lock); + spin_unlock_irq(&ctx->fault_pending_wqh.lock); /* Flush pending events that may still wait on event_wqh */ wake_up_all(&ctx->event_wqh); @@ -1136,7 +1144,7 @@ static ssize_t userfaultfd_ctx_read(struct userfaultfd_ctx *ctx, int no_wait, if (!ret && msg->event == UFFD_EVENT_FORK) { ret = resolve_userfault_fork(ctx, fork_nctx, msg); - spin_lock(&ctx->event_wqh.lock); + spin_lock_irq(&ctx->event_wqh.lock); if (!list_empty(&fork_event)) { /* * The fork thread didn't abort, so we can @@ -1182,7 +1190,7 @@ static ssize_t userfaultfd_ctx_read(struct userfaultfd_ctx *ctx, int no_wait, if (ret) userfaultfd_ctx_put(fork_nctx); } - spin_unlock(&ctx->event_wqh.lock); + spin_unlock_irq(&ctx->event_wqh.lock); } return ret; @@ -1221,14 +1229,14 @@ static ssize_t userfaultfd_read(struct file *file, char __user *buf, static void __wake_userfault(struct userfaultfd_ctx *ctx, struct userfaultfd_wake_range *range) { - spin_lock(&ctx->fault_pending_wqh.lock); + spin_lock_irq(&ctx->fault_pending_wqh.lock); /* wake all in the range and autoremove */ if (waitqueue_active(&ctx->fault_pending_wqh)) __wake_up_locked_key(&ctx->fault_pending_wqh, TASK_NORMAL, range); if (waitqueue_active(&ctx->fault_wqh)) __wake_up(&ctx->fault_wqh, TASK_NORMAL, 1, range); - spin_unlock(&ctx->fault_pending_wqh.lock); + spin_unlock_irq(&ctx->fault_pending_wqh.lock); } static __always_inline void wake_userfault(struct userfaultfd_ctx *ctx, @@ -1883,7 +1891,7 @@ static void userfaultfd_show_fdinfo(struct seq_file *m, struct file *f) wait_queue_entry_t *wq; unsigned long pending = 0, total = 0; - spin_lock(&ctx->fault_pending_wqh.lock); + spin_lock_irq(&ctx->fault_pending_wqh.lock); list_for_each_entry(wq, &ctx->fault_pending_wqh.head, entry) { pending++; total++; @@ -1891,7 +1899,7 @@ static void userfaultfd_show_fdinfo(struct seq_file *m, struct file *f) list_for_each_entry(wq, &ctx->fault_wqh.head, entry) { total++; } - spin_unlock(&ctx->fault_pending_wqh.lock); + spin_unlock_irq(&ctx->fault_pending_wqh.lock); /* * If more protocols will be added, there will be all shown |