struct file *file;
struct wait_queue_head *head;
__poll_t events;
- bool done;
- bool canceled;
struct wait_queue_entry wait;
};
struct io_splice {
struct file *file_out;
- struct file *file_in;
loff_t off_out;
loff_t off_in;
u64 len;
+ int splice_fd_in;
unsigned int flags;
};
/* store used ubuf, so we can prevent reloading */
struct io_mapped_ubuf *imu;
+ /* stores selected buf, valid IFF REQ_F_BUFFER_SELECTED is set */
+ struct io_buffer *kbuf;
+ atomic_t poll_refs;
};
struct io_tctx_node {
bool cancel_all);
static void io_uring_cancel_generic(bool cancel_all, struct io_sq_data *sqd);
-static bool io_cqring_fill_event(struct io_ring_ctx *ctx, u64 user_data,
- long res, unsigned int cflags);
+static void io_fill_cqe_req(struct io_kiocb *req, s32 res, u32 cflags);
+
static void io_put_req(struct io_kiocb *req);
static void io_put_req_deferred(struct io_kiocb *req);
static void io_dismantle_req(struct io_kiocb *req);
return atomic_dec_and_test(&req->refs);
}
-static inline void req_ref_put(struct io_kiocb *req)
-{
- WARN_ON_ONCE(!(req->flags & REQ_F_REFCOUNT));
- WARN_ON_ONCE(req_ref_put_and_test(req));
-}
-
static inline void req_ref_get(struct io_kiocb *req)
{
WARN_ON_ONCE(!(req->flags & REQ_F_REFCOUNT));
static bool io_match_task(struct io_kiocb *head, struct task_struct *task,
bool cancel_all)
+ __must_hold(&req->ctx->timeout_lock)
{
struct io_kiocb *req;
return false;
}
+static bool io_match_linked(struct io_kiocb *head)
+{
+ struct io_kiocb *req;
+
+ io_for_each_link(req, head) {
+ if (req->flags & REQ_F_INFLIGHT)
+ return true;
+ }
+ return false;
+}
+
+/*
+ * As io_match_task() but protected against racing with linked timeouts.
+ * User must not hold timeout_lock.
+ */
+static bool io_match_task_safe(struct io_kiocb *head, struct task_struct *task,
+ bool cancel_all)
+{
+ bool matched;
+
+ if (task && head->task != task)
+ return false;
+ if (cancel_all)
+ return true;
+
+ if (head->flags & REQ_F_LINK_TIMEOUT) {
+ struct io_ring_ctx *ctx = head->ctx;
+
+ /* protect against races with linked timeouts */
+ spin_lock_irq(&ctx->timeout_lock);
+ matched = io_match_linked(head);
+ spin_unlock_irq(&ctx->timeout_lock);
+ } else {
+ matched = io_match_linked(head);
+ }
+ return matched;
+}
+
static inline void req_set_fail(struct io_kiocb *req)
{
req->flags |= REQ_F_FAIL;
{
if (!(req->flags & REQ_F_INFLIGHT)) {
req->flags |= REQ_F_INFLIGHT;
- atomic_inc(¤t->io_uring->inflight_tracked);
+ atomic_inc(&req->task->io_uring->inflight_tracked);
}
}
if (def->unbound_nonreg_file)
req->work.flags |= IO_WQ_WORK_UNBOUND;
}
-
- switch (req->opcode) {
- case IORING_OP_SPLICE:
- case IORING_OP_TEE:
- if (!S_ISREG(file_inode(req->splice.file_in)->i_mode))
- req->work.flags |= IO_WQ_WORK_UNBOUND;
- break;
- }
}
static void io_prep_async_link(struct io_kiocb *req)
if (req->flags & REQ_F_LINK_TIMEOUT) {
struct io_ring_ctx *ctx = req->ctx;
- spin_lock(&ctx->completion_lock);
+ spin_lock_irq(&ctx->timeout_lock);
io_for_each_link(cur, req)
io_prep_async_work(cur);
- spin_unlock(&ctx->completion_lock);
+ spin_unlock_irq(&ctx->timeout_lock);
} else {
io_for_each_link(cur, req)
io_prep_async_work(cur);
atomic_set(&req->ctx->cq_timeouts,
atomic_read(&req->ctx->cq_timeouts) + 1);
list_del_init(&req->timeout.list);
- io_cqring_fill_event(req->ctx, req->user_data, status, 0);
+ io_fill_cqe_req(req, status, 0);
io_put_req_deferred(req);
}
}
__must_hold(&ctx->completion_lock)
{
u32 seq = ctx->cached_cq_tail - atomic_read(&ctx->cq_timeouts);
+ struct io_kiocb *req, *tmp;
spin_lock_irq(&ctx->timeout_lock);
- while (!list_empty(&ctx->timeout_list)) {
+ list_for_each_entry_safe(req, tmp, &ctx->timeout_list, timeout.list) {
u32 events_needed, events_got;
- struct io_kiocb *req = list_first_entry(&ctx->timeout_list,
- struct io_kiocb, timeout.list);
if (io_is_timeout_noseq(req))
break;
if (events_got < events_needed)
break;
- list_del_init(&req->timeout.list);
io_kill_timeout(req, 0);
}
ctx->cq_last_tm_flush = seq;
io_task_refs_refill(tctx);
}
+static __cold void io_uring_drop_tctx_refs(struct task_struct *task)
+{
+ struct io_uring_task *tctx = task->io_uring;
+ unsigned int refs = tctx->cached_refs;
+
+ if (refs) {
+ tctx->cached_refs = 0;
+ percpu_counter_sub(&tctx->inflight, refs);
+ put_task_struct_many(task, refs);
+ }
+}
+
static bool io_cqring_event_overflow(struct io_ring_ctx *ctx, u64 user_data,
- long res, unsigned int cflags)
+ s32 res, u32 cflags)
{
struct io_overflow_cqe *ocqe;
return true;
}
-static inline bool __io_cqring_fill_event(struct io_ring_ctx *ctx, u64 user_data,
- long res, unsigned int cflags)
+static inline bool __io_fill_cqe(struct io_ring_ctx *ctx, u64 user_data,
+ s32 res, u32 cflags)
{
struct io_uring_cqe *cqe;
return io_cqring_event_overflow(ctx, user_data, res, cflags);
}
-/* not as hot to bloat with inlining */
-static noinline bool io_cqring_fill_event(struct io_ring_ctx *ctx, u64 user_data,
- long res, unsigned int cflags)
+static noinline void io_fill_cqe_req(struct io_kiocb *req, s32 res, u32 cflags)
{
- return __io_cqring_fill_event(ctx, user_data, res, cflags);
+ __io_fill_cqe(req->ctx, req->user_data, res, cflags);
}
-static void io_req_complete_post(struct io_kiocb *req, long res,
- unsigned int cflags)
+static noinline bool io_fill_cqe_aux(struct io_ring_ctx *ctx, u64 user_data,
+ s32 res, u32 cflags)
+{
+ ctx->cq_extra++;
+ return __io_fill_cqe(ctx, user_data, res, cflags);
+}
+
+static void io_req_complete_post(struct io_kiocb *req, s32 res,
+ u32 cflags)
{
struct io_ring_ctx *ctx = req->ctx;
spin_lock(&ctx->completion_lock);
- __io_cqring_fill_event(ctx, req->user_data, res, cflags);
+ __io_fill_cqe(ctx, req->user_data, res, cflags);
/*
* If we're the last reference to this request, add to our locked
* free_list cache.
return req->flags & IO_REQ_CLEAN_FLAGS;
}
-static void io_req_complete_state(struct io_kiocb *req, long res,
- unsigned int cflags)
+static inline void io_req_complete_state(struct io_kiocb *req, s32 res,
+ u32 cflags)
{
if (io_req_needs_clean(req))
io_clean_op(req);
}
static inline void __io_req_complete(struct io_kiocb *req, unsigned issue_flags,
- long res, unsigned cflags)
+ s32 res, u32 cflags)
{
if (issue_flags & IO_URING_F_COMPLETE_DEFER)
io_req_complete_state(req, res, cflags);
io_req_complete_post(req, res, cflags);
}
-static inline void io_req_complete(struct io_kiocb *req, long res)
+static inline void io_req_complete(struct io_kiocb *req, s32 res)
{
__io_req_complete(req, 0, res, 0);
}
-static void io_req_complete_failed(struct io_kiocb *req, long res)
+static void io_req_complete_failed(struct io_kiocb *req, s32 res)
{
req_set_fail(req);
io_req_complete_post(req, res, 0);
link->timeout.head = NULL;
if (hrtimer_try_to_cancel(&io->timer) != -1) {
list_del(&link->timeout.list);
- io_cqring_fill_event(link->ctx, link->user_data,
- -ECANCELED, 0);
+ io_fill_cqe_req(link, -ECANCELED, 0);
io_put_req_deferred(link);
return true;
}
link->link = NULL;
trace_io_uring_fail_link(req, link);
- io_cqring_fill_event(link->ctx, link->user_data, res, 0);
+ io_fill_cqe_req(link, res, 0);
io_put_req_deferred(link);
link = nxt;
}
req->flags &= ~REQ_F_ARM_LTIMEOUT;
if (link && link->opcode == IORING_OP_LINK_TIMEOUT) {
io_remove_next_linked(req);
- io_cqring_fill_event(link->ctx, link->user_data,
- -ECANCELED, 0);
+ io_fill_cqe_req(link, -ECANCELED, 0);
io_put_req_deferred(link);
posted = true;
}
}
ctx_flush_and_put(ctx, &locked);
+
+ /* relaxed read is enough as only the task itself sets ->in_idle */
+ if (unlikely(atomic_read(&tctx->in_idle)))
+ io_uring_drop_tctx_refs(current);
}
static void io_req_task_work_add(struct io_kiocb *req)
for (i = 0; i < nr; i++) {
struct io_kiocb *req = state->compl_reqs[i];
- __io_cqring_fill_event(ctx, req->user_data, req->result,
- req->compl.cflags);
+ __io_fill_cqe(ctx, req->user_data, req->result,
+ req->compl.cflags);
}
io_commit_cqring(ctx);
spin_unlock(&ctx->completion_lock);
req = list_first_entry(done, struct io_kiocb, inflight_entry);
list_del(&req->inflight_entry);
- __io_cqring_fill_event(ctx, req->user_data, req->result,
- io_put_rw_kbuf(req));
+ io_fill_cqe_req(req, req->result, io_put_rw_kbuf(req));
(*nr_events)++;
if (req_ref_put_and_test(req))
static bool __io_complete_rw_common(struct io_kiocb *req, long res)
{
- if (req->rw.kiocb.ki_flags & IOCB_WRITE)
+ if (req->rw.kiocb.ki_flags & IOCB_WRITE) {
kiocb_end_write(req);
+ fsnotify_modify(req->file);
+ } else {
+ fsnotify_access(req->file);
+ }
if (res != req->result) {
if ((res == -EAGAIN || res == -EOPNOTSUPP) &&
io_rw_should_reissue(req)) {
return false;
}
+static inline int io_fixup_rw_res(struct io_kiocb *req, unsigned res)
+{
+ struct io_async_rw *io = req->async_data;
+
+ /* add previously done IO, if any */
+ if (io && io->bytes_done > 0) {
+ if (res < 0)
+ res = io->bytes_done;
+ else
+ res += io->bytes_done;
+ }
+ return res;
+}
+
static void io_req_task_complete(struct io_kiocb *req, bool *locked)
{
unsigned int cflags = io_put_rw_kbuf(req);
- long res = req->result;
+ int res = req->result;
if (*locked) {
struct io_ring_ctx *ctx = req->ctx;
{
if (__io_complete_rw_common(req, res))
return;
- __io_req_complete(req, issue_flags, req->result, io_put_rw_kbuf(req));
+ __io_req_complete(req, issue_flags, io_fixup_rw_res(req, res), io_put_rw_kbuf(req));
}
static void io_complete_rw(struct kiocb *kiocb, long res, long res2)
if (__io_complete_rw_common(req, res))
return;
- req->result = res;
+ req->result = io_fixup_rw_res(req, res);
req->io_task_work.func = io_req_task_complete;
io_req_task_work_add(req);
}
req->flags |= REQ_F_ISREG;
kiocb->ki_pos = READ_ONCE(sqe->off);
- if (kiocb->ki_pos == -1 && !(file->f_mode & FMODE_STREAM)) {
- req->flags |= REQ_F_CUR_POS;
- kiocb->ki_pos = file->f_pos;
+ if (kiocb->ki_pos == -1) {
+ if (!(file->f_mode & FMODE_STREAM)) {
+ req->flags |= REQ_F_CUR_POS;
+ kiocb->ki_pos = file->f_pos;
+ } else {
+ kiocb->ki_pos = 0;
+ }
}
kiocb->ki_hint = ki_hint_validate(file_write_hint(kiocb->ki_filp));
kiocb->ki_flags = iocb_flags(kiocb->ki_filp);
kiocb->ki_complete = io_complete_rw;
}
+ /* used for fixed read/write too - just read unconditionally */
+ req->buf_index = READ_ONCE(sqe->buf_index);
+ req->imu = NULL;
+
if (req->opcode == IORING_OP_READ_FIXED ||
req->opcode == IORING_OP_WRITE_FIXED) {
- req->imu = NULL;
+ struct io_ring_ctx *ctx = req->ctx;
+ u16 index;
+
+ if (unlikely(req->buf_index >= ctx->nr_user_bufs))
+ return -EFAULT;
+ index = array_index_nospec(req->buf_index, ctx->nr_user_bufs);
+ req->imu = ctx->user_bufs[index];
io_req_set_rsrc_node(req);
}
req->rw.addr = READ_ONCE(sqe->addr);
req->rw.len = READ_ONCE(sqe->len);
- req->buf_index = READ_ONCE(sqe->buf_index);
return 0;
}
unsigned int issue_flags)
{
struct io_kiocb *req = container_of(kiocb, struct io_kiocb, rw.kiocb);
- struct io_async_rw *io = req->async_data;
-
- /* add previously done IO, if any */
- if (io && io->bytes_done > 0) {
- if (ret < 0)
- ret = io->bytes_done;
- else
- ret += io->bytes_done;
- }
if (req->flags & REQ_F_CUR_POS)
req->file->f_pos = kiocb->ki_pos;
unsigned int cflags = io_put_rw_kbuf(req);
struct io_ring_ctx *ctx = req->ctx;
+ ret = io_fixup_rw_res(req, ret);
req_set_fail(req);
if (!(issue_flags & IO_URING_F_NONBLOCK)) {
mutex_lock(&ctx->uring_lock);
static int io_import_fixed(struct io_kiocb *req, int rw, struct iov_iter *iter)
{
- struct io_ring_ctx *ctx = req->ctx;
- struct io_mapped_ubuf *imu = req->imu;
- u16 index, buf_index = req->buf_index;
-
- if (likely(!imu)) {
- if (unlikely(buf_index >= ctx->nr_user_bufs))
- return -EFAULT;
- index = array_index_nospec(buf_index, ctx->nr_user_bufs);
- imu = READ_ONCE(ctx->user_bufs[index]);
- req->imu = imu;
- }
- return __io_import_fixed(req, rw, iter, imu);
+ if (WARN_ON_ONCE(!req->imu))
+ return -EFAULT;
+ return __io_import_fixed(req, rw, iter, req->imu);
}
static void io_ring_submit_unlock(struct io_ring_ctx *ctx, bool needs_lock)
ret = nr;
break;
}
+ ret += nr;
if (!iov_iter_is_bvec(iter)) {
iov_iter_advance(iter, nr);
} else {
- req->rw.len -= nr;
req->rw.addr += nr;
+ req->rw.len -= nr;
+ if (!req->rw.len)
+ break;
}
- ret += nr;
if (nr != iovec.iov_len)
break;
}
return -EAGAIN;
}
+ req->result = iov_iter_count(iter);
/*
* Now retry read with the IOCB_WAITQ parts set in the iocb. If
* we get -EIOCBQUEUED, then we'll get a notification when the
copy_iov:
iov_iter_restore(iter, state);
ret = io_setup_async_rw(req, iovec, inline_vecs, iter, false);
- return ret ?: -EAGAIN;
+ if (!ret) {
+ if (kiocb->ki_flags & IOCB_WRITE)
+ kiocb_end_write(req);
+ return -EAGAIN;
+ }
+ return ret;
}
out_free:
/* it's reportedly faster than delegating the null check to kfree() */
if (unlikely(req->ctx->flags & IORING_SETUP_IOPOLL))
return -EINVAL;
- sp->file_in = NULL;
sp->len = READ_ONCE(sqe->len);
sp->flags = READ_ONCE(sqe->splice_flags);
-
if (unlikely(sp->flags & ~valid_flags))
return -EINVAL;
-
- sp->file_in = io_file_get(req->ctx, req, READ_ONCE(sqe->splice_fd_in),
- (sp->flags & SPLICE_F_FD_IN_FIXED));
- if (!sp->file_in)
- return -EBADF;
- req->flags |= REQ_F_NEED_CLEANUP;
+ sp->splice_fd_in = READ_ONCE(sqe->splice_fd_in);
return 0;
}
static int io_tee(struct io_kiocb *req, unsigned int issue_flags)
{
struct io_splice *sp = &req->splice;
- struct file *in = sp->file_in;
struct file *out = sp->file_out;
unsigned int flags = sp->flags & ~SPLICE_F_FD_IN_FIXED;
+ struct file *in;
long ret = 0;
if (issue_flags & IO_URING_F_NONBLOCK)
return -EAGAIN;
+
+ in = io_file_get(req->ctx, req, sp->splice_fd_in,
+ (sp->flags & SPLICE_F_FD_IN_FIXED));
+ if (!in) {
+ ret = -EBADF;
+ goto done;
+ }
+
if (sp->len)
ret = do_tee(in, out, sp->len, flags);
if (!(sp->flags & SPLICE_F_FD_IN_FIXED))
io_put_file(in);
- req->flags &= ~REQ_F_NEED_CLEANUP;
-
+done:
if (ret != sp->len)
req_set_fail(req);
io_req_complete(req, ret);
static int io_splice(struct io_kiocb *req, unsigned int issue_flags)
{
struct io_splice *sp = &req->splice;
- struct file *in = sp->file_in;
struct file *out = sp->file_out;
unsigned int flags = sp->flags & ~SPLICE_F_FD_IN_FIXED;
loff_t *poff_in, *poff_out;
+ struct file *in;
long ret = 0;
if (issue_flags & IO_URING_F_NONBLOCK)
return -EAGAIN;
+ in = io_file_get(req->ctx, req, sp->splice_fd_in,
+ (sp->flags & SPLICE_F_FD_IN_FIXED));
+ if (!in) {
+ ret = -EBADF;
+ goto done;
+ }
+
poff_in = (sp->off_in == -1) ? NULL : &sp->off_in;
poff_out = (sp->off_out == -1) ? NULL : &sp->off_out;
if (!(sp->flags & SPLICE_F_FD_IN_FIXED))
io_put_file(in);
- req->flags &= ~REQ_F_NEED_CLEANUP;
-
+done:
if (ret != sp->len)
req_set_fail(req);
io_req_complete(req, ret);
{
struct io_ring_ctx *ctx = req->ctx;
- if (!req->file)
- return -EBADF;
-
if (unlikely(ctx->flags & IORING_SETUP_IOPOLL))
return -EINVAL;
if (unlikely(sqe->addr || sqe->ioprio || sqe->buf_index ||
req->sync.len);
if (ret < 0)
req_set_fail(req);
+ else
+ fsnotify_modify(req->file);
io_req_complete(req, ret);
return 0;
}
kfree(nxt);
if (++i == nbufs)
return i;
+ cond_resched();
}
i++;
kfree(buf);
} else {
list_add_tail(&buf->list, &(*head)->list);
}
+ cond_resched();
}
return i ? i : -ENOMEM;
ret = io_add_buffers(p, &head);
if (ret >= 0 && !list) {
- ret = xa_insert(&ctx->io_buffers, p->bgid, head, GFP_KERNEL);
+ ret = xa_insert(&ctx->io_buffers, p->bgid, head,
+ GFP_KERNEL_ACCOUNT);
if (ret < 0)
__io_remove_buffers(ctx, head, p->bgid, -1U);
}
async_msg = req->async_data;
req->flags |= REQ_F_NEED_CLEANUP;
memcpy(async_msg, kmsg, sizeof(*kmsg));
- async_msg->msg.msg_name = &async_msg->addr;
+ if (async_msg->msg.msg_name)
+ async_msg->msg.msg_name = &async_msg->addr;
/* if were using fast_iov, set it to the new one */
if (!async_msg->free_iov)
async_msg->msg.msg_iter.iov = async_msg->fast_iov;
if (unlikely(req->ctx->flags & IORING_SETUP_IOPOLL))
return -EINVAL;
+ if (unlikely(sqe->addr2 || sqe->file_index))
+ return -EINVAL;
+ if (unlikely(sqe->addr2 || sqe->file_index || sqe->ioprio))
+ return -EINVAL;
sr->umsg = u64_to_user_ptr(READ_ONCE(sqe->addr));
sr->len = READ_ONCE(sqe->len);
if (unlikely(req->ctx->flags & IORING_SETUP_IOPOLL))
return -EINVAL;
+ if (unlikely(sqe->addr2 || sqe->file_index))
+ return -EINVAL;
+ if (unlikely(sqe->addr2 || sqe->file_index || sqe->ioprio))
+ return -EINVAL;
sr->umsg = u64_to_user_ptr(READ_ONCE(sqe->addr));
sr->len = READ_ONCE(sqe->len);
accept->nofile = rlimit(RLIMIT_NOFILE);
accept->file_slot = READ_ONCE(sqe->file_index);
- if (accept->file_slot && ((req->open.how.flags & O_CLOEXEC) ||
- (accept->flags & SOCK_CLOEXEC)))
+ if (accept->file_slot && (accept->flags & SOCK_CLOEXEC))
return -EINVAL;
if (accept->flags & ~(SOCK_CLOEXEC | SOCK_NONBLOCK))
return -EINVAL;
int error;
};
-static int __io_async_wake(struct io_kiocb *req, struct io_poll_iocb *poll,
- __poll_t mask, io_req_tw_func_t func)
-{
- /* for instances that support it check for an event match first: */
- if (mask && !(mask & poll->events))
- return 0;
-
- trace_io_uring_task_add(req->ctx, req->opcode, req->user_data, mask);
-
- list_del_init(&poll->wait.entry);
-
- req->result = mask;
- req->io_task_work.func = func;
+#define IO_POLL_CANCEL_FLAG BIT(31)
+#define IO_POLL_REF_MASK GENMASK(30, 0)
- /*
- * If this fails, then the task is exiting. When a task exits, the
- * work gets canceled, so just cancel this request as well instead
- * of executing it. We can't safely execute it anyway, as we may not
- * have the needed state needed for it anyway.
- */
- io_req_task_work_add(req);
- return 1;
+/*
+ * If refs part of ->poll_refs (see IO_POLL_REF_MASK) is 0, it's free. We can
+ * bump it and acquire ownership. It's disallowed to modify requests while not
+ * owning it, that prevents from races for enqueueing task_work's and b/w
+ * arming poll and wakeups.
+ */
+static inline bool io_poll_get_ownership(struct io_kiocb *req)
+{
+ return !(atomic_fetch_inc(&req->poll_refs) & IO_POLL_REF_MASK);
}
-static bool io_poll_rewait(struct io_kiocb *req, struct io_poll_iocb *poll)
- __acquires(&req->ctx->completion_lock)
+static void io_poll_mark_cancelled(struct io_kiocb *req)
{
- struct io_ring_ctx *ctx = req->ctx;
-
- /* req->task == current here, checking PF_EXITING is safe */
- if (unlikely(req->task->flags & PF_EXITING))
- WRITE_ONCE(poll->canceled, true);
-
- if (!req->result && !READ_ONCE(poll->canceled)) {
- struct poll_table_struct pt = { ._key = poll->events };
-
- req->result = vfs_poll(req->file, &pt) & poll->events;
- }
-
- spin_lock(&ctx->completion_lock);
- if (!req->result && !READ_ONCE(poll->canceled)) {
- add_wait_queue(poll->head, &poll->wait);
- return true;
- }
-
- return false;
+ atomic_or(IO_POLL_CANCEL_FLAG, &req->poll_refs);
}
static struct io_poll_iocb *io_poll_get_double(struct io_kiocb *req)
return &req->apoll->poll;
}
-static void io_poll_remove_double(struct io_kiocb *req)
- __must_hold(&req->ctx->completion_lock)
+static void io_poll_req_insert(struct io_kiocb *req)
{
- struct io_poll_iocb *poll = io_poll_get_double(req);
+ struct io_ring_ctx *ctx = req->ctx;
+ struct hlist_head *list;
- lockdep_assert_held(&req->ctx->completion_lock);
+ list = &ctx->cancel_hash[hash_long(req->user_data, ctx->cancel_hash_bits)];
+ hlist_add_head(&req->hash_node, list);
+}
- if (poll && poll->head) {
- struct wait_queue_head *head = poll->head;
+static void io_init_poll_iocb(struct io_poll_iocb *poll, __poll_t events,
+ wait_queue_func_t wake_func)
+{
+ poll->head = NULL;
+#define IO_POLL_UNMASK (EPOLLERR|EPOLLHUP|EPOLLNVAL|EPOLLRDHUP)
+ /* mask in events that we always want/need */
+ poll->events = events | IO_POLL_UNMASK;
+ INIT_LIST_HEAD(&poll->wait.entry);
+ init_waitqueue_func_entry(&poll->wait, wake_func);
+}
+
+static inline void io_poll_remove_entry(struct io_poll_iocb *poll)
+{
+ struct wait_queue_head *head = smp_load_acquire(&poll->head);
+ if (head) {
spin_lock_irq(&head->lock);
list_del_init(&poll->wait.entry);
- if (poll->wait.private)
- req_ref_put(req);
poll->head = NULL;
spin_unlock_irq(&head->lock);
}
}
-static bool __io_poll_complete(struct io_kiocb *req, __poll_t mask)
- __must_hold(&req->ctx->completion_lock)
+static void io_poll_remove_entries(struct io_kiocb *req)
+{
+ struct io_poll_iocb *poll = io_poll_get_single(req);
+ struct io_poll_iocb *poll_double = io_poll_get_double(req);
+
+ /*
+ * While we hold the waitqueue lock and the waitqueue is nonempty,
+ * wake_up_pollfree() will wait for us. However, taking the waitqueue
+ * lock in the first place can race with the waitqueue being freed.
+ *
+ * We solve this as eventpoll does: by taking advantage of the fact that
+ * all users of wake_up_pollfree() will RCU-delay the actual free. If
+ * we enter rcu_read_lock() and see that the pointer to the queue is
+ * non-NULL, we can then lock it without the memory being freed out from
+ * under us.
+ *
+ * Keep holding rcu_read_lock() as long as we hold the queue lock, in
+ * case the caller deletes the entry from the queue, leaving it empty.
+ * In that case, only RCU prevents the queue memory from being freed.
+ */
+ rcu_read_lock();
+ io_poll_remove_entry(poll);
+ if (poll_double)
+ io_poll_remove_entry(poll_double);
+ rcu_read_unlock();
+}
+
+/*
+ * All poll tw should go through this. Checks for poll events, manages
+ * references, does rewait, etc.
+ *
+ * Returns a negative error on failure. >0 when no action require, which is
+ * either spurious wakeup or multishot CQE is served. 0 when it's done with
+ * the request, then the mask is stored in req->result.
+ */
+static int io_poll_check_events(struct io_kiocb *req)
{
struct io_ring_ctx *ctx = req->ctx;
- unsigned flags = IORING_CQE_F_MORE;
- int error;
+ struct io_poll_iocb *poll = io_poll_get_single(req);
+ int v;
+
+ /* req->task == current here, checking PF_EXITING is safe */
+ if (unlikely(req->task->flags & PF_EXITING))
+ io_poll_mark_cancelled(req);
+
+ do {
+ v = atomic_read(&req->poll_refs);
- if (READ_ONCE(req->poll.canceled)) {
- error = -ECANCELED;
- req->poll.events |= EPOLLONESHOT;
+ /* tw handler should be the owner, and so have some references */
+ if (WARN_ON_ONCE(!(v & IO_POLL_REF_MASK)))
+ return 0;
+ if (v & IO_POLL_CANCEL_FLAG)
+ return -ECANCELED;
+
+ if (!req->result) {
+ struct poll_table_struct pt = { ._key = poll->events };
+
+ req->result = vfs_poll(req->file, &pt) & poll->events;
+ }
+
+ /* multishot, just fill an CQE and proceed */
+ if (req->result && !(poll->events & EPOLLONESHOT)) {
+ __poll_t mask = mangle_poll(req->result & poll->events);
+ bool filled;
+
+ spin_lock(&ctx->completion_lock);
+ filled = io_fill_cqe_aux(ctx, req->user_data, mask,
+ IORING_CQE_F_MORE);
+ io_commit_cqring(ctx);
+ spin_unlock(&ctx->completion_lock);
+ if (unlikely(!filled))
+ return -ECANCELED;
+ io_cqring_ev_posted(ctx);
+ } else if (req->result) {
+ return 0;
+ }
+
+ /*
+ * Release all references, retry if someone tried to restart
+ * task_work while we were executing it.
+ */
+ } while (atomic_sub_return(v & IO_POLL_REF_MASK, &req->poll_refs));
+
+ return 1;
+}
+
+static void io_poll_task_func(struct io_kiocb *req, bool *locked)
+{
+ struct io_ring_ctx *ctx = req->ctx;
+ int ret;
+
+ ret = io_poll_check_events(req);
+ if (ret > 0)
+ return;
+
+ if (!ret) {
+ req->result = mangle_poll(req->result & req->poll.events);
} else {
- error = mangle_poll(mask);
- }
- if (req->poll.events & EPOLLONESHOT)
- flags = 0;
- if (!io_cqring_fill_event(ctx, req->user_data, error, flags)) {
- req->poll.events |= EPOLLONESHOT;
- flags = 0;
+ req->result = ret;
+ req_set_fail(req);
}
- if (flags & IORING_CQE_F_MORE)
- ctx->cq_extra++;
- return !(flags & IORING_CQE_F_MORE);
+ io_poll_remove_entries(req);
+ spin_lock(&ctx->completion_lock);
+ hash_del(&req->hash_node);
+ spin_unlock(&ctx->completion_lock);
+ io_req_complete_post(req, req->result, 0);
}
-static inline bool io_poll_complete(struct io_kiocb *req, __poll_t mask)
- __must_hold(&req->ctx->completion_lock)
+static void io_apoll_task_func(struct io_kiocb *req, bool *locked)
{
- bool done;
+ struct io_ring_ctx *ctx = req->ctx;
+ int ret;
- done = __io_poll_complete(req, mask);
- io_commit_cqring(req->ctx);
- return done;
+ ret = io_poll_check_events(req);
+ if (ret > 0)
+ return;
+
+ io_poll_remove_entries(req);
+ spin_lock(&ctx->completion_lock);
+ hash_del(&req->hash_node);
+ spin_unlock(&ctx->completion_lock);
+
+ if (!ret)
+ io_req_task_submit(req, locked);
+ else
+ io_req_complete_failed(req, ret);
}
-static void io_poll_task_func(struct io_kiocb *req, bool *locked)
+static void __io_poll_execute(struct io_kiocb *req, int mask)
{
- struct io_ring_ctx *ctx = req->ctx;
- struct io_kiocb *nxt;
+ req->result = mask;
+ if (req->opcode == IORING_OP_POLL_ADD)
+ req->io_task_work.func = io_poll_task_func;
+ else
+ req->io_task_work.func = io_apoll_task_func;
- if (io_poll_rewait(req, &req->poll)) {
- spin_unlock(&ctx->completion_lock);
- } else {
- bool done;
+ trace_io_uring_task_add(req->ctx, req->opcode, req->user_data, mask);
+ io_req_task_work_add(req);
+}
- if (req->poll.done) {
- spin_unlock(&ctx->completion_lock);
- return;
- }
- done = __io_poll_complete(req, req->result);
- if (done) {
- io_poll_remove_double(req);
- hash_del(&req->hash_node);
- req->poll.done = true;
- } else {
- req->result = 0;
- add_wait_queue(req->poll.head, &req->poll.wait);
- }
- io_commit_cqring(ctx);
- spin_unlock(&ctx->completion_lock);
- io_cqring_ev_posted(ctx);
+static inline void io_poll_execute(struct io_kiocb *req, int res)
+{
+ if (io_poll_get_ownership(req))
+ __io_poll_execute(req, res);
+}
- if (done) {
- nxt = io_put_req_find_next(req);
- if (nxt)
- io_req_task_submit(nxt, locked);
- }
- }
+static void io_poll_cancel_req(struct io_kiocb *req)
+{
+ io_poll_mark_cancelled(req);
+ /* kick tw, which should complete the request */
+ io_poll_execute(req, 0);
}
-static int io_poll_double_wake(struct wait_queue_entry *wait, unsigned mode,
- int sync, void *key)
+static int io_poll_wake(struct wait_queue_entry *wait, unsigned mode, int sync,
+ void *key)
{
struct io_kiocb *req = wait->private;
- struct io_poll_iocb *poll = io_poll_get_single(req);
+ struct io_poll_iocb *poll = container_of(wait, struct io_poll_iocb,
+ wait);
__poll_t mask = key_to_poll(key);
- unsigned long flags;
- /* for instances that support it check for an event match first: */
- if (mask && !(mask & poll->events))
- return 0;
- if (!(poll->events & EPOLLONESHOT))
- return poll->wait.func(&poll->wait, mode, sync, key);
+ if (unlikely(mask & POLLFREE)) {
+ io_poll_mark_cancelled(req);
+ /* we have to kick tw in case it's not already */
+ io_poll_execute(req, 0);
- list_del_init(&wait->entry);
+ /*
+ * If the waitqueue is being freed early but someone is already
+ * holds ownership over it, we have to tear down the request as
+ * best we can. That means immediately removing the request from
+ * its waitqueue and preventing all further accesses to the
+ * waitqueue via the request.
+ */
+ list_del_init(&poll->wait.entry);
- if (poll->head) {
- bool done;
-
- spin_lock_irqsave(&poll->head->lock, flags);
- done = list_empty(&poll->wait.entry);
- if (!done)
- list_del_init(&poll->wait.entry);
- /* make sure double remove sees this as being gone */
- wait->private = NULL;
- spin_unlock_irqrestore(&poll->head->lock, flags);
- if (!done) {
- /* use wait func handler, so it matches the rq type */
- poll->wait.func(&poll->wait, mode, sync, key);
- }
+ /*
+ * Careful: this *must* be the last step, since as soon
+ * as req->head is NULL'ed out, the request can be
+ * completed and freed, since aio_poll_complete_work()
+ * will no longer need to take the waitqueue lock.
+ */
+ smp_store_release(&poll->head, NULL);
+ return 1;
}
- req_ref_put(req);
- return 1;
-}
-static void io_init_poll_iocb(struct io_poll_iocb *poll, __poll_t events,
- wait_queue_func_t wake_func)
-{
- poll->head = NULL;
- poll->done = false;
- poll->canceled = false;
-#define IO_POLL_UNMASK (EPOLLERR|EPOLLHUP|EPOLLNVAL|EPOLLRDHUP)
- /* mask in events that we always want/need */
- poll->events = events | IO_POLL_UNMASK;
- INIT_LIST_HEAD(&poll->wait.entry);
- init_waitqueue_func_entry(&poll->wait, wake_func);
+ /* for instances that support it check for an event match first */
+ if (mask && !(mask & poll->events))
+ return 0;
+
+ if (io_poll_get_ownership(req))
+ __io_poll_execute(req, mask);
+ return 1;
}
static void __io_queue_proc(struct io_poll_iocb *poll, struct io_poll_table *pt,
* if this happens.
*/
if (unlikely(pt->nr_entries)) {
- struct io_poll_iocb *poll_one = poll;
+ struct io_poll_iocb *first = poll;
/* double add on the same waitqueue head, ignore */
- if (poll_one->head == head)
+ if (first->head == head)
return;
/* already have a 2nd entry, fail a third attempt */
if (*poll_ptr) {
pt->error = -EINVAL;
return;
}
- /*
- * Can't handle multishot for double wait for now, turn it
- * into one-shot mode.
- */
- if (!(poll_one->events & EPOLLONESHOT))
- poll_one->events |= EPOLLONESHOT;
+
poll = kmalloc(sizeof(*poll), GFP_ATOMIC);
if (!poll) {
pt->error = -ENOMEM;
return;
}
- io_init_poll_iocb(poll, poll_one->events, io_poll_double_wake);
- req_ref_get(req);
- poll->wait.private = req;
+ io_init_poll_iocb(poll, first->events, first->wait.func);
*poll_ptr = poll;
}
pt->nr_entries++;
poll->head = head;
+ poll->wait.private = req;
if (poll->events & EPOLLEXCLUSIVE)
add_wait_queue_exclusive(head, &poll->wait);
add_wait_queue(head, &poll->wait);
}
-static void io_async_queue_proc(struct file *file, struct wait_queue_head *head,
+static void io_poll_queue_proc(struct file *file, struct wait_queue_head *head,
struct poll_table_struct *p)
{
struct io_poll_table *pt = container_of(p, struct io_poll_table, pt);
- struct async_poll *apoll = pt->req->apoll;
-
- __io_queue_proc(&apoll->poll, pt, head, &apoll->double_poll);
-}
-
-static void io_async_task_func(struct io_kiocb *req, bool *locked)
-{
- struct async_poll *apoll = req->apoll;
- struct io_ring_ctx *ctx = req->ctx;
-
- trace_io_uring_task_run(req->ctx, req, req->opcode, req->user_data);
-
- if (io_poll_rewait(req, &apoll->poll)) {
- spin_unlock(&ctx->completion_lock);
- return;
- }
-
- hash_del(&req->hash_node);
- io_poll_remove_double(req);
- apoll->poll.done = true;
- spin_unlock(&ctx->completion_lock);
-
- if (!READ_ONCE(apoll->poll.canceled))
- io_req_task_submit(req, locked);
- else
- io_req_complete_failed(req, -ECANCELED);
-}
-
-static int io_async_wake(struct wait_queue_entry *wait, unsigned mode, int sync,
- void *key)
-{
- struct io_kiocb *req = wait->private;
- struct io_poll_iocb *poll = &req->apoll->poll;
-
- trace_io_uring_poll_wake(req->ctx, req->opcode, req->user_data,
- key_to_poll(key));
-
- return __io_async_wake(req, poll, key_to_poll(key), io_async_task_func);
-}
-
-static void io_poll_req_insert(struct io_kiocb *req)
-{
- struct io_ring_ctx *ctx = req->ctx;
- struct hlist_head *list;
- list = &ctx->cancel_hash[hash_long(req->user_data, ctx->cancel_hash_bits)];
- hlist_add_head(&req->hash_node, list);
+ __io_queue_proc(&pt->req->poll, pt, head,
+ (struct io_poll_iocb **) &pt->req->async_data);
}
-static __poll_t __io_arm_poll_handler(struct io_kiocb *req,
- struct io_poll_iocb *poll,
- struct io_poll_table *ipt, __poll_t mask,
- wait_queue_func_t wake_func)
- __acquires(&ctx->completion_lock)
+static int __io_arm_poll_handler(struct io_kiocb *req,
+ struct io_poll_iocb *poll,
+ struct io_poll_table *ipt, __poll_t mask)
{
struct io_ring_ctx *ctx = req->ctx;
- bool cancel = false;
+ int v;
INIT_HLIST_NODE(&req->hash_node);
- io_init_poll_iocb(poll, mask, wake_func);
+ io_init_poll_iocb(poll, mask, io_poll_wake);
poll->file = req->file;
poll->wait.private = req;
ipt->error = 0;
ipt->nr_entries = 0;
+ /*
+ * Take the ownership to delay any tw execution up until we're done
+ * with poll arming. see io_poll_get_ownership().
+ */
+ atomic_set(&req->poll_refs, 1);
mask = vfs_poll(req->file, &ipt->pt) & poll->events;
- if (unlikely(!ipt->nr_entries) && !ipt->error)
- ipt->error = -EINVAL;
+
+ if (mask && (poll->events & EPOLLONESHOT)) {
+ io_poll_remove_entries(req);
+ /* no one else has access to the req, forget about the ref */
+ return mask;
+ }
+ if (!mask && unlikely(ipt->error || !ipt->nr_entries)) {
+ io_poll_remove_entries(req);
+ if (!ipt->error)
+ ipt->error = -EINVAL;
+ return 0;
+ }
spin_lock(&ctx->completion_lock);
- if (ipt->error || (mask && (poll->events & EPOLLONESHOT)))
- io_poll_remove_double(req);
- if (likely(poll->head)) {
- spin_lock_irq(&poll->head->lock);
- if (unlikely(list_empty(&poll->wait.entry))) {
- if (ipt->error)
- cancel = true;
+ io_poll_req_insert(req);
+ spin_unlock(&ctx->completion_lock);
+
+ if (mask) {
+ /* can't multishot if failed, just queue the event we've got */
+ if (unlikely(ipt->error || !ipt->nr_entries)) {
+ poll->events |= EPOLLONESHOT;
ipt->error = 0;
- mask = 0;
}
- if ((mask && (poll->events & EPOLLONESHOT)) || ipt->error)
- list_del_init(&poll->wait.entry);
- else if (cancel)
- WRITE_ONCE(poll->canceled, true);
- else if (!poll->done) /* actually waiting for an event */
- io_poll_req_insert(req);
- spin_unlock_irq(&poll->head->lock);
+ __io_poll_execute(req, mask);
+ return 0;
}
- return mask;
+ /*
+ * Release ownership. If someone tried to queue a tw while it was
+ * locked, kick it off for them.
+ */
+ v = atomic_dec_return(&req->poll_refs);
+ if (unlikely(v & IO_POLL_REF_MASK))
+ __io_poll_execute(req, 0);
+ return 0;
+}
+
+static void io_async_queue_proc(struct file *file, struct wait_queue_head *head,
+ struct poll_table_struct *p)
+{
+ struct io_poll_table *pt = container_of(p, struct io_poll_table, pt);
+ struct async_poll *apoll = pt->req->apoll;
+
+ __io_queue_proc(&apoll->poll, pt, head, &apoll->double_poll);
}
enum {
struct io_ring_ctx *ctx = req->ctx;
struct async_poll *apoll;
struct io_poll_table ipt;
- __poll_t ret, mask = EPOLLONESHOT | POLLERR | POLLPRI;
- int rw;
+ __poll_t mask = EPOLLONESHOT | POLLERR | POLLPRI;
+ int ret;
if (!req->file || !file_can_poll(req->file))
return IO_APOLL_ABORTED;
return IO_APOLL_ABORTED;
if (def->pollin) {
- rw = READ;
mask |= POLLIN | POLLRDNORM;
/* If reading from MSG_ERRQUEUE using recvmsg, ignore POLLIN */
(req->sr_msg.msg_flags & MSG_ERRQUEUE))
mask &= ~POLLIN;
} else {
- rw = WRITE;
mask |= POLLOUT | POLLWRNORM;
}
- /* if we can't nonblock try, then no point in arming a poll handler */
- if (!io_file_supports_nowait(req, rw))
- return IO_APOLL_ABORTED;
-
apoll = kmalloc(sizeof(*apoll), GFP_ATOMIC);
if (unlikely(!apoll))
return IO_APOLL_ABORTED;
req->apoll = apoll;
req->flags |= REQ_F_POLLED;
ipt.pt._qproc = io_async_queue_proc;
- io_req_set_refcount(req);
- ret = __io_arm_poll_handler(req, &apoll->poll, &ipt, mask,
- io_async_wake);
- spin_unlock(&ctx->completion_lock);
+ ret = __io_arm_poll_handler(req, &apoll->poll, &ipt, mask);
if (ret || ipt.error)
return ret ? IO_APOLL_READY : IO_APOLL_ABORTED;
return IO_APOLL_OK;
}
-static bool __io_poll_remove_one(struct io_kiocb *req,
- struct io_poll_iocb *poll, bool do_cancel)
- __must_hold(&req->ctx->completion_lock)
-{
- bool do_complete = false;
-
- if (!poll->head)
- return false;
- spin_lock_irq(&poll->head->lock);
- if (do_cancel)
- WRITE_ONCE(poll->canceled, true);
- if (!list_empty(&poll->wait.entry)) {
- list_del_init(&poll->wait.entry);
- do_complete = true;
- }
- spin_unlock_irq(&poll->head->lock);
- hash_del(&req->hash_node);
- return do_complete;
-}
-
-static bool io_poll_remove_one(struct io_kiocb *req)
- __must_hold(&req->ctx->completion_lock)
-{
- bool do_complete;
-
- io_poll_remove_double(req);
- do_complete = __io_poll_remove_one(req, io_poll_get_single(req), true);
-
- if (do_complete) {
- io_cqring_fill_event(req->ctx, req->user_data, -ECANCELED, 0);
- io_commit_cqring(req->ctx);
- req_set_fail(req);
- io_put_req_deferred(req);
- }
- return do_complete;
-}
-
/*
* Returns true if we found and killed one or more poll requests
*/
{
struct hlist_node *tmp;
struct io_kiocb *req;
- int posted = 0, i;
+ bool found = false;
+ int i;
spin_lock(&ctx->completion_lock);
for (i = 0; i < (1U << ctx->cancel_hash_bits); i++) {
list = &ctx->cancel_hash[i];
hlist_for_each_entry_safe(req, tmp, list, hash_node) {
- if (io_match_task(req, tsk, cancel_all))
- posted += io_poll_remove_one(req);
+ if (io_match_task_safe(req, tsk, cancel_all)) {
+ hlist_del_init(&req->hash_node);
+ io_poll_cancel_req(req);
+ found = true;
+ }
}
}
spin_unlock(&ctx->completion_lock);
-
- if (posted)
- io_cqring_ev_posted(ctx);
-
- return posted != 0;
+ return found;
}
static struct io_kiocb *io_poll_find(struct io_ring_ctx *ctx, __u64 sqe_addr,
return NULL;
}
+static bool io_poll_disarm(struct io_kiocb *req)
+ __must_hold(&ctx->completion_lock)
+{
+ if (!io_poll_get_ownership(req))
+ return false;
+ io_poll_remove_entries(req);
+ hash_del(&req->hash_node);
+ return true;
+}
+
static int io_poll_cancel(struct io_ring_ctx *ctx, __u64 sqe_addr,
bool poll_only)
__must_hold(&ctx->completion_lock)
{
- struct io_kiocb *req;
+ struct io_kiocb *req = io_poll_find(ctx, sqe_addr, poll_only);
- req = io_poll_find(ctx, sqe_addr, poll_only);
if (!req)
return -ENOENT;
- if (io_poll_remove_one(req))
- return 0;
-
- return -EALREADY;
+ io_poll_cancel_req(req);
+ return 0;
}
static __poll_t io_poll_parse_events(const struct io_uring_sqe *sqe,
return 0;
}
-static int io_poll_wake(struct wait_queue_entry *wait, unsigned mode, int sync,
- void *key)
-{
- struct io_kiocb *req = wait->private;
- struct io_poll_iocb *poll = &req->poll;
-
- return __io_async_wake(req, poll, key_to_poll(key), io_poll_task_func);
-}
-
-static void io_poll_queue_proc(struct file *file, struct wait_queue_head *head,
- struct poll_table_struct *p)
-{
- struct io_poll_table *pt = container_of(p, struct io_poll_table, pt);
-
- __io_queue_proc(&pt->req->poll, pt, head, (struct io_poll_iocb **) &pt->req->async_data);
-}
-
static int io_poll_add_prep(struct io_kiocb *req, const struct io_uring_sqe *sqe)
{
struct io_poll_iocb *poll = &req->poll;
static int io_poll_add(struct io_kiocb *req, unsigned int issue_flags)
{
struct io_poll_iocb *poll = &req->poll;
- struct io_ring_ctx *ctx = req->ctx;
struct io_poll_table ipt;
- __poll_t mask;
- bool done;
+ int ret;
ipt.pt._qproc = io_poll_queue_proc;
- mask = __io_arm_poll_handler(req, &req->poll, &ipt, poll->events,
- io_poll_wake);
-
- if (mask) { /* no async, we'd stolen it */
- ipt.error = 0;
- done = io_poll_complete(req, mask);
- }
- spin_unlock(&ctx->completion_lock);
-
- if (mask) {
- io_cqring_ev_posted(ctx);
- if (done)
- io_put_req(req);
- }
- return ipt.error;
+ ret = __io_arm_poll_handler(req, &req->poll, &ipt, poll->events);
+ if (!ret && ipt.error)
+ req_set_fail(req);
+ ret = ret ?: ipt.error;
+ if (ret)
+ __io_req_complete(req, issue_flags, ret, 0);
+ return 0;
}
static int io_poll_update(struct io_kiocb *req, unsigned int issue_flags)
{
struct io_ring_ctx *ctx = req->ctx;
struct io_kiocb *preq;
- bool completing;
- int ret;
+ int ret2, ret = 0;
spin_lock(&ctx->completion_lock);
preq = io_poll_find(ctx, req->poll_update.old_user_data, true);
- if (!preq) {
- ret = -ENOENT;
- goto err;
+ if (!preq || !io_poll_disarm(preq)) {
+ spin_unlock(&ctx->completion_lock);
+ ret = preq ? -EALREADY : -ENOENT;
+ goto out;
}
+ spin_unlock(&ctx->completion_lock);
- if (!req->poll_update.update_events && !req->poll_update.update_user_data) {
- completing = true;
- ret = io_poll_remove_one(preq) ? 0 : -EALREADY;
- goto err;
- }
+ if (req->poll_update.update_events || req->poll_update.update_user_data) {
+ /* only mask one event flags, keep behavior flags */
+ if (req->poll_update.update_events) {
+ preq->poll.events &= ~0xffff;
+ preq->poll.events |= req->poll_update.events & 0xffff;
+ preq->poll.events |= IO_POLL_UNMASK;
+ }
+ if (req->poll_update.update_user_data)
+ preq->user_data = req->poll_update.new_user_data;
- /*
- * Don't allow racy completion with singleshot, as we cannot safely
- * update those. For multishot, if we're racing with completion, just
- * let completion re-add it.
- */
- completing = !__io_poll_remove_one(preq, &preq->poll, false);
- if (completing && (preq->poll.events & EPOLLONESHOT)) {
- ret = -EALREADY;
- goto err;
+ ret2 = io_poll_add(preq, issue_flags);
+ /* successfully updated, don't complete poll request */
+ if (!ret2)
+ goto out;
}
- /* we now have a detached poll request. reissue. */
- ret = 0;
-err:
- if (ret < 0) {
- spin_unlock(&ctx->completion_lock);
+ req_set_fail(preq);
+ io_req_complete(preq, -ECANCELED);
+out:
+ if (ret < 0)
req_set_fail(req);
- io_req_complete(req, ret);
- return 0;
- }
- /* only mask one event flags, keep behavior flags */
- if (req->poll_update.update_events) {
- preq->poll.events &= ~0xffff;
- preq->poll.events |= req->poll_update.events & 0xffff;
- preq->poll.events |= IO_POLL_UNMASK;
- }
- if (req->poll_update.update_user_data)
- preq->user_data = req->poll_update.new_user_data;
- spin_unlock(&ctx->completion_lock);
-
/* complete update request, we're done with it */
io_req_complete(req, ret);
-
- if (!completing) {
- ret = io_poll_add(preq, issue_flags);
- if (ret < 0) {
- req_set_fail(preq);
- io_req_complete(preq, ret);
- }
- }
return 0;
}
return PTR_ERR(req);
req_set_fail(req);
- io_cqring_fill_event(ctx, req->user_data, -ECANCELED, 0);
+ io_fill_cqe_req(req, -ECANCELED, 0);
io_put_req_deferred(req);
return 0;
}
if (get_timespec64(&data->ts, u64_to_user_ptr(sqe->addr)))
return -EFAULT;
+ INIT_LIST_HEAD(&req->timeout.list);
data->mode = io_translate_timeout_mode(flags);
hrtimer_init(&data->timer, io_timeout_get_clock(data), data->mode);
up.nr = 0;
up.tags = 0;
up.resv = 0;
+ up.resv2 = 0;
io_ring_submit_lock(ctx, !(issue_flags & IO_URING_F_NONBLOCK));
ret = __io_register_rsrc_update(ctx, IORING_RSRC_FILE,
}
/* Still need defer if there is pending req in defer list. */
+ spin_lock(&ctx->completion_lock);
if (likely(list_empty_careful(&ctx->defer_list) &&
!(req->flags & REQ_F_IO_DRAIN))) {
+ spin_unlock(&ctx->completion_lock);
ctx->drain_active = false;
return false;
}
+ spin_unlock(&ctx->completion_lock);
seq = io_get_sequence(req);
/* Still a chance to pass the sequence check */
kfree(io->free_iov);
break;
}
- case IORING_OP_SPLICE:
- case IORING_OP_TEE:
- if (!(req->splice.flags & SPLICE_F_FD_IN_FIXED))
- io_put_file(req->splice.file_in);
- break;
case IORING_OP_OPENAT:
case IORING_OP_OPENAT2:
if (req->open.filename)
* forcing a sync submission from here, since we can't
* wait for request slots on the block side.
*/
- if (ret != -EAGAIN)
+ if (ret != -EAGAIN || !(req->ctx->flags & IORING_SETUP_IOPOLL))
break;
cond_resched();
} while (1);
static void io_req_task_link_timeout(struct io_kiocb *req, bool *locked)
{
struct io_kiocb *prev = req->timeout.prev;
- int ret;
+ int ret = -ENOENT;
if (prev) {
- ret = io_try_cancel_userdata(req, prev->user_data);
+ if (!(req->task->flags & PF_EXITING))
+ ret = io_try_cancel_userdata(req, prev->user_data);
io_req_complete_post(req, ret ?: -ETIME, 0);
io_put_req(prev);
} else {
/* when returns >0, the caller should retry */
static inline int io_cqring_wait_schedule(struct io_ring_ctx *ctx,
struct io_wait_queue *iowq,
- signed long *timeout)
+ ktime_t timeout)
{
int ret;
if (test_bit(0, &ctx->check_cq_overflow))
return 1;
- *timeout = schedule_timeout(*timeout);
- return !*timeout ? -ETIME : 1;
+ if (!schedule_hrtimeout(&timeout, HRTIMER_MODE_ABS))
+ return -ETIME;
+ return 1;
}
/*
{
struct io_wait_queue iowq;
struct io_rings *rings = ctx->rings;
- signed long timeout = MAX_SCHEDULE_TIMEOUT;
+ ktime_t timeout = KTIME_MAX;
int ret;
do {
if (get_timespec64(&ts, uts))
return -EFAULT;
- timeout = timespec64_to_jiffies(&ts);
+ timeout = ktime_add_ns(timespec64_to_ktime(ts), ktime_get_ns());
}
if (sig) {
}
prepare_to_wait_exclusive(&ctx->cq_wait, &iowq.wq,
TASK_INTERRUPTIBLE);
- ret = io_cqring_wait_schedule(ctx, &iowq, &timeout);
+ ret = io_cqring_wait_schedule(ctx, &iowq, timeout);
finish_wait(&ctx->cq_wait, &iowq.wq);
cond_resched();
} while (ret > 0);
struct io_ring_ctx *ctx = node->rsrc_data->ctx;
unsigned long flags;
bool first_add = false;
+ unsigned long delay = HZ;
spin_lock_irqsave(&ctx->rsrc_ref_lock, flags);
node->done = true;
+ /* if we are mid-quiesce then do not delay */
+ if (node->rsrc_data->quiesce)
+ delay = 0;
+
while (!list_empty(&ctx->rsrc_ref_list)) {
node = list_first_entry(&ctx->rsrc_ref_list,
struct io_rsrc_node, node);
spin_unlock_irqrestore(&ctx->rsrc_ref_lock, flags);
if (first_add)
- mod_delayed_work(system_wq, &ctx->rsrc_put_work, HZ);
+ mod_delayed_work(system_wq, &ctx->rsrc_put_work, delay);
}
static struct io_rsrc_node *io_rsrc_node_alloc(struct io_ring_ctx *ctx)
ret = wait_for_completion_interruptible(&data->done);
if (!ret) {
mutex_lock(&ctx->uring_lock);
- break;
+ if (atomic_read(&data->refs) > 0) {
+ /*
+ * it has been revived by another thread while
+ * we were unlocked
+ */
+ mutex_unlock(&ctx->uring_lock);
+ } else {
+ break;
+ }
}
atomic_inc(&data->refs);
static int io_sqe_files_unregister(struct io_ring_ctx *ctx)
{
+ unsigned nr = ctx->nr_user_files;
int ret;
if (!ctx->file_data)
return -ENXIO;
+
+ /*
+ * Quiesce may unlock ->uring_lock, and while it's not held
+ * prevent new requests using the table.
+ */
+ ctx->nr_user_files = 0;
ret = io_rsrc_ref_quiesce(ctx->file_data, ctx);
+ ctx->nr_user_files = nr;
if (!ret)
__io_sqe_files_unregister(ctx);
return ret;
}
skb->sk = sk;
+ skb->scm_io_uring = 1;
nr_files = 0;
fpl->user = get_uid(current_user());
refcount_add(skb->truesize, &sk->sk_wmem_alloc);
skb_queue_head(&sk->sk_receive_queue, skb);
- for (i = 0; i < nr_files; i++)
- fput(fpl->fp[i]);
+ for (i = 0; i < nr; i++) {
+ struct file *file = io_file_from_index(ctx, i + offset);
+
+ if (file)
+ fput(file);
+ }
} else {
kfree_skb(skb);
+ free_uid(fpl->user);
kfree(fpl);
}
io_ring_submit_lock(ctx, lock_ring);
spin_lock(&ctx->completion_lock);
- io_cqring_fill_event(ctx, prsrc->tag, 0, 0);
- ctx->cq_extra++;
+ io_fill_cqe_aux(ctx, prsrc->tag, 0, 0);
io_commit_cqring(ctx);
spin_unlock(&ctx->completion_lock);
io_cqring_ev_posted(ctx);
static int io_queue_rsrc_removal(struct io_rsrc_data *data, unsigned idx,
struct io_rsrc_node *node, void *rsrc)
{
+ u64 *tag_slot = io_get_tag_slot(data, idx);
struct io_rsrc_put *prsrc;
prsrc = kzalloc(sizeof(*prsrc), GFP_KERNEL);
if (!prsrc)
return -ENOMEM;
- prsrc->tag = *io_get_tag_slot(data, idx);
+ prsrc->tag = *tag_slot;
+ *tag_slot = 0;
prsrc->rsrc = rsrc;
list_add(&prsrc->list, &node->rsrc_list);
return 0;
struct io_ring_ctx *ctx = req->ctx;
struct io_fixed_file *file_slot;
struct file *file;
- int ret, i;
+ int ret;
io_ring_submit_lock(ctx, !(issue_flags & IO_URING_F_NONBLOCK));
ret = -ENXIO;
if (ret)
goto out;
- i = array_index_nospec(offset, ctx->nr_user_files);
- file_slot = io_fixed_file_slot(&ctx->file_table, i);
+ offset = array_index_nospec(offset, ctx->nr_user_files);
+ file_slot = io_fixed_file_slot(&ctx->file_table, offset);
ret = -EBADF;
if (!file_slot->file_ptr)
goto out;
if (file_slot->file_ptr) {
file = (struct file *)(file_slot->file_ptr & FFS_MASK);
- err = io_queue_rsrc_removal(data, up->offset + done,
- ctx->rsrc_node, file);
+ err = io_queue_rsrc_removal(data, i, ctx->rsrc_node, file);
if (err)
break;
file_slot->file_ptr = 0;
err = -EBADF;
break;
}
- *io_get_tag_slot(data, up->offset + done) = tag;
+ *io_get_tag_slot(data, i) = tag;
io_fixed_file_set(file_slot, file);
err = io_sqe_file_register(ctx, file, i);
if (err) {
static void *io_mem_alloc(size_t size)
{
- gfp_t gfp_flags = GFP_KERNEL | __GFP_ZERO | __GFP_NOWARN | __GFP_COMP |
- __GFP_NORETRY | __GFP_ACCOUNT;
+ gfp_t gfp = GFP_KERNEL_ACCOUNT | __GFP_ZERO | __GFP_NOWARN | __GFP_COMP;
- return (void *) __get_free_pages(gfp_flags, get_order(size));
+ return (void *) __get_free_pages(gfp, get_order(size));
}
static unsigned long rings_size(unsigned sq_entries, unsigned cq_entries,
static int io_sqe_buffers_unregister(struct io_ring_ctx *ctx)
{
+ unsigned nr = ctx->nr_user_bufs;
int ret;
if (!ctx->buf_data)
return -ENXIO;
+ /*
+ * Quiesce may unlock ->uring_lock, and while it's not held
+ * prevent new requests using the table.
+ */
+ ctx->nr_user_bufs = 0;
ret = io_rsrc_ref_quiesce(ctx->buf_data, ctx);
+ ctx->nr_user_bufs = nr;
if (!ret)
__io_sqe_buffers_unregister(ctx);
return ret;
i = array_index_nospec(offset, ctx->nr_user_bufs);
if (ctx->user_bufs[i] != ctx->dummy_ubuf) {
- err = io_queue_rsrc_removal(ctx->buf_data, offset,
+ err = io_queue_rsrc_removal(ctx->buf_data, i,
ctx->rsrc_node, ctx->user_bufs[i]);
if (unlikely(err)) {
io_buffer_unmap(ctx, &imu);
struct io_buffer *buf;
unsigned long index;
- xa_for_each(&ctx->io_buffers, index, buf) {
+ xa_for_each(&ctx->io_buffers, index, buf)
__io_remove_buffers(ctx, buf, index, -1U);
- cond_resched();
- }
}
static void io_req_cache_free(struct list_head *list)
{
io_sq_thread_finish(ctx);
- if (ctx->mm_account) {
- mmdrop(ctx->mm_account);
- ctx->mm_account = NULL;
- }
-
/* __io_rsrc_put_work() may need uring_lock to progress, wait w/o it */
io_wait_rsrc_data(ctx->buf_data);
io_wait_rsrc_data(ctx->file_data);
#endif
WARN_ON_ONCE(!list_empty(&ctx->ltimeout_list));
+ if (ctx->mm_account) {
+ mmdrop(ctx->mm_account);
+ ctx->mm_account = NULL;
+ }
+
io_mem_free(ctx->rings);
io_mem_free(ctx->sq_sqes);
{
struct io_kiocb *req = container_of(work, struct io_kiocb, work);
struct io_task_cancel *cancel = data;
- bool ret;
-
- if (!cancel->all && (req->flags & REQ_F_LINK_TIMEOUT)) {
- struct io_ring_ctx *ctx = req->ctx;
- /* protect against races with linked timeouts */
- spin_lock(&ctx->completion_lock);
- ret = io_match_task(req, cancel->task, cancel->all);
- spin_unlock(&ctx->completion_lock);
- } else {
- ret = io_match_task(req, cancel->task, cancel->all);
- }
- return ret;
+ return io_match_task_safe(req, cancel->task, cancel->all);
}
static bool io_cancel_defer_files(struct io_ring_ctx *ctx,
spin_lock(&ctx->completion_lock);
list_for_each_entry_reverse(de, &ctx->defer_list, list) {
- if (io_match_task(de->req, task, cancel_all)) {
+ if (io_match_task_safe(de->req, task, cancel_all)) {
list_cut_position(&list, &ctx->defer_list, &de->list);
break;
}
return percpu_counter_sum(&tctx->inflight);
}
-static void io_uring_drop_tctx_refs(struct task_struct *task)
-{
- struct io_uring_task *tctx = task->io_uring;
- unsigned int refs = tctx->cached_refs;
-
- if (refs) {
- tctx->cached_refs = 0;
- percpu_counter_sub(&tctx->inflight, refs);
- put_task_struct_many(task, refs);
- }
-}
-
/*
* Find any io_uring ctx that this task has registered or done IO on, and cancel
- * requests. @sqd should be not-null IIF it's an SQPOLL thread cancellation.
+ * requests. @sqd should be not-null IFF it's an SQPOLL thread cancellation.
*/
static void io_uring_cancel_generic(bool cancel_all, struct io_sq_data *sqd)
{
cancel_all);
}
- prepare_to_wait(&tctx->wait, &wait, TASK_UNINTERRUPTIBLE);
+ prepare_to_wait(&tctx->wait, &wait, TASK_INTERRUPTIBLE);
+ io_run_task_work();
io_uring_drop_tctx_refs(current);
+
/*
* If we've seen completions, retry without waiting. This
* avoids a race where a completion comes in before we did
schedule();
finish_wait(&tctx->wait, &wait);
} while (1);
- atomic_dec(&tctx->in_idle);
io_uring_clean_tctx(tctx);
if (cancel_all) {
+ /*
+ * We shouldn't run task_works after cancel, so just leave
+ * ->in_idle set for normal exit.
+ */
+ atomic_dec(&tctx->in_idle);
/* for exec all current's requests should be gone, kill tctx */
__io_uring_free(current);
}
return -EINVAL;
if (copy_from_user(&arg, argp, sizeof(arg)))
return -EFAULT;
+ if (arg.pad)
+ return -EINVAL;
*sig = u64_to_user_ptr(arg.sigmask);
*argsz = arg.sigmask_sz;
*ts = u64_to_user_ptr(arg.ts);
__u32 tmp;
int err;
- if (up->resv)
- return -EINVAL;
if (check_add_overflow(up->offset, nr_args, &tmp))
return -EOVERFLOW;
err = io_rsrc_node_switch_start(ctx);
memset(&up, 0, sizeof(up));
if (copy_from_user(&up, arg, sizeof(struct io_uring_rsrc_update)))
return -EFAULT;
+ if (up.resv || up.resv2)
+ return -EINVAL;
return __io_register_rsrc_update(ctx, IORING_RSRC_FILE, &up, nr_args);
}
return -EINVAL;
if (copy_from_user(&up, arg, sizeof(up)))
return -EFAULT;
- if (!up.nr || up.resv)
+ if (!up.nr || up.resv || up.resv2)
return -EINVAL;
return __io_register_rsrc_update(ctx, type, &up, up.nr);
}
if (len > cpumask_size())
len = cpumask_size();
- if (copy_from_user(new_mask, arg, len)) {
+ if (in_compat_syscall()) {
+ ret = compat_get_bitmap(cpumask_bits(new_mask),
+ (const compat_ulong_t __user *)arg,
+ len * 8 /* CHAR_BIT */);
+ } else {
+ ret = copy_from_user(new_mask, arg, len);
+ }
+
+ if (ret) {
free_cpumask_var(new_mask);
return -EFAULT;
}
BUILD_BUG_ON(sizeof(new_count) != sizeof(ctx->iowq_limits));
- memcpy(ctx->iowq_limits, new_count, sizeof(new_count));
+ for (i = 0; i < ARRAY_SIZE(new_count); i++)
+ if (new_count[i])
+ ctx->iowq_limits[i] = new_count[i];
ctx->iowq_limits_set = true;
ret = -EINVAL;