* before writing the tail (using smp_load_acquire to read the tail will
* do). It also needs a smp_mb() before updating CQ head (ordering the
* entry load(s) with the head store), pairing with an implicit barrier
- * through a control-dependency in io_get_cqring (smp_store_release to
+ * through a control-dependency in io_get_cqe (smp_store_release to
* store head will do). Failure to do so could lead to reading invalid
* CQ entries.
*
#include <linux/task_work.h>
#include <linux/pagemap.h>
#include <linux/io_uring.h>
+#include <linux/tracehook.h>
#define CREATE_TRACE_POINTS
#include <trace/events/io_uring.h>
#define IORING_MAX_ENTRIES 32768
#define IORING_MAX_CQ_ENTRIES (2 * IORING_MAX_ENTRIES)
+#define IORING_SQPOLL_CAP_ENTRIES_VALUE 8
-/*
- * Shift of 9 is 512 entries, or exactly one page on 64-bit archs
- */
-#define IORING_FILE_TABLE_SHIFT 9
-#define IORING_MAX_FILES_TABLE (1U << IORING_FILE_TABLE_SHIFT)
-#define IORING_FILE_TABLE_MASK (IORING_MAX_FILES_TABLE - 1)
-#define IORING_MAX_FIXED_FILES (64 * IORING_MAX_FILES_TABLE)
+/* only define max */
+#define IORING_MAX_FIXED_FILES (1U << 15)
#define IORING_MAX_RESTRICTIONS (IORING_RESTRICTION_LAST + \
IORING_REGISTER_LAST + IORING_OP_LAST)
+#define IO_RSRC_TAG_TABLE_SHIFT (PAGE_SHIFT - 3)
+#define IO_RSRC_TAG_TABLE_MAX (1U << IO_RSRC_TAG_TABLE_SHIFT)
+#define IO_RSRC_TAG_TABLE_MASK (IO_RSRC_TAG_TABLE_MAX - 1)
+
#define IORING_MAX_REG_BUFFERS (1U << 14)
#define SQE_VALID_FLAGS (IOSQE_FIXED_FILE|IOSQE_IO_DRAIN|IOSQE_IO_LINK| \
IOSQE_IO_HARDLINK | IOSQE_ASYNC | \
IOSQE_BUFFER_SELECT)
+#define IO_REQ_CLEAN_FLAGS (REQ_F_BUFFER_SELECTED | REQ_F_NEED_CLEANUP | \
+ REQ_F_POLLED | REQ_F_INFLIGHT | REQ_F_CREDS)
+
+#define IO_TCTX_REFS_CACHE_NR (1U << 10)
struct io_uring {
u32 head ____cacheline_aligned_in_smp;
* Written by the application, shouldn't be modified by the
* kernel.
*/
- u32 cq_flags;
+ u32 cq_flags;
/*
* Number of completion events lost because the queue was full;
* this should be avoided by the application by making sure
};
struct io_file_table {
- /* two level table */
- struct io_fixed_file **files;
+ struct io_fixed_file *files;
};
struct io_rsrc_node {
struct io_rsrc_data {
struct io_ring_ctx *ctx;
- u64 *tags;
+ u64 **tags;
+ unsigned int nr;
rsrc_put_fn *do_put;
atomic_t refs;
struct completion done;
unsigned long state;
struct completion exited;
- struct callback_head *park_task_work;
};
-#define IO_IOPOLL_BATCH 8
#define IO_COMPL_BATCH 32
#define IO_REQ_CACHE_SIZE 32
#define IO_REQ_ALLOC_BATCH 8
-struct io_comp_state {
- struct io_kiocb *reqs[IO_COMPL_BATCH];
- unsigned int nr;
- unsigned int locked_free_nr;
- /* inline/task_work completion list, under ->uring_lock */
- struct list_head free_list;
- /* IRQ completion list, under ->completion_lock */
- struct list_head locked_free_list;
-};
-
struct io_submit_link {
struct io_kiocb *head;
struct io_kiocb *last;
/*
* Batch completion logic
*/
- struct io_comp_state comp;
+ struct io_kiocb *compl_reqs[IO_COMPL_BATCH];
+ unsigned int compl_nr;
+ /* inline/task_work completion list, under ->uring_lock */
+ struct list_head free_list;
- /*
- * File reference cache
- */
- struct file *file;
- unsigned int fd;
- unsigned int file_refs;
unsigned int ios_left;
};
struct io_ring_ctx {
+ /* const or read-mostly hot data */
struct {
struct percpu_ref refs;
- } ____cacheline_aligned_in_smp;
- struct {
+ struct io_rings *rings;
unsigned int flags;
unsigned int compat: 1;
unsigned int drain_next: 1;
unsigned int eventfd_async: 1;
unsigned int restricted: 1;
+ unsigned int off_timeout_used: 1;
+ unsigned int drain_active: 1;
+ } ____cacheline_aligned_in_smp;
+
+ /* submission data */
+ struct {
+ struct mutex uring_lock;
/*
* Ring buffer of indices into array of io_uring_sqe, which is
* array.
*/
u32 *sq_array;
+ struct io_uring_sqe *sq_sqes;
unsigned cached_sq_head;
unsigned sq_entries;
- unsigned sq_mask;
- unsigned sq_thread_idle;
- unsigned cached_sq_dropped;
- unsigned cached_cq_overflow;
- unsigned long sq_check_overflow;
+ struct list_head defer_list;
- /* hashed buffered write serialization */
- struct io_wq_hash *hash_map;
+ /*
+ * Fixed resources fast path, should be accessed only under
+ * uring_lock, and updated through io_uring_register(2)
+ */
+ struct io_rsrc_node *rsrc_node;
+ struct io_file_table file_table;
+ unsigned nr_user_files;
+ unsigned nr_user_bufs;
+ struct io_mapped_ubuf **user_bufs;
- struct list_head defer_list;
+ struct io_submit_state submit_state;
struct list_head timeout_list;
+ struct list_head ltimeout_list;
struct list_head cq_overflow_list;
-
- struct io_uring_sqe *sq_sqes;
- } ____cacheline_aligned_in_smp;
-
- struct {
- struct mutex uring_lock;
- wait_queue_head_t wait;
+ struct xarray io_buffers;
+ struct xarray personalities;
+ u32 pers_next;
+ unsigned sq_thread_idle;
} ____cacheline_aligned_in_smp;
- struct io_submit_state submit_state;
-
- struct io_rings *rings;
-
- /* Only used for accounting purposes */
- struct mm_struct *mm_account;
+ /* IRQ completion list, under ->completion_lock */
+ struct list_head locked_free_list;
+ unsigned int locked_free_nr;
const struct cred *sq_creds; /* cred used for __io_sq_thread() */
struct io_sq_data *sq_data; /* if using sq thread polling */
struct wait_queue_head sqo_sq_wait;
struct list_head sqd_list;
- /*
- * If used, fixed file set. Writers must ensure that ->refs is dead,
- * readers must ensure that ->refs is alive as long as the file* is
- * used. Only updated through io_uring_register(2).
- */
- struct io_rsrc_data *file_data;
- struct io_file_table file_table;
- unsigned nr_user_files;
-
- /* if used, fixed mapped user buffers */
- struct io_rsrc_data *buf_data;
- unsigned nr_user_bufs;
- struct io_mapped_ubuf **user_bufs;
-
- struct user_struct *user;
-
- struct completion ref_comp;
-
-#if defined(CONFIG_UNIX)
- struct socket *ring_sock;
-#endif
-
- struct xarray io_buffers;
-
- struct xarray personalities;
- u32 pers_next;
+ unsigned long check_cq_overflow;
struct {
unsigned cached_cq_tail;
unsigned cq_entries;
- unsigned cq_mask;
- atomic_t cq_timeouts;
- unsigned cq_last_tm_flush;
- unsigned cq_extra;
- unsigned long cq_check_overflow;
+ struct eventfd_ctx *cq_ev_fd;
+ struct wait_queue_head poll_wait;
struct wait_queue_head cq_wait;
+ unsigned cq_extra;
+ atomic_t cq_timeouts;
struct fasync_struct *cq_fasync;
- struct eventfd_ctx *cq_ev_fd;
+ unsigned cq_last_tm_flush;
} ____cacheline_aligned_in_smp;
struct {
spinlock_t completion_lock;
+ spinlock_t timeout_lock;
+
/*
* ->iopoll_list is protected by the ctx->uring_lock for
* io_uring instances that don't use IORING_SETUP_SQPOLL.
struct list_head iopoll_list;
struct hlist_head *cancel_hash;
unsigned cancel_hash_bits;
- bool poll_multi_file;
+ bool poll_multi_queue;
} ____cacheline_aligned_in_smp;
- struct delayed_work rsrc_put_work;
- struct llist_head rsrc_put_llist;
- struct list_head rsrc_ref_list;
- spinlock_t rsrc_ref_lock;
- struct io_rsrc_node *rsrc_node;
- struct io_rsrc_node *rsrc_backup_node;
- struct io_mapped_ubuf *dummy_ubuf;
-
struct io_restriction restrictions;
- /* exit task_work */
- struct callback_head *exit_task_work;
+ /* slow path rsrc auxilary data, used by update/register */
+ struct {
+ struct io_rsrc_node *rsrc_backup_node;
+ struct io_mapped_ubuf *dummy_ubuf;
+ struct io_rsrc_data *file_data;
+ struct io_rsrc_data *buf_data;
+
+ struct delayed_work rsrc_put_work;
+ struct llist_head rsrc_put_llist;
+ struct list_head rsrc_ref_list;
+ spinlock_t rsrc_ref_lock;
+ };
/* Keep this last, we don't need it for the fast path */
- struct work_struct exit_work;
- struct list_head tctx_list;
+ struct {
+ #if defined(CONFIG_UNIX)
+ struct socket *ring_sock;
+ #endif
+ /* hashed buffered write serialization */
+ struct io_wq_hash *hash_map;
+
+ /* Only used for accounting purposes */
+ struct user_struct *user;
+ struct mm_struct *mm_account;
+
+ /* ctx exit and cancelation */
+ struct llist_head fallback_llist;
+ struct delayed_work fallback_work;
+ struct work_struct exit_work;
+ struct list_head tctx_list;
+ struct completion ref_comp;
+ };
};
struct io_uring_task {
/* submission side */
+ int cached_refs;
struct xarray xa;
struct wait_queue_head wait;
const struct io_ring_ctx *last;
spinlock_t task_lock;
struct io_wq_work_list task_list;
- unsigned long task_state;
struct callback_head task_work;
+ bool task_running;
};
/*
struct hrtimer timer;
struct timespec64 ts;
enum hrtimer_mode mode;
+ u32 flags;
};
struct io_accept {
struct sockaddr __user *addr;
int __user *addr_len;
int flags;
+ u32 file_slot;
unsigned long nofile;
};
struct list_head list;
/* head of the link, used by linked timeouts only */
struct io_kiocb *head;
+ /* for linked completions */
+ struct io_kiocb *prev;
};
struct io_timeout_rem {
/* timeout update */
struct timespec64 ts;
u32 flags;
+ bool ltimeout;
};
struct io_rw {
struct io_open {
struct file *file;
int dfd;
+ u32 file_slot;
struct filename *filename;
struct open_how how;
unsigned long nofile;
struct filename *filename;
};
+struct io_mkdir {
+ struct file *file;
+ int dfd;
+ umode_t mode;
+ struct filename *filename;
+};
+
+struct io_symlink {
+ struct file *file;
+ int new_dfd;
+ struct filename *oldpath;
+ struct filename *newpath;
+};
+
+struct io_hardlink {
+ struct file *file;
+ int old_dfd;
+ int new_dfd;
+ struct filename *oldpath;
+ struct filename *newpath;
+ int flags;
+};
+
struct io_completion {
struct file *file;
- struct list_head list;
u32 cflags;
};
REQ_F_BUFFER_SELECT_BIT = IOSQE_BUFFER_SELECT_BIT,
/* first byte is taken by user flags, shift it to not overlap */
- REQ_F_FAIL_LINK_BIT = 8,
+ REQ_F_FAIL_BIT = 8,
REQ_F_INFLIGHT_BIT,
REQ_F_CUR_POS_BIT,
REQ_F_NOWAIT_BIT,
REQ_F_NEED_CLEANUP_BIT,
REQ_F_POLLED_BIT,
REQ_F_BUFFER_SELECTED_BIT,
- REQ_F_LTIMEOUT_ACTIVE_BIT,
REQ_F_COMPLETE_INLINE_BIT,
REQ_F_REISSUE_BIT,
REQ_F_DONT_REISSUE_BIT,
+ REQ_F_CREDS_BIT,
+ REQ_F_REFCOUNT_BIT,
+ REQ_F_ARM_LTIMEOUT_BIT,
/* keep async read/write and isreg together and in order */
- REQ_F_ASYNC_READ_BIT,
- REQ_F_ASYNC_WRITE_BIT,
+ REQ_F_NOWAIT_READ_BIT,
+ REQ_F_NOWAIT_WRITE_BIT,
REQ_F_ISREG_BIT,
/* not a real bit, just to check we're not overflowing the space */
REQ_F_BUFFER_SELECT = BIT(REQ_F_BUFFER_SELECT_BIT),
/* fail rest of links */
- REQ_F_FAIL_LINK = BIT(REQ_F_FAIL_LINK_BIT),
+ REQ_F_FAIL = BIT(REQ_F_FAIL_BIT),
/* on inflight list, should be cancelled and waited on exit reliably */
REQ_F_INFLIGHT = BIT(REQ_F_INFLIGHT_BIT),
/* read/write uses file position */
REQ_F_POLLED = BIT(REQ_F_POLLED_BIT),
/* buffer already selected */
REQ_F_BUFFER_SELECTED = BIT(REQ_F_BUFFER_SELECTED_BIT),
- /* linked timeout is active, i.e. prepared by link's head */
- REQ_F_LTIMEOUT_ACTIVE = BIT(REQ_F_LTIMEOUT_ACTIVE_BIT),
/* completion is deferred through io_comp_state */
REQ_F_COMPLETE_INLINE = BIT(REQ_F_COMPLETE_INLINE_BIT),
/* caller should reissue async */
/* don't attempt request reissue, see io_rw_reissue() */
REQ_F_DONT_REISSUE = BIT(REQ_F_DONT_REISSUE_BIT),
/* supports async reads */
- REQ_F_ASYNC_READ = BIT(REQ_F_ASYNC_READ_BIT),
+ REQ_F_NOWAIT_READ = BIT(REQ_F_NOWAIT_READ_BIT),
/* supports async writes */
- REQ_F_ASYNC_WRITE = BIT(REQ_F_ASYNC_WRITE_BIT),
+ REQ_F_NOWAIT_WRITE = BIT(REQ_F_NOWAIT_WRITE_BIT),
/* regular file */
REQ_F_ISREG = BIT(REQ_F_ISREG_BIT),
+ /* has creds assigned */
+ REQ_F_CREDS = BIT(REQ_F_CREDS_BIT),
+ /* skip refcounting if not set */
+ REQ_F_REFCOUNT = BIT(REQ_F_REFCOUNT_BIT),
+ /* there is a linked timeout that has to be armed */
+ REQ_F_ARM_LTIMEOUT = BIT(REQ_F_ARM_LTIMEOUT_BIT),
};
struct async_poll {
struct io_poll_iocb *double_poll;
};
+typedef void (*io_req_tw_func_t)(struct io_kiocb *req, bool *locked);
+
struct io_task_work {
- struct io_wq_work_node node;
- task_work_func_t func;
+ union {
+ struct io_wq_work_node node;
+ struct llist_node fallback_node;
+ };
+ io_req_tw_func_t func;
};
enum {
struct io_shutdown shutdown;
struct io_rename rename;
struct io_unlink unlink;
+ struct io_mkdir mkdir;
+ struct io_symlink symlink;
+ struct io_hardlink hardlink;
/* use only after cleaning per-op data, see io_clean_op() */
struct io_completion compl;
};
/* used with ctx->iopoll_list with reads/writes */
struct list_head inflight_entry;
- union {
- struct io_task_work io_task_work;
- struct callback_head task_work;
- };
+ struct io_task_work io_task_work;
/* for polled requests, i.e. IORING_OP_POLL_ADD and async armed poll */
struct hlist_node hash_node;
struct async_poll *apoll;
struct io_wq_work work;
+ const struct cred *creds;
+
/* store used ubuf, so we can prevent reloading */
struct io_mapped_ubuf *imu;
};
},
[IORING_OP_WRITE] = {
.needs_file = 1,
+ .hash_reg_file = 1,
.unbound_nonreg_file = 1,
.pollout = 1,
.plug = 1,
},
[IORING_OP_RENAMEAT] = {},
[IORING_OP_UNLINKAT] = {},
+ [IORING_OP_MKDIRAT] = {},
+ [IORING_OP_SYMLINKAT] = {},
+ [IORING_OP_LINKAT] = {},
};
+/* requests with any of those set should undergo io_disarm_next() */
+#define IO_DISARM_MASK (REQ_F_ARM_LTIMEOUT | REQ_F_LINK_TIMEOUT | REQ_F_FAIL)
+
static bool io_disarm_next(struct io_kiocb *req);
-static void io_uring_del_task_file(unsigned long index);
+static void io_uring_del_tctx_node(unsigned long index);
static void io_uring_try_cancel_requests(struct io_ring_ctx *ctx,
struct task_struct *task,
- struct files_struct *files);
-static void io_uring_cancel_sqpoll(struct io_sq_data *sqd);
-static struct io_rsrc_node *io_rsrc_node_alloc(struct io_ring_ctx *ctx);
+ 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_put_req(struct io_kiocb *req);
-static void io_put_req_deferred(struct io_kiocb *req, int nr);
+static void io_put_req_deferred(struct io_kiocb *req);
static void io_dismantle_req(struct io_kiocb *req);
-static void io_put_task(struct task_struct *task, int nr);
-static struct io_kiocb *io_prep_linked_timeout(struct io_kiocb *req);
static void io_queue_linked_timeout(struct io_kiocb *req);
static int __io_register_rsrc_update(struct io_ring_ctx *ctx, unsigned type,
struct io_uring_rsrc_update2 *up,
unsigned nr_args);
static void io_clean_op(struct io_kiocb *req);
-static struct file *io_file_get(struct io_submit_state *state,
+static struct file *io_file_get(struct io_ring_ctx *ctx,
struct io_kiocb *req, int fd, bool fixed);
static void __io_queue_sqe(struct io_kiocb *req);
static void io_rsrc_put_work(struct work_struct *work);
static void io_req_task_queue(struct io_kiocb *req);
-static void io_submit_flush_completions(struct io_comp_state *cs,
- struct io_ring_ctx *ctx);
-static bool io_poll_remove_waitqs(struct io_kiocb *req);
+static void io_submit_flush_completions(struct io_ring_ctx *ctx);
static int io_req_prep_async(struct io_kiocb *req);
+static int io_install_fixed_file(struct io_kiocb *req, struct file *file,
+ unsigned int issue_flags, u32 slot_index);
+static enum hrtimer_restart io_link_timeout_fn(struct hrtimer *timer);
+
static struct kmem_cache *req_cachep;
static const struct file_operations io_uring_fops;
}
EXPORT_SYMBOL(io_uring_get_socket);
+static inline void io_tw_lock(struct io_ring_ctx *ctx, bool *locked)
+{
+ if (!*locked) {
+ mutex_lock(&ctx->uring_lock);
+ *locked = true;
+ }
+}
+
#define io_for_each_link(pos, head) \
for (pos = (head); pos; pos = pos->link)
+/*
+ * Shamelessly stolen from the mm implementation of page reference checking,
+ * see commit f958d7b528b1 for details.
+ */
+#define req_ref_zero_or_close_to_overflow(req) \
+ ((unsigned int) atomic_read(&(req->refs)) + 127u <= 127u)
+
+static inline bool req_ref_inc_not_zero(struct io_kiocb *req)
+{
+ WARN_ON_ONCE(!(req->flags & REQ_F_REFCOUNT));
+ return atomic_inc_not_zero(&req->refs);
+}
+
+static inline bool req_ref_put_and_test(struct io_kiocb *req)
+{
+ if (likely(!(req->flags & REQ_F_REFCOUNT)))
+ return true;
+
+ WARN_ON_ONCE(req_ref_zero_or_close_to_overflow(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));
+ WARN_ON_ONCE(req_ref_zero_or_close_to_overflow(req));
+ atomic_inc(&req->refs);
+}
+
+static inline void __io_req_set_refcount(struct io_kiocb *req, int nr)
+{
+ if (!(req->flags & REQ_F_REFCOUNT)) {
+ req->flags |= REQ_F_REFCOUNT;
+ atomic_set(&req->refs, nr);
+ }
+}
+
+static inline void io_req_set_refcount(struct io_kiocb *req)
+{
+ __io_req_set_refcount(req, 1);
+}
+
static inline void io_req_set_rsrc_node(struct io_kiocb *req)
{
struct io_ring_ctx *ctx = req->ctx;
percpu_ref_put(ref);
}
-static bool io_match_task(struct io_kiocb *head,
- struct task_struct *task,
- struct files_struct *files)
+static bool io_match_task(struct io_kiocb *head, struct task_struct *task,
+ bool cancel_all)
{
struct io_kiocb *req;
if (task && head->task != task)
return false;
- if (!files)
+ if (cancel_all)
return true;
io_for_each_link(req, head) {
return false;
}
-static inline void req_set_fail_links(struct io_kiocb *req)
+static inline void req_set_fail(struct io_kiocb *req)
+{
+ req->flags |= REQ_F_FAIL;
+}
+
+static inline void req_fail_link_node(struct io_kiocb *req, int res)
{
- if (req->flags & REQ_F_LINK)
- req->flags |= REQ_F_FAIL_LINK;
+ req_set_fail(req);
+ req->result = res;
}
static void io_ring_ctx_ref_free(struct percpu_ref *ref)
return !req->timeout.off;
}
+static void io_fallback_req_func(struct work_struct *work)
+{
+ struct io_ring_ctx *ctx = container_of(work, struct io_ring_ctx,
+ fallback_work.work);
+ struct llist_node *node = llist_del_all(&ctx->fallback_llist);
+ struct io_kiocb *req, *tmp;
+ bool locked = false;
+
+ percpu_ref_get(&ctx->refs);
+ llist_for_each_entry_safe(req, tmp, node, io_task_work.fallback_node)
+ req->io_task_work.func(req, &locked);
+
+ if (locked) {
+ if (ctx->submit_state.compl_nr)
+ io_submit_flush_completions(ctx);
+ mutex_unlock(&ctx->uring_lock);
+ }
+ percpu_ref_put(&ctx->refs);
+
+}
+
static struct io_ring_ctx *io_ring_ctx_alloc(struct io_uring_params *p)
{
struct io_ring_ctx *ctx;
ctx->flags = p->flags;
init_waitqueue_head(&ctx->sqo_sq_wait);
INIT_LIST_HEAD(&ctx->sqd_list);
- init_waitqueue_head(&ctx->cq_wait);
+ init_waitqueue_head(&ctx->poll_wait);
INIT_LIST_HEAD(&ctx->cq_overflow_list);
init_completion(&ctx->ref_comp);
xa_init_flags(&ctx->io_buffers, XA_FLAGS_ALLOC1);
xa_init_flags(&ctx->personalities, XA_FLAGS_ALLOC1);
mutex_init(&ctx->uring_lock);
- init_waitqueue_head(&ctx->wait);
+ init_waitqueue_head(&ctx->cq_wait);
spin_lock_init(&ctx->completion_lock);
+ spin_lock_init(&ctx->timeout_lock);
INIT_LIST_HEAD(&ctx->iopoll_list);
INIT_LIST_HEAD(&ctx->defer_list);
INIT_LIST_HEAD(&ctx->timeout_list);
+ INIT_LIST_HEAD(&ctx->ltimeout_list);
spin_lock_init(&ctx->rsrc_ref_lock);
INIT_LIST_HEAD(&ctx->rsrc_ref_list);
INIT_DELAYED_WORK(&ctx->rsrc_put_work, io_rsrc_put_work);
init_llist_head(&ctx->rsrc_put_llist);
INIT_LIST_HEAD(&ctx->tctx_list);
- INIT_LIST_HEAD(&ctx->submit_state.comp.free_list);
- INIT_LIST_HEAD(&ctx->submit_state.comp.locked_free_list);
+ INIT_LIST_HEAD(&ctx->submit_state.free_list);
+ INIT_LIST_HEAD(&ctx->locked_free_list);
+ INIT_DELAYED_WORK(&ctx->fallback_work, io_fallback_req_func);
return ctx;
err:
kfree(ctx->dummy_ubuf);
return NULL;
}
+static void io_account_cq_overflow(struct io_ring_ctx *ctx)
+{
+ struct io_rings *r = ctx->rings;
+
+ WRITE_ONCE(r->cq_overflow, READ_ONCE(r->cq_overflow) + 1);
+ ctx->cq_extra--;
+}
+
static bool req_need_defer(struct io_kiocb *req, u32 seq)
{
if (unlikely(req->flags & REQ_F_IO_DRAIN)) {
struct io_ring_ctx *ctx = req->ctx;
- return seq + ctx->cq_extra != ctx->cached_cq_tail
- + READ_ONCE(ctx->cached_cq_overflow);
+ return seq + READ_ONCE(ctx->cq_extra) != ctx->cached_cq_tail;
}
return false;
}
+#define FFS_ASYNC_READ 0x1UL
+#define FFS_ASYNC_WRITE 0x2UL
+#ifdef CONFIG_64BIT
+#define FFS_ISREG 0x4UL
+#else
+#define FFS_ISREG 0x0UL
+#endif
+#define FFS_MASK ~(FFS_ASYNC_READ|FFS_ASYNC_WRITE|FFS_ISREG)
+
+static inline bool io_req_ffs_set(struct io_kiocb *req)
+{
+ return IS_ENABLED(CONFIG_64BIT) && (req->flags & REQ_F_FIXED_FILE);
+}
+
static void io_req_track_inflight(struct io_kiocb *req)
{
if (!(req->flags & REQ_F_INFLIGHT)) {
}
}
+static inline void io_unprep_linked_timeout(struct io_kiocb *req)
+{
+ req->flags &= ~REQ_F_LINK_TIMEOUT;
+}
+
+static struct io_kiocb *__io_prep_linked_timeout(struct io_kiocb *req)
+{
+ if (WARN_ON_ONCE(!req->link))
+ return NULL;
+
+ req->flags &= ~REQ_F_ARM_LTIMEOUT;
+ req->flags |= REQ_F_LINK_TIMEOUT;
+
+ /* linked timeouts should have two refs once prep'ed */
+ io_req_set_refcount(req);
+ __io_req_set_refcount(req->link, 2);
+ return req->link;
+}
+
+static inline struct io_kiocb *io_prep_linked_timeout(struct io_kiocb *req)
+{
+ if (likely(!(req->flags & REQ_F_ARM_LTIMEOUT)))
+ return NULL;
+ return __io_prep_linked_timeout(req);
+}
+
static void io_prep_async_work(struct io_kiocb *req)
{
const struct io_op_def *def = &io_op_defs[req->opcode];
struct io_ring_ctx *ctx = req->ctx;
- if (!req->work.creds)
- req->work.creds = get_current_cred();
+ if (!(req->flags & REQ_F_CREDS)) {
+ req->flags |= REQ_F_CREDS;
+ req->creds = get_current_cred();
+ }
req->work.list.next = NULL;
req->work.flags = 0;
{
struct io_kiocb *cur;
- io_for_each_link(cur, req)
- io_prep_async_work(cur);
+ if (req->flags & REQ_F_LINK_TIMEOUT) {
+ struct io_ring_ctx *ctx = req->ctx;
+
+ spin_lock(&ctx->completion_lock);
+ io_for_each_link(cur, req)
+ io_prep_async_work(cur);
+ spin_unlock(&ctx->completion_lock);
+ } else {
+ io_for_each_link(cur, req)
+ io_prep_async_work(cur);
+ }
}
-static void io_queue_async_work(struct io_kiocb *req)
+static void io_queue_async_work(struct io_kiocb *req, bool *locked)
{
struct io_ring_ctx *ctx = req->ctx;
struct io_kiocb *link = io_prep_linked_timeout(req);
struct io_uring_task *tctx = req->task->io_uring;
+ /* must not take the lock, NULL it as a precaution */
+ locked = NULL;
+
BUG_ON(!tctx);
BUG_ON(!tctx->io_wq);
/* init ->work of the whole link before punting */
io_prep_async_link(req);
+
+ /*
+ * Not expected to happen, but if we do have a bug where this _can_
+ * happen, catch it here and ensure the request is marked as
+ * canceled. That will make io-wq go through the usual work cancel
+ * procedure rather than attempt to run this request (or create a new
+ * worker for it).
+ */
+ if (WARN_ON_ONCE(!same_thread_group(req->task, current)))
+ req->work.flags |= IO_WQ_WORK_CANCEL;
+
trace_io_uring_queue_async_work(ctx, io_wq_is_hashed(&req->work), req,
&req->work, req->flags);
io_wq_enqueue(tctx->io_wq, &req->work);
static void io_kill_timeout(struct io_kiocb *req, int status)
__must_hold(&req->ctx->completion_lock)
+ __must_hold(&req->ctx->timeout_lock)
{
struct io_timeout_data *io = req->async_data;
if (hrtimer_try_to_cancel(&io->timer) != -1) {
+ if (status)
+ req_set_fail(req);
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_put_req_deferred(req, 1);
+ io_put_req_deferred(req);
}
}
-static void __io_queue_deferred(struct io_ring_ctx *ctx)
+static void io_queue_deferred(struct io_ring_ctx *ctx)
{
- do {
+ while (!list_empty(&ctx->defer_list)) {
struct io_defer_entry *de = list_first_entry(&ctx->defer_list,
struct io_defer_entry, list);
list_del_init(&de->list);
io_req_task_queue(de->req);
kfree(de);
- } while (!list_empty(&ctx->defer_list));
+ }
}
static void io_flush_timeouts(struct io_ring_ctx *ctx)
+ __must_hold(&ctx->completion_lock)
{
- u32 seq;
-
- if (list_empty(&ctx->timeout_list))
- return;
-
- seq = ctx->cached_cq_tail - atomic_read(&ctx->cq_timeouts);
+ u32 seq = ctx->cached_cq_tail - atomic_read(&ctx->cq_timeouts);
- do {
+ spin_lock_irq(&ctx->timeout_lock);
+ while (!list_empty(&ctx->timeout_list)) {
u32 events_needed, events_got;
struct io_kiocb *req = list_first_entry(&ctx->timeout_list,
struct io_kiocb, timeout.list);
list_del_init(&req->timeout.list);
io_kill_timeout(req, 0);
- } while (!list_empty(&ctx->timeout_list));
-
+ }
ctx->cq_last_tm_flush = seq;
+ spin_unlock_irq(&ctx->timeout_lock);
}
-static void io_commit_cqring(struct io_ring_ctx *ctx)
+static void __io_commit_cqring_flush(struct io_ring_ctx *ctx)
{
- io_flush_timeouts(ctx);
+ if (ctx->off_timeout_used)
+ io_flush_timeouts(ctx);
+ if (ctx->drain_active)
+ io_queue_deferred(ctx);
+}
+static inline void io_commit_cqring(struct io_ring_ctx *ctx)
+{
+ if (unlikely(ctx->off_timeout_used || ctx->drain_active))
+ __io_commit_cqring_flush(ctx);
/* order cqe stores with ring update */
smp_store_release(&ctx->rings->cq.tail, ctx->cached_cq_tail);
-
- if (unlikely(!list_empty(&ctx->defer_list)))
- __io_queue_deferred(ctx);
}
static inline bool io_sqring_full(struct io_ring_ctx *ctx)
{
struct io_rings *r = ctx->rings;
- return READ_ONCE(r->sq.tail) - ctx->cached_sq_head == r->sq_ring_entries;
+ return READ_ONCE(r->sq.tail) - ctx->cached_sq_head == ctx->sq_entries;
}
static inline unsigned int __io_cqring_events(struct io_ring_ctx *ctx)
return ctx->cached_cq_tail - READ_ONCE(ctx->rings->cq.head);
}
-static inline struct io_uring_cqe *io_get_cqring(struct io_ring_ctx *ctx)
+static inline struct io_uring_cqe *io_get_cqe(struct io_ring_ctx *ctx)
{
struct io_rings *rings = ctx->rings;
- unsigned tail;
+ unsigned tail, mask = ctx->cq_entries - 1;
/*
* writes to the cq entry need to come after reading head; the
* control dependency is enough as we're using WRITE_ONCE to
* fill the cq entry
*/
- if (__io_cqring_events(ctx) == rings->cq_ring_entries)
+ if (__io_cqring_events(ctx) == ctx->cq_entries)
return NULL;
tail = ctx->cached_cq_tail++;
- return &rings->cqes[tail & ctx->cq_mask];
+ return &rings->cqes[tail & mask];
}
static inline bool io_should_trigger_evfd(struct io_ring_ctx *ctx)
return !ctx->eventfd_async || io_wq_current_is_worker();
}
+/*
+ * This should only get called when at least one event has been posted.
+ * Some applications rely on the eventfd notification count only changing
+ * IFF a new CQE has been added to the CQ ring. There's no depedency on
+ * 1:1 relationship between how many times this function is called (and
+ * hence the eventfd count) and number of CQEs posted to the CQ ring.
+ */
static void io_cqring_ev_posted(struct io_ring_ctx *ctx)
{
- /* see waitqueue_active() comment */
- smp_mb();
-
- if (waitqueue_active(&ctx->wait))
- wake_up(&ctx->wait);
+ /*
+ * wake_up_all() may seem excessive, but io_wake_function() and
+ * io_should_wake() handle the termination of the loop and only
+ * wake as many waiters as we need to.
+ */
+ if (wq_has_sleeper(&ctx->cq_wait))
+ wake_up_all(&ctx->cq_wait);
if (ctx->sq_data && waitqueue_active(&ctx->sq_data->wait))
wake_up(&ctx->sq_data->wait);
if (io_should_trigger_evfd(ctx))
eventfd_signal(ctx->cq_ev_fd, 1);
- if (waitqueue_active(&ctx->cq_wait)) {
- wake_up_interruptible(&ctx->cq_wait);
+ if (waitqueue_active(&ctx->poll_wait)) {
+ wake_up_interruptible(&ctx->poll_wait);
kill_fasync(&ctx->cq_fasync, SIGIO, POLL_IN);
}
}
smp_mb();
if (ctx->flags & IORING_SETUP_SQPOLL) {
- if (waitqueue_active(&ctx->wait))
- wake_up(&ctx->wait);
+ if (waitqueue_active(&ctx->cq_wait))
+ wake_up_all(&ctx->cq_wait);
}
if (io_should_trigger_evfd(ctx))
eventfd_signal(ctx->cq_ev_fd, 1);
- if (waitqueue_active(&ctx->cq_wait)) {
- wake_up_interruptible(&ctx->cq_wait);
+ if (waitqueue_active(&ctx->poll_wait)) {
+ wake_up_interruptible(&ctx->poll_wait);
kill_fasync(&ctx->cq_fasync, SIGIO, POLL_IN);
}
}
/* Returns true if there are no backlogged entries after the flush */
static bool __io_cqring_overflow_flush(struct io_ring_ctx *ctx, bool force)
{
- struct io_rings *rings = ctx->rings;
- unsigned long flags;
bool all_flushed, posted;
- if (!force && __io_cqring_events(ctx) == rings->cq_ring_entries)
+ if (!force && __io_cqring_events(ctx) == ctx->cq_entries)
return false;
posted = false;
- spin_lock_irqsave(&ctx->completion_lock, flags);
+ spin_lock(&ctx->completion_lock);
while (!list_empty(&ctx->cq_overflow_list)) {
- struct io_uring_cqe *cqe = io_get_cqring(ctx);
+ struct io_uring_cqe *cqe = io_get_cqe(ctx);
struct io_overflow_cqe *ocqe;
if (!cqe && !force)
if (cqe)
memcpy(cqe, &ocqe->cqe, sizeof(*cqe));
else
- WRITE_ONCE(ctx->rings->cq_overflow,
- ++ctx->cached_cq_overflow);
+ io_account_cq_overflow(ctx);
+
posted = true;
list_del(&ocqe->list);
kfree(ocqe);
all_flushed = list_empty(&ctx->cq_overflow_list);
if (all_flushed) {
- clear_bit(0, &ctx->sq_check_overflow);
- clear_bit(0, &ctx->cq_check_overflow);
- ctx->rings->sq_flags &= ~IORING_SQ_CQ_OVERFLOW;
+ clear_bit(0, &ctx->check_cq_overflow);
+ WRITE_ONCE(ctx->rings->sq_flags,
+ ctx->rings->sq_flags & ~IORING_SQ_CQ_OVERFLOW);
}
if (posted)
io_commit_cqring(ctx);
- spin_unlock_irqrestore(&ctx->completion_lock, flags);
+ spin_unlock(&ctx->completion_lock);
if (posted)
io_cqring_ev_posted(ctx);
return all_flushed;
}
-static bool io_cqring_overflow_flush(struct io_ring_ctx *ctx, bool force)
+static bool io_cqring_overflow_flush(struct io_ring_ctx *ctx)
{
bool ret = true;
- if (test_bit(0, &ctx->cq_check_overflow)) {
+ if (test_bit(0, &ctx->check_cq_overflow)) {
/* iopoll syncs against uring_lock, not completion_lock */
if (ctx->flags & IORING_SETUP_IOPOLL)
mutex_lock(&ctx->uring_lock);
- ret = __io_cqring_overflow_flush(ctx, force);
+ ret = __io_cqring_overflow_flush(ctx, false);
if (ctx->flags & IORING_SETUP_IOPOLL)
mutex_unlock(&ctx->uring_lock);
}
return ret;
}
-/*
- * Shamelessly stolen from the mm implementation of page reference checking,
- * see commit f958d7b528b1 for details.
- */
-#define req_ref_zero_or_close_to_overflow(req) \
- ((unsigned int) atomic_read(&(req->refs)) + 127u <= 127u)
-
-static inline bool req_ref_inc_not_zero(struct io_kiocb *req)
+/* must to be called somewhat shortly after putting a request */
+static inline void io_put_task(struct task_struct *task, int nr)
{
- return atomic_inc_not_zero(&req->refs);
-}
+ struct io_uring_task *tctx = task->io_uring;
-static inline bool req_ref_sub_and_test(struct io_kiocb *req, int refs)
-{
- WARN_ON_ONCE(req_ref_zero_or_close_to_overflow(req));
- return atomic_sub_and_test(refs, &req->refs);
+ if (likely(task == current)) {
+ tctx->cached_refs += nr;
+ } else {
+ percpu_counter_sub(&tctx->inflight, nr);
+ if (unlikely(atomic_read(&tctx->in_idle)))
+ wake_up(&tctx->wait);
+ put_task_struct_many(task, nr);
+ }
}
-static inline bool req_ref_put_and_test(struct io_kiocb *req)
+static void io_task_refs_refill(struct io_uring_task *tctx)
{
- WARN_ON_ONCE(req_ref_zero_or_close_to_overflow(req));
- return atomic_dec_and_test(&req->refs);
-}
+ unsigned int refill = -tctx->cached_refs + IO_TCTX_REFS_CACHE_NR;
-static inline void req_ref_put(struct io_kiocb *req)
-{
- WARN_ON_ONCE(req_ref_put_and_test(req));
+ percpu_counter_add(&tctx->inflight, refill);
+ refcount_add(refill, ¤t->usage);
+ tctx->cached_refs += refill;
}
-static inline void req_ref_get(struct io_kiocb *req)
+static inline void io_get_task_refs(int nr)
{
- WARN_ON_ONCE(req_ref_zero_or_close_to_overflow(req));
- atomic_inc(&req->refs);
+ struct io_uring_task *tctx = current->io_uring;
+
+ tctx->cached_refs -= nr;
+ if (unlikely(tctx->cached_refs < 0))
+ io_task_refs_refill(tctx);
}
static bool io_cqring_event_overflow(struct io_ring_ctx *ctx, u64 user_data,
* or cannot allocate an overflow entry, then we need to drop it
* on the floor.
*/
- WRITE_ONCE(ctx->rings->cq_overflow, ++ctx->cached_cq_overflow);
+ io_account_cq_overflow(ctx);
return false;
}
if (list_empty(&ctx->cq_overflow_list)) {
- set_bit(0, &ctx->sq_check_overflow);
- set_bit(0, &ctx->cq_check_overflow);
- ctx->rings->sq_flags |= IORING_SQ_CQ_OVERFLOW;
+ set_bit(0, &ctx->check_cq_overflow);
+ WRITE_ONCE(ctx->rings->sq_flags,
+ ctx->rings->sq_flags | IORING_SQ_CQ_OVERFLOW);
+
}
ocqe->cqe.user_data = user_data;
ocqe->cqe.res = res;
* submission (by quite a lot). Increment the overflow count in
* the ring.
*/
- cqe = io_get_cqring(ctx);
+ cqe = io_get_cqe(ctx);
if (likely(cqe)) {
WRITE_ONCE(cqe->user_data, user_data);
WRITE_ONCE(cqe->res, res);
unsigned int cflags)
{
struct io_ring_ctx *ctx = req->ctx;
- unsigned long flags;
- spin_lock_irqsave(&ctx->completion_lock, flags);
+ spin_lock(&ctx->completion_lock);
__io_cqring_fill_event(ctx, req->user_data, res, cflags);
/*
* If we're the last reference to this request, add to our locked
* free_list cache.
*/
if (req_ref_put_and_test(req)) {
- struct io_comp_state *cs = &ctx->submit_state.comp;
-
if (req->flags & (REQ_F_LINK | REQ_F_HARDLINK)) {
- if (req->flags & (REQ_F_LINK_TIMEOUT | REQ_F_FAIL_LINK))
+ if (req->flags & IO_DISARM_MASK)
io_disarm_next(req);
if (req->link) {
io_req_task_queue(req->link);
}
io_dismantle_req(req);
io_put_task(req->task, 1);
- list_add(&req->compl.list, &cs->locked_free_list);
- cs->locked_free_nr++;
+ list_add(&req->inflight_entry, &ctx->locked_free_list);
+ ctx->locked_free_nr++;
} else {
if (!percpu_ref_tryget(&ctx->refs))
req = NULL;
}
io_commit_cqring(ctx);
- spin_unlock_irqrestore(&ctx->completion_lock, flags);
+ spin_unlock(&ctx->completion_lock);
if (req) {
io_cqring_ev_posted(ctx);
static inline bool io_req_needs_clean(struct io_kiocb *req)
{
- return req->flags & (REQ_F_BUFFER_SELECTED | REQ_F_NEED_CLEANUP |
- REQ_F_POLLED | REQ_F_INFLIGHT);
+ return req->flags & IO_REQ_CLEAN_FLAGS;
}
static void io_req_complete_state(struct io_kiocb *req, long res,
static void io_req_complete_failed(struct io_kiocb *req, long res)
{
- req_set_fail_links(req);
- io_put_req(req);
+ req_set_fail(req);
io_req_complete_post(req, res, 0);
}
-static void io_flush_cached_locked_reqs(struct io_ring_ctx *ctx,
- struct io_comp_state *cs)
+static void io_req_complete_fail_submit(struct io_kiocb *req)
{
- spin_lock_irq(&ctx->completion_lock);
- list_splice_init(&cs->locked_free_list, &cs->free_list);
- cs->locked_free_nr = 0;
- spin_unlock_irq(&ctx->completion_lock);
+ /*
+ * We don't submit, fail them all, for that replace hardlinks with
+ * normal links. Extra REQ_F_LINK is tolerated.
+ */
+ req->flags &= ~REQ_F_HARDLINK;
+ req->flags |= REQ_F_LINK;
+ io_req_complete_failed(req, req->result);
}
-/* Returns true IFF there are requests in the cache */
+/*
+ * Don't initialise the fields below on every allocation, but do that in
+ * advance and keep them valid across allocations.
+ */
+static void io_preinit_req(struct io_kiocb *req, struct io_ring_ctx *ctx)
+{
+ req->ctx = ctx;
+ req->link = NULL;
+ req->async_data = NULL;
+ /* not necessary, but safer to zero */
+ req->result = 0;
+}
+
+static void io_flush_cached_locked_reqs(struct io_ring_ctx *ctx,
+ struct io_submit_state *state)
+{
+ spin_lock(&ctx->completion_lock);
+ list_splice_init(&ctx->locked_free_list, &state->free_list);
+ ctx->locked_free_nr = 0;
+ spin_unlock(&ctx->completion_lock);
+}
+
+/* Returns true IFF there are requests in the cache */
static bool io_flush_cached_reqs(struct io_ring_ctx *ctx)
{
struct io_submit_state *state = &ctx->submit_state;
- struct io_comp_state *cs = &state->comp;
int nr;
/*
* locked cache, grab the lock and move them over to our submission
* side cache.
*/
- if (READ_ONCE(cs->locked_free_nr) > IO_COMPL_BATCH)
- io_flush_cached_locked_reqs(ctx, cs);
+ if (READ_ONCE(ctx->locked_free_nr) > IO_COMPL_BATCH)
+ io_flush_cached_locked_reqs(ctx, state);
nr = state->free_reqs;
- while (!list_empty(&cs->free_list)) {
- struct io_kiocb *req = list_first_entry(&cs->free_list,
- struct io_kiocb, compl.list);
+ while (!list_empty(&state->free_list)) {
+ struct io_kiocb *req = list_first_entry(&state->free_list,
+ struct io_kiocb, inflight_entry);
- list_del(&req->compl.list);
+ list_del(&req->inflight_entry);
state->reqs[nr++] = req;
if (nr == ARRAY_SIZE(state->reqs))
break;
return nr != 0;
}
+/*
+ * A request might get retired back into the request caches even before opcode
+ * handlers and io_issue_sqe() are done with it, e.g. inline completion path.
+ * Because of that, io_alloc_req() should be called only under ->uring_lock
+ * and with extra caution to not get a request that is still worked on.
+ */
static struct io_kiocb *io_alloc_req(struct io_ring_ctx *ctx)
+ __must_hold(&ctx->uring_lock)
{
struct io_submit_state *state = &ctx->submit_state;
+ gfp_t gfp = GFP_KERNEL | __GFP_NOWARN;
+ int ret, i;
- BUILD_BUG_ON(IO_REQ_ALLOC_BATCH > ARRAY_SIZE(state->reqs));
-
- if (!state->free_reqs) {
- gfp_t gfp = GFP_KERNEL | __GFP_NOWARN;
- int ret;
+ BUILD_BUG_ON(ARRAY_SIZE(state->reqs) < IO_REQ_ALLOC_BATCH);
- if (io_flush_cached_reqs(ctx))
- goto got_req;
+ if (likely(state->free_reqs || io_flush_cached_reqs(ctx)))
+ goto got_req;
- ret = kmem_cache_alloc_bulk(req_cachep, gfp, IO_REQ_ALLOC_BATCH,
- state->reqs);
+ ret = kmem_cache_alloc_bulk(req_cachep, gfp, IO_REQ_ALLOC_BATCH,
+ state->reqs);
- /*
- * Bulk alloc is all-or-nothing. If we fail to get a batch,
- * retry single alloc to be on the safe side.
- */
- if (unlikely(ret <= 0)) {
- state->reqs[0] = kmem_cache_alloc(req_cachep, gfp);
- if (!state->reqs[0])
- return NULL;
- ret = 1;
- }
- state->free_reqs = ret;
+ /*
+ * Bulk alloc is all-or-nothing. If we fail to get a batch,
+ * retry single alloc to be on the safe side.
+ */
+ if (unlikely(ret <= 0)) {
+ state->reqs[0] = kmem_cache_alloc(req_cachep, gfp);
+ if (!state->reqs[0])
+ return NULL;
+ ret = 1;
}
+
+ for (i = 0; i < ret; i++)
+ io_preinit_req(state->reqs[i], ctx);
+ state->free_reqs = ret;
got_req:
state->free_reqs--;
return state->reqs[state->free_reqs];
io_put_file(req->file);
if (req->fixed_rsrc_refs)
percpu_ref_put(req->fixed_rsrc_refs);
- if (req->async_data)
+ if (req->async_data) {
kfree(req->async_data);
- if (req->work.creds) {
- put_cred(req->work.creds);
- req->work.creds = NULL;
+ req->async_data = NULL;
}
}
-/* must to be called somewhat shortly after putting a request */
-static inline void io_put_task(struct task_struct *task, int nr)
-{
- struct io_uring_task *tctx = task->io_uring;
-
- percpu_counter_sub(&tctx->inflight, nr);
- if (unlikely(atomic_read(&tctx->in_idle)))
- wake_up(&tctx->wait);
- put_task_struct_many(task, nr);
-}
-
static void __io_free_req(struct io_kiocb *req)
{
struct io_ring_ctx *ctx = req->ctx;
io_dismantle_req(req);
io_put_task(req->task, 1);
- kmem_cache_free(req_cachep, req);
+ spin_lock(&ctx->completion_lock);
+ list_add(&req->inflight_entry, &ctx->locked_free_list);
+ ctx->locked_free_nr++;
+ spin_unlock(&ctx->completion_lock);
+
percpu_ref_put(&ctx->refs);
}
static bool io_kill_linked_timeout(struct io_kiocb *req)
__must_hold(&req->ctx->completion_lock)
+ __must_hold(&req->ctx->timeout_lock)
{
struct io_kiocb *link = req->link;
- /*
- * Can happen if a linked timeout fired and link had been like
- * req -> link t-out -> link t-out [-> ...]
- */
- if (link && (link->flags & REQ_F_LTIMEOUT_ACTIVE)) {
+ if (link && link->opcode == IORING_OP_LINK_TIMEOUT) {
struct io_timeout_data *io = link->async_data;
io_remove_next_linked(req);
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_put_req_deferred(link, 1);
+ io_put_req_deferred(link);
return true;
}
}
req->link = NULL;
while (link) {
+ long res = -ECANCELED;
+
+ if (link->flags & REQ_F_FAIL)
+ res = link->result;
+
nxt = link->link;
link->link = NULL;
trace_io_uring_fail_link(req, link);
- io_cqring_fill_event(link->ctx, link->user_data, -ECANCELED, 0);
- io_put_req_deferred(link, 2);
+ io_cqring_fill_event(link->ctx, link->user_data, res, 0);
+ io_put_req_deferred(link);
link = nxt;
}
}
{
bool posted = false;
- if (likely(req->flags & REQ_F_LINK_TIMEOUT))
+ if (req->flags & REQ_F_ARM_LTIMEOUT) {
+ struct io_kiocb *link = req->link;
+
+ 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_put_req_deferred(link);
+ posted = true;
+ }
+ } else if (req->flags & REQ_F_LINK_TIMEOUT) {
+ struct io_ring_ctx *ctx = req->ctx;
+
+ spin_lock_irq(&ctx->timeout_lock);
posted = io_kill_linked_timeout(req);
- if (unlikely((req->flags & REQ_F_FAIL_LINK) &&
+ spin_unlock_irq(&ctx->timeout_lock);
+ }
+ if (unlikely((req->flags & REQ_F_FAIL) &&
!(req->flags & REQ_F_HARDLINK))) {
posted |= (req->link != NULL);
io_fail_links(req);
* dependencies to the next request. In case of failure, fail the rest
* of the chain.
*/
- if (req->flags & (REQ_F_LINK_TIMEOUT | REQ_F_FAIL_LINK)) {
+ if (req->flags & IO_DISARM_MASK) {
struct io_ring_ctx *ctx = req->ctx;
- unsigned long flags;
bool posted;
- spin_lock_irqsave(&ctx->completion_lock, flags);
+ spin_lock(&ctx->completion_lock);
posted = io_disarm_next(req);
if (posted)
io_commit_cqring(req->ctx);
- spin_unlock_irqrestore(&ctx->completion_lock, flags);
+ spin_unlock(&ctx->completion_lock);
if (posted)
io_cqring_ev_posted(ctx);
}
return __io_req_find_next(req);
}
-static void ctx_flush_and_put(struct io_ring_ctx *ctx)
+static void ctx_flush_and_put(struct io_ring_ctx *ctx, bool *locked)
{
if (!ctx)
return;
- if (ctx->submit_state.comp.nr) {
- mutex_lock(&ctx->uring_lock);
- io_submit_flush_completions(&ctx->submit_state.comp, ctx);
+ if (*locked) {
+ if (ctx->submit_state.compl_nr)
+ io_submit_flush_completions(ctx);
mutex_unlock(&ctx->uring_lock);
+ *locked = false;
}
percpu_ref_put(&ctx->refs);
}
-static bool __tctx_task_work(struct io_uring_task *tctx)
+static void tctx_task_work(struct callback_head *cb)
{
+ bool locked = false;
struct io_ring_ctx *ctx = NULL;
- struct io_wq_work_list list;
- struct io_wq_work_node *node;
+ struct io_uring_task *tctx = container_of(cb, struct io_uring_task,
+ task_work);
- if (wq_list_empty(&tctx->task_list))
- return false;
+ while (1) {
+ struct io_wq_work_node *node;
- spin_lock_irq(&tctx->task_lock);
- list = tctx->task_list;
- INIT_WQ_LIST(&tctx->task_list);
- spin_unlock_irq(&tctx->task_lock);
+ if (!tctx->task_list.first && locked && ctx->submit_state.compl_nr)
+ io_submit_flush_completions(ctx);
- node = list.first;
- while (node) {
- struct io_wq_work_node *next = node->next;
- struct io_kiocb *req;
+ spin_lock_irq(&tctx->task_lock);
+ node = tctx->task_list.first;
+ INIT_WQ_LIST(&tctx->task_list);
+ if (!node)
+ tctx->task_running = false;
+ spin_unlock_irq(&tctx->task_lock);
+ if (!node)
+ break;
- req = container_of(node, struct io_kiocb, io_task_work.node);
- if (req->ctx != ctx) {
- ctx_flush_and_put(ctx);
- ctx = req->ctx;
- percpu_ref_get(&ctx->refs);
- }
+ do {
+ struct io_wq_work_node *next = node->next;
+ struct io_kiocb *req = container_of(node, struct io_kiocb,
+ io_task_work.node);
+
+ if (req->ctx != ctx) {
+ ctx_flush_and_put(ctx, &locked);
+ ctx = req->ctx;
+ /* if not contended, grab and improve batching */
+ locked = mutex_trylock(&ctx->uring_lock);
+ percpu_ref_get(&ctx->refs);
+ }
+ req->io_task_work.func(req, &locked);
+ node = next;
+ } while (node);
- req->task_work.func(&req->task_work);
- node = next;
+ cond_resched();
}
- ctx_flush_and_put(ctx);
- return list.first != NULL;
-}
-
-static void tctx_task_work(struct callback_head *cb)
-{
- struct io_uring_task *tctx = container_of(cb, struct io_uring_task, task_work);
-
- clear_bit(0, &tctx->task_state);
-
- while (__tctx_task_work(tctx))
- cond_resched();
+ ctx_flush_and_put(ctx, &locked);
}
-static int io_req_task_work_add(struct io_kiocb *req)
+static void io_req_task_work_add(struct io_kiocb *req)
{
struct task_struct *tsk = req->task;
struct io_uring_task *tctx = tsk->io_uring;
enum task_work_notify_mode notify;
- struct io_wq_work_node *node, *prev;
+ struct io_wq_work_node *node;
unsigned long flags;
- int ret = 0;
-
- if (unlikely(tsk->flags & PF_EXITING))
- return -ESRCH;
+ bool running;
WARN_ON_ONCE(!tctx);
spin_lock_irqsave(&tctx->task_lock, flags);
wq_list_add_tail(&req->io_task_work.node, &tctx->task_list);
+ running = tctx->task_running;
+ if (!running)
+ tctx->task_running = true;
spin_unlock_irqrestore(&tctx->task_lock, flags);
/* task_work already pending, we're done */
- if (test_bit(0, &tctx->task_state) ||
- test_and_set_bit(0, &tctx->task_state))
- return 0;
+ if (running)
+ return;
/*
* SQPOLL kernel thread doesn't need notification, just a wakeup. For
* will do the job.
*/
notify = (req->ctx->flags & IORING_SETUP_SQPOLL) ? TWA_NONE : TWA_SIGNAL;
-
if (!task_work_add(tsk, &tctx->task_work, notify)) {
wake_up_process(tsk);
- return 0;
+ return;
}
- /*
- * Slow path - we failed, find and delete work. if the work is not
- * in the list, it got run and we're fine.
- */
spin_lock_irqsave(&tctx->task_lock, flags);
- wq_list_for_each(node, prev, &tctx->task_list) {
- if (&req->io_task_work.node == node) {
- wq_list_del(&tctx->task_list, node, prev);
- ret = 1;
- break;
- }
- }
+ tctx->task_running = false;
+ node = tctx->task_list.first;
+ INIT_WQ_LIST(&tctx->task_list);
spin_unlock_irqrestore(&tctx->task_lock, flags);
- clear_bit(0, &tctx->task_state);
- return ret;
-}
-
-static bool io_run_task_work_head(struct callback_head **work_head)
-{
- struct callback_head *work, *next;
- bool executed = false;
-
- do {
- work = xchg(work_head, NULL);
- if (!work)
- break;
-
- do {
- next = work->next;
- work->func(work);
- work = next;
- cond_resched();
- } while (work);
- executed = true;
- } while (1);
-
- return executed;
-}
-
-static void io_task_work_add_head(struct callback_head **work_head,
- struct callback_head *task_work)
-{
- struct callback_head *head;
-
- do {
- head = READ_ONCE(*work_head);
- task_work->next = head;
- } while (cmpxchg(work_head, head, task_work) != head);
-}
-static void io_req_task_work_add_fallback(struct io_kiocb *req,
- task_work_func_t cb)
-{
- init_task_work(&req->task_work, cb);
- io_task_work_add_head(&req->ctx->exit_task_work, &req->task_work);
+ while (node) {
+ req = container_of(node, struct io_kiocb, io_task_work.node);
+ node = node->next;
+ if (llist_add(&req->io_task_work.fallback_node,
+ &req->ctx->fallback_llist))
+ schedule_delayed_work(&req->ctx->fallback_work, 1);
+ }
}
-static void io_req_task_cancel(struct callback_head *cb)
+static void io_req_task_cancel(struct io_kiocb *req, bool *locked)
{
- struct io_kiocb *req = container_of(cb, struct io_kiocb, task_work);
struct io_ring_ctx *ctx = req->ctx;
- /* ctx is guaranteed to stay alive while we hold uring_lock */
- mutex_lock(&ctx->uring_lock);
+ /* not needed for normal modes, but SQPOLL depends on it */
+ io_tw_lock(ctx, locked);
io_req_complete_failed(req, req->result);
- mutex_unlock(&ctx->uring_lock);
}
-static void __io_req_task_submit(struct io_kiocb *req)
+static void io_req_task_submit(struct io_kiocb *req, bool *locked)
{
struct io_ring_ctx *ctx = req->ctx;
- /* ctx stays valid until unlock, even if we drop all ours ctx->refs */
- mutex_lock(&ctx->uring_lock);
- if (!(current->flags & PF_EXITING) && !current->in_execve)
+ io_tw_lock(ctx, locked);
+ /* req->task == current here, checking PF_EXITING is safe */
+ if (likely(!(req->task->flags & PF_EXITING)))
__io_queue_sqe(req);
else
io_req_complete_failed(req, -EFAULT);
- mutex_unlock(&ctx->uring_lock);
-}
-
-static void io_req_task_submit(struct callback_head *cb)
-{
- struct io_kiocb *req = container_of(cb, struct io_kiocb, task_work);
-
- __io_req_task_submit(req);
}
static void io_req_task_queue_fail(struct io_kiocb *req, int ret)
{
req->result = ret;
- req->task_work.func = io_req_task_cancel;
-
- if (unlikely(io_req_task_work_add(req)))
- io_req_task_work_add_fallback(req, io_req_task_cancel);
+ req->io_task_work.func = io_req_task_cancel;
+ io_req_task_work_add(req);
}
static void io_req_task_queue(struct io_kiocb *req)
{
- req->task_work.func = io_req_task_submit;
+ req->io_task_work.func = io_req_task_submit;
+ io_req_task_work_add(req);
+}
- if (unlikely(io_req_task_work_add(req)))
- io_req_task_queue_fail(req, -ECANCELED);
+static void io_req_task_queue_reissue(struct io_kiocb *req)
+{
+ req->io_task_work.func = io_queue_async_work;
+ io_req_task_work_add(req);
}
static inline void io_queue_next(struct io_kiocb *req)
__io_free_req(req);
}
+static void io_free_req_work(struct io_kiocb *req, bool *locked)
+{
+ io_free_req(req);
+}
+
struct req_batch {
struct task_struct *task;
int task_refs;
static void io_req_free_batch_finish(struct io_ring_ctx *ctx,
struct req_batch *rb)
{
- if (rb->task)
- io_put_task(rb->task, rb->task_refs);
if (rb->ctx_refs)
percpu_ref_put_many(&ctx->refs, rb->ctx_refs);
+ if (rb->task)
+ io_put_task(rb->task, rb->task_refs);
}
static void io_req_free_batch(struct req_batch *rb, struct io_kiocb *req,
if (state->free_reqs != ARRAY_SIZE(state->reqs))
state->reqs[state->free_reqs++] = req;
else
- list_add(&req->compl.list, &state->comp.free_list);
+ list_add(&req->inflight_entry, &state->free_list);
}
-static void io_submit_flush_completions(struct io_comp_state *cs,
- struct io_ring_ctx *ctx)
+static void io_submit_flush_completions(struct io_ring_ctx *ctx)
+ __must_hold(&ctx->uring_lock)
{
- int i, nr = cs->nr;
- struct io_kiocb *req;
+ struct io_submit_state *state = &ctx->submit_state;
+ int i, nr = state->compl_nr;
struct req_batch rb;
- io_init_req_batch(&rb);
- spin_lock_irq(&ctx->completion_lock);
+ spin_lock(&ctx->completion_lock);
for (i = 0; i < nr; i++) {
- req = cs->reqs[i];
+ struct io_kiocb *req = state->compl_reqs[i];
+
__io_cqring_fill_event(ctx, req->user_data, req->result,
req->compl.cflags);
}
io_commit_cqring(ctx);
- spin_unlock_irq(&ctx->completion_lock);
-
+ spin_unlock(&ctx->completion_lock);
io_cqring_ev_posted(ctx);
+
+ io_init_req_batch(&rb);
for (i = 0; i < nr; i++) {
- req = cs->reqs[i];
+ struct io_kiocb *req = state->compl_reqs[i];
- /* submission and completion refs */
- if (req_ref_sub_and_test(req, 2))
+ if (req_ref_put_and_test(req))
io_req_free_batch(&rb, req, &ctx->submit_state);
}
io_req_free_batch_finish(ctx, &rb);
- cs->nr = 0;
+ state->compl_nr = 0;
}
/*
io_free_req(req);
}
-static void io_put_req_deferred_cb(struct callback_head *cb)
-{
- struct io_kiocb *req = container_of(cb, struct io_kiocb, task_work);
-
- io_free_req(req);
-}
-
-static void io_free_req_deferred(struct io_kiocb *req)
-{
- req->task_work.func = io_put_req_deferred_cb;
- if (unlikely(io_req_task_work_add(req)))
- io_req_task_work_add_fallback(req, io_put_req_deferred_cb);
-}
-
-static inline void io_put_req_deferred(struct io_kiocb *req, int refs)
+static inline void io_put_req_deferred(struct io_kiocb *req)
{
- if (req_ref_sub_and_test(req, refs))
- io_free_req_deferred(req);
+ if (req_ref_put_and_test(req)) {
+ req->io_task_work.func = io_free_req_work;
+ io_req_task_work_add(req);
+ }
}
static unsigned io_cqring_events(struct io_ring_ctx *ctx)
{
struct io_buffer *kbuf;
+ if (likely(!(req->flags & REQ_F_BUFFER_SELECTED)))
+ return 0;
kbuf = (struct io_buffer *) (unsigned long) req->rw.addr;
return io_put_kbuf(req, kbuf);
}
static inline bool io_run_task_work(void)
{
- /*
- * Not safe to run on exiting task, and the task_work handling will
- * not add work to such a task.
- */
- if (unlikely(current->flags & PF_EXITING))
- return false;
- if (current->task_works) {
+ if (test_thread_flag(TIF_NOTIFY_SIGNAL) || current->task_works) {
__set_current_state(TASK_RUNNING);
- task_work_run();
+ tracehook_notify_signal();
return true;
}
io_init_req_batch(&rb);
while (!list_empty(done)) {
- int cflags = 0;
-
req = list_first_entry(done, struct io_kiocb, inflight_entry);
list_del(&req->inflight_entry);
if (READ_ONCE(req->result) == -EAGAIN &&
!(req->flags & REQ_F_DONT_REISSUE)) {
req->iopoll_completed = 0;
- req_ref_get(req);
- io_queue_async_work(req);
+ io_req_task_queue_reissue(req);
continue;
}
- if (req->flags & REQ_F_BUFFER_SELECTED)
- cflags = io_put_rw_kbuf(req);
-
- __io_cqring_fill_event(ctx, req->user_data, req->result, cflags);
+ __io_cqring_fill_event(ctx, req->user_data, req->result,
+ io_put_rw_kbuf(req));
(*nr_events)++;
if (req_ref_put_and_test(req))
struct io_kiocb *req, *tmp;
LIST_HEAD(done);
bool spin;
- int ret;
/*
* Only spin for completions if we don't have multiple devices hanging
* off our complete list, and we're under the requested amount.
*/
- spin = !ctx->poll_multi_file && *nr_events < min;
+ spin = !ctx->poll_multi_queue && *nr_events < min;
- ret = 0;
list_for_each_entry_safe(req, tmp, &ctx->iopoll_list, inflight_entry) {
struct kiocb *kiocb = &req->rw.kiocb;
+ int ret;
/*
* Move completed and retryable entries to our local lists.
break;
ret = kiocb->ki_filp->f_op->iopoll(kiocb, spin);
- if (ret < 0)
- break;
+ if (unlikely(ret < 0))
+ return ret;
+ else if (ret)
+ spin = false;
/* iopoll may have completed current req */
if (READ_ONCE(req->iopoll_completed))
list_move_tail(&req->inflight_entry, &done);
-
- if (ret && spin)
- spin = false;
- ret = 0;
}
if (!list_empty(&done))
io_iopoll_complete(ctx, nr_events, &done);
- return ret;
+ return 0;
}
/*
* If we do, we can potentially be spinning for commands that
* already triggered a CQE (eg in error).
*/
- if (test_bit(0, &ctx->cq_check_overflow))
+ if (test_bit(0, &ctx->check_cq_overflow))
__io_cqring_overflow_flush(ctx, false);
if (io_cqring_events(ctx))
goto out;
* very same mutex.
*/
if (list_empty(&ctx->iopoll_list)) {
+ u32 tail = ctx->cached_cq_tail;
+
mutex_unlock(&ctx->uring_lock);
io_run_task_work();
mutex_lock(&ctx->uring_lock);
- if (list_empty(&ctx->iopoll_list))
+ /* some requests don't go through iopoll_list */
+ if (tail != ctx->cached_cq_tail ||
+ list_empty(&ctx->iopoll_list))
break;
}
ret = io_do_iopoll(ctx, &nr_events, min);
*/
if (percpu_ref_is_dying(&ctx->refs))
return false;
+ /*
+ * Play it safe and assume not safe to re-import and reissue if we're
+ * not in the original thread group (or in task context).
+ */
+ if (!same_thread_group(req->task, current) || !in_task())
+ return false;
return true;
}
#else
}
#endif
-static void __io_complete_rw(struct io_kiocb *req, long res, long res2,
- unsigned int issue_flags)
+static bool __io_complete_rw_common(struct io_kiocb *req, long res)
{
- int cflags = 0;
-
if (req->rw.kiocb.ki_flags & IOCB_WRITE)
kiocb_end_write(req);
if (res != req->result) {
if ((res == -EAGAIN || res == -EOPNOTSUPP) &&
io_rw_should_reissue(req)) {
req->flags |= REQ_F_REISSUE;
- return;
+ return true;
}
- req_set_fail_links(req);
+ req_set_fail(req);
+ req->result = res;
}
- if (req->flags & REQ_F_BUFFER_SELECTED)
- cflags = io_put_rw_kbuf(req);
- __io_req_complete(req, issue_flags, res, cflags);
+ return false;
+}
+
+static void io_req_task_complete(struct io_kiocb *req, bool *locked)
+{
+ unsigned int cflags = io_put_rw_kbuf(req);
+ long res = req->result;
+
+ if (*locked) {
+ struct io_ring_ctx *ctx = req->ctx;
+ struct io_submit_state *state = &ctx->submit_state;
+
+ io_req_complete_state(req, res, cflags);
+ state->compl_reqs[state->compl_nr++] = req;
+ if (state->compl_nr == ARRAY_SIZE(state->compl_reqs))
+ io_submit_flush_completions(ctx);
+ } else {
+ io_req_complete_post(req, res, cflags);
+ }
+}
+
+static void __io_complete_rw(struct io_kiocb *req, long res, long res2,
+ unsigned int issue_flags)
+{
+ if (__io_complete_rw_common(req, res))
+ return;
+ __io_req_complete(req, issue_flags, req->result, io_put_rw_kbuf(req));
}
static void io_complete_rw(struct kiocb *kiocb, long res, long res2)
{
struct io_kiocb *req = container_of(kiocb, struct io_kiocb, rw.kiocb);
- __io_complete_rw(req, res, res2, 0);
+ if (__io_complete_rw_common(req, res))
+ return;
+ req->result = res;
+ req->io_task_work.func = io_req_task_complete;
+ io_req_task_work_add(req);
}
static void io_complete_rw_iopoll(struct kiocb *kiocb, long res, long res2)
if (unlikely(res != req->result)) {
if (!(res == -EAGAIN && io_rw_should_reissue(req) &&
io_resubmit_prep(req))) {
- req_set_fail_links(req);
+ req_set_fail(req);
req->flags |= REQ_F_DONT_REISSUE;
}
}
* find it from a io_do_iopoll() thread before the issuer is done
* accessing the kiocb cookie.
*/
-static void io_iopoll_req_issued(struct io_kiocb *req, bool in_async)
+static void io_iopoll_req_issued(struct io_kiocb *req)
{
struct io_ring_ctx *ctx = req->ctx;
+ const bool in_async = io_wq_current_is_worker();
+
+ /* workqueue context doesn't hold uring_lock, grab it now */
+ if (unlikely(in_async))
+ mutex_lock(&ctx->uring_lock);
/*
* Track whether we have multiple files in our lists. This will impact
* different devices.
*/
if (list_empty(&ctx->iopoll_list)) {
- ctx->poll_multi_file = false;
- } else if (!ctx->poll_multi_file) {
+ ctx->poll_multi_queue = false;
+ } else if (!ctx->poll_multi_queue) {
struct io_kiocb *list_req;
+ unsigned int queue_num0, queue_num1;
list_req = list_first_entry(&ctx->iopoll_list, struct io_kiocb,
inflight_entry);
- if (list_req->file != req->file)
- ctx->poll_multi_file = true;
+
+ if (list_req->file != req->file) {
+ ctx->poll_multi_queue = true;
+ } else {
+ queue_num0 = blk_qc_t_to_queue_num(list_req->rw.kiocb.ki_cookie);
+ queue_num1 = blk_qc_t_to_queue_num(req->rw.kiocb.ki_cookie);
+ if (queue_num0 != queue_num1)
+ ctx->poll_multi_queue = true;
+ }
}
/*
else
list_add_tail(&req->inflight_entry, &ctx->iopoll_list);
- /*
- * If IORING_SETUP_SQPOLL is enabled, sqes are either handled in sq thread
- * task context or in io worker task context. If current task context is
- * sq thread, we don't need to check whether should wake up sq thread.
- */
- if (in_async && (ctx->flags & IORING_SETUP_SQPOLL) &&
- wq_has_sleeper(&ctx->sq_data->wait))
- wake_up(&ctx->sq_data->wait);
-}
-
-static inline void io_state_file_put(struct io_submit_state *state)
-{
- if (state->file_refs) {
- fput_many(state->file, state->file_refs);
- state->file_refs = 0;
- }
-}
-
-/*
- * Get as many references to a file as we have IOs left in this submission,
- * assuming most submissions are for one file, or at least that each file
- * has more than one submission.
- */
-static struct file *__io_file_get(struct io_submit_state *state, int fd)
-{
- if (!state)
- return fget(fd);
+ if (unlikely(in_async)) {
+ /*
+ * If IORING_SETUP_SQPOLL is enabled, sqes are either handle
+ * in sq thread task context or in io worker task context. If
+ * current task context is sq thread, we don't need to check
+ * whether should wake up sq thread.
+ */
+ if ((ctx->flags & IORING_SETUP_SQPOLL) &&
+ wq_has_sleeper(&ctx->sq_data->wait))
+ wake_up(&ctx->sq_data->wait);
- if (state->file_refs) {
- if (state->fd == fd) {
- state->file_refs--;
- return state->file;
- }
- io_state_file_put(state);
+ mutex_unlock(&ctx->uring_lock);
}
- state->file = fget_many(fd, state->ios_left);
- if (unlikely(!state->file))
- return NULL;
-
- state->fd = fd;
- state->file_refs = state->ios_left - 1;
- return state->file;
}
static bool io_bdev_nowait(struct block_device *bdev)
* any file. For now, just ensure that anything potentially problematic is done
* inline.
*/
-static bool __io_file_supports_async(struct file *file, int rw)
+static bool __io_file_supports_nowait(struct file *file, int rw)
{
umode_t mode = file_inode(file)->i_mode;
return true;
return false;
}
- if (S_ISCHR(mode) || S_ISSOCK(mode))
+ if (S_ISSOCK(mode))
return true;
if (S_ISREG(mode)) {
if (IS_ENABLED(CONFIG_BLOCK) &&
return file->f_op->write_iter != NULL;
}
-static bool io_file_supports_async(struct io_kiocb *req, int rw)
+static bool io_file_supports_nowait(struct io_kiocb *req, int rw)
{
- if (rw == READ && (req->flags & REQ_F_ASYNC_READ))
+ if (rw == READ && (req->flags & REQ_F_NOWAIT_READ))
return true;
- else if (rw == WRITE && (req->flags & REQ_F_ASYNC_WRITE))
+ else if (rw == WRITE && (req->flags & REQ_F_NOWAIT_WRITE))
return true;
- return __io_file_supports_async(req->file, rw);
+ return __io_file_supports_nowait(req->file, rw);
}
static int io_prep_rw(struct io_kiocb *req, const struct io_uring_sqe *sqe)
unsigned ioprio;
int ret;
- if (!(req->flags & REQ_F_ISREG) && S_ISREG(file_inode(file)->i_mode))
+ if (!io_req_ffs_set(req) && S_ISREG(file_inode(file)->i_mode))
req->flags |= REQ_F_ISREG;
kiocb->ki_pos = READ_ONCE(sqe->off);
!kiocb->ki_filp->f_op->iopoll)
return -EOPNOTSUPP;
- kiocb->ki_flags |= IOCB_HIPRI;
+ kiocb->ki_flags |= IOCB_HIPRI | IOCB_ALLOC_CACHE;
kiocb->ki_complete = io_complete_rw_iopoll;
req->iopoll_completed = 0;
} else {
if (req->flags & REQ_F_CUR_POS)
req->file->f_pos = kiocb->ki_pos;
- if (ret >= 0 && kiocb->ki_complete == io_complete_rw)
+ if (ret >= 0 && check_reissue)
__io_complete_rw(req, ret, 0, issue_flags);
else
io_rw_done(kiocb, ret);
- if (check_reissue && req->flags & REQ_F_REISSUE) {
+ if (check_reissue && (req->flags & REQ_F_REISSUE)) {
req->flags &= ~REQ_F_REISSUE;
if (io_resubmit_prep(req)) {
- req_ref_get(req);
- io_queue_async_work(req);
+ io_req_task_queue_reissue(req);
} else {
- int cflags = 0;
-
- req_set_fail_links(req);
- if (req->flags & REQ_F_BUFFER_SELECTED)
- cflags = io_put_rw_kbuf(req);
- __io_req_complete(req, issue_flags, ret, cflags);
+ req_set_fail(req);
+ __io_req_complete(req, issue_flags, ret,
+ io_put_rw_kbuf(req));
}
}
}
req->rw.kiocb.ki_flags &= ~IOCB_WAITQ;
list_del_init(&wait->entry);
-
- /* submit ref gets dropped, acquire a new one */
- req_ref_get(req);
io_req_task_queue(req);
return 1;
}
return true;
}
-static int io_iter_do_read(struct io_kiocb *req, struct iov_iter *iter)
+static inline int io_iter_do_read(struct io_kiocb *req, struct iov_iter *iter)
{
if (req->file->f_op->read_iter)
return call_read_iter(req->file, &req->rw.kiocb, iter);
return -EINVAL;
}
+static bool need_read_all(struct io_kiocb *req)
+{
+ return req->flags & REQ_F_ISREG ||
+ S_ISBLK(file_inode(req->file)->i_mode);
+}
+
static int io_read(struct io_kiocb *req, unsigned int issue_flags)
{
struct iovec inline_vecs[UIO_FASTIOV], *iovec = inline_vecs;
kiocb->ki_flags |= IOCB_NOWAIT;
/* If the file doesn't support async, just async punt */
- if (force_nonblock && !io_file_supports_async(req, READ)) {
+ if (force_nonblock && !io_file_supports_nowait(req, READ)) {
ret = io_setup_async_rw(req, iovec, inline_vecs, iter, true);
return ret ?: -EAGAIN;
}
if (req->flags & REQ_F_NOWAIT)
goto done;
/* some cases will consume bytes even on error returns */
+ iov_iter_reexpand(iter, iter->count + iter->truncated);
iov_iter_revert(iter, io_size - iov_iter_count(iter));
ret = 0;
} else if (ret == -EIOCBQUEUED) {
goto out_free;
} else if (ret <= 0 || ret == io_size || !force_nonblock ||
- (req->flags & REQ_F_NOWAIT) || !(req->flags & REQ_F_ISREG)) {
+ (req->flags & REQ_F_NOWAIT) || !need_read_all(req)) {
/* read all, failed, already did sync or don't want to retry */
goto done;
}
kiocb->ki_flags |= IOCB_NOWAIT;
/* If the file doesn't support async, just async punt */
- if (force_nonblock && !io_file_supports_async(req, WRITE))
+ if (force_nonblock && !io_file_supports_nowait(req, WRITE))
goto copy_iov;
/* file path doesn't support NOWAIT for non-direct_IO */
} else {
copy_iov:
/* some cases will consume bytes even on error returns */
+ iov_iter_reexpand(iter, iter->count + iter->truncated);
iov_iter_revert(iter, io_size - iov_iter_count(iter));
ret = io_setup_async_rw(req, iovec, inline_vecs, iter, false);
return ret ?: -EAGAIN;
struct io_rename *ren = &req->rename;
const char __user *oldf, *newf;
+ if (unlikely(req->ctx->flags & IORING_SETUP_IOPOLL))
+ return -EINVAL;
+ if (sqe->ioprio || sqe->buf_index || sqe->splice_fd_in)
+ return -EINVAL;
if (unlikely(req->flags & REQ_F_FIXED_FILE))
return -EBADF;
req->flags &= ~REQ_F_NEED_CLEANUP;
if (ret < 0)
- req_set_fail_links(req);
+ req_set_fail(req);
io_req_complete(req, ret);
return 0;
}
struct io_unlink *un = &req->unlink;
const char __user *fname;
+ if (unlikely(req->ctx->flags & IORING_SETUP_IOPOLL))
+ return -EINVAL;
+ if (sqe->ioprio || sqe->off || sqe->len || sqe->buf_index ||
+ sqe->splice_fd_in)
+ return -EINVAL;
if (unlikely(req->flags & REQ_F_FIXED_FILE))
return -EBADF;
req->flags &= ~REQ_F_NEED_CLEANUP;
if (ret < 0)
- req_set_fail_links(req);
+ req_set_fail(req);
io_req_complete(req, ret);
return 0;
}
-static int io_shutdown_prep(struct io_kiocb *req,
+static int io_mkdirat_prep(struct io_kiocb *req,
const struct io_uring_sqe *sqe)
{
-#if defined(CONFIG_NET)
+ struct io_mkdir *mkd = &req->mkdir;
+ const char __user *fname;
+
if (unlikely(req->ctx->flags & IORING_SETUP_IOPOLL))
return -EINVAL;
- if (sqe->ioprio || sqe->off || sqe->addr || sqe->rw_flags ||
- sqe->buf_index)
+ if (sqe->ioprio || sqe->off || sqe->rw_flags || sqe->buf_index ||
+ sqe->splice_fd_in)
return -EINVAL;
+ if (unlikely(req->flags & REQ_F_FIXED_FILE))
+ return -EBADF;
- req->shutdown.how = READ_ONCE(sqe->len);
+ mkd->dfd = READ_ONCE(sqe->fd);
+ mkd->mode = READ_ONCE(sqe->len);
+
+ fname = u64_to_user_ptr(READ_ONCE(sqe->addr));
+ mkd->filename = getname(fname);
+ if (IS_ERR(mkd->filename))
+ return PTR_ERR(mkd->filename);
+
+ req->flags |= REQ_F_NEED_CLEANUP;
return 0;
-#else
- return -EOPNOTSUPP;
-#endif
}
-static int io_shutdown(struct io_kiocb *req, unsigned int issue_flags)
+static int io_mkdirat(struct io_kiocb *req, int issue_flags)
{
-#if defined(CONFIG_NET)
- struct socket *sock;
+ struct io_mkdir *mkd = &req->mkdir;
int ret;
if (issue_flags & IO_URING_F_NONBLOCK)
return -EAGAIN;
- sock = sock_from_file(req->file);
- if (unlikely(!sock))
- return -ENOTSOCK;
+ ret = do_mkdirat(mkd->dfd, mkd->filename, mkd->mode);
- ret = __sys_shutdown_sock(sock, req->shutdown.how);
+ req->flags &= ~REQ_F_NEED_CLEANUP;
if (ret < 0)
- req_set_fail_links(req);
+ req_set_fail(req);
io_req_complete(req, ret);
return 0;
-#else
- return -EOPNOTSUPP;
-#endif
}
-static int __io_splice_prep(struct io_kiocb *req,
+static int io_symlinkat_prep(struct io_kiocb *req,
const struct io_uring_sqe *sqe)
{
- struct io_splice* sp = &req->splice;
- unsigned int valid_flags = SPLICE_F_FD_IN_FIXED | SPLICE_F_ALL;
+ struct io_symlink *sl = &req->symlink;
+ const char __user *oldpath, *newpath;
+
+ if (unlikely(req->ctx->flags & IORING_SETUP_IOPOLL))
+ return -EINVAL;
+ if (sqe->ioprio || sqe->len || sqe->rw_flags || sqe->buf_index ||
+ sqe->splice_fd_in)
+ return -EINVAL;
+ if (unlikely(req->flags & REQ_F_FIXED_FILE))
+ return -EBADF;
+
+ sl->new_dfd = READ_ONCE(sqe->fd);
+ oldpath = u64_to_user_ptr(READ_ONCE(sqe->addr));
+ newpath = u64_to_user_ptr(READ_ONCE(sqe->addr2));
+
+ sl->oldpath = getname(oldpath);
+ if (IS_ERR(sl->oldpath))
+ return PTR_ERR(sl->oldpath);
+
+ sl->newpath = getname(newpath);
+ if (IS_ERR(sl->newpath)) {
+ putname(sl->oldpath);
+ return PTR_ERR(sl->newpath);
+ }
+
+ req->flags |= REQ_F_NEED_CLEANUP;
+ return 0;
+}
+
+static int io_symlinkat(struct io_kiocb *req, int issue_flags)
+{
+ struct io_symlink *sl = &req->symlink;
+ int ret;
+
+ if (issue_flags & IO_URING_F_NONBLOCK)
+ return -EAGAIN;
+
+ ret = do_symlinkat(sl->oldpath, sl->new_dfd, sl->newpath);
+
+ req->flags &= ~REQ_F_NEED_CLEANUP;
+ if (ret < 0)
+ req_set_fail(req);
+ io_req_complete(req, ret);
+ return 0;
+}
+
+static int io_linkat_prep(struct io_kiocb *req,
+ const struct io_uring_sqe *sqe)
+{
+ struct io_hardlink *lnk = &req->hardlink;
+ const char __user *oldf, *newf;
+
+ if (unlikely(req->ctx->flags & IORING_SETUP_IOPOLL))
+ return -EINVAL;
+ if (sqe->ioprio || sqe->rw_flags || sqe->buf_index || sqe->splice_fd_in)
+ return -EINVAL;
+ if (unlikely(req->flags & REQ_F_FIXED_FILE))
+ return -EBADF;
+
+ lnk->old_dfd = READ_ONCE(sqe->fd);
+ lnk->new_dfd = READ_ONCE(sqe->len);
+ oldf = u64_to_user_ptr(READ_ONCE(sqe->addr));
+ newf = u64_to_user_ptr(READ_ONCE(sqe->addr2));
+ lnk->flags = READ_ONCE(sqe->hardlink_flags);
+
+ lnk->oldpath = getname(oldf);
+ if (IS_ERR(lnk->oldpath))
+ return PTR_ERR(lnk->oldpath);
+
+ lnk->newpath = getname(newf);
+ if (IS_ERR(lnk->newpath)) {
+ putname(lnk->oldpath);
+ return PTR_ERR(lnk->newpath);
+ }
+
+ req->flags |= REQ_F_NEED_CLEANUP;
+ return 0;
+}
+
+static int io_linkat(struct io_kiocb *req, int issue_flags)
+{
+ struct io_hardlink *lnk = &req->hardlink;
+ int ret;
+
+ if (issue_flags & IO_URING_F_NONBLOCK)
+ return -EAGAIN;
+
+ ret = do_linkat(lnk->old_dfd, lnk->oldpath, lnk->new_dfd,
+ lnk->newpath, lnk->flags);
+
+ req->flags &= ~REQ_F_NEED_CLEANUP;
+ if (ret < 0)
+ req_set_fail(req);
+ io_req_complete(req, ret);
+ return 0;
+}
+
+static int io_shutdown_prep(struct io_kiocb *req,
+ const struct io_uring_sqe *sqe)
+{
+#if defined(CONFIG_NET)
+ if (unlikely(req->ctx->flags & IORING_SETUP_IOPOLL))
+ return -EINVAL;
+ if (unlikely(sqe->ioprio || sqe->off || sqe->addr || sqe->rw_flags ||
+ sqe->buf_index || sqe->splice_fd_in))
+ return -EINVAL;
+
+ req->shutdown.how = READ_ONCE(sqe->len);
+ return 0;
+#else
+ return -EOPNOTSUPP;
+#endif
+}
+
+static int io_shutdown(struct io_kiocb *req, unsigned int issue_flags)
+{
+#if defined(CONFIG_NET)
+ struct socket *sock;
+ int ret;
+
+ if (issue_flags & IO_URING_F_NONBLOCK)
+ return -EAGAIN;
+
+ sock = sock_from_file(req->file);
+ if (unlikely(!sock))
+ return -ENOTSOCK;
+
+ ret = __sys_shutdown_sock(sock, req->shutdown.how);
+ if (ret < 0)
+ req_set_fail(req);
+ io_req_complete(req, ret);
+ return 0;
+#else
+ return -EOPNOTSUPP;
+#endif
+}
+
+static int __io_splice_prep(struct io_kiocb *req,
+ const struct io_uring_sqe *sqe)
+{
+ struct io_splice *sp = &req->splice;
+ unsigned int valid_flags = SPLICE_F_FD_IN_FIXED | SPLICE_F_ALL;
if (unlikely(req->ctx->flags & IORING_SETUP_IOPOLL))
return -EINVAL;
if (unlikely(sp->flags & ~valid_flags))
return -EINVAL;
- sp->file_in = io_file_get(NULL, req, READ_ONCE(sqe->splice_fd_in),
+ 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;
if (ret != sp->len)
- req_set_fail_links(req);
+ req_set_fail(req);
io_req_complete(req, ret);
return 0;
}
static int io_splice_prep(struct io_kiocb *req, const struct io_uring_sqe *sqe)
{
- struct io_splice* sp = &req->splice;
+ struct io_splice *sp = &req->splice;
sp->off_in = READ_ONCE(sqe->splice_off_in);
sp->off_out = READ_ONCE(sqe->off);
req->flags &= ~REQ_F_NEED_CLEANUP;
if (ret != sp->len)
- req_set_fail_links(req);
+ req_set_fail(req);
io_req_complete(req, ret);
return 0;
}
if (unlikely(ctx->flags & IORING_SETUP_IOPOLL))
return -EINVAL;
- if (unlikely(sqe->addr || sqe->ioprio || sqe->buf_index))
+ if (unlikely(sqe->addr || sqe->ioprio || sqe->buf_index ||
+ sqe->splice_fd_in))
return -EINVAL;
req->sync.flags = READ_ONCE(sqe->fsync_flags);
end > 0 ? end : LLONG_MAX,
req->sync.flags & IORING_FSYNC_DATASYNC);
if (ret < 0)
- req_set_fail_links(req);
+ req_set_fail(req);
io_req_complete(req, ret);
return 0;
}
static int io_fallocate_prep(struct io_kiocb *req,
const struct io_uring_sqe *sqe)
{
- if (sqe->ioprio || sqe->buf_index || sqe->rw_flags)
+ if (sqe->ioprio || sqe->buf_index || sqe->rw_flags ||
+ sqe->splice_fd_in)
return -EINVAL;
if (unlikely(req->ctx->flags & IORING_SETUP_IOPOLL))
return -EINVAL;
ret = vfs_fallocate(req->file, req->sync.mode, req->sync.off,
req->sync.len);
if (ret < 0)
- req_set_fail_links(req);
+ req_set_fail(req);
io_req_complete(req, ret);
return 0;
}
const char __user *fname;
int ret;
+ if (unlikely(req->ctx->flags & IORING_SETUP_IOPOLL))
+ return -EINVAL;
if (unlikely(sqe->ioprio || sqe->buf_index))
return -EINVAL;
if (unlikely(req->flags & REQ_F_FIXED_FILE))
req->open.filename = NULL;
return ret;
}
+
+ req->open.file_slot = READ_ONCE(sqe->file_index);
+ if (req->open.file_slot && (req->open.how.flags & O_CLOEXEC))
+ return -EINVAL;
+
req->open.nofile = rlimit(RLIMIT_NOFILE);
req->flags |= REQ_F_NEED_CLEANUP;
return 0;
static int io_openat_prep(struct io_kiocb *req, const struct io_uring_sqe *sqe)
{
- u64 flags, mode;
+ u64 mode = READ_ONCE(sqe->len);
+ u64 flags = READ_ONCE(sqe->open_flags);
- if (unlikely(req->ctx->flags & IORING_SETUP_IOPOLL))
- return -EINVAL;
- mode = READ_ONCE(sqe->len);
- flags = READ_ONCE(sqe->open_flags);
req->open.how = build_open_how(flags, mode);
return __io_openat_prep(req, sqe);
}
size_t len;
int ret;
- if (unlikely(req->ctx->flags & IORING_SETUP_IOPOLL))
- return -EINVAL;
how = u64_to_user_ptr(READ_ONCE(sqe->addr2));
len = READ_ONCE(sqe->len);
if (len < OPEN_HOW_SIZE_VER0)
{
struct open_flags op;
struct file *file;
- bool nonblock_set;
- bool resolve_nonblock;
+ bool resolve_nonblock, nonblock_set;
+ bool fixed = !!req->open.file_slot;
int ret;
ret = build_open_flags(&req->open.how, &op);
op.open_flag |= O_NONBLOCK;
}
- ret = __get_unused_fd_flags(req->open.how.flags, req->open.nofile);
- if (ret < 0)
- goto err;
+ if (!fixed) {
+ ret = __get_unused_fd_flags(req->open.how.flags, req->open.nofile);
+ if (ret < 0)
+ goto err;
+ }
file = do_filp_open(req->open.dfd, req->open.filename, &op);
- /* only retry if RESOLVE_CACHED wasn't already set by application */
- if ((!resolve_nonblock && (issue_flags & IO_URING_F_NONBLOCK)) &&
- file == ERR_PTR(-EAGAIN)) {
+ if (IS_ERR(file)) {
/*
- * We could hang on to this 'fd', but seems like marginal
- * gain for something that is now known to be a slower path.
- * So just put it, and we'll get a new one when we retry.
+ * We could hang on to this 'fd' on retrying, but seems like
+ * marginal gain for something that is now known to be a slower
+ * path. So just put it, and we'll get a new one when we retry.
*/
- put_unused_fd(ret);
- return -EAGAIN;
- }
+ if (!fixed)
+ put_unused_fd(ret);
- if (IS_ERR(file)) {
- put_unused_fd(ret);
ret = PTR_ERR(file);
- } else {
- if ((issue_flags & IO_URING_F_NONBLOCK) && !nonblock_set)
- file->f_flags &= ~O_NONBLOCK;
- fsnotify_open(file);
- fd_install(ret, file);
+ /* only retry if RESOLVE_CACHED wasn't already set by application */
+ if (ret == -EAGAIN &&
+ (!resolve_nonblock && (issue_flags & IO_URING_F_NONBLOCK)))
+ return -EAGAIN;
+ goto err;
}
+
+ if ((issue_flags & IO_URING_F_NONBLOCK) && !nonblock_set)
+ file->f_flags &= ~O_NONBLOCK;
+ fsnotify_open(file);
+
+ if (!fixed)
+ fd_install(ret, file);
+ else
+ ret = io_install_fixed_file(req, file, issue_flags,
+ req->open.file_slot - 1);
err:
putname(req->open.filename);
req->flags &= ~REQ_F_NEED_CLEANUP;
if (ret < 0)
- req_set_fail_links(req);
+ req_set_fail(req);
__io_req_complete(req, issue_flags, ret, 0);
return 0;
}
struct io_provide_buf *p = &req->pbuf;
u64 tmp;
- if (sqe->ioprio || sqe->rw_flags || sqe->addr || sqe->len || sqe->off)
+ if (sqe->ioprio || sqe->rw_flags || sqe->addr || sqe->len || sqe->off ||
+ sqe->splice_fd_in)
return -EINVAL;
tmp = READ_ONCE(sqe->fd);
if (head)
ret = __io_remove_buffers(ctx, head, p->bgid, p->nbufs);
if (ret < 0)
- req_set_fail_links(req);
+ req_set_fail(req);
/* complete before unlock, IOPOLL may need the lock */
__io_req_complete(req, issue_flags, ret, 0);
struct io_provide_buf *p = &req->pbuf;
u64 tmp;
- if (sqe->ioprio || sqe->rw_flags)
+ if (sqe->ioprio || sqe->rw_flags || sqe->splice_fd_in)
return -EINVAL;
tmp = READ_ONCE(sqe->fd);
__io_remove_buffers(ctx, head, p->bgid, -1U);
}
if (ret < 0)
- req_set_fail_links(req);
+ req_set_fail(req);
/* complete before unlock, IOPOLL may need the lock */
__io_req_complete(req, issue_flags, ret, 0);
io_ring_submit_unlock(ctx, !force_nonblock);
const struct io_uring_sqe *sqe)
{
#if defined(CONFIG_EPOLL)
- if (sqe->ioprio || sqe->buf_index)
+ if (sqe->ioprio || sqe->buf_index || sqe->splice_fd_in)
return -EINVAL;
if (unlikely(req->ctx->flags & IORING_SETUP_IOPOLL))
return -EINVAL;
return -EAGAIN;
if (ret < 0)
- req_set_fail_links(req);
+ req_set_fail(req);
__io_req_complete(req, issue_flags, ret, 0);
return 0;
#else
static int io_madvise_prep(struct io_kiocb *req, const struct io_uring_sqe *sqe)
{
#if defined(CONFIG_ADVISE_SYSCALLS) && defined(CONFIG_MMU)
- if (sqe->ioprio || sqe->buf_index || sqe->off)
+ if (sqe->ioprio || sqe->buf_index || sqe->off || sqe->splice_fd_in)
return -EINVAL;
if (unlikely(req->ctx->flags & IORING_SETUP_IOPOLL))
return -EINVAL;
ret = do_madvise(current->mm, ma->addr, ma->len, ma->advice);
if (ret < 0)
- req_set_fail_links(req);
+ req_set_fail(req);
io_req_complete(req, ret);
return 0;
#else
static int io_fadvise_prep(struct io_kiocb *req, const struct io_uring_sqe *sqe)
{
- if (sqe->ioprio || sqe->buf_index || sqe->addr)
+ if (sqe->ioprio || sqe->buf_index || sqe->addr || sqe->splice_fd_in)
return -EINVAL;
if (unlikely(req->ctx->flags & IORING_SETUP_IOPOLL))
return -EINVAL;
ret = vfs_fadvise(req->file, fa->offset, fa->len, fa->advice);
if (ret < 0)
- req_set_fail_links(req);
+ req_set_fail(req);
__io_req_complete(req, issue_flags, ret, 0);
return 0;
}
{
if (unlikely(req->ctx->flags & IORING_SETUP_IOPOLL))
return -EINVAL;
- if (sqe->ioprio || sqe->buf_index)
+ if (sqe->ioprio || sqe->buf_index || sqe->splice_fd_in)
return -EINVAL;
if (req->flags & REQ_F_FIXED_FILE)
return -EBADF;
ctx->buffer);
if (ret < 0)
- req_set_fail_links(req);
+ req_set_fail(req);
io_req_complete(req, ret);
return 0;
}
if (unlikely(req->ctx->flags & IORING_SETUP_IOPOLL))
return -EINVAL;
if (sqe->ioprio || sqe->off || sqe->addr || sqe->len ||
- sqe->rw_flags || sqe->buf_index)
+ sqe->rw_flags || sqe->buf_index || sqe->splice_fd_in)
return -EINVAL;
if (req->flags & REQ_F_FIXED_FILE)
return -EBADF;
ret = filp_close(file, current->files);
err:
if (ret < 0)
- req_set_fail_links(req);
+ req_set_fail(req);
if (file)
fput(file);
__io_req_complete(req, issue_flags, ret, 0);
if (unlikely(ctx->flags & IORING_SETUP_IOPOLL))
return -EINVAL;
- if (unlikely(sqe->addr || sqe->ioprio || sqe->buf_index))
+ if (unlikely(sqe->addr || sqe->ioprio || sqe->buf_index ||
+ sqe->splice_fd_in))
return -EINVAL;
req->sync.off = READ_ONCE(sqe->off);
ret = sync_file_range(req->file, req->sync.off, req->sync.len,
req->sync.flags);
if (ret < 0)
- req_set_fail_links(req);
+ req_set_fail(req);
io_req_complete(req, ret);
return 0;
}
kfree(kmsg->free_iov);
req->flags &= ~REQ_F_NEED_CLEANUP;
if (ret < min_ret)
- req_set_fail_links(req);
+ req_set_fail(req);
__io_req_complete(req, issue_flags, ret, 0);
return 0;
}
ret = -EINTR;
if (ret < min_ret)
- req_set_fail_links(req);
+ req_set_fail(req);
__io_req_complete(req, issue_flags, ret, 0);
return 0;
}
kfree(kmsg->free_iov);
req->flags &= ~REQ_F_NEED_CLEANUP;
if (ret < min_ret || ((flags & MSG_WAITALL) && (kmsg->msg.msg_flags & (MSG_TRUNC | MSG_CTRUNC))))
- req_set_fail_links(req);
+ req_set_fail(req);
__io_req_complete(req, issue_flags, ret, cflags);
return 0;
}
if (req->flags & REQ_F_BUFFER_SELECTED)
cflags = io_put_recv_kbuf(req);
if (ret < min_ret || ((flags & MSG_WAITALL) && (msg.msg_flags & (MSG_TRUNC | MSG_CTRUNC))))
- req_set_fail_links(req);
+ req_set_fail(req);
__io_req_complete(req, issue_flags, ret, cflags);
return 0;
}
accept->addr_len = u64_to_user_ptr(READ_ONCE(sqe->addr2));
accept->flags = READ_ONCE(sqe->accept_flags);
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)))
+ return -EINVAL;
+ if (accept->flags & ~(SOCK_CLOEXEC | SOCK_NONBLOCK))
+ return -EINVAL;
+ if (SOCK_NONBLOCK != O_NONBLOCK && (accept->flags & SOCK_NONBLOCK))
+ accept->flags = (accept->flags & ~SOCK_NONBLOCK) | O_NONBLOCK;
return 0;
}
struct io_accept *accept = &req->accept;
bool force_nonblock = issue_flags & IO_URING_F_NONBLOCK;
unsigned int file_flags = force_nonblock ? O_NONBLOCK : 0;
- int ret;
+ bool fixed = !!accept->file_slot;
+ struct file *file;
+ int ret, fd;
if (req->file->f_flags & O_NONBLOCK)
req->flags |= REQ_F_NOWAIT;
- ret = __sys_accept4_file(req->file, file_flags, accept->addr,
- accept->addr_len, accept->flags,
- accept->nofile);
- if (ret == -EAGAIN && force_nonblock)
- return -EAGAIN;
- if (ret < 0) {
+ if (!fixed) {
+ fd = __get_unused_fd_flags(accept->flags, accept->nofile);
+ if (unlikely(fd < 0))
+ return fd;
+ }
+ file = do_accept(req->file, file_flags, accept->addr, accept->addr_len,
+ accept->flags);
+ if (IS_ERR(file)) {
+ if (!fixed)
+ put_unused_fd(fd);
+ ret = PTR_ERR(file);
+ if (ret == -EAGAIN && force_nonblock)
+ return -EAGAIN;
if (ret == -ERESTARTSYS)
ret = -EINTR;
- req_set_fail_links(req);
+ req_set_fail(req);
+ } else if (!fixed) {
+ fd_install(fd, file);
+ ret = fd;
+ } else {
+ ret = io_install_fixed_file(req, file, issue_flags,
+ accept->file_slot - 1);
}
__io_req_complete(req, issue_flags, ret, 0);
return 0;
if (unlikely(req->ctx->flags & IORING_SETUP_IOPOLL))
return -EINVAL;
- if (sqe->ioprio || sqe->len || sqe->buf_index || sqe->rw_flags)
+ if (sqe->ioprio || sqe->len || sqe->buf_index || sqe->rw_flags ||
+ sqe->splice_fd_in)
return -EINVAL;
conn->addr = u64_to_user_ptr(READ_ONCE(sqe->addr));
ret = -EINTR;
out:
if (ret < 0)
- req_set_fail_links(req);
+ req_set_fail(req);
__io_req_complete(req, issue_flags, ret, 0);
return 0;
}
struct io_poll_table {
struct poll_table_struct pt;
struct io_kiocb *req;
+ int nr_entries;
int error;
};
static int __io_async_wake(struct io_kiocb *req, struct io_poll_iocb *poll,
- __poll_t mask, task_work_func_t func)
+ __poll_t mask, io_req_tw_func_t func)
{
- int ret;
-
/* for instances that support it check for an event match first: */
if (mask && !(mask & poll->events))
return 0;
list_del_init(&poll->wait.entry);
req->result = mask;
- req->task_work.func = func;
+ req->io_task_work.func = func;
/*
* If this fails, then the task is exiting. When a task exits, the
* of executing it. We can't safely execute it anyway, as we may not
* have the needed state needed for it anyway.
*/
- ret = io_req_task_work_add(req);
- if (unlikely(ret)) {
- WRITE_ONCE(poll->canceled, true);
- io_req_task_work_add_fallback(req, func);
- }
+ io_req_task_work_add(req);
return 1;
}
{
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_irq(&ctx->completion_lock);
+ spin_lock(&ctx->completion_lock);
if (!req->result && !READ_ONCE(poll->canceled)) {
add_wait_queue(poll->head, &poll->wait);
return true;
if (poll && poll->head) {
struct wait_queue_head *head = poll->head;
- spin_lock(&head->lock);
+ spin_lock_irq(&head->lock);
list_del_init(&poll->wait.entry);
if (poll->wait.private)
req_ref_put(req);
poll->head = NULL;
- spin_unlock(&head->lock);
+ spin_unlock_irq(&head->lock);
}
}
-static bool io_poll_complete(struct io_kiocb *req, __poll_t mask)
+static bool __io_poll_complete(struct io_kiocb *req, __poll_t mask)
__must_hold(&req->ctx->completion_lock)
{
struct io_ring_ctx *ctx = req->ctx;
if (req->poll.events & EPOLLONESHOT)
flags = 0;
if (!io_cqring_fill_event(ctx, req->user_data, error, flags)) {
- io_poll_remove_waitqs(req);
req->poll.done = true;
flags = 0;
}
if (flags & IORING_CQE_F_MORE)
ctx->cq_extra++;
- io_commit_cqring(ctx);
return !(flags & IORING_CQE_F_MORE);
}
-static void io_poll_task_func(struct callback_head *cb)
+static inline bool io_poll_complete(struct io_kiocb *req, __poll_t mask)
+ __must_hold(&req->ctx->completion_lock)
+{
+ bool done;
+
+ done = __io_poll_complete(req, mask);
+ io_commit_cqring(req->ctx);
+ return done;
+}
+
+static void io_poll_task_func(struct io_kiocb *req, bool *locked)
{
- struct io_kiocb *req = container_of(cb, struct io_kiocb, task_work);
struct io_ring_ctx *ctx = req->ctx;
struct io_kiocb *nxt;
if (io_poll_rewait(req, &req->poll)) {
- spin_unlock_irq(&ctx->completion_lock);
+ spin_unlock(&ctx->completion_lock);
} else {
bool done;
- done = io_poll_complete(req, req->result);
+ done = __io_poll_complete(req, req->result);
if (done) {
+ io_poll_remove_double(req);
hash_del(&req->hash_node);
} else {
req->result = 0;
add_wait_queue(req->poll.head, &req->poll.wait);
}
- spin_unlock_irq(&ctx->completion_lock);
+ io_commit_cqring(ctx);
+ spin_unlock(&ctx->completion_lock);
io_cqring_ev_posted(ctx);
if (done) {
nxt = io_put_req_find_next(req);
if (nxt)
- __io_req_task_submit(nxt);
+ io_req_task_submit(nxt, locked);
}
}
}
struct io_kiocb *req = wait->private;
struct io_poll_iocb *poll = io_poll_get_single(req);
__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))
list_del_init(&wait->entry);
- if (poll && poll->head) {
+ if (poll->head) {
bool done;
- spin_lock(&poll->head->lock);
+ 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(&poll->head->lock);
+ 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);
struct io_kiocb *req = pt->req;
/*
- * If poll->head is already set, it's because the file being polled
- * uses multiple waitqueues for poll handling (eg one for read, one
- * for write). Setup a separate io_poll_iocb if this happens.
+ * The file being polled uses multiple waitqueues for poll handling
+ * (e.g. one for read, one for write). Setup a separate io_poll_iocb
+ * if this happens.
*/
- if (unlikely(poll->head)) {
+ if (unlikely(pt->nr_entries)) {
struct io_poll_iocb *poll_one = poll;
+ /* double add on the same waitqueue head, ignore */
+ if (poll_one->head == head)
+ return;
/* already have a 2nd entry, fail a third attempt */
if (*poll_ptr) {
+ if ((*poll_ptr)->head == head)
+ return;
pt->error = -EINVAL;
return;
}
*/
if (!(poll_one->events & EPOLLONESHOT))
poll_one->events |= EPOLLONESHOT;
- /* double add on the same waitqueue head, ignore */
- if (poll_one->head == head)
- return;
poll = kmalloc(sizeof(*poll), GFP_ATOMIC);
if (!poll) {
pt->error = -ENOMEM;
*poll_ptr = poll;
}
- pt->error = 0;
+ pt->nr_entries++;
poll->head = head;
if (poll->events & EPOLLEXCLUSIVE)
__io_queue_proc(&apoll->poll, pt, head, &apoll->double_poll);
}
-static void io_async_task_func(struct callback_head *cb)
+static void io_async_task_func(struct io_kiocb *req, bool *locked)
{
- struct io_kiocb *req = container_of(cb, struct io_kiocb, task_work);
struct async_poll *apoll = req->apoll;
struct io_ring_ctx *ctx = req->ctx;
- trace_io_uring_task_run(req->ctx, req->opcode, req->user_data);
+ trace_io_uring_task_run(req->ctx, req, req->opcode, req->user_data);
if (io_poll_rewait(req, &apoll->poll)) {
- spin_unlock_irq(&ctx->completion_lock);
+ spin_unlock(&ctx->completion_lock);
return;
}
hash_del(&req->hash_node);
io_poll_remove_double(req);
- spin_unlock_irq(&ctx->completion_lock);
+ spin_unlock(&ctx->completion_lock);
if (!READ_ONCE(apoll->poll.canceled))
- __io_req_task_submit(req);
+ io_req_task_submit(req, locked);
else
io_req_complete_failed(req, -ECANCELED);
}
ipt->pt._key = mask;
ipt->req = req;
- ipt->error = -EINVAL;
+ ipt->error = 0;
+ ipt->nr_entries = 0;
mask = vfs_poll(req->file, &ipt->pt) & poll->events;
+ if (unlikely(!ipt->nr_entries) && !ipt->error)
+ ipt->error = -EINVAL;
- spin_lock_irq(&ctx->completion_lock);
+ spin_lock(&ctx->completion_lock);
+ if (ipt->error || (mask && (poll->events & EPOLLONESHOT)))
+ io_poll_remove_double(req);
if (likely(poll->head)) {
- spin_lock(&poll->head->lock);
+ spin_lock_irq(&poll->head->lock);
if (unlikely(list_empty(&poll->wait.entry))) {
if (ipt->error)
cancel = true;
WRITE_ONCE(poll->canceled, true);
else if (!poll->done) /* actually waiting for an event */
io_poll_req_insert(req);
- spin_unlock(&poll->head->lock);
+ spin_unlock_irq(&poll->head->lock);
}
return mask;
}
-static bool io_arm_poll_handler(struct io_kiocb *req)
+enum {
+ IO_APOLL_OK,
+ IO_APOLL_ABORTED,
+ IO_APOLL_READY
+};
+
+static int io_arm_poll_handler(struct io_kiocb *req)
{
const struct io_op_def *def = &io_op_defs[req->opcode];
struct io_ring_ctx *ctx = req->ctx;
struct async_poll *apoll;
struct io_poll_table ipt;
- __poll_t mask, ret;
+ __poll_t ret, mask = EPOLLONESHOT | POLLERR | POLLPRI;
int rw;
if (!req->file || !file_can_poll(req->file))
- return false;
+ return IO_APOLL_ABORTED;
if (req->flags & REQ_F_POLLED)
- return false;
- if (def->pollin)
+ return IO_APOLL_ABORTED;
+ if (!def->pollin && !def->pollout)
+ return IO_APOLL_ABORTED;
+
+ if (def->pollin) {
rw = READ;
- else if (def->pollout)
+ mask |= POLLIN | POLLRDNORM;
+
+ /* If reading from MSG_ERRQUEUE using recvmsg, ignore POLLIN */
+ if ((req->opcode == IORING_OP_RECVMSG) &&
+ (req->sr_msg.msg_flags & MSG_ERRQUEUE))
+ mask &= ~POLLIN;
+ } else {
rw = WRITE;
- else
- return false;
+ mask |= POLLOUT | POLLWRNORM;
+ }
+
/* if we can't nonblock try, then no point in arming a poll handler */
- if (!io_file_supports_async(req, rw))
- return false;
+ if (!io_file_supports_nowait(req, rw))
+ return IO_APOLL_ABORTED;
apoll = kmalloc(sizeof(*apoll), GFP_ATOMIC);
if (unlikely(!apoll))
- return false;
+ return IO_APOLL_ABORTED;
apoll->double_poll = NULL;
-
- req->flags |= REQ_F_POLLED;
req->apoll = apoll;
-
- mask = EPOLLONESHOT;
- if (def->pollin)
- mask |= POLLIN | POLLRDNORM;
- if (def->pollout)
- mask |= POLLOUT | POLLWRNORM;
-
- /* If reading from MSG_ERRQUEUE using recvmsg, ignore POLLIN */
- if ((req->opcode == IORING_OP_RECVMSG) &&
- (req->sr_msg.msg_flags & MSG_ERRQUEUE))
- mask &= ~POLLIN;
-
- mask |= POLLERR | POLLPRI;
-
+ 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);
- if (ret || ipt.error) {
- io_poll_remove_double(req);
- spin_unlock_irq(&ctx->completion_lock);
- return false;
- }
- spin_unlock_irq(&ctx->completion_lock);
- trace_io_uring_poll_arm(ctx, req->opcode, req->user_data, mask,
- apoll->poll.events);
- return true;
+ spin_unlock(&ctx->completion_lock);
+ if (ret || ipt.error)
+ return ret ? IO_APOLL_READY : IO_APOLL_ABORTED;
+
+ trace_io_uring_poll_arm(ctx, req, req->opcode, req->user_data,
+ mask, apoll->poll.events);
+ return IO_APOLL_OK;
}
static bool __io_poll_remove_one(struct io_kiocb *req,
if (!poll->head)
return false;
- spin_lock(&poll->head->lock);
+ 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(&poll->head->lock);
+ spin_unlock_irq(&poll->head->lock);
hash_del(&req->hash_node);
return do_complete;
}
-static bool io_poll_remove_waitqs(struct io_kiocb *req)
+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 (req->opcode != IORING_OP_POLL_ADD && do_complete) {
- /* non-poll requests have submit ref still */
- req_ref_put(req);
- }
- return do_complete;
-}
-
-static bool io_poll_remove_one(struct io_kiocb *req)
- __must_hold(&req->ctx->completion_lock)
-{
- bool do_complete;
-
- do_complete = io_poll_remove_waitqs(req);
if (do_complete) {
io_cqring_fill_event(req->ctx, req->user_data, -ECANCELED, 0);
io_commit_cqring(req->ctx);
- req_set_fail_links(req);
- io_put_req_deferred(req, 1);
+ req_set_fail(req);
+ io_put_req_deferred(req);
}
-
return do_complete;
}
* Returns true if we found and killed one or more poll requests
*/
static bool io_poll_remove_all(struct io_ring_ctx *ctx, struct task_struct *tsk,
- struct files_struct *files)
+ bool cancel_all)
{
struct hlist_node *tmp;
struct io_kiocb *req;
int posted = 0, i;
- spin_lock_irq(&ctx->completion_lock);
+ spin_lock(&ctx->completion_lock);
for (i = 0; i < (1U << ctx->cancel_hash_bits); i++) {
struct hlist_head *list;
list = &ctx->cancel_hash[i];
hlist_for_each_entry_safe(req, tmp, list, hash_node) {
- if (io_match_task(req, tsk, files))
+ if (io_match_task(req, tsk, cancel_all))
posted += io_poll_remove_one(req);
}
}
- spin_unlock_irq(&ctx->completion_lock);
+ spin_unlock(&ctx->completion_lock);
if (posted)
io_cqring_ev_posted(ctx);
if (unlikely(req->ctx->flags & IORING_SETUP_IOPOLL))
return -EINVAL;
- if (sqe->ioprio || sqe->buf_index)
+ if (sqe->ioprio || sqe->buf_index || sqe->splice_fd_in)
return -EINVAL;
flags = READ_ONCE(sqe->len);
if (flags & ~(IORING_POLL_UPDATE_EVENTS | IORING_POLL_UPDATE_USER_DATA |
if (flags & ~IORING_POLL_ADD_MULTI)
return -EINVAL;
+ io_req_set_refcount(req);
poll->events = io_poll_parse_events(sqe, flags);
return 0;
}
ipt.error = 0;
io_poll_complete(req, mask);
}
- spin_unlock_irq(&ctx->completion_lock);
+ spin_unlock(&ctx->completion_lock);
if (mask) {
io_cqring_ev_posted(ctx);
bool completing;
int ret;
- spin_lock_irq(&ctx->completion_lock);
+ spin_lock(&ctx->completion_lock);
preq = io_poll_find(ctx, req->poll_update.old_user_data, true);
if (!preq) {
ret = -ENOENT;
ret = 0;
err:
if (ret < 0) {
- spin_unlock_irq(&ctx->completion_lock);
- req_set_fail_links(req);
+ spin_unlock(&ctx->completion_lock);
+ req_set_fail(req);
io_req_complete(req, ret);
return 0;
}
}
if (req->poll_update.update_user_data)
preq->user_data = req->poll_update.new_user_data;
- spin_unlock_irq(&ctx->completion_lock);
+ 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_links(preq);
+ req_set_fail(preq);
io_req_complete(preq, ret);
}
}
return 0;
}
+static void io_req_task_timeout(struct io_kiocb *req, bool *locked)
+{
+ req_set_fail(req);
+ io_req_complete_post(req, -ETIME, 0);
+}
+
static enum hrtimer_restart io_timeout_fn(struct hrtimer *timer)
{
struct io_timeout_data *data = container_of(timer,
struct io_ring_ctx *ctx = req->ctx;
unsigned long flags;
- spin_lock_irqsave(&ctx->completion_lock, flags);
+ spin_lock_irqsave(&ctx->timeout_lock, flags);
list_del_init(&req->timeout.list);
atomic_set(&req->ctx->cq_timeouts,
atomic_read(&req->ctx->cq_timeouts) + 1);
+ spin_unlock_irqrestore(&ctx->timeout_lock, flags);
- io_cqring_fill_event(ctx, req->user_data, -ETIME, 0);
- io_commit_cqring(ctx);
- spin_unlock_irqrestore(&ctx->completion_lock, flags);
-
- io_cqring_ev_posted(ctx);
- req_set_fail_links(req);
- io_put_req(req);
+ req->io_task_work.func = io_req_task_timeout;
+ io_req_task_work_add(req);
return HRTIMER_NORESTART;
}
static struct io_kiocb *io_timeout_extract(struct io_ring_ctx *ctx,
__u64 user_data)
- __must_hold(&ctx->completion_lock)
+ __must_hold(&ctx->timeout_lock)
{
struct io_timeout_data *io;
struct io_kiocb *req;
static int io_timeout_cancel(struct io_ring_ctx *ctx, __u64 user_data)
__must_hold(&ctx->completion_lock)
+ __must_hold(&ctx->timeout_lock)
{
struct io_kiocb *req = io_timeout_extract(ctx, user_data);
if (IS_ERR(req))
return PTR_ERR(req);
- req_set_fail_links(req);
+ req_set_fail(req);
io_cqring_fill_event(ctx, req->user_data, -ECANCELED, 0);
- io_put_req_deferred(req, 1);
+ io_put_req_deferred(req);
+ return 0;
+}
+
+static clockid_t io_timeout_get_clock(struct io_timeout_data *data)
+{
+ switch (data->flags & IORING_TIMEOUT_CLOCK_MASK) {
+ case IORING_TIMEOUT_BOOTTIME:
+ return CLOCK_BOOTTIME;
+ case IORING_TIMEOUT_REALTIME:
+ return CLOCK_REALTIME;
+ default:
+ /* can't happen, vetted at prep time */
+ WARN_ON_ONCE(1);
+ fallthrough;
+ case 0:
+ return CLOCK_MONOTONIC;
+ }
+}
+
+static int io_linked_timeout_update(struct io_ring_ctx *ctx, __u64 user_data,
+ struct timespec64 *ts, enum hrtimer_mode mode)
+ __must_hold(&ctx->timeout_lock)
+{
+ struct io_timeout_data *io;
+ struct io_kiocb *req;
+ bool found = false;
+
+ list_for_each_entry(req, &ctx->ltimeout_list, timeout.list) {
+ found = user_data == req->user_data;
+ if (found)
+ break;
+ }
+ if (!found)
+ return -ENOENT;
+
+ io = req->async_data;
+ if (hrtimer_try_to_cancel(&io->timer) == -1)
+ return -EALREADY;
+ hrtimer_init(&io->timer, io_timeout_get_clock(io), mode);
+ io->timer.function = io_link_timeout_fn;
+ hrtimer_start(&io->timer, timespec64_to_ktime(*ts), mode);
return 0;
}
static int io_timeout_update(struct io_ring_ctx *ctx, __u64 user_data,
struct timespec64 *ts, enum hrtimer_mode mode)
- __must_hold(&ctx->completion_lock)
+ __must_hold(&ctx->timeout_lock)
{
struct io_kiocb *req = io_timeout_extract(ctx, user_data);
struct io_timeout_data *data;
req->timeout.off = 0; /* noseq */
data = req->async_data;
list_add_tail(&req->timeout.list, &ctx->timeout_list);
- hrtimer_init(&data->timer, CLOCK_MONOTONIC, mode);
+ hrtimer_init(&data->timer, io_timeout_get_clock(data), mode);
data->timer.function = io_timeout_fn;
hrtimer_start(&data->timer, timespec64_to_ktime(*ts), mode);
return 0;
return -EINVAL;
if (unlikely(req->flags & (REQ_F_FIXED_FILE | REQ_F_BUFFER_SELECT)))
return -EINVAL;
- if (sqe->ioprio || sqe->buf_index || sqe->len)
+ if (sqe->ioprio || sqe->buf_index || sqe->len || sqe->splice_fd_in)
return -EINVAL;
+ tr->ltimeout = false;
tr->addr = READ_ONCE(sqe->addr);
tr->flags = READ_ONCE(sqe->timeout_flags);
- if (tr->flags & IORING_TIMEOUT_UPDATE) {
- if (tr->flags & ~(IORING_TIMEOUT_UPDATE|IORING_TIMEOUT_ABS))
+ if (tr->flags & IORING_TIMEOUT_UPDATE_MASK) {
+ if (hweight32(tr->flags & IORING_TIMEOUT_CLOCK_MASK) > 1)
+ return -EINVAL;
+ if (tr->flags & IORING_LINK_TIMEOUT_UPDATE)
+ tr->ltimeout = true;
+ if (tr->flags & ~(IORING_TIMEOUT_UPDATE_MASK|IORING_TIMEOUT_ABS))
return -EINVAL;
if (get_timespec64(&tr->ts, u64_to_user_ptr(sqe->addr2)))
return -EFAULT;
struct io_ring_ctx *ctx = req->ctx;
int ret;
- spin_lock_irq(&ctx->completion_lock);
- if (!(req->timeout_rem.flags & IORING_TIMEOUT_UPDATE))
+ if (!(req->timeout_rem.flags & IORING_TIMEOUT_UPDATE)) {
+ spin_lock(&ctx->completion_lock);
+ spin_lock_irq(&ctx->timeout_lock);
ret = io_timeout_cancel(ctx, tr->addr);
- else
- ret = io_timeout_update(ctx, tr->addr, &tr->ts,
- io_translate_timeout_mode(tr->flags));
+ spin_unlock_irq(&ctx->timeout_lock);
+ spin_unlock(&ctx->completion_lock);
+ } else {
+ enum hrtimer_mode mode = io_translate_timeout_mode(tr->flags);
+
+ spin_lock_irq(&ctx->timeout_lock);
+ if (tr->ltimeout)
+ ret = io_linked_timeout_update(ctx, tr->addr, &tr->ts, mode);
+ else
+ ret = io_timeout_update(ctx, tr->addr, &tr->ts, mode);
+ spin_unlock_irq(&ctx->timeout_lock);
+ }
- io_cqring_fill_event(ctx, req->user_data, ret, 0);
- io_commit_cqring(ctx);
- spin_unlock_irq(&ctx->completion_lock);
- io_cqring_ev_posted(ctx);
if (ret < 0)
- req_set_fail_links(req);
- io_put_req(req);
+ req_set_fail(req);
+ io_req_complete_post(req, ret, 0);
return 0;
}
if (unlikely(req->ctx->flags & IORING_SETUP_IOPOLL))
return -EINVAL;
- if (sqe->ioprio || sqe->buf_index || sqe->len != 1)
+ if (sqe->ioprio || sqe->buf_index || sqe->len != 1 ||
+ sqe->splice_fd_in)
return -EINVAL;
if (off && is_timeout_link)
return -EINVAL;
flags = READ_ONCE(sqe->timeout_flags);
- if (flags & ~IORING_TIMEOUT_ABS)
+ if (flags & ~(IORING_TIMEOUT_ABS | IORING_TIMEOUT_CLOCK_MASK))
+ return -EINVAL;
+ /* more than one clock specified is invalid, obviously */
+ if (hweight32(flags & IORING_TIMEOUT_CLOCK_MASK) > 1)
return -EINVAL;
+ INIT_LIST_HEAD(&req->timeout.list);
req->timeout.off = off;
+ if (unlikely(off && !req->ctx->off_timeout_used))
+ req->ctx->off_timeout_used = true;
if (!req->async_data && io_alloc_async_data(req))
return -ENOMEM;
data = req->async_data;
data->req = req;
+ data->flags = flags;
if (get_timespec64(&data->ts, u64_to_user_ptr(sqe->addr)))
return -EFAULT;
data->mode = io_translate_timeout_mode(flags);
- hrtimer_init(&data->timer, CLOCK_MONOTONIC, data->mode);
- if (is_timeout_link)
- io_req_track_inflight(req);
+ hrtimer_init(&data->timer, io_timeout_get_clock(data), data->mode);
+
+ if (is_timeout_link) {
+ struct io_submit_link *link = &req->ctx->submit_state.link;
+
+ if (!link->head)
+ return -EINVAL;
+ if (link->last->opcode == IORING_OP_LINK_TIMEOUT)
+ return -EINVAL;
+ req->timeout.head = link->last;
+ link->last->flags |= REQ_F_ARM_LTIMEOUT;
+ }
return 0;
}
struct list_head *entry;
u32 tail, off = req->timeout.off;
- spin_lock_irq(&ctx->completion_lock);
+ spin_lock_irq(&ctx->timeout_lock);
/*
* sqe->off holds how many events that need to occur for this
list_add(&req->timeout.list, entry);
data->timer.function = io_timeout_fn;
hrtimer_start(&data->timer, timespec64_to_ktime(data->ts), data->mode);
- spin_unlock_irq(&ctx->completion_lock);
+ spin_unlock_irq(&ctx->timeout_lock);
return 0;
}
return ret;
}
-static void io_async_find_and_cancel(struct io_ring_ctx *ctx,
- struct io_kiocb *req, __u64 sqe_addr,
- int success_ret)
+static int io_try_cancel_userdata(struct io_kiocb *req, u64 sqe_addr)
{
- unsigned long flags;
+ struct io_ring_ctx *ctx = req->ctx;
int ret;
+ WARN_ON_ONCE(!io_wq_current_is_worker() && req->task != current);
+
ret = io_async_cancel_one(req->task->io_uring, sqe_addr, ctx);
- spin_lock_irqsave(&ctx->completion_lock, flags);
if (ret != -ENOENT)
- goto done;
+ return ret;
+
+ spin_lock(&ctx->completion_lock);
+ spin_lock_irq(&ctx->timeout_lock);
ret = io_timeout_cancel(ctx, sqe_addr);
+ spin_unlock_irq(&ctx->timeout_lock);
if (ret != -ENOENT)
- goto done;
+ goto out;
ret = io_poll_cancel(ctx, sqe_addr, false);
-done:
- if (!ret)
- ret = success_ret;
- io_cqring_fill_event(ctx, req->user_data, ret, 0);
- io_commit_cqring(ctx);
- spin_unlock_irqrestore(&ctx->completion_lock, flags);
- io_cqring_ev_posted(ctx);
-
- if (ret < 0)
- req_set_fail_links(req);
+out:
+ spin_unlock(&ctx->completion_lock);
+ return ret;
}
static int io_async_cancel_prep(struct io_kiocb *req,
return -EINVAL;
if (unlikely(req->flags & (REQ_F_FIXED_FILE | REQ_F_BUFFER_SELECT)))
return -EINVAL;
- if (sqe->ioprio || sqe->off || sqe->len || sqe->cancel_flags)
+ if (sqe->ioprio || sqe->off || sqe->len || sqe->cancel_flags ||
+ sqe->splice_fd_in)
return -EINVAL;
req->cancel.addr = READ_ONCE(sqe->addr);
struct io_tctx_node *node;
int ret;
- /* tasks should wait for their io-wq threads, so safe w/o sync */
- ret = io_async_cancel_one(req->task->io_uring, sqe_addr, ctx);
- spin_lock_irq(&ctx->completion_lock);
+ ret = io_try_cancel_userdata(req, sqe_addr);
if (ret != -ENOENT)
goto done;
- ret = io_timeout_cancel(ctx, sqe_addr);
- if (ret != -ENOENT)
- goto done;
- ret = io_poll_cancel(ctx, sqe_addr, false);
- if (ret != -ENOENT)
- goto done;
- spin_unlock_irq(&ctx->completion_lock);
/* slow path, try all io-wq's */
io_ring_submit_lock(ctx, !(issue_flags & IO_URING_F_NONBLOCK));
break;
}
io_ring_submit_unlock(ctx, !(issue_flags & IO_URING_F_NONBLOCK));
-
- spin_lock_irq(&ctx->completion_lock);
done:
- io_cqring_fill_event(ctx, req->user_data, ret, 0);
- io_commit_cqring(ctx);
- spin_unlock_irq(&ctx->completion_lock);
- io_cqring_ev_posted(ctx);
-
if (ret < 0)
- req_set_fail_links(req);
- io_put_req(req);
+ req_set_fail(req);
+ io_req_complete_post(req, ret, 0);
return 0;
}
{
if (unlikely(req->flags & (REQ_F_FIXED_FILE | REQ_F_BUFFER_SELECT)))
return -EINVAL;
- if (sqe->ioprio || sqe->rw_flags)
+ if (sqe->ioprio || sqe->rw_flags || sqe->splice_fd_in)
return -EINVAL;
req->rsrc_update.offset = READ_ONCE(sqe->off);
mutex_unlock(&ctx->uring_lock);
if (ret < 0)
- req_set_fail_links(req);
+ req_set_fail(req);
__io_req_complete(req, issue_flags, ret, 0);
return 0;
}
return io_renameat_prep(req, sqe);
case IORING_OP_UNLINKAT:
return io_unlinkat_prep(req, sqe);
+ case IORING_OP_MKDIRAT:
+ return io_mkdirat_prep(req, sqe);
+ case IORING_OP_SYMLINKAT:
+ return io_symlinkat_prep(req, sqe);
+ case IORING_OP_LINKAT:
+ return io_linkat_prep(req, sqe);
}
printk_once(KERN_WARNING "io_uring: unhandled opcode %d\n",
static u32 io_get_sequence(struct io_kiocb *req)
{
- struct io_kiocb *pos;
- struct io_ring_ctx *ctx = req->ctx;
- u32 total_submitted, nr_reqs = 0;
+ u32 seq = req->ctx->cached_sq_head;
- io_for_each_link(pos, req)
- nr_reqs++;
-
- total_submitted = ctx->cached_sq_head - ctx->cached_sq_dropped;
- return total_submitted - nr_reqs;
+ /* need original cached_sq_head, but it was increased for each req */
+ io_for_each_link(req, req)
+ seq--;
+ return seq;
}
-static int io_req_defer(struct io_kiocb *req)
+static bool io_drain_req(struct io_kiocb *req)
{
+ struct io_kiocb *pos;
struct io_ring_ctx *ctx = req->ctx;
struct io_defer_entry *de;
int ret;
u32 seq;
+ if (req->flags & REQ_F_FAIL) {
+ io_req_complete_fail_submit(req);
+ return true;
+ }
+
+ /*
+ * If we need to drain a request in the middle of a link, drain the
+ * head request and the next request/link after the current link.
+ * Considering sequential execution of links, IOSQE_IO_DRAIN will be
+ * maintained for every request of our link.
+ */
+ if (ctx->drain_next) {
+ req->flags |= REQ_F_IO_DRAIN;
+ ctx->drain_next = false;
+ }
+ /* not interested in head, start from the first linked */
+ io_for_each_link(pos, req->link) {
+ if (pos->flags & REQ_F_IO_DRAIN) {
+ ctx->drain_next = true;
+ req->flags |= REQ_F_IO_DRAIN;
+ break;
+ }
+ }
+
/* Still need defer if there is pending req in defer list. */
if (likely(list_empty_careful(&ctx->defer_list) &&
- !(req->flags & REQ_F_IO_DRAIN)))
- return 0;
+ !(req->flags & REQ_F_IO_DRAIN))) {
+ ctx->drain_active = false;
+ return false;
+ }
seq = io_get_sequence(req);
/* Still a chance to pass the sequence check */
if (!req_need_defer(req, seq) && list_empty_careful(&ctx->defer_list))
- return 0;
+ return false;
ret = io_req_prep_async(req);
if (ret)
- return ret;
+ goto fail;
io_prep_async_link(req);
de = kmalloc(sizeof(*de), GFP_KERNEL);
- if (!de)
- return -ENOMEM;
+ if (!de) {
+ ret = -ENOMEM;
+fail:
+ io_req_complete_failed(req, ret);
+ return true;
+ }
- spin_lock_irq(&ctx->completion_lock);
+ spin_lock(&ctx->completion_lock);
if (!req_need_defer(req, seq) && list_empty(&ctx->defer_list)) {
- spin_unlock_irq(&ctx->completion_lock);
+ spin_unlock(&ctx->completion_lock);
kfree(de);
- io_queue_async_work(req);
- return -EIOCBQUEUED;
+ io_queue_async_work(req, NULL);
+ return true;
}
trace_io_uring_defer(ctx, req, req->user_data);
de->req = req;
de->seq = seq;
list_add_tail(&de->list, &ctx->defer_list);
- spin_unlock_irq(&ctx->completion_lock);
- return -EIOCBQUEUED;
+ spin_unlock(&ctx->completion_lock);
+ return true;
}
static void io_clean_op(struct io_kiocb *req)
kfree(req->sr_msg.kbuf);
break;
}
- req->flags &= ~REQ_F_BUFFER_SELECTED;
}
if (req->flags & REQ_F_NEED_CLEANUP) {
case IORING_OP_WRITE_FIXED:
case IORING_OP_WRITE: {
struct io_async_rw *io = req->async_data;
- if (io->free_iovec)
- kfree(io->free_iovec);
+
+ kfree(io->free_iovec);
break;
}
case IORING_OP_RECVMSG:
case IORING_OP_UNLINKAT:
putname(req->unlink.filename);
break;
+ case IORING_OP_MKDIRAT:
+ putname(req->mkdir.filename);
+ break;
+ case IORING_OP_SYMLINKAT:
+ putname(req->symlink.oldpath);
+ putname(req->symlink.newpath);
+ break;
+ case IORING_OP_LINKAT:
+ putname(req->hardlink.oldpath);
+ putname(req->hardlink.newpath);
+ break;
}
- req->flags &= ~REQ_F_NEED_CLEANUP;
}
if ((req->flags & REQ_F_POLLED) && req->apoll) {
kfree(req->apoll->double_poll);
struct io_uring_task *tctx = req->task->io_uring;
atomic_dec(&tctx->inflight_tracked);
- req->flags &= ~REQ_F_INFLIGHT;
}
+ if (req->flags & REQ_F_CREDS)
+ put_cred(req->creds);
+
+ req->flags &= ~IO_REQ_CLEAN_FLAGS;
}
static int io_issue_sqe(struct io_kiocb *req, unsigned int issue_flags)
const struct cred *creds = NULL;
int ret;
- if (req->work.creds && req->work.creds != current_cred())
- creds = override_creds(req->work.creds);
+ if ((req->flags & REQ_F_CREDS) && req->creds != current_cred())
+ creds = override_creds(req->creds);
switch (req->opcode) {
case IORING_OP_NOP:
case IORING_OP_UNLINKAT:
ret = io_unlinkat(req, issue_flags);
break;
+ case IORING_OP_MKDIRAT:
+ ret = io_mkdirat(req, issue_flags);
+ break;
+ case IORING_OP_SYMLINKAT:
+ ret = io_symlinkat(req, issue_flags);
+ break;
+ case IORING_OP_LINKAT:
+ ret = io_linkat(req, issue_flags);
+ break;
default:
ret = -EINVAL;
break;
if (creds)
revert_creds(creds);
-
if (ret)
return ret;
-
/* If the op doesn't have a file, we're not polling for it */
- if ((ctx->flags & IORING_SETUP_IOPOLL) && req->file) {
- const bool in_async = io_wq_current_is_worker();
-
- /* workqueue context doesn't hold uring_lock, grab it now */
- if (in_async)
- mutex_lock(&ctx->uring_lock);
+ if ((ctx->flags & IORING_SETUP_IOPOLL) && req->file)
+ io_iopoll_req_issued(req);
- io_iopoll_req_issued(req, in_async);
+ return 0;
+}
- if (in_async)
- mutex_unlock(&ctx->uring_lock);
- }
+static struct io_wq_work *io_wq_free_work(struct io_wq_work *work)
+{
+ struct io_kiocb *req = container_of(work, struct io_kiocb, work);
- return 0;
+ req = io_put_req_find_next(req);
+ return req ? &req->work : NULL;
}
static void io_wq_submit_work(struct io_wq_work *work)
struct io_kiocb *timeout;
int ret = 0;
+ /* one will be dropped by ->io_free_work() after returning to io-wq */
+ if (!(req->flags & REQ_F_REFCOUNT))
+ __io_req_set_refcount(req, 2);
+ else
+ req_ref_get(req);
+
timeout = io_prep_linked_timeout(req);
if (timeout)
io_queue_linked_timeout(timeout);
+ /* either cancelled or io-wq is dying, so don't touch tctx->iowq */
if (work->flags & IO_WQ_WORK_CANCEL)
ret = -ECANCELED;
}
/* avoid locking problems by failing it from a clean context */
- if (ret) {
- /* io-wq is going to take one down */
- req_ref_get(req);
+ if (ret)
io_req_task_queue_fail(req, ret);
- }
}
-#define FFS_ASYNC_READ 0x1UL
-#define FFS_ASYNC_WRITE 0x2UL
-#ifdef CONFIG_64BIT
-#define FFS_ISREG 0x4UL
-#else
-#define FFS_ISREG 0x0UL
-#endif
-#define FFS_MASK ~(FFS_ASYNC_READ|FFS_ASYNC_WRITE|FFS_ISREG)
-
static inline struct io_fixed_file *io_fixed_file_slot(struct io_file_table *table,
- unsigned i)
+ unsigned i)
{
- struct io_fixed_file *table_l2;
-
- table_l2 = table->files[i >> IORING_FILE_TABLE_SHIFT];
- return &table_l2[i & IORING_FILE_TABLE_MASK];
+ return &table->files[i];
}
static inline struct file *io_file_from_index(struct io_ring_ctx *ctx,
{
unsigned long file_ptr = (unsigned long) file;
- if (__io_file_supports_async(file, READ))
+ if (__io_file_supports_nowait(file, READ))
file_ptr |= FFS_ASYNC_READ;
- if (__io_file_supports_async(file, WRITE))
+ if (__io_file_supports_nowait(file, WRITE))
file_ptr |= FFS_ASYNC_WRITE;
if (S_ISREG(file_inode(file)->i_mode))
file_ptr |= FFS_ISREG;
file_slot->file_ptr = file_ptr;
}
-static struct file *io_file_get(struct io_submit_state *state,
- struct io_kiocb *req, int fd, bool fixed)
+static inline struct file *io_file_get_fixed(struct io_ring_ctx *ctx,
+ struct io_kiocb *req, int fd)
{
- struct io_ring_ctx *ctx = req->ctx;
struct file *file;
+ unsigned long file_ptr;
- if (fixed) {
- unsigned long file_ptr;
+ if (unlikely((unsigned int)fd >= ctx->nr_user_files))
+ return NULL;
+ fd = array_index_nospec(fd, ctx->nr_user_files);
+ file_ptr = io_fixed_file_slot(&ctx->file_table, fd)->file_ptr;
+ file = (struct file *) (file_ptr & FFS_MASK);
+ file_ptr &= ~FFS_MASK;
+ /* mask in overlapping REQ_F and FFS bits */
+ req->flags |= (file_ptr << REQ_F_NOWAIT_READ_BIT);
+ io_req_set_rsrc_node(req);
+ return file;
+}
- if (unlikely((unsigned int)fd >= ctx->nr_user_files))
- return NULL;
- fd = array_index_nospec(fd, ctx->nr_user_files);
- file_ptr = io_fixed_file_slot(&ctx->file_table, fd)->file_ptr;
- file = (struct file *) (file_ptr & FFS_MASK);
- file_ptr &= ~FFS_MASK;
- /* mask in overlapping REQ_F and FFS bits */
- req->flags |= (file_ptr << REQ_F_ASYNC_READ_BIT);
- io_req_set_rsrc_node(req);
- } else {
- trace_io_uring_file_get(ctx, fd);
- file = __io_file_get(state, fd);
+static struct file *io_file_get_normal(struct io_ring_ctx *ctx,
+ struct io_kiocb *req, int fd)
+{
+ struct file *file = fget(fd);
- /* we don't allow fixed io_uring files */
- if (file && unlikely(file->f_op == &io_uring_fops))
- io_req_track_inflight(req);
- }
+ trace_io_uring_file_get(ctx, fd);
+ /* we don't allow fixed io_uring files */
+ if (file && unlikely(file->f_op == &io_uring_fops))
+ io_req_track_inflight(req);
return file;
}
+static inline struct file *io_file_get(struct io_ring_ctx *ctx,
+ struct io_kiocb *req, int fd, bool fixed)
+{
+ if (fixed)
+ return io_file_get_fixed(ctx, req, fd);
+ else
+ return io_file_get_normal(ctx, req, fd);
+}
+
+static void io_req_task_link_timeout(struct io_kiocb *req, bool *locked)
+{
+ struct io_kiocb *prev = req->timeout.prev;
+ int ret;
+
+ if (prev) {
+ ret = io_try_cancel_userdata(req, prev->user_data);
+ io_req_complete_post(req, ret ?: -ETIME, 0);
+ io_put_req(prev);
+ } else {
+ io_req_complete_post(req, -ETIME, 0);
+ }
+}
+
static enum hrtimer_restart io_link_timeout_fn(struct hrtimer *timer)
{
struct io_timeout_data *data = container_of(timer,
struct io_ring_ctx *ctx = req->ctx;
unsigned long flags;
- spin_lock_irqsave(&ctx->completion_lock, flags);
+ spin_lock_irqsave(&ctx->timeout_lock, flags);
prev = req->timeout.head;
req->timeout.head = NULL;
if (!req_ref_inc_not_zero(prev))
prev = NULL;
}
- spin_unlock_irqrestore(&ctx->completion_lock, flags);
+ list_del(&req->timeout.list);
+ req->timeout.prev = prev;
+ spin_unlock_irqrestore(&ctx->timeout_lock, flags);
- if (prev) {
- io_async_find_and_cancel(ctx, req, prev->user_data, -ETIME);
- io_put_req_deferred(prev, 1);
- io_put_req_deferred(req, 1);
- } else {
- io_req_complete_post(req, -ETIME, 0);
- }
+ req->io_task_work.func = io_req_task_link_timeout;
+ io_req_task_work_add(req);
return HRTIMER_NORESTART;
}
{
struct io_ring_ctx *ctx = req->ctx;
- spin_lock_irq(&ctx->completion_lock);
+ spin_lock_irq(&ctx->timeout_lock);
/*
* If the back reference is NULL, then our linked request finished
* before we got a chance to setup the timer
data->timer.function = io_link_timeout_fn;
hrtimer_start(&data->timer, timespec64_to_ktime(data->ts),
data->mode);
+ list_add_tail(&req->timeout.list, &ctx->ltimeout_list);
}
- spin_unlock_irq(&ctx->completion_lock);
+ spin_unlock_irq(&ctx->timeout_lock);
/* drop submission reference */
io_put_req(req);
}
-static struct io_kiocb *io_prep_linked_timeout(struct io_kiocb *req)
-{
- struct io_kiocb *nxt = req->link;
-
- if (!nxt || (req->flags & REQ_F_LINK_TIMEOUT) ||
- nxt->opcode != IORING_OP_LINK_TIMEOUT)
- return NULL;
-
- nxt->timeout.head = req;
- nxt->flags |= REQ_F_LTIMEOUT_ACTIVE;
- req->flags |= REQ_F_LINK_TIMEOUT;
- return nxt;
-}
-
static void __io_queue_sqe(struct io_kiocb *req)
+ __must_hold(&req->ctx->uring_lock)
{
- struct io_kiocb *linked_timeout = io_prep_linked_timeout(req);
+ struct io_kiocb *linked_timeout;
int ret;
+issue_sqe:
ret = io_issue_sqe(req, IO_URING_F_NONBLOCK|IO_URING_F_COMPLETE_DEFER);
/*
* doesn't support non-blocking read/write attempts
*/
if (likely(!ret)) {
- /* drop submission reference */
if (req->flags & REQ_F_COMPLETE_INLINE) {
struct io_ring_ctx *ctx = req->ctx;
- struct io_comp_state *cs = &ctx->submit_state.comp;
+ struct io_submit_state *state = &ctx->submit_state;
- cs->reqs[cs->nr++] = req;
- if (cs->nr == ARRAY_SIZE(cs->reqs))
- io_submit_flush_completions(cs, ctx);
- } else {
- io_put_req(req);
+ state->compl_reqs[state->compl_nr++] = req;
+ if (state->compl_nr == ARRAY_SIZE(state->compl_reqs))
+ io_submit_flush_completions(ctx);
+ return;
}
+
+ linked_timeout = io_prep_linked_timeout(req);
+ if (linked_timeout)
+ io_queue_linked_timeout(linked_timeout);
} else if (ret == -EAGAIN && !(req->flags & REQ_F_NOWAIT)) {
- if (!io_arm_poll_handler(req)) {
+ linked_timeout = io_prep_linked_timeout(req);
+
+ switch (io_arm_poll_handler(req)) {
+ case IO_APOLL_READY:
+ if (linked_timeout)
+ io_unprep_linked_timeout(req);
+ goto issue_sqe;
+ case IO_APOLL_ABORTED:
/*
* Queued up for async execution, worker will release
* submit reference when the iocb is actually submitted.
*/
- io_queue_async_work(req);
+ io_queue_async_work(req, NULL);
+ break;
}
+
+ if (linked_timeout)
+ io_queue_linked_timeout(linked_timeout);
} else {
io_req_complete_failed(req, ret);
}
- if (linked_timeout)
- io_queue_linked_timeout(linked_timeout);
}
-static void io_queue_sqe(struct io_kiocb *req)
+static inline void io_queue_sqe(struct io_kiocb *req)
+ __must_hold(&req->ctx->uring_lock)
{
- int ret;
+ if (unlikely(req->ctx->drain_active) && io_drain_req(req))
+ return;
- ret = io_req_defer(req);
- if (ret) {
- if (ret != -EIOCBQUEUED) {
-fail_req:
- io_req_complete_failed(req, ret);
- }
- } else if (req->flags & REQ_F_FORCE_ASYNC) {
- ret = io_req_prep_async(req);
- if (unlikely(ret))
- goto fail_req;
- io_queue_async_work(req);
- } else {
+ if (likely(!(req->flags & (REQ_F_FORCE_ASYNC | REQ_F_FAIL)))) {
__io_queue_sqe(req);
+ } else if (req->flags & REQ_F_FAIL) {
+ io_req_complete_fail_submit(req);
+ } else {
+ int ret = io_req_prep_async(req);
+
+ if (unlikely(ret))
+ io_req_complete_failed(req, ret);
+ else
+ io_queue_async_work(req, NULL);
}
}
struct io_kiocb *req,
unsigned int sqe_flags)
{
- if (!ctx->restricted)
+ if (likely(!ctx->restricted))
return true;
if (!test_bit(req->opcode, ctx->restrictions.sqe_op))
static int io_init_req(struct io_ring_ctx *ctx, struct io_kiocb *req,
const struct io_uring_sqe *sqe)
+ __must_hold(&ctx->uring_lock)
{
struct io_submit_state *state;
unsigned int sqe_flags;
int personality, ret = 0;
+ /* req is partially pre-initialised, see io_preinit_req() */
req->opcode = READ_ONCE(sqe->opcode);
/* same numerical values with corresponding REQ_F_*, safe to copy */
req->flags = sqe_flags = READ_ONCE(sqe->flags);
req->user_data = READ_ONCE(sqe->user_data);
- req->async_data = NULL;
req->file = NULL;
- req->ctx = ctx;
- req->link = NULL;
req->fixed_rsrc_refs = NULL;
- /* one is dropped after submission, the other at completion */
- atomic_set(&req->refs, 2);
req->task = current;
- req->result = 0;
- req->work.creds = NULL;
/* enforce forwards compatibility on users */
if (unlikely(sqe_flags & ~SQE_VALID_FLAGS))
return -EINVAL;
if (unlikely(req->opcode >= IORING_OP_LAST))
return -EINVAL;
- if (unlikely(!io_check_restriction(ctx, req, sqe_flags)))
+ if (!io_check_restriction(ctx, req, sqe_flags))
return -EACCES;
if ((sqe_flags & IOSQE_BUFFER_SELECT) &&
!io_op_defs[req->opcode].buffer_select)
return -EOPNOTSUPP;
+ if (unlikely(sqe_flags & IOSQE_IO_DRAIN))
+ ctx->drain_active = true;
personality = READ_ONCE(sqe->personality);
if (personality) {
- req->work.creds = xa_load(&ctx->personalities, personality);
- if (!req->work.creds)
+ req->creds = xa_load(&ctx->personalities, personality);
+ if (!req->creds)
return -EINVAL;
- get_cred(req->work.creds);
+ get_cred(req->creds);
+ req->flags |= REQ_F_CREDS;
}
state = &ctx->submit_state;
}
if (io_op_defs[req->opcode].needs_file) {
- bool fixed = req->flags & REQ_F_FIXED_FILE;
-
- req->file = io_file_get(state, req, READ_ONCE(sqe->fd), fixed);
+ req->file = io_file_get(ctx, req, READ_ONCE(sqe->fd),
+ (sqe_flags & IOSQE_FIXED_FILE));
if (unlikely(!req->file))
ret = -EBADF;
}
static int io_submit_sqe(struct io_ring_ctx *ctx, struct io_kiocb *req,
const struct io_uring_sqe *sqe)
+ __must_hold(&ctx->uring_lock)
{
struct io_submit_link *link = &ctx->submit_state.link;
int ret;
ret = io_init_req(ctx, req, sqe);
if (unlikely(ret)) {
fail_req:
+ /* fail even hard links since we don't submit */
if (link->head) {
- /* fail even hard links since we don't submit */
- link->head->flags |= REQ_F_FAIL_LINK;
- io_req_complete_failed(link->head, -ECANCELED);
- link->head = NULL;
+ /*
+ * we can judge a link req is failed or cancelled by if
+ * REQ_F_FAIL is set, but the head is an exception since
+ * it may be set REQ_F_FAIL because of other req's failure
+ * so let's leverage req->result to distinguish if a head
+ * is set REQ_F_FAIL because of its failure or other req's
+ * failure so that we can set the correct ret code for it.
+ * init result here to avoid affecting the normal path.
+ */
+ if (!(link->head->flags & REQ_F_FAIL))
+ req_fail_link_node(link->head, -ECANCELED);
+ } else if (!(req->flags & (REQ_F_LINK | REQ_F_HARDLINK))) {
+ /*
+ * the current req is a normal req, we should return
+ * error and thus break the submittion loop.
+ */
+ io_req_complete_failed(req, ret);
+ return ret;
}
- io_req_complete_failed(req, ret);
- return ret;
+ req_fail_link_node(req, ret);
+ } else {
+ ret = io_req_prep(req, sqe);
+ if (unlikely(ret))
+ goto fail_req;
}
- ret = io_req_prep(req, sqe);
- if (unlikely(ret))
- goto fail_req;
/* don't need @sqe from now on */
- trace_io_uring_submit_sqe(ctx, req->opcode, req->user_data,
- true, ctx->flags & IORING_SETUP_SQPOLL);
+ trace_io_uring_submit_sqe(ctx, req, req->opcode, req->user_data,
+ req->flags, true,
+ ctx->flags & IORING_SETUP_SQPOLL);
/*
* If we already have a head request, queue this one for async
if (link->head) {
struct io_kiocb *head = link->head;
- /*
- * Taking sequential execution of a link, draining both sides
- * of the link also fullfils IOSQE_IO_DRAIN semantics for all
- * requests in the link. So, it drains the head and the
- * next after the link request. The last one is done via
- * drain_next flag to persist the effect across calls.
- */
- if (req->flags & REQ_F_IO_DRAIN) {
- head->flags |= REQ_F_IO_DRAIN;
- ctx->drain_next = 1;
+ if (!(req->flags & REQ_F_FAIL)) {
+ ret = io_req_prep_async(req);
+ if (unlikely(ret)) {
+ req_fail_link_node(req, ret);
+ if (!(head->flags & REQ_F_FAIL))
+ req_fail_link_node(head, -ECANCELED);
+ }
}
- ret = io_req_prep_async(req);
- if (unlikely(ret))
- goto fail_req;
trace_io_uring_link(ctx, req, head);
link->last->link = req;
link->last = req;
/* last request of a link, enqueue the link */
if (!(req->flags & (REQ_F_LINK | REQ_F_HARDLINK))) {
- io_queue_sqe(head);
link->head = NULL;
+ io_queue_sqe(head);
}
} else {
- if (unlikely(ctx->drain_next)) {
- req->flags |= REQ_F_IO_DRAIN;
- ctx->drain_next = 0;
- }
if (req->flags & (REQ_F_LINK | REQ_F_HARDLINK)) {
link->head = req;
link->last = req;
{
if (state->link.head)
io_queue_sqe(state->link.head);
- if (state->comp.nr)
- io_submit_flush_completions(&state->comp, ctx);
+ if (state->compl_nr)
+ io_submit_flush_completions(ctx);
if (state->plug_started)
blk_finish_plug(&state->plug);
- io_state_file_put(state);
}
/*
}
/*
- * Fetch an sqe, if one is available. Note that sqe_ptr will point to memory
+ * Fetch an sqe, if one is available. Note this returns a pointer to memory
* that is mapped by userspace. This means that care needs to be taken to
* ensure that reads are stable, as we cannot rely on userspace always
* being a good citizen. If members of the sqe are validated and then later
*/
static const struct io_uring_sqe *io_get_sqe(struct io_ring_ctx *ctx)
{
- u32 *sq_array = ctx->sq_array;
- unsigned head;
+ unsigned head, mask = ctx->sq_entries - 1;
+ unsigned sq_idx = ctx->cached_sq_head++ & mask;
/*
* The cached sq head (or cq tail) serves two purposes:
* 2) allows the kernel side to track the head on its own, even
* though the application is the one updating it.
*/
- head = READ_ONCE(sq_array[ctx->cached_sq_head++ & ctx->sq_mask]);
+ head = READ_ONCE(ctx->sq_array[sq_idx]);
if (likely(head < ctx->sq_entries))
return &ctx->sq_sqes[head];
/* drop invalid entries */
- ctx->cached_sq_dropped++;
- WRITE_ONCE(ctx->rings->sq_dropped, ctx->cached_sq_dropped);
+ ctx->cq_extra--;
+ WRITE_ONCE(ctx->rings->sq_dropped,
+ READ_ONCE(ctx->rings->sq_dropped) + 1);
return NULL;
}
static int io_submit_sqes(struct io_ring_ctx *ctx, unsigned int nr)
+ __must_hold(&ctx->uring_lock)
{
int submitted = 0;
/* make sure SQ entry isn't read before tail */
nr = min3(nr, ctx->sq_entries, io_sqring_entries(ctx));
-
if (!percpu_ref_tryget_many(&ctx->refs, nr))
return -EAGAIN;
+ io_get_task_refs(nr);
- percpu_counter_add(¤t->io_uring->inflight, nr);
- refcount_add(nr, ¤t->usage);
io_submit_state_start(&ctx->submit_state, nr);
-
while (submitted < nr) {
const struct io_uring_sqe *sqe;
struct io_kiocb *req;
}
sqe = io_get_sqe(ctx);
if (unlikely(!sqe)) {
- kmem_cache_free(req_cachep, req);
+ list_add(&req->inflight_entry, &ctx->submit_state.free_list);
break;
}
/* will complete beyond this point, count as submitted */
if (unlikely(submitted != nr)) {
int ref_used = (submitted == -EAGAIN) ? 0 : submitted;
- struct io_uring_task *tctx = current->io_uring;
int unused = nr - ref_used;
+ current->io_uring->cached_refs += unused;
percpu_ref_put_many(&ctx->refs, unused);
- percpu_counter_sub(&tctx->inflight, unused);
- put_task_struct_many(current, unused);
}
io_submit_state_end(&ctx->submit_state, ctx);
return submitted;
}
+static inline bool io_sqd_events_pending(struct io_sq_data *sqd)
+{
+ return READ_ONCE(sqd->state);
+}
+
static inline void io_ring_set_wakeup_flag(struct io_ring_ctx *ctx)
{
/* Tell userspace we may need a wakeup call */
- spin_lock_irq(&ctx->completion_lock);
- ctx->rings->sq_flags |= IORING_SQ_NEED_WAKEUP;
- spin_unlock_irq(&ctx->completion_lock);
+ spin_lock(&ctx->completion_lock);
+ WRITE_ONCE(ctx->rings->sq_flags,
+ ctx->rings->sq_flags | IORING_SQ_NEED_WAKEUP);
+ spin_unlock(&ctx->completion_lock);
}
static inline void io_ring_clear_wakeup_flag(struct io_ring_ctx *ctx)
{
- spin_lock_irq(&ctx->completion_lock);
- ctx->rings->sq_flags &= ~IORING_SQ_NEED_WAKEUP;
- spin_unlock_irq(&ctx->completion_lock);
+ spin_lock(&ctx->completion_lock);
+ WRITE_ONCE(ctx->rings->sq_flags,
+ ctx->rings->sq_flags & ~IORING_SQ_NEED_WAKEUP);
+ spin_unlock(&ctx->completion_lock);
}
static int __io_sq_thread(struct io_ring_ctx *ctx, bool cap_entries)
to_submit = io_sqring_entries(ctx);
/* if we're handling multiple rings, cap submit size for fairness */
- if (cap_entries && to_submit > 8)
- to_submit = 8;
+ if (cap_entries && to_submit > IORING_SQPOLL_CAP_ENTRIES_VALUE)
+ to_submit = IORING_SQPOLL_CAP_ENTRIES_VALUE;
if (!list_empty(&ctx->iopoll_list) || to_submit) {
unsigned nr_events = 0;
+ const struct cred *creds = NULL;
+
+ if (ctx->sq_creds != current_cred())
+ creds = override_creds(ctx->sq_creds);
mutex_lock(&ctx->uring_lock);
if (!list_empty(&ctx->iopoll_list))
!(ctx->flags & IORING_SETUP_R_DISABLED))
ret = io_submit_sqes(ctx, to_submit);
mutex_unlock(&ctx->uring_lock);
- }
- if (!io_sqring_full(ctx) && wq_has_sleeper(&ctx->sqo_sq_wait))
- wake_up(&ctx->sqo_sq_wait);
+ if (to_submit && wq_has_sleeper(&ctx->sqo_sq_wait))
+ wake_up(&ctx->sqo_sq_wait);
+ if (creds)
+ revert_creds(creds);
+ }
return ret;
}
sqd->sq_thread_idle = sq_thread_idle;
}
+static bool io_sqd_handle_event(struct io_sq_data *sqd)
+{
+ bool did_sig = false;
+ struct ksignal ksig;
+
+ if (test_bit(IO_SQ_THREAD_SHOULD_PARK, &sqd->state) ||
+ signal_pending(current)) {
+ mutex_unlock(&sqd->lock);
+ if (signal_pending(current))
+ did_sig = get_signal(&ksig);
+ cond_resched();
+ mutex_lock(&sqd->lock);
+ }
+ return did_sig || test_bit(IO_SQ_THREAD_SHOULD_STOP, &sqd->state);
+}
+
static int io_sq_thread(void *data)
{
struct io_sq_data *sqd = data;
current->flags |= PF_NO_SETAFFINITY;
mutex_lock(&sqd->lock);
- /* a user may had exited before the thread started */
- io_run_task_work_head(&sqd->park_task_work);
-
- while (!test_bit(IO_SQ_THREAD_SHOULD_STOP, &sqd->state)) {
- int ret;
- bool cap_entries, sqt_spin, needs_sched;
-
- if (test_bit(IO_SQ_THREAD_SHOULD_PARK, &sqd->state) ||
- signal_pending(current)) {
- bool did_sig = false;
-
- mutex_unlock(&sqd->lock);
- if (signal_pending(current)) {
- struct ksignal ksig;
+ while (1) {
+ bool cap_entries, sqt_spin = false;
- did_sig = get_signal(&ksig);
- }
- cond_resched();
- mutex_lock(&sqd->lock);
- io_run_task_work();
- io_run_task_work_head(&sqd->park_task_work);
- if (did_sig)
+ if (io_sqd_events_pending(sqd) || signal_pending(current)) {
+ if (io_sqd_handle_event(sqd))
break;
timeout = jiffies + sqd->sq_thread_idle;
- continue;
}
- sqt_spin = false;
+
cap_entries = !list_is_singular(&sqd->ctx_list);
list_for_each_entry(ctx, &sqd->ctx_list, sqd_list) {
- const struct cred *creds = NULL;
+ int ret = __io_sq_thread(ctx, cap_entries);
- if (ctx->sq_creds != current_cred())
- creds = override_creds(ctx->sq_creds);
- ret = __io_sq_thread(ctx, cap_entries);
- if (creds)
- revert_creds(creds);
if (!sqt_spin && (ret > 0 || !list_empty(&ctx->iopoll_list)))
sqt_spin = true;
}
+ if (io_run_task_work())
+ sqt_spin = true;
if (sqt_spin || !time_after(jiffies, timeout)) {
- io_run_task_work();
cond_resched();
if (sqt_spin)
timeout = jiffies + sqd->sq_thread_idle;
}
prepare_to_wait(&sqd->wait, &wait, TASK_INTERRUPTIBLE);
- if (!test_bit(IO_SQ_THREAD_SHOULD_PARK, &sqd->state)) {
- list_for_each_entry(ctx, &sqd->ctx_list, sqd_list)
- io_ring_set_wakeup_flag(ctx);
+ if (!io_sqd_events_pending(sqd) && !current->task_works) {
+ bool needs_sched = true;
- needs_sched = true;
list_for_each_entry(ctx, &sqd->ctx_list, sqd_list) {
+ io_ring_set_wakeup_flag(ctx);
+
if ((ctx->flags & IORING_SETUP_IOPOLL) &&
!list_empty_careful(&ctx->iopoll_list)) {
needs_sched = false;
}
finish_wait(&sqd->wait, &wait);
- io_run_task_work_head(&sqd->park_task_work);
timeout = jiffies + sqd->sq_thread_idle;
}
- io_uring_cancel_sqpoll(sqd);
+ io_uring_cancel_generic(true, sqd);
sqd->thread = NULL;
list_for_each_entry(ctx, &sqd->ctx_list, sqd_list)
io_ring_set_wakeup_flag(ctx);
io_run_task_work();
- io_run_task_work_head(&sqd->park_task_work);
mutex_unlock(&sqd->lock);
complete(&sqd->exited);
struct io_wait_queue {
struct wait_queue_entry wq;
struct io_ring_ctx *ctx;
- unsigned to_wait;
+ unsigned cq_tail;
unsigned nr_timeouts;
};
static inline bool io_should_wake(struct io_wait_queue *iowq)
{
struct io_ring_ctx *ctx = iowq->ctx;
+ int dist = ctx->cached_cq_tail - (int) iowq->cq_tail;
/*
* Wake up if we have enough events, or if a timeout occurred since we
* started waiting. For timeouts, we always want to return to userspace,
* regardless of event count.
*/
- return io_cqring_events(ctx) >= iowq->to_wait ||
- atomic_read(&ctx->cq_timeouts) != iowq->nr_timeouts;
+ return dist >= 0 || atomic_read(&ctx->cq_timeouts) != iowq->nr_timeouts;
}
static int io_wake_function(struct wait_queue_entry *curr, unsigned int mode,
* Cannot safely flush overflowed CQEs from here, ensure we wake up
* the task, and the next invocation will do it.
*/
- if (io_should_wake(iowq) || test_bit(0, &iowq->ctx->cq_check_overflow))
+ if (io_should_wake(iowq) || test_bit(0, &iowq->ctx->check_cq_overflow))
return autoremove_wake_function(curr, mode, wake_flags, key);
return -1;
}
if (ret || io_should_wake(iowq))
return ret;
/* let the caller flush overflows, retry */
- if (test_bit(0, &ctx->cq_check_overflow))
+ if (test_bit(0, &ctx->check_cq_overflow))
return 1;
*timeout = schedule_timeout(*timeout);
const sigset_t __user *sig, size_t sigsz,
struct __kernel_timespec __user *uts)
{
- struct io_wait_queue iowq = {
- .wq = {
- .private = current,
- .func = io_wake_function,
- .entry = LIST_HEAD_INIT(iowq.wq.entry),
- },
- .ctx = ctx,
- .to_wait = min_events,
- };
+ struct io_wait_queue iowq;
struct io_rings *rings = ctx->rings;
signed long timeout = MAX_SCHEDULE_TIMEOUT;
int ret;
do {
- io_cqring_overflow_flush(ctx, false);
+ io_cqring_overflow_flush(ctx);
if (io_cqring_events(ctx) >= min_events)
return 0;
if (!io_run_task_work())
timeout = timespec64_to_jiffies(&ts);
}
+ init_waitqueue_func_entry(&iowq.wq, io_wake_function);
+ iowq.wq.private = current;
+ INIT_LIST_HEAD(&iowq.wq.entry);
+ iowq.ctx = ctx;
iowq.nr_timeouts = atomic_read(&ctx->cq_timeouts);
+ iowq.cq_tail = READ_ONCE(ctx->rings->cq.head) + min_events;
+
trace_io_uring_cqring_wait(ctx, min_events);
do {
/* if we can't even flush overflow, don't wait for more */
- if (!io_cqring_overflow_flush(ctx, false)) {
+ if (!io_cqring_overflow_flush(ctx)) {
ret = -EBUSY;
break;
}
- prepare_to_wait_exclusive(&ctx->wait, &iowq.wq,
+ prepare_to_wait_exclusive(&ctx->cq_wait, &iowq.wq,
TASK_INTERRUPTIBLE);
ret = io_cqring_wait_schedule(ctx, &iowq, &timeout);
- finish_wait(&ctx->wait, &iowq.wq);
+ finish_wait(&ctx->cq_wait, &iowq.wq);
cond_resched();
} while (ret > 0);
return READ_ONCE(rings->cq.head) == READ_ONCE(rings->cq.tail) ? ret : 0;
}
-static void io_free_file_tables(struct io_file_table *table, unsigned nr_files)
+static void io_free_page_table(void **table, size_t size)
{
- unsigned i, nr_tables = DIV_ROUND_UP(nr_files, IORING_MAX_FILES_TABLE);
+ unsigned i, nr_tables = DIV_ROUND_UP(size, PAGE_SIZE);
for (i = 0; i < nr_tables; i++)
- kfree(table->files[i]);
- kfree(table->files);
- table->files = NULL;
+ kfree(table[i]);
+ kfree(table);
}
-static inline void io_rsrc_ref_lock(struct io_ring_ctx *ctx)
+static void **io_alloc_page_table(size_t size)
{
- spin_lock_bh(&ctx->rsrc_ref_lock);
-}
+ unsigned i, nr_tables = DIV_ROUND_UP(size, PAGE_SIZE);
+ size_t init_size = size;
+ void **table;
-static inline void io_rsrc_ref_unlock(struct io_ring_ctx *ctx)
-{
- spin_unlock_bh(&ctx->rsrc_ref_lock);
+ table = kcalloc(nr_tables, sizeof(*table), GFP_KERNEL_ACCOUNT);
+ if (!table)
+ return NULL;
+
+ for (i = 0; i < nr_tables; i++) {
+ unsigned int this_size = min_t(size_t, size, PAGE_SIZE);
+
+ table[i] = kzalloc(this_size, GFP_KERNEL_ACCOUNT);
+ if (!table[i]) {
+ io_free_page_table(table, init_size);
+ return NULL;
+ }
+ size -= this_size;
+ }
+ return table;
}
static void io_rsrc_node_destroy(struct io_rsrc_node *ref_node)
kfree(ref_node);
}
+static void io_rsrc_node_ref_zero(struct percpu_ref *ref)
+{
+ struct io_rsrc_node *node = container_of(ref, struct io_rsrc_node, refs);
+ struct io_ring_ctx *ctx = node->rsrc_data->ctx;
+ unsigned long flags;
+ bool first_add = false;
+
+ spin_lock_irqsave(&ctx->rsrc_ref_lock, flags);
+ node->done = true;
+
+ while (!list_empty(&ctx->rsrc_ref_list)) {
+ node = list_first_entry(&ctx->rsrc_ref_list,
+ struct io_rsrc_node, node);
+ /* recycle ref nodes in order */
+ if (!node->done)
+ break;
+ list_del(&node->node);
+ first_add |= llist_add(&node->llist, &ctx->rsrc_put_llist);
+ }
+ spin_unlock_irqrestore(&ctx->rsrc_ref_lock, flags);
+
+ if (first_add)
+ mod_delayed_work(system_wq, &ctx->rsrc_put_work, HZ);
+}
+
+static struct io_rsrc_node *io_rsrc_node_alloc(struct io_ring_ctx *ctx)
+{
+ struct io_rsrc_node *ref_node;
+
+ ref_node = kzalloc(sizeof(*ref_node), GFP_KERNEL);
+ if (!ref_node)
+ return NULL;
+
+ if (percpu_ref_init(&ref_node->refs, io_rsrc_node_ref_zero,
+ 0, GFP_KERNEL)) {
+ kfree(ref_node);
+ return NULL;
+ }
+ INIT_LIST_HEAD(&ref_node->node);
+ INIT_LIST_HEAD(&ref_node->rsrc_list);
+ ref_node->done = false;
+ return ref_node;
+}
+
static void io_rsrc_node_switch(struct io_ring_ctx *ctx,
struct io_rsrc_data *data_to_kill)
{
struct io_rsrc_node *rsrc_node = ctx->rsrc_node;
rsrc_node->rsrc_data = data_to_kill;
- io_rsrc_ref_lock(ctx);
+ spin_lock_irq(&ctx->rsrc_ref_lock);
list_add_tail(&rsrc_node->node, &ctx->rsrc_ref_list);
- io_rsrc_ref_unlock(ctx);
+ spin_unlock_irq(&ctx->rsrc_ref_lock);
atomic_inc(&data_to_kill->refs);
percpu_ref_kill(&rsrc_node->refs);
/* kill initial ref, already quiesced if zero */
if (atomic_dec_and_test(&data->refs))
break;
+ mutex_unlock(&ctx->uring_lock);
flush_delayed_work(&ctx->rsrc_put_work);
ret = wait_for_completion_interruptible(&data->done);
- if (!ret)
+ if (!ret) {
+ mutex_lock(&ctx->uring_lock);
break;
+ }
atomic_inc(&data->refs);
/* wait for all works potentially completing data->done */
flush_delayed_work(&ctx->rsrc_put_work);
reinit_completion(&data->done);
- mutex_unlock(&ctx->uring_lock);
ret = io_run_task_work_sig();
mutex_lock(&ctx->uring_lock);
} while (ret >= 0);
return ret;
}
+static u64 *io_get_tag_slot(struct io_rsrc_data *data, unsigned int idx)
+{
+ unsigned int off = idx & IO_RSRC_TAG_TABLE_MASK;
+ unsigned int table_idx = idx >> IO_RSRC_TAG_TABLE_SHIFT;
+
+ return &data->tags[table_idx][off];
+}
+
static void io_rsrc_data_free(struct io_rsrc_data *data)
{
- kvfree(data->tags);
+ size_t size = data->nr * sizeof(data->tags[0][0]);
+
+ if (data->tags)
+ io_free_page_table((void **)data->tags, size);
kfree(data);
}
-static struct io_rsrc_data *io_rsrc_data_alloc(struct io_ring_ctx *ctx,
- rsrc_put_fn *do_put,
- unsigned nr)
+static int io_rsrc_data_alloc(struct io_ring_ctx *ctx, rsrc_put_fn *do_put,
+ u64 __user *utags, unsigned nr,
+ struct io_rsrc_data **pdata)
{
struct io_rsrc_data *data;
+ int ret = -ENOMEM;
+ unsigned i;
data = kzalloc(sizeof(*data), GFP_KERNEL);
if (!data)
- return NULL;
-
- data->tags = kvcalloc(nr, sizeof(*data->tags), GFP_KERNEL);
+ return -ENOMEM;
+ data->tags = (u64 **)io_alloc_page_table(nr * sizeof(data->tags[0][0]));
if (!data->tags) {
kfree(data);
- return NULL;
+ return -ENOMEM;
}
- atomic_set(&data->refs, 1);
+ data->nr = nr;
data->ctx = ctx;
data->do_put = do_put;
+ if (utags) {
+ ret = -EFAULT;
+ for (i = 0; i < nr; i++) {
+ u64 *tag_slot = io_get_tag_slot(data, i);
+
+ if (copy_from_user(tag_slot, &utags[i],
+ sizeof(*tag_slot)))
+ goto fail;
+ }
+ }
+
+ atomic_set(&data->refs, 1);
init_completion(&data->done);
- return data;
+ *pdata = data;
+ return 0;
+fail:
+ io_rsrc_data_free(data);
+ return ret;
+}
+
+static bool io_alloc_file_tables(struct io_file_table *table, unsigned nr_files)
+{
+ table->files = kvcalloc(nr_files, sizeof(table->files[0]),
+ GFP_KERNEL_ACCOUNT);
+ return !!table->files;
+}
+
+static void io_free_file_tables(struct io_file_table *table)
+{
+ kvfree(table->files);
+ table->files = NULL;
}
static void __io_sqe_files_unregister(struct io_ring_ctx *ctx)
fput(file);
}
#endif
- io_free_file_tables(&ctx->file_table, ctx->nr_user_files);
+ io_free_file_tables(&ctx->file_table);
io_rsrc_data_free(ctx->file_data);
ctx->file_data = NULL;
ctx->nr_user_files = 0;
}
#endif
-static bool io_alloc_file_tables(struct io_file_table *table, unsigned nr_files)
-{
- unsigned i, nr_tables = DIV_ROUND_UP(nr_files, IORING_MAX_FILES_TABLE);
-
- table->files = kcalloc(nr_tables, sizeof(*table->files), GFP_KERNEL);
- if (!table->files)
- return false;
-
- for (i = 0; i < nr_tables; i++) {
- unsigned int this_files = min(nr_files, IORING_MAX_FILES_TABLE);
-
- table->files[i] = kcalloc(this_files, sizeof(*table->files[i]),
- GFP_KERNEL);
- if (!table->files[i])
- break;
- nr_files -= this_files;
- }
-
- if (i == nr_tables)
- return true;
-
- io_free_file_tables(table, nr_tables * IORING_MAX_FILES_TABLE);
- return false;
-}
-
static void io_rsrc_file_put(struct io_ring_ctx *ctx, struct io_rsrc_put *prsrc)
{
struct file *file = prsrc->file;
if (prsrc->tag) {
bool lock_ring = ctx->flags & IORING_SETUP_IOPOLL;
- unsigned long flags;
io_ring_submit_lock(ctx, lock_ring);
- spin_lock_irqsave(&ctx->completion_lock, flags);
+ spin_lock(&ctx->completion_lock);
io_cqring_fill_event(ctx, prsrc->tag, 0, 0);
ctx->cq_extra++;
io_commit_cqring(ctx);
- spin_unlock_irqrestore(&ctx->completion_lock, flags);
+ spin_unlock(&ctx->completion_lock);
io_cqring_ev_posted(ctx);
io_ring_submit_unlock(ctx, lock_ring);
}
rsrc_data->do_put(ctx, prsrc);
- kfree(prsrc);
- }
-
- io_rsrc_node_destroy(ref_node);
- if (atomic_dec_and_test(&rsrc_data->refs))
- complete(&rsrc_data->done);
-}
-
-static void io_rsrc_put_work(struct work_struct *work)
-{
- struct io_ring_ctx *ctx;
- struct llist_node *node;
-
- ctx = container_of(work, struct io_ring_ctx, rsrc_put_work.work);
- node = llist_del_all(&ctx->rsrc_put_llist);
-
- while (node) {
- struct io_rsrc_node *ref_node;
- struct llist_node *next = node->next;
-
- ref_node = llist_entry(node, struct io_rsrc_node, llist);
- __io_rsrc_put_work(ref_node);
- node = next;
- }
-}
-
-static void io_rsrc_node_ref_zero(struct percpu_ref *ref)
-{
- struct io_rsrc_node *node = container_of(ref, struct io_rsrc_node, refs);
- struct io_ring_ctx *ctx = node->rsrc_data->ctx;
- bool first_add = false;
-
- io_rsrc_ref_lock(ctx);
- node->done = true;
-
- while (!list_empty(&ctx->rsrc_ref_list)) {
- node = list_first_entry(&ctx->rsrc_ref_list,
- struct io_rsrc_node, node);
- /* recycle ref nodes in order */
- if (!node->done)
- break;
- list_del(&node->node);
- first_add |= llist_add(&node->llist, &ctx->rsrc_put_llist);
+ kfree(prsrc);
}
- io_rsrc_ref_unlock(ctx);
- if (first_add)
- mod_delayed_work(system_wq, &ctx->rsrc_put_work, HZ);
+ io_rsrc_node_destroy(ref_node);
+ if (atomic_dec_and_test(&rsrc_data->refs))
+ complete(&rsrc_data->done);
}
-static struct io_rsrc_node *io_rsrc_node_alloc(struct io_ring_ctx *ctx)
+static void io_rsrc_put_work(struct work_struct *work)
{
- struct io_rsrc_node *ref_node;
+ struct io_ring_ctx *ctx;
+ struct llist_node *node;
- ref_node = kzalloc(sizeof(*ref_node), GFP_KERNEL);
- if (!ref_node)
- return NULL;
+ ctx = container_of(work, struct io_ring_ctx, rsrc_put_work.work);
+ node = llist_del_all(&ctx->rsrc_put_llist);
- if (percpu_ref_init(&ref_node->refs, io_rsrc_node_ref_zero,
- 0, GFP_KERNEL)) {
- kfree(ref_node);
- return NULL;
+ while (node) {
+ struct io_rsrc_node *ref_node;
+ struct llist_node *next = node->next;
+
+ ref_node = llist_entry(node, struct io_rsrc_node, llist);
+ __io_rsrc_put_work(ref_node);
+ node = next;
}
- INIT_LIST_HEAD(&ref_node->node);
- INIT_LIST_HEAD(&ref_node->rsrc_list);
- ref_node->done = false;
- return ref_node;
}
static int io_sqe_files_register(struct io_ring_ctx *ctx, void __user *arg,
struct file *file;
int fd, ret;
unsigned i;
- struct io_rsrc_data *file_data;
if (ctx->file_data)
return -EBUSY;
return -EINVAL;
if (nr_args > IORING_MAX_FIXED_FILES)
return -EMFILE;
+ if (nr_args > rlimit(RLIMIT_NOFILE))
+ return -EMFILE;
ret = io_rsrc_node_switch_start(ctx);
if (ret)
return ret;
+ ret = io_rsrc_data_alloc(ctx, io_rsrc_file_put, tags, nr_args,
+ &ctx->file_data);
+ if (ret)
+ return ret;
- file_data = io_rsrc_data_alloc(ctx, io_rsrc_file_put, nr_args);
- if (!file_data)
- return -ENOMEM;
- ctx->file_data = file_data;
ret = -ENOMEM;
if (!io_alloc_file_tables(&ctx->file_table, nr_args))
goto out_free;
for (i = 0; i < nr_args; i++, ctx->nr_user_files++) {
- u64 tag = 0;
-
- if ((tags && copy_from_user(&tag, &tags[i], sizeof(tag))) ||
- copy_from_user(&fd, &fds[i], sizeof(fd))) {
+ if (copy_from_user(&fd, &fds[i], sizeof(fd))) {
ret = -EFAULT;
goto out_fput;
}
/* allow sparse sets */
if (fd == -1) {
ret = -EINVAL;
- if (unlikely(tag))
+ if (unlikely(*io_get_tag_slot(ctx->file_data, i)))
goto out_fput;
continue;
}
fput(file);
goto out_fput;
}
- ctx->file_data->tags[i] = tag;
io_fixed_file_set(io_fixed_file_slot(&ctx->file_table, i), file);
}
if (file)
fput(file);
}
- io_free_file_tables(&ctx->file_table, nr_args);
+ io_free_file_tables(&ctx->file_table);
ctx->nr_user_files = 0;
out_free:
io_rsrc_data_free(ctx->file_data);
#endif
}
+static int io_install_fixed_file(struct io_kiocb *req, struct file *file,
+ unsigned int issue_flags, u32 slot_index)
+{
+ struct io_ring_ctx *ctx = req->ctx;
+ bool force_nonblock = issue_flags & IO_URING_F_NONBLOCK;
+ struct io_fixed_file *file_slot;
+ int ret = -EBADF;
+
+ io_ring_submit_lock(ctx, !force_nonblock);
+ if (file->f_op == &io_uring_fops)
+ goto err;
+ ret = -ENXIO;
+ if (!ctx->file_data)
+ goto err;
+ ret = -EINVAL;
+ if (slot_index >= ctx->nr_user_files)
+ goto err;
+
+ slot_index = array_index_nospec(slot_index, ctx->nr_user_files);
+ file_slot = io_fixed_file_slot(&ctx->file_table, slot_index);
+ ret = -EBADF;
+ if (file_slot->file_ptr)
+ goto err;
+
+ *io_get_tag_slot(ctx->file_data, slot_index) = 0;
+ io_fixed_file_set(file_slot, file);
+ ret = io_sqe_file_register(ctx, file, slot_index);
+ if (ret) {
+ file_slot->file_ptr = 0;
+ goto err;
+ }
+
+ ret = 0;
+err:
+ io_ring_submit_unlock(ctx, !force_nonblock);
+ if (ret)
+ fput(file);
+ return ret;
+}
+
static int io_queue_rsrc_removal(struct io_rsrc_data *data, unsigned idx,
struct io_rsrc_node *node, void *rsrc)
{
if (!prsrc)
return -ENOMEM;
- prsrc->tag = data->tags[idx];
+ prsrc->tag = *io_get_tag_slot(data, idx);
prsrc->rsrc = rsrc;
list_add(&prsrc->list, &node->rsrc_list);
return 0;
err = -EBADF;
break;
}
- data->tags[up->offset + done] = tag;
+ *io_get_tag_slot(data, up->offset + done) = tag;
io_fixed_file_set(file_slot, file);
err = io_sqe_file_register(ctx, file, i);
if (err) {
return done ? done : err;
}
-static struct io_wq_work *io_free_work(struct io_wq_work *work)
-{
- struct io_kiocb *req = container_of(work, struct io_kiocb, work);
-
- req = io_put_req_find_next(req);
- return req ? &req->work : NULL;
-}
-
static struct io_wq *io_init_wq_offload(struct io_ring_ctx *ctx,
struct task_struct *task)
{
struct io_wq_data data;
unsigned int concurrency;
+ mutex_lock(&ctx->uring_lock);
hash = ctx->hash_map;
if (!hash) {
hash = kzalloc(sizeof(*hash), GFP_KERNEL);
- if (!hash)
+ if (!hash) {
+ mutex_unlock(&ctx->uring_lock);
return ERR_PTR(-ENOMEM);
+ }
refcount_set(&hash->refs, 1);
init_waitqueue_head(&hash->wait);
ctx->hash_map = hash;
}
+ mutex_unlock(&ctx->uring_lock);
data.hash = hash;
data.task = task;
- data.free_work = io_free_work;
+ data.free_work = io_wq_free_work;
data.do_work = io_wq_submit_work;
/* Do QD, or 4 * CPUS, whatever is smallest */
struct io_uring_task *tctx;
int ret;
- tctx = kmalloc(sizeof(*tctx), GFP_KERNEL);
+ tctx = kzalloc(sizeof(*tctx), GFP_KERNEL);
if (unlikely(!tctx))
return -ENOMEM;
xa_init(&tctx->xa);
init_waitqueue_head(&tctx->wait);
- tctx->last = NULL;
atomic_set(&tctx->in_idle, 0);
atomic_set(&tctx->inflight_tracked, 0);
task->io_uring = tctx;
spin_lock_init(&tctx->task_lock);
INIT_WQ_LIST(&tctx->task_list);
- tctx->task_state = 0;
init_task_work(&tctx->task_work, tctx_task_work);
return 0;
}
WARN_ON_ONCE(!xa_empty(&tctx->xa));
WARN_ON_ONCE(tctx->io_wq);
+ WARN_ON_ONCE(tctx->cached_refs);
percpu_counter_destroy(&tctx->inflight);
kfree(tctx);
f = fdget(p->wq_fd);
if (!f.file)
return -ENXIO;
- fdput(f);
- if (f.file->f_op != &io_uring_fops)
+ if (f.file->f_op != &io_uring_fops) {
+ fdput(f);
return -EINVAL;
+ }
+ fdput(f);
}
if (ctx->flags & IORING_SETUP_SQPOLL) {
struct task_struct *tsk;
for (i = 0; i < nr_pages; i++) {
struct vm_area_struct *vma = vmas[i];
+ if (vma_is_shmem(vma))
+ continue;
if (vma->vm_file &&
!is_file_hugepages(vma->vm_file)) {
ret = -EOPNOTSUPP;
ret = io_rsrc_node_switch_start(ctx);
if (ret)
return ret;
- data = io_rsrc_data_alloc(ctx, io_rsrc_buf_put, nr_args);
- if (!data)
- return -ENOMEM;
+ ret = io_rsrc_data_alloc(ctx, io_rsrc_buf_put, tags, nr_args, &data);
+ if (ret)
+ return ret;
ret = io_buffers_map_alloc(ctx, nr_args);
if (ret) {
io_rsrc_data_free(data);
}
for (i = 0; i < nr_args; i++, ctx->nr_user_bufs++) {
- u64 tag = 0;
-
- if (tags && copy_from_user(&tag, &tags[i], sizeof(tag))) {
- ret = -EFAULT;
- break;
- }
ret = io_copy_iov(ctx, &iov, arg, i);
if (ret)
break;
ret = io_buffer_validate(&iov);
if (ret)
break;
- if (!iov.iov_base && tag) {
+ if (!iov.iov_base && *io_get_tag_slot(data, i)) {
ret = -EINVAL;
break;
}
&last_hpage);
if (ret)
break;
- data->tags[i] = tag;
}
WARN_ON_ONCE(ctx->buf_data);
}
ctx->user_bufs[i] = imu;
- ctx->buf_data->tags[offset] = tag;
+ *io_get_tag_slot(ctx->buf_data, offset) = tag;
}
if (needs_switch)
ctx->cq_ev_fd = eventfd_ctx_fdget(fd);
if (IS_ERR(ctx->cq_ev_fd)) {
int ret = PTR_ERR(ctx->cq_ev_fd);
+
ctx->cq_ev_fd = NULL;
return ret;
}
__io_remove_buffers(ctx, buf, index, -1U);
}
-static void io_req_cache_free(struct list_head *list, struct task_struct *tsk)
+static void io_req_cache_free(struct list_head *list)
{
struct io_kiocb *req, *nxt;
- list_for_each_entry_safe(req, nxt, list, compl.list) {
- if (tsk && req->task != tsk)
- continue;
- list_del(&req->compl.list);
+ list_for_each_entry_safe(req, nxt, list, inflight_entry) {
+ list_del(&req->inflight_entry);
kmem_cache_free(req_cachep, req);
}
}
static void io_req_caches_free(struct io_ring_ctx *ctx)
{
- struct io_submit_state *submit_state = &ctx->submit_state;
- struct io_comp_state *cs = &ctx->submit_state.comp;
+ struct io_submit_state *state = &ctx->submit_state;
mutex_lock(&ctx->uring_lock);
- if (submit_state->free_reqs) {
- kmem_cache_free_bulk(req_cachep, submit_state->free_reqs,
- submit_state->reqs);
- submit_state->free_reqs = 0;
+ if (state->free_reqs) {
+ kmem_cache_free_bulk(req_cachep, state->free_reqs, state->reqs);
+ state->free_reqs = 0;
}
- io_flush_cached_locked_reqs(ctx, cs);
- io_req_cache_free(&cs->free_list, NULL);
+ io_flush_cached_locked_reqs(ctx, state);
+ io_req_cache_free(&state->free_list);
mutex_unlock(&ctx->uring_lock);
}
-static bool io_wait_rsrc_data(struct io_rsrc_data *data)
+static void io_wait_rsrc_data(struct io_rsrc_data *data)
{
- if (!data)
- return false;
- if (!atomic_dec_and_test(&data->refs))
+ if (data && !atomic_dec_and_test(&data->refs))
wait_for_completion(&data->done);
- return true;
}
static void io_ring_ctx_free(struct io_ring_ctx *ctx)
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);
+
mutex_lock(&ctx->uring_lock);
- if (io_wait_rsrc_data(ctx->buf_data))
+ if (ctx->buf_data)
__io_sqe_buffers_unregister(ctx);
- if (io_wait_rsrc_data(ctx->file_data))
+ if (ctx->file_data)
__io_sqe_files_unregister(ctx);
if (ctx->rings)
__io_cqring_overflow_flush(ctx, true);
sock_release(ctx->ring_sock);
}
#endif
+ WARN_ON_ONCE(!list_empty(&ctx->ltimeout_list));
io_mem_free(ctx->rings);
io_mem_free(ctx->sq_sqes);
struct io_ring_ctx *ctx = file->private_data;
__poll_t mask = 0;
- poll_wait(file, &ctx->cq_wait, wait);
+ poll_wait(file, &ctx->poll_wait, wait);
/*
* synchronizes with barrier from wq_has_sleeper call in
* io_commit_cqring
* Users may get EPOLLIN meanwhile seeing nothing in cqring, this
* pushs them to do the flush.
*/
- if (io_cqring_events(ctx) || test_bit(0, &ctx->cq_check_overflow))
+ if (io_cqring_events(ctx) || test_bit(0, &ctx->check_cq_overflow))
mask |= EPOLLIN | EPOLLRDNORM;
return mask;
return -EINVAL;
}
-static inline bool io_run_ctx_fallback(struct io_ring_ctx *ctx)
-{
- return io_run_task_work_head(&ctx->exit_task_work);
-}
-
struct io_tctx_exit {
struct callback_head task_work;
struct completion completion;
* node. It'll be removed by the end of cancellation, just ignore it.
*/
if (!atomic_read(&tctx->in_idle))
- io_uring_del_task_file((unsigned long)work->ctx);
+ io_uring_del_tctx_node((unsigned long)work->ctx);
complete(&work->completion);
}
{
struct io_ring_ctx *ctx = container_of(work, struct io_ring_ctx, exit_work);
unsigned long timeout = jiffies + HZ * 60 * 5;
+ unsigned long interval = HZ / 20;
struct io_tctx_exit exit;
struct io_tctx_node *node;
int ret;
* as nobody else will be looking for them.
*/
do {
- io_uring_try_cancel_requests(ctx, NULL, NULL);
+ io_uring_try_cancel_requests(ctx, NULL, true);
if (ctx->sq_data) {
struct io_sq_data *sqd = ctx->sq_data;
struct task_struct *tsk;
io_sq_thread_unpark(sqd);
}
- WARN_ON_ONCE(time_after(jiffies, timeout));
- } while (!wait_for_completion_timeout(&ctx->ref_comp, HZ/20));
+ if (WARN_ON_ONCE(time_after(jiffies, timeout))) {
+ /* there is little hope left, don't run it too often */
+ interval = HZ * 60;
+ }
+ } while (!wait_for_completion_timeout(&ctx->ref_comp, interval));
init_completion(&exit.completion);
init_task_work(&exit.task_work, io_tctx_exit_cb);
/*
* Some may use context even when all refs and requests have been put,
* and they are free to do so while still holding uring_lock or
- * completion_lock, see __io_req_task_submit(). Apart from other work,
+ * completion_lock, see io_req_task_submit(). Apart from other work,
* this lock/unlock section also waits them to finish.
*/
mutex_lock(&ctx->uring_lock);
mutex_lock(&ctx->uring_lock);
}
mutex_unlock(&ctx->uring_lock);
- spin_lock_irq(&ctx->completion_lock);
- spin_unlock_irq(&ctx->completion_lock);
+ spin_lock(&ctx->completion_lock);
+ spin_unlock(&ctx->completion_lock);
io_ring_ctx_free(ctx);
}
/* Returns true if we found and killed one or more timeouts */
static bool io_kill_timeouts(struct io_ring_ctx *ctx, struct task_struct *tsk,
- struct files_struct *files)
+ bool cancel_all)
{
struct io_kiocb *req, *tmp;
int canceled = 0;
- spin_lock_irq(&ctx->completion_lock);
+ spin_lock(&ctx->completion_lock);
+ spin_lock_irq(&ctx->timeout_lock);
list_for_each_entry_safe(req, tmp, &ctx->timeout_list, timeout.list) {
- if (io_match_task(req, tsk, files)) {
+ if (io_match_task(req, tsk, cancel_all)) {
io_kill_timeout(req, -ECANCELED);
canceled++;
}
}
+ spin_unlock_irq(&ctx->timeout_lock);
if (canceled != 0)
io_commit_cqring(ctx);
- spin_unlock_irq(&ctx->completion_lock);
+ spin_unlock(&ctx->completion_lock);
if (canceled != 0)
io_cqring_ev_posted(ctx);
return canceled != 0;
io_unregister_personality(ctx, index);
mutex_unlock(&ctx->uring_lock);
- io_kill_timeouts(ctx, NULL, NULL);
- io_poll_remove_all(ctx, NULL, NULL);
+ io_kill_timeouts(ctx, NULL, true);
+ io_poll_remove_all(ctx, NULL, true);
/* if we failed setting up the ctx, we might not have any rings */
io_iopoll_try_reap_events(ctx);
struct io_task_cancel {
struct task_struct *task;
- struct files_struct *files;
+ bool all;
};
static bool io_cancel_task_cb(struct io_wq_work *work, void *data)
struct io_task_cancel *cancel = data;
bool ret;
- if (cancel->files && (req->flags & REQ_F_LINK_TIMEOUT)) {
- unsigned long flags;
+ if (!cancel->all && (req->flags & REQ_F_LINK_TIMEOUT)) {
struct io_ring_ctx *ctx = req->ctx;
/* protect against races with linked timeouts */
- spin_lock_irqsave(&ctx->completion_lock, flags);
- ret = io_match_task(req, cancel->task, cancel->files);
- spin_unlock_irqrestore(&ctx->completion_lock, flags);
+ 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->files);
+ ret = io_match_task(req, cancel->task, cancel->all);
}
return ret;
}
static bool io_cancel_defer_files(struct io_ring_ctx *ctx,
- struct task_struct *task,
- struct files_struct *files)
+ struct task_struct *task, bool cancel_all)
{
struct io_defer_entry *de;
LIST_HEAD(list);
- spin_lock_irq(&ctx->completion_lock);
+ spin_lock(&ctx->completion_lock);
list_for_each_entry_reverse(de, &ctx->defer_list, list) {
- if (io_match_task(de->req, task, files)) {
+ if (io_match_task(de->req, task, cancel_all)) {
list_cut_position(&list, &ctx->defer_list, &de->list);
break;
}
}
- spin_unlock_irq(&ctx->completion_lock);
+ spin_unlock(&ctx->completion_lock);
if (list_empty(&list))
return false;
static void io_uring_try_cancel_requests(struct io_ring_ctx *ctx,
struct task_struct *task,
- struct files_struct *files)
+ bool cancel_all)
{
- struct io_task_cancel cancel = { .task = task, .files = files, };
+ struct io_task_cancel cancel = { .task = task, .all = cancel_all, };
struct io_uring_task *tctx = task ? task->io_uring : NULL;
while (1) {
}
/* SQPOLL thread does its own polling */
- if ((!(ctx->flags & IORING_SETUP_SQPOLL) && !files) ||
+ if ((!(ctx->flags & IORING_SETUP_SQPOLL) && cancel_all) ||
(ctx->sq_data && ctx->sq_data->thread == current)) {
while (!list_empty_careful(&ctx->iopoll_list)) {
io_iopoll_try_reap_events(ctx);
}
}
- ret |= io_cancel_defer_files(ctx, task, files);
- ret |= io_poll_remove_all(ctx, task, files);
- ret |= io_kill_timeouts(ctx, task, files);
- ret |= io_run_task_work();
- ret |= io_run_ctx_fallback(ctx);
+ ret |= io_cancel_defer_files(ctx, task, cancel_all);
+ ret |= io_poll_remove_all(ctx, task, cancel_all);
+ ret |= io_kill_timeouts(ctx, task, cancel_all);
+ if (task)
+ ret |= io_run_task_work();
if (!ret)
break;
cond_resched();
}
}
-static int __io_uring_add_task_file(struct io_ring_ctx *ctx)
+static int __io_uring_add_tctx_node(struct io_ring_ctx *ctx)
{
struct io_uring_task *tctx = current->io_uring;
struct io_tctx_node *node;
/*
* Note that this task has used io_uring. We use it for cancelation purposes.
*/
-static inline int io_uring_add_task_file(struct io_ring_ctx *ctx)
+static inline int io_uring_add_tctx_node(struct io_ring_ctx *ctx)
{
struct io_uring_task *tctx = current->io_uring;
if (likely(tctx && tctx->last == ctx))
return 0;
- return __io_uring_add_task_file(ctx);
+ return __io_uring_add_tctx_node(ctx);
}
/*
* Remove this io_uring_file -> task mapping.
*/
-static void io_uring_del_task_file(unsigned long index)
+static void io_uring_del_tctx_node(unsigned long index)
{
struct io_uring_task *tctx = current->io_uring;
struct io_tctx_node *node;
unsigned long index;
xa_for_each(&tctx->xa, index, node)
- io_uring_del_task_file(index);
+ io_uring_del_tctx_node(index);
if (wq) {
/*
* Must be after io_uring_del_task_file() (removes nodes under
* uring_lock) to avoid race with io_uring_try_cancel_iowq().
*/
- tctx->io_wq = NULL;
io_wq_put_and_exit(wq);
+ tctx->io_wq = NULL;
}
}
return percpu_counter_sum(&tctx->inflight);
}
-static void io_uring_try_cancel(struct files_struct *files)
+static void io_uring_drop_tctx_refs(struct task_struct *task)
{
- struct io_uring_task *tctx = current->io_uring;
- struct io_tctx_node *node;
- unsigned long index;
-
- xa_for_each(&tctx->xa, index, node) {
- struct io_ring_ctx *ctx = node->ctx;
+ struct io_uring_task *tctx = task->io_uring;
+ unsigned int refs = tctx->cached_refs;
- /* sqpoll task will cancel all its requests */
- if (!ctx->sq_data)
- io_uring_try_cancel_requests(ctx, current, files);
+ if (refs) {
+ tctx->cached_refs = 0;
+ percpu_counter_sub(&tctx->inflight, refs);
+ put_task_struct_many(task, refs);
}
}
-/* should only be called by SQPOLL task */
-static void io_uring_cancel_sqpoll(struct io_sq_data *sqd)
+/*
+ * 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.
+ */
+static void io_uring_cancel_generic(bool cancel_all, struct io_sq_data *sqd)
{
struct io_uring_task *tctx = current->io_uring;
struct io_ring_ctx *ctx;
s64 inflight;
DEFINE_WAIT(wait);
+ WARN_ON_ONCE(sqd && sqd->thread != current);
+
if (!current->io_uring)
return;
if (tctx->io_wq)
io_wq_exit_start(tctx->io_wq);
- WARN_ON_ONCE(!sqd || sqd->thread != current);
-
atomic_inc(&tctx->in_idle);
do {
+ io_uring_drop_tctx_refs(current);
/* read completions before cancelations */
- inflight = tctx_inflight(tctx, false);
+ inflight = tctx_inflight(tctx, !cancel_all);
if (!inflight)
break;
- list_for_each_entry(ctx, &sqd->ctx_list, sqd_list)
- io_uring_try_cancel_requests(ctx, current, NULL);
-
- prepare_to_wait(&tctx->wait, &wait, TASK_UNINTERRUPTIBLE);
- /*
- * If we've seen completions, retry without waiting. This
- * avoids a race where a completion comes in before we did
- * prepare_to_wait().
- */
- if (inflight == tctx_inflight(tctx, false))
- schedule();
- finish_wait(&tctx->wait, &wait);
- } while (1);
- atomic_dec(&tctx->in_idle);
-}
-/*
- * Find any io_uring fd that this task has registered or done IO on, and cancel
- * requests.
- */
-void __io_uring_cancel(struct files_struct *files)
-{
- struct io_uring_task *tctx = current->io_uring;
- DEFINE_WAIT(wait);
- s64 inflight;
+ if (!sqd) {
+ struct io_tctx_node *node;
+ unsigned long index;
- if (tctx->io_wq)
- io_wq_exit_start(tctx->io_wq);
+ xa_for_each(&tctx->xa, index, node) {
+ /* sqpoll task will cancel all its requests */
+ if (node->ctx->sq_data)
+ continue;
+ io_uring_try_cancel_requests(node->ctx, current,
+ cancel_all);
+ }
+ } else {
+ list_for_each_entry(ctx, &sqd->ctx_list, sqd_list)
+ io_uring_try_cancel_requests(ctx, current,
+ cancel_all);
+ }
- /* make sure overflow events are dropped */
- atomic_inc(&tctx->in_idle);
- do {
- /* read completions before cancelations */
- inflight = tctx_inflight(tctx, !!files);
- if (!inflight)
- break;
- io_uring_try_cancel(files);
prepare_to_wait(&tctx->wait, &wait, TASK_UNINTERRUPTIBLE);
-
+ 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
* prepare_to_wait().
*/
- if (inflight == tctx_inflight(tctx, !!files))
+ if (inflight == tctx_inflight(tctx, !cancel_all))
schedule();
finish_wait(&tctx->wait, &wait);
} while (1);
atomic_dec(&tctx->in_idle);
io_uring_clean_tctx(tctx);
- if (!files) {
+ if (cancel_all) {
/* for exec all current's requests should be gone, kill tctx */
__io_uring_free(current);
}
}
+void __io_uring_cancel(bool cancel_all)
+{
+ io_uring_cancel_generic(cancel_all, NULL);
+}
+
static void *io_uring_validate_mmap_request(struct file *file,
loff_t pgoff, size_t sz)
{
*/
ret = 0;
if (ctx->flags & IORING_SETUP_SQPOLL) {
- io_cqring_overflow_flush(ctx, false);
+ io_cqring_overflow_flush(ctx);
- ret = -EOWNERDEAD;
if (unlikely(ctx->sq_data->thread == NULL)) {
+ ret = -EOWNERDEAD;
goto out;
}
if (flags & IORING_ENTER_SQ_WAKEUP)
}
submitted = to_submit;
} else if (to_submit) {
- ret = io_uring_add_task_file(ctx);
+ ret = io_uring_add_tctx_node(ctx);
if (unlikely(ret))
goto out;
mutex_lock(&ctx->uring_lock);
io_uring_show_cred(m, index, cred);
}
seq_printf(m, "PollList:\n");
- spin_lock_irq(&ctx->completion_lock);
+ spin_lock(&ctx->completion_lock);
for (i = 0; i < (1U << ctx->cancel_hash_bits); i++) {
struct hlist_head *list = &ctx->cancel_hash[i];
struct io_kiocb *req;
seq_printf(m, " op=%d, task_works=%d\n", req->opcode,
req->task->task_works != NULL);
}
- spin_unlock_irq(&ctx->completion_lock);
+ spin_unlock(&ctx->completion_lock);
if (has_lock)
mutex_unlock(&ctx->uring_lock);
}
rings->cq_ring_mask = p->cq_entries - 1;
rings->sq_ring_entries = p->sq_entries;
rings->cq_ring_entries = p->cq_entries;
- ctx->sq_mask = rings->sq_ring_mask;
- ctx->cq_mask = rings->cq_ring_mask;
size = array_size(sizeof(struct io_uring_sqe), p->sq_entries);
if (size == SIZE_MAX) {
if (fd < 0)
return fd;
- ret = io_uring_add_task_file(ctx);
+ ret = io_uring_add_tctx_node(ctx);
if (ret) {
put_unused_fd(fd);
return ret;
ret = xa_alloc_cyclic(&ctx->personalities, &id, (void *)creds,
XA_LIMIT(0, USHRT_MAX), &ctx->pers_next, GFP_KERNEL);
- if (!ret)
- return id;
- put_cred(creds);
- return ret;
+ if (ret < 0) {
+ put_cred(creds);
+ return ret;
+ }
+ return id;
}
static int io_register_restrictions(struct io_ring_ctx *ctx, void __user *arg,
return -EINVAL;
}
+static int io_register_iowq_aff(struct io_ring_ctx *ctx, void __user *arg,
+ unsigned len)
+{
+ struct io_uring_task *tctx = current->io_uring;
+ cpumask_var_t new_mask;
+ int ret;
+
+ if (!tctx || !tctx->io_wq)
+ return -EINVAL;
+
+ if (!alloc_cpumask_var(&new_mask, GFP_KERNEL))
+ return -ENOMEM;
+
+ cpumask_clear(new_mask);
+ if (len > cpumask_size())
+ len = cpumask_size();
+
+ if (copy_from_user(new_mask, arg, len)) {
+ free_cpumask_var(new_mask);
+ return -EFAULT;
+ }
+
+ ret = io_wq_cpu_affinity(tctx->io_wq, new_mask);
+ free_cpumask_var(new_mask);
+ return ret;
+}
+
+static int io_unregister_iowq_aff(struct io_ring_ctx *ctx)
+{
+ struct io_uring_task *tctx = current->io_uring;
+
+ if (!tctx || !tctx->io_wq)
+ return -EINVAL;
+
+ return io_wq_cpu_affinity(tctx->io_wq, NULL);
+}
+
+static int io_register_iowq_max_workers(struct io_ring_ctx *ctx,
+ void __user *arg)
+{
+ struct io_uring_task *tctx = NULL;
+ struct io_sq_data *sqd = NULL;
+ __u32 new_count[2];
+ int i, ret;
+
+ if (copy_from_user(new_count, arg, sizeof(new_count)))
+ return -EFAULT;
+ for (i = 0; i < ARRAY_SIZE(new_count); i++)
+ if (new_count[i] > INT_MAX)
+ return -EINVAL;
+
+ if (ctx->flags & IORING_SETUP_SQPOLL) {
+ sqd = ctx->sq_data;
+ if (sqd) {
+ /*
+ * Observe the correct sqd->lock -> ctx->uring_lock
+ * ordering. Fine to drop uring_lock here, we hold
+ * a ref to the ctx.
+ */
+ mutex_unlock(&ctx->uring_lock);
+ mutex_lock(&sqd->lock);
+ mutex_lock(&ctx->uring_lock);
+ tctx = sqd->thread->io_uring;
+ }
+ } else {
+ tctx = current->io_uring;
+ }
+
+ ret = -EINVAL;
+ if (!tctx || !tctx->io_wq)
+ goto err;
+
+ ret = io_wq_max_workers(tctx->io_wq, new_count);
+ if (ret)
+ goto err;
+
+ if (sqd)
+ mutex_unlock(&sqd->lock);
+
+ if (copy_to_user(arg, new_count, sizeof(new_count)))
+ return -EFAULT;
+
+ return 0;
+err:
+ if (sqd)
+ mutex_unlock(&sqd->lock);
+ return ret;
+}
+
static bool io_register_op_must_quiesce(int op)
{
switch (op) {
case IORING_REGISTER_FILES_UPDATE2:
case IORING_REGISTER_BUFFERS2:
case IORING_REGISTER_BUFFERS_UPDATE:
+ case IORING_REGISTER_IOWQ_AFF:
+ case IORING_UNREGISTER_IOWQ_AFF:
+ case IORING_REGISTER_IOWQ_MAX_WORKERS:
return false;
default:
return true;
}
}
+static int io_ctx_quiesce(struct io_ring_ctx *ctx)
+{
+ long ret;
+
+ percpu_ref_kill(&ctx->refs);
+
+ /*
+ * Drop uring mutex before waiting for references to exit. If another
+ * thread is currently inside io_uring_enter() it might need to grab the
+ * uring_lock to make progress. If we hold it here across the drain
+ * wait, then we can deadlock. It's safe to drop the mutex here, since
+ * no new references will come in after we've killed the percpu ref.
+ */
+ mutex_unlock(&ctx->uring_lock);
+ do {
+ ret = wait_for_completion_interruptible(&ctx->ref_comp);
+ if (!ret)
+ break;
+ ret = io_run_task_work_sig();
+ } while (ret >= 0);
+ mutex_lock(&ctx->uring_lock);
+
+ if (ret)
+ io_refs_resurrect(&ctx->refs, &ctx->ref_comp);
+ return ret;
+}
+
static int __io_uring_register(struct io_ring_ctx *ctx, unsigned opcode,
void __user *arg, unsigned nr_args)
__releases(ctx->uring_lock)
}
if (io_register_op_must_quiesce(opcode)) {
- percpu_ref_kill(&ctx->refs);
-
- /*
- * Drop uring mutex before waiting for references to exit. If
- * another thread is currently inside io_uring_enter() it might
- * need to grab the uring_lock to make progress. If we hold it
- * here across the drain wait, then we can deadlock. It's safe
- * to drop the mutex here, since no new references will come in
- * after we've killed the percpu ref.
- */
- mutex_unlock(&ctx->uring_lock);
- do {
- ret = wait_for_completion_interruptible(&ctx->ref_comp);
- if (!ret)
- break;
- ret = io_run_task_work_sig();
- if (ret < 0)
- break;
- } while (1);
- mutex_lock(&ctx->uring_lock);
-
- if (ret) {
- io_refs_resurrect(&ctx->refs, &ctx->ref_comp);
+ ret = io_ctx_quiesce(ctx);
+ if (ret)
return ret;
- }
}
switch (opcode) {
ret = io_register_rsrc_update(ctx, arg, nr_args,
IORING_RSRC_BUFFER);
break;
+ case IORING_REGISTER_IOWQ_AFF:
+ ret = -EINVAL;
+ if (!arg || !nr_args)
+ break;
+ ret = io_register_iowq_aff(ctx, arg, nr_args);
+ break;
+ case IORING_UNREGISTER_IOWQ_AFF:
+ ret = -EINVAL;
+ if (arg || nr_args)
+ break;
+ ret = io_unregister_iowq_aff(ctx);
+ break;
+ case IORING_REGISTER_IOWQ_MAX_WORKERS:
+ ret = -EINVAL;
+ if (!arg || nr_args != 2)
+ break;
+ ret = io_register_iowq_max_workers(ctx, arg);
+ break;
default:
ret = -EINVAL;
break;
BUILD_BUG_SQE_ELEM(28, __u32, splice_flags);
BUILD_BUG_SQE_ELEM(32, __u64, user_data);
BUILD_BUG_SQE_ELEM(40, __u16, buf_index);
+ BUILD_BUG_SQE_ELEM(40, __u16, buf_group);
BUILD_BUG_SQE_ELEM(42, __u16, personality);
BUILD_BUG_SQE_ELEM(44, __s32, splice_fd_in);
+ BUILD_BUG_SQE_ELEM(44, __u32, file_index);
BUILD_BUG_ON(sizeof(struct io_uring_files_update) !=
sizeof(struct io_uring_rsrc_update));
BUILD_BUG_ON(sizeof(struct io_uring_rsrc_update) >
sizeof(struct io_uring_rsrc_update2));
+
+ /* ->buf_index is u16 */
+ BUILD_BUG_ON(IORING_MAX_REG_BUFFERS >= (1u << 16));
+
/* should fit into one byte */
BUILD_BUG_ON(SQE_VALID_FLAGS >= (1 << 8));
BUILD_BUG_ON(ARRAY_SIZE(io_op_defs) != IORING_OP_LAST);
- BUILD_BUG_ON(__REQ_F_LAST_BIT >= 8 * sizeof(int));
+ BUILD_BUG_ON(__REQ_F_LAST_BIT > 8 * sizeof(int));
+
req_cachep = KMEM_CACHE(io_kiocb, SLAB_HWCACHE_ALIGN | SLAB_PANIC |
SLAB_ACCOUNT);
return 0;