2022-05-24 21:45:38 +03:00
|
|
|
#ifndef IOU_CORE_H
|
|
|
|
#define IOU_CORE_H
|
|
|
|
|
|
|
|
#include <linux/errno.h>
|
2022-05-25 06:54:43 +03:00
|
|
|
#include <linux/lockdep.h>
|
2022-06-16 15:57:19 +03:00
|
|
|
#include <linux/io_uring_types.h>
|
|
|
|
#include "io-wq.h"
|
2022-06-21 12:09:01 +03:00
|
|
|
#include "slist.h"
|
2022-06-16 15:57:19 +03:00
|
|
|
#include "filetable.h"
|
2022-05-24 21:45:38 +03:00
|
|
|
|
2022-06-13 16:27:03 +03:00
|
|
|
#ifndef CREATE_TRACE_POINTS
|
|
|
|
#include <trace/events/io_uring.h>
|
|
|
|
#endif
|
|
|
|
|
2022-05-25 00:21:00 +03:00
|
|
|
enum {
|
|
|
|
IOU_OK = 0,
|
|
|
|
IOU_ISSUE_SKIP_COMPLETE = -EIOCBQUEUED,
|
2022-06-30 12:12:25 +03:00
|
|
|
|
|
|
|
/*
|
2022-11-17 21:40:16 +03:00
|
|
|
* Intended only when both IO_URING_F_MULTISHOT is passed
|
|
|
|
* to indicate to the poll runner that multishot should be
|
2022-06-30 12:12:25 +03:00
|
|
|
* removed and the result is set on req->cqe.res.
|
|
|
|
*/
|
|
|
|
IOU_STOP_MULTISHOT = -ECANCELED,
|
2022-05-25 00:21:00 +03:00
|
|
|
};
|
|
|
|
|
2022-09-23 16:53:25 +03:00
|
|
|
struct io_uring_cqe *__io_get_cqe(struct io_ring_ctx *ctx, bool overflow);
|
2022-06-17 11:48:02 +03:00
|
|
|
bool io_req_cqe_overflow(struct io_kiocb *req);
|
2022-08-30 15:50:10 +03:00
|
|
|
int io_run_task_work_sig(struct io_ring_ctx *ctx);
|
2022-10-27 17:44:29 +03:00
|
|
|
int __io_run_local_work(struct io_ring_ctx *ctx, bool *locked);
|
2022-08-30 15:50:10 +03:00
|
|
|
int io_run_local_work(struct io_ring_ctx *ctx);
|
2022-06-19 14:26:05 +03:00
|
|
|
void io_req_complete_failed(struct io_kiocb *req, s32 res);
|
|
|
|
void __io_req_complete(struct io_kiocb *req, unsigned issue_flags);
|
|
|
|
void io_req_complete_post(struct io_kiocb *req);
|
2022-06-30 12:12:26 +03:00
|
|
|
bool io_post_aux_cqe(struct io_ring_ctx *ctx, u64 user_data, s32 res, u32 cflags,
|
|
|
|
bool allow_overflow);
|
2022-07-12 23:52:38 +03:00
|
|
|
bool io_fill_cqe_aux(struct io_ring_ctx *ctx, u64 user_data, s32 res, u32 cflags,
|
|
|
|
bool allow_overflow);
|
2022-06-19 14:26:05 +03:00
|
|
|
void __io_commit_cqring_flush(struct io_ring_ctx *ctx);
|
|
|
|
|
|
|
|
struct page **io_pin_pages(unsigned long ubuf, unsigned long len, int *npages);
|
|
|
|
|
|
|
|
struct file *io_file_get_normal(struct io_kiocb *req, int fd);
|
|
|
|
struct file *io_file_get_fixed(struct io_kiocb *req, int fd,
|
|
|
|
unsigned issue_flags);
|
|
|
|
|
2022-07-21 18:06:47 +03:00
|
|
|
static inline bool io_req_ffs_set(struct io_kiocb *req)
|
|
|
|
{
|
|
|
|
return req->flags & REQ_F_FIXED_FILE;
|
|
|
|
}
|
|
|
|
|
2022-11-11 19:54:08 +03:00
|
|
|
void __io_req_task_work_add(struct io_kiocb *req, bool allow_local);
|
2022-06-19 14:26:05 +03:00
|
|
|
bool io_is_uring_fops(struct file *file);
|
|
|
|
bool io_alloc_async_data(struct io_kiocb *req);
|
|
|
|
void io_req_tw_post_queue(struct io_kiocb *req, s32 res, u32 cflags);
|
|
|
|
void io_req_task_queue(struct io_kiocb *req);
|
|
|
|
void io_queue_iowq(struct io_kiocb *req, bool *dont_use);
|
|
|
|
void io_req_task_complete(struct io_kiocb *req, bool *locked);
|
|
|
|
void io_req_task_queue_fail(struct io_kiocb *req, int ret);
|
|
|
|
void io_req_task_submit(struct io_kiocb *req, bool *locked);
|
|
|
|
void tctx_task_work(struct callback_head *cb);
|
|
|
|
__cold void io_uring_cancel_generic(bool cancel_all, struct io_sq_data *sqd);
|
|
|
|
int io_uring_alloc_task_context(struct task_struct *task,
|
|
|
|
struct io_ring_ctx *ctx);
|
|
|
|
|
|
|
|
int io_poll_issue(struct io_kiocb *req, bool *locked);
|
|
|
|
int io_submit_sqes(struct io_ring_ctx *ctx, unsigned int nr);
|
|
|
|
int io_do_iopoll(struct io_ring_ctx *ctx, bool force_nonspin);
|
|
|
|
void io_free_batch_list(struct io_ring_ctx *ctx, struct io_wq_work_node *node);
|
|
|
|
int io_req_prep_async(struct io_kiocb *req);
|
|
|
|
|
|
|
|
struct io_wq_work *io_wq_free_work(struct io_wq_work *work);
|
|
|
|
void io_wq_submit_work(struct io_wq_work *work);
|
|
|
|
|
|
|
|
void io_free_req(struct io_kiocb *req);
|
|
|
|
void io_queue_next(struct io_kiocb *req);
|
2022-07-12 23:52:37 +03:00
|
|
|
void __io_put_task(struct task_struct *task, int nr);
|
2022-07-12 23:52:47 +03:00
|
|
|
void io_task_refs_refill(struct io_uring_task *tctx);
|
2022-07-27 12:30:40 +03:00
|
|
|
bool __io_alloc_req_refill(struct io_ring_ctx *ctx);
|
2022-06-19 14:26:05 +03:00
|
|
|
|
|
|
|
bool io_match_task_safe(struct io_kiocb *head, struct task_struct *task,
|
|
|
|
bool cancel_all);
|
|
|
|
|
2022-11-11 19:54:08 +03:00
|
|
|
static inline void io_req_task_work_add(struct io_kiocb *req)
|
|
|
|
{
|
|
|
|
__io_req_task_work_add(req, true);
|
|
|
|
}
|
|
|
|
|
2022-06-19 14:26:05 +03:00
|
|
|
#define io_for_each_link(pos, head) \
|
|
|
|
for (pos = (head); pos; pos = pos->link)
|
2022-06-13 16:27:03 +03:00
|
|
|
|
2022-06-20 03:25:56 +03:00
|
|
|
static inline void io_cq_lock(struct io_ring_ctx *ctx)
|
|
|
|
__acquires(ctx->completion_lock)
|
|
|
|
{
|
|
|
|
spin_lock(&ctx->completion_lock);
|
|
|
|
}
|
|
|
|
|
|
|
|
void io_cq_unlock_post(struct io_ring_ctx *ctx);
|
|
|
|
|
2022-09-23 16:53:25 +03:00
|
|
|
static inline struct io_uring_cqe *io_get_cqe_overflow(struct io_ring_ctx *ctx,
|
|
|
|
bool overflow)
|
2022-06-13 16:27:03 +03:00
|
|
|
{
|
|
|
|
if (likely(ctx->cqe_cached < ctx->cqe_sentinel)) {
|
|
|
|
struct io_uring_cqe *cqe = ctx->cqe_cached;
|
|
|
|
|
|
|
|
ctx->cached_cq_tail++;
|
|
|
|
ctx->cqe_cached++;
|
2022-06-17 11:48:05 +03:00
|
|
|
if (ctx->flags & IORING_SETUP_CQE32)
|
|
|
|
ctx->cqe_cached++;
|
2022-06-13 16:27:03 +03:00
|
|
|
return cqe;
|
|
|
|
}
|
|
|
|
|
2022-09-23 16:53:25 +03:00
|
|
|
return __io_get_cqe(ctx, overflow);
|
|
|
|
}
|
|
|
|
|
|
|
|
static inline struct io_uring_cqe *io_get_cqe(struct io_ring_ctx *ctx)
|
|
|
|
{
|
|
|
|
return io_get_cqe_overflow(ctx, false);
|
2022-06-13 16:27:03 +03:00
|
|
|
}
|
|
|
|
|
|
|
|
static inline bool __io_fill_cqe_req(struct io_ring_ctx *ctx,
|
|
|
|
struct io_kiocb *req)
|
|
|
|
{
|
|
|
|
struct io_uring_cqe *cqe;
|
|
|
|
|
2022-06-17 11:48:04 +03:00
|
|
|
/*
|
|
|
|
* If we can't get a cq entry, userspace overflowed the
|
|
|
|
* submission (by quite a lot). Increment the overflow count in
|
|
|
|
* the ring.
|
|
|
|
*/
|
|
|
|
cqe = io_get_cqe(ctx);
|
|
|
|
if (unlikely(!cqe))
|
|
|
|
return io_req_cqe_overflow(req);
|
2022-06-30 12:12:31 +03:00
|
|
|
|
|
|
|
trace_io_uring_complete(req->ctx, req, req->cqe.user_data,
|
|
|
|
req->cqe.res, req->cqe.flags,
|
|
|
|
(req->flags & REQ_F_CQE32_INIT) ? req->extra1 : 0,
|
|
|
|
(req->flags & REQ_F_CQE32_INIT) ? req->extra2 : 0);
|
|
|
|
|
2022-06-17 11:48:04 +03:00
|
|
|
memcpy(cqe, &req->cqe, sizeof(*cqe));
|
|
|
|
|
|
|
|
if (ctx->flags & IORING_SETUP_CQE32) {
|
2022-06-13 16:27:03 +03:00
|
|
|
u64 extra1 = 0, extra2 = 0;
|
|
|
|
|
|
|
|
if (req->flags & REQ_F_CQE32_INIT) {
|
|
|
|
extra1 = req->extra1;
|
|
|
|
extra2 = req->extra2;
|
|
|
|
}
|
|
|
|
|
2022-06-17 11:48:04 +03:00
|
|
|
WRITE_ONCE(cqe->big_cqe[0], extra1);
|
|
|
|
WRITE_ONCE(cqe->big_cqe[1], extra2);
|
2022-06-13 16:27:03 +03:00
|
|
|
}
|
2022-06-17 11:48:04 +03:00
|
|
|
return true;
|
2022-06-13 16:27:03 +03:00
|
|
|
}
|
|
|
|
|
2022-05-25 06:19:47 +03:00
|
|
|
static inline void req_set_fail(struct io_kiocb *req)
|
|
|
|
{
|
|
|
|
req->flags |= REQ_F_FAIL;
|
|
|
|
if (req->flags & REQ_F_CQE_SKIP) {
|
|
|
|
req->flags &= ~REQ_F_CQE_SKIP;
|
|
|
|
req->flags |= REQ_F_SKIP_LINK_CQES;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2022-05-24 21:45:38 +03:00
|
|
|
static inline void io_req_set_res(struct io_kiocb *req, s32 res, u32 cflags)
|
|
|
|
{
|
|
|
|
req->cqe.res = res;
|
|
|
|
req->cqe.flags = cflags;
|
|
|
|
}
|
|
|
|
|
2022-05-25 14:59:19 +03:00
|
|
|
static inline bool req_has_async_data(struct io_kiocb *req)
|
|
|
|
{
|
|
|
|
return req->flags & REQ_F_ASYNC_DATA;
|
|
|
|
}
|
|
|
|
|
2022-05-25 06:19:47 +03:00
|
|
|
static inline void io_put_file(struct file *file)
|
|
|
|
{
|
|
|
|
if (file)
|
|
|
|
fput(file);
|
|
|
|
}
|
|
|
|
|
2022-05-25 06:54:43 +03:00
|
|
|
static inline void io_ring_submit_unlock(struct io_ring_ctx *ctx,
|
|
|
|
unsigned issue_flags)
|
|
|
|
{
|
|
|
|
lockdep_assert_held(&ctx->uring_lock);
|
|
|
|
if (issue_flags & IO_URING_F_UNLOCKED)
|
|
|
|
mutex_unlock(&ctx->uring_lock);
|
|
|
|
}
|
|
|
|
|
|
|
|
static inline void io_ring_submit_lock(struct io_ring_ctx *ctx,
|
|
|
|
unsigned issue_flags)
|
|
|
|
{
|
|
|
|
/*
|
|
|
|
* "Normal" inline submissions always hold the uring_lock, since we
|
|
|
|
* grab it from the system call. Same is true for the SQPOLL offload.
|
|
|
|
* The only exception is when we've detached the request and issue it
|
|
|
|
* from an async worker thread, grab the lock for that case.
|
|
|
|
*/
|
|
|
|
if (issue_flags & IO_URING_F_UNLOCKED)
|
|
|
|
mutex_lock(&ctx->uring_lock);
|
|
|
|
lockdep_assert_held(&ctx->uring_lock);
|
|
|
|
}
|
|
|
|
|
2022-05-25 15:25:13 +03:00
|
|
|
static inline void io_commit_cqring(struct io_ring_ctx *ctx)
|
|
|
|
{
|
|
|
|
/* order cqe stores with ring update */
|
|
|
|
smp_store_release(&ctx->rings->cq.tail, ctx->cached_cq_tail);
|
|
|
|
}
|
|
|
|
|
2022-10-06 04:06:10 +03:00
|
|
|
/* requires smb_mb() prior, see wq_has_sleeper() */
|
|
|
|
static inline void __io_cqring_wake(struct io_ring_ctx *ctx)
|
2022-06-13 16:27:03 +03:00
|
|
|
{
|
|
|
|
/*
|
|
|
|
* 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.
|
|
|
|
*/
|
2022-10-06 04:06:10 +03:00
|
|
|
if (waitqueue_active(&ctx->cq_wait))
|
2022-06-13 16:27:03 +03:00
|
|
|
wake_up_all(&ctx->cq_wait);
|
|
|
|
}
|
|
|
|
|
2022-10-06 04:06:10 +03:00
|
|
|
static inline void io_cqring_wake(struct io_ring_ctx *ctx)
|
|
|
|
{
|
|
|
|
smp_mb();
|
|
|
|
__io_cqring_wake(ctx);
|
|
|
|
}
|
|
|
|
|
2022-05-25 18:13:39 +03:00
|
|
|
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 == ctx->sq_entries;
|
|
|
|
}
|
|
|
|
|
|
|
|
static inline unsigned int io_sqring_entries(struct io_ring_ctx *ctx)
|
|
|
|
{
|
|
|
|
struct io_rings *rings = ctx->rings;
|
|
|
|
|
|
|
|
/* make sure SQ entry isn't read before tail */
|
|
|
|
return smp_load_acquire(&rings->sq.tail) - ctx->cached_sq_head;
|
|
|
|
}
|
|
|
|
|
2022-08-30 15:50:10 +03:00
|
|
|
static inline int io_run_task_work(void)
|
2022-05-25 18:13:39 +03:00
|
|
|
{
|
2022-09-30 00:29:13 +03:00
|
|
|
if (task_work_pending(current)) {
|
|
|
|
if (test_thread_flag(TIF_NOTIFY_SIGNAL))
|
|
|
|
clear_notify_signal();
|
2022-05-25 18:13:39 +03:00
|
|
|
__set_current_state(TASK_RUNNING);
|
2022-09-30 00:29:13 +03:00
|
|
|
task_work_run();
|
2022-08-30 15:50:10 +03:00
|
|
|
return 1;
|
2022-05-25 18:13:39 +03:00
|
|
|
}
|
|
|
|
|
2022-08-30 15:50:10 +03:00
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
2022-09-03 18:52:01 +03:00
|
|
|
static inline bool io_task_work_pending(struct io_ring_ctx *ctx)
|
|
|
|
{
|
|
|
|
return test_thread_flag(TIF_NOTIFY_SIGNAL) ||
|
|
|
|
!wq_list_empty(&ctx->work_llist);
|
|
|
|
}
|
|
|
|
|
2022-08-30 15:50:10 +03:00
|
|
|
static inline int io_run_task_work_ctx(struct io_ring_ctx *ctx)
|
|
|
|
{
|
|
|
|
int ret = 0;
|
|
|
|
int ret2;
|
|
|
|
|
|
|
|
if (ctx->flags & IORING_SETUP_DEFER_TASKRUN)
|
|
|
|
ret = io_run_local_work(ctx);
|
|
|
|
|
|
|
|
/* want to run this after in case more is added */
|
|
|
|
ret2 = io_run_task_work();
|
|
|
|
|
|
|
|
/* Try propagate error in favour of if tasks were run,
|
|
|
|
* but still make sure to run them if requested
|
|
|
|
*/
|
|
|
|
if (ret >= 0)
|
|
|
|
ret += ret2;
|
|
|
|
|
|
|
|
return ret;
|
2022-05-25 18:13:39 +03:00
|
|
|
}
|
|
|
|
|
2022-10-06 23:42:33 +03:00
|
|
|
static inline int io_run_local_work_locked(struct io_ring_ctx *ctx)
|
|
|
|
{
|
2022-10-27 17:44:29 +03:00
|
|
|
bool locked;
|
|
|
|
int ret;
|
|
|
|
|
2022-10-06 23:42:33 +03:00
|
|
|
if (llist_empty(&ctx->work_llist))
|
|
|
|
return 0;
|
2022-10-27 17:44:29 +03:00
|
|
|
|
|
|
|
locked = true;
|
|
|
|
ret = __io_run_local_work(ctx, &locked);
|
|
|
|
/* shouldn't happen! */
|
|
|
|
if (WARN_ON_ONCE(!locked))
|
|
|
|
mutex_lock(&ctx->uring_lock);
|
|
|
|
return ret;
|
2022-10-06 23:42:33 +03:00
|
|
|
}
|
|
|
|
|
2022-06-15 19:33:51 +03:00
|
|
|
static inline void io_tw_lock(struct io_ring_ctx *ctx, bool *locked)
|
|
|
|
{
|
|
|
|
if (!*locked) {
|
|
|
|
mutex_lock(&ctx->uring_lock);
|
|
|
|
*locked = true;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2022-06-20 03:26:00 +03:00
|
|
|
/*
|
|
|
|
* Don't complete immediately but use deferred completion infrastructure.
|
|
|
|
* Protected by ->uring_lock and can only be used either with
|
|
|
|
* IO_URING_F_COMPLETE_DEFER or inside a tw handler holding the mutex.
|
|
|
|
*/
|
|
|
|
static inline void io_req_complete_defer(struct io_kiocb *req)
|
|
|
|
__must_hold(&req->ctx->uring_lock)
|
2022-06-15 19:33:51 +03:00
|
|
|
{
|
|
|
|
struct io_submit_state *state = &req->ctx->submit_state;
|
|
|
|
|
2022-06-20 03:26:00 +03:00
|
|
|
lockdep_assert_held(&req->ctx->uring_lock);
|
|
|
|
|
2022-06-15 19:33:51 +03:00
|
|
|
wq_list_add_tail(&req->comp_list, &state->compl_reqs);
|
|
|
|
}
|
|
|
|
|
2022-06-20 03:25:57 +03:00
|
|
|
static inline void io_commit_cqring_flush(struct io_ring_ctx *ctx)
|
|
|
|
{
|
|
|
|
if (unlikely(ctx->off_timeout_used || ctx->drain_active || ctx->has_evfd))
|
|
|
|
__io_commit_cqring_flush(ctx);
|
|
|
|
}
|
|
|
|
|
2022-07-12 23:52:37 +03:00
|
|
|
/* must to be called somewhat shortly after putting a request */
|
|
|
|
static inline void io_put_task(struct task_struct *task, int nr)
|
|
|
|
{
|
|
|
|
if (likely(task == current))
|
|
|
|
task->io_uring->cached_refs += nr;
|
|
|
|
else
|
|
|
|
__io_put_task(task, nr);
|
|
|
|
}
|
|
|
|
|
2022-07-12 23:52:47 +03:00
|
|
|
static inline void io_get_task_refs(int nr)
|
|
|
|
{
|
|
|
|
struct io_uring_task *tctx = current->io_uring;
|
|
|
|
|
|
|
|
tctx->cached_refs -= nr;
|
|
|
|
if (unlikely(tctx->cached_refs < 0))
|
|
|
|
io_task_refs_refill(tctx);
|
|
|
|
}
|
|
|
|
|
2022-07-27 12:30:40 +03:00
|
|
|
static inline bool io_req_cache_empty(struct io_ring_ctx *ctx)
|
|
|
|
{
|
|
|
|
return !ctx->submit_state.free_list.next;
|
|
|
|
}
|
|
|
|
|
|
|
|
static inline bool io_alloc_req_refill(struct io_ring_ctx *ctx)
|
|
|
|
{
|
|
|
|
if (unlikely(io_req_cache_empty(ctx)))
|
|
|
|
return __io_alloc_req_refill(ctx);
|
|
|
|
return true;
|
|
|
|
}
|
|
|
|
|
|
|
|
static inline struct io_kiocb *io_alloc_req(struct io_ring_ctx *ctx)
|
|
|
|
{
|
|
|
|
struct io_wq_work_node *node;
|
|
|
|
|
|
|
|
node = wq_stack_extract(&ctx->submit_state.free_list);
|
|
|
|
return container_of(node, struct io_kiocb, comp_list);
|
|
|
|
}
|
|
|
|
|
2022-09-08 18:56:52 +03:00
|
|
|
static inline bool io_allowed_run_tw(struct io_ring_ctx *ctx)
|
|
|
|
{
|
2022-09-08 18:56:53 +03:00
|
|
|
return likely(!(ctx->flags & IORING_SETUP_DEFER_TASKRUN) ||
|
|
|
|
ctx->submitter_task == current);
|
2022-09-08 18:56:52 +03:00
|
|
|
}
|
|
|
|
|
2022-05-24 21:45:38 +03:00
|
|
|
#endif
|