2022-06-13 16:27:03 +03:00
// SPDX-License-Identifier: GPL-2.0
# include <linux/kernel.h>
# include <linux/errno.h>
# include <linux/fs.h>
# include <linux/file.h>
# include <linux/blk-mq.h>
# include <linux/mm.h>
# include <linux/slab.h>
# include <linux/fsnotify.h>
# include <linux/poll.h>
# include <linux/nospec.h>
# include <linux/compat.h>
2023-12-01 03:57:35 +03:00
# include <linux/io_uring/cmd.h>
2024-02-13 02:42:36 +03:00
# include <linux/indirect_call_wrapper.h>
2022-06-13 16:27:03 +03:00
# include <uapi/linux/io_uring.h>
# include "io_uring.h"
# include "opdef.h"
# include "kbuf.h"
2024-03-21 00:19:44 +03:00
# include "alloc_cache.h"
2022-06-13 16:27:03 +03:00
# include "rsrc.h"
io_uring/rw: ensure poll based multishot read retries appropriately
io_read_mshot() always relies on poll triggering retries, and this works
fine as long as we do a retry per size of the buffer being read. The
buffer size is given by the size of the buffer(s) in the given buffer
group ID.
But if we're reading less than what is available, then we don't always
get to read everything that is available. For example, if the buffers
available are 32 bytes and we have 64 bytes to read, then we'll
correctly read the first 32 bytes and then wait for another poll trigger
before we attempt the next read. This next poll trigger may never
happen, in which case we just sit forever and never make progress, or it
may trigger at some point in the future, and now we're just delivering
the available data much later than we should have.
io_read_mshot() could do retries itself, but that is wasteful as we'll
be going through all of __io_read() again, and most likely in vain.
Rather than do that, bump our poll reference count and have
io_poll_check_events() do one more loop and check with vfs_poll() if we
have more data to read. If we do, io_read_mshot() will get invoked again
directly and we'll read the next chunk.
io_poll_multishot_retry() must only get called from inside
io_poll_issue(), which is our multishot retry handler, as we know we
already "own" the request at this point.
Cc: stable@vger.kernel.org
Link: https://github.com/axboe/liburing/issues/1041
Fixes: fc68fcda0491 ("io_uring/rw: add support for IORING_OP_READ_MULTISHOT")
Signed-off-by: Jens Axboe <axboe@kernel.dk>
2024-01-27 23:44:58 +03:00
# include "poll.h"
2022-06-13 16:27:03 +03:00
# include "rw.h"
struct io_rw {
/* NOTE: kiocb has the file as the first member, so don't do it here */
struct kiocb kiocb ;
u64 addr ;
u32 len ;
rwf_t flags ;
} ;
static inline bool io_file_supports_nowait ( struct io_kiocb * req )
{
return req - > flags & REQ_F_SUPPORT_NOWAIT ;
}
2022-09-07 19:51:52 +03:00
# ifdef CONFIG_COMPAT
static int io_iov_compat_buffer_select_prep ( struct io_rw * rw )
{
struct compat_iovec __user * uiov ;
compat_ssize_t clen ;
uiov = u64_to_user_ptr ( rw - > addr ) ;
if ( ! access_ok ( uiov , sizeof ( * uiov ) ) )
return - EFAULT ;
if ( __get_user ( clen , & uiov - > iov_len ) )
return - EFAULT ;
if ( clen < 0 )
return - EINVAL ;
rw - > len = clen ;
return 0 ;
}
# endif
static int io_iov_buffer_select_prep ( struct io_kiocb * req )
{
struct iovec __user * uiov ;
struct iovec iov ;
struct io_rw * rw = io_kiocb_to_cmd ( req , struct io_rw ) ;
if ( rw - > len ! = 1 )
return - EINVAL ;
# ifdef CONFIG_COMPAT
if ( req - > ctx - > compat )
return io_iov_compat_buffer_select_prep ( rw ) ;
# endif
uiov = u64_to_user_ptr ( rw - > addr ) ;
if ( copy_from_user ( & iov , uiov , sizeof ( * uiov ) ) )
return - EFAULT ;
rw - > len = iov . iov_len ;
return 0 ;
}
2024-03-19 01:13:01 +03:00
static int __io_import_iovec ( int ddir , struct io_kiocb * req ,
struct io_async_rw * io ,
unsigned int issue_flags )
{
const struct io_issue_def * def = & io_issue_defs [ req - > opcode ] ;
struct io_rw * rw = io_kiocb_to_cmd ( req , struct io_rw ) ;
2024-03-19 01:31:44 +03:00
struct iovec * iov ;
2024-03-19 01:13:01 +03:00
void __user * buf ;
2024-03-19 01:31:44 +03:00
int nr_segs , ret ;
2024-03-19 01:13:01 +03:00
size_t sqe_len ;
buf = u64_to_user_ptr ( rw - > addr ) ;
sqe_len = rw - > len ;
if ( ! def - > vectored | | req - > flags & REQ_F_BUFFER_SELECT ) {
if ( io_do_buffer_select ( req ) ) {
buf = io_buffer_select ( req , & sqe_len , issue_flags ) ;
if ( ! buf )
return - ENOBUFS ;
rw - > addr = ( unsigned long ) buf ;
rw - > len = sqe_len ;
}
2024-03-19 01:25:58 +03:00
return import_ubuf ( ddir , buf , sqe_len , & io - > iter ) ;
2024-03-19 01:13:01 +03:00
}
2024-03-19 01:31:44 +03:00
if ( io - > free_iovec ) {
nr_segs = io - > free_iov_nr ;
iov = io - > free_iovec ;
} else {
iov = & io - > fast_iov ;
nr_segs = 1 ;
}
ret = __import_iovec ( ddir , buf , sqe_len , nr_segs , & iov , & io - > iter ,
req - > ctx - > compat ) ;
if ( unlikely ( ret < 0 ) )
return ret ;
if ( iov ) {
req - > flags | = REQ_F_NEED_CLEANUP ;
io - > free_iov_nr = io - > iter . nr_segs ;
kfree ( io - > free_iovec ) ;
io - > free_iovec = iov ;
}
return 0 ;
2024-03-19 01:13:01 +03:00
}
static inline int io_import_iovec ( int rw , struct io_kiocb * req ,
struct io_async_rw * io ,
unsigned int issue_flags )
{
int ret ;
ret = __io_import_iovec ( rw , req , io , issue_flags ) ;
if ( unlikely ( ret < 0 ) )
return ret ;
2024-03-19 01:25:58 +03:00
iov_iter_save_state ( & io - > iter , & io - > iter_state ) ;
2024-03-19 01:13:01 +03:00
return 0 ;
}
static void io_rw_iovec_free ( struct io_async_rw * rw )
{
if ( rw - > free_iovec ) {
kfree ( rw - > free_iovec ) ;
2024-03-19 01:31:44 +03:00
rw - > free_iov_nr = 0 ;
2024-03-19 01:13:01 +03:00
rw - > free_iovec = NULL ;
}
}
static void io_rw_recycle ( struct io_kiocb * req , unsigned int issue_flags )
{
struct io_async_rw * rw = req - > async_data ;
2024-03-19 01:31:44 +03:00
struct iovec * iov ;
2024-03-19 01:13:01 +03:00
if ( unlikely ( issue_flags & IO_URING_F_UNLOCKED ) ) {
io_rw_iovec_free ( rw ) ;
return ;
}
2024-03-19 01:31:44 +03:00
iov = rw - > free_iovec ;
2024-03-21 00:19:44 +03:00
if ( io_alloc_cache_put ( & req - > ctx - > rw_cache , rw ) ) {
2024-03-19 01:31:44 +03:00
if ( iov )
kasan_mempool_poison_object ( iov ) ;
2024-03-19 01:13:01 +03:00
req - > async_data = NULL ;
req - > flags & = ~ REQ_F_ASYNC_DATA ;
}
}
static void io_req_rw_cleanup ( struct io_kiocb * req , unsigned int issue_flags )
{
/*
* Disable quick recycling for anything that ' s gone through io - wq .
* In theory , this should be fine to cleanup . However , some read or
* write iter handling touches the iovec AFTER having called into the
* handler , eg to reexpand or revert . This means we can have :
*
* task io - wq
* issue
* punt to io - wq
* issue
* blkdev_write_iter ( )
* - > ki_complete ( )
* io_complete_rw ( )
* queue tw complete
* run tw
* req_rw_cleanup
* iov_iter_count ( ) < - look at iov_iter again
*
* which can lead to a UAF . This is only possible for io - wq offload
* as the cleanup can run in parallel . As io - wq is not the fast path ,
* just leave cleanup to the end .
*
* This is really a bug in the core code that does this , any issue
* path should assume that a successful ( or - EIOCBQUEUED ) return can
* mean that the underlying data can be gone at any time . But that
* should be fixed seperately , and then this check could be killed .
*/
if ( ! ( req - > flags & REQ_F_REFCOUNT ) ) {
req - > flags & = ~ REQ_F_NEED_CLEANUP ;
io_rw_recycle ( req , issue_flags ) ;
}
}
static int io_rw_alloc_async ( struct io_kiocb * req )
{
struct io_ring_ctx * ctx = req - > ctx ;
struct io_async_rw * rw ;
2024-03-21 00:19:44 +03:00
rw = io_alloc_cache_get ( & ctx - > rw_cache ) ;
if ( rw ) {
2024-03-19 01:31:44 +03:00
if ( rw - > free_iovec ) {
kasan_mempool_unpoison_object ( rw - > free_iovec ,
rw - > free_iov_nr * sizeof ( struct iovec ) ) ;
req - > flags | = REQ_F_NEED_CLEANUP ;
}
2024-03-19 01:13:01 +03:00
req - > flags | = REQ_F_ASYNC_DATA ;
req - > async_data = rw ;
goto done ;
}
if ( ! io_alloc_async_data ( req ) ) {
rw = req - > async_data ;
rw - > free_iovec = NULL ;
2024-03-19 01:31:44 +03:00
rw - > free_iov_nr = 0 ;
done :
2024-03-19 01:13:01 +03:00
rw - > bytes_done = 0 ;
return 0 ;
}
return - ENOMEM ;
}
static int io_prep_rw_setup ( struct io_kiocb * req , int ddir , bool do_import )
{
struct io_async_rw * rw ;
int ret ;
if ( io_rw_alloc_async ( req ) )
return - ENOMEM ;
if ( ! do_import | | io_do_buffer_select ( req ) )
return 0 ;
rw = req - > async_data ;
ret = io_import_iovec ( ddir , req , rw , 0 ) ;
if ( unlikely ( ret < 0 ) )
return ret ;
2024-03-19 01:25:58 +03:00
iov_iter_save_state ( & rw - > iter , & rw - > iter_state ) ;
2024-03-19 01:13:01 +03:00
return 0 ;
}
static int io_prep_rw ( struct io_kiocb * req , const struct io_uring_sqe * sqe ,
int ddir , bool do_import )
2022-06-13 16:27:03 +03:00
{
2022-08-11 10:11:15 +03:00
struct io_rw * rw = io_kiocb_to_cmd ( req , struct io_rw ) ;
2022-06-13 16:27:03 +03:00
unsigned ioprio ;
int ret ;
rw - > kiocb . ki_pos = READ_ONCE ( sqe - > off ) ;
/* used for fixed read/write too - just read unconditionally */
req - > buf_index = READ_ONCE ( sqe - > buf_index ) ;
ioprio = READ_ONCE ( sqe - > ioprio ) ;
if ( ioprio ) {
ret = ioprio_check_cap ( ioprio ) ;
if ( ret )
return ret ;
rw - > kiocb . ki_ioprio = ioprio ;
} else {
rw - > kiocb . ki_ioprio = get_current_ioprio ( ) ;
}
2023-07-08 19:03:52 +03:00
rw - > kiocb . dio_complete = NULL ;
2022-06-13 16:27:03 +03:00
rw - > addr = READ_ONCE ( sqe - > addr ) ;
rw - > len = READ_ONCE ( sqe - > len ) ;
rw - > flags = READ_ONCE ( sqe - > rw_flags ) ;
2024-03-19 01:13:01 +03:00
return io_prep_rw_setup ( req , ddir , do_import ) ;
2023-11-06 17:41:17 +03:00
}
2024-03-19 01:13:01 +03:00
int io_prep_read ( struct io_kiocb * req , const struct io_uring_sqe * sqe )
2023-11-06 17:41:17 +03:00
{
2024-03-19 01:13:01 +03:00
return io_prep_rw ( req , sqe , ITER_DEST , true ) ;
2023-11-06 17:41:17 +03:00
}
2024-03-19 01:13:01 +03:00
int io_prep_write ( struct io_kiocb * req , const struct io_uring_sqe * sqe )
2023-11-06 17:41:17 +03:00
{
2024-03-19 01:13:01 +03:00
return io_prep_rw ( req , sqe , ITER_SOURCE , true ) ;
}
static int io_prep_rwv ( struct io_kiocb * req , const struct io_uring_sqe * sqe ,
int ddir )
{
const bool do_import = ! ( req - > flags & REQ_F_BUFFER_SELECT ) ;
2023-11-06 17:41:17 +03:00
int ret ;
2022-09-07 19:51:52 +03:00
2024-03-19 01:13:01 +03:00
ret = io_prep_rw ( req , sqe , ddir , do_import ) ;
2023-11-06 17:41:17 +03:00
if ( unlikely ( ret ) )
return ret ;
2024-03-19 01:13:01 +03:00
if ( do_import )
return 0 ;
2023-11-06 17:41:17 +03:00
/*
* Have to do this validation here , as this is in io_read ( ) rw - > len
* might have chanaged due to buffer selection
2022-09-07 19:51:52 +03:00
*/
2024-03-19 01:13:01 +03:00
return io_iov_buffer_select_prep ( req ) ;
}
2022-09-07 19:51:52 +03:00
2024-03-19 01:13:01 +03:00
int io_prep_readv ( struct io_kiocb * req , const struct io_uring_sqe * sqe )
{
return io_prep_rwv ( req , sqe , ITER_DEST ) ;
}
int io_prep_writev ( struct io_kiocb * req , const struct io_uring_sqe * sqe )
{
return io_prep_rwv ( req , sqe , ITER_SOURCE ) ;
2022-06-13 16:27:03 +03:00
}
2024-03-19 01:13:01 +03:00
static int io_prep_rw_fixed ( struct io_kiocb * req , const struct io_uring_sqe * sqe ,
int ddir )
2023-11-06 17:43:16 +03:00
{
2024-03-19 01:13:01 +03:00
struct io_rw * rw = io_kiocb_to_cmd ( req , struct io_rw ) ;
2023-11-06 17:43:16 +03:00
struct io_ring_ctx * ctx = req - > ctx ;
2024-03-19 01:13:01 +03:00
struct io_async_rw * io ;
2023-11-06 17:43:16 +03:00
u16 index ;
int ret ;
2024-03-19 01:13:01 +03:00
ret = io_prep_rw ( req , sqe , ddir , false ) ;
2023-11-06 17:43:16 +03:00
if ( unlikely ( ret ) )
return ret ;
if ( unlikely ( req - > buf_index > = ctx - > nr_user_bufs ) )
return - EFAULT ;
index = array_index_nospec ( req - > buf_index , ctx - > nr_user_bufs ) ;
req - > imu = ctx - > user_bufs [ index ] ;
io_req_set_rsrc_node ( req , ctx , 0 ) ;
2024-03-19 01:13:01 +03:00
io = req - > async_data ;
2024-03-19 01:25:58 +03:00
ret = io_import_fixed ( ddir , & io - > iter , req - > imu , rw - > addr , rw - > len ) ;
iov_iter_save_state ( & io - > iter , & io - > iter_state ) ;
2024-03-19 01:13:01 +03:00
return ret ;
}
int io_prep_read_fixed ( struct io_kiocb * req , const struct io_uring_sqe * sqe )
{
return io_prep_rw_fixed ( req , sqe , ITER_DEST ) ;
}
int io_prep_write_fixed ( struct io_kiocb * req , const struct io_uring_sqe * sqe )
{
return io_prep_rw_fixed ( req , sqe , ITER_SOURCE ) ;
2023-11-06 17:43:16 +03:00
}
2023-09-11 22:35:42 +03:00
/*
* Multishot read is prepared just like a normal read / write request , only
* difference is that we set the MULTISHOT flag .
*/
int io_read_mshot_prep ( struct io_kiocb * req , const struct io_uring_sqe * sqe )
{
2023-11-06 23:39:08 +03:00
struct io_rw * rw = io_kiocb_to_cmd ( req , struct io_rw ) ;
2023-09-11 22:35:42 +03:00
int ret ;
2023-11-03 18:26:13 +03:00
/* must be used with provided buffers */
if ( ! ( req - > flags & REQ_F_BUFFER_SELECT ) )
return - EINVAL ;
2024-03-19 01:13:01 +03:00
ret = io_prep_rw ( req , sqe , ITER_DEST , false ) ;
2023-09-11 22:35:42 +03:00
if ( unlikely ( ret ) )
return ret ;
2023-11-06 23:39:08 +03:00
if ( rw - > addr | | rw - > len )
return - EINVAL ;
2023-09-11 22:35:42 +03:00
req - > flags | = REQ_F_APOLL_MULTISHOT ;
return 0 ;
}
2022-06-13 16:27:03 +03:00
void io_readv_writev_cleanup ( struct io_kiocb * req )
{
2024-03-19 01:13:01 +03:00
io_rw_iovec_free ( req - > async_data ) ;
2022-06-13 16:27:03 +03:00
}
static inline loff_t * io_kiocb_update_pos ( struct io_kiocb * req )
{
2022-08-11 10:11:15 +03:00
struct io_rw * rw = io_kiocb_to_cmd ( req , struct io_rw ) ;
2022-06-13 16:27:03 +03:00
if ( rw - > kiocb . ki_pos ! = - 1 )
return & rw - > kiocb . ki_pos ;
if ( ! ( req - > file - > f_mode & FMODE_STREAM ) ) {
req - > flags | = REQ_F_CUR_POS ;
rw - > kiocb . ki_pos = req - > file - > f_pos ;
return & rw - > kiocb . ki_pos ;
}
rw - > kiocb . ki_pos = 0 ;
return NULL ;
}
# ifdef CONFIG_BLOCK
2024-04-25 18:04:32 +03:00
static void io_resubmit_prep ( struct io_kiocb * req )
2022-06-13 16:27:03 +03:00
{
struct io_async_rw * io = req - > async_data ;
2024-04-25 18:04:32 +03:00
iov_iter_restore ( & io - > iter , & io - > iter_state ) ;
2022-06-13 16:27:03 +03:00
}
static bool io_rw_should_reissue ( struct io_kiocb * req )
{
umode_t mode = file_inode ( req - > file ) - > i_mode ;
struct io_ring_ctx * ctx = req - > ctx ;
if ( ! S_ISBLK ( mode ) & & ! S_ISREG ( mode ) )
return false ;
if ( ( req - > flags & REQ_F_NOWAIT ) | | ( io_wq_current_is_worker ( ) & &
! ( ctx - > flags & IORING_SETUP_IOPOLL ) ) )
return false ;
/*
* If ref is dying , we might be running poll reap from the exit work .
* Don ' t attempt to reissue from that path , just let it fail with
* - EAGAIN .
*/
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
2024-04-25 18:04:32 +03:00
static void io_resubmit_prep ( struct io_kiocb * req )
2022-06-13 16:27:03 +03:00
{
}
static bool io_rw_should_reissue ( struct io_kiocb * req )
{
return false ;
}
# endif
2023-08-17 17:13:31 +03:00
static void io_req_end_write ( struct io_kiocb * req )
2022-06-13 16:27:03 +03:00
{
if ( req - > flags & REQ_F_ISREG ) {
2023-08-17 17:13:34 +03:00
struct io_rw * rw = io_kiocb_to_cmd ( req , struct io_rw ) ;
2022-06-13 16:27:03 +03:00
2023-08-17 17:13:34 +03:00
kiocb_end_write ( & rw - > kiocb ) ;
2022-06-13 16:27:03 +03:00
}
}
2022-10-11 18:06:23 +03:00
/*
* Trigger the notifications after having done some IO , and finish the write
* accounting , if any .
*/
static void io_req_io_end ( struct io_kiocb * req )
{
struct io_rw * rw = io_kiocb_to_cmd ( req , struct io_rw ) ;
if ( rw - > kiocb . ki_flags & IOCB_WRITE ) {
2023-08-17 17:13:31 +03:00
io_req_end_write ( req ) ;
2022-10-11 18:06:23 +03:00
fsnotify_modify ( req - > file ) ;
} else {
fsnotify_access ( req - > file ) ;
}
}
2022-06-13 16:27:03 +03:00
static bool __io_complete_rw_common ( struct io_kiocb * req , long res )
{
if ( unlikely ( res ! = req - > cqe . res ) ) {
if ( ( res = = - EAGAIN | | res = = - EOPNOTSUPP ) & &
io_rw_should_reissue ( req ) ) {
2022-10-11 18:06:23 +03:00
/*
* Reissue will start accounting again , finish the
* current cycle .
*/
io_req_io_end ( req ) ;
2024-03-07 22:53:24 +03:00
req - > flags | = REQ_F_REISSUE | REQ_F_BL_NO_RECYCLE ;
2022-06-13 16:27:03 +03:00
return true ;
}
req_set_fail ( req ) ;
req - > cqe . res = res ;
}
return false ;
}
2022-09-13 15:21:23 +03:00
static inline int io_fixup_rw_res ( struct io_kiocb * req , long res )
2022-09-09 14:11:49 +03:00
{
struct io_async_rw * io = req - > async_data ;
/* add previously done IO, if any */
if ( req_has_async_data ( req ) & & io - > bytes_done > 0 ) {
if ( res < 0 )
res = io - > bytes_done ;
else
res + = io - > bytes_done ;
}
return res ;
}
2023-06-02 17:41:46 +03:00
void io_req_rw_complete ( struct io_kiocb * req , struct io_tw_state * ts )
2022-09-29 19:57:05 +03:00
{
2023-07-08 19:03:52 +03:00
struct io_rw * rw = io_kiocb_to_cmd ( req , struct io_rw ) ;
struct kiocb * kiocb = & rw - > kiocb ;
if ( ( kiocb - > ki_flags & IOCB_DIO_CALLER_COMP ) & & kiocb - > dio_complete ) {
long res = kiocb - > dio_complete ( rw - > kiocb . private ) ;
io_req_set_res ( req , io_fixup_rw_res ( req , res ) , 0 ) ;
}
2022-10-11 18:06:23 +03:00
io_req_io_end ( req ) ;
2022-11-04 13:59:40 +03:00
2024-03-19 01:00:30 +03:00
if ( req - > flags & ( REQ_F_BUFFER_SELECTED | REQ_F_BUFFER_RING ) )
req - > cqe . flags | = io_put_kbuf ( req , 0 ) ;
2022-11-04 13:59:40 +03:00
2024-03-19 01:13:01 +03:00
io_req_rw_cleanup ( req , 0 ) ;
2023-03-27 18:38:15 +03:00
io_req_task_complete ( req , ts ) ;
2022-09-29 19:57:05 +03:00
}
2022-06-13 16:27:03 +03:00
static void io_complete_rw ( struct kiocb * kiocb , long res )
{
struct io_rw * rw = container_of ( kiocb , struct io_rw , kiocb ) ;
struct io_kiocb * req = cmd_to_io_kiocb ( rw ) ;
2023-07-08 19:03:52 +03:00
if ( ! kiocb - > dio_complete | | ! ( kiocb - > ki_flags & IOCB_DIO_CALLER_COMP ) ) {
if ( __io_complete_rw_common ( req , res ) )
return ;
io_req_set_res ( req , io_fixup_rw_res ( req , res ) , 0 ) ;
}
2022-09-29 19:57:05 +03:00
req - > io_task_work . func = io_req_rw_complete ;
2023-04-06 16:20:12 +03:00
__io_req_task_work_add ( req , IOU_F_TWQ_LAZY_WAKE ) ;
2022-06-13 16:27:03 +03:00
}
static void io_complete_rw_iopoll ( struct kiocb * kiocb , long res )
{
struct io_rw * rw = container_of ( kiocb , struct io_rw , kiocb ) ;
struct io_kiocb * req = cmd_to_io_kiocb ( rw ) ;
if ( kiocb - > ki_flags & IOCB_WRITE )
2023-08-17 17:13:31 +03:00
io_req_end_write ( req ) ;
2022-06-13 16:27:03 +03:00
if ( unlikely ( res ! = req - > cqe . res ) ) {
if ( res = = - EAGAIN & & io_rw_should_reissue ( req ) ) {
2024-03-07 22:53:24 +03:00
req - > flags | = REQ_F_REISSUE | REQ_F_BL_NO_RECYCLE ;
2022-06-13 16:27:03 +03:00
return ;
}
req - > cqe . res = res ;
}
/* order with io_iopoll_complete() checking ->iopoll_completed */
smp_store_release ( & req - > iopoll_completed , 1 ) ;
}
2024-01-10 20:05:32 +03:00
static inline void io_rw_done ( struct kiocb * kiocb , ssize_t ret )
{
/* IO was queued async, completion will happen later */
if ( ret = = - EIOCBQUEUED )
return ;
/* transform internal restart error codes */
if ( unlikely ( ret < 0 ) ) {
switch ( ret ) {
case - ERESTARTSYS :
case - ERESTARTNOINTR :
case - ERESTARTNOHAND :
case - ERESTART_RESTARTBLOCK :
/*
* We can ' t just restart the syscall , since previously
* submitted sqes may already be in progress . Just fail
* this IO with EINTR .
*/
ret = - EINTR ;
break ;
}
}
INDIRECT_CALL_2 ( kiocb - > ki_complete , io_complete_rw_iopoll ,
io_complete_rw , kiocb , ret ) ;
}
2022-06-16 12:21:57 +03:00
static int kiocb_done ( struct io_kiocb * req , ssize_t ret ,
2022-06-13 16:27:03 +03:00
unsigned int issue_flags )
{
2022-08-11 10:11:15 +03:00
struct io_rw * rw = io_kiocb_to_cmd ( req , struct io_rw ) ;
2022-09-09 14:11:49 +03:00
unsigned final_ret = io_fixup_rw_res ( req , ret ) ;
2022-06-13 16:27:03 +03:00
2023-08-29 01:47:31 +03:00
if ( ret > = 0 & & req - > flags & REQ_F_CUR_POS )
2022-06-13 16:27:03 +03:00
req - > file - > f_pos = rw - > kiocb . ki_pos ;
2022-06-16 12:21:57 +03:00
if ( ret > = 0 & & ( rw - > kiocb . ki_complete = = io_complete_rw ) ) {
if ( ! __io_complete_rw_common ( req , ret ) ) {
2022-10-11 18:06:23 +03:00
/*
* Safe to call io_end from here as we ' re inline
* from the submission path .
*/
io_req_io_end ( req ) ;
2022-09-09 14:11:49 +03:00
io_req_set_res ( req , final_ret ,
2022-06-16 12:21:57 +03:00
io_put_kbuf ( req , issue_flags ) ) ;
2024-03-19 01:13:01 +03:00
io_req_rw_cleanup ( req , issue_flags ) ;
2022-06-16 12:21:57 +03:00
return IOU_OK ;
}
} else {
2022-06-13 16:27:03 +03:00
io_rw_done ( & rw - > kiocb , ret ) ;
2022-06-16 12:21:57 +03:00
}
2022-06-13 16:27:03 +03:00
if ( req - > flags & REQ_F_REISSUE ) {
req - > flags & = ~ REQ_F_REISSUE ;
2024-04-25 18:04:32 +03:00
io_resubmit_prep ( req ) ;
2024-03-19 01:13:01 +03:00
return - EAGAIN ;
2022-06-13 16:27:03 +03:00
}
2022-06-16 12:21:57 +03:00
return IOU_ISSUE_SKIP_COMPLETE ;
2022-06-13 16:27:03 +03:00
}
static inline loff_t * io_kiocb_ppos ( struct kiocb * kiocb )
{
return ( kiocb - > ki_filp - > f_mode & FMODE_STREAM ) ? NULL : & kiocb - > ki_pos ;
}
/*
* For files that don ' t have - > read_iter ( ) and - > write_iter ( ) , handle them
* by looping over - > read ( ) or - > write ( ) manually .
*/
static ssize_t loop_rw_iter ( int ddir , struct io_rw * rw , struct iov_iter * iter )
{
struct kiocb * kiocb = & rw - > kiocb ;
struct file * file = kiocb - > ki_filp ;
ssize_t ret = 0 ;
loff_t * ppos ;
/*
* Don ' t support polled IO through this interface , and we can ' t
* support non - blocking either . For the latter , this just causes
* the kiocb to be handled from an async context .
*/
if ( kiocb - > ki_flags & IOCB_HIPRI )
return - EOPNOTSUPP ;
if ( ( kiocb - > ki_flags & IOCB_NOWAIT ) & &
! ( kiocb - > ki_filp - > f_flags & O_NONBLOCK ) )
return - EAGAIN ;
ppos = io_kiocb_ppos ( kiocb ) ;
while ( iov_iter_count ( iter ) ) {
2023-03-29 18:16:45 +03:00
void __user * addr ;
size_t len ;
2022-06-13 16:27:03 +03:00
ssize_t nr ;
2023-01-05 22:07:32 +03:00
if ( iter_is_ubuf ( iter ) ) {
2023-03-29 18:16:45 +03:00
addr = iter - > ubuf + iter - > iov_offset ;
len = iov_iter_count ( iter ) ;
2023-01-05 22:07:32 +03:00
} else if ( ! iov_iter_is_bvec ( iter ) ) {
2023-03-29 18:16:45 +03:00
addr = iter_iov_addr ( iter ) ;
len = iter_iov_len ( iter ) ;
2022-06-13 16:27:03 +03:00
} else {
2023-03-29 18:16:45 +03:00
addr = u64_to_user_ptr ( rw - > addr ) ;
len = rw - > len ;
2022-06-13 16:27:03 +03:00
}
2023-03-29 18:16:45 +03:00
if ( ddir = = READ )
nr = file - > f_op - > read ( file , addr , len , ppos ) ;
else
nr = file - > f_op - > write ( file , addr , len , ppos ) ;
2022-06-13 16:27:03 +03:00
if ( nr < 0 ) {
if ( ! ret )
ret = nr ;
break ;
}
ret + = nr ;
if ( ! iov_iter_is_bvec ( iter ) ) {
iov_iter_advance ( iter , nr ) ;
} else {
rw - > addr + = nr ;
rw - > len - = nr ;
if ( ! rw - > len )
break ;
}
2023-03-29 18:16:45 +03:00
if ( nr ! = len )
2022-06-13 16:27:03 +03:00
break ;
}
return ret ;
}
/*
* This is our waitqueue callback handler , registered through __folio_lock_async ( )
* when we initially tried to do the IO with the iocb armed our waitqueue .
* This gets called when the page is unlocked , and we generally expect that to
* happen when the page IO is completed and the page is now uptodate . This will
* queue a task_work based retry of the operation , attempting to copy the data
* again . If the latter fails because the page was NOT uptodate , then we will
* do a thread based blocking retry of the operation . That ' s the unexpected
* slow path .
*/
static int io_async_buf_func ( struct wait_queue_entry * wait , unsigned mode ,
int sync , void * arg )
{
struct wait_page_queue * wpq ;
struct io_kiocb * req = wait - > private ;
2022-08-11 10:11:15 +03:00
struct io_rw * rw = io_kiocb_to_cmd ( req , struct io_rw ) ;
2022-06-13 16:27:03 +03:00
struct wait_page_key * key = arg ;
wpq = container_of ( wait , struct wait_page_queue , wait ) ;
if ( ! wake_page_match ( wpq , key ) )
return 0 ;
rw - > kiocb . ki_flags & = ~ IOCB_WAITQ ;
list_del_init ( & wait - > entry ) ;
io_req_task_queue ( req ) ;
return 1 ;
}
/*
* This controls whether a given IO request should be armed for async page
* based retry . If we return false here , the request is handed to the async
* worker threads for retry . If we ' re doing buffered reads on a regular file ,
* we prepare a private wait_page_queue entry and retry the operation . This
* will either succeed because the page is now uptodate and unlocked , or it
* will register a callback when the page is unlocked at IO completion . Through
* that callback , io_uring uses task_work to setup a retry of the operation .
* That retry will attempt the buffered read again . The retry will generally
* succeed , or in rare cases where it fails , we then fall back to using the
* async worker threads for a blocking retry .
*/
static bool io_rw_should_retry ( struct io_kiocb * req )
{
struct io_async_rw * io = req - > async_data ;
struct wait_page_queue * wait = & io - > wpq ;
2022-08-11 10:11:15 +03:00
struct io_rw * rw = io_kiocb_to_cmd ( req , struct io_rw ) ;
2022-06-13 16:27:03 +03:00
struct kiocb * kiocb = & rw - > kiocb ;
/* never retry for NOWAIT, we just complete with -EAGAIN */
if ( req - > flags & REQ_F_NOWAIT )
return false ;
/* Only for buffered IO */
if ( kiocb - > ki_flags & ( IOCB_DIRECT | IOCB_HIPRI ) )
return false ;
/*
* just use poll if we can , and don ' t attempt if the fs doesn ' t
* support callback based unlocks
*/
2024-03-28 15:27:24 +03:00
if ( io_file_can_poll ( req ) | |
! ( req - > file - > f_op - > fop_flags & FOP_BUFFER_RASYNC ) )
2022-06-13 16:27:03 +03:00
return false ;
wait - > wait . func = io_async_buf_func ;
wait - > wait . private = req ;
wait - > wait . flags = 0 ;
INIT_LIST_HEAD ( & wait - > wait . entry ) ;
kiocb - > ki_flags | = IOCB_WAITQ ;
kiocb - > ki_flags & = ~ IOCB_NOWAIT ;
kiocb - > ki_waitq = wait ;
return true ;
}
static inline int io_iter_do_read ( struct io_rw * rw , struct iov_iter * iter )
{
struct file * file = rw - > kiocb . ki_filp ;
if ( likely ( file - > f_op - > read_iter ) )
2023-08-28 18:13:18 +03:00
return file - > f_op - > read_iter ( & rw - > kiocb , iter ) ;
2022-06-13 16:27:03 +03:00
else if ( file - > f_op - > read )
return loop_rw_iter ( READ , rw , iter ) ;
else
return - EINVAL ;
}
2022-06-17 00:22:18 +03:00
static bool need_complete_io ( struct io_kiocb * req )
2022-06-13 16:27:03 +03:00
{
return req - > flags & REQ_F_ISREG | |
S_ISBLK ( file_inode ( req - > file ) - > i_mode ) ;
}
static int io_rw_init_file ( struct io_kiocb * req , fmode_t mode )
{
2022-08-11 10:11:15 +03:00
struct io_rw * rw = io_kiocb_to_cmd ( req , struct io_rw ) ;
2022-06-13 16:27:03 +03:00
struct kiocb * kiocb = & rw - > kiocb ;
struct io_ring_ctx * ctx = req - > ctx ;
struct file * file = req - > file ;
int ret ;
2024-01-29 06:52:21 +03:00
if ( unlikely ( ! ( file - > f_mode & mode ) ) )
2022-06-13 16:27:03 +03:00
return - EBADF ;
2023-06-20 14:32:31 +03:00
if ( ! ( req - > flags & REQ_F_FIXED_FILE ) )
2023-06-20 14:32:32 +03:00
req - > flags | = io_file_get_flags ( file ) ;
2022-06-13 16:27:03 +03:00
iov_iter work, part 1 - isolated cleanups and optimizations.
One of the goals is to reduce the overhead of using ->read_iter()
and ->write_iter() instead of ->read()/->write(); new_sync_{read,write}()
has a surprising amount of overhead, in particular inside iocb_flags().
That's why the beginning of the series is in this pile; it's not directly
iov_iter-related, but it's a part of the same work...
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
-----BEGIN PGP SIGNATURE-----
iHUEABYIAB0WIQQqUNBr3gm4hGXdBJlZ7Krx/gZQ6wUCYurGOQAKCRBZ7Krx/gZQ
6ysyAP91lvBfMRepcxpd9kvtuzWkU8A3rfSziZZteEHANB9Q7QEAiPn2a2OjWkcZ
uAyUWfCkHCNx+dSMkEvUgR5okQ0exAM=
=9UCV
-----END PGP SIGNATURE-----
Merge tag 'pull-work.iov_iter-base' of git://git.kernel.org/pub/scm/linux/kernel/git/viro/vfs
Pull vfs iov_iter updates from Al Viro:
"Part 1 - isolated cleanups and optimizations.
One of the goals is to reduce the overhead of using ->read_iter() and
->write_iter() instead of ->read()/->write().
new_sync_{read,write}() has a surprising amount of overhead, in
particular inside iocb_flags(). That's the explanation for the
beginning of the series is in this pile; it's not directly
iov_iter-related, but it's a part of the same work..."
* tag 'pull-work.iov_iter-base' of git://git.kernel.org/pub/scm/linux/kernel/git/viro/vfs:
first_iovec_segment(): just return address
iov_iter: massage calling conventions for first_{iovec,bvec}_segment()
iov_iter: first_{iovec,bvec}_segment() - simplify a bit
iov_iter: lift dealing with maxpages out of first_{iovec,bvec}_segment()
iov_iter_get_pages{,_alloc}(): cap the maxsize with MAX_RW_COUNT
iov_iter_bvec_advance(): don't bother with bvec_iter
copy_page_{to,from}_iter(): switch iovec variants to generic
keep iocb_flags() result cached in struct file
iocb: delay evaluation of IS_SYNC(...) until we want to check IOCB_DSYNC
struct file: use anonymous union member for rcuhead and llist
btrfs: use IOMAP_DIO_NOSYNC
teach iomap_dio_rw() to suppress dsync
No need of likely/unlikely on calls of check_copy_size()
2022-08-03 23:50:22 +03:00
kiocb - > ki_flags = file - > f_iocb_flags ;
2022-06-13 16:27:03 +03:00
ret = kiocb_set_rw_flags ( kiocb , rw - > flags ) ;
if ( unlikely ( ret ) )
return ret ;
2022-11-02 18:18:24 +03:00
kiocb - > ki_flags | = IOCB_ALLOC_CACHE ;
2022-06-13 16:27:03 +03:00
/*
* If the file is marked O_NONBLOCK , still allow retry for it if it
* supports async . Otherwise it ' s impossible to use O_NONBLOCK files
* reliably . If not , or it IOCB_NOWAIT is set , don ' t retry .
*/
if ( ( kiocb - > ki_flags & IOCB_NOWAIT ) | |
( ( file - > f_flags & O_NONBLOCK ) & & ! io_file_supports_nowait ( req ) ) )
req - > flags | = REQ_F_NOWAIT ;
if ( ctx - > flags & IORING_SETUP_IOPOLL ) {
if ( ! ( kiocb - > ki_flags & IOCB_DIRECT ) | | ! file - > f_op - > iopoll )
return - EOPNOTSUPP ;
kiocb - > private = NULL ;
2022-11-02 18:18:24 +03:00
kiocb - > ki_flags | = IOCB_HIPRI ;
2022-06-13 16:27:03 +03:00
kiocb - > ki_complete = io_complete_rw_iopoll ;
req - > iopoll_completed = 0 ;
} else {
if ( kiocb - > ki_flags & IOCB_HIPRI )
return - EINVAL ;
kiocb - > ki_complete = io_complete_rw ;
}
return 0 ;
}
2023-09-11 22:31:56 +03:00
static int __io_read ( struct io_kiocb * req , unsigned int issue_flags )
2022-06-13 16:27:03 +03:00
{
2024-03-19 01:13:01 +03:00
bool force_nonblock = issue_flags & IO_URING_F_NONBLOCK ;
2022-08-11 10:11:15 +03:00
struct io_rw * rw = io_kiocb_to_cmd ( req , struct io_rw ) ;
2024-03-19 01:13:01 +03:00
struct io_async_rw * io = req - > async_data ;
2022-06-13 16:27:03 +03:00
struct kiocb * kiocb = & rw - > kiocb ;
2024-03-19 01:13:01 +03:00
ssize_t ret ;
2022-06-13 16:27:03 +03:00
loff_t * ppos ;
2024-03-19 01:13:01 +03:00
if ( io_do_buffer_select ( req ) ) {
ret = io_import_iovec ( ITER_DEST , req , io , issue_flags ) ;
2022-06-13 16:27:03 +03:00
if ( unlikely ( ret < 0 ) )
return ret ;
}
2024-03-19 01:13:01 +03:00
2022-06-13 16:27:03 +03:00
ret = io_rw_init_file ( req , FMODE_READ ) ;
2024-03-19 01:13:01 +03:00
if ( unlikely ( ret ) )
2022-06-13 16:27:03 +03:00
return ret ;
2024-03-19 01:25:58 +03:00
req - > cqe . res = iov_iter_count ( & io - > iter ) ;
2022-06-13 16:27:03 +03:00
if ( force_nonblock ) {
/* If the file doesn't support async, just async punt */
2024-03-19 01:13:01 +03:00
if ( unlikely ( ! io_file_supports_nowait ( req ) ) )
return - EAGAIN ;
2022-06-13 16:27:03 +03:00
kiocb - > ki_flags | = IOCB_NOWAIT ;
} else {
/* Ensure we clear previously set non-block flag */
kiocb - > ki_flags & = ~ IOCB_NOWAIT ;
}
ppos = io_kiocb_update_pos ( req ) ;
ret = rw_verify_area ( READ , req - > file , ppos , req - > cqe . res ) ;
2024-03-19 01:13:01 +03:00
if ( unlikely ( ret ) )
2022-06-13 16:27:03 +03:00
return ret ;
2024-03-19 01:25:58 +03:00
ret = io_iter_do_read ( rw , & io - > iter ) ;
2022-06-13 16:27:03 +03:00
if ( ret = = - EAGAIN | | ( req - > flags & REQ_F_REISSUE ) ) {
req - > flags & = ~ REQ_F_REISSUE ;
2024-03-19 01:13:01 +03:00
/* If we can poll, just do that. */
if ( io_file_can_poll ( req ) )
2022-06-13 16:27:03 +03:00
return - EAGAIN ;
/* IOPOLL retry should happen for io-wq threads */
if ( ! force_nonblock & & ! ( req - > ctx - > flags & IORING_SETUP_IOPOLL ) )
goto done ;
/* no retry on NONBLOCK nor RWF_NOWAIT */
if ( req - > flags & REQ_F_NOWAIT )
goto done ;
ret = 0 ;
} else if ( ret = = - EIOCBQUEUED ) {
2022-06-16 12:21:57 +03:00
return IOU_ISSUE_SKIP_COMPLETE ;
2022-06-13 16:27:03 +03:00
} else if ( ret = = req - > cqe . res | | ret < = 0 | | ! force_nonblock | |
2022-06-17 00:22:18 +03:00
( req - > flags & REQ_F_NOWAIT ) | | ! need_complete_io ( req ) ) {
2022-06-13 16:27:03 +03:00
/* read all, failed, already did sync or don't want to retry */
goto done ;
}
/*
* Don ' t depend on the iter state matching what was consumed , or being
* untouched in case of error . Restore it and we ' ll advance it
* manually if we need to .
*/
2024-03-19 01:25:58 +03:00
iov_iter_restore ( & io - > iter , & io - > iter_state ) ;
2022-06-13 16:27:03 +03:00
do {
/*
* We end up here because of a partial read , either from
* above or inside this loop . Advance the iter by the bytes
* that were consumed .
*/
2024-03-19 01:25:58 +03:00
iov_iter_advance ( & io - > iter , ret ) ;
if ( ! iov_iter_count ( & io - > iter ) )
2022-06-13 16:27:03 +03:00
break ;
io - > bytes_done + = ret ;
2024-03-19 01:25:58 +03:00
iov_iter_save_state ( & io - > iter , & io - > iter_state ) ;
2022-06-13 16:27:03 +03:00
/* if we can retry, do so with the callbacks armed */
if ( ! io_rw_should_retry ( req ) ) {
kiocb - > ki_flags & = ~ IOCB_WAITQ ;
return - EAGAIN ;
}
2024-03-19 01:25:58 +03:00
req - > cqe . res = iov_iter_count ( & io - > iter ) ;
2022-06-13 16:27:03 +03:00
/*
* Now retry read with the IOCB_WAITQ parts set in the iocb . If
* we get - EIOCBQUEUED , then we ' ll get a notification when the
* desired page gets unlocked . We can also get a partial read
* here , and if we do , then just retry at the new offset .
*/
2024-03-19 01:25:58 +03:00
ret = io_iter_do_read ( rw , & io - > iter ) ;
2022-06-13 16:27:03 +03:00
if ( ret = = - EIOCBQUEUED )
return IOU_ISSUE_SKIP_COMPLETE ;
/* we got some bytes, but not all. retry. */
kiocb - > ki_flags & = ~ IOCB_WAITQ ;
2024-03-19 01:25:58 +03:00
iov_iter_restore ( & io - > iter , & io - > iter_state ) ;
2022-06-13 16:27:03 +03:00
} while ( ret > 0 ) ;
done :
/* it's faster to check here then delegate to kfree */
2023-09-11 22:31:56 +03:00
return ret ;
}
int io_read ( struct io_kiocb * req , unsigned int issue_flags )
{
int ret ;
ret = __io_read ( req , issue_flags ) ;
if ( ret > = 0 )
return kiocb_done ( req , ret , issue_flags ) ;
return ret ;
2022-06-13 16:27:03 +03:00
}
2023-09-11 22:35:42 +03:00
int io_read_mshot ( struct io_kiocb * req , unsigned int issue_flags )
{
2023-11-06 23:39:09 +03:00
struct io_rw * rw = io_kiocb_to_cmd ( req , struct io_rw ) ;
2023-09-11 22:35:42 +03:00
unsigned int cflags = 0 ;
int ret ;
/*
* Multishot MUST be used on a pollable file
*/
2024-01-29 06:08:24 +03:00
if ( ! io_file_can_poll ( req ) )
2023-09-11 22:35:42 +03:00
return - EBADFD ;
ret = __io_read ( req , issue_flags ) ;
2024-04-01 20:27:33 +03:00
/*
* If the file doesn ' t support proper NOWAIT , then disable multishot
* and stay in single shot mode .
*/
if ( ! io_file_supports_nowait ( req ) )
req - > flags & = ~ REQ_F_APOLL_MULTISHOT ;
2023-09-11 22:35:42 +03:00
/*
* If we get - EAGAIN , recycle our buffer and just let normal poll
* handling arm it .
*/
if ( ret = = - EAGAIN ) {
2023-11-06 23:39:09 +03:00
/*
* Reset rw - > len to 0 again to avoid clamping future mshot
* reads , in case the buffer size varies .
*/
if ( io_kbuf_recycle ( req , issue_flags ) )
rw - > len = 0 ;
2024-03-12 17:29:47 +03:00
if ( issue_flags & IO_URING_F_MULTISHOT )
return IOU_ISSUE_SKIP_COMPLETE ;
2023-09-11 22:35:42 +03:00
return - EAGAIN ;
}
/*
* Any successful return value will keep the multishot read armed .
*/
2024-04-01 20:27:33 +03:00
if ( ret > 0 & & req - > flags & REQ_F_APOLL_MULTISHOT ) {
2023-09-11 22:35:42 +03:00
/*
* Put our buffer and post a CQE . If we fail to post a CQE , then
* jump to the termination path . This request is then done .
*/
cflags = io_put_kbuf ( req , issue_flags ) ;
2023-11-06 23:39:09 +03:00
rw - > len = 0 ; /* similarly to above, reset len to 0 */
2023-09-11 22:35:42 +03:00
2024-03-19 01:00:31 +03:00
if ( io_req_post_cqe ( req , ret , cflags | IORING_CQE_F_MORE ) ) {
io_uring/rw: ensure poll based multishot read retries appropriately
io_read_mshot() always relies on poll triggering retries, and this works
fine as long as we do a retry per size of the buffer being read. The
buffer size is given by the size of the buffer(s) in the given buffer
group ID.
But if we're reading less than what is available, then we don't always
get to read everything that is available. For example, if the buffers
available are 32 bytes and we have 64 bytes to read, then we'll
correctly read the first 32 bytes and then wait for another poll trigger
before we attempt the next read. This next poll trigger may never
happen, in which case we just sit forever and never make progress, or it
may trigger at some point in the future, and now we're just delivering
the available data much later than we should have.
io_read_mshot() could do retries itself, but that is wasteful as we'll
be going through all of __io_read() again, and most likely in vain.
Rather than do that, bump our poll reference count and have
io_poll_check_events() do one more loop and check with vfs_poll() if we
have more data to read. If we do, io_read_mshot() will get invoked again
directly and we'll read the next chunk.
io_poll_multishot_retry() must only get called from inside
io_poll_issue(), which is our multishot retry handler, as we know we
already "own" the request at this point.
Cc: stable@vger.kernel.org
Link: https://github.com/axboe/liburing/issues/1041
Fixes: fc68fcda0491 ("io_uring/rw: add support for IORING_OP_READ_MULTISHOT")
Signed-off-by: Jens Axboe <axboe@kernel.dk>
2024-01-27 23:44:58 +03:00
if ( issue_flags & IO_URING_F_MULTISHOT ) {
/*
* Force retry , as we might have more data to
* be read and otherwise it won ' t get retried
* until ( if ever ) another poll is triggered .
*/
io_poll_multishot_retry ( req ) ;
2023-09-11 22:35:42 +03:00
return IOU_ISSUE_SKIP_COMPLETE ;
io_uring/rw: ensure poll based multishot read retries appropriately
io_read_mshot() always relies on poll triggering retries, and this works
fine as long as we do a retry per size of the buffer being read. The
buffer size is given by the size of the buffer(s) in the given buffer
group ID.
But if we're reading less than what is available, then we don't always
get to read everything that is available. For example, if the buffers
available are 32 bytes and we have 64 bytes to read, then we'll
correctly read the first 32 bytes and then wait for another poll trigger
before we attempt the next read. This next poll trigger may never
happen, in which case we just sit forever and never make progress, or it
may trigger at some point in the future, and now we're just delivering
the available data much later than we should have.
io_read_mshot() could do retries itself, but that is wasteful as we'll
be going through all of __io_read() again, and most likely in vain.
Rather than do that, bump our poll reference count and have
io_poll_check_events() do one more loop and check with vfs_poll() if we
have more data to read. If we do, io_read_mshot() will get invoked again
directly and we'll read the next chunk.
io_poll_multishot_retry() must only get called from inside
io_poll_issue(), which is our multishot retry handler, as we know we
already "own" the request at this point.
Cc: stable@vger.kernel.org
Link: https://github.com/axboe/liburing/issues/1041
Fixes: fc68fcda0491 ("io_uring/rw: add support for IORING_OP_READ_MULTISHOT")
Signed-off-by: Jens Axboe <axboe@kernel.dk>
2024-01-27 23:44:58 +03:00
}
2023-09-11 22:35:42 +03:00
return - EAGAIN ;
}
}
/*
* Either an error , or we ' ve hit overflow posting the CQE . For any
* multishot request , hitting overflow will terminate it .
*/
io_req_set_res ( req , ret , cflags ) ;
2024-03-19 01:13:01 +03:00
io_req_rw_cleanup ( req , issue_flags ) ;
2023-09-11 22:35:42 +03:00
if ( issue_flags & IO_URING_F_MULTISHOT )
return IOU_STOP_MULTISHOT ;
return IOU_OK ;
2022-06-13 16:27:03 +03:00
}
int io_write ( struct io_kiocb * req , unsigned int issue_flags )
{
2024-03-19 01:13:01 +03:00
bool force_nonblock = issue_flags & IO_URING_F_NONBLOCK ;
2022-08-11 10:11:15 +03:00
struct io_rw * rw = io_kiocb_to_cmd ( req , struct io_rw ) ;
2024-03-19 01:13:01 +03:00
struct io_async_rw * io = req - > async_data ;
2022-06-13 16:27:03 +03:00
struct kiocb * kiocb = & rw - > kiocb ;
ssize_t ret , ret2 ;
loff_t * ppos ;
ret = io_rw_init_file ( req , FMODE_WRITE ) ;
2024-03-19 01:13:01 +03:00
if ( unlikely ( ret ) )
2022-06-13 16:27:03 +03:00
return ret ;
2024-03-19 01:25:58 +03:00
req - > cqe . res = iov_iter_count ( & io - > iter ) ;
2022-06-13 16:27:03 +03:00
if ( force_nonblock ) {
/* If the file doesn't support async, just async punt */
if ( unlikely ( ! io_file_supports_nowait ( req ) ) )
2024-03-19 01:13:01 +03:00
goto ret_eagain ;
2022-06-13 16:27:03 +03:00
2024-03-28 15:27:24 +03:00
/* Check if we can support NOWAIT. */
2022-06-17 00:22:18 +03:00
if ( ! ( kiocb - > ki_flags & IOCB_DIRECT ) & &
2024-03-28 15:27:24 +03:00
! ( req - > file - > f_op - > fop_flags & FOP_BUFFER_WASYNC ) & &
( req - > flags & REQ_F_ISREG ) )
2024-03-19 01:13:01 +03:00
goto ret_eagain ;
2022-06-13 16:27:03 +03:00
kiocb - > ki_flags | = IOCB_NOWAIT ;
} else {
/* Ensure we clear previously set non-block flag */
kiocb - > ki_flags & = ~ IOCB_NOWAIT ;
}
ppos = io_kiocb_update_pos ( req ) ;
ret = rw_verify_area ( WRITE , req - > file , ppos , req - > cqe . res ) ;
2024-03-19 01:13:01 +03:00
if ( unlikely ( ret ) )
2022-06-16 12:21:57 +03:00
return ret ;
2022-06-13 16:27:03 +03:00
2023-08-17 17:13:34 +03:00
if ( req - > flags & REQ_F_ISREG )
kiocb_start_write ( kiocb ) ;
2022-06-13 16:27:03 +03:00
kiocb - > ki_flags | = IOCB_WRITE ;
if ( likely ( req - > file - > f_op - > write_iter ) )
2024-05-21 23:11:44 +03:00
ret2 = req - > file - > f_op - > write_iter ( kiocb , & io - > iter ) ;
2022-06-13 16:27:03 +03:00
else if ( req - > file - > f_op - > write )
2024-03-19 01:25:58 +03:00
ret2 = loop_rw_iter ( WRITE , rw , & io - > iter ) ;
2022-06-13 16:27:03 +03:00
else
ret2 = - EINVAL ;
if ( req - > flags & REQ_F_REISSUE ) {
req - > flags & = ~ REQ_F_REISSUE ;
ret2 = - EAGAIN ;
}
/*
* Raw bdev writes will return - EOPNOTSUPP for IOCB_NOWAIT . Just
* retry them without IOCB_NOWAIT .
*/
if ( ret2 = = - EOPNOTSUPP & & ( kiocb - > ki_flags & IOCB_NOWAIT ) )
ret2 = - EAGAIN ;
/* no retry on NONBLOCK nor RWF_NOWAIT */
if ( ret2 = = - EAGAIN & & ( req - > flags & REQ_F_NOWAIT ) )
goto done ;
if ( ! force_nonblock | | ret2 ! = - EAGAIN ) {
/* IOPOLL retry should happen for io-wq threads */
if ( ret2 = = - EAGAIN & & ( req - > ctx - > flags & IORING_SETUP_IOPOLL ) )
2024-03-19 01:13:01 +03:00
goto ret_eagain ;
2022-06-17 00:22:18 +03:00
if ( ret2 ! = req - > cqe . res & & ret2 > = 0 & & need_complete_io ( req ) ) {
2022-06-17 00:22:19 +03:00
trace_io_uring_short_write ( req - > ctx , kiocb - > ki_pos - ret2 ,
req - > cqe . res , ret2 ) ;
2022-06-17 00:22:18 +03:00
/* This is a partial write. The file pos has already been
* updated , setup the async struct to complete the request
* in the worker . Also update bytes_done to account for
* the bytes already written .
*/
2024-03-19 01:25:58 +03:00
iov_iter_save_state ( & io - > iter , & io - > iter_state ) ;
2024-03-19 01:13:01 +03:00
io - > bytes_done + = ret2 ;
2022-06-17 00:22:18 +03:00
2022-06-24 19:24:45 +03:00
if ( kiocb - > ki_flags & IOCB_WRITE )
2023-08-17 17:13:31 +03:00
io_req_end_write ( req ) ;
2024-03-19 01:13:01 +03:00
return - EAGAIN ;
2022-06-17 00:22:18 +03:00
}
2022-06-13 16:27:03 +03:00
done :
2024-03-23 05:41:18 +03:00
return kiocb_done ( req , ret2 , issue_flags ) ;
2022-06-13 16:27:03 +03:00
} else {
2024-03-19 01:13:01 +03:00
ret_eagain :
2024-03-19 01:25:58 +03:00
iov_iter_restore ( & io - > iter , & io - > iter_state ) ;
2024-03-19 01:13:01 +03:00
if ( kiocb - > ki_flags & IOCB_WRITE )
io_req_end_write ( req ) ;
return - EAGAIN ;
2022-06-13 16:27:03 +03:00
}
}
2022-09-21 14:17:47 +03:00
void io_rw_fail ( struct io_kiocb * req )
{
int res ;
res = io_fixup_rw_res ( req , req - > cqe . res ) ;
io_req_set_res ( req , res , req - > cqe . flags ) ;
}
2022-06-13 16:27:03 +03:00
int io_do_iopoll ( struct io_ring_ctx * ctx , bool force_nonspin )
{
struct io_wq_work_node * pos , * start , * prev ;
2023-03-20 22:49:26 +03:00
unsigned int poll_flags = 0 ;
2022-06-13 16:27:03 +03:00
DEFINE_IO_COMP_BATCH ( iob ) ;
int nr_events = 0 ;
/*
* Only spin for completions if we don ' t have multiple devices hanging
* off our complete list .
*/
if ( ctx - > poll_multi_queue | | force_nonspin )
poll_flags | = BLK_POLL_ONESHOT ;
wq_list_for_each ( pos , start , & ctx - > iopoll_list ) {
struct io_kiocb * req = container_of ( pos , struct io_kiocb , comp_list ) ;
2022-09-03 00:16:29 +03:00
struct file * file = req - > file ;
2022-06-13 16:27:03 +03:00
int ret ;
/*
* Move completed and retryable entries to our local lists .
* If we find a request that requires polling , break out
* and complete those lists first , if we have entries there .
*/
if ( READ_ONCE ( req - > iopoll_completed ) )
break ;
2022-08-23 19:14:41 +03:00
if ( req - > opcode = = IORING_OP_URING_CMD ) {
2022-09-03 00:16:29 +03:00
struct io_uring_cmd * ioucmd ;
2022-08-23 19:14:41 +03:00
2022-09-03 00:16:29 +03:00
ioucmd = io_kiocb_to_cmd ( req , struct io_uring_cmd ) ;
2022-09-03 00:18:05 +03:00
ret = file - > f_op - > uring_cmd_iopoll ( ioucmd , & iob ,
poll_flags ) ;
2022-09-03 00:16:29 +03:00
} else {
struct io_rw * rw = io_kiocb_to_cmd ( req , struct io_rw ) ;
ret = file - > f_op - > iopoll ( & rw - > kiocb , & iob , poll_flags ) ;
}
2022-06-13 16:27:03 +03:00
if ( unlikely ( ret < 0 ) )
return ret ;
else if ( ret )
poll_flags | = BLK_POLL_ONESHOT ;
/* iopoll may have completed current req */
if ( ! rq_list_empty ( iob . req_list ) | |
READ_ONCE ( req - > iopoll_completed ) )
break ;
}
if ( ! rq_list_empty ( iob . req_list ) )
iob . complete ( & iob ) ;
else if ( ! pos )
return 0 ;
prev = start ;
wq_list_for_each_resume ( pos , prev ) {
struct io_kiocb * req = container_of ( pos , struct io_kiocb , comp_list ) ;
/* order with io_complete_rw_iopoll(), e.g. ->result updates */
if ( ! smp_load_acquire ( & req - > iopoll_completed ) )
break ;
nr_events + + ;
req - > cqe . flags = io_put_kbuf ( req , 0 ) ;
2024-03-19 01:13:01 +03:00
if ( req - > opcode ! = IORING_OP_URING_CMD )
io_req_rw_cleanup ( req , 0 ) ;
2022-06-13 16:27:03 +03:00
}
if ( unlikely ( ! nr_events ) )
return 0 ;
pos = start ? start - > next : ctx - > iopoll_list . first ;
wq_list_cut ( & ctx - > iopoll_list , prev , start ) ;
2023-08-25 01:53:29 +03:00
if ( WARN_ON_ONCE ( ! wq_list_empty ( & ctx - > submit_state . compl_reqs ) ) )
return 0 ;
ctx - > submit_state . compl_reqs . first = pos ;
__io_submit_flush_completions ( ctx ) ;
2022-06-13 16:27:03 +03:00
return nr_events ;
}
2024-03-19 01:13:01 +03:00
2024-03-21 00:19:44 +03:00
void io_rw_cache_free ( const void * entry )
2024-03-19 01:13:01 +03:00
{
2024-03-21 00:19:44 +03:00
struct io_async_rw * rw = ( struct io_async_rw * ) entry ;
2024-03-19 01:13:01 +03:00
2024-03-19 01:31:44 +03:00
if ( rw - > free_iovec ) {
kasan_mempool_unpoison_object ( rw - > free_iovec ,
rw - > free_iov_nr * sizeof ( struct iovec ) ) ;
io_rw_iovec_free ( rw ) ;
}
2024-03-19 01:13:01 +03:00
kfree ( rw ) ;
}