* [PATCHSET v2] io_uring: Allow relative lookups
@ 2020-02-09 17:11 Jens Axboe
2020-02-09 17:11 ` [PATCH 1/3] io-wq: add support for inheriting ->fs Jens Axboe
` (2 more replies)
0 siblings, 3 replies; 6+ messages in thread
From: Jens Axboe @ 2020-02-09 17:11 UTC (permalink / raw)
To: io-uring
Due to an oversight on my part, AT_FDCWD lookups only work when the
lookup can be done inline, not async. This patchset rectifies that,
aiming for 5.6 for this one as it would be a shame to have openat etc
without that.
Just 3 small simple patches - add io-wq support for passing in fs and
setting it, and add a ->needs_fs to the opcode table list of
requirements for opcodes that could potentially do lookups.
Last patch just ensures we allow AT_FDCWD.
--
Jens Axboe
^ permalink raw reply [flat|nested] 6+ messages in thread* [PATCH 1/3] io-wq: add support for inheriting ->fs 2020-02-09 17:11 [PATCHSET v2] io_uring: Allow relative lookups Jens Axboe @ 2020-02-09 17:11 ` Jens Axboe 2020-02-11 3:31 ` Sasha Levin 2020-02-09 17:11 ` [PATCH 2/3] io_uring: grab ->fs as part of async preparation Jens Axboe 2020-02-09 17:11 ` [PATCH 3/3] io_uring: allow AT_FDCWD for non-file openat/openat2/statx Jens Axboe 2 siblings, 1 reply; 6+ messages in thread From: Jens Axboe @ 2020-02-09 17:11 UTC (permalink / raw) To: io-uring; +Cc: Jens Axboe, stable Some work items need this for relative path lookup, make it available like the other inherited credentials/mm/etc. Cc: stable@vger.kernel.org # 5.3+ Signed-off-by: Jens Axboe <axboe@kernel.dk> --- fs/io-wq.c | 19 +++++++++++++++---- fs/io-wq.h | 4 +++- 2 files changed, 18 insertions(+), 5 deletions(-) diff --git a/fs/io-wq.c b/fs/io-wq.c index cb60a42b9fdf..58b1891bcfe5 100644 --- a/fs/io-wq.c +++ b/fs/io-wq.c @@ -16,6 +16,7 @@ #include <linux/slab.h> #include <linux/kthread.h> #include <linux/rculist_nulls.h> +#include <linux/fs_struct.h> #include "io-wq.h" @@ -59,6 +60,7 @@ struct io_worker { const struct cred *cur_creds; const struct cred *saved_creds; struct files_struct *restore_files; + struct fs_struct *restore_fs; }; #if BITS_PER_LONG == 64 @@ -141,13 +143,17 @@ static bool __io_worker_unuse(struct io_wqe *wqe, struct io_worker *worker) worker->cur_creds = worker->saved_creds = NULL; } - if (current->files != worker->restore_files) { + if ((current->files != worker->restore_files) || + (current->fs != worker->restore_fs)) { __acquire(&wqe->lock); spin_unlock_irq(&wqe->lock); dropped_lock = true; task_lock(current); - current->files = worker->restore_files; + if (current->files != worker->restore_files) + current->files = worker->restore_files; + if (current->fs != worker->restore_fs) + current->fs = worker->restore_fs; task_unlock(current); } @@ -311,6 +317,7 @@ static void io_worker_start(struct io_wqe *wqe, struct io_worker *worker) worker->flags |= (IO_WORKER_F_UP | IO_WORKER_F_RUNNING); worker->restore_files = current->files; + worker->restore_fs = current->fs; io_wqe_inc_running(wqe, worker); } @@ -476,9 +483,13 @@ static void io_worker_handle_work(struct io_worker *worker) if (work->flags & IO_WQ_WORK_CB) work->func(&work); - if (work->files && current->files != work->files) { + if ((work->files && current->files != work->files) || + (work->fs && current->fs != work->fs)) { task_lock(current); - current->files = work->files; + if (work->files && current->files != work->files) + current->files = work->files; + if (work->fs && current->fs != work->fs) + current->fs = work->fs; task_unlock(current); } if (work->mm != worker->mm) diff --git a/fs/io-wq.h b/fs/io-wq.h index 50b3378febf2..f152ba677d8f 100644 --- a/fs/io-wq.h +++ b/fs/io-wq.h @@ -74,6 +74,7 @@ struct io_wq_work { struct files_struct *files; struct mm_struct *mm; const struct cred *creds; + struct fs_struct *fs; unsigned flags; }; @@ -81,10 +82,11 @@ struct io_wq_work { do { \ (work)->list.next = NULL; \ (work)->func = _func; \ - (work)->flags = 0; \ (work)->files = NULL; \ (work)->mm = NULL; \ (work)->creds = NULL; \ + (work)->fs = NULL; \ + (work)->flags = 0; \ } while (0) \ typedef void (get_work_fn)(struct io_wq_work *); -- 2.25.0 ^ permalink raw reply related [flat|nested] 6+ messages in thread
* Re: [PATCH 1/3] io-wq: add support for inheriting ->fs 2020-02-09 17:11 ` [PATCH 1/3] io-wq: add support for inheriting ->fs Jens Axboe @ 2020-02-11 3:31 ` Sasha Levin 0 siblings, 0 replies; 6+ messages in thread From: Sasha Levin @ 2020-02-11 3:31 UTC (permalink / raw) To: Sasha Levin, Jens Axboe, io-uring; +Cc: Jens Axboe, stable, stable Hi, [This is an automated email] This commit has been processed because it contains a -stable tag. The stable tag indicates that it's relevant for the following trees: 5.3+ The bot has tested the following trees: v5.5.2, v5.4.18. v5.5.2: Failed to apply! Possible dependencies: 05f3fb3c5397 ("io_uring: avoid ring quiesce for fixed file set unregister and update") 15b71abe7b52 ("io_uring: add support for IORING_OP_OPENAT") 66f4af93da57 ("io_uring: add support for probing opcodes") b5dba59e0cf7 ("io_uring: add support for IORING_OP_CLOSE") d3656344fea0 ("io_uring: add lookup table for various opcode needs") d63d1b5edb7b ("io_uring: add support for fallocate()") eddc7ef52a6b ("io_uring: add support for IORING_OP_STATX") f86cd20c9454 ("io_uring: fix linked command file table usage") v5.4.18: Failed to apply! Possible dependencies: 561fb04a6a22 ("io_uring: replace workqueue usage with io-wq") 771b53d033e8 ("io-wq: small threadpool implementation for io_uring") ba5290ccb6b5 ("io_uring: replace s->needs_lock with s->in_async") ba816ad61fdf ("io_uring: run dependent links inline if possible") c5def4ab8494 ("io-wq: add support for bounded vs unbunded work") c826bd7a743f ("io_uring: add set of tracing events") fcb323cc53e2 ("io_uring: io_uring: add support for async work inheriting files") NOTE: The patch will not be queued to stable trees until it is upstream. How should we proceed with this patch? -- Thanks, Sasha ^ permalink raw reply [flat|nested] 6+ messages in thread
* [PATCH 2/3] io_uring: grab ->fs as part of async preparation 2020-02-09 17:11 [PATCHSET v2] io_uring: Allow relative lookups Jens Axboe 2020-02-09 17:11 ` [PATCH 1/3] io-wq: add support for inheriting ->fs Jens Axboe @ 2020-02-09 17:11 ` Jens Axboe 2020-02-11 3:31 ` Sasha Levin 2020-02-09 17:11 ` [PATCH 3/3] io_uring: allow AT_FDCWD for non-file openat/openat2/statx Jens Axboe 2 siblings, 1 reply; 6+ messages in thread From: Jens Axboe @ 2020-02-09 17:11 UTC (permalink / raw) To: io-uring; +Cc: Jens Axboe, stable This passes it in to io-wq, so it assumes the right fs_struct when executing async work that may need to do lookups. Cc: stable@vger.kernel.org # 5.3+ Signed-off-by: Jens Axboe <axboe@kernel.dk> --- fs/io_uring.c | 28 ++++++++++++++++++++++++++++ 1 file changed, 28 insertions(+) diff --git a/fs/io_uring.c b/fs/io_uring.c index 1a3ca6577a10..2a7bb178986e 100644 --- a/fs/io_uring.c +++ b/fs/io_uring.c @@ -75,6 +75,7 @@ #include <linux/fsnotify.h> #include <linux/fadvise.h> #include <linux/eventpoll.h> +#include <linux/fs_struct.h> #define CREATE_TRACE_POINTS #include <trace/events/io_uring.h> @@ -611,6 +612,8 @@ struct io_op_def { unsigned not_supported : 1; /* needs file table */ unsigned file_table : 1; + /* needs ->fs */ + unsigned needs_fs : 1; }; static const struct io_op_def io_op_defs[] = { @@ -653,12 +656,14 @@ static const struct io_op_def io_op_defs[] = { .needs_mm = 1, .needs_file = 1, .unbound_nonreg_file = 1, + .needs_fs = 1, }, [IORING_OP_RECVMSG] = { .async_ctx = 1, .needs_mm = 1, .needs_file = 1, .unbound_nonreg_file = 1, + .needs_fs = 1, }, [IORING_OP_TIMEOUT] = { .async_ctx = 1, @@ -689,6 +694,7 @@ static const struct io_op_def io_op_defs[] = { .needs_file = 1, .fd_non_neg = 1, .file_table = 1, + .needs_fs = 1, }, [IORING_OP_CLOSE] = { .needs_file = 1, @@ -702,6 +708,7 @@ static const struct io_op_def io_op_defs[] = { .needs_mm = 1, .needs_file = 1, .fd_non_neg = 1, + .needs_fs = 1, }, [IORING_OP_READ] = { .needs_mm = 1, @@ -733,6 +740,7 @@ static const struct io_op_def io_op_defs[] = { .needs_file = 1, .fd_non_neg = 1, .file_table = 1, + .needs_fs = 1, }, [IORING_OP_EPOLL_CTL] = { .unbound_nonreg_file = 1, @@ -907,6 +915,16 @@ static inline void io_req_work_grab_env(struct io_kiocb *req, } if (!req->work.creds) req->work.creds = get_current_cred(); + if (!req->work.fs && def->needs_fs) { + spin_lock(¤t->fs->lock); + if (!current->fs->in_exec) { + req->work.fs = current->fs; + req->work.fs->users++; + } else { + req->work.flags |= IO_WQ_WORK_CANCEL; + } + spin_unlock(¤t->fs->lock); + } } static inline void io_req_work_drop_env(struct io_kiocb *req) @@ -919,6 +937,16 @@ static inline void io_req_work_drop_env(struct io_kiocb *req) put_cred(req->work.creds); req->work.creds = NULL; } + if (req->work.fs) { + struct fs_struct *fs = req->work.fs; + + spin_lock(&req->work.fs->lock); + if (--fs->users) + fs = NULL; + spin_unlock(&req->work.fs->lock); + if (fs) + free_fs_struct(fs); + } } static inline bool io_prep_async_work(struct io_kiocb *req, -- 2.25.0 ^ permalink raw reply related [flat|nested] 6+ messages in thread
* Re: [PATCH 2/3] io_uring: grab ->fs as part of async preparation 2020-02-09 17:11 ` [PATCH 2/3] io_uring: grab ->fs as part of async preparation Jens Axboe @ 2020-02-11 3:31 ` Sasha Levin 0 siblings, 0 replies; 6+ messages in thread From: Sasha Levin @ 2020-02-11 3:31 UTC (permalink / raw) To: Sasha Levin, Jens Axboe, io-uring; +Cc: Jens Axboe, stable, stable Hi, [This is an automated email] This commit has been processed because it contains a -stable tag. The stable tag indicates that it's relevant for the following trees: 5.3+ The bot has tested the following trees: v5.5.2, v5.4.18. v5.5.2: Failed to apply! Possible dependencies: 05f3fb3c5397 ("io_uring: avoid ring quiesce for fixed file set unregister and update") 15b71abe7b52 ("io_uring: add support for IORING_OP_OPENAT") 3a6820f2bb8a ("io_uring: add non-vectored read/write commands") 3e4827b05d2a ("io_uring: add support for epoll_ctl(2)") 4840e418c2fc ("io_uring: add IORING_OP_FADVISE") b5dba59e0cf7 ("io_uring: add support for IORING_OP_CLOSE") d3656344fea0 ("io_uring: add lookup table for various opcode needs") d63d1b5edb7b ("io_uring: add support for fallocate()") eddc7ef52a6b ("io_uring: add support for IORING_OP_STATX") v5.4.18: Failed to apply! Possible dependencies: 15b71abe7b52 ("io_uring: add support for IORING_OP_OPENAT") 2665abfd757f ("io_uring: add support for linked SQE timeouts") 3e4827b05d2a ("io_uring: add support for epoll_ctl(2)") 3fbb51c18f5c ("io_uring: move all prep state for IORING_OP_CONNECT to prep handler") 4840e418c2fc ("io_uring: add IORING_OP_FADVISE") 561fb04a6a22 ("io_uring: replace workqueue usage with io-wq") 771b53d033e8 ("io-wq: small threadpool implementation for io_uring") 8ed8d3c3bc32 ("io_uring: any deferred command must have stable sqe data") 9adbd45d6d32 ("io_uring: add and use struct io_rw for read/writes") ad8a48acc23c ("io_uring: make req->timeout be dynamically allocated") b29472ee7b53 ("io_uring: make IORING_OP_TIMEOUT_REMOVE deferrable") ba5290ccb6b5 ("io_uring: replace s->needs_lock with s->in_async") ba816ad61fdf ("io_uring: run dependent links inline if possible") c826bd7a743f ("io_uring: add set of tracing events") e47293fdf989 ("io_uring: move all prep state for IORING_OP_{SEND,RECV}_MGS to prep handler") fbf23849b172 ("io_uring: make IORING_OP_CANCEL_ASYNC deferrable") fcb323cc53e2 ("io_uring: io_uring: add support for async work inheriting files") NOTE: The patch will not be queued to stable trees until it is upstream. How should we proceed with this patch? -- Thanks, Sasha ^ permalink raw reply [flat|nested] 6+ messages in thread
* [PATCH 3/3] io_uring: allow AT_FDCWD for non-file openat/openat2/statx 2020-02-09 17:11 [PATCHSET v2] io_uring: Allow relative lookups Jens Axboe 2020-02-09 17:11 ` [PATCH 1/3] io-wq: add support for inheriting ->fs Jens Axboe 2020-02-09 17:11 ` [PATCH 2/3] io_uring: grab ->fs as part of async preparation Jens Axboe @ 2020-02-09 17:11 ` Jens Axboe 2 siblings, 0 replies; 6+ messages in thread From: Jens Axboe @ 2020-02-09 17:11 UTC (permalink / raw) To: io-uring; +Cc: Jens Axboe Don't just check for dirfd == -1, we should allow AT_FDCWD as well for relative lookups. Signed-off-by: Jens Axboe <axboe@kernel.dk> --- fs/io_uring.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/fs/io_uring.c b/fs/io_uring.c index 2a7bb178986e..e6247b94c29d 100644 --- a/fs/io_uring.c +++ b/fs/io_uring.c @@ -4543,7 +4543,7 @@ static int io_req_needs_file(struct io_kiocb *req, int fd) { if (!io_op_defs[req->opcode].needs_file) return 0; - if (fd == -1 && io_op_defs[req->opcode].fd_non_neg) + if ((fd == -1 || fd == AT_FDCWD) && io_op_defs[req->opcode].fd_non_neg) return 0; return 1; } -- 2.25.0 ^ permalink raw reply related [flat|nested] 6+ messages in thread
end of thread, other threads:[~2020-02-11 3:31 UTC | newest] Thread overview: 6+ messages (download: mbox.gz / follow: Atom feed) -- links below jump to the message on this page -- 2020-02-09 17:11 [PATCHSET v2] io_uring: Allow relative lookups Jens Axboe 2020-02-09 17:11 ` [PATCH 1/3] io-wq: add support for inheriting ->fs Jens Axboe 2020-02-11 3:31 ` Sasha Levin 2020-02-09 17:11 ` [PATCH 2/3] io_uring: grab ->fs as part of async preparation Jens Axboe 2020-02-11 3:31 ` Sasha Levin 2020-02-09 17:11 ` [PATCH 3/3] io_uring: allow AT_FDCWD for non-file openat/openat2/statx Jens Axboe
This is a public inbox, see mirroring instructions for how to clone and mirror all data and code used for this inbox; as well as URLs for NNTP newsgroup(s).