Hi Greg/stable, Can you queue these up? Two patches for each stable kernel, all named appropriately. Thanks! -------- Forwarded Message -------- Subject: Re: [PATCH 5.10/5.15] io_uring: fix registered files leak Date: Wed, 13 Mar 2024 18:40:29 -0600 From: Jens Axboe <axboe@xxxxxxxxx> To: Fedor Pchelkin <pchelkin@xxxxxxxxx> CC: Pavel Begunkov <asml.silence@xxxxxxxxx>, Greg Kroah-Hartman <gregkh@xxxxxxxxxxxxxxxxxxx>, stable@xxxxxxxxxxxxxxx, io-uring@xxxxxxxxxxxxxxx, linux-kernel@xxxxxxxxxxxxxxx, Alexey Khoroshilov <khoroshilov@xxxxxxxxx>, lvc-project@xxxxxxxxxxxxxxxx, Nikita Zhandarovich <n.zhandarovich@xxxxxxxxxx>, Roman Belyaev <belyaevrd@xxxxxxxxx> Hi, OK, here they are. Two patches attached for every stable kernel, that gets rid of the remnants of the SCM related code: 5.4 5.10 and 5.15 (same patches) 6.1 6.6 6.7 Would appreciate if Fedor and Pavel could give them a once over, but I think they are all fine. It's just deleting the code... Barring no complaints, Greg can you queue them up? Thanks! -- Jens Axboe -- Jens Axboe
From 53e9e0ced4e4146ff57635c8ee65a57f3ed33fe3 Mon Sep 17 00:00:00 2001 From: Jens Axboe <axboe@xxxxxxxxx> Date: Wed, 13 Mar 2024 18:33:48 -0600 Subject: [PATCH 1/2] io_uring/unix: drop usage of io_uring socket Commit a4104821ad651d8a0b374f0b2474c345bbb42f82 upstream. Since we no longer allow sending io_uring fds over SCM_RIGHTS, move to using io_is_uring_fops() to detect whether this is a io_uring fd or not. With that done, kill off io_uring_get_socket() as nobody calls it anymore. This is in preparation to yanking out the rest of the core related to unix gc with io_uring. Signed-off-by: Jens Axboe <axboe@xxxxxxxxx> --- fs/io_uring.c | 18 +++++------------- include/linux/fs.h | 6 +++--- net/core/scm.c | 2 +- net/unix/scm.c | 4 +--- 4 files changed, 10 insertions(+), 20 deletions(-) diff --git a/fs/io_uring.c b/fs/io_uring.c index 9de8961763b0..cd1858cc6320 100644 --- a/fs/io_uring.c +++ b/fs/io_uring.c @@ -381,19 +381,6 @@ static struct kmem_cache *req_cachep; static const struct file_operations io_uring_fops; -struct sock *io_uring_get_socket(struct file *file) -{ -#if defined(CONFIG_UNIX) - if (file->f_op == &io_uring_fops) { - struct io_ring_ctx *ctx = file->private_data; - - return ctx->ring_sock->sk; - } -#endif - return NULL; -} -EXPORT_SYMBOL(io_uring_get_socket); - static void io_ring_ctx_ref_free(struct percpu_ref *ref) { struct io_ring_ctx *ctx = container_of(ref, struct io_ring_ctx, refs); @@ -3794,6 +3781,11 @@ static const struct file_operations io_uring_fops = { .fasync = io_uring_fasync, }; +bool io_is_uring_fops(struct file *file) +{ + return file->f_op == &io_uring_fops; +} + static int io_allocate_scq_urings(struct io_ring_ctx *ctx, struct io_uring_params *p) { diff --git a/include/linux/fs.h b/include/linux/fs.h index e009b52ab6b0..272f261894b1 100644 --- a/include/linux/fs.h +++ b/include/linux/fs.h @@ -3614,11 +3614,11 @@ extern int generic_fadvise(struct file *file, loff_t offset, loff_t len, int advice); #if defined(CONFIG_IO_URING) -extern struct sock *io_uring_get_socket(struct file *file); +bool io_is_uring_fops(struct file *file); #else -static inline struct sock *io_uring_get_socket(struct file *file) +static inline bool io_is_uring_fops(struct file *file) { - return NULL; + return false; } #endif diff --git a/net/core/scm.c b/net/core/scm.c index 5525c14f33f1..a442bf63cd48 100644 --- a/net/core/scm.c +++ b/net/core/scm.c @@ -105,7 +105,7 @@ static int scm_fp_copy(struct cmsghdr *cmsg, struct scm_fp_list **fplp) if (fd < 0 || !(file = fget_raw(fd))) return -EBADF; /* don't allow io_uring files */ - if (io_uring_get_socket(file)) { + if (io_is_uring_fops(file)) { fput(file); return -EINVAL; } diff --git a/net/unix/scm.c b/net/unix/scm.c index e881a6e78af5..51b623de3be5 100644 --- a/net/unix/scm.c +++ b/net/unix/scm.c @@ -33,10 +33,8 @@ struct sock *unix_get_socket(struct file *filp) /* PF_UNIX ? */ if (s && sock->ops && sock->ops->family == PF_UNIX) u_sock = s; - } else { - /* Could be an io_uring instance */ - u_sock = io_uring_get_socket(filp); } + return u_sock; } EXPORT_SYMBOL(unix_get_socket); -- 2.43.0
From a21d37be81f8dc59d4cda4aac37312f0774e4b71 Mon Sep 17 00:00:00 2001 From: Jens Axboe <axboe@xxxxxxxxx> Date: Wed, 13 Mar 2024 18:36:18 -0600 Subject: [PATCH 2/2] io_uring: drop any code related to SCM_RIGHTS Commit 6e5e6d274956305f1fc0340522b38f5f5be74bdb upstream. This is dead code after we dropped support for passing io_uring fds over SCM_RIGHTS, get rid of it. Signed-off-by: Jens Axboe <axboe@xxxxxxxxx> --- fs/io_uring.c | 47 +++-------------------------------------------- 1 file changed, 3 insertions(+), 44 deletions(-) diff --git a/fs/io_uring.c b/fs/io_uring.c index cd1858cc6320..2c793e4ccf09 100644 --- a/fs/io_uring.c +++ b/fs/io_uring.c @@ -63,7 +63,6 @@ #include <linux/net.h> #include <net/sock.h> #include <net/af_unix.h> -#include <net/scm.h> #include <linux/anon_inodes.h> #include <linux/sched/mm.h> #include <linux/uaccess.h> @@ -264,10 +263,6 @@ struct io_ring_ctx { struct async_list pending_async[2]; -#if defined(CONFIG_UNIX) - struct socket *ring_sock; -#endif - struct list_head task_list; spinlock_t task_lock; }; @@ -3067,20 +3062,10 @@ static int io_cqring_wait(struct io_ring_ctx *ctx, int min_events, static void __io_sqe_files_unregister(struct io_ring_ctx *ctx) { -#if defined(CONFIG_UNIX) - if (ctx->ring_sock) { - struct sock *sock = ctx->ring_sock->sk; - struct sk_buff *skb; - - while ((skb = skb_dequeue(&sock->sk_receive_queue)) != NULL) - kfree_skb(skb); - } -#else int i; for (i = 0; i < ctx->nr_user_files; i++) fput(ctx->user_files[i]); -#endif } static int io_sqe_files_unregister(struct io_ring_ctx *ctx) @@ -3575,13 +3560,6 @@ static void io_ring_ctx_free(struct io_ring_ctx *ctx) io_sqe_files_unregister(ctx); io_eventfd_unregister(ctx); -#if defined(CONFIG_UNIX) - if (ctx->ring_sock) { - ctx->ring_sock->file = NULL; /* so that iput() is called */ - sock_release(ctx->ring_sock); - } -#endif - io_mem_free(ctx->rings); io_mem_free(ctx->sq_sqes); @@ -3833,45 +3811,26 @@ static int io_allocate_scq_urings(struct io_ring_ctx *ctx, /* * Allocate an anonymous fd, this is what constitutes the application * visible backing of an io_uring instance. The application mmaps this - * fd to gain access to the SQ/CQ ring details. If UNIX sockets are enabled, - * we have to tie this fd to a socket for file garbage collection purposes. + * fd to gain access to the SQ/CQ ring details. */ static int io_uring_get_fd(struct io_ring_ctx *ctx) { struct file *file; int ret; -#if defined(CONFIG_UNIX) - ret = sock_create_kern(&init_net, PF_UNIX, SOCK_RAW, IPPROTO_IP, - &ctx->ring_sock); - if (ret) - return ret; -#endif - ret = get_unused_fd_flags(O_RDWR | O_CLOEXEC); if (ret < 0) - goto err; + return ret; file = anon_inode_getfile("[io_uring]", &io_uring_fops, ctx, O_RDWR | O_CLOEXEC); if (IS_ERR(file)) { put_unused_fd(ret); - ret = PTR_ERR(file); - goto err; + return PTR_ERR(file); } -#if defined(CONFIG_UNIX) - ctx->ring_sock->file = file; - ctx->ring_sock->sk->sk_user_data = ctx; -#endif fd_install(ret, file); return ret; -err: -#if defined(CONFIG_UNIX) - sock_release(ctx->ring_sock); - ctx->ring_sock = NULL; -#endif - return ret; } static int io_uring_create(unsigned entries, struct io_uring_params *p) -- 2.43.0
From 25b0645afe6459437da6cff71546984597f209da Mon Sep 17 00:00:00 2001 From: Jens Axboe <axboe@xxxxxxxxx> Date: Tue, 19 Dec 2023 12:30:43 -0700 Subject: [PATCH 1/2] io_uring/unix: drop usage of io_uring socket Commit a4104821ad651d8a0b374f0b2474c345bbb42f82 upstream. Since we no longer allow sending io_uring fds over SCM_RIGHTS, move to using io_is_uring_fops() to detect whether this is a io_uring fd or not. With that done, kill off io_uring_get_socket() as nobody calls it anymore. This is in preparation to yanking out the rest of the core related to unix gc with io_uring. Signed-off-by: Jens Axboe <axboe@xxxxxxxxx> --- include/linux/io_uring.h | 9 +++++---- io_uring/io_uring.c | 13 ------------- io_uring/io_uring.h | 1 - net/core/scm.c | 2 +- net/unix/scm.c | 4 +--- 5 files changed, 7 insertions(+), 22 deletions(-) diff --git a/include/linux/io_uring.h b/include/linux/io_uring.h index aefb73eeeebf..b3a32687f514 100644 --- a/include/linux/io_uring.h +++ b/include/linux/io_uring.h @@ -93,6 +93,7 @@ int io_uring_cmd_sock(struct io_uring_cmd *cmd, unsigned int issue_flags); void io_uring_cmd_mark_cancelable(struct io_uring_cmd *cmd, unsigned int issue_flags); struct task_struct *io_uring_cmd_get_task(struct io_uring_cmd *cmd); +bool io_is_uring_fops(struct file *file); #else static inline int io_uring_cmd_import_fixed(u64 ubuf, unsigned long len, int rw, struct iov_iter *iter, void *ioucmd) @@ -111,10 +112,6 @@ static inline void io_uring_cmd_do_in_task_lazy(struct io_uring_cmd *ioucmd, void (*task_work_cb)(struct io_uring_cmd *, unsigned)) { } -static inline struct sock *io_uring_get_socket(struct file *file) -{ - return NULL; -} static inline void io_uring_task_cancel(void) { } @@ -141,6 +138,10 @@ static inline struct task_struct *io_uring_cmd_get_task(struct io_uring_cmd *cmd { return NULL; } +static inline bool io_is_uring_fops(struct file *file) +{ + return false; +} #endif #endif diff --git a/io_uring/io_uring.c b/io_uring/io_uring.c index 59f5791c90c3..06bd8795a87d 100644 --- a/io_uring/io_uring.c +++ b/io_uring/io_uring.c @@ -177,19 +177,6 @@ static struct ctl_table kernel_io_uring_disabled_table[] = { }; #endif -struct sock *io_uring_get_socket(struct file *file) -{ -#if defined(CONFIG_UNIX) - if (io_is_uring_fops(file)) { - struct io_ring_ctx *ctx = file->private_data; - - return ctx->ring_sock->sk; - } -#endif - return NULL; -} -EXPORT_SYMBOL(io_uring_get_socket); - static inline void io_submit_flush_completions(struct io_ring_ctx *ctx) { if (!wq_list_empty(&ctx->submit_state.compl_reqs) || diff --git a/io_uring/io_uring.h b/io_uring/io_uring.h index c9992cd7f138..0d66a7058dbe 100644 --- a/io_uring/io_uring.h +++ b/io_uring/io_uring.h @@ -61,7 +61,6 @@ struct file *io_file_get_fixed(struct io_kiocb *req, int fd, unsigned issue_flags); void __io_req_task_work_add(struct io_kiocb *req, unsigned flags); -bool io_is_uring_fops(struct file *file); bool io_alloc_async_data(struct io_kiocb *req); void io_req_task_queue(struct io_kiocb *req); void io_queue_iowq(struct io_kiocb *req, struct io_tw_state *ts_dont_use); diff --git a/net/core/scm.c b/net/core/scm.c index 7dc47c17d863..737917c7ac62 100644 --- a/net/core/scm.c +++ b/net/core/scm.c @@ -105,7 +105,7 @@ static int scm_fp_copy(struct cmsghdr *cmsg, struct scm_fp_list **fplp) if (fd < 0 || !(file = fget_raw(fd))) return -EBADF; /* don't allow io_uring files */ - if (io_uring_get_socket(file)) { + if (io_is_uring_fops(file)) { fput(file); return -EINVAL; } diff --git a/net/unix/scm.c b/net/unix/scm.c index 6ff628f2349f..822ce0d0d791 100644 --- a/net/unix/scm.c +++ b/net/unix/scm.c @@ -35,10 +35,8 @@ struct sock *unix_get_socket(struct file *filp) /* PF_UNIX ? */ if (s && ops && ops->family == PF_UNIX) u_sock = s; - } else { - /* Could be an io_uring instance */ - u_sock = io_uring_get_socket(filp); } + return u_sock; } EXPORT_SYMBOL(unix_get_socket); -- 2.43.0
From e742473ca70a1fb2492451bf4959d0e94a0c1c7c Mon Sep 17 00:00:00 2001 From: Jens Axboe <axboe@xxxxxxxxx> Date: Tue, 19 Dec 2023 12:36:34 -0700 Subject: [PATCH 2/2] io_uring: drop any code related to SCM_RIGHTS Commit 6e5e6d274956305f1fc0340522b38f5f5be74bdb upstream. This is dead code after we dropped support for passing io_uring fds over SCM_RIGHTS, get rid of it. Signed-off-by: Jens Axboe <axboe@xxxxxxxxx> --- include/linux/io_uring_types.h | 3 - io_uring/filetable.c | 11 +-- io_uring/io_uring.c | 32 +------ io_uring/rsrc.c | 169 +-------------------------------- io_uring/rsrc.h | 15 --- 5 files changed, 10 insertions(+), 220 deletions(-) diff --git a/include/linux/io_uring_types.h b/include/linux/io_uring_types.h index 239a4f68801b..335eca49dc8b 100644 --- a/include/linux/io_uring_types.h +++ b/include/linux/io_uring_types.h @@ -358,9 +358,6 @@ struct io_ring_ctx { struct wait_queue_head rsrc_quiesce_wq; unsigned rsrc_quiesce; - #if defined(CONFIG_UNIX) - struct socket *ring_sock; - #endif /* hashed buffered write serialization */ struct io_wq_hash *hash_map; diff --git a/io_uring/filetable.c b/io_uring/filetable.c index e7d749991de4..6e86e6188dbe 100644 --- a/io_uring/filetable.c +++ b/io_uring/filetable.c @@ -87,13 +87,10 @@ static int io_install_fixed_file(struct io_ring_ctx *ctx, struct file *file, io_file_bitmap_clear(&ctx->file_table, slot_index); } - ret = io_scm_file_account(ctx, file); - if (!ret) { - *io_get_tag_slot(ctx->file_data, slot_index) = 0; - io_fixed_file_set(file_slot, file); - io_file_bitmap_set(&ctx->file_table, slot_index); - } - return ret; + *io_get_tag_slot(ctx->file_data, slot_index) = 0; + io_fixed_file_set(file_slot, file); + io_file_bitmap_set(&ctx->file_table, slot_index); + return 0; } int __io_fixed_fd_install(struct io_ring_ctx *ctx, struct file *file, diff --git a/io_uring/io_uring.c b/io_uring/io_uring.c index 06bd8795a87d..f8d145fb40bb 100644 --- a/io_uring/io_uring.c +++ b/io_uring/io_uring.c @@ -60,7 +60,6 @@ #include <linux/net.h> #include <net/sock.h> #include <net/af_unix.h> -#include <net/scm.h> #include <linux/anon_inodes.h> #include <linux/sched/mm.h> #include <linux/uaccess.h> @@ -2939,13 +2938,6 @@ static __cold void io_ring_ctx_free(struct io_ring_ctx *ctx) io_rsrc_node_destroy(ctx, ctx->rsrc_node); WARN_ON_ONCE(!list_empty(&ctx->rsrc_ref_list)); - -#if defined(CONFIG_UNIX) - if (ctx->ring_sock) { - ctx->ring_sock->file = NULL; /* so that iput() is called */ - sock_release(ctx->ring_sock); - } -#endif WARN_ON_ONCE(!list_empty(&ctx->ltimeout_list)); io_alloc_cache_free(&ctx->rsrc_node_cache, io_rsrc_node_cache_free); @@ -3867,32 +3859,12 @@ static int io_uring_install_fd(struct file *file) /* * Allocate an anonymous fd, this is what constitutes the application * visible backing of an io_uring instance. The application mmaps this - * fd to gain access to the SQ/CQ ring details. If UNIX sockets are enabled, - * we have to tie this fd to a socket for file garbage collection purposes. + * fd to gain access to the SQ/CQ ring details. */ static struct file *io_uring_get_file(struct io_ring_ctx *ctx) { - struct file *file; -#if defined(CONFIG_UNIX) - int ret; - - ret = sock_create_kern(&init_net, PF_UNIX, SOCK_RAW, IPPROTO_IP, - &ctx->ring_sock); - if (ret) - return ERR_PTR(ret); -#endif - - file = anon_inode_getfile_secure("[io_uring]", &io_uring_fops, ctx, + return anon_inode_getfile_secure("[io_uring]", &io_uring_fops, ctx, O_RDWR | O_CLOEXEC, NULL); -#if defined(CONFIG_UNIX) - if (IS_ERR(file)) { - sock_release(ctx->ring_sock); - ctx->ring_sock = NULL; - } else { - ctx->ring_sock->file = file; - } -#endif - return file; } static __cold int io_uring_create(unsigned entries, struct io_uring_params *p, diff --git a/io_uring/rsrc.c b/io_uring/rsrc.c index f521c5965a93..4818b79231dd 100644 --- a/io_uring/rsrc.c +++ b/io_uring/rsrc.c @@ -24,7 +24,6 @@ struct io_rsrc_update { }; static void io_rsrc_buf_put(struct io_ring_ctx *ctx, struct io_rsrc_put *prsrc); -static void io_rsrc_file_put(struct io_ring_ctx *ctx, struct io_rsrc_put *prsrc); static int io_sqe_buffer_register(struct io_ring_ctx *ctx, struct iovec *iov, struct io_mapped_ubuf **pimu, struct page **last_hpage); @@ -157,7 +156,7 @@ static void io_rsrc_put_work(struct io_rsrc_node *node) switch (node->type) { case IORING_RSRC_FILE: - io_rsrc_file_put(node->ctx, prsrc); + fput(prsrc->file); break; case IORING_RSRC_BUFFER: io_rsrc_buf_put(node->ctx, prsrc); @@ -402,23 +401,13 @@ static int __io_sqe_files_update(struct io_ring_ctx *ctx, break; } /* - * Don't allow io_uring instances to be registered. If - * UNIX isn't enabled, then this causes a reference - * cycle and this instance can never get freed. If UNIX - * is enabled we'll handle it just fine, but there's - * still no point in allowing a ring fd as it doesn't - * support regular read/write anyway. + * Don't allow io_uring instances to be registered. */ if (io_is_uring_fops(file)) { fput(file); err = -EBADF; break; } - err = io_scm_file_account(ctx, file); - if (err) { - fput(file); - break; - } *io_get_tag_slot(data, i) = tag; io_fixed_file_set(file_slot, file); io_file_bitmap_set(&ctx->file_table, i); @@ -675,22 +664,12 @@ void __io_sqe_files_unregister(struct io_ring_ctx *ctx) for (i = 0; i < ctx->nr_user_files; i++) { struct file *file = io_file_from_index(&ctx->file_table, i); - /* skip scm accounted files, they'll be freed by ->ring_sock */ - if (!file || io_file_need_scm(file)) + if (!file) continue; io_file_bitmap_clear(&ctx->file_table, i); fput(file); } -#if defined(CONFIG_UNIX) - if (ctx->ring_sock) { - struct sock *sock = ctx->ring_sock->sk; - struct sk_buff *skb; - - while ((skb = skb_dequeue(&sock->sk_receive_queue)) != NULL) - kfree_skb(skb); - } -#endif io_free_file_tables(&ctx->file_table); io_file_table_set_alloc_range(ctx, 0, 0); io_rsrc_data_free(ctx->file_data); @@ -718,137 +697,6 @@ int io_sqe_files_unregister(struct io_ring_ctx *ctx) return ret; } -/* - * Ensure the UNIX gc is aware of our file set, so we are certain that - * the io_uring can be safely unregistered on process exit, even if we have - * loops in the file referencing. We account only files that can hold other - * files because otherwise they can't form a loop and so are not interesting - * for GC. - */ -int __io_scm_file_account(struct io_ring_ctx *ctx, struct file *file) -{ -#if defined(CONFIG_UNIX) - struct sock *sk = ctx->ring_sock->sk; - struct sk_buff_head *head = &sk->sk_receive_queue; - struct scm_fp_list *fpl; - struct sk_buff *skb; - - if (likely(!io_file_need_scm(file))) - return 0; - - /* - * See if we can merge this file into an existing skb SCM_RIGHTS - * file set. If there's no room, fall back to allocating a new skb - * and filling it in. - */ - spin_lock_irq(&head->lock); - skb = skb_peek(head); - if (skb && UNIXCB(skb).fp->count < SCM_MAX_FD) - __skb_unlink(skb, head); - else - skb = NULL; - spin_unlock_irq(&head->lock); - - if (!skb) { - fpl = kzalloc(sizeof(*fpl), GFP_KERNEL); - if (!fpl) - return -ENOMEM; - - skb = alloc_skb(0, GFP_KERNEL); - if (!skb) { - kfree(fpl); - return -ENOMEM; - } - - fpl->user = get_uid(current_user()); - fpl->max = SCM_MAX_FD; - fpl->count = 0; - - UNIXCB(skb).fp = fpl; - skb->sk = sk; - skb->destructor = io_uring_destruct_scm; - refcount_add(skb->truesize, &sk->sk_wmem_alloc); - } - - fpl = UNIXCB(skb).fp; - fpl->fp[fpl->count++] = get_file(file); - unix_inflight(fpl->user, file); - skb_queue_head(head, skb); - fput(file); -#endif - return 0; -} - -static __cold void io_rsrc_file_scm_put(struct io_ring_ctx *ctx, struct file *file) -{ -#if defined(CONFIG_UNIX) - struct sock *sock = ctx->ring_sock->sk; - struct sk_buff_head list, *head = &sock->sk_receive_queue; - struct sk_buff *skb; - int i; - - __skb_queue_head_init(&list); - - /* - * Find the skb that holds this file in its SCM_RIGHTS. When found, - * remove this entry and rearrange the file array. - */ - skb = skb_dequeue(head); - while (skb) { - struct scm_fp_list *fp; - - fp = UNIXCB(skb).fp; - for (i = 0; i < fp->count; i++) { - int left; - - if (fp->fp[i] != file) - continue; - - unix_notinflight(fp->user, fp->fp[i]); - left = fp->count - 1 - i; - if (left) { - memmove(&fp->fp[i], &fp->fp[i + 1], - left * sizeof(struct file *)); - } - fp->count--; - if (!fp->count) { - kfree_skb(skb); - skb = NULL; - } else { - __skb_queue_tail(&list, skb); - } - fput(file); - file = NULL; - break; - } - - if (!file) - break; - - __skb_queue_tail(&list, skb); - - skb = skb_dequeue(head); - } - - if (skb_peek(&list)) { - spin_lock_irq(&head->lock); - while ((skb = __skb_dequeue(&list)) != NULL) - __skb_queue_tail(head, skb); - spin_unlock_irq(&head->lock); - } -#endif -} - -static void io_rsrc_file_put(struct io_ring_ctx *ctx, struct io_rsrc_put *prsrc) -{ - struct file *file = prsrc->file; - - if (likely(!io_file_need_scm(file))) - fput(file); - else - io_rsrc_file_scm_put(ctx, file); -} - int io_sqe_files_register(struct io_ring_ctx *ctx, void __user *arg, unsigned nr_args, u64 __user *tags) { @@ -897,21 +745,12 @@ int io_sqe_files_register(struct io_ring_ctx *ctx, void __user *arg, goto fail; /* - * Don't allow io_uring instances to be registered. If UNIX - * isn't enabled, then this causes a reference cycle and this - * instance can never get freed. If UNIX is enabled we'll - * handle it just fine, but there's still no point in allowing - * a ring fd as it doesn't support regular read/write anyway. + * Don't allow io_uring instances to be registered. */ if (io_is_uring_fops(file)) { fput(file); goto fail; } - ret = io_scm_file_account(ctx, file); - if (ret) { - fput(file); - goto fail; - } file_slot = io_fixed_file_slot(&ctx->file_table, i); io_fixed_file_set(file_slot, file); io_file_bitmap_set(&ctx->file_table, i); diff --git a/io_uring/rsrc.h b/io_uring/rsrc.h index 08ac0d8e07ef..7238b9cfe33b 100644 --- a/io_uring/rsrc.h +++ b/io_uring/rsrc.h @@ -75,21 +75,6 @@ int io_sqe_files_unregister(struct io_ring_ctx *ctx); int io_sqe_files_register(struct io_ring_ctx *ctx, void __user *arg, unsigned nr_args, u64 __user *tags); -int __io_scm_file_account(struct io_ring_ctx *ctx, struct file *file); - -static inline bool io_file_need_scm(struct file *filp) -{ - return false; -} - -static inline int io_scm_file_account(struct io_ring_ctx *ctx, - struct file *file) -{ - if (likely(!io_file_need_scm(file))) - return 0; - return __io_scm_file_account(ctx, file); -} - int io_register_files_update(struct io_ring_ctx *ctx, void __user *arg, unsigned nr_args); int io_register_rsrc_update(struct io_ring_ctx *ctx, void __user *arg, -- 2.43.0
From cbe948d42e1ff43e77d60f21823d3d798f92835e Mon Sep 17 00:00:00 2001 From: Jens Axboe <axboe@xxxxxxxxx> Date: Tue, 19 Dec 2023 12:36:34 -0700 Subject: [PATCH 2/2] io_uring: drop any code related to SCM_RIGHTS This is dead code after we dropped support for passing io_uring fds over SCM_RIGHTS, get rid of it. Signed-off-by: Jens Axboe <axboe@xxxxxxxxx> --- include/linux/io_uring_types.h | 3 - io_uring/filetable.c | 11 +-- io_uring/io_uring.c | 32 +------ io_uring/rsrc.c | 169 +-------------------------------- io_uring/rsrc.h | 15 --- 5 files changed, 10 insertions(+), 220 deletions(-) diff --git a/include/linux/io_uring_types.h b/include/linux/io_uring_types.h index 5fd664fb71c8..731beb3198c4 100644 --- a/include/linux/io_uring_types.h +++ b/include/linux/io_uring_types.h @@ -347,9 +347,6 @@ struct io_ring_ctx { struct list_head io_buffers_pages; - #if defined(CONFIG_UNIX) - struct socket *ring_sock; - #endif /* hashed buffered write serialization */ struct io_wq_hash *hash_map; diff --git a/io_uring/filetable.c b/io_uring/filetable.c index e7d749991de4..6e86e6188dbe 100644 --- a/io_uring/filetable.c +++ b/io_uring/filetable.c @@ -87,13 +87,10 @@ static int io_install_fixed_file(struct io_ring_ctx *ctx, struct file *file, io_file_bitmap_clear(&ctx->file_table, slot_index); } - ret = io_scm_file_account(ctx, file); - if (!ret) { - *io_get_tag_slot(ctx->file_data, slot_index) = 0; - io_fixed_file_set(file_slot, file); - io_file_bitmap_set(&ctx->file_table, slot_index); - } - return ret; + *io_get_tag_slot(ctx->file_data, slot_index) = 0; + io_fixed_file_set(file_slot, file); + io_file_bitmap_set(&ctx->file_table, slot_index); + return 0; } int __io_fixed_fd_install(struct io_ring_ctx *ctx, struct file *file, diff --git a/io_uring/io_uring.c b/io_uring/io_uring.c index 747d2a6ecab8..2744d72f1085 100644 --- a/io_uring/io_uring.c +++ b/io_uring/io_uring.c @@ -60,7 +60,6 @@ #include <linux/net.h> #include <net/sock.h> #include <net/af_unix.h> -#include <net/scm.h> #include <linux/anon_inodes.h> #include <linux/sched/mm.h> #include <linux/uaccess.h> @@ -2931,13 +2930,6 @@ static __cold void io_ring_ctx_free(struct io_ring_ctx *ctx) io_rsrc_node_destroy(ctx, ctx->rsrc_node); WARN_ON_ONCE(!list_empty(&ctx->rsrc_ref_list)); - -#if defined(CONFIG_UNIX) - if (ctx->ring_sock) { - ctx->ring_sock->file = NULL; /* so that iput() is called */ - sock_release(ctx->ring_sock); - } -#endif WARN_ON_ONCE(!list_empty(&ctx->ltimeout_list)); io_alloc_cache_free(&ctx->rsrc_node_cache, io_rsrc_node_cache_free); @@ -3825,32 +3817,12 @@ static int io_uring_install_fd(struct file *file) /* * Allocate an anonymous fd, this is what constitutes the application * visible backing of an io_uring instance. The application mmaps this - * fd to gain access to the SQ/CQ ring details. If UNIX sockets are enabled, - * we have to tie this fd to a socket for file garbage collection purposes. + * fd to gain access to the SQ/CQ ring details. */ static struct file *io_uring_get_file(struct io_ring_ctx *ctx) { - struct file *file; -#if defined(CONFIG_UNIX) - int ret; - - ret = sock_create_kern(&init_net, PF_UNIX, SOCK_RAW, IPPROTO_IP, - &ctx->ring_sock); - if (ret) - return ERR_PTR(ret); -#endif - - file = anon_inode_getfile_secure("[io_uring]", &io_uring_fops, ctx, + return anon_inode_getfile_secure("[io_uring]", &io_uring_fops, ctx, O_RDWR | O_CLOEXEC, NULL); -#if defined(CONFIG_UNIX) - if (IS_ERR(file)) { - sock_release(ctx->ring_sock); - ctx->ring_sock = NULL; - } else { - ctx->ring_sock->file = file; - } -#endif - return file; } static __cold int io_uring_create(unsigned entries, struct io_uring_params *p, diff --git a/io_uring/rsrc.c b/io_uring/rsrc.c index dde501abd719..95230921b16d 100644 --- a/io_uring/rsrc.c +++ b/io_uring/rsrc.c @@ -24,7 +24,6 @@ struct io_rsrc_update { }; static void io_rsrc_buf_put(struct io_ring_ctx *ctx, struct io_rsrc_put *prsrc); -static void io_rsrc_file_put(struct io_ring_ctx *ctx, struct io_rsrc_put *prsrc); static int io_sqe_buffer_register(struct io_ring_ctx *ctx, struct iovec *iov, struct io_mapped_ubuf **pimu, struct page **last_hpage); @@ -157,7 +156,7 @@ static void io_rsrc_put_work(struct io_rsrc_node *node) switch (node->type) { case IORING_RSRC_FILE: - io_rsrc_file_put(node->ctx, prsrc); + fput(prsrc->file); break; case IORING_RSRC_BUFFER: io_rsrc_buf_put(node->ctx, prsrc); @@ -402,23 +401,13 @@ static int __io_sqe_files_update(struct io_ring_ctx *ctx, break; } /* - * Don't allow io_uring instances to be registered. If - * UNIX isn't enabled, then this causes a reference - * cycle and this instance can never get freed. If UNIX - * is enabled we'll handle it just fine, but there's - * still no point in allowing a ring fd as it doesn't - * support regular read/write anyway. + * Don't allow io_uring instances to be registered. */ if (io_is_uring_fops(file)) { fput(file); err = -EBADF; break; } - err = io_scm_file_account(ctx, file); - if (err) { - fput(file); - break; - } *io_get_tag_slot(data, i) = tag; io_fixed_file_set(file_slot, file); io_file_bitmap_set(&ctx->file_table, i); @@ -675,22 +664,12 @@ void __io_sqe_files_unregister(struct io_ring_ctx *ctx) for (i = 0; i < ctx->nr_user_files; i++) { struct file *file = io_file_from_index(&ctx->file_table, i); - /* skip scm accounted files, they'll be freed by ->ring_sock */ - if (!file || io_file_need_scm(file)) + if (!file) continue; io_file_bitmap_clear(&ctx->file_table, i); fput(file); } -#if defined(CONFIG_UNIX) - if (ctx->ring_sock) { - struct sock *sock = ctx->ring_sock->sk; - struct sk_buff *skb; - - while ((skb = skb_dequeue(&sock->sk_receive_queue)) != NULL) - kfree_skb(skb); - } -#endif io_free_file_tables(&ctx->file_table); io_file_table_set_alloc_range(ctx, 0, 0); io_rsrc_data_free(ctx->file_data); @@ -718,137 +697,6 @@ int io_sqe_files_unregister(struct io_ring_ctx *ctx) return ret; } -/* - * Ensure the UNIX gc is aware of our file set, so we are certain that - * the io_uring can be safely unregistered on process exit, even if we have - * loops in the file referencing. We account only files that can hold other - * files because otherwise they can't form a loop and so are not interesting - * for GC. - */ -int __io_scm_file_account(struct io_ring_ctx *ctx, struct file *file) -{ -#if defined(CONFIG_UNIX) - struct sock *sk = ctx->ring_sock->sk; - struct sk_buff_head *head = &sk->sk_receive_queue; - struct scm_fp_list *fpl; - struct sk_buff *skb; - - if (likely(!io_file_need_scm(file))) - return 0; - - /* - * See if we can merge this file into an existing skb SCM_RIGHTS - * file set. If there's no room, fall back to allocating a new skb - * and filling it in. - */ - spin_lock_irq(&head->lock); - skb = skb_peek(head); - if (skb && UNIXCB(skb).fp->count < SCM_MAX_FD) - __skb_unlink(skb, head); - else - skb = NULL; - spin_unlock_irq(&head->lock); - - if (!skb) { - fpl = kzalloc(sizeof(*fpl), GFP_KERNEL); - if (!fpl) - return -ENOMEM; - - skb = alloc_skb(0, GFP_KERNEL); - if (!skb) { - kfree(fpl); - return -ENOMEM; - } - - fpl->user = get_uid(current_user()); - fpl->max = SCM_MAX_FD; - fpl->count = 0; - - UNIXCB(skb).fp = fpl; - skb->sk = sk; - skb->destructor = io_uring_destruct_scm; - refcount_add(skb->truesize, &sk->sk_wmem_alloc); - } - - fpl = UNIXCB(skb).fp; - fpl->fp[fpl->count++] = get_file(file); - unix_inflight(fpl->user, file); - skb_queue_head(head, skb); - fput(file); -#endif - return 0; -} - -static __cold void io_rsrc_file_scm_put(struct io_ring_ctx *ctx, struct file *file) -{ -#if defined(CONFIG_UNIX) - struct sock *sock = ctx->ring_sock->sk; - struct sk_buff_head list, *head = &sock->sk_receive_queue; - struct sk_buff *skb; - int i; - - __skb_queue_head_init(&list); - - /* - * Find the skb that holds this file in its SCM_RIGHTS. When found, - * remove this entry and rearrange the file array. - */ - skb = skb_dequeue(head); - while (skb) { - struct scm_fp_list *fp; - - fp = UNIXCB(skb).fp; - for (i = 0; i < fp->count; i++) { - int left; - - if (fp->fp[i] != file) - continue; - - unix_notinflight(fp->user, fp->fp[i]); - left = fp->count - 1 - i; - if (left) { - memmove(&fp->fp[i], &fp->fp[i + 1], - left * sizeof(struct file *)); - } - fp->count--; - if (!fp->count) { - kfree_skb(skb); - skb = NULL; - } else { - __skb_queue_tail(&list, skb); - } - fput(file); - file = NULL; - break; - } - - if (!file) - break; - - __skb_queue_tail(&list, skb); - - skb = skb_dequeue(head); - } - - if (skb_peek(&list)) { - spin_lock_irq(&head->lock); - while ((skb = __skb_dequeue(&list)) != NULL) - __skb_queue_tail(head, skb); - spin_unlock_irq(&head->lock); - } -#endif -} - -static void io_rsrc_file_put(struct io_ring_ctx *ctx, struct io_rsrc_put *prsrc) -{ - struct file *file = prsrc->file; - - if (likely(!io_file_need_scm(file))) - fput(file); - else - io_rsrc_file_scm_put(ctx, file); -} - int io_sqe_files_register(struct io_ring_ctx *ctx, void __user *arg, unsigned nr_args, u64 __user *tags) { @@ -897,21 +745,12 @@ int io_sqe_files_register(struct io_ring_ctx *ctx, void __user *arg, goto fail; /* - * Don't allow io_uring instances to be registered. If UNIX - * isn't enabled, then this causes a reference cycle and this - * instance can never get freed. If UNIX is enabled we'll - * handle it just fine, but there's still no point in allowing - * a ring fd as it doesn't support regular read/write anyway. + * Don't allow io_uring instances to be registered. */ if (io_is_uring_fops(file)) { fput(file); goto fail; } - ret = io_scm_file_account(ctx, file); - if (ret) { - fput(file); - goto fail; - } file_slot = io_fixed_file_slot(&ctx->file_table, i); io_fixed_file_set(file_slot, file); io_file_bitmap_set(&ctx->file_table, i); diff --git a/io_uring/rsrc.h b/io_uring/rsrc.h index 08ac0d8e07ef..7238b9cfe33b 100644 --- a/io_uring/rsrc.h +++ b/io_uring/rsrc.h @@ -75,21 +75,6 @@ int io_sqe_files_unregister(struct io_ring_ctx *ctx); int io_sqe_files_register(struct io_ring_ctx *ctx, void __user *arg, unsigned nr_args, u64 __user *tags); -int __io_scm_file_account(struct io_ring_ctx *ctx, struct file *file); - -static inline bool io_file_need_scm(struct file *filp) -{ - return false; -} - -static inline int io_scm_file_account(struct io_ring_ctx *ctx, - struct file *file) -{ - if (likely(!io_file_need_scm(file))) - return 0; - return __io_scm_file_account(ctx, file); -} - int io_register_files_update(struct io_ring_ctx *ctx, void __user *arg, unsigned nr_args); int io_register_rsrc_update(struct io_ring_ctx *ctx, void __user *arg, -- 2.43.0
From b706187d02eacef15922de5fafedd65bc2db0576 Mon Sep 17 00:00:00 2001 From: Jens Axboe <axboe@xxxxxxxxx> Date: Tue, 19 Dec 2023 12:30:43 -0700 Subject: [PATCH 1/2] io_uring/unix: drop usage of io_uring socket Commit a4104821ad651d8a0b374f0b2474c345bbb42f82 upstream. Since we no longer allow sending io_uring fds over SCM_RIGHTS, move to using io_is_uring_fops() to detect whether this is a io_uring fd or not. With that done, kill off io_uring_get_socket() as nobody calls it anymore. This is in preparation to yanking out the rest of the core related to unix gc with io_uring. Signed-off-by: Jens Axboe <axboe@xxxxxxxxx> --- include/linux/io_uring.h | 10 +++++----- io_uring/io_uring.c | 13 ------------- io_uring/io_uring.h | 1 - net/core/scm.c | 2 +- net/unix/scm.c | 4 +--- 5 files changed, 7 insertions(+), 23 deletions(-) diff --git a/include/linux/io_uring.h b/include/linux/io_uring.h index 106cdc55ff3b..f99ff6de926c 100644 --- a/include/linux/io_uring.h +++ b/include/linux/io_uring.h @@ -46,7 +46,6 @@ int io_uring_cmd_import_fixed(u64 ubuf, unsigned long len, int rw, struct iov_iter *iter, void *ioucmd); void io_uring_cmd_done(struct io_uring_cmd *cmd, ssize_t ret, ssize_t res2, unsigned issue_flags); -struct sock *io_uring_get_socket(struct file *file); void __io_uring_cancel(bool cancel_all); void __io_uring_free(struct task_struct *tsk); void io_uring_unreg_ringfd(void); @@ -82,6 +81,7 @@ static inline void io_uring_free(struct task_struct *tsk) __io_uring_free(tsk); } int io_uring_cmd_sock(struct io_uring_cmd *cmd, unsigned int issue_flags); +bool io_is_uring_fops(struct file *file); #else static inline int io_uring_cmd_import_fixed(u64 ubuf, unsigned long len, int rw, struct iov_iter *iter, void *ioucmd) @@ -100,10 +100,6 @@ static inline void io_uring_cmd_do_in_task_lazy(struct io_uring_cmd *ioucmd, void (*task_work_cb)(struct io_uring_cmd *, unsigned)) { } -static inline struct sock *io_uring_get_socket(struct file *file) -{ - return NULL; -} static inline void io_uring_task_cancel(void) { } @@ -122,6 +118,10 @@ static inline int io_uring_cmd_sock(struct io_uring_cmd *cmd, { return -EOPNOTSUPP; } +static inline bool io_is_uring_fops(struct file *file) +{ + return false; +} #endif #endif diff --git a/io_uring/io_uring.c b/io_uring/io_uring.c index ea772a02c140..747d2a6ecab8 100644 --- a/io_uring/io_uring.c +++ b/io_uring/io_uring.c @@ -175,19 +175,6 @@ static struct ctl_table kernel_io_uring_disabled_table[] = { }; #endif -struct sock *io_uring_get_socket(struct file *file) -{ -#if defined(CONFIG_UNIX) - if (io_is_uring_fops(file)) { - struct io_ring_ctx *ctx = file->private_data; - - return ctx->ring_sock->sk; - } -#endif - return NULL; -} -EXPORT_SYMBOL(io_uring_get_socket); - static inline void io_submit_flush_completions(struct io_ring_ctx *ctx) { if (!wq_list_empty(&ctx->submit_state.compl_reqs) || diff --git a/io_uring/io_uring.h b/io_uring/io_uring.h index c8cba7831083..411c883b37a9 100644 --- a/io_uring/io_uring.h +++ b/io_uring/io_uring.h @@ -61,7 +61,6 @@ struct file *io_file_get_fixed(struct io_kiocb *req, int fd, unsigned issue_flags); void __io_req_task_work_add(struct io_kiocb *req, unsigned flags); -bool io_is_uring_fops(struct file *file); bool io_alloc_async_data(struct io_kiocb *req); void io_req_task_queue(struct io_kiocb *req); void io_queue_iowq(struct io_kiocb *req, struct io_tw_state *ts_dont_use); diff --git a/net/core/scm.c b/net/core/scm.c index 7dc47c17d863..737917c7ac62 100644 --- a/net/core/scm.c +++ b/net/core/scm.c @@ -105,7 +105,7 @@ static int scm_fp_copy(struct cmsghdr *cmsg, struct scm_fp_list **fplp) if (fd < 0 || !(file = fget_raw(fd))) return -EBADF; /* don't allow io_uring files */ - if (io_uring_get_socket(file)) { + if (io_is_uring_fops(file)) { fput(file); return -EINVAL; } diff --git a/net/unix/scm.c b/net/unix/scm.c index 6ff628f2349f..822ce0d0d791 100644 --- a/net/unix/scm.c +++ b/net/unix/scm.c @@ -35,10 +35,8 @@ struct sock *unix_get_socket(struct file *filp) /* PF_UNIX ? */ if (s && ops && ops->family == PF_UNIX) u_sock = s; - } else { - /* Could be an io_uring instance */ - u_sock = io_uring_get_socket(filp); } + return u_sock; } EXPORT_SYMBOL(unix_get_socket); -- 2.43.0
From 42d0449f65e6632ddc638929c53676fb776d8412 Mon Sep 17 00:00:00 2001 From: Jens Axboe <axboe@xxxxxxxxx> Date: Wed, 13 Mar 2024 18:10:12 -0600 Subject: [PATCH 1/2] io_uring/unix: drop usage of io_uring socket Commit a4104821ad651d8a0b374f0b2474c345bbb42f82 upstream. Since we no longer allow sending io_uring fds over SCM_RIGHTS, move to using io_is_uring_fops() to detect whether this is a io_uring fd or not. With that done, kill off io_uring_get_socket() as nobody calls it anymore. This is in preparation to yanking out the rest of the core related to unix gc with io_uring. Signed-off-by: Jens Axboe <axboe@xxxxxxxxx> --- include/linux/io_uring.h | 10 +++++----- io_uring/io_uring.c | 13 ------------- io_uring/io_uring.h | 1 - net/core/scm.c | 2 +- net/unix/scm.c | 4 +--- 5 files changed, 7 insertions(+), 23 deletions(-) diff --git a/include/linux/io_uring.h b/include/linux/io_uring.h index a1484cdb3158..a8f3058448ea 100644 --- a/include/linux/io_uring.h +++ b/include/linux/io_uring.h @@ -42,11 +42,11 @@ void io_uring_cmd_done(struct io_uring_cmd *cmd, ssize_t ret, ssize_t res2, unsigned issue_flags); void io_uring_cmd_complete_in_task(struct io_uring_cmd *ioucmd, void (*task_work_cb)(struct io_uring_cmd *, unsigned)); -struct sock *io_uring_get_socket(struct file *file); void __io_uring_cancel(bool cancel_all); void __io_uring_free(struct task_struct *tsk); void io_uring_unreg_ringfd(void); const char *io_uring_get_opcode(u8 opcode); +bool io_is_uring_fops(struct file *file); static inline void io_uring_files_cancel(void) { @@ -71,6 +71,10 @@ static inline int io_uring_cmd_import_fixed(u64 ubuf, unsigned long len, int rw, { return -EOPNOTSUPP; } +static inline bool io_is_uring_fops(struct file *file) +{ + return false; +} static inline void io_uring_cmd_done(struct io_uring_cmd *cmd, ssize_t ret, ssize_t ret2, unsigned issue_flags) { @@ -79,10 +83,6 @@ static inline void io_uring_cmd_complete_in_task(struct io_uring_cmd *ioucmd, void (*task_work_cb)(struct io_uring_cmd *, unsigned)) { } -static inline struct sock *io_uring_get_socket(struct file *file) -{ - return NULL; -} static inline void io_uring_task_cancel(void) { } diff --git a/io_uring/io_uring.c b/io_uring/io_uring.c index 35894955b454..cf7dd62da0e3 100644 --- a/io_uring/io_uring.c +++ b/io_uring/io_uring.c @@ -153,19 +153,6 @@ static void __io_submit_flush_completions(struct io_ring_ctx *ctx); static struct kmem_cache *req_cachep; -struct sock *io_uring_get_socket(struct file *file) -{ -#if defined(CONFIG_UNIX) - if (io_is_uring_fops(file)) { - struct io_ring_ctx *ctx = file->private_data; - - return ctx->ring_sock->sk; - } -#endif - return NULL; -} -EXPORT_SYMBOL(io_uring_get_socket); - static inline void io_submit_flush_completions(struct io_ring_ctx *ctx) { if (!wq_list_empty(&ctx->submit_state.compl_reqs)) diff --git a/io_uring/io_uring.h b/io_uring/io_uring.h index 019600570ee4..59e6f755f12c 100644 --- a/io_uring/io_uring.h +++ b/io_uring/io_uring.h @@ -52,7 +52,6 @@ static inline bool io_req_ffs_set(struct io_kiocb *req) } void __io_req_task_work_add(struct io_kiocb *req, bool allow_local); -bool io_is_uring_fops(struct file *file); bool io_alloc_async_data(struct io_kiocb *req); void io_req_task_queue(struct io_kiocb *req); void io_queue_iowq(struct io_kiocb *req, bool *dont_use); diff --git a/net/core/scm.c b/net/core/scm.c index e762a4b8a1d2..a877c4ef4c25 100644 --- a/net/core/scm.c +++ b/net/core/scm.c @@ -105,7 +105,7 @@ static int scm_fp_copy(struct cmsghdr *cmsg, struct scm_fp_list **fplp) if (fd < 0 || !(file = fget_raw(fd))) return -EBADF; /* don't allow io_uring files */ - if (io_uring_get_socket(file)) { + if (io_is_uring_fops(file)) { fput(file); return -EINVAL; } diff --git a/net/unix/scm.c b/net/unix/scm.c index e8e2a00bb0f5..d1048b4c2baa 100644 --- a/net/unix/scm.c +++ b/net/unix/scm.c @@ -34,10 +34,8 @@ struct sock *unix_get_socket(struct file *filp) /* PF_UNIX ? */ if (s && sock->ops && sock->ops->family == PF_UNIX) u_sock = s; - } else { - /* Could be an io_uring instance */ - u_sock = io_uring_get_socket(filp); } + return u_sock; } EXPORT_SYMBOL(unix_get_socket); -- 2.43.0
From 083974002b6f45bb84090a9aed025d8a3d8479cb Mon Sep 17 00:00:00 2001 From: Jens Axboe <axboe@xxxxxxxxx> Date: Wed, 13 Mar 2024 18:15:05 -0600 Subject: [PATCH 2/2] io_uring: drop any code related to SCM_RIGHTS Commit 6e5e6d274956305f1fc0340522b38f5f5be74bdb upstream. This is dead code after we dropped support for passing io_uring fds over SCM_RIGHTS, get rid of it. Signed-off-by: Jens Axboe <axboe@xxxxxxxxx> --- include/linux/io_uring_types.h | 3 - io_uring/filetable.c | 10 +-- io_uring/io_uring.c | 31 +------ io_uring/rsrc.c | 151 +-------------------------------- io_uring/rsrc.h | 15 ---- 5 files changed, 8 insertions(+), 202 deletions(-) diff --git a/include/linux/io_uring_types.h b/include/linux/io_uring_types.h index f5b687a787a3..37aeea266ebb 100644 --- a/include/linux/io_uring_types.h +++ b/include/linux/io_uring_types.h @@ -330,9 +330,6 @@ struct io_ring_ctx { struct list_head io_buffers_pages; - #if defined(CONFIG_UNIX) - struct socket *ring_sock; - #endif /* hashed buffered write serialization */ struct io_wq_hash *hash_map; diff --git a/io_uring/filetable.c b/io_uring/filetable.c index b80614e7d605..4660cb89ea9f 100644 --- a/io_uring/filetable.c +++ b/io_uring/filetable.c @@ -95,12 +95,10 @@ static int io_install_fixed_file(struct io_ring_ctx *ctx, struct file *file, needs_switch = true; } - ret = io_scm_file_account(ctx, file); - if (!ret) { - *io_get_tag_slot(ctx->file_data, slot_index) = 0; - io_fixed_file_set(file_slot, file); - io_file_bitmap_set(&ctx->file_table, slot_index); - } + *io_get_tag_slot(ctx->file_data, slot_index) = 0; + io_fixed_file_set(file_slot, file); + io_file_bitmap_set(&ctx->file_table, slot_index); + return 0; err: if (needs_switch) io_rsrc_node_switch(ctx, ctx->file_data); diff --git a/io_uring/io_uring.c b/io_uring/io_uring.c index cf7dd62da0e3..415248c1f82c 100644 --- a/io_uring/io_uring.c +++ b/io_uring/io_uring.c @@ -60,7 +60,6 @@ #include <linux/net.h> #include <net/sock.h> #include <net/af_unix.h> -#include <net/scm.h> #include <linux/anon_inodes.h> #include <linux/sched/mm.h> #include <linux/uaccess.h> @@ -2628,12 +2627,6 @@ static __cold void io_ring_ctx_free(struct io_ring_ctx *ctx) WARN_ON_ONCE(!list_empty(&ctx->rsrc_ref_list)); WARN_ON_ONCE(!llist_empty(&ctx->rsrc_put_llist)); -#if defined(CONFIG_UNIX) - if (ctx->ring_sock) { - ctx->ring_sock->file = NULL; /* so that iput() is called */ - sock_release(ctx->ring_sock); - } -#endif WARN_ON_ONCE(!list_empty(&ctx->ltimeout_list)); if (ctx->mm_account) { @@ -3438,32 +3431,12 @@ static int io_uring_install_fd(struct io_ring_ctx *ctx, struct file *file) /* * Allocate an anonymous fd, this is what constitutes the application * visible backing of an io_uring instance. The application mmaps this - * fd to gain access to the SQ/CQ ring details. If UNIX sockets are enabled, - * we have to tie this fd to a socket for file garbage collection purposes. + * fd to gain access to the SQ/CQ ring details. */ static struct file *io_uring_get_file(struct io_ring_ctx *ctx) { - struct file *file; -#if defined(CONFIG_UNIX) - int ret; - - ret = sock_create_kern(&init_net, PF_UNIX, SOCK_RAW, IPPROTO_IP, - &ctx->ring_sock); - if (ret) - return ERR_PTR(ret); -#endif - - file = anon_inode_getfile_secure("[io_uring]", &io_uring_fops, ctx, + return anon_inode_getfile_secure("[io_uring]", &io_uring_fops, ctx, O_RDWR | O_CLOEXEC, NULL); -#if defined(CONFIG_UNIX) - if (IS_ERR(file)) { - sock_release(ctx->ring_sock); - ctx->ring_sock = NULL; - } else { - ctx->ring_sock->file = file; - } -#endif - return file; } static __cold int io_uring_create(unsigned entries, struct io_uring_params *p, diff --git a/io_uring/rsrc.c b/io_uring/rsrc.c index 7ada0339b387..ac658cfa89c6 100644 --- a/io_uring/rsrc.c +++ b/io_uring/rsrc.c @@ -494,11 +494,6 @@ static int __io_sqe_files_update(struct io_ring_ctx *ctx, err = -EBADF; break; } - err = io_scm_file_account(ctx, file); - if (err) { - fput(file); - break; - } *io_get_tag_slot(data, i) = tag; io_fixed_file_set(file_slot, file); io_file_bitmap_set(&ctx->file_table, i); @@ -762,22 +757,12 @@ void __io_sqe_files_unregister(struct io_ring_ctx *ctx) for (i = 0; i < ctx->nr_user_files; i++) { struct file *file = io_file_from_index(&ctx->file_table, i); - /* skip scm accounted files, they'll be freed by ->ring_sock */ - if (!file || io_file_need_scm(file)) + if (!file) continue; io_file_bitmap_clear(&ctx->file_table, i); fput(file); } -#if defined(CONFIG_UNIX) - if (ctx->ring_sock) { - struct sock *sock = ctx->ring_sock->sk; - struct sk_buff *skb; - - while ((skb = skb_dequeue(&sock->sk_receive_queue)) != NULL) - kfree_skb(skb); - } -#endif io_free_file_tables(&ctx->file_table); io_file_table_set_alloc_range(ctx, 0, 0); io_rsrc_data_free(ctx->file_data); @@ -805,134 +790,11 @@ int io_sqe_files_unregister(struct io_ring_ctx *ctx) return ret; } -/* - * Ensure the UNIX gc is aware of our file set, so we are certain that - * the io_uring can be safely unregistered on process exit, even if we have - * loops in the file referencing. We account only files that can hold other - * files because otherwise they can't form a loop and so are not interesting - * for GC. - */ -int __io_scm_file_account(struct io_ring_ctx *ctx, struct file *file) -{ -#if defined(CONFIG_UNIX) - struct sock *sk = ctx->ring_sock->sk; - struct sk_buff_head *head = &sk->sk_receive_queue; - struct scm_fp_list *fpl; - struct sk_buff *skb; - - if (likely(!io_file_need_scm(file))) - return 0; - - /* - * See if we can merge this file into an existing skb SCM_RIGHTS - * file set. If there's no room, fall back to allocating a new skb - * and filling it in. - */ - spin_lock_irq(&head->lock); - skb = skb_peek(head); - if (skb && UNIXCB(skb).fp->count < SCM_MAX_FD) - __skb_unlink(skb, head); - else - skb = NULL; - spin_unlock_irq(&head->lock); - - if (!skb) { - fpl = kzalloc(sizeof(*fpl), GFP_KERNEL); - if (!fpl) - return -ENOMEM; - - skb = alloc_skb(0, GFP_KERNEL); - if (!skb) { - kfree(fpl); - return -ENOMEM; - } - - fpl->user = get_uid(current_user()); - fpl->max = SCM_MAX_FD; - fpl->count = 0; - - UNIXCB(skb).fp = fpl; - skb->sk = sk; - skb->scm_io_uring = 1; - skb->destructor = unix_destruct_scm; - refcount_add(skb->truesize, &sk->sk_wmem_alloc); - } - - fpl = UNIXCB(skb).fp; - fpl->fp[fpl->count++] = get_file(file); - unix_inflight(fpl->user, file); - skb_queue_head(head, skb); - fput(file); -#endif - return 0; -} - static void io_rsrc_file_put(struct io_ring_ctx *ctx, struct io_rsrc_put *prsrc) { struct file *file = prsrc->file; -#if defined(CONFIG_UNIX) - struct sock *sock = ctx->ring_sock->sk; - struct sk_buff_head list, *head = &sock->sk_receive_queue; - struct sk_buff *skb; - int i; - - if (!io_file_need_scm(file)) { - fput(file); - return; - } - - __skb_queue_head_init(&list); - - /* - * Find the skb that holds this file in its SCM_RIGHTS. When found, - * remove this entry and rearrange the file array. - */ - skb = skb_dequeue(head); - while (skb) { - struct scm_fp_list *fp; - fp = UNIXCB(skb).fp; - for (i = 0; i < fp->count; i++) { - int left; - - if (fp->fp[i] != file) - continue; - - unix_notinflight(fp->user, fp->fp[i]); - left = fp->count - 1 - i; - if (left) { - memmove(&fp->fp[i], &fp->fp[i + 1], - left * sizeof(struct file *)); - } - fp->count--; - if (!fp->count) { - kfree_skb(skb); - skb = NULL; - } else { - __skb_queue_tail(&list, skb); - } - fput(file); - file = NULL; - break; - } - - if (!file) - break; - - __skb_queue_tail(&list, skb); - - skb = skb_dequeue(head); - } - - if (skb_peek(&list)) { - spin_lock_irq(&head->lock); - while ((skb = __skb_dequeue(&list)) != NULL) - __skb_queue_tail(head, skb); - spin_unlock_irq(&head->lock); - } -#else fput(file); -#endif } int io_sqe_files_register(struct io_ring_ctx *ctx, void __user *arg, @@ -986,21 +848,12 @@ int io_sqe_files_register(struct io_ring_ctx *ctx, void __user *arg, goto fail; /* - * Don't allow io_uring instances to be registered. If UNIX - * isn't enabled, then this causes a reference cycle and this - * instance can never get freed. If UNIX is enabled we'll - * handle it just fine, but there's still no point in allowing - * a ring fd as it doesn't support regular read/write anyway. + * Don't allow io_uring instances to be registered. */ if (io_is_uring_fops(file)) { fput(file); goto fail; } - ret = io_scm_file_account(ctx, file); - if (ret) { - fput(file); - goto fail; - } file_slot = io_fixed_file_slot(&ctx->file_table, i); io_fixed_file_set(file_slot, file); io_file_bitmap_set(&ctx->file_table, i); diff --git a/io_uring/rsrc.h b/io_uring/rsrc.h index acaf8dad0540..85f145607c62 100644 --- a/io_uring/rsrc.h +++ b/io_uring/rsrc.h @@ -77,21 +77,6 @@ int io_sqe_files_unregister(struct io_ring_ctx *ctx); int io_sqe_files_register(struct io_ring_ctx *ctx, void __user *arg, unsigned nr_args, u64 __user *tags); -int __io_scm_file_account(struct io_ring_ctx *ctx, struct file *file); - -static inline bool io_file_need_scm(struct file *filp) -{ - return false; -} - -static inline int io_scm_file_account(struct io_ring_ctx *ctx, - struct file *file) -{ - if (likely(!io_file_need_scm(file))) - return 0; - return __io_scm_file_account(ctx, file); -} - int io_register_files_update(struct io_ring_ctx *ctx, void __user *arg, unsigned nr_args); int io_register_rsrc_update(struct io_ring_ctx *ctx, void __user *arg, -- 2.43.0
From 63fcaffd0b3d0db784cbe3950549766936355e84 Mon Sep 17 00:00:00 2001 From: Jens Axboe <axboe@xxxxxxxxx> Date: Wed, 13 Mar 2024 17:59:01 -0600 Subject: [PATCH 2/2] io_uring: drop any code related to SCM_RIGHTS Commit 6e5e6d274956305f1fc0340522b38f5f5be74bdb upstream. This is dead code after we dropped support for passing io_uring fds over SCM_RIGHTS, get rid of it. Signed-off-by: Jens Axboe <axboe@xxxxxxxxx> --- io_uring/io_uring.c | 211 +------------------------------------------- 1 file changed, 3 insertions(+), 208 deletions(-) diff --git a/io_uring/io_uring.c b/io_uring/io_uring.c index 541e6a6985ec..386b32e6fd9a 100644 --- a/io_uring/io_uring.c +++ b/io_uring/io_uring.c @@ -62,7 +62,6 @@ #include <linux/net.h> #include <net/sock.h> #include <net/af_unix.h> -#include <net/scm.h> #include <linux/anon_inodes.h> #include <linux/sched/mm.h> #include <linux/uaccess.h> @@ -440,9 +439,6 @@ struct io_ring_ctx { /* Keep this last, we don't need it for the fast path */ struct { - #if defined(CONFIG_UNIX) - struct socket *ring_sock; - #endif /* hashed buffered write serialization */ struct io_wq_hash *hash_map; @@ -8151,15 +8147,6 @@ static void io_free_file_tables(struct io_file_table *table) static void __io_sqe_files_unregister(struct io_ring_ctx *ctx) { -#if defined(CONFIG_UNIX) - if (ctx->ring_sock) { - struct sock *sock = ctx->ring_sock->sk; - struct sk_buff *skb; - - while ((skb = skb_dequeue(&sock->sk_receive_queue)) != NULL) - kfree_skb(skb); - } -#else int i; for (i = 0; i < ctx->nr_user_files; i++) { @@ -8169,7 +8156,6 @@ static void __io_sqe_files_unregister(struct io_ring_ctx *ctx) if (file) fput(file); } -#endif io_free_file_tables(&ctx->file_table); io_rsrc_data_free(ctx->file_data); ctx->file_data = NULL; @@ -8321,170 +8307,11 @@ static struct io_sq_data *io_get_sq_data(struct io_uring_params *p, return sqd; } -#if defined(CONFIG_UNIX) -/* - * Ensure the UNIX gc is aware of our file set, so we are certain that - * the io_uring can be safely unregistered on process exit, even if we have - * loops in the file referencing. - */ -static int __io_sqe_files_scm(struct io_ring_ctx *ctx, int nr, int offset) -{ - struct sock *sk = ctx->ring_sock->sk; - struct scm_fp_list *fpl; - struct sk_buff *skb; - int i, nr_files; - - fpl = kzalloc(sizeof(*fpl), GFP_KERNEL); - if (!fpl) - return -ENOMEM; - - skb = alloc_skb(0, GFP_KERNEL); - if (!skb) { - kfree(fpl); - return -ENOMEM; - } - - skb->sk = sk; - skb->scm_io_uring = 1; - - nr_files = 0; - fpl->user = get_uid(current_user()); - for (i = 0; i < nr; i++) { - struct file *file = io_file_from_index(ctx, i + offset); - - if (!file) - continue; - fpl->fp[nr_files] = get_file(file); - unix_inflight(fpl->user, fpl->fp[nr_files]); - nr_files++; - } - - if (nr_files) { - fpl->max = SCM_MAX_FD; - fpl->count = nr_files; - UNIXCB(skb).fp = fpl; - skb->destructor = unix_destruct_scm; - refcount_add(skb->truesize, &sk->sk_wmem_alloc); - skb_queue_head(&sk->sk_receive_queue, skb); - - for (i = 0; i < nr; i++) { - struct file *file = io_file_from_index(ctx, i + offset); - - if (file) - fput(file); - } - } else { - kfree_skb(skb); - free_uid(fpl->user); - kfree(fpl); - } - - return 0; -} - -/* - * If UNIX sockets are enabled, fd passing can cause a reference cycle which - * causes regular reference counting to break down. We rely on the UNIX - * garbage collection to take care of this problem for us. - */ -static int io_sqe_files_scm(struct io_ring_ctx *ctx) -{ - unsigned left, total; - int ret = 0; - - total = 0; - left = ctx->nr_user_files; - while (left) { - unsigned this_files = min_t(unsigned, left, SCM_MAX_FD); - - ret = __io_sqe_files_scm(ctx, this_files, total); - if (ret) - break; - left -= this_files; - total += this_files; - } - - if (!ret) - return 0; - - while (total < ctx->nr_user_files) { - struct file *file = io_file_from_index(ctx, total); - - if (file) - fput(file); - total++; - } - - return ret; -} -#else -static int io_sqe_files_scm(struct io_ring_ctx *ctx) -{ - return 0; -} -#endif - static void io_rsrc_file_put(struct io_ring_ctx *ctx, struct io_rsrc_put *prsrc) { struct file *file = prsrc->file; -#if defined(CONFIG_UNIX) - struct sock *sock = ctx->ring_sock->sk; - struct sk_buff_head list, *head = &sock->sk_receive_queue; - struct sk_buff *skb; - int i; - __skb_queue_head_init(&list); - - /* - * Find the skb that holds this file in its SCM_RIGHTS. When found, - * remove this entry and rearrange the file array. - */ - skb = skb_dequeue(head); - while (skb) { - struct scm_fp_list *fp; - - fp = UNIXCB(skb).fp; - for (i = 0; i < fp->count; i++) { - int left; - - if (fp->fp[i] != file) - continue; - - unix_notinflight(fp->user, fp->fp[i]); - left = fp->count - 1 - i; - if (left) { - memmove(&fp->fp[i], &fp->fp[i + 1], - left * sizeof(struct file *)); - } - fp->count--; - if (!fp->count) { - kfree_skb(skb); - skb = NULL; - } else { - __skb_queue_tail(&list, skb); - } - fput(file); - file = NULL; - break; - } - - if (!file) - break; - - __skb_queue_tail(&list, skb); - - skb = skb_dequeue(head); - } - - if (skb_peek(&list)) { - spin_lock_irq(&head->lock); - while ((skb = __skb_dequeue(&list)) != NULL) - __skb_queue_tail(head, skb); - spin_unlock_irq(&head->lock); - } -#else fput(file); -#endif } static void __io_rsrc_put_work(struct io_rsrc_node *ref_node) @@ -8595,12 +8422,6 @@ static int io_sqe_files_register(struct io_ring_ctx *ctx, void __user *arg, io_fixed_file_set(io_fixed_file_slot(&ctx->file_table, i), file); } - ret = io_sqe_files_scm(ctx); - if (ret) { - __io_sqe_files_unregister(ctx); - return ret; - } - io_rsrc_node_switch(ctx, NULL); return ret; out_fput: @@ -9557,12 +9378,6 @@ static void io_ring_ctx_free(struct io_ring_ctx *ctx) WARN_ON_ONCE(!list_empty(&ctx->rsrc_ref_list)); WARN_ON_ONCE(!llist_empty(&ctx->rsrc_put_llist)); -#if defined(CONFIG_UNIX) - if (ctx->ring_sock) { - ctx->ring_sock->file = NULL; /* so that iput() is called */ - sock_release(ctx->ring_sock); - } -#endif WARN_ON_ONCE(!list_empty(&ctx->ltimeout_list)); if (ctx->mm_account) { @@ -10499,32 +10314,12 @@ static int io_uring_install_fd(struct io_ring_ctx *ctx, struct file *file) /* * Allocate an anonymous fd, this is what constitutes the application * visible backing of an io_uring instance. The application mmaps this - * fd to gain access to the SQ/CQ ring details. If UNIX sockets are enabled, - * we have to tie this fd to a socket for file garbage collection purposes. + * fd to gain access to the SQ/CQ ring details. */ static struct file *io_uring_get_file(struct io_ring_ctx *ctx) { - struct file *file; -#if defined(CONFIG_UNIX) - int ret; - - ret = sock_create_kern(&init_net, PF_UNIX, SOCK_RAW, IPPROTO_IP, - &ctx->ring_sock); - if (ret) - return ERR_PTR(ret); -#endif - - file = anon_inode_getfile("[io_uring]", &io_uring_fops, ctx, - O_RDWR | O_CLOEXEC); -#if defined(CONFIG_UNIX) - if (IS_ERR(file)) { - sock_release(ctx->ring_sock); - ctx->ring_sock = NULL; - } else { - ctx->ring_sock->file = file; - } -#endif - return file; + return anon_inode_getfile("[io_uring]", &io_uring_fops, ctx, + O_RDWR | O_CLOEXEC); } static int io_uring_create(unsigned entries, struct io_uring_params *p, -- 2.43.0
From 7c36fae9614f48c971c458f3bb12b97d3414abab Mon Sep 17 00:00:00 2001 From: Jens Axboe <axboe@xxxxxxxxx> Date: Wed, 13 Mar 2024 17:54:49 -0600 Subject: [PATCH 1/2] io_uring/unix: drop usage of io_uring socket Commit a4104821ad651d8a0b374f0b2474c345bbb42f82 upstream. Since we no longer allow sending io_uring fds over SCM_RIGHTS, move to using io_is_uring_fops() to detect whether this is a io_uring fd or not. With that done, kill off io_uring_get_socket() as nobody calls it anymore. This is in preparation to yanking out the rest of the core related to unix gc with io_uring. Signed-off-by: Jens Axboe <axboe@xxxxxxxxx> --- include/linux/io_uring.h | 10 +++++----- io_uring/io_uring.c | 18 +++++------------- net/core/scm.c | 2 +- net/unix/scm.c | 4 +--- 4 files changed, 12 insertions(+), 22 deletions(-) diff --git a/include/linux/io_uring.h b/include/linux/io_uring.h index 649a4d7c241b..55d09f594cd1 100644 --- a/include/linux/io_uring.h +++ b/include/linux/io_uring.h @@ -6,9 +6,9 @@ #include <linux/xarray.h> #if defined(CONFIG_IO_URING) -struct sock *io_uring_get_socket(struct file *file); void __io_uring_cancel(bool cancel_all); void __io_uring_free(struct task_struct *tsk); +bool io_is_uring_fops(struct file *file); static inline void io_uring_files_cancel(void) { @@ -26,10 +26,6 @@ static inline void io_uring_free(struct task_struct *tsk) __io_uring_free(tsk); } #else -static inline struct sock *io_uring_get_socket(struct file *file) -{ - return NULL; -} static inline void io_uring_task_cancel(void) { } @@ -39,6 +35,10 @@ static inline void io_uring_files_cancel(void) static inline void io_uring_free(struct task_struct *tsk) { } +static inline bool io_is_uring_fops(struct file *file) +{ + return false; +} #endif #endif diff --git a/io_uring/io_uring.c b/io_uring/io_uring.c index 55fd6d98fe12..541e6a6985ec 100644 --- a/io_uring/io_uring.c +++ b/io_uring/io_uring.c @@ -1116,19 +1116,6 @@ static struct kmem_cache *req_cachep; static const struct file_operations io_uring_fops; -struct sock *io_uring_get_socket(struct file *file) -{ -#if defined(CONFIG_UNIX) - if (file->f_op == &io_uring_fops) { - struct io_ring_ctx *ctx = file->private_data; - - return ctx->ring_sock->sk; - } -#endif - return NULL; -} -EXPORT_SYMBOL(io_uring_get_socket); - static inline void io_tw_lock(struct io_ring_ctx *ctx, bool *locked) { if (!*locked) { @@ -10445,6 +10432,11 @@ static const struct file_operations io_uring_fops = { #endif }; +bool io_is_uring_fops(struct file *file) +{ + return file->f_op == &io_uring_fops; +} + static int io_allocate_scq_urings(struct io_ring_ctx *ctx, struct io_uring_params *p) { diff --git a/net/core/scm.c b/net/core/scm.c index e762a4b8a1d2..a877c4ef4c25 100644 --- a/net/core/scm.c +++ b/net/core/scm.c @@ -105,7 +105,7 @@ static int scm_fp_copy(struct cmsghdr *cmsg, struct scm_fp_list **fplp) if (fd < 0 || !(file = fget_raw(fd))) return -EBADF; /* don't allow io_uring files */ - if (io_uring_get_socket(file)) { + if (io_is_uring_fops(file)) { fput(file); return -EINVAL; } diff --git a/net/unix/scm.c b/net/unix/scm.c index e8e2a00bb0f5..d1048b4c2baa 100644 --- a/net/unix/scm.c +++ b/net/unix/scm.c @@ -34,10 +34,8 @@ struct sock *unix_get_socket(struct file *filp) /* PF_UNIX ? */ if (s && sock->ops && sock->ops->family == PF_UNIX) u_sock = s; - } else { - /* Could be an io_uring instance */ - u_sock = io_uring_get_socket(filp); } + return u_sock; } EXPORT_SYMBOL(unix_get_socket); -- 2.43.0