Merge tag 'io_uring-6.7-2023-12-08' of git://git.kernel.dk/linux

Pull io_uring fixes from Jens Axboe:
 "Two minor fixes for issues introduced in this release cycle, and two
  fixes for issues or potential issues that are heading to stable.

  One of these ends up disabling passing io_uring file descriptors via
  SCM_RIGHTS. There really shouldn't be an overlap between that kind of
  historic use case and modern usage of io_uring, which is why this was
  deemed appropriate"

* tag 'io_uring-6.7-2023-12-08' of git://git.kernel.dk/linux:
  io_uring/af_unix: disable sending io_uring over sockets
  io_uring/kbuf: check for buffer list readiness after NULL check
  io_uring/kbuf: Fix an NULL vs IS_ERR() bug in io_alloc_pbuf_ring()
  io_uring: fix mutex_unlock with unreferenced ctx

Author: torvalds

io_uring/io_uring.c

@@ -271,13 +271,15 @@ static __cold void io_fallback_req_func(struct work_struct *work)
 	struct io_kiocb *req, *tmp;
 	struct io_tw_state ts = { .locked = true, };
 
+	percpu_ref_get(&ctx->refs);
 	mutex_lock(&ctx->uring_lock);
 	llist_for_each_entry_safe(req, tmp, node, io_task_work.node)
 		req->io_task_work.func(req, &ts);
 	if (WARN_ON_ONCE(!ts.locked))
 		return;
 	io_submit_flush_completions(ctx);
 	mutex_unlock(&ctx->uring_lock);
+	percpu_ref_put(&ctx->refs);
 }
 
 static int io_alloc_hash_table(struct io_hash_table *table, unsigned bits)
@@ -3146,12 +3148,7 @@ static __cold void io_ring_exit_work(struct work_struct *work)
 	init_completion(&exit.completion);
 	init_task_work(&exit.task_work, io_tctx_exit_cb);
 	exit.ctx = ctx;
-	/*
-	 * Some may use context even when all refs and requests have been put,
-	 * and they are free to do so while still holding uring_lock or
-	 * completion_lock, see io_req_task_submit(). Apart from other work,
-	 * this lock/unlock section also waits them to finish.
-	 */
+
 	mutex_lock(&ctx->uring_lock);
 	while (!list_empty(&ctx->tctx_list)) {
 		WARN_ON_ONCE(time_after(jiffies, timeout));

io_uring/kbuf.c

@@ -636,8 +636,8 @@ static int io_alloc_pbuf_ring(struct io_ring_ctx *ctx,
 	ibf = io_lookup_buf_free_entry(ctx, ring_size);
 	if (!ibf) {
 		ptr = io_mem_alloc(ring_size);
-		if (!ptr)
-			return -ENOMEM;
+		if (IS_ERR(ptr))
+			return PTR_ERR(ptr);
 
 		/* Allocate and store deferred free entry */
 		ibf = kmalloc(sizeof(*ibf), GFP_KERNEL_ACCOUNT);
@@ -756,15 +756,15 @@ void *io_pbuf_get_address(struct io_ring_ctx *ctx, unsigned long bgid)
 
 	bl = __io_buffer_get_list(ctx, smp_load_acquire(&ctx->io_bl), bgid);
 
+	if (!bl || !bl->is_mmap)
+		return NULL;
 	/*
 	 * Ensure the list is fully setup. Only strictly needed for RCU lookup
 	 * via mmap, and in that case only for the array indexed groups. For
 	 * the xarray lookups, it's either visible and ready, or not at all.
 	 */
 	if (!smp_load_acquire(&bl->is_ready))
 		return NULL;
-	if (!bl || !bl->is_mmap)
-		return NULL;
 
 	return bl->buf_ring;
 }

io_uring/rsrc.h

@@ -77,17 +77,10 @@ int io_sqe_files_register(struct io_ring_ctx *ctx, void __user *arg,
 
 int __io_scm_file_account(struct io_ring_ctx *ctx, struct file *file);
 
-#if defined(CONFIG_UNIX)
-static inline bool io_file_need_scm(struct file *filp)
-{
-	return !!unix_get_socket(filp);
-}
-#else
 static inline bool io_file_need_scm(struct file *filp)
 {
 	return false;
 }
-#endif
 
 static inline int io_scm_file_account(struct io_ring_ctx *ctx,
 				      struct file *file)

net/core/scm.c

@@ -26,6 +26,7 @@
 #include <linux/nsproxy.h>
 #include <linux/slab.h>
 #include <linux/errqueue.h>
+#include <linux/io_uring.h>
 
 #include <linux/uaccess.h>
 
@@ -103,6 +104,11 @@ 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)) {
+			fput(file);
+			return -EINVAL;
+		}
 		*fpp++ = file;
 		fpl->count++;
 	}