aio-posix: fix race between io_uring CQE and AioHandler deletion

When an AioHandler is enqueued on ctx->submit_list for removal, the
fill_sq_ring() function will submit an io_uring POLL_REMOVE operation to
cancel the in-flight POLL_ADD operation.

There is a race when another thread enqueues an AioHandler for deletion
on ctx->submit_list when the POLL_ADD CQE has already appeared. In that
case POLL_REMOVE is unnecessary. The code already handled this, but
forgot that the AioHandler itself is still on ctx->submit_list when the
POLL_ADD CQE is being processed. It's unsafe to delete the AioHandler at
that point in time (use-after-free).

Solve this problem by keeping the AioHandler alive but setting a flag so
that it will be deleted by fill_sq_ring() when it runs.

Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
Reviewed-by: Kevin Wolf <kwolf@redhat.com>
Message-ID: <20251104022933.618123-2-stefanha@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>

diff --git a/util/fdmon-io_uring.c b/util/fdmon-io_uring.c
index b0d68bdc44dda39d311f12d4e8addc6f35a63fae..ad89160f313d3976a6cb67f78c6b191e9aa128cc 100644 (file)
--- a/util/fdmon-io_uring.c
+++ b/util/fdmon-io_uring.c
@@ -52,9 +52,10 @@ enum {
     FDMON_IO_URING_ENTRIES  = 128, /* sq/cq ring size */
 
     /* AioHandler::flags */
-    FDMON_IO_URING_PENDING  = (1 << 0),
-    FDMON_IO_URING_ADD      = (1 << 1),
-    FDMON_IO_URING_REMOVE   = (1 << 2),
+    FDMON_IO_URING_PENDING            = (1 << 0),
+    FDMON_IO_URING_ADD                = (1 << 1),
+    FDMON_IO_URING_REMOVE             = (1 << 2),
+    FDMON_IO_URING_DELETE_AIO_HANDLER = (1 << 3),
 };
 
 static inline int poll_events_from_pfd(int pfd_events)
@@ -218,6 +219,16 @@ static void fill_sq_ring(AioContext *ctx)
         if (flags & FDMON_IO_URING_REMOVE) {
             add_poll_remove_sqe(ctx, node);
         }
+        if (flags & FDMON_IO_URING_DELETE_AIO_HANDLER) {
+            /*
+             * process_cqe() sets this flag after ADD and REMOVE have been
+             * cleared. They cannot be set again, so they must be clear.
+             */
+            assert(!(flags & FDMON_IO_URING_ADD));
+            assert(!(flags & FDMON_IO_URING_REMOVE));
+
+            QLIST_INSERT_HEAD_RCU(&ctx->deleted_aio_handlers, node, node_deleted);
+        }
     }
 }
 
@@ -241,7 +252,12 @@ static bool process_cqe(AioContext *ctx,
      */
     flags = qatomic_fetch_and(&node->flags, ~FDMON_IO_URING_REMOVE);
     if (flags & FDMON_IO_URING_REMOVE) {
-        QLIST_INSERT_HEAD_RCU(&ctx->deleted_aio_handlers, node, node_deleted);
+        if (flags & FDMON_IO_URING_PENDING) {
+            /* Still on ctx->submit_list, defer deletion until fill_sq_ring() */
+            qatomic_or(&node->flags, FDMON_IO_URING_DELETE_AIO_HANDLER);
+        } else {
+            QLIST_INSERT_HEAD_RCU(&ctx->deleted_aio_handlers, node, node_deleted);
+        }
         return false;
     }
 
@@ -347,10 +363,13 @@ void fdmon_io_uring_destroy(AioContext *ctx)
             unsigned flags = qatomic_fetch_and(&node->flags,
                     ~(FDMON_IO_URING_PENDING |
                       FDMON_IO_URING_ADD |
-                      FDMON_IO_URING_REMOVE));
+                      FDMON_IO_URING_REMOVE |
+                      FDMON_IO_URING_DELETE_AIO_HANDLER));
 
-            if (flags & FDMON_IO_URING_REMOVE) {
-                QLIST_INSERT_HEAD_RCU(&ctx->deleted_aio_handlers, node, node_deleted);
+            if ((flags & FDMON_IO_URING_REMOVE) ||
+                (flags & FDMON_IO_URING_DELETE_AIO_HANDLER)) {
+                QLIST_INSERT_HEAD_RCU(&ctx->deleted_aio_handlers,
+                                      node, node_deleted);
             }
 
             QSLIST_REMOVE_HEAD_RCU(&ctx->submit_list, node_submitted);