netfs: Fix AIO error handling when doing write-through
authorDavid Howells <dhowells@redhat.com>
Tue, 21 May 2024 13:36:27 +0000 (14:36 +0100)
committerChristian Brauner <brauner@kernel.org>
Fri, 24 May 2024 11:34:06 +0000 (13:34 +0200)
commit2c6b531020f0590db3b6b4950a41c692e9aa4f4a
tree818ba301a6e5605d512bce6d80a6cc061dea9294
parent9b038d004ce95551cb35381c49fe896c5bc11ffe
netfs: Fix AIO error handling when doing write-through

If an error occurs whilst we're doing an AIO write in write-through mode,
we may end up calling ->ki_complete() *and* returning an error from
->write_iter().  This can result in either a UAF (the ->ki_complete() func
pointer may get overwritten, for example) or a refcount underflow in
io_submit() as ->ki_complete is called twice.

Fix this by making netfs_end_writethrough() - and thus
netfs_perform_write() - unconditionally return -EIOCBQUEUED if we're doing
an AIO write and wait for completion if we're not.

Fixes: 288ace2f57c9 ("netfs: New writeback implementation")
Signed-off-by: David Howells <dhowells@redhat.com>
Link: https://lore.kernel.org/r/295052.1716298587@warthog.procyon.org.uk
cc: Jeff Layton <jlayton@kernel.org>
cc: Enzo Matsumiya <ematsumiya@suse.de>
cc: netfs@lists.linux.dev
cc: v9fs@lists.linux.dev
cc: linux-afs@lists.infradead.org
cc: linux-cifs@vger.kernel.org
cc: linux-fsdevel@vger.kernel.org
Signed-off-by: Christian Brauner <brauner@kernel.org>
fs/netfs/write_issue.c