fs/nfs/io: make nfs_start_io_*() killable

[ Upstream commit 38a125b3 ]

This allows killing processes that wait for a lock when one process is
stuck waiting for the NFS server.  This aims to complete the coverage
of NFS operations being killable, like nfs_direct_wait() does, for
example.

Signed-off-by: Max Kellermann <max.kellermann@ionos.com>
Signed-off-by: Trond Myklebust <trond.myklebust@hammerspace.com>
Stable-dep-of: 9eb90f43 ("NFS: Serialise O_DIRECT i/o and truncate()")
Signed-off-by: Sasha Levin <sashal@kernel.org>