On Fri, 2024-08-23 at 10:14 -0400, Olga Kornievskaia wrote:
> states_show() relied on sc_type field to be of valid type
> before calling into a subfunction to show content of a
> particular stateid. But from commit 3f29cc82a84c we
> split the validity of the stateid into sc_status and no longer
> changed sc_type to 0 while unhashing the stateid. This
> resulted in kernel oopsing as something like
> nfs4_show_open() would derefence sc_file which was NULL.
>
> To reproduce: mount the server with 4.0, read and close
> a file and then on the server cat /proc/fs/nfsd/clients/2/states
>
> [ 513.590804] Call trace:
> [ 513.590925] _raw_spin_lock+0xcc/0x160
> [ 513.591119] nfs4_show_open+0x78/0x2c0 [nfsd]
> [ 513.591412] states_show+0x44c/0x488 [nfsd]
> [ 513.591681] seq_read_iter+0x5d8/0x760
> [ 513.591896] seq_read+0x188/0x208
> [ 513.592075] vfs_read+0x148/0x470
> [ 513.592241] ksys_read+0xcc/0x178
>
> Fixes: 3f29cc82a84c ("nfsd: split sc_status out of sc_type")
> Signed-off-by: Olga Kornievskaia <okorniev@xxxxxxxxxx>
> ---
> fs/nfsd/nfs4state.c | 3 +++
> 1 file changed, 3 insertions(+)
>
> diff --git a/fs/nfsd/nfs4state.c b/fs/nfsd/nfs4state.c
> index c3def49074a4..8351724b8a43 100644
> --- a/fs/nfsd/nfs4state.c
> +++ b/fs/nfsd/nfs4state.c
> @@ -2907,6 +2907,9 @@ static int states_show(struct seq_file *s, void *v)
> {
> struct nfs4_stid *st = v;
>
> + if (!st->sc_file)
> + return 0;
> +
> switch (st->sc_type) {
> case SC_TYPE_OPEN:
> return nfs4_show_open(s, st);
OPEN stateids are the only ones that stick around for a while after
we've closed them out (and that's only for v4.0, IIRC). The others all
only release the file after unhashing.
I think it'd be better to still display that stateid, and just not do
any of the bits in nfs4_show_open that require ->sc_file. We have the
"admin-revoked" string that we display now when SC_STATUS_ADMIN_REVOKED
is set. Maybe we can display "closed" or something in this case?
--
Jeff Layton <jlayton@xxxxxxxxxx>
