Re: [PATCH v2 09/10] nfsd: Fix delegation revocation

[Christoph Hellwig <hch@infradead.org>]

> -/* Called under the state lock. */

You're removing a comment hat only becomes true after your patch :)

But I agree that it's obsolete, the lockdep_assert_held is better than
the comment.

> @@ -685,11 +686,10 @@ static void revoke_delegation(struct nfs4_delegation *dp)
>  	struct nfs4_client *clp = dp->dl_stid.sc_client;
>  
>  	if (clp->cl_minorversion == 0)
> -		destroy_delegation(dp);
> +		destroy_revoked_delegation(dp);

Seems a little confusing as we didn't turn the code into a REVOKED_DELEG
stateid type yet, but given that destroy_revoked_delegation doesn't
care this is probably fine.

>  	else {
> -		unhash_delegation(dp);
>  		dp->dl_stid.sc_type = NFS4_REVOKED_DELEG_STID;
> -		list_add(&dp->dl_recall_lru, &clp->cl_revoked);
> +		list_move(&dp->dl_recall_lru, &clp->cl_revoked);

What protects access to cl_revoked, btw?

Otherwise this looks fine to me,

Reviewed-by: Christoph Hellwig <hch@lst.de>

(would have been a little nicer to read if you had kept
destroy_revoked_delegation, otherwise diff obsfucates the fact that it
didn't change at all..