Re: [PATCH v2] nfsd: Fix races between nfsd4_cb_release() and nfsd4_shutdown_callback()

From: "J. Bruce Fields" <bfields@fieldses.org>
To: "J. Bruce Fields" <bfields@redhat.com>
Cc: Trond Myklebust <trondmy@hammerspace.com>,
	"linux-nfs@vger.kernel.org" <linux-nfs@vger.kernel.org>
Subject: Re: [PATCH v2] nfsd: Fix races between nfsd4_cb_release() and nfsd4_shutdown_callback()
Date: Thu, 7 Nov 2019 17:27:12 -0500	[thread overview]
Message-ID: <20191107222712.GB10806@fieldses.org> (raw)
In-Reply-To: <20191029214705.GA29280@fieldses.org>

On Tue, Oct 29, 2019 at 05:47:05PM -0400, J. Bruce Fields wrote:
> On Fri, Oct 25, 2019 at 11:33:36AM -0400, bfields wrote:
> > On Fri, Oct 25, 2019 at 11:21:19AM -0400, J. Bruce Fields wrote:
> > > I thought I was running v2, let me double-check....
> > 
> > Yes, with v2 I'm getting a hang on generic/013.
> > 
> > I checked quickly and didn't see anything interesting in the logs,
> > otherwise I haven't done any digging.
> 
> Reproduceable just with ./check -nfs generic/013.  The last thing I see
> in wireshark is an asynchronous COPY call and reply.  Which means it's
> probably trying to do a CB_OFFLOAD.  Hm.

Oh, I think it just needs the following.

--b.

diff --git a/fs/nfsd/nfs4callback.c b/fs/nfsd/nfs4callback.c
index fb71e7f9d0d9..e49604701a71 100644
--- a/fs/nfsd/nfs4callback.c
+++ b/fs/nfsd/nfs4callback.c
@@ -1026,8 +1026,8 @@ static bool nfsd41_cb_get_slot(struct nfsd4_callback *cb, struct rpc_task *task)
 			return false;
 		}
 		rpc_wake_up_queued_task(&clp->cl_cb_waitq, task);
-		cb->cb_holds_slot = true;
 	}
+	cb->cb_holds_slot = true;
 	return true;
 }
 
