xprtrdma: Fix handling of RDMA_ERROR replies

The RPC client currently doesn't handle ERR_CHUNK replies correctly.
rpcrdma_complete_rqst() incorrectly passes a negative number to
xprt_complete_rqst() as the number of bytes copied. Instead, set
task->tk_status to the error value, and return zero bytes copied.

In these cases, return -EIO rather than -EREMOTEIO. The RPC client's
finite state machine doesn't know what to do with -EREMOTEIO.

Additional clean ups:
- Don't double-count RDMA_ERROR replies
- Remove a stale comment

Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Cc: <stable@kernel.vger.org>
Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>

diff --git a/net/sunrpc/xprtrdma/rpc_rdma.c b/net/sunrpc/xprtrdma/rpc_rdma.c
index 2081c8fbfa48d5f3a553bfecf4feb15f0459978f..935bbef2f7bef1388f33697e05d465296131511e 100644 (file)
--- a/net/sunrpc/xprtrdma/rpc_rdma.c
+++ b/net/sunrpc/xprtrdma/rpc_rdma.c
@@ -1349,8 +1349,7 @@ rpcrdma_decode_error(struct rpcrdma_xprt *r_xprt, struct rpcrdma_rep *rep,
                        be32_to_cpup(p), be32_to_cpu(rep->rr_xid));
        }
 
-       r_xprt->rx_stats.bad_reply_count++;
-       return -EREMOTEIO;
+       return -EIO;
 }
 
 /* Perform XID lookup, reconstruction of the RPC reply, and
@@ -1387,13 +1386,11 @@ out:
        spin_unlock(&xprt->queue_lock);
        return;
 
-/* If the incoming reply terminated a pending RPC, the next
- * RPC call will post a replacement receive buffer as it is
- * being marshaled.
- */
 out_badheader:
        trace_xprtrdma_reply_hdr(rep);
        r_xprt->rx_stats.bad_reply_count++;
+       rqst->rq_task->tk_status = status;
+       status = 0;
        goto out;
 }