I found that injecting disconnects with v4.18-rc resulted in random failures of the multi-threaded git regression test.
The root cause appears to be that, after a reconnect, the RPC/RDMA transport is waking pending RPCs before the transport has posted enough Receive buffers to receive the Replies. If a Reply arrives before enough Receive buffers are posted, the connection is dropped. A few connection drops happen in quick succession as the client and server struggle to regain credit synchronization.
This regression was introduced with commit 7c8d9e7c8863 ("xprtrdma: Move Receive posting to Receive handler"). The client is supposed to post a single Receive when a connection is established because it's not supposed to send more than one RPC Call before it gets a fresh credit grant in the first RPC Reply [RFC 8166, Section 3.3.3].
Unfortunately there appears to be a longstanding bug in the Linux client's credit accounting mechanism. On connect, it simply dumps all pending RPC Calls onto the new connection. It's possible it has done this ever since the RPC/RDMA transport was added to the kernel ten years ago.
Servers have so far been tolerant of this bad behavior. Currently no server implementation ever changes its credit grant over reconnects, and servers always repost enough Receives before connections are fully established.
The Linux client implementation used to post a Receive before each of these Calls. This has covered up the flooding send behavior.
I could try to correct this old bug so that the client sends exactly one RPC Call and waits for a Reply. Since we are so close to the next merge window, I'm going to instead provide a simple patch to post enough Receives before a reconnect completes (based on the number of credits granted to the previous connection).
The spurious disconnects will be gone, but the client will still send multiple RPC Calls immediately after a reconnect.
Addressing the latter problem will wait for a merge window because a) I expect it to be a large change requiring lots of testing, and b) obviously the Linux client has interoperated successfully since day zero while still being broken.
Fixes: 7c8d9e7c8863 ("xprtrdma: Move Receive posting to ... ") Signed-off-by: Chuck Lever chuck.lever@oracle.com --- net/sunrpc/xprtrdma/verbs.c | 5 ++--- 1 file changed, 2 insertions(+), 3 deletions(-)
Hi stable@ -
This fix has been merged into v4.19 as upstream commit 8d4fb8ff427a ("xprtrdma: Fix disconnect regression"). It addresses a regression in v4.18. I expected it to go into late v4.18-rc, which is why there is no "cc: stable" on the original submission.
Could you please apply it to 4.18.y ? Thank you!
diff --git a/net/sunrpc/xprtrdma/verbs.c b/net/sunrpc/xprtrdma/verbs.c index 042bb24..1386c6e 100644 --- a/net/sunrpc/xprtrdma/verbs.c +++ b/net/sunrpc/xprtrdma/verbs.c @@ -279,7 +279,6 @@ ++xprt->rx_xprt.connect_cookie; connstate = -ECONNABORTED; connected: - xprt->rx_buf.rb_credits = 1; ep->rep_connected = connstate; rpcrdma_conn_func(ep); wake_up_all(&ep->rep_connect_wait); @@ -754,6 +753,7 @@ }
ep->rep_connected = 0; + rpcrdma_post_recvs(r_xprt, true);
rc = rdma_connect(ia->ri_id, &ep->rep_remote_cma); if (rc) { @@ -772,8 +772,6 @@
dprintk("RPC: %s: connected\n", __func__);
- rpcrdma_post_recvs(r_xprt, true); - out: if (rc) ep->rep_connected = rc; @@ -1170,6 +1168,7 @@ struct rpcrdma_req * list_add(&req->rl_list, &buf->rb_send_bufs); }
+ buf->rb_credits = 1; buf->rb_posted_receives = 0; INIT_LIST_HEAD(&buf->rb_recv_bufs);