diff options
| author | Chuck Lever <chuck.lever@oracle.com> | 2017-10-20 10:48:36 -0400 |
|---|---|---|
| committer | Anna Schumaker <Anna.Schumaker@Netapp.com> | 2017-11-17 13:47:57 -0500 |
| commit | 01bb35c89d90abe6fd1c0be001f84bbdfa7fa7d1 (patch) | |
| tree | 527e65074e175980f7ec757f189bd77f11f1b694 /net/sunrpc/xprtrdma/rpc_rdma.c | |
| parent | 0ba6f37012db2f88f881cd818aec6e1886f61abb (diff) | |
| download | op-kernel-dev-01bb35c89d90abe6fd1c0be001f84bbdfa7fa7d1.zip op-kernel-dev-01bb35c89d90abe6fd1c0be001f84bbdfa7fa7d1.tar.gz | |
xprtrdma: RPC completion should wait for Send completion
When an RPC Call includes a file data payload, that payload can come
from pages in the page cache, or a user buffer (for direct I/O).
If the payload can fit inline, xprtrdma includes it in the Send
using a scatter-gather technique. xprtrdma mustn't allow the RPC
consumer to re-use the memory where that payload resides before the
Send completes. Otherwise, the new contents of that memory would be
exposed by an HCA retransmit of the Send operation.
So, block RPC completion on Send completion, but only in the case
where a separate file data payload is part of the Send. This
prevents the reuse of that memory while it is still part of a Send
operation without an undue cost to other cases.
Waiting is avoided in the common case because typically the Send
will have completed long before the RPC Reply arrives.
These days, an RPC timeout will trigger a disconnect, which tears
down the QP. The disconnect flushes all waiting Sends. This bounds
the amount of time the reply handler has to wait for a Send
completion.
Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
Diffstat (limited to 'net/sunrpc/xprtrdma/rpc_rdma.c')
| -rw-r--r-- | net/sunrpc/xprtrdma/rpc_rdma.c | 26 |
1 files changed, 25 insertions, 1 deletions
diff --git a/net/sunrpc/xprtrdma/rpc_rdma.c b/net/sunrpc/xprtrdma/rpc_rdma.c index 853dede..4fdeaac 100644 --- a/net/sunrpc/xprtrdma/rpc_rdma.c +++ b/net/sunrpc/xprtrdma/rpc_rdma.c @@ -534,6 +534,11 @@ rpcrdma_unmap_sendctx(struct rpcrdma_sendctx *sc) for (count = sc->sc_unmap_count; count; ++sge, --count) ib_dma_unmap_page(ia->ri_device, sge->addr, sge->length, DMA_TO_DEVICE); + + if (test_and_clear_bit(RPCRDMA_REQ_F_TX_RESOURCES, &sc->sc_req->rl_flags)) { + smp_mb__after_atomic(); + wake_up_bit(&sc->sc_req->rl_flags, RPCRDMA_REQ_F_TX_RESOURCES); + } } /* Prepare an SGE for the RPC-over-RDMA transport header. @@ -667,6 +672,8 @@ map_tail: out: sc->sc_wr.num_sge += sge_no; + if (sc->sc_unmap_count) + __set_bit(RPCRDMA_REQ_F_TX_RESOURCES, &req->rl_flags); return true; out_regbuf: @@ -704,6 +711,8 @@ rpcrdma_prepare_send_sges(struct rpcrdma_xprt *r_xprt, return -ENOBUFS; req->rl_sendctx->sc_wr.num_sge = 0; req->rl_sendctx->sc_unmap_count = 0; + req->rl_sendctx->sc_req = req; + __clear_bit(RPCRDMA_REQ_F_TX_RESOURCES, &req->rl_flags); if (!rpcrdma_prepare_hdr_sge(&r_xprt->rx_ia, req, hdrlen)) return -EIO; @@ -1305,6 +1314,20 @@ void rpcrdma_release_rqst(struct rpcrdma_xprt *r_xprt, struct rpcrdma_req *req) if (!list_empty(&req->rl_registered)) r_xprt->rx_ia.ri_ops->ro_unmap_sync(r_xprt, &req->rl_registered); + + /* Ensure that any DMA mapped pages associated with + * the Send of the RPC Call have been unmapped before + * allowing the RPC to complete. This protects argument + * memory not controlled by the RPC client from being + * re-used before we're done with it. + */ + if (test_bit(RPCRDMA_REQ_F_TX_RESOURCES, &req->rl_flags)) { + r_xprt->rx_stats.reply_waits_for_send++; + out_of_line_wait_on_bit(&req->rl_flags, + RPCRDMA_REQ_F_TX_RESOURCES, + bit_wait, + TASK_UNINTERRUPTIBLE); + } } /* Reply handling runs in the poll worker thread. Anything that @@ -1384,7 +1407,8 @@ void rpcrdma_reply_handler(struct rpcrdma_rep *rep) dprintk("RPC: %s: reply %p completes request %p (xid 0x%08x)\n", __func__, rep, req, be32_to_cpu(rep->rr_xid)); - if (list_empty(&req->rl_registered)) + if (list_empty(&req->rl_registered) && + !test_bit(RPCRDMA_REQ_F_TX_RESOURCES, &req->rl_flags)) rpcrdma_complete_rqst(rep); else queue_work(rpcrdma_receive_wq, &rep->rr_work); |
