Patch "nvme-rdma: don't complete requests before a send work request has completed" has been added to the 4.14-stable tree
Subsystems (2)
This is a note to let you know that I've just added the patch titled
nvme-rdma: don't complete requests before a send work request has completed
to the 4.14-stable tree which can be found at:
http://www.kernel.org/git/?p=linux/kernel/git/stable/stable-queue.git;a=summary
The filename of the patch is:
nvme-rdma-don-t-complete-requests-before-a-send-work-request-has-completed.patch
and it can be found in the queue-4.14 subdirectory.
If you, or anyone else, feels it should not be added to the stable tree, please let <[email protected]> know about it.
Quoted 1 line
From foo@baz Thu Feb 1 13:45:42 CET 2018
From: Sagi Grimberg <[email protected]>
Date: Thu, 23 Nov 2017 17:35:22 +0200
Subject: nvme-rdma: don't complete requests before a send work request has completed
From: Sagi Grimberg <[email protected]>
[ Upstream commit 4af7f7ff92a42b6c713293c99e7982bcfcf51a70 ]
In order to guarantee that the HCA will never get an access violation (either from invalidated rkey or from iommu) when retrying a send operation we must complete a request only when both send completion and the nvme cqe has arrived. We need to set the send/recv completions flags atomically because we might have more than a single context accessing the request concurrently (one is cq irq-poll context and the other is user-polling used in IOCB_HIPRI).
Only then we are safe to invalidate the rkey (if needed), unmap the host buffers, and complete the IO.
--- drivers/nvme/host/rdma.c | 28 ++++++++++++++++++++++++---- 1 file changed, 24 insertions(+), 4 deletions(-)
drivers/nvme/host/rdma.c +24 -4
@@ -67,6 +67,9 @@ struct nvme_rdma_request {struct nvme_request req;struct ib_mr *mr;struct nvme_rdma_qe sqe; +union nvme_result result; +__le16 status; +refcount_t ref;struct ib_sge sge[1 +NVME_RDMA_MAX_INLINE_SEGMENTS ];u32 num_sge;int nents;
@@ -1177,6 +1180,7 @@ static int nvme_rdma_map_data(struct nvm req->num_sge =1 ; req->inline_data = false; req->mr->need_inval = false; +refcount_set (&req->ref,2 );/* send and recv completions */ c->common.flags |=NVME_CMD_SGL_METABUF ;
@@ -1213,8 +1217,19 @@ static int nvme_rdma_map_data(struct nvmstatic void nvme_rdma_send_done (struct ib_cq *cq,struct ib_wc *wc) { -if (unlikely (wc->status !=IB_WC_SUCCESS )) +struct nvme_rdma_qe *qe = + container_of(wc->wr_cqe, struct nvme_rdma_qe, cqe); +struct nvme_rdma_request *req = + container_of(qe, struct nvme_rdma_request, sqe); +struct request *rq =blk_mq_rq_from_pdu (req); + +if (unlikely (wc->status !=IB_WC_SUCCESS )) {nvme_rdma_wr_error (cq, wc,"SEND" ); +return ; + } + +if (refcount_dec_and_test (&req->ref)) +nvme_end_request (rq, req->status, req->result); }/*
@@ -1359,14 +1374,19 @@ static int nvme_rdma_process_nvme_rsp(st } req =blk_mq_rq_to_pdu (rq); -if (rq->tag == tag) - ret =1 ; + req->status = cqe->status; + req->result = cqe->result;if ((wc->wc_flags &IB_WC_WITH_INVALIDATE ) && wc->ex.invalidate_rkey == req->mr->rkey) req->mr->need_inval = false; -nvme_end_request (rq, cqe->status, cqe->result); +if (refcount_dec_and_test (&req->ref)) { +if (rq->tag == tag) + ret =1 ; +nvme_end_request (rq, req->status, req->result); + } +return ret; }
Patches currently in stable-queue which might be from [email protected] are
queue-4.14/nvmet-fc-correct-ref-counting-error-when-deferred-rcv-used.patch queue-4.14/nvme-fabrics-introduce-init-command-check-for-a-queue-that-is-not-alive.patch queue-4.14/nvme-fc-check-if-queue-is-ready-in-queue_rq.patch queue-4.14/nvme-loop-check-if-queue-is-ready-in-queue_rq.patch queue-4.14/nvme-rdma-don-t-complete-requests-before-a-send-work-request-has-completed.patch