From 54a485e9ec084da1a4b32dcf7749c7d760ed8aa5 Mon Sep 17 00:00:00 2001 From: Mike Marciniszyn Date: Tue, 28 Jul 2020 14:38:48 -0400 Subject: IB/rdmavt: Fix RQ counting issues causing use of an invalid RWQE The lookaside count is improperly initialized to the size of the Receive Queue with the additional +1. In the traces below, the RQ size is 384, so the count was set to 385. The lookaside count is then rarely refreshed. Note the high and incorrect count in the trace below: rvt_get_rwqe: [hfi1_0] wqe ffffc900078e9008 wr_id 55c7206d75a0 qpn c qpt 2 pid 3018 num_sge 1 head 1 tail 0, count 385 rvt_get_rwqe: (hfi1_rc_rcv+0x4eb/0x1480 [hfi1] <- rvt_get_rwqe) ret=0x1 The head,tail indicate there is only one RWQE posted although the count says 385 and we correctly return the element 0. The next call to rvt_get_rwqe with the decremented count: rvt_get_rwqe: [hfi1_0] wqe ffffc900078e9058 wr_id 0 qpn c qpt 2 pid 3018 num_sge 0 head 1 tail 1, count 384 rvt_get_rwqe: (hfi1_rc_rcv+0x4eb/0x1480 [hfi1] <- rvt_get_rwqe) ret=0x1 Note that the RQ is empty (head == tail) yet we return the RWQE at tail 1, which is not valid because of the bogus high count. Best case, the RWQE has never been posted and the rc logic sees an RWQE that is too small (all zeros) and puts the QP into an error state. In the worst case, a server slow at posting receive buffers might fool rvt_get_rwqe() into fetching an old RWQE and corrupt memory. Fix by deleting the faulty initialization code and creating an inline to fetch the posted count and convert all callers to use new inline. Fixes: f592ae3c999f ("IB/rdmavt: Fracture single lock used for posting and processing RWQEs") Link: https://lore.kernel.org/r/20200728183848.22226.29132.stgit@awfm-01.aw.intel.com Reported-by: Zhaojuan Guo Cc: # 5.4.x Reviewed-by: Kaike Wan Signed-off-by: Mike Marciniszyn Tested-by: Honggang Li Signed-off-by: Jason Gunthorpe --- include/rdma/rdmavt_qp.h | 19 +++++++++++++++++++ 1 file changed, 19 insertions(+) (limited to 'include') diff --git a/include/rdma/rdmavt_qp.h b/include/rdma/rdmavt_qp.h index c4369a6c2951..2f1fc23602cb 100644 --- a/include/rdma/rdmavt_qp.h +++ b/include/rdma/rdmavt_qp.h @@ -305,6 +305,25 @@ struct rvt_rq { spinlock_t lock ____cacheline_aligned_in_smp; }; +/** + * rvt_get_rq_count - count numbers of request work queue entries + * in circular buffer + * @rq: data structure for request queue entry + * @head: head indices of the circular buffer + * @tail: tail indices of the circular buffer + * + * Return - total number of entries in the Receive Queue + */ + +static inline u32 rvt_get_rq_count(struct rvt_rq *rq, u32 head, u32 tail) +{ + u32 count = head - tail; + + if ((s32)count < 0) + count += rq->size; + return count; +} + /* * This structure holds the information that the send tasklet needs * to send a RDMA read response or atomic operation. -- cgit