From 90cee52f2e780345d3629e278291aea5ac74f40f Mon Sep 17 00:00:00 2001 From: Dust Li Date: Tue, 28 Dec 2021 17:03:24 +0800 Subject: [PATCH] net/smc: don't send CDC/LLC message if link not ready We found smc_llc_send_link_delete_all() sometimes wait for 2s timeout when testing with RDMA link up/down. It is possible when a smc_link is in ACTIVATING state, the underlaying QP is still in RESET or RTR state, which cannot send any messages out. smc_llc_send_link_delete_all() use smc_link_usable() to checks whether the link is usable, if the QP is still in RESET or RTR state, but the smc_link is in ACTIVATING, this LLC message will always fail without any CQE entering the CQ, and we will always wait 2s before timeout. Since we cannot send any messages through the QP before the QP enter RTS. I add a wrapper smc_link_sendable() which checks the state of QP along with the link state. And replace smc_link_usable() with smc_link_sendable() in all LLC & CDC message sending routine. Fixes: 5f08318f617b ("smc: connection data control (CDC)") Signed-off-by: Dust Li Signed-off-by: David S. Miller --- net/smc/smc_core.c | 2 +- net/smc/smc_core.h | 6 ++++++ net/smc/smc_llc.c | 2 +- net/smc/smc_wr.c | 4 ++-- net/smc/smc_wr.h | 2 +- 5 files changed, 11 insertions(+), 5 deletions(-) diff --git a/net/smc/smc_core.c b/net/smc/smc_core.c index 387d28b2f8dd..55ca175e8d57 100644 --- a/net/smc/smc_core.c +++ b/net/smc/smc_core.c @@ -647,7 +647,7 @@ static void smcr_lgr_link_deactivate_all(struct smc_link_group *lgr) for (i = 0; i < SMC_LINKS_PER_LGR_MAX; i++) { struct smc_link *lnk = &lgr->lnk[i]; - if (smc_link_usable(lnk)) + if (smc_link_sendable(lnk)) lnk->state = SMC_LNK_INACTIVE; } wake_up_all(&lgr->llc_msg_waiter); diff --git a/net/smc/smc_core.h b/net/smc/smc_core.h index 59cef3b830d8..d63b08274197 100644 --- a/net/smc/smc_core.h +++ b/net/smc/smc_core.h @@ -415,6 +415,12 @@ static inline bool smc_link_usable(struct smc_link *lnk) return true; } +static inline bool smc_link_sendable(struct smc_link *lnk) +{ + return smc_link_usable(lnk) && + lnk->qp_attr.cur_qp_state == IB_QPS_RTS; +} + static inline bool smc_link_active(struct smc_link *lnk) { return lnk->state == SMC_LNK_ACTIVE; diff --git a/net/smc/smc_llc.c b/net/smc/smc_llc.c index b102680296b8..3e9fd8a3124c 100644 --- a/net/smc/smc_llc.c +++ b/net/smc/smc_llc.c @@ -1630,7 +1630,7 @@ void smc_llc_send_link_delete_all(struct smc_link_group *lgr, bool ord, u32 rsn) delllc.reason = htonl(rsn); for (i = 0; i < SMC_LINKS_PER_LGR_MAX; i++) { - if (!smc_link_usable(&lgr->lnk[i])) + if (!smc_link_sendable(&lgr->lnk[i])) continue; if (!smc_llc_send_message_wait(&lgr->lnk[i], &delllc)) break; diff --git a/net/smc/smc_wr.c b/net/smc/smc_wr.c index 79a7431f534e..df1dc225cbab 100644 --- a/net/smc/smc_wr.c +++ b/net/smc/smc_wr.c @@ -188,7 +188,7 @@ void smc_wr_tx_cq_handler(struct ib_cq *ib_cq, void *cq_context) static inline int smc_wr_tx_get_free_slot_index(struct smc_link *link, u32 *idx) { *idx = link->wr_tx_cnt; - if (!smc_link_usable(link)) + if (!smc_link_sendable(link)) return -ENOLINK; for_each_clear_bit(*idx, link->wr_tx_mask, link->wr_tx_cnt) { if (!test_and_set_bit(*idx, link->wr_tx_mask)) @@ -231,7 +231,7 @@ int smc_wr_tx_get_free_slot(struct smc_link *link, } else { rc = wait_event_interruptible_timeout( link->wr_tx_wait, - !smc_link_usable(link) || + !smc_link_sendable(link) || lgr->terminating || (smc_wr_tx_get_free_slot_index(link, &idx) != -EBUSY), SMC_WR_TX_WAIT_FREE_SLOT_TIME); diff --git a/net/smc/smc_wr.h b/net/smc/smc_wr.h index f353311e6f84..48ed9b08ac7a 100644 --- a/net/smc/smc_wr.h +++ b/net/smc/smc_wr.h @@ -62,7 +62,7 @@ static inline void smc_wr_tx_set_wr_id(atomic_long_t *wr_tx_id, long val) static inline bool smc_wr_tx_link_hold(struct smc_link *link) { - if (!smc_link_usable(link)) + if (!smc_link_sendable(link)) return false; atomic_inc(&link->wr_tx_refcnt); return true;