diff --git a/include/uapi/linux/smc.h b/include/uapi/linux/smc.h index 0e11ca421ca49344fd03763c3123f44bfed2c970..0f7f87c70baf5905ce37eb414880500745f4f1a0 100644 --- a/include/uapi/linux/smc.h +++ b/include/uapi/linux/smc.h @@ -33,4 +33,213 @@ enum { /* SMC PNET Table commands */ #define SMCR_GENL_FAMILY_NAME "SMC_PNETID" #define SMCR_GENL_FAMILY_VERSION 1 +/* gennetlink interface to access non-socket information from SMC module */ +#define SMC_GENL_FAMILY_NAME "SMC_GEN_NETLINK" +#define SMC_GENL_FAMILY_VERSION 1 + +#define SMC_PCI_ID_STR_LEN 16 /* Max length of pci id string */ + +/* SMC_GENL_FAMILY commands */ +enum { + SMC_NETLINK_GET_SYS_INFO = 1, + SMC_NETLINK_GET_LGR_SMCR, + SMC_NETLINK_GET_LINK_SMCR, + SMC_NETLINK_GET_LGR_SMCD, + SMC_NETLINK_GET_DEV_SMCD, + SMC_NETLINK_GET_DEV_SMCR, + SMC_NETLINK_GET_STATS, + SMC_NETLINK_GET_FBACK_STATS, +}; + +/* SMC_GENL_FAMILY top level attributes */ +enum { + SMC_GEN_UNSPEC, + SMC_GEN_SYS_INFO, /* nest */ + SMC_GEN_LGR_SMCR, /* nest */ + SMC_GEN_LINK_SMCR, /* nest */ + SMC_GEN_LGR_SMCD, /* nest */ + SMC_GEN_DEV_SMCD, /* nest */ + SMC_GEN_DEV_SMCR, /* nest */ + SMC_GEN_STATS, /* nest */ + SMC_GEN_FBACK_STATS, /* nest */ + __SMC_GEN_MAX, + SMC_GEN_MAX = __SMC_GEN_MAX - 1 +}; + +/* SMC_GEN_SYS_INFO attributes */ +enum { + SMC_NLA_SYS_UNSPEC, + SMC_NLA_SYS_VER, /* u8 */ + SMC_NLA_SYS_REL, /* u8 */ + SMC_NLA_SYS_IS_ISM_V2, /* u8 */ + SMC_NLA_SYS_LOCAL_HOST, /* string */ + SMC_NLA_SYS_SEID, /* string */ + __SMC_NLA_SYS_MAX, + SMC_NLA_SYS_MAX = __SMC_NLA_SYS_MAX - 1 +}; + +/* SMC_NLA_LGR_V2 nested attributes */ +enum { + SMC_NLA_LGR_V2_VER, /* u8 */ + SMC_NLA_LGR_V2_REL, /* u8 */ + SMC_NLA_LGR_V2_OS, /* u8 */ + SMC_NLA_LGR_V2_NEG_EID, /* string */ + SMC_NLA_LGR_V2_PEER_HOST, /* string */ +}; + +/* SMC_GEN_LGR_SMCR attributes */ +enum { + SMC_NLA_LGR_R_UNSPEC, + SMC_NLA_LGR_R_ID, /* u32 */ + SMC_NLA_LGR_R_ROLE, /* u8 */ + SMC_NLA_LGR_R_TYPE, /* u8 */ + SMC_NLA_LGR_R_PNETID, /* string */ + SMC_NLA_LGR_R_VLAN_ID, /* u8 */ + SMC_NLA_LGR_R_CONNS_NUM, /* u32 */ + __SMC_NLA_LGR_R_MAX, + SMC_NLA_LGR_R_MAX = __SMC_NLA_LGR_R_MAX - 1 +}; + +/* SMC_GEN_LINK_SMCR attributes */ +enum { + SMC_NLA_LINK_UNSPEC, + SMC_NLA_LINK_ID, /* u8 */ + SMC_NLA_LINK_IB_DEV, /* string */ + SMC_NLA_LINK_IB_PORT, /* u8 */ + SMC_NLA_LINK_GID, /* string */ + SMC_NLA_LINK_PEER_GID, /* string */ + SMC_NLA_LINK_CONN_CNT, /* u32 */ + SMC_NLA_LINK_NET_DEV, /* u32 */ + SMC_NLA_LINK_UID, /* u32 */ + SMC_NLA_LINK_PEER_UID, /* u32 */ + SMC_NLA_LINK_STATE, /* u32 */ + __SMC_NLA_LINK_MAX, + SMC_NLA_LINK_MAX = __SMC_NLA_LINK_MAX - 1 +}; + +/* SMC_GEN_LGR_SMCD attributes */ +enum { + SMC_NLA_LGR_D_UNSPEC, + SMC_NLA_LGR_D_ID, /* u32 */ + SMC_NLA_LGR_D_GID, /* u64 */ + SMC_NLA_LGR_D_PEER_GID, /* u64 */ + SMC_NLA_LGR_D_VLAN_ID, /* u8 */ + SMC_NLA_LGR_D_CONNS_NUM, /* u32 */ + SMC_NLA_LGR_D_PNETID, /* string */ + SMC_NLA_LGR_D_CHID, /* u16 */ + SMC_NLA_LGR_D_PAD, /* flag */ + SMC_NLA_LGR_V2, /* nest */ + __SMC_NLA_LGR_D_MAX, + SMC_NLA_LGR_D_MAX = __SMC_NLA_LGR_D_MAX - 1 +}; + +/* SMC_NLA_DEV_PORT nested attributes */ +enum { + SMC_NLA_DEV_PORT_UNSPEC, + SMC_NLA_DEV_PORT_PNET_USR, /* u8 */ + SMC_NLA_DEV_PORT_PNETID, /* string */ + SMC_NLA_DEV_PORT_NETDEV, /* u32 */ + SMC_NLA_DEV_PORT_STATE, /* u8 */ + SMC_NLA_DEV_PORT_VALID, /* u8 */ + SMC_NLA_DEV_PORT_LNK_CNT, /* u32 */ + __SMC_NLA_DEV_PORT_MAX, + SMC_NLA_DEV_PORT_MAX = __SMC_NLA_DEV_PORT_MAX - 1 +}; + +/* SMC_GEN_DEV_SMCD and SMC_GEN_DEV_SMCR attributes */ +enum { + SMC_NLA_DEV_UNSPEC, + SMC_NLA_DEV_USE_CNT, /* u32 */ + SMC_NLA_DEV_IS_CRIT, /* u8 */ + SMC_NLA_DEV_PCI_FID, /* u32 */ + SMC_NLA_DEV_PCI_CHID, /* u16 */ + SMC_NLA_DEV_PCI_VENDOR, /* u16 */ + SMC_NLA_DEV_PCI_DEVICE, /* u16 */ + SMC_NLA_DEV_PCI_ID, /* string */ + SMC_NLA_DEV_PORT, /* nest */ + SMC_NLA_DEV_PORT2, /* nest */ + SMC_NLA_DEV_IB_NAME, /* string */ + __SMC_NLA_DEV_MAX, + SMC_NLA_DEV_MAX = __SMC_NLA_DEV_MAX - 1 +}; + +/* SMC_NLA_STATS_T_TX(RX)_RMB_SIZE nested attributes */ +/* SMC_NLA_STATS_TX(RX)PLOAD_SIZE nested attributes */ +enum { + SMC_NLA_STATS_PLOAD_PAD, + SMC_NLA_STATS_PLOAD_8K, /* u64 */ + SMC_NLA_STATS_PLOAD_16K, /* u64 */ + SMC_NLA_STATS_PLOAD_32K, /* u64 */ + SMC_NLA_STATS_PLOAD_64K, /* u64 */ + SMC_NLA_STATS_PLOAD_128K, /* u64 */ + SMC_NLA_STATS_PLOAD_256K, /* u64 */ + SMC_NLA_STATS_PLOAD_512K, /* u64 */ + SMC_NLA_STATS_PLOAD_1024K, /* u64 */ + SMC_NLA_STATS_PLOAD_G_1024K, /* u64 */ + __SMC_NLA_STATS_PLOAD_MAX, + SMC_NLA_STATS_PLOAD_MAX = __SMC_NLA_STATS_PLOAD_MAX - 1 +}; + +/* SMC_NLA_STATS_T_TX(RX)_RMB_STATS nested attributes */ +enum { + SMC_NLA_STATS_RMB_PAD, + SMC_NLA_STATS_RMB_SIZE_SM_PEER_CNT, /* u64 */ + SMC_NLA_STATS_RMB_SIZE_SM_CNT, /* u64 */ + SMC_NLA_STATS_RMB_FULL_PEER_CNT, /* u64 */ + SMC_NLA_STATS_RMB_FULL_CNT, /* u64 */ + SMC_NLA_STATS_RMB_REUSE_CNT, /* u64 */ + SMC_NLA_STATS_RMB_ALLOC_CNT, /* u64 */ + SMC_NLA_STATS_RMB_DGRADE_CNT, /* u64 */ + __SMC_NLA_STATS_RMB_MAX, + SMC_NLA_STATS_RMB_MAX = __SMC_NLA_STATS_RMB_MAX - 1 +}; + +/* SMC_NLA_STATS_SMCD_TECH and _SMCR_TECH nested attributes */ +enum { + SMC_NLA_STATS_T_PAD, + SMC_NLA_STATS_T_TX_RMB_SIZE, /* nest */ + SMC_NLA_STATS_T_RX_RMB_SIZE, /* nest */ + SMC_NLA_STATS_T_TXPLOAD_SIZE, /* nest */ + SMC_NLA_STATS_T_RXPLOAD_SIZE, /* nest */ + SMC_NLA_STATS_T_TX_RMB_STATS, /* nest */ + SMC_NLA_STATS_T_RX_RMB_STATS, /* nest */ + SMC_NLA_STATS_T_CLNT_V1_SUCC, /* u64 */ + SMC_NLA_STATS_T_CLNT_V2_SUCC, /* u64 */ + SMC_NLA_STATS_T_SRV_V1_SUCC, /* u64 */ + SMC_NLA_STATS_T_SRV_V2_SUCC, /* u64 */ + SMC_NLA_STATS_T_SENDPAGE_CNT, /* u64 */ + SMC_NLA_STATS_T_SPLICE_CNT, /* u64 */ + SMC_NLA_STATS_T_CORK_CNT, /* u64 */ + SMC_NLA_STATS_T_NDLY_CNT, /* u64 */ + SMC_NLA_STATS_T_URG_DATA_CNT, /* u64 */ + SMC_NLA_STATS_T_RX_BYTES, /* u64 */ + SMC_NLA_STATS_T_TX_BYTES, /* u64 */ + SMC_NLA_STATS_T_RX_CNT, /* u64 */ + SMC_NLA_STATS_T_TX_CNT, /* u64 */ + __SMC_NLA_STATS_T_MAX, + SMC_NLA_STATS_T_MAX = __SMC_NLA_STATS_T_MAX - 1 +}; + +/* SMC_GEN_STATS attributes */ +enum { + SMC_NLA_STATS_PAD, + SMC_NLA_STATS_SMCD_TECH, /* nest */ + SMC_NLA_STATS_SMCR_TECH, /* nest */ + SMC_NLA_STATS_CLNT_HS_ERR_CNT, /* u64 */ + SMC_NLA_STATS_SRV_HS_ERR_CNT, /* u64 */ + __SMC_NLA_STATS_MAX, + SMC_NLA_STATS_MAX = __SMC_NLA_STATS_MAX - 1 +}; + +/* SMC_GEN_FBACK_STATS attributes */ +enum { + SMC_NLA_FBACK_STATS_PAD, + SMC_NLA_FBACK_STATS_TYPE, /* u8 */ + SMC_NLA_FBACK_STATS_SRV_CNT, /* u64 */ + SMC_NLA_FBACK_STATS_CLNT_CNT, /* u64 */ + SMC_NLA_FBACK_STATS_RSN_CODE, /* u32 */ + SMC_NLA_FBACK_STATS_RSN_CNT, /* u16 */ + __SMC_NLA_FBACK_STATS_MAX, + SMC_NLA_FBACK_STATS_MAX = __SMC_NLA_FBACK_STATS_MAX - 1 +}; #endif /* _UAPI_LINUX_SMC_H */ diff --git a/net/smc/Makefile b/net/smc/Makefile index 79f53cc7d8dcaa7c2f21c9e1be1cc15b1bd72b89..4444145a8ab3dd5c64d3538c03daaf128a9ea906 100644 --- a/net/smc/Makefile +++ b/net/smc/Makefile @@ -2,5 +2,5 @@ obj-$(CONFIG_SMC) += smc.o obj-$(CONFIG_SMC_DIAG) += smc_diag.o smc-y := af_smc.o smc_pnet.o smc_ib.o smc_clc.o smc_core.o smc_wr.o smc_llc.o -smc-y += smc_cdc.o smc_tx.o smc_rx.o smc_close.o smc_ism.o +smc-y += smc_cdc.o smc_tx.o smc_rx.o smc_close.o smc_ism.o smc_netlink.o smc_stats.o smc-$(CONFIG_SYSCTL) += smc_sysctl.o diff --git a/net/smc/af_smc.c b/net/smc/af_smc.c index 6a4a19f57a39b203810d19ade5fa7aa16a33206e..777636bc5e3da1f1111d8d86fb9fbd80a3c5479b 100644 --- a/net/smc/af_smc.c +++ b/net/smc/af_smc.c @@ -45,10 +45,12 @@ #include "smc_ib.h" #include "smc_ism.h" #include "smc_pnet.h" +#include "smc_netlink.h" #include "smc_tx.h" #include "smc_rx.h" #include "smc_close.h" #include "smc_sysctl.h" +#include "smc_stats.h" static DEFINE_MUTEX(smc_server_lgr_pending); /* serialize link group * creation on server @@ -622,6 +624,37 @@ static void smc_link_save_peer_info(struct smc_link *link, link->peer_mtu = clc->r0.qp_mtu; } +static void smc_stat_inc_fback_rsn_cnt(struct smc_sock *smc, + struct smc_stats_fback *fback_arr) +{ + int cnt; + + for (cnt = 0; cnt < SMC_MAX_FBACK_RSN_CNT; cnt++) { + if (fback_arr[cnt].fback_code == smc->fallback_rsn) { + fback_arr[cnt].count++; + break; + } + if (!fback_arr[cnt].fback_code) { + fback_arr[cnt].fback_code = smc->fallback_rsn; + fback_arr[cnt].count++; + break; + } + } +} + +static void smc_stat_fallback(struct smc_sock *smc) +{ + mutex_lock(&smc_stat_fback_rsn); + if (smc->listen_smc) { + smc_stat_inc_fback_rsn_cnt(smc, fback_rsn.srv); + fback_rsn.srv_fback_cnt++; + } else { + smc_stat_inc_fback_rsn_cnt(smc, fback_rsn.clnt); + fback_rsn.clnt_fback_cnt++; + } + mutex_unlock(&smc_stat_fback_rsn); +} + /* must be called under rcu read lock */ static void smc_fback_wakeup_waitqueue(struct smc_sock *smc, void *key) { @@ -747,7 +780,7 @@ static void smc_fback_replace_callbacks(struct smc_sock *smc) write_unlock_bh(&clcsk->sk_callback_lock); } -static int smc_switch_to_fallback(struct smc_sock *smc) +static int smc_switch_to_fallback(struct smc_sock *smc, int reason_code) { int rc = 0; @@ -758,6 +791,8 @@ static int smc_switch_to_fallback(struct smc_sock *smc) } smc->use_fallback = true; + smc->fallback_rsn = reason_code; + smc_stat_fallback(smc); if (smc->sk.sk_socket && smc->sk.sk_socket->file) { smc->clcsock->file = smc->sk.sk_socket->file; smc->clcsock->file->private_data = smc->clcsock; @@ -780,13 +815,12 @@ static int smc_connect_fallback(struct smc_sock *smc, int reason_code) { int rc = 0; - rc = smc_switch_to_fallback(smc); + rc = smc_switch_to_fallback(smc, reason_code); if (rc) { /* fallback fails */ if (smc->sk.sk_state == SMC_INIT) sock_put(&smc->sk); /* passive closing */ return rc; } - smc->fallback_rsn = reason_code; smc_copy_sock_settings_to_clc(smc); smc->connect_nonblock = 0; if (smc->sk.sk_state == SMC_INIT) @@ -801,6 +835,7 @@ static int smc_connect_decline_fallback(struct smc_sock *smc, int reason_code, int rc; if (reason_code < 0) { /* error, fallback is not possible */ + this_cpu_inc(smc_stats->clnt_hshake_err_cnt); if (smc->sk.sk_state == SMC_INIT) sock_put(&smc->sk); /* passive closing */ return reason_code; @@ -808,6 +843,7 @@ static int smc_connect_decline_fallback(struct smc_sock *smc, int reason_code, if (reason_code != SMC_CLC_DECL_PEERDECL) { rc = smc_clc_send_decline(smc, reason_code, version); if (rc < 0) { + this_cpu_inc(smc_stats->clnt_hshake_err_cnt); if (smc->sk.sk_state == SMC_INIT) sock_put(&smc->sk); /* passive closing */ return rc; @@ -933,7 +969,7 @@ static int smc_find_proposal_devices(struct smc_sock *smc, ini->smc_type_v1 = SMC_TYPE_N; } /* else RDMA is supported for this connection */ } - if (smc_ism_v2_capable && smc_find_ism_v2_device_clnt(smc, ini)) + if (smc_ism_is_v2_capable() && smc_find_ism_v2_device_clnt(smc, ini)) ini->smc_type_v2 = SMC_TYPE_N; /* if neither ISM nor RDMA are supported, fallback */ @@ -1191,7 +1227,7 @@ static int smc_connect_check_aclc(struct smc_init_info *ini, /* perform steps before actually connecting */ static int __smc_connect(struct smc_sock *smc) { - u8 version = smc_ism_v2_capable ? SMC_V2 : SMC_V1; + u8 version = smc_ism_is_v2_capable() ? SMC_V2 : SMC_V1; struct smc_clc_msg_accept_confirm_v2 *aclc2; struct smc_clc_msg_accept_confirm *aclc; struct smc_init_info *ini = NULL; @@ -1216,9 +1252,9 @@ static int __smc_connect(struct smc_sock *smc) version); ini->smcd_version = SMC_V1; - ini->smcd_version |= smc_ism_v2_capable ? SMC_V2 : 0; + ini->smcd_version |= smc_ism_is_v2_capable() ? SMC_V2 : 0; ini->smc_type_v1 = SMC_TYPE_B; - ini->smc_type_v2 = smc_ism_v2_capable ? SMC_TYPE_D : SMC_TYPE_N; + ini->smc_type_v2 = smc_ism_is_v2_capable() ? SMC_TYPE_D : SMC_TYPE_N; /* get vlan id from IP device */ if (smc_vlan_by_tcpsk(smc->clcsock, ini)) { @@ -1262,6 +1298,7 @@ static int __smc_connect(struct smc_sock *smc) if (rc) goto vlan_cleanup; + SMC_STAT_CLNT_SUCC_INC(aclc); smc_connect_ism_vlan_cleanup(smc, ini); kfree(buf); kfree(ini); @@ -1637,6 +1674,7 @@ static void smc_listen_out_err(struct smc_sock *new_smc) { struct sock *newsmcsk = &new_smc->sk; + this_cpu_inc(smc_stats->srv_hshake_err_cnt); if (newsmcsk->sk_state == SMC_INIT) sock_put(&new_smc->sk); /* passive closing */ newsmcsk->sk_state = SMC_CLOSED; @@ -1654,12 +1692,11 @@ static void smc_listen_decline(struct smc_sock *new_smc, int reason_code, else smc_conn_free(&new_smc->conn); if (reason_code < 0 || - smc_switch_to_fallback(new_smc)) { + smc_switch_to_fallback(new_smc, reason_code)) { /* error, no fallback possible */ smc_listen_out_err(new_smc); return; } - new_smc->fallback_rsn = reason_code; if (reason_code && reason_code != SMC_CLC_DECL_PEERDECL) { if (smc_clc_send_decline(new_smc, reason_code, version) < 0) { smc_listen_out_err(new_smc); @@ -1683,7 +1720,7 @@ static int smc_listen_v2_check(struct smc_sock *new_smc, if (pclc->hdr.version > SMC_V1) ini->smcd_version |= ini->smc_type_v2 != SMC_TYPE_N ? SMC_V2 : 0; - if (!smc_ism_v2_capable) { + if (!smc_ism_is_v2_capable()) { ini->smcd_version &= ~SMC_V2; goto out; } @@ -2001,7 +2038,7 @@ static void smc_listen_work(struct work_struct *work) { struct smc_sock *new_smc = container_of(work, struct smc_sock, smc_listen_work); - u8 version = smc_ism_v2_capable ? SMC_V2 : SMC_V1; + u8 version = smc_ism_is_v2_capable() ? SMC_V2 : SMC_V1; struct socket *newclcsock = new_smc->clcsock; struct smc_clc_msg_accept_confirm *cclc; struct smc_clc_msg_proposal_area *buf; @@ -2019,13 +2056,11 @@ static void smc_listen_work(struct work_struct *work) /* check if peer is smc capable */ if (!tcp_sk(newclcsock->sk)->syn_smc) { - rc = smc_switch_to_fallback(new_smc); - if (rc) { + rc = smc_switch_to_fallback(new_smc, SMC_CLC_DECL_PEERNOSMC); + if (rc) smc_listen_out_err(new_smc); - } else { - new_smc->fallback_rsn = SMC_CLC_DECL_PEERNOSMC; + else smc_listen_out_connected(new_smc); - } return; } @@ -2102,6 +2137,7 @@ static void smc_listen_work(struct work_struct *work) } smc_conn_save_peer_info(new_smc, cclc); smc_listen_out_connected(new_smc); + SMC_STAT_SERV_SUCC_INC(ini); goto out_free; out_unlock: @@ -2319,10 +2355,9 @@ static int smc_sendmsg(struct socket *sock, struct msghdr *msg, size_t len) if (msg->msg_flags & MSG_FASTOPEN) { /* not connected yet, fallback */ if (sk->sk_state == SMC_INIT && !smc->connect_nonblock) { - rc = smc_switch_to_fallback(smc); + rc = smc_switch_to_fallback(smc, SMC_CLC_DECL_OPTUNSUPP); if (rc) goto out; - smc->fallback_rsn = SMC_CLC_DECL_OPTUNSUPP; } else { rc = -EINVAL; goto out; @@ -2334,10 +2369,12 @@ static int smc_sendmsg(struct socket *sock, struct msghdr *msg, size_t len) goto out; } - if (smc->use_fallback) + if (smc->use_fallback) { rc = smc->clcsock->ops->sendmsg(smc->clcsock, msg, len); - else + } else { rc = smc_tx_sendmsg(smc, msg, len); + SMC_STAT_TX_PAYLOAD(smc, len, rc); + } out: release_sock(sk); return rc; @@ -2372,6 +2409,7 @@ static int smc_recvmsg(struct socket *sock, struct msghdr *msg, size_t len, } else { msg->msg_namelen = 0; rc = smc_rx_recvmsg(smc, msg, NULL, len, flags); + SMC_STAT_RX_PAYLOAD(smc, rc, rc); } out: @@ -2561,9 +2599,7 @@ static int smc_setsockopt(struct socket *sock, int level, int optname, case TCP_FASTOPEN_NO_COOKIE: /* option not supported by SMC */ if (sk->sk_state == SMC_INIT && !smc->connect_nonblock) { - rc = smc_switch_to_fallback(smc); - if (!rc) - smc->fallback_rsn = SMC_CLC_DECL_OPTUNSUPP; + rc = smc_switch_to_fallback(smc, SMC_CLC_DECL_OPTUNSUPP); } else { rc = -EINVAL; } @@ -2573,6 +2609,7 @@ static int smc_setsockopt(struct socket *sock, int level, int optname, sk->sk_state != SMC_LISTEN && sk->sk_state != SMC_CLOSED) { if (val) { + SMC_STAT_INC(!smc->conn.lnk, ndly_cnt); smc_tx_pending(&smc->conn); cancel_delayed_work(&smc->conn.tx_work); } @@ -2583,6 +2620,7 @@ static int smc_setsockopt(struct socket *sock, int level, int optname, sk->sk_state != SMC_LISTEN && sk->sk_state != SMC_CLOSED) { if (!val) { + SMC_STAT_INC(!smc->conn.lnk, cork_cnt); smc_tx_pending(&smc->conn); cancel_delayed_work(&smc->conn.tx_work); } @@ -2718,10 +2756,11 @@ static ssize_t smc_sendpage(struct socket *sock, struct page *page, goto out; } release_sock(sk); - if (smc->use_fallback) + if (smc->use_fallback) { rc = kernel_sendpage(smc->clcsock, page, offset, size, flags); - else { + } else { + SMC_STAT_INC(!smc->conn.lnk, sendpage_cnt); lock_sock(sk); rc = smc_tx_sendpage(smc, page, offset, size, flags); release_sock(sk); @@ -2774,6 +2813,7 @@ static ssize_t smc_splice_read(struct socket *sock, loff_t *ppos, flags = MSG_DONTWAIT; else flags = 0; + SMC_STAT_INC(!smc->conn.lnk, splice_cnt); rc = smc_rx_recvmsg(smc, NULL, pipe, len, flags); } out: @@ -2886,10 +2926,14 @@ static int __init smc_init(void) smc_ism_init(); smc_clc_init(); - rc = smc_pnet_init(); + rc = smc_nl_init(); if (rc) goto out_pernet_subsys; + rc = smc_pnet_init(); + if (rc) + goto out_nl; + rc = -ENOMEM; smc_tcp_ls_wq = alloc_workqueue("smc_tcp_ls_wq", WQ_UNBOUND | WQ_HIGHPRI, 0); @@ -2903,10 +2947,16 @@ static int __init smc_init(void) if (!smc_close_wq) goto out_alloc_hs_wq; + rc = smc_stats_init(); + if (rc) { + pr_err("%s: smc_stats_init fails with %d\n", __func__, rc); + goto out_alloc_wqs; + } + rc = smc_core_init(); if (rc) { pr_err("%s: smc_core_init fails with %d\n", __func__, rc); - goto out_alloc_wqs; + goto out_smc_stat; } rc = smc_llc_init(); @@ -2958,6 +3008,8 @@ static int __init smc_init(void) proto_unregister(&smc_proto); out_core: smc_core_exit(); +out_smc_stat: + smc_stats_exit(); out_alloc_wqs: destroy_workqueue(smc_close_wq); out_alloc_hs_wq: @@ -2966,6 +3018,8 @@ static int __init smc_init(void) destroy_workqueue(smc_tcp_ls_wq); out_pnet: smc_pnet_exit(); +out_nl: + smc_nl_exit(); out_pernet_subsys: unregister_pernet_subsys(&smc_net_ops); @@ -2981,9 +3035,11 @@ static void __exit smc_exit(void) destroy_workqueue(smc_close_wq); destroy_workqueue(smc_tcp_ls_wq); destroy_workqueue(smc_hs_wq); + smc_stats_exit(); proto_unregister(&smc_proto6); proto_unregister(&smc_proto); smc_pnet_exit(); + smc_nl_exit(); unregister_pernet_subsys(&smc_net_ops); rcu_barrier(); } diff --git a/net/smc/smc_clc.c b/net/smc/smc_clc.c index c9450ab0e23b2ccb65fb83d3342f668bc5aa7ce4..838470f897d0404217f5c8df6694f75c6c5bcbea 100644 --- a/net/smc/smc_clc.c +++ b/net/smc/smc_clc.c @@ -775,6 +775,11 @@ int smc_clc_send_accept(struct smc_sock *new_smc, bool srv_first_contact, return len > 0 ? 0 : len; } +void smc_clc_get_hostname(u8 **host) +{ + *host = &smc_hostname[0]; +} + void __init smc_clc_init(void) { struct new_utsname *u; diff --git a/net/smc/smc_clc.h b/net/smc/smc_clc.h index 8992949900e9f89db98326e4dde3041b644d20db..49291909ffca0902f1b95ed261809096944ace43 100644 --- a/net/smc/smc_clc.h +++ b/net/smc/smc_clc.h @@ -329,5 +329,6 @@ int smc_clc_send_confirm(struct smc_sock *smc, bool clnt_first_contact, int smc_clc_send_accept(struct smc_sock *smc, bool srv_first_contact, u8 version); void smc_clc_init(void) __init; +void smc_clc_get_hostname(u8 **host); #endif diff --git a/net/smc/smc_core.c b/net/smc/smc_core.c index 4bea24ac974f60a83294fe4148ff92824dbeb9ca..de0fda76a49b498cf9ead0188a9c0e960dec080e 100644 --- a/net/smc/smc_core.c +++ b/net/smc/smc_core.c @@ -16,6 +16,8 @@ #include #include #include +#include +#include #include #include #include @@ -31,12 +33,14 @@ #include "smc_close.h" #include "smc_ism.h" #include "smc_sysctl.h" +#include "smc_netlink.h" +#include "smc_stats.h" #define SMC_LGR_NUM_INCR 256 #define SMC_LGR_FREE_DELAY_SERV (600 * HZ) #define SMC_LGR_FREE_DELAY_CLNT (SMC_LGR_FREE_DELAY_SERV + 10 * HZ) -static struct smc_lgr_list smc_lgr_list = { /* established link groups */ +struct smc_lgr_list smc_lgr_list = { /* established link groups */ .lock = __SPIN_LOCK_UNLOCKED(smc_lgr_list.lock), .list = LIST_HEAD_INIT(smc_lgr_list.list), .num = 0, @@ -64,6 +68,16 @@ static inline struct list_head *smc_lgr_list_head(struct smc_link_group *lgr, return &smc_lgr_list.list; } +static void smc_ibdev_cnt_inc(struct smc_link *lnk) +{ + atomic_inc(&lnk->smcibdev->lnk_cnt_by_port[lnk->ibport - 1]); +} + +static void smc_ibdev_cnt_dec(struct smc_link *lnk) +{ + atomic_dec(&lnk->smcibdev->lnk_cnt_by_port[lnk->ibport - 1]); +} + static void smc_lgr_schedule_free_work(struct smc_link_group *lgr) { /* client link group creation always follows the server link group @@ -206,6 +220,367 @@ static void smc_lgr_unregister_conn(struct smc_connection *conn) write_unlock_bh(&lgr->conns_lock); } +int smc_nl_get_sys_info(struct sk_buff *skb, struct netlink_callback *cb) +{ + struct smc_nl_dmp_ctx *cb_ctx = smc_nl_dmp_ctx(cb); + char hostname[SMC_MAX_HOSTNAME_LEN + 1]; + char smc_seid[SMC_MAX_EID_LEN + 1]; + struct smcd_dev *smcd_dev; + struct nlattr *attrs; + u8 *seid = NULL; + u8 *host = NULL; + void *nlh; + + nlh = genlmsg_put(skb, NETLINK_CB(cb->skb).portid, cb->nlh->nlmsg_seq, + &smc_gen_nl_family, NLM_F_MULTI, + SMC_NETLINK_GET_SYS_INFO); + if (!nlh) + goto errmsg; + if (cb_ctx->pos[0]) + goto errout; + attrs = nla_nest_start(skb, SMC_GEN_SYS_INFO); + if (!attrs) + goto errout; + if (nla_put_u8(skb, SMC_NLA_SYS_VER, SMC_V2)) + goto errattr; + if (nla_put_u8(skb, SMC_NLA_SYS_REL, SMC_RELEASE)) + goto errattr; + if (nla_put_u8(skb, SMC_NLA_SYS_IS_ISM_V2, smc_ism_is_v2_capable())) + goto errattr; + smc_clc_get_hostname(&host); + if (host) { + memcpy(hostname, host, SMC_MAX_HOSTNAME_LEN); + hostname[SMC_MAX_HOSTNAME_LEN] = 0; + if (nla_put_string(skb, SMC_NLA_SYS_LOCAL_HOST, hostname)) + goto errattr; + } + mutex_lock(&smcd_dev_list.mutex); + smcd_dev = list_first_entry_or_null(&smcd_dev_list.list, + struct smcd_dev, list); + if (smcd_dev) + smc_ism_get_system_eid(smcd_dev, &seid); + mutex_unlock(&smcd_dev_list.mutex); + if (seid && smc_ism_is_v2_capable()) { + memcpy(smc_seid, seid, SMC_MAX_EID_LEN); + smc_seid[SMC_MAX_EID_LEN] = 0; + if (nla_put_string(skb, SMC_NLA_SYS_SEID, smc_seid)) + goto errattr; + } + nla_nest_end(skb, attrs); + genlmsg_end(skb, nlh); + cb_ctx->pos[0] = 1; + return skb->len; + +errattr: + nla_nest_cancel(skb, attrs); +errout: + genlmsg_cancel(skb, nlh); +errmsg: + return skb->len; +} + +static int smc_nl_fill_lgr(struct smc_link_group *lgr, + struct sk_buff *skb, + struct netlink_callback *cb) +{ + char smc_target[SMC_MAX_PNETID_LEN + 1]; + struct nlattr *attrs; + + attrs = nla_nest_start(skb, SMC_GEN_LGR_SMCR); + if (!attrs) + goto errout; + + if (nla_put_u32(skb, SMC_NLA_LGR_R_ID, *((u32 *)&lgr->id))) + goto errattr; + if (nla_put_u32(skb, SMC_NLA_LGR_R_CONNS_NUM, lgr->conns_num)) + goto errattr; + if (nla_put_u8(skb, SMC_NLA_LGR_R_ROLE, lgr->role)) + goto errattr; + if (nla_put_u8(skb, SMC_NLA_LGR_R_TYPE, lgr->type)) + goto errattr; + if (nla_put_u8(skb, SMC_NLA_LGR_R_VLAN_ID, lgr->vlan_id)) + goto errattr; + memcpy(smc_target, lgr->pnet_id, SMC_MAX_PNETID_LEN); + smc_target[SMC_MAX_PNETID_LEN] = 0; + if (nla_put_string(skb, SMC_NLA_LGR_R_PNETID, smc_target)) + goto errattr; + + nla_nest_end(skb, attrs); + return 0; +errattr: + nla_nest_cancel(skb, attrs); +errout: + return -EMSGSIZE; +} + +static int smc_nl_fill_lgr_link(struct smc_link_group *lgr, + struct smc_link *link, + struct sk_buff *skb, + struct netlink_callback *cb) +{ + char smc_ibname[IB_DEVICE_NAME_MAX]; + u8 smc_gid_target[41]; + struct nlattr *attrs; + u32 link_uid = 0; + void *nlh; + + nlh = genlmsg_put(skb, NETLINK_CB(cb->skb).portid, cb->nlh->nlmsg_seq, + &smc_gen_nl_family, NLM_F_MULTI, + SMC_NETLINK_GET_LINK_SMCR); + if (!nlh) + goto errmsg; + + attrs = nla_nest_start(skb, SMC_GEN_LINK_SMCR); + if (!attrs) + goto errout; + + if (nla_put_u8(skb, SMC_NLA_LINK_ID, link->link_id)) + goto errattr; + if (nla_put_u32(skb, SMC_NLA_LINK_STATE, link->state)) + goto errattr; + if (nla_put_u32(skb, SMC_NLA_LINK_CONN_CNT, + atomic_read(&link->conn_cnt))) + goto errattr; + if (nla_put_u8(skb, SMC_NLA_LINK_IB_PORT, link->ibport)) + goto errattr; + if (nla_put_u32(skb, SMC_NLA_LINK_NET_DEV, link->ndev_ifidx)) + goto errattr; + snprintf(smc_ibname, sizeof(smc_ibname), "%s", link->ibname); + if (nla_put_string(skb, SMC_NLA_LINK_IB_DEV, smc_ibname)) + goto errattr; + memcpy(&link_uid, link->link_uid, sizeof(link_uid)); + if (nla_put_u32(skb, SMC_NLA_LINK_UID, link_uid)) + goto errattr; + memcpy(&link_uid, link->peer_link_uid, sizeof(link_uid)); + if (nla_put_u32(skb, SMC_NLA_LINK_PEER_UID, link_uid)) + goto errattr; + memset(smc_gid_target, 0, sizeof(smc_gid_target)); + smc_gid_be16_convert(smc_gid_target, link->gid); + if (nla_put_string(skb, SMC_NLA_LINK_GID, smc_gid_target)) + goto errattr; + memset(smc_gid_target, 0, sizeof(smc_gid_target)); + smc_gid_be16_convert(smc_gid_target, link->peer_gid); + if (nla_put_string(skb, SMC_NLA_LINK_PEER_GID, smc_gid_target)) + goto errattr; + + nla_nest_end(skb, attrs); + genlmsg_end(skb, nlh); + return 0; +errattr: + nla_nest_cancel(skb, attrs); +errout: + genlmsg_cancel(skb, nlh); +errmsg: + return -EMSGSIZE; +} + +static int smc_nl_handle_lgr(struct smc_link_group *lgr, + struct sk_buff *skb, + struct netlink_callback *cb, + bool list_links) +{ + void *nlh; + int i; + + nlh = genlmsg_put(skb, NETLINK_CB(cb->skb).portid, cb->nlh->nlmsg_seq, + &smc_gen_nl_family, NLM_F_MULTI, + SMC_NETLINK_GET_LGR_SMCR); + if (!nlh) + goto errmsg; + if (smc_nl_fill_lgr(lgr, skb, cb)) + goto errout; + + genlmsg_end(skb, nlh); + if (!list_links) + goto out; + for (i = 0; i < SMC_LINKS_PER_LGR_MAX; i++) { + if (!smc_link_usable(&lgr->lnk[i])) + continue; + if (smc_nl_fill_lgr_link(lgr, &lgr->lnk[i], skb, cb)) + goto errout; + } +out: + return 0; + +errout: + genlmsg_cancel(skb, nlh); +errmsg: + return -EMSGSIZE; +} + +static void smc_nl_fill_lgr_list(struct smc_lgr_list *smc_lgr, + struct sk_buff *skb, + struct netlink_callback *cb, + bool list_links) +{ + struct smc_nl_dmp_ctx *cb_ctx = smc_nl_dmp_ctx(cb); + struct smc_link_group *lgr; + int snum = cb_ctx->pos[0]; + int num = 0; + + spin_lock_bh(&smc_lgr->lock); + list_for_each_entry(lgr, &smc_lgr->list, list) { + if (num < snum) + goto next; + if (smc_nl_handle_lgr(lgr, skb, cb, list_links)) + goto errout; +next: + num++; + } +errout: + spin_unlock_bh(&smc_lgr->lock); + cb_ctx->pos[0] = num; +} + +static int smc_nl_fill_smcd_lgr(struct smc_link_group *lgr, + struct sk_buff *skb, + struct netlink_callback *cb) +{ + char smc_host[SMC_MAX_HOSTNAME_LEN + 1]; + char smc_pnet[SMC_MAX_PNETID_LEN + 1]; + char smc_eid[SMC_MAX_EID_LEN + 1]; + struct nlattr *v2_attrs; + struct nlattr *attrs; + void *nlh; + + nlh = genlmsg_put(skb, NETLINK_CB(cb->skb).portid, cb->nlh->nlmsg_seq, + &smc_gen_nl_family, NLM_F_MULTI, + SMC_NETLINK_GET_LGR_SMCD); + if (!nlh) + goto errmsg; + + attrs = nla_nest_start(skb, SMC_GEN_LGR_SMCD); + if (!attrs) + goto errout; + + if (nla_put_u32(skb, SMC_NLA_LGR_D_ID, *((u32 *)&lgr->id))) + goto errattr; + if (nla_put_u64_64bit(skb, SMC_NLA_LGR_D_GID, lgr->smcd->local_gid, + SMC_NLA_LGR_D_PAD)) + goto errattr; + if (nla_put_u64_64bit(skb, SMC_NLA_LGR_D_PEER_GID, lgr->peer_gid, + SMC_NLA_LGR_D_PAD)) + goto errattr; + if (nla_put_u8(skb, SMC_NLA_LGR_D_VLAN_ID, lgr->vlan_id)) + goto errattr; + if (nla_put_u32(skb, SMC_NLA_LGR_D_CONNS_NUM, lgr->conns_num)) + goto errattr; + if (nla_put_u32(skb, SMC_NLA_LGR_D_CHID, smc_ism_get_chid(lgr->smcd))) + goto errattr; + memcpy(smc_pnet, lgr->smcd->pnetid, SMC_MAX_PNETID_LEN); + smc_pnet[SMC_MAX_PNETID_LEN] = 0; + if (nla_put_string(skb, SMC_NLA_LGR_D_PNETID, smc_pnet)) + goto errattr; + + v2_attrs = nla_nest_start(skb, SMC_NLA_LGR_V2); + if (!v2_attrs) + goto errattr; + if (nla_put_u8(skb, SMC_NLA_LGR_V2_VER, lgr->smc_version)) + goto errv2attr; + if (nla_put_u8(skb, SMC_NLA_LGR_V2_REL, lgr->peer_smc_release)) + goto errv2attr; + if (nla_put_u8(skb, SMC_NLA_LGR_V2_OS, lgr->peer_os)) + goto errv2attr; + memcpy(smc_host, lgr->peer_hostname, SMC_MAX_HOSTNAME_LEN); + smc_host[SMC_MAX_HOSTNAME_LEN] = 0; + if (nla_put_string(skb, SMC_NLA_LGR_V2_PEER_HOST, smc_host)) + goto errv2attr; + memcpy(smc_eid, lgr->negotiated_eid, SMC_MAX_EID_LEN); + smc_eid[SMC_MAX_EID_LEN] = 0; + if (nla_put_string(skb, SMC_NLA_LGR_V2_NEG_EID, smc_eid)) + goto errv2attr; + + nla_nest_end(skb, v2_attrs); + nla_nest_end(skb, attrs); + genlmsg_end(skb, nlh); + return 0; + +errv2attr: + nla_nest_cancel(skb, v2_attrs); +errattr: + nla_nest_cancel(skb, attrs); +errout: + genlmsg_cancel(skb, nlh); +errmsg: + return -EMSGSIZE; +} + +static int smc_nl_handle_smcd_lgr(struct smcd_dev *dev, + struct sk_buff *skb, + struct netlink_callback *cb) +{ + struct smc_nl_dmp_ctx *cb_ctx = smc_nl_dmp_ctx(cb); + struct smc_link_group *lgr; + int snum = cb_ctx->pos[1]; + int rc = 0, num = 0; + + spin_lock_bh(&dev->lgr_lock); + list_for_each_entry(lgr, &dev->lgr_list, list) { + if (!lgr->is_smcd) + continue; + if (num < snum) + goto next; + rc = smc_nl_fill_smcd_lgr(lgr, skb, cb); + if (rc) + goto errout; +next: + num++; + } +errout: + spin_unlock_bh(&dev->lgr_lock); + cb_ctx->pos[1] = num; + return rc; +} + +static int smc_nl_fill_smcd_dev(struct smcd_dev_list *dev_list, + struct sk_buff *skb, + struct netlink_callback *cb) +{ + struct smc_nl_dmp_ctx *cb_ctx = smc_nl_dmp_ctx(cb); + struct smcd_dev *smcd_dev; + int snum = cb_ctx->pos[0]; + int rc = 0, num = 0; + + mutex_lock(&dev_list->mutex); + list_for_each_entry(smcd_dev, &dev_list->list, list) { + if (list_empty(&smcd_dev->lgr_list)) + continue; + if (num < snum) + goto next; + rc = smc_nl_handle_smcd_lgr(smcd_dev, skb, cb); + if (rc) + goto errout; +next: + num++; + } +errout: + mutex_unlock(&dev_list->mutex); + cb_ctx->pos[0] = num; + return rc; +} + +int smcr_nl_get_lgr(struct sk_buff *skb, struct netlink_callback *cb) +{ + bool list_links = false; + + smc_nl_fill_lgr_list(&smc_lgr_list, skb, cb, list_links); + return skb->len; +} + +int smcr_nl_get_link(struct sk_buff *skb, struct netlink_callback *cb) +{ + bool list_links = true; + + smc_nl_fill_lgr_list(&smc_lgr_list, skb, cb, list_links); + return skb->len; +} + +int smcd_nl_get_lgr(struct sk_buff *skb, struct netlink_callback *cb) +{ + smc_nl_fill_smcd_dev(&smcd_dev_list, skb, cb); + return skb->len; +} + void smc_lgr_cleanup_early(struct smc_connection *conn) { struct smc_link_group *lgr = conn->lgr; @@ -305,6 +680,15 @@ static u8 smcr_next_link_id(struct smc_link_group *lgr) return link_id; } +static void smcr_copy_dev_info_to_link(struct smc_link *link) +{ + struct smc_ib_device *smcibdev = link->smcibdev; + + snprintf(link->ibname, sizeof(link->ibname), "%s", + smcibdev->ibdev->name); + link->ndev_ifidx = smcibdev->ndev_ifidx[link->ibport - 1]; +} + int smcr_link_init(struct smc_link_group *lgr, struct smc_link *lnk, u8 link_idx, struct smc_init_info *ini) { @@ -321,6 +705,8 @@ int smcr_link_init(struct smc_link_group *lgr, struct smc_link *lnk, lnk->link_idx = link_idx; lnk->smcibdev = ini->ib_dev; lnk->ibport = ini->ib_port; + smc_ibdev_cnt_inc(lnk); + smcr_copy_dev_info_to_link(lnk); lnk->path_mtu = ini->ib_dev->pattr[ini->ib_port - 1].active_mtu; atomic_set(&lnk->conn_cnt, 0); smc_llc_link_set_uid(lnk); @@ -364,6 +750,7 @@ int smcr_link_init(struct smc_link_group *lgr, struct smc_link *lnk, clear_llc_lnk: smc_llc_link_clear(lnk, false); out: + smc_ibdev_cnt_dec(lnk); put_device(&ini->ib_dev->ibdev->dev); memset(lnk, 0, sizeof(struct smc_link)); lnk->state = SMC_LNK_UNUSED; @@ -785,6 +1172,7 @@ void __smcr_link_clear(struct smc_link *lnk) smc_ib_destroy_queue_pair(lnk); smc_ib_dealloc_protection_domain(lnk); smc_wr_free_link_mem(lnk); + smc_ibdev_cnt_dec(lnk); smc_lgr_put(lnk->lgr); /* lgr_hold in smcr_link_init() */ memset(lnk, 0, sizeof(struct smc_link)); lnk->state = SMC_LNK_UNUSED; @@ -1841,6 +2229,7 @@ static int __smc_buf_create(struct smc_sock *smc, bool is_smcd, bool is_rmb) struct smc_link_group *lgr = conn->lgr; struct list_head *buf_list; int bufsize, bufsize_comp; + bool is_dgraded = false; struct rw_semaphore *lock; /* lock buffer list */ if (is_rmb) @@ -1864,6 +2253,8 @@ static int __smc_buf_create(struct smc_sock *smc, bool is_smcd, bool is_rmb) /* check for reusable slot in the link group */ buf_desc = smc_buf_get_slot(bufsize, lock, buf_list); if (buf_desc) { + SMC_STAT_RMB_SIZE(is_smcd, is_rmb, bufsize); + SMC_STAT_BUF_REUSE(is_smcd, is_rmb); break; /* found reusable slot */ } @@ -1874,9 +2265,16 @@ static int __smc_buf_create(struct smc_sock *smc, bool is_smcd, bool is_rmb) if (PTR_ERR(buf_desc) == -ENOMEM) break; - if (IS_ERR(buf_desc)) + if (IS_ERR(buf_desc)) { + if (!is_dgraded) { + is_dgraded = true; + SMC_STAT_RMB_DOWNGRADED(is_smcd, is_rmb); + } continue; + } + SMC_STAT_RMB_ALLOC(is_smcd, is_rmb); + SMC_STAT_RMB_SIZE(is_smcd, is_rmb, bufsize); buf_desc->used = 1; down_write(lock); list_add(&buf_desc->list, buf_list); diff --git a/net/smc/smc_core.h b/net/smc/smc_core.h index ab410fe62fac157b75061de9f809e35dd0e0769a..739ead4360b0eeaa007ed41a11d9581151d89fbd 100644 --- a/net/smc/smc_core.h +++ b/net/smc/smc_core.h @@ -13,7 +13,10 @@ #define _SMC_CORE_H #include +#include +#include #include +#include #include "smc.h" #include "smc_ib.h" @@ -128,6 +131,8 @@ struct smc_link { refcount_t refcnt; /* link reference count */ struct smc_link_group *lgr; /* parent link group */ struct work_struct link_down_wrk; /* wrk to bring link down */ + char ibname[IB_DEVICE_NAME_MAX]; /* ib device name */ + int ndev_ifidx; /* network device ifindex */ enum smc_link_state state; /* state of link */ struct delayed_work llc_testlink_wrk; /* testlink worker */ @@ -387,6 +392,45 @@ static inline bool smc_link_active(struct smc_link *lnk) return lnk->state == SMC_LNK_ACTIVE; } +static inline void smc_gid_be16_convert(__u8 *buf, u8 *gid_raw) +{ + sprintf(buf, "%04x:%04x:%04x:%04x:%04x:%04x:%04x:%04x", + be16_to_cpu(((__be16 *)gid_raw)[0]), + be16_to_cpu(((__be16 *)gid_raw)[1]), + be16_to_cpu(((__be16 *)gid_raw)[2]), + be16_to_cpu(((__be16 *)gid_raw)[3]), + be16_to_cpu(((__be16 *)gid_raw)[4]), + be16_to_cpu(((__be16 *)gid_raw)[5]), + be16_to_cpu(((__be16 *)gid_raw)[6]), + be16_to_cpu(((__be16 *)gid_raw)[7])); +} + +struct smc_pci_dev { + __u32 pci_fid; + __u16 pci_pchid; + __u16 pci_vendor; + __u16 pci_device; + __u8 pci_id[SMC_PCI_ID_STR_LEN]; +}; + +static inline void smc_set_pci_values(struct pci_dev *pci_dev, + struct smc_pci_dev *smc_dev) +{ + smc_dev->pci_vendor = pci_dev->vendor; + smc_dev->pci_device = pci_dev->device; + snprintf(smc_dev->pci_id, sizeof(smc_dev->pci_id), "%s", + pci_name(pci_dev)); +#if IS_ENABLED(CONFIG_S390) + { /* Set s390 specific PCI information */ + struct zpci_dev *zdev; + + zdev = to_zpci(pci_dev); + smc_dev->pci_fid = zdev->fid; + smc_dev->pci_pchid = zdev->pchid; + } +#endif +} + struct smc_sock; struct smc_clc_msg_accept_confirm; struct smc_clc_msg_local; @@ -438,6 +482,10 @@ struct smc_link *smc_switch_conns(struct smc_link_group *lgr, struct smc_link *from_lnk, bool is_dev_err); void smcr_link_down_cond(struct smc_link *lnk); void smcr_link_down_cond_sched(struct smc_link *lnk); +int smc_nl_get_sys_info(struct sk_buff *skb, struct netlink_callback *cb); +int smcr_nl_get_lgr(struct sk_buff *skb, struct netlink_callback *cb); +int smcr_nl_get_link(struct sk_buff *skb, struct netlink_callback *cb); +int smcd_nl_get_lgr(struct sk_buff *skb, struct netlink_callback *cb); static inline struct smc_link_group *smc_get_lgr(struct smc_link *link) { diff --git a/net/smc/smc_diag.c b/net/smc/smc_diag.c index f15fca59b4b2654251a375bc522d88a12b450ccc..4dd4f9e694da2f71298fbcbe95de33c98a393cef 100644 --- a/net/smc/smc_diag.c +++ b/net/smc/smc_diag.c @@ -31,19 +31,6 @@ static struct smc_diag_dump_ctx *smc_dump_context(struct netlink_callback *cb) return (struct smc_diag_dump_ctx *)cb->ctx; } -static void smc_gid_be16_convert(__u8 *buf, u8 *gid_raw) -{ - sprintf(buf, "%04x:%04x:%04x:%04x:%04x:%04x:%04x:%04x", - be16_to_cpu(((__be16 *)gid_raw)[0]), - be16_to_cpu(((__be16 *)gid_raw)[1]), - be16_to_cpu(((__be16 *)gid_raw)[2]), - be16_to_cpu(((__be16 *)gid_raw)[3]), - be16_to_cpu(((__be16 *)gid_raw)[4]), - be16_to_cpu(((__be16 *)gid_raw)[5]), - be16_to_cpu(((__be16 *)gid_raw)[6]), - be16_to_cpu(((__be16 *)gid_raw)[7])); -} - static void smc_diag_msg_common_fill(struct smc_diag_msg *r, struct sock *sk) { struct smc_sock *smc = smc_sk(sk); diff --git a/net/smc/smc_ib.c b/net/smc/smc_ib.c index 1413665f31153a3fd28cf7464708c2f43e85cfee..2e0d7aeefc41b754a54dbda481389f698b787f10 100644 --- a/net/smc/smc_ib.c +++ b/net/smc/smc_ib.c @@ -25,6 +25,7 @@ #include "smc_core.h" #include "smc_wr.h" #include "smc.h" +#include "smc_netlink.h" #define SMC_MAX_CQE 32766 /* max. # of completion queue elements */ @@ -326,6 +327,171 @@ int smc_ib_create_protection_domain(struct smc_link *lnk) return rc; } +static bool smcr_diag_is_dev_critical(struct smc_lgr_list *smc_lgr, + struct smc_ib_device *smcibdev) +{ + struct smc_link_group *lgr; + bool rc = false; + int i; + + spin_lock_bh(&smc_lgr->lock); + list_for_each_entry(lgr, &smc_lgr->list, list) { + if (lgr->is_smcd) + continue; + for (i = 0; i < SMC_LINKS_PER_LGR_MAX; i++) { + if (lgr->lnk[i].state == SMC_LNK_UNUSED || + lgr->lnk[i].smcibdev != smcibdev) + continue; + if (lgr->type == SMC_LGR_SINGLE || + lgr->type == SMC_LGR_ASYMMETRIC_LOCAL) { + rc = true; + goto out; + } + } + } +out: + spin_unlock_bh(&smc_lgr->lock); + return rc; +} + +static int smc_nl_handle_dev_port(struct sk_buff *skb, + struct ib_device *ibdev, + struct smc_ib_device *smcibdev, + int port) +{ + char smc_pnet[SMC_MAX_PNETID_LEN + 1]; + struct nlattr *port_attrs; + unsigned char port_state; + int lnk_count = 0; + + port_attrs = nla_nest_start(skb, SMC_NLA_DEV_PORT + port); + if (!port_attrs) + goto errout; + + if (nla_put_u8(skb, SMC_NLA_DEV_PORT_PNET_USR, + smcibdev->pnetid_by_user[port])) + goto errattr; + memcpy(smc_pnet, &smcibdev->pnetid[port], SMC_MAX_PNETID_LEN); + smc_pnet[SMC_MAX_PNETID_LEN] = 0; + if (nla_put_string(skb, SMC_NLA_DEV_PORT_PNETID, smc_pnet)) + goto errattr; + if (nla_put_u32(skb, SMC_NLA_DEV_PORT_NETDEV, + smcibdev->ndev_ifidx[port])) + goto errattr; + if (nla_put_u8(skb, SMC_NLA_DEV_PORT_VALID, 1)) + goto errattr; + port_state = smc_ib_port_active(smcibdev, port + 1); + if (nla_put_u8(skb, SMC_NLA_DEV_PORT_STATE, port_state)) + goto errattr; + lnk_count = atomic_read(&smcibdev->lnk_cnt_by_port[port]); + if (nla_put_u32(skb, SMC_NLA_DEV_PORT_LNK_CNT, lnk_count)) + goto errattr; + nla_nest_end(skb, port_attrs); + return 0; +errattr: + nla_nest_cancel(skb, port_attrs); +errout: + return -EMSGSIZE; +} + +static bool smc_nl_handle_pci_values(const struct smc_pci_dev *smc_pci_dev, + struct sk_buff *skb) +{ + if (nla_put_u32(skb, SMC_NLA_DEV_PCI_FID, smc_pci_dev->pci_fid)) + return false; + if (nla_put_u16(skb, SMC_NLA_DEV_PCI_CHID, smc_pci_dev->pci_pchid)) + return false; + if (nla_put_u16(skb, SMC_NLA_DEV_PCI_VENDOR, smc_pci_dev->pci_vendor)) + return false; + if (nla_put_u16(skb, SMC_NLA_DEV_PCI_DEVICE, smc_pci_dev->pci_device)) + return false; + if (nla_put_string(skb, SMC_NLA_DEV_PCI_ID, smc_pci_dev->pci_id)) + return false; + return true; +} + +static int smc_nl_handle_smcr_dev(struct smc_ib_device *smcibdev, + struct sk_buff *skb, + struct netlink_callback *cb) +{ + char smc_ibname[IB_DEVICE_NAME_MAX]; + struct smc_pci_dev smc_pci_dev; + struct pci_dev *pci_dev; + unsigned char is_crit; + struct nlattr *attrs; + void *nlh; + int i; + + nlh = genlmsg_put(skb, NETLINK_CB(cb->skb).portid, cb->nlh->nlmsg_seq, + &smc_gen_nl_family, NLM_F_MULTI, + SMC_NETLINK_GET_DEV_SMCR); + if (!nlh) + goto errmsg; + attrs = nla_nest_start(skb, SMC_GEN_DEV_SMCR); + if (!attrs) + goto errout; + is_crit = smcr_diag_is_dev_critical(&smc_lgr_list, smcibdev); + if (nla_put_u8(skb, SMC_NLA_DEV_IS_CRIT, is_crit)) + goto errattr; + if (smcibdev->ibdev->dev.parent) { + memset(&smc_pci_dev, 0, sizeof(smc_pci_dev)); + pci_dev = to_pci_dev(smcibdev->ibdev->dev.parent); + smc_set_pci_values(pci_dev, &smc_pci_dev); + if (!smc_nl_handle_pci_values(&smc_pci_dev, skb)) + goto errattr; + } + snprintf(smc_ibname, sizeof(smc_ibname), "%s", smcibdev->ibdev->name); + if (nla_put_string(skb, SMC_NLA_DEV_IB_NAME, smc_ibname)) + goto errattr; + for (i = 1; i <= SMC_MAX_PORTS; i++) { + if (!rdma_is_port_valid(smcibdev->ibdev, i)) + continue; + if (smc_nl_handle_dev_port(skb, smcibdev->ibdev, + smcibdev, i - 1)) + goto errattr; + } + + nla_nest_end(skb, attrs); + genlmsg_end(skb, nlh); + return 0; + +errattr: + nla_nest_cancel(skb, attrs); +errout: + genlmsg_cancel(skb, nlh); +errmsg: + return -EMSGSIZE; +} + +static void smc_nl_prep_smcr_dev(struct smc_ib_devices *dev_list, + struct sk_buff *skb, + struct netlink_callback *cb) +{ + struct smc_nl_dmp_ctx *cb_ctx = smc_nl_dmp_ctx(cb); + struct smc_ib_device *smcibdev; + int snum = cb_ctx->pos[0]; + int num = 0; + + mutex_lock(&dev_list->mutex); + list_for_each_entry(smcibdev, &dev_list->list, list) { + if (num < snum) + goto next; + if (smc_nl_handle_smcr_dev(smcibdev, skb, cb)) + goto errout; +next: + num++; + } +errout: + mutex_unlock(&dev_list->mutex); + cb_ctx->pos[0] = num; +} + +int smcr_nl_get_device(struct sk_buff *skb, struct netlink_callback *cb) +{ + smc_nl_prep_smcr_dev(&smc_ib_devices, skb, cb); + return skb->len; +} + static void smc_ib_qp_event_handler(struct ib_event *ibevent, void *priv) { struct smc_link *lnk = (struct smc_link *)priv; @@ -588,6 +754,49 @@ static void smc_ib_cleanup_per_ibdev(struct smc_ib_device *smcibdev) static struct ib_client smc_ib_client; +static void smc_copy_netdev_ifindex(struct smc_ib_device *smcibdev, int port) +{ + struct ib_device *ibdev = smcibdev->ibdev; + struct net_device *ndev; + + if (!ibdev->ops.get_netdev) + return; + ndev = ibdev->ops.get_netdev(ibdev, port + 1); + if (ndev) { + smcibdev->ndev_ifidx[port] = ndev->ifindex; + dev_put(ndev); + } +} + +void smc_ib_ndev_change(struct net_device *ndev, unsigned long event) +{ + struct smc_ib_device *smcibdev; + struct ib_device *libdev; + struct net_device *lndev; + u8 port_cnt; + int i; + + mutex_lock(&smc_ib_devices.mutex); + list_for_each_entry(smcibdev, &smc_ib_devices.list, list) { + port_cnt = smcibdev->ibdev->phys_port_cnt; + for (i = 0; i < min_t(size_t, port_cnt, SMC_MAX_PORTS); i++) { + libdev = smcibdev->ibdev; + if (!libdev->ops.get_netdev) + continue; + lndev = libdev->ops.get_netdev(libdev, i + 1); + if (lndev) + dev_put(lndev); + if (lndev != ndev) + continue; + if (event == NETDEV_REGISTER) + smcibdev->ndev_ifidx[i] = ndev->ifindex; + if (event == NETDEV_UNREGISTER) + smcibdev->ndev_ifidx[i] = 0; + } + } + mutex_unlock(&smc_ib_devices.mutex); +} + /* callback function for ib_register_client() */ static int smc_ib_add_dev(struct ib_device *ibdev) { @@ -627,6 +836,7 @@ static int smc_ib_add_dev(struct ib_device *ibdev) if (smc_pnetid_by_dev_port(ibdev->dev.parent, i, smcibdev->pnetid[i])) smc_pnetid_by_table_ib(smcibdev, i + 1); + smc_copy_netdev_ifindex(smcibdev, i); pr_warn_ratelimited("smc: ib device %s port %d has pnetid " "%.16s%s\n", smcibdev->ibdev->name, i + 1, diff --git a/net/smc/smc_ib.h b/net/smc/smc_ib.h index d08c58229d4c91e0d55c959871fec5cd53b40429..ead57976d354d3f6a2a5345e03e63c6b3488e67c 100644 --- a/net/smc/smc_ib.h +++ b/net/smc/smc_ib.h @@ -30,6 +30,7 @@ struct smc_ib_devices { /* list of smc ib devices definition */ }; extern struct smc_ib_devices smc_ib_devices; /* list of smc ib devices */ +extern struct smc_lgr_list smc_lgr_list; /* list of linkgroups */ struct smc_ib_device { /* ib-device infos for smc */ struct list_head list; @@ -53,11 +54,15 @@ struct smc_ib_device { /* ib-device infos for smc */ atomic_t lnk_cnt; /* number of links on ibdev */ wait_queue_head_t lnks_deleted; /* wait 4 removal of all links*/ struct mutex mutex; /* protect dev setup+cleanup */ + atomic_t lnk_cnt_by_port[SMC_MAX_PORTS]; + /* number of links per port */ + int ndev_ifidx[SMC_MAX_PORTS]; /* ndev if indexes */ }; struct smc_buf_desc; struct smc_link; +void smc_ib_ndev_change(struct net_device *ndev, unsigned long event); int smc_ib_register_client(void) __init; void smc_ib_unregister_client(void); bool smc_ib_port_active(struct smc_ib_device *smcibdev, u8 ibport); @@ -90,4 +95,5 @@ void smc_ib_sync_sg_for_device(struct smc_link *lnk, int smc_ib_determine_gid(struct smc_ib_device *smcibdev, u8 ibport, unsigned short vlan_id, u8 gid[], u8 *sgid_index); bool smc_ib_is_valid_local_systemid(void); +int smcr_nl_get_device(struct sk_buff *skb, struct netlink_callback *cb); #endif diff --git a/net/smc/smc_ism.c b/net/smc/smc_ism.c index 8e33c0128d73a8a52f97b6405d7eb32d560b9927..967712ba52a0d36d525ce808ad1d0ffb62bf5837 100644 --- a/net/smc/smc_ism.c +++ b/net/smc/smc_ism.c @@ -15,13 +15,14 @@ #include "smc_core.h" #include "smc_ism.h" #include "smc_pnet.h" +#include "smc_netlink.h" struct smcd_dev_list smcd_dev_list = { .list = LIST_HEAD_INIT(smcd_dev_list.list), .mutex = __MUTEX_INITIALIZER(smcd_dev_list.mutex) }; -bool smc_ism_v2_capable; +static bool smc_ism_v2_capable; /* Test if an ISM communication is possible - same CPC */ int smc_ism_cantalk(u64 peer_gid, unsigned short vlan_id, struct smcd_dev *smcd) @@ -51,6 +52,12 @@ u16 smc_ism_get_chid(struct smcd_dev *smcd) return smcd->ops->get_chid(smcd); } +/* HW supports ISM V2 and thus System EID is defined */ +bool smc_ism_is_v2_capable(void) +{ + return smc_ism_v2_capable; +} + /* Set a connection using this DMBE. */ void smc_ism_set_conn(struct smc_connection *conn) { @@ -201,6 +208,97 @@ int smc_ism_register_dmb(struct smc_link_group *lgr, int dmb_len, return rc; } +static int smc_nl_handle_smcd_dev(struct smcd_dev *smcd, + struct sk_buff *skb, + struct netlink_callback *cb) +{ + char smc_pnet[SMC_MAX_PNETID_LEN + 1]; + struct smc_pci_dev smc_pci_dev; + struct nlattr *port_attrs; + struct nlattr *attrs; + int use_cnt = 0; + void *nlh; + + nlh = genlmsg_put(skb, NETLINK_CB(cb->skb).portid, cb->nlh->nlmsg_seq, + &smc_gen_nl_family, NLM_F_MULTI, + SMC_NETLINK_GET_DEV_SMCD); + if (!nlh) + goto errmsg; + attrs = nla_nest_start(skb, SMC_GEN_DEV_SMCD); + if (!attrs) + goto errout; + use_cnt = atomic_read(&smcd->lgr_cnt); + if (nla_put_u32(skb, SMC_NLA_DEV_USE_CNT, use_cnt)) + goto errattr; + if (nla_put_u8(skb, SMC_NLA_DEV_IS_CRIT, use_cnt > 0)) + goto errattr; + memset(&smc_pci_dev, 0, sizeof(smc_pci_dev)); + smc_set_pci_values(to_pci_dev(smcd->dev.parent), &smc_pci_dev); + if (nla_put_u32(skb, SMC_NLA_DEV_PCI_FID, smc_pci_dev.pci_fid)) + goto errattr; + if (nla_put_u16(skb, SMC_NLA_DEV_PCI_CHID, smc_pci_dev.pci_pchid)) + goto errattr; + if (nla_put_u16(skb, SMC_NLA_DEV_PCI_VENDOR, smc_pci_dev.pci_vendor)) + goto errattr; + if (nla_put_u16(skb, SMC_NLA_DEV_PCI_DEVICE, smc_pci_dev.pci_device)) + goto errattr; + if (nla_put_string(skb, SMC_NLA_DEV_PCI_ID, smc_pci_dev.pci_id)) + goto errattr; + + port_attrs = nla_nest_start(skb, SMC_NLA_DEV_PORT); + if (!port_attrs) + goto errattr; + if (nla_put_u8(skb, SMC_NLA_DEV_PORT_PNET_USR, smcd->pnetid_by_user)) + goto errportattr; + memcpy(smc_pnet, smcd->pnetid, SMC_MAX_PNETID_LEN); + smc_pnet[SMC_MAX_PNETID_LEN] = 0; + if (nla_put_string(skb, SMC_NLA_DEV_PORT_PNETID, smc_pnet)) + goto errportattr; + + nla_nest_end(skb, port_attrs); + nla_nest_end(skb, attrs); + genlmsg_end(skb, nlh); + return 0; + +errportattr: + nla_nest_cancel(skb, port_attrs); +errattr: + nla_nest_cancel(skb, attrs); +errout: + nlmsg_cancel(skb, nlh); +errmsg: + return -EMSGSIZE; +} + +static void smc_nl_prep_smcd_dev(struct smcd_dev_list *dev_list, + struct sk_buff *skb, + struct netlink_callback *cb) +{ + struct smc_nl_dmp_ctx *cb_ctx = smc_nl_dmp_ctx(cb); + int snum = cb_ctx->pos[0]; + struct smcd_dev *smcd; + int num = 0; + + mutex_lock(&dev_list->mutex); + list_for_each_entry(smcd, &dev_list->list, list) { + if (num < snum) + goto next; + if (smc_nl_handle_smcd_dev(smcd, skb, cb)) + goto errout; +next: + num++; + } +errout: + mutex_unlock(&dev_list->mutex); + cb_ctx->pos[0] = num; +} + +int smcd_nl_get_device(struct sk_buff *skb, struct netlink_callback *cb) +{ + smc_nl_prep_smcd_dev(&smcd_dev_list, skb, cb); + return skb->len; +} + struct smc_ism_event_work { struct work_struct work; struct smcd_dev *smcd; diff --git a/net/smc/smc_ism.h b/net/smc/smc_ism.h index 8048e09ddcf86295ee101532f033f340d134a59f..113efc7352ed67b953105a8bf7ee6a95d9313bc8 100644 --- a/net/smc/smc_ism.h +++ b/net/smc/smc_ism.h @@ -10,6 +10,7 @@ #define SMCD_ISM_H #include +#include #include #include "smc.h" @@ -20,9 +21,6 @@ struct smcd_dev_list { /* List of SMCD devices */ }; extern struct smcd_dev_list smcd_dev_list; /* list of smcd devices */ -extern bool smc_ism_v2_capable; /* HW supports ISM V2 and thus - * System EID is defined - */ struct smc_ism_vlanid { /* VLAN id set on ISM device */ struct list_head list; @@ -52,5 +50,7 @@ int smc_ism_write(struct smcd_dev *dev, const struct smc_ism_position *pos, int smc_ism_signal_shutdown(struct smc_link_group *lgr); void smc_ism_get_system_eid(struct smcd_dev *dev, u8 **eid); u16 smc_ism_get_chid(struct smcd_dev *dev); +bool smc_ism_is_v2_capable(void); void smc_ism_init(void); +int smcd_nl_get_device(struct sk_buff *skb, struct netlink_callback *cb); #endif diff --git a/net/smc/smc_netlink.c b/net/smc/smc_netlink.c new file mode 100644 index 0000000000000000000000000000000000000000..6fb6f96c1d1775a5dbd30aaff1afe64ee40df348 --- /dev/null +++ b/net/smc/smc_netlink.c @@ -0,0 +1,96 @@ +// SPDX-License-Identifier: GPL-2.0 +/* + * Shared Memory Communications over RDMA (SMC-R) and RoCE + * + * Generic netlink support functions to interact with SMC module + * + * Copyright IBM Corp. 2020 + * + * Author(s): Guvenc Gulce + */ + +#include +#include +#include +#include +#include +#include + +#include "smc_core.h" +#include "smc_ism.h" +#include "smc_ib.h" +#include "smc_stats.h" +#include "smc_netlink.h" + +#define SMC_CMD_MAX_ATTR 1 + +/* SMC_GENL generic netlink operation definition */ +static const struct genl_ops smc_gen_nl_ops[] = { + { + .cmd = SMC_NETLINK_GET_SYS_INFO, + /* can be retrieved by unprivileged users */ + .dumpit = smc_nl_get_sys_info, + }, + { + .cmd = SMC_NETLINK_GET_LGR_SMCR, + /* can be retrieved by unprivileged users */ + .dumpit = smcr_nl_get_lgr, + }, + { + .cmd = SMC_NETLINK_GET_LINK_SMCR, + /* can be retrieved by unprivileged users */ + .dumpit = smcr_nl_get_link, + }, + { + .cmd = SMC_NETLINK_GET_LGR_SMCD, + /* can be retrieved by unprivileged users */ + .dumpit = smcd_nl_get_lgr, + }, + { + .cmd = SMC_NETLINK_GET_DEV_SMCD, + /* can be retrieved by unprivileged users */ + .dumpit = smcd_nl_get_device, + }, + { + .cmd = SMC_NETLINK_GET_DEV_SMCR, + /* can be retrieved by unprivileged users */ + .dumpit = smcr_nl_get_device, + }, + { + .cmd = SMC_NETLINK_GET_STATS, + /* can be retrieved by unprivileged users */ + .dumpit = smc_nl_get_stats, + }, + { + .cmd = SMC_NETLINK_GET_FBACK_STATS, + /* can be retrieved by unprivileged users */ + .dumpit = smc_nl_get_fback_stats, + }, +}; + +static const struct nla_policy smc_gen_nl_policy[2] = { + [SMC_CMD_MAX_ATTR] = { .type = NLA_REJECT, }, +}; + +/* SMC_GENL family definition */ +struct genl_family smc_gen_nl_family __ro_after_init = { + .hdrsize = 0, + .name = SMC_GENL_FAMILY_NAME, + .version = SMC_GENL_FAMILY_VERSION, + .maxattr = SMC_CMD_MAX_ATTR, + .policy = smc_gen_nl_policy, + .netnsok = true, + .module = THIS_MODULE, + .ops = smc_gen_nl_ops, + .n_ops = ARRAY_SIZE(smc_gen_nl_ops) +}; + +int __init smc_nl_init(void) +{ + return genl_register_family(&smc_gen_nl_family); +} + +void smc_nl_exit(void) +{ + genl_unregister_family(&smc_gen_nl_family); +} diff --git a/net/smc/smc_netlink.h b/net/smc/smc_netlink.h new file mode 100644 index 0000000000000000000000000000000000000000..5ce2c0a89ccda1430d2afcce8184b38911d08f31 --- /dev/null +++ b/net/smc/smc_netlink.h @@ -0,0 +1,32 @@ +/* SPDX-License-Identifier: GPL-2.0 */ +/* + * Shared Memory Communications over RDMA (SMC-R) and RoCE + * + * SMC Generic netlink operations + * + * Copyright IBM Corp. 2020 + * + * Author(s): Guvenc Gulce + */ + +#ifndef _SMC_NETLINK_H +#define _SMC_NETLINK_H + +#include +#include + +extern struct genl_family smc_gen_nl_family; + +struct smc_nl_dmp_ctx { + int pos[3]; +}; + +static inline struct smc_nl_dmp_ctx *smc_nl_dmp_ctx(struct netlink_callback *c) +{ + return (struct smc_nl_dmp_ctx *)c->ctx; +} + +int smc_nl_init(void) __init; +void smc_nl_exit(void); + +#endif diff --git a/net/smc/smc_pnet.c b/net/smc/smc_pnet.c index 30bae60d626c63b1b37d7f4d919c4b184ba5acf4..eb0971a1797adb72dbea802ee54e02a5484e5af6 100644 --- a/net/smc/smc_pnet.c +++ b/net/smc/smc_pnet.c @@ -828,9 +828,11 @@ static int smc_pnet_netdev_event(struct notifier_block *this, case NETDEV_REBOOT: case NETDEV_UNREGISTER: smc_pnet_remove_by_ndev(event_dev); + smc_ib_ndev_change(event_dev, event); return NOTIFY_OK; case NETDEV_REGISTER: smc_pnet_add_by_ndev(event_dev); + smc_ib_ndev_change(event_dev, event); return NOTIFY_OK; case NETDEV_UP: smc_pnet_add_base_pnetid(net, event_dev, ndev_pnetid); diff --git a/net/smc/smc_rx.c b/net/smc/smc_rx.c index 44583bf2df8e487cde9cdc1d58de46bd4d8fde4d..de6e3ac5244ea98095cfb537c1ddc58d882a7f04 100644 --- a/net/smc/smc_rx.c +++ b/net/smc/smc_rx.c @@ -21,6 +21,7 @@ #include "smc_cdc.h" #include "smc_tx.h" /* smc_tx_consumer_update() */ #include "smc_rx.h" +#include "smc_stats.h" /* callback implementation to wakeup consumers blocked with smc_rx_wait(). * indirectly called by smc_cdc_msg_recv_action(). @@ -285,6 +286,7 @@ static int smc_rx_recv_urg(struct smc_sock *smc, struct msghdr *msg, int len, conn->urg_state == SMC_URG_READ) return -EINVAL; + SMC_STAT_INC(!conn->lnk, urg_data_cnt); if (conn->urg_state == SMC_URG_VALID) { if (!(flags & MSG_PEEK)) smc->conn.urg_state = SMC_URG_READ; @@ -361,6 +363,12 @@ int smc_rx_recvmsg(struct smc_sock *smc, struct msghdr *msg, timeo = sock_rcvtimeo(sk, flags & MSG_DONTWAIT); target = sock_rcvlowat(sk, flags & MSG_WAITALL, len); + readable = atomic_read(&conn->bytes_to_rcv); + if (readable >= conn->rmb_desc->len) + SMC_STAT_RMB_RX_FULL(!conn->lnk); + + if (len < readable) + SMC_STAT_RMB_RX_SIZE_SMALL(!conn->lnk); /* we currently use 1 RMBE per RMB, so RMBE == RMB base addr */ rcvbuf_base = conn->rx_off + conn->rmb_desc->cpu_addr; diff --git a/net/smc/smc_stats.c b/net/smc/smc_stats.c new file mode 100644 index 0000000000000000000000000000000000000000..44491016707d04093fbb0ae156ee3b438d99b4f5 --- /dev/null +++ b/net/smc/smc_stats.c @@ -0,0 +1,406 @@ +// SPDX-License-Identifier: GPL-2.0-only +/* + * Shared Memory Communications over RDMA (SMC-R) and RoCE + * + * SMC statistics netlink routines + * + * Copyright IBM Corp. 2021 + * + * Author(s): Guvenc Gulce + */ +#include +#include +#include +#include +#include +#include +#include +#include "smc_netlink.h" +#include "smc_stats.h" + +/* serialize fallback reason statistic gathering */ +DEFINE_MUTEX(smc_stat_fback_rsn); +struct smc_stats __percpu *smc_stats; /* per cpu counters for SMC */ +struct smc_stats_reason fback_rsn; + +int __init smc_stats_init(void) +{ + memset(&fback_rsn, 0, sizeof(fback_rsn)); + smc_stats = alloc_percpu(struct smc_stats); + if (!smc_stats) + return -ENOMEM; + + return 0; +} + +void smc_stats_exit(void) +{ + free_percpu(smc_stats); +} + +static int smc_nl_fill_stats_rmb_data(struct sk_buff *skb, + struct smc_stats *stats, int tech, + int type) +{ + struct smc_stats_rmbcnt *stats_rmb_cnt; + struct nlattr *attrs; + + if (type == SMC_NLA_STATS_T_TX_RMB_STATS) + stats_rmb_cnt = &stats->smc[tech].rmb_tx; + else + stats_rmb_cnt = &stats->smc[tech].rmb_rx; + + attrs = nla_nest_start(skb, type); + if (!attrs) + goto errout; + if (nla_put_u64_64bit(skb, SMC_NLA_STATS_RMB_REUSE_CNT, + stats_rmb_cnt->reuse_cnt, + SMC_NLA_STATS_RMB_PAD)) + goto errattr; + if (nla_put_u64_64bit(skb, SMC_NLA_STATS_RMB_SIZE_SM_PEER_CNT, + stats_rmb_cnt->buf_size_small_peer_cnt, + SMC_NLA_STATS_RMB_PAD)) + goto errattr; + if (nla_put_u64_64bit(skb, SMC_NLA_STATS_RMB_SIZE_SM_CNT, + stats_rmb_cnt->buf_size_small_cnt, + SMC_NLA_STATS_RMB_PAD)) + goto errattr; + if (nla_put_u64_64bit(skb, SMC_NLA_STATS_RMB_FULL_PEER_CNT, + stats_rmb_cnt->buf_full_peer_cnt, + SMC_NLA_STATS_RMB_PAD)) + goto errattr; + if (nla_put_u64_64bit(skb, SMC_NLA_STATS_RMB_FULL_CNT, + stats_rmb_cnt->buf_full_cnt, + SMC_NLA_STATS_RMB_PAD)) + goto errattr; + if (nla_put_u64_64bit(skb, SMC_NLA_STATS_RMB_ALLOC_CNT, + stats_rmb_cnt->alloc_cnt, + SMC_NLA_STATS_RMB_PAD)) + goto errattr; + if (nla_put_u64_64bit(skb, SMC_NLA_STATS_RMB_DGRADE_CNT, + stats_rmb_cnt->dgrade_cnt, + SMC_NLA_STATS_RMB_PAD)) + goto errattr; + + nla_nest_end(skb, attrs); + return 0; + +errattr: + nla_nest_cancel(skb, attrs); +errout: + return -EMSGSIZE; +} + +static int smc_nl_fill_stats_bufsize_data(struct sk_buff *skb, + struct smc_stats *stats, int tech, + int type) +{ + struct smc_stats_memsize *stats_pload; + struct nlattr *attrs; + + if (type == SMC_NLA_STATS_T_TXPLOAD_SIZE) + stats_pload = &stats->smc[tech].tx_pd; + else if (type == SMC_NLA_STATS_T_RXPLOAD_SIZE) + stats_pload = &stats->smc[tech].rx_pd; + else if (type == SMC_NLA_STATS_T_TX_RMB_SIZE) + stats_pload = &stats->smc[tech].tx_rmbsize; + else if (type == SMC_NLA_STATS_T_RX_RMB_SIZE) + stats_pload = &stats->smc[tech].rx_rmbsize; + else + goto errout; + + attrs = nla_nest_start(skb, type); + if (!attrs) + goto errout; + if (nla_put_u64_64bit(skb, SMC_NLA_STATS_PLOAD_8K, + stats_pload->buf[SMC_BUF_8K], + SMC_NLA_STATS_PLOAD_PAD)) + goto errattr; + if (nla_put_u64_64bit(skb, SMC_NLA_STATS_PLOAD_16K, + stats_pload->buf[SMC_BUF_16K], + SMC_NLA_STATS_PLOAD_PAD)) + goto errattr; + if (nla_put_u64_64bit(skb, SMC_NLA_STATS_PLOAD_32K, + stats_pload->buf[SMC_BUF_32K], + SMC_NLA_STATS_PLOAD_PAD)) + goto errattr; + if (nla_put_u64_64bit(skb, SMC_NLA_STATS_PLOAD_64K, + stats_pload->buf[SMC_BUF_64K], + SMC_NLA_STATS_PLOAD_PAD)) + goto errattr; + if (nla_put_u64_64bit(skb, SMC_NLA_STATS_PLOAD_128K, + stats_pload->buf[SMC_BUF_128K], + SMC_NLA_STATS_PLOAD_PAD)) + goto errattr; + if (nla_put_u64_64bit(skb, SMC_NLA_STATS_PLOAD_256K, + stats_pload->buf[SMC_BUF_256K], + SMC_NLA_STATS_PLOAD_PAD)) + goto errattr; + if (nla_put_u64_64bit(skb, SMC_NLA_STATS_PLOAD_512K, + stats_pload->buf[SMC_BUF_512K], + SMC_NLA_STATS_PLOAD_PAD)) + goto errattr; + if (nla_put_u64_64bit(skb, SMC_NLA_STATS_PLOAD_1024K, + stats_pload->buf[SMC_BUF_1024K], + SMC_NLA_STATS_PLOAD_PAD)) + goto errattr; + if (nla_put_u64_64bit(skb, SMC_NLA_STATS_PLOAD_G_1024K, + stats_pload->buf[SMC_BUF_G_1024K], + SMC_NLA_STATS_PLOAD_PAD)) + goto errattr; + + nla_nest_end(skb, attrs); + return 0; + +errattr: + nla_nest_cancel(skb, attrs); +errout: + return -EMSGSIZE; +} + +static int smc_nl_fill_stats_tech_data(struct sk_buff *skb, + struct smc_stats *stats, int tech) +{ + struct smc_stats_tech *smc_tech; + struct nlattr *attrs; + + smc_tech = &stats->smc[tech]; + if (tech == SMC_TYPE_D) + attrs = nla_nest_start(skb, SMC_NLA_STATS_SMCD_TECH); + else + attrs = nla_nest_start(skb, SMC_NLA_STATS_SMCR_TECH); + + if (!attrs) + goto errout; + if (smc_nl_fill_stats_rmb_data(skb, stats, tech, + SMC_NLA_STATS_T_TX_RMB_STATS)) + goto errattr; + if (smc_nl_fill_stats_rmb_data(skb, stats, tech, + SMC_NLA_STATS_T_RX_RMB_STATS)) + goto errattr; + if (smc_nl_fill_stats_bufsize_data(skb, stats, tech, + SMC_NLA_STATS_T_TXPLOAD_SIZE)) + goto errattr; + if (smc_nl_fill_stats_bufsize_data(skb, stats, tech, + SMC_NLA_STATS_T_RXPLOAD_SIZE)) + goto errattr; + if (smc_nl_fill_stats_bufsize_data(skb, stats, tech, + SMC_NLA_STATS_T_TX_RMB_SIZE)) + goto errattr; + if (smc_nl_fill_stats_bufsize_data(skb, stats, tech, + SMC_NLA_STATS_T_RX_RMB_SIZE)) + goto errattr; + if (nla_put_u64_64bit(skb, SMC_NLA_STATS_T_CLNT_V1_SUCC, + smc_tech->clnt_v1_succ_cnt, + SMC_NLA_STATS_PAD)) + goto errattr; + if (nla_put_u64_64bit(skb, SMC_NLA_STATS_T_CLNT_V2_SUCC, + smc_tech->clnt_v2_succ_cnt, + SMC_NLA_STATS_PAD)) + goto errattr; + if (nla_put_u64_64bit(skb, SMC_NLA_STATS_T_SRV_V1_SUCC, + smc_tech->srv_v1_succ_cnt, + SMC_NLA_STATS_PAD)) + goto errattr; + if (nla_put_u64_64bit(skb, SMC_NLA_STATS_T_SRV_V2_SUCC, + smc_tech->srv_v2_succ_cnt, + SMC_NLA_STATS_PAD)) + goto errattr; + if (nla_put_u64_64bit(skb, SMC_NLA_STATS_T_RX_BYTES, + smc_tech->rx_bytes, + SMC_NLA_STATS_PAD)) + goto errattr; + if (nla_put_u64_64bit(skb, SMC_NLA_STATS_T_TX_BYTES, + smc_tech->tx_bytes, + SMC_NLA_STATS_PAD)) + goto errattr; + if (nla_put_u64_64bit(skb, SMC_NLA_STATS_T_RX_CNT, + smc_tech->rx_cnt, + SMC_NLA_STATS_PAD)) + goto errattr; + if (nla_put_u64_64bit(skb, SMC_NLA_STATS_T_TX_CNT, + smc_tech->tx_cnt, + SMC_NLA_STATS_PAD)) + goto errattr; + if (nla_put_u64_64bit(skb, SMC_NLA_STATS_T_SENDPAGE_CNT, + smc_tech->sendpage_cnt, + SMC_NLA_STATS_PAD)) + goto errattr; + if (nla_put_u64_64bit(skb, SMC_NLA_STATS_T_CORK_CNT, + smc_tech->cork_cnt, + SMC_NLA_STATS_PAD)) + goto errattr; + if (nla_put_u64_64bit(skb, SMC_NLA_STATS_T_NDLY_CNT, + smc_tech->ndly_cnt, + SMC_NLA_STATS_PAD)) + goto errattr; + if (nla_put_u64_64bit(skb, SMC_NLA_STATS_T_SPLICE_CNT, + smc_tech->splice_cnt, + SMC_NLA_STATS_PAD)) + goto errattr; + if (nla_put_u64_64bit(skb, SMC_NLA_STATS_T_URG_DATA_CNT, + smc_tech->urg_data_cnt, + SMC_NLA_STATS_PAD)) + goto errattr; + + nla_nest_end(skb, attrs); + return 0; + +errattr: + nla_nest_cancel(skb, attrs); +errout: + return -EMSGSIZE; +} + +int smc_nl_get_stats(struct sk_buff *skb, + struct netlink_callback *cb) +{ + struct smc_nl_dmp_ctx *cb_ctx = smc_nl_dmp_ctx(cb); + struct smc_stats *stats; + struct nlattr *attrs; + int cpu, i, size; + void *nlh; + u64 *src; + u64 *sum; + + if (cb_ctx->pos[0]) + goto errmsg; + nlh = genlmsg_put(skb, NETLINK_CB(cb->skb).portid, cb->nlh->nlmsg_seq, + &smc_gen_nl_family, NLM_F_MULTI, + SMC_NETLINK_GET_STATS); + if (!nlh) + goto errmsg; + + attrs = nla_nest_start(skb, SMC_GEN_STATS); + if (!attrs) + goto errnest; + stats = kzalloc(sizeof(*stats), GFP_KERNEL); + if (!stats) + goto erralloc; + size = sizeof(*stats) / sizeof(u64); + for_each_possible_cpu(cpu) { + src = (u64 *)per_cpu_ptr(smc_stats, cpu); + sum = (u64 *)stats; + for (i = 0; i < size; i++) + *(sum++) += *(src++); + } + if (smc_nl_fill_stats_tech_data(skb, stats, SMC_TYPE_D)) + goto errattr; + if (smc_nl_fill_stats_tech_data(skb, stats, SMC_TYPE_R)) + goto errattr; + if (nla_put_u64_64bit(skb, SMC_NLA_STATS_CLNT_HS_ERR_CNT, + stats->clnt_hshake_err_cnt, + SMC_NLA_STATS_PAD)) + goto errattr; + if (nla_put_u64_64bit(skb, SMC_NLA_STATS_SRV_HS_ERR_CNT, + stats->srv_hshake_err_cnt, + SMC_NLA_STATS_PAD)) + goto errattr; + + nla_nest_end(skb, attrs); + genlmsg_end(skb, nlh); + cb_ctx->pos[0] = 1; + kfree(stats); + return skb->len; + +errattr: + kfree(stats); +erralloc: + nla_nest_cancel(skb, attrs); +errnest: + genlmsg_cancel(skb, nlh); +errmsg: + return skb->len; +} + +static int smc_nl_get_fback_details(struct sk_buff *skb, + struct netlink_callback *cb, int pos, + bool is_srv) +{ + struct smc_nl_dmp_ctx *cb_ctx = smc_nl_dmp_ctx(cb); + int cnt_reported = cb_ctx->pos[2]; + struct smc_stats_fback *trgt_arr; + struct nlattr *attrs; + int rc = 0; + void *nlh; + + if (is_srv) + trgt_arr = &fback_rsn.srv[0]; + else + trgt_arr = &fback_rsn.clnt[0]; + if (!trgt_arr[pos].fback_code) + return -ENODATA; + nlh = genlmsg_put(skb, NETLINK_CB(cb->skb).portid, cb->nlh->nlmsg_seq, + &smc_gen_nl_family, NLM_F_MULTI, + SMC_NETLINK_GET_FBACK_STATS); + if (!nlh) + goto errmsg; + attrs = nla_nest_start(skb, SMC_GEN_FBACK_STATS); + if (!attrs) + goto errout; + if (nla_put_u8(skb, SMC_NLA_FBACK_STATS_TYPE, is_srv)) + goto errattr; + if (!cnt_reported) { + if (nla_put_u64_64bit(skb, SMC_NLA_FBACK_STATS_SRV_CNT, + fback_rsn.srv_fback_cnt, + SMC_NLA_FBACK_STATS_PAD)) + goto errattr; + if (nla_put_u64_64bit(skb, SMC_NLA_FBACK_STATS_CLNT_CNT, + fback_rsn.clnt_fback_cnt, + SMC_NLA_FBACK_STATS_PAD)) + goto errattr; + cnt_reported = 1; + } + + if (nla_put_u32(skb, SMC_NLA_FBACK_STATS_RSN_CODE, + trgt_arr[pos].fback_code)) + goto errattr; + if (nla_put_u16(skb, SMC_NLA_FBACK_STATS_RSN_CNT, + trgt_arr[pos].count)) + goto errattr; + + cb_ctx->pos[2] = cnt_reported; + nla_nest_end(skb, attrs); + genlmsg_end(skb, nlh); + return rc; + +errattr: + nla_nest_cancel(skb, attrs); +errout: + genlmsg_cancel(skb, nlh); +errmsg: + return -EMSGSIZE; +} + +int smc_nl_get_fback_stats(struct sk_buff *skb, struct netlink_callback *cb) +{ + struct smc_nl_dmp_ctx *cb_ctx = smc_nl_dmp_ctx(cb); + int rc_srv = 0, rc_clnt = 0, k; + int skip_serv = cb_ctx->pos[1]; + int snum = cb_ctx->pos[0]; + bool is_srv = true; + + mutex_lock(&smc_stat_fback_rsn); + for (k = 0; k < SMC_MAX_FBACK_RSN_CNT; k++) { + if (k < snum) + continue; + if (!skip_serv) { + rc_srv = smc_nl_get_fback_details(skb, cb, k, is_srv); + if (rc_srv && rc_srv != -ENODATA) + break; + } else { + skip_serv = 0; + } + rc_clnt = smc_nl_get_fback_details(skb, cb, k, !is_srv); + if (rc_clnt && rc_clnt != -ENODATA) { + skip_serv = 1; + break; + } + if (rc_clnt == -ENODATA && rc_srv == -ENODATA) + break; + } + mutex_unlock(&smc_stat_fback_rsn); + cb_ctx->pos[1] = skip_serv; + cb_ctx->pos[0] = k; + return skb->len; +} diff --git a/net/smc/smc_stats.h b/net/smc/smc_stats.h new file mode 100644 index 0000000000000000000000000000000000000000..c159acd932c5226bad601fb8bb2708ab401b1661 --- /dev/null +++ b/net/smc/smc_stats.h @@ -0,0 +1,259 @@ +/* SPDX-License-Identifier: GPL-2.0 */ +/* + * Shared Memory Communications over RDMA (SMC-R) and RoCE + * + * Macros for SMC statistics + * + * Copyright IBM Corp. 2021 + * + * Author(s): Guvenc Gulce + */ + +#ifndef NET_SMC_SMC_STATS_H_ +#define NET_SMC_SMC_STATS_H_ +#include +#include +#include +#include +#include + +#include "smc_clc.h" + +#define SMC_MAX_FBACK_RSN_CNT 30 + +extern struct smc_stats __percpu *smc_stats; /* per cpu counters for SMC */ +extern struct smc_stats_reason fback_rsn; +extern struct mutex smc_stat_fback_rsn; + +enum { + SMC_BUF_8K, + SMC_BUF_16K, + SMC_BUF_32K, + SMC_BUF_64K, + SMC_BUF_128K, + SMC_BUF_256K, + SMC_BUF_512K, + SMC_BUF_1024K, + SMC_BUF_G_1024K, + SMC_BUF_MAX, +}; + +struct smc_stats_fback { + int fback_code; + u16 count; +}; + +struct smc_stats_reason { + struct smc_stats_fback srv[SMC_MAX_FBACK_RSN_CNT]; + struct smc_stats_fback clnt[SMC_MAX_FBACK_RSN_CNT]; + u64 srv_fback_cnt; + u64 clnt_fback_cnt; +}; + +struct smc_stats_rmbcnt { + u64 buf_size_small_peer_cnt; + u64 buf_size_small_cnt; + u64 buf_full_peer_cnt; + u64 buf_full_cnt; + u64 reuse_cnt; + u64 alloc_cnt; + u64 dgrade_cnt; +}; + +struct smc_stats_memsize { + u64 buf[SMC_BUF_MAX]; +}; + +struct smc_stats_tech { + struct smc_stats_memsize tx_rmbsize; + struct smc_stats_memsize rx_rmbsize; + struct smc_stats_memsize tx_pd; + struct smc_stats_memsize rx_pd; + struct smc_stats_rmbcnt rmb_tx; + struct smc_stats_rmbcnt rmb_rx; + u64 clnt_v1_succ_cnt; + u64 clnt_v2_succ_cnt; + u64 srv_v1_succ_cnt; + u64 srv_v2_succ_cnt; + u64 sendpage_cnt; + u64 urg_data_cnt; + u64 splice_cnt; + u64 cork_cnt; + u64 ndly_cnt; + u64 rx_bytes; + u64 tx_bytes; + u64 rx_cnt; + u64 tx_cnt; +}; + +struct smc_stats { + struct smc_stats_tech smc[2]; + u64 clnt_hshake_err_cnt; + u64 srv_hshake_err_cnt; +}; + +#define SMC_STAT_PAYLOAD_SUB(_tech, key, _len, _rc) \ +do { \ + typeof(_tech) t = (_tech); \ + typeof(_len) l = (_len); \ + int _pos; \ + typeof(_rc) r = (_rc); \ + int m = SMC_BUF_MAX - 1; \ + this_cpu_inc((*smc_stats).smc[t].key ## _cnt); \ + if (r <= 0 || l <= 0) \ + break; \ + _pos = fls64((l - 1) >> 13); \ + _pos = (_pos <= m) ? _pos : m; \ + this_cpu_inc((*smc_stats).smc[t].key ## _pd.buf[_pos]); \ + this_cpu_add((*smc_stats).smc[t].key ## _bytes, r); \ +} \ +while (0) + +#define SMC_STAT_TX_PAYLOAD(_smc, length, rcode) \ +do { \ + typeof(_smc) __smc = _smc; \ + typeof(length) _len = (length); \ + typeof(rcode) _rc = (rcode); \ + bool is_smcd = !__smc->conn.lnk; \ + if (is_smcd) \ + SMC_STAT_PAYLOAD_SUB(SMC_TYPE_D, tx, _len, _rc); \ + else \ + SMC_STAT_PAYLOAD_SUB(SMC_TYPE_R, tx, _len, _rc); \ +} \ +while (0) + +#define SMC_STAT_RX_PAYLOAD(_smc, length, rcode) \ +do { \ + typeof(_smc) __smc = _smc; \ + typeof(length) _len = (length); \ + typeof(rcode) _rc = (rcode); \ + bool is_smcd = !__smc->conn.lnk; \ + if (is_smcd) \ + SMC_STAT_PAYLOAD_SUB(SMC_TYPE_D, rx, _len, _rc); \ + else \ + SMC_STAT_PAYLOAD_SUB(SMC_TYPE_R, rx, _len, _rc); \ +} \ +while (0) + +#define SMC_STAT_RMB_SIZE_SUB(_tech, k, _len) \ +do { \ + typeof(_len) _l = (_len); \ + typeof(_tech) t = (_tech); \ + int _pos; \ + int m = SMC_BUF_MAX - 1; \ + if (_l <= 0) \ + break; \ + _pos = fls((_l - 1) >> 13); \ + _pos = (_pos <= m) ? _pos : m; \ + this_cpu_inc((*smc_stats).smc[t].k ## _rmbsize.buf[_pos]); \ +} \ +while (0) + +#define SMC_STAT_RMB_SUB(type, t, key) \ + this_cpu_inc((*smc_stats).smc[t].rmb ## _ ## key.type ## _cnt) + +#define SMC_STAT_RMB_SIZE(_is_smcd, _is_rx, _len) \ +do { \ + typeof(_is_smcd) is_d = (_is_smcd); \ + typeof(_is_rx) is_r = (_is_rx); \ + typeof(_len) l = (_len); \ + if ((is_d) && (is_r)) \ + SMC_STAT_RMB_SIZE_SUB(SMC_TYPE_D, rx, l); \ + if ((is_d) && !(is_r)) \ + SMC_STAT_RMB_SIZE_SUB(SMC_TYPE_D, tx, l); \ + if (!(is_d) && (is_r)) \ + SMC_STAT_RMB_SIZE_SUB(SMC_TYPE_R, rx, l); \ + if (!(is_d) && !(is_r)) \ + SMC_STAT_RMB_SIZE_SUB(SMC_TYPE_R, tx, l); \ +} \ +while (0) + +#define SMC_STAT_RMB(type, _is_smcd, _is_rx) \ +do { \ + typeof(_is_smcd) is_d = (_is_smcd); \ + typeof(_is_rx) is_r = (_is_rx); \ + if ((is_d) && (is_r)) \ + SMC_STAT_RMB_SUB(type, SMC_TYPE_D, rx); \ + if ((is_d) && !(is_r)) \ + SMC_STAT_RMB_SUB(type, SMC_TYPE_D, tx); \ + if (!(is_d) && (is_r)) \ + SMC_STAT_RMB_SUB(type, SMC_TYPE_R, rx); \ + if (!(is_d) && !(is_r)) \ + SMC_STAT_RMB_SUB(type, SMC_TYPE_R, tx); \ +} \ +while (0) + +#define SMC_STAT_BUF_REUSE(is_smcd, is_rx) \ + SMC_STAT_RMB(reuse, is_smcd, is_rx) + +#define SMC_STAT_RMB_ALLOC(is_smcd, is_rx) \ + SMC_STAT_RMB(alloc, is_smcd, is_rx) + +#define SMC_STAT_RMB_DOWNGRADED(is_smcd, is_rx) \ + SMC_STAT_RMB(dgrade, is_smcd, is_rx) + +#define SMC_STAT_RMB_TX_PEER_FULL(is_smcd) \ + SMC_STAT_RMB(buf_full_peer, is_smcd, false) + +#define SMC_STAT_RMB_TX_FULL(is_smcd) \ + SMC_STAT_RMB(buf_full, is_smcd, false) + +#define SMC_STAT_RMB_TX_PEER_SIZE_SMALL(is_smcd) \ + SMC_STAT_RMB(buf_size_small_peer, is_smcd, false) + +#define SMC_STAT_RMB_TX_SIZE_SMALL(is_smcd) \ + SMC_STAT_RMB(buf_size_small, is_smcd, false) + +#define SMC_STAT_RMB_RX_SIZE_SMALL(is_smcd) \ + SMC_STAT_RMB(buf_size_small, is_smcd, true) + +#define SMC_STAT_RMB_RX_FULL(is_smcd) \ + SMC_STAT_RMB(buf_full, is_smcd, true) + +#define SMC_STAT_INC(is_smcd, type) \ +do { \ + if ((is_smcd)) \ + this_cpu_inc(smc_stats->smc[SMC_TYPE_D].type); \ + else \ + this_cpu_inc(smc_stats->smc[SMC_TYPE_R].type); \ +} \ +while (0) + +#define SMC_STAT_CLNT_SUCC_INC(_aclc) \ +do { \ + typeof(_aclc) acl = (_aclc); \ + bool is_v2 = (acl->hdr.version == SMC_V2); \ + bool is_smcd = (acl->hdr.typev1 == SMC_TYPE_D); \ + if (is_v2 && is_smcd) \ + this_cpu_inc(smc_stats->smc[SMC_TYPE_D].clnt_v2_succ_cnt); \ + else if (is_v2 && !is_smcd) \ + this_cpu_inc(smc_stats->smc[SMC_TYPE_R].clnt_v2_succ_cnt); \ + else if (!is_v2 && is_smcd) \ + this_cpu_inc(smc_stats->smc[SMC_TYPE_D].clnt_v1_succ_cnt); \ + else if (!is_v2 && !is_smcd) \ + this_cpu_inc(smc_stats->smc[SMC_TYPE_R].clnt_v1_succ_cnt); \ +} \ +while (0) + +#define SMC_STAT_SERV_SUCC_INC(_ini) \ +do { \ + typeof(_ini) i = (_ini); \ + bool is_v2 = (i->smcd_version & SMC_V2); \ + bool is_smcd = (i->is_smcd); \ + if (is_v2 && is_smcd) \ + this_cpu_inc(smc_stats->smc[SMC_TYPE_D].srv_v2_succ_cnt); \ + else if (is_v2 && !is_smcd) \ + this_cpu_inc(smc_stats->smc[SMC_TYPE_R].srv_v2_succ_cnt); \ + else if (!is_v2 && is_smcd) \ + this_cpu_inc(smc_stats->smc[SMC_TYPE_D].srv_v1_succ_cnt); \ + else if (!is_v2 && !is_smcd) \ + this_cpu_inc(smc_stats->smc[SMC_TYPE_R].srv_v1_succ_cnt); \ +} \ +while (0) + +int smc_nl_get_stats(struct sk_buff *skb, struct netlink_callback *cb); +int smc_nl_get_fback_stats(struct sk_buff *skb, struct netlink_callback *cb); +int smc_stats_init(void) __init; +void smc_stats_exit(void); + +#endif /* NET_SMC_SMC_STATS_H_ */ diff --git a/net/smc/smc_tx.c b/net/smc/smc_tx.c index 3a4ce37c9f826e7b720437e57a1e93c990433bf5..d6b9ccb88e65a61a63e129d1f7929a0ecc9a0d83 100644 --- a/net/smc/smc_tx.c +++ b/net/smc/smc_tx.c @@ -27,6 +27,7 @@ #include "smc_close.h" #include "smc_ism.h" #include "smc_tx.h" +#include "smc_stats.h" #define SMC_TX_WORK_DELAY 0 @@ -44,6 +45,8 @@ static void smc_tx_write_space(struct sock *sk) /* similar to sk_stream_write_space */ if (atomic_read(&smc->conn.sndbuf_space) && sock) { + if (test_bit(SOCK_NOSPACE, &sock->flags)) + SMC_STAT_RMB_TX_FULL(!smc->conn.lnk); clear_bit(SOCK_NOSPACE, &sock->flags); rcu_read_lock(); wq = rcu_dereference(sk->sk_wq); @@ -195,9 +198,19 @@ int smc_tx_sendmsg(struct smc_sock *smc, struct msghdr *msg, size_t len) goto out_err; } + if (sk->sk_state == SMC_INIT) + return -ENOTCONN; + + if (len > conn->sndbuf_desc->len) + SMC_STAT_RMB_TX_SIZE_SMALL(!conn->lnk); + + if (len > conn->peer_rmbe_size) + SMC_STAT_RMB_TX_PEER_SIZE_SMALL(!conn->lnk); + + if (msg->msg_flags & MSG_OOB) + SMC_STAT_INC(!conn->lnk, urg_data_cnt); + while (msg_data_left(msg)) { - if (sk->sk_state == SMC_INIT) - return -ENOTCONN; if (smc->sk.sk_shutdown & SEND_SHUTDOWN || (smc->sk.sk_err == ECONNABORTED) || conn->killed) @@ -485,8 +498,10 @@ static int smc_tx_rdma_writes(struct smc_connection *conn, /* destination: RMBE */ /* cf. snd_wnd */ rmbespace = atomic_read(&conn->peer_rmbe_space); - if (rmbespace <= 0) + if (rmbespace <= 0) { + SMC_STAT_RMB_TX_PEER_FULL(!conn->lnk); return 0; + } smc_curs_copy(&prod, &conn->local_tx_ctrl.prod, conn); smc_curs_copy(&cons, &conn->local_rx_ctrl.cons, conn);