X-Git-Url: https://git.whamcloud.com/?p=fs%2Flustre-release.git;a=blobdiff_plain;f=lnet%2Fklnds%2Fsocklnd%2Fsocklnd.h;h=08bbab9386c2db4cd53cc0cd5ef00f088978280e;hp=052b9eefe88d275e0e24ff5f7f519a699c6b824f;hb=71b2476e4ddb95aa42f4a0ea3f23b1826017bfa5;hpb=aab1d832130ee5c181cf7e0e5aa555244d150b00 diff --git a/lnet/klnds/socklnd/socklnd.h b/lnet/klnds/socklnd/socklnd.h index 052b9ee..08bbab9 100644 --- a/lnet/klnds/socklnd/socklnd.h +++ b/lnet/klnds/socklnd/socklnd.h @@ -1,7 +1,7 @@ /* * Copyright (c) 2003, 2010, Oracle and/or its affiliates. All rights reserved. * - * Copyright (c) 2011, 2013, Intel Corporation. + * Copyright (c) 2011, 2017, Intel Corporation. * * Author: Zach Brown * Author: Peter J. Braam @@ -31,51 +31,49 @@ #define DEBUG_PORTAL_ALLOC #define DEBUG_SUBSYSTEM S_LND -#include #include #include #include #include #include +#include #include #include #include #include +#include +#include #include #include #include #include #include #include +#include #include #include -#include -#include #include #include -#ifdef HAVE_TCP_SENDPAGE_USE_SOCKET -# define cfs_tcp_sendpage(sk, page, offset, size, flags) \ - tcp_sendpage((sk)->sk_socket, page, offset, size, flags) -#else /* !HAVE_TCP_SENDPAGE_USE_SOCKET */ -# define cfs_tcp_sendpage(sk, page, offset, size, flags) \ - tcp_sendpage(sk, page, offset, size, flags) -#endif /* HAVE_TCP_SENDPAGE_USE_SOCKET */ +#include + +#ifndef NETIF_F_CSUM_MASK +# define NETIF_F_CSUM_MASK NETIF_F_ALL_CSUM +#endif /* assume one thread for each connection type */ #define SOCKNAL_NSCHEDS 3 #define SOCKNAL_NSCHEDS_HIGH (SOCKNAL_NSCHEDS << 1) -#define SOCKNAL_PEER_HASH_SIZE 101 /* # peer lists */ -#define SOCKNAL_RESCHED 100 /* # scheduler loops before reschedule */ -#define SOCKNAL_INSANITY_RECONN 5000 /* connd is trying on reconn infinitely */ -#define SOCKNAL_ENOMEM_RETRY CFS_TICK /* jiffies between retries */ +#define SOCKNAL_PEER_HASH_BITS 7 /* log2 of # peer_ni lists */ +#define SOCKNAL_INSANITY_RECONN 5000 /* connd is trying on reconn infinitely */ +#define SOCKNAL_ENOMEM_RETRY 1 /* seconds between retries */ -#define SOCKNAL_SINGLE_FRAG_TX 0 /* disable multi-fragment sends */ -#define SOCKNAL_SINGLE_FRAG_RX 0 /* disable multi-fragment receives */ +#define SOCKNAL_SINGLE_FRAG_TX 0 /* disable multi-fragment sends */ +#define SOCKNAL_SINGLE_FRAG_RX 0 /* disable multi-fragment receives */ -#define SOCKNAL_VERSION_DEBUG 0 /* enable protocol version debugging */ +#define SOCKNAL_VERSION_DEBUG 0 /* enable protocol version debugging */ /* risk kmap deadlock on multi-frag I/O (backs off to single-frag if disabled). * no risk if we're not running on a CONFIG_HIGHMEM platform. */ @@ -85,33 +83,25 @@ # define SOCKNAL_RISK_KMAP_DEADLOCK 1 #endif -struct ksock_sched_info; - -typedef struct /* per scheduler state */ -{ - spinlock_t kss_lock; /* serialise */ - struct list_head kss_rx_conns; /* conn waiting to be read */ +/* per scheduler state */ +struct ksock_sched { + /* serialise */ + spinlock_t kss_lock; /* conn waiting to be written */ - struct list_head kss_tx_conns; + struct list_head kss_rx_conns; + struct list_head kss_tx_conns; /* zombie noop tx list */ - struct list_head kss_zombie_noop_txs; - wait_queue_head_t kss_waitq; /* where scheduler sleeps */ + struct list_head kss_zombie_noop_txs; + /* where scheduler sleeps */ + wait_queue_head_t kss_waitq; /* # connections assigned to this scheduler */ - int kss_nconns; - struct ksock_sched_info *kss_info; /* owner of it */ -#if !SOCKNAL_SINGLE_FRAG_RX - struct page *kss_rx_scratch_pgs[LNET_MAX_IOV]; -#endif -#if !SOCKNAL_SINGLE_FRAG_TX || !SOCKNAL_SINGLE_FRAG_RX - struct iovec kss_scratch_iov[LNET_MAX_IOV]; -#endif -} ksock_sched_t; - -struct ksock_sched_info { - int ksi_nthreads_max; /* max allowed threads */ - int ksi_nthreads; /* number of threads */ - int ksi_cpt; /* CPT id */ - ksock_sched_t *ksi_scheds; /* array of schedulers */ + int kss_nconns; + /* max allowed threads */ + int kss_nthreads_max; + /* number of threads */ + int kss_nthreads; + /* CPT id */ + int kss_cpt; }; #define KSOCK_CPT_SHIFT 16 @@ -119,17 +109,16 @@ struct ksock_sched_info { #define KSOCK_THREAD_CPT(id) ((id) >> KSOCK_CPT_SHIFT) #define KSOCK_THREAD_SID(id) ((id) & ((1UL << KSOCK_CPT_SHIFT) - 1)) -typedef struct /* in-use interface */ -{ - __u32 ksni_ipaddr; /* interface's IP address */ +struct ksock_interface { /* in-use interface */ + int ksni_index; /* Linux interface index */ + struct sockaddr_storage ksni_addr; /* interface's address */ __u32 ksni_netmask; /* interface's network mask */ int ksni_nroutes; /* # routes using (active) */ int ksni_npeers; /* # peers using (passive) */ char ksni_name[IFNAMSIZ]; /* interface name */ -} ksock_interface_t; +}; -typedef struct -{ +struct ksock_tunables { /* "stuck" socket timeout (seconds) */ int *ksnd_timeout; /* # scheduler threads in each pool while starting */ @@ -150,18 +139,16 @@ typedef struct int *ksnd_keepalive_count; /* # probes */ int *ksnd_keepalive_intvl; /* time between probes */ int *ksnd_credits; /* # concurrent sends */ - int *ksnd_peertxcredits; /* # concurrent sends to 1 peer */ - int *ksnd_peerrtrcredits; /* # per-peer router buffer credits */ - int *ksnd_peertimeout; /* seconds to consider peer dead */ + int *ksnd_peertxcredits; /* # concurrent sends to 1 peer_ni */ + int *ksnd_peerrtrcredits; /* # per-peer_ni router buffer credits */ + int *ksnd_peertimeout; /* seconds to consider peer_ni dead */ int *ksnd_enable_csum; /* enable check sum */ int *ksnd_inject_csum_error; /* set non-zero to inject checksum error */ int *ksnd_nonblk_zcack; /* always send zc-ack on non-blocking connection */ unsigned int *ksnd_zc_min_payload; /* minimum zero copy payload size */ int *ksnd_zc_recv; /* enable ZC receive (for Chelsio TOE) */ int *ksnd_zc_recv_min_nfrags; /* minimum # of fragments to enable ZC receive */ -#ifdef CPU_AFFINITY int *ksnd_irq_affinity; /* enable IRQ affinity? */ -#endif #ifdef SOCKNAL_BACKOFF int *ksnd_backoff_init; /* initial TCP backoff */ int *ksnd_backoff_max; /* maximum TCP backoff */ @@ -169,42 +156,42 @@ typedef struct #if SOCKNAL_VERSION_DEBUG int *ksnd_protocol; /* protocol version */ #endif -#if defined(CONFIG_SYSCTL) && !CFS_SYSFS_MODULE_PARM - struct ctl_table_header *ksnd_sysctl; /* sysctl interface */ -#endif -} ksock_tunables_t; + int *ksnd_conns_per_peer; /* for typed mode, yields: + * 1 + 2*conns_per_peer total + * for untyped: + * conns_per_peer total + */ +}; -typedef struct -{ +struct ksock_net { __u64 ksnn_incarnation; /* my epoch */ - spinlock_t ksnn_lock; /* serialise */ struct list_head ksnn_list; /* chain on global list */ - int ksnn_npeers; /* # peers */ - int ksnn_shutdown; /* shutting down? */ - int ksnn_ninterfaces; /* IP interfaces */ - ksock_interface_t ksnn_interfaces[LNET_MAX_INTERFACES]; -} ksock_net_t; + atomic_t ksnn_npeers; /* # peers */ + struct ksock_interface ksnn_interface; /* IP interface */ +}; +/* When the ksock_net is shut down, this (negative) bias is added to + * ksnn_npeers, which prevents new peers from being added. + */ +#define SOCKNAL_SHUTDOWN_BIAS (INT_MIN+1) /** connd timeout */ #define SOCKNAL_CONND_TIMEOUT 120 /** reserved thread for accepting & creating new connd */ #define SOCKNAL_CONND_RESV 1 -typedef struct -{ +struct ksock_nal_data { int ksnd_init; /* initialisation state */ int ksnd_nnets; /* # networks set up */ struct list_head ksnd_nets; /* list of nets */ - /* stabilize peer/conn ops */ + /* stabilize peer_ni/conn ops */ rwlock_t ksnd_global_lock; /* hash table of all my known peers */ - struct list_head *ksnd_peers; - int ksnd_peer_hash_size; /* size of ksnd_peers */ + DECLARE_HASHTABLE(ksnd_peers, SOCKNAL_PEER_HASH_BITS); - int ksnd_nthreads; /* # live threads */ + atomic_t ksnd_nthreads; /* # live threads */ int ksnd_shuttingdown; /* tell threads to exit */ /* schedulers information */ - struct ksock_sched_info **ksnd_sched_info; + struct ksock_sched **ksnd_schedulers; atomic_t ksnd_nactive_txs; /* #active txs */ @@ -217,7 +204,7 @@ typedef struct /* reaper sleeps here */ wait_queue_head_t ksnd_reaper_waitq; /* when reaper will wake */ - cfs_time_t ksnd_reaper_waketime; + time64_t ksnd_reaper_waketime; /* serialise */ spinlock_t ksnd_reaper_lock; @@ -234,11 +221,11 @@ typedef struct /* # connds connecting */ int ksnd_connd_connecting; /** time stamp of the last failed connecting attempt */ - long ksnd_connd_failed_stamp; + time64_t ksnd_connd_failed_stamp; /** # starting connd */ unsigned ksnd_connd_starting; /** time stamp of the last starting connd */ - long ksnd_connd_starting_stamp; + time64_t ksnd_connd_starting_stamp; /** # running connd */ unsigned ksnd_connd_running; /* serialise */ @@ -248,68 +235,55 @@ typedef struct struct list_head ksnd_idle_noop_txs; /* serialise, g_lock unsafe */ spinlock_t ksnd_tx_lock; - -} ksock_nal_data_t; +}; #define SOCKNAL_INIT_NOTHING 0 #define SOCKNAL_INIT_DATA 1 #define SOCKNAL_INIT_ALL 2 -/* A packet just assembled for transmission is represented by 1 or more - * struct iovec fragments (the first frag contains the portals header), - * followed by 0 or more lnet_kiov_t fragments. +/* A packet just assembled for transmission is represented by 1 + * struct iovec fragment - the portals header - followed by 0 + * or more struct bio_vec fragments. * - * On the receive side, initially 1 struct iovec fragment is posted for + * On the receive side, initially 1 struct kvec fragment is posted for * receive (the header). Once the header has been received, the payload is - * received into either struct iovec or lnet_kiov_t fragments, depending on - * what the header matched or whether the message needs forwarding. */ - -struct ksock_conn; /* forward ref */ -struct ksock_peer; /* forward ref */ -struct ksock_route; /* forward ref */ -struct ksock_proto; /* forward ref */ + * received into struct bio_vec fragments. + */ +struct ksock_conn; /* forward ref */ +struct ksock_conn_cb; /* forward ref */ +struct ksock_proto; /* forward ref */ + +struct ksock_tx { /* transmit packet */ + struct list_head tx_list; /* queue on conn for transmission etc */ + struct list_head tx_zc_list; /* queue on peer_ni for ZC request */ + refcount_t tx_refcount; /* tx reference count */ + int tx_nob; /* # packet bytes */ + int tx_resid; /* residual bytes */ + int tx_niov; /* # packet kvec frags */ + int tx_nkiov; /* # packet page frags */ + unsigned short tx_zc_aborted; /* aborted ZC request */ + unsigned short tx_zc_capable:1; /* payload is large enough for ZC */ + unsigned short tx_zc_checked:1; /* Have I checked if I should ZC? */ + unsigned short tx_nonblk:1; /* it's a non-blocking ACK */ + struct bio_vec *tx_kiov; /* packet page frags */ + struct ksock_conn *tx_conn; /* owning conn */ + struct lnet_msg *tx_lnetmsg; /* lnet message for lnet_finalize() */ + time64_t tx_deadline; /* when (in secs) tx times out */ + struct ksock_msg tx_msg; /* socklnd message buffer */ + int tx_desc_size; /* size of this descriptor */ + enum lnet_msg_hstatus tx_hstatus; /* health status of tx */ + struct kvec tx_hdr; /* virt hdr */ + struct bio_vec tx_payload[0]; /* paged payload */ +}; -typedef struct /* transmit packet */ -{ - struct list_head tx_list; /* queue on conn for transmission etc */ - struct list_head tx_zc_list; /* queue on peer for ZC request */ - atomic_t tx_refcount; /* tx reference count */ - int tx_nob; /* # packet bytes */ - int tx_resid; /* residual bytes */ - int tx_niov; /* # packet iovec frags */ - struct iovec *tx_iov; /* packet iovec frags */ - int tx_nkiov; /* # packet page frags */ - unsigned short tx_zc_aborted; /* aborted ZC request */ - unsigned short tx_zc_capable:1; /* payload is large enough for ZC */ - unsigned short tx_zc_checked:1; /* Have I checked if I should ZC? */ - unsigned short tx_nonblk:1; /* it's a non-blocking ACK */ - lnet_kiov_t *tx_kiov; /* packet page frags */ - struct ksock_conn *tx_conn; /* owning conn */ - lnet_msg_t *tx_lnetmsg; /* lnet message for lnet_finalize() */ - cfs_time_t tx_deadline; /* when (in jiffies) tx times out */ - ksock_msg_t tx_msg; /* socklnd message buffer */ - int tx_desc_size; /* size of this descriptor */ - union { - struct { - struct iovec iov; /* virt hdr */ - lnet_kiov_t kiov[0]; /* paged payload */ - } paged; - struct { - struct iovec iov[1]; /* virt hdr + payload */ - } virt; - } tx_frags; -} ksock_tx_t; - -#define KSOCK_NOOP_TX_SIZE ((int)offsetof(ksock_tx_t, tx_frags.paged.kiov[0])) - -/* network zero copy callback descriptor embedded in ksock_tx_t */ +#define KSOCK_NOOP_TX_SIZE ((int)offsetof(struct ksock_tx, tx_payload[0])) /* space for the rx frag descriptors; we either read a single contiguous * header, or up to LNET_MAX_IOV frags of payload of either type. */ -typedef union { - struct iovec iov[LNET_MAX_IOV]; - lnet_kiov_t kiov[LNET_MAX_IOV]; -} ksock_rxiovspace_t; +union ksock_rxiovspace { + struct kvec iov[LNET_MAX_IOV]; + struct bio_vec kiov[LNET_MAX_IOV]; +}; #define SOCKNAL_RX_KSM_HEADER 1 /* reading ksock message header */ #define SOCKNAL_RX_LNET_HEADER 2 /* reading lnet message header */ @@ -318,61 +292,64 @@ typedef union { #define SOCKNAL_RX_LNET_PAYLOAD 5 /* reading lnet payload (to deliver here) */ #define SOCKNAL_RX_SLOP 6 /* skipping body */ -typedef struct ksock_conn -{ - struct ksock_peer *ksnc_peer; /* owning peer */ - struct ksock_route *ksnc_route; /* owning route */ - struct list_head ksnc_list; /* stash on peer's conn list */ - cfs_socket_t *ksnc_sock; /* actual socket */ - void *ksnc_saved_data_ready; /* socket's original data_ready() callback */ - void *ksnc_saved_write_space; /* socket's original write_space() callback */ - atomic_t ksnc_conn_refcount; /* conn refcount */ - atomic_t ksnc_sock_refcount; /* sock refcount */ - ksock_sched_t *ksnc_scheduler; /* who schedules this connection */ - __u32 ksnc_myipaddr; /* my IP */ - __u32 ksnc_ipaddr; /* peer's IP */ - int ksnc_port; /* peer's port */ - signed int ksnc_type:3; /* type of connection, - * should be signed value */ - unsigned int ksnc_closing:1; /* being shut down */ - unsigned int ksnc_flip:1; /* flip or not, only for V2.x */ - unsigned int ksnc_zc_capable:1; /* enable to ZC */ - struct ksock_proto *ksnc_proto; /* protocol for the connection */ +struct ksock_conn { + struct ksock_peer_ni *ksnc_peer; /* owning peer_ni */ + struct ksock_conn_cb *ksnc_conn_cb; /* owning conn control block */ + struct list_head ksnc_list; /* on peer_ni's conn list */ + struct socket *ksnc_sock; /* actual socket */ + void *ksnc_saved_data_ready; /* socket's original + * data_ready() cb */ + void *ksnc_saved_write_space; /* socket's original + * write_space() cb */ + refcount_t ksnc_conn_refcount; /* conn refcount */ + refcount_t ksnc_sock_refcount; /* sock refcount */ + struct ksock_sched *ksnc_scheduler; /* who schedules this + * connection */ + struct sockaddr_storage ksnc_myaddr; /* my address */ + struct sockaddr_storage ksnc_peeraddr; /* peer_ni's address */ + signed int ksnc_type:3; /* type of connection, + * should be signed + * value */ + unsigned int ksnc_closing:1; /* being shut down */ + unsigned int ksnc_flip:1; /* flip or not, only for V2.x */ + unsigned int ksnc_zc_capable:1; /* enable to ZC */ + const struct ksock_proto *ksnc_proto; /* protocol for the connection */ /* READER */ /* where I enq waiting input or a forwarding descriptor */ struct list_head ksnc_rx_list; - cfs_time_t ksnc_rx_deadline; /* when (in jiffies) receive times out */ + time64_t ksnc_rx_deadline; /* when (in seconds) receive times out */ __u8 ksnc_rx_started; /* started receiving a message */ __u8 ksnc_rx_ready; /* data ready to read */ __u8 ksnc_rx_scheduled;/* being progressed */ __u8 ksnc_rx_state; /* what is being read */ int ksnc_rx_nob_left; /* # bytes to next hdr/body */ int ksnc_rx_nob_wanted; /* bytes actually wanted */ - int ksnc_rx_niov; /* # iovec frags */ - struct iovec *ksnc_rx_iov; /* the iovec frags */ - int ksnc_rx_nkiov; /* # page frags */ - lnet_kiov_t *ksnc_rx_kiov; /* the page frags */ - ksock_rxiovspace_t ksnc_rx_iov_space;/* space for frag descriptors */ - __u32 ksnc_rx_csum; /* partial checksum for incoming data */ - void *ksnc_cookie; /* rx lnet_finalize passthru arg */ - ksock_msg_t ksnc_msg; /* incoming message buffer: - * V2.x message takes the - * whole struct - * V1.x message is a bare - * lnet_hdr_t, it's stored in - * ksnc_msg.ksm_u.lnetmsg */ - + int ksnc_rx_niov; /* # kvec frags */ + struct kvec *ksnc_rx_iov; /* the kvec frags */ + int ksnc_rx_nkiov; /* # page frags */ + struct bio_vec *ksnc_rx_kiov; /* the page frags */ + union ksock_rxiovspace ksnc_rx_iov_space;/* space for frag descriptors */ + __u32 ksnc_rx_csum; /* partial checksum for incoming + * data */ + struct lnet_msg *ksnc_lnet_msg; /* rx lnet_finalize arg*/ + struct ksock_msg ksnc_msg; /* incoming message buffer: + * V2.x message takes the + * whole struct + * V1.x message is a bare + * struct lnet_hdr, it's stored + * in ksnc_msg.ksm_u.lnetmsg + */ /* -- WRITER -- */ /* where I enq waiting for output space */ struct list_head ksnc_tx_list; /* packets waiting to be sent */ struct list_head ksnc_tx_queue; /* next TX that can carry a LNet message or ZC-ACK */ - ksock_tx_t *ksnc_tx_carrier; - /* when (in jiffies) tx times out */ - cfs_time_t ksnc_tx_deadline; + struct ksock_tx *ksnc_tx_carrier; + /* when (in seconds) tx times out */ + time64_t ksnc_tx_deadline; /* send buffer marker */ int ksnc_tx_bufnob; /* # bytes queued */ @@ -382,93 +359,93 @@ typedef struct ksock_conn /* being progressed */ int ksnc_tx_scheduled; /* time stamp of the last posted TX */ - cfs_time_t ksnc_tx_last_post; -} ksock_conn_t; + time64_t ksnc_tx_last_post; +}; -typedef struct ksock_route -{ - struct list_head ksnr_list; /* chain on peer route list */ - struct list_head ksnr_connd_list; /* chain on ksnr_connd_routes */ - struct ksock_peer *ksnr_peer; /* owning peer */ - atomic_t ksnr_refcount; /* # users */ - cfs_time_t ksnr_timeout; /* when (in jiffies) reconnection can happen next */ - cfs_duration_t ksnr_retry_interval; /* how long between retries */ - __u32 ksnr_myipaddr; /* my IP */ - __u32 ksnr_ipaddr; /* IP address to connect to */ - int ksnr_port; /* port to connect to */ - unsigned int ksnr_scheduled:1; /* scheduled for attention */ - unsigned int ksnr_connecting:1;/* connection establishment in progress */ - unsigned int ksnr_connected:4; /* connections established by type */ - unsigned int ksnr_deleted:1; /* been removed from peer? */ - unsigned int ksnr_share_count; /* created explicitly? */ - int ksnr_conn_count; /* # conns established by this route */ -} ksock_route_t; +#define SOCKNAL_CONN_COUNT_MAX_BITS 8 /* max conn count bits */ + +struct ksock_conn_cb { + struct list_head ksnr_connd_list;/* chain on ksnr_connd_routes */ + struct ksock_peer_ni *ksnr_peer; /* owning peer_ni */ + refcount_t ksnr_refcount; /* # users */ + time64_t ksnr_timeout; /* when (in secs) reconnection + * can happen next + */ + time64_t ksnr_retry_interval;/* secs between retries */ + int ksnr_myiface; /* interface index */ + struct sockaddr_storage ksnr_addr; /* IP address to connect to */ + unsigned int ksnr_scheduled:1;/* scheduled for attention */ + unsigned int ksnr_connecting:1;/* connection in progress */ + unsigned int ksnr_connected:4;/* connections by type */ + unsigned int ksnr_deleted:1; /* been removed from peer_ni? */ + unsigned int ksnr_ctrl_conn_count:1; /* # conns by type */ + unsigned int ksnr_blki_conn_count:8; + unsigned int ksnr_blko_conn_count:8; + int ksnr_conn_count;/* total # conns for this cb */ + +}; #define SOCKNAL_KEEPALIVE_PING 1 /* cookie for keepalive ping */ -typedef struct ksock_peer -{ - struct list_head ksnp_list; /* stash on global peer list */ - cfs_time_t ksnp_last_alive; /* when (in jiffies) I was last alive */ - lnet_process_id_t ksnp_id; /* who's on the other end(s) */ - atomic_t ksnp_refcount; /* # users */ - int ksnp_sharecount; /* lconf usage counter */ - int ksnp_closing; /* being closed */ - int ksnp_accepting;/* # passive connections pending */ - int ksnp_error; /* errno on closing last conn */ - __u64 ksnp_zc_next_cookie;/* ZC completion cookie */ - __u64 ksnp_incarnation; /* latest known peer incarnation */ - struct ksock_proto *ksnp_proto; /* latest known peer protocol */ +struct ksock_peer_ni { + struct hlist_node ksnp_list; /* stash on global peer_ni list */ + time64_t ksnp_last_alive;/* when (in seconds) I was last alive */ + struct lnet_process_id ksnp_id; /* who's on the other end(s) */ + refcount_t ksnp_refcount; /* # users */ + int ksnp_closing; /* being closed */ + int ksnp_accepting; /* # passive connections pending */ + int ksnp_error; /* errno on closing last conn */ + __u64 ksnp_zc_next_cookie;/* ZC completion cookie */ + __u64 ksnp_incarnation; /* latest known peer_ni incarnation */ + const struct ksock_proto *ksnp_proto; /* latest known protocol */ struct list_head ksnp_conns; /* all active connections */ - struct list_head ksnp_routes; /* routes */ + struct ksock_conn_cb *ksnp_conn_cb; /* conn control block */ struct list_head ksnp_tx_queue; /* waiting packets */ - spinlock_t ksnp_lock; /* serialize, g_lock unsafe */ + spinlock_t ksnp_lock; /* serialize, g_lock unsafe */ /* zero copy requests wait for ACK */ struct list_head ksnp_zc_req_list; - cfs_time_t ksnp_send_keepalive; /* time to send keepalive */ - lnet_ni_t *ksnp_ni; /* which network */ - int ksnp_n_passive_ips; /* # of... */ - __u32 ksnp_passive_ips[LNET_MAX_INTERFACES]; /* preferred local interfaces */ -} ksock_peer_t; + time64_t ksnp_send_keepalive; /* time to send keepalive */ + struct lnet_ni *ksnp_ni; /* which network */ + int ksnp_n_passive_ips; /* # of... */ + __u32 ksnp_passive_ips[LNET_INTERFACES_NUM]; /* preferred local interfaces */ +}; -typedef struct ksock_connreq -{ +struct ksock_connreq { /* stash on ksnd_connd_connreqs */ struct list_head ksncr_list; /* chosen NI */ - lnet_ni_t *ksncr_ni; + struct lnet_ni *ksncr_ni; /* accepted socket */ - cfs_socket_t *ksncr_sock; -} ksock_connreq_t; + struct socket *ksncr_sock; +}; -extern ksock_nal_data_t ksocknal_data; -extern ksock_tunables_t ksocknal_tunables; +extern struct ksock_nal_data ksocknal_data; +extern struct ksock_tunables ksocknal_tunables; #define SOCKNAL_MATCH_NO 0 /* TX can't match type of connection */ #define SOCKNAL_MATCH_YES 1 /* TX matches type of connection */ #define SOCKNAL_MATCH_MAY 2 /* TX can be sent on the connection, but not preferred */ -typedef struct ksock_proto -{ +struct ksock_proto { int pro_version; /* version number of protocol */ - int (*pro_send_hello)(ksock_conn_t *, ksock_hello_msg_t *); /* handshake function */ - int (*pro_recv_hello)(ksock_conn_t *, ksock_hello_msg_t *, int);/* handshake function */ - void (*pro_pack)(ksock_tx_t *); /* message pack */ - void (*pro_unpack)(ksock_msg_t *); /* message unpack */ - ksock_tx_t *(*pro_queue_tx_msg)(ksock_conn_t *, ksock_tx_t *); /* queue tx on the connection */ - int (*pro_queue_tx_zcack)(ksock_conn_t *, ksock_tx_t *, __u64); /* queue ZC ack on the connection */ - int (*pro_handle_zcreq)(ksock_conn_t *, __u64, int); /* handle ZC request */ - int (*pro_handle_zcack)(ksock_conn_t *, __u64, __u64); /* handle ZC ACK */ - int (*pro_match_tx)(ksock_conn_t *, ksock_tx_t *, int); /* msg type matches the connection type: + int (*pro_send_hello)(struct ksock_conn *, struct ksock_hello_msg *); /* handshake function */ + int (*pro_recv_hello)(struct ksock_conn *, struct ksock_hello_msg *, int);/* handshake function */ + void (*pro_pack)(struct ksock_tx *); /* message pack */ + void (*pro_unpack)(struct ksock_msg *); /* message unpack */ + struct ksock_tx *(*pro_queue_tx_msg)(struct ksock_conn *, struct ksock_tx *); /* queue tx on the connection */ + int (*pro_queue_tx_zcack)(struct ksock_conn *, struct ksock_tx *, __u64); /* queue ZC ack on the connection */ + int (*pro_handle_zcreq)(struct ksock_conn *, __u64, int); /* handle ZC request */ + int (*pro_handle_zcack)(struct ksock_conn *, __u64, __u64); /* handle ZC ACK */ + int (*pro_match_tx)(struct ksock_conn *, struct ksock_tx *, int); /* msg type matches the connection type: * return value: * return MATCH_NO : no * return MATCH_YES : matching type * return MATCH_MAY : can be backup */ -} ksock_proto_t; +}; -extern ksock_proto_t ksocknal_protocol_v1x; -extern ksock_proto_t ksocknal_protocol_v2x; -extern ksock_proto_t ksocknal_protocol_v3x; +extern const struct ksock_proto ksocknal_protocol_v1x; +extern const struct ksock_proto ksocknal_protocol_v2x; +extern const struct ksock_proto ksocknal_protocol_v3x; #define KSOCK_PROTO_V1_MAJOR LNET_PROTO_TCP_VERSION_MAJOR #define KSOCK_PROTO_V1_MINOR LNET_PROTO_TCP_VERSION_MINOR @@ -491,51 +468,40 @@ static inline __u32 ksocknal_csum(__u32 crc, unsigned char const *p, size_t len) } static inline int -ksocknal_route_mask(void) -{ - if (!*ksocknal_tunables.ksnd_typed_conns) - return (1 << SOCKLND_CONN_ANY); - - return ((1 << SOCKLND_CONN_CONTROL) | - (1 << SOCKLND_CONN_BULK_IN) | - (1 << SOCKLND_CONN_BULK_OUT)); -} - -static inline struct list_head * -ksocknal_nid2peerlist (lnet_nid_t nid) +ksocknal_conn_cb_mask(void) { - unsigned int hash = ((unsigned int)nid) % ksocknal_data.ksnd_peer_hash_size; + if (!*ksocknal_tunables.ksnd_typed_conns) + return BIT(SOCKLND_CONN_ANY); - return (&ksocknal_data.ksnd_peers [hash]); + return (BIT(SOCKLND_CONN_CONTROL) | + BIT(SOCKLND_CONN_BULK_IN) | + BIT(SOCKLND_CONN_BULK_OUT)); } static inline void -ksocknal_conn_addref (ksock_conn_t *conn) +ksocknal_conn_addref(struct ksock_conn *conn) { - LASSERT (atomic_read(&conn->ksnc_conn_refcount) > 0); - atomic_inc(&conn->ksnc_conn_refcount); + refcount_inc(&conn->ksnc_conn_refcount); } -extern void ksocknal_queue_zombie_conn (ksock_conn_t *conn); -extern void ksocknal_finalize_zcreq(ksock_conn_t *conn); +extern void ksocknal_queue_zombie_conn(struct ksock_conn *conn); +extern void ksocknal_finalize_zcreq(struct ksock_conn *conn); static inline void -ksocknal_conn_decref (ksock_conn_t *conn) +ksocknal_conn_decref(struct ksock_conn *conn) { - LASSERT (atomic_read(&conn->ksnc_conn_refcount) > 0); - if (atomic_dec_and_test(&conn->ksnc_conn_refcount)) + if (refcount_dec_and_test(&conn->ksnc_conn_refcount)) ksocknal_queue_zombie_conn(conn); } static inline int -ksocknal_connsock_addref (ksock_conn_t *conn) +ksocknal_connsock_addref(struct ksock_conn *conn) { - int rc = -ESHUTDOWN; + int rc = -ESHUTDOWN; read_lock(&ksocknal_data.ksnd_global_lock); if (!conn->ksnc_closing) { - LASSERT(atomic_read(&conn->ksnc_sock_refcount) > 0); - atomic_inc(&conn->ksnc_sock_refcount); + refcount_inc(&conn->ksnc_sock_refcount); rc = 0; } read_unlock(&ksocknal_data.ksnd_global_lock); @@ -544,146 +510,153 @@ ksocknal_connsock_addref (ksock_conn_t *conn) } static inline void -ksocknal_connsock_decref (ksock_conn_t *conn) +ksocknal_connsock_decref(struct ksock_conn *conn) { - LASSERT (atomic_read(&conn->ksnc_sock_refcount) > 0); - if (atomic_dec_and_test(&conn->ksnc_sock_refcount)) { + if (refcount_dec_and_test(&conn->ksnc_sock_refcount)) { LASSERT (conn->ksnc_closing); - libcfs_sock_release(conn->ksnc_sock); + sock_release(conn->ksnc_sock); conn->ksnc_sock = NULL; ksocknal_finalize_zcreq(conn); } } static inline void -ksocknal_tx_addref (ksock_tx_t *tx) +ksocknal_tx_addref(struct ksock_tx *tx) { - LASSERT (atomic_read(&tx->tx_refcount) > 0); - atomic_inc(&tx->tx_refcount); + refcount_inc(&tx->tx_refcount); } -extern void ksocknal_tx_prep (ksock_conn_t *, ksock_tx_t *tx); -extern void ksocknal_tx_done (lnet_ni_t *ni, ksock_tx_t *tx); +extern void ksocknal_tx_prep(struct ksock_conn *, struct ksock_tx *tx); +extern void ksocknal_tx_done(struct lnet_ni *ni, struct ksock_tx *tx, int error); static inline void -ksocknal_tx_decref (ksock_tx_t *tx) +ksocknal_tx_decref(struct ksock_tx *tx) { - LASSERT (atomic_read(&tx->tx_refcount) > 0); - if (atomic_dec_and_test(&tx->tx_refcount)) - ksocknal_tx_done(NULL, tx); + if (refcount_dec_and_test(&tx->tx_refcount)) + ksocknal_tx_done(NULL, tx, 0); } static inline void -ksocknal_route_addref (ksock_route_t *route) +ksocknal_conn_cb_addref(struct ksock_conn_cb *conn_cb) { - LASSERT (atomic_read(&route->ksnr_refcount) > 0); - atomic_inc(&route->ksnr_refcount); + refcount_inc(&conn_cb->ksnr_refcount); } -extern void ksocknal_destroy_route (ksock_route_t *route); +extern void ksocknal_destroy_conn_cb(struct ksock_conn_cb *conn_cb); static inline void -ksocknal_route_decref (ksock_route_t *route) +ksocknal_conn_cb_decref(struct ksock_conn_cb *conn_cb) { - LASSERT (atomic_read (&route->ksnr_refcount) > 0); - if (atomic_dec_and_test(&route->ksnr_refcount)) - ksocknal_destroy_route (route); + if (refcount_dec_and_test(&conn_cb->ksnr_refcount)) + ksocknal_destroy_conn_cb(conn_cb); } static inline void -ksocknal_peer_addref (ksock_peer_t *peer) +ksocknal_peer_addref(struct ksock_peer_ni *peer_ni) { - LASSERT (atomic_read (&peer->ksnp_refcount) > 0); - atomic_inc(&peer->ksnp_refcount); + refcount_inc(&peer_ni->ksnp_refcount); } -extern void ksocknal_destroy_peer (ksock_peer_t *peer); +extern void ksocknal_destroy_peer(struct ksock_peer_ni *peer_ni); static inline void -ksocknal_peer_decref (ksock_peer_t *peer) +ksocknal_peer_decref(struct ksock_peer_ni *peer_ni) +{ + if (refcount_dec_and_test(&peer_ni->ksnp_refcount)) + ksocknal_destroy_peer(peer_ni); +} + +static inline int ksocknal_timeout(void) +{ + return *ksocknal_tunables.ksnd_timeout ?: lnet_get_lnd_timeout(); +} + +static inline int ksocknal_conns_per_peer(void) { - LASSERT (atomic_read (&peer->ksnp_refcount) > 0); - if (atomic_dec_and_test(&peer->ksnp_refcount)) - ksocknal_destroy_peer (peer); + return *ksocknal_tunables.ksnd_conns_per_peer ?: 1; } -int ksocknal_startup (lnet_ni_t *ni); -void ksocknal_shutdown (lnet_ni_t *ni); -int ksocknal_ctl(lnet_ni_t *ni, unsigned int cmd, void *arg); -int ksocknal_send (lnet_ni_t *ni, void *private, lnet_msg_t *lntmsg); -int ksocknal_recv(lnet_ni_t *ni, void *private, lnet_msg_t *lntmsg, - int delayed, unsigned int niov, - struct iovec *iov, lnet_kiov_t *kiov, +int ksocknal_startup(struct lnet_ni *ni); +void ksocknal_shutdown(struct lnet_ni *ni); +int ksocknal_ctl(struct lnet_ni *ni, unsigned int cmd, void *arg); +int ksocknal_send(struct lnet_ni *ni, void *private, struct lnet_msg *lntmsg); +int ksocknal_recv(struct lnet_ni *ni, void *private, struct lnet_msg *lntmsg, + int delayed, unsigned int niov, + struct bio_vec *kiov, unsigned int offset, unsigned int mlen, unsigned int rlen); -int ksocknal_accept(lnet_ni_t *ni, cfs_socket_t *sock); - -extern int ksocknal_add_peer(lnet_ni_t *ni, lnet_process_id_t id, __u32 ip, int port); -extern ksock_peer_t *ksocknal_find_peer_locked (lnet_ni_t *ni, lnet_process_id_t id); -extern ksock_peer_t *ksocknal_find_peer (lnet_ni_t *ni, lnet_process_id_t id); -extern void ksocknal_peer_failed (ksock_peer_t *peer); -extern int ksocknal_create_conn (lnet_ni_t *ni, ksock_route_t *route, - cfs_socket_t *sock, int type); -extern void ksocknal_close_conn_locked (ksock_conn_t *conn, int why); -extern void ksocknal_terminate_conn (ksock_conn_t *conn); -extern void ksocknal_destroy_conn (ksock_conn_t *conn); -extern int ksocknal_close_peer_conns_locked (ksock_peer_t *peer, - __u32 ipaddr, int why); -extern int ksocknal_close_conn_and_siblings (ksock_conn_t *conn, int why); -extern int ksocknal_close_matching_conns(lnet_process_id_t id, __u32 ipaddr); -extern ksock_conn_t *ksocknal_find_conn_locked(ksock_peer_t *peer, - ksock_tx_t *tx, int nonblk); - -extern int ksocknal_launch_packet(lnet_ni_t *ni, ksock_tx_t *tx, - lnet_process_id_t id); -extern ksock_tx_t *ksocknal_alloc_tx(int type, int size); -extern void ksocknal_free_tx (ksock_tx_t *tx); -extern ksock_tx_t *ksocknal_alloc_tx_noop(__u64 cookie, int nonblk); -extern void ksocknal_next_tx_carrier(ksock_conn_t *conn); -extern void ksocknal_queue_tx_locked (ksock_tx_t *tx, ksock_conn_t *conn); -extern void ksocknal_txlist_done(lnet_ni_t *ni, struct list_head *txlist, +int ksocknal_accept(struct lnet_ni *ni, struct socket *sock); + +int ksocknal_add_peer(struct lnet_ni *ni, struct lnet_process_id id, + struct sockaddr *addr); +struct ksock_peer_ni *ksocknal_find_peer_locked(struct lnet_ni *ni, + struct lnet_process_id id); +struct ksock_peer_ni *ksocknal_find_peer(struct lnet_ni *ni, + struct lnet_process_id id); +extern void ksocknal_peer_failed(struct ksock_peer_ni *peer_ni); +extern int ksocknal_create_conn(struct lnet_ni *ni, + struct ksock_conn_cb *conn_cb, + struct socket *sock, int type); +extern void ksocknal_close_conn_locked(struct ksock_conn *conn, int why); +extern void ksocknal_terminate_conn(struct ksock_conn *conn); +extern void ksocknal_destroy_conn(struct ksock_conn *conn); +extern int ksocknal_close_peer_conns_locked(struct ksock_peer_ni *peer_ni, + struct sockaddr *peer, int why); +extern int ksocknal_close_conn_and_siblings(struct ksock_conn *conn, int why); +int ksocknal_close_matching_conns(struct lnet_process_id id, __u32 ipaddr); +extern struct ksock_conn *ksocknal_find_conn_locked(struct ksock_peer_ni *peer_ni, + struct ksock_tx *tx, int nonblk); + +extern int ksocknal_launch_packet(struct lnet_ni *ni, struct ksock_tx *tx, + struct lnet_process_id id); +extern struct ksock_tx *ksocknal_alloc_tx(int type, int size); +extern void ksocknal_free_tx(struct ksock_tx *tx); +extern struct ksock_tx *ksocknal_alloc_tx_noop(__u64 cookie, int nonblk); +extern void ksocknal_next_tx_carrier(struct ksock_conn *conn); +extern void ksocknal_queue_tx_locked(struct ksock_tx *tx, struct ksock_conn *conn); +extern void ksocknal_txlist_done(struct lnet_ni *ni, struct list_head *txlist, int error); -extern void ksocknal_notify (lnet_ni_t *ni, lnet_nid_t gw_nid, int alive); -extern void ksocknal_query (struct lnet_ni *ni, lnet_nid_t nid, cfs_time_t *when); extern int ksocknal_thread_start(int (*fn)(void *arg), void *arg, char *name); -extern void ksocknal_thread_fini (void); -extern void ksocknal_launch_all_connections_locked (ksock_peer_t *peer); -extern ksock_route_t *ksocknal_find_connectable_route_locked (ksock_peer_t *peer); -extern ksock_route_t *ksocknal_find_connecting_route_locked (ksock_peer_t *peer); -extern int ksocknal_new_packet (ksock_conn_t *conn, int skip); -extern int ksocknal_scheduler (void *arg); -extern int ksocknal_connd (void *arg); -extern int ksocknal_reaper (void *arg); -extern int ksocknal_send_hello (lnet_ni_t *ni, ksock_conn_t *conn, - lnet_nid_t peer_nid, ksock_hello_msg_t *hello); -extern int ksocknal_recv_hello (lnet_ni_t *ni, ksock_conn_t *conn, - ksock_hello_msg_t *hello, lnet_process_id_t *id, - __u64 *incarnation); -extern void ksocknal_read_callback(ksock_conn_t *conn); -extern void ksocknal_write_callback(ksock_conn_t *conn); - -extern int ksocknal_lib_zc_capable(ksock_conn_t *conn); -extern void ksocknal_lib_save_callback(cfs_socket_t *sock, ksock_conn_t *conn); -extern void ksocknal_lib_set_callback(cfs_socket_t *sock, ksock_conn_t *conn); -extern void ksocknal_lib_reset_callback(cfs_socket_t *sock, ksock_conn_t *conn); -extern void ksocknal_lib_push_conn (ksock_conn_t *conn); -extern int ksocknal_lib_get_conn_addrs (ksock_conn_t *conn); -extern int ksocknal_lib_setup_sock (cfs_socket_t *so); -extern int ksocknal_lib_send_iov (ksock_conn_t *conn, ksock_tx_t *tx); -extern int ksocknal_lib_send_kiov (ksock_conn_t *conn, ksock_tx_t *tx); -extern void ksocknal_lib_eager_ack (ksock_conn_t *conn); -extern int ksocknal_lib_recv_iov (ksock_conn_t *conn); -extern int ksocknal_lib_recv_kiov (ksock_conn_t *conn); -extern int ksocknal_lib_get_conn_tunables (ksock_conn_t *conn, int *txmem, - int *rxmem, int *nagle); +extern void ksocknal_thread_fini(void); +extern void ksocknal_launch_all_connections_locked(struct ksock_peer_ni *peer_ni); +extern struct ksock_conn_cb *ksocknal_find_connectable_conn_cb_locked(struct ksock_peer_ni *peer_ni); +extern struct ksock_conn_cb *ksocknal_find_connecting_conn_cb_locked(struct ksock_peer_ni *peer_ni); +extern int ksocknal_new_packet(struct ksock_conn *conn, int skip); +extern int ksocknal_scheduler(void *arg); +extern int ksocknal_connd(void *arg); +extern int ksocknal_reaper(void *arg); +int ksocknal_send_hello(struct lnet_ni *ni, struct ksock_conn *conn, + lnet_nid_t peer_nid, struct ksock_hello_msg *hello); +int ksocknal_recv_hello(struct lnet_ni *ni, struct ksock_conn *conn, + struct ksock_hello_msg *hello, + struct lnet_process_id *id, + __u64 *incarnation); +extern void ksocknal_read_callback(struct ksock_conn *conn); +extern void ksocknal_write_callback(struct ksock_conn *conn); + +extern int ksocknal_lib_zc_capable(struct ksock_conn *conn); +extern void ksocknal_lib_save_callback(struct socket *sock, struct ksock_conn *conn); +extern void ksocknal_lib_set_callback(struct socket *sock, struct ksock_conn *conn); +extern void ksocknal_lib_reset_callback(struct socket *sock, + struct ksock_conn *conn); +extern void ksocknal_lib_push_conn(struct ksock_conn *conn); +extern int ksocknal_lib_get_conn_addrs(struct ksock_conn *conn); +extern int ksocknal_lib_setup_sock(struct socket *so); +extern int ksocknal_lib_send_hdr(struct ksock_conn *conn, struct ksock_tx *tx, + struct kvec *scratch_iov); +extern int ksocknal_lib_send_kiov(struct ksock_conn *conn, struct ksock_tx *tx, + struct kvec *scratch_iov); +extern void ksocknal_lib_eager_ack(struct ksock_conn *conn); +extern int ksocknal_lib_recv_iov(struct ksock_conn *conn, + struct kvec *scratchiov); +extern int ksocknal_lib_recv_kiov(struct ksock_conn *conn, struct page **pages, + struct kvec *scratchiov); +extern int ksocknal_lib_get_conn_tunables(struct ksock_conn *conn, int *txmem, + int *rxmem, int *nagle); extern int ksocknal_tunables_init(void); -extern void ksocknal_tunables_fini(void); -extern int ksocknal_lib_tunables_init(void); -extern void ksocknal_lib_tunables_fini(void); -extern void ksocknal_lib_csum_tx(ksock_tx_t *tx); +extern void ksocknal_lib_csum_tx(struct ksock_tx *tx); -extern int ksocknal_lib_memory_pressure(ksock_conn_t *conn); -extern int ksocknal_lib_bind_thread_to_cpu(int id); +extern int ksocknal_lib_memory_pressure(struct ksock_conn *conn); #endif /* _SOCKLND_SOCKLND_H_ */