X-Git-Url: https://git.whamcloud.com/?a=blobdiff_plain;f=lustre%2Finclude%2Flustre_import.h;h=9338a97e4635784fd2bd8b286d996a102a24c064;hb=73e0d8cff3ff7fc6cd57badab434f9511b125704;hp=4f6e83f88f862211c0a6912f7d1203aa43e9870b;hpb=f81e107d0a1d486328cc4011770be7c1cd5fcddf;p=fs%2Flustre-release.git diff --git a/lustre/include/lustre_import.h b/lustre/include/lustre_import.h index 4f6e83f..9338a97 100644 --- a/lustre/include/lustre_import.h +++ b/lustre/include/lustre_import.h @@ -1,6 +1,4 @@ -/* -*- mode: c; c-basic-offset: 8; indent-tabs-mode: nil; -*- - * vim:expandtab:shiftwidth=8:tabstop=8: - * +/* * GPL HEADER START * * DO NOT ALTER OR REMOVE COPYRIGHT NOTICES OR THIS FILE HEADER. @@ -26,36 +24,71 @@ * GPL HEADER END */ /* - * Copyright 2008 Sun Microsystems, Inc. All rights reserved + * Copyright (c) 2002, 2010, Oracle and/or its affiliates. All rights reserved. * Use is subject to license terms. + * + * Copyright (c) 2011, Whamcloud, Inc. */ /* * This file is part of Lustre, http://www.lustre.org/ * Lustre is a trademark of Sun Microsystems, Inc. */ +/** \defgroup obd_import PtlRPC import definitions + * Imports are client-side representation of remote obd target. + * + * @{ + */ #ifndef __IMPORT_H #define __IMPORT_H +/** \defgroup export export + * + * @{ + */ + #include #include -/* Adaptive Timeout stuff */ +/** + * Adaptive Timeout stuff + * + * @{ + */ #define D_ADAPTTO D_OTHER #define AT_BINS 4 /* "bin" means "N seconds of history" */ #define AT_FLG_NOHIST 0x1 /* use last reported value only */ struct adaptive_timeout { - time_t at_binstart; /* bin start time */ - unsigned int at_hist[AT_BINS]; /* timeout history bins */ - unsigned int at_flags; - unsigned int at_current; /* current timeout value */ - unsigned int at_worst_ever; /* worst-ever timeout value */ - time_t at_worst_time; /* worst-ever timeout timestamp */ - spinlock_t at_lock; + time_t at_binstart; /* bin start time */ + unsigned int at_hist[AT_BINS]; /* timeout history bins */ + unsigned int at_flags; + unsigned int at_current; /* current timeout value */ + unsigned int at_worst_ever; /* worst-ever timeout value */ + time_t at_worst_time; /* worst-ever timeout timestamp */ + cfs_spinlock_t at_lock; }; +struct ptlrpc_at_array { + cfs_list_t *paa_reqs_array; /** array to hold requests */ + __u32 paa_size; /** the size of array */ + __u32 paa_count; /** the total count of reqs */ + time_t paa_deadline; /** the earliest deadline of reqs */ + __u32 *paa_reqs_count; /** the count of reqs in each entry */ +}; + +#define IMP_AT_MAX_PORTALS 8 +struct imp_at { + int iat_portal[IMP_AT_MAX_PORTALS]; + struct adaptive_timeout iat_net_latency; + struct adaptive_timeout iat_service_estimate[IMP_AT_MAX_PORTALS]; +}; + + +/** @} */ + +/** Possible import states */ enum lustre_imp_state { LUSTRE_IMP_CLOSED = 1, LUSTRE_IMP_NEW = 2, @@ -69,6 +102,7 @@ enum lustre_imp_state { LUSTRE_IMP_EVICTED = 10, }; +/** Returns test string representation of numeric import state \a state */ static inline char * ptlrpc_import_state_name(enum lustre_imp_state state) { static char* import_state_names[] = { @@ -81,71 +115,140 @@ static inline char * ptlrpc_import_state_name(enum lustre_imp_state state) return import_state_names[state]; } +/** + * List of import event types + */ enum obd_import_event { IMP_EVENT_DISCON = 0x808001, IMP_EVENT_INACTIVE = 0x808002, IMP_EVENT_INVALIDATE = 0x808003, IMP_EVENT_ACTIVE = 0x808004, IMP_EVENT_OCD = 0x808005, + IMP_EVENT_DEACTIVATE = 0x808006, + IMP_EVENT_ACTIVATE = 0x808007, }; +/** + * Definition of import connection structure + */ struct obd_import_conn { - struct list_head oic_item; + /** Item for linking connections together */ + cfs_list_t oic_item; + /** Pointer to actual PortalRPC connection */ struct ptlrpc_connection *oic_conn; + /** uuid of remote side */ struct obd_uuid oic_uuid; - __u64 oic_last_attempt; /* jiffies, 64-bit */ + /** + * Time (64 bit jiffies) of last connection attempt on this connection + */ + __u64 oic_last_attempt; }; -#define IMP_AT_MAX_PORTALS 8 -struct imp_at { - int iat_portal[IMP_AT_MAX_PORTALS]; - struct adaptive_timeout iat_net_latency; - struct adaptive_timeout iat_service_estimate[IMP_AT_MAX_PORTALS]; +/* state history */ +#define IMP_STATE_HIST_LEN 16 +struct import_state_hist { + enum lustre_imp_state ish_state; + time_t ish_time; }; +/** + * Defintion of PortalRPC import structure. + * Imports are representing client-side view to remote target. + */ struct obd_import { + /** Local handle (== id) for this import. */ struct portals_handle imp_handle; - atomic_t imp_refcount; + /** Reference counter */ + cfs_atomic_t imp_refcount; struct lustre_handle imp_dlm_handle; /* client's ldlm export */ + /** Currently active connection */ struct ptlrpc_connection *imp_connection; + /** PortalRPC client structure for this import */ struct ptlrpc_client *imp_client; - struct list_head imp_pinger_chain; - struct list_head imp_zombie_chain; /* queue for destruction */ + /** List element for linking into pinger chain */ + cfs_list_t imp_pinger_chain; + /** List element for linking into chain for destruction */ + cfs_list_t imp_zombie_chain; - /* Lists of requests that are retained for replay, waiting for a reply, + /** + * Lists of requests that are retained for replay, waiting for a reply, * or waiting for recovery to complete, respectively. + * @{ */ - struct list_head imp_replay_list; - struct list_head imp_sending_list; - struct list_head imp_delayed_list; + cfs_list_t imp_replay_list; + cfs_list_t imp_sending_list; + cfs_list_t imp_delayed_list; + /** @} */ + /** obd device for this import */ struct obd_device *imp_obd; + + /** + * some seciruty-related fields + * @{ + */ struct ptlrpc_sec *imp_sec; - struct semaphore imp_sec_mutex; + cfs_mutex_t imp_sec_mutex; cfs_time_t imp_sec_expire; + /** @} */ + + /** Wait queue for those who need to wait for recovery completion */ cfs_waitq_t imp_recovery_waitq; - atomic_t imp_inflight; - atomic_t imp_unregistering; - atomic_t imp_replay_inflight; - atomic_t imp_inval_count; + /** Number of requests currently in-flight */ + cfs_atomic_t imp_inflight; + /** Number of requests currently unregistering */ + cfs_atomic_t imp_unregistering; + /** Number of replay requests inflight */ + cfs_atomic_t imp_replay_inflight; + /** Number of currently happening import invalidations */ + cfs_atomic_t imp_inval_count; + /** Numbner of request timeouts */ + cfs_atomic_t imp_timeouts; + /** Current import state */ enum lustre_imp_state imp_state; + /** History of import states */ + struct import_state_hist imp_state_hist[IMP_STATE_HIST_LEN]; + int imp_state_hist_idx; + /** Current import generation. Incremented on every reconnect */ int imp_generation; + /** Incremented every time we send reconnection request */ __u32 imp_conn_cnt; + /** + * \see ptlrpc_free_committed remembers imp_generation value here + * after a check to save on unnecessary replay list iterations + */ int imp_last_generation_checked; + /** Last tranno we replayed */ __u64 imp_last_replay_transno; + /** Last transno committed on remote side */ __u64 imp_peer_committed_transno; + /** + * \see ptlrpc_free_committed remembers last_transno since its last + * check here and if last_transno did not change since last run of + * ptlrpc_free_committed and import generation is the same, we can + * skip looking for requests to remove from replay list as optimisation + */ __u64 imp_last_transno_checked; + /** + * Remote export handle. This is how remote side knows what export + * we are talking to. Filled from response to connect request + */ struct lustre_handle imp_remote_handle; - cfs_time_t imp_next_ping; /* jiffies */ - __u64 imp_last_success_conn; /* jiffies, 64-bit */ + /** When to perform next ping. time in jiffies. */ + cfs_time_t imp_next_ping; + /** When we last succesfully connected. time in 64bit jiffies */ + __u64 imp_last_success_conn; - /* all available obd_import_conn linked here */ - struct list_head imp_conn_list; + /** List of all possible connection for import. */ + cfs_list_t imp_conn_list; + /** + * Current connection. \a imp_connection is imp_conn_current->oic_conn + */ struct obd_import_conn *imp_conn_current; - /* Protects flags, level, generation, conn_cnt, *_list */ - spinlock_t imp_lock; + /** Protects flags, level, generation, conn_cnt, *_list */ + cfs_spinlock_t imp_lock; /* flags */ unsigned long imp_no_timeout:1, /* timeouts are disabled */ @@ -154,13 +257,15 @@ struct obd_import { imp_replayable:1, /* try to recover the import */ imp_dlm_fake:1, /* don't run recovery (timeout instead) */ imp_server_timeout:1, /* use 1/2 timeout on MDS' OSCs */ - imp_initial_recov:1, /* retry the initial connection */ - imp_initial_recov_bk:1, /* turn off init_recov after trying all failover nids */ + imp_delayed_recovery:1, /* VBR: imp in delayed recovery */ + imp_no_lock_replay:1, /* VBR: if gap was found then no lock replays */ + imp_vbr_failed:1, /* recovery by versions was failed */ imp_force_verify:1, /* force an immidiate ping */ imp_pingable:1, /* pingable */ imp_resend_replay:1, /* resend for replay */ - imp_recon_bk:1, /* turn off reconnect if all failovers fail */ - imp_last_recon:1; /* internally used by above */ + imp_no_pinger_recover:1,/* disable normal recovery, for test only. */ + imp_need_mne_swab:1, /* need IR MNE swab */ + imp_force_reconnect:1; /* import must be reconnected instead of chouse new connection */ __u32 imp_connect_op; struct obd_connect_data imp_connect_data; __u64 imp_connect_flags_orig; @@ -178,8 +283,14 @@ struct obd_import { typedef void (*obd_import_callback)(struct obd_import *imp, void *closure, int event, void *event_arg, void *cb_data); +/** + * Structure for import observer. + * It is possible to register "observer" on an import and every time + * something happens to an import (like connect/evict/disconnect) + * obderver will get its callback called with event type + */ struct obd_import_observer { - struct list_head oio_chain; + cfs_list_t oio_chain; obd_import_callback oio_cb; void *oio_cb_data; }; @@ -205,18 +316,22 @@ static inline unsigned int at_timeout2est(unsigned int val) return (max((val << 2) / 5, 5U) - 4); } -static inline void at_init(struct adaptive_timeout *at, int val, int flags) { - memset(at, 0, sizeof(*at)); +static inline void at_reset(struct adaptive_timeout *at, int val) { at->at_current = val; at->at_worst_ever = val; at->at_worst_time = cfs_time_current_sec(); - at->at_flags = flags; - spin_lock_init(&at->at_lock); } +static inline void at_init(struct adaptive_timeout *at, int val, int flags) { + memset(at, 0, sizeof(*at)); + cfs_spin_lock_init(&at->at_lock); + at->at_flags = flags; + at_reset(at, val); +} +extern unsigned int at_min; static inline int at_get(struct adaptive_timeout *at) { - return at->at_current; + return (at->at_current > at_min) ? at->at_current : at_min; } -int at_add(struct adaptive_timeout *at, unsigned int val); +int at_measured(struct adaptive_timeout *at, unsigned int val); int import_at_get_index(struct obd_import *imp, int portal); extern unsigned int at_max; #define AT_OFF (at_max == 0) @@ -226,4 +341,8 @@ struct obd_export; extern struct obd_import *class_exp2cliimp(struct obd_export *); extern struct obd_import *class_conn2cliimp(struct lustre_handle *); +/** @} import */ + #endif /* __IMPORT_H */ + +/** @} obd_import */